INFECTION AND IMMUNITY, Oct. 2004, p. 58925902 0019-9567/04/$08.00 0 DOI: 10.1128/IAI.72.10.58925902.2004 Copyright 2004, American Society for Microbiology.

. All Rights Reserved.

Vol. 72, No. 10

A New Cruzipain-Mediated Pathway of Human Cell Invasion by Trypanosoma cruzi Requires Trypomastigote Membranes
Isabela M. Aparicio, Julio Scharfstein, and Ana Paula C. A. Lima*
Instituto de Biof sica Carlos Chagas Filho, Bloco G, CCS, Universidade Federal do Rio de Janeiro, Ilha do Fundao, Rio de Janeiro, Brazil
Received 19 October 2003/Returned for modication 17 January 2004/Accepted 24 June 2004

The intracellular protozoan Trypanosoma cruzi causes Chagas disease, a chronic illness associated with cardiomyopathy and digestive disorders. This pathogen invades mammalian cells by signaling them through multiple transduction pathways. We previously showed that cruzipain, the main cysteine protease of T. cruzi, promotes host cell invasion by activating kinin receptors. Here, we report a cruzipain-mediated invasion route that is not blocked by kinin receptor antagonists. By testing different strains of T. cruzi, we observed a correlation between the level of cruzipain secreted by trypomastigotes and the capacity of the pathogen to invade host cells. Consistent with a role for cruzipain, the cysteine protease inhibitor N-methylpiperazineurea-Phe-homophenylalanine-vinylsulfone-benzene impaired the invasion of human smooth muscle cells by strains Dm28c and X10/6 but not by the G isolate. Cruzipain-rich supernatants of Dm28c trypomastigotes enhanced the infectivity of isolate G parasites twofold, an effect which was abolished by the cysteine protease inhibitor L-trans-epoxysuccinyl-leucylamido-(4-guanidino)butane and by thapsigargin, a drug that induces depletion of the intracellular Ca2 stores. The enhancement due to Dm28 supernatants was abolished upon cruzipain immunodepletion, and the activity was restored by puried cruzipain. In contrast, supernatants from isolate G trypomastigotes (with low levels of cruzipain) or supernatants from Dm28c epimastigotes or puried cruzipain alone did not enhance parasite invasion, indicating that the protease is required but not sufcient to engage this invasion pathway. We provide evidence that activation of this pathway requires cruzipainmediated processing of a trypomastigote molecule associated with parasite-shed membranes. Our results couple cruzipain to host cell invasion through a kinin-independent route and further suggest that high-level cruzipain expression may contribute to parasite infectivity. The agellated protozoan Trypanosoma cruzi is the causative agent of Chagas disease, a pathology characterized by chronic inammation associated with cardiomyopathy and/or digestive disorders (29). When transmitted to mammals by hematophagous triatomines, metacyclic trypomastigotes are capable of invading diverse cell types, where they differentiate into amastigotes. These parasite forms then multiply by binary division in the cell cytoplasm before they retransform into trypomastigotes. After 3 to 5 days, the infected cells burst, releasing infective trypomastigotes into adjacent tissues. The penetration of nonphagocytic cells by T. cruzi is a complex event involving multiple signaling pathways, which depend on the parasite isolate-host cell combination or the nature of the infective form (metacyclic versus tissue culture trypomastigotes) (reviewed in reference 2). One of the best-characterized invasion pathways requires elevation of the host cell intracellular Ca2 concentration, which leads to synaptotagmin VII-dependent lysosome migration and fusion to the parasite attachment site, an event that precedes the formation of the parasitophorous vacuole (4, 25). Two parasite peptidases induce Ca2 transients in the host cell by signaling through the following G-protein-coupled receptors (GPCRs): (i) oligopeptidase B (OPB), a serine peptidase; and (ii) cruzipain, the parasites main papain-like cysteine protease (CP) (1, 22, 23).
* Corresponding author. Mailing address: Instituto de Biof sica Carlos Chagas Filho, Bloco G, CCS, UFRJ, Ilha do Fundao, Rio de Janeiro, RJ, CEP 21-944-900 Brazil. Phone: (55)(21)(2562-6709). Fax: (55)(21)2280-8193. E-mail: anapaula@biof.ufrj.br. 5892

Cruzipains comprise a family of closely related isoforms expressed as zymogens which undergo maturation upon proteolytic removal of the N-terminal domain (5, 9, 15, 16). These enzymes are abundantly expressed throughout the parasites life cycle and accumulate in acidic lysosome-like organelles designated reservosomes. In previous studies in which membrane-permeable synthetic irreversible CP inhibitors were used, we and others have associated cruzipains activity with the growth and differentiation of epimastigotes and amastigotes (13, 19). Although these drugs partially impaired host cell invasion by trypomastigotes, their lack of selectivity and easy access to host cell intracellular compartments precluded identication of a denite role for cruzipain in invasion. Later, the three-dimensional structure of the recombinant form of a family prototype, cruzain, enabled investigators to design more selective and highly potent synthetic inhibitors (18) which protected mice from lethal infections with T. cruzi (10). Recently, kinin peptides and the cognate GPCRs B2 and B1were identied as members of a cruzipain-driven activation pathway involved in T. cruzi signaling and invasion of endothelial cells and cardiomyocytes (23, 24, 27). These studies revealed that the activation of the B2 constitutive receptors by trypomastigotes is modulated by the angiotensin converting enzyme, a potent kinin-degrading peptidase (24). The use of captopril, an angiotensin converting enzyme inhibitor, potentiates invasion of cells expressing B2 receptors. However, in previous studies, CP inhibitors impaired host cell invasion in culture conditions that did not favor overt activation of the kinin system (19). In this study we revisited this issue, and in

