Topoisomerases: Enzymatic Control of Supercoiling

A bacterium replicating its chromosome must unwind two intertwined strands. These strands are 4.6 million base pairs long. Unwinding happens roughly 1,300 times per second at the replication fork. This must occur without letting torsional stress stall the machinery or snap the DNA. It solves this problem not by brute force, but through a family of enzymes. These enzymes do something no other class of protein does. They cut the sugar-phosphate backbone, pass DNA through the break, and reseal it. All the while, they keep every atom of chemical bond energy in escrow. These are the topoisomerases. They are the molecular machines that make DNA Supercoiling and Topology a manageable variable. It becomes dynamic and enzymatically controlled, rather than a fatal topological trap. The problem they solve is genuinely topological, not merely geometric. Duplex DNA is a double helix. Because of this, any process requiring strand separation generates compensatory over- or under-winding. This includes replication, transcription, and recombination. The winding occurs ahead of and behind the moving enzyme. It is quantified by the linking number $Lk$ . A closed circular or torsionally constrained DNA molecule cannot relax this strain by simple rotation. The linking number is a topological invariant. It can only be changed by breaking and rejoining a strand. Topoisomerases are the only enzymes that break this invariant on purpose. They do so in a controlled, catalytic, and fully reversible way. They change $Lk$ in discrete steps of exactly $\pm 1$ (Type I) or $\pm 2$ (Type II) per catalytic cycle. The chemistry that makes this possible is a covalent enzyme–DNA intermediate. This is a transient phosphotyrosine linkage (or, in some Type II subfamilies, phosphotyrosine at the opposite polarity). The enzyme's own active-site tyrosine attacks a scissile phosphodiester bond. This breaks the DNA backbone while trapping the bond energy of the broken phosphodiester. That energy is captured in the new phosphotyrosine bond. This is the same general chemistry treated in The Chemistry of DNA Hydrolysis and Strand Cleavage. It is nucleophilic substitution at phosphorus. Here it is redirected so the enzyme itself, rather than water, is the incoming nucleophile. The "hydrolysis" is deferred rather than completed. The strand break is therefore never thermodynamically free. The enzyme sits at a covalent tollbooth. Religation restores exactly the phosphodiester bond that was broken. This article works through the mechanisms of Type IA, Type IB, and Type II topoisomerases. It covers full stereochemical and energetic detail. Topics include the transesterification chemistry at the active-site tyrosine. They also include the strand-passage or strand-rotation topological step. The article covers the ATP-coupled clamp mechanism of Type II enzymes. It also explores how each enzyme class is tuned by evolution and by drugs. These drugs include quinolones, etoposide, and camptothecin. Such drugs trap the covalent intermediate. They convert a housekeeping enzyme into a DNA-damaging poison.

Key takeaways

Topoisomerases relax supercoiling by transiently breaking one or both DNA strands via a phosphotyrosine covalent intermediate, never releasing the stored phosphodiester bond energy.
Type IA enzymes cut one strand, pass the other through the gap, and change $Lk$ by $\pm 1$ per cycle, using a 5′-phosphotyrosine linkage and requiring a single-stranded region.
Type IB enzymes (and topo V) nick one strand via a 3′-phosphotyrosine linkage and allow the intact strand to swivel through the break, relaxing supercoils processively without ATP.
Type II enzymes (gyrase, topo IV, human topo IIα/β) break both strands of one duplex (the "G-segment"), pass a second duplex (the "T-segment") through the gap, and change $Lk$ by $\pm 2$ , using ATP hydrolysis to drive a protein-clamp cycle.
DNA gyrase is unique in introducing negative supercoils using ATP hydrolysis to power a directional strand-passage step, essential for bacterial replication and transcription.
Clinically vital drugs — fluoroquinolones, camptothecin, etoposide — work by trapping the covalent enzyme–DNA complex, converting the topoisomerase into a DNA-damaging poison rather than simply inhibiting its catalysis.
Key idea: topoisomerases change DNA topology by transiently converting a phosphodiester bond into a phosphotyrosine bond, conserving the free energy of the backbone so that strand passage and religation are freely reversible.

The core insight

Every topoisomerase reaction is a transesterification, not a hydrolysis: an active-site tyrosine hydroxyl attacks the scissile phosphate in-line, displacing the 3′- (or 5′-) oxyanion of the adjacent nucleotide and forming a covalent phosphotyrosine–DNA intermediate that conserves the free energy of the broken phosphodiester bond, so that after strand passage the same tyrosine oxygen can be displaced by the liberated DNA oxyanion to reseal the backbone with no net energy input (Type I) or with the energy input reserved entirely for the mechanical clamp cycle (Type II, via ATP).

