Role of Ubiquilin Associated with Protein-disulfide Isomerase in the Endoplasmic Reticulum in Stress-induced Apoptotic Cell Death*

Up-regulation of several stress proteins such as heat-shock proteins and glucose-regulated proteins participate in tolerance against environmental stress. Previously, we found that protein-disulfide isomerase (PDI) is specifically up-regulated in response to hypoxia/brain ischemia in astrocytes. In addition, the overexpression of this gene into neurons protects against apoptotic cell death induced by hypoxia/brain ischemia. To address the detailed function of PDI, we screened for proteins that interact with PDI using the yeast two-hybrid system. We report here that PDI interacts with ubiquilin, which has a ubiquitin-like domain and a ubiquitin-associated domain. Interestingly, ubiquilin is also up-regulated in response to hypoxia in glial cells with a time course similar to that of PDI induction. In hypoxia-treated glial cells, the endogenous ubiquilin and PDI were almost completely co-localized, suggesting that ubiquilin is an endoplasmic reticulum-associated protein. Overexpression of this gene in neuronal cells resulted in significant inhibition of the DNA fragmentation triggered by hypoxia, but not that induced by nitric oxide or staurosporine. Moreover, ubiquilin has the ability to attenuate CHOP induction by hypoxia. These observations suggested that ubiquilin together with PDI have critical functions as regulatory proteins for CHOP-mediated cell death, and therefore up-regulation of these proteins may result in acquisition of tolerance against ischemic stress in glial cells.

Cellular adaptations to environmental change constitute critical protective pathways. One such essential biosynthetic response is the induction of diverse stress-associated proteins such as heat-shock proteins (HSPs) 1 or glucose-regulated proteins (GRPs) (1,2). These biosynthetic responses have been hypothesized to contribute to maintenance of cellular homeostasis during adaptation to altered environmental conditions. In particular, neurons are fragile and very sensitive to stress, whereas glial cells show tolerance to stress in the brain (3)(4)(5).
For example, brain ischemic stress results in neuronal apoptosis, but has no effect on glial cells. Under these conditions, several proteins such as HSPs and GRPs are induced in glial cells. Previously, we reported that protein-disulfide isomerase (PDI, EC 5.3.4.1) is up-regulated in response to hypoxia/brain ischemia in glial cells (6). PDI is a multifunctional protein mainly located in the endoplasmic reticulum (ER) (7,8). During protein folding in the ER, PDI catalyzes thiol/disulfide exchange, including disulfide bond formation and rearrangement reactions (9). PDI has two domains with homology to the small, redox-active protein, thioredoxin (10). The thiol/disulfide centers of the two thioredoxin-like domains function as two independent active sites (11). Increased PDI expression in neurons resulted in both attenuation of the loss of cell viability in vitro and reduction of the number of DNA-fragmented cells in the rat hippocampal CA1 subregion in vivo. Therefore, it has been suggested that up-regulated PDI has an essential role in glial cell survival under severe conditions.
The ER is an organelle in which secretory proteins are folded and processed before export from the cell (12). The ER undergoes stress responses when unfolded immature proteins accumulate (13,14). Although severe ER stress can result in apoptosis through ER-specific caspase-12 (15), the ER withstands relatively mild stress through the regulation of expression of stress proteins such as GRPs, oxygen-regulated proteins, and PDI. These stress proteins behave as molecular chaperones, i.e. functional proteins that assist in the maturation and transport of unfolded secretory proteins (16 -21), suggesting that the neuroprotective effect of PDI may be partially based on its chaperone activity (22)(23)(24).
In the present study, we attempted to isolate PDI-interacting proteins as protective factors against hypoxia by yeast twohybrid screening. Our results indicated that ubiquilin interacts with PDI both in vitro and in vivo. Moreover, we showed that ubiquilin is up-regulated in response to hypoxia and suppresses hypoxia-induced neuronal cell death, suggesting that the interaction of ubiquilin with PDI may contribute to adaptive responses to ischemic stress, thereby ultimately contributing to enhanced survival of neurons.

EXPERIMENTAL PROCEDURES
Materials-Rabbit polyclonal anti-ubiquilin antibody was produced by immunization with synthetic peptide containing the N-terminal 15 amino acids of ubiquilin, followed by affinity chromatography using immobilized protein A. Mouse monoclonal anti-GFP antibody was purchased from MBL (Japan). Anti-rabbit IgG and mouse IgG conjugated with Alexa 488 or 594, and fluorescein di-␤-D-galactopyranoside were purchased from Molecular Probes. Anti-FLAG M5 monoclonal antibody and anti-HA polyclonal antibody were purchased from Sigma and Santa Cruz, respectively. pEGFP and the yeast two-hybrid system were from CLONTECH. All other reagents were obtained from Sigma.
