L. interrogans serovar Manilae strain UP-MMC-NIID (Koizumi and Watanabe, 2004) that was passaged no more than three times to maintain reproducible virulence (Toma et al., 2014), and L. biflexa serovar Patoc strain Patoc 1 were routinely stationary cultured in Ellinghausen-McCullough-Johnson-Harris (EMJH) broth at 30°C. L. interrogans was diluted 20-fold, and L. biflexa was diluted 40-fold in fresh EMJH and cultured for 3 days with shaking at 30°C for cell infection experiments.

RPTEC/TERT1 (American Type Culture Collection, ATCC^® CRL-4031™) cells were grown in Dulbecco’s Modified Eagle Medium/Nutrient Mixture F-12 Ham (DMEM/F-12 Ham; Sigma-Aldrich, St. Louis, MI, USA) supplemented with 5 pM triiodothyronine, 10 ng/mL recombinant human epidermal growth factor, 3.5 μg/mL ascorbic acid, 5 μg/mL transferrin, 5 μg/mL insulin, 8.65 ng/mL sodium selenite and 100 μg/mL G418. The cells were seeded at a density of 1 × 10⁶ cells/well in polyethylene terephthalate hanging cell culture inserts with a pore size of 3 μm (Falcon^®; Corning, New York, NY, USA) in the upper chamber of a Falcon^® Companion six-well tissue culture plate (Corning). Cells were maintained in a humidified incubator at 37°C and 5% CO₂ for 7 to 10 days after reaching confluence to allow monolayer maturation, and the medium was exchanged every 2 days. The transepithelial electrical resistance (TEER) of the monolayers was measured using a Millicell-ERS cell resistance indicator (MilliporeSigma, Burlington, MA, USA). After subtraction of the value of a cell-free filter (blank), the mean TEER value was expressed as Ωcm². The TEERs of cells before infection were designated as the baseline values. The percentage TEER, relative to the baseline value, was calculated using the following formula: (TEER of experimental wells/baseline TEER of experimental wells) ×100%.

The cells were infected with leptospires at a multiplicity of infection (MOI) of 100 from the basolateral side containing supplement-free DMEM/F-12 Ham medium. In the experiments involving the use of inhibitors, the inhibitors were added 30 min before infection. The inhibitors used were 30 μM dynasore (Sigma-Aldrich), 100 nM bafilomycin A1 (Bioviotica, Liestal, Switzerland), 50 μM chloroquine (Sigma-Aldrich), 10 μM MG-132 (Sigma-Aldrich), 200 nM bortezomib (FUJIFILM Wako, Osaka, Japan), 20 μM benzyloxycarbonyl-Val-Ala-Asp(OMe)-fluoromethylketone (Z-VAD-FMK, a pan-caspase inhibitor) (Tocris Bioscience, Bristol, UK), 20 μM benzyloxycarbonyl-Asp(OMe)-Glu(OMe)-Val-Asp(OMe)-fluoromethylketone (Z-DEVD-FMK, a caspase-3-specific inhibitor) (Tocris Bioscience), or 50 μM MDL-28170 (calpain inhibitor III) (Tokyo Chemical Industry, Tokyo, Japan). Cells were then incubated at 37°C and 5% CO₂ during infection and lysed for proteome analysis and immunoblotting, or fixed for proximity ligation assay (PLA), immunostaining, and electron microscopy analysis.

Liquid chromatography with tandem mass spectrometry (LC-MS/MS) analyses were performed using Nano LC (UltiMate^® 3000; Dionex, Sunnyvale, CA, USA) coupled with Q Exactive plus (Thermo Scientific, Waltham, MA, USA) (Yates et al., 1995). Instrument operation and data acquisition were performed using Xcalibur Software (version 3.1.66.1; Thermo Fisher Scientific, Waltham, MA, USA). Samples were reduced and alkylated as previously described (Peng et al., 2011). Proteins were diluted 3-fold with 50 mM ammonium bicarbonate (pH 8.5) and digested with trypsin at an enzyme-to-substrate ratio of 1:50 (wt/wt, modified sequencing grade, Promega, Madison, WI, USA) for 16 h at 30°C. Digested peptides were directly injected. The peptides were separated on a 500 × 0.075 mm capillary reversed-phase column (CERI, Tokyo, Japan) at a flow rate of 500 nL/min and a column temperature of 60°C. The mobile phase comprised water with 0.1% (v/v) formic acid (eluent A) and 80% acetonitrile with 0.1% (v/v) formic acid (eluent B). The peptides were separated by a linear gradient of up to 35% eluent B in 340 min for a 6 h gradient run.

