Dislocation of an Endoplasmic Reticulum Membrane Glycoprotein Involves the Formation of Partially Dislocated Ubiquitinated Polypeptides*

Accumulation of improperly folded polypeptides in the endoplasmic reticulum (ER) can trigger a stress response that leads to the export of aberrant proteins into the cytosol and their ultimate proteasomal degradation. Human cytomegalovirus encodes a type I glycoprotein, US11, that binds to nascent MHC class I heavy chain molecules and causes their dislocation from the ER to the cytosol where they are degraded by the proteasome. Examination of US11-mediated class I degradation has identified a host of cellular proteins involved in the dislocation reaction, including the cytosolic AAA ATPase p97, the membrane protein Derlin-1, and the E3 ubiquitin ligase Sel1L. However, the intermediate steps occurring between the initiation of dislocation and full extraction of the misfolded substrate into the cytosol are not known. We demonstrate that US11 itself undergoes ER export and proteasomal degradation and utilize this system to define multiple steps of US11 dislocation. Treatment of US11-expressing cells with proteasome inhibitor resulted in the accumulation of glycosylated and ubiquitinated species as well as a deglycosylated US11 intermediate. Subcellular fractionation of proteasome-inhibited US11 cells demonstrated that deglycosylated intermediates continued to be integrated within the ER membrane, suggesting that the proteasome functions in the latter steps of dislocation. The data supports a model in which US11 is modified with ubiquitin, whereas the transmembrane region is integrated in the ER membrane, and deglycosylation occurs before complete dislocation.

The endoplasmic reticulum (ER) 3 of eukaryotic cells is the site of synthesis of membrane and secretory proteins. Polypeptides are cotranslationally inserted into the ER lumen where they can be modified by the addition of asparagine-linked gly-can precursors. Proteins achieve proper tertiary structure with the aid of ER resident molecules such as calnexin, calreticulin, ERp57, and protein disulfide isomerase (1)(2)(3)(4). Around 30% of newly synthesized proteins are unable to assemble into their final functional conformation or complex and are degraded (5). Improperly assembled proteins are recognized by the ER quality control machinery and exported from the ER in a process called dislocation and are degraded by the proteasome. The ability of the cell to dispose of misfolded proteins is critical for cell survival (6). The accumulation of misfolded proteins within the ER can result in the activation of an ER stress signaling pathway that aims to restore cellular homeostasis through transcriptional and translational alterations. This may result in upregulation of ER folding chaperones and inhibition of eIF2␣-dependent translation (7,8).
The removal and degradation of aberrant proteins in the ER is a multifaceted process that likely occurs through several steps. First, degradation substrates are recognized by ER sensors and targeted for removal via an aqueous membrane pore referred to as the "dislocon." Ubiquitination of ER degradation substrates is carried out by cytosolic or ER membrane bound E3 ubiquitin ligases. Once exposed to the cytosolic environment, N-linked glycans can be cleaved by the cytosolic enzyme peptide N-glycanase. Finally, the ER degradation substrates are delivered to the proteasome and destroyed.
Model systems have been established in mammalian cells to study the export of misfolded ER substrates. For example, a mutant form of the cystic fibrosis transmembrane conductance regulator, a multimembrane-spanning protein, is targeted for proteasomal destruction (9). Degradation of type I membrane proteins have been investigated using the human cytomegalovirus (HCMV)-encoded type I glycoproteins US2 and US11 that bind directly to nascent MHC class I heavy chains in the ER and catalyze their export and subsequent destruction by the proteasome (10 -12). Unassembled ␣ subunits of the T cell antigen receptor are also exported and degraded (13,14). Proteasome degradation of soluble proteins in the ER has been studied via the export of mutant forms of the ␣1-antitrypsin protein (15)(16)(17). These are only some examples of ER proteins marked for disposal. In fact, ER quality control is important in the folding and maturation of many proteins that could give rise to disease states (18,19).
Dislocation of type I glycoproteins targeted for degradation in mammalian cells has been partially defined as a multistep process using HCMV US2 and US11. Class I heavy chains are highly unstable in US2-and US11-expressing cells, displaying a half-life of 1-2 min, and degradation can be attenuated through the use of proteasome inhibitors (10). Deglycosylation of class I heavy chains by cytosolic peptide N-glycanase occurs promptly upon exposure of the polypeptide to the cytoplasmic environment in cells that have been treated with proteasome inhibitor (10). MHC class I heavy chains accumulate as deglycosylated, polyubiquitinated intermediates before degradation (20,21). In this context class I heavy chains exist as distinct populations after treatment with proteasome inhibitor: glycosylated, ER resident molecules, ubiquitinated membrane-bound species, and deglycosylated cytosolic intermediates. However, the role of ubiquitination in substrate extraction is not fully understood, and at what point during the dislocation reaction ubiquitination occurs remains an outstanding question. The accumulation of dislocation intermediates during proteasome inhibition provides a window through which the steps of dislocation can be delineated. US11 co-opts the host cellular dislocation machinery to degrade MHC class I heavy chains. US11 activates a mild ER stress response to facilitate class I heavy chain degradation (22). The mammalian ER transmembrane protein Derlin-1, a homologue of yeast Der1p, has been implicated in the US11-mediated dislocation of class I heavy chains (23,24) and forms a complex with p97 (the mammalian homologue of yeast Cdc48p) in association with Ufd1⅐Npl4 (23,24). A requirement for Sel1L, the mammalian ortholog of yeast Hrd3p, has been demonstrated for class I degradation in US11-expressing cells (25). The addition of ubiquitin to dislocation substrates is a critical step in ER dislocation, and ubiquitin-conjugated MHC class I heavy chains accumulate in US11-expressing cells (20,21,26). p97 binds to ubiquitinated MHC class I heavy chains and is thought to use its ATP hydrolysis activity to extract membrane-bound substrates (27,28). Additionally, these cellular components of dislocation may be important in CFTR degradation as well (29,30), implying that the pathways utilized by US11 likely reflect those employed by the cell to export endogenous misfolded ER substrates.
