The amino acid sequences of TC-2 (TVAG_272260) and the TC-2Δ11 deletion mutant, as well as the target CPs TvCP2 (TVAG_057000) and TvCP39 (TVAG_298080) (https://trichdb.org/trichdb/app/) (Release 63 01 May 2024) (Alvarez-Jarreta et al., 2024) and papain (PDB:1CVZ) (Tsuge et al., 1999), were modeled using the AlphaFold3 structure prediction server (https://alphafoldserver.com/) (Abramson et al., 2024). The subsequent analysis, preparation, and visualization of the models were performed using ChimeraX software (https://www.rbvi.ucsf.edu/chimerax) (Meng et al., 2023). The models were validated using the MolProbity server (http://molprobity.biochem.duke.edu/index.php) (Williams et al., 2018). Molecular docking of each of the inhibitors (TC-2 and TC-2Δ11) with each target CP was performed using the AlphaFold3 server (Abramson et al., 2024), Prodigy server (https://rascar.science.uu.nl/prodigy/) (Xue et al., 2016) and Area Affinity server (https://affinity.cuhk.edu.cn/) (Yang et al., 2022).

The TC-2 gene was previously cloned within the expression vector pCold I. This construct was subsequently transformed into E. coli BL21 (DE3), and its expression was induced as previously reported (Puente-Rivera et al., 2014). For the expression of rTC-2Δ11, the TC-2 gene sequence (TVAG_272260) was modified by deleting its first 11 amino acids, which include the four N-terminal cysteines. The DNA sequence was optimized for its expression in E. coli and synthesized by Synbio Technologies (LLC, NJ, USA). It was cloned between the NcoI and XhoI restriction sites within pUC57. The optimized sequence was subsequently subcloned and inserted into the pCri8a expression vector (Addgene plasmid # 61317; http://n2t.net/addgene:61317; RRID: Addgene_61317) (Goulas et al., 2014). The pCri8a construct was transformed into E. coli BL21 (DE3), and clones were analyzed by plasmid double digestion with the NcoI and XhoI restriction enzymes and selected for subsequent protein expression. The expression of both proteins was first induced in a flask culture. Transformed E. coli BL21 (DE3) clones were grown in 0.5 L of Luria–Bertani (LB) media (casein peptone 10 g/L, yeast extract 5 g/L, NaCl 10 g/L) or yeast extract tryptone media (2TY) (16 g/L tryptone, 10 g/L yeast extract, 5 g/L NaCl) supplemented with ampicillin (100 µg/mL) or kanamycin (50 µg/mL) as selective antibiotics. The bacterial culture was grown at 37°C and 200 rpm until the optical density (OD₆₀₀) reached 0.6. Recombinant protein expression was induced by adding a final concentration of 0.2 mM β-D-1-thiogalactopyranoside (IPTG) for 12 h with the culture maintained at 30°C and 200 rpm.

To purify the recombinant proteins from a single production batch, the expression of rTC-2 and rTC-2Δ11 was scaled up to a 3 L jar bioreactor (BioFlo/Celligen 310 controller, Eppendorf) using a fed-batch culture. Briefly, the volume of medium necessary to inoculate a bioreactor was taken from a seed inoculum. Fermentation was started with an OD₆₀₀ of 0.05 in 2 L of BSM medium (3.5 g/L KH₂PO₄, 5 g/L K₂HPO₄, 3.5 g/L (NH₄)₂HPO₄, 4 mL/L 1M MgSO₄, 20 g/L glycerol, 5 g/L yeast extract, and 1 mL/L trace metals) with the appropriate antibiotic added for the selection of each protein and 0.5 mL/L of 10% Antifoam 204 (Sigma). The culture was grown at 37°C for approximately 12 h until glycerol was depleted. Protein expression was induced with 0.5 mM IPTG for 16 h at 18°C, and 50% glycerol was fed simultaneously at a rate of 3 mL/L*h.

Biomass was recovered by centrifugation in Sorvall LYNX (Thermo Fisher Scientific, USA) at 4248× g for 30 min. The induction of expression was validated using 15% SDS−PAGE. Lysis was then performed by adding 20 mL of lysis buffer (50 mM Tris-HCl [pH 8.0], 500 mM NaCl, 5 mM imidazole, 10% glycerol, and 0.02% sodium azide) per gram of wet biomass, followed by the addition of lysozyme to a final concentration of 0.5 mg/mL and incubated for 30 min at 37°C and 220 rpm. Then, the sample was immersed in an ice water bath (4°C) and sonicated using a 550 Sonic Dismembrator (Thermo Scientific, USA) at 30% of the wave amplitude by 6 x 30 s pulse with 30 s interval between each sonication cycle. The soluble fraction was recovered by centrifugation at 20 216 × g for 30 min at 4°C. Purification was performed by affinity chromatography on prepacked Ni-Sepharose columns (Cytiva, USA). The column was equilibrated with lysis buffer, and the protein was eluted with elution buffer (50 mM Tris-HCl [pH 8.0], 500 mM NaCl, 500 mM imidazole, 10% glycerol, and 0.02% sodium azide). For long-term storage at -80°C, sterile glycerol was added to the protein mixture at a final concentration of 25%. The final protein concentration was determined by a bicinchoninic acid (BCA) assay (Thermo Fisher, USA). The samples were analyzed by 15% SDS−PAGE under reducing or nonreducing conditions. Recombinant proteins were subjected to dynamic light scattering (DLS), thermal shift (TS), and size exclusion chromatography (SEC) after a buffer exchange (50 mM Tris-HCl pH 8.0, 100 mM NaCl, 0.02% NaN₃) with gel filtration on PD-10 columns (Cytiva, UK).

