# The Proton-Motive Force and ATP Synthase

> Electron transport and ATP synthesis are one coupled machine, joined by a proton gradient across the inner mitochondrial membrane. Complexes I, III and IV spend the energy of falling electrons to pump protons uphill; ATP synthase lets those protons fall back through a rotary motor that physically turns to make ATP. The stored energy has an electrical part and a pH part, and the cell sets the pace by how fast ATP synthase drains the gradient. Experiments that built a gradient out of nothing but light and showed that it made ATP proved the idea, after Peter Mitchell proposed it in 1961.

Course: Biochemistry (Central Metabolism). Level: university elective. About 27 minutes.
Lesson page (link this one when citing): https://stemhop.com/lessons/proton-motive-force/

Builds on: Oxidative Phosphorylation: The Electron-Transport Chain · Membranes: The Impermeable Sheet That Builds Itself · Metabolism: Basic Concepts and Design

## What you'll learn

- Write the proton-motive force Δp as the sum of an electrical term (ΔΨ) and a chemical term (ΔpH), and explain why both the electrical (ΔΨ, ~0.15 V, matrix negative) and chemical (~0.5–1 pH unit) components matter — students who drop ΔΨ lose the larger part of the energy.
- Explain chemiosmotic coupling: how the proton gradient links electron flow to ATP synthesis so that blocking one (cyanide or oligomycin) stops or sharply slows the other, and cite the Racker-Stoeckenius reconstitution and Jagendorf acid-bath experiments as the proof (Mitchell's 1961 proposal is context).
- Describe ATP synthase as a real rotary motor — Fo c-ring rotor + a-subunit half-channels, F1 α3β3γ head — and walk Boyer's binding-change mechanism (O→L→T), explaining that energy is spent releasing tightly bound ATP, not forming it.
- Predict the effect of any ETC inhibitor, the ATP-synthase inhibitor oligomycin, or an uncoupler on three readouts (O2 consumption, ATP output, gradient size), and explain why oligomycin raises the gradient while uncouplers raise O2 consumption.
- Account for the ~30–32 ATP per glucose, justify why the number is approximate (shuttle choice and the proton tax on ATP export), and connect respiratory/acceptor control (ADP acting on the chain through the synthase and Δp) and energy charge (tuning fuel supply upstream) to the cell's real-time regulation of the rate.

**Give isolated mitochondria ADP and they start using oxygen faster, though no part of the respiratory chain ever touches ADP.**

The record comes from an [oxygen electrode](https://stemhop.com/methods/respirometry-and-calorimetry/), a probe that tracks the oxygen left in a chamber of mitochondria. With fuel but no ADP, the oxygen line slopes down slowly. Add a little ADP and the slope steepens at once: State 3, respiration with ADP. When the ADP has all become ATP, the slope flattens again: State 4, respiration with the ADP used up. The chain burns oxygen; ATP synthase, a separate protein, uses ADP. How does the chain know there is ADP?

In the [last chapter](https://stemhop.com/lessons/oxidative-phosphorylation/) Complexes I, III and IV pumped protons out of the matrix, like water raised behind a dam. The water is the protons, its height is the **proton-motive force**, Δp, and the turbine is ATP synthase, a real rotary motor a few nanometers wide.

> **The one idea to carry through this lesson.** Every proton that crosses the inner membrane is priced by one shared number, the proton-motive force Δp, worth about 19 kJ per mole of protons. The pumps earn Δp and must push against it, so they slow as it rises. ATP synthase and any leak spend it, and they speed up as it rises. Δp settles where earning equals spending. To predict any drug, disease or change in demand, ask which side it changed (earning, spending through the synthase, or spending through a leak), then follow Δp to the other side.

This holds for an intact, sealed inner membrane near steady state, where Δp is the only link between the chain and the synthase. It tells you which way each rate moves, not by how much. One exception: if the chain stops completely, the synthase can run backward (IF1, near the end).

*The film above previews this lesson’s last section: coupling, the poisons, the oxygen trace and IF1.*

## How we know: a gradient alone makes ATP

In 1961 Peter Mitchell proposed that the energy linking electron flow to ATP synthesis is stored not in a high-energy chemical intermediate but in a proton gradient across a sealed membrane: the chemiosmotic hypothesis. Two kinds of experiment settled it.

Jagendorf and Uribe soaked chloroplast membranes in acid, then jumped them into base: a hand-built proton gradient, with no light and no electron transport. They made ATP in the dark. Racker and Stoeckenius rebuilt vesicles from only purified lipids, bacteriorhodopsin (a light-driven proton pump borrowed from a salt-loving microbe) and ATP synthase. Shine light, and the pump builds a gradient, the gradient drives the synthase, and ATP appears, with no respiratory chain anywhere. With many other results, these experiments won the argument, and in 1978 Mitchell the Nobel Prize.

