Exploring nonstructural NS5B protein to design a multi-epitope-based vaccine against HCV: immuno-informatics approach

Abstract

Hepatitis C Virus (HCV) is a significant global health threat that leads to chronic liver disease, cirrhosis, and hepatocellular carcinoma. While antiviral therapies have advanced over the years, they face challenges such as the high cost, drug resistance, and the absence of a prophylactic vaccine. This study focused on developing a new multi-epitope-based vaccine using in-silico methods targeting the conserved NS5B protein of HCV, known for its crucial role in viral replication. Identification and screening of B-cell and T-cell epitopes as highly antigenic, immunogenic, and non-allergenic were performed using immuno-informatics tools. The epitopes were conjugated with adjuvants and linkers (EAAAK, GPGPG, and AAY) to improve immunogenicity and stability. The physicochemical properties, antigenicity, and population coverage of the vaccine construct were evaluated. Molecular docking and dynamics simulations indicated stable and significant interactions with TLRs that ensured robust innate immunity activation. Computational immune simulations also showed potent activation of humoral and cellular immune responses, suggesting that the vaccine could protect across a range of genetically diverse populations. Designed to address the challenge of HCV’s variable and highly conserved nature, this vaccine aims to target HCV immune response while minimizing virus escape effectively. The study also highlights the success of computational vaccinology in expediting vaccine development in a time and cost-efficient manner. The proposed vaccine construct shows significant potential, and further studies are warranted to assess its immunogenic properties, safety, and efficacy. Comparative molecular dynamics simulations of the apo-receptor, vaccine construct, and their complex revealed that binding induces structural stabilization and dampens flexibility at key interface residues. These results validate the docking predictions and underscore the robustness of the designed vaccine–TLR5 interaction. This study generates further understanding toward the goal of developing a broad prophylactic strategy against HCV that is readily available and potentially amenable to future clinical use, offering hope in the fight against HCV.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40203-025-00448-9.

Introduction

Hepatitis C Virus (HCV) remains one of the most prevalent global health problems and is a significant cause of chronic liver diseases, cirrhosis, and hepatocellular carcinoma (HCC) worldwide. Although high-quality antiviral treatment is available, barriers remain to universal disease control, such as cost, accessibility, and drug resistance (Said and El-Sayed 2022). In addition, the lack of a prophylactic vaccine adds to the global burden of HCV, emphasizing the importance of developing a vaccine (Rzymski et al. 2024).

The acute phase of HCV infection is usually asymptomatic, with fatigue, nausea, abdominal pain, and jaundice becoming more pronounced as the disease progresses (Liu and Kao 2023). Untreated infections can progress to chronic infections associated with severe complications, including liver fibrosis, cirrhosis, and increased HCC risk (Solitano et al. 2021). While Direct-Acting Antivirals (DAAs) have transformed HCV treatment and achieved cure rates of over 90%, they do not provide long-term immunity or protection against reinfection (Marco et al. 2025). Furthermore, the use of expensive treatments continues to pose a challenge in resource-poor circumstances, highlighting the importance of an efficacious vaccine to mitigate the global impact of HCV (Elbahrawy et al. 2023).

The NS5B RNA-dependent RNA polymerase, a key viral protein in HCV replication, plays a crucial role in the production of the viral RNA genome, which is essential for replication (Elghoneimy et al. 2021). Its conserved nature among the HCV genotypes makes it an ideal target for epitope-based vaccine approaches, reducing the risk of immune-invasive mutants (Anwar et al. 2022). NS5B, one of the most conserved and immunogenic epitopes, is being investigated in plasmid DNA to address concerns about the broad epidemic. Using epitope prediction tools, it is possible to identify regions of NS5B that trigger potent T-cell and B-cell responses, providing evidence of the potential vaccine-targeting of NS5B.

Vaccine design, guided by immune-informatics, has progressed tremendously and led to new prophylactic approaches. In-silico methods allow the accurate prediction and selection of antigenic epitopes that are non-toxic, non-allergenic, and immunogenic. Combined immunodominant regions enhance humoral and cellular immune responses; multi-epitope-based vaccines are beneficial (Naveed et al. 2021). These vaccines are arguably safer than conventional formulations, which may cause adverse reactions and/or unwanted immune activation. Computational tools also enable population coverage analysis so analysts can check if a vaccine construct accommodates a broader range of the diversity covered. In-silico tools have become indispensable because of their speed in development, cost-effectiveness, and predictive validity in vaccine research (Saldanha et al. 2023).

Recent advances in computational biology have introduced new paradigms, such as reverse vaccinology and structural vaccinology, which utilize entire pathogen genomes and 3D protein modeling to pinpoint new vaccine targets (Naveed et al. 2021). Artificial intelligence and machine learning strengthen in-silico pipelines by predicting antigenicity, allergenicity, and epitope–HLA binding and discovering immunogenic patterns in large datasets (Ali et al. 2025). Additionally, new cloud-based high-throughput computational platforms allow the simulation of host–pathogen interactions at the systems biology scale and the integration of immune-informatics and systems biology (Naveed et al. 2025). These advances go far beyond classical epitope mapping as they provide novel, comprehensive insights into the dynamics of the immune response and facilitate the rapid transition of vaccine candidates like HCV NS5B multi-epitope constructs from in-silico design to experimental validation.

In this study, we report the design of a new multi-epitope-based vaccine that targets the HCV NS5B protein. This design process is guided by immune-informatics and molecular docking-based methods. The goal is to identify immunodominant T-cell and B-cell epitopes in NS5B, evaluate their potential antigenicity, and design a vaccine that can generate robust humoral and cellular immune responses. The safety profile of the vaccine is also assessed by predicting its allergenicity and toxicity. To enhance immunogenicity, adjuvants and linkers are incorporated into the vaccine design. Molecular docking and simulation studies are then performed to confirm and study the interaction of the vaccine construct with immune receptors, as well as its stability and efficacy. In conclusion, this work provides a comprehensive framework for further experimental validation and transfer to clinical studies, aiming to address the global problem of HCV infection.

Methodology

For the construction of vaccine candidate, several immunoinformatic tools were used. The overview of the methodology roadmap is provided in Fig. 1.

Fig. 1.

Roadmap for the construction of a vaccine candidate against HCV

Data collection

In the initial stage of in-silico vaccine development, the sequence of HCV non-structural protein NS5B, with accession number YP_001469630.1, was retrieved from NCBI (https://www.ncbi.nlm.nih.gov/genbank/) in the FASTA format against HCV genotype 2 (Geer et al. 2010).

Secondary structure analysis

To evaluate the physicochemical characteristics of a multi-epitope-based vaccine construct, parameters such as aliphatic index, grand average of hydropathicity, theoretical pI (isoelectric point), amino acid composition, molecular weight, half-life, instability index, and GRAVY were analyzed using the ProtParam tool, which is accessible at (https://web.expasy.org/protparam/) (Azimi et al. 2024). For disulfide and cysteine prediction of target protein, DIANNA 1.1 was employed, available at (https://bioinformatics.bc.edu/clotelab/DiANNA/) (Ferrè and Clote 2006).

