Wild-type human (h) P2X(3) receptors expressed in HEK293 cells responded to the prototypic agonist alpha,beta-methylene ATP (alpha,beta-meATP) with rapidly desensitizing inward currents and an increase in the intracellular Ca(2+) concentration. In contrast to electrophysiological recordings, Ca(2+) microfluorimetry showed a lower maximum of the concentration-response curve of alpha,beta-meATP in the transiently than in the permanently transfected HEK293 cells. However, the concentrations causing 50% of the maximum possible effect (EC(50) values) were identical, when measured with either method. In order to determine the role of certain conserved, positively charged amino acids in the nucleotide binding domains (NBD-1-4) of hP2X(3) receptors for agonist binding, the lysine-63, -65, -176 and -299 as well as the arginine-281 and -295 residues were substituted by the neutral amino acid alanine. We observed no effect of alpha,beta-meATP at the K63A, K176A, R295A, and K299A mutants, and a marked decrease of agonist potency at the K65A and R281A mutants. The P2X(3) receptor antagonist 2',3'-O-trinitrophenyl-ATP (TNP-ATP) blocked the effect of alpha,beta-meATP at the wild-type hP2X(3) receptor with lower affinity than at the mutant K65A, indicating an interference of this mutation with the docking of the antagonist with its binding sites. The use of confocal fluorescence microscopy in conjunction with an antibody raised against the extracellular loop of the hP2X(3) receptor documented the expression of all mutants in the plasma membrane of HEK293 cells. Eventually, we modelled the possible agonist and antagonist binding sites NBD-1-4 of the hP2X(3) subunit by using structural bioinformatics. This model is in complete agreement with the available data and integrates results from mutagenesis studies with geometry optimization of the tertiary structure predictions of the receptor.