Oligosialogangliosides inhibit GM2- and GD3-synthesis in isolated Golgi vesicles from rat liver

The effect of end-product gangliosides (GD1a, GT1b, GQ1b) on the activities of two key enzymes in ganglioside biosynthesis, namely GM2-synthase and GD3-synthase in rat liver Golgi apparatus, has been investigated in detergent-free as well as in detergent-containing assays. In detergent-free intact Golgi vesicles, phosphatidylglycerol was used as a stimulant. This phospholipid was earlier shown to stimulate the activity of GM2-synthase without disrupting the vesicular intactness; it has, however, no effect on GD3-synthase (Yusuf, H.K.M., Pohlentz, G., Schwarzmann, G. & Sandhoff, K. (1983) Eur. J. Biochem. 134, 47-54). In the presence of this stimulant, all higher gangliosides inhibited the activity of GM2-synthase, the inhibition being more profound with increasing negative charge of the inhibiting gangliosides. These inhibitions are unspecific, but they do not exclude an end-product regulation of ganglioside biosynthesis. In detergent-solubilized Golgi membranes, on the other hand, the inhibition pattern was completely different. Here, ganglioside GD1a was the strongest inhibitor of GM2-synthase, followed by GM1 and GM2, but GT1b also inhibited this enzyme appreciably, in fact more strongly than GM1 or GM2. On the other hand, GQ1b had no effect at all. Conversely, GD3-synthase activity was most strongly inhibited by GQ1b, followed by GT1b, but GD1a also inhibited this enzyme almost as strongly as GT1b. These latter findings indicate that feed-back control of the a- and the b-series pathways of ganglioside biosynthesis is probably not specific, but the pathways appear to be inhibited more preferably by their respective end-products than by any other gangliosides of the same of the other series.     

Identity of GA1, GM1a and GD1b synthase in Golgi vesicles from rat liver

Synthesis of ganglioside GD1b from ganglioside GD2 was demonstrated using Golgi membranes isolated from rat liver. Competition experiments using gangliosides GA2, GM2 and GD2 as substrates, and as mutual inhibitors for ganglioside galactosyltransferase activity in preparations of Golgi vesicles derived from rat liver, suggested that galactosyl transfer to these three compounds, leading to gangliosides GA1, GM1a and GD1b respectively, is catalyzed by one enzyme. These results strengthen the hypothesis that the main site for the regulation of ganglioside biosynthesis occurs within the reaction sequence LacCer----GA3----GD3----GT3.     

Identity of GD1C, GT1a and GQ1b synthase in Golgi vesicles from rat liver

Competition experiments using GM1b, GD1a and GT1b as substrates, and as mutual inhibitors for ganglioside sialyltransferase activity in preparations of Golgi vesicles derived from rat liver, suggested that sialyl transfer to these three respective compounds, leading to gangliosides GD1C, GT1a and GQ1b, respectively, is catalyzed by one enzyme. These results are incorporated into a model for ganglioside biosynthesis and its regulation.     

Purification to homogeneity of GD3 synthase and partial purification of GM3 synthase from rat brain

A CMP-sialic acid: GM3 sialyltransferase (GD3 synthase) and a CMP-sialic acid: LacCer sialyltransferase (GM3 synthase) have been purified 10,000- and 3,000-fold, respectively, from the Triton X-100 extract of rat brain. The two enzymes were purified and resolved by affinity chromatography on two successive CDP-Sepharose columns by NaCl gradient elution. Final purification of GD3 synthase was achieved by specific elution from a 'GM3 acid'-Sepharose column with buffer containing GM3. Sodium dodecylsulfate-gel electrophoresis of GD3 synthase revealed a single major protein band with an apparent molecular weight of 55,000.     

Purification and properties of GM2 synthase, UDP-N-acetylgalactosamine: GM3 beta-N-acetylgalactosaminyltransferase from rat liver

A UDP-N-acetylgalactosamine:ganglioside GM3 beta-N-acetylgalactosaminyltransferase which catalyzes the conversion of ganglioside GM3 to GM2 has been purified over 6300-fold from a Triton X-100 extract of rat liver particulate fractions by hydrophobic chromatography and affinity chromatography on GM3-acid-Sepharose. The purified enzyme has two identical subunits of 64,000 daltons. The enzyme has a pH optimum of pH 6.7-6.9 and requires divalent cations such as Mn2+ and Ni2+. In studies on substrate specificity GM3 containing N-acetylneuraminic acid (GM3(NeuAc] and GM3 containing N-glycolylneuraminic acid were both good acceptors for the purified enzyme. The plots of the activity of transferase as a function of GM3(NeuAc) showed sigmoidal relationships. The oligosaccharide of GM3, sialyllactose, was also a good acceptor, which indicates that the preferred acceptor substrate has the possible structure NeuAc alpha 2- or NeuGc alpha 2-3 Gal beta 1-4Glc-OR.     

