THEJOURNAL OF BIOLOGICAL
CHEMISTRY
Vol. 257, No. 14. Issue of July 25, pp. 8055-8060,
1982
Printed in U.S.A.
Adenine Nucleotide Binding Sites on Beef Heart F1-ATPase
SPECIFICITY OF COOPERATIVE INTERACTIONS BETWEEN CATALYTIC SITES*
(Received for publication, February 16, 1982)
Carlo M. NalinS and Richard L. Cross
From the Department of Biochemistry, State Uniuersity of New York, Upstate Medical Center, Syracuse, New York 13210
5’-Adenylyl-P,y-imidodiphosphate
is a close structural analog of ATPthathasbeenfoundusefulincharacterizing
substrate binding sites of ATP-utilizing enzymes (Yount et
al., 1971). The analog binds tightly to the terminal coupling
factor, F1,’ of mitochondrial oxidative phosphorylation (Pe* This research was supported by Research Grant GM 23152 from
the National Institutes of Health, United States Public Health Service. The costs of publication of this article were defrayed in part by
the payment of page charges. This article must therefore be hereby
marked “aduertisement” in accordance with 18 U.S.C. Section 1734
solely to indicate this fact.
$Present address, Section of Biochemistry, Molecular and Cell
Biology, Division of Biological Sciences, Cornell University, Ithaca,
New York 14853.
’ The abbreviations used are: F1, the terminal coupling factor of
oxidative phosphorylation and photophosphorylation. Soluble F,isolated from beef heart mitochondria is used in this study. AMP-PNP,
5”adenylyl-B,y-imidodiphosphate; F1.AMP-PNP, soluble complex
between the enzyme and a single mol of AMP-PNP bound a t an
exchangeable site; efrapeptin, also referred to in the literature as
nefsky, 1974; Philoand Selwyn, 1974). However, the p,yimidodiphosphate bond cannot becleaved to products which
dissociate rapidly during normalcatalysis. Hence, AMP-PNP
forms a stable dead-end complex with the enzyme.
Inhibition of ATP hydrolysis by the analog is competitive
with respect to ATP (Penefsky, 1974; Melnick et al., 1975)
and is reversed under conditionsfor ATP synthesis(Penefsky,
1979, Chernyak andKozlov, 1979). We have previously shown
that, althoughF, contains threecopies of the catalytic subunit,
the bindingof a single mol of AMP-PNP/mol of FI at a high
affiiity site (& = 18 nM) is sufficient for complete inhibition
(Cross andNalin, 1982).
In the current study,
we have examined the possibility that
soluble F, containing a single AMP-PNP bound at a catalytic
site may serve as a useful model in assessing the cooperative
interactions between catalytic subunits that are required for
product release during
oxidative phosphorylation. Theeffects
of binding ADP, Pi, or ATP on the dissociation of the F,.
AMP-PNP complex can be determined for single turnover
events, since under the condition used AMP-PNP does not
rebind toF1once released. The resultsof these measurements
are consistent with isotope
exchange data that have
led to the
formulation of a mechanism for ATP synthesisin which the
release of ATP from a catalytic site requires the simultaneous
binding of ADP andPi at an adjacent interacting site (Kayalar
et al., 1977).
EXPERIMENTAL PROCEDURES
Muterials
ATP, ADP, AMP, phosphoenolpyruvate, pyruvate kinase, hexokinase, glyceraldehyde-3-phosphate dehydrogenase, lactate dehydrogenase, defatted bovine serum albumin, and Sephadex G-50-80 were
obtained from Sigma Chemical Co. and Boehringer Mannheim Biochemicals. AMP-PNP, [3H]AMP-PNP, [I4C]ADP,and [‘*C]ATPwere
obtained from ICN Pharmaceuticals. Efrapeptin was a gift from Dr.
R. L. Hamill of Eli Lilly Co. Allother chemicals were of reagent grade
quality.
Methods
Preparation of Fl-Beef heart mitochondria were prepared by the
method of Smith (1967) and stored at -70 “C. Mitochondrial F, was
prepared and stored as described by Knowles and Penefsky (1972).
Prior to use, an aliquot of the ammonium sulfate suspension of the
enzyme was centrifuged at room temperature in a Beckman Airfuge,
and thepellet was dissolved in asmall volume of Mg-buffer. An equal
volume of saturated ammonium sulfate was added, the precipitate
was collected by centrifugation, and theenzyme was dissolved in Mgbuffer. Residual ammonium sulfate was removed by centrifugation of
100-pl aliquots through Sephadex columns (Penefsky, 1977) equilibrated in Mg-buffer.
Nucleotide-promoted Dissociation of the Fl.AMP-PNP Complex-F1 with a single tightly bound r3H]AMP-PNP was prepared by
efrastatin and A23871; Mg-buffer, 150 mM sucrose, 10 mM 4-(2-hydroxyethy1)-1-piperazineethanesulfonicacid (pH 8.0), 0.1 mM EDTA,
2 mMMgC12.
