ATP synthase, a spinning machine

ATP synthase does more than let H⁺ through the membrane. Part of it spins like a tiny turbine, and that spinning is what forces ADP and Pᵢ together into ATP.

A gradient is all it needs

An H⁺ gradient alone can power ATP synthase. The electron transport chain’s role is to build that gradient, and a gradient made any other way works just as well. Racker and Stoeckenius showed this by purifying two proteins and inserting them into vesicles, tiny hollow spheres of membrane:

• ATP synthase, turned so its ATP-making part faced the outside, where ADP and Pᵢ were added
• bacteriorhodopsin, a protein from certain archaea that pumps H⁺ when light shines on it, turned so it pumped H⁺ into the vesicles

In the dark, no ATP was made. In the light, bacteriorhodopsin pumped H⁺ in, and ATP synthase made plenty of ATP as the H⁺ flowed back out. The vesicles had no electron transport chain, so the gradient alone must have powered ATP synthesis.

The parts of ATP synthase

ATP synthase is assembled from many protein pieces called subunits, grouped into two main portions. The subunits in the membrane have Latin letters (a, b, c); those that project into the matrix have Greek letters (α, β, γ, δ, ε).

How H⁺ flow turns the rotor

H⁺ crosses ATP synthase one ion at a time. Each incoming H⁺ slips through a half-channel in the a subunit and latches onto a c subunit, while a neighboring c subunit drops the H⁺ it picked up earlier into a second half-channel that leads to the matrix. Repeated over and over, this grab-and-release turns the c ring clockwise, and the γ and ε subunits, which sit on the ring, turn with it.

The a, b, δ, α, and β subunits stay still. So the γ stalk turns inside a fixed knob, like an axle turning inside its housing.

Three shapes of a β subunit

Each time the γ subunit turns 120° (a third of a circle), it changes how it touches each of the three β subunits, and each β subunit changes its shape, or conformation. A β subunit cycles through three conformations, a bit like a hand that grips, squeezes hard, then lets go:

The next 120° turn returns the β subunit to conformation 1, and the cycle starts again. At any moment, the three β subunits are in three different conformations. A 3D structure of the knob supported this model: it showed one β subunit holding ADP, one holding ATP, and one empty.

Watching it spin

ATP synthase is far too small to see with a light microscope, so another team (Yoshida and Kinosita) made its rotation visible:

• They removed the membrane part and stuck the rest (one γ, three α, and three β subunits) to a glass slide, with γ pointing up.
• To the γ subunit they attached a fluorescent (glowing) actin filament, a protein fiber long enough to see under a fluorescence microscope.
• With no membrane part, there was no H⁺ gradient, so they added ATP instead. ATP synthase can run backward, breaking ATP down, and that also turns the γ subunit.

With ATP present, the filaments spun. In the control, with no ATP added, nothing moved. The conclusion: ATP synthase is a rotary machine.

Check yourselfIs the electron transport chain always needed for ATP synthase to make ATP?

No. ATP synthase needs only an H⁺ electrochemical gradient. In the vesicle experiment, bacteriorhodopsin built the gradient using light, and ATP synthase made ATP with no electron transport chain present.

Check yourselfSuppose a mutation kept a β subunit from reaching conformation 3. What would happen at that site?

Conformation 3 is the only shape that binds ATP weakly enough to release it. The finished ATP would stay stuck, no new ADP and Pᵢ could bind, and that site would stop producing ATP.