Chicken c-erbA (Ck-c-erbA) cDNA (1250 base pairs), a cellular homologue of the avian erythroblastosis virus v-erbA gene, encodes a 408 amino acid protein which binds L-T3 and its analogs with affinities similar to that of endogenous thyroid hormone nuclear receptors. By analogy with steroid receptors, Ck-c-erbA(Met1-Val408) contains an A and B domain (amino acids 1-50); a putative DNA binding C domain (amino acids 51-118); a hydrophilic D domain (amino acids 119-189); and a putative ligand binding E domain (amino acids 187-408). To further characterize the ligand binding region of Ck-c-erbA, two deletion mutations were constructed: Ck-c-erbA(Met120-Val408) which encodes a 289 amino acid protein lacking regions A, B, and C; and Ck-c-erbA-(Met199-Val408) which encodes a 210 amino acid protein lacking regions A, B, C, D, and the first 12 amino acids of the E region. The in vitro translation products ([35S]methionine) of cDNA transcripts of a human placental c-erbA, Ck-c-erbA (Met1-Val408), and Ck-c-erbA(Met120-Val408) efficiently bind L-[125I]T3, whereas Ck-c-erbA(Met199-Val408) does not bind L-[125I]T3. In frame substitution of the last 14 C-terminal amino acids of Ck-c-erbA(Met1-Val408) for the last 7 C-terminal amino acids of v-erbA reduces but does not eliminate L[125I]T3 binding. These results indicate that a broad region of the E domain is important for ligand binding.(ABSTRACT TRUNCATED AT 250 WORDS)