|Home | About | Journals | Submit | Contact Us | Français|
The unfolded protein response (UPR) monitors and maintains protein homeostasis in the endoplasmic reticulum (ER). In budding yeast, the UPR is a transcriptional regulatory pathway that is quiescent under normal conditions. Under conditions of acute ER stress, activation of UPR targets is essential for cell viability. How individual target genes contribute to stress tolerance is unclear. Uncovering these roles is hampered because most targets also play important functions in the absence of stress. To differentiate stress-specific roles from everyday functions, a single target gene was uncoupled from UPR control by eliminating its UPR-specific regulatory element. Through this approach, the UPR remains intact, aside from its inability to induce the designated target. Applying the strategy to the major ER chaperone Kar2p/BiP revealed the physiological function of increasing its cellular concentration. Despite hundreds of target genes under UPR control, we show that activation of KAR2 is indispensable to alleviate some forms of ER stress. Specifically, activation is essential to dispose misfolded proteins that are otherwise toxic. Surprisingly, induced BiP/Kar2p molecules are dedicated to alleviating stress. The inability to induce KAR2 under stress had no effect on its known housekeeping functions.
All organisms are subject to conditional changes that can cause disequilibrium of internal systems. Nutrient deprivation, hypoxia, disease, chemical and radiation exposure, and abrupt changes in pH and temperature can activate complex regulatory circuits known as stress pathways. These pathways also play general homeostatic roles and can be activated in response to natural changes during development and the aging process (Powers et al., 2009 ; Douglas and Dillin, 2010 ; Haigis and Yankner, 2010 ). The unfolded protein response (UPR) is a stress-inducible pathway that monitors and maintains multiple functions of the endoplasmic reticulum (ER). A functioning UPR is critical because the ER is the site of synthesis for about one-third of the proteome and most membrane lipids (Malhotra and Kaufman, 2007 ; Mori, 2009 ; Rutkowski and Hegde, 2010 ).
ER membranes contain a UPR sensor protein called Ire1 that defines a pathway conserved in all eukaryotes. A single-span membrane protein, Ire1 uses its luminal domain to detect ER disequilibrium. In budding yeast, direct binding to unfolded proteins drives Ire1 dimerization and transphosphorylation, but a less direct mechanism for activation might be used in mammals (Kimata et al., 2004 ; Credle et al., 2005 ; Zhou et al., 2006 ; Kimata et al., 2007 ; Gardner and Walter, 2011 ). Subsequent oligomerization activates its cytosolic RNase domain to cleave an inhibitory intron from Hac1 (yeast) or XBP-1 (metazoans) pre-mRNAs (Kimata et al., 2007 ; Korennykh et al., 2009 ). The Hac1 and XBP-1 proteins are transcription factors that elevate expression of UPR target genes (Cox and Walter, 1996 ; Mori et al., 1996 ; Shen et al., 2001 ; Calfon et al., 2002 ). Lower eukaryotes depend solely on the Ire1 pathway, whereas metazoans have two additional sensors that generate distinct outputs. One regulates the activity of a second UPR transcription factor, called ATF-6, that is normally silenced by sequestration. The third sensor, PERK, temporarily inhibits general translation by phosphorylating eIF2α, thereby allowing the ER to restore homeostasis by reducing the load of newly synthesized proteins (Mori, 2009; Rutkowski and Hegde, 2010 ).
Transcriptional profiling revealed the surprising breadth of the UPR target gene repertoire in both budding yeast and mammals. What was believed to be a pathway that regulates chaperones is actually a comprehensive regulatory circuit(s) that can remodel cellular physiology. In yeast, about half of the genes with known functions act in the secretory pathway. They regulate protein translocation, protein folding, ER-associated protein degradation (ERAD), vesicle trafficking, endocytosis, glycosylation, ion homeostasis, lipid biosynthesis, and vacuolar (lysosomal) degradation. The remaining genes function at points throughout the cell. Their roles in ER homeostasis and stress tolerance are largely unexplored (Travers et al., 2000 ; Lee et al., 2003 ).
Although many UPR-regulated genes are essential, the regulatory factors are themselves nonessential in budding yeast and in some animal cells. This reflects the observation that the pathway is inactive under normal conditions (Cox et al., 1993 , 1997 ). In this state, target genes are expressed at basal levels calibrated independent of the UPR. Of individual targets, ERO1 and ERV29 mutants exhibit sensitivity to ER stressors. ERO1 mutants are supersensitive to reducing agents and ERV29 mutants are sensitive to misfolded proteins in the ER (Frand and Kaiser, 1998 ; Pollard et al., 1998 ; Spear and Ng, 2003 ; Haynes et al., 2004 ). Their phenotypes suggest roles in ER stress tolerance. However, the mutants are deficient in the housekeeping functions of ER protein oxidation (ERO1) and COPII vesicle cargo sorting (ERV29; Frand and Kaiser, 1998 ; Pollard et al., 1998 ; Belden and Barlowe, 2001 ). Thus the contribution of indirect effects caused by the loss of normal function is likely substantial but unknown. As such, loss-of-function mutants are not suited to assess the roles of induced target gene products. For this purpose, we devised a method to analyze the stress functions of individual gene targets. A strain of Saccharomyces cerevisiae was engineered to uncouple the UPR-specific regulation of the major target gene KAR2. KAR2 encodes the yeast orthologue of the highly conserved ER chaperone BiP, a member of the Hsp70 family. In these cells, neither intrinsic Kar2p function nor the integrity of the UPR is compromised. The experiments presented here establish a method to understand the role of induced gene products independent of their normal basal activities.
