Identification and validation of specific B-cell epitopes of hantaviruses associated to hemorrhagic fever and renal syndrome

Background Orthohantavirus infection is a neglected global health problem affecting approximately 200,000 people/year, spread by rodent hosts and associated to fatal human diseases, such as hemorrhagic fever with renal syndrome (HFRS) and orthohantavirus cardiopulmonary syndrome (HCPS). Circulation of HFRS-associated orthohantaviruses, such as Seoul, Gou, Amur, Dobrava and Hantaan, are supposed to be restricted to Eurasian countries even though their hosts can be a worldwide distribution. Few confirmed HFRS orthohantavirus infections in humans have been reported in American countries, but due to lower medical awareness of the symptoms of this zoonosis, it could be associated to viral underreporting or to misdiagnosis with several tropical hemorrhagic diseases. Serological evidence of orthohantavirus infections, using enzyme-linked immunosorbent assay for the presence of immunoglobulin M and G against recombinant nucleoprotein protein, remains as an essential assay for viral surveillance. In this study, we aimed to identify in silico immunogenic B-cell linear epitopes present on orthohantavirus nucleoprotein that are exclusive to HFRS-related species. Methodology/Principal findings In silico analysis were performed using Seoul orthohantavirus nucleoprotein (SHNP) sequence as a model. Linear B-cell-epitopes on SHNP and its immunogenicity were predicted by BepiPred-2.0 and Vaxijen algorithms, respectively. The conservancy of predicted epitopes was compared with the most clinically relevant HFRS or HCPS-associated orthohantavirus, aiming to identify specific sequences from HFRS-orthohantavirus. Peptide validation was carried out by ELISA using Balb/c mice sera immunized with purified recombinant rSHNP. Peptides cross-reactivity against HCPS orthohantavirus were evaluated using immunized sera from mice injected with recombinant Juquitiba orthohantavirus nucleoprotein (rJHNP). Conclusion/Significance In silico analysis revealed nine potential immunogenic linear B-cell epitopes from SHNP; among them, SHNP(G72-D110) and SHNP(P251-D264) showed a high degree of sequence conservation among HFRS-related orthohantavirus and were experimentally validated against rSHNP-IMS and negatively validated against rJHNP-IMS. Taken together, we identified and validated two potential antigenic B-cell epitopes on SHNP, which were conserved among HFRS-associated orthohantavirus and could be applied to the development of novel immunodiagnostic tools for orthohantavirus surveillance.


Background
Orthohantavirus infection is a neglected global health problem affecting approximately 200,000 people/year, spread by rodent hosts and associated to fatal human diseases, such as hemorrhagic fever with renal syndrome (HFRS) and orthohantavirus cardiopulmonary syndrome (HCPS). Circulation of HFRS-associated orthohantaviruses, such as Seoul, Gou, Amur, Dobrava and Hantaan, are supposed to be restricted to Eurasian countries even though their hosts can be a worldwide distribution. Few confirmed HFRS orthohantavirus infections in humans have been reported in American countries, but due to lower medical awareness of the symptoms of this zoonosis, it could be associated to viral underreporting or to misdiagnosis with several tropical hemorrhagic diseases. Serological evidence of orthohantavirus infections, using enzyme-linked immunosorbent assay for the presence of immunoglobulin M and G against recombinant nucleoprotein protein, remains as an essential assay for viral surveillance. In this study, we aimed to identify in silico immunogenic Bcell linear epitopes present on orthohantavirus nucleoprotein that are exclusive to HFRSrelated species.

Methodology/Principal findings
In silico analysis were performed using Seoul orthohantavirus nucleoprotein (SHNP) sequence as a model. Linear B-cell-epitopes on SHNP and its immunogenicity were predicted by BepiPred-2.0 and Vaxijen algorithms, respectively. The conservancy of predicted epitopes was compared with the most clinically relevant HFRS or HCPS-associated orthohantavirus, aiming to identify specific sequences from HFRS-orthohantavirus. Peptide validation was carried out by ELISA using Balb/c mice sera immunized with purified recombinant rSHNP. Peptides cross-reactivity against HCPS orthohantavirus were

