Immunoproteomic Approach of Extracellular Antigens From Paracoccidioides Species Reveals Exclusive B-Cell Epitopes

Fungi of the Paracoccidioides genus are the etiological agents of paracoccidioidomycosis (PCM), a systemic mycosis restricted to the countries of Latin America. Currently, the Paracoccidioides complex is represented by Paracoccidioides lutzii, Paracoccidioides americana, Paracoccidioides brasiliensis, Paracoccidioides restrepiensis, and Paracoccidioides venezuelensis. Even with advances in techniques used for diagnosing fungal diseases, high rates of false-positive results for PCM are still presented. Additionally, there is no efficient antigen that can be used to follow up the efficiency of patient treatment. The immunoproteomic is considered a powerful tool for the identification of antigens. In addition, antigens are molecules recognized by the immune system, which make them excellent targets for diagnostic testing of diseases caused by microorganisms. In this vein, we investigated which antigens are secreted by species representing Paracoccidioides complex to increase the spectrum of molecules that could be used for future diagnostic tests, patient follow-up, or PCM therapy. To identify the profile of antigens secreted by Paracoccidioides spp., immunoproteomic approaches were used combining immunoprecipitation, followed by antigen identification by nanoUPLC-MSE-based proteomics. Consequently, it was possible to verify differences in the exoantigen profiles present among the studied species. Through a mass spectrometry approach, it was possible to identify 79 exoantigens in Paracoccidioides species. Using bioinformatics tools, two unique exoantigens in P. lutzii species were identified, as well as 44 epitopes exclusive to the Paracoccidioides complex and 12 unique antigenic sequences that can differentiate between Paracoccidioides species. Therefore, these results demonstrate that Paracoccidioides species have a range of B-cell epitopes exclusive to the complex as well as specific to each Paracoccidioides species. In addition, these analyses allowed us the identification of excellent biomarker candidates for epidemiology screening, diagnosis, patient follow-up, as well as new candidates for PCM therapy.


INTRODUCTION
Paracoccidioidomycosis (PCM) is a systemic mycosis restricted to the countries of central and south America and is considered one of the most important endemic mycoses in this region, especially in Brazil (Restrepo et al., 2001). The disease is caused by the fungal species that occupy the genus Paracoccidioides: Paracoccidioides brasiliensis, Paracoccidioides lutzii, Paracoccidioides americana, Paracoccidioides restrepiensis, and Paracoccidioides venezuelensis (Teixeira et al., 2009;Munoz et al., 2016;Turissini et al., 2017).
In the environment, Paracoccidioides spp. develop as filamentous structures (hyphae) and when under stress conditions and/or lack of nutrients, the hyphae can produce infectious propagules called conidia. PCM is acquired when an individual inhales conidia or fragments of hyphae that may reach the pulmonary alveoli, giving rise to the yeast form of the fungus, which is considered the parasitic form of Paracoccidioides (Wanke and Londero, 1994;Lacaz et al., 2002). Thus, Paracoccidioides spp. are characterized as dimorphic and thermally dependent fungi, presenting a saprobiotic mycelial phase and a parasitic yeast phase (Teixeira et al., 2014). Due to these characteristics, when found in ambient or cultured conditions in vitro, where temperatures vary around 22-27 • C, Paracoccidioides spp. grow as mycelium. When the mycelia or conidia are housed in the tissues of the host or incubated in vitro at around 36 • C, the dimorphic transition to the yeast phase occurs (McEwen et al., 1987;Franco et al., 1989;Brummer et al., 1993;Queiroz-Telles, 1994;Teixeira et al., 2009).
The development of PCM can occur immediately after contact with the fungus or can take years to be triggered. PCM can manifest itself in two clinical forms: acute/subacute (juvenile) and chronic (adult form) (Morejon et al., 2009). The acute and subacute forms of PCM account for 5-25% of cases, being observed in some states of Brazil, such as Maranhão, Goiás, Minas Gerais, Pará, and São Paulo. The chronic form presents in most cases, with a prevalence of 74-96%, manifesting mainly in male adults working in agricultural areas, aged between 30 and 60 years old. The male sex is usually more affected, with a rate of 22 men for every woman (Mendes, 1994;Costa et al., 2013). On the other hand, some patients present clinical manifestations compatible with the acute or subacute form, associated with other clinical manifestations generally observed in the chronic form, making the classification doubtful. In general, these patients demonstrate a very widespread disease with intense suppression of cellular immunity, which is referred to as mixed PCM form. In addition, PCM is a systemic disease whose host response to the infecting agent consists of a chronic granulomatous inflammatory process, which can progress to a fibrous process, characterized as anatomical and functional sequelae in the affected organs, particularly in the lungs (Shikanai-Yasuda et al., 2017. Notification of PCM's cases is not compulsory, and consequently, epidemiological approaches to PCM present certain obstacles. Among these obstacles are the difficulties in recognizing the infection acquired by Paracoccidioides species due to the poor capacity of laboratorial diagnosis in endemic areas. These issues lead to the low amount of PCM's epidemic data, which are obtained from hospital records, epidemiological research reports, case series, and mortality data (Martinez, 2015(Martinez, , 2017. Based on data obtained from care services for PCM patients, the incidence of the disease in endemic areas ranges from one to three cases per 100,000 inhabitants and three to four new cases per 1,000,000 inhabitants per year (Lacaz et al., 2002;Martinez, 2015).
For the diagnosis of PCM, biological materials, such as sputum, biopsy of injured tissues, material from lymph nodes, mucous membranes, and urine can be used, which can be analyzed by optical microscopy. For a long time, immunodiagnosis was not recommended for specific diagnosis of the disease as it presented low specificity, being more commonly indicated for patient follow-up (Shikanai-Yasuda et al., 2006). Despite that, in the Paracoccidioides's literature, there is a robust number of studies that have contributed to the improvement of PCM immunodiagnosis, with the development of serological tests that present high sensitivity and specificity (that reached 90-100%), as revised by Zancope- Oliveira et al. (2014). Even with these advances, the latest "Brazilian Guidelines for the Clinical Management of Paracoccidioidomycosis" still points to the fact that patients with histoplasmosis, aspergillosis, and leishmaniasis may have false-positive results for PCM. Additionally, determining which species of Paracoccidioides is the causative agent of the disease is still a restricted approach to some research centers, which limits the expansion of disease epidemiological data (Shikanai-Yasuda et al., 2017). Altogether, these facts point to the need for determination of specific Paracoccidioides antigens, for which the immunoproteomics approach is of great value.
To determine specific antigens of Paracoccidioides spp., we employed the immunoproteomics approach for identification and characterization of secreted antigens (exoantigens) by representative species of the Paracoccidioides complex. With this, it was possible to verify differences in the exoantigen profiles among the species used in this study. Through mass spectrometry tools, it was possible to identify 79 antigens secreted by isolates of Paracoccidioides. In addition, bioinformatics tools were used to analyze which antigens are secreted by alternative or classic pathways, as well as which antigenic epitopes are present in these molecules. In addition, we identified which molecules are shared among the isolates under study and which molecules are unique to each representative of the Paracoccidioides complex. These results demonstrate that Paracoccidioides spp. have several proteins that can be recognized by the human immune system. In addition, these analyses allowed us the identification of new possible molecules for PCM diagnosis, patient followups, and therapy.

