Reglucosylation of N-Linked Glycans Is Critical for Calnexin Assembly with T Cell Receptor (TCR) α Proteins but Not TCRβ Proteins

Association of calnexin with newly synthesized glycoproteins involves recognition of monoglucosylated glycans, generated in the endoplasmic reticulum via initial removal of two glucose (Glc) residues from immature glycan chains by glucosidase enzymes (Glc trimming), or addition of a single Glc residue to fully trimmed glycans by glucosyltransferase enzymes (reglucosylation). While it has been established that creation of monoglucosylated glycans is important for chaperone binding, it is unknown if most proteins require both deglucosylation and reglucosylation for calnexin assembly or if initial Glc trimming is sufficient. Here, we studied the deglucosylation and reglucosylation of two related glycoproteins, the α and β subunits of the T cell receptor (TCR) complex, and their assembly with calnexin in BW thymoma cells. Our data demonstrate that TCRα/β glycoproteins undergo multiple cycles of Glc removal and addition within the endoplasmic reticulum and that numerous reglucosylated proteins assemble with calnexin, including TCRα/β glycoproteins. Importantly, the current study shows that TCRβ proteins, but not TCRα proteins, effectively associate with calnexin under conditions of functional Glc trimming but impaired reglucosylation. These data demonstrate that reglucosylated proteins associate with lectin-like chaperones in vivo and provide evidence that reglucosylation is of differential importance for the association of individual, indeed similar, glycoproteins with calnexin.

Upon insertion into the lumen of the endoplasmic reticulum (ER), 1 many polypeptides are modified by addition of oligosaccharide chains to asparagine residues, a cellular process termed N-linked glycosylation (1). Immature glycan chains on nascent polypeptides have the structure Glc 3 Man 9 GlcNAc 2 (Glc ϭ glucose; Man ϭ mannose; GlcNAc ϭ N-acetyl glucosamine) and are initially processed by the sequential action of two ER enzymes, glucosidase I and glucosidase II (2). Within the last few years, it has become apparent that processing of Glc residues, or Glc trimming, is an important step in the associ-ation of many glycoproteins with the molecular chaperone calnexin (3)(4)(5)(6)(7) and, more recently, calreticulin (8 -10). Calnexin and calreticulin are transmembrane and lumenal ER proteins, respectively, that are two known members of a family of endogenous lectin-like proteins that recognize partially trimmed, monoglucosylated (Glc 1 Man 9 GlcNAc 2 ) glycan chains on newly synthesized glycoproteins (8,9,11).
Stable interaction of calnexin and calreticulin with newly synthesized glycoproteins is believed to occur by a two-step process involving initial binding of monoglucosylated glycan chains, followed by protein:protein interactions, which stabilize these associations (12,13). Monoglucosylated Glc 1 Man 9 -GlcNAc 2 glycan chains may be generated via two mechanisms: cotranslational processing of immature Glc 3 Man 9 GlcNAc 2 glycans by glucosidase I and glucosidase II enzymes (the Glc trimming pathway), or reglucosylation of fully trimmed Man 8 -9 GlcNAc 2 glycans by UDP-glucose:glycoprotein-dependent glucosyltransferase enzymes (the reglucosylation or salvage pathway) (4, 14 -17). Recent studies demonstrate that glucosyltransferase effectively reglucosylates denatured but not native glycoproteins in vitro (18 -20), suggesting that reglucosylation plays an important role in the ER quality control system. Indeed, it has been proposed that glucosyltransferase functions as a major sensor for incompletely folded proteins in the ER and that glucosylation of nascent proteins stops upon adoption of the correct conformation (16,17,19).
