19.2: ATP Synthesis

Search Fundamentals of Biochemistry

Learning Goals 

(Learning goals written by Claude, Sonnet 4.6, Anthropic)

Architecture and Rotary Mechanism of ATP Synthase

• Describe the two-part architecture of F₁Fo ATP synthase — identifying the membrane-embedded Fo proton translocator (c-ring, a-subunit, and peripheral stalk) and the soluble F₁ catalytic unit (α₃β₃ ring with central γ subunit) — and explain the analogy to an electric motor in which Fo is the rotor-driven electrical motor, and F₁ is the chemical generator.
• Explain Boyer's binding change mechanism (Loose-Open-Tight, LOT model), describing how the three interconvertible conformational states of the αβ dimer pairs differ in substrate affinity and catalytic activity, how γ subunit rotation drives sequential LOT transitions, and how phosphate binding before ADP in the synthetic direction prevents unproductive ATP binding at a high ATP:ADP ratio.
• Describe the direct experimental evidence for γ subunit rotation — including Noji's actin filament fluorescence experiment and the colloidal gold nanoparticle experiment — explaining how these single-molecule observations demonstrated counterclockwise 120° steps coupled to individual ATP hydrolysis events and confirmed that mechanical work equals the free energy of ATP hydrolysis.

Proton Translocation through F_o

• Describe the mechanism of proton translocation through the Fo subunit, explaining the roles of the critical Asp/Glu residue on each c-subunit (e.g., Asp61 in E. coli, Glu58 in bovine), the conserved Arg residue in the a-subunit, and the two aqueous half-channels (lumenal and matrix) that allow protons to access and leave the c-ring without crossing the hydrophobic membrane core.
• Explain how the sequential protonation and deprotonation of c-ring Glu/Asp residues at the lumenal and matrix half-channels — separated by the positively charged Arg residue in the a-subunit that prevents short-circuit proton flux — generates an electrostatic force on the deprotonated c-ring Glu that drives net counterclockwise c-ring rotation and hence γ subunit rotation and ATP synthesis.
• Explain the mechanisms of inhibition by oligomycin A (blocking the a-subunit/c-ring interface and proton translocation) and dicyclohexylcarbodiimide/DCCD (covalently modifying the protonated Asp61/Glu59 in a single c-subunit, halting all c-ring rotation), and connect the abnormally high pKa of this residue to its hydrophobic membrane environment.

Thermodynamics of ATP Synthesis

• Derive the proton-motive force (∆p) as the sum of the chemical potential contribution from the pH gradient (ΔG = 2.303RT·ΔpH per proton) and the electrical potential contribution (ΔG = ZFΔψ), calculate the total free energy available per proton translocated using physiological values (ΔpH = −1.4, Δψ = −0.14 V), and demonstrate that the combined electrochemical driving force is thermodynamically sufficient to drive the endergonic synthesis of ATP (ΔG°′ ≈ +30 kJ/mol).

Introduction

ATP synthase, also called F₁F_oATPase, is a rotary motor enzyme. This enzyme is found in the inner mitochondrial membrane, the analogous thylakoid membranes of chloroplasts, and the bacterial cell membrane. The enzyme consists of two parts: the membrane-bound Fo, a proton translocator, and the F₁ part, which has catalytic activity (ATP synthesis or ATP hydrolysis). The F_o part can be considered a rotary electrical motor powered by proton flow, while the F₁ part acts as a rotary chemical motor powered (in reverse) by ATP hydrolysis. The F_o part is named because it is sensitive to oligomycin (note that it should be read as F_"letter" o, not F_zero). The F₁ part is named after its elution in Fraction 1 from a column chromatography column.

Many, but especially Paul D. Boyer and John E. Walker, elucidated its mechanism and structure. They, along with Jen Skou, who studied the mechanism of the analogous Na/K-ATPase, were awarded the Nobel Prize for their work.

Let's begin with the known structure of this complex membrane-bound rotary enzyme, then work toward understanding how it functions. The structure of the bovine F₁F_oATPase is shown in Figure \(\PageIndex{1}\).

Figure \(\PageIndex{1}\): Bovine F₁F_oATPase.

In the previous section, we discussed how oxidative electron transport in the mitochondria is accompanied by proton transport against a concentration gradient from the matrix through the inner membrane to the intermembrane space. ATP synthesis from ADP and P_i is endergonic and requires an energy source. The thermodynamically favored collapse of the pH gradient across the mitochondrial inner membrane provides that energy source. Protons flow through the F_o membrane-spanning C helices from the intermembrane space (not shown in Figure 1) to the matrix. This powers ATP synthesis from ADP and P_i by the F₁ part of the enzyme in the matrix lumen.

Figure \(\PageIndex{2}\) shows an interactive iCn3D model of the entire bovine mitochondrial ATP synthase (5ARA). The subunits are shown in the same color as Figure 1 above.

