Cysteine protease 30 (CP30) contributes to adhesion and cytopathogenicity in feline Tritrichomonas foetus

Accepted Manuscript Title: Cysteine protease 30 (CP30) contributes to adhesion and cytopathogenicity in feline Tritrichomonas foetus Authors: Emily N. Gould, Richard Giannone, Stephen A. Kania, M. Katherine Tolbert PII: DOI: Reference:

S0304-4017(17)30334-5 http://dx.doi.org/doi:10.1016/j.vetpar.2017.07.034 VETPAR 8430

To appear in:

Veterinary Parasitology

Received date: Revised date: Accepted date:

22-5-2017 29-7-2017 31-7-2017

Please cite this article as: Gould, Emily N., Giannone, Richard, Kania, Stephen A., Tolbert, M.Katherine, Cysteine protease 30 (CP30) contributes to adhesion and cytopathogenicity in feline Tritrichomonas foetus.Veterinary Parasitology http://dx.doi.org/10.1016/j.vetpar.2017.07.034 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Cysteine protease 30 (CP30) contributes to adhesion and cytopathogenicity in feline Tritrichomonas foetus

Emily N. Gould1, Richard Giannone2, Stephen A. Kania1, M. Katherine Tolbert1*

University of Tennessee College of Veterinary Medicine, Department of Small Animal Clinical Sciences, Knoxville, TN

1. The University of Tennessee College of Veterinary Medicine, Knoxville, TN; Department of Small Animal Clinical Sciences (Gould, Tolbert) and Biomedical and Diagnostic Sciences (Kania).

2. Oak Ridge National Laboratory, Oak Ridge, TN.

*Corresponding author: M. Katherine Tolbert DVM PhD DACVIM ([email protected]); Department of Small Animal Clinical Sciences, 2407 River Dr., University of Tennessee College of Veterinary Medicine, Knoxville, TN. Tel.:+1 865 974 8387; fax: +1 865 974 5554.

Highlights



Bovine T. foetus (TF) cysteine protease 30 (CP30) is identified in feline TF.



CP30 is present in feline TF and non-pathogenic Pentatrichomonas hominis (PH).



Inhibition of CP30 reduces adhesion and cytotoxicity to host epithelium in vitro.



Confocal microscopy shows differences in CP30 localization between TF and PH.



CP30 represents a novel target for treatment of feline T. foetus.

Abstract Tritrichomonas foetus (T. foetus) is a flagellated protozoan parasite that is recognized as a significant cause of diarrhea in domestic cats with a prevalence rate as high as 30%. No drugs have been shown to consistently eliminate T. foetus infection in all cats. Cysteine proteases (CPs) have been identified as mediators of T. foetus-induced adhesion-dependent cytotoxicity to the intestinal epithelium. These CPs represent novel targets for the treatment of feline trichomonosis. However, cats also produce CPs that are part of life-critical systems. Thus, parasitic CPs need to be selectively targeted to reduce the potential for host toxicity. Previous studies have demonstrated the importance of a specific CP, CP30, in mediating bovine and human trichomonad cytopathogenicity. This CP has also recently been identified in feline T. foetus, although the function of this protease in the feline genotype remains unknown. Therefore, the study objectives were to characterize the presence of CP30 in feline T. foetus isolates and to evaluate the effect of targeted inhibition of CP30 on feline T. foetus-induced adhesion dependent cytotoxicity. The presence of CP30 in feline T. foetus isolates was identified by In gel zymography and proteomic analysis, indirect immunofluorescence (IF), and flow cytometry using a rabbit polyclonal antibody that targets bovine T. foetus CP30 (α-CP30). The effect of inhibition of CP30 activity on T. foetus adhesion and cytotoxicity was determined using CFSE-labeled feline T. foetus and crystal violet spectrophotometric assays in a previously validated co-culture model. CP30 expression was confirmed in all feline T. foetus isolates tested by all assays. Targeted inhibition of feline T. foetus CP30 resulted in decreased T. foetus adhesion to and cytotoxicity towards IPEC-J2 monolayers compared to rabbit IgG-treated T. foetus isolates. These studies establish that CP30 is expressed by feline T. foetus isolates and may be an important virulence factor in the cytopathogenicity of feline T. foetus. The results of these studies provide strong

evidence-based justification for investigation of CP30 as a novel target for the treatment of feline trichomonosis.

Abbreviations CP30 Cysteine protease 30 IPEC-J2 Porcine intestinal epithelial cell CFSE Carboxyfluorescein succinimidyl ester Dil 1,1'-dioctadecyl-3,3,3'3'-tetramethylindocarbocyanine perchlorate Keywords: Cat; tritrichomonas foetus; pentatrichomonas hominis; confocal, microscopy; protease

1. Introduction Tritrichomonas foetus (T. foetus) is a flagellated protozoal pathogen responsible for chronic diarrhea in domestic cats. Feline trichomonosis is associated with waxing and waning colitis characterized by foul smelling, mucoid to hemorrhagic diarrhea. Although T. foetus is recognized to have up to a 30% prevalence and global distribution (Gookin et al., 2004), it still poses a therapeutic challenge for veterinarians. Ronidazole, a 5-nitroimidazole drug, is the only therapy recognized to clear infection in affected cats (Gookin et al., 2006; Upcroft et al., 2006). However, drug safety and the recognition of ronidazole-resistant strains of T. foetus (Gookin et al., 2010; LeVine et al., 2011; Rosado et al., 2007; Xenoulis et al., 2013) dictate the need for investigation of T. foetus virulence factors and development of targeted therapies for feline trichomonosis. Cellular proteases such as cysteine and serine proteases are identified virulence factors for pathogenic protozoa including Trichomonas vaginalis (Mendoza-Lopez et al., 2000), bovine T. foetus (Singh et al., 2004), Entamoeba histolytica (Lauwaet et al., 2003), Toxoplasma gondii