Plasmids-Full-length ubiquilin and PDI cDNAs and several truncated mutants were isolated from human neuroblastoma RNA by RT-PCR. For yeast two-hybrid assay, full-length PDI and ubiquilin cDNA and several truncated fragments were subcloned into pAS2-1 and pACT2 (CLONTECH), respectively. The PDI expression vector-fused HA tag in its C-terminal (PDI-HA) was engineered. The cDNAs encoding ubiquilin and calreticulin tagged with FLAG and the HA epitope at the N-terminal (FLAG-ubiquilin and HA-calreticulin) were subcloned into mammalian expression vector pCR3.1, respectively.
Yeast Two-hybrid Assay-The pAS2-1/PDI plasmid, which contained the full sequence of the PDI coding region, was used for the yeast two-hybrid library screening to isolate PDI-interacting proteins. The two-hybrid screen was performed with a pretransformed cDNA library derived from the human fetal kidney mRNA (CLONTECH) according to the manufacturer's protocol. GAL4-BD/p53 and GAL4-AD/SV40 T-antigen were used as positive controls in this assay.
In Vitro Binding Assay-GST and GST-ubiquilin were purified from a 1-liter culture of Escherichia coli BL21(DE3) transformed with pGEX-6P-1 and pGEX-GST-ubiquilin, respectively, as described previously (25). Lysates from SK-N-MC cells were first incubated with 3 g of GST or GST-ubiquilin for 4 h in buffer containing 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 0.25% gelatin, 1 mM EDTA, 0.1% Nonidet P-40, and 0.02% NaN 3 and then applied to a glutathione-Sepharose 4B column (Amersham Biosciences). After extensive washing of the column, the proteins recovered from the resin were separated in a 10% polyacrylamide gel containing SDS, blotted onto a nitrocellulose filter, and reacted with an anti-PDI antibody.
Exposure to Hypoxic Environment-When the cells became subconfluent, they were cultured in a mixture of 5% CO 2 and the balance of N 2 in a humidified incubator (ANX-1, HIRASAWA, Tokyo, Japan) at 37°C within a sealed, anaerobic, gloved cabinet containing a catalyst to scavenge free oxygen as described previously (6,26,27). Oxygen tension in the chamber and medium were measured using a oxygen analyzer (Teledyne) and a blood gas analyzer (ABL-2, Radiometer, Sweden), respectively. The oxygen levels in the cabinet were measured with a monitor sensitive to oxygen concentrations of Ͻ10 ppm throughout the incubation period. Oxygen tension in the culture medium at 15 min, 30 min, 1 h, 6 h, and 48 h after transfer into the hypoxic chamber fell to 4.3, 3.0, 2.2, 2.0, and 2.0 Ϯ 0.2%, respectively. The medium was then changed, and the cells were cultured under normal conditions for the indicated periods (reoxygenation).
Immunoprecipitation and Western Blotting Analysis-SK-N-MC cells were washed twice with ice-cold phosphate-buffered saline, lysed in lysis buffer (50 mM Tris-HCl, (pH 7.5), 150 mM NaCl, 2 mM EGTA, 0.1% Triton X-100) containing protease inhibitors (1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, and 10 g/ml leupeptin), and centrifuged at 15,000 ϫ g for 30 min at 4°C. The supernatant was incubated with protein G-Sepharose for 1 h at 4°C and then centrifuged for 10 min. The resultant supernatant was incubated for 16 h with the indicated antibodies, which had been precoupled with protein G-Sepharose. The immunoprecipitates were washed five times with lysis buffer, boiled for 5 min with SDS sample buffer, and subjected to SDS-PAGE. After electrophoresis, proteins were blotted onto nitrocellulose membranes. The membranes were blocked with 10 mM Tris-HCl (pH 7.5), 100 mM NaCl, and 0.1% Tween 20 containing 5% nonfat milk for 1 h and then treated with primary antibody (1:1000 dilution) followed by horseradish peroxidase-conjugated secondary antibody. The immunoreactive bands were revealed by chemiluminescence (ECL Western detection kit; Amersham Biosciences). To examine the kinetics of ubiquilin induction by hypoxia in the CCF-STTG1 cells, the cells were cultured in a mixture of 5% CO 2 and the balance of the N 2 in a humidified anaerobic incubator at 37°C for the indicated periods as described (26 -28).
