Cdc48 and Cofactors Npl4-Ufd1 Are Important for G1 Progression during Heat Stress by Maintaining Cell Wall Integrity in Saccharomyces cerevisiae

The ubiquitin-selective chaperone Cdc48, a member of the AAA (ATPase Associated with various cellular Activities) ATPase superfamily, is involved in many processes, including endoplasmic reticulum-associated degradation (ERAD), ubiquitin- and proteasome-mediated protein degradation, and mitosis. Although Cdc48 was originally isolated as a cell cycle mutant in the budding yeast Saccharomyces cerevisiae, its cell cycle functions have not been well appreciated. We found that temperature-sensitive cdc48-3 mutant is largely arrested at mitosis at 37°C, whereas the mutant is also delayed in G1 progression at 38.5°C. Reporter assays show that the promoter activity of G1 cyclin CLN1, but not CLN2, is reduced in cdc48-3 at 38.5°C. The cofactor npl4-1 and ufd1-2 mutants also exhibit G1 delay and reduced CLN1 promoter activity at 38.5°C, suggesting that Npl4-Ufd1 complex mediates the function of Cdc48 at G1. The G1 delay of cdc48-3 at 38.5°C is a consequence of cell wall defect that over-activates Mpk1, a MAPK family member important for cell wall integrity in response to stress conditions including heat shock. cdc48-3 is hypersensitive to cell wall perturbing agents and is synthetic-sick with mutations in the cell wall integrity signaling pathway. Our results suggest that the cell wall defect in cdc48-3 is exacerbated by heat shock, which sustains Mpk1 activity to block G1 progression. Thus, Cdc48-Npl4-Ufd1 is important for the maintenance of cell wall integrity in order for normal cell growth and division.

The diverse functions of Cdc48/p97 are mediated by specific cofactors. The binary complex Npl4-Ufd1 is associated with ER membrane and required for degradation of ER proteins [5]. Npl4 contains NZF domain that binds polyubiquitin chain [6]. The Nterminal domain of Ufd1 also has a higher affinity toward polyubiquitin than monoubiquitin [7]. Cdc48 coupled with Npl4-Ufd1 functions in retrograde translocation of proteins from ER for degradation (ERAD) [8]. Cdc48/p97 also binds a family of proteins containing a ubiquitin-related (UBX) domain that is structurally similar to ubiquitin [9]. Ubx1, also known as Shp1 (Suppressor of high copy protein phosphatase 1) [10], Ubx2, Ubx4, Ubx6, and Ubx7 serve as cofactors for Cdc48 in ubiquitindependent protein degradation [11]. Cdc48-Shp1 is also important for chromosome bi-orientation [12]. On the other hand, the mammalian homolog of Shp1, p47, is involved in membrane fusion [13].
Budding yeast Cdc48 was originally isolated as a cell cycle mutant that arrested in mitosis at the restrictive temperature [4]. Cdc48/p97 appears to have multiple functions in the cell cycle. In budding yeast, Cdc48 is required for passing Start, the cell cycle commitment point in G1, by degrading the G1-cyclin-dependent kinase inhibitor Far1 [14]. In fission yeast, Cdc48 is required for the metaphase-to-anaphase transition by stabilizing Separase [15], the enzyme that cleaves cohesin components to separate sister chromatids. We have previously demonstrated that budding yeast Cdc48 and its cofactor Shp1 promote chromosome bi-orientation by balancing Aurora B activity [12]. In addition, Cdc48/p97 together with Npl4-Ufd1 has been shown to participate in spindle disassembly during mitotic exit [16], although the result is controversial [17]. p97 is also important for the formation of a closed nuclear envelope and nuclear expansion following nuclear envelope formation [18]. Cdc48/p97 itself is regulated in the cell cycle. The protein is primarily associated with membranes of organelles such as the ER and the Golgi [1]. In G1 phase, a fraction of Cdc48 enters the nucleus in a phosphorylationdependent manner [19]. The change of Cdc48 localization during the cell cycle likely reflects its multiple functions.
