Ceramide induces activation of the mitochondrial/caspases pathway in Jurkat and SCC61 cells sensitive to gamma-radiation but activation of this sequence is defective in radioresistant SQ20B cells

PURPOSE

To clarify the molecular mechanisms leading to radiation-induced apoptosis or resistance, the kinetics (1-48 h) and sequence of events triggered in response to 10 Gy irradiation were investigated in three cell lines displaying a gradient of sensitivity to 7-rays.

MATERIALS AND METHODS

Ceramide levels were measured by high performance liquid chromatography (HPLC). Mitochondrial function was evaluated in terms of transmembrane potential (delta(psi)m), reactive oxygen species (ROS) and glutathione levels analysed by flow cytometry or HPLC. Caspase activation was assessed by immunoblotting, and apoptosis by flow cytometry.

RESULTS

In Jurkat radiosensitive cells and SCC61 adherent cells with intermediate radiosensitivity, the degree of delayed ceramide release was directly related to their propensity to undergo apoptosis. Transduction of the death signal was mediated by a drop in delta(psi)m and glutathione levels, ROS accumulation and activation of effector caspases. Experiments conducted with caspase inhibitors, bongkrekic acid, or DL-PDMP indicated that ceramide triggers mitochondrial collapse, followed by the activation of caspases-9, -8 and -3, and poly(ADP-ribose)polymerase cleavage. In SQ20B radioresistant cells, gamma-radiation did not induce ceramide generation or subsequent activation of the mitochondrial/caspase apoptotic pathway.

CONCLUSIONS

Ceramide appears to be a determining factor in the commitment phase of radiation-induced apoptosis. When ceramide is not generated, the whole pathway is ineffective and resistance to apoptosis may result.

Comparison of methods used to study cell death in an adherent tumoral cell line with moderate clonogenic radiosensitivity

Our objective was to compare different methods for studying programmed cell death in adherent H460 non-small lung cancer cells of moderate clonogenic radiosensitivity. The major effect of gamma-radiation was found to be the release of cells from the substratum. The different methods gave complementary and unexpected information: a) with the TUNEL method, a few non-apoptotic cells were found in the culture medium; b) with the flow cytometry after propidium iodide labeling, some hypodiploid cells which remained attached to the substratum were apoptotic, as demonstrated by the effect of a caspase inhibitor; c) with the annexin V labeling, the detached cells were demonstrated either necrotic or very late apoptotic; d) the mitochondria transmembrane potential (deltapsim), measurements demonstrated that the mitochondria were implicated in cell death induced by gamma-radiation. These data illustrate the need to use several complementary methods in the study of apoptosis in adherent cells exposed to gamma-radiation.

Temporal relationships between ceramide production, caspase activation and mitochondrial dysfunction in cell lines with varying sensitivity to anti-Fas-induced apoptosis

To clarify the chronology of events leading to anti-Fas-induced apoptosis, and the mechanisms of resistance to this death effector, we compared the response kinetics of three tumour cell lines that display varying sensitivity to anti-Fas (based on levels of apoptosis), in terms of ceramide release, mitochondrial function and the caspase-activation pathway. In the highly sensitive Jurkat cell line, early caspase-8 activation, observed from 2 h after treatment, was chronologically associated with an acute depletion of glutathione and the cleavage of caspase-3 and poly-ADP ribosyl polymerase (PARP), followed by a progressive fall in the mitochondrial transmembrane potential (Delta(psi)m), between 4 and 48 h after treatment. Ceramide levels began to increase 2 h after the addition of anti-Fas (with no increase during the first hour), and increased continuously to 640% of control cells at 48 h. In the moderately sensitive SCC61 adherent cells, comparable results were observed, though with lower levels of ceramide and a delay in the response kinetics, with apoptotic cells becoming flotant. Finally, despite early cleavage of caspase-8 at 2 h, and a sustained level of activation until 48 h, no apoptotic response was observed in anti-Fas-resistant SQ20B cells. This was confirmed by a lack of ceramide generation and mitochondrial changes, and by the absence of any detectable cleavage of caspase-3 or PARP. Inhibition of caspase processing, and amplification of endogenous ceramide signalling by pharmacological agents, allowed us to establish the order of cellular events, locating ceramide release after caspase-8 activation and before caspase-3 activation, and demonstrating a direct involvement for ceramide release in mitochondrial dysfunction. Furthermore, these experiments provide strong arguments for the role of endogenous ceramide as a key executor of apoptosis, rather than as a consequence of membrane alterations.

