|Home | About | Journals | Submit | Contact Us | Français|
The yeast cell wall is an extracellular structure that is dependent on secretory and membrane proteins for its construction. We investigated the role of protein quality control mechanisms in cell wall integrity and found that the unfolded protein response (UPR) and, to a lesser extent, endoplasmic reticulum (ER)-associated degradation (ERAD) pathways are required for proper cell wall construction. Null mutation of IRE1, double mutation of ERAD components (hrd1Δ and ubc7Δ) and ire1Δ, or expression of misfolded proteins show phenotypes similar to mutation of cell wall proteins, including hypersensitivity to cell wall-targeted molecules, alterations to cell wall protein layer, decreased cell wall thickness by electron microscopy, and increased cellular aggregation. Consistent with its important role in cell wall integrity, UPR is activated by signaling through the cell wall integrity mitogen-activated protein (MAP) kinase pathway during cell wall stress and unstressed vegetative growth. Both cell wall stress and basal UPR activity is mediated by Swi6p, a regulator of cell cycle and cell wall stress gene transcription, in a manner that is independent of its known coregulatory molecules. We propose that the cellular responses to ER and cell wall stress are coordinated to buffer the cell against these two related cellular stresses.
The yeast cell wall is an essential structure that protects the cell from lysis during budding, mating, and periods of environmental stress (Klis et al., 2002 ). The yeast cell wall is a complex lattice of protein and carbohydrate that is constructed from, and by, proteins delivered by the secretory pathway (Lesage and Bussey, 2006 ). Accordingly, genome-wide screens for genes involved in yeast cell wall biosynthesis have consistently identified protein and vesicular trafficking genes (Firon et al., 2004 ). In addition, cellular secretion has been specifically shown to be required for the incorporation of mannoprotein (Conde et al., 2003 ), 1,3-β-glucan (Abe et al., 2003 ), 1,6-β-glucan (Brown et al., 1993 ), and chitin (Santos and Synder, 1997 ) into the yeast cell wall.
Protein quality control, in contrast, is one important aspect of cellular secretion that has not been extensively studied in terms of its role in yeast cell wall integrity. Misfolded or misprocessed proteins are toxic to the cell; consequently, mechanisms to identify, refold, and/or remove such proteins are crucial for cellular viability (Sayeed and Ng, 2005 ). Two general mechanisms of protein quality control are operative in yeast as well as in higher eukaryotes. The first mechanism is endoplasmic reticulum (ER)-associated degradation (ERAD; Meusser et al., 2005 ). Through ERAD, proteins resistant to chaperone-mediated refolding are identified, retro-translocated from the ER, tagged with ubiquitin, and, ultimately, degraded by the 26S proteasome. ERAD is constitutively active and, during unstressed vegetative growth, seems sufficient to process the load of misfolded proteins in yeast (Spear and Ng, 2001 ).
When the cell encounters conditions that increase unfolded proteins, a second mechanism called the unfolded protein response (UPR) is activated to compensate for elevated levels of ER stress (Ron and Walter, 2007 ). UPR is an ER-to-nucleus signaling pathway that is initiated by ER stress and induces the transcription of a large number of genes. In yeast, UPR is triggered when unfolded proteins are detected by the transmembrane sensor Ire1p. Ire1p contains protein kinase and endoribonuclease activities that are essential to its role in UPR (Cox et al., 1993 ; Mori et al., 1993 ). Ire1p oligomerizes in the presence of unfolded proteins and undergoes autophosphorylation, which in turn activates its RNase activity (Shamu and Walter, 1996 ). Ire1p RNase activity is specific for the mRNA of the transcription factor Hac1p, its only known substrate. In yeast, HAC1u mRNA (“u” for uninduced) is constitutively transcribed but is not translated due to the presence of an inhibitory intron. Activated Ire1p removes the intron from HAC1u and tRNA ligase rejoins the two exons to generate HAC1i (”i“ for induced). HAC1i is then efficiently translated and the resulting Hac1p transcription factor translocates to the nucleus where it initiates the transcriptional program of UPR (Travers et al., 2000 ).
Initially, the role of UPR was believed to be limited to protein quality control, but it has become clear that UPR plays a much broader role in cellular physiology (Sayeed and Ng, 2005 ). For example, UPR has been linked to cytokinesis (Bicknell et al., 2007 ), autophagy (Bernales et al., 2006 ), haploid tolerance (Lee et al., 2003 ), pseudohyphal growth (Schroder et al., 2000 ), and lipid biosynthesis and membrane homeostasis (Cox et al., 1997 ). Consistent with its broad role in cellular physiology, the transcriptional program of UPR includes genes involved in a wide range of cellular processes, including protein folding, ERAD, protein trafficking, lipid biosynthesis, and cell wall architecture (Travers et al., 2000 ).
