Enhancement of endoplasmic reticulum (ER) degradation of misfolded Null Hong Kong alpha1-antitrypsin by human ER mannosidase I.

Misfolded glycoproteins synthesized in the endoplasmic reticulum (ER) are degraded by cytoplasmic proteasomes, a mechanism known as ERAD (ER-associated degradation). In the present study, we demonstrate that ERAD of the misfolded genetic variant-null Hong Kong alpha1-antitrypsin is enhanced by overexpression of the ER processing alpha1,2-mannosidase (ER ManI) in HEK 293 cells, indicating the importance of ER ManI in glycoprotein quality control. We showed previously that EDEM, an enzymatically inactive mannosidase homolog, interacts with misfolded alpha1-antitrypsin and accelerates its degradation (Hosokawa, N., Wada, I., Hasegawa, K., Yorihuzi, T., Tremblay, L. O., Herscovics, A., and Nagata, K. (2001) EMBO Rep. 2, 415-422). Herein we demonstrate a combined effect of ER ManI and EDEM on ERAD of misfolded alpha1-antitrypsin. We also show that misfolded alpha1-antitrypsin NHK contains labeled Glc1Man9GlcNAc and Man5-9GlcNAc released by endo-beta-N-acetylglucosaminidase H in pulse-chase experiments with [2-3H]mannose. Overexpression of ER ManI greatly increases the formation of Man8GlcNAc, induces the formation of Glc1Man8GlcNAc and increases trimming to Man5-7GlcNAc. We propose a model whereby the misfolded glycoprotein interacts with ER ManI and with EDEM, before being recognized by downstream ERAD components. This detailed characterization of oligosaccharides associated with a misfolded glycoprotein raises the possibility that the carbohydrate recognition determinant triggering ERAD may not be restricted to Man8GlcNAc2 isomer B as previous studies have suggested.

The synthesis of glycoproteins containing N-glycans begins in the endoplasmic reticulum (ER) 1 with the addition of the Glc 3 Man 9 GlcNAc 2 precursor of N-linked glycans to nascent polypeptide chains. Subsequently, the ER lectin chaperones, calnexin or calreticulin, specifically bind to monoglucosylated N-glycans, promoting proper glycoprotein folding. This ER quality control process is mediated by glucosidases I and II that remove the glucose residues from Glc 3 Man 9 GlcNAc 2 and by UDP-glucose:glycoprotein glucosyltransferase that reglucosylates incompletely folded glycoproteins. Correctly folded glycoproteins exit the ER to their final destinations, whereas misfolded glycoproteins are readily degraded (for reviews see Refs. [1][2][3][4]. Many terminally misfolded proteins in the ER are degraded by cytoplasmic proteasomes, a mechanism known as ERAD (for reviews see Refs. [5][6][7]. Experimental evidence in yeast and mammalian cells suggests that ER ␣1,2-mannosidase I (ER ManI) that primarily removes the middle branch mannose from Man 9 GlcNAc 2 to form Man 8 GlcNAc 2 isomer B (Man8B) (8,9) acts as a signal triggering ERAD of glycoproteins (10 -12). Disruption of the ER ␣-mannosidase gene in yeast (10,11) and inhibition of ER ManI in mammalian cells (13,14) both prevent ERAD. We recently reported the molecular cloning of mouse EDEM (ER degradation enhancing ␣-mannosidase-like protein) and its involvement in glycoprotein ERAD. It was suggested that EDEM is a putative lectin, which most likely binds Man8B on misfolded ␣ 1 -antitrypsin and accelerates its degradation (15). Two groups also reported the yeast homolog of EDEM and its function in yeast ERAD (16,17), indicating that the ERAD mechanisms involved in the degradation of misfolded glycoproteins are similar in yeast and mammals. Both ER ManI and EDEM are ER resident transmembrane proteins containing characteristic signature motifs of class I ␣1,2-mannosidases (glycosylhydrolase family 47). Despite this significant sequence similarity, EDEM lacks ␣1,2mannosidase activity with Man 9 GlcNAc as substrate (15). In addition, these two proteins differ in their response to ER stress. EDEM is induced by various forms of ER stress, but ER ManI is not (15).
