The proteasome homeostasis in Saccharomyces cerevisiae is regulated by a negative feedback circuit in which the transcription activator Rpn4 upregulates the proteasome genes and is rapidly degraded by the assembled proteasome. Previous studies have shown that rpn4Δ cells are sensitive to a variety of stresses. However, the contribution of the loss of Rpn4-induced proteasome expression to the rpn4Δ phenotypes remains unclear because Rpn4 controls numerous genes other than the proteasome genes. Here we construct a yeast strain in which one of the essential proteasome genes, PRE1, is no longer induced by Rpn4. We show that the active proteasome level is lower in this strain than in the wild-type counterpart. Moreover, we demonstrate that loss of Rpn4-induced proteasome expression leads to cell-cycle delay in G2/M and sensitizes cells to various stresses. To our knowledge, this is the first report that explicitly reveals the physiological function of Rpn4-induced proteasome expression. This study also provides a tool for understanding the interactions between proteasome homeostasis and other cellular processes.
THE proteasome is responsible for degradation of abnormal proteins and regulators of growth and other cellular processes (Voges et al. 1999; Pickart and Cohen 2004; Groll et al. 2005). The 26S proteasome is a multi-subunit protease, consisting of a proteolytic core (20S core or 20S proteasome) capped at one or both ends by the 19S regulatory particle (also known as PA700). The 20S core is composed of 28 subunits arranged as a barrel-shaped stack of four layers, each of seven different subunits in an α7β7β7α7 configuration. Both exterior layers contain one copy of seven different α-type subunits. Likewise, both interior layers contain one copy of seven different β-type subunits. The 20S core possesses three types of catalytic activities, including trypsin-like, chymotrypsin-like, and peptidylglutamyl peptide hydrolase activities. These catalytic activities provided by three β-subunits are located inside the chamber of the 20S core. The 19S regulatory particle, also a multi-subunit complex, recruits and unfolds ubiquitylated substrates before their translocation into the 20S core for degradation. Recent studies have shown that a number of ubiquitylating and deubiquitylating enzymes are associated with the proteasome, suggesting that the proteasome is not just a machine for digesting proteins, but may also play an important role in specifying the protein substrates to be degraded (Kleijnen et al. 2000; Xie and Varshavsky 2000, 2002; Farrás et al. 2001; Jäger et al. 2001; Chen and Madura 2002; Verma et al. 2002; Yao and Cohen 2002; Hanna et al. 2006; Demartino and Gillette 2007; Ravid and Hochstrasser 2007). In addition to the 19S regulatory particle, several other proteins or complexes are also able to bind and activate the 20S core, including Blm10 or PA200 and the PA28 family proteins including PA28α, PA28β, and PA28γ (Demartino and Gillette 2007).
Although the structure and functions of the proteasome have been vigorously investigated, the mechanism regulating proteasome gene expression has just begun to emerge. Recent studies demonstrated that RPN4 (also named SON1 and UFD5) encodes a transcription activator that induces the proteasome genes (Mannhaupt et al. 1999; Jelinsky et al. 2000; Xie and Varshavsky 2001). An Rpn4-binding site, a 9-bp motif known as PACE (proteasome-associated control element), is found in the promoters of the proteasome genes. Interestingly, Rpn4 is an extremely short-lived protein (t1/2 ≤ 2 min) and degraded by the proteasome via ubiquitin-dependent and ubiquitin-independent pathways (Xie and Varshavsky 2001; Ju and Xie 2004). Moreover, stabilization of Rpn4 by inhibition of the proteasome activity results in an increase in expression of the proteasome genes (Ju et al. 2004; London et al. 2004). Together, these observations led to a model in which the proteasome homeostasis is regulated by a negative feedback circuit. On the one hand, Rpn4 upregulates the proteasome genes; on the other hand, Rpn4 is rapidly degraded by the assembled/active proteasome. The Rpn4-proteasome negative feedback circuit provides an efficient and sensitive means to gauge the proteasome abundance. Subsequent studies showed that a similar negative feedback mechanism also exists in higher eukaryotes, including humans (Fleming et al. 2002; Wójcik and Demartino 2002; Lundgren et al. 2003; Meiners et al. 2003; Xu et al. 2008).
