ABSTRACT:The detergent CHAPS was found to be the preferable surfactant for the efficient purification and reconstitution of the Torpedo californica nicotinic acetylcholine receptor (AChR). The main result is that the incorporation of the AChR proteins into lipid vesicles by CHAPS dialysis was strongly dependent on the salt and protein concentrations. As monitored by sucrose gradients, by electron microscopy, and by agonist-induced lithium ion flux, the best reconstitution yields were obtained in 0.5 M NaCl at a protein concentration of 0.5 g/L and in 0.84 M NaCl at 0.15 g/L protein. Electron micrographs of receptor molecules, which were incorporated into vesicles, showed single, nonaggregated dimer (M, = 580 000) and monomer (M, = 290 000) species. CHAPS dialysis at NaCl concentrations C0.5 M largely reduced the receptor incorporation concomitant with protein aggregation. Electron micrographs of these preparations revealed large protein sheets or ribbons not incorporated into vesicles. The analysis of static and dynamic light scattering demonstrated that the detergent-solubilized AChR molecules aggregate at low lipid contents (5500 phospholipids/AChR dimer), independent of the salt concentration. AChR proteins eluted from an affinity column with a solution containing 8 mM CHAPS (but no added lipid) still contained 130 f 34 tightly bound phospholipids per dimer. The aggregates (about 10 dimers on the average) could be dissociated by readdition of lipid and, interestingly, also by increasing the CHAPS concentration up to 15 mM. This value is much higher than the CMC of CHAPS = 4.0 f 0.4 mM, which was determined by surface tension measurements. The data clearly suggest protein-micelle interactions in addition to the association of monomeric detergents with proteins. Furthermore, the concentration of the (free) monomeric CHAPS at the vesicle-micelle transformation in 0.5 M NaCl ([Owlc = 3.65 mM) was higher than in 50 mM NaCl ( [Owlc = 2.8 mM). However, it is suggested that the main effect of high salt concentrations durig the reconstitution process is an increase of the fusion (rate) of the ternary protein/lipid/CHAPS complexes with mixed micelles or with vesicular structures, similar to the salt-dependent fusion of vesicles.A necessary and essential step in the detailed characterization of intrinsic membrane proteins is the detergent solubilization and the reconstitution of the protein species in artificial lipid membranes. Impressive examples are channel t We gratefully acknowledge financial support by the Deutsche For-

The voltage-gated chloride channel (CIC-0) from the electric organ of Torpedo californica was purified by immunoaffinity chromatography. A polyclonal antibody was shown to specifically recognize the CIC-0 channel (M(r) 85,000-90,000) in a Western blot of total membrane proteins. As monitored by immunoprecipitation, the formation of antibody-antigen complexes in solution strongly depends on the detergent composition. The highest yield of precipitated CIC-0 was obtained from an incubation mixture containing both an anionic detergent, cholate or lauryl sulfate, and the zwitterionic detergent CHAPS. In contrast, immuno-precipitation of CIC-0 was largely reduced when cholate was exchanged for the nonionic detergent Triton x-100, suggesting that the efficient formation of immuno-complexes is favored by negatively charged detergent. In initial immunopurification experiments, in addition to CIC-0 a major contaminating polypeptide of M(r) approximately 115,000 was copurified, which represents the SITS-binding protein [Jentsch et al. (1989) Biochem. J. 261, 155]. Purification of CIC-0 could be increased from about 35% up to 60% homogeneity when immunoaffinity chromatography was performed in the presence of N-acetylglucosamine. Therefore the highly glycosylated SITS-binding protein most likely interacts with the CIC-0 protein via its carbohydrate parts. The purified CIC-0 channel was found to be phosphorylated by PKA in vitro to a level of 0.35-0.4 mol of phosphate incorporated per mol of CIC-0. Proteolytic digestion with endoproteinase GluC and HPLC-separation revealed two major phosphopeptides, which could be identified by amino acid sequence analysis as different size fragments of the same consensus phosphorylation site. Comparison of the peptide sequences with the deduced protein sequence of CIC-0 [Jentsch et al. (1990) Nature 348, 510; O'Neill et al. (1991) Biochem. Biophys. Acta 1129, 131] indicates serine 600 as the phosphorylated residue. Therefore, our results provide strong evidence that CIC-0 is phosphorylated at this single site by PKA in vitro.