Voet D., Voet Ju.G. Biochemistry

Подождите немного. Документ загружается.

brings Asp 85 closer to the Schiff base N atom. The depro-

tonated retinal straightens and, in doing so, moves upward

(toward the inside) by 0.7 to 1.0 Å. It thereby pushes

against the F helix, causing its inside (cytoplasmic) end to

tilt outward from the channel by ⬃3.5 Å and the G helix to

partially replace it.

3. The movement of the F helix opens up the central

channel on the inside of the membrane, admitting several

water molecules that form a hydrogen bonded chain be-

tween Asp 96 and the Schiff base. One of these water mol-

ecules hydrogen bonds with Asp 96 so as to lower its pK.

This permits Asp 96 to protonate the Schiff base via the in-

termediacy of the chain of hydrogen bonded water mole-

cules, thereby yielding the N state.

4. Asp 96 is reprotonated by the cytoplasmic solution.

The loops forming bacteriorhodopsin’s inside surface bear

numerous charged residues and hence it appears that they

function as “antennas” to capture protons from the alka-

line cytoplasmic medium.Asp 85 transfers its proton to the

extracellular medium via a hydrogen bonded network that

includes several bound water molecules. This process is fa-

cilitated by a preceding 1.6-Å displacement of the Arg 82

side chain toward a complex of residues that includes Glu

194 and Glu 204, which reduces the pK of this complex.The

retinal then relaxes, via the O state, to its original all-trans

form and helices F and G return to their original positions,

thereby re-forming the ground state of the protein and

completing the catalytic cycle.

The retinal, which occupies the center of the protein chan-

nel, thereby acts as a one-way proton valve. The vectorial

nature of this process arises from the unidirectional series

of conformational changes made by the photoexcited reti-

nal as it relaxes to its ground state. The protein’s principal

proton pumping motions are remarkably small, involving

group movements of ⬃1 Å or less in response to the light-

induced flexing of the retinal.These,nevertheless, cause pK

changes in various residues that facilitate proton transfer as

well as making and breaking hydrogen bonded networks of

protein groups and water molecules in the proper sequence

to transport a proton.A similar mechanism is likely to oper-

ate in COX, but with its conformational changes motivated

by redox reactions rather than by photoexcitation.

i. COX Has Two Proton Translocation Channels

Two channels that are candidates for translocating pro-

tons from the inside to the vicinity of the O

-reducing cen-

ter have been described in bovine, P. denitrificans, and Rb.

sphaeroides COX (Fig. 22-35). These channels, which are

both contained in Subunit I, are named the K- and

D-channels after their respective key residues (K319 and

D91 in the bovine numbering scheme, which we shall use in

the subsequent discussion). Both putative channels are

similar in character to that present in bacteriorhodopsin in

that they consist of chains of hydrogen bonded and poten-

tially hydrogen bonded protein groups, bound water mole-

cules, and water-filled cavities.

The K-channel leads from K319, which is exposed to the

inside, to Y244, the residue that is the proposed substrate

electron and proton donor in the reaction forming the

P state (Step 4 of Fig. 22-27). The K319M mutant has an

extremely low activity (⬍0.05% of wild type), which is

not increased by supplying additional protons from the out-

side. Hence it appears that the K-channel is not connected

to the putative exit channel (Fig. 22-35) that leads to the

outside. It therefore seems likely that the K-channel only

functions to supply chemical protons to the O

-reduction

center.

The entrance to the D-channel is within a region on the

protein surface that appears likely to act as a proton-

gathering antenna. The mutation of D91 to any noncar-

boxylate residue eliminates proton pumping but reduces

the rate of O

reduction to only 45% of wild type (in

Section 22-3. Oxidative Phosphorylation 851

Figure 22-34 Proton pump of bacteriorhodopsin. See the text

for a description of this mechanism.The protein is represented

by its seven transmembrane helices,A through G (with helices D

and E omitted for clarity in all but the upper left panel), and

several mechanistically important side chains.The retinal is

drawn with the approximate color of the complex in its various

spectroscopically characterized states. Red arrows indicate

proton movements, blue arrows indicate the movements of

groups of atoms, and the “paddle” attached to helix F represents

the bulky side chains that must move aside to open the

cytoplasmic channel. [After Kühlbrandt,W., Nature 406, 569 (2000).]

Ground state

all trans retinal protonated

N intermediate

13-cis retinal protonated

Inside

Outside

Membrane

light

L intermediate

13-cis retinal protonated

Late M intermediate

13-cis retinal neutral

–

Asp 96

Asp 85

Arg 82

JWCL281_c22_823-870.qxd 7/2/10 11:11 AM Page 851

E. coli). Evidently, the D-channel, in series with the exit

channel, is the proton pumping channel. Moreover, the

D-channel, which extends to the vicinity of the heme

–Cu

binuclear center, is also the conduit for the chemi-

cal protons required for the second part of the reaction

cycle (Steps 5 and 6 of Fig. 22-27).

