O-linked Carbohydrate Chains Research Articles

Inhibition of selectin-dependent tumor cell adhesion to endothelial cells and platelets by blocking O-glycosylation of these cells

Expression of sialosyl-Le x (SLe x) and sialosyl-Le a (SLe a) on tumor cell lines HL60, Colo205, and U937 was greatly suppressed by application of benzyl-α-GalNAc for inhibition of O-linked carbohydrate chain extension, which resulted in reduced adhesion of tumor cells to activated endothelial cells or platelets mediated by ELAM-1 (E-selectin) or GMP-140 (P-selectin). Inhibitors or modifiers of N-glycosylation had no effect on expression of SLe x or SLe a in these tumor cells. These findings suggest the possibility that targeting of O-glycosylation inhibitors or modifiers to tumor cells may effectively suppress metastatic potential.

Studies on the structure and function of IL-6 and its receptor

Interleukins area group of signalling molecules involved in the communication between cells. They are produced and secreted by many different cell types, particularly by those of the immune system. Interleukin-6 (IL-6) 1 elicits a wide spectrum of biological functions. In general, IL-6 represents a growth and differentiation factor acting on B-cells, T-cells, lymphoma cells, hybridoma/plasmacytoma cells, hematopoietic stem cells, and hepatocytes (for review see [1]). IL-6 consists of 184 amino acids. It is synthesized as a larger polypeptide precursor with a signal peptide of 28 amino acids. Studies on the structures of Nand O-linked carbohydrate chains and their location within the polypeptide backbone of human IL-6 synthesized and ;ecreted by NIH/3T3 cells stably transfected with a human IL-6-cDNA have shown that only one (Asn 45) of the two potential N-glycosylation sites carries an oligosaccharide side chain. Thr 138 was identified as Oglycosylation site. At the cDNA level we constructed deletion mutants of human IL-6, lacking one, two, three or four amino acids from the C-terminus. In the mouse B9 cell proliferation assay as well as in the 7-fibrinogen induction assay with the human hepatoma cell line HepG2 the removal of the last amino acid (Met 184) resulted in the loss of 80% of biological activity. When three amino acids were removed from the C-terminus, IL-6 activity was completely abbrogated. Studies with various point mutations at the C-terminus of IL-6 indicated that an c~helical structure and a positive charge (Arg 182) are essential for biological activity [2-4]. In order to define domains of interleukin-6 responsible for species specific receptor recognition, we have constructed a whole series of chimeric human/mouse interleukin-6 proteins. This approach is based on the fact that human IL-6 is active on human and murine cells whereas murine IL-6 is only active on murine cells. The proteins have been expressed in E. coli, purified to homogeneity and tested for bioactivity on human and murine cells. It turned out that a region in the NH2-terminal part of IL-6 (loop between helix A and B or helix B) as well as the carboxyterminal helix D (nomenclature taken from the threedimensional structure prediction for IL-6 of Bazan [5]) is impor-

Salivary Gland Antigens of Ixodes dammini Are Glycoproteins That Have Interspecies Cross-Reactivity

Serum from rabbits and rats exposed to Ixodes dammini adults and larvae, respectively, contain antibodies to a large number of antigens in salivary gland homogenates from adult ticks that cross react with the antigens of a closely related species, Ixodes scapularis, and significantly with Dermacentor variabilis antigens. The salivary gland antigens of I. dammini are glycoproteins composed of both N- and O-linked carbohydrate chains. The antigenic determinants reside in both the polypeptide and carbohydrate chains.

Altered O-glycan synthesis in lymphocytes from patients with Wiskott-Aldrich syndrome.