8055
Downloaded from www.jbc.org by guest, on October 30, 2009
Cooperative interactions between nucleotide binding
sites on beef heart mitochondrial F1-ATPasehave been
studied by measuring substrate-promoted release of 5‘adenylyl-p,y-imidodiphosphate(AMP-PNP) from a single high affinity site. The site is initially loaded by
incubating F1 with an equimolar amount of the nonhydrolyzable ATP analog. When unbound r3H]AMP-PNP
is removed and the complex diluted to a concentration
below the K,, release of ligand shows an apparent
absolute requirement for medium ADP. Release is b1phasic with the extent of release during the initial
rapid
phase dependent on the concentration of mediumADP.
Although phosphate alone has no effect, it enhances
the rapid phase of ADP-promoted release over 2-fold
with a half-maximal effect at 60 pM Pi. The binding of
efrapeptin (A23871) to theF1-AMP-PNP complex completely prevents ADP-promoted dissociation.
Although AMP-PNP release also occurs in the presence of medium ATP, the F1- AMP-PNP complex does
not dissociate if an ATP-regenerating system of sufficient capacity to prevent accumulation of medium ADP
is added. Consistent with an inability of nucleoside
triphosphate to promote release is the failure of medium, nonradioactive AMP-PNP to affect retention of
the 3H-labeledligand.
The stability of the F1*AMP-PNP complexin the absence of medium nucleotide and the highly specific
ability of ADP plus Pi to promote rapid release of the
ATP analog are interpreted as support for an ATP
synthesis mechanism that requires substrate binding
at one catalytic site for product releasefrom an adjacent interacting site
(Kayalar, C., Rosing, J., and Boyer,
P. D. (1977) J. Biol. Chem 252,2486-2491).
8056
Fl-ATPaseSubunit Interactions
I
IOOOpM
’.
J
5
10
15
20
25
30
Time, min.
FIG. 1. ADP-promoted dissociation of F1-AMP-PNP.A complex containing 0.9molof
r3H]AMP-PNP bound/mol of F1 was
prepared and resolved of unbound nucleotide as described under
“Methods.” At zero time, the complex was diluted to 4 pgof Fdml
with Mg-buffer containing 0.25 mg of albumin/& and ADP at the
micromolar concentration indicated. Most incubations (open symbols)
also contain 10 pg of hexokinase/ml, 1 mM glucose, 20 mM AMP. At
the times indicated, duplicate aliquots were subjected to column
centrifugation and the effluents analyzed for bound [3H]AMP-PNP.
The fraction of the F1.AMP-PNP complex remaining is plotted on a
semilog scale against the time of incubation.
ADP, several precautions were taken to eliminate ATP from
the incubations. Synthesis of ATP by trace amounts of adenylate kinase that may contaminate beef heart F, was prevented by 20 mM AMP. In addition, hexokinase and glucose
were added to remove any ATP introduced as acontaminant
with AMP or ADP. Although these additions have little effect
on ADP-promoted dissociation of F1. AMP-PNP, they do
stimulate a slow release of AMP-PNP in controls lacking
added ADP (Fig. 1, compare open and closed circles). This
was determined to be due to theintroduction of trace amounts
of ADP with 20 mM AMP.
Additional experiments were performed to determine the
cause of the biphasic release of F1-bound AMP-PNP. The
onset of the slow phase shown in Fig. 1 is not due to the
rebinding of dissociated [3H]AMP-PNP,since results identical
with those shown are obtained when nonradioactive AMPPNP in200-fold molar excess to bound [3H]AMP-PNP is
added simultaneously with ADP. The slow phase of release
shown in Fig. 1 isalso not the result of depletion of the
RESULTS
promotor. Analyses of the ADP remaining after 20 min of
incubation show that at each concentration of ADP tested
ADP-promoted Release of F,-bound [3H]AMP-PNP-lncubation of F1with equimolar amounts of AMP-PNP results less than 4% is converted to AMP. The biphasic response does
in the binding of ligand at a single high affinity site, K d = 18 not appear to be due to pre-existing heterogeneity in the
nM (Cross and Nalin, 1982). Removal of bound ligand and enzyme-ligand complex. The extent of the fast phase is dedilution of the F1.AMP-PNP complex to a concentration pendent on the amount of ADP present (Fig. l).In addition,
below the K d fails to result in a measurable rate of dissociation identical patterns of AMP-PNP release are obtained when
(Fig, 1, closed circles), Even when incubated under these ADP-promoted release is measured immediately after prepaconditions for 5 h, less than 8%of the AMP-PNP is released ration of the F1.AMP-PNP complex or after aging the com(data not shown). In contrast to the stability of the complex plex for 200 min (data not shown). These results indicate that
in the absence of medium nucleotides, the addition of ADP a conformational change rendering a fraction of the FI.
causes a biphasic release of AMP-PNP. The extent of release AMP-PNP complex less susceptible to promoted release ocin the rapid phase increases with increase in ADP concentra- curs only in the presence of medium nucleotide.
