Cotranslocational Degradation Protects
the Stressed Endoplasmic Reticulum
from Protein Overload
Seiichi Oyadomari,1 Chi Yun,1 Edward A. Fisher,2,3 Nicola Kreglinger,2,3 Gert Kreibich,3
Miho Oyadomari,1 Heather P. Harding,1,4 Alan G. Goodman,5 Hanna Harant,6 Jennifer L. Garrison,7
Jack Taunton,7 Michael G. Katze,5 and David Ron1,2,3,*
1
Skirball Institute of Biomolecular Medicine, New York University School of Medicine, New York, NY 10016 USA
Department of Medicine, New York University School of Medicine, New York, NY 10016 USA
3
Department of Cell Biology, New York University School of Medicine, New York, NY 10016 USA
4
Department of Pharmacology, New York University School of Medicine, New York, NY 10016 USA
5
Department of Microbiology, School of Medicine, University of Washington, Seattle, WA 98195 USA
6
Novartis Institutes for BioMedical Research, A-1235 Vienna, Austria
7
Department of Cellular and Molecular Pharmacology, University of California, San Francisco, San Francisco, CA 94107 USA
*Contact: [email protected]
DOI 10.1016/j.cell.2006.06.051
2
SUMMARY
The ER’s capacity to process proteins is limited, and stress caused by accumulation of unfolded and misfolded proteins (ER stress) contributes to human disease. ER stress elicits
the unfolded protein response (UPR), whose
components attenuate protein synthesis, increase folding capacity, and enhance misfolded protein degradation. Here, we report
that P58IPK/DNAJC3, a UPR-responsive gene
previously implicated in translational control,
encodes a cytosolic cochaperone that associates with the ER protein translocation channel
Sec61. P58IPK recruits HSP70 chaperones to
the cytosolic face of Sec61 and can be crosslinked to proteins entering the ER that are delayed at the translocon. Proteasome-mediated
cytosolic degradation of translocating proteins
delayed at Sec61 is cochaperone dependent.
In P58IPK / mice, cells with a high secretory
burden are markedly compromised in their ability to cope with ER stress. Thus, P58IPK is a key
mediator of cotranslocational ER protein degradation, and this process likely contributes to ER
homeostasis in stressed cells.
INTRODUCTION
Early steps in the biogenesis of secreted and transmembrane proteins take place in association with the endoplasmic reticulum (ER). Newly synthesized polypeptides
or segments thereof are translocated into the ER lumen,
where chaperones and protein-modifying enzymes promote their conversion into functional proteins that exit
the organelle (Sitia and Braakman, 2003; Anken et al.,
2005). The ER’s capacity to process proteins is relatively
limited, and the stress caused by accumulation of unfolded and misfolded proteins (ER stress) contributes to
a number of important human diseases (Kaufman, 2002).
Cells respond to ER stress by activating an unfolded
protein response (UPR), which limits new protein synthesis and promotes the expression of genes that enhance
the organelle’s capacity to process unfolded proteins (Patil and Walter, 2001). Degradation of misfolded proteins
represents a third mechanism by which cells limit the burden of unfolded proteins in their ER and assumes great importance when ER function is compromised. A dedicated
machinery targets misfolded proteins to a retrotranslocation channel for subsequent proteasomal degradation in
the cytosol. Early work suggested that the anterograde
translocation of nascent proteins and retrograde translocation of completed misfolded ones were mediated by
the same channel; however, it is presently believed that
these two processes are separated in time and place
and that the degradation of misfolded ER proteins is
largely a posttranslocational event (Tsai et al., 2002;
Meusser et al., 2005).
ApoB100 is a notable exception to this rule. This large
secreted protein normally assembles with lipids in the lumen of the ER of liver cells to form a lipoprotein particle
(Fisher and Ginsberg, 2002; Hussain et al., 2003). Lipidation is required for folding of apoB100. A block in lipidation
leads to cotranslocational degradation of apoB100, as
the protein remains associated with the Sec61 channel
(Mitchell et al., 1998; Pariyarath et al., 2001). Furthermore,
whereas the extraction of conventional misfolded lumenal
proteins from the ER is mediated by the multifunctional
ATPase p97/Cdc48 (Ye et al., 2001), the degradation of
Cell 126, 727–739, August 25, 2006 ª2006 Elsevier Inc. 727
apoB100 requires cytosolic chaperones of the HSP70 and
HSP90 class (Fisher et al., 1997; Gusarova et al., 2001).
The dependence of apoB100 degradation on cytosolic
chaperones and the association of this process with the
Sec61 channel present intriguing similarities to the process by which binding of lumenal HSP70 promotes the
translocation of proteins into the ER (Matlack et al.,
1999). The latter process requires a translocon-associated DnaJ domain containing cochaperone to stimulate
ATP hydrolysis and peptide binding by the lumenal
HSP70 (Misselwitz et al., 1998); however, no such factor
has been found for cotranslocational degradation of
apoB100.
P58IPK was first discovered as an inhibitor of the virally
induced eIF2a kinase PKR (Barber et al., 1994). The encoding gene P58IPK/DnaJC3 was subsequently found to
be UPR inducible (Lee et al., 2003), and P58IPK also binds
and inhibits the ER stress-inducible eIF2a kinase PERK
(Yan et al., 2002; Van Huizen et al., 2003). Given the latter’s
role in attenuating protein synthesis in ER-stressed cells
(Harding et al., 1999), these findings implicated P58IPK in
recovery of protein synthesis following dissipation of ER
stress. However, the phenotype of the P58IPK knockout
is not easy to reconcile with a role restricted to attenuating
eIF2a phosphorylation (Ladiges et al., 2005).
P58IPK/DNAJC3 contains a functional DnaJ domain that
promotes ATP hydrolysis by HSP70 chaperones (Melville
et al., 1999). The bulk of P58IPK consists of nine tetratricopeptide (TPR) repeats by which it binds eIF2a kinase(s)
(Lee et al., 1994) and the highly conserved C-terminal segments of cytosolic HSP70 and HSP90 proteins (Melville
et al., 1999; Brychzy et al., 2003). The study below was designed to address the possibility that, in addition to its
known role as an inhibitor of eIF2a kinases, P58IPK might
also collaborate with cytosolic chaperone networks that
function in ER-stressed cells.
RESULTS
P58IPK’s PERK-Independent Function
PERK’s C. elegans homolog, pek-1, is required for the survival of worms lacking the IRE1 (ire-1) arm of the unfolded
protein response (Shen et al., 2001). Given that P58IPK inhibits PERK (Yan et al., 2002), we were surprised that, not
only did attenuation of worm’s P58IPK homolog (dnj-7) by
RNAi fail to rescue the phenotype of (partial) inactivation of
pek-1 in an ire-1 mutant strain (Figures 1A5, 1A6, and
1A7), but also, on its own, dnj-7(RNAi) markedly reduced
the fitness of ire-1 mutant animals (Figure 1A8). To follow
up on this hint of pek-1 independent function(s) of DNJ-7/
P58IPK, we compared the effects of dnj-7(RNAi) in wildtype and pek-1-deleted worms. dnj-7(RNAi) increased
basal activity of the ER stress-inducible hsp-4::gfp (BiP)
reporter in both wild-type and pek-1-deleted worms. Furthermore, dnj-7(RNAi) enhanced hsp-4::gfp expression in
worms exposed to the ER stress-promoting drug tunicamycin, regardless of their pek-1 genotype (Figures 1B
and 1C).
728 Cell 126, 727–739, August 25, 2006 ª2006 Elsevier Inc.
These observations indicate that DNJ-7/P58IPK’s role in
ER-stressed cells is not restricted to antagonizing PEK-1/
PERK’s kinase activity. Furthermore, growth retardation
by dnj-7(RNAi) in the mutant ire-1 background suggested
that attenuated DNJ-7/P58IPK expression affects HSP-4
expression indirectly, by increased burden of misfolded
proteins rather than directly by enhancing UPR signaling.
Subsequent experiments were designed to understand
these PERK-independent functions of P58IPK.
P58IPK Is a Peripheral Membrane Protein of the
Rough ER that Associates with SEC61
and HSP70 Chaperones
Recombinant P58IPK expressed in COS1 cells colocalized
with ER markers (Yan et al., 2002). We confirmed this colocalization by immunostaining the endogenous protein in
mouse fibroblasts (Figure S1A in the Supplemental Data
available with this article online) and extended this analysis by examining P58IPK’s association with subcellular
fractions in mouse liver, a rich source of the protein. The
vast majority of P58IPK from the postmitochondrial supernatant was recovered in the buoyant membrane fraction
by equilibrium density centrifugation. This association
was disrupted by carbonate extraction, which indicates
that P58IPK is a peripheral membrane protein (Figure 2A).
Hepatocytes have well-differentiated smooth and rough
ERs, which are easily separable. P58IPK comigrated with
the rough ER markers Sec61a and ribophorin I following equilibrium density gradient centrifugation (Figure 2B).
Interestingly, P58IPK was also membrane associated in
PERK / cells (Figure S1B), suggesting that membrane
anchoring of P58IPK entails interactions with a different
rough ER transmembrane protein.
P58IPK liberated from membranes by digitonin and other
mild detergents remained associated with a relatively
large heterogeneous complex that migrates through a velocity gradient as a broad peak (Figure 2C). Unfortunately,
none of the available anti-P58IPK antisera or antibodies
recognize the protein in this intact native complex, precluding its immunopurification. However, following denaturation and complex disassembly, P58IPK is readily recognizable both by the previously described monoclonal
antibodies and by the polyclonal antiserum we raised to
the bacterially expressed protein (data not shown).
To further characterize this membrane-associated
P58IPK-containing complex, we radiolabeled wild-type
and P58IPK / fibroblasts, solubilized the membranes in
detergent, and added the lysine-reactive reversible crosslinker DSP to the lysate. After quenching the crosslinker,
the complex was denatured in urea and SDS, which dissociates noncovalently bound partners and exposes P58IPK
to our antibody. The denaturants were diluted and the
radiolabeled P58IPK was immunoprecipitated along with
covalently bound partners. The addition of crosslinker
yielded a P58IPK-dependent doublet of 40 kDa, along
with other weaker bands (Figure 3A, compare lanes 4
and 5 with lanes 11 and 12). This doublet comigrated
with radiolabeled Sec61a, a central component of the
Figure 1. PERK/pek-1-Independent Activity of P58IPK/dnj-7
(A) Photomicrographs obtained 72 hr after
hatching of populations of wild-type (WT) and
ire-1 mutant (ire-1mut) C. elegans that developed from fertilized eggs placed on bacterial
lawns expressing the indicated RNAi gene constructs. The percentage of animals in the population that reached the L4 stage of development by 72 hr is indicated below the image.
