Use of exoglycosidases from Mercenaria mercenaria (hard shelled clam) as a tool for structural studies of glycosphingolipids and glycoproteins

Modification of the erythrocyte surface in rats bearing Yoshida ascites sarcoma is brought about by a tumour variant of alpha2-macroglobulin

Erythrocytes from the circulation of rats bearing Yoshida ascites sarcoma exhibit higher concanavalin A (ConA)-mediated agglutinability than those from normal animals. A tetrameric glycoprotein of subunit molecular mass 170 kDa, purified from the cell-free ascites fluid, was found to confer higher ConA-mediated agglutinability on erythrocytes in vitro. An antiserum to this tumour-derived protein failed to detect any cross-reactive component in normal rat plasma or in any of the normal tissues examined. An immunoreactive protein was, however, detected in blood plasma when the acute-phase reaction was stimulated by injection of turpentine. The cross-reactive acute-phase protein was purified by ConA-affinity, gel-filtration and ion-exchange chromatography, and identified as alpha2-macroglobulin. The acute-phase protein and the protein obtained from the ascites fluid have identical or very similar native and subunit molecular masses, subunit arrangement and pI. They both are able to inhibit trypsin and, as a consequence, acquire greater mobility in native PAGE. In addition, the two proteins bind to rat erythrocytes non-specifically, and in similar amounts. However, despite these similarities, the acute-phase protein is unable to enhance the agglutinability of erythrocytes. The two proteins differ in their carbohydrate content, but this differential glycosylation is not the cause of the difference in their surface modification activity. The chemically deglycosylated proteins show a small but consistent difference in the size of their polypeptides. Their tryptic peptide maps, although largely similar, show some differences, as do their amino acid compositions. It is probable that the proteins are independent members of the same (alpha-macroglobulin) family. The rat embryo is also found to express a soluble protein consisting of a 170 kDa polypeptide that cross-reacts with the antibody to the tumour-derived protein. The purified embryo protein is able to alter the ConA-mediated agglutinability of erythrocytes in vitro, and also yields a tryptic peptide map that is identical to that of the tumour-derived protein. The modification of the host cell surface in the tumour-bearing rats is thus caused by what appears to be a tumour (oncofetal?) variant of alpha2-macroglobulin.

Biosynthesis in vitro of neolactotetraosylceramide by a galactosyltransferase from mouse T-lymphoma: purification and kinetic studies; synthesis of neolacto and polylactosamine core

The galactosyltransferase, GalT-4, which catalyses the biosynthesis in vitro of neolactotetraosylceramide, nLcOse4Cer (Gal beta 1-4GlcNAc beta 1-3Gal beta 1-4Glc-Cer) from lactotriaosylceramide, LcOse3Cer (Glc NAc beta 1-3Gal beta 1-4Glc-Cer), and UDP-galactose has been purified 107 500-fold from a mineral oil induced mouse T-lyphoma P-1798, using affinity columns. The purified enzyme is partially stabilized in the presence of phospholipid liposomes. Two closely migrating protein bands of apparent molecular weights 56 kDa and 63 kDa were observed after sodium dodecyl sulfate polyacrylamide gel electrophoresis of highly purified mouse GalT-4. These two protein bands, when subjected to limited proteolysis, resulted in three peptides with identical mobilities indicating amino acid sequence identity between the proteins. Both protein bands from P-1798 gave a positive immunostain when tested with polyclonal antibody against bovine lactose synthetase (UDP-Gal:Glc beta 4-galactosyltransferase) following Western blot analysis on nitrocellulose paper. The enzyme has a pH optimum between 6.5 and 7.0 and like all other galactosyltransferases, GalT-4 has absolute requirements for divalent cation (Mn2+). The K(m) values for the substrate LcOse3Cer and donor UDP-galactose are 110 and 250 microM, respectively. Substrate competition studies with LcOse3Cer and either asialo-agalacto-alpha 1-acid glycoprotein or N-acetylglucosamine revealed that these reactions might be catalysed by the same protein. The only other glycolipid which showed acceptor activity toward the purified GalT-4 was iLcOse5Cer (GlcNAc beta 1-1-3Gal beta 1-4Lc3), the precursor for polylactosamine antigens. However, competition studies with these two active substrates using the most purified enzyme fraction, revealed that these two reactions might be catalysed by two different proteins since the experimental values were closer to the theoretical values calculated for two enzymes. Interestingly however, it seems that the GalT-4 from P-1798 has an absolute requirement for an N-acetylglucosamine residue in the substrate since the lyso-derivative (GlcNH2 beta 1-3Gal beta 1-4Glc-sphingosine) of the acceptor glycolipid LcOse3Cer is completely inactive as substrate while the K(m) and Vmax of the reacetylated substrate (GlcNAc beta 1-3Gal beta 1-4Glc-acetylsphingosine) was comparable with LcOse3Cer. Autoradiography of the radioactive product formed by purified P-1798 GalT-4 confirmed the presence of nLcOse4Cer, as the product cochromatographed with authentic glycolipid. The monoclonal antibody IB-2, specific for nLcOse4Cer, also produced a positive immunostained band on TLC as well as giving a positive ELISA when tested with radioactive product obtained using a highly purified enzyme from mouse P-1798 T-lymphoma.

