HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 275, Issue 52, 40757-40764, December 29, 2000

This Article Abstract Full Text (PDF) All Versions of this Article: 275/52/40757    most recent M001073200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Tokunaga, F. Articles by Arvan, P. Search for Related Content PubMed PubMed Citation Articles by Tokunaga, F. Articles by Arvan, P. Social Bookmarking                      What's this?

Endoplasmic Reticulum (ER)-associated Degradation of Misfolded N-Linked Glycoproteins Is Suppressed upon Inhibition of ER Mannosidase I^*

Fuminori Tokunaga, Charles Brostrom^§, Takehiko Koide, and Peter Arvan^¶

From the  Department of Life Science, Himeji Institute of Technology, Harima Science Garden City, Hyogo 678-1277, Japan, the ^§ Department of Pharmacology, University of Medicine and Dentistry of New Jersey, Piscataway, New Jersey 08854, and the ^¶ Division of Endocrinology, Department of Developmental and Molecular Biology, Albert Einstein College of Medicine, Bronx, New York 10461

Received for publication, February 9, 2000, and in revised form, September 6, 2000

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

To examine the role of early carbohydrate recognition/trimming reactions in targeting endoplasmic reticulum (ER)-retained, misfolded glycoproteins for ER-associated degradation (ERAD), we have stably expressed the cog thyroglobulin (Tg) mutant cDNA in Chinese hamster ovary cells. We found that inhibitors of ER mannosidase I (but not other glycosidases) acutely suppressed Cog Tg degradation and also perturbed the ERAD process for Tg reduced with dithiothreitol as well as for -carboxylation-deficient protein C expressed in warfarin-treated baby hamster kidney cells. Kifunensine inhibition of ER mannosidase I also suppressed ERAD in castanospermine-treated cells; thus, suppression of ERAD does not require lectin-like binding of ER chaperones calnexin and calreticulin to monoglucosylated oligosaccharides. Notably, the undegraded protein fraction remained completely microsome-associated. In pulse-chase studies, kifunensine-sensitive degradation was still inhibitable even 1 h after Tg synthesis. Intriguingly, chronic treatment with kifunensine caused a 3-fold accumulation of Cog Tg in Chinese hamster ovary cells and did not lead to significant induction of the ER unfolded protein response. We hypothesize that, in a manner not requiring lectin-like activity of calnexin/calreticulin, the recognition or processing of a specific branched N-linked mannose structure enhances the efficiency of glycoprotein retrotranslocation from the ER lumen.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Whether genetic or acquired, endoplasmic reticulum (ER)¹ storage diseases begin with the synthesis of misfolded versions of exportable proteins within the ER compartment (1). Generally, these proteins are retained in the ER (2) pending targeting for ER-associated degradation (ERAD) (3), at which point they are retrotranslocated to the cytosol (4, 5) and degraded via the ubiquitin/proteasome pathway (6-8). For many secretory proteins, the misfolded versions may enter the ER lumen completely before a final cellular disposition is made (9-11). Thus, it is not surprising that molecular chaperones acting in the ER lumen have been implicated as participants in the retention and ERAD processes. Of the ER chaperones proposed to function in targeting for retrotranslocation and ERAD, calnexin has been the most often implicated (12-17). Other studies, however, have suggested that ERAD is independent of calnexin or that association with calnexin actually prevents ERAD (18-24).

The binding of calnexin (and calreticulin) to most exportable substrates occurs via a lectin interaction that requires exposure of monoglucosylated oligosaccharide side chains (25, 26). In normal mammalian cells, this association is prevented by pretreatment with castanospermine (CAS), an inhibitor of glucosidases I and II (21, 27, 28). By contrast, the calnexin interaction is enhanced in cells pretreated with 1-deoxymannojirimycin (DMM), the general -mannosidase inhibitor that prevents the formation of Man_5-7 glycans, which are less efficient substrates for monoglucosylation (29, 30) and may exhibit diminished calnexin-binding affinity (31). Using these inhibitors as well as other approaches, several groups have begun to explore the role of early carbohydrate trimming reactions in the targeting of ER-retained glycoproteins for the ERAD process.

In particular, reports indicate that for yeast prepro--factor expressed in mammalian cells (32), the Hong Kong null form of ₁-antitrypsin (33), or the -subunit of the T-cell antigen receptor (34), treatment of cells with DMM results in marked inhibition of the ERAD process. In detailed studies of the Hong Kong null form of ₁-antitrypsin, when ERAD was inhibited by cycloheximide treatment, calnexin dissociation and post-translational oligosaccharide trimming were blocked; by contrast, when ERAD was inhibited by DMM, the fraction of antitrypsin dissociated from calnexin appeared normal, whereas mannose trimming was nevertheless blocked (33). One simple interpretation of these data might be that ERAD proceeds by a pathway (blocked at different points by different inhibitors) that may involve calnexin dissociation, followed by additional steps that include trimming to Man₈. In such a case, the ultimate targeting for retrotranslocation might be expected to involve ER residents other than calnexin. However, in subsequent provocative experiments using a C-terminally truncated mutant form of ₁-antitrypsin, it was argued that a monoglucosylated Man₈ calnexin-bound form is the immediate precursor for ERAD (35). Indeed, pretreatment of cells with CAS to prevent the lectin-like calnexin association was reported to result in an ERAD process for mutant antitrypsin that cannot be inhibited by DMM (33) (nor even the proteasome inhibitor lactacystin (35)). This has led to a hypothesis that rebinding to calnexin might be an essential step for targeting misfolded N-linked glycoproteins for proteasomal proteolysis.