In this report, we show that overexpression of human ER ManI accelerates the degradation of the terminally misfolded ␣ 1 -antitrypsin genetic variant-null Hong Kong (A1AT NHK) (18). We also show the combined effects of ER ManI and EDEM on the degradation of misfolded ␣ 1 -antitrypsin. Glycan analysis on misfolded NHK shows that overexpression of ER ManI greatly increases the formation of Man 8 GlcNAc 2 and Glc 1 Man 8 GlcNAc 2 and also stimulates trimming of N-linked oligosaccharides to Man 5 GlcNAc 2 in vivo. These experiments suggest that the misfolded glycoprotein interacts both with ER ManI and EDEM before being recognized by downstream ERAD components.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfection-Human HEK 293 cells were cultured in DMEM supplemented with 10% fetal bovine serum, and transfection of plasmids encoding human ER ManI, mouse EDEM, and human ␣ 1 -antitrypsin-null Hong Kong were performed using FuGENE 6 transfection reagents (Roche Molecular Biochemicals) according to the protocol recommended by the manufacturer. Approximately 1.5 ϫ 10 5 cells were plated on poly-L-lysine-coated 3.5-cm dishes. After 24 h of plating, 0.5 g of plasmid encoding ER ManI, EDEM, or pMH vector were mixed with 1 g of plasmid encoding the ␣ 1 -antitrypsin variant A1AT NHK for transfection, and cells were pulse-labeled and harvested 36 h post-transfection.
Reagents-Kifunensine, an inhibitor of ER ManI, was kindly provided by Fujisawa Pharmaceutical Co. (Osaka, Japan) and added to the culture medium at a concentration of 5 g/ml for 4 h prior to pulselabeling. The proteasome inhibitor lactacystin was purchased from Kyowa Medics Co. (Tokyo, Japan) and added to the culture medium at a concentration of 1 mM for 4 h prior to pulse-labeling. The inhibitors were also present in the medium during the pulse and the chase periods.
Plasmid Construction-Human ER ManI cDNA was subcloned into the pMH vector (Roche Molecular Biochemicals) by PCR. The entire open reading frame (9) was amplified from cDNA using a sense primer containing a HindIII site and a Kozak sequence (5Ј-AAAAAAGCTTC-CACCATGGCTGCCTGCGAGGGCAGGAG-3Ј) and an antisense primer containing a NotI site and omitting the stop codon (5Ј-AAA-AAAAAGCGGCCGCGCTGCAGGGGTCCAGATAGGCAGAG-3Ј). The open reading frame amplicon was digested with HindIII and NotI and cloned into the HindIII/NotI sites of pMH in-frame with the C-terminal HA-tag. Mouse EDEM cDNA was tagged with HA at its C terminus in pCMV-SPORT2 vector (Invitrogen) as described previously (15). The NHK variant was constructed as described elsewhere (15).
Metabolic Labeling and Immunoprecipitation-Cells were preincubated in DMEM lacking methionine (Invitrogen) for 30 min and labeled with Expre 35 S 35 S protein labeling mixture (PerkinElmer Life Sciences) for 15 min at a concentration of 8.2 MBq/ml. The cells were then incubated in DMEM during the chase period. After washing twice with phosphate-buffered saline lacking Ca 2ϩ and Mg 2ϩ , cells were incubated on ice for 20 min in lysis buffer (1% Nonidet P-40, 150 mM NaCl, 50 mM Tris-HCl, pH 8.0, supplemented with protease inhibitors) (15). For immunoprecipitation, Nonidet P-40 soluble cell lysates were mixed with various antibodies (1/50 dilution), and the immune complexes were collected using either Protein A-or Protein G-Sepharose beads (Amersham-Pharmacia, Amersham, UK) (19). One-half of the cell lysate was used for immunoprecipitation using antibody against A1AT, and the other half was incubated with anti-HA antibody. In cells co-transfected with NHK and ER ManI or EDEM, the radioactivity of immunoprecipitated NHK decreased compared with cells transfected with NHK and pMH. To obtain approximately equal signals for the immunoprecipitated NHK at chase 0 h, the gels were exposed 2 or 3 times longer for samples co-transfected with ER ManI or EDEM than for those cotransfected with pMH. Alternatively in some experiments, one-half or one-third of the cell lysates was used for immunoprecipitation in pMH transfected cells in comparison with ER ManI or EDEM transfected cells. The conditions are indicated in the figure legends. The immune complex was boiled in Laemmli's sample buffer containing 0.1 M dithiothreitol and separated by 10% SDS-PAGE, unless otherwise indicated in the legends. Radioactivity was quantified by exposing the gels to a PhosphorImager (Storm, Amersham Biosciences), and the gels were exposed to x-ray films (HR-HA, Fuji Photo Film Co. Ltd., Japan) afterward.