Where this lives today

DNA gyrase and topoisomerase IV are the direct molecular targets of the fluoroquinolone class of antibiotics (ciprofloxacin, levofloxacin, moxifloxacin), among the most widely prescribed antibacterials in clinical use, and gyrase's bacterial-specific structure (absent from the human genome, which has no gyrase-like negative-supercoiling activity) is precisely why these drugs achieve high selectivity for bacterial over human cells. Newer-generation fluoroquinolones and entirely new chemical classes are actively being developed specifically to counter quinolone-resistance mutations that arise in the drug-binding pocket at the gyrase–DNA interface, a direct pharmacological consequence of the cleavage-complex trapping mechanism worked out above. In oncology, human topoisomerase I is the target of camptothecin and its clinically used derivatives irinotecan and topotecan, while human topoisomerase IIα is targeted by etoposide, doxorubicin, and related anthracyclines — all of which work by the same conceptual mechanism as fluoroquinolones (stabilizing the covalent cleavage complex) despite acting on structurally unrelated enzyme families, illustrating a striking case of convergent pharmacological strategy across the Type IB and Type II enzyme classes. Because these drugs preferentially kill rapidly dividing cells (which have high topoisomerase turnover to support replication), they remain front-line chemotherapeutics for colorectal, lung, ovarian, and other cancers, though their genotoxicity also creates a well-documented risk of secondary, treatment-related leukemias arising from topoisomerase II-mediated chromosomal translocations. Beyond pharmacology, topoisomerases are essential tools in molecular biology and biotechnology: Type IB topoisomerase (from vaccinia virus) is the basis of "TOPO cloning" kits, in which the enzyme is supplied pre-charged with a covalent 3′-phosphotyrosine linkage on a linearized vector, allowing extremely rapid, ligase-free insertion of PCR products by simple religation chemistry — a direct laboratory exploitation of the same reversible transesterification mechanism described above. Understanding of Type II strand-passage mechanisms has also informed the design of decatenation assays essential to studying chromosome segregation, and continues to guide structural biology efforts (cryo-EM, single-molecule magnetic tweezers) aimed at resolving the transient, ATP-driven conformational states of the two-gate mechanism in real time.

Ask the tutor: How does the twin-supercoiled-domain model quantitatively determine how many topoisomerase molecules must act per second to keep pace with a replication fork moving at 1,000 bp/s?

The physical picture — mechanism, topology, and energetics

Consider a plasmid with a fixed linking number $Lk$ , partitioned as

Lk = Tw + Wr

where $Tw$ is the helical twist and $Wr$ is the writhe — the topic developed fully in the treatment of supercoiling topology. A negatively supercoiled plasmid ( $Lk < Lk_0$ , where $Lk_0$ is the relaxed linking number) has excess negative writhe that favors strand opening, which is exactly why bacterial chromosomes are held slightly underwound: it assists origin melting and RNA polymerase's local unwinding. But this convenient underwinding must be actively maintained and locally adjusted, and unrestrained positive supercoiling ahead of a translocating polymerase or replisome would otherwise stall the machinery entirely — this is the twin-supercoiled-domain problem. Topoisomerases are the enzymes that both impose and relieve this stress catalytically. The chemistry that allows this is remarkably restrained. In all topoisomerases, an active-site tyrosine's phenolic oxygen performs an in-line nucleophilic attack on the phosphorus of a specific phosphodiester bond in the DNA backbone. This is a direct analogue of the associative (in-line, $S_N2$ -like) mechanism at phosphorus discussed for nuclease and RNase chemistry, but with one crucial difference: instead of water or the 2′-OH acting as nucleophile (as in hydrolytic or RNase-type cleavage), the nucleophile is the enzyme's own tyrosyl oxygen, and the leaving group is a DNA oxyanion rather than a small molecule. Because the tyrosine remains covalently tethered to the phosphate, no bond energy is dissipated to solvent — the "cut" phosphodiester bond has simply been relocated onto the protein, ready for the reverse reaction. Two broad polarities exist. Type IA and Type II enzymes attack such that the tyrosine ends up 5′-phosphotyrosine linked, with a free 3′-OH left on the DNA. Type IB enzymes (eukaryotic topoisomerase I, and the distantly related Type IC enzyme topoisomerase V) attack to leave a 3′-phosphotyrosine linkage, with a free 5′-OH. This polarity difference dictates the entire downstream mechanism: 5′-phosphotyrosine enzymes must physically pass a second DNA strand or duplex through the transient gap (strand-passage mechanism), whereas 3′-phosphotyrosine Type IB enzymes can instead let the intact strand rotate freely around the phosphodiester backbone of the other, uncut strand (a controlled-rotation, or "swivelase," mechanism) — no protein-mediated strand passage is even needed.

Figure

Two polarities of the phosphotyrosine intermediate: Type IA/II (5′-linkage, strand-passage) versus Type IB (3′-linkage, controlled rotation)

The energetics of the covalent intermediate deserve emphasis because they are the entire reason this chemistry is biologically viable. A phosphodiester bond hydrolyzed by water is essentially irreversible under cellular conditions (large negative $\Delta G$ , and the two product ends — a 3′-OH and a 5′-phosphate, or vice versa — diffuse apart). A phosphodiester bond transesterified onto a tyrosine, by contrast, is isoenergetic with the original phosphodiester bond to a first approximation (both are phosphate diester linkages, differing mainly in the identity of the leaving/attacking oxygen), so the reverse reaction — nucleophilic attack by the liberated DNA 3′-oxyanion (or 5′-oxyanion) on the phosphotyrosine phosphorus — is essentially barrier-matched and thermodynamically neutral. This is why topoisomerases do not need ATP to break and reseal a single strand: the "break" step and the "reseal" step are simply the forward and reverse of the same transesterification equilibrium, with strand passage or rotation occurring in between while the enzyme holds the broken ends in register.