The study presented here identifies HCMV US11 as an uncharacterized ER dislocation and proteasome degradation substrate and explores the impact of proteasome inhibition on the dislocation reaction. A population of glycosylated US11 molecules that are modified with ubiquitin accumulate under conditions of proteasome inhibition. Notably, deglycosylated membrane-bound US11 intermediates accumulate after treatment with a proteasome inhibitor, suggesting that proteasome function impacts the final steps of US11 removal from the ER membrane. The findings presented here further define the dislocation reaction and demonstrate that US11 ubiquitination and deglycosylation occur before its extraction from the ER membrane. Ubiquitination is followed by deglycosylation and complete dislocation.
cDNA Constructs-The US11 site-directed cDNA constructs (US11 K200R , US11 Q192L , US11 Q192L,K200R ) were generated from the ligation of two PCR fragments in which one fragment contained the mutated codon (32). All of the constructs were engineered with restriction sites permissive for insertion into pcDNA3.1(ϩ) and subsequent subcloning into the retrovirus vector pLgPW, which contained EGFP as a selection marker.
Retrovirus Transduction-Retrovirus vectors encoding US11 wild-type or mutant constructs were transiently transfected simultaneously with vector containing the VSV-G envelope into gp2-293 cells via a lipid-based protocol to generate pseudotyped Moloney Murine Leukemia Virus amphotropic retrovirus as described previously (32). Enhanced green fluorescent protein-expressing cells were sorted using a Vantage high speed cell sorter (Mount Sinai Flow Cytometry Facility) 3-5 days post-infection.
Endoglycosidase H and peptide N-glycanase sensitivity was determined as per the manufacturer's protocol (New England Biolabs). In brief, polypeptide samples were incubated in 1ϫ denaturing buffer (0.5% SDS, 0.04 M dithiothreitol) followed by the addition of 10ϫ G5 buffer (0.5 M sodium citrate, pH 5.5, for endoglycosidase H) or 10ϫ G7 buffer (0.5 M sodium citrate, pH 7.5, 10% Nonidet P-40, for peptide N-glycanase) and 1000 units of endoglycosidase H or 500 units of peptide N-glycanase. Enzymatic reactions were carried out at 37°C for 1.5 h.
Metabolic Labeling of Cells and Pulse-chase Analysis-U373 and US11 cells were starved in methionine/cysteine-free Dulbecco's modified Eagle's medium for 45 min at 37°C. Metabolic labeling was carried out with 500 mCi of [ 35 S] methionine/cysteine (1200 Ci/mM; PerkinElmer Life Sciences)/ml at 37°C for 10 min. For experiments performed in the presence of proteasome inhibitor, cells were incubated with 50 M ZL 3 VS for 1 h at 37°C. After labeling, the chase was performed in Dulbecco's modified Eagle's medium containing an excess of 25 mM cold methionine and 50 g/ml cyclohexamide for the indicated times. Cells were pelleted at each chase point by centrifugation at 12,000 rpm for 15 s at 4°C and lysed in 1.5 ml of 1ϫ Nonidet P-40 lysis buffer on ice for 15 min with intermittent vortexing. The cell lysates were precleared by the addition of 3 l of normal mouse serum/ml, 3 l of normal rabbit serum/ml and subsequently incubated with 50 l of Pansorbin cells/ml for 1 h at 4°C. Pansorbin cells were removed through centrifugation at 15,000 ϫ g for 2 min, and immunoprecipitations for US11 polypeptides were performed as described above. The immunoprecipitates were separated by SDS-PAGE (12.5%), and the radioactive signal in the polyacrylamide gel was enhanced using Autofluor (National Diagnostics). The dried polyacrylamide gel was exposed to autoradiography film for 3 ( Flow Cytometry Analysis-Quantitative flow cytometry analysis of surface expressed MHC class I molecules in US11 WT and US11 K200R cells was performed using the monoclonal antibody W6/32. The cells were then incubated with Alexa647-conjugated anti-mouse IgG (Molecular Probes), and fluorescence was measured using a Cytomics FC 500 Flow Cytometer (Beckman Coulter). The data were analyzed using Flow Jo software (Tree Star, Inc). The levels of surface class I molecules are represented by plots of normalized cell number (% of cells) versus fluorescence signal. The % of cells was calculated by converting the maximum cell number at the respective peak fluorescence value to 100%.
Subcellular Fractionation-US11 cells were untreated or treated with 5 M ZL 3 VS for 14 h. Cells were detached using trypsin and resuspended in 1ϫ homogenization buffer (100 mM Tris, 150 mM NaCl, 250 mM sucrose, 1.5 mg/ml aprotinin, 1 M leupeptin, and 2 mM phenylmethylsulfonyl fluoride) and mechanically homogenized using a 12-m ball bearing homogenizer (Isobiotec, Hiedelberg, Germany) with 12 passes. Unbroken cells and other debris were pelleted at 15,000 ϫ g for 10 min at 4°C, and the supernatants were treated with 4.5 M urea (final) for 10 min on ice. Supernatants were centrifuged at 120,000 ϫ g for 1 h at 4°C. Pellets from both centrifugation steps were lysed directly in 1ϫ SDS sample buffer, and the 12,000 ϫ g supernatant was concentrated using an Amicon Ultra centrifugal filter (10,000 M r cut-off) at 15,000 ϫ g at 4°C until a final volume of around 100 l. Supernatants and pellets were lysed directly in 1ϫ SDS sample buffer and resolved using SDS-PAGE (12.5%).

