Regulated translocation of neutral sphingomyelinase-2 to the plasma membrane drives insulin resistance in steatotic hepatocytes.

Obesity-associated diabetes is linked to the accumulation of ceramide in various organs, including the liver. The exact mechanisms by which ceramide contributes to diabetic pathology are unclear, but one proposed scenario is that ceramide accumulation may inhibit insulin signaling pathways. It is unknown however whether the excess ceramide is generated proximal to the insulin receptor, that is, at the plasma membrane (PM), where it could affect the insulin signaling pathway directly, or the onset of insulin resistance is due to ceramide-induced mitochondrial dysfunction and/or lipotoxicity. Using hepatic cell lines and primary cultures, gain- and loss- of function approach, and state-of-the art lipid imaging, this study shows that PM-associated neutral sphingomyelinase 2 (nSMase2) regulates ceramide homeostasis in fat-loaded hepatocytes and drives the onset of insulin resistance. Our results provide evidence of a regulated translocation of nSMase2 to the PM which leads to local generation of ceramide and insulin resistance in cells treated with palmitic acid (PAL), a type of fat commonly found in diabetogenic diets. Oleic acid, which also causes accumulation of lipid droplets, does not induce nSMase2 translocation and insulin resistance. Experiments using the acyl-biotin exchange method to quantify protein palmitoylation show that cellular PAL abundance regulates the rate of nSMase2 palmitoylation. Furthermore, while inhibition of nSMase2 with GW4869 prevents PAL-induced insulin resistance, the overexpression of wild type nSMase2 but not palmitoylation-defective mutant protein potentiates the suppressive effect of PAL on insulin signaling. Overall, this study identifies nSMase2 as a novel component of the mechanism of insulin resistance onset in fat-loaded hepatocytes, that is, cell-autonomous and driven by PAL.

Incorporation and visualization of azido-functionalized N-oleoyl serinol in Jurkat cells, mouse brain astrocytes, 3T3 fibroblasts and human brain microvascular endothelial cells.

The synthesis and biological evaluation of azido-N-oleoyl serinol is reported. It mimicks biofunctional lipid ceramides and has shown to be capable of click reactions for cell membrane imaging in Jurkat and human brain microvascular endothelial cells.

Multi-enzyme kinetic analysis of glycolipid biosynthesis.

Gangliosides are acidic glycosphingolipids synthesized sequentially by a series of glycosyltransferases acting in parallel biosynthetic pathways. While most glycosyltransferases are highly specific, some, however, may catalyze equivalent steps in each pathway using different gangliosides as substrates (e.g. N-acetylgalactosaminyltransferase, sialyltransferase-IV). A multi-enzyme kinetic analysis was developed on the condition that serial enzymatic reactions operate below substrate saturation. A multi-enzyme kinetic analysis enabled a simultaneous calculation of the Vmax/Km value of each enzyme derived from the equilibrium concentration of the respective substrate. Substrate concentrations [S] were determined by radioactive labelling of gangliosides in intact cells with the precursor sugars [14C]galactose and [14C]glucosamine, followed by high-performance thin-layer chromatography and autoradiography of the radiolabelled glycolipids. On the basis of Michaelis-Menten kinetics, Vmax/Km values were derived from [S] by a system of linear equations. The procedure was used to analyze the development of the glycolipid composition during differentiation of rat gliomaxmurine neuroblastoma (NG108-15) cells. The Vmax/Km values calculated by multi-enzyme kinetic analysis were consistent with the kinetic data obtained with solubilized enzymes. Application of multi-enzyme kinetic analysis to published data on the correlation of enzyme activities with ganglioside levels in various cell lines and tissues indicated the validity of this method for analysis of the glycolipid biosynthesis, in particular, of its initial steps. On the basis of the kinetic analysis, it is suggested that the cell lines can be divided into two groups with respect to the substrate pools of GM3 used by sialyltransferase-II and N-acetylgalactosaminyltransferase-I. The first group encompasses the majority of the neuroblastoma cell lines and the embryonic rat brain where the two enzymes share a common pool of GM3. In the second group, the two enzymes do not compete for the same pool of GM3, indicating a different subcellular localization of CMP-NeuAc:GM3 alpha2-8-sialyltransferase and UDP-N-acetylgalactosaminyl:GM3 N-acetylgalactosaminyltransferase. In this study, the theory of a multi-enzyme kinetic analysis is discussed and its application to analysis of the glycolipid biosynthesis in neuroblastoma cells is demonstrated. A multi-enzyme kinetic analysis can be applied to other biosynthetic pathways and provides the advantage of analyzing kinetic data with intact cells or tissue samples.