Figure

Energy diagram: transesterification traps bond energy in phosphotyrosine, unlike hydrolysis

The core ideas — linking number arithmetic and the strand-passage law

The defining quantitative feature of each topoisomerase class is the integer by which it changes the linking number per catalytic cycle, and the sign convention (which strands cross which way) that determines whether the enzyme relaxes or introduces supercoils.

\Delta Lk_{\text{Type IA}} = \pm 1 \quad \text{per strand-passage event}

\Delta Lk_{\text{Type IB}} = \text{any integer, achieved by multi-step rotation before religation}

\Delta Lk_{\text{Type II}} = \pm 2 \quad \text{per strand-passage event, driven by ATP hydrolysis}

where:

$Lk$ is the linking number of the two strands of a closed or topologically constrained DNA molecule,
a Type IA or Type II "event" is one full catalytic cycle: cleavage of one strand (IA) or both strands of a duplex (II), passage of a second strand or duplex through the gap, and religation,
the sign of $\Delta Lk$ for Type II gyrase is negative (introduces negative supercoils) uniquely among topoisomerases, powered directly by the free energy of ATP hydrolysis; all other Type II enzymes (topo IV, eukaryotic topo II) only relax supercoils (or decatenate/unknot), moving $Lk$ toward $Lk_0$ regardless of sign.

Reading: these are not empirical curve-fits; they follow directly from the geometry of strand passage. Passing one single strand through a break in one strand changes the number of times the two strands cross each other, when the DNA is later resealed and viewed as a closed ribbon, by exactly one unit — hence $\pm 1$ for Type IA. Passing an entire duplex (two strands, both continuous) through a double-strand break changes the crossing number by exactly two units — hence $\pm 2$ for Type II. The magic of Type IB's "any integer" comes from the fact that, once one strand is nicked, the intact complementary strand can rotate freely (a swivel, not a discrete passage event), so many turns of relaxation can occur before religation recaptures whatever $Lk$ happens to be instantaneously present. The free energy stored in supercoiling itself is well described, for a closed circular DNA of $N$ bp with excess linking number $\Delta Lk = Lk - Lk_0$ , by the harmonic torsional-energy relation

\Delta G_{\text{sc}} = \frac{K}{2}\left(\frac{\Delta Lk}{Lk_0}\right)^2 \cdot N \cdot RT

where $K$ is an empirical torsional stiffness constant (typically $K \approx 10$ in units consistent with this normalization for physiological superhelical densities), $N$ is the number of base pairs, and $\sigma = \Delta Lk / Lk_0$ is the superhelical density (physiologically $\sigma \approx -0.05$ to $-0.07$ in bacteria). This quadratic dependence is the reason a cell cannot simply tolerate unregulated positive supercoiling ahead of a fork: the energetic penalty grows as the square of the accumulated twist, and only continuous topoisomerase activity keeps $\sigma$ within the narrow physiological window.

Derivation AType IA mechanism: single-strand passage and the ±1 rule

Goal. Show mechanistically, step by step, why a Type IA topoisomerase changes $Lk$ by exactly one unit per cycle, using the concrete example of E. coli topoisomerase I acting on negatively supercoiled DNA.

Step 1 — Substrate recognition. Type IA enzymes require a region of single-stranded DNA, which is why they act preferentially on negatively supercoiled substrates (negative supercoiling favors local strand separation) or on ssDNA generated during replication and recombination. The enzyme binds at the ss/dsDNA junction, gripping one strand (call it strand G, for "gate") while positioning the other transiently displaced strand (strand T, for "transfer") nearby.

Step 2 — Nucleophilic attack and cleavage. The active-site tyrosine's phenolic oxygen performs an in-line $S_N2$ -like attack on the phosphorus of a specific phosphodiester bond in strand G. The attack proceeds through a trigonal bipyramidal transition state at phosphorus, with the tyrosyl oxygen entering apically opposite the 5′-oxygen leaving group. This displaces the 5′-oxygen of the downstream nucleotide, breaking the strand and forming a covalent 5′-phosphotyrosine linkage: the enzyme is now attached to the 5′-phosphate of the cut site, and the 3′-OH of the upstream nucleotide is left free. Divalent Mg²⁺ in the active site stabilizes the developing negative charge on the pentacovalent transition state, exactly as in two-metal-ion-assisted phosphoryl transfer.

Step 3 — Gate opening and strand passage. With strand G now nicked but covalently held by the enzyme (so the ends cannot diffuse apart), a conformational change opens a physical gate in the protein. Strand T — a single strand from elsewhere in the same or another DNA molecule — is captured and threaded through the transient gap in strand G. This is an enzyme-mediated, ATP-independent conformational "breathing" that requires no external energy input because no covalent bonds beyond the reversible phosphotyrosine linkage are altered during passage itself.

