- Email: [email protected]
Properties of the Proteasome Activator Subunit PA28α and its Des-Tyrosyl Analog
ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS
Vol. 359, No. 2, November 15, pp. 283–290, 1998 Article No. BB980918
Properties of the Proteasome Activator Subunit PA28a and its Des-Tyrosyl Analog Sherwin Wilk,1 Wei-Er Chen, and Ronald P. Magnusson Department of Pharmacology, Mount Sinai School of Medicine, One Gustave L. Levy Place, New York, New York 10029
Received July 1, 1998, and in revised form August 17, 1998
The proteasome activator protein PA28 or 11 S regulator may play an important role in facilitating the generation of peptides for presentation by the MHC class I system. PA28 is composed of two homologous subunits termed a and b. Removal of the carboxyl terminal tyrosine of the a subunit of PA28 abolishes activity (X. Song et al., 1997, J. Biol. Chem. 272, 27994 – 28000). To explore the structural basis of this effect the des-tyrosyl analog of PA28a prepared by site-directed mutagenesis and PA28a were expressed at high levels in a baculovirus system and purified by FPLC. Destyrosyl-PA28a neither stimulated the proteasome nor competed with PA28a for binding to the proteasome. Hydrophobic interaction chromatography revealed that the hydrophobicity of the mutant protein was considerably greater than PA28a. When the mutant protein was chromatographed on a calibrated Superose 6 column a mixture of approximately 25% oligomer and 75% monomer was found. The oligomer weakly stimulated the proteasome but this molecule was labile. Very low concentrations of SDS (0.005%) dissociated PA28a and abolished its stimulatory activity. It is concluded that the lack of activity of des-tyrosylPA28a is due to conformational changes resulting in dissociation and that the oligomeric form of PA28a is required for activation. © 1998 Academic Press Key Words: proteasome; multicatalytic proteinase complex; proteasome activator PA28; 11 S regulator; MHC class-I.
The 20 S proteasome is the catalytic core of the major extralysosomal proteolytic system. Recent studies have demonstrated that the proteasome participates in many fundamental cellular processes which include cell cycle regulation (1); activation and inactivation of oncogenic proteins, transcription factors, and regula1 To whom correspondence should be addressed. Fax: (212) 8310114. E-mail: [email protected].
0003-9861/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.
tory enzymes (2); and the generation of antigenic peptides for presentation by the MHC class-I system (3). The active sites of this barrel-shaped macromolecule are sequestered within a central chamber. The crystal structure of the proteasome from Thermoplasma acidophilum shows a narrow entrance port at each terminal ring; however, in the crystal structure of the yeast enzyme even this narrow entrance is sealed (4). This barrier to the active sites explains the inability of the 20 S proteasome to degrade native proteins (5) and accounts for the property earlier described as latency (6). The catalytic properties of the 20 S proteasome are modified by association with cellular regulatory proteins. A 19 S protein complex combines with the 20 S proteasome to form a “26 S” proteasome which can recognize and degrade ubiquitinated proteins in an ATP-dependent reaction (7). A separate regulatory protein termed PA28 or 11 S regulator acting as a positive allosteric modifier markedly stimulates proteasomecatalyzed hydrolysis of peptides but not of proteins (8, 9). PA28 is composed of two homologous 30-kDa subunits termed a and b (10). Biochemical studies suggest that PA28 is a hexamer composed of alternating a and b subunits (11), but X-ray crystallography of the recombinant a subunit reveals a heptameric structure (12). The recombinant a subunit possesses full stimulatory activity (13), whereas the b subunit by itself has either no (14) or reduced stimulatory activity (15), but by associating with the a subunit forms a molecule with a lower Kact than PA28a (16). There is evidence that PA28, which is a g-interferon up-regulated protein, plays a role in facilitating antigen presentation (17). In addition, recent studies have shown that the immunosuppressant drug rapamycin inhibits the expression of PA28 subunits (18). Structure–function studies on the PA28 regulator have shed some light on the residues necessary for its binding to the proteasome (14). It has further been shown by carboxypeptidase Y treatment and by mu283
284
WILK, CHEN, AND MAGNUSSON
tagenesis that the carboxyl terminal tyrosyl residue of the a subunit is essential for activation (14, 19). The basis for the loss of function when this tyrosyl residue is removed has not been explored at a structural level. We have found that the a subunit of PA28 can be expressed at a very high level in the baculovirus system. The expressed protein is readily extracted from Sf-9 cells after gentle homogenization with a hypotonic buffer and can be rapidly purified to essential homogeneity by a facile FPLC technique. We have used this expression system to prepare recombinant des-tyrosylPA28a for study. We document structural changes rendering des-tyrosyl-PA28a functionally inactive. Hydrophobic interaction chromatography reveals greatly enhanced hydrophobicity of the mutant protein compared to PA28a. FPLC on a calibrated Superose 6 column demonstrates that the mutant protein is largely dissociated. We further show that very low concentrations of SDS inactivate recombinant PA28a and that the basis for inactivation is dissociation of the oligomer. These studies demonstrate that PA28a must exist in an oligomeric form in order to activate the proteasome. MATERIALS AND METHODS Materials. Bovine pituitary 20S proteasome was purified as previously described (20). Antiserum against PA28a was a kind gift of Dr. K. Tanaka, Tokyo Metropolitan Institute of Medical Science. Succinyl-Leu-Leu-Val-Tyr-7-amido-4-methylcoumarin (suc-LLVYAMC)2 was purchased from Novabiochem, San Diego, CA. The QuikChange site-directed mutagenesis kit and XL2 Ultracompetent cells were obtained from Stratgene, Inc., La Jolla, CA. BaculoGold linear DNA was obtained from Pharmingen, San Diego, CA. pVL1393 was a kind gift from Dr. Max Summers, Texas Agricultural Experimental Station. Oligonucleotides were synthesized by Integrated DNA Technologies, Inc., Coralville, IA. Measurement of enzymatic activity. Enzymatic activity was measured with the substrate suc-LLVY-AMC. Enzyme and substrate in the presence or absence of PA28 proteins at the concentrations designated in the figure legends were incubated in a 0.05 M Tris–HCl buffer, pH 7.5, at a final volume of 100 ml. Reactions were terminated with 100 ml 10% TCA and the chromogen released was measured spectrophotometrically at 560 nm by a diazotization procedure (21). Cloning of PA28a and its expression in a baculovirus system. The sequence of the human a-subunit of the proteasome activator (13) was used to search the NCBI expressed sequence tag (est) data base (dbest). This search revealed a human est (clone 25849) with a sequence identical to the input sequence and containing the 59 start codon. This est derived from a human infant brain library was described as protein I-5111, interferon-g induced. The insert size of 1.1 kb was sufficiently large to encode the entire reading frame of PA28a. The clone (Lafmid BA vector) was ordered from American Type Culture Collection (ATTC No. 352436). Sequencing confirmed that the cDNA contained the entire coding sequence of PA28a. The nucleotide sequence was identical to that published (13) with the exception of the codon AGC for serine 55 of human PA28a which was AAC in the est. This resulted in a serine–asparagine mutation, likely representing a polymorphism. 2 Abbreviations used: suc-LLVY-AMC, succinyl-Leu-Leu-Val-Tyr7-amido-4-methylcoumarin; TCA, trichloroacetic acid; dbest, expressed sequence tag data base; DMSO, dimethyl sulfoxide.
