Development of scintillation proximity assays for E-selectin and their application in testing potential antagonists.

E-selectin, a cell adhesion molecule expressed on endothelial cells, is involved in the trafficking of neutrophils to areas of inflammation. The tetrasaccharide sialyl Lewis x (sLe(x)) and other analogues have been shown to be weak affinity antagonists. To study the structure/activity relationship of these weak affinity antagonists, we have developed several scintillation proximity assays (SPAs) for E-selectin. Two of these involve immobilizing E-selectin to streptavidin-coated SPA beads through a biotinylated anti-E-selectin monoclonal antibody. These beads are incubated with 125I-labeled carcinoembryonic antigen (an sLe(x)-containing protein) or 3H-labeled HL-60 cells and the amount of bound ligand is quantitated by counting in a beta-scintillation counter. In addition, we have developed a method to prepare a functionally active biotinylated E-selectin, which can be directly coupled to SPA beads and assayed for ligand binding. These SPAs are sensitive, reproducible, and suitable for screening antagonists and studying structure/activity relationships of lead compounds. By using the SPA, we have also showed that an sLe(x) polyacrylamide polymer is 10 times more potent than the monomer, suggesting that the potency of E-selectin antagonists can be greatly enhanced by multivalent presentation.

Modulation of E-selectin structure/function by metal ions. Studies on limited proteolysis and metal ion regeneration.

E-selectin is a member of the selectin family of proteins that recognize carbohydrate ligands in a Ca(2+)-dependent manner. In order to better understand the role of Ca2+ in E-selectin-ligand interactions, we examined the E-selectin structure by limited proteolysis. Apo-Lec-EGF-CR6, a Ca(2+)-free form of soluble E-selectin containing the entire extracellular domain, was sensitive to limited proteolysis by Glu-C endoproteinase. Amino-terminal sequencing analysis of the proteolytic fragments revealed that the major cleavage site is at Glu98 which is in the loop (residues 94-103) adjacent to the Ca2+ binding region of the lectin domain. Upon Ca2+ binding, Lec-EGF-CR6 was protected from proteolysis. This Ca(2+)-dependent protection was further augmented upon sialyl Lewis x (sLex) ligand binding. These results implied that Ca2+ binding to E-selectin induces a conformational change and perhaps facilitates ligand binding. The sLex-bound complex in turn stabilizes Ca2+ binding. Lec-EGF-CR6 contains only one high-affinity Ca2+ site (Kd = approximately 3.5 microM) as determined by equilibrium dialysis. In addition, we found that Ba2+ was a potent antagonist in blocking Lec-EGF-CR6-mediated HL-60 cell adhesion. By competitive equilibrium dialysis and proteolysis analysis, we demonstrated that Ba2+ bound to apo-Lec-EGF-CR6 5-fold tighter than Ca2+ and abolished ligand binding activity. Sr2+ also bound to apo-Lec-EGF-CR6 tighter than Ca2+. However, Sr(2+)-regenerated Lec-EGF-CR6 showed 50% ligand binding activity. Mg2+ bound to apo-Lec-EGF-CR6 with much weaker affinity than Ca2+ and did not show any activity. Thus, E-selectin function can be modulated by different metal ions.

A ligand binding assay for E-selectin.

We describe here a cell-free ligand binding assay for E-selectin. The assay involves immobilizing soluble E-selection onto microtiter plates and incubating with 125I-labeled carcinoembryonic antigen (CEA) which carries the tetrasaccharide sialyl Lewis x (sLex). The bound CEA is eluted by ethylene glycol bis(beta-aminoethyl ether)-N,N,N',N',-tetraacetic acid and monitored by radioactivity. The binding is abolished by preincubation of either CEA with an anti-sLex antibody or E-selectin with a neutralizing anti-E-selectin antibody, indicating that this is an E-selectin/sLex-specific interaction. The binding is Ca2+ and pH dependent (the optimal pH at 7.0) and also requires sialic acid. Removal of sialic acid from CEA by neuraminidase digestion abrogates the binding. When the protein but not carbohydrate constituent of CEA was denatured by reduction and alkylation, the binding was partially eliminated, indicating that the interaction may also be dependent on either protein conformation or carbohydrate orientation. The assay is simple, sensitive, and comparable to other methods as indicated by an IC50 of 1 mM for sLex. Thus, the assay should be useful for screening antagonists and studying E-selectin structure/function.

