Blastocystis Legumain Is Localized on the Cell Surface, and Specific Inhibition of Its Activity Implicates a Pro-survival Role for the Enzyme*

Programmed cell death (PCD) is crucial for cellular growth and development in multicellular organisms. Although distinct PCD features have been described for unicellular eukaryotes, homology searches have failed to reveal clear PCD-related orthologues among these organisms. Our previous studies revealed that a surface-reactive monoclonal antibody (mAb) 1D5 could induce multiple PCD pathways in the protozoan Blastocystis. In this study, we identified, by two-dimensional gel electrophoresis and mass spectrometry, the target of mAb 1D5 as a surface-localized legumain, an asparagine endopeptidase that is usually found in lysosomal/acidic compartments of other organisms. Recombinant Blastocystis legumain displayed biphasic pH optima in substrate assays, with peaks at pH 4 and 7.5. Activity of Blastocystis legumain was greatly inhibited by the legumain-specific inhibitor carbobenzyloxy-Ala-Ala-AAsn-epoxycarboxylate ethyl ester (APE-RR) (where AAsn is aza-asparagine) and moderately inhibited by mAb 1D5, cystatin, and caspase-1 inhibitor. Interestingly, inhibition of legumain activity induced PCD in Blastocystis, observed by increased externalization of phosphatidylserine residues and in situ DNA fragmentation. In contrast to plants, in which legumains have been shown to play a pro-death role, legumain appears to display a pro-survival role in Blastocystis.

Programmed cell death in the unicellular protozoa is now accepted as a well established phenomenon. Several stereotypic apoptotic morphological markers similar to those observed in apoptotic metazoan cells have been described in human parasitic protozoa such as Leishmania amazonensis, Leishmania donovani, Trypanosoma cruzi, Trypanosoma brucei, Trypanosoma rhodesiense, Plasmodium falciparum, and Blastocystis (1,2). Despite a wealth of information on the organelles and cytochemical features involved in protozoan PCD, there is a scarcity of information on PCD 3 -related molecular mediators. Our earlier studies showed that Blastocystis undergoes programmed cell death when exposed to the surface-reactive monoclonal antibody mAb 1D5 with typical features of apoptotic cells (3,4). mAb 1D5 was shown to target a 30-kDa protein found on the plasma membrane of Blastocystis (5)(6)(7). This protein is functionally important, but not all cells within a clonal population would be susceptible to the cytotoxic effects of mAb 1D5 (6). These results suggest that this protein may have multiple localizations and is potentially important for cell survival. In this study, an asparaginyl cysteine protease legumain was identified as the mAb 1D5 targeting protein. Legumain is a recently described lysosomal protease, well conserved and present in plants, mammals, helminth worms, and the protozoan Trichomonas vaginalis (8 -12). The active site of legumain contains the catalytic dyad His-Gly-spacer-Ala-Cys, a characteristic shared with caspases, aspartyl cysteine proteases important as molecular mediators of apoptosis cascades (13). Legumains have specificity for the hydrolysis of bonds on the carboxyl side of asparagines (14) but have also been reported to cleave after aspartate residues, evidenced by their ability to cleave caspase-1 substrates (15,16). In plants, legumain, also termed vacuolar processing enzyme (VPE), is localized within the plant cell vacuole and is essential for hypersensitive-response cell death, a mechanism of resistance to pathogen infection (15). In mammals, it is predominantly localized in the lysosomes and involved in processing bacterial and endogenous peptides for major histocompatibility complex class II presentation. Legumain is overexpressed in many types of tumors and is associated with enhanced tissue invasion and metastasis. In these cases, the enzyme can also be found on the cell surface, co-localized with integrin ␤1 (17).
