Novel Zebrafish Mono-α2,8-sialyltransferase (ST8Sia VIII): An Evolutionary Perspective of α2,8-Sialylation.

The mammalian mono-α2,8-sialyltransferase ST8Sia VI has been shown to catalyze the transfer of a unique sialic acid residues onto core 1 O-glycans leading to the formation of di-sialylated O-glycosylproteins and to a lesser extent to diSia motifs onto glycolipids like GD1a. Previous studies also reported the identification of an orthologue of the ST8SIA6 gene in the zebrafish genome. Trying to get insights into the biosynthesis and function of the oligo-sialylated glycoproteins during zebrafish development, we cloned and studied this fish α2,8-sialyltransferase homologue. In situ hybridization experiments demonstrate that expression of this gene is always detectable during zebrafish development both in the central nervous system and in non-neuronal tissues. Intriguingly, using biochemical approaches and the newly developed in vitro MicroPlate Sialyltransferase Assay (MPSA), we found that the zebrafish recombinant enzyme does not synthetize diSia motifs on glycoproteins or glycolipids as the human homologue does. Using comparative genomics and molecular phylogeny approaches, we show in this work that the human ST8Sia VI orthologue has disappeared in the ray-finned fish and that the homologue described in fish correspond to a new subfamily of α2,8-sialyltransferase named ST8Sia VIII that was not maintained in Chondrichtyes and Sarcopterygii.

Evidence for splice transcript variants of TMEM165, a gene involved in CDG.

BACKGROUND: Defects in TMEM165 gene cause a type-II Congenital Disorder of Glycosylation affecting Golgi glycosylation processes. TMEM165 patients exhibit psychomotor retardation, important osteoporosis, scoliosis, irregular epiphyses and thin bone cortex. TMEM165 protein is highly conserved in evolution and belongs to the family of UPF0016 membrane proteins which could be an unique group of Ca2+/H+ antiporters regulating Ca2+ and pH homeostasis and mainly localized in the Golgi apparatus. METHODS: RT-PCR from human brain tissues revealed TMEM165 splice-transcript variants. mRNA expression was analyzed by RT-Q-PCR. Expression plasmids allowed us to visualize isoform proteins and their subcellular localization. Their functions on glycosylation were achieved by looking at the gel mobility of highly glycosylated proteins in cells overexpressing isoforms. RESULTS: In this study, we highlight, as previously shown for other ion channels, the existence of TMEM165 splice-transcripts isoforms, in particular the Short-Form (SF) and the Long-Form (LF) transcripts, leading to a 129 aa and 259 aa protein isoform, respectively. These proteins both localize in the endoplasmic reticulum and have different effects on glycosylation compared to the wild-type protein (324 aa). We also point out that the SF is expressed at low levels in all human cells and tissues checked, excepted in brain, and forms homodimer. The LF was only expressed in the temporal lobe of human brain. GENERAL SIGNIFICANCE: The finding of numerous splice variants could lead to a family of TMEM165 isoforms. This family of TMEM165 splice transcripts could participate in the fine regulation of TMEM165 isoforms' functions and localizations.

Glucokinase expression is regulated by glucose through O-GlcNAc glycosylation.

