Introduction

Balamuthia mandrillaris is a free-living amoeba that was discovered in 1986 in a mandrill baboon that died of neurological disease at the San Diego Zoo in California (Visvesvara et al. 1990). This amoeba is known to be the causative agent of a rare disease named granulomatous amoebic encephalitis (GAE), similar to the infection caused by Acanthamoeba spp., and infections in other organ systems, including the skin (Yera et al. 2015). GAE is usually a fatal infection of the central nervous system (CNS) that can affect both immunocompromised and immunocompetent individuals, with a mortality rate exceeding 95% (Mittal and Alsinaidi 2017). Approximately, 200 cases have been reported globally, with 96 cases documented in the Americas (Takei et al. 2018).

This microorganism has been isolated from environmental samples such as water, soil, and dust, as well as from animals (monkeys, gorillas, horses, and dogs) and humans (Lares-Jiménez et al. 2014; Schuster et al. 2003; Visvesvara et al. 2007). GAE typically occurs through the inhalation of contaminated dust or aerosols containing the amoeba or via exposure of broken skin to contaminated dust, soil, or water (Martinez and Visvesvara 1997), where the amoeba tries to reach the CNS either through the neuro-olfactory route or hematogenous spread (Matin et al. 2008). However, the precise mechanisms employed by B. mandrillaris to transmigrate the blood–brain barrier are still unclear (Kiderlen and Laube 2004). The incubation period is not well-defined, but the course of the infections is subacute to chronic, usually ending in death because the diagnosis of the disease is commonly post-mortem, as it is not recognized globally, and there is a lack of proper diagnosis and effective treatment (Siddiqui and Khan 2015; Yi et al. 2021). Moreover, GAE symptoms can be confused with other diseases like viral encephalitis or bacterial meningitis (Siddiqui and Khan 2015).

Regarding the molecular aspects of B. mandrillaris, susceptibility to drug treatments has been a challenging area with only a limited number of targets validated or characterized at the molecular level (Phan et al. 2021). In the context of drug susceptibility, a few validated targets include the heat shock protein 90-alpha, 3-hydroxy-3-methylglutaryl-CoA reductase, and lanosterol 14-alpha demethylase. These targets were identified through drug susceptibility tests using well-known antiparasitic compounds. Notably, the reconstruction and annotation of the provisional proteome of B. mandrillaris, derived from RNA-seq data, enabled the amplification, cloning, and validation of these targets through direct sequencing methods.

In contrast to the limited molecular characterization of targets in previous works, our current study builds upon this foundation by identifying specific immunogenic bands through a murine model. This approach serves to complement the existing knowledge, especially in the quest for potential biomarkers crucial for the early and accurate diagnosis of GAE caused by B. mandrillaris.

Materials and methods

Animals

All procedures performed in this study, which involved BALB/c mice, were in accordance with the Mexican Federal Regulations for Animal Experimentation and Care (NOM-062-ZOO-1999, Ministry of Agriculture, Mexico City, Mexico) and approved by the ethical standards of the Institutional Animal Care and Use Committee (number of approval ESM-CICUAL-ADEM-05/27–09-2019).

Amoebic cultures

Axenic cultures of B. mandrillaris strain ITSON-1 were grown in 75-cm2 culture flasks using BMI medium supplemented with 10% fetal bovine serum at 37 °C. Naegleria fowleri strain ATCC 30808 and Acanthamoeba spp. from a keratitis case were cultured axenically in 25-cm2 flasks using Bacto-Casitone medium at the same conditions (Lares-Jiménez et al. 2018; Rojas-Ortega et al. 2023).

Amoeba harvests and obtention of total extract for one-dimensional electrophoresis (1-DE)

Total extracts from the amoebas were obtained according to Lares-Jiménez et al. (2018) and Gutiérrez-Sánchez et al. (2020) with few modifications. Briefly, the trophozoites were chilled at 4 °C for 10 min. Then, the flasks were shaken to dislodge amoebas and centrifugated three times at 3500 RPM for 10 min with phosphate-buffered saline (PBS) washes between them. At the end, the pellet was suspended in 3 ml of PBS containing p-hydroxymercuribenzoic to prevent proteolysis. The sample was immersed in boiling water for 5 min and disrupted by sonication with one 5-s pulse at 100 W of amplitude.

B. mandrillaris protein analysis by SDS–polyacrylamide gel electrophoresis (SDS-PAGE)

The total extract was analyzed using SDS-PAGE, with 20 µg loaded onto a 12.5% gel and run at 110 V for 120 min. Some gels were stained with Coomassie blue to identify bands, while others were electroblotted at 400 mA for 1 h onto a nitrocellulose membrane for western blotting (Gutiérrez-Sánchez et al. 2020). Total extracts of N. fowleri and Acanthamoeba spp. were also analyzed by 1-D western blotting to compare protein patterns among the different amoeba species.

Protein extraction for two-dimensional electrophoresis (2-DE)

Protein extraction was done according to Gutiérrez-Sánchez et al. (2020) with few modifications. The pellet obtained after seven harvests was suspended with a solution of 5 ml of phenol and 5 ml of extraction buffer. The organic phase was added with 25 ml of ammonium acetate in methanol and stored at − 20 °C for 24 h. Finally, the pellet was dried at room temperature and dissolved in 400 µl of isoelectric focusing buffer, and it was centrifugated at 14,000 RPM for 5 min to obtain clear supernatant. The sample was stored at − 80 °C.

2-DE of B. mandrillaris

2-DE separated the previously obtained proteins following the method described by Herrera-Díaz et al. (2018) with some modifications. In general, 350 µg of total protein of B. mandrillaris were applied to an 11-cm IPG strip with a pH range of 3–10 for 10 min at room temperature in a rehydration tray. Subsequently, the tray was transferred to a PROTEIN i12 IEF isoelectric focusing unit for active rehydration (12 h at 20 °C). It was immediately followed by four focusing steps: 250 V for 30 min (fast), 1000 V for 2 h (gradient), 10,000 V (fast), and 1000 V maintained until a cumulative voltage of 45–47 kVh was reached. The washed strip underwent 2-DE on 12.5% gel in a SE 600 Ruby System at 25 °C with a voltage of 50 V for 22 h. Some gels were cut into four parts for transfer to a nitrocellulose membrane to perform 2-D western blotting, while others were stained with colloidal Coomassie for 16 h to achieve appropriate spot visualization.