VOL. 72, 2004

CRUZIPAIN-MEDIATED HOST CELL INVASION BY T. CRUZI

5893

this paper we describe a new cruzipain-mediated invasion route, which is not related to the kinin pathway. We demonstrated that invasion of human smooth muscle cells by T. cruzi isolates Dm28c and X10/6, but not by the G isolate, is largely dependent on the activity of cruzipain secreted by trypomastigotes into the extracellular millieu. Furthermore, we obtained evidence that the extracellular enzyme acts on a trypomastigote-associated molecule, leading to more efcient invasion of host cells by isolate G trypomastigotes. Taken together, these results connect cruzipain to host cell signaling and invasion through an alternative route and suggest that the endogenous levels of this enzyme may contribute to T. cruzi infectivity.
MATERIALS AND METHODS Cell cultures. Vero and LLCMK2 were cultivated in Dulbeccos modied Eagles medium (DMEM) (Sigma), and rabbit aorta endothelial cells (provided by H. Nader, UNIFESP, Sao Paulo, Brazil) and CHO cells were cultivated in Hams F-12 medium (Sigma), both of which were supplemented with 10% fetal calf serum (FCS) (Gibco BRL), at 37C in a humidied 5% CO2 atmosphere. Primary cultures ( 20 passages) of human smooth muscle cells originated from the stomach of an adult male and were purchased from the Cell Bank of Rio de Janeiro (Rio de Janeiro, Brazil). These cells were cultivated in DMEM supplemented with 10% FCS as described above. Tissue culture trypomastigotes were obtained from the supernatants of infected LLCMK2 cells cultivated in DMEM supplemented with 2% FCS. To ensure infectivity, Dm28c and Sylvio X10/6 trypomastigotes were inoculated into Swiss mice, and blood-derived trypomastigotes were used to reestablish in vitro cultures. We were not able to recover any trypomastigotes from mice inoculated with the G isolate, conrming previous ndings that infection with this parasite is subpatent (30). Epimastigotes were cultivated in liver infusion tryptose supplemented with 10% FCS at 28C until the mid-log phase. Antibodies. Rabbit anti-cruzipain serum was obtained as described previously (16). Anti-OPB serum was a gift from N. Andrews (Yale University, New Haven, Conn.). T. cruzi lysates and puried proteins. Log-phase epimastigotes or trypomastigotes were washed twice with Hanks balanced salt solution containing 1 mM glucose (HBSS) and collected by centrifugation at 3,000 g, and the parasite pellets were kept at 20C until they were processed. The pellets were resuspended in 10 mM Na2HPO4150 mM NaCl (pH 7.2) (PBS), and lysates were obtained by repeated cycles of freezing and thawing, followed by addition of Triton X-100 to a concentration of 1%. After incubation for 10 min on ice, the samples were cleared by centrifugation at 13,000 g for 15 min, and the protein concentration of the soluble fraction was determined by using a Dc-Protein assay kit II (Bio-Rad). Cruzipain was puried from strain Dm28c epimastigotes as described by Murta et al. (22). Western blots and biotinN-Pip-F-hF-VSPh blots. For Western blots, 50- g portions of parasite lysates were resolved on sodium dodecyl sulfate (SDS)11% polyacrylamide gel electrophoresis (PAGE) gels, transferred to nitrocellulose, and blocked with 9% nonfat milk in PBS containing 0.05% Tween 20. Antisera (anti-cruzipain and anti-OPB antisera) were incubated at a 1:1,000 dilution in PBS containing 0.05% Tween 20 for 1 h. The densitometry of the reactive bands was analyzed by using the ImageQuant 5.2 program (see Fig. 3D). For biotin N-methylpiperazine-urea-Phe-homopheylalanine-vinylsulfone-benzene (N-PipF-hF-VSPh) blots, 50- g portions of lysates were incubated for 30 min at 37C with 5 mM dithiothreitol (DTT) (Bio-Rad) to activate CP, since these enzymes need to maintain the active-site cysteine in its reduced state to ensure activity. Then the samples were supplemented with 10 M biotinN-Pip-F-hF-VSPh and incubated for 60 min at 37C. Reactions were stopped by addition of SDS-PAGE sample buffer (62.5 mM Tris-HCl [pH 6.8], 2% SDS, 10% glycerol, 5% -mercaptoethanol, 0.012% bromophenol blue), and the mixtures were subsequently boiled for 3 min. Samples were loaded into SDS11% PAGE gels, transferred to nitrocellulose membranes, blocked with nonfat milk, and incubated for 60 min with alkaline phosphatase-conjugated streptavidin. Enzymatic assays. T. cruzi supernatants and lysates were tested for activity by using the uorogenic synthetic peptide substrate caproyl-leucyl-(S-benzyl)-cysteinyl-7-amido-4-methylcoumarin [-NH2-(Cap)L-(SBzl)C-AMC], synthesized as previously described (7), or carbobenzoxy-phenylalanyl-arginyl-7-amido-4methylcoumarin (Sigma) (see Fig. 8A,inset), at a concentration of 10 M in a solution containing 50 mM Na2HPO4, 100 mM NaCl, 5 mM EDTA (pH 6.5), 5%

dimethyl sulfoxide (DMSO), and 2.5 mM DTT at 37C. Hydrolysis of the substrate was monitored by measuring the uorescence with excitation at 380 nm and emission at 440 nm by using an F4500 uorimeter (Hitachi). The initial velocities were calculated as described by Lima et al. (16). T. cruzi supernatants. Tissue culture trypomastigotes or log-phase epimastigotes were washed twice in HBSS and incubated in this solution at a concentration of 2 107 parasites/ml for 2 h at 37C (trypomastigotes) or 28C (epimastigotes). The parasites were centrifuged at 3,000 g for 15 min, and parasite-free supernatants were ltered with 0.22- m-pore-size lters (Millipore). The supernatants were kept at 20C. Catalytically active CP must retain the active-site cysteine in a reduced state. Therefore, supernatants were incubated with 2.5 mM DTT for 5 min at 37C to activate CP before use. In preactivation experiments (see Fig. 7 and 8), supernatants were incubated with 2.5 mM DTT at 37C for 3 h, and this was followed by inactivation of cruzipain by addition of L-trans-epoxysuccinylleucylamido-(4-guanidino)butane (E-64) to a nal concentration of 30 M. Soluble and membrane fractions from trypomastigote supernatants were fractionated by centrifugation at 100,000 g for 1 h at 4C. The pellet (membrane fraction) was resuspended in HBSS containing 0.1% bovine serum albumin (BSA) (Sigma) to the original volume. The supernatants were fractionated as described above (see Fig. 8B), and the membrane fractions were resuspended to the original volume in HBSS and subsequently incubated at 37C for 3 h with 2.5 mM DTT and 5 nM puried cruzipain or with the enzyme in the presence of 30 M E-64, as indicated below. Following incubation, the samples were supplemented with 30 M E-64 before addition to invasion assay mixtures. Immunoprecipitation. Fifty microliters of a protein A-agarose slurry (Sigma) was washed three times with PBS (pH 7.2) and incubated overnight in the same buffer containing 150 g of puried rabbit immunoglobulin G (IgG) anti-cruzipain or puried rabbit IgG from nonimmunized controls. The beads were washed six times with 1 ml of PBS0.05% Tween 20 and incubated for 2 h with 600 l of trypomastigote supernatant, obtained as described above. The beads were then washed twice and resuspended in 120 l of PBS; 100 l was assayed for CP activity. Invasion assays. Invasion assays were performed as described previously (23). Briey, the invasion assays were performed in 0.5 ml of DMEM containing 0.1% BSA (Sigma) for 3 h at 37C, and the number of intracellular parasites was estimated on Giemsa-stained coverslips by microscopy. Trypomastigotes were added at a parasite/host cell ratio of 3:1 unless indicated otherwise. In experiments in which puried cruzipain was used, the stock enzyme (200 nM) was incubated with 2.5 mM DTT at 37C for 5 min and diluted to obtain a nal concentration of 5 nM in the invasion medium. In experiments in which T. cruzi supernatants were used, the samples were incubated with 2.5 mM DTT for 5 min at 37C, and 100- l aliquots were added to host cells immediately before addition of parasites. HBSS supplemented with 2.5 mM DTT was used as a control in experiments in which no supernatants were added. Statistical analyses were done by using analysis of variance, and a signicance level of P 0.05 was used. Mobilization of intracellular calcium stores. Semiconuent human smooth muscle cells plated on coverslips were treated with 0.5 M thapsigargin in DMEM containing 10% FCS for 40 min at 37C. The cells were subsequently washed three times with HBSS, and the invasion assay was performed in DMEMBSA at a parasite/host cell ratio 50:1 for 1 h at 37C. This time interval was used to avoid the reloading of intracellular stocks with calcium after the removal of thapsigargin.

