Antibodies directed against thyroxine-binding globulin (TBG) have been used to screen a human liver Xgtll expression library. A 1.46-kilobase clone was identified which encodes nearly the complete amino acid sequence, beginning at amino acid 17 of the mature protein. To complete the protein sequence, the cDNA clone was used to identify a genomic clone coding for TBG in a human X chromosome library. The overlapping recombinant clones contained an open reading frame coding for 415 amino acids followed by a polyadenylylation signal (AATAAA) located 275 nucleotides from a TAG termination codon. Beginning at residue 21, the deduced amino acid sequence agrees closely with the known NH2-terminal sequence of the mature peptide. The preceding 20 amino acid residues are hydrophobic in character and presumably represent a leader sequence. Four glycosylation sites were identified, corresponding to the number determined for the purified protein. DNA blot hybridization revealed a single-copy gene, which by chromosomal analysis was found to be located on the long arm of the X chromosome. Unexpectedly, the nucleotide sequence of TBG is closely homologous to those encoding the plasma serine antiproteases a1-antichymotrypsin and a1-antitrypsin. However, there is little overall homology between TBG and transthyretin (prealbumin), the other major thyroxine-binding protein of human plasma.In most vertebrate species, thyroxine-binding globulin (TBG) has been reported to be the major thyroxine-binding protein (1). It is a glycoprotein (molecular mass about 54 kDa) that is synthesized in the liver as a polypeptide of 45 kDa (2). TBG represents a relatively rare translation product of total liver mRNA, accounting for only about 0.018% of total acidprecipitable radioactivity (2). Efforts to obtain detailed information about the structure of TBG and its thyroxinebinding site have been hampered by the presence of multiple electrophoretic forms and the inability to crystallize the protein (3, 4).In this report, we describe the use of serum containing polyclonal antibodies directed against human TBG to isolate cDNA clones coding for TBG from a human liver Xgtll expression library (5). To complete the coding sequence, one of these clones was used to identify a genomic clone from a human X chromosome library. Residues 21-40 of the deduced amino acid sequence correspond closely to the NH2-terminal amino acid sequence of mature TBG reported by Cheng (6). Amino acids 1-20 are very hydrophobic in nature and presumably represent the leader sequence of the propeptide. Unexpectedly, TBG shows a high degree of homology to a1-antichymotrypsin and a1-antitrypsin, and appears to belong to the serpin superfamily, most of which are serine proteinase inhibitors (7,8