If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Biological Models Laboratory, Department of Biochemistry and Molecular Biology, College of Medicine, University of the Philippines Manila, Manila, Philippines
Biological Models Laboratory, Department of Biochemistry and Molecular Biology, College of Medicine, University of the Philippines Manila, Manila, Philippines
Biological Models Laboratory, Department of Biochemistry and Molecular Biology, College of Medicine, University of the Philippines Manila, Manila, Philippines
Biological Models Laboratory, Department of Biochemistry and Molecular Biology, College of Medicine, University of the Philippines Manila, Manila, Philippines
African Swine Fever (ASF) is caused by a highly contagious and fatal hemorrhagic virus, which in 2019 alone in Asia, has killed 8 million pigs with a devastating estimated economic loss amounting to $130 billion. There were attempts to control ASFV transmission; however, developed vaccines failed to produce lasting immunity. Currently, no vaccine has been approved yet. This study designed a novel multi-epitope and multi-antigenic vaccine using open-access bioinformatics tools. B-cell, helper-T and cytotoxic T-cell epitopes were predicted using consensus sequences from ASFV genotypes of antigens p12, p17, p22, p54, p72, and CD2v, and combined with adjuvants and linkers to form the ASF vaccine. Analyses revealed that the ASF vaccine is stable, antigenic, non-allergenic, and not cross-reactive. Docking of SLA-1 to CTL-HTL regions of the developed vaccine revealed that it effectively binds to SLA-1, a vital process in priming an effective immune response. Immune simulations demonstrated that the designed ASF vaccine can elicit primary and secondary immune responses, and stimulate the production of effector immune cells and cytokines. Overall, these results revealed that the designed multi-epitope and multi-antigenic ASF vaccine is potentially effective and warrants further in vitro and in vivo studies to confirm its protective function against ASFV infection.
African Swine Fever Virus is an emerging DNA arbovirus (arthropod-borne) of global concern in the pig industry, causing a fatal hemorrhagic disease among domestic pigs [
]. ASFV mainly infects swines and warthogs, and can be harbored by ticks, highlighting its limited host range and its non-infectivity to humans due to its low mutation rates [
]. The transmission cycle of ASFV is illustrated in Fig. 1. Soft ticks (Ornithodoros) act as vectors and reservoirs of ASFV, where the virus replicates in their midgut [
]. Highly fatal (100% mortality rates in domestic swine) ASFV genotype II was found to produce non-specific manifestations (e.g., diarrhea, fever) in infected swine, while swine with non-fatal ASFV genotype I was often asymptomatic.
Fig. 1Transmission Cycle of ASFV. In the tick's life cycle, ASFV can be transmitted in different modes: sexual (male to female), transstadial (offspring to parent) and transovarial (parent to offspring). These infected ticks may transmit the virus to domestic pigs (Sus scrofa domesticus), warthogs (Phacochoerus africanus) and wild boar (Sus scrofa). Other modes of transmission include livestock-wildlife interaction, swill feeding, animal trade, fomites, intra-farm outbreaks, and sylvatic cycle (warthog to ticks and vice versa).
]. The outer capsid is mainly composed of p72 major capsid proteins forming hexameres and other minor capsid proteins such as the penton protein that contributes to its icosahedral shape [
]. Meanwhile, the inner envelope follows the icosahedral shape of the capsid and is composed of single-pass transmembrane proteins such as p12, p17 and p54 [
]. The genome size of ASFV ranges from 170 to 190 kilobase pairs which consists of 151 to 167 open reading frames encoding 68 virion-associated proteins [
]. Diagnosis and genotyping of ASFV strains mainly use p72 gene, whereas recent studies suggested the use of other discriminative genes such as p54 and pB602L genes [
The socio-economic impact of African Swine Fever (ASF) to swine industries worldwide made it a notifiable disease by the World Organisation for Animal Health [
Kumagai N. African swine fever in Asia and the Pacific. Presented at: African swine fever: An unprecedented global threat - A challenge to livelihoods, food security and biodiversity. Call for action; October 26-30, 2020; Online Webinar. https://www.fao.org/3/cb2145en/cb2145en.pdf.