Northern Blotting Analysis-Multiple human tissue Northern blots (CLONTECH) were hybridized for 16 h at 42°C using the ubiquilin cDNA probe labeled with [ 32 P]dCTP. Filters were washed twice in 2ϫ SSC, 0.05% SDS for 15 min at room temperature and in 1ϫ SSC, 0.1% SDS for 15 min at 68°C. The levels of ubiquilin mRNA were analyzed using a FUJI BAS 2000.
Immunohistochemistry-Human CCF-STTG1 astrocytoma cells were transiently transfected with FLAG-ubiquilin, GFP-fused ubiquilin (GFP-ubiquilin), and PDI-HA using Effectene transfection reagent (Qiagen). Forty-eight h after transfection, cells were fixed in 4% paraformaldehyde/phosphate-buffered saline for 10 min, permeabilized for 10 min in phosphate-buffered saline containing 0.5% Triton X-100, and then incubated for 1 h in 10% normal blocking serum. Cells were treated with anti-FLAG M5 monoclonal antibody, anti-GFP monoclonal antibody, and/or anti-HA polyclonal (1:100 dilution) for 1 h. After the wash, the cells were incubated with anti-mouse antibody conjugated with Alexa 488 and/or anti-rabbit antibody conjugated with Alexa 594 (1:200 dilution) for 1 h. All images were taken on a Zeiss LSM 510 laser-scanning confocal microscope.
DNA Fragmentation-For the DNA fragmentation assay, the cells were lysed in lysis buffer (10 mM Tris-HCl, pH 7.4, 5 mM EDTA, 0.5% Triton X-100) and incubated for 20 min at 4°C. The samples were centrifuged at 27,000 ϫ g for 15 min at 4°C. The supernatants were incubated with 40 g/ml proteinase K for 30 min at 37°C and extracted with equal volumes of phenol, phenol/chloroform (1:1 v/v), and chloroform. The DNA was precipitated from the supernatants with a 1/10 volume of 3 M sodium acetate (pH 5.2) and 2 volumes of ethanol and treated with 40 g/ml RNase A for 1 h at 37°C. The recovered DNA was then analyzed by electrophoresis on a 1.5% agarose gel and visualized with 0.5 g/ml ethidium bromide.
Assessment of Cell Viability-Cell viability was estimated by two methods (6). Initially, viability was measured by counting the stained viable cells for ␤-galactosidase activity derived from transfection of that gene. Briefly, 0.3 g of lacZ plus 1.25 g of each gene were co-transfected into SK-N-MC cells using the transfection reagent SuperFect (Qiagen) and incubated for 24 h. After hypoxic challenge, cells were fixed with 1% glutaraldehyde for 10 min, rinsed three times with phosphate-buffered saline, and stained in 5-bromo-4-chloro-3-indolyl-␤-D-galactopyranoside (X-gal) buffer (0.5 mg/ml X-gal, 4 mM K 4 [Fe(CN) 6 ], 4 mM K 3 [Fe(CN) 6 ], 1 mM MgCl 2 , 10 mM KCl, 0.1% Triton X-100 in 0.1 M sodium phosphate buffer, pH 7.5) at 37°C for 2 h. The number of stained blue cells in cultures challenged by hypoxia versus that in normoxic cultures was counted by two investigators blind to the experimental treatments, and the results are expressed as the percentage of hypoxia/normoxia.
As another strategy, the viability was estimated by the lactate dehydrogenase (LDH) leakage method using a cytotoxicity detection kit (Roche Molecular Biochemicals) according to the manufacturer's protocol. LDH activity was measured as the optimal density at 492 nm, and LDH leakage (%) was defined as the ratio of LDH activity in the culture medium to the total activity (% ϭ (extracellular activity)/(extracellular activity ϩ remaining cellular activity)).