The only molecular defect reported for the X-linked immunodeficiency Wiskott-Aldrich syndrome (WAS) is the abnormal electrophoretic behavior of the major T lymphocyte sialoglycoprotein CD43. Since the 70 to 80 O-linked carbohydrate chains of CD43 are known to influence markedly its electrophoretic mobility, we analyzed the structure and the biosynthesis of O-glycans of CD43 in lymphocytes from patients with WAS. Immunofluorescence analysis with the carbohydrate dependent anti-CD43 antibody T305 revealed that in 10 out of the 12 WAS patients tested increased numbers of T lymphocytes carry on CD43 an epitope which on normal lymphocytes is expressed only after activation. Other activation antigens were absent from WAS lymphocytes. Western blots of WAS cell lysates displayed a high molecular mass form of CD43 which reacted with the T305 antibody and which could be found on in vivo activated lymphocytes but was absent from normal unstimulated lymphocytes. To examine the O-glycan structures, carbohydrate labeled CD43 was immunoprecipitated and the released oligosaccharides identified. WAS lymphocyte CD43 was found to carry predominantly the branched structure NeuNAc alpha 2----3Gal beta 1----3 (NeuNAc alpha 2----3Gal beta 1----4G1cNAc beta 1----6) GalNAcOH whereas normal lymphocytes carry the structure NeuNAc alpha 2----3Gal beta 1----3 (NeuNAc alpha 2----6) GalNAcOH. Only after activation NeuNAc alpha 2----3Gal beta 1----3 (NeuNAc alpha 2----3Gal beta 1----4GlcNAc beta 1----6) GalNAcOH becomes the principal oligosaccharide on CD43 from normal lymphocytes. Analyzing the six glycosyltransferases involved in the biosynthesis of these O-glycan structures it was found that in WAS lymphocytes high levels of beta 1----6 N-acetyl-glucosaminyl transferase are responsible for the expression of NeuNAc alpha 2----3Gal beta 1----3 (NeuNAc alpha 2----3Gal beta 1----4GlcNAc beta 1----6) GalNAcOH on CD43. The gene responsible for WAS has not yet been identified but the results presented in this study suggest that the primary defect in WAS may affect a gene which is involved in the regulation of O-glycosylation.

Recombinant carbohydrate and selenomethionyl variants of human choriogonadotropin

Recombinant human choriogonadotropin and selenomethionyl human choriogonadotropin (rhCG and SehCG) were expressed in baculovirus expression system by coinfection of SF9 insect cells by recombinant viruses, AcMNPV-hCG alpha and AcMNPV-hCG beta containing hCG alpha and hCG beta cDNAs. The expression efficiency of both rhCG and SehCG was quite high. The association of the alpha and beta subunits into a dimer was apparently complete since no detectable amount of rhCG beta was found in the rhCG eluate from the monoclonal hCG beta antibody immunoaffinity column. Both rhCG and SehCG preparations were homogeneous as indicated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and reverse-phase high performance liquid chromatography. The apparent molecular mass of rhCG and SehCG on sodium dodecyl sulfate-polyacrylamide gel electrophoresis under nonreducing conditions was about 38 kDa while under reducing conditions the heterodimer dissociated to yield beta and alpha subunits with molecular masses of 22.5 and 18 kDa, respectively. The carbohydrate analysis of rhCG showing the presence of 2.1, 3.3, 7.38, 4.2, and 27.8 residues of Fuc, GalNAC, GlcNAC, Gal, and Man, respectively, per mole of the hormone was consistent with the presence of 4 N-linked high mannose type carbohydrate hydrate and 4 O-linked simple carbohydrate chains, probably made up of Gal-GalNAC. Despite the altered glycosylation, rhCG demonstrated close similarity to the native urinary hCG in amino acid composition, receptor binding, and in its ability to stimulate cAMP and steroidogenesis. This indicates that there is no specificity of carbohydrate required for biological activity. Furthermore, it implies that the alteration from the complex to high mannose type carbohydrates in rhCG does not affect its proper folding. Finally, amino acid analysis of SehCG showed that 84% of methionine residues in rhCG were replaced by selenomethionine.