The cDNA was excised from the plasmid by digestion with HindIII and NotI. The baculovirus transfer vector pVL1393 was digested with BamH1 and Not1. The PA28a cDNA was subcloned into pVL1393 with the aid of the BamH1-HindIII linkers 59GATCCGGCGCCA39 and 59AGCTTGGCGCCG39. The resulting transfer vector was cotransfected with BaculoGold linearized AcMNPV DNA. The recombinant virus was cloned by two rounds of limiting dilution in 96-well dishes. Preparation of the des-tyrosyl PA28a mutant. Site directed mutagenesis was performed on the PA28a clone from ATCC with the aid of the Stratagene Quik-Change kit. Oligonucleotides were designed in which the codon for Tyr (TAT) was mutated to a stop codon (TAG). The oligonuclotides synthesized were for the forward direction 59GGGAGAGAGGGCTCTCTCTAGATCATTCCCTTTGTTTC39 and for the reverse direction 59GAAACAAAGGGAATGATCTAGAGAGAGCCCTCTCTCCC39. Temperatures for the PCR were set at 95°C for denaturation, 55°C for annealing and 68°C for extension. PCR was run for 15 cycles in the presence of pfu DNA polymerase on an Ericomp cycler. Sequencing (Utah State University Biotechnology Center, Logan, UT) confirmed the mutation and the mutated desTyr-PA28a was subcloned into the baculovirus transfer vector and recombinant virus prepared as above. FPLC purification of recombinant PA28a and des-tyrosyl-PA28a. Approximately 1 3 108 virus-infected cells were gently mixed for 10 min at 4°C with 10 ml 10 mM Tris–EDTA, pH 7.5. The mixture was centrifuged at 1100g for 10 min and the supernatant was carefully removed. The supernatant was subjected to a second centrifugation at 12,000g for 10 min and then removed for FPLC purification. FPLC was performed on a Pharmacia system. A 3-ml aliquot (10 mg protein) was loaded onto a Mono Q column equilibrated with 10 mM Tris–EDTA, pH 7.5. The column was washed with the same buffer until the absorbance at 280 nm reached baseline. Proteins were eluted with a 20-ml gradient established between 0 M and 0.4 M NaCl in the same buffer. Samples of 0.5 ml were collected and monitored for stimulation of proteasome (PA28a) or immunoreactivity against an anti-PA28a antiserum (des-tyrosyl-PA28a). In addition PA28a and des-tyrosyl-PA28a were readily detected by Coomas-
FIG. 1. SDS–PAGE analysis (12.5% gel) of the FPLC purification of des-tyrosyl-PA28a. Lanes 1 and 5, molecular mass markers in kDa. Lane 2, Sf-9 cell extract. Lane 3, After mono Q FPLC. Lane 4, after phenyl Superose FPLC. Chromatographic conditions as described under Materials and Methods. Proteins stained with Coomassie blue.
PROTEASOME ACTIVATOR SUBUNIT PA28a
285
FIG. 2. Elution profile of PA28a (I) and des-tyrosyl PA28a (II) from a phenyl Superose column. The samples after Uno Q chromatography were applied to the phenyl Superose column and elution was conducted as described under Materials and Methods. Peak a, PA28a; peak b, des-tyrosyl-PA28a. g represents start of elution by 20% glycerol. Peaks detected by absorption at 280 nm. sie blue staining. The PA28a-containing fractions were combined and solid ammonium sulfate was added to achieve a final concentration of 1 M. The solution was applied to a phenyl Superose column equilibrated with 10 mM Tris–EDTA, pH 7.5, 1 M ammonium sulfate (buffer A). The column was washed with buffer A until the absorbance at 280 nm reached baseline. A linear gradient was established between 10 ml buffer A and 10 ml 10 mM Tris–EDTA, pH 7.5. Samples of 0.5 ml were collected. Elution of PA28a was monitored by measuring the activation of the hydrolysis of sucLLVY-AMC. Elution of des-tyrosyl-PA28a was monitored by Western blotting. After the gradient ended, a buffer of 20% glycerol, 10 mM Tris–EDTA, pH 7.5, was required to elute this protein. Des-tyrosyl-PA28a was judged to be homogeneous after phenyl Superose chromatography (Fig. 1). PA28a which eluted from phenyl Superose earlier than des-tyrosylPA28a contained minor protein contaminants. These were removed by chromatography on a Superose 6 column. Active samples from the phenyl Superose column were pooled and concentrated to 200 ml. The concentrated sample was applied to a Superose 6 column equilibrated with 10 mM Tris–EDTA, pH 7.5, and PA28a was eluted with the same buffer. Gel electrophoretic analysis of aliquots after each step of the purification procedure for des-tyrosyl-PA28a is shown in Fig. 1. A scale-up of the purification procedure was achieved by substituting a Uno Q6 column (Bio-Rad) for the mono Q column. In this case 80 mg protein was applied, and the volume of the gradient was increased to 60 ml. The remaining steps of the purification procedure were unchanged. Determination of relative molecular mass (Mr) on a calibrated Superose 6 column. A Superose 6 column was calibrated with catalase (232,000), aldolase (158,000), bovine serum albumin (67,000), and soybean trypsin inhibitor (20,000). The total column volume was 24 ml and the void volume was measured with blue dextran as 7.0 ml. Mr was determined by a plot of log Mr vs K (K 5 V 2 Vo/Vt 2 Vo, where Vo is the void volume, Vt the total volume, and V the elution volume of the protein of interest).
Gel electrophoresis. Nondissociating polyacrylamine gel electrophoresis was run on 8% gels in a 0.375 M Tris–HCl buffer, pH 8.8. Proteins were stained with Coomassie blue.
RESULTS
Expression of Des-Tyrosyl-PA28a in a Baculovirus System Des-tyrosyl-PA28a was expressed at very high levels in infected Sf-9 cells. The protein was readily released from the cell pellet by gentle mixing with a hypotonic buffer and was visible as a major band in the crude extract after SDS–PAGE (Fig. 1). Control cells infected with an unrelated virus lacked an equivalent band (not shown). The protein was purified to apparent homogeneity by a rapid two-step HPLC procedure on mono Q and phenyl Superose columns. The Hydrophobicity of Des-Tyrosyl-PA28a Is Significantly Greater than that of PA28a Nonmutated PA28a was also expressed at very high levels in the baculovirus system. Both proteins were subjected to the same FPLC purification. After FPLC on a mono Q column, both proteins eluted similarly. However a marked difference in their elution patterns after FPLC on a phenyl Superose column was observed. PA28a eluted at about 11 ml of the 20 ml gradient. Des-tyrosyl-PA28a did not elute under these
286
WILK, CHEN, AND MAGNUSSON
FIG. 3. Stimulation of the proteasome-catalyzed hydrolysis of suc-LLVY-AMC by recombinant PA28a purified from Sf-9 cells. Incubation mixtures in a total volume of 100 ml contained 1.7 mg proteasome, suc-LLVY-AMC at concentrations indicated on the abscissa, 0.05 M Tris–HCl, pH 7.5, and either 2.8 mg PA28a (F) or no activator (h). Samples were incubated for 30 min and activity was determined as described (20). Velocity is given in absorbance units at 560 nm.
