1. Introduction
HIV-1 and HIV-2 are two retroviruses that cause AIDS in humans [
1]. These viruses have various similarities in terms of replication, transmission, and clinical signs [
2]. HIV-1 is the primary cause of HIV infections globally, resulting in 680,000 fatalities in 2020 and a total of 37.7 million individuals living with HIV globally [
3]. It is worth noting that over eighty percent of individuals infected with HIV-1 gain the virus via mucosal exposure [
4]. As the pandemic continues to progress, the generation of a vaccine to combat this virus has become a top priority for researchers. However, the substantial genetic diversity of HIV gives rise to various genetic subtypes, constituting challenges for vaccine development [
5]. The surface glycoproteins gp120 and gp41, present in the envelope (Env) of the virus, display excessive genetic variability. These glycoproteins are crucial in promoting the attachment of the virus into the T cells of the host [
6]. Gp120 is extremely significant in the process of virus infection as it brings about the identification of specific receptors, including DC-SIGN, heparan-sulfate proteoglycans, and CD4 on the T cells of the host [
7]. One strategy is to avoid viral attacks into CD4+ cells and thereby prevent infection by blocking the formation of the gp120-CD4 complex [
8]. Furthermore, the production of neutralizing antibodies has garnered significant attention as they have the potential to bind to the junction of the external and internal domains of the gp120-CD4 binding site to rival CD4 receptors [
9]. Different experiments have found that neutralizing antibodies can effectively inhibit the activity of gp120, making this glycoprotein a promising candidate for the development of attentive and potent HIV vaccines [
4]. A subgroup of individuals living with HIV, indicated as elite controllers, acquire the incredible potential to remain asymptomatic and sustain high CD4+ cell counts for prolonged periods without the need for antiretroviral therapy (ART) [
10]. These elite controllers have excesses of CD4+ and CD8+ cells, according to the analysis, that secrete IFN-γ, a cytokine that induces Th1 immune response, recommending its role in controlling HIV infection [
11]. Moreover, the stimulation of CD4+ and CD8+ cells has shown effectiveness in combating HIV infection [
12]. However, conventional strategies for designing and developing HIV vaccines have thus far been unsuccessful. As a unique and promising strategy, the therapeutic vaccination of individuals already infected with HIV-1 has appeared to prevent disease progression to AIDS [
13].
A therapeutic vaccine against HIV-1 is meant to activate more effective and broader immune responses, especially targeting conserved viral epitopes, compared to the immune responses developed during natural infection. Therefore, the development of victorious HIV-1 therapeutic vaccine candidates has faced challenges, principally due to incompetent delivery systems or the suboptimal design of immunogens [
13]. One promising strategy for the development of HIV-1 therapeutic vaccines involves the design and assemblage of artificial multiepitope immunogens. These immunogens are designed by choosing from prevalent viral antigens and a diverse array of immunostimulatory, protective, and T-cell epitopes. The objective is to provoke effective immune feedback against HIV-1 infection [
14]. By employing this strategy, scientists hope to overcome earlier limitations and boost the effectiveness of therapeutic vaccines against HIV-1. Accordingly, in silico approaches that assist the discovery of potentially immunogenic peptides, known as epitopes, from the linear protein sequence are employed. Moreover, docking and molecular dynamics (MD) simulations can distinguish between different epitopes by assessing disparities in affinity and the stability of the peptide–protein complex on the major histocompatibility complex (MHC), including both MHC-I and MHC-II [
15]. To analyze the dendrimer-G4-PAMAM-peptide complexes, Rodrguez-Fonseca and colleagues used 3D models of HIV-1 gp120. Female BALB/c mice received these complexes intravenously, either as individual peptides or as complexes. The research revealed that the immune response to the peptides was triggered at both the systemic and mucosal levels and that dendrimer–peptide complexes produced better IgG and IgA responses in serum and nasal washes [
16]. Due to the significant role played by gp120 in host attachment and pathogenicity, we employed it as a primary antigen in the development of an HIV-1 vaccine. The focus of our study was on eliciting robust T-cell responses, specifically targeting the CTL and HTL. We present an effective in silico design strategy for designing a recombinant vaccine. To achieve this, we examined the conserved epitopes of gp120 and carefully selected suitable CTL and HTL epitopes capable of eliciting B lymphocyte responses against the gp120 protein. Subsequently, we designed the vaccine construct incorporating appropriate linkers and the Gb-1 adjuvant to enhance the induction of cellular immunity.
