 |
INTRODUCTION |
Conventional muscle myosin II polymerizes into filaments and is
designed to interact with actin as part of an ensemble, and the kinetic
properties of myosins isolated from these tissues reflect this role.
Their so-called "duty cycle," i.e. the length of time
the myosin spends in a force- or motion-producing state, is relatively
low compared with the overall cycle time determined by the ATPase
activity. This feature allows for speed of contraction, and much of the
cycle is spent in a state that is dissociated from actin. In contrast,
unconventional myosins are nonfilamentous, and most of these classes of
myosin will probably operate in much smaller groups or potentially even
individually. Because of these different functional roles, the kinetic
properties of these motors are expected to be quite different from the
well characterized myosin IIs.
Murine myosin V is a member of the class of unconventional myosins that
is implicated in organelle movement and membrane trafficking based on a
number of cellular and genetic studies. Mutations in murine myosin V
result in a range of defects, from impaired pigment granule movement,
resulting in a dilute coat color, to a lack of smooth endoplasmic
reticulum in the dendritic spines of Purkinje cells, which may be the
cause of the neurological defect that results in early postnatal death
(reviewed in Ref. 1).
Myosin V is particularly interesting from several points of view. It is
a dimeric molecule that has an unusually long neck, three times that of
myosin II. This region of the molecule has been proposed to act as a
lever arm that ultimately results in relative sliding of actin and
myosin. The neck region contains six IQ motifs that have the consensus
sequence (IQXXIRGXXXR) for binding of calmodulin
or myosin light chains. Thus, the potential for
calcium-dependent regulation of motor activity also exists for this myosin. Because its cellular role may require it to work alone
in the cell, myosin V has also been proposed to be a processive motor
that undergoes multiple ATPase cycles before dissociating from its
actin track. In addition, the globular tail domain beyond the
coiled-coil region is believed to be the cargo-binding domain (reviewed
in Ref. 1).
The full-length dimeric myosin V has been isolated from chick brain
(2), but the quantities that can be obtained are limited, precluding a
detailed characterization of the molecule. The strategy we chose was to
express shorter monomeric constructs of murine myosin V using the
baculovirus/insect cell expression system. We then used these simpler
monomeric molecules to begin our characterization of myosin V, much as
proteolytic fragments of conventional myosin II have been used to
obtain a wealth of information about these myosins.
Here we show that the
MD(2IQ)1 construct (motor
domain and two IQ motifs plus two calmodulins) has the potential to be
regulated by calcium but only under conditions where excess calmodulin
is not present. Thus, the mechanism of inhibition appears to be
dissociation of a bound calmodulin. In addition, transient kinetics
were used to measure many of the elementary steps in the acto-MD(2IQ)
ATPase cycle. These measurements suggest that the monomeric MD(2IQ)
construct is not kinetically processive.
| |
MATERIALS AND METHODS |
Design and Purification of Constructs--
Three constructs were
analyzed here. All were derived from the murine myosin V clone (3)
(accession number X57377), which was a generous gift from Nancy
Jenkins. The motor domain (MD) ends at Asp764, MD(1IQ) ends
at Thr795, and MD(2IQ) ends at Thr820. A
dimeric MD(2IQ) construct was made by adding the following sequences
after Thr820: a 32-amino acid leucine zipper, followed by a
portion of the smooth muscle myosin rod containing the S2.2 epitope.
The latter piece of DNA was derived from the "0-heptad zipper"
construct described by Trybus et al. (4). Flag epitope
(DYKDDDDK) was cloned at the C terminus of all heavy chain constructs
to facilitate purification.
Recombinant baculovirus encoding for the different heavy chain
constructs was isolated by conventional protocols (5). Sf9 cells
in suspension culture were infected with recombinant virus coding for
the heavy chain and harvested 65-75 h later. For MD(1IQ) and MD(2IQ),
there was sufficient endogenous cellular calmodulin to saturate the IQ
motifs. Because calmodulin from diverse sources (chicken, rat, and
human) is completely conserved at the amino acid level, we were able to
infect with only one virus. Recombinant proteins were isolated on an
anti-FLAG affinity column (Sigma). Typical yields were 1-1.5 mg of
protein/109 Sf9 cells (300-ml suspension culture).
Densitometry--
Gel images were captured in digital format
using a Kodak Digital Science DC120 zoom digital camera, and the band
intensity was quantified using the Kodak Digital Science 1D image
analysis software package.
Labeling of Actin with Pyrene-Iodoacetamide--
Actin (1 mg/ml
in 10 mM imidazole, pH 7, 10 mM NaCl, 1 mM MgCl2, 1 mM MgATP) was fully
reduced by the addition of 2 mM DTT at pH 8 and then
dialyzed overnight into 10 mM imidazole, pH 7.8, 0.1 M KCl, 1 mM MgCl2, 1 mM
NaN3. A 2-fold molar excess of pyrene-iodoacetamide (48 µM) (Molecular Probes, Inc., Eugene, OR) was added from a 4 mM stock in dimethyl formamide. After 20 h of
labeling in the dark on ice, the actin was pelleted in the Beckman
TL-100 at 350,000 × g for 1 h. The pellet was
resuspended in 10 mM HEPES, pH 7, 0.1 M NaCl, 5 mM MgCl2, 1 mM EGTA, 1 mM NaN3, 1 mM DTT. Protein concentration was determined by Bradford reagent with bovine serum albumin as a standard.
Steady-state ATPase Assays--
Actin-activated ATPase assays
were performed at 37 °C in 10 mM imidazole, pH 7, 8 mM KCl, 1 mM MgCl2, 1 mM EGTA, 1 mM DTT, 1 mM
NaN3 with or without 1.5 mM CaCl2.
