Studies on mannose-containing glycopeptides from a normal and an SV40 transformed human cell.

The oligosaccharide moiety of cell-surface mannose-labelled glycopeptides from a normal (WI38) and an SV40 transformed cell (W118Va) have been investigated using specific glycosidases. Partially purified mannose-containing glycopeptides were separated into acidic and neutral species by high voltage paper electrophoresis. Endo-beta-N-acetylgucosaminidase D, in the presence of three exoglycosidases, released from the acidic glycopeptides of non-growing cells a product completely absent in growing cells. However, the acidic species from growing WI18 Va and WI38 were found to be similar in the products released by enzyme digestion. The neutral species from growing normal cells contained a proportion of the glycopeptides resistant to endoglycosidase D while those from the non-growing cells were almost free of these resistant species. The SV40 transformed cells were further enriched, when compared to normal cells (WI38), in these neutral resistant species. We suggest that the oligomannosyl core of the majority of the susceptible species contains three mannose residues while that of the resistant species contains between six and eight.     

Studies on mannose-containing glycopeptides from a normal and an SV40 transformed human cell

The oligosaccharide moiety of cell-surface mannose-labelled glycopeptides from a normal (WI38) and an SV40 transformed cell (WI18 Va) have been investigated using specific glycosidases. Partially purified mannose-containing glycopeptides were separated into acidic and neutral species by high voltage paper electrophoresis. Endo-β-N-acetylglucosaminidase D, in the presence of three exoglycosidases, released from the acidic glycopeptides of non-growing cells a product completely absent in growing cells. However, the acidic species from growing WI18 Va and WI38 were found to be similar in the products released by enzyme digestion. The neutral species from growing normal cells contained a proportion of the glycopeptides resistant to endoglycosidase D while those from the non-growing cells were almost free of these resistant species. The SV40 transformed cells were further enriched, when compared to normal cells (WI38), in these neutral resistant species. We suggest that the oligomannosyl core of the majority of the susceptible species contains three mannose residues while that of the resistant species contains between six and eught.     

Membrane glycopeptides from virus-transformed hamster fibroblasts and the normal counterpart.

Comparisons of membrane glycopeptides from baby hamster kidney fibroblasts (BHK21/C13) and a clone transformed by Rous sarcoma virus (C13/B4) were made by using cells metabolically labeled with radioactive D-glucose and L-fucose. Most of the glycopeptides were metabolically labeled with both the general and the specific glycoprotein precursors. The glycopeptides obtained from the cell surface by controlled trypsinization were representative of the surface membrane as shown by comparing them with those of purified membrane preparations. The trypsin-removable glycopeptides from both cell types were further processed and examined by successive chromatography on Sephadex G-50 and DEAE-cellulose. The chromatographic distribution patterns showed that each cell type had glycopeptides of similar characteristics, although the proportions of the glycopeptides differed dramatically between the two cell types. After transformation there was an increase in the larger, more highly charged glycopeptides. This was verified by the increased sialic acid content in these glycopeptides. Some of the glycopeptides were homogeneous after the size and charge separations, since a variety of procedures did not separate them further. The apparent homogeneity and reasonably few species obtained may be due to the methods of isolation, with the procedures selecting particular glycopeptides from the external portion of the membrane. These results corroborate the concept and show for the first time that virus transformation is accompanied by an increase in certain species of glycopeptides rather than de novo synthesis.     

Alterations of acidic and neutral glycopeptides during the post-natal growth period of the mouse

We have examined the electrophoretic pattern of mannose-labelled glycopeptides derived from liver, pancreas, heart, brain, lung and kidney. The various organs were removed from white Swiss mice of 4–5, 20 and 40 days of age as well as from adult animals. Our data demonstrated that each organ had a characteristic ratio of acidic to neutral glycopeptides and that this ratio, in some cases, was substantially modified as the animal grew, consistent with the hypothesis that these changes might play an important role in the development and physiological function of organs.     

Lectin affinity chromatography of glycopeptides and oligosaccharides from normal and lectin-resistant Chinese-hamster ovary cells.