Antigenicity analysis

For the Antigenicity analysis, which predicts whether the chosen protein is capable of making antigens that are crucial in vaccine designing. The tool Vexijen 2.0 (https://www.ddg-pharmfac.net/vaxijen/VaxiJen/VaxiJen.html) was taken into account for this analysis (Doytchinova and Flower 2007). Higher antigenic score and non-allergenic nature of the protein make it exceptional among other proteins.

Allergenicity prediction

Allergenicity and toxicity analysis of the targeted protein is an important step. For this, the allergenicity was confirmed by AllerTOP V 2.0 (https://www.ddg-pharmfac.net/AllerTOP/) while toxicity was checked through ToxiPred (https://webs.iiitd.edu.in/oscadd/toxipred/supple.php) (Dimitrov et al. 2014; Sharma et al. 2022).

Epitope prediction

The MHC Binding Tools within the IEDB framework predict peptide binding to Major Histocompatibility Complex (MHC) molecules, which is useful for T-cell response research and vaccine development. For the B cell, the Antibody Epitope prediction IEDB tool (http://tools.iedb.org/main/bcell/) was used, which helped in discovering regions that B cells recognize and direct antibody response investigations and vaccine development (Yan et al. 2024). Tools like ElliPro and DiscoTope 3.0 predicted conformational epitopes using protein structure information, aiding in our understanding of antibody-antigen interactions (Ponomarenko et al. 2008; Høie et al. 2024). The IEDB tool for T-cell epitope prediction mainly focuses on MHC I & MHC II binding prediction, considering their distinct roles in presenting antigens to CD8 + and CD4 + T cells, respectively. The set of epitopes filtered and sorted was done on an Excel sheet for profiling.

Construction of multi-epitope-based vaccine

The multi-epitope-based vaccine was constructed with a blend of B-cell, MHC I & MHC II epitopes along with linkers such as EAAAK, GPGPG, and AAY (Naveed et al. 2022a). For the adjuvant, the synthetic construct MOM gene was used at the N-terminal of the vaccine construct with a 6x His tag that was required for the binding of the histidine protein (Naveed et al. 2024a).

Evaluation of immunogenic, allergenic, & physicochemical analysis of multi-epitope-based vaccine construct

The sequence is assembled and presented for the Antigenicity (VaxiJen), Allergenicity (AllerTOP), and toxicity analysis to ToxinPred (https://webs.iiitd.edu.in/raghava/toxinpred/multi_submit.php) (Doytchinova and Flower 2007; Dimitrov et al. 2014; Sharma et al. 2022). The solubility check analysis of vaccine sequence in E. Coli was done by the SOLUPROT tool (https://loschmidt.chemi.muni.cz/soluprot/) and the solubility score was above 0.5 which is a must to be the best candidate for developing expression in E. Coli (Hon et al. 2021). The Protparam tool was used for the physicochemical analysis of our vaccine structure and provided the molecular composition, weight, half-life prediction, and Gravy for hydrophobicity or hydrophilicity.

Physicochemical properties analysis

To evaluate the physicochemical characteristics of multi-epitope-based vaccine construct, parameters such as aliphatic index, grand average of hydropathicity, theoretical pI (isoelectric point), amino acid composition, molecular weight, half-life, instability index, and GRAVY were analyzed using the ProtParam tool, which is accessible at (https://web.expasy.org/protparam/) (Azimi et al. 2024). For disulfide and cysteine prediction of target protein, DIANNA 1.1 was employed, available at (https://bioinformatics.bc.edu/clotelab/DiANNA/) (Ferrè and Clote 2006).

Disulfide bond engineering

The Disulfide by Design 2.13 (DbD2) Server (http://cptweb.cpt.wayne.edu/DbD2/) was used for disulfide engineering for improved structural integrity of the designed multi-epitope-based vaccine. The input file contained the three-dimensional structure of the homology modeled final vaccine construct in PDB format. The program scanned the intra-chain residue pairs systematically in a defined order to locate regions that could be mutated to cysteine and form disulfide bonds. The default geometric limitations of optimal Chi3 angle and default tolerance that was applied of + 97°/−87° (± 30°), and the Cα– Cβ– S angle > 114.6° (± 10°) were also used. Pairs of candidate residues that had the lowest energy scores, and hence higher feasibility of disulfide bond formation, were screened from the rest based on dihedral angles, bond geometry, and energy values. For disulfide bond formation, the lowest energy pairs were below ~ 3.0 kcal/mol. These pairs were identified, and further structural validation was applied because they were predicted to be outside the epitope or linker regions.

Secondary & tertiary structure of vaccine construct’s extrapolation & validation

The vaccine sequence was further utilized to predict its secondary structure from RaptorX (http://raptorx.uchicago.edu/), and the 3D structure analysis was predicted through (https://yanglab.qd.sdu.edu.cn/trRosetta/) (Källberg et al. 2012; Du et al. 2021). The 3D model was refined using the GalaxyRefine (https://galaxy.seoklab.org/cgi-bin/submit.cgi?%20type=REFINE) tool to enhance its score (Heo et al. 2013). The result provides the improved 3D structure of the vaccine with an increased score above 90%. For further verification, the Ramachandran Plot, which is present on the UCLA website (https://saves.mbi.ucla.edu/), was used (Naveed et al. 2024b). For predicting conformational B–cell epitopes, we used ELLIPRO and DiscoTope (http://tools.iedb.org/ellipro/) because their servers have epitope propensity scoring augmented by structural data, and perform better than purely sequence-based tools. ELLIPRO, for example, predicts epitopes because of structural protrusion and solvent accessibility in the three-dimensional space, and DiscoTope adds amino acid surface accessibility statistics from experimentally corroborated epitopes, along with the surface accessibility of the molecules. Thus, they are better than BepiPred or DiscoTope, which are based on sequence-only prediction, and even COBEpro.

Molecular docking of the vaccine construct with the receptor

To ensure an effective immune response, the vaccine construct needs to interact properly with the immune receptor molecule. The 3D structure of the TLR5 receptor (3J0A) was retrieved from the Protein Data Bank (PDB) (https://www.rcsb.org/search) (Burley et al. 2023). The docking analysis was done by Cluspro 2.0 (https://cluspro.bu.edu/login.php), which provided the 3D structure of the binding between the ligand and the receptor. Among the predicted models, model 0 was identified as the best and was selected for further analysis (Kozakov et al. 2017).

Molecular dynamics simulation analysis of dock-complex

Molecular dynamics (MD) simulations, the precise computational methods required for studying biological systems at the atomic scale, provide a wealth of information about the structure, interaction, conformational changes, and biomechanical dynamics of biomolecules in a simulated environment. In this study, MD simulations (using Gromacs 2020.4) were instrumental in understanding the molecular-level structural properties and the interactions between the model immune receptors and the vaccine protein candidate (Spoel et al. 2005). The vaccine protein and human receptor structures were meticulously optimized to prepare a system that was then solvated in a cubic box of TIP3P water molecules and neutralized with counter ions to ensure charge balance (Price and Brooks 2004).

