A pan-genomic analysis based multi-epitope vaccine development by targeting Stenotrophomonas maltophilia using reverse vaccinology method: an in-silico approach

Abstract

Antibiotic resistance in bacteria leads to high mortality rates and healthcare costs, a significant concern for public health. A colonizer of the human respiratory system, Stenotrophomonas maltophilia is frequently associated with hospital-acquired infections in individuals with cystic fibrosis, cancer, and other chronic illnesses. The importance of this study is underscored by its capacity to meet the critical demand for effective preventive strategies against this pathogen, particularly among susceptible groups of cystic fibrosis and those undergoing cancer treatment. In this study, we engineered a multi-epitope vaccine targeting S. maltophilia through genomic analysis, reverse vaccination strategies, and immunoinformatic techniques by examining a total of 81 complete genomes of S. maltophilia strains. Our investigation revealed 1945 core protein-coding genes alongside their corresponding proteomic sequences, with 191 of these genes predicted to exhibit virulence characteristics. Out of the filtered proteins, three best antigenic proteins were selected for epitope prediction while seven epitopes each from CTL, HTL, and B cell were chosen for vaccine development. The vaccine was refined and validated, showing highly antigenic and desirable physicochemical features. Molecular docking assessments revealed stable binding with TLR-4. Molecular dynamic simulation demonstrated stable dynamics with minor alterations. The originality of this investigation is rooted in the thorough techniques aimed at designing a vaccine that directly targets S. maltophilia, a microorganism of considerable clinical relevance that currently lacks an available vaccine. This study not only responds to a pressing public health crisis but also lays the groundwork for subsequent research endeavors focused on the prevention of S. maltophilia outbreaks. Further evidence from studies in mice models is needed to confirm immune protection against S. maltophilia.

Introduction

Antibiotics possess remarkable potency in their ability to target and eradicate or hinder the proliferation of harmful microorganisms, particularly bacteria. Nevertheless, the reckless and improper use of antibiotics has given rise to a concerning phenomenon known as antibiotic resistance. This resistance has led to a grave global epidemic, rendering previously manageable diseases arduous to treat and resulting in severe consequences (Salam et al. 2023). In 1943, a gram-negative bacterium called Stenotrophomonas maltophilia was first discovered in a sample of human pleural fluid. Initially referred to as Bacterium bookeri, it was later renamed Pseudomonas maltophilia in 1961 (Brooke 2012). This microorganism has been found in various environments such as soil, water, plants, and animals, and it also tends to colonize the respiratory tract of humans. However, S. maltophilia is most commonly associated with infections acquired in hospitals, particularly in individuals with cystic fibrosis, cancer, and other chronic ailments(Aedh, n.d.). Furthermore, S. maltophilia exhibits resistance to multiple drugs, making it a formidable microbial organism (Raad et al. 2023). Few outbreaks of S. maltophilia infections have been documented in healthcare facilities and hospitals. For instance, in one outbreak that occurred in a hemodialysis center in Brazil, 21 patients were infected with S. maltophilia, and three of them succumbed to the infection (Guyot et al. 2013). Further investigations revealed that the outbreak was caused by contaminated water used in the hemodialysis equipment. Stenotrophomonas maltophilia was recognized as an emerging bacterium in 2012, as documented in a study published in the journal Clinical Microbiology Reviews (Brooke 2014). The World Health Organization (WHO) has classified S. maltophilia as one of the most significant multidrug-resistant pathogenic bacteria found in hospitals (Chang et al. 2015). S. maltophilia happens to be an aerobic bacterium belonging to the Xanthomonadaceae family. It possesses the remarkable ability to move, and its dimensions range from 0.5–1.5 µm in length and 0.2–0.5 µm in width (Brooke 2012; Shovon et al. 2023). The S. maltophilia species is made up of two well-known subspecies: S. maltophilia subspecies: maltophilia and S. maltophilia subspecies: rhizophila. Among these, the more prevalent one is the S. maltophilia subspecies maltophilia, which can be found across the world. On the other hand, S. maltophilia subspecies rhizophila is a less commonly found species that is often discovered in soil and water. Once this bacterium enters the human body, it can spread to various regions such as the circulatory system, urinary tract, and brain (Fernandes et al. 2023; Shovon et al. 2023). In some cases, S. maltophilia can cause severe medical conditions like sepsis and meningitis. Immunocompromised individuals, including those with cystic fibrosis, cancer, or HIV, are more susceptible to S. maltophilia infections. Additionally, those who have recently undergone surgery or have medical devices inserted in their bodies are also at a higher risk of infection (Serra Neto et al. 2023). As of now, there is no commercially available vaccine for S. maltophilia. The absence of a vaccine can be attributed to several factors. One possible reason is that S. maltophilia is a unique bacterium, making it difficult to design a vaccine that can effectively target a wide range of strains. Moreover, S. maltophilia is highly adaptable and can quickly develop antibiotic resistance (Terlizzi et al. 2023). To develop a potential vaccine candidate against S. maltophilia, scientists are utilizing pan-genomic analysis and reverse vaccinology. Pan-genomics involves studying the entire protein repertoire of different S. maltophilia strains to identify common proteins that play a role in its virulence (Nageeb & Hetta 2023; Shovon et al. 2023). By targeting these conserved regions, it may be possible to create vaccines that are effective against multiple strains. On the other hand, reverse vaccinology employs computational methods to predict the proteins that are exposed on the surface or secreted by the bacterium and interact with the host’s immune system (Albaqami et al. 2023). This approach allows us to swiftly identify potential vaccine targets without the need for pathogen culture. In this study, pan-genomic studies and reverse vaccinology tactics were employed to develop an in silico multi-epitope-based vaccine against S. maltophilia. We first identified the core proteome of the bacterium and then predicted candidate vaccine proteins using various filters. The vaccination targets were further analyzed to determine potential B and T-cell epitopes (Irfan et al. 2022). Eventually, a chimeric vaccine consisting of multiple epitopes was developed. This proposed vaccine design could prove valuable for experimental researchers working towards the development of an effective vaccine against S. maltophilia.

Materials and methods

Sequence retrieval and pangenome analysis

The NCBI server, an online platform (https://www.ncbi.nlm.nih.gov), was utilized to procure the complete proteome sequences of Stenotrophomonas maltophilia in the highly efficient FASTA format. A comprehensive investigation was conducted on 81 strains of Stenotrophomonas maltophilia, within the extensive collection of 1441 proteomic entries discovered on NCBI. Employing the powerful BPGA software (version 1.3), a comprehensive pan-genomic analysis was executed on the downloaded FASTA datasets, aiming to unveil potential candidates for an innovative vaccine(Chaudhari et al. 2016). Through this process, the pan-genomic inquiries successfully unveiled the core, accessory, and unique proteins that are present across all the strains under scrutiny. These core proteins, essential for the bacterium’s survival and reproduction, exhibit remarkable conservation throughout all strains within the same species(Bell 2017). Consequently, a vaccination strategy that targets these core proteins has a higher probability of achieving success against a wide array of bacterial strains, while remaining resilient to mutations. Hence, to fulfill the objective of discovering a fitting vaccine, a meticulous screening process was implemented, encompassing all the indispensable proteins.

