Sulfated N-linked oligosaccharides in mammalian cells. III. Characterization of a pancreatic carcinoma cell surface glycoprotein with N- and O-sulfate esters on asparagine-linked glycans

In the preceding two papers, we described two new classes of sulfated N-linked oligosaccharides isolated from total cellular 35SO4-labeled macromolecules of different mammalian cell lines. The first class carries various combinations of sialic acids and 6-O-sulfate esters on typical complex-type chains, while the second carries heparin and heparan-like sequences. In this study, we have characterized a sulfophosphoglycoprotein of 140 kDa from FG-Met-2 pancreatic cancer cells whose oligosaccharides share some properties of both these classes. The molecule was localized to the cell surface by electron microscopy using a monoclonal antibody (S3-53) and by cell surface 125I-labeling. Metabolic labeling of the cells with radioactive glucosamine, methionine, inorganic sulfate, or phosphate all demonstrated a single 140-kDa molecule. Pulse-chase analysis and tunicamycin treatment indicated the glycosylation of a putative primary translation product of 110 kDa via an intermediate (120 kDa) to the mature form (140 kDa). Digestion with peptide:N-glycosidase F (PNGaseF) indicated a minimum of four N-linked glycosylation sites. PNGaseF released more than 90% of the [6-3H]GlcNH2 label and 40-70% of 35SO4 label from the immunoprecipitated 140-kDa molecule. The isolated oligosaccharides were characterized as described in the preceding two papers. The majority of [6-3H]GlcNH2-labeled molecules were susceptible to neuraminidase. More than 50% of the 35SO4 label was associated with only 5-10% of the 3H-labeled chains. Some of the sulfated chains were partly sialylated molecules with four to five negative charges. Treatment with nitrous acid released about 25% of the 35SO4 label as free sulfate, together with 6% of the [6-3H]GlcNH2 label, indicating the presence of N-sulfated glucosamine residues. Some of these oligosaccharides were degraded by heparinase and heparitinase. Therefore, while they are not as highly charged as typical heparin or heparan chains, they must share structural features that permit recognition by the enzymes. Thus, this 140-kDa glycoprotein contains at least four asparagine-linked chains substituted with a heterogeneous mixture of sulfated sequences. The heterogeneity of these molecules is as extensive as that described for whole-cell sulfated N-linked oligosaccharides in the preceding two papers.     

Sulfated N-linked oligosaccharides in mammalian cells. I. Complex-type chains with sialic acids and O-sulfate esters

The structures of sulfated N-linked oligosaccharides have been reported for a few specific proteins. We recently demonstrated that such oligosaccharides occur in many different types of tissue culture cell lines (Freeze, H. H., and Varki, A. (1986) Biochem. Biophys. Res. Commun. 140, 967-973). Here we report improved methods to metabolically label cell lines with 35SO4 and to release sulfated N-linked oligosaccharides with peptide:N-glycosidase F as well as the partial structure of some of these novel oligosaccharides. The released 35SO4-labeled chains from Chinese hamster ovary (CHO) cells and bovine pulmonary artery endothelial cells (CPAE) were characterized by gel filtration, anion exchange and lectin affinity chromatography, and various enzymatic and chemical treatments. Each cell line contains a class of sulfated oligosaccharide chains bearing from two to six negative charges in varying combinations of O-sulfate esters and sialic acids. These molecules represent a significant proportion of both the total 35SO4 label and the total anionic N-linked oligosaccharides. They are also relatively enriched in a CHO mutant that is deficient in glycosaminoglycan chain synthesis. Lectin affinity chromatography of such molecules from CPAE cells indicates that the majority are sialylated multiantennary complex-type chains. The sulfate esters are exclusively of the primary type. Sequential exoglycosidase digestions, including beta-hexosaminidase A treatment at low pH, demonstrate that at least one-third of these sulfate esters are found in the following structure, (formula; see text) where R is the remainder of the underlying oligosaccharide, and SA is sialic acid. In addition to these molecules, a more highly charged group of sulfated N-linked oligosaccharides sharing structural features with glycosaminoglycans was found in CPAE cells, but not in CHO cells. These are described in the following paper (Sundblad, G., Holojda, S., Roux, L., Varki, A., and Freeze, H. H. (1988) J. Biol. Chem. 263, 8890-8896).     

