Giardia lamblia trophozoites assemblage A (strain WB C6; ATCC No. 30957) and assemblage B (strain GS H7; ATCC 50581) were cultivated using modified TYI-S-33 medium supplemented with 10% adult bovine serum and 0.5 mg/mL bovine bile (Diamond et al., 1978; Keister, 1983). The antibiotic piperacillin (100 µg/mL) was added during routine culturing of the parasite (Gillin et al., 1989). Growth was initiated by adding ∼10⁵ trophozoites/mL in the culture medium and continued to grow until the cells became 80–90% confluent (∼48 h). To produce cysts, trophozoites from WB isolates were cultured in high bile medium, as previously described by Keister (1983). The trophozoites and cysts forms of human Giardia-isolate H3 (assemblage B), collected from gerbils, were purchased from Waterborne Inc. (New Orleans, LA), and HT29 cells lines were obtained from the ATCC.

LRs from Giardia trophozoites and cysts were identified by immunostaining with cholera toxin B (CTXB) and GM1 antibody. Briefly, cells were harvested by centrifugation, washed in PBS, and stained with CTXB or GM1 antibody as previously described (De Chatterjee et al., 2015). For CTXB labeling, _~1x10⁷ cells (trophozoites and cysts) were fixed in 4% formaldehyde, followed by immunostaining with CTXB conjugated to Alexa Fluor 488 (1 μg/mL) for 10 min, and labeling with CTXB antibody (1:200 dilution) for 15 min, as recommended by the manufacturer. LRs were also immunostained with GM1 antibody. For this, cells were incubated overnight with GM1 antibody (1:50 dilution) in cold (6-10°C) and then with a secondary antibody conjugated to Alexa Fluor 568 (1:500 dilution) for 1 h, at room temperature (RT). Both CTXB- and GM1-labeled cells were mounted with ProLong Gold antifading reagent with DAPI (Invitrogen) and visualized by confocal microscopy (Carl Zeiss LSM 700 confocal microscope). For the drug treatments, trophozoites were incubated with nystatin (27 μM), oseltamivir (20 μM), and myriosin (27 μM) for 30 min, at 37°C, before staining CTXB and GM1 antibody, as described by De Chatterjee et al. (2015).

LRs from Giardia (after treatment with nystatin and oseltamivir) were isolated following the method described by De Chatterjee et al. (2015). Briefly, ~1x10⁹ trophozoites were collected from the TYI-S-33 growth medium (supplemented with bovine serum and bile), washed in PBS, and incubated with HRP-conjugated CTXB for 1 h at 4°C (Blank et al., 2002; Graham, 2002). Trophozoites were resuspended in 5 mM Tris-HCl buffer (pH 7.8) containing EDTA (2 mM), DTT (0.4 mM), Triton X-100 (1%), and Halt Protease Inhibitor Cocktail (Thermo Fisher Scientitific). Lysate was placed at the bottom of Optiprep™ density gradient (35, 30, 25, 20, and 0%). Samples were centrifuged at 200,000 xg for 4 h, in a Sorvall T-865.1 fixed-angle rotor (Sorvall WX Ultra Series Centrifuge; Thermo Fisher Scientific, IL). Fractions of 1 mL were collected from the top of the gradient. To investigate which LR fractions were enriched in cholesterol and GM1, a cholesterol measurement kit (Invitrogen) and GM1 antibody-based ELISA assays were employed as described by De Chatterjee et al. (2015). Cholesterol- and GM1-enriched-LR fractions were subjected to proteomic analysis, as described below.

Large and small extracellular vesicles (EVs) from Giardia trophozoites were isolated by the method described by Gavinho et al. (2020). Briefly, Giardia cells were grown to confluency in 75-cm² flasks, as described above. Approximately 1 x 10⁷ trophozoites were harvested via an ice chill, and the cells were then pelleted by centrifugation at 2,400 xg, for 5 min at 4°C. For the drug treatments, cells were resuspended in TYI-S-33 medium (pH 7.1) without serum that contained nystatin (27 μM), oseltamivir (20 μM), and myriocin (27 μM) for 30 min at 37°C, followed by centrifugation at 2,400 xg at 4°C, for 5 min. The trophozoites (~1x10⁷ cells) were washed three times in PBS and resuspended in the incubation buffer (PBS containing 5 mM L-cysteine, 5 mM glucose, and 1 mM CaCl₂, pH 7.1) and incubated for 3 h at 37°C to allow the secretion of EVs. The viability was checked every hour by staining with propidium iodide and flow cytometry (Cytomics FC500; Beckman Coulter) analysis ( Figure S1 ) following the protocol from Ruiz-Medina et al. (2019). Cell debris were removed by concurrent low speed centrifugation at 600 xg (5 min) and 4000xg (30 min) at 4°C, respectively. Vesicles were isolated by differential centrifugation; large vesicles (LVs, microvesicle-like) were collected from the supernatant after 15,000 xg (60 min) at 4°C and small vesicles (SVs, exosome-like) were collected from pellet after centrifugation at 100,000 xg for 4 h, at 4°C (Gavinho et al., 2020).