(Dowse et al., 2008; Pszenny et al., 2012), and Leishmania spp (de Araujo Soares et al., 2003; Mahmoudzadeh-Niknam and McKerrow, 2004). We previously showed that feline T. foetus possess cell-associated cysteine proteases (CP) that promote adhesion of the parasite to the intestinal epithelium. Broad-spectrum inhibition of feline T. foetus cysteine protease (CP) activity ameliorated T. foetus-induced adhesion-dependent cytotoxicity towards the intestinal epithelium. (Tolbert et al., 2014) Moreover, targeted inhibition of specific cysteine proteases largely prevented feline T. foetus-induced intestinal epithelial cell destruction, suggesting that specific CPs might play a larger role than others in mediating host cellular cytopathogenicity. (Tolbert et al., 2016) In bovine venereal trichomonosis, cysteine protease 30, (CP30), also known as CP8, is one of the most abundant CPs (Mallinson et al., 1995) and mediates adhesion-dependent cytotoxicity towards urogenital epithelial cells. (Mendoza-Lopez et al., 2000; Singh et al., 2005; Singh et al., 2004). A recent proteomic study investigating CPs in both feline and bovine T. foetus identified CP30 as the most abundant CP, independent of genotype (Stroud et al., 2017). Despite the recognized role of CP30 as a potent virulence factor in venereal trichomonosis, no studies have evaluated the function of CP30 in feline trichomonosis. This information is instrumental in the targeting of feline CP30 as a novel therapeutic. Therefore, the objectives of this study were to evaluate the localization of CP30 in isolates of feline T. foetus and determine the function of CP30 in promoting feline T. foetus cytopathogenicity. We also evaluated for the presence of CP30 in feline P. hominis, a nonpathogenic intestinal trichomonad. We hypothesized that CP30 would be conserved across feline T. foetus isolates but absent in P. hominis, and that CP30 would promote adhesion and cytotoxicity of feline T. foetus towards the intestinal epithelium. For these reasons, we also proposed that it represents a novel therapeutic target for the treatment of feline trichomonosis.

In these studies, we demonstrate that CP30 isconserved in feline T. foetus isolates and present in P. hominis isintracellularly located in both pathogenic feline T. foetus and nonpathogenic P. hominis, Targeted inhibition of CP30 decreases feline T. foetus-induced host cellular pathogenicity. Moreover, CP30 was not identified in host intestinal epithelial cells used in these assays, which suggests that CP30 may be a safe protease to selectively inhibit. These findings support further investigation of CP30 as a novel target for the treatment of feline trichomonosis.

2. Experimental Methods and Design 2.1 T. foetus and P. hominis isolates: Five (A, F, Ja, JT, Sti) different isolates of feline T. foetus (to account for strain variability), including one isolate acquired from a ronidazole-treated cat (JT), one bovine T. foetus isolate (positive control known to express CP30), and one feline P. hominis isolate (negative control; presumptively nonpathogenic trichomonad that has been demonstrated to lack CP activity (Tolbert et al., 2014) were used. Isolates of T. foetus and P. hominis were obtained from the feces of naturally infected cats. The bovine T. foetus isolate 40204, originally obtained from a naturally infected bull, was generously donated from Auburn University. Trichomonads were cultivated at 37°C in modified Diamond’s media with 10% inactivated equine serum, penicillin, and streptomycin as previously described (Tolbert et al., 2014). Only mid-log phase trichomonads that had been passaged no greater than 10 times since cryopreservation were used for all assays. Similarly, only trichomonads with >95% viability as assessed by motility were utilized. 2.2 IPEC-J2 cells: The porcine jejunal epithelial cell line (IPEC-J2), a non-transformed primary cell line originally isolated from neonatal piglet jejunum, was cultured as previously described (Tolbert et al., 2016;

Tolbert et al., 2014; Tolbert et al., 2013). Briefly, IPEC-J2 monolayers were cultivated in culture media that included Advanced Dulbecco’s minimal essential medium: Nutrient Mixture F-12 (DMEM/F12) supplemented with 5 g/ml each of insulin, transferrin, and selenium, EGF (5 ng/ml), penicillin (50,000 IU/ml), streptomycin (50,000 mg/ml), and 5% fetal bovine serum and incubated at 37°C in 5% CO2. Prior to co-culture studies with feline T. foetus isolates, IPEC-J2 were seeded on 24-well polystyrene plates and allowed to reach confluence as previously described. Immediately prior to infection studies, culture media was replaced with co-culture media, which contained the same constituents as culture media but was devoid of serum to prevent replication of trichomonads and epithelial cells during infection studies. Uninfected IPEC-J2 monolayers and monolayers infected with unlabeled rabbit IgG (isotype control; Southern Biotech, Birmingham, AL)-treated T. foetus were used as negative and positive controls, respectively, for all co-culture assays. IPEC-J2 cells were used at passage numbers 47-60. 2.3 Anti-CP30 antibody: For all immunoassays, rabbit serum containing a polyclonal antibody directed against T. vaginalis CP30 was used. Antibody was isolated from the serum of rabbits previously immunized with the 30 kDa proteinase (Mendoza-Lopez et al., 2000). CP30 has been demonstrated to have high homology between T. vaginalis and bovine T. foetus and the antibody cross-reacts with bovine CP30 (Singh et al., 2004). Rabbit serum containing α-CP30 antibody was quantified according to manufacturers’ instructions using a commercial enzyme-linked immunosorbent assay (ELISA) (Abcam, Cambridge, MA) for rabbit IgG. All antibodies were stored as aliquots at -20°C to prevent antibody degradation from repetitive freeze-thaw cycles. 2.4 In gel zymography assay:

