Path of the polypeptide in bacteriorhodopsin ( purple membrane / diffraction / protein folding
D. Engelman,R. Henderson,A. McLachlan,B. A. WALLACEt
Abstract:An attempt has been made to fit the amino acid sequence of bacteriorhodopsin to the three-dimensional density map of the molecule. First, seven segments of the sequence were selected as being probable transmembrane a helices. Then each of the 5040 possible ways of fitting these seven segments into the seven regions of helical density in the map were evaluated based on the criteria of connectivity of the nonhelical link regions, charge neutralization, and total scattering density per helix. A single model that may be experimentally tested emerged as the most probable. Bacteriorhodopsin is a transmembrane protein found in Halobacter halobium. Light energy absorbed by retinal bound to the protein is used to pump protons across the membrane, and the proton gradient is then used as an energy source (for review see ref. 1). The molecules of this protein exist naturally in a highly ordered two-dimensional array. A low-resolution map (2) of the electron-scattering density of purple membrane shows that each bacteriorhodopsin molecule is composed of seven rods of density oriented perpendicular to the membrane surface. These are believed to represent a-helical segments of the polypeptide that cross the membrane. The molecular boundary of a single bacteriorhodopsin molecule is now known from observation of the same molecule in a different crystal environment (3). The complete amino acid sequence of bacteriorhodopsin has been determined (4) and portions of it have been independently confirmed (5, 6). Ovchinnikov et al. (4) used the proteolytic cleavage points of the protein in the native membrane to propose an arrangement of the polypeptide in the membrane. This consisted of seven a-helical segments of polypeptide, each of which contained between 26 and 32 amino acids, with short nonhelical segments of up to 8 amino acids linking them. Adjacent helices in the sequence had opposite orientations in the membrane. Additional cleavage sites at the NH2 and COOH termini determined by Walker et al. (6) provide further constraints on the polypeptide arrangement. A useful step towards a more detailed description of the structure of bacteriorhodopsin would result from a determination of the way in which the sequence fits into the density map. There are, a priori, 5040 (7!) ways of doing this. There would be 10,080 ways if the orientation of the sequence and density map were not known. However, the correlation of electron diffraction with freeze fracture (7, 8) has established the orientation of the map with respect to the outer surface of the cell. The top of the model in our convention is the cytoplasmic surface. Similarly, digestion experiments on inside-out vesicles have demonstrated that the carboxyl end of the sequence is on the cytoplasmic surface of the membrane (4, 9). In this paper we examine the 5040 possible models by using three criteria: the lengths of links between helix ends, the formation of ion pairs in the protein interior, and the electron scattering power of each helix. We reject a large number of models as unrealistic and propose a single model as the most probable. RESULTS AND DISCUSSION Arrangement of polypeptide in membrane Any considerations limiting the ways in which a model can be constructed depend on having a reliable idea of the arrangement of the polypeptide segments across the membrane. We must, therefore, make an estimate of the exact lengths of helical segments and link regions connecting them before examining ways of fitting the polypeptide segments of the sequence into the density map. Fig. 1 shows a choice of helical segments rather different from that proposed by Ovchinnikov et al. (4). In deciding the exact positions of the ends of the helices, we have endeavored to construct an arrangement conservatively, where the only residues included in the helical segments are those that fit strong criteria of hydrophobicity and inaccessibility to proteolytic cleavage. The accessibility to proteolytic cleavage is clearly a property of surface residues; the hydrophobicity in the helical segments is preferred on energetic grounds for residues that are either buried inside the protein or below the surface of the lipid bilayer. Correspondingly, the occurrence of a high density of charged and polar residues is taken as a strong indication of a nonhelical link region. We have also tried to avoid assumptions that result in either unusually short or long helices because the known membrane thickness, the absence of substantial surface projections, and the three-dimensional density map suggest that the helices should be comparable in length. As a result, the lengths of our nonhelical link regions are rather longer than may eventually be found. For example, they are on average three residues longer than those proposed by Ovchinnikov et al. (4). Several features of Fig. 1 are important, some of which have already been noted (4). First, most of the charged residues (Asp, Glu, Lys, and Arg) are on one or the other surface. On the outer surface of the membrane, there are six charged residues that are either entirely accessible or are close enough to the end of a helix so that their side chains can reach the solvent. Similarly, there are 19 charged residues at the cytoplasmic surface. Second, there are nine charged residues in Fig. 1 that are sufficiently far from either membrane surface to make direct interaction with water unlikely. These charged side chains would be energetically very difficult to bury except as selfneutralizing ion pairs. Of these nine, four may naturally form ion pairs within the same helix because they are separated by either three or four residues in the sequence. These are (Arg-82 and Asp-85) and (Asp-211 and Lys-215). Other ion pairs may * Present address: Department of Molecular Biophysics and Biochemistry, Yale University, Box 1937 Yale Station, New Haven, CT 06520. t Present address: Department of Biochemistry, College of Physicians and Surgeons, Columbia University, New York, NY 10032. 2023 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U. S. C. §1734 solely to indicate this fact. 2024 Biophysics: Engelman et al. Proc. Natl. Acad. Sci. USA 77 (1980) PLEGLW LAAL ILE \ SER SER Ks>PRO