1067
Product Inhibition of the Actomyosin
Subfragment-1 ATPase in Skeletal, Cardiac,
and Smooth Muscle
Jean S. Drew, Vijay A. Harwalkar, and Leonard A. Stein
Downloaded from http://circres.ahajournals.org/ by guest on June 18, 2017
We studied product inhibition of the actin-activated ATPase of myosin subfragment-1 (S-i) from the three
types of muscle tissue: skeletal, cardiac, and smooth. Increasing levels of [MgADP] in the 0-1-mM range
caused significant inhibition of the actin-activated MgATPase activity of cardiac and gizzard but not
skeletal muscle S-1. When total nucleotide concentration ([ATP]+ [ADP]) was kept constant at 1 mM,
ATPase activity was inhibited by 50% at an ADP/ATP ratio of 6: 1 for cardiac S-1 and 3:1 for gizzard S-1.
For skeletal S-1, however, even a 19:1 ratio did not cause 50%o inhibition of ATPase activity. The observed
effect was not due to changes in pH or inorganic phosphate concentration, nor could it be explained by
substrate (ATP) depletion. In the absence of actin, ADP had little or no inhibitory effect on the ATPase
activity of S-1, and these observations imply that ADP is competing directly for the ATP binding site of the
actin-Si complexes of cardiac and smooth muscle S-1. ADP has previously been shown to be a weak
competitive inhibitor of the ATPase activity in skeletal muscle. The current data imply that ADP is a very
effective competitive inhibitor for the actin-activated ATPase activity of cardiac and gizzard S-1 and,
therefore, that ADP may be a physiologically important modulator of contractile activity in cardiac and
smooth muscle. (Circulation Research 1992;71:1067-1077)
KEY WoRDs * ADP * subfragment-1 * ATPase * gizzard * cardiac muscle
C ontraction of both striated and smooth muscle
results from the cyclic interaction of myosin
and actin. This interaction is driven by the
hydrolysis of ATP to ADP and inorganic phosphate (Pi)
by the myosin ATPase. To gain an understanding of the
actin-myosin interaction from a molecular standpoint,
biochemists have turned to the biochemical kinetics of
the actin-activated myosin ATPase activity, which is the
in vitro correlate to contraction in vivo. Kinetic studies
of the myosin ATPase activity of striated myosin have
led to the development of a six-state kinetic model
capable of accounting for the great majority of the
available pre-steady-state and steady-state data' (see
Figure la), and recent physiological data also appear to
support this model.2-5 In the six-state model, when
actomyosin subfragment-1 (acto-S-1) is added to ATP, a
rapid equilibrium (equilibrium binding constant K3,
Figure la) is established between actin-bound and
actin-free subfragment-1 (S-1)-ATP complexes (i.e.,
M-ATP, where M is S-1; Figure la). Hydrolysis of ATP
then occurs on the surface of the myosin molecule in
From the Departments of Medicine and Physiology, State
University of New York at Stony Brook.
Supported by American Heart Association, New York State
Affiliate, Inc., grant 90-OllG (L.A.S.), and American Heart
Association, New York State Affiliate, Inc., postdoctoral fellowship 89-204F (J.S.D.). L.A.S. was the recipient of a Research
Career Development Award from the National Institutes of
Health.
Address for correspondence: Dr. Jean S. Drew, Laboratory of
Molecular Cardiology, HSC T17, Room 090, Department of
Medicine, SUNY at Stony Brook, Stony Brook, NY 11794-8171.
Received January 6, 1992; accepted June 19, 1992.
both the actin-free and actin-bound pathways (see
Figure la, M-ATP <-> M-ADP-Pi,, and A-M-ATP
A-M-ADP-Pi,, respectively, where A is actin), and the
rate of this "rapid" hydrolysis is very fast compared with
the steady-state ATPase rate. After this rapid hydrolysis, a conformational change occurs within the myosinATP complex (M-ADP-Pil
M-ADP-P12 and A-MADP-Pi,-> A-M-ADP-Pi2), and this transition is rate
limiting to steady-state hydrolysis. After this conformational change, the products of hydrolysis, ADP and Pi,
are released from the myosin surface, and the cycle
begins again in the presence of ATP. In this model,
actin activates the ATPase activity by binding to the
M-ADP-Pi complex and by increasing the rate of product release (i.e., product release is much slower in the
absence of actin).
To simplify the mathematical modeling of the sixstate model, it is usually assumed that the "on" rate of
ATP to acto-S-1 is very rapid and that the "of' rate for
the products of hydrolysis is not rate limiting for hydrolysis. Hence, there is no significant population of the
states A-M or A-M-ADP during the steady-state hydrolysis of ATP. However, routine studies in our laboratory
on the steady-state hydrolysis of ATP by acto-S-1 of
cardiac and smooth muscle origin show that the actinactivated ATPase activity decreases with time as the
reaction progresses. This is not due to the protein
denaturing, because addition of ATP returns the rate
toward normal. Skeletal S-1 shows a fall in rate when
the substrate is virtually exhausted, but in the cardiac
and smooth muscle systems, the rate falls at ATP
concentrations normally considered to be saturating.
Circulation Research Vol 71, No 5 November 1992
1068
a^}
ZI-ASX!
tt4
tS
^
urn-w1
*
W
()
tS
t4
Z3
-n
Downloaded from http://circres.ahajournals.org/ by guest on June 18, 2017
1-E-AP *
i-u-mA-i1
acto-S-1 found that ADP release was an order of
magnitude slower in cardiac than in skeletal acto-S-1.13
However, the "off" rate measured was still one to two
t
orders of magnitude greater than the V,,,,, of the steadystate ATPase activity and, therefore, was not rate
limiting for the ATPase cycle.
jL-jK-jol3.-P:L2
,
The purpose of the present study was to evaluate
ADP inhibition of the actin-activated ATPase of isolated myosin S-1 from the three distinct muscle tissues:
cardiac, skeletal, and smooth. Smooth muscle S-1 has
1ZD1111
traditionally been neglected as a kinetic model for
smooth muscle myosin. However, we have observed
recently that unphosphorylated gizzard S-1 has an actinN-A
activated Vma,, similar to that of phosphorylated heavy
*
=pmeromyosin (HMM) and an MgATPase rate in the
absence of actin that is comparable to that of phosphortffiylated HMM in the extended conformation.'4,15 Thus,
we believe that S-1 can indeed be a useful model, along
with HMM, for investigating actin-myosin interactions
*
u-Dw-u2
*
&-M-IDP-1
in smooth muscle.
