THE 72-KDA PROTEIN OF NAEGLERIA FOWLERI PLAYS AN IMPORTANT ROLE IN THE ADHESION OF TROPHOZOITES TO BALB/C MICE NASAL EPITHELIUM

Protozoal infections of the central nervous system (CNS) are the leading causes of morbidity and mortality worldwide. These statistics include diseases caused by protozoan members of the group of free-living amoebae (FLA; Kielian, 2009), which are organisms that are distributed throughout the world, of which Naegleria fowleri, Acanthamoeba spp., and Balamuthia mandrillaris are highly pathogenic nonobligate parasites that cause CNS diseases. Among the species belonging to these genera, infection by N. fowleri is the most important, since it occurs mainly in healthy children and young people causing mortality >98%. Naegleria fowleri is the etiologic agent of primary amoebic meningoencephalitis (PAM), an acute and fulminant disease of the CNS that results in death within 3 − 7 days after infection (Schuster and Visvesvara, 2004). Naegleria fowleri infections have been reported mainly in people with recent swimming activities in water bodies where N. fowleri trophozoites are found (Jamerson et al., 2012). The infection begins with the passage of trophozoites through the nasal cavity where the first contact occurs with the mucus layer of the olfactory epithelium, reaching the nasal epithelium and, later, the extracellular matrix (Carrasco-Yepez et al., 2014). In the first hours after infection, it has been observed that N. fowleri can bind to extracellular matrix glycoproteins such as type I collagen, fibronectin, and laminin-I (Jamerson et al., 2012), penetrating the olfactory neuroepithelium, migrating through the olfactory nerves and crossing the cribriform plate until reaching the olfactory bulbs (Jarolim et al., 2000; Rojas-Hernández et al., 2004). Several mechanisms of invasion by N. fowleri have been described, such as the secretion of pore-forming proteins, proteases, and glycoproteins that mediate adhesion (Aldape et al., 1994; Herbst et al., 2002; Serrano-Luna et al., 2007; Shibayama et al., 2013). As in other pathogens, there is evidence that the adhesion of N. fowleri to host cells is an important factor in the cytopathogenicity process (Cervantes-Sandoval et al., 2010). However, little is known about the molecules involved in the adhesion of the amoeba to the host epithelium. The existence of integrin-like proteins present in the trophozoites of N. fowleri has been demonstrated. Such proteins allow them to adhere to the components of the extracellular matrix (ECM). One of these proteins has a molecular weight of 60 kDa, which is a fibronectin-binding protein. When this protein is blocked with an anti-integrin antibody, the adhesion capability of trophozoites is inhibited (Han et al., 2004). The presence of another integrin-type protein has also been reported with a molecular weight of 70 kDa. This protein can be blocked with anti-integrin antibodies that inhibit the adhesion of trophozoites to ECM components such as laminin-1, collagen I, and fibronectin (Jamerson et al., 2012). In another study, the survival of mice intranasally exposed to lectin-pretreated amoebas was higher than that of control mice (Carrasco-Yepez et al., 2013). The authors suggest N. fowleri glycoproteins containing α-d-mannose and α-d-glucose could participate in trophozoite adhesion to mouse nasal cavity cells and, consequently, prevent infection by trophozoites. Glycoproteins with relative molecular weights of 70, 73, and 75 kDa might be responsible for adhesion (Carrasco-Yepez et al., 2013). Various molecules have been characterized that are implicated in the pathogenesis of N. fowleri, which are involved in the invasion and adhesion of trophozoites, mainly to the nervous tissue and to elements that make up the extracellular matrix and the lamina propria. However, the molecules that participate directly in the adhesion of N. fowleri trophozoites to the mouse nasal epithelium have not been described. In this work we demonstrate that there is a protein on the surface of N. fowleri trophozoites that has direct interaction with proteins of the nasal epithelium of BALB/c mice and plays an important role in the adhesion process.

Epithelial cells were obtained from 8–12-wk-old male BALB/c mice. White New Zealand breed rabbits, 2.5–3 kg, were used to obtain the antisera. Experimental animal use was endorsed by the Ethics Commission of the FES Iztacala with folio number 1294 and governed under the official Mexican standard NOM-062-ZOO-1999, Technical specifications for production, care, and use of laboratory animals.

The ATCC30808 N. fowleri strain was cultured at 37 C in bactocasitone medium (Difco, Le Pont-de-Claix, France) supplemented with 10% fetal bovine serum (GIBCO, Grand Island, New York) and 0.01% penicillin/streptomycin (Corning, Cellgro, Arizona). Parasites were harvested in the log phase at 72 hr and washed twice with phosphate-buffered saline (PBS).

Nasal epithelial cells were obtained from the nasal cavity of the mouse (nasal-associated lymphoid tissue, nostrils, septum, and lateral walls). Tissues were disintegrated and incubated in RPMI-1640 medium (Sigma-Aldrich, St. Louis, Missouri) at 37 C for 30 min under constant shaking. Suspended cells were filtered through a nylon mesh to remove debris, washed with RPMI-1640, and pelleted by centrifugation at 560 g for 10 min at 4 C. The pellet was suspended in 4 ml of 20% Percoll (Pharmacia Fine Chemicals, Piscataway, New Jersey) and superimposed in 4 ml of Percoll 40. The discontinuous 20-40% Percoll gradient was centrifuged at 560 g for 30 min at room temperature. Cells were recovered from the gradient interface, washed with PBS, and pelleted by centrifugation at 560 g for 10 min at 4 C. Pellets were suspended in 10 mM p-hydroxymercuribenzoic acid (PHMB) as a protease inhibitor (Carrasco-Yepez et al., 2018).

Total amoebae and epithelial cell extracts (NfEx and EpiEx, respectively) were obtained by freezing–thawing and using octyl-β-glucoside (1 mg/ml) for the solubilization of membrane proteins. Extracts were centrifuged to remove the cellular debris, PHMB was added and extracts were stored at −70 C until use. The protein concentration of the cell extract was determined by the Bradford technique.