3. Discussion
The field of bioinformatics has emerged as a pivotal force in the realm of vaccine development, owing to its ability to predict immunogenic peptides that facilitate the development of vaccines that are both effective and safe. By utilizing peptide predictors, the costly and unwanted side effects associated with vaccines derived from attenuated pathogens, be they living or inert, can be minimized [
24]. While remarkable progress has been made in combating AIDS through antiretroviral treatments, the urgent need for an effective HIV-1 vaccine remains paramount in our efforts to curb this perilous global epidemic [
25]. Regrettably, most previously devised vaccines have fallen short of expectations, largely due to their limited potency against the promptly mutating nature of the virus, as they have predominantly targeted a narrow range of HIV genotypes [
26]. According to studies using the in silico developed multiple-epitope EP HIV-1090 vaccine, these vaccines’ inability to elicit strong cellular and HTL feedback is another reason why they are ineffective at preventing HIV-1 [
12], as well as for their failure to provoke broadly neutralizing antibodies, as demonstrated by experiments conducted on BALB/c mice using three multiple-epitope vaccines [
27]. Additionally, the lack of appropriate cytokine stimulation and the failure to stimulate the coveted innate immune feedback further contribute to the suboptimal performances of these vaccines. A truly effective vaccine must have the ability to stimulate specific immune responses against HIV-1 by enhancing both CTL and HTL activities, as CTL-mediated feedback expresses a key role in the management of viral infections. Furthermore, the significance of HTL-mediated immunity cannot be ignored since it is vital for supporting antibody immune responses and encouraging a functional CD8+ cytotoxic T lymphocyte (CTL) response, which, in turn, provides protection against the virus and a decrease in viral load [
28]. It is quite possible to improve the design of immunogens for HIV-1 vaccines by developing innovative immunoinformatics methods for the study of HIV-1 and the discovery of multi-functional T-cell epitopes, mainly when paired with in vivo studies [
29]. Strong cellular and HTL feedback can be elicited using multiple-epitope-based vaccines. Notably, the methods used in these studies have paved the way for the suggestion of a multiple-epitope vaccine that can successfully provoke broadly neutralizing antibodies, stimulate the production of the appropriate cytokine like IFN-γ, and trigger the desired innate response via docking with TLRs, complemented by the addition of an adjuvant Gb-1 at the vaccine’s C-terminal. Due to the encouraging results of several studies, in silico techniques have become more important in the construction of multiple-epitope vaccines. However, a multiple-epitope vaccine against the Onchocerciasis disease caused by microfilaria was effectively developed in an in silico investigation [
30]. Similar to this, a multiple-epitope vaccination against MERS was created using an immunoinformatic technique [
31]. A multiple-epitope vaccine against brucellosis showed encouraging T-cell responses [
32]. Furthermore, pre-clinical studies using animal models and a multiple-epitope Epstein–Barr virus vaccination showed encouraging outcomes [
33]. Similar to our current work, a multiple-epitope vaccine that is specially designed to target HIV-1 has also been created [
34].
The current immunoinformatic strategy was used to design a vaccine against HIV-1, and the focus was directed toward the envelope glycoprotein, specifically gp120, which involves attachment to the host receptor CD4 and enhances the virulence. Glycoproteins, prominently displayed on the surfaces of viruses, have long been prime candidates for vaccine development, as these surface antigens serve as the initial point of contact with the host’s immune cells. The current trial embraced a non-traditional and efficient procedure rooted in computational biology, harnessing the wealth of genomic data available, to design a multiple-epitope-based vaccine targeting diverse strains of HIV-1. Given the proven efficacy of immunoinformatic techniques in vaccine design, the principal goal of this work was to design a vaccine that might lessen the worldwide burden of various cancers brought on by the virus. The 3D structure and predicted epitopes were built using the consensus sequence. Several possibilities were produced when lineal epitopes for MHC-I and -II binding areas were predicted, and ElliProt was used for non-linear prediction. Furthermore, the 3D visualization of gp120 highlighted epitopes that may be recognized by antibodies. Considering that lower IC50 values imply better affinity, the use of a QSAR model, which forecasts IC50 values, made it easier to categorize peptides according to their affinity for the major histocompatibility complex (MHC).