Ubiquilin Interacts with PDI in Yeast-
To identify the proteins interacting with PDI, we performed a yeast two-hybrid library screening using full-length PDI as bait. Approximately 1 ϫ 10 7 independent clones in the human kidney cDNA expression library were screened on selective medium, and then the ␤-galactosidase assay was performed to eliminate false positives. Several independent clones were isolated as determined by activation of reporter genes. Sequence analysis demonstrated that a clone was identical to ubiquilin and the remaining five clones were unknown proteins. The interaction between ubiquilin and PDI was analyzed precisely by the yeast two-hybrid assay. Ubiquilin is composed of at least four domains (a ubiquitin-like (UB) region, two Asn-Pro repeat regions, and a ubiquitin-associated region) (Fig. 1A). To determine the PDI-binding region of ubiquilin, various deletion constructs fused to GAL4 were used for the two-hybrid assay with full-length PDI as bait. The wild-type ubiquilin, 1-535, 1-420, and 318 -535 mutants bound to PDI, whereas the N-terminal-(1-317) half-fragment spanning amino acids of ubiquilin did not (Fig. 1A). Next, we evaluated the ubiquilinbinding region of PDI using the yeast two-hybrid system. PDI is composed of at least three domains (two thioredoxin-like regions (a and aЈ), two unknown regions (b and bЈ), and a putative Ca 2ϩ -binding region (C)). As shown in Fig. 1B, PDI interacted specifically with ubiquilin through its C-terminal-(441-509) containing the putative Ca 2ϩ -binding region.
Distribution of Ubiquilin mRNA in Human Tissues-To examine ubiquilin mRNA expression in different tissues, we performed Northern blotting analysis using a radioactive probe corresponding to full-length ubiquilin cDNA. This probe detected a major transcript of ϳ4.5-kb in all human tissues. Particularly, high levels of ubiquilin mRNA expression were seen in the adult heart, brain, and placenta, and in fetal brain, liver, lung, and kidney (Fig. 2).
Interaction between Ubiquilin and PDI in Vitro and in Vivo-To confirm whether PDI interacts with ubiquilin, we performed both in vitro pull-down assay and co-immunoprecipitations using lysates in mammalian cells transfected with PDI and ubiquilin. Cell lysates were incubated with GST or GST fused with ubiquilin, and PDI was detected in proteins eluted from the beads binding to GST-ubiquilin but not from those binding to GST alone ( Fig. 2A). To determine whether the interaction between PDI and ubiquilin occurs in vivo, SK-N-MC cells were transiently transfected with FLAG-ubiquilin together with PDI-HA. Western blotting analysis showed that PDI-HA and FLAG-ubiquilin were expressed in each transfectant (Fig. 2B). Whole cell lysates immunoprecipitated with control IgG, anti-FLAG mAb, or anti-HA pAb were then analyzed by Western blotting using HA or FLAG antibody. As show in Fig. 2B, PDI-HA protein was detected in immunoprecipitates with anti-FLAG from cells co-transfected with PDI-HA and FLAG-ubiquilin. FLAG-ubiquilin protein was also detected in immunoprecipitates with anti-HA.
Co-localization of PDI with Ubiquilin in the Cells-PDI has been reported to be localized specifically in the lumen of the ER. Based on the above observations, we speculated that ubiquilin was at least partly localized in the ER. To determine the cellular localizations of PDI and ubiquilin, expression vectors for GFP-or FLAG-ubiquilin and PDI-HA were transfected into human CCF-STTG1 cells. Two days after transfection, the cells were treated with anti-GFP or FLAG and anti-HA antibody, and the proteins were stained by the Alexa 488-and Alexa 594-conjugated second antibodies. On indirect immunofluorescence labeling of GFP/FLAG-ubiquilin, both GFPubiquilin and FLAG-ubiquilin (green) were mainly observed in the cytoplasm (Fig. 3C). In addition, localization of ubiquilin (green) was completely coincidental with those of PDI (red) and calreticulin (red) as an ER marker in the cytoplasm as shown by the yellow color (Fig. 3C). These results, i.e. the perinuclear staining and peripheral lace-like network, suggested that overexpressed ubiquilin may be associated with the ER.
