Functional Analysis of Toxoplasma gondii Protease Inhibitor 1*

We have characterized a Kazal family serine protease inhibitor, Toxoplasma gondii protease inhibitor 1 (TgPI- 1), in the obligate intracellular parasite Toxoplasma gondii . TgPI-1 contains four inhibitor domains predicted to inhibit trypsin, chymotrypsin, and elastase. Antibodies against recombinant TgPI-1 detect two polypeptides, of 43 and 41 kDa, designated TgPI-1 43 and TgPI-1 41 , in tachyzoites, bradyzoites, and sporozoites. TgPI-1 43 and TgPI-1 41 are secreted constitutively from dense granules into the excreted/secreted antigen fraction as well as the parasitophorous vacuole that T. gondii occupies during intracellular replication. Recombinant TgPI-1 inhibits trypsin, chymotrypsin, pancreatic elastase, and neutrophil elastase. Immunoprecipitation studies with anti-rTgPI-1 antibodies reveal that recombinant TgPI-1 forms a complex with trypsin that is dependent on interactions with the active site of the protease. TgPI-1 is the first anti-trypsin/chymotrypsin

Toxoplasma gondii commonly infects humans and occasionally causes opportunistic disease. Recrudescence of a latent infection in immunodeficient individuals can result in encephalitis (1). Transplacental transmission of T. gondii can cause spontaneous abortion, mental retardation, and blindness (2).
T. gondii is primarily acquired through the ingestion of sporulated oocysts, containing sporozoites, shed by the definitive host (felids) or by ingestion of undercooked meat harboring bradyzoite tissue cysts. The cyst wall is digested during transit through the gastrointestinal tract, releasing the bradyzoites/ sporozoites, which penetrate the intestinal epithelium where they immediately differentiate into rapidly dividing tachyzoites. Tachyzoites disseminate and proliferate during the acute stage of infection before differentiating into bradyzoites, which encyst in muscle tissue and the central nervous system thereby establishing a chronic infection (3).
Because T. gondii transits through the digestive tract and is resistant to physiological levels of trypsin (4), it has been speculated that T. gondii secretes protease inhibitors that aid in protecting the parasite from the proteolytic enzymes, trypsin and chymotrypsin, found within the lower intestine. Previously, TgPI 1 (5) was identified in T. gondii. We have independently identified the same protease inhibitor, termed TgPI-1. Until now, expression of protease inhibitor polypeptides has not been demonstrated in bradyzoites or sporozoites, stages of the parasite most likely to be exposed to the proteolytic environment of the digestive tract.
As obligate intracellular parasites, host cell invasion and subsequent proliferation are essential to the lifecycle of T. gondii. During the lytic cycle of replication parasite attachment to the host cell is accompanied by the discharge of micronemes, small cigar-shaped vesicles located at the apical end of the parasite. Next, rhoptry (ROP) proteins are discharged and serve to aid in the establishment of the parasitophorous vacuole (PV), a protected compartment within the host cell in which the parasites reside. Once inside the host cell, the parasite releases proteins from dense granules (DGs) into the PV. Although ten DG proteins have been identified (6,7), the precise functions of DG proteins remain poorly understood.
N-terminal sequencing of proteins found in excreted/secreted antigen (ESA) fractions reveal the presence of TgPI-1, a protein belonging to the Kazal family of serine protease inhibitors. This family of protease inhibitors is defined by a set of conserved disulfide bonds that form a rigid structure exposing the reactive site P1 residue for interaction with its target protease (8 -10). TgPI-1 contains four inhibitor domains, which are predicted to inhibit chymotrypsin, trypsin, and elastase.
Here we show that TgPI-1 protein is expressed in T. gondii tachyzoites, bradyzoites, and sporozoites and is secreted from DGs. Recombinant TgPI-1 (rTgPI-1) inhibits trypsin, chymotrypsin, pancreatic elastase, and neutrophil elastase. The ability of TgPI-1 to target a broad range of proteases, along with its expression and secretion profile, suggests that it may play a critical function in the lifecycle of T. gondii.
