20 S proteasome from Saccharomyces cerevisiae is responsive to redox modifications and is S-glutathionylated.

The 20 S proteasome core purified from Saccharomyces cerevisiae is inhibited by reduced glutathione (GSH), cysteine (Cys), or the GSH precursor gamma-glutamylcysteine. Chymotrypsin-like activity was more affected by GSH than trypsin-like activity, whereas the peptidylglutamyl-hydrolyzing activity (caspase-like) was not inhibited by GSH. Cys-sulfenic acid formation in the 20 S core was demonstrated by spectral characterization of the Cys-S(O)-4-nitrobenzo-2-oxa-1,3-diazole adduct, indicating that 20 S proteasome Cys residues might react with reduced sulfhydryls (GSH, Cys, and gamma-glutamylcysteine) through the oxidized Cys-sulfenic acid form. S-Glutahionylation of the 20 S core was demonstrated in vitro by GSH-biotin incorporation and by decreased alkylation with monobromobimane. Compounds such as N-ethylmaleimide (-S-sulfhydril H alkylating), dimedone (-SO sulfenic acid H reactant), or 7-chloro-4-nitrobenzo-2-oxa-1,3-diazole (either -SH or -SOH reactant) highly inhibited proteasomal chymotrypsin-like activity. In vivo experiments revealed that 20 S proteasome extracted from H(2)O(2)-treated cells showed decreased chymotrypsin-like activity accompanied by S-glutathionylation as demonstrated by GSH release from the 20 S core after reduction with NaBH(4). Moreover, cells pretreated with H(2)O(2) showed decreased reductive capacity assessed by determination of the GSH/oxidized glutathione ratio and increased protein carbonyl levels. The present results indicate that at the physiological level the yeast 20 S proteasome is regulated by its sulfhydryl content, thereby coupling intracellular redox signaling to proteasome-mediated proteolysis.

The proteasome is an essential proteolytic complex in eukaryotic cells where it is responsible for the degradation of many cellular proteins. It plays an important role in cell-cycle regulation, cell signaling, including apoptosis, and elimination of abnormal proteins generated by mutation (1,2) and oxidative damage (3)(4)(5).
In recent years, many publications have reported the reversible S-glutathionylation of a discrete number of proteins (6). Protein S-glutathionylation seems to play an essential role in redox regulation. This form of regulation has direct effects on both enzyme activity and the ability of transcription and replication factors to bind DNA targets. The mechanism of protein glutathionylation has evolved in recent years from the strict belief that this event would take place solely when intracellular GSSG 1 levels increased upon oxidative stress, with the formation of mixed disulfides between protein Cys-SH residues and GSSG (7,8), to the present and more complete understanding of protein sulfhydryl chemistry and evidence showing formation of protein-Cys-SOH derivatives (9) that are prone to Sglutathionylation by reduced GSH (10).
Either individual enzyme activity or global cellular responses can be rapidly controlled by the oxidation of protein-Cys-SH residues (reviewed in Ref. 6) generating Cys-SOH, Cys-SO 2 H, and Cys-SO 3 H acid forms (11), where Cys-SOH is susceptible to S-thionylation and reversibly reduced to Cys-SH (11)(12)(13)(14). Cys-SOH formation and S-glutathionylation during enzyme catalysis and redox signaling are novel cofactors in the context of redox regulation (14).
In a recent publication (15) it was demonstrated that the reduced and oxidized forms of GSH modulate the chymotrypsin-like activity of purified 20 S proteasome extracted from mammalian cells. In the present report we show that the activity of the 20 S proteasome purified from the yeast Saccharomyces cerevisiae is also sensitive to GSH, though in a different way from that observed in the mammalian proteasome. The 20 S proteasome extracted from yeast is inhibited by reduced GSH and S-glutathionylated in vitro, as well as in vivo, when cells are submitted to oxidative challenge. Considering that the proteasome plays important role in cell signaling regulation by hydrolysis of many proteins involved in cascade events of the cellular regulatory pathways, it is not surprising that its activity may be regulated by its Cys-SH residues redox status.