The N. fowleri extracts were resolved by SDS-PAGE in 10% preparative gels, under reducing conditions. The gels were fixed and stained with Coomassie blue. From these gels, proteins ranging from 63 to 100 kDa, as well as the band of approximately 72 kDa, were extracted. The fragments were subjected to electro-elution, using elution buffer (25 mmol L⁻¹ base Tris, 192 mmol L⁻¹ glycine, 0.1% SDS) an Electro-Eluter device (Model 422, Bio-Rad, Hercules, California) with a PowerPac power source (Bio-Rad) at a constant current of 8–10 mA/tube for 4 hr at room temperature. The eluted proteins were recovered in an approximate volume of 800 μl, and 600 μl of PHMB was added. The samples were subjected to equilibrium dialysis overnight to remove SDS using PBS. Proteins were concentrated using centrifuge filters (Amicon Ultra-15, Darmstadt, Germany) with 30-kDa cut-off size membranes. Finally, the protein concentration was determined by the Bradford method. The purity of the protein was evaluated by electrophoresis in a 10% SDS-PAGE gel. Samples were stored at −40 C. These proteins were also used to obtain polyclonal IgG antibodies, using the immunization scheme described below.

A total of 3 New Zealand White rabbits were used for each antigen (N. fowleri extracts, epithelial cell extracts, and 72-kDa protein). The immunization scheme used was: On day 1, a first dose of 400 mg of antigen + Freund’s complete adjuvant (Sigma Chemical Co.), 1:1 dilution was administered subcutaneously. After 7 days, a second dose of 400 mg of antigen + incomplete Freud’s adjuvant in a 1:1 dilution was administered by the same route. Seven days later, 400 mg of antigen in 5 ml of saline solution was administered through the intramuscular route. On day 28, serum was obtained by cardiac puncture. IgG purification was performed with a protein A column, following the procedure recommended by the manufacturer (Pierce® Protein A Columns, Thermo Scientific, Waltham, Massachusetts).

For the study of the protein–protein interactions, we followed the protocol proposed by Hall (2015). For Western blot assay, N. fowleri (Nf) trophozoite extracts (20 μg/well) were resolved by 12% or 10% SDS-PAGE and transferred to a nitrocellulose membrane. They were blocked at room temperature with blocking buffer (PBS–0.01% Tween with 5% fat-free milk), with shaking for 1 hr. Subsequently, membranes were incubated for 2 hr with anti–N. fowleri IgG or antiepithelial cell IgG (anti-Epi IgG), diluted 1:250 in blocking buffer and washed 3 times with PBS–0.01% Tween. Finally, the membranes were incubated for 1 hr with the horseradish peroxidase (HRP) –conjugated secondary antibodies (dilution 1:1,000) in blocking buffer. They were washed 3 times with PBS–0.01% Tween and the enzymatic activity was evidenced with the substrate solution (0.1% H₂O₂, 17.5% methanol, 0.15% 4-chloro-α-naphthol, and 82.5% PBS). For the overlay tests, an incubation step was included with epithelial cell extract (40 μg), in blocking buffer for 2 hr before incubation with the primary antibody. Subsequently, membranes were washed 3 times with PBS–0.01% Tween and then incubated with the anti-Epi IgG (1:100 dilution) and subsequently incubated with the secondary antibody conjugated with HRP. Finally, they were developed with the same substrate (Hall, 2015).

One hundred micrograms of total Nf extract were prepared with the ReadyPrep™ 2-D Starter Kit (Bio-Rad) following the manufacturer’s recommendations. For the first dimension, 7 cm immobilized pH gradient (IPG) 3-10 strips were rehydrated with N. fowleri protein extract. Proteins were separated by isoelectric point (pI) via isoelectric focusing (IEF) using a PROTEAN i12 IEF cell system (Bio-Rad), programmed as per manufacturer recommendations. For the second dimension, the IPG strips were balanced and placed on 10% SDS-PAGE gels and the proteins were separated by molecular weight. Subsequently, the gels were transferred to a nitrocellulose membrane for immunodetection. Three membranes were used to perform the following treatments: (1) Nf extracts resolved in 2 dimensions, transferred to a nitrocellulose membrane and incubated with anti–72-kDa IgG. (2) Nf extracts resolved in 2 dimensions, transferred to nitrocellulose membranes, incubated with epithelial cell extracts and treated with anti-Epi IgG. (3) Nf extracts resolved in 2 dimensions, transferred to a nitrocellulose membrane and incubated with anti-Epi IgG.

Bands of interest, resulting from Western blot analysis and overlay assay by 1D were extracted from the 10% polyacrylamide gels and cut into small pieces. Subsequently, they were treated with 5% (v/v) acetic acid and 50% (v/v) methanol for 12 hr. Once destained, the gels were washed with deionized water and incubated for 10 min in 0.1 M ammonium bicarbonate. Subsequently, 0.05 M ditiotreitol was added over 45 min as a reducing agent. Time elapsed, 0.03 M iodoacetamide was added and incubated for 2 hr at room temperature in the dark. After incubation, the pieces were washed 3 times with 0.1 M ammonium bicarbonate and dehydrated with 100% acetonitrile under vacuum. Protein digestion was performed by adding 30 μl of 20 ng/μl modified porcine trypsin solution (Promega, Madison, Wisconsin) in 0.05 M ammonium bicarbonate, followed by incubation for 24 hr at 37 C. The resulting peptides were extracted twice in 50% (v/v) acetonitrile and 5% (v/v) formic acid for 30 min with sonication. The volume obtained was decreased by evaporation in a vacuum centrifuge and adjusted to 20 μl with 1% (v/v) formic acid. Peptides were analyzed by mass spectrometry using an integrated nano-LC-ESI MS/MS: quadrupole/light time, Synapt G2 high-definition mass spectrometer (Waters Corporation, Milford, Massachusetts) equipped with a NanoLockSpray ion source. The instrument was connected online to a NanoAcquity Ultra Performance Liquid Chromatograph (UPLC) (Waters Corporation). Two percent acetonitrile in Milli Q water with 0.1% formic acid (mobile phase A) and 98% acetonitrile in Milli Q water with 0.1% formic acid (mobile phase B) were used as a binary solvent system in a column of C18 UPLC symmetric capture (5 mm, 180 mm × 20 mm, Waters Corporation). The samples were desalted, concentrated, and washed with 100% of the mobile phase A, at a flow rate of 15 μl min. After 3 min, the capture column was switched online with an analytical column. Peptides were separated on a BEH, C18 UPLC column (1.7 mm, 75 mm × 100 mm, Waters Corporation) using a linear 40% B gradient over a 30-min period at a flow rate of 0.3 μl/min, followed by a 98% wash of mobile phase B (Escalante et al., 2018).