Studies using cell-free assays have determined that the size of the glycan chain, e.g. the Man core, is a critical factor in their modification by glucosyltransferase enzymes; glycopeptides containing Man 8 -9 GlcNAc 2 glycans are reglucosylated much more efficiently than those having Man 7 GlcNAc 2 glycans, which are poorly recognized by glucosyltransferase (16). In contrast, although Man core size is important for interaction of oligosaccharide chains with calnexin (which associates preferentially with Glc 1 Man 9 GlcNAc 2 glycans), it is not crucial, as calnexin also binds smaller Glc 1 Man 5-7 GlcNAc 2 species with significant efficiency (12). Spiro and colleagues (10) recently reported similar findings regarding calreticulin association with glycans. Like calnexin and calreticulin chaperones, glucosyltransferase is believed to interact with both glycan and protein domains on incompletely folded proteins; this interaction, interestingly, involves recognition of hydrophobic stretches on the polypeptide backbone and the innermost Glc-NAc residue on the oligosaccharide chain (19). Hebert et al. (15) recently provided evidence that viral hemagglutinin glycoproteins synthesized in cell-free systems are processed by glucosidase I and II and glucosyltransferase enzymes and interact with calnexin. However, because no inhibitors of glucosyltransferase currently exist, the relative importance of the reglucosylation pathway for calnexin assembly remains to be determined. Indeed, it has never been demonstrated that reglucosylated proteins associate with lectin-like chaperones in intact cells of any type (21).
Glc trimming and calnexin association are initial molecular events in the translation of T cell antigen receptor (TCR) proteins, occurring coincident with or immediately after their translocation into the ER and preceding their assembly into multisubunit protein complexes (14,22). Clonotypic TCR␣ and TCR␤ polypeptides are members of the immunoglobulin gene superfamily and share several structural features, including conserved cysteine residues and post-translational modification by a similar number of N-linked glycan chains (6,23). Interestingly, oligosaccharide processing and calnexin association is critical for stabilizing nascent TCR␣ proteins within the ER, but is not required for the stability of analogous TCR␤ proteins (6). This apparent paradox may relate to the suggestion that, unlike the TCR␤ transmembrane domain, the transmembrane region of TCR␣ does not function efficiently as a stop transfer region within the lipid bilayer of the ER and may rely on calnexin association for membrane stabilization (6,24).
To further our understanding of the role of glycan processing and chaperone association in the quality control system of the ER, we studied the reglucosylation of TCR␣/␤ proteins and their assembly with calnexin in BW thymoma cells. These data demonstrate that TCR␣/␤ proteins undergo cycles of deglucosylation and reglucosylation in the ER and show that numerous reglucosylated proteins assemble with calnexin in intact cells, including TCR␣/␤ proteins. Finally, the results in the current study indicate that reglucosylation is critical for calnexin assembly with TCR␣ proteins but not TCR␤ proteins. In experiments using Msn, cells were cultured overnight in medium containing 10 mM Msn at 37°C in 5% CO 2 ; cell viability was identical in medium-and Msn-treated cell cultures (data not shown). In experiments using both Msn and Cas, Cas was included (100 g/ml final concentration) during the final 3 h of culture.

Deglucosylation and Reglucosylation of TCR␣/␤ Proteins in
BW Cells-To study reglucosylation of TCR␣/␤ proteins in BW cells, [ 3 H]galactose was utilized as a radioactive tracer of Glc residues on oligosaccharide chains (30,31). As illustrated in Fig. 1 the ER ( Fig. 1) (1,32,33). In our studies, radiolabeling of TCR␣/␤ proteins with [ 3 H]galactose was restricted to newly translated and pre-existent proteins localized within the ER, since TCR␣/␤ glycoproteins made in BW cells are retained within the ER as unassembled chains (6).