Figure \(\PageIndex{2}\): Bovine mitochondrial ATP synthase (5ARA). (Copyright; author via source). Click the image for a popup or use this external link: https://structure.ncbi.nlm.nih.gov/i...3YUBWrX91nnua9

The enzyme is reversible. If protons flow down a concentration gradient through F_o, ATP is synthesized by F₁. Alternatively, ATP hydrolysis by F₁ drives proton transport through F_o and against a concentration gradient. Isolated F₁ can only break down ATP and not synthesize it.

We call this structure a rotary enzyme, so what rotates? A rotary "axle" protein complex comprised of the γ, ε, and δ subunits (some also describe the γ, ε, and δ subunits as part of F₁) rotates. Its interactions with the multimeric complex of α-helices (shown in purple in the figures above) in the main F_o complex cause the "c-ring complex" in F_o to rotate as well. The F₁ part does not rotate because of the conformational stability of the β subunit and the connection to the long alpha helices of the D and B1 proteins, which comprise the "stator (the stationary) part of an electric motor", which keeps F₁ stationary.

Hence, the enzyme acts like an electric motor, whose components are shown in Figure \(\PageIndex{3}\).

Figure \(\PageIndex{3}\): Electric motor with rotor, shaft, and stator. http://www.cdmmotor.com/en/article-22982-28485.html

Now, let's examine the mechanisms and evidence for each part of this remarkable enzyme.

F₁

The F₁ unit (with a quaternary structure of a₃β₃ forming a hexagonal ringed structure with a central cavity occupied by a gamma subunit)is about 80 angstroms from the Fo subunit, and both are connected to the rod-shaped γ subunit, which spans the center of the a₃β₃ ring. Energy transduction (necessary to capture the negative free energy change associated with the collapse of the proton gradient to drive the positive free energy change for ATP synthesis) occurs between the two subunits.

Figure \(\PageIndex{4}\) shows a top view of the α (red) and β (green) subunits in the a₃β₃ F₁ complex.

Figure \(\PageIndex{4}\): Top view of the α (red) and β (green) subunits in the a₃β₃

The γ subunit is shown coming out of the plane of the ring. ATP is shown in gold spheres, ADP in purple, and the smaller SO₄^2- (from a crystallizing agent) is shown in CPK spheres. They are sandwiched between the α and β subunits.

Boyer, in the absence of the complete structure of the F₁F_o ATP synthase, was able to deduce from experimental evidence that the a₃β₃ complex, which can be viewed as three aβ dimers (with catalysis occurring between subunits of individual dimers where ATP and ADP bind), has three different, interconvertible conformation defined as a Loose (L), Open (O) and Tight (T) states, with names describing the strength of substrate binding in each dimer.

• O - open state with very low affinity for substrates and has no catalytic activity;
• L - loose state with low affinity for substrates and also no catalytic activity;
• T - tight state with high affinity for substrates and with catalytic activity.

The three states rotate not physically around a central axis but conformationally, depending on their interactions with the γ subunit, which binds perpendicularly at the central junction of the a₃β₃ ring. Changes in the orientation of the central γ subunit due to its rotation with respect to the a₃β₃ ring cause the conformation of the O, L, and T states to change in situ with the orientation of the rotating γ subunit

The conversion of the LOT conformations, their binding of substrates (ADP and P_i), the conversion of bound ADP and P_i, and the release of the product (ATP) proposed by Boyer are shown in Figure \(\PageIndex{5}\).

Figure \(\PageIndex{5}\): Boyer's three-state conformational model (L-O-T) for ATP synthesis

The collapse of the proton gradient (i.e., the proton-motive force) causes the γ subunit to rotate like a crankshaft relative to the F1 subunit, forcing the β subunit to change conformation from the T to the O (releasing ATP) and then to the L (binding ADP and Pi) states. The γ subunit does not appear to undergo any significant conformational change on ATP hydrolysis, as evidenced by tritium exchange studies of amide protons.

To prove that the γ subunit rotates, you'd have to observe a single molecule. Since the γ subunit was too small to discern its rotation visually, Noji et al. covalently attached a fluorescein-labeled actin filament to the γ subunit (near where F_o would bind). The entire F1 molecule was fixed to a glass slip through a His-tag such that the a₃β₃ ring was immobilized. The γ subunit was free to rotate, as evidenced by fluorescence from the attached actin filament observed under a microscope.  Figure \(\PageIndex{6}\) describes this experiment and the outcomes.

Figure \(\PageIndex{6}\): The direct observation of the γ rotation in the F₁ motor. Noji and Yoshida. JBC (2001) DOI:https://doi.org/10.1074/jbc.R000021200. Creative Commons Attribution (CC BY 4.0)

Panel A illustrates the experimental setup for observing γ rotation using an optical microscope. The F₁ motor, tagged with 10 His residues at the N-terminus of the β subunit, was immobilized upside down on a coverslip coated with nickel-nitrilotriacetic acid (Ni-NTA). An actin filament (green) labeled with fluorescent dyes and biotins was attached to the biotinylated γ subunit (gray) through streptavidin (blue).

