Kinesin hydrolyses one ATP per 8-nm step

Mark J. Schnitzer*† & Steven M. Block†‡

Departments of * Physics and † Molecular Biology, and ‡ Princeton Materials Institute, Princeton University, Princeton, New Jersey 08544, USA

Nature 24 Juli 1997, vol. 388, pp. 386-390

Basics

Kinesin:Two-headed ATP driven motor protein that moves along the

microtubules in discrete steps of 8 nm.

Central question:How many molecules of ATP are consumed per step?

Method:Using the processivity of kinesin, statistical analysis of intervals

between steps at limiting ATP and studies of fluctuations in motor speed as a function of [ATP]

Setup

Silica beads (0.5 m) with nonspecifically bound kinesin captured in an optical trap and deposited onto immobilized microtubules bound to the coverglass. Subsequent movements recorded with optical-trapping interferometry.

Low concentrations of kinesin protein ensures maximum one kinesin per bead.

Applied force: 7 fN nm-1 => mean force < 0.9 pN (< 15 % stall)

Velocity vs. [ATP]

Michaelis-Menten kinetics

yielded kcat = 680 ± 31 nms-1 and Km = 62 ± 5 M.

[ATP] independent coupling ratio (# ATP hydrolysed per advance).

Poisson statistics f = 1 - exp(-C) confirms that single molecules suffice to move beads. (2 = 0.7)

Stepping length

Advance of kinesin molecules in clear increments of 8 nm (minority of other step sizes cannot be excluded).

At limiting [ATP] kcat/Km = 11 ± 1 nms-1 M-1 implying stepping rate of 1.4 ± 0.1 M-1 s-1

ATP per step

Exponential distribution => solitary, rate-limiting biochemical reaction (ATP binding) i.e. kinesin requires only one ATP per step.

If (n) ATP molecules were needed the distribution would be a convolution of n exponentials.

Data well fit by single exponential (reduced 2 = 0.6) with a rate of 1.1 ± 0.1 M-1 s-1

(Two exponentials gave 2 = 1.4 with a rate of 0.8 ± 0.06 M-1 s-1)

Fluctuation analysis

For single processive motors, fluctuations about the average speed reflect underlying enzyme stochasticity.

Randomness parameter, r, is a dimensionless measure of the temporal irregularity between steps. For motors that step a distance, d, and whose positions are functions of time, x(t), it’s:

Since both numerator and denominator increase linearly in time, their ratio approaches a constant. The reciprocal of this constant, r-1, supplies a continuous measure of the number of rate-limiting transitions per step.

Robust to sources of thermal and instrumental noise and without need to identify individual stepwise transitions.

Control

Test of determinability of r performed with both simulated and real data.

Simulated: Stochastic staircase records and gaussian white noise.

Real: 2 mM AMP-PNP (non-hydrolysable ATP analogue) and movement of stage.

Both tests positive i.e. stepping statistics could be distinguished.

RandomnessSaturating ATP value implies minimum two

rate-limiting transitions per step. Biochemical pathways also predict r ≈ ½ with assumption of one hydrolysis per step.

For limiting [ATP], r rises through 1, reflecting a single rate-limiting transition once per advance (ATP binding).

Why is r > 1 at limiting [ATP]?Not heterogeneity in ATP binding rate, bead

size or stiffness of bead-kinesin linkage, futile hydrolysis, sticking or transient inactive states.

Maybe backwards movement (7 %) and/or double step (16 %).

Conclusion

Kinesin hydrolyses only one ATP per 8 nm step. Models consistent with this result is:

1) Alternating 16-nm steps by each of the two heads.2) Two shorter substeps of which only one is ATP-dependent and the ATP-independent substep

must be at least as fast as kcat (tsubstep ≈ 15 ms and beneath resolution).

3) Alternatively the ATP-independent substep might be load dependent with a rate slowed with increasing load.

Another future challenge lies in understanding the molecular basis of kinesin movement, since the motor domain of kinesin is quite small (4.5 x 4.5 x 7.0 nm) compared to the 8 nm step.