We downloaded proteomes of pneumocystis pneumonia of 2 different strains Pneumocystis jirovecii (Human pneumocystis pneumonia agent) (SE8) and Pneumocystis jirovecii (strain RU7) (Human pneumocystis pneumonia agent) (RU7) from the UNIPROT server (www.uniprot.org). For both of these strains, redundancy and non-redundancy analyses were carried out [18]. The Reference genome of Pneumocystis jirovecii (strain SE8) consists of 3419 proteins which are then analyzed to shortlist antigen, allergen, and immunogenic proteins.

In addition to the role of proteins in active transport, outer membrane proteins (OMPs) also play an important role in the pathogenicity of fungus, the permeability barrier of the outer membrane (OM), efflux pumps, and the diffusion of nutrients as well as outer membrane proteins (OMPs) have a role in the preservation of the structural integrity of the membrane. Due to their antigenic and immunogenic properties, fungal OMPs are very versatile and have been selected for use in the production of mRNA vaccines. To get a complete subtractive analysis, the first thing we did was predict the cellular localization of 3419 proteins by using PSORT-II. This allowed us to have a more in-depth look at the data. This software is able to classify the localization of proteins as being extracellular, outer membrane, cytoplasmic, periplasmic, or inner membrane.

The candidates for the vaccination that are homologous to proteins found in other organisms are being investigated as a potential target. It is possible that the selection of these protein targets will reduce the likelihood of cross-reactivity in the host cell. We carried out a BLASTp (https://blast.ncbi.nlm.nih.gov/Blast.cgi?PAGE=Proteins) study of 982 proteins in the National Center for Biotechnology Information (NCBI) database against the proteome of the host Homo sapiens in order to remove homologous sequences between the human host and the pathogen [19]. If the E-value (expectation value) was 104 or less, the hits were deemed to be homologous proteins, and they were discarded from the study.

We were able to validate the proteins localization within the cell by utilizing two different servers at the same time (PSORT-II). Through the utilization of Signal P 4.037, we were able to generate hypotheses on the secretary function of 918 non-human comparable proteins [20]. Those forecasts turned out to be spot on. The presence of N-terminal signal sequences allowed for the prediction of proteins that were going to be exported from the inner membrane to the periplasm in accordance with the sec pathway, which is also known as the secretory pathway. This prediction was made possible because the sec pathway is also known as the secretory pathway. Because the sec route is also known as the secretory pathway, this prediction was made possible by the fact that it has both names. Following the narrowing down of the available alternatives, the proteins that generated a strong signal were selected for further analysis.

The non-classical pathway is responsible for transporting proteins from the periplasm to the extracellular region, whereas the classical system moves proteins from the cytosol to the periplasm. Proteins that are released through non-classical pathways do not have signal peptides, although they can still be identified in the extracellular region of fungus. With the assistance of the programme Secretome P 2.0 (https://services.healthtech.dtu.dk/services/SecretomeP-2.0/), 50 proteins were omitted from the analysis due to the non-classical secreted extracellular proteins they contained [21]. It was determined that the proteins with a SecP score of 0.5 or higher were non-classically secreted. Using the Ensembl database (https://www.ensembl.org/index.html), BLAST was performed on the finalized 10 transmembrane proteins chosen, non-redundant proteomes.

The experimentally identified essential proteins are included in the database of essential genes (DEG), which was used to establish which pathogens require which essential proteins [22]. In DEG, a BLASTp search was carried out using a cut-off of 1*10${}^4$ E-value, a bit score of 100, and the BLOSUM matrix [23]. This resulted in the shortlisting of seven Pneumocystis jirovecii OMPs.

Virulence factors that are connected with transmembrane proteins play a significant part in the interaction between the host and the pathogen and affect the host’s defence system. The VFDB database (http://www.mgc.ac.cn/VFs/search_VFs.htm) is utilised in order to identify OMPs that are linked with virulence [24]. The 7 Pneumocystis jirovecii proteins that were identified were BLASTp against the VFDB protein core dataset (R1) with an E-value of 1 × 10${}^4$ and a default cut-of bit score of $>$100. It was hypothesised that each of the 7 proteins would be infectious.

The Kyoto encyclopaedia of genes and genome’s (KEGG) pathway database provides an explanation of the function of proteins within several metabolic pathways [25]. In order to determine which metabolic pathways are specific to the human host and which are shared with pneumocystis pneumonia, we have manually compared all of the enumerated metabolic pathways. The 7 proteins were separated into two groups, one for their role in unique, pathogen-specific pathways, and the other for their role in common, shared pathways between the pathogen and the host. We have also determined the fundamental role of these 7 proteins using either analysis relying on pathways or analysis independent of pathways.

