The ATPase Pathway That Drives the Kinesin-14 Kar3Vik1 Powerstroke

Background: Kar3Vik1 binds side-by-side microtubule protofilaments and utilizes a minus-end-directed powerstroke.

Results: Microtubule collision occurs through Vik1 followed by Kar3 binding and ADP release, which destabilize Vik1 and generate the intermediate poised for ATP binding.

Conclusion: The transient Kar3Vik1 two-head-bound state intermediate was identified.

Significance: This study provides new insights into force generation by kinesin-14 motors.

Abstract

Kar3, a Saccharomyces cerevisiae microtubule minus-end-directed kinesin-14, dimerizes with either Vik1 or Cik1. The C-terminal globular domain of Vik1 exhibits the structure of a kinesin motor domain and binds microtubules independently of Kar3 but lacks a nucleotide binding site. The only known function of Kar3Vik1 is to cross-link parallel microtubules at the spindle poles during mitosis. In contrast, Kar3Cik1 depolymerizes microtubules during mating but cross-links antiparallel microtubules in the spindle overlap zone during mitosis. A recent study showed that Kar3Vik1 binds across adjacent microtubule protofilaments and uses a minus-end-directed powerstroke to drive ATP-dependent motility. The presteady-state experiments presented here extend this study and establish an ATPase model for the powerstroke mechanism. The results incorporated into the model indicate that Kar3Vik1 collides with the microtubule at 2.4 μm⁻¹ s⁻¹ through Vik1, promoting microtubule binding by Kar3 followed by ADP release at 14 s⁻¹. The tight binding of Kar3 to the microtubule destabilizes the Vik1 interaction with the microtubule, positioning Kar3Vik1 for the start of the powerstroke. Rapid ATP binding to Kar3 is associated with rotation of the coiled-coil stalk, and the postpowerstroke ATP hydrolysis at 26 s⁻¹ is independent of Vik1, providing further evidence that Vik1 rotates with the coiled coil during the powerstroke. Detachment of Kar3Vik1 from the microtubule at 6 s⁻¹ completes the cycle and allows the motor to return to its initial conformation. The results also reveal key differences in the ATPase cycles of Kar3Vik1 and Kar3Cik1, supporting the fact that these two motors have distinctive biological functions.

Introduction

Kinesin-14 represents a subfamily of kinesins that are nonprocessive, promote microtubule (MT)² minus-end-directed force generation, and contain C-terminal motor domains that are dimerized through an N-terminal coiled coil (1–7). Unlike the well known N-terminal motor domain kinesins that use an asymmetric hand-over-hand mechanism for MT plus-end-directed processive stepping (8–11), kinesin-14s use an MT minus-end-directed rotation or bending of the coiled-coil stalk to generate force (12–14). This rotation coupled to ATP turnover is designated the powerstroke and is used to slide one MT relative to another (15–19). During mitosis, kinesin-14s are known to cross-link parallel MTs at the spindle poles but cross-link antiparallel interpolar MTs in the spindle overlap zone (20–24). Furthermore, at the spindle poles and the spindle overlap zone, MT plus-end-directed kinesin-5 is present and provides an outward force to counterbalance the inward force of kinesin-14 to achieve spindle integrity yet allow MT dynamics and other motor and protein interactions required for chromosome segregation (25–28).

Saccharomyces cerevisiae kinesin-14 Kar3 is an intriguing model to study kinesin-14 mechanochemistry and head-head communication because Kar3 forms heterodimers with either Cik1 or Vik1, both of which lack a nucleotide binding site, and Kar3Cik1 and Kar3Vik1 have different physiological roles during the yeast life cycle (21, 22, 29–32). Moreover, the C-terminal domain of Vik1 exhibits the structural fold of a kinesin motor domain, and it can bind MTs with high affinity. However, the absence of a nucleotide binding site indicates that nucleotide cannot modulate the interactions of Vik1 with the MT lattice (33).

Recent studies using equilibrium binding, cryo-EM, x-ray crystallography, and site-directed cross-links at the base of the coiled coil with analysis by in vitro motility assays clearly demonstrated that both Kar3 and Vik1 interact with the MT for motility (14). Furthermore, an unprecedented mode of MT binding was revealed for Kar3Vik1 by high resolution unidirectional metal shadowing, which strongly emphasizes the surface features when viewed by EM. The images of the shadowed MT·Kar3Vik1 complexes in the presence of ADP, used to mimic the beginning of the cycle when Kar3Vik1 collides with the MT, revealed that Kar3 and Vik1 bound side-by-side MT protofilaments rather than binding along a single protofilament in a head-to-tail fashion (14). Because the complexes were formed with ADP, the results suggested that the initial event of MT collision occurred through Vik1; however, this collision complex was not observed through cryo-EM. Docking of the Kar3Vik1 x-ray structure into the cryo-EM maps captured a nucleotide-free prepowerstroke state in which one head was bound to the MT, presumably Kar3, and the second head was detached from the MT with the coiled-coil stalk pointing toward the MT plus-end. In contrast, the ATP state, generated by AMPPNP, showed a new position for the coiled-coil stalk, rotated ∼90° and pointed toward the MT minus-end. In the AMPPNP state, the Vik1 head appeared detached from the MT and in close contact to the coiled coil, suggesting that Vik1 rotates with the coiled coil upon uptake of ATP (14, 34).

In the study presented here, we pursued a mechanistic analysis using presteady-state methodologies to define the ATPase cycle and the sequence of events as the heads interact with the MT to generate force. Although our previous structural studies (14, 33) captured the stable intermediates, this analysis identifies the transient intermediates and associates specific chemical steps of ATP turnover with structural transitions required to generate motility. Although the results reveal many similarities between the ATPase pathways for Kar3Cik1, Kar3Vik1, and Drosophila melanogaster Ncd, suggesting a conserved mechanism for kinesin-14s, there are distinct differences between KarVik1 and Kar3Cik1 that may be related to their different biological roles. In addition, the results provide insights to understand intermolecular communication mediated through the neck coiled coil for Ncd, Kar3Vik1, and Kar3Cik1.