Regulation of glycosyltransferases in ganglioside biosynthesis by phosphorylation and dephosphorylation.

The biosynthesis of gangliosides is known to be under strict metabolic control. One level of control is through post-translational modification of the glycosyltransferases responsible for their biosynthesis. Thus, the activities of several sialyltransferases have been demonstrated to be downregulated by the action of protein kinase C (PKC) in cell-free and intact cell systems. This modulatory effect can be reversed at least in part by the action of membrane-bound phosphatases. In contrast, the activity of N-acetylgalactosaminyltransferase can be upregulated by the action of protein kinase A (PKA) in cultured cells. In addition, studies from several laboratories have demonstrated that phosphorylation of certain glycosyltransferases can affect their intracellular processing and translocation. Thus, modulation of glycosyltransferases by phosphorylation and dephosphorylation should represent an important regulatory mechanism for ganglioside biosynthesis.

Probing quantum coherence in a biological system by means of DNA amplification.

As a result of rapid decoherence, quantum effects in biological systems are usually confined to single electron or hydrogen delocalizations. In principle, molecular interactions at high temperatures can be guided by quantum coherence if embedded in a dynamics preventing decoherence. This was experimentally investigated by analyzing the thermodynamics, kinetics, and quantum mechanics of the primer/template duplex formation during DNA amplification by polymerase chain reaction. The structures of the two oligonucleotide primers used for amplification of a cDNA template were derived either from a repetitive motif or a fractal distribution of nucleotide residues. Contrary to the computer-based calculation of the primer melting temperatures (T(m)) that predicted a higher T(m) for the non-fractal primer due to nearest-neighbor effects, it was found that the T(m) of the non-fractal primer was actually 2 degrees C lower than that of its fractal counterpart. A thermodynamic analysis of the amplification reaction indicated that the primer annealing process followed Bose-Einstein instead of Boltzmann statistics, with an additional binding potential of mu=500 J/mol or 10(-21) J/molecule due to a superposition of binding states within the primer/template duplex. The temporal evolution of the Bose-Einstein state was determined by enzyme kinetic analysis of the association of the primer/template duplex to Taq polymerase. Assuming that collision with the enzyme interrupted the superposition, it was found that the Bose-Einstein state lasted for t(dec)=0.7x10(-12) s, corresponding to the energy dispersion (DeltaE) of quantum coherent states (mu=DeltaE>/=h/t(dec)). A quantum mechanical analysis revealed that the coherent state was stabilized by almost vanishing separation energies between distinct binding states during a temperature-driven shifting of the two DNA strands in the primer/template duplex. The additional binding potential is suggested to arise from a short-lived electron tunneling as the result of overlapping orbitals along the axis of the primer/template duplex. This effect was unique to the fractal primer due to the number of binding states that remained almost constant, irrespective of the size of shifting. It is suggested that fractal structures found in proteins or other macromolecules may facilitate a short-lived quantum coherent superposition of binding states. This may stabilize molecular complexes for rapid sorting of correct-from-false binding, e.g. during folding or association of macromolecules. The experimental model described in this paper provides a low-cost tool for simulating and probing quantum coherence in a biological system.

Prenatal alcohol exposure triggers ceramide-induced apoptosis in neural crest-derived tissues concurrent with defective cranial development.