What is the mechanism that couples O

reduction to

proton pumping in COX? Unfortunately, X-ray crystallog-

raphy has provided little guidance in answering this ques-

tion because the X-ray structures of only a few different

states of COX have as yet been determined and because

the resolution of several of these structures is too low to re-

liably reveal small structural differences between them.

Consequently, the mechanisms that have been proposed to

explain how COX pumps protons are largely inferences

based on limited structural information, interpretation of

site-directed mutagenesis experiments, spectroscopic data,

theoretical considerations, and chemical intuition. Several

ingenious although largely phenomenological models of

how COX pumps protons have been proposed.These mod-

els agree that at least one proton is pumped during each of

Steps 5 and 6 in the COX reaction cycle (Fig. 22-27) but dis-

agree as to where in the reaction cycle the remaining two

protons are pumped and how the protein actuates this

process. Clearly there is still much to learn about how COX

carries out its function.

j. The Horizontal Helix of Complex I’s Peripheral

Arm Functions Like a Piston.

Complex I pumps four protons out of the matrix for

every electron pair it translocates from NADH to CoQ.

One of these protons appears to be transported at the in-

terface between the peripheral and transmembrane arms

of Complex I (Fig. 2-18) in a way that presumably is driven

by the conformational changes in the peripheral arm as it

translocates electrons from NADH to CoQ. But how are

these conformational changes coupled to the far more dis-

tant antiporter-like subunits, Nqo12, 13, and 14, which pre-

sumably pump one proton each out of the matrix? The low

resolution of the Complex I X-ray structure (Fig. 22-18)

precludes the visualization of the membrane arm’s side

chains and hence a detailed mechanism for proton trans-

port. However, the unusual 110-Å-long horizontal helix

that appears to link the antiporter-like subunits to the con-

formational changes in the peripheral arm has led Sazanov

to postulate that this helix functions like a piston to couple

conformational changes in the peripheral arm to those in

the antiporter-like channels.

C. Mechanism of ATP Synthesis

See Guided Exploration 21: F

–ATP synthase and the binding

change mechanism

The proton-motive force across the mito-

chondrial membrane is harnessed in the synthesis of ATP

by proton-translocating ATP synthase (also known as

–ATPase, Complex V, and F-type H

ⴙ

–ATPase). In the

following subsections we discuss the location and structure

of this ATP synthase and the mechanism by which it har-

nesses proton flux to drive ATP synthesis.

a. Proton-Translocating ATP Synthase Is a

Multisubunit Transmembrane Protein

Proton-translocating ATP synthase consists of two ma-

jor substructures comprising 8 to 13 different subunits.

Electron micrographs of mitochondria (Fig. 22-36) show

lollipop-shaped structures studding the matrix surface of

the inner mitochondrial membrane (Fig. 22-36a). Similar

entities have been observed lining the inner surface of the

bacterial plasma membrane and in chloroplasts (Section

24-2Da). Sonication of the inner mitochondrial membrane

yields sealed vesicles, submitochondrial particles, from

which the “lollipops” project (Fig. 22-36b) and which can

carry out ATP synthesis.

Efraim Racker discovered that the proton-translocating

ATP synthase from submitochondrial particles comprises

two functional units, F

and F

is a water-insoluble trans-

membrane protein composed of as many as eight different

types of subunits (although only three in E. coli) that con-

tains a proton translocation channel. F

is a water-soluble

852 Chapter 22. Electron Transport and Oxidative Phosphorylation

Outside

Inside

exit channel

heme a

copper B

hydrophobic

cavity

D-channel

K-channel

Asp 369

Asp 364

Arg 439

Arg 438

His 368

His 291

Tyr 244

Thr 316

Lys 319

Ser 255

Ser 157

Glu 242

Ser 156

Asp 91

Figure 22-35 The proton-translocating channels in bovine

COX. The enzyme is viewed parallel to the membrane with the

matrix below. The four rectangles delineate the proposed proton

entry and exit conduits. Single and double circles, respectively,

represent water molecules that are observed in X-ray structures

or that theoretical methods suggest are likely to be present.

[After a drawing by Mårten Wikström, University of Helsinki,

Helsinki, Finland.]

JWCL281_c22_823-870.qxd 7/5/10 10:27 AM Page 852

peripheral membrane protein, composed of five types of

subunits, that is easily dissociated from F

by treatment

with urea. Solubilized F

can hydrolyze ATP but cannot

synthesize it (hence the name ATPase). Submitochondrial

particles from which F

has been removed by urea treat-

ment no longer exhibit the lollipops in their electron mi-

crographs (Fig. 22-36c) and lack the ability to synthesize

ATP. If, however, F

is added back to these F

-containing

submitochondrial particles, their ability to synthesize ATP

is restored and their electron micrographs again exhibit the

lollipops. Thus the lollipops are the F

particles. Higher res-

olution cryoelectron micrograph–based images of the ATP

synthase from bovine heart mitochondria reveal that its F

and F

components are joined by both an ⬃50-Å-long cen-

tral stalk and a less substantial peripheral stalk (Fig. 22-37).

