Evidence for the existence of both proteasomes and a novel high molecular weight peptidase in Entamoeba histolytica.

To screen for high molecular weight proteases in Entamoeba histolytica, we subjected a soluble amebal extract to density gradient centrifugation and tested the fractions for activity against the chymotryptic peptide substrate, Suc-leucyl-leucyl-valyl-tyrosyl-4-methylcoumaryl-7-amide. Two peaks of activity, of approximately 11 and 20 S, were clearly separated. Based on SDS-electrophoretic pattern and immunoblot analysis, we ascribe the 20 S activity to proteasomes. The 11 S protein was purified from amebal homogenates by a series of chromatographic steps. As determined by molecular sieve chromatography and nondenaturing gel electrophoresis, the native complex had an apparent Mr of 385,000 +/- 10%. On SDS gels, the purified enzyme exhibited a single band of Mr 62,000 that yielded a single N-terminal sequence, indicating that the preparation was homogeneous and that the native complex consisted of six very similar or identical subunits. The enzyme preferred peptides with aromatic residues at the P1 position and had low but distinct activity toward azocasein. We conclude that the 11 S enzyme is a novel high molecular weight protease that is distinct from proteasomes.

To screen for high molecular weight proteases in Entamoeba histolytica, we subjected a soluble amebal extract to density gradient centrifugation and tested the fractions for activity against the chymotryptic peptide substrate, Suc-leucyl-leucyl-valyl-tyrosyl-4-methylcoumaryl-7-amide. Two peaks of activity, of approximately 11 and 20 S, were clearly separated. Based on SDS-electrophoretic pattern and immunoblot analysis, we ascribe the 20 S activity to proteasomes. The 11 S protein was purified from amebal homogenates by a series of chromatographic steps. As determined by molecular sieve chromatography and nondenaturing gel electrophoresis, the native complex had an apparent M r of 385,000 ؎ 10%. On SDS gels, the purified enzyme exhibited a single band of M r 62,000 that yielded a single N-terminal sequence, indicating that the preparation was homogeneous and that the native complex consisted of six very similar or identical subunits. The enzyme preferred peptides with aromatic residues at the P 1 position and had low but distinct activity toward azocasein. We conclude that the 11 S enzyme is a novel high molecular weight protease that is distinct from proteasomes.
Entamoeba histolytica is a parasitic protozoon that resides in the human gut. It frequently occurs in developing countries, causes amebic dysentery, and may lead to the formation of tumor-like abscesses in liver and spleen (1). As to its cell biology, E. histolytica is a low eukaryote that lacks mitochondria and a well defined endoplasmatic reticulum/Golgi apparatus. Morphologically conspicuous is the enormous amount of vacuoles in the amebae. These occupy about 40% of the total cell volume and are functionally equivalent to both the lysosomes and the cytotoxic vesicles of higher eukaryotic cells (2,3). Besides its function in cellular metabolism, protein degradation is essential for a range of regulatory processes and for the elimination of unstable or abnormal proteins (4). In the last decade, new insights have emerged into the mechanisms of intracellular protein metabolism. Generally, there appear to be two major pathways for protein degradation: one is lysosomal and employs the action of thiol-and aspartyl-dependent cathepsins (5), and the other is cytosolic and functions with the aid of high molecular weight proteases. The latter are represented by the Lon and ClpAP proteases in prokaryotes and the 20 S proteasome or multicatalytic protease (MCP), 1 which is the proteolytic core of a 26 S protease, in eukaryotes (4,6). The 20 S proteasome has also been found in some Archaea (7) and in at least one Eubacterium (8). Protein degradation by Lon, ClpAP, and the 26 S protease is ATP dependent, and protein substrates generally have to be ubiquitinated prior to degradation by the 26 S protease (9). By contrast, small peptides are cleaved by Lon, ClpP, and the 20 S protease in the absence of ATP and, as to the latter, without the need for ubiquitin tagging (10). Consequently, fluorogenic peptides have been used for screening purposes and as model substrates to probe the specificity of these high molecular weight proteases (11,12). For proteasomes, the pattern of activity toward the different substrates is variable. In particular, upon stimulation by ␥-interferon, mammalian proteasomes exhibit an increase in chymotryptic activity, which may serve to generate peptides for antigen presentation and appears to be due both to the replacement of certain proteasomal ␤-subunits by others and to an enhanced response to activator protein (6,13,14). The threedimensional structure of an archaeal 20 S proteasome, which should be similar to its eukaryotic counterpart, has recently been solved (7). It consists of four stacked rings with seven subunits each: two rings of 22.3-kDa ␤-subunits sandwiched between two rings of 25.8-kDa ␣-subunits. Remarkably, the bacterial ClpP seems to possess a similar 7-fold symmetry (15). The active site residue of the proteasome has been identified as a ␤-chain threonine both by site-directed mutagenesis (16) and by N-terminal modification (17).
