Concentration dependence of variability in growth rates of microtubules

Growth and shortening of microtubules in the course of their polymerization and depolymerization have previously been observed to occur at variable rates. To gain insight into the meaning of this prominent variability, we studied the way in which its magnitude depends on the growth rate of experimentally observed and computer-simulated microtubules. The dynamic properties of plus-ended microtubules nucleated by pieces of Chlamydomonas flagellar axonemes were observed in real time by video-enhanced differential interference contrast light microscopy at differing tubulin concentrations. By means of a Monte Carlo algorithm, populations of microtubules were simulated that had similar growth and dynamic properties to the experimentally observed microtubules. By comparison of the experimentally observed and computer-simulated populations of microtubules, we found that 1) individual microtubules displayed an intrinsic variability that did not change as the rate of growth for a population increased, and 2) the variability was approximately fivefold greater than predicted by a simple model of subunit addition and loss. The model used to simulate microtubule growth has no provision for incorporation of lattice defects of any type, nor sophisticated geometry of the growing end. Thus, these as well as uncontrolled experimental variables were eliminated as causes for the prominent variability.