Step 4 — Religation. Once strand T has passed through, the gate recloses around strand G. The free 3′-OH of strand G (generated in Step 2) now performs the reverse nucleophilic attack: it attacks the phosphotyrosine phosphorus in-line, displacing the tyrosine and reforming the original phosphodiester bond — chemically identical to the one broken in Step 2, but now with strand T looped through the DNA duplex on the other side of the (resealed) break.

Step 5 — Topological bookkeeping. Before the cycle, strands G and T were unlinked with respect to this local crossing (or linked $n$ times). After one strand of G has been broken, T threaded through, and G resealed, the local crossing number between the two duplex axes has changed by exactly one full crossing — geometrically, threading one strand through a single-strand gap can only add or remove one intersection when the gap recloses, because the gap admits only one strand's width at a time. Since $Lk$ is (by definition) the signed sum of crossings in a planar projection divided by two, and Type IA passage changes the strand-strand crossing count by one full unit directly (not by two, as in double-strand passage), the linking number changes by exactly

\Delta Lk = \pm 1

with the sign fixed by the chirality of the pre-existing supercoil (negative supercoils are relaxed toward zero, i.e., $Lk$ increases by 1 toward $Lk_0$ , for E. coli topo I and topo III).

Step 6 — Energy accounting. No ATP was consumed. The phosphodiester bond broken in Step 2 is chemically identical to the phosphodiester bond reformed in Step 4; the only net change in the system is the topological state of the DNA (one fewer negative supercoil) and the release of the corresponding torsional strain energy $\Delta G_{\text{sc}}$ computed from the quadratic supercoiling-energy relation. The enzyme is a topological catalyst: it lowers the kinetic barrier to changing $Lk$ without altering the equilibrium of the phosphodiester bond itself.

Go deeper · Derivation B · 3 minType II mechanism: double-strand passage, the ATP-gated clamp, and the ±2 rule

Derivation BType II mechanism: double-strand passage, the ATP-gated clamp, and the ±2 rule

Goal. Derive why Type II topoisomerases change $Lk$ in units of exactly two, and show how ATP hydrolysis couples to directional strand passage in DNA gyrase, using the two-gate ("G-segment/T-segment") model common to gyrase, topo IV, and eukaryotic topo II.

Step 1 — Architecture. A Type II topoisomerase functions as a homodimer (or heterotetramer, as in gyrase's GyrA₂GyrB₂) with three sequential protein gates along the DNA path: an N-gate (ATPase domain, binds ATP and captures the T-segment), a DNA-gate (the catalytic core, binds and cleaves the G-segment), and a C-gate (exit gate for the T-segment). The enzyme first binds the "gate" duplex segment, G, bending it sharply (~150° in gyrase) across the dimer interface.

Step 2 — Double-strand cleavage. Two active-site tyrosines, one from each protomer, simultaneously attack the phosphates of both strands of the G-segment, staggered by four base pairs (producing a 4-nucleotide 5′ overhang upon cleavage, as is characteristic of Type II enzymes). Each attack is the same in-line transesterification described for Type IA: tyrosyl oxygen attacks phosphorus, Mg²⁺-stabilized transition state, 5′-phosphotyrosine linkage forms on each strand, 3′-OH ends are liberated on each side of the now fully double-strand-broken G-segment. Crucially, the four-base 5′ overhangs mean the two covalent linkages are on opposite strands displaced by a short single-stranded stagger, not blunt — this staggered break is what allows the DNA gate to reseal in exact register after strand passage.

Step 3 — ATP binding and N-gate closure. A second DNA duplex, the transfer segment T, is captured by the ATPase (GyrB/topo II ATPase) domains. Binding of two ATP molecules (one per protomer) to the GHKL-fold ATPase domains induces dimerization of the N-gate, trapping the T-segment inside the enzyme, above the now-open DNA-gate.

Step 4 — Strand passage through the double-strand break. The trapped T-segment is translocated through the transient double-strand break in the G-segment. This is a large, ATP-hydrolysis-coupled conformational wave: ATP hydrolysis (not binding alone) drives the T-segment through the gate and toward the C-gate, and inorganic phosphate release resets the ATPase domains for the next cycle. This is the step that consumes chemical energy; unlike Type IA/IB religation, resealing the double-strand break itself is thermoneutral (again a transesterification reversal), but the directional, processive movement of an entire duplex through another requires the mechanical work supplied by ATP hydrolysis, roughly 2 ATP hydrolyzed per full catalytic cycle.

Step 5 — Religation. With the T-segment now on the far side of the G-segment, the two free 3′-OH ends of the G-segment attack their respective phosphotyrosine linkages in-line, exactly reversing Step 2 chemistry on both strands, resealing the G-segment as an intact duplex — but now with the T-segment topologically on the opposite side.

