Highly purified subsynaptic membrane fragments prepared from Torpedo marmorata electric organ (specific activity, >4 imol of Na'a nigricoilis a-[3H toxin per mg of protein) exhibit, on sodium dodecyl sulfate/polyacrylamide gel electrophoresis, two major protein bands of apparent molecular weight 40,000 and 43,000, respectively. Dissolution of these membranes by the. nondenaturing detergents Triton X-100 and Berol 043 followed by standard fractionation yielded (j)the 9S acetycholine-receptor protein which still binds the cW4ltoxin and after further purification-yielded, in the presence of sodium dodecyl sulfate, the 40,OOWdalton component, covalently labeled by the affinity reagent 4-(N-maleimido)phenyl[3H]trimethylammonium; only serine was found as the NHrterminal amino acid of this protein; and (iil) a high molecular weight aggregate named 43,000 protein which was resolved in denaturing gels almost exclusively as the 43,000-dalton band. In the absence of detergents, the 43,000 protein binds compounds known to interact with the acetylcholine ionophore: a fluorescent local anesthetic quinacrine and histrionicotoxin (apparent dissociation constant, 7 i 1 x 10-7 M). The regulation of quinacrine fluorescence by carbamylcholine, observed in the intact membrane, no longer occurs with the isolated 43,000 component. The elementary functional unit that accounts for the regulation by acetylcholine (ACh) of cation translocation through an excitable membrane comprises, a priori, two main categories of sites: (i) the ACh receptor site, which binds cholinergic agonists, antagonists (1), and snake venom a-toxins (2); and (ii) a site involved in the selective permeation of small cations. It was further postulated that these two sites were topographically distinct, although coupled by "allosteric" interactions and most likely carried by different polypeptide entities referred to as the ACh-receptor protein sensu stricto, and the ACh ionophore (3) [or ion conductance modulator (4)]. The ACh-receptor protein has been isolated and purified from fish electric organ and skeletal muscle in several laboratories (see ref. 5) on the basis of its ability to bind the snake. a-toxins.Local anesthetics such as prilocaine, tetracaine, and Quotane affect in vwo (6) and in vitro (7) the permeability response to bath-applied ACh in a manner rather different from that of "competitive" antagonists like curare or Flaxedil. These local anesthetics block the response by decreasing the amplitude without changing the apparent dissociation constant for the agonists. When individual current pulses corresponding to single ionophore openings are recorded in the presence of one of these anesthetics, the square contour of the pulse becomes chopped by multiple fast blocking events that are interpreted as representing the repetitive and reversible binding of the local anesthetic directly to the ion gate (E. Neher, personal com-