Proteasomes (Multi-protease Complexes) as 20 S

0 downloads 0 Views 3MB Size Report
From the Institute for Enzyme Research, The University of Tokushima, Tokushima 770, Japan .... The Proteasomes (Eukaryotic Multi-protease Complexes).
Vol. 263, No. 31. Issue of November 5, PP. 1620!3-16217,1988 Printed in U.S.A.

THEJOURNALOF BIOLOGICAL CHEMISTRY Q 1988 by The American Society for Biochemistry and Molecular Biology, Inc.

Proteasomes (Multi-protease Complexes) as 20 S Ring-shaped Particles in a Variety of Eukaryotic Cells* (Received for publication, May 9, 1988)

Keiji TanakaS, TetsuroYoshimura, Atsushi Kumatori, and Akira Ichihara From the Institute for Enzyme Research, The University of Tokushima, Tokushima 770,Japan

Atsushi Ikai andMasaaki Nishigai From the Department of Biophysics and Biochemistry, Faculty of Science, University of Tokyo, Bunkyo-ku, Tokyo 113, Japan

Keiichi Kameyama and Toshio Takagi From the Institute for Protein Research, Osaka University, Suita 565, Japan

Latentmulticatalyticprotease complexes, named Recently we (1, 2) and others (3-17) have reported very proteasomes, were purified to apparent homogeneity large intracellularproteasesin various mammalian cells. from various eukaryotic sources, such as human, rat, Studies have shown that these large proteases have quite and chicken liver, Xenopus laevis ovary, and yeast different functional and structural characteristics from other (Saccharomyces cerevisiae), and their functional and well-known proteases. These large proteases have some comstructural properties werecompared. They showed la- mon properties, such as a latent form and multiple peptidase tency in breakdown of [methyLSH]casein, but were activities that are catalyzed by at least three independent greatly activated in various ways, such as by addition active sites within a single complex (1, 8,11).Furthermore, of polylysine. They all degraded three types of fluoro- these enzymes are all unusually large, with a sedimentation genic oligopeptides at the carboxyl side of basic, neu- coefficient of approximately 20 S and a molecular mass of tral, and acidic amino acids, and the three cleavage reactions showed different spectra for inhibition, sug- 700-800 kDa (2, 18) and consist of multiple nonidentical gesting that they had three distinct active sites. The subunits (19). By electron microscopy they appear to have a proteasomes all seemed to be seryl endopeptidases with symmetrical ring shape (2, 18, 19) that is consistent with a similar pH optima in the weakly alkaline region. Their simulated model obtained by x-ray small-angle scattering physicochemical properties, such as their sedimenta- analysis (19). Based on these findings, we have named these tion coefficients (19 S to 22 S ) , diffusion coefficients large multifunctional protease complexes proteasomes (19, (2.0-2.6 X lo-’ cm2 s-’), molecular masses (700-900 20). Previously (I), we found by quantitative enzyme immukDa),andcirculardichroicspectra,weresimilar. Their amino acid compositions were also very similar. noassay that these proteasomes were abundant in all ratcells and tissuesexamined. Although these multicatalytic proteases Electron microscopy showed that theyhadsimilar well-defined symmetrical morphology, appearing tobe have been found mainly in various mammalian cells (1,8,ll), ring-shaped particles with a small hole in the center. similar types of large alkaline proteases have also been found All the proteasomes seemed to be multisubunit com- in fish muscle (21, 22) and yeast (23). Therefore, one way of plexes consisting of 15-20 polypeptides with molecu- studying the biological role of these proteases is to determine lar masses of 22-33 kDa and isoelectric points of pH their distribution. In this work, we purified similar types of 3-10, but they showed species-specific differences in multi-protease complexes from human, rat, and chicken liver, subunit multiplicity. Moreover, they differed immu- Xenopus laevis ovary, and yeast, and demonstrated that they nologically, as shown by Ouchterlony tests andimmu- are functionally and structurallyrelated enzymes. These findnoblotting analyses, although cross-immunoreactivi- ings indicate that proteasomes are ubiquitously distributed as ties of some subunits or domains were observed. These results indicate that thesizes and shapesof these pro- large ring-shaped 20 S particles in avariety of eukaryotic cells teasomes have been highly conserved during evolution, ranging from human to yeast cells. Interestingly, the size, shape, and subunit multiplicity of but that they show species-specific differences in immunoreactivities and subunit structures. Thus protea- these proteasomes are remarkably similar to those of the ringmmes with similar structure and function seem to be shaped particles of 20 S found previously in a wide variety of eukaryotic organisms (24-38), including ribonucleoprotein ubiquitously distributed in eukaryotic organisms ranging from man to yeast. This distribution implies the particles with specific RNA species (24-27, 30-33, 38), 19 S general importanceof these proteasomes for proteoly- particles with certain heat-shock proteins (25, 26), and parsis. ticles with tRNA processing nuclease activities (29). Very recently we found that 19 S particles in HeLa cells are identical to mammalian proteasomes (20). Moreover, the identity of 19 S ribonucleoprotein particles with a multicatal* This work was supported in part by grants from the Ministry of ytic proteinase from rat muscle was reported by Falkenburg Education, Science, and Culture of Japan and the Foundation for et al. (39). However, it is unknown whether these particles Enzyme Application, Osaka. The costs of publication of this article were defrayed in part by the payment of page charges. This article are all identical, since these miniparticles differ in immunomust therefore be hereby marked ‘‘advertisement” in accordance with logical reactivities and subunit compositions (31, 38). In this paper, we show that the ubiquity and interspecies structural 18 U.S.C. Section 1734 solely to indicate this fact. $ To whom correspondence should be addressed. variations of proteasomes account for these immunological

16209

16210

The Proteasomes (EukaryoticMulti-protease Complexes)

and structural differences of the various 20 S particles reported so far, and discuss the identity of proteasomes with the subcellular 20 S particles in various eukaryotic cells. EXPERIMENTALPROCEDURES