One data set was collected in 10 mM HEPES, pH 7, 0.1 M NaCl, 5 mM MgCl2, 1 mM EGTA, 1 mM NaN3, 1 mM DTT at 20 °C for comparison with the transient
kinetic measurements. Inorganic phosphate was determined
colorimetrically (6) at six time points per actin concentration, using
SDS to stop the reaction.
Motility Assay--
The motility assay was performed at 30 °C
in 25 mM imidazole, pH 7.5, 25 mM KCl, 4 mM MgCl2, 1 mM EGTA, 0.5%
methylcellulose, 1 mM MgATP, 10 mM DTT, 3 mg/ml
glucose, 0.1 mg/ml glucose oxidase, 0.018 mg/ml catalase, with or
without 1.5 mM CaCl2, essentially as described
in Ref. 7. Monoclonal anti-FLAG antibody, or antibody S2.2 (8) in the
case of the dimeric MD(2IQ), was used for attachment to the
nitocellulose coverslip. For measurements with excess calmodulin, the
motility solutions also contained 0.1 mg/ml calmodulin.
Transient Kinetic Experiments--
All kinetic experiments were
done in 10 mM HEPES, pH 7, 0.1 M NaCl, 5 mM MgCl2, 1 mM EGTA, 1 mM NaN3, 1 mM DTT at 20 °C using a Kin-Tek stopped flow spectophotometer and a 100-watt mercury lamp.
The concentration of MD(2IQ) was 0.5-2 µM, and when
appropriate, actin was added at 1.2-1.3 times the MD2(IQ)
concentration. For tryptophan fluorescence, the exciting beam was
passed through a 294-nm interference filter, and the emission was
detected after passing through a 340-nm interference filter. For 90°
light scattering, the exciting beam was passed through a 294-nm
interference filter, and the emission was detected with a 294-nm
interference filter. Mant nucleotides and pyrene actin were excited
using a 360-nm interference filter, and emission was detected with a
400-nm cut-off filter. Mant nucleotides (2' or 3') were purchased from
Molecular Probes. All nucleotide stocks were prepared with an equimolar amount of magnesium. All of the transients shown are the average of
three or four independent mixings. The signal averaging and fitting was
done using Kin-Tek software. Note that the final concentrations of
protein and nucleotide after mixing equal volumes of the two syringes
are half the initial values.
Interpretation of Kinetic Data--
Single exponential data were
fit to the equation y = c + a1(exp
1t), and
double exponentials were fit to y = c + a1(exp1t) + a2(exp2t), where
c is a constant and a1 and
a2 are the amplitudes of the signal. The data
are interpreted in terms of standard schemes developed from analysis of
conventional myosin II (9, 10) In Scheme
1, Ki values are
equilibrium association constants (k+i/k
i), and
k+i and ki are rate constants.
In Scheme 2, KA, KD, KAD, and
KDA are equilibrium dissociation constants.
KAD = k5A/K4Ak5A and KD = k5/K4k5.
In all three schemes, A represents actin and M
represents the MD(2IQ) construct.
| |
RESULTS |
Description of the Constructs--
Three monomeric constructs
derived from the murine myosin V clone were expressed in Sf9
cells. MD contains only the catalytic domain (heavy chain residues
1-764). MD(1IQ) contains the motor domain plus one IQ motif for
binding of calmodulin (heavy chain residues 1-795), while MD(2IQ)
contains the motor domain plus 2IQ motifs (heavy chain residues 1-820)
(Fig. 1). The purified MD(1IQ) and
MD(2IQ) constructs have bound calmodulin, which shows its
characteristic calcium-dependent shift in mobility in
SDS-polyacrylamide gel electrophoresis (Fig.
2). The purified constructs show a
stoichiometry of calmodulin/heavy chain that is consistent with the
number of calmodulin-binding motifs present in the construct, within
experimental error (1.3 for MD(1IQ) and 2.2 for MD(2IQ), average slope
from five different gel loadings).
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 1.
Schematic diagram of myosin V. Shown is
a schematic diagram of murine myosin V, adapted from Ref. 2. The
molecule is dimeric and nonfilamentous. The three monomeric constructs
expressed here are MD, MD(1IQ), and MD(2IQ), followed by the amino acid
numbers they contain in brackets. Calmodulin molecules are
indicated by the filled black shapes.
The amino acid number at the head/rod junction, at the end of the first
coiled-coil rod region, and at the end of the full-length molecule are
also indicated in brackets.
|
|
View larger version (31K):
[in this window]
[in a new window]
|
Fig. 2.
Expressed murine myosin V motor domain with
two IQ motifs and bound calmodulin. Shown in an SDS-gel are smooth
muscle myosin purified from tissue (lane 1) and
expressed myosin V motor domain with two IQ motifs and bound
calmodulin, MD(2IQ) (lanes 2 and 3).
Calmodulin undergoes its characteristic calcium-dependent
shift in mobility (lane 2, calcium;
lane 3, EGTA). SDS-15% polyacrylamide gel was
used.
|
|
Steady-state Activity--
The steady-state ATPase activity as a
function of actin concentration was assessed for MD, MD(1IQ), and
MD(2IQ) in the presence or absence of calcium (37 °C, 10 mM imidazole, pH 7, 8 mM KCl, 1 mM
MgCl2, 1 mM EGTA, 1 mM DTT, 1 mM NaN3, with or without 1.5 mM
CaCl2) to determine if the activity of any of these
constructs is regulated by calcium. As expected, calcium had no effect
on the activity of MD (Vmax = 6.3 ± 0.3 s1, Kapp = 0.3 ± 0.07 µM) (Fig. 3A).