Initially, BW cells were cultured in medium or the glucosidase inhibitor Cas (34) and radiolabeled with [ 35 S]methionine or [ 3 H]galactose. Cells were solubilized in Nonidet P-40 and lysates immunoprecipitated with anti-TCR␤ mAb. As expected, the vast majority of TCR␤ proteins synthesized in Cas-treated [ 35 S]methionine-labeled groups showed decreased migration on SDS-PAGE gels relative to TCR␤ proteins in medium-treated groups (Fig. 2, compare lanes 1 and 2), due to blockade of co-translational removal of Glc residues from nascent glycan chains. TCR␤ proteins in medium-treated [ 3 H]galactose-labeled groups showed mobility similar to that for TCR␤ proteins in medium-treated [ 35  , and a lower band with mobility similar to that of TCR␤ proteins in medium-treated groups (Fig. 2, compare lanes 3 and 4), which was not observed in [ 35 S]methionine-labeled Cas groups (Fig. 2, lane 2). We reasoned that the lower TCR␤ band in [ 3 H]galactose-labeled cells represented pre-existent TCR␤ proteins that had undergone Glc trimming prior to the addition of Cas, but contained radioactive Glc residues added via the reglucosylation pathway. To confirm this idea, [ 3 H]galactose-labeling experiments were performed in the presence of the de novo protein synthesis inhibitor Chx. Chx treatment effectively inhibited Ͼ99% of de novo protein synthesis in BW cells as determined by parallel [ 35 S]methionine-labeling experiments (data not shown). Radiolabeling of TCR␤ proteins with [ 3 H]galactose in BW groups without Cas was not significantly affected by Chx (Fig. 2, compare lanes 3 and 5), and importantly, the upper TCR␤ species in Cas-treated [ 3 H]galactose-labeled groups was completely sensitive to Chx, whereas the lower species was not (Fig. 2, compare lanes 4 and 6). These results verify that the upper and lower TCR␤ species in [ 3 H]galactose-labeled groups represent newly synthesized and pre-existent TCR␤ proteins, respectively. Consistent with conversion of [ 3 H]galactose into [ 3 H]glucose, which is incorporated into glycan chains on pre-existent proteins within the ER, the radioactive signal on TCR␤ glycoproteins in medium/Chx-treated groups (the species shown in Fig. 2, lane 5) was sensitive to digestion with endoglycosidase H, which cleaves immature glycan chains, but resistant to digestion with jack bean mannosidase (data not shown), specific for fully trimmed glycans devoid of Glc residues (22).
Next, we evaluated if Glc residues added via the reglucosylation pathway persisted on nascent TCR␣/␤ proteins in BW cells. BW cells were radiolabeled with [ 3 H]galactose in the presence of Chx for 30 min and chased in nonradioactive medium containing Chx for various time periods in the presence or absence of Cas. Cells were solubilized in Nonidet P-40 and lysates sequentially immunoprecipitated with anti-TCR␤ mAb, followed by anti-TCR␣ mAb, to capture TCR␤ and TCR␣ proteins in BW lysates, respectively, which exist in this cell type as unassembled, free proteins due to low synthesis of CD3␦ chains (6). As shown in Fig. 3A, the relative amounts of radiolabeled TCR␣/␤ proteins visualized in chase groups was significantly decreased relative to pulse groups (Fig. 3A, lanes 1-4), indicative of removal of radiolabeled glycans from TCR␣/␤ proteins during the chase period by glucosidase enzymes. Indeed, inclusion of Cas during the chase period prevented the disappearance of radiolabeled TCR␣/␤ proteins (Fig. 3A, compare lanes  2-4 and 5-7). These results further support our conclusion that FIG. 2. Reglucosylation of TCR␣/␤ proteins in BW cells. BW cells were cultured in the presence of medium or Cas (100 g/ml) for 1 h at 37°C and radiolabeled with [ 35 S]methionine or [ 3 H]galactose for 30 min at 37°C. Where indicated, cells were pretreated with 1 mM Chx for 3 min prior to radiolabeling; the presence of Chx and/or Cas was maintained during the radiolabeling procedure. Cells were solubilized in 1% Nonidet P-40 and lysates immunoprecipitated with anti-TCR␤ mAb; precipitates were analyzed on SDS-PAGE gels under reducing conditions. Effectiveness of Chx treatment in inhibiting de novo protein synthesis was verified by parallel [ 35 S]methionine-labeling experiments, which showed that Ͼ99% of protein synthesis in BW cells was inhibited under these conditions (data not shown). The position of TCR␤ proteins is indicated.

FIG. 3. Reglucosylated glycans do not persist on TCR␣/␤ proteins in BW cells.