Introduction
Orthohantaviruses belong to genus Orthohantavirus, of the family Hantaviridae, which comprises a diverse group of infectious negative-sense single-stranded RNA viruses. They are zoonotic in nature and transmitted to humans through the inhalation of contaminated aerosols present in the excreta, saliva, and urine of infected rodent animals [1]. Orthohantavirus infections are a neglected global health problem, affecting approximately 200,000 people/year [1], and are associated with serious and fatal diseases in humans, such as hemorrhagic fever with renal syndrome (HFRS) and orthohantavirus cardiopulmonary syndrome (HCPS). The clinical manifestations of the disease are correlated to orthohantavirus species/genotypes and also it is intimately associated with rodent reservoirs that have a specific geographical distribution worldwide. HFRS is caused by Old World orthohantaviruses such as Hantaan, Amur and Gou virus in Asia and by Seoul orthohantavirus (SEOV) worldwide. Besides, orthohantavirus New World species such as Sin Nombre orthohantavirus in North America, and Andes, Juquitiba, Laguna Negra, and other related virus in Latin America, such as Juquitiba genotype are responsible for the more severe HCPS [2][3][4][5].
Among hantavirus associated to HFRS, the SEOV is spread worldwide by Rattus norvegicus and Rattus rattus, two rodent species present on the five continents, which are also implicated in leptospirosis outbreaks [6]. In Americas, we highlight that SEOV has already been detected in both rodents, R. norvegicus and R. rattus, in Argentina [7], United States [8] and Brazil [7,9]. Nonetheless, few cases of HFRS have been reported in Brazil [10], where this syndrome could be misdiagnosed as other tropical infectious diseases like dengue, yellow fever and leptospirosis. This data combined with the absence of a specific surveillance for SEOV and others HFRS-associated orthohantavirus in Western countries, and with the inability to confirm cases on site, by laboratory diagnosis, make urgent the development of new diagnostic tools.
In this context, the orthohantavirus nucleoprotein (NP) is the major antigen that elicits early serological responses in infected humans and has been used as a biomarker to develop antibodies for epidemiological surveillance in regions where various orthohantavirus species cocirculate [11,12]. This protein consists of about 429-amino acid residues, is largely expressed in initial infection and is highly conserved among orthohantavirus [13]. Since the aminoacid sequence of HFRS-associated orthohantavirus are not identical, previous works tried to develop immunochromatographic antibody tests using a combination of different orthohantavirus nucleoproteins [14] or a combination of nucleoprotein and other peptides [15] in order to detect all Eurasian pathogenic orthohantavirus infections. Recently, Kalaiselvan and collaborators suggested that in silico predicted B-cell epitopes on NP could be used to identify specific markers to the HFRS [16] and HCPS-associated orthohantavirus [17]. However, the experimental validation of these epitopes remains unexplored. By this way, this study aimed to predict B-cell epitopes on Seoul orthohantavirus nucleoprotein (SHNP), that are exclusively conserved among HFRS-associated orthohantavirus, using a combination of the three most used prediction algorithms, and to validate the epitopes predicted as specific to SEOV or as conserved among HFRS orthohantavirus, against sera of mice immunized with recombinant SHNP.

Sequence data
To predict possible antigenic properties and the three-dimensional (3D) structure of SHNP (Seoul virus BjHD01, Accession Number AY627049.2) using bioinformatic tools, the entire sequence of SHNP (NCBI ID: AAT45728.1) was downloaded from the NCBI website (www. ncbi.nlm.nih.gov/protein) and used for analyses.
Firstly, the prediction of linear B-cell epitopes was carried out using the web server Bepipred 2.0, which uses a Random Forest algorithm trained on epitopes and non-epitope amino acids determined by crystal structures from a protein sequence. For each FASTA input sequence, the server outputs a prediction score for each amino acid. To determine potential Bcell linear epitopes, we utilized the recommended cutoff of 0.5, ensuring a specificity of 57% and sensibility of 59% [18]. Therefore, the Bepipred score represents the average of the scores of at least nine consecutive amino acids above the cut-off. Sequences with BepiPred score above 0.5 were considered as potential linear B-cell epitopes and analyzed by Vaxijen.
VaxiJen is the first server for alignment-independent prediction of protective antigens. It was developed to allow antigens classification solely based on the physicochemical properties of proteins without recourse to sequence alignment. Bacterial, viral and tumour protein datasets were used to derive models for prediction of whole protein antigenicity, showing prediction accuracy from 70% to 89% [19]. To evaluate the antigenicity of predicted epitopes, we used the default cut-off (0.4), suggested to viral antigens. Therefore, sequences with BepiPred score above 0.5 and Vaxijen score above 0.4 were considered potential linear B-cell epitopes and evaluated to specificity.