Screening and selection of potential vaccine candidates

The primary focus of vaccine candidates lies in potent proteins that can elicit an immune response within the host body. To identify the most effective candidates, the Core Virulent Factor Database (VFDB) (http://www.mgc.ac.cn/VFs/) was consulted, using the BLASTP program to search for necessary proteins (Liu et al. 2022; Yang et al. 2008). The proteins that exhibited a sequence identification of less than 30% and a bit score exceeding 100 were selected as the most potent contenders for vaccine development while proteins that did not meet these specified parameters were disregarded (Tahir ul Qamar et al. 2021). Surface proteins, being the first point of contact for the immune system, were considered the most likely to trigger an immune response and therefore retained (60). To determine the subcellular localization, the filtered proteomes underwent analysis using the PSORTb server (https://www.psort.org/psortb/) (version 3.0.3) (Ullah et al. 2023; Yu et al. 2010). All proteins except for surface proteins were eliminated. These remaining proteins were deemed strong candidates for vaccine design due to their lower transmembrane helicase content. The transmembrane helicase content was further assessed using the HMMTOP (http://www.enzim.hu/hmmtop/index.php) and TMHMM 2.0 (https://services.healthtech.dtu.dk/services/TMHMM-2.0/) servers, eliminating any proteins that displayed values greater than 0 and 1 (Krogh et al. 2001; Yu et al. 2010). The next step involved investigating the filtered proteins for various physicochemical parameters, such as molecular weight, atomic composition, instability index, theoretical isoelectric point (pI), amino acid composition, aliphatic index, grand average of hydropathicity (GRAVY), and estimated half-life. This analysis was carried out using the ProtParam tool (https://web.expasy.org/protparam/) (Duvaud et al. 2021). Proteins with a thermal stability value exceeding 40 were considered the most suitable candidates for vaccine development, as they could be easily purified. To assess the antigenicity and allergenicity, Vaxijen and AllerTOP v.2.0 servers were employed (Dimitrov et al. 2014; Doytchinova and Flower 2007). The antigenicity threshold was set to 0.4, and proteins with antigenicity scores surpassing 0.6 were identified as ideal vaccine candidates, as they exhibit high reactivity to antibodies and T-cells in the host immune system. Proteins with low antigenicity and allergenicity were discarded. From the pool of eligible proteins, the top three highly antigenic proteins were selected for further investigation.

Immune cell epitope prediction

MHC class I and class II molecules are actively involved in presenting antigens to T cells, a specific type of white blood cell that plays a significant role in building resistance against infections. MHC class I molecules present antigens that originate within the cell to CD8 + T cells, while MHC class II molecules present antigens that originate outside the cell to CD4 + T cells. MHC molecules often contain tiny peptides as antigens, which are produced by proteins that are typically broken down by the immune system (61). To predict MHC-I and MHC-II alleles, a reference set of MHC molecules from the IEDB server (https://www.iedb.org/) was utilized (Irfan et al. 2022; Vita et al. 2019). The IEDB recommended 2020.09 (NetMHCpan EL 4.1) approach and the IEDB recommended 2023.05 (NetMHCIIpan EL 4.1) method were employed for predicting epitopes for MHC class I and class II, respectively (Khazaei and Moghadamizad 2022). B cells play a crucial role in the adaptive immune response and are frequently targeted in vaccine development. Vaccines designed to elicit a strong antibody-mediated, or humoral, immune response often focus on stimulating B cells to produce specific antibodies against the infection. To predict linear B-cell epitopes, the Bepipred linear epitope prediction 2.0 method was utilized through the IEDB server. The predicted epitopes were subsequently selected for further evaluation of their antigenic properties. Additionally, AllerTOP 2.0 was used to identify allergic conditions, and the ToxinPred tool (http://crdd.osdd.net/raghava/toxinpred/) was utilized for toxicity prediction (Gupta et al. 2013). The chosen epitopes were then used to construct a multi-epitope vaccine.

Multi-epitope vaccine construction

The vaccine was ingeniously developed by incorporating multiple highly antigenic epitopes from both MHC class I and class II as well as linear B cells. The MHC class I, MHC class II, and linear B-cell epitopes were skillfully interconnected using the AAY (Ala-Ala-Tyr), GPGPG (Gly-Pro-Gly-Pro-Gly), and KK (bi-lysine) linkers, respectively (Alom et al. 2021). These linkers not only enhance the stability, immunogenicity, and effectiveness of the vaccine but also act as guardians against the emergence of new epitopes (Bhattacharjee et al. 2023). To further augment the potency of the multi-epitope vaccine, the cholera toxin B subunit sequence (Uniprot ID—E9RIX3) was appended to the N-terminal of the vaccine construct via an EAAAK linker (Stratmann 2015). The objective was to amplify the efficacy of the vaccine by utilizing the cholera toxin B subunit (CTB) as an adjuvant, renowned for its ability to enhance immune responses. CTB serves a critical role in developing vaccines against intestinal infections due to its remarkable affinity for the intestinal tract and its capability to modulate the Th1/Th2 balance (Fan et al. 2022). The developed vaccine underwent a comprehensive evaluation of its physiological and chemical properties using the ExPASy ProtParam server. The ExPASy Protparam, an exceptional web-based service, was employed to assess the molecular weight, number of amino acids, instability index, Grand average of hydropathicity (GRAVY), theoretical PI value, and aliphatic index of the vaccine.

Secondary and tertiary structure prediction

PsiPred (http://bioinf.cs.ucl.ac.uk/psipred/) and Sopma (https://npsa-pbil.ibcp.fr/cgi-bin/npsa_automat.pl?page=/NPSA/npsa_sopma.html) servers were employed to predict the secondary structure (Deléage 2017; Fan et al. 2022). PSIPRED and SOPMA employ two distinct strategies to predict secondary structures in proteins. PSIPRED is a more traditional technique that scrutinizes PSI-BLAST output using a feed-forward neural network (Farzan et al. 2023). SOPMA, on the other hand, is an innovative approach that analyzes multiple sequence alignment of proteins by employing a hidden Markov model (Chakma et al. 2023). The anticipation of the tertiary structure was carried out using Robetta (https://robetta.bakerlab.org/) (Kim et al. 2004). Robetta, established by the Baker Laboratory at the University of Washington, is an online platform for 3D modeling and prediction of protein structure. By employing sophisticated algorithms and computational techniques, it generates three-dimensional models of proteins based on their amino acid compositions.

Structure refinement and validation

The tertiary structure of the developed vaccine was enhanced through the utilization of the GalaxyRefine tool, which catalyzed refinement. Employing the GalaxyRefine tool is paramount in the realm of in silico vaccine development, as it plays a pivotal role in ensuring that the vaccine protein is properly folded and possesses the necessary configuration to interact with the immune system (Ko et al. 2012). To assess the constructed 3D structure, the SAVESv6.0 server (https://saves.mbi.ucla.edu/) and the ProSA-web server (https://prosa.services.came.sbg.ac.at/prosa.php) were employed (Laskowski et al. 1993; Wiederstein and Sippl 2007). These servers implicate the Z-score value and the Ramachandran plot score to estimate standard deviations. The Ramachandran plot depicts the phi and psi angles of each residue within a protein, showcasing their configuration on a two-dimensional plane. The phi angle corresponds to the rotation of the N–C bond, while the psi angle corresponds to the rotation along the C–C bond. The Ramachandran plot provides a visual representation of the permissible regions for these angles for each type of residue and the Z-score of a residue serves as an indicator of how much it deviates from the mean value of the Ramachandran plot for its respective residue type (Sobolev et al. 2020). Residues with a Z-score of two or higher are considered to be located within unfavorable areas of the Ramachandran plot.

Disulfide engineering and codon optimization

Disulfide manipulation is a technique that utilizes computational modeling to predict the optimal placement of disulfide bonds within a protein (Craig & Dombkowski 2013). This process is critical in the virtual development of vaccines as it enhances the protein’s ability to trigger an immune response and maintain its stability. During the disulfide engineering phase, the Disulphide by Design 2 (version 2.13) webserver (http://cptweb.cpt.wayne.edu/DbD2/) was widely employed to introduce disulfide bonds into the vaccine structure (Craig & Dombkowski 2013). To ensure optimal expression of the vaccine in Escherichia coli, the Java Codon Adaptation tool (JCat) (http://www.jcat.de/) was utilized to optimize the in-silico codon (Grote et al. 2005). A high expression level in E. coli can be inferred when the vaccine exhibits a favorable codon adaptation index (CAI) value and GC content (Khan et al. 2021). For the designated expression system, the GC content should ideally fall within the range of 30% to 70% with a CAI score of 1 (Srivastava et al. 2023).

