An mRNA-based reverse-vaccinology strategy to stimulate the immune response against Nipah virus in humans using fusion glycoproteins

The zoonotic pathogen, Nipah virus, is considered a potential healthcare threat due to its high mortality rates and detrimental symptoms like encephalitis. Ribavirin, an antiviral drug helps in overcoming the number of casualties and reducing the mortality rate, but no long-lasting solution has been proposed yet putting global health security in jeopardy. Given the cognizance of mRNA-based vaccines as safe and efficacious preventative strategies against pathogens, the current study has utilized the reverse-vaccinology approach coupled with immunoinformatics to propose an mRNA-based vaccine candidate against the Nipah virus. To ensure the effectiveness of the vaccine candidate against all strains of Nipah and associated viruses, three fusion glycoproteins from Nipah and Hendra viruses were selected. A total of 30 potential epitopes, 10 B-cell-, 10 MHC-I-, and 10 MHC-II-specific, were screened for the construct. The finalized epitopes were highly antigenic with scores ranging from 0.75 to 1.7615 at a threshold of 0.4 for viruses and non-homologous to Homo sapiens eradicating any chance of immune tolerance. The construct, with a World population coverage of 97.2%, was structurally stable, thermo-stable, and hydrophilic with indices of 32.91, 93.62, and –0.002, respectively. The vaccine candidate’s tertiary structure was predicted with a TM score of 0.131 and the refined model displayed superlative RAMA improvement (98.2) and MolProbity score (0.975). A quality factor of 93.5421% further validated the structural quality and stability. A prompt and stable immune response was also simulated, and the vaccine candidate was shown to eliminate from the body within the first five days of injection. Immune complexes count of 7000 mg/mL was predicted against the antigen with a small but non-significant danger signal, countered by the cytokines. Lastly, strong molecular interactions of the vaccine candidate with TLR-3 (331.09 kcal/mol) and TLR-4 (-333.31 kcal/mol) and molecular dynamics simulation analysis authenticated the immunogenic potential of the vaccine candidate. This vaccine candidate can serve as a foundation for future in-vitro and in-vivo trials to minimize or eradicate the diseases associated with the Nipah virus or the Henipaviral family.


INTRODUCTION
Nipah Viral Disease is a zoonotic disease that has a high mortality rate of 91% (Pillai et al., 2020).The disease first emerged in 1998 in Malaysia and spread to neighboring countries including South-East Asia, like Bangladesh, Singapore, and India (Aditi & Shariff, 2019).It was isolated and identified at the University of Malaysia, Faculty of Medicine in 1999.As this virus was first spread in the Malaysian village Sungai Nipah, in pig farms and people associated with them, so scientists named this virus as Nipah virus (Pillai et al., 2020).This virus belongs to the Paramyxoviridae family and genus Henipa viruses.This virus is classified into two strains, the first one was isolated from Malaysia and Cambodia, and the other one was isolated from India and Bangladesh (Mourya et al., 2018).This virus leads to the development of rapid progressive illness in the human respiratory tract and causes encephalitis in the brain.Initially, because of the encephalitis-related symptoms, scientists thought of this disease as Japanese encephalitis, but it was confirmed later that the Nipah virus causes a different disease (Singh et al., 2019).The main host reservoirs for Henipa viruses are fruit bat Pteropus.These bats manipulate different fruits like date palms by secreting saliva and urine that is further consumed by animals.The most recent outbreak ruptured in Kerala, India with a loss of 17 human lives (Weingartl et al., 2009).The Nipah virus causes severe breathing issues, destroying the human lung structures and the membrane of the cerebrum.In this condition, the brain enlarges, causing memory loss and severe pain (Rockx et al., 2010).Its symptoms include brain encephalitis, headache, muscular pain, and respiratory disorders.Nausea, giddiness, and fever are also clinically significant (Reddy, 2018).According to the Centers for Disease Control (CDC), the Nipah virus has the potential to create a global emergency and, hence, deserves the immediate attention of the scientific community (Fischer et al., 2018).Ribavirin, an antiviral drug helps in overcoming the number of casualties and reducing the mortality rate, but no long-lasting solution has been proposed yet (Deb et al., 2019).Vaccines are an efficacious solution to fight this deadly viral agent.mRNA-based vaccines are among the most effective candidates utilized to develop protective immunity against the pathogen.The vaccine shots activate the immune system, and the B-cells and T-cells develop immunologic memory for when a real pathogen attacks the host the immune system recognizes and destroys it.The mRNA vaccines are further appreciated for being cost-effective and timesaving (Kumar et al., 2015).With computational approaches, the labor intensity and process complexity are minimized (Naveed et al., 2022;Naveed et al., 2023a;Naveed et al., 2023b), making mRNA vaccines worthy solutions against viruses like the Nipah virus.
The current work is based on the proposition of an mRNA vaccine utilizing in-silico and immunoinformatics tools.The computational approach is considered to safely predict the immune response against the proposed vaccines.Three transmembrane proteins of the Nipah virus were selected for their significant role in the early interactions with the hosts.The selection of the transmembrane proteins was based on their antigenicity, allergenicity, non-toxicity, and non-resemblance with human proteins.Potential epitopes following the criteria of Naveed and others (Naveed et al., 2022a;Naveed et al., 2022b) are shortlisted and the construct is evaluated based on its population coverage, immune response simulations, molecular docking with respective toll-like receptor(s), and physicochemical properties.