Detection of B. mandrillaris antigenic proteins by 1-D and 2-D western blotting

The nitrocellulose membranes previously obtained by 1-DE and 2-DE were processed according to Gutiérrez-Sánchez et al. (2020) with few modifications. They were blocked with 0.05% PBS-Tween (PBS-T) plus 2% bovine serum albumin (BSA) for 1 h with constant agitation. Subsequently, the membranes were incubated with hyperimmune rabbit anti-B. mandrillaris serum at a 1:1000 dilution for 1-D and 1:100 dilution for 2-D western blotting overnight at 4 °C. Next, the membranes were incubated with anti-rabbit IgG peroxidase-conjugated at a 1:1000 dilution for 4 h at 4 °C in the dark. Afterward, the blot was visualized by adding the substrate solution (4-chloro-naphthol/methanol/H2O2). For the purified 50-kDa band of B. mandrillaris, the gel was loaded with 10 µg, and the following mouse sera were separately added: anti-50 kDa band, anti-total extract, and anti-control at a 1:100 dilution. Then, they were incubated with mouse anti-IgG and anti-IgA (1:1000).

50-kDa band purification by electro-elution

The antibody-recognized antigenic protein band was matched with the corresponding protein band on stained Coomassie Blue gel images to excise the protein band. The excised gel pieces (30 gels per electro-elution) were subjected to electro-elution at a constant voltage of 60 mA for 5 h according to Gutiérrez-Sánchez et al. (2023) with few modifications. Then, the eluted sample was transferred into a dialysis membrane for 24 h in PBS. Subsequently, centrifugation was carried out using a 30-kDa MWCO concentrator tube at 3000 RPM for 20 min at 8 °C until a constant volume of approximately 500 µl was obtained. The protein concentration was determined using Nanodrop.

Identification of antigenic spots by mass spectrometry

The samples were processed as previously described by Duarte Escalante et al. (2018) with some modifications. The immunogenic spots selected from the 2-DE gels were cut and treated with 5% acetic acid and 50% methanol for 12 h. They were then washed with deionized water and incubated for 10 min in 0.1 M ammonium bicarbonate, after which 50 mM DTT was added as a reducing agent for 45 min. Subsequently, 30 mM iodoacetamide was added and incubated for 1 h at room temperature in the dark. The pieces were washed three times with ammonium bicarbonate and dehydrated with 100% acetonitrile under vacuum. The digestion was performed by adding 30 µl of modified porcine trypsin solution at 20 ng/µl, followed by an 18-h incubation at 37 °C. Finally, the analysis was performed using the integrated nano-LC-ESI MS/MS system.

Data processing and protein identification

The data processing was carried out using the Protein Lynx Global Server version 2.5.1; Waters (PLGS) server and software, where a PLGS score > 96% confidence was accepted as correct. For protein identification, databases extracted from the website https://www.uniprot.org were used for Acanthamoeba, Dictyostelium discoideum, and Amoebozoa.

Immunization of BALB/c mice and serum collection

Male BALB/c mice, aged 8–10 weeks, were used for the three experimental groups. The band of interest (50 kDa) and the total extract were administered intraperitoneally on days 1, 7, and 14. The dose used for each immunization in the 50-kDa band group was 10 µg/µl plus 1 µg/µl of cholera toxin (CT) as an adjuvant. For the total extract group, the dose was 100 µg/µl plus 1 µg/µl of CT. The control group received 30 µl of PBS. After the third immunization, the mice were sacrificed on day 21, and serum was collected from them (Carrasco-Yepez et al. 2014; Rojas-Hernández et al. 2004).

Detection of antibodies by the ELISA method

The levels of the anti-50-kDa band or anti-total extract antibodies in serum samples were evaluated using the ELISA technique, according to the methodology reported by Rojas-Ortega et al. (2023) with few modifications. Initially, 96-well plates were coated with 100 µl of carbonate buffer pH 9.6, mixed with purified 50-kDa band (5 µg/well) or with total extract (10 µg/well), and incubated overnight at 4 °C. After incubation, the plates were washed thrice with PBS-T and blocked with 2% BSA in PBS-T for 4 h at 37 °C. The plates were incubated overnight at 4 °C with mouse anti-50-kDa band, anti-total extract, and anti-control sera. All sera were used at a 1:100 dilution. Subsequently, peroxidase-conjugated anti-IgA and anti-IgG mouse antibodies were added at a 1:500 dilution and incubated overnight at 4 °C. The plates were washed, and the developing solution (o-phenylenediamine 0.4 mg/ml, 0.04% H2O2 in 50 mM phosphate-citrate buffer pH 5.2) was added. The reaction was stopped with 2.5 M sulfuric acid. Absorbance was measured at 490 nm.

Immunolocalization of the 50-kDa band in B. mandrillaris trophozoites using a confocal microscopy technique

Subsequently, 5 × 105 trophozoites were placed on a microscope slide in 200 µl of BMI medium and incubated for 30 min at 37 °C. Subsequently, the methodology reported by Rojas-Ortega et al. (2023) was applied. Briefly, six samples were treated with 100 µl of 0.1% PBS-Triton at room temperature for 5 min to permeabilize the cells, while six other samples were left untreated. Afterward, mouse anti-50-kDa band or rabbit anti-total extract serum was applied at a 1:100 dilution and incubated at 4 °C for 2 h. Next, fluorochrome-coupled antibodies (rabbit anti-IgG Alexa 488® (green)/mouse anti-IgG Texas Red (red)) were applied at a 1:1000 dilution and incubated for 1 h in the dark. After this time, 5 µl of DAPI was added, and the cells were visualized in a confocal fluorescence microscope.

Lectin blot

For the detection of glycoproteins, we followed the modified protocol proposed by Carrasco-Yepez et al. (2013), where 20 µg of total extract of B. mandrillaris were separated by SDS-PAGE (12.5%) and electroblotted (400 mA for 1 h) onto a nitrocellulose membrane. Then, nitrocellulose membrane strips were blocked with 2% BSA diluted in 0.05% PBS-T and incubated overnight at 4 °C. The strips were treated with biotinylated lectins: Galanthus nivalis (specific for α-D-mannose), Pisum sativum (specific for α-D-mannose and α-D-glucose) and Tetragonolobus lotus (specific for α-D-fucose) at a concentration of 5 µg/ml, diluted 1:500 in PBS-T, and incubated for 4 h at 4 °C in the dark. The strips were incubated with streptavidin-bound horseradish peroxidase, and enzyme activities were detected with the substrate solution after being incubated for 5 min.

