In a search for genes encoding the serine peptidases prostasin and testisin, which are expressed mainly in prostate and testis, respectively, we identified a related, novel gene. Sequencing of cDNA allowed us to deduce the full amino acid sequence of the human gene product, which we term "pancreasin" because it is transcribed strongly in the pancreas. The idiosyncratic 6-exon organization of the gene is shared by a small group of tryptic proteases, including prostasin, testisin, and gamma-tryptase. Like the other genes, the pancreasin gene resides on chromosome 16p. Pancreasin cDNA predicts a 290-residue, N-glycosylated, serine peptidase with a typical signal peptide, a 12-residue activation peptide cleaved by tryptic hydrolysis, and a 256-amino acid catalytic domain. Unlike prostasin and other close relatives, human pancreasin and a nearly identical chimpanzee homologue lack a carboxyl-terminal membrane anchor, although this is present in 328-residue mouse pancreasin, the cDNA of which we also cloned and sequenced. In marked contrast to prostasin, which is 43% identical in the catalytic domain, human pancreasin is transcribed strongly in pancreas (and in the pancreatic ductal adenocarcinoma line, HPAC) but weakly or not at all in kidney and prostate. Antibodies raised against pancreasin detect cytoplasmic expression in HPAC cells. Recombinant, epitope-tagged pancreasin expressed in Chinese hamster ovary cells is glycosylated and secreted as an active tryptic peptidase. Pancreasin's preferences for hydrolysis of extended peptide substrates feature a strong preference for P1 Arg and differ from those of trypsin. Pancreasin is inhibited by benzamidine and leupeptin but resists several classic inhibitors of trypsin. Thus, pancreasin is a secreted, tryptic serine protease of the pancreas with novel physical and enzymatic properties. These studies provide a rationale for exploring the natural targets and roles of this enzyme.