A snapshot from a gliding assay is shown in Figure 2. A good microtubule density is 1-10 microtubules per field of view; substantially more will result in mistracking as microtubules cross each other. A plot of the 11 microtubule trajectories from the gliding assay in Figure 2 is shown in Figure 3. Typical trajectories are 10 to 30 μm long; some trajectories have gaps where one microtubule crosses another. These trajectories may be discarded from analysis.
A single, long microtubule trajectory is shown in Figure 4A with example tangent angles at two positions along the trajectory. The difference between tangent angles separated by a fixed distance is related to the persistence length; the tangent angle as a function of position along the microtubule trajectory is shown in Figure 4B.
The tangent angle data from many microtubule trajectories is combined to calculate a single persistence length for all microtubules in a given experiment. Figure 5 shows a plot of θs> versus contour length s for the gliding assay of Figure 2. At large contour length values, very few values of cosθ are measured; hence the average is highly variable. The weighted fit conforms to the short s, high-precision data more closely than the long s, low-precision data.
The value of persistence length from these 11 trajectories, 500 ± 40 μm (± standard error of the mean), is representative of persistence lengths for relatively short microtubules2. Similar experiments give a range of persistence lengths between 300 and 1,000 μm.
Troubleshooting
If microtubules are completely absent, increase the concentration of microtubules in step 3.9 to 0.5 mg/ml. If microtubules are still missing, re-polymerize microtubules. Make sure paclitaxel is present in each solution microtubules are diluted into, otherwise microtubules will depolymerize.
If microtubules are freely diffusing in solution but not binding to surface, increase the kinesin concentration in step 3.7 and use AMP-PNP instead of ATP to ensure kinesin binds microtubules irreversibly. If microtubules still do not bind, make new streptavidin stock and buffer, followed by new biotin-BSA buffer. Finally, purify new biotinylated kinesin.
If microtubules bind but do not move, replace ATP stock solution with fresh ATP.
If fluorophores photobleach too quickly, reduce illumination intensity. If fluorophores still photobleach too quickly, replace oxygen scavenging system with fresh stock.
If fluorophores are too dim, increase illumination intensity.

Figure 1. Cartoon of the microtubule gliding assay. Kinesin enzymes are specifically bound to a coverslip by a biotin-streptavidin linkage. Microtubules are sparsely labeled with organic fluorophores. Upon addition of ATP, microtubules are pushed by the kinesin motors. The leading free end of the microtubule fluctuates due to thermal forces in solution; these fluctuations are used to calculate the microtubule persistence length. The length-scale of the microtubule probed by these experiments is the length of the free end.

Figure 2. Typical microscope snapshot of a gliding assay, taken via TIRF microscopy. Microtubules are sparsely decorated with single fluorophores. Scale bar is 5 μm.

Figure 3. Microtubule trajectories from the image sequence shown in Figure 2. Each microtubule trajectory combines many single fluorophore trajectories (about 10 on average), and have been thinned to one point per 100 nm.

Figure 4. Calculating the tangent angles to a trajectory. (A) Trajectory of one microtubule, with example tangent angles (θ) shown. (B) Tangent angle as a function of position along the microtubule trajectory. These data are used to calculate the average angles used in Eq. 1.

Figure 5. Calculating persistence length from tangent angles. (dots) Plot of θs>versus contour length s for the 11 trajectories shown in Figure 3. (solid line) Fit to Eq. 1. For this group of microtubules, the persistence length is 500 ± 40 μm. For long contour lengths (above 10 μm or so), the data are highly variable due to limited statistics.