The Q Exactive plus (Thermo Fisher Scientific) was operated in data-dependent (dd) mode with full scans acquired at a resolution of 35,000 at 350 m/z and with dd-MS/MS scans acquired at a resolution of 17,500. The mass spectrometer was operated in positive mode in the scan range of 350 –1,800 m/z. Fixed first m/z is 150 in dd-MS/MS scans. The 10 most abundant isotope patterns with a charge ≥2 from the survey scan were selected using an isolation window of 1.6 m/z. The maximum ion injection times for the full scan and the dd-MS/MS scans were 20 ms and 100 ms, respectively, and the automatic gain control for the full and the dd-MS/MS scans were 3E6 and 1E5, respectively. Repeat sequencing of peptides was kept to a minimum by the dynamic exclusion of the sequenced peptides for 20 s.

The database search was performed against the UniProt database of humans (Taxonomy ID:9606, released in June 2021) and NCBInr database of L. interrogans serovar Manilae (Taxonomy ID: 214675, released on 06/29/2021) using Proteome Discoverer version 2.5 (PD2.5, Thermo Fisher Scientific) with the MASCOT (Matrix Science, Boston, MA, USA) and Sequest (Thermo Fisher Scientific) search engine software (Mann and Wilm, 1994; Perkins et al., 1999; Shalit et al., 2015). Search parameters were as follows: fixed modifications, carbamidomethyl; variable modifications, oxidation (M); missed cleavages, up to 1; monoisotopic peptide tolerance, 10 ppm; and MS/MS tolerance, 0.6 Da. The PD2.5 search result files were uploaded into Scaffold (version 5.0.1, Proteome Software Inc., Portland, OR, USA) for protein identification and label-free quantitation based on an intensity-based absolute quantification value (Spinelli et al., 2012). The two LC-MS/MS technical replicates (n = 2) of each group were combined in Scaffold. The in-built normalization function and Student’s t-test was used to calculate the fold changes and p-values of protein abundances between two groups (i.e., non-infected group and infected group at 4 h or 24 h post-infection (p.i.)). Proteins that were detected with 99.0% probability (protein FDR = 0.2%) assigned by ProteinProphet (Keller et al., 2002) and containing at least two peptides that were detected with Mascot Threshold were considered positive identifications and quantified. The proteomics dataset has been deposited to the ProteomeXchange and jPOST (Okuda et al., 2017) with the dataset identifiers PXD040838 and JPST002081, respectively.

For the PLA ( Figure 1 ), non-infected or infected cells were methanol-fixed, permeabilized, and blocked with blocking buffer (5% bovine serum albumin (BSA), 1% Triton X-100 in Tris-buffered saline (TBS; 50 mM Tris and 150 mM NaCl (pH 7.4)) for 15 min. The antibodies used for analyses were rabbit monoclonal anti-Ecad (1:50; #3195, Cell Signaling Technology (CST) Danvers, MA, USA) and multi-epitope cocktail mouse monoclonal anti-p0071 (ready-to-use, #651166, ProGen, Baden-Wurttemberg, Germany). PLA was conducted using the DuoLink PLA kit (Sigma-Aldrich) with Detection Reagents Green by following the manufacturer’s protocol.

