Development of an epitope-based vaccine from mycoplasma genitalium adhesion protein: addressing antibiotic resistance through immune-informatics

Abstract

Mycoplasma genitalium is increasingly recognized for its role in severe health conditions, including sexually transmitted infections, ovarian and prostate cancer. The adhesion protein plays a crucial role in the pathogen’s ability to attach to and invade host cells, making it a key target for vaccine development. The need to develop a vaccine against M. genitalium stems from its rising antibiotic resistance, limited treatments and effectiveness. This study focuses on the design and computational evaluation of adhesion protein-based epitope vaccine. Through an immunoinformatic approach, multiple novel cytotoxic T lymphocyte (CTL), helper T lymphocyte (HTL), and linear B-cell epitopes were identified from the adhesion protein, demonstrating strong antigenic, non-allergenic, and immunogenic properties. The vaccine construct’s 3D structure was validated using Ramachandran plot analysis, ProSA, and ERRAT servers, confirming its stability and suitability. Molecular docking studies revealed a high binding affinity of the vaccine with the TLR-2 receptor, further supported by 100 ns molecular dynamics (MD) simulations that confirmed the structural stability and robust interaction of the vaccine with immune receptors. In silico immune simulations using the C-ImmSim server demonstrated the vaccine’s potential to elicit strong humoral and cell-mediated immune responses. Codon optimization for expression in E. coli using the pET-29a(+) vector predicted efficient production of the vaccine. The comprehensive computational analysis, underscores the potential of this epitope-based vaccine as a promising candidate against M. genitalium infections. However, the study emphasizes the necessity of in vitro and in vivo experiments to validate the vaccine’s efficacy and safety before advancing to clinical trials.

Introduction

Sexually transmitted infections are among the most common acute conditions worldwide.¹  Mycoplasma genitalium is implicated in sexually transmitted infections, particularly linked to non-gonococcal urethritis in men. In women, it contributes to inflammatory reproductive tract syndrome such as pelvic inflammatory disease, cervicitis and infertility, underscoring its significant impact on reproductive health. The bacterium plays a crucial role in the manifestation of diverse clinical conditions in both genders within the context of sexually transmitted infections.^2,3 The rising incidence of multidrug-resistant M. genitalium has raised it to the foremost position among sexually transmitted infections (STIs), ranking second only to Chlamydia trachomatis⁴ with more biosafety concerns.⁵ A recent CDC report has flagged this bacterium on the “watch list” of antibiotic resistance threats, highlighting its alarming propensity to develop resistance to multiple antibiotics. Of particular concern are its resistance patterns against macrolides and fluoroquinolones, crucial antibiotics commonly employed in the treatment of various STIs.⁶ The increasing resistance of conventional STI antibiotics emphasizes the urgency of developing a preventive vaccine that could be instrumental in providing a proactive and resilient defense against M. genitalium infections, offering a crucial avenue for effective prevention in the face of growing antibiotic challenges.

M. genitalium is recognized for its varied flask-shaped pleiomorphic structure, obligate intracellular and slow-growing bacterium. It is the smallest self-replicating prokaryote, containing 580 kb long genome which encodes 485 different bacterial proteins.⁷ It was first isolated from a male with non-gonococcal urethritis in 1981 by Tully et al.⁸ Remarkably, it became one of the initial bacterial genomes to be completely sequenced in 1995,⁹ marking a significant breakthrough in understanding the genetic intricacies of this sexually transmitted pathogen. The adhesion P110 protein, known for binding to sialic-acid human receptors, particularly with the neuraminic acid moiety of oligosaccharides, plays a critical role in mycoplasma cytoadherence which is an essential step for infectivity.¹⁰

P140 and P110, found to be mutually stabilized, engage in an interaction resulting in the formation of a transmembrane complex named “Nap.” This complex accumulates at the terminal organelle, a polar structure in M. genitalium.¹¹ Dr Tully and colleagues described the term “Nap” originally refers to the short fibers adorning the outer most surface of terminal organelle, elucidating a structural aspect of M. genitalium’s unique membrane organization.¹² The Nap, formed by the interaction of P140 and P110, play a crucial role in gliding motility and adherence within pneumoniae cluster of mycoplasmas.¹³ Besides their roles in locomotion and adherence, during cell division, P110 and P140 are also involved in terminal organelle replication.¹⁴ Notably, these adhesins serve as primary targets for host antibodies during M. genitalium infections,¹⁵ emphasizing their significance in the host-pathogen interaction. So, developing a vaccine targeting these adherence proteins could potentially provide protective immunity against M. genitalium, offering a strategic avenue for preventive measures. At present, no vaccine against M. genitalium has received FDA approval,¹⁶ emphasizing the need for ongoing research and development in the quest for effective preventive strategies.

Here, an immunoinformatic approach was systematically employed to propose a novel vaccine against M. genitalium. The strategy involved the identification of immune-potentiating, non-toxic and non-allergic regions within adhesion protein of bacterium. Through the integration of linkers, potential epitopes were strategically connected to a potent adjuvant, resulting in the development of a chimeric vaccine. The final construct comprised epitopes from B-cell, cytotoxic T and helper T lymphocyte. Additionally, comprehensive physicochemical analysis, structure validation, docking studies, and MD predicted stability assessments were performed to evaluate the safety and efficacy of the proposed vaccine. The computational selection of vaccine candidates presents a promising avenue for experimental validation in wet labs, offering potential insights into the mitigation of M. genitalium-induced infections.

Material and methods

Retrieval of M. Genitalium reference proteome

The reference adhesion protein FASTA sequence of M. genitalium (G-37/strain ATCC 335300/NCTC 10195) was obtained from the Protein Data Bank (PDB) online server with the accession number 6R41. This specific strain was chosen due to its well-conserved nature, prevalence, and clinical significance compared to other Mycoplasma strains. The retrieved FASTA sequences of protein was then uploaded to the VaxiJen server. The protein was systematically ranked and filtered based on their antigenic scores, with a threshold set at >0.5. Following this, predictions regarding the allergenic nature and toxicity of the protein were carried out using the Allertop and ToxinPred online servers, contributing valuable potential insights into immunogenic and harmful properties of the identified adhesion protein.

CTL epitope prediction

Initially, the NetCTL 1.2 online server, which uses artificial neural network algorithms, was employed to predict MHC-class I binding epitopes.¹ Subsequently, the predicted epitopes were subjected to further analysis by submitting the NetCTL-generated results to the VaxiJen v2.0 and ToxinPred servers,² aimed at predicting protective and non-toxic antigens. Following the screening process with VaxiJen and the ToxinPred servers, the resulting epitopes were further analyzed for immunogenicity using the IEDB server.

