Sts1 Can Overcome the Loss of Rad23 and Rpn10 and Represents a Novel Regulator of the Ubiquitin/Proteasome Pathway*

A rad23Δ rpn10Δ double mutant accumulates multi-Ub proteins, is deficient in proteolysis, and displays sensitivity to drugs that generate damaged proteins. Overexpression of Sts1 restored normal growth in rad23Δ rpn10Δ but did not overcome the DNA repair defect of rad23Δ. To understand the nature of Sts1 suppression, we characterized sts1-2, a temperature-sensitive mutant. We determined that sts1-2 was sensitive to translation inhibitors, accumulated high levels of multi-Ub proteins, and caused stabilization of proteolytic substrates. Additionally, ubiquitinated proteins that were detected in proteasomes were inefficiently cleared in sts1-2. Despite these proteolytic defects, overall proteasome activity was increased in sts1-2. We propose that Sts1 is a new regulatory factor in the ubiquitin/proteasome pathway that controls the turnover of proteasome substrates.

A major pathway of intracellular protein degradation involves the attachment of a multi-ubiquitin (multi-Ub) 3 chain to proteolytic substrates. Ubiquitinated proteins are degraded by the multicatalytic proteasome (8). The mechanism of substrate recognition, ubiquitination, and proteasome-specific hydrolysis is well described. However, the events that promote the recognition of multi-Ub substrates by the proteasome are not well defined. For instance, it is unclear whether multi-Ub substrates form unaided affinity-based interactions with the proteasome in vivo or whether regulatory proteins deliver substrates to the proteasome. It is also unclear whether compositionally distinct proteasomes promote the degradation of unique classes of substrates.
Rad23 and Rpn10 are multi-Ub chain-binding proteins that play an important role in the degradation of substrates by the proteasome (1)(2)(3)5). Although a significant proportion of purified proteasomes contains Rpn10 (9), only a small fraction contains Rad23 (10). In an effort to understand the role of Rad23, we generated a yeast strain lacking both Rad23 and Rpn10 and discovered that the double mutant is severely growth impaired and displays significant proteolytic defects (3). Rad23 contains two autonomous sequences: UbL, which binds the proteasome, and UBA domains that interact with multiubiquitin chains. Based on its ability to bind both ubiquitinated proteins and the proteasome, we proposed that Rad23 translocated proteolytic substrates to the proteasome (2). In this model, Rpn10 would represent an ideal proteasome receptor for Rad23-specific substrates. We also determined that expression of high levels of the UbL domain caused stabilization of a reporter protein (2). A simple interpretation of this result is that UbL/proteasome binding interferes with the normal docking of native Rad23 with the proteasome, consistent with a shuttle-factor model. In contrast, expression of a Rad23 derivative lacking the UbL domain, but containing UBA sequences, retains efficient interaction with multiubiquitinated proteins but not the proteasome (2). We conclude from these findings that Rad23 binds multiubiquitin proteins, before it interacts with the proteasome. These two activities are functionally linked because single amino acid mutations in either UbL or UBA domains cause a proteolytic defect. Further support for the shuttle-factor model was described recently (11). It has also been proposed that Rad23 and Rpn10 function as alternate multi-Ub chainbinding receptors in the proteasome (4,7).
To understand the proteolytic defect of rad23⌬ rpn10⌬, we sought high copy suppressors of its cold temperature growth defect. We isolated Sts1, a poorly characterized protein that lacks distinctive motifs that might offer insight into a biochemical function. However, previous studies showed that Sts1 is required for efficient chromosome segregation (12), maintaining rRNA stability, and suppressing an ER/Golgi secretion defect (13). Moreover, a mutation in STS1 suppressed the mRNA processing defect of an rna15 mutant (14). Sts1 also interacts with proteasome subunit Rpn11, and Srp1, a nuclear localization signal-binding protein (15). The effects of these seemingly unrelated genetic and biochemical interactions have not been resolved. However, the finding that high level expression of Sts1 restored the degradation of a proteolytic substrate in an srp1 mutant suggested an unexpected role in proteolysis. Additionally, a distantly related protein in Schizosaccharomyces pombe (Cut8) regulates the subcellular trafficking of proteasomes (16,17), providing further evidence of a likely proteolytic function for Sts1.
In this report, we show that overexpression of Sts1 suppressed the defects of rad23⌬ rpn10⌬. However, the DNA repair defect associated with rad23⌬ was not suppressed, and sts1-2 is not sensitive to UV light. (We note that this is contrasted by the DNA damage sensitivity of an S. pombe cut8 mutant). We demonstrate that sts1-2 mutant accumulated high levels of multi-Ub proteins, displayed heightened sensitivity to drugs that generate damaged proteins, and failed to degrade test substrates of the ubiquitin/proteasome system. A physiological substrate (Sic1) is also stabilized in sts1-2. We also determined that multiubiquitinated proteins could bind the proteasome in sts1-2 but were inefficiently degraded. These compelling results indicate that Sts1 encodes a previously unrecognized factor that influences degradation by the ubiquitin/proteasome pathway.

EXPERIMENTAL PROCEDURES
Yeast Strains and Plasmids-Yeast rad23⌬ rpn10⌬ was transformed with a galactose-inducible yeast cDNA library (18), and STS1 was isolated as a high copy suppressor of the growth defect at 13°C. FLAG-STS1 and FLAG-RPN10 were generated by PCR, and expression was induced from the P CUP1 promoter by the addition of 100 M CuSO 4 to the growth medium. SIC1 was amplified from yeast genomic DNA and cloned into YEplac181, containing two tandem hemagglutinin (HA) epitopes. The gene was expressed from the P CUP1 promoter. DNA encoding sts1-2 was amplified from NA25 (sts1-2 ts ; Ref. 13), and all constructs were verified by DNA sequencing.