]. Due to the persisting crisis brought about by ASF epidemics, it is imperative to explore vaccine strategies to protect against ASF and control ASFV transmission. Early attempts to develop vaccines failed to induce protective immune responses against ASFV infection. Recent studies reported conflicting results on ASFV-specific neutralizing antibodies whilst non-neutralizing ASFV antibodies may even lead to swine mortality [
]. Despite the availability of genomic sequences and numerous immunogenic antigens of various ASFV strains, the functions of other ASFV proteins or viral-coding genes remain to be elucidated. These characterized proteins, especially those exposed on the surface of the virions, may be potent candidates for vaccine development. Exploring these immunogenic antigens, this study designed a multi-epitope and multi-antigenic vaccine composed of predicted B-cell, T-helper and Cytotoxic T-cell epitopes from protective ASFV surface antigens. The designed ASF vaccine was evaluated for its immunogenicity, allergenicity, and cross-reactivity. Its physicochemical properties and its secondary and tertiary structures were estimated. Lastly, immune simulations were done to predict its ability to elicit immune responses.
Methodology
Retrieval of ASFV sequences
The methodological framework of the in silico design of the multi-epitope ASF Vaccine is illustrated in Fig. 3. Fifty-six (56) African Swine Fever Virus (ASFV) isolates from different countries were retrieved from the NCBI Virus Database. The protein sequences of p12, p17, p22, p54, p72, and CD2v from these isolates were aligned using Clustal Omega [
]. Conserved fragments were identified through protein variability analysis using the Protein Variability Server with a Shannon entropy threshold of 1.0 which were used for screening of B-cell, Helper T-cell and Cytotoxic T-cell epitopes [
The identified conserved sequences from six antigens of ASFV were used to screen for linear B lymphocyte (BL) epitopes. Several tools in Immune Epitope Database (IEDB) such as Emini Surface Accessibility (ESA), BepiPred Linear Epitope 2.0 (BLE), Kolaskar & Tongaonkar Antigenicity (KTA), and ABCPred Server were used to generate linear BL epitopes [
]. Epitopes that overlapped with glycosylation and cleavage sites were excluded. BL epitopes with 7–20 amino acid residues obtained from at least 2 servers were used for further evaluation.
Linear helper T lymphocyte epitope prediction
The consensus sequences from the six antigens of ASFV were assessed for predicted MHC II binding using IEDB analysis resource: MHC II Binding Prediction set with IEDB recommended 2.22. Since only human MHC II alleles are available in the prediction tool, common human MHC II alleles that are orthologs to their swine counterpart alleles were used to predict epitope peptides. Helper T-lymphocyte (HTL) epitopes with IC50 < 500 and length of 11 amino acids were considered for further evaluations.
Linear cytotoxic T lymphocyte epitope prediction
From the consensus sequences obtained for the six ASFV antigens, proteasomal cleavage/ TAP transport/ MHC class I combined predictor tool in IEDB was employed for cytotoxic T-lymphocytes (CTL) epitopes [
]. In addition to antigen processing (proteasomal cleavage), TAP (transporter associated with antigen processing protein) transport, MHC class I binding, and T cell receptor-MHC interaction are included in the prediction outcomes for potential CTL epitopes. Common SLA class I molecules (SLA -1, -2, and -3) are utilized in the prediction. The output of MHC Class I binding predictions include the possible epitopes with predicted binding affinities (IC50), proteasome, TAP, MHC and total scores. The predicted epitopes were filtered using these parameters: epitope length of 9 to 11 residues, IC50 <500 nM, TAP score >1.0, proteasome-processing score > 1.0, and those covering the MHC I SLA alleles.
Evaluation of predicted epitopes
Predicted BL, HTL, and CTL epitopes were evaluated for antigenicity, allergenicity, and cross-reactivity [
]. 16–17 Antigenic ASFV epitopes were identified using Vaxijen 2.0 Server with a threshold e” 0.5 and using virus as a target. Epitopes with potential allergenic sites were identified using Allergen FP v.1.0 [
]. The predicted epitopes were assessed for autoimmune reactions, exploring for possible hits against swine protein sequences (Sus scrofa, Sus scrofa domesticus, and Sus scrofa domestica) in UniProtKB and SwissProt databases using protein-protein BLAST (BLASTp). Default parameters in BLASTp excluding models, non-redundant refseq proteins, and uncultured/environmental sample sequences were used. Epitopes that passed all evaluationswere used in creating the ASF multi-epitope vaccine.