US11 Is Degraded in a Proteasomal-dependent Manner-
The degradation of type I membrane glycoproteins is a multistep process that involves both membrane and cytosolic protein complexes. The relationship of the proteasome to the extraction of degradation substrates is not well defined. The current study addresses how proteasome inhibition impacts dislocation of the type I ER membrane protein HCMV US11.
Our initial experiment examined the degradation kinetics of US11 using pulse-chase analysis to determine the half-life of US11 protein. U373 cells and cells that stably express US11 were metabolically labeled with [ 35 S]methionine for 10 min followed by a chase of up to 120 min. US11 was immunoprecipitated from cell lysates using anti-US11 serum at the 0, 30, 60, and 120 min chase points. The immunoprecipitates were resolved using SDS-PAGE and visualized by fluorography. US11 is a 215-amino acid glycoprotein with a predicted molecular mass of ϳ25 kDa. As expected, US11 molecules were not recovered from U373 lysates (Fig. 1A, lanes 1-4). US11 polypeptides recovered from US11 cells decreased throughout the course of the chase (Fig. 1A, lanes 5-8) and were largely absent from the 120-min chase point (Fig. 1A, lane 8). Quantification of the recovered US11 reveals that it has a half-life of around 60 min (Fig. 1B). The instability of US11 molecules throughout the course of the chase demonstrates that US11 was targeted for proteolysis.
The instability of US11 molecules is reminiscent of that observed for other ER degradation targets (10,13) suggesting that US11 may also be targeted for proteasomal degradation. To address this hypothesis, U373 and US11 cells treated with or without the proteasome inhibitor ZL 3 VS (50 M, 1 h) were pulsed with [ 35 S]methionine for 10 min and chased for up to 60 min (Fig. 1C). US11 was recovered from cell lysates at the 0-, 30-, and 60-min chase points using anti-US11 serum. Immunoprecipitates were subjected to SDS-PAGE and fluorography (Fig. 1C). US11 was not recovered from U373 lysates ( Fig. 1C, lanes 1-3 and 7-9). In US11 cells untreated with ZL 3 VS, the amount of US11 decreased throughout the chase times (Fig. 1C, lanes 4 -6) in keeping with our previous findings (Fig. 1A). However, when US11 cells were treated with proteasome inhibitor, relatively equal amounts of US11 were recovered throughout the chase points (Fig. 1C, lanes [11][12]. Notably, a faster migrating species was recovered from the US11 precipitates from the 60-min chase point (Fig.  1C, lane 12, denoted by the asterisk). Treatment with proteasome inhibitors typically results in the accumulation of faster-migrating, deglycosylated class I heavy chains in US11-expressing cells (10). The appearance of a faster-migrating US11 polypeptide upon proteasome inhibition suggests that deglycosylated US11 molecules may accumulate under these conditions. These data demonstrate that US11 was targeted for proteasomal destruction. Degradation of viral proteins may be a strategy evolved by HCMV to prevent a massive ER stress response induced by overexpression of the US11 glycoprotein in the host ER, as US11 has been shown provoke a mild ER stress response (22).

US11 Degradation Intermediates
Accumulate upon Proteasome Inhibition-A major difference between US11 degradation and US11-mediated class I degradation is the kinetics of the reaction. In US11expressing cells, class I heavy chains are degraded with a halflife of 1-2 min (10), whereas the half-life of US11 is around 60 min (Fig. 1). The appearance of degradation intermediates is likely dependent upon the efficiency of proteasome inhibition and the rate of dislocation. To determine conditions that allow for accumulation of US11 degradation intermediates, US11 was immunoprecipitated from US11 cells that had been treated with ZL 3 VS for up to 12 h. Treatment with a minute amount of proteasome inhibitor for a prolonged period of time was not toxic to cells (data not shown). Immunoprecipitates were subsequently separated using SDS-PAGE and visualized by immunoblot analysis using anti-US11 serum ( Fig. 2A). As expected, US11 was not recovered from U373 cells ( Fig. 2A, lanes 1-4). An increasing amount of glycosylated US11 was recovered from US11 cells treated with proteasome inhibitor, with nearly equal amounts of glycosylated US11 recovered from cells treated for 6 and 12 h ( Fig.  2A, lanes 6-8). Interestingly, a faster migrating polypeptide was also recovered from US11 cells treated with ZL 3 VS for 6 and 12 h ( Fig. 2A, lanes 7 and 8). Therefore, accumulation of glycosylated and deglycosylated US11 species after long periods of proteasome inhibition is representative of US11 degradation intermediates.
Notably, high molecular weight US11-reactive polypeptides were observed after 3 h of proteasomal inhibition ( Fig. 2A, lane 6, denoted by arrows). These molecules increased in molecular weight during the time course ( Fig. 2A, lanes 6 -8, denoted by arrow) and are likely US11 molecules modified with ubiquitin. Indeed, slower migrating US11 species are modified with ubiquitin, as evidenced by immunoblot of US11 immunoprecipitates using anti-ubiquitin monoclonal antibody (Fig. 2B, compare lanes 3 and 4). In addition, US11 recovered from ZL 3 VS-treated cells followed by immunoblot using polyclonal anti-ubiquitin antibody shows a discrete pattern of ubiquitin-reactive bands similar to that observed in the anti-US11 immunoblot (Fig. 2C). It could be the case that the polyclonal antibody recognizes polyubiquitin chains more readily than monoubiquitin modifications and, thus, gives rise to the observed distinct high molecular weight US11 species (Fig. 2C,  lane 2). Collectively, these data suggest that longer periods of proteasome inhibition are required to observe US11 degradation intermediates.