Translocation of UDP-N-acetylglucosamine into vesicles derived from rat liver rough endoplasmic reticulum and Golgi apparatus

A mixture of UDP-N-acetylglucosamine labeled with different radioisotopes in the uridine and glucosamine was used to show that the intact sugar nucleotide was translocated across the membrane of vesicles derived from rat liver rough endoplasmic reticulum (RER) and Golgi apparatus. Translocation was dependent on temperature, saturable at high concentrations of sugar nucleotide, and inhibited by treatment of vesicles with proteases, suggesting protein carrier mediated transport. Translocation of UDP-GlcNAc by RER-derived vesicles appeared to be specific since these vesicles were unable to translocate UDP-galactose, in contrast to those derived from the Golgi apparatus. Preliminary results suggest that the mechanism of UDP-GlcNAc translocation into RER-derived vesicles is via a coupled exchange with lumenal nucleoside monophosphate. This is similar to the recently postulated mechanism for translocation of sugar nucleotides into vesicles derived from the Golgi apparatus.     

Orientation of glycoprotein galactosyltransferase and sialyltransferase enzymes in vesicles derived from rat liver Golgi apparatus

UDP-galactose: N-acetylglucosamine galactosyltransferase (GT) and CMP- sialic:desialylated transferrin sialyltransferse (ST) activities of rat liver Golgi apparatus are membrane-bound enzymes that can be released by treatment with Triton X-100. When protein substrates are used to assay these enzymes in freshly prepared Golgi vesicles, both activities are enhanced about eightfold by the addition of Triton X-100. When small molecular weight substrates are used, however, both activities are only enhanced about twofold by the addition of detergent. The enzymes remain inaccessible to large protein substrates even after freezing and storage of the Golgi preparation for 2 mo in liquid nitrogen. Accessibility to small molecular and weight substrates increases significantly after such storage. GT and ST activities in Golgi vesicles are not destroyed by treatment with trypsin, but are destroyed by this treatment if the vesicles are first disrupted with Triton X-100. Treatment of Golgi vesicles with low levels of filipin, a polyene antibiotic known to complex with cholesterol in biological membranes, also results in enhanced trypsin susceptibility of both glycosyltransferases. Maximum destruction of the glycosyltransferase activities by trypsin is obtained at filipin to total cholesterol weight ratios of approximately 1.6 or molar ratios of approximately 1. This level of filipin does not solubilize the enzymes but causes both puckering of Golgi membranes visible by electron microscopy and disruption of the Golgi vesicles as measured by release of serum albumin. When isolated Golgi apparatus is fixed with glutaraldehyde to maintain the three-dimensional orientation of cisternae and secretory vesicles, and then treated with filipin, cisternal membranes on both cis and trans faces of the apparatus as well as secretory granule membranes appear to be affected about equally. These results indicate that liver Golgi vesicles as isolated are largely oriented with GT and ST on the luminal side of the membranes, which corresponds to the cisternal compartment of the Golgi apparatus in the hepatocyte. Cholesterol is an integral part of the membrane of the Golgi apparatus and its distribution throughout the apparatus is similar to that of both transferases.     

Cholesterol-containing lactose derived neoglycolipids serve as acceptors for sialyltransferases from rat liver Golgi vesicles

The cholesterol-containing lactose derived neoglycolipids -Lactosylcholesterol, Cholesteryl--lactosylpropane-1,3-diol, 3-Cholesteryl-1--lactosylglycerol, 2-Cholesteryl-1--lactosylglycerol, 2,3-Dicholesteryl-1--lactosylglycerol, 1-Deoxy-1-cholesterylethanolaminolactitol, 1-Deoxy-1-cholesteryl (N-acetyl)-ethanolaminolactitol, 1-Deoxy-1-cholesterylphosphoethanolaminolactitol, and 1-Deoxy-1-cholesterylphospho (N-acetyl)-ethanolaminolactitol were synthesized and used as acceptors for sialytransferases from rat liver Golgi vesicles. Relative activities with the neoglycolipids as acceptors varied from 28 to 163% compared to those obtained with the authentic acceptor lactosylceramide. Product identification by thin layer chromatography and fast atom bombardment mass spectrometry showed that the neoglycolipids yielded mono- and disialylated products. The results of competition experiments suggested that lactosylceramide and the neoglycolipids were sialylated by the same enzymes.     