Whole organism trichomonads were sonicated and quantified with a bicinchoninic acid (BCA) assay according to manufacturer’s instructions (Thermo Scientific, Rockford, IL). For purposes of demonstrating cysteine protease activity in feline T. foetus protein lysate, the cysteine protease inhibitor, E64 (300 µM), was applied to live trichomonads (isolate F) for 15 min at 37°C immediately prior to protein extraction, electrophoresis, and In gel zymography as previously described (Tolbert et al., 2014). Zones of proteolysis were detected by the visualization of clear bands (i.e. areas where gelatin was digested through protease activity) against a stained background (i.e. areas where protease activity was not present). 2.5 Gel-band identification via LC-MS/MS-based shotgun proteomics: In gel zymography bands were processed for liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis in order to identify their constituent proteins. Gel bands corresponding to areas of protease enzymatic activity were excised from the Coomassie stained gel and in-gel tryptic digestion performed as previously described (Shevchenko et al., 2006). Tryptic peptides obtained from processed gel bands were acidified and filtered through a 10 kDa MWCO spin filter (Vivaspin 500, GE Healthcare) and analyzed by LC-MS/MS analysis using a LTQ-Orbitrap Pro Mass Spectrometer (Thermo Scientific) coupled with a u3000 HPLC (Dionex). Briefly, 50 ul of band-derived trypic peptides were loaded onto a 150 micron inner diameter (ID) MudPIT back column packed with 5 cm strong-cation exchange chromatographic resin (Luna SCX; Phenomenex). Following a 10 min wash with solvent A (95% LC-MS-grade H2O, 5% acetonitrile, 0.1% formic acid), the back column was placed in-line with an in-house pulled, 100 micron ID nanospray emitter packed with 15 cm of reversed-phase (RP) resin (Kinetex C18; Phenomenex). Positively-charged peptides were moved from the SCX back column to emitter with a 5 min pulse of 500 mM ammonium acetate (in solvent A). Peptides were then eluted from the C18 packed

emitted via organic gradient to 0-100% solvent B (30% LC-MS-grade H2O, 70% acetonitrile, 0.1% formic acid) over 120 min. Eluting peptides underwent nanospray ionization (NSI) directly into the LTQ-Orbitrap MS whereby high accuracy peptide mass determination followed by fragmentation analysis (MS/MS) allowed for robust and accurate (< 10 ppm) peptide sequence determination via protein database search described below. 2.6 Genomic and Proteomic Analysis: Peptide fragmentation spectra generated from the LC-MS/MS analysis were searched against the T. foetus expressed sequence tag database (Morin-Adeline et al., 2014) (appended with E. coli strain K12, common contaminants, and reversed/decoy protein sequences to serve as distractors) with Myrimatch protein identification software (Tabb et al., 2007). Peptide-spectrum matches (PSM) were filtered by false-discovery rate (< 1% at the peptide- and protein-level) and assembled to proteins based on the information available from the feline T. foetus EST database using IDPicker v 3.0. (Ma et al., 2009). Protein abundance was derived from MS1/parent peptide chromatographic area-under-the-curve and compared across all processed gel bands. 2.7 Indirect immunofluorescence: Immunofluorescence (IF) was used as a qualitative assay to confirm the conservation of CP30 across feline T. foetus isolates prior to co-localization, adhesion, and cytotoxicity assays. Mid-log phase whole organism trichomonads were pelleted at 1500 x g for 5 minutes, washed in 1X PBS, and re-suspended to 1 x 107 trichomonads/mL. A hydrophobic barrier pen (Vector Laboratories, Burlingame, CA) was used to create barriers on slides pre-treated for electrostatic adherence (Thermo Fisher Scientific, Rockford, IL). 3 x 106 trichomonads were applied via direct pipetting to areas denoted by the barrier pen and allowed to adhere to slides at room temperature for one hour. All samples were fixed in acetone at -20°C for 15 min and then allowed to air dry for three

minutes. Slides were rinsed briefly in 1X PBS for five minutes and incubated for one hr at room temperature (RT; 23°C) in block buffer (1X PBS, 5% goat serum and 2% BSA). Slides were then incubated in either primary α-CP30 antibody diluted 1:100 in block buffer or an equivalent concentration of unlabeled rabbit IgG for three hours in a humidified chamber at RT. Following incubation with primary antibody or isotype control, slides were rinsed thrice in 1X PBS for five minutes followed by incubation in fluorescein isothiocyanate (FITC)-conjugated F(ab’)2 goat antirabbit IgG (secondary antibody; Jackson Immuno Research, West Grove, PA) diluted 1:50 in block buffer for one hour at RT. Immunofluorescence was also performed on IPEC-J2 cells adhered to 24-well polystyrene plates. With the exception of the initial adherence steps, primary and secondary antibody were added in the exact same fashion. Following three rinses with 1X PBS for five minutes each, the nuclear counterstain, DAPI (i.e. 4’6-diamidino-2phenylindole) (Vectashield, Vector Laboratories, Burlingame, CA) was applied for visualization of either epithelial and trichomonad nuclei. A mercury epifluorescence microscope (Nikon® Digital Sight DS, Melville, NY) was used for detection of fluorescence. 2.8 Flow cytometry: Mid-log phase trichomonads, cultured for 48 hr, were centrifuged at 1500 x g for five min, washed once in 1X PBS (pH 7.4) and re-suspended to 5 x 106 trichomonads/mL in either sodium azide or 10% buffered formalin. Isolates were divided into 1 mL aliquots for the following samples: T. foetus receiving both primary α-CP30 antibody and secondary antibody, T. foetus receiving an equivalent concentration of isotype control (i.e. unlabeled rabbit IgG), T. foetus receiving only secondary antibody (negative control) and T. foetus receiving no antibodies (autofluorescence). Conical tubes (15 mL) preserved the shape of mid-log phase trichomonads and were used for all flow assays. Isolates were divided into 1.0 mL aliquots for each treatment group, pelleted, and 10

µL of primary α-CP30 antibody or equivalent amount of rabbit IgG was added to the pellet for a final dilution of 1:100 per mL of re-suspended trichomonads. All samples were incubated on ice for 30 minutes and washed once with 1.0 mL of 1X PBS. Fluorescein isothiocyanate (FITC)conjugated anti-rabbit IgG was applied at the same concentration of 1:100 directly to the pellet, incubated for 30 min on ice, washed once more, and re-suspended to a final volume of 1.0 mL for analysis. A total of 10,000 events (i.e. trichomonads) were analyzed for each treatment population, and each isolate was analyzed in triplicate. All data were acquired on an Attune® acoustic focusing flow cytometer (Applied Biosystems, Thermo Fisher Scientific, Rockford, IL). 2.9 Confocal microscopy and co-localization assays: Mid-log phase protozoa were prepared exactly as described above for IF, but 0.17 mm coverslips (Thermo Fisher Scientific, Rockford, IL) designed for confocal microscopy were used. A Leica SP8 epifluorescence confocal microscope (Leica Microsystems Inc., Buffalo Grove, IL) was utilized for acquisition of all images. Co-localization assays were performed using the membrane surface marker dye, Dil (1,1'-dioctadecyl-3,3,3'3'-tetramethylindocarbocyanine perchlorate; Thermo Fisher Scientific, Rockford, IL), at a concentration of 0.5 mg/uL and the lysosomal/intracellular marker, LysoTrackerRed DND99 dye (Thermo Fisher Scientific, Rockford, IL) at a final concentration of 1.0 uM. All co-localization assays were performed identical to IF assays, but following incubation with secondary antibody, each slide was incubated with the respective co-localization dye for 30 minutes in a dark, humidified chamber prior to staining with DAPI. For both feline T. foetus (isolate A) and P. hominis, the effect of the presence of the intestinal epithelium on CP30 expression was evaluated with co-localization assays performed before and after exposure to secreted products of IPEC-J2 cells. IPEC-J2 monolayers were grown to confluence in 24-well plates. 20 x 106 mid-log phase trichomonads/mL of co-culture