A, BW cells were radiolabeled with [ 3 H]galactose for 30 min at 37°C and chased in medium without radiolabel at 37°C for the time period indicated in the absence or presence of Cas. Note that the presence of Chx (1 mM) was maintained throughout the entire experiment and that Cas was present only during the chase period. Cells were solubilized in 1% Nonidet P-40 and lysates sequentially immunoprecipitated with anti-TCR␤ mAb, followed by anti-TCR␣ mAb; precipitates were analyzed as in Fig. 2A To determine if TCR glycoproteins undergo multiple cycles of Glc removal (deglucosylation) and Glc addition (reglucosylation) in the ER, BW cells were radiolabeled with [ 3 H]galactose for 15 min, chased in nonradioactive medium for up to 30 min, and then labeled a second time with [ 3 H]galactose; the presence of Chx was maintained throughout the entire experiment. As previously noted in Fig. 3A, decreased amounts of radiolabeled TCR␣/␤ proteins existed in chase groups relative to initial pulse groups (Fig. 3B, lanes 1-3), due to removal of radiolabeled [ 3 H]Glc saccharides by glucosidases during the chase period. Most importantly, we found that TCR␤ glycoproteins, and to a lesser extent TCR␣ proteins, were radiolabeled in the secondary pulse (Fig. 3B, lane 4), under conditions where new protein synthesis was completely abrogated (verified by parallel [ 35 S]methionine labeling experiments; data not shown). These results show that TCR␣/␤ glycoproteins undergo multiple cycles of deglucosylation and reglucosylation in the ER, a finding that is consistent with current models of protein folding and quality control (15). The molecular basis for differential radiolabeling of TCR␣ and TCR␤ proteins in secondary pulse periods remains to be determined, but does not result from differences in protein survival, as the half-lives of TCR␣/␤ chains are quite comparable in BW cells (6), even during extended chase periods (data not shown); rather, these results most likely reflect accelerated folding of TCR␣ proteins relative to TCR␤ proteins in BW cells.
Assembly of Reglucosylated TCR␣/␤ Proteins with Calnexin-To determine if reglucosylated TCR␣/␤ proteins were associated with calnexin and calreticulin, digitonin lysates of [ 3 H]galactose-labeled BW cells were precipitated with anticalnexin and anti-calreticulin Abs and analyzed on SDS-PAGE gels. As shown in Fig. 4A, numerous reglucosylated proteins were associated with calnexin in BW lysates (Fig. 4A, lane 1), including proteins that comigrated with clonotypic TCR␣/␤ proteins (Fig. 4A, lanes 3 and 4). The identity of reglucosylated TCR proteins in anti-calnexin precipitates was confirmed by analysis on two-dimensional NEPHGE/SDS-PAGE gels (Fig.  4B) and immunoprecipitation/release/recapture experiments and preclearing studies (data not shown). Interestingly, with the exception of an unknown 70-kDa protein, which coprecipitated equivalently with calnexin and calreticulin molecules in BW lysates (Fig. 4A, asterisk), markedly fewer reglucosylated proteins existed in anti-calreticulin precipitates compared to anti-calnexin precipitates (Fig. 4A, compare lanes 1 and 2). The significance of this finding is unclear but may reflect transient interaction of glycoproteins with calreticulin relative to calnexin, which we have recently observed regarding the association of nascent TCR␣/␤ proteins with calnexin, calreticulin chaperones in BW cells (35). Importantly, these results demonstrate that numerous reglucosylated proteins associate with calnexin in BW cells, including TCR␣/␤ proteins.
Synthesis and Processing of TCR␣/␤ Proteins Containing Truncated Glycan Chains-It has been shown that efficiency of reglucosylation is a function of glycan chain length (Man core size) (16). Therefore, we evaluated processing of TCR␣/␤ proteins in BW cell types that synthesize truncated glycan chains. For these studies, mutant BW cells were utilized, BWE, that synthesize glycoproteins with short Glc 3 Man 5 GlcNAc 2 glycans due to deficient formation of dolichol-P-mannose (26, 36), a key intermediate in the terminal stages of glycan biosynthesis (1,34). In addition, parental BW cells were treated with Msn, a chain terminator of glycan elongation resulting in transfer of truncated Glc 3 Man 5-7 GlcNAc 2 glycan chains to glycoproteins (37,38).