A neuraminidase from Streptococcus sanguis that can release O-acetylated sialic acids

The naturally occurring sialic acids can have different types of N- and O-substitutions, resulting in more than 20 known isomers and compounds. Most methods for the detailed study of these various sialic acids require that the molecules be first released from their alpha-glycosidic linkage. When mild acid hydrolysis is used for this purpose, significant destruction of O-substituent groups occur. On the other hand, the presence of O-substituent groups renders the sialic acid molecule partially or completely resistant to the action of the currently known neuraminidase. To circumvent this problem, we searched for a neuraminidase whose activity is not affected by O-substitution. We reasoned that because Streptococcus sanguis from the human oral cavity is continually exposed to O-substituted sialic acids, its extracellular neuraminidase might not be blocked by O-substitution. We therefore purified this enzyme 3100-fold (56% yield) using ammonium sulfate precipitation, N-(p-aminophenyl)oxamic acid-agarose affinity chromatography, and chromatography on quaternary aminoethyl (QAE)-Sephadex, sulfopropyl (SP)-Sephadex, and Sephacryl S-200. The purified preparation is free of other significant glycosidase activities and proteolytic activities. It is capable of quantitatively releasing all the O-acetylated sialic acids that we studied with the single exception of the 4-O-acetylated sialic acid of equine submaxillary mucin. The activity of the enzyme is also not restricted by the type pf sialic acid linkage or the nature of the underlying oligosaccharide. However, it has maximal activity on gangliosides only in the presence of detergents. The general properties of this enzyme are described and its substrate specificities are contrasted with those of the commonly used neuraminidase from Vibrio cholerae.     

O-acetylation and de-O-acetylation of sialic acids. O-acetylation of sialic acids in the rat liver Golgi apparatus involves an acetyl intermediate and essential histidine and lysine residues--a transmembrane reaction?

Isolated intact rat liver Golgi vesicles utilize [acetyl-3H]coenzyme A to add 3H-O-acetyl esters to sialic acids of internally facing endogenous glycoproteins. During this reaction, [3H]acetate also accumulates in the vesicles, even though the vesicles are impermeant to free acetate. On the other hand, entry of intact AcCoA into the lumen of the vesicles could not be demonstrated, and permeabilization of the vesicles did not alter the reaction substantially (Diaz, S., Higa, H. H., Hayes, B. K., and Varki, A. (1989) J. Biol. Chem. 264, 19416-19426). When vesicles prelabeled with [acetyl-3H] coenzyme A are permeabilized with saponin, we can demonstrate a [3H]acetyl intermediate in the membrane that can transfer label to the 7- and 9-positions of exogenously added free N-acetylneuraminic acid but not to glucuronic acid or CMP-N-acetylneuraminic acid. This labeled acetyl intermediate represents a significant portion of the radioactivity incorporated into the membranes during the initial incubation and cannot be accounted for by nonspecifically "trapped" acetyl-CoA in the permeabilized vesicles. There was no evidence for involvement of acetylcarnitine or acetyl phosphate as an intermediate. The overall acetylation reaction appears to involve two steps. The first step (utilization of exogenous acetyl-CoA to form the acetyl intermediate) is inhibited by coenzyme A-SH (apparent Ki = 24-29 microM), whereas the second (transfer from the acetyl intermediate to sialic acid) is not affected by millimolar concentrations of the nucleotide. Studies with amino acid-modifying reagents indicate that 1 or more histidine residues are involved in the first step of the acetylation reaction. Diethylpyrocarbonate (which can react with both nonsubstituted and singly acetylated histidine residues) also blocks the second reaction, indicating that the acetyl intermediate on both sides of the membrane involves histidine residue(s). Taken together with data presented in the preceding paper, these results indicate that the acetylation of sialic acids in Golgi vesicles may occur by a transmembrane reaction, similar to that described for the acetylation of glucosamine in lysosomes (Bame, K. J., and Rome, L. H. (1985) J. Biol. Chem. 260, 11293-11299). However, several features of this Golgi reaction distinguish it from the lysosomal one, including the nature and kinetics of the reaction and the additional involvement of an essential lysine residue. The accumulation of free acetate in the lumen of the vesicles during the reaction may occur by abortive acetylation (viz. transfer of label from the acetyl intermediate to water). It is not clear if this is an artifact that occurs only in the in vitro reaction.     

