Nucleosome

A bead the cell uses two metres of

Each of the four core histones is small (102–135 residues) and shares the same fold — three alpha helices linked by two short loops, a motif called the histone fold. Two histone-fold pairs interlock head-to-tail to make a dimer; in the octamer, an H3-H4 pair handshakes another H3-H4 to form a central tetramer, and one H2A-H2B dimer docks onto each face. The four-helix bundles of H3 and H4 sit at the tetramer's core, and H2A-H2B dimers latch on the sides like wings. The whole assembly is symmetric — flip it and the structure looks identical.

The DNA wrap is left-handed and 1.65 turns long, covering 147 base pairs. Fourteen contacts hold it in place: at every full helical turn (about every 10 bp) the DNA minor groove faces the histone surface, where an arginine from one of the histones reaches in and bonds the phosphate backbone. There are no sequence-specific contacts — the histones grip whatever DNA is presented — but they prefer flexible sequences and reject stiff poly(dA:dT) tracts.

This wrap is the cell's first compaction step. 147 bp of B-form DNA is 50 nm long stretched out; wound around a 7-nm-tall octamer it spans about 11 nm. Multiply that across the 6 billion base pairs of a diploid genome and you have already shrunk DNA roughly seven-fold. Stack the nucleosomes into a 30-nm fibre with H1 and you reach about 50-fold; loop that fibre on the chromosome scaffold and you reach 10,000-fold compaction at metaphase.

Core composition: who is what, where

Component	Stoichiometry	Length (residues)	Position in particle	Distinguishing feature
Histone H3	2 copies	135	Central tetramer core, contacts H4	Long N-terminal tail (residues 1-40); CENP-A variant at centromeres
Histone H4	2 copies	102	Central tetramer core, contacts H3	Most evolutionarily conserved protein known; H4K16ac is the global accessibility switch
Histone H2A	2 copies	129	Lateral dimers, contacts H2B	Long C-terminal tail; H2A.X is phosphorylated at DSBs (γ-H2AX); H2A.Z marks promoters
Histone H2B	2 copies	126	Lateral dimers, contacts H2A	Short tail; H2BK120ub on gene bodies activates transcription
DNA	147 bp	—	Wraps 1.65 left-handed turns around octamer	Minor groove faces histone surface every ~10 bp
Histone H1 (linker)	0 or 1 per nucleosome	~220	Caps entry/exit DNA	Forms the chromatosome (167 bp protected); 11 mammalian variants

The numbers above are the textbook ones. The reality has a few wrinkles: H1 occupancy is sub-stoichiometric in active chromatin, the linker length varies (yeast averages ~20 bp, mammals ~50 bp, sea-urchin sperm over 100 bp), and many chromosomal regions use histone variants that change the rules.

Assembly behind the replication fork

New nucleosomes are not built spontaneously. After replication, the parental octamers split — roughly half stay with the leading-strand daughter and half with the lagging strand — and new histones (synthesised in S phase by an entire cluster of replication-coupled genes) fill the gaps. The chaperone CAF-1 binds PCNA at the fork and delivers the H3-H4 tetramer first; NAP1 and the FACT complex then add the H2A-H2B dimers. The whole process takes seconds and runs in lockstep with replication.

Two consequences follow. First, half the histones on a daughter chromosome are old (and therefore carry the parental modifications), which is the substrate that lets reader-writer enzymes copy the marks onto new histones in the hours after replication. Second, the cell needs vast quantities of new histones every S phase — histones are among the most abundant proteins synthesised during cell division, and disrupting their cluster (HIST1 on chromosome 6 in humans) is lethal.

How nucleosomes are positioned

If nucleosomes were placed randomly, the cell could not know in advance whether a given gene was accessible. In practice positioning is highly non-random and determined by three forces. Sequence preferences favour flexible GC-rich DNA and reject stiff poly(dA:dT) tracts — these tracts at promoters create nucleosome-free regions a few hundred base pairs wide. Statistical positioning packs subsequent nucleosomes against any barrier; a transcription factor or a stable nucleosome anchors a phased array of nucleosomes downstream. ATP-dependent remodellers — SWI/SNF, ISWI, CHD, INO80 — burn ATP to slide, eject, or load nucleosomes, often against the local thermodynamic preference.

The result is a stereotyped layout at most genes: a nucleosome-free region (NFR) ~150 bp wide at the promoter, flanked by a +1 nucleosome positioned over the transcription start site and a -1 nucleosome upstream. RNA polymerase II then has to displace the +1 nucleosome to begin transcription — which is one reason the +1 is enriched for the variant H2A.Z and H3.3, both of which destabilise the wrap.

