Chromatin Fiber and Higher-Order Folding

Chromatin compaction proceeds hierarchically. The first level of packing, DNA wound twice around a histone octamer to form the nucleosome core particle, shortens DNA roughly 6-fold. Strung together with linker DNA and linker histone H1, this produces the 10-nm "beads-on-a-string" fiber — itself only a modest compaction. To achieve the roughly 10,000-fold compaction needed to fit a chromosome into a nucleus, chromatin must fold further. Historically, biochemists proposed that nucleosome arrays coil into a thicker, ~30-nm fiber, visible by electron microscopy in vitro. Whether this fiber exists as a discrete, regular structure inside living cells, versus an irregular, liquid-like, dynamically folding polymer, has been one of the more contentious debates in the field — addressed below. Beyond local fiber folding, chromatin organizes over much longer distances through loops: a stretch of DNA between two points, brought into proximity, with everything in between bulging out as a loop. Loops let a gene's promoter contact a distant enhancer tens or hundreds of kilobases away, looping out the intervening DNA. Many loops share a common mechanism — extrusion by the cohesin complex, stopped by the DNA-binding protein CTCF — and nest within larger self-interacting neighborhoods called topologically associating domains (TADs), which in turn cluster into even larger-scale genome compartments. Overlaid on this architecture is a fundamental distinction in chromatin state. Euchromatin is the less condensed, more accessible configuration, enriched in active genes, acetylated histones (loosening histone-DNA contacts), and early-replicating DNA. Heterochromatin is densely packed, gene-poor or silenced, enriched in methylated histones (particularly H3K9me3) and, in many organisms, methylated DNA, and it replicates late in S phase. Heterochromatin is subdivided into constitutive heterochromatin — permanently condensed regions like centromeres and telomeres, present in essentially the same state in every cell type — and facultative heterochromatin — regions silenced in some cells but active in others, such as one entire X chromosome in female mammalian cells. This state is not written in DNA sequence; it is written in chromatin, making it a heritable but reversible layer of regulation — the molecular basis of much of what is studied under Histones and the Histone Code.

Step 1 — From nucleosomes to a folded fiber. The 10-nm fiber, an array of nucleosomes connected by linker DNA (typically 20–90 bp, giving a nucleosome repeat length around 147 + 20–90 = roughly 165–240 bp), is the substrate for further folding. Linker histone H1 binds at the nucleosome dyad and to the two entering/exiting DNA strands, drawing them together and promoting compaction. In vitro, purified nucleosome arrays with H1 fold into a fiber roughly 30 nm in diameter, in which nucleosomes pack against each other in a zigzag or solenoidal arrangement (models differ on the exact geometry — a two-start zigzag, where alternating nucleosomes stack, versus a one-start interdigitated solenoid). This packing compacts DNA a further roughly 6-fold beyond the 10-nm fiber.

Step 2 — Does the 30-nm fiber exist in living cells? This is genuinely unsettled and worth stating plainly. Classical electron microscopy of isolated chromatin, and later cryo-EM of reconstituted nucleosome arrays with H1, clearly show a compact ~30-nm fiber in vitro. But when nuclei are cryo-preserved without chemical fixation or dehydration artifacts and imaged by cryo-electron tomography, regular 30-nm fibers are largely absent from interphase nuclei. Instead, chromatin appears as a disordered, "polymer melt" of interconnected nucleosome chains of variable, often lower, local density — a flexible, irregularly folded fiber rather than a rigid uniform solenoid. Current consensus treats the 30-nm fiber as a real, demonstrable in vitro folding motif and a useful conceptual intermediate, but not the dominant physiological state of bulk interphase chromatin — the story is one of dynamic, heterogeneous folding rather than a single fixed higher-order structure.

Step 3 — Loop extrusion by cohesin. Independent of, and layered on top of, local fiber folding, chromatin forms loops by an active, ATP-driven mechanism. The cohesin complex — a ring made of SMC1, SMC3, RAD21 (kleisin), and STAG1/2 (SA1/2) — loads onto chromatin (aided by the loader NIPBL–MAU2) and uses ATP hydrolysis to reel DNA through itself, extruding a progressively growing loop symmetrically from both sides of the ring. Extrusion continues until the cohesin ring encounters a boundary element: the zinc-finger protein CTCF bound to its DNA recognition motif. CTCF sites are directional (the motif has an orientation), and cohesin extrusion halts efficiently only when it meets a CTCF site in convergent orientation relative to the other anchor — explaining why most stable loops connect CTCF sites pointing toward each other. The cohesin-unloading factor WAPL eventually releases cohesin, making loops dynamic rather than permanent; loss of WAPL causes loop over-extension ("vermicelli" chromosomes).