Parasite Culture and Purification-T. gondii strain RH and 2F tachyzoites were serially passaged in vitro in human foreskin fibroblasts grown in DMEM supplemented with 10% fetal bovine serum, 2 mM glutamine, and penicillin/streptomycin (50 g/ml). Parasites were purified by membrane filtration as described previously (16). Bradyzoites and sporozoites were purified as described previously (17,18).
RT-PCR on Tachyzoite and Bradyzoite RNA-Total RNA isolated from encysted bradyzoites and tachyzoites was digested with DNase, and the absence of DNA was checked by PCR before reverse transcription as described (19). One microgram of purified total RNA was reverse-transcribed for 1 h at 42°C in a buffer containing 1 M oligo(dT) 18 primer, 2 mM dNTP, and 25 units of avian myeloblastosis virus reverse transcriptase. The reaction mixture was heat inactivated at 70°C for 15 min. PCR amplifications were performed using 5 units of TaqDNA polymerase in 100-l reaction volumes containing 10 mM Tris-HCl (pH 9.0), 50 mM KCl, 0.1% Triton X-100, 1.5 mM MgCl 2 , 200 M dNTP, and 100 pM of each primer. Thermal cycling conditions were: 1) denaturation at 94°C for 1 min; 2) annealing at 50 -60°C (depending on each primer pair); 3) elongation at 72°C for 2 min; 4) at the end, an additional extension was done at 72°C for 10 min. The primer pair for TgPI-1 was as follows: FTgPI-1.69 (GTGTTTGCTTCGCCCGAAACG) and RTgPI-1.369 (GTTGAGTCGCAATTCGCGAGT).
Identification of TgPI-1 in ESA-Large-scale preparation of ESA proteins was performed by incubating 5 ϫ 10 9 2F strain tachyzoites in 15 ml of DMEM containing 2 mM glutamine, 10 mM HEPES, and 1% ethanol at 37°C for 20 min followed by cooling on ice 5 min. Parasites were removed by centrifugation (1000 ϫ g, 10 min, 4°C), and the supernatant was concentrated to 400 l using C-20 concentrators according to the manufacturer's instructions.
ESA proteins were separated in the first dimension using the PROTEAN isoelectric focusing system employing pH gradient strips. 150 g of ESA protein was mixed with rehydration buffer containing 8 M urea, 4% CHAPS, 0.2% carrier ampholytes, and 2 mM tri-butyl phosphine. Rehydration and isoelectric focusing were performed according to the manufacturer's instructions using an 11-cm ReadyStrip TM (pH 4 -7). Following focusing, proteins were reduced and carboxymethylated by successive 10-min treatments with equilibration buffer containing 2% dithiothreitol followed by 2.5% iodoacetamide. Proteins were resolved in the second dimension on a 10 -20% linear gradient SDS-PAGE gel and stained with colloidal Coomassie Blue (20).
The protein spot corresponding to TgPI-1 (preliminarily identified by Western blotting) was excised, digested in-gel with sequencing grade trypsin, and solvent-extracted as described previously (21). Extracted peptides were dried down and redissolved in 2 l of 50% acetonitrile and 0.3% trifluoroacetic acid, mixed with matrix (␣-cyano-4-hydroycinnamic acid), and deposited on the mass spectrometry (MS) target plate according to a general two-layer method (22). MS was performed on a matrixassisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometer (PerSeptive Biosystem Voyager DE-STR). The spectra of peptide mass fingerprints were acquired in the positive reflection mode with delayed extraction. The spectra were calibrated by external standard sample (ProteoMass peptide calibration kit) deposited on the MS plate adjacent to the Tg-PI-1 spot. Peptide mass fingerprinting data from MALDI-TOF were used to search against NCBI data base via the MS-Fit algorithm (prospector. ucsf.edu). The search resulted in a match to TgPI-1 (nine matching peptides at mass tolerance of 50 ppm, Mowse score 9.15 ϫ 10 5 , 22% sequence coverage).
Secretion Assays-Secretion analyses were performed essentially as described previously (23). Briefly, 3 ϫ 10 9 tachyzoites were resuspended in invasion media (DMEM, 1.5 g/liter sodium bicarbonate, 20 mM HEPES, pH 7.4, and 3% fetal bovine serum) and incubated for 2 min at 37°C with 100 M BAPTA-AM or invasion media alone before being transferred to an ice water bath. Parasites were removed by centrifugation (twice at 1000 ϫ g for 3 min at 4°C), and culture supernatants were concentrated 20-fold in C-10 concentrators, aliquoted, and stored at Ϫ80°C until use.