Purification, properties, and immunological characterization of GalT-3 (UDP-galactose: GM2 ganglioside, beta 1-3 galactosyltransferase) from embryonic chicken brain

A beta 1-3 galactosyltransferase (GalT-3; UDP-Gal; GM2 beta 1-3 galactosyltransferase) was purified over 5100-fold from 19-day-old embryonic chicken brain homogenate employing detergent solubilization, alpha-lactalbumin Sepharose, Q-Sepharose, UDP-hexanolamine Sepharose, and GalNAc beta 1-4Gal beta-Synsorb column chromatography. The purified enzyme was resolved into two bands on reducing gels with apparent molecular weights of 62 kDa and 65 kDa, respectively. GalT-3 activity was also localized in the same regions by activity gel analysis and sucrose-density gradient centrifugation of a detergent-solubilized extract of 19-day-old embryonic chicken brain. Purified GalT-3 exhibited apparent Kms of 33 microM, and 14.4 mM with respect to the substrates GM2, UDP-galactose, and MnCl2, respectively. Substrate specificity studies with the purified enzyme and a variety of glycosphingolipids, glycoproteins, and synthetic substrates revealed that the enzyme was highly specific only for the glycosphingolipid acceptors, GM2 and GgOse3Cer (asialo-GM2). Ovine-asialo-agalacto submaxillary mucin inhibited the transfer of galactose to GM2 but did not act as an acceptor in the range of concentrations tested. Polyclonal antibodies raised against purified GalT-3 inhibited GalT-3 activity in vitro and Western-immunoblot analysis of purified GalT-3 showed immunopositive bands at 62 and 65 kDa.

Reconstituted Sendai virus envelopes as biological carriers: dual role of F protein in binding and fusion with liver cells

We have assessed the potential of reconstituted Sendai viral envelopes containing only the fusion protein (F-virosomes) as biological carriers for the delivery of drugs and macromolecules. [125I]lysozyme entrapped in F-virosome is used to study its distribution in various organs of Balb/c mouse in vivo as a function of dose and time. F-virosomes injected intravenously are rapidly cleared from circulation. A major percentage (55-60%) of vesicle contents is delivered to liver at 15 min after injection, showing thereby the liver to be the major site for the accumulation of vesicles. Uptake of virosomes by liver is found to reach a near saturation level at a dose of 0.5 mg F-protein associated with virosomes. In competition studies, the inhibitory effect of asialofetuin on the uptake of F-virosomes suggests the involvement of asialoglycoprotein receptor in its recognition by hepatic parenchymal cells. Incorporation of asialoganglioside-GM1 in the F-virosomes enhanced the uptake by about 1.6-fold. The observed specific interaction of hepatic receptor with F-protein containing a terminal galactose moiety is further supported by degalactosylation of F-virosomes with hard-shelled clam exoglycosidase. The uptake of degalactosylated F-virosomes by liver is found to be significantly reduced. The subcellular radioactivity profile in liver cells exhibits a considerable decrease in cytosolic localisation of the degalactosylated F-virosomal contents with a concomitant increase in their accumulation in lysosomal/mitochondrial fraction as compared to the untreated virosomes. Trypsinized and heat-treated F-virosomes also reflect similar subcellular distribution profile as that of degalactosylated virosomes. Moreover, F-virosomes are able to interact and deliver [125I]lysozyme to the HepG2 cells in culture in the presence of a potent inhibitor of endocytotic process. These results indicate the involvement of specific binding of F-proteins with hepatic receptors followed by their fusion with the membrane of liver cells in the delivery of [125I]lysozyme. The findings reported here open up the possibility of using F-virosomes with defined specificity as fusogenic vehicles for efficient delivery of drugs and biologically active macromolecules both in vivo and in vitro.

