In silico Prediction of Epitope Based Vaccine Candidates against Powassan Virus Infection

Egyptian Academic Journal of Biological Sciences is the official English language journal of the Egyptian Society for Biological Sciences, Department of Entomology, Faculty of Sciences Ain Shams University. C. Physiology & Molecular Biology journal is one of the series issued twice by the Egyptian Academic Journal of Biological Sciences, and is devoted to publication of original papers that elucidate important biological, chemical, or physical mechanisms of broad physiological significance. www.eajbs.eg.net Provided for non-commercial research and education use.


POWV, epitope prediction, toxicity prediction, population coverage analysis
Powassan virus (POWV) is responsible for encephalitis and severe neurological sequelae globally.Peptide target based designing offers a promising therapeutic invention for the eradication of viral infection.Immunoinformatics serves as a powerful tool to screen and select antigenic peptide sequences as potential epitopes for binding affinity with HLA alleles.In the present study, a computational pipeline was developed for the identification of B-cell and T-cell epitopes for suitable vaccine candidates.Further, immunogenicity and physico-chemical prediction studies enable the discrimination between antigens and non-antigens.Considering the population setting globally, population coverage analysis was also performed for the identification of possible binding alleles (MHC class-I and MHC class-II) of T-cell epitopes.This computational prediction analysis will enhance our understanding of B-cell/T-cell immune response and assist in selecting the antigenic peptide(s) for the formulation of antigen based diagnostic kit or peptide based subunit vaccine design against POWV.
As per the previously published literature, phylogenomic-related studies reveal more than 80% similarity in phylogenetic relationship among deer tick virus (DTV) and POWV (Ebel et al., 2010).
Molecular epidemiology deciphers the organization of 11kb single-stranded, positive-sense RNA genome encoding seven nonstructural proteins and three structural proteins: capsid (C) protein, premembrane (preM) protein, and envelope (E) glycoprotein (Hermance and Thangamani, 2017).Despite being completely sequenced, the vague knowledge regarding causing infection and survival in the host raised the quest for the development of novel vaccines.Therefore, in this study, an in silico approach was employed for epitope identification (Somvanshi and Seth, 2009;Singh et al., 2009) and characterization, which provides a solid platform for future researchers to develop potential vaccines against Powassan virus.The identification of B-cell and Tcell epitopes (HLA class-I CD8+ T-cells and HLA class-II CD4+ T-cells) to unwire the underlying viral pathogenesis has been performed nearly a decade ago (Somvanshi and Seth, 2009;Singh et al., 2009;Somvanshi et al., 2008a;Somvanshi et al., 2008b).In order to overcome the limitation of presence of extreme polymorphisms among maximal population setting, population coverage analysis was performed which will help in the discovery of novel vaccine against Powassan virus.

Data-set Collection:
The complete protein sequences of Powassan virus were retrieved from the NCBI genome database (https://www.ncbi.nlm.nih.gov/genome/) in FASTA format.The retrieved sequences were further subjected to immunogenicity and epitope prediction.

Physico-Chemical Characterization:
The assessment of several physicochemical properties of the proteins/peptides using sequences was performed by using an online server at EXPASY Bioinformatics Suite, ProtParam (Gasteiger et al., 2005).The physico-chemical properties included molecular weight, amino acid composition, extinction coefficient, theoretical pI, and grand average of hydropathicity, aliphatic index, and instability index (Gasteiger et al., 2005).

Immunogenicity Prediction:
The evaluation of immunogenicity of the protein sequences was performed by using VaxiJen V2.0 server (http://www.ddg-pharmfac.net/vaxijen/ Vaxi-Jen/VaxiJen.html) which enables the discrimination between antigens and non-antigens by predicting protective antigens, tumor antigens, and subunit vaccines with the precision level of 70 to 89% employing the underlying Auto Cross Covariance (ACC) algorithm.

Epitope Prediction:
ABCpred server (www.imtech.res.in/abcpred) was employed for the prediction of B-cell epitopes from the primary protein sequences.This server enables B-cell epitope prediction using artificial neural network strategy using n-1 combinations for n number of possible outcomes.Additionally, a combinatorial machine learning platform NetCTL 1.2 server enables the identification of T-cell epitopes (MHC class-I and MHC class-II) for Powassan virus protein sequences.The properties of the possible identified epitopes were calculated by using Peptide Property Calculator available at https://www.genscript.com.

Toxicity Assessment:
The toxicity check of the identified epitopes was performed by using Toxin Pred server (Gupta et al., 2013).This server takes into account the frequency and probability of amino acids at a particular position by generating a quantitative matrix using Support Vector Machine (SVM).

Population Coverage Analysis:
Taken into consideration the population coverage globally, Immune Epitope Database and ANALYSIS RESOURCE (IEDB) Population Coverage tool available at http://tools.immuneepitope.org/tools/population/iedb_input was used for the identification of possible binding alleles (MHC class-I and MHC class-II) of Tcell epitopes (Bui et al., 2006).