Extracellular parasites were purified as described above, fixed in 3% formaldehyde, 0.027% glutaraldehyde for 30 min, 4°C, washed twice with 10 ml of PBS, 4°C, and resuspended in 1 ml of PBS. Fixed parasites were added to eight-well chamber slides precoated with poly-L-lysine (0.01% final), incubated at 25°C for 25 min, and then permeabilized with 0.1% Triton X-100, PBS for 15 min at 25°C. Antibody staining was done exactly as described for intracellular parasites.
Vacuole Fractionations-Membranous fractions were isolated from vacuoles as described previously (25). Briefly, tachyzoites were inoculated onto HFF monolayers and incubated 16 h. Monolayers were washed twice with PBS to remove extracellular tachyzoites, and vacuoles were lysed by passage through 23-and 25-gauge syringes. Liberated tachyzoites were removed by centrifugation at 1,000 ϫ g for 10 min at 25°C. The low speed supernatant was centrifuged at 100,000 ϫ g 2 h at 4°C to separate membranous fractions (pellet) and soluble components (supernatant). Equivalent amounts of pellet and supernatant were analyzed by Western blotting.
Kinetic Analysis-Proteases were incubated with increasing concentrations of inhibitor in a volume of 5 l and incubated at 4°C (trypsin, 0.59 pmol; human neutrophil elastase, 20 pmol) or 25°C (pancreatic elastase, 11.58 pmol; chymotrypsin, 0.56 pmol) for 15 min before addition of 175 l of assay buffer (chymotrypsin assay buffer: 0.1 M Tris-Cl, pH 7.4, containing 10 mM CaCl 2 , 1 mM Suc-Ala-Ala-Pro-Phe-pNA; trypsin assay buffer: 50 mM Tris-Cl, pH 7.8, containing 0.1 M NaCl, 0.05% polyethylene glycol 4000, 0.4 mM Suc-Ala-Ala-Pro-Lys-pNA, pancreatic elastase assay buffer; 0.1 M Tris-Cl, pH 8.0, containing 150 mM NaCl, 1 mM Suc-Ala-Ala-Pro-Phe-pNA; human neutrophil elastase buffer: 0.2 mM Tris-Cl, pH 8.0, 0.2 mM Suc-Ala-Ala-Ala-pNA). Experiments were performed at least twice, in triplicate, and reaction velocities were linear over the course of the reaction. Initial velocities were measured by monitoring absorbance at 405 nm on a V max TM Molecular Devices kinetic microplate reader. K i app were determined graphically by plotting where V o is the velocity in the absence of inhibitor, V i is the reaction velocity in the presence of inhibitor and [I] is inhibitor concentration. K i app were converted to K i according to the following formula: Varying concentrations of substrate were incubated with a fixed amount of protease under the conditions described above, and initial velocities were measured by monitoring absorbance at 405 nm. Double-reciprocal Lineweaver-Burke plots of 1/[v] versus 1/[s] were used to determine K m of each substrate for its partner protease.

SDS-PAGE and Western
Blotting-SDS-PAGE was performed on 12.5% minigels (Bio-Rad) under reducing conditions, and proteins were transferred to nitrocellulose membranes for 40 min at 16 V using a Transblot SD semi-dry transfer cell (Bio-Rad). Western blots were developed by chemiluminescence as described previously (26).

TgPI-1 belongs to the Kazal Family of Protease Inhibitors-
Previously, putative secretory proteins were identified utilizing sub-cellular fractionation techniques and isolation of excreted/ secreted antigens (ESA) in parasite culture supernatants (27). N-terminal sequencing of putative secretory proteins revealed the presence of T. gondii protease inhibitor 1 (TgPI-1) in ESA. BLAST analysis of the T. gondii expressed sequence tag (EST) database with the TgPI-1 nucleotide sequence reveals that the TgPI-1 sequence was present in the database with 13 tachyzoite ESTs and no bradyzoite ESTs.
