F26 Epigenetics Topic 5: DNA Methylation

Overview and Chemical Basis of DNA Methylation

Chemical reaction of cytosine methylation and demethylation

Definition of DNA Methylation:
- DNA methylation is an epigenetic modification involving the covalent addition of a methyl group ( $\text{--CH}_3$ ) to the carbon-5 position of the cytosine ring, forming 5-methylcytosine ( $\text{5mC}$ ).
Enzymatic Catalysis:
- The addition of the methyl group is catalyzed by a specialized class of enzymes designated as DNA methyltransferases (DNMTs).
Methyl Donors and Biosynthetic Pathway:
- Direct Methyl Donor: S-adenosyl-methionine (SAM / AdoMet) serves as the direct methyl donor in the enzymatic reaction.
- Reaction Byproduct: Transfer of the methyl group from AdoMet to cytosine yields S-adenosyl-homocysteine (AdoHcy).
- Chemical Reaction: $\text{Cytosine} + \text{AdoMet} \xrightarrow{\text{DNA methyltransferase}} \text{5-methylcytosine} + \text{AdoHcy}$
- Demethylation Reaction: $\text{5-methylcytosine} \xrightarrow{\text{DNA demethylase}} \text{Cytosine}$
- Dietary Precursors: The cellular pool of SAM is directly sustained by essential dietary methyl nutrients, including:
  - Methionine
  - Folate
  - Betaine
  - Choline

Classification of DNA Methylation by Sequence Context

CG Methylation:
- Represents the predominant form of DNA methylation in mammalian genomes.
- Approximately $70\text{--}80\%$ of all CG dinucleotide sites across the mammalian genome are methylated.
- Sequence Symmetry: CG sites are palindromic/symmetric ( $5'\text{--CG--}3'$ mirrored by $3'\text{--GC--}5'$ on the complementary strand).
CHG Methylation:
- Context notation: $\text{H} = \text{A, T, or C}$ .
- Primarily specific to plant genomes.
- Sequence Symmetry: CHG sites are palindromic/symmetric ( $5'\text{--CHG--}3'$ mirrored by $3'\text{--GHC--}5'$ on the complementary strand).
CHH Methylation:
- Context notation: $\text{H} = \text{A, T, or C}$ .
- Predominantly found in plants and present at extremely low levels ( $\sim 2\%$ ) in mammalian cells.
- Sequence Asymmetry: CHH sites are non-palindromic/asymmetric.
Mechanistic Significance of Sequence Symmetry:
- Symmetric Contexts (CG and CHG): The identical sequence context on both antiparallel DNA strands ensures that following semi-conservative replication, the pre-existing parental strand serves as a spatial template to direct accurate maintenance methylation onto the newly synthesized daughter strand.
- Asymmetric Contexts (CHH): Because the complementary strand lacks an opposing cytosine in the identical context, replication inherently leaves no methylated partner strand template. Consequently, CHH methylation cannot be copied directly by maintenance methyltransferases and must be re-established de novo during each cell cycle.

Dynamics of De Novo and Maintenance Methylation

Overview of de novo, maintenance, and active demethylation pathways

De Novo Methylation:
- Refers to the targeted placement of novel methyl groups onto previously unmethylated double-stranded DNA substrates.
- Establishes new epigenetic patterns during embryonic development, cell differentiation, and germline specification.
Semi-Conservative Replication and Hemimethylated DNA:
- During DNA replication, the double helix unwinds and each parental strand guides the synthesis of a complementary daughter strand.
- Because newly incorporated nucleotides are unmethylated, replication temporarily converts fully methylated DNA into hemimethylated DNA (where only the parental strand retains methyl marks).
Maintenance Methylation:
- Recognizes hemimethylated DNA intermediates at the replication fork and selectively methylates the newly synthesized daughter strand, restoring full symmetrical methylation.
- Preserves epigenetic memory faithfully across somatic cell divisions.
Passive Dilution vs. Active Demethylation:
- Passive Dilution: Occurs when maintenance methyltransferases fail to restore methyl groups across successive replication cycles, leading to progressive, cell-division-dependent loss of DNA methylation.
- Active Demethylation: Enzymatic removal of methyl groups independent of DNA replication. TET (Ten-Eleven Translocation) dioxygenases catalyze the oxidation of 5-methylcytosine to 5-hydroxymethylcytosine (5hmC) as an intermediate step toward full demethylation.