Compartmental Organization of the Synthesis of GM3, GD3, and GM2 in Golgi Membranes from Neural Retina Cells

The relationship among lactosylceramide-(LacCer), GD3- and GM2-synthases and between the two last transferases and their common GM3 acceptor was investigated in intact Golgi membrane from chick embryo neural retina cells at early (8-days) and late (14 days) stages of the embryonic development. [³H]Gal was incorporated into endogenous glucosylceramide by incubation of Golgi membranes with UDP-[³H]Gal. Conversion of the synthesized [³H]Gal-LacCer into GM3, and of the latter into GD3, GM2 and GD2 was examined after a second incubation step with unlabeled CMP-NeuAc and/or UDP-GalNAc. With CMP-NeuAc, most [³H]Gal-LacCer was converted into GM3 in either 8- or 14- day membranes. However, while about 90% of GM3 was converted into GD3 in 8-day membranes, only about 25% followed this route in 14-day membranes. With CMP-NeuAc and UDP-GalNAc, about 90% of GM3 was used for synthesis of GM2 in 14-day membranes, while in 8-day membranes about 80% followed the route to GD3, and a part to GD2. Performing the second incubation step in the presence of increasing detergent concentrations showed that conversion of GM3 to GM2 was inhibited at concentrations lower than those required for inhibition of LacCer to GM3 conversion. Taken together, results indicate that transfer steps leading to synthesis of GM3, GD3, GM2 and GD2 from LacCer are functionally coupled in the Golgi membranes, and that GD3- and GM2-synthases compete in a common compartment for using a fraction of GM3 as substrate. In this competition, the relative activities of the transferases and their relative saturation with the respective donor sugar nucleotides, are important factors influencing conversion of GM3 toward either GD3 or GM2.     

Precursor-product relationship between GM1 and GD1a biosynthesized from exogenous GM2 ganglioside in rat liver

The demonstration of a precursor-product relationship in the course of GM1 and GD1a biosynthesis is described in the present paper. We injected rats with GM2 gangliosides [GalNAc beta 1----4(NeuAc alpha 2----3)Gal beta 1----4Glc beta 1----1'Cer] of brain origin, which were isotopically radiolabeled on the GalNAc ([GalNAc-3H]GM2) or sphingosine ([Sph-3H]GM2) residue. We then compared the time-courses of GM1 and GD1a biosynthesis in the liver after the administration of each radiolabeled GM2 derivative. After the administration of [GalNAc-3H]GM2, GM1, and GD1a were both present as doublets, that could be easily resolved on TLC. The lower spot of each doublet was identified as a species having the typical rat brain ceramide moiety and represented gangliosides formed through direct glycosylation of the injected GM2. The upper spot of each doublet was identified as a species having the typical rat liver ceramide moiety and represented gangliosides formed through recycling of the [3H]GalNAc residue, released during ganglioside catabolism. After the administration of [Sph-3H]GM2, only ganglioside with the rat brain ceramide moiety were found, that represented the sum of ganglioside formed through direct glycosylation and those formed through recycling of some sphingosine-containing fragments. In each case, the time-course of GM1 and GD1a biosynthesis exhibited a precursor-product relationship. The curve obtained from the direct glycosylation showed a timing delay with respect to those obtained from recycling of GM2 fragments. These results are consistent with the hypothesis that the sequential addition of activated sugars to a sphingolipid precursor is a dissociative process, catalyzed by physically independent enzymatic activities.     