Panel B shows the rotary movement of an actin filament observed from the bottom, the membrane side, with an epifluorescent microscope. Length from the axis to tip, 2.6 μm; rotary rate, 0.5 revolutions per second; the time interval between images, 133 ms.

The actin filament rotated only in the presence of ATP. It rotated only counterclockwise, indicating that the motion was not random but a specific motion of the γ subunit. At extremely low ATP concentrations, rotation occurred only in 120° increments, implying one step per ATP molecule hydrolyzed. (Remember the β subunits are separated by 120 ^° .) As rotation occurs, the actin filament experiences viscous resistance to movement. He calculated that for a single 120° step caused by hydrolysis of a single ATP molecule, the work was 80 piconewtons, approximately the free energy of hydrolysis of a single ATP molecule.

Later experiments using a colloidal gold nanoparticle (40 nm in diameter, with lower frictional resistance to movement) instead of an actin filament showed the same result. At low [ATP], the motor rotates in 120^o steps. At high [ATP], the rotation rate becomes continuous and saturates (with Michaelis/Menten kinetics) at 130 revolutions per second.

Other experiments using immobilized ATPase and magnetic tweezers have investigated the timing of substrate binding and product release during reverse operation (ATP hydrolysis). On rotation of the γ subunit, the three binding sites change properties. In hydrolysis, ATP binds to the open site and helps promote the 120-degree rotation. Next, ATP is hydrolyzed. In the final step, products dissociate. Pi dissociation occurs last from the third site. Hence, each of the three beta-binding sites has a different role. One binds the substrate, one catalyzes the reaction, and the third releases products. Assuming the synthesis pathway is the reverse of the ATPase reaction, the final release of Pi in ATP cleavage predicts that Pi binds first in the synthetic direction. This would, in turn, preclude ATP binding, which matters because its concentration during synthesis can be 10 times that of ADP. As Pi is bound first, only ADP, not ATP, can bind next.

The γ subunit's rotation plays a "catalytic" role by inducing cyclic conformational changes in the β subunit of the synthase. Can ATP synthesis occur without the gamma subunit by a mechanism that involves a less proficient, but concerted set of cyclic changes in beta subunit conformation? It can! Uchihashi et al. have used high-speed atomic force microscopy (AFM) to study the a₃β₃ ring of the F1 subunit in the absence of the gamma subunit. They found that upon ATP hydrolysis, the beta subunits underwent conformational changes in the same counterclockwise rotary direction as when the gamma subunit was present. This is illustrated in Figure \(\PageIndex{7}\).

Figure \(\PageIndex{7}\): AFM Study of Conformational Changes in F1 "gammaless" subunit

These experiments conclusively show that the F1 subunit is effectively a rotary motor, with the gamma subunit acting as a rotor in the stationary hexagonal ring composed of the three pairs of alpha/beta subunits, which acts as the stator (the stationary part of an electric rotary motor).

The amino acids involved in ATP synthesis/hydrolysis are still not clearly defined, but Glu 190 on the beta subunit acts as a general base. Figure \(\PageIndex{8}\) shows bound ADP and the proximity of Glu 188. Ala 158 is thought to move towards the active site after a conformational change, with the nonpolar methyl side chain displacing an adjacent water molecule, which could leave as a product of ATP synthesis.

Figure \(\PageIndex{8}\): Active site residues for the synthesis of ATP. ADP is shown in spacefill, and CPK colors. Mg²⁺ is shown as a gray sphere (1E79).

F_o and Proton Transfer

The mechanism by which the proton gradient drives ATP synthesis involves a complex coupling of the F_o and F₁ subunits. Proton translocation occurs through the interface of the inner mitochondrial (or cell membrane of bacteria) membrane proteins C1 and A (or beta), whose structures from the bovine enzyme are shown again in Figure \(\PageIndex{8}\).

 

Figure \(\PageIndex{8}\): Inner membrane proteins C1 and A (or beta)

 

For protons to translocate through the membrane, there must be an aqueous channel on the matrix and lumenal sides (bounded by the inner and outer mitochondrial membranes). So we have to understand two features:

• How does the rotary axle part of the complex interact with the C1 hexameric ring, causing the ring to rotate in the membrane?
• The nature of the water channels and the pathway for proton translocation.

