THE JOURNAL OF BIOLOGICAL CHEMISTRY
Printed m U. S. A.
Vol. 257, No. 20, Issue of October 25. pp. 12101-121(16, 1982
Mechanism of ATP Hydrolysis by Beef Heart Mitochondrial ATPase
RATE ENHANCEMENTS RESULTING FROM COOPERATIVE INTERACTIONS BETWEEN MULTIPLE
CATALYTIC SITES*
(Received for publication, May 26, 1982)
Richard L. Cross$, Charles Grubmeyerg, and Harvey S . Penefsky
From the Department of Biochemistry, The Public Health Research Institute of the city of New York, h c . ,
New York, New York 10016
The very slow turnover rate for [y3’P]ATP hydroly-
sis at a single catalytic site on soluble mitochondrial
ATPase (F1) (uni-site catalysis) is accelerated over 10‘-
fold when additional nonradioactive ATP is added to
allow binding at multiple catalytic sites. This rate en-
hancement increases turnover to V,, with no detect-
able lag, thus demonstrating that the high affinity cat-
alytic site previously characterized (Grubmeyer, C.,
Cross, R. L., and Penefsky, H. S . (1982) J. Biol. Chem
257, 12092-12100) is a normal catalytic site.
When ATP binding at a second site is rate limiting
for hydrolysis approximately 0.6 mol of ATP per mol of
F1 is detected at catalytic sites. This level is predicted
from the equilibrium distribution of substrate and
products bound at a single site. The rate of ATP binding
to a second catalytic site (6 X 10‘ M” s-’) is the same
as the rate of binding at the first site. A K,,, = 30 PM and
a V,, = 300 s-’ were measured under conditions that
allow two sites to hydrolyze ATP. At higher ATP con-
centrations an additional K, = 150 PM and V,, = 600
s-’ were measured suggesting the possibility of three
functional catalytic sites on F1.
The rate constants for the reaction mechanism de-
veloped here and in the accompanying paper success-
fully predict a lag time of about 10 ms in the early
reaction kinetics of hydrolysis at 10 PM ATP. Hence,
prior to addition of substrate F1 appears to be in a fully
active form.
The 10‘-fold rate enhancement obtained when sub-
strate binds at a second catalytic site on the enzyme is
accompanied by a 30-fold increase in the rate of ATP
cleavage. However, the principal effect of substrate
binding at the second site is a 10‘-fold increase in the
rate of product release from the first site. These results
provide a dramatic example of the potential magnitude
of catalytic site interactions.
The molecular mechanism of soluble mitochondrial ATPase
is a problem of considerable importance to the role of the
membrane-bound enzyme in ATP synthesis during oxidative
phosphorylation. Under conditions for ATP hydrolysis that
allow turnover at only a single site on the molecule (uni-site
* This research was supported in part by Research Grants GM
21731 and 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.
$ On leave from the Department of Biochemistry, State University
of New York, Upstate Medical Center, Syracuse, New York 13210.
0 Present address, Department of Biology, New York University,
Washington Square East, New York, NY 10003.
catalysis) it was shown (1) that substrate binds rapidly with
very high affinity (K, = 1 O I 2 “I). Binding is followed by rapid
development of an equilibrium between bound ATP and
bound hydrolysis products, ADP and Pi. The equilibrium
constant is 0.5. These observations support the concept that
enzyme-bound ATP is formed from ADP and Pi with almost
zero change in free energy and that the primary function of
energy input during oxidative phosphorylation is to promote
release of product ATP from catalytic sites on F1’ (reviewed
in Ref. 2).
It is clear from evidence currently available that the mul-
tisubunit ATPase contains more than one catalytic site.