The TgPI-1 gene is predicted to encode a 30-kDa protein containing four protease inhibitor domains and an N-terminal signal sequence (Fig. 1A). TgPI-1 is predicted to be a serine protease inhibitor belonging to the Kazal family whose members characteristically have multiple protease specificity domains (8 -10). This family of protease inhibitors is defined by a set of conserved disulfide bonds that form a rigid structure exposing the reactive site P1 residue for interaction with its target protease (Fig. 1A). Alignment of the inhibitor domains of TgPI-1 with other Kazal inhibitor domains reveals a highly conserved placement of cysteine residues predicted to form the disulfide bonds important to the three-dimensional structure of each domain (Fig. 1B). Kazal inhibitors act as pseudosubstrates, binding tightly to their target proteases (8 -10).
TgPI-1 mRNA Is Detected in Both Bradyzoites and Tachyzoites, and TgPI-1 Protein Is Expressed in Tachyzoites, Bradyzoites, and Sporozoites-RT-PCR was used to assess whether mRNA for TgPI-1 was present in tachyzoites, the rapidly dividing form of the parasite, and bradyzoites, the slow growing form of the parasite. RNA was purified from tachyzoites and bradyzoites and used in RT-PCR with TgPI-1specific primers. With both tachyzoite and bradyzoite mRNA, RT-PCR with TgPI-1-specific primers yielded the expected product of 300 bp ( Fig. 2A, lanes 3 and 4). As a control, the housekeeping gene T. gondii superoxide dismutase was included ( Fig. 2A, lanes 1 and 2). These results indicate that mRNA for TgPI-1 is present in both tachyzoites and bradyzoites.
TgPI-1 43 and TgPI-1 41 Secretion into ESA Is Unaffected by BAPTA-AM-TgPI-1 possesses a hydrophobic N-terminal sequence that has the properties of a signal sequence, suggesting that it might be a secretory protein. In an effort to determine whether TgPI-1 was secreted, two-dimensional gels of T. gondii ESA were transferred to nitrocellulose and probed with R␣rTgPI-1 (Fig. 3A). The polypeptide recognized by R␣rTgPI-1 was excised from a sister gel stained with colloidal Coomassie Blue (Fig. 3B), digested with trypsin, and subjected to matrixassisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF). The resulting peptide mass fingerprint was used to search the NCBI database and confirm the identity of TgPI-1. These results suggest that TgPI-1 is secreted by T. gondii.
ESA fractions consist predominately of secreted micronemal and DG proteins. Calcium chelators such as BAPTA-AM specifically block microneme secretion without markedly affecting DG secretion (16). To determine if TgPI-1 43 and TgPI-1 41 are of micronemal or DG origin, ESA fractions obtained from untreated and BAPTA-AM treated parasites were resolved by SDS-PAGE and probed with R␣rTgPI-1. As a control, release of the micronemal protein, TgMIC2 (115 kDa), which is proteo- FIG. 1. TgPI-1 belongs to the Kazal family of serine protease inhibitors. A, TgPI-1 contains a signal sequence and four inhibitor domains. TgPI-1 has an N-terminal signal sequence (SP) and four inhibitor domains that have been modeled according to the threedimensional structure of leech-derived trypsin inhibitor using Swiss-Model software. This model is intended to schematically depict the general structure of the inhibitor domains and has not been rigorously optimized. Each domain is defined by three disulfide bonds, generating a rigid structure in which the P1 residue (indicated by the single-letter amino acid code) is left exposed to interact with its cognate protease lytically processed to sTgMIC2 (95 kDa) coincident with its release into ESA (23), was assessed. Secretion of sTgMIC2 into ESA is completely blocked when parasites are pretreated with BAPTA-AM (Fig. 3C, lane 4). In contrast to TgMIC2, the DG protein, GRA1, is present in extracellular tachyzoites (Fig. 3D, lane 1), as well as ESA from untreated and BAPTA-AM-treated parasites (Fig. 3D, lanes 3 and 4), indicating the constitutive nature of its secretion. Like GRA1, TgPI-1 43 and TgPI-1 41 are present in the ESA of untreated as well as BAPTA-AMtreated parasites (Fig. 3E, lanes 1, 3, and 4). These results are consistent with TgPI-1 43 and TgPI-1 41 being secreted into ESA via DGs.