EXPERIMENTAL PROCEDURES

Buffer and Experimental Conditions

The experiments were performed at 25 °C in ATPase buffer (20 mm Hepes, pH 7.2 with KOH, 5 mm magnesium acetate, 0.1 mm EDTA, 0.1 mm EGTA, 50 mm potassium acetate, 1 mm dithiothreitol, 5% sucrose). On the morning of each experiment, the tubulin was cold depolymerized, clarified, and polymerized with 1 mm GTP at 34 °C. The MTs were stabilized with 40 μm paclitaxel, and the MT concentrations are reported as paclitaxel-stabilized tubulin polymer. All reported concentrations of nucleotides and nucleotide analogs also include the equivalent concentration of magnesium acetate. Errors are reported as S.E.

Kar3Vik1

The GCN4-KarVik1 heterodimeric kinesin-14 used in this study was characterized previously (14). The truncated KAR3 gene was cloned into pET-24d (Novagen; kanamycin selection). When expressed, this plasmid yields amino acid residues MSVKELEDKVEELLSKNYHLENEVARLKKLV-Lys³⁵³–Lys⁷²⁹ with a predicted M_r of 46,151. The truncated VIK1 gene was cloned into pKLC37, a plasmid modified from pET-16b (Novagen; ampicillin selection). Vik1 is expressed as MSYYHHHHHHDYDIPTSENLYFQGASMSVKELEDKVEELLSKNHLENEVARLKKLV-Ser³⁴¹–Thr⁶⁴⁷ with a predicted M_r of 42,724. The sequence MSVKELEDKVEELLSKNYHLENEVARLKKLV of Kar3 and Vik1 represents the GCN4 leucine zipper domain for dimerization (35). Kar3 and Vik1 were co-expressed in Escherichia coli BL21-CodonPlus(DE)-RIL as described previously (33, 36), yielding a Kar3Vik1 heterodimer that formed through the GCN4 leucine zipper with a predicted M_r of 88,875.

For the experiments, Kar3Vik1 was clarified (Beckman Coulter TLA 100 rotor, 3 min, 90,000 rpm, 4 °C), and the motor concentration was determined by the Bio-Rad protein assay with IgG as the standard. Concentration is reported as the Kar3Vik1 heterodimer with one nucleotide binding site per Kar3Vik1.

Formation of the Nucleotide-free MT·Kar3Vik1 Complex

Nucleotide-free Kar3Vik1 was generated as described previously for Kar3Cik1 (36) and used to measure the kinetics of ATP binding, ATP hydrolysis, and ATP-promoted MT·Kar3Vik1 dissociation. Briefly, the MT·Kar3Vik1 complex was preformed in the presence of 100 μm MgADP followed by an apyrase treatment (0.02 unit/ml; grade VII, Sigma-Aldrich) for 90 min. This approach assumes that Vik1 will bind to the MT in the presence of ADP with the Kar3·ADP head detached from the MT or weakly bound but tethered to the MT through Vik1. The apyrase VII isoform selectively converts ADP to AMP, which has a very weak affinity for the Kar3 active site such that the MT·Kar3Vik1 complex is essentially nucleotide-free. The concentration of apyrase is sufficiently low that apyrase does not compete with Kar3Vik1 for ATP in the experiments.

Kinetics of MantATP Binding

The nucleotide-free MT·Kar3Vik1 complex was rapidly mixed with mantATP in an SF-2003 KinTek Corp. stopped-flow instrument (see Fig. 1), and the fluorescence enhancement was monitored as a function of time with excitation at 360 nm and detection at 450 nm via a Semrock 409-nm long pass filter (36, 37). The exponential increase in fluorescence is correlated with mantATP binding because the fluorescence of mant nucleotides is enhanced by the hydrophobic environment of the active site. Each averaged transient was fit to a single exponential function plus a linear phase. The observed rates of the initial exponential phase of the transients were plotted as a function of increasing mantATP concentration. These data were fit to a linear function,

where k_obs is the observed rate of the initial exponential phase of the fluorescence enhancement, k₊₁ defines the second-order rate constant for mantATP binding, and k₋₁ represents the observed rate constant of mantATP dissociation as obtained from the y intercept. The equilibrium dissociation constant was determined by K_d = k₋₁/k₊₁ (Scheme 1).

FIGURE 1.

MantATP binding kinetics. The nucleotide-free MT·Kar3Vik1 complex was rapidly mixed in the stopped-flow instrument with increasing concentrations of mantATP. Final concentrations were 5 μm Kar3Vik1 and 12.5 μm MTs for 7.5–40 μm mant-ATP and 2.5 μm Kar3Vik1 and 12.5 μm MTs for 2.5–7.5 μm mantATP. A, representative transients at 5, 10, and 15 μm mantATP (bottom to top) show an exponential increase in fluorescence as a function of time. Each transient was fit to a single exponential function plus a linear term. B, the observed rates of the initial exponential phase of the transients were plotted as a function of mantATP concentration, and the fit of the data provided a second-order rate constant of mantATP binding, k₊₁ of 0.8 ± 0.04 μm⁻¹ s⁻¹ and k₋₁ of 7.4 ± 0.9 s⁻¹.

