The cathepsin B of Toxoplasma gondii, toxopain-1, is critical for parasite invasion and rhoptry protein processing.

Cysteine proteinases play a major role in invasion and intracellular survival of a number of pathogenic parasites. We cloned a single copy gene, tgcp1, from Toxoplasma gondii and refolded recombinant enzyme to yield active proteinase. Substrate specificity of the enzyme and homology modeling identified the proteinase as a cathepsin B. Specific cysteine proteinase inhibitors interrupted invasion by tachyzoites. The T. gondii cathepsin B localized to rhoptries, secretory organelles required for parasite invasion into cells. Processing of the pro-rhoptry protein 2 to mature rhoptry proteins was delayed by incubation of extracellular parasites with a cathepsin B inhibitor prior to pulse-chase immunoprecipitation. Delivery of cathepsin B to mature rhoptries was impaired in organisms with disruptions in rhoptry formation by expression of a dominant negative micro1-adaptin. Similar disruption of rhoptry formation was observed when infected fibroblasts were treated with a specific inhibitor of cathepsin B, generating small and poorly developed rhoptries. This first evidence for localization of a cysteine proteinase to the unusual rhoptry secretory organelle of an apicomplexan parasite suggests that the rhoptries may be a prototype of a lysosome-related organelle and provides a critical link between cysteine proteinases and parasite invasion for this class of organism.

The protozoan, Toxoplasma gondii, is an obligate intracellular parasite that can invade and replicate within any nucleated cell of vertebrate hosts, including humans (1)(2)(3). Invasion by T. gondii tachyzoites is mediated by the sequential regulated release of specialized secretory organelles of the parasite including the micronemes, rhoptries, and dense granules (4). In early observations, the penetration of host cells by tachyzoites was enhanced by the addition of partially purified lysosomal enzymes (5). In addition, proteinases have been implicated in host cell invasion in other members of the Apicomplexa such as Plasmodium (6) and Eimeria tenella (7). We now show that a cathepsin B, toxopain-1, is strongly implicated in T. gondii invasion and that infection can be interrupted with specific cathepsin B inhibitors.
Another unusual feature of Toxoplasma is the absence of a morphologically identifiable lysosomal system. In higher eukaryotic cells, acidic cathepsins in lysosomes are important in protein processing and breakdown. The mammalian precursor of cathepsin B is targeted to the lysosomal compartment by mannose 6-P for proteolytic activation (8). Thus, the apparent lack of a lysosomal system in Toxoplasma raises a number of questions regarding cellular proteinase functions within the parasite. The rhoptries are club-shaped organelles located at the apical end of the parasites with no known counterpart outside of the phylum Apicomplexa (9). Although nine rhoptry proteins (ROP1-9) 1 have been identified to date (10), a definitive function has only been determined for ROP2, which mediates binding of host mitochondria to the parasitophorous membrane (11). How rhoptry proteins directly contribute to the invasion process is still not clear. T. gondii rhoptry proteins are synthesized as prepro proteins that are subject to proteolytic cleavage to remove the presequence and the proregion at the N terminus and then processed to their mature forms (11)(12)(13)(14). Rhoptries and pre-rhoptries are the only acidified organelles identified in the parasite (15), contain a high lipid to protein ratio (16), and scavenge sterols from the host cell (17). We now show that the T. gondii cathepsin B, toxopain-1, localizes to the rhoptries, and inhibition of cathepsin B activity or disruption of ROP protein targeting leads to mislocalization of toxopain-1 and abnormal rhoptry biogenesis. This is the first evidence for localization and function of any proteinase to the unusual rhoptry secretory organelle of an apicomplexan parasite.

EXPERIMENTAL PROCEDURES
Reagents-All reagents were purchased from Sigma Chemical Co. unless otherwise specified. All AMC and synthetic AMC peptide substrates were obtained from Enzyme System Products (Livermore, CA). Synthetic peptide inhibitors were a gift from Prototek Inc. (Dublin, CA).