The N. fowleri trophozoites were cultured on glass coverslips in bactocasitone medium for 30 min at 37 C. The coverslips were washed with PBS and fixed with 2% paraformaldehyde for 20 min at 37 C. Subsequently, they were washed with PBS and incubated with the anti–72-kDa IgG primary antibody or anti-cathepsin IgG as control at a 1:5,000 dilution overnight at 4 C. They were washed with PBS and then incubated with Alexa 647 or Alexa 488 conjugated anti-rabbit IgG secondary antibody, respectively. Incubation with the secondary antibodies was carried out overnight at 4 C. Finally, the coverslips were washed with PBS and mounted in Vectashield mounting medium with 4′ 6-diamidino-2-phenylindole (DAPI). For visualization, images were collected and analyzed with an Axioscop 2 mot plus confocal fluorescence microscope (Carl Zeiss, Jena, Germany).

5 × 10⁴ cells per well of the human colon carcinoma cell line HCT116 (with epithelial morphology) were placed in a 94-well plate and allowed to grow until reaching confluence in supplemented RPMI medium for 24 hr at 37 C with 5% CO₂. Naegleria fowleri trophozoites (5 × 10⁴) were incubated with the anti–72-kDa IgG antibody at a 1:5,000 dilution for 10 min at room temperature (RT). As control groups, trophozoites without prior antibody treatment and trophozoites treated with nonimmunized rabbit IgG were used. Subsequently, the 3 treatments were added to the HTC116 cell culture and incubated for 1 hr at 37 C with 5% CO₂ to allow the trophozoites to adhere. Finally, the medium was removed, and the cells were fixed with 3.7% paraformaldehyde for 20 min at RT. The plate was washed with PBS and trophozoites were stained using anti–N. fowleri IgG antibodies, and subsequently incubated with a secondary antibody conjugated with HRP and highlighted with the substrate 3,3′ diaminobenzidine (DAB). The plate was washed with PBS and the cells were counted under a bright field microscope (Han et al., 2004; Jamerson et al., 2012).

Mice were inoculated by nasal route with 5 × 10⁴ trophozoites that had been preincubated with the anti–72-kDa IgG antibody Then, trophozoites were centrifuged and 30 μl of PBS were added (Han et al., 2004; Jamerson et al., 2012). The control group was inoculated with trophozoites that were not pretreated. Mice were monitored for 30 days and, finally, the percentage survival curve was performed.

Ten micrograms of the ∼72-kDa protein were resolved in 10% polyacrylamide gels, transferred to a nitrocellulose membrane, and blocked with 0.05% PBS-T-1% BSA for 1 hr at RT. Subsequently, membranes were incubated with the biotinylated lectins from Tritricum vulgaris, Canavalia ensiformis, Galanthus nivalis, and Artocarpus integrifolia (Sigma-Aldrich, St. Louis, Missouri), at a concentration of 5 μg/ml (1:500 dilution) in PBS-T and incubated overnight at 4 C. Membranes were washed and incubated with peroxidase-conjugated streptavidin (Pierce, Rockford, Illinois). The enzymatic activity was detected with the substrate (0.1% H₂O₂, 17.5% methanol, 0.15% 4-chloro-α-naphthol, and 82.5% PBS; Carrasco-Yepez et al., 2013).

Data processing was performed using the global ProteinLynx version 2.4 server and software (Waters Corporation) with a Protein Lynx Global Server (PLGS) (Waters Corporation). PLGS score of >95% confidence was accepted as correct. The UNIPROT database (https://www.uniprot.org) was used to collate data. The peptides were matched with the theoretical peptides of reported proteins from N. fowleri.

The extracts of N. fowleri obtained and nasal epithelium cells were resolved in 12% polyacrylamide gels and subsequently stained with Coomassie blue. Protein integrity was observed in both extracts, where bands with a relative molecular weight from 11 to 180 kDa were observed (Fig. 1A, lanes 1 and 2).

Western blot analysis shows the reactivity of the anti–N. fowleri antibodies with the immobilized extracts of N. fowleri (Fig. 1B, lane 1) and the reactivity of the anti-epithelial antibodies with the immobilized extracts of epithelial cells (Fig. 1B, lane 3). Bands ranging from 11 to 180 kDa were observed in both extracts. Bands with greater intensity were recognized at 180, 75, 42, 28, 19, and 12 kDa in N. fowleri extracts, and 75-, 17-, and 15-kDa bands were strongly labeled for the epithelial extract. Nonimmunized rabbit IgG (control) recognized only 1 band corresponding to 30 kDa, which was lightly stained, in N. fowleri extracts (Fig. 1B, lane 2), and 2 faint bands corresponding to 36 and 44 kDa were recognized in the epithelial cell extracts (Fig. 1B, lane 4).

