Ultrastructural Analysis of Microtubule Ends.

Abstract

Microtubules can be detected in light microscopes, but the limited resolution of these instruments means that the polymers appear as lines whose width is defined by the diffraction of light. Much important work on microtubule dynamics has been accomplished by light microscopy, but the details of microtubule end structure are not accessible in such studies. Slight variations in fluorescence intensity, etc. have been used to comment on the structure of dynamic ends, and the combination of light microscopy with laser tweezers has provided insight into aspects of microtubule elongation. However, for views that reveal structural details of the pathways for microtubule growth and shortening, electron microscopy has been of great value. Here, we describe methods for using electron microscopes to look at the ends of microtubules as they grow and shrink, both in vivo and in vitro. The key problems to be overcome for ultrastructural study of microtubule dynamics are those of reliable sample preparation. Dynamic microtubules are labile and can therefore be modified by preparative methods. Our chapter follows the premise that rapid freezing, which converts sample water into vitreous ice, is the best approach for sample preparation. Therefore, all of the methods described involve finding optimal conditions for sample vitrification, and then getting the frozen sample into a form suitable for electron microscopy. We also posit that the end of a microtubule must be considered in three dimensions, so we employ electron tomography as a way to get the necessary information. The methods described for the study of microtubules in cells employ rapid freezing, freeze-substitution fixation, plastic embedding, serial sectioning, and tomography of stained samples. The methods for following microtubule growth in vitro employ sample preparation on holy grids, blotting, and plunge-freezing, followed by electron cryo-tomography. Quantification of structure from both approaches is accomplished by segmentation and analysis of graphic models.