Parasite and Host Cell Cultures-Human foreskin fibroblasts (HFF) were initially cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (Irvine Scientific, Irvine, CA) and 2 mM glutamine and maintained subsequently in the same medium with 2% fetal bovine serum. The RH strain of T. gondii was obtained from Dr. John Boothroyd (Stanford University, CA). Tachyzoites of the RH strain were maintained in confluent monolayers of HFF cells.
Cloning of the Cathepsin B Gene-A cathepsin B family proteinase gene, tgcp1, was amplified from cDNA with primers based on conserved sequences of eukaryotic cathepsins (18) and on partial Expressed Sequence Tag (EST) sequence data from the Toxoplasma genome project (EST clone: TgESTzy53g12r.1) (19). The lambda phage clone was obtained from Genome Systems (St. Louis, MO) in vivo excised with ExAssist (Stratagene, La Jolla, CA) as helper phage and sequenced. The remaining sequence was obtained by 3Ј-and 5Ј-RACE (Invitrogen). Total cellular RNA from T. gondii was isolated using RNAzol reagent and transcribed into single-stranded cDNA using Superscript II reverse transcriptase and tgcp1-specific reverse primer R1 (5Ј-GAG TCA TCA TAT CTC TCT TGA CG-3Ј). Excess dNTPs and primer were removed from single-stranded cDNA, and a homopolymeric tail of dCs was added to the end of single-stranded cDNAs using terminal deoxynucleotidyl transferase. The 5Ј-end of tgcp1 cDNA was then amplified from dCtailed cDNA using the abridged anchor primer and a specific nested primer F2 (5Ј-CTT CTG TGC TGG CGA AG-3Ј). For 3Ј-RACE, cDNA was synthesized from total RNA using the oligo(dT) primer. The 3Ј-end of the tgcp1 cDNA was PCR-amplified with the tgcp1-specific forward primer F5 (5Ј-GTG TCG GGG GCC TTC ATG GTC-3Ј) and the oligo(dT) primer. Finally, the complete gene for T. gondii cathepsin B was amplified by RT-PCR with specific primers against the 5Ј-and 3Ј-untranslated sequences, cloned into the TA vector (Invitrogen) and sequenced. All nucleotide sequences and predicted amino acid sequences were compared with known sequences in the data base using the BLASTX, BLASTN, and BLASTP programs at the National Center for Biotechnology Information. Southern hybridization of Toxoplasma genomic DNA with a tgcp1 gene probe was performed according to previously described protocols (20).
Homology Modeling of Toxoplasma Cathepsin B-The sequence of toxopain-1 was aligned with human liver lysosomal cathepsin B, rat cathepsin B, and rat procathepsin B. Existing x-ray crystal structures of human liver lysosomal cathepsin B with CA030 inhibitor (PDB ID: 1CSB), human liver lysosomal cathepsin B with no inhibitor (PDB ID: 1HUC), and rat cathepsin B (PDB ID: 1CTE) were used as threedimensional templates for the modeling of Toxoplasma cathepsin B. MODELLER, a module of the Quanta software package, was used to generate homology-based models of the enzyme.
Expression and Refolding of Recombinant Toxopain-1-To express recombinant enzyme, the sequence of the pro-and mature toxopain-1 was amplified and subcloned into pBAD/Thio TOPO (Invitrogen). The plasmid was transformed into Escherichia coli codon plus cells (Stratagene) in SOB media and induced with 0.02% arabinose with 100 g/ml ampicillin. The cell pellet was solubilized in 6 M GdnHCl, 0.1 M NaH 2 PO 4 , 0.01 M Tris, pH 8.0, 10 mM ␤-mercaptoethanol, sonicated, and the soluble proteins applied to a nickel affinity column (Invitrogen ProBond), washed with the same buffer at pH 6.3, and eluted at pH 4.5. The protein in the fractions was determined by the Bradford reaction and the fractions screened by SDS-PAGE. One liter of culture yielded ϳ30 mg of the thioredoxin-TgCP1 fusion protein. Thioredoxin-TgCP1 fusion protein (74.5 mg or 1 M) was slowly added to 1 liter of 50 mM MES, pH 6.0, 30 mM NaCl, 10 mM MgCl 2 , 10 mM CaCl 2 , 750 mM arginine, 500 mM GdnHCl, 5 mM reduced glutathione, and 0.5 mM oxidized glutathione and stirred in the cold for 48 h. The sample was dialyzed against Trisbuffered saline, pH 7.5 and concentrated to 30 ml. The sample was re-dialyzed against 0.01 M Tris, pH 7.5 and separated by ion exchange FPLC on a Mono Q column.