Proteins were modeled using the CHARMM36m force field, and solvation was accomplished using the TIP3P water model (Price and Brooks 2004). The system, designed to mimic a cubic simulation box with a 1.0 nm buffer from the protein complex to the box edge, underwent a thorough preparation process. Before the simulation, the system was meticulously relaxed by performing energy minimization using the steepest descent method to relieve steric clashes. Two equilibration steps were then performed, ensuring the system was equilibrated at 300 K for 100 ps in the NVT ensemble and at 300 K and 1 bar for an additional 100 ps in the NPT ensemble to equilibrate the pressure and the volume (Zaheer et al. 2023).

After equilibration, a 100 ns production run was maintained under an NPT ensemble to analyze the vaccine protein candidate and human receptor dynamics. The Parrinello-Rahman barostat was used to maintain a pressure of 1 bar. The V-rescale thermostat was used to maintain a temperature of 300 K. The long-range electrostatic interactions were treated by the use of the Particle Mesh Ewald (PME) method, and a cut-off of 1.0 nm was applied for both van der Waals and electrostatic interactions (Abraham and Gready 2011). Numerical integration was performed with a leapfrog integrator using a time step of 2 fs for stability and accuracy in the simulation.

Molecular dynamics simulation analysis of receptor protein and the vaccine candidate

To further assess intrinsic stability, additional molecular dynamics (MD) simulations were performed on the unbound TLR5 receptor (apo form) and the standalone vaccine construct. Simulations were carried out in OpenMM employing the AMBER ff19SB force field under standard conditions of 300 K temperature and 1 atm pressure (Rahman et al. 2025). Each system underwent 20,000 steps of energy minimization to eliminate steric clashes, followed by equilibration before production. A 100-nanosecond trajectory was generated for both systems (Zaheer et al. 2023; Kashefolgheta et al. 2021). The proteins were solvated in an orthorhombic box using the SPC water model with a 10 Å buffer, and neutralized with 0.15 M Na⁺/Cl⁻ counter-ions to mimic physiological conditions (Zaheer et al. 2023). Post-simulation analyses included the evaluation of structural and dynamic properties such as root-mean-square deviation (RMSD), root-mean-square fluctuation (RMSF), radius of gyration (Rg), and principal component analysis (PCA) to characterize conformational stability and flexibility (Naveed et al. 2025; Zaheer et al. 2023).

Expression analysis

The expression analysis of the vaccine construct was predicted by screening out the nucleotide sequence of the vaccine through the tool EMBOSS Backtranseq (https://www.bioinformatics.nl/cgi-bin/emboss/backtranseq?%20_pref_hide_optional=1) (Rice et al. 2000). The sequence derived from EMBOSS is improved by the codon optimization IDT tool (https://sg.idtdna.com/CodonOpt) (Rice et al. 2000). The hindrance and lower score of the sequence were improved with the more accurate sequence of the vaccine construct. The optimized sequence is further used in Snapgene (https://www.snapgene.com/snapgene-viewer) for the amplification and cloning of the vaccine sequence by E. Coli (Naveed et al. 2022b).

Immune simulation analysis

The potential immune response of the vaccine construct in the human body was simulated using the C-IMMSIM tool (https://kraken.iac.rm.cnr.it/C-IMMSIM/index.php) (Castiglione and Bernaschi 2004; Naveed et al. 2024c). The tool demonstrates the production of antibodies, stimulation of both innate and adaptive immune cells, as well as their presumed count and lifespan. Additionally, it provided graphical predictions of the activation of danger signals, cytokines, and chemokines activation in response to the vaccine.

Results

Identification of target protein

The Poly protein sequence was retrieved from NCBI, consisting of structural and non-structural protein sequences. The NS5B RNA-dependent RNA polymerase of strain HC-J6CH has a highly mutated sequence containing 590 amino acids, along with a 65.5 kDa molecular weight possessing transmembrane helices. The antigenic value of the targeted protein was 0.48, which indicates that predicting its epitopes for our research would be significant. Furthermore, the viral protein was non-allergenic and non-toxic. Table 1. Shows the physicochemical properties of the protein.

Table 1.

Physiochemical properties of NS5B RNA-Dependent RNA Polymerase

Category of protein accession number	Trans-membrane Helices	AA no	MW	GRAVY index	Aliphatic index	Instability index	Theoretical PI	AA score
YP_001469630.1		590	65,543.71	-0.134	87.17	39.86	9.24	0.486

Prediction of epitopes

The Immune Epitope Database (IEDB) identified 18 B-cell Epitopes for NS5B, out of which 4B-cell epitopes qualified for vaccine construction depending on their solubility and antigenic properties, as discussed in Table 2. Furthermore, the IEDB tool predicted a vast quantity of epitopes of both MHC class I and MHC class II alleles, consisting of almost 25,000 to 35,000. An exploratory investigation was carried out to identify a subset of 20 epitopes with an IC50 value of less than 100 for every single MHC class. Subsequently, a screening process was carried out, taking into account each epitope’s antigenicity, non-allergenicity, and non-toxicity After conducting a population coverage analysis of 91.35% worldwide population coverage, as listed in Tables 3 and 4 for MHC class I and II, respectively, five epitopes from each protein were chosen as ultimate targets, along with the most common HLA alleles of those proteins that were recommended by the IEDB server. Table 2 provides a comprehensive summary of B-cell epitopes, and Tables 3 and 4, respectively summarize MHC I & II epitopes with their alleles.

Table 2.

B-cells qualified epitopes with antigenic and non-allergen properties

	B-cell-specific epitopes
Protein ID	Epitope sequence	Antigenic score
YP_001469630.1	LAASKVTARLLTMEE	0.51
	VFCVDPTKGGK	1.9576
	QGTEEDERN	0.8939
	VKTKLKLTPLPEARLLDLSSW	1.3864

Table 3.

MHC I class epitopes qualified for vaccine construct with an IC50 value less than 100

MHC-I-specific epitopes
Protein ID	Epitope sequence	Antigenicity score	Alleles	IC50 value
YP_001469630.1	AVMGASYGF	0.4816	HLA-A*32:01, HLA-A*02:06, HLA-B*15:01	12.69
	FVGVGLFLL	0.9567	HLA-A*02:06, HLA-A*02:01	23.81
	HSVSRARPR	0.7778	HLA-A*31:01, HLA-A*68:01	21.55
	MSYSWTGALI	0.8224	HLA-A*68:02, HLA-A*24:02, HLA-B*58:01	18.23
	LAASKVTAR	1.0528	HLA-A*31:01, HLA-A*68:01	50.16

Table 4.