T cells, which include both HTLs and CTLs, are essential for triggering an efficient immune response and protecting the host against viral infections. However, vaccines targeting CTL responses alone have shown lesser effectiveness compared to those targeting both CTLs and HTLs. Additionally, because of their specific qualities, these multiple-epitope vaccines have significant benefits over conventional and single-epitope vaccinations: (i) Different MHC class T-cell TCRs are capable of recognizing a wide variety of self and MHC class II epitopes that come from different T cell subsets. (ii) Interacting humoral and cellular immune responses can be triggered concurrently by overlapping CTL, HTL, and B-cell epitopes, promoting an all-encompassing and well-coordinated defense. (iii) Including an adjuvant in the formulation of the vaccine enables persistent immune feedback with increased immunogenicity. (iv) The difficulties resulting from in vitro antigen expression and pathogen culture can be avoided by eliminating the accompanying difficulties. Using the IEDB prediction service, T-cell epitopes that can recognize a variety of MHC class I and class II molecules were carefully chosen to further boost the immune feedback. In Tond epitopes with a greater affinity for binding and low percentile rankings (IC50), this selection method looked for them. On the other hand, exposed antigenic epitopes immediately recognize B-cell receptors (BCRs), which then stimulate the production of antibodies that are specific to those epitopes. This study used the IEDB prediction service to find linear or continuous B-cell epitopes. Overall, 106 T-cell and B-cell epitopes were ultimately chosen after the first screening phase based on the favorable results. The most promising choices among these epitopes, which had highly antigenic qualities while being free of allergens, toxins, and homologies with the human proteome, were subsequently chosen after going through multiple rigorous rounds of screening. Additionally, the HTL epitopes’ capacities to elicit cytokine responses, including IFN-γ, IL-4, and IL-10 responses, were evaluated. Gb-1, an adjuvant, was used to boost the vaccine’s strong immunogenicity. Using the linkers GTG, GSG, GGGGS, and GGTGG, the epitopes were joined at the proper locations. Adjuvants and linkers were used to strengthen the vaccine’s structure and improve its immunogenicity, antigenicity, and durability.
Long peptides have demonstrated superior efficacy in eliciting immune responses compared to short peptides [
35]. However, there remains uncertainty regarding whether artificially anticipated multiple-epitope constructs, comprising T- and B-cell epitopes, can be adequately displayed and processed to activate targeted immunity. Consequently, an essential aspect in enhancing multiple-epitope-based vaccines is the ability to predict TAP transport and proteasomal cleavage. According to the findings from the NetCTL 1.2 server, it appears that all chosen epitopes, including CD8+ and CD4+ T-cells, are accessible for immune feedback during the antigen processing and presentation carried out by competent APCs. Evaluating the immune feedback triggered by these epitopes is crucial for their judicious preference in vaccine design. Since specific cytokines can be induced by particular residues and motifs within an epitope, employing in silico cytokine prediction tools provides a comprehensive overview of the capacity of T-cell epitopes to stimulate heterogenous cytokines in a straightforward, rapid, and cost-effective way compared to in vitro and in vivo immunological evaluation [
36]. IFN-γ is a cytokine that serves as the hallmark of adaptive and innate immunity, displaying antiviral, immune regulatory, and anti-tumor properties. IFN-γ secretion plays a pivotal role in the Th1 response and is crucial for diminishing the viral load of HIV-1 [
37]. In this particular investigation, we evaluated the capacity of IFN-γ and the production of cytokines for each selected HTL epitope. To accomplish this, we employed the IFNepitope server to forecast peptides that could induce IFN-γ through binding to MHC class II molecules. The majority of our HTL epitopes demonstrated positive induction of IFN-γ and cytokine production, as indicated by their predicted SVM scores. The production of IFN-γ is closely related to the immunogenicity of HIV-specific T-cells and the elicitation of Th1 responses [
38]. Some studies described that IL-10 possesses anti-HIV activity by inhibiting the release of inflammatory cytokines [
39]. Furthermore, additional research has shown that T-cells that secrete IL-10 contribute to the reduction in HIV replication in pregnant women [
40].
Vaccine immunogens must be produced to combat the antigenic diversity of HIV-1 and the various HLA tissue types. To increase the population coverage rate in our investigation, we used peptides with different epitopes and numerous HLA binding specificities. MHC class I epitope population coverage was 90.23%, whereas MHC class II population coverage was 72.95%. Moreover, particularly in areas with a high prevalence of HIV-1, the multiepitope structures showed a significant cumulative population coverage. The size of the final designed epitopes showed an epitope conservation score of 75% among the M group of HIV-1 subtypes. The probability of viral immune evasion was decreased by high conservation across HIV-1 subtypes, which also offered wider protection. Utilizing the Vaxigen, AllerTOP v.2.0, and Toxinpred servers, respectively, the antigenicity, allergenicity, and toxicity of the predicted epitopes and the final construct of the HIV-1 gp120 vaccine sequence were evaluated. The outcomes demonstrated the antigenicity, non-allergenicity, and non-toxicity of the vaccine protein. The ProtParam program from the ExPASy website was used to examine the physicochemical characteristics of each predicted epitope and the proposed construct. With a high pI value of 10.13, the HIV-1 gp120 vaccine demonstrated acceptable stability and is considered to exist within the normal range. The vaccine showed a half-life beyond the 10 h in a prokaryotic E. coli culture system and 30 h in mammalian reticulocytes in vitro, indicating the possibility of the stable and scalable production of the vaccine in this system. A GRAVY value of -0.741 further demonstrated the vaccine’s hydrophilic nature and increased water solubility. High solubility was predicted via SolPro for the HIV-1 gp120 vaccine.