Although the role of UPR and, more generally, secretory protein quality control in yeast cell wall biosynthesis has not been extensively studied, a number of reports have provided evidence for a link between these two important processes in yeast. First, ER stress has been shown to trigger signaling through the cell wall integrity (CWI) mitogen-activated protein kinase (MAPK) signaling cascade (Bonilla and Cunningham, 2003 ), the most important mediator of the cellular response to cell wall stress (Levin, 2005 ). Second, null mutations in the components of the CWI pathway are hypersensitive to ER stress (Chen et al., 2005 ). Third, Slt2/Mpk1p (hereafter noted as Mpk1), MAPK of the CWI pathway, stabilizes Hac1p during the concurrent application of heat stress and ER stress, leading to increased UPR activation during periods of dual stress (Pal et al., 2007 ). Fourth, activation of Pkc1p has been reported to increase the degradation of misfolded carboxypeptidase Y (CPY*), suggesting cell wall stress may increase the capacity of the quality control systems during periods of stress (Nita-Lazar and Lennarz, 2005 ). Finally, cytosolic molecular chaperones such as Ydj1p, an Hsp40 chaperone (Wright et al., 2007 ), and Sse1p, an Hsp110 chaperone (Shaner et al., 2008 ), affect cell wall integrity through interactions with CWI pathway components such as Mpk1p (Sse1p) and Pkc1p (Ydj1p).
Here, we show that ire1Δ mutants as well as strains expressing misfolded proteins have defects in cell wall integrity. In addition, double mutants that combine ire1Δ with mutations in components of the ERAD pathway show synthetic cell wall defects. We also demonstrate that cell wall stress activates UPR in a process dependent on both Ire1p and components of the CWI pathway. Finally, we provide evidence that both cell wall stress-induced and basal UPR activity is mediated by Swi6p in a process that is independent of its known transcriptional coactivators. Together, these results indicate that UPR and CWI signaling pathways form an interdependent regulatory circuit that allows yeast to respond to both intracellular ER stress and extracellular cell wall stress.
Yeast strains used in this study are listed in Table 1 and were derived from either S288C (BY4741) or W303 (CRY1) genetic backgrounds. Single yeast mutants in the BY4741 background were obtained from the yeast deletion collection (Invitrogen, Carlsbad, CA). Yeast strains were grown in yeast peptone dextrose (YPD) or synthetic drop out medium prepared according to standard recipes (Burke et al., 2000 ). Expression of alleles from galactose-regulated promoters was induced by preculturing strains in 2% raffinose for 8 h before addition of 2% galactose. All incubations were performed at 30°C unless noted otherwise. DH5α Escherichia coli cells (Invitrogen) were used for routine plasmid manipulations. All chemicals were obtained from Sigma-Aldrich (St. Louis, MO) and used as received. Oligonucleotides were synthesized by Integrated DNA Technologies (Corralville, IA).
The IRE1 open reading frame was deleted by standard polymerase chain reaction (PCR)-based one step gene-replacement with a KANMX cassette generated from the deletion collection ire1Δ::KAN strain (Johnston et al., 2002 ). Correct integration was determined by PCR with primers flanking the sequence used to generate the knockout cassette. All ire1Δ mutants displayed the expected hypersensitivity to deoxyglucose (Back et al., 2005 ). The KANMX cassette was switched to the NATMX cassette according to a published procedure (Tong and Boone, 2006 ).
pJC104 contains four copies of the unfolded protein response element (4X-UPRE) fused to the lacZ gene in a CEN/URA3 plasmid and was generously provided by P. Walter (Cox and Walter, 1996 ). pMB4 contains the HAC1 promoter and open reading frame fused to a 6X hemagglutinin (HA) tag at the 3′ end in a CEN/LEU2 plasmid and was a gift of K. Cunningham (Johns Hopkins University, Baltimore, MD; Bonilla and Cunningham, 2003 ). pBD1265 contains SWI6 under the control of its endogenous promoter in a 2μ plasmid with a LEU2 marker and was a gift of L. Breeden (University of Washington, Seattle, WA).
Agar plates supplemented with Calcofluor white (CFW), Congo red, caffeine, and caspofungin were prepared according to recipes reported previously (Krysan et al., 2005 ). Overnight cultures of each strain were adjusted to a cell density of 1 OD600 unit. A three-step, 10-fold dilution series was prepared and spotted on to plates by using a metal frogger (VP Scientific, San Diego, CA). The plates were incubated at 30°C or 37°C for 2–5 d and photographed. Two independent isolates of each strain were tested in duplicate for each phenotype. Zymolyase sensitivity was assayed as described previously (de Nobel et al., 2000 ).
Yeast were grown overnight to early stationary phase in YPD. The culture was diluted in an equal volume of water and placed on a microscope slide. The number of cell aggregates with >10 cells per high-power field was counted for 50 high-power fields on three separate days for each strain. Reported values are the mean and SD of the three separate replicates.
Dithiothreitol (DTT)-extracts of intact cells were prepared according to a previously described protocol (Klis et al., 2007 ). Briefly, cells were grown to early stationary phase (OD600 ~10) overnight in YPD, harvested by centrifugation (300 × g), washed with water, and resuspended (1 OD600 equivalent/μl) in extraction buffer (50 mM Tris, pH 7.5, and 5 mM DTT). The cell suspension was shaken in a multi-vortex apparatus for 2 h at 4°C. The supernatant was removed, mixed with 2× Laemelli sample buffer, boiled for 5 min at 100°C, and fractionated by SDS-polyacrylamide gel electrophoresis (PAGE) (4–15% gradient gel; Bio-Rad, Philadelphia, PA). Fractionated proteins were visualized by silver stain according to the manufacturer's protocol (Silver Snap kit; Bio-Rad). Gels were photographed, and images were processed using Adobe Photoshop software (Adobe Systems, Mountain View, CA).