Selection of Target Proteins
The Nipah and Hendra Henipavirus genomes were retrieved from the NCBI database and downloaded in FastA format.CELLO2GO, an online server available at (http://cello.life.nctu.edu.tw/cello2go/), was utilized to study the gene ontology and localization of all the proteins present in the genomes.The conservation and homology evaluation was performed using blastp available at (https://blast.ncbi.nlm.nih.gov/Blast.cgi?PAGE=Proteins).It ensured that the selected proteins based on protein localization are non-homologous to the human host and are conserved across Henipaviruses.Lastly, the online servers Vaxijen2.0,available at (http://www.ddg-pharmfac.net/vaxijen/VaxiJen/Vaxi-Jen.html), and AllerTop, 2.0 accessed at (https://www.ddg-pharmfac.net/AllerTOP/method.html), were utilized to predict the antigenicity and allergenicity of the shortlisted proteins, respectively.

Protein Selection & Sequence Retrieval
The online server UniProt (www.uniprot.org)was used to retrieve the Nipah virus virulent protein sequences.Based on different factors like higher antigenic capacity, non-allergenicity, non-toxicity, and non-resemblance with Homo sapiens, three trans-membrane fusion glycoproteins (one belonging to Hendra virus and two belonging to the identified strains of Nipah virus) were selected for a potential vaccine candidate.

Prediction of the MHC-1 Epitopes
MHC-1 binding prediction tool of IEDB (http:// tools.iedb.org/mhci/)was used to predict the conserved epitopes with the NetMHCpan BA 4.0 version (Reynisson et al., 2020).MHC source specie was selected as human and epitopes were screened based on higher antigenicity value, non-allergenicity, and a specified range of IC 50 values lower than 100.

Prediction of MHC-II Epitopes
MHC-II binding prediction tool (http://tools.iedb.org/mhcii/) was used to predict the conserved epitopes.NetMHCIIpan 4.0 BA version (Reynisson et al., 2020) was utilized to predict the conserved epitopes.Epitopes were screened based on higher antigenicity value, nonallergenicity, and a specified range of IC 50 values lower than 100.

Population Coverage
The evaluation of population coverage of the selected MHC-I and MHC-II epitopes was carried out with the IEDB population coverage tool (http://tools.iedb.org/population/).All the selected epitopes with their corresponding alleles were included in the input file.On the input page, the number of epitopes was changed accordingly.A combined MHC-I and MHC-II epitopes input file was provided, and the World was selected as the target population.

Vaccine Construct
The finalized epitopes of B-and T-cells were reversetranslated and fused together with the help of linkers, adjuvants, a MITD sequence, polyA-tail, 5' and 3' UTRs, and methyl-guanosine cap at the start of the construct to enable it for in-host expression.