Certain bacterial F

–ATPases translocate Na

⫹

ions

rather than protons.

Section 22-3. Oxidative Phosphorylation 853

Figure 22-37 Cryoelectron microscopy–based image of

–ATPase from bovine heart mitochondria. [Courtesy of John

Rubinstein, University of Toronto, Canada, and John Walker and

Richard Henderson, MRC Laboratory of Molecular Biology,

Cambridge, U.K.]

Figure 22-36 Electron micrographs and interpretive drawings

of the mitochondrial membrane at various stages of dissection.

(a) Cristae from intact mitochondria showing their F

“lollipops”

projecting into the matrix. [From Parsons, D.F., Science 140, 985

Advancement of Science. Used by permission.]

(b) Submitochondrial particles, showing their outwardly

projecting F

lollipops. Submitochondrial particles are prepared

by the sonication (ultrasonic disruption) of inner mitochondrial

membranes. [Courtesy of Peter Hinkle, Cornell University.]

[Courtesy of Efraim Racker, Cornell University.]

particles

Outer

membrane

Matrix

Inner

membrane

sonication

urea

reconstitution

particles F

particles

Sonication

cleavage lines

Mitochondrion

Submitochondrial

particle

Membranous

vesicles

(a)

(b)

(c)

JWCL281_c22_823-870.qxd 7/2/10 11:11 AM Page 853

b. The X-Ray Structure of F

Reveals the

Basis of Its Lollipoplike Structure

Mitochondrial F

is an ␣

␤

␥␦ε nonamer in which the ␤

subunits contain the catalytic sites for ATP synthesis and

the ␦ subunit is required for binding of F

to F

. The X-ray

structure of F

from bovine heart mitochondria, deter-

mined by John Walker and Andrew Leslie, reveals that this

⬃400-kD protein is a 100-Å-high and 100-Å-wide spheroid

that is mounted on a 50-Å-long stem (Fig. 22-38a,b). F

’s ␣

and ␤ subunits (553 and 528 residues), which are 20% iden-

tical in sequence and have nearly identical folds, are

arranged alternately, like the segments of an orange, about

the upper portion of a 114-Å-long, 74-residue ␣ helix

formed by the ␥ subunit (298 residues). The C-terminus of

this helix protrudes into a 15-Å-deep dimple that is cen-

trally located at the top of the spheroid. The lower half of

854 Chapter 22. Electron Transport and Oxidative Phosphorylation

Figure 22-38 X-ray structures of F

–ATPase from bovine

heart mitochondria. (a) F

–ATPase in complex with ADP and

ATP drawn as a semitransparent ribbon diagram viewed parallel

to the membrane with the matrix above. The ␣, ␤, ␥, ␦, and ε

subunits are pink, yellow, blue, orange, and green, respectively,

and the nucleotides are shown in space-filling form with ATP C

green,ADP C cyan, N blue, O red, and P orange. (b) The

structure in Part a rotated 90° about the horizontal axis such that

the protein is viewed along its pseudo-3-fold axis from the

matrix.The ␦ and ε subunits have been deleted for clarity. Note

that although the ␣

and ␤

subunits both normally bind ATP,

in this structure they have bound ADP. (c) A surface diagram of

the inner portion of the ␣

␤

assembly through which the

C-terminal helix of the ␥ subunit penetrates as viewed from the

matrix.The surface is colored according to its electrical potential,

with positive potentials blue, negative potentials red, and neutral

white. Note the absence of charge on the inner surface of this

sleeve. That part of the ␥ subunit’s C-terminal helix which

contacts this sleeve is similarly devoid of charge. [Parts a and b

based on an X-ray structure by Andrew Leslie and John Walker,

MRC Laboratory of Molecular Biology, Cambridge, U.K. PDBid

1E79. Part c from Abrahams, J.P., Leslie,A.G.W., Lutter, R., and

Walker, J.E., Nature 370, 621 (1994). PDBid 1BMF.]

See Interactive Exercise 19

(a)

(b)

(c)

JWCL281_c22_823-870.qxd 10/19/10 8:08 AM Page 854

the helix forms a bent left-handed antiparallel coiled coil

with the N-terminal segment of the ␥ subunit. This coiled

coil forms much of the ⬃50-Å-long central stalk seen in

cryoEM-based images of the F

–ATPase such as Fig.22-37.

