| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
UMR USTL/CNRS, Laboratoire de Glycobiologie Structurale et Fonctionnelle, Villeneuve d’Ascq, France
1Correspondence: UMR USTL/CNRS 8576, Laboratoire de Glycobiologie Structurale et Fonctionnelle, IFR 147, 59655 Villeneuve d’Ascq, France. E-mail: tony.lefebvre{at}univ-lille1.fr
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: OGT RNA interference • heat shock • proteasomal degradation
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. This process is observed in both healthy and injured cells. In addition, the ubiquitin-proteasome pathway also plays an important role in the regulation of the level of proteins with a short half-life, such as the cyclins acting during the cell cycle (2)
or transcription factors (3)
. The proteasome contributes also in the elimination of chemically or metabolically damaged proteins that contain hydrophobic peptide segments, which have a tendency to aggregate. These proteins are toxic for cell homeostasis and require a rapid degradation by a cellular mechanism known as the ubiquitin-proteasome system (UPS) (4)
.
It has been recently speculated that proteins could be protected against proteasomal degradation by O-linked N-acetylglucosaminylation (O-GlcNAc) (5
6
7
8)
.O-GlcNAc is the major glycosylation type found within the cytosolic and nuclear compartments of eukaryotic cell (9
, 10)
. The O-GlcNAc process results in the covalent linkage performed by the O-GlcNAc transferase (OGT) of a single N-acetylglucosamine residue to a serine or a threonine residue of substrate proteins. This O-GlcNAc protein modification does not require a strict consensus sequence on the polypeptide chain; however, OGT needs a specific peptidic environment: usually O-GlcNAc linkage occurs near a proline residue and in serine/threonine-enriched regions. Interestingly, OGT modifies sequences that are similar to PEST sequences (peptide segments enriched in proline, glutamic acid, serine, and threonine residues), the latter being a signal targeting the proteins through the ubiquitin-proteasome pathway (11)
. Activation of PEST sequences frequently occurs after phosphorylation (12)
. Because there exists a relation between O-GlcNAc and phosphorylation, it can be assumed that O-GlcNAc counteracts the phosphorylation effect in order to protect proteins. Indeed, it has been recently demonstrated that modification of p53 at serine 149 by O-GlcNAc prevents its degradation by decreasing its phosphorylation at threonine 155 leading to a blockade of ubiquitination (13)
. Another study has shown that the O-GlcNAc glycosylation of Sp1 reduced its susceptibility to proteasomal destruction as studied in cells cultured in a medium supplemented with glucosamine and that on the contrary, glucose starvation was shown to diminish the half-life of this transcription factor (5)
. Another example that strongly supports a protective function of O-GlcNAc is that of the β-estrogen-receptor (β-ER), which comprises a O-GlcNAc motif in a peptide sequence with a high PEST sequence score (6)
. The glycosylated form of β-ER was more resistant to proteasomal degradation than the unglycosylated form, whereas a mutant that mimicked a constitutive phosphorylated β-ER form was more sensitive to the degradation. Recently, it has also been demonstrated that the proteasome itself could be regulated through the O-GlcNAc glycosylation of the 19S-regulatory subunit (7)
. This inhibition of proteasome by O-GlcNAc adds another glycosylation-mediated regulatory level of protein protection. Finally, our group has demonstrated that the constitutive form of the 70-kDa heat-shock protein family (Hsc70) and the inducible one (Hsp70) displayed a lectin activity toward O-GlcNAc residues (14
, 15)
. This affinity was enhanced when cells were stressed (15
, 16)
. Accordingly, one can hypothesize that OGT and Hsp70 can act in synergy to prevent proteins from aggregation and degradation: first, by modifying substrates with O-GlcNAc and second, by Hsp70 binding on the glycosylated proteins.