Analysis of Physiochemical Properties
The physicochemical properties of the vaccine construct were predicted by the ProtParam tool of the Ex-Pasy web server (https://web.expasy.org/protparam/).Toxicity prediction was done using Toxinpred.The Vaxijen tool was run to predict the antigenicity of the vaccine construct and to check the vaccine construct's allergenicity.
Prediction of Secondary Structure of the Proposed mRNA Vaccine PSIPRED (http://bioinf.cs.ucl.ac.uk/psipred/) was used for the prediction of the secondary structure.The mRNA vaccine candidate sequence was first translated to its primary protein sequence using EXPASY translate, and this primary sequence was used as an input sequence on PSIPRED.Out of a variety of analyses available on the server, the cartoon structure along with the basic prediction was interpreted.The tertiary structure of the mRNA was predicted using the online server trRosetta https://yanglab.nankai.edu.cn/trRosetta/.This online server for protein docking has been continuously extended and refactored.Its advantages include interoperability between modeling capabilities and performance (Leman et al., 2020).Refinement of the tertiary structure of our proposed mRNA vaccine was done by Galaxy WEB's Galaxy REFINE https://galaxy.seoklab.org/.

Validation of Tertiary Structure of the Proposed mRNA Vaccine
After refinement of the tertiary structure, we validated our proposed mRNA tertiary structure by Ramachandran plot analysis using the PROCHECK server (https://saves.mbi.ucla.edu/),whereas the structure quality was validated using ERRAT quality score on the PRO-CHECK server.

Immune Simulations
To validate whether the mRNA-based vaccine construct has the potential to elicit a stable immune response, C-IMMSIM analyses were performed through the webserver available at https://kraken.iac.rm.cnr.it/C-IMMSIM/index.php?page=1.An immunogenic response of single-dose injection was calculated.

Cloning of the Construct
The mRNA-based construct was converted into the DNA sequence after codon optimization using JCAT (http://www.jcat.de/Result.jsp).The optimized construct was cloned in the pBluescribe vector at the BsoB1 site with the removal of overhangs using Snapgene https:// www.snapgene.com/.The methodology is schematically represented in Fig. 1.

Target Proteins' Selection
Analyzing 2335 proteins present in the genomes of Nipah and Hendra Henipavirus on the CELLO2GO tool, we selected several pathogenic transmembrane proteins.Out of 108 transmembrane proteins, 3 transmembrane proteins were selected for vaccine design.

Selection and Sequence Retrieval of Transmembrane Proteins
Based on different factors like higher antigenic capacity, non-allergenicity, non-toxicity and non-resemblance with a human protein, three transmembrane fusion glycoproteins  were selected for vaccine design.UniProt protein database was used for the sequence retrieval of fusion glycoproteins under specific allocated UniProt IDs given in Table 1.

Prediction of B cells Epitopes
After protein selection and their sequence retrieval, we screened the predicted B-epitopes according to their higher antigenic capacity, non-allergenicity, non-homology, and non-toxicity.Ten B-cells epitopes, shown in Table 2, extracted from three selected fusion glycoproteins were included in the vaccine construct.All the shortlisted epitopes were non-allergenic and antigenic with a score between 0.7 and 1.21, having a threshold of 0.4.

Prediction of MHC-1 Epitopes
For the prediction of MHC-I-restricted epitopes, all alleles of the HLA dataset were selected, and the peptide length was taken as 9 and 10.The epitopes were sorted based on predicted IC50 values.All other parameters were kept as default.The shortlisted epitopes were screened based on higher antigenicity value, allergenicity and an IC50 value <100 giving a total of 10 epitopes, provided in Table 3.

Prediction of MHC-II epitopes
The length of the predicted MCH-2 cell epitopes was set at 15 and all other parameters were kept as default.The predicted epitopes were saved as an XHTML output table and peptides are sorted according to their adjusted rank.Further screening relied upon antigenicity values and non-allergenicity.The finalized epitopes, shown in Table 4, having a higher antigenicity value, non-allergenicity and IC50 values <100 were selected for the vaccine construct.The restricting alleles for both the MHC-I-restricted and MHC-II-restricted epitopes were recorded for population coverage analysis.

Vaccine Construct
The screened epitopes with a cumulative World population coverage of 97.2%, shown in Fig. 6A, were converted to the mRNA sequence, and fused together using universal linkers (EAAAK, GPGPG, KK, AAY).The 5'  and 3' components along with the RpfE adjuvant and the Kozak sequence were added to the construct, shown in Fig. 1A, for stability, improved immunogenicity, and ease of entry into the host cells.

