Caco-2 (human colorectal adenocarcinoma) from the C2BBe1 lineage and MDCK (Madin Darby canine kidney) type I (Cereijido et al., 1978; Sambuy et al., 2005) cells (passages: 15–37 and 20–100, respectively) were grown in DMEM medium (Gibco) supplemented with penicillin (100 IU/ml), streptomycin (100 mg/ml) (in vitro), 10% fetal bovine serum (Gibco), and insulin (0.08 U/ml) (Eli Lilly), at 37°C in a 5% CO₂ atmosphere (Natoli et al., 2012). Cell confluence was verified by TEER measurement and by phase contrast microscopy. E. histolytica trophozoites, strain HM1:IMSS, clone A (Orozco et al., 1983) were axenically cultured at 37°C in TYI-S-33 medium and harvested during logarithmic growth phase (Diamond et al., 1978). All experiments presented here were performed at least three independent times by duplicate.

To produce and activate rEhCP1, rEhCP2, rEhCP5, and rEhCP112 recombinant proteases, E. coli BL21 DE3 bacteria were transformed with the pQE80L-ppEhcp112 or pet15-ppEhcpa1 or pet15-ppEhcpa2 or pet15-ppEhcpa5 constructions (kindly donated by Dr. Jaime Ortega from the Department of Biotechnology and Bioengineering, CINVESTAV-IPN, México). Recombinant proteins were induced with 1 mM IPTG and recovered from the inclusion bodies, using the solubilization buffer (20 mM sodium phosphate, pH 7.4, 0.5 M NaCl, 10 mM imidazole, and 8 M urea). Recombinant enzymes were purified using a HisPur cobalt resin (Thermo Fisher Scientific) and refolded in a PD-10 desalting column (GE Healthcare Science) with the refolding buffer (5 mM CaCl₂, 0.02% SDS, and 100 mM Tris–HCl pH 8.0). Recombinant proteins were activated with 0.05% β-mercaptoethanol (β-ME), during 20 min at 25°C. The proteolytic activity of the enzymes was confirmed by zymogram assays in 12% polyacrylamide gels (PAGE) co-polymerized with 0.1% gelatin and stained with Coomassie blue solution. We also used the Z-Phe-Arg substrate degradation and releasing of 7-amino-4-methylcoumarin (AMC) was measured in a fluorimeter (Fluoroskan Ascent FL) by excitation and emission wavelengths at 355 and 460 nm, respectively. Bovine serum albumin (BSA) was used as a negative control.

For EhCP112 immunodetection, rabbit polyclonal antibodies against a specific EhCP112 C-terminus peptide (N-431KYHSNSTYVQFYNHT444-C) were used (Betanzos et al., 2013). For immunoprecipitation assays, a polyclonal antibody against EhCP112 (α-EhCP112) was obtained after three immunizations (each 2 weeks) of New Zealand male rabbits with 300 μg of rEhCP112p diluted in TiterMax® (Sigma) adjuvant. Before immunization, rabbits were bled to obtain pre-immune serum (PS). Other primary antibodies used were: mouse monoclonal α-actin (kindly donated by Dr. Jose Manuel Hernández from the Department of Cellular Biology, CINVESTAV-IPN, México), mouse monoclonal α-claudin-1 (Invitrogen, cat number: 51-9000), α-claudin-2, α-claudin-4 (Thermo Scientific, cat number: 32-5600 and 32-9400, respectively), α-occludin (Invitrogen, cat number: 33-1500), α-mucin-2 (Thermo Scientific, cat number: MA1-35701) and α-GAPDH (Santa Cruz, cat number: sc-365062) antibodies; and rabbit polyclonal α-ZO-2, α-ZO-1 (Invitrogen, cat number: 37-4700 and 40-2200, respectively) anti-α/β tubulin (α-tubulin) (Cell Signaling, cat number: 2148S) antibodies. Secondary antibodies derived from goat included: α-rabbit and α-mouse HRP-labeled IgGs (1:10,000) (Life technologies G21234 and 62–6,520, respectively); and α-rabbit and α-mouse FITC and TRITC-labeled IgGs (1:150) (Zymed, 65–6,111 and Life technologies F-2761, respectively).

rEhCP112a (10–30 μg/cm²) or trophozoites (10⁵/cm²; 2:1 cells to trophozoites ratio) or 5 mM EDTA or refolding buffer were added to the apical side of confluent Caco-2 or MDCK cells and incubated for different times at 37°C.