For initial analysis of infected cells, RPTECs were lysed with a mild lysis buffer: 20 mM Tris-HCl (pH 7.4), 150 mM NaCl, and 0.1% Nonident P-40, supplemented with a protease inhibitor cocktail (Roche, Basel, Switzerland). The cell lysates were collected with a cell scrapper for centrifugation, and supernatants were processed for immunoblotting ( Figure 2 ). For further analysis (all other immunoblots), cell lysates were obtained with RIPA buffer (Nacalai, Kyoto, Japan) [50 mM Tris-HCl buffer (pH7.6), 150 mM NaCl, 1% Nonidet P40, 0.5% sodium deoxycholate, and 0.1% sodium dodecyl sulphate (SDS)], supplemented with a protease inhibitor cocktail (Roche). Cell lysates obtained with RIPA buffer were analyzed without centrifugation to consider those proteins that are in the insoluble fraction owing to their association with the cytoskeletal proteins network (Sebastián et al., 2021). Cell lysates were mixed with Laemmli sample buffer (Laemmli, 1970) and heated for 10 min at 100°C. The RIPA cell lysates were sonicated for a total of 3 min (5s sonication with 5s intervals), and the samples were separated using 7.5% SDS-gel electrophoresis and further processed for immunoblots. Amersham ImageQuant 800 Imaging System (GE Healthcare, Chicago, IL, USA) was used to visualize the protein bands.

Primary antibodies used for immunoblotting included rabbit polyclonal anti-p0071 (1:1000, #A304-649AA, Bethyl Laboratories, Montgomery, TX, USA); mouse monoclonal antibodies: anti-p120-ctn (1:2500; sc-23872, Santa Cruz Biotechnology, Inc., Dallas, TX, USA), anti-α-catenin (1:1000; sc-9988) and anti-ubiquitin (1:5000, sc-8017); and rabbit monoclonal antibodies: anti-β-tubulin (1:1000; CST#2128), anti-E-cad (1:1000; CST#3195), anti-plakoglobin (1:1000; CST#7550), and anti-β-catenin (1:1000; CST#8480). The secondary antibodies were a horseradish peroxidase (HRP)-anti-rabbit antibody (1:10,000; #111-035-144, Jackson ImmunoResearch (JIR), West Grove, PA, USA) and HRP-anti-mouse antibody (1:10,000; JIR#715-005-150).

RPTECs were fixed in cold methanol for 15 min, permeabilized, and blocked with the same blocking buffer used for PLA for 15 min to analyze the localization of AJC proteins. For the immunostaining of the lysosomal-associated membrane protein-1 (LAMP-1) and for F-actin staining, infected cells were fixed with 2% paraformaldehyde in phosphate-buffered saline for 5 h at 4°C and washed twice with TBS. The filter membranes were detached from the hanging culture inserts before immunostaining.

Antibodies were diluted in TBS containing 1% BSA and 0.1% Triton X-100, except in LAMP-1 immunostainings wherein TBS containing 0.2% saponin and 10% Blocking One (Nacalai) was used as the antibody dilution buffer (Toma and Suzuki, 2020). The primary antibodies for bacteria were diluted rabbit polyclonal anti-L. interrogans antibody (1:500; kindly provided by Sharon YAM Villanueva, University of the Philippines) (Villanueva et al., 2014) and anti-L. biflexa antibody (1:500; Affinity BioReagents, Golden, CO, USA). RPTEC proteins were immunostained with multi-epitope cocktail mouse monoclonal anti-p0071 (ready-to-use, #651166, ProGen); mouse monoclonal antibodies: anti-p120-ctn (1:50; sc-373751), anti-LAMP1 (1:200; sc-20011), anti-acetylated tubulin (1:100; #T7451, Sigma-Aldrich) and, anti-E-cad (1:50; # 610182, BD Transduction Laboratories, Franklin Lakes, NJ, USA); rabbit monoclonal anti-E-cad (1:50; CST#3195) and anti-p120-ctn (1:300; CST#59854); and rabbit polyclonal anti-p0071 (1:1000, #A304-649AA, Bethyl Laboratories). F-actin was stained with a 1:500 dilution of rhodamine phalloidin (Abcam, Cambridge, UK). Cells were counterstained to label the DNA with a 1:100 dilution of TO-PRO™-3 (Invitrogen, Waltham, MA, USA). Secondary antibodies used for immunofluorescence analysis included anti-rabbit Alexa 488- (1:100; JIR#711-545-152), Cy3- (1:100; JIR#711-165-152), or Alexa 647-conjugated antibody (1:100, JIR#711-605-152); and anti-mouse Alexa 488- (1:100; JIR#715-545-151) or Cy3-conjugated antibody (1:100; JIR#715-165-151). The compiled Z-stack images were acquired using a Leica TCS-SPE confocal laser scanning microscope after mounting with SlowFade™ Diamond Antifade Mountant (Invitrogen) using LEICA LAS AF acquisition software (version 2.6.0.7266, Leica Microsystems CMS GmbH, Germany). Quantification of the fluorescence intensity of immunostained proteins at cell-cell junctions were measured from maximum projections of Z-stacks images using ImageJ software.