In this study, we report for the first time the detailed characterization of Blastocystis legumain protein as a target of a pre-* This work was supported by Biomedical Research Council Grant 05/1/21/ 19/378, the National Medical Research Council (06May107), and the Infectious Disease Program. □ S The on-line version of this article (available at http://www.jbc.org) contains supplemental Figs. S1 and S2. The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) ACO24555 and GU124590. 1 Both authors contributed equally to this work. 2 To whom correspondence should be addressed: 5  viously described cytotoxic mAb. Blastocystis legumain was expressed in Escherichia coli, and the purified recombinant protein was used to define enzymatic properties. Cellular localization of this protein was investigated by anti-legumain antibody and mAb 1D5. The involvement of legumain in programmed cell death was investigated by protease inhibitor assays, annexin V, and TUNEL apoptosis assays to determine the extent of phosphatidylserine externalization and DNA fragmentation, respectively. The data strongly suggest that Blastocystis legumain has a key role in the regulation of Blastocystis cell death.

EXPERIMENTAL PROCEDURES
Parasite Cell Culture-Blastocystis isolate B (subtype 7) was isolated from the stool of a local patient. Parasite cells was cultured in Iscove's modified Dulbecco's medium containing 10% inactivated horse serum and incubated anaerobically at 37°C in an Anaerojar (Oxoid). 4-Day-old cells at log phase were used for all experiments. Cell number was counted with a hemocytometer, and their viability was examined morphologically by a light microscope.
Two-dimensional Electrophoresis-Parasite lysate was prepared by three freeze-thaw cycles in liquid nitrogen and a 37°C water bath. Cell debris was removed by centrifugation at 14,500 ϫ g for 10 min at 4°C. Total proteins from clarified cell lysate were precipitated by 10% trichloroacetic acid. The dried pellet was dissolved in 20 mM Tris-Cl buffer, pH 7.0, and protein concentration was determined by the Bradford assay (Sigma). The first dimension IEFs were performed on 18-cm IEF gel strips (pH 4 -7, Bio-Rad) following the manufacturer's instructions. Briefly, the dried strips were rehydrated at room temperature overnight with 375 l of rehydration buffer (7 M urea, 2 M thiourea, 4% CHAPS, 50 mM dithiothreitol, 3% IPG buffer, 0.02% bromphenol blue) and 50 g of protein. IEF strips were loaded onto a Protean IEF cell (Bio-Rad) by using the following parameters: 500 V for 500 V-h; 1000 V for 500 V-h; 8000 V gradient for 4500 V-h; and 8000 V for 64,000 V-h. Subsequently, strips were soaked in 10 ml of equilibration solution (7 M urea, 0.05 M Tris-Cl, pH 6.8, 30% glycerol, 2% SDS) supplied with 0.1% dithiothreitol for 15 min, removed, and further soaked in equilibration solution containing 0.25% iodoacetamide and 0.02% bromphenol blue for 15 min. Strips were laid on top of a 10% SDS-polyacrylamide gel (Protean II XI, Bio-Rad), and electrophoresis was performed at 5°C. Current was kept constant at 7 mA per gel for 1 h and then increased to 25 mA per gel for another 6 h. Protein was visualized by Coomassie Brilliant Blue staining.
Protein Identification-A gel slice encompassing the protein of interest was cut out, soaked in 200 l of 50% (v/v) acetonitrile with 50 mM ammonium bicarbonate, and incubated at 37°C for 30 min to destain. The gel slice was dried for 5 min using a SpeedVac concentrator and soaked in 20 l of trypsin (3.33 ng l Ϫ1 , diluted in 50 mM ammonium bicarbonate) at 30°C overnight or 37°C for 4 h. Peptides were extracted using 50% (v/v) acetonitrile with 0.5% (v/v) trifluoroacetic acid (200 l) followed by sonication (10 min) at 37°C. The solution was dried using a SpeedVac concentrator and redissolved in 50% (v/v) acetonitrile with 0.5% (v/v) trifluoroacetic acid (1.5 l). Pep-tides were desalted using a ZipTip (Millipore) and mixed with an equal volume of matrix solution (␣-cyano-4-hydroxycinnamic acid (5 mg) dissolved in 50% (v/v) acetonitrile with 0.5% (v/v) trifluoroacetic acid (1 ml)) and spotted (1 l) onto a MALDI plate (384 Opti-TOF, Applied Biosystems). After air drying, spots were identified by a combination of mass spectrometry followed by tandem mass spectrometry (10 largest peaks) using an ABI 4800 MALDI-TOF/TOF mass spectrometer (Applied Biosystems). Peptides derived from trypsin were used as an internal standard. The tandem mass spectrometry results were analyzed by searching the SwissProt Protein Data Base using Mascot version 1.9.