Pulse-Chase Kinetics of ATP Binding

The nucleotide-free MT·Kar3Vik1 complex was rapidly mixed in the KinTek Corp. chemical quench-flow instrument with ATP plus trace [α-³²P]ATP and 200 mm KCl (syringe concentration) for times ranging from 5 to 500 ms (see Fig. 2 and Refs. 36 and 37). The reaction was subsequently chased with 30 mm nonradiolabeled ATP (syringe concentration) and allowed to proceed for 6 s (>10 ATP turnovers). The reaction was expelled from the instrument into a 1.5-ml tube containing 100 μl of 22 n formic acid to terminate the reaction and release nucleotide from the active site. This experimental design quantifies [α-³²P]ATP stably bound at the active site that proceeds through ATP hydrolysis. Thus, the pulse-chase experiment measures the kinetics of ATP binding. The products [α-³²P]ADP + P_i were separated from [α-³²P]ATP by thin-layer chromatography for quantification. The concentration of [α-³²P]ADP product formed was plotted as a function of time. Each transient was fit to the burst equation.

where A₀ is the amplitude of the exponential burst phase representing the concentration of ATP stably bound at the active site that proceeded through ATP hydrolysis, k_b is the observed rate of the exponential burst phase, and t is the time in seconds. The k_slow is the rate of the linear phase (μm·s⁻¹) and when divided by the Kar3Vik1 concentration approximates the steady-state turnover. The observed rates of the exponential burst phase were plotted as a function of ATP concentration, and the data were fit to Equation 3,

where K₁ defines the equilibrium association constant for complex formation (K_d = 1/K₁) and k₊₁′ represents the first-order rate constant for the ATP-promoted isomerization (Scheme 1).

FIGURE 2.

Pulse-chase kinetics of ATP binding. The nucleotide-free MT·Kar3Vik1 complex was rapidly mixed in the chemical quench-flow instrument with increasing concentrations of [α-³²P]ATP plus KCl and chased with unlabeled ATP. Final concentrations were 10 μm Kar3Vik1 and 30 μm MTs for 25–125 μm [α-³²P]ATP, 5 μm Kar3Vik1 and 30 μm MTs for 5–40 μm [α-³²P]ATP, and 100 mm KCl. A, representative transients display an ATP concentration-dependent presteady-state burst of [α-³²P]ADP·P_i product formation followed by a linear phase of product formation. Each transient was fit to Equation 2. B, the observed exponential rates of the presteady-state burst phase were plotted as a function of ATP concentration and fit to Equation 3, providing a K₁k₊₁′ of 18.6 ± 0.03 μm⁻¹ s⁻¹, K₁ of 0.34 ± 0.03 μm⁻¹, k₊₁′ of 54.9 ± 1.0 s⁻¹, and an apparent K_d,ATP of 2.9 μm. C, the amplitude of each burst phase (as the fraction of Kar3Vik1 active sites) was plotted as a function of ATP concentration and fit to a hyperbolic function, providing a maximum burst amplitude of 1.07 ± 0.02 per Kar3Vik1 heterodimer and apparent K_d,ATP of 25.2 ± 1.3 μm. B and C include data from multiple experiments.

Acid Quench Kinetics of ATP Hydrolysis

The nucleotide-free MT·Kar3Vik1 complex was rapidly mixed in the KinTek Corp. chemical quench-flow instrument with ATP, trace [α-³²P]ATP, and 200 mm KCl (syringe concentration) for times ranging from 5 to 500 ms (see Fig. 3 and Refs. 36 and 37). The reaction was subsequently quenched with 10 n formic acid. The products [α-³²P]ADP + P_i were separated from [α-³²P]ATP by thin-layer chromatography for quantification. The concentration of [α-³²P]ADP product formed was plotted as a function of time, and each transient was fit to Equation 2 where A₀ is the amplitude of the exponential burst phase representing the formation of [α-³²P]ADP·P_i at the Kar3 active site during the first ATP turnover, k_b is the observed rate of the initial exponential phase that corresponds to the rate of product formation at the active site during the first ATP turnover, t is the time in seconds, and k_slow is the rate of the linear phase (μm·s⁻¹), which corresponds to steady-state turnover at 100 mm KCl. The observed rates (Fig. 3B) and amplitudes of the exponential burst phase (Fig. 3C) were plotted as a function of MgATP concentration, and each data set was fit to a hyperbolic function. The maximum rate constant obtained from the hyperbolic fit of the data in Fig. 3B provides the rate constant of ATP hydrolysis, k₊₂ (Scheme 1).

FIGURE 3.

Acid quench kinetics of ATP hydrolysis. The nucleotide-free MT·Kar3Vik1 complex was rapidly mixed in the rapid quench instrument with increasing concentrations of [α-³²P]ATP plus KCl. Final concentrations were 5 μm Kar3Vik1 and 30 μm MTs for 5–80 μm [α-³²P]ATP, 10 μm Kar3Vik1 and 30 μm MTs for 10–200 μm [α-³²P]ATP, and 100 mm KCl. A, representative transients display a time-dependent presteady-state burst of [α-³²P]ADP·P_i product formation followed by a linear phase of subsequent ATP turnovers. Each transient was fit to Equation 2. B, the observed exponential rates of the presteady-state burst phase were plotted as a function of ATP concentration, and the hyperbolic fit provided a rate constant of ATP hydrolysis, k₊₂, of 25.9 ± 1.4 s⁻¹ and an apparent K_d,ATP of 34.4 ± 5.1 μm. C, the amplitude of each exponential burst phase (as the fraction of Kar3Vik1 active sites) was plotted as a function of ATP concentration and fit to a hyperbola, providing a maximum burst amplitude of 0.91 ± 0.03 per Kar3Vik1 heterodimer and a K_d,ATP of 16 ± 1.6 μm. B and C include data from multiple experiments.