HTL and linear epitope prediction

In the screening of HTL epitopes, the study employed the IEDB tool “MHC-II Binding Predictions” as its primary resource. This tool utilizes a combination of three distinct methods, namely Combinatorial Library, Sturniolo and SMM-align to predict MHC-II binding affinity for peptide sequences. To determine the percentile rank for each peptide, the scores of the peptides were compared against the scores of randomly selected peptide of 15-mer from SwissProt database, totaling five million sequences. The Consensus method then generates a rank by computing the median percentile rank derived from the three individual methods.³

B CELL epitope prediction

Epitopes with most immunogenic potential were predicted within adhesion protein for identification of epitopes crucial for vaccine design and immunodiagnostics. To achieve this, a range of tools from IEDB were utilized. These tools included the Kolaskar and Tongaonkar antigenicity scale, which assesses antigenicity based on amino acid sequences, the Emini surface accessibility prediction method, which predicts surface exposure of amino acid residues, and the Karplus and Schulz flexibility prediction, which evaluates protein flexibility. Additionally, the Bepipred linear epitope prediction analysis was employed to identify linear epitopes within the protein sequence.⁴

Epitopes assembly

Linear B-cell and Helper T lymphocyte (HTL) epitopes were fused together using GPGPG linker peptides, while CTL epitopes were linked with AAY linkers. To enhance immunogenicity, the N-terminal position of the vaccine construct was connected to a sequence comprising 125 amino acid residues from the L7/50S ribosomal L12 peptide, serving as an adjuvant, whereas the C-terminal portion was linked with 6HIS linker. The construct underwent evaluation for allergenicity using AllerTOP (2.0) online server, which assesses basic properties related to allergenicity. Subsequently, the construct was analyzed using the VaxiJen server to determine its potential as an antigen capable of eliciting an immune response.⁵ This integrated approach aimed to design a vaccine construct with enhanced immunogenicity and reduced risk of allergenicity for the development of effective vaccines against M. genitalium.

Profiling of immunogenicity and Allergenicity

The vaccine components were rigorously assessed for their antigenicity, toxicity, and allergenicity to ensure their safety and efficacy. The antigenicity of predicted epitopes was evaluated using the VaxiJen 2.0 server. This tool is the first to independently predict protective antigens based just on the physicochemical characteristics of proteins. This method does not require sequence alignment for antigen prediction.¹⁷ An antigenicity threshold of 0.5 was chosen for evaluation while keeping bacteria as target organism. To assess potential toxicity, the ToxinPred server was employed, known for its capability to calculate various physicochemical properties alongside toxicity prediction. Additionally, allergenicity prediction was performed using the AllerTOP server, which applies auto cross covariance (ACC) to transform amino acid sequences into equal length of uniform vectors.¹⁸ These analyses ensure that the predicted epitopes not only elicited an immune response but also remained non-toxic and non-allergenic, crucial factors in the development of safe and effective vaccines.

Physiochemical properties prediction

In this study, the physicochemical properties of the fusion peptide were evaluated to understand its behavior and functionality in diverse environments. The ProtParam tool, accessible via the ExPasy server, was employed to compute various intrinsic properties of the fusion peptide using its primary sequence. These properties involved heoretical isoelectric point (pI), molecular weight, atomic compositions, instability index, estimated half-life, GRAVY, extinction coefficient, and aliphatic index.¹⁹ Each of these parameters provides important information regarding composition and characteristics of the fusion peptide, aiding in the assessment of its stability, structure, and potential function.

Computation of secondary structure

The secondary structure of vaccine construct’s was evaluated through the GOR IV program, revealing insights into its molecular arrangement. The program categorized the secondary structure into three states: coil (C), helix (H), and extended or beta sheets (D).²⁰ For each amino acid, the analysis assigned a likelihood value indicating the probability of its specific secondary structure. Despite the inherent complexity of such predictions, the GOR IV program achieved a mean accuracy of 64.4%, providing a valuable overview of the structural composition and aiding in understanding its potential functional characteristics.²¹

Computation of tertiary structure (3D)

After joining epitopes, trRosetta was employed, leveraging advanced algorithms to predict protein structures with remarkable speed and accuracy. trRosetta combines direct energy minimizations and restrained Rosetta, facilitated by deep layers of residual neural network. This network estimates the orientation and distance distributions between residues, effectively combining constraints from both residue-residue distances and orientations. Unlike methods solely dependent on distance information, trRosetta’s utilization of both distances and orientations provides a more comprehensive understanding of protein structure, enriching the predictive capabilities and enhancing the fidelity of the vaccine’s structure.²²

3D model refinement

The GalaxyRefine web server is well known for refining. It is built on a refinement approach that has been rigorously validated in CASP10-based refinement experiments, GalaxyRefine utilizes a combination of repacking and subsequent MD simulation techniques to relax the protein structure. By applying these methods, GalaxyRefine effectively enhances the quality of global and local structures.²³

3D structure quality and validation analysis

Tertiary structure validation is a critical stage in the 3D model construction as it helps in identifying potential errors and inaccuracies in 3D models.²⁴ To ensure the reliability of the constructed model for the vaccine peptide, several validation tools were utilized. ProSA-web server was employed initially for assessment of overall quality of the protein 3D structure by estimating a total quality score, represented as a Z score. Z-score deviations falling outside the range expected for native proteins indicate possible structural errors. The ERRAT online web-server was utilized to analyze the interactions between non-bonded atoms, particularly in predicting high-resolution crystallography structures.²⁵ The quality of the modeled structure was further evaluated using the Verify 3D server, which provided an overall quality factor. Furthermore, a Ramachandran plot was utilized to assess the validation of 3D structure by indicating the amino acid residues (%) in accepted allowed and rejected disallowed regions. By using these validation tools, we can comprehensively evaluate the reliability and accuracy of tertiary structure of constructed vaccine peptide, ensuring its suitability for further experimental studies.

Docking and interaction studies of vaccine with TLR

The blind molecular docking of 3D vaccine construct and the Toll-like receptor (TLR) receptor was conducted using the ClusPro 2.0 automated web-based program. ClusPro 2.0 facilitates the docking process through three main steps: FFT correlation for the docking of rigid bodies, clustering of docked structure complexes based on RMSD values, and refinement of the docked complex.²⁶ The output of ClusPro 2.0 provides the top 10 best docked complex models for various coefficients. Subsequently, the PDBsum server was utilized to provide structural summaries of all proteins in the Protein Data Bank (PDB), predicting interactions within the docked complex.²⁷ This insight into protein–protein interactions includes details such as salt bridges, hydrogen bonds, non-bonded contacts, and disulfide bonds. To further analyze the molecular interactions, PDBePisa and HawkDock online servers were employed. PDBePisa was used to calculate the energy of solvation, while HawkDock was utilized to predict the MMGBSA of vaccine-receptor docked complex.

Simulated immune response

The C-ImmSim in silico method, available through an online simulation server, played a fundamental role in evaluating immune response profile of the vaccine constructs. This tool provides a comprehensive assessment of both humoral and cellular responses elicited by the vaccine construct, offering valuable insights into its potential efficacy in combating diseases. Simulation parameters were set at default values at different time to ensure a robust and detailed analysis.

Reverse transcription and computational cloning into vector backbone

For an effective plasmid construction that carries the multiple epitope sequence, NovoPro server (https://www.novoprolabs.com/tools/codon-optimization) was employed. This server offers an organism-specific and codon-adjusted form of the DNA sequence of interest, providing essential parameters such as the CAI (codon adaptation index) and GC content. The CAI serves as a measure of how well the codon usage of the DNA sequence matches that of the host organism, with an ideal value of 1.0 indicating optimal adaptation. Additionally, the GC content is important for stability and expression efficiency, with an optimal range typically falling between 30% and 70%. Following the analysis on NovoPro, SnapGene software was utilized for successful cloning into the pET28a (+) expression vector. Moreover, XhoI and BamHI restriction sites were strategically incorporated into the design to facilitate the seamless integration of the chimera into the pET28a (+) vector.

