Dr. Stanley Dunn

Distinguished University Professor, Adjunct Professor, and Emeritus Professor
B.Sc. University of Nebraska
Ph.D. University of California
Postdoctoral Fellow, Cornell University
Office: Medical Sciences Building, Rm. 324
Phone: 519.661.3055
E-mail: sdunn@uwo.ca

Molecular Mechanics of ATP Synthase

ATP synthase occupies a central position in bioenergetics, since it couples the protonmotive across a membrane to the synthesis/hydrolysis of ATP in the processes of oxidative phosphorylation and photophosphorylation. Because of its remarkable rotational mechanism it has been called the world's smallest motor. Discovering how the structural features of ATP synthase enable this function is the focus of our research program. ATP synthase is sometimes viewed in mechanical terms as a pair of motors linked through a common, central driveshaft and a peripheral stator stalk.

The membrane-embedded F0 sector couples rotation of a ring of c subunits to the movement of protons across the membrane; their path involves the handoff of protons between the a and c subunits. The peripheral F1 sector couples rotation of a γε subcomplex to the synthesis or hydrolysis of ATP. In this case conformational changes in the three active sites at αβ interfaces are linked to rotation of an asymmetrical coiled coil of γ (not seen in the diagram) that extends up the centre of the α3β3 hexamer like a spindle. Noncovalent interactions link γε to the c ring, creating a common driveshaft, so that either motor can drive the other in reverse. For the direction of rotation shown, transmembrane H+ movement down a proton potential gradient drives the synthesis of ATP from ADP and phosphate. However, the reaction is reversible, and ATP hydrolysis by F1 can drive the movement of protons in the opposite direction, generating or enhancing protonmotive force. This reaction is linked to rotation of the γεc10 rotor in the opposite direction. Living organisms have both a high protonmotive force and a high energy of ATP hydrolysis, so in vivo the two motors are constantly trying to turn the rotor in opposite directions. A device to ensure that the other subunits of either F1 or F0 do not simply turn with the rotor is essential to the mechanism; this is one function of the peripheral stator stalk. Thus, the two linkages between the distant motors are essential to coupling energy transduction between the protonmotive force and ATP, and interest is increasingly focussed on these parts of the complex.

Despite the usefulness of the mechanical model, ATP synthase is made of protein rather than hardened metal, and we know that efficient energy transduction requires both elastic distortion of the elements and other factors that are not yet understood. The stator stalk features a novel, but incompletely characterized, right-handed coiled coil formed by the two b subunits, while the ε subunit has a recently discovered, potentially regulatory, ATP binding site and undergoes dramatic conformational changes. Mutations to either b or ε can dramatically affect energy coupling, even though their function in linking the motors is uncompromised. These findings show that the functions of the two stalks go well beyond simply holding the complex together.

The  b2δ "stator stalk"

The 156-residue b subunit of Fo has a hydrophobic N-terminal membrane spanning domain while the remainder of the polypeptide is predominantly polar in nature, and is essential for binding F1 to F0. ATP synthase from some species-notably photosynthetic or pathogenic bacteria-contain two distinct, but related, b-type subunits, called b and b', present as a heterodimer, but a homodimer of identical b subunits is seen in the prototypical complex from E. coli and in many other species. After removing the sequence encoding the first 25 residues of that protein we expressed the remainder as a soluble protein and characterized it as a highly extended, helical dimer. An interaction with the δ subunit of F1 was identified and characterized by analytical ultracentrifugation. Further work has shown that the very C-terminal residues of b are essential for binding to δ and hence to F1, identifying the C-terminal part of the polypeptide as the δ- or F1-binding domain. Other work showed that residues 53-122 provide the interactions essential for dimerization, leading to the domain model shown below, where residues between the membrane-spanning domain and the dimerization domain are labelled the "tether" domain. Little is known about the structure or function of this region, but it contains an essential, conserved residue Arg-36 and is likely associated with polar loops of the a subunit.

Multiple sequence alignments of b-type subunits reveal an unusual 11-residue (hendecad) repeat pattern in the dimerization domain. The positions called a and h are usually occupied by small residues, especially alanine, while the d and e positions are often filled by larger hydrophobic residues. Most other positions are usually occupied by polar or charged residues.