In E. histolytica, a range of cysteine proteinases of the papain type has been extensively characterized both on the genomic and on the protein level (18,19). These enzymes are localized in the lysosome-like vacuoles mentioned above. By contrast, nothing is known yet about cytoplasmic protein degradation in this organism or about the proteases involved. In this study, we present evidence indicating that E. histolytica contains both proteasomes and a novel, unrelated high molecular weight protease.

EXPERIMENTAL PROCEDURES
Cell Culture-E. histolytica, strain HM1:IMSS, was cultured at 37°C in TYI-S-33 medium (20) supplemented with 100 units/ml penicillin and 100 g/ml streptomycin. After 3 days, cells were chilled on ice and centrifuged at 300 ϫ g for 10 min at 4°C. Cells were resuspended in phosphate-buffered saline containing the cysteine proteinase inhibitor E-64 (final concentration, 0.1 M) and disrupted by sonication. The homogenate was centrifuged at 100,000 ϫ g for 1 h at 4°C, and the supernatant (soluble extract) was used for further experiments.
Peptidase and Protease Assays-Routinely, peptidolytic activity was * This study was supported by the Deutsche Forschungsgemeinschaft and by the Fonds der Chemischen Industrie. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 (21). In a fluorescence cuvette, 10 l of substrate (5 mM dissolved in dimethyl sulfoxide), up to 40 l of 1% (w/v) SDS, and 5-20 l of sample were added to 1 ml of 10 mM Tris-HCl, pH 7.8, at 25°C. The rate of release of 7-amino-4-methylcoumarin was determined by following the increase of fluorescence at 380 nm excitation and 460 nm emission. The cleavage of N-t-Boc-LSTR-MCA was determined under identical conditions. With benzyloxycarbonyl-leucyl-leucyl-glutamyl-2-naphthylamide as substrate (end concentration, 0.1 mM), wavelengths were adjusted to 333 nm excitation and 410 nm emission, respectively (11). Proteolytic activity was determined at 37°C with azocasein (final concentration, 10 mg/ml) as a substrate (22).
Density Gradient Centrifugation-Sucrose density gradient centrifugation was performed in 37-ml polyallomer tubes using a swing rotor (SW 28, Beckman). Soluble extract containing E-64 was loaded onto a linear gradient of 15-40% sucrose and centrifuged at 27,000 rpm for 42 h at 4°C. Fractions of 1 ml were collected and analyzed for enzymatic activity and protein content.
Purification of the 11 S Peptidase-Soluble extract containing E-64 was applied onto a column of DEAE-cellulose (Whatman; bed volume, 30 ml) equilibrated with 10 mM Tris-HCl, pH 7.8, and eluted with a linear gradient of 0 -0.5 M NaCl in the same buffer. Suc-LLVY-MCA hydrolyzing activity was desorbed at about 220 mM NaCl (Fig. 1A). Active fractions were pooled, concentrated by ultrafiltration (Filtron), and subjected to gel filtration by fast protein liquid chromatography using Sephacryl S-300 (Pharmacia Biotech Inc.; bed volume, 120 ml) in 10 mM Tris-HCl, pH 7.8, containing 0.1 M NaCl. As shown in Fig. 1B, fractions containing peptide-hydrolyzing activity eluted as a major peak followed by a second peak of lower activity at higher retention time. The active fractions of the major peak were equilibrated to 10 mM Tris-HCl, pH 7.8, by ultrafiltration and dilution and then submitted to fast protein liquid chromatography anion exchange chromatography over a MonoQ column equilibrated with the same buffer. The protease was eluted by a broken linear gradient of 0 -0.5 M NaCl as indicated in Fig.  1C and a flow rate of 0.5 ml/min over a time range of 55 min. Fractions containing Suc-LLVY-MCA hydrolyzing activity, which desorbed at about 240 mM NaCl, were pooled, concentrated by ultrafiltration, and chromatographed twice on a Superose 6 column (bed volume, 24 ml) with 10 mM Tris-HCl, pH 7.8, containing 0.1 M NaCl. As shown in Fig.  1D, the Suc-LLVY-MCA cleaving fractions corresponded to a distinct peak of protein absorption and were clearly separated from a residual impurity of lower retention time. These fractions were concentrated and stored at Ϫ20°C in 10 mM Tris-HCl, pH 7.8.
Analytical Molecular Sieve Chromatography-The apparent molecular weight of the native 11 S enzyme was determined by molecular sieve chromatography over Sephacryl S-300 after calibration of the column with the high molecular weight standards thyroglobulin (M r 669,000), apoferritin (M r 443,000), catalase (M r 240,000), and aldolase (M r 158,000).