Substrate specificity of acetyl-CoA-carboxylase from the rat liver

The interaction between acetyl-CoA fragments and rat liver acetyl-CoA carboxylase was studied. It was found that the 3'-phosphate group did not interfere with the enzyme interaction since the substrate properties of acetyl-dephospho-CoA and acetyl-CoA are nearly identical. The non-nucleotide substrate analogs S-acetyl-pantethin and its 4'-phosphate) also displayed substrate properties (V = 1.5% and 15% of the V for acetyl-CoA carboxylation respectively). The nucleotide fragment of the acetyl-CoA molecule produced an appreciable effect on the thermodynamics of this substrate interaction with the enzyme. Its physiological role consists in all probability, in the activation and propes orientation of the acetyl group in the enzyme active center. The far more pronounced substrate properties of S-acetyl pantethin 4'-phosphate and the inhibitory properties of pantethin 4'-phosphate (compared to non-phosphorylated analogs) suggest the essential role of the beta-phosphate residue of ADP in the acetyl-CoA binding to the enzyme. The data obtained suggest also that the hydrophobic region responsible for the acyl radical binding, has a site which specifically recognizes the beta-mercaptoethyl residue of the CoA pantethin fragment. The pivotal role in the acetyl-CoA carboxylase interaction with the substrate is ascribed to the productive binding of the acetyl radical; the contribution of individual fragment of the CoA molecule is variable.     

Biosynthesis of gangliosides from asialogangliosides in rat liver Golgi vesicles

Biosynthesis of glycolipids GA2, GA1, GM1b, and GD1c was studied in Golgi vesicles isolated from rat liver. Sequential addition of N-acetylgalactosamine, galactose and two sialic acid residues to lactosylceramide led to the endproduct GD1c. Activities of the corresponding glycosyltransferases were shown to be present in isolated Golgi vesicles and their respective kinetic data were determined. The products of each reaction were characterized by their mobility on thin-layer chromatography, by enzymic degradation to their respective precursors, and in case of GM1b by FAB mass spectrometry.     

Substrate specificity of liver calmodulin-dependent glycogen synthase kinase

A number of proteins were tested as potential substrates for purified rabbit liver calmodulin-dependent glycogen synthase kinase. It was found that liver phenylalanine hydroxylase and several brain proteins including tyrosine hydroxylase, microtubule-associated protein 2, and synapsin I were readily phosphorylated. Brain tubulin was very poorly phosphorylated. These results suggest that calmodulin-dependent glycogen synthase kinase may be a more general protein kinase involved in the regulation of several cellular Ca²⁺-dependent functions.     

Translocation of adenosine 3'-phosphate 5'-phosphosulfate into rat liver Golgi vesicles

The translocation of adenosine 3'-phosphate 5'-phosphosulfate (PAPS) across rat liver Golgi-derived vesicles has been studied. Vesicles of the same topographical orientation as in vivo were incubated with a mixture of [adenine-8-3H]PAPS and [35S]PAPS. The tritium to radiolabeled sulfur ratio of the incubation medium was 1.73 +/- 0.03 while that in the vesicles was 1.82 +/- 0.13. This strongly suggests that the entire PAPS molecule was being translocated across the Golgi vesicle membrane even though intact PAPS could not be detected within the vesicles. Translocation of PAPS resulted in accumulation of solutes within vesicles. This accumulation was temperature dependent, saturable (apparent Km = 0.7 microM; Vmax = 25 pmol/mg of protein/10 min), and inhibited by the substrate analogue 3',5'-ADP but not by 2',5'-ADP. Translocation of PAPS was inhibited following treatment of Golgi vesicles with Pronase under conditions in which the activity of a lumenal Golgi membrane marker such as sialyltransferase was not. This result is consistent with the existence of a PAPS carrier protein, portions of which face the cytoplasmic side of the Golgi membrane.     

Chemoenzymatic synthesis of 2-azidoethyl-ganglio-oligosaccharides GD3, GT3, GM2, GD2, GT2, GM1, and GD1a