Subsets of sialylated, sulfated mucins of diverse origins are recognized by L-selectin. Lack of evidence for unique oligosaccharide sequences mediating binding

Previous studies have shown that the mucin-type polypeptidesGlyCAM-1, CD34, and MAdCAM-1 can function as ligands for L-selectinonly when they are synthesized by the specialized high-endothelialvenules (HEV) of lymph nodes. Since sialylation, sulfation,and possibly fucosylation are required for generating recognition,we reasoned that other mucins known to have such componentsmight also bind L-selectin. We show here that soluble mucinssecreted by human colon carcinoma cells, as well as those derivedfrom human bronchial mucus can bind to human L-selectin in acalcium-dependent manner. As with GlyCAM-1 synthesized by lymphnode HEY, 2–3 linked sialic acids and sulfation seem toplay a critical role in generating this L-selectin binding.In each case, only a subset of the mucin molecules is recognizedby L-selectin. Binding is not destroyed by boiling, suggestingthat recognition may be based primarily upon carbohydrate structures.Despite this, O-linked oligosaccharide chains released fromthese ligands by beta-elimination do not show any detectablebinding to L-selectin. Following protease treatment of the ligands,binding persists in a subset of the resulting fragments, indicatingthat specific recognition is determined by certain regions ofthe original mucins. How ever, O-linked oligosaccharides releasedfrom the subset of non-binding mucin fragments do not show verydifferent size and charge profiles compared to those that dobind. Furthermore, studies with polylactosamine-degrading endoglycosidasessuggest that the core structures involved in generating bindingcan vary among the different ligands. Taken together, thesedata indicate that a single unique oligosaccharide structuremay not be responsible for high-affinity binding. Rather, diversemucins with sialylated, sulfated, fucosylated lactosamine-typeO-linked oligosaccharides can generate high-affinity L-selectinligands, but only when they present these chains in unique spacingand/or clustered combinations, presumably dictated by the polypeptidebackbone. L-selectin mucins sialic sialic acid sulfate adhesion     

Demonstration of the heterozygous state for I-cell disease and pseudo-Hurler polydystrophy by assay of N-acetylglucosaminylphosphotransferase in white blood cells and fibroblasts

The biochemical abnormalities of I-cell disease (mucolipidosis II) and pseudo-Hurler polydystrophy (mucolipidosis III) can be explained by a deficiency of the enzyme UDP-N-acetylglucosamine:lysosomal enzyme N-acetylglucosamine-1-phosphotransferase. We demonstrate here that obligate heterozygotes for these autosomal recessive diseases have intermediate levels of this enzymatic activity in homogenates of peripheral blood white cells and in extracts from cultured fibroblasts. This finding provides further evidence that the enzyme deficiency is the primary genetic defect in these diseases. In addition, the previous observation that obligate heterozygotes for mucolipidosis III have elevations of total serum beta-hexosaminidase outside the range of normal was confirmed. In studies of three pedigrees of patients with mucolipidosis III, these techniques were used to score individuals at risk for the carrier state.     

Biosynthesis and turnover of O-acetyl and N-acetyl groups in the gangliosides of human melanoma cells

We and others previously described the melanoma-associated oncofetal glycosphingolipid antigen 9-O-acetyl-GD3, a disialoganglioside O-acetylated at the 9-position of the outer sialic acid residue. We have now developed methods to examine the biosynthesis and turnover of disialogangliosides in cultured melanoma cells and in Golgi-enriched vesicles from these cells. O-Acetylation was selectively expressed on di- and trisialogangliosides, but not on monosialogangliosides, nor on glycoprotein-bound sialic acids. Double-labeling of cells with [3H]acetate and [14C]glucosamine introduced easily detectable labels into each of the components of the ganglioside molecules. Pulse-chase studies of such doubly labeled molecules indicated that the O-acetyl groups turn over faster than the parent molecule. When Golgi-enriched vesicles from these cells were incubated with [acetyl-3H]acetyl-coenzyme A, the major labeled products were disialogangliosides. [Acetyl-3H]O-acetyl groups were found at both the 7- and the 9-positions, indicating that both 7-O-acetyl GD3 and 9-O-acetyl GD3 were synthesized by the action of O-acetyltransferase(s) on endogenous GD3. Analysis of the metabolically labeled molecules confirmed the existence of both 7- and 9-O-acetylated GD3 in the intact cells. Surprisingly, the major 3H-labeled product of the in vitro labeling reaction was not O-acetyl-GD3, but GD3, with the label exclusively in the sialic acid residues. Fragmentation of the labeled sialic acids by enzymatic and chemical methods showed that the 3H-label was exclusively in [3H]N-acetyl groups. Analyses of the double-labeled sialic acids from intact cells also showed that the 3H-label from [3H]acetate was exclusively in the form of [3H]N-acetyl groups, whereas the 14C-label was at the 4-position. Pulse-chase analysis of the 3H/14C ratio showed that the N-acetyl groups of both GD3 and of the monosialoganglioside GM3 were turning over faster than the parent molecules. Selective periodate oxidation showed that both the inner and outer sialic acid residues of GD3 incorporated 3H-label in the in vitro reaction, and showed similar turnover of N-acetylation in the pulse-chase study. Taken together, these results indicate that both the O- and N-acetyl groups of the sialic acid residues of gangliosides turn over faster than the parent molecules. They also demonstrate a novel re-N-acetylation reaction that predicts the existence of de-N-acetyl gangliosides in melanoma cells.     

