The movement of single GFP-tagged motors was visualized by total internal reflection fluorescence (TIRF) microscopy. Surprisingly, despite the dramatic changes to the neck linker, all of the extended kinesins moved processively, taking >100 steps along the microtubule without dissociating. The average run lengths were similar to wild-type kinesin (WT) (Figures 1B and S2). Even the longest insertion (26P) displayed a run length that was 60% of WT. These findings contrast with the results from prior ATPase studies by Hackney et al. (2003), which suggested that insertions to the neck region severely impair kinesin processivity. Thus, our direct single molecule measurements show that kinesin can maintain its high processivity even after diminishing the mechanical tension between its catalytic domains.
The extended kinesins, however, moved significantly slower than WT. The velocity of the proline insertion constructs decreased progressively with the number of inserted proline residues, with the longer inserts (13P, 19P, and 26P) moving at ∼5-fold lower rates than WT (Figure 1C). The more flexible 14GS insertion resulted in an even greater decrease (10-fold lower than WT). Slow movement of the constructs was not due to impaired ATP hydrolysis, since the maximal MT-stimulated ATP turnover rates (kcat) were similar between WT and the extended kinesins (Figure 1D). This result indicates that the extended kinesins are less efficient in converting ATP hydrolysis energy into productive unidirectional motion. An estimated coupling ratio (Figure 1E) revealed that WT kinesin is at least 80% efficient in converting ATPs hydrolyzed into forward steps, whereas the extended kinesins show a much lower coupling efficiency (as low as ∼10% for the long extended kinesins).