Studies on the biosynthesis of glyco-sphingolipids in cultured mouse neuroblastoma cells: characterization and acceptor specificities of N-acetylneuraminyl- and N-acetylgalactosaminyltransferases

Alterations in neuroblastoma ganglioside synthesis by induction of GD1b synthase by retinoic acid

Recent findings link increased expression of the structurally complex 'b' pathway gangliosides GD1b, GT1b, GQ1b (CbG) to a favourable clinical and biological behaviour in human neuroblastoma (NB). Seeking a model to probe these observations, we evaluated four human NB cell lines. Very low CbG content (4-10%) in three of the four cell lines (LAN-5, LAN-1, SMS-KCNR) reflected the ganglioside pattern observed in the most aggressive NB tumours. Pharmacological alterations of complex ganglioside synthesis in vitro by a 5-7 day exposure to 5-10 microM retinoic acid, which is employed in maintenance therapy of disseminated NB, included markedly increased (i) relative expression of CbG (6.6+/-2.0-fold increase, P=0.037), (ii) relative expression of the analogous 'a' pathway gangliosides, termed CaG (6.4+/-1.4-fold increase in GM1a and GD1a; P=0.010), and (iii) total cellular ganglioside content (2.0-6.3-fold), which in turn amplified the accumulation of structurally complex gangliosides. Substantial increases (2.7-2.9-fold) in the activity of GD1b/GM1a synthase (beta-1,3-galactosyltransferase), which initiates the synthesis of CbG and CaG, accompanied the all-trans retinoic acid (ATRA)-induced ganglioside changes. Thus, increased CbG synthesis in NB cell lines is attributable to a specific effect of ATRA, namely induction of GD1b/GM1a synthase activity. Since the shift towards higher expression of CbG and CaG during retinoic acid-induced cellular differentiation reflects a ganglioside pattern found in clinically less-aggressive tumours, our studies suggest that complex gangliosides may play a role in the biological and clinical behaviour of NB.

1H NMR studies of a biosynthetic lacto-ganglio hybrid glycosphingolipid: confirmation of structure, interpretation of "anomalous" chemical shifts, and evidence for interresidue amide-amide hydrogen bonding

Glycosphingolipids bearing GlcNAc beta 1----3 and GalNAc beta 1----4 linked to beta-Gal of lactosylceramide (lacto-ganglio hybrids), first isolated from a murine myelogenous leukemia cell line [Kannagi, R., Levery, S. B., & Hakomori, S. (1984) J. Biol. Chem. 259, 8444-8451], have since been found as normal components of mullet roe and English sole liver. In order to clarify the biosynthetic pathways responsible for its occurrence both as a product of normal tissues and as a possible mammalian cancer-associated antigen, the lacto-ganglio hybrid core structure LcGg4Cer was synthesized from Lc3Cer using a GalNAc beta 1----4 transferase preparation from English sole liver. A preliminary characterization of the enzyme, which may be identical to the GalNAc T-1 responsible for synthesis of GM2 ganglioside, is presented. The enzymatically synthesized product was analyzed by 1- and 2-D 1H NMR spectroscopy, confirmining its primary structure as GalNAc beta 1----4-(GlcNAc beta 1----3)Gal beta 1----4Glc beta 1----1Cer. In addition to assigning all nonexchangeable glycosyl proton resonances, measurements of several properties of the amide NH protons, including chemical shift, coupling constants, exchange rates, and temperature shift coefficients, were obtained and compared to those in the simpler constituent triglycosylceramides, Lc3- and Gg3Cer. An approximate three-dimensional structure for LcGg4Cer is proposed, consistent with all data obtained, which should be useful in discussing the results of 1H NMR analysis of compounds containing this core tetrasaccharide. The structure is characterized by an unusual arrangement of terminal N-acetylhexosamine residues, resulting in a pi-H hydrogen-bonding interaction between their acetamido groups.

Coordinate regulation of ganglioside glycosyltransferases in differentiating NG108-15 neuroblastoma x glioma cells

The enzymatic basis for ganglioside regulation during differentiation of NG108-15 mouse neuroblastoma x rat glioma hybrid cells was studied. This cell line contains four gangliosides that lie along the same biosynthetic pathway: GM3, GM2, GM1, and GD1a. Chemically induced neuronal differentiation of NG108-15 cells led to an 80% drop in the steady-state level of their major ganglioside, GM3, a sixfold increase in the level of a minor ganglioside, GM2 (which became the predominant ganglioside of differentiated cells); and relatively little change in the levels of GM1 and GD1a, which lie further along the same biosynthetic pathway. The enzymatic basis for this selective change in ganglioside expression was investigated by measuring the activity of two glycosyltransferases involved in ganglioside biosynthesis. UDP-N-acetylgalactosamine: GM3 N-acetylgalactosaminyltransferase (GM2-synthetase) activity increased fivefold during butyrate-induced differentiation, whereas UDP-galactose: GM2 galactosyltransferase (GM1-synthetase) activity decreased to 10% of its control level. Coordinate regulation of these two glycosyltransferases appears to be primarily responsible for the selective increase of GM2 expression during NG108-15 differentiation.