In this report, we first corroborate previous published data (17)
showing that after thermal stress O-GlcNAc level was enhanced: this rapid O-GlcNAc rise occurs in parallel to the ubiquitin pathway activation. In addition, we show that proteasome inhibition stabilizes protein ubiquitination but not protein O-GlcNAc-modification. Modulation of the O-GlcNAc content by chemical inhibitors or by OGT RNA interference affects ubiquitination. Surprisingly, E1, an ubiquitin-activating enzyme that initiates the process of protein ubiquitination, is modified with O-GlcNAc and physically interacts with Hsp70. Both, E1 glycosylation and E1-Hsp70 interaction vary with cell culture conditions and with induced stress, suggesting a possible regulation of protein ubiquitination by E1 O-GlcNAc modification.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Thermal stress was induced by placing cells at 42°C in a 5% CO2 enriched atmosphere either for 24 h or for the indicated time periods (See Fig. 1
; kinetic experiments).
|
MG132 (N-carbobenzoxyl-Leu-Leu-leucinal, 8 mM stock solution in dimethyl sulfoxide at –80°C; Sigma) was used at a final concentration of 8 µM. PUGNAc [O-(2-acetamido-2-deoxy-D-glucopyranosylidene)amino N-phenylcarbamate] was used at a final concentration of 250 µM, and forskolin was used at a final concentration of 50 µM.
Immunoprecipitation and coimmunoprecipitation
HepG2 cells were first washed with 10 ml of cold phosphate-buffered saline (PBS). For immunoprecipitation experiments, cells were lysed on ice with lysis buffer [10 mM Tris/HCl, 150 mM NaCl, 1% Triton X-100 (v/v), 0.5% sodium deoxycholate (w/v), 0.1% sodium dodecyl sulfate (w/v) and protease inhibitors, pH 7.4]. Cell extracts were then centrifuged at 20,000 g for 30 min at 4°C. Supernatants were incubated together with the mouse monoclonal anti-O-GlcNAc antibody [RL2; Affinity Bioreagents (Ozyme), St. Quentin Yvelines, France] at a final dilution of 1:400 or with the rabbit polyclonal anti-ubiquitin (Stressgen, Tebu-bio, Le Perray en Yvelines, France) at a final dilution of 1:500 and placed at 4°C overnight. For immunoprecipitation of E1, the polyclonal anti-E1 antibody (ab34711; Abcam, Cambridge, UK) was used at a final concentration of 1:500. Antibody-bound proteins were recovered after adding either 30 µl of Sepharose-labeled protein G for RL2, or 30 µl of Sepharose-labeled protein A for anti-ubiquitin and anti-E1 (GE Healthcare, Templemars, France) for 1 h at 4°C. Beads were gently centrifuged for 1 min and subsequently washed with the following buffers: lysis buffer, lysis buffer supplemented with 500 mM NaCl, lysis buffer/TNE (10 mM Tris/HCl, 150 mM NaCl, and 1 mM EDTA, pH 7.4) (v/v), and finally with TNE alone.
For coimmunoprecipitation, cells were lysed on ice in a lysis buffer containing 20 mM Tris/HCl, 150 mM NaCl, 0.5% Nonidet P-40 (v/v), and protease inhibitors, pH 8.0. Whole cell extracts were centrifuged at 20,000 g for 30 min at 4°C, and supernatants were collected. Immunoprecipitation with the anti-Hsp70 antibody (Stressgen) was carried out at a final concentration of 5 µl/ml overnight at 4°C, followed by an incubation with Sepharose-labeled protein A for 1 h at 4°C. Beads were gently centrifuged for 1 min and washed 4x for 5 min each with the lysis buffer.
Controls for the immunoprecipitation specificities were performed with normal mouse IgG1 (Sigma) and with normal rabbit IgG (Santa Cruz Biotechnologies, Santa Cruz, CA, USA).