CpG Islands (CGIs) and Unmethylated Context Recognition

CpG island promoter architecture in active vs silenced states

Genomic distribution of Kdm2b binding and domain structures of CXXC domain proteins

Characteristics of CpG Islands (CGIs):
- Defined as genomic regions spanning $20\,\text{bp}$ to $2\,\text{kb}$ in length characterized by a significantly higher density of CG dinucleotides compared to the genome average.
- Methylation Status: CGIs are generally maintained in an unmethylated (hypomethylated) state despite the surrounding genome being globally hypermethylated.
- Promoter Localization: Approximately $60\%$ of human gene promoters contain CpG islands.
- Tissue-Specific Gene Promoters: Promoters regulating tissue-specific genes typically contain a lower density of CGIs (low CGI density) and are heavily methylated (hypermethylated) in non-expressing cell types.
CXXC Domain Proteins:
- Proteins containing the specialized CXXC zinc-finger domain specifically recognize and bind unmethylated CpG motifs within CGIs.
- Key Examples: KDM2B, KDM2A, and CFP1.
- Repressive Recruited Functions: CXXC-containing proteins like KDM2B bind unmethylated CGIs (such as those across the Hoxa developmental gene cluster, including Hoxa1 through Hoxa11) and recruit chromatin-modifying complexes that deposit repressive marks to achieve gene silencing or regulate chromatin organization.
Domain Structures of CXXC-Containing Factors:
- Cfp1 (656 amino acids): Contains PHD1, CXXC, A, B, S, C, and PHD2 domains.
- Kdm2a (1162 amino acids): Contains JmjC, CXXC, and PHD domains.
- Kdm2b (1336 amino acids): Contains JmjC, CXXC, PHD, F-box, and LRR domains.

Models Explaining CGI Escape from Methylation

Three models explaining CGI resistance to methylation

Model A: Direct Inhibition:
- Transcription factors or CGI-binding protective proteins occupy the unmethylated CGI promoter, physically blocking DNA methyltransferases (DNMTs) from accessing the underlying DNA sequence.
Model B: Active Demethylation:
- DNA methyltransferases stochastically methylate CGI cytosines, but dedicated DNA demethylases (DM / TET enzymes) continuously and rapidly remove these methyl groups, sustaining a hypomethylated steady state.
Model C: Steric Hindrance / Binding Inhibition via H3K4me3 and DNMT3L:
- Active transcription at CGIs maintains high levels of histone H3 lysine 4 trimethylation (H3K4me3).
- The regulatory factor DNMT3L functions as a molecular sensor for histone H3 tail modifications. Unmethylated H3K4 recruits DNMT3L to stimulate de novo methylation by DNMT3A/3B.
- The presence of the H3K4me3 mark sterically inhibits DNMT3L binding, thereby preventing the recruitment and activation of the de novo methylation machinery at active CpG islands.

Functional Impact on Promoters and Enhancers

Promoter methylation responses in LCP, ICP, and HCP classes

Promoter Classification by CpG Density:
- Low CpG Promoters (LCP): Contain low CpG density. LCPs remain active regardless of their DNA methylation status.
- Intermediate CpG Promoters (ICP): Contain intermediate CpG density. ICPs acquire differentiation-dependent methylation during developmental choices and are suppressed when methylated, making them central to tissue-specific expression programs.
- High CpG Promoters (HCP): Characterized by high CpG density (CGIs). HCPs are almost universally unmethylated in normal somatic tissues, but undergo profound, robust silencing when hypermethylated.
Enhancer Methylation Dynamics:
- Hypomethylation at distal enhancer elements directly correlates with elevated target gene transcription.
- Dynamic methylation changes at enhancers are often more potent determinants of cell-type-specific gene expression programs than promoter methylation changes.