Study of O-sialylation of glycoproteins in C6 glioma cells treated with retinoic acid

When treated with retinoic acid in vivo, C6 glioma cells show an enhancement of CMP-Neu5Ac:Gal beta 1-3 GalNAc-R alpha-2,3 sialyltransferase activity. A 300 kDa glycoprotein was detected by lectin affinoblotting in retinoic acid-treated C6 cells which stained weakly or not at all in control cells. Comparative studies with different lectins demonstrated that this glycoprotein contains alpha 2,3 Neu5Ac Gal-GalNAc O-glycan moieties. Cultures in the presence of an inhibitor of O-glycan synthesis (N-acetylgalactosaminide alpha-O-benzyl) demonstrated that enhancement of staining of the 300 kDa glycoprotein was not due to the increase of the alpha 2,3 sialytransferase but to the de novo synthesis of the polypeptide chain of this glycoprotein.

Study of retinoic acid effect upon retinoic acid receptors beta (RAR-beta) in C6 cultured glioma cells

Using monoclonal antibodies against the RAR-alpha and RAR-beta retinoic receptors, we demonstrated that these receptors were present together in C6 glioma cells as two isoforms of 50 and 55 kDa. For RAR-beta, the 50 kDa isoform predominated (60 to 80% of the total of the two isoforms). After a treatment for 48 h with retinoic acid 10 microM, the 55 kDa form was enhanced, while no effect was observed either on RAR-alpha isoforms from C6 cells and on both RAR-alpha and RAR-beta forms from neuroblastoma SKN SH SY5Y used as a control. Using purified neuronal and glial rat brain nuclei, we showed that the 55 kDa isoform from RAR-beta predominated in glial cells. These results suggest that retinoic acid treatment of C6 cells led to a partial differentiation, the enhancement of the heavy form of RAR-beta being a marker of this phenomenon.

[Serum phospholipase A2 activity in osteoarthritis]

Serum phospholipase A2 activity in 67 osteoarthritis patients and 17 controls was determined using a radiolabeled specific substrate. Serum phospholipase A2 activity was significantly higher in osteoarthritis patients (115 +/- 73.6 dpm/h/ml) than in controls (45 +/- 25 dpm/h/ml) (p = 0.002). In 41 osteoarthritis patients, serum phospholipase A2 activity was unrelated to age, time since onset of osteoarthritis symptoms, duration of morning stiffness, Lequesne's index, roentgenographic stage of osteoarthritis, number of joints with osteoarthritis, erythrocyte sedimentation rate, or serum C-reactive protein levels. In 12 osteoarthritis patients who were evaluated twice at a mean interval of 46 days, changes in serum phospholipase A2 activity were unrelated to changes in Lequesne's index. Blind evaluation of long-term joint space loss was performed in 14 patients; serum phospholipase A2 activity increased only in those patients with progressive joint space loss, but the difference was not statistically significant as compared with the controls. These data suggest that serum phospholipase A2 activity is not useful in practice as a marker for osteoarthritis.

Study of O-glycan sialylation in C6 cultured glioma cells: regulation of a beta-galactoside alpha 2,3 sialyltransferase activity by Ca2+/calmodulin antagonists and phosphatase inhibitors

We have demonstrated that the alpha 2,3 sialyltransferase (alpha 2,3 ST) from C6 cultured glioma cells was inhibited in vivo by W-7 and related Ca2+/Calmodulin (Ca/CaM) antagonists while protein kinase C effectors had no effect. Dephosphorylation of alpha 2,3 ST by the wide specificity alkaline phosphatase led to inactivation indicating that the enzyme is phosphorylated. The serine/threonine protein phosphatase inhibitors okadaic acid and Calyculin A led also to an inhibition of alpha 2,3 ST activity. In addition, Ca/CaM antagonists and phosphatase inhibitors led both to an inhibition of a alpha 2,3 sialoglycoprotein from C6 glioma cells as demonstrated with lectin affinity blotting. A concerted regulatory mechanism with phosphorylation/dephosphorylation of alpha 2,3 ST is then postulated.