Extraction and Purification of the 20 S Proteasome-Yeast cells were disrupted according to the protocol described in Ref. 16. The 20 S proteasome from untagged strain was purified by sequential chromatographs on DEAE-Sepharose, Sephacryl TM S-400 column (Amersham Biosciences), and a monoQ column in an HPLC system. The purification was performed according to a method described previously (17), except that gel filtration was used instead of a sucrose gradient, and the monoQ column was used in the HPLC instead of the fast protein liquid chromatography system. The 20 S proteasome core from the RJD1144 strain with a FLAGHis 6 -tagged PRE1 subunit was purified by affinity chromatography or, when specified, by immunoprecipitation. Affinity chromatography was performed using a HiTrap TM chelating HP column (Amersham Biosciences) attached to a P1 peristaltic pump (Amersham Biosciences) according to the manufacturer's protocol. The proteasome was eluted from the column with 300 mM imidazole. Active fractions from all steps of all preparations were assayed by the degradation of the fluorogenic peptide s-LLVY-MCA and confirmed by inhibition with tri-leucine vinyl sulfone. Final enriched fractions were pooled, concentrated, and rediluted twice in a Centricon CM-30 apparatus (Amicon). Purification by immunoprecipitation was performed according to the protocol described in Ref. 16 with the anti-FLAG M2 affinity gel freezer-safe antibody (Sigma) immobilized on agarose beads. The 20 S proteasome was eluted from the agarose beads, when specified, by incubation with the FLAG® peptide (Sigma) according to the manufacturer's protocol. The final preparations were analyzed by SDS-PAGE and by electrophoresis on a 5% non-denaturing polyacrylamide gel according to the protocol described in Ref. 17. The three different 20 S proteasome preparations described here were used for the assays. The results shown below (in vivo and in vitro) refer to the preparations obtained from strains containing FLAGHis 6 -tagged PRE1; however, we did not observe any difference compared with preparations from strains containing untagged PRE1 (data not shown).
Hydrolysis Assay of Fluorogenic Peptides-20 S proteasome (0.5-3 g) or cell extracts (400 -500 g protein/ml) were incubated at 37°C in 10 mM Tris/HCl buffer, pH 7.8, containing 20 mM KCl and 5 mM MgCl 2 , here referred to as standard buffer. Incubation was started by the addition of 25-50 M of the peptides (other additions are specified in the legends to the figures). The reaction was stopped by adding 4 volumes of 0.1 M sodium borate, pH 9, containing 7.5% ethanol. Fluorescence emission was recorded at 430 nm (excitation at 365 nm). MCA liberated from the substrates was calculated from a standard curve of free MCA.
Oxidation of 20 S Proteasome with H 2 O 2 -When specified, purified 20 S proteasome preparations were treated with H 2 O 2 . Proteasome concentration in these experiments was typically 50 g or as otherwise specified, and the preparation was incubated for 30 min at room temperature in the presence of 5 mM H 2 O 2 with or without 100 M DTPA in standard buffer. After incubation, excess H 2 O 2 was removed by three cycles of centrifugation and redilution with standard buffer through Microcon CM-10 filters (Amicon). Aliquots of the 20 S proteasome were taken for further incubation or for hydrolysis assay after determination of protein concentration.
Reduction of 20 S Proteasome with NaBH 4 -After isolation of the 20 S core with the anti-FLAG antibody, as described above, the agarose beads were washed three times according to the manufacturer's protocol. NaBH 4 reduction was performed with the 20 S proteasome bound to the anti-FLAG-agarose bead complex incubated in the presence of 20 mM NaBH 4 buffered in 50 mM Tris/HCl, pH 8.6, for 40 min at 37°C. After incubation, the suspension was centrifuged, the supernatant was utilized for GSH determination (described below), and the beads were washed twice with standard buffer. The proteolysis assay, when specified, was performed with the 20 S proteasome still bound to the beads.