Carbohydrate variant of the recombinant beta-subunit of human choriogonadotropin expressed in baculovirus expression system

The beta-subunit of human choriogonadotropin (hCG) has two complex type N-linked and four O-linked carbohydrate chains. To further evaluate the specificity of the carbohydrate moiety on the hCG function, we have expressed hCG beta subunit in the baculovirus insect cell system to modify its carbohydrate structures. The recombinant hCG beta (rhCG beta) was efficiently secreted in the medium and was purified to homogeneity by immunoaffinity chromatography using a highly specific monoclonal antibody against hCG beta. The homogeneity of the recombinant subunit was established by sodium dodecyl sulfate-polyacrylamide gel electrophoresis performed under reducing and nonreducing conditions and reverse phase high performance liquid chromatography. rhCG beta had molecular weights of 22,500 and 33,000 under reducing and nonreducing conditions, respectively. Digestion with N-glycanase cleaved the Mr = 22,500 protein to 18,000, while digestion with Endo H or Endo F yielded an additional protein band of 20,500 on sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The carbohydrate analysis by pulse amperometry yielded the relative number of 2.5, 2.4, 3.7, and 11.3 residues of fucose, N-acetylgalactosamine, galactose, and mannose, respectively, based on a value of 4 residues for N-acetylglucosamine. Lectin binding studies showed rhCG beta to bind with concanavalin A with a high affinity and not with wheat germ agglutinin. In the studies with endoglycosidases together with the carbohydrate analysis and lectin binding properties, rhCG beta appears to have two high mannose-type N-linked and three to four O-linked carbohydrate simple disaccharide chains. The carbohydrate modification of the beta-subunit did not alter its immunopotency and its ability to combine with hCG alpha. The reconstituted hormone made up of rhCG beta and hCG alpha was found to be similar to hCG in biological properties such as receptor binding and in its ability to stimulate cAMP and steroidogenesis.

Proclotting enzyme from horseshoe crab hemocytes. cDNA cloning, disulfide locations, and subcellular localization.

Proclotting enzyme is an intracellular serine protease zymogen closely associated with an endotoxin-sensitive hemolymph coagulation system in limulus. Its active form, clotting enzyme, catalyzes conversion of coagulogen to insoluble coagulin gel. We present here the cDNA and amino acid sequences, disulfide locations, and subcellular localization of proclotting enzyme. The isolated cDNA for proclotting enzyme consists of 1,501 base pairs. The open reading frame of 1,125 base pairs encodes a sequence comprising 29 amino acid residues of prepro-sequence and 346 residues of the mature protein with a molecular mass of 38,194 Da. Three potential glycosylation sites for N-linked carbohydrate chains were confirmed to be glycosylated. Moreover, the zymogen contains six O-linked carbohydrate chains in the amino-terminal light chain generated after activation. The cleavage site that accompanies activation catalyzed by trypsin-like active factor B, proved to be an Arg-Ile bond. The resulting carboxyl-terminal heavy chain is composed of a typical serine protease domain, with a sequence similar to that of human coagulation factor XIa (34.5%) or factor Xa (34.1%). The light chain has a unique disulfide-knotted domain which shows no significant homology with any other known proteins. Thus, this proclotting enzyme has a mammalian serine protease domain and a structural domain not heretofore identified in coagulation and complement factors. Immunohistochemical studies showed that the proclotting enzyme is localized in large granules of hemocytes.

Identification of a novel glycoprotein (AGp110) involved in interactions of rat liver parenchymal cells with fibronectin.