rEhCP112p was labeled with Alexa 647 fluorescent dye, following the manufacturer recommendations (Molecular Probes). Briefly, 100 μg of rEhCP112p were equilibrated with 0.1 M sodium bicarbonate to allow the adequate succinimidyl ester reaction with primary amines of the protein, to form stable dye-protein conjugates. Confluent Caco-2 cells were grown on coverslips and Alexa-labeled rEhCP112p was added for 0, 0.5, 1, and 5 min to the apical side of the cells. After washing five times with PBS, pH 6.8, cells were fixed with absolute ethanol for 30 min at −20°C and nuclei were stained for 5 min with 2.5 μg/ml of 4′,6-diamidino-2-phenylindole (DAPI) (Zymed). Cells were mounted using Vectashield (Vector laboratories) and analyzed in a confocal microscope (Leica TCS_SP5_MO) through Z-stack sections of 0.5 μm and xy- and zy-planes. In all cases, 10 fields were analyzed per condition. Representative images were selected for each time.

Different concentrations of rEhCP112a (0–30 μg/cm²) were added to the apical side of confluent Caco-2 or MDCK cells grown on transwell filters (6.5 mm of diameter and 0.4 μm pore) (Corning) and then, TEER was measured using an EVOM epithelial voltmeter (World Precision Instruments) during 90 min. In some assays, rEhCP112a was pre-incubated for 10 min with 20 μg/ml of E-64 or 10 μg of α-EhCP112 antibody, before the enzyme was added to Caco-2 cells. rEhCPA1, rEhCP2 or rEhCP5 (10 μg/cm²) or trophozoites (10⁵/cm²; 2:1 ratio) or papain were added to Caco-2 cells. Each transwell measurement was normalized accordingly to its initial value (above 1,000 Ω/cm²) before treatment (Betanzos et al., 2013). During the TEER assays, cells were maintained at 37°C to avoid temperature changes. To verify the plasma membrane integrity, treated cells were stained by 2 μg/ml of propidium iodide (PI) for 10 min. We also verified cell viability using the Sytox green reagent (Thermo Scientific), following the manufacturer instructions. Cells were washed and observed through a laser confocal microscope.

Ehcp112 complete gene was PCR amplified with specific primers containing BamHI and KpnI restriction sites (forward: 5′-GGGGTACCATGACAGCGATTGTTGTCGCTTTTTT-3′, reverse: 5′-CGGGATCCTTACTTATCGTTCGTCATCCTTGTAATCGATTGTATGATTGTAGAATTGG-3′) under the following conditions: 95°C 15 s, 60°C 30 s, 72°C 90 s during 30 cycles. Then, the PCR product was cloned into pJET1.2 plasmid (Fermentas) according to the manufacturer recommendations. The Ehcp112/pJET construction was digested and the Ehcp112 fragment was subcloned in the pExEhNeo plasmid (Hamann et al., 1995). This plasmid was amplified in E. coli DH5α, purified by Qiagen Midi Kit (Qiagen) and automatically sequenced. Transfection was performed as described (Avalos-Padilla et al., 2015). Transfected trophozoites were selected with increasing concentrations of G-418 (a neomycin analog) until stable growth was achieved (20 μg/ml). Overexpression of the Ehcp112 gene was verified by RT-PCR by specific primers and using s2 ribosomal gene as internal control. WB assays were performed using α-EhCP112 (1:5,000) and α-actin (1:3,000) antibodies. Laser confocal microscopy assays were performed using α-EhCP112 (1:100) antibodies in paraformaldehyde fixed and triton X-100 permeabilized trophozoites and nuclei were stained by propidium iodide.

Paracellular permeability of Caco-2 and MDCK cells was determined by red ruthenium dye (RR) penetration and by FITC-dextran diffusion assays. For RR penetration assays, cell monolayers were incubated for 30 min with rEhCP112a (10 μg/cm²) or with trophozoites or EDTA (5 mM) or refolding buffer. Then, cells were fixed with 2.5% glutaraldehyde and incubated for 1 h at room temperature (RT) with 1% osmium tetraoxide and 5 μg/ml of RR. Samples were dehydrated by incubation in increasing concentrations of ethanol and propylene oxide and embedded in Polybed resin (Polysciences). Ultrathin sections (60 nm) were obtained and analyzed through a Jeol 1011 transmission electron microscope.