conditions and a 20% glycerol-containing buffer was required for elution (Fig. 2). After phenyl Superose chromatography, the preparation of des-tyrosyl-PA28a was judged to be homogeneous by SDS–PAGE, whereas trace impurities were present in the PA28a eluate. These were removed by FPLC on a Superose 6 column. Des-Tyrosyl-PA28a Neither Stimulates Nor Binds to the 20 S Proteasome Recombinant PA28a markedly stimulated the hydrolysis of suc-LLVY-AMC by purified bovine pituitary 20 S proteasome, consistent with the results of other laboratories (14, 15). This effect was due to a lowering of the Km and an increase in the Vmax (Fig. 3). As reported by Song et al., (14) des-tyrosyl-PA28a did not activate the proteasome. To determine if des-tyrosylPA28a binds to the proteasome, competition experiments between PA28a and des-tyrosyl-PA28a were run. In these experiments a fixed amount of the mutant protein was added to mixtures of proteasome plus varying amounts of PA28a and the hydrolysis of suc-
LLVY-AMC was compared to parallel controls containing no mutant protein. As shown in Fig. 4 des-tyrosylPA28a did not compete with PA28a for binding to the proteasome. Des-Tyrosyl-PA28a Exists Predominantly as an Monomer It was possible that the failure of des-tyrosyl-PA28a to activate the proteasome or to compete with PA28a for binding to the proteasome was due to its inability to form heptamers. Accordingly the Mr’s of PA28a and the des-tyrosyl mutant were determined on a calibrated Superose 6 column. As shown in Figure 5, PA28a eluted before des-tyrosyl-PA28a. The elution of PA28a was slightly greater than catalase, indicating an Mr of about 270,000. Two peaks were found for des-tyrosylPA28a. The first eluting peak amounting to about 25% of the total protein had an elution volume identical to PA28a. The second and major peak eluted at an Mr of 38,000. The material in the first eluting peak was able to moderately stimulate the proteasome-catalyzed hydrolysis of suc-LLVY-AMC (approximately twofold);
PROTEASOME ACTIVATOR SUBUNIT PA28a
287
FIG. 4. Effect of des-tyrosyl-PA28a on the PA28a stimulation of the hydrolysis of suc-LLVY-AMC by the 20 S proteasome. Incubation mixtures (total volume of 100 ml) contained PA28a at the amounts shown on the abscissa, 2.1 mg 20 S proteasome, 4 ml 10 mM suc-LLVY-AMC in DMSO, 0.05 M Tris–HCl buffer, pH 7.5, and either 0.7 mg des-tyrosyl-PA28a (■) or no des-tyrosyl-PA28a (E). Samples were incubated at 37°C for 30 min and activity determined as described (20).
however, this protein was quite labile. No stimulation was found after overnight storage on ice, indicating that it had undergone further dissociation. Monomeric Des-Tyrosyl-PA28a Can Form Oligomers The monomeric des-tyrosyl-PA28a fraction from the Superose 6 column (Fig. 5) was concentrated approximately 10-fold and subjected to nondissociating polyacrylamide gel electrophoresis. Two protein bands of approximate equal intensity were found (Fig. 6). These results demonstrate that concentration of the monomeric protein facilitated formation of the oligomer. Activation of the 20 S Proteasome by PA28a Requires an Oligomeric Structure When PA28a was incubated with the proteasome and substrate in the presence of 0.01% SDS, activation was totally abolished. The effect of varying concentrations of SDS on activation of the hydrolysis of sucLLVY-AMC by PA28a was then determined. Concentrations of SDS up to 0.01% did not affect the basal
activity of the proteasome (Fig. 7). However a concentration of 0.005% SDS totally abolished activation by PA28a. The concentration dependence of the SDS effect was steep. A concentration of 0.001% SDS had no effect and an approximately 50% reduction in activation occurred at an SDS concentration of about 0.003%. It was observed that the activity of the proteasome in the presence of both PA28a and SDS at concentrations of 0.005% or 0.01% was lower than in the presence of SDS alone. To directly demonstrate that low concentrations of SDS dissociate PA28a, this protein was chromatographed on the calibrated Superose 6 column in the presence of SDS. When run in the presence of 0.002% SDS, a mixture of approximately 65% oligomer, 15% monomer and 20% of an intermediate species was found (Fig. 8, panel I). The intermediate species corresponds to an Mr of approximately 100,000. At this concentration of SDS, stimulation was lowered by approximately one-third (Fig. 7). When run in the presence of 0.005% SDS, only about 10% oligomer remained. A major peak emerged at a position between the intermediate species and the monomer and pre-
288
WILK, CHEN, AND MAGNUSSON
FIG. 5. Elution volumes of PA28a and des-tyrosyl-PA28a on a calibrated Superose 6 column. The column was equilibrated with 0.05 M Tris–EDTA, pH 7.5 and 0.1 M NaCl and calibrated as described under Materials and Methods. (I) PA28a; (II) des-tyrosyl-PA28a. Proteins were detected by absorbance at 280 nm. The elution volumes in milliliters are shown as are points at which samples were injected onto the column (inj.).
sumably represents a mixture of both species. At 0.005% SDS, PA28a does not activate the proteasome. DISCUSSION
The proteasome activator PA28 has been suggested to play an important role in facilitating the ability of
FIG. 6. Nondissociating polyacrylamde gel electrophoresis of destyrosyl-PA28a. Monomeric des-tyrosyl-PA28a from the Superose 6 column was concentrated approximately 10-fold and subjected to gel electrophoresis as described under Materials and Methods. Proteins were stained with Coomassie blue.
the 20 S proteasome to process peptides for presentation by the MHC class I system (13, 22). Indirect evidence in support of this role is provided by the g-interferon up-regulation of PA28 (13) and by the inhibition of the expression of its mRNA by the immunosuppressant drug rapamycin (18). Direct evidence for a role in antigen processing was obtained by demonstrating enhancement of antigen presentation following transfection of mouse fibroblasts with the gene encoding PA28a (17). Studies on the relationship of PA28 structure to its function are of importance in determining the mechanism of proteasome activation and facilitating the rational development of drugs that may potentiate or block the PA28-proteasome interaction. It has been reported that the carboxyl terminus of PA28 is essential for its activity (19). Removal of one or more amino acids by lysosomal carboxypeptidase abolishes activity. More recently, elimination of the carboxyl terminal tyrosine by site-directed mutagenesis yielded a molecule which neither stimulated nor bound to the proteasome (14). Here we confirm and extend these observations. A rapid FPLC procedure was developed to purify PA28a and des-tyrosyl-PA28a to apparent homogeneity from infected Sf-9 cell lysates. The lysates contained high levels of the expressed proteins. Des-tyrosyl-PA28a did not stimulate the chymot-
PROTEASOME ACTIVATOR SUBUNIT PA28a
289
FIG. 7. Effect of SDS on proteasome activation by PA28a. Incubation mixtures (total volume of 100 ml) contained 1.7 mg 20 S proteasome, 4 ml 10 mM suc-LLVY-AMC in DMSO, SDS at concentrations shown on the abscissa, and either 0.7 mg PA28a (F) or no PA28a (h) and 0.05 M Tris–HCl, pH 7.5. Velocity is expressed in arbitrary absorbance units at 560 nm.
rypsinlike activity of the 20 S proteasome nor did this protein compete with PA28a for binding to the proteasome. The basis for this dramatic effect brought about by the deletion of a single amino acid was explored.
