Detergent binding as a measure of hydrophobic surface area of integral membrane proteins.

The mode by which four nonionic detergents interact with integral membrane proteins was studied by relating detergent binding to characteristic detergent monolayer and micellar dimensions and to the surface area of the hydrophobic sector of the proteins. The detergents used (C12E8, Triton X-100, dodecyl-beta-D-maltoside, and dodecyldimethylamineoxide) are efficient membrane solubilizers, which, according to our hydrodynamic analysis, form small approximately spherical micelles. Their binding by protomeric forms of bacteriorhodopsin, photosynthetic reaction center, sarcoplasmic reticulum Ca(2+)-ATPase, and cytochrome oxidase was measured after equilibration and delipidation, both by the use of successive chromatographies on silica gel and the use of agarose gel columns in combination with DEAE-cellulose chromatography. It was found that, despite detergent binding by silica gel, both chromatographic procedures gave equivalent and consistent results. The data obtained show systematic trends; thus dodecyldimethylamineoxide exhibited the highest binding levels, followed by dodecylmaltoside, whereas the polyethylene glycol detergents C12E8 and Triton X-100 had the lowest degree of binding. This ranking order is inversely related to the cross-sectional areas of these detergents in monolayers at an air-water interface and in micelles. Binding was reduced by aggregation of protomers, whereas the presence of strongly bound residual lipid slightly increased it. Among the membrane proteins with a high molecular mass (reaction center, Ca(2+)-ATPase, and cytochrome oxidase), relative binding could be related to the size of their hydrophobic sectors. On the other hand, bacteriorhodopsin bound somewhat more detergent per transmembrane helix, probably because of less steric hindrance, caused by the absence of a sizable hydrophilic domain in this protein. Theoretical calculations indicated monolayer binding of detergent to the hydrophobic surface of membrane proteins to be a better model for interaction with membrane proteins than binding of the detergents in micellar form.