SDS-PAGE, Western blot analysis, and antibody staining
Equal amounts of extracted proteins were subjected to Western blot analysis. Samples were analyzed by 10% SDS-PAGE under reducing conditions, and proteins were electroblotted on nitrocellulose sheet (GE Healthcare). Equal loading was verified using Ponceau red staining and by the use of anti-tubulin or anti-actin antibodies. Membranes were first saturated for 45 min with 5% nonfatty acid milk in Tris-buffered saline (TBS) -Tween buffer [20 mM Tris/HCl, 150 mM NaCl, and 0.05% Tween (v/v), pH 8.0]. Mouse monoclonal anti-O-GlcNAc (RL2), mouse monoclonal anti-ubiquitin (Sigma), and mouse monoclonal anti-E1 (2G2.3–5, Sigma) antibodies were used at a final dilution of 1:1000. Rabbit polyclonal anti-OGT (AL28; kind gift from Gerald W. Hart, Johns Hopkins Institute, Baltimore, MD, USA) was used at a dilution of 1:2000; rabbit polyclonal anti-ubiquitin antibody, at a dilution of 1:10,000; rabbit polyclonal anti-E1 antibody (ab34711; Abcam), at a dilution of 1:5000; rabbit polyclonal anti-tubulin antibody, at a dilution of 1:500; and rabbit polyclonal anti-actin, at a dilution of 1:10,000. The specificity of RL2 staining was ensured at different stages of the experiments by incubating the first antibody with 0.5 M of free N-acetylglucosamine. Membranes were incubated with the different antibodies overnight at 4°C, then washed 3x with TBS-Tween for 10 min and incubated with either an anti-rabbit or an anti-mouse horseradish peroxidase-labeled secondary antibody at a dilution of 1:10,000 for 1 h. Finally, 3 washes of 10 min each were performed with TBS-Tween, and the detection was carried out with enhanced chemiluminescence (GE Healthcare).
Densitometry analyses of the Western blots were done with the GeneTools software (File version: 3.07.03; Syngene, Cambridge, UK).
RNA interference (RNAi)
RNAis were designed for the human OGT sequence. Oligonucleotides were designed, synthesized, and purchased from Eurogentec (Angers, France). Double-stranded oligonucleotide (named 1153) used in this study was the following: 5'GGA GGC UAU UCG AAU CAG U 3' (antisense sequence: 5'ACU GAU UCG AAU AGC CUC C 3'). As a negative control, RNAi oligonucleotides designed for green fluorescent protein (GFP) were used (sense sequence: 5' GAA CGG CAU CAA GGU GAA CTT 3'; antisense sequence: 5' GUU CAC CUU GAU GCC GUU CTT 3'). DreamFect reagent (OZ Biosciences, Marseille, France; 8 µl) was diluted with serum- and antibiotic free Opti-MEM I Reduced Serum Medium (1x) liquid with GlutaMAX I medium (Opti-MEM/GlutaMAX; Invitrogen) to a final volume of 100 µl. RNAi (2 µg) was diluted with Opti-MEM/GlutaMAX to a final volume of 100 µl. The 100 µl diluted transfection reagent and the 100 µl diluted RNAi solution were then mixed and incubated for 20 min. The 200 µl RNAi solution was added to HeLa cells maintained in 1.8 ml of Opti-MEM/GlutaMAX per well in a 6-well plate. This procedure was repeated every 24 h for 4 days. Ninety-six hours after incubation with the oligonucleotides, OGT expression and activity were tested by immunoblotting, either with rabbit anti-OGT antibodies (AL28) to test OGT levels or with RL2 antibodies to check O-GlcNAc levels. Viability of the transfected cells was enquired by duplicating the transfections in a 24-well plate (all volumes and quantities were divided by 2), using the trypan blue exclusion method (16)
.
Sambucus nigra agglutinin (SNA) staining
After OGT silencing, crude cellular extracts were run on a 10% SDS-PAGE and electroblotted; membranes were then incubated with the digoxigenin-coupled SNA (DIG Glycan differentiation kit; Roche, Meylan, France) at a final concentration of 1:1000 in TBS-Tween for 1 h at room temperature. This incubation was followed by 3 washes of 10 min each with TBS-Tween. The membranes were incubated with an alkaline phosphatase-labeled anti-digoxigenin secondary antibody at a final concentration of 1:1000 in TBS-Tween for 1 h. After 3 washes of 10 min each with TBS-Tween, proteins were visualized by the addition of 5-bromo-4-chloro-3-indolyl-phosphate and nitro-blue tetrazolium.
