(Received for publication, September 9, 1996, and in revised form, October 29, 1996)
From the Experimental Immunology Branch, NCI, National Institutes of Health, Bethesda, Maryland 20892-1360
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
and
subunits of the T cell receptor (TCR) complex, and their
assembly with calnexin in BW thymoma cells. Our data demonstrate that
TCR
/
glycoproteins undergo multiple cycles of Glc removal and
addition within the endoplasmic reticulum and that numerous reglucosylated proteins assemble with calnexin, including TCR
/
glycoproteins. Importantly, the current study shows that TCR
proteins, but not TCR
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 Glc3Man9GlcNAc2 (Glc = glucose; Man = mannose; GlcNAc = 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 association of many glycoproteins with the molecular chaperone calnexin (3-7) 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 (Glc1Man9GlcNAc2) 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 Glc1Man9GlcNAc2 glycan chains may be generated via two mechanisms: cotranslational processing of immature Glc3Man9GlcNAc2 glycans by glucosidase I and glucosidase II enzymes (the Glc trimming pathway), or reglucosylation of fully trimmed Man8-9GlcNAc2 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 Man8-9GlcNAc2 glycans are reglucosylated much more efficiently than those having Man7GlcNAc2 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 Glc1Man9GlcNAc2 glycans), it is not crucial, as calnexin also binds smaller Glc1Man5-7GlcNAc2 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 GlcNAc 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.
BW parental cells, BWPHAR2.7 cells (25), and BWE cells (26) were maintained by weekly passage in RPMI 1640 medium containing 5% fetal calf serum at 37 °C in 5% CO2. BWPHAR2.7 cells and BWE cells were generously provided by Drs. Rosalind Kornfeld and Carolyn M. Knoll (Washington University, St. Louis, MO) and Drs. Stuart Kornfeld and Carolyn M. Knoll (Washington University, St. Louis, MO), respectively. Castanospermine (Cas) and mannosamine (Msn) were purchased from Calbiochem and Sigma, respectively.
Metabolic Labeling and Mannosamine/Castanospermine TreatmentMetabolic pulse-labeling of BW cells with [35S]methionine was performed as described previously (22). Briefly, cells were incubated in methionine-free RPMI 1640 medium (Biofluids, Rockville, MD) containing 10% fetal calf serum and 1 mCi/ml [35S]methionine (Tran35S-label; ICN, Irvine, CA) for 30 min at 37 °C in 5% CO2. For [3H]galactose labeling, cells were incubated in Glc-free RPMI 1640 medium (Life Technologies, Inc.) containing 10% dialyzed fetal calf serum, 5 mM sodium pyruvate (Life Technologies, Inc.), and 0.5 mCi/ml [6-3H]galactose (ICN) for 15-30 min at 37 °C in 5% CO2. Where indicated, [3H]galactose labeling was performed in the presence of cycloheximide (Chx; Sigma). For these studies, cells were incubated in RPMI 1640 medium containing 5% fetal calf serum and 1 mM Chx for 2 min at 37 °C in 5% CO2, centrifuged, and resuspended in Glc-free medium containing [3H]galactose and 1 mM Chx; radiolabeling was then performed as described above. Effectiveness of Chx treatment in blocking new protein synthesis was verified by parallel experiments using [35S]methionine. In experiments using Msn, cells were cultured overnight in medium containing 10 mM Msn at 37 °C in 5% CO2; 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 µg/ml final concentration) during the final 3 h of culture.
Cell Lysis and ImmunoprecipitationCells were solubilized
in 1% Nonidet P-40 (Calbiochem) lysis buffer (20 mM Tris,
300 mM NaCl, plus protease inhibitors) or 1% digitonin
(Wako, Kyoto, Japan) lysis buffer (20 mM Tris, 150 mM NaCl, plus protease inhibitors) at 1 × 108 cells/ml for 20 min at 4 °C. Cell lysates were
clarified by centrifugation to remove insoluble material and
immunoprecipitated with the appropriate antibodies pre-absorbed to
protein A-Sepharose beads; sequential immunoprecipitation and
immunoprecipitation/release/recapture techniques were performed as
described previously (22). The following monoclonal antibodies (mAb)
were used in this study: H28-710, specific for TCR (27), and
H57-597, specific for TCR
(28); the following antisera were used:
SPA-860 anti-calnexin Ab (Stressgen Biotechnologies Corp., Victoria,
BC, Canada), anti-calreticulin Ab (Affinity Bioreagents Inc., Neshanic
Station, NJ), and anti-glucosyltransferase Abs directed against both
native and denatured glucosyltransferase (kindly provided by Dr.
Armando Parodi, Buenos Aires, Argentina).
One- and two-dimensional SDS-PAGE gel electrophoresis were performed as described previously (29). Gels containing [35S]methionine-labeled material were processed for autoradiography using dimethyl sulfoxide (Sigma), saturated with 2,5-diphenyloxazole (Aldrich). Gels containing [3H]galactose-labeled material were processed for autoradiography using EN3HANCE (DuPont). Immunoblotting experiments were performed as described previously (29).