Consensus repeat domains of E-selectin enhance ligand binding.

To study the structural characteristics of E-selectin necessary for mediating cell adhesion, we examined the role of the consensus repeat (CR) domains in E-selectin function. Soluble constructs containing different numbers of CR domains were stably expressed in Chinese hamster ovary cells, purified to homogeneity, and characterized. The minimum functional unit of soluble E-selectin consisted of the lectin (Lec) and epidermal growth factor (EGF) domains alone (Lec-EGF) as indicated by its ability to mediate in vitro HL-60 cell adhesion. However, E-selectin containing all six CR domains (Lec-EGF-CR6) at its COOH terminus was the most potent in blocking neutrophil or HL-60 cell adhesion to either immobilized E-selectin or cytokine-stimulated human umbilical vein endothelial cells. This increased potency of Lec-EGF-CR6 in blocking cell adhesion was not due to CR-mediated oligomerization of the protein. Lec-EGF-CR6 was most likely monomeric in solution, as judged by gel filtration fast protein liquid chromatography, membrane ultrafiltration, and chemical cross-linking analysis. Therefore, although the lectin and EGF domains are necessary and sufficient for mediating cell adhesion, the additional six CR domains, present in native E-selectin, contribute to the enhanced binding of E-selectin to its ligand.

Immunochemical detection of adenine nucleotide-binding proteins with antibodies to 5'-p-fluorosulfonylbenzoyladenosine.

5'-p-Fluorosulfonylbenzoyladenosine (FSBA) is a useful reagent for the affinity labeling of adenine nucleotide binding proteins. We have developed an immunochemical approach to the detection of proteins that have been covalently modified with FSBA, which provides an alternative to the use of a radiolabeled ligand. Antibodies have been prepared against FSBA-modified glutamate dehydrogenase and purified by chromatography on ATP-agarose. The resulting affinity-purified antibodies react on Western blots only with proteins that have been labeled previously with the affinity reagent. The degree of immunoreactivity on Western blots correlates well with the extent of covalent modification as shown by studies on the modification and inhibition of the catalytic subunit of cAMP-dependent protein kinase. In crude cellular extracts, numerous proteins can be labeled with FSBA and then detected by using this approach. The labeling and subsequent detection of these proteins can be blocked by including an excess of MgATP, which competes with FSBA for nucleotide-binding sites. The labeling of specific proteins in crude mixtures is saturable, as shown by labeling studies of p56lck, a protein-tyrosine kinase that is abundantly expressed in membranes from the T lymphoma cell line LSTRA.

Detection of protein kinase activity in sodium dodecyl sulfate-polyacrylamide gels.

A procedure is described for identifying protein kinase activity in protein samples following electrophoresis on sodium dodecyl sulfate-polyacrylamide gels. Protein kinase activity is detected by renaturation of the enzymes within the gel followed by phosphorylation with [gamma-32P]ATP of either substrates included in the polyacrylamide gel or of the kinase itself. Then, after removal of the unreacted [gamma-32P]ATP by washing the gel in the presence of an anion-exchange resin, the positions (Mr) of the protein kinase activity are visualized by autoradiography. Studies using a purified catalytic subunit of cAMP-dependent protein kinase indicate that enzyme concentrations as low as 0.01 microgram can easily be detected on gels containing 1 mg/ml casein. The technique is also useful for identifying active subunits of multisubunit enzymes. The active subunit of casein kinase II, for example, can readily be determined by renaturing the dissociated enzyme in gels containing casein. Putative protein kinases present in crude mixtures of proteins can also be detected following separation by gel electrophoresis and can be characterized on the basis of molecular weight and identity of the phosphorylated amino acid. Using this technique, at least three major protein kinases were detected in a mixture of proteins prepared by subfraction of red blood cell membranes.