MHC II class epitopes qualified for vaccine construct

MHC-II-specific epitopes
Protein ID	Epitopes sequence	Amino acid score	Alleles	IC50 value
YP_001469630.1	AKEVRSLSGRAVNHI	0.4011	HLA-DRB1*01:01/ HLA-DRB1*09:01/ HLA-DRB1*13:02/ HLA-DRB1*15:01	11.5
	EVRSLSGRAVNHIKS	0.5309	HLA-DRB1*01:01/ HLA-DRB5*01:01	21.7
	KEVRSLSGRAVNHIK	0.4885	HLA-DRB1*01:01/ HLA-DRB5*01:01/ HLA-DRB1*13:02	13.5
	MGASYGFQYSPAQRV	0.4505	HLA-DRB1*07:01/ HLA-DRB1*04:01/ HLA-DRB1*13:02/ HLA-DRB1*01:01/ HLA-DRB1*04:05/ HLA-DRB1*09:01/ HLA-DRB5*01:01/ HLA-DRB3*01:01	8.6
	YHNKVYCTTTKSASL	0.7482	HLA-DRB1*07:01/ HLA-DRB1*11:01	46.2

For MHC-I, the ANN4.0 algorithm was used, while in the case of MHC-II, the NN-Align method was used. Although we got more than 25,000 epitopes in each case, the epitope filtering was done on the basis of the lowest IC50 value, non-allergen nature, high antigenicity, and non-toxic nature. The epitopes that we selected were positive in all these analyses and were then considered for making a vaccine candidate.

Population coverage analysis

The estimated total coverage of the epitopes is 91.35% of the population, as shown in Table 5; Fig. 2. Interestingly, although the technique used only predicted half of the alleles, the evaluation of MHC class II allele coverage was limited. It is therefore expected that the true population coverage of MHC class II alleles will exceed the current estimate.

Table 5.

Population coverage analysis

Population area	Class	% Coverage	Average_hitb
World	MHC I	71.85	1.06
	MHC II	69.26	1.59
	Combine MHC I & MHC II	91.35	2.64

Fig. 2.

Combined population coverage of the number of epitopes hits/HLA of MHC I & MHC II

Vaccine construct

The candidate vaccine construct exhibits several components, as shown in Fig. 3. The construct is developed by the attachment of epitopes of B-cells, MHC I & MHC II through KK, GGGS, and GPGPG linkers, respectively, along with a 6X Histidine molecule. To boost immunogenicity and solubility, the synthetic MOM gene, acting as an adjuvant, is attached at the N-terminal via an AAY linker. The linkers preserved and improved the structural stability and pliability of the epitopes. Positively charged lysine residues designated KK linkers are added to vaccines to increase their solubility in aqueous solutions and to increase their intrinsic flexibility, which allows the vaccine to migrate across related protein domains. The EAAAK linker contributes to the immunogenicity; meanwhile, the GPGPG sequence serves as a flexible spacer in multi-epitope-based vaccine constructs, effectively separating different peptide epitopes. This separation helps prevent steric hindrance, ensuring that each epitope is appropriately processed and presented by the immune system.

Fig. 3.

AAY, KK, GGGS, and GPGPG Linkers were used for the creation of the final vaccine candidate, comprising B-cell, MHC class I & II epitopes

Immunogenicity & physicochemical interpretation

The vaccine construct demonstrated substantial antigenicity with a score of 0.6322, indicating its potential to induce an immune response. Furthermore, it was confirmed to be non-toxic and non-allergenic, ensuring its safety profile. The construct comprised 269 amino acids and was expected to be electrically neutral with an isoelectric point of 10.41. According to the predictions of Expasy ProtParam, the vaccine candidate has an estimated half-life of 4.4 h in human reticulocytes, suggesting stability and rapid clearance from the body after administration, as detailed in Table 6. Moreover, the solubility score expressed to be 0.889 allocates the hydrophilic nature of the vaccine candidate, which ensures that the construct can be completely dissolved in aqueous solutions. Table 6 provides additional insights into the physicochemical properties of the vaccine construct, detailing its composition and behavior.

Table 6.

Physicochemical analysis of vaccine candidate

S. no	Physiochemical properties	Results
1)	Number of amino acids	269
2)	Molecular Weight	29,835.96
3)	Theoretical protrusion index (PI)	10.41
4)	The estimated half-life is: (mammalian reticulocytes, in vitro)	4.4 h
	(Yeast, in vivo)	> 20 h
	(Escherichia coli, in vivo)	> 10 h
5)	Instability index	37.02
6)	Aliphatic index	70.37
7)	GRAVY (Grand average hydropathicity)	-0.625
8)	Solubility upon overexpression	0.889(soluble)
9)	Antigenic score	0.6322
10)	Allergenicity	Non-allergen

Disulfide bond engineering

The results from the Disulfide by Design 2.13 program show that vaccine constructs epitope models allowed for the formation of several residue pairs at particular geometries favorable for cysteine substitution and subsequent formation of disulfide bonds. Considering the studied angulation pairs of residues Chi3 associating with the 3D geometry of the bonds and the energy of the formed bonds, the more the bond energy, the more favorable. Among the predicted pairing based on the lowest energetic conformation (2.6–3.3 kcal/mol), pairs A1–A163 (ALA–ILE), A3–A53 (TYR–ALA), and A7–A57 (ASN–VAL) exhibited Chi3 angles in the optimum ang genes (± 97/−87). Hence, the configuration of such bonds would be minimally dense and carry considerably less deformation within the architecture of the multi-epitope, as shown in Table 7. On the other hand, pairs A10–A21 (GLN–CYS) and A24–A71 (LEU–CYS) were notable outliers as they were both classified as angular criteria bonds and energetic 4.5 kcal/mol. Although A10 and A21 bonds can be classified as sterically and locally stable bonds, A24 and A71 are bonded with some restrictive rigidity. Overall, pairs with energy values around 3.0 kcal/mol are more preferred, whereas bonds with energy greater than 5.0 kcal/mol are more likely to be speculative, needing additional structural confirmations.

Table 7.

Disulfide bond engineering analysis

Res1 chain	Res1 Seq #	Res1 AA	Res2 chain	Res2 Seq #	Res2 AA	Chi3	Energy	Sum B-factors
A	1	ALA	A	163	ILE	-100.09	2.7	0
A	3	TYR	A	53	ALA	-84.81	2.63	0
A	7	ASN	A	57	VAL	-105.37	3.35	0
A	10	GLN	A	21	CYS	116	7.56	0
A	24	LEU	A	71	CYS	121.1	4.51	0
A	31	GLU	A	77	GLY	-87.4	3.98	0
A	37	SER	A	42	LYS	-102.7	4.05	0
A	59	ALA	A	99	LEU	118.15	6.42	0
A	60	ARG	A	105	ALA	73.27	3.02	0
A	63	THR	A	101	PRO	123.5	2.47	0
A	63	THR	A	105	ALA	-89.07	3.15	0
A	81	LYS	A	84	THR	123.54	8.17	0
A	102	LEU	A	105	ALA	-106.13	2.43	0
A	109	ASP	A	112	SER	102.97	0.77	0
A	113	TRP	A	137	LEU	114.3	4.09	0
A	116	ALA	A	133	GLY	93.56	2.28	0
A	116	ALA	A	136	LEU	-103.29	3.31	0
A	117	VAL	A	133	GLY	-80.73	2.36	0
A	120	ALA	A	129	PHE	-88.01	1.83	0
A	131	GLY	A	161	ALA	107.87	6.35	0
A	145	ALA	A	150	ALA	117.05	2.15	0
A	173	ALA	A	213	LYS	-105.8	6.51	0
A	184	SER	A	198	VAL	-86.12	2.42	0
A	188	VAL	A	194	GLY	-104.3	1.84	0
A	193	PRO	A	197	GLU	96.16	2.1	0
A	193	PRO	A	205	ALA	96.7	2.28	0
A	197	GLU	A	202	SER	119.02	4.43	0
A	231	MET	A	235	TYR	-67.91	6.79	0
A	239	TYR	A	244	ARG	-107.1	3.24	0
A	255	THR	A	258	LYS	95.21	1.69	0