Physiochemical Properties Prediction
The physicochemical properties computed for the vaccine candidate by translating the construct into its peptide sequence indicated that it was antigenic (0.8746), non-allergenic, stable (32.91), thermostable (93.62), hydrophilic (-0.002), non-toxic, and had 683 amino acids.Table 5 discusses the analysis and indication of the physicochemical properties of peptides.

Prediction of Secondary Structure
The secondary structure predicted by PSIPRED, shown in Fig. 2B, depicts the alpha-helices, beta-turns, and coils of the vaccine candidate.Most of the amino acid residues were predicted as coils indicating the vaccine candidate's transmembrane localization.

Prediction and Refinement of Tertiary Structure
trRosetta predicted the tertiary model of the vaccine candidate with a TM-score of 0.131.Since, this was below the ideal TM score, the model was refined and GalaxyREFINE predicted 5 models, out of which the first one was selected (depicted in Fig. 2C) based on the most improved RAMA scores (98.2) indicating the torsion angle distribution of the vaccine candidate.Other metrics like MolProbity (0.975) that evaluates the model quality at both protein and nucleic-acid levels and least RMSD (0.356) calculating the difference in protein backbone conformation in the initial and final structures were also considered for model selection.

Validation of Tertiary Structure
ERRAT calculated the overall quality factor score of 93.5421 and the Ramachandran plot (Fig. 2D) indicated that most of the residues lie in the most favored region.541 residues constituted 96.8% of the total protein, residues in the additional allowed region were 2.3%, in the generally allowed region were 0.4% and in the disallowed region were 0.5%.

Immune Simulation Response
C-IMMSIM predicted a potent and stable immune response against the injected antigen.The graph in Fig. 2E illustrates that the server predicted the antigen to be eliminated from the host system within the first 5 days after escalating to 700 000 antigen count per mL on the first day.The immune complexes (IgG+IgM) were generated when the antigen count was almost 0 and escalated up to 7000 for the first 15 days of injection.The complex count decreased afterward but became steady at 3600 after a month indicating stability and longevity of the immune response.The graph in Fig. 2F illustrates the cytokine production and danger-signal produced by the body in response to the antigen.A small danger signal was produced lasting for up to 25 days but was not significant.The interleukin-2 count escalated up to 250 000 mg/mL on the 5th day and steadily decreased afterward.It was clear from the graph that the increase in IL-2 count was a direct response to the signal validating the potent response.IFN-gamma was also readily produced by the immune system.

Molecular Docking Analysis
HDOCK webserver predicted 10 models of the docking poses of mRNA vaccine construct with TLR-3 (PDB ID: 3ULV) and 10 models with TLR-4 (PDB ID:4R7N).With TLR-3 docking, the first model was selected having a docking score of -331.09kcal/mol and a confidence score of 0.9740, shown in Fig. 3A.With TLR-4 docking, the first model (Fig. 3B) was selected for having a dock-ing score of -333.31kcal/mol and a confidence score of 0.975.

Molecular Dynamics Simulations
With an eigenvalue of 2.95646e-06 and 4.0524e-06, illustrated in Figs.4d and 5d respectively, the docked complex (Figs.4a, 5a) was predicted to be stiff with little dynamic residues.The B-factor (shown in Figs.4b  and 5b) also showed minimal flexibility, except for the last few residues.The deformability plot shown in Figs.4c and 5c was found consistent with the eigenvalue and the B-factor plot indicating that only a few residues towards the end of the docked complex were flexible/deformable.The variance and co-variance map (Figs. 4e,5e and Figs. 4f and 5f,respectively) show stable amino acid-pair interactions validating strong molecular interactions.Lastly, the elastic network map showed that the atoms in the docked complex are closely linked to each other, demonstrated in Figs.4g and 5g, reflecting on the covariance findings.

Expression Analysis
The CAI score of 0.95518 and GC content of 67.057% reflected the maximum codon optimization of the construct sequence, illustrated in Fig. 6B.pBluescribe plasmid was utilized for the in-silico cloning experiment.The plasmid was cut by the AHD-I enzyme producing sticky ends, enabling unidirectional cloning of the construct into the vector.Figure 6C shows the cloning process, whereas Fig. 6D represents the cloned vector.

