Identification of Trichomonas vaginalis Cysteine Proteases That Induce Apoptosis in Human Vaginal Epithelial Cells*

A secreted cysteine protease (CP) fraction from Trichomonas vaginalis is shown here to induce apoptosis in human vaginal epithelial cells (HVEC) and is analyzed by mass spectrometry. The trichomonad parasite T. vaginalis causes one of the most common non-viral sexually transmitted infection in humans, trichomoniasis. The parasite as well as a secreted cysteine protease (CP) fraction, isolated by affinity chromatography followed by Bio-Gel P-60 column chromatography, are shown to induce HVEC apoptosis, as demonstrated by the Cell Death Detection ELISAPLUS assay and annexin V-fluorescein isothiocyanate flow cytometry analyses. Initiation of apoptosis is correlated with protease activity because the specific CP inhibitor E-64 inhibits both activities. SDS-PAGE analysis of the CP fraction reveals triplet bands around 30 kDa, and matrix-assisted laser desorption ionization time-of-flight MS indicates two closely associated peaks of molecular mass 23.6 and 23.8 kDa. Mass spectral peptide sequencing of the proteolytically digested CPs results in matches to previously reported cDNA clones, CP2, CP3, and CP4 (Mallinson, D. J., Lockwood, B. C., Coombs, G. H., and North, M. J. (1994) Microbiology 140, 2725-2735), as well as another sequence with high homology to CP4 (www.tigr.org). These last two species are the most abundant components of the CP fraction. The present results, suggesting that CP-induced programmed cell death may be involved in the pathogenesis of T. vaginalis infection in vivo, may have important implications for therapeutic intervention.

McLaughlin and Mü ller (19) reported the first purification of a CP from a trichomonad parasite, Trichomonas foetus, in 1979, but it was not until the studies of Coombs, North, and their colleagues (8, 20 -23) that the field began to flourish. They showed the presence of several CPs in the extracellular medium and demonstrated that T. vaginalis secretes CPs (7,21). Their work culminated in the cloning of four T. vaginalis CPs (7) using sequences based on homologies within the broad family of CPs (GenBank TM accession numbers: X77218, X77219, X77220 and X77221). Two of the clones appear to be full-length. The other two are partial, although near fulllength. Subsequently, additional CPs were cloned by Garber et al. (24) (accession number X70823) and Leon-Sicairos et al. (25, 26; accession numbers AY371180, AY371181, and AY463679).
Several years ago, we established human (H) and bovine (B) vaginal epithelial cell (VEC) culture systems that display hostparasite specificity (27,28). We showed that trichomonad infection of VECs in vitro results in cell detachment followed by cell destruction (27,28). We now show that host cell destruction is the result of apoptosis, which is induced by CPs secreted by T. vaginalis. Apoptosis is an important and well regulated form of cell death that occurs under a variety of physiological and pathological conditions (29). It has been studied in detail as a response to several bacterial and viral infections (30) and it is apparent that viral, bacterial, and protozoan pathogens have evolved a variety of different strategies to modulate host cell apoptosis (31,32). The studies presented here are the first to demonstrate that extracellular T. vaginalis CPs induce HVEC apoptosis. We have identified these CPs using mass spectrometric analysis and comparison to the GenBank and TIGR data bases. Preliminary sequence data were obtained from The Institute for Genomic Research through the website at www.tigr.org.

EXPERIMENTAL PROCEDURES
Trichomonad Parasites-T. vaginalis (UR1) isolates were obtained from a symptomatic patient attending an STD clinic. and S10-RR015942 (to C. E. C.). Sequencing of Trichomonas vaginalis was accomplished with support from the National Institutes of Health NIAID. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
T. vaginalis Soluble Fraction (SF)-PBS-washed trichomonads were resuspended in trichomonad incubation buffer (PBS with 10 mM HEPES, and 0.05% L-ascorbic acid, pH 6.2) and incubated at 37°C for 2.5 h, as reported earlier (33,34). Preparations were used only when Ͼ95% parasites were motile after incubation. The suspension was centrifuged at 12,000 ϫ g for 15 min and the supernatant was filtered sequentially through 0.45-and 0.22-m filters followed by additional centrifugation for 2 h at 50,000 ϫ g. The soluble fraction (devoid of membrane debris) was concentrated using centrifugal concentrators (Centriplus 10, 10K MWCO, or Jumbosep 10K, PALL).