We have identified an integral membrane glycoprotein in rat liver that mediates adhesion of cultured hepatocytes on fibronectin substrata. The protein was isolated by affinity chromatography of detergent extracts on wheat germ lectin-Agarose followed by chromatography of the WGA binding fraction on fibronectin-Sepharose. The glycoprotein (AGp110), eluted at high salt concentrations from the fibronectin column, has a molecular mass of 110 kD and a pI of 4.2. Binding of immobilized AGp110 to soluble rat plasma fibronectin required Ca2+ ions but was not inhibited by RGD peptides. Fab' fragments of immunoglobulins raised in rabbits against AGp110 reversed the spreading of primary hepatocytes attached onto fibronectin-coated substrata, but had no effect on cells spread on type IV collagen or laminin substrata. The effect of the antiserum on cell spreading was reversible. AGp110 was detected by immunofluorescence around the periphery of the ventral surface of substratum attached hepatocytes, and scattered on the dorsal surface. Immunohistochemical evidence and Western blotting of fractionated liver plasma membranes indicated a bile canalicular (apical) localization of AGp110 in the liver parenchyma. Expression of AGp110 is tissue specific: it was found mainly in liver, kidney, pancreas, and small intestine but was not detected in stomach, skeletal muscle, heart, and large intestine. AGp110 could be labeled by lactoperoxidase-catalyzed surface iodination of intact liver cells and, after phase partitioning of liver plasma membranes with the detergent Triton X-114, it was preferentially distributed in the hydrophobic phase. Treatment with glycosidases indicated extensive sialic acid substitution in at least 10 O-linked carbohydrate chains and 1-2 N-linked glycans. Immunological comparisons suggest that AGp110, the integrin fibronectin receptor and dipeptidyl peptidase IV, an enzyme involved in fibronectin-mediated adhesion of hepatocytes on collagen, are distinct proteins.

Subcellular distribution of acetylcholinesterase forms in chromaffin cells

The presence of acetylcholinesterase (AChE) in chromaffin granules has been controversial for a long time. We therefore undertook a study of AChE molecular forms in chromaffin cells and of their distribution during subcellular fractionation. We characterized four main AChE forms, three amphiphilic forms (Ga1, Ga2 and Ga4), and one non-amphiphilic form (Gna4). Each form shows the same molecular characteristics (sedimentation, electrophoretic migration, lectin interactions) in the different subcellular fractions. All forms are glycosylated and seem to possess both N-linked and O-linked carbohydrate chains. There are differences in the structure of the glycans carried by the different forms, as indicated by their interaction with some lectins. Glycophosphatidylinositol-specific phospholipases C converted the Ga2 form, but not the other amphiphilic forms, into non-amphiphilic derivatives. The distinct patterns of AChE molecular forms observed in various subcellular compartments indicate the existence of an active sorting process. Gna4 was concentrated in fractions of high density, containing chromaffin granules. We obtained evidence for the existence of a lighter fraction also containing chromogranin A, tetrabenazine-binding sites and Gna4 AChE, which may correspond to immature, incompletely loaded granules or to partially emptied granules. The distribution of Gna4 during subcellular fractionation suggested that this form is largely, but not exclusively, contained in chromaffin granules, the membranes of which may contain low levels of the three amphiphilic forms.

Structure determination of the major N‐ and O‐linked carbohydrate chains of the β subunit from equine chorionic gondotropin

The carbohydrate moieties of equine chorionic gonadotropin alpha and beta subunits were released from the protein backbones by successive treatments with peptide-N4-(N-acetyl-beta-glucosaminyl)asparagine amidase F and alkaline borohydride and then fractionated by FPLC and HPLC. The major N- and O-linked glycans of the beta subunit were characterized by 500-MHz 1H-NMR spectroscopy, showing a remarkable structural heterogeneity for the N-glycosidically linked chains, comprising mono-, di-, tri- and tri'-antennary N-acetyllactosamine type of glycans, being partly alpha 1-6 fucosylated at the Asn-bound GlcNAc residue and having alpha 2-6 and alpha 2-3 linked N-acetyl- and N-acetyl-4-O-acetylneuraminic acid residues as sialic acid constituents. Significant differences in this respect were detected for the partially characterized glycans of the alpha subunit. The major part of the O-linked carbohydrate chains, occurring solely in the beta subunit, is formed by tri-, tetra-, penta- and hexa-saccharides. There are indications for the presence of oligo(N-acetyllactosamine) units in both the N- and O-linked glycans of the beta subunit.

