Origin of a "bridge" intron in the gene for a two-domain globin

Red cells of the clam Barbatia reeveana express two hemoglobins, one composed of 16- to 17-kDa chains and the other of 35-kDa chains. The nucleotide sequence of the cDNA encoding the 35-kDa chain shows that the polypeptide has two very similar heme-binding domains, which are joined without use of an additional bridging sequence. Two novel introns occur in the gene for the two-domain globin: one, the "precoding" intron, is located two bases 5' from the start codon, and the other, a "bridge" intron, separates the DNA sequences encoding the two domains. Close correspondence exists between the 3' end of the precoding intron and the 3' end of the bridge intron and between parts of the 3' noncoding region of the cDNA for the two-domain globin and the 5' end of the bridge intron. These observations indicate that the bridge intron arose by unequal crossing-over between two identical or very similar genes for a single-domain globin. This conclusion, together with the proposal that exons were initially independent "minigenes" [Gilbert, W. (1987) Cold Spring Harbor Symp. Quant. Biol. 52, 901-905], suggests that many introns may have evolved from the 5' noncoding region of one gene and/or the 3' noncoding region of a second gene. This hypothesis implies that splice junctions would be associated with the original NH2 and COOH termini of proteins and provides an explanation for the observation that splice junctions usually map to protein surfaces. They do so because most NH2- and COOH-terminal residues are usually located on or near the surfaces of proteins.