| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Molecular Pathology and Medicine. DIBIT-San Raffaele Scientific Institute, 20132 Milan, Italy
1Correspondence: Department of Molecular Pathology and Medicine, DIBIT-San Raffaele Scientific Institute, Via Olgettina 58, 20132 Milan, Italy. E-mail: r.sitia{at}hsr.it
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: IgM • quality control • redox regulation • secretion
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
for review). Proteins that fail to fold
or assemble are retained in the ER and eventually degraded
(2)
.
A large body of evidence now indicates that cytosolic proteasomes are
responsible for the degradation of both membrane and soluble
ER-synthesized proteins (3
4
5
6
7
8
9
10
11
12
13)
. This implies that
proteins targeted for degradation must be translocated across the ER
membrane to reach the cytosol where proteasomes reside
(14)
. By the use of yeast mutant strains and proteasome
inhibitors in mammalian cells, it has been possible to recapitulate the
following chain of events in the life of a short-lived ER protein (refs
15
16
17
and references therein): substrates enter the ER
lumen, as demonstrated by the fact that they often undergo
N-linked glycosylation, and associate with chaperones and/or
physiological assembly partners localized in the ER (e.g., Class I
heavy chains with ß2 microglobulin). They can be then ubiquitinated
and deglycosylated by ubiquitin-conjugating enzymes and N-glycanases
associated with the cytosolic face of the ER membrane
(18
19
20)
, before being degraded by proteasomes.
Sec61, an essential component of the translocon that mediates the entry
of proteins into the ER (21)
, also seems to be involved in
the retrograde translocation, or dislocation, toward the cytosol
(22
, 23)
. If similar translocon complexes can mediate both
the entry and the exit from the ER, the question arises as to what
determines the directionality of transport across the ER membrane
(24
, 25)
.
A feature of ER-associated degradation that has long been puzzling is
the absence of intermediates in the process. This implies that once
initiated, degradation proceeds rapidly. Thus, systems that act as
pulling forces are likely to be operative to facilitate the extraction
of proteins from the ER. Ubiquitination (9
, 26)
and
binding to cytosolic molecules, including N-glycanase or the
proteasomes themselves (27)
, might play a role in the
process of dislocation.
To investigate the molecular mechanisms underlying the translocation of soluble proteins from the ER to the cytosol, we have chosen assembly-deficient myeloma cells that do not produce immunoglobulin (Ig) light (L) chains. In these cells, all newly synthesized Ig µ and J subunits are rapidly degraded from the ER. Evidence is presented indicating that the extraction of these soluble short-lived polypeptides from the ER lumen is coupled to their degradation.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
29
30
31)
. WEHI-231, a B lymphoma, and WEHI-231-J 3E5, a
transfectant expressing J chains (32)
kindly donated by
Dr. Ron Corley (Boston, Mass.), were maintained in RPMI1640
supplemented with 10% fetal calf serum and 50 µM 2-mercaptoethanol.
A rabbit anti-mouse (RAM) antiserum specific for J chains was a
kind gift of Dr. R. M. E. Parkhouse (Pirbright Laboratory,
Surrey, U.K.) (33)
. Anti-hsp70 monoclonal antibodies
(SPA-820) were provided by Dr. G. Multhoff (GSF; Munich, Germany),
whereas anti-calnexin and anti-ERp72 were obtained through the
generosity of Drs. A. Helenius (ETH, Zurich, CH) and M. Green (St.
Louis University, St. Louis, Mo.), respectively. Horseradish
peroxidase-conjugated goat anti-mouse µ, goat anti-mouse Ig, or goat
anti-rabbit Ig were from Southern Biotechnology Associates, Inc.
(Birmingham, Ala.). Purified RAM-µ antibodies and an
anti-calreticulin antiserum were from Zymed and Affinity BioReagents
(Golden, Colo.), respectively.