Purification and Molecular Characterization of FAP, a Feto-acinar Protein Associated with the Differentiation of Human Pancreas

This work describes the purification of FAP, a feto-acinar pancreatic protein associated with the ontogenesis, differentiation, and transformation of the human exocrine pancreas. The protein was purified to homogeneity from pancreatic juices of patients with pancreatic pathology by a two-step chromatographic procedure which consisted of size exclusion on Sephacryl S-200 and affinity on heparin-Sepharose. The final preparation gave a single band at Mr 110,000 by sodium dodecyl sulfate-polyacrylamide gel electrophoresis after Coomassie stain or autoradiography of the 125I-labeled protein. Immunodetection with the murine monoclonal antibody mAb J28 in nitrocellulose replicas demonstrated a main Mr 110,000 component and trace components in the Mr 100,000-80,000 range. The immunopattern was identical to that in the original crude pancreatic secretion, therefore showing that the molecular characteristics of the protein, i.e. molecular mass, microheterogeneity, and immunoreactivity, were not altered along the purification procedure. FAP was identified as an acidic protein (isoelectric point 4.2-4.8) consisting of a single polypeptide chain having no free SH residues. Analysis of the amino acid composition showed a high proline content. Twenty-two residues of the N-terminal sequence were determined. No significant homology between this peptide and other proteins was found following a search of the NBRF-18 data bank. Sugar analysis showed the presence of mannose which is consistent with N-linked carbohydrate chains and an unusual high ratio in N-acetylgalactosamine residues suggesting the presence of many O-linked carbohydrate chains. Sequential deglycosylation with neuraminidase, hexosaminidase, and O-glycanase yield a single Mr 58,000 peptide showing that, relative to a molecular mass of 110,000, the carbohydrate moiety of FAP accounts for at least 47% of its apparent Mr by sodium dodecyl sulfate-polyacrylamide gel electrophoresis.

Expression of Human Interleukin-2 in Recombinant Baby Hamster Kidney, Ltk−, and Chinese Hamster Ovary Cells: Structure of O-linked carbohydrate chains and their location within the polypeptide

Abstract The similarity or identity of O-glycosylation in glycoproteins from natural sources or produced in heterologous cell lines, a central problem for the development of many biotechnologically relevant production processes, was examined using interleukin-2 (IL-2) as a model. Human interleukin-2 was constitutively expressed in several mammalian cell lines in high amounts. The recombinant proteins were purified to homogeneity and their carbohydrate structures were analyzed. Only the NeuAc alpha 2-3Gal beta 1-3[NeuAc alpha 2-6]GalNAc oligosaccharide structure or the NeuAc alpha 2-3Gal beta 1-3GalNAc were found in all IL-2 preparations secreted from recombinant Ltk-, Chinese hamster ovary, and baby hamster kidney cell lines. The O-linked chains were exclusively linked to Thr in position 3 of the polypeptide chain which is the carbohydrate attachment site in natural human IL-2. The proportions of O-glycosylated versus nonglycosylated forms of the protein secreted by each recombinant cell line were independent of productivity or of cell culture conditions. Our results show that O-glycosylated human IL-2 can be produced by applying recombinant DNA technology in heterologous cell lines with the same type of post-translational modification that is observed for the protein secreted from natural T lymphocytes.

Structural analysis of the carbohydrate chains of a mouse monoclonal IgM antibody.

A mouse monoclonal IgM antibody, directed against human blood group B determinant, was isolated from hybridoma culture growth medium. Chemical analysis indicated presence of N- and O-linked oligosaccharides. The N- and O-linked carbohydrate chains were liberated using two different conditions of reductive alkaline degradation. Structural analysis was carried out on the isolated chains using chemical analysis, 500-MHz 1H-NMR spectroscopy and fast-atom-bombardment mass spectrometry. The following composite structures of the N-linked chains were found: (formula; see text) where R = OH for biantennary structures and R = Neu5Ac alpha 2-3Gal beta 1-4 GlcNAc beta 1- or Neu5Ac alpha 2-3Gal beta 1-3[Neu5Ac alpha 2-6]GlcNAc beta 1- for triantennary structures. The O-linked oligosaccharides, found in the light chains, were shown to have the structure Neu5Ac alpha 2-3Gal beta 1-3GalNAc. The native IgM antibody could be separated on a concanavalin-A-Sepharose column into two subfractions, differing in the presence of a high-mannose-type oligosaccharide.

