Calcium Regulation of Myosin-I Tension Sensing

Abstract

Myo1b is a myosin that is exquisitely sensitive to tension. Its actin-attachment lifetime increases > 50-fold when its working stroke is opposed by 1 pN of force. The long attachment lifetime of myo1b under load raises the question: how are actin attachments that last >50 s in the presence of force regulated? Like most myosins, forces are transmitted to the myo1b motor through a light-chain binding domain that is structurally stabilized by calmodulin, a calcium-binding protein. Thus, we examined the effect of calcium on myo1b motility using ensemble and single-molecule techniques. Calcium accelerates key biochemical transitions on the ATPase pathway, decreases the working-stroke displacement, and greatly reduces the ability of myo1b to sense tension. Thus, calcium provides an effective mechanism for inhibiting motility and terminating long-duration attachments.

Introduction

Myosins are molecular motors that use the energy stored in ATP to perform mechanical work on actin filaments. ATP-dependent conformational changes in the myosin motor domain are amplified into a power stroke by the light-chain binding domain (LCBD), which acts as the motor's lever arm. The LCBD (also known as the regulatory domain) of most myosins is composed of one or more IQ motifs, which are α-helical sequences of ∼23 amino acids, each with a core consensus sequence that binds calmodulin and calmodulin-like proteins (i.e., light chains).

We recently found the actin-attachment lifetime of the widely expressed myosin-I isoform myo1b to be highly force-sensitive (1,2), i.e., actin-attachment lifetimes increase substantially with forces that resist the working stroke. This force sensitivity depends on the length of the LCBD (2). Isoforms with longer LCBDs are more force-sensitive, consistent with the LCBD rotation being coupled to the force-sensitive transition (2). The extraordinarily long attachment lifetime of myo1b under load raises the question: how are attachments that last >50 s in the presence of force regulated?

It has been proposed that calcium binding to LCBD-associated calmodulins is a potential cellular mechanism for regulating motility (3–6), especially since it appears that light-chain binding to the α-helical LCBD may provide the mechanical stiffness needed for force generation (7). Indeed, it has been demonstrated in vitro that calcium binding weakens the affinity of calmodulin for the LCBD of some myosins, resulting in the uncoupling of motile and ATPase activities (5,8–10). Additionally, it has been shown that processive stepping of myosin-V is decreased when the free calcium concentration is increased, presumably as a result of calcium-induced changes in LCBD mechanics (11,12). Thus, it is possible that calcium regulates myo1b tension sensing, because calcium binding to one or more of the calmodulins bound to the LCBD could directly lead to a reduction in the mechanical stiffness of the LCBD and/or effective lever arm length. This change in LCBD mechanics could reduce the ability of the lever arm to bear load, and potentially release the myo1b from prolonged actin-binding events in the presence of force. However, although calcium regulation of myosin tension sensing has been proposed as a model, it has not been shown experimentally.

In this study, we examined the effect of calcium on the biochemical and mechanical properties of myo1b through a series of ensemble-level and single-molecule experiments. Experiments were performed with the “a” splice isoform of myo1b truncated after the sixth IQ motif (myo1b^a) and with a construct truncated after the IQ motif closest to the motor domain (myo1b^IQ) (Fig. 1 A) (2). The results of these experiments establish that calcium modulates myo1b kinetics at physiological temperatures (37°C). Furthermore, these results directly demonstrate that calcium decreases the unitary displacement and force sensitivity of myo1b, confirming the hypothesis that calcium can regulate the extent to which load affects myosin kinetics.

Figure 1.

Myo1b steady-state ATPase and motility rates are calcium-sensitive. (A) Expressed myo1b protein constructs. Diagrams show the relationship of the myo1b motor domain (large rectangles) to the calmodulin-binding IQ motifs (smaller numbered rectangles). (B) Calcium dependence of the actin-activated steady-state ATPase rate for myo1b^IQ in the presence of 2 mM MgATP and 20 μM actin at 37°C. The Ca²⁺_F concentration at half-maximal ATPase activation is K_0.5 = 0.42 ± 0.019 μM. Each point represents the average of three experiments. (C) Ca²⁺_F dependence of myo1b^IQ and myo1b^a powered gliding of actin filaments in the in vitro motility assay in the presence of 2 mM MgATP at 37°C. Error bars are the standard deviation of the velocity.

