Search

The leukocyte adhesion molecule L-selectin, which mediates the initial steps of leukocyte attachment to vascular endothelium, is intensely glycosylated. Different glycoforms of L-selectin are expressed on different leukocyte subsets and differences in L-selectin glycosylation appear to be correlated with the leukocyte's ability to attach to different endothelial targets. In the present study we addressed the question whether glycosylation of L-selectin influences L-selectin-ligand interactions. To obtain different glycoforms of L-selectin, recombinant proteins were expressed both in the baby hamster kidney (BHK) cell line and in the human myelogenous cell line K562, resulting in sL-sel[BHK] or sL-sel[K562], respectively. The glycosylation characteristics of the purified proteins were determined. The most striking differences in glycosylation were seen in the terminal sialylation. Each of the two proteins carried sialic acids in the alpha 2-3 position, while alpha 2-6-bound sialic acids were found exclusively on sL-sel[K562]. To investigate their adhesive properties, both recombinant sL-selectins were used in cell adhesion assays and interactions with the ligands present on various hematopoietic cell lines or activated human cardiac microvascular endothelial cells were examined. The binding capacity of sL-sel[K562] was about 1.6 fold higher compared to sL-sel[BHK] under static as well as under flow conditions. These findings indicate that the terminal sialylation pattern of L-selectin modulates its binding characteristics.

In the present study, we have analyzed a previously identified constitutively active pituitary adenylate cyclase activating polypeptide (PACAP) type I (PAC1) receptor with a deletion of the single amino acid residue Glu(261) (Y.-J. Cao, G. Gimpl, F. Fahrenholz, A mutation of second intracellular loop of pituitary adenylate cyclase activating polypeptide type I receptor confers constitutive receptor activation, FEBS Lett. 469 (2000)). This glutamic acid residue is highly conserved within the second intracellular loop of class II G protein-coupled receptors and may thus be of importance for many members of this receptor class. To explore the molecular characteristics of this mutant receptor, we performed photoaffinity labeling using previously defined photoreactive PACAP analogues. In COS cells, the PAC1 receptor was expressed in two differently glycosylated forms: a M(r) 75,000 and a M(r) 55,000 form. According to partial deglycosylation, at least three carbohydrate chains may exist in the rat PAC1 receptor expressed in COS cells. The constitutively active PAC1 receptor was expressed at the surface of COS-7 cells at the same density as the wild-type receptor. With respect to the different photoreactive PACAP analogues, the labeling specificity was the same for the wild-type versus mutant receptor: (125)I-[Lys(15)(pBz(2))]-PACAP-27 and (125)I-[Bpa(22)]-PACAP-27 were efficiently incorporated into each of the receptors, whereas (125)I-[Bpa(6)]-PACAP-27 labeled each of the receptors only to a negligible extent. This suggests that both receptors have the same or at least a very similar hormone binding site which is in close contact to Tyr(22) and Lys(15) located in the carboxy-terminal alpha-helical region of the PACAP-27 molecule. However, in comparison with the wild-type PAC1 receptor, the constitutively active receptor showed a markedly (approx. 6--8-fold) enhanced photoaffinity labeling efficiency in particular of the high glycosylated form. The enzymatically deglycosylated rat PAC1 receptor was efficiently labeled by photoreactive PACAP analogues. In contrast, nonglycosylated PAC1 receptors produced by tunicamycin treatment of the transfected COS-7 cells showed a 30-fold lower affinity for PACAP-27 and were capable of signal transduction with 30--50-fold lower potency as compared with the glycosylated PAC1 receptors.