Using a comparative subtractive proteomics technique, we were able to exclude 7 of P. jirovecii most pathogenic proteins from its 3419 total protein sequences, as previously studied. These seven OMPs were taken into consideration as possible targets for the construction of an mRNA vaccine. To begin, an antigenicity analysis was performed with the use of the VaxiJen server [26]. Antigenic proteins were determined to be those that had a score value greater than 0.5, and these proteins were taken into consideration for further investigation.

The diverse protein expressions in the fungal cell will be affected by the mRNA vaccination that is directed against the proteins of the outer- and trans membrane. These vaccinations may have an effect on the intra-species interaction partners, which would result in a reduction in the pathogenesis of a Pneumocystis jirovecii infection. The entire outer membrane protein-protein intra-species interaction analysis was carried out with the help of the protein interaction database STRING (https://string-db.org/) in order to locate the proteins network for identified OMPs [27].

When pathogenic proteins enter the proteasomal complex of the host cell, they naturally form epitopes. T-cells respond to MHC class I epitopes to activate the immune response against the pathogen. In order to find MHC -Peptide complexes that can generate an immune response in the host cell, we used major histocompatibility (MHC) class I (MHC I) immunogenicity prediction tool, the IEDB website (http://tools.immuneepitope.org/immunogenicity/). The epitopes that were finally selected for further investigation were examined using this server with the default settings. Prediction tools’ other parameters were left at their default values. For further investigation, we focused on the epitopes that received a high score for their immunogenicity. Every single one of the outer membrane and transmembrane proteins was evaluated using IEDB prediction algorithms separately. The best-predicted protein epitopes were chosen for further analysis.

Epitopes that attach to MHC class II molecules can be identified using the MHC Class II prediction tool http://tools.iedb.org/mhcii/[28]. The Helper T-Lymphocytes epitopes were predicted under the MHC-II IEDB server NetMHCpan BA 4.1 module by selecting full set human allele with the default values. Stabilization matrix alignment and the average relative binding matrix were the two consensus methods used to determine which epitopes of MHC II would be the most effective.

Antigenicity and toxicity were investigated in MHC I and II epitopes. Both analyses made use of the VaxiJen 2.0 and ToxinPred tools (https://webs.iiitd.edu.in/raghava/toxinpred/) [29]. The physicochemical features of VaxiJen servers ($>$0.5) can be used to predict antigenic behaviour of epitopes. To find out if the immune response in the host cell targeted only fungal cells or also the host cells, researchers employed the ToxinPred tool.

The recognition of epitopes by B cells is necessary to start the process of humoral immunity. ABCPred server, were used to make predictions on B-cell epitopes [30]. The shortlisted proteins were delivered with threshold of 0.51. The epitopes that were chosen were further take into consideration in the vaccine construction process.

In order to build an mRNA vaccine that is targeted to OMPs and transmembrane proteins, we connected all of the selected epitopes (HTL, CTL, and B epitopes) using amino acid linkers GGGS, GPGPG, HEYGAEALERAG, and CYY Adjuvant ‘EAAK’ was added at the N-terminus of the vaccine construct which improved their immunogenicity. In order for the mRNA vaccine to be effective, it needs to have a Kozak sequence. In addition, two structures known as tPA (P00750) and MITD (Q8WV92) were inserted into the construct’s 5’ region and the mRNA vaccine’s 3’ locus, respectively. All of the proteins that make up adjuvants act as agonists to various TLR complexes, which polarises CTL responses and has a powerful immunomodulating effect [31].

The vaccine construct was evaluated on the basis of their antigenicity, allergenicity, and solubility prediction methods in order to check that vaccine design would be most appropriate for use. The vaccine was run through the AlgPred server so that we could determine how allergic the vaccine was. Both ANTIGENpro and the VaxiJen 2.0 server were used to make the antigenicity prediction for vaccine constructions. Using the SOLpro service, we were able to forecast the solubility of vaccine construction probability ($\ge$0.5).

The Expasy ProtParam server was utilised in order to define the physiochemical qualities of the vaccine constructions [32]. This site is utilised by a large number of people in order to calculate the number of amino acids, PI values, molecular weight, aliphatic index, instability index, and hydropathicity GRAVY values. The protein’s instability index serves as a predictor of the protein’s stability. The thermostability of vaccines can be understood through their aliphatic index values. The hydrophilic or hydrophobic properties of a protein can be deduced from its GRAVY [33].

For the prediction of the secondary structure of the vaccine, PSIPRED server is used [34]. This server predicts the alpha, beta helix and the coil structure of the protein with the accuracy of 81.6%.