Electron Microscopy-Drops of the purified 11 S protease were applied to carbon film-covered glow-discharged grids and negatively stained with aqueous 1% uranyl acetate (23). Electron micrographs were recorded with a Philips EM 420.
Gel Electrophoreses-SDS-PAGE was carried out in 15% polyacrylamide gels according to Ref. 24. Native gels (7.5% polyacrylamide) were run at 4°C with a Tris/glycine system, pH 8.9, according to Ref. 25. The Suc-LLVY-MCA hydrolyzing activity was detected by overlaying the gel strip with 2 ml of 50 mM Tris-HCl buffer, pH 7.8, containing 40 l of a 5 mM substrate solution, and visualizing the fluorescence of the released 7-amino-4-methylcoumarin on a UV fluorescent table equipped with a 354-nm lamp.
N-terminal Sequencing-The purified 11 S enzyme was subjected to SDS-PAGE and blotted on polyvinylidene difluoride membrane (Bio-Rad). The protein band was cut out and N-terminally sequenced on an Applied Biosystems 473A protein sequencer.
Immunoblotting-Gels were blotted on nitrocellulose sheets in 0.1 M CAPS-HCl, pH 11.5, and the immunoreaction was performed according to a protocol of Hoefer Scientific Instruments. Antibodies against proteasomes of Thermoplasma acidophilum, Dictyostelium discoideum, and rat proteasome were kindly provided by A. Grziwa and M. Nesper (Martinsried) and B. Dahlmann (Dü sseldorf), respectively. The MCP231 antibody against proteasome ␣-chains (26) was a generous gift of K. Hendil (Copenhagen).

Identification of Two High Molecular Weight Proteases-To
test for the presence of non-thiol proteases in E. histolytica, a soluble amebic extract (100,000 ϫ g supernatant) containing the cysteine proteinase inhibitor E-64 was incubated with the fluorogenic chymotrypsin substrate, Suc-LLVY-MCA. A release of fluorophore could readily be observed. As shown in Fig. 2, this peptidolytic activity was stimulated by over 60% by SDS, with an optimum at an SDS concentration of about 0.4 mg/ml. Beyond the optimum, the curve exhibited a flat declining plateau and then decreased steeply at concentrations exceeding 0.8 mg/ml SDS. Stimulation of peptidolytic activity by SDS has typically been found for proteasomes (27). We therefore set out to investigate whether these complexes were present in E. histolytica. To this end, a soluble amebal extract was subjected to sucrose density gradient centrifugation, and the fractions were tested for chymotryptic activity. As shown in Fig. 3 (closed circles), two Suc-LLVY hydrolyzing peaks with sedimentation velocities of approximately 11 and 20 S were clearly separated. Only the second peak exhibited significant activity toward the tryptic substrate, BOC-LSTR-MCA (Fig. 3, open squares). The SDSelectrophoretic pattern of the 20 S peak (Fig. 3, inset A) was restricted to a series of bands between 25 and 30 kDa and roughly corresponded to that found for proteasomes from a number of organisms (21), whereas the 11 S fraction exhibited bands over a much wider range (data not shown) and obviously contained a crude mixture of proteins. In immunoblot experiments, we tested the reactions of a cell homogenate and of the 11 and 20 S peaks with an antibody (MCP231) against a sequence motif common to ␣-chains of proteasomes (26). As shown in Fig. 3, inset B, the 20 S peak (lane 3) exhibited a series of cross-reacting bands around 30 kDa. By contrast, neither the 11 S fraction (lane 2) nor the purified 11 S protease (lane 1; see below) cross-reacted with the antibody. These findings indicated that E. histolytica contained both 20 S proteasomes and an 11 S peptidase.
Purification and Subunit Composition of the 11 S Peptidase-At this stage, it seemed possible that the 11 S enzyme was a catalytically active proteasomal fragment that lacked ␣-chains; alternatively, it could be a distinct protease that had nothing to do with proteasomes. To distinguish between these two possibilities, we set out to purify and characterize the 11 S enzyme. Purification was achieved by two anion exchange and two molecular sieve chromatography steps ( Fig. 1; Table I).