To understand the physiological role of target gene activation, the most direct approach is to uncouple it from the regulatory circuit. For the UPR, we chose to examine KAR2 because its routine functions and promoter regulatory elements are well defined. Kar2p is required for protein translocation, folding, ERAD, and nuclear membrane fusion during yeast mating (Rose et al., 1989 ; Vogel et al., 1990 ; Sanders et al., 1992 ; Simons et al., 1995 ; Matlack et al., 1997 ). Its role in promoting ER stress tolerance, however, is unclear. The KAR2 promoter contains a 22–base pair sequence called the unfolded protein response element (UPRE) that binds Hac1p and confers UPR-specific regulation (Mori et al., 1992 , 1996 ; Kohno et al., 1993 ; Cox and Walter, 1996 ). Mutation of the UPRE eliminated UPR regulation without affecting basal expression of a lacZ reporter gene (Kohno et al., 1993 ). These studies inspired a strategy to uncouple UPR regulation of target genes by simply disrupting the UPRE. Initially, we mutated the KAR2 UPRE in its native chromosomal locus. However, the mutation had little effect on its UPR induction, suggesting regulatory redundancy (data not shown). The difference from earlier studies might reflect their use of limited promoter elements on a plasmid. There the promoter included 245 base pairs of KAR2 5′ noncoding sequences, which is sufficient for full UPR regulation. To better adapt the earlier results to our system, we engineered two versions of the KAR2 gene into a yeast integrating plasmid (see Materials and Methods). The first contains the intact wild-type open reading frame and 284 base pairs of its promoter (UPRE-KAR2). This fragment was previously defined as a fully functional KAR2 clone (Vogel et al., 1990 ). The second is identical except that the UPRE has been replaced with nonspecific sequences that we term upred-KAR2 (the d denotes defective). Although this is technically a mutant gene, “KAR2” is in upper case to emphasize that the encoded protein is wild type. Through several steps, the constructs were integrated into the ura3-1 locus of a strain containing a kar2::KANX insertion that replaces the entire KAR2 coding sequence. The resulting CHY220 and CHY438 strains carry the upred-KAR2 and UPRE-KAR2 alleles as the sole genes expressing Kar2p, respectively.
To test whether the integrated upred-KAR2 allele complements the growth of the kar2::KANX knockout under normal growth conditions, wild-type and upred-KAR2 cells were spotted onto rich (yeast extract/peptone/dextrose [YPD]) and synthetic media (SC) in decreasing concentrations. The engineered strain grows indistinguishably from wild type under both conditions (Figures 1A and and2).2). Next, because Kar2p plays a role in modulating activity of the unfolded protein response (Kimata et al., 2004 ; Pincus et al., 2010 ), the general integrity of the pathway in response to stress was examined. For this purpose, wild-type and upred-KAR2 cells were mock treated or treated with tunicamycin. Tunicamycin is a strong inducer of ER stress by inhibiting N-linked glycosylation (Cox et al., 1993 ). HAC1 mRNA splicing was analyzed by RT-PCR as a direct measure of UPR activation. In the presence of tunicamycin, HAC1 message is spliced in the upred-KAR2 strain to an equal extent as wild type (Figure 1B). In the absence of stress, HAC1 mRNA is found primarily in the unspliced form, with no significant difference between the strains. From these data, we conclude that the upred-KAR2 allele fully supports growth in the absence of stress and does not generally compromise the integrity of the UPR pathway.
Next steady-state Kar2 protein levels were measured after treating wild-type and upred-KAR2 cells with tunicamycin. Equal cell numbers were harvested after 0, 1, 2, and 3 h of treatment and whole-cell extracts prepared. Proteins were separated by SDS–PAGE and analyzed by quantitative immunoblotting. As shown in Figure 1C, Kar2 protein levels rose steadily during the time course in wild-type cells, topping out at greater than fourfold by 3 h. By contrast, no significant change was observed in the upred-KAR2 strain other than a slight increase after 1 h. Northern blot analysis confirmed that the upred-KAR2 gene is unresponsive to ER stress at the transcriptional level (Supplemental Figure S1).
Although it is well established that moderate stress conditions can kill UPR-deficient strains, little is known of how the UPR output confers tolerance (Cox et al., 1993 ). The upred-KAR2 strain allows the direct assessment of KAR2's role. Using standard tests established by Walter and coworkers, we challenged wild-type and upred-KAR2 strains to moderate levels of tunicamycin and the reducing agent dithiothreitol (DTT; Cox et al., 1993 ). As shown in Figure 2A, wild-type cells grew well under both forms of stress, whereas cells lacking Ire1p were killed. The upred-KAR2 cells, however, exhibited differential sensitivities to the agents. Tunicamycin treatment was toxic, whereas DTT was well tolerated. The effect is specific to the inability to activate Kar2p synthesis because the UPRE-KAR2 strain, differing from upred-KAR2 only by its functional UPRE, tolerated both stress conditions. These data show that KAR2 up-regulation is required for some but not all forms of ER stress. In agreement with our data, the Brodsky lab recently reported that expression of KAR2 from the UPR-independent TEF1 promoter confers tolerance to moderate DTT treatment (Vembar et al., 2010 ). Other stress conditions, however, were not examined.
Next we challenged upred-KAR2 cells to ectopic expression of the model misfolded protein carboxypeptidase Y (CPY*), a well-studied substrate of the ERAD pathway (Finger et al., 1993 ; Wolf and Schafer, 2005 ). Hemagglutinin (HA) epitope–tagged CPY* (referred to as CPY* for simplicity) driven by the tightly controlled GAL1 promoter was introduced into wild-type, UPRE-KAR2, and upred-KAR2 strains. Induction of CPY* using this construct strongly induces the UPR (Spear and Ng, 2003 ). Use of CPY* as a specific form of ER stress provides several benefits over that of chemical agents. Directed expression specifies misfolding only to a molecule not needed by the cell. This avoids the pleiotropic folding defects caused by chemical agents that can lead to unintended and nonspecific effects. By use of this strategy, the form and extent of the stress can be carefully controlled by varying CPY* levels. Furthermore, neutralization of aberrant protein toxicity, an important aspect of stress tolerance, can be analyzed by examining the fate of CPY*. Here the severity of stress was adjusted by using two centromeric plasmids instead of a multicopy plasmid to reduce expression variability from cell to cell. As shown in Figure 2B, all strains grew equally well on glucose-containing media, which represses CPY* expression. However, expression of CPY* on galactose-containing media, well tolerated by wild type and UPRE-KAR2 strains, was lethal to upred-KAR2 cells (Figure 2B, CPY*OE and 2C). This is a direct effect of the misfolded protein because upred-KAR2 cells not expressing CPY* grew well under the same conditions (Figure 2B, vector control).