Growth Assays and Sensitivity to Translation Inhibitors-Yeast cells were grown in selective medium (ϳ18 h), transferred to fresh medium, and grown to the exponential phase. The cultures were normalized to a density of A 600 ϳ 1.0, and 10-fold serial dilutions were spotted on synthetic medium containing 2% glucose or galactose, and 1 g/ml L-canavanine, 1 mM paromomycin, or 1 mM neomycin. The plates were incubated at 13, 24, 30, or 37°C for 3-4 days.
UV Survival Assays-The cells were grown in YPD to midlog phase (ϳ10 7 cells/ml), washed with sterile water, and resuspended at A 600 of 1. Ten-fold serial dilutions were spread on YPD plates and irradiated at one J/m 2 /s of predominantly 254-nm UV light, using an American Ultraviolet Co. germicidal lamp. Following irradiation, the plates were wrapped with aluminum foil and incubated at 30°C until colonies appeared (typically 3-4 days, depending on the genetic background). The number of colonies was counted, and the average of duplicate experiments was plotted.
Preparation of Protein Extracts-Yeast cultures were suspended in 250 l of lysis buffer (50 mM HEPES, pH 7.5, 150 mM NaCl, 5 mM EDTA, 1% Triton X-100) containing protease inhibitors (Roche Applied Science) and lysed by disruption with glass beads. The protein extracts were centrifuged at 14,000 ϫ g for 3 min at 4°C, and protein concentration was determined using the Bradford assay (Bio-Rad).
Immunoprecipitation and Immunoblotting-One milligram of total protein was adjusted to 1 ml using lysis buffer and combined with 20 l anti-FLAG M2-agarose beads (Sigma). The samples were incubated for 2 h at 4°C, and FLAG-tagged proteins were immunoprecipitated. The immobilized proteins were separated in SDS/PAGE, transferred to nitrocellulose, and incubated with antibodies. The antibodies used were as follows: anti-ubiquitin and anti-FLAG (Sigma), anti-HA (Roche Applied Science), and anti-␤-gal (Promega). We generated anti-Rpt1 antibodies using recombinant glutathione S-transferase-Rpt1 (Pocono Rabbits). Antibodies against Pab1 were a gift from S. Peltz (University of Medicine and Dentistry of New Jersey). Antibodies against Rpn10 and Rpn12 were generously provided by D. Skowyra (St. Louis University). The immunoblotting reactions were detected with Chemi-luminescence Reagent Plus (Perkin-Elmer) and quantified using Kodak ID-3.5 imaging software.
Pulse-Chase Measurement of Protein Stability-Protein stability measurements were described previously (19), using the EXPRE 35 S 35 S protein labeling reagent (Perkin-Elmer). Fifty milliliter cultures of exponentially growing cells were pelleted and resuspended in 0.4 ml of labeling buffer (50 mM sodium phosphate, pH 7.0, 2% glucose). EXPRE 35 S 35 S protein labeling mix (0.5 mCi) was added, and the suspension was incubated for 5 min at 30°C. The cells were washed with 1 ml of water and resuspended in 0.4 ml of chase buffer (YPD-glucose medium containing unlabeled L-methionine and L-cysteine, and 0.5 mg/ml cycloheximide). An aliquot (0.1 ml) of the suspension was immediately withdrawn and frozen in liquid nitrogen. The rest of the suspension was incubated at 30°C, and 0.1-ml aliquots were withdrawn at the indicated times and frozen. Trichloroacetic acid precipitation and scintillation counting estimated incorporation of 35 S into acid-insoluble proteins. Lysates containing equal 35 S cpm were adjusted to equal volume and combined with anti-␤-gal antibodies to immunoprecipitate ␤-gal proteolytic substrates (20). Autoradiograms were quantified using Kodak imaging software.
To examine Sic1-HA levels, we arrested growth of STS1 and sts1-2 in the G 1 phase by growth into stationary phase. Yeast cells were then suspended for 30 min in medium containing 100 M copper sulfate to induce expression of Sic1-HA. The cells were pelleted, washed, and resuspended in fresh medium containing 15 g/ml nocodozole to permit entry into mitotic growth but cause arrest in the G 2 phase of the cell cycle. Aliquots of the cultures were withdrawn periodically and examined microscopically to gauge release from G 1 arrest and to prepare protein extract. Equal amounts of protein were resolved by SDS/PAGE, and immunoblots were incubated with antibodies against the HA epitope and anti-Rad23 antibodies.
Proteasome Activity-Total protein extract was prepared from STS1 and sts1-2 that were grown at either 23 or 37°C (for 2 h) and incubated with the fluorogenic substrate SUC-Leu-Leu-Val-Tyr-AMC (Boston Biochem) to measure chymotryptic activity. Following incubation at 30°C for 1 h, fluorescence units were determined using a Turner 700 fluorometer. Control reactions containing the proteasome inhibitor epoxomicin allowed us to specifically quantify proteasome-hydrolytic activity.
Microscopy-Yeast strains were grown in synthetic medium and fixed in 5% ethanol, and the images were captured using a Zeiss microscope.

Sts1
Is a Dosage Suppressor of rad23⌬ rpn10⌬-Loss of either Rad23 or Rpn10 does not cause a significant growth defect.
