Lecture Notes on ATP Synthase and Oxidative Phosphorylation
use Garrett and Grisham 2nd ed. for illustrations
last modified Oct. 23, 2001 4:30 PM
David Nelson
The last two lectures focused on events that take place in mitochondria. To clarify
what we have been talking about, here is a decent picture of a mitochondrion. The outer
membrane is permeable to small molecules including nucleotides, because it has the protein
porin. The inner membrane is where all the electron transport complexes are found. The
internal space is the matrix. Approximately 800 proteins that are coded by nuclear genes
and made on cytoplasmic ribosomes have to be imported across this double membrane
system and then folded or assembled in the mitochondrion. There are a number of essential
genes that are specific for the assembly of electron transport complexes, but they do not
exist in the final complex structure.
In our last lecture, we covered electron transport from NADH and FADH2 through the
complexes of the electron transport chain eventually arriving at oxygen. This process is
called RESPIRATION. During this process, protons were pumped across the
mitochondrial inner membrane to the periplasmic space. We followed the path of the
electrons completely, but we really did not talk too much about the protons.
Now I want to go over what happens to the protons. As the electrons enter this pathway,
they all encounter iron-sulfur centers that cannot accept protons, so the protons are
separated from the electrons early in the pathway. In complex I (NADH dehydrogenase)
protons and electrons follow different routes. When NADH passes its single proton and
two electrons to FMN, one additional proton has to come with it to form FMNH2. This
extra proton comes from the matrix environment. When the FMNH2 transfers its electrons
(one at a time) to the first FeS center, the protons of FMNH2 are released. It is not
known what happens to these protons. They may be the pumped protons, or they may be
released back to the matrix. The other three flavoproteins that feed electrons to
ubiquinone do not pump protons, so the FMNH2 protons are probably equivalent to the FADH2
protons that do not get pumped. What we do know is that 4 protons are eventually pumped
for each pair of electrons that passes through complex I.
Remember Peter Rich's concept of charge compensation. Placing an electron into a protein
interior is thermodynamically expensive so taking a proton up at the same time can reduce
the cost. Presumably the proton interacts with a side chain or part of a peptide
backbone that can become protonated and in that way compensates for the electron's
negative charge (COO- to COOH, or NH2 to NH3+, etc.). In bacteriorhodopsin, which is
a light driven proton pump, relays of side chains like aspartic acid form a conduit that
the proton can follow through the protein. This has been called a bucket brigade by G.
Cecchini. I have seen a video of a molecular dynamics simulation of this process where
you can watch the protons flowing through the protein, jumping from side chain to side
chain, forming and breaking hydrogen bonds as they go. Changes in pKs of the side
chains and shifts in water molecule hydrogen bonds are also important. Something like
this has to happen in the three complexes I, III and IV that are proton pumps.
Now lets go back to complex I. The electrons have passed to the iron-sulfur centers and
will find their way to a ubiquinone binding site that has an oxidized ubiquinone bound.
This site is near the matrix side of the membrane. As the electrons are transferred to
the ubiquinone it has to take up two protons, and these come from the matrix.
The ubiquinone is released and diffuses to the P center of complex III (bc1 complex).
P stands for the positive side of the membrane, N for the negative side.
When the protons are released here, before the electrons can flow, these protons are
destined to be pumped out of the mitochondrion. At the N center near the matrix side of
the membrane, oxidized ubiquinone is bound and receives electrons from the b hemes, one at
a time during the Q cycle. The reduced ubiquinone has to pick up two protons from the
matrix to form UQH2. Four protons get pumped and two are taken up for each pair of
electrons that move to cytochrome c. Remember that each electron that goes to cytochrome
c comes from a separate ubiquinone, because the electrons from a single ubiquinone split
up and go different ways.
In cytochrome c oxidase no protons are donated by cytochrome c. The 4 protons that
react with oxygen and the 4 pumped protons come from the matrix. Robert Gennis
has mutants in bacterial complex IV that interfere with proton pumping and at least
three of these mutants he argues lie in the path of proton flow. These kind of experiments
in conjunction with the crystal structures have allowed identification of
proton channels in these proteins. One of the pathways in complex IV is called the D-
pathway since it starts with a conserved aspartate residue (D in the one letter code).
Another side chain (glutamate or E) in this pathway is required, but it can been moved to
An adjacent helix in the structure and it still works at about 5% efficiency.