RPTECs were infected for 16 h and pre-fixed with 2.5% glutaraldehyde in phosphate buffer (0.1 M, pH 7.4), and then post-fixed with 2% osmium tetroxide and 1.5% potassium hexacyanoferrate(II) in water for 1 h. Specimens were stained overnight en bloc with a 2% (w/v) uranyl acetate aqueous solution, dehydrated with serially increasing concentrations of ethanol (30–100%), and embedded in epoxy resin according to standard protocol.

Volume microscopy was performed using a focused ion beam scanning electron microscope (Helios 650; FEI Company, Eindhoven, The Netherlands) (Kizilyaprak et al., 2019). Milling was performed with a gallium ion beam at 30 kV and 2.5 nA, and the imaging was done at 1.5 keV, 800 pA, 6144 × 4096-pixel frame size, 6 μs dwell time, with 70 Å pixel size and 200 Å section thickness (Kizilyaprak et al., 2015). After milling at 52°, the fresh surface was tilted to 90° and imaged normally to the electron beam (see accompanying video of Kizilyaprak et al., 2015). The images were aligned with IMOD (Kremer et al., 1996) using the “Align Serial Sections/Blend Montages” function.

Band intensities in immunoblots were analyzed using ImageJ software (version 1.53). Reacting protein bands were normalized with β-tubulin as a loading control, and the relative expression level of each protein was calculated by considering the ratio of analyzing protein/β-tubulin in non-infected RPTECs as one. Statistical significance was determined using an unpaired two-tailed Student t-test. All the data were presented as the mean of at least three independent determinations per experimental condition. Differences were considered significant at a p-value of < 0.05.

L. interrogans induces E-cad endocytosis, resulting in AJ disassembly. However, a transcriptomics approach revealed that AJ-associated proteins were not differentially expressed, suggesting that post-transcriptional events are critical during AJ disruption by L. interrogans (Sebastián et al., 2021). In this study, to understand the mechanism of AJ disruption, we analyzed the changes in protein profiles in response to L. interrogans infection via a proteome analysis of non-infected and infected RPTECs (dataset identifiers in the repositories are described in the Materials and Methods section). Comparing protein profiles at 4 h p.i. and 24 h p.i. indicated that 106 RPTEC proteins were decreased and 47 were increased with a fold change of <0.5 and >2, respectively ( Tables S1 , S2 ). Notably, a multifunctional protein coordinating cell adhesion with cytoskeletal organization, p0071 (also known as plakophilin-4) (Keil and Hatzfeld, 2014), was decreased 50 times at 24 h p.i. ( Table S1 ). We first focused on p0071 and analyzed whether it is associated with E-cad in RPTECs as reported in other cell types (Hofmann et al., 2009; Medvetz et al., 2012) using an in situ PLA. Non-infected RPTECs showed green spots surrounding the cells, representing p0071/E-cad interactions; these green spots decreased in L. interrogans-infected RPTECs ( Figure 1A ). Dual immunofluorescence indicated the co-localization of p0071 and E-cad in non-infected RPTECs which was decreased in L. interrogans-infected RPTECs ( Figure 1B ).