Assays of Cysteine Proteinase Activity-Proteinase activity in column fractions or tachyzoite lysates were measured by cleavage of peptide substrates. T. gondii RH strain tachyzoites were harvested from human fibroblast monolayers, purified by filtration through 3.0-m nucleopore filters (Costar, Cambridge, MA) and resuspended at 10 9 cells/ml in 50 mM Tris buffer, pH 7.4. Following three freeze-thaw cycles and a 15-s sonication, the proteinase activity was measured in the soluble fraction. Substrate specificity was tested for the liberation of the fluorescent leaving group, AMC, from the synthetic peptide substrates to determine the preferred cleavage of the P1 and P2 sites (21). The initial rate of substrate hydrolysis (nmol/min/mg protein) is based on the rate of increase of fluorescence using a Labsystems Fluoroskan Spectrofluorometer.
The inhibitor profile (IC 50 ) was determined by monitoring inhibition of hydrolysis of Z-Arg-Arg-AMC substrate in the presence of serial dilutions of the inhibitors following incubation for 30 min at room temperature. The IC 50 was calculated as the concentration of inhibitor resulting in 50% inhibition of proteinase activity compared with noninhibited controls. The peptidyl ketone inhibitors, PRT2005 and PRT2253 (a kind gift from Prototek), with a Phe-hPhe substitution in the P2 and P1 residues, were used in all subsequent inhibitor studies.
Antibody Production-Recombinant toxopain-1 was expressed in E. coli and purified by nickel-nitrilotriacetic acid affinity column chromatography as above. Polyclonal antibody to recombinant toxopain-1 was produced by immunizing rabbits three times with 100 g of recombinant protein mixed with Titermax Gold Adjuvant (Sigma). The antiserum was affinity-purified by adsorption and desorption to recombinant toxopain-1 on nitrocellulose membranes as described previously (22).
Monoclonal antibodies were produced previously by immunizing BALB/c mice with 50 g of recombinant ACP1 (amebic cysteine proteinase 1) using RIBI adjuvant and boosted once intravenously (23). Spleens were harvested and fused with NSO myeloma cells. Three monoclonal antibodies, 1.7, 1.15, and 1.17 were subsequently found to cross-react with T. gondii lysates by immunoblots and ELISA.
Effect of Cysteine Proteinase Inhibitors on Invasion by T. gondii-To evaluate the effect of specific proteinase inhibitors on invasion, tachyzoites (5 ϫ 10 4 to 1 ϫ 10 6 ) were preincubated 30 min at 37°C in medium alone or medium containing 20 M PRT2005 or PRT2253. They were then added directly to fibroblast monolayers on chamber slides (Lab Tek, Nunc) for 2 h. Following washing, the cells were fixed in 4% formaldehyde for 10 min, stained with acridine orange (5 g/ml), and the number of infected cells was determined by fluorescent microscopy.
Immunofluorescence Microscopy-HFF cells were grown overnight on coverslips and infected with 2 ϫ 10 5 tachyzoites per well for 24 -48 h. Monolayers were washed with phosphate-buffered saline (PBS) and fixed for 15 min in 3% paraformaldehyde (w/v) in PBS at room temperature. Cells were washed in PBS, permeabilized for 5 min in 0.25% Triton X-100 in PBS, and washed three additional times in PBS. Cells were incubated for 1 h with primary antibody diluted in PBS with 3% bovine serum albumin using anti-ROP2/3/4 mAb T3 4A7 (14) or antihemagglutinin (HA) epitope tag monoclonal antibody for overexpressing wild-type Tg1-HA, dominant negative Tg1-HA(D176A) 2 or ROP2⌬-HA mutant parasites (24). Toxopain-1 localization was performed with mAb CP1.7 (2 mg/ml) directly labeled with Alexa 594 per kit instructions (Molecular Probes, Eugene, OR) before adding the second antibody. After three washes in PBS, cells were incubated with FITC-conjugated secondary antibodies and washed in PBS. The slides were mounted and examined by epifluorescent microscopy.