PiEnhancement of ADP-promoted Dissociation of F1tion, whereas the rate of release in the slow phase appears t o
AMP-PNP-At concentrations of 1 mM or less, inorganic
be independent of ADP concentration (Fig. 1).
To ensure that the effects observed in Fig. 1 are due to phosphate fails to promote dissociation of F1. AMP-PNP dur-
Downloaded from www.jbc.org by guest, on October 30, 2009
incubating 0.8 p~ C3H]AMP-PNP (specificactivity = 4000 cpm/pmol)
with 0.3 mg of Fl/& in Mg-buffer. After 90 min, the F1.AMP-PNP
complex was resolved of unbound C3H]AMP-PNP using centrifuge
columns equilibrated in Mg-buffer. Under these conditions, AMPPNP does not exchange with “endogenous” nucleotides bound at
noncatalytic sites, but rather binds a t what, appears to be one of the
multiple copies of the catalytic site (Cross and Nalin, 1982).
The enzyme was then diluted to between 10 and 12 nM using Mgbuffer containing 0.25 mg of albumin/ml. In experiments designed to
test the ability of ADP to promote release of r3H]AMP-PNP, the
reaction mixtures also contained 20 mM AMP, 10 pgof hexokinase/
ml, 1 mM glucose. In measurements of the effect of ATP and AMPPNP on dissociation of the complex, 10 m~ phosphoenolpyruvate
and pyruvate kinase at the concentrations indicated were added as
an ATP regeneration system. MgClz was maintained in 2 mM excess
to the concentration of adenosine di- and triphosphate in all experiments.
At the times indicated, 100-pl aliquots of the incubations were
transferred to centrifuge column extension tubes that contained 10
p l of10mgof
albumin/& to minimize loss of Fl on the columns
(Cross and Nalin, 1982). Samples were centrifuged for 2 min a t 100
X g through columns equilibrated in Mg-buffer. The time of incubation corresponds to the time elapsed between the dilution of the
enzyme. AMP-PNP complex into Mg-buffer containing the promoters
and the time centrifugation was initiated. Column effluents were
collected directly in scintillation vials. Bound AMP-PNP iscalculated
from the tritium that elutes from the column, corrected for incomplete
recovery of Fl. This correction amounted to less than 15%of the F,
applied as determined in controls using 3H-pyridoxylated F1 (Koga
and Cross, 1982).
For each experiment, identical reaction mixtures were prepared
containing [’HIAMP-PNP equivalent to theamount added as the F1.
AMP-PNP complex but lacking Fl. No counts above background
were found in the column effluents, establishing that, subsequent to
release from F1, AMP-PNP does not bind to the kinases or to other
proteins present inthe reaction mixtures.
Despite the addition of20mM
AMP to incubations containing
ADP, it is possible that contaminating adenylate kinase activity in
either the F, (Penefsky et al., 1975) or hexokinase preparations might
slowly deplete the medium ADP. To assess this possibility, incubations were prepared containing 10 nM FI. AMP-PNP, 20 mM AMP, 5
pg hexokinase/ml, 10 mM glucose, and [I4C]ADP at various concentrations between 15 PM and 1.0 mM. Reaction was stopped after 20
min by addition of 1.0 M perchloric acid, and 14C-labelednucleotides
were analyzed by thin layer chromatography as described below.
Other Methods-The concentrations of nucleotide stock solutions
were determined from the absorbance at 259 nmusing a molar
extinction coefficient of 15,400.The purity of nucleotide solutions was
analyzed by thin layer chromatography on polyethyleneimine-cellulose plates (MC/B 5504) using 1 M LiCl (Randerath and Randerath,
1964). Radiochemical purity was determined using a Packard radiochromatogram scanner.
Protein was determined by a modified Lowry procedure (Hartree,
1972) after removal of interfering substances (Bensadoun and Weinstein, 1976) using defatted bovine serum albumin as a protein standard. Values determined for F, were divided by 1.18 to convert to dry
weight (Penefsky and Warner, 1965) and the molar concentration of
FI was calculated using M, = 347,000 (Knowles and Penefsky, 1972).
f - A T P a s e Subunit Interactions
-
;;;-r3 ,I
:
l
Y
-0)
._
,
,
,
1.25
0)
LT
1.00
0.2
0.4 0.6
[Phosphate],
0.8
1.0
mM
FIG. 2. Phosphate enhancement of ADP-promoted dissociation of F1oAMP-PNP. A complex containing 0.9 to 1.0 mol of [’HI
AMP-PNP/mol of F, was prepared and resolved of unbound nucleotide as described under “Methods.” At zero time, the complex was
diluted to 4 pg of Fl/ml with Mg-buffer containing albumin, hexokinase, glucose, and AMP as described for Fig. 1. The incubation also
contained 100 PM ADP and P, at the concentration indicated. After 2
min, duplicate aliquots from each incubation were removed and
subjected to column centrifugation to separate bound from unbound
ligand. Different symbols represent independent experiments. Incubation of the complex in buffer containing 100 PM ADP but lacking
added P, resulted in an average release of 23%of the [‘HIAMP-PNP
(1.0 on the y - a x i s ) .
