Deglycosylation of glycoproteins by trifluoromethanesulfonic acid

Neutral and acidic human tracheobronchial mucin. Isolation and characterization of core protein

Human bronchial mucin from a patient suffering from chronic bronchitis was solubilized in aqueous solution containing sodium azide and protease inhibitors and purified by Sepharose 4B and 2B column chromatography. The mucin was further purified by cesium bromide density gradient centrifugation. Sodium dodecyl sulfate-polyacrylamide gel (7.5%) electrophoresis of this material showed high-molecular-weight mucin component(s) at the top of the gel. Chemical analysis of this preparation indicated a typical mucin profile of amino acids and carbohydrates. Ion-exchange chromatography resulted in resolution of the purified mucin into neutral and acidic fractions. Comparison of the chemical composition of these two fractions showed higher mole percentage of threonine, serine, sialic acid, and sulfate in the acidic fraction. Chemical deglycosylation of the purified mucin preparation with trifluoromethane sulfonic acid was carried out at 20 degrees C for 3 1/2 h. Sialic acid, fucose, galactose, and N-acetylglucosamine were completely removed, whereas traces of N-acetylgalactosamine were still detected. High-pressure liquid chromatography of the deglycosylated products from native, neutral, and acidic mucin preparations resulted in a principal peptide, P1, with identical amino acid composition. Cyanogen bromide (CNBr) treatment of the peptide P1 from neutral and acidic mucins and subsequent fractionation of the fragments by high-pressure liquid chromatography resulted in similar peptide profiles. The P1 peptide fraction was further subjected to high-pressure liquid chromatography in a second solvent system, which resulted in two peaks, P1a and P1b. Gel filtration of both peptides in 6 M guanidine hydrochloride indicated a single peak with molecular weight of approximately 97 kDa. The amino acid profile of the two peptides was dominated by high levels of threonine, serine, and proline, which combined accounted for nearly 39% of the total residues, and in most respects, the profile resembled that of native mucin. End-group analysis of the peptide P1a indicated a blocked N-terminus, whereas serine was found to be the N-terminal amino acid in the peptide P1b. Rabbit antibodies prepared against the peptide P1 from native tracheal mucin reacted strongly with neutral and acidic mucin as well as the mucin from human colon. Both neutral and acidic human tracheal mucins were immunologically reactive with mouse monoclonal antibody HMPFG-2, which was prepared against human mammary mucin. However, the response of this antibody to human colonic mucin was rather weak.

Sodium channel polypeptides in central nervous systems of various insects identified with site directed antibodies

Immunoprecipitation, radiophosphorylation and SDS-PAGE autoradiography enable the characterization of sodium channel polypeptides in the central nervous system of insects belonging to four phylogenetically distinct orders: grasshoppers, cockroaches, flies and moth larvae. It has been shown that the insect sodium channels: (1) Are recognized by the previously described (Gordon et al. (1988) Biochemistry 27, 7032-7038) site directed antibodies corresponding to a highly conserved segment linking the homologous domains III and IV in the vertebrate sodium channel alpha subunits. (2) Serve as substrates for phosphorylation by cAMP-dependent protein kinase. (3) Are devoid of disulfide linkage to smaller subunits unlike sodium channels in vertebrate brain. (4) Are glycoproteins as shown in the grasshopper by the decrease of apparent molecular weight following endoglycosidase F treatment and specific binding to the lectins concanavalin A and wheat germ agglutinin. (5) Reveal a diversity with regard to their (a) apparent molecular masses which range from 240 to 280 kDa and (b) V8 proteinase digestion phosphopeptides indicating either differences in the positioning of the enzymatic cleavage and/or phosphorylation sites. These results provide the first evidence for structural diversity of sodium channel subtypes among various insect orders and are compared to their mammalian counterparts.

Electric properties of photoaffinity-labelled pancreatic A-subtype cholecystokinin

Although the isoelectric point of a protein is very important, electric focusing of intrinsic membrane proteins in polyacrylamide or agarose gels often fails. The recently introduced Bio-Rad Rotofor cell allowed isoelectric focusing of such a protein, cholecystokinin (CCK) receptor. Both the isoelectric point and the molecular weight (Mr) of pancreatic CCK receptor were determined. For this purpose, membrane CCK receptor was photoaffinity labelled by a cleavable agonist probe, subsequently prepurified on immobilized wheat germ agglutinin and analysed by sodium dodecyl sulphate polyacrylamide gel electrophoresis and isoelectrofocusing in solution in the presence of Nonidet P-40. CCK receptor was identified at Mr 85,000-100,000, whereas its deglycosylated product was shifted to Mr 42,000. Further, the isoelectric points of the glycosylated and deglycosylated forms of CCK receptor were pH 4.8 and 4.3, respectively. A knowledge of the isoelectric point should help in characterizing better CCK receptor heterogeneity and/or in purifying CCK receptor proteins.