Fetal alcohol syndrome (FAS) is caused by maternal alcohol consumption during pregnancy. The reason why specific embryonic tissues are sensitive toward ethanol is not understood. We found that in neural crest-derived cell (NCC) cultures from the first branchial arch of E10 mouse embryos, incubation with ethanol increases the number of apoptotic cells by fivefold. Apoptotic cells stain intensely for ceramide, suggesting that ceramide-induced apoptosis mediates ethanol damage to NCCs. Apoptosis is reduced by incubation with CDP-choline (citicoline), a precursor for the conversion of ceramide to sphingomyelin. Consistent with NCC cultures, ethanol intubation of pregnant mice results in ceramide elevation and increased apoptosis of NCCs in vivo. Ethanol also increases the protein level of prostate apoptosis response 4 (PAR-4), a sensitizer to ceramide-induced apoptosis. Prenatal ethanol exposure is concurrent with malformation of parietal bones in 20% of embryos at day E18. Meninges, a tissue complex derived from NCCs, is disrupted and generates reduced levels of TGF-ß1, a growth factor critical for bone and brain development. Ethanol-induced apoptosis of NCCs leading to defects in the meninges may explain the simultaneous presence of cranial bone malformation and cognitive retardation in FAS. In addition, our data suggest that treatment with CDP-choline may alleviate the tissue damage caused by alcohol.

A novel isoform of prostate apoptosis response 4 (PAR-4) that co-distributes with F-actin and prevents apoptosis in neural stem cells.

The elevated expression of prostate apoptosis response-4 (PAR-4) induces apoptosis in differentiating mouse embryonic stem (ES) cells. In embryoid body (EB) cells and the E15.5 stage of embryonic mouse brain, PAR-4 is expressed as two isoforms (38 and 33 kDa). Using mouse EB-derived RNA as a template we have cloned and characterized a novel isoform of PAR-4 (PAR-4/p33) that lacks exon 3 and shows a bona fide splice junction of exons 2 and 4. The molecular mass for PAR-4/p33 is estimated to be 33 kDa, corresponding to the short form found in the EB cells and E15.5 mouse brain. The fluorescent fusion protein of PAR-4/p33 is mainly found in the cytosol and is co-distributed with F-actin filaments, while that of the 38 kDa full length PAR-4/p38 is predominantly translocated to the nucleus. In contrast to the full length PAR-4 (PAR-4/p38), ectopic expression of PAR-4/p33 does not result in the activation of caspase 3 and the induction of apoptosis. PAR-4/p33 forms a complex with PAR-4/p38, which inhibits its nuclear translocation and the induction of apoptosis. PAR-4/p33 is suggested to be a dominant negative isoform of PAR-4/p38 and may regulate PAR-4-dependent apoptosis.

Isolation of a cytosolic beta-glucosidase from calf liver and characterization of its active site with alkyl glucosides and basic glycosyl derivatives.

A beta-glucosidase/beta-galactosidase with Mr 52,500 was isolated from calf liver cytosol by a four-step procedure incorporating affinity chromatography on N-(9-carboxynonyl)-deoxynojirimycin-AH-Sepharose. Its pH optimum was at 5.8 with half-maximal activity at pH 3.5 and 8.6. Affinity for gluco compounds expressed by Km or Ki of substrates and inhibitors was 2- to 10-fold higher than for the corresponding galacto compounds. Alkyl glucosides were hydrolyzed with lower Vmax than p-nitrophenyl and 4-methylumbelliferyl glucosides, but due to their higher affinity the alkyl glucosides displayed values for kcat/Km of the same magnitude of the aryl glucosides when the alkyl chains were longer than octyl. Glucosylsphingosine was bound with Ki (= Km) 2.2 microM and hydrolyzed with a Vmax that was 50-fold lower than the Vmax for 4-methylumbelliferyl beta-glucoside. The product sphingosine was inhibitory with Ki 0.30 microM. A systematic study with alkyl glucosides and glucosylamines defined the aglycon site as a narrow, strongly hydrophobic cleft able to accommodate up to 10 methylene groups. Each CH2 group contributed 3.1 kJ/mol to the standard free energy of binding. The inhibition by gluco- and galactosylamine and by 1-deoxynojirimycin and its D-galacto analog was approximately 200-fold better than by corresponding nonbasic compounds. pH dependence of the inhibition and comparison with permanently cationic glycosyl derivatives showed that the nonprotonated form was the inhibiting species. This feature puts the cytosolic beta-glucosidase in the large class of glycoside hydrolases which strongly bind basic glycosyl derivatives by their protonation at the active site and formation of a shielded ion pair with the carboxylate of an aspartic or glutamic side chain.