Western Blotting-mAb 1D5, a murine IgM monoclonal antibody produced in this laboratory (7), was purified using an affinity column (Affiland). Anti-human legumain was purchased commercially (Cell Signaling). Parasite cells were lysed in buffer (25 mM Tris, pH 7.5, 150 mM NaCl, 0.1% Triton X-100, 1 mM dithiothreitol, 1 mM EDTA, 1 mM EGTA, 10 mM NaF, 20 mM ␤-glycerophosphate, 1 mM Na 3 VO 4 , Roche Applied Science complete protease inhibitor), and the soluble fraction was obtained by centrifugation (14,500 ϫ g, 10 min) at 4°C. The total protein concentration was determined by the Bradford assay. 50 g of total protein or 5 g of purified protein was electrophoresed on a 10% SDS-polyacrylamide gel and transferred onto a polyvinylidene difluoride membrane (GE Healthcare) by semi-dry blotting (Bio-Rad), blocked with 2% bovine serum albumin in TBS-T (triethanolamine-buffered saline with 0.1% Tween), incubated with primary antibody, washed, incubated with secondary antibody, visualized by chemiluminescence (ECL Plus, GE Healthcare), and exposed to an x-ray film (Eastman Kodak Co.).
Legumain Cloning and Expression-PCR primers were synthesized according to the peptide sequences from mass spectrometry as follows: P1 (5Ј-TTTGTNTCNATNGCNCTNCTN-CGN-3Ј and complementary sequence); P4 (5Ј-GATGACTTT-CANGCNACNCTNAAG-3Ј and complementary sequence); P6 (5Ј-TTTAACTACGANCANCAGTCNAGC-3Ј and complementary sequence); P7 (5Ј-TACCTNGAAGAGATGTCN-CTNCGN-3Ј and complementary sequence), where N represents A, T, G, or C. Forward and reverse primers were picked randomly. In total, 12 primer pair combinations were used for PCR. PCRs were performed with the Faststart high fidelity PCR system (Roche Applied Science) according to the manufacturer's protocol with Blastocystis cDNA library as template. PCR products were electrophoresed in a 1% agarose gel. Appropriate PCR products were excised, purified from the gel, and cloned into pCR2.1 (Invitrogen) by TOPO cloning for sequencing. Five positive plasmids were sequenced with M13 ϩ and M13 Ϫ primers, and consensus sequence was obtained. This product was amplified again by PCR using primers (forward, 5Ј-TTAGGA-TCCATGTTTGTGAGTATCGCCTTACT-3Ј, and reverse, 5Ј-ATCTCGAGACGAAGGCACATCTCCTC-3Ј). These primers contained BamHI and XhoI restriction sites to allow directional cloning into E. coli expression vector pGEX-6p-1 (GE Healthcare). pGEX-6p-1 containing the Blastocystis legumain gene was transformed into E. coli BL21(DE3) and amplified in 500 ml of LB containing 100 g ml Ϫ1 ampicillin at 37°C and 200 rpm. After A 600 reached 0.6 -0.8, isopropyl 1-thio-␤-D-galactopyranoside at a final concentration of 0.5 mM was added. Cells were cultured at 16°C overnight and then harvested by centrifuging at 5000 ϫ g for 10 min. The pellets were resuspended in 20 ml of cold PBS and disrupted by sonication. Protein expression was examined using a 10% SDS-polyacrylamide gel.
Purification of Recombinant Legumain-The lysate of transformed BL21(DE3) cells was clarified by centrifuging at 25,000 ϫ g for 20 min at 4°C, incubated with 2 ml of glutathi-one-Sepharose 4B slurry (GE Healthcare) at room temperature for 2 h, and washed with 5 volumes of PBS. Legumain was eluted with 2 volumes of elution buffer (10 mM reduced glutathione, PBS). The eluant was concentrated and purity assessed by SDS-PAGE.