Cell cycle progression is mainly governed by cyclin-dependent kinases (CDK). Coupled with G1 or mitotic cyclins, the CDK activity drives G1/S transition or mitotic entry, respectively. Budding yeast has three G1 cyclins encoded by CLN1, CLN2, and CLN3 [20]. These G1 cyclins share redundant functions, as cells can live on just one of the cyclins [21]. The expression of these genes is induced as cells traverse G1. The mRNA and protein of CLN3 constantly exist during the cell cycle and are modestly induced at late G1 [22]. On the other hand, Cln1 and Cln2 are present at low levels at early G1 and transiently induced before Start [20,22]. The induction of CLN1 and CLN2 is mediated through the SCB and MCB sequences in their promoters that bind transcription factors Swi4/Swi6 (SBF) and Swi4/Mbp1 (MBF) complexes, respectively [23]. Activation of both SBF and MBF is dependent on the kinase activity of Cln3-Cdc28 [23]. Cln3-Cdc28 also inactivates Whi5, the transcription suppressor that inactivates SBF specifically by direct binding until G1 [24,25]. Thus, Cln proteins promote their own expression through a positive feedback loop [26,27,28].
The Start commitment point in G1 phase of budding yeast is controlled by nutrient availability, cell size, and the presence of mating pheromone [29]. In addition, heat shock transiently inhibits Start [30]. Heat shock generates misfolded or aggregated proteins that trigger heat shock response pathway featuring the induced synthesis of a set of evolutionarily conserved heat shock proteins [31]. Many of the heat shock proteins are chaperones that help protein folding [31]. The induction of heat shock proteins are primarily mediated by heat shock factor (HSF) and transcription factors Msn2 and Msn4 that bind to heat shock elements (HSE) and stress response element (STRE), respectively, in the promoter of many heat-inducible genes [32]. In addition, heat shock activates the MAP kinase homolog Mpk1 that maintains cell wall integrity and prevents cell lysis when cells are grown at elevated temperature [33]. Mpk1 is downstream of Pkc1 that regulates a protein kinase cascade in which Bck1 (MEK kinase) activates Mkk1 and Mkk2 (MAP kinase-kinase) that in turn activate Mpk1 [34]. This cell wall integrity pathway senses the cell surface defects through cell surface proteins Wsc1 [35] and Mid2 [36]. Mpk1 phosphorylates and activates transcription factor Rlm1 that regulates the expression of many genes involved in the cell wall biogenesis [37]. Mpk1 is also important for the heat shock-induced transcription through the HSE and STRE elements [33]. In addition, the SBF transcription factor is a target of Mpk1 [38], providing a link between heat stress and cell cycle control.
In this study, we examine the cell cycle function of Cdc48 and show that Cdc48 and the cofactor Npl4-Ufd1 complex are important for maintaining the cell wall integrity during heat stress to allow G1 progression.

G1 delay of cdc48-3 at high temperature
In order to understand the cell cycle function of Cdc48, we have examined the phenotypes of the temperature-sensitive cdc48-3 mutant. The mutant was largely arrested at mitosis with 2N DNA content at 37uC as determined by fluorescenceactivated cell sorter (FACS) analysis (Fig. 1A). Interestingly, a small fraction of cells contained 1N DNA at 38.5uC, indicating that G1 progression was perturbed. To analyze the cell cycle progression, we first arrested the cells at G1 with a-factor and then released the arrest at 38.5uC. FACS analysis showed that most of the wild-type cells completed DNA replication by 100 min after the release (Fig. 1B). However, less than 50% cdc48-3 cells completed DNA replication by 100 min and some cells still contained 1N DNA even at 140 min (Fig. 1B). Examination of the cell morphology showed that more than 50% of wild-type cells have budded by 80 min after release from a-factor arrest and more than 80% of the cells have become large-budded after 140 min (Fig. 1C). However, only ,50% of cdc48-3 cells have budded with only 20% large-budded cells at 140 min (Fig. 1C). These results show that cdc48-3 mutant was delayed at G1/S transition at 38.5uC.
Reduced expression of G1 cyclin G1 progression is controlled by the accumulation of G1 cyclins encoded by CLN1, CLN2, and CLN3. Because CLN1 and CLN2 are transiently induced before G1/S transition, we measured the promoter activities of CLN1 and CLN2 during a synchronous cell cycle by reporter assays. Figure 2 shows that the luciferase activity driven by the CLN1 promoter increased about 20 folds at 120 min after release from G1 arrest in the wild-type cells at 38.5uC, whereas the activity increased only 5 folds in cdc48-3. CLN1 promoter activity was slightly lower in cdc48-3 than that in the wild-type cells at 37uC during 100 min after release from G1 arrest (Fig. 2). The rise of the activity after 100 min in the wildtype cells likely reflected the second cell cycle. cdc48-3 cells did Figure 2. CLN1 promoter activity is reduced in cdc48-3 at high temperature. Wild-type and cdc48-3 cells carry Renilla reniformis and Pyrophorus plagiophthalamus luciferases under the control of CLN1 and CLN2 promoters, respectively. Cells were first arrested at G1 with a-factor, and then released into the cell cycle at 38.5uC or 37uC. Luciferase activities were measured in triplicates at the indicated times after the release. The activities were normalized to that at time 0. The plot shows the average activities in fold increase and the standard deviation. doi:10.1371/journal.pone.0018988.g002 not show further increase of CLN1 promoter activity after 100 min, because the mutant was arrested at mitosis at 37uC [12]. Unlike CLN1 promoter, CLN2 promoter-driven luciferase activities were comparable in the wild-type and cdc48-3 cells at both 38.5uC and 37uC (Fig. 2). This result suggests that CLN1, but not CLN2, promoter activity was affected in cdc48-3 at 38.5uC.
