The sequence and conformation of human pancreatic procarboxypeptidase A2. cDNA cloning, sequence analysis, and...

A full-length cDNA clone coding for human pancreatic preprocarboxypeptidase A2 has been isolated from a lambda gt 11 human pancreatic library. Expression clones were identified by specific interaction with antisera raised against the native protein. The open reading frame of the polynucleotide sequence is 1254 base pairs in length and encodes a protein of 417 amino acids. This cDNA includes a short leader signal peptide of 16 amino acids and a 94-amino acid-long activation segment. The amino acid sequence shows 89% identity to that of rat procarboxypeptidase A2, the only A2 form sequenced so far, and 64% identity to that of human procarboxypeptidase A1. The newly determined sequence was modeled to the three-dimensional crystal structures of both bovine carboxypeptidase A and porcine procarboxypeptidase A1 by a novel distance geometry approach. Biases in the modeling were avoided by relying exclusively on automatic procedures and by using random structures as starting points. Information taken from the known homologous structures refers only to the backbone since no explicit data describing the conformation of side chains were transferred. Ten structures of human carboxypeptidase A2 were determined on the basis of each of the two known crystal structures. The root-mean-square distance for the backbone atoms between the 10 structures and their mean for 237 selected residues is 0.7 A when starting from the bovine protein and 0.8 A for 251 selected residues when starting from the porcine protein. The 94 residue-long activation segment was also determined in the modeling based on the porcine zymogen; its structure is well defined but not its orientation with respect to the enzyme moiety. The model obtained for human procarboxypeptidase A2 is discussed with respect to the specificity and activation of the enzyme.