DNA Supercoiling

Consider a segment of double-stranded DNA whose ends are not free — either because it is a covalently closed circle (as in most bacterial chromosomes and plasmids) or because it is anchored at both ends to a protein scaffold (as in eukaryotic chromatin loops). In such a topologically constrained molecule, the two strands are linked around each other a fixed integer number of times, called the linking number, Lk. Lk counts, in effect, how many times one strand's path crosses the imaginary surface bounded by the other strand — an integer that is a true topological invariant: it cannot change by bending, twisting, or stretching the DNA, only by transiently breaking one or both strands and resealing them. Relaxed B-form DNA has about 10.5 base pairs per helical turn, so a relaxed closed circle of N base pairs has a reference linking number Lk₀ = N/10.5. If the actual linking number Lk differs from Lk₀, the molecule is said to be supercoiled, and the difference is captured by two related quantities:

$\sigma$, the superhelical density, is the standard way to report supercoiling because it is length-independent: a value of σ = −0.06, typical of bacterial chromosomal DNA, means the molecule is underwound by 6% relative to relaxed B-DNA regardless of how many base pairs it contains. Negative σ (Lk < Lk₀, too few helical turns) is called negative supercoiling; positive σ is positive supercoiling. The central equation of the field, first articulated rigorously by James White and Brock Fuller in the early 1970s building on Călugăreanu's mathematical theorem, is:

Twist (Tw) is the number of times the two strands wind around the local helix axis — essentially the helical repeat integrated along the molecule's length. Writhe (Wr) is the number of times the helix axis itself crosses over in three-dimensional space — the coiling of the coil, visible as the interwound, plectonemic superhelix you get when you twist a rubber tube. The crucial point is that Lk is fixed (topological), while Tw and Wr are geometric and can trade off against each other continuously: the same underwound linking number deficit can be expressed as local unwinding of the double helix (reduced Tw, which favors strand melting) or as global coiling of the axis (negative Wr, the classic interwound plectoneme), or any combination. This partition — which the cell can shift back and forth without breaking any bonds — is what makes supercoiling both a store of mechanical energy and a switch that couples global DNA shape to local helix opening.

Step 1 — Generating the strain. Every process that requires transient strand separation — replication and transcription chief among them — locally reduces Tw as the helicase or RNA polymerase unwinds the duplex. But if the ends of the DNA are constrained, Lk cannot change to compensate, so the unwinding at the enzyme is compensated by compensatory over-twisting or positive writhing elsewhere in the molecule. This is the origin of the classic twin-domain model proposed by Liu and Wang (1987): a translocating polymerase or helicase generates positive supercoiling ahead of it (over-winding, since local unwinding must be "paid for" topologically by extra winding elsewhere) and negative supercoiling behind it. Because the enzyme complex itself is large and rotationally hindered (often anchored to other machinery or membranes), the DNA — not the enzyme — is forced to rotate, generating this twin-domain torsional stress.

Step 2 — Partitioning strain between twist and writhe. Once Lk deviates from Lk₀, the excess (or deficit) distributes between Tw and Wr according to the relative energetic costs of each deformation, roughly in a 2:1 ratio favoring writhe in typical circular plasmids (i.e., about two-thirds of the linking number difference is absorbed as writhe, one-third as altered twist) — though the exact partition depends on ionic conditions, DNA length, and protein binding. Under substantial negative supercoiling, some of that deficit can instead be absorbed by locally melting a few base pairs (unwinding, i.e., extreme local reduction of twist) or by a local transition to alternative conformations, since negative supercoiling favors any structure with lower twist than B-DNA — including strand-separated bubbles, and, in G/C-rich alternating sequences, left-handed Z-DNA (DNA Conformations: A, B, and Z Forms), which has an intrinsically negative and opposite-handed twist relative to B-DNA.

Step 3 — Three-dimensional shape: plectonemes. When writhe dominates, negatively supercoiled circular DNA does not curl into a simple flat circle; it collapses into an interwound, plectonemic superhelix — two DNA duplexes wrapped around each other's axis like the twisted strands of a telephone cord, branching into a tree-like structure of superhelical loops in larger molecules. This is directly visible in electron micrographs (a technique central to the discovery of supercoiling itself) and explains the compact, twisted shapes seen for natural bacterial plasmids under the electron microscope, in contrast to the open circles produced when a single strand nick releases the strain.