Densitometric analysis by ImageLab and statical analysis

The gels and western blots were scanned using the calibrated GS-900 densitometer (Bio-Rad). The Image Lab software was used to determine changes in the densitometry of the bands and spots. Statical analyses were performed using Prism GraphPad 8.3.0 software. Two-way analysis of variance unpaired test with Tukey post-test. P-values of p < 0.05 were considered significant.

Results

1-D and 2-D profiles

To describe the 1-D protein profile of B. mandrillaris, polyacrylamide gels at 12.5% were employed. The outcomes are presented in Fig. 1a. A band pattern ranging from 150 to 10 kDa was obtained. In relation to the intensity of the band, we observed that bands at 63, 50, 41, 23, 16, and 10 kDa were mostly expressed compared to the other bands (Fig. 1a, red arrows).

Fig. 1
figure 1

Protein expression of total B. mandrillaris extract. Total extracts of B. mandrillaris were subjected to electrophoreses in a SDS-PAGE 12.5% polyacrylamide gels (110 V, 120 min) and b two-dimensional using IPG ReadyStrip/pH 3–10, 12.5% polyacrylamide gels (50 V, 22 h)

Full size image

Furthermore, the total extract of B. mandrillaris was separated through 2-DE 12.5% over a pH range of 3–10 (Fig. 1b). Approximately 100 spots were observed between 191 and 11 kDa within the pH range of 4–7, exhibiting a pattern closely resembling that obtained in the 1-D gel (Fig. 1b). Spots at 68, 65, 50, 49, 33, 29, 20, 13, and 11 kDa displayed a higher recognition intensity (Fig. 1b, blue arrowheads). Meanwhile, the 41-kDa spot exhibited the highest intensity compared to the rest of the spots (Fig. 1b, red arrowhead).

The pink and yellow arrow and arrowhead in both the 1-D and 2-D gels (Fig. 1a and b) highlight specific bands and spots corresponding to findings in a prior publication by Kucerova et al. (2011). Notably, the pink arrow and arrowhead refer to a 30-kDa band, while the yellow arrow and arrowhead highlight a 60-kDa band. The purple arrowhead points to a 55-kDa band observed only in the 2-D gel (Fig. 1b), which was specific to the CDC:V416 strain by Kucerova et al. (2011).

Furthermore, features marked by blue boxes (S1-S3) in the 2-D gel (Fig. 1b) were selectively excised for subsequent mass spectrometry identification, as discussed later. Briefly, in S1, proteins identified included elongation factor 1-alpha, zinc finger protein, and myosin heavy chain kinase. S2 showed the presence of calmodulin and a predictive protein, while S3 revealed a membrane protein.

Western blots 1-D and 2-D

Through 1-D immunoblot analysis, we identified antigens shared by B. mandrillaris with other free-living amoeba species, such as N. fowleri and Acanthamoeba spp., by using rabbit anti-B. mandrillaris antibodies. In Fig. 2a, the 1-D blot reveals the predominant recognition of the serum IgG against the total extract of B. mandrillaris, displaying bands in the range of 270 to 30 kDa (Fig. 2a). Notably, the pink arrow highlights the 50-kDa band, which exhibits distinct recognition intensity, serving as a justification for its subsequent purification in the 1-D gel. Importantly, this 50-kDa band is unique to B. mandrillaris and is not present in other amoebas, further motivating its purification for detailed analysis.

Fig. 2
figure 2

Immunogenic analysis of B. mandrillaris proteins. a Total extracts of B. mandrillaris, N. fowleri, and Acanthamoeba spp. were subjected to electrophoresis in SDS-PAGE 12.5% polyacrylamide gels (110 V, 120 min), and b B. mandrillaris was subjected to two-dimensional electrophoresis using IPG ReadyStrip/pH 3–10, 12.5% polyacrylamide gels (50 V, 22 h), then the gels were transferred to a nitrocellulose membrane (400 mA, 1 h). The blots were probed with rabbit anti-B. mandrillaris serum

Full size image

For the total extract of N. fowleri, four bands at 95, 70, 56, 33, and 31 kDa were recognized (Fig. 2a). Regarding the total extract of Acanthamoeba spp., anti-B. mandrillaris antibodies identified more bands than N. fowleri across a broader range (127 to 30 kDa) (Fig. 2a). When comparing the immunogenicity among the three free-living amoebas, it was found that B. mandrillaris shares bands at 70, 56, 33, and 31 kDa with N. fowleri (Fig. 2a, green arrows); whereas with Acanthamoeba spp., it presents a greater number of common bands, including those at 127, 113, 60, 36, 34, and 30 kDa (Fig. 2a, red arrows).

Regarding the 2-D immunoblot, the antigens of B. mandrillaris recognized by the antibodies are observed in spots within a range of 130 to 16 kDa. However, enhanced recognition was found in the range of 90 to 26 kDa at a pH between 4 and 7. Specifically, spots at 74, 58, 50, 49, 43, 42, and 33 kDa exhibited the highest reaction intensity (Fig. 2b, blue arrowheads). The blue boxes marked as S1–S3 correspond to the areas highlighted in the 2-D gel (Fig. 1b), aligning with the spots selected for mass spectrometry analysis.

Protein identification by shotgun and mass spectrometry

To ascertain the polypeptides constituting the total extract of B. mandrillaris, and considering the limited B. mandrillaris database, shotgun results were compared with the Acanthamoeba database. Interestingly, only three proteins were identified: actin 1, plastin 3 T isoform putative, and S-phase kinase-associated protein putative, detailed in Table 1.

Table 1 Proteins identified by shotgun from the total extract of B. mandrillaris using the Acanthamoeba genus database
Full size table

Therefore, a comparison was made with the Amoebozoa phylum database, which included genera such as Arcella, Cavenderia, Dictyostelium, Entamoeba, Heterostelium, Naegleria, Paulinella, Vermamoeba, Amoeba, Sappinia, Ministeria, Acanthamoeba, among others. This comparison resulted in the identification of 24 proteins, as presented in Table 2.