HIV-1 gp120 constructs were 3D-modeled using the RoseTTAFold web tool and then refined using the GalaxyRefine tool to make sure the protein closely matched its native or natural structure. Our results showed that after refining the vaccine construct’s, the final accuracy and the anticipated 3D structures’ quality both improved. The Ramachandran plot and the Z-score observed during the 3D and refined structure assessment of the HIV-1 gp120 vaccine validated the good outcomes. This analysis indicated that the vaccine possessed a structurally adequate form and should be effective as a vaccine. The vaccine’s Z-score was -6.02, which is in the acceptable range for experimentally validated X-ray crystal structures of proteins. Notably, the highest concentration of amino acids was observed in the analysis’ preferred regions [
41]. Additionally, protein–protein docking analysis between TLRs and our predicted constructs was carried out. By identifying pathogens and subsequently eliciting adaptive immune feedback, TLRs have a significant impact on activating the innate immune system [
42]. TLR2 and TLR4 detect virus structural proteins and release inflammatory cytokines as a result. Additionally, TLR3 is in charge of the activation of dendritic cells caused by HIV-1. [
43].
In our study, an in silico assay was conducted to examine the correlation between the HIV-1 gp120 construct and various TLRs, as mentioned in the methodology. The findings showed low energy scores and significant binding affinity between the vaccine design and TLRs. Contact investigation using contact maps displayed different patterns of interchain residue-to-residue association in the TLR-vaccine complexes. Specifically, the TLR2-vaccine complex exhibited a greater number of interchain contacts in various domains of the proteins. Three particular areas on the contact map in the TLR3 and TLR8 vaccine complexes showed interchain interactions between various domains. Fewer connections were seen in the TLR4–vaccine complex, indicating a less robust link between the TLR and vaccine chains. Interchain interactions in the TLR5–vaccine complex were restricted to closely related residues, but the TLR8–vaccine complex showed different vaccine chain residues making contact with the TLR chain. These findings suggest that the designed vaccine constructs have the potential to activate TLRs and downstream pathways, leading to the production of pro-inflammatory cytokines to counter the HIV-1 infection. A potent HIV-1 therapeutic vaccine should stimulate cellular and humoral immune feedback. A TRIF-dependent signaling cascade that is started by TLR-3 activation can lead to the transcription of inflammatory genes. [
44]. This stimulation in DCs leads to their maturation into potent immunostimulatory cells capable of productively cross-priming T-cells [
43]. In the context of HIV-1 infection, the stimulation of DCs is reliant on TLR-3 activation [
45]. On the other hand, TLR-4 activation can result in the induction of IL-6, which has the potential to re-emerge the virus from its dormant stage [
46]. Additionally, TLR-10 has been associated with enhanced HIV-1 infection [
47].
MD simulations were conducted to ensure the binding efficiency and equilibrium of the vaccine–receptor complex. These simulations allowed us to observe the correlation of the vaccine in association with the TLR receptors over time. The results indicated that the vaccine design was able to effectively occupy the TLR receptors with minimal energy, suggesting a strong binding affinity. Additionally, normal-mode evaluation performed using iMODS showed that the vaccine complex exhibited elevated eigenvalues, indicating a lower extent of deformability. This suggests that the vaccine construct is relatively rigid and less prone to conformational changes. The deformability graph further supported this observation, indicating that the vaccine construct is likely to maintain its structural integrity and stability. Furthermore, atomistic simulations provided valuable insights into the stability and conformational dissimilarities of the physiological systems. By examining several descriptors derived from the trajectory data of molecular dynamics simulations, we were able to study the conformational variations in and stability of the vaccine–receptor complex in a more precise manner [
48]. These simulations provide a deeper understanding of the dynamic manner and structural characteristics of the vaccine construct in complex with the TLR receptors.
To enhance the mRNA of the vaccine, the JCAT was employed, with the E. coli strain K-12 selected as the cell culture system. This tool was used to analyze the translation capacity of the HIV-1 gp120 vaccine. The results yielded a CAI value of 0.60 and a GC content of 61.90%. The current values were deemed acceptable, as a CAI value above 0.60 and a GC content ranging from 30 to 70% are considered favorable scores. During codon optimization in the pET plasmid, NcoI and XhoI restriction enzymes were utilized to cleave the N and C termini, respectively. The presence of 6xHis tags in the cloned plasmid allows for the post-translational purification of the vaccine. Furthermore, the RNAfold (version 2.1.0) software was employed to analyze the secondary structure of the HIV-1 gp120 mRNA. The software produced a minimum free energy value of -378.55 kcal/mol, indicating a more stable vaccine within the body. This stability is significant during vaccine construction. Moreover, in silico assessments of the host immune responses to our designed multiple-epitope vaccine demonstrated significant findings. The C-ImmSim tool predicted the activation of B cells and T cells, leading to prolonged memory. The presence of IgG1 and IgG2 showed Th1 and Th2 responses to HIV-1 antigens, suggesting significant protection against HIV-1 infection. However, results from the ICM server indicated a significant increase in the level of Th1 cells correlates with an elevation in the level of CTL. The current findings show that the vaccine construct, encompassing the immunogenic B- and T-cell epitopes of gp120, holds promise for the development of multiple-epitope vaccines against HIV-1. However, further in vitro and in vivo investigations are compulsory to fully assess the potential of our designed multiple-epitope-based vaccines for combating HIV-1.