Multi-epitope ASF vaccine construction
The ASF vaccine was constructed using the evaluated epitopes. Overlapping epitopes of CTL and HTL from the same protein were shortlisted using BL epitopes as reference. CTL epitopes were also shortlisted using HTL epitopes as template, if the overlapping CTL and HTL epitopes did not overlap with any BL epitopes. Overlapping epitopes for each protein from the same type of lymphocyte were also combined into continuous peptides. The combined epitopes were arranged to form the protein vaccine. Valine was added in the N-terminus and two potential immune-adjuvant peptide sequences were used, specifically the A6 peptide (sequence: TNGDILNYY) and the F3 peptide (sequence: SVDSPTITY). In the constructed vaccine, A6 peptide was linked to F6 using an AAY linker since both peptides are CTL epitopes. The adjuvant peptides are linked to overlapped CTL epitopes using AAY linkers, while HTL and BL overlapping epitopes are linked using GPGPG linkers.
Evaluation of the constructed multi-epitope ASF vaccine
Antigenicity and allergenicity of the constructed vaccine was assessed using Vaxijen 2.0 and AllergenFP v1.0 servers, respectively. The vaccine was also checked for cross-reactivity against Sus scrofa (taxid: 9823), Sus scrofa domestica (taxid: 9825), and Sus scrofa domesticus (taxid: 9825) whilst excluding other models (XM/XP), non-redundant RefSeq proteins (WP) and uncultured/environmental sample sequences using BLASTp. The physicochemical properties were obtained using the Protparam tool [
]. GOR IV created the secondary structure percentage profile specifically marking regions predicted as random coils, alpha helix, and extended strands. The tertiary structure was modeled using the Galaxy TBM tool and was refined further using Galaxy Refine server [
The refined structure was docked to a cleaned crystal structure of swine major histocompatibility complex class 1 SLA-1 0401 from Research Collaboratory for Structural Bioinformatics Protein Data Bank (RCSB PDB ID: 3QQ3) using ClusPro server [
Crystal structure of swine major histocompatibility complex class I SLA-1*0401 and identification of 2009 pandemic swine-origin influenza A H1N1 virus cytotoxic T lymphocyte epitope peptides.
]. The ASF Vaccine-SLA-1 complex with the least binding energy was chosen and modeled in USCF Chimera. The binding energy and affinity for the complex was evaluated at 37 °C by the PRODIGY web server [
]. Default simulation parameters were used in the system for a prophylactic vaccine injected two times in three weeks intervals (time steps set at 1 and 63). The immune response profile of the vaccine was compared with that of the adjuvant alone.
In silico cloning optimization
For vaccine production, in silico cloning optimization was employed for the ASF vaccine. Using competent Escherichia coli K12 strains in silico, cloning optimization was performed using Java Codon Adaptation Tool (JCAT) [
]. JCAT allows the adaptation of a gene codon to sequenced prokaryotes approximating the codon adaptation index. Additional options which avoided prokaryotic ribosome binding sites, cleavage sites of restriction enzymes and rho-independent transcription terminators were applied. Moreover, the GC-content of the improved sequence of the vaccine was also computed.
Results
Conserved sequence analysis, epitope prediction and evaluation
The conserved amino acid sequences of the ASFV antigens were chosen based on their location in the virion and existing reports on its immunogenicity against ASF. The resulting conserved sequences of p12, p17, p22, p54, p72, and CD2v antigens were used for BL, HTL and CTL epitope prediction. The predicted epitopes were evaluated for antigenicity, allergenicity, and cross-reactivity and those that passed all the evaluations were used in constructing the multi-epitope vaccine. A total of 36 epitopes were obtained: 12 BL epitopes, 20 HTL epitopes and 4 CTL epitopes.
ASF multi-epitope vaccine construct and properties
The multi-epitope vaccine (Fig. 4) was constructed using the BL, CTL, and HTL evaluated epitopes connected by linkers. The ASF vaccine consists of 398 residues generated from 36 selected epitopes that were merged into 18 epitope fragments arranged from N-terminus to C-terminus: CTL, HTL, BL epitopes, respectively.