Heparan sulfate proteoglycan from human tubular basement membrane. Comparison with this component from the glomerular basement membrane

Heparan sulfate proteoglycan (HSPG) was extracted from human tubular basement membrane (TBM) with guanidine and purified by ion-exchange chromatography and gel filtration. The glycoconjugate was sensitive to heparitinase and resistant to chondroitinase ABC, had an apparent molecular mass of 200-400 kDa and consisted of 70% protein and 30% glycosaminoglycan. The amino acid composition was characterized by its high content of glycine, proline, alanine and glutamic acid. Hydrolysis with trifluoromethanesulfonic acid yielded core proteins of 160 and 110 kDa. The heparan sulfate (HS) chains obtained after alkaline NaBH4 treatment had a molecular mass of about 18 kDa. Results of heparitinase digestion and HNO2 treatment suggest a clustering of sulfate groups in the distal portion of the HS side chains. These chemical data are comparable to those obtained previously on glomerular basement membrane (GBM) HSPG (Van den Heuvel et al. (1989) Biochem. J. 264, 457-465). Peptide patterns obtained after trypsin, clostripain or V8 protease digestion of TBM and GBM HSPG preparations showed a large similarity. Polyclonal antisera and a panel of monoclonal antibodies raised against both HSPG preparations and directed against the core protein showed complete cross-reactivity in ELISA and on Western blots. They stained all basement membranes in an intense linear fashion in indirect immunofluorescence studies on human kidneys. Based on these biochemical and immunological data we conclude that HSPGs from human GBM and TBM are identical, or at least very closely related, proteins.

Isolation, characterization, and cellular insertion of the flagella from two strains of the archaebacterium Methanospirillum hungatei

In high (45 mM)-phosphate medium, Methanospirillum hungatei strains GP1 and JF1 grew as very long, nonmotile chains of cells that did not possess flagella. However, growth in lower (3 or 30 mM)-phosphate medium resulted in the production of mostly single cells and short chains that were motile by means of two polar tufts of flagella, which transected the multilayered terminal plug of the cell. Electron microscopy of negatively stained whole mounts revealed a flagellar filament diameter of approximately 10 nm. Flagellar filaments were isolated from either culture fluid or concentrated cell suspensions that were subjected to shearing. Flagellar filaments were sensitive to treatment with both Triton X-100 and Triton X-114 at concentrations as low as 0.1% (vol/vol). The filaments of both strains were composed of two flagellins of Mr 24,000 and 25,000. However, variations in trace element composition of the medium resulted in the production of a third flagellin in strain JF1. This additional flagellin appeared as a ladderlike smear on sodium dodecyl sulfate-polyacylamide gels with a center of intensity of Mr 35,000 and cross-reacted with antisera produced from filaments containing only the Mr-24,000 and -25,000 flagellins. On sodium dodecyl sulfate-polyacrylamide gels, all flagellins stained by the thymol-sulfuric acid and Alcian blue methods, suggesting that they were glycosylated. This was further supported by chemical deglycosylation of the strain JF1 flagellins, which resulted in a reduction in their apparent molecular weight on sodium dodecyl sulfate-polyacylamide gels. Heterologous reactions to sera raised against the flagella from each strain were limited to the Mr-24,000 flagellins.

Characterization of Trichinella spiralis antigens sharing an immunodominant, carbohydrate-associated determinant distinct from phosphorylcholine

The biochemical and immunochemical characteristics of T. spiralis molecules (group II antigens) sharing an immunodominant epitope were examined. Six major proteins, ranging from 43-68 kDa, and from pI 5.0-6.3, express the determinant. Together, they account for at least 3% by weight of the total protein in L1 larval homogenate. The antigens are glycosylated. Following periodate oxidation, they reacted with biotin aminocaproyl hydrazide, and treatment with trifluoromethanesulfonic acid decreased their Mr. Deglycosylated group II antigens lost immunoreactivity with a monoclonal antibody specific for the determinant, and oligosaccharides released by treatment with mild base blocked binding of the monoclonal antibody to native antigens. The determinant on one of the group II antigens (43 kDa) was removed by N-glycanase. Neither phosphorylcholine nor antibody to phosphorylcholine interfered with monoclonal antibody binding to native group II antigens. Together, these results suggest that the immunodominant group II antigen epitope is associated with N- and O-linked oligosaccharides, and that it is not phosphorylcholine.

Recovery of an epitope recognized by a novel monoclonal antibody from airway lavage during experimental induction of chronic bronchitis

Prolonged exposure of dogs to high concentrations of SO2 gas results in a syndrome with many of the characteristics of human chronic bronchitis, including cough and chronic mucous hypersecretion as well as airway obstruction. We developed and used a novel monoclonal antibody, GB-4B, raised against epithelial glycoprotein isolated from human hypersecretory mucus to probe airway lavage samples from dogs before and during prolonged exposure to SO2 gas. There were relatively low mean titers of the epitope recognized by GB-4B in airway lavage fluid as evidenced by enzyme-linked immunosorbent assay before exposure to SO2 gas. After 25 to 50 wk of SO2 exposure, the dogs showed a significant increase in pulmonary resistance and there was a significant increase in the titer of the epitope in the airway lavage fluid. Using the same antibody immunohistochemical analysis of airway tissues from SO2-exposed dogs revealed patchy staining of the mucous glands and airway secretory cells and dense staining along the airway surface; airway tissue from control dogs and one SO2-exposed dog whose lavage fluid did not contain the epitope showed little or no staining. These data demonstrate that similar mucin epitopes appear in airway lavage fluid under hypersecretory conditions in both animals and humans. The epitope may have utility as a marker of chronic mucous hypersecretion.

Differential splicing generates a nervous system-specific form of Drosophila neuroglian

We recently described the characterization and cloning of Drosophila neuroglian, a member of the immunoglobulin superfamily. Neuroglian contains six immunoglobulin-like domains and five fibronectin type III domains and shows strong sequence homology to the mouse neural cell adhesion molecule L1. Here we show that the neuroglian gene generates at least two different protein products by tissue-specific alternative splicing. The two protein forms differ in their cytoplasmic domains. The long form is restricted to the surface of neurons in the CNS and neurons and some support cells in the PNS; in contrast, the short form is expressed on a wide range of other cells and tissues. Thus, whereas the mouse L1 gene appears to encode only one protein that functions largely as a neural cell adhesion molecule, its Drosophila homolog, the neuroglian gene, encodes at least two protein forms that may play two different roles, one as a neural cell adhesion molecule and the other as a more general cell adhesion molecule involved in other tissues and imaginal disc morphogenesis.

