| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Division of Endocrinology (P.S.K.), University of Cincinnati College of Medicine, Cincinnati, Ohio 45267; and Division of Endocrinology (P.A.) and Department of Developmental and Molecular Biology, Albert Einstein College of Medicine, Bronx, New York 10461
Correspondence: Address correspondence to: Peter Arvan, M.D., Ph.D., Division of Endocrinology, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, New York 10461.
 |
Abstract |
|---|
 Top Abstract I. IntroductionII. ER Molecular Chaperones,...III. Models of ER...IV. Endocrinopathies as Models...V. Summary: A Proposed...References |
|---|
|
I. Introduction |
|---|
|
TopAbstractI. IntroductionII. ER Molecular Chaperones,...III. Models of ER...IV. Endocrinopathies as Models...V. Summary: A Proposed...References |
|---|
Recent years have witnessed the identification of numerous inborn errors of metabolism that affect secretory or plasma membrane proteins. In many instances, mutations causing minor changes in protein primary structure lead to intracellular retention of the affected proteins, suggesting that proper conformation is critical for protein transport as well as biological activity. Nowadays, scientists are increasingly combining molecular and biochemical analyses in the hopes of identifying precisely how mutations produce folding defects that lead to abnormal protein trafficking. In this report, we review defective protein export as the cause of a variety of endocrinopathies that fall into the category of Endoplasmic Reticulum Storage Diseases (ERSDs) (11). These disorders include certain forms of congenital hypothyroid goiter, osteogenesis imperfecta, diabetes insipidus, familial hypercholesterolemia, and others. In each case, the disease leads to accumulation in the ER of a critically important protein that is unable to reach its target site and therefore is unable to perform its physiologically intended function.
Morphological studies of cells affected by ERSDs routinely reveal expansion and dilation of the ER compartment, which may in part be due to accumulation of misfolded exportable proteins. Moreover, the compensatory response in such cells also frequently includes a selective induction of the synthesis (and supranormal accumulation) of several ER resident proteins that are thought to participate in ER quality control. Most of these proteins are considered to be molecular chaperones, a subtopic that is considered further in Section II.
This review will provide an endocrinologist’s perspective of protein folding in the ER. From this vantage point, we consider how subtle mutations in the coding sequences of polypeptide hormone precursors and other exportable proteins, in conjunction with ER quality control, can lead to defective protein trafficking, causing a disease phenotype. Finally, we review specific representative endocrinopathies in greater detail, as a means to highlight the underlying similarities and differences in phenotypes and modes of transmission of ERSDs, with an eye toward identifying future directions of endocrine investigation.
B. Protein folding in the ER
Until recently, the principle of protein biogenesis relied
entirely on the hypothesis that each peptide chain can self-assemble
into a stable, low free-energy conformation, based solely on
information contained within the primary structure (12, 13). The idea
that few additional factors were required for proper folding was
supported by studies of the renaturation of small polypeptides in
dilute solution at reduced temperature. However, folding of proteins in
living cells (14), especially within the ER compartment, occurs under
highly restrictive conditions unique to this microenvironment (15, 16).
For one thing, the protein concentration in the lumen of the ER may be
as high as 100 mg/ml (17). Second, secretory proteins are translocated
into the ER as they are being translated; thus, the NH2
termini of secretory proteins routinely begin to fold in the ER before
the COOH termini have even been synthesized (18, 19). Third, the ER is
the site for de novo addition of N-linked core
oligosaccharides, as well as initial carbohydrate processing that may
exert both direct and indirect effects on glycoprotein folding (see
Section I.D). Fourth, the composition of ions and small
molecules in the intraluminal environment of the ER is highly
regulated: this includes levels of calcium (20) and ATP (21). Finally,
in mammalian cells, the ER is a more oxidizing environment than the
soluble cytoplasm, owing to the transport of oxidized glutathione (22),
which fosters the formation of disulfide bridges (23). All of these
conditions, but particularly the oxidizing environment and
extraordinary concentration of nascent (unfolded) polypeptides, lead to
an increased possibility for improper intra- and intermolecular
associations.
In spite of these obstacles, a high fraction of newly synthesized secretory and plasma membrane proteins are successfully folded and exported from the ER. Indeed, while a fraction of the initial cohort of newly made exportable proteins may enter novel misfolded states, for endogenous proteins in general, most of the cohort follows a "statistically-most-probable" folding pathway, proceeding through a predictable series of conformational intermediates en route to the native state (24, 25, 26, 27, 28). It is believed that this process has evolved via cotranslational domain-dependent folding (19) in conjunction with the actions of compartment-specific molecular chaperones.
C. Supervised folding: the concept of molecular chaperones and
folding catalysts
It is believed that the overall speed and efficiency of exportable
protein folding is enhanced through a combination of interactions with
another group of proteins resident to the ER. This group includes
members of highly conserved families of molecular chaperones, as well
as others to be described, some of which are viewed primarily as
folding catalysts. Found in every living cell, molecular chaperones
were originally defined as families of proteins that assist in the
self-assembly of other chains but are generally not part of the final
functional unit (29). Thus, by definition, classical molecular
chaperones interact only transiently with their "substrate"
proteins (for further definition, see Section II and Fig. 1
, below). Of the ER molecular
chaperones, several are known to be members of the larger family of
heat shock proteins (HSPs), conserved even to prokaryotes (which lack
ER and other cytoplasmic organelles). Heat shock is only one of several
different types of stress that can cause protein denaturation, and
HSP60, HSP70, and HSP90 classes of heat shock proteins, so named for
their approximate molecular weights, are known to play crucial
compensatory roles that allow cell survival in the face of stress, by
limiting and potentially reversing aggregation of misfolded proteins
(30). In the cells of higher eukaryotes, each intracellular
compartment, including the ER, has its own subset of HSPs and other
chaperones (except for the absence of an identifiable HSP60 class
member in the ER). In the HSP70 class, all members share remarkable
homology in their "substrate"-binding regions—and the same is true
of the HSP90 class. Thus members within a given class tend to differ
mostly within the region of the chaperone that specifies targeting to a
particular subcellular compartment. However, certain metabolic insults
may tend to trigger a chaperone stress response limited more or less
selectively to a specific compartment (31, 32, 33). Indeed, glucose
deprivation is a relatively selective stimulant of the ER stress
response; thus, certain ER chaperones also go under the name of GRPs,
for glucose-regulated proteins (34).
|
The accumulation of proteins in the ER membrane may trigger a second
distinct signal transduction pathway, recently termed the ER-overload
response (36). In this case, increasing presence of either misfolded or
nascent membrane proteins is capable of causing the activation of
another transcriptional factor, nuclear factor-
B, which regulates a
different cascade of gene expression (49). Some forms of cellular
stress can trigger both unfolded protein and ER-overload responses,
while others are selective for only one signaling pathway. In addition
to these transcriptional mechanisms, cells appear to be able to
posttranslationally regulate ER chaperone activity, to a variable
extent, via ADP ribosylation (50), phosphorylation (51, 52, 53, 54, 55), as well as
formation of oligomeric chaperone complexes (56).