Effect of non-steroidal anti-inflammatory drugs (NSAIDS) on glycosyltransferase activity from human osteoarthritic cartilage

The effect of non-steroidal anti-inflammatory drugs (NSAIDs) on the activity of glycosyltransferases required for the synthesis of the polysaccharide chains of proteoglycans, was studied in human osteoarthritic cartilage in vitro. Using exogenous acceptors, salicylate and indomethacin suppressed the activity of glucuronyl- and xylosyltransferases in a concentration-dependent manner, but had little effect on N-acetylgalactosaminyl- and galactosyltransferases. When used at a concentration derived from the values found in the synovial fluid, salicylate, indomethacin and chloroquine significantly suppressed the activity of glucuronyl- and xylosyltransferases, while tiaprofenic acid, paracetamol (acetaminophen), floctafenine, ketoprofen, ibuprofen and tenoxicam had no effect on the enzymes. An alteration of some glycosyltransferases could explain the reported suppressive effect of some NSAIDs on cartilage proteoglycan synthesis.

Study of O-glycan sialylation in C6 cultured glioma cells: evidence for post-translational regulation of a beta-galactoside alpha 2,3 sialyltransferase activity by N-glycosylation

We have studied the Gal beta 1-3GalNAc-R alpha 2,3 sialyltransferase from C6 glioma cells transferring Neu5Ac from CMP-Neu5Ac onto O-glycans of glycoproteins. Using synchronized C6 glioma cells, we showed that the alpha 2,3 sialyltransferase activity was inhibited by tunicamycin to a greater extend than DNA and protein biosynthesis suggesting inhibition of N-glycosylation of this enzyme. Additional demonstration of N-glycosylation of the alpha 2,3 sialytransferase was provided through ConA-Sepharose binding. Treatment of partially purified alpha 2,3 sialytransferase by peptide-N-glycosidase F showed a significative inhibition demonstrating that N-glycan moiety is required for complete activity of the C6 glioma cell alpha 2,3 sialyltransferase.

[Cytokines, prostaglandin E2, phospholipase A and metalloproteases in synovial fluid in osteoarthritis]

Concentrations of prostaglandin E2, interleukin 1 beta, interleukin 6 and tumor necrosis factor alpha, phospholipase A2, collagenase and proteoglycanase activity were determined in synovial fluid from 26 patients with osteoarthrosis of the knee and 10 with rheumatoid arthritis. Osteoarthrosis synovial fluid was characterised by the absence of interleukin 1 beta while tumour necrosis factor alpha and interleukin 6 were present in relatively large amounts, by a very high phospholipase A2 activity contrasting with a very low concentration of prostaglandin E2, and by a collagenase/proteoglycanase activity only slightly less constant and high as in rheumatoid arthritis. In osteoarthrosis patients, the interleukin 6 concentration, but not that of tumor necrosis factor alpha, was correlated with the collagenase and proteoglycanase activity of synovial fluid.

Calmodulin-dependent collagenase and proteoglycanase activities in chondrocytes from human osteoarthritic cartilage

Chondrocyte metalloproteinases appear to play a major role in the development of osteoarthritis. The intracellular post-traductional mechanisms regulating collagenase and proteoglycanase are not known. Calmodulin antagonists including phenothiazine and sulfonamide derivatives significantly increased proteoglycanase activity and decreased collagenase activity. H-7, a specific inhibitor of protein kinase C, had no effect on the two metalloproteinase activities, and calmodulin was ineffective in in vitro assays upon metalloproteinase activities. We postulate that collagenase and proteoglycanase activities are controlled by calmodulin-dependent regulation.

