The Dynamic Architecture of Euchromatin, Heterochromatin, and ATP-Dependent Chromatin Remodeling: From Multi-Scale Polymer Mechanics to Mendelian Chromatinopathies

1. Introduction & Paradigm Shift: From Cytological Dichotomy to Multi-Scale Polymer Physics

Nearly a century ago, Emil Heitz (1928) introduced the foundational cytological distinction between euchromatin—the lightly staining nuclear fraction that decondenses during interphase telophase—and heterochromatin, the densely staining, refractory nuclear compartments that remain compacted throughout the cell cycle (Passarge, 1979; Yunis & Yasmineh, 1971). For decades, textbook models interpreted this dichotomy through a rigid structural hierarchy: 147 base pairs of DNA wrapped in 1.65 left-handed superhelical turns around a histone octamer (Luger et al., 1997) were proposed to fold sequentially into regular 11-nm 'beads-on-a-string' arrays, 30-nm solenoidal or two-start zigzag fibers (Routh et al., 2008), 120-nm chromonema fibers, and 300–700-nm metaphase chromatids. In this classical framework, euchromatin was viewed as an unfolded 11-nm template passively permissive to RNA polymerase II (Pol II), whereas heterochromatin was conceptualized as acrystalline 30-nm steric barrier that physically occluded transcription factors.

Over the past fifteen years, cryo-electron microscopy of vitreous sections (CEMOVIS), small-angle X-ray scattering (SAXS), super-resolution single-nucleosome tracking, and ChromEMT (chromatin electron microscopy tomography) have overturned the hierarchical 30-nm fiber paradigm in living mammalian cells (Ou et al., 2017; Maeshima et al., 2016). Using an osmiophilic diaminobenzidine (DAB) photo-oxidation polymer selectively deposited onto DRAQ5-stained DNA, Ou et al. (2017) reconstructed the three-dimensional ultrastructure of individual chromatin chains across interphase and mitotic human cells at nanometer resolution. Their tomograms demonstrated that in situ chromatin does not form regular 30-nm fibers in either active or silenced compartments; instead, the genome exists as a flexible, disordered 5-to-24-nm curvilinear chain that bends at variable persistence lengths and packs at distinct three-dimensional chromatin volume concentrations (CVC)—ranging from 12–20% in euchromatic territories to 35–50% in lamina-associated and pericentromeric heterochromatin (Ou et al., 2017). Consequently, functional genome partitioning is governed not by crystalline secondary fiber geometry, but by local nucleosome packing density, viscoelastic polymer dynamics, and the local balance between ATP-consuming chromatin motors and multivalent reader-writer phase separation (Gibson et al., 2019; Strom et al., 2017; Larson et al., 2017).

The compartment distinction is also written in protein composition. A systematic comparison of the molecular complexes resident at euchromatin, heterochromatin, and centromeric chromatin shows that each compartment is constituted by a distinct repertoire of chromatin factors and complexes, and that these repertoires in turn shape how accessible the underlying DNA is (Morrison & Thakur, 2021).

1.1 Physical Foundations of the Nucleosome Core Particle and Disordered 10-nm Chains

At the base of the chromatin polymer lies the nucleosome core particle (NCP), whose 2.8 Å X-ray crystal structure by Luger et al. (1997) revealed 146–147 bp of B-form DNA wrapped 1.65 times around a central (H3–H4)₂ tetramer flanked by two H2A–H2B dimers. The octamer is stabilized by histone-fold handshake motifs and four-helix bundles, while the minor groove of the DNA superhelix is electrostatically anchored every 10.2–10.5 bp at fourteen superhelical locations (SHL ± 0.5 to ± 6.5) via arginine side chains that insert into AT-rich minor grooves (Luger et al., 1997). Despite this formidable electrostatic stability (net octamer charge of approximately +148 e neutralizing roughly half of the -294 e DNA phosphate backbone charge), single-molecule Förster resonance energy transfer (smFRET) and optical trap unzipping experiments demonstrate that nucleosomes undergo spontaneous thermal site exposure ('DNA breathing'). Entry-exit DNA segments (SHL ± 6.5 to ± 4.5) transiently unwrap from the histone surface on timescales of τ_unwrap ≈ 10–250 ms with an equilibrium accessibility probability of K_eq ≈ 10⁻²–10⁻¹, whereas the dyad region (SHL 0) exhibits K_eq < 10⁻⁴ and requires active ATP-dependent remodeling to become accessible (Clapier & Cairns, 2009; Narlikar et al., 2013).

Crucially, two structural features of the NCP mediate inter-nucleosomal packing within the crowded nucleoplasm (~100–400 mg/mL macromolecular density). First, the H2A–H2B acidic patch—a negatively charged cleft formed by six glutamate and two aspartate residues (E56, E61, E64, D90, E91, E92 on H2A; E102, E110 on H2B)—serves as the universal docking surface for the positively charged basic tail of histone H4 (residues K16–R23, KRHRKVLR) from an adjacent nucleosome, as well as for chromatin remodelers (ISWI, INO80, BAF) and viral tethering factors (Luger et al., 1997; Shogren-Knaak et al., 2006). Second, the ten intrinsically disordered N- and C-terminal histone tails project through minor-groove channels beyond the DNA gyres, acting as electrostatic polyelectrolyte bridges and modular scaffolds for covalent post-translational modifications (Strahl & Allis, 2000; Kouzarides, 2007). In physiological divalent cation concentrations (≥ 1–2 mM Mg²⁺) and in the presence of macromolecular crowders, inter-digitation of 10-nm fibers via H4 tail–acidic patch contacts drives polymer melt condensation without requiring regular 30-nm helical symmetry (Maeshima et al., 2016; Gibson et al., 2019).

1.2 Resolution of the 30-nm Fiber Controversy: Cryo-EM, SAXS, and Ionic Crowding

Understanding why reconstituted nucleosome arrays readily fold into regular 30-nm two-start zigzag or interdigitated solenoidal fibers in dilute buffers (Routh et al., 2008) whereas native mammalian interphase and mitotic chromosomes lack 30-nm fibers in situ (Ou et al., 2017; Maeshima et al., 2016) requires analyzing polyelectrolyte polymer physics at physiological chromatin volume concentrations (CVC = 20–50%, corresponding to 80–250 mg/mL nucleosome density). In dilute solution (<0.5 mg/mL), intra-fiber H4 tail–acidic patch interactions and linker histone H1 stem formation occur exclusively between i and i+1 or i+2 nucleosomes along the same DNA chain, enforcing regular helical periodicity that generates a prominent 30-nm X-ray scattering peak. However, once nucleosome arrays are concentrated above the polymer overlap threshold (c^ ≈ 15–30 mg/mL), inter-fiber nucleosome contacts become entropically and enthalpically favored over intra-fiber folding (Maeshima et al., 2016). Just as concentrated linear polymers form an interpenetrating polymer melt* in which intra-chain excluded-volume and short-range helical correlations are screened out, inter-digitation of histone tails between neighboring 10-nm chains dissolves regular 30-nm periodicity into a disordered, liquid-like 10-nm meshwork.

Furthermore, native genomic chromatin differs fundamentally from homogeneous 601 Widom tandem repeats in three biochemical respects that frustrate regular 30-nm folding: (1) heterogeneous nucleosome repeat lengths (NRLs) that vary from ~165 bp in actively transcribed domains to ~195–205 bp in constitutive heterochromatin; (2) heterogeneous combinations of core histone variants (H2A.Z, H3.3, macroH2A, CENP-A), non-histone architectural proteins (HMGB1/2, HMGN1/2, PARP1, MeCP2), and dynamic H4K16ac patches (Shogren-Knaak et al., 2006); and (3) continuous mechanical torque and nucleosome sliding generated by RNA polymerase II transcription and ATP-dependent chromatin remodelers (Clapier & Cairns, 2009). Thus, rather than relying on a crystalline 30-nm secondary structure, living cells regulate genome accessibility by modulating the mesh size (ξ ≈ 15–40 nm) and viscoelastic relaxation time (τ,relax) of interdigitated 10-nm clutches (Ou et al., 2017; Maeshima et al., 2016).

