Immunochemical studies on the reactivities and combining sites of the d-galactopyranose- and 2-acetamido-2-deoxy-d-galactopyranose-specific lectin purified from Sophora japonica seeds

Sophora japonica lectin agglutinates human B erythrocytes strongly and A₁ erythrocytes weakly. Bivalent metal ions such as Ca²⁺, Mn²⁺, or Mg²⁺ were shown to be essential for hemagglutinating and precipitating activities. At optimal concentrations of bivalent metal ions, hemagglutinating activity was highest between pH 8.5 and 9.0 and decreased sharply below pH 8.5, whereas precipitating capacity was maximal between pH 6.7 and 9.5. The combining site of the S. japonica lectin was explored by quantitative precipitin and precipitin inhibition assays. This lectin showed substantial differences in precipitation with several blood group B substances ascribable to heterogeneity resulting from incomplete biosynthesis of their carbohydrate side chains. The lectin precipitated moderately well with A₁ substance and precursor blood group I fractions (OG). It precipitated weakly or not at all with A₂, H, or Le^a substances. In inhibition assays, glycosides of dGalNAc were about five to six times better than those of dGal; dGalNAc itself was about six times better than dGal. Nitrophenyl glycosides were all substantially better than the methyl glycosides, indicating a hydrophobic contribution to the site subterminal to the nonreducing moiety. Although nitrophenyl β-glycosides were much better than the corresponding α-glycosides, the methyl α-and βDGalNAcp were equal in activity as were methyl α- and βDGalp. Among the oligosaccharides tested, the β-linked N-tosyl-l-serine glycoside of dGalβ1 → 3dGalNAc was best and was as active as p-nitrophenyl βDGalNAcp, whereas dGalβ1 → 3dGalNAc α-N-tosyl serine and the nitrophenyl and phenyl α-glycosides of dGalβ1 → 3dGalNAc were much less active, suggesting that the hydrophobic moiety and/or a subterminal dGalNAc β-linked and substituted on carbon 3 play an important role in binding and that a β-linked glycoside of dGalβ1 → 3dGalNAc may be an essential requirement for binding. The results of inhibition studies with other oligosaccharides indicate that a subterminal dGlcNAc substituted on carbon 3 or 4 by dGalβ may contribute somewhat to binding and that whether the dGlcNAc is linked β1 → 3 or β1 → 6 to a third sugar does not contribute to or interfere with binding. The β1 → 3 linkage of the terminal dGal to the subterminal amino sugar is significant since dGalβ1 → 4dGlcNAc was one-half as active as the corresponding β1 → 3-linked compound and the subterminal sugar must be unsubstituted for optimal binding. N-Acetyllactosamine was 50% more active than lactose, indicating that the subterminal N-acetamido group was also contributing significantly to binding. A variety of other sugars, glycosides, and oligosaccharides showed little or not activity. From the oligosaccharides available, the combining size of this lectin would appear to be least as large a β-linked disaccharide and most complementary to dGalβ1 → 3dGalNAc β-linked to tosyl-l-serine the most active compound tested.