Targeted delivery of hygromycin B using reconstituted Sendai viral envelopes lacking hemagglutinin-neuraminidase

Hygromycin B was encapsulated in reconstituted Sendai viral envelopes containing only the fusion (F) protein (F-virosomes). Incubation of loaded F-virosomes with cultured HepG2 cells resulted in fusion mediated delivery of hygromycin B to the cell cytoplasm, as was inferred from inhibition of DNA synthesis. Binding of the F-virosomes to HepG2 cells was mediated by the interaction of terminal beta-galactose residues of fusion protein with asialoglycoprotein receptor on HepG2 cells, subsequently leading to fusion between the two membranes. The cytotoxic effect of hygromycin B enclosed in F-virosomes was comparable with that of F,HN-virosomes containing both hemagglutinin-neuraminidase (HN) and F protein and F,HNred-virosomes containing HN whose disulfide bonds were irreversibly reduced (HNred). Hygromycin B loaded fusogenic liposomes were prepared by coreconstituting the viral envelope containing only fusion protein with exogenous lipids. These fusogenic liposomes were found to be more active than F-virosomes at the same fusion protein concentrations.

The influence of the nature of the asparagine 289-linked oligosaccharide on the activation by urokinase and lysine binding properties of natural and recombinant human plasminogens

Several strategies have been used to obtain recombinant (r) human plasminogens (HPg) containing different oligosaccharide side chains on its sole N-linked glycosylation site, present at Asn289. The approaches included expression of the cDNA for HPg in insect cell lines under various conditions, addition of glycosidase inhibitors during expression, and purification of specific glycoforms of HPg using affinity chromatography on an insolubilized lectin column. The activation kinetics for urokinase (UK) of each of the purified HPgs, as well as their relative abilities to bind to the ligand, epsilon-aminocaproic acid (EACA), were then determined. Removal of both N- and O-linked oligosaccharide from HPg resulted in a slight increase in the Kcat/Km for its activation, while a glycoform containing tetrasialyl-tetra-antennary complex oligosaccharide on Asn289 was a slightly poorer substrate for UK than plasma HPg, which contains bisialyl-biantennary complex carbohydrate on Asn289. The most dramatic differences were observed for HPgs with high mannose-type glycans on Asn289. (Man9GlcNAc2)-HPg possessed only approximately 6% of the kcat/Km of plasma HPg, whereas (Glc3Man9GlcNAc2)-HPg did not undergo activation at a significant rate by UK. Differences were also found in the relative abilities of the HPg glycoforms to interact with EACA. The most effective interactions were observed with HPgs containing complex-type glycans, and the least effective binding was found for HPgs with high mannose-type oligosaccharides. The full range of the binding effects is represented by a fourfold difference between HPg containing tetrasialyl-tetra-antennary glycan and HPg with (Glc3Man9GlcNAc2) assembled on Asn289. These results clearly demonstrate that the nature of the N-linked glycan assembled on HPg dramatically influences its ability to be activated by UK and to bind to omega-amino acid effector molecules.