For FITC-dextran assays, 3 mg/ml of FITC-dextran (2–4 kDa) (Sigma Aldrich) were added to the apical side of confluent epithelial cells. Then, rEhCP112a (10 μg/cm²) or trophozoites or EDTA or refolding buffer were added to the upper chamber of transwells and incubated for 90 min at 37°C with gentle shaking. Samples from the basal chamber were collected each 30 min and the diffused fluorescent tracer was measured in the fluorimeter by excitation and emission wavelengths at 492 and 520 nm, respectively. Emission values were converted to FITC-dextran concentration, using a standard curve (Matter and Balda, 2003).

BSA-BODIPY FL C₁₂-Sphingomyelin (Molecular Probes) (50 μg) re-suspended in P buffer (10 mM HEPES pH 7.4, 145 mM NaCl, 1 mM sodium pyruvate, 10 mM glucose, 3 mM CaCl₂, and 3.4 mg of ultrapure BSA) was added to confluent Caco-2 cell monolayers grown on coverslips and then, incubated for 10 min at 4°C. After washing three times with P buffer, rEhCP112a (10 μg/cm²) or refolding buffer or EDTA were added to the apical side of the cells. Samples were processed for immunofluorescence confocal microscopy, using the α-occludin antibody as a TJs marker, as described below. Images were acquired in the zy-plane, to visualize Bodipy diffusion (Matter and Balda, 2003).

Caco-2 cells were incubated for 2, 10, and 30 min at 37°C with 10 μg/cm² of rEhCP112a or E-64 (20 μg/ml for 10 min) pre-treated rEhCP112a or live trophozoites. After washing twice with PBS, cells were scrapped with a rubber policeman and lysed in RIPA buffer [40 mM Tris-HCl pH 7.6, 150 mM NaCl, 2 mM EDTA, 10% glycerol, 1% Triton X-100, 0.5% sodium deoxycholate, 0.2% SDS, 1 mM PMSF and 1 mM Complete™ protease inhibitor cocktail (Roche)] under continuous and vigorous shaking. Extracts were sonicated three times for 5 s, centrifuged for 15 min at 15,300 × g to eliminate cellular debris and boiled with sample buffer (50 mM Tris-HCl pH 6.8, 2% SDS, 10% glycerol, 1% β-mercaptoethanol, 12.5 mM EDTA, and 0.02 % bromophenol blue; Betanzos et al., 2013). Protein samples were separated by 6, 8, 10, or 15% sodium dodecyl sulfate (SDS)-PAGE, transferred to nitrocellulose membranes and incubated 1 h with 5% non-fat milk. Membranes were incubated overnight (ON) with α-EhCP112 (1:5,000) or α-occludin (1:1,000) or α-claudin-1 (1:1,500) or α-claudin-2 (1:500) or α-claudin-4 (1:500) or α-ZO-1 (1:500) or α-ZO-2 (1:400) or α-Tub (1:3,000) antibodies. After washing, membranes were incubated for 1 h with the corresponding HRP-labeled secondary antibodies (1:10,000). Protein bands were visualized by chemiluminescence using the Pierce ECL-Plus kit (Thermo Fisher) and the MicroChemi system (Biostep). Densitometry analyses were performed using tubulin as loading control and employing the ImageJ software.

Confluent Caco-2 cells grown on coverslips were incubated with 10 μg/cm² of rEhCP112a during 2–30 min. After washing five times with PBS pH 6.8, cells were fixed with absolute ethanol for 30 min at −20°C, blocked with 0.5% BSA and incubated ON with α-EhCP112 (1:150) and α-claudin-1 (1:50) or α-claudin-2 (1:40) or α-occludin (1:100) or α-ZO-1 (1:50) antibodies. Cells were washed and incubated with specie-specific FITC- or TRITC-labeled antibodies (1:150) (Zymed). DAPI was added to stain nuclei and cells were mounted and analyzed by confocal microscopy as above.