Expression prediction in E.Coli

The solubility prediction for the vaccine construct was performed using the SoluProt tool. SoluProt is an advanced web-based tool designed for predicting the solubility of proteins expressed in Escherichia coli. By analyzing sequence-based features, the tool estimates the likelihood of successful soluble protein expression in E. coli. The results from SoluProt provided valuable insights into the potential expression efficiency of the vaccine construct, aiding in the assessment of its suitability for further experimental validation.

Molecular dynamics simulation

The MD simulation was conducted using GROMACS under the CHARMM36 jul-2021 force field, known for its accurate representation of protein-ligand and protein-water interactions.²⁸ The TIP3P water model was selected for solvation, providing a balanced approach to describe water-protein interactions effectively. Then complex was placed in a cubic water box, ensuring a minimum distance of 10 Å from any box boundary to prevent interaction with its periodic image. The system was then solvated using TIP3P water molecules. To achieve electroneutrality, the system was neutralized by adding an appropriate number of NaCl ions. Next energy minimization was performed using the steepest-descent algorithm to eliminate steric clashes and optimize the geometry of the solvated system. The energy minimization process continued until a maximum force below 1,000 kJ/mol/nm was achieved, ensuring that the system reached a stable, low-energy state, suitable for subsequent equilibration.²⁹ The system was equilibrated in two phases to ensure stability before the production MD simulation. Initially, the system was equilibrated under the canonical ensemble (NVT), where volume, number of particles and temperature were kept constant. The system was maintained at 300 K for 100 picoseconds, with velocity rescaling to achieve the desired temperature. Following the NVT phase, equilibration was performed under the isothermal-isobaric ensemble (NPT), maintaining constant Number of particles (N), Pressure (P), and Temperature (T). This phase was conducted for 100 picoseconds at 1 bar pressure to stabilize the pressure and density of the system.³⁰ The Parrinello-Rahman barostat was used for pressure control. Trajectories were saved every 2 femtoseconds during both equilibration phases to monitor system behavior. After equilibration, a production MD simulation of 100 nanoseconds was performed. The conditions from the equilibration phases (300 K temperature,1 bar pressure) were maintained to ensure consistent environmental settings throughout the production run.

Results

Adhesion protein sequence retrieval

The FASTA sequence of M. genitalium (G-37/strain ATCC 335300/NCTC 10195) was retrieved from NCBI database containing 916 amino acid residues. Figure 1 illustrates the FASTA sequence, schematic representation and 3D model of vaccine construct.

Fig. 1.

Schematic diagram and 3D structure of vaccine construct.

B cell epitopes

After careful protein selection and retrieval of their sequences, we rigorously screened the predicted B-epitopes based on several criteria, including their antigenic potential, non-allergenicity, lack of homology, and absence of toxicity.³¹ Among the numerous candidates, 15 B-cell epitopes were identified and extracted from submitted sequence. These epitopes (Table 1) were ultimately integrated into the vaccine construct. Crucially, all selected epitopes exhibited non-allergenic properties and demonstrated significant antigenicity, with scores ranging from 0.50 to 1.5, surpassing the predetermined threshold of 0.4. This selection process ensures the inclusion of epitopes that are highly likely to evoke robust immune responses while minimizing potential allergic reactions or adverse effects, thereby enhancing the efficacy and safety profile of the vaccine construct.

Table 1.

Selected B-cell epitopes for vaccine construct.

Sequences	Start	End	Length	Antigenicity score	Allergenicity	Toxicity
DSSYQIADQTIH	9	28	12	0.535	Non	Non
ESSGSNNISFDSTSQGEK	68	85	18	1.5409	Non	Non
KYQLDVPN	107	114	8	0.923	Non	Non
EQPLTKYTLNSSLAKEK GKTQREVHLGSGQANQ WTSQRNQHDLNNNPSP NASTGFKLTTGNAYRK LSESWPIYEPIDGTKQG KGKDSSGWSSTEENEA KNDAPSVSGGGSSSGT FNKYLNTKQ	129	251	123	1.124	Non	Non
TNLDWGEDKQKQFVE NQLGYKETTSTNSHNF HSKSFTQP	321	359	39	0.5432	Non	Non
GNHTNNGTTGPIKTAY PVKKDQKSTVK	449	475	27	1.106	Non	Non
LINATPLNSYGDEG	479	492	14	0.81	Non	Non
NYNFKSNQERLPSR	502	515	14	1.193	Non	Non
ELRSAKSSADSTGSDT KVNWSNTQSRYLPV PYNYSEGIIDADGFK RPENRGASVTTFSG LKSIAPDG	529	595	67	0.818	Non	Non
VGLKAGIDPNPVMSG KKA	605	622	18	1.097	Non	Non
NPGNDSLLSTTDNNI APISFSFTPFTAAESA VDLTTFKEVTYNQ	640	683	44	0.504	Non	Non
SSLKPSHDGKQTPVTD	694	709	16	0.917	Non	Non
IELNQDQATTTLDVAP	723	738	16	1.143	Non	Non
EEFSAVIAKDSDQNKI DIYKNNNGLFEIDTQL SNSVATNNGGLAPSY TENR	762	812	51	0.5593	Non	Non
QFLDGNFYDLNS	881	892	12	0.743	Non	Non

MHC epitopes

In the prediction of MHC-I and II restricted epitopes, a comprehensive approach was undertaken by selecting all alleles within the HLA dataset. The epitopes were sorted based on predicted IC₅₀ values,³² while keeping all other parameters at default settings as shown in Table 2. Following this comprehensive screening process, the epitopes after shortlisting were further evaluated based on their antigenicity values, allergenicity, and an additional criterion of having an IC₅₀ value less than 50. As a result, 4 and 9 epitopes were identified as MHC-I and II respectively, as outlined in Table 2. Selected MHC-1 & MHC-II epitopes for vaccine construct.

Table 2.

Selected MHC-1 & MHC-II epitopes for vaccine construct.

