Spg5 Protein Regulates the Proteasome in Quiescence

Background: The ubiquitin-proteasome system is the major pathway for selective protein degradation in eukaryotes.

Results: Spg5 is induced in quiescence, binds proteasome components, and is required for proteasome function in quiescence.

Conclusion: Spg5 is a novel proteasome regulator, which promotes proteasome function in quiescence.

Significance: Spg5 provides a link between the proliferative state of the cell and the ubiquitin-proteasome pathway.

Abstract

The ubiquitin-proteasome system is the major pathway for selective protein degradation in eukaryotes. Despite extensive study of this system, the mechanisms by which proteasome function and cell growth are coordinated remain unclear. Here, we identify Spg5 as a novel component of the ubiquitin-proteasome system. Spg5 binds the regulatory particle of the proteasome and the base subassembly in particular, but it is excluded from mature proteasomes. The SPG5 gene is strongly induced in the stationary phase of budding yeast, and spg5Δ mutants show a progressive loss of viability under these conditions. Accordingly, during logarithmic growth, Spg5 appears largely dispensable for proteasome function, but during stationary phase the proteasomes of spg5Δ mutants show both structural and functional defects. This loss of proteasome function is reflected in the accumulation of oxidized proteins preferentially in stationary phase in spg5Δ mutants. Thus, Spg5 is a positive regulator of the proteasome that is critical for survival of cells that have ceased to proliferate due to nutrient limitation.

Introduction

All living cells exist in either a state of active proliferation or a state of quiescence. Although proliferation has garnered greater attention, the majority of cells, including most human cells, are quiescent (1). In quiescence, the cessation of growth is accompanied by profound modifications of metabolic and cellular regulatory pathways. The depletion of nutrients from a fixed culture of Saccharomyces cerevisiae is marked by a distinct inflection of the growth curve. This transition, the diauxic shift, represents the transition from fermentation to oxidative metabolism. Over the course of days, yeast cells enter a true period of quiescence, referred to as stationary phase, to reflect the complete cessation of net growth. Under these conditions, protein synthesis and degradation continue, but at markedly reduced rates, and remain finely balanced as amino acid precursors are increasingly depleted from the media and must be supplied through the breakdown of endogenous proteins. Autophagy, which provides for nonspecific breakdown of cytoplasmic proteins, as well as selective degradation of organelles such as mitochondria, is induced by nutrient limitation and becomes critical for cell survival under these conditions (2–5). Cells can survive in this quiescent state for extended periods of time and upon nutrient repletion will in time restore growth-associated regulatory circuits. Thus, budding yeast provides a tractable model for studying quiescence, and indeed there are broad parallels between quiescence in yeast and in higher organisms (1, 6).

The ubiquitin-proteasome system is the major pathway of selective protein degradation in eukaryotes (7). Although considerable progress has been made in understanding how protein biosynthesis is coordinated with growth and quiescence (8), little is known about the relationship between cell growth and protein turnover. The proteasome is a 2.5-MDa multisubunit protease whose proteolytic active sites are sequestered inside the cylindrical central portion of the complex known as the core particle (CP)⁴ (9). At either end of the CP is a regulatory particle (RP) whose function is to recognize proteins requiring destruction and assist the CP in their disposal. Proteins are designated for proteasome-mediated degradation primarily by the covalent attachment of ubiquitin. Among the functions of the regulatory particle are recognition and binding of this ubiquitin marker, unfolding of the protein substrate, insertion of the unfolded protein into the CP, and removal of the ubiquitin targeting signal (9).

Given the size and complexity of the proteasome, it is perhaps not surprising that there exists a large group of accessory factors, sometimes referred to as proteasome chaperones, which assist in its assembly and maturation. Some of these factors function primarily in CP biogenesis (Ump1 and Pba1–4), whereas others function in RP assembly (Nas6, Hsm3, Rpn14, and Nas2) (10). A common feature of these chaperones is their ability to interact with nascent proteasome species, while being excluded for the most part from mature, fully assembled proteasomes.

We report here the identification of a quiescence-specific regulator of the proteasome, Spg5. This protein directly binds the regulatory particle of the proteasome and is preferentially expressed in stationary phase. Its molecular function appears largely dispensable under conditions of nutrient repletion. By contrast, after nutrient limitation has led to the cessation of growth, the absence of Spg5 results in progressive loss of viability, likely reflecting the essential function of Spg5 in the ubiquitin-proteasome system in quiescence.