Materials-The compounds used were as follows: Fast Flow QSepharose, Mono-Q packed columns (Pharmacia LKB Biotechnology Inc.), Cosmosil5C4-300 packed column (Nakara Chemicals, Kyoto), Suc'-Leu-Leu-Val-Tyr-MCA (PeptideInstitute, Inc., Mino), CbzAla-Arg-Arg-MNAand Cbz-Leu-Leu-Glu-NA (gifts from Dr. L. Waxman), "'I-protein A (8.3 pCi/rg, Du Pont/New England Nuclear), molecular weight markers (Pharmacia), and PI markers (Oriental Yeast Co., Tokyo). Assay of Protease Activity-Protease activity was assayed as described previously (1). The degradation of [meth~l-~HIcasein was determined by measuring the radioactivity of the acid-soluble fragments cleaved. The hydrolytic activities on various fluorogenic substrates were determined by measuring the fluorescence of groups liberated from these peptides. Purification of Proteasomes from Various Eukaryotic Cells-Proteasomes from various sources, such as human, rat,and chicken liver, X. &vis ovary, and yeast were purified by sequential chromatographies on Fast Flow Q-Sepharose, Bio-Gel A-1.5m, hydroxylapatite, heparin-Sepharose CL-GB, and FPLC Mono-Q columns. The procedures used were essentially similar to those reported previously (I), but two steps were improved: Fast Flow Q-Sepharose chromatography and FPLC with a Mono-Q column were used as the first and final steps, respectively. Details of the modifications of the procedure are as follows. Fast Flow Q-Sepharose chromatography was carried out a t a high flow rate (200 ml/h). The crude extract was applied to a column (5 X 60 cm) equilibrated with 25 mM Tris-HC1 buffer (pH 7.5) containing 1 mM 2-mercaptoethanol and 20%glycerol, and proteins were eluted with a linear gradientof 0-0.8 M NaCI. Fractions with high activity were concentrated by ultrafiltration on a PM-10 membrane (Amicon) to about 25 ml and thenapplied to a Bio-Gel A1.5m molecular sieving column. The enzymes from X. lueuis and yeast were concentrated as described above, whereas those from other sources were concentrated by precipitation with polyethylene glycol 6000 (15%), because the protein concentration was higher. Subsequent steps on Bio-Gel A-1.5m, hydroxylapatite, and heparin-Sepharose CL-GB columns were performed as described previously (1).A Pharmacia FPLC system with a Mono-Q column was used as the final step of purification of all these enzymes. The enzyme preparations obtained by heparin-Sepharose CL-GB chromatography were directly subjected to Pharmacia FPLC on a Mono-Q column (1 X 10 cm) equilibrated with 25 mM Tris-HC1 buffer (pH 7.5) containing 1 mM dithiothreitol and 20% glycerol.Elution was carried out ata flow rate of 1ml/min with a linear gradient of 0-0.8 M KC1 in the same buffer. The enzymes from human, rat, chicken, and Xenopus were eluted from the column as single symmetrical peaks at the same concentration of about 340 mM KCl, whereas the enzyme from yeast was eluted at a higher KC1 concentration (400 mM). This procedure was completed within 2 h a t room temperature, and all other procedures were performed at 4 "C. Starting with 100 g (wet weight) of tissues from various sources, averages of 5.1 mg (human liver), 4.5 mg (rat liver), 2.6 mg (chicken liver), 2.9 mg (X. luevis ovary), and 3.1 mg(yeast) of purified enzymes were obtained. The final enzyme preparations were apparently homogeneous,judging by their single symmetrical profiles on Mono-Q FPLC and theirsingle protein bands onelectrophoresis in polyacrylamide gel under nondenaturing conditions. Biochemical Analyses-SDS-PAGE was carried out by the method of Laemmli (40) in 12.5% slab gels with or without 0.1% SDS. Isoelectric focusing in polyacrylamide gel with urea was done by the method of O'Farrell (41), and that in a column according to the manufacturer's protocol. Proteins were detected with Coomassie Brilliant Blue. The proteins used as molecular weight markers were phosphorylase b (94,000), BSA (67,000), aldolase (43,000), carbonic anhydrase (30,000), soybean trypsin inhibitor (20,100), and a-lactal-

' The abbreviations used are: SUC-,succinyl; Cbz, N-benzyloxycarbonyl; MCA, methylcoumarylamide; NA, 2-naphthylamide; MNA, 4methoxy-2-naphthylamide;SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; HPLC, high performance liquid chromatography; BSA, bovine serum albumin; FPLC, fast protein liquid chromatography.