Isolation of CPs from SF-SF was applied to a bacitracin affinity chromatography column according to methods reported earlier (21,33), except Affi-Gel 10 (Bio-Rad) was used as the support matrix. Concentrated SF was diluted (1:3) with sodium acetate buffer (20 mM, pH 4.0) and applied to the column (equilibrated in sodium acetate buffer) at a flow rate of 0.3 ml/min. The column was washed (0.8 ml/min) with 20 mM sodium acetate buffer until A 280 read zero. Material bound to the column was eluted with 0.1 M Tris-HCl (pH 7.0), 1.0 M NaCl, 25% 2-propanol as described by Thomford et al. (34). The eluted fractions containing protein (A 280 ) were pooled and dialyzed for 2 h at 4°C against water and concentrated using centrifugal filtration devices. The concentrated bacitracin-bound fraction was applied to a Bio-Gel P-60 column (1 ϫ 60 cm) equilibrated with 0.1 M ammonium acetate (pH 6.0), and fractions were collected. Two protein peaks (molecular mass ϳ30 and 60 kDa) were detected and the appropriate fractions were combined and concentrated using centrifugal devices. Protein concentrations were determined with Bradford reagent. Preparations typically yielded 100 -200 g of the 60-kDa fraction and ϳ600 g of the 30-kDa fraction (CP30) from 4 liters of parasite culture (ϳ2.2 ϫ 10 10 parasites). The two fractions were subjected to SDS-PAGE analysis on 12% polyacrylamide gels, according to Laemmli (35), using a 3% stacking gel, and stained with Coomassie Blue.
Mass Spectrometry-The masses of the intact proteins in the CP fraction were determined by MALDI-TOF MS on a Bruker Reflex IV instrument, with irradiation from a nitrogen laser (337 nm) and with 2,5-dihydroxybenzoic acid and 5-methoxysalicylic acid (9:1) as the matrix. The mass scale was calibrated with bovine serum albumin and myoglobin. Interpretation was done using XTOF 5.1.1 (Bruker). The fractions were digested with trypsin or endoproteinase Glu-C in solution after reduction with dithiothreitol (DTT), or in-gel after reduction with DTT and alkylation with iodoacetamide. The digests were first screened by MALDI-TOF MS as described above, with 2,5-dihydroxybenzoic acid or ␣-cyano-4-hydroxycinnamic acid as matrix and Bruker peptide standard (m/z 757-3147) for calibration. Peptides were sequenced using an Applied Biosystems Sciex Pulsar i QoTOF mass spectrometer (QStar) either by electrospray ionization (ESI) (nanospray) MS using self-pulled glass tips, or following SDS-PAGE and in-gel digestion, by LC-MS/MS using an online Waters CapLC system with a Michrom Magic C18 (5 m, 200 Å, 0.1 ϫ 150 mm) column. The data were analyzed using QAnalyst (Applied Biosystems) software, GPMaw 5.02 (Lighthouse data, Denmark), using the predicted CP sequences published by Mallinson et al. (8) and Sicairos et al. (25), and BLAST searches of the NCBI (www.ncbi. nlm.nih.gov/BLAST) and TIGR genome sequencing (www.tigr.org/ tdb/e2k1/tvg) data bases.
Cysteine Protease Assays-Cysteine protease activity was determined essentially as reported by Thomford et al. (34) using Z-RR-AMC (n-carbobenzoxyl-arginyl-arginyl-7-amido-4-methylcoumarin) as substrate. Briefly, 1 g of protein was added to 980 l of 0.5 M Tris-HCl, 0.15 M NaCl, 5 mM DTT (pH 7.6), and 10 l of Z-RR-AMC (1 mg/ml in Me 2 SO, 15 M final concentration) and the fluorescence was continuously monitored at an excitation wavelength of 380 nm and emission wavelength of 450 nm.
