Treatment of kidney microvillar membranes with the non-ionic detergent Triton X-114 at 0 degrees C, followed by low-speed centrifugation, generated a detergent-insoluble pellet and a detergent-soluble supernatant. The supernatant was further fractionated by phase separation at 30 degrees C into a detergent-rich phase and a detergent-depleted or aqueous phase. Those ectoenzymes with a covalently attached glycosyl-phosphatidylinositol (G-PI) membrane anchor were recovered predominantly (greater than 73%) in the detergent-insoluble pellet. In contrast, those ectoenzymes anchored by a single membrane-spanning polypeptide were recovered predominantly (greater than 62%) in the detergent-rich phase. Removal of the hydrophobic membrane-anchoring domain from either class of ectoenzyme resulted in the proteins being recovered predominantly (greater than 70%) in the aqueous phase. This technique was also applied to other membrane types, including pig and human erythrocyte ghosts, where, in both cases, the G-PI-anchored acetylcholinesterase partitioned predominantly (greater than 69%) into the detergent-insoluble pellet. When the microvillar membranes were subjected only to differential solubilization with Triton X-114 at 0 degrees C, the G-PI-anchored ectoenzymes were recovered predominantly (greater than 63%) in the detergent-insoluble pellet, whereas the transmembrane-polypeptide-anchored ectoenzymes were recovered predominantly (greater than 95%) in the detergent-solubilized supernatant. Thus differential solubilization and temperature-induced phase separation in Triton X-114 distinguished between G-PI-anchored membrane proteins, transmembrane-polypeptide-anchored proteins and soluble, hydrophilic proteins. This technique may be more useful and reliable than susceptibility to release by phospholipases as a means of identifying a G-PI anchor on an unpurified membrane protein.
Phase separation of the receptor for immunoglobulin E and its subunits in Triton X-114.