(Biochemistry 2000 Dec 26;39(51):15847-50.) Evidence now suggests that pumped protons and
chemical protons can enter through the D-pathway. There is also a K-pathway, named for a
conserved lysine residue (K) near the beginning of the path. At least one proton enters
through the K-pathway. These two pathways are from the matrix to the active site. There
is also an exit pathway (maybe multiple routes) that takes the pumped protons out to the
intermembrane space, but the exit path is not as well defined.
In April of 1997, a report appeared (PNAS 94, 4526-4531 1997) that mutations in two
different subunits of cytochrome c oxidase COX1 and COX2 correlate with late onset
Alzheimer's disease. These are mitochondrially encoded subunits. A defect in cytochrome
oxidase can reduce the level of ATP in a cell and produce unwanted toxic oxygen
metabolites like superoxide anion and hydroxyl ion. Brain tissue is especially dependent
on energy and is vulnerable to a mitochondrial defect.
OXIDATIVE PHOSPHORYLATION
Respiration is very tightly coupled to synthesis of ATP in a process called oxidative
phosphorylation. The synthesis of ATP uses the energy stored in the proton and
membrane potential gradient to drive a molecular motor called the ATP synthase. This is
another protein complex located in the inner mitochondrial membrane, sometimes called
complex V, but more often called the F0F1 ATPase, named for the reverse reaction that is
easier to measure. This protein has a lollipop structure in electron microscopic images
(fig 21.23 p. 694). The lollipop part is the F1 portion and it can be removed from the
membranes by treatment with urea. The F1 part contains the catalytic
subunits. It has been crystallized and the structure solved. See fig. 21.24 p. 695).
The FO part is membrane bound (the 0 in FO stands for oligomycin,
an inhibitor that binds to this part of the protein). It contains the proton channel.
Part of FO has been crystallized from the yeast ATP synthase,
and it contains the stalk and 10 c subunits, but subunits a and b are missing.
Science 286 1700-1705 Nov. 26 1999. A model of the whole structure is given in Fig. 21.25 p. 695.
When the F1 part of the ATPase is removed from the membranes, the electron transport
chain can still work and oxygen can be reduced to water. The oxidative phosphorylation
(synthesis of ATP) cannot work, and the ATP synthesis is said to be uncoupled from
respiration. There are several ways that mitochondria can become uncoupled. The easiest
and the most troublesome to people who isolate mitochondria, is that their membranes can
be leaky. Not with big holes, but it does not take much to let protons through and destroy
the proton gradient. Another way is the addition of a chemical uncoupler. These small
molecules actually serve as proton shuttles that carry a proton across the membrane down
the gradient. They are protonated on the cytosolic side where the pH is lower (H+
concentration is higher) Then they diffuse across the membrane as uncharged species and
release the proton to the matrix, where the pH is higher. At this point the uncoupler is
charged and cannot go back across the membrane for another proton, so these are one way
proton carriers that accumulate in the matrix. Chemical uncouplers include CCCP,
2,4-dinitrophenol, FCCP (fig. 21.31 p. 700)
If the F1 part of the ATP synthase is added back to the membranes where it was originally,
the two parts reassemble and ATP synthesis can once again be coupled to electron transport
and respiration. This is called a reconstitution experiment.
Oxidative phosphorylation is an unusual process, that depends on the integrity of the inner
mitochondrial membrane. Peter Mitchell, who first proposed this mechanism, is highly
regarded by bioenergeticists because he had the insight to come up with this model which is
called the chemiosmotic hypothesis. He also proposed the Q cycle as a means for pumping
protons. He predicted that the inner membrane would be impermeable to protons on
theoretical grounds and it is. This permits the controlled leak of protons down the gradient
through the ATP synthase to make ATP. How this is done is a very intensely studied area,
and quite a bit is known. Because ATP synthesis is done indirectly, oxidative
phosphorylation is distinct from ATP synthesis seen in glycolysis or GTP synthesis seen in
the TCA cycle at succinyl CoA synthetase. Those direct syntheses of a high energy bond
are called substrate level phosphorylations to distinguish them from oxidative
phosphorylation.