4. Methodology
4.1. Protein Sequence and Multiple-Sequence Exploration
GenBank was initially searched for gp120 protein sequences connected to HIV. To determine the mutant and conserved portions of the gp120 protein, these sequences were then submitted to a multiple-sequence alignment using the STRAP program. The muscle server was used to create a consensus sequence.
4.2. B-Cell Epitope Prediction
B-cell epitopes are key players in vaccine design as they are recognized by the immune system. The current research employed two tools, namely IEDB Bepipred linear epitope prediction tool and BepiPred-2.0, to identify linear B cell epitopes within the gp120 protein sequence. Default thresholds were utilized during the analysis [
49,
50]. BepiPred-2.0 is a tool that anticipates B cell epitopes by employing a random forest algorithm with epitope trains obtained from antibody–antigen protein structures. This procedure combines a sizable number of linear epitopes acquired from the IEDB network with sequence-based epitope prediction exploiting 3D structures. In addition, the B-cell epitopes discovered by IEDB and BepiPred-2.0 were validated using iBCE-EL. We used a web-based analysis tool for linear B-cell epitopes called iBCE-EL. It utilizes a mix of dipeptide and physicochemical characteristics, very randomized tree and gradient boosting techniques, and a combination of amino acid configuration and physicochemical properties as input features. iBCE-EL forecasts a peptide’s class and probability values when given the peptide [
51]. It is important to note that almost 90% of B-cell epitopes are thought to be discontinuous, which means they are composed of sequences that are close while being far from one another in terms of long-distance pathogenic protein sequences. In the final revised vaccination 3D structural model, we used the ElliPro tool to analyze the conformational B-cell epitopes. The geometrical characteristics of protein structure were analyzed via ElliPro. ElliPro has the greatest AUC value of 0.732 among conformational B-cell epitope prediction methods, making it a very accurate technique for detecting antibody epitopes in protein antigens.
4.3. Cytotoxic T-Cell Epitope
To identify MHC-I-binding alleles, both commonly occurring and less frequent ones, the epitopes were analyzed using the NN-align tool in the IEDB analysis tool. The identification of MHC-I-binding alleles was based on the following parameters such as a peptide length of 8–12 amino acids and an IC50 value of less than 200. Peptides with IC50 values below 50 nM were regarded as occupying stronger affinity, while those below 500 nM showed moderate affinity, and those below 5000 nM had a poor affinity in this context. A lower IC50 value, therefore, denotes a greater affinity. The IEDB server is a comprehensive resource that was also utilized to predict the processing score, proteasomal cleavage, TAP value, and MHC-I-binding value of the selected epitopes and their corresponding alleles via the NN-align algorithm.
4.4. Class I Immunogenicity Assessment
MHC-peptides with potential for immunogenicity within the infected host cell were evaluated by using the IEDB MHC I immunogenicity prediction server. The server’s default settings were used for the evaluation of the shortlisted epitopes. The epitopes with a high immunogenicity value were then chosen for additional experiments.
4.5. Helper T Cell Epitope
Helper T-lymphocytes (HTLs), which identify foreign antigens and stimulate B-cells and cytotoxic T-cells, are essential for the immune system’s response to infectious diseases. The Immune Epitope Database was employed to analyze MHC class II T-cell epitopes. Based on either the percentile rank or the MHC binding affinity, the server predicted the binders. We used the IEDB-recommended combinatorial method to determine HTL epitopes. This strategy incorporated NN-align, SMM-align, CombLib, Sturniolo, and NetMHCIIpan techniques [
52]. The final selection of epitopes was based on the epitopes’ scores (lower scores indicating better binding), capacity to release IFN-γ, ability to produce emergent characteristics, and an IC50 value of less than 500 nm.
4.6. Selection of Top Epitopes
The selection of thresholds for the epitope selection process was based on previous studies. The aim was to strike a balance between achieving a high sensitivity to obtain a large number of high-quality epitopes and maintaining specificity. Individual filters were applied to each epitope based on the results of various tests conducted using a dataset of experimentally confirmed epitopes. This dataset primarily consisted of HIV-1 data, as it provided a comprehensive collection of immunogenic and non-immunogenic epitopes derived from a meta-analysis of multiple studies focusing on MHC class I epitopes.
