Skip to main content
. 2020 Dec 21;9:e62876. doi: 10.7554/eLife.62876

Figure 6. KIF21B prevents centrosome stalling induced by force balance during polarization by limiting MT length.

(A) The T cell model is initialized to a steady state MT network in 300 s (Phase I) and is then extended to a polarizing model (Phase II). In this transition, the cell shape is changed to include a flat side with curved corners with 50 dynein units attached, and the centrosome confinement (red) is released to allow centrosome movement and translocation. (B) Snapshots of modeled T cells during centrosome polarization in the KIF21B-pausing model with the presence of 10 (top row) or 0 (bottom row) KIF21B motors. The simulations were stopped after the centrosomes reached a polarized position <2.5 µm away from the synapse. The final snapshot in each row indicates the frame of the simulation when the centrosome was fully polarized. Timepoints are indicated below each snapshot. (C) Quantification of polarization time (the time the centrosome needs in Phase II to reach a polarization position) per simulated cell for the non-pausing model at indicated catastrophe rates. Some cells did not polarize within the maximum simulation time (800 s) as indicated in the upper part of the graph. n = 30 simulated cells per condition. (D) Quantification of polarization time (the time the centrosome needs in Phase II to reach a polarization position) per simulated cell for the KIF21B-pausing model with indicated numbers of KIF21B motors present. Some cells did not polarize within the maximum simulation time (800 s) as indicated in the upper part of the graph. n = 30 simulated cells per condition. (E) Trajectories showing the centrosome-synapse distance during Phase II over time along the vertical axis of the simulated T cell. Trajectories are color coded for the number of KIF21B motors present (see legend). Shown are 15 runs per condition. (F) Quantification of force imbalance within the KIF21B-pausing model. Force imbalance is calculated from the forces experienced by dynein at a given time point, projected on the horizontal axis. Force imbalance was characterized as the absolute difference of projected forces pointing to the left and forces pointing to the right, normalized by the sum of all forces. This quantity is determined for each timestep between 0 and the polarization time for every repeat and binned to create the according histogram. Histograms are color coded for the number of KIF21B motors present in the simulation (see legend). n = 30 simulated cells per condition. (G) Quantification of the mean number of MTs that are bound by dynein to the synapse within the KIF21B-pausing model. Violin plot distributions show the total number of MTs bound (purple), and the absolute difference (|MTL-MTR|, yellow) between the number of MTs bound on the right versus left side of the cell’s vertical axis. Within each violin, a boxplot indicates the quartiles (black) with the inner quartile range (25–75%) indicated with the thickened region. Median values are indicated for each violin (white dot). Values were obtained in the first 10 s of Phase II of each simulation. n = 30 simulated cells per condition. (H) Trajectories of centrosome-synapse distance over time (solid lines) plotted together with trajectories of the absolute difference |MTL-MTR| between the number of MT bound to dynein and passing the nucleus on the left or on the right (dashed lines) in the same color for the corresponding runs. To make the |MTL-MTR| more legible, the mean and 95% confidence interval is plotted averaged over three frames (1.2 s). Trajectories are color coded for the number of KIF21B motors present in the simulation (see legend).

Figure 6—source data 1. A CSV file with numerical data of polarization time as a function of catastrophe rates represented in Figure 6C.
Figure 6—source data 2. A CSV file with numerical data of polarization time as a function of the numbers of KIF21B motors represented in Figure 6D.
Figure 6—source data 3. A CSV file with numerical data of the distance from the centrosome to the synapse as a function of time for different numbers of KIF21B motors, shown in Figure 6E.
Figure 6—source data 4. A CSV file with numerical data of force imbalance per time trace and per time point for different numbers of KIF21B motors, as plotted in Figure 6F.
Figure 6—source data 5. A CSV file with numerical data of the mean numbers of MT bound to dynein for different numbers of KIF21B motors, as shown in Figure 6G.
Figure 6—source data 6. A CSV file with numerical data of time traces of the centrosome-synapse distance and time traces of the difference between the number of MTs bound by dynein passing along the right and the left side of the nucleus.
The data is given for different numbers of KIF21B motors and shown in Figure 6H.

Figure 6.

Figure 6—figure supplement 1. KIF21B affects the balance of force in T cell polarization.

Figure 6—figure supplement 1.

Histogram showing the force on dynein at every timestep. The forces on dynein along the x-axis are summed, where left-pointing forces are taken as negative. This sum is taken for every recorded time between 0 and the polarization time of the simulation and binned to create histograms. Histograms are color coded for the number of KIF21B motors present in the simulation (see legend).
Figure 6—figure supplement 1—source data 1. A CSV file with numerical data of summed horizontal forces on dynein per time trace and per time point for different numbers of KIF21B motors, as shown in Figure 6—figure supplement 1.
Appendix 1—table 1. Parameters used for simulations.
Figure 6—video 1. Recording of a simulated T cell in the ‘non-pausing’ model with 10 KIF21B motors added.
Download video file (25.8MB, mp4)
The T cell simulation consists of two phases. At first, the model is initialized to a steady state MT network in the first 300 s, during which the centrosome is confined to the red box at one side of the T cell. Next, the cell shape is changed to include 50 membrane-anchored dynein motors (blue) at a flat immunological synapse surface, and centrosome confinement is released to allow centrosome translocation. KIF21B motors are included throughout the entire simulation (orange). Snapshots of this run are shown in Figure 6B (upper row). Growing MT plus ends are indicated by yellow arrowheads, depolymerizing MT plus ends are indicated by red arrowheads, and KIF21B units by orange circles.
Figure 6—video 2. Recording of a simulated T cell in the ‘non-pausing’ model with 0 KIF21B motors added.
Download video file (42.4MB, mp4)
The T cell simulation consists of two phases. At first, the model is initialized to a steady state MT network in the first 300 s, during which the centrosome is confined to the red box at one side of the T cell. Next, the cell shape is changed to include 50 membrane-anchored dynein motors (blue) at a flat immunological synapse surface, and centrosome confinement is released to allow centrosome translocation. Snapshots of this run are shown in Figure 6B (bottom row). Growing MT plus ends are indicated by yellow arrowheads and depolymerizing MT plus ends are indicated by red arrowheads.