One famous reconstitution experiment provided evidence that the chemiosmotic hypothesis
was correct. In this experiment
performed by Efriam Racker and Walther Stoeckenius, bacteriorhodopsin was reconstituted into
phospholipid vesicles with the FOF1 ATPase. When light was applied, the bacteriorhodopsin
pumped protons into the vesicles, since it is a light driven proton pump. Fortunately for the
experiment, the orientation of the pump could be prevented from being random and there was net
pumping into the vesicles not out from them. With this proton gradient, the ATP synthase was
able to make ATP. This was an elegant experiment that showed there was no high energy
intermediate, or no specific interaction with the electron transport complexes. Just a simple
proton gradient could power ATP synthesis.
THE STRUCTURE OF ATP SYNTHASE
The whole molecule has a dumbbell appearance. (fig. 21.25 p. 695, 21.30 p. 699). The
membrane part and the F1 part are connected by a 50 angstrom stalk. The number
of subunits in the FO part varies depending on the source, but there are only 3 FO subunits in
E coli. They are called a, b and c and they exist in the ratio 1:2:10-12. The c subunit is very
interesting, since it is small and exists in so many copies. It has only two transmembrane
segments and one of these has a critical aspartate residue in the middle of the helix. This
aspartate is required for the proton channel. A chemical called dicyclohexyl carbodiimide
reacts with this side chain and kills the enzyme. In fact it is often called the DCCD binding
protein. In an interesting experiment Robert Fillingame mutated the E. coli subunit c
aspartate residue to eliminate the carboxyl group. He then selected for regain of function
mutants that could make ATP again. What he found was the negative charge from the carboxyl
reappeared in the protein structure, but it came back as a glutamate on the other
transmembrane helix at about the same depth into the membrane. This carboxyl is protonated at
some point during the transport of protons through the channel and this is required to make
the protein work.
In yeast the three subunits that are equivalent to the three E. coli FO subunits are all encoded
in the mitochondrial genome. This is just another example of the mitochondrion keeping
control of these very critical proteins and not giving them up to the nuclear genome. Doug
Wallace, an expert in mitochondrial diseases, wrote in Scientific American that the
symbiont (the mitochondrion) ensures its own survival by keeping its fingers on the jugular
vein of cellular energy flow.
The FO part of the ATPase includes about half of the long stalk. The other half comes from
the F1 particle. In the stalk region, there is a protein called the oligomycin sensitivity
conferring protein or OSCP. OSCP is required for sensitivity to oligomycin, an antibiotic
inhibitor of the ATPase. However, oligomycin binds to the FO part of the protein, not to
the OSCP. The other portions of the stalk are just being characterized.
The F1 particle has been crystallized and the structure is known. This work earned John
Walker the Nobel Prize in Chemistry in 1997. The subunit composition of F1
is alpha 3, beta 3, gamma, delta, epsilon. Alpha and beta alternate like 6 slices of an
orange to make a ball. Only the beta subunits are active catalytic subunits. The alpha
subunits have binding sites for nucleotide, but they are not catalytic. In the crystal
structure, delta and epsilon are not visible, but gamma has two long helices that run
through the center of the ball made by alpha and beta. This long helical gamma subunit
sticks out the bottom of the ball formed by alpha and beta to form part of the stalk. It is
known that the gamma subunit rotates relative to the F1 ball during ATP synthesis, so that
gamma is like a molecular axle and the central hole in F1 is like a molecular bearing. It is
very important to this model that the gamma subunit is not symmetrical. It has significant
differences in its interactions with the three catalytic beta subunits, and this causes
distortions or conformational changes in the beta subunits. This is important as we will see
below.
THE BINDING CHANGE MECHANISM
The model of how the ATP synthase works is called the binding change mechanism.
This mechanism has three main tenets or postulates. Paul Boyer received the Nobel Prize
in Chemistry in 1997 for devising this model and the experiments to confirm it.
1. The first postulate is that the energy of the proton gradient is not used to form ATP but
to release ATP from a very tight binding site where it forms spontaneously. The release of
this tightly bound ATP requires the energy of the proton gradient converted into the
mechanical energy of protein conformational changes probably brought about by
interactions with the gamma subunit and cooperative changes with the other alpha and beta
subunits.
2. The second postulate says that the three catalytic sites are each in a unique conformation
and the conformations are interconvertible. They represent three different stages of the
catalytic cycle. One site L is loose, with ADP and Pi bound. ATP forms without added
energy input. Afterwards the contacts with the newly formed ATP become very strong.
This tight conformation T then needs the energy from the proton gradient to release the
ATP. The third state is the open or O state before nucleotide is bound.