Processing of Rhoptry Proteins-To assess the effect of proteinase inhibitors on rhoptry protein processing in vivo, purified tachyzoites (2.5 ϫ 10 8 ) were washed in prewarmed methionine/cysteine-free (Met Ϫ / Cys Ϫ ) Dulbecco's modified Eagle's medium and incubated for an additional 30 min at 37°C in medium alone or containing 20 M inhibitor PRT2253S. Tachyzoites were pulse-labeled with 300 Ci/ml of [ 35 S]methionine/[ 35 S]cysteine (PerkinElmer Life Sciences, Boston, MA) for 15 min in the presence or absence of 20 M PRT2253S and chased in complete medium for 0 -60 min before harvesting. Tachyzoites were washed with PBS containing 1 mM MgCl 2 and 1 mM CaCl 2 with inhibitor and lysed in 1% SDS, and the radiolabeled rhoptry proteins were immunoprecipitated with mAb T3 4A7 and protein A-Sepharose 4B beads (Amersham Biosciences). The beads were separated by 12% SDS-PAGE, the gel impregnated with fluorographic enhancement solution (Amplify, Amersham Biosciences), dried, and exposed to x-ray film for autoradiography.
Tg1 Mutants-Tg1 mutants were produced by PCR with an HA epitope tag introduced into Tg1 between codons 231 and 232. 2 The resulting D176A mutation had been shown to alter the binding of the rat 2-chain to tyrosine-binding motifs (25). Tg1-HA and the mutated Tg1-HA(D176A) were ligated into pNTP/Sec for transient transfection and into pDHFR-TS/C3M2M3 for stable transfection as previously described (24).
Electron Microscopy-For cryoimmunoelectron microscopic localization, monolayers of fibroblast cells were infected with RH tachyzoites for 24 h, fixed in 0.8% paraformaldehyde in 0.25 M HEPES buffer pH 7.4, for 48 h, scraped into 1% bovine serum albumin solution, and pelleted in 10% fish skin gelatin. The pellets were infiltrated overnight with 2.3 M sucrose and frozen with liquid nitrogen, and ultrathin cryosections were transferred to grids. Sections were incubated with monospecific polyclonal antibodies to toxopain-1. After washing, sections were incubated with 5 or 10 nmol of protein A-gold conjugates. Following washing, sections were postfixed in 1% glutaraldehyde and contrasted with 0.2% uranyl acetate in 2% methyl cellulose. Images were recorded with a Philips 410 Transmission electron microscope.
For transmission electron microscopic studies, infected monolayers were grown in the presence of media with or without inhibitor for 48 h (20 M PRT2005), fixed in 2.5% glutaraldehyde in 0.1 M sodium cacodylate, pH 7.4 for 30 min, washed and postfixed in 1% osmium tetroxide. Fixed samples were embedded in Epon and thin-sectioned for transmission electron microscopy. Images were recorded with a Philips Tcenai 10 Transmission electron microscope.

RESULTS
Cloning of tgcp1-A full-length cathepsin B gene, tgcp1, was amplified from RH tachyzoite cDNA based on homology to conserved sequences of eukaryotic cysteine proteinases and a partial sequence in the T. gondii EST data base (19). The completed transcript of 3236 nucleotides encodes a preprocathepsin B precursor of 569 amino acids (M r ϭ 62,165), comprising a predicted N-terminal signal sequence, a propeptide domain, and mature enzyme (Fig. 1). The predicted Met start codon conforms to a Kozak eukaryotic consensus sequence (AN-NATGG) for initiation of translation. The sequences leading the ATG start codon (TTTTTTTCACCAAGGAAAAAATGG) conform closely to the T-stretches found in many T. gondii genes. The N-terminal 34 amino acids of the prepeptide region possess a putative signal sequence consisting of hydrophobic amino acids. The putative cleavage site between the signal peptide and propeptide region is predicted by the (Ϫ3, Ϫ1) rule using the SignalP algorithm of von Heijne (26).