Step 6 — Exit and topological bookkeeping. The T-segment exits through the C-gate, and the enzyme releases ADP + Pi, resetting to the initial state. Passing an entire duplex (both strands, rigidly held together) through a double-strand break changes the crossing number between the two DNA duplex axes by exactly two full crossings (one duplex has two strands, and complete passage of a rigid rod-like duplex through a gap changes the projected linking number contribution by 2, in contrast to the single-strand-width passage of Type IA). Hence

\Delta Lk = \pm 2 \quad \text{per catalytic cycle}

Step 7 — Directionality in gyrase. For most Type II enzymes (topo IV, human topo IIα/β) the sign of $\Delta Lk$ is not fixed by the enzyme itself but by thermodynamics: strand passage occurs preferentially in whichever direction reduces $|\Delta Lk|$ (i.e., relaxes supercoiling, whether positive or negative), because the bent G-segment geometry in the absence of specialized subunits has no directional bias. DNA gyrase is exceptional: its unique GyrA "C-terminal domain" wraps DNA around the enzyme in a defined, positive chiral sense (a positive DNA wrap), which means the T-segment captured for passage is always presented in the orientation that removes a positive crossing (equivalently, adds a negative supercoil) when passed through the gate. This directional wrapping, combined with ATP hydrolysis supplying the free energy to drive strand passage against the resulting increase in torsional stored energy, is precisely why gyrase alone among topoisomerases can actively introduce negative supercoils, at the expense of ATP:

\Delta G_{\text{net}} = \Delta G_{\text{ATP hydrolysis}} + \Delta G_{\text{sc}}(\Delta Lk = -2) < 0

with $\Delta G_{\text{ATP hydrolysis}} \approx -30.5\ \text{kJ/mol}$ per ATP under cellular conditions supplying more than enough free energy to overcome the positive $\Delta G_{\text{sc}}$ penalty of adding two more negative supercoils to an already negatively supercoiled plasmid.

Worked examples

Example 1 — Foundational: relaxing a single Type IA event on a small plasmid. Take a circular plasmid of $N = 4{,}200$ bp with $Lk_0 = N/10.5 = 400.0$ (using 10.5 bp per helical turn as the relaxed B-form value) and an initial linking number $Lk = 386$ (negatively supercoiled). The superhelical density is

\sigma = \frac{Lk - Lk_0}{Lk_0} = \frac{386 - 400}{400} = -0.035

A single Type IA relaxation event changes $Lk$ by exactly $+1$ (toward $Lk_0$ , since the plasmid is negatively supercoiled and E. coli topo I only relaxes negative supercoils):

Lk' = 386 + 1 = 387, \qquad \sigma' = \frac{387-400}{400} = -0.0325

Using the harmonic supercoiling free energy with $K \approx 10$ (dimensionless empirical stiffness constant commonly used for physiological superhelical densities) and $RT = 2.48\ \text{kJ/mol}$ at 298 K, the energy released by this single step is

\Delta G_{\text{sc}}(\sigma) = \frac{K}{2}\sigma^2 \cdot N \cdot RT

\Delta G_{\text{sc}}(-0.035) = \frac{10}{2}(0.035)^2 (4200)(2.48\ \text{kJ/mol}) = 5(0.001225)(4200)(2.48)

= 5 \times 0.001225 \times 4200 \times 2.48 = 63.8\ \text{kJ/mol}

\Delta G_{\text{sc}}(-0.0325) = 5(0.0325)^2(4200)(2.48) = 5(0.001056)(4200)(2.48) = 55.0\ \text{kJ/mol}

The single catalytic step releases $63.8 - 55.0 = 8.8\ \text{kJ/mol}$ of stored torsional strain energy — with no ATP consumed, entirely from relaxing the DNA itself, exactly as the mechanism predicts: the phosphodiester bond broken and remade is energy-neutral, and the released free energy comes purely from the topology of the DNA, not the chemistry of the enzyme.

Limiting check: if $\sigma = 0$ (fully relaxed, $Lk = Lk_0$ ), $\Delta G_{\text{sc}} = 0$ — correctly predicting that a relaxed plasmid provides no thermodynamic driving force for further topoisomerase action, consistent with Type IA and Type IB enzymes acting until $\sigma \to 0$ and then stalling (an equilibrium point, not a forced endpoint).

Example 2 — Applied: gyrase-driven negative supercoiling against ATP cost. E. coli DNA gyrase actively introduces negative supercoils into relaxed DNA. Starting from a relaxed 4,200 bp plasmid ( $Lk_0 = 400$ , $\sigma = 0$ ), each gyrase catalytic cycle changes $Lk$ by $-2$ and consumes 2 ATP (one per GyrB ATPase subunit, per the two-gate mechanism). After $n$ cycles,

Lk_n = 400 - 2n, \qquad \sigma_n = \frac{-2n}{400} = -0.005\,n

Suppose gyrase acts for $n = 7$ cycles (an asymmetric, non-round choice to track precisely): $Lk_7 = 400 - 14 = 386$ , and $\sigma_7 = -0.035$ — reproducing the physiological superhelical density used in Example 1, a sanity check that these numbers are self-consistent with real bacterial values. The torsional energy stored after 7 cycles, from the same formula, is (as computed above) $\Delta G_{\text{sc}} = 63.8\ \text{kJ/mol}$ per mole of plasmid. The ATP cost is