After obtaining both the extracts and the anti–N. fowleri and anti-epithelial antibodies, protein interaction tests were performed between both extracts (Fig. 2). The N. fowleri extracts were resolved by 12% SDS-PAGE, transferred to nitrocellulose membranes, and incubated with 40 μg of the epithelial cell extract. After 2 hr of incubation, the anti-epithelial antibody (1:100 dilution) was added. In Figure 2A, lane 1, we show the Western blot where N. fowleri extracts were immobilized and treated with anti–N. fowleri antibody to confirm the integrity of the amoeba extracts that were used in the overlay assay. Meanwhile, in lane 2, an anti-epithelial antibody was used to detect a possible cross-reactivity of the antibodies with the amoeba extracts. We observe in N. fowleri extracts, as previously shown in Figure 1, that there is a pattern of bands that range from 11 to 180 kDa that were recognized by anti–N. fowleri IgG (Fig. 2, lane 1). Meanwhile, when N. fowleri extracts were analyzed, 2 faint bands at 68 and 57 kDa were recognized as anti-epithelial cells (Fig. 2, lane 2). Lanes 3 and 4, Figure 2A, show the results of the interaction tests. In lane 3, which corresponds to the overlay assay, a recognition pattern with bands ranging from 63 to 100 kDa of the N. fowleri extract is observed, mainly bands of approximately 90, 75, 72, and 66 kDa (Fig. 2A, lane 3). These N. fowleri bands interact with proteins from epithelial extracts. We clearly can support this interaction between N. fowleri and epithelial proteins, as when lane 3 is compared with lane 2 of the same figure, it is observed that these proteins (90, 75, 72, and 66 kDa) are not being recognized by the anti-epithelial antibody, confirming that this result is not the consequence of a cross-reaction given by the direct recognition of the anti-epithelial antibody towards these protein bands of N. fowleri.

Furthermore, the bands corresponding to the molecular weights of approximately 90, 75, and 72 detected in the overlay are comparable regarding the molecular weight with those that were identified in the protein band pattern of the N. fowleri extract (Fig. 2A, lane 1, gray arrows).

To analyze whether epithelial extract proteins may interfere with the interaction between N. fowleri proteins and their corresponding anti–N. fowleri antibodies, we incubated the interaction of immobilized N. fowleri proteins with epithelial extracts plus the anti–N. fowleri antibody. In this interaction (Fig. 2, lane 4) we observe a possible blockage by the binding of the epithelial cell extract. The interference given by the epithelial proteins was determined by the comparison between lane 1 and lane 4 regarding the decrease in the intensity of pre-existing bands. In connection with this, we can observe a lower intensity in the recognition of the bands that correspond to the molecular weight range of 90, 75, 72, and 66 kDa (lane 4) compared with their respective control (lane 1).

The intensity differences in these bands were analyzed by densitometry (Fig. 2B) and we can observe that there was a significant decrease (P < 0.001) when the extract of the epithelial cells interfered with the binding of the anti–N. fowleri antibody, mainly in bands of 90 and 72 kDa from lanes 1 and 4 (Fig. 2A).

In Figure 2, panel C we examined the opposite reaction, whether immobilized epithelial cell extract interacted with soluble N. fowleri extract. In lanes 1 and 2, Figure 2C, Western blot assays were performed to observe the recognition of the proteins of the epithelial cell extract and the cross-reaction of the anti–N. fowleri antibody. In lanes 3 and 4, Figure 2C, the result of the interaction tests of the N. fowleri extracts and epithelial cells is observed. In lane 3, Figure 2C, which corresponds to the overlay, we observed that N. fowleri extracts recognize different bands of epithelial extracts with molecular weights of 180, 70, 63, and 35 kDa. We note the 70-kDa band had greater intensity.

As we have previously shown for N. fowleri, in lane 4, Figure 2C we found that blocking proteins from epithelial extracts influenced by the N. fowleri extracts may interfere with interactions between epithelial extracts and their specific antibodies (anti-epithelial). For example, the polypeptide band of 70 kDa that was mainly recognized in lane 1 by anti-epithelial antibody, now in lane 4 is barely detected. The same happens for the 63- and 35-kDa bands, which were found to be more strongly labeled in Figure 2C, lane 1 and when we added N. fowleri extracts, these bands were weakly labeled (Fig. 2C, lane 4). These results were also corroborated by a densitometric analysis (Fig. 2D).

As we found some proteins of N. fowleri that were recognized by the epithelial extracts in the protein–protein overlay results, specifically bands of an approximate molecular weight of 90, 75, 72, and 66 kDa (Fig. 2A, lane 3), we decided to slice a section of the 10% polyacrylamide gels corresponding to this molecular weight range and to purify the proteins via electroelution.

It is worth mentioning that to avoid discarding some protein bands of interest, we purified such range from the 63–100-kDa bands. Subsequently, the electroeluted proteins were resolved by 10% SDS-PAGE. One gel was stained with Coomassie blue and the other was transferred to a nitrocellulose membrane to observe the integrity and purity of the range obtained (Fig. 3). In Figure 3A, lane 1, we observed the protein patterns after N. fowleri extract separation, which was taken as a reference for the band’s integrity, meanwhile in Figure 3A, lane 2, we observed the result of the purification of the range of proteins from 63 to 100 kDa. In the immunoblot, a band pattern is observed within the molecular weight range of 11 to 180 kDa which is recognized by the anti–N. fowleri IgG (Fig. 3B, lane 1). The bands of interest of 90 and 72 kDa, which were purified within the range, were detected by the same antibody whereas bands of 75 and 66 kDa were not detected in the immunoblot; however, these bands are present in the purification range from polyacrylamide gel (Fig. 3A, lane 2).