Fig. 4ASF Vaccine multi-epitope diagram. The designed vaccine was organized according to the epitopes involved interspersed with respective linkers. The N-terminal contains a valine followed by the adjuvants (pink) with a B cell/HTL linker (beige) in between, while an adjuvant linker (blue) links the adjuvant region to the CTL epitopes (purple) that is interspersed with CTL linkers (green). HTL epitopes (blue green) follows and then the BL epitopes (sky blue) and both are interspersed with B cell/HTL linkers (beige).
The ASF vaccine was predicted to be antigenic (0.5344) in Vaxijen 2.0 Server using “virus” as a target and non-allergenic with a Tanimoto Similarity index of 0.82. Additionally, the vaccine had no significant similarity with swine proteins. Protparam analysis revealed that the vaccine has a molecular weight of 42, 485. 82 g/mol and a theoretical isoelectric point (pI) of 10.10. The total number of negatively (Asp + Glu) and positively (Arg + Lys) charged residues were 15 and 45, respectively. The computed extinction coefficient was 54,570 M−1 cm−1 at 280 nm measurement in water and absorbance of 1.284 if all pairs of cysteine residues form cystines. On the other hand, if all cysteine residues are reduced the extinction coefficient is 54,320 M−1 cm−1 at 280 nm measurement in water. Its estimated half-life of the vaccine is 100 h in mammalian reticulocytes (in vitro), >20 h in yeast (in vivo), and >10 h in Escherichia coli (in vivo). The resulting instability index is 35.83 (<40 threshold) signifying that the vaccine is classified as a stable protein. The vaccine is hydrophilic as revealed by its grand average of hydropathicity is -0.356.
ASF vaccine secondary and tertiary structures
The secondary structure composition of the ASF vaccine is shown in Fig. 5. The most common structures are random coils (purple), followed by extended strands (red) and least being alpha helices (blue). The random coil regions are interspersed within different epitope divisions of the ASF vaccine which correlates with the identified disordered regions (1–7, 93–119, 133–174, 185–243, 255–267, 273–289, 296–308, 318–389). These regions are flexible, less likely to form stable folds, and act as binding regions for B-cell receptors. The best tertiary structure model of the ASF vaccine (Fig. 6) resulted in an elongated peptide with alpha helices interspersed between random coils. It is subdivided into 4 regions: F3-A6 Adjuvant, CTL, HTL and BL epitopes. The conformation with the least interaction energy by ASF Vaccine to SLA-1 docking was shown in Fig. 6. The estimated binding energy (∆G) and binding affinity (Kd) for the complex are -49.3712 kj/mol and 4.5E-09 mol/L, respectively. In addition, residues 42-118 are the interaction points of ASF Vaccine with SLA-1 which coincides with the predicted linear epitopes from the vaccine.
Fig. 5Graphical representations of secondary structures of the ASF vaccine. Predicted secondary structures in GOR IV include alpha helices (8.54%) in blue, extended strands (28.64%) in red and random coils (62.81%) in purple.
Immune response profile and codon optimization of ASF vaccine
Results from the immune simulations showed that the ASF Vaccine was able to elicit an immune response in contrast to the adjuvant alone (Fig. 7). The primary and secondary immune response was evidently higher in the ASF Vaccine than the adjuvant alone. This trend was also evident with that of the B-cell and cytotoxic T-cell population. Interestingly, populations of helper T-cells and most of the cytokines were relatively higher for the adjuvant alone than the vaccine. Meanwhile for the codon optimization, the resulting CAI-value obtained from the JCAT tool for ASFV multi-epitope vaccine codon usage optimization is 0.9361 which is inside the CAI range of 0.9 to 1. The GC-content of the improved vaccine codon sequence is 54.94% which is within the 30–70% range. Codon optimization of ASFV multi-epitope vaccine with an improved DNA codon sequence length of 1198 base pairs allows the vaccine to be highly expressed in E. coli K12.