Culture of HVECs-HVECs were cultured as previously reported (27). HVECs were maintained at 37°C in an atmosphere of 5% CO 2 in Williams' complete medium (27,28) and were cultured to confluence and subcultured. Williams' complete medium for normal growth contained fetal bovine serum (10%), 2 mM L-glutamine, epidermal growth factor (10 ng/ml), insulin (5 g/ml), transferrin (5 g/ml), selenium (5 ng/ml), and antibiotic (streptomycin, 100 g/ml)-antimycotic (amphotericin B, 0.25 g/ml). The purity of HVECs was determined by anticytokeratin monoclonal antibody. The identification of squamous epi-thelium in HVEC cultures was performed by immunostaining with C23 antibody (obtained from Dr. R. Wu, University of California, Davis) as reported earlier (27,28). Cell purity is routinely 98 -99%. The only contaminants present were vaginal fibroblasts, which are not killed in response to parasites. For experimental studies, cells were subcultured in 96/48/24-well cluster plates or in 6-well plates. In some experiments, the cells were also cultured on a glass coverslip (Fisher Scientific) placed in a 6-well plate containing Williams' medium. All experiments were performed when cells were 70 -80% confluent and proliferating. When employed as controls, bovine cells (BVECs) were cultured similarly, as reported earlier (28).
Cytotoxicity Assay-HVECs were exposed to trichomonad parasites, SF, or the isolated CP fraction for 4 -20 h at 37°C. In control experiments, parasites or the CP fraction were omitted, or HVECs were treated with the homogenous CP30 isolated from the bovine parasite T. foetus. The WST-1 assay (Roche Diagnostics) was used to measure cytotoxicity/viability of HVECs, according to the manufacturers instructions. Cytotoxicity is calculated as 1 Ϫ E/C, where E/C is the ratio of absorbance of the formazan reading at 450 nm for experimental (E) versus control (C) samples. Data are derived from quadruplicate samples in three separate experiments.
In some experiments protease inhibitors such as aprotinin (Sigma Apoptosis Assays-A variety of methods, described below, were employed to evaluate apoptotic cell death of HVECs. Camptothecin (Fluka; 5 g/ml) was used as a positive control to induce apoptosis.
DNA Fragmentation-DNA fragmentation was quantitatively evaluated by Cell Death Detection ELISA PLUS (Roche Molecular Biochemicals), according to the manufacturers instructions. The enrichment of mono-and oligonucleosomes released into the cytoplasm was calculated as the ratio of the absorbance of the sample cells/absorbance of control cells. The enrichment factor was used as a parameter of apoptosis and shown on the y axis as mean Ϯ S.D. of triplicate experiments (36). An enrichment factor of 1 represents background or spontaneous apoptosis (generally ϳ6 -10%). Following incubation of HVECs in 48-or 24-well plates with appropriate agents the supernatant was collected and centrifuged, and the pellet was washed with PBS. The cell pellet and washed cells remaining in the wells were lysed and added to microtiter plates as described in the ELISA kit. The supernatant was not used in the ELISA when HVECs were infected with live parasites because it contains parasites along with dead HVECs.
Plasma Membrane Asymmetry-The annexin V-FITC apoptosis detection kit (BD Pharmingen) was used to measure plasma membrane asymmetry. Following treatment for 6 h in 6-well plates, HVECs were harvested by the addition of 0.25% trypsin, 5.3 mM EDTA for 2 min at 37°C (28). Trypsin was inactivated by addition of Williams' medium, cells were collected by centrifugation at 200 ϫ g and the pellet was washed with PBS. Washed cells were resuspended in the binding buffer and stained with both annexin V-FITC and propidium iodide (PI), according to the manufacturers protocol, and analyzed by flow cytometry. The FACStar plus flow cytometer (BD Bioscience) was set for FL 1 (annexin) versus FL 2 (PI) bivariate analysis. Data from 10,000 cells/sample were collected. The quadrants were set based on the population of healthy, unstained cells in untreated samples. CellQuest analysis of the data were used to calculate the percentage of the cells in the respective quadrants.