However, loss of both proteins caused pleiotropic defects (3) that were suppressed by overexpressing Sts1 (P GAL1 ::STS1). Normal growth was restored in rad23⌬ rpn10⌬ at 13, 23, and 30°C (Fig. 1A). Significantly, the growth improvement caused by overexpressing Sts1 was comparable with that observed by expressing Rad23 in rad23⌬ rpn10⌬. High levels of Sts1 also restored growth of rad23⌬ rpn10⌬ in medium containing canavanine, neomycin or paromomycin (Fig. 1B), which cause increased levels of damaged proteins. Because defects in the ubiquitin/proteasome pathway often cause sensitivity to translation inhibitors (21,22), these results suggest that Sts1 functions in protein degradation.
Rad23 is required for efficient nucleotide excision repair (23), because rad23⌬ is sensitive to UV-induced DNA damage. However, the UV sensitivity of rad23⌬ is not exacerbated in rad23⌬ rpn10⌬ (3), indicating that the DNA repair functions of Rad23 do not overlap with its proteolytic activities with Rpn10. It is significant in this regard that Sts1 suppressed only the growth and proteolytic functions of Rad23 and not the UV sensitivity of rad23⌬ rpn10⌬ (Fig. 1C). rad23⌬ rpn10⌬ cells display a large, multi-budded phenotype and are noticeably delayed in the G 2 phase of the cell cycle (3), consistent with a failure to properly execute cell division. Overexpression of Sts1 eliminated the aberrant morphology of rad23⌬ rpn10⌬ cells and restored normal growth (Fig. 1D).
We examined the stability of FLAG-Sts1 by pulse-chase methods, and determined that the protein is rapidly degraded, with an in vivo half-life of only a few minutes (Fig. 1E). Surpris-ingly, the stability of FLAG-Sts1 was further reduced in rad23⌬ rpn10⌬ but was stabilized in proteasome mutant pre1-1 pre2-2. The sts1-2 mutant protein was much more unstable and not detected in our stability measurements. These findings suggested that the dosage suppression of rad23⌬ rpn10⌬ might be related to restoring normal levels of Sts1. The ability of Sts1 to suppress the aberrant phenotypes of rad23⌬ rpn10⌬ suggests that it might perform a regulatory role in the ubiquitin/proteasome pathway. To understand the mechanism of suppression of rad23⌬ rpn10⌬, we investigated the biochemical properties of Sts1.
Sts1 Does Not Bind Multiubiquitinated Proteins-Proteins bearing structural similarity to Rad23 have been described (6, 24 -26), and several, including yeast Dsk2 (27) and Ddi1 (28), perform overlapping roles with Rad23. Moreover, these proteins can bind both multi-Ub proteins and proteasome. Because Rad23 and Rpn10 also bind both multi-Ub chains and proteasomes, we speculated that a protein with similar binding properties might suppress rad23⌬ rpn10⌬. Therefore, we investigated whether Sts1 possessed either of these biochemical properties of Rad23 and Rpn10.
Protein extracts were prepared from wild type cells expressing FLAG-Sts1, FLAG-Rad23, FLAG-Rpn10, or Pre1-FLAG, and applied to FLAG-agarose beads ( Fig. 2A). The immunoprecipitated proteins were separated in SDS-PAGE and transferred to nitrocellulose. We note that the levels of FLAG-Rad23 and FLAG-Sts1 are quite similar. The immunoblot was incubated with antibodies against ubiquitin ( Fig. 2A), and as expected, high levels of Ub cross-reacting material were copurified with FLAG-Rad23 (lane 6; FLAG-IP) and FLAG-Rpn10 (lane 7). Lower levels of ubiquitinated protein were also co-purified with proteasomes (lane 5; Pre1-FLAG). However, no significant interaction with multi-Ub proteins was detected with FLAG-Sts1 (lane 8), demonstrating that, unlike Rad23 and Rpn10, it does not bind ubiquitin. The prominent ϳ70-kDa band detected in lane 8 is a nonspecific interaction and is also seen in the control immunoprecipitation from an extract lacking a FLAG-tagged protein (Ctrl). The expression of the FLAGtagged proteins is shown in the lower panels following incubation of the nitrocellulose filter with anti-FLAG antibodies (left) and after staining the filter with Ponceau S (right). These findings suggest that Sts1 is not a substrate shuttle-factor or receptor for multi-Ub proteins. The expression of high levels of Sts1 resulted in a small increase in the overall levels of high molecular weight multi-Ub proteins in a wild type strain (Extracts; FIGURE 1. Sts1 can suppress rad23⌬ rpn10⌬. A, yeast cultures were grown to exponential phase and adjusted to equal density, and 10-fold dilutions were spotted on agar medium containing galactose. The plates were incubated at the temperatures indicated. B, The same cultures were also examined on similar medium supplemented with canavanine, neomycin, and paromomycin. C, Wild type (WT), rad23⌬ rpn10⌬, and rad23⌬ rpn10⌬ containing a plasmid expressing P GAL1 ::STS1 were adjusted to A 600 ϭ 1, and ϳ1 ϫ 10 6 to 1 ϫ 10 3 cells were plated on medium containing galactose. The plates were exposed to 254-nm UV light, wrapped in foil, and incubated at 30°C. The averaged values from two independent experiments were plotted. D, yeast strains indicated were grown in galactose medium, and cell morphology was examined by microscopy. The aberrant morphology of rad23⌬ rpn10⌬ is evident in the right panels. E, the stability of 35 S FLAG-Sts1 was determined by pulse-chase methods in the strains indicated. The chase time points are in min. Significant stabilization was observed in the proteasome mutant pre1-1 pre2-2. 1 and 4). In contrast, overexpression of Rad23 resulted in a highly significant increase in the levels of ubiquitinated proteins (lane 2). The implication of these effects is unclear.