Mutant PA28a bound much more tightly to a phenyl Superose column than PA28a. This very tight binding facilitated purification of the mutant protein to apparent homogeneity after only two columns. The markedly
FIG. 8. Superose 6 gel filtration of PA28a in the presence of low concentrations of SDS. (I) Column equilibrated with 0.002% SDS, 0.05 M Tris–HCl. pH 7.5, 0.1 M NaCl. (II) Column equilibrated with 0.005% SDS, 0.05 M Tris–HCl, pH 7.5, 0.1 M NaCl. The elution volumes in milliliters are shown as are points at which samples were injected onto the column (inj.). Proteins were detected by absorbance at 280 nm.
290
WILK, CHEN, AND MAGNUSSON
enhanced hydrophobicity reflects major conformational changes presumably resulting in the exposure of buried hydrophobic side chains. When this protein was subjected to gel filtration on a calibrated Superose 6 column a mixture of approximately 25% oligomer and 75% monomer was found. The Mr’s calculated from the calibration curve exceeded the theoretical values; however, the ratio of Mr’s of oligomer to monomer was 7.1, in agreement with the heptameric structure of PA28a as revealed by X-ray crystallography (12). When the monomeric form was concentrated about 10-fold and subjected to nondissociating gel electrophoresis, two protein bands were found (Fig. 6). Therefore the mutant protein can reversibly undergo a monomer– oligomer transition with the oligomeric state being favored by an increase in the protein concentration. The oligomer was capable of moderately stimulating the proteasome but its ability to stimulate did not survive overnight storage on ice. Therefore the des-tyrosylPA28a oligomer is quite labile, and failure to bind to the proteasome could be attributed to dissociation to the monomer. The Superose 6 experiment (Fig. 5) shows that even for PA28a, a small amount of monomer can be detected. The difference is that for the mutant protein, the oligomer–monomer equilibrium is shifted in favor of the monomer. We determined that the ability of PA28a to activate the chymotrypsinlike activity of the proteasome was highly sensitive to very low concentrations of SDS. The effect of SDS is due to a dissociation of PA28a and not to an effect on the proteasome. It has long been known that SDS stimulates proteasome activity presumably by inducing conformational changes (23). However the stimulatory effect of SDS occurs at a concentration of 0.02– 0.04%. Inhibition of the activation of the proteasome by PA28a occurs at an SDS concentration one order of magnitude lower (Fig. 7). There is no stimulation of proteasome basal activity at these concentrations. Surprisingly, PA28a in the presence of low concentrations of SDS slightly inhibits the proteasome. We interpret this to be due to a competition of a nonactivating protein (in this case PA28a monomer) with suc-LLVY-AMC for binding to the proteasome. Superose 6 chromatography of PA28a run in the presence of 0.002% or 0.005% SDS clearly shows PA28a dissociation with a loss of stimulation approximately equal to the percentage loss of oligomer. Therefore these studies directly demonstrate that the heptameric form of PA28a is necessary for activation of the proteasome. This conclusion is supported by the recent study of Realini et al., (15) who demonstrate that although the b-subunit of PA28 exists in solution predominantly as a monomer, activation is observed at relatively high concentrations and is presumably due to a monomer– oligomer equilibrium. In addition, in a recent mutagenesis study by Zhang et al. (24), all mutants that failed
to form heptamers were inactive. The requirement of an oligomeric structure for activity is reminiscent of the proteasome itself. Partial or total dissociation of the proteasome results in loss of activity (25). ACKNOWLEDGMENT These studies were supported by NIH Grant NS-29936.
REFERENCES 1. Pagano, M. (1997) FASEB J. 11, 1067–1075. 2. Weissman, A. M. (1997) Immunol. Today 18, 189 –198. 3. Tanaka, K., Tanahashi, N., Tsurumi, C., Yokota, K-Y. and Shimbara, N. (1997) Adv. Immunol. 64, 1–38. 4. Groll, M., Ditzel, L., Lowe, J., Stock, D., Bochtler, M., Bartunik, H. D., and Huber, R. (1997) Nature 386, 463– 471. 5. Baumeister, W., Walz, J., Zuhl, F., and Seemuller, E. (1998) Cell 92, 367–380. 6. Tanaka, K., Ii, K., Ichihara, A., Waxman, L., and Goldberg, A. L. (1986) J. Biol. Chem. 261, 15197–15203. 7. Coux, O., Tanaka, K., and Goldberg, A. L. (1996) Annu. Rev. Biochem. 65, 801– 847. 8. Ma, C. P., Slaughter, C. A., and DeMartino, G. N. (1992) J. Biol. Chem. 267, 10515–10523. 9. Dubiel, W., Pratt, G., Ferrell, K., and Rechsteiner, M. (1992) J. Biol. Chem. 267, 22369 –22377. 10. Mott, J. D., Pramanik, B. C., Moomaw, C. R., Afendis, S. J., DeMartino, G. N., and Slaughter, C. A. (1994) J. Biol. Chem. 269, 31466 –31471. 11. Ahn, K., Erlander, M., Leturcq, D., Peterson., P. A., Fruh, K., and Yang, Y. (1996) J. Biol. Chem. 271, 18237–18242. 12. Johnston, S. C., Whitby, F. G., Realini, C., Rechsteiner, M., and Hill, C. P. (1997) Protein Sci. 6, 2469 –2473. 13. Realini, C., Dubiel, W., Pratt, G., Ferrell, K., and Rechsteiner, M. (1994) J. Biol. Chem. 269, 20727–20732. 14. Song, X., von Kampen, J., Slaughter, C. A., and DeMartino, G. N. (1997) J. Biol. Chem. 272, 27994 –28000. 15. Realini, C., Jensen, C. C., Zhang, Z., Johnston, S. C., Knowlton, J. R., Hill, C. P., and Rechsteiner, M. (1997) J. Biol. Chem. 272, 25483–25492. 16. Kuehn, L., and Dahlmann, B. (1996) FEBS Lett. 394, 183–186. 17. Groettrup, M., Soza, A., Eggers, M., Kuehn, L., Dick, T. P., Schild, H., Rammensee, H-G., Koszinowski, U. H., and Kloetzel, P-M. (1996) Nature 381, 166 –168. 18. Wang, X., Omura, S., Szeweda, L. I., Yang, Y., Berard, J., Seminaro, J., and Wu, J. (1997) Eur. J. Immunol. 27, 2781–2786. 19. Ma, C.-P., M., Willy, P. J., Slaughter, C. A., and DeMartino, G. N. (1993) J. Biol. Chem. 268, 22514 –22519. 20. Orlowski, M., and Michaud, C. (1989) Biochemistry 28, 9270 – 9278. 21. Wilk, S., and Orlowski, M. (1980) J. Neurochem. 35, 1172–1182. 22. Tanaka, K., Tanahashi, N., Tsurumi, C., Yokota, K-Y., and Shimbara, N. (1997) Adv. Immunol. 64, 1–38. 23. Orlowski, M., and Wilk, S. (1981) Biochem. Biophys. Res. Commun. 101, 814 – 822. 24. Zhang, Z., Clawson, A., Realini, C., Jensen, C. C., Knowlton, J. R., Hill, C. P., and Rechsteiner, M. (1998) Proc. Natl. Acad. Sci. USA 95, 2807–2811. 25. Figueiredo-Pereira, M., Yu, B., and Wilk, S. (1994) J. Biol. Chem. 269, 621– 626.