An alignment of dimerization domain sequences from 16 b and b' squences, and analysis of the hydrophobic periodicity by Fourier transformation. Note the peak periodicity at 3.67 (3/11), indicating 3 turns of the helix relative to the hydrophobic face per 11 residues.

The a, d, e, and h positions should form one surface on a helix (right), and this has been seen in a crystal structure of a 62-122 fragment (left). Since this alanine-rich hydrophobic strip veers slightly to the right as you move up the helix, dimerization about this interface would form a novel parallel two stranded right-handed coiled coil.

No high resolution structure of the dimeric form is yet available, but disulfide formation studies have confirmed that the a, d, e and h positions do lie at the helix-helix interface, and furthermore imply that the helices are offset by about 1.5 turns of the helix (right), rather than being in-register, as they are in a left-handed coiled coil. This makes b2 intrinsically asymmetric. We designate the two positions as bN, to denote the subunit that is offset N-terminally, and bC.

The asymmetry of b2 means that the two b subunits will interact differently with α3β3δ and ac10. Our initial studies to test this (Wood and Dunn, 2007, JBC) using mutations at the C-termini imply that bN is more directly involved in the interaction with δ and plays a larger role in binding F1, compared to bC.

The γε "rotor"

Recently we fused a series of protein domains of different sizes onto the C-terminus of the ε subunit. As the rotor gets larger, some point should be reached at which it is unable to pass inside the b2δ stator, and rotation should be sterically blocked. Strains carrying a fusion of the 12-kDa cytochrome b retained growth on acetate media, indicating functional oxidative phosphorylation, while fusions of 18 kDa (flavodoxin) or larger could not grow on acetate despite normal assembly of the ATP synthase. These results provide the first evidence for operation of the rotational mechanism in vivo. Interestingly, membranes containing ATP synthase with the larger fusions exhibited an ATPase activity that was uncoupled from proton pumping, implying that the C-terminus of ε is essential for maintaining efficient coupling in the complex. We have previously shown that this part of ε interacts with a specific site on one of the catalytic β subunits. Further studies of rotation are underway, making use of chemical cross-linking and fluorescent resonance energy transfer (FRET) techniques.

How do the rotor and stator function?

In order to store the energy released by the 3-4 protons that move sequentially across the membrane for each ATP that is produced, the rotor and stator must both be capable of elastic deformation, allowing them to store energy like springs. The large globular domain of γ that sits between α3β3 and c10 appears to be the major flexible element in the rotor. Whether ε also contributes to energy storage, or acts primarily to monitor ATP abundance and/or catalytic nucleotide binding site occupancy, is uncertain. The role of ε in regulation and coupling remains controversial.

We postulate that the unusual, right-handed helical interaction between the b subunits was selected by nature because this relationship makes the stator stalk better able to resist the torque imposed on F1 by the turning of the rotor. When the two motors push against each other, the left-handed coiled coil of γ in the rotor and the right-handed coiled coil of b2 in the stator stalk will both become wound more tightly (see diagram below), consistent with elastic storage of energy, whereas coiled coils of the opposite turns would be loosened, preventing elastic storage of energy. Consistent with this idea, replacing part of the hendecad sequence of b with well-characterized left-handed coiled coil segments, to produce chimeric subunits, results in an uncoupled form of ATP synthase (Bi et al., 2008). In contrast, sequences from unrelated proteins showing the characteristic 11-residue repeat pattern can be inserted to make functional chimeras, and even functional heterodimeric b chimeras have been produced. We believe that these will be particularly useful for elucidating roles of bN and bC.

Functional relevance of right-handedness in the stator stalk by structural stabilization to torque: A thought experiment where either ATP hydrolysis (left) or else proton flow (right) is permitted, but not both. Operation of either motor tightens both the left-handed coiled coil of γ and the right-handed coiled coil of b2.

Current and future studies

In order to obtain direct evidence of how the rotational mechanism operates so efficiently, it is essential to better understand the structure of the stator, the interactions of the rotor and stator with α3β3 and to learn how these interactions change during the catalytic cycle. Studies in our laboratory, supported by the Canadian Institutes of Health Research, are designed to provide this information so that the molecular mechanism of this extraordinarily important and complex enzyme may be understood. Our approaches include bioinformatics, site-directed mutagenesis, deletion analysis, chemical cross-linking, engineered disulfide bonds, fusion proteins, biophysical analysis, and collaborative high resolution structural analysis.


See complete list of publications from PubMed.