A bioactive peptide containing a glutamine-linked oligosaccharide was chemo-enzymatically synthesized by use of the solid-phase method of peptide synthesis and the transglycosylation activity of endo-beta-N-acetylglucosaminidase. Substance P, a neuropeptide, is an undecapeptide containing two L-glutamine residues. A substance P derivative with an N-acetyl-D-glucosamine residue attached to the fifth or sixth L-glutamine residue from the N-terminal region was chemically synthesized. A sialo complex-type oligosaccharide derived from a glycopeptide of hen egg yolk was added to the N-acetyl-D-glucosamine moiety of the substance P derivative using the transglycosylation activity of endo-beta-N-acetylglucosaminidase from Mucor hiemalis, and a substance P derivative with a sialo complex-type oligosaccharide attached to the L-glutamine residue was synthesized. This glycosylated substance P was biologically active, although the activity was rather low, and stable against peptidase digestion. The oligosaccharide moiety attached to the L-glutamine residue of the peptide was not liberated by peptide-N(4)-(N-acetyl-beta-D-glucosaminyl) asparagine amidase F.

Most beta-1,4-galactosyltransferase (beta-1,4-GalT)-knockout mice die after birth. Although several defects were found transiently in these animals, the primary cause of death is obscure. Not only beta-1,4-linked galactose residues on N-glycans, but also beta-1, 4-GalT activities were found in some of the tissues. Recently, five human genes which encode beta-1,4-GalTs have been cloned, and the possible presence of such novel beta-1,4-GalTs in mice is considered to bring about survival of the mutant animal beyond birth. In order to understand the semi-lethal nature of this animal, it is inevitable to clarify how individual novel beta-1,4-GalTs are involved in the biosynthesis of glycoconjugates based on their acceptor-substrate specificities.

We studied the interaction of n-octyl-beta-d-glucopyranoside-solubilized VIP receptors (VIPR) with wheat germ agglutinin and found that the addition of the lectin to the detergent extract led to the formation of aggregates that could be pelleted by high speed centrifugation. Resuspension of the pellet in the presence of the competing trisaccharide, N,N', N"-triacetylchitotriose (TAC), dissociated the lectin from the complex without altering the precipitability of VIPR. The final pellet (referred to as TAC pellet) contained an average of 12% of total protein and 96% of total VIP binding activity with a typical rank order of potency for VIP-related peptides. Lipid analysis and electron microscopic examination indicated that the precipitated material was composed of lipid vesicles. VIPR molecules were shown to be integrally inserted in the liposomes because they could not be dissociated from the vesicles at pH 11 or with high salt concentration. By comparing the liposome-associated VIP binding activity in the presence and absence of detergent and by showing accessibility of VIPR to PNGase F, it was concluded that VIP binding sites were not simply trapped within the reconstituted vesicles but likely exposed at the external surface of the liposomes.

Variants of the S49 mouse lymphoma cell line exhibit multiple lesions along the pathway of cyclic AMP generation in response to beta2-adrenergic stimulation. Two such variants, beta(p) and beta(d), are characterized by decreased receptor binding and mRNA expression, 50% and 25% of wild-type receptor expression, respectively. The rate of beta2-adrenergic receptor synthesis was measured and found to be decreased in the beta d cells vis-a-vis the rate in wild type cells. The molecular mass of the beta2-adrenergic receptor in the S49 wild-type, beta(p) and beta(d) variant cells was estimated by labeling the receptor with the photoaffinity probe [125I]iodocyanopindololdiazirine. Receptor size was found to be 67,000 and 47,000 Da in the wild-type and 60,000 and 42,000 in the two variant cells. This 6 kDa discrepancy in mass was abolished upon treatment of labeled cell extracts with N-glycosidase F, suggesting the possibility of either N-terminal truncation or altered glycosylation of the receptor in the variant cells. To distinguish between these possibilities, we sequenced the beta2-adrenergic receptor gene and two kilobases of the 5'-non-coding region. No differences were found in the coding region of the gene from wild-type, beta(p) and beta(d) S49 cells suggesting that both the diminished expression and the decreased size of beta2-adrenergic receptor in the beta(p) and beta(d) S49 variants are related to impaired glycosylation of the receptor. This hypothesis was substantiated by the reduced retention of the variant cells' beta2-adrenergic receptor on immobilized WGA. Furthermore, growth of the S49 cells in the presence of the alpha-mannosidase II inhibitor, swainsonine, preferentially impaired the ability of the receptors derived from the variant cells to bind to WGA. These results imply that altered expression and glycosylation of G-protein-linked receptors occur as a consequence of one or more mutations outside the receptor's open reading frame.