NaOH extracts of isolated cell wall material were prepared according to a published procedure (Kapteyn et al., 1999 ). Briefly, cells were cultivated and harvested as described for the DTT extracts. Cell walls were isolated, suspended in β-mercaptoethanol/SDS extraction buffer (50 mM Tris, pH 7.8, 0.1M EDTA, 2% SDS, and 40 mM β-mercaptoethanol), and heated to 100°C for 5 min. The cell walls were pelleted, washed with water three times, and resusupended in 30 mM NaOH (1 mg wet cell wall/μl). The cell wall suspension was incubated overnight at 4°C. The resulting supernatant was removed, mixed with 2× Laemelli sample buffer, and processed as described for the DTT extracts.
Strains were grown to stationary phase and harvested. Yeast were fixed and processed according to previously published methods but with modifications to the resin type and resin infiltration/embedding steps (Gammie et al., 1998 ). Briefly, the yeast were fixed 30 min in 40 mM potassium phosphate buffered 2.0% glutaraldehyde containing 1 mM CaCl2, 1 mM MgCl2, and 0.2 M sorbitol. The yeast were pelleted, resuspended in buffer, incubated for 4 h in 4.0% potassium permanganate, rinsed four times in water, resuspended in sodium periodate, washed in potassium phosphate buffer and ammonium phosphate, rinsed in water, resuspended in 2.0% uranyl acetate (covered with foil to exclude light), and incubated overnight at 4°C on a rotator plate. The yeast were then washed in water and dehydrated in 60-min steps in a graded series of ethanol:water solutions to 100:0%. Spurr epoxy resin was substituted for the LR White used in the published procedure (Gammie et al., 1998 ), and the infiltration of the resin involved overnight incubations (on rotator) in an increasing concentration series (1:1, 1:2, 1:3, 1:4, and 100% resin) of Spurr resin diluted in 100% ethanol. The yeast were pelleted into fresh resin in Eppendorf tubes and polymerized overnight at 60°C. Thin sections were cut at 70 nm onto grids (no staining of grids) and examined with a Hitachi 7650 transmission electron microscopy by using a Gatan Erlangshen 1000SW digital camera.
Overnight cultures of yeast strains harboring pJC104 were diluted to 0.1 OD600 in YPD and grown for two-doublings (~3 h). Following previous experiments using this reporter (Cox and Walter, 1996 ) to assay UPR response, cultures were exposed to stressor for 5 h, harvested, and the pellets flash-frozen in liquid nitrogen. Cell lysates were prepared and β-galatosidase activity determined as described previously (Burke et al., 2000 ). The specific activity for the extracts is expressed as nanomoles per milligram of protein per minute (Miller units). Each reported value represents the mean of at least two experiments with independent transformants of the strain performed in duplicate or triplicate. Error is expressed as SE of the mean.
HAC1 mRNA splicing was assayed using a modification of a reported protocol (Bicknell et al., 2007 ). Overnight cultures were grown to mid-log phase (0.5–1.0 OD600) in YPD and treated with tunicamycin or CFW for 1 h. The cells were harvested, and total RNA was isolated using an RNA Easy kit (QIAGEN, Valencia, CA) according to the manufacturer's protocol. cDNA was generated from 2 μg of total RNA using the SuperScript III First Strand Synthesis system (Invitrogen) and oligo(dT) primers according to the manufacturer's protocol. HAC1u and HAC1i cDNA was amplified by PCR using primers flanking the intron (Bicknell et al., 2007 ): HAC1F, 5′-TAGAGGGATTTCCAGAGCACG and HAC1R, 5′-TCATTGAAGTGATGAAGAAATC and the following thermocycle conditions: 94°C, 1 min; 25 cycles of [94°C, 1 min; 54°C, 1 min; 65°C 1 min]; 65°C, 7 min. Amplicons were analyzed by electrophoresis on 2% agarose gels. ACT1 cDNA was amplified as a loading control by using previously reported primers and conditions (Zhong and Greenberg, 2003 ). All RT-PCR experiments were performed at least three times with independent strain isolates.
Hac1p levels were determined with cells harboring pMB4 following the procedure described previously (Bonilla and Cunningham, 2003 ). Briefly, logarithmic phase cultures in synthetic dextrose medium lacking leucine were treated with either tunicamycin or CFW, incubated for 2 h, harvested, and flash frozen. Cell extracts (25–50 OD600 equivalents) were prepared by glass beads lysis in sorbitol breaking buffer (300 mM sorbitol, 100 mM NaCl, 5 mM MgCl2, 10 mM Tris pH 7.5, mini-complete protease inhibitors [Roche Diagnostics, Indianapolis, IN]). Extracts were normalized by protein concentration (Bradford assay; Bio-Rad), fractionated by SDS-PAGE electrophoresis (10% gel), and transferred to nitrocellulose membranes. The membranes were blocked in TBST (50 mM Tris, pH 7.5, 150 mM NaCl, and 0.05% Tween 20) with 5% nonfat dry milk for 2 h at room temperature and then incubated overnight at 4°C in the same buffer containing a 1:5000 dilution of mouse anti-HA antibody (12CA5; GE healthcare, Little Chalfont, Buckinghamshire, United Kingdom). The TBST-washed membranes were then incubated with a 1:5000 dilution horseradish peroxidase-conjugated anti-mouse secondary antibody (GE Healthcare) in TBST with 5% nonfat dry milk, washed with TBST and developed for chemiluminescence with an ECL-Plus detection kit (GE Healthcare).