Expression of Ubiquilin in Human Astrocytoma CCF-STTG1 Cells-We demonstrated previously that PDI is up-regulated in response to hypoxia in astrocytes (6). Thus, we investigated whether ubiquilin is also induced by hypoxic stress in astrocytes. Using anti-ubiquilin IgG raised against the N-terminal synthetic peptide, immunoblotting studies were performed on lysates of cultured human CCF-STTG1 astrocytes exposed to hypoxia and hypoxia followed by reoxygenation. Ubiquilin was not detected or was detected only at a very low level in astrocytes in the quiescent state, was detected 12 h after hypoxia, and the level continued to increase at least 48 h. There were no significant changes in cell viability during this period (data not shown). Following replacement of hypoxic cultures into the normoxic environment, the intensity of the ubiquilin band decreased gradually (Fig. 4A). We next investigated whether this phenomenon was accompanied with an increase in mRNA induction (Fig. 4B). Hypoxia-induced ubiquilin mRNA expression was detected at 12 h after hypoxia and was sustained for 12 h after reoxygenation, decreasing gradually thereafter. As upregulation of the endogenous ubiquilin in response to hypoxia was detected by immunoblotting, we attempted to elucidate the subcellular localization of ubiquilin by indirect fluorescent immunocytochemistry. Under normoxic conditions, immunostaining of astrocytes with anti-ubiquilin and anti-PDI antibodies showed little and slight staining, respectively (Fig. 4C), consistent with the protein levels observed by Western blotting analysis (Fig. 4B). Immunofluorescent staining of up-regulated endogenous ubiquilin revealed bright perinuclear staining as well as peripheral stained structures (Fig. 4C). The pattern was reminiscent of the ER staining. In addition, we assessed the co-localization of ubiquilin and PDI, a marker protein of the ER. The pattern of immunoreactivity with anti-PDI antibody was nearly identical to that with anti-ubiquilin antibody as shown by the yellow signal (Fig. 4C).
Ubiquilin Suppresses Hypoxia-induced DNA Fragmentation in Neurons-Previously, we reported that up-regulated PDI has a protective effect against the loss of cell viability induced by hypoxia (6). Therefore, we speculated that ubiquilin may also have a protective role against hypoxic stress. Initially, we established transfectants that stably expressed ubiquilin in SH-SY5Y neuroblastoma cells. As shown in Fig. 5A, hypoxia resulted in DNA ladder formation in mock-transfected SH-SY5Y cells. However, there was little fragmentation by hypoxia in ubiquilin-transfected cells. Interestingly, overexpression of ubiquilin showed a protective effect against apoptosis induced by only hypoxia, but not other apoptosis inducers such as NOC18 (a nitric oxide donor) and staurosporine (Fig. 5B). In addition, we examined whether co-expression of PDI and ubiquilin in neuronal cells has an additive effect on hypoxiainduced loss of cell viability. As shown in Fig. 5C, co-expression of both ubiquilin and PDI rendered the cells more resistant against hypoxia than cells expressing either ubiquilin or PDI alone.
Both Ubiquilin and PDI Negatively Regulate CHOP Induction by Hypoxia-It is well known that CHOP (a C/EBP homolog) is induced by several ER stresses and plays a critical role in cell death. We examined the effects of ubiquilin and PDI on ER stress-induced CHOP induction in neuronal cells. In mocktransfected SK-N-MC cells, kinetic analysis revealed that CHOP mRNA was detectable after 24 h of hypoxia treatment, and the high level of this induction was sustained for up to 36 h. On treatment with hypoxic stress, the loss of cell viability was observed after 24 h coincident with the time course of CHOP induction. After 36 h of stress treatment, the cells were mostly nonviable (Fig. 6). In contrast, the CHOP induction by hypoxia was selectively delayed in ubiquilin-or PDI-overexpressed neuronal cells. In particular, the induction was not detected for up to 36 or 24 h in ubiquilin-or PDI-transfected cells, respectively. At these time points, ubiquilin and PDI significantly attenuated the loss of cell viability. These results suggested that ubiquilin inhibits ER stress-induced neuronal cell death via attenuation of CHOP induction.

Isolation and Identification of Ubiquilin, Which
Interacts with PDI-The aim of this study was to isolate and identify proteins capable of interacting with PDI because we have reported previously that PDI up-regulated by hypoxia/brain ischemia may be involved in the acquisition of tolerance against these stresses. In the present study, we carried out yeast twohybrid screening to isolate proteins and consequently characterized an interaction between PDI and ubiquilin. This interaction was possibly mediated via the asparagine-proline (NP) repeat region in the C terminus (residues 365-476) of ubiquilin and the C terminus (residues 441-509) of PDI containing a putative Ca 2ϩ -binding region (Fig. 1). Ubiquilin has sequence similarities to a region of ubiquitin in its N terminus and to the yeast protein Dsk2, Xenopus protein XDRP1, mouse protein PLIC-1/DA41, PLIC-2, and mouse protein UBIN (29 -35). All of these proteins possess a well conserved ubiquitin-like domain at their N termini and ubiquitin-associated domain at their C termini, suggesting that both are involved in targeting and degradation of proteins by the ubiquitin-proteasome pathway.

FIG. 3. Ubiquilin interacts with PDI in vitro and in vivo.