Fig. 7C-ImmSim simulation profile of the ASF Vaccine and adjuvant alone. A) Immunoglobulin production and specific subclass (colored peaks) in response to injected antigen (black peaks); B) B-cell population per isotype; C) Population of HTL memory and not memory cells; D) CTL populations at the active, resting, and anergic state and: E) Counts induced cytokines and interleukins while the insert plot shows IL-2 level and Simpson Index (D)
Fig. 7C-ImmSim simulation profile of the ASF Vaccine and adjuvant alone. A) Immunoglobulin production and specific subclass (colored peaks) in response to injected antigen (black peaks); B) B-cell population per isotype; C) Population of HTL memory and not memory cells; D) CTL populations at the active, resting, and anergic state and: E) Counts induced cytokines and interleukins while the insert plot shows IL-2 level and Simpson Index (D)
Vaccination remains to be a powerful tool in improving animal health, reducing transmission, and mitigating the impacts of diseases on livestock production [
]. There are also studies that made use of immunogenic ASFV structural proteins, such as p30, p54, and p72, wherein immunization of the purified recombinant proteins, recombinant DNA, or a combination of the two have showed partial to no significant protection and only robust immune responses against the proteins was observed [
The African swine fever virus proteins p54 and p30 are involved in two distinct steps of virus attachment and both contribute to the antibody-mediated protective immune response.
]. This ineffectiveness may be attributed to factors such as high mutability of protein targets and lack of clearly defined targets that could elicit humoral immunity. There also exist research gaps on ASFV infection and immunity, strain variation, and ASFV protein functions, which hampers the development of an effective vaccine. Another aspect to consider is the fact that ASFV contains numerous immunogenic antigens, and some other proteins or viral-coding gene functions remain elusive, which accentuates the need to explore the immune response against these ASFV antigens [
]. Moreover, neutralizing antibodies plays an important role in providing a high degree of immune protection and conferring long lasting immunity. Recent studies reported conflicting results on ASFV-specific neutralizing antibodies whilst non-neutralizing ASFV antibodies may even cause swine mortality.
In the advent of biotechnology, production of subunit vaccines is sought-after since it explores small amounts of protein or glycoprotein virus fragments that are able to induce an effective immune response. Development of subunit vaccines first requires identification of immunogenic antigens; however, the conventional flow of antigen screening is a tedious process. In silico screening prior to conducting in vitro or in vivo assays is advisable to prevent problems on antigen expression [
]. Open-access immunoinformatics tools were utilized in this study to design, predict and evaluate ASF Vaccine from previously studied immunogenic antigens, prior to actual vaccine production.
In silico vaccine design/development
The vaccine was designed using predicted BL, HTL and CTL epitopes from the conserved sequences of six antigens from different ASFV isolates. Conserved sequences from each antigen were identified through the Protein Variability Server, an online based tool that makes use of variability metrics to calculate for the absolute site variability in a given multiple protein-sequence alignments [
]. Conserved fragments used for screening of B-cell, Helper T-cell and Cytotoxic T-cell epitopes were identified using a Shannon entropy threshold of 1.0, to assure that these sequences are highly conserved. The values of Shannon entropy (H) vary from 0–4.332 [
]. Thus, a Shannon entropy threshold of 1 was in the study to assure that the viral sequences that would be used for downstream analysis are highly conserved. In a rapidly evolving pathogen that makes use of sequence variability for immune evasion, this tool is able to identify distinct patterns of conserved antigenic residues which can be used as protein targets in an epitope vaccine design [
]. The approach of creating a multiple epitope vaccine is considered advantageous since it includes MHC-restricted epitopes that can be identified by T-cell receptors from a number of T-cell subsets [
Immunogenic antigens were chosen for epitope prediction, including p12, p17, p22, p54, p72 and CD2v. These antigens are situated in the capsid, inner envelope, and outer envelope for ease of recognition of immune cells. There have been attempts to identify protective antigens among the ASFV antigens, which include p72 and p54 that were able to induce neutralizing monoclonal and polyclonal antibody production [
]. Meanwhile, CD2v antigen, involved in hemadsorption of ASFV-infected cells, was shown to induce production of hemagglutination inhibitory antibodies and prevent ASFV infection in vitro [
]. Similarly, surface proteins p12, p17, and p22 were among the newly identified immunogenic antigens upon immunization by DNA primed and recombinant vaccinia virus [
Immunization of oigs by DNA prime and recombinant vaccinia virus boost to identify and rank african swine fever virus immunogenic and protective proteins.