Dual label indirect immunofluorescence of extracellular tachyzoites was used to confirm the DG localization of TgPI-1. R␣rTgPI-1 exhibited punctate staining throughout the parasite (Fig. 4A), and this staining pattern perfectly overlaps with the distribution of the DG protein, GRA1 (as detected with mAb Tg17-43). Punctate granular staining was also observed with single labeling for TgPI-1 and GRA1 (data not shown). These results confirm the DG localization of TgPI-1 in extracellular tachyzoites.
TgPI-1 Secretion into the PV Follows the Kinetics of DG Discharge-After T. gondii has invaded host cells, DG proteins discharge from lateral sites near the apical end of the parasite, diffuse and surround the parasites, and eventually disperse throughout the PV. To date, ten DG proteins (GRA1 through GRA8 and two isoforms of NTPase) have been identified. It has been speculated that DG proteins play a role in the maintenance and remodeling of the PV as well as nutrient acquisition (28). Dual label immunofluorescence of timed invasions was used to assess the kinetics of TgPI-1 secretion into the PV.
Freshly isolated tachyzoites were inoculated onto a monolayer of human foreskin fibroblasts (HFF) for 1 min and incubated for varying intervals of time before they were fixed, permeabilized, and probed with R␣rTgPI-1 and a mouse mAb against GRA1 (mAbTg17-43) or GRA4 (WU# 950). Ten minutes post-invasion, R␣rTgPI-1 and mAb Tg17-43 exhibited intense staining on either side of the apical end of the parasites (Fig. 4B) in 45% (45/101) of the PVs observed. At 60 min post-invasion, TgPI-1 and GRA1 are redistributed, completely surrounding the parasite residing within the PV (Fig. 4C) in 82% (125/152) of PVs, and by 24 h post-invasion, TgPI-1 and GRA4 completely filled the PV (Fig. 4D) in virtually all PVs (not quantified). Similar patterns were also observed in singlelabeling experiments using R␣rTgPI-1, mAb Tg17-43, or WU# 950 (data not shown). These results indicate that the kinetics of TgPI-1 secretion into the PV follows the timing of DG discharge.
TgPI-1 Remains Soluble after Secretion into the PV-After secretion into the PV, DG proteins can be targeted to the PV membrane (29,30) or the intravacuolar network of tubular membranes within the PV (25) or remain soluble (31). The solubility characteristics of a DG protein are important for predicting whether the protein will remain confined to the host cell after parasite egress or whether it will be released to

FIG. 3. TgPI-1 secretion into ESA is calcium independent. A,
Western blot of two-dimensional gel electrophoresis of ESA. ESA fractions were separated in the first dimension using PROTEAN isoelectric focusing system employing pH gradient strips (pH 4 -7) and then resolved in the second dimension on a 10 -20% linear gradient SDS-PAGE gel, transferred to nitrocellulose, and probed with rabbit TgPI-1 antibodies. An asterisk marks residual signal from a previous blot. B, colloidal Coomassie stain of two-dimensional gel electrophoresis of ESA. A sister two-dimensional gel of ESA was run as described in A and stained with colloidal Coomassie. The circled band is TgPI-1. potentially act in the extracellular environment of the tissues surrounding the site of infection. Differential centrifugation was used to determine the intravacuolar localization of TgPI-1 43 and TgPI-1 41 .
HFF monolayers were infected with tachyzoites for 24 h and washed to remove extracellular parasites. Vacuoles containing 8 -16 parasites were mechanically lysed and centrifuged at low speed to remove the intact parasites. The low speed supernatant was centrifuged at 100,000 ϫ g to pellet membranes containing PV membrane and tubular membrane networks. The soluble and membrane fractions were analyzed by Western blotting. As controls, GRA1, a soluble vacuole component (31), and GRA2, which distributes evenly between the membrane and soluble fractions (32), were analyzed. As expected, GRA1 was observed exclusively within the soluble fraction (Fig. 5A, lanes 2  and 3), whereas GRA2 was distributed equally between the membrane and soluble fractions (Fig. 5B, lanes 2 and 3). TgPI-1 43 and TgP1-1 41 were detected exclusively within the soluble fractions (Fig. 5C, lanes 2 and 3). These results indicate that TgPI-1 43 and TgPI-1 41 remain soluble after secretion into the PV. In addition, TgPI-1 43 and TgP1-1 41 do not appear to be proteolytically processed or otherwise modified upon their release into the PV.