Molecular docking analysis

Molecular docking serves as a creative computational technique for predicting the optimal alignment of the ligand with the receptor, allowing it to form a stable and harmonious complex (Grinter and Zou 2014). In this comprehensive study, we undertook the docking of specifically selected MHC-I and MHC-II binding epitopes that have been identified as potential candidates for vaccine development in conjunction with the MHC-I and MHC-II proteins, thereby facilitating a deeper understanding of their interactions. The crystal structures corresponding to both the MHC-I and MHC-II proteins were meticulously retrieved from the RCSB Protein Data Bank (PDB) server, whereby the PDB identification number for the MHC-I protein was designated as 1TMC, while the MHC-II protein was assigned the PDB ID of 3L6F. The execution of the molecular docking procedure, which involved the binding epitopes of MHC-I and MHC-II alongside their respective protein counterparts, was performed through the utilization of the ClusPro 2.0 server, ensuring that advanced computational techniques were employed to achieve accurate results. Furthermore, to effectively engage in this sophisticated molecular docking process, the ClusPro 2.0 server served as an essential tool for facilitating the docking interactions between the receptor proteins and the vaccine candidates that had been developed, thereby enhancing the overall efficacy of the study (Desta et al. 2020). The 3D structure of TLR4 (commonly known as Toll-like receptor 4) was initially obtained in PDB format from the protein data bank (RCSB) (https://www.rcsb.org/). TLR4, renowned for its remarkable effectiveness in identifying diverse pathogen-associated molecular patterns (PAMPs), such as lipopolysaccharides (LPS) present on the surface of Gram-negative bacteria, plays a pivotal role as an immune receptor (Vaure & Liu 2014). The ClusPro 2.0 server was entrusted with the task of analyzing the refined 3D model of the multi-epitope vaccine, which was submitted as the ligand, while TLR4 itself served as the receptor. To gain valuable insight and meticulously examine the bonds established between the ligand and receptor residues within the docked complex, we used the LigPlot program (version 2.2.7) and the BIOVIA Discovery Studio 2022 Client 22.1, two powerful tools that provided us with comprehensive analytical capabilities (Roman & Mark 2011).

Molecular dynamic simulation analysis

A molecular dynamics simulation lasting 50 ns was carried out to determine the consistency of the binding of vaccine-receptor complex structures. The "Desmond v3.6 Program" by Schrödinger was employed within a Linux framework for this purpose (Biswas et al. 2022; Nur Kabidul Azam et al. 2024). The predefined TIP3P aqueous strategy was utilized to establish a predetermined volume with an orthorhombic periodic bounding box shape (Molla et al. 2023). To neutralize the electric power in the structure, ions were randomly dispersed across its chemical solvent environment. The system’s framework was subsequently reduced and made comfortable using the protocol that applied the force field constants OPLS3e, which are included in the Desmond package (Biswas et al. 2022). Each Isothermal-Isobaric ensemble (NPT) assembly utilized overall Nose–Hoover temperature combinations and an isotropic technique, and it was maintained at a temperature of 300 K and atmospheric pressure (1,01325 bar) (Nur Kabidul Azam et al. 2024). This was followed by 50 PS grabbing pauses with an efficiency of 1.2 kcal/mol. The fidelity of this MD simulation was evaluated throughout the entire simulation using the Simulations Interaction Diagram (SID) from the Desmond modules in the Schrödinger suite (MA et al. 2022).

The root mean square deviation (RMSD) that arises from molecular dynamics (MD) simulations functions as an indicator of the average distances resulting from the removal of one molecule from a system over a predetermined duration, in comparison to a reference value (Abdurrahman et al. 2022). In the context of interactions between vaccines and receptors, the RMSD of vaccine atoms fitting into the receptor, across each time frame, is added to the RMSD of protein structural atoms, such as C, foundation, sidechain, and bulkier elements, subsequent to the comparison against the benchmark time (as observed in our 50 ns MD simulation analysis) (Tahir ul Qamar et al. 2021). Concerning the protein complex, the root mean square fluctuation (RMSF) has predominantly been utilized to identify and monitor regional changes in translational structure.

Immune simulation analysis

The ability to promptly anticipate and analyze immune responses to vaccine candidates makes in silico immune simulations crucial for vaccine development. These simulations aid in expediting the development process and simplifying the selection of effective vaccines for further evaluation. This is achieved by identifying the optimal antigen components, doses, and formulations. The developed vaccine was submitted to C-IMMSIM (https://kraken.iac.rm.cnr.it/C-IMMSIM/) in fasta format to assess the potential immune response following vaccination (Stolfi et al. 2022). The results of the analysis were subsequently retrieved for closer examination. In this particular case, a minimum of 28 days between doses was specified (Palatnik-de-Sousa et al. 2022). Three simulated injections were administered at time intervals of 1, 84, and 168, with each time step representing 8 h in real life. The default values of the remaining server parameters were retained.

Codon adaptation and in silico cloning

In order to facilitate the expression of a foreign gene within a designated host organism, it is imperative to perform codon optimization tailored to the specific characteristics of that host. Consequently, the genetic construct was uploaded to the JCat server for the purpose of codon adaptation. In this instance, we selected the commonly utilized E. coli K12 as the host organism, and the entire procedure was executed while adhering to three principal criteria: (1) avoidance of restriction enzyme cleavage sites, (2) exclusion of prokaryotic ribosome binding sites, and (3) prevention of rho-independent transcription termination. The resultant adapted sequence was assessed based on the codon adaptation index (CAI) and the guanine–cytosine (GC) content. Ultimately, the optimized nucleotide sequence was employed for in silico cloning into the pET28a (+) expression vector. The entire in silico cloning process was conducted utilizing SnapGene v5.2.3 software.

Results

Pre-screening phase

Stenotrophomonas maltophilia possesses 1441 proteomic data in the database of the National Center for Biotechnology Information (NCBI). Among these records, 81 entries encompass the complete proteome of different strains. We have successfully retrieved all 81 entries from the NCBI server, where the sizes of pathogen strains varied from 4.1 to 5 Mb, and the average GC contents ranged from 66.1 to 67% (Table 1). In addition to providing an overview of horizontal gene transfer and insights into the species’ evolution, the pan-genomic analysis of the 81 entries obtained from NCBI has further revealed this bacterial species’ core, accessory, and unique gene pools. Figure 1 depicts the number of genomes encompassing pan-genome families and core genome families. Based on the pan-genome analysis, all 81 strains exhibited 1945 core proteins, and on average, there were 1856 accessory proteins, as shown in Table 1. To gain a better understanding of genome evolution, gene orthology, genome complexity, and the identification of pathogenic and therapeutic sequences, the core genes, which are shared by all species’ genomes, are extensively explored. The core genome set refers to the collection of sequences that are shared by all strains, while the pan-genome encompasses all strain genomic sequences (Figs. 2, 3).

Table 1.