Blood glucose fluctuates with the fasting-feeding cycle. One of the liver's functions is to maintain blood glucose concentrations within a physiological range. Glucokinase (GCK) or hexokinase IV, is the main enzyme that regulates the flux and the use of glucose in the liver leading to a compensation of hyperglycemia. In hepatocytes, GCK catalyzes the phosphorylation of glucose into glucose-6-phosphate. This critical enzymatic reaction is determinant for the metabolism of glucose in the liver which includes glycogen synthesis, glycolysis, lipogenesis and gluconeogenesis. In liver, simultaneous increase of glucose and insulin enhances GCK activity and gene expression, changes its subcellular location and interaction with regulatory proteins. The post-translational O-linked ß-N-acetylglucosaminylation (O-GlcNAcylation) acts as a glucose-sensitive modification and is believed to take part in hepatic glucose sensing by modifying key regulatory proteins. Therefore, we aimed to determine whether GCK is modified by O-GlcNAcylation in the liver of mice and investigated the role that this modification plays in regulating GCK protein expression. We demonstrated that endogenous GCK expression correlated with O-GlcNAc levels in the pathophysiological model ob/ob mice. More specifically, in response to the pharmacological inhibition of O-GlcNAcase (OGA) contents of GCK increased. Using the GlcNAc specific lectin succinylated-WGA and click chemistry labeling approaches, we demonstrated that GCK is modified by O-GlcNAcylation. Further, we demonstrated that siRNA-mediated Ogt knock-down not only decreases O-GlcNAc content but also GCK protein level. Altogether, our in vivo and in vitro results demonstrate that GCK expression is regulated by nutrient-sensing O-GlcNAc cycling in liver.

Evidence for an imbalance between tau O-GlcNAcylation and phosphorylation in the hippocampus of a mouse model of Alzheimer's disease.

Intracellular accumulation of hyperphosphorylated tau protein is linked to neuronal degeneration in Alzheimer's disease (AD). Mounting evidence suggests that tau phosphorylation and O-N-acetylglucosamine glycosylation (O-GlcNAcylation) are mutually exclusive post-translational modifications. O-GlcNAcylation depends on 3-5% of intracellular glucose that enters the hexosamine biosynthetic pathway. To our knowledge, the existence of an imbalance between tau phosphorylation and O-GlcNAcylation has not been reported in animal models of AD, as yet. Here, we used triple transgenic (3xTg-AD) mice at 12 months, an age at which hyperphosphorylated tau is already detected and associated with cognitive decline. In these mice, we showed that tau was hyperphosphorylated on both Ser396 and Thr205 in the hippocampus, and to a lower extent and exclusively on Thr205 in the frontal cortex. Tau O-GlcNAcylation, assessed in tau immunoprecipitates, was substantially reduced in the hippocampus of 3xTg-AD mice, with no changes in the frontal cortex or in the cerebellum. No changes in the expression of the three major enzymes involved in O-GlcNAcylation, i.e., glutamine fructose-6-phosphate amidotransferase, O-linked ß-N-acetylglucosamine transferase, and O-GlcNAc hydrolase were found in the hippocampus of 3xTg-AD mice. These data demonstrate that an imbalance between tau phosphorylation and O-GlcNAcylation exists in AD mice, and strengthens the hypothesis that O-GlcNAcylation might be targeted by disease modifying drugs in AD.

Detection and identification of O-GlcNAcylated proteins by proteomic approaches.

O-GlcNAcylation (O-linked beta-N-acetylglucosaminylation) is a widespread PTM confined within the nuclear, the cytosolic, and the mitochondrial compartments of eukaryotes. Recently, O-GlcNAcylation has been also detected in the close vicinity of plasma membranes particularly in lipid microdomains. The detection of this PTM can be easily done if appropriate controls and precautions are taken using a wide variety of tools including lectins, antibodies, or click-chemistry-based methods. In contrast, the identification of the proteins bearing O-GlcNAc moieties and the localization of the precise sites of O-GlcNAcylation remain challenging. This is due to the lability of the glycosidic bond between hydroxyl group of serine or threonine and N-acetylglucosamine using conventional fragmentation techniques such as CID. To tentatively overcome this technical limitation, electron-capture dissociation, or electron-transfer dissociation MS/MS are now used. Thanks to these breakthroughs, a large number of O-GlcNAc sites have been identified to date but these methodologies remain far from being used in routine.

Integrative view of α2,3-sialyltransferases (ST3Gal) molecular and functional evolution in deuterostomes: significance of lineage-specific losses.