BW and BWE cells were cultured in medium or Msn and where indicated, the glucosidase inhibitor Cas was included in cultures to inhibit removal of Glc residues from nascent glycan

FIG. 4. Association of reglucosylated glycoproteins with calnexin. A, BW cells were radiolabeled with [ 3 H]galactose for 30 min at
37°C in the presence of cycloheximide, solubilized in 1% digitonin, and lysates immunoprecipitated with anti-calnexin Ab, anti-calreticulin Ab, or sequentially immunoprecipitated with anti-TCR␤ mAb, followed by anti-TCR␣ mAb. Precipitates were analyzed on SDS-PAGE gels under reducing conditions. The positions of TCR␣/␤ proteins are indicated; asterisk denotes an unknown 70-kDa reglucosylated protein present in anti-calnexin and anti-calreticulin precipitates of BW lysates that was not isolated with control antisera (data not shown). In the experiments presented in B, BW cells were radiolabeled with [ 3 H]galactose for 30 min at 37°C in the presence of cycloheximide; digitonin lysates were precipitated with anti-calnexin Ab or with a mixture of anti-TCR␣ plus anti-TCR␤ mAbs; precipitates were analyzed on two-dimensional NEPHGE/SDS-PAGE gels under reducing conditions. The positions of TCR␣/␤ proteins are indicated. chains. As shown in Fig. 5A, mobility of TCR␣/␤ proteins synthesized in Msn-treated BW cells and untreated BWE cells was increased relative to TCR␣/␤ proteins made in untreated BW cells (Fig. 5A, compare lanes 1, 3, and 5) and migration of TCR␣/␤ proteins from Msn-treated BW cells was slightly retarded compared to TCR␣/␤ proteins from untreated BWE cells (Fig. 5A, compare lanes 3 and 5); migration of TCR␣/␤ proteins in BWE cells was relatively unaffected by Msn treatment, as expected (Fig. 5A, compare lanes 5 and 7). TCR␣/␤ proteins in Cas groups migrated with decreased mobility relative to their respective control groups without Cas (Fig. 5A, compare even and odd numbered lanes), indicating that TCR␣/␤ proteins containing truncated glycan chains are processed by Glc trimming enzymes. The Glc trimming status of truncated glycan chains was also compared in parental BW cells and glucosidase II-deficient BWPHAR2.7 cells (6, 25); TCR␣/␤ proteins synthesized in Msn-treated parent BW cells migrated with increased mobility relative to those made in Msn-treated BWPHAR2.7 cells (Fig. 5B, compare lanes 3 and 4), indicating that TCR␣/␤ proteins having truncated glycans are effectively processed to at least the monoglucosylated stage. Taken together, these data are consistent with the synthesis of TCR␣/␤ proteins containing Man 9 core glycans in untreated BW cells, Man 5-7 core glycans in Msn-treated BW cells, and Man 5 core glycans in both untreated and Msn-treated BWE cells; the data show that TCR␣/␤ proteins having truncated glycans are effectively processed by the Glc trimming pathway.
Reglucosylation of TCR␣/␤ proteins containing truncated glycan chains was studied in [ 3 H]galactose labeling experi-ments performed in the presence of cycloheximide. As shown in Fig. 6A, reglucosylation of TCR␣/␤ proteins was markedly decreased in mutant BWE cells relative to parental BW cells (Fig.  6A), which did not result from decreased glucosyltransferase expression as determined by immunoblotting (Fig. 6B). Similar results were obtained regarding reglucosylation of TCR␣/␤ proteins and glucosyltransferase expression in medium-versus Msn-treated parental BW cells (data not shown). These data demonstrate that reglucosylation of TCR␣/␤ proteins having truncated glycans is severely impaired relative to those having standard length glycan chains.