The number of c monomers in the c-ring can vary between 8 and 17. For example, the bovine ATP synthase in Figures 1 and 2 has eight c-monomers. Each has a critical proton donor and acceptor near the a- or beta-F_o chain. In the bovine case, it is glutamate 58. (The side chains in the bovine structure presented in Figure 2 are not shown since they were not resolved in the cryoEM structure.) In E. Coli, which we will explore further below, it is aspartate 61. In yeast, it is at position 59. They are positioned near the center of the membrane helices. Mutations in these residues cause significant decreases in enzyme activity. An E56Q mutation in some bacilli species prevents proton pumping and ATP synthesis. Another key charged group, arginine, in the α or β Fo protein is also important.

Figure \(\PageIndex{9}\) shows an interactive iCn3D model of the E. Coli A₁-C₁₂ subcomplex of F₁F_O ATP Synthase (1C17).

Figure \(\PageIndex{9}\): E. Coli A₁-C₁₂ Subcomplex OF F₁F_O ATP Synthase (1C17). (Copyright; author via source). Click the image for a popup or use this external link: https://structure.ncbi.nlm.nih.gov/i...sfRJBHEBiC5sy9

The key aspartate 61 (sticks and labeled) in each C ring subunit (red) and the proximity of two of the aspartate 61s to arginine 210 in the A subunit (brown) are evident.

Two classic ATP synthase inhibitors (structures shown below) bind the F_o subunit. One, oligomycin A, binds between the a and c subunits and blocks the proton transport activity of the F_o subunit. The O protein at the top of the F₁ complex is called the Oligomycin-sensitivity-conferring protein (OSCP), even though oligomycin does not bind there. The soluble F₁ by itself is not sensitive to oligomycin, but when it's linked to F_o in part through the O or OSCP peripheral stalk protein, it becomes sensitive to oligomycin.

Another inhibitor, dicyclohexylcarbodiimide, reacts with a protonated Asp 61 in c subunits of F_o. It does so even at pH 8.0, indicating that the pKa of Asp61 is significantly higher than usual. This might occur if the Asp is in a very hydrophobic environment. Modifying only one Asp61 in one c-subunit is sufficient to stop F_o activity. The protonated carboxyl group donates a proton to a nitrogen atom in DCCD, which then reacts with the deprotonated Asp to form an O-acyl isourea derivative. Figure \(\PageIndex{10}\) shows the structures of oligomycin A and DCCD, inhibitors of proton transport by F_o.

 

Figure \(\PageIndex{10}\): structures of oligomycin A and DCCD

Figure \(\PageIndex{11}\) shows an interactive iCn3D model of the F₁F_oATPase c10 ring with bound oligomycin (4F4S)

Figure \(\PageIndex{11}\): E. Coli A₁-C₁₂ Subcomplex OF F₁F_O ATP Synthase (1C17). (Copyright; author via source). Click the image for a popup or use this external link: https://structure.ncbi.nlm.nih.gov/i...pJw1pFr8jTPmK8

Only two of the c-subunits are shown interacting with a single oligomycin molecule. The actual complex has 10 c-subunits and eight oligomycins bound. The oligomycin is shown in spacefill with CPK colors. The two c-subunits are shown in gray. The side chains of the two c-subunits that interact with oligomycin are shown as CPK-colored sticks. They are conserved between humans and yeast. Note the single H-bond between the critical Glu 59 and oligomycin. The green colors indicate hydrophobic parts of the c-subunits, which interact with the mostly nonpolar parts of the antibiotic.

How do these key Asp or Glu side chains participate in proton translocation across the membrane and c-ring rotation, and how do they access the matrix and lumenal water channels?

We can start to understand the answer to these question by examining the structure of the c-ring and its interaction with the a-chain from Bacillus PS3 ATP synthase, as shown in Figure \(\PageIndex{12}\), which show the key Glu59 residues in each of the c subunits of the 10-mer ring along with Arg 176 and two other glutamates (Glu 223 and Glu 162) in the A ring.

Figure \(\PageIndex{12}\): Schematic picture of the a-subunit and c-ring of F_o. Mitome et al. (2022). eLife 11:e69096. https://doi.org/10.7554/eLife.69096. Creative Commons Attribution License

Panel (a) shows the ac₁₀ part of the F_o region, with key amino acids shown as colored spheres (numbered according to the yeast protein).

Panel (b) shows a cartoon of a simulation model developed by Mitome et al (ibid). The Glu 59s are shown as green circles and labeled a-j. The gray represents the membrane.

Protons move between the protonated Glu 59 in c subunits proximal to the deprotonated Glu 223 and Glu 162 in a hydrogen-bond "handshake". The Glu 59s in the c chains in the 10-mer ring are likely to be protonated, given their position in the center of the membrane. In contrast, the proximal Glu 223 and 162 are more likely to be initially deprotonated, given their proximity to aqueous channels, where they "hand off" their protons to complete the translocation from the intermembrane luminal space to the matrix. Now, once Glu59 is deprotonated, it carries a negative charge.