Adding 1–2 additional pairs will likely improve the conformational rigidity of the vaccine construct, which will improve thermal stability and shelf-life. None of the proposed scaffolds overlapping with disulfide sites contained core epitope residues or linker sequences, which suggests that immunogenic accessibility and functional pliability will be maintained. This suggests that disulfide engineering acceptably intersects with antigen presentation. Disulfide bonds, however, are still form-fitted bonds that require an oxidizing cellular environment. Thus, additional silico refinement using molecular dynamics simulations paired with empirical testing will be needed to assess the stabilizing effect.

Prediction of secondary structure

The PSIPRED analysis predicted that the majority of the protein's amino acids form alpha helices, with a smaller portion contributing to the formation of beta sheets and coil structures (Fig. 4A). Further analysis through MEMSAT revealed that most of the construct’s residues are cytoplasmic, which is a good indication of cellular penetration (Fig. 4B). As claimed by membrane-spanning sections, the amino acid residues from 118 to 133 were discovered to be membrane-spanning (Fig. 4C). According to the SOPMA tool, the number of α- helices is 42.38%, extended strands are 17.10%, and random coils are 32.34% (Fig. 4D & E).

Fig. 4.

Characteristics of the vaccine construct. A PSIPRED annotation of the secondary structure indicating helices (pink), coils (grey), and sheets (yellow); B MEMSTAT prediction of vaccine localization in the cell by amino acids; C MEMSTAT depiction of vaccine localization in the cell; D and E α-helices, β-sheets, and irregular coils in the secondary structure predicted by SOPMA server

Tertiary structure prediction & validation

Among the five models predicted by trRosetta for refining the vaccine construct, Model 1 was chosen for further validation and analysis (Fig. 5A). Initially, the configuration had a Rama-favored score of 91.9%. Following refinement, this score improved significantly to 94%, indicating an enhanced conformational state. The Ramachandran plot demonstrated that the majority of the amino acids in the vaccine construct were positioned in the preferred region, accounting for 94% of the total amino acid count (Fig. 5B).

Fig. 5.

A 3D Refined structure of vaccine construct by tRosetta. B Ramachandran Plot of vaccine candidate

Epitope validation analysis

The computational analysis showed that the vaccine candidate experienced expert proteasomal cleavage at a key point of 0.3, as shown in Fig. 6. The tool ELLIPRO revealed a total of twelve epitopes, five linear and seven discontinuous conformational epitopes, depicted in Table 8 and Fig. 7. This finding emphasizes the possibility of the vaccine candidate to stimulate T- and B-cell immune responses. The interactions between continuous and discontinuous B-cell epitopes during vaccine development demonstrated their adaptability and ability to interact with antibodies.

Fig. 6.

Proteasomal cleavage is represented by Netchop

Table 8.

Predicted number of discontinuous epitopes of the vaccine candidate

Predicted discontinuous epitopes
Sr. No	Residue	No. of residue	z-score
1	A:Y248, A:H249, A:N250, A:K251, A:V252, A:Y253, A:C254, A:T255, A:T256, A:T257, A:K258, A:S259, A:A260, A:S261, A:L262, A:H265	16	0.898
2	A:A177, A:K178, A:E179, A:V180, A:R181, A:S182, A:L183, A:S184, A:G185, A:R186, A:A187, A:V188, A:N189, A:H190, A:I191, A:G192, A:P193, A:G194, A:P195, A:G196, A:E197, A:V198, A:R199, A:S200, A:L201, A:S202, A:G203, A:R204, A:A205, A:V206, A:N207, A:H208, A:I209, A:S211, A:K212	35	0.835
3	A:H264, A:H267, A:H268	3	0.773
4	A:A145, A:R146, A:P147, A:R148, A:E149	5	0.698
5	A:C14, A:G18, A:G19, A:R20, A:C21, A:K29, A:E31, A:L62, A:T63, A:M64, A:E65, A:E66, A:K67, A:K68, A:V69, A:F70, A:C71, A:V72, A:D73, A:P74, A:T75, A:G77, A:G78, A:K79, A:K80, A:K81, A:Q82, A:G83, A:T84, A:E85, A:E86, A:D87, A:E88, A:R89, A:N90, A:K91, A:K92, A:V93, A:K94, A:T95, A:K96, A:L97, A:K98, A:L99, A:T100, A:P101, A:L102, A:R106	48	0.673
6	A:H226, A:I227, A:K230	3	0.658
7	A:S240, A:A242, A:Q243, A:V245, A:K246	5	0.579

Fig. 7.

Discontinuous epitopes predicted by ELLIPRO (1–7) show the discontinuous epitopes of HCV. The yellow area represents the epitopes, and the grey structure is the protein

Molecular docking with TLR5 of vaccine construct and protein–protein interaction demonstration

By employing the tool Cluspro 2.0, the vaccine construct’s molecular docking was done with TLR5. The top docked results with TLR5 had the highest members of 53 and the lowest energy score of -1315.9. The docked complex is shown in Fig. 8A. For performing docking, we chose TLR-5 specifically because TLR-5 and TLR-4 are the two receptors that are present on the surface of cells. They are able to attach the epitopes immediately and trigger the immune response in the body, whereas other TLRs are present within the cell which make it difficult to recognize the epitope. The docked complex of TLR5 and vaccine construct is further used for MD simulations because of their lower energy of interaction.

Fig. 8.

A Docked complex of vaccine candidate and TLR-5. B Interacting residue between chain A (vaccine construct) and Chain C (TLR5) demonstrated hydrogen bonding and non-bonded contacts with salt bridges. C All types of bonds present between the docked complex are shown in this image