After treatment with nystatin and oseltamivir, the size (diameter) and number of EVs in each EV-enriched fraction were measured by nanoparticle tracking analysis (NTA), using a NanoSight Model LM14C (Malvern Panalytical Ltd., Malvern, UK) (Bayer-Santos et al., 2013). Samples were diluted 1:10 in triple-filtered PBS. NTA 3.2 Dev Build 3.2.16 software was used for the analysis, with the following settings: 30 sec capture for a total of three captures, Camera Level: 14, Slider Shutter: 1259, Slider Gain: 366, FPS 25.0. Four independent experiments in triplicate were conducted, and the results were represented by bar graphs with appropriate statistical analysis. Representative figures directly acquired from the NanoSight are also shown in the results section.

Giardia samples were fixed in paraformaldehyde (PFA, 4%) at 4°C for 10 min, followed by vigorous washing. For LR staining, ~3x10⁶ trophozoites/mL were immunostained with CTXB-AF647 conjugate (dilution 1:1000) (Thermo Fisher Scientific, C34778) at 4°C for 10 min, followed by washing in PBS. The cells were then treated for 15 min at 4°C with a 1:200 dilution of anti-CTXB and 1% normal goat serum for 5 min. Cells were washed and samples were embedded in 4% low-gelling temperature agarose (Sigma-Aldrich, A9414) on a coverslip. Imaging buffer (50 mM Tris, 10% glucose, 10 mM NaCl, 40 mg/mL catalase, 500 mg/mL glucose oxidase and 10 mM cysteamine) (Sigma-Aldrich, M9768) was administered to the agarose-embedded samples for 15 min prior to capturing images following the Lin et al. (Lin et al., 2016) method. Data was collected using an oil-immersion objective (PlanApo N, 150x/1.45 NA; Olympus) in an oblique illumination setup on an Olympus IX-71 microscope outfitted with an objective-based TIRF illuminator. A 637-nm laser (Thorlabs, laser diode HL63133DG) with custom-built collimation lenses was used to excite the samples. A self-registration technique was used to ensure that minimal drift occurred during data collecting (Graus et al., 2018). We utilized a maximum likelihood-based technique from (Smith et al., 2010) to fit probe positions from raw dSTORM data. As reported in Graus et al. (2018), we used normal fitting with a Cramer-Rao lower bound (CRLB)-based, fit-accuracy threshold of 0.2 pixels or 22 nm.

Cluster analysis: the hierarchical single emitter hypothesis (H-SET) was used to estimate the placement of single emitter probes and reduce errors in subsequent cluster identification processes, as reported by Graus et al. (2018). The first H-SET pass condensed clusters of observations of several blinking fluorophores into single estimates of real fluorophore localizations. The second H-SET pass completed any multiple localization-collapse procedures left incomplete in the first pass and applied a density-based spatial clustering of applications with noise cluster-identification technique (DB-SCAN) (Ester et al., 1996). DB-SCAN has two parameters: epsilon, a distance metric that measures the maximum distance between two points within a cluster, and minpoints, the minimum number of points needed to form a cluster. By running 100 H-SET iterations over a variety of epsilons and minpoints, these values were optimized. The epsilon and minpoints values used for the analysis of the G. lamblia images were 29 nm and 3 points, respectively. Singlet clusters are defined as one or two localizations and multi clusters are defined as three or more.

