Activation of membrane P2X₇ receptors by extracellular ATP results in the opening within several milliseconds of an integral ion channel that is permeable to small cations. If the ATP application is maintained for several seconds two further sequelae occur: there is a gradual increase in permeability to the larger cation N-methyl-D-glucamine, and the cationic propidium dye YO-PRO-1 (quinolinium,4-[(3-methyl-2-(3H)-benzoxazolylidene)methyl]-1-[3-(triethylammonio)propyl]di-iodide) enters the cell. The similarity in the time course of these two events has led to the widespread view that N-methyl-D-glucamine and YO-PRO-1 enter through a common permeation pathway, the "dilating" P2X₇ receptor pore. Here we provide two independent lines of evidence against this view. We studied single human embryonic kidney cells expressing rat P2X₇ receptors by patch clamp recordings of membrane current and by fluorescence measurements of YO-PRO-1 uptake. Firstly, we found that maintained application of the ATP analog did not cause any increase in N-methyl-D-glucamine permeability when the extracellular solution contained its normal sodium concentration, although YO-PRO-1 uptake was readily observed. Secondly, we deleted a cysteine-rich 18 amino acid segment in the intracellular juxtamembrane region of the P2X₇ receptor. This mutated receptor showed normal YO-PRO-1 uptake but had no permeability to N-methyl-D-glucamine. Taken together, the clear differential effects of extracellular sodium ions or of mutating the receptor strongly suggest that N-methyl-D-glucamine and YO-PRO-1 do not enter the cell by the same permeation pathway.