Evidence that beta-amyloid protein in Alzheimer's disease is not derived by normal processing

The beta-amyloid protein (beta/A4), derived from a larger amyloid precursor protein (APP), is the principal component of senile plaques in Alzheimer's disease. APP is an integral membrane glycoprotein and is secreted as a carboxyl-terminal truncated molecule. APP cleavage, which is a membrane-associated event, occurred at a site located within the beta/A4 region. This suggests that an intact amyloidogenic beta/A4 fragment is not generated during normal APP catabolism. Therefore, an early event in amyloid formation may involve altered APP processing that results in the release and subsequent deposition of intact beta/A4.

Role of the putative "link" glycopeptide of intestinal mucin in binding of piliated Escherichia coli serotype O157:H7 strain CL-49

Purified rat intestinal mucin was used to identify mucin-binding sites for type 1-piliated Escherichia coli O157:H7 strain CL-49 isolated from a patient with hemorrhagic colitis and hemolytic uremic syndrome. Optimum binding of bacteria in a microtiter binding assay occurred with a mucin coating concentration of 15 micrograms (protein)/150 microliters. In hapten inhibition studies, several nonmucin glycoproteins bearing exposed mannosyl residues in N-linked oligosaccharides were effective inhibitors, as was rat mucin. The same glycoproteins caused bacterial aggregation. High-molecular-mass glycoproteins of the mucin were separated from its 118-kilodalton "link" glycopeptide fraction, and the latter was shown to be the mucin-binding component for E. coli CL-49 and its purified type 1 pili. This was confirmed in hemagglutination inhibition studies. Treatment of the link glycopeptide with jack bean alpha-mannosidase or endo-beta-N-acetylglucosaminidase H destroyed bacterial binding activity. Chemical or enzymatic modifications of intact rat mucin were undertaken to evaluate the normal accessibility of the link glycopeptide receptors to E. coli CL-49. Deglycosylation with trifluoromethane-sulfonic acid abolished binding, whereas pronase digestion had no effect. Reduction and alkylation as well as lipid extraction enhanced bacterial binding by the mucin, presumably by causing greater exposure of receptor sites. In summary, our binding studies revealed, for the first time, that intestinal mucin bears oligomannosyl receptors for type 1 pili and that these receptors are located on N-linked oligosaccharides of the 118-kilodalton link glycopeptide region of the mucin. Our experiments suggest the receptors are normally partly "covered" by noncovalently bound lipid. In addition, release of the link component from the rest of the mucin by disulfide bond reduction causes greater exposure of specific bacterium-binding sites.

Glycan research on barley, maize, oats, and sorghum grain alpha-amylases: comparison with rice alpha-amylase

alpha-Amylases from germinated maize, oats, rice, and sorghum were isolated by glycogen precipitation and hydrophobic interaction chromatography. Several methods were used for the detection of glycoproteins, including barley alpha-amylase isozymes purified as previously described and using the rice alpha-amylase as a positive control for glycosylation. Affinoblotting using concanavalin A, immunoblotting using a xylose-specific serum which reacts with complex N-linked glycans, and endo-beta-N-acetylglucosaminidase H treatment of amylases gave negative results for maize, oats, sorghum, and barley. However, after deglycosylation with trifluoromethanesulfonic acid, the molecular weight of one maize alpha-amylase constituent was clearly decreased. The same result was obtained after beta-elimination in mild conditions. Together these results indicated probable O-linked glycosylation of one maize alpha-amylase when barley, oats, and sorghum alpha-amylases did not appear to be glycosylated. Chemical deglycosylation of rice alpha-amylase resulted in the production of two polypeptides with different molecular weights.

Monoclonal antibody recognizing an apparent peptide epitope of human seminal plasma glycoprotein and exhibiting sperm immobilizing activity

A hybridoma (3B2-F7) has been established which secretes a monoclonal antibody (Mab) directed against a peptide determinant of human seminal plasma glycoprotein (HSP-gP). The deglycosylation of HSP-gP was performed chemically with TFMS hydrolysis and enzymatically in the presence of detergent and further treated with periodic acid after fixing deglycosylated HSP on plastic wells. The Mab 3B2-F7 (IgM, kappa) exhibited sperm immobilization activity (256 units of SI50) and inhibited sperm binding to human zona pellucida. Human epididymis, pancreatic islets of Langerhan's and distal tubulus of kidney were strongly labelled whilst other tissues were essentially negative by avidin-biotin complex tissue staining with this Mab. The antigen epitope to the Mab was in the 36 kDa molecule of human HSP-gP. The antigenic determinant recognized by Mab 3B2-F7 was destroyed by six different proteases, but was resistant to N-glycanase and other carbohydrate splitting enzymes. This epitope is therefore likely to be composed of a polypeptide chain. Peptide fragments after proteolysis of the HSP molecule with Staph. aureus V8 protease and trypsin retained antigenicity, hence the epitope corresponding to the Mab may be a peptide chain and not dependent on the conformational structure of the polypeptide.