For labeling with [ 3 H]mannose, cells were preincubated in DMEM containing 1 mM glucose (Invitrogen) for 30 min and labeled with 7.4 MBq/ml D-[2-3 H]mannose for 1 h. Cells were then chased for the times indicated in DMEM containing 25 mM glucose (Invitrogen) supplemented with 5 mM mannose (20). Cell lysis, immunoprecipitation, and SDS-PAGE were performed as described for the labeling with 35 Sprotein-labeling mixture, and the samples were blotted onto a PVDF membrane (Bio-Rad) in 5 mM sodium tetraborate buffer. The position of the [ 3 H]mannose-labeled NHK bands on the PVDF membrane were identified by exposing the membrane to a PhosphorImager.
Antibodies-Antiserum against ␣ 1 -antitrypsin was purchased from Cappel (ICN Pharmaceuticals, Inc.), and purified IgG against the HAtag was obtained from Santa Cruz Biotechnology, Inc. Human ER ManI antibody was prepared by immunizing rabbits with a keyhole limpet hemocyanin-conjugated synthetic peptide (GRRDVEVKPADRHNLLR-PET). The antiserum was affinity-purified using the peptide. For the antibody against mouse EDEM, multiple antigenic peptide was synthesized using the following peptide sequence, DERRYSLPLKSIYMRQID. Whole rabbit serum was used for the experiments.
Western Blotting-Nonidet P-40-soluble cell lysates or immunoprecipitates prepared as described above (see "Metabolic Labeling and Immunoprecipitation") were adjusted to 1 ϫ Laemmli's buffer containing 0.1 M dithiothreitol and separated by 10% SDS-PAGE. For Western blotting, 40 g of protein was loaded per lane. For Western blotting following immunoprecipitation, 400 g of protein of Nonidet P-40-soluble cell lysate was used as starting material. After blotting to a nitrocellulose membrane, antigen-antibody complexes were detected by ECL (Amersham Biosciences) and exposed to x-ray films.
Characterization of Oligosaccharides-The [ 3 H]mannose-labeled NHK bands on the PVDF membranes were cut into small pieces. The pieces were first rinsed with methanol and then with 0.1 M sodium citrate, pH 5.5, three times for 15 min at 37°C, and the washes were discarded. The membrane pieces were then incubated with Endo-H (New England Biolabs) in the same buffer at 37°C for 48 h. Endo-H (1000 units) was added at 0, 12 and 36 h of incubation, and the membrane pieces were washed three times with water. The extracts and washes were combined, boiled for 3 min, and lyophilized. The labeled oligosaccharides completely released by this procedure were dissolved in water, mixed with standard 14 C-labeled Glc 3 Man 9 GlcNAc and fractionated by HPLC as described previously (21). In some cases, an internal standard of 14 C-labeled Glc 1 Man 8 GlcNAc was also used. It was prepared by treatment of labeled Glc 1 Man 9 GlcNAc for 45 min at 37°C with 0.2 mg of purified recombinant ␣1,2-mannosidase of Saccharomyces cerevisiae (22).