EXPERIMENTAL PROCEDURES

Yeast Strains

Yeast strains are listed in Table 1. Standard techniques were used for strain constructions and transformations (12). Yeast were cultured at 30 °C unless otherwise noted. YPD medium consisted of 1% yeast extract, 2% Bacto-peptone, and 2% dextrose.

TABLE 1.

Yeast strains

The plasmid pEL36 encodes RPT1-TEV-ProA. YCplac33 encodes an empty vector (URA3). pJH149 encodes a centromeric yeast expression vector for SPG5.

Strain	Genotype	Ref.
SUB62	MATa lys2-801 leu2-3, 2-112 ura3-52 his3-Δ200 trp1-1	33
SDL133	MATa lys2-801 leu2-3, 2-112 ura3-52 his3-Δ200 trp1-1 rpn11::RPN11-TEV-ProA (HIS3)	14
SDL135	MATa lys2-801 leu2-3, 2-112 ura3-52 his3-Δ200 trp1-1 pre1::PRE1-TEV-ProA (HIS3)	14
SY36	MATa lys2-801 leu2-3, 2-112 ura3-52 his3-Δ200 trp1-1 rpt1:: (HIS3) pEL36 (TRP1)	14
sJH92	MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0	34
sJH301	MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 spg5::KAN	34
sJH303	MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 RPN11::RPN11-Tev-ProA (ClonNAT)	This study
sJH304	MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 spg5::KAN RPN11::RPN11-Tev-ProA (ClonNAT)	This study
sJH370	MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 [YCplac33]	This study
sJH371	MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 spg5::KAN [YCplac33]	This study
sJH372	MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 spg5::KAN [pJH149]	This study
sJH386	MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 RPN11::RPN11-Tev-ProA (ClonNAT) [YCplac33]	This study
sJH387	MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 spg5::KAN RPN11::RPN11-Tev-ProA (ClonNAT) [YCplac33]	This study
sJH388	MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 spg5::KAN RPN11::RPN11-Tev-ProA (ClonNAT) [pJH149]	This study

Plasmids

The coding sequence of SPG5 was cloned into the vector pGEX-4T-1 to generate plasmid pJH101, which allows for expression and purification of GST-Spg5 from bacteria. The coding sequence of SPG5, along with 500 bp prior to the start codon and 200 bp following the termination codon, was cloned into the yeast centromeric vector YCplac33 (URA3) to create pJH149.

Recombinant Proteins

Recombinant GST-Spg5 was expressed from plasmid pJH101 in Escherichia coli Rosetta 2 (DE3) cells (Novagen). To prepare partially purified lysates, cells were harvested and resuspended in phosphate-buffered saline (PBS) supplemented with protease inhibitors, lysed by a French press, and clarified by centrifugation at 21,000 × g at 4 °C for 25 min. The protease inhibitors were aprotinin solution (Sigma), leupeptin (2 μg/ml), antipain (4 μg/ml), benzamidine (20 μg/ml), chymostatin (2 μg/ml), pepstatin (2 μg/ml), and 4-(2-aminoethyl)benzenesulfonyl fluoride (1 μm).

To prepare purified protein, GST-SPG5 was expressed in Rosetta 2 (DE3) cells. Cells were harvested and resuspended in PBS supplemented with an additional 750 mm NaCl and protease inhibitors (as above), lysed by a French press, and clarified by centrifugation at 21,000 × g at 4 °C for 25 min. Clarified lysate was incubated with glutathione-Sepharose 4B resin (GE Healthcare) at 4 °C for 1.5 h. The resin was then washed with 100 bed volumes of PBS containing 750 mm NaCl, followed by 10 bed volumes of GST elution buffer containing 50 mm Tris-HCl (pH 7.5), 1 mm DTT, and 150 mm NaCl. The recombinant protein was then eluted by incubating the resin with elution buffer containing 15 mm reduced glutathione for 20 min at room temperature.