Immunofluorescence experiment
HepG2 and HeLa cells were grown on coverglasses for 48 h and afterward washed twice with cold PBS. For RNAi experiments, the amount of each reagent was adjusted to a 6-well plate. Cells were fixed in 3% of paraformaldehyde in cold PBS for 15 min and subsequently washed with PBS. Excess of paraformaldehyde was eliminated with a solution of 50 mM ammonium chloride for 10 min. After washing with PBS, cells were permeabilized with 0.1% Triton-X100 for 5 min. Nonspecific sites were blocked with goat serum. Coverglasses were then incubated for 30 min with either anti-O-GlcNAc (RL2), anti-OGT (AL28), or anti-ubiquitin, each at a dilution of 1:100 in a 10% goat serum solution (in PBS). After 3 washes with PBS, the coverglasses were covered with fluorescein isothiocyanate (FITC) and Texas Red-labeled secondary antibodies (dilution 1:50). Nuclei were specifically stained with 4',6'-diamidino-2-phenylindole (DAPI). Cells were visualized using an Axioplan 2 imaging microscope (Zeiss, Jena, Germany) and an Axio Cam HRc camera (AxioVision; Zeiss).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Heat-shock treatment induces changes in levels of O-GlcNAc and ubiquitin
We were interested in establishing the evolution of ubiquitination and O-GlcNAc glycosylation of cells during thermal stress situations. For that purpose, we investigated the ubiquitin and O-GlcNAc modifications of proteins in HepG2 cells after exposure to thermal stress (42°C) for increasing time periods. This experiment was performed either in the presence or in the absence of MG132, a reversible proteasome inhibitor (Fig. 1)
. For each condition, the amounts of O-GlcNAc and ubiquitin were evaluated by Western blot analysis (Fig. 1A
). Afterward, the blots were densitometrically measured, and the values were accordingly reported in histograms (Fig. 1B
). As illustrated in Fig. 1A, B
, the content of both ubiquitin and O-GlcNAc rapidly increased in HepG2 cells during thermal stress (Fig. 1A
, lane 5; 
5 min after onset of stress). However, ubiquitin-containing proteins reached their maximal level after 10 min of exposure to 42°C (Fig. 1A
, lane 10), whereas the levels of O-GlcNAc increased steadily to a maximum level at 15 min of exposure to 42°C (Fig. 1)
. These results are consistent with data previously published by Zachara et al. (17)
showing that O-GlcNAc levels increased rapidly in response to a large variety of stresses. Intriguingly, in contrast to ubiquitin, the O-GlcNAc responses were insensitive to the inhibition of the proteasome (Fig. 1A, B
). To verify the specificity of the O-GlcNAc staining, blots were restained with RL2 in the presence of free N-acetylglucosamine. All of the RL2 staining was abrogated by the free N-acetylglucosamine, showing that the staining was specific.
O-GlcNAc and ubiquitin modify proteins concomitantly
In our quest to better understand the putative relation or interplay between ubiquitin and O-GlcNAc, we examined the nuclear and the cytosolic distribution of O-GlcNAc and ubiquitinated proteins. These studies were performed by immunofluorescence microscopy, using the monoclonal anti-O-GlcNAc antibody RL2 in conjunction with an FITC-coupled secondary antibody and a polyclonal anti-ubiquitin antibody in conjunction with a Texas Red-coupled secondary antibody. The nucleus was stained with DAPI (blue coloration). As shown in Fig. 2
A, ubiquitin labeling (
-Ub) was mainly present in the cytosolic compartment. By superimposing the images of ubiquitinated stained proteins and DAPI staining (Fig. 2A
; -Ub/DAPI) a few red spots were also visible in the nucleus. In contrast, as shown in merged pictures with DAPI, O-GlcNAc glycosylation was predominantly localized in the nucleus. A more diffuse green color was also visible in the cytoplasm. By superimposing both the ubiquitin and the O-GlcNAc stained images, colocalization of the two PTMs was shown to be restricted to the cytoplasm (orange).