The proteasome is quite abundant in the cell. It remains unclear if such a high abundance is of any physiological relevance. For example, it is not known if the cell is sensitive to a subnormal level of proteasome expression. It is also unclear whether the proteasome abundance maintained by Rpn4 is important for cell survival under stressed conditions. Although early studies have shown that rpn4Δ mutants are hypersensitive to a variety of stresses (Jelinsky et al. 2000; Gasch et al. 2001; Owsianik et al. 2002; Ju et al. 2004; London et al. 2004; Hahn et al. 2006; Haugen et al. 2004; Yokoyama et al. 2006), it is difficult to conclude that these phenotypes result from downregulation of the proteasome genes because Rpn4 also regulates numerous nonproteasome genes (Mannhaupt et al. 1999; Jelinsky et al. 2000). In this study we constructed a yeast strain in which PRE1 encoding one of the essential proteasome subunits is no longer induced by Rpn4. We found that the active proteasome level is lower in this strain than in the wild-type counterpart. Cell-cycle analysis showed that downregulation of PRE1 delays G2/M exit. Moreover, we demonstrated that loss of Rpn4-induced proteasome expression sensitizes cells to different stresses. This study explicitly reveals for the first time the physiological function of Rpn4-induced proteasome expression.
MATERIALS AND METHODS
Yeast strains used in this study included JD52 (MATa his3-Δ200 leu2-3, 112 lys2-801 trp1-Δ63 ura3-52), EJY140 (an rpn4Δ∷LEU2 derivative of JD52), YXY206 (MATa trp1-Δ63 ura3-52 his3-Δ200 leu2-3,112 lys2-801 PRE1-FLAG-His6∷YIplac211), and YXY210 (MATa trp1-Δ63 ura3-52 his3-Δ200 leu2-3,112 lys2-801 rpn4Δ∷LEU2 PRE1-FLAG-His6∷YIplac211). Details of these strains were described previously (Johnson et al. 1995; Ju et al. 2004).
Construction of PACE-less PRE1 yeast strains:
The knock-in vector used to generate strains expressing PRE1 from a PACE-less promoter was constructed as follows. PCR with primers YX714 (ATCCTCGAGCCTGGGTTCTGACTA) and YX715 (TCCAAGCTTGTAAAGATTTCGCTGCGAAAG) was used to amplify a PRE1 promoter fragment from −575 to −143, immediately upstream of the PACE motif (−142 to −134). This fragment (“fragment 1”) carries an XhoI site at the 5′-end and a HindIII site at the 3′-end. Another PCR with primers YX716 (ATTAAGCTTAAATAAAGAAAAGTGAATATTGAACAC) and YX717 (ACTGGATCCCCTAATTGACTTGGCTAATTC) amplified a fragment (“fragment 2”) from −133 to +280 of PRE1, with HindIII and BamHI sites at the 5′- and 3′-ends, respectively. Fragment 1 was digested with XhoI and HindIII, whereas fragment 2 was cut with HindIII and BamHI. These two fragments were then subcloned into a XhoI/BamHI-cut RS303 vector to get plasmid RS303-PRE1(right arm). A third PCR with primers YX718 (ACTGGATCCCCTAATTGACTTGGCTAATTC) and YX719 (CCTTCTAGAATCCTGTACACGGATGCC) amplified a fragment (“fragment 3”) from −98 to +40 of PRE1, with BamHI at 5′- and XbaI at 3′-ends, respectively. Fragment 3 was digested with BamHI and XbaI and subcloned into a BamHI/XbaI-cut RS303-PRE1(right arm) vector to get plasmid RS303-PRE1(knock-in). BamHI-linearized the RS303-PRE1(knock-in) vector [left arm (−98 to +40 of PRE1)-HIS3-Ampr-right arm (−575 to +280 of PRE1 