Materials and Methods

Reagents, proteins, and buffers

Rabbit skeletal muscle actin was prepared and gel filtered (13). Actin for transient kinetics experiments was labeled with N-(1-pyrenyl) iodoacetamide (pyrene-actin) and gel filtered (14). All actin was stabilized with a molar equivalent of phalloidin (Sigma, St. Louis, MO). Calmodulin was expressed in bacteria and purified (15). ADP and ATP concentrations were determined spectrophotometrically before each experiment by absorbance at 259 nm (ε₂₅₉ = 15,400 M⁻¹cm⁻¹). Steady-state and transient experiments were performed in KMg25 buffer (10 mM Mops, 25 mM KCl, 1 mM MgCl₂, 1 mM EGTA, 1 mM DTT). The pH of the buffer was adjusted to 7.0 at 25°C, and changed by <0.1 pH units when the temperature was increased to 37°C. CaCl₂ was added to KMg25 to attain free calcium concentrations as calculated using the MaxChelator program (16).

Myosin-I expression and purification

Myo1b^IQ and myo1b^a expression constructs were prepared as previously described (2,9). A 15-amino-acid AviTag sequence for site-specific biotinylation and a FLAG sequence for purification were inserted at the C-termini of the two constructs (9,17). Myo1b constructs were expressed and purified from Sf9 cells that were coinfected with baculovirus containing recombinant myo1b and calmodulin as previously described (18). Concentrated proteins were site-specifically biotinylated with 20 μg/mL biotin ligase in the presence of 10 mM MgATP and 50 μM biotin at 30°C for 1 h. Free biotin and ATP were removed by MonoQ column chromatography and dialysis.

Kinetic measurements

Transient kinetic and steady-state ATPase measurements were made with an Applied Photophysics (Surrey, UK) SX.18MV stopped-flow apparatus as previously described (19). Unless stated otherwise, all concentrations are given as final after mixing. Rigor solutions of actomyo1b contained apyrase (0.01 U/mL) when loaded into the stopped-flow apparatus to ensure that the mixtures were free of contaminating ADP and ATP. Reported errors are the standard errors of the parameter estimates obtained from regression analysis. Details on the kinetic measurements can be found in the Supporting Material.

In vitro motility assays

In vitro motility assays were performed at 37°C as described previously (9,20). Briefly, glass coverslips were coated with nitrocellulose, and flow cells were assembled with the use of double-stick tape. Solutions were added sequentially to the motility chambers as follows: 0.1 mg/mL streptavidin in water (1 min); 1 mg/mL bovine serum albumin in KMg25 (2 × 5 min); 250 nM biotinylated myo1b and 10 μM calmodulin in KMg25 (2 min); KMg25 buffer containing 5 mM MgATP; KMg25 buffer (3 chamber volumes); 40 nM tetramethyl rhodamine isothiocyanate-phalloidin-labeled F-actin in KMg25; and activation buffer (KMg25 with 1 mg/mL glucose, 5 mM MgATP, 10 μM calmodulin, 1 mg/mL bovine serum albumin, 192 U/ml glucose oxidase, 48 μg/mL catalase (Roche), 0.5% methyl cellulose, and the desired concentration of calcium). Note that rhodamine-phalloidin fluorescence was substantially decreased at 37°C compared with experiments performed at 20°C. The rate of actin filament gliding was determined using Metamorph (Molecular Devices, Sunnyvale, CA). A moving filament was defined as one moving continuously in a directional manner. Only filaments moving in a directional manner were used for the analysis of motility, and thus the observed reduction in velocity upon addition of calcium is due to a reduction in the myosin sliding velocity and not to the number of moving filaments.

Single-molecule measurements

Motility chambers and beads for bead-actin-bead dumbbells were prepared as previously described (1,2). Solutions were added sequentially to the motility chambers as follows: 0.1 mg/mL streptavidin in water (3 min); 1 mg/mL bovine serum albumin in KMg25 (2 × 5 min); 1–5 nM biotinylated myo1b and 10 μM calmodulin in KMg25 (5 min); and 1 nM rhodamine-phalloidin-labeled F-actin in KMg25 with 1 mg/mL glucose, 25 μM ATP, 20 μM calmodulin, 1 mg/mL bovine serum albumin, 192 U/ml glucose oxidase, and 48 μg/mL catalase (Roche). Beads coated with N-ethyl maleimide (NEM)-modified myosin-II were added to one side of the chamber to replace ∼1/4 the volume of the chamber. The chamber was sealed with silicon vacuum grease (Dow Corning, Midland, MI).