media were separated from confluent IPEC-J2 cells by superimposition of 3.0 um pore culture polyester plate inserts (Corning™ permeable inserts, Tewksbury, MA) and incubated for 16 hours. Co-localization assays using 0.5 mg/mL and 1.0 mM concentrations of Dil and LysoTrackerRed, respectively, were performed immediately prior to, and following co-culture in the transwell inserts. These concentrations were used for all trichomonads. 2.10

CSFE adhesion assays:

To determine if CP30 promotes adhesion of feline T. foetus to the intestinal epithelium, adhesion assays were performed using uninfected monolayers and monolayers infected with unlabeled rabbit IgG or α-CP30-treated T. foetus. Carboxyfluorescein succinimidyl ester (CFSE) (CellTrace™ CFSE cell proliferation kit, Thermo Fisher Scientific, Rockford, IL) was used to label viable trichomonads as previously described (Tolbert et al., 2013). IPEC-J2 monolayers were cultivated to confluence in 24-well polystyrene plates prior to infection with T. foetus. IPEC-J2 monolayers were inoculated with 10 x 106 T. foetus per well, with all trichomonad groups pretreated with either α-CP30 (1:100-1:500) or an equivalent concentration of unlabeled rabbit IgG prior to infection. A third group containing only (1:100-1:500) α-CP30 but lacking T. foetus was used as additional control to evaluate for non-specific binding of primary antibody to the IPEC-J2 cells. Infected and uninfected groups of IPEC-J2 cells were maintained at 37°C in 5% CO2. After six hours of co-culture, an established time for adhesion based on a previously validated co-culture model (Tolbert et al., 2013), the cells were washed gently with 37°C DPBS twice and counterstained with DAPI. Adherent trichomonads in individual wells were counted in six high power fields (HPFs) using an epifluorescence microscope (Nikon® Digital Sight DS, Melville, NY), with the average of six HPFs representative of one replicate. All assays were performed with

a minimum of four replicate cultures per treatment group and repeated in triplicate experiments. Mean or median adherence were compared among groups. 2.11

Crystal violet cytotoxicity assays:

To provide a quantitative analysis of the effect of CP30 on T. foetus-induced epithelial cytotoxicity, crystal violet (CV) assays were performed as previously described (Tolbert et al., 2013) on uninfected monolayers and monolayers infected with rabbit IgG-treated or α-CP30 antibody treated T. foetus. IPEC-J2 monolayers were cultivated to confluence on 24-well polystyrene plates prior to co-culture with 10 x 106 T. foetus for 24 hrs. Prior to co-culture, trichomonads were treated with α-CP30 antibody or rabbit IgG. After 24 hours of co-culture, monolayers were washed and stained with CV as previously described (Tolbert et al., 2014). Monolayers were solubilized in 1% sodium dodecyl sulfate in 50% ethanol and the staining intensity of the solubilized monolayers was measured by a spectrophotometer (570nm wavelength). Assays were performed in a minimum of six replicate cultures and repeated in triplicate experiments. Mean cytotoxicity was compared among experimental groups (T. foetus with or without α-CP30 or rabbit IgG treatment) and uninfected monolayers. 2.12

Statistical Analysis:

Raw data was interpreted first for normality with a Shapiro Wilk W Statistic and equal variance with a Brown-Forsythe followed by analysis with either parametric or non-parametric statistics where appropriate using SigmaStat 13 (Jandell Scientific). Either a Student’s T-test or Wilcoxon Rank Sum was used to determine differences between groups in all adhesion assays with an analysis of variance (ANOVA) used to evaluate for differences among groups for cytotoxicity. For all analysis, P < 0.05 was considered significant.

Descriptive data were generated for

immunofluorescence, flow cytometry, proteomic analysis, and confocal microscopy assays.

3. Results 3.1 Feline T. foetus demonstrates CP30 activity: To confirm the presence of CP30 in feline T. foetus , In gel zymography followed by proteomic analysis using mass spectrometric was performed (Tolbert et al., 2014). A high abundance of sequences consistent with conserved CP30 were identified in the clear zones of proteolysis caused by lower molecular weight proteases corresponding to CPs (Fig 1). 3.2 CP30 is intracellularly located in feline T. foetus and P. hominis isolates, but absent in intestinal epithelial cells: Indirect immunofluorescence and flow cytometry were used as qualitative and semi-quantitative assays, respectively, to confirm the presence of CP30 in feline T. foetus (Figs 2 and 3) and P. hominis (Fig 2) isolates. Immunofluorescence was also used to evaluate for the presence of CP30 in IPEC-J2 cells (Fig 4). All feline T. foetus isolates tested were determined to express CP30. Differentiation between surface and intracellular localization was made following the use of two different fixatives. No fluorescence was detected when a sodium azide fixative, which fixes cells without allowing intracellular permeabilization of antibodies, was applied prior to application of α-CP30 antibody. In contrast, all T. foetus isolates were positive for fluorescence with α-CP30 following the use of 10% buffered formalin as a fixative. Unlike sodium azide based fixatives, formalin acts as a simultaneous permeabilizing agent and allows antibodies to traverse cellular membranes, confirming the intracellular location of CP30 in both the bovine and feline T. foetus isolates. The mean fluorescence following application of α-CP30 exceeded that of any other control groups, confirming the presence of CP30 in both bovine and feline T. foetus (Fig 3). The mean fluorescence of the negative and isotype control groups were equivalent. P. hominis also