Materials and Methods
kA
.
ci!WJ
t'
k.AP
i-K~ ~
b.
W
Pi
t
cDwi
FIGURE 1. Six-state model describing the molecular interactions of actin-activated myosin A TPase activity. M, subfragment-l; A, actin; Pi, inorganic phosphate. Subscripts 1 and 2
indicate a conformational change in the M-ADP-Pi or A-MADP-P, that is rate limiting for actin-activated ATPase activity. Panel a: Model in the absence ofinhibitory concentrations
ofADP. Panel b: Model in the presence of inhibitory concentrations of ADP. K3, actin-S-1 equilibrium binding constant;
kATp, ATP "on" rate for acto-S-1; k ATp, ATP "offl' rate for
acto-S-; kADp, ADP "on" rate for acto-S-1; k ADP, ADP "off'
rate for acto-S-1.
These findings suggest that product inhibition is responsible for the fall in ATPase activity observed for cardiac
and smooth muscle acto-S-1, and this was the impetus
for the present studies. One possible model for product
inhibition is that ADP is competing directly with ATP
for the ATP binding site. Because of the data presented
herewith, we favor this model as shown in Figure lb.
Recently, there have been several reports of the
effects of ADP on the actin-myosin interaction. In prior
solution studies, ADP appeared to be a weak competitive inhibitor of the skeletal acto-S-1 MgATPase.6 Inhibition by ADP has also been reported for thiophosphorylated gizzard myosin using an in vitro motility assay.7 In
cardiac muscle in vivo, hypoxic ventilation is accompanied by severe depression of cardiac function as well as
a 50% increase in cardiac tissue ADP level.8 In smooth
muscle cells in vitro, ADP increases tension development and slows down crossbridge detachment.9-11 Also,
the local concentration of ADP at the contractile apparatus may rise significantly during contraction in vivo.8,12
These findings suggest that ADP inhibition may be a
physiologically important modulator of ATPase activity
in cardiac and smooth muscle tissues. Prior kinetic
studies on the rate of ADP release from bovine myosin
Proteins
Actin and skeletal myosin were prepared from rabbit
skeletal muscle,16 and cardiac myosin was purified from
swine cardiac tissue.' Smooth muscle myosin was isolated from frozen turkey gizzards.17-19 S-1 from skeletal
and cardiac myosin was obtained by chymotryptic digestion' and from gizzard myosin by papain digestion.19 We
have also prepared gizzard S-1 by digestion with Staphylococcus aureus protease.20 The two gizzard subfragments differ in that the 20-kd regulatory light chain
(LC20) and the 26-kd S-1 heavy chain fragment were
cleaved by papain but not S. aureus protease, and this
cleavage was associated with threefold to fourfold
strengthened actin affinity for papain-digested S-1.14
However, there were no differences in the effects of
ADP on papain versus S. aureus protease-digested S-1
kinetics. In the present study, we report the results for
papain-digested S-1.
Chymotrypsin was obtained from Worthington Biochemical Corp., Freehold, N.J., and papain was obtained from Sigma Chemical Co., St. Louis, Mo.
Kinetic Studies
ATPase assays were performed as described previously by (y-_"P)ATP hydrolysis assay.3 ATPase data
were analyzed by a hyperbolic least-squares fitting
procedure.3,2'
For the inhibition assays, mathematical analysis of
Figure lb predicted a simple hyperbolic dependence of
ATPase activity on [ATP], if the experiments were
performed with the sum of [ADP] and [ATP] kept
constant. Thus, instead of performing experiments at
several fixed levels of [ADP] and varying [ATP], we
varied both [ADP] and [ATPI from 0 to 1 mM, such that
the total nucleotide concentration, [ATP]+ [ADP], was
kept constant at 1 mM and the ionic strength was kept
constant at 13 mM. In all experiments for which [ADP]
and [ATP] were varied, the reported concentrations
were corrected for ATP split during the experiment. It
is of note that the data obtained in this way will be
applicable to the determination of the model in Figure
lb, once techniques are devised to measure kinetic rate
Drew et al ADP Inhibition of Acto-S-1 ATPase
1069
TABLE 1. Steady-State Actin-Activated Mg2` ATPase Activity of Skeletal, Cardiac, and Gizzard
Subfragment-1
Vmax
KATPase
KBinding
Temperature
Muscle type
(sec-')
(,uM actin)
(OC)
(,uM actin)
Skeletal
3.33
2.7
11.6
15
Cardiac
1.74
5.1
26.0
15
Smooth
1.40
10.8
26.5
25
Vma,c, maximum velocity; KATPaS., activation constant of actin-activated ATPase activity; KBinding,
dissociation constant of subfragment-1 to actin.
Conditions were as follows: 10 mM imidazole (pH 7.0), 1 mM MgCl2 (free), 0.5 or 1 mM MgATP, and
1 mM dithiothreitol; ionic strength maintained at 13 mM with added 2.5 mM KCl when necessary.
constant k ATP directly and accurately (see Figure lb),
and this will be the subject of a subsequent report.