Caco-2 cell monolayers were incubated with 10 μg/cm² of rEhCP112p for 10 min at 37°C and then, cells were lysed with RIPA/HO buffer (1:1 vol/vol), HO buffer was composed of 50 mM HEPES pH 7.5, 150 mM NaCl, 1 mM EGTA, 1.5 mM MgCl₂, 10% glycerol, 1% Triton X-100, and 1 mM Complete™ protease inhibitor cocktail and incubation was carried out for 20 min at 4°C. Then, lysates were incubated for 3 h with Protein G-Sepharose (Invitrogen) to clear unspecific bound proteins, washed with HO buffer and centrifuged at 500 × g for 2 min. Supernatants were incubated ON at 4°C with rabbit α-EhCP112 antibody or rabbit pre-immune serum. Next, samples were mixed with rProtein G-Sepharose for 3 h, washed five times with HO buffer, boiled with sample buffer and centrifuged at 15,300 × g for 15 min (Betanzos et al., 2013). Supernatants were analyzed by SDS-PAGE and WB, using α-EhCP112 and anti-TJ proteins antibodies.

Evan's blue-based in vivo colon permeability assays were performed as described (Lange et al., 1994; Citalán-Madrid et al., 2017). Pathogen-free C57/BL6 male mice (6–8 weeks old, ~25 g each, n = 6) were intraperitoneally anesthetized with 125 mg/kg of ketamine hydrochloride (Sanofi) and 12.5 mg/kg of xylazine (Phoenix Scientific) and then, rectally inoculated with rEhCP112a (50 μg in 300 μl of refolding buffer) or trophozoites (10⁶ cells in 300 μl) or with refolding buffer alone (300 μl) for 30 min. After laparotomy, a 22G polyethylene tube was inserted into the colon adjacent to the cecum and ligated. Then, remaining stool was rinsed out, and 1 ml of 1.5% Evan's blue dye (Sigma Aldrich) was instilled for 15 min. After washing with PBS until the perianal washout was clear, animals were euthanized and the colon extracted. The colon was longitudinally opened and rinsed again with PBS, followed by 1 ml of 6 mM N-acetylcysteine to remove dye within the mucus. Colons were incubated in 2 ml of formamide ON at RT and the extracted dye was spectrophotometrically measured at 610 nm. Values were expressed as arbitrary units per gram of tissue.

Colon of mice treated as above were scraped in the luminal side (Perreault and Beaulieu, 1996) and the epithelium layer was re-suspended in RIPA buffer with protease inhibitors (1 mM PMSF and 1 mM Complete™ protease inhibitor cocktail) for 30 min under continuous and vigorous shaking at 4°C. Tissue extracts were sonicated three times for 30 s, centrifuged for 15 min at 15,300 × g to eliminate undissolved cellular debris and analyzed by WB assays.

Mouse colon samples were placed in tissue freezing medium (Leica Biosystems) and snap-frozen in liquid nitrogen for 5 min. Frozen tissue sections (10 μm) were mounted in gelatin embedded coverslips, frozen at −70°C during 1 week, fixed with absolute ethanol for 30 min at −20°C and processed for immunofluorescence staining as above.

All data shown are representative of three independent experiments performed by duplicate unless stated otherwise. GraphPad Prism 5 software was used for statistical analysis. Data were analyzed by two tailed Student t-test. Statistical significance was assumed when ^*p < 0.05, ^**p < 0.01, or ^***p < 0.001. All results are displayed as mean with standard error.

The Centre for Research and Advanced Studies (CINVESTAV) fulfills the standard of the Mexican Official Norm (NOM-062-ZOO-1999) “Technical Specifications for the Care and Use of Laboratory Animals” based on the Guide for the Care and Use of Laboratory Animals “The Guide,” 2011, NRC, USA with the Federal Register Number BOO.02.03.02.01.908, awarded by the National Health Service, Food Safety and Quality (SENASICA) belong to the Animal Health Office of the Secretary of Agriculture, Livestock, Rural Development, Fisheries and Food (SAGARPA), an organization that verifies the state compliance of such NOM in Mexico. The Institutional Animal Care and Use Committee (IACUC/ethics committee) from CINVESTAV as the regulatory office for the approval of research protocols, involving the use of laboratory animals and in fulfillment of the Mexican Official Norm, has reviewed and approved all animal experiments (Protocol Number 0505-12, CICUAL 001).