RESULTS Highly infective T. cruzi isolates depend on functional CP for host cell invasion. The membrane-permeable irreversible synthetic inhibitor N-Pip-F-hF-VSPh was used as a tool to explore the participation of cruzipain in invasion routes that are not related to the activation of kinin receptors. To establish the most appropriate model, we analyzed the invasion of several cell types in vitro by tissue culture trypomastigotes of three T. cruzi isolates (Fig. 1A). The isolates tested could be clearly divided into highly infective isolates (strains Dm28c and X10/6) or an isolate with a low level of infectivity (strain G), irrespective of the cell type tested. The infectivity of strain G parasites was consistently lower than that of Dm28c parasites, even with prolonged interaction times (Fig. 1B) or at different parasite/host cell ratios (Fig. 1C). Primary cultures of human

FIG. 1. Differential infectivity of T. cruzi isolates. (A) Tissue culture trypomastigotes were incubated with monolayers containing different cell types at a parasite/host cell ratio of 3:1 in the appropriate medium (see Materials and Methods for details) supplemented with 0.1% BSA for 3 h at 37C. (B and C) Invasion of human smooth muscle cells by the Dm28c and G isolates as described above for panel A at different interaction times (B) or at different parasite/host cell ratios (C). The data are representative of the data from three independent experiments. 5894

VOL. 72, 2004

CRUZIPAIN-MEDIATED HOST CELL INVASION BY T. CRUZI

5895

FIG. 2. Highly infective T. cruzi isolates depend on functional CP for host cell invasion. (A) Invasion of human smooth muscle cells by Dm28c, X10/6, or G trypomastigotes in DMEM containing 0.1% BSA for 3 h at 37C at a parasite/cell ratio of 3:1 in the presence of 1% DMSO and in the presence of 1% DMSO and 10 M N-Pip-F-hFVSPh. (B and C) Invasion of human smooth muscle cells by strain Dm28c (B) or by strain G (at a 10:1 parasite/host cell ratio) (C) in the presence of different concentrations of N-Pip-F-hF-VSPh. The data are representative of the data from three independent experiments.

smooth muscle cells were then chosen as a model for the subsequent studies because (i) smooth muscle is one of the tissues damaged during natural T. cruzi infections, (ii) there is little information pertaining to the invasion of human cells by this pathogen, and (iii) in contrast to established cell lines, primary cultured cells are more representative of the situation found in vivo. Addition of 10 M N-Pip-F-hF-VSPh during T. cruzi inter-

action with human smooth muscle cells (Fig. 2A) decreased invasion by the highly infective isolates Dm28c and X10/6 by 60%, indicating that functional CP are required for invasion of these cells. In contrast, this CP inhibitor had no effect on the residual infectivity of the G isolate (Fig. 2A). This observation was corroborated by dose dependence experiments, which showed that there was a 30% reduction in invasion by Dm28c at an inhibitor concentration of 100 nM (Fig. 2B). The penetration of isolate G parasites was not altered even at high concentrations of the compound, indicating that these trypomastigotes employ invasion mechanisms that do not depend on the activity of CP (Fig. 2C). The participation of kinin receptors was ruled out since the kinin receptor antagonists Hoe 140 and [Leu8]des-Arg9-BK, directed to the B2 and B1 receptor subtypes, respectively, had no effect on the invasion of smooth muscle cells by either Dm28c or G trypomastigotes (data not shown). Highly infective trypomastigotes secrete elevated levels of cruzipain. It has been shown previously that functional cruzipain, which is the primary target of N-Pip-F-hF-VSPh, is released into the extracellular medium by trypomastigotes (23). The presence of an active peptidase in the supernatants of Dm28c trypomastigotes was conrmed by enzymatic assays by using a synthetic peptidyl substrate, which revealed that there was continuous release of the enzyme for up to 1 h (Fig. 3A). The activity of the extracellular enzyme was then stable for at least 3 additional hours. The identication of this peptidase as a CP was conrmed by incubation with the specic CP inhibitor E-64. Quantication of active CP in supernatants of different isolates showed that Dm28c and X10/6, whose infectivities were affected by N-Pip-F-hF-VSPh, released ve to six times more functional protease than the G isolate released (Fig. 3B). The levels of intracellular CP activity were likewise distinct; they were about threefold higher in Dm28c and X10/6 trypomastigotes than in G trypomastigotes (Fig. 3C). Analysis of the intracellular cruzipain contents of these parasites by Western blotting revealed slightly reduced levels of the enzyme in G trypomastigotes compared with Dm28c trypomastigotes (Fig. 3D). In parallel, evaluation of the levels of a serine peptidase (OPB) associated with host cell invasion by Western blotting revealed no signicant difference among the three isolates (Fig. 3E). Specic antibodies were then used to conrm that cruzipain was the active CP released by Dm28c trypomastigotes (Fig. 4). These antibodies reacted strongly with puried cruzipain (Fig. 4A, lane 3) and recognized two major bands in lysates from epimastigotes (Fig. 4A, lane 2) and from trypomastigotes (Fig. 4A, lane 1). The difference in the apparent molecular masses of these proteins may have been due to variable posttranslational modications encountered in cruzipain isoforms (6, 15). Dm28c supernatants were then subjected to immunoprecipitation by using anti-cruzipain IgG or control IgG, and the peptidase activity bound to the resin was monitored by measuring the hydrolysis of a synthetic substrate (Fig. 4B). E-64 completely inhibited substrate hydrolysis, conrming that this process was mediated by CP (data not shown). The CP activity present in supernatants was immunoprecipitated by anti-cruzipain IgG but not by control IgG (Fig. 4B). Incubation of supernatants with biotin-coupled N-Pip-F-hF-VSPh, followed by immunoprecipitation with anti-cruzipain IgG, provided fur-

5896

APARICIO ET AL.