4.7. Proteasomal Cleavage/TAP Transport
NetCTL 1.2 was employed for proteasomal cleavage and transporter-associated antigen processing (TAP). The website uses an algorithm that provides predictions for the efficiency of TAP transport, MHC class I binding, and proteasomal C-terminal cleavage. The default parameters gave TAP transport ability, with a weight of 0.05, and C-terminal cleavage, with a weight of 0.15.
4.8. IFN-γ Cytokine Inducer Prediction
The website server is used to predict the epitopes that cause IFN-γ cytokine production. The intracellular Th1 immune responses that IFN-γ stimulates are known to produce antiviral action in both the innate and adaptive immune systems. This service was used to anticipate MHC class II-binding peptides that may cause CD4+ T cells to produce IFN-γ. To prioritize IFN-γ over other cytokines, the server settings were changed to choose the Support Vector Machine (SVM)-based approach.
4.9. Determination of Population Coverage
The coverage rate of particular epitopes in the population was measured using the population coverage tool on the IEDB website. The global human population was evaluated via the given tool. Predicting the right epitopes for various HLA bindings can be aided by estimating population coverage. Varied ethnic groups may have varied frequencies of epitopes with stronger HLA binding affinities, which aids us in overcoming the drawbacks of MHC restriction in T-cell responses. In this study, the MHC-I and MHC-II HLA binding alleles were assessed for the gp120 protein. Moreover, the IEDB online server epitope conservation analysis tool was employed to anticipate the conserved cross-reactive epitopes by determining the identity of a specific peptide sequence among various HIV-1 subtypes in group M.
4.10. Antigenicity, Allergenicity, and Solubility Analysis
The AntigenPro server and the VaxiJen 2.0 server were employed to confirm antigenicity for the final vaccine and its components. To convert protein sequences into constant vectors of primary amino acid characteristics, VaxiJen uses the auto cross-covariance (ACC) transformation [
53]. AntigenPro, on the other hand, is a pathogen-independent, sequence-based predictor of protein antigenicity [
54]. The AllergenFP 1.0 and the AllerTOP 2.0 tools were employed to analyze the allergenicity of the final vaccine and its components. To differentiate allergens from non-allergens, AllergenFP employs a binary classifier. Five E-descriptors characterize the dataset, and the strings are converted into constant vectors using the ACC transformation [
55]. The ACC transformation and E-descriptors are also used by AllerTOP. The SolPro server and the Protein-Sol server were used to assess solubility. SolPro uses SVM to predict a protein sequence’s solubility with a tenfold cross-validation accuracy of more than 74%. Protein-Sol, on the other hand, is based on data from a cell-free production system for Escherichia coli protein solubility [
56].
4.11. Toxicity and Physicochemical Properties Assessment
The ToxinPred tool was employed to anticipate the toxicity of the final vaccine and its components. ToxinPred ranks toxicity and non-toxicity using an SVM model [
57]. The ExPASy ProtParam tool was employed to forecast the physicochemical properties of the final vaccine and its components. Hydropathicity, charge, half-life, instability index, pI (theoretical isoelectric point value), and molecular weight are among these qualities [
58].
4.12. Hydropathy Analysis of Epitopes
In the case of HIV-1 vaccines, all epitopes must possess a hydrophilic nature, meaning they should be present on the surface. This is because hydrophilic epitopes can provoke immune feedback within the host cell. To analyze the hydrophobicity of the epitopes, the GRAVY (grand average of hydropathy) score was measured through the ProtParam server. Calculating the total hydropathy values of all the amino acids in the protein and dividing that number by the total number of amino acid residues yields the GRAVY score. A positive GRAVY score specified a hydrophobic protein, while a negative value specified the presence of a hydrophilic region.
4.13. MHC Restricted Alleles through Cluster Analysis
We used the IEDB server to narrow down the selection of epitopes restricted to MHC class I and class II. To validate the predictions, we also employed the MHCcluster v2.0 server, which provided additional verification. This server presents the relationship between peptides and HLA functionality through a static heat map visualization [
59].
4.14. Multiepitope Subunit Vaccine Design
The epitopes that displayed the most promising characteristics, including high antigenicity, non-allergenic and non-toxic characteristics, a lack of homology to the human proteome, and conservation across selected strains, were carefully examined. Their potential to stimulate cytokines was assessed, and only the HTL epitopes competent in eliciting cytokine were chosen for vaccine development. The most promising epitopes that fulfilled the required standard and passed the above analysis were considered for constructing a multiple-epitope vaccine against HIV-1. To design the vaccine, the best-selected B-cell, CTL, and HTL epitopes were connected using an adjuvant Gb-1 (GP-2 binding peptide) sequence and different linkers. The Gb-1 peptide was identified in our laboratory through the utilization of the phage display screening method. It was discovered that the Gb-1 peptide possesses immune stimulatory properties and specifically triggers a Th2 immune response. The adjuvant plays a crucial role in enhancing the antigenicity, immunogenicity, stability, and longevity of the vaccine. The adjuvant was potentially linked to the epitopes using GTG linker and the inclusion of the Gb-1 sequence improves the ability of the immune feedback induced by the vaccine. Subsequently, the B-cell, CTL, and HTL epitopes were linked in a consecutive sequence using GTG, GSG, GGTGG, and GGGGS linkers.