Step 4 — Removing or resetting the strain: topoisomerases. Because Lk can only change by breaking and resealing a strand, the cell relies on dedicated enzymes — topoisomerases — to manage supercoiling. Type I topoisomerases cut one strand, pass the other through the break, and reseal it, changing Lk in steps of one; type II topoisomerases (including bacterial DNA gyrase) cut both strands, pass an intact duplex segment through the double-strand break, and reseal, changing Lk in steps of two. Gyrase is unusual among topoisomerases in actively introducing negative supercoils using energy from ATP hydrolysis, rather than merely relaxing existing strain — this is why bacterial DNA maintains a steady-state σ of roughly −0.05 to −0.07 even at rest. This machinery is explored in depth under Topoisomerases and Linking Number; here the essential point is that supercoiling and topoisomerase action are two faces of one phenomenon — you cannot understand either without the other.

Step 5 — Biological payoff: negative supercoiling promotes opening. Underwinding (reduced Tw relative to relaxed DNA) directly lowers the energetic cost of locally separating the strands, because strand separation removes twist entirely from that stretch of duplex. This is why replication origins (like bacterial oriC) and many promoters are flanked by AT-rich sequences prone to melting, and why these elements open more readily, and transcribe or fire more efficiently, on negatively supercoiled templates than on relaxed ones — a fact demonstrated repeatedly in vitro using purified RNA polymerase and topoisomer ladders of defined σ. In eukaryotes, chromatin looping and nucleosome wrapping (Nucleosome Structure) constrain much of the negative supercoiling as restrained writhe around histone octamers rather than as free interwound plectonemes, but transcription still generates dynamic, unrestrained twin-domain supercoiling that must be resolved by topoisomerases I and II.

Vinograd and the discovery of superhelical DNA (1965). Jerome Vinograd and colleagues, studying polyoma virus DNA and bacterial plasmid DNA by analytical ultracentrifugation, noticed something odd: covalently closed circular DNA sedimented significantly faster than the same DNA after a single nick was introduced into one strand. A nicked circle relaxes to an open, floppy ring; the closed circle, they proposed, must be constrained into a more compact, coiled conformation — what Vinograd named "superhelical" DNA, by direct analogy to the over- or under-wound coiling of a twisted rubber band. Electron micrographs confirmed the prediction directly: closed circular DNA appeared as tightly interwound, branched structures, while nicked or linear DNA of identical sequence appeared as simple open circles or rods. This was the first direct physical evidence that topological constraint, not merely sequence or protein binding, could dictate a DNA molecule's global shape — and that closure of the strands was the essential trigger.

Depurination and dye-binding titrations establishing negative supercoiling (Vinograd, Bauer, and colleagues, late 1960s–1970s). To determine the sign of natural supercoiling (not just its existence), Vinograd's group titrated closed circular DNA with intercalating dyes such as ethidium bromide. Intercalation locally unwinds the double helix (each intercalated dye molecule reduces the local twist by a small, defined angle), and because Lk is fixed in a closed circle, unwinding by intercalation must be compensated by an increase in writhe. At low dye concentrations, intercalation into negatively supercoiled DNA removes negative writhe (making the molecule less compact and slower-sedimenting, up to a point), whereas the same treatment applied to already-relaxed or positively supercoiled DNA would only add compaction. The observed titration curves — sedimentation coefficient dropping to a minimum (fully relaxed) and then rising again as dye continued to unwind the helix past neutrality into positive supercoiling — showed unambiguously that native bacterial and viral plasmid DNA is negatively supercoiled in vivo, with σ typically in the range of −0.05 to −0.07. This dye-titration method remains a standard technique for measuring superhelical density today.