Genomic Targets and Physiological Consequences of DNA Methylation

Mechanism of transposon silencing via DNA methylation

Centromeric repeat methylation and X-chromosome inactivation

Comparison of Mendelian and Imprinted gene expression

Lifecycle and maintenance of genomic imprints

Transposon Silencing:
- Transposable elements ("jumping genes") constitute up to $40\%$ of the mammalian genome and an even higher fraction of plant genomes (e.g., maize/corn).
- Unchecked transposition threatens genomic integrity by causing insertional mutagenesis (gene knockouts), structural rearrangements, or altered expression of neighboring host genes.
- Transposons are routinely silenced by dense DNA methylation. Loss of methylation triggers transposon reactivation, increased transposition, and abnormal transcription of adjacent genomic regions.
Centromeric Repeat Methylation and ICF Syndrome:
- Centromeric repetitive DNA sequences possess intrinsic promoter activity capable of transcription but are held in a transcriptionally silent heterochromatic state by dense DNA methylation.
- Loss of centromeric DNA methylation causes ICF syndrome (Immunodeficiency, Centromeric instability, and Facial anomalies syndrome).
X-Chromosome Inactivation:
- Serves as a dosage compensation mechanism in female mammals to equalize X-linked gene expression relative to males.
- DNA methylation reinforces long-term gene silencing on the inactive X chromosome ( $\text{Xi}$ ).
- Because DNA methylation acts in concert with multiple redundant silencing marks (such as Xist RNA recruitment and repressive histone modifications), targeted loss of DNA methylation results in only partial or moderate reactivation of $\text{Xi}$ -linked genes.
Genomic Imprinting:
- Genomic imprinting is an epigenetic process resulting in monoallelic, parent-of-origin-specific gene expression.
- Either the maternal or paternal allele is selectively methylated and silenced during gametogenesis (oogenesis vs. spermatogenesis).
- The inherited parental methylation pattern (imprint) is preserved in somatic cells of the offspring following fertilization, while primordial germ cells erase and re-establish the imprint based on the biological sex of the individual.
- Disruption of imprinting patterns causes severe developmental disorders.

Molecular Mechanisms of Gene Silencing by DNA Methylation

Chromatin alteration between open active state and condensed silent state

Direct interference with transcription factor binding by DNA methylation

MeCP2 recruitment of corepressor complex and HDACs

Mechanism 1: Alteration of Chromatin Architecture:
- Hypomethylated DNA / Open Euchromatin: Low DNA methylation correlates with an open, accessible chromatin configuration, characterized by acetylated histones, binding of SWI/SNF chromatin-remodeling complexes, recruitment of Histone Acetyltransferases (HATs), and engagement of RNA Polymerase II with transcription factors/co-activators.
- Hypermethylated DNA / Condensed Heterochromatin: Heavy DNA methylation promotes chromatin condensation into closed heterochromatin, characterized by recruitment of Histone Deacetylases (HDACs) and Histone Methyltransferases (HMTs), leading to deacetylated histones and transcriptional arrest.
Mechanism 2: Direct Inhibition of Transcription Factor Binding:
- Methylation at specific cytosine bases within transcription factor recognition motifs physically sterically hinders factor binding in vitro.
- Transcription factors directly obstructed by DNA methylation include:
  - AP-2
  - c-MYC
  - E2F
Mechanism 3: Recruitment of Repressive Corepressor Complexes via MBD Proteins:
- Methylated cytosines serve as binding platforms for specialized Methyl-CpG-Binding Domain (MBD) proteins, including MeCP1 and MeCP2.
- MeCP2 Action: MeCP2 selectively binds methylated DNA (e.g., $5'\text{--tcaagcgcgcgcgcgcta--}3'$ ) and recruits corepressor complexes containing HDACs and corepressor proteins.
- This physical bridging converts a primary DNA methylation mark directly into repressive histone modifications (deacetylation), driving local chromatin condensation and gene silencing.

Enzymatic Machinery Across Kingdoms

De Novo DNA Methylation:
- Mammals: Catalyzed by DNMT3A and DNMT3B.
- Plants: Catalyzed by DOMAINS REARRANGED METHYLASES 1 and 2 (DRM1 & DRM2).
- RNA-Mediated Pathways: Non-coding RNA pathways, specifically RNA-directed DNA methylation (RdDM) and RNA interference (RNAi), direct de novo methyltransferases to specific genomic loci.
CG Methylation Maintenance:
- Mammals: Executed by DNA METHYLTRANSFERASE 1 (DNMT1).
- Plants: Executed by METHYLTRANSFERASES 1 (MET1).
CHG Methylation Maintenance:
- Plants (Plant-Specific): Maintained by CHROMETHYLASES 3 (CMT3).
CHH Methylation Maintenance / Re-establishment:
- Plants: Maintained via continuous de novo action of DRM2 guided by RdDM (and partially sustained by CMT3).