Simulations suggest that 2-3 deprotonated Glu59 residues have a negative charge proximal to the a-subunit. One likely takes up a proton, leading to ATP synthesis. The orange arrows show the direction of proton flow. ATP is hydrolyzed if the c-ring rotates clockwise, while counter-clockwise rotation leads to ATP synthesis.

Let's look at the water channel forming at the interface between the c-ring and an a-subunit and an additional protein, ASA6, from the unicellular green alga Polytomella sp. Figure \(\PageIndex{13}\) shows the interactions and nature of the water channels.

 

Figure \(\PageIndex{13}\): Two aqueous channels in F_o. Niklas Klusch, Bonnie J Murphy, Deryck J Mills, Özkan Yildiz, Werner Kühlbrandt (2017) Structural basis of proton translocation and force generation in mitochondrial ATP synthase eLife 6:e33274. https://doi.org/10.7554/eLife.33274. Creative Commons Attribution License

Subunit a is shown in blue, the c₁₀-ring in yellow, and the associated ASA 6 protein in brick. The channels are shown as potential surfaces (red: negative; blue: positive; grey: neutral). (A) and (B) display a 5 Å slice of the c₁₀-ring at the level of the protonated cGlu111.

Panel(A) shows a 5 Å slice of the c₁₀-ring at the level of the protonated c-chain Glu111, with the lumenal channel seen from the crista lumen.

Panel (B) shows a 5 Å slice of the c₁₀-ring at the level of the protonated c-chain Glu111, with the matrix channel seen from the matrix.

Panel (C) shows a side view of both channels from the c-ring, with outer c-ring helices shown in transparent yellow. Lumenal channel, left; matrix channel, right. The strictly conserved a-chain Arg239 in helix 5 (H5) separates the lumenal and matrix channels.

Panel (D) shows that the lumenal channel passes through the H5/H6 hairpin at the small side chains aAla246, aGly247 (H5), and aAla292 (H6) (green).

Panel (E) shows H4, the N-terminal half of H5, and the connecting H4/H5 loop at the matrix channel.

Figure \(\PageIndex{14}\) a probable proton pathway through the F_o subcomplex

Figure \(\PageIndex{14}\): Proton pathway through the F_o subcomplex. Klusch et al., ibid.

The a-subunit is shown in blue, the adjacent c-ring helices in transparent yellow, and the aqueous channels in translucent grey.

Panel (A) shows the movement of protons (red arrow) in the luminal channel. They likely move between Glu172, His248, His252, and Glu288 in the a-chain (dashed red ellipse) through the H5/H6 helix hairpin at the small side chains of a-chain Ala246 and Gly247 (H5) and Ala292 (H6) (green) to c-chain Glu111 in the rotor ring c-subunits (red circles).

Panel (B) shows that a-chain Arg239 (blue circle) is located halfway between the lumenal channel on the left and the matrix channel on the right, forming a seal to prevent proton leakage. c-ring helices (transparent yellow) with cGlu111 are seen in the foreground.

Panel (C) shows the movement of protons (dashed red arrow) in the matrix channel as they pass directly from the deprotonated c-chain Glu111 to the pH 8 matrix.

A final summary view showing c-ring rotation powered by the collapse of the pH gradient between the lumenal (intermembrane space) channel (pink) and matrix channel (light blue) is shown in Figure \(\PageIndex{15}\).

Figure \(\PageIndex{15}\): c-ring rotation is powered by the potential gradient between the lumenal channel (pink) and the matrix channel (light blue). Klusch et al., ibid.

The c-ring (yellow) and the membrane-intrinsic four-helix bundle of subunit a (blue) are drawn to scale, as seen from the matrix. Protons (red) pass from the cristal lumen below the projection plane through the lumenal channel between H5 and H6 to protonate Glu111 of c-subunit A.  Subunit J is deprotonated by the higher pH of the matrix channel. The positively charged a-chain Arg239 will likely interact with the deprotonated c-chain Glu111 during its short passage to the lumenal channel. The lumenal and matrix channels approach within 5–7 Å. A protonmotive force of 200 mV across the closely spaced channels creates a local electrostatic field of 40-100 million V/m, depending on the protein's dielectric. The field exerts a force on the deprotonated c-chain Glu111, causing a net counterclockwise rotation of the c-ring (grey arrow). Scale bar, 10 Å.

Click the images below to view a series of stunning videos (in a new window) showing proton translocation across the channel. (Niklas Klusch, Bonnie J Murphy, Deryck J Mills, Özkan Yildiz, Werner Kühlbrandt (2017) Structural basis of proton translocation and force generation in mitochondrial ATP synthase eLife 6:e33274. https://doi.org/10.7554/eLife.33274. Creative Commons Attribution License)

The three-dimensional arrangement of subunit a (blue), c-ring (yellow), with the lumenal channel in pink and matrix channel in light blue.
Arrangement of channel-lining side chains for the lumenal channel. Sidechains in stick representation are colored as subunit a, blue; c-ring, yellow; ASA 6, brick; lipids, grey. Channels are shown as potential surfaces (red, negative; blue, positive; grey, neutral).
Arrangement of channel-lining side chains for the matrix channel. Side chains in stick representation are colored as subunit a, blue; c-ring, yellow; ASA 6, brick. Channels are shown as potential surfaces (red, negative; blue, positive; grey, neutral).