INFECT. IMMUN.

FIG. 3. Highly infective T. cruzi isolates contain higher levels of functional CP. (A) Time course of CP release by trypomastigotes. Dm28c trypomastigotes were incubated in HBSS at 37C. Aliquots (200 l) were collected at different times and processed as described in Materials and Methods, and the peptidase activity was detected by using 10 M -NH2-(Cap)Leu-(SBzl)Cys-AMC as a substrate. (B) Quantication of CP released by Dm28c, X10/6, or G trypomastigotes. Supernatants were obtained after incubation of parasites in HBSS at 37C for 2 h. The peptidase activity was determined as described above for panel A. (C) Quantication of intracellular CP in Dm28c, X10/6, or G trypomastigotes. Trypomastigote lysates (30 g) were used in enzymatic assays as described above for panel A. (D) Cruzipain accumulation in trypomastigotes: Western blot of trypomastigote lysates (50 g) obtained by using anti-cruzipain antiserum at a 1:1,000 dilution. The numbers below the blot indicate the values obtained by densitometry of the bands. Cruzipain puried from epimastigotes was included as a control. (E) Quantication of OPB in trypomastigotes: Western blot of trypomastigote lysates (50 g) obtained by using anti-OPB antiserum at a 1:1,000 dilution. The densitometry values are indicated below the blot. MW, molecular mass.

VOL. 72, 2004

CRUZIPAIN-MEDIATED HOST CELL INVASION BY T. CRUZI

5897

FIG. 4. Cruzipain is the CP released by trypomastigotes. (A) Western blot of T. cruzi Dm28c lysates obtained by using anti-cruzipain antiserum. Lane 1, trypomastigotes; lane 2, epimastigotes; lane 3, puried cruzipain. (B) Supernatants were subjected to immunoprecipitation with anti-cruzipain or control IgG, and the peptidase activity bound to the beads was determined by using 10 M -NH2-(Cap)Leu(SBzl)Cys-AMC as a substrate. (C) Supernatants (lane 1) or puried cruzipain (lane 2) were supplemented with 5 mM DTT and incubated with 10 M biotinN-Pip-F-hF-VSPh at 37C for 1 h. The samples were subsequently subjected to immunoprecipitation with anti-cruzipain IgG and visualized by Western blotting after incubation of the membrane with alkaline phosphatase-streptavidin conjugates. ctrl, control; MW, molecular mass.

FIG. 5. CP released by trypomastigotes mediates host cell invasion. (A) Invasion of human smooth muscle cells by G trypomastigotes at a parasite/cell ratio of 4:1 in DMEM containing 0.1% BSA for 3 h at 37C. The invasion assays were performed in the presence or absence of Dm28c trypomastigote supernatants supplemented with 2.5 mM DTT or with the reducing agent and 30 M E-64, as indicated. The controls received HBSS containing 2.5 mM DTT. (B) Invasion of human smooth muscle cells by G trypomastigotes at parasite/host cell ratios 4:1, 10:1, and 20:1. The data are representative of the data from three independent experiments.

ther evidence that both cruzipain species released by trypomastigotes were capable of binding to this inhibitor (Fig. 4C, lane 1). Cruzipain released by trypomastigotes mediates host cell invasion. The results described above indicate that CP participate in an invasion mechanism that is predominantly used by Dm28c and X10/6 but not by the G isolate. The limited amount of cruzipain liberated by the latter organism may be insufcient for engagement of the CP-dependent pathway. This possibility was evaluated by performing assays for smooth muscle cell invasion by G trypomastigotes in the presence of cruzipain-rich supernatants from Dm28c trypomastigotes (Fig. 5). The Dm28c supernatants caused a twofold increase in the invasion of G parasites through a CP-dependent mechanism, since the

effect was completely abolished by E-64, an irreversible CP inhibitor (Fig. 5A). Although E-64 is a broad-range CP inhibitor, due to its hydrophilic nature its action should be restricted to the extracellular environment, in contrast to the activity of the hydrophobic compound N-Pip-F-hF-VSPh. The effect of E-64 was observed only in the presence of supernatants, ruling out participation of host cell papain-like CP in the basal invasion mechanism used by G trypomastigotes. The Dm28c supernatants consistently increased the infectivity of G trypomastigotes twofold, even when different parasite/host cell ratios were used (Fig. 5B). To demonstrate that cruzipain was in fact the CP responsible for this potentiation of infectivity, the invasion assays were performed by using Dm28c supernatants from which cruzipain was depleted by immunoprecipitation (Fig. 6). Addition of untreated supernatants or of supernatants subjected to immunoprecipitation with rabbit control IgG enhanced the invasion of the G isolate about twofold compared

5898

APARICIO ET AL.

INFECT. IMMUN.

FIG. 6. Extracellular cruzipain is required for host cell invasion: smooth muscle cell invasion by G trypomastigotes at a parasite/cell ratio of 4:1 in DMEM containing 0.1% BSA for 3 h at 37C. Assays were performed in the presence (striped bar) or in the absence (open bar) of supernatant from Dm28c trypomastigotes or in the presence of supernatant subjected to immunoprecipitation with anti-cruzipain IgG (solid bars) or with control IgG (gray bar) as described in Materials and Methods. In some experiments, puried cruzipain was supplemented with 2.5 mM DTT, incubated for 5 min at 37C, and added to the monolayers at a concentration of 5 nM. E-64 was added at a concentration of 30 M. The data are representative of the data from two independent experiments. FIG. 7. Cruzipain-mediated invasion requires a trypomastigote-associated factor. (A) Host cell invasion by G trypomastigotes at a parasite/cell ratio of 20:1 in DMEM containing 0.1% BSA for 3 h at 37C in the presence or in the absence of supernatants from Dm28c or G trypomastigotes (Tryp), Dm28c epimastigotes (Epi), obtained in the same way, or in the presence of puried cruzipain at a nal concentration of 5 nM (solid bar). The enzyme was supplemented with 2.5 mM DTT and incubated for 5 min at 37C for activation prior to addition to the assay mixture. (B) Assays performed as described above for panel A in the presence or in the absence (open bar) of supernatants from Dm28c trypomastigotes (solid bars) or supernatants from G trypomastigotes (gray bars). The processed samples consisted of Dm28c supernatants (solid bar, processed) incubated with 2.5 mM DTT for 3 h at 37C and subsequently supplemented with 30 M E-64 prior to addition to the assay mixture or supernatants from G trypomastigotes (gray bar, cruzipain) supplemented with 5 nM puried cruzipain and 2.5 mM DTT.