Production of Extracellular Extracts
Yeasts of each Paracoccidioides spp. were transferred to the Fava Netto's liquid medium and maintained at 36 • C under agitation at 120 rpm for 96 h. After this period, cells were washed in phosphate buffered saline (PBS) and counted in Neubauer's chamber, and the viability was measured using Trypan blue dye. For inoculum, 5 × 10 6 viable cells/ml were used in liquid medium McVeigh-Morton (MMcM) (Restrepo and Jimenez, 1980) and incubated at 120 rpm for 24 h at 36 • C. After incubation, the cells were centrifuged at 2,000 × g for 20 min. The supernatant obtained after centrifugation was subjected to vacuum filtration on 0.22-µm membranes (Millipore, United States). The filtered samples were concentrated using a 10-kDa ultrafiltration system (Amicon R , Millipore, United States) and then washed three times with PBS (Weber et al., 2012). The quantification of extracellular extracts was determined by the Bradford method (Bradford, 1976).

Cell Lysis Detection Assay Using Polymerase Chain Reaction (PCR)
The genomic DNA of the isolates was obtained according to the protocol described by Sambrook and Russell (2001). For the PCR reactions, the supernatant of sample secretomes (2 µl) and the genomic DNA obtained were used. The reactions were performed in 40 cycles of 94 • C for 30 s, 55 • C for 30 s, and 72 • C for 2 min. The 681 base pair PCR products were generated using oligonucleotides (sense 5 -GACATGCGTGATATCGACTTG-3 and antisense 5 -GTGCGCCATGCCATTCT-3 ) for the formamidase coding gene (GenBank accession number AY163575). PCR amplicons were detected by 1% (w/v) agarose gel electrophoresis using the GelRed intercalating agent (Biotium TM , United States). To verify the sensitivity of PCR, a genomic DNA curve containing five dilutions (50 ng, 5 ng, 50 pg, 5 pg, and 1 pg) was constructed for each Paracoccidioides species, as described by Weber et al. (2012).

One-Dimensional Gel Electrophoresis (1-D SDS-PAGE)
The integrity of the extracellular extracts was verified by 1-D SDS-PAGE (Laemmli, 1970). Thirty micrograms of extracellular extracts from each isolate were prepared with sample buffer [0.2 M Tris-HCl, pH 6.8, 40% (w/v) SDS, 2% (w/v) β-mercaptoethanol, and bromophenol blue traces] and heated in the thermoblock at 100 • C for 10 min. Subsequently, the samples were submitted to 12% 1-D SDS-PAGE. As reference, the lowmolecular-weight marker (GE Healthcare, United Kingdom) was used. Soon after this step, the gels were stained by Coomassie Blue. ImageQuant 300 (GE Healthcare, United Kingdom) was used to obtain the images.

Production of Polyclonal Antibodies
Balb/c mice from 8 to 12 weeks old were provided by bioterium of the Instituto de Patologia Tropical e Saúde Pública, da Universidade Federal de Goiás (IPTSP-UFG). The animals were handled according to Conselho Nacional de Controle de Experimentação Animal (CONCEA-MCTIC) and submitted to the Ethics Committee of Comissão de Ética no Uso de Animais (CEUA-UFG) under the registry number 030/2016. Four groups containing five animals each were immunized with extracellular extract samples to induce the production of polyclonal antibodies. Separately, each group was immunized using proteins secreted by P. lutzii, P. americana, P. brasiliensis, and P. restrepiensis. For the immunizations, 50 µg of extracellular extracts were used in three doses with intervals of 15 days using the complete Freund's adjuvant (Sigma-Aldrich, United States) for the first dose and the incomplete Freund's adjuvant (Sigma-Aldrich, United States) for the second and third immunizations. The negative control was obtained from immunized mice only with Freund's adjuvant as previously described. After the immunizations, the animals were euthanized, and the whole blood of the animals was collected. Whole blood was incubated at 37 • C for 10 min, 4 • C for 10 min, and centrifuged at 500 × g for 10 min. After centrifugation, the immunized sera containing the polyclonal antibodies were collected.