with the PACE motif substituted by a HindIII site)] was transformed into strains YXY206 (for C-terminally FLAG-His6 tagged Pre1) and JD52 (for untagged Pre1). The disruption vector was integrated via recombination into the promoter region of PRE1 downstream of the PACE motif, separating the native PRE1 promoter from its coding sequences by the RS303 backbone. PRE1 was therefore expressed from a PACE-less (ΔPACE) promoter. HIS+ transformants were isolated and site-specific recombinants were verified. Specifically, genomic DNA was prepared from the HIS+ isolates and subjected to two PCR analyses. The first PCR with primers YX474 (ATCGGATCCTGATAGTTTGAGCCTGGG) and T7 was used to confirm site-specific integration of the left arm. Primer YX474 corresponds to −587 to −566 of the PRE1 promoter, whereas T7 primer anneals to the RS303 backbone downstream from the left arm. An ∼800-bp PCR product was generated from the desired recombinants but not from the parental strains. The second PCR with primers YX714 and YX717 amplified a PRE1 fragment (−575 to +280) from both wild-type and PACE-less PRE1 alleles. However, only the PCR product from the ΔPACE allele could be cut into two pieces by HindIII. Strains YXY362 and YXY368 derived from YXY206 and JD52 were confirmed to express PRE1 from a ΔPACE promoter. The phenotypes of the ΔPACE strains (slow growth and poor survival under stressed conditions) were corrected by transformation of a low-copy plasmid expressing PRE1 from its native promoter (−575 to −1), including the PACE motif.
Chromatin immunoprecipitation (ChIP) assays were carried out as described (Ausubel et al. 1998). In brief, wild-type (JD52) and ΔPACE (YXY368) strains expressing untagged or C-terminally FLAG-tagged Rpn4 or Rpn4*, a stabilized version of Rpn4, from the CUP1 promoter on vector RS424 were induced with 0.1 mm CuSO4 for 6 hr. After in vivo crosslinking of DNA and proteins with formaldehyde, cells were lysed and genomic DNA was sonicated to generate fragments ∼0.5 kb in length. FLAG-tagged Rpn4 or Rpn4* and bound DNA fragments were immunoprecipitated by anti-FLAG agarose (Sigma, St. Louis). After crosslink reversal, DNA fragments were purified and amplified by real-time quantitative PCR using an ABI StepOne instrument (Applied Biosystems). The primers (P1 and P2) used for real-time PCR corresponded to the positions of −98 to −81 and +259 to +279 of the PRE1 gene, respectively. The ChIP signals were calculated by normalizing the amount of PCR product from the immunoprecipitates to that from input samples before immunoprecipitation; this was expressed as the relative ratio against the nonspecific signal obtained from untagged Rpn4, which was set at 1.0.
The enzymatic activity of β-galactosidase in liquid yeast culture was determined by o-nitrophenyl-β-d-galactopyranoside (ONPG) assay as described (Ausubel et al. 1998), using the chromogenic substrate ONPG. The plasmids expressing Rpn4172-229-βgal and Rpn4172-229/K187R-β-gal from the CUP1 promoter were previously described (Ju and Xie 2006; Ju et al. 2008). For induction of the CUP1 promoter, yeast cells grown to OD600 of 0.3–0.5 were treated with CuSO4 at a final concentration of 0.1 mm for 6–7 hr.