Active site directed inhibition of a cytosolic beta-glucosidase from calf liver by bromoconduritol B epoxide and bromoconduritol F.

Hydrolysis of p-nitrophenyl-beta-D-glucoside by cytosolic beta-glucosidase proceeds with retention of the anomeric configuration. Whereas inactivation of the enzyme by the glucosidase inhibitor conduritol B epoxide (CBE) was extremely slow (ki(max)/Ki 0.57 M-1 min-1) it reacted 130 times more rapidly with 6-bromo-6-deoxy-CBE (Br-CBE). The beta-glucosidase could be labeled with [3H]Br-CBE; incorporation of 1 mol inhibitor/mol enzyme resulted in complete loss of activity. Most of the bound inhibitor was released after denaturation and treatment with ammonia as (1,3,4/2,5,6)-6-bromocyclohexanepentol, thus demonstrating the formation of an ester bond with an active site carboxylate by trans-diaxial opening of the epoxide ring. It was concluded from the Ki values for the epoxide inhibitors and for coduritol B with the cytosolic enzyme and corresponding data for the lysosomal beta-glucosidase that the unusually low reactivity with CBE and Br-CBE is probably due to the inability of the cytosolic enzyme to effectively donate a proton to the epoxide oxygen. An extremely rapid inactivation of the cytosolic beta-glucosidase was caused by bromoconduritol F ((1,2,4/3)-1-bromo-2,3,4-trihydroxycyclohex-5-ene) with ki(max)/Ki 10(5) M-1 min-1. In contrast with the Br-CBE-inhibited enzyme the beta-glucosidase inhibited by bromoconduritol F was subject to spontaneous reactivation with t1/2 approximately 20 min.

Intracellular activity of lysosomal glucosylceramidase measured with 4-nonylumbelliferyl beta-glucoside.

Enzymatic activity of lysosomal glucosyl-ceramidase was determined in intact murine hybridoma and macrophage cells with the synthetic substrate nonylumbeliferyl-beta-glucoside (NUG). The substrate was applied as complex with bovine serum albumin (two binding sites, Kd 2.2 +/- 0.3 microM). The transport of the artificial substrate from medium to the enzyme was explored by measurements of substrate concentrations in cellular membranes and of endocytosis rate relative to substrate hydrolysis. The results indicated that, after enrichment in the plasma membrane, the substrate is mainly transported by simple diffusion. Release of nonylumberlliferone monitored fluorimetrically after disintegration of the cells in borate buffer containing Triton X-100 at pH 9.5 showed that 10(8) cells of both cell lines hydrolysed 1-1.5 nmol substrate/min at a total concentration of 0.1 mM NUG in the medium. Substrate hydrolysis was prevented by preincubating the cells with conduritol B epoxide (CBE), a specific active site-directed inhibitor of lysosomal glucosylceramidase. The substrate concentration at the site of the enzyme and maximal activity were evaluated by the inhibiting effect of the substrate on the inactivation rate by conduritol B epoxide. The rate of inhibitor uptake measured with bromo-[3H]conduritol B epoxide was shown to be not rate-limiting for the inactivation reaction. The molar concentration of the enzyme was determined by labeling with bromo-[3H]conduritol B epoxide. Comparison of the maximal intracellular activity with that of the enzyme after disintegration and activation by taurocholate showed a 20-fold lower activity in the native environment.(ABSTRACT TRUNCATED AT 250 WORDS)

Man9-mannosidase from human kidney is expressed in COS cells as a Golgi-resident type II transmembrane N-glycoprotein.

