We also observed an inherent asymmetry in the force-induced movement, with faster velocities observed for an equal magnitude pull directed toward the microtubule plus end, compared with the minus end (Figure 5A). This result is consistent with the lower unbinding forces measured when kinesin is pulled toward the plus end, compared with the minus end (Uemura and Ishiwata, 2003) (see Discussion).
We next tested whether an applied force can induce movement in the presence of AMPPNP (a nonhydrolyzable ATP analog), which causes kinesin to bind more tightly to microtubules than in nucleotide-free conditions (Kawaguchi and Ishiwata, 2001), and ADP, which induces a weak binding state. With 1 mM AMPPNP, a 9 pN force was required to induce stepping toward the plus end, whereas higher forces of 12 pN were required to produce minus-end-directed stepping (Figure 5B). A 6 pN load, which induced stepping under nucleotide-free conditions, did not lead to stepping with 1 mM AMPPNP (Figure S11). In the presence of ADP, much lower forces of ∼1 pN were sufficient to move kinesin toward the plus end (see Experimental Procedures) (Figure 5C). The velocity of movement was dependent on the ADP concentration (Figure 5C, insert), which likely reflects the relative time that the microtubule-bound kinesin head spends in a nucleotide-free state (resistant to detachment with a 1 pN load) versus a weakly bound ADP state (see also Uemura and Ishiwata, 2003). Once again, a higher load (2 pN) was required to drive the movement toward the minus end in the presence of ADP. In contrast to Uemura and Ishiwata (2003), we found that lower forces (1–2 pN) are sufficient to allow head detachment with bound ADP, compared with their reported average unbinding forces of 3–4 pN. The higher force measured by Uemura and Ishiwata is likely explained by their use of a constantly moving trap (instead of a force-feedback trap), which leads to a rapidly increasing load and the measurement of larger forces (depending on the time constant of detachment). In summary, these experiments reveal that the amount of force required for stepping is most likely determined by the affinity of the rear kinesin head for the microtubule. In addition, there is a clear directional asymmetry for detachment under all nucleotide conditions, with less force required for trailing head detachment when the pull is in the normal direction of kinesin movement.
We next explored whether a pulling force might substitute for a force-inducing conformational change of the neck linker. To test this idea, we deleted the native neck linker sequence that docks onto the catalytic core and added 19 prolines to act as a spacer between the catalytic core and coiled-coil (19P-NL) (Figure 6A). This construct displayed no detectable movement in single-molecule assays with 1 mM ATP (Figure 6B). Remarkably, an external force (3 or 6 pN) from the optical trap caused the “immotile” 19P-NL mutant to step along the microtubule (Figure 6C) with a similar rate to that of WT in the absence of ATP. Load-induced movement in the presence of 1 mM ATP for 19P-NL was ∼4-fold faster (84 ± 6 nm/s) than that under nucleotide-free conditions. Although immotile, 19P-NL has ATPase activity (5-fold lower than WT), which enables its trailing head to transit from a tight (nucleotide-free) to a weak (ADP) binding state. As described above, the ADP state is more susceptible to detachment under load, which most likely accounts for the higher velocities of 19-NL in the presence of ATP. In summary, these results show a kinesin lacking its mechanical element and ATP, can undergo 8 nm unidirectional stepping, if external tension is applied to the motor.