Proteomic comparison of human and great ape blood plasma reveals conserved glycosylation and differences in thyroid hormone metabolism

Most blood plasma proteins are glycosylated. These glycoproteins typically carry sialic acid-bearing sugar chains, which can modify the observed molecular weights and isoelectric points of those proteins during electrophoretic analyses. To explore changes in protein expression and glycosylation that occurred during great ape and human evolution, we subjected multiple blood plasma samples from all these species to high-resolution proteomic analysis. We found very few species-specific differences, indicating a remarkable degree of conservation of plasma protein expression and glycosylation during approximately 12 million years of evolution. A few lineage-specific differences in protein migration were noted among the great apes. The only obvious differences between humans and all great apes were an apparent decrease in transthyretin (prealbumin) and a change in haptoglobin isoforms (the latter was predictable from prior genetic studies). Quantitative studies of transthyretin in samples of blood plasma (synthesized primarily by the liver) and of cerebrospinal fluid (synthesized locally by the choroid plexus of the brain) confirmed approximately 2-fold higher levels in chimpanzees compared to humans. Since transthyretin binds thyroid hormones, we next compared plasma thyroid hormone parameters between humans and chimpanzees. The results indicate significant differences in the status of thyroid hormone metabolism, which represent the first known endocrine difference between these species. Notably, thyroid hormones are known to play major roles in the development, differentiation, and metabolism of many organs and tissues, including the brain and the cranium. Also, transthyretin is known to be the major carrier of thyroid hormone in the cerebrospinal fluid, likely regulating delivery of this hormone to the brain. A potential secondary difference in retinoid (vitamin A) metabolism is also noted. The implications of these findings for explaining unique features of human evolution are discussed.     

Enhanced 3-O-sulfation of galactose in Asn-linked glycans and Maackia amurensis lectin binding in a new Chinese hamster ovary cell line

We report the characterization of two Chinese hamster ovary cell lines that produce large amounts of sulfated N-linked oligosaccharides. Clones 26 and 489 were derived by stable transfection of the glycosaminoglycan-deficient cell mutant pgsA-745 with a cDNA library prepared from wild-type cells. Peptide:N-glycanase F released nearly all of the sulfate label, indicating that sulfation had occurred selectively on the Asn-linked glycans. Hydrazinolysis followed by nitrous acid treatment at pH 4 and borohydride reduction yielded reduced sulfated disaccharides that comigrated with standard Gal3SO4beta1-4anhydromannitol. The disaccharides were resistant to periodate oxidation but became sensitive after the sulfate group was removed by methanolysis, indicating that the sulfate was located at C3 of the galactose residues. Maackia amurensis lectin bound to the sulfated glycopeptides on the cell surface and in free form, even after sialidase treatment. This finding indicates that the lectin requires only a charged group at C3 of the galactose unit and not an intact sialic acid. Growth of cells with chlorate restored sialidase sensitivity to lectin binding, indicating that sulfation and sialylation occurred largely at the same sites. The enhanced sulfation was due to elevated sulfotransferase activity that catalyzed transfer of sulfate from phosphoadenosine-5'-phosphosulfate to Galbeta1-4(3)GlcNAcbeta-O-naphthalenemethanol.