Human placental ATP diphosphohydrolase (ATP-DPH), has been previously characterized as an azide-sensitive, Ca(2+)- or Mg(2+)-dependent triphospho- and diphosphonucleosidase which migrates as an 82 kDa protein band on SDS-PAGE (Christoforidis, S. et al. (1995) Eur. J. Biochem. 234, 66-74). In this paper we have studied the subcellular localization of placental ATP-DPH by differential centrifugation and flotation experiments. Using specific enzymatic markers it was found that ATP-DPH is localized on plasma membrane. ATP-DPH was found to be a highly N-glycosylated protein which is a common post-translational modification of plasma membrane proteins. Extensive incubation of the native pure enzyme with N-glycosidase F resulted in the elimination of the 82 kDa form and the concurrent formation of a deglycosylated product of 57.5 kDa and four other intermediate products, indicating the presence of at least five N-glycosylation sites within the ATP-DPH molecule. The partially deglycosylated sample retained its activity in solution and in native gel electrophoresis and activity staining.

Lamb pregastric lipase (LPGL) was purified from pharyngeal tissues. The purification procedure was based on an aqueous extract containing 0.7% Tween 80 which was chromatographed on DEAE-cellulose anion-exchanger and adsorbed on HA-Ultrogel followed by gel filtration on Ultrogel AcA-54. The final enzymatic preparation, where the overall activity recovery was 3%, showed a single protein band on SDS-PAGE with a molecular mass of 50 kDa. LPGL is a glycoprotein containing approx. 14% (w/w) of carbohydrate. Extensive deglycosylation using peptide N-glycosidase F yielded a protein with an apparent molecular mass of 43 kDa. An uncontrolled proteolysis of LPGL during the purification lead to a 45 kDa form which was previously observed in human lysosomal acid lipase (HLAL) and rabbit gastric lipase (RGL). The labile bond X54-Leu55 was identified. Isoelectric focusing of LPGL reveals a major band corresponding to an isoelectric point of 4.8. The pure enzyme displayed specific activities of 950 U mg-1, 300 U mg-1 and 30 U mg-1 at pH 6.0, using tributyroylglycerol, trioctanoylglycerol and trioleoylglycerol as substrates, respectively. Using Western blot analysis, a cross-immunoreactivity of LPGL was observed with purified anti-human gastric lipase polyclonal antibodies. Determination of the amino-acid sequence of 62 residues revealed a high degree of homology with other known preduodenal lipases.

11 beta-hydroxysteroid dehydrogenase (11-HSD) catalyzes the interconversion of corticosterone and 11-dehydrocorticosterone in rats, or cortisol and cortisone in humans. The 'liver' or 'Type I' isozyme is a widely distributed glycoprotein that utilizes NADP+ as a co-factor. To study the role of glycosylation in maintaining enzymatic activity, we introduced mutations into the two potential N-linked glycosylation sites (asparagine-X-serine, residues 158-160 and 203-205) predicted from the rat cDNA sequence. Mutagenesis was performed by a PCR based technique, and wild-type (WT) and mutant cDNAs were expressed in Chinese hamster ovary cells after cloning into the pCMV4 vector. At each putative glycosylation site, asparagine (N) was changed to glutamine (Q) or aspartic acid (D), and serine (S) changed to alanine (A). All three modifications of the first site (N158Q, N158D, S160A) had minimal (75-100% of WT) effects on dehydrogenase activity and caused a mild (50-75% of WT) decrease in reductase activity. In contrast, mutations at the second site had marked effects, with N203Q and N203D completely abolishing both dehydrogenase and reductase activities and S205A decreasing both activities to about 20% of WT. The double mutation of S160A and S205A also abolished all activity, even though the enzyme carrying each mutation alone was, at least, partially active. The results suggest that N203 (which is highly but not completely conserved in short chain dehydrogenase enzymes) is essential for activity of 11-HSD. N-linked glycosylation may be necessary for full activity or stability of the enzyme.