Mathematical Analysis of Inhibition Data
Downloaded from http://circres.ahajournals.org/ by guest on June 18, 2017
The data presented are shown to fit hyperbolic kinetics quite well, and it is of interest to use the hyperbolic
fitting parameters to determine the ratio of [ADP]/
[ATP] that gives 50% inhibition of the ATPase, ([ADP]/
[ATP])50%. If it is assumed that the total nucleotide
concentration, [ADP]+[ATP], is 1 mM, our model
predicts that the hyperbolic dependence of ATPase
activity on [ATP] can be written as
1/VATp=Fl+F2(l/[ATPI)
(1)
where VATP is the ATPase rate and F1 and F2 depend on
the kinetic parameters in Figure 1, the ratio of the rapid
equilibrium binding constants KADp/KATp, and the kinetic rate constants k ADP and k ATP. F, and F2 are
experimentally determined from the data using a leastsquares fitting procedure based on Nelder and Mead.21
In our system, the maximal ATPase rate occurs, by
definition, at [ATP] = 1 mM and [ADP] =0. Substituting
1 for [ATP] in Equation 1, we find the maximum rate to
be given by
V1 mM ATP = 1/(F1 + F2)
Hence, the [ATP] that gives half the maximum rate can
be determined from
Vl/2=Vl
mM
ATP/2= 1/[2(F1 +F2)]=[ATP]/(F2+F,[ATP])
At 50% inhibition we find that
(rADPW[ATP])so%=(F1 +F2)/F2
(2)
Another way to view the system is as a simple competitive
inhibition system with the
phenomenological equation
V=Vma[SV(Km(l + [I]/Ki) + [5])
(3)
where V is the observed ATPase rate, Vm,. is the rate
extrapolated to infinite [ATP], [S] is [ATP], [I] is [ADP],
Km is the apparent activation constant of the ATPase
activity for ATP, and Ki is the apparent inhibition constant for ADP. If we then require that [ATP] + [ADP]= 1
and assign F, and F2 analogously to Equation 1, i.e., if
F1=(1-Km/Ki)/Vm and F2=(Km+KmIKi)/Vm, we find
that
KmIKi=F2/(F2+F) -FlKm/(F2+FF)
(4)
If Km can be assumed to be very small (i.e., less than 0.01
mM), then Ki/Km is approximately equal to ([ADP]/
[ATP])5o%. Note that the ratio Km/Ki is not the ratio of
the ADP and ATP binding constants K"p/KATp but is
related to this ratio in a complex manner. The relation
involves k ATP and k ADP, the ATP and ADP "off' rates,
respectively, as well as all the other rate constants in the
model given in Figure lb.
Results
Kinetic Characterization of Skeletal, Cardiac, and
Smooth Muscle Proteins
In Table 1, the kinetic constants Vm., KATPase (the
activation constant of actin-activated ATPase activity),
and KBjndifg (the dissociation constant of S-1 to actin) for
the skeletal, smooth, and cardiac muscle preparations
used are listed. The constants given for skeletal and
cardiac muscle are in agreement with previously reported values from this laboratory, whereas the values
for smooth muscle acto-S-1 are the first to be reported
from this laboratory.
Inhibition of Actin-Activated A TPase Activity During
Hydrol,ysis Over Time
Figure 2 shows the progressive inhibition of the
actin-activated ATPase of cardiac S-1 as the hydrolysis
of 1 mM ATP goes to completion. The ATPase activity
was reduced by 20-30% of its initial value when 50% of
the ATP had been hydrolyzed and by 50% when 80% of
the ATP had been hydrolyzed. A concentration of 0.2
mM ATP would be near saturating for cardiac S-1 in the
absence of added ADP,22 yet significant inhibition of
ATPase activity was observed. The observed inhibition
was unlikely to be secondary to a change in pH, since
only a small drop in pH was observed in these experiments (approximately a 0.2-pH drop after 1 mM ATP
was hydrolyzed to completion). However, measurement
of pH in the presence of relatively concentrated protein
solutions is difficult. Hence, we decided to control pH
directly by hydrolyzing less than 0.1 mM ATP in each
experiment, using the time course of complete hydrolysis of 1 mM ATP (Figure 2) as a control value.
ADP Inhibition of Skeletal, Cardiac, and Gizzard S-]
ATPases: Eliminating the Effect of Variable pH and
Ionic Strength
The next set of experiments was designed to mimic
the presumed changes in [ADP] over time, as seen in
Figure 2, by individually controlling [ADP] and [ATP].
Since we wished to compare all the experiments performed in a quantitative fashion, it was mandatory to
keep the ionic strength and pH constant. In addition, it
was important to work at maximal extraction efficiency
in the ('y-32P)Pi extraction procedure. These goals were
accomplished as follows. First, we performed the inhi-
Circulation Research Vol 71, No 5 November 1992
1070
1.75
-
FIGURE 2.
Graph showing actin-activated
Mg2`ATPase of cardiac subfragment-1: Hydrolysis of 1 mMATP to completion. ATP to 1 mM
was added to 1 ,uM cardiac subfragment-1 and
20 (x -x), 10 (+ --- +), 5 (a-m), 3.3 (x --- x),
or 2.5 (+ - +) ,uM actin. The ATPase rate was
monitored until hydrolysis was nearly complete
(approximately 1 hour). Lines are hyperbolic
least-squares fits of the data. Conditions were as
follows: 10 mMimidazole (pH 7.0), 2 mMMgCl2
(1 free), 1 mM dithiothreitol, and 1 mMATP, at
15'C.
04
0
43
U
4
Downloaded from http://circres.ahajournals.org/ by guest on June 18, 2017
Fraction ATP Hydrolyzed
bition assays with only very limited ATP hydrolysis at
each level of [ADP] and [ATP]. This led to a high
extraction efficiency of hydrolyzed Pi in the isobutanol/
toluene extraction assay and circumvented any significant pH change. Second, ionic strength was kept constant by correcting the assay medium at each level of
[ADP] and [ATP] with KCl or KPi.
Another important feature in the design of the experiments was that, in order to maximize the interaction of
actin and myosin, it was necessary to keep ionic strength
low.23 This requirement severely limited the range of
possible nucleotide and salt concentrations (including
[KPiJ and [ADP]) that could be used. Therefore, instead of either monitoring complete hydrolysis of 1 mM
ATP as in Figure 2 or keeping [ADP] constant while
varying [ATP], we varied [ADP] and [ATP] from 0 to 1
mM such that the total nucleotide concentration,
[ATP]+[ADP], was kept constant at 1 mM. The ionic
strength was kept constant at 13 mM by using either KCl
or KPi with the following assumptions: the ionic
strengths of 1 mM MgATP and 1 mM MgADP are 5.0
and 4.0 mM, respectively, and the ionic strengths of 1
mM KPi or KCl are 1.6 mM and 1 mM, respectively. In
these experiments, less than 0.1 mM ATP was hydrolyzed at each particular [ADP] and [ATP], and the pH
remained constant in the presence of 6-10 mM imidazole. In all experiments for which [ADP] and/or [ATP]
were varied, the reported concentrations were corrected
for ATP split during the experiment.