To test this hypothesis, we have now examined cell culture expression of the cog thyroglobulin (Tg) mutant (a L2263P substitution in Tg that causes congenital goiter in homozygous mice) and -carboxylation-deficient protein C, two misfolded glycoproteins that serve as models of genetic and acquired ER storage diseases, respectively (36, 37). As with mutant antitrypsin (38), ERAD of both proteins is significantly impaired upon specific inhibition of ER mannosidase I. However, in cells pretreated with CAS so that lectin-like binding of calnexin is blocked, the ERAD process is still proteasomal; moreover, it is still largely blocked upon inhibition of ER mannosidase I. Although the data do not exclude enhanced calnexin binding under specified experimental circumstances, they signify a more general role for recognition and trimming of specific N-linked mannose residues in the ERAD process for at least a subset of glycoproteins.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Materials-- Lactacystin was the generous gift of Dr. S. Omura (Kitasato Institute, Tokyo, Japan). CAS, chloroquine, DMM, tunicamycin, N-acetyl-L-leucyl-L-leucyl-L-norleucinal, and cycloheximide were from Sigma. Benzyloxycarbonyl-L-leucyl-L-leucyl-L-leucinal and E-64d ((L-3-trans-ethoxycarbonyloxirane-2-carbonyl)-L-leucine (3-methylbutyl)amide) were from The Peptide Institute (Osaka, Japan). Kifunensine (KIF) was from Toronto Research Chemicals (Toronto, Canada). 1,4-Dideoxy-1,4-imino-D-mannitol hydrochloride was purchased from Oxford GlycoSciences (Abingdon, United Kingdom). Endoglycosidase H was from New England Biolabs Inc. (Beverly, MA). Polyclonal rabbit antiserum against rat Tg was from Dr. P. R. Larsen (Brigham and Women's Hospital, Boston, MA). Polyclonal antiserum against human protein C was developed as described previously (37). Polyclonal antisera to BiP, calnexin, ERp72, calreticulin, and GRP94 were kindly provided by Dr. P. Kim (University of Cincinnati, Cincinnati, OH). Polyclonal antibodies to protein-disulfide isomerase and ER60 were the kind gift of Dr. T. Wileman (Pirbright Laboratories, Surrey, United Kingdom). Polyclonal antiserum to UDP-glucose:glycoprotein glucosyltransferase was graciously provided by Dr. A. Parodi (Fundacion Campomar, Buenos Aires, Argentina). Zysorbin was from Zymed Laboratories Inc. (South San Francisco, CA). [³⁵S]Methionine/cysteine (Expre³⁵S³⁵S) was from NEN Life Science Products. Methionine/cysteine-deficient DMEM and other tissue culture reagents, protease inhibitors, and stock chemicals were from Sigma.

Cell Culture, Cog Tg Transfection, and Selection of Stable Tg-expressing Clones-- CHO and BHK cells were maintained in DMEM containing 10% fetal bovine serum. The parental CHO cells employed were the W5 line generously provided by Dr. P. Stanley (Albert Einstein College of Medicine, Bronx, NY). The Cog Tg (L2263P) cDNA (36) was subcloned into the pCB6 vector encoding neomycin resistance and in which Tg expression is driven by the cytomegalovirus immediate early promoter. CHO cells were transfected using calcium phosphate precipitation. Transfectants were selected by growth in G418, and clonal expression of Tg was screened by using ³⁵S labeling, followed by Tg immunoprecipitation. The stably transfected BHK cells expressing protein C have been described previously (37).

Metabolic Labeling, Immunoprecipitation, and Gel Electrophoresis-- For pulse-chase experiments, cells were starved for 30 min in Met/Cys-free DMEM and 10% dialyzed fetal calf serum and labeled with Expre³⁵S³⁵S at a concentration of 100-250 µCi/ml. At the end of the pulse-labeling period, the cells were washed and chased with complete DMEM supplemented with Met/Cys. Unless otherwise indicated, inhibitors of glycoprotein processing or proteolysis were added 30 min before pulse labeling and throughout the labeling and chase period. For long-term labeling to approach steady state, the cells were labeled for 2 or 3 days in complete growth medium. In this case, when the cells were treated with labeling medium containing a drug, the medium was changed each day, and the same was done for control samples. At the end of each experiment, the medium was collected, and cells were lysed in 0.1% SDS and 1% Nonidet P-40 plus an anti-protease mixture (39). Samples were immunoprecipitated using rabbit anti-rat Tg or anti-human protein C antiserum, followed by 30 µl of Zysorbin. After washing, the immunoprecipitates were dissociated by heating in SDS gel sample buffer. After SDS-PAGE, radioactive band intensity was measured using a Storm 850 PhosphorImager (Molecular Dynamics, Inc.). In some instances, autoradiographic exposures were made using Kodak XAR films at 80 °C.