bumin (14,400). Various acetylated cytochromes c with PI values of 4.1,4.9,6.4,8.3,9.7, and 10.6 wereused as PI markers. Reverse-phase HPLC was performed as follows: Samples of0.5 mg of purified enzymes were applied directly to a Cosmosil 5C4-300 column (10 X 250 mm) equilibrated with 0.05% trifluoroacetic acid. Gradient elution was performed with an increasing concentration of acetonitrile containing 0.05% trifluoroacetic acid in a Waters model 141 HPLC system a t a flow rate of 1 ml/min. The elution was monitored as absorbance at 280 and 215 nm. Amino acid analysis was performed as described previously (2). The amount of cysteic acid was not determined. Physicochemical Analyses-Analytical ultracentrifugation and measurements of quasi-elastic light scattering and circular dichroism were performed as described previously (2). Electron Microscopy-Details of the method were described elsewhere (19). A sample of about 100 pg of purified enzyme was applied to a TSK G4000SW column (Toyo Soda, Tokyo) connected to an HPLC system (Jasco, Tokyo) and materials were eluted with 100 mM ammonium acetate buffer (pH 7.4). The elution was monitored by measuring the absorbance a t 280 nm. Fractions in the major peak containing monomeric forms of the enzyme were collected for electron microscopic examination. The enzyme was negatively stained on a glow-discharged, carbon-coated Formvar membrane with 3% uranyl acetate at pH 4.5. Negatively stained samples were examined in a Hitachi H7000 electron microscope at a direct magnification of X 50,000. Immunological Analysis-Antisera against various proteasomes were raised in rabbits, and their IgG fractions were prepared as described previously (I). The standard procedure was used for Ouchterlony double-diffusion tests (42). Immunoelectrophoretic blot analysis was performed by a modification of the method of Towbin et al. (43). Briefly, samples (25pgof protein) separated by SDSPAGE in 12.5% gel weretransferred electrophoretically to Durapore membranes (Millipore) with a semi-dry electroblotter (Sartorius). The membranes were treated with 3% BSA and then with various anti-proteasome antibodies (10 pglml) and '"I-protein A (0.25 pCi/ ml). They were then washed extensively with 0.05% Tween 20 solution and autoradiographed at -70 "C. concentration was measured by the Protein Assay-Protein method of Bradford (44) with BSA as a standard. The protein concentration of the purified enzyme was calculated from the absorbance at 280 nm, assuming E!m : values of 11.2 (human liver), 9.6 (rat liver), 10.8 (chicken liver), 12.3 (X.lueuis ovary) and 7.4 (yeast proteasomes). These values were determined with a refractometer as described by Lodola et at. (45) using a solution of BSA of known absorbance as a standard. RESULTS

I. Functional Propertiesof Proteasomes Latency-We have reported that a high molecular weight protease from rat liver is present in a latent state in cells and can be activatedin various ways (1). Therefore, we examined whether the enzymes from other species also have latent proteinase activity. Table I shows that all the purified enzymes have very lowactivity for the breakdown of [3H]casein, but that their activities are strongly activated by addition of poly-L-lysine, the most effective activator of the ratliver enzyme. These enzymes were also activated in artificial ways such as by heat treatment and trypsin digestion, as reported previously for the enzyme from rat liver (data not shown). Thus, these results clearly indicate that all these enzymes show latent activity for protein hydrolysis. Substrate Specificity-The activities of the enzymes from five different species on low molecular weight peptides were examined. The enzymes degraded three types of fluorogenic oligopeptideswithacidic, basic, andneutralhydrophobic amino acidsat their carboxyl termini; namely, Cbz-Leu-LeuGlu-NA, Cbz-Ala-Arg-Arg-MNA, andSuc-Leu-Leu-Val-TyrMCA. Fig. 1 shows the pH dependence of their activities on these typical substrates. The pHoptimafor Suc-Leu-LeuVal-Tyr-MCA and Cbz-Leu-Leu-Glu-NA were in the range of 8.0-8.5, whereas those for Cbz-Ala-Arg-Arg-MNAwere in

16211

The Proteasomes (Eukuaryotic Multi-protease Complexes) TABLE I Latent proteolytic activityof purified proteasoms Purified enzymes (5 pg of protein) were used for assays. Poly-L-lysine (34 kDa) was addedat a final concentration of 0.5 mdml. Data are means f S.D. for three different exDeriments. Degrading activity of [methyl-3H]casein Addition liver

Human liver

Rat liver

Chicken

1.9 f 0.3 18.4 f 2.2

3.8 f 0.5 23.1 f 3.6

2.5 f 0.4 18.0 k 4.0

x.loeuis ovarv

Yeast

%/h

None Poly-L-lysine

PH

FIG.1. Effect of pH on the activities of various enzymes on the three fluorogenic substrates. Tris-HC1 (0.1M) was used, and the pH of the assay mixture was measured directly after increasing the volume to 1 ml. SDS ata final concentration of 0.05% was added to the assay mixtures with Suc-Leu-Leul-Val-Tyr-MCA (left panel) and Cbz-Leu-Leu-Glu-NA (right panel),but not with Cbz-Ala-ArgArg-MNA (middle panel). Values for human (O), rat (O), chicken (A), Xenopus (V), and yeast ( X ) proteasomes are shown as percentages of the maximum activity. the range of9.0-9.5. The pH dependence of the cleavage reactions of these three peptides were very similar for all enzymes, suggesting that all eukaryotic enzymes can be regarded as functionally similar types of proteases. Interestingly, the activities for degrading Cbz-Leu-Leu-Glu-NA and Suc-Leu-Leu-Val-Tyr-MCA were markedly increased by 0.02-0.05% SDS, but the degrees of stimulation of different preparations of the enzymes varied greatly (3-30-fold). In contrast, the hydrolytic activity for Cbz-Ala-Arg-Arg-MNA was partially reduced by the addition of a low concentration of SDS (data not shown). Inhibitor Sensitivity-We then examined the effects of various inhibitors on the activities on the three substrates. The inhibitor spectra of the enzymes from all sources were similar (data not shown). Leupeptin and chymostatin preferentially inhibitedthe cleavages of Cbz-Ala-Arg-Arg-MNAand Suc-Leu-Leu-Val-Tyr-MCA, respectively. In contrast, these two inhibitors had no effect on the hydrolysis of Cbz-LeuLeu-Glu-NA. These three activities of all enzymes were inhibited, although to different extents, by sulfhydryl-blocking agents, such asN-ethylmaleimide and 5,5'-dithiobis-(2-nitrobenzoate), but Ep-475, a specific cysteine-protease inhibitor, had no effect on any of the activities. Therefore, the enzymes cannot be considered to be thiol proteases, although thiol residues appear to play a critical role in their activities. Phenylmethylsulfonyl fluoride and diisopropyl fluorophosphate inhibited the three hydrolytic reactions to different extents. Since these compounds are covalent modifiers of serine residues of enzyme proteins, the enzymes are thought to be serine proteases. In addition, hemin had different effects on these three activities. These findings strongly suggest that the hydrolyses of these threepeptides are catalyzed a t distinct catalytic sites,as suggested previously (1,8, 11).Thus, all the data are compatible with the idea that the enzymes all have multiple catalytic sites within a single protein complex.