Epitope	Sequences	Allele	Start	End	Length	IC50	Antigenicity score	Allergenicity	Toxicity
MHC-1	QENASNKPTW	HLA-B*44:02 HLA-*44:03 HLA-B*58:01 HLA-B*57:01 HLA-B*53:01 HLA-B*40:01 HLA-B*15:01 HLA-A*23:01 HLA-A*24:02 HLA-*30:01 HLA-A*30:02 HLA-*32:01 HLA-B*08:01 HLA-B*51:01 HLA-B*35:01 HLA-A*31:01 HLA-A*26:01 HLA-B*07:02 HLA-A*03:01 HLA-A*68:01 HLA-A*01:01 HLA-A*11:01 HLA-A*02:03 HLA-A*02:01 HLA-A*02:06 HLA-A*68:02 HLA-A*33:01	308	317	10	9.49	1.142	Non	Non
	SIANFSVGLK	HLA-A*11:01 HLA-A*68:01 HLA-A*03:01 HLA-A*31:01 HLA-A*68:02 HLA-A*30:01 HLA-A*02:06 HLA-A*02:03 HLA-A*02:01 HLA-A*30:02 HLA-A*32:01 HLA-A*33:01 HLA-A*26:01 HLA-B*58:01 HLA-B*08:01 HLA-B*15:01 HLA-B*57:01 HLA-B*07:02 HLA-A*01:01 HLA-B*40:01 HLA-B*44:02 HLA-B*51:01 HLA-A*24:02 HLA-B*53:01 HLA-B*35:01 HLA-B*44:03 HLA-A*23:01	599	608	10	8.84	0.7184	Non	Non
	SMWYWVVER	HLA-A*31:01 HLA-A*33:01 HLA-A*68:01 HLA-A*11:01 HLA-A*03:01 HLA-A*32:01 HLA-A*02:01 HLA-A*02:06 HLA-A*02:03	297	305	9	3.96	0.7323	Non	Non
		HLA-A*30:02 HLA-B*15:01 HLA-A*30:01 HLA-A*23:01 HLA-B*35:01 HLA-B*44:02 HLA-A*68:02 HLA-A*01:01 HLA-B*40:01 HLA-B*57:01 HLA-B*58:01 HLA-A*26:01 HLA-A*24:02 HLA-B*07:02 HLA-B*51:01 HLA-B*44:03 HLA-B*08:01 HLA-B*53:01
	YDLNSTIAGV	HLA-A*02:03 HLA-A*02:01 HLA-A*02:06 HLA-A*68:02 HLA-B*15:01 HLA-A*30:02 HLA-A*26:01 HLA-A*01:01 HLA-A*03:01 HLA-A*31:01 HLA-B*44:02 HLA-A*30:01 HLA-B*58:01 HLA-B*40:01 HLA-B*57:01 HLA-B*08:01 HLA-B*35:01 HLA-A*23:01 HLA-B*51:01 HLA-B*44:03 HLA-A*11:01 HLA-A*68:01 HLA-A*33:01 HLA-B*07:02 HLA-B*53:01 HLA-A*24:02 HLA-A*32:01	888	897	10	2.93	0.7117	Non	Non
MHC-II	DTKVNWSNTQSRYLP	HLA-DRB1*07:01 HLA-DRB1*09:01 HLA-DRB1*01:01 HLA-DRB1*13:02 HLA-DRB5*01:01 HLA-DRB1*04:01 HLA-DRB1*04:05 HLA-DRB3*02:02 HLA-DRB1*08:02 HLA-DRB1*03:01 HLA-DRB4*01:01 HLA-DRB1*15:01 HLA-DRB1*11:01 HLA-DRB3*01:01 HLA-DRB1*12:01	543	557	14	7.2	0.9498	Non	Non
	EKPSYVVEFTNSTNI	HLA-DRB1*07:01 HLA-DRB1*01:01 HLA-DRB1*04:05	84	98	14	5.0	0.4401	Non	Non
		HLA-DRB1*13:02 HLA-DRB3*01:01 HLA-DRB1*04:01 HLA-DRB1*09:01 HLA-DRB1*08:02 HLA-DRB1*15:01 HLA-DRB3*02:02 HLA-DRB5*01:01 HLA-DRB1*11:01 HLA-DRB1*12:01 HLA-DRB1*03:01
	GGVVRLNFNPGNDSL	HLA-DRB3*02:02 HLA-DRB1*13:02 HLA-DRB4*01:01 HLA-DRB1*04:01 HLA-DRB1*01:01 HLA-DRB1*04:05 HLA-DRB1*08:02 HLA-DRB1*12:01 HLA-DRB1*09:01 HLA-DRB3*01:01 HLA-DRB1*03:01 HLA-DRB1*15:01 HLA-DRB5*01:01 HLA-DRB1*11:01 HLA-DRB1*07:01	632	646	14	7.5	1.0108	Non	Non
	GPIKTAYPVKKDQKS	HLA-DRB5*01:01 HLA-DRB1*01:01 HLA-DRB4*01:01 HLA-DRB1*08:02 HLA-DRB1*11:01 HLA-DRB1*13:02 HLA-DRB1*03:01 HLA-DRB1*09:01 HLA-DRB1*15:01 HLA-DRB1*04:01 HLA-DRB3*02:02 HLA-DRB1*04:05 HLA-DRB1*07:01 HLA-DRB1*12:01 HLA-DRB3*01:01	458	472	14	9.5	0.6887	Non	Non
	KPSYVVEFTNSTNIG	HLA-DRB1*07:01 HLA-DRB1*04:05 HLA-DRB1*04:01 HLA-DRB3*01:01 HLA-DRB1*08:02 HLA-DRB1*13:02 HLA-DRB3*02:02 HLA-DRB1*01:01 HLA-DRB1*15:01 HLA-DRB1*11:01 HLA-DRB4*01:01 HLA-DRB1*12:01 HLA-DRB1*03:01	85	99	14	5.5	0.5099	Non	Non
	KYTLNSSLAKEKGKT	HLA-DRB3*02:02 HLA-DRB1*04:01 HLA-DRB5*01:01 HLA-DRB1*11:01 HLA-DRB1*01:01 HLA-DRB1*04:05 HLA-DRB3*01:01 HLA-DRB1*03:01 HLA-DRB1*09:01	134	148	14	8.4	0.8921	Non	Non
		HLA-DRB1*08:02 HLA-DRB1*07:01 HLA-DRB1*12:01 HLA-DRB1*13:02 HLA-DRB4*01:01 HLA-DRB1*15:01
	PSYVVEFTNSTNIGI	HLA-DRB1*07:01 HLA-DRB1*04:05 HLA-DRB1*04:01 HLA-DRB1*13:02 HLA-DRB1*08:02 HLA-DRB3*01:01 HLA-DRB3*02:02 HLA-DRB1*01:01 HLA-DRB1*15:01 HLA-DRB1*09:01 HLA-DRB5*01:01 HLA-DRB1*11:01 HLA-DRB4*01:01 HLA-DRB1*12:01 HLA-DRB1*03:01	86	100	14	4.3	0.4856	Non	Non
	RGGVVRLNFNPGNDS	HLA-DRB3*02:02 HLA-DRB1*13:02 HLA-DRB4*01:01 HLA-DRB1*04:01 HLA-DRB1*08:02 HLA-DRB1*04:05 HLA-DRB1*12:01 HLA-DRB3*01:01 HLA-DRB1*03:01 HLA-DRB1*11:01 HLA-DRB1*15:01 HLA-DRB5*01:01 HLA-DRB1*01:01 HLA-DRB1*09:01 HLA-DRB1*07:01	631	645	14	9.4	1.2855	Non	Non

Vaccine construct

Figure 1 demonstrates that the sequence begins with an adjuvant at the N-terminal, which is linked to the N-terminal of the first linker segment, composed of 5 residues. This linker ensures proper spacing and structural orientation before joining to the B cell epitope segment at its C-terminal. The B cell epitopes are connected at their N-terminal to the first linker and extend until their C-terminal, where another 5-residue linker is joined. This second linker bridges the B cell epitopes with the MHC-I epitopes, joining at the C-terminal of the linker to the N-terminal of the MHC-I epitopes. After the MHC-I epitopes, a 3-residue linker follows, connecting at its N-terminal to the C-terminal of the MHC-I segment. This linker then joins to the N-terminal of the MHC-II epitopes, ensuring proper presentation and spacing for immune recognition. The sequence concludes with a histidine tag at the C-terminal end, linked directly to the C-terminal of the MHC-II epitopes, allowing for easy purification and detection of the vaccine protein. Each of these joining is designed to maintain the structural integrity and functional efficacy of the vaccine.