The [3H]mannose-labelled glycopeptides from two lectin-resistant lines of Chinese-hamster ovary cells were fractionated by chromatography on lentil lectin-Sepharose and concanavalin A-agarose columns and subsequently analysed by gel filtration in comparison with the glycopeptides of the parental cell line. Essentially all of the [3H]mannose-labelled asparaginyl-oligosaccharides from the 'single-mutant' cells selected for resistance to phytohaemagglutinin and the 'double-mutant' cells selected for additional resistance to concanavalin A were not bound to lentil lectin, whereas approximately one-sixth of the parental-cell glycopeptides were bound and specifically eluted with alpha-methyl mannoside. These bound and eluted glycopeptides represented a specific subset of the complex acidic-type asparaginyl-oligosaccharides. The percentage of radiolabelled glycopeptides and oligosaccharides from each cell line that were specifically bound to concanavalin A was consistent with the relative sensitivities of the three cell lines to this lectin. The major radiolabelled species in the endoglycosidase digest of the 'double-mutant'-cell glycopeptides (Man4GlcNAc1-size neutral oligosaccharides) were not bound to concanavalin A, whereas essentially all of the other neutral-type oligosaccharides were bound. In addition, the larger neutral-type oligosaccharides (Man8--9GlcNAc1) were more strongly bound to concanavalin A than were either the smaller neutral-type or the di-antennary acidic-type structures.     

Characterization of glycopeptides isolated from membranes of F9 embryonal carcinoma cells

From cells of a nullipotential line of embryonal carcinoma was isolated a membrane fraction enriched in the cell surface F9 antigen. More than 40% of the radioactive fucose and galactose incorporated by cells into nondialyzable material was recovered in this membrane preparation, corresponding to an approximately 10-fold purification of the labeled material. Extreme heterogeneity of membrane glycoproteins labeled with these sugars was revealed by sodium dodecyl sulfate gel electrophoresis. Glycopeptides prepared by extensive pronase digestion of membranes labeled with fucose or galactose showed properties similar to those already described for fucose-labeled glycopeptides from whole cells. Namely, large glycopeptides eluted near the excluded volume of Sephadex G-50 column were the predominant glycopeptide species, while complex glycopeptides of molecular weight around 2500 were minor components. Therefore, these large glycopeptides, characteristic of embryonal carcinoma cells, are derived mainly from a variety of glycoproteins closely associated with the membrane system, most probably cell-surface membrane of the cells. The large glycopeptides were also significantly labeled with glucosamine, but only slightly with mannose; major components of mannose-labeled glycopeptides from the membranes were high-mannose glycopeptides of low molecular weight. Several experiments excluded the possibility that the larg glycopeptides are mucopolysaccharides, glycolipids or mucin-type glycoproteins with short oligosaccharide chains.     

Elucidation of glycoprotein structures by unspecific proteolysis and direct nanoESI mass spectrometric analysis of ZIC-HILIC-enriched glycopeptides

Protein glycosylation was explored by direct nanoESI MS and MS/MS analysis of ZIC-HILIC-enriched proteolytic glycopeptides without further separation or purification. In a previous publication, we demonstrated that a direct MS-based analysis of proteolytic glycopeptides is feasible for a number of proteins (Henning , S. J. Mass Spectrom. 2007 , 42 , 1415 - 21). This method has now been refined for two aspects: (1) separation of glycopeptides by use of ZIC-HILIC SPE and (2) the use of unspecific proteases like thermolysin, elastase, or a trypsin/chymotrypsin mixture leading per se to a mass-based separation, that is, small nonglycosylated peptides and almost exclusively glycopeptides at higher m/z values. Furthermore, the glycopeptides produced by the above proteases in general contain short peptide backbones thus improving-probably due to their higher hydrophilicity--the ZIC-HILIC-based separation. The combination of unspecific proteolysis, glycopeptide separation, and their direct MS analysis was successfully accomplished for probing glycoproteins carrying high-mannose type (ribonuclease B), neutral (asialofetuin), and acidic (haptoglobin and α1-acid glycoprotein) complex type glycans as well as for glycopeptides derived from glycoprotein mixtures and, finally, for exploring the glycosylation of a human IgG preparation. Our results show that the presented method is a fast, facile, and inexpensive procedure for the elucidation of protein N-glycosylation.     