Eukaryotic nucleosome vs bacterial nucleoid

	Eukaryotic nucleosome	Bacterial nucleoid
Packaging proteins	Histones (H2A, H2B, H3, H4) + H1	HU, IHF, H-NS, Fis (NAPs)
Wrapping geometry	147 bp, 1.65 left-handed turns	Variable; HU bends DNA but does not wrap fully
Compaction	~10,000× at metaphase	~1,000× in the nucleoid
Regulation	Post-translational marks on histone tails	Supercoiling, NAP binding, growth-phase control
Replication-coupled assembly	CAF-1 deposits new H3-H4 at the fork	None; nucleoid is dynamic, not modular
Variants	CENP-A, H3.3, H2A.Z, macroH2A, H2A.X	Different NAPs at different growth phases

Archaea blur this distinction. They use proteins related to eukaryotic histones (HMfA, HMfB) and form roughly 60-bp wraps in tetramer particles — looking like an evolutionary midpoint between bacterial nucleoids and eukaryotic nucleosomes.

Why this structure matters

• Cancer. Mutations in histones themselves are oncogenic. H3K27M (lysine-to-methionine) in pediatric diffuse intrinsic pontine glioma poisons EZH2 and globally erases H3K27me3; H3G34V in giant-cell tumour of bone disrupts H3K36 methylation locally. Both are gain-of-function, dominant-negative oncohistones.
• Centromeres. The kinetochore — the platform that grabs spindle microtubules during mitosis — assembles only on nucleosomes containing the H3 variant CENP-A. Without CENP-A loading at G1, sister chromatids cannot segregate.
• Pioneer factors. Most transcription factors cannot bind DNA wrapped on a nucleosome. A small set — FOXA1, OCT4, GATA, KLF4 — can, and they recruit remodellers to evict the nucleosome, opening chromatin for the next round of factors. They are why developmental cell-fate decisions can override pre-existing chromatin.
• Drug action. Topoisomerase poisons (etoposide, doxorubicin), HDAC inhibitors (vorinostat), BET-bromodomain inhibitors (JQ1) and EZH2 inhibitors (tazemetostat) all act through the nucleosome's contacts and modifications.
• Sperm chromatin. Sperm package most of their DNA on protamines instead of histones, achieving denser compaction; the small fraction (~5–15% in humans) that remains on nucleosomes is at developmentally important genes — an inheritable signal beyond the genetic code.

Histone variants and their niches

• CENP-A (CenH3). Replaces H3 at every functional centromere. Required for kinetochore assembly. Loaded by HJURP in early G1.
• H3.3. Replication-independent variant deposited by HIRA at active genes and by ATRX/DAXX at telomeres and pericentric heterochromatin. Differs from canonical H3 by only four residues but is functionally distinct.
• H2A.Z. Marks promoters, enhancers, and the +1 nucleosome of active genes; destabilises the wrap and primes the locus for activation.
• H2A.X. Phosphorylated at S139 to γ-H2AX within minutes of a DNA double-strand break, marking megabases around the lesion for repair-factor recruitment.
• macroH2A. Carries an extra C-terminal macrodomain; coats the inactive X chromosome and senescent heterochromatin.
• TH2B and protamines. Replace H2B and the entire octamer respectively during spermiogenesis, achieving dense, transcriptionally silent packaging in sperm.

Pitfalls and easy misreadings

• "DNA is just stored in nucleosomes." Nucleosomes do not store DNA, they regulate access to it. Every step of transcription, replication, and repair has to negotiate them.
• "The 30-nm fibre is the in-vivo state." Recent ChromEM and cryo-electron tomography work suggests the 30-nm fibre may be a salt-induced artefact and that interphase chromatin is more like a disordered liquid than a regular helical fibre. The textbook picture is still useful as a teaching device but not literal.
• "All four core histones are equivalent." H4 is one of the most conserved proteins in biology — only two residue differences between yeast and human — because H4K16 acetylation is the master internucleosomal contact. H3 has the most variants. H2A and H2B are the most readily exchanged in and out of nucleosomes.
• "Nucleosomes block transcription completely." RNA polymerase II elongates through nucleosomes routinely, with the help of FACT and elongation factors that disassemble and reassemble the wrap behind the polymerase. The block is at initiation, not elongation, in most cases.
• "H1 is just a clip." H1 has 11 variants in humans, each with different tissue expression and chromatin preferences; H1 depletion in mouse embryonic stem cells causes specific transcription defects, not just global decompaction.