6.6 f 0.5 34.7 f 5.9

8.2 f 0.6 35.4 f 5.2

II. Structural Properties of Proteasomes Physicochemical Properties-First, we examined the physicochemical properties of proteasomes purified from various cells and tissues. The results are summarized in Table 11. All the proteasomes sedimented as single components with sedimentation coefficients (szo,,,,) of 19 S to 22 S. Measurements by quasi-elastic light scattering gave diffusion coefficients (Dz0,,,,)of2.02-2.60 X cm2 s" for these enzymes. Their effective hydrodynamic (Stokes) radii were calculated to be 82-106 A according to theEinstein-Stokes relation. From the the molecular masses of these enzymes values of s20,,,, and Dz0,,,,, were estimated to be 710-910 kDa by Svedberg's equation. These results indicate that all the enzymes in species ranging from man to yeast have a similar large size. The similarity in size of these proteasomes was supported by the observation that all the enzymes had almost the same retention time on molecular sieve chromatography on FPLC on a Superose 6 column (data not shown). The PI value of the yeast enzyme was estimated to be 4.6 and those of the native enzymes of human, rat, and chicken liver, and X. h v i s were found to be approximately 5.0 by isoelectric focusing. Morphology-We have reported that rat liver proteasomes appear byoelectron microscopy to be Fing-shaped particles of 160 f 10 A in diameter $nd 110 f 7 A in height with a small central hollow of 10-30 A in diameter (19). The enzyme from rat skeletal muscle was reported to have similar morphology (18).Here, we examined the morphology of proteasomes from other eukaryotic cells. Fig. 2 shows electron micrographs of the enzymes purified from human, chicken, Xenopus, and yeast after negative staining with uranyl acetate. Interestingly, all these enzymes showedvery similar morphology, appearing as ring-shaped particles with a well-defined symmetrical structure and a small hole in the center. The particle diametersof proteasomes of four different species ( N = 100) were measured at higher magnification (Fig. 2, righthundphotogruphs). The diameters (average f S.D.) of human, chicken, and yeast pcoteasomes were, respectively, 127 f 7, 124 k 7, and 123 f 7 A, whereas that of Xenopus proteasomes was significantly larger, being 153 f 12 A, with a broader distribution of values, like that of rat liver proteasomes. Since these proteasomes showed no difference in either molecular weight or Stokes radius, as described in the previous section, the difference in the average diameter seen by electron microscopy could be due to distortion of molecular shape during preparation of rat andXenopus proteasomes. The internal structure of proteasomes is difficult to study due to complexity of subunit proteins. But two interesting observations were made. First, in many cases, particularly in human proteasomes, we detected semiglobular units of approximately 30-50 A in diameter along the central hole (Fig. 2A, arrows). These units were too large to be single subunit proteins and probably consisted of several proteins. Second, in about one-fourth to one-third of the proteasomes of all four species, one or two slits connecting the central hole to the

The Proteasomes (Eukaryotic Multi-proteaseComplexes)

16212

TABLEI1 Physicochemical properties of proteasomes Values of S Z O , ~Dzo,w, , and PI were measured as described under "ExperimentalProcedures" and in Ref. 2. Values for molecular weight were estimated from values of s ~ , and ~ , those of D20,u,using the Svedberg equation. Stokes . of molecular ellipiticity ([6']zz0.,, based on residue molarity) were radii were calculated from values of D z o , ~Values calculated from circular dichroism measurements. source

Enzyme

SZ0.U

s Human liver -9,000 2.28 5.0 21.8 Rat liver" 2.50 19.85.0 Chicken liver 2.02 20.0 X . laevis ovary 2.15 19.6 2.60 20.0 Yeast (S. cereuisiae) Data are from Tanaka et al. (2).

D20.l"

M,

x 10" X X

outside were clearly visible (Fig. 2, B and D, and arrows in

D).These observations suggested that the proteasomes are rather fragile or flexible moleculesconsisting of 6-8 substructures, each containing several subunit proteins. Subunit Structure-We then examined the subunit structures of these proteasomes. When the purified enzymes were denatured with 1%SDS and subjected to electrophoresis on polyacrylamide gel with 0.1% SDS, they gave multiple bands in the range of 22-33 kDa (Fig. 3). The patternswere similar in the presence and absence of a sulfhydryl-reducing reagent (data not shown). Although the enzymes from different species were all separated into smaller polypeptides with a similar molecular weight range, the exact electrophoretic mobilities or sizes of the resulting polypeptides were somewhat different. Moreover, on isoelectric focusing after dissociation with 8 M urea, these enzymes showed strikingly different PI values of 3-11, and clear species differences in patterns (PI variations) (Fig. 4). Moreover, when these multiple components were separated by reverse-phase HPLC on a Cosmosil5C4-300 column (Fig. 5), they showed different elution profiles: the profiles of those from Xenopus laevis and yeast were especially different from those of the other three species. These differences indicate that proteasomes from different sources have minor interspecies differences in subunit multiplicity. Spectral Properties-The absorption spectra of these five proteasomes resembled each other and exhibited a maximum at 278 nm with a small shoulder at 283 nm (data not shown). The extinction coefficients, at 280 nm were calculated to be 11.2 (human liver), 9.6 (rat liver), 10.8 (chicken liver), 12.3 (X. laevis ovary), and 7.4 (yeast). No peak or shoulder was detected at around 260 nm in any preparation,indicating that they were all free from nucleic acids. The circular dichroism spectra in the far-ultraviolet region for all proteasomes showed the double minimum characteristics of an cy-helical structure (Fig. 6). The spectrum of the rat liver enzyme has been reported (2). The minimum values of the mean residual ellipticity at 220 nm were calculated to be -9,800 (human), -12,000 (rat), -10,300 (chicken), -11,600 (Xenopus), and -9,600 (yeast). The similar patterns of these CD spectra suggest that all these proteasomes are similar in secondary structure. Amino Acid Composition-The amino acid compositions of the five proteasomes are shown in Table 111. All the enzymes except the yeast enzyme had remarkably similar amino acid compositions. The yeast enzyme differed in certain amino acids such as Asx,Arg, Ile, Phe, Tyr, Met, and Pro. The higher content of Asx and lower content of Arg of the yeast enzyme may account for its acidic nature shown by its PI value (Table I).