To study reglucosylation of TCR/
proteins in BW
cells, [3H]galactose was utilized as a radioactive tracer
of Glc residues on oligosaccharide chains (30, 31). As illustrated in
Fig. 1, exogenous [3H]galactose
internalized by cells can radiolabel glycoproteins containing
N-linked oligosaccharides via three major pathways: (i)
conversion into UDP-[3H]galactose, the sugar donor for
galactosyltransferase enzymes which transfer galactose residues to
mature, complex type oligosaccharides in the trans Golgi; (ii)
epimerization of UDP-[3H]galactose to
UDP-[3H]glucose, the sugar donor for
UDP-glucose:glycoprotein-dependent glucosyltransferase
enzymes that transfer Glc residues to high mannose glycans on
incompletely folded glycoproteins in the ER; and (iii) conversion of
UDP-[3H]glucose into
dolicholphospho-[3H]glucose, which is incorporated into
nascent Glc3Man9GlcNAc2 glycans
that are cotranslationally added to newly synthesized polypeptides in
the ER (Fig. 1) (1, 32, 33). In our studies, radiolabeling of
TCR
/
proteins with [3H]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 [35S]methionine
or [3H]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
[35S]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
[3H]galactose-labeled groups showed mobility similar to
that for TCR
proteins in medium-treated
[35S]methionine-labeled groups (Fig. 2, compare
lanes 1 and 3); interestingly, however, two
TCR
species were visible in Cas-treated groups of [3H]galactose-labeled cells: an upper band corresponding
to newly synthesized TCR
proteins containing untrimmed glycan chains
(Fig. 2, lane 4), analogous to the species observed in
[35S]methionine-labeled Cas groups (Fig. 2, lane
2), 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
[35S]methionine-labeled Cas groups (Fig. 2, lane
2). We reasoned that the lower TCR
band in
[3H]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,
[3H]galactose-labeling experiments were performed in the
presence of the de novo protein synthesis inhibitor Chx. Chx
treatment effectively inhibited >99% of de novo protein
synthesis in BW cells as determined by parallel
[35S]methionine-labeling experiments (data not shown).
Radiolabeling of TCR
proteins with [3H]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 [3H]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
[3H]galactose-labeled groups represent newly synthesized
and pre-existent TCR
proteins, respectively. Consistent with
conversion of [3H]galactose into
[3H]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 [3H]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
radiolabeled [3H]galactose saccharides are converted into
[3H]Glc residues, which are transferred to pre-existent
TCR
/
proteins via the reglucosylation pathway in BW cells.
Moreover, these studies show that [3H]Glc residues do not
persist on reglucosylated TCR
/
glycoproteins in BW cells but are
rapidly removed within 15 min of their addition.
To determine if TCR glycoproteins undergo multiple cycles of Glc
removal (deglucosylation) and Glc addition (reglucosylation) in the ER,
BW cells were radiolabeled with [3H]galactose for 15 min,
chased in nonradioactive medium for up to 30 min, and then labeled a
second time with [3H]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
[3H]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
[35S]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.
To determine if reglucosylated TCR/
proteins were
associated with calnexin and calreticulin, digitonin lysates of
[3H]galactose-labeled BW cells were precipitated with
anti-calnexin 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
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 Glc3Man5GlcNAc2 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
Glc3Man5-7GlcNAc2 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 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 Man9 core glycans in untreated BW cells, Man5-7 core glycans in Msn-treated BW cells, and
Man5 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 [3H]galactose labeling experiments
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
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
[35S]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 observed in BWE cells, nascent TCR
proteins, but not TCR
proteins, were efficiently assembled with
calnexin in Msn-treated 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
Finally, the stability of newly synthesized TCR/
proteins containing truncated glycan chains was examined in
[35S]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.
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 Glc3Man5GlcNAc2 glycan
chains are generated via the alternative pathway of glycoprotein
biosynthesis, which is utilized by many cell types following glucose
starvation or ATP depletion (42-44). Glucosylated
Man5GlcNAc2 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 Glc1Man5GlcNAc
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 Man5-7 core glycans compared to BW cells making
typical Man9 core glycans. While decreased reglucosylation
of TCR
proteins is confounded by the differential stability of
TCR
proteins under these conditions, decreased reglucosylation 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 glucosyltransferase 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 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 Glc1-3Man5-7GlcNAc2 glycans less efficiently than
Glc1-3Man9GlcNAc2 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 investigation. 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.
This work is dedicated to Flora Ann Kirby Casteel.
We thank Drs. Juan Bonifacino, Richard Hodes, Armando Parodi, Randall Ribaudo, Paul Roche, and Dave Segal for critical reading of the manuscript. We are especially grateful to Drs. Stuart Kornfeld and Carolyn Knoll for providing BWE cells and to Dr. Armando Parodi for helpful discussion and for providing anti-glucosyltransferase Abs.