Brefeldin A, cycloheximide (CHX), diamide, dithiothreitol (DTT),
leupeptin, N-ethyl maleimide (NEM), phenylmethanesulfonyl fluoride
(PMSF), and TritonX100 were purchased from Sigma (St. Louis, Mo.);
endoglycosidases H and F, N-tosyl-L-phenylalanine chloromethyl ketone
(TPCK), N-acetyl-leucyl-leucyl-norleucinal (ALLN), and DNase I were
from Boehringer (Mannheim, Germany). Carboxybenzyl-leucyl-leucyl-leucyl
vinylsulfone (ZL3VS) and
carboxybenzyl-leucyl-leucyl-leucinal (ZL3H), two
proteasome inhibitors (6
, 34
, 35)
, were kind gifts of Drs.
M. Bogyo and H. Ploegh (Harvard University, Cambridge, Mass.).
Unless otherwise indicated, ALLN was used at 260 µM, diamide at 1 mM, leupeptin at 10 µM, TPCK at 250 µM, ZL3VS and ZL3H at 50 µM.
Immunoprecipitation (IP), gel electrophoresis, and Western
blotting (WB)
Pulse-chase, IP, and gel electrophoresis were performed as
described previously (30)
. To detect interactions among ER
proteins, cells were pulse labeled for 2 h with
[35S] methionine and cysteine, lysed in 1%
Nonidet P-40, and immunoprecipitated with rabbit anti-µ, anti-J,
anti-calnexin, or anticalreticulin antibodies and protein A Sepharose.
After three washes, immunoprecipitates were eluted by heating at 95°C
for 5 min in 2% sodium dodecyl sulfate (SDS), diluted in
phosphate-buffered saline to a final concentration of 0.02% SDS, and
reimmunoprecipitated with µ- or J-specific antibodies or with protein
A Sepharose alone as a control.
In most experiments, cells were exposed to 
1 mM iodoacetamide or NEM
before lysis to prevent disulfide interchange. Aliquots of the lysates,
treated or not with endoglycosidases H or F (Boehringer-Mannheim), were
resolved by SDS-PAGE (polyacrylamide gel electrophoresis) under
reducing (samples were heated for 5 min at 95°C in the presence of 50
mM DTT and alkylated with 200 mM iodoacetamide at room temperature) or
nonreducing (samples were treated only with iodoacetamide) conditions
and transferred to nitrocellulose. Filters were decorated with the
appropriate antibodies and developed with the ECL detection system
(Amersham Italy, Milan) as recommended by the supplier. In some
experiments, filters were cut horizontally to correspond to the 43 kDa
marker; the upper part was decorated with anti-µ, and the lower with
anti-J.
Films were scanned by an automated densitometer (Molecular Dynamics, Sunnyvale, Calif.) and individual bands quantitated by ImageQuant software. At least two exposures of each film were analyzed.
Subcellular fractionation and protease protection assays
About 3 x 107 cells were washed once
in phosphate-buffered saline, once in homogenization buffer (10 mM
TrisCl, pH 7.4, 250 mM sucrose, 0.1 mM PMSF [1 mM NEM was added in the
fractionation of N[µ1] cells]), and resuspended in 500 µl of the
same buffer. Cells were homogenized by passing them 25–30 times
through a G22 needle and spun three times at 1000 g at 4°C
to remove nuclei and residual unbroken cells. The postnuclear
supernatant (PNS) was spun at 10,000 g (10K) for 30 min at
4°C to generate a pellet (P 10) and a supernatant (S 10). Part of the
latter was fractionated by centrifugation at 200,000 g for
1 h at 4°C so as to yield a pellet (P 200) and a supernatant (S
200). Pellets were resuspended in 0.1% SDS, 0.2% TritonX100, 1 mM
NEM, 0.1 mM PMSF, and 10 µg/ml DNase I.
Individual fractions were normalized with respect to their total protein concentration (determined by Bio-Rad protein assay as recommended by the supplier) and analyzed by SDS-PAGE and Western blotting.