A, after incubation of human SK-N-MC cell lysates with glutathione-agarose beads bound with either GST or GST-ubiquilin, the binding ability was analyzed by Western blotting analysis using anti-PDI mAb. B, co-immunoprecipitation of PDI and ubiquilin protein was performed in SK-N-MC cells transfected with PDI-HA (pCR3.1-PDI-HA) alone or together with FLAG-ubiquilin (pCR3.1-FLAG-ubiquilin). Forty-eight h after transfection, total cell lysates were analyzed by Western blotting (WB) to check expression of PDI (third panel) and ubiquilin (fourth panel) proteins. Equal amounts of cell lysates were immunoprecipitated with normal mouse IgG or anti-FLAG mAb and normal rabbit IgG or anti-HA pAb, respectively. The immune complex bound to protein G-Sepharose beads was then washed with lysis buffer and washing buffer, resolved by SDS-PAGE, and analyzed by Western blotting using anti-HA pAb (top panel) and anti-FLAG mAb (second panel), respectively. C, cellular localization of overexpressed ubiquilin, PDI, and calreticulin. Human CCF-STTG1 cells co-overexpressing GFP-ubiquilin and PDI-HA, FLAG-ubiquilin, and PDI-HA or calreticulin-HA were fixed and subjected to indirect immunofluorescence staining with anti-GFP mAb and anti-HA pAb, anti-FLAG mAb and anti-HA pAb, respectively. The green signal (ubiquilin) was obtained with anti-mouse IgG Alexa 488-conjugated secondary Ab, whereas the red signal (PDI or calreticulin) was obtained with anti-rabbit IgG Alexa 594-conjugated secondary Ab. Superimposing two colors (merged) resulted in a yellow signal, indicating co-localization of the two proteins.
In addition, these proteins were found to contain numerous regularly spaced asparagine-proline (NP) repeats, although the precise biological relevance of this motif remains to be determined. Endogenous ubiquilin has been shown to be localized in both the nucleus and cytoplasm of HeLa cells; however, the ubiquilin staining pattern is variable in different cell types (29). Interestingly, ubiquilin had a fine punctate appearance with hints of association to a network-like pattern, possibly the ER (29). Here, endogenous ubiquilin up-regulated by hypoxic stress in glial cells revealed both staining of perinuclear and peripheral structures, a pattern reminiscent of the ER (Fig.  4C). In agreement with this observation, double staining with anti-ubiquilin and PDI antibodies showed that the endogenous ubiquilin antigen up-regulated by hypoxia was co-localized almost perfectly with the PDI antigen in glial cells (Fig. 4C). Ubiquilin has been isolated as a protein that interacts with presenilin (29). In fact, PDI and presenilin are localized in the ER lumen and membrane, respectively. Based on these observations, we speculated that overexpressed ubiquilin was local-ized to the ER, or alternatively endogenous ubiquilin up-regulated by hypoxia may be primarily located in or associated with the ER. Ubiquilin/DA41 has a putative transmembrane domain ( 147 FGLGGLGGLAGLSSLGL 163 ) (33), and UBIN, a novel ubiquilin-like protein, was suggested to be localized to the cytosolic side of the ER membrane, perhaps in a membranebound form (35). In general, the ER lumen proteins such as GRP and PDI have an ER-retrieval signal motif, KDEL, or KDEL-like sequence at the C terminus (36), although ubiquilin lacks such a motif. However, Herp also lacks this sequence at the C terminus but is localized in the ER membrane (37). The observation that ubiquilin was co-localized with presenilin and PDI suggested that ubiquilin may be localized in the ER via an as yet unknown mechanism.
Up-regulation of Ubiquilin in Response to Hypoxic Stress-Previously, we demonstrated that exposure to hypoxia/brain ischemia results in significant enhancement of PDI mRNA and protein levels in glial cells but not neurons. In the present study, we attempted to detect the endogenous ubiquilin in neuronal or glial cells. Unfortunately, the level of endogenous ubiquilin was very low in lysates from several glial cell types in the quiescent state (Fig. 4A). We speculated that ubiquilin may also be up-regulated in response to hypoxic stress as seen in PDI. As shown in Fig. 4, A and B, both the protein and mRNA levels were increased in astrocytes subjected to hypoxia similarly to PDI. It has recently been reported that Herp, a novel membrane-associated ER protein, is a target gene for unfolded protein response-stimulated transcription such as GRP, PDI, and calnexin (37). Hypoxia causes the accumulation of immature protein, which may eventually initiate cell death caused because of ER dysfunction (9). Therefore, ubiquilin could be a target for unfolded protein response-induced gene expression.