To test the hypothesis that ER quality control mechanisms affect cell wall biosynthesis, we compared the growth of an ire1Δ mutant to WT in the presence of the cell wall perturbing agents CFW and Congo red. ire1Δ (W303 background) is hypersensitive to CFW (Figure 1A) and Congo red (Supplemental Figure S1A). Because the phenotypes of ire1Δ and cell wall-related mutants can vary with strain background (Larson et al., 2002 ), we also tested an ire1Δ mutant constructed in the S288c background. Although the S288c-derived ire1Δ strain was less sensitive to CFW than the W303 ire1Δ mutant, it showed a growth defect when exposed to the combined stress of CFW and elevated growth temperature (Figure 1B).
A well known difference between the W303 and S288c strain backgrounds is a polymorphism at the SSD1 locus. SSD1 is involved in cell wall integrity (Kaeberlein and Guarente, 2002 ) and W303 has a hypofunctional, truncated allele (ssd1-d), whereas S288c has a functional, full-length allele (ssd1-v). As shown in Figure 1B, the S288c-derived ssd1Δire1Δ mutant recapitulates the CFW phenotype of W303-derived ire1Δ, suggesting that the background dependence of ire1Δ CFW sensitivity is related to differences in the functionality of SSD1. The hypersensitivity of ire1Δ mutants to cell wall perturbation and the genetic interaction of IRE1 with SSD1, a gene involved in cell wall integrity, support our hypothesis that IRE1 is required for proper yeast cell wall biosynthesis.
Because Ire1p is required for the cell to effectively degrade unfolded proteins during periods of stress (Casagrande et al., 2000 ), it seemed likely that the cell wall defects associated with ire1Δ mutants may be due to the accumulation of misfolded or misprocessed proteins. If that was correct, then expression of a misfolded protein would be expected to also cause cell wall defects. To test this hypothesis, we examined the growth of strains expressing the well-studied misfolded protein, carboxypeptidase Y (CPY*) from a galactose-inducible promoter on plates containing CFW. Spear and Ng have shown that expression of CPY* at this level saturates ER quality control mechanisms and causes severe ER stress (Spear and Ng, 2003 ). In wild-type (WT) cells, high level expression of CPY* causes hypersensitivity to CFW (Figure 1C), indicating that saturation of ER quality control mechanisms with a misfolded protein compromises cell wall integrity. It is important to note that, although ire1Δ mutants in some strain backgrounds grow poorly on galactose medium, the growth of ire1Δ on YPG is only slightly reduced relative to WT cells and that its growth is clearly decreased by the addition of CFW to YPG, indicating that the phenotype is due to the presence of CFW (Larson et al., 2002 ).
To confirm the effect of misfolded protein accumulation on cell wall integrity, we also tested a second misfolded protein, the ERAD substrate referred to as KWW. KWW is a chimeric fusion of the luminal portion of the simian virus 5 hemagglutinin-neuraminidase ectodomain (KHN) with the transmembrane and cytosolic domains of Wsc1p (KWW = KHN/Wsc1p transmembrane domain/Wsc1p cytosolic domain; Vashit and Ng, 2004 ). Indeed, expression of KWW from the ACT1 promoter also increased the CFW sensitivity of both WT and ire1Δ cells, demonstrating that this effect is not unique to CPY* (Supplemental Figure S1B). These results suggest that uncompensated ER stress generated by either loss of UPR or the accumulation of a misfolded protein leads to defects in cell wall integrity.
Next, we examined the stability of ire1Δ and CPY*-expressing strains to treatment with the cell wall-degrading enzyme, zymolyase to confirm the cell wall effects of ER stress by using a second standard cell wall phenotype (de Nobel et al., 2000 ). Consistent with the spot dilution assays, ire1Δ, GAL-CPY*, and ire1Δ GAL-CPY* are more easily lysed by zymolyase than WT (Figure 1D). We note that ire1Δ GAL-CPY* strains do not grow on galactose-containing agar plates (Spear and Ng, 2003 ) but, in our hands, slowly grow for four to five doublings in liquid medium, permitting the assays shown here and below.
Although the ER is involved in a variety of cellular processes that could, in principle, affect cell wall integrity, it seemed likely that cell wall protein homeostasis might be most directly affected by uncompensated ER stress. To test this hypothesis, we carried out a set of experiments to characterize the protein fraction of the cell wall in ER-stressed strains. As shown in Figure 2A, the total amount of protein per milligram of cell wall is increased in these mutants; indeed, twice as much protein is present in the cell wall of the ire1ΔGAL-CPY* strain relative to WT. Consistent with this finding, ire1ΔGAL-CPY* also stained much more intensely than WT when treated with fluorescein isothiocyanate-labeled concanavalin A, a lectin that binds mannoprotein (Supplemental Figure S2A). Furthermore, ER-stressed strains formed large cellular aggregates (Figure 2B) and displayed slightly increased flocculation (Supplemental Figure S2B). Because intracellular interactions responsible for aggregation are mediated by the outermost, mannoprotein layer of the cell wall (Ferreira et al., 2006 ), this result suggests that the surface properties of ER-stressed cells are quite different from WT cells.