To the extent that HSP classes of ER chaperones monitor protein folding, the transcriptional regulation described above is a mechanism designed to ensure that adequate availability of these supervisory molecules is maintained at all times. The importance of available free chaperones is underscored by the recent understanding that even after dissociation from a given chaperone, incompletely folded "substrate" proteins routinely rebind to the same or another copy of that chaperone, contributing to an increase in the bound chaperone fraction. Although the structural basis for the binding of different chaperones appears to vary, the property of cyclic association-dissociation is a common feature ranging from the bacterial chaperonins (57) to ER chaperones that are primarily involved in recognition of polypeptide (58) or carbohydrate moieties (9). Importantly, the fact that molecular chaperones act on "substrate" proteins does not violate the principle of self-assembly. As a rule, classic chaperones interact with many different "substrates" without conferring steric information to influence the final folded structure. However, by interacting with nascent chains, chaperones prevent (and may even reverse) undesirable protein-protein interactions; this increases the chances that newly made proteins will have the opportunity to achieve their native structure. Because ER chaperone associations are based on recognition of features enriched in incompletely folded versions of exportable proteins, associations of ER chaperones are usually (but not always) at their highest levels immediately upon nascent chain translocation into the ER, and terminated before export of the "substrate" protein from this compartment.
Not every interaction with individual ER resident proteins will enhance folding speed or even folding efficiency (see Section III), although some available data tend to suggest enhancement of efficiency at the possible expense of delayed protein folding (59). Promotion of productive folding is likely to be the net effect of interactions with both chaperones and folding catalysts. However, the extent to which roles played by the binding of individual chaperones are overlapping vs. unique (60), and how the different chaperones may cooperate in the folding process for a wide variety of proteins (14), remains largely unknown.
D. Co- and posttranslational modifications are factors that can
influence folding
Although the flow of genetic information ends when the primary
structure has been synthesized, co- and posttranslational
modifications, under the influence of local environmental factors
(discussed in Section I.B), can also affect the folding
outcome for many exportable proteins. Of course, many important
modification steps (e.g., terminal glycosylation, sulfation,
phosphorylation, and dibasic proteolytic cleavage events) take place as
proteins are transported through the Golgi complex, which can
significantly alter protein destination and biological function.
However, with the exception of proteolytic cleavage, Golgi processing
activities generally have fewer effects on underlying protein
conformation than the processing activities of the ER, which are the
subject of this section.
One of the most important ER modifications is the proteolytic cleavage
of the predominantly hydrophobic 
20 amino acid signal sequence (61)
by the signal peptidase complex; the signal peptide is degraded after
it has served to target the nascent chain into the ER translocation
channel (62). Indeed, failure to remove the signal peptide generally
results in severe, irreversible misfolding of secretory proteins (63).
Also, as the nascent chain enters the ER lumen, the rapid collapse of its hydrophobic domains is accompanied by the ordered formation of intramolecular disulfide bonds (26, 64), which stabilize secondary and tertiary structure and can be critical for maintaining a biologically active conformation (65). Indeed, most of the cysteine residues of exportable proteins eventually form disulfides (23), while similar covalent bonds are not observed in cytosolic proteins or in the cytosolic domains of transmembrane proteins.
Within the oxidizing ER environment, reactive thiols also can form mispaired disulfide bonds. Subsequent reshuffling and correction of aberrant disulfide bonds (see Section II.D) may represent one of the rate-limiting steps in protein folding (23). Studies of exportable proteins mutated to lack specific cysteine residues have provided additional evidence for the importance of correct disulfide bond formation. Similarly, treatment of live cells with dithiothreitol or other membrane-permeant reductants results in unfolding of many newly made proteins; upon removal of dithiothreitol, reduced proteins begin to properly refold, reoxidize, and ultimately leave the ER, albeit at a slower rate (66, 67, 68, 69, 70, 71). Intermolecular disulfide bonds may also be important in the formation and maintenance of quaternary structure (28, 72).
One of the next most important ER modifications is the
addition of N-linked carbohydrates to glycoproteins (73).
A large preassembled, oligosaccharide containing two
N-acetylglucosamines, nine mannoses, and three terminal
glucoses (74) is transferred cotranslationally from a dolichol-linked
intermediate to an asparagine residue (of the consensus sequence,
Asn-X-Ser/Thr), as the nascent polypeptide emerges through the
translocation channel in the ER membrane (Fig. 2). ER membrane proteins known as
ribophorins (75), as part of a protein complex encoded by at least
seven genes (76), assist in the catalysis of this initial glycosylation
reaction. Further ER carbohydrate modifications then occur through the
actions of glucosidases and other processing enzymes (Fig. 2, discussed
in Section II.C). Although not found on all exported
proteins, carbohydrate moieties often assist in the folding, stability,
and solubility of nascent exportable polypeptides (77), and in some
cases glycosylation is required for the folding of subunits that occurs
before oligomeric assembly (78). Thus, it is not surprising that
inhibition of N-linked glycosylation frequently leads to misfolding and
aggregation of nascent chains. Once fully folded, however, removal of
sugar groups generally has little impact on protein solubility and
conformation.
|
Other covalent modifications that can also have profound effects
on protein folding may be limited to certain subgroups of exported
proteins. For example, cotranslational hydroxylation of proline within
the tripeptide Gly-Pro-Pro repeats of collagen, catalyzed by an ER
enzyme known as prolyl-4-hydroxylase, is essential for triple-helix
formation and stability (80, 81). Hydroxylation of lysine side chains
also plays an important role in collagen stability (82, 83). Similarly,
carboxylation of glutamyl residues by a vitamin K-dependent mechanism
plays an important structural role in the stability of a number of
blood-clotting factors and calcium-binding proteins (84, 85). Moreover,
covalent attachment of a glycosylphosphatidylinositol anchor near the
carboxy termini of a certain subset of exportable proteins is
associated with the proteolytic cleavage of the C-terminal 20 amino
acids (86, 87); failure to remove these residues can lead to protein
aggregation and transport incompetence (88). Other roles for
glycosylphosphatidylinositol anchors in protein folding and ER export
per se currently remain unknown.
|
II. ER Molecular Chaperones, Folding Catalysts, and Molecular Escorts |
|---|
|
TopAbstractI. IntroductionII. ER Molecular Chaperones,...III. Models of ER...IV. Endocrinopathies as Models...V. Summary: A Proposed...References |
|---|
). In conjunction with the schema in Fig. 1, we define these molecules
as follows: 1) ER chaperones are simply viewed as binding proteins
whose association with exportable protein "substrates" is regulated
by the concentrations of the two components and their binding affinity
in a bimolecular interaction. 2) Folding catalysts are true enzymes,
which also physically interact with substrate proteins, but in so doing
they lower the activation energy required for a discrete conformational
change within an exportable protein. In the secretory pathway,
chaperones and folding catalysts tend to reside predominantly in the
ER, where they serve their primary biological functions, while 3)
transport subunits and molecular escorts routinely accompany their
"substrate" proteins out of the ER, and persistent interaction may
even be required for ER exit. Note that we have not attempted to review
each individual molecule listed in Fig. 1.