Figure 3 shows the normalized actin-activated ATPase
for each protein, with the actin adjusted to a level that
gave approximately half-maximal activation in the absence
of added ADP (see Table 1). This provides a direct
comparison of cardiac, smooth, and skeletal acto-S-1 at
the KATPae of each protein tested, i.e., at the same level of
activation by actin. These data indicate that for a given
level of actin activation, ADP inhibits the ATPase of
cardiac and smooth muscle S-1 to a much greater extent
than it does skeletal muscle S-1. Both cardiac and gizzard
S-1 exhibited significant ATPase inhibition at low levels of
[ADP] (' 0.5 mM), whereas skeletal S-1 did not. In
addition, these data demonstrate that the inhibition observed in Figure 2 cannot be explained by changes in ionic
strength or pH, since both were controlled in the experimental design.
ATPase Inhibition Related to Substrate
(ATP) Depletion
We next attempted to determine how much of the
observed inhibition could be due to substrate (ATP)
depletion alone. In a separate set of paired experiments,
[MgATP] was varied between 0.1 and 1.0 mM in the
presence or absence of added MgADP, and ionic
strength was maintained at a constant level using KCl.
Actual [ATP] varied from 0.95 to 0.05 mM, once the
correction was made for ATP split during the experiment. There was little or no decrease in ATPase activity
in the absence of added ADP, even at [ATP] as low as
0.05 mM (Figure 4, upper curves). Thus, substrate
depletion alone could not explain the reduced ATPase
rates observed in the presence of increasing levels of
[ADP] (Figure 4, lower curves).
ATPase Inhibition Related to Increased [PJ
We then tested whether Pi could be responsible for
part of the observed inhibitory effect on the ATPase
activity. The strategy here was to detect an inhibitory
response to [Pi] by working under conditions involving
[ADP] in which significant inhibition was evident (i.e.,
at >0.5 mM ADP), and then vary [KPi] by exchanging
KPi with KCI while keeping ionic strength constant. In
this case, if Pi were responsible for at least part of the
inhibition, a definite change in ATPase activity would
be observed. Figure 5 shows that 1.5 mM KP, had no
significant effect compared with 2.4 mM KCI on the
acto-S-1 ATPase activity for any of the three muscle
types. Neither Vmax nor KATp.,e was affected by KPi.
Thus, increasing [Pi] was not responsible for the inhibition observed in Figure 2. Other studies using cardiac
S-1 and KP1 concentrations up to 2.5 mM (requiring an
Drew et al ADP Inhibition of Acto-S-1 ATPase
1.20
4i.1I
04
4J
Ii
'4
U
pi
Elf
'4
.4
Downloaded from http://circres.ahajournals.org/ by guest on June 18, 2017
[ADP] (mM)
imidazole concentration of only 3 mM to keep ionic
strength constant) showed no inhibition of ATPase
activity by KPi (data not shown).
ADP Inhibition of the Actin-Activated ATPase
Activity at Different Actin Concentrations
Figure 6 shows plots at "low," "intermediate," and
"high" actin concentrations for the three S-1 types. The
choices of actin concentrations were made to see if actin
had any significant effect on the inhibition observed.
The data were analyzed using a least-squares hyperbolic
fitting program.2' This procedure determined the values
for F, and F2 from Equation 1 (see "Materials and
Methods"). Table 2 lists these constants for the data in
Figure 6.
It is evident from Figure 6 that the data fit a hyperbolic plot reasonably well, strongly supporting the prediction of the models used (Figure 1) that the relation
between ATPase and [ATP] is hyperbolic. Table 2 also
lists the ADP to ATP ratios that gave 50% inhibition of
the rate observed at 1 mM ATP and 0 ADP ([ADP]/
[ATP]1O%, see "Materials and Methods"). The analysis
confirms our observation from Figures 3 and 6 that
there is little or no ADP inhibition of skeletal S-1
ATPase activity, regardless of actin concentration. An
ADP to ATP ratio of 40-50:1 would be necessary to
obtain 50% inhibition of the skeletal S-1 ATPase activity. Cardiac and smooth muscle S-1 ATPase activities
were inhibited 50% at quite low ADP to ATP ratios,
4-8 :1 for cardiac muscle and 2-4 :1 for smooth muscle.
In addition, for cardiac and smooth muscle S-1, 50%
inhibition occurred at smaller ADP to ATP ratios as the
actin concentration was raised. This suggests that the
inhibition affects mainly the acto-S-1 complex, and this
finding fits well with the model proposed in Figure lb.
We were also able to estimate values for KmIKi from
the data in Figure 6 (see "Materials and Methods").
The value of Km varies with the S-1 type as well as the
actin concentration. To determine the ratio of these
1071
FIGURE 3. Graph showing inhibition of the
actin-activated Mg`+ATPase of subfragment-1
with increasing concentrations of ADP. ATP
hydrolysis was monitored with varying concentrations of ADP and ATP. The total nucleotide
concentration was kept constant at 1 mM (i.e.,
[ADP]+[ATP]=1 mM), and only .10% of the
ATP was hydrolyzed during the time course ofthe
experiments. Lines are hyperbolic least-squares
fits of the data. Data were normalized to the
ATPase rate in the absence of added ADP
([ADP]=0) as calculated from the hyperbolic
fits. Results are shown for each subfragment-1
type at an actin concentration that caused halfmaximal activation in the absence ofadded ADP
(see Table 1). These data are also included in
Figure 6 without normalization. Conditions
were as follows: 10 mM imidazole, 2 mM
MgCI2 (1 =free), 1 mM dithiothreitol, and
[ADP]+[ATP] =1 mM. *---m, Skeletalsubfragment-1 and 2 ,uM actin; ci - cl, cardiac subfragment-1 and 4 ,uM actin; +--..- -+, gizzard
subfragment-1 and 10 ,uM actin.
apparent activation constants, it is necessary to know
Km. We do not have exact values for K., which for this
analysis is the [ATP] that gives one half of the product
release rate at a given actin concentration. However, we
can estimate Km as follows. Acto-S-1 ATPase activity
involves two series of steps: a rapid hydrolysis step
(including ATP binding) and a product release step.
