Fullerene‐sp2‐Iminosugar Balls as Multimodal Ligands for Lectins and Glycosidases: A Mechanistic Hypothesis for the Inhibitory Multivalent Effect

Concerted functioning of lectins and carbohydrate‐processing enzymes, mainly glycosidases, is essential in maintaining life. It was commonly assumed that the mechanisms by which each class of protein recognizes their cognate sugar partners are intrinsically different: multivalency is a characteristic feature of carbohydrate–lectin interactions, whereas glycosidases bind to their substrates or substrate‐analogue inhibitors in monovalent form. Recent observations on the glycosidase inhibitory potential of multivalent glycomimetics have questioned this paradigm and led to postulate an inhibitory multivalent effect. Here the mechanisms at the origin of this phenomenon have been investigated. A D ‐gluco‐configured sp²‐iminosugar glycomimetic motif, namely 1‐amino‐5N,6O‐oxomethylydenenojirimycin (1N‐ONJ), behaving, simultaneously, as a ligand of peanut agglutinin (PNA) lectin and as an inhibitor of several glycosidases, has been identified. Both the 1N‐ONJ–lectin‐ and 1N‐ONJ–glycosidase‐recognition processes have been found to be sensitive to multivalency, which has been exploited in the design of a lectin–glycosidase competitive assay to explore the implication of catalytic and non‐glycone sites in enzyme binding. A set of isotropic dodecavalent C₆₀‐fullerene–sp²‐iminosugar balls incorporating matching or mismatching motifs towards several glycosidases (inhitopes) was synthesized for that purpose, thereby preventing differences in binding modes arising from orientational preferences. The data supports that: 1) multivalency allows modulating the affinity and selectivity of a given inhitope towards glycosidases; 2) multivalent presentation can switch on the inhibitory capacity for some inhitope–glycosidase pairs, and 3) interactions of the multivalent inhibitors with non‐glycone sites is critical for glycosidase recognition. The ensemble of results point to a shift in the binding mode on going from monovalent to multivalent systems: in the first case a typical ′′key–lock′′ model involving, essentially, the high‐affinity active site can be assumed, whereas in the second, a lectin‐like behavior implying low‐affinity non‐glycone sites probably operates. The differences in responsiveness to multivalency for different glycosidases can then be rationalized in terms of the structure and accessibility of the corresponding carbohydrate‐binding regions.