Microsomal Sedimentation Assay-- CHO cells expressing Cog Tg in 35-mm dishes were pulse-labeled and chased for 6 h. Then, the cells were scraped into 400 µl of 50 mM Hepes buffer, pH 7.4, containing 0.25 M sucrose and protease inhibitors and homogenized by 15 passages through a 27-gauge needle. Nuclei and unbroken cells were removed by centrifugation at 1100 × g for 10 min, and post-nuclear supernatants were further centrifuged at 50,000 rpm for 1 h in a Beckman TL-100 centrifuge. Precipitates (representing the microsomal fraction) and supernatants (representing the cytosolic fraction) were analyzed by immunoprecipitation, SDS-PAGE, and phosphorimaging.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Degradation of Cog Tg by CHO Cells-- To further explore the ERAD process, we stably expressed the cog Tg cDNA, which encodes the mouse Tg L2263P mutant, in CHO cells. As in COS cells (36), the Cog Tg protein was not secreted into the medium (Fig. 1A). Instead, after a 30-min pulse labeling with ³⁵S-labeled amino acids, following a lag time of ~2 h during which no degradation was apparent, newly synthesized Cog Tg thereafter was degraded with a subsequent half-time of 2-3 h (Fig. 1B). Based on this, recovery of Cog Tg at the 1-h chase time was generally taken as a reference point for most subsequent experiments. At all chase times, the undegraded intracellular Tg remaining was sensitive to digestion with endoglycosidase H (data not shown). Notably, the degradation was blocked by N-acetyl-L-leucyl-L-leucyl-L-norleucinal and, more specifically, by benzyloxycarbonyl-L-leucyl-L-leucyl-L-leucinal (MG-132) and lactacystin, indicating a fate of proteasomal proteolysis (Fig. 2). Thus, Cog Tg degradation in CHO cells displayed the classic hallmarks of ERAD (3). Moreover, in initial studies (data not shown), we found that, in common with several other N-linked glycoproteins (32-34), degradation of newly synthesized Cog Tg was inhibited by DMM. For these reasons, we examined KIF, a specific inhibitor of mannosidase I activity (40) in the ER (35, 41).

View larger version (27K): [in this window] [in a new window]   Fig. 1.   Intracellular degradation of Cog Tg in CHO cells. A, stably transfected CHO cells expressing Cog Tg were pulse-labeled for 30 min with 100 µCi/ml Expre35S35S and chased for the times indicated. Cell extracts and media were immunoprecipitated with anti-rat Tg antiserum and analyzed by SDS-4% PAGE under reducing conditions. B, shown is the quantitation of radioactive band intensities from A. Immunoreactive Tg in the pulse-labeled cell extracts was taken as 100%; the relative radioactivities of cell extracts () and media () are shown.

View larger version (52K): [in this window] [in a new window]   Fig. 2.   Proteasome-mediated degradation of Cog Tg in CHO. CHO cells expressing Cog Tg were pulse-labeled as described in the legend Fig. 1 and chased for 8 h in the absence or presence of various inhibitors. Upper panel, the cell extracts were immunoprecipitated for Tg and analyzed by SDS-4% PAGE and phosphorimaging. Lane 1 is a 1-h chased sample with no inhibitor. Lanes 2-8 were chased for 8 h in the presence of the following: no inhibitor (lane 2), 100 µM chloroquine (lane 3), 30 mM NH4Cl (lane 4), 100 µM E-64d (lane 5), 100 µM N-acetyl-L-leucyl-L-leucyl-L-norleucinal (ALLN; lane 6), 20 µM benzyloxycarbonyl-L-leucyl-L-leucyl-L-leucinal (LLL; lane 7), or 20 µM lactacystin (lane 8). Secretion of Cog Tg into the medium was not detected in the presence of any of these inhibitors (not shown). Lower panel, shown is the quantitation of radioactive band intensities from the upper panel, taking Tg radioactivity at the 1-h time (lane 1) as 100%.

By 16 h of chase in control CHO cells, newly synthesized Cog Tg was degraded completely (Fig. 3A, lane 2). However, this was substantially prevented in the presence of 200 µM KIF (Fig. 3A, lane 3). Furthermore, substantial suppression of Cog Tg degradation could be obtained with concentrations of KIF as small as 5 µM (Fig. 3B).

View larger version (43K): [in this window] [in a new window]   Fig. 3.   Pathway of proteasomal degradation of Cog Tg is inhibited by KIF, yet does not require lectin-like binding of calnexin. A, CHO cells stably expressing Cog Tg were pulse-labeled as described in the legend to Fig. 1. Lane 1 is a pulse-labeled sample with no inhibitor (control (Con)), and lanes 2-5 were chased for 16 h in the presence of no inhibitor (lane 2), 0.2 mM KIF (lane 3), 1 mM CAS (lane 4), or the combined addition of 0.2 mM KIF and 1 mM CAS (lane 5). Cells were pretreated with the inhibitors for 30 min before labeling, and treatment was continued throughout the chase. The cell extracts (left panel) and media (right panel) were immunoprecipitated for Tg and analyzed by SDS-4% PAGE and phosphorimaging. B, using the same labeling conditions as described for A, a dose response to KIF was performed. C, a protocol similar to that described for A was performed in the presence or absence of CAS plus no inhibitor (Con), a cysteine protease inhibitor (E-64d), or a proteasome-specific inhibitor (lactacystin).