Proteasome Inhibition Induces US11-deglycosylated Intermediates-US11 is an ER resident glycoprotein that possesses a single N-linked glycan at amino acid 73 (34,35). Cleavage of the N-linked glycan by the cytosolic enzyme peptide N-glycanase during dislocation would result in a faster migrating polypeptide. Is the faster-migrating US11 species observed during proteasome inhibition ( Fig. 2A) a deglycosylated degradation intermediate? To address this question, US11 molecules were treated with endoglycosidase H, a bacterial enzyme that cleaves high mannose asparagine-linked residues from glycoproteins (36). U373 and US11 cells untreated or treated with ZL 3 VS (5 M, 14 h) were lysed in 2% SDS. Half of the lysates were subjected to endoglycosidase H digestion. Cell lysates were then resolved via SDS-PAGE and US11, class I heavy chains, and GAPDH proteins were visualized by an immunoblot using their cognate antibodies (Fig. 3A). Immunoblot using anti-GAPDH antibody confirms that equal levels of protein were loaded onto the gel (Fig. 3A, lanes 17-24). US11 was not observed in U373 cell lysates (Fig. 3A, lanes 1-4). In lysates from US11-expressing cells, a glycosylated form of US11 (Fig. 3A, lane 5) was sensitive to endoglycosidase H digestion resulting in faster migrating US11-reactive polypeptides (Fig. 3A, lane 6). Both the glycosylated form of US11 and the faster-migrating species were observed in US11 cells treated with ZL 3 VS (Fig. 3A, lane  7). After incubation with endoglycosidase H, glycosylated US11 has a mobility similar to the faster migrating US11-reactive species (Fig. 3A, lane 8), suggesting that the faster migrating species was most likely a deglycosylated US11 intermediate. These data demonstrate that the accumulation of a faster migrating US11 species can be used as a diagnostic marker for the dislocation reaction.
To verify the effectiveness of proteasome inhibition, MHC class I heavy chains were visualized by immunoblot using antiheavy chain serum (Fig. 3A, lanes 9 -16). Class I heavy chains traffic through the ER and Golgi, where their N-linked glycans become further modified by Golgi-resident glycosylating enzymes (37,38). Thus, class I heavy chains were unaffected by endoglycosidase H (Fig. 3A, lanes 9 and 10). When U373 cells were treated with ZL 3 VS, a small fraction of class I heavy chains were retained in the ER due to the inhibition of proteasomal peptide processing. This is evidenced by the susceptibility of a minute amount of class I heavy chains to endoglycosidase H (Fig. 3A, lane 12). However, the majority of class I heavy chains observed during steady state in U373 cells has trafficked past the ER, rendering them resistant to endoglycosidase H treat-FIGURE 2. US11 degradation intermediates accumulate upon proteasome inhibition. A, US11 proteins were recovered from US11 cells treated with ZL 3 VS (2.5 M) for 0, 3, 6, or 12 h using anti-US11 serum. The precipitates were resolved by SDS-PAGE followed by immunoblot using anti-US11 serum. U373 cells were utilized as a control. Glycosylated (ϩCHO), faster migrating (*), and high molecular weight US11-reactive species (arrows) are indicated. B, U373 and US11 cells were treated with ZL 3 VS (2.5 M, 12 h) and lysed in 1% SDS, and US11 proteins were immunoprecipitated using anti-US11 serum. Recovered proteins were separated via SDS-PAGE and visualized by immunoblot using mouse monoclonal anti-ubiquitin antibody. High molecular weight ubiquitin-reactive polypeptides are indicated. C, US11 cells were treated with ZL 3 VS (2.5 M, 14 h), lysed and immunoprecipitated as in C, and immunoblotted using rabbit polyclonal anti-ubiquitin antibody. Immunoglobulin heavy chains (Ig HC) are noted. ment (Fig. 3A, lanes 10 and 12). In contrast, class I heavy chains were degraded in a US11-dependent manner and were, therefore, not observed in US11-expressing cells (Fig. 3A, lanes 13  and 14). When proteasome activity was impaired by treatment with ZL 3 VS, faster migrating endoglycosidase H-resistant class I heavy chains were observed (Fig. 3A, lane 16). These results indicate that, as expected, deglycosylated class I heavy chains accumulated upon proteasome inhibition of US11-expressing cells.
A one-dimensional isoelectric focusing experiment was performed in conjunction with glycosidase treatment to confirm that the US11 species lacking a glycan were in fact deglycosylated forms of US11 (Fig. 3B). Cleavage of glycoproteins by peptide N-glycanase results in the conversion of asparagine to aspartic acid (10,11), whereas endoglycosidase H treatment does not change the amino acid composition of the polypeptide.