Table 2 Proteins identified by shotgun from the total extract of B. mandrillaris using the Amoebozoa phylum database
Full size table

Three spots detected in 2-D gel and/or immunoblot were selected and excised from the 2-D gel (Figs. 1b and 2b; S1, S2, S3; blue boxes) and compared with the D. discoideum database. The selection criteria were particularly influenced by the outcomes of the shotgun analysis, detailed in Table 2, where polypeptides with similar molecular weights and isoelectric points were identified. This targeted approach aimed to explore or confirm the presence of specific polypeptides of interest, aligning with the functional aspects under investigation. Following Nano-LC-ESI MS/MS analysis, the results were compared with the D. discoideum database due to its closer relationship with B. mandrillaris, as demonstrated in Table 2. At the same time, the results were further cross-referenced with Amoebozoa databases, leading to the identification of various polypeptides (Table 3).

Table 3 Identification of proteins by Nano-LC-ESI MS/MS from spots of B. mandrillaris compared with the D. discoideum specie and the Amoebozoa phylum
Full size table

Immunization modulates levels of anti-B. mandrillaris antibodies

Considering that the 50-kDa band was recognized with great intensity by the IgG antibodies in B. mandrillaris, while in N. fowleri and Acanthamoeba spp. this band did not appear (Fig. 2a, pink arrow), we decided to assess whether this band could induce a response in mice following immunization with the 50-kDa band plus CT and with the total extract of B. mandrillaris plus CT. Seven days after the last immunization, mice were sacrificed, and serum samples were obtained. Subsequently, IgG and IgA levels against the 50 kDa band, and the total extract were analyzed using the ELISA technique.

When we analyzed the IgG and IgA antibody response in the serum samples from the groups immunized against the antigens of the total extract of B. mandrillaris (Fig. 3a), the group of mice immunized with the total extract plus CT showed significantly higher IgG levels (p < 0.0001) compared to both the group immunized with the 50-kDa band plus CT and the control group (Fig. 3a).

Fig. 3
figure 3

Response of anti-B. mandrillaris antibodies in immunized mice. BALB/c mice were immunized intraperitoneally with 100 µg/µl of total B. mandrillaris extract plus 1 µg/µl of CT, or with 10 µg/µl of the 50-kDa protein band plus 1 µg/µl of CT, three times at 7-day intervals. The control group received 30 µl of PBS. Antibody response against a total B. mandrillaris extract, and b 50 kDa protein band of B. mandrillaris were analyzed by ELISA. c Total extracts of B. mandrillaris and the 50-kDa protein band were subjected to electrophoresis in SDS-PAGE 12.5% polyacrylamide gels (110 V, 120 min) and transferred to a nitrocellulose membrane (400 mA, 1 h). Serum of immunized mice with total extract of B. mandrillaris plus CT and the 50 kDa band plus CT were incubated to detect the recognition pattern of the IgG and IgA antibodies. a and b The data expressed absorbance values at 490 nm mean ± SD of each experimental group (n = 5). **** indicate a significant difference of p < 0.0001, *** indicate a significant difference of p < 0.0005, and ** indicate a significant difference of p < 0.005 between control and immunized mice. +  +  +  + indicate a significant difference of p < 0.0001, and + indicate significant difference of p < 0.05 between immunized groups. c The meaning of numbers is described in the text

Full size image

On the other hand, when we measured the IgG response against the 50-kDa band (Fig. 3b), the immunized groups exhibited a higher response compared to the control. However, although the response of mice immunized with the 50-kDa band plus CT was greater compared to when measuring the response against the total extract, the response against the 50-kDa band in mice immunized with the total extract plus CT was significantly higher than in mice immunized with the 50-kDa protein plus CT (p < 0.005) (Fig. 3b).

Although we measured the IgA response against the total extract and the 50-kDa protein band (Fig. 3a and b, respectively), the levels were very low for both the immunized and the control groups. Additionally, no significant differences were found between these groups.

To determine the antigens of B. mandrillaris, including the 50-kDa protein band, recognized by antibodies present in the sera of mice immunized with the total extract plus CT and the 50 kDa band plus CT, 1-D western blotting was conducted. The results are depicted in Fig. 3c. In strip 1, it is evident that the IgG from the group of mice immunized with the total extract plus CT recognizes bands ranging from 250 to 29 kDa in the total extract of B. mandrillaris. Recognition of the 50- and 45-kDa bands was more intense than the other bands (Fig. 3c, strip 1, red arrows). Meanwhile, IgG from immunized mice with 50-kDa band plus CT recognized a smaller number of B. mandrillaris bands, in the range of 78 to 41 kDa, with the bands at 54, 50, 48, 46, and 41 kDa showing the highest intensity (Fig. 3c, strip 2, red arrows), while the rest of the bands at 78, 73, 64, and 60 kDa were recognized very weakly. The IgA from both immunized groups recognized four bands of total B. mandrillaris extract, those with weights of 50, 48, 46, and 41 kDa (Fig. 3c, strips 3 and 4).

Regarding the recognition by IgG and IgA antibodies from immunized mice of the immobilized 50 kDa band, we observed that the IgG from immunized mice with the total extract plus CT recognized three bands at 54, 50, and 46 kDa (Fig. 3c, strip 5). Meanwhile, the IgG from immunized mice only with the 50 kDa band plus CT recognized six bands at 54, 50, 48, 42, 41, and 37 kDa, with the 50-kDa band showing the highest intensity (Fig. 4c, strip 6, red arrow). Regarding the IgA from both immunized groups, only the 50-kDa band was recognized (Fig. 3c, strips 7 and 8, red arrows).

It is worth noting that both groups of immunized mice recognize the 50-kDa band in both total extracts and the purified 50-kDa band of B. mandrillaris. Furthermore, the recognition by IgG was greater than the recognition by IgA, as expected, given that IgG is the most abundant antibody in serum.

