Force- and length-dependent catastrophe activities explain interphase microtubule organization in fission yeast

Abstract

The cytoskeleton is essential for the maintenance of cell morphology in eukaryotes. In fission yeast, for example, polarized growth sites are organized by actin, whereas microtubules (MTs) acting upstream control where growth occurs. Growth is limited to the cell poles when MTs undergo catastrophes there and not elsewhere on the cortex. Here, we report that the modulation of MT dynamics by forces as observed in vitro can quantitatively explain the localization of MT catastrophes in Schizosaccharomyces pombe. However, we found that it is necessary to add length-dependent catastrophe rates to make the model fully consistent with other previously measured traits of MTs. We explain the measured statistical distribution of MT-cortex contact times and re-examine the curling behavior of MTs in unbranched straight tea1Delta cells. Importantly, the model demonstrates that MTs together with associated proteins such as depolymerizing kinesins are, in principle, sufficient to mark the cell poles.

Figure 1

(A) S. pombe strain expressing GFP–tubulin and the nuclear pore marker Nup85–GFP. (B) The 3D simulation contains a spherical nucleus radius of 1.3 μm and MT bundles attached to it. (C) The cell of half-length 5.5 μm is a cylinder closed by half-spheres, of radius 1.6 μm. In each bundle, four MTs overlap near their minus ends, where they are linked to the nucleus. For more information, see Supplementary information.

Figure 2

Trait success–failure diagram. (A) To evaluate model FL, we systematically varied the growth speed v₀ (x axis) and catastrophe rate c₀ (y axis). Each symbol depicts the outcome of one simulation, summarizing its conformity with the 10 traits of Table I (see legend). (B–D) Snapshots from the simulations circled in red in (A).

Figure 3

Simulations compared with experiments. (A) MT contact times with the cortex at the cell poles. Bars: histogram of 303 measured events from 343 min of live imaging. Line: distribution predicted by model FL, for the standard parameter values (see Supplementary information). (B) Predicted density of catastrophes in the cell, as a function of longitudinal (∣x∣) and radial (R=√y²+z²) positions. Top: bundle catastrophes occur mostly at the cell poles as observed. Bottom: hidden catastrophes are more distributed. (C) Fluorescence microscopy images of wild-type (left) and tea1Δ cells (right) with GFP–α2 tubulin. Compared with wild type, MTs are curling more strongly in tea1Δ cells. (D) Simulated and (E) experimental quantification of the MT curling phenotype: the angle between the cell axis and the MT plus end, at the time of catastrophe, is plotted as a function of cell diameter and cell length. Open symbols on the experimental plot correspond to wild-type cells, and closed symbols to tea1Δ cells. The best linear fit to the data is shown, as well as the envelope of many similar fits obtained for other simulation results, which differ only by their random number seed. The slope of the best fit and its standard deviation are indicated. (F) Distribution of measured cell widths. Source data is available for this figure at www.nature.com/msb