|
We hypothesized that some O-GlcNAc proteins could be ubiquitinated concurrently. To verify this hypothesis, we performed immunoprecipitation experiments. HepG2 cell extracts were immunoprecipitated either with the anti-O-GlcNAc or the anti-ubiquitin antibodies. Subsequently, anti-O-GlcNAc and anti-ubiquitin precipitated proteins were analyzed by Western blot and visualized both with the anti-ubiquitin antibodies (Fig. 2B
, bottom panel) and the anti-O-GlcNAc antibody (top panel). We observed that some proteins could be both O-GlcNAc glycosylated and modified with ubiquitin. The staining of the two antibodies was more intense when cells were incubated both with glucosamine and MG132. The specificity of the different immunoprecipitation experiments performed with MG132 was ensured by using normal mouse IgG1 and normal rabbit IgG (Fig. 2B
, IP mIgG and rIgG). Staining of the whole cell lysates (WCL) with the anti-O-GlcNAc and the anti-ubiquitin is also shown. The specificity of RL2 was checked by incubating the antibody with free N-acetylglucosamine. Equal loading was ascertained by staining with the anti-actin antibody.
Perturbation of intracellular O-GlcNAc content leads to variations in ubiquitination status
In our next set of experiments, we wanted to examine whether the levels of MG132-stabilized ubiquitinated proteins were modulated by O-GlcNAc content increase and decrease. For that purpose, we decided to modulate the efficiency of the hexosamine biosynthetic pathway (HBP) which produces UDP-GlcNAc (the donor for O-GlcNAc addition). O-GlcNAc decrease was induced either by glucose starvation (cell viability: 83.4±3.57 vs. 93.1±5.65% for control cells) or by treating cells with an activator of adenyl cyclase namely forskolin (cell viability: 93.2±3.42%). Activation of adenyl cyclase leads to the activation of PKA, which, in turn, blocks the HBP. O-GlcNAc increase was performed by culturing cells in presence of extra glucosamine (cell viability: 94.9±0.79%). Surprisingly, ubiquitin and O-GlcNAc levels were reduced both by deprivation of glucose and treatment with forskolin (Fig. 3
A). As illustrated in Fig. 3A
by the histograms of the densitometrical measured values adjusted to tubulin expression, the levels of reduction were different for ubiquitinated and for O-GlcNAc modified proteins (Fig. 3A
). Interestingly, the treatment of HepG2 with glucosamine rescued both the O-GlcNAc level and, to a lesser extent, the ubiquitin level that was diminished by glucose deprivation, while the ubiquitin level remained less than that of control cells. Such an increase of ubiquitination provoked by glucosamine was previously shown in Fig. 2B
.
|
Next, we wondered whether treating the HepG2 cells with PUGNAc (cell viability: 84.45±15.71%), an inhibitor of O-GlcNAcase used to increase O-GlcNAc levels, also exerted an influence on the MG132-stabilized ubiquitination levels of proteins within HepG2 cells. Accordingly, crude cell extracts from cells grown in the presence and absence of PUGNAc were analyzed by Western blot analysis. Blots were hybridized both with an anti O-GlcNAc and an anti-ubiquitin. As illustrated by Fig. 3B
, both PTMs were shown to be increased by the PUGNAc treatment. The latter results and our observations with forskolin, glucose starvation, and glucosamine reinforce our hypothesis that O-GlcNAc and ubiquitin PTMs could be linked to each other. In a control experiment, the specificity of RL2 was ensured by incubating the antibody with free GlcNAc (Fig. 3A
, right panel; B, middle panel). Although for the latter, the staining was not completely abolished, the remaining signal could be attributed to the necessity of anti-O-GlcNAc antibodies to recognize both the glycosyl part of the epitope and the peptidic backbone, the latter being not chased out by the free sugar. Densitometrical analyses of the two experiments are represented below each set of Fig. 3
. Ubiquitin and O-GlcNAc signals were normalized vs. tubulin or actin signal intensities.