Step 4 — TADs and compartments. The nested output of many overlapping extruded loops is a topologically associating domain (TAD): a chromosomal region, typically hundreds of kilobases to a few megabases, within which DNA sequences contact each other far more frequently than they contact sequences outside the domain. TAD boundaries are enriched for CTCF and cohesin binding. Functionally, TADs constrain enhancer-promoter communication — an enhancer inside a TAD preferentially finds promoters within the same TAD, insulated from enhancers in neighboring TADs. At a still larger scale (megabases to tens of megabases), chromosomal regions segregate into two spatial compartments visible in genome-wide contact maps — A compartments (open, active, euchromatic, associate with other A regions) and B compartments (closed, heterochromatic, associate with other B regions), reflecting a phase-separation-like spatial demixing of active and inactive chromatin, independent of the cohesin/CTCF loop-extrusion machinery (compartments persist, or are even enhanced, in cohesin-depleted cells).

Step 5 — Establishing euchromatin versus heterochromatin. Chromatin state is written and read by combinations of enzymes:

• Writers deposit modifications: histone acetyltransferases (HATs) add acetyl groups to lysines (e.g., H3K9ac, H4K16ac), neutralizing the lysine's positive charge, weakening histone-DNA and nucleosome-nucleosome contacts, and opening chromatin. Histone methyltransferases such as SUV39H1 deposit H3K9me3 (a hallmark of constitutive heterochromatin), while EZH2 (within the Polycomb Repressive Complex 2, PRC2) deposits H3K27me3 (a hallmark of facultative heterochromatin).
• Erasers remove marks: histone deacetylases (HDACs) strip acetyl groups, re-condensing chromatin; histone demethylases (e.g., LSD1, JMJD family) remove methyl marks.
• Readers bind marks and propagate the state: HP1 (heterochromatin protein 1) binds H3K9me3 via its chromodomain, then dimerizes and recruits SUV39H1 to methylate neighboring nucleosomes, spreading the mark along the chromosome in a self-templating, positive-feedback loop — a molecular explanation for position-effect variegation, where a gene relocated next to heterochromatin is stochastically silenced in a clonal, spreading pattern.

Step 6 — Constitutive versus facultative heterochromatin, and X-inactivation as the model case. Constitutive heterochromatin (pericentromeric and telomeric repeats) is maintained by the H3K9me3–HP1–SUV39H1 loop in essentially all cell types, silencing repetitive and transposon-derived sequences to protect genome integrity. Facultative heterochromatin is cell-type-specific and developmentally programmed. The clearest example is the mammalian inactive X chromosome: the long non-coding RNA Xist, transcribed from one X chromosome early in female development, coats that chromosome in cis, recruits PRC2 to deposit H3K27me3, excludes RNA polymerase II, and drives the entire chromosome into a compact, late-replicating Barr body — silencing roughly 1,000 genes on one X while leaving the homologous genes active on the other X, achieving dosage compensation between XX females and XY males.

Heitz (1928) — the cytological discovery of heterochromatin. Emil Heitz, examining moss (Pellia) and other plant chromosomes stained with basic dyes, noticed that some chromosomal regions remained densely stained and condensed throughout interphase, while most chromatin decondensed and stained lightly once mitosis ended. He coined the terms heterochromatin (stays condensed) and euchromatin (decondenses) purely from microscopy, decades before any molecular mechanism was known. This was a landmark observation because it established, from cytology alone, that chromatin exists in at least two heritable structural states correlated with gene activity — the phenomenological foundation for everything mechanistic described above.