Sialyltransferases are responsible for the synthesis of a diverse range of sialoglycoconjugates predicted to be pivotal to deuterostomes' evolution. In this work, we reconstructed the evolutionary history of the metazoan α2,3-sialyltransferases family (ST3Gal), a subset of sialyltransferases encompassing six subfamilies (ST3Gal I-ST3Gal VI) functionally characterized in mammals. Exploration of genomic and expressed sequence tag databases and search of conserved sialylmotifs led to the identification of a large data set of st3gal-related gene sequences. Molecular phylogeny and large scale sequence similarity network analysis identified four new vertebrate subfamilies called ST3Gal III-r, ST3Gal VII, ST3Gal VIII, and ST3Gal IX. To address the issue of the origin and evolutionary relationships of the st3gal-related genes, we performed comparative syntenic mapping of st3gal gene loci combined to ancestral genome reconstruction. The ten vertebrate ST3Gal subfamilies originated from genome duplication events at the base of vertebrates and are organized in three distinct and ancient groups of genes predating the early deuterostomes. Inferring st3gal gene family history identified also several lineage-specific gene losses, the significance of which was explored in a functional context. Toward this aim, spatiotemporal distribution of st3gal genes was analyzed in zebrafish and bovine tissues. In addition, molecular evolutionary analyses using specificity determining position and coevolved amino acid predictions led to the identification of amino acid residues with potential implication in functional divergence of vertebrate ST3Gal. We propose a detailed scenario of the evolutionary relationships of st3gal genes coupled to a conceptual framework of the evolution of ST3Gal functions.

O-GlcNAcylation stabilizes ß-catenin through direct competition with phosphorylation at threonine 41.

Dysfunctions in Wnt signaling increase ß-catenin stability and are associated with cancers, including colorectal cancer. In addition, ß-catenin degradation is decreased by nutrient-dependent O-GlcNAcylation. Human colon tumors and colons from mice fed high-carbohydrate diets exhibited higher amounts of ß-catenin and O-GlcNAc relative to healthy tissues and mice fed a standard diet, respectively. Administration of the O-GlcNAcase inhibitor thiamet G to mice also increased colonic expression of ß-catenin. By ETD-MS/MS, we identified 4 O-GlcNAcylation sites at the N terminus of ß-catenin (S23/T40/T41/T112). Furthermore, mutation of serine and threonine residues within the D box of ß-catenin reduced O-GlcNAcylation by 75%. Interestingly, elevating O-GlcNAcylation in human colon cell lines drastically reduced phosphorylation at T41, a key residue of the D box responsible for ß-catenin stability. Analyses of ß-catenin O-GlcNAcylation mutants reinforced T41 as the most crucial residue that controls the ß-catenin degradation rate. Finally, inhibiting O-GlcNAcylation decreased the ß-catenin/α-catenin interaction necessary for mucosa integrity, whereas O-GlcNAcase silencing improved this interaction. These results suggest that O-GlcNAcylation regulates not only the stability of ß-catenin, but also affects its localization at the level of adherens junctions. Accordingly, we propose that O-GlcNAcylation of ß-catenin is a missing link between the glucose metabolism deregulation observed in metabolic disorders and the development of cancer.

Assessing ER and Golgi N-glycosylation process using metabolic labeling in mammalian cultured cells.

Modifications of N-glycosylation in disease states are common and illustrate the crucial requirement of glycosylation in human biology. Mainly based on glycan permethylation and the use of mass spectrometry analysis, we can easily understand that many different methods to analyze the N-glycome have seen the day. While extremely powerful, these methods are mainly used to analyze qualitative variations of N-glycosylation of human serum proteins and do not necessarily reflect the glycosylation status of derived mammalian cultured cells. This chapter summarizes two methods that we are routinely using in our laboratory to assess the ER and Golgi N-glycosylation process. The proposed methodology allows pinpointing ER as well as Golgi glycosylation deficiencies in mammalian cultured cells. The first approach is based on direct metabolic labeling of cultured mammalian cells with [2-(3)H] mannose followed by sequential extraction and HPLC analysis of the purified oligosaccharides. The second one is based on the copper-catalyzed azide alkyne cycloaddition (CuAAC) strategy. We propose the use of alkyne-tagged sialic acid (SialNAl) to visualize the Golgi glycosylation efficiency. Their metabolic incorporation into newly synthesized glycoproteins can then be chemoselectively coupled to complementary azide-functionalized fluorophores, and visualized by using confocal laser scanning microscopy. To summarize, we present here a detailed description of our know-how in the field of ER and Golgi N-glycosylation.