Serotonin-stimulated phospholipase A2 and collagenase activation in chondrocytes from human osteoarthritic articular cartilage

We have previously described several receptors on the chondrocyte membrane. In an attempt to further characterize the coupling mechanisms of serotoninergic receptors, here we examined the involvement of serotonin in the phospholipase A2 activity. Serotonin dose-dependently stimulated phospholipase A2. This activation enhanced collagenase type II activity and had no effect on proteoglycanase activity.

Effect of desipramine on a glycoprotein sialyltransferase activity in C6 cultured glioma cells

The tricyclic antidepressant desipramine, when added to culture medium, gave rise in C6 rat glioma cells to a decrease of the activity of the enzyme asialofetuin sialyltransferase. The inhibition was dose and time dependent and was observed in both multiplying cells and cells blocked with 2 mM thymidine or depletion of amino acids. This inhibition was rather specific to the sialyltransferase, as under the conditions where this enzyme was inhibited up to 70%, other enzymes such as dolichol phosphate mannose synthetase, glutamine synthetase, and glycerol phosphate dehydrogenase remained unaffected. This inhibition was not reversed after removal of desipramine from the medium and was not observed by direct addition of desipramine to the sialyltransferase incubation assay. Under the same conditions, W-7 [N-(6-aminohexyl)-5-chloro-1-naphthalenesulfonamide], which is known to be a potent calmodulin antagonist and an inhibitor of calmodulin-dependent kinases, gave the same concentration-dependent inhibition profile of sialyltransferase as desipramine, whereas H-7 [1-(5-isoquinolinylsulfonyl)-2-methylpiperazine], which is an inhibitor of protein kinase C and cyclic nucleotide-dependent kinases, had no effect. So, it is suggested that desipramine inhibits the sialyltransferase activity in C6 glioma cells through a calmodulin-dependent system.

Cartilage degradative enzymes in human osteoarthritis: effect of a nonsteroidal antiinflammatory drug administered orally

The activity of stromelysin and collagenase was determined in fibrillated human OA cartilage using labeled proteoglycans and type II collagen as substrates. In vitro paracetamol had no effect on metalloprotease whereas TA induced a significant inhibition of stromelysin. In cartilage and synovium from nine patients treated with TA and nine patients treated with paracetamol during 8 weeks before surgery for hip OA, stromelysin activity was significantly lower in the TA than in the paracetamol group. The results suggest that TA has a potential chondroprotective effect in OA.

Effect of retinoic acid on two glycosyltransferase activities in C6 cultured glioma cells

1. Activity of two glycosyltransferases was studied in retinoic acid-treated C6 cultured glioma cells. 2. The beta-galactoside alpha 2,3-sialyltransferase transferring N-acetylneuramin onto the O-glycans residues of glycoproteins was activated up to twice after chronic treatment (from 24 to 96 hr) with all-trans retinoic acid. 3. No effect was observed for shorter treatments. 4. On the opposite, the N-glycan galactosyltransferase activity remained unchanged whatever the length of retinoic acid treatment was. 5. The activatory effect was not dependent on isomery, as all-trans and 13-cis retinoic acid isomers were both activators of the C6 glioma cell sialyltransferase. 6. Measurement of adhesion of retinoic acid-treated cells using labelled plasma membranes showed an enhancement of adhesion in correlation with enhancement of sialyltransferase activity.

Evidence for an O-glycan sialylation system in brain. Characterization of a beta-galactoside alpha 2,3-sialyltransferase from rat brain regulating the expression of an alpha-N-acetylgalactosaminide alpha 2,6-sialyltransferase activity

