Single kinesin motor molecules were observed to buckle the microtubule
s along which they moved in a modified in vitro gliding assay. In this
assay a central portion of the microtubule was clamped to the glass s
ubstrate via biotin-streptavidin bonds, while the plus end of the micr
otubule was free to interact with motors adsorbed at low density to th
e substrate. A statistical analysis of the length of microtubules buck
led by single motors showed a decreasing probability of buckling for l
oads greater than 4-6 pN parallel to the filament. This is consistent
with kinesin stalling forces found in other experiments. A detailed an
alysis of some buckling events allowed us to estimate both the magnitu
de and direction of the loading force as it developed a perpendicular
component tending to pull the motor away from the microtubule. We also
estimated the motor speed as a function of this changing vector force
. The kinesin motors consistently reached unexpectedly high speeds as
the force became nonparallel to the direction of motor movement. Our r
esults suggest that a perpendicular component of load does not hinder
the kinesin motor, but on the contrary causes the motor to move faster
against a given parallel load. Because the perpendicular force compon
ent speeds up the motor but does no net work, perpendicular force acts
as a mechanical catalyst for the reaction. A simple explanation is th
at there is a spatial motion of the kinesin molecule during its cycle
that is rate-limiting under load; mechanical catalysis results if this
motion is oriented away from the surface of the microtubule.