Caspase Activation-Ac-DEVD-AFC (Bio-Rad) was used as the substrate to detect caspase activation fluorometrically in HVECs treated with CPs, E-64-treated CPs, and camptothecin. HVEC lysates were prepared and incubated with the substrate in cell lysis buffer as recommended by the manufacturer. In addition, anti-ACTIVE® caspase-3 polyclonal antibody (Promega Corp.) was employed to detect an active form of caspase-3 in apoptotic HVECs. Immunocytochemical analysis was performed according to the manufacturers instructions. The involvement of caspases in HVEC apoptosis was further assessed by the incubation of HVECs with the apoptosis inhibitor Z-VAD-fmk (Enzyme System Product, Inc.). The inhibitor (75 M final concentration) was added 45 min prior to activation (with CPs) and cytotoxicity/apoptosis were evaluated by the WST-1 assay.

RESULTS
T. vaginalis Infection Induces Apoptosis in HVECs-In earlier reports (27,28), we demonstrated that incubation of trichomonad parasites with cultured VECs (HVECs and BVECs) leads to host cell death in a strikingly species-specific manner. We subsequently observed that a SF, obtained as described under "Experimental Procedures," is cytotoxic to host cells in a species-specific manner. SF obtained from T. vaginalis causes the complete destruction of HVECs in a time and concentration dependent manner (not shown).
Thus having shown earlier that T. vaginalis parasites are cytopathogenic to HVECs (27), we have now investigated whether the observed cell death is the result of apoptosis. As shown in Fig. 1B, HVECs undergo apoptosis when exposed to live T. vaginalis parasites or to the SF fraction. Camptothecin was used as a positive control to show apoptosis in HVECs. Although only the data obtained by a single technique for assaying apoptosis is shown in Fig. 1, we obtained identical results with flow cytometry using annexin V-FITC, MitoCapture (JC-1 dye, BioVision Inc.), and DNA fragmentation using bisbenzamide assays.
It was reported a number of years ago that trichomonads secrete a number of hydrolytic enzymes (23). Of particular note was the demonstration that CPs are present in the secreted fraction (Refs. 21-23; see below). As shown in Fig. 1A, heat treatment virtually eliminates the cytotoxicity of SF. Cytotoxicity requires a reducing agent (e.g. cysteine), but is diminished by E-64, a specific inhibitor of CPs, and by TLCK, which inhibits both cysteine and serine proteases. Other protease inhibitors, including leupeptin (0.63-2.5 M), aprotinin (46 -180 nM), and phenylmethylsulfonyl fluoride (3.4 M-1.3 mM) had no effect on SF-induced HVEC cell death. The SF fraction from the bovine parasite (TF-SF) is not cytotoxic to HVECs, nor does it induce apoptosis (Fig. 1). These results suggest that CPs in T. vaginalis SF are the initiators of HVEC apoptosis.
Isolation of Cytotoxic Cysteine Proteases-To examine the hypothesis that CPs are the cytotoxic agents in SF and may initiate parasite-induced HVEC apoptosis, we employed bacitracin affinity chromatography to purify active components in SF, as reported earlier for parasite CPs (21,34). The vast majority of the A 280 absorbing material in SF did not bind to the bacitracin affinity column and showed no cytotoxic activity (data not shown). The bound material was eluted as described under "Experimental Procedures." SDS-PAGE analysis, followed by Coomassie Blue staining, of this fraction showed a single band around 60 kDa and three closely spaced bands around 30 kDa. These two major fractions were subsequently separated on a Bio-Gel P-60 column and analyzed by SDS-PAGE, as shown in Fig. 2A. The number of bands in the approximately 30-kDa fraction (CP30) was variable, suggesting the possibility of proteolytic degradation. MALDI-TOF MS analysis of this fraction (Fig. 3) showed a peak centered at 23.8 kDa with a clear shoulder at 23.6 kDa and a high mass tail.
Protease activity of the CP30 fraction was measured fluorometrically using Z-RR-AMC as substrate. Protease activity required the presence of reducing agents such as cysteine (data not shown), or DTT, and was inhibited by E-64 and leupeptin (Fig. 2B), thus confirming its identity as a CP.