3. The third postulate is that conformational changes at the three sites are driven by rotation
of the asymmetric gamma subunit relative to the F1 ball.
EVIDENCE FOR THE THREE POSTULATES.
The first tenet was shown to be correct by O18 water exchange experiments. If the ATP
synthase is incubated with ADP, Pi and O18 water, Paul Boyer showed that O18 gets into
free phosphate. The only way this can happen is by forming ATP and then rehydrolyzing
it. The O18 incorporation occurred without a proton gradient, therefore, ATP was formed
without energy input from the gradient. ATP was not released unless the gradient was
applied, so the energy is needed to release the ATP not to form the ATP.
The crystal structure immediately supported the second tenet, because the three catalytic
sites were each in a different conformation.
The third tenet, that the protein rotates has now been proved by Richard Cross and others.
He has made a mutant with cysteine added into the beta subunits very near a naturally
occurring cysteine in the gamma subunit. If the cysteine mutation is added to the beta
subunit without oxidation, nothing happens and ATP is made. If both cysteines are present
together and the disulfide is formed by oxidation, then no ATP can be made. The protein is
locked and cannot turn. (Proc. Natl. Acad. Sci USA 92, 10964-10968 1995). This
experiment was done with soluble F1. A similar experiment was done with membrane
bound FOF1 ATPase to show the relevance to the membrane bound intact ATP synthase
(Biochim. Biophys Acta 1275, 96-100 1996).
There is an argument against the rotation interpretation that says the conformational changes
in alpha and beta are large enough that a disulfide might prevent them even if the protein
does not rotate. In this kind of model the ball sits in one place and beats like a heart.
To try and prove that the protein actually rotates, a more sophisticated experiment has been
designed. The disulfide linked ATPase has been dissociated and reconstituted with beta
subunits that have the cysteine mutation and epitope tags. The F1 protein has been
reconstituted with the FO part. The disulfide was reduced and the protein was allowed to
make ATP. After a brief time the disulfide was reformed by oxidation and the epitope tags
were used to detect the location of beta-gamma cross linked subunits. If no rotation had
occurred then the epitope tag should not be covalently attached to gamma. The original
untagged subunit should reform the cross link. However if the F1 portion rotates, then
some of the epitope tagged beta should now be in a position to cross link with gamma.
This will be detectable on SDS gels by an increase in the size of the epitope tagged protein.
And that is what was seen. (Proc. Natl. Acad. Sci USA 94, 10583-10587 1997).
The same type of experiment was done with cysteines added between epsilon and beta.
These proved that epsilon is part of the rotor, just like gamma. (J Biol Chem. 1998 Nov
27;273, 31765-9). Another crosslinking experiment from Fillingame's lab showed that
epsilon is in contact with the c subunit (J Biol Chem 1999 Jun 11;274(24):17011-6)
So there is evidence for rotation. Richard Cross is now doing another experiment to
actually measure the rate of the rotation. He has engineered a tryptophan into the protein
near the site where the disulfide was. On the other subunit he has added a label that will
quench the tryptophan fluorescence when it is close. In this experiment, the fluorescence
of the tryptophan should vary with the rotation of the F1 part. The period of the fluctuation
in fluorescence should be equal to the period of rotation. This experiment is being done
now, but the results are not in yet. One problem with an experiment like this is
synchronization. It will be necessary for all the ATPases to start at the same time, or there
will only be an average fluorescence. This can be achieved by using caged ATP that can be
released by a pulse of light, so that all the ATPases will see the ATP at the same time.
In March of 1997, another very simple and elegant experiment showed the rotation of
the gamma subunit directly. Hiroyuki Noji and collaborators engineered ten histidines onto
the N-terminal of the the ATPase beta subunit and expressed these his-tagged proteins in
bacteria. The F1 part of the ATPase was purified and attached to a glass slide that was
coated with nickel. The ATPase beta subunits in the F1 complex bound to the slide
through their his tags and immobilzed the F1 particle upside down with the gamma subunit
facing up. The gamma subunit was mutated so it had only one cys residue near its end.
The cys was used as a site for biotinylation and attachment of streptavidin. Finally, a
biotinylated flourescent actin filament was added to the complex and it bound to the
streptavidin. Fluorescence of the actin filament tagged to the gamma subunit was observed
in a flourescence microscope. When ATP was added to the slide, the filaments spun
around. It moved at about 4 cycles per second, slower than expected for the free enzyme,
(about 20 cycles per cecond) but it has a big drag on it from the actin filament.