For protease protection assays, the PNS from N[µ1] cells, prepared in the absence of any protease inhibitor, was brought to 5 mM Tris-HCl pH 8.0 and 2 mM CaCl2 and digested at 4°C with 10 µg/ml proteinase K (Boehringer-Mannheim) in the presence or absence of 1% Nonidet P-40. After 60 min of incubation, PMSF and a mixture of protease inhibitors (Complete, Boehringer-Mannheim) were added to block proteolysis. The control sample was performed by adding protease inhibitors before incubation with proteinase K. Samples were then analyzed by WB before or after centrifugation at 10,000 g.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(30)
. In the absence of L chains, IgM
polymers cannot be assembled, and µ and J subunits are retained in
the ER and eventually degraded. To investigate and compare the
degradation of µ and J chains, N[µ1] were pulsed for 5 min with
radioactive amino acids and chased for different times before
immunoprecipitation with µ- or J- specific polyclonal antibodies.
After a lag of ~30 min, the intensity of both µ and J bands
decreased with similar kinetics (Fig. 1
). As previously observed for µ chains (30)
, the
degradation of J chains was only slightly affected by the presence of
brefeldin A during the chase (not shown), indicating that transport to
the distal parts of the secretory pathway is not required for
degradation.
|
Inhibitors of cytosolic proteasomes prevent the degradation of both
µ and J chains
To characterize the mechanisms of µ and J chain degradation, we
tested the effect of a panel of protease inhibitors on the disposal of
the µ and J pools after blocking protein synthesis with cycloheximide
(Fig. 2A
). Three drugs known to inhibit cytosolic proteasomes—ALLN,
ZL3H, and ZL3VS (35
, 34)
—efficiently prevented the degradation of both µ and J,
whereas leupeptin and TPCK were not effective. Thus, cytosolic
proteasomes seem to be involved in the disposal of the two Ig subunits.
However, although the kinetics (Fig. 1)
and mode of disposal were
similar, a notable difference between the two chains was evident when
the electrophoretic patterns were compared: whereas the gel mobility of
the anti-µ reactive band was not affected by the three proteasome
inhibitors, bands of faster mobility
(Mr of ~23 and 21 kDa) became
detectable with anti-J antibodies when proteasome activity was blocked
(Fig. 2A
, lanes 3–5). These bands were barely detectable in
untreated cells (lane 1) or after treatment with leupeptin or TPCK
(lanes 6, 7), where a single species of ~25 kDa was decorated by
anti-J antibodies.
|
Pulse-chase assays (Fig. 2B
) confirmed the involvement of
cytosolic proteasomes in the degradation of µ and J chains, and
suggested a precursor–product relationship among the three J isoforms
in N[µ1] cells, the 23 kDa form becoming detectable after 60 min of
chase and the fastest migrating one after longer chase times (90–120
min). Unlike in the case of J, no bands of altered electrophoretic
mobility were detected by anti-µ antibodies in either the Western
blot (Fig. 2A
) or the pulse-chase assay (Fig. 2B
).
The sequential appearance of the faster migrating J isoforms was
followed by Western blot analyses of NSO cells exposed to proteasome
inhibitors for different times (Fig. 2C
). The 23 kDa species
became detectable after 15 min of culture with ZL3H (lane
2), whereas a longer time was required to detect the 21 kDa band (lane
5). The addition of CHX did not alter the kinetics of appearance of the
two isoforms (data not shown), indicating that de novo
protein synthesis is not required for the generation of the 21 kDa
fragment. The two J chain isoforms may consist of intermediates in the
process of degradation, accumulating when proteasome function is
inhibited. Their detection at later times in pulse-chase assays (Fig. 2B
) parallels the lag observed in degradation (see Fig. 1
),
suggesting that newly made polypeptides are somehow protected from
degradation.
To further confirm that the two faster bands derive from J chains, we
analyzed a WEHI-231 B lymphoma transfectant expressing J chains (Fig. 2D
). In WEHI-231-J 3E5 cells, J chains are inefficiently
incorporated into IgM polymers (32
; C. Fagioli and R.
Sitia, unpublished results) and are degraded (compare lanes 2 and 3).