KAR2 induction by the UPR is required for ER stress tolerance, but how is this achieved by elevating Kar2p levels? Kar2p/BiP binds misfolded proteins in vivo (Hurtley et al., 1989 ; de Silva et al., 1990 ; Machamer et al., 1990 ; Ng et al., 1990 ), so the simplest explanation is to maintain adequate levels needed for its essential functions. To examine its functional integrity during ER stress, we compared protein translocation and folding efficiencies of upred-KAR2 cells to wild type. Pulse-label analysis is used to reveal translocation defects through the appearance of cytosolic precursors (Deshaies and Schekman, 1987 ; Hann and Walter, 1991 ). As shown in Figure 3A, substrates of the signal recognition particle (SRP)–dependent (DPAP B) and SRP-independent pathways (Gas1p) translocate efficiently in both strains under stress conditions (Ng et al., 1996 ). To evaluate whether protein folding is compromised, pulse-chase analysis was performed to measure the processing of the glycosylphosphatidylinositol-anchored protein Gas1p. The maturation of Gas1p is well characterized, and failure to fold causes its retention in the ER (Fujita et al., 2006 ). In this experiment, Gas1p reached its Golgi form rapidly in both strains, indicative of normal folding (Figure 3B).
A chemical-based method was next applied as an independent measure of protein folding efficiency in upred-KAR2 cells. The reagent maleimide-polyethlene glycol 5000 (Mal-PEG) attacks and covalently attaches to free cysteinyl sulfhydryls but not disulfides. Because correct disulfide bond formation depends on protein conformation, we used Mal-PEG to develop an assay that monitors the folding state of highly oxidized secretory proteins (Wang and Ng, 2010 ). The assay takes advantage of each attached moiety, increasing the target molecular weight by 5 kDa, a change easily detectable by SDS–PAGE (Tsai et al., 2002 ). Gas1p's 14 cysteine residues make it an excellent candidate for the assay. As a control, the assay was performed with CPY* as the target. As shown in Figure 3C, Mal-PEG–modified CPY* migrates as a high–molecular weight smear, even after a long chase, reflecting its conformational heterogeneity. DTT treatment of cells to prevent CPY* oxidation resulted in a single, high–molecular weight species after Mal-PEG treatment, which demonstrates modification efficiency under assay conditions (Figure 3C, +DTT/+Mal-PEG lanes). Applying the assay to Gas1p from pulse-labeled control cells resulted in a slightly shifted band, followed by a smear that reflects a mixture of folding intermediates (Figure 3D, lane 6). After a cold chase, and coincident with folded molecules transported from the ER, Gas1p became completely resistant to Mal-PEG. This is consistent with the formation of seven disulfide bonds. The reagent is not limiting under these assay conditions because reduced Gas1p (Figure 3D, +DTT) was fully modified and migrated as a single species near the top of the gel. When the assay was applied to Gas1p from upred-KAR2 cells under CPY*-induced stress, the Golgi-modified form was resistant to Mal-PEG, confirming that folding was unperturbed (Figure 3D, lanes 15 and 16).
These experiments demonstrate that an elevated level of Kar2p is not required to maintain its essential functions during ER stress. Next we tested whether the primary function of KAR2 activation is to promote the removal of aberrant proteins. The basis for this notion comes from findings that the UPR facilitates stress tolerance by activating ERAD and vacuolar pathways (Spear and Ng, 2003 ). Kar2p is required for ERAD but its role in the ER-to-vacuole pathway is unknown (Brodsky et al., 1999 ; Plemper and Wolf, 1999 ). The interplay of these pathways under stress conditions is illustrated in Figure 4. When overexpressed, CPY* is degraded equally well in wild-type cells and in the Δcue1 ERAD mutant (Figure 4B, upper left). This is due to the activation and use of the ER-to-vacuole pathway when ERAD is defective or saturated (Spear and Ng, 2003 ). The level of CPY* expression used in this study is sufficient to saturate even a UPR-enhanced ERAD pathway. The contribution of the vacuole is shown by measuring CPY* turnover in a Δpep4 strain deficient in vacuolar proteases (Figure 4B, upper right). The partial stabilization reflects the fraction of CPY* accumulated in the vacuole that cannot be degraded by UPR-enhanced ERAD (Spear and Ng, 2003 ). By contrast, CPY* turnover was severely curtailed in upred-KAR2 cells, suggesting that both degradation pathways are compromised when Kar2p is limiting (Figure 4B, lower left). This notion was confirmed when CPY* turnover was not further compromised in the upred-KAR2/Δpep4 or upred-KAR2/Δcue1 double-mutant strains (Figure 4B, lower right, and Supplemental Figure S3).
The pulse-chase experiments suggested that the ability to clear misfolded proteins might be disrupted in Kar2p-limiting cells, and this, in turn, reduces their fitness to tolerate ER stress. To test the idea, we applied a protein clearance assay. Wild-type and upred-KAR2 cells were preloaded with CPY* by galactose induction, followed by a glucose shift to terminate synthesis. Clearance efficiency was analyzed by immunoblotting and by indirect immunofluorescence. As shown in Figure 5A, CPY* efficiently cleared from wild-type cells but persisted in upred-KAR2, even in absence of ongoing synthesis. CPY* colocalizes with Kar2p, showing that accumulation occurs in the ER (Figure 5B). The inability of upred-KAR2 cells to use the dual degradation pathways becomes apparent when the assay is applied to the same strains also lacking PEP4. By eliminating vacuolar degradation while maintaining transport, this strain reveals the portion of CPY* normally degraded by the vacuole (Spear and Ng, 2003 ). After preloading, CPY* was observed in both ER (colocalization with Kar2p) and vacuoles (non–Kar2p-staining compartments) of Δpep4 cells. In wild-type cells, vacuolar CPY* is difficult to observe because it rapidly degrades there (Kawaguchi et al., 2010 ; Figure 5). Note that the strong CPY* vacuolar staining in Δpep4 cells is due to their accumulation during the load (Figure 6, upper left). Shortly after the glucose shift, ER CPY* was rapidly cleared by ERAD and by trafficking to the vacuole. By contrast, CPY* in upred-KAR2/Δpep4 cells was concentrated primarily in the ER after preloading (Figure 6, upred-KAR2/Δpep4). This result indicates that UPR regulation of KAR2 is critical for ER-to-vacuole trafficking of CPY*. This view is supported by its persistence in the ER long after the termination of synthesis (Figure 6, far right). This is a consequence of stress. Low levels of CPY*-expressed upred-KAR2 cells are cleared efficiently, consistent with the general functionality of basal Kar2p under nonstress conditions (unpublished results). Taken together, these data show that UPR-elevated Kar2p is essential for the clearance of misfolded proteins through ERAD and vacuolar degradation pathways.