Continuous treatment with CAS (which blocks glucosidases I and II and thereby prevents lectin-like binding of calnexin to newly synthesized glycoproteins) did not suppress the ERAD process for Cog Tg (Fig. 3A, lane 4), in agreement with previous reports (18, 20, 21, 23). Inhibition of glucose trimming after CAS treatment was confirmed by Cog Tg retardation upon SDS-PAGE (Fig. 3C, first two lanes) (42). In contrast with the behavior of truncated ₁-antitrypsin (35), lactacystin (and other proteasomal inhibitors not shown) stabilized Cog Tg regardless of CAS treatment (Fig. 3C, last two lanes). More importantly, unlike for mutant ₁-antitrypsin (33), suppression of glycoprotein ERAD with KIF was dominant over the ERAD that accompanied CAS treatment (Fig. 3A, lane 5). In these cells (treated with CAS in the presence or absence of KIF), SDS-PAGE mobility of Cog Tg was analyzed on 20-cm-long slab gels to look for small mobility differences between different glyco forms. At 2 h of chase, a small but definite KIF-mediated retardation of SDS-PAGE mobility was observed (Fig. 4, compare lanes 2 and 3), which preceded the suppression of Cog Tg degradation that became apparent only at later chase times (Fig. 3A, compare lanes 4 and 5). This shift in Cog Tg electrophoretic mobility must reflect inhibition of mannose trimming because no difference in mobility between KIF-treated and KIF-untreated samples persisted after digestion with jack bean mannosidase (Fig. 4, lanes 4 and 5). Thus, the data strongly indicate that after KIF treatment, Cog Tg protected from ERAD had failed to undergo trimming from Man₉ to Man₈.

View larger version (47K): [in this window] [in a new window]   Fig. 4.   SDS-PAGE mobility of Cog Tg after treatment with 1 mM castanospermine ± 0.2 mM kifunensine. CHO cells were pretreated with the indicated sugar-processing inhibitors and pulse-labeled as described in the legend to Fig. 1. After a 2-h chase, the cell lysates were immunoprecipitated for Tg, and some were digested further with peptide N-glycosidase F (P) or jack bean mannosidase (J). The samples were finally analyzed by 4% SDS-PAGE in a long (20 cm) slab gel format to discriminate small mobility differences.

Neither KIF alone nor KIF + CAS treatment was toxic to CHO cells, as neither treatment prevented the secretory trafficking of non-mutant Tg (Fig. 5). Moreover, following CAS treatment in the presence or absence of KIF, coprecipitation of Cog Tg with calnexin was undetectable (data not shown). It has been demonstrated for mutant ₁-antitrypsin that enhanced calnexin binding to monoglucosylated Man₉ oligosaccharides occurs following post-pulse protein synthesis inhibition with cycloheximide, and this correlates with protection from ERAD (33). Although cycloheximide (20 µM) treatment also had a stabilizing effect on Cog Tg, this effect was quite modest and tended to be limited to early chase times (see Fig. 8). Taken together, these data suggest that, for Cog Tg, the lectin-like binding of calnexin is not a major determinant for ERAD and that KIF inhibits a calnexin-independent step in targeting the misfolded glycoprotein for proteasomal degradation.

View larger version (48K): [in this window] [in a new window]   Fig. 5.   Mannose and/or glucose trimming does not affect the secretion of wild-type Tg. A, CHO cells expressing wild-type (Wt) Tg were labeled as described in the legend to Fig. 1 and chased for 16 h in the absence or presence of inhibitors. The cell extracts (left panel) and media (right panel) were immunoprecipitated for Tg and analyzed by SDS-4% PAGE and phosphorimaging. Lane 1 is a 1-h chased sample with no inhibitor. Lanes 2-4 are 16-h chased samples with no inhibitor (lane 2), 0.2 mM KIF (lane 3), and 0.2 mM KIF plus 1 mM CAS (lane 4). B, the relative radioactivities in the cells (black bars) and media (white bars) from A are shown.

Undegraded Cog Tg Is Not Liberated into the Free Cytosol-- Wiertz et al. (43) have suggested that oligosaccharide processing may occur in the cytosol at a step preceding proteasomal proteolysis, as evidenced by the fact that lactacystin treatment leads to accumulation of an undegraded cytosolic intermediate that can be detected in soluble supernatants after microsomal sedimentation. We therefore performed similar experiments at several chase times for Cog Tg in control CHO cells or in cells treated with lactacystin, CAS, or CAS + KIF. As shown in Fig. 6, lactacystin and CAS + KIF treatments led to Cog Tg stability (Cell lysate panels). However, at no time could any detectable Cog Tg intermediate be found liberated into the free cytosolic fraction. In additional trypsin protection experiments (data not shown), we found that Cog Tg and wild-type Tg were equally protected in microsomal fractions. Moreover, following the protocol described by Johnston et al. (44), we did not detect the presence of insoluble complexes (aggresomes) of newly synthesized Cog Tg accumulated after 10 h of chase in the presence of proteasomal inhibition (Fig. 7). Thus, it appears that a bottleneck for Cog Tg degradation occurs at or prior to the retrotranslocation process (see "Discussion").

View larger version (28K): [in this window] [in a new window]   Fig. 6.   Microsomal sedimentation assay for Cog Tg. Stably transfected CHO cells expressing Cog Tg were pulse-labeled for 30 min with 250 µCi/ml Expre35S35S and chased for up to 6 h in the absence or presence of 20 mM lactacystin, 1 mM CAS, or 1 mM CAS plus 0.2 mM KIF. KIF and/or CAS was added first 30 min before pulse labeling, whereas lactacystin was added only during the chase. To detect Cog Tg in whole cell lysates, a set of parallel samples were lysed and immunoprecipitated for Tg. Microsomal sedimentation was performed as described under "Experimental Procedures." Microsomal pellets (P) and cytosolic supernatant (S) fractions were each immunoprecipitated for Tg. All samples were finally analyzed by SDS-4% PAGE and phosphorimaging. Note that under all conditions, no cytosolic Tg was recovered at any chase time.