In a time course experiment, we quantitatively analyzed p0071 by conducting immunoblotting assays of non-infected and L. biflexa- or L. interrogans- infected RPTECs. p0071 protein levels gradually decreased in L. interrogans- and L. biflexa-infected RPTECs until 12 h p.i. However, in L biflexa-infected RPTECs, p0071 levels were recovered to basal levels at 24 h p.i. In contrast, in L. interrogans-infected RPTECs, p0071 continuously decreased during the course of infection and was significantly decreased at 24 h p.i. (~ 30% of L. biflexa-infected RPTECs, p <0.01) ( Figures 2A, B ). Immunofluorescence analysis showed that p0071 was localized at AJs in non-infected and L. biflexa-infected RPTECs but it was displaced from cell-cell junctions in L. interrogans-infected RPTECs at 24 h p.i. ( Figures 2C–E ). Leptospiral immunostaining showed that the bacterial population interacting with epithelial cells was higher in L. interrogans-infected RPTECs than in L. biflexa-infected RPTECs ( Figure 2D ). The monolayer integrity was disrupted in L. interrogans-infected RPTECs as indicated by the evaluation of TEER, which reduced to ~25% of the initial TEER ( Figure 2F ). These results suggest that the p0071 decrease is associated with L. interrogans pathogenicity related to AJ disassembly.

Comparison of proteomes at 4 h p.i. and 24 h p.i. showed that, in addition to p0071, three ctns with armadillo domains were decreased at 24 h p.i.: p120-ctn, β-ctn (scaffold proteins of the AJ), and plakoglobin (a desmosomal scaffold protein), with a decrease of 10, 5, and 3 times, respectively ( Table S1 ). Among the AJ proteins, E-cad and α-ctn (a protein that links the AJ to the cytoskeleton) also showed a decrease of 10 and 3 times, respectively ( Table S1 ). For a more accurate quantitative analysis, we compared these proteins levels in total cellular lysates of L. biflexa- and L. interrogans-infected RPTECs at 4 h and 24 h p.i. with proteins levels of non-infected RPTECs. The three independent experiments showed that p120-ctn was significantly decreased at 24 h p.i. in L. interrogans-infected RPTECs (p <0.05), whereas E-cad, β-ctn, α-ctn, and plakoglobin were not decreased ( Figures 3A, B ). Therefore, we ruled out the possibility of E-cad-Hakai-dependent ubiquitination and proteasomal degradation as a trigger event of the L. interrogans-induced AJ disassembly (Hartsock and Nelson, 2012). Immunofluorescence analysis showed that L. interrogans infection induced significant p120-ctn and E-cad displacement from the AJs (p <0.01; Figures 3C, D ). Dual immunofluorescence analysis of p0071 and plakoblogin at 24 h p.i. showed that, in L. interrogans-infected RPTECs, plakoglobin was not significantly displaced from the cell-cell junctions (p >0.05; Figures S1A, B ). These results suggest that L interrogans specifically induces mislocalization and decrease in the levels of the p120-ctn sub-family proteins (p0071 and p120-ctn) that interact with the cytoplasmic E-cad JMD to displace the AJ from the plasma membrane.

We hypothesized that p0071 and p120-ctn degradation may be the trigger and upstream event of the clathrin-mediated endocytosis of E-cad found in our previous study (Sebastián et al., 2021). Dynasore-pre-treated RPTECs were infected with L. biflexa or L. interrogans and total cell lysates were analyzed via immunoblotting to evaluate this hypothesis. Inhibition of endocytosis with dynasore significantly prevented p0071 and p120-catenin degradation (p <0.01) ( Figures 4A, B ). Dynasore treatment did not affect localization of p0071, p120-ctn, and E-cad in non-infected RPTECs ( Figure S2 ). Subsequently, we analyzed the effect of dynasore pre-treatment in infected RPTECs by immunofluorescence and found that dynasore prevented the displacement of p0071 and p120-ctn from the AJs ( Figures 4C, D ). These results suggest that p0071 and p120-ctn degradation does not trigger AJ disassembly and an endocytic event precedes p0071 and p120-ctn degradation.