Kar2p functions by binding directly to unfolded peptide segments to facilitate protein translocation and folding (Vogel et al., 1990 ; Sanders et al., 1992 ; Brodsky et al., 1995 ; Panzner et al., 1995 ; Simons et al., 1995 ; Matlack et al., 1999 ). In ERAD, its interaction keeps substrates in a soluble state required for degradation (Nishikawa et al., 2001 ). Because UPR-regulated Kar2p is required for ER stress tolerance, we measured its substrate interactions in upred-KAR2 cells. Coimmunoprecipitation experiments revealed that Kar2p complexed with CPY* was markedly reduced compared with wild type (Figure 7A). Because Kar2p essential functions are intact under these conditions (Figure 3), this result shows that stress-regulated Kar2p is deployed to chaperone misfolded proteins. We next analyzed how the reduction in chaperone–substrate complexes affects CPY* solubility. For this, a microsomal fraction was prepared from wild-type, Δire1, and upred-KAR2 cells overexpressing CPY*. The membranes were solubilized in detergent and centrifuged at high speed to separate soluble components (supernatant) from insoluble aggregates (pellet). Overexpressed CPY* in Δire1 cells is entirely insoluble (Figure 7B, lanes 7–9), a condition that causes rapid cell death (Spear and Ng, 2003 ). This contrasts with moderately expressed CPY* in wild-type cells, which fractionates entirely in the soluble fraction (Nishikawa et al., 2001 ; Spear and Ng, 2003 ). In wild-type cells, the majority of overexpressed CPY* fractionates in the insoluble fraction, indicating significant aggregation. This shows that existing stress pathways, when functioning properly, can tolerate the presence of some ER protein aggregates, as previously reported (Kruse et al., 2006a ). In upred-KAR2 cells, CPY* was found almost entirely in the pellet fraction, reflecting increased aggregation associated with the reduction of Kar2p–substrate interactions. Although it is tempting to speculate that the increase over wild type represents a threshold for toxicity, this notion remains to be tested.
The data presented support the notion that a key role of Kar2p in ER stress tolerance is to clear the ER of misfolded proteins. However, it was unclear whether the accumulation of misfolded CPY in the ER is in itself toxic or the consequence of also limiting Kar2p. Recently an independent study in our laboratory allowed a direct test of the question. We determined the structural elements of CPY* essential for its packaging into COPII vesicles. One preexisting variant, abcD-CPY* (CPY* lacking its three amino-proximal glycans), is entirely defective in COPII-mediated transport (Kawaguchi et al., 2010 ). Expressed at moderate levels, abcD-CPY* degrades efficiently by ERAD and exhibits no toxicity (Spear and Ng, 2005 ). Expressing a single copy of abcD-CPY* from the GAL1 promoter, however, leads to cell death even in wild-type (Kawaguchi et al., 2010 ; Figure 8A). To determine whether toxicity is a consequence of compromised protein clearance caused by disrupting the export signal, we analyzed wild-type strains expressing CPY* or the abcD variant. In pulse-chase experiments, CPY* is turned over efficiently (single copy of the GAL1-CPY* gene), whereas abcD-CPY* is stable (Figure 8B). In these cells, the proteins are continuously expressed from the GAL1 promoter. We next applied the clearance test. After preloading of wild-type cells, CPY* was rapidly cleared after glucose repression (Figure 8C). Clearance was more efficient here than in the experiment in Figure 5 because of lower CPY* expression. By contrast, abcD-CPY* persisted in the ER over long periods (Figure 8, B–D). Note that the export-defective abcD-CPY* in wild-type cells behaves similarly to CPY* in upred-KAR2 cells (compare Figures 5 and and8).8). To verify that the inability to exit through the ER-to-vacuole pathway is the cause of toxicity, we also analyzed the effects of ABCd-CPY*. This variant exhibits the reciprocal phenotype to abcD-CPY*, in that it is defective in ERAD (the D-glycan is part of the ERAD signal) but proficient in transport (Kostova and Wolf, 2005 ; Spear and Ng, 2005 ). Overexpression of ABCd-CPY* was well tolerated, emphasizing the importance of the vacuolar pathway under stress. By subverting the cell's ability to clear misfolded proteins at the substrate level, these data demonstrate that their accumulation in the ER is intrinsically toxic.
How the UPR promotes ER stress tolerance and cellular homeostasis remains unclear. Part of the problem lies in the complexity of the regulatory output. Even in simple budding yeast, the regulation of nearly 400 genes is dependent, at least in part, on the UPR (Travers et al., 2000 ; Patil et al., 2004 ). Although reports abound on the genetic and functional analyses of individual UPR targets, their physiological contributions to stress tolerance remain largely unexplored. The inherent problems of using coding sequence loss-of-function mutations can be overcome by uncoupling their regulation by the UPR. For UPR target genes, this can be accomplished simply by destroying their respective UPREs.
The BiP/Kar2p chaperone is a major UPR target that carries out essential constitutive functions. Therefore specifically disabling its UPR regulation allowed functional assessment of the induced state. The inability to boost BiP/Kar2p to stress levels (about fourfold for CPY* overexpression) led to poor substrate turnover and reduced cell viability when stressed. Analyses of CPY* showed decreased chaperone association and a concomitant increase in substrate aggregation. It was previously reported that ER chaperone mutants lead to substrate aggregation and a defect in ERAD (Nishikawa et al., 2001 ). Surprisingly, we observed considerable CPY* aggregation even in stress-tolerant wild-type cells. This contrasts with good solubility reported when CPY* is expressed at moderate levels (Nishikawa et al., 2001 ; Spear and Ng, 2003 ). Although it is tempting to speculate that aggregated CPY* traffics to the vacuole using the autophagic pathway as shown for α1-antitrypsin and fibrinogen aggregates (Kruse et al., 2006a , 2006b ), excess CPY* is degraded efficiently in all autophagy mutants tested (E. D. Spear and D. T. W. Ng, unpublished results). Because ERAD requires substrates to be soluble and COPII vesicles cannot accommodate large aggregates (Rivera et al., 2000 ), BiP/Kar2p might play a role in resolving aggregated species needed for turnover. Although this assertion remains to be tested, the bacterial DnaK and eukaryotic Hsp70 proteins, homologues of BiP/Kar2p, function to disaggregate misfolded proteins during heat stress (Skowyra et al., 1990 ; Glover and Lindquist, 1998 ; Goloubinoff et al., 1999 ).