Enhanced Degradation of A1AT NHK by Overexpression of ER ManI-
To examine the effect of ER mannosidase I on ERAD of misfolded glycoproteins, human ER ManI cDNA was transfected into HEK 293 cells along with the A1AT NHK. A1AT is a serum protein belonging to a serine protease inhibitor superfamily. Mutations in ␣ 1 -antitrypsin cause emphysema or liver cirrhosis (23). The NHK genetic variant of A1AT misfolds terminally in the ER and is degraded by ERAD (12,15). Transfection of human ER ManI greatly enhanced the degradation of misfolded A1AT NHK in comparison with mock transfected cells (Fig. 1A, closed arrow, compare lanes 1-3 with lanes 4 -6). NHK was not secreted into the medium of cells co-transfected with ER ManI (data not shown) nor of cells co-transfected with vector alone (15,24). Quantitative analysis revealed that the t1 ⁄2 of transfected NHK in 293 cells was ϳ100 min. Co-transfection of ER ManI shortened the t1 ⁄2 of NHK to about 50 min (Fig. 1C).
NHK immunoprecipitated from cell lysates transfected with ER ManI migrated slightly faster on SDS-PAGE than that from mock transfected cells (Fig. 1A, closed arrow). Kifunensine inhibited the degradation of NHK in cells co-transfected with ER ManI (Fig. 2, A and C) and also resulted in a small decrease in NHK electrophoretic mobility, consistent with inhibition of ER ␣1,2-mannosidase activity ( Fig. 2A, black  arrow, compare lanes 1-3 with lanes 4 -6). Therefore overexpression of ER ManI enhances ERAD of misfolded NHK by accelerating mannose trimming from its N-linked glycans. This conclusion was further supported by NHK oligosaccharide analyses described below.
Co-immunoprecipitation of ER ManI with NHK was detected in cells transfected with ER ManI (Fig. 1A, open arrow). Notably, the transfected ER ManI was degraded very rapidly (Fig.  1B, lanes 4 -6, and see below). A broad uncharacterized protein band of ϳ170 kDa was co-immunoprecipitated with ER ManI using anti-HA antibody (Fig. 1B, lanes 4 -6). This 170-kDa band decreased in parallel with the HA-tagged ER ManI during chase.
The degradation of co-transfected ER ManI was not inhibited by the addition of kifunensine (Fig. 2B, compare lanes 1-3 with lanes 4 -6), in accordance with the fact that human ER ManI is not a glycoprotein (8,9). Kifunensine also decreased the mobility of the 170-kDa co-immunoprecipitated protein band (Fig.  2B, compare lanes 1 and 4), suggesting that the 170-kDa band is glycosylated.
Inhibition of NHK Degradation by Lactacystin-The effect of the proteasome inhibitor lactacystin on NHK degradation was then examined. The accelerated degradation of NHK in cells transfected with ER ManI was repressed by lactacystin ( Fig.  3A, arrow, compare lanes 1-3 with lanes 4 -6), indicating that the enhanced NHK degradation induced by ER ManI overexpression occurs through the ERAD pathway. Again, it is notable that the transfected ER ManI is rapidly degraded (Fig. 3B,  lanes 1-3, open arrow, see also Figs. 1B and 2B). General protein synthesis examined by the incorporation of [ 35 S]methionine into the trichloroacetic acid-insoluble fraction was not inhibited by transfection of ER ManI. The degradation of ER ManI is independent of proteasomes (Fig. 3B, compare lanes  1-3 with lanes 4 -6). The degradation of ER ManI was not inhibited by other protease inhibitors including serine or cys-teine protease inhibitors, calpain inhibitor, nor aspartic protease inhibitor. 2 Combined Effects of ER ManI and EDEM on NHK Degradation-Recently, we reported that mouse EDEM, which has sequence homology to ␣1,2-mannosidases but lacks enzyme activity with Man 9 GlcNAc as substrate, accelerates ERAD of misfolded NHK when transfected into 293 cells (15). EDEM was shown to interact with NHK in co-immunoprecipitation experiments and was postulated to recognize misfolded glycoproteins to sort them for retrotranslocation. We first compared the effects of transfection with ER ManI and with EDEM on ERAD of NHK. ER ManI and EDEM expressed individually shortened the intracellular half-life of NHK (Fig.  4A, closed arrow, compare lanes 1-3 with lanes 4 -6 or 7-9). Co-immunoprecipitation of either ER ManI or EDEM was detected with NHK, using antibody raised against A1AT (Fig.  4A, lanes 4 -6, open arrow showing ER ManI; lanes 7-9, thin arrow indicating EDEM). In the immunoprecipitates using anti-HA antibody, protein bands of ϳ170 and 150 kDa on SDS-PAGE were observed in ER ManI and EDEM transfected cells, respectively (Fig. 4B, lanes 4 -6, open arrowhead for ER ManI; lanes 7-9, closed arrowhead for EDEM). Although transfected ER ManI is rapidly degraded, transfected 2 N. Hosokawa, I. Wada, and K. Nagata, unpublished observation. EDEM is stable during the 2-h chase period (Fig. 4B, compare  lanes 4 -6 with lanes 7-9).