Golgi-mediated post-translational processing of secretory acid phosphatase by Leishmania donovani promastigotes

Monensin, an inhibitor of Golgi function, was used to investigate the role of this cell compartment in the glycosylation of Leishmania donovani promastigote secretory acid phosphatase (EC 3.1.3.2). Monensin-treated cells demonstrated morphological changes in the Golgi complex and secreted enzyme with an altered electrophoretic mobility: two discrete bands of approximately 95 and 110 kDa were found, as compared to the heterodisperse nature of the enzyme from untreated controls. Chemical deglycosylation by mild acid hydrolysis resulted in a similar effect on the electrophoretic mobility of purified extracellular enzyme. Acid phosphatase was also treated with N-glycosidase F (EC 3.5.1.52) to remove N-linked oligosaccharides. The altered lectin-binding properties of the enzyme after these two treatments demonstrated that an unusual type of galactose-containing acid-labile carbohydrate was present in secretory acid phosphatase in addition to the N-linked oligosaccharides. Further, experiments with 32P-labelled enzyme indicated that phosphodiester bonds were the structural component responsible for the sensitivity of this carbohydrate to mild acid hydrolysis. Cumulatively, these results demonstrated that a novel form of Golgi-mediated posttranslational modification had occurred to the secretory acid phosphatase presumably by the addition of an acid-labile phosphoglycan.

Identification of a highly reactive threonine residue at the active site of gamma-glutamyl transpeptidase

gamma-Glutamyl transpeptidase [(5-glutamyl)-peptide:amino-acid 5-glutamyltransferase, EC 2.3.2.2], an enzyme of major importance in glutathione metabolism, was inactivated by treating it with L-(alpha S,5S)-alpha-amino-3-chloro-4,5-dihydro-5-[3-14C]isoxazoleacetic acid. This selective reagent binds stoichiometrically to the enzyme; more than 90% of the label was bound to its light subunit. Enzymatic digestion of the light subunit gave a 14C-labeled peptide that corresponds to amino acid residues 517-527 of the enzyme and two incomplete digestion products that contain this labeled peptide moiety. The radioactivity associated with this peptide was released with threonine-523 during sequencing by the automated gas-phase Edman method. The light subunit contains 14 other threonine residues and a total of 19 serine residues; these were not labeled. Threonine-523 is situated in the enzyme in an environment that greatly increases its reactivity, indicating that other amino acid residues of the enzyme must also participate in the active-site chemistry of the enzyme.

Peptidoglycan-associated polypeptides of Mycobacterium tuberculosis

Important protein-based immunoreactivities have long been associated with the cell wall core of mycobacteria. In order to explore the molecular basis of such activities, purified cell walls of Mycobacterium tuberculosis were extracted with sodium dodecyl sulfate to produce an insoluble residue composed of the mycolylarabinogalactan-peptidoglycan complex and about 2% of unextractable protein. Treatment of the product from an avirulent strain of M. tuberculosis with trifluoromethanesulfonic acid released a single polypeptide with a molecular size of 23 kilodaltons, accounting for all of the insoluble cell wall protein. Extensive purification and then analysis of the 23-kilodalton protein demonstrated the absence of diaminopimelic acid, muramic acid, or other peptidoglycan components, pointing to either a novel linkage between protein and peptidoglycan or a noncovalent but tenacious association. The released 23-kilodalton protein showed amino acid homology and other similarities to the outer membrane protein OmpF of Escherichia coli. Although a similar product was released in small quantities from cell walls of the virulent M. tuberculosis Erdman and H37Rv by lysozyme treatment, the cell walls of virulent bacilli were dominated by the presence of poly-alpha-L-glutamine, accounting for as much as 10% of their weight. The poly-alpha-L-glutamine was successfully separated from the cell wall proper, demonstrating again the absence of a covalent association between peptidoglycan and the polymer. The antigenicity of these products is demonstrated, and their roles vis-a-vis analogous polypeptides from other bacteria in immunogenicity, pathogenicity, and bacterial physiology are discussed.

Deglycosylation of neutral and acidic human colonic mucin

Human colonic mucin has been isolated from normal colonic mucosa by a phenol-water extraction procedure and purified by Sepharose 2B column chromatography. The mucin was further purified by cesium bromide density gradient centrifugation. Sodium dodecyl sulfate-polyacrylamide gel (5%) electrophoresis of this material showed high-molecular-weight mucin component(s) at the top of the gel. Chemical analyses of this preparation indicated a typical mucin profile of amino acids and carbohydrates. Ion-exchange chromatography resulted in the separation of two major fractions, one being more acidic than the other. Chemical deglycosylation of the purified preparation at 20 degrees C for 3 1/2 showed loss of sialic acid, fucose, galactose, and N-acetylglucosamine, whereas traces of N-acetylgalactosamine were still detected. High-pressure liquid chromatography of the deglycosylated material resulted in the purification of a major peptide, P1, with high levels of threonine, serine, and proline, resembling, in most respects, the profile of native mucin. The molecular weight of the peptide was determined to be approximately 97 kDa and serine was the single NH2 terminus.

Enzymatic generation of triplet acetone by deglycosylated horseradish peroxidase

A study of the effects of deglycosylation of horseradish peroxidase on protein conformation, as well as on its catalytic activity of oxidation of isobutyraldehyde or its enol form to triplet acetone and formic acid, was performed. The loss of carbohydrates leads to structural modifications of this enzyme. This is confirmed by a change in the circular dichroism spectrum, an increase in tryptophan's environment polarity, and a loss of the chiral specificity toward D- and L-tryptophan. Deglycosylation does not destroy either the peptide backbone or the amino acid residues and does not affect the heme group content of the protein. The rates of oxygen uptake and light emission observed when horseradish peroxidase oxidizes isobutyraldehyde or the trimethylsilyl enol ether form of the latter are reduced when the enzyme is 70% deglycosylated. Concomitantly, the acting deglycosylated enzyme becomes inactivated during the course of the reaction. It appears that the carbohydrate moiety plays an important role in the protection of the peroxidase from damaging effects induced by triplet acetone and in the stabilization of the three-dimensional structure of this enzyme.