Proteasome Purification

Proteasome purification was carried out by affinity methods as described previously (13, 14). Briefly, the appropriate strains were grown in YPD or minimal media for the indicated amounts of time. For standard purified proteasomes, cells were grown overnight (16–20 h) and were in logarithmic phase growth (doubling time of ∼2.0 h) at the time of harvesting. For experiments in stationary phase, cells were cultured for 6 days (YPD) or 10 days (minimal medium). Clarified lysates were prepared, and proteasome complexes were immobilized on IgG resin. Proteasome holoenzyme was prepared by washing with purification buffer (50 mm Tris-HCl [pH 7.5], 1 mm ATP, and 5 mm MgCl₂) supplemented with 50 mm NaCl. RP and CP were prepared by incubating immobilized proteasomes for 1 h at room temperature in purification buffer containing 500 mm NaCl and then washing with the same buffer. Lid and base were prepared by incubating immobilized proteasomes for 1 h at room temperature in purification buffer containing 1 m NaCl, and then washing with the same buffer. For Spg5 binding assays, GST-Spg5 lysate or purified protein was added to the immobilized complexes as indicated, incubated at 4 °C for 1 h, washed with 100 bed volumes of purification buffer containing 50 mm NaCl, and eluted via tobacco etch virus protease.

Native Gel Analysis

Native gel analysis was performed by nondenaturing PAGE (4%) as described previously (15). In-gel proteasome activity assays were performed by incubating gels with 100 μm suc-LLVY-AMC at 30 °C for 5–15 min in proteasome purification buffer. LLVY hydrolyzing activity was visualized by ultraviolet transillumination. For Western blotting, gels were transferred to PVDF membranes and processed by standard immunohistochemical techniques.

Quantitative Real Time PCR

Strains were grown in YPD for the indicated length of time. Cells were lysed using acid-washed glass beads, and RNA was isolated by phenol/chloroform extraction. Total RNA was purified using the RNeasy method (Qiagen) according to the manufacturer's instructions, and RNA integrity was verified by evaluation of ribosomal RNA by agarose gel electrophoresis. Two hundred nanograms of RNA were reverse-transcribed and assayed in quadruplicate using TaqMan Gene Expression Assay and analyzed as described previously (16).

Yeast Spot Assays

Cultures were grown in YPD for the indicated amount of time, diluted in sterile water, normalized to an A₆₀₀ of 0.2, and spotted in 3-fold serial dilutions onto the indicated plates, which were grown at the indicated temperatures for 2–5 days.

Detection of Oxidized Proteins

The presence of protein carbonyls was assayed by Western blot analysis according to the manufacturer's instructions (OxyBlot protein oxidation detection kit, Millipore). Briefly, cultures were grown in YPD or synthetic media lacking the appropriate nutrient for the indicated amount of time. Cells were harvested and washed with 50 mm Tris-HCl (pH 8.0), 1 mm EDTA, 5 mm MgCl₂, and 1 mm ATP, then resuspended in 10 ml of the same buffer. Resuspended cells were lysed using a French press; the lysate was clarified by centrifugation at 21,000 × g for 25 min and filtered through cheesecloth. Fifteen μg of total protein was reacted with 2,4-dinitrophenylhydrazine and separated on a 4–12% BisTris gel (Invitrogen). After transferring to a PVDF membrane, protein carbonyls were visualized by the chemiluminescence of a secondary antibody attached to a primary antibody specific to dinitrophenylhydrazone-derivatized residues (OxyBlot).

RESULTS

Spg5 Binds the Regulatory Particle but Not Proteasome Holoenzyme

Proteome-wide protein-protein interaction studies suggested that Spg5 might interact with several proteasome subunits (17, 18). To test for an interaction between Spg5 and the proteasome, we used an affinity purification method to isolate yeast proteasomes (14), and we challenged them with recombinant GST-tagged Spg5 expressed in bacteria. However, we were unable to detect any binding of Spg5 to the proteasome (Fig. 1 and data not shown). Concerned that endogenous Spg5 might have blocked recognition of GST-Spg5, we repeated the experiment with proteasomes purified from an spg5Δ strain, but we still failed to detect any binding of GST-Spg5 (data not shown).

FIGURE 1.