Co-transfection of 293 cells with both ER ManI and EDEM further enhanced ERAD of NHK, compared with ER ManI alone, showing that there is a combined effect of these two ER resident proteins on NHK degradation (Fig. 5A, compare lanes  1-3 with lanes 4 -6). Enhancement of degradation of NHK was also evident in cells co-transfected with both ER ManI and EDEM compared with EDEM alone (Fig. 5A, compare lanes  4 -6 with lanes 7-9). Because of the rapid degradation of NHK observed following transfection of both ER ManI and EDEM in this experiment, cell lysates were examined at shorter chase periods. The levels of expression of ER ManI, EDEM, and NHK were reduced when both ER membrane proteins and substrate proteins were co-transfected. However, when the proteasome inhibitor lactacystin was added, similar amounts of labeled NHK were recovered from the immunoprecipitates of cell lysates (Fig. 5B, lanes 3, 4, 7, and 8), whereas only a very small amount of NHK was detected in cells overexpressing both ER ManI and EDEM after a 45-min chase in the absence of lactacystin (Fig. 5B, lane 6). These results confirm that the reduced amount of labeled NHK immunoprecipitated from cells cotransfected with ER ManI and EDEM was actually due to degradation by the proteasomes and not due to reduced synthesis of NHK.
We then investigated whether ER ManI and EDEM form a complex with each other. We prepared antisera against synthetic peptides from both ER ManI and EDEM. Each antiserum recognized its corresponding overexpressed protein by immunoprecipitation and Western blot analysis but hardly de-

FIG. 2. Kifunensine inhibits the degradation of NHK in ER ManI transfected cells.
Kifunensine (5 g/ml) was added to the culture medium 4 h prior to metabolic labeling and was present in the medium during the pulse-labeling and chase period. Both NHK and ER ManI were transfected in all the cells. Immunoprecipitates were prepared as in Fig. 1. Arrow notations are as described in Fig. 1. A and B, autoradiogram of the immunoprecipitates using antibody (Ab) against A1AT (A) and HA-tag (B), respectively. C, quantification of NHK within the cell. The radioactivity of NHK at 0 h chase was set arbitrarily to 1.0, and the means with standard deviations of three independent experiments were shown. tected endogenous levels of the proteins. Accordingly, we cotransfected HA-tagged ER ManI with untagged EDEM or untagged ER ManI with HA-tagged EDEM. We performed immunoprecipitation of labeled cell lysates, as well as immunoprecipitation of cell lysates followed by Western blotting using either HA-tag or synthetic peptide antibodies. We did not detect any interaction of overexpressed ER ManI with EDEM (data not shown). The co-immunoprecipitated band observed in cells transfected with ER ManI was consistently larger than that observed in EDEM transfected cells (Fig. 4B, lanes 4 -9,  open and closed arrowheads), indicating that the proteins differ.