Man9-mannosidase, an alpha 1,2-specific exo-enzyme involved in N-linked oligosaccharide processing, has been cloned recently from a human kidney cDNA library [Bause, E., Bieberich, E., Rolfs, A., Völker, C. & Schmidt, B. (1993) Eur. J. Biochem. 217, 533-540]. Transient expression in COS 1 cells of the enzyme resulted in a more than 20-fold increase of a catalytic activity cleaving specifically alpha 1,2-mannosidic linkages in [14C]Man9-GlcNAc2 or [14C]Man5-GlcNAc2. Man9-mannosidase is expressed as a N-glycoprotein with a molecular mass of 73 kDa. Its enzymic activity is metal ion dependent and inhibited strongly by 1-deoxymannojirimycin (50% at 100 microM). Proteolytic studies with the membrane-associated form of Man9-mannosidase support the view that the enzyme is a type II transmembrane protein as predicted from its cDNA sequence. Several lines of evidence suggest that Man9-mannosidase, as expressed, is N-glycosylated at one of three potential Asn-Xaa-Thr/Ser/Cys acceptor sites. Approximately 50% of the N-linked oligosaccharide chains are removed by endoglycosidase H treatment, whereas complete deglycosylation of the enzyme is observed, when transfected cells were cultured in the presence of the Golgi mannosidase II inhibitor swainsonine, indicating that the sugar moiety of Man9-mannosidase is processed partially by Golgi-resident enzymes. This observation is consistent with the results of indirect immunofluorescence studies, pointing to a localization of the Man9-mannosidase predominantly in the juxtanuclear Golgi region. This localization clearly differs from that of pig liver Man9-mannosidase which appears to be located in the endoplasmic reticulum and transient vesicles.

Protein-ribosome-mRNA display: affinity isolation of enzyme-ribosome-mRNA complexes and cDNA cloning in a single-tube reaction.

An enzyme-ribosome-mRNA complex was specifically purified by binding to the immobilized enzyme substrate and the cDNA was cloned in a single-tube reaction by one-step reverse transcription-PCR. The ganglioside GM3, used by sialyltransferase II (ST-II) as a substrate, was coated on a 96-well microtiter plate and ST-II was in vitro transcribed and translated from a cDNA library. The isolation of an enzyme-specific protein-ribosome (PRIME) complex was achieved with as little as 0.1 ng ST-II-specific cDNA in 5 microg of a total plasmid preparation or with the cDNA prepared from sublibraries previously inoculated at a density of 2000 clones/culture well. The affinity purification of the PRIME complex was highly specific for GM3 and did not result in cDNA amplification when a different ganglioside (GM1) was used for coating of the microtiter plate. The amplified cDNA was used for cloning or a second round of ribosome display, providing a fast analysis of enzyme affinity to multiple substrates. PRIME display can be used for host-free cDNA cloning from mRNA or cDNA libraries and for binding site mapping of the in vitro translated protein. The use of a single-tube reaction in ligand-coated microtiter plates indicates the versatility of PRIME display for cDNA cloning by automated procedures.

Regulation of apoptosis during neuronal differentiation by ceramide and b-series complex gangliosides.