(B) Fluorescent photomicrographs of the GFP
signal of hsp-4::gfp transgenic animals of
wild-type (WT) or pek-1-deleted (pek-1D) genotypes exposed to the indicated RNAi constructs or tunicamycin (30 mg/ml for 16 hr).
(C) Immunoblots of GFP protein and endogenous HSP-4/BiP (detected with anti-HDEL antibody) in lysates of untreated or tunicamycintreated hsp-4::gfp transgenic animals of the
indicated genotype and RNAi exposure. The
asterisk marks an unidentified protein detected
by the anti-HDEL antibody.
translocon (Figure 3A, lanes 17 and 18) and was reactive
with anti-Sec61a antiserum, as revealed by the contents
of an anti-Sec61a immunoprecipitation performed sequentially on the anti-P58IPK immunoprecipitate after reversal of the crosslink (Figure 3A, lane 15). Furthermore,
P58IPK copurified with digitonin-solubilized native translocons, isolated by immunoprecipitation with a Sec61b antiserum in the absence of crosslinker (Figure 3B). These
observations suggest that P58IPK molecules are associated with the translocon.
The tetratricopeptide repeats (TPR) in P58IPK closely resemble those found in TPR2, a protein of unknown function that binds avidly to HSP70 chaperones. P58IPK has
also been noted to form a complex with HSP70 in vitro
(Melville et al., 1999), an observation that we confirmed
(data not shown). TPR-containing proteins bind the conserved C-terminal tails of cytoplasmic HSP70 chaperones
(Scheufler et al., 2000), and consistent with this observation, P58IPK was co-eluted with HSP70 by excess ATP
from an ATP-agarose affinity matrix loaded with detergent-solubilized ER membranes (Figure 3C). We also
compared the colocalization of HSP70 chaperones with
translocon components in velocity gradients prepared
with digitonin-solubilized ER membranes from wild-type
and P58IPK / mouse liver. This mild solubilization procedure, which preserves the integrity of the translocon (Gorlich et al., 1992), yielded abundant comigrating HSP70 in
the wild-type but not in the P58IPK / sample (Figure 3D)
and suggests that P58IPK plays an important role in recruiting chaperones to the cytosolic face of the ER translocon.
P58IPK and Cotranslocational Degradation
of ER Proteins
The retrotranslocation of misfolded/misassembled proteins from the ER and their degradation in the cytoplasm
requires numerous proteins on the cytosolic face of the
ER (Tsai et al., 2002; Meusser et al., 2005); it is also a process whose disruption is predicted to promote accumulation of misfolded proteins in the ER lumen. Therefore, we
measured the effect of the P58IPK genotype on degradation of ectopically expressed T cell receptor a chain
(TCRa), a well-characterized indicator of the activity of
the ER retrotranslocation/degradation apparatus (Yu
et al., 1997). In pulse-chase labeling experiments, newly
synthesized TCRa decayed with similar kinetics in transfected wild-type and P58IPK / cells (Figures 4A and
4B). While the P58IPK genotype did not affect the decay
of mature TCRa after the pulse, the amount of labeled
Cell 126, 727–739, August 25, 2006 ª2006 Elsevier Inc. 729
Figure 2. P58IPK Is a Peripheral Membrane Protein of the
Rough ER
(A) Immunoblots of P58IPK, ribophorin I (RPN 1, an integral ER membrane marker), and eIF2a (a cytosolic marker) in postmitochondrial supernatant (PMS) or membrane fraction (MB) of liver from untreated and
tunicamycin-treated mice; where indicated, the floated membranes
were extracted with Na2CO3.
(B) Immunoblot of P58IPK, ribophorin I (RPN I), and Sec61a (a rough ER
[RER] marker) and Cyp1A1 (a smooth ER [SER] marker) content of fractions of a self-generated OptiPrep equilibrium gradient loaded with
membranes procured by floatation from tunicamycin-treated mouse
liver.
(C) Immunoblots of ribophorin I (RPN I), protein disulfide isomerase
(PDI, a lumenal ER marker), and P58IPK in fractions of a glycerol (velocity) gradient of digitonin-solubilized membranes of tunicamycintreated mouse liver. Where indicated, the solubilized membranes
were further exposed to RIPA or Na2CO3 before loading on the gradient. The migration of globular proteins of known size through the gradient is indicated.
TCRa present at the end of the (short) pulse was reproducibly higher in the P58IPK / cells. This is unlikely to reflect
differences in transfection efficiency or global protein synthesis, as accumulation of radiolabeled neomycin phosphotransferase (NPT, encoded by a separate gene on
730 Cell 126, 727–739, August 25, 2006 ª2006 Elsevier Inc.
the TCRa plasmid) was similar in the two genotypes. Furthermore, the inhibitory effect of P58IPK on the accumulation of 35S-labeled nascent TCRa following a brief labeling
pulse was recapitulated in transfected COS1 cells (Figure S2). Together, these experiments suggested that
P58IPK is dispensable for the posttranslocational degradation of mature TCRa but affects early step(s) in the biogenesis of ER proteins, prompting us to further explore the
latter aspect.
Interfering with lipidation by blocking the microsomal
triglyceride transfer protein (MTP) disrupts apoB100
translocation into the ER and renders it a substrate for
an unconventional degradation process that initiates at
the translocon and results in the delivery of the nascent
chain to the cytoplasm for proteasomal degradation
(Fisher and Ginsberg, 2002). Given evidence that P58IPK
functions in early steps of ER protein biogenesis, we
chose to explore its effects on apoB100 degradation. Primary hepatocytes from wild-type and P58IPK / mice
were pulse-chase labeled in the presence or absence of
an MTP inhibitor. The MTP inhibitor markedly reduced
the level of full-length apoB100 that accumulated at the
end of the pulse and destabilized the protein. This destabilization was prevented by treating cells with a proteasome inhibitor, as reported previously (Fisher et al.,
1997). These effects of the MTP inhibitor were attenuated
in the P58IPK / cell, as reflected in the accumulation of
more full-length apoB100 at the end of the pulse and
marked stabilization of the protein during the chase (Figures 4C and 4D). ApoB48, an alternative product of the
same gene whose biosynthesis is less dependent on lipidation, was less affected by the P58IPK mutation, and the
P58IPK genotype had no effect on total protein synthesis
in the presence or absence of the MTP inhibitor (data
not shown). Furthermore, MTP inhibition increased the
amount of newly synthesized apoB recovered by sequential co-immunoprecipitation in a crosslinked complex with
P58IPK (Figure 4E). Together, these observations suggest
a direct role for P58IPK in the degradation of apoB100
molecules whose translocation is compromised by defective lipid transfer.
ApoB100 is a special, ligand-dependent secreted protein; however, the observation that the amount of TCRa
accumulating at the end of a short labeling pulse was
also negatively affected by the expression of P58IPK
(Figure 4A and Figure S2) suggested that the cochaperone
may contribute to the degradation of conventional ER
proteins early during their biogenesis. However, detection
of the nonglycosylated TCRa intermediate, predicted to
serve as P58IPK’s substrate in this degradation process,
was unreliable (data not shown). Therefore, we sought
a more robust system in which to test whether P58IPK’s
role in cotranslocational degradation is restricted to
apoB100 or if the cochaperone contributes generally to
the degradation of other ER proteins that translocate
inefficiently.
Derivatives of a naturally occurring cyclodepsipeptide,
HUN-7293, have been reported to selectively block the
Figure 3. P58IPK Is Associated with the ER Translocon and Cytosolic HSP70 Chaperones
(A) Autoradiograph of metabolically labeled P58IPK and associated crosslinked proteins immunoprecipitated from wild-type (P58+/+) and P58IPK
knockout (P58 / ) mouse fibroblast lysate with the indicated antisera. The crosslink was reversed, and the individual radiolabeled proteins in half
of the immunoprecipitated sample were resolved by SDS-PAGE (lanes 1–14). The other half was subsequently immunoprecipitated with antiserum
to Sec61a (lanes 15 and 16) and loaded alongside an anti-Sec61a immunoprecipitate of uncrosslinked labeled lysate (lanes 17 and 18). The position of
P58IPK and Sec61a bands is indicated.
(B) Immunoblot of P58IPK and Sec61a in complexes recovered by an anti-Sec61b immunoprecipitate from digitonin-solubilized purified membrane
fractions of liver from tunicamycin-treated wild-type (+/+) and P58IPK knockout ( / ) mice. Anti-CHOP antiserum was used to control for the specificity of the immunoprecipitation procedure, and the Ponceau S stain of the blot reveals the levels of immunoprecipitating IgG heavy chain (IgG HC) in
the samples.
(C) Immunoblot of HSP70 and P58IPK content of complexes purified by ATP-agarose affinity chromatography, or agarose affinity chromatography (as
a specificity control) from detergent-solubilized membranes of tunicamycin-treated mouse liver. The complexes were eluted with MgCl2 and ATP as
indicated.
(D) Immunoblot of P58IPK, Sec61a, HSP70, and protein disulfide isomerase (PDI) in fractions of a sucrose velocity gradient loaded with rough ER
membranes purified from liver of tunicamycin-treated wild-type (P58+/+) and P58IPK knockout (P58 / ) mice and solubilized with digitonin under
conditions that preserve translocon integrity.
Cell 126, 727–739, August 25, 2006 ª2006 Elsevier Inc. 731
Figure 4. P58IPK Is Required for Degradation of Nonlipidated apoB100 in Hepatocytes
(A) Autoradiograph of 35S-labeled HA epitopetagged T cell receptor a chain (TCRa) ectopically expressed in transfected wild-type
(P58+/+) and P58IPK knockout (P58 / ) mouse
fibroblasts and immunoprecipitated from the
cell lysate at the end of the 30 min metabolic labeling pulse or after the indicated period of cold
chase. Radiolabeled neomycin phosphotransferase II (NPT II) expressed from the same plasmid serves as an internal control.
(B) Plot of the TCRa signal from (A).
(C) Autoradiograph of apoB immunoprecipitated from wild-type and P58IPK / cultured
primary hepatocytes in a pulse-chase labeling
experiment conducted in the presence or absence of a microsomal transfer protein inhibitor
(MTP-I) or a proteasome inhibitor (P-I). The positions of apoB100 and apoB48 are indicated.
(D) Plot of the apoB100/apoB48 signal from (C).