We present evidence for the existence in rat brain of several sialyltransferases able to sialylate sequentially asialofetuin. [14C]Sialylated glycans of asialofetuin were analyzed by gel filtration. Three types of [14C]sialylated glycans were synthesized: N-glycans and monosialylated and disialylated O-glycans. The varying effects of N-ethylmaleimide, lysophosphatidylcholine (lysoPtdCho) and trypsin, were helpful in the identification of these different sialyltransferases. One of them, selectively inhibited by N-ethylmaleimide, was identified as the Neu5Ac alpha 2----3Gal beta 1----3GalNAc-R:alpha 2----6 sialyltransferase previously described [Baubichon-Cortay, H., Serres-Guillaumond, M., Louisot, P. and Broquet, P. (1986) Carbohydr. Res. 149, 209-223]. This enzyme was responsible for the synthesis of disialylated O-glycans. LysoPtdCho and trypsin selectively inhibited the enzyme responsible for the synthesis of monosialylated O-glycan. N-ethylmaleimide, lysoPtdCho and trypsin did not inhibit Neu5Ac transfer onto N-glycans, giving evidence for three different molecular species. To identify the enzyme responsible for monosialylated O-glycan synthesis, we used another substrate: Gal beta 1----3GalNAc--protein obtained after galactosylation of desialylated ovine mucin by a GalNAc-R:beta 1----3 galactosyltransferase from porcine submaxillary gland. This acceptor was devoid of N-glycans and of NeuAc in alpha 2----3 linkages on the galactose residue. When using N-ethylmaleimide we obtained the synthesis of only one product, a monosialylated structure. After structural analysis by HPLC on SAX and SiNH2 columns, we identified this product as Neu5Ac alpha 2----3Gal beta 1----3GalNAc. The enzyme leading to synthesis of this monosialylated O-glycan was identified as a Gal beta 1----3GalNAc-R:alpha 2----3 sialyltransferase. When using lysoPtdCho and trypsin, sialylation was completely abolished, although the Neu5Ac alpha 2----3Gal beta 1----3GalNAc-R:alpha 2----6 sialyltransferase was not inhibited. We provided thus evidence for the interpendence between the two enzymes, the alpha 2----3 sialyltransferase regulates the alpha 2----6 sialyltransferase activity since it synthesizes the alpha 2----6 sialyltransferase substrate.

Different reactivity of two brain sialyltransferases towards sulfhydryl reagents. Evidence for a thiol group involved in the nucleotide-sugar binding site of the NeuAc alpha 2-3Gal beta 1-3GalNAc alpha(2-6)sialyltransferase

We have studied the amino-acid residues involved in the catalytic activity of two distinct brain sialyltransferases acting on fetuin and asialofetuin. These two enzymes were strongly inhibited by N-bromosuccinimide, a specific blocking reagent for tryptophan residues. This result suggests the involvement of such residues in the catalytic process of the two sialyltransferases. Furthermore, chemical modifications by various sulfhydryl reagents led to a strong inhibition of the fetuin sialyltransferase while the asialofetuin sialyltransferase was only slightly inhibited. For a more thorough understanding of the thiol inactivation mechanism of the fetuin sialyltransferase, we studied in more detail the reactivity of this enzyme with NEM (N-ethylmaleimide), an irreversible reagent. The time-dependent inactivation followed first-order kinetics and these kinetic data afforded presumptive evidence for the binding of 1 mol NEM per mol of enzyme. Only CMP-NeuAc protected the enzyme against NEM inactivation effectively. MnCl2 did not enhance the protective effect of CMP-NeuAc. The modifications of the fetuin sialyltransferase kinetic parameters by NEM showed a competitive mechanism between NEM and CMP-NeuAc. The results suggest the involvement of a sulfhydryl residue in or near the nucleotide-sugar binding site of the fetuin sialyltransferase (but we could not excluded that CMP-NeuAc binding may induce a change in conformation of the protein, leading to a decreased accessibility of this thiol group located near the nucleotide-sugar binding site). This SH group is essential to the enzyme activity, which is not the case for the asialofetuin sialyltransferase.

Study of the fucosyltransferase system in intestinal microsomes

Fucosyltransferase activity of rat small intestine microsomes is solubilized by 0.5% Triton X-100. The solubilized activity can be purified up to 8,300-fold using DEAE-cellulose and affinity chromatography on GDP-Sepharose. At this step, chromatography on Sephadex G15 separates different specificities: N-acetylglucosaminide-alpha-(1,3)-fucosyltransferase acting on asialoserotransferrin, and galactoside-alpha-(1,2)-fucosyltransferase acting on O-glycans of asialofetuin. The use of small saccharidic acceptors also indicates the presence of a N-acetylglucosaminide-alpha-(1,4)-fucosyltransferase and of a very weak glucose-alpha-(1,3)-fucosyltransferase activity. These activities are tightly bound to concanavalin A-Sepharose, suggesting that they are supported by N-glycosylproteins.