MD(1IQ) showed lower activity at subsaturating actin concentrations in
the presence of calcium (Vmax = 8.5 ± 1.5 s1; Kapp = 2.7 ± 1.0 µM) compared with that seen in EGTA
(Vmax = 8.7 ± 0.5 s1;
Kapp = 1.0 ± 0.2 µM), but
this was an effect on Kapp and not on
Vmax (Fig. 3B). In contrast, the
activity of MD(2IQ) was high in the presence of EGTA
(Vmax = 7.4 ± 0.6 s1,
Kapp = 1.2 ± 0.3 µM) and
inhibited to a value of approximately 1 s1 in the
presence of calcium (Fig. 3C). In all instances, the motor activity in the absence of actin was <0.1 s1.
View larger version (13K):
[in this window]
[in a new window]
|
Fig. 3.
The actin-activated ATPase activity of
MD(2IQ) is inhibited by calcium. Steady-state ATPase assays as a
function of actin concentration for MD (A), MD(1IQ)
(B), and MD(2IQ) (C) in the absence
(filled circles) or presence (open circles) of calcium are shown. MD had a
Vmax = 6.3 ± 0.3 s1 and a
Kapp = 0.3 ± 0.07 µM.
MD(1IQ) showed the same Vmax but a higher
Kapp in the presence of calcium
(Vmax = 8.5 ± 1.5 s1;
Kapp = 2.7 ± 1.0 µM)
compared with that seen in EGTA (Vmax = 8.7 ± 0.5 s1; Kapp = 1.0 ± 0.2 µM). In contrast, the activity of MD(2IQ) was high in the
presence of EGTA (Vmax = 7.4 ± 0.6 s1; Kapp = 1.2 ± 0.3 µM) and decreased to ~1 s1 in the
presence of calcium. Conditions were as follows: 10 mM
imidazole, pH 7, 8 mM KCl, 1 mM
MgCl2, 1 mM EGTA, 1 mM DTT, 1 mM NaN3, with or without 1.5 mM
CaCl2 at 37 °C.
|
|
The ATPase values reported here for the monomeric constructs are lower
than reported for the native dimeric molecule isolated from tissue (27 s1 in the presence of calcium) (11). Thus, we engineered
a double-headed MD(2IQ) construct that was dimerized by virtue of the
addition of a leucine zipper, a strategy we have previously used (4). Dimerization of our expressed construct did not increase, and in fact
somewhat reduced, the activity in EGTA (Vmax = 3.2 ± 0.32 s1; Kapp = 1.6 ± 0.4 µM). As with the monomeric MD(2IQ)
construct, the addition of calcium decreased the activity (0.4 s1). The addition of calmodulin in the presence of
calcium increased the ATPase activity to ~70% of the value in EGTA.
Therefore, dimerization per se does not cause enhanced
activity in calcium.
Rescue of ATPase Activity by Exogenous Calmodulin--
Calcium
could decrease the enzymatic activity of the MD(2IQ) complex either by
changing the conformation of the bound calmodulin or by causing its
dissociation from the heavy chain. To discriminate between these two
possibilities, 12 µM exogenous calmodulin was added to
the ATPase solution. The addition of calmodulin to MD(2IQ) in the
presence of calcium resulted in a 10-fold activation of actin-activated
ATPase activity (Table I). The simplest
explanation of these data is that calmodulin dissociation from site 2 (with site 1 adjacent to the motor domain) causes the
calcium-dependent decrease in activity.
View this table:
[in this window]
[in a new window]
|
Table I
Restoration of actin-activated ATPase activity by the addition of
calmodulin
Conditions were as follows: 10 mM imidazole, pH 7, 8 mM KCl, 1 mM MgCl2, 1 mM
EGTA, 1 mM DTT, 1 mM NaN3, 4 µM actin, with or without 1.5 mM
CaCl2, with or without 12 µM calmodulin,
37 °C.
|
|
Dissociation of some calmodulin from the MD(2IQ) complex in the
presence of calcium was also observed by an actin pelleting assay.
MD(1IQ) retained all of its bound calmodulin both in the presence and
absence of calcium. In contrast, MD(2IQ) released some of its
calmodulin into the supernatant in the presence of calcium. Note that
the protein concentration used in the pelleting assay is nearly
100-fold greater than that used for the ATPase assays, and thus it is
likely that more calmodulin is dissociated during activity measurements.
Motility of Expressed Constructs--
The motility of the MD,
MD(1IQ), and MD(2IQ) constructs were examined in the presence and
absence of calcium (Table II). All three
constructs moved actin in the absence of calcium, whether or not
exogenous calmodulin was added. There was a trend toward faster
movement as the length of the neck region increased. The rate of
motility of the dimerized MD(2IQ) was the same as that of the monomeric
MD(2IQ).
View this table:
[in this window]
[in a new window]
|
Table II
Motility of expressed constructs
Numbers in parentheses indicate number of filaments. A minimum of two
independent preparations was used for each construct. Conditions were
as follows: 10 mM imidazole, pH 7.5, 25 mM KCl,
4 mM MgCl2, 1 mM EGTA, 1 mM
NaN3, with or without 1.5 mM CaCl2, with or
without 6 µM calmodulin, 30 °C.
|
|
Calcium had no effect on the rates of motility for MD and MD(1IQ),
consistent with the ATPase measurements, which showed no effect of
calcium on Vmax. Paralleling the inhibition of
ATPase activity observed for MD(2IQ) in the presence of calcium, no
actin movement was observed under these conditions. The addition of 6 µM calmodulin to the motility solutions restored movement
to levels comparable with that observed in EGTA, consistent with the
restoration of ATPase activity that occurred upon the addition of
exogenous calmodulin.