Lipid analysis of gestational day E14.5 mouse brain revealed elevation of ceramide to a tissue concentration that induced apoptosis when added to the medium of neuroprogenitor cells grown in cell culture. Elevation of ceramide was coincident with the first appearance of b-series complex gangliosides (BCGs). Expression of BCGs by stable transfection of murine neuroblastoma (F-11) cells with sialyltransferase-II (ST2) resulted in a 70% reduction of ceramide-induced apoptosis. This was most likely due to an 80% reduced expression of prostate apoptosis response-4 (PAR-4). PAR-4 expression and apoptosis were restored by preincubation of ST2-transfected cells with N-butyl deoxinojirimycin (NB-DNJ) or PD98059, two inhibitors of ganglioside biosynthesis or p42/44 mitogen-activated protein (MAPK) kinase, respectively. In sections of day E14.5 mouse brain, the intermediate zone showed intensive staining for complex gangliosides, but only low staining for apoptosis (TUNEL) and PAR-4. Apoptosis and PAR-4 expression, however, were elevated in the ventricular zone which only weakly stained for complex gangliosides. Whole cell patch clamping revealed a 2-fold increased calcium influx in ST2-transfected cells, the blocking of which with nifedipine restored apoptosis to the level of untransfected cells. In serum-free culture, supplementation of the medium with IGF-1 was required to maintain MAPK phosphorylation and the anti-apoptotic effect of BCG expression. BCG-enhanced calcium influx and the presence of insulin-like growth factor-1 may thus activate a cell survival mechanism that selectively protects developing neurons against ceramide-induced apoptosis by up-regulation of MAPK and reduction of PAR-4 expression.

Effect of N-glycosylation on turnover and subcellular distribution of N-acetylgalactosaminyltransferase I and sialyltransferase II in neuroblastoma cells.

Gangliosides are sialylated glycosphingolipids whose biosynthesis is catalyzed by a series of endoplasmic reticulum (ER)- and Golgi-resident glycosyltransferases. Protein expression, processing, and subcellular localization of the key regulatory enzymes for ganglioside biosynthesis, sialyltransferase II (ST-II) and N-acetylgalactosaminyltransferase I (GalNAcT), were analyzed upon transient expression of the two enzymes in the neuroblastoma cell lines NG108-15 and F-11. The enzymes were endowed with a C-terminal epitope tag peptide (FLAG) for immunostaining and immunoaffinity purification using a FLAG-specific antibody. Mature ST-II-FLAG and GalNAcT-FLAG were expressed as N-glycoproteins with noncomplex oligosaccharides. ST-II-FLAG was distributed to the Golgi apparatus, whereas GalNAcT-FLAG was found in the ER and Golgi. Inhibition of early N-glycoprotein processing with castanospermine resulted in a distribution of ST-II-FLAG to the ER, whereas that of GalNAcT-FLAG remained unaltered. In contrast to GalNAcT, the activity of ST-II and the amount of immunostained enzyme were reduced concomitantly by 75% upon incubation with castanospermine. This was due to a fourfold increased turnover of ST-II-FLAG, which was not found with GalNAcT-FLAG. The ER retention and increased turnover of ST-II-FLAG were most likely due to its inability to bind to calnexin upon inhibition of early N-glycoprotein processing. Calnexin binding was not observed for GalNAcT-FLAG, indicating a differential effect of N-glycosylation on the turnover and subcellular localization of the two glycosyltransferases.

N-acylated serinol is a novel ceramide mimic inducing apoptosis in neuroblastoma cells.

A novel structural analog of ceramide was synthesized by N-acylation of serinol (2-amino-1,3-propanediol) and studied for its effects on glycolipid biosynthesis and cell differentiation of neuroblastoma cells. Incubation with N-palmitoylated serinol (C16-serinol) increased the concentration of endogenous ceramide by 50-80% and caused apoptosis in rapidly dividing low density cells but not in confluent cells. Cell death was not suppressed by simultaneous incubation with phorbol ester, known to antagonize ceramide-induced apoptosis by activation of protein kinase C (PKC). Purification of potential target proteins of C16-serinol was achieved by affinity chromatography of a protein preparation from rat brain on immobilized C16-serinol. A gel activity assay revealed that the eluate from C16-serinol-Sepharose contained three serine/threonine-specific protein kinases with molecular masses of 50, 70, and 95 kDa. The 70-kDa protein was immunostained on a Western blot using a PKCzeta-specific antibody. The purified PKCzeta could be activated directly by C16-serinol in an in vitro phosphorylation assay. Induction of apoptosis in neuroblastoma cells was suppressed by inhibition of PKCzeta with Gö 6983. Our overall results indicate that apoptosis in neuroblastoma cells induced by C16-serinol was at least partially mediated by activation of PKCzeta on condition of ongoing cell division. N-Acylated serinols may thus be useful for induction of apoptosis in mitotic cells and may be of therapeutic potential for treatment of cancer in the nervous system.

