Cooperative binding of KIF5C to GDP microtubules. (A–D) The binding affinity of the KIF5C motor domain (monomer) to GDP or GMPCPP microtubules (MTs) was examined using single-molecule assays in the absence of nucleotide (A and B) or in the presence of 1 mM AMPPNP (C and D). (A and C) Examples of the raw results. Each image shows the KIF5C motor domain (magenta) bound on GDP microtubules (left columns) or GMPCPP microtubules (right columns) in the absence of nucleotide (A) or in the presence of 1 mM AMPPNP (C). The concentration of the kinesin used for each experiment is shown on the left side of the columns. (B and D) Distribution of the density of bound KIF5C on the microtubules. The results were fitted with a mixed Gaussian model to examine the heterogeneity in binding affinity. Two-peak distribution was evident with nucleotide-free KIF5C and GDP microtubules (B, left). See Tables S4, S5, and S6 for details. (E–H)K_d Dissociation constants of KIF5C. Affinity (equilibrium K_d) plotted against the concentration of KIF5C monomeric motor domain. Error bars indicate SEM (see Materials and methods for details). (E–G) Affinity of the wild-type KIF5C motor domain to GDP, GMPCPP, or Taxol-stabilized GDP microtubules. The nucleotide conditions for kinesin and the microtubule conditions are indicated in each panel. (H) Affinity of the L11 mutant KIF5C (K11C, the L11-α4 junction was replaced with the corresponding sequence of KIF1A). This mutant showed constant affinity to both GDP and GMPCPP microtubules, which is consistent with our previous report that this mutant cannot discriminate axonal microtubules from dendritic microtubules in neurons (Nakata et al., 2011). The numbers of the observed microtubules and bound KIF5 molecule are summarized in Tables S4, S5, and S6.