Single-molecule interactions were recorded at 20 ± 2°C using the three-bead assay geometry in a dual-beam optical trap system as previously described (1,2,21). Trap stiffnesses were ∼0.022 pN/nm. Bead-actin-bead assemblies were first pretensioned to ∼2.5 pN. The actin was then lowered onto the surface of a pedestal via a piezoelectric stage controller, and moved onto the surface of pedestals to scan for actomyo1b interactions. Upon observation of interactions, data were filtered at 1 kHz and digitized with a 2 kHz sampling rate for up to 10 min. The force dependence of actomyo1b attachment lifetimes was measured with the use of a feedback system that applies a dynamic load to the actomyo1b to keep the actin filament near its isometric position during the myosin working stroke as previously described (2,21). Covariance threshold selection was used to detect single-molecule interactions as previously described (1). Ensemble averages of interactions that were synchronized at the times the interactions started or ended were constructed as described elsewhere (22,23). The best values for the parameters described in Eq. 1 were found with the use of a maximum likelihood estimation (MLE) routine with modified exponential decay probability distribution functions as described previously (2). Data sets from each experimental condition were fit independently, rather than globally as in Laakso et al. (2).

Results

Calcium sensitivity of key steps in the ATPase cycle of myo1b

Calcium binding to the calmodulin associated with the IQ motif closest to the motor domain modulates the ATPase kinetics of the myosin (8,9), so we used myo1b^IQ to determine the effect of calcium on the actin-activated ATPase cycle at 37°C (Table 1, Fig. 1, and Fig. S1). In previous studies, investigators measured the effect of calcium on some key steps of the ATPase cycle at 20°C (8,24); however, here we obtained a more complete characterization at physiological temperatures. The rates of key steps in the ATPase pathway were also determined at 20°C to allow for comparison of biochemical kinetics measured in the stopped flow with mechanical kinetics measured in the optical trap (see below; Table 1 and Fig. S2).

Table 1.

Rate and equilibrium constants for the actomyo1b^IQ ATPase cycle in the absence and presence of 100 μM free calcium at 20°C and 37°C

Calcium dependence of ATPase rate∗			Phosphate release†
				− Calcium	+ Calcium
K0.5Ca (μM)	0.42 ± 0.019 (37°C)		k+4′ (s−1)	0.59 ± 0.056 (37°C)	1.4 ± 0.032 (37°C)
nHill	2.6 ± 0.24 (37°C)		K9 (μM)	55 ± 9.8 (37°C)	5.0 ± 0.44 (37°C)
ATP binding			ADP release
	− Calcium	+ Calcium		− Calcium	+ Calcium
1/K1′ (μM)‡	330 ± 25 (37°C)	236 ± 52 (37°C)	K5′ (μM)‡	0.84 ± 0.069 (37°C)	0.85 ± 0.11 (37°C)
	290 ± 35 (20°C)	340 ± 65 (20°C)		0.56 ± 0.0090 (20°C)	1.3 ± 0.015 (20°C)
k+2′ (s−1)‡	500 ± 15 (37°C)	770 ± 60 (37°C)	k+5′ (s−1)‡	6.7 ± 0.064 (37°C)	11 ± 0.11 (37°C)
	33 ± 1.2 (20°C)	79 ± 4.6 (20°C)		2.1± 0.13 (20°C)	6.4± 0.29 (20°C)
K1′k+2′ (μM−1s−1)§	1.5 ± 0.12 (37°C)	3.3 ± 0.77 (37°C)	k-5′ (μM−1s−1)§	8.0 ± 0.65 (37°C)	13 ± 1.7 (37°C)
	0.12 ± 0.015 (20°C)	0.23 ± 0.046 (20°C)		3.8 ± 0.24 (20°C)	4.7 ± 0.22 (20°C)
Nucleotide-free isomerization step			ATP hydrolysis
	− Calcium	+ Calcium		− Calcium	+ Calcium
Kα‡	3.7 ± 0.65 (37°C)	2.1 ± 0.22 (37°C)	k3app (s−1)¶	41 ± 4.9 (37°C)	39 ± 3.2 (37°C)
	2.6 ± 0.48 (20°C)	5.4 ± 1.2 (20°C)
k+α (s−1)‡	13 ± 1.0 (37°C)	14 ± 1.5 (37°C)
	3.1 ± 0.26 (20°C)	11 ± 2.2 (20°C)
k-α(s−1)§	3.5 ± 0.67 (37°C)	6.7 ± 1.0 (37°C)
	1.2 ± 0.24 (20°C)	2.1 ± 0.62 (20°C)

KMg25 (10 mM MOPS (pH 7.0), 25 mM KCl, 1 mM EGTA, 1 mM DTT, 1 mM MgCl₂, ±100 μM Ca²⁺_F). Transient kinetic data for myo1b^IQ in the absence of calcium at 37°C were published previously (19).