\Delta G_{\text{ATP}} = 7\ \text{cycles} \times 2\ \text{ATP/cycle} \times (-30.5\ \text{kJ/mol per ATP}) = -427\ \text{kJ/mol}

Comparing magnitudes: the cell spends 427 kJ/mol of ATP free energy to store only 63.8 kJ/mol as recoverable torsional strain — an efficiency of roughly 15%, with the rest dissipated as heat and consumed in the large conformational (clamp-opening, clamp-closing, translocation) work of the strand-passage cycle itself. This is expected: gyrase is not a free-energy-conserving machine for storing supercoiling energy per se; it is a motor that does mechanical and chemical work to move the DNA into a metastable, biologically useful topological state (a state the cell then exploits kinetically, e.g., to promote local strand opening at origins and promoters), not to bank the ATP energy as retrievable torsional energy.

Limiting check: as $n \to Lk_0/2 = 200$ , $\sigma \to -1$ , which is far beyond any physiologically tolerated superhelical density (real $\sigma$ saturates around $-0.05$ to $-0.09$ because the DNA denatures, plectonemes form, or gyrase itself reaches a supercoiling/relaxation steady state with topoisomerase I and IV activity) — correctly signaling that this simple linear model must break down and be superseded by a proper equilibrium between gyrase's supercoiling activity and relaxing topoisomerases (topo I, topo IV) well before such extreme values are reached, exactly as observed in cells.

Example 3 — Exam/competition level: quinolone poisoning and the cleavage-complex equilibrium. Fluoroquinolone antibiotics (e.g., ciprofloxacin) act by trapping the covalent gyrase(or topo IV)–DNA cleavage complex — stabilizing the double-strand-broken, phosphotyrosine-linked intermediate rather than allowing religation, converting the enzyme into a cellular poison that snaps chromosomes when collided with by replication forks. Model the cleavage complex as a two-state equilibrium between the religated (closed, "R") and cleaved (open, covalent, "C") enzyme–DNA states, with equilibrium constant

K_{eq} = \frac{[C]}{[R]} = \exp\left(-\frac{\Delta G_{C \to R}}{RT}\right)

In the absence of drug, suppose the cleaved state is disfavored, with $\Delta G_{C \to R} = -12.0\ \text{kJ/mol}$ (religation strongly favored), giving

K_{eq}^{\text{no drug}} = \exp\left(\frac{12.0}{2.48}\right) = \exp(4.84) = 126

so only a fraction $f_C = 1/(1+K_{eq}) = 1/127 = 0.0079$ , i.e. 0.79%, of enzyme–DNA complexes are in the cleaved state at any instant — consistent with the normally low steady-state level of DNA breaks gyrase produces during ordinary catalysis. Ciprofloxacin intercalates at the cleavage site and hydrogen-bonds to the enzyme and the exposed DNA bases, selectively stabilizing state C by binding much more tightly to the cleaved complex than to the religated one. Suppose the drug binds C with $K_d^C = 2\ \mu M$ and R with $K_d^R = 200\ \mu M$ (a 100-fold selectivity, a realistic and asymmetric — not round-number-coincidental — ratio for fluoroquinolones), at a free drug concentration $[D] = 20\ \mu M$ (a typical achievable intracellular concentration). The drug binding shifts the effective equilibrium according to

K_{eq}^{\text{+drug}} = K_{eq}^{\text{no drug}} \times \frac{1 + [D]/K_d^C}{1 + [D]/K_d^R}

= 126 \times \frac{1 + 20/2}{1 + 20/200} = 126 \times \frac{11}{1.1} = 126 \times 10 = 1260

so the fraction cleaved rises to

f_C^{\text{+drug}} = \frac{1260}{1261} = 0.99921, \text{ i.e., essentially all complexes are now trapped in the cleaved state.}

This tenfold shift in equilibrium (from $K_{eq} = 126$ to $1260$ ) converts a rare, fleeting intermediate (0.79% cleaved) into the dominant, essentially permanent state (99.9% cleaved) — a small, chemically modest difference in relative binding affinity ( $K_d^C \ll K_d^R$ ) producing an enormous biological effect, because these trapped double-strand breaks, when converted to permanent breaks by a colliding replication fork, are lethal to the bacterium. This is exactly the mechanism by which fluoroquinolones function as bactericidal drugs, and it is the same conceptual trap (stabilizing the covalent cleavage complex rather than blocking substrate binding) used by human topoisomerase-targeting anticancer drugs like camptothecin (traps Type IB topo I) and etoposide (traps Type II topo IIα/β).

Limiting check: as $[D] \to 0$ , the ratio $(1+[D]/K_d^C)/(1+[D]/K_d^R) \to 1$ , recovering $K_{eq}^{\text{+drug}} \to K_{eq}^{\text{no drug}}$ exactly — correctly predicting that in the absence of drug the enzyme behaves as an ordinary, low-cleavage-fraction catalyst, and that the poisoning effect scales with drug concentration and selectivity, not with some intrinsic toxicity of the enzyme itself.