Immunoblotting Analysis
Immunoblotting was performed as described by de Oliveira et al. (2018) with some modifications. The extracellular extract samples (30 µg) were submitted to 1-D SDS-PAGE at 12% with subsequent transfer to nitrocellulose membranes. The primary antibodies (1:500) were incubated for 2 h, and the secondary antibody (1:20,000) (alkaline phosphatase-labeled mouse anti-IgG; Sigma-Aldrich, United Kingdom) was incubated for 2 h. The membranes were then washed twice with PBS and once with alkaline phosphatase buffer for 15 min. The chromogenic substrate solution for alkaline phosphatase containing BCIP (5-bromo-4-chloro-3-indolyl phosphate) and NBT (tetrazolium-nitroblue chloride) (Sigma-Aldrich, United Kingdom) was used.

Coupling of IgG to Protein G-Sepharose and Immunoprecipitation
Protein G-Sepharose R 4B resin (Invitrogen, United States) was used for affinity chromatography. The resin was washed three times with PBS and 1.5 ml of immunized or control serum at a concentration of 1.5 µg/µl. The resin was subsequently added to the affinity columns, incubated at 4 • C for 30 min to promote the binding of the antibodies to protein G, which is covalently linked to Sepharose beads. After ligation, the columns were washed three times with PBS. Then, 1.5 mg extracellular extracts from each Paracoccidioides species were added to the columns, maintained for 16 h at 4 • C for the antigen-antibody binding. Afterward, the columns were washed three times using PBS, then the elution solution buffer [glycine 0.1 M, 0.02% sodium azide (w/v), pH 2.6] was added and incubated for 10 min. Subsequently, the supernatants containing antigens were obtained. The experimental procedure of immunoprecipitation is summarized in Supplementary Figure S1.

Tryptic Digestion of Exoantigens
Antigens of the Paracoccidioides species obtained by immunoprecipitation were subjected to tryptic digestion. Briefly, 10 µl of 50 mM ammonium bicarbonate pH 8.5 were added to 50 µg of the protein extracts and subsequently subjected to tryptic digestion as previously described (Murad and Rech, 2012;Bailao et al., 2014;de Curcio et al., 2017;de Oliveira et al., 2018). Then, 25 µl of 0.2% (w/v) RapiGest SF TM surfactant (Waters Corporation, United Kingdom) was added to the protein extracts and vortexed with subsequent incubation at 80 • C for 15 min. After this incubation, 2.5 µl of 100 mM dithiothreitol (GE Healthcare, United States) was added and incubated at 60 • C for 30 min. Then, 2.5 µl of 300 mM iodoacetamide (GE Healthcare, United States) was added and maintained at room temperature for 30 min. Subsequently, 10 µl of 50 mM trypsin (Promega, United States) solution was added and incubated for 16 h at 37 • C. To precipitate the RapiGest SF TM , 10 µl of 5% trifluoroacetic acid (Sigma-Aldrich, United Kingdom) was added and incubated for 90 min at 37 • C. The samples were centrifuged at 20,000 × g for 30 min at 6 • C, and the supernatants were collected. The supernatants were dried by using speed vacuum (Eppendorf, GER). All peptides obtained were resuspended in 80 µl of a solution containing 20 mM ammonium formate and Fosforilase MassPREP TM Digestion Standard (100 fmol/µl for PbEpm83, 400 fmol/µl for Pb01, 100 fmol/µl for Pb02, and 100 fmol/µl for Pb18) as the endogenous standard.

Immunoproteomic NanoUPLC-MS E Data Processing
After nanoUPLC-MS E acquisition, the data processing was performed by using Protein Lynx Global Server version 3.0.2 software (Waters Corporation, United Kingdom). The obtained spectra were compared with sequences deposited in the database of Pb01 (P. lutzii), Pb03 (P. americana), Pb18 (P. brasiliensis), and PbCnh (P. restrepiensis) 1 to identify the peptides.

Bioinformatics Analyses
After data processing and protein identification, the identified proteins were subjected to in silico analyses. The identified proteins were functionally categorized based on UniProt 2 database. To predict the subcellular localization of the secreted antigens, WoLF PSORT software 3 was employed. Expasy software 4 was used for isoelectric point prediction. For the prediction of proteins secreted by classical and non-classical pathways, SignalP 4.1 5 and SecretomeP 2.0 6 were, respectively, used. Regarding SignalP and SecretomeP software, values greater than or equal to 0.5 were considered indicative of secretion. For prediction of linear B-cell epitopes, we used the software BCPREDS 7 and ABCpred 8 .
To analyze and generate the Venn diagram, we used OrthoVenn software 9 and Draw Venn Diagram 10 . BLASTp (Basic Local Alignment Search Tool) software 11 and ClustalX 2.0 and R software 12 were used for the homology analysis of the identified exoantigens and heat map generation, respectively. ProtScale software 13 was used for prediction of hydrophilic peptides. The three-dimensional (3-D) modeling of the peptides was performed using I-Tasser 14 and the figures were visualized by using PyMol version 2.3 software 15 . To verify energy minimization and total structural refinement of 3-D models, the GalaxyWEB 16 and ModRefiner 17 servers were used.