Recombinant TgPI-1 Is a Broad-spectrum Serine Protease Inhibitor-Based on homology with other Kazal family mem-bers, TgPI-1 is predicted to contain four serine protease inhibitor domains (Fig. 1A). In the Kazal family, the cognate protease is dictated by the amino acid occupying the P1 position. TgPI-1 has two domains where arginine residues, predicted to inhibit trypsin, occupy the P1 position and two additional domains containing leucine residues at the P1 position. Leucine at P1 has been observed to inhibit chymotrypsin and elastase (reviewed in Refs. 8, 10, and 33).
Recombinant TgPI-1 (rTgPI-1) was expressed, purified from Escherichia coli, and tested for inhibitory activity against trypsin, chymotrypsin, and elastase. Proteases were incubated with increasing amounts of inhibitor, and K i values were determined by monitoring cleavage of the peptide substrate.
As predicted on the basis of the P1 position of each domain, rTgPI-1 inhibited chymotrypsin, trypsin, pancreatic elastase, and neutrophil elastase. The strongest inhibitory activity was observed against trypsin and chymotrypsin with K i values of 0.035 and 0.35 nM, respectively (Table I). Recombinant TgPI-1 (rTgPI-1) also inhibits pancreatic elastase as well as human neutrophil elastase. Although the K i values for pancreatic elastase (K i ϭ 15 nM) and human neutrophil elastase (K i ϭ 49 nM) were markedly higher than the K i values observed for both trypsin and chymotrypsin (Table I), both proteases were significantly inhibited by rTgPI-1. Pancreatic elastase was inhibited 80% when preincubated with a 2-fold molar excess of inhibitor, whereas neutrophil elastase was inhibited 50% in the presence of a 2-fold molar excess and almost 80% when 4-fold FIG. 4. The kinetics of TgPI-1 secretion into the PV follows DG discharge. Dual labeling indirect immunofluorescence of T. gondii. Freshly harvested extracellular tachyzoites (A) were fixed, and detergent permeabilized and stained with antibodies to GRA1 (mAbTg17-43) and TgPI-1 (R␣rTgPI-1). For intracellular parasites, monolayers were fixed, semi-permeabilized to reveal the PV contents, and stained with antibodies to DG proteins GRA1 (mAb Tg17-43) at 10 min (B) and 1 h (C) post-invasion or GRA4 (mAb 4G1.AH11) at 24 h (D) postinvasion R␣rTgPI-1. Oregon-green conjugated goat anti-rabbit and Texas-red conjugated goat anti-mouse antibodies were used for detection. Merged images were deconvolved using C-Imaging Systems software and overlaid onto the corresponding phase contrast images. Bar, 5 m.  -43). B, GRA2 is equally distributed between the soluble and membrane fractions. The membrane from A was stripped and probed with antibodies against GRA2 (R␣GRA2). C, TgPI-1 43 and TgPI-1 41 remain in the soluble vacuole fraction. The membrane from A was stripped and probed with R␣rTgPI-1. molar excess of inhibitor was added (data not shown).
It is possible that the inhibition observed was actually due to degradation of the test protease by a contaminating E. coli protease that co-purified with our inhibitors. To rule out this possibility, proteases were incubated with the inhibitors under the same conditions used to assess inhibition. Protease/inhibitor mixes were then separated on SDS-PAGE gels and stained to confirm that the test protease was intact. In all cases, the protease amounts were unchanged after incubation with the inhibitors (data not shown).