Rat alpha-fetoprotein (RAFP) shows two discrete electrophoretic variants, slow and fast, with molecular weights of 72,500 and 69,500, respectively. To elucidate the structural basis of this heterogeneity, we developed the expression system of RAFP cDNA and its mutants with the use of cultured mouse fibroblasts and a retrovirus vector. This produced recombinant wild-type and three mutant proteins and has several advantages over the use of Escherichia coli or Saccharomyces cerevisiae. The mutants were designed to lack either one or both of the two possible glycosylation sites on RAFP. Electrophoretic analyses relating to the glycosylation states of the recombinant proteins as well as of natural RAFP produced by rat cells before and after digestion with glycopeptidase F indicated that the slow variant has carbohydrate units located at Asn-93 and Asn-229 and the fast one has one at Asn-229.

Bovine zona pellucida (ZP) glycoproteins from ovarian egg emerged as three bands with molecular mass of 78 kDa, 64 kDa and 21 kDa in SDS-PAGE under reducing conditions. Endo-beta-galactosidase (E beta G) digestion of the glycoproteins yielded five products with molecular mass of 76 kDa (E beta G-76), 68 kDa (E beta G-68), 63 kDa (E beta G-63), 47 kDa (E beta G-47) and 21 kDa (E beta G-21) under the same conditions. The N-terminal amino acid sequences of E beta G-76 and E beta G-21 were identical. This fact together with the results of diagonal SDS-PAGE indicated that E beta G-21 (N-terminal region) is linked to E beta G-63 (C-terminal region) through disulfide bond to form E beta G-76. Immunoblot analysis using anti-pig ZP protein antibodies revealed that bovine E beta G-76, E beta G-68 and E beta G-47 correspond to pig PZP2, PZP3 alpha and PZP3 beta glycoproteins, respectively. The E beta G-76 and E beta G-68 components were shown to be specifically cleaved during fertilization.

The sodium-independent GLUT1 glucose transporter is expressed in high density in human erythrocytes and in tissues which serve a barrier function. In the polarized endothelial cells of the brain capillaries, which comprise the blood-brain barrier (BBB), GLUT1 is expressed on both apical and basolateral membranes; however, in the epithelium of the choroid plexus, GLUT1 expression is restricted to the basolateral surface. The present study examined whether these differences in subcellular localization of GLUT1 at the BBB and choroid plexus could be correlated with differential N-linked or O-linked glycosylation of the protein. Western blot analysis of solubilized brain capillaries (BC) and choroid plexus (CP) revealed that while the BC GLUT1 had an average molecular mass identical to that of the purified human erythrocyte transporter (54 kDa), the CP GLUT1 was of lower molecular mass (47 kDa). Treatment of brain capillaries and choroid plexus with N-glycanase resulted in a shift in the mobility of the GLUT1 of both samples to a lower molecular mass of 42 kDa; however, in contrast, treatment with O-glycanase produced no change in the mobility patterns of GLUT1, but did result in O-linked deglycosylation of another BBB marker, gamma-glutamyl transpeptidase. In conclusion, BBB and choroid plexus GLUT1 are subject to differential N-linked glycosylation with the protein having an N-linked carbohydrate side chain of higher molecular mass at the BBB in comparison to the choroid plexus.

A hybrid heterodimeric alkaline phosphatase expressed in KB cells, consisting of placental and intestinal (fetal) subunits, was purified by use of two different immunoaffinity columns using the monoclonal antibodies 2HIMS-1 and HPMS-1. The closely related subunits were found to yield a dimeric active enzyme glycosylated as the mature heterodimeric forms. This enzyme displays intermediate properties to the placental and intestinal (fetal) isozymes with regard to heat stability, inhibition patterns with amino acids and amino acid derivatives, as well as reactivity with monoclonal antibodies specific for human alkaline phosphatase isozymes. Peptide fragments obtained from the hybrid enzyme after cyanogen bromide cleavage belong to either the placental or intestinal (meconial) isozyme as evaluated by SDS polyacrylamid gel electrophoresis, and the N-terminal amino acid sequences, corresponding to the placental and intestinal subunits, can be identified in the peptide fragments. By N-glycanase digestion or tunicamycin treatment, the molecular mass of the subunits was reduced to 62 kDa compared to 69 kDa for the native ones. The results confirm that some cell lines can synthesize hybrid alkaline phosphatases.