(E) Quantification of radiolabeled apoB associated with P58IPK following crosslinking with
DSP and recovery by sequential immunoprecipitation (mean ± SEM), first with anti-P58IPK
serum and then (after reversal of the crosslink)
by anti-apoB serum. Where indicated, the cells
were exposed to oleic acid (OA, to promote lipidation) or MTP inhibitor (MTP-I, to block lipidation) and proteasome inhibitor (P-I) to block
apoB degradation. The crosslinker (DSP) was
omitted in sample 5, and the specific antiP58IPK and anti-apoB sera were replaced by
the irrelevant anti-CHOP serum in samples 6
and 7. The amount of radiolabeled (apoB) protein recovered in the sequential immunopurification, measured by scintillation counting, is
expressed as a fraction of the total radiolabeled
apoB, recovered by direct immunoprecipitation from a parallel sample. The inset is an immunoblot of P58IPK and ribophorin I, demonstrating that the experimental conditions used
here did not affect protein levels in these cells.
translocation of human vascular cell adhesion molecule-1
(VCAM-1) into the ER. The stalled VCAM-1 (which lacks
glycans and migrates faster than the mature form on
SDS-PAGE) is rapidly degraded normally but accumulates
at the translocon if proteasomes are inhibited (Besemer
et al., 2005; Garrison et al., 2005). To determine if P58IPK
affects the fate of human VCAM-1 in cells treated with
a translocation inhibitor, we compared the effects of an
inhibitor, CAM741 (Besemer et al., 2005), on the fate of
metabolically labeled human VCAM-1 expressed from a
recombinant retrovirus in wild-type and P58IPK / mouse
fibroblasts. Micromolar concentrations of CAM741 led to
accumulation of a higher-mobility form of human VCAM1 in the mutant but not in the wild-type cells (Figure 5A,
lanes 5 and 6). A species of indistinguishable mobility accumulated in both wild-type and mutant CAM741-treated
732 Cell 126, 727–739, August 25, 2006 ª2006 Elsevier Inc.
cells when exposed to a proteasome inhibitor (Figure 5A,
lanes 7 and 14). The latter corresponds to the labile, nonglycosylated pre-VCAM-1 whose translocation is blocked
by the inhibitor (Besemer et al., 2005; Garrison et al.,
2005). A slightly faster-migrating species, presumed to reflect translocated but unglycosylated VCAM-1, whose
signal peptide had been cleaved, was observed in cells
treated with the glycosylation inhibitor tunicamycin
(Figure 5A, lane 15; and Figure S3). Furthermore, at lower
concentrations (50–200 nM) CAM741 was less effective in
blocking the accumulation of mature, glycosylated
VCAM-1 in cells lacking P58IPK (Figure 5A, compare lanes
2–4 with 9–11).
COS1 cells express low levels of detectable endogenous P58IPK (Figure 5B); therefore, we examined the effect
of enforced expression of P58IPK on the fate of VCAM-1 in
Figure 5. P58IPK Promotes Cotranslocational Extraction of Human VCAM-1 in
Cells Exposed to a Specific Translocation Inhibitor
(A) Autoradiogram of metabolically labeled human VCAM-1 expressed from a recombinant
retrovirus in P58IPK+/+ and P58IPK / mouse fibroblasts. The cells were exposed to the indicated concentration of CAM741 (‘‘CAM’’), tunicamycin (Tm), or proteasome inhibitor (P-I) for
the duration of the 4 hr labeling pulse, followed
by lysis and immunoprecipitation with a rabbit
immune serum that recognizes both translocated (glycosylated [‘‘G’’]) and nontranslocated
(nonglycosylated [‘‘N’’]) VCAM-1. The mature
VCAM-1 signal was quantified in each lane
and is expressed as a fraction of the signal in
untreated cells.
(B) Autoradiogram of metabolically labeled
human VCAM-1 expressed by transfection in
COS1 cells or coexpressed together with wildtype P58IPK (FL) or mutant P58IPK lacking the effector J domain (DJ). The cells were treated with
CAM741 as in (A). The lower panel is an autoradiograph of P58IPK (full-length [FL] and mutant
[DJ]), immunopurified from the flow-through of
the VCAM-1 immunopurification.
(C) Autoradiogram of radiolabeled VCAM-1 associated with P58IPK in cotransfected COS1
cells, following crosslinking with DSP and recovery by sequential immunoprecipitation, first
with anti-P58IPK serum and then (after reversal
of the crosslink and denaturation of the sample)
by anti-VCAM-1 serum. The cells were exposed
to CAM741 (1 mM to inhibit VCAM-1 translocation) and proteasome inhibitor to block degradation of the stalled VCAM-1. CAM741 was
omitted in lane 2, the crosslinker (DSP) was
omitted in lane 3, the specific anti-P58IPK and
anti-VCAM sera were replaced with the irrelevant anti-CHOP serum in lanes 7 and 8, and
a truncated form of VCAM-1 lacking the signal
peptide (VCAM-1DSP) was used in lanes 9 and
10. Lanes 11–15 are of VCAM-1 immunopurified from the flow-through of lanes 1–3 and
9–10, as indicated. The asterisk marks an unidentified protein that is variably immunopurified with the VCAM-1 antisera and may represent a degradation product.
these cells (characterization of VCAM-1 proteins in COS1
cells is presented in Figure S3). Wild-type, but not mutant
P58IPKDJ (lacking the effector J domain), sensitized COS1
cells to the inhibitory effects of CAM741 on mature human
VCAM-1 expression (Figure 5B), indicating that P58IPK
can be limiting for the inhibitory effect of CAM741.
To further characterize the interaction between P58IPK
and VCAM-1, we determined the effect of CAM741 on
the recovery of VCAM-1 in a crosslinked complex with
P58IPK by sequential immunoprecipitation from radiola-
beled COS1 cells that had been treated with a proteasome
inhibitor. A crosslinker-dependent and antisera-dependent radiolabeled band of similar mobility to nonglycosylated VCAM-1 was recovered in complex with P58IPK in
cells treated with CAM741 (Figure 5C). Neither mature
VCAM-1 nor a truncated VCAM-1 lacking the signal peptide interacted with P58IPK, arguing against postlysis associations. Together, these observations point to a direct role
for P58IPK in extraction of nascent human VCAM-1 from
the translocon under conditions in which translocation is
Cell 126, 727–739, August 25, 2006 ª2006 Elsevier Inc. 733
partially inhibited and in delivery of the stalled protein to
the downstream degradation machinery when translocation is completely inhibited.
An ER Stress Disease Model Uncovers
a Requirement for P58IPK
To determine if P58IPK’s role in degrading proteins that are
delayed at the translocon correlates with a physiologically
significant impairment in coping with ER stress, we turned
to the P58IPK knockout mice. As an isolated mutation in
a C57BL/6 background, P58IPK+/ animals are normal
and P58IPK / mice have a very mild phenotype, manifesting as a mild delay in postnatal growth in both sexes and
subclinical glucose intolerance developing in males by
6 months of age, which progresses to fasting hyperglycemia at 1 year of age in some animals (Ladiges et al., 2005;
Figure 6).
The ‘‘Akita’’ mutation, C96Y in the A chain of the Ins2
gene product, leads to pro-insulin misfolding and compromised ER function. The b cell adapts to this stress by
activating the UPR, and on the C57BL/6 background
Ins2+/C96Y ‘‘Akita’’ animals are fairly well compensated;
females are virtually asymptomatic and males develop
progressive glucose intolerance culminating in a welltolerated nonketotic chronic hyperglycemia (Oyadomari
et al., 2002). Induction of P58IPK is part of the adaptation
of Ins2+/C96Y ‘‘Akita’’ b cells to ER stress, as reflected in elevated P58IPK staining in the affected endocrine pancreas
(Figure 6A). Therefore, we chose to examine the importance of P58IPK in this well-characterized ER stressmediated disease model.
Compared to Ins2+/C96Y;P58IPK+/ mice (which are
themselves indistinguishable from Ins2+/C96Y;P58IPK+/+),
Ins2C96Y/+;P58IPK / mice were dramatically compromised: their growth as pups was severely retarded, and
overt hyperglycemia set in within days after weaning in
both sexes (Figure 6B). The elevated blood glucose was
not met by a corresponding increase in plasma insulin;
rather, plasma insulin levels were very low (Figure 6C),
a finding that correlates with dramatic loss of islet mass
(Figure 6D). These findings indicate that P58IPK has an important role in preserving function of ER-stressed b cells.
The phenotype of the compound mutant mice is consistent with a role for P58IPK-mediated cotranslocational
degradation in promoting survival of ER-stressed b cells.
However, the evanescence of the islets in the compound
mutant mice precluded more detailed characterization of
P58IPK’s role in that system. Therefore, to explore the potential effect of ER stress on P58IPK activity, we compared
the physical interaction between P58IPK and metabolically
labeled nascent proteins in unstressed and ER-stressed
cultured fibroblasts. Confining the analysis to nascent
proteins favors detection of interactions between P58IPK
and its clients (over interactions with radiolabeled mature
cofactors). To avoid the confounding effects of changing
levels of P58IPK, we induced ER stress with a brief pulse
of thapsigargin (an agent that adversely affects the protein
folding environment in cells within minutes of application)
734 Cell 126, 727–739, August 25, 2006 ª2006 Elsevier Inc.
and we minimized the inhibitory effect of ER stress on
protein synthesis by conducting the experiment in the
presence of serum (Harding et al., 1999).
Cells were untreated or pretreated with thapsigargin for
30 min followed by a brief 10 min labeling pulse conducted
in the presence of a low concentration of puromycin. The
latter integrates randomly into radiolabeled nascent
chains, releasing them from the ribosome with a covalent
bound C-terminal puromycin moiety that is reactive with
a specific antiserum (Zhang et al., 1997). Following solubilization with digitonin and crosslinking with DSP, the lysate
was denatured in 2% SDS, disrupting noncovalent interactions. The SDS was diluted and P58IPK (and associated
radiolabeled proteins) was immunoprecipitated. The antiP58IPK immune complex was disrupted with 2% SDS,
crosslinks reversed with DTT, and the radiolabeled, nascent, puromycinylated proteins were captured by a
second immunoprecipitation reaction with an antisera to
puromycin and quantified by scintillation counting. Interestingly, exposure to thapsigargin resulted in a 1.6-fold
increase in the fraction of puromycinylated (nascent) proteins recovered in complex with endogenous P58IPK
(Figure 7A).