Analysis of immunogenicity and Allergenicity

The in-silico analysis of the designed vaccine against Mycoplasma genitalium using various bioinformatics tools yields promising results. The VaxiJen 2.0 tool predicts the vaccine sequence as a probable antigen with a high score of 0.8530, indicating its potential to elicit a strong immune response. The ToxinPred analysis confirms that all regions of the vaccine are non-toxic, which is crucial for ensuring the safety of the vaccine candidate. Additionally, AllerTop v2.0 predicts the sequence as a probable non-allergen, further supported by its similarity to a known non-allergenic protein (UniProtKB accession number P25054). These results suggest that the designed vaccine is not only likely to be effective in generating an immune response but also safe and unlikely to cause allergic reactions or toxicity in humans, making it a strong candidate for further development.

Physiochemical properties

The ProtParam analysis of the designed vaccine provides valuable insights into its biochemical properties as summarized in table 3. The molecular weight of 73,513.14 Da and predicted isoelectric point (pI) of 6.59 suggests that the protein is slightly acidic and may have a net negative charge at physiological pH. The near-equal number of negatively charged and positively charged residues indicates a balanced distribution of charged amino acids, which can influence the protein’s solubility and interaction with other molecules. The extinction coefficient of 88,240 at 280 nm, whether in oxidized or reduced states, indicates strong absorbance, which is useful for quantifying the protein in solution. The estimated half-life varies significantly across organisms, being relatively short in mammalian cells (1 hour) and yeast (30 min), but much longer in Escherichia coli (>10 h), suggesting the vaccine might be more stable in bacterial systems. For therapeutic use, PEGylation or glycosylation can extend half-life. These approaches could enhance protein expression and practical application. The instability index of 29.93 classifies the protein as stable, meaning it is likely to maintain its structure and function under physiological conditions. The aliphatic index of 56.51 suggests moderate thermostability, and the negative GRAVY score of −0.892 indicates that the protein is hydrophilic, implying good solubility in aqueous environments. Overall, these properties suggest that the designed vaccine is stable, soluble, and likely to be suitable for further experimental studies and potential applications.

Table 3.

Predicted physiochemical properties of vaccine construct on protparam.

Physiochemical properties	Values
Number of amino acids	673
Theoretical pI	6.59
Molecular weight	73513.14
Instability index	29.93 (stable)
Aliphatic index	56.51
Total positively charged amino acids	(Arg + Lys): 68
Total negatively charged amino acids	(Asp + Glu): 71
Grand average of hydropathicity (GRAVY)	−0.892
Coefficient of extinction (in M−1 cm−1 at 280 nm)	88240
Half-life in prokaryotic cell	>10 Hrs

Secondary structure analysis

The secondary structure prediction of vaccine construct targeting the M. genitalium adhesion protein epitopes, as analyzed using the GOR IV tool, indicates a structure primarily composed of random coils (67.01%), followed by extended strands (22.14%) and a smaller fraction of alpha helices (10.85%). The absence of​ pi helices, helices, beta bridges, bend regions, beta turns and ambiguous states indicates a relatively simple secondary structure. The results show that random coils dominate the sequence, particularly in regions that likely correspond to flexible and less structured segments, which could include linker regions and loops. Extended strands (Fig. 1) are distributed throughout the sequence, contributing to the formation of beta sheets, which could be involved in stabilizing the protein structure or facilitating protein–protein interactions. Alpha helices appear intermittently, potentially contributing to the structural integrity of certain regions or playing roles in epitope presentation. Overall, this secondary structure profile suggests that the vaccine design is flexible with some regions of ordered structure, which may be beneficial for epitope recognition and immune system activation. However, the high percentage of random coils might also indicate a degree of structural flexibility that could affect the constructed vaccine stability.

3D vaccine model

The trRosetta results for the 3D structure modeling of the designed vaccine, shown in fig. 1, provide a comprehensive analysis of inter-residue distances and orientations, which are crucial for accurate structural prediction. The contact map, distance, and angular predictions (omega, theta, phi) suggest that the structure has well-defined regions of close contact and specific angular orientations, indicating a likely stable fold. The templates used for modeling, identified by HHsearch against the PDB70 database, show high confidence scores (100.0), with good coverage (ranging from 68.9% to 61.5%) and identity (67.3% to 57.9%) between the vaccine sequence and known protein structures. The high Z-scores further indicate strong structural similarity between the modeled vaccine and the template structures, suggesting that the predicted model is reliable.

The primary template used, 6RUT_G, covers 68.9% of the vaccine sequence with an identity of 67.3%, providing a solid basis for the 3D model. The other templates contribute additional structural insights, helping to refine the model. The alignment data reveal that the secondary structure includes regions of helices and coils, with some predicted disordered regions, which may contribute to the flexibility or functional aspects of the vaccine.

Overall, the trRosetta modeling results suggest that the vaccine has a well-defined and reliable 3D structure, based on homologous protein templates, which supports its potential stability and functionality as an antigen. This structural model can now be used for further refinement and validation studies, as well as for understanding the vaccine’s interaction with immune system components.

Refinement of 3D vaccine model

After modeling the 3D structure of the vaccine was further refined using the Galaxy online server. According to results, GDT-HA (Global Distance Test - High Accuracy) values for the refined models are slightly lower than the initial model but still remain high (around 0.91 to 0.92), indicating that the overall structural quality is well-maintained after refinement. The RMSD values, which measure the average deviation of atom positions between the initial and refined models, range from 0.499 to 0.530 Å, indicating minimal changes in the structure and good consistency with the original model. The MolProbity scores, which assess the quality of the atomic model by considering geometric parameters, improved slightly after refinement, with values ranging from 2.183 to 2.289 compared to the initial score of 2.309. The Clash scores, which measure steric overlaps, decreased in the refined models (ranging from 11.5 to 13.2) compared to the initial score of 12.5, indicating reduced atomic clashes and improved structural quality. The percentage of poor rotamers is low across all models, with minor improvements observed post-refinement. The Ramachandran favored values, which indicate the percentage of amino acid residues in favored regions of the Ramachandran plot, improved significantly, ranging from 87.9% to 88.2% in the refined models compared to 82.4% in the initial model. This suggests better overall backbone geometry and conformational quality in the refined structures. Overall, the refinement process on the Galaxy server successfully improved the geometric and steric quality of the vaccine model, resulting in a more reliable and accurate 3D structure that is likely to be functionally and structurally robust.