Spg5 binds to the RP and base but not mature proteasomes. A, binding of bacterially expressed GST-Spg5 (pJH101) to purified proteasome or proteasome subcomplexes. Visualization is by anti-GST antibody. Input represents ∼3% of the total GST-Spg5 present in each binding assay. GST-Spg5 was added in excess of the proteasome species. B, electrophoretic profile of the proteasome and proteasome subcomplexes from A, as visualized by Coomassie staining. (26 S, RP, and lid were generated from strain SDL133; base from SY36; CP from SDL135). Note that our CP preparations have a minor contaminant that comigrates with GST-Spg5, which may account for the weak band of this mobility in the 2nd lane, because this band is not detected by antibody to GST (A). C, binding of purified GST-Spg5 expressed in bacteria to the base subcomplex, as visualized by Coomassie stain (upper panel) or Western blot with ant-GST antibody (lower panel). The asterisk indicates a fragment of GST-Spg5 that comigrates with bona fide GST. The double asterisk indicates a nonspecific band.

The proteasome comprises two subcomplexes, the 28-subunit core particle (CP) and the 19-subunit RP, the latter itself consisting of two subcomplexes, the lid and the base (9). Reasoning that Spg5, although not binding to fully assembled proteasomes, might recognize a proteasome subcomplex, we next tested binding of Spg5 to the CP and the RP. We found that Spg5 bound to the RP, but not the CP, and within the RP, the major binding determinant for Spg5 proved to be the base (Fig. 1A). In fact, Spg5 bound to the base more strongly than to the RP, a binding pattern that has been seen for other base-interacting proteins (14). The purified proteasome complexes were also visualized by Coomassie staining (Fig. 1B), which, in addition to confirming the purity and uniform loading of the proteasome species, highlighted the abundance of bound Spg5. By visual inspection, Spg5 appeared to approach 1:1 stoichiometry with integral subunits of the base. Binding proved reversible, as bound Spg5 could be liberated from the base by high salt concentrations (500 mm NaCl; data not shown).

The binding studies described above utilized partially purified bacterial lysates expressing GST-Spg5. Further purification of GST-Spg5 initially proved difficult under standard conditions. However, when high salt concentrations were included during purification, higher yields of purified GST-Spg5 were obtained (Fig. 1C, 4th lane). Using this purified GST-Spg5, we next tested whether the binding of Spg5 to proteasome subassemblies was in fact direct. We were able to detect significant binding of Spg5 to the base, as demonstrated by Coomassie staining and Western blotting with anti-GST antibodies (Fig. 1C). There was no binding of Spg5 to beads alone (Fig. 1C), and free GST showed no binding to the base subcomplex (Fig. 1C and data not shown).

Proteasomes from spg5Δ Mutants Are Similar to Wild-type under Nutrient-rich Conditions

We next sought to determine the effects of deleting the SPG5 gene on proteasomes. To do this, we purified proteasomes from wild-type and spg5Δ strains. The electrophoretic profiles of these proteasomes were similar (Fig. 2A). Succinyl-LLVY-AMC, a tetrapeptide substrate of the proteasome, allowed us to assay the proteolytic activity of the purified proteasomes; again, we were unable to detect a significant difference (Fig. 2B). The purified proteasome, analyzed by nondenaturing gel electrophoresis, appeared to consist of predominantly RP₂CP and RP₁CP forms, with a small amount of CP, which likely dissociated after purification. However, we did not see significant differences in either the enzymatic activity (Fig. 2C) or the composition of these species in the spg5Δ mutant (Fig. 2D).

FIGURE 2.

spg5Δ proteasomes are similar to those of wild type in logarithmic or late-logarithmic phase cultures. A, electrophoretic profile of proteasomes purified from wild-type (SJH303) and spg5Δ strains (SJH304), visualized by Coomassie staining. B, suc-LLVY-AMC hydrolyzing activity of proteasomes (5 μg) from A. Control, buffer only. C, native gel electrophoresis coupled with in-gel suc-LLVY-AMC assay highlights proteolytically active proteasome species, as indicated, for the purified proteasomes from A. D, electrophoretic profile of the proteasome species, separated by native gel electrophoresis as in C, and visualized by Coomassie staining.

SPG5 Is Induced in Stationary Phase

SPG5 was first identified in a genome-wide, microarray-based evaluation of stationary phase transcription (19). Given our inability to detect an effect of the spg5Δ mutation on proteasome function, we sought to examine the dependence of SPG5 transcription on the growth state of the culture in greater detail. Using quantitative real time PCR, we found that SPG5 is progressively induced in stationary phase (Fig. 3A), with an ∼10-fold induction over logarithmic phase growth after 5 days of culture, and with statistically significant induction seen as early as after 2 days of culture (Fig. 3A). In contrast, an unrelated species, the 18 S ribosomal RNA (rRNA), showed no time-dependent increase in abundance.