Two proteins with alpha-galactosidase activity, alpha-galactosidase A (alpha-GalA) and alpha-galactosidase B (alpha-GalB, or alpha-N-acetylgalactosaminidase; alpha-NAGA) have a high homology of amino-acid sequence. Point mutations of the alpha-GalA gene have been reported only in the exons 1, 2 or 6. In this study, the exon 1-2 and/or 6 sequences of alpha-GalA cDNA were partly substituted by the corresponding regions of alpha-GalB cDNA, and three chimeric proteins were prepared by the baculovirus expression system: CMB12 with substitution at the exon 1-2 region, CMB6 at the exon 6 region, and CMB126 at both regions. They all preserved alpha-GalA antigenicity. Their kinetic properties toward 4-methylumbelliferyl alpha-galactopyranoside were compared with those of alpha-GalA. The catalytic activity was slightly low in CMB12, decreased to 1/10 in CMB6, and restored to a significant degree in CMB126. Km was more than 4-fold higher for CMB6 and CMB126 than for alpha-GalA. The pH optimum was 4.0 for both CMB12 and alpha-GalA, 4.8 for CMB6, and 4.6 for CMB126 and alpha-GalB. The catalytic activity was inhibited most by galactosamine in CMB6, and less in alpha-GalB, CMB126, alpha-GalA and CMB12 in decreasing order. The 50% inhibition concentrations of melibiose (Gal alpha 1-6Glc) and methyl alpha-galactopyranoside were 2.5- to 3-fold higher for CMB126 than for alpha-GalA. These results indicate that the low affinity of CMB126 to the substrate was caused by a reduced affinity to terminal alpha-linked galactose. We conclude that (1) the two regions encoded by exons 1-2 and 6 contribute to the alpha-galactosidic cleavage, and (2) an increase in Km of CMB6 or CMB126, with chimeric substitutions at the exon 6 region, was caused by a loss of affinity toward terminal alpha-linked galactose.

In an approach to examine the lectin-hypothesis in the pathogenesis of coeliac disease, the presence of lectin-like components in three wheat gluten preparations known to induce coeliac disease, gliadin, Frazer fraction III and an acetic acid/ethanol extract of gluten, was investigated. Lectin-like components in these wheat gluten preparations were traced in binding studies employing a variety of model glycoproteins glycosylated with the different types of N-linked oligosaccharides, i.e., those of the high mannose-, complex- and hybrid-type. Binding affinity of wheat proteins to these glycoproteins was analyzed by affinity dotting and blotting techniques and was compared to that of the well characterized lectins Galanthus nivalis agglutinin, Concanavalin A and wheat germ agglutinin. Though the three wheat gluten preparations exhibited binding reactivity for distinct model glycoproteins, no correlation was found between the type of N-glycosylation of the model glycoproteins and their binding capability to the different wheat gluten preparations. Moreover, binding of the three gluten preparations to the model glycoproteins could not be inhibited by competitive saccharides (methyl-alpha-D-mannopyranoside, N-acetyl-D-glucosamine, mannan). Enzymatic deglycosylation of the ligand glycoproteins with endo-beta-N-acetylglucosaminidase H (Endo H, EC 3.2.1.96) or peptide N-glycosidase F (PNGase F, EC 3.5.1.52) abolished their binding reactivity for the plant lectins, but did not affect binding of the wheat gluten preparations. These results give no evidence for the presence of lectin-like components in wheat gluten preparations and do question the 'lectin hypothesis' of coeliac disease.