Endoglycosidase H-sensitive glycopeptides in eleven different animal cells.

We have examined the distribution of mannose-labelled glycopeptides in eleven different animal cells grown in vitro. All of the cells examined contained endoglycosidase H-sensitive species of high and low molecular weight, associated with the cell material and with the cell surface; however, the distribution between the two pools was different, suggesting a 'sorting out' of glycoproteins. Another conclusion from our studies is that the oligosaccharide processing known to occur during or after membrane glycoprotein translation is incomplete in a high percentage of mannose-containing N-linked oligosaccharides of cell surfaces. There was no consistent correlation between the relative amounts of endoglycosidase H-sensitive and -resistant glycopeptides and whether the cells were normal, virus-transformed or tumour-derived.     

Synthetic glycopeptides for the development of tumour-selective vaccines.

Based on structural information reported for the tumour-associated epithelial mucin MUC1, glycopeptides have been synthesized which contain tumour-associated saccharide antigens. such as the Thomsen-Friedenreich (T), TN or sialyl TN antigen. in combination with peptide sequences of the tandem repeat region of MUC1. Solid-phase syntheses have been carried out using N-Fmoc protected O-glycosyl serine and threonine building blocks and an allylic anchor which is stable to basic and acidic conditions, but can be cleaved under neutral conditions in a palladium(0)-catalysed allyl transfer reaction. In addition. a (2-3)sialyl T antigen threonine building block was prepared by a chemoenzymatic strategy and used in the synthesis of an N-terminal glycopeptide antigen of leukosialin (CD43). The proliferation of cytotoxic T cells could be induced using a construct consisting of a MUC1-glycopeptide antigen and a T cell epitope.     

Comparison of glycopeptides from control and virus-transformed baby hamster kidney fibroblasts

Glucosamine-labeled glycopeptides from control and virus-transformed BHK fibroblasts were characterized by size, lectin affinity, charge, and composition. As already demonstrated, on the basis of elution position on a column of Sephadex G-50, transformed cells contained a greater proportion of large glycopeptides than did control cells. Transformed cells also contained a larger proportion of glycopeptides which do not bind to Con A-Sepharose. By sequential chromatography on Sephadex G-50, Con A-Sepharose, and DEAE-Sephadex, approximately 40 individual peaks were partially or completely resolved. If sialic acid was removed from the glycopeptides prior to analysis by ion-exchange chromatography, 95% of the glycopeptides from control cells and 85% of the glycopeptides from transformed cells were no longer bound by DEAE-Sephadex. It was concluded that the DEAE-Sephadex elution properties of the glycopeptides are determined almost entirely by the sialic acid content of the molecules. A comparison of the profiles of control and transformed cell glycopeptides simultaneously eluting from columns of DEAE-Sephadex revealed that the differences between the two cells were largely quantitative; however, the possibility of the existence of qualitative differences as well cannot be excluded. In particular, there was one component present on the surface of transformed cells that was virtually absent in control cells. It was degraded by nitrous acid hydrolysis and heparinase and appeared to be heparan sulfate like material. After fractionation, each isolated glycopeptide population was analyzed for carbohydrate and, in some cases, amino acid content. The apparently larger glycopeptides, group A, the dominant population in transformed cells, were found to contain 3 to 4 mannose residues/glycopeptide when the sugars were normalized to sialic acid content. On the basis of the same criteria, group B glycopeptides contained 4-6 mannose residues/glycopeptide. The carbohydrate and amino acid compositions of the glycopeptides from transformed cells were, with a few exceptions, similar to those from control cells. Some isolated glycopeptides appeared to contain both O-glycosidic anad N-glycosidic linkages on the same oligopeptide.     