We have synthesized several ganglio-oligosaccharide structures using glycosyltransferases from Campylobacter jejuni. The enzymes, alpha-(2-->3/8)-sialyltransferase (Cst-II), beta-(1-->4)-N-acetylgalactosaminyltransferase (CgtA), and beta-(1-->3)-galactosyltransferase (CgtB), were produced in large-scale fermentation from Escherichia coli and further characterized based on their acceptor specificities. 2-Azidoethyl-glycosides corresponding to the oligosaccharides of GD3 (alpha-D-Neup5Ac-(2-->8)-alpha-D-Neup5Ac-(2-->3)-beta-D-Galp-(1-->4)-beta-D-Glcp-), GT3 (alpha-D-Neup5Ac-(2-->8)-alpha-D-Neup5Ac-(2-->8)-alpha-D-Neup5Ac-(2-->3)-beta-D-Galp-(1-->4)-beta-D-Glcp-), GM2 (beta-D-GalpNAc-(1-->4)-[alpha-D-Neup5Ac-(2-->3)]-beta-D-Galp-(1-->4)-beta-D-Glcp-), GD2 (beta-D-GalpNAc-(1-->4)-[alpha-D-Neup5Ac-(2-->8)-alpha-D-Neup5Ac-(2-->3)]-beta-D-Galp-(1-->4)-beta-D-Glcp-), GT2 (beta-D-GalpNAc-(1-->4)-[alpha-D-Neup5Ac-(2-->8)-alpha-D-Neup5Ac-(2-->8)-alpha-D-Neup5Ac-(2-->3)]-beta-D-Galp-(1-->4)-beta-D-Glcp-), and GM1 (beta-D-Galp-(1-->3)-beta-D-GalpNAc-(1-->4)-[alpha-D-Neup5Ac-(2-->3)]-beta-D-Galp-(1-->4)-beta-D-Glcp-) were synthesized in high yields (gram-scale). In addition, a mammalian alpha-(2-->3)-sialyltransferase (ST3Gal I) was used to sialylate GM1 and generate GD1a (alpha-D-Neup5Ac-(2-->3)-beta-D-Galp-(1-->3)-beta-D-GalpNAc-(1-->4)-[alpha-D-Neup5Ac-(2-->3)]-beta-D-Galp-(1-->4)-beta-D-Glcp-) oligosaccharide. We also cloned and expressed a rat UDP-N-acetylglucosamine-4'epimerase (GalNAcE) in E. coli AD202 cells for cost saving in situ conversion of less expensive UDP-GlcNAc to UDP-GalNAc.     

The activity of GD3 synthase modulates the ganglioside pattern in rat liver

Variations of the ganglioside composition in the livers of Wistar rats correlated with the activity of GD3 synthase in the corresponding liver homogenates. With increasing enzyme activity, higher proportions of b-series gangliosides (GD3, GD1b, GT1b) were detected. No significant changes in the activity of GM2 synthase or GM1 synthase were observed, indicating a regulatory function for GD3 synthase in this tissue. Young animals showed an average GD3 synthase activity of 0.5-0.6 nmol.h-1.mg protein-1 without sex-dependent variations. Among the older animals, however, males expressed an activity five-fold higher than females, suggesting that this enzyme might be affected by hormones.     

Substrate specificities of a bacterial sialidase and rat liver ganglioside GM3 sialyltransferase

Sialidases cleave off sialic acid residues from the oligosaccharide chain of gangliosides in their catabolic pathway while sialyltransferases transfer sialic acid to the growing oligosaccharide moiety in ganglioside biosynthesis. Ganglioside G_M3 is a common substrate for both types of enzymes, for sialidase acting on ganglioside G_M3 as well as for ganglioside G_D3 synthase. Therefore, it is possible that both enzymes recognize similar structural features of the sialic acid moiety of their common substrate, ganglioside G_M3. Based on this idea we used a variety of G_M3 derivatives as glycolipid substrates for a bacterial sialidase (Clostridium perfringens) and for G_D3 synthase (of rat liver Golgi vesicles). This study revealed that those G_M3 derivatives that were poorly degraded by sialidase also were hardly recognized by sialyltransferase (G_D3 synthase). This may indicate similarities in the substrate binding sites of these enzymes.     

Purification and properties of two enzymes catalyzing galactose transfer to GM2 ganglioside from rat liver Golgi

An enzyme activity which catalyzed the transfer of galactose from UDP-galactose to GM2 ganglioside was demonstrated in rat liver homogenate and enriched 38-fold in specific activity by preparation of Golgi membranes. This activity could be solubilized from Golgi membranes by sonication and extraction with 1% Triton X-100. The solubilized activity catalyzed the formation of GM1 ganglioside and was completely dependent upon the addition of acceptor. The rate of galactose incorporation was constant for up to 5 h at 30 degrees C. This enzyme activity was further purified by gel filtration on Sepharose CL-6B and ion exchange chromatography on DEAE-Sepharose. The elution position on gel filtration corresponded to a molecular weight for the enzyme of 38,000 which was in good agreement with that obtained by sedimentation velocity studies. Ion exchange chromatography resolved GM2 ganglioside galactosyltransferase into two species. The more basic enzyme (I) comprising 28% of the recovered activity was not retarded by the column, whereas enzyme II was eluted from the resin following the application of a salt gradient. Net purification was 120- to 140-fold for each enzyme with a total recovery of 42%. Unlike the activity in the Golgi extract, the purified enzymes I and II were labile to freezing and could be stored at -20 degrees C only in the presence of 50% glycerol. Both enzymes I and II had similar molecular weights and Michaelis constants and both had a strict requirement for Mn2+. Properties which distinguish the two enzymes included pH optima (enzyme I 7.0, enzyme II 6.0) and surfactant requirements. Neither enzyme was active following removal of Triton X-100 from the preparation. Among a series of glycolipids tested for ability to serve as substrates for galactose transfer only GM2 and asialo-GM2 ganglioside served as acceptors.     