Vol. 359, No. 2, November 15, pp. 283–290, 1998 Article No. BB980918
Properties of the Proteasome Activator Subunit PA28a and its Des-Tyrosyl Analog Sherwin Wilk,1 Wei-Er Chen, and Ronald P. Magnusson Department of Pharmacology, Mount Sinai School of Medicine, One Gustave L. Levy Place, New York, New York 10029
Received July 1, 1998, and in revised form August 17, 1998
The proteasome activator protein PA28 or 11 S regulator may play an important role in facilitating the generation of peptides for presentation by the MHC class I system. PA28 is composed of two homologous subunits termed a and b. Removal of the carboxyl terminal tyrosine of the a subunit of PA28 abolishes activity (X. Song et al., 1997, J. Biol. Chem. 272, 27994 – 28000). To explore the structural basis of this effect the des-tyrosyl analog of PA28a prepared by site-directed mutagenesis and PA28a were expressed at high levels in a baculovirus system and purified by FPLC. Destyrosyl-PA28a neither stimulated the proteasome nor competed with PA28a for binding to the proteasome. Hydrophobic interaction chromatography revealed that the hydrophobicity of the mutant protein was considerably greater than PA28a. When the mutant protein was chromatographed on a calibrated Superose 6 column a mixture of approximately 25% oligomer and 75% monomer was found. The oligomer weakly stimulated the proteasome but this molecule was labile. Very low concentrations of SDS (0.005%) dissociated PA28a and abolished its stimulatory activity. It is concluded that the lack of activity of des-tyrosylPA28a is due to conformational changes resulting in dissociation and that the oligomeric form of PA28a is required for activation. © 1998 Academic Press Key Words: proteasome; multicatalytic proteinase complex; proteasome activator PA28; 11 S regulator; MHC class-I.
The 20 S proteasome is the catalytic core of the major extralysosomal proteolytic system. Recent studies have demonstrated that the proteasome participates in many fundamental cellular processes which include cell cycle regulation (1); activation and inactivation of oncogenic proteins, transcription factors, and regula1 To whom correspondence should be addressed. Fax: (212) 8310114. E-mail: [email protected].
0003-9861/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.
tory enzymes (2); and the generation of antigenic peptides for presentation by the MHC class-I system (3). The active sites of this barrel-shaped macromolecule are sequestered within a central chamber. The crystal structure of the proteasome from Thermoplasma acidophilum shows a narrow entrance port at each terminal ring; however, in the crystal structure of the yeast enzyme even this narrow entrance is sealed (4). This barrier to the active sites explains the inability of the 20 S proteasome to degrade native proteins (5) and accounts for the property earlier described as latency (6). The catalytic properties of the 20 S proteasome are modified by association with cellular regulatory proteins. A 19 S protein complex combines with the 20 S proteasome to form a “26 S” proteasome which can recognize and degrade ubiquitinated proteins in an ATP-dependent reaction (7). A separate regulatory protein termed PA28 or 11 S regulator acting as a positive allosteric modifier markedly stimulates proteasomecatalyzed hydrolysis of peptides but not of proteins (8, 9). PA28 is composed of two homologous 30-kDa subunits termed a and b (10). Biochemical studies suggest that PA28 is a hexamer composed of alternating a and b subunits (11), but X-ray crystallography of the recombinant a subunit reveals a heptameric structure (12). The recombinant a subunit possesses full stimulatory activity (13), whereas the b subunit by itself has either no (14) or reduced stimulatory activity (15), but by associating with the a subunit forms a molecule with a lower Kact than PA28a (16). There is evidence that PA28, which is a g-interferon up-regulated protein, plays a role in facilitating antigen presentation (17). In addition, recent studies have shown that the immunosuppressant drug rapamycin inhibits the expression of PA28 subunits (18). Structure–function studies on the PA28 regulator have shed some light on the residues necessary for its binding to the proteasome (14). It has further been shown by carboxypeptidase Y treatment and by mu283
284
WILK, CHEN, AND MAGNUSSON
tagenesis that the carboxyl terminal tyrosyl residue of the a subunit is essential for activation (14, 19). The basis for the loss of function when this tyrosyl residue is removed has not been explored at a structural level. We have found that the a subunit of PA28 can be expressed at a very high level in the baculovirus system. The expressed protein is readily extracted from Sf-9 cells after gentle homogenization with a hypotonic buffer and can be rapidly purified to essential homogeneity by a facile FPLC technique. We have used this expression system to prepare recombinant des-tyrosylPA28a for study. We document structural changes rendering des-tyrosyl-PA28a functionally inactive. Hydrophobic interaction chromatography reveals greatly enhanced hydrophobicity of the mutant protein compared to PA28a. FPLC on a calibrated Superose 6 column demonstrates that the mutant protein is largely dissociated. We further show that very low concentrations of SDS inactivate recombinant PA28a and that the basis for inactivation is dissociation of the oligomer. These studies demonstrate that PA28a must exist in an oligomeric form in order to activate the proteasome. MATERIALS AND METHODS Materials. Bovine pituitary 20S proteasome was purified as previously described (20). Antiserum against PA28a was a kind gift of Dr. K. Tanaka, Tokyo Metropolitan Institute of Medical Science. Succinyl-Leu-Leu-Val-Tyr-7-amido-4-methylcoumarin (suc-LLVYAMC)2 was purchased from Novabiochem, San Diego, CA. The QuikChange site-directed mutagenesis kit and XL2 Ultracompetent cells were obtained from Stratgene, Inc., La Jolla, CA. BaculoGold linear DNA was obtained from Pharmingen, San Diego, CA. pVL1393 was a kind gift from Dr. Max Summers, Texas Agricultural Experimental Station. Oligonucleotides were synthesized by Integrated DNA Technologies, Inc., Coralville, IA. Measurement of enzymatic activity. Enzymatic activity was measured with the substrate suc-LLVY-AMC. Enzyme and substrate in the presence or absence of PA28 proteins at the concentrations designated in the figure legends were incubated in a 0.05 M Tris–HCl buffer, pH 7.5, at a final volume of 100 ml. Reactions were terminated with 100 ml 10% TCA and the chromogen released was measured spectrophotometrically at 560 nm by a diazotization procedure (21). Cloning of PA28a and its expression in a baculovirus system. The sequence of the human a-subunit of the proteasome activator (13) was used to search the NCBI expressed sequence tag (est) data base (dbest). This search revealed a human est (clone 25849) with a sequence identical to the input sequence and containing the 59 start codon. This est derived from a human infant brain library was described as protein I-5111, interferon-g induced. The insert size of 1.1 kb was sufficiently large to encode the entire reading frame of PA28a. The clone (Lafmid BA vector) was ordered from American Type Culture Collection (ATTC No. 352436). Sequencing confirmed that the cDNA contained the entire coding sequence of PA28a. The nucleotide sequence was identical to that published (13) with the exception of the codon AGC for serine 55 of human PA28a which was AAC in the est. This resulted in a serine–asparagine mutation, likely representing a polymorphism. 2 Abbreviations used: suc-LLVY-AMC, succinyl-Leu-Leu-Val-Tyr7-amido-4-methylcoumarin; TCA, trichloroacetic acid; dbest, expressed sequence tag data base; DMSO, dimethyl sulfoxide.