 

In summary, F₁F_oATP synthase is a rotary enzyme that ultimately couples the collapse of a proton gradient (a chemical potential gradient contributing to the transmembrane electrical potential) to a chemical (phosphorylation) step. The rotor, which contacts both the F_o proton pore and F1 synthase, moves relative to both subunits, coupling them. Both the Fo axle and c-ring rotate. The interaction of the delta protein with the F₁ a₃β₃ hexamer does not cause a₃β₃ to rotate as a whole, as does the c-ring. Instead, it causes sequential rotation of the three aβ dimers, transitioning from the Loose to the Open to the Tight (LOT) conformations. The F_o pore can hence be considered an electrical motor and the F₁ synthase a chemical motor. Carrying the analogy of a motor even further, the F_o electrical motor turns the F₁ chemical motor into a generator, not of electricity but of ATP. The figure and link below, taken from the Protein Data Bank, provide more in-depth information about this nanomotor.

The concerted movement of the c-ring and axle protein with respect to the not physically rotating a₃β₃ F₁ part can be seen in Figure \(\PageIndex{16}\), which is derived from the cryoEM structure of different rotated states of bovine ATP synthase.

Figure \(\PageIndex{16}\): Structure and conformational states of the bovine mitochondrial ATP synthase (5ARA, 5ARE, 5ARH, 5ARI, 5FIJ, 5FIK, 5FIL)

A phenomenal video is available showing the detailed steps in ATP synthesis by ATP synthase. Click on Figure \(\PageIndex{17}\) to get the link to the video.

 

Figure \(\PageIndex{17}\): ATP synthase in action. Muzzey and Lue, HHMI Institute. For educational and non-commercial use only.

Researchers can purify this enzyme and bacteriorhodopsin and reconstitute them into lipid vesicles. If these proteins are properly oriented, light can drive proton translocation into the vesicle lumen, generating a transmembrane potential. The proton gradient can collapse as protons move from the lumen through the F_o part of F₁F_o ATP synthase, causing ATP synthesis from ADP and Pi in the outside solution. This is shown in Figure \(\PageIndex{18}\).

Figure \(\PageIndex{18}\):. Light-driven ATP production. Ahmad et al. ACS Synthetic Biology 2021. https://pubs.acs.org/doi/10.1021/acssynbio.1c00071. https://creativecommons.org/licenses/by/4.0/

Panel(A) shows a cartoon of the light-driven ATP synthesis in lipid vesicles. Panel (B) shows the increase in [ATP] over time.

 

Proton Gradient Collapse and ATP Synthesis - Thermodynamics

Mathematical analyses indicate that the free energy change associated with proton-gradient collapse can readily power the endergonic synthesis of ATP. Consider a typical pH gradient (-1.4 pH units) across the inner membrane of respiring mitochondria (with the outside having a lower pH than the inside, making the inside more depleted in protons). A chemical potential difference in protons exists across the membrane. However, another factor determines the thermodynamic driving force for proton translocation across the membrane. A transmembrane potential exists across the inner membrane of the mitochondria, as it does across most membranes. The source of the membrane potential will be discussed in the signal transduction chapter (Chapter 28). The inside is more negative than the outside, creating a transmembrane electrical potential across the membrane of about -0.14 V. Clearly, protons would be attracted to the other side of the membrane (into the matrix) by this potential difference, which then augments the chemical potential difference. A simple mathematical derivation shows that, indeed, a proton gradient can supply enough energy for ATP synthesis, especially when coupled to a transmembrane electrical potential. We reshow a small image of the complex and its activities in Figure \(\PageIndex{19}\) to help you through this section

Figure \(\PageIndex{19}\): ATP synthase and its activities. Noji, H., Ueno, H. & Kobayashi, R. Correlation between the numbers of rotation steps in the ATPase and proton-conducting domains of F- and V-ATPases. Biophys Rev 12, 303–307 (2020). https://doi.org/10.1007/s12551-020-00668-7. Creative Commons Attribution 4.0 International License. http://creativecommons.org/licenses/by/4.0/.

Chemical Potential (use G instead of μ)

Since most students use free energy rather than chemical potential (free energy/mol at constant T and P), we will derive these equations using free energy G.

