Selection of galectin‐3 ligands derived from genetically encoded glycopeptide libraries

Abstract

In this article, we used genetically encoded fragment‐based discovery (GE‐FBD) approach to identify glycopeptides that bind to the carbohydrate recognition domain of the human galectin‐3 (G3C). We generated 6 chemically identical phage libraries Ser‐[X]4‐Gly‐Gly‐Gly, built on variable combinations of redundant Ser and Gly codons. Oxime ligation of hydroxylamine derivatives of galactose (Gal), glucose (Glu), mannose (Man), rhamnose (Rha), and xylose (Xyl) produced a glycopeptide library in which both peptide and glycan can be decoded via DNA sequencing. Screening of this library against G3C identified 1062 combinations of monosaccharides and peptides that exhibited a significant (P\u2009<\u2009.05) enrichment on G3C and not control selections. Glycopeptides Gal‐WKPE, Gal‐WHVP, and Gal‐LSMA displayed on phage exhibited up to 63‐fold increase in binding potency to G3C when compared to phage displaying random glycopeptide or nonglycosylated SWKPE, SWHVP, and SLSMA. This work mapped the boundary conditions of the GE‐FBD approach with respect to the affinity of individual fragments. We observed that fragments with no detectable affinity (Glu, Xyl, and Rha) diverted the selection toward ligands that bind to G3C equally well with or without the glycan. Weak fragments (Gal, 10 mM) could effectively steer the selection toward G3C ligands in which glycan and peptide bind synergistically.