E. histolytica is a phagocytic and cytolytic parasite in which several molecules contribute to its pathogenicity (Thibeaux et al., 2013). The EhCPs family is composed by 50 members in the genome (Que and Reed, 2000) (http://amoebadb.org) and some of them play significant roles during the intestine invasion by this parasite. First, we compared the enzymatic activity of the β-mercaptoethanol-activated recombinant proteases EhCP1, EhCP2, EhCP5, and EhCP112 (rEhCP1, rEhCP2, rEhCP5, and rEhCP112) on Z-Phe-Arg, a specific substrate for L-cathepsin proteases (Ocadiz et al., 2005; Meléndez-López et al., 2007; Quintas-Granados et al., 2009; Hou et al., 2010; Ocadiz-Ruiz et al., 2013; Cornick et al., 2016). All these enzymes degraded the substrate with different efficiency, including papain, the positive control, while BSA, the negative control showed no activity (Figure S1A). Subsequently, we tested the effect of the active enzymes on TEER of confluent Caco-2 cell monolayers. In these experiments, rEhCP112 exhibited the highest effect, dropping TEER by 65% after 90 min interaction with Caco-2 cells (Figure S1B).

To explore the molecular mechanism of EhCP112 action without interference by other parasite proteins, we used the purified precursor rEhCP112 (rEhCP112p) and the active protease (rEhCP112a). WB assays revealed the expected 52 and 35 kDa bands corresponding to the precursor and active forms, respectively (Figure 1A). Zymograms evidenced that only the active 35 kDa protein degraded gelatin (Figure 1A, inset). However, both forms degraded the Z-Phe-Arg substrate in a concentration dependent manner (Figure 1A), indicating that rEhCP112p was partially auto-activated after refolding (Ocadiz et al., 2005; Quintas-Granados et al., 2009). rEhCP112a presented higher activity in this assay. Then, we incubated Caco-2 cells with different rEhCP112a concentrations and measured TEER at different times. Results showed that the enzyme caused a drop in TEER in a time and dose-dependent manner (Figure 1B). Live trophozoites (10⁵/cm²; 2:1 cells to trophozoites ratio), used as control, decreased the TEER values by 35 and 80% after 2 and 30 min, respectively; similar to the damage produced by 30 μg/cm² of the enzyme (Figure 1B). Next, we explored whether TJs damage compromised the viability of Caco-2 cells and the integrity of the plasma membrane. TEER measurement revealed that cells in permanent contact with the protease presented a continuous drop in TEER, whereas cells in which rEhCP112a was removed, gradually recovered TEER (Figure 1B, arrow). Of note, a 30 min incubation with the active protease did not lead to cell death in Caco-2 cells (Figure 1C). These findings highlight that at this time and dose, the drop in TEER produced by the enzyme was reversible and did neither compromise membrane integrity nor cell viability. Incubation of rEhCP112a (10 μg/cm²) with the α-EhCP112 antibody or E-64 protease inhibitor, before being added to the Caco-2 cells, prevented the drop in TEER (Figure 1D), evidencing the specific effect of EhCP112 on TJs.

To further corroborate the EhCP112 effect on gate functions, we generated trophozoites stably overexpressing pNeoEhcp112. RT-PCR confirmed an increase of the ehcp112 transcript in pNeoEhcp112-transfected trophozoites in comparison with the control pNeo (Figure 2A). WB assays showed higher amounts of the different protease protein forms, including the precursor and active enzyme, in comparison to the control (Figures 2B,C). Confocal microscopy images also showed more EhCP112 in the cytoplasm (Figure 2D). EhCP112-overexpressing trophozoites caused a significantly higher drop in TEER in Caco-2 cells than trophozoites transfected with pNeo only. Antibodies against EhCP112 significantly prevented the TEER decrease (Figure 2E). The effect on TEER correlated with the increase of the enzyme in transfected trophozoites. However, to corroborate that the drop in TEER was not directly related to cellular damage at the time tested, we quantified cell viability and confirmed that 98% of Caco-2 cells were still viable after treatment with trophozoites (Figure 2F).

Gate function comprises control of both ion and macromolecular flux (Liang and Weber, 2014). To investigate whether rEhCP112a enzyme also disturbs the paracellular permeability for macromolecules, we performed TEM assays using RR dye, an electrodense marker. TEM images revealed that RR did not enter into the intercellular space in Caco-2 monolayers incubated with the active enzyme (Figure 3A). In addition, TEM images of cell monolayers in contact with trophozoites evidenced that RR was also restricted to the apical side of Caco-2 cells (Figure 3A). As expected, EDTA, used as positive control, allowed RR diffusion indicating opening of junctions (Gonzalez-Mariscal et al., 1985; Figure 3A). FITC-dextran (2–4 kDa) did not diffuse either through the intercellular space in Caco-2 cells treated with rEhCP112 or trophozoites (Figure 3B); further indicating that the macromolecular flux was not altered by the enzyme.