(Nature 386, 217-219 1997 news and views article, and 299-302 research article.)
A more recent experiment repeats this strategy with intact F1FO and couples the
actin fiber to the c subunits. The result is the same, showing that the c
subunits rotate. Science 286, 1722-1724 1999.
NEW FINDINGS ABOUT THE STRUCTURE OF THE ATP SYNTHASE
Most of the data we have talked about concerns the F1 part of the ATPase. What about the
rest of the complex. Some additional experiments have been done in 1997 and 1998 that
clarify the mechanism of the whole complex. Once again, disulfide bonds have been used
to show interaction of subunits and rotation or lack of rotation between subunits. Two of
the subunits of the F1 particle that we have not discussed are the delta and epsilon subunits.
The question is do these subunits rotate with gamma or stay fixed with alpha and beta in the
ball structure. by introducing more site directed cysteines at critical places, the motion or
lack of motion of these subunits has been probed. Experiments with delta crosslinked to
alpha show that no effect is found when the crosslink is formed by oxidation (J. Biol.
Chem. 272, 16652-16656 1997). Therefore delta is fixed relative to alpha. It is argued
that delta is part of the stator or fixed part of the motor. For gamma to turn relative to the
alpha and beta ball, some part of the structure must connect the ball to the FO membrane
structure, otherwise the ball would spin too and no ATP would be made. Similar
experiments with epsilon show that linking epsilon to alpha kills the activity, so epsilon
moves with gamma in the interior of the ball structure.(J. Biol. Chem. 272, 19621-19624
1997; J. Biol. Chem. 1998 Nov 27;273, 31765-9.)
The FO part of the enzyme is also critical to function. Since there are 10-12 c subunits in FO
and only one a and two b subunits, it is important to know which of these move and which
don't. Experiments like the ones described above using cys linkages have shown that
subunit b binds to delta and is part of the stator, while subunits c are rigidly fixed to gamma
and are part of the rotor. The single a subunit is also part of the stator. Therefore, the ball
is fixed to the membrane subunit a through delta and b. This makes up the stator, while
gamma and epsilon are coupled with the 10-12 c subunits to make the rotor.
In November of 2000, two groups reported the x-ray crystal structure of the stalk subunits
gamma, delta and epsilon from bovine heart and the equivalent subunits from E. coli. (Nature Structural
Biology 7, 1002 (commentary by Fillingame) 1051 (E. coli) 1055 (bovine heart). The structure identified an
ARG (R75) that probably interacts with an acidic sequence DELSEED in the beta subunit with five
aspartate and glutamate residues. This interaction may provide a catch mechanism to stabilize the protein
at 120o intervals as it turns. A foot for the stalk was also described that would bind to the
ring of c subunits and join them to the stalk, since they are both part of the rotor. The only part of
the ATPase missing in the structures now is the a and b subunits that connect the membrane part of the
stator to the ball of F1. (please note that bovine delta is not the same as delta in E. coli
that anchors the b subunits of the stator to the alpha-beta ball. The bovine delta subunit is equivalent to
the E. coli epsilon subunit).
CONSIDERING THE ELECTRON TRANSPORT CHAIN AND ATPase
AS A WHOLE INPUTS AND OUTPUTS AND THE P/O RATIO
There are some aspects of the electron transport chain and FOF1 ATPase that need to be
looked at from the perspective of a system. So far we have examined each part as a unique
structure. One property that depends on the whole chain is the P/O ratio, or how many
ATP molecules are produced per oxygen consumed. The traditional way to score this is
number of ATP produced per pair of electrons sent down the chain. Since it takes 4
electrons to reduce molecular oxygen to two waters, the P/O ratio is really ATP produced/
atom of oxygen. This number is dependent on how many protons are pumped per site in
the chain and how many are consumed to make one ATP. The best estimates are 4 protons
per electron pair at complex I, 4 protons per electron pair at complex III (2 for each turn of
the Q cycle) and 2 protons per electron pair at cytochrome oxidase (This amounts to 4
protons for complete reduction of O2 to water).
The estimate for protons consumed per ATP synthesized is 4. There is an interesting
theory about the ATPase that might explain this number. It is called the elevator model.
Remember that the FO part had 10-12 c subunits, each with a single DCCD binding aspartate.