The hexosamine biosynthetic pathway and O-GlcNAcylation drive the expression of ß-catenin and cell proliferation.

The short half-life protooncogene ß-catenin acquires a remarkable stability in a large subset of cancers, mainly from mutations affecting its proteasomal degradation. In this sense, colorectal cancers (CRC) form a group of pathologies in which early steps of development are characterized by an aberrant expression of ß-catenin and an uncontrolled proliferation of epithelial cells. Diet has long been described as an influence in the emergence of CRC, but the molecular events that link metabolic disorders and CRC remain elusive. Part of the explanation may reside in hexosamine biosynthetic pathway (HBP) flux. We found that fasted mice being force-fed with glucose or glucosamine leads to an increase of ß-catenin and O-GlcNAcylation levels in the colon. MCF7 cells possessing intact Wnt/ß-catenin signaling heavily expressed ß-catenin when cultured in high glucose; this was reversed by the HBP inhibitor azaserine. HBP inhibition also decreased the expression of ß-catenin in HT29 and, to a lesser extent, HCT116 cells. The same observation was made with regard to the transcriptional activity of ß-catenin in HEK293 cells. Inhibition of HBP also blocked the glucose-mediated proliferation capacity of MCF7 cells, demonstrating that glucose affects both ß-catenin expression and cell proliferation through the HBP. The ultimate element conducting these events is the dynamic posttranslational modification O-GlcNAcylation, which is intimately linked to HBP; the modulation of its level affected the expression of ß-catenin and cell proliferation. In accordance with our findings, we propose that metabolic disorders correlate to CRC via an upregulation of HBP that reverberates on high O-GlcNAcylation levels including modification of ß-catenin.

[Signaling and metabolic predispositions linked to the colorectal cancer]. / Signalisation et prédispositions métaboliques liées au cancer colorectal.

The setting up and the progression of the colorectal cancer (CCR) follow a sequence of events that are spatio-temporally rigorously orchestrated. The failures that specifically target the signaling pathways responsible for the cancerization of the colorectal mucosa have been well described and among these it seems that a dysregulation of the Wnt/ß-catenin pathway is involved in the triggering of near 90 % of the cases. It has been also described that several risk factors linked to metabolic disorders (feeding, insulin resistance, metabolic syndrome, etc.) predispose individuals to CCR but no rational explanations were given. We propose that, since it is implicated in the control of the insulin pathway among other actions, the nutritional sensor O-GlcNAcylation may be the element linking these metabolic disorders to CCR.

Identification of phosphorylated oligosaccharides in cells of patients with a congenital disorders of glycosylation (CDG-I).

Protein N-glycosylation is initiated by the dolichol cycle in which the oligosaccharide precursor Glc(3)Man(9)GlcNAc(2)-PP-dolichol is assembled in the endoplasmic reticulum (ER). One critical step in the dolichol cycle concerns the availability of Dol-P at the cytosolic face of the ER membrane. In RFT1 cells, the lipid-linked oligosaccharide (LLO) intermediate Man(5)GlcNAc(2)-PP-Dol accumulates at the cytosolic face of the ER membrane. Since Dol-P is a rate-limiting intermediate during protein N-glycosylation, continuous accumulation of Man(5)GlcNAc(2)-PP-Dol would block the dolichol cycle. Hence, we investigated the molecular mechanisms by which accumulating Man(5)GlcNAc(2)-PP-Dol could be catabolized in RFT1 cells. On the basis of metabolic labeling experiments and in comparison to human control cells, we identified phosphorylated oligosaccharides (POS), not found in human control cells and present evidence that they originate from the accumulating LLO intermediates. In addition, POS were also detected in other CDG patients' cells accumulating specific LLO intermediates at different cellular locations. Moreover, the enzymatic activity that hydrolyses oligosaccharide-PP-Dol into POS was identified in human microsomal membranes and required Mn(2+) for optimal activity. In CDG patients' cells, we thus identified and characterized POS that could result from the catabolism of accumulating LLO intermediates.