FIGURE 3.

Stationary phase transcriptional induction of SPG5. A, quantitative real time PCR analysis of SPG5 transcription after 2 and 5 days of culture, expressed as fold-induction relative to exponential phase growth. The 18 S ribosomal RNA serves as a control. Error bars indicate standard deviations of experiments carried out in triplicate. RNA was generated from strain SUB62. White bars, 18 S RNA; black bars, SPG5 RNA. B, growth of wild-type (SJH92) and spg5Δ (SJH301) strains on rich medium at the indicated temperatures after growth for 28 or 60 days in rich medium at 30 °C.

To evaluate the physiologic significance of SPG5 transcriptional induction, we monitored survival of yeast cells after prolonged periods of culture in liquid medium. To assay survival, cells were spotted onto fresh plates, which were then incubated further. After 28 days of culture at 30 °C, we were unable to identify a survival difference relative to wild type (Fig. 3B). However, after 60 days of culture at 30 °C, the spg5Δ mutant showed a decrease in survival when plated onto fresh media at 30 °C, which was more pronounced at 37 °C (Fig. 3B). Previous studies, which involved initially culturing cells at 37 °C, yielded a much quicker loss of viability; under these conditions, spg5Δ mutants remained viable after 9 days of growth but showed a 1000-fold loss in viability at day 16 (19). These findings suggest that induction of SPG5 during stationary phase is part of an adaptive response to quiescence. It is worth noting that many proteasome hypomorphic mutants show decreased viability at 37 °C, presumably due to increased demand on the ubiquitin-proteasome system at that temperature. Furthermore, these results could provide a physiologic basis for the lack of overt functional defects in proteasomes purified from the spg5Δ mutant (Fig. 2), because these proteasomes were purified after overnight culture (i.e. 16–20 h of growth). These cells show a doubling time of approximately 2 h (data not shown) and thus are in the logarithmic to late logarithmic phase of growth and are unlikely to exhibit the stationary phase induction of SPG5.

Spg5 Is Required for Proteasome Structure and Function in Quiescence

We next purified proteasomes from wild-type and spg5Δ cells that had been cultured for 6 days and thus were in true quiescence, with a doubling time that was so long as to be difficult to quantify. The profiles of the wild-type and mutant purified proteasomes were strikingly different (Fig. 4A). Upon nondenaturing gel electrophoresis followed by an in-gel LLVY activity assay, wild-type proteasomes appeared similar to those seen in conventional preparations (i.e. those prepared from logarithmic or late logarithmic cultures), showing predominantly RP₂CP and RP₁CP forms, with a small amount of free CP and Blm10-CP (see Fig. 2C). By contrast, proteasomes purified from spg5Δ cells showed only RP₂CP, with a near-complete absence of the other active species (Fig. 4A). Furthermore, despite having loaded an equivalent amount of purified material (see Fig. 4C), the enzymatic activity of RP₂CP was diminished in the spg5Δ mutant (Fig. 4A). Evaluation of this same gel with Coomassie staining confirmed these findings (Fig. 4B). Remarkably, when these same materials were analyzed by denaturing electrophoresis, the overall proteasome profiles were similar between the wild-type and the spg5Δ mutant (Fig. 4C), implying that the defects seen in the spg5Δ mutant are likely related to a failure of assembly or stability of the proteasome, rather than to the loss or gain of individual subunits in the spg5Δ proteasomes. Finally, we evaluated the enzymatic activity of the purified proteasomes in a solution phase LLVY hydrolysis assay. The proteasomes purified from the spg5Δ mutant showed a strong defect in the chymotrypsin-like activity of the CP as compared with those of wild type (Fig. 4D). This defect is unlikely to represent direct modulation of the chymotrypsin-like activity of the proteasome; rather this defect likely reflects broader structural and functional defects of the proteasome holoenzyme, as observed in Fig. 4B.