The isolated CP30 fraction is also biologically active, as it is cytotoxic toward HVECs. Cytotoxicity is concentration dependent and is inhibited by E-64, indicating that apoptosis is closely correlated with CP activity (Fig. 4). Leupeptin (62.5-250 M, not shown) also inhibits cytotoxicity. The specific activity (based on 50% cell death in cytotoxic assays) of this fraction is increased about 13-fold relative to SF. In contrast, the 60-kDa fraction showed less than 5% the cytotoxicity and CP enzymatic activity of the CP30 fraction. In control experiments, HVECs treated with the purified, homogeneous CP30 (33) from the bovine trichomonad, T. foetus, showed no cytotoxic effects (Fig. 4).
The isolated CP30 fraction from T. vaginalis induces HVEC apoptosis, as shown in Figs. 5 and 6. Fig. 5 shows the results of DNA fragmentation analysis. When the CP fraction was inactivated, either by treatment with E-64 or by heating, there was no measurable increase in apoptosis above background levels. Purified T. foetus CP30, obtained from the bovine pathogen, did not induce apoptosis in HVECs, a result entirely consistent with our previous observations on species specificity (27,28,33). Fig. 6 shows flow cytometric analysis of HVECs treated with the CP fraction and stained with annexin V-FITC. Treated HVECs (panel B) showed a significant increase in the annexinϩ cell population compared with control cells (panel A).
In contrast, HVECs treated with the CP fraction and E-64 (panel C) showed no increase in activity and showed normal morphology. In addition, HVECs undergoing apoptosis in the presence of the CP fraction showed activation of caspase-3, the ultimate caspase in apoptosis, and apoptosis was inhibited 92% by Z-VAD-fmk (75 M), a specific caspase inhibitor (data not shown).
Identification of Specific CPs-After it had been demonstrated that the isolated CP fraction is biologically active and that the activity is coincident with its CP activity, the isolated CP fraction was proteolytically digested with either trypsin or endo-Glu-C, analyzed by MALDI-TOF MS peptide mapping, and partially sequenced by ESI MS/MS, as described under "Experimental Procedures." Sequenced peptides and their assignments are listed in Table I. Especially in the digest mixture obtained with endo-Glu-C, many unspecific cleavages were detected, most likely because of autolysis of the CPs. Two peptides were found 2 Da below the predicted mass and probably feature a S-S bridge between 2 Cys residues; they were most likely reoxidized after DTT was removed prior to trypsin treatment. The major peptide signals found in ESI MS (Fig. 7) and MALDI MS (not shown) spectra correspond to CP4, the sequence of which is based on prediction from a cDNA clone reported by Mallinson et al. (7). The peptide maps provided by MS of the protein digests are largely consistent with the predicted CP4 sequence (Fig. 8). However, the prominent peptide NSWGTTWGEK is not completely matched by any predicted sequence. Because we observed no signal for the expected peptide NSWGTAWGEK from CP4, we speculate that the CP4 in the parasite strain used by us is slightly different from that studied by Mallinson et al. (7), or that there may be a 1-base pair sequencing error in the data bases, leading to the prediction of A rather than T in the peptide.
A few peptides correspond to CP3. Signals corresponding to CP3 in the MS spectra of the tryptic digest have only low abundance. Further peptides of significant abundance, such as VTGYVNVVEGDEKDLATK (Fig. 9), were only matched by a protein sequence very recently predicted in the TIGR data base as entry 55250.m00050 (Fig. 8). We call it CPT here; for final nomenclature of the TIGR sequence, we await the official publication by the consortium. In another preparation, which was not treated with TLCK, the peptide FMLTADYPYTAR, which matches only CP2, was found (not listed in Table I). Minor signals are present in the digests that could not be successfully sequenced. Although some of the sequences in Table I could also correspond to other partial sequences of predicted CPs, there is no unique evidence for the presence of any CPs other than those listed in Table I.