The reduced CLN1 promoter activity in cdc48-3 at 38.5uC suggests that the G1 delay may result from reduced levels of G1 cyclins. To test this possibility, we expressed CLN1 or CLN2 through the MET3 promoter in cdc48-3. The cells were released from G1 arrest in methionine-free medium to induce CLN1 or CLN2 expression. Without additional CLN1 or CLN2, cdc48-3 cells in methionine-free medium traversed G1 slowly at 38.5uC, with only 20% of the cells budded at 2.5 hr after release from the G1 arrest (Fig. 3A). Upon expression of CLN1 through MET3 promoter, more than 60% of cdc48-3 cells budded at 1.5 hr after G1 release (Fig. 3A). The expression of CLN2 in cdc48-3 also expedited G1 progression, with ,35% of the cells budded at 1.5 hr after G1 release (Fig. 3A). Western blots showed that the ectopically expressed Cln1 and Cln2 proteins can be detected by 30 min after induction (Fig. 3B). These results show that overexpression of either Cln1 or Cln2 protein can partially rescue the G1 delay of cdc48-3 at high temperature and that Cln1 is more effective in driving G1 progression than Cln2 is under this condition.
Prolonged activation of Mpk1 in cdc48-3 Heat shock is known to transiently arrest yeast cells in G1, raising the possibility that the G1 delay of cdc48-3 at 38.5uC may be a consequence of heat stress. We thus examined Mpk1, a MAPK family member and a component of the cell wall integrity pathway that is activated by phosphorylation in response to perturbation of the cell wall from various stress conditions including heat shock. We monitored phosphorylated Mpk1 with a phospho-MAPK antibody that recognizes several phosphorylated MAPK members. In wild-type cells arrested at G1 with a-factor, Mpk1 phosphorylation increased when the growth temperature was shifted from 25uC to 38.5uC (Fig. 4, top panel, compare lanes 1 and 2). After release from the arrest at 38.5uC, the phosphorylation remained for 20 min and gradually declined afterwards (Fig. 4, top panel). Phosphorylation of Mpk1 in cdc48-3 was enhanced both at 25uC and upon temperature up-shift in comparison with that in the wild-type cells, and it sustained for at least 2 hr after release from a-factor arrest ( Fig. 4, top panel). The level of Mpk1 phosphorylation declined faster at 37uC than that at 38.5uC (Fig. 4, middle panel). At 37uC, Mpk1 in cdc48-3 was still phosphorylated to a higher degree than that in the wild-type cells (Fig. 4, middle panel). At 25uC, Mpk1 phosphorylation disappeared shortly after release from a-factor arrest in both wild-type and cdc48-3 cells (Fig. 4, bottom panel). The anti-phospho-MAPK antibody also recognized phosphorylated Fus3, a mating-specific MAPK member that arrests the cell cycle through transcriptional repression of CLN1 and CLN2 and through posttranscriptional inhibition of Cln3 [39]. Fus3 phosphorylation was present in cells arrested with a-factor and quickly disappeared upon release from the arrest in both wildtype and cdc48-3 cells at all temperatures tested (Fig. 4). The enhanced phosphorylation of Mpk1 in cdc48-3 at high temperature suggests that the heat stress may be exacerbated in cdc48-3 mutant.