Sts1 Functions in the Ubiquitin/Proteasome System
Sts1 Can Bind the Proteasome-Because Sts1/sts1-2 proteins are rapidly degraded (Fig. 1E), we surmised that their expression in a proteasome mutant would ease their biochemical characterization. Therefore, to determine whether Sts1 interacted with the proteasome, we expressed either FLAG-Sts1 or FLAG-sts1-2 in pre1-1 pre2-2 (Fig. 2B). Protein extracts were prepared and incubated with FLAG-agarose. Immunoblotting showed that a 19 S proteasome subunit, Rpt1, was co-purified with FLAG-Sts1 and FLAG-sts1-2 from cells grown at 23 and 37°C ( Fig. 2B; FLAG-IP). The same filter was also incubated with antibodies against Rpn12, and a similar interaction was observed. The expression of Rpt1 and Rpn12 was similar in STS1 and sts1-2. These studies suggest that the defect of the sts1-2 mutant protein is not caused by deficient interaction with the proteasome or altered proteasome abundance. Although the interaction between Sts1 and the proteasome is weak, we note that in a control reaction (Ctrl), we did not detect nonspecific purification of proteasome subunits on FLAG-agarose.
Stabilization of Proteolytic Substrates-The interaction between Sts1 and the proteasome led us to investigate whether it affected the degradation of proteolytic substrates. We transformed STS1 and sts1-2 with plasmids expressing the engineered substrates Arg-␤-gal and Ub-Pro-␤-gal. We used pulsechase methods to compare the stabilities of these substrates to Met-␤-gal, a stable reporter protein. Pioneering studies by Varshavsky and co-workers (20, 29) using these reagents uncovered the N-end rule pathway of substrate targeting, as well as the ubiquitin fusion degradation pathway. Yeast cells were grown to exponential phase at the semi-permissive temperature (30°C) and then incubated in medium containing [ 35 S]methionine. Following 5 min of labeling, aliquots were withdrawn, and protein extracts were examined, as described under "Experimental Procedures." As expected, Ub-Pro-␤-gal was rapidly degraded in a wild type strain. However, Ub-Pro-␤-gal was strongly stabilized in sts1-2 (Fig. 3A). The band intensities were quantified, and we determined that the half-life of Ub-Pro-␤-gal over the initial 10 min of chase was ϳ6 min in the wild type strain and ϳ30 min in sts1-2. Arg-␤-gal is an N-end rule substrate and was found to be similarly stabilized in sts1-2 (Fig. 3B). In contrast, Met-␤-gal was stable in both STS1 and sts1-2. Collectively, these studies showed that substrates of the proteasome are inefficiently degraded in sts1-2.
To further investigate a role for Sts1 in protein degradation, we examined the stability of a physiological substrate, Sic1-HA. Sic1 is a key regulator of entry into the cell cycle, and its degradation by the ubiquitin/proteasome pathway is tightly regulated. We generated Sic1-HA and expressed the protein in STS1 and sts1-2 using the copper-inducible P CUP1 promoter. Because sts1-2 showed no growth defect at low temperature, we grew STS1 and sts1-2 at 23°C (Fig. 3C) to increase the efficiency of G 1 arrest. We confirmed that ϳ90% of the cells were arrested in G 1 following nutrient deprivation. To increase the level of Sic1-HA, the G 1 arrested cells were incubated for 30 min with 100 M CuSO 4 . To stimulate reentry of these G 1 arrested cells containing higher levels of Sic1-HA into mitotic growth, they were transferred to fresh prewarmed medium (30°C) containing 15 g/ml nocodozole. This approach permitted cells to exit the G 1 FIGURE 2. Sts1 binds the proteasome, but not multiubiquitinated proteins. A, a wild type strain was transformed with plasmids expressing Pre1-FLAG (lanes 1 and 5), FLAG-Rad23 (lanes 2 and 6), FLAG-Rpn10 (lanes 3 and 7), or FLAG-Sts1 (lanes 4 and 8). Equal amounts of total protein extract were prepared from these strains, as well as a control (Ctrl) strain lacking a FLAGtagged protein. Protein extracts were examined by immunoblotting to determine the levels of multiubiquitinated proteins (lanes 1-4). Similarly, equal amounts of extract were applied to FLAG-agarose to immunoprecipitate the FLAG-tagged proteins and measure the interaction with multiubiquitinated proteins (lanes 5-8). The nitrocellulose filter described in A above was stained with Ponceau S (lower right panel) and probed with antibodies against FLAG (lower left panel) to determine the expression level and purification of the FLAG-tagged proteins. (FLAG-Rad23 migrates close to the immunoglobulin heavy chain and is therefore not readily detected by Ponceau S staining.) B, because of the rapid degradation of Sts1 and sts1-2, we expressed both FLAGtagged proteins in the proteasome mutant pre1-1 pre2-2. Protein extracts were prepared from cells incubated at either 23°C (lanes 1 and 2) or 37°C (lanes 3 and 4), incubated with FLAG-agarose, and examined by immunoblotting with antibodies against proteasome subunits Rpt1 and Rpn12. The expression of these proteasome subunits was not affected significantly by temperature (Extract), or by expression of FLAG-Sts1 (lanes 1 and 3), or FLAG-sts1-2 (lanes 2 and 4). Furthermore, similar amounts of Rpt1 and Rpn12 were co-precipitated with FLAG-Sts1 and FLAG-sts1-2 (FLAG-IP). Ponceau S staining (lower panel) showed equivalent expression levels of the FLAG-tagged proteins. Asterisks indicate the positions of the immunoglobulin heavy and light chains. mw, molecular weight. phase and cease growth in G 2 and provided a way to monitor the degradation of Sic1-HA within a single cell cycle.