Although the 60-kDa fraction did not exhibit strong CP enzyme activity or cytoxicity it was also subjected to mass spectro-metric analysis. The molecular mass was determined to be about 53 kDa by MALDI-TOF MS (data not shown); peptide sequence analysis by ESI-MS/MS and peptide mapping by MALDI-FTMS published elsewhere (37) were used to identify it as a S-adenosylhomocysteine hydrolase (accession number U40872). DISCUSSION We have shown that co-incubation of T. vaginalis with HVECs causes extensive HVEC destruction within hours, but the human parasite has no effect on BVECs (27). In contrast, T. foetus parasites have no effect on HVECs, but rapidly destroy BVECs (27,28). Since those original observations, we found that soluble molecules obtained from conditioned medium (SF) also cause cell destruction in a species-specific manner. Therefore, we set out to identify and characterize the molecule(s) responsible for host cell destruction. In the present study, we show that CPs with molecular masses of ϳ24 kDa, obtained from SF, induce HVEC apoptosis. We recently reported (33) that a similar purified, homogeneous CP obtained from T. foetus, causes BVEC apoptosis; but as shown here, the T. foetus CP30 has no effect on HVECs. Conversely, as we demonstrated previously, the T. vaginalis preparation reported here has no effect on BVECs (33).
The ability of the isolated CPs to induce HVEC apoptosis is closely linked to CP activity, as it is inhibited by E-64 and requires a reducing agent for full activity (see Fig. 2B). The data provide overwhelming evidence that the biologically active agent consists of one or several CPs, because enzymatic assays using synthetic peptide substrates require a reducing agent, and are inhibited by a specific CP inhibitor, E-64. Finally, with the exception of a few experimental artifacts also found in negative control experiments (e.g. trypsin and endo Lys-C autolysis products, keratins), all of the peptides identified by extensive mass spectral peptide sequence analysis are derived from putative CPs.
North, Coombs and their colleagues (7, 8, 21-23) examined a number of secreted proteases from both T. vaginalis and T. foetus. Using a combination of SDS-PAGE analyses and synthetic substrates they demonstrated that several different proteases are present in the secreted fraction (21)(22)(23). Multiple proteases have been observed in parasite cell lysates and as many as 23 protease species have been observed on two-dimensional SDS-PAGE of T. vaginalis extracts (38). It has been suggested, however, that many of the observed species are procedural artifacts and/or are the products of post-translational modification. In fact, Garber and Lemchuk-Favel (39,40) demonstrated that a 60-kDa CP from T. vaginalis fragments into 23-and 43-kDa species. The soon to be published T. vaginalis genome should directly address some of these issues.
Garber et al. (24) reported the cloning of a cDNA from a GT 11 library representing a portion of a secreted 60-kDa CP, whereas Leon-Sicairos et al. (25,26) reported the cloning of three CPs, including an apparently intracellular CP, tvcp12. Although the approach used by Mallinson et al. (7) to clone T. vaginalis CPs did not allow the investigators to identify the CPs as being secreted, our MALDI-TOF MS (Fig. 3)   Peptides were sequenced after (A) trypsin digest in solution followed by nanospray MS/MS, (B) trypsin in gel digest followed by LC-MS/MS, or (C) Glu-C digest in solution followed by LC-MS/MS. The isobaric Ile and Leu residues have been assigned after the observed peptide sequence was aligned with the predicted sequence. Residues in italics are only matches in mass, whereas the fragment signals between them were not observed. ESI (nanospray) MS spectrum of an insolution tryptic digest of TLCK-treated Tv30, with the assignments indicated for the more abundant signals. These data suggest that CP4 is the major component in the fraction. The signal marked CP4? is the NSWGTTWGEK peptide we assume to be a part of our CP4 sequence (see text). The numbers shown in parentheses correspond to the amino acid positions as shown in Fig. 8. from the data presented here, as well as comparisons among the various CP sequences, that this family of T. vaginalis CPs are highly homologous, and belong to the papain family (clan CA) of CPs. Many of the CPs in this family are biosynthesized as pro-enzymes and their activation is accompanied by removal of about 100 amino acids (ϳ10 kDa) at the N terminus. The predicted molecular masses of the CPs identified here, based on sequences found in the genome, are close to 33-36 kDa. The molecular masses of these CPs, as determined by MALDI-TOF MS analysis (23.6 and 23.8 kDa), most likely represent the mature, processed, forms of the enzymes CPT and CP4, respectively.