Cellular Localization of Blastocystis Legumain-Cells (1 ϫ 10 7 ) were washed with cold PBS and incubated in Iscove's modified Dulbecco's medium containing 5 g of fluorescein diacetate at 37°C for 10 min, washed, and fixed with 3.7% (v/v) formaldehyde on ice for 30 min. To permeabilize cells, cells were incubated with 0.1% Triton X-100 for 5 min. Cells were incubated with primary antibody (mAb 1D5 or anti-legumain, 1.2 g ml Ϫ1 in PBS containing 3% bovine serum albumin) for 2 h at room temperature, washed, and incubated with AlexaFluor 594-conjugated secondary antibody (anti-mouse IgM or anti-goat IgG, 2 g ml Ϫ1 in PBS containing 3% bovine serum albumin) for 1 h at room temperature. The cells were washed and stained with 4Ј,6-diamidino-2-phenylindole and viewed under a confocal microscopy (Olympus FV500). Blue (CBB)-stained two-dimensional electrophoresis gel of total Blastocystis proteins (left) and Western blot (WB) analysis of total proteins probed with mAb 1D5 (right). A mAb 1D5-reactive spot (arrow) was used to identify the corresponding antigen from a Coomassie Brilliant Blue-stained gel and excised for mass spectrometric identification. B, peptide sequences from mass spectrometry are underlined. PCR was performed with primer combinations, and the appropriate-sized PCR product was sequenced, and the protein sequence was obtained. FIGURE 2. Multiple sequence alignment of the catalytic domain of the legumain family. The sequence segments containing known or putative catalytic residues in the legumain family were aligned by Vector NTI. Note the blocks of four predominantly hydrophobic residues (gray) located 2 or 3 residues preceding the catalytic residues (marked with asterisk). Other residues in the catalytic domains were highly conserved in the legumain family. Sequences were compared, and their accession numbers were as follows: mouse (NP_035305), rat (NP_071562), human (AAH03061), bovine (NP_776526), frog (NP_ 001005720), zebra fish (NP_999924), Schistosoma (CAB71158), rice (BAC41386), tobacco (CAE84598), Blastocystis (ACO24555), Hemonchus (CAJ45481), and Trichomonas (AAQ93040).
Annexin V Apoptosis Assay-Loss of membrane permeability and exposure of phosphatidylserine (PS) were detected using an annexin V-fluorescein isothiocyanate apoptosis detection kit (BioVision) following the manufacturer's instructions. In brief, 4-h treated (with antibody or inhibitors mentioned previously) or untreated cells (1 ϫ 10 7 ) were washed with cold PBS and incubated with 0.21 g ml Ϫ1 annexin V-fluorescein isothiocyanate and 2.5 g ml Ϫ1 propidium iodide at room temperature for 10 min. The cells were washed, resuspended in 500 l of PBS, and analyzed by a flow cytometer (DAKO CyAn ADP) to read green and red fluorescence.
TUNEL Assay for DNA Fragmentation-TUNEL was performed using APO-bromodeoxyuridine kit (Invitrogen) following the manufacturer's instructions. In brief, 24-h treated (with anti-body or the inhibitors mentioned previously) or untreated cells (1 ϫ 10 7 ) were washed with cold PBS and fixed in 1 ml of 3.7% (v/v) formaldehyde on ice for 30 min, washed, and resuspended in 1 ml of 70% cold ethanol at Ϫ20°C for 30 min to permeabilize cells. The cells were washed and incubated with 50 l of labeling solution at 37°C for 1 h. After rinsing, cells were stained with AlexaFluor 488-conjugated anti-bromodeoxyuridine antibody at room temperature for 30 min. The cells were washed, resuspended in 500 l of PBS, and analyzed by a flow cytometer (DAKO CyAn ADP) to read green fluorescence.