Degree of Association in the Presence of MgATP--
An unusual
feature of the myosin V constructs is the high degree of association
with actin in the presence of MgATP, even for these monomeric
constructs. Both pelleting and light scattering experiments (Fig.
4) show that near physiologic ionic
strength, most of the MD(2IQ) is associated with actin in the presence
of MgATP. Under the same conditions, smooth or skeletal muscle myosin II would be >90% dissociated from actin.
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 4.
MD(2IQ) in the presence of MgATP shows a high
degree of association with actin at physiologic ionic strength.
The degree of association of MD(2IQ) with actin was determined by 90°
light scattering at 295 nm. The value in the absence of MgATP
(open circle) was assigned as 100% association
as determined by actin pelleting assays. 1 mM MgATP was
then added (filled circles), followed by further
additions of NaCl. The filled triangle shows that
smooth muscle S1 at 0.1 M NaCl, 1 mM MgATP
would be completely dissociated. Conditions were as follows: 1 µM actin, 1 µM MD(2IQ), 10 mM
HEPES, pH 7, varying NaCl, 5 mM MgCl2, 1 mM EGTA, 1 mM NaN3, 1 mM MgATP at 20 °C.
|
|
Rate of ATP Binding to MD(2IQ)--
The MD(2IQ) construct was
selected for further analysis by transient kinetics. Table
III summarizes the values that were
obtained. All of these experiments were conducted in 10 mM
HEPES, pH 7, 0.1 M NaCl, 5 mM
MgCl2, 1 mM EGTA, 1 mM
NaN3, 1 mM DTT at 20 °C for comparison with
values previously obtained for conventional muscle myosins (reviewed in
Refs. 9 and 10). The maximal steady-state actin-activated ATPase for
MD(2IQ) under the above conditions was 3.3 ± 0.9 s1, with a Kapp ~20
µM (note that the conditions for Fig. 3 are 8 mM KCl and 37 °C).
View this table:
[in this window]
[in a new window]
|
Table III
Summary of rate constants (see Scheme 1)
Conditions are as follows: 10 mM HEPES, pH 7, 0.1 M NaCl, 5 mM MgCl2, 1 mM
EGTA, 1 mM NaN3, 1 mM DTT at 20 °C.
|
|
The rate of nucleotide binding to MD(2IQ) was followed by the increase
in intrinsic tryptophan fluorescence (Fig.
5). At low ATP concentrations, the rate
of the transient increased linearly with ATP, defining
K1k2 (Scheme 1) as 7 × 105 M1 s1.
Fiting the data to a hyperbola gave a a maximum rate of 198 ± 13 s1, which is likely to be a measure of the rate of the
hydrolysis step (k3 + k3), by analogy with studies on other myosins where the maximum rate of the fluorescence change has been equated with
the rate of ATP cleavage determined by quench flow measurements (12).
The amplitude of the fluorescence signal decreased from 5% at low ATP
concentrations to 3% at high ATP concentrations. This pattern is
consistent with a two-step binding scheme, with approximately half of
the fluorescence change attributed to binding and the other half to
hydrolysis. Binding of ADP caused no change in fluorescence.
View larger version (13K):
[in this window]
[in a new window]
|
Fig. 5.
Rate of ATP hydrolysis as determined by
intrinsic tryptophan fluorescence. A, 2 µM MD(IQ) was mixed with MgATP, and the increase in
tryptophan fluorescence as a function of time (jagged trace) was fitted to a single exponential (smooth line). Note that the concentration of ATP following mixing
is half the initial value. B, these rates are plotted as a
function of ATP concentration. The initial slope defines
K1k2 = 7 × 105 M1 s1. Fitting
the data to a hyperbola gave a maximum rate of 198 ± 13 s1. Conditions were as follows: 10 mM HEPES,
pH 7, 0.1 M NaCl, 5 mM MgCl2, 1 mM EGTA, 1 mM NaN3, 1 mM DTT at 20 °C.
|
|
The source of the tryptophan(s) that contribute to the change upon
nucleotide binding would be expected to be conserved among various
myosins. The MD(2IQ) construct has eight tryptophan residues, seven of
which are in the catalytic domain. Two of these Trp residues are
conserved between smooth, skeletal, Dictyostelium, and
myosin V; they are equivalent to Trp441 and
Trp512 in the smooth muscle numbering system. The varying
amplitude of the signal enhancement seen with different myosins may
reflect variable contributions of the nonconserved Trp residues to the base-line signal, since their level of fluorescence is dependent on
their environment.
Mant Nucleotide Binding to MD(2IQ) and
Acto-MD(2IQ)--
Fluorescent nucleotides have proved to be a useful
way of determining the rate of nucleotide binding to myosins,
particularly when the intrinsic tryptophan signal is small or
nonexistent. Similar to other myosins, nucleotide binding to MD(2IQ)
caused an increase in mant fluorescence. The rate of mant-ATP binding to MD(2IQ) was similar to that determined for ATP, with an apparent second order rate constant of 7 × 105
M1 s1 (Fig.
6C, open
circles). The use of the fluorescent nucleotide also allowed
the value for mant-ADP binding to be determined, for which there was no
corresponding intrinsic tryptophan signal. The signal was biphasic,
with the rate as a function of nucleotide concentration defining
apparent second order rate constants of 4 × 106
M1 s1 and 4 × 105 M1 s1 (Fig.
6A, open circles).
View larger version (33K):
[in this window]
[in a new window]
|
Fig. 6.
Rate of mant-ADP and mant-ATP binding to
MD(2IQ) and acto-MD(2IQ). The rate of mant-ADP (A) or
mant-ATP binding (C) to MD(2IQ) (open circles) or acto-MD(2IQ) (filled circles) as a function of nucleotide concentration is shown.