In the presence of ZL3H, degradation is inhibited and two
bands of ~23 and 21 kDa accumulate in the lysates (lane 4). These
bands were not detected in untransfected WEHI-231 cells (lane 1),
excluding artifactual cross-reactions of the antibodies used.
Accumulation of deglycosylated J chains in the presence of
proteasome inhibitors
By analogy with other systems (22)
, the appearance of
faster migrating J isoforms could reflect removal of N-glycans during
the process of retrotranslocation from the ER to the cytosol. In
agreement with this, the slower of the two isoforms accumulating in
ZL3H-treated NSO cells (Fig. 3A
, lane 3) comigrated with unglycosylated J chains
synthesized in the presence of tunicamycin (lane 1, TM). Moreover,
treatment with endoglycosidases H or F (Fig. 3B
) did not
alter the electrophoretic mobility of the two bands induced by
proteasome inhibitors while shifting that of glycosylated J chains, the
only isoform detectable in untreated cells (compare lanes 1 and 2 in
Fig. 3B
). In view of the precursor–product relationships
revealed by the pulse-chase assays (Fig. 2B
), the 23 kDa
band accumulating in the presence of proteasome inhibitors likely
consists of J chains in which the single N-linked glycan has
been removed. The faster J isoform (21 kDa) might consist of a
proteolytic fragment generated by protease(s) resistant to the
inhibitors (36)
. In both N[µ1] and NSO cells, its
relative abundance varied according to the nature of the inhibitor
used, being most prominent after treatment with ZL3H (Fig. 2A
).
|
Deglycosylated J chains are in a reduced state
Although the structure of J chains has not yet been resolved, this
polypeptide is known to form intrachain disulfide bonds
(37)
. These can be revealed by the different
electrophoretic mobility of glycosylated J chains present in untreated
NSO cells resolved under reducing or nonreducing conditions (compare
lanes 1 and 2 in Fig. 4
). In contrast, the two deglycosylated isoforms accumulating in the
presence of proteasome inhibitors displayed the same mobility with or
without DTT (lanes 3–4), indicating that reduction of the substrate
precedes degradation.
|
Deglycosylated J chains accumulate in the cytosol when proteasomes
are inhibited, whereas µ chains are retained in the ER lumen
The above results suggested that J chains synthesized,
N-glycosylated, and oxidized in the ER lumen are retrotranslocated to
the cytosol to be degraded by proteasomes. To determine the
localization of the different J isoforms, we performed subcellular
fractionation of NSO cells treated with or without ZL3H for
2 h Aliquots from individual fractions, normalized for their total
protein contents, were resolved by SDS-PAGE and analyzed by WB with
antibodies against different marker proteins. As shown in Fig. 5A
, hsp70 was detected only in the S 10 and S 200 fractions,
where cytosolic proteins are expected to accumulate. In contrast, the
distribution of calnexin (clnx), a membrane protein of the ER
(38)
, was limited to the microsome containing fractions
(mainly in P 10). Some ERp72, a soluble ER resident protein, was also
present in the cytosolic fractions, suggesting that some microsomes
were broken during the fractionation procedure. The distribution of
these three markers was similar in ZL3H-treated or control
cells (right and left panels, respectively). On the other hand, notable
differences were evident when blots were decorated with anti-J
antibodies. In untreated cells, only glycosylated J chains were
detected. Consistent with their expected localization in the ER lumen,
most of them accumulated in the P 10 fraction. Their pattern was
similar to that of ERp72, although much less glycosylated J chains were
detected in the cytosolic fractions than ERp72. Since we failed to
detect Nonidet P-40 insoluble J chains (C. Fagioli and R. Sitia,
unpublished results; Fig. 6
, below), these findings may suggest that J chains interact with ER
membrane protein(s). When proteasomal function was inhibited, the bands
corresponding to deglycosylated J chains became detectable in the
cytosolic fractions. In particular, the 21 kDa isoform was found
exclusively in the soluble fractions.