to the basal invasion levels (Fig. 6). In contrast, addition of the cruzipain-depleted supernatants resulted in invasion levels similar to those observed in controls (Fig. 6). Addition of 5 nM puried cruzipain to the depleted supernatants fully restored the ability to potentiate invasion, which was again blocked by E-64. These results demonstrate that cruzipain present in Dm28c supernatants is indeed the CP responsible for engagement of an invasion route, which is decient in trypomastigotes of the G isolate. Recently, it was reported that the prolyl oligopeptidase Tc80 (POP-Tc80), secreted by trypomastigotes, participates in the invasion of rodent muscle cells by T. cruzi (12). The up-regulation of the invasion of human muscle cells by G trypomastigotes caused by Dm28c supernatants was not affected by N-benzyloxycarbonyl-prolinal-prolinal-dimethylacetal (Z-ProPro-DMA), an inhibitor of POP-Tc80, while it was totally abolished by addition of N-Pip-F-hF-VSPh (data not shown). This result argues against the involvement of POP-Tc80 in this cruzipain-dependent invasion pathway. It is worth mentioning that although Z-Pro-Pro-DMA binds to cruzipain (Ki 31 M), the concentration used in the assay (10 M) was threefold less than the Ki for cruzipain and 10-fold greater than the Ki for POP-Tc80, which favored binding to the latter enzyme. Cruzipain acts on a trypomastigote molecule. The role of cruzipain in mediating host cell invasion was further addressed in assays in which supernatants from G trypomastigotes or from Dm28c epimastigotes, which are extremely rich in cruzipain, were used as sources of functional enzyme (Fig. 7A). As

expected, the addition of G-derived supernatants, which contained little cruzipain, did not affect the invasion indices for the parasites. Remarkably, Dm28c epimastigote supernatants behaved in a similar fashion, indicating that the presence of active cruzipain is required but not sufcient for engagement of this invasion pathway. This observation was further conrmed by addition of puried cruzipain to the invasion assay mixtures, which did not potentiate the infectivity of G trypomastigotes (Fig. 7A). As shown above (Fig. 6), puried cruzipain restored the ability of cruzipain-depleted supernatants to up-regulate invasion by G parasites. These observations strongly suggest that extracellular cruzipain is required to process a substrate(s) which is present in the supernatants of trypomastigotes but is absent in the supernatants of epimastigotes.

VOL. 72, 2004

CRUZIPAIN-MEDIATED HOST CELL INVASION BY T. CRUZI

5899

To evaluate if cruzipain in fact processes a trypomastigotederived molecule and releases a putative invasion factor, Dm28c supernatants were incubated at 37C in the presence of reducing agents, which enabled cruzipain to act on its substrate(s). The reducing environment was necessary to maintain the active-site cysteine in its reduced state, thus ensuring catalytic activity. The enzyme was subsequently inactivated by E-64 prior to addition of the processed supernatants to the invasion assay mixtures (Fig. 7B). Interestingly, Dm28c supernatants that were preprocessed as described above and were devoid of functional cruzipain increased the infectivity of G parasites to the same levels as those observed with the unprocessed supernatants containing active enzyme, when they were compared to the controls to which no supernatant was added (Fig. 7B). Supernatants that were preprocessed in the presence of E-64 could not potentiate G infectivity (data not shown). These results indicate that the putative invasion factor generated by proteolysis is the nal product required to mediate invasion by T. cruzi. To test whether cruzipain-dependent invasion was compromised in G parasites solely due to defective secretion of the enzyme, G supernatants were supplemented with exogenous cruzipain prior to addition to assay mixtures. In contrast to Dm28c supernatants, neither G supernatants alone nor the sample supplemented with puried cruzipain was able to potentiate the infectivity of these parasites (Fig. 7B). This result indicates that supernatants derived from G trypomastigotes contain low levels not only of cruzipain but also of its putative substrate(s) and/or of cofactors required for engagement of the invasion pathway. It is well documented that T. cruzi continuously sheds part of its plasma membrane into the extracellular medium in the form of small vesicles, which contain several cell surface molecules (11). Therefore, the supernatants used in this study should have consisted of a mixture of soluble and vesicleassociated components. The distribution of extracellular cruzipain and its substrate(s) (and/or cofactors) in soluble and membrane fractions was checked after centrifugation of supernatants at 100,000 g. After fractionation, the peptidase activity was detected mainly in the soluble fraction, indicating that the majority of the functional extracellular cruzipain was not associated with shed vesicles (Fig. 8A, inset). The distribution of the trypomastigote-associated cruzipain substrate(s) in soluble and membrane fractions was subsequently checked by invasion assays. While unprocessed Dm28c supernatants enhanced the basal level of invasion of G parasites twofold, the isolated soluble or membrane fractions had no effect on G infectivity (Fig. 8A). Since cruzipain was detected in the soluble fraction, this result suggests that the enzymes putative substrate(s) is membrane associated. This assumption was further conrmed by preprocessing of supernatants, followed by enzyme inactivation and fractionation by centrifugation at 100,000 g. The soluble fraction, but not the membrane fraction (Fig. 8A), of preprocessed supernatants increased the infectivity of G trypomastigotes about twofold. Identication of cruzipain as the enzyme capable of releasing the invasion factor from Dm28c trypomastigote membranes was further addressed by incubation of 100,000- -g fractions with puried cruzipain, followed by enzyme inactivation prior to addition to invasion assay mixtures (Fig. 8B). In agreement with this, the level of invasion of host cells by G trypomastigotes was in-