The studies described in Fig. 3C showed that the rapid elimination of Sic1-HA following exit from G 1 (in both STS1 and sts1-2) under semi-permissive conditions (30°C) obscured any proteolytic defect of sts1-2. Therefore, in the study described in Fig. 3D, we grew STS1 and sts1-2 to a stationary phase at 30°C to cause partial defect in sts1-2. The cultures were then transferred to 37°C and Sic1-HA levels examined. In contrast to the study described in Fig. 3C, Sic1-HA levels remained elevated in sts1-2 at 37°C for at least 60 min. Sic1-HA was detectable for up to 240 min (Fig. 3D). The filters were subsequently incubated with antibodies against Rad23, and similar levels were detected in all of the samples that were examined. We are aware that the stabilization of Sic1-HA might be a consequence of a cell cycle progression delay in sts1-2, and further study will be required to resolve this possibility. Nonetheless, the stabilization of ␤-gal substrates (as well as Sic1-HA) is consistent with our hypothesis that Sts1 is a previously uncharacterized regulatory factor in the ubiquitin/proteasome system. Because Sts1 affects the stability of substrates of distinct targeting mechanisms, we propose that it functions after proteolytic substrates have been targeted by the ubiquitin-conjugating apparatus.
We confirmed the temperature-sensitive growth defect of sts1-2 at 37°C (Fig. 4A, lower left panel) and verified that this defect could be overcome by expressing FLAG-Sts1 from a plasmid (Fig. 4A, lower right panel). In contrast, growth of sts1-2 at 37°C was not restored by the FLAG-sts1-2 mutant protein (lower right panel).
Sensitivity to Translation Inhibitors-Mutations in the Ub/proteasome pathway frequently result in poor growth on medium containing translation inhibitors. One interpretation of this observation is that the Ub/proteasome pathway is required for the degradation of nascent, misfolded proteins that are generated by the ribosome. To determine whether Sts1 played a general role in the Ub/proteasome system, we examined growth in the presence of drugs that induce protein damage. 10-fold serial dilutions of exponential phase cultures of FIGURE 3. sts1-2 is unable to degrade substrates of the ubiquitin/proteasome system. A, the proteolytic substrate Ub-Pro-␤-gal was expressed in STS1 and sts1-2 at 30°C, and its stability determined by pulse-chase methods. 35 S-Labeled Ub-Pro-␤-gal was immunoprecipitated, and its abundance was determined at 0, 10, 30, and 60 min after translation was inhibited with cycloheximide. The position of 14 C-labeled molecular weight markers is shown. B, the stabilities of Arg-␤-gal and the stable protein and Met-␤-gal were also determined as described above. An arrowhead indicates the position of the ␤-gal substrate, whereas an asterisk indicates the position of a previously described 90-kDa degradation product. C, the expression of Sic1-HA at 23°C in STS1 and sts1-2 was induced in stationary phase by the addition of 100 M CuSO 4 to the growth medium. Yeast cells were then resuspended in fresh medium containing 15 g/ml nocodozole, and the samples were withdrawn at the times indicated and examined by immunoblotting. The nitrocellulose filter was incubated with antibodies against HA (upper panel) and Rad23 (lower panel). D, the expression of Sic1-HA in STS1 and sts1-2 was also examined at 37°C, as described in C above.

FIGURE 4. sts1-2 is sensitive to translation inhibitors but not DNA damage.
A, the temperature-sensitive growth defect of sts1-2 was compared with STS1 at 23 and 37°C (left panel). The sts1-2 strain was transformed with 2-micron plasmids expressing either FLAG-Sts1 or FLAG-sts1-2, from the copperinducible P CUP1 promoter. Yeast strains were plated on selective medium and incubated for 3 days at 23 and 37°C. FLAG-Sts1 but not FLAG-sts1-2 suppressed the temperature-sensitive growth defect of sts1-2. B, The growth of sts1-2 was examined on medium containing cycloheximide. 10-fold dilutions were spotted, and growth was examined following 5 days of growth at 23°C. C, to determine whether Rad23 and Rpn10 performed functionally overlapping roles with Sts1, we simultaneously overexpressed Rad23 and Rpn10 in sts1-2 and examined the growth defect at 37°C. Precultures were grown in glucose containing medium (uninduced) and 10-fold dilutions were spotted on agar medium containing either glucose or galactose (induced) and incubated at 23°C and 37°C for 4 days. D, the UV sensitivity of sts1-2 was compared with STS1 at 23°C (upper panel) and after preincubation at 37°C for 2 h (lower panel). The average values from two independent experiments were plotted.

Sts1 Functions in the Ubiquitin/Proteasome System
STS1 and sts1-2 were grown (at 23°C) on medium containing cycloheximide or paromomycin (Fig. 4B). We found that sts1-2 is exceedingly sensitive to cycloheximide and also displayed poor growth on medium containing paromomycin (data not shown). We speculate that this defect will be more severe at the semi-permissive temperature.
Sts1 Cannot be Functionally Replaced by Rad23 and Rpn10-To determine whether Sts1 and Rad23/Rpn10 operated in a functionally redundant manner, we simultaneously overexpressed both P GAL1 ::RAD23 and P GAL1 ::RPN10 in the temperature-sensitive sts1-2 mutant (14). However, the growth defect of sts1-2 at 37°C was not suppressed (Fig. 4C), indicating that Sts1 cannot be replaced by either Rad23 or Rpn10. This finding also suggests that the mechanism of suppression achieved by Sts1 occurred by an alternate mechanism.