FIGURE 4.

spg5Δ proteasomes from stationary phase cells are defective. A, native gel electrophoresis coupled with in-gel suc-LLVY-AMC assay highlights proteolytically active proteasome species, as indicated, for wild-type (SJH303) and spg5Δ (SJH304) proteasomes. B, electrophoretic profile of the proteasome species, separated by native gel electrophoresis as in A, and visualized by Coomassie staining. C, denaturing gel electrophoresis of the proteasome species as in A, visualized by Coomassie staining. D, suc-LLVY-AMC hydrolyzing activity of proteasomes (3.5 μg) from A. Control, buffer only. E, native gel electrophoresis coupled with in-gel suc-LLVY-AMC assay shows that restoration of SPG5 expression corrects the proteasome defects of the spg5Δ mutant. Strains: wild type with empty vector (SJH386), spg5Δ with empty vector (SJH387), and spg5Δ with SPG5 CEN plasmid (SJH388).

To verify that that the proteasome defects seen in the spg5Δ mutant were truly due to loss of Spg5, we generated a low copy centromeric (CEN) plasmid that expresses SPG5 under the control of its own promoter, and we transformed proteasome-tagged wild-type and spg5Δ cells with either an empty vector or the SPG5-expressing plasmid. We then cultured cells for 10 days in minimal media to ensure expression of the plasmids and purified proteasomes. We again visualized the proteasome species by native gel electrophoresis followed by an in-gel LLVY hydrolysis assay. As in Fig. 4A, we detected a strong difference in the proteasome profiles between wild-type and spg5Δ proteasomes (Fig. 4E). Moreover, when we restored SPG5 expression to spg5Δ cells, the proteasome profile was comparable with that seen in wild-type cells (Fig. 4E). Specifically, expression of SPG5 restored the presence of RP₁CP and free CP bands. These results indicate that the proteasome defects seen in spg5Δ mutants specifically reflect loss of the Spg5 protein.

In addition to the major enzymatically active species we could detect (RP₂CP, RP₁CP, Blm10-CP, and CP), native gel analysis followed by Coomassie staining showed a small number of faint bands that might represent additional proteasome species. We therefore pursued immunoblotting to further analyze these species. Evaluation with an antibody to a core particle subunit (α7) showed a profile comparable with that of the LLVY-based in-gel activity assay (Fig. 5A). In contrast, when we examined Rpt4, a subunit of the base of the regulatory particle, we detected two new species predominantly present in the spg5Δ proteasomes, with near absence in those of wild type (Fig. 5A). These two species, by their composition and electrophoretic properties (27), appear to represent free RP and free base (Fig. 5A). Examination of a lid subunit (Rpn8) also identified the free RP, as expected, which again predominated in the mutant proteasomes (Fig. 5A). A second band was also detected, running just below that of the free base, which appears to represent the free lid (Fig. 5A). In contrast to the RP and the base, the free lid was present in both the wild-type and spg5Δ proteasome samples, and its increased abundance may be related to preferential purification given that the proteasome affinity tag is present on the lid subunit Rpn11.

FIGURE 5.

spg5Δ proteasomes from stationary phase cells show an altered proteasome subcomplex profile. A, wild-type and spg5Δ proteasomes as in Fig. 4, analyzed by native gel electrophoresis and visualized by suc-LLVY-AMC in-gel activity assay or immunoblotting with antibodies to the CP subunit α7, the base subunit Rpt4, and the lid subunit Rpn8, as indicated. B, denaturing gel electrophoresis of the proteasome species, visualized by Coomassie staining.

Spg5 Is Distinct from Known RP Chaperones

Our data identify Spg5 as novel proteasome-interacting protein with some features reminiscent of several of the recently described RP chaperones: Nas6, Hsm3, and Rpn14. Like those proteins, Spg5 binds to the RP and specifically the base but appears to be excluded from mature proteasomes. Transcriptional regulation of the genes encoding known RP chaperones has not previously been evaluated. Examination of Nas6, Hsm3, Rpn14, as well as a fourth RP chaperone, Nas2, showed no transcriptional induction in stationary phase, in contrast to SPG5 (Fig. 6). Thus, Spg5 is distinct from the known RP chaperones in its pattern of gene expression.

FIGURE 6.

Regulation of SPG5 expression is distinct from that of known RP chaperones. Quantitative real time PCR analysis of transcription of the indicated genes after 5 days of culture, expressed as fold-induction relative to exponential phase growth. The 18 S ribosomal RNA serves as an endogenous control. Error bars indicate standard deviations of experiments carried out in triplicate. RNA was generated from strain SUB62.