Limiting cases, scope & common traps

Regime / Enzyme class	Behavior	Common trap
Type IA (bacterial topo I, topo III; also topo V)	Requires ssDNA; $\Delta Lk = \pm 1$ /cycle; 5′-phosphotyrosine; no ATP	Assuming all Type I enzymes are ATP-independent — true for IA and IB, but eukaryotic topo III (a IA) still needs ssDNA regions, unlike IB
Type IB (eukaryotic topo I; topo V is a distinct IC fold with similar chemistry)	Controlled rotation/swivel; 3′-phosphotyrosine; changes $Lk$ by variable integers per binding event; no ATP	Confusing "no strand passage" with "no strand break" — Type IB still nicks DNA covalently, it just doesn't need a protein gate for the intact strand
Type II (gyrase, topo IV, human topo IIα/β)	Double-strand break; $\Delta Lk = \pm 2$ /cycle; ATP-dependent; 5′-phosphotyrosine	Thinking ATP powers the bond chemistry — it powers the mechanical clamp/translocation step, not the (thermoneutral) transesterification itself
DNA gyrase specifically	Uniquely introduces negative supercoils using ATP + directional DNA wrap	Assuming all Type II enzymes can supercoil — topo IV and human topo II can only relax/decatenate, never actively supercoil
$\sigma \to 0$ (relaxed DNA)	$\Delta G_{\text{sc}} \to 0$ ; Type I/topo IV activity has no thermodynamic driving force	Forgetting that cells maintain steady-state $\sigma \neq 0$ via a dynamic balance of gyrase (adding negative supercoils) versus topo I/IV (relaxing), not a static endpoint
Drug-trapped cleavage complex	Enzyme becomes a DNA-damaging poison, not simply "inhibited"	Calling quinolones/camptothecin "topoisomerase inhibitors" without qualification — mechanistically they are catalytic-cycle poisons that stabilize a normally transient covalent intermediate

The chemist behind it

James C. Wang discovered DNA topoisomerase I in E. coli in 1971, identifying an activity he called the "ω protein" that could relax supercoiled DNA without any added energy source — a result that, at the time, was almost paradoxical, since biologists assumed any enzyme capable of manipulating a topologically constrained polymer as complex as supercoiled DNA would require an external energy input. Wang's insight — confirmed over the following decade through painstaking biochemical fractionation, linking-number gel assays, and eventually crystallography — was that the enzyme achieves its effect not by hydrolyzing the phosphodiester bond outright but by transiently transesterifying it onto a specific active-site tyrosine, an idea that reframed the entire field's understanding of how nucleic acid backbone chemistry could be harnessed reversibly. Wang went on to characterize eukaryotic topoisomerases I and II, established the two broad structural/mechanistic superfamilies (Type I and Type II) that remain the organizing framework for the field, and their essential roles in replication, transcription, and chromosome segregation. His laboratory's structural and mechanistic work, alongside that of Nicholas Cozzarelli (who elucidated DNA gyrase's ATP-dependent supercoiling mechanism and its role in bacterial chromosome topology) and Martin Gellert (co-discoverer of gyrase itself, in 1976), established the entire conceptual vocabulary — cleavage complex, strand passage, ATP-gated clamp — that is now standard in molecular biology and pharmacology.

Origins · 1 minJames C. Wang and how this was discovered

James C. Wang

1936–2011 · Discoverer of DNA topoisomerase I; established the transesterification mechanism and Type I/Type II classification framework

Wang trained as a physical chemist before turning to nucleic acid enzymology, and his approach to topoisomerases retained a physicist's insistence on quantitative, topologically rigorous bookkeeping — using linking number gel-shift assays to count, literally one unit at a time, how many times an enzyme changed $Lk$ per binding event. This quantitative rigor is what first revealed the sharp mechanistic distinction between $\Delta Lk = \pm 1$ (Type I) and $\Delta Lk = \pm 2$ (Type II) enzymes, decades before either class was crystallized.

History

- 1971 — James Wang discovers E. coli topoisomerase I (the ω protein), the first topoisomerase identified.

- 1976 — Martin Gellert and colleagues discover DNA gyrase, the first Type II topoisomerase and the only one that actively introduces negative supercoils using ATP.

- 1979–1980s — Nicholas Cozzarelli's laboratory dissects gyrase's ATP-coupled strand-passage mechanism and its ±2 linking-number change per cycle.

- 1984 — Human topoisomerase I identified as the cellular target of camptothecin, establishing the drug-trapping ("poison") paradigm for topoisomerase-targeted chemotherapy.

- 1996 — Crystal structure of a 67 kDa fragment of human topoisomerase I bound to DNA (Stewart, Champoux, Hol) reveals the enzyme clamped around the duplex, rationalizing the controlled-rotation mechanism structurally.

Why don't topoisomerases need ATP to break and reseal DNA in Type I enzymes?