^∗Determined from the calcium-dependent change in the steady-state ATPase rate and fitting of the Hill equation to the data.

^†Phosphate-binding protein.

^‡Pyrene-actin fluorescence.

^§Calculated.

^¶MantATP.

The myo1b^IQ steady-state ATPase rate in the presence of 20 μM actin increased ∼10-fold from 0.16 ± 0.027 s⁻¹ to 1.4 ± 0.016 s⁻¹ at free calcium (Ca²⁺_F) concentrations > 1 μM (Fig. 1 B). This increase is the result of an accelerated rate of phosphate release (k₊₄′) and an increase in the apparent affinity of M.ADP.P_i for actin (K₉; Scheme 1; see below). We determined the half-maximal activation of the ATPase activity by calcium by fitting the Hill equation to the ATPase data, yielding K_0.5 = 0.42 ± 0.019 μM (Table 1):We determined the rate and equilibrium constants for key steps in the actomyosin ATPase pathway (Scheme 1) in the presence and absence of 100 μM Ca²⁺_F using stopped-flow kinetics (Fig. 2 and Table 1; experimental details are provided in the Supporting Material). Calcium increases the rates of ATP binding (K₁′k₊₂′), phosphate release (k₊₄′), and ADP release (k₊₅′) two- to threefold (Fig. 2 and Table 1) but does not affect the rate of the ATP hydrolysis step (k₃^app; Fig. S1 and Table 1). Although the rate constant for phosphate release (k₊₄′) increases (Fig. 2 and Table 1), it remains the rate-limiting step of the ATPase cycle.

Scheme 1.

Figure 2.

Kinetic analysis of the (A and B) rates of ATP binding, (C and D) P_i release, and (E and F) ADP release for myo1b^IQ in the presence (red, ■) and absence (orange, •) of 100 μM Ca²⁺_F. (A) The rate of ATP binding was measured by rapidly mixing pyrene-actomyo1b^IQ with ATP and measuring the increase in pyrene fluorescence as the myosin dissociated from actin. The plot shows the fluorescence increase as a function of time obtained after mixing 1.0 μM actomyo1b^IQ with 150 μM ATP. (B) The rates obtained from the fast phase of the fluorescence transient were plotted against ATP concentration and fit to a hyperbolic function, as represented by the continuous lines. (C) The rate of phosphate release was measured by monitoring the fluorescence produced when free phosphate bound to phosphate-binding protein. The plot shows a transient increase in phosphate-binding protein fluorescence after 1 μM ATP was mixed with 3 μM myo1b^IQ that had been premixed with 10 μM actin. (D) The rate of phosphate release measured as a function of actin concentration. Each data point is the average of one to six transients. (E) The rate of ADP release was measured from the increase in fluorescence after 1 mM ATP was mixed with 0.15 μM pyrene-actomyo1b^IQ that was preequilibrated with varying ADP concentrations. The plot shows the fluorescence transient observed in the presence 10 μM ADP. (F) The rate of the slow phase of ATP-induced dissociation as a function of ADP concentration provides a measurement of the rate of ADP release (k_+5′) at ADP concentrations > 10 μM.

Calcium inhibits actin gliding and decreases the myo1b working stroke

We measured the ensemble-level mechanical activities of myo1b^a and myo1b^IQ using an in vitro motility assay (Fig. 1 C). Myo1b proteins were site-specifically attached to streptavidin-coated coverslips via a biotinylation tag positioned directly C-terminal to the IQ motifs (9), and actin gliding rates were measured in the presence of 0 – 10 μM Ca²⁺_F. The motility rates of myo1b^a (120 ± 5.1 nm/s) at low calcium were ∼3-fold faster than those of myo1b^IQ (38 ± 3.1 nm/s) as a result of the proteins having different lever-arm lengths (2,9). Unidirectional actin gliding was eliminated for both constructs at >1 μM Ca²⁺_F, despite the increased ATPase activity (Fig. 1, B and C). At high calcium concentrations, the actin filaments appeared to reptate and diffuse without a directional bias imposed by the myosin (Movie S1).