Electron microscopy and nuclease digestion — defining fiber levels. In the 1970s, chromatin spread on grids and shadowed or negatively stained for electron microscopy revealed the 10-nm "beads-on-a-string" appearance (Olins and Olins, 1974; Woodcock, 1973), directly visualizing nucleosomes for the first time. Subsequent work by Finch and Klug (1976), using purified chromatin fibers under physiological salt, showed condensation into a thicker ~30-nm fiber, and proposed the solenoid model. Micrococcal nuclease digestion of chromatin — cutting linker DNA preferentially over nucleosome-protected DNA — produced a ladder of DNA fragments in multiples of the nucleosome repeat length (roughly 200 bp per step) on a gel, providing biochemical confirmation of the regularly repeating nucleosome array underlying the fiber. Decades later, cryo-electron tomography of intact, vitrified nuclei (Eltsov et al., 2008; Ou et al., 2017, using ChromEMT staining) found no widespread regular 30-nm fibers in vivo, instead showing chromatin as flexible, disordered 5–24 nm diameter chains at variable local densities — the key evidence that reoriented the field away from a rigid universal 30-nm solenoid model toward a "polymer melt" view.

Position-effect variegation — heterochromatin spreading is a real, stochastic, propagating process. In Drosophila, chromosomal inversions or translocations that relocate the white gene (normally producing red eye pigment) next to pericentromeric heterochromatin produce eyes with patches of red and white — a mottled, variegated phenotype, first studied extensively by Hermann Muller in the 1930s. This variegation arises because heterochromatin spreads into the relocated white gene in some cell lineages but not others, stochastically, then that state clonally propagates through cell division. Genetic screens for modifiers of this variegation identified Su(var) genes — several of which turned out to encode SUV39 methyltransferase homologs and HP1 itself — directly linking a classical genetic phenomenon to the molecular machinery of H3K9 methylation and HP1 spreading described in the mechanism above.

Chromosome conformation capture (3C) and Hi-C — genome-wide proof of loops and TADs. Job Dekker and colleagues (2002) developed 3C, in which chromatin is crosslinked with formaldehyde (freezing spatially proximal DNA segments together), digested with a restriction enzyme, then re-ligated under dilute conditions so that crosslinked, spatially adjacent fragments preferentially ligate to each other rather than to distant fragments. The resulting ligation frequency between any two loci is a direct proxy for their 3D spatial proximity. Scaling 3C genome-wide with high-throughput sequencing produced Hi-C (Lieberman-Aiden et al., 2009), generating genome-wide contact maps. Hi-C first revealed the A/B compartment structure (alternating active/inactive megabase domains) and, at finer resolution (Dixon et al., 2012; Nora et al., 2012), revealed TADs as squares of enriched self-contact along the diagonal of the contact matrix, with sharp boundaries enriched for CTCF binding.

Depletion experiments — proving cohesin and CTCF cause loops, not just correlate with them. The causal role of cohesin and CTCF was nailed down by acute protein degradation. Using auxin-inducible degron systems to rapidly deplete cohesin subunits (Rao et al., 2017) or CTCF (Nora et al., 2017) within hours, researchers showed that Hi-C loops and TADs disappear almost immediately upon depletion, while A/B compartments persist or even sharpen. Re-expressing the protein restored loops within hours. This directly demonstrated that cohesin and CTCF actively generate loops and TADs in real time — not merely mark pre-existing structures — while compartmentalization arises from an independent, likely self-association/phase-separation-based mechanism.

Example 1 — Reading a Hi-C contact map to diagnose a disease-causing structural variant. A patient with a limb malformation (polydactyly, extra digits) carries a heterozygous 1.5 Mb deletion on chromosome 2, removing a CTCF-bound TAD boundary that normally separates the epha4 gene's regulatory domain from a neighboring TAD containing several limb enhancers (a real, well-studied case; the actual disease locus involves the EPHA4/*WNT6/IHH TAD in mice and humans). Before the deletion: the enhancer, inside its own TAD, only contacts promoters within that TAD (its contact map shows a discrete high-contact square bounded by the CTCF site). After the deletion: the boundary is gone, and Hi-C shows the two TADs fuse into a single large domain — the enhancer, no longer insulated, now makes ectopic contact with EPHA4 in the neighboring domain, driving EPHA4* misexpression during limb development. Reasoning through this: the disease is not caused by a mutation in any coding sequence at all — the deleted DNA contains no genes — but purely by loss of the insulating loop-boundary, illustrating that TAD architecture itself is functionally load-bearing, and its disruption is sufficient to cause a congenital malformation.

Example 2 — Quantifying compaction across the hierarchy. Estimate the total linear compaction from naked B-form DNA to the metaphase chromosome, working level by level.