Rapid hydrolysis is the step that is affected by [ATP],
and this rate can be estimated by stopped-flow fluorescence burst measurements. By relating one half of Vma.
(from Table 1) to reported values for the Km for the
burst rate and fluorescence burst rates in the presence
of actin (References 22 and 24 and L.A. Stein and J.S.
Drew, unpublished data), we were able to estimate that
Km is approximately 1 ,uM for the three types of S-1 at
infinite actin (0.5-2 ,uM for skeletal and cardiac S-1 and
0.25-1 ,M for gizzard S-1). Therefore, if we ignore the
small Fl(Km) term from Equation 4 (see "Materials and
Methods"), the estimates for KmIKi will be within
approximately 5%. This leaves us with estimates for
Km/Ki given by F2/(F1+ F2), which is equal to the inverse
of [ADP]/[ATP]50%, and these values are listed in Table
2. Furthermore, by using 1 ,uM as an estimate for Km,
[ADP]/[ATP]5o% (Table 2) approximates Ki for ADP in
the presence of actin.
Discussion
The present study demonstrates significant inhibition
of the actin-activated MgATPase of cardiac and gizzard
S-1 associated with an increase in the ADP to ATP ratio.
Additional experiments demonstrated that neither high
levels of [Pi] (1.5-2.5 mM) nor substrate depletion (low
[ATP]) was associated with the observed ATPase inhibition for any S-1 type. As predicted by analysis of the
kinetic model for ATPase activity (Figure lb), there was
a simple hyperbolic relation between [ATP] and ATPase
activity, given that [ADP] + [ATP] = 1 mM. Hyperbolic
fits of the data indicated that 50% inhibition of the
actin-activated ATPase activity occurred at an ADP-to-
Circulation Research Vol 71, No 5 November 1992
1072
1.20
ai .oo
$4
.34
n4 0.60
4'0
to
a16 0.50
94N
6100.40
'.4i
0.20
m
0.00
Downloaded from http://circres.ahajournals.org/ by guest on June 18, 2017
[ATP] (am)
1.20
a
FIGURE 4. Graphs showing effect of substrate
(ATP) depletion on the actin-activated Mg2`ATPase
of skeletal (top panel), cardiac (middle panel), and
gizzard (bottom panel) subfragment-l. ATPase activity was measured in the presence of decreasing
amounts of A TP with or without increasing amounts
of ADP added. Results are normalized to the initial
rate in the presence of 1 mMATP and 0 mMADP.
Conditions were as follows: 10 mM imidazole, 2 mM
MgCl2, 1 mM dithiothreitol, and 1-0.1 mM ATP;
[A TP] corrected for amount hydrolyzed by 5 minutes;
and ionic strength maintained at 13 mM with added
KCl. o, No ADP added; *, 0-0.9 mMADP added
such that [ADP] + [ATP]=1 mM.
.34
p
4i
Ua
a
p#
94
'-I
[AYP] (am)
1.20
a
1.00
p
4i
.4J
Ph
0.8
6
.6
P4E4
'e6
0.40
"I1
*d 0.20
XlO
0
0.00
-
1.00
0.40
[ATP] (m)
0.co
Drew et al ADP Inhibition of Acto-S-1 ATPase
Downloaded from http://circres.ahajournals.org/ by guest on June 18, 2017
ATP ratio of 3:1 for gizzard S-1 and an ADP-to-ATP
ratio of 6:1 for cardiac S-1. By way of contrast, an
ADP-to-ATP ratio of 40-50:1 would be required for
50% inhibition of skeletal muscle acto-S-1. Inhibition was
greater at higher actin concentrations, suggesting that
ADP is exerting its effect primarily on actin-bound
complexes, but this is also partially due to the asymmetry
in the model for ATP and ADP. ADP only associates and
dissociates, whereas ATP is also hydrolyzed.
From Figure lb (see also "Materials and Methods"),
it is obvious that exact quantitation of KADp/KATp is not
possible until considerably more kinetic experiments are
performed. Both the ATP and ADP "off' rates, k ATP
and k ADP, must be determined, and this is outside the
scope of the present report. Preliminary studies in our
laboratory confirm that the ADP "off' rate for porcine
cardiac acto-S-1 at 15°C is approximately 60/sec, in close
agreement with the value for bovine cardiac S-1 obtained previously.13 In addition, preliminary data with
gizzard acto-S-1 at 25°C reveals an ADP "off' rate that
is about two orders of magnitude faster than Vma, and,
therefore, is not rate limiting for the ATPase cycle. Only
one prior study to measure the ATP "off' rate exists in
the literature,25 and it used quench-flow kinetics. The
technique used in these experiments was a difficult one,
and the interpretation of the data was model dependent
and also depended on several assumptions. However,
this technique can be adapted to our system, making it
possible to develop an interpretation of the data according to the model in Figure lb.
We were able to estimate KmIKi to within approximately 5% for the three muscle types at different actin
concentrations from the present data. This serves as a
good approximation of the relative affinity of ADP
versus ATP for the acto-S-1 active site. Qualitatively,
these data indicate that the relative "affinity"' for ADP
versus ATP is quite strong for cardiac and smooth
muscle acto-S-1 (approximately 0.2 and 0.3, respectively) but very weak (approximately 0.02) for skeletal
acto-S-1. To estimate the value of Ki would require a
determination of Km. A good estimate of this constant
would be difficult to obtain directly because at the very
low [ATP] necessary to estimate Km, the [ADP] resulting from hydrolysis would be high enough to inhibit the
ATPase, especially for cardiac and smooth muscle S-1.