A. Binding protein (BiP)
BiP (also known as GRP78), the most studied ER molecular
chaperone, is a member of the HSP70 class (89, 90, 91), a major
calcium-binding protein in the ER lumen (92), and an essential gene
product that even the simplest eukaryotes cannot live without (93, 94).
Like other members of the HSP70 class, BiP possesses a peptide-binding
groove lined with hydrophobic side chains, which is thought to interact
optimally with hepta- or octapeptides enriched in aromatic/hydrophobic
residues in alternating positions — a feature common to many naturally
occurring peptide chains (95, 96, 97). In properly folded globular
proteins, such features are normally buried in the hydrophobic core or
are used at the interface between subunits for protein oligomerization.
This helps to account for the observation that although BiP interacts
with a remarkably wide range of exportable proteins, it tends to
associate most strongly with unfolded, unassembled, or aberrant
polypeptides (25, 98, 99, 100, 101, 102, 103, 104, 105, 106).
In addition to the role of BiP in posttranslational folding, this chaperone has been proposed to be involved in the translocation of nascent polypeptides across the ER membrane (107, 108, 109, 110, 111). Further, BiP plays an indirect but key role in the fusion of nuclear membranes during fertilization between haploid yeast cells (112, 113). Although the physiological relevance of these functions for mammalian cells in vivo remains to be determined (114), it is presumed from these studies that luminal BiP binding serves to stabilize nascent polypeptide chains as well as the luminal domains of certain endogenous ER membrane proteins.
Misfolded exportable proteins, such as those produced by mutation or abnormal glycosylation, have been found to interact with BiP for periods long after their synthesis (e.g., Refs. 115–117). Such prolonged interaction occurs via persistent reassociation of the chaperone (58) to BiP-binding sites that cannot be properly buried within the hydrophobic core of misfolded polypeptides. Because each round of BiP dissociation is coupled to a round of ATP hydrolysis (56, 118), ATP consumption in the ER of cells that have accumulated misfolded polypeptides is expected to increase. Further, misfolding of exportable proteins can be found in cells depleted of ATP (66, 119); this observation has been used to suggest that cyclical binding of BiP and other ATP-dependent chaperones facilitates protein maturation (15). However, it must be pointed out that because ATP hydrolysis is required for polypeptide release from BiP (120), the depletion of ATP—or the use of BiP mutants that cannot hydrolyze ATP (121, 122, 123, 124)—produces defective chaperone dissociation that is likely to inhibit the maturation and ER egress of exportable proteins (see Section III). Thus, the results of experiments using these approaches cannot and do not independently establish that BiP binding normally promotes folding or export from the ER under physiological conditions, although such a conclusion has been suggested based on in vivo protein refolding after heat shock (125).
En route to normal folding, assembly, and exit from the ER,
BiP binding to exportable proteins is typically observed only
transiently (25, 126, 127). Importantly, existence of such transient
interaction also cannot distinguish models in which BiP binding is
thought to promote folding and export from those where it is thought to
delay folding and export (see Section III). In addition,
because individual BiP-binding sites are represented by relatively
small stretches of primary structure, incompletely folded versions of
large polypeptides may expose multiple potential BiP-binding sites
during protein folding, simultaneously or in series (71). For example,
at a moment in time in the steady state, the average stoichiometry of
association between BiP and nascent thyroglobulin (a 330-kDa
monomer) has been estimated at 10:1 in the thyroid ER (25).
Potentially, the earliest folding intermediates of nascent secretory
proteins may bind even more than the average number of BiP molecules,
while later, more folded forms are likely to associate with
progressively fewer BiP molecules.
As described in an earlier section, BiP levels are transcriptionally regulated by the unfolded protein response. Importantly, ordinary physiological dynamics of the production of exportable proteins is sufficient to regulate the synthesis of BiP and other ER chaperones (128, 129, 130).
Lastly, coupled with the recent identification of 14 hsp70 family members in the yeast genome (59), a second, novel hsp70 member in the ER, LHS-1, has been described, which is a nonessential gene that exhibits partial BiP-like function (131). However, it is possible that in mammalian cells the function of the yeast LHS-1 is subserved by its non-hsp70 homolog, GRP170 (132).
B. GRP94
GRP94 (also known as endoplasmin), the product of a single gene
(133), is a member of the HSP90 class (134) and is also a major ER
luminal calcium-binding protein (92, 135) that is transcriptionally
coregulated with BiP under most conditions (34, 35). Considerably less
is known about the peptide-binding specificity of GRP94, but by analogy
to other members of the HSP90 class, GRP94 is likely to act in a
cooperative way (136, 137) in associating with nascent polypeptides
that have exposed unfolded patches. For instance, GRP94 might be one of
the luminal proteins whose binding enhances completion of nascent chain
translocation into the ER lumen (110). Further, BiP and GRP94 are found
in ternary complexes in which they are simultaneously involved in
direct interactions with exportable proteins (138, 139), although their
precise association-dissociation kinetics may not be identical (140).
Like BiP, GRP94 can also be found to interact with misfolded exportable proteins for prolonged periods after synthesis (117, 138, 141, 142, 143), as well as for transient periods with normal proteins maturing in the export pathway (140, 143). Individual polypeptides may expose more than one potential GRP94 binding site during folding, either simultaneously or in series (140). Sequence analysis has revealed two potential ATP-binding sites in GRP94 (53), which exhibits increased binding to polypeptide "substrates" under conditions of ATP depletion (33, 144). This and other observations led to the conclusion that GRP94 binds ATP and exhibits weak ATP hydrolytic activity (21, 145). However, a recent study has demonstrated that GRP94 interactions with peptides are nucleotide-independent, and that ATP binding and hydrolysis are not inherent properties of this chaperone (146). Such an interpretation is of particular interest because an adenine nucleotide-binding site has recently been unambiguously identified in the cytoplasmic homolog, hsp90 (147). Thus, although the molecular mechanism remains poorly understood, it is presumed that GRP94, like BiP, can undergo cycles of unbinding and rebinding to peptide "substrates" (15). It should be noted that thus far (148), such binding of GRP94 has not been demonstrated to result in enhancement of folding or export of proteins from the ER (139).
As noted above, GRP94 levels are transcriptionally coregulated with those of BiP as part of the unfolded protein response (55, 130, 149). However, there are notable examples where disproportionate changes in one chaperone over the other is observed (150), indicating subtleties in transcriptional and/or translational regulation that are not well understood.