1
I
0
I0
20
3’0
40 io
[Phosphate] , pM
FIG. 3. Effect of a Pi trap on ADP-promoted dissociation of
Fl.AMP-PNP. Mg-buffer containing albumin, hexokinase, glucose,
and AMP (as described in Fig. 1) and 100 p~ ADP with (0,W) or
without (0,
0 ) added Pi was incubated for30 min with (U, W) or
without (0,
0 )a P,-utilizing, coupled enzyme system consisting of 0.1
mM glyceraldehyde 3-phosphate, 0.1 mM NAD+, 50 pg of glyceraldehyde-3-phosphate dehydrogenase/ml, 0.1 mM pyruvate, 50 pg of lactate dehydrogenase/ml. FI containing 0.9 mol of [’HIAMP-PNP
bound/mol of enzyme was then added to a final concentration of 5
pg of Fl/ml. After 2 min, duplicate aliquots were subjected to column
centrifugation to separate bound from unbound ligand. Incubation of
the complex in buffer containing 100 ~ L MADP but lacking added P,
resulted in release of 27% of the [3H]AMP-PNP (1.0on the y - a x i s ) .
ligand (closed squares and triangles). As a control, both Pi
and ADPwere added simultaneously at zero time (circles).
The delayed addition of Pi to anincubation containing ADP
results in a second rapid phase of AMP-PNP release. The
total amountof FI AMP-PNP thatdissociates during the two
bursts is approximately equal to the amount thatdissociates
during the single rapid phasein the othertwo samples. Hence,
it appears that the
order of addition of ADP and Pi has little
effect on the magnitude of the rapid phase. The rate of the
slow phase is affected, however, with the delayed addition of
ADP giving the fastest rate (Fig. 4, solid squares).
Efi-apeptin Preuents ADP-promoted Dissociation of Fl
AMP-PNP-Efrapeptin is a peptide antibiotic that inhibits
oxidative phosphorylation as a consequence of a specific interaction with F1 (Lardy et al., 1975; Cross and Kohlbrenner,
1978). We have previously shown by equilibrium binding
measurements that efrapeptin blocks access of two out of the
three exchangeable nucleotide sites on FI tomedium nucleotide (Cross and Nalin, 1982). Because of this property, efrapeptin was used to furtherassess the natureof ADP-promoted
dissociation of FI.AMP-PNP. Although efrapeptin alone appears to have little effect on the stability of the single, tightly
bound AMP-PNP, it completely prevents ADP-promoted release in the presence (Table I) or absence (data not shown)of
Pi. These results indicate that ADP-promoted dissociation
requires a specific interaction of ADP with one or both of the
remaining catalytic sites notoccupied by AMP-PNP.
Inability of ATP and AMP-PNP
to Promote Rapid Dissociation of FI-AMP-PNP-It was of interest to determine
whether ATP can also promote AMP-PNP release from soluble F,. Duringa 2-min incubation, ATP causes approximately the same amount
of dissociation as thatobtained with
an equivalent amount of ADP. However, the effectiveness of
ATP is significantly reduced by the addition of increasing
concentrations of pyruvate kinase in the presence of phosphoenolpyruvate (Fig. 5). At 50 p~ ATP, dissociation of the F, .
AMP-PNP complex is actually prevented by addition of pyruvate kinase at the highest concentrations tested.
By several criteria, the ability of pyruvate kinase in this
experiment to prevent dissociation of FI. AMP-PNP
is due to
its ability to prevent accumulation of medium ADP rather
than to bind all medium nucleotide. Sephadex column centrifugation of samples containing 50 p~ [14C]ATP and the
-
Downloaded from www.jbc.org by guest, on October 30, 2009
ing a 2-min incubation. However, in this concentration range,
Pi causes a considerable enhancement of the rapid phase of
ADP-promoted release (Fig. 2). Using 100 PM ADP, amaximal
2.2-fold enhancement is obtained at 1 mM Pi, with a halfmaximal effect at 60 p~ Pi. In contrast,neither 1 mM sodium
sulfate nor 1 mM potassium acetate affects ADP-promoted
release of AMP-PNP under the conditions tested in Fig. 2.