Several controls were included: the signal detected by
sequential immunoprecipitation in cells radiolabeled in
the absence of puromycin was negligible (Figure S4),
and the experiment was conducted in P58IPK / cells, in
which thapsigargin did not increase recovery of puromycinylated proteins in the anti-P58IPK immune complex. Exposure of wild-type cells to arsenite, an agent that mimics
the effects of ER stress on the translational apparatus
(reflected in eIF2a phosphorylation [see inset]) but causes
little or no ER stress, very modestly increased complex
formation (1.1-fold). Finally, combining, before crosslinking, lysates made from radiolabeled, thapsigargin, and
puromycin-treated P58IPK / cells and nonlabeled lysates
from thapsigargin-treated wild-type cells yielded a low
background signal, suggesting that the interactions immortalized by the crosslinking procedure do not reflect
a postlysis artifact. While other interpretations cannot be
excluded, the observation that most of the endogenous
P58IPK in cells is associated with RER membranes (Figure 2), suggests that the puromycinylated proteins associated with it are translocation substrates. This experiment
is therefore consistent with an ER stress-mediated enhancement of P58IPK’s interactions with its client proteins.
DISCUSSION
Previously characterized as a stoichiometric inhibitor of
the stress-inducible eIF2a kinases PKR and PERK,
P58IPK emerges from these studies as having a broader
role in the ER.
P58IPK and Degradation of Stalled
Translocation Substrates
Studies of several model proteins have shown that their
degradation occurs after dissociation from the translocon
Figure 6. Loss of P58IPK Sensitizes Mice
to ER Stress-Induced Diabetes Mellitus
(A) Immunostaining of P58IPK and insulin in sections of pancreas obtained from wild-type
(Ins2+/+) and ‘‘Akita’’ mutant mice (Ins2+/C96Y).
The H33258 signal reports on all the nuclei in
the field.
(B) Time-dependent changes in blood glucose
levels and body weight of male and female
mice with the indicated P58IPK and Ins2 genotypes. Shown are means ± SD (n = 6; *p <
0.05 for Ins2C96Y;P58+/ versus Ins2C96Y;
P58 / ). Note that the compound mutant animals all died between 6 and 12 weeks.
(C) Plots of plasma insulin and blood glucose of
3-week-old female mice. Shown are means ±
SD (n = 5; *p < 0.05).
(D) Photomicrographs of the pancreas of
3-week-old female mice with the indicated
genotypes.
and release into the lumen (in the case of soluble proteins)
or after lateral gating and partitioning into the lipid bilayer
(in the case of integral membrane proteins). A retrotranslocation apparatus with components conserved in all eukaryotes identifies the misfolded protein and recruits it
to a channel that is likely distinct from the translocon.
Following channel engagement and with the assistance
of cytosolic partners, the misfolded protein is retrotranslocated, polyubiquitinated, and delivered to the proteasome
for degradation (Tsai et al., 2002; Meusser et al., 2005). We
have no evidence that P58IPK contributes to this process;
rather, P58IPK helps degrade proteins at the translocon, as
Cell 126, 727–739, August 25, 2006 ª2006 Elsevier Inc. 735
Figure 7. Increased Association of Endogenous P58IPK with Nascent Polypeptides in ER-Stressed Cells
(A) Quantification of radiolabeled nascent (puromycinylated) proteins associated with endogenous P58IPK following crosslinking with DSP
and recovery by sequential immunoprecipitation, first with anti-P58IPK serum and then (after
reversal of the crosslink and dissociation of the
immune complex) by anti-puromycin serum.
Where indicated, mouse fibroblasts of wildtype and knockout P58IPK genotype were
treated with thapsigargin (Tg, 100 nM, to induce ER stress) or arsenite (25 mM, as a cytosolic stress control). The crosslinker (DSP)
was omitted from sample 3, and radiolabeled
extracts from / cells were combined with
nonlabeled extracts from +/+ cells prior to
crosslinking in sample 7 (labeled <=>). The
amount of radiolabeled puromycinylated protein recovered in the sequential immunopurification was measured by scintillation counting
and is expressed as a fraction of the total radiolabeled puromycinylated proteins, recovered
by direct immunoprecipitation. Shown are the
mean and range of an experiment performed
in duplicate and repeated four times. The inset
is that of a P58IPK and P-eIF2a blot of lysates of
cells exposed to thapsigargin or arsenite.
(B) Cartoon depicting P58IPK’s hypothesized
role at the translocon. In unstressed cells with
ample reserves of lumenal chaperones (BiP
and associated J domain cofactors), protein
translocation proceeds efficiently. The ribosome-translocon junction is closed, and
P58IPK is inactive (left panel). In ER-stressed
cells (right panel), accumulation of misfolded
protein challenges lumenal chaperone reserves. The resulting inefficient protein translocation favors disruption of the ribosome-translocon seal
and exposure of the cytosolic portion of the translocating polypeptide to P58IPK. The latter recruits cytosolic HSP70s to the translocon and presents
the client protein to them, thus promoting its extraction from the lumen through ATP hydrolysis, possibly by a ratcheting mechanism involving sequential binding of HSP70 molecules.
exemplified by its role in the degradation of apoB100
and VCAM-1.
ApoB100 degradation occurs while the protein remains
associated with the translocon (Mitchell et al., 1998; Pariyarath et al., 2001); however, the process is incompletely
understood. Especially obscure is the mechanism by
which cytosolic proteins such as the chaperones known
to assist in apoB100 degradation (Fisher et al., 1997; Gusarova et al., 2001) might gain access to the polypeptide.
Engagement of the substrate at the translocon shields it
from reactive probes applied to the cytosol side, findings
that imply a tight ribosome-translocon junction (Crowley
et al., 1994; Jungnickel and Rapoport, 1995). However,
mounting evidence suggests that the seal on the ribosomal side might be modulated by events in the channel
or possibly more distally in the lumen of the ER. For example, the signal sequence can influence the nascent chain’s
accessibility on the cytosolic side; when appended to
a model translocated substrate, the efficient signal sequence from prolactin specifies a tight ribosome-translocon junction, whereas the weaker signal sequence of the
736 Cell 126, 727–739, August 25, 2006 ª2006 Elsevier Inc.
prion-related protein specifies a weaker seal (Fons et al.,
2003). Translocating apoB is likewise accessible to
probes applied to the cytosolic side (Hegde and Lingappa,
1996).
While the relationship of accessibility on the cytosolic
side to lipidation (and folding) on the lumenal side has
not been explored directly, it is conceivable that misfolding associated with impaired lipidation retards translocation and disrupts the seal at the ribosome translocon junction. This, in turn, is predicted to expose the protein to
P58IPK, which is supported by the observation that inhibitors of lipidation increase the amount of radiolabeled apoB
recovered in complex with P58IPK (Figure 4E). We propose
that P58IPK uses its C-terminal DnaJ cochaperone domain
to promote HSP70 chaperone binding to exposed regions
of proteins whose translocation is delayed for any reason.
Cycles of P58IPK-coordinated cytosolic chaperone binding extract the misfolded substrate from the translocon
by a ratcheting mechanism similar to that used by lumenal
chaperones to promote anterograde translocation (Matlack et al., 1999).
The fate of VCAM-1 in cells exposed to low concentrations of the translocation inhibitor suggests that translocation and P58IPK-mediated extraction (and degradation) are
competing processes and that the partitioning of the substrate between these two fates is sensitive to the level of
P58IPK (and therefore modulated by the UPR). Accumulation of cytosolic VCAM-1 was only observed in knockout
cells exposed to high concentrations of the inhibitor; at
lower concentrations of inhibitor reduced expression of
mature VCAM-1 was also observed in the mutant cells,
but this was not associated with the accumulation of the
nonglycosylated form (Figure 5A, lanes 3 and 4). These
findings are consistent with the existence of (saturable)
P58IPK-independent mechanisms for degrading substrates that fail to translocate and would otherwise
accumulate on the cytosolic face of the ER membrane.
Cotranslocational Degradation and ER Stress
Cotranslational translocation can proceed without lumenal chaperones, at least in vitro (Gorlich and Rapoport,
1993). However, their contribution to the efficiency of
the process is suggested by observations that BiP and
its translocon-associated J domain-containing cofactor,
Sec63, are required in an in vitro reconstituted system
that translocates proteins posttranslationally (Matlack
et al., 1999) and in an in vitro system for cotranslational
translocation (Brodsky et al., 1995). Furthermore, inactivation of BiP/Kar2p by a ts mutation leads to accumulation of
translocation substrates in the yeast cytosol in vivo (Vogel
et al., 1990). Based on these observations we propose
that, under conditions of ER stress, translocation is compromised by limited lumenal chaperone reserve. As a consequence, diverse ER proteins, delayed at the translocon,
become substrates for extraction and degradation mediated by P58IPK and its cytosolic HSP70 partners. This
speculation is supported by the observation that ER stress
increases the association of nascent (puromycinylated)
proteins with P58IPK (Figure 7A). Therefore, we propose
that the induction of P58IPK in ER-stressed cells represents an attempt to shift the equilibrium in favor of extraction from the translocon and cytosolic degradation of
proteins as outlined in cartoon form (Figure 7B).
P58IPK knockout results in increased misfolded protein
burden in the ER lumen and higher levels of ER stress signaling (Figure 1B and Figure S5). The markedly impaired
survival of b cells of P58IPK / ;Ins2+/C96Y ‘‘Akita’’ mice,
compared with ‘‘Akita’’ mice with a functional P58IPK allele, is consistent with the importance of cotranslocational
degradation in protecting the stressed ER of secretory
cells (though other roles for P58IPK cannot be excluded).
This protective effect might reflect a reduction in protein
flux into the stressed ER’s lumen. Alternatively, as the
folding of complex proteins often entails interactions between distant segments on the linear polypeptide chain,
stalled translocating proteins remain unfolded and engage
chaperones on the lumenal side. The first line of defense
against this threat is the PERK-mediated attenuation of
translation initiation, which has no negative consequences
on the processivity of translation or translocation and
therefore on the rate of protein folding. Our studies suggest that P58IPK induction represents a second line of defense, which by extracting stalled translocation substrates
frees the bound lumenal chaperones to engage in other
tasks and clears the translocon from obstructing difficult-to-fold polypeptides.
EXPERIMENTAL PROCEDURES
Cell Culture, Transfection, and Treatment
P58IPK / and P58IPK+/+ mouse embryo fibroblasts were prepared
from embryonic day 13.5 progeny of heterozygous knockout mice (Ladiges et al., 2005). Radiolabeled HA-tagged TCRa, NHK-AT, and wildtype or DJ mutant (1-393) mouse P58IPK was expressed by transient
transfection using the FuGENE 6 lipid-mediated gene transfer (Roche)
in fibroblasts and the DEAE-dextran method for COS1 cells and detected by immunoprecipitation as previously described (Yu et al.,
1997). Mammalian expression plasmids for human VCAM-1 and its
truncated derivative have been previously described (Besemer et al.,
2005; Garrison et al., 2005).