Analysis of glycopeptides as borate complexes by polyacrylamide gel electrophoresis

A new method using polyacrylamide gel electrophoresis in a Tris-borate buffer to analyze Pronase-derived glycopeptides is described. Examination of immunoglobulin, Sindbis virus, and ovalbumin-radiolabeled glycopeptides by this system demonstrates a pattern similar to that seen after Bio-Gel-P-6 chromatography and, in addition, exposes a heterogeneity in the immunoglobulin and Sindbis virus glycopeptides not apparent after gel filtration. The resolution of glycopeptides by gel electrophoresis depends on the inclusion of borate ions in the sample, the gel, and the electrophoresis buffer. The borate ions react with neutral sugars, converting them to charged complexes which migrate during electrophoresis. The number of borate ions bound to a glycopeptide is a function of the composition, sequence, and linkages of the carbohydrates. Gel electrophoresis of glycopeptides in a borate buffer has several advantages: (1) The method requires no new equipment or special skills beyond those necessary for conventional polyacrylamide gel electrophoresis; (2) when performed on a slab gel, up to 24 samples can be analyzed simultaneously; and (3) since detection is by radio-autography, small amounts of radiolabeled glycopeptides can be visualized by prolonging the exposure time. These characteristics are advantageous for studies of glycopeptides based on digestion products resulting from incubations with specific exo- and endo-glycosidases. Untreated glycopeptides have been compared on the same gel with glycopeptides sequentially treated with different glycosidases to gain structural information.     

Analysis of IgG glycopeptides by alkaline borate gel filtration chromatography

A protocol for the characterization of IgG glycopeptides is described. Central to this scheme is the novel application of an alkaline borate buffer to gel filtration chromatography. The use of this buffer significantly enhances the resolution of glycopeptides. Furthermore, it results in the separation of a unique size class of glycopeptides derived from IgG secreted by murine hybridomas. Although predominantly neutral, these glycopeptides differ both qualitatively and quantitatively by lectin affinity chromatography from the other glycopeptides which are presumably derived from the Fc portion of IgG.     

BHK21 fibroblast aggregation inhibited by glycopeptides from the cell surface.

Glycopeptides were removed by trypsinization from the surface of baby hamster kidney cells (line BHK21-C13), digested by pronase and separated into 2 fractions by exclusion chromatography. The addition of small amounts of either glycopeptide fraction to shaken suspensions of lightly trypsinzied cells inhibited their rapid aggregation, but one fraction was more active than the other and in higher concentrations it was able to inhibit aggregation completely. After this fraction was purified by high-voltage electrophoresis one subfraction also inhibited aggregation. The effect of the glycopeptides increased following their pretreatment with neuraminidase, but preincubation with periodiate or galactose oxidase destroyed all activity. Galactose oxidase also inhibited cell aggregation directly. Similar glycopeptides from virus-transformed BHK21 cells, oligosaccharides and intact and desialysed human urinary glycoproteins had comparatively little or no effect on BHK21 cell aggregation. The results suggest terminal beta-galactosides and possible alpha-galactosides, and to some extent a particular substructure of cell surface heteroglycans are necessary for their inhibitory activity. The parent, plasma membrane of glycoproteins might serve as adhesive binding sites in cell cohesion, but some evidence indicates cell surface sialyl- and galactosyltransferases may not ordinarily act as their complementary binding receptors.     

Developmentally-related changes in surface membrane glycopeptides of murine haemopoietic cells.

We have carried out a comparative study of mature murine granulocytes with two immature haemopoietic cell lines (multipotential cells, FDCP-Mix, and granulocyte progenitor cells, FDCP-2) with respect to the structure and composition of their surface membrane glycopeptides. The glycopeptides were labelled biosynthetically by incubation of the cells for 1-3 days with [3H]glucosamine. Cell-associated glycopeptides were released by treatment with trypsin and the trypsin extract was exhaustively digested with Pronase to remove most residual peptide. Radiolabelled materials were fractionated by chromatography on lectin affinity columns connected in the series: lentil lectin (LCA), concanavalin A (Con A) and wheat germ agglutinin (WGA). Lectin-binding glycopeptides were eluted with appropriate competing sugars and further analysed by gel filtration, base/borohydride elimination and susceptibility to degradation by glycosidases including endo-beta-galactosidase. Abundant quantities of N-linked polylactosamine-type glycopeptides, which bound only to the WGA columns, were identified on mature granulocytes but the molecules were highly-branched (i.e. resistant to endo-beta-galactosidase). In contrast, there seemed to be very little branching in the polylactosamine chains from FDCP-2 cells, whilst corresponding carbohydrates from multipotential FDCP-Mix cells gave evidence for both linear and branched domains in the same, large complex glycans. O-Linked tetrasaccharides of general structure: NeuAc-Gal-(NeuAc)-GalNAc were found in clusters on WGA-binding glycopeptides from all cell types, these components being especially prominent on mature granulocytes. FDCP-2 cells were distinguished by the presence of monosialylated and non-sialylated counterparts of the foregoing tetrasaccharides. The relative amount of LCA-binding glycopeptides was low on FDCP-Mix cells by comparison with FDCP-2 cells and mature granulocytes. Our findings therefore demonstrate that notable differences in gross composition and molecular fine structure of surface membrane glycopeptides are detectable in haemopoietic cells at different stages of development. The relationship of these differences to the biological properties of cell surfaces remains to be established.     