Targeting Machinery and Epigenetic Cross-Talk

Histone Modification Cross-Talk:
- H3K4 Methylation Blocks De Novo DNA Methylation:
  - The regulatory protein Dnmt3L acts as a molecular sensor for histone H3 lysine 4 (H3K4) methylation status.
  - In the absence of H3K4 methylation (unmethylated H3K4), Dnmt3L binds the histone tail and recruits DNMT3A/3B to execute de novo DNA methylation.
  - If H3K4 is methylated (e.g., H3K4me3 at active promoters), Dnmt3L binding is blocked, suppressing de novo DNA methylation.
- Mutual Reinforcement of H3K9 and CHG Methylation in Plants:
  - KRYPTONITE, an H3K9 methyltransferase, specifically binds methylated CHG DNA sites and catalyzes histone H3 lysine 9 methylation (H3K9me2).
  - CMT3 (Chromomethylase 3) contains chromo and BAH domains that specifically recognize nucleosomes containing H3K9 dimethylation and catalyze CHG DNA methylation.
  - This creates a self-reinforcing forward feedback loop that locks target plant genomic regions into a silenced state.

Non-Coding RNA-Directed DNA Methylation and Gene Silencing

Overview of Non-Coding RNAs (ncRNAs):
- Small non-coding RNA molecules generated from double-stranded RNA (dsRNA) or long single-stranded RNA precursors that do not encode proteins.
- Major Structural Classes:
  - Small interfering RNAs (siRNA)
  - MicroRNAs (miRNA)
  - Piwi-interacting RNAs (piRNA)
  - Long non-coding RNAs (lncRNA)
Mechanisms of Gene Silencing:
- Transcriptional Gene Silencing (TGS): ncRNAs guide DNA methyltransferases and histone-modifying enzymes directly to complementary DNA sequences via base-pairing interactions (RNA-directed DNA methylation / RdDM), establishing stable covalent DNA and histone marks.
- Post-Transcriptional Gene Silencing (PTGS): ncRNAs target mRNA transcripts in the cytoplasm for cleavage, degradation, or translational repression.

Summary of Learning Objectives

Objective 1 (Definition, Enzymes, Donor, Precursors): DNA methylation is the addition of a methyl group to cytosine catalyzed by DNA methyltransferases (DNMTs). SAM (AdoMet) is the direct methyl donor, synthesized via dietary inputs including methionine, folate, betaine, and choline.
Objective 2 (Sequence Contexts & Symmetry): CG predominates in mammals; CHG and CHH occur in plants. Palindromic symmetry in CG and CHG allows post-replication maintenance via hemimethylated templates, whereas asymmetric CHH lacks complementary strand templates and requires continuous re-establishment.
Objective 3 (De Novo vs. Maintenance Machinery): De novo methylation (DNMT3A/3B in mammals; DRM1/2 in plants) establishes new marks; maintenance methylation (DNMT1 in mammals; MET1 in plants; CMT3 for CHG in plants) copies marks onto hemimethylated DNA post-replication.
Objective 4 (CpG Islands & CXXC Recognition): CGIs are hypomethylated promoter stretches present in \n\sim 60\%\n of human genes. Tissue-specific promoters have lower CGI density and undergo developmental hypermethylation. Unmethylated CGIs are recognized by CXXC domain proteins (e.g., KDM2B) that recruit repressive marks.
Objective 5 (CGI Protection Models): Three proposed models for CGI methylation resistance include direct factor inhibition, active demethylation (TET/DM), and steric hindrance via H3K4me3 blocking Dnmt3L sensor recruitment.
Objective 6 (Promoters & Enhancers): LCPs remain active regardless of methylation; ICPs acquire tissue-specific developmental methylation; HCPs are silenced when methylated. Enhancer hypomethylation strongly correlates with gene activation and frequently exerts greater transcriptional control than promoter methylation.
Objective 7 (Functional Consequences of Loss): Methylation represses transposons (preventing genomic instability/transposition), centromeric repeats (loss causes ICF syndrome), $\text{Xi}$ genes (loss causes partial reactivation), and imprinted alleles (enforcing parent-of-origin expression).
Objective 8 (Mechanisms of Transcriptional Change): Methylation alters gene expression by condensing chromatin (heterochromatin formation), directly blocking transcription factor binding (AP-2, c-MYC, E2F), or recruiting MBD proteins (MeCP1/2) that assemble HDAC-corepressor complexes.
Objective 9 (Epigenetic Targeting & Cross-Talk): De novo methylation is regulated by histone marks: H3K4me3 blocks the Dnmt3L sensor, whereas plant H3K9me2 and CHG methylation form a self-reinforcing feedback loop via KRYPTONITE and CMT3.
Objective 10 (Non-Coding RNAs & Silencing): Non-coding RNAs (siRNA, miRNA, piRNA, lncRNA) target specific sequence contexts via RdDM to establish TGS (DNA methylation) or mediate PTGS (mRNA cleavage/translation repression).