FIG. 8. Cruzipain-mediated invasion requires trypomastigote membranes. (A) Host cell invasion by G trypomastigotes at a parasite/ cell ratio of 20:1 in DMEM containing 0.1% BSA for 3 h at 37C in the presence or in the absence (open bar) of supernatants from Dm28c trypomastigotes. Untreated supernatants were added to the invasion assay mixture in full (solid bar) or after fractionation by centrifugation at 100,000 g (gray bars). S, soluble fraction; M, membrane fraction. Processed supernatants (striped bars) were incubated with 2.5 mM DTT for 3 h at 37C and subsequently supplemented with 30 M E-64, followed by fractionation by centrifugation at 100,000 g prior to addition to the invasion assay mixture. S, soluble fraction; M, membrane fraction. (Inset) Peptidase activity of unprocessed supernatants (100 l) from Dm28c trypomastigotes after fractionation by centrifugation at 100,000 g, as determined with 10 M carbobenzoxy-phenylalanyl-arginyl-7-amido-4-methylcoumarin. The graph shows the initial velocity. S, soluble fraction; M, membrane fraction. The data are representative of the data from two independent experiments. (B) Host cell invasion by G trypomastigotes at a parasite/cell ratio of 4:1 in DMEM containing 0.1% BSA for 3 h at 37C. Supernatants from Dm28c trypomastigotes were fractionated by centrifugation at 100,000 g, and the membrane fraction was incubated with 2.5 mM DTT and 5 nM puried cruzipain for 3 h at 37C and subsequently supplemented with 30 M E-64 before addition to the assay mixtures. Open bar, saline; solid bar, untreated fraction; gray bar, membrane fraction treated with cruzipain; striped bar, membrane fraction treated with cruzipain which was preinactivated by 30 M E-64.

creased in the presence of membrane fractions which were pretreated with puried cruzipain compared to the basal invasion levels. Untreated membrane fractions or the fractions treated with cruzipain inactivated by E-64 were not capable of enhancing invasion by the G strain. Taken together, these data indicate that the liberation of a factor that mediates host cell invasion by cruzipain requires trypomastigote membranes. Finally, since the engagement of kinin receptors by cruzipain triggers an increase in the free intracellular Ca2 concentration in host cells, we next veried that a similar signaling event was implicated in the cruzipain-dependent invasion route de-

5900

APARICIO ET AL.

INFECT. IMMUN.

FIG. 9. Increase in infectivity induced by Dm28c trypomastigote supernatants is inhibited by thapsigargin: host cell invasion by G trypomastigotes at a parasite/cell ratio of 50:1 in DMEM containing 0.1% BSA for 1 h at 37C in the presence or in the absence of supernatants from Dm28c trypomastigotes. Where indicated, the host cells were pretreated with 0.5 M thapsigargin in DMEM containing 10% fetal calf serum for 40 min at 37C. The coverslips were washed three times with HBSS before addition of trypomastigotes. The low level of infectivity resulted from the short time interval used in the assay (1 h). The data are representative of the data from three independent experiments.

scribed here (23). The levels of invasion of G parasites were not affected by prior treatment of host cells with thapsigargin, indicating that these trypomastigotes do not depend on the mobilization of intracellular Ca2 stores for invasion (Fig. 9). In contrast, the potentiation of invasion observed in the presence of Dm28 supernatants was completely abolished by host cell pretreatment with thapsigargin, suggesting that the supernatant-mediated invasion route requires mobilization of host cell intracellular Ca2 stores. Taken together, these results associate cruzipain with host cell invasion by an alternative mechanism. DISCUSSION In vitro cellular systems employed in studies of T. cruzi interactions with mammalian cells have indicated that intracellular Ca2 -mediated activation responses transduced by GPCRs can enhance parasite invasion of nonphagocytic cells (reviewed in reference 2). Previous efforts to characterize the mechanisms leading to GPCR activation focused on OPB (1, 3). More recently, it has been demonstrated that T. cruzi trypomastigotes rely on kinin liberation by cruzipain to efciently invade host cells that express high levels of kinin receptors (17, 23, 27). Here, we report a role for cruzipain in the invasion of human smooth muscle cells through a kinin-independent route. Analysis of the in vitro infectivity of three T. cruzi isolates by using several host cell types allowed discrimination of highly infective trypomastigotes and trypomastigotes with low levels of infectivity. The use of a synthetic irreversible CP inhibitor with partial selectivity for cruzipain, N-Pip-F-hF-VSPh, demonstrated that highly infective isolates Dm28c and X10/6, but not the G isolate with a low level of infectivity, are very dependent

on the activity of CP for invasion. Quantication of the functional CP secreted by trypomastigotes of the three isolates revealed a direct correlation between the infective potential and the level of CP present in the supernatant. The CP activities present in parasite lysates were also lower in the G isolate with a lower level of infectivity than in Dm28c or X10/6. The processes leading to reduced levels of extracellular CP activity in supernatants of the G isolate were not investigated in detail. The Western blot of trypomastigote lysates revealed that the intracellular level of cruzipain in the G isolate was slightly less than the level in Dm28c, although the activity found in the supernatants was vefold lower. Since the control of cruzipain expression for the different stages of T. cruzi appears to be primarily posttranscriptional (28), additional mechanisms, such as the rate of zymogen processing combined with inefcient trafcking and secretion of the mature enzyme, could contribute to the decreases in the levels of extracellular cruzipain. Diminished levels of active cruzipain could also result from strain-dependent differences in the expression of chagasin, a recently described endogenous CP inhibitor of T. cruzi (21). It is possible that G trypomastigotes regulate cruzipain expression also at the transcriptional level. These possibilities are currently being addressed. Characterization of secreted cruzipain by using a biotinylated inhibitor (biotinN-Pip-F-hF-VSPh) revealed that trypomastigotes secrete at least two heterogeneous cruzipain species. It was demonstrated previously that alternative cruzipain isoforms may be expressed by the mammalian stages of T. cruzi and that at least one of them, cruzipain 2, differs from cruzipain 1/cruzain with respect to substrate specicity and afnity for inhibitors (15, 16). It is possible that multiple secreted cruzipain isoforms act synergistically to process a common precursor, or alternatively, they may act on different natural substrates. Irrespective of the nature of cruzipain isoforms, our data demonstrate that the activity of extracellular cruzipain(s) is required for generation of a putative invasion factor. In agreement with this, we showed that the cruzipain-rich supernatants derived from Dm28 trypomastigotes increased invasion by the G isolate twofold. The increase due to Dm28c supernatants was abolished by immunodepletion of the enzyme from supernatants or by inhibition with E-64. Interestingly, the supernatants of G trypomastigotes could not potentiate invasion. These parasites secrete signicantly less cruzipain than Dm28c secretes, but the addition of exogenous cruzipain did not complement this phenotype, suggesting that the cells may not produce sufcient amounts of the invasion factor precursor. Besides the decit of functional extracellular cruzipain, it is possible that additional molecules involved in invasion are also down-regulated in G trypomastigotes. The levels of OPB are apparently not altered in the G isolate compared to strain Dm28c or X10/6, while the levels of the adhesion molecule Tc85 are slightly higher in the former organism (data not shown). In agreement with this, experiments to quantify the adhesion capability of these parasites in prexed smooth muscle cells demonstrated that they adhere at even higher rates than Dm28c (data not shown). Generation of the presumed invasion factor was not affected by addition of the Z-Pro-Pro-DMA inhibitor, indicating that the serine protease POP-Tc80 (12) and cruzipain act on different invasion routes. However, we cannot rule out the possibility that cruzipain and OPB may act synergistically. It was