Yeast cells were grown to OD600 of 0.8–1.2, harvested, and resuspended in lysis buffer (150 mm NaCl, 50 mm Tris.Cl, 5 mm EDTA, 1% Triton X-100, pH 7.5) plus protease inhibitor mix (Roche, Indianapolis). Yeast extracts were prepared using the glass-bead vortexing method (Ju et al. 2004). Protein concentrations were measured by Bradford assay. Approximately 20 μg of each extract was separated by SDS–PAGE (12% gel), followed by immunoblotting with a monoclonal anti-FLAG antibody (Sigma) and detection with the Odyssey infrared imaging system according to the manufacturer's instruction (Li-Cor Biosciences, Lincoln, NE). The blots were reprobed with an anti-yeast α-tubulin antibody (Chemicon, Temecula, CA) to verify equal loading.
Cell-cycle analysis with flow cytometry:
For analysis of synchronized cells, exponentially growing cultures (OD600 ≈ 1.0) were treated with 5 μg/ml α-factor for 2 hr and then washed and allowed to resume growth in YPD. Aliquots were withdrawn at 20-min intervals after release from G1 arrest and processed for flow cytometric DNA analysis as described previously with minor modifications (Xie and Varshavsky 2001). Briefly, cells were fixed in 70% ethanol and incubated with RNase A (0.2 mg/ml) at 37° for 3 hr, followed by treatment with Proteinase K (0.2 mg/ml) at 37° for 1 hr. Cells were then stained with SYBR Green (1:10,000) and analyzed by Becton Dickinson FACScan. Unsynchronized cells from mid-log phase cultures were also subjected to DNA flow cytometry to detect the distribution of cells in different phases.
Cell growth and survival assays:
Cell growth was assessed by serial dilution assays. Yeast cells were grown to exponential phase and normalized by optical density. Fivefold serial dilutions of cells were spotted on selective plates and incubated at 30°. Cell survival rates were measured by colony formation assays. Overnight yeast cultures were diluted and continued to grow to OD600 of 0.8–1.2 before subjected to stresses. For methyl methanesulfonate (MMS) and tert-butyl hydroperoxide (t-BuOOH) stresses, the yeast cells were treated for 1 hr at 30°, followed by extensive washing and seeding on synthetic complete (SC) plates. Approximately 50,000 cells treated with 0.3% MMS or 30 mm t-BuOOH and 5000 cells treated with 0.1% MMS or 15 mm t-BuOOH were spread on SC plates. Approximately 500 untreated cells were spread on SC plates as a control. The formed colonies were counted 72 hr after plating. For the stresses with DTT and CdCl2, ∼5000 yeast cells were directly seeded on the plates containing the respective stressing agents. For ultraviolet (UV) stress, ∼100,000 cells (for 200 J/m2) and 5000 cells (for 100 J/m2) were plated and incubated for 4 hr before exposure to UV radiation. The plates were kept in the dark afterward. The survival rates of treated cells were normalized against that of untreated cells, which was set at 100%. Each survival assay was repeated at least three times.