Single mutations in subunit a stop transport, so subunit a probably forms part of the proton
channel. The interesting fact is that mutation of even 1 of the 10-12 subunits of c will also
stop transport. This suggests that the a subunit rotates relative to the c subunits to form a
complete proton pathway. The elevator model suggests that protons start their passage through
the membrane in the a subunit. At some point they cannot continue unless they move onto one
of the c subunits which then carries the proton as the ATPase rotor moves 1/12th (or 1/10th) of a full
circle. This opens up the proton path in subunit a once again for another proton to leave its c
subunit and exit. By this model, the rotor would have to turn 1/12th (or 1/10th) of a circle for every
proton, or 10-12 protons could pass through per revolution. We already know that three ATPs are
made per revolution of the F1 relative to gamma, so that gives 3/12 = 1 ATP per 4 protons. If
the number of c subunits were variable, then there might be a variable stoichiometry. If there
were 10 c subunits the math would be 3.3 protons per ATP, It does not have to be an integer
number if the number of c subunits is variable.
This elevator model is a speculative model that is not supported with hard evidence like the
disulfide experiments.
Back to the P/O ratio. For electron transport from NADH to O2, there are 10 protons
pumped at 4 protons/ATP = 2.5 ATP/NADH. For transport starting with FADH2, only 6
protons are pumped at 4/ATP = 1.5 ATP/FADH2. There is a way to add electrons directly
to cytochrome c and bypass complex III. Ascorbic acid can reduce a compound called
tetramethyl-para-phenylenediamine (TMPD) which can reduce cytochrome c. If this
method is used, 2 protons are pumped per pair of electrons (2 cytochrome cs) for a ratio of
2/4 = 0.5 ATP/pair of electrons.
By using different substrates and inhibitors, it is possible to tell where a complex lies in the
electron transport pathway. In practice, NADH and FADH2 are not added in experiments.
Succinate can be added that enters the TCA cycle at complex II and makes FADH2. This
bypasses complex I. beta-hydroxybutyrate is oxidized by beta-hydroxybutyrate
dehydrogenase to form NADH that feeds into the pathway at complex I. You should know
where these substrates enter the chain.
If I told you a new complex had been isolated from chili peppers that accepted electrons
from some exotic compound X in chilis, you should be able to interpret the results of
adding in inhibitors and substrates along with P/O ratios to place the complex in the
electron transport chain. If compound X is added and you measure a P/O ratio of 1.5 that
is completely blocked by antimycin, but not affected by rotenone, you should be able to tell
that this complex is like complex II and it feeds in electrons upstream of complex III, but
does not interact with complex I. If you add compound X and the P/O ratio is 2.5 and
blocked by all known inhibitors, then it must be upstream of complex I and probably is a
compound X dehydrogenase that produces NADH.
The inputs and outputs of glucose oxidation (Table 21.4 p. 705) depend on the
method used to transport reducing equivalents from the cytosol into the mitochondrion. We
already talked about the glycerophosphate shuttle, with two forms of glycerophosphate
dehydrogenase, one membrane bound and the other cytosolic. These enzymes fed electrons from
NADH into the electron transport chain at ubiquinone, so some of the available energy was
lost. There is another way to get NADH reducing equivalents across the inner membrane
that keeps the electrons entering the path at the level of NADH. This is called the malate-
aspartate shuttle.
This shuttle is illustrated in fig 21.34 p. 703. To get the NADH reducing
equivalents across the membrane, they
are first transferred to oxaloacetate to form malate by the malate dehydrogenase in the
cytosol. Malate moves into the mitochondrion on a carrier that exchanges malate for alpha-
ketoglutarate. Once inside, the mitochondrial malate dehydrogenase oxidizes the malate back
to oxaloacetate and NADH is formed. This gets the NADH inside the mitochondrion, but it does
not complete the shuttle. So far oxaloacetate has moved in and alpha-ketoglutarate has moved
out. We need to restore these to the original side of the membrane where they started. This
is done by aspartate amino transferase that converts oxaloacetate in the matrix to aspartate
and glutamate to alpha-ketoglutarate. A second carrier now exchanges the aspartate for
glutamate. Now everything is correct on the matrix side. On the cytosolic side aspartate
transaminates alpha-ketoglutarate to glutamate and reforms oxaloacetate. All is back to
normal now except that NADH has moved inside the mitochondrion.
Notice this is an example of the same enzymes existing in two different compartments.