To begin to characterize the cell walls of ER-stressed mutants, we extracted whole cells with 5 mM DTT, a procedure that liberates noncovalently and dithiol-linked cell wall proteins and that yields increased amounts of extractable proteins when the cell wall is defective (Hagen et al., 2004 ). SDS-PAGE fractionation of DTT extracts from equivalent numbers of cells (20 OD600 equivalents/lane) showed that increased amounts of protein are released by DTT in ER-stressed mutants as detected by silver staining (Figure 2C). In addition, the spent media from cultures of ire1Δ, GAL-CPY*, and ire1ΔGAL-CPY* contains more protein than that from WT cultures (Figure 2D). Increased release of protein into the culture medium is also a phenotype of strains with mutations that affect cell wall proteins (Richard et al., 2002 ). Thus, these two experiments indicate that uncompensated ER stress causes changes to the protein component of the cell wall that are similar to those observed when cell wall proteins are mutated, supporting the idea that ER quality control mechanisms may be required for the proper function of cell wall proteins.
To further explore the effects of uncompensated ER stress on the protein component of the cell wall, we examined the fraction of proteins covalently attached to the cell wall through an alkali-sensitive ester linkage with 1,3-β-glucan (Kapteyn et al., 1999 ) in ER-stressed strains. In contrast to the noncovalently bound cell wall protein fraction, ER stress seems to decrease the amount of protein that is released from cell walls by alkali extraction (Figure 2E). An important caveat to the cell wall extraction and protein release experiments is that no internal protein loading control is available; thus, the samples were normalized by cell number or cell wall weight. Consequently, we cannot rule out the possibility that differences in the ratio of cell wall material to cell number may contribute to some of the observed changes in protein levels. With these limitations in mind, our results are consistent with similar experiments performed with cell wall mutants (Kapteyn et al., 1999 ; Hager et al., 2004). Therefore, the data support the notion that ER stress has a variety of affects on cell wall protein homeostasis. Furthermore, the fact that multiple fractions of cell wall protein are altered by ER stress suggests that these cell wall defects are unlikely to be due to misprocessing of a single class of cell wall proteins or components.
As a final means to characterize the cell wall of ER-stressed mutant strains, we examined the cells by electron microscopy (Figure 3). WT, ire1Δ, GAL-CPY*, and ire1ΔGAL-CPY* strains all displayed the characteristic bilayer cell wall with a translucent, carbohydrate region sandwiched between two electron-dense mannoprotein layers (Osumi, 1998 ). Although no gross, general morphological defects were observed in the mutant strains, two more subtle differences between WT and ER-stressed strains are apparent. First, the outer mannoprotein layer of ire1Δ and ire1ΔGAL-CPY*, in particular, seems less electron dense. Additionally, fewer electron dense mannoprotein fibrils emanate from the cell surface of these strains (Figure 3).
Second, the distance between the electron-dense layers of the cell wall of the ire1Δ and ire1ΔGAL-CPY* cells is decreased in comparison with either WT or GAL-CPY*. The distance between electron-dense layers in WT cells was similar to that reported for the same W303-WT strain (105 nm [SD 20], Wright et al., 2007 ), whereas the translucent layer in the ire1Δ and ire1ΔGAL-CPY* strains was thinner (75 nm [SD 19] and 46 nm [SD 5], respectively). Because the translucent layer of the cell wall is mainly carbohydrate, we measured the total carbohydrate content of cell walls isolated from the mutants and, consistent with the morphological observations, there is less carbohydrate in the cell wall of ire1Δ (25% reduced) and ire1ΔGAL-CPY* (48% reduced) relative to WT and GAL-CPY* (Supplemental Figure S2C). Thinning of the cell wall has also been observed in Candida albicans mutants lacking glycosylphosphatidylinositol (GPI)-linked cell wall proteins, indicating that our findings are likely to represent significant alterations to cell wall structure (Plaine et al., 2008 ). Together, our data support the notion that uncompensated ER stress causes a variety of changes to the cell wall and, hence, seems to affect several processes involved in cell wall biosynthesis.
The genes involved in ERAD are up-regulated by UPR and are also required for maintaining ER quality control during unstressed, vegetative growth. We, therefore, wondered whether mutations of ERAD components would cause cell wall defects. Because ERAD and ire1Δ mutants display synthetic phenotypes (Friedlander et al., 2000 ; Travers et al., 2000 ), we also constructed ERAD/UPR double mutants to determine whether loss of both quality control mechanisms would cause cell wall defects. ERAD has three subtypes that are distinguished by whether the lesion is in the luminal (ERAD-L), membrane (ERAD-M), or cytosolic (ERAD-C) portion of the misfolded protein (Carvalho et al., 2006 ; Ismail and Ng, 2006). To test the effects of these different subtypes, we constructed three ERAD single and ERAD/ire1Δ double mutants in the S288c background: hrd1Δ/ire1Δ hrd1Δ (ERAD-L and -M); doa10Δ/ire1Δdoa10Δ (ERAD-C), and ubc7Δ/ire1Δubc7Δ (common to ERAD-L, -M, and -C).
As shown in Figure 4, A and B, none of the single ERAD mutants was hypersensitive to Congo red or zymolyase digestion. However, ire1Δhrd1Δ and ire1Δubc7Δ were hypersensitive to all three conditions, with ire1Δhrd1Δ being affected most severely. hrd1Δ ire1Δ and ubc7Δ ire1Δ also showed increased DTT-extractable cell wall protein (Figure 4C) in a pattern similar to either CPY* expression or ire1Δ mutation in the W303 background (see Figure 2C). ire1Δdoa10Δ, in contrast, showed no increased sensitivity to cell wall perturbation and no changes in DTT-extractable protein.