The cyclical arrangement and structural similarities of

’s ␣ and ␤ subunits gives it both pseudo-3-fold and

pseudo-6-fold rotational symmetry. Nevertheless, the pro-

tein is asymmetric.This is in part due to the presence of the

␥, ␦, and ε subunits but, more importantly, because each of

the ␣ and ␤ subunits have a somewhat different conforma-

tion.Thus, one ␤ subunit (designated ␤

) normally binds a

molecule of ATP, the second (␤

) binds ADP, and the

third (␤

) has an empty and distorted binding site. The ␣

subunits all normally bind ATP, although they also differ

conformationally from one another (however, ␣

and ␤

both bind ADP in Fig. 22-38a,b). The ATP and ADP bind-

ing sites each lie at a radius of ⬃20 Å near an interface be-

tween adjacent ␣ and ␤ subunits and, in fact, all incorpo-

rate a few residues from the adjacent subunit. The ␣ and ␤

subunits both contain two sequence motifs, the Walker A

motif (GXXXXGKT/S, where X is any residue) and the

Walker B motif (R/KXXXGXXL/VhhD, where h is a hy-

drophobic residue), that participate in ATP binding and

which occur widely in nucleotide-binding proteins of all

descriptions.

The ␦ and ε subunits (168 and 51 residues) are wrapped

about the base of the ␥ subunit’s coiled coil. The X-ray

structure of E. coli F

resembles its bovine counterpart.

Note, however, that in an unfortunate confusion of nomen-

clature, the E. coli ε subunit is the homolog of the mito-

chondrial ␦ subunit; the E. coli ␦ subunit is the counterpart

of the mitochondrial oligomycin-sensitivity conferral pro-

tein (OSCP; see Problem 11), and the mitochondrial ε sub-

unit has no counterpart in either bacterial or chloroplast

ATP synthases.

c. The c Subunits of F

Form a Transmembrane Ring

The F

component of the F

–ATPase from E. coli con-

sists of three transmembrane subunits, a, b, and c (271, 161,

and 89 residues) that form an a

complex. However,

the number c subunits per F

assembly varies from 10 to 15,

depending on the species. Mitochondrial F

additionally

contains one copy each of three different subunits, d, F

and OSCP (256, 108, and 213 residues), as well as several

“minor” subunits, e, f, g, and A6L, of unknown function.

Subunits a and A6L are encoded by mitochondrial genes.

A variety of evidence indicates that the hydrophobic c sub-

units associate to form a ring with the ab

unit located at its

periphery (see below). The sequence of the a subunit sug-

gests that this highly hydrophobic peptide forms five trans-

membrane helices.

The X-ray structure of the c

unit from the Na

⫹

translocating F

–ATPase of the gram-negative bac-

terium Ilyobacter tartaricus, determined by Peter Dim-

roth, reveals an 11-fold symmetric cylindrical assembly

that is 70 Å high and ⬃50 Å in diameter (Fig. 22-39). Each

of its identical, largely hydrophobic subunits consists al-

most entirely of an inner helix that defines the length of

the cylinder and an outer helix that is somewhat shorter.

Both helices are bent in the vicinity of the Na

⫹

ion-binding

Section 22-3. Oxidative Phosphorylation 855

Figure 22-39 X-ray structure of the c

assembly of the I.

tartaricus Na

⫹

-translocating F

–ATPase. (a) Ribbon diagram

as viewed along the plane of the plasma membrane with the

cytosol above. Each subunit is colored in rainbow order from its

N-terminus (blue) to its C-terminus (red). Bound Na

⫹

ions are

represented by purple spheres.The parallel lines delineate the

inferred position of the plasma membrane. (b) View from the

cytosol parallel to the assembly’s 11-fold axis in which each

subunit is drawn in worm form and colored as in Part a.The

minimally ⬃20-Å-diameter central channel presumably contains

a lipid bilayer. [Based on an X-ray structure by Peter Dimroth,

ETH, Zürich, Switzerland. PDBid 1YCE.]

(a)

(b)

JWCL281_c22_823-870.qxd 7/2/10 11:11 AM Page 855

site, thereby conferring an hourglasslike shape on the c

assembly (Fig. 22-39a). Each Na

⫹

ion is liganded by

residues from both helices of a given subunit as well as

those of the clockwise neighboring outer helix as seen in

Fig. 22-29b.

In vertebrates, the peripheral stalk (Fig. 22-37) consists

of the OSCP, b, d, and F

subunits, whereas in most

prokaryotes, it consists of the OSCP homolog ␦ and two

copies of b. The X-ray structure of most of the cytosolic

portion of a complex of bovine subunits b, d, and F

, con-

stituting 54% of the extramembrane segment of the pe-

ripheral stalk, determined by Leslie and Walker, reveals

that this b fragment forms a 160-Å-long and curved ␣ he-

lix, and that the other subunit fragments are also mainly

helical (Fig. 22-40). The curvature of this assembly closely

matches that of the peripheral stalk in cryoEM-based im-

ages of the F

–ATPase (e.g., Fig. 22-37). Such cryoEM-

based images, together with the foregoing X-ray structures

and the NMR structures of ␦ and the transmembrane seg-

ment of b, both from E. coli, has enabled the construction

of a composite model of the E. coli F

–ATPase (Fig. 22-

41), which, of course, resembles the mitochondrial assem-

bly. Note the extensive contacts between ␥ and ε and the

top of the c cylinder.