4.15. Assessment of Physicochemical Properties of the Vaccine Construct
A crucial aspect of vaccine development is ensuring higher antigenicity, which guarantees the identification of the vaccine by the host immune system, leading to the stimulation of immune cells and succeeding immune response [
60]. To make sure that the vaccination does not cause an allergic reaction in the host, allergenicity testing must be carried out. Additionally, to guarantee the safety and potency of the vaccine, a thorough assessment of several physicochemical variables is required. VaxiJen v2.0, with an accuracy criterion of 0.4, was employed to observe the antigenicity of the created vaccine. Using AlgPred and AllerTop v2.0, the allergenicity of designed vaccine was also evaluated. The AlgPred service uses a variety of techniques to assess antigenicity and forecast probable allergenicity by comparing the resemblance of common epitopes in protein regions.
The allergenicity of designed vaccine was predicted using the MEME/MAST motif prediction in AllerTop v2.0. ProtParam was once more used to ascertain the physicochemical characteristics of the vaccines, including variables like the pI value, half-life, and the GRAVY value. Additionally, the SOLpro tool of the SCRATCH protein predictor was used to forecast the solubility of the vaccine construct while maintaining the default settings throughout the prediction, and the results were confirmed using the Protein-Sol tool. The solubility factor was essential in making sure that the vaccine was sufficiently soluble after being given to the host. A vaccine’s efficacy might be harmed even if it is very effective if it clumps into insoluble substances. Protein-Sol uses a quick sequence-based procedure to determine the findings, whereas the SolPro service uses the SVM method to predict solubility. As a result, it is possible to estimate the solubility of protein sequences using any of these services with confidence [
56].
4.16. Secondary and Tertiary Structure Analysis of the Vaccine
As a protein’s functioning depends on a certain structural conformation, it is essential to assess the secondary and tertiary structures of a designed vaccine. Using the online PRISPRED program, the secondary structure of the vaccine was ascertained after antigenicity and allergenicity testing while maintaining all the default settings. PRISPRED is a server designed to assess the secondary structure of biomolecules, as well as their transmembrane topology, helices, folds, and domain recognition [
61]. Two-dimensional (2D) structural analysis was carried out by utilizing the SOPMA and Phyre2 servers for correlation and extra confirmation to further validate the results. To construct the 3D design of the vaccine, we used the RoseTTAFold (BOINC version 7.6.22) software. RoseTTAFold is a neural network-based method that incorporates a “three-track” approach. It considers patterns in protein sequences, interactions between amino acids within the protein, and potential 3D structures. This comprehensive approach allows the complex to inclusively link the chemical components of a protein and its folded structure. RoseTTAFold has achieved high accuracies comparable to those of DeepMind. In the RoseTTAFold architecture, information flows back and forth between the one-, two-, and three-dimensional levels, enabling a holistic understanding of a protein’s structure [
62].
4.17. Refinement and Validation of 3D Vaccine Structure
The computation of vaccine design using computer-based methods is important, particularly when there is limited experimental evidence available. While designing 3D models is not always sufficient for ensuring accuracy in biomedical applications, it is important to have experimental data to verify the accuracy. Through the process of 3D structure refinement, it became feasible to enhance the fidelity of primitively designed structures and rectify local flaws while preserving the vaccine’s fundamental structure as closely as possible. The GalaxyRefine tool was used to improve the 3D vaccination architecture. This server employs a CASP10-tested approach for refinement and dynamics simulation, which resulted in upgraded structures [
63]. Moreover, achieving precise and uniform refinement of 3D vaccine models remains a difficult task, especially when working with high-resolution data. Therefore, to validate the accuracy of the vaccine design, Ramachandran plots were designed using the PROCHECK server [
64]. These plots analyzed the allowed and disallowed dihedral angles (psi and phi) of the amino acid configurations, taking into account the van der Waals radius of the side chain. Additionally, the ProSA-web tool was utilized, which employs various statistical methods to produce a Z-score that serves as a measure of protein structure validation. The Z-score indicates the standard of the protein structure being analyzed. By comparing the Z-score of the query protein with the range of Z-scores found in the PDB database for experimentally determined protein chains, we could determine the quality and consistency of the query protein’s structure [
65].