*Figure (AI-generated illustration).* Two experiments supporting chemiosmosis: Racker and Stoeckenius’s light-driven vesicles (main) and Jagendorf and Uribe’s acid-bathed chloroplast membranes in the dark (inset), each making ATP from a gradient with no respiratory chain. Image: a lipid vesicle containing bacteriorhodopsin and ATP synthase. Light drives protons inward through bacteriorhodopsin; protons return to the outside through the membrane portion of ATP synthase, powering ATP formation from ADP and phosphate at its external head. An inset shows acid-treated thylakoid membranes transferred into base in darkness, leaving an acidic lumen that drives ATP synthesis.

**Predict.** The lit vesicles are making ATP. Now add an **uncoupler**, a compound that carries protons straight back across the membrane. The light stays on. What happens to ATP synthesis?

> **Check your prediction.** It stops, or nearly. The pump still earns, but the leak spends faster than the pump can rebuild, so the gradient collapses and the synthase has nothing to spend. Pump and synthase share nothing but the gradient.

## What the dam stores: Δp, in one number

Two pulls draw a proton back into the matrix. The first is **electrical**: the pumps carry positive charge out, leaving the matrix about 150 mV negative, which attracts the positive protons back. This is the **membrane potential**, ΔΨ (“delta psi”). The second is **chemical**: the matrix is about 0.5 to 1 pH unit more alkaline, so protons are scarcer inside and drift in. The matrix is both *negative* and *alkaline*, so both pulls point inward.

But one is in millivolts and the other in pH units, and you cannot add 150 to 0.75. The bridge: at body temperature, a tenfold concentration difference, exactly one pH unit, pushes as hard as 61.5 mV ([log scales](https://stemhop.com/math-toolkit/#log-scales)).

> **Equation card: How hard the gradient pulls each proton back into the matrix, with the electrical and pH parts in one unit.**
>
>     Δp = |ΔΨ| + 61.5 mV × ΔpH
>
> In words: The proton-motive force equals the size of the membrane potential, plus 61.5 millivolts for every pH unit of difference across the membrane.
>
> - Δp: the proton-motive force: the total pull on a proton, in one unit (mV)
> - ΔΨ: the membrane potential; the bars mean we use its size (mV)
> - ΔpH: matrix pH minus intermembrane-space pH; about 0.5 to 1 (pH units)
> - 61.5 mV: the voltage worth one pH unit at 37 °C (2.303RT/F) (mV per pH unit)
>
> Sanity check: Both terms add. If the pH were equal on both sides, Δp would be ΔΨ alone. Textbooks that write a minus sign use a different sign convention, not different physics.
> Try one number Take ΔΨ = 150 mV and ΔpH = 0.75. Chemical part: 61.5 × 0.75 = 46 mV. Δp = 150 + 46 = 196 mV, about 0.2 V. The electrical part is 150 / 196, about three-quarters. Energy per mole of protons: F × Δp = 96.5 kJ per mole per volt × 0.196 V = 18.9 kJ/mol. Across ΔpH 0.5 to 1, this is about 17 to 20 kJ/mol.
> Where it comes from optional The electrical work to move a mole of protons through ΔΨ is F × ΔΨ; the chemical work across a ΔpH is 2.303 × R × T × ΔpH. Add them and divide by F: Δp = ΔΨ + (2.303RT/F) × ΔpH. At 310 K, 2.303 × 8.314 × 310 / 96,485 = 0.0615 V.

The electrical term is the larger, so forgetting it is the most common mistake here.

**Predict.** A mitochondrion is treated so that the pH is equal on both sides of its inner membrane, but its membrane potential rises to 180 mV. Can it still make ATP?

> **Check your prediction.** Yes. Δp = 180 + 61.5 × 0 = 180 mV, about 17 kJ per mole of protons, at the low end of the usual range. The synthase responds to the total pull. A student who tracks only the pH would predict no ATP, throwing away all of the energy.

### One price for every proton

That 19 kJ per mole is a price both machines face at once.

**For the synthase it is income.** Making ATP and delivering it to the cytosol costs about 50 to 60 kJ/mol in a working cell, far more than the 30.5 kJ/mol standard value, because the cell keeps ATP high and ADP low. 55 / 19 is about 3, so **at least about three protons** must fall per ATP delivered. The machine actually spends about 3.7 (counted below). That margin keeps it running forward; a machine spending exactly the minimum would sit at equilibrium and make nothing.