Rat T-kininogens were characterized in plasma, urine and liver slice incubation medium in different physiological conditions using monoclonal antibodies that allow to distinguish between the two forms of T-kininogen. T-kininogen purified from the plasma of both normal and inflamed Wistar and Sprague-Dawley rats was found to contain the two forms of T-kininogen, TI and TII, separated by non-denaturing polyacrylamide gel electrophoresis. The two forms were also found in the plasma of several strains of normal and inflamed rats, except in the plasma of the Buffalo rat which contained only TII-kininogen. The two forms of T-kininogen were also found in the media in which liver slices from inflamed and non-inflamed wistar rats had been incubated. The urine T-kininogen of normal rats was chiefly TI-kininogen, but both forms were found in the urine of inflamed rats. T-kininogen in the plasma of normal and inflamed rats was further characterized by chromatography on Con A-Ultrogel. In normal plasma, we observed a ratio of non-retained to retained T-kininogen of 0.41. The retained T-kininogen was eluted as two peaks, one eluted with 45 mM and the other with 120 mM alpha-methyl-D-glucoside. The ratio of non-adsorbed to adsorbed T-kininogen in inflamed Wistar rat plasma was 1.40 and the retained material was almost exclusively in a single peak, which eluted with 50 mM alpha-methyl-D-glucoside. The non-adsorbed and adsorbed fractions contained both forms of T-kininogen, but the protein bands in the non-retained fraction had greater mobilities on non-denaturing PAGE. SDS-PAGE analysis of T-kininogen deglycosylated by N-glycosidase F showed a major band with a molecular mass of 50 kDa, whereas the molecular mass of the native form was 66 kDa. We concluded that both forms of T-kininogen exist in the non-inflamed and the inflamed rat plasma, except for the Buffalo rat, and that T-kininogen displays a different chromatographic pattern on Con A-Ultrogel after inflammation suggesting altered glycosylation.

Arylsulfatase A purified from human placenta contained an unreported component with an apparent molecular mass of 7 kDa in addition to the two known components with apparent molecular masses of 58 and 50 kDa. The detailed relationship between the 58 kDa component and the 50 kDa component is as yet unknown. The present study was undertaken to define the structure of the subunits of the sulfatase. The N-terminal sequence of the 50 kDa component was identical to that of the 58 kDa component. Furthermore, the peptide maps of the 50 kDa component, which was separately digested with trypsin and Achromobacter proteinase I, were quite similar to those of the 58 kDa one. Through sequence analysis of the incompatible peaks in the peptide maps, the 50 kDa component was found to lack a sequence from Val-445 to the C-terminus. On the other hand, the N-terminal sequence of the 7 kDa component began with Ala-448, though there was a minor sequence commencing with Thr-449. These observations suggest that the 50 and 7 kDa components were produced by limited proteolysis near the C-terminus of the 58 kDa component. Through analysis using unreducing SDS-PAGE, the 58 and the 7 kDa components were found to be linked by disulphide bonds. Arylsulfatase A purified from human liver was also composed of the same subunits as the placental one. This finding suggests that human arylsulfatase A undergoes similar proteolytic processing regardless of the tissue involved.

The effect of carbohydrate removal on the properties of the lysosomal enzyme alpha-L-fucosidase has been investigated by comparatively characterizing N-glycanase-treated and mock-treated control fucosidases. N-Glycanase treatment removed approx. 90% of the carbohydrate from purified native human liver fucosidase as determined by carbohydrate assay after gel filtration on Sephadex G-50, and by Western blotting with a lectin-digoxigenin conjugate and densitometric scanning. Removal of carbohydrate from fucosidase does not affect its catalytic activity, its Km value for synthetic substrate, its recognition and rate of hydrolysis of three natural substrates, or its gross conformation as determined by circular dichroism. However, loss of carbohydrate led to significantly decreased activity at acidic pH values (3.1-4.7), a 0.6 pH unit shift to a more neutral optimum and decreased thermostability. The decreased activity at acidic pH values and the more neutral pH optimum of deglycosylated fucosidase suggest that the presence of carbohydrate is physiologically significant in allowing fucosidase to perform its catabolic function more efficiently in the acidic milieu of the lysosome.