Proteins in the 63–100-kDa range, which were previously purified from the total extract of N. fowleri, were immobilized on nitrocellulose membranes, incubated with 40 μg of epithelial cell extract, and treated with anti–N. fowleri and anti-epithelial antibodies (Fig. 4). Lanes 1 and 2 correspond to Western blot assays for the recognition of 63–100-kDa N. fowleri proteins (Fig. 4A, lane 1) as well as their possible cross-reaction of the anti-epithelial antibody (Fig. 4A, lane 2). Overlay assays where 63–100-kDa N. fowleri proteins were immobilized on nitrocellulose membranes, incubated with epithelial cell extract, and treated with the anti-epithelial antibody are shown in Figure 4A, lane 3. Two bands, with a relative electrophoretic mobility of 72 and 66 kDa, are visible. These bands were not detected in the N. fowleri extract (63–100 kDa) by anti-epithelial antibody (Fig. 4A, lane 2). However, after interaction with the epithelial protein extracts these N. fowleri protein bands (72 and 66 kDa) can be detected by the anti-epithelial antibody. This suggests an interaction between purified N. fowleri proteins with the epithelial cell extracts and rules out the possibility of a cross-reaction between N. fowleri proteins belonging to this fraction with epithelial proteins is also discarded.

To confirm such interaction between N. fowleri proteins with epithelial cell proteins, as was shown in lane 3, we again carried out an assay using an anti–N. fowleri antibody instead of the anti-epithelial (Fig. 4, lane 4). There is an interference by the epithelial cell extract as the anti–N. fowleri antibody was not capable of recognizing the 72- and 66-kDa bands at the same intensity as it was observed in lane 1, where the epithelial extract was not added. The intensity of the bands corresponding to 72 and 66 kDa was confirmed by densitometric analysis (Fig. 4B) where we compared lane 1 (without epithelial cell extracts) with lane 4 (with epithelial cell extracts).

Based on Western blots, 72- and 66-kDa bands from N. fowleri were found to interact with proteins from epithelial cell extracts (Fig. 4, lane 3); such bands were extracted from polyacrylamide gels previously stained with Coomassie blue. The proteins obtained from the bands were analyzed by mass spectrometry nano LC–ESI-MSMS and the peptides were compared with the peptides reported for N. fowleri on the UNIPROT database (Table I). Analysis of the approximately 72-kDa band revealed 4 major proteins: actin type 1 (29 peptides, 56.8% coverage, 41 kDa MW), actin type 2 (25 peptides, 46.1% coverage, 41 kDa MW), membrane protein (22 peptides, 52.5% coverage, 20 kDa MW), and the heat shock protein 70 Hsp70 (19 peptides, 38.5% coverage, 71 kDa MW). For the band of approximately 66 kDa the analysis revealed 5 main proteins: actin type 1 (43 peptides, 76.8% coverage, 41 kDa MW), actin type 2 (37 peptides, 54.99% coverage, 41 kDa MW), ATP synthase F1 alpha subunit (47 peptides, 48.2% coverage, 62 kDa MW), membrane protein (38 peptides, 65.2% coverage, 20 kDa MW), and heat shock protein 70 Hsp70 (20 peptides, 41.3% coverage, 71 kDa MW; reliability percentage 95%).

As the 72-kDa polypeptide band was found to interact with epithelial cells with great intensity by 1-DE analysis (Figs. 2A, 4A, lane 3), an anti–72-kDa rabbit polyclonal IgG antibody was produced and used to detect the specific molecule that could act as an adhesin from this band. The antibody purified recognized the 72-kDa polypeptide band in N. fowleri, but not in the nonpathogenic species, Naegleria gruberi (Fig. 5A). Once the specificity was confirmed by 1-dimensional Western blot, we searched protein spots from N. fowleri of the same molecular weight (72 kDa) through a 2-dimensional Western blot; proteins that could be participating in the interaction with epithelial lysates. The results showed spots of N. fowleri with a molecular weight from 35 to 180 kDa and isoelectric points ranged from pH 3 to pH 10 (Fig. 5B), observing different numbers and intensity of spots, which depended on the interaction with epithelial extract and antibody used (Fig. 5B–D). Firstly, N. fowleri antigens were incubated with the rabbit polyclonal anti–72-kDa antibody and we detected 4 large spots of N. fowleri (Fig. 5B) corresponding to a molecular weight of 72 kDa and at an isoelectric point between pH 3.1–3.8 (square 1), 3.9–4.7 (square 2), 5–5.6 (square 3), and 5.7–6.5 (square 4). Thus, we corroborated that the antibody produced in rabbits recognizes proteins with a molecular weight of 72 kDa. However, we also found other spots with different molecular weights such as 30 (arrow 9), 35 (arrow 8), 50 (arrow 7), 63 (arrow 6b, c), and 90 kDa (arrow 5), which were recognized by the same polyclonal antibody.

The recognition by the epithelial cells extract towards N. fowleri proteins is observed in the overlay assay (Fig. 5, panel C) where some proteins of N. fowleri are shown interacting with the epithelial extract; it is evidenced through the spots with different molecular weights and isoelectric point such as 30 kDa with an isoelectric point of pH 3.5–4 (square 9), 35 kDa with an isoelectric point of pH 3.5–4 (square 8), 50 kDa with an isoelectric point of pH 5.5–6 (square 7), 63 kDa with an isoelectric point of pH 3–4 (square 6a) and 72 kDa, which had an a wide range regarding to its isoelectric point (pH 3–6.5; Fig. 5C, squares 1, 2, 3, and 4).

Finally, one of the membranes was incubated only with the anti-epithelial antibody to discard cross-reaction where such an antibody could recognize N. fowleri antigens. However, when we compare the spots recognized in the overlay (Fig. 5C) with those spots recognized directly by the antibody specific to epithelial proteins (Fig. 5D), it is observed that in certain areas the reaction was not a result of cross reactivity. For example, the spots corresponding to the molecular weights of 30, 35, and 50 kDa were detected in the overlay result (Fig. 5C, squares 9, 8, and 7, respectively). However, they were not detected when the anti-epithelial antibody was used directly (Fig. 5D). In contrast, the spots corresponding to the molecular weight of 72 kDa showed a difference regarding the intensity of the signal. These results were corroborated with a densitometric analysis to compare spots 1, 2, 3, and 4 (Fig. 5E). Furthermore, the comparison of the labeled spots of 30, 35, 50, and 63 kDa was carried out (Fig. 5F).