The cDNA was excised from the plasmid by digestion with HindIII and NotI. The baculovirus transfer vector pVL1393 was digested with BamH1 and Not1. The PA28a cDNA was subcloned into pVL1393 with the aid of the BamH1-HindIII linkers 59GATCCGGCGCCA39 and 59AGCTTGGCGCCG39. The resulting transfer vector was cotransfected with BaculoGold linearized AcMNPV DNA. The recombinant virus was cloned by two rounds of limiting dilution in 96-well dishes. Preparation of the des-tyrosyl PA28a mutant. Site directed mutagenesis was performed on the PA28a clone from ATCC with the aid of the Stratagene Quik-Change kit. Oligonucleotides were designed in which the codon for Tyr (TAT) was mutated to a stop codon (TAG). The oligonuclotides synthesized were for the forward direction 59GGGAGAGAGGGCTCTCTCTAGATCATTCCCTTTGTTTC39 and for the reverse direction 59GAAACAAAGGGAATGATCTAGAGAGAGCCCTCTCTCCC39. Temperatures for the PCR were set at 95°C for denaturation, 55°C for annealing and 68°C for extension. PCR was run for 15 cycles in the presence of pfu DNA polymerase on an Ericomp cycler. Sequencing (Utah State University Biotechnology Center, Logan, UT) confirmed the mutation and the mutated desTyr-PA28a was subcloned into the baculovirus transfer vector and recombinant virus prepared as above. FPLC purification of recombinant PA28a and des-tyrosyl-PA28a. Approximately 1 3 108 virus-infected cells were gently mixed for 10 min at 4°C with 10 ml 10 mM Tris–EDTA, pH 7.5. The mixture was centrifuged at 1100g for 10 min and the supernatant was carefully removed. The supernatant was subjected to a second centrifugation at 12,000g for 10 min and then removed for FPLC purification. FPLC was performed on a Pharmacia system. A 3-ml aliquot (10 mg protein) was loaded onto a Mono Q column equilibrated with 10 mM Tris–EDTA, pH 7.5. The column was washed with the same buffer until the absorbance at 280 nm reached baseline. Proteins were eluted with a 20-ml gradient established between 0 M and 0.4 M NaCl in the same buffer. Samples of 0.5 ml were collected and monitored for stimulation of proteasome (PA28a) or immunoreactivity against an anti-PA28a antiserum (des-tyrosyl-PA28a). In addition PA28a and des-tyrosyl-PA28a were readily detected by Coomas-
FIG. 1. SDS–PAGE analysis (12.5% gel) of the FPLC purification of des-tyrosyl-PA28a. Lanes 1 and 5, molecular mass markers in kDa. Lane 2, Sf-9 cell extract. Lane 3, After mono Q FPLC. Lane 4, after phenyl Superose FPLC. Chromatographic conditions as described under Materials and Methods. Proteins stained with Coomassie blue.
PROTEASOME ACTIVATOR SUBUNIT PA28a
285
FIG. 2. Elution profile of PA28a (I) and des-tyrosyl PA28a (II) from a phenyl Superose column. The samples after Uno Q chromatography were applied to the phenyl Superose column and elution was conducted as described under Materials and Methods. Peak a, PA28a; peak b, des-tyrosyl-PA28a. g represents start of elution by 20% glycerol. Peaks detected by absorption at 280 nm. sie blue staining. The PA28a-containing fractions were combined and solid ammonium sulfate was added to achieve a final concentration of 1 M. The solution was applied to a phenyl Superose column equilibrated with 10 mM Tris–EDTA, pH 7.5, 1 M ammonium sulfate (buffer A). The column was washed with buffer A until the absorbance at 280 nm reached baseline. A linear gradient was established between 10 ml buffer A and 10 ml 10 mM Tris–EDTA, pH 7.5. Samples of 0.5 ml were collected. Elution of PA28a was monitored by measuring the activation of the hydrolysis of sucLLVY-AMC. Elution of des-tyrosyl-PA28a was monitored by Western blotting. After the gradient ended, a buffer of 20% glycerol, 10 mM Tris–EDTA, pH 7.5, was required to elute this protein. Des-tyrosyl-PA28a was judged to be homogeneous after phenyl Superose chromatography (Fig. 1). PA28a which eluted from phenyl Superose earlier than des-tyrosylPA28a contained minor protein contaminants. These were removed by chromatography on a Superose 6 column. Active samples from the phenyl Superose column were pooled and concentrated to 200 ml. The concentrated sample was applied to a Superose 6 column equilibrated with 10 mM Tris–EDTA, pH 7.5, and PA28a was eluted with the same buffer. Gel electrophoretic analysis of aliquots after each step of the purification procedure for des-tyrosyl-PA28a is shown in Fig. 1. A scale-up of the purification procedure was achieved by substituting a Uno Q6 column (Bio-Rad) for the mono Q column. In this case 80 mg protein was applied, and the volume of the gradient was increased to 60 ml. The remaining steps of the purification procedure were unchanged. Determination of relative molecular mass (Mr) on a calibrated Superose 6 column. A Superose 6 column was calibrated with catalase (232,000), aldolase (158,000), bovine serum albumin (67,000), and soybean trypsin inhibitor (20,000). The total column volume was 24 ml and the void volume was measured with blue dextran as 7.0 ml. Mr was determined by a plot of log Mr vs K (K 5 V 2 Vo/Vt 2 Vo, where Vo is the void volume, Vt the total volume, and V the elution volume of the protein of interest).
Gel electrophoresis. Nondissociating polyacrylamine gel electrophoresis was run on 8% gels in a 0.375 M Tris–HCl buffer, pH 8.8. Proteins were stained with Coomassie blue.
RESULTS
Expression of Des-Tyrosyl-PA28a in a Baculovirus System Des-tyrosyl-PA28a was expressed at very high levels in infected Sf-9 cells. The protein was readily released from the cell pellet by gentle mixing with a hypotonic buffer and was visible as a major band in the crude extract after SDS–PAGE (Fig. 1). Control cells infected with an unrelated virus lacked an equivalent band (not shown). The protein was purified to apparent homogeneity by a rapid two-step HPLC procedure on mono Q and phenyl Superose columns. The Hydrophobicity of Des-Tyrosyl-PA28a Is Significantly Greater than that of PA28a Nonmutated PA28a was also expressed at very high levels in the baculovirus system. Both proteins were subjected to the same FPLC purification. After FPLC on a mono Q column, both proteins eluted similarly. However a marked difference in their elution patterns after FPLC on a phenyl Superose column was observed. PA28a eluted at about 11 ml of the 20 ml gradient. Des-tyrosyl-PA28a did not elute under these
286
WILK, CHEN, AND MAGNUSSON
FIG. 3. Stimulation of the proteasome-catalyzed hydrolysis of suc-LLVY-AMC by recombinant PA28a purified from Sf-9 cells. Incubation mixtures in a total volume of 100 ml contained 1.7 mg proteasome, suc-LLVY-AMC at concentrations indicated on the abscissa, 0.05 M Tris–HCl, pH 7.5, and either 2.8 mg PA28a (F) or no activator (h). Samples were incubated for 30 min and activity was determined as described (20). Velocity is given in absorbance units at 560 nm.