Identification of mAb 1D5 Target Protein-Total Blastocystis
proteins were separated by two-dimensional electrophoresis and probed with mAb 1D5. A distinct spot of 30 kDa, consistent with our previous studies (6, 7), was observed (Fig. 1A). The corresponding protein spot in the Coomassie Brilliant Blue-stained gel was analyzed by mass spectrometry. In total, seven tryptic peptides were sequenced (Fig. 1B, underlined). Among them, peptides 2, 3, and 5 matched the catalytic domain of legumain sequences from human, rat, mouse, and bovine origins. This suggests that the protein analyzed might be a member of the legumain family. Peptides 1, 4, 6, and 7 did not match any conserved regions of legumain and might be unique to Blastocystis legumain and located at the N or C terminus. Degenerate PCR primers were designed according to the sequences of these seven peptides resulting in 12 pairs of PCR primer combinations. The longest PCR product of ϳ1 kb was cloned and sequenced. It included 990 bp and encoded a 33-amino acid protein (GenBank TM accession number ACO24555) with a theoretical molecular mass of 37.5 kDa and pI value of 4.9, as calculated by the Vector NTI program. The full-length gene was obtained by 5Ј-rapid amplification of cDNA ends and 3Ј-rapid amplification of cDNA ends (1194 bp, supplemental Fig. S1; GenBank TM accession number GU124590). This sequence included a highly conserved catalytic domain that had 40 -60% similarity to that of legumains from other origins and by ClustalW was shown to possess unique N-and C-terminal regions (data not shown). As with other legumains, the Blastocystis orthologue presumably codes for a precursor that is auto-cleaved at the C-terminal asparagine residue (18), possibly at Asn-285 to result in the mature form (30-kDa protein).
Blastocystis legumain contains a catalytic dyad with the motif His-Gly-spacer-Ala-Cys, which is also found in other legumains (Fig. 2) and in caspases (13). Interestingly, the first 16 amino acids (KFV-SIALLRVLALAAA) include a putative secretary signal peptide, which was also observed in the legumain sequence of the Chinese liver fluke Clonorchis sinensis (19) and not from legumains of other origins.  Expression, Purification, and Verification of Legumain-Blastocystis legumain (truncated form missing 66 amino acids at the 3Ј end when compared with the full-length sequence) was expressed in E. coli (Fig. 3). The appearance and accumulation of the overexpressed protein band in E. coli total extracts after 1-h (Fig. 3, lane 2) and 2h (lane 3) isopropyl 1-thio-␤-Dgalactopyranoside induction could be detected in comparison with total extracts before induction (lane 1). About 70% overexpressed protein was insoluble, and soluble recombinant protein was purified using the N-terminal GST tag (Fig. 3, lane 4). The GST-tagged legumain was about 65 kDa and its purity was about 95%. The purified recombinant protein was confirmed by anti-GST antibody on the Western blot (Fig. 3, lane 5). To confirm that the purified recombinant protein was legumain and that this was a bona fide target of mAb 1D5, Western blots of the recombinant protein were probed with anti-human legumain (IgG) or mAb 1D5 with Blastocystis cell lysate as a reference and control. Results showed that there were bands at the expected positions (Fig. 3, recombinant legumain, lanes 6 and 8, 65 kDa, and cell lysate, lanes 7 and 9, 30 kDa). The variations in band intensities may be due to mAb 1D5 and anti-human legumain possessing different affinities for precursor and mature forms of legumain. Furthermore, another truncated legumain gene missing 15 amino acids at the 5Ј end, which removes a putative secretary signal peptide, was expressed in E. coli using the same induction method. The His 6 -tagged protein was then purified, and activity and inhibitor assays were performed to study the effect of the signal peptide on enzyme activity. The results (supplemental Fig. S2) showed that legumain with or without signal peptide displayed similar protease activities.
pH Dependence of Legumain Activity-Purified recombinant legumain was incubated at 30°C with pH buffers in the range of pH 3.0 -8.8, and activity against legumain-specific substrate Z-Ala-Ala-Asn-NHMec was measured. Interestingly, two distinct peaks, one at pH 4.0 and the other one at pH 7.5, could be observed (Fig. 4). The reason for this unusual biphasic activity profile is not known but may be related to the localization of Blastocystis legumain in intracellular compartments and on the parasite surface (see below).