Validation of 3D modelled vaccine

In quality assessment as shown in fig. 2, the ERRAT analysis yielded a score of 68%, which, while not optimal, indicates that the majority of the model’s structure is reasonably accurate, though certain regions (highlighted in red) may have significant errors. The Verify 3D score of 88.86% reflects a high percentage of residues that are compatible with their environments in the 3D structure, suggesting that most of the model aligns well with expected structural features based on its sequence. The Ramachandran plot analysis shows total 79.2% located residues in the most favored regions, which is a positive indicator of the model’s overall structural quality, though 3.8% of residues in disallowed regions highlight potential conformational issues that may need addressing. Further validation using the ProSA web tool provides a Z-score of −5.22, which is within the typical range for native proteins of similar size, reinforcing the model’s global structural integrity. The local model quality graph from ProSA shows variations in knowledge-based energy, with most regions displaying acceptable values, though some fluctuations indicate that certain areas of the model might require refinement. Together, these analyses suggest that the model is largely reliable, with a generally favorable structure, though it does contain specific regions that could benefit from further optimization to improve overall accuracy.

Fig. 2.

Quality assessment and structure validation of 3D model of vaccine.

Docking analysis

The docking results from ClusPro 2.0 for the interaction between the vaccine construct and the TLR-2 receptor are summarized in the table and shown in fig. 3 which demonstrated 11 distinct clusters of docking poses. Each cluster represents a group of similar poses, with the number of members indicating how many poses are within each cluster. Clusters with a higher number of members suggest more stable and frequently occurring interactions during the docking simulation. The Weighted Score is a key parameter, with lower scores indicating more favorable binding energy and potentially stronger interactions. The “Lowest Energy” poses within each cluster are of particular interest, as they represent the most energetically favorable configurations. For instance, Cluster 0, which has the highest number of members (130), has a lowest energy score of −1296.2, indicating a very strong potential interaction between the vaccine and TLR-2 in this pose. Other clusters, such as Cluster 6 and Cluster 4, also show low energy scores (−1283.0 and − 1223.4, respectively), suggesting other possible strong binding interactions. The “Center” of each cluster represents an average pose, and these scores are generally higher (less negative) than the lowest energy scores, indicating that while the average pose in a cluster is stable, the most favorable interaction within that cluster is even more so. The docking results suggest that the vaccine has several potential binding modes with TLR-2, with multiple clusters showing strong binding affinities. Cluster 0, in particular, appears to be the most promising, given its large number of members and very favorable lowest energy score, making it a primary candidate for further investigation in the context of vaccine efficacy and receptor binding.

Fig. 3.

Docked complex and interacting residues of vaccine with TLR-2 receptor.

Interaction analysis

The results from PDBsum and PDBePISA provide a detailed analysis of the interactions between the vaccine construct (Chain A) and the TLR-2 receptor (Chain B). A total of 8 salt bridges were identified between the chains as illustrated in fig. 3. Notable interactions include the salt bridge between Lys343, Glu255, Asp434, Asp421, Glu464, Asp427, Asp466 (by Chain A) and Asp160 (Chain B) with a distance of 2.53 Å, and between Glu255 (Chain A) and Asp160, Arg321, Lys422, Arg447, Arg486, Lys488, Arg508 (Chain B) at 2.76 Å. These salt bridges contribute to the stability of the complex by providing strong ionic interactions between the negatively charged aspartate/glutamate residues and positively charged lysine/arginine residues. Additionally, 31 hydrogen bonds forming between the two chains, with distances ranging from 2.7 Å to 3.2 Å. Key residues involved in hydrogen bonding include Tyr496, Lys343, Arg256, Asn253, Gln254, Glu255, Gly297, Ile298, Thr371, Ala372, Thr368, Pro369, Ala376, Asp434, Asp421, Gln426, ILE465, Thr467 (Chain A) with Ser39, Asp160, Arg320, Arg321, Tyr323, Tyr326, Lys347, Asn370, Tyr376, Gln396, His398, Lys422, Arg447, ASN468, Arg486, LYS488, Arg508, Thr532, LYS561 (Chain B). These hydrogen bonds further stabilize the interaction by allowing for specific and directional interactions between the side chains of the amino acids. The interface between the vaccine construct and TLR-2 involves a significant number of residues, covering an interface area of approximately 1760 Å². The interface area is reported as approximately 2004.4 Å² with very low solvation energy gained upon complex formation which was −3.4 kcal/mol for Chain B and − 7.7 kcal/mol for Chain A, indicating that while the interaction is specific, it may not involve extensive desolvation. The ΔiG value, which represents the change in Gibbs free energy upon binding, is −11.1 kcal/mol with a P-value of 0.985, suggesting that the binding is thermodynamically favorable and specific. The low P-value indicates that the interaction is statistically significant, reinforcing the stability of the vaccine-TLR-2 complex. The results indicate a strong interaction between the vaccine and TLR-2 receptor, characterized by multiple salt bridges and hydrogen bonds that stabilize the complex. The interface is substantial, with a large number of residues involved in interactions, further supported by the favorable solvation energy and ΔiG values. This suggests that the vaccine construct is well-suited for binding to TLR-2, which may enhance its immunogenic potential by effectively engaging the receptor and triggering the desired immune response. The MMGBSA (Molecular Mechanics Generalized Born Surface Area) value of −125.05 kcal/mol for the docked complex of vaccine with the TLR-2 receptor, as calculated using the HawkDock server, suggests a strong and favorable binding affinity between the vaccine and the receptor. MMGBSA is a computational method used to estimate the binding free energy of a complex, where more negative values indicate stronger interactions. A value of −125.05 kcal/mol is significantly negative, indicating that the vaccine forms a stable complex with TLR-2, which is crucial for eliciting an immune response. This strong interaction implies that the designed vaccine is likely to effectively engage with TLR-2, potentially leading to a robust activation of the immune response, making it a promising candidate for further development.

Reverse transcription and computational cloning

The codon optimization for the vaccine construct was performed using the JCat server. The CAI improved from 0.63 before optimization to 0.80 after optimization, indicating a significant enhancement in the potential expression efficiency of the construct in the target host. The GC content, which was 63.20% before optimization, was reduced to 50.17% after optimization, aligning closer to the ideal range for many expression systems, particularly in organisms where high GC content might reduce expression efficiency. After optimization the relative adaptiveness was more consistent and higher codon usage, which suggests that the optimized sequence now uses codons that are more frequently utilized by the host organism, leading to potentially higher expression levels. The GC content percentages along the sequence before and after optimization show a more uniform and reduced GC content, which could minimize issues like secondary structure formation in mRNA that can hinder translation. Overall, the optimization process appears to have successfully adjusted the codon usage and GC content to improve the likelihood of efficient and robust expression of the vaccine construct in the E.coli as shown in fig. 4.

Fig. 4.

In silico cloning of vaccine construct.

Expression prediction analysis

The predicted solubility value for the vaccine construct was 0.718, as determined by the SoluProt tool. This value indicates a moderately high probability of the constructed vaccine that will be expressed in a soluble form when produced in E. coli. A solubility score above 0.5 generally suggests a good likelihood of soluble expression, which is crucial for the stability and functionality of the protein in experimental and industrial applications. The value of 0.718 reflects a favorable outcome, implying that the construct has a relatively strong potential for successful expression without forming insoluble aggregates, which is a common issue in protein expression studies. This high solubility prediction supports the further development of the vaccine construct, increasing its viability for experimental validation and potential use in vaccine production.