Immunodetection of rTC-2 and rTC-2Δ11 was performed by Western blot (WB) assays, as previously reported (Puente-Rivera et al., 2014). Briefly, 4 µg rTC-2 and rTC-2Δ11 proteins were separated by 15% SDS-PAGE, and transferred onto a nitrocellulose (NC) membrane. The NC membrane was blocked with 5% skim milk for 18 h at 4°C and incubated with the Rα-rTC-2 primary antibody (1:500), as previously described (Puente-Rivera et al., 2014); washed with Tris-buffered saline (TBS, 20 mM Tris-HCl, 500 mM NaCl)-0.5% Tween 20; and incubated with a goat anti-rabbit peroxidase-conjugated secondary antibody (1:3000) (Bio-Rad, USA). Protein detection was performed by chromogenic method with 4-chloro-naphtol (0.5 mg/mL) in TBS and 0.05% H₂O₂.

The hydrodynamic diameter (Dh) of both recombinant proteins was determined using DLS under reducing (1, 5, or 15 mM DTT) or nonreducing conditions for samples after 1, 3, and 6 months of storage at 4°C. Samples of recombinant proteins (1 mg/mL) on storage buffer (50 mM Tris-HCl pH 8.0, 100 mM NaCl, 0.02% NaN₃) were centrifuged at 17 949 × g for 10 min (4°C) and filtered through 0.22 µm membrane. Measurements were performed using a Zetasizer nano ZSP (Malvern Panalytical, USA) in 173° backscatter mode at 25°C with a 40 µL cuvette [ZEN0040] (Brand, Germany). Two replicates were obtained for each sample, and at least three data acquisitions were conducted. The data were analyzed using the Zetasizer software (v.7.12; Malvern Instruments).

The TS assay was performed in 50 µL of Tris buffer (50 mM Tris-HCl pH 8.0, 100 mM NaCl, 0.02% NaN₃) containing 20 µM protein and 1x Sypro orange (Invitrogen, USA). Assays were performed using a real-time qPCR Gentier 48E thermal cycler (Tianlong, China). The temperature gradient was adjusted from 25 to 95°C with 1°C/min increments. Each sample was run in triplicate.

The molecular weight and size were determined using analytical SEC for rTC-2 and rTC-2Δ11 in samples stored for 1 and 3 months. SEC was performed using a Superdex 200 pg HiLoad 26/600 column (GE Healthcare Bioscience, Sweden) on an ÄKTA Pure 25 system (GE Healthcare, USA). Before sample injection, the column was calibrated with gel filtration standards [1511901] (Bio-Rad). Five milligrams of protein were loaded on three runs, and the proteins were resolved with elution buffer (50 mM Tris-HCl, pH 8.0; 100 mM NaCl; and 0.02% NaN₃ at a velocity of 29.38 cm/h). The results were analyzed with Unicorn software (v7.1, GE Healthcare, USA).

The ability of rTC-2 and rTC-2Δ11 to inhibit protease activity was evaluated using papain and T. vaginalis PRE, and the fluorogenic substrates Z-Phe-Arg-MCA (Sigma) and E-64 (trans-Epoxysuccinyl-L-leucylamido(4-guanidino)butane) as CP control inhibitor. Papain (2 ng/µL) was activated with buffer (50 mM Tris, pH 6.5; 5 mM DTT) for 10 min at 25°C and incubated with rTC-2, rTC-2Δ11, or E-64 (0.7 µm). The reaction started upon the addition of the fluorogenic substrate Z-Phe-Arg-MCA (40 µM), as previously reported (Puente-Rivera et al., 2014). The fluorescence was measured at an excitation wavelength (λ) of 355 nm and emission wavelength (λ) of 460 nm on a SpectraMax Gemini EM spectrofluorometer (Molecular Devices). For the inhibition assays of PRE proteolytic activity, 2 x 10⁷ parasites from the CNCD 188 trichomonad isolate resuspended in a PBS pH 8.0 were lysed with 0.5% sodium deoxycholate (DOC) in the absence of protease inhibitors and centrifuged in a 10% sucrose gradient at 16 200 x g, for 30 min at 4°C. The supernatant was recovered, and PRE protein concentration was estimated at A₂₈₀ against a previous calibration curve. Immediately, the inhibition assays were done as described for papain by using 20 µg of PRE incubated in activation buffer (50 mM NaOAc, pH 5.0; 4 mM EDTA; and 8 mM DTT) with different concentrations (0 µM, 0.7 µM, 2 µM, or 2.7 µM) of rTC-2, rTC-2Δ11, or E-64 used as control CP inhibitor. The time course of the fluorescence intensity for each assay was plotted using a GraphPad Prism 8.0.0 (GraphPad Software, Boston, Massachusetts USA, www.graphpad.com).