Cellular Localization of Blastocystis Legumain-To investigate the localization of legumain in Blastocystis, mAb 1D5 and anti-human legumain were used as primary antibodies to stain nonpermeabilized and permeabilized Blastocystis cells. 4Ј,6-Diamidino-2-phenylindole (blue, nuclear staining) and fluorescein diacetate (green, staining cytoplasm and central vacuole) were included to visualize cellular structures. Blastocystis cells have a large central vacuole that occupies ϳ90% of the cell volume and forces the cytoplasm and organelles into a very thin peripheral rim (20). Both antibodies revealed surface labeling in nonpermeabilized samples, whereas intracellular localization within vesicular-like structures was observed on permeabilized samples (Fig. 5). This result suggests that Blastocystis legumain localizes on the cell surface as well as in the cytoplasm and central vacuole.
Regulation of Legumain Protease Activity-Recombinant legumain protease activity was performed in PBS buffer with fluorogenic substrate Z-Ala-Ala-Asn-NHMec in the presence of mAb 1D5 or inhibitor. From the results (Fig. 6), the legumain-specific inhibitor carbobenzyloxy-Ala-Ala-AAsnepoxycarboxylate ethyl ester (APE-RR) showed over 90% inhibition when compared with MOCK control; monoclonal antibody mAb 1D5 showed about 20% inhibition; cystatin showed about 25% inhibition, and caspase-1 inhibitor Ac-YVAD-CMK had only a 10% inhibitory effect. There were no inhibitory effects for nonspecific IgM and cathepsin B inhibitor Z-FA-fluoromethyl ketone even when the final concentration of the inhibitor was increased by 5-fold (data not shown).
Annexin V Apoptosis Assay-Previous studies showed that mAb 1D5 induces PCD features in Blastocystis (2,21). We asked if the induction of PCD was triggered by inhibition of surface legumain activity. Blastocystis cells were incubated for 4 h with antibody mAb 1D5 or protease inhibitors described above. To ensure reaction specificity, competitive inhibition assays were included. This involved the preincubation of mAb 1D5 or legumain inhibitor APE-RR with purified recombinant legumain protein (r-legumain) at a molar ratio of 1:2 for 30 min before adding to cells. Annexin V-fluorescein isothiocyanate staining was used to detect flipped PS as an early signal of cell apoptosis. From the results (Fig. 7), 18.29% of the population was annexin V-positive in mAb 1D5-treated cells; 49.53% of the population was annexin V-positive in cystatin-treated cells; and 47.04% of the population was annexin V-positive in legumain inhibitor-treated cells. The extensive annexin V-positive population seen in mAb 1D5 and legumain inhibitor-treated cells was abrogated when mAb 1D5 or legumain inhibitor was preincubated with purified r-legumain. There was no significant increase of annexin V-positive cells for cells treated with caspase-1 and cathepsin B inhibitor. The percentage of annexin V-positive cells in mAb 1D5-treated cells might be less than the actual value, which can probably be attributed to the blockage of annexin V access to PS by the surrounding antibodies.
TUNEL Assay-To confirm the correlation between legumain protease activity with cell death, the TUNEL assay was employed to quantify in situ DNA fragmentation, a late stage PCD event. Blastocystis cells were incubated for 24 h with antibody mAb 1D5, protease inhibitors, preneutralized mAb 1D5, or legumain inhibitor. As shown in Fig. 8, 90.11% of the population was TUNEL-positive in cystatin-treated cells; 80.25% of the population was TUNEL-positive in legumain inhibitortreated cells; and 47.17% of the population was TUNEL-positive when cells were treated with mAb 1D5. Similar to the trend observed in the annexin V apoptosis assay, there was no significant increase in TUNEL-positive cells when these were treated with caspase-1 inhibitor (2.5%) or cathepsin B inhibitor (0.97%). As also observed for the annexin V assays, the r-legumain-preincubated mAb 1D5 and the legumain inhibitor showed no significant increase in TUNEL-positive cells.