We determined the size of the working stroke and the kinetics of the interactions of myo1b^IQ and myo1b^a with actin at the single-molecule level using an optical trap with a three-bead configuration (1,21,25) (Fig. 3; see Materials and Methods). It was previously shown that the myo1b working stroke occurs in two substeps (1,26). Therefore, we determined average substep sizes by ensemble averaging the time courses of unitary interactions synchronized at their start times (time-forward averages) to determine the average displacement of the first substep, and at their end times (time-reverse averages) to determine the average displacement of their total step (Fig. 4) (22,23). In the absence of calcium and in the presence of 25 μM ATP, the displacements were similar to the values reported previously (see Table 3) (2), with total steps of 5.2 ± 0.63 nm and 13 ± 0.57 nm for myo1b^IQ and myo1b^a, respectively. The addition of 9 μM Ca²⁺_F to myo1b^IQ and myo1b^a resulted in fivefold (1.1 ± 0.47 nm) and 15-fold (0.89 ± 0.34 nm) reductions in total displacements, respectively (see Fig. 4 and Table 3), suggesting dramatic calcium-induced changes to the lever-arm structure.

Figure 3.

Representative traces showing single-molecule actomyosin interactions in the optical trap for (A) myo1b^IQ and (B) myo1b^a. Data were collected with the use of the three-bead assay (see Materials and Methods). The number of points in the traces was reduced 30-fold for presentation. Each trace shows 50 s of data. Note the reduction in the size of the working stroke and in the attachment durations in the presence of calcium. The expanded region shows 5 s of data (without reducing the number of points) to highlight the presence of binding events in the presence of calcium.

Figure 4.

Calcium dependence of the (A and B) average working stroke substeps and (C and D) attachment durations for (A and C) myo1b^IQ and (B and D) myo1b^a. (A and B) Ensemble averages of single interactions showing the calcium dependence of myo1b displacements. (C and D) Cumulative frequency distributions of actin-attachment durations for (C) myo1b^IQ and (D) myo1b^a in the absence (•) and presence (■) of 9 μM Ca²⁺_F. Attachment durations were measured for each single-molecule binding event determined by covariance threshold analysis (see Materials and Methods). Solid lines are single exponential functions that were fit to the data (Table 1). Experiments were performed in the optical trap (20 ± 2°C) with 25 μM ATP.

Table 3.

Displacements and actin-detachment kinetics for myo1b^IQ and myo1b^a,∗

	− Calcium		+ Calcium
	Myo1bIQ	Myo1ba	Myo1bIQ	Myo1ba
Total (nm)	5.2 ± 0.63	13 ± 0.57	1.1 ± 0.47	0.89 ± 0.34
Step 1 (nm)	4.2 ± 0.53	6.5 ± 0.53	0.72 ± 0.48	0.17 ± 0.39
Step 2 (nm)	1.0 ± 0.82	6.5 ± 0.78	0.38 ± 0.67	0.72 ± 0.52
Actin detachment rates for myo1bIQ and myo1ba,∗
	− Calcium		+ Calcium
	Myo1bIQ	Myo1ba	Myo1bIQ	Myo1ba
kdet (s−1)	1.1 ± 0.023	0.23 ± 0.0040	2.6 ± 0.049	1.7 ± 0.011
Calcium dependence of Myo1b-IQ and Myo1ba force sensitivity∗,†,‡
	− Calcium		+ Calcium
	Myo1bIQ	Myo1ba	Myo1bIQ	Myo1ba
kg0 (s−1)	0.77 +0.12/−0.27	0.62 +0.49/−0.24	1.5 +0.59/−0.34	0.67 +0.13/−0.13
ki (s−1)	NA	0.014 +0.0044/−0.0034	NA	NA
ddet (nm)	4.1 +0.39/−1.1	21 +5.5/−5.1	0.96 +0.75/−0.81	2.9 +0.51/−0.46

^∗Performed in the optical trap in KMg25 the presence of 25 μM ATP ± 9 μM Ca²⁺_F at 20°C.

^†Force was applied with an isometric optical clamp (see Materials and Methods).

^‡Fit to a probability distribution function and optimized with an MLE routine; 90% confidence interval, N = 1000 simulations (see Materials and Methods).

We determined the attachment durations of single actomyosin interactions using the optical trap. The actomyo1b^a detachment rate (k_det) in the presence of 25 μM ATP increased in the presence of calcium from 0.23 ± 0.0040 s⁻¹ to 1.7 ± 0.011 s⁻¹ (Fig. 4), whereas k_det from actomyo1b^IQ increased from 1.1 ± 0.023 s⁻¹ to 2.6 ± 0.049 s⁻¹. The rate of actomyo1b^a dissociation is similar to that of actomyo1b^IQ in the presence of calcium, consistent with a conformational change in the first calmodulin modulating the changes in the actomyosin kinetics.