Monitoring of Paracoccidioides Species Cell Integrity Based on PCR Analysis
To verify if cell lysis would influence the profile of proteins secreted by the P. lutzii (Pb01), P. americana (Pb02), P. brasiliensis (Pb18), and P. restrepiensis (PbEpm83) species, the PCR technique was used, as described by Weber et al. (2012), with some modifications.
The formamidase gene of Paracoccidioides species was analyzed. The sensitivity of the technique was attested to by constructing a standard curve using genomic DNA (samples from 50-ng to 1-pg dilutions) and primers to formamidase gene, generating a 681 base pair amplicon, which was evaluated by agarose gel electrophoresis.
These results demonstrate that this assay was able to amplify extremely low amounts of genomic DNA (1 pg); nevertheless, no amplifications were detected in the supernatant of secretome samples of the isolates under study (Supplementary Figure S2). It was possible to observe that the extracellular extracts did not present PCR-detectable contamination and, accordingly, there are no detectable contaminants from the intracellular compartment since the gene that encodes the formamidase was not amplified in the samples analyzed during the PCR tests (Supplementary Figure S2). This demonstrates that even in the case of any cell lysis, it was undetectable by PCR and thus cannot influence the profile of samples during the proteomics analyses.

Profile of Extracellular Extracts Evaluated by 1-D SDS-PAGE
Extracellular extracts of P. lutzii, P. americana, P. brasiliensis, and P. restrepiensis were analyzed by 12% 1-D SDS-PAGE. After separation, the proteins of isolates depicted a molecular mass distribution between 103 and 16 kDa (Supplementary Figure S3).
High abundance of extracellular protein species by P. lutzii, P. Americana, and P. restrepiensis species was found in the 45-kDa range, demonstrating that the profile of the secreted proteins between the analyzed species presented some similarities (Supplementary Figure S3). Among these, the one with the highest abundance of proteins in the 45-kDa range was the P. americana species. The P. brasiliensis species did not present high expression of these proteins in the 45-kDa range (Supplementary Figure S3).

Exoantigen Profiles Assayed by Immunoblotting
Analyses were performed with the objective of obtaining the profile of exoantigens of Paracoccidioides species. For this, the extracellular extracts of the P. lutzii, P. americana, P. brasiliensis, and P. restrepiensis species were used. Initially, extracellular extracts were subjected to 1-D SDS-PAGE with subsequent transfer to nitrocellulose membranes. Subsequently, immunoblotting assays were performed using antibodies obtained from control and immunized mice (Figure 1).
These results demonstrated that immunoblottings probed with antibodies obtained from immunized mice only with Freund's adjuvant (controls) did not show any reactivity with the proteins of Paracoccidioides species. On the other hand, the serum obtained from mice immunized with secreted protein extracts from P. lutzii, P. americana, P. brasiliensis, and P. restrepiensis depicts reactivity against Paracoccidioides exoproteomes (Figure 1). In addition, by means of these data, it was possible to observe differences in antigenic profiles among the species representative of the Paracoccidioides complex. The antigenic profile of P. lutzii showed a wide range of antigenic proteins between 105 and 19 kDa. For the P. americana species, a range of antigens between 99 and 14 kDa were presented. Antigens in the range from 99 to 37 kDa were observed in samples from P. brasiliensis species. Regarding P. restrepiensis, it is possible to observe that the profile of the antigenic proteins varies between 105 and 21 kDa (Figure 1). As a result, it was possible to verify that the secreted protein extracts from Paracoccidioides species present antigenic molecules that induce the production of different antibodies.

Immunoprecipitation of Paracoccidioides Exoantigens
After analyzing the differences in the antigenic profiles between Paracoccidioides species, we used immunoprecipitation approach to purify the antibody-reactive exoantigens (Supplementary Figure S1) by using the secretome samples from P. lutzii, P. americana, P. brasiliensis, and P. restrepiensis species. The immunoprecipitated samples were separated by 12% 1-D SDS-PAGE and demonstrate the presence of immunoreactive proteins, which proves the antigen-antibody binding (Figure 2). These profile differences are observed according to the molecular weight of the bands of proteins that are visualized in the electrophoresis (Figure 2). However, unlike what was seen regarding the immunized serum, the control sample showed no differences in protein profiles. The absence of immunoreactive proteins in the control sample demonstrates that there was no anti-secretome antibody production (Figure 2).
When comparing the exoantigens identified in Paracoccidioides species by software OrthoVenn, seven proteins from P. lutzii were exclusives in comparison with other fungus species described in our study. Four, seven, and 17 exoantigens of P. americana, P. brasiliensis, and P. restrepiensis were exclusive to these isolates, respectively (Figure 5 and Supplementary  Table S5). The proteins 2-methylcitrate synthase mitochondrial, malate dehydrogenase NAD-dependent, and HSP 70-like protein were identified in all species during the immunoproteomics analyses (Supplementary Table S6). Other proteins, such as endo-1,3(4)-beta-glucanase, dihydrolipoyl dehydrogenase, and glutamate carboxypeptidase, were identified in common between P. americana, P. brasiliensis, and P. restrepiensis species (Supplementary Table S6).