displayed positive fluorescence for CP30 via indirect immunofluorescence (Fig 2) and flow cytometry. In contrast, IPEC-J2 cells lacked positive fluorescence following application of α-CP30. (Fig 4), DPBS control, or isotype control (unlabeled rabbit IgG). 3.3 Differential localization of CP30 in feline T. foetus and Pentatrichomonas hominis following exposure to IPEC-J2 cells: Immunofluorescence and flow cytometry demonstrated the presence of CP30 in both feline T. foetus and P. hominis, however, CP30 is an active enzyme in T. foetus alone (Tolbert et al, 2014). Confocal microscopy was used to interrogate whether there are differences in cellular localization (i.e. membrane versus intracellular) of CP30 between the two trichomonads that might help to explain the lack of activity of CP30 in P. hominis. Whereas the distribution of CP30 in T. foetus was diffuse (Fig 5A), P. hominis had a more localized, segmented pattern of CP30 distribution (Fig 5B). Both T. foetus and P. hominis CP30 co-localized to intracellular and membrane locations, with a predominant membrane localization of enzyme for both trichomonads (Fig 6A,D). Coculture and co-localization assays were performed to further explore changes in localization of CP30 between trichomonads, such as differences in ability to switch from a predominantly intracellular to extracellular localization after exposure to IPEC-J2 cells. Following exposure to secreted products of IPEC-J2 cells, the membrane distribution of CP30 in feline T. foetus consistently increased (Fig 6B,C), whereas membrane co-localization in P. hominis was subjectively more heterogeneous following co-culture (Fig 6E,F). 3.4 Feline T. foetus CP30 contributes to adhesion and exertion of cytotoxicity towards intestinal epithelial monolayers:

To investigate the role of CP30 in promoting adhesion and cytotoxicity of feline T. foetus to the intestinal epithelium, feline T. foetus were treated with α-CP30 prior to co-culture with IPEC-J2 monolayers. Inhibition of T. foetus CP30 had a dose-dependent effect on T. foetus adhesion. Groups treated with concentrations at or above 1:100 α-CP30 had significantly fewer numbers of T. foetus adhered to IPEC-J2 cells when compared to rabbit IgG-treated T. foetus (P<0.05) (Fig 7). Groups treated with lower concentrations of α-CP30 (i.e. 1:250, 1:500) had no significant differences in adhesion compared to isotype-control groups (Fig 7). Cytotoxicity was also reduced with inhibition of feline T. foetus CP30 activity. α-CP30-treated T. foetus-infected IPEC-J2 monolayers showed significantly less cytotoxicity (P < 0.001) as compared to rabbit-IgG treated T. foetus-infected monolayers following 24 hours of co-culture (Fig 8).

4. Discussion Despite differences in organ tropism, feline Tritrichomonas foetus utilizes similar virulence factors as the urogenital pathogens, Trichomonas vaginalis and bovine Tritrichomonas foetus. For example, cysteine proteases are potent virulence factors in vitro for both intestinal and urogenital trichomonosis. Broad-spectrum inhibition of cysteine proteases would likely result in host toxicity, thus a targeted approach must be employed. As inhibition of specific cysteine proteases, such as CP30 and CP65 (Alvarez-Sanchez et al., 2000), have already been shown to significantly ameliorate cytopathogenicity in both human and bovine venereal trichomonosis (Arroyo and Alderete, 1995; Singh et al., 2005), evaluation of specific cysteine proteases as potential targets for treatment of intestinal trichomonosis was warranted. Our first objective was to determine which specific CP should be targeted in order to have the maximal effect on amelioration of T. foetus-induced intestinal cytopathogenicity. We have

previously demonstrated that feline T. foetus CPs are lower molecular weight proteases. Thus, we employed In gel zymography to isolate proteases. We excised and digested the high and lower molecular weight proteases and pursued tandem mass spectrometry and genomic and proteomic analyses to survey the cysteine proteases present. Using these techniques, we identified that a sequence of T. vaginalis and bovine T. foetus CP30, also known as CP8, was present and highly abundant in feline T. foetus in the lower molecular weight proteases, which is where we have previously demonstrated the highest area of CP activity. This is in agreement with studies that have identified CP30 as highly abundant in both bovine (Mallinson et al., 1995) and feline (Stroud et al., 2017) T. foetus isolates, as well as an important virulence factor in bovine trichomonosis (Singh et al., 2005). Using immunofluorescence and flow cytometric analyses, we confirmed that CP30 was conserved across the feline T. foetus isolates utilized in our assays, as has been shown for bovine (Mallinson et al., 1995; Singh et al., 2005; Thomford et al., 1996) and, more recently, feline T. foetus (Stroud et al., 2017) Our next objective was to evaluate the role that feline T. foetus CP30 plays in adhesion to and cytotoxicity towards host intestinal epithelial cells. While induction of cytotoxicity is multifactorial, direct contact with the host cell is a critical first step towards exertion of trichomonad cytotoxicity (Alderete and Garza, 1988) (Singh et al., 1999) (Tolbert et al., 2014). As feline T. foetus cysteine proteases help to promote adhesion-dependent cytotoxicity towards intestinal epithelial cells in vitro (Tolbert et al., 2014) and CP30, specifically, facilitates T. vaginalis attachment to human vaginal epithelial cells (Arroyo and Alderete, 1989, 1995; Mendoza-Lopez et al., 2000), we hypothesized that CP30 played a role in feline T. foetus adhesion and cytotoxicity to the intestinal epithelium. Indeed, targeted inhibition CP30 significantly reduced feline T. foetus adhesion to and cytotoxicity towards intestinal epithelial monolayers in a dose-