**For the pumps it is a cost.** Electrons passing from one NADH to oxygen release about 220 kJ per mole at standard potentials (about 210 in the cell), and the chain pumps about ten protons per NADH. At Δp = 200 mV those ten protons cost 10 × 19.3 = 193 kJ, about 88% of the standard total. What is left over, only a few tens of kJ (about 27 at standard potentials, about 17 in the cell), is the pumps’ **margin**. The chain runs close to its limit: if Δp climbs to about 220 mV, the margin is gone and electron flow slows to a trickle. That is **back-pressure**.

*Figure (AI-generated illustration).* One price for every proton. The pumps earn the proton-motive force (Δp) and must push against it; ATP synthase and any leak spend it. Δp settles where earning equals spending, so a change on one side reaches the other only through Δp (inset: adding ADP versus oligomycin). Image: the inner mitochondrial membrane with Complexes I, III and IV pumping protons out (earning) on the left, and ATP synthase and a heat-releasing leak letting protons back in (spending) on the right. A gold gauge in the middle stands for the shared proton-motive force. An inset compares adding ADP, where spending rises, the gauge dips slightly and the pumps speed up, with oligomycin, where the synthase is blocked, the gauge rises and the pumps slow.

## A motor, not a metaphor

ATP synthase has two parts named, unhelpfully, Fo and F1.

Fo sits in the membrane: a ring of c-subunits (the rotor) pressed against a stationary a-subunit that holds two offset half-channels. A proton enters one half-channel from the intermembrane space and lands on a conserved acidic residue on the c-ring. A charged group cannot sit in the oily core of the membrane; only a c-subunit whose acid carries a proton can turn away from the a-subunit into the lipid. The proton-motive force biases the ring’s random jostling. Protons board easily from the acidic, positive side and leave easily into the alkaline, negative matrix, so the ring steps mostly one way. The voltage and the pH difference both supply that bias, which is why the acid-bath chloroplasts could turn theirs with a pH jump alone. Nearly a full turn later, that c-subunit returns to the a-subunit and releases its proton into the matrix through the second half-channel.

F1 is the catalytic head in the matrix: three α and three β subunits arranged like orange segments around a bent, asymmetric γ shaft. The c-ring and γ are bolted together, so the proton flow that spins the rotor spins the shaft. γ’s off-center bend presses differently on each β subunit at any instant, holding each in a different shape.

*Figure (AI-generated illustration).* ATP synthase couples proton movement to rotation. Protons enter one a-subunit half-channel from the intermembrane space, bind to the c-ring, travel around with it and leave through the other half-channel into the matrix. The c-ring turns the gamma shaft within the matrix-facing F1 head, which contains three alpha and three beta subunits. Image: ATP synthase spans the inner mitochondrial membrane with its F1 head in the matrix and its c-ring connected to the gamma shaft. One inset shows six alternating alpha and beta subunits around the asymmetric shaft. A second schematic inset shows separate a-subunit half-channels connecting the intermembrane space and matrix to offset sites at the c-ring interface.

## Boyer’s inside-out insight: where the energy goes

Paul Boyer asked: what if it takes no energy to *make* ATP at the catalytic site, but it takes energy to get the ATP *off*?

That is the binding-change mechanism. Each β subunit cycles through three states as γ turns: **Open** (releases finished ATP, takes in ADP and Pi), **Loose** (ADP and Pi held loosely) and **Tight** (ATP forms spontaneously and is held so firmly it won’t leave). Each 120° turn of γ moves all three forward one state and releases one ATP, so a full turn releases three. John Walker later caught the three β subunits in a single crystal, one holding an ATP analogue, one ADP, one empty: a frozen snapshot of Boyer’s cycle.

So where does Δp’s energy go? On the enzyme, making the bond costs almost nothing either way. The gradient pays for the rest: above all, prying the finished ATP loose, and, as later work showed, driving the binding of ADP and Pi.

*Figure (AI-generated illustration).* Boyer’s binding-change cycle. Each 120° turn of the bent γ shaft moves every β subunit one state forward: Loose to Tight (ATP forms with almost no energy cost), Tight to Open (the finished ATP is released), Open to Loose (new ADP and phosphate bind). The gradient pays mainly for the release and for binding, not for making the bond. Image: three top-down views of the ATP synthase F1 head. In each, three beta subunits around a bent gamma shaft hold different states: open and empty, loose with ADP and phosphate, and tight with ATP. With each 120-degree turn of gamma every beta subunit moves one state forward, and the open subunit releases a finished ATP.

