Mathematical Biology Seminar: Microtubule assembly dynamics at the nanoscale

Microtubules are intracellular polymers that dynamically grow and shorten at their ends via the stochastic addition and loss of αβ-tubulin heterodimers, a highly regulated process that underlies many fundamental cellular processes, including chromosome segregation and cell polarization. Previously, the rates of tubulin subunit exchange at the ends of growing microtubules have been estimated using a 1D linear growth theory, which assumes that tubulin dissociation occurs at a constant rate regardless of the free subunit concentration. We now find via 2D molecular-level simulations that the tubulin dissociation rate during microtubule growth is not expected to be constant, but rather will increase with increasing free subunit concentration. This effect is due to a concentration-dependent bias in simulated microtubule tip structures, as has been experimentally observed. As a consequence, we predict theoretically that the published subunit addition and loss rates at growing microtubule ends in vitro have been consistently underestimated in the literature by an order-of-magnitude. We then test this prediction experimentally via TIRF-microscopy and via a laser-tweezers assay with near-molecular resolution, and find that the variance in the assembly rate in vitro is too high to be consistent with the previous low kinetic rate estimates. In contrast, the 2D model, with kinetic rates that are an order-of-magnitude higher than the 1D model kinetic rates, quantitatively predicts a priori the variance and its concentration dependence. We conclude that net assembly is the result of a relatively small difference between large rates of subunit addition and loss, both of which occur at near-kHz rates, far faster than previously believed. More generally, our theoretical analysis demonstrates that the fixed off rate originally used in the 1D model of Oosawa, and assumed in most subsequent models, is problematic for self-assembled polymers having both lateral and longitudinal bonding interactions between subunits. Our results imply a major revision of how microtubule assembly is likely regulated in vivo.