O-Mannosylation of recombinant human insulin-like growth factor I (IGF-I) produced in Saccharomyces cerevisiae

A glycosylated form of recombinant human insulin-like growth factor I (IGF-I) expressed in Saccharomyces cerevisiae was shown to contain mannose as the only carbohydrate constituent. All oligosaccharide chains of the glycoprotein could be released by mild alkaline treatment, and separated from the protein by gel-permeation chromatography on Bio-Gel P-2. The structures of these O-linked carbohydrate chains were determined by 500-MHz 1H-NMR spectroscopy, affording the disaccharide Manα1-2Man as the major component and the tetrasaccharide Manα1-3Manα1-2Manα1-2Man as a minor component. Reference oligosaccharides were prepared from mannoproteins released from the cell wall of S. cerevisiae X2180 (α-wild type). In addition to previously reported structures, ranging from mannose to mannotetraose, the pentasaccharide Manα1-3Manα1-3Manα1-2Manα1-2Man was identified in the cell wall mannoprotein.

O-Glycosylation of the α-Subunit Does not Limit the Assembly of Chorionic Gonadotropin αβ Dimer in Human Malignant and Nonmalignant Trophoblast Cells*

Biosynthetic experiments were carried out in cultures of human malignant trophoblast cells (the JAR cell line) and in explants of normal first trimester human placental tissue to test the hypothesis that the O-glycosylation of the glycoprotein hormone alpha-subunit at Thr-39 regulates the assembly of the CG alpha beta dimer. This modification of alpha has been shown to ablate its ability to combine in vitro with the beta-subunit of bovine LH and might explain why JAR cells and placental explants secrete uncombined alpha- and beta-subunits in addition to the hCG alpha beta dimer. We have previously detected an O-linked carbohydrate chain at Thr-39 in preparations of secreted free alpha-subunit, but not dimer CG alpha from JAR culture medium. We report here evidence that the O-glycosylation of alpha does not regulate the biosynthetic assembly of the hCG dimer in cultures of JAR choriocarcinoma cells or first trimester placental explants. The intracellular precursor forms of alpha and beta that accumulate in the endoplasmic reticulum and combine in that compartment are not yet modified with O-linked carbohydrate, as determined by measurements of their [3H]galactosamine content after biosynthetic labeling of amino sugars with [3H]glucosamine. Furthermore, only half of the free alpha-subunit secreted by JAR cells and less than 10% of free alpha secreted by 10-week-old placental explants received the O-linked chain. This was shown by determining the ratio of the unglycosylated and glycosylated forms of the tryptic peptide from free alpha that contains the O-glycosylation site (residues 36-42). Based on these findings, we make the following conclusions. 1) O-Glycosylation of alpha-subunit is a late event in the secretory pathway of trophoblasts compared to the rapid combination in the rough endoplasmic reticulum of hCG subunit precursors to form alpha beta dimer. 2) Association of alpha with beta precludes the subsequent addition of the glycan to alpha at Thr-39. 3) The alpha molecules that fail to combine with beta in the endoplasmic reticulum are substrates for the later addition of O-linked carbohydrate, presumably in the Golgi complex, but only a fraction of the free alpha molecules are modified with O-linked carbohydrate.