We also examined the cell wall ultrastructure of the ire1Δhrd1Δ mutant (Supplemental Figure S3). Although the cell wall thickness was not decreased appreciably, the outer cell wall layer of ire1Δhrd1Δ cells is less electron dense than WT, and, similar to the mutants shown in Figure 3, the hair-like mannoprotein fibrils decorating the outer surface of the cells are less prominent in the ire1Δhrd1Δ cell wall. Together, the cell wall defects shown by double mutants involving specific ERAD components and ire1Δ provide additional evidence that ER quality control mechanisms are required to preserve cell wall integrity. Ire1p, however, seems to be more important for cell wall integrity than any of the single ERAD components examined.
Because our data clearly demonstrate UPR is involved in the maintenance of cell wall integrity, it seemed plausible that cell wall stress may induce UPR. Although ER stress is known to activate the CWI pathway (Bonilla and Cunningham, 2003 ; Chen et al., 2005 ), to our knowledge, cell wall stress has not been shown to trigger UPR. To test this hypothesis, we used a sensitive reporter of UPR-mediated transcription that contains four copies of the unfolded protein response element (UPRE) fused to the E. coli β-galactosidase gene (lacZ) on a centromeric plasmid (Cox and Walter, 1996 ). Treatment of WT cells harboring the 4X-UPRE reporter with cell wall perturbing agents CFW and caffeine induced reporter activity six- and ninefold, respectively (Figure 5A). Although this level of UPR activation is three- to fourfold below that induced by DTT (a compound that generates misfolded proteins; Back et al., 2005 ), UPR is clearly induced by exposure of cells to cell wall stress. The UPR reporter activity induced by both CFW and caffeine is abolished by null mutation of IRE1 (Figure 5A), indicating that cell wall stress activates the reporter through UPR and not via an alternate, Ire1p-independent pathway.
To confirm these observations, we next examined the effect of other cell wall stressors on UPR activity. Incubation at 37°C is a mild cell wall stress that activates CWI signaling, and, accordingly, causes a modest induction of UPR. gas1Δ is a mutant with significant cell defects that result in constitutive activation CWI signaling pathway as a compensatory response (de Nobel et al., 2000 ; Popolo and Vai, 1999 ); the fact that gas1Δ also shows increased basal UPR, strongly supports the notion UPR is induced by cell wall stress.
The cellular effects of cell wall stress are frequently suppressed by the addition of an osmotic support such as 1 M sorbitol to the growth medium of cell–wall-stressed cells (Levin, 2005 ). Indeed, CFW treatment of cells in YPD + 1 M sorbitol induced less UPR activity compared with standard YPD, providing further support for the notion that cell wall damage, and not some other cellular CFW effect, triggers UPR (Figure 5B).
To confirm that cell wall stress activates UPR, we next examined the effect of CFW on HAC1u splicing by using a sensitive RT-PCR assay (Bicknell et al., 2007 ). As shown in Figure 5C, CFW treatment results in an increase in spliced HAC1i, the key molecular event mediated by Ire1p. CFW treatment also leads to the translation of hemagglutinin-tagged Hac1p as demonstrated by the Western blot shown in Figure 5D. Together, these data demonstrate for the first time that cell wall stress induces UPR.
The CWI MAP kinase pathway is the most important regulator of the cell wall stress response in Saccharomyces cerevisiae; therefore, it seemed a likely candidate to mediate cell wall stress-induced UPR (Levin, 2005 ). Consistent with this hypothesis, treatment of a strain lacking MPK1, the MAP kinase of the CWI pathway, with sublethal CFW (20 μg/ml) did not increase UPR activity (Figure 6A). In addition, the basal level of UPR activity during unstressed vegetative growth was decreased twofold. To confirm that Mpk1p contributes to the regulation of basal UPR activity, we compared the extent of HAC1u splicing in unstressed, logarithmic phase WT and mpk1Δ cells by RT-PCR (Figure 6B). HAC1 splicing under these conditions was reduced by one-half in mpk1Δ cells compared with WT. Bicknell et al. (2007) recently showed that vegetative cells express low levels of basal UPR activity and that it is required for efficient cytokinesis. The CWI pathway is activated during new bud formation; hence, a low level of Mpk1p activity is present during vegetative growth (Zarzov et al., 1996 ). Thus, the partial dependence of basal UPR on Mpk1p provides further support for the notion that the CWI pathway regulates UPR.
The fact that basal and cell wall stress-activated UPR requires both Mpk1p and Ire1p suggested that these two proteins may function in a common regulatory pathway. Consistent with this model, the ire1Δmpk1Δ double mutant is as sensitive to CFW as the mpk1Δ single mutant, indicating that the ire1Δ and mpk1Δ mutations are epistatic (Figure 6C). Furthermore, these results also indicate that a linear pathway may connect CWI signaling to UPR during cell wall stress and vegetative growth.