GSH Determination-Intracellular GSH was extracted by lysing the cell pellets in 1 volume of glass beads and 2 volumes of 3.5% sulfosalicylic acid. The suspension was vortexed for 20 min at 4°C in a multifold vortex and centrifuged at maximum speed in a microcentrifuge. This procedure was repeated twice, and the supernatants were combined. Total GSH, as well as GSSG, was assayed according to a protocol described previously (15). The determination was performed by reaction with DTNB in the presence of glutathione reductase and NADPH. Samples for GSSG measurement were incubated previously for 1 h with NEM after adjusting the pH to 7. GSH released from 20 S proteasome was detected as follows: 20 S proteasome isolated by immunoprecipitation was reacted with NaBH 4 as described above. The supernatant recovered was filtered through a Microcon CM-10 filter. The filtrate was acidified to remove remaining NaBH 4 (reaction with NaBH 4 was performed at pH 8.6), and the pH was adjusted to 7 prior to the addition of DTNB, glutathione reductase, and NADPH. GSH detection was increased in the presence of glutathione reductase and NADPH comparing to detection in the presence of DTNB only. GSH concentration was calculated from standard curves of GSSG. Intracellular reduced GSH was determined by the difference between total GSH and GSSG.
GSH-Biotin Preparation-GSH-biotin derivative was achieved as described in Ref. 18. GSH was reacted with sulfo-NHS-biotin (EZ-Link TM NHS-LC-Biotin; Pierce) at 1/1 molar ratio in 50 mM NaHCO 3 , pH 8.5, for 1 h at room temperature. The reaction was stopped by the addition of 5ϫ excess NH 4 CO 3 compared with initial biotin concentration.
Proteasome S-Modification by the GSH-Biotin Derivative-H 2 O 2treated or untreated 20 S proteasome (450 g) was incubated with 5 mM GSH-biotin for 20 min at room temperature. Excess GSH-biotin was removed by filtration through a Microcon CM-10 filter followed by redilution three times in standard buffer. After washing, protein was diluted in standard buffer, mixed with streptavidin-agarose beads (100 l/mg protein), and incubated for 60 min at 4°C. The beads were washed by four cycles of centrifugation/redilution in standard buffer containing 150 mM NaCl, and the remaining protein was eluted from the beads by incubation for 30 min in standard buffer containing 0.1% SDS and 10 mM DTT, followed by centrifugation at maximum speed for 15 min. The supernatants were submitted to 12.5% SDS-PAGE after determination of protein concentration.
Cell Viability-Aliquots of the cell suspension were diluted to an A 600 of 0.2 immediately after the treatments, followed by a further 5,000ϫ redilution. Aliquots of 100 l of the final suspension were spread on YPD-agar plates (five per sample) and incubated at 30°C for 48 h. After incubation, colonies were counted manually. Cell viability assays were repeated at least five times.
Assay for the Protein Carbonyl Group-Carbonyl group formation in the proteins was determined by derivatization with dinitrophenylhydrazine and quantified by spectrophotometry as described in Ref. 19.
Protein Determination-Protein concentration was determined with Bradford reagent (Bio-Rad).

Yeast 20 S Proteasomal Activity Is Inhibited by Sulfhydryl
Compounds-Purified 20 S proteasome core isolated from S. cerevisiae was assayed for hydrolysis with the s-LLVY-MCA substrate in the presence of GSH, GSSG, and Cys ( Fig. 1). GSSG had a slight effect on 20 S proteasomal activity at any concentration tested. However, GSH or Cys strongly inhibited the proteasomal activity at 5 mM concentration. The GSH precursor GC also inhibited chymotrypsin-like activity in a pattern similar to that obtained by Cys treatment (results not shown). The trypsin-like activity determined by the hydrolysis of the fluorogenic substrate t-butoxycarbonyl-Leu-Lys-Arg-MCA was 50% decreased in the presence of 10 mM GSH (results not shown). The trypsin-like specific activity (mol MCA/min/ mg) represented less than 5% of the chymotrypsin-like activity measured under the experimental conditions described here (the same protocols were used for both determinations). The peptidylglutamyl-hydrolizing activity assayed by the hydrolysis of the fluorogenic peptide carbobenzoxy-Leu-Leu-Glu-MCA was not affected by any of the sulfhydryl compounds tested (GSH, GC, or Cys) at any concentration or by GSSG (result not shown).
