DNA Supercoiling

What supercoiling actually is

Take a coiled telephone cord and twist one end extra before you let it go: it springs into loops that fold over each other. DNA does the same thing. The double helix is already a coil — two strands wound around each other once every ~10.5 base pairs in the standard B-form. When something forces that helix to wind tighter (over-wound) or looser (under-wound) than its relaxed state, the molecule relieves the strain by coiling the entire helix axis upon itself. Those higher-order loops are supercoils, and the property of being over- or under-wound is supercoiling.

The reason this matters biologically is that DNA in a cell is not a free-floating noodle that can spin away its strain. In bacteria the chromosome is a single covalently closed circle; in eukaryotes the linear chromosomes are clamped into looped domains anchored to a protein scaffold. In both cases the ends are fixed, so the molecule is topologically constrained — any winding you add or remove is trapped until an enzyme cuts the backbone. That trapped winding is stored torsional energy, and the cell exploits it: under-wound DNA is poised to pop open, which is exactly what every promoter, origin, and repair site needs.

The topology: linking number, twist, and writhe

Supercoiling is governed by one elegant identity from mathematical topology, valid for any closed ribbon:

Lk = Tw + Wr

• Lk — linking number. The number of times one strand crosses the other in a closed molecule. It is an integer topological invariant: you cannot change it by bending, stretching, or wiggling the DNA. The only way to change Lk is to break a strand, pass DNA through the gap, and reseal — which is precisely what topoisomerases do.
• Tw — twist. The local helical winding of the two strands about the duplex axis. Relaxed B-DNA has one twist per ~10.5 base pairs, so a 5,250 bp circle has Tw ≈ 500.
• Wr — writhe. The coiling of the helix axis in 3-D space — the visible loops and crossings of a supercoil. A perfectly relaxed circle laid flat has Wr = 0.

Because Lk is locked, twist and writhe trade off against each other. If you unwind the helix (lower Tw), the axis must writhe by the same amount (raise Wr's magnitude) to keep Lk constant. That is why forcibly opening a few base pairs at a promoter doesn't just relax — it injects a compensating supercoil somewhere else in the loop. The relaxed reference is Lk₀ ≈ (base pairs) / 10.5, and the deviation ΔLk = Lk − Lk₀ is what we colloquially call "the number of supercoils." Negative ΔLk means under-wound; positive means over-wound.

How torsional strain builds and gets drained

The day-to-day source of supercoiling is the molecular machines that have to pry the strands apart. A replication fork or an elongating RNA polymerase unwinds roughly one helical turn for every 10.5 base pairs it advances. Since the DNA ends are anchored, that unwinding cannot dissipate down the molecule — it accumulates as positive (over-wound) supercoils ahead of the moving machine and negative (under-wound) supercoils behind it. This is the twin-supercoiled-domain model of Leroy Liu and James Wang (1987).

The numbers are brutal. A bacterial replication fork moves at about 1,000 base pairs per second; at one turn per 10.5 bp that demands the helix rotate roughly 100 times every second. Without relief, the positive torsion ahead of the fork would climb until the strands physically cannot be separated — replication would stall within a few kilobases. The cell solves this with topoisomerases, enzymes that transiently cleave the backbone, let DNA pass through, and reseal:

• Type I (e.g. E. coli Topo I, eukaryotic Topo I). Nick one strand, let the other strand rotate or pass through the break, reseal. Changes Lk in steps of ±1. Requires no ATP — relaxing a strained molecule is energetically downhill, so the enzyme just harnesses the stored torsion.
• Type II (e.g. DNA gyrase, Topo IV, eukaryotic Topo II). Cut both strands to make a transient double-strand gate, pass an intact duplex through it, reseal. Changes Lk in steps of ±2 and consumes ATP (~2 ATP per cycle). Type II enzymes can also unlink intertwined daughter chromosomes (decatenation) and untie knots — jobs a single-strand nick cannot do.
• DNA gyrase (a special type II, bacteria only). Uses ATP not merely to relax but to actively introduce negative supercoils, wrapping ~130 bp of DNA in a positive loop and inverting it through the gate. Gyrase is what keeps the bacterial genome under-wound; eukaryotes have no gyrase and instead store negative supercoiling in nucleosomes.

The under-wound steady state and who maintains it

Cells do not aim for relaxed DNA — they deliberately keep the genome negatively supercoiled, at a superhelical density σ (sigma) of about -0.05 to -0.07. Negative supercoiling stores free energy that helps the helix open locally, so promoters fire more easily, origins melt more readily, and recombination intermediates form. The cell runs the genome like a slightly loosened spring, ready to release.

In bacteria this is a homeostatic tug-of-war: DNA gyrase pumps negative supercoils in (using ATP), while Topo I relaxes excess negative supercoils. The ratio of gyrase to Topo I activity sets σ, and the genes for both enzymes are themselves regulated by supercoiling — a built-in feedback loop. Stresses such as anaerobic shift, osmotic shock, or temperature change alter σ and reprogram large blocks of genes, so supercoiling acts as a global, fast-acting transcriptional switch.

Eukaryotes get net negative supercoiling for free from chromatin. Each nucleosome wraps ~147 base pairs of DNA in 1.65 left-handed turns around the histone octamer, constraining about -1 supercoil per nucleosome. With a nucleosome every ~200 bp, a chromosome stores enormous negative writhe in its packaging rather than as free torsional strain — which is why eukaryotes never needed a gyrase. The famous "linking number paradox" (nucleosomes look like they wrap ~1.65 turns but constrain only ~1.0) is resolved by the over-twisting of DNA on the histone surface.

