Low molecular weight divalent and trivalent ligands have also been generated to probe ConA ligand recognition. As had been shown with the hepatic lectin (see Section V.C) [96,97] and influenza virus hemagglutinin (see Sections VI.E-VI.F) , low molecular weight ligands can achieve high potencies, displaying inhibitory potencies 100- to 1000-fold higher than those of the corresponding monovalent saccharides. To investigate the effect of epitope spacing and orientation in ConA binding, Roy and coworkers constructed a series of low molecular weight bivalent ligands in which the saccharide epitopes were appended through linkers differing in length and composition (Fig. 15) [89,90]. The ability of these bivalent molecules to inhibit ConA binding to immobilized mannan (a yeast glycoprotein) was tested in an enzyme-linked lectin assay (ELLA) (Fig. 8). The structure of the linking group had subtle effects on the ConA inhibitory properties of the bivalent molecules. The bivalent ligand with a six-carbon linker, 17c, found to be superior in this study, was 5.3 times more effective per saccharide at inhibition than a-^-nitrophenyl-manno-pyranoside . Roy and coworkers also examined the effects of higher order valency in a related series of related ligands (Fig. 16) . Bi-, tri-, and tetravalent ligands, 18, 19, and 20, were 138, 210, and 231 times more active than a-Me-mannose, respectively, when analyzed on a molar scale. However this translates to 69-, 70-, and 57.8-fold increases in potency when determination is made on a per-saccharide-residue basis. On a multivalent display, additional ligands that do not contribute productively to binding can detract from the observed potency . This has also been observed in the case of polyacrylamide displays targeting influenza virus, where maximal activity is observed at a low ligand density (see Sections VI.A-VI.B) [99-101].
The abilities of small bi-, tri-, and tetravalent molecules to inhibit ConA likely arise from their propensities to cross-link lectin multimers. Molecular modeling of Roy's ligands suggests that the spacers linking the saccharide recognition epitopes
are not long enough to span the 65 A required to simultaneously occupy two binding sites within a single ConA tetramer (D. Mann, unpublished results, 1999). Another potential mode of action for these ligands is suggested by Brewer's finding that many oligosaccharides and glycopeptides can form cross-linked lectin complexes . A turbidimetric assay monitoring the formation of cross-linked ConA precipitates supports this model for Roy's ligands (Fig. 7). Thus, the best small-molecule inhibitors of ConA may be those that can cross-link the lectin. Dendrimers that exhibit enhanced inhibitory activity toward ConA also appear to function by this mechanism [87,88]. Regardless of their mode of action, data generated with the bivalent ligands highlight the necessity for proper spacing between ligands to elicit optimal multi-valent interactions.
Was this article helpful?