P-glycoprotein (P-gp) is thought to transport anti-cancer drugs and to be responsible for the multidrug-resistant (MDR) phenotype. Immunohistochemistry reveals that P-gp is also expressed in normal human tissues, such as the adrenal gland, kidney, liver, and the capillary endothelium of the brain and testis. However, little is known about the structural and functional variations of P-gp in these tissues. With immunoblotting and photoaffinity labeling, we found that the molecular mass of P-gp in these tissues varied between 130-140 kDa. To clarify the post-translational modification of P-gp, we studied the biosynthesis of P-gp in a human multidrug-resistant cell line (KB-C2). We found that P-gp was produced in KB-C2 cells as a 125 kDa precursor and was slowly processed (t1/2 = 45-60 min) to the mature form of 140 kDa. In the presence of tunicamycin, a 120 kDa form of P-gp was synthesized and this form was no longer processed. Treating the 125 kDa precursor form with endo-beta-N-acetylglucosaminidase H (Endo H) and the 140 kDa mature form with N-glycanase diminished the molecular size of P-gp to that of the tunicamycin-treated form. N-Glycanase almost completely removed [3H]glucosamine labeling from P-gp. These data indicate that the major modification of P-gp is N-linked glycosylation. P-gps from KB-C2 cells, kidney and adrenal gland had a different lectin-binding capacity. There seems to be a variety of N-linked glycosylations in tissue and tumor P-gps.

The human erythrocyte glucose transporter is a fully integrated membrane glycoprotein having only one N-linked carbohydrate chain on the extracellular part of the molecule. Several authors have suggested the involvement of the carbohydrate moiety in glucose transport, but not definitive results have been published to date. Using transport glycoproteins reconstituted in proteoliposomes, kinetic studies of zero-trans influx were performed before and after N-glycanase treatment of the proteoliposomes: this enzymatic treatment results in a 50% decrease of the Vmax. The orientation of transport glycoproteins in the lipid bilayer of liposomes was investigated and it appears that about half of the reconstituted transporter molecules are oriented properly. Finally, it could be concluded that the release of the carbohydrate moiety from the transport glycoproteins leads to the loss of their transport activity.

The possible noninvolvement of the carbohydrate moiety of human fibrinogen in the clotting mechanism was examined by eliminating the neutral sugar chains from desialylated fibrinogen by almond glycopeptidase digestion. 40% of the total neutral sugars was removed from the desialylated fibrinogen. The neutral sugars from both the beta- and gamma-polypeptide chains were released equally. The protein moiety of the glycopeptidase-digested fibrinogen was found to be intact. No significant change was observed in the thrombin time(fibrinogen clottability) of the resultant fibrinogen. The results suggest that the carbohydrate moiety of fibrinogen is not involved in the clotting mechanism. Oligosaccharide was detected in the glyopeptidase digest of desialylated fibrinogen by thin-layer chromatography (TLC), and was found to be identical with those released quantitatively from the peptic digests of beta- and gamma-polypeptide chains. The structure of the sugar chain was identified tentatively as Gal2-GlcNAc2-Man3-GlcNAc2, by sequential exoglycosidase digestion and quantitative analysis of carbohydrate components.

The glycopeptidase preparation that has been isolated from almond emulsin and acts on beta-aspartylglycosylamine linkages in glycopeptides was separated into three active fractions by DEAE-cellulose column chromatography. The three discrete species of glycopeptidase (Groups A, B and C) have been purified 30-, 136-, and 99-fold, respectively. The optimum pH value of Group A was 6.0 and those of Groups B and C, 5.0. Isoelectric points of Groups A, B and C were pH 7.7, 8.6 and 8.7, respectively. All three glycopeptidases hydrolyzed quantitatively glycopeptides with 3-11 amino acid residues prepared from stem bromelain, ovalbumin and ovotransferrin. Group C preferred glycopeptides with shorter peptide chains, whereas Groups A and B preferred those with longer chains. Glycopeptidase Group A also hydrolyzed intact glycoproteins such as stem bromelain, ovalbumin, Taka-amylase A and desialylated human transferrin.