Sialyltransferase-1 in a human malignant glioma cell line. Kinetic characteristics and effect of human interferon-beta

Sialyltransferase-1 activity was studied in cultured 12-18 human glioma cells. The apparent Km and Vmax with variable LacCer concentrations were 32 microM and 197 pmoles/mg protein/hr and with variable CMP-NeuAc concentrations were 172 microM and 877 pmoles/mg protein/hr., respectively. The pH optimum towards exogenous LacCer was 6.0 and towards endogenous acceptors was 6.2. The optimum protein:detergent ratio was 1:1. Human beta interferon (1000 units/ml medium) increased sialyltransferase-1 activity only slightly on a protein basis but increased it 47% on a per cell basis. These results demonstrate that one of the biochemical effects of beta-interferon on 12-18 human glioma cells is to stimulate ganglioside synthesis.

Effects of lectin activation on sialyltransferase activities in human lymphocytes

The effects of phytohemagglutinin (PHA) stimulation on the activities of sialyltransferase 1 (SAT-1), and sialyltransferase 3 (SAT-3), in human lymphocytes were investigated in vitro. For SAT-1 and SAT-3, respectively, the apparent Km values with variable CMP-NeuAc concentrations were 0.19 and 0.015 mM and with variable LacCer were 0.075 and 0.17 mM. Progressive increases in the activities of SAT-1 and SAT-3 were detected in lymphocytes stimulated with PHA, whereas no increase was observed in control lymphocytes incubated in culture medium alone. These increased activities occurred within 18-36 h of incubation and preceded optimum lymphocyte proliferation. Intact lymphocytes were needed for the lectin-stimulated increase of sialyltransferase activities because neither concanavalin A nor phytohemagglutinin added to the broken cell preparation modulated SAT-1 activity. The glycolipid products formed as a result of these enzymatic reactions in the presence of endogenous and exogenous acceptors were tentatively identified by thin-layer chromatography and autofluorography. The addition of exogenous LacCer to the SAT-1 assay resulted in the radiolabeling of a small amount of ganglioside GM1b (3.4%), but GM3 was the major labeled product (96%). When GgOse4Cer was added to the SAT-3 assay, 32% GM3 and 24.6% GM1b were detected while 44% consisted of glycolipids not labeled in assays performed without exogenous acceptors. Of the radioactivity transferred to endogenous acceptors, 81.3% was in GM3 and 14.6% in GM1b. These results demonstrate that the modulation of sialyltransferase activity occurs earlier than cellular activation.

Ganglioside glycosyltransferase assay using ion-exchange chromatography

A rapid method for determination of ganglioside glycosyltransferase activity has been developed employing ion-exchange chromatography. Using 13-day chick brain as a source of uridine diphospho-N-acetyl-D-galactosamine: ganglioside GM3 N-acetylgalactosaminyltransferase (ganglioside GM2 synthetase), we were able to accurately measure transfer of N-[3H]acetylgalactosamine (GalNAc) from UDP-[3H]GalNAc to GM3 by application of the reaction mixture to small columns of DEAE-Sepharose and elution of radiolabeled GM2 reaction product with 10 mM potassium acetate in methanol. This method proved to be as reliable and sensitive as previously published assays but requires less time and fewer manipulations.

Studies on the turnover and subcellular localization of membrane gangliosides in cultured neuroblastoma cells

To compare the subcellular distribution of endogenously synthesized and exogenous gangliosides, cultured murine neuroblastoma cells (N1E-115) were incubated in suspension for 22 h in the presence of D-[1-3H]galactose or [3H]GM1 ganglioside, transferred to culture medium containing no radioisotope for periods of up to 72 hr, and then subjected to subcellular fractionation and analysis of lipid-sialic acid and radiolabeled ganglioside levels. The results indicated that GM2 and GM3 were the principal gangliosides in the cells with only traces of GM1 and small amounts of disialogangliosides present. About 50% of the endogenously synthesized radiolabelled ganglioside in the four major subcellular membrane fractions studied was recovered from plasma membrane and only 10-15% from the crude mitochondrial membrane fraction. In contrast, 45% of the exogenous [3H]GM1 taken up into the same subcellular membrane fractions was recovered from the crude mitochondrial fraction; less than 15% was localized in the plasma membrane fraction. The results are similar to those obtained from previously reported studies on membrane phospholipid turnover. They suggest that exogenous GM1 ganglioside, like exogenous phosphatidylcholine, does not intermix freely with any quantitatively major pool of endogenous membrane lipid.

