We have determined the sequences ofthe recombinant DNA inserts of three bacterial plasmid cDNA clones containing most of the rat a-fetoprotein mRNA. The resultant nucleotide sequence of a-fetoprotein was exhaustively compared to the nucleotide sequence of the mRNA encoding rat serum albumin. These two mRNAs have extensive homology (50%) throughout and the same intron locations. The amino acid sequence of rat a-fetoprotein has been deduced from the nucleotide sequence, and its comparison to rat serum albumin's amino acid sequence reveals a 34% homology. The regularly spaced positions of the cysteines found in serum albumin are conserved in rat a-fetoprotein, indicating that these two proteins may have a similar secondary folding structure. These homologies indicate that a-fetoprotein and serum albumin were derived by duplication of a common ancestral gene and constitute a gene family. a-Fetoprotein (AFP) and serum albumin are major plasma proteins both synthesized in mammalian liver parenchymal cells. AFP is also produced in the embryonic yolk sac (1) and is a single chain glycoprotein with a molecular weight of70,000 containing 4% carbohydrate (2-6). AFP synthesis is associated with developmental and neoplastic processes (7-9), and it is the dominant plasma protein of the mammalian fetus. After birth the concentration ofAFP decreases to a trace level in the serum of healthy adult mammals (10, 11). On the other hand, production of serum albumin, a single chain polypeptide with a molecular weight of66,000, increases severalfold after birth to become the predominant protein in the adult serum (7,12,13). This inverse relationship in expression of AFP and serum albumin and the physical similarities between these two proteins suggest that AFP may be the fetal analog of serum albumin.The developmental alterations in the expression of AFP and serum albumin are of interest as an example ofeukaryotic gene regulation (14). Analysis of the sequence and the structural organization ofthese two genes should elucidate the evolutionary relationship between AFP and serum albumin, and may also aid in understanding their regulation. The mRNA nucleotide sequence and the structural organization of the rat serum albumin gene have been established by sequence analysis (15)(16)(17). The amino acid sequence data available for AFP have been compared to those for serum albumin and a degree of internal homology has been shown to exist between these two proteins (18-20). The structural organizations ofthe mouse AFP and serum albumin genes have been estimated from electron micrographs of R-loops, and there is some similarity in this regard between the rat albumin gene and the mouse AFP and albumin genes (21,22). The mouse AFP amino acid sequence has been deduced from the nucleotide sequence of its mRNA. A comparison of this sequence to that of human and bovine albumin revealed a 32% homology and regularly spaced cysteines (23). These homologies suggest that AFP and serum albumin are related, possibly derived by the duplication of a c...