B and D, the observed fluorescence increase when
mant nucleotides bind to acto-MD(2IQ). Fits are shown by the
smooth line through the experimental trace
(jagged line). The rates were biphasic for
mant-ADP binding and single exponentials for mant-ATP binding. The
apparent second order rate constant for binding to MD(2IQ) is 4 × 106 M1 s1 and
4 × 105 M1 s1
for mant-ADP and 7 × 105 M1
s1 for mant-ATP. Actin had little effect on these values.
Conditions were as follows: 10 mM HEPES, pH 7, 0.1 M NaCl, 5 mM MgCl2, 1 mM EGTA, 1 mM NaN3, 1 mM DTT at 20 °C.
|
|
The rate of mant-ADP and mant-ATP binding to acto-MD(2IQ) was also
determined (Fig. 6, A and C, filled
circles). Binding of mant-ATP fit a single exponential
signal (Fig. 6D), while binding of mant-ADP was best fit by
two exponentials (Fig. 6B). The values that were obtained
were very similar to those obtained in the absence of actin, implying
that actin has little effect on the rate at which nucleotide binds.
Rate of ADP Release from MD(2IQ)--
Because ATP but not ADP
binding caused a change in tryptophan fluorescence, this signal could
be used to determine the rate of ADP release from MD(2IQ). 1 µM MD(2IQ) with varying ADP concentration (4-14
µM) was mixed with 400 µM ATP. The single
exponential rise in fluorescence that was observed occurred at a rate
of approximately 8-12 s1, much lower than the ~100
s1 determined at this ATP concentration in the absence of
ADP (Fig. 7A). Thus, ADP
dissociates from MD(2IQ) at approximately 10 s1
(k5).
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 7.
Rate of ADP release from MD(2IQ) and
acto-MD(2IQ). A, 2 µM MD(2IQ), 14 µM ADP was mixed with 400 µM MgATP, and the
increase in intrinsic tryptophan fluorescence as a function of time
(jagged line) was fitted to a single exponential
(smooth line). B, 0.7 µM
acto-MD(2IQ), 60 µM ADP was mixed with 1 mM
MgATP, and the decrease in light scattering as a function of time
(jagged line) was fitted to a single exponential
(smooth line).
|
|
The rate of mant-ADP release from MD(2IQ), determined by mixing MD(2IQ)
and 14 µM mant-ADP with 1 mM ATP, was 0.6 s1, an order of magnitude slower than observed for ADP.
The dissociation constant of mant-ADP for MD(2IQ), estimated from the
ratio of off to on rate constants is ~0.2 µM (0.6 s1/4 × 106 M1
s1). If the rate of binding of ADP and mant-ADP to
MD(2IQ) is similar, mant-ADP would bind tighter to MD(2IQ) than ADP
does because of a reduced rate of nucleotide release. A 10-fold tighter
binding of mant-ADP compared with ADP was observed with both smooth and skeletal S1 (13).
Rate of Dissociation of Acto-MD(2IQ) by ATP--
Light scattering
was used to determine the rate of dissociation of acto-MD(2IQ) by ATP
(Fig. 8). One µM
acto-MD(2IQ) was mixed with increasing concentrations of ATP. The
apparent second order rate constant for binding
(K1Ak2A)
was 7 × 105 M1
s1. There was little deviation from a straight line until
quite high ATP concentrations, but the extrapolated value estimated from a fit to a hyperbola was approximately 850 s1. This
value could be an underestimate of the maximum rate of dissociation
(kd).
View larger version (14K):
[in this window]
[in a new window]
|
Fig. 8.
ATP rapidly dissociates acto-MD(2IQ).
A, 0.7 µM acto-MD(2IQ) was mixed with MgATP,
and the decrease in light scattering as a function of time
(jagged line) was fitted to a single exponential
(smooth line). Note that the concentration of ATP
following mixing is half the initial value. B, the rates are
plotted as a function of ATP concentration. The apparent second order
rate constant for binding
(K1Ak2A)
was 7 × 105 M1
s1, and the extrapolated maximum value estimated from a
fit to a hyperbola was ~850 s1. Conditions were as
follows: 10 mM HEPES, pH 7, 0.1 M NaCl, 5 mM MgCl2, 1 mM EGTA, 1 mM NaN3, 1 mM DTT at
20 °C.
|
|
Rate of ADP Release from Acto-MD(2IQ)--
Three different signals
were used for measuring the rate of ADP release from acto-MD(2IQ). The
first method, which used pyrene-labeled actin, also allowed one to
determine the affinity of ADP for acto-MD(2IQ). The fluorescence of
pyrene-labeled actin is quenched when it is tightly bound to the motor
domain. Binding of ATP causes an increase in fluorescence, while
binding of ADP causes no change in fluorescence (Fig.
9A). When the MD(2IQ)·ADP
complex (0-16 µM ADP) is mixed with 1 mM
MgATP, the increase in fluorescence is biphasic (Fig. 9A),
with the amplitude of the fast phase decreasing and the amplitude of
the slow phase increasing with ADP concentration (Fig. 9B). The fast phase is due to binding of ATP to pyrene-acto-MD(2IQ) that has
no ADP bound, while the slow phase is due to binding of ATP to
pyrene-acto-MD(2IQ) that has ADP bound. The amplitude dependence of the
signal fitted a hyperbola with a midpoint ~2 µM,
reflecting the affinity of acto-MD(2IQ) for ADP
(KAD). The rate of the slow phase, either when the
transient was fitted to two exponentials or at higher ADP when the
transient was fitted to a single exponential, was between 13.5 and 14.5 s1, which is the rate of ADP release from acto-MD(2IQ)
(k5A).
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 9.