The PDBSUM database was utilized to visualize the protein–protein interaction within the docked complex between the chains of the vaccine construct and the receptor. The investigation discovered that Chain A of the vaccine established 16 hydrogen bonds with Chain C of TLR5 Fig. 8B. The formation of these bonds took place at multiple residue coordinates, namely [190–43, 263–614, 265–615, 115–627, 115–627, 119–640, 122–640, 123–641, 123–641, 125–641, 126–641, 123–642, 265–646, 244–647, 244–647 & 258–653]. A hydrogen bond was identified between His190 and Gln43, exhibiting a distance of 3.04 Å. On the other hand, His263 and Asp614 are 2.93 Å. Moreover, His265-Ser615 bonded at 2.78 Å. Lys115- Glu627 forms two hydrogen bonds at 2.58 Å and 2.62 Å, respectively. At the same time, Gly119-Phe640 and Tyr122-Phe640 bonded at distances of 3.33 Å and 2.92 Å. Also, Gly123 and Ser-641 made two hydrogen bonds at distances of 3.19 Å and 2 79 Å. In addition, Gly125-Ser641, Gly126-Ser641, and Gly123-Leu642 formed a bond at distances of 2 91 Å, 304 Å, and 3 17 Å, respectively. His265-Cys646 bonded at 3.22 Å, Lys258-Phe653 at 2.51 Å, and Arg244-Thr647 at distances of 2.89 Å and 2.71 Å, creating two hydrogen bonds. The remaining were 173 non-bonded contacts, and 1 salt bridge demonstrated at Lys115-Glu627, bonded at a distance of 2.58 Å. Additionally, 24 non-bonded contacts between Chain A (vaccine construct)-B (TLR5) were observed, with an interference area of 248 Ų and 271 Ų, respectively. Strong interactions were detected, with the majority of hydrogen bond lengths between immune receptors and vaccination ranging between 2 and 3 Å. The 2D interaction diagram can be seen in Fig. 8C. These findings reinforce the claim of the necessity of multiple hydrogen bond interactions between the epitopes and the MHC complex for effective binding to MHC I and MHC II proteins.

Molecular dynamics simulation analysis of dock-complex

The RMSD value, which indicates the stability of the protein–protein complex over a time interval, is a crucial parameter in our study. The value of 1.5 Å, as shown in Fig. 9A, suggests that the complex experiences minor conformational changes during the simulation. While such fluctuations are common in dynamic systems, the fact that the RMSD value remains relatively low indicates that the complex maintains its structural integrity and does not undergo significant destabilizing shifts. This result implies a certain degree of rigidity in the vaccine construct and human receptor, a quality that means the protein can maintain its functional architecture, a crucial factor for its potential efficacy.

Fig. 9.

The graphical representation of molecular dynamics simulations of vaccine construct-TLR 5 complexes done through GROMACS software; A Root mean square deviation (RMSD) of docked complexes; B Root mean square fluctuation (RMSF); C Rate of gyration of docked complexes

The RMSF analysis illustrated in Fig. 9B reveals the dynamism of individual residues within the protein folds by comparing side-chain movements. In this plot, areas with high peaks correspond to higher flexible residues while lower range corresponds to more stable regions. Such information is essential to pinpoint the more dynamic regions of the vaccine construct and human receptor that may be optimized for interaction/binding with other molecules. The identification of flexible regions that are involved in receptor binding could signal a target for therapeutic intervention or a site for modification. In the RMSF graph, the results show the stability of important functional regions, which is significant for understanding the stability of the behavior of the vaccine construct, underlining the importance of this research.

The Rg analysis (Fig. 9C), a measure used to gauge the compactness of the protein–protein complex, tracks the distribution of atoms away from the center of mass. Complementarily, the oscillation of the Rg values around 19.5 Å indicates that the complex does not undergo significant conformational changes during MD simulations and maintains a compact structure throughout the simulations. As the Rg value oscillates, this may be due to local folding changes; however, in general, because the Rg does not change significantly in the context of this aggregate, it strongly indicates that the protein–protein complex is not significantly unfolded or expanded. This result suggests that the folded complex maintains a compact conformation crucial for efficient interaction and could lead to therapeutic action for the specific vaccine construct and human receptor.

Molecular dynamics simulation analysis of receptor protein and the vaccine candidate

To comprehensively evaluate structural stability, molecular dynamics simulations were performed on the apo-TLR5 receptor, the vaccine construct alone, and the receptor–vaccine complex. Comparative analyses of RMSD, RMSF, radius of gyration, and PCA trajectories provided insight into the intrinsic and interaction-induced dynamics of the systems.

The apo-TLR5 receptor rapidly equilibrated, stabilizing at ~ 0.15–0.20 nm after 20 ns, with only minor deviations thereafter, highlighting its intrinsic rigidity. The vaccine construct exhibited higher fluctuations, with RMSD rising steadily to ~ 0.30–0.35 nm, consistent with the flexible and chimeric nature of multi-epitope assemblies. In contrast, the receptor–vaccine complex displayed an intermediate profile, reaching ~ 0.25 nm and maintaining steady oscillations. This suggests that complex formation dampens the intrinsic flexibility of the vaccine while inducing modest structural adjustments in the receptor to accommodate binding.

Per-residue fluctuations revealed distinct patterns across systems. Apo-TLR5 remained rigid overall (< 0.15 nm), with mobility confined to surface loops and termini. The vaccine construct displayed markedly higher RMSF peaks (> 1.5 nm), reflecting linker- and epitope-associated flexibility. Binding in the complex redistributed fluctuations: key receptor interface residues showed reduced mobility, while vaccine epitopes near the interface became stabilized. This mutual damping effect highlights structural complementarity and stable intermolecular contacts.

The apo-receptor preserved a compact organization with an Rg of ~ 1.50 nm. The vaccine construct fluctuated between 1.52 and 1.65 nm, reflecting transient expansion before converging to a more compact fold. The complex exhibited stable Rg values (~ 1.55–1.60 nm) with reduced amplitude compared to the free vaccine, suggesting that binding promotes a more compact and stable global architecture.

The conformational landscape of apo-TLR5 was restricted to overlapping clusters, indicating minimal large-scale motions. The vaccine construct sampled a broader space, exploring multiple conformational basins before settling into a dominant cluster in the later simulation stages. Upon complex formation, the conformational sampling became more restrained, with the trajectory points clustering into a narrower, stable basin. This reduction in conformational entropy underscores the stabilizing effect of receptor engagement on the vaccine structure (Figs. 10, 11).

Fig. 10.

Apo-protein MD simulation throughout 100 ns A RMSD, B RSMF, C Radius of Gyration, and D Principal component analysis

Fig. 11.

Vaccine candidate MD simulation throughout 100 ns A RMSD, B RSMF, C Radius of Gyration, and D Principal component analysis

Codon optimization & insilico vaccine cloning

The codons of amino acids were optimized with the help of the Codon Optimization Tool for better cloning and expression in the host system, which in this case is E. Coli, shown in Fig. 12. The GC content present in the first 800 bp sequence of the vaccine construct is 64%. To facilitate the cloning process, XhoI and AgeI restriction sites were strategically incorporated at the beginning and the end of the codon-optimized sequence. Moreover, an 807 bp sequence was cloned with a PET-9a vector with restriction sites present, which made the clone sequence of the vaccine go from the length of 4341 bp to 5148 bp.

Fig. 12.