For transmission electron microscopy (TEM) analysis, samples were fixed in 2% paraformaldehyde/2.5% glutaraldehyde in 100-mM sodium cacodylate buffer for 2 h at room temperature. The samples were then embedded in 2.5% agarose and post-fixed in 2% osmium tetroxide (Ted Pella Inc. Redding, CA) for 1 h at RT. After three washes with deionized water, samples were stained en bloc in 1% aqueous uranyl acetate (Electron Microscopy Sciences, Hatfield, PA) for 1 h. Samples were then rinsed in distilled water, dehydrated in a graded series of ethanol, and embedded in Eponate 12 resin (Ted Pella Inc). The samples were cut into 95-nm sections with a Leica Ultracut UCT ultramicrotome (Leica Microsystems Inc., Bannockburn, IL), stained with uranyl acetate and lead citrate, and viewed on a JEOL 1200 EX transmission electron microscope (JEOL USA Inc., Peabody, MA) equipped with an AMT 8-megapixel digital camera and AMT Image Capture Engine V602 software (Advanced Microscopy Techniques, Woburn, MA). Released vesicles were isolated by the method of Evans-Osses et al. (2017) and examined by TEM. Briefly, trophozoites were resuspended in the incubation buffer and incubated for 3 h at 37°C as described above. Cells and debris were separated by centrifugation at 600 xg (5 min) and 4000 x g (30 min) at 4°C, respectively. Vesicles (containing LVs and SVs) were collected and mixed with glutaraldehyde (1% final concentration). Samples were then allowed to absorb onto freshly glow discharged formvar/carbon-coated copper grids for 10 min. Grids were then washed in dH₂O and negative stained with 1% aqueous uranyl acetate (Ted Pella Inc.) for 1 min. Excess liquid was gently wicked off and grids were allowed to air dry. Samples were viewed on a transmission electron microscope.

The large and small vesicles (LVs and SVs) secreted by Giardia were isolated by differential density- gradient centrifugation, as described above, and subjected to proteomic analysis by one-dimensional liquid chromatography-tandem mass spectrometry (1D-LC-MS/MS) (Szempruch et al., 2016; Ribeiro et al., 2018; Cronemberger-Andrade et al., 2020; Cortes-Serra et al., 2020). Briefly, 1 μL of digested peptides was injected into a custom packed AQUA 5μm,125Å, C18 (Phenomenex, cat no. 04A-4299) 20 cm, 15 ± 1 μm, PicoFritEmitter (New Objective) equilibrated with optima grade (Thermo Fisher Scientific) 5% solvent B (90% acetonitrile, 0.1%formic acid), and 95% solvent A (0.1% formic acid). Peptides were applied and washed for 10 min continuously with 5% solvent B at a flow ratio of 0.5 μL/min maintained throughout the run before applying the elution gradient. Solvent B was increased to 35% over 85 min, and then increased to 95% over 5 min to maintain a high organic plateau for 9 min. The elution was decreased to 5% solvent B for 1 min and maintained at 5% solvent B for re-equilibration prior to the next sample injection for a total runtime of 120 min. The mass spectrometer was equipped with a NanoSpray Flex ion source (Thermo Fisher Scientific). Peptides were analyzed in a top 10 data-dependent MS2 acquisition by a Q-Exactive Plus Hybrid Quadrupole-Orbitrap mass spectrometer (Thermo Fisher Scientific). Full scan parameters were set to resolution of 70,000, 3e6 AGC target, and a scan range of 350 to 1600 m/z. MS2 parameters were set to 17, resolution of 500,1e5 AGC target, isolation window at 3.0 m/z, (N) CE: 30, charge exclusion: unassigned, 1, >8.

A proteomic analysis of gLRs was carried out by two-dimensional (2D) LC-MS/MS. All experiments were conducted with a U300 Q-Executive mass spectrometer and Ultimate 3000 RS HPLC System. A 10-h, 5-step 2D-LC-MS/MS method was run with 5-µl injections of ammonium acetate from the auto sampler (100, 250, 500, 750, and 1000 mM). One-hundred-fifty micromolar fused silica was used and packed with 5 cm of SCX material (Luna). We used the following instrumental conditions: top 10, runtime: 120 min, default charge state: 2, resolution: 70,000, AGC target: 3e6, maximum IT: 100ms, scan range 400 to 1600 m/z, for dd-MS2 resolution: 17,500, AGC target 1e5, maximum IT: 50 ms, Minimum AGC target: 8e3, dynamic exclusion: 15s. Solvent A was 95% acetonitrile, 5% water, and 0.01% formic acid. Solvent B was 5% water, 95% acetonitrile, and 0.01% formic acid. Solvent C was 500 mM ammonium acetate in solvent A. Run gradients were (t=0-5min, 5% B; t=5-5.1 min, per pulse [20, 40, 60, 80, and 100%] C; t=5.1-7.5 min, [20, 40, 60, 80, and 100%] C; t=7.5-8 min, 5% B; t=8-15 min, 5% B; t=15-120 min, 50% B). Wash gradients (t=0-5 min, 5% B; t=5-20 min, 95% B; t=20-22 min, 5% B; t=22-25 min, 5% B; t=25-40 min, 95% B; t=40-42 min, 5% B; t=42-60 min, 5%) were performed with a flow rate of 300 µL/min.