**Predict.** Detach F1 from the membrane, so there is no Fo and no gradient, and give it ATP. What does it do?

> **Check your prediction.** It splits ATP, and γ spins the opposite way from synthesis: without the gradient, the machine runs the easy way, downhill. The gradient does not make the chemistry possible; it sets the direction. Noji and Yoshida used exactly this: ATP-fed α3β3γ heads fixed to glass spun an attached fluorescent actin filament in discrete 120° steps, one per ATP (figure below). Later, Itoh and colleagues turned γ with magnets in the synthesis direction, and ATP appeared.

*Figure (AI-generated illustration).* Watching ATP-driven rotation. An isolated F1 head from a thermophilic bacterium is fixed to glass, and a fluorescent actin filament is attached to its gamma shaft. ATP hydrolysis drives counterclockwise rotation when viewed from the shaft side, with one ATP consumed per 120° step under stepping-assay conditions. Image: an isolated bacterial F1 head fixed to glass, with fluorescent actin attached to the projecting gamma shaft. ATP hydrolysis powers rotation. An inset viewed from the shaft side shows the filament at three successive positions 120° apart, with counterclockwise arrows and one ATP hydrolyzed per step.

## Counting protons and ATP

The vertebrate c-ring has 8 subunits, the smallest ring and lowest proton cost per ATP known. Eight protons and three ATP per turn: 8 / 3 ≈ 2.7 H⁺ per ATP. But the ATP must leave the matrix. The ATP–ADP translocase, one of the inner membrane’s most abundant proteins, swaps ATP⁴⁻ out for ADP³⁻ in, an electrogenic 1:1 exchange in which ΔΨ pulls one net negative charge outward: a proton tax on every ATP shipped. The phosphate carrier imports Pi with a proton. Together, the two carriers cost about one more proton-equivalent, for about 3.7 H⁺ per exported ATP, just above the floor of about three.

Cytoplasmic NADH can’t cross the inner membrane, so a shuttle hands its electrons in. The **glycerol-3-phosphate shuttle** delivers them to FAD and CoQ, bypassing Complex I, so that NADH is worth only about 1.5 ATP. The **malate-aspartate shuttle** regenerates matrix NADH, worth about 2.5 ATP. Many tissues run both, and the mix depends on enzyme levels and metabolic state.

On exams, use NADH ≈ 2.5 and FADH₂ ≈ 1.5, which gives **about 30–32 ATP per glucose**. Those rounded values assume about 4 H⁺ per ATP. The c8 machine’s 3.7 would allow about 2.7 and 1.6, or about 32–34 per glucose (depending on the shuttle) before any leak, but measured P/O ratios in isolated mitochondria come out near the exam values (Going deeper, “Shuttles and measured yields”). The word “approximately” is doing real work.

*Interactive on the lesson page:* ATP synthase as a molecular gearbox

## Setting the rate, and breaking it on purpose

*Where the film calls coupling a relationship, read two machines sharing one Δp. Where it says energy charge governs the rate, read: ADP reaches the chain through the synthase and Δp, and energy charge is read by enzymes upstream.*

Back to the opening trace.

**Predict.** Resting mitochondria have fuel but almost no ADP. You add ADP. What happens first to Δp, then to the pumps, then to oxygen use?

> **Check your prediction.** ADP lets the synthase spend, so Δp dips. A lower Δp lowers the price of each proton the pumps push out, their margin grows, electrons flow faster, and oxygen use rises: State 3. When the ADP is used up, spending drops, Δp climbs back toward the pumps’ ceiling, and the chain idles: State 4. The chain feels ADP only through Δp.

Δp itself moves only a little. A 10 mV dip in Δp saves about 10 kJ on the pumps’ ten protons, widening their few-tens-of-kJ margin by a third or more. So Δp barely rises when ADP runs out; small changes in Δp drive large changes in flow.

This is **acceptor (respiratory) control**: ADP availability is a main control on the pace, alongside substrate supply and other signals. Energy charge, a gauge of the cell’s balance of ATP, ADP and AMP, works upstream instead, telling enzymes such as phosphofructokinase-1 (PFK-1) how much fuel to send.

Poisons dissect the coupling, and the table they make is heavily tested. Read each on three gauges, O₂ consumption, ATP output and gradient size, and ask which side it changed.

**Predict.** Respiring mitochondria get **oligomycin**, which plugs Fo. Minutes later they also get an uncoupler (DNP or FCCP). Predict all three gauges after each.