In vitro biosynthesis of GbOse4Cer (globoside) and GM2 ganglioside by the (1-->3) and (1-->4)-N-acetyl beta-D-galactosaminyltransferases from embryonic chicken brain. Solubilization, purification, and characterization of the transferases

(1-->4)-N-Acetyl-beta-D-galactosaminyltransferase (GalNAcT-1) and (1-->3)-N-acetyl-beta-D-galactosaminyltransferase (GalNAcT-2), which are involved in the in vitro biosynthesis of GM2 and GbOse4Cer glycosphingolipids, respectively, have been solubilized and separated by differential detergent extraction from a membrane preparation of 19-day-old embryonic chicken brain. The separated GalNAcT-1 activity had a pH optima of 7.8-8.0, and the separated GalNAcT-2 activity a single pH optimum of 7.2. Furthermore, the partially purified GalNAcT-2 preparation catalyzed the transfer of N-acetylgalactosamine from UDP-D-[3H]GalNAc to only GbOse3Cer and nLcOse5Cer. Both GalNAcT-1 and GalNAcT-2 activities were purified to approximately 316- and 428-fold, respectively, by use of UDP-hexanolamine-Sepharose 4B affinity-column chromatography. However, the partially purified GalNAcT-1 preparation appeared to be active only with GM3, lactosylceramide, and lactotriaosylceramide. The proposed linkage of the N-acetylgalactosamine unit incorporated into GM3 is beta-D-GalpNAc-(1-->4)-GM3 from the isolation of [3H]threitol after hydrolysis of the desialylated, lead tetraacetate-treated, enzymic product, beta-D-GalpNAc-(1-->4)-beta-D-[6-3H]Galp-(1-->4)-beta-D-Glcp-(1-->1)-Cer . In addition, beta-D-GalpNAc-(1-->3)-GbOse3Cer was produced, as shown by the identification of 2,4,6-tri-O-methyl-galactose after permethylation and hydrolysis of the GalNAcT-2 enzymic product, GalpNAc-[6-3H]Galp--->Gal-->Glc-->Cer.

In vitro galactosylation of rhodopsin and opsin: kinetics, properties and characterization

At best, only trace amounts of galactose have been detected as constituents of rhodopsin as analysed by several laboratories. Nevertheless, the enzymatic galactosylation of rhodopsin proceeds readily in vitro, a process which can be catalysed by galactosyltransferases from several sources. Little information is available, however, concerning the properties of the in vitro reaction. We have examined characteristics of the latter process with the hope of shedding light on the capacity of the retina to carry out this reaction. Kinetic properties of the galactosyltransferases of bovine and embryonic chick retinas, bovine milk and rat liver-Golgi were examined using rhodopsin, opsin, N-acetylglucosamine and ovalbumin as exogenous acceptors. All of these studies demonstrated the very limited activity of the galactosyltransferases of the retina as compared to the milk and rat liver systems. The subcellular distribution of the galactosyltransferases of bovine retina was examined. The influence of compounds that might modulate the reaction was also examined. alpha-Lactalbumin, a modifier of the galactosyltransferase in milk, acted as a competitive inhibitor of the galactosylation of opsin. Analogs of vitamin A, shown to inhibit galactosyltransferases in other systems, did not have this effect on the galactosylation of opsin. Similarly, mixing experiments could not demonstrate the presence of endogenous material that inhibited the reaction in the retina. The conformation of the visual pigment was shown to influence the reaction. After bleaching by visible light, opsin was preferred over rhodopsin as an acceptor of galactose by the galactosyltransferases of bovine and embryonic chick retinas and by rat liver. This distinction was only minimally demonstrated by the milk enzyme. The galactosylation of ovalbumin was unaffected by conditions of light or dark by any of the enzymes. While the mode ratio of galactose to rhodopsin after catalysis by the milk enzyme was about 1.6, this ratio was only about 0.01 after reaction with the enzyme from bovine retina. The linkage of galactose in enzymatically galactosylated rhodopsin and opsin was beta(1-4).