Non-homologous Exoantigens of Paracoccidioides Species
A comparative analysis was carried out to verify which molecules were identified as exclusive and conserved among the studied Paracoccidioides species. BLASTp-NCBI algorithm was used, and the genome of P. venezuelensis was inserted to analyze its exoantigens homology within the Paracoccidioides complex. The exoantigens hypothetical protein (PAAG_05807) and glutamate-1-semialdehyde 2,1-aminomutase (PAAG_06925) were exclusive to P. lutzii when their homology was compared to the other species of Paracoccidioides (Table 1). In addition, the exoantigen hypothetical protein (PAAG_12630) of P. lutzii showed low homology to P. brasiliensis (72%) and P. restrepiensis (76%), but the molecule was not homologous to P. americana and P. venezuelensis. All other exoantigens analyzed have strong homology between the species of Paracoccidioides, demonstrating the presence of these conserved proteins among Paracoccidioides species (Supplementary Table S7).

B-Cell Epitope Prediction of Exoantigens of Paracoccidioides Complex
Epitopes are specific regions of an antigen to which an antibody binds. These peptide regions are presented by MHC molecules to T lymphocytes (Abbas et al., 2008). Bioinformatics tools were used during analyses to predict and characterize which B-cell epitopes are present in the exoantigens identified by our immunoproteomics analysis. BCPREDs and ABCpred software were used together to verify if the predicted epitopes would be identified by both tools, leading to confirmation and strengthening of the experimentally obtained data. Initially, the proteins were submitted to BCPREDS software, where the epitopes with the highest score were selected (Supplementary Table S9). Predicted epitopes were also counted for each identified exoantigen, totaling 93 epitopes present in 15 proteins of P. lutzii and 158 epitopes present in 14 proteins of P. americana. For P. brasiliensis, 143 epitopes were predicted in 17 proteins, and there were 232 epitopes in 33 proteins of P. restrepiensis. All of these predicted epitopes are listed in Supplementary Table S9.
Subsequently, the epitopes were analyzed using ABCpred software, where the sequences predicted by this tool were selected and compared to the sequences predicted by BCPREDS software. This analysis was performed to verify if these two software tools share the same results, strengthening our experimental data (Supplementary Table S9). Thus, all epitopes predicted by both tools were verified and showed high score values, except for the exoantigens PAAG_12701 of P. lutzii, which did not show epitope prediction in any software.
In addition, the amount of predicted epitopes for each identified exoantigen was also evaluated. Interestingly, one of the exoantigens that had the largest number of predicted epitopes was endo-1,3 (4) beta-glucanase with 23 epitopes. This antigen was identified in P. americana, P. brasiliensis, and P. restrepiensis species. Next, aminopeptidase 2 presented 18 epitopes and was identified in P. americana and P. brasiliensis species. Heat shock proteins such as HSP70 and HSP60 presented 17 and 10 predicted epitopes, respectively, and were identified in all species in this study. The glutamate carboxypeptidase was another exoantigen that showed large numbers of predicted epitopes. A total of 11 epitopes were predicted in this molecule and identified in P. brasiliensis and P. restrepiensis species. Also, in P. brasiliensis and P. restrepiensis, the exoantigen dihydrolipoyl dehydrogenase was analyzed and presented nine predicted epitopes (Supplementary Table S9).

Specific B-Cell Epitopes of Paracoccidioides Species
After verifying which exoantigens were homologous to antigenic molecules in other organisms, the homology of the B lymphocyte epitopes predicted by BCPREDS and ABCpred software was verified to identify which antigenic sequences were conserved or not in other organisms. For this, all sequences of the epitopes listed in Supplementary Table S9 were analyzed against homologous epitopes of other important fungi and pathogenic bacteria. Proteins with homology to human molecules were also analyzed to identify and avoid the process of cross-reactivity during future tests. Initially, species such as C. neoformans, H. capsulatum, S. schenckii, C. immitis, C. albicans, A. fumigatus, M. tuberculosis, E. coli, S. pneumoniae, L. braziliensis, and H. sapiens were selected. All sequences of proteins homologous to the exoantigens identified during the immunoproteomic were obtained from the UniProt database. ClustalX software was used to perform the alignment and all of the sequences were manually checked. After analysis of all identified epitopes, we selected only those that did not present homology with other analyzed organisms. In addition, the epitope topology in the protein structure is important. To check which epitopes were hydrophilic and can be present in the protein surface, we used ProtScale software, as performed by Ejazi et al. (2018). Thus, we selected only the hydrophilic, non-homologous to other related organisms and exclusive epitopes of the Paracoccidioides complex ( Table 2). Therefore, five, 11, 10, and 18 epitopes of P. lutzii, P. americana, P. brasiliensis, and P. restrepiensis, respectively, were obtained ( Table 2). In addition, the homology of all epitopes among Paracoccidioides species was evaluated to confirm its presence in all species of this fungus.