Antisera, Immunoblotting, Immunoprecipitation,
and Immunostaining
Antiserum to mouse P58IPK (NY1165) was raised by injecting a rabbit
with full-length mouse P58IPK expressed in E. coli as a GST fusion protein and cleaved by thrombin. Its use in immunostaining is described is
detail in Figure S1A. Insulin immunoreactivity was detected as previously described (Oyadomari et al., 2002).
Rabbit anti-mouse CYP1A1 was a gift of Virginia Black (NYU), antiserum to puromycin was a gift of Peter Walter (UCSF), and antiserum to
human VCAM-1 and ribophorin I were raised in rabbit against the bacterially expressed protein. Other antisera were purchased from commercial sources and used according to the vendor’s specifications:
rabbit anti-human a-1-Antitrypsin (DAKO), rabbit anti-GFP (Molecular
Probe), mouse anti-HSP70/HSC70 (Stressgen), rabbit anti-neomycin
phosphotransferase II (NPT II; US Biological), mouse anti-PDI (Stressgen), rabbit anti-Sec61a (Affinity BioReagents), and anti-phosphoeIF2a (BioSource).
Subcellular Fractionation, Protein Purification, and Crosslinking
The liver of the untreated or tunicamycin-treated mice (i.p. injection,
1 mg/kg body weight, 16 hr) was mechanically disrupted by Teflon
glass homogenizer in homogenization buffer (20 mM Tris-HCl [pH
7.5], 5 mM MgCl2, 1 mM dithiothreitol, 250 mM sucrose, 10 mg/ml cycloheximide, 1 mM PMSF, 4 mg/ml aprotonin, and 2 mg/ml pepstatin A),
and the postmitochondrial supernatant (PMS) was isolated by centrifugation. The PMS was adjusted to 30% OptiPrep (Axis-Shield), and total
membranes were floated by centrifugation (250,000 3 g, 3 hr) on 25%
OptiPrep. After dilution of the OptiPrep and pelleting (100,000 3 g,
1 hr), the membranes were resuspended in homogenization buffer
containing 20% OptiPrep and subjected to equilibrium density centrifugation (350,000 3 g, 2 hr) in a self-generated OptiPrep gradient to
separate rough and smooth ER. Rough ER was recovered by pelleting
the relevant membrane fraction, translocons were solubilized in 3%
digitonin (Calbiochem), and the soluble material was subjected to nonequilibrium (velocity) gradient centrifugation in a 15%–50% sucrose
gradient as described (Gorlich et al., 1992).
HSP70 proteins were purified from the digitonin-solubilized membrane preparation by ATP-agarose affinity chromatography and eluted
in 5 mM MgCl2 with or without 20 mM ATP.
Crosslinking of P58IPK to translocon components was assayed in
wild-type or P58IPK / fibroblasts labeled for 16 hr with 35S-Translabel,
solubilized in 3% digitonin, and exposed to varying concentrations of
dithio-bis-succinimidylpropionate (DSP, Pierce) on ice for 60 min. After
Cell 126, 727–739, August 25, 2006 ª2006 Elsevier Inc. 737
quenching the unreacted DSP with 100 mM Tris-HCl (pH 7.5), the
sample was denatured by exposure to 6 M urea and 2% SDS at
37 C. Following 203 dilution of the denaturants, P58IPK was immunoprecipitated in RIPA buffer. The crosslink was reversed by incubating
the sample in 100 mM dithiothreitol, 6 M urea, and 2% SDS at 37 C (we
found that heating the sample led to substantial loss of Sec61a immunoreactivity). Half the sample was set aside for loading on an SDSPAGE and the other half was diluted again and immunoprecipitated
with antiserum to Sec61a, reduced, and denatured as described
above and applied to the same SDS-PAGE. Under these conditions,
Sec61a migrates as a doublet on SDS-PAGE (given the lack of S-S
bonds or glycosylation sites, we attribute this heterogeneity to incomplete denaturation at the relatively low temperature required to retain
solubility).
Detection of metabolically labeled apoB was adapted from the published procedure (Fisher et al., 1997) to accommodate the use of primary hepatocytes isolated from C57BL/6 adult male mice of wildtype and P58IPK / genotypes and cultured on type I collagen-coated
plates in Waymouth’s medium supplemented with 0.1 nM insulin. Cells
were switched to DMEM with 10% of the normal methionine and cysteine content 30 min before addition of 35S-TRANSlabel (MP Biomedicals) at 250 mCi/ml for 30 min, followed by cold chase. Cells were pretreated with 10 mM lactacystin (Boston Biochem) for 30 min and then
treated with the MTP inhibitor (50 mM CP10447; Pfizer) and 10 mM lactacystin as indicated. Radiolabeled apoB was immunoprecipitated
with antiserum raised to mouse apoB.
Radiolabeled apoB in complex with P58IPK was isolated from rat
McA-RH7777 hepatoma cells by adapting a procedure to detect complexes of apoB and Sec61 (Pariyarath et al., 2001). The cells were metabolically labeled in the presence of proteasome inhibitor as above,
treated with digitonin (Calbiochem), and exposed to DSP crosslinker.
Quenching, complex disruption, and sequential immunoprecipitation
were conducted as described above (for Sec61a). The total amount
of apoB synthesized was recovered by immunoprecipitation from
a parallel sample of cells + supernatant, and the fraction of apoB recovered in complex with P58IPK was determined as the ratio of the
two measurements.
VCAM-1 biogenesis was followed by pulse-labeling and immunoprecipitation. Wild-type and P58IPK / mouse fibroblasts were infected with a recombinant retrovirus encoding human VCAM-1,
whereas COS1 cells were transfected with a plasmid encoding
VCAM-1, alongside plasmids encoding wild-type or DJ mutant mouse
P58IPK. Two days after gene transduction, the cells were exposed to
varying concentrations of CAM741 (50 nM–10 mM) in the presence or
absence of proteasome inhibitor or tunicamycin, as described (Besemer et al., 2005), radiolabeled for 4 hr, and immunoprecipitated
with a rabbit polyclonal serum raised against the bacterially expressed
protein that detects both native and nonnative human VCAM-1.
Complexes between human VCAM-1 and P58IPK were detected in
cotransfected COS1 cells as described above. CAM741 (1 mM) and
MG132 (10 mM) were added to the culture media 30 min before and
throughout the 2 hr 35S-TRANSlabel (ICN; 250 mCi/ml) labeling pulse.
Complex solubilization, crosslinking, and sequential immunoprecipitation were performed as described above for the analysis of P58IPK
and Sec61a complexes.
The methods used to purify and detect complexes between puromycinylated (nascent) proteins and P58IPK in mouse fibroblasts are
described in detail in the Results section. Metabolic labeling was
performed in methionine minus DMEM with 10% dialyzed fetal calf
serum. Puromycin (Calbiochem; 0.3 mM) and 35S-TRANSlabel (ICN;
250 mCi/ml) were added for 10 min before solubilization in digitonin
and crosslinking.
C. elegans Genetics and Procedures
The ire-1(zc14)II mutant and hsp-4::gfp(zcIs4)V BiP-promoter GFP reporter fusion strains are available from the Caenorhabditis Genetics
Center. The pek-1 (PERK) deletion strain pek-1(zcdf2)X was bred into
738 Cell 126, 727–739, August 25, 2006 ª2006 Elsevier Inc.
the hsp-4::gfp(zcIs4)V background using standard procedures. To
study the effect of gene inactivation on larval development, gravid homozygous ire-1(zc14)II mutant animals were treated with bleach to release fertilized eggs, 100 of which were placed on a lawn E. coli transformed with the pPD129.36 plasmid expressing double-stranded RNA
to pek-1 or dnj-7 (P58IPK), or a mixture of both. The plates were photographed 72 hr later, and the fraction of progeny that reached the L4
stage was determined by counting individual animals.
Experimental Diabetes Mellitus in the Mouse
All experiments in mice were approved by the IACUC. ‘‘Akita’’ mice
bearing the C96Y mutation in Ins2 were purchased from Jackson
Labs and were bred onto the P58IPK knockout background (Ladiges
et al., 2005) to create Ins2C96Y/+;P58IPK /+ animals. As P58IPK+/
mice are indistinguishable from the wild-type, Ins2C96Y/+;P58IPK /+
animals were then crossed with Ins2+/+;P58IPK / to yield compound
animals with combinations of Ins2+/+, Ins2C96Y/+, P58IPK+/ , and
P58IPK / genotypes. Glycemic control was assessed weekly by measuring morning blood glucose. Plasma insulin levels were measured at
sacrifice. Pancreata were fixed in 10% formaldehyde, paraffin embedded, sectioned in 5 mm, and stained with hematoxylin and eosin or immunostained for insulin and P58IPK. Statistical analysis was performed
by the unpaired two-tailed Student’s t test.
Supplemental Data
The Supplemental Data include five figures and can be found with this
article online at http://www.cell.com/cgi/content/full/126/4/727/DC1/.
ACKNOWLEDGMENTS
We thank Jennifer Hui, Kazutaka Araki, Rivka Jungreis, and Cristina
Villagra (NYU) for their assistance with various phases of this project;
Berndt Oberhauser and Shirley Wang (Novartis) for synthesis of
CAM741 and construction of the VCAM1 plasmids; Ron Kopito (Stanford University) for the HA-tagged TCRa clone; Virginia Black (NYU) for
the antiserum to CYP1A1; Peter Walter (UCSF) for the anti-puromycin
antibody; and Ramanujan Hegde (NIH) and Cole Haynes (NYU) for advice. Supported by NIH grants DK47119 and ES08681 to D.R.,
HL58541 to E.A.F., and fellowships from JDRF and the Uehara Memorial Foundation to S.O. H.H. is employed by Novartis and owns stock
options and restricted stock units. The authors have no other conflicting financial interests.
Received: November 14, 2005
Revised: May 1, 2006
Accepted: June 8, 2006
Published: August 24, 2006
REFERENCES
Anken, E., Braakman, I., and Craig, E. (2005). Versatility of the endoplasmic reticulum protein folding factory. Crit. Rev. Biochem. Mol.
Biol. 40, 191–228.
Barber, G.N., Thompson, S., Lee, T.G., Strom, T., Jagus, R., Darveau,
A., and Katze, M.G. (1994). The 58-kilodalton inhibitor of the interferoninduced double-stranded RNA-activated protein kinase is a tetratricopeptide repeat protein with oncogenic properties. Proc. Natl. Acad.
Sci. USA 91, 4278–4282.
