The Bacterial Photosynthetic Reaction Center as a Model for Membrane Proteins

Membrane proteins participate in many fundamental cellular processes. Until recently, an understanding of the function and properties of membrane proteins was hampered by an absence of structural information at the atomic level. A landmark achievement toward understanding the structure of membrane proteins was the crystallization (1) and structure determination (2-5) the photosynthetic reaction center (RC) from the purple bacteria Rhodopseudomonas viridis, followed by that of the RC from Rhodobacter sphaeroides (6-17). The RC is an integral membrane protein-pigment complex, which carries out the initial steps of photosynthesis (reviewed in 18). RCs from the purple bacteria Rps. viridis and Rb. sphaeroides are composed of three membrane-associated protein subunits (designated L, M, and H), and the following cofactors: four bacteriochlorophylls (Bchl or B), two bacteriopheophytins (Bphe or phi), two quinones, and a nonheme iron. The cofactors are organized into two symmetrical branches that are approximately related by a twofold rotation axis (2, 8). A central feature of the structural organization of the RC is the presence of 11 hydrophobic alpha-helixes, approximately 20-30 residues long, which are believed to represent the membrane-spanning portion of the RC (3, 9). Five membrane-spanning helixes are present in both the L and M subunits, while a single helix is in the H subunit. The folding of the L and M subunits is similar, consistent with significant sequence similarity between the two chains (19-25). The L and M subunits are approximately related by the same twofold rotation axis that relates the two cofactor branches. are the first membrane proteins to be described at atomic resolution; consequently they provide an important model for discussing the folding of membrane proteins. The structure demonstrates that alpha-helical structures may be adopted by integral membrane proteins, and provides confirmation of the utility of hydropathy plots in identifying nonpolar membrane-spanning regions from sequence data. An important distinction between the folding environments of water-soluble proteins and membrane proteins is the large difference in water concentration surrounding the proteins. As a result, hydrophobic interactions (26) play very different roles in stabilizing the tertiary structures of these two classes of proteins; this has important structural consequences. There is a striking difference in surface polarity of membrane and water-soluble proteins. However, the characteristic atomic packing and surface area appear quite similar. computational method is described for defining the position of the RC in the membrane (10). After localization of the RC structure in the membrane, surface residues in contact with the lipid bilayer were identified. As has been found for soluble globular proteins, surface residues are less well conserved in homologous membrane proteins than the buried, interior residues. Methods based on the variability of residues between homologous proteins are described (13) ; they are useful (a) in defining surface helical regions of membrane and water-soluble proteins and (b) in assigning the side of these helixes that are exposed to the solvent. A unifying view of protein structure suggests that water-soluble proteins may be considered as modified membrane proteins with covalently attached polar groups that solubilize the proteins in aqueous solution.