|
|
As the removal of the J chain glycans was coupled to
retrotranslocation, the absence of deglycosylated µ chains in
ZL3H-treated N[µ1] cells could reflect their retention
in the ER lumen, which would make them inaccessible to the glycanase
activity located in the cytosol. Therefore, to compare the subcellular
localization of µ and J chains, the above experiment was repeated
with N[µ1] cells. The distribution of calnexin (not shown) and J
chains was similar in ZL3H-treated NSO (Fig. 5A
)
and N[µ1] (Fig. 5B
) cells. In contrast to J, very few
(if any) µ chains were detectable in the cytosolic fractions, and all
displayed the mobility of glycosylated molecules.
These results suggested that µ chains fail to retrotranslocate
efficiently to the cytosol when proteasome function is inhibited.
However, if µ chains formed aggregates on dislocation into the
cytosol, these could codistribute with microsomes during
centrifugation. To discriminate between these two possibilities,
protease protection assays were performed on the PNS of
ZL3H-treated N[µ1] cells (Fig. 6
). In these assays, molecules that are present in the lumen should be
protected from proteolytic digestion unless the integrity of the
membrane is perturbed by detergents, whereas molecules that wholly or
in part translocated to the cytosol should be sensitive. In the absence
of detergent, most µ chains present in the PNS are resistant to
proteinase K (compare lanes 1 and 2, Fig. 6A
). In contrast,
an anti-µ immunoreactive fragment of ~60 kDa was generated when
detergent was added (lane 3, Fig. 6A
). These results
indicated that most µ chains indeed accumulate in the lumen in the ER
in a conformation that is in part resistant to proteases. Aliquots of
the PNS samples were then centrifuged at 10,000 g (Fig. 6B
). As expected, the deglycosylated J chain isoforms
present in the cytosol were entirely sensitive to protease, whereas the
25 kDa glycosylated species was resistant unless detergent was added
(Fig. 6B
, lanes 5, 6). No anti-J immunoreactive peptides
were detected, suggesting that unassembled J chains are retained in the
ER in a protease-sensitive conformation. Digestion with trypsin yielded
similar results (not shown). Like the deglycosylated J isoforms, the
small amounts of µ chains present in the S 10 fraction (Fig. 6B
, lane 2) were digested by proteinase K (Fig. 6B
, lane 4). Addition of detergent allowed all µ and J
chains to undergo proteolysis, excluding the possibility that µ
chains formed protease resistant aggregates when degradation is
prevented. Also in this experiment, more ERp72 molecules than
glycosylated µ and J chains were found in the S 10 fractions (not
shown). Taken together, these data indicate that, when proteasome
function is inhibited, a fraction of J chains can be dislocated to the
cytosol and undergo deglycosylation, whereas virtually all µ chains
are retained in the ER lumen in a conformation that is partially
resistant to proteases. Accordingly, when cells were incubated in the
presence of proteasome inhibitors, the staining obtained with
fluorescent anti-µ increased in intensity, without significant
changes in the pattern, which remained largely overimposable to PDI
distribution (data not shown).
Interactions of µ and J chains with calnexin and calreticulin
The above results suggested that active proteasomes are required
for extraction of µ chains from the ER lumen, whereas some J chains
can be exported to the cytosol in the presence of inhibitors. This may
reflect different interactions of the two Ig subunits with ER
chaperones. As µ and J chains carry five and one N-linked
glycans, respectively, we investigated their association with calnexin
and calreticulin, two chaperones known to bind monoglucosylated
proteins in the ER (1)
by reimmunoprecipitation assays
(Fig. 7A
). Lanes 1 and 2 show the amount of immunoprecipitated µ
and J chains that can be reimmunoprecipitated with the corresponding
antibodies. A fraction of µ chains were present in the eluates of
anticalnexin (lane 3) and anticalreticulin (lanes 6). In contrast,
neither antibody coprecipitated J chains (lanes 4 and 7). The µ-like
band present in lane 7 is probably an artifact, as it was not detected
in other experiments. Hence, calnexin and calreticulin seem to
associate more with µ than with J chains.