The maintenance of essential Kar2p/BiP functions when the molecule is limiting was an unanticipated finding. This cannot be explained by functional redundancies because ER protein translocation and folding activities are eliminated by KAR2 temperature-sensitive loss-of-function alleles (Vogel et al., 1990 ; Simons et al., 1995 ). Elevating transcription of UPR target genes with overlapping functions could contribute. However, UPR activation by itself is insufficient to fully compensate because a constitutively activated UPR does not suppress KAR2 ts alleles (Supplemental Figure S2). A more provocative model has the chaperone deployed according to priority, with essential functions being the highest. This can be accomplished through differential binding affinities to substrates and/or cofactors (e.g., DnaJ family proteins). Indeed, when BiP/Kar2p is limiting, we observe a dramatic decrease of the chaperone associated with CPY* (Figure 7A). Although this result is consistent with the established functions of Kar2p, we were surprised to find that DTT-induced ER stress was well tolerated by the upred-KAR2 strain. This result suggests that the up-regulation of UPR-regulated oxidoreductase genes like ERO1 and PDI1 sufficiently restored oxidative protein folding so the need to neutralize misfolded proteins by Kar2p did not exceed its basal capacity (Frand and Kaiser, 1998 ; Pollard et al., 1998 ; Travers et al., 2000 ).
Even as disease models have unequivocally shown that aberrant proteins can be toxic, the molecular mechanisms of proteotoxicity remain mostly unresolved. One promising mechanism was deduced by inserting a disease-associated protein into an easily tractable genetic system. High α-synuclein levels, whose cytosolic aggregates are associated with some forms of Parkinson disease, shut down the essential secretory pathway in budding yeast (Cooper et al., 2006 ). Although the effect is not yet directly linked to the human disease, it illustrates how aberrant proteins can be harmful in unexpected ways. More recently, engineered amyloid-like proteins were shown to form interactions with key cellular factors and to disrupt the stress response (Olzscha et al., 2011 ). This raises the potential of widespread dysfunctions that can be profoundly harmful when combined. In this study, high levels of CPY* (albeit lower than in the α-synuclein model) are well tolerated in wild-type cells, indicating that the specified stress is calibrated below the toxicity threshold. These conditions were instrumental to reveal how the UPR regulation of KAR2 is a protective mechanism required to clear toxic proteins. Two independent lines of evidence indicate this to be a specific role of BiP/Kar2p stress regulation. First, only aberrant protein clearance is disrupted when KAR2 is uncoupled from the UPR (Figure 3, A and B). Second, crippling misfolded CPY's ability to enter the ER-to-vacuole pathway increases its intrinsic toxicity, evident from its ability to kill even wild-type cells (Figure 8). Without the alternative route facilitated by Kar2p, even reduced expression (single-copy gene) causes massive ER accumulation and cell death.
Although the data show that accumulation of CPY* in the ER is toxic, the mechanism of toxicity remains unclear. In previous studies, overexpression of CPY* was shown to kill cells lacking the cargo sorting receptor Erv29p (Spear and Ng, 2003 ; Haynes et al., 2004 ). Erv29p is required for transport of many folded cargo proteins like CPY and pro–α factor and is also needed for CPY* to use the ER-to-vacuole pathway (Belden and Barlowe, 2001 ; Spear and Ng, 2003 ). Of interest, Cooper and coworkers reported that overexpression of CPY* in Δerv29 cells leads to the accumulation of reactive oxygen species (ROS) that is associated with activation of the UPR (Haynes et al., 2004 ). Although this is an attractive mechanism, whether ROS contribute to the death of upred-KAR2 cells remains to be determined. Fundamentally, the results from the Δerv29 experiments cannot be easily extended to the present study. upred-KAR2 cells differ, in that all protein trafficking genes are intact and, accordingly, display no defects in the trafficking of normal proteins even under stress.
We developed a simple method to analyze the stress-activated state of a specific UPR target gene. Its application to KAR2 revealed a specific role of elevated chaperone levels to mitigate stress by ridding misfolded proteins through the ERAD and ER-to-vacuole pathways. Because the UPR regulates genes with functions as diverse as glycosylation and proteasome biogenesis, it is a major homeostatic pathway of the cell (Travers et al., 2000 ; Lee et al., 2003 ). This strategy can be used to analyze the role of nearly any UPR-regulated gene in cell physiology and, in principle, can be extended to other transcriptionally inducible pathways.
Yeast strains used in this study are listed in Supplemental Table S1. Monoclonal α-HA antibodies (HA.11) were purchased from Covance Research Products (Richmond, CA). Monoclonal α-Pgk1p antibodies were purchased from Molecular Probes (Eugene, OR). Polyclonal α-Kar2p and α-Sec61p antibodies were generously provided by Peter Walter (University of California, San Francisco, San Francisco, CA). Polyclonal α-Gas1p antibodies were raised against a glutathione S-transferase fusion protein containing the N-terminal amino acids 40–289 of Gas1p (Spear and Ng, 2003 ). α-DPAP B antibodies were a kind gift of Tom Stevens (University of Oregon, Eugene, OR). Goat horseradish peroxidase (HRP)–conjugated α-mouse immunoglobulin G (IgG) antibodies were purchased from Pierce Biotechnology (Rockford, IL). Donkey HRP-conjugated α-rabbit IgG antibodies were purchased from Jackson ImmunoResearch Laboratories (West Grove, PA). Goat α-rabbit IRDye 800 and goat α-mouse IRDye 680 were purchased from LI-COR Biosciences (Lincoln, NB). Alexa Fluor 488 goat α-mouse IgG and Alexa Fluor 594 goat α-rabbit IgG were purchased from Molecular Probes.
pES28. pES28 carries the GAL1 promoter–regulated, HA-tagged CPY* gene in YCp50 (Spear and Ng, 2003 ).
pCB11 and pCH49. An EcoRI/SalI fragment containing the intact, GAL1 promoter–regulated, HA-tagged CPY* gene from pES67 was inserted into pRS313 and pRS314 to create pCB11 and pCH49, respectively (Sikorski and Hieter, 1989 ; Spear and Ng, 2003 ).
pCH66. GAL1 promoter–regulated, HA-tagged abcD-CPY* was constructed by digestion of pES147 (Spear and Ng, 2005 ) with AccI and treatment with T4 DNA polymerase, followed by digestion with SphI. The fragment was ligated into BamHI (T4 DNA polymerase [DNAP]-treated)/SphI–digested pTS210, which carries the GAL1/10 promoter cassette (Marschall et al., 1996 ).