Affinity of ADP for acto-MD(2IQ) and its rate
of dissociation determined with pyrene-actin. A, 0.5 µM acto-MD(2IQ) and 0, 4, 8, or 16 µM MgADP
was mixed with 1 mM MgATP, and the increase in pyrene
fluorescence as a function of time (jagged line)
was fitted to two exponentials, which vary in relative amplitude as the
ADP concentration is increased. The flat trace
was produced by mixing pyrene-labeled actin MD(2IQ) with ADP to show
that ADP alone did not dissociate the complex. B, the
amplitudes of the slow and fast phases are plotted as a function of ADP
concentration. The fit to a hyperbola defines an affinity of ADP for
acto-MD(2IQ) as ~2 µM. Conditions were as follows: 10 mM HEPES, pH 7, 0.1 M NaCl, 5 mM
MgCl2, 1 mM EGTA, 1 mM
NaN3, 1 mM DTT at 20 °C.
|
|
Light scattering was the second method used to measure the rate of ADP
release. The acto-MD(2IQ)·ADP complex was mixed with ATP. The ADP
concentration was 5-60 µM, severalfold higher than the
dissociation constant, and the ATP concentration was 1 mM, so that the rate at which it binds to and dissociates the acto-MD(2IQ) complex is fast compared with the rate of ADP release. The light scattering decrease occurred at a rate of approximately 17-22 s1 (k5A) (Fig.
7B), compared with greater than 200 s1 in the
absence of ADP at this ATP concentration. This value also defines the
rate at which ADP dissociates from acto-MD(2IQ)
(k5A).
The above experiment was repeated with mant-ADP bound to acto-MD(2IQ).
The signal was the decrease in mant fluorescence as the fluorescent
nucleotide dissociated from the active site and was replaced with
nonfluorescent ATP. The rate of the fluorescence decrease was in the
range of 17-19 s1, in good agreement with the value
obtained by light scattering. Thus, ADP and mant-ADP have similar rates
of dissociation from acto-MD(2IQ), in contrast to the different rates
of dissociation they have from MD(2IQ). This pattern of behavior was
also observed for smooth and skeletal S1 (13).
An apparent second order rate constant of 6-11 × 106
M1 s1 for ADP binding to
acto-MD(2IQ) is predicted from a dissociation constant of 2 µM and an off rate of 13-22 s1 (13-22
s1/2 µM).
Effect of ADP on the Affinity of Actin for MD(2IQ)--
The
equilibrium dissociation constant of actin for MD(2IQ) and for
MD(2IQ)·ADP was determined in the stopped flow (KA and KDA in Scheme 2). Phalloidin-stabilized
pyrene-actin (30 nM) was incubated with varying
concentrations of MD(2IQ) or MD(2IQ)·50 µM ADP. The
complex was then dissociated by mixing with 10 µM ATP in
the stopped flow. The final voltage of the fluorescence signal is
constant, since it reflects the fluorescence of 30 nM unbound actin. The initial voltage of the signal decreased as the
MD(2IQ) concentration increased, consistent with a higher degree of
quenching as more motor binds to the pyrene-actin (14). The amplitudes
of the signals, plotted as a function of MD(2IQ) concentration, were
fitted to a hyperbola. The dissociation constant was 61 ± 19 nM in the presence of ADP, and 43 ± 12 nM in the absence of ADP (Fig.
10A, Table
IV).
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 10.
Affinity of MD(2IQ) for actin in the
presence or absence of ADP. A, phalloidin-stabilized,
pyrene-labeled actin (30 nM) plus varying concentrations of
MD(2IQ) was mixed with 10 µM MgATP. The amplitude of the
signal increases with increasing MD(2IQ) concentration. Half-maximal
saturation was obtained at 43 nM. The same experiment was
performed, but the acto-MD(2IQ) complex also contained 50 µM ADP. Half-maximal saturation was obtained at 61 nM. These values define the dissociation constant of actin
from acto-MD(2IQ) or from acto-MD(2IQ)·ADP. B, the rate of
association of actin with 0.25 µM MD(2IQ) in the absence
(filled circles) or presence (open circles) of 50 µM ADP was measured by light
scattering. Conditions were as follows: 10 mM HEPES, pH 7, 0.1 M NaCl, 5 mM MgCl2, 1 mM EGTA, 1 mM NaN3, 1 mM DTT at 20 °C.
|
|
The rate at which MD(2IQ) (with or without 50 µM ADP)
binds to actin was determined by light scattering. The rate decreased from 4.8 × 106 M1
s1 to 3.0 × 106
M1 s1 in the presence of ADP
(Fig. 10B), showing that the presence of ADP slowed the
binding to actin less than 2-fold.
| |
DISCUSSION |
Calcium-dependent Regulation of Myosin V's Motor
Activity
Studies on tissue-purified myosin V showed that the motility and
ATPase activity of this molecule is regulated by calcium (2). To gain
insight into the mechanistic basis for this
calcium-dependent regulation, we first established the
minimal motor whose activity is regulated by calcium. Such an approach
with smooth muscle myosin showed that the minimal molecule that is
regulated by phosphorylation has two heads and a 15-nm tail, suggesting
that both head-head and head-rod interactions are necessary to obtain
the "off" state (4). In the case of myosin V, the motility and
Vmax of MD(1IQ) were independent of calcium, but
the activity and motility of the monomeric MD(2IQ) construct were
substantially inhibited by calcium. Calcium could exert its effect on
activity either by altering the conformation of bound calmodulin (like
phosphorylation affects the activity of the smooth muscle regulatory
light chain) or by dissociation of calmodulin from the heavy chain.