dependent manner. While it is likely that other proteases and virulence factors play a role in cytotoxicity as complete abolishment of host cell destruction was not observed, inhibition of CP30 represents a promising new adjunct therapeutic for feline trichomonosis. When searching for a novel therapeutic target, considering the potential for host cell toxicity is also an important factor. Although pathogenic organisms utilize CPs to exert damage to their hosts, mammalian cells also normally express and rely upon these proteases to maintain normal metabolic functions such as apoptosis, antigen presentation, processing of proenzymes and hormones and regulation of normal cellular turnover (Dubin, 2005; Rzychon et al., 2004). This places an emphasis on identifying a therapeutic target that specifically targets individual CPs utilized by the parasite and not host cells. For this reason, we evaluated for the presence of CP30 in IPEC-J2 monolayers. Porcine epithelial cells were used as a model for feline intestinal epithelial cells given that no in vitro feline intestinal epithelial cell lines currently exist. Nonetheless, this is an appropriate substitute for feline intestinal epithelial cells given that the porcine trichomonad, Tritrichomonas suis, has a similar tropism for the gastrointestinal tract, is genetically similar to feline T. foetus (Slapeta et al., 2012; Tachezy et al., 2002) and feline T. foetus isolates consistently cause destruction of these cells (Tolbert et al., 2014; Tolbert et al., 2013). Using immunofluorescence, we determined that IPEC-J2 do not possess CP30. While further studies are necessary to confirm these findings, the presence and activity of CP30 in multiple feline isolates, as well as absence in IPEC-J2 cells, makes CP30 an attractive target for development of a novel therapeutic. This work also evaluated CP30 in the nonpathogenic feline intestinal trichomonad, P. hominis. Our earlier studies demonstrated that, unlike feline T. foetus, P. hominis attaches poorly to the intestinal epithelium (Tolbert et al., 2013). The reason for this difference is unknown. Studies

comparing pathogenic Entamoeba histolytica with the non-pathogenic species, E. dispar, have found that differences in metallopeptidase expression might contribute to differences in pathogenicity. (Meyer et al., 2016) We have previously demonstrated that P. hominis does not maintain active CP enzymes (Tolbert et al., 2014). Lack of active CPs might contribute to P. hominis being avirulent. In the current study, we sought to determine if differences in cellular localization of CP30 both prior to and after exposure to the intestinal epithelium might contribute to lack of CP30 enzymatic activity in P. hominis. It is possible that exposure to host epithelial cells stimulates an upregulation or increased expression of CP30 in T. foetus but not P. hominis. This work did not evaluate for quantitative differences in CP30 before and after exposure to IPEC-J2 cells, but functional differences (e.g. changes in upregulation or activation) in CPs between species might provide insight into how T. foetus CP30 exerts cytotoxicity. Further study to determine if the differences noted between surface and intracellular localizations of CP30 in T. foetus versus P. hominis contribute to differences in cytopathogenicity is warranted. In conclusion, this study was the first to identify that CP30 is present in and conserved across multiple feline T. foetus isolates. It also affirmed the role of feline CP30 as similar to that of human T. vaginalis and bovine T. foetus because targeted inhibition of this protease resulted in significantly decreased adhesion and cytotoxicity to in vitro intestinal epithelial cells. CP30 represents an attractive novel therapeutic target for feline trichomonosis. Acknowledgments This project was supported by a Miller Trust Grant from the Winn Feline Foundation (Tolbert MT15-005). The contents of this publication are solely the responsibility of the authors and do not necessarily represent the views of Winn. We would like to thank Drs. Bibhuti Singh (Upstate Medical University), Jody Gookin (North Caroline State University), and Chance

Armstrong (Auburn University) for their generous donation of antibodies and trichomonad isolates. We would also like to acknowledge John Dunlap (University of Tennessee) and Jonathan Wall (University of Tennessee Medical Center) for their assistance in acquisition of immunofluorescence and CFSE-adhesion images, as well as Dianne Trent, Mabre Brand, and Rachel Dickson (University of Tennessee) for their assistance with culture and flow cytometry assays.

References

Alderete, J.F., Garza, G.E., 1988. Identification and properties of Trichomonas vaginalis proteins involved in cytadherence. Infection and immunity 56, 28-33. Alvarez-Sanchez, M.E., Avila-Gonzalez, L., Becerril-Garcia, C., Fattel-Facenda, L.V., Ortega-Lopez, J., Arroyo, R., 2000. A novel cysteine proteinase (CP65) of Trichomonas vaginalis involved in cytotoxicity. Microbial pathogenesis 28, 193-202. Arroyo, R., Alderete, J.F., 1989. Trichomonas vaginalis surface proteinase activity is necessary for parasite adherence to epithelial cells. Infection and immunity 57, 2991-2997. Arroyo, R., Alderete, J.F., 1995. Two Trichomonas vaginalis surface proteinases bind to host epithelial cells and are related to levels of cytoadherence and cytotoxicity. Archives of medical research 26, 279-285. de Araujo Soares, R.M., dos Santos, A.L., Bonaldo, M.C., de Andrade, A.F., Alviano, C.S., Angluster, J., Goldenberg, S., 2003. Leishmania (Leishmania) amazonensis: differential expression of proteinases and cell-surface polypeptides in avirulent and virulent promastigotes. Experimental parasitology 104, 104-112. Dowse, T.J., Koussis, K., Blackman, M.J., Soldati-Favre, D., 2008. Roles of proteases during invasion and egress by Plasmodium and Toxoplasma. Sub-cellular biochemistry 47, 121-139. Dubin, G., 2005. Proteinaceous cysteine protease inhibitors. Cellular and molecular life sciences : CMLS 62, 653-669. Gookin, J.L., Copple, C.N., Papich, M.G., Poore, M.F., Stauffer, S.H., Birkenheuer, A.J., Twedt, D.C., Levy, M.G., 2006. Efficacy of ronidazole for treatment of feline Tritrichomonas foetus infection. Journal of Veterinary Internal Medicine 20, 536-543. Gookin, J.L., Stauffer, S.H., Dybas, D., Cannon, D.H., 2010. Documentation of in vivo and in vitro aerobic resistance of feline Tritrichomonas foetus isolates to ronidazole. Journal of Veterinary Internal Medicine 24, 1003-1007. Gookin, J.L., Stebbins, M.E., Hunt, E., Burlone, K., Fulton, M., Hochel, R., Talaat, M., Poore, M., Levy, M.G., 2004. Prevalence of and risk factors for feline Tritrichomonas foetus and giardia infection. Journal of clinical microbiology 42, 2707-2710. Lauwaet, T., Oliveira, M.J., Callewaert, B., De Bruyne, G., Saelens, X., Ankri, S., Vandenabeele, P., Mirelman, D., Mareel, M., Leroy, A., 2003. Proteolysis of enteric cell villin by Entamoeba histolytica cysteine proteinases. The Journal of biological chemistry 278, 22650-22656. LeVine, D.N., Papich, M.G., Gookin, J.L., Davidson, G.S., Davis, J.L., Hayes, R.B., 2011. Ronidazole pharmacokinetics after intravenous and oral immediate-release capsule administration in healthy cats. Journal of feline medicine and surgery 13, 244-250. Ma, Z.Q., Dasari, S., Chambers, M.C., Litton, M.D., Sobecki, S.M., Zimmerman, L.J., Halvey, P.J., Schilling, B., Drake, P.M., Gibson, B.W., Tabb, D.L., 2009. IDPicker 2.0: Improved protein assembly with high discrimination peptide identification filtering. Journal of proteome research 8, 3872-3881. Mahmoudzadeh-Niknam, H., McKerrow, J.H., 2004. Leishmania tropica: cysteine proteases are essential for growth and pathogenicity. Experimental parasitology 106, 158-163. Mallinson, D.J., Livingstone, J., Appleton, K.M., Lees, S.J., Coombs, G.H., North, M.J., 1995. Multiple cysteine proteinases of the pathogenic protozoon Tritrichomonas foetus: identification of seven diverse and differentially expressed genes. Microbiology (Reading, England) 141 ( Pt 12), 30773085.