C. Calnexin and calreticulin
Calnexin, the only major molecular chaperone of the ER that is an
integral membrane protein (151), is a single-spanning, calcium-binding
protein of the ER membrane (152, 153). The CNX1 gene, a homolog that is
believed to encode a form of calnexin from the yeast
Saccharomyces pombe, is an essential gene whose critical
function is contained within its ER luminal domain (154, 155).
The binding/recognition function of mammalian calnexin has been an area
of intense interest; almost immediately it was realized that treatment
of cells with tunicamycin, a condition that causes severe misfolding of
glycoproteins (and generally causes their increased binding to BiP),
interferes with the binding of many such glycoproteins to calnexin
(153). Nevertheless, calnexin does indeed bind to misfolded, mutant
proteins and to folding intermediates of proteins en route
to export (156, 157, 158). The carbohydrate dependence of calnexin binding
has led to the proposal that calnexin is a lectin (159, 160) serving as
part of a "chaperone apparatus" that includes two independent
enzyme activities: glucosidase II and
UDP-glucose:glycoprotein-glucosyltransferase (UGGT) (161). This
proposal is based upon knowledge of the processing pathway for N-linked
carbohydrates in the ER (Fig. 2).
A 14-saccharide unit is initially added en bloc to N-linked
consensus acceptor sites in glycoproteins. This oligosaccharide
includes two N-acetyl glucosamine (GlcNac) residues anchored
in series to an asparagine in the exportable polypeptide (Fig. 2).
Attached to this disaccharide is a triantennary structure comprised of
nine mannoses (called Man9). At one antenna of the
Man9 are a string of three terminal glucose residues (74).
Normally, the three terminal glucoses are removed from N-linked
oligosaccharides before glycoprotein exit from the ER, by the
sequential action of glucosidase I (which removes the outermost
glucose) and glucosidase II (which sequentially removes the remaining
two glucose residues) (162). However, a single terminal glucose residue
is restored onto Man9 if the UGGT enzyme "senses" the
glycoprotein to be unfolded (163, 164). This sensing involves
interaction of the UGGT enzyme both with exposed hydrophobic residues
as well as an exposed GlcNac residue at the base of the peptide-bound
oligosaccharide (165). Calnexin in turn has been proposed to
preferentially bind to the
Glucose1-Man9-GlcNac2 form of
exportable glycoproteins in the ER (77, 166). All carbohydrate-binding
activity of calnexin resides in its luminal domain (167). Recently,
association of the calnexin luminal domain with monoglucosylated RNase
B in vitro was shown to be very dynamic, suggesting that
calnexin undergoes rapid on/off binding to monoglucosylated
glycoproteins (167). Glucosidase II removal of terminal glucoses is not
likely to occur while glycoproteins are bound to calnexin (167);
however, during the period when glycoproteins have been released from
calnexin, glucosidase II can remove the terminal glucose such that
rebinding to calnexin cannot occur. In this view, cycles of rebinding
to calnexin (168, 169) are triggered solely by the action of the UGGT
enzyme.
In contrast with the view of calnexin acting solely as a lectin in the ER lumen, others have suggested chaperone function involving the transmembrane, nonlectin portion of calnexin (170, 171) or chaperone binding to transmembrane domains of exportable proteins (172). In addition it has been shown that calnexin-"substrate" complexes are not dissociated even when the oligosaccharide is cleaved from the polypeptide upon endoglycosidase digestion (166, 173, 174), suggesting that some interactions with calnexin might occur in a glycan-independent manner (175). Further, certain unglycosylated proteins have also been shown (71, 176, 177) to associate with calnexin. Calnexin binding to unglycosylated proteins occurs commonly with protein aggregates that have been suggested to be separate from the productive maturation pathway (178). However, there is reason to believe that protein aggregates may indeed participate in productive maturation (25, 179); moreover, such aggregates bound to calnexin are reversible in vivo, leading to successful export from the ER (71). Thus, while the capability of calnexin to interact with monoglucosylated glycoproteins is unequivocally established (167), whether this represents the complete story of its role in physiological protein maturation continues to be debated (180, 181).
Calnexin is thought to be in close proximity to nascent chains upon their entry into the ER lumen; thus, it is not surprising that calnexin is hypothesized to play a cotranslational as well as posttranslational role in the folding of exportable proteins (71, 182). During the cotranslational period, calnexin binding may act to shield reactive free thiols from forming incorrect disulfide bonds, although this improvement in folding efficiency may actually slow down kinetic progression through the folding pathway (168, 183). Of course, these features are not discordant with observations that posttranslational rebinding of calnexin to monoglucosylated side chains (184) is also correlated with productive folding of exportable glycoproteins (169).
Calreticulin (181, 185), a homolog of calnexin, is another major calcium binding protein (92, 135), the topology of which is entirely luminal. Although its function as part of an ER chaperone apparatus has been studied far less than that of calnexin (186), it shares with calnexin the capability of recognizing monoglucosylated glycoproteins, with partially overlapping specificity (187, 188). Calreticulin and calnexin both go through cycles of binding and release, indicating roles in monitoring and assisting protein folding (168, 169). Importantly, however, calreticulin is a full participant in the transcriptionally regulated, ER unfolded protein response (189), a feature noted to be absent for calnexin (151). Calreticulin also plays a role in intracellular calcium signaling (190, 191), in transcriptional regulation of steroid-sensitive gene expression (192), and in a number of other important cellular functions (reviewed separately in Ref.185).
Calmegin is the newest member of the calnexin family, found only in the testis, where it is thought to perform a function in the folding of proteins that are necessary for the ability of sperm to adhere to and fertilize eggs (193).
D. Disulfide isomerase and prolyl isomerase: families of folding
catalysts
Protein disulfide isomerase (PDI), another major component of the
ER lumen, is a true foldase, in that it catalyzes thiol-disulfide
interchange with a broad substrate specificity, and it shows strong
homology with bacterial thioredoxin (23, 194, 195). The regulation of
PDI synthesis overlaps only partially with those of BiP, GRP94,
calreticulin, and other ER residents that exhibit the unfolded protein
response (189, 196). Instead, PDI expression seems to be proportional
to the flux of nascent chains into the ER [i.e., especially
high in secretory tissues (197, 198, 199)], and its activity may be further
regulated posttranslationally (200). Recent reports indicate that PDI
may undergo dimerization, autophosphorylation, and ATP hydrolysis that
is stimulated in the presence of denatured polypeptides (201). PDI
increases both the rate and efficiency of proper folding and export of
proteins that contain disulfide bonds (195, 202).