Changes in cell-surface fucose-containing glycopeptides and adhesion of cultured intestinal epithelial cells as a function of cell density.

Confluent cultured intestinal epithelial cells displayed greater adhesion to the substratum than did subconfluent cells. Subconfluent and confluent cells were labelled with [3H]fucose for 24h and the cell-surface components were released by mild Pronase treatment. After extensive Pronase digestion, cell-surface and cell-residue glycopeptides were fractionated on Bio-Gel P-6. The cell surface contained a higher proportion of lower-molecular-weight glycopeptides than the residue. No significant difference in elution pattern was found between total cell-surface glycopeptides of subconfluent and confluent cells. However, confluent cells contained almost twice as much [3H]-fucose-labelled glycopeptides that were bound to concanavalin A-Sepharose and were subsequently eluted with 20mM-methyl alpha-D-glucopyranoside as subconfluent cells. When the bound glycopeptides were chromatographed on Bio-Gel P-6, it was found that confluent cells contained a larger proportion of lower-molecular-weight glycopeptides than subconfluent cells. This difference in size was eliminated after treatment of glycopeptides with sialidase. When growth of subconfluent cells was inhibited with a non-toxic concentration of retinoic acid, no significant effect on the elution pattern of [3H]fucose-labelled glycopeptides was observed on either Bio-Gel P-6 or concanavalin A-Sepharose. No significant difference was found in the total [3H]fucose-labelled glycoproteins from subconfluent and confluent cells by two-dimensional gel electrophoresis. It is suggested that the differences in [3H]fucose-labelled glycopeptides between subconfluent and confluent cells are cell-density-dependent rather than growth-dependent, and that these differences are likely to result from some changes in glycosylation mechanism(s). Furthermore, the differences in cell-surface glycopeptides may be related to the changes in the adhesion of the cells to the substratum.     

Cell-density-dependent changes in cell-surface glycopeptides and in adhesion of cultured intestinal epithelial cells

We studied mannose-containing glycopeptides and glycoproteins of subconfluent and confluent intestinal epithelial cells in culture. Cells were labelled with d-[2-³H]mannose for 24h and treated with Pronase or trypsin to release cell-surface components. The cell-surface and cell-residue fractions were then exhaustively digested with Pronase and the resulting glycopeptides were fractionated on Bio-Gel P-6, before and after treatment with endo-β-N-acetylglucosaminidase H to distinguish between high-mannose and complex oligosaccharides. The cell-surface glycopeptides were enriched in complex oligosaccharides as compared with residue glycopeptides, which contained predominantly high-mannose oligosaccharides. Cell-surface glycopeptides of confluent cells contained a much higher proportion of complex oligosaccharides than did glycopeptides from subconfluent cells. The ability of the cells to bind [³H]concanavalin A decreased linearly with increasing cell density up to 5 days in culture and then remained constant. When growth of the cells was completely inhibited by either retinoic acid or cortisol, no significant difference was observed in the ratio of complex to high-mannose oligosaccharides in the cell-surface glycopeptides of subconfluent cells. Only minor differences were found in total mannose-labelled glycoproteins between subconfluent and confluent cells by two-dimensional gel analysis. The adhesion of the cells to the substratum was measured at different stages of growth and cell density. Subconfluent cells displayed a relatively weak adhesion, which markedly increased with increased cell density up to 6 days in culture. It is suggested that alterations in the structure of the carbohydrates of the cell-surface glycoproteins are dependent on cell density rather than on cell growth. These changes in the glycopeptides are correlated with the changes in adhesion of the cells to the substratum.     