Membrane fractions from US11 cells untreated (Fig. 3B,  lanes 1-3) or treated (Fig. 3B, lanes 4 -6) with ZL 3 VS (5 M, 14 h) were lysed in 0.5% SDS. Lysates were undigested (Fig. 3B,  lanes 1 and 4) or digested with endoglycosidase H (Fig. 3B, lanes  2 and 5) or peptide N-glycanase (Fig. 3B, lanes 3 and 6). US11 was recovered using anti-US11 serum, and the immunoprecipitates were resolved using a one-dimensional isoelectric focusing polyacrylamide gel. In a similar experiment, US11 cells untreated with ZL 3 VS were lysed in 0.5% SDS, and US11 polypeptides were subjected to digestion with endoglycosidase H or peptide N-glycanase and resolved by SDS-PAGE (Fig. 3C). The US11 polypeptides were subjected to immunoblot analysis using anti-US11 serum (Fig. 3, B and C). The mature glycosylated US11 protein migrates as a single polypeptide with a predicted pI of ϳ5.5 (Fig. 3B, lane 1) and molecular mass of around 27 kDa (Fig. 3C, lanes 3 and 5). Treatment of US11 cell lysates with endoglycosidase H did not alter the pI or migration of glycosylated US11 on IEF gel (Fig. 3B, lane 2) but caused US11 to migrate as a faster migrating deglycosylated intermediate on SDS-polyacrylamide gel (Fig. 3C, lane 6). On the other hand, US11 recovered from peptide N-glycanase-treated lysates migrated faster on an IEF gel, as its pI was changed to ϳ5.3 (Fig. 3B, lane 3). Peptide N-glycanase-treated US11 polypeptides migrated as deglycosylated species when resolved by SDS-PAGE (Fig. 3C, lane 7). US11 molecules from ZL 3 VS-treated cells migrate as two discrete bands on IEF that correspond to glycosylated and deglycosylated versions of the molecule (Fig. 3, B, lane 4, C, lane 4). Treatment with endoglycosidase H does not affect the migration of US11 polypeptides by IEF (Fig. 3B, lane 5), whereas US11 migrates as a faster migrating species when resolved by SDS-PAGE (Fig. 3C, lane 6). US11 recovered from the sample treated with peptide N-glycanase migrates at the same position as the deglycosylated US11 intermediate under both electrophoretic conditions (Fig. 3, B, lane 6, C, lane 7). Taken together, these results demonstrate that US11 was deglycosylated by cytosolic enzyme peptide N-glycanase as it was exported from the ER to the cytosol during the dislocation reaction.

A US11 Mutant That Is Unable to Mediate Class I Heavy Chain Destruction Is Targeted for Proteasomal Degradation-Because
US11 is targeted for proteasomal degradation in a manner similar to class I heavy chains, it is possible that that US11 and heavy chains are destroyed in a co-degradational manner. To examine this possibility, a US11 mutant (US11 Q192L ) that is incapable of targeting class I molecules for proteolysis (39) was generated, and its stability was analyzed in the presence of proteasome inhibitor. U373, US11 WT , and US11 Q192L cells were treated up to 12 h with ZL 3 VS, lysed in 1% SDS, and immuno-  1 and 4), treated with endoglycosidase H (lanes 2 and 5) or peptide N-glycanase (lanes 3 and 6), and resolved on a one-dimensional isoelectric focusing gel. The glycosylated (ϩCHO) and deglycosylated (ϪCHO) US11 polypeptides were visualized by anti-US11 immunoblot. C, US11 was recovered from U373 and US11 cells untreated with ZL 3 VS, and the recovered polypeptides were untreated (lane 5) or treated with endoglycosidase H (lane 6) or peptide N-glycanase (lane 7). The immunoprecipitates and SDS cell lysates were resolved by SDS-PAGE, and proteins were visualized by anti-US11 immunoblot. precipitated with anti-US11 serum. The precipitates were subsequently analyzed by immunoblot analysis using anti-US11 serum (Fig. 4). As expected, US11 was not recovered from U373 cells (Fig. 4, lanes 1 and 2). Increased amounts of glycosylated US11 WT and deglycosylated intermediate were recovered from lysates of cells treated with ZL 3 VS (Fig. 4, lanes 3-5). Similarly, increased amounts of glycosylated US11 Q192L molecules were recovered from cells at the 0-, 6-, and 12-h time points (Fig. 4,  lanes 6 -8). Deglycosylated US11 Q192L was recovered at 6 and 12 h of inhibition (Fig. 4, lanes 7-8), similar to deglycosylated US11 WT (Fig. 4, lanes 4 -5). Slightly less amounts of deglycosylated US11 Q192L were recovered as compared with US11 WT (Fig. 5C, lanes 7-8 and 4 -5). This suggests that the kinetics of US11 Q192L degradation may be delayed compared with wild-type US11, perhaps due to prolonged association with class I heavy chains in the ER (data not shown (39)).
These data indicate that class I heavy chain degradation is not a prerequisite for US11 degradation. Lysine Residue in the Cytoplasmic Tail of US11 Is Not Required to Initiate Its Degradation-Is ubiquitination of US11 required to initiate its extraction from the ER membrane? US11 possesses a single lysine residue in its cytoplasmic region (amino acids 200) that could be a target for the cytosolic ubiquitin-conjugating machinery. To investigate whether ubiquitination of lysine 200 is critical for the initiation of US11 degradation, we generated a site-directed mutant where lysine 200 was mutated to arginine (US11 K200R ). Before examining US11 K200R stability, its ability to down-regulate the surface expression of class I heavy chains was examined by flow cytometry using an antibody directed against properly folded class I molecules (W6/ 32, Fig. 5A). US11 K200R -expressing cells (Fig. 5A, thin solid line) reduced cell surface class I molecules to levels similar to wild-type US11 (Fig. 5A, dashed line) when compared with class I expression in U373 cells (Fig. 5A, shaded peak). A further confirmation of class I degradation in US11 K200R cells was observed by immunoblot analysis, where deglycosylated class I heavy chains accumulated in ZL 3 VStreated US11 K200R cells (data not shown). These results demonstrate that US11 K200R is functionally active  U373 (lanes 1 and 2), US11 WT (lanes 3-5), and US11 Q192L (lanes 6 -8) cells were treated with ZL 3 VS (5 M) for 0, 6, or 12 h followed by lysis in 1% SDS and immunoprecipitation with anti-US11 serum. The recovered proteins were separated by SDS-PAGE and subjected to anti-US11 immunoblot. Glycosylated (ϩCHO) and deglycosylated (ϪCHO) US11 proteins are indicated. and can down-regulate class I molecules similar to its wild-type counterpart.