We noticed that a loading control performed with tubulin showed that PUGNAc-treated cells expressed a slightly higher amount of tubulin than control cells (data not shown), whereas no significant difference could be seen with actin. This discrepancy can be reflected to the modification of alpha-tubulin with O-GlcNAc (18)
and consequently to its greater stability in the presence of PUGNAc.
OGT silencing impairs ubiquitination and decreases cell viability
Because modulation of O-GlcNAc contents induces concomitant changes in the level of ubiquitinated proteins, we investigated the putative role of OGT in the ubiquitination process. To attain such insight, OGT knockdown was performed using the RNAi technology. HeLa cells were transfected with OGT RNAi and incubated at 37°C for 96 h. A control experiment was done with oligonucleotides used to knock down the GFP expression. As a negative control, HeLa cells were transfected exclusively with the DreamFect reagent, indicated as the vehicle in Fig. 4
. OGT silencing, as well as the decrease of O-GlcNAc-bearing proteins, was followed by immunoblotting (Fig. 4A
). Beta-tubulin was used to ensure equal loading of the samples. To investigate whether OGT silencing did not perturbate N- and O-glycosylations, we performed a lectin staining with SNA. This lectin binds specifically to sialic acid linked in 2,6 to a galactose residue. The lectin blot demonstrated that OGT silencing did not disturb N- and O-glycosylation pathways (Fig. 4B
). The efficiency of OGT silencing and, accordingly, changes in O-GlcNAc incorporation was also visualized using immunofluorescent microscopy (Fig. 5
A). First, we observed that the knockdown of OGT expression decreased cell viability, as assessed by counting living cells (Fig. 5A
; DAPI staining), which reinforces the hypothesis that OGT is essential for cell life (17)
. Then, the top and middle panels of Fig. 5A
show that OGT silencing impaired OGT expression and reduced O-GlcNAc levels, respectively, as expected; but especially, we observed that OGT knockdown decreased the ubiquitination of proteins (Fig. 5A
, bottom panel). These results were validated by immunoblotting; control cells (Fig. 5B
, 37°C) and heat-shocked cells (Fig. 5B
, 42°C) were analyzed according to their O-GlcNAc and ubiquitin contents. Both at 37°C and at 42°C, cells exhibited less O-GlcNAc and ubiquitin levels when OGT was knocked down. It has to be noted that for a temperature of 42°C, cells transfected with the 1153 oligonucleotide were less viable than mock cells, demonstrating that O-GlcNAc is also an essential PTM for the resistance of the cells to heat (Fig. 5C
) (similar repercussions on cell viability when O-GlcNAc was reduced or blocked were previously obtained by Zachara et al., ref. 17
). The latter experiments reinforce the hypothesis that ubiquitination and O-GlcNAc glycosylation processes may have a common regulation in normal and injured cells.
|
|
E1 O-GlcNAc status
Because of the common features existing between ubiquitination and O-GlcNAc, we hypothesized that the ubiquitination process could be regulated by O-GlcNAc. Accordingly, we sought to determine whether E1, the ubiquitin-activating enzyme that initiates the ubiquitination process, could be modified with O-GlcNAc (Fig. 6
). HepG2 cells were cultured under different conditions, including normal conditions, in the presence of 25 mM glucose (Ctrl); heat-shocked at 42°C (HS); without glucose (w/o Glc); and without glucose but supplemented with 5 mM glucosamine (w/o Glc+GlcNH2). After cell lysis, cell extracts were immunopurified either with the anti-O-GlcNAc antibody (Fig. 6A
) or with the polyclonal anti-E1 antibodies (pE1) (Fig. 6B
). E1 was found to be slightly O-GlcNAc modified in control cells and in glucose-deprived cells (Fig. 6A, B
), and no detectable band was found in heat-shocked cells. In contrast, when cells were treated with glucosamine, O-GlcNAc modification of E1 was enhanced. Interestingly, as illustrated in Fig. 6A
, we showed that E1 coimmunoprecipitated with Hsp70. In addition, the interaction between E1 and Hsp70 is dependent on the O-GlcNAc status of E1, because it was greater when cells were incubated with glucosamine.