One question raised by these results was how the reduced form of GSH and other thiols tested (Cys and GC), but not GSSG, inhibited 20 S proteasomal activity. Our working hypothesis was that Cys-SH residues in the 20 S proteasome structure are probably oxidized to the Cys-SOH form, susceptible to S-glutathionylation by GSH, as described elsewhere (6,20), but not by GSSG, according to the reaction,

Yeast 20 S Proteasome S-Glutathionylation
To test this hypothesis, purified 20 S proteasome preparations were incubated in the presence of 5 mM H 2 O 2 and 100 M DTPA. The iron chelator DTPA was used to prevent the Fenton reaction and consequently the generation of the very reactive hydroxyl radical, which may produce nonspecific protein oxidation, in addition to iron-catalyzed thiol oxidation and generation of several protein-sulfur derivatives (21). Thus, in the absence of iron or another transition metal, sulfhydryl oxidation to Cys-SOH may prevail, though further oxidation of Cys-SOH to Cys-SO 2 H and Cys-SO 3 H is expected (11). According to our results, 20 S proteasome treatment with H 2 O 2 in the presence of DTPA decreased chymotrypsin-like activity to 80% of the original level ( Fig. 2A), whereas when H 2 O 2 treatment was performed in the absence of DTPA the activity was reduced to 60% (data not shown). When 20 S proteasome was pretreated with H 2 O 2 plus DTPA the chymotrypsin-like activity was much more affected by GSH ( Fig. 2A). In fact, after H 2 O 2 /DTPA pretreatment, inhibition by GSH occurred at concentrations as low as 0.01 mM (35% inhibition), and GSH at 1 mM promoted stronger inhibition (60%; see Fig. 2A) than that verified without pretreatment (Fig. 1). In contrast, GSSG did not affect 20 S proteasome activity (Fig. 2A). These results are in agreement with our hypothesis in the reaction shown above.
In contrast to H 2 O 2 , DTT treatment enhanced proteasomal activity by about 15-20% (Fig. 2B). Moreover, it sensitized the proteasome to GSH incorporation after H 2 O 2 treatment. When 20 S proteasome samples were treated with 10 mM DTT, followed by incubation with H 2 O 2 /DTPA, the chymotrypsin-like activity was even more sensitive to GSH when compared with samples not preincubated with DTT ( Fig. 2A). In this condition, proteasomal activity was decreased to 50 and 5% in the presence of 0.5 or 1 mM GSH, respectively (Fig. 2B), whereas when the same GSH concentrations were employed without DTT pretreatment, chymotrypsin-like activity was decreased only to 60 and 40%, respectively ( Fig. 2A). Cys-SOH already present in 20 S proteasome was probably reduced to Cys-SH by DTT, with the consequent prevention of hyperoxidation of Cys-SOH to Cys-SO 2 H or Cys-SO 3 H by H 2 O 2 treatment.
As expected according to our hypothesis described in the reaction shown above, when the 20 S core was pretreated only with DTT, no alteration by GSH was observed in its activity (Fig. 2B). Also, proteasome reduced by DTT was not inhibited by GSSG (result not shown) as could be expected, because GSSG is able to react with sulfhydryl groups to form mixed disulfides (7,8). Reduced Cys-20 S might be prevented from reacting with GSSG by structural constraints imposed by nearby groups.
Proteasomal Activity Is Inhibited by Cys-SH and Cys-SOH Reactants-To demonstrate that modification of Cys-20 S core residues is responsible for the inhibition of chymotrypsin-like activity, we next tested this activity in the presence of Cys-SH and Cys-SOH reactants, such as NBD, dimedone, and NEM (Table I). We observed that chymotrypsin-like activity was inhibited 40 and 30% by 50 M NBD and 1 mM NEM, respectively. NEM alkylates Cys-SH whereas NBD is incorporated into both Cys-SH and Cys-SOH. This might be, at least in part, the reason why NBD is more potent than NEM in terms of chymotrypsin-like activity inhibition. After inhibition by NBD, 20 S proteasomal activity was recovered by incubation with 5 mM DTT (result not shown). Because DTT treatment leads to NBD release (9), this result indicates that inhibition was promoted by Cys conjugation.
