We used a series of adamantane derivatives to probe the structure of the phencyclidine locus in either the resting or desensitized state of the nicotinic acetylcholine receptor (AChR). Competitive radioligand binding and photolabeling experiments using well-characterized noncompetitive antagonists such as the phencyclidine analogue [piperidyl-3,4-3H(N)]-N-[1-(2-thienyl)cyclohexyl]-3,4-piperidine ([3H]TCP), [3H]ethidium, [3H]tetracaine, [14C]amobarbital, and 3-(trifluoromethyl)-3-(m-[125I]iodophenyl)diazirine ([125I]TID) were performed. Thermodynamic and structure-function relationship analyses yielded the following results. (1) There is a good structure-function relationship for adamantane amino derivatives inhibiting [3H]TCP or [3H]tetracaine binding to the resting AChR. (2) Since the same derivatives inhibit neither [14C]amobarbital binding nor [125I]TID photoincorporation, we conclude that these positively charged molecules preferably bind to the TCP locus, perhaps interacting with αGlu262 residues at position M2-20. (3) The opposite is true for the neutral molecule adamantane, which prefers the TID (or barbiturate) locus instead of the TCP site. (4) The TID site is smaller and more hydrophobic (it accommodates neutral molecules with a maximal volume of 333 ± 45 Å3) than the TCP locus, which has room for positively charged molecules with volumes as large as 461 Å3 (e.g., crystal violet). This supports the concept that the resting ion channel is tapering from the extracellular mouth to the middle portion. (5) Finally, although both the hydrophobic environment and the size of the TCP site are practically the same in both states, there is a more obvious cutoff in the desensitized state than in the resting state, suggesting that the desensitization process constrains the TCP locus. A plausible location of neutral and charged adamantane derivatives is shown in a model of the resting ion channel.