To determine which components of the CWI pathway are involved in cell wall stress-induced UPR, we measured CFW-induced 4XUPRE-lacZ reporter activity in a set of CWI-pathway mutants (Table 2). Initially, we tested strains lacking the putative cell wall stress receptors Mid2p and Wsc1p (Figure 7A). One difference between Mid2p and Wsc1p is that Mid2p mediates CFW-induced CWI pathway activity, whereas Wsc1p does not (de Nobel et al., 2000 ). As shown in Table 2, CFW-induced UPR activation was decreased in the mid2Δ mutant, but WT activity was retained in the wsc1Δ mutant. RT-PCR assays confirmed that CFW-induced splicing was decreased in the mid2Δ mutant but not in wsc1Δ (Figure 7B). That Mid2p mediates both CFW-induced cell wall gene expression and UPR supports the idea that these two transcriptional responses represent separate outputs of CWI signaling.
Three transcription factors/regulators are controlled in part by the CWI pathway: Rlm1p, Swi4p, and Swi6p (Levin, 2005 ). Basal UPR activity was decreased twofold in the absence of Rlm1p (Table 2). In contrast, CFW treatment of rlm1Δ induced the same relative increase in UPR as observed with WT (~5-fold), indicating that Rlm1p may play a role in CWI-dependent basal UPR, but that it does not seem to mediate cell wall–stress-induced UPR. Swi4p and Swi6p form a dimeric transcription factor called SBF that plays a role in CWI and cell cycle signaling (Levin, 2005 ; Madden et al., 1997 ). The swi4Δ mutant displayed basal and CFW-induced UPR activity similar to WT, whereas UPR activation was significantly lower in the swi6Δ mutant under both conditions (Table 2 and Figure 7C). In addition, introduction of a multicopy plasmid expressing SWI6 from its endogenous promoter into a swi6Δ strain complemented the HAC1 splicing defect and, in fact, increased basal and CFW-induced UPR by RT-PCR (Figure 7C). These results strongly suggest that Swi6p is required for basal and cell wall stress-induced UPR in a manner that is independent of the Swi4p and hence SBF.
In addition to Swi4p, Swi6p has two other well-characterized binding partners, Mbp1p and Stb1p. Mb1p interacts with Swi6p to form a second dimeric transcription factor known as the MBF complex (Koch et al., 1993 ), whereas Stb1p regulates the timing of transcription at Start through its interaction with Swi6p (Ho et al., 1999 ). As shown in Table 2, CFW treatment of either mbp1Δ or stb1Δ activated the UPR reporter to the same extent as WT strains, indicating that the role of Swi6p in cell wall stress-induced UPR is independent of its three best characterized binding partners (Swi4p, Mbp1p, and Stb1p).
We have demonstrated that uncompensated ER stress leads to cell wall defects in yeast. Protein glycosylation, GPI-anchor biosynthesis, and 1,6-β-glucan synthesis (Lesage and Bussey, 2006 ) are examples of the wide range of essential cell wall-related processes localized to the secretory pathway. Therefore, it seems most likely that the cell wall consequences of ER stress are the result of defects in multiple steps in cell wall biosynthesis as opposed to a single defect in a specific step. Among the many possible causes of ER stress-related cell wall defects, we suggest that at least some of the defects may be due to the delivery of misfolded or misprocessed mannoproteins to the cell wall.
At a very simple level, delivery of normally degraded misfolded proteins to the cell wall would explain the increased amount of cell wall protein observed in ER-stressed mutants (Figure 2A). Furthermore, misfolded or misprocessed cell wall proteins may not be suitable substrates for enzymes that incorporate cell wall proteins into the carbohydrate lattice of the cell wall. Similarly, misfolded cell wall proteins may not be able to carry out key structural functions. Consistent with this model, mutants under uncompensated ER stress display cell wall phenotypes that mimic mutants lacking cell wall proteins (e.g., increased DTT extractable proteins and excessive protein release into culture medium (Figure 2, B and C). Obviously, more detailed analyses of the cell wall of these mutants is required to characterize the specific mechanisms by which compromised ER quality control causes cell wall defects.
Although UPR seems to be the most important protein quality control mechanism for cell wall integrity, loss of specific ERAD components affects cell wall integrity when combined with loss of UPR function. Specifically, mutation of ERAD components involved in luminal or membrane proteins quality control (ERAD-L, M) cause cell wall defects, whereas loss of cytosolic ERAD (ERAD-C) does not affect the cell wall. This indicates that misfolded protein stress within the secretory pathway is most damaging to cell wall integrity.
Because ERAD/UPR double mutants (ire1Δhrd1Δ and ire1Δubc7Δ) also display synthetic noncell wall phenotypes (Friedlander et al., 2000 ; Travers et al., 2000 ), an alternate explanation for the synthetic cell wall phenotypes is that they represent a general fitness defect. However, other groups have shown that hrd1Δ, ubc7Δ, and doa10Δ display similar levels of basal UPR activation (Friedlander et al., 2000 ; Bicknell et al., 2007 ), indicating that the mutations cause comparable levels of ER stress. Because these ERAD mutants activate UPR similarly yet have quite different effects on the cell wall, we propose that the cell wall phenotypes of ERAD/UPR double mutants are the result of effects on cell wall biosynthesis rather than due to a decreased global fitness. If this model is correct, then the cellular consequences of misfolded protein stress are determined, in part, by the cellular location (ER vs. cytoplasm) subjected to stress.