d. The Binding Change Mechanism:

Proton-Translocating ATP Synthase Is

Driven by Conformational Changes

The mechanism of ATP synthesis by proton-translocating

ATP synthase can be conceptually broken down into three

phases:

1. Translocation of protons carried out by F

2. Catalysis of formation of the phosphoanhydride

bond of ATP carried out by F

3. Coupling of the dissipation of the proton gradient

with ATP synthesis, which requires interaction of F

and F

The available evidence supports a mechanism for ATP

formation, proposed by Boyer, that resembles the confor-

mational coupling hypothesis of oxidative phosphorylation

(Section 22-3A). However, the conformational changes in

the ATP synthase that power ATP formation are generated

by proton translocation rather than by direct electron

transfer,as proposed in the original formulation of the con-

formational coupling hypothesis.

is proposed to have three interacting catalytic pro-

tomers, each in a different conformational state: one that

binds substrates and products loosely (L state), one that

binds them tightly (T state), and one that does not bind

them at all (open or O state). The free energy released on

proton translocation is harnessed to interconvert these

three states.The phosphoanhydride bond of ATP is synthe-

sized only in the T state and ATP is released only in the O

state. The reaction involves three steps (Fig. 22-42):

1. Binding of ADP and P

to the “loose” (L) binding

site.

2. A free energy–driven conformational change that

converts the L site to a “tight” (T) binding site that cat-

alyzes the formation of ATP.This step also involves confor-

mational changes of the other two subunits that convert

856 Chapter 22. Electron Transport and Oxidative Phosphorylation

Figure 22-40 X-ray structure of a portion of the peripheral

stalk of bovine F

–ATPase. This protein fragment is drawn in

ribbon form embedded in its semitransparent molecular surface,

with b (residues 79–183) magenta, d (residues 3–123) cyan, and

(residues 5–70) orange. The N- and C-termini of each subunit

are indicated. [Based on an X-ray structure by Andrew Leslie

and John Walker, MRC Laboratory of Molecular Biology,

Cambridge, U.K. PDBid 2CLY.]

Figure 22-41 A composite model of the E. coli F

–ATPase.

This model is based on the X-ray structure of the E. coli F

subunit (PDBid 1JNV), which resembles that of bovine F

(Fig. 22-38), the structures displayed in Figs. 22-39 and 22-40, and

the NMR structures of ␦ and the transmembrane segment of the

b from E. coli (PDBids 2A7U and 1B9U). The structures of a

and the so-called hinge region of b are unknown. [Courtesy of

Peter Dimroth, ETH, Zürich, Switzerland.]

JWCL281_c22_823-870.qxd 7/2/10 11:11 AM Page 856

the ATP-containing T site to an “open” (O) site and con-

vert the O site to an L site.

3. ATP is synthesized at the T site on one subunit while

ATP dissociates from the O site on another subunit. On the

surface of the active site, the formation of ATP from ADP

and P

entails little free energy change, that is, the reaction

is essentially at equilibrium. Consequently, the free energy

supplied by the proton flow primarily facilitates the release

of the newly synthesized ATP from the enzyme; that is, it

drives the T S O transition, thereby disrupting the

enzyme–ATP interactions that had previously promoted

the spontaneous formation of ATP from ADP ⫹ P

in the

T site.

How is the free energy of proton transfer coupled to

the synthesis of ATP? Boyer proposed that the binding

changes are driven by the rotation of the catalytic assembly,

␣

␤

with respect to other portions of the F

–ATPase.This

hypothesis is supported by the X-ray structure of F

.Thus,

the closely fitting nearly circular arrangement of the ␣

and ␤ subunits’ inner surface about the ␥ subunit’s helical

C-terminus is reminiscent of a cylindrical bearing rotating

in a sleeve (Fig. 22-38c). Indeed, the contacting hydropho-

bic surfaces in this assembly are devoid of the hydrogen

bonding and ionic interactions that would interfere with

their free rotation; that is, the bearing and sleeve appear

to be “lubricated.” Moreover, the central cavity in the

␣

␤

assembly (Fig. 22-38a) would permit the passage of

the ␥ subunit’s N-terminal helix within the core of this

particle during rotation. Finally, the conformational dif-

ferences between F

’s three catalytic sites appear to be

correlated with the rotational position of the ␥ subunit.

Apparently the ␥ subunit, which is thought to rotate within

the fixed ␣

␤

assembly, acts as a molecular cam shaft in

linking the proton-motive force–driven rotational motor to

the conformational changes in the catalytic sites of F

.This

concept is also supported by molecular dynamics simula-

tions (Section 9-4) by Leslie, Walker, and Martin Karplus,

which indicate that the conformational changes in the ␤

subunits arise from both steric and electrostatic interac-

tions with the rotating ␥ subunit.