View larger version (65K): [in this window] [in a new window]   Fig. 7.   Cog Tg is not detected in detergent-insoluble aggregates. Stably transfected CHO cells expressing Cog Tg cells were pulse-labeled for 30 min with 200 µCi/ml Expre35S35S and either lysed without chase or chased for 10 h in the presence of 20 µM lactacystin. Cells were lysed in nondenaturing immunoprecipitation buffer (N) containing 1% Nonidet P-40, 0.5% deoxycholate, and a protease inhibitor mixture (see "Experimental Procedures") for 30 min on ice. After centrifugation at 13,000 × g for 15 min, the pellet was resolubilized by sonication in the presence 1% SDS (S). The SDS was then diluted to 0.05% with immunoprecipitation buffer, and both soluble and resolubilized fractions were immunoprecipitated with anti-Tg antiserum, followed by reducing SDS-4% PAGE and phosphorimaging. The labeled bands seen below the Tg position were not recovered in all experiments.

Generality of Kifunensine Effects-- So far, reports of kifunensine inhibition of ERAD are limited to just a few glycoproteins such as a C-terminally truncated version of ₁-antitrypsin (35) that contains a primary structural defect confined to a discrete protein domain. By contrast, treatment of cells with dithiothreitol at low concentrations (1 mM) prevents the formation of nascent disulfide bonds (45) and produces massive misfolding of Tg by preventing formation of many of the 60 intrachain disulfide linkages spanning much of the polypeptide (46). In CHO cells treated with 600 µM dithiothreitol, Cog Tg was reduced (Fig. 8B) and, during an 8-h chase, was degraded to at least the same extent as in untreated cells (Fig. 8A). Although post-pulse protein synthesis inhibition provided no suppression of ERAD under reducing conditions, KIF treatment effectively blocked the degradation of misfolded Tg (Fig. 8A, lower four panels).

View larger version (24K): [in this window] [in a new window]   Fig. 8.   Effects of combined addition of cycloheximide, dithiothreitol, and KIF on ERAD of Cog Tg. A, CHO cells expressing Cog Tg were pulse-labeled for 30 min with 100 µCi/ml Expre35S35S and chased for the times indicated in the absence (Control) or presence of inhibitors at the concentrations indicated. The cell extracts were immunoprecipitated for Tg and analyzed by reducing SDS-4% PAGE and autoradiography. B, CHO cells were pulse-labeled and analyzed at the earliest time as described for A, except that samples finally alkylated with 10 mM iodoacetamide were analyzed by SDS-4% PAGE under nonreducing conditions. CHX, cycloheximide; DTT, dithiothreitol.

To gain an independent assessment of the effect of mannosidase inhibition on ERAD, we examined recombinant protein C expressed in BHK cells. -Glutamyl carboxylation is a post-translational modification that promotes proper folding of protein C and other vitamin K-dependent clotting factors in the ER, whereas warfarin inhibition of -carboxylation results in misfolding and degradation of -carboxylation-deficient protein C (37) via a proteasomal mechanism (47). As for Cog Tg, a block in glucosidase activity (by CAS) did not stabilize the misfolded glycoprotein, but -carboxylation-deficient protein C was significantly stabilized in cells treated with DMM (data not shown) and, more specifically, with KIF (Fig. 9). KIF-mediated suppression of ERAD was also observed in CAS-treated BHK cells, indicating (as for Cog Tg) a dominant effect of the mannosidase I inhibitor (Fig. 9). Essentially identical observations were made in stably transfected CHO cells expressing protein C (data not shown).

View larger version (54K): [in this window] [in a new window]   Fig. 9.   Kifunensine also suppresses the intracellular degradation of -carboxylation-deficient protein C. BHK cells expressing protein C, cultured in the presence of 10 µM warfarin, were pulse-labeled for 30 min with 100 µCi/ml Expre35S35S (upper panel) or chased for 8 h in the absence (control (Con)) or presence of 1 mM CAS, 0.2 mM KIF, or 1 mM CAS plus 0.2 mM KIF. Cells were pretreated with the inhibitors for 30 min before labeling, and treatment was continued throughout the chase. At 8 h of chase, newly synthesized protein C in cell extracts (middle panel) and media (lower panel) was detected by immunoprecipitation, followed by SDS-9% PAGE and phosphorimaging. Of note, 1 mM DMM suppressed ERAD of -carboxylation-deficient protein C similarly to KIF, whereas 1 mM 1,4-dideoxy-1,4-imino-D-mannitol hydrochloride had no effect (not show).

In cells treated with tunicamycin, unglycosylated Cog Tg and -carboxylation-deficient protein C misfold and are degraded intracellularly with severely impaired anterograde transport through the secretory pathway (48, 49). However, as shown in Fig. 10, treatment with KIF had no protective effect on the intracellular degradation of these unglycosylated secretory proteins. These data are consistent with the hypothesis that KIF-mediated suppression of degradation is selective for glycoprotein substrates.