MT·Kar3Vik1 Dissociation Kinetics

The nucleotide-free MT·Kar3Vik1 complex was rapidly mixed with ATP, AMPPNP, ADP, and buffer in the stopped-flow instrument, and the change in turbidity at 340 nm was monitored as a function of time (see Fig. 4 and Refs. 36 and 37). The decrease in the turbidity signal is associated with dissociation of the complex because of the decrease in mass of the MTs when Kar3Vik1 detaches and because the motors in solution do not contribute significantly to the turbidity signal. The observed rates of the initial exponential phase of the ATP transients were plotted as an increase in ATP concentration, and the hyperbolic fit of the data provided k₊₃, the rate constant of ATP-promoted Kar3Vik1 detachment from the MT (Scheme 1). To test whether ATP hydrolysis was required for MT·Kar3Vik1 dissociation, the experiment was repeated with buffer, ADP, and the nonhydrolyzable ATP analog AMPPNP.

FIGURE 4.

ATP-promoted dissociation kinetics of the MT·Kar3Vik1 complex. A, the nucleotide-free MT·Kar3Vik1 complex was preformed and rapidly mixed in the stopped-flow instrument with either a nonhydrolyzable ATP analog, AMPPNP (orange), buffer (black), ATP (red), or MgADP (blue). Final concentrations were 2.5 μm Kar3Vik1; 5.5 μm MTs; 1000 μm ATP, ADP, or AMPPNP; and 100 mm KCl. B, the dissociation experiment was performed with increasing concentrations of ATP plus KCl. Final concentrations were 3.5 μm Kar3Vik1 and 7.5 μm MTs for 3.5–200 μm ATP, 2.5 μm Kar3Vik1 and 6 μm MTs for 2.5–100 μm ATP, and 100 mm KCl. The observed exponential rates of the turbidity decrease of each transient were plotted as a function of ATP concentration and fit to a hyperbola, providing a k₊₃ of 6.2 ± 0.1 s⁻¹ and a K½_,ATP of 3.7 ± 0.3 μm.

MT Kar3Vik1 Association Kinetics

Kar3Vik1 was rapidly mixed with MTs in the stopped-flow instrument, and the turbidity signal at 340 nm was monitored as a function of time (see Fig. 5, supplemental Fig. S1, and Refs. 36 and 37). The initial exponential rates of the transients were plotted as a function of increasing MT concentration, and the data were fit to a linear function,

where k_obs is the observed rate for the initial exponential phase of the averaged transients, k₊₄ defines the second-order rate constant for MT association, and k₋₄ corresponds to the observed rate constant of motor dissociation as determined by the y intercept (Scheme 1).

FIGURE 5.

MT Kar3Vik1 association kinetics. Kar3Vik1 with ADP tightly bound at the Kar3 active site was rapidly mixed in the stopped-flow instrument with increasing concentrations of MTs, and the increase in turbidity was monitored. A, representative transients are shown at 1.5, 2, and 2.5 μm MTs, and individual transients were fit to a double exponential function. Final concentrations were 2.5 μm Kar3Vik1 for 2.5–12.5 μm MTs and 1.25 μm Kar3Vik1 for 1.25–2.5 MTs. B, the observed rates of the initial exponential phase of each transient were plotted as a function of MT concentration, and the linear fit of the data provided a k₊₄ of 2.4 ± 0.1 μm⁻¹ s⁻¹. The observed rates of the second slow phase (red ▴) were 1–3 s⁻¹ and were not MT concentration-dependent. C, Comparison of the association kinetics in the presence or absence of 1 mm ADP. Final concentrations: 2.5 μm Kar3Vik1, 5 μm MTs, ± 1 mm ADP.

Kinetics of MantADP Release after MT Collision

Kar3Vik1 at 5 μm was incubated with 30 μm mantADP, and this complex was subsequently mixed in the stopped-flow instrument with MTs plus 2 mm ATP (see Fig. 6 and Refs. 36 and 37). The fluorescence change was monitored over time with excitation at 360 nm and detection at 450 nm via a Semrock 409-nm long pass filter (36, 37). The transients were fit to a double exponential function, and the observed rate of the initial exponential phase was plotted as a function of increasing concentrations of MTs. The hyperbolic fit of the data provided the maximum rate constant, k₊₅, of mantADP release (Scheme 1).

FIGURE 6.

MantADP release kinetics. The Kar3Vik1·mantADP complex was rapidly mixed in the stopped-flow instrument with increasing concentrations of MTs plus ATP, and the decrease in fluorescence was monitored. Final concentrations were 2.5 μm Kar3Vik1 for 2.5–50 μm MTs, 1 μm Kar3Vik1 for 1–2.5 MTs, 15 μm mantADP, and 1 mm ATP. A, representative transients at 2.5, 5, and 15 μm MTs are shown (top to bottom), and each was fit to a double exponential function. B, the observed rates of the initial fast exponential phase of the transients were plotted as a function of MT concentration. The fit of the data to a hyperbola provided a k_max of 14.4 ± 0.3 s⁻¹ and a K½_,MTs of 7.7 ± 0.5 μm. The observed rates of the second slow phase (red ■) showed no concentration dependence and varied from 1 to 3 s⁻¹.

RESULTS

We previously proposed a model for Kar3Vik1 MT interactions based on a powerstroke mechanism for nonprocessive movement (14, 33). This model (Scheme 1 and Fig. 7) was used to design the experiments presented here and to test hypotheses that resulted from the earlier studies. Table 1 presents the constants determined experimentally for Kar3Vik1 in comparison with the constants for Kar3Cik1 reported previously (36).

FIGURE 7.

ATPase Model for Kar3Vik1 Force Generation. The cycle begins as Kar3Vik1·ADP collides with the MT and productive MT binding occurs by Vik1 followed by Kar3·ADP association and ADP release (E0–E3). The release of ADP tightens the MT affinity of Kar3 such that this conformational change destabilizes the interaction of Vik1 with the MT (E2–E3) and forms the E3 intermediate poised for ATP binding. ATP binding at Kar3 promotes rotation of the coiled-coil stalk toward the MT minus-end with Vik1 in association with the coiled coil, resulting in the large conformational change for force production (E3–E4). ATP hydrolysis occurs independently of Vik1, and phosphate release coupled to Kar3Vik1 detachment from the MT completes the cycle (E4–E5). Once detached from the MT, Kar3Vik1 returns to the E0 state for another round of the ATPase cycle. This model assumes that Kar3Vik1 is nonprocessive with one powerstroke per ATP turnover, requiring the coordinated interactions of both Vik1 and Kar3 with the MT. Vik1MHD, Vik1 motor homology domain; Kar3MD, Kar3 motor domain.