Proteome Discover (PD) 2.5.0.400 (Thermo Fisher Scientific) was utilized to process the raw files from the 1D- and 2D-LC-MS/MS analyses. Database for Giardia (30,894 sequences) was downloaded from UniProtKB (http://www.uniprot.org/) on 21 April, 2021. A contaminant data set composed of trypsin autolysis fragments, keratins, and standards found in the CRAPome (Mellacheruvu et al., 2013) repository and in-house contaminants were run in parallel. The PD analysis parameters were as follows: a false-discovery rate (FDR) of 1%, HCD MS/MS, fully tryptic peptides only, up to 2 missed cleavages, a parent-ion mass of 10 ppm (monoisotopic); fragment mass tolerance of 0.6 Da (in Sequest) and 0.02 Da (in PD 2.1.1.21) (monoisotopic). Two-high confidence peptides per protein were applied for identification. The PD dataset was processed through Scaffold Q+ 4.8.2 (Proteome Software, Portland, OR) for protein quantification. A protein threshold of 99%, peptide threshold of 95%, and a minimum number of 2 peptides were used for protein validation.

To conduct the Giardia attachment assay, HT-29 (human adenocarcinoma cell line) cells were grown in a four-well chamber slide (Nunc 155383 Chambered Cover Glass, Thermo Fisher Scientific) in McCoy’s medium supplemented with 10% fetal bovine serum (FBS). Giardia trophozoites were cultured in regular growth medium supplemented with bile and adult bovine serum (Gillin et al., 1989) described in Materials and Methods. Cells were pre-treated with one of the following inhibitors—i.e., nystatin (27 μM), oseltamivir (20 μM), methyl-β-cyclodextrin (MBCD, 1 mM) or myriocin (27 μM) for 30 min at 37°C. Attached trophozoites were harvested by chilling the flasks in icy water, collected by centrifugation (2,500 xg), washed in ice-cold PBS, and added (~0.5x10⁷ trophozoites) to the confluent monolayer of HT-29 cells. Giardia trophozoites and HT-29 cells were co-incubated for 3 h at 37°C in incomplete McCoy’s medium. Slides were removed from the chambers, washed, fixed in 4% paraformaldehyde, and incubated with TSA 417 antibody (TSA417, polyclonal, 1:100 dilution- obtained from Dr. Frances Gillin, UCSD; Gillin et al., 1990) followed by reacting with anti-rabbit secondary antibody conjugated to Alexa fluor. DAPI was used to stain the DNA of HT29 cells. Attached trophozoites were

Mouse experiments were performed under the vertebrate animal protocol approval of the Institutional Animal Care and Use Committee (IACUC) of the University of Texas at El Paso. Briefly, 3-4-week-old mice (C57BL/6) were given a cocktail of three antibiotics—i.e., ampicillin, neomycin (G418), and vancomycin following the methodologies described earlier by Solaymani-Mohammadi and Singer (2011) and Bartelt et al. (2013). All antibiotics were given at a concentration of 1mg/mL ad libitum by mixing with drinking water. The antibiotic treatment was initiated three days (72 h) prior to infecting the mice with Giardia and continued throughout the experiment. C57BL/6 mice (5 per group) were then infected with ~1 x 10⁷ luciferase-expressing WB trophozoites (resuspended in 0.1 mL PBS) to establish the infection. The firefly luciferase expressing PNT5 construct (Pan et al., 2009) was obtained from Dr. Chin-Huang Sun (Taiwan). This plasmid (PNT5) originated from 5’Δ5N-pac construct prepared by Dr. Steve Singer and his colleagues (Singer et al., 1998). The firefly luciferase gene was inserted at the NcoI/EcoRI sites. The resulting plasmid, PNT5-Luc+, contained the luciferase gene under the control of the α2-tubulin promoter (Pan et al., 2009). G. lamblia (WB isolate, assemblage A) was transfected with PNT5-Luc+, selected under neomycin and gavaged into mice. The treatment was initiated with oseltamivir after 48 h of infection. The doses of oseltamivir were 1.5 and 3 mg/kg per mouse, respectively. Drugs were resuspended in 0.1 mL PBS and gavaged every day for 13 days. Metronidazole (3.0 mg/kg) was used as a positive control. Mice were imaged through IVIS Lumina III In Vivo Imaging System (Perkin Elmer) to track the colonization of trophozoites in the intestine, and the parasite load was monitored by counting cysts collected from fecal materials. The method of sucrose floatation technique described by Roberts-Thomson et al. (1976) was used (with some modifications) to isolate cysts. Briefly, ~200 mg of fecal material from each mouse was collected and resuspended in 2.5mL of sterile PBS and subjected to homogenization. The fecal suspension was layered on 3mL sucrose solution (0.85M). The samples were centrifuged at 1,200 rpm for 10 min at 4°C. Cysts were collected from the water-sucrose interface with a Pasteur pipette (~700 μL) and counted under a microscope with the help of a hemocytometer.