Phosphorylation, glycosylation, and proteolytic activity of the 52-kD estrogen-induced protein secreted by MCF7 cells

We have studied the posttranslational modifications of the 52-kD protein, an estrogen-regulated autocrine mitogen secreted by several human breast cancer cells in culture (Westley, B., and H. Rochefort, 1980, Cell, 20:353-362). The secreted 52-kD protein was found to be phosphorylated mostly (94%) on high-mannose N-linked oligosaccharide chains, and mannose-6-phosphate signals were identified. The phosphate signal was totally removed by alkaline phosphatase hydrolysis. The secreted 52-kD protein was partly taken up by MCF7 cells via mannose-6-phosphate receptors and processed into 48- and 34-kD protein moieties as with lysosomal hydrolases. By electron microscopy, immunoperoxidase staining revealed most of the reactive proteins in lysosomes. After complete purification by immunoaffinity chromatography, we identified both the secreted 52-kD protein and its processed cellular forms as aspartic and acidic proteinases specifically inhibited by pepstatin. The 52-kD protease is secreted in breast cancer cells under its inactive proenzyme form, which can be autoactivated at acidic pH with a slight decrease of molecular mass. The enzyme of breast cancer cells, when compared with cathepsin D(s) of normal tissue, was found to be similar in molecular weight, enzymatic activities (inhibitors, substrates, specific activities), and immunoreactivity. However, the 52-kD protein and its cellular processed forms of breast cancer cells were totally sensitive to endo-beta-N-acetylglucosaminidase H (Endo H), whereas several cellular cathepsin D(s) of normal tissue were partially Endo H-resistant. This difference, in addition to others concerning tissue distribution, mitogenic activity and hormonal regulation, strongly suggests that the 52-kD cathepsin D-like enzyme of breast cancer cells is different from previously described cathepsin D(s). The 52-kD estrogen-induced lysosomal proteinase may have important functions in facilitating the mammary cancer cells to proliferate, migrate, and metastasize.

Study of circadian correlations between acetylcholine muscarinic receptor and brain glycosyltransferases by multivariate analysis

Circadian variations of the acetylcholine muscarinic receptor and some glycosyltransferases were studied in brain using multivariate analysis. Highly significant correlations exist between fucosyltransferase, sialyltransferase and galactosyltransferase and to a lesser extent between both of these enzymes and acetylcholine receptor. No correlation appeared between these enzymes and dolichol phosphate mannose synthase.

Different reactivity to lysophosphatidylcholine, DIDS and trypsin of two brain sialyltransferases specific for O-glycans: a consequence of their topography in the endoplasmic membranes

Some properties of two distinct rat brain sialyltransferases, acting on fetuin and asialofetuin, respectively, were investigated. These two membrane-bound enzymes were both strongly inhibited by charged phospholipids. Neutral phospholipids were without effect except lysophosphatidylcholine (lysoPC) which modulated these two enzymes in a different way. At 5 mM lysoPC, the fetuin sialyltransferase was solubilized and highly activated while the asialofetuin sialyltransferase was inhibited. Preincubation of brain microsomes with 4,4'-diisothiocyanatostilbene-2,2'-disulfonic acid (DIDS), known as a specific anion inhibitor and a non-penetrating probe, led to a moderate inhibition of the asialofetuin sialyltransferase just as in the case of the ovomucoid galactosyltransferase (used here as a marker for the luminal side of the Golgi membrane); under similar conditions, the fetuin sialyltransferase was strongly inhibited. In the presence of Triton X-100, which induced a disruption of membranes, all three enzymes were strongly inhibited by DIDS. Trypsin action on intact membranes showed that asialofetuin sialyltransferase, galactosyltransferase and fetuin sialyltransferase were all slightly inhibited. After membrane disruption by Triton X-100, the first two enzymes were completely inactivated by trypsin while the fetuin sialyltransferase was quite insensitive to trypsin treatment. From these data, we suggest that the fetuin sialyltransferase, accessible to DIDS, is an external enzyme, oriented closely towards the cytoplasmic side of the brain microsomal vesicles (endoplasmic and Golgi membranes), whereas the asialofetuin sialyltransferase is an internal enzyme, oriented in a similar manner to the galactosyltransferase. Moreover, the anion site (nucleotide sugar binding site) of the fetuin sialyltransferase must be different from its active site, as this enzyme, when solubilized, is strongly inhibited by DIDS while no degradation is observed in the presence of trypsin.