Novel O-Linked Carbohydrate Chains in the Cellulase Complex (Cellulosome) of Clostridium thermocellum: 3-O-methyl-N-acetylglucosamine as a constituent of a glycoprotein

Alkaline borohydride treatment of the cellulosome of Clostridium thermocellum yielded two major oligosaccharide-alditols, namely D-Galp-beta(1----4)-D-GalOH and (formula; see text) The compounds, isolated via gel permeation chromatography and high performance liquid chromatography, were analyzed by monosaccharide analysis, methylation analysis, gas-liquid chromatography/mass spectrometry, fast atom bombardment/mass spectrometry, and one- and two-dimensional 500-MHz (COSY, HOHAHA, ROESY) 1H NMR spectroscopy. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis combined with blotting technology indicated that the tetrasaccharide is mainly associated with one of the cellulosome subunits.

Specific and nonspecific inhibition of adhesion of oral actinomyces and streptococci to erythrocytes and polystyrene by caseinoglycopeptide derivatives.

Various caseinoglycopeptide derivatives prepared from mammalian milk were evaluated as inhibitors of hemagglutinations mediated by Actinomyces viscosus Ny1, Streptococcus sanguis OMZ9, and, for comparative purposes, plant lectins from Arachis hypogaea and Bauhinia purpurea. It was found that recognition of the beta-D-galactose-(1----3)-2-acetamido-2-deoxy-D-galactose carbohydrate chain by Actinomyces viscosus Ny1 organisms and Arachis hypogaea and B. purpurea agglutinins had similar structural requirements; in all cases, the desialylated bovine caseinoglycomacropeptide, on which several units of the above mentioned disaccharide are clustered, behaved as the most potent hemagglutination inhibitor. By contrast, none of the preparations tested inhibited erythrocyte agglutination by S. sanguis OMZ9. Thus, the desialylated bovine caseinoglycomacropeptide acts as a potent and specific inhibitor of oral Actinomyces adhesion to cell membranes (a soft surface) and could be used as a probe for the study of recognition mechanisms mediated by Actinomyces galactose-binding lectins. During the present study, both native and desialylated variants of the same bovine glycomacropeptide also totally prevented the adhesion of Actinomyces viscosus Ny1, S. sanguis OMZ9, and S. mutans OMZ176 to polystyrene surfaces. Comparative evaluations of various structurally different compounds gave the following results. Neither mono- nor disaccharides related to caseinoglycopeptide carbohydrates prevented adhesion; highly positively or negatively charged polypeptides and polysaccharides were either not or only moderately active. Besides these glycomacropeptides, an inhibitory activity was also exhibited by other mucin-type glycoproteins carrying short O-linked carbohydrate chains (including bovine submaxillary mucin), polyethylene glycol, and bovine serum albumin. Consequently, caseinoglycopeptide prevention of oral bacterial adhesion to polystyrene tubes (a hard surface) takes place with no species specificity and can be compared to nonspecific inhibition exhibited by various polymers with very different structural characteristics.

Differential processing of colony-stimulating factor 1 precursors encoded by two human cDNAs.

The biosynthesis of macrophage colony-stimulating factor 1 (CSF-1) was examined in mouse NIH-3T3 fibroblasts transfected with a retroviral vector expressing the 554-amino-acid product of a human 4-kilobase (kb) CSF-1 cDNA. Similar to results previously obtained with a 1.6-kb human cDNA that codes for a 256-amino-acid CSF-1 precursor, the results of the present study showed that NIH-3T3 cells expressing the product of the 4-kb clone produced biologically active human CSF-1 and were transformed by an autocrine mechanism when cotransfected with a vector containing a human c-fms (CSF-1 receptor) cDNA. The 4-kb CSF-1 cDNA product was synthesized as an integral transmembrane glycoprotein that was assembled into disulfide-linked dimers and rapidly underwent proteolytic cleavage to generate a soluble growth factor. Although the smaller CSF-1 precursor specified by the 1.6-kb human cDNA was stably expressed as a membrane-bound glycoprotein at the cell surface and was slowly cleaved to release the extracellular growth factor, the cell-associated product of the 4-kb clone was efficiently processed to the secreted form and was not detected on the plasma membrane. Digestion with glycosidic enzymes indicated that soluble CSF-1 encoded by the 4-kb cDNA contained both asparagine(N)-linked and O-linked carbohydrate chains, whereas the product of the 1.6-kb clone had only N-linked oligosaccharides. Removal of the carbohydrate indicated that the polypeptide chain of the secreted 4-kb cDNA product was longer than that of the corresponding form encoded by the smaller clone. These differences in posttranslational processing may reflect diverse physiological roles for the products of the two CSF-1 precursors in vivo.