Lipid rafts (LRs) are cholesterol- and sphingolipid (SL)-enriched nanoscale compartments of membranes that participate in various biological processes. We have previously shown that gLRs are enriched in cholesterol and monosialoganglioside 1 (GM1) (De Chatterjee et al., 2015). In the current experiment, to visualize raft domains, we used fluorescently conjugated cholera toxin B (CTXB). CTXB binds with GM1 glycolipids (Blank et al., 2002), and therefore, it is conceivable that GM1-enriched membrane domain should react with CTXB. The confocal microscopy analysis ( Figure 1 ) reveals that the trophozoite plasma membranes immunostained with Alexa fluor 488-conjugated CTXB. It is also clear from the images that staining with CTXB produces green punctate-raft-like structures in the membranes of trophozoites (WB isolate). In addition, the periphery of the ventral disc (VD) of WB isolate (trophozoites) also reacts faintly with CTXB ( Figure 1 A ) (De Chatterjee et al., 2015), indicating the presence of small amount of membrane components (especially the GM1 glycolipid) could be associated with the VD (Chavez and Martinez-Palomo, 1995). LRs are present in the plasma membrane of in vitro-derived, water-resistant cysts (WB isolate). The immunostaining of cysts with CTXB (green) and cyst-wall protein (CWP) antibody (conjugated with Alexa Fluor 568) shows the presence of raft domains in the cyst plasma membrane beneath the cyst walls (red) ( Figure 1 B ). Rafts are also present in GS and H3 isolates, and they all show the same punctate structures ( Figures 1 C, E ). However, the staining of VD with CTXB is not noticeable in these isolates. Figure 1 D shows the staining of LRs in GS isolate (trophozoite) with GM1 antibody conjugated with Alexa Fluor 568 (red), which shows a similar labeling pattern with CTXB ( Figure 1 C ), supporting the observation that CTXB binds with GM1 lipid of gLRs. ( Figure 1 F ) is the staining of in vivo-derived cysts (H3 isolate) collected from gerbils. Interestingly, in this cyst, LRs are associated with the plasma membrane as well as the endomembranes concentrated in the cytoplasm. Here, no attempt was made to stain the cyst wall and therefore the wall not visible in the figure ( Figure 1 F ).

Because of their dynamic nature, the size of LRs in the eukaryotic cells can vary from 10 to 200 μm (Day and Kenworthy, 2015). Since Giardia is an early-diverging eukaryote and gLRs participate in encystation and cyst production (De Chatterjee et al., 2015), we wanted to determine the size of raft domains in this parasite. We also asked if nystatin and oseltamivir could be effective in disassembling gLRs at the nanoscale range. We employed single molecular localization-based super-resolution microscopy. More specifically, an analysis was done with direct stochastic-optical reconstruction microscopy (dSTORM). Single-molecule-based, super-resolution techniques use photo-switchable fluorophores, single-molecule localization, temporal separation, and image reconstruction, which achieved an optical resolution of down to ~20 nm in the imaging plane (Betzig et al., 2006; Heilemann et al., 2008; Heintzmann and Huser, 2017). dSTORM imaging identified CTXB-labeled membrane structures consisting of multiple (≥3) CTXB labels. These structures were termed “multiclusters” and are interpreted as representative of gLRs. The mean radius of CTXB-labeled multiclusters in untreated Giardia was 19.29 nm, as shown in Figure 2D . Figure 2A shows the staining of GM1-enriched gLRs by CTXB and the overall distribution of raft sizes after treatment with nystatin and oseltamivir. Giardial cells were treated with the LR disruptors—nystatin (27 μM) and oseltamivir (20 μM)—to determine any changes in the LR-size distribution. Myriocin (27 μM), an inhibitor of 3-keto-sphinganine synthesis (Chen et al., 1999), was used as a nonspecific binding reference (De Chatterjee et al., 2015). When cells were treated with LR disruptors, the multicluster density of GM1-enriched LRs significantly decreased following treatment with oseltamivir or nystatin ( Figure 2B ). Subsequently, the mean density of singlet GM1 clusters increased in oseltamivir-treated samples but not with nystatin ( Figure 2C ). At this point, it is not clear why nystatin that disrupts multiple GM1 densities ( Figure 2A ) fails to influence the singlet GM1 density. Figure 2D shows the mean cluster radius of samples treated with either LR disruptor has significantly decreased. This provides evidence that treatment with oseltamivir and nystatin affects the size and distribution of LR microdomains in Giardia.