However, we were able to estimate Km indirectly from
other kinetic constants (see "Results") and found that
K. is approximately 1 ,uM for the three types of S-1 in
the presence of actin. This gives us approximations for
Ki of 50, 4-8, and 2-4 ,M for skeletal, cardiac, and
gizzard S-1, respectively. These values compare favorably with Ki values estimated from in situ studies with
skeletal (120 ,uM)6 and smooth muscle tissue (2-5
,uM)"1 as well as an ADP association constant for
cardiac acto-S-1 in vitro (4 1IM).13
Another point that needs to be examined is whether
the new data presented here could have any effect on
the prior interpretations of kinetic data with cardiac
proteins from our laboratory. We do not believe this to
be the case, because we were aware very early that the
ATPase activity of cardiac acto-S-1 was initially linear
and then appeared to slow down late in the reaction.
Skeletal acto-S-1 also slowed down but not until 90+%
of the ATP was split, whereas for cardiac S-1 it was clear
that slowing occurred earlier. We also found very early
1073
that, unlike results obtained with skeletal proteins, the
turbidity of cardiac acto-S-1 in the presence of ATP
began to rise very slowly after approximately half of the
ATP was split. Hence, all of our reported studies were
carried out under conditions in which reasonable linearity was ensured, i.e., conditions in which less than
50% of the ATP was split and [ADP] was insufficient to
lead to significant inhibition.
Physiological Significance of ADP Inhibition
Within the structural environment of the skeletal
muscle fiber, high concentrations of ADP decrease
ATPase activity but increase isometric tension.26 High
concentrations of Pi, however, decrease tension and
slightly decrease the isometric ATPase but do not
inhibit Vma. in solution.27 These observations are consistent with a model suggesting that crossbridge attachment is weak before the release of phosphate and then
strong while ADP remains bound.28 In this model, high
concentrations of Pi would favor the weakly bound
A-M-ADP-Pi state (low tension), whereas high [ADP]
would favor the strongly bound A-M-ADP state (high
tension, slow ATPase).27 In the solution system of
biochemical kinetics, [Pi] had little or no effect on
ATPase activity, and this is probably due to the fact that
the A-M-ADP state is always at its minimal free-energy
conformation in solution as soon as Pi is released. In the
fiber, however, after Pi is released, the A-M-ADP state
is in a strong binding conformation far away from its
minimal free-energy conformation, and work must be
performed to bring the A-M-ADP state to its minimal
free-energy state. For this reason, Pi binding to A-MADP in solution can be very different from Pi binding to
the A-M-ADP state in vivo after Pi is released.
Based on the existence of copious enzymes to convert
ADP to ATP in skeletal muscle, it is generally accepted
that [ADP] cannot increase sufficiently to cause ATPase
inhibition in vivo.27 However, the situation for cardiac
and smooth muscle may be quite different. In the living
rat, hypoxic ventilation (8% or 10% 02) results in severe
depression of cardiac function, decline in high-energy
phosphate (ATP and creatine phosphate) turnover, and
an increase in [ADP] from 58% to 71% of total nucleotide.8 This value for [ADP] is in the range of significant
ATPase inhibition in vitro (see Figure 6, middle panel).
In addition, the measured nucleotide concentrations are
tissue concentrations. Local [ADP] at the myofilament
may be higher because of compartmentalization.8,12
Thus, ADP inhibition may contribute to the decline in
high-energy phosphate turnover with cardiac hypoxia
and could conceivably be involved in the impaired
diastolic relaxation (prolonged strong binding state?)
observed with myocardial ischemia.29
In skinned smooth muscle tissue, elevated ADP
causes contraction in the absence of calcium and myosin
light chain (MLC) phosphorylation9 and is associated
with increased Ca2` sensitivity10 and slow relaxation."
In intact smooth muscle tissue, ATP and phosphocreatine tissue contents do not change significantly during
development of isometric force, but there is a small but
significant increase in ADP content.12 This again may
reflect a larger increase at the contractile apparatus that
is due to functional compartmentalization,12 with the
possibility
that ADP
reaches
concentrations
high
enough (.>50% of total nucleotide, according to our in
Circulation Research Vol 71, No 5 November 1992
1074
U
12
U
-
p,, .o0
43
.44
o.7
'4
j
0.5
o.w
W40.251
Downloaded from http://circres.ahajournals.org/ by guest on June 18, 2017
aottal (P-1)
1/
4
*
m
33
43>
.4
FIGURE 5. Graphs showing effect of inorganic
phosphate (Pi) on the actin-activated Mg2`ATPase
of subfragment-1. Steady-state ATPase measurements were obtained in the presence of ADP with
added KPi, KCl, or imidazole. Lines are double
reciprocal plots of hyperbolic fits of the raw data.
Conditions were as follows: 2 mM MgCI2 (1 =free),
.1 mM dithiothreitol, apP5/proximately 15 mMmM,
ionic strength, and KPi, KCl,
and imidazole as indicated. Top panel: Skeletal
subfragment-1 with 8 mM imidazole, 0.2 mMATP,
and 0.8 mMADP, along with 1.5 mMKPi (m) or2.4
mM KCl (x) at 15°C. Middle panel: Cardiac
subfragment-1 with 0.25 mM ATP and 0.75 mM
ADP, along with 10 mM imidazole and 0. 75 mM
KPi (o), 7 mM imidazole and 1. 75 mM KPi (m), or
10 mM imidazole and 1.125 mM KCl (x) at 150C.
Bottom panel: Gizzard subfragment-1 with 0.4 mM
0.50 ATP, 0.6 mMADP, and 10 mM imidazole, along
with 1.5 mM KP, (i) or 2.4 mM KCl (x) at 25°C.
[ADP]+[ATP]1=
/>
2
d
800
0.10
I
I
I
0.20
0.30
0.40
1/ CACtifJ (pMI 1)
4
U~~~~~~~~~~~~~~
@3
4J 2
-
p#
8.00
0.05
0.10
1L/ [ActiXl]
0.15
(PM-,1)
0.20
0. 25
Drew et al ADP Inhibition of Acto-S-1 ATPase
1075
2.50
_ 2.00
43
1.00
4J
LW
'I
O.W
oo
0.50.