In contrast to our findings with Caco-2 cells, earlier studies demonstrated that RR penetrates through the paracellular space in MDCK cell monolayers incubated with live trophozoites (Martinez-Palomo et al., 1985). Thus, we evaluated the effect of rEhCP112 on MDCK cell monolayers. Our results showed that, in agreement to the results reported by Martinez-Palomo et al. (1985), but unlike our data in Caco-2 cells, macromolecular flux in MDCK cells increased after incubation with trophozoites (Figure 3C). By contrast, rEhCP112a had no effect on MDCK paracellular flux, but also increased ion flux as in Caco-2 cells (Figures 1B, 3C). These results show that E. histolytica trophozoites have differential effects on different types of epithelia. It is currently unknown what causes these discrepancies, but it is likely that differential expression of TJ proteins plays a role (Elkouby-Naor and Ben-Yosef, 2010; Lu et al., 2013).

In addition to the gate function, TJs retain plasma membrane molecules in either the apical or basolateral compartment, thus maintaining cell polarization (fence function; Zihni et al., 2016). When the fence function is disrupted, proteins and lipids that constitute the apical and basolateral membranes diffuse freely. To explore the effect of EhCP112 on fence function in Caco-2 cells, apical lipids were stained using fluorescently-labeled Bodipy FC-Sphingomyelin, incubated with rEhCP112a, and examined by confocal microscopy. We found that after 10 min, rEhCP112a caused diffusion of sphingomyelin to the basolateral membrane implying disruption of the fence function (Figure 4). Cells treated with the refolding buffer as negative control, indeed maintained the lipid at the apical membrane. Occludin did not shift significantly to the basolateral region after rEhCP112a treatment (Figure 4), strengthening the hypothesis that this TJ protein is not affected by EhCP112. These results indicate that rEhCP112 affects ion flux and fence functions of TJs.

Some pathogens provoke delocalization or degradation of TJ proteins, affecting gate and fence functions (Sousa et al., 2005; Hodges and Gill, 2010; Maia-Brigagao et al., 2012). Based on these findings, we analyzed the route that Alexa 647-labeled rEhCP112 follows after its contact with the apical side of Caco-2 cells. In these experiments we employed the precursor form of the enzyme, to slower the degradation process. Confocal microscopy revealed that after 5 s of interaction, rEhCP112p was already distributed along the apical cell surface (Figure 3A). This was more evident in zy-plane images (Figures 5A,B). At 30 s, rEhCP112p appeared concentrated in the apical region between two cells, but remained outside of the cells. After 1 min of incubation, fluorescent rEhCP112p was present within the intercellular space, indicating its affinity for the paracellular route. After 5 min, rEhCP112p still appeared along the lateral membrane and now some of the enzyme already entered the cells (Figures 5A,B). Moreover, we detected co-localization of the protease with occludin at TJ corroborating that EhCP112 indeed interacts with TJs (Figure 5C). We explored whether the inactivated enzyme was also able to bind to and penetrate into the cells, but only detected a very poor signal along the apical surface of Caco-2 cell monolayers suggesting that enzyme activity could be important for both binding to the apical surface and subsequent paracellular entry (Figure 5D). To confirm this, we recovered the supernatants of Caco-2 cells incubated with rEhCP112p and rEhCP112p pre-treated with E-64. WB assays showed that EhCP112p-E-64 remained in the supernatant, whereas EhCP112p amount was lower because the enzyme bound to Caco-2 cells (Figure 5E). These Results don't discard that the enzyme could also follow the transcellular route to penetrate the Caco-2 cell monolayers. Intriguingly, images showed the rEhCP112 entrance into the cell after 5 min of interaction (Figure 5A); however, IP staining (Figure 1C) revealed no membrane damage. This could be explained by the fact that EhCP112 is also internalized by caveolae and clathrin coated vesicles as shown recently by Hernández-Nava et al. (2017) in MDCK cell monolayers.