The k_det-value measured for actomyo1b^a in the absence of calcium is fourfold slower (0.23 ± 0.004 s⁻¹; Fig. 4) than predicted by transient kinetic experiments conducted at 20°C (i.e., taking into account both the second-order ATP binding rate and the ADP release rate, one would expect a detachment rate of 0.98 s⁻¹; Table 2). This discrepancy is due to the finding that k_det for actomyo1b^a in the absence of calcium is extraordinarily sensitive to load (2), and that the small loads encountered in the optical trap are sufficient to slow the actin-detachment kinetics of myo1b^a (see below). The kinetics of myo1b^IQ are substantially less force-dependent (2), and the small load imposed by the optical trap has little effect on the attachment lifetime (Tables 1 and 3). Interestingly, the k_det-values for actomyo1b^IQ and actomyo1b^a in the presence of calcium are consistent with the rates determined in unloaded biochemistry experiments, suggesting that calcium is affecting the force-sensing properties of actin-bound myo1b^a.

Table 2.

Rate and equilibrium constants for key steps of the actomyo1b^a ATPase cycle in the absence and presence of 100 μM calcium at 20°C^∗

ATP binding
	− Calcium	+ Calcium
1/K1′ (μM)†	420 ± 29	480 ± 32
k+2′ (s−1)†	31 ± 0.70	70 ± 1.6
K1′k+2′ (μM−1s−1)‡	0.074 ± 0.0054	0.15 ± 0.011
ADP release
	− Calcium	+ Calcium
K5′ (μM)†	0.57 ± 0.034	1.48 ± 0.027
k+5′ (s−1)†	2.1 ± 0.13	5.0 ± 0.44
k-5′ (μM−1s−1)‡	3.6 ± 0.031	3.4 ± 0.30
Nucleotide-free isomerization step
	− Calcium	+ Calcium
Kα†	1.5± 0.090	2.8 ± 0.26
k+α (s−1)†	2.4 ± 0.39	7.9 ± 0.30
k-α (s−1)‡	1.6 ± 0.28	2.8 ± 0.28

^∗KMg25 (10 mM MOPS (pH 7.0), 25 mM KCl, 1mM EGTA, 1mM DTT, 1mM MgCl₂, ±100μM Ca²⁺_F, 20°C).

^†Pyrene-actin fluorescence.

^‡Calculated.

Calcium decreases actin-attachment lifetimes in the presence of load

We determined the effect of force on the myo1b^IQ and myo1b^a actin-detachment kinetics in the presence and absence of calcium in the optical trap using a feedback system that applies a dynamic load to actomyosin to keep the actin near its isometric position (21). The load applied to myosin depends on the displacement caused by its initial strong binding to actin, and by the size and direction of the working stroke. The actin-attachment durations of myo1b^IQ and myo1b^a increase with increasing forces, with myo1b^a attachment durations increasing more steeply with force than myo1b^IQ, as shown previously (Fig. 5) (2). The attachment durations of myo1b^a increase until a plateau is reached, at which point the attachment durations are force-independent due to detachment of M.ADP from actin (Fig. 5) (1). In the presence of 9 μM Ca²⁺_F, myo1b^IQ and myo1b^a attachment durations are substantially less force-dependent than in the absence of calcium (Fig. 5), and no plateau is apparent for either myo1b^a or myo1b^IQ.

Figure 5.

Force sensitivity of (A and B) myo1b^IQ and (C and D) myo1b^a in the absence (orange) and presence (red) of 9 μM Ca²⁺_F. (A and C) Attachment durations as a function of applied force in the absence (orange) and presence (red) of 9 μM Ca²⁺_F. Force-dependent data sets acquired in the absence of calcium were published previously (2). Equation 1 was fit to the data, yielding values presented in Table 3. (B and D) Attachment durations were ordered by the average interaction force, and the inverse average attachment durations of 10 consecutive points were plotted. The solid lines show plots of Eq. 1 using the best-fit parameters (Table 3). (E) Model showing the working stroke of myo1b^a in the absence (left) and presence (right) of calcium. The calmodulin bound to the first IQ motif in the LCBD is shown as a colored oval. The remainder of the LCBD is shown as a dashed line. In the absence of calcium, a conformational change in the converter region of the motor domain rotates the LCBD to generate a 13-nm working stroke. In the presence of calcium, the LCBD is not rigidly coupled to the converter region, resulting in a decreased working stroke and reduced force sensitivity. General disordering of the entire LCBD in the presence of calcium is depicted as a wavy line, but this disordering has not been shown directly.