PDI1 is a gene essential for viability in yeast (203). There are different opinions regarding the significance of this observation. First, critical to the disulfide isomerase catalytic activity are two -CXXC- motifs. However, polypeptide binding by PDI has been reported not to involve these motifs (204), and PDI also has been shown to exhibit the potential for assisting folding of proteins that do not possess disulfide bonds (205). Further, certain deletions in PDI that leave residual disulfide isomerase activity are nevertheless lethal, while cells carrying a variant PDI in which both-CGHC-active sites are disrupted (i.e., no measurable isomerase activity in vitro) remain viable (206) and able to assist in protein folding (207). On the other hand, the ER lumen appears to have other proteins (discussed further, below) that can function as disulfide isomerases in the absence of PDI1 enzyme activity (208). With this in mind, it is interesting that Escherichia coli thioredoxin can complement null mutants of yeast PDI only if thioredoxin is mutated to contain a reactive CXXC motif (209). These and similar studies have led some to conclude that the essential function of PDI is in fact to unscramble nonnative disulfide bonds (210). Nevertheless, PDI is recognized to be multifunctional; it is for instance a well-recognized subunit of the prolyl-hydroxylase complex that is involved in collagen synthesis (211), and it heterodimerizes with the 97-kDa subunit of the microsomal triglyceride transfer protein complex (212).
Peptidylprolyl isomerase (PPI) catalyzes cis-trans isomerization of proline side chains, and enhances the rate of protein folding in vitro (213). PPIs are also termed immunophilins and are comprised of cyclophilins (which bind the immunosuppressant cyclosporin A) and FKBPs (which bind the immunosuppressant FK506). The ER luminal FKBPs are transcriptionally regulated with other members of the unfolded protein response (149, 214).
E. ERp72 and ER60
ERp72, one of the more recently described luminal chaperones
(215), is also a calcium-binding protein in the ER (135, 208) and is a
member of the PDI superfamily (see above). Although ERp72 is not an
essential gene product, its overexpression can rescue nonviable cells
deficient for PDI (216). Indeed, ERp72 contains three copies of the
-CXXC- active site motif found in PDI (215). While it seems plausible
that ERp72 may be able to exhibit limited PDI-like activity (208, 217),
it also may be that its ability to rescue cells lacking PDI could be
associated with the reported ability of ERp72 to assist in the
degradation of proteins that cannot fold in the ER (218). ERp72 has
been shown to interact with misfolded versions of exportable proteins
(33, 219, 220), but evidence for its interaction with normal protein
folding intermediates in vivo is not well established (143).
Nevertheless, ERp72 is clearly regulated with other proteins exhibiting
the ER unfolded protein response (196, 221).
ER60/calregulin, which also contains thioredoxin-like sequences and thus shares homology with ERp72 and PDI, has been implicated in the ER-associated degradation of misfolded proteins (222). Although initially hypothesized to be a phosphoinositide-specific phospholipase C (223), this has not been independently confirmed. By contrast, although additional study is clearly needed, some evidence for a thiol-dependent protease activity of ER60 is accumulating (224). Moreover, it has recently been reported that processing of N-linked carbohydrates in exportable protein "substrates" may be required for their association with ER60 (225), which has been proposed to act in conjunction with either calnexin or calreticulin (226).
F. HSP47
Heat shock protein 47 (HSP47) is a collagen-specific stress
protein that performs chaperone function during folding and assembly of
newly synthesized procollagen molecules (227). Because type I
procollagen forms a triple helix beginning at the carboxy terminus (see
Section IV.C), much of its folding must occur after
translocation of the nascent chain has been completed. Early
association of HSP47 is facilitated by its binding to the
amino-terminal globular domain of the collagen propeptide (228), but
the chaperone appears to remain associated throughout most, if not all,
subsequent stages of tertiary structural maturation (229, 230).
Although it may act in concert with other, more ubiquitous ER
chaperones (144, 231), the services of HSP47 appear to be unique to
cells secreting collagen and its homologs (232). Because of this, HSP47
falls into a "gray area" of protein-specific ER chaperones, as is
the case with the microsomal triglyceride transfer protein (see Fig. 1
and Section IV.D, below).
G. Molecular escorts: pro-peptides, transport subunits,
receptor-associated protein (RAP), and 7B2
Figure 1 provides a schematic to categorize a number of additional
polypeptides that provide helper function to the protein export
pathway, yet should nevertheless be distinguished from conventional ER
chaperones. These include the propeptides of many polypeptide hormones,
transport subunits, and a subgroup of proteins for which we propose to
adopt the name "molecular escorts" (233). There are growing numbers
of examples of proteins in each of these subgroups, but we mention only
one or two representative examples. We emphasize that these
distinctions may change over time, as molecular information about the
roles of helper proteins in the folding pathways of individual proteins
becomes clearer.
Importantly, not all polypeptides that participate in the folding and
trafficking of exportable proteins are ER residents. Specifically, the
propeptide regions of many exportable proteins themselves serve
primarily a structural, rather than a functional, role (234). Thus,
folding in the absence of the propeptide, or in the presence of a
mutated propeptide, may perturb the conformational maturation and ER
export of polypeptide hormone precursors (235). Such proregions have
been referred to as intramolecular chaperones, but a distinction should
be made between this idea and classic ER chaperones (Fig. 1) which
(except under unusual circumstances) are not transported down the
secretory pathway, and whose turnover in the cell is far slower, and
which routinely associate with folding intermediates of more than one
kind of "substrate" protein.
Not unlike the situation with the propeptides are subunits of
oligomeric proteins that function primarily in polypeptide transport
and stabilization, while playing a relatively minor role in subsequent
biological activity. A case in point is the 
-subunit of the
glycoprotein hormones (LH, FSH, CG, TSH), which combines noncovalently
with ß-subunit in the ER (236). Both subunits must be partially
folded before subunit assembly (64, 237, 238, 239), and the specific regions
used by - and ß-subunits to combine have begun to be mapped
(240, 241, 242, 243). When expressed by themselves, ß-subunits tend to be
relatively poorly secreted, whereas when coexpressed with , a much
higher fraction of ß is secreted, as heterodimers (244). Although a
heterodimeric structure (and therefore, the presence of a subunit) is
needed for biological activity (245), the evidence indicates that the
-subunit, which is common to all of these hormones and is therefore
unlikely to provide biological specificity, plays an important role in
export of the different glycoprotein hormones, in particular for FSH
and for TSH (246).
Another subgroup of molecules that should be distinguished from conventional ER chaperones is represented by the "molecular escorts" (233). For illustrative purposes, we review only two members of this subgroup.
RAP is a 40 kDa polypeptide which interacts shortly after the
biosynthesis of low-density lipoprotein receptor (LDLR)-related protein
(LRP), and RAP travels with LRP molecules, having the potential to
remain associated on the cell surface (247). By itself, RAP is a
protein that is believed to be retained within the ER (248). Further,
LRP in the absence of RAP is also defective for ER exit (233). However
RAP association with LRP allows both partners to exit the ER (249),
presumably by maintaining a given LRP in a favorable conformation (250)
and by preventing premature association of LRP ligands in the ER (251),
which could lead to receptor retention and/or degradation. Instead,
dissociation of RAP, when it occurs, takes place after the proteins
have reached the cell surface, where LRP ligands can displace the
escort.