Because Mpk1 is a component of the cell wall integrity pathway, the enhanced phosphorylation of Mpk1 in cdc48-3 indicates a defect in the cell wall at high temperature. We tested this possibility by including 1 M sorbitol in the medium to increase the osmolarity which is known to protect the cell wall and prevent cell lysis in mutants defective in the cell wall integrity pathway. Without sorbitol addition less than 40% of cdc48-3 cells were budded at 120 min after release from G1 arrest, whereas more than 90% of the cells were budded in the presence of sorbitol (Fig. 5A). Sorbitol addition also accelerated DNA replication in cdc48-3 at 38.5uC, with a small lag compared to the wild-type cells (Fig. 5A). In addition, reporter assays showed that sorbitol treatment increased the CLN1 promoter activity in cdc48-3 at 38.5uC, although the activity was still slightly lower than that in the wild-type cells (Fig. 5B). On the other hand, CLN2 promoter activity at 38.5uC was not affected by sorbitol addition (Fig. 5B). These results show that high osmolarity can rescue the G1 delay of cdc48-3, which implies that cdc48-3 was defective in maintaining the cell wall integrity during heat shock. Npl4 and Ufd1 are involved in G1 progression Because Cdc48 executes its diverse functions through specific cofactors, we searched for the cofactors of Cdc48 involved in G1 progression. The known Cdc48 cofactors include Npl4-Ufd1 complex and a family of UBX domain-containing proteins. The deletion mutants of the UBX family proteins did not display specific G1 delay at high temperature (data not shown), whereas the temperature-sensitive npl4-1 and ufd1-2 mutants were much delayed in both budding and DNA replication upon release from a-factor arrest at 38.5uC (Fig. 6A). Similar to cdc48-3, the promoter activity of CLN1, but not CLN2, was reduced in npl4-1 and ufd1-2 at 38.5uC, but not at 37uC (Fig. 6B). These results indicate that Npl4-Ufd1 complex mediates the function of Cdc48 in G1.
Cells grown at high temperature may accumulate denatured proteins that need to be folded by chaperones or be degraded by the ubiquitin-proteasome system. Because Cdc48 and Npl4-Ufd1 complex are important for ERAD, the G1 delay of cdc48-3, npl4-1, and ufd1-2 cells at 38.5uC may be related to their ERAD function. We thus examined the deletion of two other components of the ERAD system, the ubiquitin-conjugation enzyme UBC7 and the ubiquitin-protein ligase HRD1 [40]. FACS analysis showed that ubc7D and hrd1D mutants had only a small delay in DNA replication during a synchronized cell cycle at 38.5uC (Fig. 7A). This result suggests that defects in ERAD do not impact on G1 progression during heat shock and that the G1 delay of cdc48-3, npl4-1, and ufd1-2 mutants at 38.5uC is probably independent of their functions in ERAD. To determine if the G1 delay involves other aspects of protein degradation, we examined the deletion of DOA1, a ubiquitin-binding protein that bridges Cdc48 to its substrates for protein degradation [41]. FACS analysis showed that doa1D behaved essentially the same as the wild-type cells at 38.5uC (Fig. 7A). In addition, cdc48-3, npl4-1, and ufd1-2 cells at 38.5uC contained more ubiquitincojugated proteins than did wild-type cells, whereas ubc7D, hrd1D, and doa1D cells did not (Fig. 7B). We also analyzed additional ERAD components, including ubiquitin ligase Doa10 [40], the cytosolic Hsp70 chaperone Ssa1 involved in ERAD of a membrane protein [42], and Ubx2 that recruits Cdc48 to ERAD ubiquitin ligase [40]. Because ubx2D deletion mutant grew very slowly in our strain background, we conditionally controlled the expression of UBX2 with a galactose-inducible promoter that can be suppressed with glucose. Similar to ubc7D and hrd1D cells, doa10D, ssa1D, and Ubx2-depleted cells did not accumulate ubiquitin conjugates (Fig. S1). Furthermore, additional mutations in the ERAD system did not abolish the ubiquitin conjugates in cdc48-3 (Fig.  S1), indicating that the proteins were ubiquitylated by an ERAD-independent mechanism.
The accumulation of ubiquitylated proteins may lead to Mpk1 activation and G1 arrest in cdc48-3. To test this possibility, we examined the effect of adding 1 M sorbitol to the medium during temperature shift to 38.5uC, which suppressed G1 delay and enhanced CLN1 promoter in cdc48-3 (Fig. 5). Figure 7C shows that addition of sorbitol greatly reduced the level of phosphorylated Mpk1 in both wild-type and cdc48-3 cells without an obvious effect on the level of ubiquitin conjugates in cdc48-3 cells. Therefore, accumulation of ubiquitylated proteins per se does not cause Mpk1 activation and G1 arrest in cdc48-3.
Our results of sorbitol treatment suggest that a cell wall defect is likely the direct cause of Mpk1 phosphorylation and G1 delay in cdc48-3. Thus, we tested the sensitivity of cdc48-3 to two cell wall perturbing agents, Calcofluor white and Congo red. Figure 8A shows that, unlike wild-type cells, cdc48-3 cells were unable to grow on YPD plates containing either Calcofluor white or Congo red even at 25uC or 30uC. This result suggests that the cell wall is defective in cdc48-3.