Cloning and expression of glucosidase I from human hippocampus.

Glucosidase I, the first enzyme in the N-linked oligosaccharide processing pathway, cleaves the distal alpha 1,2-linked glucose residue from the Glc3-Man9-GlcNAc2 oligosaccharide precursor highly specifically. A human hippocampus cDNA library was screened against oligonucleotide probes, generated by PCR using primers derived from the amino acid sequences of tryptic peptides of pig liver glucosidase I. Two independent lambda clones were isolated which allowed the construction of a full-length glucosidase I cDNA of 2881 bp. This cDNA construct encodes, in a single open reading frame, a polypeptide of 834 amino acids corresponding to a molecular mass of 92 kDa. The 92-kDa protein contains a single N-glycosylation site of the Asn-Xaa-Thr/Ser type at Asn655, as well as a strongly hydrophobic sequence close to its N-terminus (amino acids 38-58) which, most likely, functions as a transmembrane anchor. The amino acid sequences of all tryptic peptides of the pig liver enzyme were found, with little deviation, within the coding sequence. This demonstrates the authenticity of the cDNA construct and the close evolutionary relationship between the enzymes from human hippocampus and pig liver. In contrast, the nucleotide and amino acid sequence revealed no homology with other processing enzymes cloned so far. Transfection of COS 1 cells with the glucosidase I cDNA construct resulted in overexpression (about fourfold) of enzymic activity, which was inhibited strongly by 1-deoxynojirimycin or N,N-dimethyl-deoxynojirimycin. The expressed enzyme, with a molecular mass of 95 kDa, is degraded by endoglycosidase H to a 93-kDa form, indicating that it carries a high-mannose oligosaccharide chain at Asn655. The presence of this glycan is in line with the localization of glucosidase I in the lumen of the endoplasmic reticulum, shown by immunofluorescence microscopy. The hydrophobicity profile as well as the removal by trypsin of an approximately 4-kDa polypeptide from the membrane-associated glucosidase I in intact microsomal structures, supports the view that the enzyme is a type-II transmembrane glycoprotein, which contains a short cytosolic tail of approximately 37 amino acids, followed by a single transmembrane domain and a large C-terminal catalytic domain located on the luminal side of the endoplasmic reticulum membrane.

Combinatorial PCR approach to homology-based cloning: cloning and expression of mouse and human GM3-synthase.

GM3-synthase, also known as sialyltransferase I (ST-I), catalyzes the transfer of a sialic acid residue from CMP-sialic acid onto lactosylceramide to form ganglioside GM3. In order to clone this enzyme, as well as other sialyltransferases, we developed an approach that we termed combinatorial PCR. In this approach, degenerate primers were designed on the basis of conserved sequence motifs of the ST3 family of sialyltransferases (STs). The nucleotide sequence of the primers was varied to cover all amino acid variations occurring in each motif. In addition, in some primers the sequence was varied to cover possible homologous substitutions that are absent in the available motifs. A panel of cDNA from 12 mouse and 8 human tissues was used to enable cloning of tissue- and stage-specific sialyltransferases. Using this approach, the fragments of 11 new putative sialyltransferases were isolated and sequenced so far. Analysis of the expression pattern of a particular sialyltransferase across the panel of cDNA from the different tissues provided information about the tissue specificity of ST expression. We chose two new ubiquitously expressed human and mouse STs to clone full-length copies and to assay for GM3-synthase activity. One of the STs, which exhibited the highest homology to ST3 Gal III, showed activity toward lactosylceramide (LacCer) and was termed ST3 Gal V according to the suggested nomenclature [1]. The other ubiquitously expressed sialyltransferase was termed ST3Gal VI. All isolated sialyltransferases were screened for alternatively spliced forms (ASF). Such forms were found for both human ST3Gal V and ST3Gal VI in human fetal brain cDNA library. The detailed cloning strategy, functional assay, and full length cDNA and protein sequences of GM3 synthase (ST3Gal V, or ST-I) are presented.