Evaluation of conservation degree of linear B-cell epitopes
Sequences identified as potential linear B-cell epitopes were aligned to nucleoprotein sequences of orthohantavirus associated to HFRS or HCPS for similarity comparison. and Sin Nombre virus (SNV; Uniprot ID: Q89462) were the HCPS selected viruses for the conservation degree comparison of the predicted epitopes. All NP sequences were aligned and compared to SHNP and the degree of conservation among them were evaluated by BioEdit Sequence Alignments Editor, Version 7.0.9.0 [20].

Peptide synthesis
After consensus analysis of the in silico prediction tools, the selected sequences were synthesized by fluorenylmethoxycarbonyl (F-moc) solid-phase chemistry [21,22] (GenOne Biotechnologies, Brazil). Analytical chromatography of the peptide demonstrated a purity of >95% and mass spectrometric analysis also indicated estimated masses corresponding to the molecular masses of the peptides (SHNP (

Expression and purification of the recombinant hantavirus nucleoproteins
The cDNA codifying the Seoul or Juquitiba orthohantavirus genotypes nucleoprotein (429 and 428 amino acids, respectively) were cloned into plasmid pET21a(+) to express the nucleoprotein with a six-histidine tag on its C-terminus (Genscript, USA). This construct was used to transform into BL21DE3 strain from E. coli and the protein was expressed during 4 h upon the addition of 1 mM IPTG. The cells were pelleted by centrifugation at 5,000 g for 15 min at 4 o C and stored at -80 o C until use. The cellular pellet was resuspended using the following buffer composition: 50 mM sodium phosphate, pH 6,0, 500 mM NaCl, 1 mM β-mercaptoethanol, 5% glycerol, 10 mM imidazol, EDTA-free protease inhibitor cocktail (Roche, Germany). The cellular lysis was obtained by the addition of 1 mg/mL lysozyme under stirring at 4˚C for 30 min, Next, 20 μg/mL DNase and 2 mM MgCl 2 were added and the solution was incubated for more 30 min. The lysate was sonicated 15 times, alternating 15-seg on-and-off cycles at 4˚C. After sonication, the lysate was then centrifugated at 27000 g for 20 min to obtain the protein pellet inclusion body. This pellet was dissolved in buffer containing 50 mM sodium phosphate, pH 6,0, 500 mM NaCl, 1 mM β-mercaptoethanol, 5% glycerol, 10 mM imidazol and 8 M urea, overnight stirring at 4˚C. The cleared lysate was loaded onto a HisTrap HP column connected to the Ä kta Purifier HPLC. Proteins were eluted with a linear gradient of imidazole (10-500 mM) prepared in the same buffer composition and the purified nucleoprotein was refolded by dialysis using a cut-off pore of MWCO of 3,500 Da and the same buffer composition during overnight at 4˚C, making three changes using 1-L buffer each time. Recombinant SHNP or JHNP (rSHNP and rJHNP, respectively) were obtained with high purity and yield as analyzed by 15% SDS-PAGE and, as expected, a unique band of about 50 kDa was observed. The refolding process of the rSHNP or rJHNP was confirmed by fluorescence spectroscopy, which was used to monitor the intrinsic tryptophan fluorescence in the buffer containing 8M urea and after the dialysis process.

Ethics statement
This study was conducted in accordance with the Brazilian National Animal Care Ethical Council. All experiments were approved and permitted by the Ethics Committee for Animal Experimentation of the Oswaldo Cruz Foundation (CEUA-FIOCRUZ, Protocol N˚LW-13/ 16).

