Nucleosome Structure

A nucleosome consists of a protein core — the histone octamer — and the DNA wound around it. The octamer is built from two copies each of four small, highly conserved, positively charged proteins: H2A, H2B, H3, and H4. These assemble first into an (H3–H4)₂ tetramer, which then binds two H2A–H2B dimers, one on each face, to form the disc-shaped octamer. Around this disc, 147 base pairs of double-stranded DNA wrap in 1.65 left-handed superhelical turns, contacting the histone surface at roughly 14 discrete points spaced about every 10 base pairs — matching the helical repeat of B-form DNA so that the same face of the DNA backbone always faces the protein. The complex of core particle plus the intervening linker DNA (the stretch connecting one nucleosome to the next, typically 20–90 bp depending on cell type and species) constitutes a nucleosome repeat, historically measured as roughly 165–225 bp per unit in different tissues, with ~200 bp a common textbook average in somatic cells. A fifth histone, the linker histone H1 (or its variant H5 in some cells), binds at the DNA entry/exit point and the adjoining linker DNA, helping to seal the two turns and promote further folding into higher-order fibers — a topic taken up separately. It is essential to distinguish two nested levels of description. The nucleosome core particle is the minimal, biochemically stable unit isolated after nuclease trimming: octamer plus 147 bp. The nucleosome in the broader chromatin sense includes some or all of the linker DNA and is the true repeating unit visualized as "beads on a string" in the electron microscope. Wrapping DNA into nucleosomes achieves a compaction ratio of about sevenfold in linear length at this first level alone (147 bp of B-DNA, which is ~50 nm of extended contour length, condensed to a particle ~10 nm in diameter and ~6 nm thick) — the first of several hierarchical folding steps that together compact the genome roughly 10,000-fold.

Step 1 — Histone fold and dimer formation. Each core histone shares a common tertiary structural motif, the histone fold: three alpha helices (α1–α2–α3) connected by two loops, arranged so that pairs of histones associate head-to-tail in a "handshake" motif. H3 pairs with H4, and H2A pairs with H2B, each forming a stable heterodimer through this handshake interface. All four core histones are small (roughly 11–15 kDa) and unusually rich in the basic residues lysine and arginine, which neutralize the negatively charged DNA phosphate backbone.

Step 2 — Tetramer and octamer assembly. Two H3–H4 dimers associate through a four-helix bundle formed by their two H3 subunits, creating a stable (H3–H4)₂ tetramer. This tetramer alone can organize roughly 3/4 turn of DNA and is deposited first during replication-coupled nucleosome assembly. Two H2A–H2B dimers then dock onto this tetramer, one on each side, via a second four-helix-bundle interface (this time between H4 and H2B), completing the octamer. The octamer is not a symmetric hexamer or a random aggregate — it has precise twofold (C2) symmetry about an axis, called the dyad axis, that passes through the center of the H3–H3′ four-helix bundle and through the midpoint of the wrapped DNA.

Step 3 — DNA wrapping. DNA does not wrap passively; it is bent sharply and repeatedly around the octamer's superhelical path, with a diameter of roughly 8.4 nm (measured pitch-to-pitch of the turns is about 2.4–2.6 nm). This bending is energetically costly for a stiff double helix (persistence length ~50 nm, far greater than the ~5 nm radius of curvature demanded here), and it is precisely the histone octamer's curved, positively charged surface that lowers this energetic cost by neutralizing phosphate charge repulsion and providing a scaffold that matches the required curvature. Roughly 121 direct hydrogen bonds and numerous electrostatic contacts occur between histone side chains/backbone amides and the DNA phosphate backbone, distributed at ~10 bp intervals corresponding to the points where the DNA minor groove faces inward toward the octamer.

Step 4 — Sequence preference without sequence specificity. Because contacts are made mostly with the sugar-phosphate backbone rather than with bases in the major groove, nucleosomes do not require a strict consensus sequence to form. However, bending 147 bp this sharply is easier at some sequences than others: AT-rich stretches tend to sit where the minor groove faces inward (easier to compress), and GC-rich stretches favor the minor groove facing outward. This gives nucleosomes a statistical, rotationally phased sequence preference (a ~10 bp periodicity of dinucleotide steps like AA/TT/TA alternating with GC steps) rather than a fixed binding motif — in contrast to sequence-specific DNA-binding proteins like transcription factors.

Step 5 — Linker DNA, H1, and higher packing. The DNA entering and exiting the core particle is the linker DNA. Linker histone H1 binds near the dyad axis at the entry/exit point, clamping the two DNA strands together and organizing an additional ~20 bp, which promotes folding of the nucleosome array into a more condensed structure (historically described as a 30 nm fiber, now understood to be a more heterogeneous, irregularly folded state in vivo — foreshadowing the discussion of higher-order chromatin folding).