Since the addition of exogenous calmodulin rescued the activity of
MD(2IQ) to 70% of the activity obtained in EGTA, dissociation appears to be the mechanism for inhibition of actin-activated ATPase activity. The simplest interpretation of these data is that calmodulin
dissociates from site 2 (the second site away from the motor domain).
This interpretation assumes that the properties of site 1 are the same in MD(1IQ) as in MD(2IQ), but it is possible that interactions between
calmodulins on adjacent sites modulate the properties of the individual sites.
Comparison with Myosin V Isolated from Tissue
Our results differ from those obtained with dimeric chick myosin V
isolated from tissue, which showed an actin-activated ATPase activity
of 2 s1 in EGTA and 27 s1 in calcium with
excess calmodulin (37 °C) (11). Our ATPase values in calcium were
never an order of magnitude higher than the rates obtained in EGTA,
even with excess calmodulin. A dimerized MD(2IQ) construct also did not
show a higher ATPase activity in calcium plus calmodulin than in EGTA.
With regard to motility, calcium stopped movement of both the expressed
MD(2IQ) and the tissue-isolated chick myosin V, provided that excess
calmodulin was not present (2). In both cases, the addition of
calmodulin partially or wholly restored motility to the level seen in
EGTA. Paradoxically, activity and motility are therefore uncoupled in
the case of the tissue-isolated chick myosin V (high activity and low
motility in calcium; low activity but high motility in EGTA), while
they are coupled for MD(2IQ) (low activity and low motility in calcium;
high activity and high motility in EGTA). Brush border myosin I, a
single-headed motor with three IQ motifs, showed a pattern of activity
similar to that observed with the expressed MD(2IQ). In the presence of
calcium, ATPase activity decreased 5-fold and motility ceased, but
exogenous calmodulin restored both of these activities (15).
It is striking that the rate of motility (~0.3 µm/s at 24 °C)
supported by the full-length construct (2) is similar to that of the
expressed MD(2IQ) (~0.3 µm/s at 30 °C). These results raise the
possibility that the entire lever arm may not be functioning as a rigid
lever that would allow faster motility. Alternatively, the longer
necked constructs may have a correspondingly higher dwell time on
actin, in which case a potentially longer unitary step size would be
balanced by a longer time spent attached to actin (velocity = step
size/dwell time).
Is the Activity of the Myosin V Motor Regulated by
Calcium in Vivo?
Mutants of Myo2p, a member of the class V myosins in yeast,
provide some insights into the potential of this motor to be regulated by calcium. Perhaps most strikingly, a yeast strain that contained a
mutant myosin in which the 6IQ motifs were deleted was still able to
support growth at 90% that of the wild-type strain (16). Since Myo2p
is essential for growth, this result suggests that a motor that lacks
the potential to be regulated by calcium via calmodulin binding to the
neck still functions well. Furthermore, inactivation of the
calcium-bind ing sites of
calmodulin has little effect on the yeast strains carrying this
mutation (17). Both of these studies imply that
calcium-dependent regulation of myosin V's activity is not
essential for proper functioning of myosin V in the cell.
An alternative mechanism for regulation of myosin V's activity is that
proteins or lipids on the surface of the cargo vesicle affect myosin
V's activity via the tail domain. The potential for this to occur was
suggested by studies in which myosin V was isolated bound to its
associated vesicles. Only after treatment with a dilute detergent was
motility observed, suggesting that inhibitory factors were present on
the vesicles (18).
Transient Kinetic Studies of MD(2IQ)
Common Features--
The kinetics of several steps in the ATPase
cycle of myosin V do not differ significantly from that seen with other
myosins (Table III). The rate of dissociation of acto-MD(2IQ) by ATP is fast and extrapolates to approximately 850 s1, and the
rate of ATP hydrolysis is approximately 200 s1. Thus,
like other myosins, myosin V will first dissociate from actin and then
hydrolyze ATP. Second order rate constants for nucleotide binding lie
within the range of previously established values. The rate of ADP
release from acto-MD(2IQ) is fairly slow, between 13 and 22 s1, but this value is still faster than the maximal
velocity from steady-state ATPase measurements
(Vmax = 3.3 ± 0.9 s1 at 0.1 M NaCl and 20 °C) and is therefore not the rate-limiting step in the cycle. The affinity of ADP for acto-MD(2IQ) is high but
close to that observed with smooth muscle S1 (13) or brush border
myosin I (19) (Table IV).
Unique Features--
Myosin V MD(2IQ) differs in several
significant ways from smooth and skeletal S1, particularly in the
coupling between actin and nucletide binding (Tables III and IV). Actin
accelerates the rate of ADP release from MD(2IQ) by only 2-fold,
compared with a 10-fold increase with smooth S1 and a 250-fold increase
with skeletal S1 (Table IV). In addition, the presence of ADP does not
appreciably alter the affinity of MD(2IQ) for actin or its rate of
binding to actin (Fig. 10). In contrast, MgADP weakens the affinity of
skeletal S1 for actin to a large extent (~50-fold), and the affinity
of smooth S1 for actin by ~4-fold. By calculation, one can also infer
that the affinity of ADP for MD(2IQ) and acto-MD(2IQ) is similar and
strong (1-2 µM). In the case of smooth and skeletal S1,
the tighter binding of ADP for the actin-free motor drives the
dissociation of acto-S1 by nucleotides.
