Synthesis · Complexes · Gradient · Yield · assembled from standard bioenergetics and structural work, not a reading of one paper
This is a synthesis, not a close reading: it assembles the standard account of oxidative phosphorylation into a model you can run and count. One turnover on screen is one O2 molecule, which is four electrons and two NADH. Every particle is scheduled from the published stoichiometry, so the totals come out right if you pause and tally them.
The chain does not make ATP. It builds a gradient, and ATP synthase spends it. The famous yields of 2.5 and 1.5 are a division, protons pumped divided by protons per ATP, and the last panel does that division in front of you rather than asserting it.
What is real here and what is not: the complexes, their proton stoichiometry, the midpoint potentials, and the c-ring count are measured quantities with sources below. The rates are not. Flux, leak, and the speed the gradient settles at come from coefficients chosen so the simulation moves at a watchable pace, and the readout panel shows the expressions so you can see which is which.
This page describes mechanism. It is not medical advice and reading it creates no clinical relationship. Inhibitors are named because they are how the chain was mapped and how it is studied; several are lethal poisons, and nothing here is a guide to handling them. Clinical questions belong with a clinician.
The simulation could not start: React did not load, so the membrane model has nothing to render into. Everything else on this page is static and still readable.
| Quantity | Value here | Tier | Where it comes from |
|---|---|---|---|
| H+ pumped, Complex I | 4 per 2 e- | Measured | The consensus value, supported by the structures showing three antiporter-like subunits plus the E-channel. Two further protons are taken onto quinone and are carried, not pumped. |
| H+ per 2 e-, whole chain | 10 NADH, 6 succinate | Measured | 4 + 4 + 2 and 0 + 4 + 2 across Complexes I, III and IV. Both totals include the protons ferried on quinone and released at the Qo site. |
| Midpoint potentials | -320 to +816 mV | Measured | Standard biochemical values at pH 7 for the NADH/NAD+, Q/QH2, cytochrome c and O2/H2O couples. The ladder panel is drawn to scale from them. |
| c-ring stoichiometry | 8 subunits | Measured | The human mitochondrial ring, resolved structurally. It gives 8 protons for 3 ATP, about 2.7 per ATP in the matrix, and the ring size differs between species. |
| Cost per cytosolic ATP | ~4 H+ | Simplified | 2.7 at the ring plus roughly one more for the adenine nucleotide translocase and the phosphate carrier. Rounding to 4 is what produces the familiar 2.5 and 1.5, and both figures are conventions rather than constants of nature. |
| Electron flux, leak, settling rate | chosen | Chosen | The coefficients in the formula block under the readouts are tuned so the gradient settles at a watchable speed and the panels move at a countable pace. They are not kinetic constants and no rate on this page is a measurement. |
| Readout units | relative | Chosen | Oxygen consumption is a fraction of the model's own uncoupled maximum, ATP synthesis is in arbitrary units, and heat is a residual. Only the proton-motive force is on a scale you could compare with a real measurement, and even then loosely. |
| Number of mitochondria | 1 | Simplified | One idealised inner membrane at steady supply. No supercomplex organisation, no cristae geometry, no calcium, no substrate depletion, and no variation between tissues or people. |
| Particle count on screen | exact | Measured | The animation schedules particles from the stoichiometry above rather than for looks, so the counter under the diagram and a hand tally of one turnover agree. This is the one place the visual is meant to be checked. |
The organising idea is Peter Mitchell's chemiosmotic hypothesis: the chain and the synthase are not chemically coupled at all, only through a proton gradient across a sealed membrane. It was received badly for the better part of a decade and won the 1978 Nobel Prize in Chemistry. The Q cycle, which explains how Complex III doubles its proton yield out of a two-electron donor feeding a one-electron acceptor, is Mitchell's too. Paul Boyer's binding change mechanism, which says the synthase spends its energy releasing ATP rather than making it, shared the 1997 prize with John Walker's structure of the enzyme.
Run the model with NADH alone and the ledger is 10 protons per pair of electrons. Switch to succinate and it is 6, because Complex II sits in a 14 mV gap between succinate and quinone that is worth about 2.7 kJ per mole, against roughly 17 kJ to push one proton across a 180 mV membrane. The energy is simply not there, which is a more satisfying answer than calling it an exception.
The inhibitors are the reason any of this is known. Each one cuts the chain at a defined point and leaves everything upstream reduced and everything downstream oxidised, and the step where that flips is a crossover point. The crossover probe reproduces the classical experiment: rotenone at Complex I, antimycin at the Qi site of Complex III, cyanide at the a3 heme of Complex IV, oligomycin at the synthase, and an uncoupler that leaves the chain running flat out while nothing is captured.
Where the model stops being trustworthy. Every rate is invented, so the speed at which the gradient recovers, the exact height of the readouts, and the shape of the leak curve carry no evidence. The membrane is one idealised patch with unlimited substrate, so nothing here shows what happens when fuel runs out or when supercomplexes reorganise. And the P/O ratio is displayed to two decimals, which is more precision than the underlying convention deserves: the value depends on the translocase cost, the c-ring count of the species, and how much the membrane leaks, all of which are approximations here. The stoichiometry and the ordering are what survive; the numbers with decimal points are for feel.
Scope: this page models one idealised inner mitochondrial membrane at steady substrate supply, to show the ordering of the carriers, the proton bookkeeping, and where the conventional ATP yields come from. Stoichiometry, midpoint potentials, structural counts and inhibitor targets are attributable to the sources above. Every rate, every readout scale except the proton-motive force, and the settling behaviour of the gradient are properties of a phenomenological model with chosen coefficients, not measurements and not predictions for any tissue, species or person. Nothing here is medical advice.