Downloaded from http://circres.ahajournals.org/ by guest on June 18, 2017
8.o
0.20
0.40
0.60
0.80
1.00
[ADP] (mu)
1.75
1.75
FIGURE 6. Graphs showing inhibition of the
actin-activated Mg2`ATPase of subfragment-1 with
increasing concentrations of ADP and varied concentrations of actin. ATP hydrolysis was monitored
with varying concentrations of ADP and ATP. The
_U.. ^~ __total nucleotide concentration was kept constant at
1 mM (i.e., [ADP]+[ATP]=1 mM), and only
s10% of the ATP was hydrolyzed during the time
course of the experiments. Lines are hyperbolic
least-squares fits of the data. Conditions were as
follows: 10 mM imidazole, 2 mMMgCl2 (1 =free), 1
mM dithiothreitol, and [ADP]+[ATP]=1 mM.
Top panel: Skeletal subfragment-1 (0.1 ,M) and
12 (x-x ), 5 (m---m), or 2 (+-+) ,uM actin at
*x
15°C. Middle panel: Cardiac subfragment-1 (0.1
,uM) and 24 (m--- m), 8 (x -x), 4 (+---+), or 2
1.50
*. 4~~~r_
1 i.25
-___
43
pa 1.00 x-
x
00.75
*75
9
0
p. 0.50
94
*
0.25
(rn-rm) uM actin at 15°C. Bottom panel: Gizzard
0°08.0°
0.20
0.40
02DP0
0.0
060
100
0.)
__1.00
43 0.75
94
P.*
0.25
0.20
0.40
0.0
[ADP] (mMl)
(0.2
,uM)
and 20
(m--),
10 (x---x), or 5 (+ - +) pM actin at 25°C.
1.25
0.0810
subfragment-1
0.80
1.00
1076
Circulation Research Vol 71, No 5 November 1992
TABLE 2. Values Derived From Hyperbolic Fits of Data From Figure 6
[Actin] (pM)
Muscle type
F2
F,
2
Skeletal
0.875
0.018
5
0.509
0.014
12
0.411
0.009
2
1.616
0.226
Cardiac
4
1.027
0.198
0.774
8
0.146
24
0.544
0.174
5
1.548
0.497
Smooth
10
1.080
0.476
20
0.435
0.473
[ADP]/[ATP]50%*
49:1
38:1
47:1
8:1
6:1
6:1
4:1
4:1
3:1
2:1
Km/Kit
0.020
0.027
0.021
0.123
0.162
0.159
0.242
0.243
0.306
0.521
F1 and F2, defined by 1/VATp=Fl+(F2/[ATPI) where VATP is the observed ATPase rate; [ADP]/[ATPJ50%,
[ADP]/[ATP] ratio that gives 50% inhibition of the ATPase rate observed at 1 mM ATP, given [ADP]+[ATPI=1 mM;
K,, apparent activation constant of the actin-activated ATPase activity for ATP; Ki, apparent inhibition constant of the
actin-activated ATPase activity for ADP. Values were derived given that 1/V=F1 +(F2/[ATP]) where V is the observed
Downloaded from http://circres.ahajournals.org/ by guest on June 18, 2017
ATPase rate.
*Value estimated by (F1+F2)/F2. See text for derivation. Assuming K,, is approximately 1 ,M, this value also
provides an estimate (in ,uM) for Ki.
tValue estimated by F2/(F1+F2). See text for derivation.
vitro data) to modulate force, ATPase activity, and
shortening velocity in vivo.
Some smooth muscles, including gizzard, have a biphasic pattern of contraction.30,31. An initial transient
phase characterized by high [Ca2+], high levels of MLC
phosphorylation, and rapid shortening velocity and
stress development is followed by long-term stress maintenance at low (but suprabasal) levels of [Ca2+], MLC
phosphorylation, ATPase activity, and shortening velocity. The second phase, termed "latch," has been adequately described by a four-state physiological model in
which phosphorylated crossbridges maintain force with
rapid crossbridge cycling while dephosphorylated crossbridges maintain force with very slow crossbridge detachment.30 Although the latch state has been described
well physiologically, the biochemical basis for this phenomenon remains obscure. For example, it is not understood how unphosphorylated crossbridges would
maintain tension at low but suprabasal levels of [Ca2+]
but permit fast relaxation at basal [Ca2+]. Proposed
calcium-regulated thin-filament proteins such as caldesmon or calponin may be involved.32 Additionally or
alternatively, an increase in the ADP to ATP ratio
occurring with activation may promote or prolong tension maintenance by unphosphorylated crossbridges.
The latch model predicts a linear relation between
shortening velocity and MLC phosphorylation. However, when comparing the relation between shortening
velocity and MLC phosphorylation for various tissue
types, there are variations that cannot be explained by
the small observed differences in the inherent ATPase
rates of the purified myosins.33. In addition, there are
differences in the relation between shortening velocity
and MLC phosphorylation in a given tissue when stimulated by different agonists.33 Again, these differences
may be related to Ca2+-regulated thin-filament proteins,
but ADP may be an additional or alternative modulator
of the crossbridge cycling rate.
Acknowledgment
We wish to thank Marianne P. White for expert technical
assistance.