> **Check your prediction.** **Oligomycin** removes spending through the synthase. Δp *rises* as back-pressure, and the pumps slow until earning matches the small remaining leak. O₂ down to the leak rate, ATP down, gradient **up**.
>
> **The uncoupler** adds a large leak. Δp drains without making ATP, the pumps’ margin opens wide, and the chain races to refill the gradient. O₂ **up**, usually past the starting rate; gradient down; ATP still zero. Given alone, an uncoupler gives O₂ up, ATP down, gradient down.

**ETC inhibitors** change the earning side: rotenone at Complex I, antimycin A at Complex III, cyanide or CO at Complex IV. Pumping stops: O₂ down, ATP down, gradient down. (That holds in isolated mitochondria whose fuel enters the chain upstream of the block, such as pyruvate plus malate with rotenone, and with no ATP to run the synthase backward; see Going deeper, “When blocking the chain does not empty the gradient.”)

Run all three on the dam below and predict each gauge before you click.

*Interactive on the lesson page:* The mitochondrial dam & turbine — coupling, predicted

Biology runs its own uncoupler. **UCP1 (thermogenin)** in brown fat deliberately leaks protons to make heat instead of ATP, the basis of nonshivering thermogenesis: O₂ rises while ATP output stays low.

The earning side can also stop. In ischemia, when oxygen runs short, the chain stalls and the gradient collapses, so the synthase can run backward, burning ATP to hold Δp up. The **IF1 inhibitor protein** locks the rotor at low pH, stopping that reverse spin and saving the cell’s ATP.

We have now followed one glucose from the gut to the last drop of ATP. Next we leave the mitochondrion for the chloroplast, where the same proton-gradient trick runs on sunlight instead of food.

## Going deeper

### Mitchell, Glynn House and the Nobel banquet

The high-energy intermediate was the molecule everyone was hunting and no one could find, and Mitchell’s alternative met years of resistance. Forced out of Edinburgh by illness, he restored a half-ruined Cornish mansion, founded a small private research institute there, and kept measuring protons with his collaborator Jennifer Moyle. At the 1978 Nobel banquet he said he was “not a little astonished.”

### Signs in the Δp equation

The lesson’s equation uses sizes, so both terms come out positive. Measured as inside minus outside, ΔΨ ≈ −150 mV and the equation reads Δp = ΔΨ − 61.5 mV × ΔpH ≈ −196 mV. Same size; only the sign bookkeeping differs.

### Where energy charge fits

Energy charge, ([ATP] + ½[ADP]) / ([ATP] + [ADP] + [AMP]), is held near 0.85–0.90. This fuel gauge tunes phosphofructokinase-1 (PFK-1), isocitrate dehydrogenase and pyruvate dehydrogenase, while the chain itself feels ADP only through the synthase and Δp.

### Shuttles and measured yields

The glycerol-3-phosphate shuttle is prominent in brown fat and some skeletal muscle and brain cells; the malate-aspartate shuttle is major in heart and liver, and also active in muscle and brain. Measured P/O ratios fall below the c8 ceiling mainly because of proton leak and slip: some protons return without making ATP.

### Rotor and speed details

The c-ring’s acidic residue is a glutamate in mitochondria and an aspartate in E. coli. Swap the long, draggy actin filament for a tiny bead and the motor spins roughly 130 revolutions per second, about 400 ATP per second.

### When blocking the chain does not empty the gradient

Rotenone blocks Complex I, but electrons from succinate still enter at Complex II. And in intact cells making ATP by glycolysis, the reversed synthase can spend that ATP to keep ΔΨ partly up after the chain is blocked.

### More on UCP1, IF1 and old gradients

Brown-fat thermogenesis can be lit up on a PET-CT in a cold adult. Pigs carry a broken UCP1 gene, which is why piglets rely on shivering. IF1 is overexpressed in some cancers, where its role is still debated. The proton gradient is older than mitochondria: chloroplasts and most bacterial flagella run on it too.

### Leber hereditary optic neuropathy as a respiratory-chain disease

Mitochondrial DNA mutations show the same principle. mtDNA is maternally inherited and present in many copies, so disease often depends on **heteroplasmy**, the mix of mutant and normal genomes in a cell, and on the mutant fraction, and strikes high-demand tissues first. LHON, most often caused by the m.11778G>A mutation in ND4 (a Complex I subunit), can blind otherwise healthy young adults, mostly men, over weeks, though many carriers never lose vision.

---

From STEMhop (https://stemhop.com), free lessons by Christopher Deppmann. Lesson text © the authors. Practice questions: https://stemhop.com/lessons/proton-motive-force/