Like human cathepsin B, toxopain-1 contains the active site triad of cysteine, histidine, and asparagine at positions 30, 211, and 231, respectively. Toxopain-1 has 12 Cys residues at similar positions to those of human cathepsin B to participate in formation of 6 disulfide bridges. In cathepsin B family proteases, there is a conserved motif Gly-Cys-Asn-Gly-Gly (residues 70 -74 in human); this motif is also present in toxopain-1 (residues 73-77). Toxopain-1 has an intact occluding loop (His 116 and His 117 ) and likely possesses exopeptidase activity in addition to its endopeptidase function. Unlike the cysteine proteinases of Trypanosoma brucei and Trypanosoma cruzi, toxopain-1 has a much shorter C-terminal extension (ϳ30 residues) (31).
Southern blot hybridization of parasite genomic DNA with a toxopain-1 probe revealed two bands of ϳ5200 and 3000 base pairs when restricted with EcoRI and a single band of ϳ19,000 base pairs when cut with HindIII, indicating a single copy gene (data not shown).
Homology Modeling of Toxoplasma Cathepsin B-The model structure generated for Toxoplasma cathepsin B (Fig. 2) was very similar to that of the templates, human liver lysosomal cathepsin B and rat cathepsin B. Overall, the root meansquared deviation between the backbone of the model structures was less than 1 Å when superimposed on the template starting structures. Close examination of the active site region of the models shows a high degree of similarity. In particular, the catalytic triad residues of the models and templates are nearly perfectly superimposable. In human cathepsin B, the substrate-binding cleft located between the L and R domains of the enzyme is occluded by a loop (residues 104 -114). The essential His 110 -His 111 moiety in the loop provides positively charged anchors for binding of dipeptidyl carboxylpeptidase substrates for the exopeptidase activity, which is only present in cathepsin B.
Expression and Purification of Active, Recombinant Toxopain-1-To express recombinant protein for refolding, we amplified the region encoding the promature sequence and subcloned the fragment into the pBAD/Thio vector, which optimizes expression of soluble protein and allows purification by nickel affinity chromatography of ϳ30 mg of toxopain-1/ liter. Using a partial factorial refolding screen spanning (32), we found that maximal soluble protein was recovered (73% input protein) after 48 h in a buffer containing 50 mM MES, pH 6.0, 500 mM guanidine hydrochloride, 750 mM arginine, and a 10:1 ration of reduced and oxidized glutathione. Following dialysis against Tris-buffered saline, concentration and purification by ion exchange chromatography, proteolytic activity was associated with a single 32-kDa band, consistent with the mature enzyme (Fig. 3).
We compared the substrate specificity of the soluble cysteine proteinase activity in tachyzoite lysates to the recombinant enzyme. Substrate specificity was tested by cleavage of synthetic peptide substrates containing Boc-X-X-4-amino-7-methylcoumarin, where X is a positively charged or neutral amino acid residue, to assess the preferred cleavage of the P1 and P2 sites. The initial velocity of cleavage of Z-Arg-Arg-AMC (modeling the specificity of cathepsin B) was greater than seven times that of Z-Phe-Arg-AMC (substrate of cathepsin L) (Fig. 4).
Cysteine Proteinase Inhibitors Block Toxopain-1 and Parasite Invasion-Ten new peptidyl ketone cysteine proteinase inhibitors (from Prototek) were screened for their activity against cysteine proteinases in T. gondii lysates by measuring the IC 50 . The inhibitors are di-and tripeptides bearing activated ketones, which are irreversible and covalently modify the enzyme. Seven inhibitors were active at concentrations less than 100 M (data not shown). PRT2005 and PRT2253, two of the most active inhibitors with an IC 50 Ͻ 1 M were used in the subsequent studies.