Mendoza-Lopez, M.R., Becerril-Garcia, C., Fattel-Facenda, L.V., Avila-Gonzalez, L., Ruiz-Tachiquin, M.E., Ortega-Lopez, J., Arroyo, R., 2000. CP30, a cysteine proteinase involved in Trichomonas vaginalis cytoadherence. Infection and immunity 68, 4907-4912. Meyer, M., Fehling, H., Matthiesen, J., Lorenzen, S., Schuldt, K., Bernin, H., Zaruba, M., Lender, C., Ernst, T., Ittrich, H., Roeder, T., Tannich, E., Lotter, H., Bruchhaus, I., 2016. Overexpression of Differentially Expressed Genes Identified in Non-pathogenic and Pathogenic Entamoeba histolytica Clones Allow Identification of New Pathogenicity Factors Involved in Amoebic Liver Abscess Formation. PLoS pathogens 12, e1005853. Morin-Adeline, V., Lomas, R., D, O.M., Stack, C., Conesa, A., Lapeta, J., 2014. Comparative transcriptomics reveals striking similarities between the bovine and feline isolates of Tritrichomonas foetus: consequences for in silico drug-target identification. BMC genomics 15, 955. Pszenny, V., Davis, P.H., Zhou, X.W., Hunter, C.A., Carruthers, V.B., Roos, D.S., 2012. Targeted disruption of Toxoplasma gondii serine protease inhibitor 1 increases bradyzoite cyst formation in vitro and parasite tissue burden in mice. Infection and immunity 80, 1156-1165. Rosado, T.W., Specht, A., Marks, S.L., 2007. Neurotoxicosis in 4 cats receiving ronidazole. Journal of Veterinary Internal Medicine 21, 328-331. Rzychon, M., Chmiel, D., Stec-Niemczyk, J., 2004. Modes of inhibition of cysteine proteases. Acta biochimica Polonica 51, 861-873. Shevchenko, A., Tomas, H., Havlis, J., Olsen, J.V., Mann, M., 2006. In-gel digestion for mass spectrometric characterization of proteins and proteomes. Nature protocols 1, 2856-2860. Singh, B.N., Hayes, G.R., Lucas, J.J., Beach, D.H., Gilbert, R.O., 2005. In vitro cytopathic effects of a cysteine protease of Tritrichomonas foetus on cultured bovine uterine epithelial cells. American journal of veterinary research 66, 1181-1186. Singh, B.N., Lucas, J.J., Beach, D.H., Shin, S.T., Gilbert, R.O., 1999. Adhesion of Tritrichomonas foetus to bovine vaginal epithelial cells. Infection and immunity 67, 3847-3854. Singh, B.N., Lucas, J.J., Hayes, G.R., Kumar, I., Beach, D.H., Frajblat, M., Gilbert, R.O., Sommer, U., Costello, C.E., 2004. Tritrichomonas foetus induces apoptotic cell death in bovine vaginal epithelial cells. Infection and immunity 72, 4151-4158. Slapeta, J., Muller, N., Stack, C.M., Walker, G., Lew-Tabor, A., Tachezy, J., Frey, C.F., 2012. Comparative analysis of Tritrichomonas foetus (Riedmuller, 1928) cat genotype, T. foetus (Riedmuller, 1928) cattle genotype and Tritrichomonas suis (Davaine, 1875) at 10 DNA loci. International journal for parasitology 42, 1143-1149. Stroud, L.J., Slapeta, J., Padula, M.P., Druery, D., Tsiotsioras, G., Coorssen, J.R., Stack, C.M., 2017. Comparative proteomic analysis of two pathogenic Tritrichomonas foetus genotypes: there is more to the proteome than meets the eye. International journal for parasitology 47, 203-213. Tabb, D.L., Fernando, C.G., Chambers, M.C., 2007. MyriMatch: highly accurate tandem mass spectral peptide identification by multivariate hypergeometric analysis. Journal of proteome research 6, 654-661. Tachezy, J., Tachezy, R., Hampl, V., Sedinova, M., Vanacova, S., Vrlik, M., Van Ranst, M., Flegr, J., Kuldaa, J., 2002. Cattle pathogen tritrichomonas foetus (Riedmuller, 1928) and pig commensal Tritrichomonas suis (Gruby & Delafond, 1843) belong to the same species. The Journal of eukaryotic microbiology 49, 154-163. Thomford, J.W., Talbot, J.A., Ikeda, J.S., Corbeil, L.B., 1996. Characterization of extracellular proteinases of Tritrichomonas foetus. J Parasitol 82, 112-117. Tolbert, M.K., Brand, M.D., Gould, E.N., 2016. In vitro effects of cysteine protease inhibitors on Trichomonas foetus-induced cytopathic changes in porcine intestinal epithelial cells. American journal of veterinary research 77, 890-897.