Selective enrichment of sialic acid-containing glycopeptides using titanium dioxide chromatography with analysis by HILIC and mass spectrometry

The terminal monosaccharide of cell surface glycoconjugates is typically a sialic acid (SA), and aberrant sialylation is involved in several diseases. Several methodological approaches in sample preparation and subsequent analysis using mass spectrometry (MS) have enabled the identification of glycosylation sites and the characterization of glycan structures. In this paper, we describe a protocol for the selective enrichment of SA-containing glycopeptides using a combination of titanium dioxide (TiO(2)) and hydrophilic interaction liquid chromatography (HILIC). The selectivity of TiO(2) toward SA-containing glycopeptides is achieved by using a low-pH buffer that contains a substituted acid such as glycolic acid to improve the binding efficiency and selectivity of SA-containing glycopeptides to the TiO(2) resin. By combining TiO(2) enrichment of sialylated glycopeptides with HILIC separation of deglycosylated peptides, a more comprehensive analysis of formerly sialylated glycopeptides by MS can be achieved. Here we illustrate the efficiency of the method by the identification of 1,632 unique formerly sialylated glycopeptides from 817 sialylated glycoproteins. The TiO(2)/HILIC protocol requires 2 d and the entire procedure from protein isolation can be performed in <5 d, depending on the time taken to analyze data.     

The covalent structure of individual N-linked glycopeptides from ovomucoid and asialofetuin

In order to explore whether individual N-linked glycans in a given glycoprotein may be processed to different end products and at the same time prepare a number of well characterized glycopeptides as substrates for glycopeptide hydrolases, we have prepared the individual glycopeptides representing the four major glycosylation sites in ovomucoid and the three sites in asialofetuin. The individual glycopeptides were characterized by amino acid sequence determination before and after removal of the glycan by peptide:N-glycanase (amidase), and the liberated glycans were subjected to mass spectrometric analysis. As expected from available sugar analyses of the individual glycans in ovomucoid, no major differences were detected between the four glycosylation sites in this glycoprotein, but a definite trend toward less processed (less extensively branched) species was observed in going from site 1 to 4. In fetuin, for which the glycan pool is known to be made up of about two-thirds triantennary and one-third biantennary structures, the analysis of the three glycopeptides gave triantennary to biantennary ratios of 75/25, 67/33, and 70/30, respectively, demonstrating that the three sites are processed to a very similar, albeit perhaps not identical, extent. All the glycopeptides obtained in these studies, including the CNBr-produced glycopeptide from ovalbumin, were purified by a set series of steps, gel filtration on Sephadex G-50 followed by ion-exchange chromatography on DE52 and/or reverse phase high performance liquid chromatography. Based on the results, these procedures appear to have general application for the preparation of glycopeptides.     

Characterization of fucosyl glycopeptides from cell surface and cellular material of rat fibroblasts

Approximately 70% of fucose-labeled glycopeptides from the cell surface and cellular material of rat fibroblasts (3Y1B cells) were hydrolyzed by endo-β-N-acetylglucosaminidase D in the presence of neuraminidase, β-glactosidase and β-N-acetylglucosaminidase. Structure of the suspceptible glycopeptides were found to be very similar to non-membrane glycopeptides of the complex heteropolysaccharide unit, such as the sialylated glycopeptides of thyroglobulin. On the other hand, the resistant glycopeptides were also refractory toward endo-β-N-acethylglucosaminidase H and α-mannosidase, and appeared to be a mixture of glycopeptides with unique structures.     

Partial characterization of rat hepatoma cell-surface glycopeptides possessing concanavalin A receptor activity.

Papain digestion of Novikoff or AS-30D rat hepatoma cells released glycopeptides from the cell surface. That portion of the glycopeptides accessible to Sephadex G-50 was digested with pronase and the component glycopeptides partially resolved by ion-exchange chromatography. Each tumor type yielded two well resolved sialoglycopeptide fractions which possessed concanavalin A receptor activity. The amino acid and saccharide composition of these low molecular weight (3,100 ± 300 daltons) sialoglycopeptides was determined.