2. Hierarchical and Polymer-Physical Organization of Euchromatin and the 3D Genome

Above the scale of single nucleosomes and local 10-nm clutches (4–12 nucleosomes revealed by STORM super-resolution imaging; Ricci et al., 2015), the interphase genome is organized by two orthogonal physical mechanisms that operate simultaneously but independently: ATP-dependent cohesin loop extrusion, which generates kilobase-to-megabase topologically associating domains (TADs), and affinity-driven block-copolymer microphase separation, which segregates chromosomes into multi-megabase A (euchromatic) and B (heterochromatic) compartments (Lieberman-Aiden et al., 2009; Fudenberg et al., 2016; Rao et al., 2017; Nora et al., 2017).

2.1 Cohesin-Mediated Loop Extrusion and CTCF Boundary Insulation

High-resolution chromosome conformation capture (Hi-C) maps reveal self-interacting genomic blocks of 200 kb to 2 Mb—termed TADs or loop domains—across which enhancer-promoter contact probabilities are elevated 2-to-5-fold relative to inter-domain loci of equal linear genomic separation s (Lieberman-Aiden et al., 2009; Fudenberg et al., 2016). As formalized by the loop extrusion model of Fudenberg et al. (2016), the ring-shaped SMC complex cohesin (comprising SMC1, SMC3, RAD21, and STAG1/2, loaded by NIPBL-MAU2) binds chromatin and extrudes progressively larger cis-DNA loops at ATP-dependent rates of v_ext ≈ 0.5–1.5 kb·s⁻¹ until arrested by the 11-zinc-finger insulator protein CTCF bound in a convergent motif orientation (where the N-terminus of CTCF, containing the YDF motif that binds the STAG1/2–RAD21 interface, faces the approaching cohesin motor). The steady-state architecture of a loop extrusion domain is governed by the dimensionless ratio of the mean extruded loop length λ = v_ext · τ_res (where τ_res ≈ 15–25 min is the WAPL-regulated cohesin residence time) to the mean spacing d between loaded cohesin motors:

λ = v_ext · τ_res ≈ 120–300 kb, P(s) ∝ s^(−γ) (γ ≈ 0.75 inside TADs vs. γ ≈ 1.08–1.25 between TADs)

Definitive causal proof that loop extrusion and compartmentalization represent mechanistically distinct processes emerged from acute auxin-inducible degron (AID) experiments. When Rao et al. (2017) acutely degraded the cohesin kleisin subunit RAD21 in human cells, all loop domains and TAD boundaries vanished within 60 minutes, yet plaid A/B compartmentalization actually strengthened—demonstrating that active cohesin extrusion continually mixes and counteracts fine-scale epigenetic phase segregation. Conversely, when Nora et al. (2017) acutely degraded CTCF in mouse embryonic stem cells, TAD boundary insulation collapsed and enhancer-promoter specificity was lost across domain borders, whereas large-scale A/B compartments remained intact.

2.2 Euchromatic Biochemical Signatures, Enhancer Hubs, and Viscoelastic Chromatin Motion

Within the A compartment (located preferentially in the nuclear interior and surrounding nuclear speckles), euchromatin is defined by a characteristic biochemical landscape that lowers inter-nucleosomal attraction and facilitates transcriptional initiation and elongation (Heintzman et al., 2009; Buenrostro et al., 2013; Kouzarides, 2007). Active promoters feature a 150–300 bp nucleosome-depleted region (NDR)—quantified genome-wide via ATAC-seq hyperactive Tn5 transposase insertion (Buenrostro et al., 2013)—flanked by well-positioned -1 and +1 nucleosomes enriched in the replication-independent histone variants H2A.Z and H3.3, alongside high levels of H3K4me3 (deposited by COMPASS/MLL1–4 complexes), H3K9ac, and H4K16ac. Distal enhancers and super-enhancers are demarcated by monomethylated H3K4me1 (deposited by MLL3/MLL4 [KMT2C/KMT2D]) coupled with H3K27ac (catalyzed by p300/CBP; Heintzman et al., 2009). Across actively transcribed gene bodies, SetD2-mediated H3K36me3 recruits DNMT3B to maintain intragenic CpG methylation and suppresses cryptic intragenic transcription initiation.

Biophysically, single-nucleosome tracking shows that euchromatic chromatin behaves as a subdiffusive viscoelastic polymer network whose mean-squared displacement scales as MSD(Δ t) = 4 D_α Δ t^α with an anomalous exponent α ≈ 0.45–0.55 (consistent with a Rouse polymer subject to topological and cross-linking constraints; Maeshima et al., 2016). Remarkably, active transcription and super-enhancer condensate assembly (mediated by the intrinsically disordered regions of BRD4, MED1, and Pol II C-terminal domain) can locally tether euchromatic loops into condensed yet liquid-like transcription hubs, constraining long-range diffusion while maintaining rapid internal factor exchange (Gibson et al., 2019).

Accessibility measurements in living cells reinforce this dynamic picture. Using an in vivo methyltransferase accessibility assay in living human cells, Prajapati et al. (2025) found that both euchromatin and heterochromatin are fully accessible at the nucleosome level, with only centromeric α-satellite DNA partly inaccessible; the far lower accessibilities reported previously appear to be a property of isolated nuclei rather than of living cells, indicating that nucleosomes remain highly dynamic in vivo even within heterochromatic territories.

2.3 Nuclear Lamina-Associated Domains (LADs), Nucleolar-Associated Domains (NADs), and Nuclear Speckles

Spatial positioning of euchromatic and heterochromatic compartments within the spherical or ellipsoidal geometry of the interphase nucleus is anchored by two major structural scaffolds: the nuclear lamina at the periphery and nuclear bodies (nucleoli, SON/SRRM2-rich nuclear speckles, and PML/Cajal bodies) in the interior (Lieberman-Aiden et al., 2009). DamID mapping of Lamin B1 contacts has demonstrated that ~35–40% of the mammalian genome is organized into ~1,000–1,400 lamina-associated domains (LADs) spanning 0.1 to 10 Mb each. LADs are gene-poor, A/T-rich, late-replicating B-compartment territories enriched in H3K9me2 (deposited by G9a/GLP [EHMT2/EHMT1]), peripheral H3K9me3, and H3K27me3 at LAD borders. Physical tethering of LADs to the nuclear envelope relies on a bipartite mechanism: in proliferating cells, the inner nuclear membrane protein Lamin B Receptor (LBR) binds directly to HP1α/γ and H4K20me2 via its Tudor domain, whereas in differentiated cells, Lamin A/C and LEM-domain proteins (Emerin, MAN1, LAP2β) anchor heterochromatin via Barrier-to-Autointegration Factor (BAF/BANF1) and tissue-specific transmembrane netrins.

Conversely, the euchromatic A compartment subdivides into two radial sub-compartments revealed by sub-kilobase Hi-C and TSA-seq: A1 euchromatin, which clusters directly around nuclear speckles (interchromatin granule clusters scaffolded by the intrinsically disordered proteins SON and SRRM2 alongside MALAT1 lncRNA) and exhibits the highest gene density, GC content, H3K4me3/H3K36me3 enrichment, and co-transcriptional pre-mRNA splicing efficiency; and A2 euchromatin, which occupies intermediate radial positions between the lamina and nuclear speckles and displays moderate transcriptional activity (Lieberman-Aiden et al., 2009; Nuebler et al., 2018). Crucially, transcriptional activation of a developmental gene located within a facultative LAD triggers local histone hyperacetylation (H3K27ac/H4K16ac) and SWI/SNF remodeling that detaches the locus from the nuclear lamina and relocates it inward toward the A1/A2 euchromatic interior.