For Molecular Docking, the vaccine three-dimensional structure using Robetta. The Procheck, ERRAT, and Ramachandran Plot, at the saves website of University of California Los Angeles (UCLA, https://saves.mbi.ucla.edu/) were utilized to verify the proteins’ estimated structural analyses. The alleles were downloaded from Protein Data Bank PDB RCSB. Cluspro was used for the docking of vaccine with TLR-3 and TLR-4 complex [35]. For the Molecular Dynamics Simulation, I-Mod was used [36].

The proteomes of several strains of Pneumocystis jirovecii were retrieved through the UniPort service, which has a list of two strains. Manual redundancy and non-redundancy assessments were performed on the proteomes of both Pneumocystis jirovecii strains. In this work, the proteome of Pneumocystis jirovecii (strain SE8) is used as a reference proteome.

The UniPort server was used to download the protein sequences from the whole proteome of Pneumocystis jirovecii (strain SE8). There are a total of 3419 proteins in this strain of the Pneumocystis jirovecii. The UniPort Id for this Pneumocystis jirovecii strain SE8 is UP000010422. Table 1 shows the proteins that are extracted from the whole proteome of this strain.

From the reference proteome, the potential proteins for vaccine design can be selected by filtering the 3419 proteins on the base of their subcellular location. By using PSORT-II, we filtered 884 proteins out of 3419 proteins of the Pneumocystis jirovecii (strain SE8). These proteins are present in the sub-cellular location of the fungal cell. To avoid vaccination cross-reactivity with human host cells, the effective vaccine should be assembled against non-human homologous proteins. These non-human homologous fungal proteins were used to construct a vaccine design that only interacts with Pneumocystis jirovecii proteins.

The selection of potential proteins depends on their antigenicity and allergenicity of the proteins from Pneumocystis jirovecii fungus were considered as displayed in Table 1. As vaccine targets, essential proteins will have a significant impact on the fungus, hence we examined targeting proteins using the DEG server as displayed in Table 1. One protein, L0PAX6, an uncharacterized protein, is thought to be crucial for the survival of the Pneumocystis jirovecii strain.

STRING predictions of these selected proteins of Pneumocystis jiroveciishow various intra-species interactions between the proteins were revealed. We may be able to grasp the protein networking of proteins with the aid of all interaction studies. An mRNA vaccine against these proteins may affect all interacting network proteins, increasing the vaccine’s efficacy, all protein networking is presented in Fig. 1, except the second uncharacterized protein with UniPort Id L0PAX6.

Fig. 1.

STRING interaction of targeted proteins.

An mRNA vaccine against the virulence factor-associated proteins will boost the protein’s potency and effectiveness. The DFVF (Database of Virulence Factors in Fungal Pathogens) analysis in Table 1 revealed that all selected proteins are associated with Pneumocystis jirovecii virulence. As a result, these proteins are regarded as a crucial target for inhibiting Pneumocystis jirovecii pathogenesis.

To shortlist the identified proteins involved in various pathways, a manual comparison was performed. These findings uncover seven distinct pathogen pathways prevalent in both pathogen and host Pneumocystis jirovecii. As demonstrated in Table 1, four of the seven identified proteins were found to be metabolic pathway dependent.

The IEDB server was used for the binding prediction of MHC-I of 7 selected proteins. For analysis, epitopes with a greater affinity (IC 50 nM) and a good percentile rank (0.2) were chosen along with ANN predicted method. To increase accuracy, we eliminated epitopes with Inhibitory Concentration 50 (IC50) values smaller than 500. 80 MHC class I epitopes from various sources were tested for class I immunogenicity, allergenicity, and toxicity. A high immunogenicity score indicated the ability to excite T cells as well as promote cell-mediated immunity. In the immunogenicity prediction, the antigenicity score of selected 14 epitopes ranges from 3.0104 to 0.5226 as shown in Table 2.

Pneumocystis jirovecii seven selected proteins were evaluated to predict the potential MHC-II epitopes. The IEDB server was used to test MHC-II binding predictions for all seven identified proteins. 14 epitopes were chosen for vaccine design based on their percentile rank and IC50 value (50 nM). As stated in Table 3, only those epitopes that were studied as antigenic, non-toxic, and non-allergenic were chosen.

Using the ABCpred server, we identified the top two epitopes from each targeted protein sequence. Furthermore, we solely considered antigenic, non-toxic, and non-allergenic epitopes for vaccine development. Table 4 indicates that we have chosen 14 B-cell epitopes to include in the vaccine design, beginning with the targeted proteins.