]. Early studies conducted by Zsak et al. reported that antibodies from pigs who recovered from ASFV protection experiment conferred immunity against ASF [
]. In vivo and in vitro studies observed the potential protective effects of antibodies through different modes of actions such as complement-mediated cell lysis or antibody dependent cellular cytotoxicity [
]. With these, it is evident that antibodies could provide protection against ASF therefore, BL epitopes were included in the vaccine design. Platforms such as Emini Surface Accessibility (ESA), BepiPred Linear Epitope 2.0 (BLE), Kolaskar & Tongaonkar Antigenicity (KTA) tools found in Immune Epitope Database (IEDB), and ABCPred Server were used to generate linear B-Cell epitopes. Emini Surface Accessibility is a tool based on Emini's surface accessibility scale which predicts the accessibility profile of a hexapeptide sequence showing its probability of being found on the surface [
]. KTA is a semi-empirical approach that utilizes the physicochemical properties and frequency of occurrence of amino acid residues in predicting epitopes [
]. Additionally, ABCPred is a tool used to predict linear B-Cell epitopes in a given sequence through the use of artificial neural networks. Exploring the combination of these platforms enabled us to predict for B-cell epitopes that can be included in the designed ASF Vaccine [
Cytotoxic T lymphocytes are important effector cells of the adaptive immune system, which are known to interact with MHC class I molecules present on antigen presenting cells. In swine immunity, MHC-I interacting with CTLs are referred to as SLA Class I. These CTLs cause cytotoxicity to virus-infected cells through the action of perforins, granzymes and cytotoxic granules. As reported by Denyer et al. in 2006, there are two phenotypes of CTLs involved in ASF immune response, both of which cause cytolysis to infected cells [
Perforin expression can define CD8 positive lymphocyte subsets in pigs allowing phenotypic and functional analysis of natural killer, cytotoxic T, natural killer T and MHC un-restricted cytotoxic T-cells.
]. ASF outcomes are relative to the infecting viral strain, denoting the existence of more virulent and less virulent ASFV strains. In a study conducted by Oura et al., depletion of CD8+ T lymphocytes resulted in impaired protective immunity leading to increased swine fatalities [
]. This demonstrates the importance of CTL in conferring protective immunity against ASF. Proteasomal cleavage/TAP transport/MHC class I combined predictor tool in IEDB was employed for CTL epitopes. This tool applies the concept that antigens undergo proteasomal cleavage, transported and finally presented on MHC class I molecules [
]. In addition to antigen processing (proteasomal cleavage), TAP (transporter associated with antigen processing protein) transport, MHC class I binding, then T cell receptor-MHC interaction are included in the prediction outcomes for potential CTL epitopes. MHC molecules are polymorphic and since the aim is to produce a vaccine for swine, common SLA class I molecules (SLA-1, -2, and -3) are utilized in the prediction. All six antigens were used in the SLA-I epitope prediction; however, only p22, CD2v and p72 yielded epitopes that passed the set parameters. After further screening, three p72 epitopes and one p22 epitope remained. To compensate for the reduced number of epitopes, two in silico predicted CTL epitope capable of stimulating interferon-gamma production in vitro were added as vaccine adjuvants [
]. With such, epitope length of 9–11 amino acids length were considered as part of the cutoffs. IC50 is presented in a percentile rank, which signifies specific binders having lower percentile ranks; hence, IC50 values less than 500 nM are ideal binders, which is the threshold associated with immunogenicity [
]. Meanwhile, proteasome-processing scores evaluate the generation of peptides with corresponding C-terminus. TAP transport score, on the other hand, calculates the generation of peptide or N-terminally extended precursors. Both the proteasome-processing and TAP transport scores are predictors that were tested using TOC curves, which utilizes type positive and false positive predictions for every threshold and an area under the ROC curve is measured for the predictor's performance. Hence, a random predictor obtains a score of 0.5 and a perfect predictor would score >1 [
Helper T-lymphocytes orchestrate the activation of macrophages for phagocytosis, BL for antibody secretion, and CTL to eliminate virus-infected cells. This is achieved by interacting with MHC class II molecules on antigen presenting cells (APC). Unlike SLA class I, SLA class II displays strong homology to HLA class II and SLA II genes, DR and DQ are constitutively expressed on professional APC surfaces [
]. Physiological similarity between humans and pigs is the rationale the latter is considered as a biomedical model for toxicology, drug screening and heart and gut function [
]. For this reason, common HLA class II genes were used in the study. The epitopes and their binding affinity with MHC II alleles were determined through IEDB recommended 2.2 found in the IEDB server. This method has an improved performance in peptide binding prediction since it was based on an extended data set of quantitative MHC–peptide binding affinity [
]. The peptides binding to MHC II are not constrained in size as reported by the works of Yassai et al., and Nelson et al. and as such can vary from 11–30 amino acids long according to Rammensee et al. [
]. For this reason the authors chose to set the parameters to the lowest amino acid parameter to sample the virion peptides in the smallest possible length to detect short binding fragments that are capable of binding with MHC II. The method of sampling the smallest peptide is then compensated by the additional step of peptide conjugation of screened peptides based on IC50 values of less than 500 nm. Effectively, this means that the HTL epitopes that are conjugated are not merely a series of 11 epitopes combined together, but a compilation of Linear HTL epitopes present in each virion part considered.