conditions and a 20% glycerol-containing buffer was required for elution (Fig. 2). After phenyl Superose chromatography, the preparation of des-tyrosyl-PA28a was judged to be homogeneous by SDS–PAGE, whereas trace impurities were present in the PA28a eluate. These were removed by FPLC on a Superose 6 column. Des-Tyrosyl-PA28a Neither Stimulates Nor Binds to the 20 S Proteasome Recombinant PA28a markedly stimulated the hydrolysis of suc-LLVY-AMC by purified bovine pituitary 20 S proteasome, consistent with the results of other laboratories (14, 15). This effect was due to a lowering of the Km and an increase in the Vmax (Fig. 3). As reported by Song et al., (14) des-tyrosyl-PA28a did not activate the proteasome. To determine if des-tyrosylPA28a binds to the proteasome, competition experiments between PA28a and des-tyrosyl-PA28a were run. In these experiments a fixed amount of the mutant protein was added to mixtures of proteasome plus varying amounts of PA28a and the hydrolysis of suc-
LLVY-AMC was compared to parallel controls containing no mutant protein. As shown in Fig. 4 des-tyrosylPA28a did not compete with PA28a for binding to the proteasome. Des-Tyrosyl-PA28a Exists Predominantly as an Monomer It was possible that the failure of des-tyrosyl-PA28a to activate the proteasome or to compete with PA28a for binding to the proteasome was due to its inability to form heptamers. Accordingly the Mr’s of PA28a and the des-tyrosyl mutant were determined on a calibrated Superose 6 column. As shown in Figure 5, PA28a eluted before des-tyrosyl-PA28a. The elution of PA28a was slightly greater than catalase, indicating an Mr of about 270,000. Two peaks were found for des-tyrosylPA28a. The first eluting peak amounting to about 25% of the total protein had an elution volume identical to PA28a. The second and major peak eluted at an Mr of 38,000. The material in the first eluting peak was able to moderately stimulate the proteasome-catalyzed hydrolysis of suc-LLVY-AMC (approximately twofold);
PROTEASOME ACTIVATOR SUBUNIT PA28a
287
FIG. 4. Effect of des-tyrosyl-PA28a on the PA28a stimulation of the hydrolysis of suc-LLVY-AMC by the 20 S proteasome. Incubation mixtures (total volume of 100 ml) contained PA28a at the amounts shown on the abscissa, 2.1 mg 20 S proteasome, 4 ml 10 mM suc-LLVY-AMC in DMSO, 0.05 M Tris–HCl buffer, pH 7.5, and either 0.7 mg des-tyrosyl-PA28a (■) or no des-tyrosyl-PA28a (E). Samples were incubated at 37°C for 30 min and activity determined as described (20).
however, this protein was quite labile. No stimulation was found after overnight storage on ice, indicating that it had undergone further dissociation. Monomeric Des-Tyrosyl-PA28a Can Form Oligomers The monomeric des-tyrosyl-PA28a fraction from the Superose 6 column (Fig. 5) was concentrated approximately 10-fold and subjected to nondissociating polyacrylamide gel electrophoresis. Two protein bands of approximate equal intensity were found (Fig. 6). These results demonstrate that concentration of the monomeric protein facilitated formation of the oligomer. Activation of the 20 S Proteasome by PA28a Requires an Oligomeric Structure When PA28a was incubated with the proteasome and substrate in the presence of 0.01% SDS, activation was totally abolished. The effect of varying concentrations of SDS on activation of the hydrolysis of sucLLVY-AMC by PA28a was then determined. Concentrations of SDS up to 0.01% did not affect the basal
activity of the proteasome (Fig. 7). However a concentration of 0.005% SDS totally abolished activation by PA28a. The concentration dependence of the SDS effect was steep. A concentration of 0.001% SDS had no effect and an approximately 50% reduction in activation occurred at an SDS concentration of about 0.003%. It was observed that the activity of the proteasome in the presence of both PA28a and SDS at concentrations of 0.005% or 0.01% was lower than in the presence of SDS alone. To directly demonstrate that low concentrations of SDS dissociate PA28a, this protein was chromatographed on the calibrated Superose 6 column in the presence of SDS. When run in the presence of 0.002% SDS, a mixture of approximately 65% oligomer, 15% monomer and 20% of an intermediate species was found (Fig. 8, panel I). The intermediate species corresponds to an Mr of approximately 100,000. At this concentration of SDS, stimulation was lowered by approximately one-third (Fig. 7). When run in the presence of 0.005% SDS, only about 10% oligomer remained. A major peak emerged at a position between the intermediate species and the monomer and pre-
288
WILK, CHEN, AND MAGNUSSON
FIG. 5. Elution volumes of PA28a and des-tyrosyl-PA28a on a calibrated Superose 6 column. The column was equilibrated with 0.05 M Tris–EDTA, pH 7.5 and 0.1 M NaCl and calibrated as described under Materials and Methods. (I) PA28a; (II) des-tyrosyl-PA28a. Proteins were detected by absorbance at 280 nm. The elution volumes in milliliters are shown as are points at which samples were injected onto the column (inj.).
sumably represents a mixture of both species. At 0.005% SDS, PA28a does not activate the proteasome. DISCUSSION
The proteasome activator PA28 has been suggested to play an important role in facilitating the ability of
FIG. 6. Nondissociating polyacrylamde gel electrophoresis of destyrosyl-PA28a. Monomeric des-tyrosyl-PA28a from the Superose 6 column was concentrated approximately 10-fold and subjected to gel electrophoresis as described under Materials and Methods. Proteins were stained with Coomassie blue.
the 20 S proteasome to process peptides for presentation by the MHC class I system (13, 22). Indirect evidence in support of this role is provided by the g-interferon up-regulation of PA28 (13) and by the inhibition of the expression of its mRNA by the immunosuppressant drug rapamycin (18). Direct evidence for a role in antigen processing was obtained by demonstrating enhancement of antigen presentation following transfection of mouse fibroblasts with the gene encoding PA28a (17). Studies on the relationship of PA28 structure to its function are of importance in determining the mechanism of proteasome activation and facilitating the rational development of drugs that may potentiate or block the PA28-proteasome interaction. It has been reported that the carboxyl terminus of PA28 is essential for its activity (19). Removal of one or more amino acids by lysosomal carboxypeptidase abolishes activity. More recently, elimination of the carboxyl terminal tyrosine by site-directed mutagenesis yielded a molecule which neither stimulated nor bound to the proteasome (14). Here we confirm and extend these observations. A rapid FPLC procedure was developed to purify PA28a and des-tyrosyl-PA28a to apparent homogeneity from infected Sf-9 cell lysates. The lysates contained high levels of the expressed proteins. Des-tyrosyl-PA28a did not stimulate the chymot-
PROTEASOME ACTIVATOR SUBUNIT PA28a
289
FIG. 7. Effect of SDS on proteasome activation by PA28a. Incubation mixtures (total volume of 100 ml) contained 1.7 mg 20 S proteasome, 4 ml 10 mM suc-LLVY-AMC in DMSO, SDS at concentrations shown on the abscissa, and either 0.7 mg PA28a (F) or no PA28a (h) and 0.05 M Tris–HCl, pH 7.5. Velocity is expressed in arbitrary absorbance units at 560 nm.
rypsinlike activity of the 20 S proteasome nor did this protein compete with PA28a for binding to the proteasome. The basis for this dramatic effect brought about by the deletion of a single amino acid was explored.