3-D Analysis of Exclusive Epitopes of Paracoccidioides Species
All exoantigens that showed exclusive epitopes common to all Paracoccidioides species (Table 2) had their 3-D structure built by using I-TASSER software and were refined by ModRefiner and GalaxyWEB algorithms. The position of each identified  epitope was evaluated in exoantigen models and whether they were present in the surface of the model or not, which possibly facilitates antigen-antibody binding. The epitopes used for the analysis are displayed in Table 2, totaling two, four, three, and five exoantigens for P. lutzii, P. americana, P. brasiliensis, and P. restrepiensis, respectively. For P. lutzii, the exoantigen nucleic acid-binding protein (PAAG_04814 -3 epitopes) was evaluated. During the analysis of the 3-D model of this exoantigen, it was possible to verify that one epitope was on the surface ( Figure 7A). In P. americana, molecular modeling of four exoantigens was made: serine protease (PABG_00534, one epitope), aminopeptidase 2 (PABG_00710, one epitope), endonuclease G, mitochondrial (PABG_04696, two epitopes), and endo-1,3(4)-beta-glucanase (PABG_12341, seven epitopes). After analysis, it was verified that all epitopes analyzed in the exoantigens were present externally to the analyzed models ( Figure 7B). Regarding P. brasiliensis predicted models, the exoantigens aminopeptidase 2 (PADG_03149, one epitope), carbonic anhydrase (PADG_07674, two epitopes), and endo-1,3(4)-beta-glucanase (PADG_12370, seven epitopes) had their epitopes analyzed in the 3-D models obtained, and it was also possible to verify the presence of the antigens on the surface (Figure 7C). In P. restrepiensis, five exoantigens had their epitopes analyzed in generated 3-D models. Thus, the exoantigens oxidoreductase (A0A1D2JMM5, 3 epitopes), endonuclease G mitochondrial (A0A1E2XUM2, one epitope), endo-1,3(4)-beta-glucanase (A0A1E2Y5K2, seven epitopes), ATP synthase subunit D (A0A1E2YCS7, five epitopes), and carbonic anhydrase (A0A1E2YFV5, two epitopes) were analyzed, evidencing that all predicted epitopes were present in the exposed regions in the 3-D models ( Figure 7D).

Exclusive Epitopes of Members of the Paracoccidioides Complex
After identification of the exclusive epitopes of Paracoccidioides, the level of homology of these epitopes was investigated among species. Thus, the FASTA sequences were obtained of proteins of the Paracoccidioides species that were homologous to exoantigens identified during the immunoproteomics assays. For this analysis, homologous molecules of the species P. lutzii, P. americana, P. brasiliensis, P. restrepiensis, and P. venezuelensis were selected. ClustalX software was used to perform the alignment and the homology of these epitopes in the Paracoccidioides complex was manually analyzed. Consequently, non-homologous epitopes to other pathogens were identified that were exclusive to each species of Paracoccidioides (Table 3).
In P. lutzii, seven epitopes from two proteins, hypothetical protein (PAAG_05807) and hypothetical protein (PAAG_12630),   were identified (Table 3). In P. brasiliensis, three predicted epitopes were exclusive. Two epitopes were present in malate dehydrogenase NAD-dependent (PADG_08054) and one in endo-1,3(4)-beta-glucanase (PADG_12370) ( Table 3). For P. restrepiensis, two epitopes were identified and belong to ER lumen protein retaining receptor 2 (A0A1E2Y490) ( Table 3). Identification of these epitopes needs to be tested to be used during epidemiological monitoring of the disease, supporting the recognition of which species of Paracoccidioides is causing PCM.
FIGURE 7 | Prediction of the three-dimensional (3-D) model of the exoantigens of Paracoccidioides species by I-TASSER. The models were generated using the structure of the exoantigens identified during the immunoproteomics that presented non-homologous epitopes to other pathogenic organisms. The 3-D representation of predicted non-homologous epitopes was performed after structural refinement and energy minimization using ModRefiner and GalaxyWEB software. 3-D view of the non-homologous and hydrophilic epitopes predicted in (A) P. lutzii, (B) P. americana, (C) P. brasiliensis, and (D) P. venezuelensis.