Chemical modification studies and measurement of the num-
ber of nucleotide-binding sites indicate that there are three
copies of the p subunit (3, 4) and that this subunit contains
the catalytic site (reviewed in Refs. 5 and 6). Direct evidence
for the presence of at least two catalytic sites on F1 was
provided by isotope-trap experiments with the ATP analog
TNP-ATP (7). Strong cooperativity between catalytic sites
also was demonstrable with both TNP-ATP and ATP. For
example, hydrolysis of TNP-ATP bound in only one catalytic
site was accelerated as much as 20-fold when a hydrolyzable
nucleotide was made available to the second site (8). Studies
of the I8O exchange between water and phosphate formed
during hydrolysis of ATP also are compatible with the pres-
ence of more than one catalytic site on the enzyme (9) and
with the suggestion of catalytic site cooperativity (10).
This paper undertakes an examination of the properties of
catalysis that result when ATP is made available to multiple
catalytic sites on the enzyme. Net turnover at a single site is
enhanced over 106-fold when substrate binds to additional
catalytic sites. These experiments also show that during
steady state hydrolysis of ATP at low concentrations, signifi-
cant and predictable levels of ATP are bound in catalytic
sites. Successful predictions of presteady state kinetics and
substrate modulation of the l8O exchange catalyzed by F1 are
made using the rate constants provided in this and an accom-
panying paper ( 1).
MATERIALS AND METHODS
ATP and CDTA were purchased from Sigma. 32Pi, enzyme grade,
was obtained from ICN. [y3’P]ATP was prepared as described by
Glynn and Chappell (11) with a specific activity in the range of lo5 to
lo6 cpm/nmol. The radiochemical purity of ATP was determined by
analytical thin layer chromatography as described (12). Adenine
nucleotide concentrations were determined from the absorbance of
solutions at 259 nm using a millimolar extinction coefficient of 15.4.
Radioactivity was measured by scintillation counting (13).
triphosphatase; TNP-ATP, 2’,3’-0-(2,4,6-trinitrophenyl)ATP; CDTA,
’ The abbreviations used are: F,, soluble mitochondrial adenosine
trans-1,2-diaminocyclohexane-N,N,N’,N’-tetraacetic acid.
12101
12102 Modes of Catalysis by R-ATPase
F, was isolated as described (14) and equilibrated on a centrifuge
column (13) with Mg buffer or CDTA buffer. Mg buffer contained
0.25 M sucrose, 40 mM 4-morpholineethanesulfonic acid, 40 mM Tris,
1 mM KzHP04, and 0.5 mM MgS04, pH 8. CDTA buffer was identical
with Mg buffer except that MgS04 was replaced by 2.5 m~ CDTA.
Immediately after column centrifugation, MgS04 was added in 0.5
mM excess to CDTA. As shown earlier (1) the adenine nucleotide
content of samples of Fl prepared in Mg buffer or in CDTA buffer
was 3.5 or 2.8 mol/mol of F1, respectively.
Quenched flow experiments were carried out in a Precision Syringe
Ram, model UI-1001 (Update Instrument, Inc., Madison WI). The
mixer and sampling valve configurations used in individual experi-
ments are described in detail elsewhere (1). Experimental details are
given in the figure legends.
The isotope trap technique (15) was used to measure [y-”PIATP
bound to catalytic sites on F1 in the experiments shown in Figs. 1 and
2. FI was first incubated with [Y-~~PIATP to form an F1. ”P complex,
followed by addition of a large excess of nonradioactive ATP. Reaction
was stopped by addition of perchloric acid, and 3>P, was measured as
described below. The calculated and measured efficiency of the trap
in preventing further binding of [Y-~~PIATP was excellent. Thus the
only [y-3ZP]ATP that could hydrolyze during the cold chase was that
which was already bound to the enzyme. This amount was determined
by comparison to the ”Pi present in a sample that was acid quenched
at the same time as the cold chase was initiated.