Recombinant TgPI-1 Forms a Stable Complex with Trypsin through Active Site Interactions-The closely related Kazal and
Kunitz inhibitors act as pseudosubstrates consisting of an exposed reactive loop, which fits in the active site cleft of the target enzyme (33). In most cases binding is rapid and tight. If TgPI-1 inhibits trypsin by the mechanism described above, it should be possible to isolate trypsin-rTgPI-1 complexes. Additionally, modification of the trypsin active site should inhibit complex formation. Trypsin-rTgPI-1 complexes were isolated by incubation of inhibitor with enzyme followed by immunoprecipitation with mouse anti-rTgPI-1 antibodies. Immunoprecipitated material was then analyzed by Western blotting with rabbit anti-trypsin antibodies. As expected for complex formation, purified trypsin co-immunoprecipitated with rTgPI-1 and mouse anti-rTgPI-1 antibodies (Fig. 6A, lane 3). Trypsin was not detected when mouse anti-rTgPI-1 antibodies were used in immunoprecipitations in the absence of rTgPI-1 (Fig. 6A, lane  2). Preincubation of trypsin with (p-4-amidinophenyl)methylsulfonyl fluoride, a compound that covalently modifies the trypsin active site (34), blocked complex formation (Fig. 6B, lane 4). Furthermore, preincubation of trypsin with soybean trypsin inhibitor, an inhibitor known to interact with the active site of trypsin (35,36), blocked complex formation between rTgPI-1 and trypsin (Fig. 6B, lane 3). Preincubation of trypsin with buffers alone did not interfere with the ability of rTgPI-1 to form a complex with trypsin (Fig. 6B, lanes 1 and 2). These experiments indicate that rTgPI-1 forms a complex with purified trypsin through active site interactions. DISCUSSION Serine proteases and their cognate inhibitors play fundamental roles in cellular biology with functions ranging from digestion, apoptosis, proteolytic processing, wound healing, and cellular and extracellular remodeling (8 -10, 33). Serine protease inhibitors can be grouped into several families, including the SERPIN, Kazal, and Kunitz families of inhibitors, based on their structural homologies and mechanism of inhibition. The SERPINS have been most extensively studied, whereas little work has been published on the cellular biology of the Kazal family of inhibitors.
Kazal inhibitors have been identified in a wide variety of organisms, including birds, mammals, insects, and leeches (8,(13)(14)(15). Because Kazal inhibitors act as pseudosubstrates, they have been used extensively to elucidate interactions between serine proteases and their substrates (8 -10, 33); however, little work has centered on the in vivo function of this family of inhibitors. We have identified a Kazal inhibitor, TgPI-1, in T. gondii, a parasite amenable to genetic manipulation and easily cultured in vitro under a wide variety of conditions.
Western blots indicate that two polypeptides, named TgPI-1 43 and TgPI-1 41 according to their apparent molecular weight, are detected in Toxoplasma cell lysates by antibodies raised against recombinant TgPI-1. Currently, the relationship between TgPI-1 43 and TgPI-1 41 is unknown. TgPI-1 43 may represent the full-length polypeptide that is proteolytically processed to the smaller form, TgPI-1 41 . Alternatively, it is possible that the full-length TgPI-1 41 is glycosylated to TgPI-1 43 . The ratio between TgPI-1 43 and TgPI-1 41 can vary between lysate preparations, although it is unclear why this is the case. The presence or absence of protease inhibitors in the parasite purification protocol does not affect the ratio of TgPI-1 43 to TgPI-1 41 (data not shown), suggesting that the two species are not generated during sample preparation.
TgPI-1 43 and TgPI-1 41 are secreted into ESA fractions by extracellular tachyzoites in a BAPTA-AM-insensitive manner, a characteristic of DG proteins. Additionally, indirect immunofluorescence of T. gondii as it invades host cells reveals that TgPI-1 is secreted into the parasitophorous vacuole by intracellular parasites. It is likely that TgPI-1 is also secreted by  bradyzoites and sporozoites in a manner similar to that observed in tachyzoites. Dense granule proteins are constitutively secreted by extracellular tachyzoites, and secretion into the PV is slightly up-regulated upon invasion (24). In vitro studies indicate that rTgPI-1 is active against trypsin, chymotrypsin, pancreatic elastase, and neutrophil elastase. Knowledge regarding the expression and secretion of TgPI-1 allows us to predict what types of host cell tissues and proteases this broad-spectrum inhibitor is likely to encounter.