The KAR2 gene is toxic to Escherichia coli if inserted into bacterial high-copy-number vectors (Rose et al., 1989 ). This required the exclusive use of low copy plasmids. To generate the upred-KAR2 strain, we used pMR397 as a base, which contains the intact KAR2 gene in pMR366 shuttle vector (Vogel et al., 1990 ). PCR-based mutagenesis was used to replace sequences containing unfolded protein response element (5′-GGAACTGGACAGCGTGTCGA-3′; Mori et al., 1992 ) with the nonspecific sequence 5′-GTTCTCATGTTTGACAGCTT-3′ derived from a pBR322 intergenic region. The resulting plasmid, pCB10, is a functioning bacterial–yeast shuttle vector that carries the upred-KAR2 allele. To generate a genome-integrating version of pCB10, pMR397 was digested with BglII and KpnI to release sequences containing the autonomous replicating sequence and centromere from the plasmid. The vector was ligated following treatment with T4 DNAP, resulting in the pCB20 intermediate, which also functions as an integrating vector for the wild-type UPRE-KAR2 allele. A PvuI/NcoI fragment containing KAR2 and flanking sequences from pCB20 were replaced with the corresponding sequences from pCB10 to generate pCB18, the upred-KAR2 integrating vector.
To create the upred-KAR2 strain, MS785 (MATα, kar2L148::LEU2, pMR397; Rose et al., 1989 ) was crossed to YJL183 (MATa, ura3Δ99, leu2Δ1, trpΔ99, ade2-101ochre; Ng et al., 1996 ) to generate a KAR2/kar2L148::LEU2 heterozygous diploid. The pMR397 plasmid was dropped from the diploid strain by counterselection using 5-fluoroorotic acid. Next pCB18 was cleaved at its unique NcoI site in the URA3 marker gene and transformed into the diploid strain. Uracil prototrophs containing the integrated allele URA3::upred-KAR2 were sporulated for tetrad dissection. A haploid strain containing the kar2L148::LEU2 and URA3::upred-KAR2 alleles (scored as leucine and uracil prototrophs) was isolated and backcrossed to the W303 strain (MATa, ura3-1, leu2-3, his3-11, trp1-1, can1-100, ade2-1) six times. The identical procedure was performed in parallel using pCB20 to generate the UPRE-KAR2 control strain. Because the kar2L148::LEU2 allele is not a complete knockout, we discovered that an amino-terminal fragment of Kar2p containing the full-length signal sequence was expressed in these cells and causing ER stress. To eliminate the expression of this mutant fragment, the Δkar2::KanMX allele (complete deletion of KAR2-coding sequences) was amplified by PCR from a KAR2/Δkar2::KanMX heterozygous strain (Winzeler et al., 1999 ) using CH20 (5′-AGGAACTGGACAGCGTGTCGAA-3′) and CH21 (5′-CAACCTTGAAGCTTCCAGCAGC-3′) primers. The purified PCR product was used directly to knock out the kar2L148::LEU2 alleles in both upred-KAR2 and UPRE-KAR2 strains to create CHY220 and CHY438, respectively.
Cells were grown at 30°C in YPD and harvested at early log phase (≤0.4 OD600/ml). Cells were treated with 2.5 μg/ml tunicamycin or DMSO for 1 h. Two OD600 units of cells were harvested, and total RNA was isolated using an RNeasy Kit (Qiagen, Valencia, CA) according to the manufacturer's protocol. cDNA was prepared from 1 μg of total RNA using the SuperScript III First Strand Synthesis system (Invitrogen, Carlsbad, CA) and oligo(dT) primers according to the manufacturer's protocol. Unspliced and spliced cDNAs were amplified by PCR using primers flanking the intron: HAC1F (GT97), 5′-TCGCACTCGTCGTCTGATA, and HAC1R (GT102), 5′-TCATGAAGTGATGAAGAAATCATTCACT1, cDNA was amplified as a loading control by using primers, Act1F (GT103), 5′-GGTTGCTGCTTTGGTTATTGA, and Act1R (GT104), 5′-TTTTGACCCATACCGACCAT. Products were separated by 1% agarose gel electrophoresis and images acquired with a Gel Doc system (Bio-Rad, Richmond, CA). Quantification was performed using ImageQuant TL software (GE Healthcare Life Sciences, Uppsala, Sweden). All data reflect three independent experiments, with the SD indicated as error bars. A representative gel image from a single experiment is shown.
Test strains were transformed with pCB11 and pCH49 or a single copy of pES28 or pCH66 (each of these plasmids carry a single copy of GAL1-CPY*). Cells were grown at 30°C overnight in SC medium containing the appropriate amino acids and 3% raffinose to mid log phase. Cells were harvested, washed, and transferred into SC medium containing the appropriate amino acids and 2% galactose for 6 h to induce CPY* synthesis before further processing. Δpep4 and upred-KAR2/Δpep4 cells were grown in galactose-containing media for 8 h before processing for indirect immunofluorescence microscopy.
