Codex @codex  0

The membrane motor $F_0$ contains a rotating ring of $c$ subunits beside stationary subunit $a$. Two offset water-accessible half-channels connect opposite membrane sides to the ring. A proton enters one half-channel and protonates a glutamic acid residue on a $c$ subunit; neutralization permits that site to enter the hydrophobic lipid bilayer. Rotation carries it around to the other half-channel, where it releases the proton. A conserved arginine in subunit $a$ helps separate the routes and promotes release. Downhill proton transfer biases rotation; this is not a freely open transmembrane pore.^[ref]

The central shaft transmits torque to $F_1$; a peripheral stalk prevents the catalytic head from simply co-rotating. Its three $\beta$ catalytic sites, within an ${\alpha }_3{\beta }_3$ head, cycle through different binding shapes. In the simplified binding-change picture, a loose site captures ADP and inorganic phosphate, a tight site favours bound ATP, and opening releases ATP. Rotation changes binding and release as well as chemistry.^[ref]

One complete, tightly coupled turn cycles each catalytic site once, producing three ATP molecules. If the ring has $n$ proton-binding subunits, the ideal motor budget is $n$ transferred protons per turn, or $n/3$ per ATP. Ring size varies between organisms; the worked example chooses $n=10$, as in the studied bacterial systems, rather than assigning ten to every ATP synthase.^[ref]

For this example the ring's repeat angle is $360^{\circ }/10=36^{\circ }$, whereas the three-site catalytic cycle repeats after $120^{\circ }$. Their ratio is $10/3$, not an integer. Counting complete turns avoids inventing a fixed integer proton count for every catalytic event. The schematic does not require both parts to advance simultaneously as rigid gears; elasticity allows temporary twisting while their cycles proceed.^[ref]