To further link cdc48-3 to cell wall defect, we examined genetic interaction between cdc48-3 and components of the cell wall integrity pathway, including Mpk1, Pkc1 (the upstream kinase of Mpk1), Rho1 (a G-protein and a regulator for Pkc1), and Rom2 (GDP/GTP exchange factor for Rho1). The single mutants of these components were growth-defective at elevated temperatures, which can be suppressed by the addition of 1 M sorbitol in the medium up to 35uC (Fig. 8B and data not shown). Their double mutants with cdc48-3 showed more severe growth phenotype than did the single mutants (Fig. 8B). In the presence of sorbitol the double mutants can grow at 30uC but behaved as cdc48-3 alone at 32uC (Fig. 8B). The synthetic phenotype in the double mutants and its suppression by sorbitol suggest that failure to activate the cell wall integrity pathway in cdc48-3 compromises cell viability and that cdc48-3 normally activates this signaling pathway to repair its cell wall defect.

Discussion
Mild heat shock is known to transiently arrest yeast cells at G1 [30]. Herein we report that mutations in Cdc48 and its cofactors Npl4 and Ufd1 prolong the G1 delay in the budding yeast Saccharomyces cerevisiae at 38.5uC. This delay is due to a low CLN1, but not CLN2, promoter activity. Several lines of evidence support that the G1 delay of cdc48-3 at 38.5uC is a consequence of cell wall defect. First, phosphorylation of Mpk1, a MAPK family member important for the cell wall integrity pathway, is increased in cdc48-3 at 38.5uC. Second, the CLN1 promoter activity and the G1 delay in cdc48-3 are rescued by an increase of osmolarity in the medium to protect the cell wall. Furthermore, cdc48-3 is hypersensitive to cell wall perturbing agents and is synthetic-sick with mutations in the cell wall integrity signaling pathway. Our study suggests that Cdc48 is important for cell wall integrity. The cell wall defect in cdc48-3 is probably exacerbated at high temperature (38.5uC) to a degree that over-activates the cell wall integrity pathway and delays G1 progression.
We show that G1 cyclins CLN1 and CLN2 are differentially regulated in cdc48-3 at 38.5uC. The activity of CLN1 promoter is lower in cdc48-3 mutant at 38.5uC than that in the wild-type cells, whereas the CLN2 promoter activities are comparable in both strains. Both CLN1 and CLN2 promoters are activated by Swi4/ Swi6 (SBF) and Mbp1/Swi6 (MBF) complexes that recognize multiple SCB and MCB sequences, respectively, in the upstream regions of CLN1 and CLN2 [43]. Thus, CLN1 and CLN2 are known to be regulated similarly during the cell cycle. Our study provides a rare example that these genes can be regulated differently. Heat stress is known to induce the transcription repressor Xbp1 that shares homology with Swi4 and Mbp1 in the DNA-binding domain [44]. Despite the structural similarity, the DNA recognition sequence of Xbp1 is distinct from Swi4/Swi6 and Mbp1/Swi6 binding sites. The binding motif for Xbp1 is present in CLN1 promoter, and overexpression of Xbp1 can repress the expression of G1 cyclins and lengthen G1. We have found that deletion of XBP1 gene in cdc48-3 promoted budding at 38.5uC, compared to cdc48-3 alone (Fig. S2), suggesting that Xbp1 contributes to the suppression of CLN1 promoter in cdc48-3. How heat stress may control the activity of Xbp1 remains to be determined. Furthermore, the levels of Swi4 and Swi6 proteins are similar in wild-type and cdc48-3 cells at 38.5uC (Fig. S3). However, we cannot exclude the possibility that these transcription activators are modified differently in response to heat stress in cdc48-3 cells and that CLN1 promoter may be more sensitive to the small alteration of these proteins.