Immune simulation

The immune simulation results from the C-ImmSim tool, shown in fig. 5(B), provide insights into the immune response generated by the designed vaccine. The antigen (Ag) count, represented by the black line, peaks sharply on day 1, indicating a strong initial presence of the vaccine antigen. This is followed by a rapid decline, suggesting that the immune system is effectively recognizing and processing the antigen. The antibody (Ab) response is represented by different curves for various immunoglobulins (IgM, IgG1, IgG2). The total IgM and IgG response (orange line) peaks around day 10 and then gradually declines, indicating a strong humoral immune response with a good initial production of antibodies. IgM (green line), which is the first antibody produced in response to an antigen, shows a peak around the same time, confirming an effective primary immune response. IgG1 and IgG2 (blue and purple lines, respectively) show sustained responses, with IgG2 peaking slightly later, reflecting the development of a more specialized and mature immune response, which is crucial for long-term immunity. The low response of IgG2 (red line) suggests that while the vaccine effectively induces an immune response, the response may be more skewed towards the IgM and IgG1 pathways. Overall, the simulation suggests that the vaccine is capable of inducing a robust and sustained immune response, with a strong initial reaction followed by the development of long-term immunity. This indicates the vaccine’s potential efficacy in providing protection against M. genitalium.

Fig. 5.

A and B) immune simulation and superimposition of different frames of MD simulation.

MD simulation analysis

In fig. 6, each plot of MD simulation results of vaccine construct docked with the TLR-2 receptor provides insights into the stability, dynamics, and structural properties of the complex over time.

Fig. 6.

MD simulation graphs of docked complex.

The root mean square deviation plot

RMSD measures the overall displacement of the atoms of the vaccine-TLR-2 complex relative to the initial structure as a function of time. The RMSD starts close to 0, as expected, since it measures the deviation from the initial structure. It gradually rises until about 50 ns, indicating conformational changes and adjustments in the complex as it finds a stable binding configuration. After reaching a plateau around 1.0 nm, the RMSD shows minimal fluctuations, indicating that the system stabilizes and reaches equilibrium. The increase followed by a plateau in RMSD suggests that after initial adjustments, the vaccine-TLR-2 complex becomes stable and does not undergo considerable changes, indicating a well-formed interaction.

The root mean square fluctuation plot

RMSF provides information on the flexibility of individual residues within the system over the simulation time. The RMSF plot shows fluctuations in individual residues over the simulation time. Most residues show low RMSF values (below 0.5 nm), indicating they are relatively rigid during the simulation. However, a few regions exhibit higher fluctuations (up to 2.5–3.0 nm), likely corresponding to flexible loops or terminal ends. These regions may not be involved in the core binding interface and are more prone to movement.

The low fluctuation values across most residues indicate that the vaccine construct and TLR-2 receptor remain largely stable, with only a few flexible regions, which is typical in protein–protein interactions.

Solvent accessible surface area (SASA)

SASA measures the surface area of the complex that is accessible to the solvent over time. The SASA plot shows that the surface area exposed to the solvent increases slightly at the beginning, likely due to conformational adjustments of the complex. After around 50 ns, the SASA stabilizes, indicating that the structure reaches an equilibrium where the amount of surface exposed to the solvent remains constant. The stable SASA values after the initial phase suggest that the complex maintains a consistent structure, with no significant unfolding or collapse, further supporting the stability of the vaccine-receptor interaction.

Radius of gyration (Rg)

Rg measures the compactness of the complex. A lower Rg value reflects a compact structure, while a higher value signifies a more extended conformation. The radius of gyration fluctuates between 3.1 nm and 3.4 nm throughout the simulation, with no significant trends of increasing or decreasing. This indicates that the overall compactness of the vaccine-TLR-2 complex remains stable over time, with only minor variations in its size. The consistent radius of gyration suggests that the complex does not undergo significant structural changes, maintaining its overall shape and compactness throughout the simulation, reinforcing the idea of a stable interaction.

Hydrogen bonds

This plot tracks the number of hydrogen bonds between the vaccine and the TLR-2 receptor over the 100 ns simulation. Hydrogen bonds are crucial for the stability of the complex. he number of hydrogen bonds between the vaccine and TLR-2 fluctuates over the 100 ns simulation, ranging from 5 to 15 bonds at different points in time. The initial spike in hydrogen bonds (around 20) indicates strong initial interactions, which then stabilize as the simulation progresses. These fluctuations are normal in molecular dynamics, as hydrogen bonds can form and break as the molecules undergo conformational changes. The presence of these hydrogen bonds suggests that the vaccine construct maintains stable interactions with TLR-2, although their variability might indicate shifts in binding conformation or flexibility in the interface.

Comparative alignment of vaccine-TLR-2 complex MD conformations

The fig. 5(A) shows the alignment of the vaccine construct and TLR-2 receptor complex at different times in ns (0, 10, 50, and 100) of the MD simulation. Green, cyan,violet and yellow colors represent poses at 0 ns, 10 ns, 50 ns and 10 ns respectively. The initial docked structure (0 ns, in green) serves as the reference point, and the subsequent poses show progressive deviations from this starting conformation. The complex at 10 ns shows an initial deviation of 1.971 Å. This indicates that, within the first 10 ns, the structure has adjusted slightly, suggesting some relaxation and adaptation to the interaction environment. The RMSD value of around 2 Å is typical for early relaxation and indicates a stable interaction without significant conformational changes. The complex at 10 ns shows an increase in the RMSD to 2.429 Å, indicating further structural rearrangements. The complex continues to adjust, possibly optimizing the interaction interface between the vaccine and TLR-2. The deviation is still within a reasonable range, suggesting that the interaction is stable, and the observed changes may be due to slight reorientation of secondary structure elements. The complex at 50 to 100 ns shows nore increase in RMSD to 2.793 Å, showing additional structural deviations. The increase in RMSD suggests that the complex continues to evolve but remains stable overall. The change indicates that the protein–protein interface may still be optimizing, but the lack of large deviations suggests that the binding interface is generally maintained throughout the simulation.