Besemer, J., Harant, H., Wang, S., Oberhauser, B., Marquardt, K., Foster, C.A., Schreiner, E.P., de Vries, J.E., Dascher-Nadel, C., and Lindley, I.J. (2005). Selective inhibition of cotranslational translocation of
vascular cell adhesion molecule 1. Nature 436, 290–293.
Brodsky, J.L., Goeckeler, J., and Schekman, R. (1995). BiP and
Sec63p are required for both co- and posttranslational protein translocation into the yeast endoplasmic reticulum. Proc. Natl. Acad. Sci.
USA 92, 9643–9646.
Brychzy, A., Rein, T., Winklhofer, K.F., Hartl, F.U., Young, J.C., and
Obermann, W.M. (2003). Cofactor Tpr2 combines two TPR domains
and a J domain to regulate the Hsp70/Hsp90 chaperone system.
EMBO J. 22, 3613–3623.
Crowley, K.S., Liao, S., Worrell, V.E., Reinhart, G.D., and Johnson, A.E.
(1994). Secretory proteins move through the endoplasmic reticulum
membrane via an aqueous, gated pore. Cell 78, 461–471.
Fisher, E.A., and Ginsberg, H.N. (2002). Complexity in the secretory
pathway: The assembly and secretion of apolipoprotein B-containing
lipoproteins. J. Biol. Chem. 277, 17377–17380.
Fisher, E.A., Zhou, M., Mitchell, D.M., Wu, X., Omura, S., Wang, H.,
Goldberg, A.L., and Ginsberg, H.N. (1997). The degradation of apolipoprotein B100 is mediated by the ubiquitin-proteasome pathway and
involves heat shock protein 70. J. Biol. Chem. 272, 20427–20434.
Fons, R.D., Bogert, B.A., and Hegde, R.S. (2003). Substrate-specific
function of the translocon-associated protein complex during translocation across the ER membrane. J. Cell Biol. 160, 529–539.
Garrison, J.L., Kunkel, E.J., Hegde, R.S., and Taunton, J. (2005). A
substrate-specific inhibitor of protein translocation into the endoplasmic reticulum. Nature 436, 285–289.
Gorlich, D., and Rapoport, T.A. (1993). Protein translocation into proteoliposomes reconstituted from purified components of the endoplasmic reticulum membrane. Cell 75, 615–630.
Gorlich, D., Prehn, S., Hartmann, E., Kalies, K.U., and Rapoport, T.A.
(1992). A mammalian homolog of SEC61p and SECYp is associated
with ribosomes and nascent polypeptides during translocation. Cell
71, 489–503.
Gusarova, V., Caplan, A.J., Brodsky, J.L., and Fisher, E.A. (2001). Apoprotein B degradation is promoted by the molecular chaperones hsp90
and hsp70. J. Biol. Chem. 276, 24891–24900.
Harding, H., Zhang, Y., and Ron, D. (1999). Translation and protein
folding are coupled by an endoplasmic reticulum resident kinase.
Nature 397, 271–274.
Hegde, R.S., and Lingappa, V.R. (1996). Sequence-specific alteration
of the ribosome-membrane junction exposes nascent secretory proteins to the cytosol. Cell 85, 217–228.
Hussain, M.M., Shi, J., and Dreizen, P. (2003). Microsomal triglyceride
transfer protein and its role in apoB-lipoprotein assembly. J. Lipid Res.
44, 22–32.
Jungnickel, B., and Rapoport, T.A. (1995). A posttargeting signal sequence recognition event in the endoplasmic reticulum membrane.
Cell 82, 261–270.
Kaufman, R.J. (2002). Orchestrating the unfolded protein response in
health and disease. J. Clin. Invest. 110, 1389–1398.
Ladiges, W.C., Knoblaugh, S.E., Morton, J.F., Korth, M.J., Sopher,
B.L., Baskin, C.R., MacAuley, A., Goodman, A.G., LeBoeuf, R.C.,
and Katze, M.G. (2005). Pancreatic beta-cell failure and diabetes in
mice with a deletion mutation of the endoplasmic reticulum molecular
chaperone gene P58IPK. Diabetes 54, 1074–1081.
Lee, T.G., Tang, N., Thompson, S., Miller, J., and Katze, M.G. (1994).
The 58,000-dalton cellular inhibitor of the interferon-induced doublestranded RNA-activated protein kinase (PKR) is a member of the tetratricopeptide repeat family of proteins. Mol. Cell. Biol. 14, 2331–2342.
Lee, A.H., Iwakoshi, N.N., and Glimcher, L.H. (2003). XBP-1 regulates
a subset of endoplasmic reticulum resident chaperone genes in the
unfolded protein response. Mol. Cell. Biol. 23, 7448–7459.
Matlack, K.E., Misselwitz, B., Plath, K., and Rapoport, T.A. (1999). BiP
acts as a molecular ratchet during posttranslational transport of prepro-alpha factor across the ER membrane. Cell 97, 553–564.
Melville, M.W., Tan, S.L., Wambach, M., Song, J., Morimoto, R.I., and
Katze, M.G. (1999). The cellular inhibitor of the PKR protein kinase,
P58(IPK), is an influenza virus-activated co-chaperone that modulates
heat shock protein 70 activity. J. Biol. Chem. 274, 3797–3803.
Meusser, B., Hirsch, C., Jarosch, E., and Sommer, T. (2005). ERAD:
The long road to destruction. Nat. Cell Biol. 7, 766–772.
Misselwitz, B., Staeck, O., and Rapoport, T.A. (1998). J proteins catalytically activate Hsp70 molecules to trap a wide range of peptide sequences. Mol. Cell 2, 593–603.
Mitchell, D.M., Zhou, M., Pariyarath, R., Wang, H., Aitchison, J.D.,
Ginsberg, H.N., and Fisher, E.A. (1998). Apoprotein B100 has a prolonged interaction with the translocon during which its lipidation and
translocation change from dependence on the microsomal triglyceride
transfer protein to independence. Proc. Natl. Acad. Sci. USA 95,
14733–14738.
Oyadomari, S., Koizumi, A., Takeda, K., Gotoh, T., Akira, S., Araki, E.,
and Mori, M. (2002). Targeted disruption of the Chop gene delays endoplasmic reticulum stress-mediated diabetes. J. Clin. Invest. 109,
525–532.
Pariyarath, R., Wang, H., Aitchison, J.D., Ginsberg, H.N., Welch, W.J.,
Johnson, A.E., and Fisher, E.A. (2001). Co-translational interactions of
apoprotein B with the ribosome and translocon during lipoprotein assembly or targeting to the proteasome. J. Biol. Chem. 276, 541–550.
Patil, C., and Walter, P. (2001). Intracellular signaling from the endoplasmic reticulum to the nucleus: The unfolded protein response in
yeast and mammals. Curr. Opin. Cell Biol. 13, 349–355.
Scheufler, C., Brinker, A., Bourenkov, G., Pegoraro, S., Moroder, L.,
Bartunik, H., Hartl, F.U., and Moarefi, I. (2000). Structure of TPR domain-peptide complexes: Critical elements in the assembly of the
Hsp70-Hsp90 multichaperone machine. Cell 101, 199–210.
Shen, X., Ellis, R.E., Lee, K., Liu, C.-Y., Yang, K., Solomon, A., Yoshida,
H., Morimoto, R., Kurnit, D.M., Mori, K., and Kaufman, R.J. (2001).
Complementary signaling pathways regulate the unfolded protein
response and are required for C. elegans development. Cell 107,
893–903.
Sitia, R., and Braakman, I. (2003). Quality control in the endoplasmic
reticulum protein factory. Nature 426, 891–894.
Tsai, B., Ye, Y., and Rapoport, T.A. (2002). Retro-translocation of proteins from the endoplasmic reticulum into the cytosol. Nat. Rev. Mol.
Cell Biol. 3, 246–255.
Van Huizen, R., Martindale, J.L., Gorospe, M., and Holbrook, N.J.
(2003). P58IPK, a novel endoplasmic reticulum stress-inducible protein and potential negative regulator of eIF2a signaling. J. Biol.
Chem. 278, 15558–15564.
Vogel, J.P., Misra, L.M., and Rose, M.D. (1990). Loss of BiP/GRP78
function blocks translocation of secretory proteins in yeast. J. Cell
Biol. 110, 1885–1895.
Yan, W., Frank, C.L., Korth, M.J., Sopher, B.L., Novoa, I., Ron, D., and
Katze, M.G. (2002). Control of PERK eIF2-a kinase activity by the endoplasmic reticulum stress-induced molecular chaperone P58IPK.
Proc. Natl. Acad. Sci. USA 99, 15920–15925.
Ye, Y., Meyer, H.H., and Rapoport, T.A. (2001). The AAA ATPase
Cdc48/p97 and its partners transport proteins from the ER into the
cytosol. Nature 414, 652–656.
Yu, H., Kaung, G., Kobayashi, S., and Kopito, R.R. (1997). Cytosolic
degradation of T-cell receptor alpha chains by the proteasome.
J. Biol. Chem. 272, 20800–20804.
Zhang, J.X., Braakman, I., Matlack, K.E., and Helenius, A. (1997). Quality control in the secretory pathway: The role of calreticulin, calnexin
and BiP in the retention of glycoproteins with C-terminal truncations.
Mol. Biol. Cell 8, 1943–1954.
Cell 126, 727–739, August 25, 2006 ª2006 Elsevier Inc. 739
Cell, Volume 126
Supplemental Data
Cotranslocational Degradation Protects
the Stressed Endoplasmic Reticulum
from Protein Overload
Seiichi Oyadomari, Chi Yun, Edward A. Fisher, Nicola Kreglinger, Gert Kreibich, Miho
Oyadomari, Heather P. Harding, Alan G. Goodman, Hanna Harant, Jennifer L. Garrison,
Jack Taunton, Michael G. Katze, and David Ron
Figure S1. Colocalization of Endogenous P58IPK with an ER Marker and Preserved Membrane Association of
P58IPK in PERK-/- Cells
(A) Fluorescent photomicrographs of wildtype (P58+/+) and P58IPK knockout (P58-/-) mouse fibroblasts
immunostained with antiserum to P58IPK and the KDEL epitope common to ER lumenal proteins.
