Lincoln Taiz, University of California, Santa Cruz, California, USA

F_oF₁-ATP synthases (also called F-ATPases) are present in the inner membranes of mitochondria, chloroplasts, and bacteria. As discussed in Chapter 7 of the textbook, these large, multisubunit enzymes consist of a water-soluble catalytic complex (F₁) attached to an integral membrane protein complex (F_o) that transports protons across the membrane (see textbook Figure 7.30). The F₁ complex is composed of at least five different types of subunits (three α, three β, one γ, one δ, and one ε). When the F₁ complex is dissociated from the membrane, it is active as an ATPase. In fact, under the appropriate conditions the intact F_oF₁-ATP synthase can run in reverse and act as a proton pump, using the energy of ATP hydrolysis to move H⁺ across the membrane.

Our understanding of how ATP is synthesized was advanced by Paul Boyer at the University of California, Los Angeles, who proposed the binding-change mechanism for catalysis by F-ATPases (Web Figure 12.4.A) (Boyer 1997). From the detailed understanding of how the ATPases function, he proposed that the binding-change mechanism for ATP synthesis contains three important components:

Web Figure 12.4.A The binding-change mechanism as seen from the top of the F₁ complex. There are three catalytic sites in three different conformations: loose, open, and tight. (For clarity, only the three β subunits are shown.) Substrate (ADP + P_i) initially binds to the open site and is converted to ATP at the tight site. In step 1, rotation of the γ subunit causes a conformational change, resulting in a change in the formation of the sites. As a result, ATP is released from the enzyme. In step 2, substrate again binds to the open site, and another ATP is synthesized at the tight site. (After Duncan et al. 1995.)

1. The major energy-requiring step is not the synthesis of ATP from ADP and P_i, but the release of ATP from the enzyme.
2. Substrate is bound and products are released at three separate but interacting catalytic sites, corresponding to the three catalytic subunits (β subunits). Each catalytic site can exist in one of three conformations: tight, loose, or open.
3. The binding changes are coupled to proton transport by rotation of the γ subunit. That is, the flow of protons down their electrochemical gradient through the F_o complex causes the γ subunit to rotate.

Rotation of the γ subunit then brings about the conformational changes in the catalytic complex that allow the release of ATP from the enzyme, and the reaction is driven forward. The reverse occurs when the enzyme functions as an ATP-driven proton pump.

The first two predictions of the binding-change model are supported by many lines of evidence, mainly kinetic studies, and are now generally accepted. However, the prediction of a rotary mechanism for coupling proton flow to ATP synthesis has been more difficult to demonstrate. Two major breakthroughs have led to confirmation of the third prediction, as well.

The first breakthrough was the determination of the crystal structure of the F₁-part of bovine mitochondrial ATPase by the laboratory of John Walker in Cambridge, England (Abrahams et al. 1994). The crystal structure showed that the three catalytic β subunits differ in their conformations and in the nucleotide bound to them, consistent with the binding-change mechanism.

Even more exciting was the discovery that the γ subunit is inserted like a shaft through the center of the catalytic complex, which consists of three α subunits and three β subunits arranged alternately in a doughnut-like structure. Moreover, the interface between the γ subunit and the α and β subunits is highly hydrophobic. The hydrophobicity of the interface minimizes the interactions between the subunits, consistent with the rotation of the γ subunit within the hole formed by the catalytic complex. In other words, the γ subunit looks like a molecular bearing lubricated by a hydrophobic interface.

Although many questions were answered by elucidation of the crystal structure of the F₁-ATP synthase, the rotational model cannot be tested by a static "snapshot" of the enzyme. Definitive demonstration of rotation requires a video recording of the spinning of the enzyme in real time. But although they are large for proteins, F₁-ATP synthases are still far too small to be visualized in a light microscope.

To visualize the rotation of the enzyme, Masasuke Yoshida and his colleagues at the Tokyo Institute of Technology came up with an ingenious method to make the enzyme much larger than it is (Noji et al. 1997). As Web Figure 12.4.B shows, they attached an actin filament labeled with a fluorescent dye to the base of the γ subunit using another protein as a "glue." They then attached the F₁ complex upside down to a glass surface. If the γ subunit rotates with respect to the catalytic complex, the actin filament should swing around with it. Since the filament is very long compared to the ATP synthase (about 1 μm), its rotation should be visible in a fluorescence microscope. In other words, the fluorescently tagged actin filament, which is large enough to visualize in a light microscope, reports the rotation of the γ subunit.

Web Figure 12.4.B A method for visualizing rotation of the γ subunit. A fluorescently labeled actin filament was attached to one protruding end of the γ subunit. The F₁ complex was then attached upside down to a coverslip. When ATP was added to the coverslip, the actin filament rotated. (After Noji et al. 1997.)

The results were spectacular! When ATP was added to the modified enzyme, the actin filaments were seen to swing around in a circle at as much as 4 revolutions per second in a fluorescence microscope (in Web Figure 12.4.C the rotation rate is only 0.5 r.p.s.). For a video of the rotating ATPase, visit the following website:

http://docencia.izt.uam.mx/docencia/alva/atpaseyoshida.htm. To give some idea of scale, if you were a γ subunit, this rotation rate would be equivalent to swinging a several hundred meter-long rod around your head at 4 revolutions per second in water! However, the measured velocity is undoubtedly an underestimation of the actual velocity in vivo, because of the enormous torque required to swing such a large mass. Demonstration of the rotary motion of the γ subunit made it possible to put together a model of how the ATP synthase works (Web Figure 12.4.D). For their contributions to elucidation of the mechanism of ATP synthesis, Paul Boyer and John Walker shared half the Nobel prize in physiology or medicine in 1997. The other half went to Jens Skou for his pioneering work on the K⁺,Na⁺-ATP synthase, the mammalian counterpart of the plant plasma membrane H⁺-ATP synthase (see textbook Chapter 6).

Web Figure 12.4.C Sequential images of the rotating actin filament attached to the γ subunit, as viewed in a fluorescence microscope. There is 133 ms between the images and the rotation rate is 0.5 r.p.s. (From Noji et al. 1997, courtesy of S. Noji.)

Web Figure 12.4.D Model of the F_oF₁-ATPase, showing the attachment of the catalytic complex to the membrane via the β subunit and the δ subunit. When the reaction runs in reverse (ATP synthesis), protons diffuse through the F_o complex down their electrochemical gradient. The movement of protons through the channel drives the rotation of the entire F_o complex within the membrane. The γ subunit, which is attached to the F_o complex, then turns within the catalytic complex, causing the conformational changes that are required for ATP synthesis. It is assumed that the catalytic complex itself does not rotate, but is anchored to the membrane. The δ subunit is located on the outside of the β subunit and serves as the site of attachment of the β subunit, which anchors the catalytic complex to the membrane and prevents it from spinning. In mechanical terms, the F₁ complex and its membrane anchor act as a stationary housing, or “stator,” while γ subunit (and possibly the F_o complex) serves as the “rotor.” (From Junge et al. 1997.)