Vaccine evaluation
Aligned with the aim of reducing the tedious process of screening antigens, Vaxijen, an alignment-independent tool applying antigen prediction based on auto-cross covariance transformation approach was utilized. It relies on the chemical properties of the primary structure of antigens [
]. Each ASFV epitopes obtained and the ASF vaccine peptide were screened for their antigenicity scores. The vaccine yielded a score of 0.5344 which means that is indeed antigenic, signifying the ability of the vaccine to be recognized by antibody binding sites.
Although considered rare, the possibility of an allergic reaction from vaccinations has been a major health concern [
]. Hence, extra care should be done especially with peptide-based vaccines because it could be potentially allergenic. To assess the allergenicity of epitope candidates and the designed vaccine, the study used AllergenFP, an alignment-independent prediction tool which utilizes amino acid descriptors and novel descriptor fingerprints approach based on Tanimoto coefficient similarity search to predict allergenicity [
]. The designed vaccine scored 0.82 Tanimoto similarity index, signifying that it is non-allergenic and may prevent unwanted hypersensitivity reactions.
Vaccines have a potential to induce an autoimmunity as a consequence of cross-reactivity [
]. Local alignment algorithms, like FASTA or BLASTp, were used and the results obtained provided the degree of homology between proteins which can be cross-reactive. Screening for the similarity of the predicted epitopes and the ASF Vaccine to swine proteins revealed no significant similarity, thereby ruling out the possibility of cross-reactivity. The GC-content of the improved vaccine codon sequence has been increased, hence with codon optimization, the ASFV multi-epitope vaccine may be highly expressed in E. coli K12 host. The physicochemical properties and in silico cloning optimization are useful later, in vaccine production.
To probe the possible utility of ASF vaccine in vitro and in vivo, a molecular docking simulation was done to determine if the predicted epitope regions will interact with at least one of its natural receptors such as SLA-1. The 3QQ3 crystal structure used was captured via complexation of SLA-1 0401 with peptides derived from swine-origin influenza A virus which gives a representation of the structural basis of peptide presentation by SLA-1. Cleaning of the PDB file was done on the 3QQ3 crystal structure to ensure that the target protein alone would be used for docking [
]. Default parameters were used in protein-protein docking mode in ClusPro. This tool is a widely used protein-protein docking tool that utilizes Piper, an FFT or Fast Fourier Transform rigid docking program. Essentially, given protein pdb structures it outputs 1000 lowest energy structural confirmation out of 109 possible outcomes that underwent rotations and translations throughout. The clustering of the 1000 ligand positions are determined via the number of neighbors of a ligand in a 9 Angstroms radius from its C-alpha. The exploration of the vaccine epitope-SLA-1 interaction is necessary in the induction of cytolysis of monocytes and macrophages which are preferably infected by ASFV. This demonstrated favorable complex formation which validates the predicted binding region of ASF vaccine and is comparable to epitope-SLA binding results in literature. In future studies, the possible impact of solvents, metabolites, steric effects and allosteric effects on epitope-SLA-1 binding can be explored further, since this study only demonstrated a preliminary study of the possibility of the vaccine epitope-SLA-1 binding [
Since swine immune simulations are not freely available, CIMM Server was used. This tool allows the simulation of cellular and humoral responses of the human immune system hence, this tool could provide an overview of the immune response elicited by a mammalian immune system [
]. Immune simulations showed that the constructed vaccine was able to elicit primary and secondary immune responses and produce memory B cells and antibodies such as IgM and IgG subclasses which are important against infection. It also yielded a slightly higher HTL memory cell population compared to the adjuvant as the antigen (vaccine) concentration increases. This implies that the adjuvant alone can initiate an increased immunity, and when combined with the all epitopes as a vaccine it can induce a lasting immune response by increasing HTL memory cell populations. Lastly, immune simulations revealed that the constructed vaccine and the adjuvant alone were able to stimulate production of cytotoxic T-cells and induce interferon-gamma production and other related cytokines, demonstrating the immunogenicity of ASF Vaccine.