Establishment and maintenance of a regionalized glycoprotein distribution during early mouse development

The regionalization of the cell membranes of the mouse embryo into apical and basolateral zones has been studied using antibodies to a pair of glycoproteins expressed during the two-cell to early blastocyst stage. These antigens are found on the outer, free surface and in the underlying cortical cytoplasm, but are not detectable at areas of cell contact. In the early blastocyst stage, antigen also appears at the free surfaces of cells bordering the blastocoel. Antigen regionalization is also reestablished after experimental manipulation and appears to be a direct consequence of cell contact. Thus, blastomeres examined 4 hr after dissociation from four- and eight-cell stage embryos express antigen in cortical areas underlying newly exposed surfaces and new sites of contact between embryos in multiple-embryo aggregates lose detectable antigen within 2 to 4 hr of the formation of the contacts. Microfilaments are involved in controlling the regional expression of these glycoproteins. Incubation of embryos from the two-cell stage in medium containing cytochalasin B interferes with antigen targeting, resulting in abnormal expression of the antigens both on the surface and in the cytoplasm of the embryos. Cytochalasin B treatment of later stage embryos results in an uneven distribution of the antigen in cortical cytoplasm and prevents the complete removal of antigen from new sites of cell contact in multiple-embryo aggregates. The presence of nocodozole, which inhibits the polymerization of microtubules, had no detectable effect on the expression of the antigens. Interference with the glycosylation of these proteins, by incubation of embryos in the presence of tunicamycin, did not alter the regionalized pattern of expression.

Comparison of heparan sulfate proteoglycans from equine and human glomerular basement membranes

1. Proteoglycans extracted from human and equine glomerular basement membranes (GBM) were purified by ion-exchange chromatography and gel filtration. 2. The glycoconjugates had an apparent molecular mass of 200-400 kDa and consisted of 75% protein and 25% glycosaminoglycan. Glycosidase and HNO2 treatment and the amino sugar and sulfate composition of both proteoglycan preparations identified heparan sulfate (HS) as the predominant saccharide chain. 3. Hydrolysis with trifluoromethanesulfonic acid yielded comparable core proteins with molecular masses of ca 160 and 120 kDa. 4. The HS chains had an apparent molecular mass of 18 kDa. Results of heparitinase digestion and HNO2-treatment indicated a clustering of sulfate groups in the distal part of the HS side chains. 5. Peptide mapping after trypsin, clostripain or V8 protease digestion of radiolabeled human and equine heparan sulfate proteoglycans (HSPG) preparations with three different separation techniques showed large differences. 6. Polyclonal antisera raised against the HSPGs reacted against the core proteins. Both HSPG preparations and their antisera showed ca 40% cross-reactivity. About 50% of monoclonal antisera elicited against one HSPG preparation showed reaction with both HSPG preparations. 7. Polyclonal antisera stained all basement membranes in an intense linear fashion in indirect immunofluorescence studies of kidney sections from horse, man and various mammalian species. 8. Biochemical and immunological data indicate that HSPGs from equine and human GBM have a comparable structure, but the core proteins differ considerably.

Purification and characterization of a heterogeneous glycosylated invertase from the rumen holotrich ciliate Isotricha prostoma

The invertase (beta-fructofuranosidase, EC 3.2.1.26) of the rumen holotrich ciliate Isotricha prostoma has been purified. This is the first report of an enzyme purification from a known species of rumen protozoon. Cells were disrupted by ultrasonic treatment and the enzyme was purified from the cell-free extract by three successive liquid column chromatographies (Sepharose CL4B/octyl-Sepharose CL4B, DE52 DEAE-cellulose and concanavalin A-Sepharose 4B). This resulted in a 160-fold purification and a 15% yield. The major form of the purified enzyme was a tetramer with Mr about 350,000 that was readily dissociated by electrophoresis. The invertase was heterogeneous, as five types of monomers were shown by SDS/polyacrylamide-gel electrophoresis after denaturation. Part of this heterogeneity was due to different glycosylated forms of one of the polypeptides present in the purified enzyme. Isotricha prostoma invertase exhibited maximum activity at pH 5.5-6.0 and 50 degrees C. The kinetic properties of the purified enzyme were very similar to those of invertases from other sources such as yeast or plants (substrate and product inhibition, transferase activity).

Characterization of the hydroxyproline-rich protein core of an arabinogalactan-protein secreted from suspension-cultured Lolium multiflorum (Italian ryegrass) endosperm cells

An arabinogalactan-protein (AGP) purified from the filtrate of liquid-suspension-cultured Italian-ryegrass (Lolium multiflorum) endosperm cells by affinity chromatography on myeloma protein J539-Sepharose was deglycosylated with trifluoromethanesulphonic acid to remove polysaccharide chains that are covalently associated with hydroxyproline residues in the peptide component of the proteoglycan. The protein core, which accounts for less than 10% (w/w) of the intact proteoglycan, was purified by h.p.l.c. It has an apparent Mr of 35,000, but reacts very poorly with both Coomassie Brilliant Blue R and silver stains. Amino-acid-sequence analysis of the N-terminus of the h.p.l.c.-purified protein core and of tryptic peptides generated from the unpurified protein reveals a high content of hydroxyproline and alanine. These are sometimes arranged in short (Ala-Hyp) repeat sequences of up to six residues. Polyclonal antibodies raised against the protein core do not cross-react with native AGP, the synthetic peptide (Ala-Hyp)4, poly-L-hydroxyproline or poly-L-proline. The results suggest that the polysaccharide chains in the native AGP render the protein core of the proteoglycan inaccessible to the antibodies and that the immunodominant epitopes include domains of the protein other than those rich in Ala-Hyp repeating units.