View larger version (29K): [in this window] [in a new window]   Fig. 10.   Kifunensine does not suppress ERAD of unglycosylated Cog Tg or -carboxylation-deficient protein C. CHO cells stably expressing Cog Tg (upper panels) or BHK cells stably expressing protein C and cultured in the presence of 10 µM warfarin (lower panels) were pretreated with tunicamycin (TUN; 10 µg/ml) before and during pulse labeling. When kifunensine was employed, cells were pretreated with 0.2 mM KIF, and the treatment was continued throughout the chase. Cells were either pulse-labeled only (control (Con) and lane 1) or chased (lanes 2 and 3) for 16 h (upper panels) or 8 h (lower panels), and the lysates were immunoprecipitated for Tg or protein C and analyzed by reducing SDS-PAGE and phosphorimaging. The first lane is a control showing the N-glycosylated form of each protein as a comparison for gel mobility. The panels on the right quantify (in arbitrary units) the recovery of the unglycosylated proteins in lanes 1-3 from the gel images shown on the left. cdPC, -carboxylation-deficient protein C.

Mannosidase I Processing Is a Relatively Slow Step in the ERAD Pathway-- It has been argued that since degradation intermediates do not routinely accumulate in the cytosol during ERAD, steps leading to retrotranslocation are likely to be rate-limiting for proteolytic cleavage (50). Although it is not clear if ER mannosidase I activity is rate-limiting in glycoprotein degradation, we note that KIF addition to CHO cells stabilized Cog Tg even when added 1 h after synthesis of the mutant glycoprotein (Fig. 11), at a time before any loss of Cog Tg could yet be detected (Fig. 1B). Protection from ERAD after delayed KIF addition suggested that for a significant fraction of Cog Tg molecules, the KIF-inhibitable step had not yet occurred even 1 h after synthesis. Similar protection from ERAD was obtained by delayed KIF addition to cells pretreated with CAS (data not shown).

View larger version (28K): [in this window] [in a new window]   Fig. 11.   Mannose trimming of Cog Tg is a relatively slow step. A, CHO cells expressing Cog Tg were pretreated and pulse-labeled as described in the legend to Fig. 3 and chased for 1 h (lane 1) or 16 h (lanes 2-5) in the absence (lanes 1 and 2) or presence of 0.2 mM KIF (lane 3), 0.2 mM KIF plus 1 mM CAS (lane 4), or 0.2 mM KIF only after a 1-h chase delay (lane 5). The cells (left panel) and media (right panel) were immunoprecipitated for Tg and analyzed by reducing SDS-4% PAGE and phosphorimaging. B, shown are the relative radioactivities of the 16-h chased samples from four independent experiments (mean ± S.D.) In addition, Tg recovery after pretreatment with 1 mM CAS is shown. Con, control.

In 10 µM warfarin-treated BHK cells, -carboxylation-deficient protein C was degraded with a half-time of ~2.5 h (Fig. 12). When DMM was added up to 1 h after pulse labeling, suppression of ERAD approached that observed in cells pretreated with the mannosidase inhibitor (Fig. 13). Thus, similar to Cog Tg, the step in the ERAD pathway blocked by mannosidase inhibition had not yet occurred for most molecules even 1 h after synthesis. Longer chase times prior to DMM addition led to a declining ability to suppress ERAD (Fig. 13). Similar results were obtained with KIF treatment (data not shown). These results suggest that mannosidase I processing is a relatively slow step after misfolded glycoprotein synthesis. We note that even when DMM was used in pretreatment and throughout the chase, some degradation of -carboxylation-deficient protein C still occurred (see "Discussion").

View larger version (21K): [in this window] [in a new window]   Fig. 12.   Kinetics of intracellular disposal of -carboxylation-deficient protein C. A, stably transfected BHK cells cultured in the presence of 10 µM warfarin were pulse-labeled as described in the legend to Fig. 9 and chased for the indicated times. Protein C in the cell extracts and media was immunoprecipitated with anti-protein C antiserum and analyzed by SDS-9%-PAGE under reducing (for cells) or nonreducing (for media) conditions. B, taking the radioactivity in the pulse-labeled cell extracts as 100%, the relative radioactivities of cells () and media () are shown. cdPC, -carboxylation-deficient protein C.

View larger version (21K): [in this window] [in a new window]   Fig. 13.   Effects of delayed addition of DMM on ERAD of -carboxylation-deficient protein C. BHK cells expressing protein C, cultured in the presence of 10 µM warfarin, were pulse-labeled and chased for 8 h. 1 mM DMM was added at the various times of chase shown on the y axis. In the sample marked Control, DMM was never added. Samples were analyzed as described in the legend to Fig. 9; the relative radioactivities of cells (black bars) and media (white bars) are shown. cdPC, -carboxylation-deficient protein C.