Lipids of nervous tissue: composition and metabolism

As indicated in the Introduction, the many significant developments in the recent past in our knowledge of the lipids of the nervous system have been collated in this article. That there is a sustained interest in this field is evident from the rather long bibliography which is itself selective. Obviously, it is not possible to summarize a review in which the chemistry, distribution and metabolism of a great variety of lipids have been discussed. However, from the progress of research, some general conclusions may be drawn. The period of discovery of new lipids in the nervous system appears to be over. All the major lipid components have been discovered and a great deal is now known about their structure and metabolism. Analytical data on the lipid composition of the CNS are available for a number of species and such data on the major areas of the brain are also at hand but information on the various subregions is meagre. Such investigations may yet provide clues to the role of lipids in brain function. Compared to CNS, information on PNS is less adequate. Further research on PNS would be worthwhile as it is amenable for experimental manipulation and complex mechanisms such as myelination can be investigated in this tissue. There are reports correlating lipid constituents with the increased complexity in the organization of the nervous system during evolution. This line of investigation may prove useful. The basic aim of research on the lipids of the nervous tissue is to unravel their functional significance. Most of the hydrophobic moieties of the nervous tissue lipids are comprised of very long chain, highly unsaturated and in some cases hydroxylated residues, and recent studies have shown that each lipid class contains characteristic molecular species. Their contribution to the properties of neural membranes such as excitability remains to be elucidated. Similarly, a large proportion of the phospholipid molecules in the myelin membrane are ethanolamine plasmalogens and their importance in this membrane is not known. It is firmly established that phosphatidylinositol and possibly polyphosphoinositides are involved with events at the synapse during impulse propagation, but their precise role in molecular terms is not clear. Gangliosides, with their structural complexity and amphipathic nature, have been implicated in a number of biological events which include cellular recognition and acting as adjuncts at receptor sites. More recently, growth promoting and neuritogenic functions have been ascribed to gangliosides. These interesting properties of gangliosides wIll undoubtedly attract greater attention in the future.(ABSTRACT TRUNCATED AT 400 WORDS)

Cell-cycle dependence of a ganglioside glycosyltransferase activity and its inhibition by enkephalin in a neurotumor cell line

Rat glioma X mouse neuroblastoma hybrid neurotumor cells (NG108-15), synchronized by amino acid deprivation, showed a cell-cycle-dependent peak of activity of a ganglioside N-acetylgalactosaminyl transferase 14-24 h following release from the cell cycle block (S/G2 phase). Maximal expression of two typical lysosomal hydrolases, N-acetyl-beta-hexosaminidase and beta-galactosidase, occurred between 18 and 21 h following release (S phase), declining to G1 phase levels during the peak of N-acetylgalactosamine (GalNAc) transferase activity. In addition, glycosyltransferase activity in G2 phase cells showed an increase in apparent Vmax (suggesting the presence of more enzyme/mg of cell protein) and apparent binding affinity for uridine diphosphate N-acetylgalactosamine (UDP-GalNAc) (32 versus 14 microM) when compared to transferase activity in the G1 phase. However, the opioid peptide enkephalin [D-Ala2, D-Leu5], which inhibits ganglioside GalNAc transferase activity in unsynchronized NG108-15 cultures, was much more inhibitory in whole cells 8 h after release from the cell cycle block (G1 phase) than in cells 20 h after release (G2 phase), with 50% inhibition occurring at 2 X 10(-9) M and 2 X 10(-7) M, respectively. These results suggest that the GalNAc transferase activity is regulated in more than one way during the cell cycle, since both Vmax and Km changes are observed, and that the cyclic AMP-dependent mechanism by which opiates reduce transferase activity is receptor mediated and cell cycle dependent.