Consistent with the dependence of cell wall integrity on ER quality control, we have shown that cell wall stress activates UPR, the pathway that regulates the main compensatory response to secretory stress. Although the effect of cell wall stress on UPR has not been reported previously, Pal et al. (2007) have recently shown that CWI pathway activation by hypotonic stress increases the UPR activity of cells that are concomitantly exposed to ER stress (Pal et al., 2007 ). The synergistic increase in UPR during the simultaneous application of ER and hypotonic stress is the result of an Mpk1p-dependent lengthening of the lifetime of Hac1p. Hypotonic stress alone, however, is not sufficient to induce UPR, probably because hypotonic stress causes only a transient CWI pathway activation (Davenport et al., 1995 ). Our results indicate that more severe cell wall stress activates UPR through increased Ire1p-mediated HAC1u splicing; thus, cell wall stress seems to modulate UPR activity by directly inducing HAC1u splicing and by increasing the lifetime of Hac1p.
The CWI pathway is also activated in vegetative cells during periods of polarized growth in a process that is partially dependent on Cdc28p. Therefore, it would follow that UPR might also be activated during polarized growth if it was regulated by the CWI pathway. Although the function of UPR has traditionally thought to be limited to periods of cellular stress, Bicknell et al. have recently shown that UPR is active at low, basal levels during unstressed vegetative growth (Bicknell et al., 2007 ). Consequently, our observation that constitutive UPR activity is partially dependent on Mpk1p suggests a potential mechanism by which basal UPR is regulated and implies that this regulation maybe related to the cell cycle in yeast.
Cell wall stress signals activate UPR through components of the CWI MAPK pathway, including Mid2p, a plasma membrane-localized CWI pathway stress receptor; Mpk1p, the MAPK for the CWI pathway; and Swi6p, a transcriptional regulator modulated by the CWI pathway. Interestingly, Swi6p has also been shown to interact genetically under cell wall stress with the cytoplasmic Hsp110 chaperone Sse1p, suggesting that it could play a broader role in the cellular response to misfolded protein stress (Shaner et al., 2008 ). However, it is not clear whether the cell wall defects associated with loss of cytoplasmic chaperone function are to due to defective signaling through the CWI pathway (Shaner et al., 2008 ) or are, alternatively, the result of more direct effects on cell wall structure (Wright et al., 2007 ).
At present, it is also unclear how the signal to activate UPR is transmitted from Swi6p to Ire1p. Swi6p-mediated UPR activation is not dependent on any of its previously characterized binding partners (Swi4p, Mbp1p, or Stb1p). Because Swi6p does not seem to possess DNA binding domains (Sedgwick et al., 1998 ), a transcriptionally mediated mechanism for Swi6p-induced UPR activation would seem to require a previously unidentified Swi6p-dependent DNA binding protein. Alternatively, Swi6p could active UPR through a nontranscriptionally mediated mechanism involving interactions with other proteins or, even, through direct binding to Ire1p.
In this report, we have demonstrated that secretory stress has significant effects on cell wall integrity, particularly in the absence of an appropriate compensatory response. These observations provide a mechanistic rationale for previous reports indicating that secretory stress activates both UPR and CWI stress pathways. Recently, Cohen et al. (2008) have shown that the Hos2p/Set3p deacetylase complex mediates transmission of secretory stress signals to the CWI pathway in a process that is parallel to the initiation of UPR. This parallel relationship is consistent with the fact that mutants deficient in UPR and CWI signaling (e.g., ire1Δmpk1Δ) show synthetic hypersensitivity to ER stress.
We have shown that cell wall stress also initiates both a secretory pathway stress response (UPR) and a cell wall stress response. In contrast to the parallel activation of these responses by secretory stress, our data indicate that cell wall stress-mediated activation CWI signaling and UPR occurs via linear pathway connecting the CWI pathway to Ire1p-mediated signaling. As such, ire1Δmpk1Δ mutants show epistatic cell wall defects relative to the corresponding single mutants.
Combining our results with previous studies on the role of the CWI pathway in secretory stress responses, we propose the following model to describe the interrelated nature of the cellular response to secretory pathway stress and cell wall stress. During secretory pathway stress, the resulting ER dysfunction could lead to cell wall damage; consequently, the cell wall compensatory response is triggered to reinforce cell wall integrity. Conversely, during cell wall stress, any breakdown in ER quality control could further exacerbate cell wall damage and, thus, the folding capacity of the secretory pathway is increased through UPR activation, lessening the chances of further cell wall damage as a result of misfolded proteins. Studies on the molecular mechanisms responsible for this complex interplay between these two major cellular signaling pathways are in progress and will be reported in due course.
We thank Davis Ng (National University of Singapore), Peter Walter (University of California, San Francisco), Kyle Cunningham (Johns Hopkins University), and Linda Breeden (University of Washington) for providing strains and plasmids. We are grateful to Alison Gammie and Mark Rose (Princeton University) for providing protocols for electron microscopy. We thank Davis Ng, Frans Klis (University of Amsterdam) and Linda Breeden for helpful discussions. Davis Ng, Melanie Wellington (University of Rochester), Anuj Kumar (University of Michigan), and David Dean (University of Rochester) are thanked for critical readings of the manuscript. This work was supported by National Institutes of Health grant K08AI062978 (to D.J.K.).
This article was published online ahead of print in MBC in Press (http://www.molbiolcell.org/cgi/doi/10.1091/mbc.E08-08-0809) on October 29, 2008.