VOL. 72, 2004

CRUZIPAIN-MEDIATED HOST CELL INVASION BY T. CRUZI

5901

recently demonstrated that recombinant OPB is capable of trimming small peptides (up to 17 residues), including kininogen-based substrates (14). It would be interesting to know if both cruzipain and OPB are required to produce a common invasion factor, where OPB would act downstream from cruzipain. However, the fact that OPB is found in the cytosol argues against this hypothesis (1). We found that the invasion route enhanced by addition of Dm28c trypomastigote supernatants was sensitive to host cell treatment with thapsigargin. Interestingly, tissue culture trypomastigotes from the G isolate were unable to activate this pathway, although it has been demonstrated that G metacyclic trypomastigotes mobilize calcium from thapsigargin-sensitive stores of HeLa cells (8). However, the molecules involved in this pathway (i.e., gp85 and gp35/50) are not present in tissue culture or blood-derived trypomastigotes (8). In this study we demonstrated that the precursor molecule processed by cruzipain is associated with trypomastigotes but not with epimastigotes. This conclusion is based on the following facts: (i) epimastigote supernatants (rich in cruzipain) had no effect on host cell invasion by G trypomastigotes, (ii) the ability of cruzipain-depleted supernatants to increase the infectivity of G parasites was fully restored by addition of cruzipain puried from epimastigotes, and (iii) puried cruzipain alone was not sufcient to promote invasion by the G isolate. We obtained further evidence that the release of the invasion factor by cruzipain requires trypomastigote membranes, since (i) after fractionation, the soluble fraction of supernatants alone (in which active cruzipain was detected) was unable to promote engagement of the CP-dependent pathway, (ii) in supernatants that were preprocessed prior to fractionation, the presumed invasion factor was found in the soluble fraction, and (iii) the presumed invasion factor was present after treatment of a 100,000- -g fraction of supernatants with puried cruzipain. These results are in agreement with previous observations that membranes from trypomastigotes, but not membranes from epimastigotes, induce calcium signaling in the host cell (26). Experiments performed with kinin receptor antagonists (selective for B2 or B1) have ruled out the possibility that the kinin signaling pathway is involved in the invasion mechanism described here. In contrast to the kinin invasion route, in which cruzipain processes a plasma-derived precursor substrate, kininogen (17, 23), cruzipain acts on a trypomastigoteassociated molecule(s) present in shed membranes. A feature that distinguishes the short-lived kinin peptides from the invasion factor is the stability of the latter, which may act on host cells even in the absence of trypomastigotes. Considering the great heterogeneity displayed by T. cruzi strains (reviewed in reference 20), our ndings may be relevant in the context of natural infections, in which mammalian hosts may be simultaneously exposed to different parasite strains. In this scenario, cruzipain secreted by highly infective trypomastigotes could assist poorly infective strains, such as the G strain, to establish an infection in a host. We also observed potentiation of G infectivity by Dm28c supernatants using other cell types, such as Vero and LLCMK2 (data not shown), suggesting that the cruzipain-mediated invasion mechanism is not restricted to smooth muscle. In summary, we show that functional cruzipain secreted by trypomastigotes participates directly in host cell invasion by

releasing a soluble invasion factor from trypomastigote membranes. It remains to be determined if this route is engaged in vivo and the extent to which it may inuence parasite virulence and/or tissue tropism. Identication of the trypomastigote-associated substrate of cruzipain is being currently pursued.
ACKNOWLEDGMENTS We are grateful to Luiz Juliano (UNIFESP, Sao Paulo, Brazil) for synthesis of -NH2-(Cap)L-(SBzl)C-AMC and for supplying Z-ProPro-DMA, to Jim Palmer (Axys Pharmaceuticals) for donation of N-Pip-F-hF-VSPh, and to Norma W. Andrews (Yale University, New Haven, Conn.) for donation of anti-OPB antibodies. We thank Leila Faustino, Edna Lopes, and Alda Fidelis for technical assistance. This work was supported by CNPq/Universal, FAPERJ, and PRONEX.
References 1. Burleigh, B. A., E. V. Caler, P. Webster, and N. W. Andrews. 1997. A cytosolic serine endopeptidase from Trypanosoma cruzi is required for the generation of Ca2 signaling in mammalian cells. J. Biol. Chem. 136:609 620. 2. Burleigh, B. A., and A. M. Woolsey. 2002. Cell signaling and Trypanosoma cruzi invasion. Cell. Microbiol. 4:701711. 3. Caler, E. V., S. Vaena de Avalos, P. A. Haynes, N. W. Andrews, and B. A. Burleigh. 1998. Oligopeptidase B-dependent signaling mediates host cell invasion by Trypanosoma cruzi. EMBO J. 17:49754986. 4. Caler, E. V., S. Chakrabarti, K. T. Fowler, S. Rao, and N. W. Andrews. 2001. The exocytosis-regulatory protein synaptotagmin VII mediates cell invasion by Trypanosoma cruzi. J. Exp. Med. 193:10971104. 5. Campetella, O., J. Henriksson, L. Aslund, A. C. C. Frasch, U. Pettterson, and J. J. Cazzulo. 1992. The major cysteine proteinase (cruzipain) from Trypanosoma cruzi is encoded by multiple polymorphic tandemly organized genes located on different chromosomes. Mol. Biochem. Parasitol. 50:225 234. 6. Cazzulo, J. J., J. Martinez, and A. J. Parodi. 1992. On the posttranslational modications at the C-terminal domain of the major cysteine proteinase (cruzipain) from Trypanosoma cruzi. FEMS Microbiol. Lett. 79:411416. 7. Del Nery, E., M. A. Juliano, M. Meldal, I. Svendsen, J. Scharfstein, A. Walmsley, and L. Juliano. 1997. Characterization of the substrate specicity of the major cysteine protease (cruzipain) from Trypanosoma cruzi using a portion-mixing combinatorial library and uorogenic peptides. Biochem. J. 323:427433. 8. Dorta, M. L., A. T. Ferreira, M. E. M. Oshiro, and N. Yoshida. 1995. Ca2 signal induced by Trypanosoma cruzi metacyclic trypomastigote surface molecules implicated in mammalian cell invasion. Mol. Biochem. Parasitol. 73:285289. 9. Eakin, A. E., R. R Mills, G. Harth, J. H McKerrow, and C. S. Craik. 1992. The sequence, genomic organization and expression of the major cysteine proteinase (cruzain) from Trypanosoma cruzi. J. Biol. Chem. 267:74117420. 10. Engel, J. C., P. S. Doyle, I. Hsieh, and J. H. McKerrow. 1998. Cysteine protease inhibitors cure an experimental Trypanosoma cruzi infection. J. Exp. Med. 188:725734. 11. Goncalves, M. F., E. S. Umezawa, A. M. Katzin, W. de Souza, M. J. Alves, B. Zingales, and W. Colli. 1991. Trypanosoma cruzi: shedding of surface antigens as membrane vesicles. Exp. Parasitol. 72:4353. 12. Grellier, P., S. Vendeville, R. Joyeau, I. M. D. Bastos, H. Drobecq, F. Frappier, A. R. L. Teixeira, J. Schevel, E. Davioud-Charvet, C. Sergheraert, and J. M Santana. 2001. Trypanosoma cruzi prolyl-oligopeptidase Tc80 is involved in nonphagocytic mammalian cell invasion by trypomastigotes. J. Biol. Chem. 276:4707847086. 13. Harth, G., N. W. Andrews, A. A. Mills, J. C. Engel, R. Smith, and J. H. McKerrow. 1993. Peptide-uoromethyl ketones arrest intracellular replication and intercellular transmission of Trypanosoma cruzi. Mol. Biochem. Parasitol. 58:1724. 14. Hemerly, J. P., V. Oliveira, E. Del Nery, R. E. Morty, N. W. Andrews, M. A. Juliano, and L. Juliano. 2003. Sub-site specicity (S3,S2,S1 ,S2 and S3 ) of oligopeptidase B from Trypanosoma cruzi and Trypanosoma brucei using uorescent-quenched peptides: a comparative study and identication of specic carboxy-peptidase activity. Biochem. J. 373:933939. 15. Lima, A. P. C. A., D. C. Tessier, D. Y. Thomas, J. Scharfstein, and T. Vernet. 1994. Identication of new cysteine protease gene isoforms in Trypanosoma cruzi. Mol. Biochem. Parasitol. 67:333338. 16. Lima, A. P. C. A., F. C. G. Reis, C. Serveau, G. Lalmanach, L. Juliano, R. Menard, T. Vernet, D. Y. Thomas, A. C. Storer, and J. Scharfstein. 2001. Cysteine protease isoforms from Trypanosoma cruzi, cruzipain 2 and cruzain, present different substrate preference and susceptibility to inhibitors. Mol. Biochem. Parasitol. 114:4152.