The kinetic constants for ATP hydrolysis reported in Table I were
measured in 1-ml reaction mixtures consisting of Mg buffer, 64 pg of
pyruvate kinase, 5 mM phosphoenolpyruvate, 10 to 2000 ELM [y-”P]
ATP (together with equimolar MgS04), and F,. The amount of F,
added was sufficient to hydrolyze about 10% of the ATP in 25 s. The
reaction was stopped by addition of 100 pl of 60% perchloric acid. At
each concentration of ATP tested, a time course was constructed
between 0 and 25 s in 5-s intervals. The linear portion of the time
course was used to obtain the rate of hydrolysis.
”P, was determined by the method of Sugino and Miyoshi (16),
using modifications introduced earlier (8) and with corrections for
recovery of ”Pi. F1 protein was determined by a modified Lowry
procedure (17) or a modified biuret procedure (18). A molecular
weight of 347,000 (19) was used in all calculations.
RESULTS
In an accompanying paper, we show that ATP bound in a
single very high affinity catalytic site on F1 (K , = 10” “I) is
in rapid equilibrium with its bound hydrolysis products ADP
and Pi (1). However, net hydrolysis of bound ATP, that is
product release to form free enzyme, is very slow ( s-’ (1)).
In contrast, the addition of excess nonradioactive ATP caused
rapid hydrolysis and product release from the high affinity
catalytic site, thus clearly distinguishing this site from the
noncatalytic nucleotide-binding sites on the enzyme. In the
present study, the ability of medium nucleotide to promote
turnover of the high affinity catalytic site was used to deter-
mine whether the site is a normal catalytic site capable of
hydrolyzing ATP at the rapid steady state rates (600 s-’)
observed in the presence of saturating concentrations of sub-
strate. A complex was formed by incubating 0.15 PM [ y - 3 “ ]
ATP with a 10-fold molar excess of F1 for 2 s in the quenched-
flow apparatus. These are conditions for observing uni-site
catalysis by F1 (1). The F1 -32P complex was then mixed in a
second mixer (Fig. 1, zero time) with a large excess of nonra-
dioactive MgATP (final concentration 3.3 m ~ ) . The reaction
was allowed to proceed for periods of 5 to 40 ms before
quenching by injection into perchloric acid. It may be seen
that about 20% of the added [y-32P]ATP was hydrolyzed
during the initial 2-s incubation (zero time). This amount of
hydrolysis is consistent with rapid establishment of an equi-
librium between bound substrate and products (1). It may
also be seen in Fig. 1 that the bulk of the promoted hydrolysis
was very rapid and occurred before 5 ms of chase time had
elapsed. It is clear that at least 2 reaction half-times must
have elapsed during the fist 5 ms giving a minimum rate
constant of 300 s-’ for the fraction of ATP that hydrolyzes in
t
20 4Y 2 = = h r 0.3 M ATP32 3 IO mM ATP 1
I H -
I I I I I 1
0 5 IO 20 30 40
AGING TIME (mrec)
FIG. 1. ATP-promoted hydrolysis of [y-‘”PIATP bound in a
single site on FI. The experiment was carried out in a quenched-
flow apparatus in the push-push mode and employing two mixers.
Each of the 3 syringes contained 2.5 ml of Mg buffer with 10 mM P,
and in addition: syringe 1, 3 p~ FI; syringe 2, 0.3 p~ [Y-~’P]ATP;
syringe 3, 10 m~ MgATP. In the fwst push, equal volumes from
syringes 1 and 2 were mixed in mixer 1 and allowed to age for 2 s in
the length of hose designated “A” between mixers 1 and 2. During
this time, the F1.32P complex was formed. The second push flushed
the contents of hose A through the second mixer, where mixing with
10 mM nonradioactive ATP occurred. The reaction mixture then
passed through a second aging hose into a vial containing 0.2 ml of
60% perchloric acid, 0.8 ml of Mg buffer, 0.1 ml of 100 m~ MgATP.