Rotating assemblies are not unprecedented in biologi-

cal systems. Bacterial flagella, which function as propellers,

had previously been shown to be membrane-mounted ro-

tary engines that are driven by the discharge of a proton

gradient (Section 35-3Ib).

e. The F

–ATPase Is a Rotary Engine

In the F

–ATPase, the rotor is proposed to be an as-

sembly of the c ring and its associated ␥ and (E. coli) ε sub-

units, whereas the ab

unit and the (E. coli) ␦ subunit to-

gether with the ␣

␤

spheroid form the stator (Fig. 22-41).

The rotation of the c ring in the membrane relative to the

stationary a subunit is driven by the migration of protons

from the outside to the inside, as we discuss below. The pe-

ripheral arm (b

␦) presumably functions to hold the ␣

␤

spheroid in place while the ␥ subunit rotates inside it.

The rotation of the E. coli ␥ε–c-ring rotor with respect

to the ab

␦–␣

␤

stator has been ingeniously demonstrated

by Masamitsu Futai using techniques developed by

Kazuhiko Kinosita Jr. and Masasuke Yoshida (Fig. 22-43a).

The ␣

␤

spheroid of E. coli F

–ATPase was fixed, head

down, to a glass surface as follows: Six consecutive His

residues (a so-called His Tag; Section 6-3Dg) were muta-

genically appended to the N-terminus of the ␣ subunit,which

is located at the top of the ␣

␤

spheroid as it is drawn in

Fig. 22-38a. The His-tagged assembly was applied to a glass

surface coated with horseradish peroxidase (which, like most

proteins, sticks to glass) conjugated with Ni

2⫹

-nitriloacetic

acid [N(CH

COOH)

, which tightly binds His tags], thereby

binding the F

–ATPase with its F

side facing away from

the surface. The Glu 2 residues of this assembly’s c sub-

units, which are located on the side of the c ring facing away

from F

, had been mutagenically replaced by Cys residues,

which were then covalently linked to biotin (a coenzyme

that normally participates in carboxylation reactions; Sec-

tion 23-1Ab). A fluorescently labeled and biotinylated

(at one end) filament of the muscle protein actin (Section

35-3Ac) was then attached to the c subunit through the ad-

dition of a bridging molecule of streptavidin, a protein that

avidly binds biotin to each of four binding sites (Cys 193 of

Section 22-3. Oxidative Phosphorylation 857

Figure 22-42 Energy-dependent binding change mechanism

for ATP synthesis by proton-translocating ATP synthase. F

has

three chemically identical but conformationally distinct

interacting ␣␤ protomers: O, the open conformation, has very

low affinity for ligands and is catalytically inactive; L has loose

binding for ligands and is catalytically inactive;T has tight

binding for ligands and is catalytically active. ATP synthesis

occurs in three steps. (1) Binding of ADP and P

to site L.

(2) Energy-dependent conformational change converting

ADP + P

Energy

ATP H

ADP • P

ATP

T O

O L

ATP

O L

binding site L to T,T to O, and O to L. (3) Synthesis of ATP at

site T and release of ATP from site O. The enzyme returns to its

initial state after two more passes of this reaction sequence. The

energy that drives the conformational change is apparently

transmitted to the catalytic ␣

␤

assembly via the rotation of the

␥ε assembly (in E. coli; ␥␦ε in mitochondria), here represented

by the centrally located asymmetric pointer (green). [After Cross,

R.L., Annu. Rev. Biochem. 50, 687 (1980).]

See the Animated

Figures

JWCL281_c22_823-870.qxd 7/20/10 6:25 PM Page 857

the ␥ subunit, the only other Cys residue in the rotor, was

mutagenically replaced by Ala to prevent it from being

linked to an actin filament).

E. coli F

–ATPase can work in reverse, that is, it can

pump protons from the inside (cytoplasm) to the outside

(periplasm) at the expense of ATP hydrolysis (this enables

the bacterium to maintain its proton gradient under anaer-

obic conditions, which it uses to drive various processes

such as flagellar rotation). Thus, the foregoing preparation

was observed under a fluorescence microscope as a 5 mM

MgATP solution was infused over it. Many of the actin fila-

ments were seen to rotate (Fig. 22-43b), and always in a

counterclockwise direction when viewed looking down on

the glass surface (from the outside). This would permit the ␥

subunit to sequentially interact with the ␤ subunits in the

direction

(Figs. 22-38b and 22-42), the direction expected for ATP

hydrolysis.

␤

(O state)

␤

(L state)

␤

(T state)

In a variation of the above experiment, the ␥ subunit of

the ␣

␤

␥ complex was directly cross-linked, via its Cys 193,

to a fluorescently labeled actin filament and, in this case,

the ␤ subunits were immobilized by appended His tags. At

very low ATP concentrations (e.g., 0.02 ␮M), video images

(Fig. 22-44) revealed that the fluorescent actin filament ro-

tated counterclockwise in discrete steps of 120°, as the

binding change mechanism predicts. Moreover, the calcu-

lated frictional work done in each rotational step is very

nearly equal to the energy available from the hydrolysis of

one ATP molecule, that is, the F

–ATPase converts chem-

ical to mechanical energy with nearly 100% efficiency.