Genome statistics of S. maltophilia strains that were entirely sequenced

Organism Name	Strain	Size (Mb)	GC%	No. of core genes	No. of accessory genes	No. of unique genes	No. of exclusively absent genes
Stenotrophomonas maltophilia	2013-SM24	4.53309	66.4	1945	1178	0	0
Stenotrophomonas maltophilia	ACYCa.6E	4.64618	66.4887	1945	1177	0	0
Stenotrophomonas maltophilia	CYZ	4.51769	66.6	1945	1998	23	1
Stenotrophomonas maltophilia	ACYCe.8N	4.63822	66.3447	1945	2282	53	3
Stenotrophomonas maltophilia	ISMMS2	4.50972	66.4	1945	2011	43	0
Stenotrophomonas maltophilia	ISMMS2R	4.50972	66.4	1945	2147	70	6
Stenotrophomonas maltophilia	HW002Y	4.50894	66.5	1945	2143	87	13
Stenotrophomonas maltophilia	PEG-305	4.49551	67.2	1945	1992	0	0
Stenotrophomonas maltophilia	Col1	4.45857	66.5	1945	1992	0	0
Stenotrophomonas maltophilia	CPBW01	4.44433	66.5	1945	1706	348	32
Stenotrophomonas maltophilia	ACYCd.9D	4.43537	66.5	1945	2075	52	3
Stenotrophomonas maltophilia	ACYCb.6H	4.41762	66.5	1945	2374	0	0
Stenotrophomonas maltophilia	ACYCc.3B	4.41646	67	1945	2235	82	1
Stenotrophomonas maltophilia	ACYCb.1 K	4.41071	66.5	1945	2121	64	4
Stenotrophomonas maltophilia	ZT1	4.39147	66.5	1945	2272	54	1
Stenotrophomonas maltophilia	LH-B2	4.12196	66.8	1945	2280	30	4
Stenotrophomonas maltophilia	SG.Y2	4.12052	66.9	1945	2209	41	3
Stenotrophomonas maltophilia K279a	K279a	4.85113	66.3	1945	2255	58	4
Stenotrophomonas maltophilia D457	D457	4.76916	66.8	1945	2095	29	1
Stenotrophomonas maltophilia R551-3	R551-3	4.57397	66.3	1945	1967	51	7
Stenotrophomonas maltophilia JV3	JV3	4.54448	66.9	1945	2331	160	19
Stenotrophomonas maltophilia	T50-20	4.7777	66	1945	2099	18	6
Stenotrophomonas maltophilia	PEG-173	4.76952	66.1	1945	1964	44	0
Stenotrophomonas maltophilia	CSM2	4.73905	66.6	1945	1915	100	3
Stenotrophomonas maltophilia	W18	4.73843	66.1	1945	2060	84	1
Stenotrophomonas maltophilia	2013-SM15	4.70257	66.3	1945	2161	0	0
Stenotrophomonas maltophilia	sm454	4.68544	66.3	1945	2182	175	4
Stenotrophomonas maltophilia	DHHJ	4.68291	66.3	1945	2024	62	1
Stenotrophomonas maltophilia	Sm53	4.68244	66.3	1945	2084	21	4
Stenotrophomonas maltophilia	SKK55	4.67545	66.3	1945	1974	41	1
Stenotrophomonas maltophilia	ACYCa.1 J	4.67028	66.4	1945	2300	63	0
Stenotrophomonas maltophilia	OUC_Est10	4.66874	66.3	1945	1893	64	1
Stenotrophomonas maltophilia	454	4.66643	66.3	1945	2090	71	0
Stenotrophomonas maltophilia	AA1	4.66334	67.4	1945	2311	57	33
Stenotrophomonas maltophilia	NCTC10498	4.66135	66.4	1945	2429	34	3
Stenotrophomonas maltophilia	ACYCb.10 K	4.63699	66.2	1945	1893	21	0
Stenotrophomonas maltophilia	JZL8	4.63543	66.3	1945	1922	102	2
Stenotrophomonas maltophilia	PEG-68	4.63502	66.7	1945	2342	107	1
Stenotrophomonas maltophilia	XL133	4.63014	66.4	1945	2193	88	1
Stenotrophomonas maltophilia	FDAARGOS_649	4.62588	66.5	1945	2031	61	1
Stenotrophomonas maltophilia	2013-SM13	4.6137	67	1945	2130	78	2
Stenotrophomonas maltophilia	2013-SM12	4.61223	67	1945	2253	82	0
Stenotrophomonas maltophilia	NCTC13014	4.59263	66.3	1945	2468	0	0
Stenotrophomonas maltophilia	CF13	4.5917	66.5	1945	2049	105	2
Stenotrophomonas maltophilia	ACYCa.2H	4.58497	66.5	1945	1933	26	5
Stenotrophomonas maltophilia	KMM 349	4.5783	66.2	1945	2066	10	2
Stenotrophomonas maltophilia	FDAARGOS_507	4.57757	66.6	1945	2055	56	0
Stenotrophomonas maltophilia	2013-SM4	4.57585	66.4	1945	2102	32	0
Stenotrophomonas maltophilia	PSKL2	4.57466	66.5	1945	1848	39	2
Stenotrophomonas maltophilia	NCTC10259	4.56388	66.8	1945	1921	29	2
Stenotrophomonas maltophilia	X28	4.55422	66.5	1945	1962	159	0
Stenotrophomonas maltophilia	PEG-390	4.55402	66.4	1945	2024	43	1
Stenotrophomonas maltophilia	MER1	4.5473	66	1945	2001	139	1
Stenotrophomonas maltophilia	HT2	4.54291	66.6	1945	1854	34	0
Stenotrophomonas maltophilia	U5	4.54164	66.4	1945	1877	27	2
Stenotrophomonas maltophilia	CW002SM	4.95906	66.1	1945	2044	24	0
Stenotrophomonas maltophilia	AB550	4.94343	66.5	1945	1989	8	20
Stenotrophomonas maltophilia	SJTH1	4.93232	65.9	1945	2092	25	1
Stenotrophomonas maltophilia	NCTC10498	4.92865	66.3	1945	2079	0	0
Stenotrophomonas maltophilia	WP1-W18-CRE-01	4.92394	66.2	1945	2077	0	1
Stenotrophomonas maltophilia	WGB211	4.91368	66	1945	1957	48	1
Stenotrophomonas maltophilia	SJTL3	4.891	66.3	1945	1824	32	8
Stenotrophomonas maltophilia	SCAID WND1-2022 (370)	4.91903	66.1663	1945	2301	42	3
Stenotrophomonas maltophilia	PEG-42	4.8548	66.1	1945	2024	34	4
Stenotrophomonas maltophilia	FDAARGOS_325	4.85151	66.3	1945	2205	77	0
Stenotrophomonas maltophilia	Stenotrophomonas maltophilia 1800	4.83711	66.2	1945	2102	167	1
Stenotrophomonas maltophilia	142	4.83098	66.2	1945	2277	22	0
Stenotrophomonas maltophilia	FDAARGOS_92	4.82022	66.3	1945	2274	56	2
Stenotrophomonas maltophilia	FZD2	4.81712	66.4	1945	1943	21	8
Stenotrophomonas maltophilia	ISMMS3	4.804	66.7	1945	2373	0	0
Stenotrophomonas maltophilia	NEB515	4.78552	66.4	1945	1597	51	3
Stenotrophomonas maltophilia	GYH	4.94921	66.345	1945	2009	14	5
Stenotrophomonas maltophilia	O1	4.517	66.5	1945	2156	0	2
Stenotrophomonas maltophilia	SoD9b	4.41565	66.8	1945	1998	45	0
Stenotrophomonas maltophilia	NCTC10257	5.00426	66.1	1945	1589	61	7
Stenotrophomonas maltophilia	SM 866	5.08618	66	1945	2468	0	0
Stenotrophomonas maltophilia	NCTC10258	4.48112	66.6	1945	1960	17	0
Stenotrophomonas maltophilia	sm-RA9	5.00758	65.6	1945	2101	14	1
Stenotrophomonas maltophilia	FDAARGOS_1044	5.00427	66.1	1945	2068	42	0
Stenotrophomonas maltophilia	PEG-141	5.0023	66.1	1945	2025	22	1
Stenotrophomonas maltophilia	ICU331	4.99599	66.2	1945	2293	55	0

Fig. 1.

A schematic representation demonstrating the methods used to develop a multi-epitope vaccine against Stenotrophomonas maltophilia

Fig. 2.

81 Stenotrophomonas maltophilia genomes are shown in a pan-core plot. Each additional genome increased the size of the pangenome while decreasing the size of the core genome. The pangenome curve has maintained its rising trend. This indicates that the gene pool of S. maltophilia is likely to expand in the coming years

Fig. 3.

Shortlisted proteins by subtractive proteomics filters

Subtractive proteomics filters

The core proteins that were obtained were submitted to the VFDB database through the BLASTP search engine in order to determine the virulent proteins. Among the core proteins, a total of 191 pathogenic proteins were identified. The inclusion of virulent proteins in vaccine formulations is highly desirable due to their ability to initiate immune pathways and enhance the elicitation of safe immune responses. Since these proteins are only used in limited areas during vaccine design, their incorporation into vaccines is secure and unlikely to negatively affect human cells. The PSORTb server was then utilized to verify the localization of the selected proteins. The pathogenesis, invasion, and colonization of bacteria within host cells heavily rely on surface proteins. Furthermore, the host immune system can effectively recognize the antigenic epitopes of these proteins and initiate innate immune responses. According to the projections made by the PSORTb server, it is likely that 52 proteins are located within the cytoplasmic membrane, while five are situated in the outer membrane, leaving the remaining proteins to be localized within the cell. Subsequently, the HMMTOP and THMMTOP 2.0 servers were employed to confirm the presence of transmembrane helices in the 57 surface proteins. Only proteins with zero or one transmembrane helix were selected, while the others were discarded. Out of the 57 surface proteins that were filtered out, a total of 22 proteins met the criteria for having transmembrane helices. Conducting research and cloning on membranes with more than two helices poses a significant challenge. This assumption was made due to inadequate protein expression in vitro systems such as E. coli. Additional analysis was conducted on the selected proteins, including their molecular weight, atomic composition, instability index, theoretical isoelectric point (pI), amino acid composition, aliphatic index, grand average of hydropathicity (GRAVY), and anticipated half-life. The instability index was the most important factor in evaluating the physiochemical properties. A threshold value of 40 was set for the instability index. If a protein had an instability index greater than 40, it was considered unstable and excluded from the study. Conversely, proteins with an instability index below 40 were considered stable and subjected to further examination. Based on the instability index criteria, a total of 14 proteins were deemed stable. These 14 proteins were then further investigated for selection based on predictions of their antigenicity, allergenicity, and toxicity. Out of the initial 1945 core proteins, 11 out of the final 14 proteins were anticipated to be antigenic and non-allergenic (Table 2). The top three antigenic proteins were chosen for additional investigation.