To evaluate the effect of inhibitors on host cell invasion, tachyzoites (5 ϫ 10 4 to 1 ϫ 10 6 ) were incubated in medium alone or containing 20 M PRT2005 and allowed to invade fibroblast monolayers in chamber slides for 2 h. The slides were washed, fixed in formaldehyde for 10 min, and stained with acridine orange for counting. The number of infected cells was significantly less in the presence of the inhibitor (p Ͻ 0.05 by Student's paired t test, Fig. 5).
Cathepsin B Localizes to Rhoptries-To localize toxopain-1, we performed confocal fluorescent microscopy with monoclonal and polyclonal antibodies against toxopain-1, monoclonal antibodies against rhoptry proteins 2, 3, and 4 (14), polyclonal antibodies generated against ROP2 (33), and monoclonal antibodies against micronemes and dense granules. Studies of infected monolayers revealed that toxopain-1 localized to the rhoptries and residual body, but not to micronemes and dense granules (Fig. 6). Cryoimmunoelectron microscopy using affinity-purified polyclonal antibodies to toxopain-1 confirmed that toxopain-1 is specifically localized (Ͼ90% of labeled gold particles) to the rhoptries, endosomal vesicles (34), and secreted into the parasitophorous vacuolar space (Fig. 7). Toxopain-1 was concentrated in a narrow strip down the center of the rhoptry, in contrast to other rhoptry proteins, which are diffusely distributed.
Cysteine Proteinase Inhibitors Disrupt the Enzymatic Processing of ROP2-Because rhoptry proteins are processed prior to their delivery to mature rhoptries (25), we examined whether ROP2 processing is inhibited by specific cathepsin B inhibitors. Processing of pro-rhoptry proteins, ROP2, 3, 4 to mature rhoptry proteins was delayed Ͼ60% following preincubation of extracellular parasites with 20 M PRT2253S for 15 min at room temperature prior to pulse-chase experiments (Fig. 8).
Disruption of Tg1 Function Missorts Cathepsin B to Abnormal Rhoptries-Localization of T. gondii ROP2 is dependent on a tyrosine motif (YEQL) in the cytoplasmic tail that binds to the 1-chain of the T. gondii AP-1 clathrin adaptor (Tg1) (24). Tg1 adaptin is essential for normal biogenesis of rhoptries (24). A Tg1(D176A)-HA mutant with a mutation at a residue necessary for binding to the tyrosine motif functions as a dom- inant negative inhibitor of AP-1 function. In parasites overexpressing the dominant negative Tg1(D176A)-HA, rhoptry biogenesis is altered, and the labeling of ROP2 is decreased drastically in intensity and detectable only in a thin, discontinuous organelle spanning the apical half of the cell (Fig. 9). Cathepsin B was also mislocalized to both distorted rhoptries and to residual bodies in the parasitophorous vacuole space in the D176A mutant, but not in parasites stably overexpressing wild-type T. gondii 1-HA (data not shown). These results indicate that cathepsin B is directed to the rhoptry secretory pathway. Hence, manipulations that interfere with rhoptry biogenesis also interfere with normal targeting of cathepsin B.
Inhibition of Cathepsin B Activity Results in Abnormal Rhoptry Morphology-To determine the requirement of active cathepsin B for rhoptry biogenesis, infected monolayers were grown in the presence of 20 M PRT2005 for 48 h. Electron micrographic studies revealed that the rhoptries were small and poorly developed (Fig. 10). No changes were detected in uninfected fibroblast cells incubated with 20 M PRT2005 for 48 h (data not shown). Thus, rhoptry biogenesis is dependent on both toxopain-1 and ROP protein delivery and processing.

DISCUSSION
Cysteine proteinases are critical to invasion by a number of protozoan parasites, including Plasmodium (6) and Eimeria (7). Therefore, we evaluated the role of cysteine proteinases in the pathogenesis of toxoplasmosis. We first cloned a full-length cysteine proteinase gene by amplification from parasite cDNA by 3Ј-and 5Ј-RACE based on conserved sequences of eukaryotic cathepsins and a cDNA sequence from the T. gondii EST data base. The resultant cysteine proteinase gene, tgcp1, has the active site triad of cysteine, asparagine, and histidine (Fig. 1). The deduced amino acid sequence is 44 -46% identical to human (30) and Schistosoma (28) cathepsin B family proteinases (Fig. 1). Toxopain-1 contains an occluding loop sequence used previously to identify cathepsin B, conferring exopeptidase activity ( Fig. 1) (30) and is present as a single copy gene.