For each strain, 3.0 A600 OD equivalents of cells were disrupted in 1 ml of 10% trichloroacetic acid (TCA) and 0.4 ml of 0.5-mm zirconium beads with 2× 30-s pulses in a Mini-Beadbeater-8 cell homogenizer (BioSpec Products, Bartlesville, OK). The homogenate was transferred to a new tube and pooled with a 150 μl of bead wash. After centrifugation, the pellet was resuspended in 150 μl TCA resuspension buffer (100 mM Tris base, 3% SDS, 1 mM phenylmethylsulfonyl fluoride [PMSF]) and incubated at 100°C for 10 min. Insoluble particles are removed by centrifugation. A lysate equivalent of 0.02 A600 OD units of cells was used for each sample in Western analysis. The samples were resolved by SDS–PAGE and then transferred to nitrocellulose membrane (GE Healthcare Life Sciences). The membrane was blotted in blotting buffer (5% nonfat dry milk in PBST [137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4·2H2O, 2 mM KH2PO4, 0.1% Tween-20]). α-Kar2p and α-Pgk1p were used at 1:10,000 and 1:1000 dilutions in blotting buffer, respectively. Secondary antibodies were α-rabbit IRDye 800 and α-mouse IRDye 680 used at 1:10,000 in 0.5% nonfat dry milk in PBST. Immunoblots were dried and scanned using the LI-COR Odyssey Infrared Imaging system, which allows for simultaneous dual-wavelength detection (169-μm resolution, 0.0-nm focus offset, 1.5 intensity for 700 nm, 2.5 intensity for 800 nm).
The GAL1-CPY* gene was induced as described. After 6 h, glucose was added to the culture media to repress synthesis and incubated for the times indicated. CPY* levels were analyzed by immunoblotting as described, except that detection was performed using enhanced chemiluminescence (ECL) according to manufacturer's protocols (Pierce Biotechnology). α-HA (to detect CPY*) and α-Kar2p antibodies were used at 1:10,000 dilutions. HRP-conjugated goat α-mouse IgG and HRP-conjugated donkey and HRP-conjugated α-rabbit IgG were used at 1:10,000 dilutions.
The 3.0 A600 OD cell equivalents grown to log phase were collected and resuspended in 0.9 ml of synthetic media lacking methionine and cysteine. After 30 min of incubation at the appropriate temperature, 150 μCi of [35S]Met/Cys (Pro-Mix; Amersham Pharmacia Biotech, Piscataway, NJ) was added to cells for 5 or 10 min, as indicated. A cold chase was initiated by adding unlabeled methionine/cysteine to 2 mM. A 100 μl amount of ice-cold 100% trichloroacetic acid was added to terminate the chase. Cells were homogenized by adding 0.4 cc of 0.5-mm zirconium beads and agitation in a Mini-Beadbeater-8 cell disrupter for 2× 30-s cycles (BioSpec Products). The homogenate was transferred to a fresh tube and pooled with a subsequent 10% TCA bead wash. Following centrifugation, the pellet was resuspended in 120 μl of TCA resuspension buffer (100 mM Tris base, 3% SDS, 1 mM PMSF) and heated to 100°C for 5 min. Insoluble debris was pelleted, and 40 μl of the detergent lysate was transferred to 560 μl of IPS II (1% Triton X-100, 50 mM Tris pH 7.5, 1 mM PMSF, 1 μl of yeast protease inhibitor cocktail [Sigma-Aldrich, St. Louis, MO]) and the appropriate antibody. Following a 120-min incubation at 4°C, the sample was centrifuged and the supernatant transferred to a fresh tube containing protein A–Sepharose beads. The tube was rotated for 30 min and washed 5× with IPS I (0.2% SDS, 1% Triton X-100, 50 mM Tris pH 7.5) and 1× with phosphate-buffered saline. Immunoprecipitated proteins were eluted with gel sample buffer, separated by SDS-gel electrophoresis, and visualized/quantified using a Typhoon phosphorimager (GE Healthcare Life Sciences).
Cell labeling, homogenization, and TCA precipitation were carried out as described. The TCA precipitate pellet was washed once with cold acetone and resuspended in Mal-PEG reaction solution (5 mM Mal-PEG [Fluka, Sigma-Aldrich], 100 mM Tris, pH 7.4, 2% SDS) at 50 μl per A600 OD cell equivalent. The mixture was incubated at 100°C for 10 min with occasional agitation on a vortex mixer. The tube was then left on ice for 50 min and again heated to 100°C for 10 min. After centrifugation, the cell lysate was used for immunoprecipitation as described. For DTT-treated cells, DTT was added directly to the culture media to 10 mM for 20 min prior to metabolic labeling.
This assay was modified from that previously described by Carvalho et al. (2006 ). Briefly, 40 A600 OD units of log-phase cells were harvested, washed once with ice-cold water, and resuspended in 500 μl of TBS IP buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 2 mM PMSF, 0.3 μl of yeast protease inhibitor cocktail [Sigma-Aldrich] per A600 OD unit). Cell disruption was performed by agitation with 0.5 mm zirconium beads in the Mini-Beadbeater-8 (3× 15 s with 5-min intervals on ice). The lysate was transferred to a new tube and pooled with a subsequent 500 μl of TBS IP buffer bead wash. After centrifugation at 30,000 × g for 30 min, the pellet was collected, resuspended in 1 ml of TBS IP buffer containing 1% Triton X-100, and incubated for 30 min on ice. The detergent lysate was clarified by centrifugation at 30,000 × g for 10 min. Five μl of HA-probe mouse monoclonal IgG (Santa Cruz Biotechnology, Santa Cruz, CA) was added to the supernatant and incubated for 30 min. Twenty-five microliters of protein A–Sepharose beads (Sigma-Aldrich) was added into the tube and incubated for 2 h with rocking. Beads were washed three times with 1 ml of TBS IP buffer containing 1% Triton X-100 and once with 1 ml of TBS IP buffer (1×). Immunoprecipitated proteins were eluted by boiling in SDS-loading buffer for 10 min, resolved on SDS–PAGE, and analyzed by immunoblotting.
Wild-type, Δire1, and upred-KAR2 cells carrying the GAL1-CPY* gene were grown and galactose induced as described, except that Δire1 cell media was supplemented with 50 μg/ml myoinositol. Five OD600 units of cells were harvested, washed once in water, and resuspended in 500 μl of TNE buffer (50 mM Tris, pH7.4, 150 mM NaCl, 5 mM EDTA) containing 1 mM PMSF and 1.5 μl of yeast protease inhibitor cocktail. Cells were disrupted by beating with 0.5-mm zirconium beads (10× 1 min, with 5 min between each interval on ice) using a vortex mixer at full speed at 4°C. Low-speed centrifugation (750 × g, 5 min) was performed twice to remove cell debris. Membranes in the supernatant fraction were solubilized by adding Triton X-100 1% followed by incubation at room temperature for 5 min. Fifty microliters of this material (T, total) was saved. The remaining lysate was centrifuged at 100,000 × g for 15 min at 4°C. The supernatant fraction (S) was removed and saved. The pellet fraction (P) was resuspended in 450 μl of 3% SDS and 50 mM Tris, pH 7.5, and boiled at 100°C for 5 min. Fifty microliters of each fraction (T, S, and P) was resolved by SDS–PAGE and proteins detected using ECL-based Western analysis as described.