A role for Cdc48 in G1 has been suggested previously [14]. By using a temperature-sensitive degron-tagged cdc48-td mutant, it has been shown that Cdc48 is required for the execution of Start commitment point in Saccharomyces cerevisiae by degrading the G1cyclin-dependent kinase inhibitor Far1 [45]. In cdc48-3 mutant Far1 is still degraded with kinetics similar to that in the wild-type cells at 38.5uC (Fig. S3), indicating that the G1 delay of cdc48-3 is not due to a defect in the degradation of Far1. Together, studies with different cdc48 mutant alleles reveal that Cdc48 is important for G1 progression during a normal cell cycle and under heat stress through different mechanisms. Figure 5. High osmolarity rescues G1 defects of cdc48-3. (A) CDC48 and cdc48-3 cells were first arrested at G1 with a-factor. The cells were shifted to 38.5uC during the last 30 min of the arrest in YPD or YPD containing 1 M sorbitol. The cells were then released in the same medium at 38.5uC, and samples were taken for budding index determination and FACS analysis. (B) CDC48 and cdc48-3 cells carrying luciferases under the control of CLN1 and CLN2 promoters were grown as described above and samples were taken at the indicated times for the measurement of luciferase activities. The plot shows the average of three measurements in fold increase and the standard deviation. doi:10.1371/journal.pone.0018988.g005 Figure 6. G1 progression is delayed in npl4-1 and ufd1-2 at high temperature. (A) Wild-type, npl4-1, and ufd1-2 cells were released from G1 arrest at 38.5uC as described in Figure 1B. Budding index and FACS were analyzed after the lease. Filled diamond, no bud; open circle, small bud; filled triangle, medium/large bud. (B) Wild-type, npl4-1, and ufd1-2 cells carrying luciferases under CLN1 and CLN2 promoters were released from G1 arrest at 38.5uC or 37uC as described in Figure 1B. Samples were taken at the indicated times after the release for the measurement of luciferase activities. The plot shows the average of three measurements in fold increase and the standard deviation. doi:10.1371/journal.pone.0018988.g006 Heat shock is known to transiently arrest cells before Start by repressing transcription of CLN1 and CLN2 [46]. Deletion of any G1 cyclin genes has no significant effect on the transient arrest from heat shock at 37uC [46]. In our study we found that G1 progression is further delayed in cdc48-3 at 38.5uC, even though CLN2 expression is similar to that of wild-type cells. It is probable that cells require more G1 cyclin activity to recover from the heat stress incurred at a higher temperature.
Cdc48 is best known for its function in ERAD that is one of the protein quality control systems [47]. Secretory and membrane proteins are normally synthesized and folded in the ER. Misfolded or unassembled proteins are retained in the ER and subsequently degraded via ERAD. Heat shock may cause accumulation of misfolded or aggregated proteins in the ER that need to be removed through the ERAD pathway. Indeed, we found that cdc48-3, npl4-1, and ufd1-2 mutants accumulated higher levels of ubiquitin conjugates than did the wild-type cells at 38.5uC (Fig. 7B). Subcellular fractionation shows that the ubiquitin conjugates were enriched in the ER fractions (Fig. S4). Interestingly, mutations in other ERAD components did not abolish the accumulation of ubiquitylated proteins in cdc48-3 (Fig. S1). Thus, Cdc48-Npl4-Ufd1 is also required for degradation of ER proteins that are ubiquitylated by ERAD-independent pathway. Furthermore, deletion of ERAD components UBC7 and HRD1 has no significant effect on the G1 progression at 38.5uC, suggesting that a defect in ERAD system does not cause G1 arrest in response to heat stress. Therefore, the G1 delay of cdc48-3, npl4-1, and ufd1-2 mutants at 38.5uC is independent of their ERAD function. However, we cannot exclude the possibility that there may be other unknown or redundant ERAD components that function together with Cdc48-Npl4-Ufd1 to remove the ubiquitylated proteins from the ER.
We observe enhanced phosphorylation of Mpk1, a MAPK family member downstream of Pkc1, in cdc48-3 mutant at 38.5uC. This pathway is activated by hypotonic shock or by heat stress [48], and the activation is sustained during growth at a high temperature [33]. Because Pkc1-regulated signaling pathway is known to detect and respond to weakness in the cell wall, the sustained phosphorylation of Mpk1 in cdc48-3 suggests that the defect in the cell wall or the plasma membrane is not repaired. That the addition of sorbitol restores cell growth without affecting the overall ubiquitylation level (Figs. 5 and 7C) indicates that the G1 defect of cdc48-3 is caused by cell wall defect, rather than the accumulation of denatured proteins per se. The sensitivity of cdc48-3 to cell wall perturbing agents at permissive temperature indicates a cell wall defect that is likely exacerbated at high temperature. In addition, Mpk1 phosphorylation is also increased in the cold-sensitive cdc48-1 mutant at 14uC compared to that in the wild-type cells (Fig. S5A), which is consistent with the notion that Cdc48 is required for cells wall integrity. Mpk1 is increasingly phosphorylated with elevated temperatures and the levels are similar between the wild-type and cdc48-1 cells (Fig. S5A), showing that the cold-sensitive cdc48-1 mutant has normal response to heat shock. On the other hand, Mpk1 is phosphorylated to a higher degree in cdc48-3 than in the wild-type cells at both 37u and 38.5uC ( Fig. 4 and S5B). We believe that the heat-induced Mpk1 activation in wild-type cells elicits cell wall repair and transient G1 arrest, whereas overactivation of Mpk1 in cdc48-3 caused by failure to repair the cell wall prolongs G1 arrest.