The speed of the second push determined the aging time shown on
the abscissa. This is the time elapsed between mixing the F1-”P
complex with nonradioactive ATP and quenching in acid. The point
representing zero aging time was obtained by disconnecting the sec-
ond mixer and collecting the outflow of the first mixer directly in the
perchloric acid quench. A 5-s time point (see text) was obtained by
collecting the outflow of the second mixer in a vial containing Mg
buffer and MgATP. Perchloric acid was then added 5 s later. A
control experiment indicated that approximately 3 to 6% of the chase
ATP was hydrolyzed during the 5-51 reaction. 32Pl formed was deter-
mined as described under “Materials and Methods.” The data are
expressed as per cent hydrolysis of the added [y3’P]ATP.
this rapid phase. Since the turnover rate for net hydrolysis a t
a single site on F1 is s” (I), the acceleration in rate caused
by binding additional ATP is over 106-fold. Approximately
20% of the [y3’P]ATP initially added was hydrolyzed at a
slower rate between 40 ms and 5 s (data not shown) indicating
some heterogeneity in the FI . 32P complex. Since the biphasic
nature of the ATP-promoted hydrolysis was influenced by P,
and MgS04 concentration, the F1. 32P complex may be subject
to ligand-induced asymmetries. About 13% of the added [ y -
32P]ATP did not bind at catalytic sites prior to addition of the
nonradioactive ATP.
In view of the substrate-product equilibrium at a single site
( K = 0.5, Ref. l), one might expect that during steady state
hydrolysis, under conditions such that the rate-limiting step
is the rate of binding of ATP to a second site, catalytic site
occupancy by ATP would approach a value of 0.7. Fig. 2
shows an experiment in which the isotope trap technique (15)
was used to measure [y-”’P]ATP bound to catalytic sites on
the enzyme. At 20 n~ substrate and 2 n~ F,, a lag of about 2
s preceded the subsequent linear rate of hydrolysis (lower
curue in Fig. 2 A , “acid quench”). The duration of the lag is
consistent with the time required to build up a steady state
level of enzyme-nucleotide complex. The difference between
the ”Pi measured in the cold chase curve and in the acid
quench curve is a direct measure of [Y-~’P]ATP bound to
catalytic sites and committed to hydrolysis. The values for
bound ATP are plotted in Fig. 2B. It may be seen that a
steady state level of about 0.6 mol of [y3’P]ATP bound per
mol of F1 was reached.
Modes of Catalysis by Fl-ATPase 12103
3.01
-
LL-
5 2 .0 -
E
ACID QUENCH
0 5 IO 15 20
REACTION TIME (sec)
FIG. 2. ATP bound to catalytic sites on F1 during steady
state hydrolysis at 20 n~ ATP. A, reaction mixtures contained, in
a final volume of 1 ml , 20 IIM [Y-~’P]ATP in Mg buffer (pH 8.0). The
reaction was started by addition of 2 pmol of Fl in 20 pl of buffer. At
the indicated times the reaction was quenched by addition of 150 p1
of 60% perchloric acid, followed by 100 pl of 100 mM MgATP (0, m),
or chased by addition of the MgATP first, followed 5 s later by the
perchloric acid (0, 0). The presence of 10 pmol of nonradioactive
ATP provided an effective isotope trap for [y-”PIATP not bound in
catalytic sites since only 6 nmol of ATP was hydrolyzed during the 5-
s cold chase. 32P1 was measured as described under “Materials and
Methods.” Circles and squares represent 2 separate experiments. B,
the difference between the cold chase and acid quench samples is
plotted as a measure of [y-32P]ATP bound to catalytic sites on FI.
The rate of hydrolysis of 20 I” ATP in Fig. 2 (0.16 mol of
Pi/mol of Fl/s) is limited by the rate of binding of ATP to a
second site on F1. This point is substantiated by the linear
dependence of the rate of hydrolysis on the ATP concentra-
tion at substrate levels between 5 and 60 I“ (Fig. 3 ) . The rate
of ATP binding to a second site was determined from the
slope to be 6.3 X lo6 M”.s”. This is the same as the rate of
binding to the first site (I).