Homology modeling for toxopain-1 reveals that the overall fold of model structures is very similar to human liver lysosomal cathepsin and rat cathepsin B, particularly at the active sites (Fig. 2). Both of the Toxoplasma and human cathepsin B enzymes have a glutamic acid in the base of the S2 pocket, suggesting a substrate preference for positively charged amino acids (Fig. 2). To obtain active, recombinant enzyme, we expressed the proenzyme as a thioredoxin fusion protein in E. coli. A number of refolding protocols were evaluated (32), and the resulting purified proteolytically active fractions contained a 32-kDa band consistent with the mature enzyme. The refolded enzyme, therefore, is capable of autocatalytically removing both the prosegment and the thioredoxin fusion (Fig. 3).
The preferred substrate of toxopain-1 is for positively charged amino acids, particularly arginine, in the P1 and P2 positions, consistent with cathepsin B activity. The substrate specificities of purified rtoxopain-1 and tachyzoite lysates (Fig.  4), as well as the proteolytic activity released into the medium (data not shown) were essentially identical, suggesting that toxopain-1 is the major cathepsin B proteinase species in the parasite. To date, no other cathepsin B has been identified in the Toxo EST data base or genome project.
It is interesting to note that, in contrast to other proteases of this family, toxopain-1 shows very little specificity for Phe/Argcontaining substrates but a distinct preference for Arg/Arg substrates. The homology-based models created for the Toxoplasma protease suggest an explanation for the observed kinetic findings. If the coordinates of another rat cathepsin B structure (PDB ID: 1THE), one with a covalently bound inhibitor at the active site, are superimposed on the model structures of TgCP1, the positioning of the structure 1THE has the sequence, Z-arginine-serine-O-benzyl, with Arg in the P2 position. Superimposition upon toxopain-1 models shows that the Arg in the P2 position of the inhibitor would be ideal to form a stabilizing salt bridge with nearby Glu 255 (Fig. 2). A Phe in P2, however, would find no stabilizing interactions, and as a result would be a less favorable substrate. The sequence of rat cathepsin B (1THE) has a nearby tyrosine at position 75 that could conceivably make constructive stacking interactions with a Phe at the P2 position of a substrate. Toxopain-1 does not have this tyrosine residue, further supporting the observed substrate profile of the enzyme.
Studies of the biologic roles of cysteine proteinases have been greatly aided by the development of specific, cell-penetrating inhibitors. The peptidyl ketone inhibitors PRT2005 and PRT2253 have an IC 50 for the T. gondii proteinase of Ͻ1 M and produce no demonstrable effect on host cells. When purified tachyzoites were incubated 30 min in the presence of PRT2005 or PRT2253, invasion was significantly inhibited (Fig. 5).
Immunofluorescent and immunoelectron microscopy studies revealed that toxopain-1 localizes to the rhoptries (Figs. 6 and 7), organelles that play an essential but as yet undefined role in the invasion process of parasites. All nine rhoptry proteins are released as prepro proteins and must be processed to the mature proteins. To determine if toxopain-1 was a key enzyme for ROP processing, we first evaluated the effect of specific cysteine proteinase inhibitors on ROP processing in whole tachyzoites. Greater than 60% inhibition of processing of ROP2, 3, and 4 was detected (Fig. 8). The failure to completely block rhoptry processing might reflect incomplete penetration of the inhibitor or the possible requirement of a proteinase cascade. Ahn et al. (35) and Miller et al. 3 recently reported a subtilisin-like proteinase that localized to the rhoptries and might also play a role in ROP processing.