Additional evidence that the interaction of myosin V with actin differs
from that seen with conventional myosins comes from the use of mant
nucleotides and pyrene actin. The signals from these probes have been
interpreted in terms of a three-state model, in which actin first forms
a collision complex with the motor and then forms an attached state
(A-state; nucleotide tightly bound, actin weakly bound), followed by a
rotated or rigor-like state (R-state; nucleotide weakly bound, actin
tightly bound) (20, 21). The fluorescence of pyrene-labeled actin is
high when free or in the A-state (A*) and quenched in the R-state when the motor is tightly bound. Mant nucleotides also distinguish between
these two attached states, but in this case the fluorescence is high
when the nucleotide is tightly bound to the motor or in the A-state
(mant-ADP*) and low when unbound or in the R-state (Scheme
3)
With myosin V, the fluorescence of mant-ADP increases upon binding
to acto-MD(2IQ), indicating that nucleotide is tightly bound (Fig. 6).
In the ternary complex of pyrene actin-MD(2IQ)·ADP, the pyrene signal
is also quenched, showing that actin is tightly bound (Fig. 9). This is
unusual in that it implies that both actin and nucleotide are tightly
bound to the motor. In contrast, the addition of mant-ADP to smooth
muscle acto-S1 did not cause an increase in mant fluorescence, although
pyrene fluorescence was quenched in the ternary complex with ADP (13),
indicating that the smooth acto-S1·ADP complex was in the
"R-state" (tight actin binding, weak nucleotide binding).
Another unique feature of myosin V is that it shows a high degree of
association with actin in the presence of MgATP at physiological ionic
strength, as determined by light scattering and actin pelleting assays.
This was previously observed with the full-length dimeric molecule
isolated from tissue (11). Here we show that the high degree of
association is a property of the monomer and does not require the
presence of both heads. In contrast, smooth or skeletal S1 would be
predominantly dissociated from actin under these conditions. Although
light scattering and pelleting assays show a high degree of
association, it is interesting to note that the pyrene actin signal is
not quenched under these conditions (Fig. 9). This implies that the
actin is not tightly bound to the motor and is in the so-called weakly
bound "A-state." One region of myosin V that could contribute to a
higher degree of association with actin in the presence of ATP is the
50/20-kDa loop, which is thought to be involved in the initial
electrostatic interaction of myosin and actin. Although the sequence of
this loop varies even among myosin IIs, there are approximately a dozen
more amino acids in this region of yeast, mouse, and chicken myosin V
that are not found in other myosins. It is possible that this
additional sequence could act as a tether to produce the observed high
degree of association in the steady state.
Is Myosin V Processive?
This study also addresses the question of whether a monomeric
construct of myosin V is kinetically processive. Criteria for kinetic
processivity, which is defined as the average number of ATPase cycles
before dissociation of the actin-motor complex, were established from
recent studies with the microtubule-based motors kinesin and non-claret
disjunctional protein (NCD), and these will be applied to MD(2IQ)
(22-24).
The first criterion is the ratio of Vmax to the
rate constant of dissociation of the motor by ATP. A value <1
indicates low processivity. At 20 °C and 0.1 M NaCl, the
steady-state ATPase activity of MD(2IQ) is 3.3 s1, and
the rate of dissociation is approximately 850 s1,
yielding a value of 0.004, which implies that monomeric myosin V is not
processive. In other words, because of the fast rate of dissociation,
MD(2IQ) is more likely to dissociate from actin than to undergo another
ATPase cycle without dissociation. For comparison, the value obtained
with kinesin monomer is 0.75 (60 s1/80 s1),
and the value for kinesin dimer is >2 (26 s1/12
s1) (22, 23, 25, 26). The interpretation of these data
was that the kinesin monomer showed a small degree of processivity, while the dimer was quite processive. The kinesin dimer has also been
shown to be mechanically processive (27, 28).
A second criterion was the second order rate constant derived from
kcat/Kapp. A value of
>107 M1 s1
indicates many ATPase cycles per encounter. The processive kinesin dimer gave a value of 1.5 × 108
M1 s1 (38 s1/0.25
µM) (25). MD(2IQ) was 1.7 × 105
M1 s1 at 20 °C and 0.1 M NaCl (3.3 s1/20 µM) and
7.4 × 106 M1
s1 at 8 mM NaCl and 37 °C (7.4 s1/1 µM), both values indicating a
nonprocessive motor.
The last criteria was the ratio of the rate of
motility/Vmax. The smaller this ratio, the
higher the duty cycle and therefore the larger the degree of
processivity. This is because the rate of motility is inversely
proportional to the time attached to actin. The closer the attached
time comes to the total cycle time, the smaller the ratio of
motility/Vmax becomes. For MD(2IQ), the ratio is
0.04 (0.3 µm/s/7.4 s1), which is lower than smooth or
skeletal S1, but not as low as obtained for the processive kinesin
dimer (0.015).
In summary, the kinetic evidence does not support the idea that the
myosin V monomer is processive. These data do not exclude the
possibility that the two-headed myosin V species is processive, and we
are currently extending our kinetic studies to dimeric myosin V. Changes in several rate constants would enhance the chance that the
dimeric species could be processive. These would include increasing the
duty cycle by decreasing the ADP release rate or decreasing the rate of
dissociation of acto-MD(2IQ) by ATP. A priori, there is no
reason to assume that myosin V must be a processive motor. An equally
plausible alternative mechanism would be that several myosin V
molecules work in concert to move an organelle along an actin track
without it diffusing away from its track. The architecture of myosin V,
with its long neck region, may enhance the chance that the second head
finds a new actin binding site before the first head dissociates.
However, this sort of mechanism would be quite different from the
coordinated head action of the processive motor kinesin, in which the
heads are closely coupled and not interrupted by a light chain-binding neck region.
,
TOP
ABSTRACT
INTRODUCTION
To whom correspondence should be addressed: Dept. of Physiology
and Molecular Biophysics, Given Bldg. E205, University of Vermont,
Burlington, VT 05405. Tel.: 802-656-8750; Fax: 802-656-0747; E-mail:
trybus@salus.med.uvm.edu.