3. Constitutive and Facultative Heterochromatin: Read-Write Cycles, Phase Separation, and Non-Coding RNAs

In contrast to transcriptionally permissive euchromatin, heterochromatin partitions into two mechanistically distinct classes that occupy the nuclear periphery (lamina-associated domains, LADs) and nucleolar periphery (nucleolus-associated domains, NADs): constitutive heterochromatin, which permanently silences satellite repeats, pericentromeres, telomeres, and transposable elements via the H3K9me3–HP1 axis, and facultative heterochromatin, which reversibly silences developmental lineage-specifying genes (such as the HOX clusters) and the inactive X chromosome via the Polycomb Repressive Complexes PRC1 and PRC2 (Rea et al., 2000; Bannister et al., 2001; Lachner et al., 2001; Cao et al., 2002; Blackledge & Klose, 2021).

Heterochromatin should not be mistaken for a static endpoint of this partitioning. Reviews of its establishment and evolution emphasize that heterochromatin is neither inert nor static: it guards the genome against transposon mobilization, its sequences and regulators coevolve rapidly, and dynamic transitions between euchromatic and heterochromatic states accompany development and evolution, although much of the mechanistic evidence derives from Drosophila and yeast models (Liu et al., 2020).

3.1 Constitutive Heterochromatin: The SUV39H1/2–H3K9me3–HP1 Read-Write Feedback Loop

The molecular foundation of constitutive heterochromatin was established by the landmark discovery of Rea et al. (2000) that mammalian SUV39H1 (suppressor of variegation 3-9 homolog 1) and its paralog SUV39H2 are SET-domain histone methyltransferases that selectively catalyze trimethylation of histone H3 lysine 9 (H3K9me3). Immediately thereafter, Bannister et al. (2001) and Lachner et al. (2001) demonstrated that the N-terminal chromodomain (CD) of Heterochromatin Protein 1 (HP1α/CBX5, HP1β/CBX1, HP1γ/CBX3) forms a three-stranded aromatic cage (comprising conserved tyrosine and tryptophan residues) that selectively recognizes H3K9me2/3 with micromolar affinity (K_d ≈ 2–10 μM), while showing negligible binding to unmethylated H3 or H3K4me3.

Lateral spreading and epigenetic inheritance of constitutive heterochromatin rely on a self-propagating read-write cycle: HP1 dimerizes via its C-terminal chromoshadow domain (CSD) (K_d ≈ 150 nM), creating a symmetric nonpolar peptide-binding groove that directly recruits SUV39H1/2, the H4K20me3 methyltransferase SUV420H1/2, and the maintenance DNA methyltransferase complex DNMT1/UHRF1 (Bannister et al., 2001; Lachner et al., 2001). Furthermore, SUV39H1 itself possesses an N-terminal chromodomain that allosterically binds pre-existing H3K9me3, stimulating catalytic methylation of adjacent unmodified nucleosomes. At endogenous retroviruses (ERVs) and LINE-1 transposons in euchromatic or interspersed regions, a parallel H3K9me3 pathway is nucleated by sequence-specific KRAB-zinc finger proteins (KRAB-ZFPs), which recruit the scaffold KAP1 (TRIM28) and the bifurcated SET-domain methyltransferase SETDB1 alongside the HUSH complex (TASOR, MPP8, PPHLN1). Notably, recent studies in pluripotent and differentiating cells by Zhang et al. (2024) revealed that H3K9me3 domains are not monolithic: SUV39H1/2-, SETDB1-, and SUV39H-independent domains exhibit distinct replication-coupled maintenance kinetics that differentially govern lineage-inappropriate genes versus satellite repeats.

The chromatin remodeller ATRX illustrates that this machinery is not confined to silent compartments. Together with DAXX, ATRX deposits histone H3.3 at sites of nucleosome turnover, and improved chromatin immunoprecipitation demonstrates that ATRX also binds active euchromatic regulatory elements; mutation of ATRX reduces chromatin accessibility, histone modification, transcription factor binding, and H3.3 deposition at its binding sites, with stochastic effects restricted to subsets of erythroid cells (Truch et al., 2022).

3.2 Biophysical Mechanics of HP1α: Phase Separation, Hinge Electrostatics, and Nucleosome Core Reshaping

Despite maintaining durable transcriptional repression, HP1 molecules inside constitutive heterochromatin foci exchange with the nucleoplasm within seconds (t₁/₂ ≈ 1–10 s by FRAP). Two back-to-back studies in Nature by Strom et al. (2017) and Larson et al. (2017) proposed that constitutive heterochromatin assembly involves liquid-liquid phase separation (LLPS) of HP1α/HP1a. Larson et al. (2017) demonstrated that phosphorylation of the human HP1α N-terminal extension (NTE, serines S11–S14 by casein kinase II) relieves autoinhibition and enables electrostatic bridging between the negatively charged NTE of one HP1α dimer and the positively charged disordered hinge region (KRK-rich patch) of another, lowering the saturation concentration for droplet formation from >25 μM to C_sat ≈ 2.8–6.5 μM in the presence of DNA and H3K9me3 nucleosome arrays (Larson et al., 2017). In Drosophila embryos, Strom et al. (2017) showed that newly formed heterochromatin domains exhibit liquid-like droplet coalescence, sensitivity to 1,6-hexanediol during early nucleation, and selective exclusion of inert macromolecular probes, while maturing over developmental time into a biphasic gel/polymer-scaffolding state with an immobile structural fraction.

Crucially, Sanulli et al. (2019) discovered via hydrogen-deuterium exchange mass spectrometry (HDX-MS) and NMR spectroscopy that HP1 binding is not a passive surface decoration of rigid nucleosomes: interaction of Schizosaccharomyces pombe Swi6/HP1 with H3K9me3 nucleosomes transiently reshapes and loosens the internal histone octamer core (increasing solvent exposure of H3 αN, H2B α 1, and H4 core helices by 2-to-4-fold). By exposing buried hydrophobic and aromatic residues within the histone octamer, HP1 allosterically increases multivalent inter-nucleosomal contact valency, directly coupling internal octamer plasticity to mesoscale phase separation and chromatin compaction (Sanulli et al., 2019).

Subsequent work has refined this picture at the level of thermodynamics, molecular contacts, and material state. A theoretical model coupling self-interacting HP1 to the chromatin polymer shows that HP1's affinity for H3K9me2/3 permits coacervation far below the phase-separation thresholds measured in vitro, and that HP1–chromatin crosstalk produces anomalously slow equilibration and multiple long-lived microphase compartments, consistent with 4D microscopy in live Drosophila embryos (Tortora et al., 2023). At the molecular level, phosphorylation-driven phase separation of HP1α involves an exchange of intradimer hinge–N-terminal-extension contacts and an extended dimer conformation, and charged peptide or nucleic acid ligands enhance or disrupt droplet formation according to their sign, so that electrostatics fine-tune HP1α condensate properties (Her et al., 2022). Within HP1α–DNA condensates, HP1α itself remains liquid-like while the compacted DNA is locally constrained; the condensates resist large mechanical forces yet dissolve upon HP1β addition, and paralog-specific behavior maps to their disordered regions (Keenen et al., 2020). Integrative structural analyses (NMR, SAXS, MALS, and molecular dynamics) likewise resolve phosphorylation-dependent dynamic contacts between the HP1α N-terminal tails and basic chromodomain/hinge segments, with the phosphorylated dimer phase-separating as aggregated multimers; notably, a mutant lacking the chromoshadow domain still phase-separates through the dynamic N-tail-mediated dimer, and mutation of the basic chromodomain-end segment alters heterochromatin focus size in cells (Furukawa et al., 2025).