The critical link between rhoptry biogenesis and toxopain-1 was also demonstrated in localization studies in ROP2 or transport mutants. ROP2 is targeted to rhoptries by an evolutionarily conserved YEQL sequence in its cytoplasmic tail that binds to -chain subunits of the adaptor complex (24). In a dominant negative mutant of the 1-chain of the adaptor complex, 2 toxopain-1 localizes to distorted rhoptries and residual bodies in the parasitophorous vacuole (Fig. 9). Alternatively, when toxopain-1 is inhibited by growing infected monolayers in the presence of cell-permeant cysteine proteinase inhibitors, rhoptries are small and distorted (Fig. 10).
These studies also shed light on the cellular location of ROP protein processing. Earlier pulse-chase experiments found that processing of ROP2, 3, and 4 was completed in ϳ30 min (14) FIG. 9. Overexpressing dominant negative Tg1(D176A) altered rhoptry and cathepsin B trafficking. Indirect immunofluorescence of RH tachyzoites overexpressing wild-type (Tg1-HA) and dominant negative mutant (Tg1(D176A)-HA). Parasites were stained with mAb T3 4A7 to the ROP2 and monoclonal antibody mAb CP1.7 to cathepsin B. The corresponding phase image is shown on the right. and likely occurred in either secretory granules or in rhoptry precursors (36). Soldati et al. (12) found that both brefeldin A and low temperatures blocked processing of ROP1, so the event was likely post-Golgi. Overexpression of a ROP1-Myc fusion suggested that processing occurred in nascent rhoptries of dividing tachyzoites (12). Similarly, processing of ROP2 was proposed to occur within the secretory pathway en route to the rhoptries (11). Earlier experiments and our findings are consistent with processing of ROP1 and 2 in the late secretory pathway or in maturing rhoptries. Lysosomes have not been identified in T. gondii, but the forming and mature rhoptries are the most acidic compartment in the parasite (15). Both the processed rhoptry proteins and toxopain-1 appear to be densely packed in the mature rhoptries, resulting in the intense signals detected by confocal microscopy (Fig. 6) and immunoelectron microscopy (Fig. 7). This is the first evidence for the localization of and the identification of function of any proteinase to the unusual rhoptry organelle. These findings suggest that the rhoptries may be one of the earliest models of a lysosomerelated organelle (37). This may also prove to be an important model for rhoptry protein processing in Plasmodium as well. Similar post-translational processing occurs (38), but the exact site and proteinases involved remain to be determined.
Cathepsin Bs of higher eukaryotes are sorted to lysosomes by mannose-6-P receptors (39). Toxopain-1 does not appear to have mannose 6-P despite two potential glycosylation sites in its prodomain (data not shown). McKerrow and coworkers (40) has identified a prodomain motif in the cathepsin Ls of T. cruzi and Leishmania mexicana, which are required for targeting to lysosomes, but similar signals have not been identified in cathepsin B. We also found that toxopain-1 is released extracellularly into buffer independent of synchronized rhoptry release during invasion. The atypical location of toxopain-1 down the center of the rhoptry (Fig. 7, A and B) suggests a potential subcompartment, which might be released constitutively. Alternatively, toxopain-1 was also detected in endosomal vacuoles (Fig. 7C), which may represent an exosomal pathway of protein excretion (41). Precedence for lysosomal proteinases in a vesicle population comes from Soldati's group 4 and the localization of cathepsin D. The exact signals that target toxopain-1 to the rhoptries or for extracellular release, independently of cell invasion, are under investigation and could lead to important insights into the biogenesis and targeting to a primitive lysosomal system.
The rhoptry organelles are parasite-specific and play a crucial function in host-cell invasion and establishment of the parasitophorous vacuole. Therefore, targeting inhibitors to the unique rhoptries is particularly attractive for blocking parasite invasion and growth. We have now shown that toxopain-1 plays a key role in cellular invasion and thus may be a target for therapeutic intervention. Protease inhibitors are proving to be potent antimicrobial drugs and may provide a strategy to prevent and/or treat Toxoplasma infections. This goal is critically important for toxoplasmosis in which the current optimal therapeutic regimens are toxic, and life-long suppression is required in severely immunocompromised AIDS patients.