Role of Up-regulated Ubiquilin in Response to ER Stress-To address whether ubiquilin plays a critical role in tolerance against ischemic stress, we assessed its protective effect against hypoxia-induced apoptosis. Stable expression of ubiquilin in human neuroblastoma SH-SY5Y cells almost completely attenuated DNA ladder formation evoked by hypoxia (Fig. 5A). Interestingly, the anti-apoptotic effect of ubiquilin was observed only when the cells were challenged by hypoxia, but not following exposure to NO or staurosporine (Fig. 5B). These observations strongly suggested that ubiquilin selectively blocks ER stress-induced apoptosis in neuronal cells. In addition, co-expression of both ubiquilin and PDI exhibited more resistance against hypoxia than cells expressing either ubiquilin or PDI alone (Fig. 5C). We subsequently attempted to elucidate the mechanism by which ubiquilin/PDI inhibited hypoxia or ER stress-induced cell death. Initially, we investigated the activation of caspase-12 by Western blotting analysis, as caspase-12 has been reported to be localized in the ER and to be involved in apoptosis induced by ER stress (38 -40). However, no significant decrease in the caspase-12 level by hypoxia was observed. These results suggested that other pathways besides the caspases cascade are involved in the attenuation of apoptosis by ubiquilin. Recently, it has been demonstrated that apoptosis induced by ER stress is mediated by CHOP expression in response to unfolded protein response (41)(42)(43). Therefore, we examined whether CHOP induction occurs by treatment with hypoxia and the effects of ubiquilin or PDI. In mock-transfected cells, CHOP induction was observed 24 h after hypoxia challenge and the level was sustained for up to 36 h (Fig. 6). Consistent with the time course of CHOP induction, cell viability was reduced in a time-dependent manner. There were few viable cells at 36 h after challenge. Overexpression of either ubiquilin or PDI significantly delayed the CHOP expression and consequently attenuated cell death. Our results FIG. 4. Up-regulation of ubiquilin in response to hypoxia in glial cells. A, human astrocytoma CCF-STTG1 cells were washed and lysed. The supernatant (about 5 g of protein) was subjected to SDS-PAGE (8%) and evaluated by Western blotting using anti-ubiquilin antibody. B, RT-PCR analysis to determine ubiquilin mRNA expression in stress-induced cells. Human CCF-STTG1 astrocytes were exposed to hypoxic conditions for the indicated times. Total RNA was extracted and subjected to RT-PCR using specific primers for ubiquilin. Twenty cycles of PCR produced a linear relationship between the amount of input RNA and that of the resulting PCR product. C, immunofluorescence analysis. Human CCF-STTG1 astrocytes plated on coverslips were exposed to normoxia (upper panel) or hypoxia for 24 h (lower panel), permeabilized, and incubated with anti-ubiquilin antibody, or anti-PDI antibody. The green signal (ubiquilin) was obtained with antirabbit IgG Alexa 488-conjugated secondary antibody, whereas the red signal (PDI) was obtained with anti-mouse IgG Alexa 594-conjugated secondary antibody. Double labeling (merged) resulted in a strong yellow signal. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.
together with evidence reported previously suggested that the protective effects of ubiquilin or PDI against hypoxia-induced cell death (apoptosis) may be partly mediated by attenuation or delay of CHOP induction. Alternatively, PDI is known to be involved in the protein quality control machinery in the ER that recognizes terminally misfolded secretory proteins and targets them to the export channel in the ER membrane (10).
In the present study, we demonstrated the interaction of ubiquilin with PDI in vivo and in vitro. Co-expression of ubiquilin may be influenced by the chaperone activity of PDI. The precise mechanism of this function remains to be determined.
Ubiquilin contains a ubiquitin-like domain at its N terminus and a ubiquitin-associated domain at its C terminus (29). A number of proteins with structural similarity to ubiquitin have FIG. 5. Overexpression of ubiquilin suppresses hypoxia-induced DNA fragmentation. A, human SH-SY5Y cells were stably transfected with empty plasmid (mock) or ubiquilin cDNA. Stable clones were exposed to hypoxia for 30 h and then harvested. DNA was extracted and subjected to agarose gel (1.5%) electrophoresis. B, stable clones transfected with ubiquilin were treated with hypoxia for 30 h, NOC18 (250 M), or staurosporine (STS) (1 M) for 24 h. At the indicated times, DNA was extracted and subjected to agarose gel (1.5%) electrophoresis. The gels were then stained with 0.5 g/ml ethidium bromide for 15 min, and the fragmented DNA was visualized under ultraviolet light and then photographed. C, effects of ubiquilin and PDI on the loss of viability in response to hypoxia. Cells were transfected with 1.25 g of PDI and/or ubiquilin with empty vector (pCR3.1) using SuperFect reagent (Qiagen). Twenty-four h after transfection, each transfectant was transferred to a low oxygen chamber and incubated for another 36 h. After incubation, the loss of cell viability was estimated by the LDH leakage method, as described under "Experimental Procedures." Values represent the mean Ϯ S.E. of triplicate cultures run in parallel. Star represents significantly different values from those of mock-transfected cells (*, p Ͻ 0.05; **, p Ͻ 0.01, Student's t test).