DISCUSSION
Our previous studies showed that the cytotoxic surface-reactive mAb 1D5 induced a number of typical PCD features in Blastocystis (3,21,22), including cell shrinkage, caspase-like activity, PS externalization, mitochondrial outer membrane permeabilization, and DNA fragmentation. This study reveals that the target of mAb 1D5 is a surface legumain. Comparison of Blastocystis full-length protein sequence with other legumain sequences showed that Blastocystis legumain was 30.1% FIGURE 5. Localization of mAb 1D5 and anti-legumain. Nonpermeabilized or permeabilized cells were incubated with primary antibody (mAb 1D5 or anti-legumain) followed by incubated with AlexaFluor 594-conjugated secondary antibody (red). Nuclei were stained with 4Ј,6-diamidino-2-phenylindole (DAPI) (blue), and cytoplasm and central vacuole were stained with fluorescein diacetate (FDA) (green). A confocal microscope with differential interference contrast (DIC) optics was used to observe cell morphology. Arrows indicate healthy cells revealing surface labeling, and triangles indicate cells with compromised membranes resulting in intracellular staining. Control conditions include absence of primary antibody or using nonspecific antibody as the primary antibody. identical to human, 28.4% identical to blood fluke, and 22.8% identical to T. vaginalis legumains. Key conserved features include the catalytic His-Cys dyad, four predominantly hydrophobic amino acids closely preceding each of the catalytic residues, putative N-terminal prodomain and C-terminal extension, activity against legumain substrates, and inhibition by legumain-specific inhibitors. From the multiple sequence alignment, Blastocystis legumain possesses unique Nand C-terminal regions, suggesting downstream substrates, localization, and other properties distinct from reported legumains. Regarding pH dependence of stability and activity, mammalian and plant legumains have been found to be unstable at neutral pH (23) and to exhibit optimal activity at low pH, with the pH optimum of human legumain at pH 4.0 (14), pig legumain at pH 5.8 (23), and kidney bean legumain at pH 5.4 (23). Legumains from parasite origins, in contrast, have pH optimum near neutral pH, e.g. T. vaginalis legumain at pH 7.0 (24) and the helminths Hemonchus contortus and Schistosoma mansoni legumains at pH 7.0 (25) and 6.8 (26), respectively. Blastocystis legumain is active over the wide range of pH 4 -8 with the optimum pH at 7.5, similar to what has been reported in T. vaginalis (23). It has been shown that parasitic cysteine proteases have a broad pH profile and serve various extra-lysosomal functions as opposed to the narrow pH range of mammalian lysosomal cysteine proteases that primarily exert their activity in the lysosome (27). Interestingly, Blastocystis legumain has biphasic pH dependence, with peaks at pH 4.0 and 7.5, a property that has not been seen in legumains from all other origins, and is infrequently observed among enzymes in general. Some individuals with glucose-6-phosphate dehydrogenase deficiency harbor variant glucose-6-phosphate dehydrogenases that display biphasic pH optimum curves (28,29). In a study on renin pH dependence, it was shown that Ser-84 of human renin contributes to the biphasic pH dependence of the renin-angiotensinogen reaction (30). This biphasic characteristic was specific for human renin, whereas mouse and rat renins exhibited single peaks. It is presently unclear why Blastocystis legumain exhibits two peaks of enzyme activity. The localization of the enzyme on the cell surface and in intracellular compartments is consistent with activity peaks at neutral and acidic pH values. We postulate that Blastocystis legumain possesses multiple functions and plays important roles at the cell surface and within acidic compartments. Legumains are generally located within lysosomes of mammalian cells and vacuoles of plant cells but have been reported to localize to the cell surface of metastatic tumors (17) and on the microvillar surfaces of helminth intestinal cells (25). Surface legumains were suggested to activate local zymogens that may aid tumor invasion or participate in helminth alimentary digestion of host proteins. Blastocystis cysteine proteases are able to cleave human immunoglobulin A (31) and also induce pro-inflammatory responses in host cells (32). The presence of legumain on the parasite surface may similarly function to activate these proteases. The localization of Blastocystis legumain to the cell surface is probably promoted by its uncommon secretory signal peptide. Such a signal peptide has only been explicitly described in the legumain of the Chinese liver fluke C. sinensis (19), which was found in the excretory-secretory products as a serological antigen.
