III.E.2. Three-Dimensional Crystals

Crystalline specimens that extend in three-dimensionsose two major problems for image analysis. First, theB>specimen must be thin or be made thin enough (<100 nm) to examine in the microscope. Second, if 3D structural informations required, independent views of the crystal must be photographed and combined. The specimen may be embedded in plastic and thin-sectioned in several, known orientations, or it may be possibleo crush the crystals and produce thin fragments that reveal andequate number of independent views. Thin-sectioning often leads to unsatisfactory results because the preparation procedures are rather drastic, leading to lower resolution of structural details>5 nm). Thin sectioning methods may be essential to study crystals that exist in vivo (e.g. intracellular crystalline enclusions) if they can not be purified in vitro. Crystals grown in vitro may also require thin sectioning, or thin fragments, produced by sonicating or crushing the crystals, maybe suitable specimens.

Several 3D crystalline specimens have been studied 2D image analysis procedures (optical and computer diffraction filtering, and photographic superposition: see Tables 3 and.III.B.10.a,b; Baker, 1981) but only a few have been reconstructed three dimensions (e.g. lipoyl transuccinylase, lipovitellin-phosvitin, myosin S1, nucleosomes, and catalase). It is note worthy that, in each of these examples, structural data were obtained from-ray diffraction experiments along with microscopy. The combination of diffraction amplitudes measured by x-ray or electron diffraction with phases measured from images has proved to be valuable in several structural studies (see Table 1.III.B.8,9, Baker, 1981).