The specific Cys-SOH reagent dimedone (9) produced low inhibition when compared with the former reagents. Dimedone promoted 23% inhibition only at 10 mM concentration. On the other hand, when purified 20 S proteasome preparations were pretreated with H 2 O 2 /DTPA prior to incubation with dimedone, we observed increased proteolytic inhibition compared with the inhibition observed in the absence of H 2 O 2 /DTPA pretreatment followed by dimedone incubation (results shown in italics in Table I). Upon H 2 O 2 pretreatment, 10 mM dimedone decreased chymotrypsin-like activity to 50% of that observed in H 2 O 2 control samples.
Taken together, the results reported thus far indicate that any group located in Cys residues of the 20 S core decreases its hydrolytic activity, at least the chymotrypsin-like activity, which is considered the strongest of its activities (1). It seems that Cys residues in the 20 S proteasome must be reduced as much as possible to allow maximum activity. Nevertheless, these residues appear to oxidize easily to Cys-SOH.
Cys-SOH Formation and S-Glutathionylation-To demonstrate that Cys residues in the 20 S proteasome structure are oxidized to Cys-SOH, 20 S proteasome preparations were incubated with NBD. This compound reacts with Cys-SH, as well as with Cys-SOH. NBD adducts of Cys-SOH and Cys-SH can be distinguished by their spectra (9). The proteasome-Cys-S(O)-NBD adduct was generated by 20 S proteasome pretreatment with H 2 O 2 /DTPA followed by NBD incubation (Fig. 3, dashed line spectrum). This adduct showed maximum absorbance at 345 nm whereas the purified 20 S core not oxidized by H 2 O 2 yielded the NBD adduct with a maximum absorbance at 420 nm (Fig. 3, solid line spectrum).
We also reacted denatured 20 S proteasome preparations with the -SH reactant mBrB (22). Denatured 20 S proteasome samples were preincubated with DTT or GSH. After incubation, DTT and GSH were removed, and the samples were treated overnight with 150 M mBrB. Proteasome-Cys-bimane conjugates were detected by fluorescence emission recorded after removal of mBrB. DTT-reduced samples showed fluorescence emission at least twice as high as control samples, whereas GSH-treated samples showed reduced fluorescence, probably because 20 S proteasome mixed disulfides could not conjugate with mBrB (Fig. 4). This result indicates that GSH does not play the role of a reducing compound as DTT, but probably GSH was incorporated into the protein structure by S-conjugation.
To demonstrate effectively that the effect of GSH on proteasomal activity is because of S-glutathionylation, 20 S proteasome preparations were incubated with the GSH-biotin derivative according to the protocol described in the literature (18). This method allows the direct determination of protein S-glutathionylation, because biotinylated GSH-proteasome com-plexes can be isolated by the streptavidin affinity procedure. 20 S proteasome pretreatment with H 2 O 2 increased GSH incorporation (Fig. 5). The protein concentration determined after protein elution from the streptavidin-agarose beads was 4-fold higher in samples pretreated with H 2 O 2 . Protein recovered from control and H 2 O 2 -treated samples after elution from the streptavidin beads was 10 and 40.5 g, respectively (the amount of protein reacted with GSH-biotin was the same in both samples, i.e. 300 g). This result is direct proof that 20 S proteasome is susceptible to S-glutathionylation by means of GSH addition to Cys-SOH. Because a significant amount of S-glutathionylated 20 S proteasome was detected in the control sample, part of its Cys residues were probably already oxidized to Cys-SOH (Fig. 5). It should be emphasized that the reaction with GSH-biotin was performed under non-denaturing conditions. The preparation was brought to denaturing conditions only after incubation with and removal of GSH-biotin to displace the 20 S core from the streptavidin-agarose beads. Thus the results described in Fig. 5 do not allow us to predict how many or which subunits of the 20 S core were S-glutathiony-  lated. We tried to denature the proteasome just after incubation with GSH-biotin and prior to the streptavidin incubation to determine approximately how many and which subunits would be S-glutathionylated. However, under these conditions, the proteasome concentration of control and H 2 O 2 -treated samples recovered from the agarose beads was not sufficient to be detected on the gel by Coomassie Blue staining. This is an indication that perhaps very few or even only one subunit might be S-glutathionylated.