Construction of yeast strains expressing PRE1 from a PACE-less promoter:
Genomewide analyses have shown that Rpn4 regulates numerous genes other than the proteasome genes (Mannhaupt et al. 1999; Jelinsky et al. 2000). To study the biological significance of Rpn4-induced proteasome expression, we wanted to generate a strain in which Rpn4 remains active but the abundance of assembled/active proteasome is no longer upregulated by Rpn4. Since the proteasome subunits are nearly stoichiometrically present in Saccharomyces cerevisiae (Glickman et al. 1998; Russell et al. 1999b), we reasoned that downregulation of one of the essential subunits should reduce the level of assembled and active proteasome. Using a site-specific recombination approach, we constructed yeast strains in which the PRE1 proteasome gene is expressed from a promoter that lacks the PACE sequence (Rpn4-binding motif) at its chromosomal locus (Figure 1A). To ascertain that Rpn4 is unable to bind to the PACE-less promoter in a ΔPACE strain, we performed ChIP assays. The ΔPACE strain retains an intact PRE1 promoter, which is separated from the PRE1 open reading frame (ORF) by the 5-kb RS303 vector backbone and the ΔPACE promoter (Figure 1A). The intact PRE1 promoter may still recruit Rpn4 even though it cannot drive the expression of the PRE1 ORF. To solve this problem, we designed a pair of primers (P1 and P2) for the ChIP assays, which correspond to the region between −98 and +279 of the PRE1 gene (Figure 1A). The genomic DNA was broken down into fragments with an average size of 0.5 kb in the ChIP assays, so that the intact PRE1 promoter in the ΔPACE strain could not be amplified by PCR with primers P1 and P2. Since Rpn4 is extremely short lived in vivo, we expected that the ChIP signals would be weak. To facilitate detection of the ChIP signals, we took advantage of a stabilized mutant of Rpn4 (Rpn4*), which has the N-terminal 10 amino acids and the ubiquitin-dependent degradation signal deleted, and yet retains the transcription activity (Ju and Xie 2006 and data not shown). Wild-type and ΔPACE strains were transformed with high-copy vectors expressing C-terminally FLAG-tagged Rpn4 (Rpn4-FLAG) or Rpn4* (Rpn4*-FLAG) from the induced CUP1 promoter. Transformants overexpressing untagged Rpn4 were used as a control to normalize the nonspecific immunoprecipitation by the anti-FLAG antibody. As shown in Figure 1B, Rpn4-FLAG was readily recruited to the PRE1 promoter in the wild-type strain but not to the PACE-less promoter in the ΔPACE strain. As expected, the ChIP signal from Rpn4*-FLAG was even stronger. These results confirmed that the PACE motif is vital for recruitment of Rpn4 to the PRE1 promoter.
Diminished proteasome activity in the ΔPACE strain:
We then compared the steady-state levels of Pre1 with wild-type, ΔPACE, and rpn4Δ strains in which a FLAG-His6 (FH) epitope is attached to the C terminus of Pre1 expressed from the chromosomal locus. Immunoblotting analysis with an anti-FLAG antibody showed that the Pre1 protein level in ΔPACE is comparable to that in rpn4Δ, but markedly lower than that in the wild type (Figure 1C). These results indicate that Rpn4 cannot activate the PACE-less PRE1 promoter. To examine whether the proteasome activity, an indicator of the assembled and active proteasome level, is low in the ΔPACE strain, we measured the degradation of Rpn4172-229-β-gal in wild-type, ΔPACE, and rpn4Δ strains. Rpn4172-229-βgal is a short-lived β-gal fusion protein carrying the ubiquitin-dependent degradation signal of Rpn4 (Ju and Xie 2006). Rpn4172-229/R187K-β-gal, a stabilized derivative of Rpn4172-229-βgal with a K187R substitution, was used as a control to ensure that the disparity of β-gal activity is not due to variation of transcription and/or translation of the β-gal substrates. As shown in Figure 1D, the proteasome activity in the ΔPACE strain was similar to that in rpn4Δ, but much weaker than that in the wild-type strain. Transformation of a low-copy plasmid expressing PRE1 from its native promoter (with the PACE motif) in the ΔPACE strain completely rescued the proteasome activity. Thus, loss of Rpn4-induced expression of PRE1 leads to a decrease in the abundance of assembled and active proteasome.
Loss of Rpn4-induced proteasome expression causes cell-cycle delay at G2/M:
To explore the phenotypes caused by loss of Rpn4-induced proteasome expression, we first compared the growth of ΔPACE, wild-type, and rpn4Δ strains. Although these three strains showed no notable difference in colony formation efficiency (data not shown), ΔPACE and rpn4Δ grew considerably slower than the wild-type counterpart (Figure 2A). The slow growth of ΔPACE cells was overcome by adding back a low-copy plasmid expressing PRE1 from its native promoter. Thus, lack of Rpn4-induced proteasome expression impairs cell growth. To gain insight into the underlying mechanism, we compared the cell-cycle progression of ΔPACE and wild-type strains. Cell cultures were synchronized in G1 phase with α-factor, followed by flow cytometry of cellular DNA content as a function of time after release from G1 arrest. Clearly, ΔPACE cells displayed a delay in G2/M when compared to the wild-type cells (Figure 2B). Consistent with the data from the synchronized cultures, unsynchronized (mid-log phase) ΔPACE culture had a higher percentage of G2/M cells than the wild-type counterpart (Figure 2C). These results indicate that Rpn4-induced proteasome expression is essential for normal G2/M exit.