We next examined whether US11 K200R itself was targeted for proteasomal degradation. U373, US11, and US11 K200R cells were treated with ZL 3 VS for up to 6 h. US11 was recovered from SDS cell lysates using anti-US11 serum, and the precipitates were analyzed by an immunoblot using anti-US11 serum (Fig.  5B). As expected, US11 was not recovered from U373 cells (Fig.  5B, lanes 1 and 2), whereas increasing amounts of glycosylated and deglycosylated US11 WT molecules were recovered from cells treated with proteasome inhibitor (Fig. 5B, lanes 3-5). Significantly, US11 K200R was also immunoprecipitated at increasing levels at 3 and 6 h after ZL 3 VS treatment (Fig. 5B, lanes 7  and 8). In addition, a deglycosylated US11 K200R intermediate was recovered from cells treated for 6 h, similar to US11 WT (Fig.  5B, lane 8). These results indicate that US11 K200R undergoes proteasomal degradation in a manner similar to wild-type US11. More importantly, the data reveal that the possible ubiquitination of this substrate on lysine 200 was not required to initiate US11 degradation.
The proteasomal destruction of US11 K200R does not exclude the possibility that ubiquitination of class I heavy chain complexed with US11 is the catalyst needed to target US11 K200R for destruction. To examine whether US11 K200R degradation is independent of class I heavy chain degradation, cells that express a US11 double mutant (US11 Q192L,K200R ) that is incapable of inducing class I degradation (data not shown) were generated. U373, US11 WT , and US11 Q192L,K200R cells were treated with ZL 3 VS (5 M, 14 h) followed by lysis in 1% SDS. Whole cell lysates from equal number of cells were resolved by SDS-PAGE followed by immunoblot analysis with anti-US11 serum (Fig. 5C). As expected, US11 was not observed in U373 cell lysates (Fig. 5C, lanes 1 and 2), and increased amounts of glycosylated as well as deglycosylated US11 molecules accumulated after proteasome inhibition (Fig. 5C, lane 4). Similarly, an increased amount of glycosylated US11 Q192L,K200R as well as deglycosylated species were also observed in lysates from US11 Q192L,K200R cells treated with proteasome inhibitor (Fig.  5C, lane 6). These results indicate that US11 Q192L,K200R was targeted for proteasomal degradation and further confirms the initiation of US11 degradation was likely independent of ubiquitin modification of lysine 200 or class I degradation.
Proteasomal Inhibition Causes the Accumulation of Ubiquitinated, Glycosylated US11-High molecular weight US11-reactive species were recovered from ZL 3 VS-treated cells (Fig. 2). It is not clear whether these US11 molecules are derived from glycosylated or deglycosylated US11 species, as ubiquitination of lysine does not appear to participate in initiation of US11 dislocation (Fig. 5). To address this, US11 was recovered from cells treated with ZL 3 VS (2.5 M, 14 h) using anti-US11 serum. The immunoprecipitates were subsequently untreated or treated with endoglycosidase H and subjected to immunoblot analysis using anti-US11 serum (Fig. 6A). As expected, ubiquitinated US11 species were not recovered from untreated cells (Fig. 6A, lanes 1 and 2), whereas high molecular weight ubiquitin-modified US11 polypeptides were recovered from ZL 3 VS-treated cells (Fig. 6A, lane 3). These US11 species migrated at 152, 130, 98, 58, and 42 kDa and were sensitive to endoglycosidase H, as evidenced by their faster migration pattern after treatment (Fig. 6A, lane 4). The molecular weight of a single ubiquitin molecule is ϳ8 kDa (40), whereas that of glycosylated US11 is ϳ27 kDa (10). Therefore, judging by the mobility of the slower migrating glycosylated US11 molecules, US11 proteins were modified with multiple ubiquitins (Fig. 6A,  lane 3). These data demonstrate that a population of glycosylated US11 was modified with ubiquitin while still integrated in the ER membrane.
Lysine 200 is not required to initiate US11 WT dislocation (Fig. 5, A and B). Does the ubiquitination of glycosylated US11 WT (Fig. 6A) occur on luminal lysine residues that have become exposed to the cytosol during dislocation. To address this question, US11 K200R and US11 Q192L,K200R molecules were recovered from ZL 3 VS-treated cells (2.5 M, 14 h), digested with endoglycosidase H, and subjected to immunoblot analysis using anti-US11 serum (Fig. 6B). High molecular weight forms of US11 K200R (Fig. 6B, lanes 3 and 4) and US11 Q192LK,200R (Fig.  6B, lanes 7 and 8) accumulated in the presence of ZL 3 VS. These high molecular weight US11 K200R and US11 Q192L,K200R species were also sensitive to endoglycosidase H treatment (Fig. 6B,  lanes 4 and 8). Notably, US11 K200R and US11 Q192L,K200R that migrate at about 58 kDa were absent from the immunoblot (Fig.  6B, lanes 3 and 4 and lanes 7 and 8). It is possible that the presence of wild-type US11 that migrates at 58 kDa (Fig. 6A,  lane 3) represents a US11 molecule that has been ubiquitinated on lysine 200. The accumulation of glycosylated and ubiquitinated US11 K200R and US11 Q192L,K200R that lack lysines in the cytoplasmic tail indicates that these molecules were modified elsewhere on the polypeptide. This finding suggests that the luminal region of US11 was dislocated to the cytosol and modified with ubiquitin before its complete deglycosylation and extraction from the ER membrane. Together these data demonstrate that partially dislocated US11 molecules accumulate in the ER when proteasomal activity was impaired.
Proteasome Inhibition Blocks a Latter Step of the Dislocation Reaction-The disposal of ER substrates is a multistep process that involves the dislocation of substrates across the ER membrane followed by their proteasomal degradation in the cytosol. Our findings that proteasome inhibition causes an accumulation of both glycosylated, ubiquitinated US11 polypeptides and deglycosylated degradation intermediates suggest that the dislocation and degradation of US11 are tightly coupled processes ( Figs. 2A, 3B, and 6A). Therefore, proteasome inhibition allows for the accumulation of diverse dislocation intermediates. To further characterize the deglycosylated US11 intermediates, we examined the membrane topology of glycosylated and deglycosylated US11 proteins upon proteasome inhibition using urea treatment. Incubation of cell-derived microsomes with urea causes the release of non-integral peripheral proteins from the membrane bilayer (41). In contrast, a polypeptide imbedded within the bilayer is insensitive to urea treatment. Therefore, this protocol can be used to determine whether the US11 polypeptide is integrated within the membrane or has been fully extracted from the membrane bilayer.