4.18. Evaluation of Vaccine Disulfide Engineering
A key component of vaccine design is using multiple molecular interactions to increase the stability of protein vaccines. Disulfide by Design 2 v12.2 was employed to analyze the disulfide design of vaccine protein. This server’s purpose is to identify probable locations in a protein structure where disulfide linkages are more likely to occur. To predict protein structure, the tool uses computational methods [
66]. Its technique uses a geometric framework built from natural disulfide links to precisely evaluate the x3 torsion angle based on the 5th Cbeta–Cbeta distance.
4.19. Vaccine Construct Docking with TLRs
Docking serves as a structural configuration for both drug design and basic investigations into protein interplay. It involves predicting the structure of a complex by combining the individual protein structures through protein–protein docking. In the current experiment, protein–protein docking was examined by docking it with different TLRs, including TLR-2, 3, 4, 5, 8, and 9 (PDB ID: 2Z7X, 1ZIW 3FXI, 3J0A, 3W3G, and 3WPB). The 3D structures of both TLRs and the HIV-1 gp120 construct were then refined for energy using PyMOL v2.3.4 software. To enhance the accuracy of predictions, various online docking tools were employed for protein–protein docking assessment. The PDB structure of the vaccine, TLRs, and HIV-1 gp120 experienced the removal of all water molecules. During the docking analysis, the 3D structure was observed using the Chimera V 1.13.1 server. Next, the HIV-1 gp120 constructs, along with TLR2 and TLR8, were submitted to the HADDOCK server to evaluate their interaction areas. The complexes with the highest rankings were selected based on the lowest intermolecular binding and the lowest mean RMSD from the HADDOCK analysis, representing the whole-molecular HIV-1 gp120–TLR interaction.
The clusters of docked complexes were arranged using ClusPro 2.0 in the second phase of docking, occupying the center and lowest energy ratings. Here, E = 0.40Erep + 20.40Eatt + 600Eelec + 1.00EDARS, and this equation was used with the ClusPro tool to calculate the energy score [
67]. A higher binding affinity was indicated by a lower energy score. Molecular mechanics/generalized Born surface area (MM-GBSA) research was carried out via docking, using all default settings. Following a successful docking procedure, visualization and MD simulation investigations were used to further evaluate the TLRs docked with the vaccine construct. The Discovery Studio Visualizer was used to visualize the docked complexes. PDBsum was used to conduct a thorough investigation of protein–protein interplay inside the complexes [
68]. H-band formation was evaluated using the LigPlot+ program.
4.20. Molecular Dynamics Simulation of the Vaccine-Receptor Complex
iMOD was employed to run the molecular dynamics simulation. Such modes may be explored using iMODS, which also creates workable transition routes between two homologous structures [
69]. By using normal mode analysis (NMA), the server computes internal attributes to evaluate the stability of proteins. The main-chain deformability plot, B-factor values, eigenvalue, covariance matrix, and elastic network model are some of the characteristics used to represent protein stability. Understanding the protein’s stability and structural dynamics is possible through such analysis.
4.21. Codon Optimization and In Silico Cloning of the Vaccine Construct
Reverse translation of the desired protein was carried out to find a suitable DNA sequence that might transcribe our peptide vaccine. The intended organism was then given the altered DNA sequence after codon optimization, which allowed it to successfully express the target protein utilizing the codons from the altered DNA sequence. The Java Codon Adaptation Tool (JCat) server was employed to carry out the codon adaptation procedure for the proposed vaccination protein [
70]. The prokaryotic
E. coli strain K-12 was preferred as the target organism. During the codon adaptation process, certain considerations were made to bypass rho-independent transcription terminators, prokaryotic ribosome-binding sites, and cleavage sites for the restriction enzymes
NcoI and
XhoI. Additionally, the N- and C-termini of the optimized DNA sequence were positioned at the combined
NcoI and
XhoI restriction sites. Using the SnapGene restriction cloning tool, the modified DNA sequence was implanted between the
NcoI and
XhoI restriction sites of the pET-28a (+) vector [
71]. The DNA sequence of the pET-28a (+) vector included a single 6×His tag. The recombinant protein’s solubilization and efficient affinitative purification were made easier by the presence of this tag [
70].
4.22. Immune Simulations of Vaccine Construct
To model and forecast the immune feedback to the recombinant HIV-1 gp120 construct, the C-ImmSim server was used. It used a position-specific scoring matrix (PSSM) to evaluate immunological interactions through an agent-based model and machine learning approaches. The simulation steps domain was adjusted to 1050 to reflect the 4-week gap between the first and second doses of the vaccination. Each time step, which equaled eight hours in actual time, was set at 1 and 84 and applied to two injections. Other settings remained at their preset values during the study.