A Clone the history image. B Cloned vector of vaccine candidate in E. coli

Immune stimulation of HCV vaccine construct

Our objective was to induce a specific immunological response in the human host, and the data reflected this goal effectively. After the vaccine injection, antigen counts spiked to approximately 660,000 per mL, which then diminished by the fifth day, aligning with the predictions from the half-life study. Some patients might experience nausea within these initial five days post-vaccination. As the antigen levels decreased, the IgM and IgG counts rose significantly, peaking at 6,700 counts (arbitrary scale) on the 13th day post-injection, before gradually declining to 3,000 counts by the 35th day, as depicted in Fig. 13A. This pattern indicates a robust and sustained immune response. The levels of interferon-gamma (IFN-γ) rose significantly to 450,000 ng/mL and remained stable from days 11 to 15 following the injection. Furthermore, minimal levels of the cytokines IL-10, IL-12, and TGF-β were generated. A mild danger signal was noted on the fifth day post-injection, which corresponded with an IL-2 count that spiked to 205,000 ng/mL, peaking on the same day (Fig. 13B). The generation of memory cells exceeded expectations, with 182 cells per mm³ being produced consistently until the 35th day post-injection. The non-memory cells displayed an initial decline with minor fluctuations but then stabilized at 100 cells per mm³, as shown in Fig. 13C. As the antigen levels rose, the presenting cells also increased but then dropped to zero by the sixth day post-injection. Concurrently, active B cells rose to 320 cells per mm³ on the sixth day and maintained this level thereafter (Fig. 13D). The production of minute anergic cells in response to the injection suggests the vaccine's efficacy. Memory helper T cells showed a steady increase, peaking at 300 cells per mL on the tenth day post-injection and then stabilizing (Fig. 13E). Active TH cells gradually rose to 1,700 cells per mm³ by day 10, with a slight reduction to 1,500 cells per mm³ by day 35 (Fig. 13F). CD4 T-regulatory lymphocytes also increased, reaching 140 cells per mm³ on the fifth day before slightly decreasing to 100 cells per mm³ by day 35 (Fig. 13G). The cytotoxic T-cell population remained steady, ranging between 1,095 and 1,155 cells per mm³ without producing any anergic cells. Active CD8 T-cytotoxic lymphocytes showed a gradual increase, reaching up to 700 cells per mm³ by the 35th day (Fig. 13). In addition to the adaptive immune response, the innate immune response also met our objectives. As shown in Figs. 13J–M, nearly 210 macrophage cells per mm³ were observed. The natural killer (NK) varied from 345–373 cells per mm³ on the 10th day post-injection, with a gradual decrease till the 17th day, before fluctuating up to 360 cells per mm³ on the 23rd day, with a decrease till 330 at the 35th day post-injection. Dendritic cells ranged between 200–250 cells per mm³, and epithelial cells peaked at 400 cells per mm³.

Fig. 13.

The immune simulation analysis of the HCV vaccine construct over the span of a month; A Antigen and antibody count; B Chemokine and cytokine stimulation; C B-cell population; D B-cell population per state; E Helper T (TH) cell population; F A comparison of different TH effector cells; G The conversion of TH0 cells to effector TH1 and TH2 cells; H TC cell population; I cytotoxic T (TC) cell population per state; J Macrophage (MA) population; K Natural killer (NK) cell population; L Dendritic cell (DC) population; M Epithelial cells (EP) population

Immunization and its success in activating both arms of immunity in a vaccine model. During the early stages of immunization, the appearance of a pronounced and sustained IgG response is accompanied by the appearance of even low-toned IgM. This suggests an immune class switch, the versatility so necessary for the retention of long-term immunity. Efficient control of HCV particles is achieved through neutralization. The ability to sharply induce IFN-γ and IL-2 in a system also indicates a control Th1 specialization, skewing the system, which is major for the control of viruses due to its promotion of CTL. The immune suppression to IL-10 and TGF-β also depicts the contour of the construct, which seeks to minimize the levels of immune suppression and maximizes the chances of response. This is also shown with the equilibrium levels of effector memory B and T cells, which are more than enough. It enhances the chance of lasting immune memory and minimizes the chance of reinfection. In addition, the sustained increase in functional cytotoxic CD8 + T cells, along with the absence of reactive anergic subpopulations, indicates that the vaccine design successfully stimulates effector CTLs targeting infected hepatocytes. Active participants of the immune response, mainly the innate, along with antigen-presenting cells such as DCs, macrophages, NK cells, and even epithelial cells, in the upper arms of the immune response bear witness to immune control, antigen fatigue, and antiviral post-attack.

The evidence of the sP6-E2-TGF-b1’s potential IgG subclass promoting immunization has strong implications for designing and engineering prophylactic anti-HCV vaccines which might translate to the needed immune tethering via robust Th2 activity and supporting functional IgG subclass tiering, along with robust Th1 activity tethering and supporting “HCV-centered” anti-HCV cellular immune responses, and/or responses needed to be tethered to the IgG-based HCV neutralization activity needed to be deployed and/or used to elicit the IgG-based HCV neutralization activity tethered at the cellular level for deployment when and where needed to achieve robust innate immune activities.

Discussion

The potency of an HCV multi-epitope-based vaccine is an area of research with vast public health implications, as chronic liver diseases due to HCV are prevalent worldwide. Here, we used an immune-informatics docking-based approach to design a vaccine construct that targets conserved regions of the HCV NS5B protein. By doing so, epitopes that produced active T-cell and B-cell immune responses were identified and combined with molecular docking and dynamics simulations to validate the interaction with immune receptors. These observations complement and extend previous work focused on HCV vaccine development.

Our study confirms the findings of Zubair et al., who reported on a computational strategy that can identify conserved epitopes within the NS5B protein of HCV (Zubair et al. 2025). Our work concurs with their study, as they subjectively determined the antigenicity and stability of a vaccine construct by combining immunogenic epitopes with flexible adjuvants and linkers. Both studies demonstrated high clinical relevance and predicted adequate immune stimulation mediated via Toll-like Receptor (TLR) interactions. Similarly, Behmard et al. structural and non-structural proteins are critical to be involved in multi-epitope-based vaccine design as there is an immune advantage over vaccine candidates containing either of the protein targets alone (Behmard et al. 2022). Although their work also utilized TLR3 and TLR4 agonists to enhance immune responses, we validated the use of GPGPG and EAAAK linkers here to stabilize the epitopes and optimize immunogenicity (Naveed et al. 2022b).

Update studies on NS5B-based, as well as multi-epitope HCV vaccines such as those by Zubair et al. (2020), and Behmard et al. (2022), analyzed the capabilities of immunoinformatics-guided vaccine constructs, but they seem to be stunted by coverage prediction population limitations, failure to incorporate structural adjuvants, or lacking a variety of epitope types (Zubair et al. 2025; Behmard et al. 2022). Unlike these studies, the current construct has an increased population coverage of 91.35%, incorporates non-allergenic and non-toxic B and T cell epitopes, employs immunogenicity-enhancing linkers (EAAAK, GPGPG, AAY), and the immunogenic MOM gene adjuvant, along with T-cell and B-cell epitope integration. Further, the molecular docking and 300 ns molecular dynamics simulation go beyond previous reports in the stability and immune interaction profile of the receptor to provide further clarity into receptor stability and immune interaction profiles. All of these improvements strengthen the position of our design as a more efficient and potentially effective vaccine candidate than the previous constructs using the NS5B target or multi-epitope constructs.