The proposed construct of mRNA vaccine illustrated in Fig. 2, displayed the vaccine from N-terminal to C-terminal is: 5${}^{\prime }$ m7GCap-5${}^{\prime }$ UTR-Kozak sequence-tPA (Signal peptide)-EAAAK Linker-RpfE (Adjuvant)-GPGPG linker-MHC-II Epitopes-KK-B-Cell Epitopes-CYY-MHC-I Epitopes -MITD Sequence-Stop codon-3${}^{\prime }$ UTR-Poly (A) tail.

Fig. 2.

The layout of mRNA vaccine construct.

The VaxiJen server was used to predict the antigenicity of Vaccine design. The antigenicity of Vaccine constructs 1.0428, and non-allergen and non-toxic. The vaccine is expected to be antigenic, non-allergenic, non-toxic, and soluble. Using SOLpro, a vaccine design demonstrated excellent solubility ($>$0.9651) during heterologous expression in E. coli. Furthermore, as shown in Table 5, the ExPasy ProtParam tool was utilized to assess the physiochemical profile of the construct. Based on the results the construct of an mRNA-based vaccine is a potential candidate of the vaccine.

PSIPRED was used to predict the secondary structures of the final vaccine constructions. Fig. 3 depicts the secondary structure of the vaccine design. The vaccine construct has an alpha-helix, a beta-sheet, and a beta-turn structure. The Rosetta server predicted the vaccination model’s tertiary structure, which was verified by the Ramachandran plot.

Fig. 3.

Structure prediction and validation of vaccine construct. (A) PSIPRED server results of secondary structure of vaccine. (B) The Robetta server used to predict the Tertiary structure of vaccine. (C) The PROCHECK server used to analyze the Ramachandran Plot. (D) Z-score analysed by the Pro-SA webserver.

A vaccine can interact with various HLA-alleles found in human populations. The final vaccine design was docked with 5 distinct HLA-allele proteins, as indicated in Table 6. We investigated the vaccine design created against Pneumocystis jirovecii and established its applicability.

The projected protein model folding produces a conformational B-cell epitope. The discontinuous B-cell Epitopes were analyzed using the ElliPro server. The prediction score for all 392 residues ranged from 0.557 to 0.824. Fig. 4 shows 2D and 3D representations of conformational B-cell epitopes.

Fig. 4.

The ElliPro server of IEBD database for the prediction of seven conformational B-cell epitopes. 3D models of B-cells epitopes where yellow spheres present the conformational B-cell epitopes. (A) 10 residues with a score of 0.675. (B) 42 residues with a score of 0.802. (C) 35 residues with a score of 0.819. (D) 20 residues with a score of 0.568. (E) 13 residues with a score of 0.634. (F) 51 residues with a score of 0.615. (G) 11 residues with a score of 0.579.

We used ClusPro tool, for the determination of molecular docking and analyzing the interaction among the vaccine construct and receptors of TLR-3 and TLR-4. For the TLR-3-vaccine complex, the binding affinity was –1301.7 kcal/mol${}^{-1}$ and for TLR-4 the binding affinity was –1374.7 kcal/mol${}^{-1}$ (Fig. 5A and Fig. 6A).

Fig. 5.

Molecular dynamics simulation, Normal Mode Analysis, and receptor-ligand interactions. (A) Vaccine-TLR4 docked complex using the Cluspro server. (B) Deformability graph. (C) Eigenvalue of vaccine-TLR4 complex. (D) B-factor graph. (E) Covariance matrix. (F) Elastic network model using the iMODS server.

Fig. 6.

Molecular dynamics simulation of TLR-4 vaccine complex, Normal Mode Analysis, and receptor-ligand interactions. (A) Vaccine-TLR4 docked complex using the Cluspro server. (B) Deformability graph. (C) Eigenvalue of vaccine-TLR4 complex. (D) B-factor graph. (E) Covariance matrix. (F) Elastic network model using the iMODS server.

The molecular dynamic simulation of docked vaccine-TLR-3 and TLR-4 was performed using the iMODS server. The deformability graph, as shown in Fig. 5B and Fig. 6B, examines the deformable loci. The graphic depicts the B-factor graph, which was used to evaluate protein flexibility using the Normal Mode Analysis (NMA) technique. Fig. 5C and Fig. 6C depicts the complex eigenvalues, which signify a low deformation index, increased stiffness, and increased stability. The Fig. 5D and Fig. 6D depicts the link between the amino acid duplets that disperse in the dynamical area. The white color in Fig. 5E and Fig. 6E represents anti-correlated amino acids, the red color represents correlated amino acids, the blue color indicates atom pairs joined by springs, and the grey color in Fig. 5F and Fig. 6F represents the stiffer area.