Phase separation is also not a property of every HP1 paralog acting alone. HP1γ cannot phase-separate by itself, but the tripartite-motif protein TRIM66 binds HP1γ through a PxVxL motif and the pair co-phase-separates in vitro and in vivo; TRIM66–HP1γ granules correlate with H3K9me3-marked sites, suggesting a TRIM66–HP1γ–H3K9me3 axis in the assembly of facultative heterochromatin (Shen et al., 2025).

3.3 Facultative Heterochromatin: Canonical and Variant Polycomb Repressive Complexes (PRC1 & PRC2)

Developmental plasticity requires a distinct silencing architecture—facultative heterochromatin—that can be stably inherited across mitoses in non-expressing lineages yet rapidly erased in cells instructed to activate target loci (Blackledge & Klose, 2021). This system is orchestrated by two cooperative multi-subunit machines: Polycomb Repressive Complex 2 (PRC2), which catalyzes mono-, di-, and trimethylation of histone H3 lysine 27 (H3K27me3; Cao et al., 2002), and Polycomb Repressive Complex 1 (PRC1), an E3 ubiquitin ligase formed around the RING1A/B–PCGF heterodimer that monoubiquitinates histone H2A at lysine 119 (H2AK119ub1).

The core catalytic lobe of PRC2 comprises the SET-domain methyltransferase EZH2 (or its paralog EZH1 in post-mitotic tissues), the structural WD40-repeat subunit EED, the VEFS-domain scaffold SUZ12, and the histone-binding chaperone RBBP4/7 (Cao et al., 2002; Margueron et al., 2009). In a landmark structural and biochemical study, Margueron et al. (2009) demonstrated that the aromatic cage on the top surface of the EED β-propeller specifically binds pre-existing H3K27me3, inducing an allosteric conformational rearrangement of the EZH2 SET activation loop that stimulates its catalytic Vₘₐₓ by 3-to-7-fold. Mutations disrupting the EED aromatic cage abolish H3K27me3 domain propagation in vivo, proving that PRC2 utilizes an intrinsic read-write allosteric loop analogous to SUV39H1/HP1 (Margueron et al., 2009). Accessory subunits further partition PRC2 into two functional subcomplexes: PRC2.1 (containing PCL1–3 [PHF1, MTF2, PHF19], which bind unmethylated CpG islands via winged-helix domains, plus EPOP or PALI1) and PRC2.2 (containing JARID2 and AEBP2, which directly recognize PRC1-deposited H2AK119ub1 via a ubiquitin-interaction motif; Blackledge & Klose, 2021).

How PRC2 domains form and persist has been resolved further by rebuild and excision experiments. Inducible collapse and re-formation of PRC2 domains reveals a two-step mechanism: PRC2 is first stably recruited by JARID2 and MTF2 to a small number of spatially interacting nucleation sites, and H3K27me3 binding then allosterically activates the complex so that the mark spreads rapidly in cis and to far-cis regions; in the rebuilt domains H3K27me3 remains concentrated near the nucleation sites, whereas H3K27me2 extends distally (Oksuz et al., 2018). Whether such domains can propagate without continued targeted recruitment was tested directly in Drosophila: inserted Polycomb response elements (PREs) create extended H3K27me3 domains and repression, but after PRE excision the mark dilutes with replication and repression is lost, persisting only in cells stalled in replication — marked nucleosomes thus carry sequence-independent memory, but propagation to new nucleosomes requires sequence-specific PRC2 targeting (Laprell et al., 2017). Three-dimensional folding also participates: H3K27me3 HiChIP and ORCA imaging reveal Polycomb-associated loops spanning tens to hundreds of megabases across TADs in mouse embryonic stem cells and human induced pluripotent stem cells, anchored at Polycomb nucleation points and developmental genes, and deleting a loop anchor disrupts distant contacts and H3K27me3 deposition in cis while ectopically activating the partner gene in embryos; EZH2 requires its RNA-binding capacity for this long-range folding and spreading (Kraft et al., 2020). Finally, the mark itself, rather than mere PRC2 occupancy, is the functional agent: in embryonic stem cells expressing catalytically inactive EZH2, loss of H3K27 methylation abolishes differentiation capacity, while EZH1 can maintain repression through H3K27me2/3 and PRC1 recruitment — although altered H3K27me1 at active genes and diffuse chromatin hyperacetylation compromise developmental potential (Lavarone et al., 2019).

Rather than operating strictly via the classical unidirectional hierarchy (in which PRC2 deposits H3K27me3 first to recruit canonical PRC1), modern genetic and genomic dissection has established a bidirectional feedback network between variant PRC1 (vPRC1) and canonical PRC1 (cPRC1) (Blackledge & Klose, 2021): (1) vPRC1 complexes (incorporating PCGF1/3/5/6 and RYBP/YAF2, such as the KDM2B-containing PCGF1-PRC1 complex that scans unmethylated CpG islands) possess high intrinsic E3 ligase activity and deposit the majority of genome-wide H2AK119ub1; (2) H2AK119ub1 is read by JARID2/AEBP2 to recruit PRC2.2, which deposits H3K27me3; and (3) H3K27me3 is subsequently recognized by the chromodomain of CBX subunits (CBX2, CBX4, CBX6, CBX7, CBX8) within cPRC1 (PCGF2/4 [MEL18/BMI1] + PHC1–3 + CBX + RING1A/B). Once recruited, cPRC1 drives long-range chromatin compaction and Polycomb body formation through two synergistic mechanisms: Oligomerization of the Sterile Alpha Motif (SAM) domain of PHC1–3 into head-to-tail helical filaments, and electrostatic phase separation driven by the positively charged disordered compaction region of CBX2.

3.4 Non-Coding RNAs in Heterochromatin Nucleation: Xist lncRNA and Pericentromeric Satellite Transcripts

A recurring paradox in heterochromatin biology is that the establishment and maintenance of silent chromatin domains frequently require low-level non-coding RNA (ncRNA) transcription across the target locus itself (Blackledge & Klose, 2021). The paradigm of lncRNA-guided facultative heterochromatin assembly is mammalian X-chromosome inactivation (XCI), orchestrated by the 17–19 kb Xist (X-inactive specific transcript) lncRNA expressed exclusively from the future inactive X chromosome (Xi). Xist contains six conserved tandem repeat modules (Repeats A through F) that act as modular multivalent scaffolds: Repeat A (at the 5' end) binds the RNA-binding protein SPEN (SHARP), which recruits the SMRT/NCoR–HDAC3 corepressor complex to deacetylate active enhancers and promoters across the X chromosome and evict RNA Pol II; Repeat B (and Repeat C) binds hnRNP K, which directly recruits the PCGF3/5-vPRC1 complex to deposit H2AK119ub1; and H2AK119ub1 subsequently recruits JARID2/AEBP2-containing PRC2.2 to spread H3K27me3 across the entire inactive X territory (Blackledge & Klose, 2021). Late in XCI maintenance, the variant histone macroH2A1/2 and SMCHD1 (an SMC-hinge ATPase that merges Xi compartments into two megadomains flanking the DXZ4 macrosatellite) lock the inactive X into a mitotically heritable silent state.

Similarly, at pericentromeric constitutive heterochromatin in mammalian cells, strand-specific transcription of major satellite repeats during late G1 and S phase generates nascent satellite non-coding RNAs that remain cis-associated with pericentromeric chromatin. Both SUV39H1 (via basic residues within its chromodomain and hinge) and HP1α (via its positively charged hinge domain) bind nascent major satellite RNA transcripts, stabilizing SUV39H1 retention at pericentromeres and nucleating HP1α phase-separated foci (Larson et al., 2017; Strom et al., 2017).