Negative vs positive supercoiling

Property	Negative (under-wound)	Positive (over-wound)
Sign of ΔLk	ΔLk < 0 (fewer turns)	ΔLk > 0 (extra turns)
Effect on helix	Easier to unwind / melt	Harder to unwind, more rigid
Typical in-vivo state	σ ≈ -0.06 (most cells)	Transient, ahead of fork/polymerase
Biological role	Primes promoters, origins, recombination	By-product of tracking machines; must be removed
Who creates it	DNA gyrase (ATP); nucleosomes (eukaryotes)	Replication forks, RNA polymerase
Who removes it	Topo I (bacterial), gyrase reversal	Gyrase, Topo IV, eukaryotic Topo I & II
Extreme structures	Z-DNA, cruciforms, R-loops, strand opening	P-DNA (over-stretched), helix seizing
Thermophiles	Often near zero or positive (reverse gyrase)	Stabilizes DNA at high temperature

The numbers: sizes, energies, and rates

• Helical repeat. Relaxed B-DNA: ~10.5 bp per turn in solution (10.0 in the crystal). So Lk₀ for the 4.6 million bp E. coli chromosome is about 438,000, and the cell holds it ~6 percent below that.
• Superhelical density. σ = ΔLk / Lk₀ ≈ -0.06 in vivo. The classic plasmid pBR322 (4,361 bp) carries roughly -25 supercoils when extracted.
• Energy of supercoiling. The free energy stored scales as ΔG ≈ (1100 RT / N) × ΔLk², where N is the number of base pairs — on the order of 10 RT per supercoil for a few-kilobase plasmid. Negative supercoiling stores enough free energy to pay much of the cost of opening a promoter (~10–12 bp of melted DNA).
• Replication torsion rate. Fork at ~1000 bp/s ⇒ ~95 turns/s of unwinding to be removed; gyrase can relax at ~100 supercoils/min per enzyme, so dozens of gyrase molecules work each fork.
• Topoisomerase steps. Type I changes Lk by ±1 per cycle; type II by ±2 per cycle at ~2 ATP each. Type II decatenation prevents the catastrophic anaphase bridge that would otherwise tear sister chromosomes apart.
• Nucleosome constraint. 147 bp wrapped in ~1.65 turns → ~ -1 constrained supercoil each; ~30 million nucleosomes in a human cell store the genome's negative writhe.

Where it shows up: disease, antibiotics, and chemotherapy

• Fluoroquinolone antibiotics. Ciprofloxacin, levofloxacin and relatives bind the gyrase–DNA and Topo IV–DNA complexes after the enzyme has cut both strands but before it religates, freezing a covalent enzyme-bridged double-strand break. The trapped breaks are lethal to dividing bacteria. Resistance arises from point mutations in the gyrase A subunit (the "quinolone resistance-determining region").
• Topoisomerase-poison chemotherapy. Etoposide and teniposide trap human Topo II, while camptothecin derivatives (irinotecan, topotecan) trap human Topo I, and doxorubicin both intercalates and poisons Topo II. All convert an essential enzyme into a generator of genome-wide double- or single-strand breaks, killing fast-dividing tumor cells — at the cost of secondary leukemias from mis-rejoined breaks.
• Z-DNA and disease. Negative supercoiling can flip GC-rich tracts into left-handed Z-DNA, which is recognized by the Z-DNA-binding domain of the innate-immunity sensor ZBP1 and the editing enzyme ADAR1; mis-regulation is implicated in inflammatory disease.
• R-loops and genome instability. Negative supercoiling behind a transcribing polymerase favors the nascent RNA re-invading the template to form an R-loop. R-loops regulate transcription and replication but, when persistent, cause DNA damage linked to neurodegeneration and cancer; topoisomerases and the enzyme RNase H help resolve them.
• Hyperthermophiles. Archaea like Sulfolobus living near 80 °C use a unique reverse gyrase — a type IA topoisomerase fused to a helicase — to introduce positive supercoils, stiffening the helix against heat denaturation. It is the only known enzyme that creates positive supercoils.

Common misconceptions and pitfalls

• "Supercoiling is just DNA being messy or tangled." No — supercoiling is a precise, quantized, sign-defined property governed by Lk = Tw + Wr. Tangles and knots are a separate topological problem (also handled by type II enzymes), not the same as the smooth over/under-winding of supercoiling.
• "Twist and writhe are the same thing." They are distinct components. Twist is the strands winding around the duplex axis; writhe is the axis winding around itself. A molecule can have all its strain as twist (a flat, over-wound ring) or convert it to writhe (a coiled plectoneme) without changing Lk at all.
• "Topoisomerases use ATP to relax DNA." Only type II enzymes hydrolyze ATP, and gyrase uses it to add negative supercoils, not merely relax. Type I relaxation is ATP-independent because releasing stored torsion is downhill.
• "Linear DNA can't supercoil, so eukaryotic chromosomes don't supercoil." Free linear DNA cannot, but cellular chromatin is partitioned into anchored loop domains that are topologically closed. A transcribing polymerase in a human cell generates the same twin domains of supercoiling within its loop.
• "More base pairs always means more supercoils." What matters is superhelical density σ (supercoils normalized to length), not raw ΔLk. A small plasmid at σ = -0.06 and a giant chromosome at σ = -0.06 are equally strained per turn.
• "Cells want relaxed DNA." They don't. The under-wound σ ≈ -0.06 set point is functional, storing energy to open promoters and origins. Drugs that fully relax bacterial DNA (or that block gyrase) are lethal precisely because relaxed DNA can't do its job.