A glycoprotein that circulates in human blood, binds to the surface of platelets and white cells and also binds serotonin with high affinity and specificity has previously been purified and partially characterized. This glycoprotein has been called serotonectin. Antibodies raised against serotonectin inhibited the uptake of [3H]serotonin by platelets. We now report on the amino acid and carbohydrate composition of this protein as well as on some of the properties of the protein from which the carbohydrate moiety was removed. Serotonectin (apparent molecular weight 200 000; as judged by SDS-polyacrylamide gel electrophoresis) is an acidic protein that contains about 13% carbohydrate (w/w) consisting of mannose, galactose, glucosamine and sialic acid in a molar ratio of 2:1:4:0.8. Initial characterization suggests that serotonectin is a sialoglycoprotein of complex-type oligosaccharide N-linked to asparagine through N-acetylglucosamine. Treatment of serotonectin with neuraminidase resulted in a quantitative release of sialic acid without loss of antigenicity or binding capacity for [3H]serotonin. Treatment of desialylated serotonectin under non-denaturing conditions with almond glycopeptidase A resulted in 60-80% release of sugar. The protein moiety of the glycopeptidase-digested material showed no change in the capacity to bind [3H]serotonin and exhibited the same antigenic properties as untreated serotonectin. These data show the non-involvement of the carbohydrate moiety of human serotonectin in the mechanism of binding serotonin but the possible contribution of this moiety to a tighter interaction with the serotonectin receptor.

After pepsin digestion, all of the carbohydrates in ovalbumin were recovered in two glycopeptides, Glu-Glu-Lys-Tyr-Asn(CHO)-Leu-Thr-Ser-Val and Glu-Gln-Lys-Tyr-Asn(CHO)-Leu-Thr-Ser-Val. Almond glycopeptidase released quantitatively oligosaccharides from the glycopeptides. The products from both glycopeptides contained both the high-mannose-type oligosaccharides and the hybrid-type oligosaccharides in the same ratio. Thus, either the high-mannose-type or the hybrid-type oligosaccharide is attached to the unique asparagine residue in the ovalbumin molecule.

The beta-subunit of dog kidney (Na+ + K+)-ATPase is a sialoglycoprotein and contains three potential N-glycosylation sites. In this study, the oligosaccharide chains of purified dog kidney beta-subunit were labeled with tritium by oxidation with sodium periodate or galactose oxidase followed by NaB3H4 reduction. The beta-subunit was extensively digested by trypsin and the radioactive peptides were purified by HPLC. The enzyme, glycopeptidase A, which catalyzes the removal of N-linked oligosaccharide chains and the conversion of the glycosylated Asn residue to Asp, was used to demonstrate that a number of purified beta-subunit tryptic peptides were glycosylated. Amino-acid analysis of these beta-subunit peptides following glycopeptidase-A treatment revealed the expected Asn to Asp conversion for Asn-157, Asn-192 and Asn-264, demonstrating that all three potential N-glycosylation sites of the dog kidney beta-subunit are glycosylated. In addition, amino-acid sequence data suggest that a disulfide bond exists between Cys-158 and Cys-174.

Taka-amylase A (1,4-alpha-D-glucan glucanohydrolase, EC 3.2.1.1), which contains a single asparagine-linked oligosaccharide unit, was digested with almond glycopeptidase immobilized on Sepharose 6B at 20 degrees C for 4 h. A maximum of 10% of the parent protein was isolated as apoprotein by column chromatography on Con-A Sepharose. The characteristics of the apoprotein were compared to those of the native Taka-amylase A. The removal of the sugar chain from Taka-amylase. A caused no change in the pH-activity profile or in kinetic parameters of the hydrolysis of soluble starch. The stability of the apoprotein toward changing pH and digestion by proteases did not show any appreciable difference from that of the native Taka-amylase. These results suggest that the carbohydrate moiety of Taka-amylase A is not an essential participant in the catalysis.