PI

870,000 720,000 910,000 840,000 710,000

106 -11,600 98 82

[&O"rn degrees cm2 mol"

A

ern2 s" X94 X 10" 85

Stokes radius

4.9 5.0 4.6

-12,000 -10,300 -9,600

III. Immunological Properties of Proteasomes We next examined the immunological reactivities of these proteasomes. Polyclonal antibodies againstthe enzymes from the five species were raised in rabbits. As shown in Fig. 7, the enzymes were all clearly distinguishable by the Ouchterlony double-diffusion test, indicating species-specificdifferences in immunological reactivity. Previously, we found that the antibody against the rat liver enzyme cross-reacted with enzymes from other tissues, such as skeletal muscle, kidney, heart and brainof the same rat (I),indicating the absence of tissue specificity of proteasomes in the same species. The species-specific immunological differences seem to have been acquired during evolution. We then examined whether the species-specific differences revealed by Ouchterlony analysis were due to differences in immunoreactivity of subunit components of the various proteasomes. Fig. 8 shows the results of immunoelectrophoretic blot analyses. In all cases, the antibody against one enzyme reacted with almost all the subunits of that enzyme, although the reactivities with individual components were not correlated with theirproteincontents. Moreover, the antibody against an enzyme of one species also partially cross-reacted with some components of the enzymes of other species: the antibody against the enzyme from rat liver reacted strongly with one or two subunits of other enzymes (Fig. 8C), whereas the antibodies to theenzymes from human, chicken, X. laevis, and yeast reacted weakly, but significantly, with various components of other enzymes (Fig. 8, B , D, E, and F). These partial cross-immunoreactivities with subunit components of other enzymes observed under denatured conditions using polyclonal antibodies against native enzymes were not consistent with the results of Ouchterlony immunodiffusion analysis (Fig. 7). These differences can be explained by supposing that some subunits in each enzyme complex show the same immunoreactivity, but that these subunits are located in the complexes in such a way as notto react with antibodies. The existence of several components with similar immunoreactivities in different species suggests that some subunits, or domains, of the various proteasomes have been conserved during evolution. DISCUSSION

Physiological Role of Ubiquitously Distributed Proteasomes-In the present study, we purified proteasomes from cells of a variety of eukaryotes ranging from man to yeast. All the eukaryotic proteasomes obtained by the present study had the same enzymological function. The enzymes had two unique properties: latency of proteolytic function and multiplicity of catalytic siteswithin a single molecular complex, as

The Proteasomes (EukaryoticMulti-protease Complexes) 1

-

2

3

4

5

-

20” 0 14-

16213

-

m

””

FIG.3. SDS-PAGE analysis of proteasomes from various eukaryotes. SDS-PAGE was carried out asdescribed under “Experimental Procedures.” Lanes 1-5 show the electrophoretic patternsof the enzymes from hnuman, rat, and chicken liver, X . laeuis ovary, andyeast, respectively. The left and right lanesare M, marker proteins. Samples of 25 pg of the enzymes were used. Proteins were stained with Coomassie Brilliant Blue.

1

.

.,

2

3

4

5

:.?

I

f

8.3-

; .”

9.7-

i

10.6,

-

-

FIG.2. Electron micrographs of proteasomes. Source of proteasomes: A, human liver; I?, chicken liver; C, X . [aeois ovary; D, yeast. Bars on. the left represent 1000 A and those on the right represent 200 A. Samples were negatively stained with 3% uranyl acetate. Photographsa t right show the similarityof different particles selected a t random. Arrows in A and R-D indicate semiglobular units and slits, respectively (see text).

in rat liver proteasomes (1).Their molecular structures may beresponsiblefor this latency, since some conformational change could easily affectthe catalytic activity of the complex. The latency may be important physiologically for regulation of intracellular proteolysis. The multicatalytic activities may

.

FIG.4. Urea-isoelectric focusing analysis of proteasomes from various eukaryotes. Urea-isoelectric focusing wascarried out as described under “Experimental Procedures.” Lanes 1-5 show the isoelectric focusing patterns of the enzymes from human, rat, and chicken liver, X . laeuis ovary, and yeast, respectively. Bars on theleft show the positions of various acetylated cytochromes c used as PI markers. Samples of 25 pg of the enzymes were used. Proteins were stained with Coomassie Brilliant Blue.

be involved in cascade or cooperative reactions in the proteolytic pathway,in which they would be advantageous for rapid proteolysis, or they might be responsible for post-translational modification of various proteins, such as in processing of various precursor proteins. This work showed that proteasomes are present in a wide variety of eukaryotic organisms ranging from man to yeast, implying that these proteasomes are generally important in proteolytic functions. Proteasomes are extralysosomal proteases and probably play an essential role in the so-called nonlysosomal proteolytic pathway, which is thought to be responsible for selectivedegradation of abnormal proteins. In

The Proteasomes (Eukuryotic Multi-protease Complexes)

16214

I

I

I

Human Liver

RatLiver

8c

Xenopus Laevir Ovary

Yeast

0

e

s

P

U

-Q

>

c. 0

a

a

1

b

I

&me

am

1

la8

Liver

am.