Giardia attachment on host intestinal cells is critical for establishing replication, producing intestinal infection, and causing subsequent giardial encystation or cyst formation. We have previously reported that two repurposed drugs—i.e., nystatin and oseltamivir—inhibit cyst production by disrupting the LRs (De Chatterjee et al., 2015). Studies also indicate that MBCD, a cholesterol-binding agent and a common LR inhibitor, blocks the attachment of trophozoites on cultured Caco-2 cells (Humen et al., 2011), implicating that giardial attachment to host cells is likely to be linked to gLRs. Attachment through LRs has also been reported in case of human sperm cells on oocytes (Van Blerkom and Caltrider, 2013), and evidence suggests that LRs regulate bacterial attachment and viral entry into host cells (Chatterjee et al., 2021; Palacios-Rápalo et al., 2021). It been demonstrated that, MBCD blocks microvesicles biogenesis as well as attachment to CaCo-2 cells by Giardia (Evans-Osses et al., 2017), further supporting the idea that LRs, giardial attachment, and vesicle biogenesis are interdependent. In the current experiment, we investigated whether two repurposed drugs—i.e., nystatin and oseltamivir—inhibit trophozoite attachment on cultured intestinal epithelial cells. We also asked about the types and classes of EVs secreted by Giardia, and whether these two compounds interfere with the secretion processes.

Since the attachment of Giardia on the small intestine lumen is critical for establishing gut infection (Einarsson et al., 2016), and because MBCD (an LR inhibitor) inhibits the attachment of trohozoites to Caco-2 cells (Humen et al., 2011), we investigated whether nystatin and oseltamivir, which disassemble raft domains (De Chatterjee et al., 2015), also inhibit the attachment of Giardia. To conduct this experiment, HT29 (human colorectal adenocarcinoma) cells were cultured in a four-well chamber slide as described in the methods section. Nystatin, oseltamivir, myriocin, and MBCD were added onto the attached trophozoites, as described by De Chatterjee et al. (2015) and Humen et al. (2011). Myriocin was used as a negative control and MBCD as a positive raft disruptor. The treatment was carried out for 30 min. Cells were collected washed in PBS, added onto the confluent monolayer of HT29 cells, and allowed to co-incubate for 3h at 37°C. The washing of inhibitor-treated trophozoites minimized the chances of residual compounds that could affect the LRs of HT29 cells (Humen et al., 2011). After the co-incubation, parasites and HT29 cells were immunostained with antibody against the trophozoite-surface antigen TSA417 (red, specific for Giardia) and DAPI (blue). TSA417 is a member of VSP family and therefore its expression may not be uniform in by all trophozoites. However, in the current study the parasites were allowed to attach on HT29 cells for short period of time (30 min) and the immunostaining with anti-TSA417 was carried out after the cells were fixed in paraformaldehyde. Results shown in Figure 3A indicate that the trophozoites (red) attached to the HT29 cells (blue). Analysis also reveals that nystatin (27 μM) and oseltamivir (20 μM) inhibit this attachment significantly. For quantitative purposes, trophozoites (red) attached to HT29 cells were counted. Approximately, 50-100 co-cultured cells were examined from 5-10 different fields to identify attached cells. Figure 3B demonstrates that the trophozoite attachments are inhibited by nystatin, oseltamivir, and MBCD. However, oseltamivir is more potent than the other two LR inhibitors. Myriocin, an inhibitor of serine-palmitoyltransferase (SPT), has no effects on the attachment.

EVs are released by Giardia and other parasites for inter-cell communications and host interactions (de Souza and Barrias, 2020). TEM analysis of EVs released by trophozoites. Figures 4A , B demonstrate the blebbing of vesicles from the plasma membranes.