Rpn4-induced proteasome expression is crucial for cell survival under stressed conditions:
We further examined whether Rpn4-induced proteasome expression is important for cell survival under stressed conditions. Given the effect of the loss of Rpn4-induced proteasome expression on cell-cycle progression, we compared the sensitivity of ΔPACE, wild-type, and rpn4Δ strains to DNA-damaging agents, including UV radiation and the alkylating agent MMS. The survival rates were measured by colony formation assays (Figure 3, A and B). We found that ΔPACE cells were more sensitive to UV and MMS than wild-type cells, especially at higher doses. The survival rate of ΔPACE cells after treatment with 200 J/m2 UV or 0.3% MMS for 1 hr was ∼0.05%, whereas >1% of wild-type cells survived the same stresses. Introducing a low-copy plasmid expressing PRE1 from its native promoter in the ΔPACE cells increased the resistance to UV and MMS to a similar level as the wild-type cells. This further demonstrated that the hypersensitivity of ΔPACE cells to UV and MMS results from insufficient expression of the Pre1 proteasome subunit. Interestingly, ΔPACE cells had a comparable UV and MMS sensitivity as rpn4Δ cells. These observations suggest that loss of Rpn4-induced proteasome expression may be a major cause of the hypersensitivity of rpn4Δ cells to DNA-damaging agents (Jelinsky et al. 2000; Ju et al. 2004; London et al. 2004).
We also assessed the sensitivity of ΔPACE cells to the oxidizing agent t-BuOOH and the reducing agent dithiothreitol (DTT). Only ∼20 and 0.05% of ΔPACE cells grew to form colonies after treatment with 15 and 30 mm t-BuOOH for 1 hr, whereas >30 and 0.4% of wild-type cells survived the same treatments (Figure 3C).
ΔPACE cells were also more sensitive to DTT than wild-type cells (Figure 3D). Again, ΔPACE cells exhibited a similar sensitivity as rpn4Δ cells to t-BuOOH and DTT, and adding back the low-copy plasmid expressing PRE1 from its native promoter reduced the sensitivity of ΔPACE cells. Together, these results indicate that Rpn4-induced proteasome expression is important for cell survival in response to oxidizing and reducing stresses.
Recent study has shown that Rpn4 is vital for the cell's adaptation to stress caused by toxic metals (Haugen et al. 2004). We tested whether Rpn4-induced proteasome expression is important for cell viability in response to cadmium. Specifically, we measured the colony formation efficiency of the ΔPACE, wild-type and rpn4Δ strains in medium containing 0, 30, and 60 mm CdCl2. As shown in Figure 3E, the ΔPACE cells were significantly more sensitive to CdCl2 than the wild-type cells. Unlike other stresses, ΔPACE cells appeared relatively more resistant to CdCl2 than rpn4Δ cells. This suggests that a normal-level expression of other Rpn4 target genes may also be important for cell viability under CdCl2 stress. Nevertheless, our results indicate that Rpn4-induced proteasome expression is critical for cell survival in response to CdCl2 stress.