TABLE 1.

Comparison of ATPase parameters for the MT·Kar3Vik1 and MT·Kar3Cik1 motors

	Kar3Vik1	Kar3Cik1a
MantATP binding	k+1 = 0.8 ± 0.04 μm−1 s−1	k+1 = 2.1 ± 0.06 μm−1 s−1
	k−1 = 7.4 ± 0.9 s−1	k−1 = 16.6 ± 1.0 s−1
ATP binding (pulse-chase)	K1k+1′ = 18.6 ± 0.03 μm−1 s−1	K1k+1′ = 8.4 ± 0.02 μm−1 s−1
	k−1 not observed	k−1 not observed
	K1 = 0.34 ± 0.03 μm−1	K1 = 0.12 ± 0.01 μm−1
	k+1′ = 54.9 ± 1.0 s−1	k+1′ = 69 ± 1.4 s−1
	Kd,ATP = 2.9 μm	Kd,ATP = 8.3 μm
	A0 = 1.07 ± 0.02 per KarVik1	A0 = 0.94 ± 0.04 per KarCik1
	Kd,ATP = 25.2 ± 1.3 μm	Kd,ATP = 28.2 ± 4.6 μm
ATP hydrolysis	k+2 = 25.9 ± 1.4 s−1	k+2 = 26 ± 0.8 s−1
	Kd,ATP = 34.4 ± 5.1 μm	Kd,ATP = 35.9 ± 4.6 μm
	A0 = 0.91 ± 0.03 per Kar3Vik1	A0 = 0.56 ± 0.02 per Kar3Cik1
	Kd,ATP = 16 ± 1.6 μm	Kd,ATP = 9.8 ± 2.5 μm
MT·Kar3Vik1 dissociation	k+3 = 6.2 ± 0.1 s−1	k+3 = 11.5 ± 0.3 s−1
	K½,ATP = 3.7 ± 0.3 μm	K½,ATP = 17.6 ± 2.5 μm
MT Kar3Vik1 association	k+4 = 2.4 ± 0.1 μm−1 s−1	k+4 = 5.1 ± 0.3 μm−1 s−1
	k−4 not observed	k−4 not observed
	k+4 = 4.0 ± 0.32 μm−1 s−1 (ADP)	k+4 = 4.9 ± 0.1 μm−1 s−1 (ADP)
	k−4 = 29.1 ± 1.5 s−1	k−4 not observed
MantADP release
Fast phase	k+5 = 14.4 ± 0.3 s−1	k+5 = 110 ± 7.7 s−1
	K½,MTs = 7.7 ± 0.5 μm	K½,MTs = 37.9 ± 5.3 μm
Slow phase	kobs = 1–3 s−1	kmax = 5.0 ± 0.2 s−1
		K½,MTs = 5.6 ± 1.0 μm
Steady-state parameters	kcat = 6.7 ± 0.1 s−1 b	kcat = 5.0 ± 0.1 s−1
	Km,ATP = 65.2 ± 4.1 μmb	Km,ATP = 27.1 ± 2.9 μm
	K½,MTs = 0.7 ± 0.05 μmb	K½,MTs = 0.4 ± 0.02 μm
Motility rate	5.87 ± 0.05 μm/minb	2.69 ± 0.08 μm/min

^a From Ref. 36.

^b From Ref. 14.

ATP Binding Is Rapid and Promotes the Kar3Vik1 Coiled-coil Stalk Rotation

Substrate binding for Kar3Vik1 was determined by two approaches: fluorescent mantATP binding using the stopped-flow instrument (Fig. 1) and pulse-chase kinetics with [α-³²P]ATP using the rapid quench-flow instrument (Fig. 2). Fig. 1A shows representative transients of mantATP binding to the nucleotide-free MT·Kar3Vik1 complex, and the observed rates of the initial exponential phase were plotted as a function of increasing mantATP concentration (Fig. 1B). The fit of the data provided a second-order rate constant of mantATP binding of 0.8 μm⁻¹ s⁻¹ with the y intercept, indicating a dissociation rate of 7.4 s⁻¹.

In the second approach, the kinetics of ATP binding were measured using [α-³²P]ATP, and representative transients are shown in Fig. 2A. The initial exponential phase corresponds to the first ATP turnover with the following linear phase representing subsequent ATP turnovers. The data show that there was an increase in both the observed rate and amplitude of the initial exponential phase as a function of ATP concentration. These rates for each transient were plotted as a function of ATP concentration, and the observation that the rate becomes saturated rather than remaining linear is indicative of a rate-limiting ATP-promoted isomerization that occurs at the active site to generate the MT·Kar3Vik1*·ATP intermediate that is poised for ATP hydrolysis (Scheme 1 and Table 1). The maximum rate constant (k₊₁′) of this conformational change is 55 s⁻¹. Because the nonhydrolyzable ATP analog AMPPNP promotes rotation of the Kar3Vik1 coiled-coil stalk (14), we propose that the 55 s⁻¹ isomerization measured here represents the rate constant of the ATP-promoted stalk rotation to generate the postpowerstroke intermediate that is poised for ATP hydrolysis.

Fig. 2C shows the amplitude of each transient plotted as a function of increasing ATP concentration, and the fit of the data provides the maximum amplitude of 1 ADP·P_i per Kar3Vik1. The amplitude represents the concentration of ATP bound at the active site that proceeded through ATP hydrolysis during the first ATP turnover. Because Kar3Vik1 only has a single Kar3 ATP binding site, an amplitude of ∼1 ADP·P_i per Kar3Vik1 was expected and indicates that ATP binding is sufficiently tight that every ATP on the active sites proceeds through ATP hydrolysis.