For additional characterization, EVs were isolated following the method of Gavinho et al. (2020), with some modifications and analyzed by NTA. Nystatin (27 μM) and oseltamivir (20 μM) were used to determine whether they would influence the release of vesicles. Myriocin (20 μM) was included as a negative control (De Chatterjee et al., 2015). EV secretion by control and drug-treated trophozoites was carried out for 3 h at 37°C, in nutrient-supplemented PBS, as described in Materials and Methods section, to collect the EVs for NTA. Cell viability (control and treated) was tested during the incubation period by flow-cytometry analysis after staining with propidium iodide ( Figure S1 ). The NTA data indicate that Giardia produced two types of EV particles: LVs (mean size ~270 nm, microvesicle-like particles) and SVs (mean size ~100 nm, exosome-like particles) ( Figure 5 ). Results also show that the size and pattern of SV particles are drastically altered by nystatin and oseltamivir ( Figure 5A ). Statistical analysis demonstrates that oseltamivir-treated cells show a significant decrease in size (p<0.01), followed by nystatin-treated Giardia (p<0.05) ( Figure 5B ). The effect of myriocin is not significant. Interestingly, these compounds did not show substantial effects on LVs.

To further understand the interplay between membrane LRs and EVs in Giardia, we conducted a proteomic analysis. gLRs were isolated by density gradient centrifugation, as described by De Chatterjee et al. (2015). Trophozoites were treated with nystatin (27 μM) and oseltamivir (20 μM) and labeled with CTXB-horseradish peroxidase (HRP), lysed in 1% Triton X-100 buffer, and subjected to OptiPrep gradient centrifugation. Fractions were collected from the top of the gradient and the GM1-enriched fractions were identified as previously described (De Chatterjee et al., 2015). Fractions 3–5 (from the top) were considered as LR fractions and subjected to proteomic analysis along with the non-LR fraction (Fraction 7) for comparison. LVs and SVs were isolated as described in the Materials and Methods section (Gavinho et al., 2020), before conducting a proteomic analysis, essentially following previously described methodologies (Szempruch et al., 2016; Ribeiro et al., 2018; Cortes-Serra et al., 2020; Cronemberger-Andrade et al., 2020).

Figure 6 compares the proteomic profiles of LRs, LVs, and SVs. The Venn diagram shows that ~640 proteins are shared among the LRs, LVs, and SVs ( Figure 6A ). While only 418 proteins are shared between LRs and LVs, 60 proteins are common between LRs and SVs, and 162 proteins are shared by all three. Nystatin- and oseltamivir-induced changes in proteomic profiles are shown by the volcano plots ( Figure 6B ), and the original data are presented in Tables S1 , S2 . The analyses show the changes of protein expression in LRs, LVs, and SVs post-treatment. The drug treatments cause the up- and down-regulation of many proteins in all three components. However, most changes in the proteome profile (in comparison with untreated control) are observed in SVs ( Figure 6Be’, f’ ). Moreover, oseltamivir treatment is more effective than nystatin in altering the proteome profile. This is in good agreement with our NTA results ( Figure 5 ). Similar results with oseltamivir and LRs are also reflected in the principal component analysis (PCA) shown in Figure 7 . The PCA analysis reveals that SVs are separated from LRs and LVs, and their protein profiles are significantly affected by nystatin and oseltamivir. On the other hand, although LVs are separated from LRs, their protein expression profile is affected slightly by either of these two drugs.