In this study we demonstrated that downregulation of one of the proteasome genes by deletion of the Rpn4-binding site from the PRE1 promoter was able to reduce the active proteasome level in the cell. This result is in line with the observation that the proteasome subunits are stoichiometrically present in S. cerevisiae (Glickman et al. 1998; Russell et al. 1999b). Thus, to maintain a normal level of proteasome activity in a cell appears to require coordinate expression of all the proteasome genes. This finding is of potential clinical relevance because it provides an approach to reducing the proteasome activity in cancer cells. The current method of targeting the proteasome in cancer therapy is limited to use of proteasome inhibitors that attack the catalytic sites of the proteasome (Orlowski and Kuhn 2008). Unfortunately, recent clinical trials have shown that many types of cancer cells are resistant to proteasome inhibitors. Lowering the assembled proteasome level by knocking down one of the proteasome genes can be an important alternative to using proteasome inhibitors for inhibiting the proteasome activity in cancer cells.
We showed here that subnormal-level expression of PRE1 impairs cell growth. Previous studies have reported that proteasome mutants grow more slowly than wild-type cells. Our results, however, for the first time demonstrated that loss of Rpn4-induced proteasome expression can also result in defects in cell growth. It is interesting that ΔPACE cells grow as slowly as the rpn4Δ cells in the serial dilution assay (Figure 2A). Given that Rpn4 controls numerous genes in addition to the proteasome genes, this result suggests that activation of the proteasome genes is likely a major function of Rpn4 required for normal cell growth. However, this does not imply that normal-level expression of other Rpn4 target genes is not important for cell growth. It is possible that downregulation of the nonproteasome genes may not display a significant additive effect with subnormal-level expression of the proteasome in rpn4Δ cells. Although the details underlying the slow growth of ΔPACE cells remain to be investigated, one simple explanation is that lack of Rpn4-induced proteasome expression causes G2/M delay. It is important to note that other cell-cycle phases are not markedly delayed in the ΔPACE cells (Figure 2B). In line with our observation, previous study has shown that cim3-1 and cim5-1 proteasome mutants are also delayed in G2/M (Ghislain et al. 1993). Thus, the demand for proteasome activity is apparently higher prior to G2/M exit than in other phases. It is possible that degradation of one or more of the G2/M inhibitors requires a higher proteasome activity. However, we cannot rule out the possibility that the involvement of the proteasome in G2/M exit may be independent of its proteolytic activity.
Cell response to environmental stress is a complex process. Recent studies in S. cerevisiae have revealed a stress-response network in which Rpn4 may serve as a major regulator (Figure 4). In addition to the proteasome genes, Rpn4 upregulates other key stress response genes involved in protein ubiquitylation, DNA repair, and other cellular processes (Mannhaupt et al. 1999; Jelinsky et al. 2000; Gasch et al. 2001). Interestingly, RPN4 itself is regulated by a wide range of signals (Jelinsky et al. 2000; Owsianik et al. 2002; Haugen et al. 2004; Ju et al. 2004; London et al. 2004; Hahn et al. 2006). The RPN4 promoter carries response elements for heat-shock transcription factor (Hsf1), multidrug-resistance-related transcription factors (Pdr1 and Pdr3), and Yap1, a transcription factor that plays an important role in response to oxidation, toxic metal, and MMS. These transcription factors are activated by various environmental stresses and in turn induce the expression of the RPN4 gene (Owsianik et al. 2002; Haugen et al. 2004; Hahn et al. 2006). In support of the central role of Rpn4 in the stress response network, early studies have shown that rpn4Δ cells are hypersensitive to different stressing agents (Jelinsky et al. 2000; Gasch et al. 2001; Owsianik et al. 2002; Haugen et al. 2004; Ju et al. 2004; London et al. 2004; Hahn et al. 2006; Yokoyama et al. 2006). However, it has been difficult to determine the Rpn4 target genes whose downregulation contributes to the hypersensitivity because Rpn4 controls numerous genes.