Isolation and partial characterization of heparan sulphate proteoglycan from the human glomerular basement membrane

Heparan sulphate proteoglycan was solubilized from human glomerular basement membranes by guanidine extraction and purified by ion-exchange chromatography and gel filtration. The yield of proteoglycan was approx. 2 mg/g of basement membrane. The glycoconjugate had an apparent molecular mass of 200-400 kDa and consisted of about 75% protein and 25% heparan sulphate. The amino acid composition was characterized by a high content of glycine, proline, alanine and glutamic acid. Hydrolysis with trifluoromethanesulphonic acid yielded core proteins of 160 and 110 kDa (and minor bands of 90 and 60 kDa). Alkaline NaBH4 treatment of the proteoglycan released heparan sulphate chains with an average molecular mass of 18 kDa. HNO2 oxidation of these chains yielded oligosaccharides of about 5 kDa, whereas heparitinase digestion resulted in a more complete degradation. The data suggest a clustering of N-sulphate groups in the peripheral regions of the glycosaminoglycan chains. A polyclonal antiserum raised against the intact proteoglycan showed reactivity against the core protein. It stained all basement membranes in an intense linear fashion in immunohistochemical studies on frozen kidney sections from man and various mammalian species.

Low-density lipoprotein binding affinity of arterial wall proteoglycans: characteristics of a chondroitin sulfate proteoglycan subfraction

The characteristics of an arterial wall chondroitin sulfate proteoglycan (CS-PG) subfraction that binds avidly to low-density lipoproteins (LDL) was studied. A large CS-PG was extracted from bovine aorta intima-media under dissociative conditions, purified by density-gradient centrifugation and gel filtration chromatography, and further subfractionated by affinity chromatography on LDL-agarose. A proteoglycan subfraction, representing 25% of the CS-PG, showed an elution profile (with dissociation from LDL-agarose occurring between 0.5 and 1.0 M NaCl) corresponding to that of heparin, heretofore considered to be the most strongly binding glycosaminoglycan with LDL. The proteoglycan subfraction which migrated as a single band on composite agarose-polyacrylamide gel electrophoresis contained chondroitin 6-sulfate, chondroitin 4-sulfate and dermatan sulfate in a proportion of 70:22:8. The core protein of the proteoglycan had an apparent molecular weight of 245,000, and contained approx. 33 glycosaminoglycan chains with an average molecular weight of 32,000. The CS-PG subfraction, like heparin, formed insoluble complexes in the presence of 30 mM Ca2+. Complexing of LDL with proteoglycan resulted in two classes of interactions with 0.1 and 0.3 proteoglycan monomer bound per LDL particle characterized by an apparent Kd of 4 and 21 nM, respectively. This indicates that multiple LDL particles bind to single proteoglycan monomers even at saturation. In contrast, LDL-heparin interactions showed a major component characterized by an apparent Kd of 151 nM and a Bmax of 9 heparin molecules per LDL particle. The occurrence of a potent LDL-binding proteoglycan subfraction within the family of arterial CS-PG may be of importance in terms of lipid accumulation in atherogenesis.

Drosophila neuroglian: a member of the immunoglobulin superfamily with extensive homology to the vertebrate neural adhesion molecule L1

Drosophila neuroglian is an integral membrane glycoprotein that is expressed on a variety of cell types in the Drosophila embryo, including expression on a large subset of glial and neuronal cell bodies in the central and peripheral nervous systems and on the fasciculating axons that extend along them. Neuroglian cDNA clones were isolated by expression cloning. cDNA sequence analysis reveals that neuroglian is a member of the immunoglobulin superfamily. The extracellular portion of the protein consists of six immunoglobulin C2-type domains followed by five fibronectin type III domains. Neuroglian is closely related to the immunoglobulin-like vertebrate neural adhesion molecules and, among them, shows most extensive homology to mouse L1. Its homology to L1 and its embryonic localization suggest that neuroglian may play a role in neural and glial cell adhesion in the developing Drosophila embryo. We report here on the identification of a lethal mutation in the neuroglian gene.

Increased proteoglycan synthesis following the differentiation of F9 embryonal carcinoma cells: formation of a differentiation-specific proteoheparan sulfate

We have examined changes in proteoglycan synthesis by F9 embryonal carcinoma cells after the cells have been treated with retinoic acid or retinoic acid plus cholera toxin. Retinoic acid is known to stimulate the differentiation of this cell type to a primitive endoderm-like cell characterized by the production of basement membrane components such as type IV collagen, laminin and proteoglycans. We have now demonstrated that proteoglycan synthesis and secretion were further stimulated when cholera toxin was added in addition to retinoic acid. Moreover, media of these fully differentiated cells was found to contain a different species of proteoheparan sulfate not produced by stem cells or retinoic acid-treated cells. This proteoheparan sulfate had a high density upon CsCl gradient centrifugation. The protein core of this proteoheparan sulfate was estimated by SDS gel electrophoresis to be approximately 15,000 daltons.