Consistent with previous reports that AMP-PNPis a reversible inhibitor of F1 (Penefsky, 1974; Philo and Selwyn, 1974),
we find that therelease of AMP-PNP, promoted by ADP plus
Pi, results in a reactivation of the ATP hydrolysis activity
(data notshown).
release
The ability of Pi to enhanceADP-promoted
prompted additional experiments to determine whether dissociation of the F1.AMP-PNP complex might have an absolute requirement for medium phosphate. For this purpose,
all additions for measuring ADPreaction mixtures containing
promoted releaseexcept for F1.AMP-PNP were preincubated
with a phosphate-utilizing, coupled enzyme system to remove
any Pi that may contaminate the solutions. This treatment
had no effect on ADP-promoted release (Fig. 3, open square)
when compared to a control lacking the Pi-utilizing enzyme
system (Fig. 3, open circ2e). In order to establish the efficacy
of the Pi trap, its ability to prevent Pi enhancement of ADPpromoted release was also tested. The enhancement observed
with addition of up to 50 p~ Pi (Fig. 3, closed circles) was
eliminated by pretreatment of the reaction mixture with the
Pi-utilizing system (closed squares). These results establish
that medium Pi is not requiredfor ADP-promoted dissociation
of the F1. AMP-PNP complex.
Effects of Sequential Addition of ADP and P, on the
Dissociation of F1.AMP-PNP-The onset of the slow phase
of ADP-promoted release of AMP-PNP might result from the
accumulation of a nonproductive complex between promotor
and FI AMP-PNP. In the presence of both ADP and Pi,
maximal release of AMP-PNP may require ordered binding,
as hasbeen suggested for ATP synthesis by membrane-bound
F1 (Schuster etal., 1977). In view of these considerations, the
correct sequential addition of ADP and Pi might promote
more dissociation during the rapid phase than would either
simultaneous addition or sequential addition in the reverse
possibility, F1 .AMP-PNP
was incubated
order. TOtest for this
with 0.5 mM Pi (Fig. 4, open squares) or 100 p~ ADP (open
triangles) for 18 min before addition of the complementary
8057
8058
F1-ATPaseSubunit Interactions
-42.0
0
IO
20
30
40
TABLEI
The effect of efrapeptin on ADP + Pi-promoted dissociation of
FI.AMP-PNP
A complex containing 0.9 mol of [3H]AMP-PNP bound/mol of F1
was preincubated for 10 min a t 50 pgof Fl/ml in buffer containing
0.25 mgof albumin/ml with or without 5.9 p~ efrapeptin. At zero
time, the complex was diluted IO-fold in Mg-buffer containing albumin, hexokinase, and glucose as described in Fig. 1. Where indicated,
ADP, Pi, and 20 mM AMP were present. After 20 min, duplicate
aliquots were subjected to column centrifugation. The percentage of
the original complex that dissociates during the 20-min incubation
was determined. Control samples lacking bound AMP-PNP showed
that, at the
concentration of efrapeptin used, ATP hydrolysis activity
was inhibited by 96%.
%
’ release in 20 min
Addition
-EfraDeatin
+Efraoeotin
3
64
0
0
N o addition
1 m~ ADP, 1 mM P,
highest level of pyruvate kinase used in Fig. 5 showed that
less than 0.1%of the A T P is protein-bound. In addition, ATPdependent release is not decreased by pyruvate kinase in the
absence of phosphoenolpyruvate.
As an additional test for the ability of adenosine triphosp h a t e to promote dissociation of F1.[3H]AMP-PNP, the complex was dilutedinto buffer containing unlabeled AMP-PNP.
A slow release of radioactivity was observed over a 20-min
period (Table 11). However, trace amounts of ATP can form
spontaneously upon storageof A M P - P N P solutions (Penningroth et al., 1980) and A D P formed from such a contaminant
might be responsible forthe observed effect. Consistent with
this possibility, pyruvate kinase plus phosphoenolpyruvate
completely prevents dissociationof Fl r3H]AMP-PNP in the
presence of medium AMP-PNP (Table11).
3
I
5
15
IO
i
d
40
IIPyruvateKinase, ( rng/rnl)“
FIG. 5. Effect of pyruvate kinase on the rate of dissociation
of F1.AMP-PNP in the presence of ATP. A complex containing
0.9 to 1.0 mol of [3H]AMP-PNP bound/mol of FI was prepared and
resolved of unbound nucleotide as described under “Methods.” At
zero time, the complex was diluted to 4 pg of FJml with Mg-buffer
containing 0.25 mg of albumin/ml, ATP at 50 ,UM (O),100 p~ (A),or
200 p~ (H),10 m~ phosphoenolpyruvate, and pyruvate (Pyr) kinase
a t the concentration indicated. Duplicate aliquots were removed after
2 min and subjected to column centrifugation. Control samples containing ATP and phosphoenolpyruvate but lacking pyruvate kinase
were used to determine maximum release (1.0 on the y-axis): 19%
release at 50 p~ ATP, 30%release at 100 PM ATP, and 34%release at
200 p~ ATP.