4. ATP-Dependent Chromatin Remodeling Complexes: Mechanochemical Motors of the Genome

Because thermal DNA breathing at the nucleosomal dyad occurs with negligible probability (K_eq < 10⁻⁴) and because 14 hydrogen bonds and electrostatic salt bridges anchor DNA at the 14 superhelical locations (totalling Δ G_bind ≈ 35–45 kcal·mol⁻¹), eukaryotic cells rely on ATP-dependent chromatin remodeling complexes to slide, eject, space, or exchange histones within nucleosomes (Clapier & Cairns, 2009; Narlikar et al., 2013). All chromatin remodelers share a conserved catalytic Snf2-family Superfamily 2 (SF2) ATPase motor composed of two RecA-like lobes (DExx lobe 1 and HELICc lobe 2) separated by a family-specific insertion domain. High-resolution cryo-EM structures and single-molecule optical tweezer assays have established a unified inchworm DNA translocation mechanism: the two RecA lobes bind internal nucleosomal DNA (typically at superhelical location SHL +2, two helical turns from the dyad, or at SHL -6 in INO80) and undergo an ATP-hydrolysis-driven open-to-closed conformational cycle that pulls 1–2 bp of DNA per catalytic step from the linker entry side toward the dyad (Clapier & Cairns, 2009; He et al., 2020; Eustermann et al., 2018). Auxiliary domains flanking the ATPase core differentiate the remodelers into four major evolutionary families: SWI/SNF, ISWI, CHD, and INO80/SWR1.

4.1 The SWI/SNF (BAF, PBAF, ncBAF) Family: Nucleosome Eviction and Enhancer Accessibility

Mammalian SWI/SNF (mSWI/SNF or BAF) complexes are massive (~1.1–1.5 MDa), combinatorial 11-to-15-subunit machines powered by one of two mutually exclusive catalytic ATPases—BRG1 (SMARCA4) or BRM (SMARCA2)—which contain a C-terminal bromodomain that binds acetylated histones (Clapier & Cairns, 2009; Kadoch & Crabtree, 2015). Depending on subunit paralog assembly, mSWI/SNF partitions into three distinct assemblies: (1) canonical BAF (cBAF), defined by ARID1A/ARID1B and DPF1/2/3, which predominantly occupies distal enhancers and super-enhancers; (2) polybromo-associated BAF (PBAF), defined by PBRM1 (BAF180) (containing six tandem bromodomains and two BAH domains), ARID2 (BAF200), and BRD7, which localizes to active promoters; and (3) non-canonical BAF (ncBAF or GBAF), a smaller complex lacking SMARCB1 (BAF47) and ARID subunits but incorporating BRD9 and GLTSCR1/1L, which colocalizes with CTCF and promoter-proximal sites.

The 3.7 Å cryo-EM structure of the human nucleosome-bound cBAF complex solved by He et al. (2020) (and the corresponding PBAF structure by Yuan et al., 2022) revealed how SWI/SNF achieves high-force nucleosome sliding and eviction (generating optical trap stall forces of F_stall ≈ 12–15 pN). BAF organizes into three modules—the ATPase module (SMARCA4 engaging nucleosomal DNA at SHL +2), the Actin-related protein (ARP) bridge module (β-actin and ACTL6A/BAF53A binding the SMARCA4 HSA helix to couple ATPase torque to the base), and the Base module (comprising SMARCB1, SMARCC1/2, SMARCD1, SMARCE1, ARID1A, and DPF2). Crucially, the Base and ATPase modules sandwich the nucleosome core particle simultaneously on opposite faces: while SMARCA4 engages SHL +2, the C-terminal basic α-helix of SMARCB1 (BAF47) anchors directly into the bottom H2A–H2B acidic patch, while the SMARCA4 SnAC domain engages the top acidic patch (He et al., 2020). This bilateral grip prevents passive nucleosome rotation and allows SMARCA4 translocation to peel DNA off the octamer, driving nucleosome sliding, octamer transfer, and complete nucleosome eviction to create open NDRs (Basurto-Cayuela et al., 2024).

4.2 The ISWI and CHD Families: Allosteric Rulers for Ordered Nucleosome Spacing

Whereas SWI/SNF complexes disrupt nucleosome order to expose cis-regulatory DNA, the ISWI (Imitation Switch) and CHD (Chromodomain Helicase DNA-Binding) families act primarily as chromatin assembly and spacing machines that slide nucleosomes into regularly phased arrays without evicting histones (Clapier & Cairns, 2009; Narlikar et al., 2013). Mammalian ISWI complexes are built around the catalytic ATPases SNF2H (SMARCA5) or SNF2L (SMARCA1) paired with 1–3 regulatory subunits (forming ACF, CHRAC, WICH, NoRC, RSF, and NURF). ISWI ATPases contain a C-terminal HAND-SANT-SLIDE (HSS) domain that binds extranucleosomal linker DNA (~20–40 bp from the entry site) and two autoinhibitory motifs—AutoN and NegC—that hold the ATPase lobes in an inactive conformation in the absence of a nucleosome substrate. Binding of the basic H4 tail (residues K16–R19) to a dedicated pocket on ATPase lobe 2 relieves AutoN inhibition (explaining why H4K16ac potently inhibits ISWI-mediated sliding; Shogren-Knaak et al., 2006), while binding of the HSS domain to linker DNA relieves NegC inhibition, converting the HSS domain into a mechanical 'protein ruler' that equalizes linker lengths to ~180–200 bp nucleosome repeat lengths (Narlikar et al., 2013).

The CHD family (CHD1–CHD9) is defined by tandem N-terminal chromodomains adjacent to the Snf2 ATPase motor and partitions into three subfamilies: Subfamily I (CHD1, CHD2), in which the tandem chromodomains of CHD1 selectively recognize H3K4me2/3 at active promoters to space +1/+2 nucleosomes in the wake of RNA Pol II elongation; Subfamily II (CHD3, CHD4, CHD5), which lack DNA-binding SANT-SLIDE domains but contain tandem PHD fingers (binding unmodified H3 tails or H3K9me3) and serve as the catalytic core of the NuRD (Nucleosome Remodeling and Deacetylase) complex alongside HDAC1/2, MTA1–3, MBD2/3, and GATAD2A/B; and Subfamily III (CHD6–CHD9), which feature BRK and SANT-like domains that regulate developmentally poised enhancers (Clapier & Cairns, 2009).

4.3 The INO80 and SWR1/SRCAP Family: Split-ATPase Scaffolds for Histone Variant Exchange

The INO80/SWR1 family (comprising mammalian INO80, SRCAP [the human ortholog of yeast SWR1], and the p400/TIP60 [EP400–KAT5] complex) is uniquely distinguished by a long insertion domain within the Snf2 ATPase core that recruits the hexameric AAA+ ATPase ring RUVBL1–RUVBL2 (Pontin/Reptin; Rvb1/Rvb2) as an obligate structural stators (Clapier & Cairns, 2009; Eustermann et al., 2018). Rather than engaging SHL +2 like SWI/SNF and ISWI, the cryo-EM structure of nucleosome-bound INO80 by Eustermann et al. (2018) demonstrated that the INO80 motor binds near the nucleosomal entry gyre at SHL -6, while the ARP5–IES6 module grasps SHL -2/-3 and the lower H2A–H2B acidic patch. By pumping linker DNA into the nucleosome against the ARP5–IES6 grapnel, INO80 transiently unwraps ~15–20 bp of outer-gyre DNA, exposing the H2A–H2B dimer interface. This specialized geometry allows SRCAP and p400/TIP60 to catalyze ATP-dependent replacement of canonical H2A–H2B dimers with the variant H2A.Z–H2B at +1 promoters and DNA double-strand breaks (Qu et al., 2024), whereas INO80 catalyzes both nucleosome sliding and the reverse eviction of H2A.Z to maintain genomic stability and replication fork progression.