The foregoing system also works in reverse. An ⬃0.7-

␮m-diameter magnetic bead that was coated with strepta-

vidin was attached to the biotinylated ␥ subunit of an im-

mobilized ␣

␤

␥ complex.When the resulting assembly was

placed in a rotating magnetic field in the presence of ADP

and P

,ATP was produced when the magnetic field rotated

in the clockwise direction but hydrolyzed when it rotated

in the counterclockwise direction. This further demon-

strates that F

is a device that interconverts mechanical and

chemical energy.

f. c-Ring Rotation Is Impelled by H

ⴙ

-Induced

Conformational Changes

The foregoing structural and biochemical information

has led to the model for proton-driven rotation of the F

subunit that is diagrammed in Fig. 22-45. Protons from the

outside enter a hydrophilic channel between the a subunit

and the c ring, where they bind to a c subunit. The c ring

858 Chapter 22. Electron Transport and Oxidative Phosphorylation

Figure 22-43 Rotation of the c ring in E. coli F

–ATPase.

(a) The experimental system used to observe the rotation. See

the text for details.The blue arrow indicates the observed

direction of rotation of the fluorescently labeled actin filament

that was linked to the c ring. (b) The rotation of a 3.6-␮m-long

actin filament in the presence of 5 mM MgATP as seen in

successive video images taken through a fluorescence micro-

scope. [Courtesy of Masamitsu Futai, Osaka University, Osaka,

Japan.]

Figure 22-44 Stepwise rotation of the ␥ subunit of F

relative

to an immobilized ␣

␤

unit at low ATP concentration as

observed by fluorescence microscopy. The graph plots the

cumulative number of rotations made by a fluorescently labeled

actin filament that was linked at one end to the ␥ subunit in a

preparation similar to that diagrammed in Fig. 22-43a (but

lacking F

, ␦, and ε). Note that the actin filament rotates in

increments of 120°.This is also evident in the inset, which shows

the superposition of the centers of the actin images (the ␦

␤

␥

assembly is fixed in the center). [Courtesy of Kazuhiko Kinosita,

Jr., Keio University, Yokohama, Japan.]

Revolutions

Time (s)

a [ATP] = 0.02 μM, Actin length = 1.1 μm

0306090

(a)

(b)

JWCL281_c22_823-870.qxd 6/8/10 9:19 AM Page 858

then rotates nearly a full turn (while protons bind to suc-

cessive c subunits as they pass this input channel) until the

subunit reaches a second hydrophilic channel between the

a subunit and the c-ring that opens into the inside. There

the proton is released.Thus, E.coli F

–ATPase, which has

12 c subunits in its F

assembly and generates 3ATP per turn,

ideally forms 3/12 ⫽ 0.25 ATP for every proton it passes

from the periplasmic space (outside) to the cytosol (inside).

Organisms with more/less c subunits in their c-rotor tend

to have lesser/greater values of proton-motive force across

their membranes and hence have less/more impetus im-

parted to their c-rotor per proton passed. Hence it takes the

passage of proportionately more/less protons to generate

each ATP, as the first law of themodynamics requires.

How does the passage of protons through this system in-

duce the rotation of the c ring and hence the synthesis of

ATP? The mutation of the c subunit’s conserved Asp 61 to

Asn inactivates E. coli F

–ATPase.The a subunit’s invari-

ant Arg 210 (E. coli numbering) has been similarly impli-

cated in proton translocation. Through the mutagenic con-

version of selected residues on the a and c subunits to Cys,

Robert Fillingame has shown that the outer (C-terminal)

helix of E. coli subunit c (Fig. 22-39), which contains Asp

61, can be disulfide-cross-linked to the putative fourth he-

lix of subunit a, which contains Arg 210. Evidently, these

helices are juxtaposed at some point in the c ring’s rotation

cycle. Thus, it is postulated that the protonation of Asp 61

releases its attraction to Arg 210, thereby permitting the

c ring to rotate.

Comparison of the NMR structures of subunit c at pH 8

and pH 5 (Fig. 22-46), at which Asp 61 is, respectively, de-

protonated and protonated, reveals that its main confor-

mational change on protonation is an ⬃140° clockwise ro-

tation (as viewed from F

) of its Asp 61-containing

C-terminal helix with respect to its N-terminal helix. Since

the C-terminal helix is the c ring’s outer helix (Fig. 22-39b),

this suggests that, on protonation, the rotation of the

C-terminal helix mechanically pushes against the juxtaposed

a subunit so as to rotate the c ring in the direction indicated

in Fig. 22-45.

