Structure of the canine pancreatic colipase gene includes two protein-binding sites in the promoter region.

Characterization of a lambda phage genomic clone, CL5A, which encodes the canine pancreatic colipase gene, revealed the primary structure of 987 nucleotides (nt) of 5'-flanking sequence, 2066 nt defining the primary transcriptional unit, which is organized into three exon sequences, and 130 nt of 3'-flanking sequence. Exon 1 encodes the amino-terminal signal peptide, the propeptide (Val1-Pro-Asp-Pro-Arg), and the hydrophobic lipid-binding region (Gly6-Ile-Ile-Ile) at the amino terminus of the mature coenzyme. Exon 2 encodes carboxylate residues (Glu12 and Glu15) likely to be involved in binding of pancreatic lipase to colipase at the aqueous-lipid interface. Exon 3 encodes the hydrophobic sequence (Leu54-Tyr-Gly-Tyr-Tyr) that is essential for binding the central tightly structured disulfide-bonded region of the coenzyme to lipid. Southern blot analysis was consistent with the presence of a single-copy colipase gene and a potential colipase gene homologue. Among 16 tissues examined by Northern blot analysis, colipase expression was detected only in pancreas. Proteins contained in nuclear extracts prepared from dog pancreas conferred two regions of DNase I protection coincident for both coding and noncoding strands (positions -62 to -44 (CL-I site) and -128 to -106 (CL-II site) in the coding strand). Competition gel mobility shift experiments indicated that protein-DNA interactions that occur at colipase sites I and II are sequence- and protein-specific and unrelated to the PAN-binding sequence described in the 5'-enhancer region of the rat chymotrypsin B gene (Nelson, C., Shen, L.-P., Meister, A., Fodor, E., and Rutter, W. J. (1990) Genes & Dev. 4, 1035-1043). Nuclear extracts from pancreas and brain, but not liver, contain similar CL-I- and CL-II-binding proteins. CL-I and CL-II represent protein-binding elements that may participate as additional promoter regions in regulated expression of the colipase gene. CL-I contains a central homopolymeric d(G) sequence. CL-II shows a GC-rich region on the noncoding strand (5' GGGGGCGTGT 3') that is similar (8/9 match) to the Sp1-binding sequence.