T. gondii infection is most often acquired through the ingestion of sporozoites, encapsulated within sporulated oocysts, or by ingestion of bradyzoite-laden tissue cysts contained within the muscle tissue of an infected animal. Tachyzoites and bradyzoites are resistant to physiological concentrations of trypsin (4), and it has been hypothesized that protease inhibitors may be responsible for this resistance. In vitro studies here show that recombinant TgPI-1 potently inhibits trypsin and chymotrypsin. Bradyzoite and sporozoite expression of TgPI-1 is the first example of a protease inhibitor expressed in a stage of the parasite exposed to the digestive enzymes in the lower intestine. Furthermore, dense granule proteins secreted by bradyzoites and sporozoites (including TgPI-1) would be expected to come into contact with secreted proteases present within the small intestine (trypsin and chymotrypsin) as the parasites traverse the digestive tract. Upon examination of the lifecycle of T. gondii, however, there are other potential functions for TgPI-1.
After bradyzoites/sporozoites penetrate the intestinal epithelium, they differentiate into rapidly dividing tachyzoites, which disseminate throughout the body, proliferating within the PV of infected cells (3). Some tachyzoites probably replicate in intestinal epithelial cells shortly after initiation of infection. In this case, soluble TgPI-1 released from the vacuole upon host cell lysis could help protect the parasite from digestive proteases of the gastrointestinal tract before it enters adjacent cells. This protection would provide the parasite an opportunity to proliferate within the intestinal epithelium before disseminating throughout the host organism.
Interestingly, with the exception of the rhoptry protein, TgROP1, which is degraded within hours of invasion, none of the known DG or rhoptry (ROP) proteins is proteolytically modified after secretion into the PV. This suggests that the PV is largely devoid of proteolytic activity. However, tachyzoites express at least two surface proteases (23) and one ROP-derived protease (TgSUB2) 2 that presumably occupy the PV. Thus, TgPI-1 may serve to inactivate parasite-derived proteases in the PV, thereby preventing them from inappropriately processing vacuolar protein substrates. It is also conceivable that host-derived proteases occupy the PV and TgPI-1 might additionally function to neutralize these proteases and prevent them from degrading the vacuolar contents.
Neutrophils are one of the key immune effector cells of the innate immune response that forms the first line of defense against microbial infections. Recent studies suggest that chemokines rapidly recruit host neutrophils to the site of T. gondii infection where these cells play an essential role in controlling the early infection, possibly by secreting interleukin-12 to initiate the type 1 cytokine immune response (37)(38)(39). However, despite being capable of deploying a battery of anti-microbial products, including cytokines, proteases, reactive oxygen intermediates, and nitric oxide, neutrophils fail to prevent T. gondii tachyzoites from establishing the early infection. This suggests the parasite possesses the ability to neutralize or counteract anti-microbial products released by neutrophils. Our observa-tion that TgPI-1 effectively inhibits neutrophil elastase activity suggests that it could contribute to the ability of the parasite to survive the innate immune response and establish long term infection. TgPI-1 secreted by extracellular tachyzoites, as well as any soluble TgPI-1 released from the PV upon host cell lysis, could inhibit proteases secreted by neutrophils recruited to the site of infection. In this scenario, TgPI-1 could act locally as an anti-inflammatory agent, inhibiting neutrophil proteases such as elastase.
SERPINS, the most extensively studied family of serine protease inhibitors, have been shown to influence viral pathogenesis by dictating viral host range, decreasing inflammation and inhibiting apoptosis (11, 40 -42). In addition, a member of the Kunitz family of serine protease inhibitors expressed by the hookworm Ancylostoma ceylanicaum has been shown to inhibit chymotrypsin, pancreatic elastase, neutrophil elastase, and trypsin. The A. ceylanicaum Kunitz inhibitor (AceKI-1) may protect the parasites during their transit through the digestive system or provide protection against the host immune system (12). However, the natural target proteases for these microbial SERPINS have not been identified. Kazal inhibitors have been used extensively in vitro to study the interactions of serine proteases with their substrates and are known to bind quickly and tightly to their target proteases. Exploitation of this characteristic tight binding may allow for identification of the natural target protease of TgPI-1 in vivo through co-immunoprecipitation studies. Indeed, we have shown here that rTgPI-1 forms a complex with purified trypsin that is stable under relatively stringent immunoprecipitation conditions (1 M Tris, 1% Triton-X-100, 5% sodium deoxycholate, 0.2% SDS, 100 mM NaCl 2 , 5 mM EDTA). Additionally, the ability to generate geneknockout mutants in T. gondii should provide additional insight into the function of Kazal inhibitors in T. gondii and maybe provide evidence for functions in other systems.