DISCUSSION
In this work, several exoantigens were identified in the representatives of the Paracoccidioides complex: P. lutzii, P. americana, P. brasiliensis, and P. restrepiensis. These data provide the first large-scale identification of exoantigens of Paracoccidioides species. Some of these exoantigens have already been identified and described as antigens of pathogenic fungi, such as formamidase (Borges et al., 2005, dihydrolipoyl dehydrogenase (Landgraf et al., 2017), HSP60 (Izacc et al., 2001;Cunha et al., 2002), HSP70 (Bromuro et al., 1998;Bisio et al., 2005), serine protease Lacerda Pigosso et al., 2017), and isocitrate lyase (Cruz et al., 2011;Silva et al., 2019). This demonstrates the robustness of method used in our immunoproteomics analyses. Taking into account that our target is the characterization only of secreted antigens, we were able to produce and purify this group of proteins once no DNA from fungal cell lysis was detected in PCR-based assays. Similar results have already been observed in secretome analyses of Paracoccidioides sp. (Weber et al., 2012). Subsequently, we verified the profile of secretome obtained, where the molecular weight profiles were found to vary from 16 to 103 kDa, being compatible with the protein profile secreted by Paracoccidioides spp. (Weber et al., 2012;de Oliveira et al., 2018), which secreted a range of complex proteins of several sizes. Additionally, it was observed that Paracoccidioides species have differences in the exoantigen profiles, with an immunoreactive profile presenting differences between the studied species, although some exoantigens seem to be shared by all species. These antigenic differences may be related to the singular genetic compositions of each species analyzed, as reported (Pigosso et al., 2013;do Amaral et al., 2019). In Sporothrix spp., immunoblotting and immunoproteomics approach also have been demonstrating a dissimilar profile between species of complex, highlighting the importance of antigenic molecules as potential biomarkers of serological diagnosis and candidates for vaccine development (Rodrigues et al., 2015).
Currently, immunoprecipitation is a technique used for the purpose of purifying protein extracts, concentrating lowabundance proteins, analyzing protein-protein interactions, posttranslational modification studies, and can also be employed for antigens precipitation (Lund-Johansen et al., 2000). Employing this technique, we found the presence of immunoreactive bands when compared to control samples. Similar results were observed when immunoprecipitation was used for antigen purification with subsequent construction of new diagnostic tests for African trypanosomiasis (Sullivan et al., 2013). In addition, we found that the immunoreactive proteins depict a distinct pattern among the studied species, although some exoantigens were present in all Paracoccidioides species, as demonstrated by our proteomics analyses. It can be inferred that these differences can be related to the genetic background, mechanisms involved in the protein expression, as well as the secretion machinery of each Paracoccidioides species (Desjardins et al., 2011;Pigosso et al., 2013;de Oliveira et al., 2018;do Amaral et al., 2019;Peres da Silva et al., 2019). Thus, the use of this technique for the purification of antigenic molecules has been shown to be highly effective for obtaining exoantigens from Paracoccidioides species.
Additionally, we characterized the Paracoccidioides identified antigenic proteins in functional classes. The distribution of secreted proteins in the functional classes was consistent with previous analyses of secretome of Paracoccidioides spp. (Weber et al., 2012;de Oliveira et al., 2018) and with proteins secreted in vesicles in other pathogenic fungi (Albuquerque et al., 2008;Rodrigues et al., 2008Rodrigues et al., , 2014Oliveira et al., 2010). Also, we verified that the protein profile obtained, including mass range and isoelectric point (IP), is consistent with other Paracoccidioides sp. secretomes (Weber et al., 2012). Most of the identified antigenic proteins depicted prediction of subcellular localization in the cytoplasm and mitochondria, which is in accordance with described secretomes. These proteins should perform nonclassical functions at other cellular locations, being described as moonlighting proteins (Karkowska-Kuleta and Kozik, 2014). The secretion of these molecules seems to be related to virulence and/or host-pathogen interaction of Paracoccidioides spp. once it has been demonstrated that secretion of immunogenic molecules in other pathogens is implicated in virulence during hostpathogen interaction (Oliveira et al., 2010;Campbell et al., 2015;de Oliveira et al., 2018).
As expected, the identification of similar exoantigens among Paracoccidioides species was confirmed by mass spectrometry analysis. When comparing the identified proteins, some molecules, such as malate dehydrogenase NAD-dependent, HSP70-like protein, and 2-methylcitrate synthase, were present in all species. On the other hand, several exoantigens were identified as unique in the analyzed species. Previous analysis has shown constitutively and differentially expressed molecules between Paracoccidioides species (Pigosso et al., 2013;do Amaral et al., 2019). However, analyses of exclusive proteins among the isolates could possibly be used in the differential diagnosis of PCM, as well as epidemiology studies.
Homology analyses were used to identify unique exoantigens of Paracoccidioides species. These analyses were initially employed among the exoantigens identified by nanoUPLC-MS E , resulting in the identification of exclusive exoantigens to P. lutzii. The identification of these unique exoantigens enables its use in the differential diagnosis of PCM caused by P. lutzii. Subsequently, all exoantigen sequences were compared to sequences of other fungal and bacterial pathogens to verify the homology of these molecules to other organisms. These homology analyses were also employed on human sequences (Virginio et al., 2014;Rodrigues et al., 2015), aiming to select non-homologous molecules and to avoid possible cross-reactivity in future diagnostic tests. However, a small percentage of the exoantigens analyzed presented low homology, being represented by P. lutzii and P. americana species. Because homology analysis is performed throughout the exoantigen sequence, we decided to check the homology singly analyzing only the homology of the predicted epitopes. Thus, we identified several exclusive and conserved epitopes in the Paracoccidioides complex, since these molecules are candidates for the specific diagnosis of PCM. Similar analyses have already been performed in other works, where linear B-cell epitopes were employed for serological detection of several pathogenic organisms, such as Borrelia miyamotoiis (Tokarz et al., 2019), Toxoplasma gondii (Gatkowska et al., 2019), Treponema pallidum (Ogden et al., 2019), Leishmania spp. (Bremer Hinckel et al., 2019;Vale et al., 2019), and several human viruses (Amrun et al., 2019;Mutsvunguma et al., 2019;Yao et al., 2019). This reinforces the importance of these analyses, demonstrating the applications of linear B-cell epitopes in the diagnosis of human pathogenic organisms.
Similar to findings for other fungi, the cell wall of Paracoccidioides spp. is composed of a network of glycoproteins and polysaccharides, which act to protect the cell from environmental stressors (De Groot et al., 2005). Recent studies have demonstrated that both the yeast and mycelial phases of Paracoccidioides spp. present different amounts of β-(1,3) glucans (Camacho and Nino-Vega, 2017). However, enzymes, such as endo β-(1,3)-glucanase, act by hydrolyzing the β-glucan chains, making these enzymes extremely important because they play a fundamental role in the morphological processes of the fungal cell (Adams, 2004). Endo-1,3(4)-beta-glucanase was found as an antigen secreted in yeasts of P. americana, P. brasiliensis, and P. restrepiensis during our analyses. This enzyme has also been detected as secreted in yeast proteomes of H. capsulatum (Garfoot et al., 2017). In addition, the endo-1,3(4)-beta-glucanase showed low homology when compared to other fungi and pathogenic bacteria. Despite the low homology, we decided to investigate the level of homology between the epitopes of this molecule. During analysis, seven specific B-cell epitopes presented as exclusive for the Paracoccidioides complex, being excellent candidates for synthesis of new diagnostic tests for PCM.
Interestingly, the carbonic anhydrase (CA) of Paracoccidioides spp. was first reported as a secreted antigen in the present study. This is a metalloenzyme and acts by catalyzing the reversible hydration of CO 2 to generate HCO − 3 (Supuran, 2008). Additionally, in Paracoccidioides sp., four coding genes for carbonic anhydrase have already been described, leading to the synthesis of CA1, CA2, and CA3 proteins for Class β and CA4 protein for Class α. In addition, this enzyme was positively regulated in Paracoccidioides yeast when compared to the mycelium phase of the fungus (Tomazett et al., 2016). In Stenocarpella macrospora, CA has already been described in the mitochondrial, cytoplasmic, and extracellular compartments (Elleuche and Poggeler, 2010). CA has been identified as an exoantigen in the current work and has been associated with the extracellular compartment. In addition, CA presented exclusive epitopes, which makes this molecule a potential candidate as a PCM biomarker. Additionally, this exoantigen can be characterized as a moonlighting protein because it has a metabolic function in the cytoplasm and, when exported to the extracellular environment, it is probably related to the virulence of the fungus. In addition, CA has been described and is related to morphogenesis and growth of Paracoccidioides sp. and C. neoformans (Bahn et al., 2005;Costa et al., 2007) during the infectious processes, suggesting the importance of this molecule during hostpathogen interactions.
Aminopeptidase 2 was identified as an exoantigen during analyses. These molecules are characterized as exopeptidases and act as cleaving proteins in their N-terminal region (Gonzales and Robert-Baudouy, 1996). In Aspergillus oryzae mycelium secretome, aminopeptidase was related to collagen degradation processes (Ding et al., 2014). Furthermore, in a recent study, aminopeptidase demonstrated its antigenic potential during the expression of recombinant proteins of Fasciola hepatica when probed against antibodies from patients with fascioliasis (Mirzadeh et al., 2018) and against anti-aminopeptidase monoclonal antibodies from Taenia pisiformis (Zhang et al., 2018). The antigenicity of this molecule was also attested to in eggs and adult worms of Schistosoma mansoni following mouse infection assays (Maggioli et al., 2018). In Aspergillus terreus mycelium exoantigens, this molecule showed a high antigenicity when probed against monoclonal antibodies, demonstrating its antigenic profile (Nayak et al., 2011). During analyses, these molecules were identified in P. americana and P. brasiliensis, where one epitope was characterized as exclusive of the Paracoccidioides complex. These characteristics indicate that this molecule may be an excellent candidate to be tested as a potential specific biomarker for Paracoccidioides species infections.
Therefore, further analyses are underway to investigate the potential of these molecules as biomarkers, which may be used in the design of a rapid diagnostic test for PCM, in patient treatment follow-ups, as well as to investigate the immunogenic potential of these molecules and their possible use in the therapy of this important neglected human systemic mycosis.