To know the localization of the protein of approximately 72 kDa in the trophozoites and its role in the adhesion of N. fowleri to epithelial cells, we continued using the anti–72-kDa rabbit polyclonal IgG antibody. First, immunofluorescence assay provided evidence that 72-kDa protein is on the N. fowleri trophozoite surface as the strongest fluorescent signals were visualized mainly in the membrane and some areas like the pseudopodia structures (Fig. 6A, arrowheads). We also showed an immunofluorescence of a known protein of N. fowleri (cathepsin B), which allows us to determine the localization of our protein of interest. In this case, we observed cathepsin B whose localization is also observed in pseudopodia structures (Fig. 6B, arrows). It is worth mentioning that in these assays the trophozoites were not permeabilized before treatment with the antibodies.

Continuing with the characterization of this molecule, we also analyzed the type of carbohydrates present through a Lectin Blot assay (Fig. 6C). The protein was resolved in 10% polyacrylamide gels and transferred to a nitrocellulose membrane; it was treated with different lectins that are related to different types of carbohydrates. Firstly, we confirmed the integrity and specificity of the 72-kDa band with the anti–N. fowleri and anti–72-kDa IgG antibodies (Fig. 6C, lanes 1, 2). The lectin blot showed that there was recognition by all lectins, mainly with Con A (Fig. 6C, lane 4) and Galanthus nivalis (Fig. 6C, lane 5), indicating that the polypeptide band of 72 kDa presents a large number of mannose and glucose residues. Although T. vulgaris (Fig. 6C, lane 3) and Jacalin (Fig. 6C, lane 6) recognized the polypeptide band of 72 kDa with lower intensity, it also suggested the presence of N-acetyl-d-glucosamine and Gal/GalNac residues in this band.

On the other hand, the adhesion assay was carried out using the HCT116 cell line. Initially, we observed the control group (Fig. 6D), which consisted of the interaction of trophozoites untreated (without antibody) with the HCT116 cell culture (average of 14 trophozoites per field). Second, treatment for N. fowleri trophozoites was also included using an unrelated antibody, nonimmunized rabbit IgG (adhesion of 8 trophozoites on average per field; Fig. 6E); this group was compared with the adhesion control group of trophozoites without previous treatment (P < 0.05, ANOVA and Student’s t-test, **P = 0.008; Fig. 6G). Finally, we observed N. fowleri trophozoites that were previously treated with the anti–72-kDa antibody and subsequently placed in interaction with the HCT116 cells (Fig. 6F). Adhered trophozoites were counted under brightfield microscopy. The results show a lower number of amoebae adhered to the culture (average of 2.5 trophozoites per field), compared to the control group. There were significant differences between both groups (P < 0.05, ANOVA and Student’s t-test, ***P = 0.0002; Fig. 6G). We also compared the groups of amoebae treated with the antibodies, the group where trophozoites were incubated with nonimmunized rabbit IgG, and the group of amoebae incubated with the anti–72-kDa IgG antibody, where we clearly can observe significant differences (P < 0.05, ANOVA and Student’s t-test, *P = 0.02; Fig. 6G).

The in vitro adhesion test was supplemented with an in vivo analysis. Mice of the BALB/c strain were used, and 2 groups were made. One was inoculated with the lethal dose of amoeba without prior treatment (control group) and the second group of mice was inoculated with trophozoites previously treated with rabbit anti–72-kDa IgG antibodies. For the control group, a survival percentage of 0% was obtained on day 12. On the other hand, for the group inoculated with the trophozoites treated with the anti–72-kDa IgG antibody, survival of 33% was achieved after monitoring for 30 days (Fig. 6H).

The adhesion of different microorganisms to host tissues has been reported as one of the main virulence factors for many pathogens. These pathogens express adhesive molecules on their surfaces to initiate the interaction with host cell receptors (Patti et al., 1994). Through this mechanism, most of the pathogens, approximately 90%, can invade different tissues via the mucosal route (Marasini et al., 2014). In the case of N. fowleri, it has been shown that the adhesion of trophozoites to the nasal mucosa is the first crucial step for the development of the disease (Jarolim et al., 2000; Rojas-Hernández et al., 2004; Siddiqui et al., 2016). Up to date, little is known about the molecules involved during the contact-dependent N. fowleri pathogenesis, more specifically those proteins that N. fowleri uses to attach to the nasal mucosa and cross the first nasal barriers.

To identify a specific protein band of N. fowleri that could be involved in the adhesion to the nasal epithelial, we first employed a blot overlay assay followed by adhesion assays either in vitro or in vivo. It is worth mentioning that the traditional overlay method has been used to analyze possible adhesins in several previous studies (Renesto et al., 2006; Kesimer et al., 2009; Vellaiswamy et al., 2011; Taniguchi et al., 2021). Based on this technique, we detected 2 bands of interest, approximately 72 and 66 kDa (Fig. 4A, lane 3), which were suspected to interact directly with epithelial cell soluble extract.

According to the results yielded by the overlay assays, the selected bands were identified by nanoliquid chromatography tandem mass spectrometry (nanoLC–MS/MS). The bands of approximately 72 and 66 kDa were analyzed for their homology with known proteins. Four main proteins were obtained for both bands: actin 1, actin 2, membrane protein, and heat shock protein 70 (Hsp70). The role of the nf-actin gene has been characterized, which is related to the contact-dependent mechanism of N. fowleri. The expression of the N. fowleri nf-actin gene plays an important role in the ability to increase cell adhesion, cytotoxicity, and phagocytosis (Sohn et al., 2019). Studies on the membrane protein (Mp2CL5) have demonstrated its participation in the pathogenicity of N. fowleri and it is probably a virulence factor, due to the expression of higher levels of the protein compared to the expression in nonpathogenic species, Naegleria lovaniensis and N. gruberi (Réveiller et al., 2001; Flores-Huerta et al., 2020; Guzmán-Téllez et al., 2020). On the other hand, studies carried out for the Hsp70 protein show that this protein is a factor in the tolerance to high temperatures and plays an important role in the proliferation and cytotoxicity of N. fowleri. Moreover, this protein has already been identified in in vitro assays as a factor of cytopathogenicity whose localization is in the cytoplasm, pseudopodia, and phagocytic food-cups of N. fowleri trophozoites. The virulence levels of the amoeba have also been associated with this protein (Song et al., 2007, 2008; Zysset-Burri et al., 2014).