7B2 exhibits restrictive expression in neuroendocrine tissues, where it
is synthesized as a 25-kDa precursor protein and is released from
the regulated secretory pathway as processed proteolytic fragments. The
carboxy-terminal domain of the pro7B2 protein is responsible for
inhibiting the prohormone processing activity of prohormone convertase
2 (PC2) (252, 253), while the larger N-terminal portion of processed
7B2 contains independent stimulatory actions on the catalytic activity
of PC2 (254). Evidently, several different domains of the pro-7B2
protein may interact with the pro-form of PC2 (255). Recently, it was
determined that pro-7B2 and pro-PC2 assemble (in a 1:1 stoichiometry)
in the ER of neuroendocrine secretory cells, and this assembly leaves
the ER together (256) before dissociation in the most distal portions
of the Golgi complex (257, 258). Because pro-PC2 may be transported
through the secretory pathway more slowly or less efficiently in the
absence of pro-7B2 (259), pro-7B2 has been called a chaperone (256) but
may be better viewed as a molecular escort. Such a distinction also
represents a "gray area" since recombinant 7B2 may be able to
associate with at least one other unrelated protein, based on in
vitro studies (260).
|
III. Models of ER to Golgi Traffic Influence Models of Quality Control |
|---|
|
TopAbstractI. IntroductionII. ER Molecular Chaperones,...III. Models of ER...IV. Endocrinopathies as Models...V. Summary: A Proposed...References |
|---|
Although quality control monitoring of the secretory pathway may not occur exclusively in the ER, as described in the Overview (Section I.A), the ER contains mechanisms that are intended to help cells decide which proteins are ready for export and which are not. Since some newly synthesized proteins are destined to be permanent residents of the ER while others are destined for forward transport, at some point there must be a partitioning of proteins that remain in the ER from those that advance further (262). There is currently considerable debate about the mechanisms by which this sorting is achieved. We believe this may also bear on the question of why many misfolded secretory proteins are not exported. The following two sections propose two very different hypothetical mechanisms by which ER exit of exportable proteins might occur, and it remains possible that none, all, or parts of both mechanisms may be correct, or that other mechanisms not considered in this review (263) may come into play. Nevertheless, for the present, such hypotheses are necessary to begin to understand diseases of protein folding and secretion (264).
A. Escape from ER retention as one hypothesis to explain
anterograde protein traffic from the ER
Over the past decade, a dominant view has been that no specific
signals are required for soluble proteins to undergo forward transport,
while luminal resident proteins are maintained in the ER by specific
mechanisms, including both prevention from forward trafficking (265) as
well as retrieval of ER residents that have escaped (262, 266).
Retrieval via retrograde vesicular transport to the ER is the better
understood mechanism (267) and involves a receptor, Erd2p (268), that
binds to a C-terminal recognition motif comprised of the sequence
K-D-E-L or a close variant, which is common to luminal ER resident
proteins. A short cytosolically disposed motif serves a similar
retrieval function for ER transmembrane proteins (266, 269, 270).
However, retrieval may not be the major mechanism that retains ER
resident proteins, because even in the absence of a KDEL tail, luminal
ER molecular chaperones are very slow to exit the ER (271). This ER
retention has been attributed in some way to the calcium binding and
calcium levels in the ER lumen, possibly resulting in the formation of
an insoluble protein matrix (265). Indeed, mutation of the
calcium-binding domain of calreticulin has a profound effect on the
egress of this molecular chaperone, even when the KDEL sequence, and
receptor, remain intact (272). Moreover, new genes are now being
described that play roles in KDEL-independent retention of luminal ER
resident proteins, although these mechanisms are still not well
understood (273).
Thus, it is currently thought that the primary mechanism maintaining the localization of ER luminal resident proteins involves their direct retention within this compartment, with the KDEL signal serving as a back-up system for retrieving escaped ER chaperones. With this in mind, secretory proteins must exhibit at least two properties for their successful exit from the ER. First, in general, they must express neither an ER retention domain nor a C-terminal retrieval motif used by the luminal ER resident proteins. Second, they must not bind tightly or extensively to ER chaperones that are themselves anchored in the ER. In other words, in this model, for secretory proteins to exit the ER, they must escape from the clutches of anchored ER chaperones. Moreover, this model predicts that the residence time of newly synthesized secretory proteins will be directly proportionate to the extent of their binding to ER chaperones (139). This feature could explain why the average length of time that newly synthesized exportable proteins spend in the ER differs for each protein species (274, 275).
B. Cargo receptors as another hypothesis to explain anterograde
protein traffic from the ER
In the past 2 yr, an alternate sorting model of protein exit from
the ER has emerged (276). In this view, at or just before the time of
budding of a newly formed ER transport vesicle, certain proteins are
selectively extracted by receptors for entry into these ER transport
vesicles (277). Support for such a model is based on the idea that
cytosolic proteins, which are known to coat the surface of budding
membrane vesicles, are likely to be recruited by cytoplasmically
disposed domains of transmembrane receptors, while the luminal domains
of these receptors in turn could recognize specific ligands on the
luminal side of the membrane (278, 279). This kind of model has strong
precedent in post-Golgi trafficking (280), e.g., in the
delivery of newly synthesized lysosomal proenzymes as well as endocytic
ligands via coated vesicles (281, 282, 283). Between the ER and Golgi
compartments, two different kinds of multicomponent coat protein
complexes, known as COPI (277) and COPII (261), have been described to
be recruited to the cytosolic surface of membranes in preparation for
vesicle budding (284)—and neither class of coated vesicle includes ER
resident proteins (285, 286, 287). These features have provided an incentive
to search for anterograde sorting signals (3) and putative cargo
receptors captured by transport vesicles (288).
Recently, in mammalian cells, ERGIC53, a recycling membrane protein that functions as a mannose-binding lectin (289, 290), has been hypothesized to extract secretory proteins from the ER (262). In addition, in yeast, at least one ER membrane component, EMP24, has been postulated as a potential recycling receptor for selected exportable proteins (291). Further, loss of the yp24A protein, homologous to EMP24, is thought to inhibit the formation of transport vesicles (292). Moreover, the recent identification of other homologous proteins in higher organisms, and the demonstration that members of this p24 family of transmembrane proteins express short cytosolically disposed motifs necessary to recruit coat protein complexes (288) has led to the hypothesis that the p24 family represents one class of long-sought cargo receptors.