This correlates with other studies, which have shown the effectiveness of multi-epitope–based vaccines for other pathogens also (Naveed et al. 2021). Naveed et al. demonstrate that computational immune simulation and molecular dynamics should be used to determine the quality of the vaccine constructs, as done here. However, the prediction of antigenic, non-toxic, and non-allergenic vaccine constructs in both studies was ensured using immuno-informatics tools like VaxiJen and AllerTOP (Doytchinova and Flower 2007; Dimitrov et al. 2014). In contrast, our study addressed the particular challenges posed by the genetic diversity of HCV, which is a key factor in the virus’s ability to escape immune responses. We highlighted the selection of conserved NS5B epitopes to combat immune escape due to viral mutation, thereby contributing to the development of a more effective HCV vaccine.

The data from this study underscore the importance of targeting conserved regions of the NS5B protein with minimal genetic variability, which will maximize the efficacy of the resultant vaccine. The predicted epitopes were found to have good binding affinity with MHC-I and MHC-II molecules, activating both (cytotoxic and helper) T-cell responses. Moreover, molecular docking studies confirmed stable interactions of the vaccine construct with immune receptors, including TLR2 and TLR3. Moreover, the vaccine construct could elicit both innate and adaptive immune responses.

The adjuvants (like β-defensin) used in the vaccine increase the immunogenicity of the vaccine construct by activating innate immune pathways. Flexible linkers GPGPG and EAAAK were used to separate the receptor epitopes to increase the structural stability of the epitopes and to maintain the correct presentation of the antigen to the immune system. In-silico immune simulations predicted a balanced immune response characterized by activation of humoral as well as cellular immunity. The RMDS analysis shows that the protein–protein complex is generally stable, even though minor fluctuations can be found over time. This indicates that the vaccine construct and human receptor retain their conformation during the simulation. The RMSF data show regions of the protein that are more dynamic, especially those implicated in receptor binding, which could be important to improve the binding interaction of the vaccine protein with the immune receptor. Finally, it should be pointed out that the Radius of Gyration analyses indicated that the complex preserves a highly compacted and stable conformation, providing a sense of security and certainty in the overall structure of the vaccine construct and receptor as a viable therapeutic candidate. Collectively, these results demonstrate that apo-TLR5 is inherently rigid and structurally stable, while the designed vaccine is more dynamic due to its modular and linker-rich architecture. Importantly, complex formation balances these properties: the receptor constrains vaccine flexibility, and the vaccine induces subtle but non-disruptive adjustments in the receptor. The convergence of the complex into a stable conformational basin, with compact Rg and moderate RMSD values, validates the docking results and supports the structural feasibility of TLR5-mediated recognition of the vaccine construct. Such results must be consistent with previous studies on epitope-based vaccine designs, confirming that in-silico designed constructs are economical and capable of mounting strong immunological responses.

The comparative MD analyses proved invaluable in clarifying the stabilizing effects of vaccine binding on TLR5 and in capturing conformational adjustments at the molecular level (Daipule et al. 2020). While the unbound vaccine construct displayed pronounced flexibility across epitope- and linker-rich regions, these fluctuations were markedly reduced upon complex formation, indicating binding-induced stabilization. Conversely, the apo-TLR5 receptor exhibited intrinsic rigidity with only loop regions showing appreciable mobility; when bound to the vaccine, these loops became further stabilized, underscoring favorable interfacial interactions. The receptor–vaccine complex as a whole adopted a compact and stable conformation with restricted conformational sampling, as demonstrated by the RMSD, RMSF, Rg, and PCA analyses. Together, these findings highlight the value of comparative MD simulations in elucidating how binding events reinforce structural stability and promote conformational complementarity between designed immunogens and their target receptors.

The current study also addresses HCV’s genetic variability challenge by including epitopes that have good population coverage across HLA allele diversity. This is important, especially in light of HCV’s widespread nature and the varying genetic backgrounds of affected populations. This vaccine construct, by covering a broad population, could be an efficient candidate for universal protection, which is a key component in the potential success of any prophylactic intervention.

Conclusion

Design of a novel multi-epitope-based vaccine for the Hepatitis C Virus: The Conserved NS5B protein was used as a potential target. The vaccine construct exhibited robust antigenicity, immunogenicity, and stability, and predicted wide population coverage with low allergenicity and toxicity. Molecular docking and dynamics simulations corroborated strong interactions with immune receptors, highlighting their capacity to generate humoral and cellular immune responses. Overall, the comparative MD simulations demonstrated that binding to TLR5 confers structural stabilization and conformational complementarity, validating the robustness of the designed vaccine–receptor interaction. Though these results are promising, further experimental validation is crucial to establish the effectiveness and safety of the vaccine candidate for clinical purposes. If successful, this research will provide a powerful tool for preventing HCV infection, instilling confidence in our ability to control the virus.

Limitation and future prospective

This work is a proof of concept for the computational predictions, which might fail to represent actual biological conditions. The immunogenicity, safety, and efficacy of the vaccine’s absence of in vitro and in vivo validation require urgent verification. Moreover, genetic variation among HCV genotypes and differences in human immune responses to HCV may affect the universal applicability of any vaccine. Future work should emphasize experimental validation, establishing the vaccine’s effectiveness and safety. The construct’s applicability can be improved by optimizing it with next-level simulations and exploring adjuvants and delivery systems. However, it’s the potential impact of large-scale population studies that should make us feel hopeful about the future of vaccine development. Implementing these steps is critical for moving the vaccine from design to clinical trials to worldwide use.

Furthermore, the considered assumptions and the associated biases that come with computational methods must be acknowledged, too. Immunoinformatics features training data, algorithms, and scoring routines that may not sufficiently address the nuances of antigen processing, HLA binding, and the activation of T and B cells in vivo. Consequently, the epitope predictions may have a wide range of immunogenicity when evaluated experimentally. Moreover, the NS5B protein, largely conserved, is still subject to the same constraints as the HCV as a whole, with its prodigious genetic diversity and rapid mutation rates to the production of escape variants that erase a vaccine’s protective potential. Practical difficulties in scaling vaccine production pose another challenge, as multi-epitope-based constructs may need intricate systems for suppression, expression, purification, and formulation to maintain stability while keeping costs practical. These shortcomings reinforce the need for in-silico predictions to be paired with extensive experimental and clinical research to transform computationally designed vaccines into safe, practical options for real-world use.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization , Nouman Ali, Adeeba ; Methodology Nouman Ali.; Software, Nouman Ali, Adeeba Ali, Muhammad Sheraz Yousuf, Arsheen Rehman, Saad Muhammad Islam, Bilal Ilyas, and Aamir Ali; Validation, Adeeba Ali, and Bilal Ilyas.; Formal analysis, Muhammad Sheraz Yousuf, Arsheen Rehman and Saad Muhammad Islam.; Investigation, Nouman Ali.; Resources, Aamir Ali.; data curation Saad Muhammad Islam.;.; writing original draft preparation, Muhammad Sheraz Yousuf, Arsheen Rehman and Bilal Ilyas.; writing review and editing, Nouman Ali.; visualization, Adeeba Ali.; supervision, Nouman Ali.; project administration; Nouman Ali.

Funding

This research work received no external funding.

Data availability

All the data generated in this research study have been included in this manuscript.

Declarations

Conflict of interest

The authors declare no competing interests.