Secondary structure of a core protein from pig skin proteodermatan sulfate: CD and Fourier transform IR spectroscopic studies in solution

The secondary structure of a 38 kDa core protein from pig skin proteodermatan sulfate (PDS), was investigated in solution using CD and Fourier transform (FT) ir spectroscopy. Both techniques generally have provided complementary data on the secondary structures of proteins. CD spectral analysis has shown that the core protein contains 60% beta-turn and alpha-helical structures, the rest being "unordered" structure. FT ir data do not permit calculation of quantitative contributions of substructures, at the present time, to the overall secondary structure of the core protein. CD spectrum of the intact PDS is similar to the core protein CD spectrum.

Chemical deglycosylation on a micro-scale of membrane glycoproteins with retention of phosphoryl-protein linkages

We provide here an easy and convenient procedure for the chemical deglycosylation of membrane glycoproteins that can be done rapidly and inexpensively on large numbers of membrane samples on a microscale without prior isolation of the glycoprotein in question. We have used the procedure successfully on three different glycoproteins and additionally have shown that covalent linkages of the polypeptide to phosphoryl residues are preserved.

S-layer protein gene of Acetogenium kivui: cloning and expression in Escherichia coli and determination of the nucleotide sequence

Acetogenium kivui is anaerobically growing thermophilic bacterium with a gram-positive type of cell wall structure. The outer surface is covered with a hexagonally packed surface (S) layer. The gene coding for the S-layer polypeptide was cloned in Escherichia coli on two overlapping fragments by using the plasmid pUC18 as the vector. It was expressed under control of a cloned Acetogenium promoter or the lacZ gene. We determined the complete sequence of the structural gene. The mature polypeptide comprises 736 amino acids and is preceded by a typical procaryotic signal sequence of 26 amino acids. It i weakly acidic, weakly hydrophilic, and contains a relatively high proportion of hydroxyamino acids, including two clusters of serine and threonine residues. An N-terminal region of about 200 residues is homologous to the N-terminal part of the middle wall protein, one of the two S-layer proteins of Bacillus brevis, and there is also an internal homology within the N-terminal region of the A. kivui polypeptide.

A lectin gene encodes the alpha-amylase inhibitor of the common bean

An alpha-amylase inhibitor that inhibits insect and mammalian alpha-amylases but not plant alpha-amylases, is present in seeds of the common bean (Phaseolus vulgaris). We have purified the alpha-amylase inhibitor by using a selective heat treatment in acidic medium and affinity chromatography with porcine pancreas alpha-amylase coupled to agarose. Under sodium dodecyl sulfate gel electrophoresis, the purified inhibitor gave rise to five bands with mobilities corresponding to molecular masses ranging from 14 to 19 kDa. N-terminal sequencing (up to 15 amino acids) of the polypeptides obtained from these bands resulted in only two different sequences matching two stretches of the amino acid sequence deduced from an already described lectin gene [Hoffman, L. M. (1984) J. Mol. Appl. Gen. 2,447-453]. This gene is different from but closely related to the genes that code for phytohemagglutinin, the major lectin of bean. Further evidence based on amino acid composition, identification of a precursor, and recognition of the product of the gene (expressed in Escherichia coli) by an anti-alpha-amylase inhibitor serum confirms that the inhibitor is encoded by this or a closely related lectin gene. This finding assigns a biological function, which has been described at the molecular level, to a plant lectin gene product and supports the defense role postulated for seed lectins. The lack of homology with other families of enzyme inhibitors suggests that this may be the first member of a new family of plant enzyme inhibitors.

Asparagine-linked glycosylation of the scrapie and cellular prion proteins

Post-translational modification of the scrapie prion protein (PrP) is thought to account for the unusual features of this protein. Molecular cloning of a PrP cDNA identified two potential Asn-linked glycosylation sites. Both the scrapie (PrPSc) and cellular (PrPC) isoforms were susceptible to digestion by peptide N-glycosidase F (PNGase F) but resistant to endoglycosidase H as measured by migration in sodium dodecyl sulfate-polyacrylamide gel electrophoresis. PNGase F digestion of PrPC yielded two proteins of Mr26K and 28K; however, the 26-k species was only a minor component. In contrast, PNGase F digestion of PrPSc yielded equimolar amounts of two proteins of Mr26K and 28K. The significance of this altered stoichiometry between the 26- and 28-kDa deglycosylated forms of PrP during scrapie infection remains to be established. Both isoforms as well as PrP 27-30, which is produced by limited proteolysis of PrPSc, exhibited a reduced number of charge isomers after PNGase F digestion. The molecular weight of PrP 27-30 was reduced from 27K-30K by PNGase F digestion to 20K-22K while anhydrous hydrogen fluoride or trifluoromethanesulfonic acid treatment reduced the molecular weight to 19K-21K and 20K-22K, respectively. Denatured PrP 27-30 was radioiodinated and then assessed for its binding to lectin columns. PrP 27-30 was bound to wheat germ agglutinin (WGA) or lentil lectins and eluted with N-acetylglucosamine or alpha-methyl-mannoside, respectively. Digestion of PrP 27-30 with sialidase prevented its binding to WGA but enhanced its binding to Ricinus communis lectin. These findings argue that PrP 27-30 probably possesses Asn-linked, complex oligosaccharides with terminal sialic acids, penultimate galactoses, and fucose residues attached to the innermost N-acetyl-glucosamine. Whether differences in Asn-linked oligosaccharide structure between PrPC and PrPSc exist and are responsible for the distinct properties displayed by these two isoforms remain to be established.