ATP Hydrolysis Occurs after the Powerstroke

The acid quench experimental design was used to measure the time course of ATP hydrolysis. The representative transients shown in Fig. 3A reveal a presteady-state burst of ADP·P_i product formation at the active site during the first ATP turnover, indicating that ATP binding and hydrolysis are fast followed by a subsequent slow step. At high ATP concentrations, ATP binding occurs faster than ATP hydrolysis. Therefore, ATP binding no longer limits ATP hydrolysis, and the observed rate of the exponential phase predicts the rate of ATP hydrolysis. The rates of the initial exponential burst phase were plotted as a function of increasing ATP concentration (Fig. 3B), and the hyperbolic fit of the data provided a rate constant of ATP hydrolysis, k₊₂, of 26 s⁻¹ with an apparent K_d,ATP of 36 μm. These constants are comparable with those measured for Kar3Cik1, 26 s⁻¹ with a K_d,ATP of 34 μm (Table 1 and Ref. 36), suggesting that ATP hydrolysis is intrinsic to Kar3 and independent of Vik1 or Cik1 or alternatively that the presence of Vik1 or Cik1 affects ATP hydrolysis to the same extent. Because the rate constant is 70 s⁻¹ for the Kar3 motor domain (38, 39), we conclude that ATP hydrolysis is independent of Vik1 or Cik1, but heterodimerization through the coiled coil does have an effect.

Fig. 3C presents the burst amplitude data for the presteady-state transients. Note that the fit of the data provided a maximum burst amplitude of 0.91 ADP·P_i formed on the active site during the first turnover for each Kar3Vik1 in the experiment. The approximately full burst amplitude observed for Kar3Vik1 is in sharp contrast to the amplitude observed for Kar3Cik1 of 0.56 ADP·P_i per Kar3Cik1 (Table 1). In the case of Kar3Cik1, the modeling of the kinetic data indicated that the lowered burst amplitude was due to reversals at ATP hydrolysis, suggesting that phosphate remained at the active site for resynthesis of ATP (36). The approximately full burst amplitude result for Kar3Vik1 indicates that phosphate release is sufficiently fast or that there is an isomerization at the active site that is sufficiently fast to prevent resynthesis of ATP during the first turnover. Note that the burst amplitude for dimeric Ncd was also ∼1 ADP·P_i per Ncd active site as determined for Kar3Vik1 (40).

MT·Kar3Vik1 Complex Dissociation Occurs after ATP Hydrolysis

For most kinesins including Ncd and Kar3Cik1, ATP hydrolysis is required for the motor to detach from the MT (36, 41–43). To test the hypothesis that ATP hydrolysis must occur to generate the weak MT binding state, the preformed nucleotide-free MT·Kar3Vik1 complex was rapidly mixed in the stopped-flow instrument with the nonhydrolyzable ATP analog AMPPNP, buffer, ATP, or ADP, and turbidity was monitored. Fig. 4A shows an ATP- and ADP-promoted exponential decrease in the turbidity signal as a function of time that is correlated with Kar3Vik1 motor detachment from the MT. In contrast, the amplitudes associated with the buffer and AMPPNP transients are extremely small, indicating that ATP hydrolysis was required for the Kar3Vik1 motors to detach from the MT. For the ATP transient, there was a short lag at the start of the ATP-promoted turbidity decrease that is consistent with the time required for ATP binding and ATP hydrolysis to occur. In contrast, the ADP-promoted dissociation transient shows an immediate decrease in turbidity without a lag, suggesting that Kar3Vik1·ADP is the species that detaches from the MT.

The dissociation experiments were repeated with increasing concentrations of ATP. The observed rates for the initial exponential phase of each transient were plotted as a function of ATP concentration with the hyperbolic fit of the data providing a rate constant of MT·Kar3Vik1 complex dissociation, k₊₃, of 6.2 s⁻¹ (Table 1). Because ATP binding and ATP hydrolysis occur as fast steps, the results are consistent with the conclusion that Kar3Vik1 detachment from the MT at 6.2 s⁻¹ is rate-limiting for steady-state ATP turnover (k_cat = 6.7 s⁻¹; Table 1) and occurs after the ATP-promoted powerstroke.

MT Kar3Vik1 Collision Occurs through Vik1

Kar3Vik1 was rapidly mixed with MTs in the stopped-flow instrument, and the exponential increase in the turbidity signal was correlated with formation of the MT·Kar3Vik1 complex (Fig. 5A). The observed rate of the initial fast phase increased linearly as a function of increasing MT concentration, and the fit of the data provided an apparent second-order rate constant, k₊₄, of 2.4 μm⁻¹ s⁻¹ (Fig. 5B and Table 1). The observed rates of the second exponential phase varied from 1 to 3 s⁻¹ (Fig. 5B). The rates of the second phase were not dependent upon the MT concentration and were too slow to account for steady-state turnover at 6.7 s⁻¹. Therefore, the second phase was assumed to be an isomerization that led to an increase in turbidity but was not part of the ATPase pathway.

The second experiment used to evaluate the collision of Kar3Vik1 with the MT was to measure the kinetics of mantADP release because MT association dramatically activates ADP release from the active site of kinesins (38, 40, 44–47). A Kar3Vik1·mantADP intermediate was generated by incubating the Kar3Vik1 with mantADP at a 1:6 ratio to exchange ADP at the active site with the fluorescent analog mantADP. This Kar3Vik1·mantADP complex was then rapidly mixed in the stopped-flow instrument with MTs plus 1 mm ATP. The biphasic transients shown in Fig. 6A do exhibit a short lag, and the fluorescence decrease is correlated with mantADP leaving the hydrophobic environment of the Kar3 active site. The observed rates of the initial fast phase, when plotted as a function of MT concentration, provided a maximum rate constant of mantADP release of 14 s⁻¹ and a K½_,MTs of 7.7 μm (Fig. 6B and Table 1). The observed rates of the second slow phase were 1–3 s⁻¹ and were independent of MT concentration (Fig. 6B), suggesting an off-pathway isomerization.