Step 6 — Dynamics: not a static clamp. Nucleosomes are not frozen. DNA can partially unwrap and rewrap from the ends ("breathing"), transiently exposing internal DNA — measured by biophysical assays to occur on a timescale of tens of milliseconds — allowing transcription factors momentary access without full nucleosome removal. ATP-dependent chromatin remodeling complexes (e.g., SWI/SNF-family enzymes) can slide, evict, or restructure nucleosomes, and this dynamic repositioning is a major regulatory layer built directly on top of the static structure described here.

1. Electron microscopy: "beads on a string" (Olins & Olins, 1974; Woodcock, 1973). When chromatin was gently spread from lysed nuclei onto grids at low ionic strength and examined by transmission electron microscopy, Ada and Donald Olins, and independently Christopher Woodcock, observed a regular array of roughly spherical particles about 10 nm in diameter, connected by thin filaments of naked-looking DNA. This directly falsified the then-prevailing idea of chromatin as a smooth, uniformly coated fiber, and instead showed DNA organized into discrete, repeating particles — the first direct visual evidence for the nucleosome.

2. Micrococcal nuclease (MNase) digestion and the DNA "ladder" (Hewish & Burgoyne, 1973; Noll, 1974). Donald Hewish and Leigh Burgoyne treated isolated nuclei with an endogenous Ca²⁺/Mg²⁺-dependent nuclease and found that chromosomal DNA was cleaved preferentially at regularly spaced sites, producing DNA fragments in multiples of a fixed unit length when run on a gel — a "ladder" of bands at roughly 200, 400, 600, 800 bp and so on. Markus Noll extended this using purified micrococcal nuclease (MNase), which digests the exposed linker DNA between particles far faster than the DNA wrapped tightly around the histone core. Brief digestion released oligonucleosomes (multiples of ~200 bp); extensive digestion trimmed the linkers away entirely, converging on a limit product of ~146–147 bp, the mononucleosome core particle. This was decisive quantitative evidence: it showed that (a) chromatin DNA is organized into a repeating unit of fixed size, and (b) that unit has a nuclease-protected core surrounded by more exposed linker DNA — exactly what the beads-on-a-string model predicted, now given precise molecular dimensions.

3. Protein composition and the histone octamer stoichiometry (Kornberg, 1974; Kornberg & Thomas, 1974). Roger Kornberg synthesized the emerging biochemical and structural data — X-ray diffraction patterns of chromatin showing a repeating ~10 nm periodicity, cross-linking experiments (Thomas & Kornberg) showing histones exist as an octameric complex in solution, and the MNase digestion data — to propose in 1974 the nucleosome model essentially as understood today: DNA wound around a histone octamer of two copies each of H2A, H2B, H3, and H4. Chemical cross-linking of histones in intact chromatin, using reagents like dimethyl suberimidate, showed histones could be trapped as dimers, tetramers, and ultimately octamer-sized complexes, directly demonstrating the correct stoichiometry rather than an arbitrary assumption.

4. Atomic structure by X-ray crystallography (Richmond, Luger, et al., 1984–1997). Timothy Richmond and colleagues obtained early low-resolution (7 Å) crystal structures of the nucleosome core particle in the 1980s, sufficient to confirm the overall wrapping but not to resolve individual atoms or side chains. In 1997, Karolin Luger, Richmond, and coworkers solved the structure at 2.8 Å resolution, using a core particle reconstituted from recombinant Xenopus histones and a defined 146-bp DNA sequence. This structure directly visualized: the histone-fold handshake motifs; the exact left-handed superhelical path of the DNA (1.65 turns, 147 bp); the ~14 discrete histone–DNA contact points at ~10 bp intervals; and the protruding, unstructured N-terminal histone tails extending out of the DNA gyres — the very tails later shown to be extensively modified in the Histones and the Histone Code. This crystal structure remains the reference structure cited in essentially all subsequent chromatin biology. Together, these four independent lines of evidence — direct imaging, controlled enzymatic digestion, cross-linking stoichiometry, and atomic-resolution crystallography — converged on the same structural model, a textbook example of triangulating a molecular structure from orthogonal experimental methods.

Example 1 — Quantifying genome compaction by nucleosomes. The human diploid genome is about 6.4 × 10⁹ bp. Using the standard conversion of 0.34 nm rise per base pair for B-form DNA, the fully extended contour length is: 6.4 × 10⁹ bp × 0.34 nm/bp = 2.18 × 10⁹ nm ≈ 2.18 meters. Now consider only the nucleosome-level compaction. A nucleosome repeat length of ~200 bp (147 bp core + ~53 bp linker, a representative somatic-cell average) corresponds to a core particle roughly 11 nm in diameter, stacked into an array. Empirically, nucleosome arrays measured by EM run about 1/6 to 1/7 the length of the equivalent naked, extended DNA — i.e., ~200 bp of extended DNA (68 nm) is compacted into a linear array segment of roughly 10–11 nm. Compaction ratio ≈ 68 nm / 11 nm ≈ 6-fold, purely from nucleosome wrapping (before any 30 nm fiber-type folding or the further loop/domain and mitotic-chromosome compaction levels are included). Multiplying by the additional ~5- to 6-fold compaction attributed to higher-order fiber folding, and further orders of magnitude from loop domains and mitotic condensation, gives the commonly cited overall figure of ~10,000-fold compaction from naked DNA to a metaphase chromosome — nucleosome wrapping is the essential first step in that cascade, without which none of the subsequent folding levels would have a substrate to act on.