Method: Antiserum to mouse P58IPK (NY1165) was raised by injecting a rabbit with full-length mouse P58IPK
expressed in E. coli as a GST fusion protein and cleaved by thrombin. The antiserum reacts strongly with the
denatured protein on immunoblot (1/3000) and by immunoprecipitation, but does not recognize native P58IPK in
complex with the translocon. For use in immunostaining, the rabbit anti-mouse P58IPK antiserum was purified by
flowing through a GST column (to deplete E. coli-expressed antigens) and an immobilized P58IPK-/- liver lysate
column. Immunochemistry was performed after denaturing methanol-fixed cells by microwave irradiation for 5 min
in 5M urea and 100 mM Tris-HCl (pH 9.5), followed by heating for 20 min at 95°C to unmask the antigen. KDEL
and P58IPK immunoreactivity were detected by incubating with a mouse anti-KDEL (Stressgen) diluted 1/1000 and
purified rabbit anti-P58IPK antiserum diluted 1/500, followed by incubation with FITC-conjugated donkey antimouse and Texas red-conjugated goat anti-rabbit (all from Jackson immunoResearch) diluted to 1/500.
(B) Immunoblot of P58IPK, ribophorin I (RPN I) and eIF2α in post-mitochondrial supernatant (PMS) and membrane
fractions (MB) purified by flotation and concentration by sedimentation from the indicated organs of tunicamycintreated wildtype (+/+) and PERK knockout (-/-) mice. The lower levels of P58IPK in the input PMS of the -/samples is expected, as P58IPK mRNA is a target of the PERK-dependent integrated stress response (Harding, et al.,
2003, Mol Cell; 11:619).
Figure S2. Enforced Expression of P58IPK Does Not Affect the t1/2 of Conventional Misfolded/Misassembled
ER Proteins
(A) Autoradiograph of mutant Null Hong Kong α-1 anti-trypsin (NHK-AT) immunoprecipitated from
metabolically-labeled COS-1 cells cotransfected with plasmids expressing NHK-AT (a gift of Nobuko Hosokawa
and Kazuhiro Nagata, Kyoto University) and full-length P58IPK (FL) or P58IPK lacking the DnaJ domain (∆J). The
cells were 35S-labeled by a 30 min pulse followed by a cold-chase of the indicated time
(B) Autoradiograph of labeled T cell receptor alpha chain (TCRα) immunoprecipitated from metabolically-labeled
COS-1 cells cotransfected with plasmids expressing TCRα (a gift of Ron Kopito, Stanford University) and fulllength P58IPK (FL) or P58IPK lacking the DnaJ domain (∆J). The cells were 35S-labeled by a 30 min pulse followed
by a cold-chase of the indicated time.
(C) Autoradiogram of TCRα and NPT II immunoprecipitated from COS-1 cells transfected with combinations of
plasmids directing the expression of full-length TCRα (FL), TCRα lacking the signal peptide (∆SP), full-length
P58IPK (FL) or P58IPK lacking the DnaJ domain (∆J), following a brief 5 min labeling pulse (mean ± range).
Figure S3. Biochemical Characterization of VCAM-1 Synthesized in the Presence of Inhibitors
Autoradiogram of metabolically labeled human VCAM-1 expressed in transfected COS1 and immunoprecipitated
with a specific antiserum. Where indicated the cells were exposed to a combination of the translocation inhibitor
CAM741 and a proteasome inhibitor or the glycosylation inhibitor tunicamycin. The immunopurified proteins were
resolved by SDS-PAGE. The heterogeneity in mobility of mature VCAM-1 presumably reflects heterogenity in
post-ER sugar modifications or other post-translational modifications. Note the faster mobility and uniform size of
the non-glycosylated VCAM-1 whose translocation was inhibited by CAM741 and the fact that it migrates a bit
behind VCAM-1 from tunicamycin-treated cells, as predicted by signal sequence cleavage of the latter (similar
species were identified by Besemer et al., 2005; Garrison et al., 2005).
(References: Besemer, J., Harant, H., Wang, S., Oberhauser, B., Marquardt, K., Foster, C. A., Schreiner, E. P., de
Vries, J. E., Dascher-Nadel, C., and Lindley, I. J. (2005). Selective inhibition of cotranslational translocation of
vascular cell adhesion molecule 1. Nature 436, 290-293.
Garrison, J. L., Kunkel, E. J., Hegde, R. S., and Taunton, J. (2005). A substrate-specific inhibitor of protein
translocation into the endoplasmic reticulum. Nature 436, 285-289.)
Figure S4. Specificity of the Sequential Immunoprecipitation Procedure Used to Detect Puromycinylated
Proteins Associated with P58IPK
Quantification of radiolabeled proteins detected by TCA precipitation (samples 1 & 2) or sequential
immunoprecipitation with anti-P58IPK followed by anti-puromycin (samples 3 & 4, as in figure 7A) (mean ± range).
At the concentration used (0.3µM), puromycin has almost no effect on the total incorporation of label into newly
synthesized proteins (revealed by TCA precipitation, samples 1 & 2) but its omission from the labeling mix
markedly deceased the labeled material recovered in the sequential immunoprecipitation procedure (compare
samples 3 & 4), attesting the latter’s specificity for nascent puromycinylated proteins.
Figure S5. Higher Levels of ER Stress and Protein Misfolding in P58IPK-/- Fibroblasts
(A) (Upper panels, labeled PNS) Immunoblot of endogenous P58IPK, BiP and HSP47 in the post-nuclear supernatant
(PNS) of wildtype (P58+/+), P58IPK knockout (P58-/-) or knockout cells rescued with a P58IPK transgene (P58-/rescued) mouse fibroblasts following exposure to tunicamycin. (Lower panels, labeled HMW) Immunoblots of BiP
and HSP47 in the high molecular weight (HMW), SDS resistant fraction derived from the same samples by
exposure to 0.5% SDS and sedimentation through a 20% glycerol gradient at 150,000 x g for 1 h. The position of
glycosylated (G) and nonglycosylated (N) HSP47 is indicated.
(B) Quantification of the ratio of BiP and Hsp47 found in the high molecular weight complex (HMW) to that found
in the post-mitochondrial supernatent (PMS) in each of the samples shown above.
This assay estimates the effect of P58IPK on misfolded protein burden by comparing the incorporation of the
abundant ER chaperones BiP and HSP47 into SDS-resistant, high molecular weight complexes in mouse fibroblasts
with various P58IPK genotypes exposed to the ER stress-inducing glycosylation inhibitor tunicamycin. These
surrogate markers of protein misfolding (Marciniak et al., 2004) were detected earlier in tunicamycin-exposed
P58IPK-/- cells than in wildtype or P58IPK-/- cells rescued with a P58-expressing transgene. Unlike the tissues of
animals in which P58IPK is inducible by ER stress, P58IPK is constitutively present in the mouse fibroblasts used
here, rendering the comparison between the wildtype and rescued P58IPK-/- cells easier to interpret. The antisera to
Hsp47 was a gift of Nobuko Hosokawa and Kazuhiro Nagata (Kyoto University).
(Reference: Marciniak, S. J., Yun, C. Y., Oyadomari, S., Novoa, I., Zhang, Y., Jungreis, R., Nagata, K., Harding, H.
P., and Ron, D. (2004). CHOP induces death by promoting protein synthesis and oxidation in the stressed
endoplasmic reticulum. Genes and Development 18, 3066-3077).
Molecular Cell 23, 773–777, September 15, 2006 ª2006 Elsevier Inc.
DOI 10.1016/j.molcel.2006.08.024
Previews
Cotranslocational Degradation:
Utilitarianism in the
ER Stress Response
Recently, a new layer of the unfolded protein response
was discovered that supports the cotranslocational
degradation of nascent chains stalled in endoplasmic
reticulum translocons (Oyadomari et al., 2006).
Disruption of protein folding in the endoplasmic reticulum (ER) activates the unfolded protein response
(UPR) signal transduction pathway, causing temporary
remodeling of the ER (Schroder and Kaufman, 2005).
The balance of ER resident proteins is shifted to remove
aberrant substrates and restore the ER to an efficiently
operating maturation compartment. The UPR pathway
functions as a tripartite signal that involves (1) increasing
the expression of housekeeping proteins that can work
toward properly folding the misfolded proteins, (2) attenuating the secretory pathway load by decreasing the
expression of secretory cargo, and (3) increasing the capacity for ER-associated protein degradation or ERAD
(Figure 1). A study in the August 25, 2006, issue of Cell
by Oyadomari et al. (2006) reveals a new layer in the
UPR pathway that permits the cotranslocational degradation of secretory proteins (Oyadomari et al., 2006).
This function diminishes the biosynthetic burden on the
ER by degrading proteins at a stage earlier than previously envisioned.
The recruitment of the ER Hsp70 chaperone BiP to
misfolded proteins in the lumen of the ER sequesters
the chaperone away from three different ER membrane
proteins that act as UPR transducers known as IRE1,
ATF6, and PERK (Schroder and Kaufman, 2005). BiP release from these signaling proteins triggers a cascade of
events that initiates the UPR. The enhanced transcription of folding and degradation machinery messages is
initiated by IRE1 through its endonuclease activity and
ATF6, which is itself a transcription factor. In addition,
the kinase PERK attenuates general protein synthesis
through its phosphorylation of eIF2a.
The quality control machinery of the ER monitors the
integrity of protein maturation and sorts misfolded proteins to the ERAD pathway (Hebert et al., 2005). This process involves the posttranslational targeting of terminally misfolded proteins to a retrotranslocation channel
whose identity is an area of vigorous concern and debate
(Tsai et al., 2002). With the aid of the cytosolic AAAATPase p97/Cdc48, the misfolded protein is extruded
through the membrane conduit, where it is then polyubiquitinated and delivered to the proteasome for degradation. The cotranslocational selection of nascent chains
for degradation while they are still associated with the
Sec61 translocon seemingly bypasses the necessity for
the targeting of proteins to a translocon (Oyadomari
et al., 2006). This mechanism of degradation also frees
up anterograde channels for the translocation of housekeeping and ERAD machinery induced by UPR. While
it still remains to be determined if Sec61 is the primary
route for misfolded proteins to travel to the cytoplasm
for ERAD, this study demonstrates that Sec61 can
work in the reverse retrograde direction to retrotranslocate proteins that are stalled during the translation and
translocation processes (Oyadomari et al., 2006).
P58IPK/DNAJC3 was initially linked to UPR as an inhibitor of PERK (Yan et al., 2002). Thus, P58IPK/DNAJC3
was thought to be part of a negative feedback loop involved in restoring translation to normal physiological
levels once the stress had dissipated (Figure 1, dotted
line). However, Oyadomari et al. (2006) demonstrated
that P58IPK/DNAJC3 is peripherally associated with the
ER membrane and exists in large molecular weight complexes with the Sec61 translocon and cytosolic Hsp70
(Figure 2) (Oyadomari et al., 2006). P58IPK/DNAJC3
binds Hsp70 chaperones and stimulates its ATPase activity through its nine TPR repeat motifs and its single J
domain, both well-characterized Hsp70 binding motifs
(Melville et al., 1999).