This study also attempted to do cross protection screening; however, it was deemed not feasible as ASFV was the sole DNA virus infecting swines. Since ASF immune response is strain-specific, consensus sequences from different ASFV genotypes were used to hypothetically provide cross protection among ASFV genotypes. Utilizing open-access bioinformatics tools poses limitations as to epitope prediction, specifically in HTL epitope prediction where we utilized human orthologs of HLA Class II. Nevertheless, utilizing several in silico tools revealed promising results which warrants ASF vaccine for further in vitro and in vivo studies.
Conclusion
The designed multi-epitope and multi-antigenic ASF vaccine constructed using predicted B-cell, Helper T-cell, and Cytotoxic T-cell using 6 protective antigens (p12, p17, CD2v, p22, p54, and p72 was shown to be antigenic, immunogenic, non-allergenic and non-cross-reactive which are crucial parameters in designing a vaccine. The ASF vaccine was also recognized by SLA-1 preferentially at the CTL region, which is an essential step in the induction of a cytotoxic immune response against ASFV-infected cells. Immune simulations also showed that ASF vaccine was able to induce pertinent immune response against ASF infection. ASFV sequences were obtained globally and could potentially induce cross protection across ASFV genotypes. Utilizing several in silico tools revealed promising results which can be used as the first step in designing effective and safe ASF vaccines. Hence, the developed multi-epitope and multi-antigenic ASF vaccine warrants further screening in vitro and in vivo to confirm its protective function against ASF infection.
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
CRediT authorship contribution statement
Ara Karizza G. Buan: Visualization, Data curation, Formal analysis, Writing – original draft. Nico Alexander L. Reyes: Visualization, Data curation, Formal analysis, Writing – original draft. Ryan Nikkole B. Pineda: Visualization, Data curation, Formal analysis, Writing – original draft. Paul Mark B. Medina: Visualization, Writing – original draft, Supervision.
Data Availability
The data used for the research were obtained from publicly-available protein sequence databases. The bioinformatics tools for analyses were also accessed free of charge.
Declaration of Competing Interest
The authors declare that there has not been any conflict of interest associated with the work and no significant financial support that could influence the results of the study.
Acknowledgements
We also would like to show our appreciation to Dr. Leana Rich M. Herrera for sharing her published research work with us on which this research was adapted.
References
Gaudreault NN
Madden DW
Wilson WC
Trujillo JD
Richt JA.
African swine fever virus: an emerging DNA arbovirus.
Kumagai N. African swine fever in Asia and the Pacific. Presented at: African swine fever: An unprecedented global threat - A challenge to livelihoods, food security and biodiversity. Call for action; October 26-30, 2020; Online Webinar. https://www.fao.org/3/cb2145en/cb2145en.pdf.
Crystal structure of swine major histocompatibility complex class I SLA-1*0401 and identification of 2009 pandemic swine-origin influenza A H1N1 virus cytotoxic T lymphocyte epitope peptides.
The African swine fever virus proteins p54 and p30 are involved in two distinct steps of virus attachment and both contribute to the antibody-mediated protective immune response.
Immunization of oigs by DNA prime and recombinant vaccinia virus boost to identify and rank african swine fever virus immunogenic and protective proteins.
Perforin expression can define CD8 positive lymphocyte subsets in pigs allowing phenotypic and functional analysis of natural killer, cytotoxic T, natural killer T and MHC un-restricted cytotoxic T-cells.
00011-8&id=ga1.jpg){kind=link}
00011-8&id=gr1.jpg){kind=link}
00011-8&id=gr2.jpg){kind=link}
00011-8&id=gr3.jpg){kind=link}
00011-8&id=gr4.jpg){kind=link}
00011-8&id=gr5.jpg){kind=link}
00011-8&id=gr6.jpg){kind=link}
00011-8&id=gr7a.jpg){kind=link}
00011-8&id=gr7b.jpg){kind=link}