Table 2.

Different physicochemical analysis of virulent proteins; subcellular localization, Transmembrane helices, instability index, Antigenicity, Allergenicity, and Toxicity

ID	Subcellular localization	HMMTOP		TMHMM	Instability index	Antigenicity		Allergenicity	Toxicity
> CORE_REP| Org37_Gene2333#	Cytoplasmic Membrane 10.00	1		2	32.25	0.5526		NON	NON
> CORE_REP| Org55_Gene2869#	Outer Membrane 10.00	1		0	22.75	0.6772		NON	NON
> CORE_REP| Org61_Gene2356#	Outer Membrane 10.00	0		0	34.64	0.6396		NON	NON
> CORE_REP| Org65_Gene1269#	Cytoplasmic Membrane 9.97	0		0	34.5	0.6363		NON	NON
> CORE_REP| Org43_Gene3823#	Cytoplasmic Membrane 9.99	0		0	32.78	0.5401		NON	NON
> CORE_REP| Org40_Gene905#	Cytoplasmic Membrane 9.99	0		0	38.62	0.5198		NON	NON
> CORE_REP| Org70_Gene1716#	Outer Membrane 9.93	0		0	37.17	0.5596		NON	NON
> CORE_REP| Org60_Gene83#	Outer Membrane 9.93	0	0		37.52		0.7839	NON	NON

Epitope prediction and prioritization phase

B and T cell binding epitopes serve a crucial purpose in the field of vaccine development as they indicate specific regions of pathogens that elicit immune responses. The integration of these epitopes into vaccines ensures precise and efficient immune responses, leading to the production of antibodies and memory T cells that can provide persistent protection against specific pathogens. To anticipate the epitopes that activate CD8 + T cells, CD4 + T cells, and B cells, the three most antigenic proteins were submitted to the IEDB server (> CORE_REP|Org55_Gene2869#, > CORE_REP|Org60_Gene83#, > CORE_REP|Org54_Gene653#). The server was accessed to retrieve CD8 + T cell epitopes and CD4 + T cell epitopes based on their respective scores. Subsequently, epitopes with scores greater than 0.5 were obtained from the server for further investigation. The alleles HLA-A*01:01, HLA-A*02:01, HLA-A*02:06, HLA-A*03:01, HLA-A*11:01, HLA-B*07:02, HLA-B*08:01, and HLA-B*15:01 were utilized to predict the MHC-I epitopes. In the case of MHC-II epitopes, the alleles HLA-DRB1*07:01, HLA-DRB1*03:01, HLA-DRB3*01:01, HLA-DRB3*02:02, HLA-DRB4*01:01, and HLA-DRB5*01:01 were employed for analysis. Each B cell epitope generated by the server was collected for further examination. Through the utilization of various immunoinformatic approaches, we identified the most effective vaccine candidate as a water-soluble, non-toxic, non-allergenic, and antigenic protein. This involved the prioritization of epitope selection.

Multi-epitope vaccine construction

Based on the antigenic characteristics they possess, a selection of seven MHC-I epitopes (Table 3), seven MHC-II epitopes (Table 4), and seven B-cell epitopes (Table 5) was created in order to create a combination of multi-epitope vaccine. The epitopes were combined using AAY, GPGPG, and KK linkers, as depicted in Fig. 4, and the resulting peptide was connected to the adjuvant cholera toxin B subunit (Uniprot Id—E9RIX3) using an additional “EAAAK” linker. The constructed vaccine was then subjected to the ProtParam tool on the Expasy server in order to assess its physiochemical attributes. The instability index of the vaccine was determined to be 26.91, which classifies it as stable. Furthermore, the grand average of hydropathicity was calculated to be −0.803, indicating that the vaccine is hydrophilic. More comprehensive information regarding the physiochemical properties of the multi-epitope vaccine can be found in Table 6.

Table 3.

Selected CTL epitopes for the development of the vaccine

Allele	CD8 + Epitope	Antigenicity	Allergenicity	Toxicity
HLA-A*01:01	SSEKGKLSY	1.2644	Non-Allergen	Non-Toxic
HLA-A*03:01	RIYYPVPAY	1.4076	Non-Allergen	Non-Toxic
HLA-B*07:02	EPNWNPLAL	1.0055	Non-Allergen	Non-Toxic
HLA-A*02:06	KQQERAVNL	1.0622	Non-Allergen	Non-Toxic
HLA-A*03:01	GLQDAYAKK	1.2781	Non-Allergen	Non-Toxic
HLA-B*15:01	TLKDRNGGY	1.0801	Non-Allergen	Non-Toxic
HLA-A*01:01	DSGLDLALY	1.2433	Non-Allergen	Non-Toxic

Table 4.

Selected HTL epitopes for the development of the vaccine

Allele	CD4 + Epitope	Antigenicity	Allergenicity	Toxicity
HLA-DRB5*01:01	KAEYEKAAAENKTKSDQ	1.4446	Non-Allergen	Non-Toxic
HLA-DRB1*07:01	TTGESNFDRTTGAGISP	1.1543	Non-Allergen	Non-Toxic
HLA-DRB1*07:01	GESNFDRTTGAGISP	0.9414	Non-Allergen	Non-Toxic
HLA-DRB3*01:01	DADLTPDTQLSVGYD	0.9881	Non-Allergen	Non-Toxic
HLA-DRB5*01:01	SGVQYRVIEAGKGAK	1.2919	Non-Allergen	Non-Toxic
HLA-DRB5*01:01	GVQYRVIEAGKGAKPTQ	1.0946	Non-Allergen	Non-Toxic
HLA-DRB3*01:01	VIDADLTPDTQLSVG	0.9985	Non-Allergen	Non-Toxic

Table 5.

Selected CTL epitopes for the development of the vaccine; where AN = antigenicity, AL = allergenicity, TOX = toxicity, ES = estimated solubility

Epitope	AN	AL	TOX	ES
RAEGYSVRRTSAGTRFDLAPREIPQ	1.1842	Non-Allergen	Non-Toxic	GOOD
DTDGQMDRYNQR	1.5806	Non-Allergen	Non-Toxic	GOOD
DYQHKRANGA	1.5678	Non-Allergen	Non-Toxic	GOOD
DWKSEGEGADRAHKVT	1.4406	Non-Allergen	Non-Toxic	GOOD
NLAELTGRGEQLDIN	1.4165	Non-Allergen	Non-Toxic	GOOD
QKREQGRAQAAKAEYEKAAAENKTKSDQFIAANKAKAGVQSLPS	0.9070	Non-Allergen	Non-Toxic	GOOD
IEAGKGAKPT	1.2443	Non-Allergen	Non-Toxic	GOOD
NLKPKDKTSGNARSG	2.2945	Non-Allergen	Non-Toxic	GOOD

Fig. 4.

A schematic illustration of the final multi-epitope vaccination development

Table 6.