Example 2 — Reading an MNase digestion time-course to infer chromatin state. A researcher digests chromatin from two cell types, A and B, with MNase for identical short times and quantifies band intensities on a gel:

Cell type A shows rapid conversion mostly to mononucleosomes with little sub-nucleosomal degradation — consistent with well-positioned, closely and regularly spaced nucleosomes protecting linker DNA fairly uniformly, typical of transcriptionally quiescent, compact heterochromatin, where regular spacing allows MNase to cut cleanly at each exposed linker before it can attack a core. Cell type B shows a broader distribution, slower accumulation of mononucleosomes, and a substantial sub-nucleosomal fraction — consistent with irregular nucleosome spacing, nucleosome-depleted regions, and MNase over-digestion into individual cores at accessible, actively transcribed loci (e.g., promoters and enhancers with unstable or absent nucleosomes, allowing MNase to nibble further before the core itself is finally attacked). This kind of MNase-seq/MNase-digestion comparison is precisely how modern genomics infers nucleosome occupancy and positioning genome-wide, extending the original 1970s gel-based logic to sequencing resolution.

Roger Kornberg trained as a physical chemist before turning to chromatin structure as a postdoctoral fellow with Aaron Klug at the MRC Laboratory of Molecular Biology in Cambridge in the early 1970s — a lab already steeped in structural approaches to biological macromolecules, from Klug's work on virus structure and tRNA. Kornberg brought a biophysicist's insistence on quantitative, testable structural models to a field that had, until then, mostly described chromatin qualitatively as a fiber of unknown internal order. His 1974 synthesis of X-ray diffraction periodicities, cross-linking stoichiometry data (much of it generated with Jean Thomas), and the emerging MNase digestion ladders from Hewish, Burgoyne, and Noll produced the first explicit, falsifiable nucleosome model — a bold structural proposal made before any atomic structure existed, later vindicated in full detail by crystallography two decades on. Kornberg's father, Arthur Kornberg, had earlier won the Nobel Prize for discovering DNA polymerase; Roger's own Nobel Prize, in 2006, was awarded not for the nucleosome but for his later work solving the structure of eukaryotic RNA polymerase II and elucidating transcription — connecting directly back to the question of how a polymerase must contend with the very nucleosomal packaging he described three decades earlier.

Cancer biology and epigenetic therapy. Many oncogenic mutations target chromatin regulators rather than classical oncogenes directly — for example, mutations in the SWI/SNF nucleosome remodeling complex (subunit SMARCB1/INI1) drive pediatric malignant rhabdoid tumors, and recurrent histone gene mutations (e.g., H3K27M in diffuse midline glioma) alter nucleosome-based signaling. Drugs targeting chromatin-modifying enzymes (HDAC inhibitors, EZH2 inhibitors) act on machinery that reads or writes marks on nucleosomal histones, making nucleosome structure the physical substrate of an entire emerging drug class.

Genome-wide nucleosome mapping. MNase-seq, ATAC-seq, and DNase-seq all exploit the differential nuclease accessibility of linker versus core DNA (the same principle as the original 1970s gel ladders, now read out by high-throughput sequencing) to map nucleosome positions and accessible regulatory regions genome-wide in normal development, disease states, and drug response.

Reproductive and developmental biology. During spermatogenesis, most histones are replaced by protamines, achieving a far tighter DNA packaging than nucleosomes allow — understanding nucleosome structure clarifies exactly what must be disassembled and reassembled during this transition and during the reverse process at fertilization.

Biotechnology and synthetic biology. Recombinant nucleosome reconstitution (the same approach used by Luger and colleagues) is now a standard tool for in vitro assays of chromatin remodelers, histone chaperones, and epigenetic "reader/writer/eraser" enzymes, underpinning much of modern epigenetics drug discovery.

Forensic and ancient DNA science. Because nucleosome-wrapped DNA degrades in a characteristic ~10 bp-periodic pattern (nuclease/chemical damage occurs preferentially where DNA faces outward on the octamer), ancient DNA researchers can infer historical nucleosome positioning from degradation patterns in fossil and archaeological samples, reconstructing epigenomic states of extinct populations.

Ask the tutor: How exactly do ATP-dependent chromatin remodelers slide or evict a nucleosome along DNA?