Platelet/endothelial cell adhesion molecule-1 (PECAM-1, CD31) is a member of the immunoglobulin superfamily present on platelets, endothelial cells, and leukocytes that may function as a vascular cell adhesion molecule. The purpose of this study was to examine the role of the cytoplasmic domain in PECAM-1 function. To accomplish this, wild-type and mutated forms of PECAM-1 cDNA were transfected into murine fibroblasts and the functional characteristics of the cells analyzed. Wild-type PECAM-1 localized to the cell-cell borders of adjacently transfected cells and mediated heterophilic, calcium-dependent L-cell aggregation that was inhibitable by a polyclonal and two monoclonal anti-PECAM-1 antibodies. A mutant protein lacking the entire cytoplasmic domain did not support aggregation or move to cell-cell borders. In contrast, both forms of PECAM-1 with partially truncated cytoplasmic domains (missing either the COOH-terminal third or two thirds of the cytoplasmic domain) localized to cell-cell borders in 3T3 cells in a manner analogous to the distribution seen in cultured endothelial cells. L-cells expressing these mutants demonstrated homophilic, calcium-independent aggregation that was blocked by the polyclonal anti-PECAM-1 antibody, but not by the two bioactive monoclonal antibodies. Although changes in the cytoplasmic domain of other receptors have been shown to alter ligand-binding affinity, to our knowledge, PECAM-1 is the first example of a cell adhesion molecule where changes in the cytoplasmic domain result in a switch in the basic mechanism of adhesion leading to different ligand-binding specificity. Variations in the cytoplasmic domain could thus be a potential mechanism for regulating PECAM-1 activity in vivo.

Monoclonal antibodies (mAbs) have been produced against the chicken beta 1 subunit that affect integrin functions, including ligand binding, alpha subunit association, and regulation of ligand specificity. Epitope mapping of these antibodies was used to identify regions of the subunit involved in these functions. To accomplish this, we produced mouse/chicken chimeric beta 1 subunits and expressed them in mouse 3T3 cells. These chimeric subunits were fully functional with respect to heterodimer formation, cell surface expression, and cell adhesion. They differed in their ability to react with a panel anti-chicken beta 1 mAbs. Epitopes were identified by a loss of antibody binding upon substitution of regions of the chicken beta 1 subunit by homologous regions of the mouse beta 1 subunit. The identification of the epitope was confirmed by a reciprocal exchange of chicken and mouse beta 1 domains that resulted in the gain of the ability of the mouse subunit to interact with a particular anti-chicken beta 1 mAb. Using this approach, we found that the epitopes for one set of antibodies that block ligand binding mapped toward the amino terminal region of the beta 1 subunit. This region is homologous to a portion of the ligand-binding domain of the beta 3 subunit. In addition, a second set of antibodies that either block ligand binding, alter ligand specificity, or induce alpha/beta subunit dissociation mapped to the cysteine rich repeats near the transmembrane domain of the molecule. These data are consistent with a model in which a portion of beta 1 ligand binding domain rests within the amino terminal 200 amino acids and a regulatory domain, that affects ligand binding through secondary changes in the structure of the molecule resides in a region of the subunit, possibly including the cysteine-rich repeats, nearer the transmembrane domain. The data also suggest the possibility that the alpha subunit may exert an influence on ligand specificity by interacting with this regulatory domain of the beta 1 subunit.