Kinetics of microtubule catastrophe assessed by probabilistic analysis

Microtubules are cytoskeletal filaments whose self-assembly occurs by abrupt switching between states of roughly constant growth and shrinkage, a process known as dynamic instability. Understanding the mechanism of dynamic instability offers potential for controlling microtubule-dependent cellular processes such as nerve growth and mitosis. The growth to shrinkage transitions (catastrophes) and the reverse transitions (rescues) that characterize microtubule dynamic instability have been assumed to be random events with first-order kinetics. By direct observation of individual microtubules in vitro and probabilistic analysis of their distribution of growth times, we found that while the slower growing and biologically inactive (minus) ends obeyed first-order catastrophe kinetics, the faster growing and biologically active (plus) ends did not. The non-first-order kinetics at plus ends imply that growing microtubule plus ends have an effective frequency of catastrophe that depends on how long the microtubules have been growing. This frequency is low initially but then rises asymptotically to a limiting value. Our results also suggest that an additional parameter, beyond the four parameters typically used to describe dynamic instability, is needed to account for the observed behavior and that changing this parameter can significantly affect the distribution of microtubule lengths at steady state.