To determine the inhibition of the trichomonal cysteine protease-dependent cytotoxicity on HeLa cells, T. vaginalis (CNCD 188) was grown for one week in trypticase-yeast extract-maltose (TYM) medium supplemented with 10% heat-inactivated adult bovine serum (HIBS) at 37°C, and for HeLa cell cultures, Dulbecco’s modified Eagle’s medium (DMEM) (Invitrogen-Gibco, Carlsbad, CA, USA) supplemented with 10% HIBS was used. HeLa cell monolayers were prepared by inoculating 3.5 x 10⁴ cells per well in flat bottom 96-well culture plates, incubated at 37°C for 24 h in a 5% CO₂ atmosphere to allow the formation of confluent cell monolayers. The interaction of HeLa cells with T. vaginalis parasites was performed at a 1:5 ratio, as previously described (Alvarez-Sánchez et al., 2000). Briefly, parasites were previously incubated in interaction medium (DMEM: TYM, 2:1) without serum in the presence of rTC-2, rTC-2Δ11, E-64 as a positive inhibition control, or bovine serum albumin (BSA, an unrelated protein) as a negative inhibition control at 18 µM, 36 µM, and 54 µM concentrations. Then, the parasites were added to the HeLa cell monolayers and incubated for 2 h at 37°C and 5% CO₂ atmosphere. Monolayer destruction was measured by a colorimetric method using crystal violet, as reported (Alvarez-Sánchez et al., 2000). Eluted stain from remaining cells was quantified with a VersaMax spectrophotometer (Molecular Devices) at a wavelength of 570 nm. Assays were conducted using technical and biological triplicates. Destruction of HeLa cell monolayer by untreated parasites was taken as 100% cytotoxicity and the inhibition percentages of each group were estimated, and an analysis of variance (Three-way ANOVA) was performed to evaluate significant differences between groups using the GraphPad Prism 8.0.0 (GraphPad Software, Boston, Massachusetts, USA, www.graphpad.com).

To determine whether the TC-2Δ11 protein, which lacks the first 11 amino acids of the N-terminus, could still interact with CPs and how these interactions compared with those of wild-type TC-2, in silico analyses were performed. These analyses were used to predict interactions between T. vaginalis CPs TvCP2 and TvCP39 and papain, a CP protein from an unrelated organism. Thus, the molecular dockings were done within the correct range of the predicted template modeling (pTM) score, the interface-predicted template modeling (ipTM) score, and the AlphaFold3 ranking score ( Supplementary Table S1 ). The molecules were correctly validated for use in molecular docking in MolProbity ( Supplementary Table S2 ). Moreover, in the Ramachandran plots ( Supplementary Figure S1 ), the residues were within the favored and allowed regions for every molecule used in the molecular docking experiments except for the modeled TC-2 and TC-2Δ11 inhibitors. In the case of TC-2, two outlier residues were identified at G5 and V54, whereas in TC-2Δ11, one outlier was observed at V54 ( Supplementary Table S3 ). Although these outliers indicate unfavorable amino acid positions, models were still validated. G5 is located in the unstructured N-Terminus (1-20 residues) of TC-2, and V54 is located at the end of the beta-sheet (S48-V54), in the Ramachandran’s plots, they were found too close to the allowed region limit and both amino acid residues have small side chains that may or may not restrict free rotation along the protein.

Molecular docking of rTC-2 with papain ( Figure 1A ) revealed 11 hydrogen bond interactions ( Table 1 ). Two amino acids of rTC-2 (K52/S55) that interact with papain belong to the central domain of cystatin (QKVVSG) (Puente-Rivera et al., 2014). Binding affinity with ΔG~ (-13.1, -13.5) kcal/mol and K_d ~ (1.2 × 10^-10, 2.36 × 10^-10) suggest that this binding can occur spontaneously ( Table 2 ). In the case of the interaction between rTC-2Δ11 and papain ( Figure 1B ), three fewer hydrogen bonds interact between both proteins ( Table 1 ), suggesting a lower affinity than rTC-2 and papain given that the binding affinity between rTC-2Δ11 and papain ΔG~ (-10.3, -7.4) kcal/mol and K_d ~ (3.9 × 10^-6 - 2.6 × 10^-8) ( Table 2 ).