Chronic Mannosidase I Inhibition Causes ER Accumulation of Cog Tg-- Since acute KIF treatment caused a profound slowing of glycoprotein ERAD, we predicted that the misfolded protein would begin to accumulate in the ER. Therefore, normal CHO cells or CHO cells expressing Cog Tg were treated for 3 days with KIF (which was added to the medium each day). During this period, the cells were labeled continuously with ³⁵S-labeled amino acids at a constant low specific radioactivity. At the end of the treatment period, cells were examined by specific immunoprecipitation for their levels of Tg and a number of ER resident proteins. As shown in Fig. 14A, after daily addition of KIF for 3 days, Cog Tg showed some accumulation in the stably transfected CHO cells, and all of the accumulated Tg was sensitive to digestion with endoglycosidase H (data not shown). Quantitation showed that Cog Tg abundance after chronic KIF treatment rose ~3-fold. In control CHO cells, chronic KIF treatment itself had no significant effect on the levels of any of the ER resident proteins measured (Fig. 14A, first two lanes); more importantly, in stably transfected CHO cells, the accumulation of misfolded Cog Tg accompanying chronic KIF treatment also did not lead to a significant elevation of ER resident proteins (Fig. 14A, last two lanes). Similar results were obtained by routine Western blotting (data not shown). When pulse-chase experiments were performed to examine the fate of newly synthesized Cog Tg after 16 h in cells chronically treated daily with KIF for 3 days (Fig. 14B), a small (20%) increase in Tg degradation was observed, which involved proteasomal degradation as it was still sensitive to lactacystin (Fig. 14C). Together, the data raise the possibility of some form of adaptation to chronic mannosidase I inhibition that is associated with only modest steady-state accumulation of misfolded Tg that does not elevate ER chaperone levels.

View larger version (26K): [in this window] [in a new window]   Fig. 14.   Chronic kifunensine treatment causes accumulation of Cog Tg, but does not induce ER chaperone expression. A, parental CHO cells (Mock) or CHO cells expressing Cog Tg were metabolically labeled for 3 days (3d) with 100 µCi/ml Expre35S35S in complete DMEM in the absence or presence of 0.2 mM KIF. Cell lysates were immunoprecipitated for Tg and the various ER resident proteins indicated. Immunoprecipitations were analyzed by SDS-4% PAGE (for Tg) or SDS-8% PAGE (for the others) under reducing conditions. B, CHO cells expressing Cog Tg were either treated or untreated daily with 0.2 mM KIF for 3 days. At the end of this period, the cells were pulse-labeled and then either chased for 1 h and lysed as an estimate of the amount of Tg synthesized (P) or chased for 16 h in the presence of 0.2 mM KIF to examine the stability of Cog Tg (C). C, shown is the effect of lactacystin (+L; last lane), added during a 16-h chase, on the stability of Cog Tg in cells treated chronically with KIF as described for B. UGGT, UDP-glucose:glycoprotein glucosyltransferase.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Early carbohydrate trimming reactions have become a subject of increasing interest not only with respect to the folding of normal exportable glycoproteins and their exit from the ER (51), but also in the targeting of misfolded glycoproteins for ERAD. At the time of initial reports that ER mannosidase activity might function at an early step in the degradation of retained versions of exportable glycoproteins (32), a conceptual framework accounting for ERAD had not yet emerged. Studies in Saccharomyces cerevisiae brought forward the suggestion that calnexin might facilitate retrotranslocation to the cytosol (14). It should be noted, however, that this observation was made for unglycosylated pro--factor, for which it is unclear how the lectin-only model of calnexin activity (25, 26) would apply. At the same time, cells devoid of Mns1p (ER mannosidase I) were found to exhibit abnormally diminished degradation of a mutant form of carboxypeptidase Y (52); these results have recently been confirmed and extended to show that such degradation does not require the formation of glucosylated oligosaccharides (53). Taken together, these findings in S. cerevisiae indicate suppressive effects of ERAD as a consequence of decreased ER mannosidase I activity without an evident requirement for the lectin-like activity of calnexin.

Since the role of calnexin in S. cerevisiae may be unique to that organism (54), it is of interest to know how the foregoing reports compare with the situation for glycoprotein ERAD in higher eukaryotes. Specifically, although several groups have now found that inhibition of ER mannosidase activity can also suppress ERAD in mammalian cells (24, 32-34, 38), study of a C-terminally truncated ₁-antitrypsin expressed in mouse hepatoma cells led to the recent suggestion that ER mannosidase I specifically targets misfolded glycoproteins for disposal using a lectin-mediated process dependent upon binding monoglucosylated Man₈ oligosaccharides to calnexin (35). We therefore wished to confirm a role for early carbohydrate processing and calnexin binding in the ERAD process for Cog Tg and -carboxylation-deficient protein C, which cause inherited and acquired ER storage diseases, respectively (36, 37).

After synthesis, both misfolded glycoproteins were extensively degraded (Figs. 1 and 12) via a proteasomal mechanism. Confirming the work of Sifers and co-workers (35), we found that KIF, an inhibitor of ER mannosidase I (41), selectively suppressed ERAD of both misfolded glycoproteins (Figs. 3 and 9). More specifically, the data suggested that Cog Tg protected from ERAD failed to undergo trimming from Man₉ to Man₈ (Fig. 4). KIF treatment did not impair the anterograde traffic of native Tg or protein C from the ER (Fig. 5) (data not shown), nor did it affect the degradation of misfolded Tg or protein C that was unglycosylated after tunicamycin treatment (Fig. 10). However, KIF did appear to impair retrotranslocation of the misfolded glycoproteins to the cytosol, based upon microsomal sedimentation data (Fig. 6) and our unpublished observations that native and misfolded forms showed no difference in trypsin sensitivity in microsomal protease protection assays. However, we cannot exclude the possibility of a limited degree of retrotranslocation to the cytosol in the presence of lactacystin that is below the detection limits of our assays (23), although we could find no evidence for accumulation of insoluble Tg aggregates (Fig. 7) in cytosolic aggresomes (44). Suppressed Tg degradation was also observed in dithiothreitol-treated cells (Fig. 8); together, these data suggest that the effects of mannosidase I are not limited to specific protein domains, but seem to represent a more general feature in the ERAD process for at least a subset of glycoproteins.

