Identification of key amino acid residues determining ligand binding specificity, homodimerization and cellular distribution of human galectin-10.

Charcot-Leyden crystal protein/Gal-10, abundantly expressed in eosinophils and basophils, is related to several immune diseases. Recently, crystallographic and biochemical studies showed that Gal-10 cannot bind lactose, because a glutamate residue (Glu33) from another monomer blocks the binding site. Moreover, Gal-10 actually forms a novel dimeric structure compared to other galectins. To investigate the role that Glu33 plays in inhibiting lactose binding, we mutated this residue to glutamine, aspartate, and alanine. The structure of E33A shows that Gal-10 can now bind lactose. In the hemagglutination assay, lactose could inhibit E33A from inducing chicken erythrocyte agglutination. Furthermore, we identified a tryptophan residue (Trp127) at the interface of homodimer that is crucial for Gal-10 dimerization. The variant W127A, which exists as a monomer, exhibited higher hemagglutination activity than wild type Gal-10. The solid phase assay also showed that W127A could bind to lactose-modified sepharose-6B, whereas wild type Gal-10 could not. This indicates that the open carbohydrate-binding site of the W127A monomer can bind to lactose. In addition, the distribution of EGFP-tagged Gal-10 and its variants in HeLa cells was investigated. Because Trp72 is the highly conserved in the ligand binding sites of galectins, we used EGFP-tagged W72A to show that Gal-10 could not be transported into the nucleus, indicating that Trp72 is crucial for Gal-10 transport into that organelle.