FIG. 6. Inhibitory effects of ubiquilin and PDI on hypoxia-induced CHOP expression. Cells were transfected with 1.25 g of the vector (pCR3.1), PDI, or ubiquilin plus 0.3 g of pACT-␤-gal using SuperFect reagent (Qiagen). Twenty-four h after transfection, each transfectant was transferred to a low-oxygen chamber and further incubated for the indicated periods. Total cellular RNA was prepared and analyzed by RT-PCR as described under "Experimental Procedures." At the same time, the viability was assessed by counting the stained viable cells for ␤-galactosidase activity derived from transfection of that gene as described under "Experimental Procedures." After fixation, the cells were incubated with X-gal buffer, and the stained cells were then counted. Viability is represented as the ratio of stained cells under normoxia and hypoxic challenge. Values represent the means of triplicate cultures run in parallel and are expressed as the percentages of those under normoxic conditions in control cultures. GAPDH, glyceraldehyde-3-phosphate dehydrogenase. been identified; ubiquitin-like proteins (UBLs) are classified into two subclasses, designated as type-I and type-II UBL families (44). Type-I UBLs such as SUMO-1 and NEDD8 have conserved Gly-Gly sequences at the C terminus of the ubiquitin domain. Type-II UBLs such as Rad23p and Parkin do not have the conserved Gly-Gly sequences. Ubiquilin is considered to belong to the type-II UBL family. Type-I UBLs have been implicated in post-transcriptional protein modification and type-II UBLs remain poorly characterized. However, various roles of type-II UBLs have been recently reported. For example, the ubiquitin-like domain of Rap23p interacts with the 19 S regulatory subunit of the 26 S proteasome (45,46). In particular, 1) XDRP1 (Xenopus ubiquilin homolog) acts post-transcriptionally like a molecular chaperone and inhibits degradation of cyclin A (31); 2) Chap1 (ubiquilin 2) binds the ATPase domain of Stch, an HSP70-like protein (47); 3) PLIC/ubiquilin has been speculated to provide a link between the ubiquitination machinery and the proteasome (34); 4) ubiquilin is associated with mTOR protein kinase, which is partly involved to control protein degradation, however, the detailed function has not been determined yet (48,49). In addition, it has been proposed that the ER-associated protein degradation system eliminates misfolded ER proteins via degradation through the ubiquitin-proteasome pathway in the cytosol (50,51). These lines of evidence suggested that stress-induced ubiquilin together with PDI may contribute to protein folding and degradation via molecular chaperone activity and the proteasome. In fact, hypoxia/brain ischemia causes accumulation of immature and denatured protein, and consequently the dysfunction of the ER (9). Thus, up-regulation of ubiquilin, PDI, and other unfolded protein response-inducible proteins such as GRPs by ER stress may result in acquisition of tolerance against stresses by controlling "protein quality." In summary, we have demonstrated here that ubiquilin is a PDI-interacting protein in the ER. Furthermore, the overexpression of ubiquilin appears to confer resistance to hypoxia. Although the detailed mechanism by which ubiquilin is involved in cell survival is not yet clear, it is obvious that both ubiquilin and PDI attenuate CHOP expression induced by hypoxia in neurons. Therefore, the protective effect against apoptosis appears to be at least partly mediated by the inhibition of CHOP induction. Ubiquilin is highly expressed in neurons of the human brain and is associated with neurofibrillary tangles and Lewy bodies seen in the brains of patients with Alzheimer's disease and Parkinson's disease, respectively (29). It is of interest to address the relationship between the role of ubiquilin/ PDI and neurodegenerative diseases such as Alzheimer's and Parkinson's. At present, the detailed mechanism of action of ubiquilin is not clear. However, ubiquilin together with PDI may have a role in promoting recovery from the loss of function of "protein quality control" in neuronal diseases.