Accumulation of Oxidized Proteins in Stationary Phase in spg5Δ Mutants

To generate metabolic energy, yeast cells growing under conditions of nutrient limitation rely heavily on mitochondrial respiration, a process that generates reactive oxygen species. Oxidized proteins are subject to varying degrees of misfolding and dysfunction and are generally good substrates of the proteasome (20). We therefore investigated the status of oxidized proteins in spg5Δ mutants using a detection system that is specific for carbonyl modifications, which are generated by oxidation.

When cells were evaluated under conditions of logarithmic growth, the abundance of oxidized species in the spg5Δ mutant relative to wild type was comparable (Fig. 7A). However, in stationary phase, there was a marked increase in total oxidized proteins in the spg5Δ mutant relative to wild type. A Pgk1 loading control was similar under all conditions (Fig. 7A).

FIGURE 7.

Accumulation of oxidized proteins in stationary phase in spg5Δ mutants. A, total cellular abundance of oxidized proteins in wild-type (SJH92) and spg5Δ (SJH301) strains in logarithmic and stationary phase growth, as indicated. Pgk1, loading control. Comparable results were obtained in four independent experiments. B, total cellular abundance of oxidized protein in stationary phase. Strains: wild-type with empty vector (sJH370), spg5Δ with empty vector (sJH371), and spg5Δ complemented with a plasmid-borne SPG5 gene (sJH372). Pgk1, loading control.

To show that the increase in oxidized proteins was due to loss of Spg5 function, we again utilized the low copy CEN plasmid that expresses SPG5 under the control of its own promoter, and we transformed spg5Δ cells with either an empty vector or the SPG5-expressing plasmid. When cells were evaluated in stationary phase, we again observed an increase in oxidized proteins in the spg5Δ mutant transformed with an empty vector, and this decrease was reversed in the spg5Δ mutant complemented the SPG5 plasmid (Fig. 7B). Thus, spg5Δ cells show an accumulation of oxidized proteins during quiescence, further indicating that spg5Δ proteasomes are functionally defective in stationary phase.

DISCUSSION

Spg5 Is a Novel Component of the Ubiquitin-Proteasome System

The results presented here identify both a novel proteasome-interacting protein, Spg5, and an unexpected link between quiescence and the proteasome. Yeast cells grow logarithmically under nutrient-rich conditions, but as nutrients are depleted, the growth rate slows and eventually ceases, culminating in quiescence. Spg5 is progressively induced in stationary phase, and, reflecting the physiologic relevance of this induction, spg5Δ mutants show increasing loss of viability as stationary phase progresses. Spg5 binds the proteasome regulatory particle and its subcomplex, the base, but appears to be excluded from mature, fully assembled proteasomes. Proteasomes from spg5Δ mutants show defects in proteasome structure and function, which are supported by in vivo phenotypes that appear to reflect deficient proteasome function.

There are deep parallels between the states of quiescence experienced by yeast cells and those of higher organisms. In yeast, quiescence is characterized by a complete absence of growth, increasing reliance on oxidative metabolism, induction of autophagy, the selective repression of ribosomal genes, a strong decrease in protein synthesis, and the specific induction of quiescence-related genes (1, 6). Regulation of the proteasome may represent another important feature of quiescence. Although data are lacking from mammalian cells, studies in budding yeast as well as fission yeast, an organism considerably diverged from budding yeast, suggest a critical role as the proteasome is strictly required for cells to maintain viability in quiescence (21–23). Further supporting this idea, other studies highlight tight control over subcellular proteasome localization in quiescence (24), with nuclear proteasomes migrating to discrete cytoplasmic foci. The physiologic basis for the requirement of the proteasome in quiescence remains unclear, but one might speculate that it relates to the function of the proteasome in protein quality control or possibly a role in generating amino acids to serve as biosynthetic precursors under conditions of nutrient limitation (25). Indeed, there is evidence that pathways of protein synthesis and degradation might regulate one another under conditions of nutrient limitation (26).