Differential processing of colony-stimulating factor 1 precursors encoded by two human cDNAs.

The biosynthesis of macrophage colony-stimulating factor 1 (CSF-1) was examined in mouse NIH-3T3 fibroblasts transfected with a retroviral vector expressing the 554-amino-acid product of a human 4-kilobase (kb) CSF-1 cDNA. Similar to results previously obtained with a 1.6-kb human cDNA that codes for a 256-amino-acid CSF-1 precursor, the results of the present study showed that NIH-3T3 cells expressing the product of the 4-kb clone produced biologically active human CSF-1 and were transformed by an autocrine mechanism when cotransfected with a vector containing a human c-fms (CSF-1 receptor) cDNA. The 4-kb CSF-1 cDNA product was synthesized as an integral transmembrane glycoprotein that was assembled into disulfide-linked dimers and rapidly underwent proteolytic cleavage to generate a soluble growth factor. Although the smaller CSF-1 precursor specified by the 1.6-kb human cDNA was stably expressed as a membrane-bound glycoprotein at the cell surface and was slowly cleaved to release the extracellular growth factor, the cell-associated product of the 4-kb clone was efficiently processed to the secreted form and was not detected on the plasma membrane. Digestion with glycosidic enzymes indicated that soluble CSF-1 encoded by the 4-kb cDNA contained both asparagine(N)-linked and O-linked carbohydrate chains, whereas the product of the 1.6-kb clone had only N-linked oligosaccharides. Removal of the carbohydrate indicated that the polypeptide chain of the secreted 4-kb cDNA product was longer than that of the corresponding form encoded by the smaller clone. These differences in posttranslational processing may reflect diverse physiological roles for the products of the two CSF-1 precursors in vivo.

Characterization of bovine placental lactogen as a glycoprotein with N-linked and O-linked carbohydrate side chains.

In a previous report we showed that purified bovine placental lactogen (bPL) exists in two isoforms in the 31,000-33,000 Mr range, each with at least five isoelectric variants differing in approximately 2 orders of magnitude in isoelectric points (pI) 4-6. The multiple isoelectric variants are unique to the bovine hormone. In an effort to determine the nature of these variants endo- and exoglycohydrolase digestions were conducted to determine if this hormone was glycosylated. Analysis of peptide/N-glycosidase F and endoglycosidase F digests of radioiodinated bPL on one-dimensional gel electrophoresis showed a Mr decrease from 31,000 to 24,000 and 33,000 to 26,000 for the two isoforms. Digestion with a mixture of neuraminidase plus mixed exoglycosidases resulted in a Mr decrease of 4,000. Digestion with neuraminidase resulted in a Mr decrease of 2,000. Further analysis of peptide/N-glycosidase F- and neuraminidase-treated bPL by two-dimensional gel electrophoresis showed the isoelectric variants shifted from pI 4.4-6.3 to 4.9-8.0. The sialic acid residues on the N-linkage are responsible for the pronounced acidic character of bPL, but do not account for the residual charge heterogeneity as the different isoelectric variants persist after sialic acid removal. The apparent Mr of the protein after removal of N-linked carbohydrate residues is similar to that of PRL and GH. These enzymatic digestion results demonstrate the presence of N-linked complex oligosaccharide residues attached to the beta-amide group of an asparagine residue. Analyses of the sugar content of the molecule were consistent with the presence of one biantennary N-linked and two O-linked carbohydrate chains.