TABLE
I1
The effect of medium AMP-PNP on the rate of dissociation of
F , .AMP-PNP
A complex containing 0.9 mol of [3H]AMP-PNP bound/mol of FI
was prepared and resolved of unbound nucleotide as described under
“Methods.” At zero time, the complex was diluted to 4 pg of Fdml
with Mg-buffer containing 0.25 mg of albumin/& and nonradioactive
AMP-PNP at the concentrations given. Where indicated, 10 mM
phosphoenolpyruvate and 0.13 mgof pyruvate kinase/ml were also
present. After 20 min, duplicate aliquots were removed and subjected
to column centrifugation. The percentage of the original complex that
dissociates during the 20-min incubation was determined.
% release in 20 min
AMP-PNP concentration
Without pyruvate ki- With pyruvate kinase
nase + phosphoenolpy- + phosphoenolpynvate
mvate
PM
0
10
25
100
250
1
2
8
15
19
4
0
1
1
0
~
AMP-PNPcomplexdemonstrates
that adenosinediphosp h a t e a n dPi can bind simultaneously atoform of the soluble
enzyme that has an adenosine triphosphate analog tightly
bound at a catalytic site. These results are interpreted as
support for a mechanism for ATP synthesis by membranebound F, that involves strong cooperative interactions between multiple copies of the catalytic subunit.
The binding change mechanism for ATP synthesis by F1ATPases during oxidative phosphorylation and photophosphorylation has two essentialfeatures. The fist is that product release (Boyer et al., 1973; Boyer, 1974) and substrate
binding (Boyer et al., 1975; Jagendorf, 1975) are associated
with energization of membrane-bound F,, while reversible
DISCUSSION
synthesis of A T P at the catalytic site occurs spontaneously
The most significant observationin the current study isthe (Boyer et al., 1973). The second is that the binding of subapparent absolute requirement for ADP binding in order t o strates and release of product occur simultaneouslyon separelease an ATP analog from soluble
F1under conditions where rate but interacting catalytic subunits (Kayalar et al., 1977).
single turnover events are measured. In addition, the syner- Net A T P synthesis requires coupling the exergonic oxidationgistic effect of A D P p l u Pi
s in promoting dissociationof the F1- reduction reactions of the respiratory chain t o the endergonic
-
Downloaded from www.jbc.org by guest, on October 30, 2009
Time, rnin.
FIG. 4. Effects of sequential addition of ADP and Pi on the
dissociation of Fl*AMP-PNP.A complex containing 0.9 mol of [3H]
AMP-PNP bound/mol of FI was prepared and resolved of unbound
nucleotide as described under “Methods.” At zero time, the complex
was diluted to 4 pgof F l / d with Mg-buffer containing albumin,
hexokinase, and glucose as described for Fig. 1. Also present were 0.5
m~ Pi (O), 100 p~ ADP (A), or 0.5 rm P, plus 100 p~ ADP (0).At
18 min, 100 p~ ADP (H) or 0.5 rm PI (A)was added as indicated by
the arrows. In each, 20 m~ AMP was added to the reaction mixture
at the same time as ADP. At the times indicated, duplicate aliquots
were subjected to column centrifugation and the effluents were analyzed for bound [3H]AMP-PNP. The fraction of the F1. AMP-PNP
complex remaining is plotted on a semilog scale against the time of
incubation.
Interactions
F1-ATPaseSubunit
'C. M. Nalin and R. L. Cross, unpublished results.
through subunit cooperativity, utilize the energy available
from the tight binding of substrate at one catalytic site to
disrupt strong interactionsbetween an adjacent catalytic subunit and tightly bound product.
It is evident that substrate binding is also required for
product release when catalysis proceeds in the reverse direction. By measuring the effect of ATP concentration on the
amount of oxygen exchanged per mol of Pi formed during
ATP hydrolysis, it was determined that the rate of ATP
binding influences the rateof ADP + Pi release (Hackney and
Boyer, 1978; Choate et al., 1979). In binding measurements,
Hutton andBoyer (1979)found that ATPpromotes release of
P, from mitochondrial F1 and Adolfsen and Moudrianakis
(1976) have reported that medium AMP-PNP increases the
rate of dissociation of ADP from bacterial Fl. Finally, Grubmeyer and Penefsky (1981) have shown that the formation of
medium product from a single tightly bound trinitrophenyl
analog of ATP requires the binding of substrate at a second
catalytic site.