Experimental mice immunization
BALB/c mice (6 weeks old; n = 6) were injected intraperitoneally at day 0 with an emulsion (1:1) containing 50 μg of purified rSHNP or rJHNP, expressed in Escherichia coli, in a final volume of 100 μL with complete Freund´s adjuvant (CFA). Fourteen and twenty-eight days after initial immunization, mice were re-injected intraperitoneally with an emulsion (1:1) containing 50 μg of purified rSHNP or rJHNP in a final volume of 100 μL with incomplete Freund´s adjuvant (IFA).

Collection of the templates, in silico homology modeling and model optimization
The rationale used for template selection and in silico homology modeling of SHNP was described by Lima et al. (2016) [23] and carried out with some adaptations. Briefly, for the selection of templates, a search with BLAST and HHBlits [24,25] using the amino acid sequence of SHNP (GenBank: AAT45728.1) was performed against the expasy SWISS-MO-DEL template server [26][2] [2]. Three structures were selected (PDB ID: 5E04, 5FSG, 4FI5) with the view to obtain as close as possible complete coverage of the amino acid sequence. The server-generated models, which were built and based on the target-template alignment with ProMod3 Version 1.2.0 [27], were used to perform the 2D alignment with the sequence of SHNP and comparative modeling using MODELLERv9.19 by satisfaction of spatial restraints [28]. The lowest energy model was selected using PyMOL Version 1.8 and your 3D structure evaluated with Verify 3D [29,30] and MolProbity [31]. The energy minimization and correction of Ramachandran outliers was performed manually using COOT Version 0.8.6 [32].

The major of predicted B-cell epitopes in HNSP were highly conserved among orthohantaviruses
Nucleoprotein is a conserved protein among orthohantavirus, presenting similarity level ranging from 60% to 99%. However, when compared to orthohantavirus related to HCPS (Juquitiba genotype, Andes virus, Laguna-Negra virus and Sin Nombre virus), the nucleoprotein from orthohantavirus related to HFRS (Gou, Hantaan, Amur, Dobrava-Belgrade and Puumala) were only intermediary conserved, ranging from 62% to 74% of similarity (S1 Table).
Based on this variable degree of conservation of nucleoprotein and aiming to identify immunodominants B-cell epitopes specific to orthohantaviruses related to HFRS, we compare the similarity level of SHNP predicted epitopes among nucleoproteins of HFRS Hantavirus and HCPS Hantavirus. The levels of similarity were classified as very low (<25%;); low (25%> X>50%), intermediary (50%<X<75%) and high (75%<X<100%). The similarity level observed among the epitopes of each orthohantavirus compared was shown in Table 2. Firstly, when compared to other orthohantavirus nucleoproteins, while SHNP presented only intermediary and high similarity levels (62% to 99%), we found B-cell epitopes from very low to high similarity (11% to 100%). The majority of predicted linear Bcell epitopes on SHNP were highly conserved in comparison with orthohantavirus associated with HFRS and to orthohantavirus associated with HCPS. The epitopes SHNP (G145-K173) , SHNP (P182-G196) , SHNP (T347-F361) , SHNP (Q363-I380) , SHNP (D394-E405) presented an conservation degree ranged from 67% to 100%, being the sequence SHNP (Q363-I380) the most conserved epitope, with an similarity level of 100% when compared to the most of orthohantavirus associated to HFRS and 94% when compared to virus or isolates associated with HCPS (Table 2). Besides, the predicted epitopes SHNP (K30-E48) and SHNP (G72-D110) presented high conservation degrees (ranged from 84% to 95% and 77% to 97%, respectively) in comparison with to the most of orthohantaviruses associated to HFRS and intermediary degree (ranged from 47% to 53% and 51% to 56%, respectively) when compared to sequences of virus or isolates associated to HCPS. Remarkably, the predicted epitopes SHNP (P182-G196) and SHNP (P251-D264) were considered specific to Seoul orthohantavirus, once both presenting 100% of similarity with Gou specie and low level of conservancy (ranged from 14% to 21% and 11%; respectively) in comparison with orthohantavirus associated HCPS.