References
1. Stein LA, White MP, Annis DT. Biochemical kinetics of porcine
cardiac subfragment-1. Circ Res 1989;65:515-525
2. Stein LA, Chalovich JM: Activation of skeletal S-1 ATPase activity
by actin-tropomyosin-troponin. Biophys J 1991;60:300-407
3. Millar NC, Homsher E: The effect of phosphate and calcium on
force generation in glycerinated rabbit skeletal muscle fibers: A
steady-state and transient kinetic study. J Biol Chem 1990;265:
20234-20240
4. Metzger JM, Moss RL: Phosphate and the kinetics of force generation in skinned skeletal muscle fibers. (abstract) Biophys J 1991;
59:418a
5. Walker JW, Lu Z, Swartz DR, Moss DL: Thin filament modulation
of cross-bridge transitions measured by photogeneration of Pi in
skeletal muscle fibers. (abstract) Biophys J 1991;59:418a
6. Sleep J, Glyn H: Inhibition of myofibrillar and actomyosin subfragment 1 adenosinetriphosphatase by adenosine 5'-diphosphate,
pyrophosphate, and adenyl-5'-yl imidodiphosphate. Biochemistry
1986;25:1149-1154
7. Warshaw D, Trybus K: What is the rate limiting step for actin
filament movement by smooth muscle myosin? (abstract) Biophys J
1991;59:186a
8. Bittl JA, Balschi JA, Ingwall JS: Contractile failure and highenergy phosphate turnover during hypoxia: 3"P-NMR surface coil
studies in living rat. Circ Res 1987;60:871-878
9. Kerrick WG, Hoar PE: Non-Ca 2`-activated contraction in smooth
muscle, in Seigman MJ, Somlyo AP, Stephens NL (eds): Regulation
and Contraction of Smooth Muscle. New York, Alan R Liss, Inc,
1987, pp 437-448
10. Nishimura J, van Breeman C: Possible involvement of actomyosin
ADP complex in regulation of Ca 2+ sensitivity in a-toxin permeabilized smooth muscle. Biochem Biophys Res Commun 1989;165:
408-415
11. Nishive E, Somlyo AV, Torok K, Somlyo AP: Cross-bridge kinetics
and the effects of Mg-ADP in alpha-toxin permeabilized guineapig portal vein. (abstract) Biophys J 1992;61:A17
12. Paul RJ: Smooth muscle energetics. Annu Rev Physiol 1989;51:
331-349
13. Siemankowski RF, White HD: Kinetics of the interaction between
actin, ADP, and cardiac myosin-Sl. J Biol Chem 1984;259:
5045-5053
14. Drew JS, Stein LA, White MP, Moos C: Effect of papain digestion
time on kinetics of gizzard myosin subfragment-1. (abstract) Biophys J 1991;59:436a
15. Drew JS, White MP, Stein LA: Is S-1 a good kinetic analogue for
7S HMM in smooth muscle? (abstract) Biophys J 1992;61:A298
16. Stein LA, Schwarz R, Chock PB, Eisenberg E: Mechanism of the
actomyosin adenosine triphosphatase: Evidence that adenosine
5'-triphosphate hydrolysis can occur without dissociation of the
actomyosin complex. Biochem J 1979;18:3895-3909
Drew et al ADP Inhibition of Acto-S-1 ATPase
17. Sobieszek A, Bremel RD: Preparation and properties of vertebrate
smooth-muscle myofibrils and actomyosin. Eur J Biochem 1975;55:
49-60
18. Sellers JR, Pato M, Adelstein RS: Reversible phosphorylation of
smooth muscle myosin, heavy meromyosin, and platelet myosin.
J Biol Chem 1981;256:13137-13142
19. Trybus KM, Henry L: Monoclonal antibodies detect and stabilize
conformational states of smooth muscle myosin. J Cell Biol 1989;
109:2879-2886
20. Ikebe M, Hartshorne DJ: Proteolysis of smooth muscle myosin by
Staphylococcus aureus protease: Preparation of heavy meromyosin
and subfragment 1 with intact 20,000-dalton light chains. Biochemistry 1985;24:2380-2387
21. Nelder JA, Mead R: A simple method for function minimization.
Comput J 1965;308-313
22. Marston SB, Taylor EW: Comparison of the myosin and actomyosin ATPase mechanisms of the four types of vertebrate muscles.
J Mol Biol 1980;139:573-600
23. Chalovich JM, Stein LA, Greene LE, Eisenberg E: Interaction of
isozymes of myosin subfragment-1 with actin: Effect of ionic
strength. Biochemistry 1984;23:4885-4890
24. Marston SB, Taylor EW: Mechanism of myosin and actomyosin
ATPase in chicken gizzard smooth muscle. FEBS Lett 1978;86:
167-170
1077
25. Sleep JA, Hutton RL: Actin mediated release of ATP from a
myosin-ATP complex. Biochemistry 1984;23:4885-4890
26. Hoar PE, Mahoney CW, Kerrick WG: MgADP- increases maximum tension and Ca 2+ sensitivity in skinned rabbit soleus fibers.
Pflugers Arch 1987;410:30-36
27. Pate E, Cooke R: A model of crossbridge action: The effects of
ATP, ADP and Pi. J Muscle Res Cell Motil 1989;10:181-196
28. Eisenberg E, Hill TL: Muscle contraction and free energy transduction in biological systems. Science 1985;227:999-1006
29. Grossman W: Diastolic dysfunction in congestive heart failure.
N Engl J Med 1991;325:1557-1564
30. Hai C-M, Murphy RA: Cross-bridge phosphorylation and regulation of latch state in smooth muscle. Am J Physiol 1988;254:
C99-C106
31. Kenney RE, Hoar PE, Kerrick GL: The relationship between
ATPase activity, isometric force, and myosin light-chain phosphorylation and thiophosphorylation in skinned smooth muscle fiber
bundles from chicken gizzard. J Biol Chem 1990;265:8642-8649
32. Winder SJ, Walsh MP: Smooth muscle calponin: Inhibition of
actomyosin MgATPase and regulation by phosphorylation. J Biol
Chem 1990;265:10148-10155
33. Miller-Hance WC, Kamm KE: Force-velocity relation and myosin
light chain phosphorylation in bovine coronary arterial smooth
muscle. Circ Res 1991;69:1207-1214
Downloaded from http://circres.ahajournals.org/ by guest on June 18, 2017
Product inhibition of the actomyosin subfragment-1 ATPase in skeletal, cardiac, and
smooth muscle.
J S Drew, V A Harwalkar and L A Stein
Downloaded from http://circres.ahajournals.org/ by guest on June 18, 2017
Circ Res. 1992;71:1067-1077
doi: 10.1161/01.RES.71.5.1067
Circulation Research is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 1992 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7330. Online ISSN: 1524-4571
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circres.ahajournals.org/content/71/5/1067
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Circulation Research can be obtained via RightsLink, a service of the Copyright Clearance Center, not the
Editorial Office. Once the online version of the published article for which permission is being requested is
located, click Request Permissions in the middle column of the Web page under Services. Further information
about this process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation Research is online at:
http://circres.ahajournals.org//subscriptions/