When this experiment was performed for Kar3Cik1, there was no lag at the start of the transients, and the initial fast phase of mantADP release was 110 s⁻¹ (Table 1 and Ref. 36). Furthermore, the second phase for Kar3Cik1 showed MT concentration dependence, and the hyperbolic fit of the data provided a k_max of 5 s⁻¹ with a K½_,MTs of 5.6 μm. Because of the absence of a lag, the interpretation was that the 110 s⁻¹ rate constant represented MT collision by Kar3 for ∼⅓ of the population, and the 5 s⁻¹ rate constant represented ∼⅔ of the population of motors that collided with the MT through Cik1 followed by Kar3 MT association and mantADP release. Because steady-state ATP turnover for Kar3Cik1 was also 5 s⁻¹, the authors proposed that the release of mantADP at 5 s⁻¹ captured the rate of the structural transition for Kar3 binding to the MT after Cik1 MT collision (36).

When the mantADP release kinetics for Kar3Vik1 are interpreted in the context of the kinetics for Kar3Cik1, it is obvious that the Kar3Vik1 14 s⁻¹ rate constant cannot represent the intrinsic rate constant for mantADP release from Kar3 if this head were to collide with the MT first independently of Vik1 because it would be observed at 110 s⁻¹. Rather, the mantADP release kinetics for Kar3Vik1 indicate that the initial collision is through Vik1 followed by Kar3 binding and mantADP release at 14 s⁻¹. The observation that the K½_,MTs for Kar3Vik1 (7.7 μm) is similar to the second phase K½_,MTs for Kar3Cik1 (5.6 μm) supports this interpretation. These results also suggest that there is an asymmetry in the Kar3Vik1·ADP species that promotes MT collision through Vik1.

The Two-head-bound State Is Transient

Rank et al. (14) revealed a novel MT binding pattern in which Kar3 and Vik1 bound to adjacent MT protofilaments. This binding configuration requires that Kar3 and Vik1 be in opposite orientations on the MT lattice because the coiled coil does not unwind significantly and predicts that the E2 state with both heads bound to the MT prior to ADP release is transient (Fig. 7). This binding pattern suggests that the intermediate with both heads bound would not be stable because of the distance between MT binding sites and the structural transition that occurs with tight binding of Kar3 (48). In initial MT association experiments, we assumed that collision would occur through either Kar3 or Vik1, and to bias binding through Vik1, ADP was added to the MT syringe. Fig. 5C shows that the kinetics were biphasic, and the initial exponential phase of the ADP transient was similar to that of the transient in the absence of ADP. However, the amplitude associated with the ADP transient was about 3 times smaller than that of the transient in the absence of ADP. These results suggest that although MT collision occurred through Vik1, binding of Kar3 to the MT destabilized Vik1 such that the high concentration of ADP resulted in complete detachment of a significant population of the Kar3Vik1 motors. This interpretation was reinforced when the association kinetics were measured as a function of MT concentration in 1 mm MgADP (supplemental Fig. S1). The results showed that Kar3Vik1 collides with MTs at 4 μm⁻¹ s⁻¹; however, high ADP promoted a significant off-rate of 30 s⁻¹. This is in sharp contrast to the data in Fig. 3B where there is no evidence of an off-rate in the absence of added ADP.

The MT cosedimentation analysis of Kar3Vik1 demonstrated that Kar3Vik1 partitioned with the MTs in 1 mm MgADP and showed saturated MT binding at 1 Kar3Vik1 per 2 αβ-tubulin subunits. However, at conditions in which the MT complex was formed with nucleotide-free Kar3Vik1, the stoichiometry became ∼1:1 (14). Furthermore, the cryo-EM results showed a stoichiometry at 1:1 for nucleotide-free Kar3Vik1, and the binding was cooperative and dominated by the tight binding of the Kar3 motor domain. The results from each of these experiments are consistent with the interpretation that the E2 species (Fig. 7) is transient, and the tight binding by Kar3 after ADP release (E3) destabilizes the Vik1 MT complex due to the strain between the Kar3 and Vik1 heads and is mediated by the coiled coil.

DISCUSSION

Proposed Model of the Kar3Vik1 Powerstroke Mechanism

The model that has emerged from the structural and mechanistic studies is presented in Fig. 7. In solution, Kar3Vik1 has ADP tightly bound at the active site (E0). Kar3Vik1·ADP collides with the MT, and we propose that the productive binding state is through Vik1 followed by Kar3·ADP association and ADP release (E0–E3). Release of ADP tightens the MT affinity of Kar3 such that this conformational change destabilizes the interaction of Vik1 with the MT (E2–E3) and forms the E3 intermediate poised for ATP binding. ATP binding at Kar3 promotes rotation of the coiled-coil stalk with Vik1 in association with the coiled coil, resulting in the large conformational change for force production (E3–E4). ATP hydrolysis occurs independently of Vik1, and phosphate release coupled to Kar3Vik1 detachment from the MT completes the cycle (E4–E5). This powerstroke mechanism requires both Vik1- and Kar3-coordinated interactions with the MT mediated by head-head communication via the coiled coil. We propose that tight binding of Kar3 to the MT and the interactions of Vik1 with the coiled coil at E4 prevent slippage such that Vik1 cannot rebind to the MT and the coiled-coil stalk cannot revert back to the prepowerstroke state at E3. Furthermore, this mechanistic capability is intrinsic to the MT·Kar3Vik1 complex without requiring an MT-Kar3Vik1-MT cross-linked configuration, suggesting that the Vik1 association with the coiled coil plays an important role in preventing reversals at E4.