However, it must be pointed out that the formation of coated transport vesicles does not even require the presence of luminal contents (293). Certainly, lectin-like sorting interactions like those proposed for ERGIC53 (294) cannot solely account for the ER exit rate of secretory proteins in mammalian cells, where many secreted proteins (e.g., proinsulin) are unglycosylated. Moreover, in yeast, evidence indicates that the phenotype associated with a loss of EMP24, one of the p24 family, is not restricted to secretory protein transport but includes altered cellular handling of luminal ER chaperones as well as protein disulfide isomerase (273). Thus, the cargo receptor hypothesis, while gaining increasing interest, still has a number of obstacles to overcome.
C. What provides quality control of ER export?
As noted in the Overview (Section I.A), the ER quality
control machinery is designed to try to prevent the export of
incompletely/improperly folded versions of exportable proteins. This
might happen by one of several different mechanisms. For one, unfolded
proteins may aggregate and become insoluble, thereby becoming unable to
advance into the lumen of ER transport vesicles (295). However, not all
unfolded versions of exportable proteins are insoluble, or even
aggregated. In these cases, hypotheses designed to explain how
selective protein export is prevented depend largely on which model is
favored for the mechanism of normal protein export (reviewed above). If
anterograde traffic out of the ER requires presentation of certain
features to cargo receptors, then it is possible that unfolded proteins
fail to present the required features and thus cannot be recognized and
cannot be carried forward into Golgi-bound transport vesicles. In such
a case, ER chaperones might serve a role in promoting sufficient
folding to assist in the presentation of exportable proteins to cargo
receptors. Alternatively, if anterograde traffic requires escape from
ER chaperones, then it is possible that unfolded proteins fail to bury
chaperone recognition sites. In that case, unfolded patches exposed on
exportable proteins serve as de facto ER retention signals,
because they either promote the formation of insoluble aggregates
(which cannot advance because of intrinsic biophysical properties) or
binding to ER chaperones (which prevent protein advance because the
chaperones are anchored in the ER—see Section III.A). If
the latter view is correct, a corollary is that for all proteins
retained in the ER that are not intrinsically insoluble, the retained
protein must be bound to one or more chaperones in order that its
export be prevented. With these hypotheses in mind, it is worth
reviewing what is known about the binding of ER chaperones with respect
to helping or hindering protein export.
Based on current knowledge, there are two kinds of studies suggesting
the promotion of protein export as a consequence of binding to ER
chaperones. First is the case of calnexin and calreticulin (169, 183),
which associate with monoglucosylated carbohydrate side chains of a
wide variety of glycoproteins (167, 296) such that association is
abrogated by pretreatment of cells with castanospermine or
deoxynorjirimicin (inhibitors of carbohydrate processing that prevent
formation of monoglucosylated core sugars on glycoproteins, see
Section II.C). In this case, drug treatment can be clearly
shown to diminish the folding and transit of a subset of glycoproteins
from the ER to the Golgi complex (297, 298). A second method
capitalizes on the fact that most ER chaperones have a much longer
half-life in the cell (
1 day) than the secretory proteins with which
they interact (minutes to hours). Thus, treatment of cells with
cycloheximide, an inhibitor of protein synthesis, can allow for the
drainage of previously synthesized exportable proteins from the ER,
while the ER chaperones remain at normal concentrations. This causes
the fraction of unoccupied chaperones to increase. Because
cycloheximide effects are reversible, the drug can be washed away and
the resumption of protein synthesis allows new exportable proteins to
be introduced into an ER that now has an increased availability of
chaperones. In such a case, aggregation of newly synthesized secretory
protein has been observed to be diminished (25). By this means,
increased availability of BiP and other chaperones might enhance
protein maturation and traffic through the anterograde transport
pathway (299).
On the other hand, many of the ER chaperones described in Section
II have been specifically implicated in the retention of
exportable proteins within the ER. Indeed, association of BiP with Ig
heavy chains (90) and light chains (101) has been directly correlated
with their failure to undergo export from the ER, and indeed, the loss
of BiP-binding sites from surface or secreted Igs restores their
ability to undergo intracellular transport (300, 301)—even if they are
incompletely folded or assembled. Similarly, soluble (truncated) forms
of the T cell receptor -chain do not aggregate but exist as
monomers, and yet they are not secreted; instead, they coprecipitate
with BiP, and manipulations that cause BiP dissociation allow for
-chain secretion in vivo (302). Also along this spectrum,
unassembled subunits of oligomeric membrane proteins are typically
retained in the ER bound to BiP, but they often undergo high molecular
weight aggregation that may prevent their entry into ER export vesicles
(303, 304).
The situation with calnexin appears to be analogous: the T cell
receptor ß-chain is prevented from ER export in the absence of
assembly with -chain, and the retained ß-chain is bound to
calnexin (305). For MHC class I heavy chains that have not yet
assembled with ß2-microglobulin and antigenic peptide,
export from the ER is impeded by association with calnexin (306), even
while the assembly of heavy chains with ß2-microglobulin
is promoted (183). Similarly, dissociation from calnexin parallels the
egress of MHC class II heterotrimers from the ER, suggesting that the
chaperone is involved in retention of unfolded/unassembled subunits
(307). Moreover, in yeast cells, deletion of the calnexin homolog CNE1
does not lead to a loss of cell viability but, rather, inhibits ER
retention of selected proteins that normally cannot exit this
compartment (308).
Similar reports regarding retention of exportable proteins within the
ER have been attributed to GRP94 (138) and other ER chaperones (232).
These reports are especially common in ERSDs (caused by the presence of
mutant forms of exportable proteins) where the ER unfolded protein
response causes remarkable induction of the synthesis of these ER
chaperones to levels 1 order of magnitude above the normal
range. Certainly, in ERSDs, a higher than normal number of these
chaperones are likely to be bound to exportable proteins entrapped in
the ER. However, ERSDs cannot be used to provide definitive evidence
regarding the effects of chaperone binding on the retention of
exportable proteins, because it is difficult to ascribe how much of the
retention phenotype could be due to underlying abnormalities associated
with the mutation of the exportable protein, rather than effects of
increased chaperone binding. For this reason, scientists have attempted
to independently express BiP or other ER chaperones via constitutive
promoters, to increase the free level of ER chaperones in the absence
of an ERSD, with the intention of increasing chaperone binding to
exportable proteins. When BiP is made increasingly available in the ER,
this is sufficient to blunt or block the ER unfolded protein response
in Chinese hamster ovary (CHO) cells when an experimental stress is
imposed (309). Equally importantly, increased availability of BiP
inhibits export from the ER of those secretory proteins known to bind
BiP (309). Analogous results have recently been reported upon increased
expression of GRP94 (139).