Because the "break" is not a hydrolysis but a transesterification: the active-site tyrosine's oxygen displaces the DNA's own leaving-group oxygen, forming a phosphotyrosine bond that is chemically similar in energy to the original phosphodiester bond. The bond energy is conserved on the enzyme rather than dissipated to solvent, so the reverse reaction (the liberated DNA oxyanion attacking the phosphotyrosine phosphorus) is essentially barrier-matched and thermoneutral. No external energy input is required for either bond-breaking or bond-remaking; only the topological rearrangement (strand rotation or single-strand passage) occurs in between, and that step also requires no chemical energy, only a conformational gating motion of the protein.

Why do Type II topoisomerases need ATP if the chemistry itself is energy-neutral?

ATP hydrolysis in Type II enzymes does not power the phosphodiester cleavage/religation chemistry, which remains a thermoneutral transesterification just as in Type I enzymes. Instead, ATP binding and hydrolysis at the GHKL-fold ATPase (N-gate) domains drive the large mechanical conformational cycle: capturing the transfer (T) segment, closing the N-gate around it, and translocating the entire duplex through the transient double-strand break in the gate (G) segment. This is genuine mechanical work — moving a rigid, several-nanometer duplex through a protein-gated pore against viscous drag and DNA bending resistance — and that work is what ATP hydrolysis (roughly 2 ATP per cycle) actually pays for.

What is the biological reason bacteria need negative supercoiling maintained by gyrase?

Negative supercoiling stores free energy that assists local strand separation, which is required at replication origins (oriC) for the initial melting step and at many promoters for open-complex formation by RNA polymerase. Physiological superhelical density (σ ≈ −0.05 to −0.07) is maintained as a dynamic steady state between DNA gyrase (which actively introduces negative supercoils using ATP) and relaxing topoisomerases such as topoisomerase I (which removes excess negative supercoiling). This balance is homeostatically regulated; gyrase and topo I expression are themselves supercoiling-sensitive, forming a feedback loop that keeps σ within a narrow physiological window rather than drifting toward either extreme.

How does the twin-supercoiled-domain model relate to topoisomerase function during transcription and replication?

As RNA polymerase or a replicative helicase translocates along DNA, it must locally unwind the duplex, which mechanically generates positive supercoiling ahead of the moving complex and, if the molecule is not free to rotate (as is typical in vivo, where DNA is anchored to other proteins or itself), a compensatory build-up of negative supercoiling behind it. This is the twin-supercoiled-domain model, first proposed by Liu and Wang. Left unresolved, the ahead-of-the-fork positive supercoiling would rapidly stall the polymerase or replisome, since positive supercoiling actively opposes strand separation. Topoisomerases (topo IV and gyrase in bacteria; topo I and topo II in eukaryotes) relieve this stress in real time, working just ahead of and behind the moving complex.

Why does Type IA topoisomerase require single-stranded DNA as a substrate, unlike Type IB?

Type IA enzymes evolved from a structural fold related to certain ssDNA-binding nucleases, and their catalytic mechanism requires threading an entire second single strand (not merely allowing an intact strand to swivel) through the transient gap created in the first strand. This obligate strand-passage step needs a genuinely single-stranded segment of DNA available to serve as the transfer strand, which is why Type IA enzymes act preferentially on negatively supercoiled DNA (which favors local strand melting), replication intermediates, or recombination substrates with exposed ssDNA. Type IB enzymes, by contrast, never need a second strand to thread through anything — the intact strand simply rotates about the phosphodiester backbone of the uncut strand — so no pre-existing ssDNA region is required.

Are the phosphotyrosine linkages in Type IA/Type II enzymes (5′) and Type IB enzymes (3′) chemically different reactions?

Mechanistically they are the same reaction type — an in-line nucleophilic substitution at phosphorus by the tyrosine phenolic oxygen — but they target opposite ends of the scissile phosphodiester bond. In Type IA/II enzymes, the tyrosine oxygen displaces the 5′-oxygen leaving group, leaving a free 3′-OH on the DNA and a 5′-phosphotyrosine linkage on the enzyme. In Type IB enzymes, the tyrosine oxygen instead displaces the 3′-oxygen leaving group, leaving a free 5′-OH on the DNA and a 3′-phosphotyrosine linkage on the enzyme. This single difference in regiochemistry has large downstream mechanistic consequences: it is precisely why Type IB enzymes can support free rotation (the intact strand's phosphate backbone at the nick is unencumbered on the 3′ side) while Type IA/II enzymes require protein-mediated strand passage.

What structural feature distinguishes gyrase from other bacterial and human Type II topoisomerases?

DNA gyrase possesses a unique GyrA subunit C-terminal domain (CTD) that wraps the bound DNA around the enzyme in a defined, positive superhelical sense before strand passage occurs. This directional wrapping means that whichever segment of DNA becomes the "transfer" (T) segment is topologically presented to the gate in a fixed chirality, so that every strand-passage event removes a positive DNA crossing (equivalently, adds a negative supercoil), regardless of the DNA's initial topological state. Topoisomerase IV and human topoisomerase IIα/β lack this directional CTD wrap and instead act essentially isotropically, driving Lk toward equilibrium (relaxation, decatenation, unknotting) rather than actively supercoiling in a fixed direction.