P58IPK/DNAJC3 appears to assist in the cotranslational/translocational degradation of nascent chains
that are stalled on ribosomes and in ER translocons
(Oyadomari et al., 2006). Pharmacological arrest of
both ApoB100 and VCAM-1 within ER translocons led
to their association with P58IPK/DNAJC3. The cotranslocational degradation of these substrates was diminished
in cells derived from P582/2 mice, accelerated when
P58IPK/DNAJC3 was overexpressed, and required a functional J domain. This supports the hypothesis that
P58IPK/DNAJC3 recruits Hsp70 to the cytosolic opening
of the ER translocon stimulating its ATPase activity to
assist in the extraction of stalled nascent proteins into
the cytoplasm for proteasomal degradation (Figure 2).
UPR-induced intervention early in the maturation
process by P58IPK/DNAJC3 allows the maturation and
quality control machinery to focus its attention on the
pre-existing improperly folded proteins that triggered
the initial UPR signal. A recent study from the Weissman
laboratory has also uncovered a new early intervention
posttranscriptional process that involves the translation
branch of the UPR signal (Hollien and Weissman, 2006).
RNA encoding secretory and membrane proteins was
discovered to be degraded by the activated endonuclease, IRE1. The endonuclease activity of IRE1 was shown
to be able to process numerous transcripts in addition to
XBP-1 (or HAC1 in yeast) mRNA, the transcription factor
message processed and activated by IRE1 (Schroder
and Kaufman, 2005). The more promiscuous endonuclease activity of IRE1 provides an additional posttranscriptional method of lowering the workload of the ER
by directly reducing the number of nascent proteins
translated and translocated into the ER.
ER stress and UPR are associated with a number of
human diseases (Kaufman, 2002). Studies using P582/2
mice found that these animals were greatly disrupted
in their ability to respond to ER stress (Oyadomari
et al., 2006). Hyperglycemic P582/2 mice possessing
a mutation in the Ins2 gene were unable to elevate their
plasma insulin levels, causing b cells defects and death
Molecular Cell
774
Figure 1. The Tripartite Unfolded Protein
Response
Three primary transducers of the UPR signal,
known as ATF6, IRE1, and PERK, seek to relieve ER stress through increased expression of housekeeping proteins, increased
degradation of terminally misfolded proteins,
and suppression of translational initiation
until the aberrations have been alleviated.
P58IPK/DNAJC3 upregulation is triggered by
the UPR, is initiated through IRE1 activity,
and was originally found to act as a negative
regulator of PERK-dependent translational
attenuation (dotted line). However, a recent
study demonstrates that P58IPK/DNAJC3
assists in the cotranslational ER-associated
degradation of translocon-stalled substrates
(Oyadomari et al., 2006).
within 6 to 12 weeks of birth. This demonstrated that
P58IPK/DNAJC3 plays a central role in maintaining the
homeostasis of the whole organism.
Many questions remain as to the molecular mechanism of P58IPK/DNAJC3 substrate selection and its
assistance in the extraction of translocon-associated
polypeptides. Is the lumenal quality control process involved in substrate selection, or is this decision solely
based on cytosolic cues? Does P58IPK/DNAJC3 association with translocons require nascent chains? Is there
a conformational change that takes place within P58IPK/
DNAJC3 that supports Hsp70 recruitment? Or is activation executed by the abundance of these complexes
that is controlled by UPR? And finally, is the ability of
P58IPK/DNAJC3 to aid in the degradation of stalled
nascent chains the key role of this multifunctional
protein? Future studies will be required to increase
our understanding of this powerful and intuitive stress
response.
Bradley R. Pearse1 and Daniel N. Hebert1
1
Department of Biochemistry and Molecular Biology
Program in Molecular and Cellular Biology
University of Massachusetts
710 North Pleasant Street
Amherst, Massachusetts 01003
Figure 2. Cotranslational/Translocational Degradation Regulates the Secretory Load during the ER Stress Response
Under normal conditions in the ER, nascent chains are cotranslationally translocated into the ER lumen through the Sec61 translocon (blue). As
the polypeptide emerges from the translocon channel, it is engaged by numerous chaperones such as BiP (green) that support proper substrate folding. Under ER stress conditions and initiation of the UPR, P58IPK/DNAJC3 (red) recruits cytosolic HSP70 chaperones (purple) to the
Sec61 translocon to extract nascent chains for subsequent proteasomal degradation (yellow). This cotranslational/translocational degradation
decreases the amount of secretory cargo entering the ER lumen, allowing UPR-induced chaperones such as BiP to more effectively clear
misfolded proteins from the ER.
Previews
775
Selected Reading
Hebert, D.N., Garman, S.C., and Molinari, M. (2005). Trends Cell
Biol. 15, 364–370.
Hollien, J., and Weissman, J.S. (2006). Science 313, 104–
107.
Oyadomari, S., Yun, C., Fisher, E.A., Kreglinger, N., Kreibich, G.,
Oyadomari, M., Harding, H.P., Goodman, A.G., Harant, H., Garrison,
J.L., et al. (2006). Cell 126, 727–739.
Schroder, M., and Kaufman, R.J. (2005). Annu. Rev. Biochem. 74,
739–789.
Kaufman, R.J. (2002). J. Clin. Invest. 110, 1389–1398.
Tsai, B., Ye, Y., and Rapoport, T.A. (2002). Nat. Rev. Mol. Cell Biol.
3, 246–255.
Melville, M.W., Tan, S.L., Wambach, M., Song, J., Morimoto, R.I.,
and Katze, M.G. (1999). J. Biol. Chem. 274, 3797–3803.
Yan, W., Frank, C.L., Korth, M.J., Sopher, B.L., Novoa, I., Ron, D., and
Katze, M.G. (2002). Proc. Natl. Acad. Sci. USA 99, 15920–15925.
Molecular Cell 23, September 15, 2006 ª2006 Elsevier Inc. DOI 10.1016/j.molcel.2006.08.023
SnoRNP Biogenesis Meets
Pre-mRNA Splicing
In vertebrates, hundreds of small nucleolar RNAs
(snoRNAs) are processed from pre-mRNA introns. In
the September 1 issue of Molecular Cell, Hirose et al.
(2006) demonstrate that a spliceosomal intron binding
protein, IBP160, couples box C/D snoRNA processing
with pre-mRNA splicing in the C1 splicing complex.
In eukaryotic cells, a large number of box C/D and H/
ACA small nucleolar RNAs (snoRNAs) direct the sitespecific 20 -O-ribose methylation and pseudouridylation
of rRNAs, respectively. The snoRNAs function in the
form of small nucleolar RNPs (snoRNPs), each of which
consists of a box C/D or box H/ACA guide RNA and four
associated C/D or H/ACA snoRNP proteins. Biogenesis
of vertebrate snoRNPs is remarkable: instead of being
transcribed from independent genes, the majority of
box C/D and H/ACA snoRNAs are synthesized within
pre-mRNA introns. The intron-encoded snoRNAs are
processed from the spliced out and debranched host introns by exonucleolytic trimmings. In fact, the box C/D
and H/ACA snoRNP proteins that bind to the intron-embedded snoRNA sequences protect the termini of mature snoRNAs from the processing exonucleases. However, vertebrate introns have a rapid turnover; they are
degraded immediately upon cotranscriptional removal
from the growing pre-mRNA chain. Recent studies revealed that efficient processing of intronic snoRNAs is
promoted by active recruitment of snoRNP proteins to
the nascent intronic snoRNAs during synthesis or before
splicing of the host pre-mRNAs (Ballarino et al., 2005;
Darzacq et al., 2006; Richard et al., 2006; Yang et al.,
2005). These works also indicated that recruitment of
snoRNP proteins and early assembly of intronic
snoRNPs are physically and functionally linked to premRNA synthesis or processing. However, the molecular
mechanisms responsible for coupling intronic snoRNP
assembly with mRNA biogenesis remained speculative.
In a recent issue of Molecular Cell, Hirose et al. identify
a general splicing factor, the intron binding protein 160
(IPB160), as a crucial factor for splicing-dependent
assembly of intronic box C/D snoRNPs (Hirose et al.,
2006). Earlier, IBP160 had been reported to be associated with the purified C1 splicing complex (Jurica et al.,
2002). In the current study, depletion of IBP160 from
HeLa cells by RNA interference demonstrates that
IBP160 is essential for both cell viability and intronic
box C/D snoRNP expression. Using an in vitro system
that couples pre-mRNA splicing with snoRNP processing, Hirose and coworkers show that binding of IBP160
to pre-mRNA introns takes place in the C1 splicing complex, in which assembly of intronic box C/D snoRNPs
also occurs (Hirose et al., 2003). The authors demonstrate that IBP160 interacts with intronic nucleotides
33–40 upstream of the branch site in a sequence-independent manner. Binding of IBP160 to pre-mRNA introns
does not depend on the presence of an intronic snoRNA,
supporting the notion that IBP160 is a general splicing
factor, although its function in pre-mRNA splicing remains unknown. The position-dependent binding of
IBP160 to pre-mRNA introns may be guided by the
SAP155 spliceosomal protein that, as supported by
coimmunoprecipitation experiments, can specifically
interact with IBP160. SAP155 is a component of the U2
snRNP-associated SF3b complex that interacts with
intronic sequences up to 25 nucleotides upstream of
the branch point. Thus, SAP155 may recruit and deposit
IBP160 onto upstream intronic sequences.
How does IBP160 link intronic snoRNP assembly to
pre-mRNA splicing? It was noticed earlier that most
mammalian intronic box C/D snoRNAs are located about
70–80 nucleotides upstream of the 30 splice site (Hirose
and Steitz, 2001). Later, in vivo and in vitro processing
studies confirmed the significance of this observation,
showing that an optimal distance (about 50 nucleotides)
between the intronic box C/D snoRNA and the branch
point is crucial for efficient snoRNP assembly (Hirose
et al., 2003). Since IBP160 contacts intronic sequences
about 40 nucleotides upstream of the branch point, it
may recruit, directly or indirectly, box C/D snoRNP proteins or snoRNP assembly factors to snoRNAs positioned about 50 nucleotides upstream of the branch
point. According to another attractive scenario, upon
binding to pre-mRNA introns, IBP160 may facilitate
proper folding of neighboring intronic box C/D snoRNAs.
Supporting this possibility, IBP160 contains a helicaselike domain, and, more tellingly, processing of intronic
box C/D snoRNAs that are not optimally located within
their host introns requires extended terminal stem
structures that likely promote correct folding of
these snoRNAs. The external stems, therefore, could
compensate for the absence of IBP160 in the proximity