The structural studies published previously can only capture the more stable MT·Kar3Vik1 intermediates. Cryoelectron tomography (34) and cryo-EM with three-dimensional helical analysis (14) identified the E3 prepowerstroke intermediate and the postpowerstroke E4 intermediate. Although most structures of kinesin have revealed the solution state, the x-ray crystal structure for Kar3Vik1 (Protein Data Bank code 4ETP) identified the E3 prepowerstroke intermediate (14) but not the solution state. The results presented here provide evidence for the more transient intermediates and the structural transitions that occur. We propose that the Kar3Vik1 collides with the MT through Vik1, which promotes Kar3 binding and ADP release (Figs. 5 and 6). Both E1 and E2 are transient states that were not captured by the structural studies. The kinetics of ATP binding (Fig. 2) identified the transition from E3 to E4 at 55 s⁻¹, and the results from Fig. 3 revealed that ATP hydrolysis was intrinsic to Kar3 and independent of Vik1. The dissociation kinetics (Fig. 4) confirmed that Kar3Vik1 detached from the MT after ATP hydrolysis and therefore from the postpowerstroke state. The results also indicate that it is after detachment from the MT that Kar3Vik1 returns to the E0 state whose conformation biases it for productive MT binding through Vik1 at the start of the next ATP cycle.

Kinesin-14s: Kar3Vik1, Kar3Cik1, and Ncd

Both the structural and mechanistic results indicate that the ATPase mechanisms for Kar3Vik1 and Ncd are very similar even though Ncd is a homodimer and the nucleotide state can modulate the interactions of both heads with the MT (12–14, 33, 40, 42, 49, 50). Both Ncd and Kar3Vik1 show cooperative binding to the MT lattice such that in a population of MTs one MT will be completely saturated by Kar3Vik1 or Ncd, but other nearby MTs show no evidence of MT binding (12, 14, 33). This cooperative binding behavior is inherent to the dimer state because when similar experiments are performed with the Kar3 motor domain the monomers bind stochastically to the MT (33, 51). Asymmetry for Ncd is generated through asymmetric affinity for ADP where one head holds ADP tightly and the other holds ADP weakly when Ncd is in solution and detached from the MT, and Foster et al. (42) proposed that it was this asymmetry that biased the head that bound ADP weakly to collide and form the productive MT association complex. The Ncd prepowerstroke intermediate captured by cryo-EM is very similar to the MT·Kar3Vik1 prepowerstroke intermediate, and ATP binding mimicked by AMPPNP for both Kar3Vik1 and Ncd results in the postpowerstroke intermediate (12–14, 49). Cryo-EM for both Ncd and Kar3Vik1 showed the detached head in association with the coiled coil, providing additional evidence that this head rotates with the coiled coil to generate the postpowerstroke intermediate (12–14). ATP hydrolysis for both Ncd (40) and Kar3Vik1 showed full burst amplitude, indicating that reversals at this step to resynthesize ATP do not occur, and MT detachment for both occurred from the postpowerstroke state. Therefore, based on these results, the evidence supports a conserved mechanism for Kar3Vik1 and Ncd for cross-linking parallel MTs as would occur at the spindle poles.

In contrast, Kar3Cik1 does not exhibit the cooperative binding observed for Ncd and Kar3Vik1 (51). Instead, Cik1 targets Kar3 to the MT plus-end, and Kar3Cik1 promotes MT shortening in vitro at conditions where the MTs are not highly stabilized by paclitaxel. Kar3Cik1-promoted MT shortening does require ATP turnover, and the single exponential kinetics suggest that a single ATP turnover promotes release of a single αβ-tubulin subunit from the MT plus-end. In contrast, because of the cooperative binding, Ncd and Kar3Vik1 did not exhibit the same characteristics of MT plus-end targeting for MT shortening as observed for Kar3Cik1 (33). For Kar3Cik1, the burst amplitude for the ATP hydrolysis kinetics was reduced, and computational modeling suggested that the reduced burst amplitude was due to reversals of ATP hydrolysis to resynthesize ATP (Table 1 and Ref. 36). The MT·Kar3Cik1 dissociation kinetics promoted by ADP were quite different whereby there was no evidence that ADP promoted Kar3Cik1 detachment from the MT. Chen et al. (36) proposed that ADP, when bound to Kar3, led to the formation of the MT·Kar3Cik1·ADP intermediate in which the Cik1 head was tightly bound to the MT and the Kar3·ADP head was weakly bound or detached but tethered to the MT through Cik1. These differences would suggest that the E2 state may be more stable for Kar3Cik1 than Kar3Vik1. The last major difference was in the kinetics of MT·motor association followed by mantADP release. The kinetics for Kar3Vik1 indicated that there was an asymmetry in the dimer such that Kar3Vik1 was biased to bind through Vik1 for productive MT·Kar3Vik1 complex formation. In contrast, the kinetics for Kar3Cik1 revealed that ∼⅓ of the population collided with the MT initially by Kar3 but that the productive collision complex occurred through Cik1 followed by Kar3 MT binding (36). Despite these distinctive differences, the mechanistic studies indicate that Kar3Cik1, at least at conditions for MT cross-linking, will also use an ATP-promoted powerstroke mechanism with ATP hydrolysis intrinsic to Kar3 and occurring as the postpowerstroke intermediate.

In summary, the results presented here show that there is a conserved mechanism for kinesin-14 Ncd and Kar3Vik1 with evidence suggesting that Kar3Cik1 will show similarities. The key questions ahead are directed at Kar3Cik1 to determine whether it also exhibits an ATP-promoted powerstroke, whether a powerstroke mechanism is used during Kar3Cik1-promoted MT shortening, and the structural differences in Cik1 that change the MT interactions for specific cellular functions.