A brain sialyltransferase having a narrow specificity for O-glycosyl-linked oligosaccharide chains

The existence of a brain sialyltransferase catalyzing the specific transfer of NeuAc on native fetuin was demonstrated. This enzyme was not able to sialylate either asialofetuin or desialylated and nondesialylated orosomucoid, transferrin, and bovine submaxillary mucin. It required the presence of Mn2+ for optimal activity. Moreover, in fetuin, this activity was closely related to the proportion of NeuAc residues, but in liver tissue sialylation occurred only onto asialofetuin. In native fetuin, sialylation took place on O-glycan chains to give an O-disialyltetrasaccharidic structure. The Gal----GalNAc----protein was not an acceptor, but alpha-NeuAc-(2----3)-Gal----GalNAc----protein was, suggesting a specific transfer alpha-(2----6) to the GalNAc residue.

Involvement of phospholipids in the modulation of a membrane-bound brain fucosyltransferase

Microsomal fucosyltransferase isolated from sheep brain is strongly enhanced by charged lysophospholipids such as lysophosphatidylinositol and lysophosphatidic acid, while the corresponding phospholipids are inhibitive. Lysophosphatidylcholine (lyso-PC) also greatly increases the enzymatic activity and leads to its solubilization. Its stimulatory effect is related to the length of the fatty acyl chain involved in the lyso-PC structure: fatty acids C18 and C20 are less activating than the fatty acids C14-C16. Stimulation is restored when C18 fatty acids are unsaturated (e.g., C18:1-C18:3). Enzymatic activity enhancement is decreased when phosphatidylcholine structures are reformed by the addition of lyso-PC and the corresponding fatty acid. The physical state of these structures has no influence. These data provide evidence that bilayer structures do not modify enzymatic activity, while micellar structures formed by detergents and lysophospholipids lead to a strong increase in fucosyltransferase activity. However, lyso-PC does not interact in exactly the same way as Triton X-100. Although they both enhance the maximal velocity of fucosyltransferase for its two substrates, GDP-fucose and asialofetuin, the effect with lyso-PC is greater, and it clearly enables a better affinity for GDP-fucose. Endogenous phospholipids are also able to modify enzymatic activity. Hydrolysis of phosphatidylcholine by phospholipase A2 leads to an enzymatic stimulation.

Subcellular localisation of cerebral fucosyltransferase

GDP-fucose: asialofetuin fucosyltransferase from sheep brain was fractionated on a sucrose gradient into two activity peaks. Using purification on Ficoll adapted from the proposed method [(1980) J. Neurochem. 35, 281-296], double localisation of cerebral fucosyltransferase was confirmed and the subcellular active fractions identified as light microsomes and mitochondria.

Modulation of solubilized brain fucosyltransferase activity by phospholipids

Phospholipids interact on Triton X-100 solubilized GDP-fucose: asialofetuin fucosyltransferase (EC 2.4.1.68) isolated from sheep brain. This enzymatic activity is modulated by charged phospholipids. In particular, phosphatidic acid and analogues markedly inhibit the transfer of fucose from GDP-[14C]fucose. Kinetic studies show that phosphatidic acid interacts as a mixed inhibitor: the velocity and affinity of fucosyltransferase for the GDP-fucose and asialofetuin substrates are strongly decreased. However, this inhibitory effect is not related to stereospecificity, and the different parameters involved in the enzymatic reaction of glycosylation are not modified. The nature of fatty acids and chemical bond (ester or ether) occurring in the carbohydrate chain does not modify the behaviour of phosphatidic acid with respect to fucosyltransferase activity. Further, the physical state of phosphatidic acid (gel phase or liquid crystalline phase) has no influence. However, as the inhibition is closely pH-dependent, these data suggest that phosphatidic acid might directly interact with the active site of the enzyme and induce a conformational change.