Wang's two-dimensional gel electrophoresis and the direct visualization of topoisomers (James C. Wang, 1970s). The most decisive and widely repeated evidence comes from gel electrophoresis. Because DNA molecules of identical length and sequence but differing Lk (topoisomers) differ in compaction — more negatively supercoiled molecules are more compact and migrate faster through an agarose gel under standard conditions — a population of topoisomers differing by single units of Lk resolves into a ladder of discrete bands. Wang and colleagues used this to show that:

• Purified topoisomerases change Lk in discrete integer steps (by one for type I enzymes, by two for type II), producing a stepwise ladder of bands rather than a smear, exactly as topological theory predicts.
• Adding chloroquine (another intercalator) to the gel running buffer, at varying concentrations, shifts the whole topoisomer ladder — allowing the sign of supercoiling to be read directly from which direction bands shift, a technique still used to distinguish positive from negative supercoiling on a gel.
• Bacterial DNA gyrase, when incubated with relaxed closed circular DNA and ATP, actively converts it into a strongly negatively supercoiled ladder — the first direct biochemical demonstration that an enzyme could impose negative supercoiling rather than merely relaxing existing strain, using energy from ATP hydrolysis. This experiment (Gellert, Mizuuchi, O'Dea, and Nash, 1976, who first purified gyrase) directly established that negative supercoiling in bacteria is not a passive byproduct but an actively maintained, energy-dependent cellular state.

Together, these three lines of evidence — sedimentation behavior of closed versus nicked circles, dye-titration determination of sign and magnitude, and gel-based resolution of discrete topoisomers — converted supercoiling from a hypothetical topological abstraction into a directly measurable, quantifiable physical property of real DNA molecules.

Example 1 — Counting turns on a plasmid. A circular plasmid is 4,200 base pairs long. Assume the reference (relaxed) helical repeat is 10.5 bp/turn, and the cell maintains it at a superhelical density of σ = −0.06, typical for E. coli. Step 1: Relaxed linking number. Lk₀ = 4200 / 10.5 = 400. Step 2: Linking number deficit. ΔLk = σ × Lk₀ = −0.06 × 400 = −24. Step 3: Actual linking number. Lk = Lk₀ + ΔLk = 400 − 24 = 376. Interpretation: the plasmid's two strands cross each other's topological boundary only 376 times instead of 400 — 24 fewer helical turns than a relaxed molecule of the same length. If we imagine this deficit expressed entirely as writhe, the plasmid would form roughly 24 negative superhelical turns (in practice, partitioned perhaps ~16 turns of writhe and ~8 turns of reduced twist, given the typical ~2:1 writhe:twist partition). This is exactly the kind of number you would read off a two-dimensional topoisomer gel: a population centered on Lk = 376, with a Gaussian-like spread of neighboring topoisomers (Lk = 374, 375, 376, 377, 378…) reflecting thermal fluctuation around the enzymatically set mean. Now suppose a gyrase inhibitor (such as the fluoroquinolone antibiotic ciprofloxacin, which traps gyrase mid-reaction on the DNA) is added, blocking further negative supercoiling. Topoisomerase I (which relaxes negative supercoils by removing them one at a time, ΔLk of +1 per catalytic cycle) continues acting unopposed. Over time, Lk relaxes back toward 400. If topo I acts on this plasmid until it has added back 10 units of linking number, the new Lk = 386, new ΔLk = −14, new σ = −14/400 = −0.035 — a measurably less negatively supercoiled, more relaxed plasmid, exactly the phenotype seen when gyrase is pharmacologically inhibited in bacterial cultures, and the biochemical basis for fluoroquinolone antibiotic action.

Example 2 — Reading a topoisomer gel and inferring enzyme mechanism. A researcher incubates relaxed closed circular plasmid DNA (a single band on a gel, Lk = Lk₀) with a partially purified enzyme extract and ATP, then runs samples at 0, 5, 15, and 30 minutes on an agarose gel (no intercalator in the gel, so more compact/supercoiled topoisomers migrate faster/further). Observed result: at 5 minutes, the single starting band has resolved into a discrete ladder of new bands migrating progressively faster, spaced at regular intervals; by 30 minutes, the ladder has shifted further, centered on a much faster-migrating set of topoisomers, and no intermediate smearing is seen at any timepoint — only discrete bands appear, shifting position, never blurring together. Reasoning: 1. Discrete bands, not a smear, immediately tell us Lk changes in whole-integer steps, consistent with a strand-passage mechanism rather than any continuous or fractional process. 2. The ladder shifts toward faster migration — meaning the DNA is becoming more compact — consistent with increasing negative supercoiling (more writhe, tighter plectonemes) rather than relaxation (which would shift bands toward the slower, less compact, nicked/relaxed position). 3. The requirement for ATP, combined with progressive negative supercoiling rather than mere relaxation, identifies the enzyme as DNA gyrase (a type II topoisomerase) rather than topoisomerase I: only gyrase actively pumps negative supercoils into DNA using ATP hydrolysis; topoisomerase I would relax existing supercoiling (shift bands the opposite direction) without any ATP requirement. 4. If instead each cycle changed Lk by exactly one unit per catalytic event (visible as single-band steps rather than double-band steps on a high-resolution gel), that would additionally indicate a type I-like single-strand-passage mechanism; steps of two are the signature of the double-strand-passage mechanism used by type II enzymes like gyrase — a distinction gel resolution can, in practice, actually discriminate. This kind of experiment — essentially Wang's and Gellert's original design — is still the standard assay used today to characterize any newly discovered or engineered topoisomerase-like activity.