Association of TCR␣/␤ Proteins Containing Truncated Glycan Chains with Calnexin-Having determined that TCR␣/␤ proteins containing truncated glycan chains are effectively processed by the Glc trimming pathway but not the reglucosylation pathway, we next examined their assembly with calnexin in [ 35 S]methionine-labeling experiments. As shown in Fig. 7A, numerous cellular proteins were assembled with calnexin in both BW and BWE cells, including TCR␤ proteins (Fig.  7A, top panels); interestingly, however, calnexin assembly with nascent TCR␣ proteins was markedly reduced in BWE cells compared to parental BW cells (Fig. 7A, top panels), despite the abundant presence of TCR␣ proteins in BWE lysates (Fig. 7A,  bottom panels). Decreased assembly of TCR␣ proteins with calnexin in mutant BWE cells versus parental BW cells was confirmed by immunoprecipitation/release/recapture studies (Fig. 7B). Assembly of nascent proteins with calnexin in BWE cells was dependent on Glc trimming as it was inhibited by pretreatment with Cas (data not shown). Similar to what was FIG. 5. Synthesis and processing of TCR␣/␤ proteins containing truncated glycan chains. A, BW and BWE cells were cultured in medium or 10 mM Msn for 10-12 h at 37°C in 5% CO 2 , and radiolabeled with [ 35 S]methionine for 30 min. Where indicated, castanospermine was included in the last 3 h of culture. The presence of Msn and/or Cas was maintained during the radiolabeling period. Cells were solubilized in 1% Nonidet P-40 and lysates sequentially immunoprecipitated with anti-TCR␤ mAb, followed by anti-TCR␣ mAb; precipitates were analyzed on SDS-PAGE gels under reducing conditions. The positions of TCR␣/␤ proteins are indicated. The proteins migrating slightly above TCR␤ in BW cells treated with Msn and BWE cells are unknown, but do not represent glycoproteins, as indicated by their insensitivity to glycosidase digestion (data not shown). B, parent BW cells and glucosidase II-deficient BWPHAR2.7 cells were cultured in medium or 10 mM Msn for 10 -12 h at 37°C in 5% CO 2 , and radiolabeled with [ 35 S]methionine for 30 min; the presence of Msn was maintained during the radiolabeling period. TCR␣ and TCR␤ proteins were immunoprecipitated from lysates and analyzed as in A. Note that TCR␣/␤ proteins made in BWPHAR2.7 cells migrate with decreased mobility relative to TCR␣/␤ proteins synthesized in parental BW cells due to persistence of diglucosylated glycans on TCR␣/␤ polypeptides. Lanes 1 and 3 contain untreated parent BW cells and Msn-treated parent BW cells, respectively; lanes 2 and 4 contain untreated BWPHAR2.7 cells and Msn-treated BWPHAR2.7 cells, respectively. observed in BWE cells, nascent TCR␤ proteins, but not TCR␣ proteins, were efficiently assembled with calnexin in Msntreated BW cells (Fig. 8A), which was verified by immunoblotting of anti-TCR␣ and anti-TCR␤ precipitates with anti-calnexin Ab (Fig. 8B). Taken together, these results demonstrate that Glc trimming is sufficient for calnexin assembly with TCR␤ proteins, and numerous other unidentified cellular proteins, but that both deglucosylation and reglucosylation are required for effective formation of calnexin-TCR␣ protein complexes.