|
Subsequently, we examined how glycosylation of E1 was affected by OGT silencing in HeLa cells. As illustrated in Fig. 6C
, OGT silencing decreased E1 glycosylation. The latter results suggest that E1 glycosylation could be one of the processes by which ubiquitination and O-GlcNAc contents are linked.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 20)
. O-GlcNAc can modify numerous nucleocytoplasmic proteins that belong to diverse families, including metabolic enzymes, transcription factors, heat-shock proteins, and architectural proteins (21)
. Although the list of O-GlcNAc-modified proteins still increases, the function of this glycosylation type remains to be elucidated. However, several observations have led to the hypothesis that O-GlcNAc could stabilize proteins by preventing proteasomal degradation (5
, 6)
. The reduced protein degradation by O-GlcNAc may be due to the O-GlcNAc modification of the 19S-regulatory subunit of the proteasome (7
, 8)
, accordingly inhibiting its function or modifying the protein substrates directly, making them more stable. Indeed, experiments have shown that a wide variety of stresses (e.g., UV, hyperthermia, H2O2) inflicted on cells led to an increase of O-GlcNAc content on a large panel of proteins (22)
. Sohn et al. (23)
also showed that hyperthermia was followed by an increase in O-GlcNAc modification. In addition, these researchers demonstrated that cells overexpressing OGT support stress better than control cells. They hypothesized that OGT might behave as a chaperone by modifying unfolded hydrophobic peptide segment exposed after cell injury, which may lead concurrently to protection of degradation. Our group recently reported the GlcNAc-binding property of Hsp70 (15)
. This lectin activity occurs after thermal or nutrient stresses (15
, 16)
. Accordingly, we postulated that O-GlcNAc glycosylation of damaged proteins and Hsp70-GlcNAc binding properties act in synergy. This synergy may protect target proteins against proteasome activity by binding O-GlcNAc-exposed motifs to Hsp70. Despite these exciting observations, the function of O-GlcNAc in protein protection is far from being entirely deciphered.
In the present study, we explored the putative relation between O-GlcNAc and ubiquitin: the O-GlcNAc modification seeming to be a signal acting against protein degradation and the ubiquitination being a well-described proteasome-targeting PTM. Intriguingly, it appeared that in contrast to the reciprocal relation between O-GlcNAc and phosphorylation, no reciprocity between O-GlcNAc and ubiquitination could be demonstrated. However, when a thermal stress was applied to cells, both PTMs were enhanced (Fig. 1)
. Unlike ubiquitination, which is stabilized by MG132, no effect of MG132 could be shown on O-GlcNAc glycosylation. This suggests that O-GlcNAc proteins are not degraded by the proteasome at the same rate as ubiquitinated proteins. This last result could present a discrepancy with an earlier report (17)
. Indeed, these researchers found that cells treated with an inhibitor of the proteasome (ALLN or MG132) exhibited higher O-GlcNAc levels than untreated ones. This difference in results could be explained by the procedures used in the two studies: 1) the researchers treated the cells for 1 h, and then cells were returned to 37°C treatment for various times, whereas in the present work, we treated cells for increasing time periods (for a maximum of 15 min), and studies were immediately performed; 2) the cell type was different, as they used Cos7, and we used HepG2; and 3) the temperature used for heat-treating the cells was 45°C in the cited article and 42°C in the present work.