Involvement of some amino acid residues in the enzymatic activity of solubilized cerebral fucosyltransferase

We have checked the effect of some chemical reagents specific for amino acid residues on the activity of a solubilized cerebral glycoprotein:fucosyltransferase. Diethylpyrocarbonate, 2,3-butanedione and tetranitromethane specific for histidyl, arginyl, and tyrosyl residues respectively, were strong inhibitors of the enzymatic activity This led us to conclude that these amino acid residues are "essential residues" in the cerebral fucosyltransferase activity.

Postnatal developmental changes in fucosyltransferase activity in rat brain

GDP-fucose:asialofetuin fucosyltransferase activity was studied during the postnatal development of rat brain. The enzymatic activity was very low during the first days of life and reached a maximum level around 21 days. This increase in enzymatic activity was characterized by two periods of rapid change. A rapid increase occurred between 3 and 7 days after birth, followed by a slow increase from 7 to 17 days, then a new rapid change from 17 to 21 days. Stimulation of the enzymatic activity by Triton X-100 increased with age. The development profiles of GDP-fucose pyrophosphatase and fucosidase did not change during this period.

Substrate specificity of cerebral GDP-fucose: glycoprotein fucosyltransferase

Solubilized sheep brain fucosyltransferase was shown to transfer fucose from GDP-fucose onto glycoprotein and glycopeptide acceptors, such as asialofetuin, asialotransferrin, their glycopeptides and glycopeptides from ovalbumin, but not on to monosaccharides and disaccharides such as galactose, N-acetylglucosamine and lactose. Competition studies between asialofetuin and glycopeptide V from ovalbumin provided evidence that both substrates compete for a common enzyme active site. The position of the fucosyl linkage was then investigated. Endo-beta-N-glucosaminidase D digestion of fucosylated and acetylated glycopeptide V showed that fucose is not linked to asparagine-linked N-acetylglucosamine. Hydrazinolysis and nitrous acid deamination performed on asialofetuin and glycopeptide V proved that fucose is not linked to external galactose or N-acetylglucosamine either. Thus we assume that fucose is linked to the oligomannochitobiosyl core of the glycan, and probably to the second N-acetylglucosamine.

[Demonstration by chromatofocalization of isoenzymes of sheep brain fucosyltransferase]

Glycoprotein : fucosyltransferase from sheep brain has previously been solubilized with detergent Triton X 100 and prepurified by hydrophobic chromatography on ethyl-agarose. Additional purification was performed using a new method, namely chromatofocalization. With a small column (30 ml) of P B E 94 Pharmacia eluted by a mixture of polybuffers 76 and 94, we were able to separate quickly four isoenzymes of the cerebral fucosyltransferase in the pH range 5-8. With regard to speed and ease, this technique seems to be very useful in purification of glycosyltransferases.

[Purification of solubilized fucosyltransferase from lamb brain using ethylagarose gel]

The glycoprotein: fucosyl-transferase of the cerebral hemispheres, assayed with desialylated fetuin as exogenous acceptor, was the most active of the protein: glycosyltransferases tested in the brain. The addition of Titron X-100 to the membrane suspension, followed by high speed centrifugation, led to a solubilization of the enzyme. The use of hydrophobic chromatography on ethylagarose gave a good purification of this solubilized fucosyl-transferase, whose homogeneity has been shown by Ultrogel AcA 22, DEAE-cellulose chromatography and disc electrophoresis.

[Kinetics of the reactions of glycosylation in the biosynthesis of glycoproteins: ordered bi-bi mechanism of glycosyltransferases and cooperativity]

Glycosyl-transferases respond to a Bi-Bi ordered mechanism, with two substrates and two products, completed by positive or negative cooperativity in relation to substrates concentration.