Heat shock is known to create cell wall stress that activates the cell wall integrity pathway, leading to phosphorylation and activation of transcription factor Rlm1 by Mpk1 [37]. Rlm1 induces expression of many genes implicated in cell wall biogenesis [49]. Mutants in this pathway are deficient in cell wall construction, leading to cell lysis at elevated temperatures. The enzymes for cell wall biogenesis mostly reside in the plasma membrane or the cell wall [50] and are synthesized and modified in ER. It has been shown that proteins synthesized during heat shock are prone to denaturation and are rapidly degraded through Cdc48-Npl4-Ufd1, independently of ERAD [51]. The accumulation of ubiquitylated proteins in the ER of cdc48-3 cells (Fig. S4) suggests a possibility that some of the newly synthesized enzymes for cell wall biogenesis may be denatured and ubiquitylated in the ER during heat shock. Inability to efficiently degrade these proteins in cdc48-3 cells may perturb repair of the cell wall and sustain cell wall integrity pathway, leading to G1 arrest. This effect is specific to certain targets of Cdc48, rather than a general inhibition of protein degradation, because the level of Mpk1 phosphorylation in the proteasome mutants rpt2RF and rpt5S is comparable to that in the wild type at 38.5uC (Fig. S5C). In addition, these proteasome mutants are not sensitive to chemicals that perturb the cell wall (data not shown). These results suggest that defects in proteasomal degradation and accumulation of polyubiquitylated protein per se do not cause cell wall defect or Mpk1 overactivation. It is probable that these proteasome mutants do not accumulate polyubiquitylated proteins in the ER to perturb the synthesis or maturation of cell wall repair enzymes. The direct targets of Cdc48 in the cell wall biogenesis pathway remain to be determined in the future.

Materials and Methods
Growth of yeast YEPD medium contained 1% yeast extract, 2% bacto-peptone, and 2% glucose. Complete synthetic medium contained 0.67% yeast nitrogen base without amino acids (YNB), 16 complete supplement mixture (CSM) (Bio 101), and 2% glucose. For induction from the MET3 promoter, 16CSM was replaced by 16 CSM without methionine. To arrest cells at G1, a-factor (Sigma) was added at 1 mg/ml from a 10 mg/ml stock in DMSO and the cells were grown at 25uC for 3 hr. Cells were then shifted to the indicated temperature for 30 min. To release from G1 arrest, the cells were washed 3 times with warm water and then resuspended in pre-warmed medium at the indicated temperature. Table 1 lists the yeast strains used in this work. All strains are derivatives of W303, except cdc48-1 [1], rpt2RF, and rpt5S [52] that were used in Figure S5. npl4-1, ufd1-2, and rho1-104 mutants were backcrossed four times to W303. Gene deletions, epitope tagging, and introduction of pGAL to UBX2 were generated by PCRmediated integration [53]. For the reporter constructs, the promoter regions (1000 bp upstream of the start codon) of CLN1 and CLN2 were amplified by PCR from yeast genomic DNA and cloned at SacI and XbaI sites upstream of Renilla reniformis (RLUC) and Pyrophorus plagiophthalamus (spLUC) Luciferase, respectively, in pRS416. pCLN1-RLUC and pCLN2-spLUC regions were then removed with SacI and SmaI and cloned into the cognate sites in vectors pRS404 and pRS405, respectively. pRS404-pCLN1-RLUC and pRS405-pCLN2-spLUC were linearlized with Bsu36I and XcmI, respectively, to integrate into yeast genome at the selection markers.

FACS analysis
1 ml of yeast cells were pelleted and resuspended in ice-cold fix solution (40% ethanol, 0.1 M sorbitol, 5 mM EDTA, 5 mM sodium azide). Samples were temporarily frozen at 280uC. Cells were then washed with PBS plus 0.5% Triton X-100, and incubated with 100 mg/ml of RNaseA in 50 mM Tris-Cl (pH 8.0) overnight at 37uC. Cells were then resuspended in 100-200 ml of Sytox Green (Invitrogen, 1:400 in 38 mM sodium citrate) and briefly sonicated. Samples were diluted in 1 ml PBS and analyzed by FACSCalibur flow cytometer (BD Biosciences).