After knowing that the intercellular space is a pathway followed by EhCP112 to penetrate the epithelium, we studied the effect of the enzyme on the cellular location of TJ proteins until 30 min of incubation with rEhCP112a. Confluent Caco-2 cells were incubated with rEhCP112a for different times and then, by specific antibodies, the cellular location of claudin-1, claudin-2, occludin, and ZO-1 were analyzed through confocal microscopy. In the absence of EhCP112, the cellular proteins analyzed, except claudin-2, presented the typical honeycomb pattern already described for TJ proteins. However, after 2 min of incubation with the enzyme, claudin-1 moved to the cytoplasm and nucleus, co-localizing in some points with rEhCP112a (Figure 6A). The translocation of claudin-1 to the nucleus is documented in colon cells, after certain stimuli and metastatic transformation (Dhawan et al., 2005). In agreement with many reports on the scarce presence of claudin-2 in tight epithelia (Escaffit et al., 2005), it gave very faint signal in control confluent Caco-2 cells. Interestingly, at 2 min of incubation with EhCP112, as it has been well documented for sparse cell cultures (Escaffit et al., 2005), when the cells forming the monolayer began to loosen due to the effect of the rEhCP112a protease, fluorescence of claudin-2 increased at the cellular borders and in the cytoplasm, co-localizing in some points with the enzyme. Later, at 30 min, claudin-2 decreased again, probably due to the degradative effect of the enzyme (Figure 6B). In contrast, occludin and ZO-1 patterns appeared without changes after 30 min of interaction (Figures 6C,D), suggesting that EhCP112 has effect only on claudins. The unaltered occludin pattern is in concordance with the unaffected macromolecules flux after the enzyme treatment showed in Figure 5.

To explore the effect of rEhCP112a on TJ proteins integrity, confluent Caco-2 cells incubated with rEhCP112a were lysed and submitted to WB assays. The α-claudin-1 antibody recognized a 22 kDa band in control cells. After incubation with rEhCP112a, an 11 kDa band appeared and gradually increased upon time, suggesting degradation of claudin-1 (Figures 7A,B). To obtain more evidence on the claudin-1 degradation after contact with the protease, we treated rEhCP112a with E-64 inhibitor, before the enzyme addition to Caco-2-cells. In these experiments, the 11 kDa band did not show up, substantiating that it could be a degradation product due to the enzyme action (Figure 7C). However, the 22 kDa band maintained similar amount through the experiments, possibly due to a compensatory de novo synthesis of the protein, but this assumption needs to be experimentally proved. Similarly to confocal immunolocalization assays (Figure 6B), claudin-2 in confluent Caco-2 cell monolayers was scarce, but 2 min after incubation with rEhCP112a, when the protease has started to act on TJs, fluorescence corresponding to claudin-2 increased. At 30 min incubation the recognition of claudin-2 by the antibody, decreased (Figures 7A,B), probably due to the degradative effect of the continuous enzyme exposure. We confirmed by WB assays that claudin-2 is in higher amount in confluent Caco-2 cell cultures than in sparse cells that were not in contact with the enzyme (Figure 7D). rEhCP112a did not affect claudin-4, occludin, ZO-1, and ZO-2 (Figures 7A,B). Trophozoites degraded all proteins tested except claudin-4 and tubulin (Figures 7A,B), confirming that distinct trophozoites molecules produce dissimilar effect on host cell proteins. Intriguingly, trophozoites degraded occludin; although they did not affect paracellular flux in Caco-2 cells (Figure 3), possibly, claudin-4, a molecule that also regulates paracellular flux (Takizawa et al., 2014; Khan and Asif, 2015) performed a compensatory effect in this function, given that it was not affected by trophozoites.

The immunofluorescence experiments using Caco-2 cell monolayers and rEhCP112a showed that the enzyme affects claudin-1 and claudin-2, co-localizing with rEhCP112 and delocalizing them from their initial site in the cell. Therefore, to investigate the association of TJ proteins with the enzyme, we assessed immunoprecipitation assays employing the α-EhCP112 antibody. In these experiments, we used rEhCP112p to slow the proteolytic process; and, immediately after the immunoprecipitation with the antibody, we added protease inhibitors, to avoid proteolytic activity. In WB assays of the immunoprecipitates, the specific antibodies detected claudin-1 and claudin-2, but not claudin-4, occludin, ZO-1, and ZO-2 (Figure 8). Similarly to the immunofluorescence and protein degradation assays (Figures 6, 7), these data reinforce the assumption that claudin-1 and claudin-2 are binding targets for EhCP112 during the trophozoite attack. Again, recognition of claudin-2 in the immunoprecipitates was higher than in cells that were not in contact with the protease (Figure 8), suggesting that the interaction of the cells with rEhCP112p provoked an enrichment of this protein, in agreement with results obtained in Figures 6, 7.