4.4 Comparative Structural Mechanics of Nucleosome Sliding, Eviction, and Hexasome Generation

Recent single-molecule optical tweezers, cryo-EM snapshots across nucleotide states (apo, ADP-BeF_x, ADP), and genomic footprinting have clarified how the same conserved Snf2-family SF2 ATPase fold generates radically different structural outcomes across the four remodeler families (Clapier & Cairns, 2009; Narlikar et al., 2013; He et al., 2020; Eustermann et al., 2018). During the fundamental catalytic cycle at SHL +2, binding of the apo Snf2 ATPase lobes to the tracking and guide DNA strands induces a 1-bp register shift in the tracking strand, creating a local twist defect that pulls 1 bp of DNA from the proximal entry linker into the nucleosome core. Upon ATP binding, closure of the two RecA lobes eliminates the twist defect and propagates the 1-bp DNA bulge across the dyad toward the distal exit gyre; subsequent ATP hydrolysis and phosphate release reset the motor for the next 1-bp translocation step.

Whether this 1-bp-per-ATP DNA translocation results in ordered nucleosome sliding (ISWI/CHD1) or octamer eviction / hexasome formation (SWI/SNF) depends on how the auxiliary subunits regulate linker DNA sensing and octamer clamping (He et al., 2020; Yuan et al., 2022). In ISWI (SNF2h), once the translocated nucleosome approaches a neighboring nucleosome and the extranucleosomal linker DNA shortens below ~20 bp, the C-terminal HSS domain loses high-affinity linker engagement, allowing the autoinhibitory NegC helix to re-engage the ATPase lobes and halt translocation before the two nucleosomes collide. In stark contrast, SWI/SNF (cBAF/PBAF) lacks NegC/AutoN brakes and instead uses its rigid ARP module and bilateral SMARCB1–acidic patch clamp to continue pumping DNA at high torque (>12 pN) even when the nucleosome collides with an adjacent particle (He et al., 2020). As a result, BAF drives one nucleosome directly into the territory of its neighbor, peeling DNA off the distal H2A–H2B dimer to generate subnucleosomal hexasomes (lacking one H2A–H2B dimer) and ultimately ejecting the entire histone octamer to establish nucleosome-depleted cis-regulatory elements (Basurto-Cayuela et al., 2024).

5. Mendelian Chromatinopathies and Somatic Epigenetic Oncogenesis

The exquisite dosage sensitivity of the chromatin machinery is underscored by the Mendelian Disorders of the Epigenetic Machinery (MDEMs)—commonly termed chromatinopathies—a rapidly expanding class of >80 monogenic neurodevelopmental syndromes caused by germline heterozygous loss-of-function (haploinsufficiency) or dominant-negative missense mutations in histone writers, erasers, readers, remodelers, and architectural insulators (Fahrner & Bjornsson, 2014; Fahrner & Bjornsson, 2019). Despite affecting diverse biochemical classes, chromatinopathies share a striking clinical triad: intellectual disability / neurodevelopmental delay, growth dysregulation (either postnatal undergrowth/microcephaly or overgrowth/macrocephaly), and craniofacial/skeletal dysmorphism.

5.1 Mechanistic Paradigms of Writer-Eraser Balance and Remodeler Syndromes

As conceptualized by Fahrner & Bjornsson (2014, 2019), normal developmental gene expression depends on a dynamic steady-state equilibrium between opposing chromatin writers and erasers at target promoters and enhancers. Disrupting either side of a coupled writer-eraser pair produces convergent or mirror-image syndromic phenotypes:

• Kabuki Syndrome (KMT2D / KDM6A): Heterozygous truncating or missense mutations in KMT2D (MLL2/MLL4)—the COMPASS-like H3K4me1/2 methyltransferase at developmental enhancers (Ng et al., 2010)—or in KDM6A (UTX), the X-linked H3K27me3 demethylase that physically associates with KMT2D to remove Polycomb repression during enhancer commissioning, cause Kabuki syndrome Type 1 and Type 2, respectively. Both defects collapse the H3K4me1/H3K27ac vs. H3K27me3 balance toward excessive Polycomb silencing at neural crest and hippocampal neurogenesis loci, a defect that can be partially reversed in preclinical mouse models via ketogenic β-hydroxybutyrate HDAC inhibition (Fahrner & Bjornsson, 2019).

• Rubinstein-Taybi Syndrome (CREBBP / EP300) vs. Menke-Hennekam Syndrome: Haploinsufficiency of the master enhancer acetyltransferases CBP (CREBBP) or p300 (EP300) reduces H3K27ac and H3K18ac across activity-dependent neuronal enhancers (controlled by CREB and FOS/JUN), causing intellectual disability, broad thumbs/halluces, and postnatal growth retardation (Dancy & Cole, 2015).

• Kleefstra Syndrome (EHMT1) vs. Koolen-de Vries Syndrome (KANSL1): Germline deletions or point mutations in EHMT1 (GLP), which heterodimerizes with G9a (EHMT2) to deposit euchromatic H3K9me1/2 repression at non-neuronal and synaptic genes, cause Kleefstra syndrome (9q34.3 deletion). Conversely, heterozygous loss of KANSL1—the core scaffold of the KAT8-containing NSL acetyltransferase complex at housekeeping promoters—causes Koolen-de Vries syndrome (17q21.31 microdeletion; Koolen et al., 2012).

• Overgrowth-Intellectual Disability Syndromes (Weaver [EZH2], Sotos [NSD1], Tatton-Brown-Rahman [DNMT3A]): Constitutional partial loss-of-function mutations in EZH2 (Weaver syndrome), the H3K36me2 methyltransferase NSD1 (Sotos syndrome), or the de novo DNA methyltransferase DNMT3A (Tatton-Brown-Rahman syndrome) impair facultative heterochromatin and intergenic DNA methylation boundaries, de-repressing growth-promoting mitogen and cell-cycle pathways.

• BAFopathies (Coffin-Siris & Nicolaides-Baraitser Syndromes) and Cohesinopathies (Cornelia de Lange Syndrome): Germline mutations in mSWI/SNF subunits (ARID1B, ARID1A, SMARCB1, SMARCA4, SMARCE1) cause Coffin-Siris syndrome, while missense mutations in the SMARCA2 ATPase core cause Nicolaides-Baraitser syndrome (Kadoch & Crabtree, 2015). In parallel, mutations in the cohesin loader NIPBL (accounting for ~65% of cases) or cohesin core subunits (SMC1A, SMC3, RAD21, HDAC8) impair enhancer-promoter loop extrusion without abolishing mitotic sister chromatid cohesion, causing Cornelia de Lange syndrome.

5.2 Somatic mSWI/SNF Mutations and Synthetic Lethal Epigenetic Dependencies in Cancer

Pan-cancer genomic sequencing compiled by Kadoch & Crabtree (2015) demonstrated that subunits of the mammalian SWI/SNF (BAF/PBAF) complex are somatically mutated in over 20% of all human malignancies, rivaling TP53 in mutation frequency. Remarkably, specific subunit lesions drive highly tissue-specific tumors through distinct biochemical mechanisms: biallelic loss of SMARCB1 (BAF47/INI1) is the defining genetic lesion in >98% of malignant rhabdoid tumors and epithelioid sarcomas; ARID1A is mutated in ~50% of ovarian clear-cell carcinomas and ~30% of endometrioid carcinomas; PBRM1 (BAF180) is inactivated in ~41% of clear-cell renal cell carcinomas; and chromosomal translocation in 100% of synovial sarcomas fuses 78 C-terminal amino acids of SSX1/2/4 onto the BAF subunit SS18 (SS18–SSX), which hijacks cBAF complexes away from enhancers to evict Polycomb (PRC1/PRC2) from SOX2 and other mesenchymal oncogenes (Kadoch & Crabtree, 2015).