Comparative Catalytic Regulation by GSH among 20 S Proteasome Structures-Chymotrypsin-like activity of 20 S proteasome from Methanosarcina thermophila, an Archaeon bacterium, was assayed in the presence of GSH and NEM for comparison to its counterpart, the yeast 20 S proteasome. The Archaeon proteasome was not affected by GSH and was inhibited slightly by NEM (Table II). In the case of mammalian 20 S proteasome, both GSH and GSSG at micromolar concentrations activated the chymotrypsin-like activity, whereas it was inhibited by millimolar concentrations (15). Moreover, 20 S mammalian proteasomal activity was inhibited strongly by NEM. On the other hand, the yeast 20 S core is inhibited only by reduced GSH (Fig. 1) and is inhibited strongly by NEM (Table I) whereas the Archaeon proteasome is not affected by any GSH form and is only affected slightly by NEM (Table II).
These results indicate a possible role of Cys residues inside the 20 S core controlling its catalytic activity. Comparing Cys residue content among archaeon, yeast, and mammalian (human) 20 S proteasomes, 28, 74, and 100 (108 total Cys) reduced Cys residues are found, respectively. Most likely the differential effect of GSH addition on proteasome activity is related to non-conserved Cys residues. Our belief is that the 20 S proteasome core has evolved as an important sensor of cellular redox status by modification in its catalytic activity brought about by changes in its Cys residues, promoting the cellular response to redox alterations.
In Vivo 20 S Proteasome Glutathionylation-Our next approach was to search for in vivo 20 S proteasome S-glutathionylation. Our hypothesis was that proteasome would be sensitive to intracellular reductive capacity, being more active as the cell reductive environment is maximized. This might be an additional way for cells to trigger the signaling response to oxidative alterations. To test this hypothesis, we treated cells with increasing H 2 O 2 concentrations. After treatment, the hydrolysis of the fluorogenic substrate s-LLVY-MCA was determined in preparations of 20 S proteasome isolated from the cell extract with the anti-FLAG system and in the proteasome-free cell extract obtained after 20 S proteasome extraction. The 20 S core-anti-FLAG antibody complex bound to agarose beads was assayed for the hydrolysis of the fluorogenic peptide s-LLVY-MCA before and after incubation in the presence of 20 mM NaBH 4 to reduce the S-S bridges formed by mixed disulfides or to reduce Cys-SOH to Cys-SH. NaBH 4 can also reduce carbonyl groups or Schiff bases formed by the oxidation of some specific amino acid residues to alcohol derivatives (23). Our purpose in using NaBH 4 was to reduce oxidized Cys residues inside the 20 S core, thereby recovering proteasome sulfhydryls and promoting GSH release from the core. GSH was assayed in the supernatant recovered after reduction with NaBH 4 , as described under "Materials and Methods." The reductive cellular capacity was evaluated by the determination of the GSH/GSSG ratio. We also determined cell viability and the formation of carbonyl protein as a marker of oxidative stress (Table III).
H 2 O 2 treatment leads to a significant increase in the formation of protein carbonyl when added to the culture medium at FIG. 4. mBrB incorporation into the 20 S core after DTT and GSH preincubation. 20 S proteasome (10 g) was incubated in standard buffer for 30 min with DTT or for 10 min with GSH at room temperature at the indicated concentrations. GSH or DTT was removed, and protein samples were washed in standard buffer by filtration. The Cys-mBrB conjugate was generated by overnight incubation with 150 M mBrB at 4°C in the dark. After incubation, 20 S proteasome was precipitated with 20% trichloroacetic acid followed by washing three times with 10% trichloroacetic acid. The protein was dissolved in 5 M guanidine buffered in 0.1 M Tris/HCl, pH 7.5. Fluorescence emission was recorded at 476 nm (excitation at 400 nm). Results are expressed as means Ϯ S.D. of three independent experiments. *, p Ͻ 0.000025. **, p at least 0.001 (ANOVA).