Among the pathways involving Rpn4, the Rpn4-proteasome negative feedback circuit provides an intriguing mechanism that not only regulates the proteasome homeostasis but also gauges the expression of other Rpn4 target genes via keeping the Rpn4 protein level in check (Figure 4). Rpn4-induced proteasome expression is an essential element of this mechanism. Early studies have shown that inhibition of the proteasome activity leads to stabilization of Rpn4 and subsequent upregulation of the proteasome genes, which may compensate the impairment imposed to the proteasome (Ju et al. 2004; London et al. 2004). Consistently, deletion of RPN4 or inactivation of Rpn4 exhibits a strong synthetic lethality with mutations of several proteasome subunits (Xu and Kurjan 1997; Fujimoro et al. 1998; Ju et al. 2004). Recently, Hanna et al. (2007) suggested that depletion of ubiquitin (ubiquitin stress) may trigger an Rpn4-dependent upregulation of the proteasome genes, which, along with an increased expression of the polyubiquitin gene UBI4 and the gene encoding deubiquitylating enzyme Ubp6, enable the cell to restore a normal ubiquitin level. In this study, we were able to demonstrate that downregulation of PRE1 by deletion of the Rpn4-binding site from its promoter leads to subnormal-level expression of the assembled/active proteasome. Taking advantage of the ΔPACE strain, we showed that Rpn4-induced proteasome expression is crucial for cell survival in response to various stresses. The similar survival rates of ΔPACE and rpn4Δ cells under many of the stressed conditions strongly suggest that subnormal-level expression of the proteasome genes is a major cause of the hypersensitivity of rpn4Δ cells to the stresses.
The role of Rpn4-induced proteasome expression in stress response may involve several aspects. First, Rpn4-induced proteasome upregulation may facilitate the degradation of some regulatory proteins that otherwise inhibit the stress-response pathways under normal conditions. By doing so, the cell can maneuver a cost-effective response to environmental insults. Second, an increase in proteasome expression is likely required for rapid removal of damaged or misfolded proteins generated in large amounts during stress. This may help reestablish the cellular homeostasis. Third, the function of Rpn4-induced proteasome expression in stress response may also be in a proteolysis-independent manner. Previous study has suggested that the 19S regulatory particle facilitates DNA repair independently of proteasomal degradation (Russell et al. 1999a). Rpn4-induced expression of the 19S proteasome genes may therefore be important for repairing the DNA damaged by stress. Although the 19S proteasome genes are not downregulated in the ΔPACE strain, the assembly of the 19S regulatory particle may be impaired because the level of the 20S proteasome, which is required for efficient assembly of the 19S regulatory particle in vivo (Kusmierczyk et al. 2008), is reduced due to insufficient Pre1 subunit. Thus, the sensitivity of ΔPACE cells to DNA-damaging agents may be partly attributed to a lower level of assembled 19S regulatory particles. The ΔPACE strain may serve as an interesting model for further understanding the role of the 19S regulatory particle as well as the 20S proteasome in DNA repair.
Whereas Rpn4-induced proteasome expression is clearly an important factor for cell survival under stressed conditions, the contributions of other Rpn4 target genes cannot be ignored. For example, deletion of either RAD23 or MAG1, two DNA repair genes upregulated by Rpn4, sensitizes cells to DNA-damaging agents (Jelinsky et al. 2000). We also noted that rpn4Δ is more sensitive to CdCl2 than the ΔPACE strain. This observation suggests that upregulation of other Rpn4 target genes is also important for cell viability in response to CdCl2 stress. Further study is required to dissect different Rpn4-mediated pathways to define their roles in response to environmental stresses.
We thank Marie Piechocki for assistance in flow cytometry and the reviewers for their constructive comments on our manuscript. This study was supported by the American Cancer Society (RGS-0506401-GMC) and the National Science Foundation (MCB-0816974). Xiaogang Wang is a Ph.D. candidate of Fudan University, Shanghai, China.
- Received July 29, 2008.
- Accepted September 16, 2008.
- Copyright © 2008 by the Genetics Society of America