Tolbert, M.K., Stauffer, S.H., Brand, M.D., Gookin, J.L., 2014. Cysteine protease activity of feline Tritrichomonas foetus promotes adhesion-dependent cytotoxicity to intestinal epithelial cells. Infection and immunity 82, 2851-2859. Tolbert, M.K., Stauffer, S.H., Gookin, J.L., 2013. Feline Tritrichomonas foetus adhere to intestinal epithelium by receptor–ligand-dependent mechanisms. Veterinary parasitology 192, 75-82. Upcroft, J.A., Dunn, L.A., Wright, J.M., Benakli, K., Upcroft, P., Vanelle, P., 2006. 5-Nitroimidazole drugs effective against metronidazole-resistant Trichomonas vaginalis and Giardia duodenalis. Antimicrobial agents and chemotherapy 50, 344-347. Xenoulis, P.G., Lopinski, D.J., Read, S.A., Suchodolski, J.S., Steiner, J.M., 2013. Intestinal Tritrichomonas foetus infection in cats: a retrospective study of 104 cases. Journal of feline medicine and surgery 15, 1098-1103.

Figure 1. Low molecular weight feline T. foetus CP is identified as consistent with bovine T. foetus and human T. vaginalis cysteine protease 30. Fig 1A depicts a representative In gel zymography of cellular protein lysate from feline T. foetus (isolate F). Lane 1 represents vehicletreated protein lysate while lane 2 represents treatment with E64, a broad-spectrum CP inhibitor. Note the abolishment of the low molecular weight zone of proteolysis in the E64-treated lysate. The specific zone of proteolysis on In gel zymography is identified via tandem mass spectrometry as consistent with bovine and human trichomonad CP30 (Fig 1B). The Y-axis represents the relative protein abundance of CP30 in area under the curve in the higher molecular weight proteases (white bar) and lower molecular weight proteases (black bar) seen in Fig 1A.

Figure 2. Indirect immunofluorescence confirms the presence of CP30 in feline and bovine T. foetus isolates, as well as one feline Pentatrichomonas hominis (P. hominis) isolate. Mid-log phase feline T. foetus (isolate JT, 2A), bovine T. foetus (2B) and feline P. hominis (2C) displayed positive fluorescence following incubation with 1:100 anti-CP30 and FITC- conjugated goat anti-rabbit IgG. Isolate JT displayed negative fluorescence following incubation with unlabeled rabbit IgG and FITC-conjugated secondary only (2D). All images acquired at 400X. A total of three feline isolates (A, Ja and JT) were evaluated via indirect immunofluorescence.

Figure 3. Both feline and bovine T. foetus display positive fluorescence for CP30 via flow cytometry. Figures 3A and B display positive fluorescence of either feline isolate A (3A) or bovine isolate 2040 (3B) following application of 1:100 α-CP30 polyclonal antibody and 1:50 FITC-conjugated secondary antibody. The Y-axis represents the average trichomonad count for a particular population, and the X-axis represents (moving from left to right) the mean fluorescence on a logarithmic scale of the isolate treated either with no antibody (auto control), equivalent concentrations of isotype control (rabbit IgG) or 1:100 anti-CP30 antibody. 10,000 events were counted for each treatment group. A total of four feline isolates (A, F, Ja and Sti) were evaluated via flow cytometry.

Figure 4. Indirect immunofluorescence confirms the absence of CP30 in in vitro porcine intestinal epithelial cells (IPEC-J2). IPEC-J2 cells lack green fluorescence following incubation with either 1:100 α-CP30 and FITC- conjugated secondary (4A) or isotype control rabbit IgG and FITC-conjugated secondary (4B). All images acquired at 400X.

Figure 5. The distribution of CP30 is shown in a feline T. foetus isolate (A) and a feline Pentatrichomonas hominis (P. hominis) isolate. Mid logarithmic phase feline T. foetus (5A) and feline P. hominis (5B) display positive fluorescence but slightly different distribution patterns of CP30 following application of 1:100 anti-CP30 and 1:100 FITC-conjugated goat anti-rabbit IgG. Images were acquired at 620X.

Figure 6. Membrane co-localization of CP30 differs in feline T. foetus and Pentatrichomonas hominis (P. hominis) following exposure to IPEC-J2 monolayers. Yellow intensity represents colocalization of CP30 and membrane marker Dil [(1,1'-dioctadecyl-3,3,3'3'tetramethylindocarbocyanine perchlorate). Figures 6A and D depict localization of CP30 (conjugated to a FITC-green fluorescent secondary antibody) in both feline isolate A (6A) and feline P. hominis (6D) prior to exposure to epithelial cells. Figures 6B,C and 6E,F depict the change in yellow intensity, representing surface localization of CP30, in both feline isolate A (6B,C) and feline P. hominis (6E,F) following co-culture with in vitro IPEC-J2 cells.

Figure 7. CP30 promotes adhesion of feline trichomonads to IPEC-J2 cells following 6 hours of co-culture and is concentration dependent. The median number of adhered trichomonads to IPEC-J2 monolayers after co-culture with either 1:100, 1:250 or 1:500 concentration of respective treatment group and 10 x 106 feline T. foetus (isolate A) treated with either an equivalent volume of isotype control (white bar; unlabeled mouse IgG) or polyclonal antibody (black bar; α-CP30). Data represent n=5-8 cultures per treatment and are reported as median + SE. ** P < 0.01 between groups. All comparisons between groups determined by a MannWhitney Rank Sum Test. All assays were performed in triplicate.

Figure 8. Cysteine protease 30 promotes cytotoxicity of feline trichomonads to IPEC-J2 cells following 24 hours of co-culture. The average crystal violet absorbance of IPEC-J2 monolayers following co-culture with either just DPBS (uninfected monolayers; white bar) or 10 x 106 feline T. foetus (isolate A) treated with either an equivalent volume of isotype control (gray bar; unlabeled rabbit IgG) or 1:100 polyclonal antibody (black bar; α-CP30). Data represent n=8 cultures per treatment and are reported as mean + SD. ** P < 0.001, *** P<0.0001 between groups. Determined by a one way ANOVA and Holm-Sidak post hoc. All assays were performed in triplicate.