James C. Wang built much of the biochemical and technical foundation of this field over several decades at Harvard, where his laboratory purified the first bacterial DNA topoisomerase (originally called the ω-protein, later recognized as topoisomerase I) from E. coli in 1971, and went on to develop the two-dimensional and chloroquine-gel electrophoretic methods that made linking number changes directly countable on a gel rather than only inferable from sedimentation behavior. His group's biochemical dissection of how topoisomerases change Lk in discrete integer steps, and how this connects mechanistically to strand passage through transient breaks, turned supercoiling from an abstract topological curiosity, established by Vinograd's physical measurements, into a fully mechanistic enzymology with defined active sites, reaction intermediates, and inhibitors. Wang's parallel theoretical work, alongside physicist Nicholas Cozzarelli and mathematically minded collaborators including Brock Fuller, helped rigorously connect the biology to the underlying topology (the Lk = Tw + Wr relation), ensuring that the field's central equation was not just a useful mnemonic but a mathematically exact statement (Călugăreanu–White–Fuller theorem) about closed curves in space.

Antibiotics that target gyrase. Fluoroquinolone antibiotics (ciprofloxacin, levofloxacin, and related drugs) work by trapping bacterial DNA gyrase (and topoisomerase IV) mid-reaction, converting the normally transient double-strand break intermediate into a lethal, permanent lesion. Because gyrase is essential and structurally distinct from eukaryotic type II topoisomerases, this exploits a genuine topological vulnerability unique to bacteria, making fluoroquinolones among the most widely used antibacterial drug classes in clinical medicine, though rising resistance (via gyrase mutations that reduce drug binding) is now a major public-health concern.

Anticancer topoisomerase poisons. Human topoisomerase I and II are essential for relieving supercoiling ahead of replication forks and RNA polymerases in rapidly dividing cells. Drugs such as camptothecin derivatives (topotecan, irinotecan; topo I poisons) and etoposide (a topo II poison) stabilize the cleavage complex intermediate, generating persistent DNA breaks that preferentially kill fast-dividing cancer cells whose replication machinery is constantly running into topological strain.

Genome architecture and gene regulation. Genome-wide chromatin conformation studies show that supercoiling domains correlate with topologically associating domains (TADs) and with transcriptional activity, and that negative supercoiling accumulates behind actively transcribing genes in a pattern directly predicted by the twin-domain model — linking classical topology to modern chromosome-scale genome organization.

Plasmid biotechnology and DNA vaccines. Manufacturing of plasmid DNA for gene therapy and DNA vaccines requires strict quality control of supercoiling state, since supercoiled (rather than nicked or linear) plasmid transfects cells and expresses genes far more efficiently — a direct practical payoff of understanding why negative supercoiling favors accessible, transcribable DNA.

Understanding thermophile and hyperthermophile biology. Some archaea and thermophilic bacteria maintain positively supercoiled DNA using reverse gyrase, a unique enzyme combining a helicase and topoisomerase domain, which is thought to stabilize the double helix against heat-induced denaturation at extreme temperatures — a striking natural inversion of the more familiar negative-supercoiling strategy, illustrating that supercoiling sign itself is an evolvable, adaptive trait.

Ask the tutor: How does nucleosome wrapping specifically constrain negative supercoiling, and how is that "stored" writhe counted separately from dynamic transcription-generated supercoiling?