Characterization of the full length cDNA for murine beta-1,4-galactosyltransferase. Novel features at the 5'-end predict two translational start sites at two in-frame AUGs.

We have isolated overlapping cDNA clones representing the full length (4038 base pairs) transcript for murine beta-1,4-galactosyltransferase. The coding sequence predicts a membrane-bound glycoprotein with three distinct structural features, a large COOH-terminal domain (355 amino acids), a single transmembrane domain (20 amino acids), and a short NH2-terminal domain. Primer extension analysis, S1 protection analysis, and RNA blotting demonstrate the presence of two sets of mRNA transcripts which differ in length by about 200 base pairs. The 5' boundary of the long transcripts maps upstream of two in-frame ATGs. The 5' boundary of the short transcripts maps between these two ATGs. These results predict that two related forms of beta-1,4-galactosyltransferase of 399 and 386 amino acids are synthesized as a consequence of alternative translation initiation. The long form (399 amino acids) has an NH2-terminal extension of 13 amino acids.