J. Am. Chem. Soc. doi:10.1021/ja511237n

Carbohydrate-binding proteins such as lectins create specific but weak interactions with their glycan ligands using shallow protein surfaces and hydrophilic interactions, features that complicate inhibitor discovery. Previous research has used synergy between a carbohydrate anchoring group and a pendant ligand to develop inhibitors that retain specificity but display improved binding affinity, but generating compound analogs has generally been slow. Ng et al. now substantially accelerate this process by combining the diversity of a phage-based peptide library with a chemical modification strategy to quickly survey modified carbohydrates as lectin inhibitors. In their approach, the authors began with an N-terminal serine followed by a variable heptameric peptide; periodate oxidation of the serine group enabled attachment of a modified mannose residue on the phage surface. Panning of the resultant 108-member phage library against the lectin target concanavalin A followed by a careful comparison to control experiments led to 86 hits, allowing the authors to define a [WYF]Y[SDEA] consensus motif, which they optimized in a second focused screen to a WYDLF sequence. Biophysical analysis of hits from both rounds confirmed that the monosaccharide, linker and specific peptide sequence were all critical in achieving low-micromolar IC50 values, indicating that these results could not have been achieved by screening a peptide-only library. A crystal structure of the mannose-WYD construct with concanavalin A indicated that the peptide caused conformational rearrangements of the protein surface, opening a hydrophobic site that was not apparent in the structure bound to a three-mannose chain. Finally, testing of a ligand containing a fluorescent probe demonstrated good selectivity, as only 3 of 85 other lectins bound the inhibitor and with lower potency than did concanavalin A. This report offers a new and facile route to creating inhibitors of these challenging protein targets.