FIG. 5. 20 S proteasome S-glutathionylation by GSH-biotin.
Glutathionylated 20 S proteasome was purified from control (300 g) and H 2 O 2 -treated (300 g) samples after incubation with GSH-biotin followed by SDS-PAGE. The complete experimental procedure is described under "Materials and Methods." The gel shown is representative of three independent experiments. The left lane refers to 5 g of a standard preparation of 20 S proteasome, and the middle and right lanes refer to the GSH-biotin-incubated control (10 g) and to H 2 O 2treated samples (40.5 g), respectively.

Yeast 20 S Proteasome S-Glutathionylation
a concentration of at least 0.5 mM (Table III). At this concentration, H 2 O 2 -promoted oxidative stress was accompanied by decreased reductive capacity of the cell, as evaluated by the GSH/GSSG ratio, which dropped to 68.5% compared with that found in control cells. Under the same conditions, proteasomal activity decreased 50% before incubation with NaBH 4 and was partially recovered after reduction with NaBH 4 to 70% of control samples (Table III). Hydrolysis of the fluorogenic peptide s-LLVY-MCA, determined in the proteasome-free extract, increased by 60%. Reports in the literature (24 -26) show that mammalian cells under oxidative stress show increased proteolysis. The cited authors assume that increased proteolysis is because of increased proteasomal activity such as that seen in vitro when oxidized substrates show increased hydrolysis levels by purified 20 S proteasome preparations from mammalian cells (3). However, our results show that yeast 20 S proteasome is highly affected by oxidative stress whereas hydrolysis not dependent on proteasome is increased (Table III). The reason for the partial recovery of proteasome activity after NaBH 4 treatment (Table III) is probably the reduction of Cys-SOH or Cys-SSG to Cys-SH. However, other oxidative processes are responsive to reduction by NaBH 4 , e.g. formation of carbonyl and Schiff bases (19), but in this case the original amino acid structures are not regenerated as expected for Cys-SOH and Cys-SSG. GSH release from the 20 S core after reduction with NaBH 4 (Table III) is strong evidence that proteasomal activity might be regulated by S-glutathionylation under intracellular oxidative conditions, and consequently decreased reductive capacity, here attested by a reduced GSH/ GSSG ratio upon H 2 O 2 cell treatment.
Taken together, results obtained in vivo after cell treatment with H 2 O 2 indicate that loss of reductive cellular capacity, according to GSH/GSSG ratios and protein carbonyl levels, is associated with loss of 20 S proteasomal activity and with its S-glutathionylation. These data are an indication of redox modulation of 20 S proteasome activity. DISCUSSION It is becoming increasingly apparent that many oxidantsensitive proteins are S-glutathionylated in response to intracellular redox status (6, 18, 27) (reviewed in Ref. 28). Cys-SOH is an intermediate form involved in redox regulation and catalysis by protein sulfhydryl groups (reviewed in 6). Distinct fates for protein Cys-SOH have been considered (11,14) whereby S-glutathionylation would be one of the mechanisms modulating protein activity. Protein S-glutathionylation is a reversible process, and there is considerable evidence that GSH release is controlled enzymatically, probably by glutaredoxin (29).
In this manuscript we describe the mechanism by which 20 S proteasome is glutathionylated in vitro. Taken together, those results indicate that during moderate oxidative stress Cys residues inside the 20 S proteasome core might be oxidized to Cys-SOH, reversibly protected by S-glutathionylation, and most likely deglutathionylated after cellular recovery from oxidative stress.
From the data in Table III we calculated that one to two molar GSH was released per mol of 20 S proteasome, considering that the molecular mass of the yeast protease is ϳ700 kDa. GSH was not detected in 20 S proteasome samples obtained from control cells, probably because of the sensitivity of the assay utilized for GSH detection. The low GSH concentration released from the 20 S core is a suggestion that S-glutathionylation of 20 S proteasome at the physiological level is highly specific; probably very few Cys residues in the core are prone to S-glutathionylation during metabolic processes. Our hypothesis is that S-glutathionylation of the 20 S proteasome is coupled to the oxidation of sulfhydryls, according to the reaction shown above.