|
Pulse-chase assays (Fig. 7B
) were then performed to
investigate the kinetics of the interaction between µ chains and
calnexin. After 5 min of pulse, a µ chain band was evident in the
anticalnexin precipitates (lane 1). As expected, this band was no
longer detected after 4 h of chase (lane 2), a time when most
radioactive µ chains were degraded. When ZL3H was present
during the chase, the intensity of the µ chain band was only slightly
reduced. These results confirmed that in the absence of active
proteasomes, µ chains fail to retrotranslocate to the cytosol, and
remain glycosylated and associated with calnexin in the ER.
Redox regulation of J chain degradation
In general, disassembly and unfolding are necessary for the
translocation of proteins across a membrane (39
40
41)
. If
this were also the case for the retrotranslocation of µ and J chains,
disruption of intrachain disulfide bonds might accelerate degradation
whereas oxidation should inhibit it (42)
. Therefore, we
investigated the effects on degradation of diamide, a membrane-permeant
oxidant, and of two reducing agents, 2ME and DTT (43)
. As
shown in Fig. 8A
, 2ME increased the degradation of J chains whereas diamide
completely blocked it. Deglycosylated J chains were not detected in
diamide-treated cells (lane 6). Consistent with the possibility that
oxidation prevents the retrograde translocation of J chains to the
cytosol, addition of diamide to ZL3H-treated cells
prevented the appearance of deglycosylated J chains (Fig. 8B
). Conversely, DTT caused an increase in the relative
proportions of the deglycosylated J chains, as well as a shift toward
the 21 kDa isoform (Fig. 8C
). DTT also facilitated the
degradation of µ chains (compare lanes 2 and 3 of Fig. 8C
). However, despite its clear effects on J chains, the
presence of DTT did not induce the appearance of deglycosylated µ
chains, suggesting that reduction of intrachain SS bonds
(43)
was not sufficient to allow µ chain
retrotranslocation.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 35)
on the
degradation of unassembled Ig subunits confirm the involvement of
cytosolic proteasomes in maintaining homeostasis within the ER,
extending it to two soluble mammalian glycoproteins. The substrates
used in this study (µ and J chains) are not mutated or aberrant:
their fate is dictated by the absence of L chains, which prevents IgM
assembly and oligomerization, thereby making them substrates of the ER
quality control machinery (16)
. The disposal of µ
(30)
and J (C. Fagioli, A. M. Fra, and R. Sitia,
unpublished results) is not significantly inhibited by brefeldin A, and
both chains are sensitive to Endo-H, confirming that they need not be
transported to the Golgi to be degraded.
Although several features of the breakdown of µ and J chains
(subcellular localization, kinetics, pharmacology) are similar,
remarkable differences become evident when the proteasomal activity is
blocked. Under these conditions, faster migrating J chain isoforms
accumulate and become easily detectable, whereas the main species
present in untreated cells consist of glycosylated and oxidized J
chains localized in the ER. Pulse-chase experiments clearly establish
precursor–product relationships between the latter and the two
isoforms induced by proteasome inhibitors. The species of intermediate
mobility appears first, and likely corresponds to J chains in which the
single N-glycan has been removed. As this form is also detectable,
although at low levels, in cells that have not been treated with
proteasome inhibitors, it is likely to represent a normal intermediate
in the J chain degradation pathway. This isoform is localized primarily
in the cytosol (Figs. 5
, 6)
, where it can be further converted into the
fastest migrating isoform. The latter may represent a proteolytic
fragment generated by proteasome-independent activities, although this
remains to be established.
The opposite effects of reducing agents and diamide (Fig. 8)
suggest
that the pathways of J chain degradation are redox regulated. Reducing
agents induce µ chain degradation as well. In all likelihood,
proteins translocate more easily in the reduced and denatured state
(42)
. However, the use of uncompartmentalized redox
modulators does not allow to draw any firm conclusions, as several
steps in both the export and the proteolytic machineries are inhibited
by thiol-oxidizing reagents and, conversely, facilitated by reducing
agents (44)
.