However, when rTC-2 interacted with T. vaginalis CPs, the number of interactions increased compared with those observed for rTC-2 binding with papain ( Tables 3 , 4 ). Specifically, 14 interactions were noted between rTC-2 and TvCP2 ( Figure 2A ; Table 3 ), and 16 interactions were noted between rTC-2 and TvCP39 ( Figure 3A ; Table 4 ). The number of interactions of rTC-2Δ11 with T. vaginalis CPs was very close to that observed for rTC-2, with 12 interactions with TvCP2 ( Figure 2B ; Table 3 ), and 15 interactions found with TvCP39 ( Figure 3B ; Table 4 ). Thus, although there are fewer interactions between trichomonad CPs with rTC-2Δ11 than with rTC-2, recombinant inhibitors exhibit more interactions with T. vaginalis proteases than papain. Moreover, both rTC-2 and rTC2Δ11 also interacted with the Q51/K52/S55 conserved sites of T. vaginalis CPs ( Tables 3 , 4 ).

In addition, the ΔG values for rTC-2 and rTC-2Δ11 were greater in terms of interactions with TvCP2 and TvCP39 than with papain, and the same trend was noted for K_d ( Table 2 ). This finding indicates a greater specificity of rTC-2 for T. vaginalis proteases. Moreover, the slightly lower affinity binding ΔG and K_d values of rTC-2Δ11 with TvCP2 and TvCP39 indicate a lower affinity of the mutant toward the proteases; however, high values were also observed, indicating spontaneous binding ( Table 2 ).

The recombinant expression of rTC-2 and rTC-2Δ11 was scaled up to a 2L bioreactor to do all assays with a single batch and determine their production yield at this scale. Recombinant TC-2 and TC-2Δ11 were expressed in E. coli, producing soluble protein ( Figures 4A, D ). A volumetric yield of ~990 mg of protein per liter of culture was obtained, and up to >95% purity was achieved after purification by IMAC ( Figures 4B, E ). Unlike rTC-2, rTC-2Δ11 did not aggregate ( Figures 4C–H ). Furthermore, even under nonreducing conditions, rTC-2Δ11 shows a faint band of aggregation (a dimer) compared with that of rTC-2 ( Figures 4G, H ). In addition, WB analysis with Rα-rTC-2 antibodies confirmed that each observed band belonged specifically to the rTC-2 protein and that there were no other bacterial-contaminating proteins ( Figures 4C, F ). The expression yield of ~ 1 g protein/L culture facilitated the purification process, it was sufficient to perform all the experiments using a single protein batch, and it is suitable for large scale production.

To assess whether rTC-2Δ11 preserves its inhibitory capacity, inhibition assays were performed with papain as a model cysteine protease and with trichomonad PREs. Inhibition assays of papain proteolytic activity by rTC-2 and rTC-2Δ11 revealed that the inhibitory capacity of rTC-2Δ11 to papain activity was lost in the mutant as compared to rTC-2 in the range of concentrations tested. However, these differences were not observed when rTC-2 and rTC-2Δ11 were incubated with T. vaginalis PRE. Interestingly, even at the lowest concentrations tested, inhibition by the rTC-2Δ11 protein completely ablated the trichomonal proteolytic activity, similar to the inhibition caused by rTC-2 ( Figure 6 ). These results suggest that the inhibition of CPs from other organisms, at least papain, is less specific than the inhibition of CPs from the same parasite. Furthermore, the essential sites for the interaction of the inhibitors with the trichomonad proteases were not located in the first amino acids of the N-terminal end. Thus, the four cysteines present at this end do not play an essential role in the inhibitory capacity toward T. vaginalis proteases.

Cytotoxicity analyses were performed to verify whether rTC-2Δ11 also confers protection to HeLa cell monolayers from T. vaginalis protease activity. This is the case rTC-2, as its protective activity has already been reported (Puente-Rivera et al., 2014). Figure 7 shows that both recombinant inhibitors significantly decreased the cytotoxicity levels at all concentrations tested and at all times compared with those of BSA, an unrelated protein, which was used as a control ( Figure 7 ). Notably, both rTC-2 and rTC-2Δ11 had similar protective effects on HeLa cell monolayers against T. vaginalis protease activity. These proteins provided up to 81% and 85% protection in a concentration-dependent manner. These results indicate that, under the conditions tested, rTC-2Δ11 retains its CP inhibitory ability and protective role over HeLa cell monolayers without the first 11 amino acids. Thus, the N-terminus of TC-2 does not contain the amino acids essential for the inhibitory capacity of TC-2 against T. vaginalis CPs.