Determination of disulfide bond assignments and N-glycosylation sites of the human gastrointestinal carcinoma antigen GA733-2 (CO17-1A, EGP, KS1-4, KSA, and Ep-CAM)

The GA733-2 antigen is a cell surface glycoprotein highly expressed on most human gastrointestinal carcinoma and at a lower level on most normal epithelia, It is an unusual cell-cell adhesion protein that does not exhibit any obvious relationship to the four known classes of adhesion molecules. In this study, the disulfide-bonding pattern of the GA733-2 antigen was determined using matrix-assisted laser desorption/ionization mass spectrometry and N-terminal sequencing of purified tryptic peptides treated with 2-[2 '-nitrophenylsulfonyl]-3-methyl-3-bromoindolenine or partially reduced and alkylated, Numbering GA733-2 cysteines sequentially from the N terminus, the first three disulfide linkages are Cys(1)-Cys(4), Cys(2)-Cys(6) and Cys(3)-Cys(5), which is a novel pattern for a cysteine-rich domain instead of the expected epidermal growth factor-like disulfide structure. The next three disulfide linkages are Cys(7)-Cys(8), Cys(9)-Cys(10), and Cys(11)-Cys(12), consistent with the recently determined disulfide pattern of the thyroglobulin type 1A domain of insulin-like growth factor-binding proteins 1 and 6, Analysis of glycosylation sites showed that GA733-2 antigen contained N-linked carbohydrate but that no O-linked carbohydrate groups were detected. Of the three potential N-linked glycosylation sites, Asn(175) was not glycosylated, whereas Asn(88) was completely glycosylated, and Asn(51) was partially glycosylated, These data show that the extracellular domain of the GA733-2 antigen consists of three distinct domains; a novel cysteine-rich N-terminal domain (GA733 type 1 motif), a cysteine-rich thyroglobulin type 1A domain (GA733 type 2 motif), and a unique nonglycosylated domain without cysteines (GA733 type 3 motif).