The fate of unassembled J chain therefore seems to involve, as
described for other substrates (refs 15
16
17
and references
therein), retrotranslocation across the ER membrane, removal of
amino-terminal glycans, and degradation by cytosolic proteasomes. For
some substrates, retrotranslocation seems to be independent of
proteasome activity (26)
. In contrast, in both myeloma and
B lymphoma cells, the amount of J chain isoforms that accumulate in the
cytosol in a given time is much smaller than the fraction degraded in
the absence of proteasome inhibitors (Fig. 2B, D
). These
observations suggest that degradation and retrotranslocation of Ig
subunits are coupled phenomena. This is more evident for µ than for J
chains: in conditions in which deglycosylated J chains are easily
detectable in the cytosol, most µ chains remain in the lumen of the
ER microsomes (Fig. 6)
, glycosylated and associated with resident
chaperones (Fig. 7)
. Thus, proteasomes themselves might be one pulling
force in the extraction of certain substrates from the ER. Indeed,
proteasomes are often detected at the cytosolic face of the ER
(14)
and have been implicated in the extraction of an
artificial membrane protein from the ER (27)
. The ATPases
present in the 19S cap might be responsible for the capture of the
substrates as they emerge from the membrane, facilitating their
insertion into the proteolytic cavity as well as their extraction from
the ER. ATP breakdown could provide part of the energy for retrograde
translocation. In the presence of inhibitors, the proteolytic cavity
could be occluded, and the transport chain be slowed down.
Why is the retrotranslocation of certain polypeptides (µ
chains) inhibited to a larger extent than others when proteasomes are
not active? Unfolding is generally crucial for the translocation of
proteins across membranes (39)
(41)
. Although
the translocon seems to be capable of considerable lateral expansion
(45
, 46)
, folding, assembly, disulfide bond formation, and
other posttranslational modifications within the ER lumen are likely to
generate structures that are not compatible with translocation.
Compared to J, µ are large multidomain proteins, with five N-glycans
and numerous intra- and interchain disulfide bonds. The partial
resistance to proteases suggests a more compact conformation of µ
chains. Their extraction from the ER might require the coordinated
unfolding and reduction of the five Ig domains. However, although DTT
favors degradation, it is not sufficient to induce the accumulation of
deglycosylated µ chains when proteasomes are inactive (Fig. 8)
.
The involvement of Sec61 in both cotranslational translocation and
dislocation of ER proteins to the cytosol supports the concept that the
translocon is itself devoid of vectoriality, but rather acts as a
tunnel that can be used in both directions (24
, 47)
. The
characterization of yeast Sec61 mutants in which retrograde
translocation is selectively impaired (23
, 48)
demonstrates that import and export through the translocon are
mechanistically distinct processes. However, as in normal conditions
proteins can be transported across Sec61 in opposite directions, the
mechanisms that determine vectoriality remain to be elucidated.
Vectoriality could be the result of a ‘tug of war game’, with forces
that pull on both sides of the membrane. In this connection, BiP has
been shown to act as a ‘Brownian ratchet’ in the posttranslational
import of prepro-
factor into the ER (47)
. Likewise,
proteasomes can facilitate export. With ubiquitination (9
, 26)
and possibly other cytosolic molecules, proteasome binding
might be one of the players that act on the cytosolic face of the
membrane, the opponents being ER chaperones and physiological assembly
partners. In agreement with this model, the degradation of µ chains
is slowed down considerably in cells that coexpress L chains, even if
the µ2L2 complexes are not allowed to exit from the ER (28
, 31)
. Interactions with ER resident proteins are likely to
regulate dislocation to the cytosol (49)
. Indeed,
ribophorin mutants that fail to bind calnexin are degraded more rapidly
than their glycosylated counterparts (50)
.