Luciferase assay
Yeast cells carrying pCLN1-RLUC and pCLN2-spLUC were grown to mid-log phase and arrested at G1 stage by a-factor. The cells were shifted to 37uC or 38.5uC for 30 min before release from the arrest at the same temperature. At each time point, OD 600 of the culture was measured and another 1.5 ml of yeast cells were collected. The cells were washed with cold PBS and frozen at 280uC until use. Cell pellets were resuspended to 65 ml in PBS, and 20 ml of the sample was used for each luminescence measurement with Vector3 luminometer plus autoinjector (Perkin Elmer). Through the autoinjector, 100 ml of 1 mM coelenterazine (Promega) and D-Luciferin (Sigma) was added to the sample for R. Luciferase and sp Luciferase, respectively, with 5-sec equilibration and 10-sec integration time for the measurement of luminescence. The results were normalized with OD 600 .

Subcellular fractionation
Subcellular fractionation was performed as described [55]. Briefly, 80 OD 600 of cells were disrupted by bead-beating in STE buffer ( Ethylmaleimide, 10 mg/ml each of leupeptin, pepstatin and chymostatin). Unbroken cells were pelleted by centrifugation at 3006g for 2 min at 4uC. Total lysate from the supernatant was then subjected to centrifugation at 100,0006g for 60 min at 4uC (TLS55 rotor, Beckman, Optima Max centrifuge). The resulting supernatant and pellet represent cytosol and membrane fractions, respectively. The membrane pellet was then resuspended in STE buffer and layered on top of a continuous gradient composed of 20-60% (wt/ wt) sucrose in 5 ml of TE buffer (10 mM Tris-HCl [pH 7.6], 10 mM EDTA). After centrifugation at 100,0006g for 18 hr at 4uC (MLS50.1 rotor, Beckman, Optima Max centrifuge), fractions of 356 ml were manually collected from top to bottom of the gradient. Proteins were resolved by SDS-PAGE for Western blot analysis.

Supporting Information
Figure S1 Accumulation of ubiquitin conjugates in cdc48-3 is independent of ERAD components. Wild-type and the indicated mutant strains were grown to mid-log phase and shifted to 38.5uC for 3 hr. pGAL-UBX2 and pGAL-UBX2 cdc48-3 were first grown in YEP containing galactose to mid-log phase and then changed to YPD to suppress Ubx2 expression for 2 hr before shifting to 38.5uC for 3 hr. Cell lysates were prepared for Western blots with anti-ubiquitin (Ub) and anti-Mad2 antibodies. Mad2 serves as a loading control.
(TIF) Figure S2 Deletion of XBP1 partially rescues budding defect of cdc48-3 at 38.56C. Wild type, cdc48-3, xbp1D, and xbp1D cdc48-3 cells were grown as described in Figure 1B and their budding index at the indicated times during the cell cycle entry were determined. Filled diamond, no bud; open circle, small bud; filled triangle, medium/large bud.
(TIF) Figure S3 The protein levels of Swi4, Swi6, and Far1 are unaffected by temperature up-shift in cdc48-3. Swi4, Swi6, and Far1 were tagged at the carboxyl-terminus with 3HA at the chromosomal loci in CDC48 and cdc48-3 cells. The cells were arrested at G1 with a-factor and released into the cell cycle at 38.5uC as described in Figure 1B. Samples were taken at the indicated times after the release for Western blot with anti-HA antibody. Asterisks denote cross-reacting proteins. The migration of molecular size standard is indicated on the left. (TIF) Figure S4 Ubiquitin conjugates are enriched in the ER fraction of cdc48-3. CDC48 and cdc48-3 cells were grown to mid-log phase and shifted to 38.5uC for 3 hr. Total cell lysates (lanes 1 and 2) were prepared and then separated into cytosol (lanes 3 and 4) and membrane (lanes 5 and 6) fractions. The membrane fractions were further fractionated by continuous 20-60% (wt/wt) sucrose gradient (lanes 7-20, CDC48; lanes 21-34, cdc48-3). The fraction numbers from top to bottom of the gradient are indicated. Equivalent to 1/500 of the total lysate, cytosol, and membrane fractions as well as 1/100 of each fraction from the sucrose gradient were subjected to Western blot analysis with antibodies against ubiquitin (Ub), ER protein Sec61, and plasma membrane protein Pma1.
(TIF) Figure S5 Phosphorylation of Mpk1 is enhanced in cdc48 mutants, but not in proteasome mutants. Cells of indicated genotypes were first grown at 25uC and then shifted to 14uC for 2 days or other indicated temperatures for 3 hr. Cell lysates were prepared for Western blots with anti-phospho-Mpk1 and anti-Mad2 antibodies. Mad2 serves as a loading control. (TIF)