Sts1 Does Not Contribute to the DNA Repair Function of Rad23-Because high level expression of Sts1 did not overcome the UV sensitivity or rad23⌬ rpn10⌬ (Fig. 1C), we anticipated that it would not be required for the cellular response to DNA damage. To verify this conjecture, we measured the sensitivity of sts1-2 to 254-nm UV light, at both permissive and nonpermissive temperatures. Wild type and sts1-2 cells were grown at 23°C, incubated for 2 h at either 23 or 37°C, plated on selective agar medium, and exposed to UV light (at 1 J/m 2 /s). (The viability of sts1-2 is not affected by incubation at 37°C for this duration.) UV-irradiated cells were allowed to recover in the dark at 23°C. We found that the survival of sts1-2 was indistinguishable from the wild type strain following UV light treatment (Fig. 4D), demonstrating that Sts1 contributes to the growth and proteolytic functions of Rad23 but not its DNA repair activities. We recognize, however, that because STS1 is essential for viability, its apparent dispensability in DNA repair cannot be verified unambiguously.
Increased Levels of Multi-Ub Proteins in sts1-2-The sensitivity of sts1-2 to translational drugs, and stabilization of proteolytic substrates are consistent with a role in protein degradation. To investigate whether the targeting of proteolytic substrates was altered in sts1-2, we examined the levels of total multi-Ub proteins. Wild type and sts1-2 cells were pregrown at 23°C and then transferred to 37°C. Aliquots were withdrawn periodically, and total proteins were separated in SDS/PAGE and transferred to nitrocellulose. An immunoblot was incubated with antibodies against Ub. At 23°C the levels of multi-Ub proteins in sts1-2 were similar to the wild type strain (Fig. 5A, 0 h, lanes 1 and 2). However, after transfer to 37°C the levels of high molecular weight multi-Ub proteins increased progressively in sts1-2 (even-numbered lanes marked ts). The increased levels of multi-Ub proteins in sts1-2 is evident by 1 h (lane 4), and marked elevation was observed by 4 h (lane 8). In contrast, the levels of multi-Ub proteins in STS1 decreased in the corresponding time points at 37°C (odd-numbered lanes marked WT). Note for instance that at 6 h, only lower molecular weight ubiquitinated species were detected in STS1 (lane 9). We also investigated whether normal levels of multi-Ub proteins would be restored if cells were returned to 23°C following 6 h of incubation at 37°C. We found that ubiquitin levels remained high in sts1-2 following transfer to 23°C for 2 h ( lanes  13 and 14), although a perceptible increase in the amount of higher molecular weight species was observed during this recovery period (compare lanes 10 and 14). Prolonged incubation of the cultures at 23°C (untreated; lanes 11 and 12) had no effect on the levels of multi-Ub proteins, suggesting that the accumulation of ubiquitinated proteins at 37°C is related to the temperature-sensitive growth defect of sts1-2. The same filter  1 and 2) and then shifted to 37°C (lanes 3-10). Aliquots were removed at the times indicated (in h), and equal amounts of proteins were examined by immunoblotting to estimate the levels of multiubiquitinated proteins. The levels of multiubiquitinated proteins in STS1 and sts1-2 were similar before transfer to 37°C (lanes 1 and 2) and also after incubation of the culture at 23°C for 6 h (lanes 11 and 12). However, upon transfer to 37°C, the levels of overall multiubiquitinated proteins increased progressively in sts1-2 (even lanes labeled ts), in contrast to decreasing levels in STS1 (odd lanes labeled WT). After incubation at 37°C for 6 h, the cultures were returned to 23°C for 2 h. However, the high accumulation of multiubiquitinated proteins in sts1-2 was not restored to normal levels (lanes 13 and 14). The same filter was incubated with antibodies against Pab1, for a loading control (Ctrl). Pair-wise comparison indicated equal loading of protein extracts. Prestained molecular weight (mw) standards are shown in the left lane. B, the chymotryptic activity of the proteasome was measured using the fluorogenic substrate Suc-LLVY-AMC, using extracts prepared from STS1 and sts1-2 at both 23 and 37°C. The values represent the means from four independent measurements (p Ͻ 0.01). C, to verify that the composition of the proteasome was not significantly affected in sts1-2, proteasomes were purified using an epitope-tagged subunit in the 20 S particle (Pre1-FLAG). Following immunoprecipitation on FLAG-agarose, an immunoblot was incubated with antibodies against three subunits in the 19 S regulatory particle (Rpt1, Rpn12, and Rpn10). Equivalent levels of these 19 S subunits were copurified with Pre1-FLAG from STS1 and sts1-2 at both 23 and 37°C. DECEMBER 7, 2007 • VOLUME 282 • NUMBER 49 was probed with antibodies against Pab1 (loading control), and similar levels were detected in each set of paired extracts (Fig.  5A, compare WT to ts at each time point).

Sts1 Functions in the Ubiquitin/Proteasome System
Increased Proteasome Activity in sts1-2-Because both Sts1 and sts1-2 proteins formed equivalent interactions with the proteasome, we questioned whether the underlying defect of sts1-2 was linked to proteasome function. For instance, altered proteasome peptidase activity or reduced binding to multi-Ub proteins could underlie the defects of sts1-2. To address this question, we examined the hydrolytic activity of proteasomes. We prepared total protein extracts from yeast cells grown at 23°C. Proteasome activity was determined using a fluorogenic substrate (SUC-LLVY-AMC), which measures chymotryptic activity. We detected Ͼ2-fold increased activity in sts1-2 (**, p Ͻ 0.05) that was grown at 23°C (Fig. 5B). Because the abundance of proteasomes was similar in STS1 and sts1-2 at 23°C (Fig. 2B, lanes 1 and 3, and see below), the increased chymotryptic levels in sts1-2 might reflect a compensatory response to a defect in protein degradation. For instance, higher chymotryptic activity might enable the removal of an increased level of substrates in sts1-2. However, this process might be inefficient because Ub-Pro-␤-gal was stabilized in sts1-2 (Fig. 3A). Moreover, the inability to further increase proteasome function at 37°C might be related to the growth deficiency of sts1-2 at high temperature.