Comparing the results described here about the effect of GSH and GSSG on proteasomal activity from yeast (see Figs. 1 and 2) and M. thermophila (Table II) to earlier results (15) obtained for mammalian 20 S proteasome, it is clear that the 20 S proteasome homologues respond differentially to changes in redox conditions. In this regard, it is important to observe that the mammalian counterpart has either long-or short-lived proteins as substrates, and its localization inside cells is widespread; it is found in the cytoplasm and nucleus (30). On the other hand, the yeast proteasome is associated with the degradation of short-lived proteins and localized mainly inside the cell nucleus, and perhaps not more than 20% occurs in the cell treatment Cells were grown in YPD medium to A 600 of 0.6 -0.8 followed by a further 1-h incubation with H 2 O 2 added to the medium at the final concentrations shown. After incubation, aliquots were taken for cell viability assay. The remaining cells were harvested by centrifugation and washed twice with water. Cell extract preparation, 20 S proteasome purification by immunoprecipitation, reduction with NaBH 4 , and measurements of proteolysis, total and oxidized glutathione, GSH released from 20 S proteasome, and carbonyl proteins are described under "Materials and Methods." ND, not determined. c 20 S proteasome was purified from the cell extract by immunoprecipitation and assayed for proteolysis before and after reduction with NaBH 4 . d Proteolysis was determined in the cell extract, free of the 20 S proteasome particle, recovered after incubation two times with anti-FLAG antibody.
e GSH was released from samples of 20 S proteasome after reduction with NaBH 4 , as described in footnote c. f The protein fraction from cell extracts was precipitated with 20% trichloroacetic acid. All results are means Ϯ S.D. of four to five independent experiments. g p ϭ 0.027 (ANOVA).
cytoplasm associated with the nuclear endoplasm reticulum network (31). It is interesting to point out that GSH distribution inside yeast cells is still not clear. Although it has been demonstrated already that GSH is distributed inside the mammalian nucleus at concentrations as high as in the cytoplasm (32), its presence inside the yeast nucleus is not clear, neither is the distribution of glutarredoxin isoforms inside yeast cells clear thus far. A mammalian nuclear isoform was described already (33,34) whereas information on yeast nuclear isoforms is still lacking. Because glutaredoxin activity is coupled to direct GSH consumption, its presence inside the nucleus would be strong evidence of GSH distribution in the nucleus. Considering proteasome distribution and its role in the yeast cell, as discussed above, elucidation of GSH distribution inside yeast cells is an important matter to corroborate the findings discussed here.
Another important difference suggested in the literature (30,31) is that yeast proteasome is always capped by the 19 S regulatory unit in contrast to the finding that its counterpart, the mammalian 20 S proteasome, is found in 3-4-fold excess over the 19 S regulatory unit. In our opinion, even if it is true that the yeast proteasome is always capped with the 19 S regulator, this does not rule out the possibility that the catalytic unit represented by 20 S proteasome is regulated independently of either protein ubiquitinylation or substrate recognition by the 26 S proteasome. Our results indicated that redox regulation by glutathionylation is important in vivo (Table III). Other examples of redox regulation and S-glutathionylation have been described for other metabolic processes (27,35,36). Our results lead us to speculate that 20 S proteasomal activity can be modulated by S-glutathionylation through the increased presence of oxidants, which would be used in signaling processes. Besides, transient 20 S proteasome inhibition would decrease the hydrolysis of proteins responsible for redox signaling, e.g. AP-1-like factors. AP-1-like proteins are sensors of the redox state of the cell (37) and, as already demonstrated in mammalian cells, are degraded by the proteasome (38). The metabolic advantage of protein S-glutathionylation in redox signaling is the prevention of irreversible oxidation of the Cys thiol group to Cys-SO 2 H or Cys-SO 3 H, permitting protein reactivation by reduction. Such mechanism might work in parallel to the main mechanism controlling proteasome-mediated proteolysis, i.e. ubiquitinylation. The relationship between proteasome glutathionylation and ubiquitinylation of substrates remains to be established.