The detachment of myo1b^a from actin was modeled previously assuming force-dependent (k_g) and force-independent (k_i) detachment pathways:

Scheme 2.

The force dependence of the detachment rate (k_det) is given by

$$k_{\det }\left(F\right)=k_{\text{g}0}{e}^{-\frac{F{d}_{\det }}{kT}}+k_i,$$ (1)

where k_g0 is the rate of k_g in the absence of force (F), k is the Boltzmann constant, T is temperature, d_det is the distance parameter that reports the distance to the transition state of the force-sensitive step, and k_i is a force-independent detachment rate that is due to detachment of M.ADP from actin (1,2). We fit Eq. 1 to the data using a maximum-likelihood routine and determined the confidence limits as previously described (2). Fits of Eq. 1 to the myo1b^a attachment durations in the absence of calcium yielded values of k_g0 = 0.62 (+0.49/−0.24) s⁻¹, d_det = 21 (+5.5/−5.1) nm, and k_i = 0.014 (+0.0044/−0.0034) s⁻¹ (Table 3) (2). Because the myo1b^IQ data did not reach a plateau under the loads tested, the force-independent pathway (k_i) was not included in the fitting, which yielded values of k_g0 = 0.77 (+0.12/−0.27) s⁻¹ and d_det = 4.1 (+0.39/−1.1).

The best-fit d_det-values in the presence of 9 μM Ca²⁺_F were sharply decreased for both myo1b^IQ (d_det = 0.96 (+0.75/−0.81) nm) and myo1b^a (d_det = 2.9 (+0.51/−0.46) nm). The calcium-induced change in force sensitivity is most clearly seen in plots of the detachment rate as a function of force (Fig. 5). For example, 1 pN of force decreases k_det ∼3.5-fold for myo1b^IQ and 56-fold for myo1b^a in the absence of calcium. In the presence of calcium, 1 pN of force decreases k_det only 1.3-fold and 2.0-fold for myo1b^IQ and myo1b^a, respectively.

Discussion

In this study, we reveal the mechanism of myo1b regulation by showing that calcium affects the kinetics, unitary displacement, and force dependence of myo1b. In the presence of >1 μM free calcium, the unitary displacements of both myo1b^a and myo1b^IQ are reduced to ∼1 nm (Fig. 4 and Table 3), actin gliding is completely inhibited (Fig. 1 C), and the force dependence of both constructs is substantially reduced (Fig. 5). These results indicate that calcium is acting on the calmodulin bound closest to the motor domain to uncouple conformational changes in the motor domain from LCBD (lever arm) rotation (Fig. 5 E), in contrast to a previous report (8).

Calcium and biochemical kinetics

Calcium-bound calmodulin remains bound to the first IQ motif under our experimental conditions for myo1b^a (9); therefore, the biochemical effects of calcium are likely due to rearrangements (or loss) of interactions at the motor domain-calmodulin interface. These rearrangements are transmitted to the ATP-binding site and also likely affect the rotation and/or stiffness of the lever arm (see below). Calcium-induced conformational changes at the IQ-calmodulin interface have been observed for other myosins (10), and the modulation of kinetics by light chains was observed in previous studies (27–31).

The overall effect of calcium on the unloaded ATPase kinetics is to decrease the lifetime of the biochemical intermediates by more than threefold, without dramatically changing the duty ratio. Of note, calcium increases the affinity of the M.ADP.P_i state for actin (K₉). It also increases the rate of phosphate release (k₊₄′), which results in an increase in the steady-state ATPase activity because phosphate release is the rate-limiting step for the ATPase cycle in both the absence and presence of calcium. Although calcium increases the flux through the ATPase pathway, resulting in increased actin-detachment rates, the unloaded biochemical changes likely are of secondary importance to the dramatic calcium-induced effects on myo1b mechanics. This point is underscored by the decrease in the size of the working stroke and the inability of myo1b^a and myo1b^IQ to support motility at micromolar calcium concentrations (Fig. 1).