Physicochemical characteristics of the developed vaccine

Characteristics	Result
Number of amino acids	523
Molecular weight	56,054.94 da
Theoretical pI	9.55
Total number of negatively charged residues (Asp + Glu)	57
Total number of positively charged residues (Arg + Lys)	82
Chemical formula	C2465H3916N712O769S7
Total number of atoms	7869
Estimated half-life (mammalian reticulocytes, in vitro)	1 h
Estimated half-life (yeast, in vivo)	30 min
Estimated half-life (Escherichia coli, in vivo)	> 10 h
Instability index	26.91
Aliphatic index	60.46
Grand average of hydropathicity (GRAVY)	-0.803

Secondary and tertiary structure prediction of MEV

The investigation examined the secondary and tertiary configurations of the developed vaccine utilizing sophisticated computational tools. Specifically, the utilization of the PsiPred and Sopma servers allowed for the anticipation of the secondary structure of the immunization. The PsiPred server disclosed that the immunization consists of 40% alpha helix, 10.32% beta-strand, and 49.68% random coil, as demonstrated in the accompanying Fig. 5A. Similarly, the Sopma server anticipated that the immunization is comprised of 33.84% alpha helix, 14.34% beta-strand, and 51.82% random coil, as exhibited in the aforementioned Fig. 5B. Additionally, the Robetta server was employed to anticipate the tertiary structure of the immunization, enabling the visualization of its three-dimensional arrangement, as portrayed in the accompanying Fig. 6. In order to facilitate the visualization of the tertiary structure, the Discovery Studio 2021 software was employed. The intricate three-dimensional architecture of the multi-epitope vaccine underwent a meticulous process of prediction concerning discontinuous, also referred to as conformational, B-cell epitopes, which are crucial for eliciting an immune response. Utilizing the advanced capabilities of the Ellipro server, a comprehensive total of 255 distinct epitopes was meticulously identified and projected, thereby providing a significant insight into the potential immunogenic components of the vaccine. Furthermore, an elaborate compilation detailing the specific characteristics and features of the identified discontinuous B-cell epitopes associated with the developed vaccine is systematically presented in Table 11, which serves as a valuable resource for further investigation and understanding of the vaccine’s efficacy. This extensive analysis not only enhances our comprehension of the vaccine’s immunological profile but also lays the groundwork for future research endeavors aimed at optimizing the vaccine’s design and effectiveness in combating relevant diseases.

Fig. 5.

Secondary structure prediction of the developed multi-epitope vaccine. (A) Secondary structure prediction by PsiPred server (B) Secondary structure prediction by SOPMA server

Fig. 6.

The designed vaccine construct’s tertiary structure is shown by the red (a-helix), cyan (b-strand), and green (random coil) colors of its three domains

Vaccine refinement and structure validation

The vaccine that was developed underwent analysis using the GalaxyRefine tool, which was employed to enhance the structure of the vaccine. This tool facilitates the correct folding of the MEV and enables the MEV to interact with the immune system more efficiently. The server generated 5 revised models based on various factors such as GDT-HA, RMSD, MolProbity, clash score, poor rotamers, Rama preferred residues percentage, and GALAXY energy. After considering the information provided by the server in Table 7, Model 3 was selected as the refined candidate for the MEV. The refined 3D model of the vaccine was further assessed using the SAVESv6.0 and ProSA-web servers. The analysis of the improved vaccine’s Ramachandran plot showed that 92.3% of the residues were in the favorable zone, 5.1% were in allowed regions, and 1.6% were in prohibited regions, as depicted in Fig. 7A. In comparison, the crude model had a Z-score of -8.77 (Fig. 7B). To achieve superior 3D-1D profiling, it is recommended that at least 80% of the amino acids in a model should have a score greater than 0.1. The refined model, in this case, contains 83.75% of residues with an average 3D-1D profiling score higher than 0.1.

Table 7.

The GalaxyRefine server’s results. Based on the GDT-HA, RMSD, MolProbity, Clash score, Poor rotamers, and Rama favored parameters, Model 3 was picked as the best-refined model

Sequence Number	AA	Sequence Number	AA	Chi3	Energy	Sum B-Factors
167	TYR	197	ALA	-94.93	0.21	0
170	GLY	193	GLN	87.04	0.62	0
174	ALA	190	ALA	-83.91	1.07	0
112	VAL	119	HIS	-87.62	1.31	0
202	ALA	220	GLU	85.74	1.92	0
152	TYR	162	ALA	98.37	2.19	0

Fig. 7.

Validation of the tertiary structure of the vaccine. (A) The statistics depicted on the Ramachandran plot showcase the regions that are deemed most favorable, additional, and disallowed, with percentages of 92.3, 5.1, and 1.6%, respectively. Furthermore, (B) The refined vaccine model is represented on the ProSA-web, where it exhibits a Z-score of -8.77

Disulphide engineering and in-silico codon optimization

One of the primary objectives of protein engineering is to enhance protein stability. A reasonable approach involves augmenting the naturally occurring molecular interactions that stabilize proteins. Disulfide bridges are covalent connections that confer substantial structural stability. Based on the energy derived from Disulphide bonds, we elected to substitute 6 amino acid residues in the mutant version of the vaccine design (Fig. 8A, B). Specifically, we exclusively replaced six amino acids with energies lower than 2.20 (Kcal/mol) (Table 8). In the vaccine mutant structure, the yellow-colored stick signifies the altered pair of amino acids (Fig. 8B).

Fig. 8.

(A) Wild structure of the developed vaccine (B) mutated structure in which the inclusion of disulfide bonds is indicated by the presence of yellow bands in the mutated structure

Table 8.

Sequence number, Chi value, Energy, and sum B-Factors of amino acid (AA)

Model	GDT-HA	RMSD	MolProbity	Clash score	Poor rotamers	Rama favored
Initial	1.000	0.000	3.391	24.1	16.4	71.3
Model 1	0.9381	0.466	2.586	24.9	1.7	88.3
Model 2	0.9434	0.451	2.456	24.2	1.1	89.5
Model 3	0.9561	0.443	2.442	23.6	0.9	91.2
Model 4	0.9327	0.447	2.521	24.5	1.4	88.9
Model 5	0.9396	0.453	2.593	24.3	1.8	88.6

Subsequently, the reverse translated DNA sequence was inserted into the pET28a(+) vector (Codon Usage modified to Escherichia coli (strain K12). The improved sequence displays a CAI-value of 1, indicating its suitability as a promising vaccination candidate. The improved sequence exhibits a GC content of 50.92415551306565% (GC content should range between 30–70% for optimal vaccine candidates).

Molecular docking studies

The intricate process of molecular docking, which involves the interaction between MHC-I and MHC-II binding epitopes with their corresponding MHC-I and MHC-II proteins, was meticulously conducted utilizing the ClusPro 2.0 server, and the resultant data has been systematically presented in Table 9, where it is observed that the binding energy values exhibit a notable range extending from approximately −397 to −872, thereby indicating a significant affinity between the interacting molecules. The ClusPro 2.0 server, utilized for molecular docking, performed a complex examination of the multi-epitope vaccine and TLR4 (B chain). Consequently, 30 docked complexes with distinct cluster members were generated as a result of this extensive analysis, and the complex with the lowest energy was chosen for further investigation. Among all the docked complexes, Cluster No. 1 demonstrated the most negative score (strongest interaction), with an energy score of −1161.7 kcal/mol. Additionally, the BIOVIA Discovery studio software, a widely employed tool for visualizing intricate 3D interactions, was employed to further clarify the interaction between TLR4 (B chain) and the developed vaccine as depicted in Fig. 9A. The Ligpolot software, an efficient tool for the 2D analysis of residue interactions, was employed as demonstrated in Fig. 9B. From the analysis facilitated by Ligplot, it was discerned that the amino acid residue Arg447 located within chain B established a notable interaction through bonding with the residues Asp309 and Thr311 present within the vaccine, exhibiting bond lengths measuring 2.87 Å and 2.67 Å respectively, while concurrently, Asn417 situated within chain B formed connections with Ser317 and Asn307, demonstrating bond lengths of 2.86 Å for both interactions, and additionally, Ser445 from chain B likewise formed a bonding interaction with Asn307, characterized by a bond length of 3.03 Å, alongside Glu422 of chain B which exhibited dual bonding interactions with Arg310, characterized by bond lengths of 2.70 Å and 2.74 Å, thereby culminating in the generation of a detailed and exhaustive compilation of the various residues implicated in hydrogen bonding interactions, inclusive of their corresponding bond lengths, which has been meticulously compiled to furnish a more thorough and nuanced analysis of the intricate interactions that transpire between the vaccine and the TLR4 receptor, specifically focusing on the contributions made by chain B (Table 10).

Table 9.