Targeting and glycosylation of patatin the major potato tuber protein in leaves of transgenic tobacco

Patatin, the most abundant protein in the storage parenchyma cells of potato (Solanum tuberosum L.) tubers, is a vacuolar glycoprotein that consists of a number of closely related polypeptides and is encoded by a large gene family. To analyse the glycosylation pattern and the nature of the glycans on a single patatin polypeptide in a heterologous tissue we introduced a single chimaeric patatin gene into tobacco (Nicotiana tabacum L.) and studied its product in leaves. Patatin isolated from the leaves of transgenic tobacco plants is glycosylated at asparagine (Asn)(60), and Asn(90), but the third glycosylation site (Asn(202)) has no glycan. The two glycans are typical small complex glycans with xylose, fucose, mannose and N-acetylglucosamine in a ratio 1:1:3:2, the same ratio as found on patatin isolated from potato tubers. Expression of patatin in tobacco leaves was accompanied by the correct processing of the signal peptide, and the proper targeting of the glyco-protein to the vacuoles of mesophyll cells.

Reduced cell adhesion to fibronectin and laminin is associated with altered glycosylation of beta 1 integrins in a weakly metastatic glycosylation mutant

A weakly metastatic wheat-germ-agglutinin-resistant mutant Wa4-b1 was previously shown to be less adherent to endothelial cell extracellular matrix than the more metastatic parental B-16 melanoma cells. This report describes reduced adhesion and spreading of Wa4-b1 cells on the cell-binding domain of fibronectin (CBD) and laminin (LN). Cell surface receptors which mediate such interactions are members of the integrin family of membrane glycoproteins. An antibody that recognizes the beta 1 integrin subunit inhibited spreading on both the CBD and LN. The integrins of the mutant cells immunoprecipitated by the antibody appeared to be structurally altered, showing a greater electrophoretic mobility. The mobility difference between the parent and the mutant receptors was abolished following removal of the glycan moieties of the receptors enzymatically using glycopeptidase F, or chemically using trifluoromethanesulfonic acid, suggesting that the structural alteration of the mutant receptors is in glycosylation. The altered receptors may be responsible for the observed decrease in cell adhesion and spreading of the mutant cells to the CBD and LN. Such a decrease in Wa4-b1 cell interaction with extracellular matrix components may play a role in their decreased metastatic potential.

Serologic mapping and biochemical characterization of the carcinoembryonic antigen epitopes using fourteen distinct monoclonal antibodies

A series of 14 monoclonal anti-carcinoembryonic antigen (CEA, Mr 180,000) antibodies (MAbs) that show a strong degree of selective reactivity for human colon carcinomas versus normal adult tissues were used to construct a serological map of the CEA molecule. The MAbs were generated using extracts of colon carcinomas as immunogen and are thus given a COL designation. None of the 14 COL-MAbs tested were reactive with purified non-specific cross-reacting antigen (NCA, Mr 55,000) from normal lung, although some showed reactivity to human granulocytes. All the COL-MAbs tested were reactive with normal fecal antigen-2 (NFA-2, Mr 170,000); however, many of the COL-MAbs demonstrated a higher affinity constant to CEA than to NFA-2. Cross-competition radioimmunoassays classified the 14 COL-MAbs into 5 groups. The chemical nature of the COL-binding domains was tested using chemically or enzymatically treated CEA; all reacted with periodate-treated CEA and deglycosylated CEA, indicating that the COL-reactive epitopes appear to be of a proteinaceous nature. Heat treatment, reduction, alkylation, pepsin digestion or pronase treatment of CEA, however, gave differential results with respect to COL binding. Antibody titration experiments were carried out to define differential reactivities to colorectal carcinoma versus NCA-containing granulocyte extracts; these results were compared with results obtained using several anti-CEA MAbs that have been used in clinical trials. Granulocyte binding and biochemical studies showed that the COL MAbs may distinguish as many as 7 to 10 CEA epitopes.

The biochemical characterization of E2, a T cell surface molecule involved in rosettes

We have previously identified a molecule on the T cell surface, which, in addition to CD2 is involved in the rosette phenomenon. This is a 32-kDa single polypeptide chain which we have termed E2. The studies reported here show striking patterns on the glycosylation status of E2. It is a heavily sialylated and glycosylated molecule, the sugar moieties accounting for almost half of its relative molecular mass (Mr). It carries no N-linked sugar residues, only O-linked oligosaccharides. Despite heavy glycosylation, the molecule appears to behave homogeneously on gel electrophoresis, both in terms of Mr and pI. Neuraminidase treatment of E2 lowered its Mr to 28 kDa; this was further decreased to 18 kDa after removal of O-linked sugar residues by treatment with O-glycanase. An identical reduction in size was observed after treatment with trifluoromethane sulfonic acid, showing that the molecule carries no detectable N-linked sugar residues. Moreover, endoglycosidase F and endoglycosidase H treatment of either the immunoprecipitates from 125I surface-labeled thymocytes, or of a purified preparation of E2, did not reduce the Mr of E2, nor did tunicamycin treatment of T cells. Two-dimensional gel electrophoresis revealed two discrete spots of acidic pI (4.4 and 4.6) that were still seen after neuraminidase treatment, though they had moderately shifted. Pulse-chase experiments revealed a single 28-kDa precursor form that could have been the unsialylated molecule. Finally, sequencing 14 amino acid residues of the N-terminal side revealed no homology with known proteins. Since the sugar moieties of adhesion protein could play an important role, the results obtained in this study will prove valuable to our understanding of the role exerted by the E2 molecule.