The rate constants for the reaction mechanism developed
previously (1) and in this paper were tested for their ability to
predict the early reaction kinetics of ATP hydrolysis. The
dashed line in Fig. 4 represents a computer-assisted numerical
integration of the difference equation for Scheme 1. The rates
of binding of ATP in steps 1 and 3 were taken to be equal (6
X lo6 M” . s-’, Fig. 3 and Ref. 1). The rate constants for the
reversible hydrolysis of a single ATP on the enzyme (step 2)
were 10 s-’ in the forward direction and 20 s-’ in the reverse
(1). The rate of steps 4 plus 5 was taken to be 300 s-l as
discussed below (Table I).
It may be seen in Fig. 4 that the steady state rate at 10 PM
ATP is the same for both the predicted and the experimental
curves (about 50 turnovers per s). This rate is identical with
one obtained in experiments done by hand under the same
conditions but over longer time periods. The experimental
time course (Fig. 4) shows a lag of about 10 ms. This is in
good agreement with the predicted lag time of about 14 m s . A
number of experiments of this kind were carried out at ATP
concentrations in the range of 5 to 100 IJM. In all cases the lag
time was less than 20 ms.
In order to investigate further the hydrolysis of ATP at
multiple catalytic sites on the enzyme, the substrate depend-
ency of the reaction rate was measured over a concentration
range from 10 to 2000 IJM ATP. The nonlinearity of the
Lineweaver-Burk plots obtained from these measurements is
in agreement with the apparent negative cooperativity re-
ported by Ebel and Lardy (20). The K,,, and VmaX values
calculated from the linear portions of the plots are shown in
Table I. For purposes of comparison the Kd and turnover rate
for hydrolysis of ATP in a single site (1) are shown in line 1
of the table. The turnover rate of uni-site catalysis is too slow
0.4 I
0 2 0 40 60
ATP CONCENTRATION (nM)
FIG. 3. Rate of binding of ATP at a second site on F1. The
reaction mixtures contained 0.25 to 3 pmol of F1 and a 20-fold molar
excess of [Y-~’P]ATP in 1 ml of Mg buffer. The rate of hydrolysis was
measured by acid quenching seven samples at various times over a
period sufficient to give about 2.5 turnovers (8 to 80 s). 32P, was
measured as described (see under “Materials and Methods”). The
turnover rate is plotted versus the ATP concentration.
8 - -
LL-
1
0 50 100 150 20(
AGING TIME (rnsec)
FIG. 4. Early reaction kinetics of ATP hydrolysis by F1. The
experiment was carried out in a quenched flow apparatus. The single
mixer configuration was employed with (0) and without (X) the
sampling valve described (see under “Materials and Methods”). Both
syringes (A and B) contained 2.5 ml of Mg buffer, pH 8, and in
addition, A, 0.5 p~ FI; B, 20 p~ [y-32P]ATP. The velocity of the single
push controlled the aging time shown on the abscissa. The outflow of
the aging hose was collected in vials containing 0.15 ml of 60%
perchloric acid and 1 ml of Mg buffer. At the end of the experiment,
5O-pl samples were taken from syringe B to measure 32P, at zero time.
32P, was determined as described (see under “Materials and Meth-
ods”). -, drawn through experimental points; - - -, represents a
computer-assisted numerical integration of the difference equation
for Scheme 1 as described in the text.
ATP
-2112 ATP 1 4
ADP+ Pi
.ADP+i\ .ADP.Pi 1 ~
FI 3b !‘ATP ‘‘ATP F F
SCHEME 1
12104 Modes of Catalysis by K-ATPase
TABLE I
Kinetic constants for ATP hydrolysis at multiple catalytic sites
Mode of catalysis K d or K , V,,
M S-1
Uni-site lo"*
Bi-site" 3 X 300
Tri-site" 1.5 X 600
a Reaction kinetics were measured as described under "Materials
and Methods." The rates were analyzed usi