Let's consider the free energy of the protons on the outside (low pH, higher concentration) and in the inside (matrix, higher pH, low concentration):

\begin{equation}
\mathrm{H}^{+}{ }_{\text {out }} \longrightarrow \mathrm{H}^{+}{ }_{\text {in }} \mathrm{G}=\mu=\text { chemical potential }(\mathrm{kcal} / \mathrm{mol})
\end{equation}

Now express this as the change in free energy, ΔG

\begin{equation}
\Delta G_{\text {chem }}=G_{\text {in }}-G_{\text {out }}=G^{0}+R T \ln \left[H^{+}\right]_{\text {in }}-\left(G^{0}+R T \ln \left[H^{+}\right]_{\text {out }}\right)
\end{equation}

This can be rewritten as:

\begin{equation}
\Delta G_{\text {ohem }}=\operatorname{RT|n} \frac{\left[H^{+}\right]_{\text {in }}}{\left[H^{+}\right]_{\text {out }}}=2.303 R T\left[-\log \left[H^{+}\right]_{\text {out }}-\left(-\log \left[H^{+}\right]_{\text {in }}\right)\right]
\end{equation}

Hence

\begin{equation}
\Delta \mathrm{G}_{\text {chem }}=2.303 \mathrm{RT}\left(\mathrm{pH}_{\text {out }}-\mathrm{pH}_{\text {in }}\right)=2.303 \mathrm{RT} \Delta \mathrm{pH}
\end{equation}

In respiring mitochondria, ΔpH = -1.4, so

\begin{equation}
\Delta \mathrm{G}{ }_{\text {chem }}=2.303 \mathrm{RT} \Delta \mathrm{pH}=2.303(1.99 \mathrm{cal} / \mathrm{mol} \cdot \mathrm{K})(298 \mathrm{~K})(-1.4)=-1.91 \mathrm{kcal} / \mathrm{mol} \mathrm{} \mathrm{H}^{+}
\end{equation}

Electrical Potential

Consider the relationship between ΔG and the transmembrane electrical potential

\begin{equation}
\Delta G_{\text {elect }}=+Z F \Delta \psi
\end{equation}

where Δψ is the transmembrane potential, Z=+1 (charge on a proton), and F is the Faraday constant

This equation is similar to \begin{equation}
\Delta G \text { elect }=-n F \Delta E
\end{equation}

where ΔE is the cell potential or EMF. If ΔE is + and Δψ is -, then ΔG is -.

In respiring mitochondria, Δψ = -0.140 V (remember that 1 V = 1 J/C). This might seem small, but consider this.  On a larger scale, it is equivalent to an electrical field strength of approximately 30 million Volts per Meter.  Compare this to an electric field right before a lightning strike of about 3,000,000 V/m.   Hence

\begin{equation}
\Delta G_{\text {elect }}=+Z F \Delta \Psi=1(96.485 \mathrm{C} / \mathrm{mol})(-0.140 \mathrm{~J} / \mathrm{C}) \times(1 \mathrm{cal} / 4.18 \mathrm{~J})=-3.23 \mathrm{kcal} / \mathrm{mol} \mathrm{H}{ }^{+}
\end{equation}

 

Electrochemical potential

Now, let's add the equations to get the full driving force, the electrochemical potential:

\begin{equation}
\Delta G_{\text {tot }}=2.303 \mathrm{RT} \Delta \mathrm{pH}+\mathrm{ZF} \Delta \psi=-5.15 \mathrm{kcal} / \mathrm{mol} \mathrm{H}{ }^{+}
\end{equation}

This gives

\begin{equation}
\frac{\Delta G_{\text {tot }}}{F}=\frac{2.303 R T \Delta p H++z F \Delta \psi}{F}
\end{equation}

Hence

\begin{equation}
\Delta p=\Delta \psi+\frac{2.303 R T \Delta p H}{F}=\Delta \psi+0.059 \Delta p H=0.224 \mathrm{~V}
\end{equation}

where Δp is the protonmotive force.

Summary

(Summary written by Claude, Sonnet 4.6, Anthropic)

This chapter describes the structure, mechanism, and thermodynamics of F₁Fo ATP synthase—the molecular machine that couples the collapse of the mitochondrial proton gradient to ATP synthesis — revealing one of the most elegant solutions in biochemistry to the problem of energy transduction.

ATP synthase (F₁F_o ATPase) is a reversible rotary motor enzyme embedded in the inner mitochondrial membrane. Its two major components are mechanically and functionally coupled: the membrane-embedded F_o subcomplex, which translocates protons down their electrochemical gradient and rotates in response, and the soluble F₁ subcomplex, which uses the mechanical energy of rotation to drive ATP synthesis. The turbine-driven electrical generator analogy is precise: Fo acts as a proton-powered electrical motor whose rotation drives the F₁ chemical motor to generate ATP as its "electrical" output.