Disruption of GM2/GD2 synthase gene resulted in overt expression of 9-O-acetyl GD3 irrespective of Tis21

GM2/GD2 synthase gene knockout mice lack all complex gangliosides, which are abundantly expressed in the nervous systems of vertebrates. In turn, they have increased precursor structures GM3 and GD3, probably replacing the roles of the depleted complex gangliosides. In this study, we found that 9-O-acetyl GD3 is also highly expressed as one of the major glycosphingolipids accumulating in the nervous tissues of the mutant mice. The identity of the novel component was confirmed by neuraminidase treatment, thin layer chromatography-immunostaining, two-dimensional thin layer chromatography with base treatment, and mass spectrometry. All candidate factors reported to be possible inducer of 9-O- acetylation, such as bitamine D binding protein, acetyl CoA transporter, or O-acetyl ganglioside synthase were not up-regulated. Tis21 which had been reported to be a 9-O-acetylation inducer was partially down-regulated in the null mutants, suggesting that Tis21 is not involved in the induction of 9-O-acetyl-GD3 and that accumulated high amount of GD3 might be the main factor for the dramatic increase of 9-O-acetyl GD3. The ability to acetylate exogenously added GD3 in the normal mouse astrocytes was examined, showing that the wild-type brain might be able to synthesize very low levels of 9-O-acetyl GD3. Increased 9-O-acetyl GD3, in addition to GM3 and GD3, may play an important role in the compensation for deleted complex gangliosides in the mutant mice.     

GT3 Synthesis in the Proximal Golgi Occurs in a Compartment Different from Those for GD3 and GM3 Synthesis

Abstract: Previous studies from this laboratory have shown that synthesis of GT3, the precursor of c series gangliosides, occurs in proximal Golgi compartments, as has been shown for the synthesis of GM3 and GD3, the precursors of a and b series gangliosides, respectively. In this work we studied whether the synthesis of GM3, GD3, and GT3 occurs in the same or in different compartments of the proximal Golgi. For this, we examined in retina cells (a) the effect of monensin, a sodium ionophore that affects mostly the trans Golgi and the trans Golgi network function, on the metabolic labeling of glycolipids from [³H]Gal by cultured cells from 7- and 10-day chick embryos and (b) the labeling in vitro of endogenous glycolipids of Golgi membrane preparations from 7-day embryos incubated with UDP-[³H]Gal. In (a), 1 µM monensin produced a twofold accumulation of radioactive glucosylceramide and a decrease to ～50 and 20% of total ganglioside labeling in 7- and 10-day cells, respectively. At both ages, monensin produced a threefold accumulation of radioactive GM3 and an inhibition of >90% of GT3, GM1, GD1a, and GT1b synthesis. GD3 synthesis was inhibited ～30 and 70%, respectively, in 7- and 10-day cells. In (b), >80% of the [³H]Gal was incorporated into endogenous glucosylceramide to form radioactive lactosylceramide. About 90% of [³H]Gal-labeled lactosylceramide was converted into GM3, and most of this in turn into GD3 when unlabeled CMP-NeuAc was also present in the incubation system. Under the same conditions, however, <5% of labeled GD3 was converted into GT3. Golgi membranes incubated with CMP-[³H]NeuAc incorporated ～20% of [³H]NeuAc into endogenous GT3, and this percentage was not affected by 1 µM monensin. These results indicate that synthesis of GT3 is carried out in a compartment of the proximal Golgi different from those for lactosylceramide, GM3, and GD3 synthesis. Results from the experiments with monensin point to the cis/medial Golgi as the main compartment for coupled synthesis of lactosylceramide, GM3, and GD3 and to the trans Golgi as the main compartment for synthesis of GT3.