5902

APARICIO ET AL.

INFECT. IMMUN.
Trypanosoma cruzi is potentiated by activation of bradykinin B2-receptors. J. Exp. Med. 192:12891299. Scharsftein, J. 2003. Activation of bradykinin-receptors by Trypanosoma cruzi: a role for cruzipain in microvascular pathology, p. 111137. In J. M. Kelly (ed.), Molecular pathogenesis of Chagas disease. Landes Bioscience, Austin, Tex. Tardieux, I., P. Webster, J. Ravesloot, W. Boron, J. A. Lunan, J. E. Heuser, and N. W. Andrews. 1992. Lysosome recruitment and fusion are early events required for trypanosome invasion of mammalian cells. Cell 73:11171130. Tardieux, I., M. H. Nathanson, and N. W. Andrews. 1994. Role in host cell invasion of Trypanosoma cruzi-induced cytosolic free Ca2 transients. J. Exp. Med. 179:10171022. Todorov, A. G., D. Andrade, J. B. Pesquero, R. de C. Araujo, M. Bader, J. Stewart, L. Gera, W. Muller-Esterl, V. Morandi, R. C. Goldenberg, H. C. Neto, and J. Scharfstein. 2003. Trypanosoma cruzi induces edematogenic responses in mice and invades cardiomyocytes and endothelial cells in vitro by activating distinct kinin receptor (B1/B2) subtypes. FASEB J. 17:7375. Tomas, A. M., and J. M. Kelly. 1996. Stage-regulated expression of cruzipain, the major cysteine protease of Trypanosoma cruzi, is independent of the levels of RNA1. Mol. Biochem. Parasitol. 76:91103. Tyler, K., and M. A. Miles. 2003. American trypanosomiasis, vol. 7, p. 8196. Kluwer Academic Publishers, London, United Kingdom. Yoshida, N. 1983. Surface antigens of metacyclic trypomastigotes of Trypanosoma cruzi. Infect. Immun. 40:836839.

17. Lima, A. P. C. A., P. C. Almeida, I. L. S. Tersariol, V. Schmitz, A. H. Schmaier, L. Juliano, I. Y. Hirata, W. Muller-Esterl, J. R. Chagas, and J. Scharfstein. 2002. Heparan sulfate modulates kinin release by Trypanosoma cruzi through the activity of cruzipain. J. Biol. Chem. 277:58755881. 18. McGrath, M. E., A. E. Eakin, J. C. Engel, J. H. McKerrow, C. S. Craik, and R. A Fletterick. 1995. The crystal structure of cruzain: a therapeutic target for Chagas disease. J. Mol. Biol. 247:251259. 19. Meirelles, M. N., L. Juliano, E. Carmona, S. G. Silva, E. M. Costa, A. C. Murta, and J. Scharfstein. 1992. Inhibitors of the major cysteinyl proteinase (GP57/51) impair host cell invasion and arrest the intracellular development of Trypanosoma cruzi in vitro. Mol. Biochem. Parasitol. 52:175184. 20. Miles, M. A., M. Yeo, and M. Gaunt. 2003. Genetic diversity of Trypanosoma cruzi and the epidemiology of Chagas disease, p. 115. In J. M. Kelly (ed.), Molecular pathogenesis of Chagas disease. Landes Bioscience, Austin, Tex. 21. Monteiro, A. C. S., M. Abrahamson, A. P. C. A. Lima, M. A. Vannier-Santos, and J. Scharfstein. 2001. Identication, characterization and localization of chagasin, a tight-binding cysteine protease inhibitor in Trypanosoma cruzi. J. Cell Sci. 114:39333942. 22. Murta, A. C., P. M. Persechini, T. Souto-Padron, W. de Souza, J. A. Guimaraes, and J. Scharfstein. 1990. Structural and functional identication of GP57/51 antigen of Trypanosoma cruzi as a cysteine proteinase. Mol. Biochem. Parasitol. 43:2738. 23. Scharfstein, J., V. Schmitz, V. Morandi, M. A. Capella, A. P. C. A. Lima, A. Morrot, L. Juliano, and W. Muller-Esterl. 2000. Host cell invasion by

24.

25. 26. 27.

28. 29. 30.

Editor: J. F. Urban, Jr.