The effect of calcium on myosin mechanics

Increasing calcium concentrations slowed actin gliding in the motility assay until the filaments underwent nondirectional diffusive motion (Movie S1). The filaments remained dynamically bound to the motility surface but did not move directionally at the highest Ca²⁺_F concentrations tested (10 μM). At intermediate calcium concentrations, the gliding velocity decreased because some of the myosins stopped contributing to the power stroke (due to the reduction in the size of the working stroke). Calcium-bound myosins may introduce a drag force on myosins that do not have bound calcium, contributing to the observed reduction in gliding velocity (32). The increased affinity of the M.ADP.P_i state (K₉; Table 1) may also prevent diffusion of myosin away from actin filaments during a calcium transient.

Our previous work suggested that tension sensing occurs because force on actomyo1b prevents the rotation of the lever arm, which inhibits ADP release, subsequent ATP binding, and actin detachment (1,2). Myo1b isoforms with shorter LCBDs have shorter unitary displacements and smaller distance parameters (2). Thus, the calcium-shortened myo1b step size results in a shorter distance to the transition state and decreased force sensitivity. In support of this proposal, we note that the magnitudes of the changes in the total working stroke and d_det of myo1b^IQ are approximately the same (∼4.5-fold) upon calcium binding. However, d_det for myo1b^a decreases sevenfold and the working stroke decreases 15-fold in the presence of calcium (Table 3). This result is consistent with our previous results showing that the size of the working stroke is different from the size of d_det for myo1b^a (2). The different results for the two myo1b isoforms may stem from increased compliance in the myo1b^a LCBD due to weaker calmodulin binding (9), or from calcium substantially changing the stiffness of the lever arm of myo1b^a without making it completely compliant. It is also possible that the calcium-bound calmodulin dissociates from myosin in the presence of force. Regardless of the mechanism, our results confirm the proposal that the first light chain is not merely a passive structural element and instead plays an important role in mechanical transduction, as previously proposed for myosin-II (33).

It is worth noting that in the absence of calcium, there is another mechanical difference between myo1b^a and myo1b^IQ that was previously observed (2). The myo1b^a working stroke is composed of two steps of approximately equal size, whereas the myo1b^IQ construct has a larger first step followed by a small second step. If one were to assume that the lever arm acts as an ideal rigid rod, one would expect the ratio of sizes of the first and second substeps of the working stroke to be equal for both myo1b isoforms. Thus, it is possible that there is a nonlinear elasticity in the lever-arm itself, in the linkage between the myosin and surface, and/or the actin-bead linkage. Future measurements of the force-extension relationship may help resolve this issue.

Relationship of myo1b regulation to myo1c and cell function

The response of myo1b to calcium is different from that observed for the widely expressed short-tailed myosin-I isoform, myo1c. In the presence of calcium, myo1c's rate-limiting step shifts from phosphate release to ATP hydrolysis, which increases the population of the prehydrolysis, weak-binding intermediates (29). Additionally, a previous study showed that calcium increases the size of the myo1c working stroke (34). However, we note that at low free-calmodulin concentrations, such as those found in the cell during a calcium transient (35), calcium inhibits myo1c-driven actin gliding (10). Thus, this mechanical effect may depend strongly on the free calmodulin concentration. Further experiments are required to determine whether calcium affects myo1c mechanics in a manner similar to what we observe for myo1b.

The slow kinetics and force-sensitive properties of myo1b point to tethering, anchoring, and/or force-sensing functions for myo1b, rather than a myosin V-like transport role (1,2). Myo1b is concentrated on intracellular membranes, where it has been proposed to play a role in trafficking of membranes from the Golgi apparatus, possibly controlling membrane morphology and dynamics (36). Resting cytoplasmic calcium concentrations are generally in a range (50–100 nM) below the concentration at which myo1b motility is inhibited (K_0.5 = 420 ± 19 nM; Fig. 1). During excitation, the free calcium concentration in the cell can exceed 1 μM (37), which is above the calcium concentration at which mechanical inhibition occurs (Figs. 1 C and 3). Thus, a signaling event that results in a calcium transient will dramatically decrease the actin-attachment lifetime of myo1b working against a load (e.g., myo1b^a working against a 1 pN load will detach from actin 19-fold faster in the presence of calcium). Additionally, myo1b will not be able to translocate along actin to restore tension during the calcium transient, because motility is effectively inhibited in the presence of calcium. However, during a calcium transient, myo1b will transiently interact with actin, and its increased affinity for actin in the M.ADP.P_i state (K₉; Table 1) may help prevent diffusion of myosin away from actin filaments. Thus, calcium is an effective means of terminating the remarkably long force-induced myo1b actin attachments. Given that the majority of myosins appear to have LCBDs stabilized by calmodulin (38), this mechanism of calcium-calmodulin regulation may be a feature of other members of the myosin superfamily.