To that end, homogenates from US11 cells untreated or treated with proteasome inhibitor were generated using a ballbearing homogenizer followed by a 15,000 ϫ g centrifugation step to remove nuclei, unbroken cells, and large membrane sheets (42). The 15,000 ϫ g supernatant was untreated or treated with 4.5 M urea to release any peripheral proteins from the microsomal membranes. The supernatant was then clarified by a 120,000 ϫ g centrifugation step to isolate the membrane fraction from the cytosolic contents. The 15,000 ϫ g pellet, 120,000 ϫ g pellet, and 120,000 ϫ g supernatant were subsequently resolved by SDS-PAGE and subjected to immunoblot analysis using antibodies directed against US11, TRAM1, and p97. TRAM1 is a multimembrane-spanning ER protein that plays a role in protein processing (43) and can be used as a membrane protein control. Conversely, p97 is a soluble cytosolic protein that associates with the ER membrane (44) and can verify the effectiveness of urea treatment.
The membrane proteins US11 and TRAM1 were found in the membrane fraction independent of urea treatment (Fig. 7A,  lanes 6, 7, and 9 and lanes 11, 12, and 14, respectively). Examination of the p97 distribution reveals that some but not all of the protein associates with the membrane fraction, as evidenced by its localization in both the 120,000 ϫ g pellet and 120,000 ϫ g supernatant (Fig. 7A, lanes 2 and 3). However, treatment with 4.5 M urea caused p97 to be completely released from the microsomal surface into the cytosolic fraction (Fig.  7A, compare lanes 2 and 3 versus 4 and 5). These results confirm that urea treatment allows for the liberation of non-integral, peripheral proteins from the microsomes.
In US11 cells treated with proteasome inhibitor, TRAM1 was found only in the membrane fractions independent of urea treatment (Fig. 7B, lanes 6, 7, and 9). Glycosylated US11 was also observed exclusively in the pellet fractions (Fig. 7B,  lanes 11, 12, and 14), a result consistent with the localization of a membrane protein. On the other hand, a significant FIGURE 7. Deglycosylated US11 remains integrated within the ER membrane upon proteasomal inhibition. Subcellular homogenates from US11 cells untreated (A) or treated (B) with ZL 3 VS were subjected to 15,000 ϫ g centrifugation. The supernatant was treated without or with 4.5 M urea and centrifuged at 120,000 ϫ g. Cell pellets from the 15,000 and 120,000 ϫ g centrifugations and the 120,000 ϫ g supernatant were resolved by SDS-PAGE and subjected to immunoblot analysis using anti-p97 (lanes 1-5), anti-TRAM1 (lanes 6 -10), and anti-US11 (lanes [11][12][13][14][15] serum. Glycosylated (ϩCHO) and deglycosylated (ϪCHO) US11, p97, and TRAM1 proteins are indicated. amount of the deglycosylated US11 protein localized as a membrane protein (Fig. 7B, lanes 11, 12, and 14), whereas only a minute amount of deglycosylated US11 was observed in the cytosolic fraction (Fig. 7B, lane 13). Interestingly, treatment with urea caused only a small increase in the amount of deglycosylated US11 released as a cytosolic polypeptide (Fig. 7B, lane 15). The efficacy of urea treatment was confirmed by the exclusive localization of p97 in the 120,000 ϫ g supernatant (Fig. 7B, lane 5). Together, these results indicate that in the absence of proteasome function, deglycosylated US11 polypeptides accumulate in the ER membrane and are unable to be fully released into the cytosol. These findings demonstrate that proteasome inhibition induces the accumulation of unique degradation intermediates and supports a model in which proteasome function is involved in the final steps of extracting US11 from the ER membrane, downstream of peptide N-glycanase activity and before release into the cytosol and degradation.

DISCUSSION
Proteasomal degradation of HCMV US11 can be utilized as a new tool to define dislocation and degradation of ER type I glycoproteins. Inhibition of US11 degradation through interference of proteasome function resulted in the accumulation of both glycosylated, ubiquitinated and membrane-integrated deglycosylated US11 molecules. The formation of these distinct degradation intermediates further clarifies the mechanism of substrate extraction into the cytosol and allows for the proposal of an expanded model of ER dislocation (Fig. 8).
The US11 polypeptide is a 215-residue type I glycoprotein with a predicted luminal domain (amino acids 1-180), transmembrane domain (amino acids 181-199), and cytoplasmic tail (amino acids 200 -215). The protein is modified with an N-linked glycan (amino acids 73) and contains four lysine residues at positions 109, 123, 179, and 200 (35). The initiation of US11 degradation proceeds with the partial dislocation of US11 across the ER bilayer; however, the ER proteins or complexes that mediate the partial extraction of US11 have not yet been defined. Molecular chaperones and transmembrane proteins such as Derlin-1 might mediate in part the dislocation of US11 across the membrane bilayer. Once partially dislocated, glycosylated US11 polypeptides are ubiquitinated (Figs. 6 and Fig. 8,  step 2).
