Cryo electron microscopy (cryo-EM) relies on a physical process known as vitrification, in which the temperature of the sample is lowered so quickly that water molecules do not have time to crystallize, instead forming vitreous ice, an amorphous, glass-like solid.
By avoiding the structural disruption caused by expansion during crystallization, cryo-EM preserves biomolecular structures in their natural, fully hydrated state. It literally freezes the sample at an instant in time and can be applied to molecular complexes, whole cells, and tissues, permitting observations of molecular components in association with their functional complements in their native context. The nearly instantaneous nature of the process allows time-resolved analysis of dynamic phenomena such as molecular flexibility, conformational changes, and intermolecular interactions.
High-contrast, high-resolution, low-noise images are difficult to obtain from biological samples because they do not typically contain features such as regular periodic structures or large variations in density or atomic number that generate strong contrast in a TEM; and they do not tolerate the high beam currents or extended exposure times that would build contrast and enhance signal-to-noise ratios. Low-dose imaging refers to a combination of techniques that seeks to address this problem by extracting the maximum information from the sample for each beam electron to which it is exposed.
Two different approaches to 3-D TEM analysis—single-particle analysis and tomographic analysis—have yielded important results. Single-particle analysis combines images from many separate but identical particles to construct a 3-D model.The images of arbitrarily oriented particles are first sorted into categories by similarity, with each category representing a similar orientation.
Images within each category are averaged to create a high-quality, low-noise composite image, then the composite images are combined in a 3-D model. Theoretically, the quality of the model can be improved almost arbitrarily by increasing the number of images.
Tomography computes the 3-D structure by recombining a series of 2-D images acquired over a range of perspectives as the sample rotates in known increments about an axis perpendicular to the beam direction. Its principle advantage is that it looks at a single instance of the imaged structure and can thus be used to evaluate variations among individual particles.