Microbes have developed mechanisms to stimulate the apoptotic signal transduction cascade, which likely plays a role in pathogenesis (30 -33). Apoptotic cell death has been studied in detail as a response to several bacterial and viral infections (30) and it is apparent that viral, bacterial, and protozoan pathogens have evolved a variety of different strategies to modulate host cell apoptosis (32,33). Unlike bacterial and viral infections, relatively little is known regarding apoptotic cell death as a response to parasitic infections. There are a few published reports that show that parasitic pathogens, e.g. Acanthamoeba histolytica (41), Plasmodium falciparum (42), Trypanosoma cruzi (43), Cryptosporidium parvum (44), Toxoplasma gondii (45), and Entamoeba histolytica (46), can kill mammalian cells by an apoptotic mechanism that occurs in response to infection. However, the precise mechanisms by which individual pathogens induce cell death in specific host cells remain to be elucidated. It can be expected that the mechanism(s) by which extracellular parasites, including T. vaginalis, induce host cell apoptosis are very likely to be quite different from the mechanism(s) used by intracellular parasites. CPs have been shown to be essential for E. histolytica-induced pathology, where they destroy host tissue (46,47).
Knowledge of species-specific apoptotic pathways will have important implications for controlling or treating human and bovine trichomoniasis. Caspases activated during apoptosis cleave specific protein targets, and thus bring about the irreversible commitment to cell death (48,49). Caspase-3 plays an important role in mediating the various morphological changes associated with Fas-mediated apoptosis (49). Caspase-3 activation has been implicated in apoptotic cell death induced by E. histolytica parasites in Jurkat cells (46,47) and mosquito mid-gut epithelial cells infected by the P. falciparum parasite (50). Recently, we also reported the activation of caspase-3 in T. foetus CP30-induced apoptosis in BVECs and BUECs (33,51). Our present data provide clear endorsement for the involvement of caspase-3 in HVEC apoptosis.
This study is the first to demonstrate apoptotic cell death in HVECs in response to T. vaginalis infection. From these re-  7), Tf8 is the predicted sequence of CP8 from Mallinson et al. (8). CPT is derived from a BLAST search for VTGYVNVVEGDEKDLATK in the TIGR T. vaginalis data base. CP3# contains CP3 sequence corrected by rapid amplification of cDNA ends results (data not shown). Sequenced peptides are underlined. The residues in the active center are marked in bold. NSWGTSWGEQ would be a theoretical match in CP2, but is left unmarked because it seems less likely than the NSWGTSWGEK peptide. sults, in combination with our recent study of T. foetus and BVECs, we are gaining a better understanding of the trichomonad infectious processes at the molecular level. We recently showed (33) that T. foetus CP30 corresponds to CP8 reported by Mallinson et al. (8). Among the T. vaginalis proteins found, CP4 and CPT are most similar to the incomplete sequence of CP8 from T. foetus, when the latter is compared against the TIGR data base. Although we have yet to explicitly determine which of the T. vaginalis CPs are capable of inducing HVEC apoptosis, it is clear that there is a strong correlation between activity of CP30 fraction and apoptosis.
Our studies, revealing that CPs constitute some of the molecules involved in pathogenesis, represent an important step toward reaching a thorough understanding of the significant differences between human and bovine trichomoniasis, and of their similarities. The differences in disease progression between human and bovine trichomoniasis may be manifested by something as basic as the specificity and multiplicity of parasite CPs. Although the differences in DNA sequence and activities toward synthetic substrates between the T. vaginalis CPs and T. foetus CP30 appear to be slight, the impact on their respective host cells is dramatically species-specific. Pharmacological and immunological exploitation of the host-parasite interactions during apoptosis could lead to new forms of intervention in the disease process. Indeed, it has been suggested by several authors that parasite CPs may be useful targets for the development of novel chemotherapies (1,47). Thus, further research on parasite CP-induced apoptosis will not only provide new insights into the induction of apoptosis but may also open up new therapeutic avenues.