Discussion

The rapid rise in antibiotic resistance has pushed global health to a critical tipping point, with projections suggesting it could result in 10 million deaths annually by 2050.³³ In response, there has been a concerted global effort to develop non-antibiotic therapies, with vaccines emerging as a promising alternative for combating bacterial infections.³⁴ Traditional vaccine development, however, is often costly, time-consuming, and involves working with live pathogens, raising significant safety concerns.³⁵ Advances in genomics and proteomics have revolutionized this field, enabling the use of computational tools to design safe and effective vaccines.³⁶ This has led to the development of subunit vaccines against challenging pathogens, with subunit-epitope-based approaches showing particular promise in preventing microbial pathogenicity. These methods allow for the precise selection of vaccine candidates based on the pathogen’s proteome, resulting in products that are efficient, stable, cost-effective, and safe.¹⁶ Several studies have demonstrated the efficacy of using adhesion proteins for vaccine development through in silico methods, supporting the approach taken in this research involving M. genitalium. Ghosh et al. (2011) identified the T2544 adhesion protein in Salmonella enterica serovar Typhi, showing its role in bacterial adhesion and strong immunogenicity, making it a promising vaccine candidate.³⁷ Zhou et al. (2013) explored the ApfA protein in Actinobacillus pleuropneumoniae, highlighting its importance in virulence and its potential to induce cross-serotype immunity in mice.³⁸ Mugunthan et al. (2021) designed a multi-epitope vaccine targeting Mycoplasma gallisepticum cytoadhesion proteins, which demonstrated high antigenicity and strong immune responses, indicating its suitability as a vaccine candidate.³⁹ Schurwanz et al. (2009) focused on Mycoplasma pneumoniae adhesion proteins P1 and P30, creating a chimeric protein vaccine that significantly reduced bacterial adhesion to human cells.⁴⁰ Faezi et al. (2017) introduced a subunit vaccine derived from type IV pili of Pseudomonas aeruginosa, showing potential in promoting pathogen eradication.⁴¹ Akhtar et al. (2021) targeted the agglutinin-like sequence-3 protein of Candida auris to design a multivalent vaccine that displayed strong stability and immune receptor binding.⁴² Choudhry et al. (2024) developed a multi-epitope vaccine against Saprolegnia parasitica by targeting key adhesion proteins, demonstrating robust binding and immune response in fish.⁴³ Finally, Atapour et al. (2021) employed the Kinetoplastid membrane protein-11 in Leishmania as a target for a fusion protein vaccine, showing potential for stimulating immune responses against visceral leishmaniasis.⁴⁴ These studies collectively support the use of adhesion proteins in vaccine development, validating the approach applied to M. genitalium. Identifying immunogenic antigens is a crucial aspect of vaccine development, particularly for in-silico epitope prediction,⁴⁵ as these antigens can interact with immune cells and exhibit immuno-dominance.⁴⁶ In this study, we utilized computational methods to thoroughly assess the P110 adhesion protein for vaccine construction. The study involves predicting linear B-cell and T-cell epitopes with different bioinformatics tools to trigger both cellular and humoral immune responses.⁴⁷ Given the P110 protein’s critical role in bacterial attachment, a key initial step in infection, we selected its linear B and T-cell epitopes for vaccine prediction. These epitopes hold potential for developing vaccines that could effectively target P110 and elicit both humoral and cell-mediated immunity. Notably, predicting B-cell epitopes is vital for vaccine design,⁴⁸ as it identifies the sites of antigen–antibody interactions.⁴⁹ The selection of the P110 adhesion protein as the primary antigen is based on its critical role in M. genitalium pathogenesis. However, incorporating additional virulence factors, such as the P140 protein and Nap complexes, could further enhance vaccine efficacy by broadening immune coverage and targeting multiple mechanisms of bacterial adhesion and immune evasion.

The 3D model of the vaccine constructed by trRosetta, displayed the knowledge-based energy of the 3D vaccine model across the sequence on ProSA web. Fluctuations in the graph indicate variations in model quality at different residue positions, with peaks above zero suggesting regions of potential instability or less favorable energy, whereas regions below zero are more stable. The overall trend remains close to zero, which is generally a good sign. The scatter plot shows the Z-score of the model compared to other experimentally determined structures. The vaccine’s Z-score is around −5.22, which falls within the range of native proteins, indicating that the global quality of the model is acceptable. The distribution of points in the plot shows that models of similar size typically fall within this range, further validating the structure. Verify 3D anlysis graph shows the residue-by-residue compatibility of the model’s 3D structure with its own amino acid sequence. Green dots represent raw scores, and the blue line represents the average score across a window of residues. The majority of the scores hover above zero, indicating that most residues are in an environment consistent with their 3D structure. However, some negative scores suggest areas where the local structure may not be as well defined. The docking results from ClusPro 2.0 for the interaction between the vaccine construct and the TLR-2 receptor revealed 11 distinct clusters of docking poses, each representing a group of similar binding configurations. Cluster 0 emerged as the most promising, with the highest number of members (130) and a favorable lowest energy score, indicating a strong potential interaction. Other clusters, such as Cluster 6 and Cluster 4, also showed low energy scores, suggesting multiple viable binding modes. The overall results imply that the vaccine construct has strong affinity for TLR-2, with several stable binding poses warranting further exploration for vaccine efficacy.In docked complex total Interactions (285) which is sum of hydrogen bonds (31), salt bridges (8) and non-bonded (246) and extensive hydrogen bonding along with salt bridges as well as the significant interface area and solvation energy, all point to a strong binding between the vaccine and receptor molecules. This interaction is crucial for the activation of the TLR-2 receptor, which is known to play a role in the innate immune response. The results suggest that the vaccine construct has the potential to effectively engage TLR-2, leading to a desired immune response. The high number of interactions and the stability of the interface further support the efficacy of the vaccine design in targeting TLR-2. The MD simulation of the in-silico vaccine construct docked with the TLR-2 receptor indicates a stable interaction. RMSD analysis shows that the system stabilizes after an initial equilibration phase, while the RMSF plot highlights overall low flexibility with a few dynamic regions. The number of hydrogen bonds fluctuates but remains consistent, supporting stable interactions. The SASA and Rg confirm that the complex maintains a stable, compact conformation throughout the simulation. These results suggest the vaccine construct forms a robust and stable complex with TLR-2. The comparative alignment analysis revealed a gradual increase in RMSD indicating that the vaccine and receptor are adapting to each other, but no significant disruptions in the structure are observed. This stability is important for the potential efficacy of the vaccine, as it suggests that the vaccine construct maintains a strong interaction with the TLR-2 receptor, crucial for triggering an immune response.

Although this study provides strong computational evidence supporting the designed vaccine’s antigenicity, immunogenicity, and stability, experimental validation remains essential. Future in vitro studies should focus on expressing and purifying the vaccine candidate in a suitable host system, followed by antigenicity and immunogenicity assessments using ELISA and cytokine profiling assays. Additionally, in vivo studies in appropriate animal models should evaluate the vaccine’s immune response, safety profile, and protective efficacy against M. genitalium infection. These experimental validations will be crucial in advancing this vaccine candidate toward clinical applications.

Conclusion

This study presents a computationally designed epitope-based vaccine targeting the adhesion protein of M. genitalium, addressing the urgent need for an effective preventive strategy against this pathogen. The vaccine construct demonstrated strong antigenic, immunogenic, and structural stability characteristics, with promising interactions with immune receptors. While these findings highlight its potential, further in vitro and in vivo studies are necessary to validate its efficacy and safety before clinical translation. Future experimental validation will be crucial in advancing this vaccine toward real-world application.

Acknowledgments

The authors extend their appreciation to the researchers supporting project number (RSPD2025R726), King Saud University, Riyadh, Saudi Arabia. And the University of the Punjab, Lahore, Pakistan.

Author contributions

Conceptualization: HMR. Formal analysis: MB. Methodology: MSK. Validation: TL. Data curation: RS, HB. Investigation resources: FK. Draft preparation: SAK and critical revision of the manuscript: NA, ST.

Conflicts of interest. There are no declared conflicts of interest.

Data and material availability

Data can be provided on demand from M. Rehman: muhammad.rehman@mlt.uol.edu.pk.