The failure of the F1.AMP-PNP complex to dissociate in
the absence of medium ADP indicatesthat AMP-PNPis not
a simple competitive inhibitor with respect to ATP as indicated by standard kinetic analyses (Penefsky, 1974; Melnick
et al., 1975; Cross and Nalin, 1982). Although ATP andAMPPNP may actually compete for binding, reversal of inhibition
appears to require ADP. Hence, in coupled enzyme assays,
the apparent inhibition constants measured for AMP-PNP
may be influenced, in part, by the efficiency of the ATPregenerating system. In cases where the regenerating system
is of very high capacity, medium ADP is maintained below
the level necessary for reversal of AMP-PNP binding (Fig. 5 ) .
Similarly, the observations of Harris et al. (1978) that ATP
does not always reverse AMP-PNP inhibition of F1 and of
Schuster et al. (1975) that AMP-PNP inhibition can be noncompetitive under certain conditions may be due to their use
of high levels of pyruvate kinase.
Acknowledgments-We wish to thank Thomas Barra and Wendy
Wessels for excellent technical assistance during these studies.
REFERENCES
Adolfsen, R., andMoudrianakis, E. N. (1976) Arch. Biochem. Biophys. 172,425-433
Bensadoun, A., and Weinstein, D. (1976) Anal. Biochem. 70,241-250
Boyer, P. D. (1974) in Dynamics of Energy-Transducing Membranes
(Ernster, L., Estahrook, R. W., Slater, E.C., eds) pp. 289-301,
Elsevier, Amsterdam
Boyer, P. D. (1975) FEBS Lett. 58, 1-6
Boyer, P. D. (1979) in Membrane Bioenergetics (Lee, C. P., Schatz,
G., and Ernster, L., eds) pp. 461-479, Addison-Wesley, Reading
Boyer, P. D., Cross, R. L., and Momsen, W. (1973) Proc. Nutl. Acad.
Sei. U. S. A . 70,2837-2839
Boyer, P. D., Smith, D. J., Rosing, J., andKayalar,C.
(1975) in
Electron Transfer Chains and Oxidative Phosphorylation (Quagliariello, E., Papa, s., Palmieri, F., Slater, E. C., and Siliprandi, N.,
eds) pp. 361-372, North-Holland, Amsterdam
Chernyak, B. C., and Kozlov, I. A. (1979) FEBS Lett. 104,215-219
Choate, G . L., Hutton, R. L., and Boyer, P. D. (1979) J. Biol. Chem.
254, 286-290
Cross, R. L. (1981) Annu. Rev. Biochem. 50,681-714
Cross, R. L., and Boyer, P. D. (1975) Biochemistry 14, 392-398
Cross, R. L., and Kohlbrenner,W. E. (1978) J. Biol. Chem. 253,48654873
Cross, R. L., and Nalin, C. M. (1982) J.Biol. Chem. 257,2874-2881
Esch, F. S., and Allison, W. S. (1979) J. Biol. Chem. 254,10740-10746
Gruhmeyer, C., and Penefsky, H. S. (1981) J. Biol. Chem. 256,37283734
Hackney, D. D., and Boyer, P. D. (1978) J . Biol. Chem. 253, 31643170
Harris, D. A., Gomez-Fernandez,J. C., Klungsrayr, L., and Radda, G .
K. (1978) Biochim. Biophys. Acta 504,364-383
Hartree, E. F. (1972) Anal. Biochem. 48,422-427
Downloaded from www.jbc.org by guest, on October 30, 2009
binding changes. A transmembrane electrochemical gradient
(Mitchell, 1961) appears to function as an intermediate in
coupling these two processes (Boyer, 1975).
In the current study,
dissociation of a product analog from
soluble F, was made thermodynamically favorableby dilution
of the FI-AMP-PNP complex to a concentration below the
Kd. In this way, kinetic limitations involved in the product
release step could be examined in a simple, well defined
system.Consistentwith the tenets of the binding change
mechanism, release of the ATP analog does not occur a t a
measurable rate in the absence of medium ADP (Fig. 1).
Several lines of evidence establish that AMP-PNP release is
the result of interactions between multiple nucleotide sites
rather than being due to simple competition between ADP
and AMP-PNP for binding at a single site. Release of bound
['HIAMP-PNP is not observed in the absence of medium
nucleotide (Fig. 1) or in the presence of a large excess of
nonradioactive AMP-PNP (Table 11).In addition, ADP-promoted release is not enhanced by medium AMP-PNP (Fig.
I), butis prevented by efrapeptin (Table I) which blocks the
binding of medium nucleotide at thetwo remaining catalytic
sites not occupied by the single bound AMP-PNP (Cross and
Nalin, 1982). These observations, as well as the pronounced
specificity for ADP, are consistent with studies
of the medium
ATP + HOH oxygen exchange where release of ATP bound
to submitochondrial particles has been shown to require medium ADP (Kayalar et al., 1977).