The lysine at position 200 is a likely candidate for ubiquitination. However, the US11 mutant US11 K200R continues to be targeted for degradation and can be modified with ubiquitin by a cellular E3 ligase. Substrate cysteine (45), threonine, and serine residues (46) can be targets for viral E3 ligases MIR1 and mK3. US11 possesses a serine residue in the cytoplasmic tail at position. The possibility exists that this serine residue could accept ubiquitin moieties during dislocation; however, these non-traditional ubiquitination events have thus far only been observed for viral E3 ligases. We maintain that US11 is recognized as a degradation substrate and disposed of by cellular machinery in a manner similar to a misfolded protein. Therefore, we favor a model in which the US11 luminal domain is partially dislocated before ubiquitination and ubiquitinated on FIGURE 8. Dislocation of US11 is a multi-step process. 1, US11 is a type I glycoprotein integrated into the ER membrane. 2, glycosylated US11 is partially dislocated in two possible configurations (A or B) followed by ubiquitination of exposed lysine residues. The dislocation of US11 is probably mediated by a combination of ER factors. 3, ubiquitinated US11 would interact with p97/Npl4/Udfd1 followed by the complete extraction of the luminal domain. The N-linked glycan is removed upon exposure to the cytosol by peptide N-glycanase. 4, ubiquitin molecules would then be cleaved by de-ubiquitinating enzymes while still integrated in the ER membrane. 5, finally, US11 is completely extracted from the ER membrane and degraded by the proteasome. lysine residues by a cellular E3 ligase. Because the ubiquitinated US11 species maintain an N-linked glycan, the topology of glycosylated/ubiquitinated US11 molecules must consider that the region around the N-linked glycan is within the ER lumen.
There are three possibilities how ubiquitinated/glycosylated US11 could occur. 1) US11 could be extracted through the membrane in a C-terminal manner. This is most likely not the case because deglycosylated US11 remains integrated within the ER membrane after treatment with urea (Fig. 7B). This suggests that the transmembrane region of US11 is the last domain to be dislocated out of the ER. 2) The US11 luminal domain between amino acids 73 and the transmembrane region are dislocated across the membrane into the cytosol. This configuration is supported by the presence of multiple lysines that can be ubiquitinated.
3) The N terminus of US11 is dislocated across the membrane, and the N-terminal methionine residue is modified with ubiquitin (47). Even though it is not possible to state with certainty the exact topology of the glycosylated/ubiquitinated US11 molecules, the results support the hypothesis that the luminal domain is initially dislocated across the ER membrane in a ubiquitin-independent manner. The model of partial dislocation and subsequent ubiquitination of US11 is supported by the accumulation of membrane-bound ubiquitinated CD4 in the context of HIV-1 Vpu in yeast (48).
The role of substrate ubiquitination in dislocation is poorly understood. Ubiquitin conjugation does not appear to be critical for initiation of the process, but the data presented here support a model in which ubiquitination of the substrate is important for the latter steps of substrate extraction before deglycosylation and proteasome degradation. Attachment of ubiquitin may participate in the recruitment of dislocation machinery, including p97 or perhaps the proteasome itself.
Based upon what is known about US11-mediated degradation of MHC class I heavy chains, we predict that ubiquitinated US11 will most likely then be targeted for dislocation via the p97⅐Ufd1⅐Npl4 complex. This prediction remains to be tested, however. The luminal domain of the glycosylated/ubiquitinated US11 molecule would most likely then be completely extracted through the bilayer. Upon exposure of the N-linked glycan to the cytoplasm, peptide N-glycanase removes the glycan, generating a deglycosylated and ubiquitinated US11 species (Fig. 8, step 3). Once engaged with the dislocation machinery, ubiquitins may be removed by de-ubiquitin hydrolases in preparation for proteasome degradation (49). The accumulation of membrane-integrated deglycosylated US11 intermediates upon proteasome inhibition suggests that the transmembrane domain is dislocated in one of the final steps of dislocation and its dislocation may indirectly require proteolytic activity of the proteasome.
The 19 S proteasome cap may be important in substrate extraction from the ER (50). An interesting question is whether the 19 S complex is involved in recognition of ubiquitinated dislocation substrates as they exit the ER, a function that could be separate and upstream from subsequent proteolytic activity. The 19 S cap of the proteasome is subdivided into a base that is composed of at least 6 AAA ATPase subunits (Rpt1-6) and a lid comprised of non-ATPase subunits (for review, see Ref. 51). The 19 S base also contains two non-ATPase subunits, Rpn1 and Rpn2, that bind ubiquitin-like protein domains (52,53). Ubiquitin binding capabilities have also been demonstrated for the non-ATPase subunit Rpn10 (54,55) and the ATPase base member Rpt5 (56). De-ubiquitination of substrates must occur before proteolysis, and this activity has been described for the non-ATPase lid subunit Rpn11 (57) and the p97-associated non-proteasomal protein ataxin-3 (58,59). The non-proteasomal protein Rad23 has been proposed to play a role in delivering ubiquitinated substrates to the proteasome, as it possesses both ubiquitin binding and ubiquitin-like domains (60,61). Similarly, the ER transmembrane protein Herp could potentially interact with and recruit the proteasome to the ER via its N-terminal ubiquitin-like domain (62)(63)(64). p97 binding to ubiquitinated proteins may play a role in extracting substrates from the ER, but the dynamics between p97 and the proteasome are not clear. These proteins and complexes may be important for US11 degradation, but the role of these proteins in US11 destruction remains to be determined.
The data presented here suggests that the proteasome may potentially be involved in the latter steps of extraction of ER degradation substrates. Cellular proteasomes are ubiquitous and can bind to the Sec61 translocation channel of the ER (65). Proteasome function has been implicated in the removal of T cell receptor subunits TCR␣ and CD3-␦ (66) and unassembled chains of IgM molecules (67), indicating that the proteasome participates in the dislocation of a wide range of cellular substrates. The data presented here provide more insight into the mechanism of type I glycoprotein dislocation through the identification of novel degradation intermediates. The exact nature of the activity of the proteasome in this process remains to be further investigated.