Blastocystis legumain activity was abolished by the legumainspecific inhibitor APE-RR, an aza-Asn derivative that does not cross-react with caspases, papain, and cathepsin (33). It is moderately inhibited by cystatin and mAb 1D5 and weakly inhibited by the caspase-1 inhibitor Ac-YVAD-CMK. Cystatins are potent inhibitors of the papain-like cysteine peptidases in the unrelated family C1 but also inhibit legumain, due to a separate site on the cystatin molecule (34), although plant and mammalian legumains have been shown to cleave caspase-1 substrates (15,23). Hence, Blastocystis legumain shares a number of characteristics with legumains of other origins.
In plants, legumain is essential for hypersensitive-response cell death (15) and is also involved in fungal toxin FB1-induced plant cell death (16). Virus-induced gene silencing of VPE revealed loss of VPE and caspase-1 cleavage activities concomitant with loss of PCD features upon tobacco mosaic virus induction, resulting in increased virus proliferation. Hence, VPE-mediated cell death is important for resistance to pathogen infection. VPE targets that mediate hypersensitive-response cell death are currently unknown (35). In contrast to the plant model, our present studies suggest that Blastocystis legumain mediates pro-survival functions. This is evidenced by the significant PS externalization and in situ DNA fragmentation upon incubation with legumain inhibitors. There was a positive correlation between the extent of these PCD features with the degree of legumain inhibition (Fig. 9), with legumain-specific inhibitor and cystatin inducing the greatest amount of PS flipping and DNA fragmentation. The data indicate that PCD in Blastocystis is regulated by surface legumain activity and suggests a pro-survival role for the surface molecule. However, although cystatin only caused 45% inhibition of Blastocystis legumain activity as compared with 90% inhibitory effect by the legumain-specific inhibitor, the extent of cell death induced by cystatin was even higher than that induced by the legumain-specific inhibitor. Because cystatin is also a high affinity inhibitor of family C1 cysteine proteases, it is possible that there is a papain-like cysteine protease downstream of legumain whose inhibition triggers PCD, and by inhibiting both legumain and this hypothetical substrate protease, cystatin causes massive cell death in Blastocystis. Because cystatin is a cell-impermeable inhibitor (36), it is thus unlikely that PCD induction was due to inhibition of some other intracellular enzyme. We speculate that Blastocystis cell surface legumain may be responsible for processing and activating a downstream cysteine protease, which also localizes at the cell surface and has important function in nutrition uptake. Alternatively, cystatin may independently trigger PCD by inactivating papain-like proteases on the parasite surface. Our previous studies showed  that mAb 1D5 exposure induces PCD features rapidly, as evidenced by PS externalization and caspase-like activity by 2 h post-induction (4,22), suggesting a direct role for legumain in Blastocystis PCD. In this study, it was demonstrated that mAb 1D5 was able to inhibit Blastocystis legumain. It is known that enzymes can be inhibited by some of their specific antibodies through mechanisms such as steric hindrance and conformational changes (37). mAbs of different Ig subtypes have been used to study the role of individual cytochrome P450 isoforms. Because of their high specificity and high inhibitory activity, carefully selected mAbs were suggested to be superior to chemical inhibitors to target specific cytochrome P450 enzymes (37). mAb 1D5 is the first reported legumain-specific antibody that has an inhibitory effect on this enzyme. The mechanism by which Blastocystis legumain modulates PCD is presently unclear. Human embryonic kidney 293 cells overexpressing legumain appear more resistant to apoptosis compared with cells that have basal levels and is associated with tumor invasion and metastasis in vivo (17). More work needs to be done to unravel the pro-survival mechanism of Blastocystis legumain and to determine whether it is similar to that of tumor cells. Identification of the substrates of legumain will take us one step further in delineating the legumain-mediated PCD pathway.
This study is the first to provide detailed functional characterization of an unusual protozoan legumain. Blastocystis legumain is localized on the cell surface, has biphasic pH optima, and is associated with PCD. As there have only been a handful of protozoan genes shown to be involved in PCD, the identification of a surface protease with a possible pro-survival function sheds new light on the mechanisms of PCD among the unicellular eukaryotes.