Characterization of Oligosaccharides on NHK-To determine the structure of the oligosaccharides present on NHK, cells were labeled with [2-3 H]mannose, and the oligosaccharides released by Endo-H from immunoprecipitated NHK were analyzed by HPLC (Fig. 6). In mock transfected cells the major labeled oligosaccharides on NHK were Glc 1 Man 9 GlcNAc and Man 9 GlcNAc following pulse-labeling, along with smaller amounts of Man 8 GlcNAc (Fig. 6A). The identity of the oligosac-charides was confirmed by exhaustive treatment with jack bean ␣1,2-mannosidase followed by HPLC of the products. Glc 1 Man 9 GlcNAc yielded about 47% labeled Glc 1 Man 4 GlcNAc and mannose whereas Man 9 GlcNAc was entirely converted to mannose (data not shown). With increasing times of chase, trimming of NHK glycans to Man 7 GlcNAc, Man 6 GlcNAc (Fig.  6, B and C), and even to Man 5 GlcNAc after 4 h of chase (data not shown) was observed. Therefore, the oligosaccharides on misfolded NHK are trimmed beyond Man 8 GlcNAc 2 by endogenous ␣1,2-mannosidase activity.
Overexpression of ER ManI caused a large increase in Man 8 GlcNAc, and an appearance of Glc 1 Man 8 GlcNAc (Fig. 6, D, E, and F, and Fig. 7A). At the same time, there was  For lanes 4 -6, the film was exposed twice as long. The result of one representative experiment is shown, and the relative radioactivity of the NHK is indicated below each lane (radioactivity at chase 0 h of each transfection was arbitrarily set to 1.0). B, cells were pulse-labeled for 15 min (P) and chased for 45 min (C) with or without lactacystin. Specific immunoprecipitates using anti-A1AT antibody are shown, and the relative radioactivity was indicated below each lane (radioactivity of lane 1 was arbitrarily set to 1.0). a large decrease in the proportion of Glc 1 Man 9 GlcNAc and Man 9 GlcNAc 2 as well as an increase in Man 5-7 GlcNAc compared with mock transfected cells. An aliquot of the sample from Fig. 6F subjected to HPLC with an internal standard of Glc 1 Man 8 GlcNAc (Fig. 7A) confirmed the identity of the large peak of Glc 1 Man 8 GlcNAc and the disappearance of Man 9 GlcNAc in cells transfected with ER ManI compared with mock transfected cells. The Glc 1 Man 8 GlcNAc may have been formed by the action of ER ManI on Glc 1 Man 9 GlcNAc or by the addition of glucose to the increased Man 8 GlcNAc by UDPglucose:glycoprotein glucosyltransferase. These results show that ER ManI is capable of trimming beyond Man 8 GlcNAc 2 in vivo as was recently reported for recombinant ER ManI in vitro (25). In cells overexpressing EDEM the major labeled oligosaccharides were Glc 1 Man 9 GlcNAc and Man 9 GlcNAc, as in mock transfected cells with smaller proportions of Man 6 -8 GlcNAc (Fig. 6, G, H, and I). When a sample similar to that shown in Fig. 6I was fractionated in the presence of the Glc 1 Man 8 GlcNAc internal standard (Fig. 7B), Man 9 GlcNAc and a small amount of Glc 1 Man 8 GlcNAc were found. When cells were treated with kifunensine the formation of Man 6 -8 GlcNAc was nearly completely inhibited (Fig. 8B) showing that ER ␣1,2mannosidase II is not responsible for trimming NHK oligosaccharides in vivo.

DISCUSSION
In the present study, we have shown that overexpression of human ER ManI in HEK 293 cells enhances ERAD of the terminally misfolded glycoprotein NHK. Accelerated degradation of NHK in cells overexpressing ER ManI was inhibited by lactacystin and by kifunensine, indicating that misfolded NHK was degraded through the ERAD pathway and that overexpression of ER ManI enhanced ERAD by accelerating trimming of its N-glycans. It has been postulated that the processing of N-linked oligosaccharides to Man8B by ER ManI acts as a signal for degradation of misfolded glycoproteins (1, 7, 13, 14).  Fig. 7), and G1M9, which co-eluted with 14 C-labeled Glc 1 Man 9 GlcNAc (not shown), corresponds to Glc 1 Man 9 GlcNAc.
