Comparative analysis of retroviral and native promoters driving expression of β1,3-galactosyltransferase β3Gal-T5 in human and mouse tissues

β1,3-Galactosyltransferase β3Gal-T5 is highly expressed in the colons of humans and certain primates due to a retroviral long terminal repeat (LTR) acting as a strong promoter. Because this promoter is inactive in other human tissues or mice, we attempted to understand how adoption of a retrotransposon allowed the gene to acquire tissue-specific expression. We identified three novel 5′-UTRs of 3Gal-T5 mRNA, types A, B, and C, and found wide-spread expression of the type A transcript at much lower levels than the LTR transcript, the expression of which is restricted to organs of the gastrointestinal tract. Expression of the type C 5′-UTR transcript was mostly restricted to the ileum, where it was expressed at high levels. We cloned the 5′-flanking regions of both types A and B 5′-UTRs, found deletion constructs functionally active as promoters, and identified CCAAT-binding factor (CBF) and hepatocyte nuclear factor 1 (HNF-1) as the principal nuclear factors controlling the promoters of types A and B 5′-UTR transcripts, respectively. The CCAAT-binding factor binding site and the entire downstream sequence driving the expression of type A transcripts in humans are structurally and functionally conserved in mice, where they constitute a unique β3Gal-T5 promoter that appears to be the ancestral promoter of the gene. The HNF-1 binding motif of the second human promoter is identical to the HNF-1/Cdx binding motif of the LTR promoter but is in the antisense orientation, resulting in much lower binding affinity and promoter strength. These data may explain the successful insertion of the transposon during evolution.