RESULTS AND DISCUSSIONS
The prevalence of infections caused by Powassan virus is evolving at high pace.This sudden health burden becomes a major concern for the countries having tropical cover (Black et al., 2010).Despite several advancements in healthcare sector, a huge impact will be justified by the development of a vaccine against the infection of Powassan virus (Huang et al., 2011).
In order to appraise the role of humoral immunity against the infection of Powassan virus, a computational pipeline was developed by considering immunogenicity prediction of viral protein and the identification of B-cell and T-cell epitopes (Saha and Raghava, 2006).The antigenic property of viral proteins has been ensured using VaxiJen V2.0 at a constant threshold of 0.4 (Table 1).To combat the complexity related to the pathogenesis of Powassan virus, antigenic determinant sites were identified for B-cells and T-cells using ABCpred, an epitope identification tool based on machine learning algorithmic strategy.The occurrence of both continuous and discontinuous epitopes (Shen et al., 2015;Blythe and Flower, 2005) in viral sequences ensured greater accuracy of the results obtained.The immunogenic potential of viral envelope protein in Powassan virus with a score of 0.6884 indicated its capability to use as a probable antigen.Table 2 represents the probable B-cell epitopes for POWV and their respective immunogenic potentials.Additionally, a positive correlation between physiochemical properties (antigenicity and surface amino acid residues) of viral sequences and B-cell antigenic determining sites was observed.Combining amino acid anchoring pair composition (APC) and support vector machine (SVM) methods to obtain an area under curve of 0.847 was set as default parameter with respect to the occurrence of both continuous and discontinuous B-cell epitopes (Kori et al., 2015).
Physico-chemical characterization enables computation of instability index, extinction coefficient, GRAVY, aliphatic index, and theoretical pI of protein sequences of Powassan virus.As per the results, envelope protein and 2K protein showed least instability index; hence they were found more stable under in vitro conditions.The total amino acid composition (mainly rate of cystine residue formation) reflected the molar concentration of protein which in turn assists in determining the strength by which protein absorbs the light at a given wavelength per molar concentration.The thermostability of the protein sequences was revealed by the estimation of volume occupied by the aliphatic side chains in a protein (aliphatic index).The solubility of protein sequences was determined by computing the hydropath values ranging from -2 to +2 for most of the proteins, with the positively rated proteins being more hydrophobic (Table 1).Anchored protein C showed highest hydropathicity index and antigenicity score of 0.5251.
Further, the identification of T-cell epitopes (MHC class-I and MHC class-II) assists in the development of potential vaccines for the treatment of infection caused by Powassan virus (Table 3).The identification of T-cell receptors on antigen presenting cells (APC) responding towards the class-I and class-II molecules was performed using NetCTL 1.2 server at default parameters of 0.5, 0.89, and 0.94, for sensitivity, specificity, and accuracy (Peters and Sette, 2005;Tenzers et al., 2005).For MHC class-I alleles, IEDB database using SMM align method was employed by considering potential sites having IC 50 value < 300 nM.In addition, the antigenicity prediction of the screened epitopes was done by VaxiJen (Doytchinova and Flower, 2007).Table 3 represents the highest number of MHC class-I binding alleles are related to anchored core protein C (4) and nonstructural NS2A (7) sequences.Further, the antigenic assessment identified "YRGCKAAG" epitope with highest VaxiJen score of 2.09.In order to increase the confidence level of the prediction of MHC class-II alleles, IC 50 value of 300 nM was used as threshold, as there is a direct relationship between SMM-align prediction scores and logtransformed IC 50 binding affinity (Kori et al., 2015).Multiple binding of epitopes with various alleles enables the identification of T-cell epitopes responding towards MHC class-II.T-cell epitopes "CGRGGWSYYAASRP" and "ITSNYNIMV" of anchored core protein C showed highest VaxiJen score of 0.98 and 0.96, respectively.The toxicity assessment enabled the identification of toxic potential of the identified epitopes.The applicability of the potential vaccine at global level was assessed by performing population coverage analysis ranging between 16.84-96.38%(Table 4 and Figure 1).

Fig. 1 :
Fig. 1: Represents the average hit and PC90 analysis for MHC class-I and MHC class-II, respectively.(blue = average hit for MHC class-I, orange = PC90 for MHC class-I, grey = average hit for MHC class-II, and yellow = PC90 for MHC class-II) Overall, the current results imply the use of the identified and characterized epitopes in vaccine discovery in the future.Further analysis of anchored protein C epitopes for vaccine discovery purpose is warranted because of positive correlation between the physicochemical properties (GRAVY) and B-cell /T-cell epitopes' VaxiJen scores.