Retention of exportable upon increased expression of BiP or GRP94 might be explained by a saturation of the ER, causing the diminished availability of folding promoters and other key resident proteins in the ER (i.e., "chaperone imbalance"). As an alternative, increased chaperone expression might simply cause increased complex formation with susceptible "substrate" proteins. The former idea has been rendered less likely by recent demonstrations that the levels of a wide array of ER resident proteins are not changed as a consequence of selective overexpression of an individual ER chaperone (139, 309a). By contrast, experimental evidence favoring the alternative view is the fact that decreased expression of BiP (using antisense methodology, which does not diminish the levels of foldases or other ER resident proteins) actually increases the ER export of certain heterologous proteins expressed in CHO cells (119, 310). Moreover, these effects are selective, as overexpression of BiP, or even overexpression of mutant BiP, which cannot undergo ATP-dependent release of a "substrate" protein, fails to cause the retention of secretory proteins that lack a demonstrable BiP-binding site (124).
The message transmitted by these results is that to execute their role with respect to protein export, chaperones must not only associate with unfolded secretory proteins for a period of time—they must also dissociate during the folding process (25). As described in Section I.C, repeated cycles of binding and release to the same or different unfolded polypeptides is a common function of molecular chaperones in all compartments. This has left open the hotly debated question of what stage in the chaperone-binding cycle does the polypeptide folding actually takes place: during the period of chaperone association or during the period of release (311)? There is good reason to believe that chaperone binding, at least in some circumstances, may decrease the assembly of oligomeric subunits (168), decrease the folding of monomers, and actually promote monomer unfolding (69). Although semantically inappropriate, such behavior has been termed an "anti-chaperone" function of ER chaperones and is dependent upon the particular "substrate" and the availability and stoichiometry of chaperone binding (103).
Thus, to summarize this section, quality control of ER export (9) may be provided by any of three different mechanisms: 1) inability of an exportable protein to properly present itself to cargo receptors mediating ER exit (an increasingly popular model but one for which direct evidence is still lacking); 2) formation of protein aggregates that are biophysically unable to enter transport vesicles (for which indirect support has been obtained in only a few cases); and 3) repetitive, cyclical binding by ER chaperones, which mediates ER retention and does not necessarily facilitate folding in all cases (for which there is the largest amount of available evidence, albeit indirect). The latter model has obvious implications for ERSDs (see below), and suggests that successful export from the ER takes place only when all chaperone-binding sites on the exportable polypeptide are buried, or when ER chaperones are otherwise disabled or overwhelmed. Thus, according to this model, one simple way to think about the role of ER chaperones in the regulation of protein flow out of the ER may be the dam concept, in which the dam (comprising all available ER chaperones) serves as a central regulator. The "height" and "tightness" of the dam, represented by the levels of ER chaperones and their respective affinities for "substrate"-binding sites, regulates the escape of exportable proteins from the ER. The rigor of ER quality control is related to both parameters. Ideally, ER quality control should be sufficiently zealous to retain only mutant proteins that may have lost their primary biological activity or that may have other unwanted toxic effects. However, as we shall see further in Section IV, this quality control is not perfectly efficient in all situations. Indeed, ER quality control machinery may exhibit drastically different retention properties for different mutant protein subunits whose primary structures differ from the native primary structure only by the loss of a single free cysteine or a single disulfide bond, in some cases even leading to export that is augmented over that of the wild-type protein (79a, 312, 313), but in most cases leading to a diminution of protein export.
D. ER-associated degradation
In cells, both wild-type and mutant versions of exportable
proteins are subject to misfolding under normal or stressed conditions,
although the fraction that is irreversibly misfolded may vary (315).
The development of a system that is not only able to recognize
irreversibly misfolded proteins, but to target them for degradation, is
an essential function for cell survival, because inexorable
accumulation of undegraded misfolded proteins in the ER is likely to
clog the secretory pathway and become toxic to cells (316). A central
question to resolve for the coming decade is: although the lysosomal
compartment of cells is primarily designed for macromolecular
digestion, how do misfolded versions of exportable proteins that cannot
readily reach the lysosome (because of ER quality control restraints)
undergo degradation? Unlike autophagy (317, 318, 319, 320, 321), a phenomenon known as
ER-associated degradation, or ERAD (322), has been defined by 1)
insensitivity to classic inhibitors of lysosomal proteases and
intracellular transport blockers, 2) immunolocalization of degradative
substrates in the ER, as well as 3) lack of Golgi-type sugar processing
on the substrates (323, 324).
At this stage, most of the machinery responsible for ERAD remains largely unidentified in higher eukaryotes. With few possible exceptions (143, 324), proteolytic fragments are generally not observed as intermediates in the ERAD process, complicating its characterization. In some instances, the ER compartment itself has been implicated as the site of degradation (325), suggesting that at least some of the ERAD machinery is contained within this compartment. Further, evidence has been presented that redox events, or the availability of free thiols, may strongly influence the activity of ERAD machinery (326, 327). Indeed, ER60 and ERp72, which contain copies of the -C-X-X-C- motif conserved in PDI (described above) have been postulated either to be potential cysteine proteases or molecules that may target misfolded proteins to ERAD (218, 222, 223, 224). Additional new genes are being found that apparently influence this process (328).
One recently proposed mechanism of ERAD that is gaining great attention is the idea of dislocation of proteins from the ER membrane (329, 330, 331), or the ER lumen (322, 332), presumably via reverse translocation through the ER translocon (333), i.e., back to the cytosol, for selective protein degradation by the ubiquitin-proteasome proteolytic pathway (334, 335, 336). The steps involed in this process are far from being completely worked out, but may include initial partial proteolysis within the ER, ubiquitination, extraction from the ER, and the removal of all previously attached N-linked carbohydrates, before complete proteasomal digestion (10, 337).
Interestingly, even though essentially all exportable proteins that fail to escape the ER are eventually degraded, several studies have revealed that some proteins turn over rapidly while others disappear more slowly (338). As molecular chaperones play major roles in assisting or preventing protein degradation in other cellular compartments (339), it seems likely that they may play similar diverse roles within the ER. At this early stage in our knowledge, it should perhaps not be surprising that BiP and calnexin association have been correlated with the destruction of misfolded exportable proteins by ERAD (105, 322, 340, 341) just as others have suggested that binding of these very chaperones helps to protect from ERAD (168, 342, 343). In addition, inhibitors of proteasomal proteolysis may cause the accumulation of undegraded, misfolded secretory proteins, leading to induction of ER chaperone synthesis as part of the unfolded protein response (344). More work is clearly needed to understand how molecular chaperones help to distinguish misfolded proteins targeted for ERAD from normal early folding intermediates of exported proteins and, more specifically, the relationships between ERAD and the quality control of ER export. Whatever the answers, it would appear that ER chaperones are likely to be intimately involved.
|
IV. Endocrinopathies as Models of Defective Protein Export |
|---|
|
TopAbstractI. IntroductionII. ER Molecular Chaperones,...III. Models of ER...IV. Endocrinopathies as Models...V. Summary: A Proposed...References |
|---|
Export from the ER represents the rate-limiting step in the overall process of secretion of Tg (139). Much of the time after biosynthesis is used to convert Tg into a transport-competent form, which involves multiple processing steps within the ER (346). Cotranslationally, collapse of hy