Stability of TCR␣/␤ Proteins Containing Truncated Glycan Chains-Finally, the stability of newly synthesized TCR␣/␤ proteins containing truncated glycan chains was examined in [ 35 S]methionine-labeling experiments. As expected, the vast majority of both TCR␣ and TCR␤ proteins synthesized during a 30-min pulse in untreated parent BW cells were stable during a 30-min chase period (Fig. 9). In contrast, TCR␣ proteins made in Msn-treated BW cells showed relatively limited survival (Fig. 9), whereas stability of TCR␤ proteins was unaffected by Msn treatment (Fig. 9). Similar results were observed in untreated BWE cells in that TCR␣ proteins, but not TCR␤ proteins, were inherently unstable (Fig. 9). Consistent with these findings, significantly fewer TCR␣ proteins existed in lysates of BWE cells and Msn-treated BW cells compared to lysates of untreated BW cells as determined by immunoblotting with anti-TCR␣ mAb (data not shown). These data show that TCR␣ proteins containing truncated glycans have limited survival within the ER compared to TCR␣ proteins having standard length glycan chains, and that stability of TCR␤ proteins is unaffected by glycan core size. DISCUSSION In the current report we examined the deglucosylation and reglucosylation of TCR␣/␤ proteins and their assembly with calnexin in BW thymoma cells. The results in this study show that reglucosylated TCR␣/␤ proteins and other unidentified reglucosylated proteins associate with calnexin in vivo, providing the first evidence that reglucosylated proteins interact with calnexin chaperones in intact cells of any type (15,21). Interestingly, we found that significantly fewer reglucosylated proteins were associated with calreticulin compared to calnexin, which might reflect transient interactions of nascent glycoproteins with calreticulin relative to calnexin (8). In support of this idea, we find that association of TCR␣/␤ proteins with calreticulin is significantly transitory relative to their assembly with calnexin (35). However, we have also considered the possibility that calreticulin may preferentially associate with glycoproteins bearing monoglucosylated glycans generated via the Glc trimming pathway, which raises the interesting suggestion that proteins containing monoglucosylated species created by reglucosylation may not be conformationally identical to those derived via initial deglucosylation of nascent glycans.
The data in the current study are of important significance regarding the role of glucosidase enzymes in removal of Glc residues from reglucosylated proteins, as it has been proposed that endomannosidase enzymes localized within the intermediate compartment between the trans ER and the cis Golgi might also perform this function, providing a distal, more stringent level of quality control (10,39). Our finding that removal of Glc residues from reglucosylated TCR␣/␤ proteins was effectively blocked by castanospermine, an inhibitor of glucosidase I and glucosidase II enzymes (34), but not endomannosidase activity (40,41), indicates that glucosidase II is responsible for deglucosylation of the vast majority of reglucosylated TCR␣/␤ proteins in BW thymoma cells.
In agreement with current models of protein folding and quality control, our data indicate that TCR␣/␤ proteins undergo multiple cycles of Glc removal and addition. Interestingly, we found that comparable amounts of TCR␤ proteins were reglucosylated in consecutive labeling experiments, suggesting that no significant improvement in the folding status of TCR␤ proteins had occurred during a 30-min lag period. Regarding this issue, it is noteworthy to mention that similar amounts of reglucosylated TCR␤ proteins were associated with calnexin in consecutive labeling experiments in BW cells (data not shown), raising the consideration as to whether calnexin assembly with reglucosylated TCR␤ glycoproteins actually promotes their folding into the correct conformation, or if calnexin functions solely to retain incompletely folded glycoproteins in the ER.
Truncated Glc 3 Man 5 GlcNAc 2 glycan chains are generated via the alternative pathway of glycoprotein biosynthesis, which is utilized by many cell types following glucose starvation or ATP depletion (42)(43)(44). Glucosylated Man 5 GlcNAc 2 glycans are effectively transferred to proteins in the ER and can be processed to complex-type oligosaccharides in the Golgi (26,45). The current study shows that numerous molecules interact with calnexin in cell types synthesizing proteins with truncated glycans, providing evidence that quality control mechanisms that regulate egress of properly folded proteins from the ER to the Golgi complex are sustained under conditions where truncated glycan chains are added to newly synthesized proteins. These data are in agreement with recent findings by Spiro and co-workers, demonstrating that smaller size Glc 1 Man 5 GlcNAc glycans effectively interact with calreticulin in cell-free assays (10). The present study shows that reglucosylation of TCR␣/␤ glycoproteins is markedly decreased in BW cell types synthesizing truncated Man 5-7 core glycans compared to BW cells making typical Man 9 core glycans. While decreased reglucosylation of TCR␣ proteins is confounded by the differential stability of TCR␣ proteins under these conditions, decreased re- glucosylation of TCR␤ chains in BW cells making truncated glycans must result from differences in the recognition of TCR␤ proteins by glucosyltransferase enzymes in these cell types, as the stability of TCR␤ proteins is unaffected by glycan chain length. Indeed, our findings are in clear agreement with previously published results, showing that glycopeptides containing truncated glycans are extremely poor substrates for glu-cosyltransferase in vitro (16).