Molecular docking between T-cell epitopes and MHC molecules

	Members	Representative	Weighted score
MHC-I epitopes
SSEKGKLSY	639	Center	-483.2
		Lowest Energy	-574.5
RIYYPVPAY	677	Center	-700.2
		Lowest Energy	-819.6
EPNWNPLAL	955	Center	-678.9
		Lowest Energy	-741.9
KQQERAVNL	425	Center	-449.2
		Lowest Energy	-555.5
GLQDAYAKK	234	Center	-445.5
		Lowest Energy	-503.9
TLKDRNGGY	232	Center	-469.8
		Lowest Energy	-561
DSGLDLALY	999	Center	-636.8
		Lowest Energy	-695.4
MHC-II Epitopes
KAEYEKAAAENKTKSDQ	73	Center	-398.3
		Lowest Energy	-477.1
TTGESNFDRTTGAGISP	115	Center	-640
		Lowest Energy	-825.5
GESNFDRTTGAGISP	296	Center	-594.9
		Lowest Energy	-676.1
DADLTPDTQLSVGYD	127	Center	-519.9
		Lowest Energy	-659.4
SGVQYRVIEAGKGAK	188	Center	-669
		Lowest Energy	-803.1
GVQYRVIEAGKGAKPTQ	444	Center	-633
		Lowest Energy	-872.1
VIDADLTPDTQLSVG	368	Center	-601.2
		Lowest Energy	-765.7

Fig. 9.

(A) The visual representation of the docked complex between the multi-epitope vaccine and TLR4 was achieved through the utilization of the Discovery Studio software. The schematic depicts the vaccine construct (red colored) while the cartoon form illustrates TLR4 (appearing in a dark ash shade). (B) The LigPlot software was employed to visualize the interaction between the vaccine and chain B of TLR4. The presence of green dashed lines denotes the existence of hydrogen bonds

Table 10.

List of residues in the docked complex that are involved in hydrogen bonding between the vaccine and TLR4 (chain B)

TLR4 (chain B)	Vaccine	Bond length (A)
Arg447	Asp309	2.87
	Thr311	2.67
		2.80
Asn417	Ser317	2.86
	Asn307	2.86
Ser445	Asn307	3.03
Glu422	Arg310	2.70
		2.74

Molecular dynamic simulation

Molecular dynamics simulations (MDS) have been extensively employed in determining the conformational strength of atoms and molecules by modeling the system at an atomistic scale. With the objective of investigating the stability of a ligand in a targeted protein macromolecule, the MD simulation has been recognized as a remarkable and distinctive approach. This study intended to evaluate the capability of the vaccine to bind to the protein and to the active site cavity of the protein (Table 11). The findings of the MD simulation have been explicated based on the root means square deviation (RMSD) and root means square fluctuation (RMSF).

Table 11.

Discontinuous B cell epitope of the vaccine

Start Position	End Position	Peptide	Number of Residues	Type
464	523	KKQKREQGRAQAAKAEYEKAAAENKTKSDQFIAANKAKAGVQSLPSKKIEAGKGAKPTKK	60	Linear
416	448	ARSGKKRAEGYSVRRTSAGTRFDLAPREIPQKK	33	Linear
286	372	RVIEAGKGAKPTQGPGPGGESNFDRTTGAGISPGPGPGDADLTPDTQLSVGYDGPGPGVIDADLTPDTQLSVGKKDTDGQMDRYNQR	87	Linear
205	218	LKDRNGGYGPGPGK	14	Linear
269	281	IEAGKGAKGPGPG	13	Linear
112	120	VWNNKTPHA	9	Linear
31	45	TDLCAEYHNTQIYTL	15	Linear
68	72	KNGAI	5	Linear
154	157	AYGL	4	Linear
78	84	PGSQHID	7	Linear
22	29	YAHGTPQN	8	Linear

The RMSD of a vaccine-receptor complex system is a valuable tool that permits the determination of the average distance caused by a selected atom’s displacement over a specified period. Generally, the square root of the mean of squared errors is utilized to ascertain the extent of dissimilarity between two values, specifically the observed value and estimated value. The mean or average value varies from one frame to another, with a permissible range of 1–5 Å or 0.1–0.5 nm. However, a value outside this acceptable range indicates a significant conformational shift in the protein. Consequently, the RMSD of the vaccine candidate (orange) complex structure has been compared with the human Toll-like receptor 4 protein (PDB ID: 4G8A) to observe the changes in order, as depicted in Fig. 10A. The RMSD was found to exhibit minor fluctuations, which were perfectly acceptable in the context of vaccine development.

Fig. 10.

A simulation utilizing molecular dynamics was conducted on the complex formed by the vaccine and TLR4. Herein, different MD simulation plots show (A) root mean square deviation, and (B) root mean square fluctuation

As a result, the RMSF values of the experimental vaccine-receptor complex were computed in order to analyze the modification in protein structural flexibility, as exemplified in Fig. 10B. The RMSF values provide an indication of the deviation of the position of each atom in the protein from its average position over a specified time period. This analysis is crucial in comprehending the flexibility of the protein in response to the binding of the vaccine candidate to the receptor.

Immune response simulation

The experimental model of the immune response displayed similarities to the authentic immunological phenomena that are elicited by specific pathogens, as demonstrated in the accompanying Fig. 11. It is worth noting that the secondary and tertiary immune responses were discovered to be more robust compared to the primary immune response (Fig. 9A). The secondary and tertiary responses were characterized by heightened levels of various types of antibodies (such as IgG1 + IgG2, IgM, and IgG + IgM), which were accompanied by a decrease in the antigen burden, thus signifying the development of memory cells and the subsequent enhancement of antigen clearance upon subsequent exposures (Fig. 9A). Additionally, there was an extended lifespan observed in B-cells, cytotoxic T-cells, and helper T-cells, which indicated the switching between immune cells and the formation of IgM memory (Fig. 11B–D). The upregulation of IFN, IL-4, and IL-10 was also evident (Fig. 11E). Interestingly, the percentage (%) and quantity (cells/mm³) of Th0 type immune reaction were discovered to be lower compared to those of the Th1 type reaction (Fig. 11F). During the presentation, it was demonstrated that the movement of macrophages significantly increased, while the movement of dendritic cells was found to be predictable (Fig. 11G, H).

Fig. 11.

The vaccine-induced immune response. The illustration depicts (A) primary, secondary, and tertiary immune responses, (B) B-cell population, (C) cytotoxic T-cell population, (D) helper T-cell population, (E) induction of cytokines and interleukins, (F) Th1 mediated immune response, (G) macrophage population per state and (H) dendritic cell population per state

Codon adaptation and in silico cloning

In the pursuit of enhancing the translational efficiency of the vaccine construct, we meticulously optimized the codons that are incorporated within it, specifically tailoring them to the characteristics of the E. coli K12 strain, utilizing the sophisticated analytical capabilities provided by the JCat server. To facilitate the insertion of this adapted sequence into the pET28a (+) vector, we selected the restriction enzyme sites of EcoRI and BamHI, which were designated as the initial and terminal cut points, respectively, ensuring a precise and efficient cloning process. Consequently, the optimized vaccine construct was successfully cloned into the pET28a (+) cloning vector, a process that was expertly managed using the advanced features of SnapGene software, as illustrated in Fig. 12, depicting the entire procedure. The resultant cloning vector, after careful consideration and execution of the aforementioned steps, culminated in a final size of 6945 nucleotide base pairs (bp), a measurement that is indicative of the complexity and scale of the genetic material involved in this significant biotechnological endeavor.

Fig. 12.