Homology modeling
Search of homologous SHNP structures (GenBank: AAT45728.1) were carried out using the SWISS-MODEL server, using information from Protein Data bank (PDB), and resulted in three protein structures which were used for homology modeling: 5FSG, 4FI5 and 5E04 with sequence identity of 82.91%, 84.62% and 66.08%, respectively. The complete template description is shown in Table 3.
From the alignment of the target sequence with the template sequences, 200 models were built and the lowest energy model was selected. The Molprobity server and WinCoot software was used to evaluate the stereochemical quality of the selected model from Ramachandran plots. After energy minimization, it was possible to reach 97.19% (415/427) of residues in preferred regions, 2.58% (11/427) in allowed regions and only 0.23% (1/427)-114 GLY-outliers in Ramachandran graph. The 3D structure of the model presented a score of 75.52% on Verify 3D, which demonstrates the quality of the model obtained.
SHNP homology model and the prediction of its electrostatic surface are shown in Figs 3A and 2B. The spatial location and shape of predicted linear peptides-SHNP (G72-D110) (Fig 3C), SHNP (P251-D264) (Fig 3D) and SHNP (I266-A283) (Fig 3E)-were demonstrated in the model generated. When performing an analysis of the location of the peptides in front of the electrostatic surface of the predicted protein, it is possible to verify that the peptide SHNP (G72-D110) (Fig 3C) is at a more electronegative region (red) than SHNP (P251-D264) (Fig 3D) and SHNP (I266-A283) (Fig 3E) peptides, which are in a more neutral/electropositive region. Bars represent the optical densities of mice immunized with rSHNP (red bars, n = 3) and immunized with rJHNP (blue bars, n = 3) and non-immunized mice (green bars, n = 6). Traced line indicates the value of Cut-off, defined as 3 times the mean of non-immune optical densities. Data is presented as floating bars with lines indicate minimum, mean and maximum values. Values higher than cut-off were considered responder to antigen and compared by test-T against other groups. � represents difference between immunized groups and non-immunized mice, while + represents statistical differences between immunized groups. https://doi.org/10.1371/journal.pntd.0007915.g002