Proteasome Composition and Interaction with Sts1 Is Not Affected in sts1-2-Based on the accumulation of multi-Ub proteins, as well as higher proteasome activity in sts1-2, we questioned whether proteasome integrity was affected in the mutant strain. We purified proteasomes (using Pre1-FLAG) from STS1 and sts1-2 at both 23 and 37°C and examined the interaction between the 20 and 19 S particles. The immunoprecipitates were characterized by immunoblotting, using antibodies against the 19 S subunits Rpt1, Rpn12, and Rpn10. Rpt1 is present in the base complex, whereas Rpn12 is present in the lid complex. Previous studies suggested that Rpn10 functioned as a hinge that linked the lid to the base (30). All three subunits were co-purified with Pre1-FLAG in both STS1 and sts1-2. Equivalent expression of Pre1-FLAG was also observed in these strains. Collectively, these studies suggest that the assembly of the proteasome is likely to be unaffected in sts1-2. These studies also showed that the interaction between the Rpn10 and proteasome was unaffected in the sts1-2 mutant. Similarly, Rad23/proteasome interaction was unaffected in sts1-2 (data not shown).
The Interaction between Proteasome and Multiubiquitinated Proteins Is Altered in sts1-2-To determine whether Sts1 influenced the translocation of substrates to the proteasome, we compared the levels of multi-Ub proteins in total extract (as described in Fig. 5A) with that present in proteasomes. We expressed Pre1-FLAG in STS1 and sts1-2 and transferred the cells from 23°C to 37°C, and protein extracts were prepared and characterized by immunoblotting (as described in Fig. 5). Consistent with earlier findings (Fig. 5A), we observed significant accumulation of high molecular weight multi-Ub species in sts1-2 at 37°C (Fig. 6A). The same protein extracts were also applied to FLAG-agarose to specifically measure the levels of ubiquitinated substrates in association with proteasomes (Fig.  6B). Low levels of multiubiquitinated proteins were detected in proteasomes purified at 23°C from both STS1 and sts1-2 (lanes 1 and 2). After correcting for the level of Pre1-FLAG that was purified (Fig. 6C), we estimated that higher levels were already present in proteasomes purified from sts1-2 at 23°C (Fig. 6B,  compare lanes 1 and 2). The levels of multi-Ub proteins copurified with proteasomes from STS1 increased initially after transfer from 23 to 37°C (compare lanes 1 and 3) but remained constant thereafter, based on the level of Pre1-FLAG that was purified (Fig. 6C). In striking contrast, the level of multiubiquitinated proteins purified with proteasomes from sts1-2 increased progressively after transfer from 23 to 37°C (lanes 4, 6, 8, and 10), consistent with the higher levels in total extract (Fig.  6A). Based on these results, we propose that proteasomes purified from sts1-2 may be deficient in the clearance of multiubiquitinated substrates, following their interaction with the proteasome.

DISCUSSION
The ability of Sts1 to suppress the defects of rad23⌬ rpn10⌬ suggested a link to protein degradation. Our characterization of FIGURE 6. Multiubiquitinated species bind proteasomes in sts1-2 but are inefficiently eliminated. A, STS1 and sts1-2 were transformed with a plasmid encoding Pre1-FLAG. Precultures were grown at 23°C and then transferred to 37°C, as described in Fig. 5. Consistent with earlier results, the levels of high molecular weight (mw) multiubiquitinated proteins increased dramatically in sts1-2. B, protein extracts were incubated with FLAG-agarose to examine the corresponding levels of multiubiquitinated proteins associated with proteasomes. C, the immunoblot described in B was incubated with antibodies against FLAG to determine the levels of Pre1-FLAG that were recovered. Pairwise comparisons suggest increasing accumulation of multi-Ub proteins in proteasomes purified from sts1-2.

Sts1 Functions in the Ubiquitin/Proteasome System
the Sts1 protein and a temperature-sensitive sts1 mutant demonstrate that Sts1 performs a role in the Ub/proteasome system. Specifically, we found that Sts1 could bind the proteasome and affected proteasome activity. Furthermore, we determined that sts1-2 is highly sensitive to drugs that cause protein damage and is unable to efficiently degrade proteolytic substrates. Strikingly, sts1-2 accumulated high levels of multiubiquitinated proteins at the nonpermissive temperature, and proteasomes showed a reduced ability to clear these substrates. Pulse-chase studies confirmed the stabilization of test substrates. Moreover, the regulated turnover of Sic1-HA, which is degraded following the exit of cells from G 1 , was significantly impaired in sts1-2.
Although the stabilization of Sic1-HA in sts1-2 might be caused by defective entry into mitotic growth, it could also be the result of the aforementioned proteolytic defects.
The accumulation of multi-Ub proteins in sts1-2 does not appear to be caused by deficient proteasome activity, because chymotryptic activity was increased. However, one interpretation of our results is that an increased load of proteolytic substrates could contribute to the higher levels of multi-Ub proteins in sts1-2 (Fig. 5A). Because a major fraction of newly synthesized proteins are degraded before they mature (31,32), it is possible that the acute sensitivity of sts1-2 to translation inhibitors is caused by inefficient elimination of nascent damaged proteins.