In silico cloning of the intended vaccine into the pET-28a (+) vector. The black color shows the vector DNA while the red color shows the adapted DNA sequence of the proposed vaccine

Discussion

Stenotrophomonas maltophilia is a multidrug-resistant bacterium that poses significant challenges when treating infections caused by this opportunistic pathogen, especially among people with fragile immune systems such as cystic fibrosis or cancer patients. This resistance often presents hurdles due to antibiotics, notably beta-lactams, aminoglycosides, and chloramphenicol, being intrinsically resistant to S. maltophilia (Waters 2012). Additionally, this bacterium is susceptible to developing resistance to other antibiotics, including tetracyclines, fluoroquinolones, and trimethoprim-sulfamethoxazole (Kullar et al. 2022). Its propensity to colonize the respiratory tracts of people with cystic fibrosis is noteworthy, leading to recurring lung infections and exacerbating their medical conditions (García et al. 2023; Thornton & Parkins 2023). This frequently results in extended hospitalizations, rising healthcare expenses, and a higher likelihood of treatment failure. Furthermore, nosocomial epidemics can occur in medical facilities, emphasizing the significance of stringent infection control procedures. Healthcare professionals are investigating alternative treatment modalities, such as combination therapy and newer antimicrobial agents, to address these issues. They are also conducting research and surveillance to better understand the epidemiology and resistance patterns of S. maltophilia and to create efficient avoidance and management strategies. Vaccines are a crucial factor in saving millions of lives and controlling the spread of deadly infectious illnesses (Shovon et al. 2023). However, conventional vaccine development techniques are expensive, time-consuming, and labor-intensive, with a high probability of successive trials failing. Reverse vaccinology, which utilizes bioinformatics tools and computational methods for antigen identification, can significantly improve the immunogenicity and efficacy of vaccines by discovering new potential antigenic proteins (Jalal et al. 2022). Recently, multi-epitope vaccines against various pathogens have been developed using reverse vaccinology methods, including Candida auris, human cytomegalovirus, SARS-CoV-2, Mycobacterium tuberculosis, Helicobacter pylori, and Leishmania infantum (Alizadeh et al. 2022; Jayavel et al. 2024). In our study, pan-genomic analysis was first used to study the whole proteome of 81 strains retrieved from NCBI. The pan-genomic analysis revealed 1945 core proteins among all strains and those core proteins were targeted for further analysis. All the core proteins were checked to select virulent proteins, finding 191 proteins to be virulent. The non-virulent proteins were filtered out while the virulent proteins were further analyzed to check for the proteins that reside in the outer or cytoplasmic membrane of the bacterium (Tahir ul Qamar et al. 2021). 52 among them were found to be in the cytoplasmic membrane and 5 in the outer membrane while the rest of the proteins were filtered out. Among the surface proteins, 22 met the transmembrane helices criteria while the remaining 35 were filtered out. The 22 remaining proteins were checked for physicochemical properties to assess all important criteria; 14 of them were found to be stable proteins. Those remaining 14 proteins were then checked for antigenicity, allergenicity, and toxicity, and the top three antigenic proteins were picked for the prediction of CTL, HTL, and B cell. The 0.5 value was considered as a cut-off value to retrieve CTL and HTL epitopes provided by the IEDB server, and all the predicted B cell epitopes were retrieved to assess their antigenicity, allergenicity, toxicity, and physicochemical properties (Chand & Singh 2021; Gharbavi et al. 2021; Hossain et al. 2021). Based on these parameters, the top seven antigenic epitopes from CTL, HTL, and B cell were used for the vaccine const the epitopes were combined using AAY, GPGPG, and KK linkers, and the resulting peptide was connected to the adjuvant cholera toxin B subunit (Uniprot Id—E9RIX3) by another “EAAAK” linker (Irfan et al. 2022). The developed vaccine was then checked for physicochemical properties and the Expasy server determined the vaccine’s instability index is 26.91 which classifies the vaccine as stable. In addition to that, the vaccine’s grand average hydropathicity was predicted to be -0.803 which classifies the vaccine as hydrophilic. The secondary structure of the vaccine was predicted from two different servers. The PsiPred server predicted that the vaccine is composed of 40% alpha helix, 10.32% beta-strand, and 49.68% random coil and the Sopma server predicted that the vaccine consists of 33.84% alpha helix, 14.34% beta-strand, and 51.82% random coil. Finally, the tertiary structure of the vaccine was predicted from the Robetta server, and the 3D structure of the vaccine was visualized using Biovia Discovery Studio 2021 software. The developed vaccine was then submitted to the GalaxyRefine tool for refinement (Qadeer et al. 2021). The server provided 5 revised models, resulting in model 3 being initially picked up as the best model based on different parameters. In the validation test of the 3D structure, we discovered an adequate Z score (-8.77) and the best characteristics of the most favored, accepted, and forbidden regions for the Ramachandran plot (Islam et al. 2022; Samad et al. 2020). The vaccine construct had a CAI value of 1 and a GC content of 50.92415551306565%, according to the JCat server. CAI value greater than 0.89080 and a GC content of between 30 and 70% is considered the optimal range for target organism expression. Six disulfide bonds were added to the improved model of the vaccine’s structure using the process of disulfide engineering, which aimed to increase the stability of the structure (Al Zamane et al. 2021). The possibility of the vaccine’s infection-inhibitory activity was confirmed by molecular docking between the peptide vaccine and the TLR4 receptor’s favorable receptor for virus glycoproteins, which had the lowest energy score of -1161.7 kcal/mol. This result also suggested a possible tight interaction between the modeled vaccine’s ligand and the receptor surface (Samad et al. 2020). 50 ns dynamic simulations were conducted to evaluate the vaccine candidate’s dynamics, and the outcomes have been assessed based on the RMSD and RMSF scores. A particular molecular system’s many atomic conformations can be compared using the RMSD value. In this study, the significant flexibility and departure of vaccine candidates from the receptor structure were assessed using the RMSD value, whereas the displacement of an individual vaccine candidate’s atom with respect to the receptor structure was assessed using the RMSF of the complex structure. Finally, we conducted an immune simulation to figure out the most suitable behavior and concentration of cell parameters for successful infection elimination and the best immunological response (Samad et al. 2020). The vaccine doses generated an immunological reaction that resulted in memory B-cells (with a half-life of many months) and T-cells. As a consequence of increased aide T-cell initiation, there was sustained synthesis of IFN-gamma and IL-2 after vaccination. With this, the vaccination effectively replicated a humoral immune response to increased immunoglobulin production. Overall, the predictive methodology for multi-epitope vaccine design and its downstream investigation using molecular docking and interaction analysis provides a solid foundation for vaccine candidate development. However, more in vitro, and in vivo research is needed to determine whether the epitopes and multi-epitope vaccine proposed in this study could offer protective immunity against S. maltophilia infections.

Conclusion

S. maltophilia is an emerging pathogen that exhibits resistance to multiple drugs, thereby posing a significant threat to human health. The development of a vaccine to combat infections caused by this pathogen would be a noteworthy accomplishment. However, the intricate biology of the pathogen and the formidable challenges associated with conventional vaccine development make the task of creating a vaccine against it quite difficult. In this study, we have designed an in-silico vaccine using antigenic epitopes derived from the pathogen’s core antigenic proteins. By selecting epitopes from the core proteome, we have ensured that the vaccine provides broad-spectrum protection against all the strains of this particular bacteria. Furthermore, our vaccine has been shown to induce both humoral and cellular immunity, making it highly effective. Importantly, the epitopes included in the vaccine development are antigenic and non-allergenic, and non-toxic. The vaccine itself possesses all the desirable qualities of a practical vaccine candidate and exhibits a strong affinity and stability in binding to the TLR4 innate immune receptor, which plays a crucial role in the host immune system’s recognition and response. The use of immunoinformatic and biophysical approaches in this study has proven to be invaluable in guiding experimental studies and saving time and resources. However, due to the limitations of the tools and servers used, we recommend conducting in vitro immunological assays to evaluate the vaccine’s biological efficacy.

Abbreviations

CTL

Cytotoxic T lymphocyte

HTL

Helper T lymphocyte

TLR-4

Toll-Like Receptor 4

MHC

Major histocompatibility complex

MEV

Multi-epitope Vaccine

GDT-HA

Global Distance Test-High Accuracy

RMSF

Root mean square function

RMSD

Root mean square deviation

IFN

Interferon

IL

Interleukin

CTB

Cholera toxin B subunit

MD

Molecular dynamics

Author contributions

HJS: Writing-Original draft preparation, resources, data curation, visualization, investigation. MI: Investigation, resources. PB: Investigation, data curation. MIT: Investigation, visualization. MNHZ: Conceptualization, validation, supervision. MJH: Resources, data curation. DRP: Visualization, data curation. MNH: Conceptualization, validation, project administration. All authors have read and agreed to the published version of the manuscript.

Funding

No external or internal funding was received for this project.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.