Although we found 2 polypeptide bands (72 and 66 kDa) that directly interact with nasal epithelial cell extracts with high intensity by 1D overlay analysis (Fig. 4A, lane 3), we focused on the 72-kDa band. First, this band was not detected by the anti–72-kDa rabbit polyclonal IgG antibody in the nonpathogenic species N. gruberi (Fig. 5A), and secondly due to the result observed in the mass spectrometry analysis (Table I) where it was found that Hsp70 protein has a theoretical molecular weight of 71.3 kDa, close to the relative molecular weight of 72 kDa detected in this study. Therefore, the overlay technique was now performed on Western transfers of 2D gels to improve the resolution and to know possible specific molecules (spots) responsible for the interaction with the nasal epithelial extract. Firstly, when we immobilized N. fowleri total antigens and used the anti–72-kDa rabbit polyclonal IgG antibody as integrity control, 4 spots corresponding to 72-kDa molecular weight were found, however; it is noteworthy that other spots with different molecular weights such as 30, 35, 50, and 63 kDa were also found. This result could be due to the antibody recognition of antigenic determinants (peptides) in the same protein, which would determine the types of antibody–antigen complexes that will form (Sela-Culang et al., 2013). Therefore, in the structure of different proteins of N. fowleri probably there are repeated antigen sites that expose determinants that contribute to antibody binding. It is worth mentioning that we used an anti–72-kDa rabbit polyclonal IgG antibody; although it is true that most recognition given by this antibody is towards the 72-kDa band when we used the 1D Western blotting analysis (Fig. 5A, lane 1), when we used 2D Western blot technique the processing of proteins is different (Hajduch et al., 2005; Puente et al., 2009; Janson, 2011) resulting in recognition of additional spots. For example, in 1D Western blot, the total extract of the amoebae is used where many functional groups, mainly carbohydrates, could be recognized by the antibody as its potential binding site, increasing its specificity. On the other hand, in the 2D assay, these antibody binding sites may have changed and set antigenic determinants more exposed to the antibody resulting in the recognition of additional spots, as observed (Fig. 5B). Moreover, fragments of these proteins could migrate to molecular weights and isoelectric points different from those expected for the whole intact protein. Such fragments can also present posttranslational modifications that change their migration pattern in a 2D gel (Herrera-Díaz et al., 2018). Therefore, at the end of the process, fragments of the same protein can be identified or recognized by the same antibodies (anti–72-kDa IgG antibody) in spots corresponding to different molecular weights. After the analysis of this result was done, we continued with the interpretation of the 2-dimensional Western blot overlay (Fig. 5C) showed some N. fowleri proteins interact with the epithelial extract, as evidenced by spots with different molecular weights and isoelectric points that were detected with the anti-epithelial antibody. We focused on 4 main spots that correspond to the molecular weight of 72 kDa, located in a range of pI from pH 3.1 to pH 6.5 (spot 1 pH 3.1–3.8, spot 2 pH 3.9–4.7, spot 3 pH 5–5.6, and spot 4 pH 5.7–6.5). Additionally, we ruled out cross-reactivity by anti-epithelial antibody as the intensity of recognition is tenuous (Fig. 5D). We confirmed these results by densitometric analysis (Fig. 5E, F).

Regarding the spots corresponding to the molecular weight of 72 kDa, we can highlight that particularly the spots 2 and 3 with pI ranging from 3.9 to 5.6 are similar to the theoretical pI of Hsp70, the result observed in the mass spectrometry analysis (Table I). These spots can share sequences that might be involved in the interactions with the epithelial cell. As we mentioned previously, Hsp70 is known as a potential pathogenicity factor and is found to increase levels of expression when highly pathogenic trophozoites are compared with weakly pathogenic trophozoites, the results were obtained through comparative 2D electrophoresis (Zysset-Burri et al., 2014). Although heat shock proteins (HSPs) are intracellular chaperones that play a key role in the recovery from stress, it has been shown that Hsp70 integrates into the plasma membrane of cells after stress. Extracellular HSPs, specifically Hsp70, have been reported to activate macrophages, dendritic cells, and natural killer cells by a receptor-mediated process (Vega et al., 2008). Thus, possible binding of this N. fowleri Hsp70 protein to mice epithelial cells could be explained by this mechanism. Moreover, HSPs have been localized at the cell surface of fungal pathogens (López-Ribot et al., 1996; Nimrichter et al., 2005). For example, Hsp70 contributes to Candida–host interactions acting as a factor of invasion (Sun et al., 2010). Although we suggest that some peptides recognized by our purified anti–72-kDa IgG antibody belong to Hsp70, it would be necessary to excise each spot corresponding to the molecular weight of 72 kDa from the 2D gels to confirm by mass spectrometry the match with N. fowleri Hsp70.