The fact that Pi at low concentration is able to enhance
ADP-promoted release while having no effect in the absence
of ADP establishes the ability of ADP and Pi to bind simultaneously to F1.AMP-PNP. Since mitochondrial FI appears
to have three
copies of the catalytic subunit (Esch and
Allison,
1979; Todd et al., 1980; Cross and Nalin, 1982),we cannot rule
out the possibility that ADP and P, are bound at separate
catalytic sites. Nevertheless, if ADP and Pi do bind at the
same site, then the binding change mechanism would predict
some synthesis of F,-bound ATP during promoted release
(Boyer et al., 1973; Cross and Boyer, 1975). Using radioisotope-labeled ADP and Pi as promoters, we fail to detect bound
ATP in preliminary experiments using centrifuge columns.2
However, transient formation of ATP may occur on a time
scale not resolved by this technique, and further studies are
warranted.
It has previously been proposed thatsubstrate binding
during oxidative phosphorylation is an endergonic process
(Boyer, 1979).This is based in part on the lack of evidence for
simultaneous binding of ADP and P, to soluble F1 (Penefsky,
1977). However, previous studies utilized enzyme that lacked
a product analog bound at a catalytic site. In contrast to F1,
F,.AMP-PNP may resemble a conformational state of the
enzyme-product complex that binds ADP and Pi readily during normal catalysis. An additional argument that has been
made for an endergonic substrate binding step (Boyer, 1979)
comes from evidence linking substrate binding to the energization of membrane-bound F1 (Boyer et al., 1975; Rosing et
al., 1977). However, it appears essential that an energy-dependent conformational transition required for product release be tightly coupled to the binding of ADP and Pi at an
adjacent site. Otherwise, according to the binding change
mechanism, no ATP would be synthesized during the following portion of the enzyme cycle (Boyer, 1979; Cross, 1981). If
thesealternativeexplanations
are correct, thensubstrate
binding may actually be an exergonic process that supplements energy available from the oxidation-reduction reactions
in promoting product release. This may, in fact, illustrate a
general advantage enjoyed by oligomeric enzymes that,
8059
8060
Fl-ATPase Subunit Interactions
Hutton, R. L., and Boyer, P. D. (1979) J. Biol. Chem. 254, 99909993
Jagendorf, A. T. (1975) Fed. Proc. 34, 1718-1722
Kayalar, C., Rosing, J., and Boyer, P. D. (1977) J. Biol. Chem. 252,
2486-2491
Knowles, A. F., and Penefsky, H. S.(1972)J. Biol. Chem. 247,66176623
Koga, P. G., and Cross, R.L. (1982) Biochim. Biophys. Acta 679,
269-278
Lardy, H. A., Reed, P., and Lin, C. C. (1975)Fed. Proc. 34, 1707-1710
Melnick, R. L., Tavares de Sousa, J., Maguire, J., and Packer, L.
(1975)Arch. Biochem. Biophys. 166, 139-144
Mitchell, P. (1961) Nature (Lond.)191, 144-148
Penefsky, H. S . (1974) J.Biol. Chem. 249, 3579-3585
Penefsky, H. S.(1977) J. Biol. Chem. 252,2891-2899
Penefsky, H. S. (1979) Adv. Enzymol. Relat. Areas Mol. Biol. 49,
223-280
Penefsky, H. S., and Warner, R. C. (1965) J. Biol. Chem. 240, 46944702
Penefsky, H. S., Schwab, A., and Garrett, N. E. (1975) in Electron
Transfer Chain and OxidativePhosphorylation (Quagliariello,E.,
Papa, S.,Palmieri, F., Slater, E. C., and Siliprandi, N., eds) pp. 135147, North-Holland, Amsterdam
Penningroth, S. M., Olehnik, K., and Cheung, A. (1980)J. Biol. Chem.
255,9545-9548
Philo, R. D., and Selwyn, M. J . (1974) Biochem. J. 143, 745-749
Randerath, K., and Randerath, E. (1964) J. Chromatogr. 16,111-125
Rosing, J., Kayalar, C., and Boyer, P. D. (1977) J . Biol. Chem. 252,
2478-2485
Schuster, S. M., Ebel., R. E., and Lardy, H. A. (1975) J. Biol. Chem.
250, 7848-7853
Schuster, S. M., Reinhart, G. D., and Lardy, H. A. (1977) J . Biol.
Chem. 252,427-432
Smith, A. L. (1967) Methods Enzymol. 10,81-86
Todd, R. D., Griensenbeck, T. A., and Douglas, M. G. (1980) J. Biol.
Chem. 255,5461-5467
Yount, R. G., Ojala, D., and Babcock, D. (1971) Biochemistry 10,
2490-2496
Downloaded from www.jbc.org by guest, on October 30, 2009