Discussion
Orthohaantavirus species associated with hemorrhagic fevers are worldwide spread; however, they could be misdiagnose as viral or bacterial hemorrhagic diseases, like dengue, yellow fever or leptospirosis, especially in countries where these diseases are endemic. In this context, the development of novel, specific and broader diagnostic tools, able to recognized HFRS-associated orthohantavirus species, have become an urgent demand to public health. Immunodiagnostic tools are vital in the study of orthohantavirus infections due to the limited success with molecular detection as the viremia will be short-lived. Based on this, our study allowed the in silico identification and experimental confirmation of two B-cell linear epitopes on SHNP. These data could be helpful to improve the development of new diagnostic tools, resulting in a better and more specific HFRS-associated orthohantavirus diagnose, especially in the New World.
Recently, Kalaiselvan and collaborators suggested that 20-mers peptides in silico identified as B-cell linear epitopes may be used for development of geographic region-specific immunoassays [16], however, the experimental validation of predicted epitopes remains pending. Based on this study, we explored the nucleoprotein of Seoul orthohantavirus, using a combination of in silico approaches to predict antigenic B-cell linear epitopes, non-conserved among Identification of B-cell linear epitopes present on Seoul orthohantavirus nucleoprotein hantaviruses associated to HCPS, and to validate its immunogenicity by the reactivity of serum samples from mice immunized with the recombinant protein. Initially, we predicted nine sequences as immunogenic epitopes, in which eight (88.9%) were, entirely or partially, inserted on sequences described in Kalaiselvan's study [16]. In our point of view, the differences observed between our and his study about predicted epitopes were due the use of BepiPred 1.0, by Kalaiselvan, and BepiPred 2.0, by us. However, the high congruence of both studies reinforces the use of prediction algorithms to identify B-cell epitopes on proteins and remarks the necessity of experimental validation of these results.
Aiming to identify epitopes non-conserved among HCPS orthohantavirus, we compared the conservation of each predicted sequence among orthohantaviruses associated to HFRS and HCPS. The epitopes SHNP (K30-E48) , SHNP (G72-D110) , and SHNP (I266-A283) were highly conserved just among species associated to HFRS, presenting a conservation level ranging from moderate to low, when compared to HCPS orthohantavirus species. Besides, the epitope SHNP (P251-D264) was considered specific epitope of Seoul orthohantavirus, presenting a very high similarity when compared with Gou (100%) and low conservancies (<36%) in comparison with species associated to HCPS or HFRS. Remarkably, all epitopes considered non-conserved among HCPS orthohantavirus (SHNP (K30-E48) , SHNP (G72-D110) , SHNP (P251-D264) and SHNP (I266-A283) ) presented low to moderate conservation degrees, ranging from 17% to 59%, when compared with Puumala nucleoprotein sequence. Despite of the association of this orthohantavirus specie with HFRS [33][34][35], Puumala orthohantavirus was also associated with HCPS [36] and is considered the most prevalent pathogen associated with nephropathia epidemica [37], a mild form of HFRS. This epidemics is characterized by acute kidney injury (AKI) and thrombocytopenia [5], suggesting differences among species associated to HFRS. Besides, we observed that major of predicted sequences were highly conserved among orthohantavirus associated to both syndromes, HFRS and HCPS. Our data corroborate Kalaiselvan's previous studies [16,17], in which epitopes conserved between both orthohantavirus species, HFRS and HCPS, were in silico predicted with prevalence in C-terminal region of the nucleoprotein.
Despite the increasing use of in silico approaches to predict epitopes against several diseases, the experimental validation of predicted sequences remains scarce, despite being a critical step to the development of novel vaccines and diagnostic tools. In this context, this was the first study reporting the experimental validation of specific B-cell linear epitopes against orthohantavirus, resulting in the recognition of SHNP (G72-D110) and SHNP (P251-D264) epitopes as promising targets to further development of monoclonal antibodies and/or highly-specific tools to HFRS orthohantavirus diagnosis. Moreover, about the non-reactivity of predicted epitope SHNP (I266-A283) and considering the use of in silico strategies to select B-cell epitopes, the oligomerization mode of selected target seems to be an important aspect to epitope selection. Here, we initially predicted three specifics B-cell epitopes in HFRS orthohantavirus and evaluated their localization in tertiary monomeric structure of SHNP, considering all epitopes exposed in protein surface. Controversially, considering that nucleoprotein oligomerizes in dimeric or hexameric forms [3,38], we believed that the non-reactivity of epitope SHNP (I266-A283) could be explained by its localization in quaternary structure.
Currently, orthohantavirus commercial serological kits are based on nucleoprotein detection however the cross-reactivity among hosts infected by HCPS and/or HFRS-associated orthohantavirus could not be discarded. In our study, we experimentally immunized mice with recombinant nucleoprotein from Juquitiba genotype, a HCPS-associated orthohantavirus that mainly circulates in Brazil [39,40], eliciting a polyclonal response able to generate antibodies that recognized JHNP and SHNP (Fig 2). Based on this cross reaction and considering that Juquitiba genotype nucleoprotein is highly conserved among ANDV and SNV (95% and 86% of similarity, respectively), both orthohantavirus associated to HCPS in Americas, we believe that Juquitiba genotype could be used as HCPS model. Remarkably, despite the cross reaction among JHNP and SHNP, we identify epitopes specific to Seoul orthohantavirus. SHNP (G72-D110) and SHNP (P251-D264) epitopes were only recognized by sera from SHNP immunized mice, suggesting that these epitopes are non-conserved among HCPS-associated orthohantavirus and corroborating their potential as HFRS target. This strategy, based on experimental immunization assays using recombinant proteins, have been used to confirm in silico findings [41][42][43][44], however, considering the probable misdiagnose of Seoul orthohantavirus in Americas and the necessity of differential diagnosis of HFRS in New world, experiments using samples of naturally infected hosts (human and/or rodents) should be conducted in order to fully support the present data and to confirm the real potential of these epitopes as differential diagnostic tools.
In summary, we in silico identified and experimentally validated the sequences (72) GKNIGQDRDPTGVEPGDHLKERSALSYGNTLDLNSLDID(110) and (251)PIAGSLS GNPVNRD(264) as linear B-cell epitopes on SHNP, which could be helpful to improve the development of novel immunodiagnostic tools against HFRS-associated orthohantavirus species. These epitopes, as synthetic peptides, could be anchored on solid surface or membranes for enzyme immunoassays; or used in the development of monoclonal antibodies, applicable to tools like sandwich ELISA, immunohistochemistry and immunofluorescence tests, which may be relevant to better and more specific HFRS-associated orthohantavirus diagnose, especially in the New World.