Identification of glycoproteins

To ascertain whether the bands recognized by 1-D immunoblot, in addition to being immunogenic, exhibit the characteristic of being glycoproteins, a lectin blot was performed using streptavidin-peroxidase to determine the presence of glycoproteins with residues of α-D-mannose, α-D-glucose, and α-D-fucose in the total extract of B. mandrillaris (Fig. 4). In strip 1, a more intense recognition of bands in the range of 253 to 82 kDa is observed, while the bands at 50, 48, 47, and 39 kDa were recognized more weakly. In strip 2, more bands ranging from 270 to 37 kDa were recognized. Meanwhile, in strip 3, seven bands with relative molecular weights of 130, 82, 71, 63, 53, 46, and 44 kDa were detected. The biotinylated bands of 130 and 82 kDa were recognized by the lectins from G. nivalis, P. sativum, and T. lotus, with the 82-kDa band highlighted as the most intense (Fig. 4, green arrows). The 50-kDa band was recognized only by the lectins from G. nivalis and P. sativum, which bind to α-D-mannose and α-D-glucose (Fig. 4, strips 1 and 2, respectively; red arrows). This assay demonstrated that the immunogenicity of these bands is possibly due to their complex structures (glycoproteins capable of efficiently activating the immune system).

Fig. 4
figure 4

Detection of B. mandrillaris glycoproteins by lectins. Total extracts of B. mandrillaris were subjected to electrophoresis in SDS-PAGE 12.5% polyacrylamide gels (110 V, 120 min) and transferred to a nitrocellulose membrane (400 mA, 1 h). The nitrocellulose membrane was incubated with biotinylated lectins: 1, G. nivalis; 2, P. sativum; and 3, T. lotus. Finally, the strips were incubated with peroxidase-conjugated streptavidin

Full size image

Immunolocalization of the purified 50-kDa band

The immunodetection of the 50-kDa band of B. mandrillaris was performed using confocal microscopy. Trophozoites were incubated with specific antibodies against the 50-kDa band and the total extract of B. mandrillaris. We observe that the positive labeling for the anti-B. mandrillaris antibodies is more abundant and present in most of the cellular structure (Fig. 5a, d, white arrows). Meanwhile, when observing the staining by the anti-50-kDa band antibody, it is found in both cytoplasmic regions and the cell membrane (Fig. 5b and d). It is noteworthy that the anti-50-kDa label is focused on specific regions of the membrane (Fig. 5b and d, yellow arrows). Figure 5e–h is an enlargement of Fig. 5d marked in a red box, where we can confirm the focused expression of the 50-kDa proteins (Fig. 5f, h, yellow arrows). Although the labeling with anti-50-kDa antibodies apparently uniformly recognizes the amoeba, focal patterns of the label are observed in both the cytoplasm and the membrane with higher intensity (Fig. 5f, yellow arrows) (additional data is given in the Online Resource).

Fig. 5
figure 5

Interaction of B. mandrillaris trophozoites with anti-50-kDa protein band and anti-B. mandrillaris serum. B. mandrillaris trophozoites were fixed with 2% paraformaldehyde. a and e Trophozoites were incubated with rabbit anti-B. mandrillaris serum at a 1:100 dilution. b and f Trophozoites were incubated with mouse anti-50-kDa protein band serum at a 1:100 dilution. Subsequently, a and e Alexa Fluor 488 (green) fluorescent anti-rabbit IgG secondary antibody was applied, and b and f Texas Red (red) fluorescent anti-mouse IgG secondary antibody was applied at a 1:1000 dilution. Finally, c and g DAPI was added to visualize the slides in a confocal microscope

Full size image

Discussion

Detecting surface protein biomarkers in B. mandrillaris with immunological significance is relevant for potential applications in diagnosing GAE. However, there is a scarcity of research related to this topic. In this study, we identified 36 peptides from immunogenic spots and the total B. mandrillaris extract, such as Ras-related C3 botulinum toxin substrate 2 (Rac2), known to participate in macrophage activation and cell adhesion (Rosowski et al. 2016; Yang et al. 2001); actin, which plays a role in morphological changes, phagocytosis, and migration (Labruyère and Guillén 2006); and ADP-ribosylation factor 1, located in the plasma membrane and involved in phagocytosis (Ooi et al. 1998; Uchida et al. 2001), among others. However, of the 36 identified peptides, we decided to focus on elongation factor 1-α (eEF-1α) from the 50-kDa band, as it has been demonstrated to play a crucial role in cell adhesion in other organisms (Itagaki et al. 2012; Yang et al. 1990). This aspect could be crucial in the pathogenicity of B. mandrillaris and its ability to induce GAE.

The band recognition obtained in the 1-D western blot, using hyperimmune rabbit serum against B. mandrillaris, exhibited some differences compared to the profile in the 1-D gel, indicating specific reactions to certain antigens (Figs. 1a and 2a). In the 1-D blot, bands were observed in a range from 270 to 30 kDa, with the 50-kDa band showing greater recognition (Fig. 2a, pink arrow). Compared with the study by Kucerova et al. (2011), where hyperimmune rabbit serum was used, they detected immunodominant bands with weights of 80, 70, 50, and 25 kDa. However, we did not detect the 25-kDa band (data not shown), but bands of approximately 70 and 50 kDa were indeed detected (Fig. 2a). Specifically, the 50-kDa band proved to be highly immunogenic, aligning with our results and emphasizing its importance as an immunogen of interest (Fig. 2a, pink arrow).

It is noteworthy to mention that in the study by Kucerova et al. (2011), when using serum from a GAE clinical case, bands of 75, 50, 40, and 25 kDa were observed. Although the 40-kDa band was not obtained in our study or theirs using hyperimmune rabbit serum, the 50-kDa band exhibited a stronger reaction with the serum from a clinical case compared to rabbit serum, emphasizing the significance of this band as a subject for future investigations to better understand its role in the immune response to B. mandrillaris.

Furthermore, in the study by Matin et al. (2007a), where serum from healthy individuals was used, reactions involving antigens of approximately 148, 115, 82, 67, 60, 56, 44, 42, 40, and 37 kDa were identified. Although immunogenicity of the 50-kDa band was not observed in that study, similar to the western blot using serum from healthy individuals in the study by Kucerova et al. (2011), they found other bands that align with those obtained in our study, such as the 82-, 60-, 56-, and 44-kDa bands (Fig. 2a). The absence of the 50-kDa band in sera from healthy individuals suggests that this band could be a specific marker for B. mandrillaris infection. Its presence in positive clinical cases and hyperimmune rabbit serum indicates specific antibody recognition, supporting the idea that it could serve as a distinctive marker for B. mandrillaris infection. This is reinforced by comparing the 1-D western blot of B. mandrillaris with cell lysates of N. fowleri and Acanthamoeba spp. genotype T4, where the presence of this band was not observed using hyperimmune rabbit serum against B. mandrillaris (Fig. 2a).