DISCUSSION
Nipah virus first emerged in the late nineties, with fruit bat as its mediatory host.First, it was observed in the pig but was soon observed in the human population (Aditi & Shariff, 2019).Due to a high fatality rate in developing countries, this virus is thought to evoke a potential zoonotic pandemic like coronavirus (Kulkarni et al., 2013).Ribavirin, a drug initially used against hepatitis and liver-virus infection, has been utilized to minimize viral load in Nipah viral disease patients (Deb et al., 2019).However, the mortality rates are still high (Naveed et al., 2021).Owing to the absence of preventative strate-   et al., 2021).It has been observed that vaccines have some epigenetic impact on immune system genes, which train the immune system (Kumar et al., 2015).Likewise, the focus of this study was to develop a long-lasting memory in the form of activated B cells and T cells.The HTL induces IL-10, IL-4, and IFN-γ, antigen-presenting cells express epitopes of HTL, and the lymphocyte can secrete chemokines, all play crucial roles against the virus (Al Tbeishat et al., 2022).We observed the production of all these immune system components in response to the antigen.Except for the memory cells, all the immune cells died and were eliminated with the destroyed antigen.B cells with membrane-bound immunoglobulin receptors identified the antigen epitopes and internalized them to become presenting cells for memory-making units of the immune system.Furthermore, we noticed a hike in plasma-secreted antibodies (up to 40 000 counts) that are utilized to neutralize the antigen (vaccine candidate injection) and produce memory cells.
Our results were consistent with previous mRNA-based vaccine findings (Naveed et al., 2022;Maruggi et al., 2017), predicting an efficacious vaccine construct with the ability to elicit long-lasting protective immunity.
The tools utilized in this study have revolutionized biotechnological research.The reverse-vaccinology ap-  proach has saved time, preliminary clinical trials, and labor costs (Kumar et al., 2015).We followed this approach because of its proclaimed merits in prediction ( (Naveed et al., 2022a;Naveed et al., 2022b) as well as clinical studies (Hall et al., 2021;Pormohammad et al., 2021).The in-silico prediction of antigenicity, allergenicity, and toxicity significantly helped us in characterizing the potential epitopes.These tools have the potential to save lives at pre-clinical trials (Xinhui et al., 2021) as researchers may have already predicted the immunogenic, allergenic, or toxic outcome of a proposed vaccine before it is administered to the public (Naveed et al., 2023a;Naveed et al., 2023b).Lastly, the wide population coverage of the proposed candidate validates that it can be efficacious worldwide irrespective of race, ethnicity, or area.

Declarations
Ethics approval and consent to participate: Not applicable.
Competing interests: The authors declare no conflict of interest.
Immune response against Nipah virus in humans Prediction and Refinement of the Tertiary Structure

Figure 1 .
Figure 1.Process overview of the mRNA-based vaccine construct

FISFIIVEKK
Figure 2. (A) The final vaccine construct; (B) The predicted secondary structure of the construct; (C) The predicted tertiary structure refined by GalaxyREFINE; (D) Ramachandran plot validating the refined tertiary structure; (E) Antigen and Immune complex counts plotted over a month period; (F) Danger signal produced by the body in response to the injection and the cytokines production plotted over a month.

Figure 4 .
Figure 4.The results of the iMods simulation study of the proposed mRNA vaccine and TLR-3 docked complex.(a) MNA mobility, (b) deformability, (c) B factor, (d) Eigen values, (e) variance (green color indicated cumulative variances and purple color indicates the individual variance), (f) co-variance map (uncorrelated as white, correlated as red, and anti-correlated as blue motions) and (g) elastic network (darker grey regions indicate stiffer regions).

Figure 5 .
Figure 5.The results of the iMods simulation study of the proposed mRNA vaccine and TLR-4 docked complex.(a) MNA mobility, (b) deformability, (c) B factor, (d) Eigen values, (e) variance (green color indicated cumulative variances and purple color indicates the individual variance), (f) co-variance map (uncorrelated as white, correlated as red, and anti-correlated as blue motions) and (g) elastic network (darker grey regions indicate stiffer regions)

Figure 6 .
Figure 6.(A) Population coverage of the T-cell-specific epitopes; (B) Codon optimization chart of the proposed mRNA vaccine construct; (C) The history map of cloning, the black vector shows pBluescribe plasmid, 2040 bps long, the red line shows vaccine construct and the final construct after insertion in which red color shows gene of interest (4791 bps long); (D) The cloned Bluescribe vector with the gene of interest shown in red color