Rojas-Hernández et al. (2020), found some immunogenic polypeptide bands of N. fowleri with a relative molecular weight of 250, 100, 70, 50, and 37 kDa. Particularly, the polypeptide band corresponding to 70 kDa contained peptides that matched with Hsp70 (Q6B3P1). Therefore, it is likely that the polypeptide band analyzed in the present work shares abundant antigen determinants with the polypeptide band of 70 kDa previously analyzed by our working group. Furthermore, Gutiérrez-Sánchez et al. (2020) carried out an analysis to identify differential proteins and protein pattern recognition between N. fowleri and N. lovaniensis. The results obtained in 2-D gels and Western blot showed differences in spot intensity between these 2 species, specifically those spots with relative molecular weight of 100, 75, 50, and 19 kDa. Some spots corresponding to these molecular weights were also identified as Hsp70. These Hsp70 peptides might have a fundamental role in the attachment of the amoeba to the epithelial cells during the entry of this pathogen. It has been previously reported that a synthesized inhibitor of HSP70 protein, N-formyl-3,4-methylenedioxy-γ-butyrolactam (KNK437) inhibited the Hsp70 mRNA and the protein synthesis was also reduced in a dose-dependent manner; therefore, the damage caused by N. fowleri to CHO cells in culture is reduced (Song et al., 2008).

Regarding its localization, Hsp70 has been observed in pseudopodia regions (Song et al., 2007), structures that have been related to the adhesion of the amoeba (Sohn et al., 2019); when we analyzed the localization of peptides corresponding to the 72-kDa band, we visualized the signal on the surface of N. fowleri trophozoites including areas resembling pseudopodial structures (Fig. 6A).

Additionally, it has been reported that glycoconjugates also play an important role in the adhesion of N. fowleri to target cells (González-Robles et al., 2007). It has been described that there is a notable difference in the expression of such glycoconjugates between the trophozoites of the pathogenic species, N. fowleri, and the nonpathogenic species, N. gruberi and N. lovaniensis (Carrasco-Yepez et al., 2013; Cervantes-Sandoval et al., 2010). Glycoconjugates mainly contain α-d-glucosyl, α-d-mannosyl, α-l-fucosyl, N-Acetyl-α-d-galactosaminyl, and α-d-galactose; when these terminal surface carbohydrates are blocked, the adhesion of trophozoites and the cytotoxicity of Madin-Darby canine kidney cells are reduced (Cervantes-Sandoval et al., 2010). In addition, with an in vivo mouse model we have demonstrated that the adhesion of trophozoites to the olfactory epithelium of BALB/c mice is affected when α-d-glucose and α-d-mannose residues from N. fowleri glycoproteins are blocked, obtaining a survival of up to 40% (Carrasco-Yepez et al., 2013). This evidence suggests that the carbohydrates present in the surface proteins of N. fowleri could be recognized by endogenous lectins of the nasal epithelium, facilitating host–parasite interaction, as has been also suggested for Entamoeba histolytica with intestinal epithelial cells (Guzmán-Téllez et al., 2020). Regarding this, the present work shows through a lectin blot that the protein of approximately 72 kDa mainly presents residues of α-d-mannose and α-d-glucose and residues of N-acetyl glucosamine, α-d-galactose, and N-acetyl galactosamine (Fig. 6B), suggesting that the presence of these carbohydrates has an important role in the adhesion function. These results are corroborated by mass spectrometry analysis, where several peptides corresponding to the identified proteins had theoretical posttranslational modifications such as O-glycosylations.

To demonstrate that the polypeptide band of approximately 72 kDa participates in the adhesion process, protein-blocking assays were performed using our purified anti–72-kDa IgG antibody, both in vivo and in vitro (Fig. 6C, D). The results indicate that there is inhibition in the adhesion of N. fowleri trophozoites to the culture of HCT116 cells (Fig. 6C) when the protein is blocked with the antibody. Similar studies have been conducted for other N. fowleri proteins. Anti-integrin antibodies have been used to demonstrate their role in the adhesion of amoebae to ECM components (Han et al., 2004; Jamerson et al., 2012). In other parasites, such as Entamoeba histolytica, the parasite–host interaction through an integrin-like protein, which is blocked with an anti-integrin antibody, leads to the inhibition of amoeba adhesion to ECM components (Sengupta et al., 2001). In the in vivo test, where previously incubated trophozoites with the anti–72-kDa antibody, BALB/c mice were inoculated and a 33% survival was achieved, suggesting a partial inhibition of the adhesion of trophozoites to the nasal epithelium of the mice. These results are similar to those obtained by Carrasco-Yepez et al. (2013), where 40% mouse survival is achieved when mice are inoculated with trophozoites whose carbohydrate residues (α-d-mannose and α-d-glucose) on the surface amoebic are blocked with Pisum sativum and Concanavalin A, lectins that are specific for both carbohydrate residues (Carrasco-Yepez et al., 2013). Although this last study has suggested different glycoprotein bands that would play an important role in adhesion, the specific weight of a glycoprotein band that is responsible for this process is not reported.

Our results suggest that peptides present in the band of 72 kDa which were matched with the proteins reported in Table I, are directly involved in the interaction of Naegleria fowleri trophozoites with the nasal epithelium of BALB/c mice. Such sequences of proteins were mainly found on the surface of trophozoites. In addition, the analyzed sequences showed a protein fingerprint with proteins that have been previously associated with adhesion structures such as amoebastomas. In this way, our in vitro and in vivo blocking assay reinforce the important role of this protein in the adhesion of Naegleria fowleri to the nasal epithelium of BALB/c mice, which in consequence would facilitate the amoeba migration to the brain and cause the PAM. The peptides present in the 72-kDa protein band would play a role in the initial interactions of the amoeba turning them as potential targets for drugs or vaccine strategies against the PAM.

This work is a requirement to obtain a Ph.D. degree in the Programa de Doctorado en Ciencias Biológicas, Universidad Nacional Autónoma de México (UNAM) for B. Flores-Suárez (CVU CONACyT No. 420822), who received a research grant from the National Council of Science and Technology of Mexico (CONACyT). The authors also wish to thank L. E. I. Roberto Carlos Flores-Suárez for reviewing the use of English in this manuscript. The study was funded by the following grant: DGAPA PAPIIT-UNAM IN22519 and IN222923.