In the ‘tug of war’ scenario, the requirements for active proteasomes
would be more pronounced for proteins (i.e., µ chains) that interact
stably with ER chaperones, such as calnexin and calreticulin, BiP or
PDI (51)
. Further binding studies will be required to
determine the precise role of these and other chaperones in dictating
the fate of individual substrates. It will be of interest to identify
the lumenal, transmembrane and cytosolic components that associate with
different substrates during their translocation across the ER
membrane.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  I. Chantret and S. E H Moore Free oligosaccharide regulation during mammalian protein N-glycosylation Glycobiology, March 1, 2008; 18(3): 210 - 224. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. M. Kelly, J. K. VanSlyke, and L. S. Musil Regulation of Ubiquitin-Proteasome System mediated Degradation by Cytosolic Stress Mol. Biol. Cell, November 1, 2007; 18(11): 4279 - 4291. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. N. Hebert and M. Molinari In and Out of the ER: Protein Folding, Quality Control, Degradation, and Related Human Diseases Physiol Rev, October 1, 2007; 87(4): 1377 - 1408. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. M. Baker and D. Tortorella Dislocation of an Endoplasmic Reticulum Membrane Glycoprotein Involves the Formation of Partially Dislocated Ubiquitinated Polypeptides J. Biol. Chem., September 14, 2007; 282(37): 26845 - 26856. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Meister, U. Schubert, K. Neubert, K. Herrmann, R. Burger, M. Gramatzki, S. Hahn, S. Schreiber, S. Wilhelm, M. Herrmann, et al. Extensive Immunoglobulin Production Sensitizes Myeloma Cells for Proteasome Inhibition Cancer Res., February 15, 2007; 67(4): 1783 - 1792. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. W. Ruddock and M. Molinari N-glycan processing in ER quality control J. Cell Sci., November 1, 2006; 119(21): 4373 - 4380. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. R. Hegde, M. S. Chevalier, T. W. Wisner, M. C. Denton, K. Shire, L. Frappier, and D. C. Johnson The Role of BiP in Endoplasmic Reticulum-associated Degradation of Major Histocompatibility Complex Class I Heavy Chain Induced by Cytomegalovirus Proteins J. Biol. Chem., July 28, 2006; 281(30): 20910 - 20919. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Mattioli, T. Anelli, C. Fagioli, C. Tacchetti, R. Sitia, and C. Valetti ER storage diseases: a role for ERGIC-53 in controlling the formation and shape of Russell bodies J. Cell Sci., June 15, 2006; 119(12): 2532 - 2541. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. A. Obeng, L. M. Carlson, D. M. Gutman, W. J. Harrington Jr, K. P. Lee, and L. H. Boise Proteasome inhibitors induce a terminal unfolded protein response in multiple myeloma cells Blood, June 15, 2006; 107(12): 4907 - 4916. [Abstract] |
) and
21 kDa
(
).
B) N[µ1] were pulsed for 5 min with
[35S]-methionine and cysteine and chased for the
indicated times with (lanes 7–11) or without (lanes 2–6)
ZL3H. Cell lysates were immunoprecipitated with anti-µ
(upper panels) or anti-J (lower panels) and resolved by SDS-PAGE under
reducing conditions (10% acrylamide for the anti-µ IPs, 12% for the
anti-J). Arrows indicate µ and J chain isoforms as described above.
C) NSO cells were treated with ZL3H for the
indicated times in the absence of CHX. Aliquots of the cell lysates
corresponding to similar numbers of cells were loaded onto each lane.
The blot was developed with anti-J. Note the sequential appearance of
the two faster migrating J isoforms, indicated by arrows as detailed
above. D) WEHI-231-J cells (lanes 2–4) were pulsed for
5 min with [35S]-methionine and cysteine and chased for
4 h with or without ZL3H as indicated. Cell lysates
were immunoprecipitated with anti-J and protein A Sepharose. Lane 1
shows the anti-J immunoprecipitate obtained from pulse-labeled
untransfected WEHI-231 cells.
),
possibly reflecting different degrees of intrachain disulfide bond
formation.
).