However, there are bands that exhibit similarities with B. mandrillaris antigens. For instance, in N. fowleri, antigens of 70, 56, 33, and 31 kDa were detected (Fig. 2a, green arrows). In the case of Acanthamoeba spp., a broader recognition of bands was observed, sharing with B. mandrillaris those of 127, 113, 60, 36, 34, and 30 kDa (Fig. 2a, red arrows). The recognition of a greater number of bands in Acanthamoeba spp. compared to N. fowleri using hyperimmune rabbit serum against B. mandrillaris may be attributed to the proximity between Acanthamoeba and Balamuthia, as they belong to the same Amoebozoa supergroup (Acanthamoebidae), unlike N. fowleri, which belongs to the Excavata supergroup (Heterolobosia) (Page 1988). Additionally, it has been reported that the size and AT content in the mitochondrial genome of B. mandrillaris are more similar to those of A. castellanii (Burger et al. 1995) than those of N. fowleri (Herman et al. 2013). On the other hand, in an alignment analysis conducted with B. mandrillaris scaffolds using the nucleotide sequence database, matches with A. castellanii of around 4.3% were revealed. Furthermore, for the 28S gene, an identity of 68.5% with A. castellanii was evident, positioning it as the closest known relative to B. mandrillaris (Greninger et al. 2015).

In addition to the above, additional bands of 79, 64, 54, 37, and 32 kDa were detected solely in Acanthamoeba spp., and a band of 95 kDa in N. fowleri. This could be attributed, for example, to cross-reactivity or antigenic homologies, as sera contain a mixture of antibodies directed against multiple antigens, and it is possible that some of these exhibit reactivity with antigens from related microorganisms or share similar sequences or structures in their antigens. However, these results differ from those reported in previous studies by Kucerova et al. (2011) and Schuster et al. (2008), in which no cross-reactivity between these amoebas was observed through western blot, ELISA, and immunofluorescence. The differences observed in our study compared to the studies by Kucerova et al. (2011) and Schuster et al. (2008) may be attributed to the different techniques used, but primarily to the variability among the strains, as well as the different types of antibodies and conditions employed to conduct these assays.

Few studies have reported the protein profile of B. mandrillaris. In particular, Kucerova et al. (2011) reported a prominent 30 kDa band in eight strains isolated from humans, horses, and primates, which we also observed in both gels, although not as notably as mentioned by these authors (Fig. 1a and b, pink arrows). Additionally, they reported an intense band at approximately 55 kDa in CDC:V416 and CDC:V188 strains, both from clinical cases (Booton et al. 2003; Visvesvara et al. 1990), which we only identified in the 2-D gel (Fig. 1b, purple arrow), along with a 60-kDa band specific to the CDC:V416 strain. We similarly found this band in both the 1-D and 2-D gels, with slight differences (Fig. 1a and b, yellow arrows).

We also compared the 2-D of the total extract of B. mandrillaris with N. fowleri and Acanthamoeba spp. We found similarities in the distribution of most spots in the 2-D gels of B. mandrillaris, mainly located at pH 4.8–6.8 and pH 9.3. This pattern is similar to what has been reported for Acanthamoeba spp. and N. fowleri, although the distribution of spots in N. fowleri showed a higher number of basic proteins (pH 5.4–7.2 and pH 9.7) (Gutiérrez-Sánchez et al. 2020). In contrast, Acanthamoeba spp. presented slightly acidic spots (pH 4–6.3) (Behera et al. 2016). These variations in pH profiles could have important implications for the biology and behavior of these amoebas, as extracellular and free-living organisms tend to have slightly acidic proteomes (Kiraga et al. 2007). This might explain the difference with Acanthamoeba, which is more commonly found in the environment. Additionally, it has been mentioned that organisms more associated with the host tend to have a more basic proteome, which could explain the severity of the pathology caused by N. fowleri (Kiraga et al. 2007).

Since in a 1-D gel, a band is composed of different polypeptides, in the 2-D western blot, we analyzed the antigenicity of the total extract of B. mandrillaris. In this analysis, we observed that the spots of higher intensity were detected at relative molecular weights of 74, 58, 50, 49, 43, 42, and 33 kDa (Fig. 2b, blue arrowheads). Thus, we demonstrated that these polypeptides do not lose their immunogenicity when subjected to the 2-DE and western blotting process. Therefore, we could assume that some of these polypeptides might participate as virulence factors in B. mandrillaris.

When we sequenced the different polypeptides against the Balamuthia database (UniProt), we did not find similarity with our peptides, likely due to the limited data in this database. Therefore, we used the Acanthamoeba genus and Amoebozoa phylum databases. Acanthamoeba was chosen for its phylogenetic proximity to Balamuthia (Greninger et al. 2015), although we only found similarities with three peptides: actin 1 (ACT1), Plastin 3 T isoform putative (L8GJG4), and S-phase kinase-associated protein putative (L8GW23) (Table 1). We then had to resort to a broader database to increase our results. For this, we used the Amoebozoa phylum database, where we found similarities with 24 sequences reported for different genera within this phylum (Table 2).

In both the 1-D and 2-D gels of the total extract of B. mandrillaris, we identified a band/spot with high intensity at approximately 41 kDa (Fig. 1a and b, red arrowhead), as also reported by Matin et al. (2007a, b) in 1-D gels. This band/spot has been reported in both A. polyphaga and N. fowleri, with a very similar molecular weight and isoelectric point (pI 5.1) (Caumo et al. 2014; Gutiérrez-Sánchez et al. 2020). Compared to the Amoebozoa database, using the shotgun technique revealed that the main polypeptide present in this spot was actin (F4Q4Y6) compared to C. fasciculata (Table 2) and actin 1 (ACT1) compared to A. castellanii (Table 1).