The TJ gate function in MDCK and Caco-2 cells suffered differential impact after the trophozoites and EhCP112 attack (Figure 3), supporting the distinct susceptibility of epithelial cell lines to trophozoites molecules. In in vivo studies, we need to have in mind that the host immune response plays a double role, modulating the damage that the parasite and its molecules produce (Petri, 2008; Mortimer and Chadee, 2010) and generating defense cells and molecules that could contribute to the tissue injure. To gain further insight on the E. histolytica damage produced on animal models, we investigated the effect of rEhCP112a on C57/BL6 mice, a strain susceptible to the trophozoites colonization (Kissoon-Singh et al., 2013). In this model, we used Evan's blue tracer to determine whether rEhCP112a alters the intestinal permeability (Lange et al., 1994). Results showed that, while in control mice, treated only with the re-folding buffer no evident damage was observed in the tissues, but 30 min after inoculation of 50 μg of the enzyme resuspended in the refolding buffer, the Evan's blue tracer penetrated the mouse intestinal epithelium. Permeability of rEhCP112a-treated mice, increased twice, in comparison with refolding buffer-treated animals (Figure 9A). On the other hand, the epithelium of mice inoculated with 10⁶ trophozoites, absorbed twice Evan's blue than the rEhCP112a-treated animals, indicating that trophozoites had a higher effect on epithelial permeability.

Our results using Caco-2 cells evidenced important alterations of claudin-1 and claudin-2 due to the effect of rEhCP112a, thus, we analyzed the integrity of TJ proteins in the mouse colonic epithelium after inoculation with the protease. In agreement with results found in experiments with Caco-2 cells monolayers (Figure 7), WB assays using epithelia obtained from rEhCP112a-inoculated mice revealed that claudin-1 and claudin-2 were degraded, whereas, claudin-4, occludin, ZO-1, and ZO-2, as well as the control GAPDH, appeared without modification (Figures 9B,C). Trophozoites, had the same effect than rEhCP112a, but they also degraded ZO-2 and, in contrast with the Caco-2 cells, we did not observe degradation of occludin and ZO-1. In colonic epithelium, claudin-2 is restricted to the proliferative regions of the intestine, but is poorly expressed in differentiated cells (Escaffit et al., 2005). Here, we used colonic tissue samples containing both regions of the whole epithelium (Figures 9B,C).

To better examine the damage produced by the enzyme on the TJ proteins of intestinal epithelium, we performed confocal microscopy assays to localize claudin-1 and occludin in treated and untreated mice. Images of colon sections using α-EhCP112 and α-claudin-1 antibodies showed claudin-1 at the apical side, along the colonic epithelium of control mice. In concordance with the WB results, claudin-1 significantly decreased in rEhCP112a-treated mice and the enzyme appeared at the cellular borders and in the cytoplasm of epithelial cells (Figure 10). Occludin, localized also at cellular borders, was not altered after rEhCP112 incubation, accordingly to WB results (Figure 9B). E-cadherin, located along cellular borders of epithelial cells, below TJs, was used as a protein control damaged by rEhCP112 in intestinal epithelial cells (Hernández-Nava et al., 2017). In these experiments E-cadherin also diminished after incubation of intestinal epithelium with the protease (Figure 10).

Finally, after knowing that rEhCP112 was able to damage the intestinal epithelium, we wondered whether the enzyme was able to cross the mucosa layer that protects the intestinal epithelium. We performed confocal microscopy assays of the intestinal epithelium from rEhCP112a-treated mice, using α-mucin-2 and α-EhCP112 antibodies. Images showed that EhCP112 appeared alone in the luminal side of the intestine. Besides, α-EhCP112 antibodies also illuminated areas where mucin was abundant and decorated the epithelium, where goblet cells were present (Figure S2). Both proteins appeared close each other in the epithelium (Figure S2).

In conclusion, our results demonstrate that EhCP112 damages cultured epithelial cells and the mouse intestinal epithelium, disturbing TJs through a direct effect on claudin-1 and claudin-2. Our findings reveal a crucial role of EhCP112 enzyme during invasion of E. histolytica to the host cells and point out the potential of this molecule as a therapeutic target.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.