Complementary DNAs coding for two different endothelin (ET) receptors were isolated from a human lung cDNA library. DNA sequencing revealed that they corresponded to either the ET-1-selective receptor subtype originally identified in cows (ETA) or the nonselective subtype first found in rats (ETB). An open reading frame coding for a 427 (ETA) or a 442 (ETB) amino acid-long protein, with a transmembrane topology also found in other G protein-coupled receptors, was determined. Whereas the amino acid identity is about 90% between the characterized members of a given receptor subtype, with the best conserved parts being the hydrophobic transmembrane domains, less than 60% identity was found between the human A and B subtypes. Southern blot analysis of human genomic DNA revealed the existence of multiple hybridizing bands when using probes derived from the full-length ETA or ETB cDNAs. However, when shorter fragments were selected as probes, only one band remained detectable, which suggests the existence of a single-copy gene for each subtype of ET receptor.