To better understand the changes in the dynamics of p120-ctn subfamily proteins and E-cad at AJs, we performed immunofluorescence analysis of L. interrogans-infected RPTECs at 12 h and 24 h pi. Quantification of fluorescence intensity at cell-cell junctions showed that p0071, p120-ctn, and E-cad intensities significantly decreased at 12 h p.i. (p <0.01 for p0071, and p <0.05 for p120-ctn and E-cad), and these proteins were displaced continuously from the AJs ( Figures S3A–C ). We hypothesized that L. interrogans infection of RPTECs induces stressful stimuli that enhance E-cad endocytosis and p0071 and p120-ctn degradation through the lysosomal pathway and/or ubiquitinin-proteasomal system (UPS) (Flores-Maldonado et al., 2017). We then analyzed whether p0071 displaced from the AJs was associated with the endo-lysosomal pathway using a dual immunofluorescence staining to detect p0071 and the LAMP-1. AJ-disassociated p0071 were found in LAMP-1-positive-vacuoles, cytoplasmic LAMP-1-negative compartments, and nuclear compartments ( Figure S4A ).

Next, we analyzed the effect of pre-treating RPTECs with the lysosomal inhibitors [bafilomycin A (BF) or chloroquine (CQ)] or the proteasomal inhibitors [MG-132 (MG) or bortezomib (BZ)] before infection. All inhibitors did not significantly prevent degradation of p120-ctn family proteins (p >0.05) ( Figures S4B, C ). Notably, fragments reacting in p120-ctn immunoblots (between ~58 kDa and ~88 kDa) were detected when the proteasome was inhibited, suggesting that these fragments are p120-ctn-derived and may be targets of the UPS degradation pathway ( Figure S4B ). The immunoblots also suggested that limited proteolysis may occur upstream of the UPS since these putative UPS targets were smaller than the full-length (FL) protein. TEER decrease during L. interrogans infection of RPTECs (~20% of initial TEER) was prevented with the proteasomal inhibitors (~140% or ~95% of initial TEER with MG and BZ, respectively), but not with the lysosomal inhibitors ( Figure S4D ).

Compensation mechanisms between degradative pathways have been reported because other systems must clear cellular materials that accumulate after inhibiting one degradative system to maintain homeostasis (Kocaturk and Gozuacik, 2018). Thus, we hypothesized that some ubiquitinated proteins accumulated after adding MG and BZ ( Figure S4E ) may be translocated to the lysosomal pathway for degradation. We then analyzed whether combining lysosomal and proteasomal inhibitors can prevent p0071 and p120-ctn degradation and found that such combination did not significantly prevent p0071 degradation (p >0.05) ( Figures 5A, B ). In contrast, FL-p120-ctn degradation tended to be suppressed by combining inhibitors and was significantly prevented by the combination of CQ+BZ (p <0.05) ( Figures 5A, B ). The L. interrogans-induced TEER decrease at 18 h p.i. (~17% of initial TEER) was prevented in infected RPTECs pretreated with a combination of BF+MG (~115% of initial TEER) and CQ+MG (~134% of initial TEER) ( Figure 5C ). The effects of BZ combined with either BF or CQ were not additive in terms of inhibiting the TEER decrease. Therefore, we used the BF+MG combination for further analysis using two different immunofluorescence protocols to detect total p120-ctn ( Figure 5D ) or AJ-associated p120-ctn ( Figures 5E, F ). Confocal images and quantification of fluorescence intensity at cell-cell junctions showed that p120-ctn accumulated in LAMP-1-positive vacuoles and also localized at AJs in BF+MG-treated RPTECs ( Figures 5D , S5 ). The combination of BF+MG significantly prevented p120-ctn and E-cad displacement from the AJs (p <0.01) but not the displacement of p0071 ( Figures 5E–G ). The prevention of AJ disassembly without preventing p0071 degradation and displacement from AJs shows the redundant role of p0071 and p120-ctn as the AJ scaffold proteins. These results suggested that additional factors are involved in p0071 proteolysis.

