In order to determine the structure of a protein by X-ray crystallography, well ordered three-dimensional crystals are required. However, despite the wealth of experience accumulated in the course of the crystallization and structural analyses of several hundred soluble globular proteins and their complexes, the process of crystallization still remains something of an art, and is often the rate limiting step of any analysis. For membrane proteins that present an additional challenge by virtue of their amphipathic nature, experience is considerably more limited, and the first three-dimensional crystals suitable for X-ray analysis were only reported in 1980 (1 ,2 ). Many membrane proteins form two-dimensional arrays in situ, and these may be studied by electron microscopy and electron diffraction of tilted specimens to determine their three-dimensional structure, but only in the pioneering study of bacteriorhodopsin has the resolution of the structural analysis approached that obtainable by X-ray crystallography (3 ). The formation of two-dimensional crystalline arrays will not be considered in this chapter. The first membrane protein crystal structure, the bacterial photosynthetic reaction center complex, was solved in 1985 (4 ); this was followed by the second reaction center structure in 1986 (5 ,6 ), and more recently porin in 1991 (7 ). Since these pioneering studies, increasing numbers of membrane protein crystallizations have been reported. For recent reviews,seerefs. 8 –11 .