D. Uncoupling of Oxidative Phosphorylation

Electron transport (the oxidation of NADH and FADH

) and oxidative phosphorylation (the synthesis of ATP)

are normally tightly coupled due to the impermeability of

the inner mitochondrial membrane to the passage of pro-

tons. Thus the only way for H

⫹

to reenter the matrix is

through the F

portion of the proton-translocating ATP

Section 22-3. Oxidative Phosphorylation 859

Figure 22-45 Schematic diagram of the action of the E. coli

–ATPase. The ␥ε–c

ring complex is the rotor and the

–␣

␤

␦ complex is the stator. Rotational motion is imparted to

the rotor by the passage of protons from the outside (periplasm)

to the inside (cytoplasm). Protons entering from the outside bind

to a c subunit where it interacts with the a subunit, and exit to

the inside after the c ring has made a nearly full rotation as

indicated (black arrows), so that the c subunit again contacts the

a subunit.The b

␦ complex presumably functions to prevent the

␣

␤

assembly from rotating with the ␥ subunit. [Courtesy of

Richard Cross, State University of New York, Syracuse, New York.]

Figure 22-46 NMR structures of the c subunit of E. coli

–ATPase. The structures, which closely resemble that in Fig.

22-39, were determined in chloroform–methanol–water (4:4:1)

solution at (a) pH 8 (at which D61 is deprotonated) and (b) pH 5

(at which D61 is protonated). Selected side chains are shown to

aid in the comparison of the two structures. Note that the

C-terminal helix in the pH 8 structure has rotated by 140°

clockwise, as viewed from the top of the drawing, relative to that

in the pH 5 structure. [Courtesy of Mark Girvin,Albert Einstein

College of Medicine. PDBids (a) 1C99 and (b) 1C0V.]

JWCL281_c22_823-870.qxd 7/2/10 11:13 AM Page 859

synthase. In the resting state, when oxidative phosphoryla-

tion is minimal, the proton-motive force across the inner

mitochondrial membrane builds up to the extent that the

free energy to pump additional protons is greater than the

electron-transport chain can muster,thereby inhibiting fur-

ther electron transport. However, many compounds, in-

cluding 2,4-dinitrophenol (DNP) and carbonylcyanide-p-

trifluoromethoxyphenylhydrazone (FCCP), have been

found to “uncouple” these processes. The chemiosmotic

hypothesis has provided a rationale for understanding the

mechanism by which these uncouplers act.

The presence in the inner mitochondrial membrane of an

agent that renders it permeable to H

⫹

uncouples oxidative

phosphorylation from electron transport by providing a

route for the dissipation of the proton-motove force that

does not require ATP synthesis. Uncoupling therefore

allows electron transport to proceed unchecked even when

ATP synthesis is inhibited. DNP and FCCP are lipophilic

weak acids that therefore readily pass through membranes.

In a pH gradient, they bind protons on the acidic side of the

membrane, diffuse through, and release them on the alka-

line side, thereby dissipating the gradient (Fig. 22-47).Thus,

such uncouplers are proton-transporting ionophores (Sec-

tion 20-2C).

Even before the mechanism of uncoupling was known,

it was recognized that metabolic rates were increased by

such compounds. Studies at Stanford University in the

early part of the twentieth century documented an increase

in respiration and weight loss caused by DNP. The com-

pound was even used as a “diet pill” for several years. In

the words of Efraim Racker (A New Look at Mechanisms

in Bioenergetics, p. 155):

In spite of warnings from the Stanford scientists, some

enterprising physicians started to administer dinitrophenol to

obese patients without proper precautions. The results were

striking. Unfortunately in some cases the treatment eliminated

not only the fat but also the patients, and several fatalities were

reported in the Journal of the American Medical Association in

1929.This discouraged physicians for a while....

a. Hormonally Controlled Uncoupling in Brown

Adipose Tissue Functions to Generate Heat

The dissipation of an electrochemical H

⫹

gradient, which

is generated by electron transport and uncoupled from ATP

synthesis, produces heat. Heat generation is the physiologi-

cal function of brown adipose tissue (brown fat). This tissue

is unlike typical (white) adipose tissue in that, besides con-

taining large amounts of triacylglycerols, it contains numer-

ous mitochondria whose cytochromes color it brown. New-

born mammals that lack fur, such as humans, as well as

hibernating mammals, contain brown fat in their neck and

upper back that functions in nonshivering thermogenesis,

that is, as a “biological heating pad.” (The ATP hydrolysis

that occurs during the muscle contractions of shivering—or

860 Chapter 22. Electron Transport and Oxidative Phosphorylation

Figure 22-47 Uncoupling of oxidative phosphorylation. The

proton-transporting ionophores DNP and FCCP uncouple

oxidative phosphorylation from electron transport by

diffusion

Matrix

high pH

Mitochondrial

membrane

Cytosol

low pH

2,4-Dinitrophenol (DNP)

+ H

–

OHOH

diffusion

Carbonylcyanide-p-trifluoro-

methoxyphenylhydrazone (FCCP)

+ H

–

discharging the electrochemical proton gradient generated by

electron transport.

JWCL281_c22_823-870.qxd 3/19/10 11:00 AM Page 860