Conversely, loss of SMARCB1 in malignant rhabdoid tumors destabilizes enhancer-bound cBAF complexes and unleashes unopposed PRC2 (EZH2) activity at bivalent tumor suppressor loci (such as CDKN2A/p16INK4A). This direct biochemical antagonism between SWI/SNF and Polycomb established the clinical paradigm of epigenetic synthetic lethality: SMARCB1-deficient rhabdoid tumors and epithelioid sarcomas are selectively addicted to EZH2 catalytic activity, leading to the clinical approval of the SAM-competitive EZH2 inhibitor tazemetostat, alongside emerging proteolysis-targeting chimeras (PROTACs) directed against the residual paralogous ATPase SMARCA2 (BRM) in SMARCA4-mutant non-small-cell lung cancers and against BRD9 (ncBAF) in synovial sarcoma and SMARCB1-deficient tumors (Kadoch & Crabtree, 2015).

6. Computational Polymer Physics and Predictive Modeling of the 3D Epigenome

Integrating one-dimensional epigenomic tracks (ChIP-seq, ATAC-seq, CUT&Tag) with three-dimensional chromosome folding (Hi-C, Micro-C, SPRITE, and super-resolution chromatin tracing) has catalyzed a rigorous physical modeling framework for the cell nucleus (Imakaev et al., 2015; Fudenberg et al., 2016; Nuebler et al., 2018). Contemporary computational models divide into two complementary paradigms: bottom-up mechanistic polymer simulations (such as the Strings and Binders Switch [SBS] model, block-copolymer microphase separation, and active loop-extruding polymer dynamics) and top-down data-driven restraint and deep-learning models.

In unified polymer physics simulations developed by Mirny and colleagues (Fudenberg et al., 2016; Nuebler et al., 2018), chromosomes are represented as coarse-grained bead-spring chains (typically 1–10 kb per monomer) governed by Langevin dynamics with potential energy terms capturing excluded volume, bending stiffness, homotypic Flory-Huggins interaction energies arepsilon_AA, arepsilon_BB between chromatin states (driving A/B compartmentalization), and active non-equilibrium motor rules for cohesin loop extrusion:

U_total = U_connectivity + U_bending + ∑_(i<j) ε_(σ_i, σ_j) · f(r_ij) + W_loop-extrusion(t)

Crucially, Nuebler et al. (2018) demonstrated through polymer simulations and experimental validation in WAPL- and NIPBL-perturbed cells that loop extrusion and compartmental segregation act antagonistically: active loop extrusion continuously reels chromatin segments across epigenetic boundaries, blurring short-range A/B micro-compartments while establishing sharp CTCF-insulated TADs; conversely, halting ATP-dependent loop extrusion allows thermodynamic Flory-Huggins attractions between H3K9me3/HP1 or Polycomb domains to drive finer-scale microphase separation. Thus, the living interphase nucleus is an active, non-equilibrium viscoelastic polymer fluid whose steady-state 3D architecture is continuously maintained by ATP hydrolysis across SMC motors and Snf2 chromatin remodelers (Maeshima et al., 2016; Nuebler et al., 2018).