GAL1-CPY* expression was induced in cells as described. CPY* synthesis was terminated by adding 2% glucose and incubated for various times indicated. Formaldehyde was added directly to the culture media to 3.7%. After incubation at 30°C for 90 min, cells were washed once with 0.1 M potassium phosphate, pH 7.5, and treated with zymolyase (1 mg/ml zymolyase 20T [ICN Biomedicals, Irvine, CA] in 0.1 M potassium phosphate, pH 7.5, and 1.2 M sorbitol) for 30 min at room temperature. The resulting spheroplasts were washed twice with 0.1 M potassium phosphate, pH 7.5, and 1.2 M sorbitol. Cells were applied to a 0.1% poly-l-lysine–coated multi-well slide, incubated at room temperature for 10 min, and washed once with TBS buffer (50 mM Tris, pH 7.5, 150 mM NaCl). The slide was immersed in cold methanol for 6 min at −20°C, transferred into cold acetone for 30 s, and allowed to equilibrate in TBS for 3 min at room temperature. Thirty microliters of TBS-blocking buffer (5% nonfat dry milk in TBS buffer) was added to each well, incubated for 30 min, and then washed once with TBS. Thirty microliters of primary antibodies was applied on wells for 90 min and washed twice with TBS afterward. Thirty microliters of secondary antibodies was then applied on wells for 90 min, followed by two TBS washes. After drying, 5 μl of mounting media (phosphate-buffered saline, 90% glycerol, 1 mg/ml p-phenylenediamine, 0.025 μg/ml 4′,6-diamidino-2-phenylindole [DAPI]) was added to each well before sealing. Samples were viewed on Zeiss LSM 5 Exciter upright microscope with a Zeiss PlanApoChromatic 100× oil immersion lens (numerical aperture, 1.4; Carl Zeiss MicroImaging, Jena, Germany). Images were acquired and processed with LSM Image Browser software (Carl Zeiss MicroImaging). Primary antibodies α-HA (for CPY*) and α-Kar2p were used at 1:500 and 1:1000 dilutions, respectively, in TBS-blocking buffer. Secondary antibodies Alexa Fluor 488 goat α-mouse IgG and Alexa Fluor 594 goat α-rabbit were diluted 1:500 in blocking buffer for working concentrations.
Cells transformed with pJC31, a plasmid carrying the UPRE-LacZ reporter gene (Cox and Walter, 1996 ), were grown to log phase and treated with 2.5 μg/ml tunicamycin for 1 h. The 3.0 OD600 units were collected, washed with 1 ml of Z buffer (0.06 M Na2HPO4·7H2O, 0.04 M NaH2PO4·H2O, 0.01 M KCl, 0.001 M MgSO4·7H2O, 50 mM β-mercaptoethanol) and pelleted by low-speed centrifugation. The cells were resuspended in 50 μl of Z buffer, and then 50 μl of CHCl3 and 20 μl of 0.1% SDS were added. The mixture was vortexed hard for 20 s. Seven hundred microliters of 2.0 mg/ml ο-nitrophenyl-β-galactopyranoside (in Z buffer) was added into the mixture and incubated at 30°C for 1–10 min with time of incubation recorded. The reaction was quenched by adding 500 μl of 1 M Na2CO3. After a quick spin, the A420 OD of the supernatant was measured. β-Galactosidase activity was expressed in Miller units as 1000(A420 OD)/[tminVmL(A600 OD units of cells)] (Guarente, 1983 ).
To induce the UPR, cells were treated with 2.5 μg/ml tunicamycin or 10 mM DTT for 1 h. Total RNA was isolated using the hot phenol method. Briefly, resuspended cells in 22 ml of 50 mM NaOAc and 10 mM EDTA were disrupted with the addition of 2 ml of 10% SDS and 25 ml of phenol and shaking at 65°C. Following phase separation by centrifugation, the phenol phase was removed and the extraction was repeated by addition of 25 ml of phenol. The aqueous phase was transferred to another tube for extraction using chloroform. Total RNA was collected by ethanol precipitation from the aqueous phase. Equal amounts of RNA were separated by formaldehyde agarose gel electrophoresis and transferred to nitrocellulose using the buffer wicking method. DNA for probe preparation was amplified by PCR and purified from agarose gels. Probes were labeled with [α-32P]dCTP using Ready-To-Go DNA labeling beads (GE Healthcare Life Sciences). The KAR2-specific probe was prepared using the forward primer 5′-ACAGACTAAGCGCTGGCAAGCT-3′ and the reverse primer 5′-CAGCATGGGTAACCTTAGAGCC-3′ to obtain a 550–base pair fragment corresponding to the 5′ end of the coding sequence. The ACT1-specific probe was generated using the forward primer 5′-ATCGTCGGTAGACCAAGA CACC-3′ and the reverse primer 5′-CGAAGTCCAAGGCGACGTAACA-3′ to obtain a 559–base pair fragment corresponding to the 5′-end of the coding sequence.
We are grateful to Jeremy Brodhead, Sandy Toh, and TLL facilities for excellent technical support. We thank Randy Schekman (University of California, Berkeley, CA), Mark Rose (Princeton University, Princeton, NJ), Tom Stevens (University of Oregon, Eugene, OR), and Peter Walter (University of California, San Francisco, CA) for gifts of strains and antibodies. This research was supported by funds from the Temasek Trust and the Singapore Millennium Foundation. The early stage of this work was supported by National Institutes of Health Grant GM059171 to D.T.W.N. and facilities at the Pennsylvania State University.
This article was published online ahead of print in MBoC in Press (http://www.molbiolcell.org/cgi/doi/10.1091/mbc.E11-04-0297) on December 21, 2011.