11.

zm’ 8

4

Chicken

RetentionTime(min) FIG. 5. Separation of multiple components from proteasomes by reverse-phase HPLC. Chromatography on a Cosmosil5C4-300 reverse-phase column was carried out as described under “Experimental Procedures.” 0.025 absorbance units) and 215 nm (- - -, 0.25 Samples of0.5mgwere used for analyses at 280 nm (-, absorbance units). Theblank run, performed by injecting 200 p1 of buffer, indicates that two peaks with retention times of about 25 min and 300 min are artifacts due to injection of samples and increase in the acetonitrile concentration, respectively.

all living cells of prokaryotes and eukaryotes, rapid proteolysis is important for preventing the accumulation of unnecessary proteins, which are continuously generated in cells by mutation of genes, errors of transcription and translation, and postsynthetic damage of various proteins (46). The wide distribution of proteasomes is consistent with these ubiquitous cellular functions. But these enzymes do not appear to be present in bacteria such as Escherichia coli, suggesting that proteasomes are present only in eukaryotic cells. Interestingly, we recently found that proteasomes are partly localized in the nucleus and that their concentration is high in mammalian liver: suggesting their biological importance for nuclear proteolysis. This may beanother reason for the presence of proteasomes in only eukaryotic cells. K. Tanaka, et al., manuscript submitted for publication.

Similarity and Interspecies Structural Variation of Various Proteasomes-In the present study, we also showed that proteasomes resembled each other in gross structure. They all had unusually large molecular masses of 700-900 kDa and sedimentation values of approximately 19 S to 22 S (Table 11), and appeared to be symmetrical ring-shaped particles by electron microscopy (Fig. 2). In addition, they had similar amino acid compositions (Table 111)and secondary structures, as judged by their circular dichroic spectra (Fig. 6). Thus, the gross structures of proteasomes appear to have been highly conserved during evolution of eukaryotes. Interestingly, all the proteasomes were multisubunit complexes consisting of 15-20 small polypeptides with similar molecular masses of 22-23 kDa (Fig. 3), but with strikingly different PI values of3-10 (Fig. 4). Previously, we reported that multiple components isolated from rat liver proteasomes

The Proteasomes (Eukaryotic Multi-protease

Wavelength (nrn)

FIG.6. Circular dichroism spectra in the far-ultravioletregion. Spectra were obtained at protein concentrations of 0.9-1.5 mg/ ml. Results are averages of 16 accumulations of the enzyme from human ( A ) , chicken ( B ) , Xenopus ( C ) , and yeast (D)in solution containing 25 mM Tris-HC1 (pH 7.5), 0.1 M KC1, and 0.2 mM dithiothreitol. A cell with a light path of 0.1 cm was used.

TABLE 111 Amino acid compositionsof the proteasomes Human Rat Chicken X . loeuis Yeast liver

livef ovary liver

mol %

Asparagine or 8.4 8.78.7 8.2 aspartic acid Glutamine or 12.4 11.2 13.1 12.9 12.5 glutamic acid 5.5 5.4 5.4 5.2 Arginine 6.2 6.7 5.5 6.0 Lysine Histidine 1.9 1.8 1.9 1.9 Alanine 9.5 9.5 9.0 9.4 8.1 8.0 8.5 7.9 Glycine 8.7 Leucine 8.8 8.3 8.4 Isoleucine 5.7 5.3 5.7 5.8 Valine 7.2 7.2 7.0 7.1 3.4 3.4 3.5 3.3 Phenylalanine 4.6 4.5 4.5 4.9 Tyrosine 0.3 0.4 0.4 0.3 Tryptophan Serine 5.8 6.1 6.8 5.7 Threonine 5.4 5.8 5.5 5.5 NDb 1.4 NDb NDb Half-cystine Methionine 3.1 3.1 2.8 3.0 Proline 3.6 3.0 3.6 3.3 Data from Tanaka et al. (2). ND. not determined.

11.0

3.6 6.2 1.3 8.9 9.4 8.3 6.5 7.1 2.8 3.5 0.4 6.2 5.5 NDb 1.8 4.5

Complexes)

16215

(Fig. 8), suggesting that during evolution some subunits, or domains, of proteasomes have been conserved, although not completely, because cross-reactivity was observed between components with different molecular weights. In addition, the finding that thecross-reactivity of proteasomes of one species with antibody to proteasomes of another species was different from the reverse cross-reactivity suggested that the localizations of antigenic domains, that is, the subunit organizations, in various proteasomes are different. The resemblance of the gross structures of different proteasomes suggests a common mechanism of their subunit organization. The similarity in subunit size may be important for construction of a symmetrical complex, and theheterogeneity of the netcharges of the subunits is presumably related to the assembly of the large multisubunit complex by complementary charge interaction. Thus, the size similarity and charge variation in the subunits may be important for conservation of the proteasomes during evolution, even though the individual components are altered. For further elucidation of the species-specificity insubunit multiplicity of proteasomes, more detailed studies, such as analyses of their gene structures, are required. Identity of Proteasomes with Ring-shaped 20 S Particles Found in Various Eukaryotic Cells-Proteasomes are structurally unique in size, shape, and subunit multiplicity. Interestingly, these multi-protease complexes are quite similar to the subcellular 19 S to 22 S particles found during the past 15 years in a wide variety of eukaryotic cells, such as plants (32,38), Drosophila (24-27), Xenopus (28,29),seaurchin (30), birds (31,38), and various mammals (31-38). So far it has been very difficult to determine their similarity or identity, since they have been isolated from different cells and tissues and in different ways, and thebiological functions of most of these particles are unknown. Recently, we found that human prosomes are identical to proteasomes, judging from their immunological cross-reactivity, morphological appearance by electron microscopy, polypeptide composition, and proteolytic activities (20). Falkenburg et al. (39) also reported the identity of Drosophila 19 S ribonucleoproteins particles with rat muscle multicatalytic proteinase. In the present work, we demonstrated that proteasomes were distributed in a variety of eukaryotic cells. This finding is consistent with the observation of the ubiquitous &-t 13 rlbution . of these 19 S to 20 S particles (31, 38). Thus, all