Some bacterial pathogens, such as Helicobacter pylori (Hatakeyama and Brzozowski, 2006), can induce AJ disassembly by altering epithelial cell polarization. Therefore, we tested the possibility that L. interrogans also affects RPTEC polarization. Acetylated tubulin, which is a marker of polarized epithelial cells, and F-actin organization in non-infected RPTECs ( Figure 6A ) were compared with RPTECs infected with L. interrogans in the absence or presence of BF+MG. L. interrogans induced F-actin rearrangement; however, acetylated tubulin was not affected ( Figure 6B ). This cytoskeletal disorganization was prevented by BF+MG inhibitors ( Figure 6C ). Leptospires were observed at the apical side in DMSO-treated RPTECs, while they were observed as dots near the basal side in BF+MG-treated RPTECs ( Figures 6B, C ).

We performed focused ion beam-scanning electron microscopy (FIB/SEM) analysis, which has high image and spatial resolution, to further understand the effect of BF+MG during L. interrogans infection of RPTECs. AJ disruption by L. interrogans in the absence of inhibitors induced the appearance of large gaps within the cell monolayer. These intercellular spaces were filled with leptospires which were observed from the basal to the apical side ( Figures 7A–D and Video S1 ). Some adhesion-compromised RPTECs were apoptotic which might represent cells that extrude from the monolayer ( Figure 7A ) (Grieve and Rabouille, 2014). The x, y, and z view of infected RPTECs showed the large spaces between the cells, which were absent in the BF+MG-treated RPTECs ( Figures 7D , 8A ). In contrast, in BF+MG-treated cells, a continuous tight cell monolayer was observed ( Figure 8B and Video S2 ). Leptospires were located proximally to the basal side within intracellular compartments that were found to be LAMP-1-positive by immunofluorescence ( Figures 8A , S6 ) or were located extracellularly in a tightly controlled intercellular space ( Figures 8B, C ). Leptospires transmigrating from the basal side through the pore of the culture insert membrane which was filled with RPTEC-cytoplasm were also observed ( Figures 8D, E and Video S2 ). These results indicate that L. interrogans does not affect epithelial cell polarization, and BF+MG inhibitors can prevent the AJ disassembly, F-actin rearrangement, and monolayer destruction induced by L. interrogans.

To test the possibility that p0071 is undergoing limited proteolysis during L. interrogans infection, we evaluated the effect of pre-treating RPTECs with the pan-caspase inhibitor Z-VAD-FMK and the calpain inhibitor MDL28170 before infection. The TEER decrease during L. interrogans RPTEC infection (~20% of initial TEER) was partially prevented by Z-VAD-FMK (~50% of initial TEER), but not by MDL28170 (~20% of initial TEER) ( Figure 9A ). Z-VAD-FMK significantly prevented p0071 degradation (p <0.05) ( Figures 9B, C ), displacement of p120-ctn subfamily proteins and E-cad from the AJs ( Figures 9D–F ), and partially prevented the F-actin rearrangement induced by L. interrogans ( Figure S7 ), Since L. interrogans induces a caspase-3-dependent cell death in renal epithelial cells in animal models (Marinho et al., 2015), the effects of pre-treating RPTECs before infection with Z-DEVD-FMK were also evaluated to rule out the possibility that the inhibitory effects of Z-VAD-FMK are caspase-3-dependent events. No inhibitory effects were detected in L. interrogans-induced TEER decrease, p0071 degradation, and AJ-proteins mislocalization ( Figures 9A–F ). These results suggested that proteolysis by a protease inhibited by Z-VAD-FMK is involved in p0071 degradation and displacement of p120-ctn subfamily proteins from the AJs ( Figure 10 ).