Overall, specific activity toward the chymotrypsin peptide substrate went up by a factor 320, and 730 mg of protein (from approximately 5 ϫ 10 8 amebae) yielded about 100 g of peptidase (Table I). To test for the purity of the isolated enzyme, an aliquot was subjected to SDS-PAGE. Under these conditions, the protein appeared as a single band with an apparent M r of 62,000 (Fig. 4A, right) and yielded a single N-terminal sequence; this sequence (DNXVNVKNQLS) did not exhibit sig- nificant similarity to any other N-terminal sequence in the EMBL data base. Under non-denaturing conditions, the enzyme migrated as a single, peptidolytically active complex between the marker proteins catalase (M r 240,000) and apoferritin (M r 443,000; Fig. 4A, left). In agreement both with this observation and with its sedimentation constant, we determined its apparent M r by molecular sieve chromatography as 385,000 Ϯ 10% (n ϭ 3; see "Experimental Procedures"). Finally, under weakly denaturing conditions (1 mg/ml SDS, no boiling or reduction) the protein, which was still peptidolytically active, exhibited an apparent M r of about 180,000 (Fig. 4A, middle). From these data, we infer that the native protease is probably composed of six identical or very similar subunits of 60 -65 kDa. Under weakly denaturing conditions, three of these subunits remained associated as a metastable, active trimer half the size of the native complex. A cartoon illustrating these inferred relationships is given in Fig. 4B. An electron micrograph of the purified 11 S protease revealed compact particles with an average diameter of about 9 nm (Fig. 5).
Substrate Specificity and Inhibitor Profile of the 11 S Pepti-dase-Tables II, lII, and IV summarize some kinetic properties of the 11 S enzyme. Table II lists its relative activity toward a range of fluorogenic substrates. It can be seen that the enzyme had relatively high chymotryptic activity (i.e. it preferred aromatic residues at the P 1 position), much lower trypsin-like activity (Arg at the P 1 position; see also Fig. 1), and did not accept any of the other peptides offered as a substrate. In particular, whereas the rough extract had considerable peptidylglutamyl peptidase activity (up to 20% of the chymotryptic activity; data not shown), the purified enzyme lacked this activity altogether. Table III presents the kinetic parameters of the 11 S enzyme toward the chymotrypsin substrate. Both K m and V max were within the range reported for proteasomes (10,13). The 11 S enzyme also acted as a protease, albeit a rather inefficient one (Table IV). It shares this latter property with 20 S proteasomes (10); we suppose that, as with proteasomes, its proteolytic activity may in vivo be activated by accessory proteins. The pH optimum of both peptidolytic and proteolytic activity (a broad optimum around 6.0 -7.0; data not shown) was lower than that of proteasomes (Ն8; e.g. Ref. 28,Fig. 5). In agreement with its specificity, the enzyme was completely inhibited by chymostatin, a chymotrypsin inhibitor (Table V). Also the calpain inhibitor I was very effective, much more so     than the other inhibitors of serine and cysteine proteases that we tested. Whereas EDTA and ATP had no or little effect, enzyme activity increased by about 40% in the presence of MgCl 2 or CaCl 2 (both at 10 mM; Table V). Under all conditions tested, activity was increased up to 2-fold by the addition of 0.4 mg/ml SDS (data not shown).

DISCUSSION
In this study, we have identified two high molecular weight (11 and 20 S) proteases in E. histolytica. The heavier enzyme appears proteasome-like both from its subunit composition and from its cross-reactivity with an antibody against proteasomal ␣-subunits. By contrast, the lighter enzyme exhibited several novel characteristics. Clearly, its activity cannot be ascribed to the vesicular cysteine proteinases described in a series of earlier reports (18,19) because it was not inhibited by E-64, a specific inhibitor of these enzymes. In addition, its high molecular weight and its substrate specificity (preferred cleavage of the peptide bond of hydrophobic residues at the P 1 position) distinguish this enzyme from the known amebic proteases histolysain and amebapain, which favor arginine in P 1 and P 2 (28). Also, although peptidolytic activity was effectively inhibited by a calpain inhibitor and stimulated by Ca 2ϩ (Table V), we obviously were not dealing with a calpain, as these enzymes have a very different subunit composition and are blocked by EDTA and E-64 (29). We originally thought the 11 S enzyme might correspond to part (specifically, to the ␤-core) of the proteasome. Indeed, apart from its complete lack of peptidylglutamyl activity, the substrate specificity, kinetics, and effector profile of the 11 S enzyme (Tables II-V) were not that different from those found for proteasomes; SDS at low concentrations stimulated activity to a comparable degree, and in electron-optical images the 11 S complex had a similar diameter (Fig. 5). However, a very strong argument against the 11 S enzyme being part of the proteasome is its radically different subunit composition (identical or very similar 62-kDa subunits rather than a range of bands between 25 and 30 kDa). Of course, we cannot at present exclude a distant relationship between the two complexes. For instance, the 62-kDa subunits might be the product of natural gene duplications of sequences coding for far relatives (with deviating N-terminal ends) of proteasomal ␤-chains. However, the inferred symmetry of the 11 S particle (3-fold rather than 7-fold, see Fig. 4B) would seem to argue even against a distant relationship. Evidence has recently been reported (30) that Trypanosoma brucei, another parasitic protozoon that branched off relatively early from the main eukaryotic line (31), contains a high molecular weight protease that migrates faster than mammalian 20 S proteasomes and that exhibits a deviating substrate specificity. This enzyme could possibly be related to the E. histolytica 11 S protease.