Differential effects of glycolipid biosynthesis inhibitors on ceramide-induced cell death in neuroblastoma cells.

An in vitro model of Gaucher's disease in murine neuroblastoma x rat glioma NG108-15 cells was used to investigate the physiological effects of two specific inhibitors of glucosylceramide synthase, d,l-threo-1-phenyl-2-decanoylamino-3-morpholino-1-propanol (d,l-PDMP) and N-butyldeoxynojirimycin (NB-DNJ), which have been suggested as agents for treatment of glycolipid storage disorders. Incubation of NG108-15 cells with conduritol-B-epoxide, a covalent inhibitor of glucosylceramidase, raised the intracellular concentration of glucosylceramide (GC) by more than fourfold, indicating a glycolipid composition equivalent to that of Gaucher's cells. The level of GC was decreased, and the cells were depleted of gangliosides by postincubation with d,l-PDMP or NB-DNJ. Treatment with d,l-PDMP, but not with NB-DNJ, resulted in a dose-dependent reduction of the growth rate and eventually caused cell death in NG108-15 cells on reaching confluency. An in situ detection assay using terminal nucleotidyltransferase indicated that cell degeneration was accompanied by apoptosis. Lipid analysis by high-performance TLC revealed that on incubation with d,l-PDMP, but not with NB-DNJ, the concentration of endogenous ceramide was elevated by threefold. Ceramide elevation and apoptosis were also observed when NG108-15 cells were incubated with daunorubicin, which was previously reported to induce programmed cell death by stimulation of ceramide synthesis. Structural characterization by HPLC and subsequent laser desorption mass spectrometry revealed that the endogenous ceramide contained fatty acids with chain lengths ranging from C14:0 to C24:0. The results indicate that elevation of levels of these ceramide species by incubation with d,l-PDMP or daunorubicin induces programmed cell death in NG108-15 cells. Because ceramide accumulation and cell death were not observed on incubation with NB-DNJ, its use is suggested to be less toxic than that of d,l-PDMP for treatment of Gaucher's disease and other sphingolipid storage disorders.

Molecular cloning and primary structure of Man9-mannosidase from human kidney.

Man9-mannosidase, a processing enzyme found in the endoplasmic reticulum (ER), catalyses the removal of three distinct mannose residues from peptide-bound Man9-GlcNAc2 oligosaccharides producing a single Man6 isomer [Bause, E., Breuer, W., Schweden, J., Roesser, R. & Geyer, R. (1992) Eur. J. Biochem. 208, 451-457]. We have isolated four Man9-mannosidase-specific clones from a human kidney cDNA library and used these to construct a full-length cDNA of 3250 base pairs. A single open reading frame of 1875 nucleotides encodes a protein of approximately 71 kDa, consistent with data from immunological studies. Analysis of the coding sequence predicts that Man9-mannosidase is a type II transmembrane protein consisting of a short cytoplasmic polypeptide tail, a single transmembrane domain acting as a non-cleavable signal sequence and a large luminal catalytic domain. This domain architecture closely resembles that of other ER and Golgi-located processing enzymes, pointing to common structural motifs involved in membrane insertion and topology. The protein sequence of the Man9-mannosidase contains three potential N-glycosylation sites of which only one site is used. The amino acid sequence of several peptide regions, including a calcium-binding consensus sequence, bears striking similarities to an ER alpha-1,2-mannosidase from yeast, whereas, by contrast, no sequence similarity was detectable with rat liver ER alpha-mannosidase and Golgi alpha-mannosidase II. This finding may indicate that the mammalian alpha-mannosidases, which differ significantly in their substrate specificity, are coded for by evolutionarily unrelated genes, providing an attractive means of regulation and fine-tuning oligosaccharide processing, not only at the enzymic but also at the transcriptional level.