On the basis of our observations, we hypothesized that, after stress, proteins could be modified with both PTMs and that O-GlcNAc would transiently allow the protection of modified proteins, whereas ubiquitination targets the protein to the proteasome. Nevertheless, immunoprecipitation experiments presented in the present work show that O-GlcNAc and ubiquitin can coexist on the same proteins (Fig. 2B
). This reinforces our suggestion that ubiquitin and O-GlcNAc are capable of acting in conjunction. Conversely, we noticed some discrepancies between levels of O-GlcNAc proteins when cells were treated with glucosamine and additionally treated with MG132. Indeed, the O-GlcNAc content was higher in MG132-treated cells than in nontreated cells (Fig. 2B
). We assume that this difference is caused by a synergistic inhibition of the proteasome by MG132, in addition to its modification with O-GlcNAc (7)
; however, this remains to be elucidated.
Concluding from earlier knowledge and the results presented in the present work, we propose that the ubiquitin/O-GlcNAc ratio could be the switch for a protein to take the path of destruction or the path of repairing. We also wondered whether the ubiquitination process itself could be regulated by O-GlcNAc. Accordingly, we reasoned that when the O-GlcNAc dynamism was perturbed either upward or downward, ubiquitination contents followed the same modifications. As illustrated by our experiments, there might be, indeed, a common regulation between the two PTMs. In our quest to find the link between these two PTMs, we demonstrated that the ubiquitin-activating enzyme E1 was itself O-GlcNAc glycosylated and that this modification was modulated by cell culture conditions and stress (Fig. 6)
. We also showed an interaction between E1 and Hsp70; intriguingly, the interaction correlated with the glycosylation level of E1. Further investigations are necessary to determine whether the interaction between these two partners is regulated by the glycosylation status of E1. However, Cole and Hart (24)
have demonstrated that the synaptosomal enzymatic form of deubiquitination (ubiquitin carboxyl hydrolase-L1; UCH-L1) is O-GlcNAc modified; therefore, one can hypothesize that both the ubiquitination pathway and the deubiquitination process could be regulated by O-GlcNAc. As for E1, the effect of O-GlcNAc on this enzyme is yet unknown, but it can be assumed that it directly regulates its activity. Nevertheless, a putative regulation of protein ubiquitination through E1 O-GlcNAc modification is an interesting field to investigate.
Given recent findings, it could be interesting to indicate the nature of the change we found in the ubiquitination levels occurring after the O-GlcNAc level increase: does it affect more particularly the monoubiquitination, the multiubiquitination, or the polyubiquitination? Considering the Western blot profiles that we obtained during this study, we incline rather toward a polyubiquitination change. Indeed, it has been recently shown that according to the type of polyubiquitin chains found on a substrate, the function managed by polyubiquitination was different (25)
. For example, Lys(48)-linked chains have been shown to target proteins for proteasomal degradation, whereas Lys(63)-linked chains have been implicated in other biological functions, such as endocytosis or signal transduction. It could be envisioned that O-GlcNAc may favor one linkage rather than another; then, if the Lys(48)-linked chains are found after O-GlcNAc increase, one can suppose that the glycosylation may assist the degradation of the ubiquitinated proteins. Conversely, if rather the Lys(63)-linked chains are found, it can be speculated that the ubiquitination induced by O-GlcNAc increase enables the cell to endure stress or serves any other function.
Obviously, at this stage, more extensive investigations are needed to increase our understanding of the role played by O-GlcNAc in the proteasomal processing. But it appears quite clear that it has a protective effect on protein stability, and a comparison can be drawn between the quality control function played by calreticulin and calnexin in the endoplasmic reticulum via the gluco/degluco cycle (26
, 27)
vs. Hsp70 via the O-GlcNAc/de-O-GlcNAc cycle.
This work opens a new door on the study of protein stability and questions about the relations linking the O-GlcNAc dynamism to the ubiquitin pathway.
|
ACKNOWLEDGMENTS |
|---|
Received for publication December 18, 2007. Accepted for publication March 27, 2008.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  C. L. Elsegood, M. Chang, W. Jessup, G. M. Scholz, and J. A. Hamilton Glucose Metabolism Is Required for Oxidized LDL-Induced Macrophage Survival: Role of PI3K and Bcl-2 Family Proteins Arterioscler Thromb Vasc Biol, September 1, 2009; 29(9): 1283 - 1289. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