these particles are thought to be identical to proteasomes. However, there areseveral reports of similarity in morphology but differences in homology of these particles (25, 29, 31); namely differences in subunitcomposition and immunological reactivity of apparently similar particles. Therefore, it has sometimes been claimed that these particles are closely related, but not identical. In the present work, we found that proteasomes from a variety of eukaryotic cells exhibit interspecies variationsinsubunit compositions (Figs. 3-5) and immunological reactivities (Figs. 7 and 8 ) , suggesting that the apparently discrepant results reported previously are attribare nonidentical, as judged by peptide mapping and cell-free utable to minor structural variations of homologous moletranslation of mRNA (19). Thus, otherproteasomes are prob- cules, which may be acquired during evolution. Thus all the ably also composed of multiple, nonidentical subunits. The 20 S particles described previously can be regarded as proteamultiplicity of subunits of similar size but with widely differ- somes showing minor differences. ent PI values was a common feature of the proteasomes, but There are, however, major differences between the ringthe subunits showed species-specific differences (Figs. 3-5). shaped particles reported in theliterature and thoseobserved Moreover, purified proteasomes of various species did not in this work. Prosomes and 19 S ring-type particles containing cross-react on Ouchterlony immunodiffusion analysis (Fig. 7), heat-shock proteins have been considered to be ribonucleoshowing that theimmunoreactivities of the proteasomes were protein particles that have significant amounts (15-20% of species-specific. However, some components of proteasomes the total molecular weight) of specific RNA species (25, 27, showed partial cross-reactivity on immunoblotting analyses 31, 38). In contrast, the proteasomes demonstrated by us (2,

The Proteasones (Eukaryotic Multi-protease Complexes)

16216 Human Liver

LiverRat

Chicken Liver

1

4.

4

Xenopus Laevis Ovary

1

1

4

4

Yeast

4- .,

FIG.7. Ouchterlony double-diffusion analysis of proteasomes from various eukaryotes. Samples of 10 pg of the enzymes from the various sources indicatedat the topof each panel were placed in the center well ( E ) . Anti-serum (7 pl) to samples from human liver ( I ) , rat liver (2), chicken liver (3), X. laevis ovary (4), and yeast ( 5 )and nonimmunized control serum(6)were placed in theside wells. Samples were incubated for 2 days a t room temperature in a humidified chamber, washed for 4 days with phosphate-buffered saline with gentle shaking, and then stained withCoomassie Brilliant Blue. concentration to prevent artificial adsorption of RNA contaminants, whereas particles without RNAs have been purified by more drastic ion-exchange and hydroxylapatite chromatographic techniques. Thus particles per se may consist of a core part, in which the subunits arevery tightly assembled, and additional components that bind loosely to thecore part and are probably lost during drastic purification procedures. Functionally, the core part may be a multi-enzyme complex containing proteaseand probably nuclease, and theadditional part may contain some factors such as small RNA moieties B and/or certain heat-shock proteins, which presumably play a regulatory role in the enzymatic functions of the core part. Thus, thesubcellular 20 S particles in a variety of eukaryotic cells must be homologous. In this paper, we describe proteasomes only as multi-protease complexes; that is, particles with protease activities. However, these enzymes may also have other as yet unknown functions. Such functions are suggested from the structural similarity of proteasomes to many kinds of 20 S particles. For example, some functions of 20 S ring-shaped particles have D been reported. The ribonucleoproten particles named “prosomes,” have been characterized and arebelieved to function in the regulation of translation, because they are co-purified with the inactive mRNA complex (24) and because their RNA moiety has a common complementary sequence in different mRNAs (31). In addition, in Drosophila, 19 S ring-shaped particles are tightly associated with polyribosomes (27). Moreover, 19 S cytoplasmic particles have been shown to have small heat-shock proteins and tobe identical to theprosomes in Drosophila (25,26). Furthermore, Castano et al. (29) found F that similar particlesare co-purified as a pre-tRNAprocessing nuclease in X. lueuis. These results suggest that the 20 S particles contain small RNA species and that certain heatshock proteins may function in RNA processing, in the regFIG.8. Immunoelectrophoretic blot analysis of various pro- ulation of protein synthesis,and in the response of eukaryotic teasomes. Immunoblotting analysis was performed as described under “Experimental Procedures.” Lanes 1-5 show the electrophoretic cells to environmental stress; in other words, proteasomes patterns of the enzymesfrom human, rat, and chickenliver, X.luevis may be multifunctional enzyme complexes with both protease ovary, and yeast, respectively. Panel A shows the patternsof various and nuclease activities and may function in both RNA and proteasomes stained withCoomassie Brilliant Blue. Panels B-F show protein metabolism in all eukaryotic cells. The restriction of the patterns of various proteasomes after immunostaining with anmany biologically essential functions to one particle may be tibodies against the enzymes from human, rat, and chickenliver, X. very efficient for cellular activities. laevis ovary, and yeast, respectively.

1

2

3

4

5

E

19, and thispaper) and 5’-pre-tRNAase (29)and 22 S particles from X. laeuis (28) did not contain any RNA. These differences seem to be dueto differences in the procedures used for purification of the particles: particles containing small RNA moieties have been isolated by the mild procedure of sucrose density gradient centrifugation inthe presence of a high salt

REFERENCES 1. Tanaka, K., Ii, K., Ichihara, A., Waxman, L., and Goldberg, A. L.(1986) J. BWl. Chem. 261,15197-15203 2. Tanaka, K., Yoshimura, T., Ichihara, A., Kameyama, K., and Takagi, T. (1986) J. BWl. Chem. 261,15204-15207 3. DeMartino, G. N., and Goldberg, A. L.(1979) J. Biol. Chern. 254, 3712-3715