Mutant PA28a bound much more tightly to a phenyl Superose column than PA28a. This very tight binding facilitated purification of the mutant protein to apparent homogeneity after only two columns. The markedly
FIG. 8. Superose 6 gel filtration of PA28a in the presence of low concentrations of SDS. (I) Column equilibrated with 0.002% SDS, 0.05 M Tris–HCl. pH 7.5, 0.1 M NaCl. (II) Column equilibrated with 0.005% SDS, 0.05 M Tris–HCl, pH 7.5, 0.1 M NaCl. The elution volumes in milliliters are shown as are points at which samples were injected onto the column (inj.). Proteins were detected by absorbance at 280 nm.
290
WILK, CHEN, AND MAGNUSSON
enhanced hydrophobicity reflects major conformational changes presumably resulting in the exposure of buried hydrophobic side chains. When this protein was subjected to gel filtration on a calibrated Superose 6 column a mixture of approximately 25% oligomer and 75% monomer was found. The Mr’s calculated from the calibration curve exceeded the theoretical values; however, the ratio of Mr’s of oligomer to monomer was 7.1, in agreement with the heptameric structure of PA28a as revealed by X-ray crystallography (12). When the monomeric form was concentrated about 10-fold and subjected to nondissociating gel electrophoresis, two protein bands were found (Fig. 6). Therefore the mutant protein can reversibly undergo a monomer– oligomer transition with the oligomeric state being favored by an increase in the protein concentration. The oligomer was capable of moderately stimulating the proteasome but its ability to stimulate did not survive overnight storage on ice. Therefore the des-tyrosylPA28a oligomer is quite labile, and failure to bind to the proteasome could be attributed to dissociation to the monomer. The Superose 6 experiment (Fig. 5) shows that even for PA28a, a small amount of monomer can be detected. The difference is that for the mutant protein, the oligomer–monomer equilibrium is shifted in favor of the monomer. We determined that the ability of PA28a to activate the chymotrypsinlike activity of the proteasome was highly sensitive to very low concentrations of SDS. The effect of SDS is due to a dissociation of PA28a and not to an effect on the proteasome. It has long been known that SDS stimulates proteasome activity presumably by inducing conformational changes (23). However the stimulatory effect of SDS occurs at a concentration of 0.02– 0.04%. Inhibition of the activation of the proteasome by PA28a occurs at an SDS concentration one order of magnitude lower (Fig. 7). There is no stimulation of proteasome basal activity at these concentrations. Surprisingly, PA28a in the presence of low concentrations of SDS slightly inhibits the proteasome. We interpret this to be due to a competition of a nonactivating protein (in this case PA28a monomer) with suc-LLVY-AMC for binding to the proteasome. Superose 6 chromatography of PA28a run in the presence of 0.002% or 0.005% SDS clearly shows PA28a dissociation with a loss of stimulation approximately equal to the percentage loss of oligomer. Therefore these studies directly demonstrate that the heptameric form of PA28a is necessary for activation of the proteasome. This conclusion is supported by the recent study of Realini et al., (15) who demonstrate that although the b-subunit of PA28 exists in solution predominantly as a monomer, activation is observed at relatively high concentrations and is presumably due to a monomer– oligomer equilibrium. In addition, in a recent mutagenesis study by Zhang et al. (24), all mutants that failed
to form heptamers were inactive. The requirement of an oligomeric structure for activity is reminiscent of the proteasome itself. Partial or total dissociation of the proteasome results in loss of activity (25). ACKNOWLEDGMENT These studies were supported by NIH Grant NS-29936.
REFERENCES 1. Pagano, M. (1997) FASEB J. 11, 1067–1075. 2. Weissman, A. M. (1997) Immunol. Today 18, 189 –198. 3. Tanaka, K., Tanahashi, N., Tsurumi, C., Yokota, K-Y. and Shimbara, N. (1997) Adv. Immunol. 64, 1–38. 4. Groll, M., Ditzel, L., Lowe, J., Stock, D., Bochtler, M., Bartunik, H. D., and Huber, R. (1997) Nature 386, 463– 471. 5. Baumeister, W., Walz, J., Zuhl, F., and Seemuller, E. (1998) Cell 92, 367–380. 6. Tanaka, K., Ii, K., Ichihara, A., Waxman, L., and Goldberg, A. L. (1986) J. Biol. Chem. 261, 15197–15203. 7. Coux, O., Tanaka, K., and Goldberg, A. L. (1996) Annu. Rev. Biochem. 65, 801– 847. 8. Ma, C. P., Slaughter, C. A., and DeMartino, G. N. (1992) J. Biol. Chem. 267, 10515–10523. 9. Dubiel, W., Pratt, G., Ferrell, K., and Rechsteiner, M. (1992) J. Biol. Chem. 267, 22369 –22377. 10. Mott, J. D., Pramanik, B. C., Moomaw, C. R., Afendis, S. J., DeMartino, G. N., and Slaughter, C. A. (1994) J. Biol. Chem. 269, 31466 –31471. 11. Ahn, K., Erlander, M., Leturcq, D., Peterson., P. A., Fruh, K., and Yang, Y. (1996) J. Biol. Chem. 271, 18237–18242. 12. Johnston, S. C., Whitby, F. G., Realini, C., Rechsteiner, M., and Hill, C. P. (1997) Protein Sci. 6, 2469 –2473. 13. Realini, C., Dubiel, W., Pratt, G., Ferrell, K., and Rechsteiner, M. (1994) J. Biol. Chem. 269, 20727–20732. 14. Song, X., von Kampen, J., Slaughter, C. A., and DeMartino, G. N. (1997) J. Biol. Chem. 272, 27994 –28000. 15. Realini, C., Jensen, C. C., Zhang, Z., Johnston, S. C., Knowlton, J. R., Hill, C. P., and Rechsteiner, M. (1997) J. Biol. Chem. 272, 25483–25492. 16. Kuehn, L., and Dahlmann, B. (1996) FEBS Lett. 394, 183–186. 17. Groettrup, M., Soza, A., Eggers, M., Kuehn, L., Dick, T. P., Schild, H., Rammensee, H-G., Koszinowski, U. H., and Kloetzel, P-M. (1996) Nature 381, 166 –168. 18. Wang, X., Omura, S., Szeweda, L. I., Yang, Y., Berard, J., Seminaro, J., and Wu, J. (1997) Eur. J. Immunol. 27, 2781–2786. 19. Ma, C.-P., M., Willy, P. J., Slaughter, C. A., and DeMartino, G. N. (1993) J. Biol. Chem. 268, 22514 –22519. 20. Orlowski, M., and Michaud, C. (1989) Biochemistry 28, 9270 – 9278. 21. Wilk, S., and Orlowski, M. (1980) J. Neurochem. 35, 1172–1182. 22. Tanaka, K., Tanahashi, N., Tsurumi, C., Yokota, K-Y., and Shimbara, N. (1997) Adv. Immunol. 64, 1–38. 23. Orlowski, M., and Wilk, S. (1981) Biochem. Biophys. Res. Commun. 101, 814 – 822. 24. Zhang, Z., Clawson, A., Realini, C., Jensen, C. C., Knowlton, J. R., Hill, C. P., and Rechsteiner, M. (1998) Proc. Natl. Acad. Sci. USA 95, 2807–2811. 25. Figueiredo-Pereira, M., Yu, B., and Wilk, S. (1994) J. Biol. Chem. 269, 621– 626.