Overexpression of Man2C1 leads to protein underglycosylation and upregulation of endoplasmic reticulum-associated degradation pathway.

Unfolded glycoproteins retained in the endoplasmic reticulum (ER) are degraded via the ER-associated degradation (ERAD) pathway. These proteins are subsequently transported to the cytosol and degraded by the proteasomal complex. Although the sequential events of ERAD are well described, its regulation remains poorly understood. The cytosolic mannosidase, Man2C1, plays an essential role in the catabolism of cytosolic free oligomannosides, which are released from the degraded proteins. We have investigated the impact of Man2C1 overexpression on protein glycosylation and the ERAD process. We demonstrated that overexpression of Man2C1 led to modifications of the cytosolic pool of free oligomannosides and resulted in accumulation of small Man(2-4)GlcNAc(1) glycans in the cytosol. We further correlated this accumulation with incomplete protein glycosylation and truncated lipid-linked glycosylation precursors, which yields an increase in N-glycoprotein en route to the ERAD. We propose a model in which high mannose levels in the cytosol interfere with glucose metabolism and compromise N-glycan synthesis in the ER. Our results show a clear link between the intracellular mannose-6-phosphate level and synthesis of the lipid-linked precursors for protein glycosylation. Disturbance in these pathways interferes with protein glycosylation and upregulated ERAD. Our findings support a new concept that regulation of Man2C1 expression is essential for maintaining efficient protein N-glycosylation.

Differential effects of lobe A and lobe B of the Conserved Oligomeric Golgi complex on the stability of {beta}1,4-galactosyltransferase 1 and {alpha}2,6-sialyltransferase 1.

Initially described by Jaeken et al. in 1980, congenital disorders of glycosylation (CDG) is a rapidly expanding group of human multisystemic disorders. To date, many CDG patients have been identified with deficiencies in the conserved oligomeric Golgi (COG) complex which is a complex involved in the vesicular intra-Golgi retrograde trafficking. Composed of eight subunits that are organized in two lobes, COG subunit deficiencies have been associated with Golgi glycosylation abnormalities. Analysis of the total serum N-glycans of COG-deficient CDG patients demonstrated an overall decrease in terminal sialylation and galactosylation. According to the mutated COG subunits, differences in late Golgi glycosylation were observed and led us to address the question of an independent role and requirement for each of the two lobes of the COG complex in the stability and localization of late terminal Golgi glycosylation enzymes. For this, we used a small-interfering RNAs strategy in HeLa cells stably expressing green fluorescent protein (GFP)-tagged ß1,4-galactosyltransferase 1 (B4GALT1) and α2,6-sialyltransferase 1 (ST6GAL1), two major Golgi glycosyltransferases involved in late Golgi N-glycosylation. Using fluorescent lectins and flow cytometry analysis, we clearly demonstrated that depletion of both lobes was associated with deficiencies in terminal Golgi N-glycosylation. Lobe A depletion resulted in dramatic changes in the Golgi structure, whereas lobe B depletion severely altered the stability of B4GALT1 and ST6GAL1. Only MG132 was able to rescue their steady-state levels, suggesting that B4GALT1- and ST6GAL1-induced degradation are likely the consequence of an accumulation in the endoplasmic reticulum (ER), followed by a retrotranslocation into the cytosol and proteasomal degradation. All together, our results suggest differential effects of lobe A and lobe B for the localization/stability of B4GALT1 and ST6GAL1. Lobe B would be crucial in preventing these two Golgi glycosyltransferases from inappropriate retrograde trafficking to the ER, whereas lobe A appears to be essential for maintaining the overall Golgi structure.

Arginine 469 is a pivotal residue for the Hsc70-GlcNAc-binding property.

