Human alpha-3-N-acetylgalactosaminyltransferase has been purified 27,000,000-fold from A1 plasma by (NH4)2SO4 fractionation and affinity chromatography on Sepharose 4B, anti-human group O plasma antibodies-Sepharose 4B, and Blue Dextran-Sephadex G-25. A modified procedure in the Sepharose 4B step was developed by batch adsorption and desorption experiments. Cibacron Blue F3G-A, the chromophore of Blue Dextran, was found to bind to the enzyme. UDP is an effective inhibitor of this binding. The pure transferase has an apparent molecular weight of 35,000 as judged by SDS-PAGE in the presence of a reducing agent. The specific activity is 16 pmol/min.ng enzyme, which is comparable to that (30 pmol/min.ng enzyme) of alpha-3-N-acetylgalactosaminyltransferase from porcine submaxillary glands [Schwyzer and Hill (1977) J. Biol. Chem. 252, 2338-2355]. The apparent Km values for UDP-GalNAc, 2'-fucosyllactose, and lacto-N-fucopentaose I are 13, 270, and 350 microM, respectively. The reaction velocity was found to fall off again at high concentrations of oligosaccharide acceptor substrates. The apparent Ki values for UDP and UDP-galactose are 8.6 and 6.2 microM, respectively. The pure enzyme also catalyzes the transfer of galactose in alpha-linkage to 2'-fucosyllactose though the transfer rate of galactose is much lower than that of N-acetylgalactosamine.

Normal human plasma was found to contain beta 1-4N-acetylgalactosaminyltransferase catalyzing the transfer of N-acetylgalactosamine from UDP-GalNAc to 3'-sialyl-lactose, NeuAc alpha 2-3Gal beta 1-4Glc. The transferred N-acetylgalactosaminyl residue was cleaved from the desialylated reaction product by the beta-N-acetylhexosaminidase from jack beans. Methylation and hydrolysis of the desialylated reaction product yielded only 2,3,6-tri-O-methylgalactose and 2,3,6-tri-O-methylglucose as neutral sugars, indicating that the N-acetylgalactosaminyl residue was introduced at position C-4 of the galactosyl residue of 3'-sialyllactose. The enzyme required Mn2+ ions for its activity and showed a pH optimum between 6.5 and 8.5. By using a wide variety of oligosaccharides and glycoconjugates, the acceptor specificity of the beta 1-4N-acetylgalactosaminyltransferase was investigated. No detectable amount of N-acetylgalactosamine was transferred to either 6'-sialyllactose or lactose. The enzyme did not act on ganglioside GM3, NeuAc alpha 2-3Gal beta 1-4Glc-ceramide, suggesting that the hydrophobic ceramide portion of GM3 interferes with the enzyme reaction. On the other hand, glycoproteins carrying terminal NeuAc alpha 2-3Gal beta 1-4GlcNAc structures on their N-linked oligosaccharide chains, e.g. Tamm-Horsfall glycoprotein, were efficient acceptors.

No abstract

We measured UDP-GlcNAc:Gal (beta 1-4) Glc (or GlcNAc) (beta 1-3) N-acetylglucosaminyltransferase activities in bovine (Holstein and Jersey cow) and human colostrums, and found in human colostrums sufficient activity to study the enzyme properties while not in bovine colostrums. The properties (requirements, pH optimum, acceptor specificity and Km values for lactose and N-acetyllactosamine) of the enzyme from human colostrum were very similar to those from human serum and urine. The reaction product was hydrolyzed by beta-N-acetylhexosaminidase, indicating that the N-acetylglucosaminyl residue was beta-linked to lactose. Methylation and hydrolysis of the reaction product from lactose [3H] labeled at the terminal galactose yielded 2, 4, 6-tri-O-methyl [3H] galactose. Thus the structure of the product was demonstrated to be GlcNAc (beta 1-3) Gal (beta 1-4) Glc (lacto-N-triose II). On the other hand, bovine sera contained N-acetylglucosaminyltransferase catalyzing the transfer of N-acetylglucosamine from UDP-GlcNAc to lactose. The enzyme activities were approximately 1/6-1/4 of that contained in human serum. The presence of (beta 1-3) N-acetylglucosaminyltransferase in human colostrum and its absence in bovine colostrums, apparently corresponds with the presence and absence of oligosaccharides containing lacto-N-triose II structure in colostrum.

Various oligosaccharides containing galactose(s) and one glucosamine (or N-acetylglucosamine) residues with β1-4, α1-6 and β1-6 glycosidic bond were synthesized; Galβ1-4GlcNH 2 , Galα1-6GlcNH 2 , Galα1-6GlcNAc, Galβ1-6GlcNH 2 , Galβ1-4Galβ1-4GlcNH 2 and Galβ1-4Galβ1-4GlcNAc. Galα1-6GlcNH 2 (MelNH 2 ) and glucosamine (GlcNH 2 ) had a suppressive effect on the proliferation of K562 cells, but none of the other saccharides tested containing GlcNAc showed this effect. On the other hand, the proliferation of the human normal umbilical cord fibroblast was suppressed by none of the saccharides other than GlcNH 2 . Adding Galα1-6GlcNH 2 or glucosamine to the culture of K562 cell, the cell number decreased strikingly after 72 h. Staining the remaining cells with Cellstain Hoechst 33258, chromatin aggregation was found in many cells, indicating the occurrence of cell death. Furthermore, all of the cells were stained with Galα1-6GlcNH-FITC (MelNH-FITC). Neither the control cells nor the cells incubated with glucosamine were stained. On the other hand, when GlcNH-FITC was also added to cell cultures, some of them incubated with Galα1-6GlcNH 2 were stained. The difference in the stainability of the K562 cells by Galα1-6GlcNH-FITC and GlcNH-FITC suggests that the intake of Galα1-6GlcNH 2 and the cell death induced by this saccharide is not same as those of glucosamine. The isolation of the Galα1-6GlcNH 2 binding protein was performed by affinity chromatography (melibiose-agarose) and LC-MS/MS, and we identified the human heterogeneous ribonucleoprotein (hnRNP) A1 (34.3 kDa) isoform protein (30.8 kDa). The hnRNP A1 protein was also detected from the eluate(s) of the MelNH-agarose column by the immunological method (antihnRNP-A1 and HRP-labeled anti-mouse IgG (γ) antibodies).