Actin is one of the most abundant proteins in eukaryotic cells, participating in multiple cellular processes such as invasion and motility (Das et al. 2021). In Plasmodium, the importance of actin has been demonstrated in the formation and release of binding sites (Münter et al. 2009). In E. histolytica, it has been reported that interaction with extracellular matrix proteins activates signaling pathways, which, in turn, induce actin gene expression and its reorganization to form adhesive structures, facilitating interaction with the substrate and even the release of proteases, leading to substrate degradation and promoting trophozoite dissemination (Meza 2000). For free-living amoebas, in the case of N. fowleri, the role of actin in cellular reorganization for adhesion to external substrates, such as fibronectin, has been demonstrated (Han et al. 2004), and a similar behavior has been observed in both Acanthamoeba and Balamuthia in their interaction with laminin, collagen, and fibronectin (Rocha-Azevedo et al. 2007, 2009). Therefore, a deeper understanding of actins’ role in the organization of the cytoskeleton, adhesion to various substrates, and triggering signaling pathways could reveal fundamental clues to understanding its function in the pathogenicity of B. mandrillaris.

In the spot with a relative molecular weight of 20 kDa, we found a sequence of ADP-ribosylation factor 1 (ARF1). It has been reported that this factor in D. discoideum is bound to GDP in the Golgi apparatus but also in the cytosol and the plasma membrane, suggesting its role in vesicular traffic between the endocytic pathway and the Golgi network (Gotthardt et al. 2006; Guetta et al. 2010). A previously described function of ARF1 is its involvement in the extension and closure stages of phagosomes, suggesting its role in phagocytosis (Swanson 2008). In mammals and T. brucei, it has also been localized in endosomal structures or the plasma membrane, in addition to the cytosol (Ooi et al. 1998; Price et al. 2007), and its participation in the activation of phospholipid-modifying enzymes and the phosphatidylinositol 5-kinase pathway has been reported (Donaldson et al. 2005).

On the other hand, the chaperonin GroEL proteins with relative molecular weights of 58 kDa were identified by comparing the sequence with B. cookevillensis (A0A0Q9YJ74) and L. jeonii (CH60). GroEL chaperonins are proteins known for facilitating the correct folding of proteins (Ishii 2017), and their presence in mitochondria has been demonstrated, highlighting their relevance in endosymbiotic theory (Hayer-Hartl et al. 2016). This is interesting because free-living amoebas, mainly Acanthamoeba, have been reported to serve as reservoirs for pathogenic microorganisms belonging to the families Legionellaceae, Mycobacteriaceae, Enterobacteriaceae, Vibrionaceae, among others (Greub and Raoult 2004; Thomas et al. 2010; Winiecka-Krusnell and Linder 2001), allowing them to feed and protect themselves against extreme environments.

In general, several of the identified proteins exhibit activity in the organization of the cytoskeleton, such as the actin-bundling protein (A0A6A5BNJ8) found when compared with N. fowleri. The ARF-SAR protein family (D2W6C2), which also participates in intracellular trafficking and is a key regulator of various events such as cell division and nuclear assembly (Reiner and Lundquist 2018), or the SAC protein domain (Q55G55), which performs various functions beyond cytoskeleton organization, such as signal transduction, vesicle trafficking, among others (Zhong et al. 2005).

However, during the analysis of proteins identified by shotgun technique using the Amoebozoa database (Table 2), the presence of eEF-1α (EF1A1) stood out, showing molecular weight and pI similar to S1 (Figs. 1b and 2b), besides playing a key role in cell adhesion (Itagaki et al. 2012). Additionally, other spots of interest were observed, such as S2 (Figs. 1b and 2b), possibly related to Rac2 (A0A6B2LNZ3) and S3, although not detected in the shotgun analysis by molecular weight and pI, nor observed in the 2-D gel (Fig. 1b), was intriguing due to its significant immunoreaction in the western blot 2-D (Fig. 2b, S3). For these reasons, mass spectrometry was performed on these three spots, comparing them with the D. discoideum database, and the results are shown in Table 3. The comparison with this microorganism is because most peptides identified in the total extract of Balamuthia shown in Table 2 were obtained by comparing them with D. discoideum. Furthermore, its relationship with Acanthamoeba and Balamuthia has been demonstrated by presenting a similar proteomic profile (Phan et al. 2021). However, spots S1, S2, and S3 were also compared with the Amoebozoa database.

Therefore, it was confirmed that one of the proteins present in S1 was eEF-1α (EF1A2). Other proteins were identified, such as the zinc finger protein (F4PLI8), which has been shown to be involved in various cellular functions, including DNA repair, apoptosis regulation, stress response, and cell proliferation (Hajikhezri et al. 2020; Laity et al. 2001). In other microorganisms, such as fungi, have been reported to be involved in pathogenicity (Li and Liu 2020). We also identified the presence of myosin heavy chain kinase (D3AW24), which participates in various processes such as cell division, locomotion, and intracellular trafficking (Russ et al. 2006). It is worth mentioning that in the case of N. fowleri, it has been suggested that this protein could be involved in phagocytosis (Zysset-Burri et al. 2014). Moreover, the presence of this protein in two distinct spots on the gel suggests the existence of protein isoforms or variants. Our analysis revealed the presence of a carbamidomethyl + C(23) modification in this protein; this modification could be the reason for its presence in both spots. This observation is consistent with Casado-Vela et al. (2011), who mention that protein variants observed in different spots could represent isoforms originating from post-translational modifications or protein processing mechanisms.

In the case of S2, despite not obtaining Rac2, we identified the presence of calmodulin (A0A6A5BIA3), a protein with multiple cellular functions, such as the regulation of motility, secretion, ion transport, and cell proliferation (O'Day et al. 2020). We also found a predictive protein (D2VDE5) that, according to the UniProt database, is involved in signal transduction and is classified as a membrane protein, similar to the only protein found in S3 (D3BGP3). It would be of great interest to conduct studies on these proteins to determine if they are involved in the adhesion mechanisms and/or pathogenicity of B. mandrillaris.

Regarding eEF-1α, it is a protein recognized for playing a significant role in protein synthesis (Andersen et al. 2003) and is considered the second most abundant protein in eukaryotes after actin (Abbas et al. 2015; Condeelis 1995). eEF-1α is a protein that has demonstrated multiple functions, such as signal transduction (Ransom-Hodgkins et al. 2000), apoptosis (Lamberti et al. 2004), proteolysis (Chuang et al. 2005), cytoskeleton organization (Pittman et al. 2009), and it has even been shown to be involved in infections and pathologies, participating in viral replication (Li et al. 2009), and oncogenesis (Abbas et al. 2015).