An important question concerns the evolutionary conservation of Spg5. To date, a clear mammalian ortholog of Spg5 has not been identified by sequence-based algorithms, nor have we identified any functional motifs within SPG5 which might inform its cellular role.⁵ Nevertheless, a functional ortholog may exist as the challenges posed by quiescence in yeast are likely to be similar in higher organisms. Indeed, there is precedent for such functional conservation within the ubiquitin-proteasome system. The mammalian form of securin, an anaphase-promoting complex substrate that regulates mitotic spindle function, performs the same function as the budding yeast protein Esp1, but it shows no significant sequence homology (11).

Role of Spg5 in Proteasome Function

The precise role that Spg5 plays in proteasome function remains to be elucidated. At a minimum, the in vitro and in vivo data support a positive role for Spg5 in the ubiquitin-proteasome system. Loss of Spg5 results in defects in both the structure and function of the proteasome in vitro. At least two models could account for such findings. In the first model, Spg5 might be dispensable for initial biosynthetic proteasome assembly but necessary for the stability and integrity of the mature proteasome holoenzyme. That Spg5 fails to bind fully assembled proteasomes argues against but does not fully exclude this model. In a second model, Spg5 promotes the assembly of newly synthesized proteasomes, assuming that new proteasomes are in fact generated in deep stationary phase. Indeed, the binding pattern of Spg5 to the base and RP, but not the intact 26 S proteasome, is similar to that of most known RP chaperones. That notwithstanding, Spg5 is clearly distinguished from the known RP chaperones by virtue of its unique transcriptional regulation.

Despite the differences in transcriptional regulation between SPG5 and the genes encoding the four known RP chaperones, there are suggestions of a close relationship at a mechanistic level. Like the RP chaperones, Spg5 binds the base but does not do so when the base is incorporated into the proteasome holoenzyme. In the case of the RP chaperones, this specificity is thought to be related to their binding proximally to the RP-CP interface (27–30), which puts them in competition with CP for RP binding. The CP-proximal binding sites of the chaperones are the C-terminal domains (known as C-domains) of the proteasomal ATPases (Rpt1–Rpt6) (27, 31). Whether Spg5 binds similarly to the C-domains is currently under study. A second suggestion of similarity between Spg5 and the chaperones is that an accumulation of proteasome subassemblies can be observed in both types of mutant. As chaperones function to assist in the generation of mature protein complex, this phenotype is consistent with a chaperone function for Spg5. However, further work is required to define the nature of this defect in detail. For example, it is unclear whether de novo proteasome formation is affected in the mutant over prolonged periods of quiescence. Finally, although spg5Δ mutant proteasomes are markedly unstable, it is unclear to what extent the observed defects represent in vivo or in vitro instability.

Cellular Regulation of Spg5

Transcriptional induction of SPG5 is best understood as an adaptive response to the cellular conditions associated with quiescence. Yet why a specific adaptation of the ubiquitin-proteasome system has evolved in quiescence remains unclear. Central to this uncertainty is the fact that the proteasome and protein quality control have not been well studied under these conditions. In one model, the stresses associated with stationary phase growth might create an increased or unique set of substrates, such as oxidized proteins, requiring modification of the proteasome. In another model, the stresses associated with stationary phase might compromise the proteasome itself.

The transcriptional machinery that orchestrates SPG5 induction remains to be discovered. Equally important will be an exploration of how the need for SPG5 induction is recognized by the cell. A number of cellular alterations associated with quiescence, most notably carbon depletion, nitrogen depletion, increasing oxidation, and increasing acidity of the culture medium, are potential candidate inputs. Available data indicate that induction of SPG5 occurs in response to multiple nutrient deficiencies, including glucose, nitrogen, and phosphorus deprivation, with perhaps a more robust response in the context of glucose starvation (32). The full delineation of the stimuli resulting in SPG5 induction will be key to understanding how cells coordinate proteasome function in quiescent states.

The present results provide strong evidence that the proteasome is subject to growth phase-dependent regulation, similarly to protein biosynthetic pathways. Indeed, the concept of protein homeostasis, which involves not only the elimination of aberrant or dysfunctional proteins but an overall balance between protein synthesis and degradation, implies just this sort of proteasomal regulation. In recent years, our understanding of the complexity of the pathways integrating growth and protein synthesis has expanded tremendously. The results presented here provide a framework to help elucidate what may be an important parallel complex pathway integrating cell growth and proteasome function.