The members of the 70kDa-heat shock proteins (HSP70) family play numerous fundamental functions in the cell such as promoting the assembly of multimeric complexes or helping the correct folding of nascent proteins to take place. In numerous previous studies we demonstrated that Hsp70 and its constitutive isoform Hsc70 are endowed of a GlcNAc-binding activity. The molecular modeling of the substrate binding domain of Hsc70 and in silico docking experiments using Ser/Thr-O-GlcNAc motifs allowed to define the potential carbohydrate-recognition region and to point out the crucial position of Arg469 as an amino-acid directly interacting with the sugar moiety. We cloned a flagged Hsc70 in a pCMV.SPORT6 vector and we showed that the mutation R469A decreased the GlcNAc-binding property of the chaperone of around 70%. This is the first work reporting the localization of the GlcNAc-binding domain of a member of the HSP70 family.

Molecular phylogeny and functional genomics of beta-galactoside alpha2,6-sialyltransferases that explain ubiquitous expression of st6gal1 gene in amniotes.

Sialyltransferases are key enzymes in the biosynthesis of sialoglycoconjugates that catalyze the transfer of sialic residue from its activated form to an oligosaccharidic acceptor. ß-Galactoside α2,6-sialyltransferases ST6Gal I and ST6Gal II are the two unique members of the ST6Gal family described in higher vertebrates. The availability of genome sequences enabled the identification of more distantly related invertebrates' st6gal gene sequences and allowed us to propose a scenario of their evolution. Using a phylogenomic approach, we present further evidence of an accelerated evolution of the st6gal1 genes both in their genomic regulatory sequences and in their coding sequence in reptiles, birds, and mammals known as amniotes, whereas st6gal2 genes conserve an ancestral profile of expression throughout vertebrate evolution.

Dysregulation of the nutrient/stress sensor O-GlcNAcylation is involved in the etiology of cardiovascular disorders, type-2 diabetes and Alzheimer's disease.

O-GlcNAcylation is widespread within the cytosolic and nuclear compartments of cells. This post-translational modification is likely an indicator of good health since its intracellular level correlates with the availability of extracellular glucose. Apart from its status as a nutrient sensor, O-GlcNAcylation may also act as a stress sensor since it exerts its fundamental effects in response to stress. Several studies report that the cell quickly responds to an insult by elevating O-GlcNAcylation levels and by unmasking a newly described Hsp70-GlcNAc binding property. From a more practical point of view, it has been shown that O-GlcNAcylation impairments contribute to the etiology of cardiovascular diseases, type-2 diabetes and Alzheimer's disease (AD), three illnesses common in occidental societies. Many studies have demonstrated that O-GlcNAcylation operates as a powerful cardioprotector and that by raising O-GlcNAcylation levels, the organism more successfully resists trauma-hemorrhage and ischemia/reperfusion injury. Recent data have also shown that insulin resistance and, more broadly, type-2 diabetes can be controlled by O-GlcNAcylation of the insulin pathway and O-GlcNAcylation of the gluconeogenesis transcription factors FoxO1 and CRCT2. Lastly, the finding that AD may correspond to a type-3 diabetes offers new perspectives into the knowledge of the neuropathology and into the search for new therapeutic avenues.

Molecular cloning and characterization of the expression pattern of the zebrafish alpha2, 8-sialyltransferases (ST8Sia) in the developing nervous system.

Sialyltransferases are Golgi type II transmembrane glycoproteins involved in the biosynthesis of sialylated glycolipids and glycoproteins. These sialylated compounds play fundamental roles in the development of a variety of tissues including the nervous system. In this study, we have molecularly cloned from zebrafish sources, the orthologues of the six human alpha2,8-sialyltransferases (ST8Sia), a family of sialyltransferases implicated in the alpha2-8-mono-, oligo-, and poly-sialylation of glycoproteins and gangliosides and we have analysed their expression pattern in the embryonic zebrafish nervous system, using in situ hybridization. Our results show that all six ST8Sia exhibit distinct and overlapping patterns of expression in the developing zebrafish central nervous system with spatial and temporal regulation of the expression of these genes, which suggests a role for the alpha2-8-sialylated compounds in the developing nervous system.