In D. discoideum, eEF-1α has been reported with a molecular weight of 50 kDa (Demma et al. 1990), and it has been attributed to a role in cytoskeleton organization (Yang et al. 1990). In mammalian cells, it has also been demonstrated that eEF-1α regulates cellular motility and additionally regulates phosphatidylinositol 4-kinase activity (Yang and Boss 1994). This enzyme catalyzes the first step in the biosynthesis of phosphatidylinositol 4-phosphate (PI-4-P) and the phospholipid phosphatidylinositol-4,5-bisphosphate (PI-4,5-P2), which is a second messenger regulating the adhesion energy between the cytoskeleton and the plasma membrane (Raucher et al. 2000). These interactions between the cytoskeleton and membranes are crucial in forming and retracting filopodia, lamellipodia, neurites, and other membrane processes in response to chemotactic signals and other stimuli. For example, in E. histolytica, it was found that eEF-1α and actin concentrate in adhesion sites or lamellipodia, which could explain the trophozoite’s activity after adhesion to host cells, such as phagocytosis and the release of proteases (Zhou et al. 2020). Additionally, there are reports highlighting the role of eEF-1α as a membrane receptor (Itagaki et al. 2012) and its involvement in the migration and invasion of cancer cells (Fukai 2017).

Considering that eEF-1α may function as a virulence factor, we immunized mice with the 50-kDa band to generate antibodies targeted toward this particular band. Using the ELISA technique, we observed that the antigens present in the 50-kDa band stimulate the production of IgG antibodies with no significant differences, and to a lesser extent, we observed the production of IgA antibodies (Fig. 3a and b). However, a higher production of antibodies was observed against the total extract of B. mandrillaris. This may be due to several reasons, one of them being that the total extract contains a variety of proteins, increasing antigenic complexity and providing a broader range of epitopes to recognize and respond to. Additionally, the concentration of the purified 50-kDa band may be lower compared to the total amount of proteins present in the total extract.

To validate this observed response, we conducted a 1-D blot to assess the recognition of antigens in both the total extract and the 50-kDa band using anti-B. mandrillaris and anti-50-kDa band mouse sera. In this analysis, besides confirming the presence of anti-50-kDa antibodies, other bands were also detected using both sera (Fig. 3c). This observation could be explained by the heterogeneous composition of the 50-kDa band, which includes various polypeptides, such as the 41 kDa CCCH-type zinc finger-containing protein or the 42-kDa Acetyl CoA acetyltransferase, both identified in the same spot (Table 3, S1).

The presence of the 50-kDa band, as well as the other bands using IgA antibody (Fig. 3c, strips 3 and 4), may be because this antibody, although typically associated with mucous membranes, is also found in low concentrations in the bloodstream, either due to its diffusion from mucous membranes or its production by plasma cells in lymphoid tissues. However, despite the diversity of bands identified using anti-50-kDa band serum, it is clear that the 50-kDa band exhibited the highest immunogenicity (Fig. 3c).

Subsequently, through lectin blot using a total extract of B. mandrillaris, residues of α-D-mannose and α-D-glucose recognized by G. nivalis and P. sativum lectins were detected in the 50-kDa band (Fig. 4, strips 1 and 2, red arrows), along with α-D-fucose residues in an 82-kDa band even detected by T. lotus (Fig. 4, green arrows). However, the 50-kDa band was more intensely identified by P. sativum (Fig. 4, strip 2), indicating that this band has more α-D-glucose residues than α-D-mannose. The presence of glycoproteins is of great importance, as their presence on cell surfaces has been reported, and they are implicated as adherence factors. For example, a mannose-binding protein has been described in Acanthamoeba, and its adherence to the corneal surface is attributed to it (Garate et al. 2004). Two laminin-binding proteins have also been identified, showing affinity to host extracellular matrix proteins (Rocha-Azevedo et al. 2009). In the case of B. mandrillaris, a 100-kDa galactose-binding protein has been found, which is also attributed to its adhesion to cerebral microvascular endothelial cells (Matin et al. 2007b). With G. nivalis lectin, we identified a 100-kDa band but with α-D-mannose residues (Fig. 4, strip 1).

Finally, we employed confocal microscopy with specific antibodies against the 50-kDa band to detect its in vitro distribution in B. mandrillaris trophozoites and determine where its highest presence is localized. Our findings indicate a notable presence of the 50-kDa band throughout the cell, encompassing both intracellular regions of trophozoites and the cell membrane. However, it is essential to highlight that the 50-kDa band showed a focused distribution in both regions (Fig. 5b and f, yellow arrows) (online resource). Previous studies in E. histolytica suggest that the eEF-1α protein concentrates in the membrane in the ectoplasmic region, presenting areas of high concentration in pseudopods and adhesion zones, indicating its role in cytoskeleton regulation (Zhou et al. 2020). In P. berghei, its presence in extracellular vesicles has been documented, and it has been demonstrated that, in combination with another component of these vesicles known as histamine-releasing factor, it inhibits the antigen-specific T-cell response by interfering with key phosphorylation pathways associated with T-cell receptor signaling, thus facilitating evasion of host immune responses (Demarta-Gatsi et al. 2019). However, information available on eEF-1α is scarce, emphasizing the need for further research to explore its potential functions in the pathogenicity of B. mandrillaris.

Conclusion

Our study represents the first-ever analysis using 2-D gels of a total extract from B. mandrillaris, offering valuable insights into its proteome and marking a significant advancement in understanding the protein-related characteristics of this pathogen.

We also unveiled the significance of a band with a relative molecular weight of 50 kDa, displaying residues of α-D-mannose and α-D-glucose, suggesting glycoproteins with potential adhesive functions in this pathogen. Through mass spectrometry, we identified one of the polypeptides present as eEF-1α, a protein that may play a crucial role in the adhesion and pathogenicity of B. mandrillaris. Confocal microscopy further demonstrated a focused distribution of the 50-kDa band throughout the cell, primarily on the membrane. This distribution pattern suggests the possible involvement of the protein in specific adhesion and recognition functions in B. mandrillaris.

These findings open new avenues for exploring the biology and pathogenicity of this microorganism and provide a foundation for further research that could have implications for the development of therapeutic and diagnostic strategies. The limited existing research underscores the need for future studies to delve deeper into the functional implications of the protein-related characteristics of B. mandrillaris.