Unlike for ₁-antitrypsin, suppression of ERAD with inhibitors of ER mannosidase I occurred efficiently even in cells continuously treated with CAS (Figs. 3 and 9). Under these conditions, the lectin-like binding of calnexin to monoglucosylated oligosaccharides cannot be initiated (18, 27, 28, 55); moreover, these conditions also interfere with glycoprotein association of calreticulin (56) as well as ER60/ERp57 (57). Thus, although data implicating lectin-like calnexin association have been presented in some cases (13, 35, 58), an important conclusion from this study (in conjunction with previous results (18-23)) is that a lectin-mediated association of the calnexin/calreticulin-ERp57 complex is not a general requirement for retrotranslocation of misfolded glycoproteins to the ERAD machinery. However, since ER chaperones are likely to be associated in a network (59) and since different ERAD substrates exhibit different patterns of chaperone binding (42), it is most probable that multiple ER chaperones have the potential to participate in the targeting of different misfolded protein substrates for retrotranslocation and ERAD.

Current reports suggest plasticity in the mechanisms leading to ER targeting for proteolysis. For example, ERAD occurs also for unglycosylated (misfolded) exportable proteins (Fig. 10) (14). Additionally, there is one N-linked glycoprotein (the T-cell receptor -subunit) for which inhibition of mannosidase activity does not significantly suppress the ERAD process (34). Importantly, complete deglycosylation has been reported as an intermediate stage in ERAD for several misfolded glycoprotein substrates (43, 60-63) based on peptide N-glycosidase activity that may be localized to the cytosol (64) or to the ER (65, 66). Based on these observations, it is reasonable to propose plasticity in the ERAD pathway involving the existence of distinct, and perhaps parallel, initial targeting. In this case, a Man₈-GlcNAc₂ oligosaccharide structure specific to the "B-isoform" configuration (i.e. such that after mannosidase cleavage, a terminal mannose remains in the 1,3-linkage (41)) may directly or indirectly facilitate protein disposition via one of these routes, and when this is blocked, undegraded intermediates presumably accumulate until alternative degradative routes prevail. The existence of such alternative routes presumably accounts for the fact that eventual degradation of Cog Tg and -carboxylation-deficient protein C is observed even in the presence of KIF. Thus, two alternative hypotheses that need to be tested are as follows. 1) Although not absolutely required, mannose reCognition and trimming may be an important priming step in the pathway of deglycosylation preceding ERAD; alternatively, 2) the B-isoform of Man₈-GlcNAc₂ oligosaccharides might facilitate the presentation of misfolded glycoproteins (53) for multi-ubiquitination (6, 22, 23, 34, 67). Either or both of these activities might be associated (physically or kinetically) with retrotranslocation through the Sec61 translocon (58).

With these thoughts in mind, we note that oligosaccharide trimming by ER mannosidase I appears to be a relatively slow step in the processing of Cog Tg and -carboxylation-deficient protein C (Figs. 11 and 13). We therefore expected that Cog Tg might massively accumulate in the ER in the setting of chronic KIF treatment, which could trigger the unfolded protein response (68). However, after 3 days of KIF treatment, Cog Tg accumulated only moderately, and no clear increase in BiP, GRP94, ERp72, calreticulin, or ER60 was observed (Fig. 14A). Potentially, alternative routes directing Cog Tg to proteasomal degradation could be induced under these conditions (which may be consistent with the results of Fig. 14 (B and C)). We do not yet know if cells may exhibit a compensatory response to chronic mannosidase I blockade, such as further induction of ER mannosidase I, endomannosidase (69, 70), or other activities. We are currently pursuing these possibilities.

	ACKNOWLEDGEMENTS

We gratefully acknowledge the assistance of Drs. S. Omura, P. R. Larsen, P. Kim, T. Wileman, A. Parodi, and P. Stanley, who provided valuable reagents and antibodies. We thank Young-Nam Park for providing the Cog Tg cDNA subcloned into the pCB6 vector. We thank members of the Arvan laboratory, Drs. R. G. Spiro (Joslin Diabetes Center, Boston), and A. Chang (Albert Einstein College of Medicine) for helpful discussions.

	FOOTNOTES

^* This work was supported by National Institutes of Health Grant DK40344 (to P. A.); Grant-in-aid 09276103 for Scientific Research on Priority Areas (Molecular Chaperones) from the Ministry of Education, Science, Sports, and Culture of Japan; and a research grant from the Uehara Memorial Science Foundation (to F. T.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

To whom correspondence should be addressed: Div. of Endocrinology, Dept. of Developmental and Molecular Biology, Albert Einstein College of Medicine, 1300 Morris Park Ave., Bronx, NY 10461. Tel.: 718-430-8685; Fax: 718-430-8557; E-mail: arvan@aecom.yu.edu.

Published, JBC Papers in Press, September 12, 2000, DOI 10.1074/jbc.M001073200

	ABBREVIATIONS

The abbreviations used are: ER, endoplasmic reticulum; ERAD, endoplasmic reticulum-associated degradation; CAS, castanospermine; DMM, 1-deoxymannojirimycin; Tg, thyroglobulin; KIF, kifunensine; DMEM, Dulbecco's modified Eagle's medium; CHO, Chinese hamster ovary; BHK, baby hamster kidney; PAGE, polyacrylamide gel electrophoresis.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

</