However, all studies to date were based on the assumption that Man8B is the only oligosaccharide found on misfolded glycoproteins due to ER ManI activity and on observations of carbohydrate trimming from misfolded glycoproteins by shifts in their mobility upon SDS-PAGE. In this first study reporting HPLC analysis of the glycans on terminally misfolded NHK, the results clearly show that a mixture of oligosaccharides is found on NHK and that the accepted assumption that only Man8B is formed by ER ManI needs to be revised.
In mock transfected cells, Glc 1 Man 9 GlcNAc, Man 9 GlcNAc, and Man 8 GlcNAc were the major oligosaccharides found on NHK, but smaller amounts of Man 7 GlcNAc and Man 6 GlcNAc were also present. It is therefore possible that oligosaccharides other than Man8B, or in addition to Man8B, may act as recognition markers for ERAD. The relatively small amounts of Man 7 GlcNAc and Man 6 GlcNAc found on NHK might be the result of selective ERAD of specific NHK glycoforms bearing these oligosaccharides. The presence of Glc 1 Man 9 GlcNAc on NHK is consistent with previous studies demonstrating NHK association with calnexin (26). The results also show that the additional trimming beyond Man 8 GlcNAc is not due to ER ␣-mannosidase II because it was sensitive to kifunensine (Fig. 8).
Overexpression of ER ManI, which increased ERAD of NHK, also greatly stimulated trimming of mannose residues to Man 5 GlcNAc, Man 6 GlcNAc, Man 7 GlcNAc, and Man 8 GlcNAc, as well as the formation of Glc 1 Man 8 GlcNAc. With the exception of the latter, the same oligosaccharides were formed in mock transfected cells, showing that the stimulation of NHK degradation caused by overexpression of ER ManI is not likely caused by the formation of atypical glycans. The increased trimming of N-glycans beyond Man 8 GlcNAc observed in cells overexpressing ER ManI is consistent with previous experiments with recombinant ER ManI in vitro indicating that ER ManI is less specific than previously believed (25).
The pattern of oligosaccharides found in EDEM and mock transfected cells is very similar, but there is always a shoulder ahead of the Man 9 GlcNAc peak that corresponds to a small We demonstrated combined effects of ER ManI and of EDEM on ERAD. We previously reported that EDEM interacts with and accelerates the degradation of misfolded NHK, suggesting that EDEM may be a putative lectin, which recognizes misfolded glycoproteins for ERAD (15). NHK degradation is faster in cells co-transfected with both ER ManI and EDEM than in cells transfected with ER ManI or EDEM alone (Fig. 5A).
We also obtained evidence indicating that ER ManI and EDEM are part of different complexes because different proteins were co-immunoprecipitated with each of them, and no evidence that ER ManI and EDEM form a complex with each other was obtained. We propose a model whereby misfolded glycoproteins are recognized independently by ER ManI and by EDEM. The identities of the proteins co-immunoprecipitated with ER ManI and EDEM remain to be determined, which may aid in the elucidation of productive glycoprotein folding and in understanding the machinery leading to retrotranslocation and subsequent degradation by proteasomes.
Notably, ER ManI transfected into 293 cells was degraded very rapidly through a mechanism independent of cytoplasmic proteasomes (Fig. 3B, lanes 1-6). Despite its rapid turnover, we could detect the expression of transfected ER ManI by Western blot analysis using antibody against HA-tag or human ER ManI peptide. The level of expression of ER ManI in transfected cells is higher than that of endogenous ER ManI because we could not detect the latter by Western blot analysis using anti-peptide antibody. Although the molecular mechanism and the reason for the rapid degradation of ER ManI will require further investigation, the results suggest that the level of ER ManI expression is stringently controlled.