• Naked DNA: 0.34 nm rise per bp.
• A stretch of 200 bp (one nucleosome + typical linker) = 200 × 0.34 nm = 68 nm of naked DNA.
• Wrapped into one nucleosome (10-nm fiber): the nucleosome core particle itself is about 11 nm across; the 200 bp unit occupies roughly 11 nm of fiber length once packaged — compaction factor ≈ 68/11 ≈ 6-fold.
• Folded into the ~30-nm fiber: six nucleosomes pack per ~11 nm of fiber rise in the classic solenoid model, giving a further ~6-fold compaction, so ~36-fold cumulative from naked DNA.
• Looping/TAD organization and further mitotic condensation (addressed fully in Chromosome Structure and the Mitotic Chromosome) compacts this by roughly another 250- to 300-fold, to reach the final overall compaction of a metaphase chromosome, commonly cited as approximately 10,000-fold relative to naked B-form DNA.

Sanity check against real numbers: human chromosome 1 contains about 249 million bp of naked DNA, which at 0.34 nm/bp would stretch to 249,000,000 × 0.34 nm ≈ 8.5 cm if fully extended — yet the condensed metaphase chromosome 1 is only about 10 micrometers long, a compaction ratio of 8.5 cm / 10 μm ≈ 8,500-fold, closely matching the ~10,000-fold estimate built up from the hierarchical folding steps above.

Emil Heitz was a German botanist and cytologist working in the 1920s at a time when chromosomes could only be studied by light microscopy and histological staining — decades before DNA was even confirmed as the genetic material. Working with mosses and liverworts, Heitz systematically compared chromosome staining before, during, and after mitosis across cell types and species. His key insight was recognizing a stable, heritable structural distinction between chromosomal regions — one group of regions that "waited" in a condensed state through interphase while the rest decondensed to carry out normal cellular business. He proposed, remarkably presciently given the tools available, that this heterochromatin was functionally distinct — later work would show it corresponds to repeat-rich, largely gene-poor and gene-silenced DNA. Job Dekker, a molecular biologist whose 3C and Hi-C methods later transformed the field into a genome-wide, sequencing-based science, effectively gave Heitz's qualitative cytological distinction a quantitative, high-resolution molecular readout seven decades later.

Cancer genetics. Recurrent mutations in cohesin subunits (STAG2, RAD21, SMC1A, SMC3) and in CTCF itself are found across many cancers, including bladder cancer, myeloid leukemias, and Ewing sarcoma. Loss of CTCF boundaries can fuse neighboring TADs, allowing oncogene promoters to be captured by inappropriate enhancers — a mechanism directly analogous to the limb-malformation example above, but activating oncogenes like TAL1 in T-cell leukemia or IGF2 in various tumors instead of disrupting limb genes.

Developmental disorders. Structural variants (deletions, duplications, inversions) that remove or relocate TAD boundaries are established causes of congenital limb malformations, and are increasingly screened for in clinical genome sequencing of patients with unexplained developmental phenotypes, since standard exome sequencing (which only reads coding DNA) misses these purely architectural, non-coding lesions.

Epigenetic therapeutics. HDAC inhibitors (e.g., vorinostat, romidepsin) are approved cancer drugs that block histone deacetylases, keeping chromatin acetylated and open, re-activating silenced tumor-suppressor genes in certain lymphomas. DNA methyltransferase inhibitors (azacitidine, decitabine) work on a related principle for myelodysplastic syndromes, exploiting the link between heterochromatin marks and gene silencing.

Reproductive and stem cell biology. Understanding facultative heterochromatin and X-inactivation underlies work on induced pluripotent stem cell reprogramming (which must erase and re-establish large-scale heterochromatin domains) and on why cloned animals and iPSC-derived cells often show incomplete epigenetic reprogramming, a major technical bottleneck in regenerative medicine.

Agriculture and evolutionary genomics. Constitutive heterochromatin silencing of transposable elements is central to genome stability in crop breeding programs, where transposon reactivation during tissue culture or wide hybridization can cause unwanted mutagenesis; understanding HP1/H3K9me3 silencing informs strategies to stabilize genomes of genetically modified or hybrid crop lines.

Ask the tutor: How does the loop-extrusion/TAD system interact with the further compaction that occurs specifically during mitosis to form condensed chromosomes?