Molecular modeling and protein engineering techniques have been used to study residues within G-protein-coupled receptors that are potentially important to ligand binding and selectivity. In this study, Tyr-129 located in transmembrane domain 2 of the human endothelin (ET) type A receptor A (hETA) was targeted on the basis of differences between the hETA and type B receptor (hETB) sequences and the position of the residue on ET receptor models built using the coordinates of bacteriorhodopsin. Replacement of Tyr-129 of hETA by alanine, glutamine, asparagine, histidine, lysine, serine, or phenylalanine results in receptor variants with enhanced ET-3 and sarafotoxin 6C affinities but with unchanged ET-1 and ET-2 affinities. Except for Tyr-129-->Phe hETA, these hETA variants have two to three orders of magnitude lower binding affinity for the ETA-selective antagonist BQ123. Replacement of His-150, the residue in hETB that is analogous in sequence to Tyr-129 of hETA, by either tyrosine or alanine does not affect the affinity of peptide ligands. These results indicate that although transmembrane domain 2 is important in ligand selectivity for hETA, it does not play a significant role in the lack of ligand selectivity shown by hETB. Chimeric receptors have been constructed that further support these conclusions and indicate that at least two hETA regions contribute to ligand selectivity. Additionally, the data support an overlap in the binding site in hETA of agonists ET-3 and sarafotoxin 6C with that of the antagonist BQ123.