Glycosphingolipids of an alveolar rhabdomyosarcoma and normal human muscle. A case study

The distribution of neutral, or asialosyl-, glycosphingolipids and gangliosides in a rhabdomyosarcoma of alveolar type have been studied. Histologically, this muscle tumor is composed primarily of two cell types: one with oval or round hyperchromatic nuclei and very little cytoplasm, and one a giant cell, with multiple, peripherally placed nuclei and weakly staining eosinophillic cytoplasm. In comparing glycolipids of the rhabdomyosarcoma with normal muscle from the same leg, the striking alteration in the tumor was a virtual disappearance of ganglioside GM2. There was also a slight increase in GM3 and a decrease in GD1a. The asialosyl derivative of GM2 (GalNac-Gal-Glc-Cer) was markedly increased in the tumor. A loss of glucosylceramide was also observed. The results are discussed in terms of glycolipid metabolic changes in muscle oncogenesis and their implications.

Characterization of a glycosphingolipid beta-N-acetylgalactosaminyl-transferase activity in cultured hamster (nil) cells

The activity of a glycosphingolipid N-acetylgalactosaminyltransferase (GalNAc transferase) in cultured hamster fibroblasts (NIL-8) was characterized with respect to substrate binding, acceptor specificity, pH optimum and detergent requirements. Of the glycosphingolipid acceptors tested, transferase activity was observed only with globotriaosylceramide. The apparent Km values for uridinediphosphate-N-acetylgalactosamine and globotriasylceramide were 0.14 and 0.42 mM, respectively. The enzyme required Mn2+ for maximum activity (4 mM), and Mg2+ was not able to replace Mn2+. Of the detergents tested, sodium taurodeoxycholate gave the greatest activation of the enzyme at 1 mg/ml. A broad pH optimum (4.5-8.0) was obtained, with maximum activity at pH 6.0 in 2-(N-morpholino)ethanesulfonic acid. Globotetraosylceramide and II3-alpha-N-acetylneuraminyl-lactosylceramide inhibited transferase activity with globotriaosylceramide as substrate, but lactosylceramide had no effect on the activity with this acceptor. The major product of the assay was shown to be a tetraglycosylceramide with a terminal beta-N-acetylgalactosamine moiety by co-migration with authentic globotetraosylceramide on TLC plates and by cleavage of the labeled N-acetylgalactosamine from the product by jack bean beta-hexosaminidase.

Effects of divalent cations on the glycolipids from cultured mouse neuroblastoma cells

The influence of divalent cations on glycosphingolipid metabolism was examined in the NB41A mouse neuroblastoma clonal cell line. HPLC methods were utilized to quantitate the effects on neutral glycolipids and monosialogangliosides. NB41A cells were shown to contain GM3, GM2, GM1, GD3, and GD1a by HPLC and TLC. The neutral glycosphingolipids consisted of glucosylceramide (GlcCer), lactosylceramide (LacCer), GalNAc (beta 1 leads to 4) Gal(beta 1 leads to 4)Glc(beta 1 leads to 1)Cer (GgOse3Cer), and GalNAc(beta 1 leads to 3)Gal(alpha 1 leads to 4) Gal(beta 1 leads to 4)Glc(beta 1 leads to 1)Cer (GbOse4Cer) according to their HPLC behavior. Cells grown in the presence of 1.85 mM-EGTA showed a two-to threefold increase in GM3 whereas other glycosphingolipids were only slightly affected. When cells were grown in the presence of 1.45 mM-EGTA plus 0.4 mM-EDTA a similar increase in GM3 was observed but this change was now accompanied by decreases in GM2, GM1, GgOse3Cer. The EGTA-EDTA effects were reversed when growth was in the presence of Ca2+ sufficient to bind all chelator. Mn2+ replacement reversed the chelator effects differentially; GM2 and GM1 levels were the most sensitive to increases in Mn2+ concentration; GgOse3Cer and GbOse4Cer were also sensitive, whereas GM3 was the least affected. These results suggest calcium serves an important regulatory role on GM3 levels and that manganese concentration may regulate the levels of galactosamine-containing glycolipids in mouse NB41A neuroblastoma cells.

Possible role of cyclic AMP in the receptor-mediated regulation of glycosyltransferase activities in neurotumor cell lines

Exposure of mouse neuroblastoma cell line N4TGl to opiates or [D-Ala2,D-Leu5] enkephalin produced a naloxone-reversible inhibition of cyclic AMP synthesis and prevented, in a concentration-dependent manner, the formation of both ganglioside GM2 (GalNAc-[NeuNAc]-Gal-Glc-ceramide) from GM3 (NeuNAc-Gal-Glc-ceramide) and ganglioside GM1 (Gal-GalNAc-[NeuNAc]-Gal-Glc-ceramide) from GM2 in cell-free extracts. In contrast, the receptor-mediated elevation of intracellular cyclic AMP levels by agents such as prostaglandin E1 (in the presence of isobutylmethylxanthine) or the addition of the cyclic AMP derivatives (dibutyryl cyclic AMP) markedly stimulated the activities of UDP-GalNAc:GM3,N-acetylgalactosaminyltransferase and UDP-Gal:GM2,galactosyltransferase. An overall increase in the synthesis of gangliosides more complex than GM3 was also observed in the mouse neuroblastoma x hamster brain explant hybrid cell line NCB-20 following elevation of cyclic AMP levels by treatment with serotonin and pargyline. The data presented support the hypothesis that cyclic AMP may have a role in the regulation of sialoglycosphingolipid biosynthesis.

