Characterization and modelling of CEACAM1 interactions in cell signalling

Abstract

CEACAM1, the primordial carcino-embryonic Ag gene family member, is a transmembrane cell adhesion molecule expressed in leukocytes, epithelia, and blood vessel endothelia. As a result of differential splicing, CEACAM1 occurs as several isoforms, the two major ones being CEACAM1-L and -S, that have long (L) or short (S) cytoplasmic domains, respectively. The L:S expression ratios vary in different cells and tissues. In addition to CEACAM1, human but not rodent cells express GPI-linked CEACAM members (CEACAM5-CEACAM8).

We compared the expression patterns of CEACAM1-L, CEACAM1-S, CEACAM6, and CEACAM8 in purified populations of neutrophilic granulocytes, B lymphocytes, and T lymphocytes from rats, mice, and humans. Human granulocytes expressed CEACAM1, CEACAM6, and CEACAM8, whereas human B lymphocytes and T lymphocytes expressed only CEACAM1 and CEACAM6. Whereas granulocytes, B cells, and T cells from mice and rats expressed both CEACAM1-L and CEACAM1-S in ratios of 2.2 2.9:1, CEACAM1-S expression was totally lacking in human granulocytes, B cells, and T cells. This suggests that the GPI-linked CEACAM members have functionally replaced CEACAM1-S in human leukocytes. Support for this hypothesis was obtained from experiments in which the extracellular signal-regulated kinases Erk 1 /2 were activated by anti-CEACAM Abs.

We demonstrated that CEACAM1 and CEACAM8 are physically associated in human granulocytes. The CEACAM1/CEACAM8 complex in human cells might accordingly play a similar role as CEACAM1-L/CEACAM1-S dimers in rat cells. CEACAM1-L has been found in large molecular weight forms suggesting posttranslational covalent modification. We investigated the possibility that the cytoplasmic domain of CEACAM1-L can act as a transglutaminase substrate. Glutathione S-transferase fusion proteins of the cytoplasmic domains of rat and mouse CEACAM1-L as well as free cytoplasmic domains were converted into covalent dimers by tissue transglutaminase. Thus, the cytoplasmic domains of rat and mouse CEACAM-L are substrates for tissue transglutaminase, lending support to the notion that higher molecular weight forms of CEACAM-L are formed by transglutaminase modification.

We report a novel development of the approach to determining active concentrations based on surface plasmon resonance (SPR) technology. The method relies on changes in binding rates with varying flow rates under conditions of partial mass transport, and does not require standards of known concentrations. We introduce an analytical solution to the differential equations describing the formation of a 1: 1 bimolecular complex, taking into account both the association and dissociation reactions, under partial mass transport limitations. The accuracy, precision, and sensitivity of this approach were determined in experiments involving binding of tyrosine-phosphorylated recombinant proteins to anti-phosphotyrosine antibodies, where the active concentration could be determined independently by in vitro phosphorylation with 33P. There was an excellent agreement between the active concentrations determined by the analytical SPR-based method and by determination of the level of radioactivity of the phosphorylated protein.

We have applied surface plasmon resonance-based techniques to investigate the key patterns, kinetics and thermodynamics of the binding interaction of SHP-1 with CEACAM1-L. This required development of new reaction and curve fitting algorithms. Both the NSH2 and C-SH2 binding sites of SHP- I were shown to participate in the interaction. Furthermore, we discovered a novel phosphotyrosine binding site in the C-SH2 domain, which differed kinetically from the classical C-SH2 site. The pY488 motif was found to interact with all three binding sites in N,C-(SH2)2. The pY515 motif interacted with the two different C-SH2 sites, but not with the N-SH2 site, The interaction of N,C-(SH2)2 with a diphosphorylated CEACAM1-L (pY488 + pY515) ligand resulted in the formation of 7 dynamic complexes, all with a stoichiometry of 1: 1. Of these, 5 forms interacted via single docking, and 2 forms via double-docking utilizing both of the ITIMlike sites. This investigation provides means for continuing detailed interaction analyses of intact CEACAM1-L with intracellular signal molecules. In general, the results point to broad variations in interaction kinetics of SH2 domains, which should be considered when evaluating signal transduction mechanisms.