The Proteasomes (Eukaryotic: Multi-protease Complexes) 4. Rose, I. A., Warms, J. V. B., and Hershko, A. (1979) J. Biol. Chem. 254,8135-8138 5. Wilk, S., and Orlowski, M. (1980) J. Neurochem. 3 5 , 1172-1182 6. Edmunds, T., and Pennington, R. J. T. (1982) Znt. J. Biochem. 14,701-703 7. Ismail, F., and Gevers, W. (1983) Biochim. Biophys. Acta 7 4 2 , 399-408 8. Wilk, S., and Orlowski, M. (1983) J. Neurochem. 40,842-849 9. Rivett, A. J. (1985) J. Biol. Chem. 260, 12600-12606 10. Ray, K., and Harris, H. (1985) Proc. Natl. Acad. Sci. U. S. A. 8 2 , 7545-7549 11. Dahlmann, B., Kuehn, L., Rutschmann, M., and Reinauer, H. (1985) Biochem. J. 2 2 8 , 161-170 12. Dahlmann, B., Rutschmann, M., Kuehn, L., and Reinauer, H. (1985) Biochem. J. 228,171-177 13. McGuire, M. J., and DeMartino, G. N. (1986) Biochim. Biophys. Acta 873,279-289 14. Nojima, M., Ishiura, S., Yamamoto, T., Okuyama, T., Furuya, H., and Sugita, H. (1986) J. Biochem. (Tokyo) 99,1605-1611 15. Zolfaghari, R., Baker, C. R. F., Jr., Canizaro, P. C., Amirgholami, A., and BBhal, F. J. (1987) Biochem. J. 2 4 1 , 129-135 16. Hough, R., Pratt, G., and Rechsteiner, M. (1987) J. Biol. Chem. 262,8303-8313 17. Waxman, L., Fagan, J. M., and Goldberg, A. L. (1987) J. Biol. Chem. 262,2451-2457 18. Kopp, F., Steiner, R., Dahlmann, B., Kuehn, L., and Reinauer, H. (1986) Biochim. Biophys. Acta 8 7 2 , 253-260 19. Tanaka, K., Yoshimura, T., Ichihara, A., Ikai, A., Nishigai, M., Morimoto, M., Sato, M., Tanaka, N., Katsube, Y., Kameyama, K., and Takagi, T. (1988) J. Mol. Biol. 2 0 1 , in press 20. Arrigo,A.-P., Tanaka, K., Goldberg, A. L., and Welch, W. J. (1988) Nature 331, 192-194 21. Hase, J., Kobashi, K., Nahai, N., Mitsui, K., Iwata, K., and Takadera, T. (1980) Biochim. Biophys. Acta 6 1 1,205-213 22. Toyohara, H., Nomata, H., Makinodan, Y., and Shimizu, Y. (1987) Comp. Biochem. Physiol. 8 6 B , 99-102 23. Achstetter, T., Ehmann, C., Osaki, A., and Wolf, D. H. (1984) J. Biol. Chem. 259,13344-13348 24. Schmid, H. P., Akhayat, O., Martins de Sa, C., Puvion, F., Koehler, K., and Scherrer, K. (1984) EMBO J. 3 , 29-34

16217

25. Schuldt, C., and Kloetzel, P.-M. (1985) Deu. Biol. 1 1 0 , 65-74 26. Arrigo,A.-P., Darlix, J.-L., Khandjian, E. W., Simon, M., and Spahr, P.-F. (1985) EMBO J. 4,399-406 27. Kloetzel, P.-F., Falkenburg, P.-E., Hossl, P., and Glatzer, K. H. (1987) Exp. Cell Res. 1 7 0 , 204-213 28. Kleinschmidt, J. A., Hugle, B., Grund, C., and Franke, W.W. (1983) Eur. J. Cell Biol. 3 2 , 143-156 29. Castano, J. G., Ornberg, R., Koster, J. G., Tobian, J. A., and Zasloff, M. (1986) Cell 46,377-387 30. Akhayat, O., Grossi de Sa, F., and Infante, A.A. (1987) Proc. Natl. Acad. Sci. U. S. A. 8 4 , 1595-1599 31. Martins de Sa, C., Grossi de Sa, M.-F., Akhayat, O., Broders, F., and Scherrer, K. (1986) J. Mol. Biol. 187,479-493 32. Shelton, E., Kuff, E. L., Maxwell, E. S., and Harrington, J. T. (1970) J. Cell Biol. 4 5 , 1-8 33. Narayan, K. S., and Rounds, D. E. (1973) Nature New Biol. 2 4 3 , 146-150 34. Smulson, M. (1974) Exp. Cell Res. 8 7 , 253-258 35. Harris, J. R., and Naeem, I. (1981) Biochim. Biophys. Acta 6 7 0 , 285-290 36. Malech, H. L., and Marchesi, V. T. (1981) Biochim. Biophys. Acta 670,385-392 37. Domae, N., Harmon, F. R., Busch, R. K., Spohn, W., Subrahmanyam, C. S., and Busch, H. (1982) Life Sci. 30,469-477 38. Arrigo, A.-P., Simon, M., Darlix, J.-L., and Spahr, P.-F. (1987) J. Mol. Euol. 2 5 , 141-150 39. Falkenburg, P.-E., Haass, C., Kloetzel, P.-M., Niedel, B., Kopp, F., Kuehn, L., and Dahlmann, B. (1988) Nature 3 3 1 , 190-192 40. Laemmli, U. K. (1970) Nature 227,680-685 41. O’Farrell, P. H. (1975) J. Biol. Chem. 250,40074021 42. Ouchterlony, 0. (1967) in Handbook of ExperimentalZmmunology (Weir, C., ed) pp. 655-688, Blackwell Scientific Publications, Oxford 43. Towbin, H. S., Staehelin, J., and Gordon, J. (1979) Proc. Natl. Acad. Sci. U.S. A. 7 6 , 4350-4354 44. Bradford, M. M. (1976) Anal. Biochem. 7 2 , 248-254 45. Lodola, A., Sprogy, S. P., and Holbrook, J. J. (1978) Biochem. J. 169,577-588 46. Goldberg, A. L., and St. John,A. C. (1976) Annu. Reu. Biochem. 45,747-803