Ganglioside biosynthesis in rat liver: characterization of cytidine-5'-monophospho-n-acetylneuraminic acid:hematoside (GM3) sialyltransferase

CMP-NAcNeu:GM3 ganglioside sialytransferase (GD3 synthase) was concentrated 80-100-fold, relative to total homogenates, in Golgi apparatus fractions from rat liver. Ultrasound treatment of Golgi apparatus in a low salt medium extracted 40-60% of the original protein but did not dissociate the transferase from membranes. The acivity was greatest in the presence of certain detergents, had a pH optimum of 6.2, was stimulated by mg2+ and diacylphospholipids and was inhibited by lysophospholipids. Apparent Km values for CMP-NAcNeu and GM3 were about 0.8 and 0.2 mM, respectively. On chromatographic separation, virtually all the reaction product migrated as GD3. GD3 synthase appeared to be a glycoprotein since the activity bound to concanavalin A-Sepharose and was eluted, with increased specific activity, by alpha-methyl mannoside.

Biosynthesis in vitro of mono- and di-sialosylgangliosides from gangliotetraosylceramide by cultured cell lines and young rat brain. Structure of the products, and activity and specificity of sialosyltransferase

Incubations in vitro of GA1, labeled with 3H in the terminal D-galactopyranosyl group, with nonradioactive CMP-NeuNAc in the presence of homogenates of C21 rat brain glial cells, NIE mouse neuroblastoma cells, 3T3 mouse fibroblasts, SV 40-transformed 3T3 cells, chick embryo fibroblasts, Rous sarcoma virus-transformed chick embryo fibroblasts, and 9-day old rat brain resulted in all cases in the formation in high yield of GM1b, in which the neuraminidase-labile NeuNAc group is linked at O-3 of the terminal D-galactosyl residue, as shown by permethylation studies. No trace of the naturally occurring neuraminidase-stable GM1a was detected in any case. In addition, with NIE cells, and normal and RSV-transformed chick embryo fibroblasts, a disialosylganglioside (GD1) differing from GD1a and GD1b, and bearing only one substituent at O-3 of the terminal D-galactopyranosyl residue was formed. It was also biosynthesized from GM1b and CMP-NeuNAc by NIE and chick embryo cells but not by C21 cells, or rat brain. However, C21 cells and rat brain were capable of synthesizing GD1a from GM1a. Periodate oxidation degraded both NeuNAc groups in GD1 to a 7-carbon fragm:nt, indicating lack of substitution at O-8. GM1b could not be detected as a natural product in rat brain.

Sialyltransferases in young rat brain

1. There are more glycolipid acceptor sites for NeuNAc than for glycoproteins in 11--15 day old rat cerebra. 2. The glycolipid acceptors appear to be almost exclusively Cer-Glc-Gal and GM1 ganglioside and each is a substrate for a different sialyltransferase. 3. The sialyltransferase(s) that acted on glycoprotein could be differentiated from the ones that acted on the glycolipids. 4. The apparent Km for CMP-NeuNAc was the same for all four of the sialyltransferase reactions studied. 5. Electron microscopic examination and marker enzyme studies on continuous sucrose gradient fractions found that most of the sialyltransferase activities appeared to be localized in smooth microsomal membrane and the Golgi complex derivatives and not associated with the synaptosomes.

Localization of some glycolipid glycosylating enzymes in the Golgi apparatus of rat kidney

Cell fractions from rat kidney were isolated and studied for their ability to synthesize several possible intermediates in the biosynthesis of sulfatides and gangliosides. The enzymes studied include UDP-Gal:ceramide galactosyltransferase, UDP-Gal:glucosylceramide galactosyltransferase, UDP-Gal:galactosylceramide galactosyltransferase, and CMP-NAN:lactosylceramide sialytransferase activities. The initial glycosylation of ceramide was found to be present in all of the kidney cell fractions studied. The remaining glycosylating enzymes were largely localized in the Golgi apparatus of kidney. Thus, in addition to modifying glycoproteins for secretion, the Golgi apparatus in kidney is involved in the modification of a number of glycolipids which are destined to form cell membrane components.