Our findings that reglucosylated TCR␤ proteins were assembled with calnexin even though reglucosylation was not necessary for formation of calnexin-TCR␤ protein complexes suggest that two populations of TCR␤ proteins exist within the ER of BW cells: TCR␤ proteins that effectively assemble with calnexin immediately after Glc trimming and those that require S]methionine at 37°C; the presence of Msn was maintained during the radiolabeling period. Cells were solubilized in 1% digitonin, and lysates immunoprecipitated with anti-calnexin Ab or with a mixture of anti-TCR␣ plus anti-TCR␤ mAbs. Precipitates were analyzed on two-dimensional NEPHGE/SDS-PAGE gels under reducing conditions. The positions of calnexin (bold arrow) and TCR␣/␤ proteins are indicated. TCR␣/␤ proteins made in Msn-treated BW cells show slightly increased migration toward the basic side of the gel relative to TCR␣/␤ proteins of untreated BW cells, indicative of incorporation of amine groups into terminal Man residues on glycan chains. B, digitonin lysates of equivalent numbers of medium-and Msn-treated BW cells were sequentially immunoprecipitated with anti-TCR␤ mAb, followed by anti-TCR␣ mAb. Precipitates were analyzed on SDS-PAGE gels under reducing conditions, transferred to nitrocellulose, and immunoblotted with anti-calnexin Ab. The position of calnexin is indicated. multiple rounds of deglucosylation and reglucosylation for chaperone interaction. Interaction of TCR␤ proteins with calnexin was not significantly decreased under conditions of impaired reglucosylation, however, indicating that most TCR␤ proteins associate with calnexin following initial removal of Glc residues from nascent glycan chains. We would also note that our data are consistent with the idea that TCR␤ proteins are reglucosylated while assembled into calnexin-TCR␤ protein complexes. We think this is possible because oligosaccharides are not required to maintain calnexin-protein interactions and are readily accessible to macromolecular probes such as lectins and glycosidases (12,13,22,46). Moreover, TCR␤ proteins contain multiple glycan chains, all of which are most likely not involved in calnexin binding. Experiments designed to explore these issues are currently in progress.
Finally, the current study demonstrates that survival of newly synthesized TCR␣ proteins in the ER is severely limited under conditions where reglucosylation is severely impaired. These findings importantly extend previous results that TCR␣ proteins containing full-length glycans are rapidly degraded under conditions were Glc residues persist on glycan chains and calnexin association is impaired (6), by showing that TCR␣ proteins are specifically degraded under conditions where numerous cellular proteins, including TCR␤ proteins, effectively assemble with calnexin. Impaired calnexin assembly with TCR proteins containing truncated glycans might also be explained by the failure of glucosidase enzymes to remove Glc residues from nascent oligosaccharides, as glucosidase I and II enzymes have been shown to deglucosylate Glc 1-3 Man 5-7 GlcNAc 2 glycans less efficiently than Glc 1-3 Man 9 GlcNAc 2 glycans in vitro (47). However, as noted in the current report, Glc trimming of TCR␣/␤ proteins to monoglucosylated species was not significantly impaired in BW cell types synthesizing truncated glycans. Whether or not removal of the final, innermost Glc residue occurs more slowly on TCR␣/␤ proteins containing truncated glycans than those having standard size glycans is unknown; however, we would point that this would favor the opportunity for calnexin binding, not decrease it (12). The ineffectiveness of the TCR␣ transmembrane domain to function as a stop transfer region (24) may relate to the reason why TCR␣ proteins containing truncated glycan chains do not effectively associate with calnexin as TCR␣ proteins might require numerous rounds of deglucosylation and reglucosylation to stably assemble with calnexin in the lipid bilayer of the ER. Alternatively, reglucosylation of an unknown accessory protein may be necessary for stable interaction of TCR␣ proteins with calnexin molecules. These issues are currently under investi-gation. It will also be interesting to determine if reglucosylation is critical for calnexin assembly with other proteins that utilize lectin-dependent folding pathways in the ER, such as major histocompatibility complex class I proteins and cystic fibrosis conductance transmembrane regulator molecules.