To identify proteins that are partitioned in gLRs, LVs, and SVs, we ran an enrichment analysis to classify them pre-and post-treatment. The stacked bar in Figure 8 shows (A) percentage of enrichment; (B) biological process; (C) cellular component, and (D) molecular function. Table S3 shows the enrichment statistics and identifies the proteins belonging to each enrichment class. For example, in the category of proteins that participate in biological processes, the LR inhibitors nystatin and oseltamivir cause drastic changes in ribosomal large-subunit biogenesis ( Figure 8B ). Additionally, enrichment analysis was carried out on the protein overlaps between LR, LV, and SV groups ( Table S4 ). In LVs, the proteins involved in SNARE-complex disassembly are lost after treatment with the drugs. In SVs, while the proteins involved in cytoskeletal organizations are depleted by nystatin, the levels of ER stress-respond proteins are increased by oseltamivir ( Figure 8B ). In the cellular-component analysis ( Figure 8C ), we find no ribosomal proteins present in SVs. Microtubule-related proteins are decreased by oseltamivir and disappear after treatment with nystatin. An increase in cytosolic and cytoplasmic proteins in SVs by nystatin and oseltamivir treatment is also visible ( Figure 8C ). Finally, in the molecular-function category in SVs ( Figure 8D ), the proteins involved in peroxidase activity are not visible in treated samples. Because of the vast number of proteomic changes that occur after drug treatments, we reduced the scope of the proteomic profile comparisons and focused on potential virulence factors that have been identified in the literature and are likely to participate in host-parasite interactions (de Souza and Barrias, 2020). Table S5 is the list of virulence factors identified in LRs, LVs, and SVs by the proteomic analysis. Table S6 documents similarities and differences between the virulence-protein groups. To further illustrate this, we constructed a heatmap with virulence proteins that have been significantly up- or down-regulated when compared with their respective controls ( Figure 9 , Table S7 ). The PSM (peptide-spectrum match) values were normalized into Z-scores to better visualize changes within the data. The heatmap is divided into LRs, LVs, and SVs with and without the drug treatments, and is clustered into eight virulence factor groups. Results show that SVs express higher levels of giardins, CKs, and VSPs compared to LRs and LVs. The analysis also reveals that the LRs express higher levels of HCMPs and tenascins. On the other hand, two isotypes of delta giardins (accession numbers: A8BEZ6 and E1F6V5) are expressed ~3-4 fold higher in LVs compared to LRs and SVs. Interestingly, although nystatin and oseltamivir alter the expression of many of these proteins in LRs, LVs, and SVs, the effects of these two drugs are more robust on SV proteins. Both nystatin and oseltamivir reduced the expressions of most of the potential virulence factors, and oseltamivir is found to be more effective than nystatin ( Figure 10 ).

Because oseltamivir is more potent than nystatin in disassembling gLRs, blocking the attachment of trophozoites, and reducing the size of the exosome-like SVs, we decided to test this compound on mice infected with Giardia. Briefly, 3–4-week-old mice (C57BL/6) were infected with 1 x 10⁷ luciferase expressing (PNT5-Luc+) trophozoites (Pan et al., 2009) as described in Materials and Methods. Barash et al. (2017) also used luciferase-expressing Giardia (WB isolate) to locate high density foci of parasite in the small intestine of mice and gerbils. In the current investigation, the treatment with oseltamivir was initiated at day 2 post infection with gavage and continued until day 13. Metronidazole was used as a positive control. To assess the colonization of Giardia trophozoites in the small intestine, luciferin substrate (Gold Bio) was injected intraperitoneally, and the sedated mice were imaged through the IVIS. Figure 10A shows that the bioluminescent signals captured by IVIS as early as day 5 of post-infection, which increased steadily until day 10 before it declined by day 12. (Panel A). Other laboratories have reported that Giardia (both WB and GS) infection become significantly higher by day 4 or 5 of post-infection and peeks around day 7 to 10 (Solaymani-Mohammadi and Singer, 2011; Bartelt et al., 2013; Barash et al., 2017). The differences of infection profile in published work and in the current study could be due to the use of high doses of antibiotics (Crouch et al., 1990) and the changing of microbiome status in mice (Maarten's et al., 2021). There could be other host factors that might influence the colonization timing of trophozoites in the small intestine of mice used in this study. Nonetheless, the use antibiotics was necessary to establish infection by WB strain (Singer and Nash, 2000).

The focus of our experiment was to see if oseltamivir could reduce the colonization of parasite in mice. A cohort of five mice was imaged and the average of intensity (in total flux) was plotted against the days of post-infection. However, there were substantial variations in bioluminescent signals of individual mice (control and 1.5 mg/kg oseltamivir-treatment). The variation in bioluminescent signaling was also observed by Barash et al. (2017). Figure 10A shows that oseltamivir at 1.5 mg/kg reduced the bioluminescent signals emanating out of the small intestine effectively but was not statistically significant. However, this bioluminescent data gave us the guidelines to carry out cyst experiments described below.

For measuring the cysts released by infected animals, fecal materials from untreated and treated mice were collected and cysts were isolated by sucrose-floatation technique (described in Materials and Methods) and counted. Oseltamivir administered in a dose of 3 mg/kg shows a significant decrease of cysts in stool samples of mice ( Figures 10B , C ). The efficacy of oseltamivir is comparable to that of metronidazole, which was also administered at 3 mg/kg. Oseltamivir at 1.5 mg/kg fared significantly better than the control, showing a sharp decrease during peak infection at day 12. We found the variations of cysts counts of individual mice were minimum compared the bioluminescent study ( Figure 10A ) described above. The statistical analysis demonstrated that significant infection reduction (assessed by counting the cysts released by animals) was observed on day12 of the infection ( Figure 10D ). After day 12, the parasite load diminishes because most Giardia infections are self-limiting.