The F₁ unit consists of an α₃β₃ hexameric ring with catalysis occurring at the three αβ dimer interfaces, plus a central γ subunit that spans the ring's central cavity. Paul Boyer deduced from isotope exchange experiments — before the structure was known — that the three αβ dimer pairs exist in three interconvertible conformational states: Open (O, no substrate bound, no catalytic activity), Loose (L, low-affinity substrate binding, no catalytic activity), and Tight (T, high-affinity substrate binding, catalytically active). The γ subunit, rotating like a crankshaft within the stationary α₃β₃ ring, drives sequential O→L→T→O conformational transitions in the three dimer pairs. Substrates (ADP and Pᵢ) bind at L, are condensed to ATP at T, and ATP is released at O — with the critical feature that Pᵢ binds before ADP in the synthetic direction, preventing ATP (present at up to 10-fold excess over ADP) from occupying the site. ATP synthesis thus occurs via a conformational driving mechanism, in which energy input changes binding-site affinity rather than directly catalyzing the phosphorylation chemistry. Glu190 on the β subunit acts as a general base in catalysis.

The evidence for γ rotation is definitive. Noji and colleagues attached a fluorescently labeled actin filament to the γ subunit of F₁ immobilized through a His-tag on a coverslip and observed unidirectional counterclockwise rotation at 0.5 revolutions per second — only in the presence of ATP. At limiting [ATP], rotation occurred in discrete 120° steps, one per ATP hydrolyzed. The mechanical work performed per step (~80 piconewton-nm) matched the free energy of ATP hydrolysis. Subsequent colloidal gold nanoparticle experiments confirmed these observations with less frictional drag. Even without the γ subunit, AFM studies showed that the β subunits undergo intrinsic sequential counterclockwise conformational changes during ATP hydrolysis, demonstrating that the α₃β₃ ring is inherently a rotary motor, with γ acting as a rotor that amplifies and coordinates the intrinsic motion.

Proton translocation through F_o depends on the structural interplay between the c-ring and the a-subunit. Each c-subunit monomer contains a critical carboxylate residue (Asp61 in E. coli, Glu58/59 in bovine/yeast) positioned near the center of the membrane bilayer. Its hydrophobic environment abnormally elevates the pKa of this residue, keeping it protonated (neutral) in the lipid bilayer. The a-subunit contains two aqueous half-channels: a lumenal channel (opening to the intermembrane space/cristae lumen) and a matrix channel (opening to the mitochondrial matrix), separated by a strictly conserved, positively charged arginine residue (Arg239 in bovine; Arg210 in E. coli) that acts as a seal preventing proton "short-circuit" directly between the two channels. A proton from the lumen enters through His and Glu residues lining the lumenal half-channel and protonates a c-subunit Glu/Asp near the a-subunit interface, neutralizing its charge. As the c-ring rotates, this protonated, neutral c-subunit moves through the hydrophobic lipid environment, eventually returning to the a-subunit interface — now on the matrix side — where its abnormally high pKa drops (given the aqueous, higher-pH matrix environment), it deprotonates, and the proton is released to the matrix. The electrostatic force exerted by the positively charged Arg239 on the deprotonated, negatively charged c-subunit Glu drives the net counterclockwise rotation of the c-ring. This rotation is transmitted through the γ/ε/δ axle to the F₁ unit, driving ATP synthesis.

The two classic F_o inhibitors illuminate this mechanism. Oligomycin A binds at the a-subunit/c-ring interface, blocking proton access to the c-ring Glu and halting rotation. Dicyclohexylcarbodiimide (DCCD) reacts covalently with the protonated, hydrophobic Asp/Glu residue on a single c-subunit, permanently blocking that subunit from accepting protons and halting c-ring rotation — since one blocked subunit prevents rotation of the entire ring.

The thermodynamic basis for ATP synthesis is rigorously quantified by the proton-motive force (∆p). The total free energy available for proton translocation across the inner mitochondrial membrane has two components: a chemical potential contribution from the pH gradient (ΔG_chem = 2.303RT·ΔpH = +8.0 kJ/mol per proton at ΔpH = −1.4) and an electrical potential contribution from the membrane potential (ΔG_elec = ZFΔψ = +13.5 kJ/mol per proton at Δψ = −0.14 V). The combined electrochemical driving force provides approximately −21.5 kJ/mol per proton translocated. With approximately 10–14 c-subunits per c-ring and three ATP synthesized per revolution, the free energy of approximately 3.5 protons per ATP provides more than sufficient thermodynamic force (~75 kJ/mol) to drive the endergonic synthesis of ATP (ΔG°′ ≈ +30 kJ/mol, with actual in vivo ΔG higher), with the coupling efficiency approaching the thermodynamic limit. The equivalent local electric field between the two aqueous half-channels — approaching 40–100 million V/m — is ten- to thirty-fold higher than the field intensity just before a lightning strike, illustrating the extraordinary energetic environment within this nanometer-scale molecular machine.