Despite higher proteasome activity, the levels of multi-Ub proteins increased rapidly in total extracts at 37°C. However, the accumulation of these substrates in proteasomes in the sts1-2 mutant was delayed. One interpretation of this finding is that substrate delivery to the proteasome is normal, although their degradation is defective. The increased proteasome activity in sts1-2 could reflect a compensatory response to a defect at a different step in degradation. It is evident that substrate ubiquitination is unaffected in sts1-2 and that ubiquitinated proteins could successfully bind the proteasome. However, the interaction of ubiquitinated substrates to proteasomes is poorly understood, and it is uncertain whether these proteins interact accurately with proteasomes in sts1-2. Moreover, other steps in degradation by the proteasome, such as deubiquitination, substrate unfolding, or substrate translocation into the catalytic particle, could be affected in sts1-2.
We explored a plausible mechanism for Sts1-mediated suppression. Rad23 and Rpn10 can bind multi-Ub proteins and the proteasome. Based on these properties, we proposed that Rad23 might function as a shuttle-factor that delivered multi-Ub proteins to the proteasome. Genetic studies suggest that multi-Ub substrates may be transferred from Rad23 to Rpn10, a multiubiquitin chain-binding protein in the proteasome. Other models have also been proposed to explain the function of Rad23 and Rpn10 in promoting the degradation of substrates by the proteasome (4, 7). Sts1 does not bind either Rad23 or Rpn10 or affect their interactions with proteasomes ( Fig. 5C and data not shown). This is not surprising because high level expression of Sts1 suppressed the loss of Rad23 and Rpn10, indicating that Sts1 operates through an independent and functionally distinct mechanism. The ability of Sts1 to bind the proteasome, but not multi-Ub proteins, suggested that it might exert its effect on the proteasome. In agreement with this conjecture, we found that proteasome activity increased in sts1-2, and multiubiquitinated substrates were inefficiently cleared.
To suppress the pleiotropic defects of rad23⌬ rpn10⌬, Sts1 might improve proteasome function to restore the degradation of Rad23-and Rpn10-specific substrates. Alternatively, Sts1 might improve the targeting of proteolytic substrates to the proteasome by recruiting other shuttle factors. This model would support the hypothesis that Sts1 bypasses the defects of rad23⌬ rpn10⌬ by using an alternate pathway. Further study will be required to determine whether other shuttle-factors provide a mechanism to bypass the defects of rad23⌬ rpn10⌬.
The proteolytic defects of sts1-2 are similar to those caused by mutations in genes encoding components of the ubiquitin/ proteasome pathway. However, Sts1 is unique because proteasome activity is increased in sts1-2, despite its failure to degrade proteolytic substrates efficiently. Dual overexpression of Rad23 and Rpn10 did not suppress the temperature-sensitive growth defect of sts1-2. This nonredundancy supports the hypothesis that multiple mechanisms contribute to the removal of proteolytic substrates.
Yanagida and co-workers (16,17) reported that Cut8, the S. pombe counterpart of Sts1, could tether proteasomes to the nuclear membrane. It was proposed that multi-ubiquitination of Cut8 promoted its interaction with the proteasome and also initiated its eventual degradation. Similarly, we report that Sts1 is highly unstable in a wild type strain and is further destabilized in rad23⌬ rpn10⌬. We speculate that the dosage suppression of rad23⌬ rpn10⌬ by Sts1 might reflect restoring normal levels of this otherwise unstable protein. Despite some functional similarity, Cut8 and Sts1 may operate through distinct mechanisms, because STS1 is an essential gene, whereas Cut8 is dispensable for viability in S. pombe. Moreover, a cut8 null mutant is sensitive to DNA damage, whereas sts1-2 displayed no apparent sensitivity to UV light. At the nonpermissive temperature, the levels of multiubiquitinated proteins increased dramatically in sts1-2 but not in cut8 (17). Additionally, Cut8/proteasome interaction required multiubiquitination of Cut8, whereas proteasomes purified from S. cerevisiae contained unconjugated Sts1. (However, it is possible that the multi-Ub chain is removed from Sts1, following its interaction with the proteasome).
Srp1 is a nuclear localization signal-binding factor that can bind karyophilic proteins and facilitate their translocation into the nucleus. Specific mutations disrupted nucleocytoplasmic transport and also caused proteolytic defects. Tabb et al. (15) isolated Sts1 as a high copy dosage suppressor of srp1-49 and found that it suppressed the protein degradation defect of this mutant. Characterization of Sts1 revealed predominant localization to the nucleus (15). Studies described by Enenkel et al. (33,34) showed that a major fraction of cellular proteasomes co-localized with nuclei. Collectively, these studies are consistent with the findings of Tatebe and Yanagida (17), who reported that Cut8 was required for the nuclear targeting of proteasomes. We isolated Sts1 as a dosage suppressor of the growth and proteolytic defects of rad23⌬ rpn10⌬, a well characterized mutant in the Ub/proteasome pathway (3). Although the underlying mechanism of suppression is not known, the studies DECEMBER 7, 2007 • VOLUME 282 • NUMBER 49 JOURNAL OF BIOLOGICAL CHEMISTRY 35581 described here provide compelling evidence of a role for Sts1 in the Ub/proteasome system. Furthermore, because Sts1 can bind proteasomes but not multi-Ub chains, it is distinct from both Rad23 and Rpn10. Because both Sts1 and sts1-2 can bind the proteasome, the defect in sts1-2 might occur after Sts1/ proteasome binding. Collectively, these and other findings (16,33,35) suggest that Sts1 contributes to the ubiquitin/proteasome pathway of protein degradation. We propose that Sts1 is a novel regulatory factor that affects the activity of the proteasome.