Leguminous lectins have a conserved carbohydrate recognition site comprising four loops (A–D). Here, we randomly mutated the sequence and length of loops C and D of peanut agglutinin (PNA) and expressed the proteins on the surface of mouse green fluorescent protein (GFP)-reporter cells. Flow cytometry, limiting dilution, and cDNA cloning were used to screen for several mutated PNAs with distinct properties. The mutated PNA clones obtained using NeuAcα2-6(Galβ1-3)GalNAc as a ligand showed preference for NeuAcα2-6(Galβ1-3)GalNAc rather than non-sialylated Galβ1-3GlcNAc, whereas wild-type PNA binds to Galβ1-3GlcNAc but not sialylated Galβ1-3GalNAc. Sequence analyses revealed that for all of the glycan-reactive mutated PNA clones, (i) loop C was eight amino acids in length, (ii) loop D was identical to that of wild-type PNA, (iii) residue 127 was asparagine, (iv) residue 125 was tryptophan, and (v) residue 130 was hydrophobic tyrosine, phenylalanine, or histidine. The sugar-binding ability of wild-type PNA was increased nine-fold when Tyr125 was mutated to tryptophan, and that of mutated clone C was increased more than 30-fold after His130 was changed to tyrosine. These results provide an insight into the relationship between the amino acid sequences of the carbohydrate recognition site and sugar-binding abilities of leguminous lectins.