CONCLUSION
The differences among exoantigen expression profiles were essential to identify specific biomarkers for each Paracoccidioides species. Here, the use of an immunoproteomics approach allowed us the characterization of 15, 14, 17, and 33 exoantigens in P. lutzii, P. americana, P. brasiliensis, and P. restrepiensis, respectively. Additionally, bioinformatics analyses made it possible to perform a series of other characterizations, allowing us to verify the biological functions, homology of exoantigens, epitopes prediction, as well as strengthen the data accuracy. Two exoantigens were identified and described in this work as unique in P. lutzii. Also, 44 epitopes exclusive to the Paracoccidioides complex were mapped using bioinformatics. Regarding the epitopes that could be used for epidemiological monitoring of the disease, a total of 12 antigenic sequences were identified in exoantigens in the four Paracoccidioides species studied. Therefore, these findings demonstrate that Paracoccidioides species have a range of epitopes exclusive to the complex, as well as specific to each fungal species. Finally, the current serological tests could be improved further with combinations of two and more synthetic peptides, as well as modifications that can optimize these tests. It is important to highlight that the reactivity of these proposed epitopes is under evaluation and will be tested against clinical samples to provide a new tool for diagnosis, patient follow-ups, and/or PCM therapy.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation, to any qualified researcher.

ETHICS STATEMENT
The animal study was reviewed and approved by the Comissão de Ética no Uso de Animais (CEUA-UFG) under the registry number 030/2016.

SUPPLEMENTARY MATERIAL
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fmicb. 2019.02968/full#supplementary-material FIGURE S1 | Affinity chromatography obtained by immunoprecipitation. (A) Scheme used for purifying and obtaining the antigens using serum from immunized and control animals. (B) Scheme demonstrating the interaction of sepharose bead -G protein bound antibody and antibody antigen interaction. FIGURE S2 | Confirmation by PCR of the extracts secreted by Paracoccidioides species. The sensitivity of the PCRs to the formamidase gene were obtained using the genomic DNA of P. lutzii, P. americana, P. brasiliensis, and P. restrepiensis in five dilutions ranging from 50 ng to 1 pg (Top panel). C−: Negative control (without genomic DNA). Sec: Cell-free secretome. bp: Base pairs according to the molecular weight marker. The amplicons obtained by PCR were evaluated by 1% agarose gel electrophoresis and stained by GelRed DNA intercalator.
FIGURE S3 | Integrity analysis of the proteome profile of the secreted isolates of Paracoccidioides species by one-dimensional electrophoresis. Twenty micrograms of samples were separated by one-dimensional electrophoresis (SDS-PAGE) at 12%. P. lutzii, P. americana, P. restrepiensis, and P. brasiliensis. kDa, Kilodaltons. Coomassie Blue staining was used to visualize the proteins.
TABLE S1 | Identification of exoantigens recognized by serum from animals immunized with P. lutzii secretome.