This is fairly common.
This malate-aspartate shuttle illustrates another common feature of mitochondria. These are
the carriers. In yeast, there are 35 genes for mitochondrial carriers, some of them are
redundant, but there seem to be at least 28 different functions performed by these carriers.
In humans there is sequence evidence for at least 46 different genes. C.
elegans, whose genome is now complete, has a total of 34 carrier
genes (counted by exhaustive blast searching with all C. elegans carriers against the C.
elegans genome. Searches done on Oct. 21, 1999).
Searches of sequence databases show 33 carriers in the Candida albicans genome.
Two very important carriers in mitochondria are the phosphate carrier that brings in
phosphate for ATP synthesis in exchange for hydroxide ion, and the ADP/ATP carrier that
exchanges ADP from the cytosol for ATP freshly made in the matrix. In a resting human,
the ATP turnover amounts to about 80 pounds of ATP per day, and almost all of this goes
through the ADP/ATP carrier twice per cycle, once going in and once going out. So
approximately your body weight in adenine nucleotide passes through this protein every
day. The direction of transport is affected by the electrical membrane potential which is
inside negative. ADP and ATP move through this carrier without complexed magnesium.
The movement is electrogenic, because ATP is -4 and ADP is -3. This favors movement of
ATP out and ADP in (the most negative species goes to the positive side of the membrane).
Another member of this family is the uncoupler protein, also called thermogenin because it
generates heat. This protein transports fatty acid anions across the inner membrane from
the matrix in exchange for a counter ion like chloride or hydroxyl ion. Once the fatty acids
are released on the outer positive side of the membrane they become protonated and flip
flop back across the membrane as non-charged species. On the matrix side they
deprotonate and become charged again. This is like an uncoupler, except the fatty acids do
not have to accumulate at the matrix side since the uncoupler protein can transport them out
again. This process uncouples oxidative phosphorylation and wastes the proton gradient as
heat. It is advantageous to be able to generate heat in special situations and this protein
only occurs in brown fat adipose tissue where it is hormonally regulated. It is activated by
free fatty acids that are produced by lipolysis in response to hormones. The protein is
important in newborn mammals and hibernating animals when they need to wake up. It is
a new evolutionary invention, because it only exists in mammals. In March of 1997 a
new UCP2 protein was found in humans and everyone got very excited, because it was in
many tissues, not just brown fat. Adult humans have no brown fat, so UCP1 was pretty
useless to adults. UCP2 and now UCP3 can be turned on in adults as a way to regulate
weight by wasting energy through uncoupling oxidative phosphorylation. The search is
now on to find drugs that will activate UCP2 or UCP3.
One final consideration about mitochondria and respiration/oxidative phosphorylation is
what are the consequences of a tightly coupled system. Electron flow is linked to ATP
synthesis. If electrons do not flow no ATP can be made. That is obvious from all we have
seen so far. But the opposite is also true if no ATP can be made then electrons cannot flow
because of the tight coupling. This is not so obvious. If substrate is present, like
succinate, but no ADP or Pi is present, the ATPase cannot make ATP, but one would still
expect that electrons could flow down the chain to water. However they do not. Why?
The reason they don't is the proton gradient. When electrons flow, protons are pumped
against an existing proton gradient. Once this gradient is large enough, the energy released
by electron movements to higher reduction potential is just equal to the force applied by the
gradient and no more protons can be pumped. If the gradient is relaxed by synthesis of
ATP, then more electrons can flow. If the gradient is relaxed by an uncoupler then the
electrons can flow. Otherwise a steady state is reached and electron flow ceases.
This is illustrated in the following oxygen electrode traces. When mitochondria are
incubated with succinate as an electron donor, and ADP and Pi are present, we have rapid
oxygen consumption. This is called state 3 respiration. When the ADP runs out, the slope
drops back to a much lower level called state 4 respiration. This is not a zero slope,
because there is always some turnover of ATP in the matrix and the membranes are always
a little bit leaky to protons. The ratio of state 3 to state 4 respiration is called the respiratory
control ratio and it should be about 5-6 in good preps of isolated mitochondria. Addition of
an uncoupler produces the same state 3 level of respiration. The mitochondria show a
different appearance in electron microscope images when they are in state 3 respiration
versus state 4. The actively respiring mitochondria have a large periplasmic space and the
matrix appears condensed. In state 4 they look more typical with very little periplasmic
space.