Protein ubiquitination is modulated by O-GlcNAc glycosylation.

During the past two decades, O-GlcNAc modification of cytosolic and nuclear proteins has been intensively studied. Nevertheless, the function of this post-translational modification remains unclear. It has been recently speculated that O-GlcNAc could act as a protective signal against proteasomal degradation, both by modifying target substrates and/or by inhibiting the proteasome itself. In this work, we have investigated the putative relation between O-GlcNAc and the ubiquitin pathway. First, we showed that the level of both modifications increased rapidly after thermal stress but, unlike ubiquitinated proteins, O-GlcNAc-modified proteins failed to be stabilized by inhibiting proteasome function. Increasing O-GlcNAc levels, using glucosamine or PUGNAc, enhanced ubiquitination. Inversely, when O-GlcNAc levels were reduced, using forskolin or glucose deprivation, ubiquitination decreased. Targeted-RNA interference of O-GlcNAc transferase also reduced ubiquitination and moreover halved cell thermotolerance. Finally, we demonstrated that the ubiquitin-activating enzyme E1 was O-GlcNAc modified and that its glycosylation and its interaction with Hsp70 varied according to the conditions of cell culture. Altogether, these results show that O-GlcNAc and ubiquitin are not strictly antagonistic post-translational modifications, but rather that the former might regulate the latter, and also suggest that E1 could be one of the common links between the two pathways.

Hsp70-GlcNAc-binding activity is released by stress, proteasome inhibition, and protein misfolding.

Numerous recent works strengthen the idea that the nuclear and cytosolic-specific O-GlcNAc glycosylation protects cells against injuries. We have first investigated O-GlcNAc level and Hsp70-GlcNAc-binding activity (HGBA) behaviour after exposure of HeLa and HepG(2) cells to a wide variety of stresses. O-GlcNAc and HGBA responses were different according to the stress and according to the cell. HGBA was released for almost all stresses, while O-GlcNAc level was modified either upwards or downwards, depending to the stress. Against all expectations, we demonstrated that energy charge did not significantly vary with stress whereas UDP-GlcNAc pools were more dramatically affected even if differences in UDP-GlcNAc contents were not correlated with O-GlcNAc variations suggesting that O-GlcNAc transferase is itself finely regulated during cell injury. Finally, HGBA could be triggered by proteasome inhibition and by L-azetidine-2-carboxylic acid (a proline analogue) incorporation demonstrating that protein misfolding is one of the key-activator of this Hsp70 property.

Discrimination between lumenal and cytosolic sites of deglycosylation in endoplasmic reticulum-associated degradation of glycoproteins by using benzyl mannose in CHO cell lines.

Recent studies demonstrated that deglycosylation step is a prerequisite for endoplasmic reticulum (ER)-associated degradation of misfolded glycoproteins. Here, we report the advantages of using benzyl mannose during pulse-chase experiments to study the subcellular location of the deglycosylation step in Chinese hamster ovary (CHO) cell lines. Benzyl mannose inhibited both the ER-to-cytosol transport of oligomannosides and the trimming of cytosolic-labeled oligomannosides by the cytosolic mannosidase in vivo. We pointed out the occurrence of two subcellular sites of deglycosylation. The first one is located in the ER lumen, and led to the formation of Man8GlcNAc2 (isomer B) in wild-type CHO cell line and Man4GlcNAc2 in Man-P-Dol-deficient cell line. The second one was revealed in CHO mutant cell lines for which a high rate of glycoprotein degradation was required. It occurred in the cytosol and led to the liberation of oligosaccharides species with one GlcNAc residue and with a pattern similar to the one bound onto glycoproteins. The cytosolic deglycosylation site was not specific for CHO mutant cell lines, since we demonstrated the occurrence of cytosolic pathway when the formation of truncated glycans was induced in wild-type cells.