The improved characterization of membrane proteins in recent years is leading to more carefully controlled experiments to determine the nature of their interaction with phospholipids. Often it is only a minor segment of the protein which interacts directly with the membrane, the main functional part of the molecule being in the aqueous phase, closely resembling related water-soluble proteins (e.g. cytochrome b5 resembles cytochrome c; Mathews et al., 1971). It is often possible to release this segment intact by proteolysis. For example, active fragments of influenza haemagglutinin (Brand & Skehel, 1971) and neuraminidase (Wrigley et al., 1973), of cytochrome b5 (Spatz & Strittmatter, 1971) and of cytochrome b5 reductase (Spatz & Strittmatter, 1973) have been obtained in this way. The lipid-binding segment is usually short (30-100 residues) and it is not known whether it has a relatively non-specific anchoring function or whether it plays a more positive organizational role. The sequence and conformation of these segments should soon be known; for the moment we can suggest that, like other hydrophobic peptides in non-polar solvents, the conformation is probably helical. When lipid and detergent are removed from these proteins they form small irregular rosettes, aggregated through their hydrophobic tails, and do not lose their biological activity. Intrinsic membrane proteins with gating or transport function [e.g. acetylcholine receptors, rhodopsin, ion-transporting ATPases (adenosine triphosphatases)] interact much more extensively with lipid and are consequently more difficult to handle. It is usually impossible to obtain them in an active state without bound lipid and/or detergent. Removal of detergent leads to extensive irregular aggregates, so that it has proved difficult or impossible to characterize them by hydrodynamic methods or to determine the size of the basic functional unit. The sort of questions we would like to answer concern the extent of the hydrophobic interaction between lipid and protein, the sequence and conformation of the protein, the specificity of its interaction with lipid and its stability in the absence of lipid. It should then be possible to make a realistic assessment of the function of the hydrophobic domain. Does it have active carrier or catalytic function or does it merely provide an aqueous channel for communication across the bilayer? We have been attempting to answer some of these questions for the Caz+-transporting ATPase of sarcoplasmic reticulum by proteolytic fragmentation (Thorley-Lawson & Green, 1973) and by controlled delipidation (Hardwicke & Green, 1974). By way of introduction it should be said that about half of this protein (mol.wt. 115OOO) projects into the aqueous phase, while the remainder can be seen as an intramembrane particle in freeze-fractured preparations. When dissolved in minimal deoxycholate it can be purified as a lipoprotein (80-100 mol of phospholipid/mol), which regenerates vesicles on dilution or dialysis (Maclennan et al., 1971). We have recently lowered the lipid content to less than 5 mol/mol without irreversible inactivation by carefully controlled washing of the protein, solubilized in Triton X-100 and adsorbed on DEAE-cellulose, in the presence of Ca2+ and glycerol. In this way, we have defined threeconformational states of the ATPase in termsofactivityandappearance in electron micrographs. These may be termed active, reversibly inactivated and inactive conformations. The active preparations (25-30 mol of P, 50 mol of Triton X-l00/ mol) were obtained by immediate elution with a salt gradient (cf. Migala & Hasselbach, 1972). Removal of detergent gave active vesicles similar in appearance to the parent membrane, but somewhat fragmented. Washing with 2 column vo1. of 0.2 %Triton X-100 gave fully activatable preparations (3-5mol of P, 70mol of Triton X-l00/mol) after similar elution. The initial activity was less than 5 % of that obtained after various lipids were added. Removal of Triton gave small (2CMOnm) membranous fragments with 6nm projections. Attempts to remove the last few molecules of lipid by more extensive washing have given totally inactive preparations, which require high salt con-
Read full abstract