References

1. [1] Passarge E (1979). Emil Heitz and the concept of heterochromatin: longitudinal chromosome differentiation was recognized fifty years ago. American Journal of Human Genetics 31(2):106–115. PMID 377956.
2. [2] Yunis JJ, Yasmineh WG (1971). Heterochromatin, satellite DNA, and cell function. Science 174(4015):1200–1209. PMID 4942044.
3. [3] Luger K, Mäder AW, Richmond RK, Sargent DF, Richmond TJ (1997). Crystal structure of the nucleosome core particle at 2.8 Å resolution. Nature 389(6648):251–260. PMID 9305837.
4. [4] Routh A, Sandin S, Rhodes D (2008). Nucleosome repeat length and linker histone stoichiometry determine chromatin fiber structure. Proceedings of the National Academy of Sciences USA 105(26):8872–8877. PMID 18579775.
5. [5] Ou HD, Phan S, Deerinck TJ, Thor A, Ellisman MH, O'Shea CC (2017). ChromEMT: Visualizing 3D chromatin structure and compaction in interphase and mitotic cells. Science 357(6349):eaag0025. PMID 28751582.
6. [6] Maeshima K, Ide S, Hibino K, Sasai M (2016). Liquid-like behavior of chromatin. Current Opinion in Genetics & Development 37:36–45. PMC8091949.
7. [7] Gibson BA, Doolittle LK, Schneider MWG, Jensen LE, Gamarra N, Henry L, Gerlich DW, Redding S, Rosen MK (2019). Organization of chromatin by intrinsic and regulated phase separation. Cell 179(2):470–484.e21. PMID 31543265.
8. [8] Strom AR, Emelyanov AV, Mir M, Fyodorov DV, Darzacq X, Karpen GH (2017). Phase separation drives heterochromatin domain formation. Nature 547(7662):241–245. PMID 28636597.
9. [9] Larson AG, Elnatan D, Keenen MM, Trnka MJ, Johnston JB, Burlingame AL, Agard DA, Redding S, Narlikar GJ (2017). Liquid droplet formation by HP1α suggests a role for phase separation in heterochromatin. Nature 547(7662):236–240. PMC10636758.
10. [10] Morrison, Thakur (2021). Molecular Complexes at Euchromatin, Heterochromatin and Centromeric Chromatin. International Journal of Molecular Sciences.
11. [11] Clapier CR, Cairns BR (2009). The biology of chromatin remodeling complexes. Annual Review of Biochemistry 78:273–304. PMID 19355820.
12. [12] Narlikar GJ, Sundaramoorthy R, Owen-Hughes T (2013). Mechanisms and functions of ATP-dependent chromatin-remodeling enzymes. Cell 154(3):490–503. PMC9157391.
13. [13] Shogren-Knaak M, Ishii H, Sun JM, Pazin MJ, Davie JR, Peterson CL (2006). Histone H4-K16 acetylation controls chromatin structure and protein interactions. Science 311(5762):844–847. PMID 16469925.
14. [14] Strahl BD, Allis CD (2000). The language of covalent histone modifications. Nature 403(6765):41–45. PMID 10638745.
15. [15] Kouzarides T (2007). Chromatin modifications and their function. Cell 128(4):693–705. PMID 17320507.
16. [16] Lieberman-Aiden E, van Berkum NL, Williams L, et al. (2009). Comprehensive mapping of long-range interactions reveals folding principles of the human genome. Science 326(5950):289–293. PMID 19815776.
17. [17] Fudenberg G, Imakaev M, Lu C, Goloborodko A, Abdennur N, Mirny LA (2016). Formation of chromosomal domains by loop extrusion. Cell Reports 15(9):2038–2049. PMID 27210764.
18. [18] Rao SSP, Huang SC, Glenn St Hilaire B, et al. (2017). Cohesin loss eliminates all loop domains. Cell 171(2):305–320.e24. PMID 28985562.
19. [19] Nora EP, Goloborodko A, Valton AL, et al. (2017). Targeted degradation of CTCF decouples local insulation of chromosome domains from genomic compartmentalization. Cell 169(5):930–944.e22. PMID 28525758.
20. [20] Heintzman ND, Hon GC, Hawkins RD, et al. (2009). Histone modifications at human enhancers reflect global cell-type-specific gene expression. Nature 459(7243):108–112. PMID 19295514.
21. [21] Buenrostro JD, Giresi PG, Zaba LC, Chang HY, Greenleaf WJ (2013). Transposition of native chromatin for fast and sensitive epigenomic profiling of open chromatin, DNA-binding proteins and nucleosome position. Nature Methods 10(12):1213–1218. PMID 24097267.
22. [22] Prajapati, Xu, Eriksson, Clark (2025). Nucleosome dynamics render heterochromatin accessible in living human cells. Nature Communications.
23. [23] Rea S, Eisenhaber F, O'Carroll D, et al. (2000). Regulation of chromatin structure by site-specific histone H3 methyltransferases. Nature 406(6796):593–599. PMID 10949293.
24. [24] Bannister AJ, Zegerman P, Partridge JF, Miska EA, Thomas JO, Allshire RC, Kouzarides T (2001). Selective recognition of methylated lysine 9 on histone H3 by the HP1 chromo domain. Nature 410(6824):120–124. PMID 11242054.
25. [25] Lachner M, O'Carroll D, Rea S, Mechtler K, Jenuwein T (2001). Methylation of histone H3 lysine 9 creates a binding site for HP1 proteins. Nature 410(6824):116–120. PMID 11242053.
26. [26] Cao R, Wang L, Wang H, Xia L, Erdjument-Bromage H, Tempst P, Jones RS, Zhang Y (2002). Role of histone H3 lysine 27 methylation in Polycomb-group silencing. Science 298(5595):1039–1043. PMID 12351676.
27. [27] Blackledge NP, Klose RJ (2021). The molecular principles of gene regulation by Polycomb repressive complexes. Nature Reviews Molecular Cell Biology 22(12):815–833. PMID 34400841.
28. [28] Liu, Ali, Zhou (2020). Establishment and evolution of heterochromatin. Annals of the New York Academy of Sciences.
29. [29] Zhang J, Donahue G, Gilbert MB, Lapidot T, Nicetto D, Zaret KS (2024). Distinct H3K9me3 heterochromatin maintenance dynamics govern different gene programs and repeats in pluripotent cells. Nature Cell Biology 26(12):2115–2128. DOI 10.1038/s41556-024-01547-z
30. [30] Truch et al. (2022). The chromatin remodeller ATRX facilitates diverse nuclear processes, in a stochastic manner, in both heterochromatin and euchromatin. Nature Communications.
31. [31] Sanulli S, Trnka MJ, Dharmarajan V, et al. (2019). HP1 reshapes the nucleosome core to promote phase separation of heterochromatin. Nature 575(7782):390–394. PMID 31618757.
32. [32] Tortora, Brennan, Karpen, Jost (2023). HP1-driven phase separation recapitulates the thermodynamics and kinetics of heterochromatin condensate formation. Proceedings of the National Academy of Sciences USA.
33. [33] Her et al. (2022). Molecular interactions underlying the phase separation of HP1α: role of phosphorylation, ligand and nucleic acid binding. Nucleic Acids Research.
34. [34] Keenen et al. (2020). HP1 proteins compact DNA into mechanically and positionally stable phase separated domains. eLife.
35. [35] Furukawa et al. (2025). A dynamic structural unit of phase-separated heterochromatin protein 1α as revealed by integrative structural analyses. Nucleic Acids Research.
36. [36] Shen et al. (2025). TRIM66-HP1γ remodels the chromatin through phase separation. Biophysics Reports.
37. [37] Margueron R, Justin N, Ohno K, et al. (2009). Role of the polycomb protein Eed in the propagation of repressive histone marks. Nature 461(7265):762–767. PMID 19759537.
38. [38] Oksuz et al. (2018). Capturing the Onset of PRC2-Mediated Repressive Domain Formation. Molecular Cell.
39. [39] Laprell, Finkl, Müller (2017). Propagation of Polycomb-repressed chromatin requires sequence-specific recruitment to DNA. Science.
40. [40] Kraft et al. (2020). Polycomb-mediated genome architecture enables long-range spreading of H3K27 methylation. Proceedings of the National Academy of Sciences USA.
41. [41] Lavarone et al. (2019). Dissecting the role of H3K27 acetylation and methylation in PRC2 mediated control of cellular identity. Nature Communications.
42. [42] He S, Wu Z, Tian Y, Yu Z, Yu J, Wang X, Li J, Liu B, Xu Y (2020). Structure of nucleosome-bound human BAF complex. Science 367(6480):875–881. PMID 32001526.
43. [43] Eustermann S, Schall K, Kostrewa D, Lakomek K, Strauss M, Moldt M, Hopfner KP (2018). Structural basis for ATP-dependent chromatin remodelling by the INO80 complex. Nature 556(7701):386–390. PMID 29643509.
44. [44] Bernstein BE, Mikkelsen TS, Xie X, et al. (2006). A bivalent chromatin structure marks key developmental genes in embryonic stem cells. Cell 125(2):315–326. PMID 16630819.
45. [45] Kadoch C, Crabtree GR (2015). Mammalian SWI/SNF chromatin remodeling complexes and cancer: Mechanistic insights gained from human genomics. Science Advances / Nature Genetics 47(6):560–570. PMID 26601204.
46. [46] Yuan J, Chen K, Zhang W, Chen Z (2022). Structure of human chromatin-remodeling PBAF complex bound to a nucleosome. Nature 605(7908):166–171. PMC9741621.
47. [47] Basurto-Cayuela L, Guerrero-Martínez JA, Gómez-Marín E, et al. (2024). SWI/SNF-dependent genes are defined by their chromatin landscape. Cell Reports 43(3):113855. DOI 10.1016/j.celrep.2024.113855
48. [48] Qu K, Chen K, Wang H, Li X, Chen Z (2022/2024). Structure of the human TIP60/NuA4 histone exchange and acetyltransferase complex. Nature / PMC 611:608–614. PMC11330486.
49. [49] Fahrner JA, Bjornsson HT (2014). Mendelian disorders of the epigenetic machinery: tipping the balance of chromatin states. Annual Review of Genomics and Human Genetics 15:269–293. PMC4579332.
50. [50] Fahrner JA, Bjornsson HT (2019). Mendelian disorders of the epigenetic machinery: postnatal malleability and therapeutic prospects. Human Molecular Genetics 28(R2):R254–R264. PMID 31595951.
51. [51] Ng SB, Bigham AW, Buckingham KJ, et al. (2010). Exome sequencing identifies MLL2 mutations as a cause of Kabuki syndrome. Nature Genetics 42(9):790–793. PMID 20711175.
52. [52] Dancy BM, Cole PA (2015). Protein lysine acetylation by p300/CBP. Chemical Reviews 115(6):2419–2452. PMID 25594381.
53. [53] Koolen DA, Kramer JM, Neveling K, et al. (2012). Mutations in the chromatin modifier gene KANSL1 cause the 17q21.31 microdeletion syndrome. Nature Genetics 44(6):639–641. PMID 22544363.
54. [54] Imakaev MV, Fudenberg G, Mirny LA (2015). Modeling chromosomes: Beyond pretty pictures. FEBS Letters 589(20 Pt A):3031–3036. PMC6512960.
55. [55] Nuebler J, Fudenberg G, Imakaev M, Abdennur N, Mirny LA (2018). Chromatin organization by an interplay of loop extrusion and compartmental segregation. Proceedings of the National Academy of Sciences USA 115(29):E6697–E6706. PMC6055145.
56. [56] Alabert C, Barth TK, Reverón-Gómez N, et al. (2015). Two distinct modes for propagation of histone PTMs across the cell cycle. Genes & Development 29(6):585–590. PMID 25792596.
57. [57] Charlton SJ, Flury V, Kanoh Y, et al. (2024). The fork protection complex promotes parental histone recycling and epigenetic memory. Cell 187(18):5029–5047.e21. DOI 10.1016/j.cell.2024.07.017