Background
Streptococcus pneumoniae (pneumococcus), an opportunistic Gram + bacterium, resides normally on the throat or nasopharynx of healthy people and is transmitted through the carriers’ respiratory droplets [
1]. It can cause life-threatening bacterial infections such as meningitis, sepsis, and pneumonia if it invades sterile regions of the body [
2]. Moreover, pneumococcus is one of the leading causes of severe secondary infections following viral respiratory diseases such as coronavirus [
3]. According to reports from the WHO, pneumococcal (Pnc) infections claimed the lives of 1.6 million people in both developed and developing nations in 2005, which happened a lot in the children < 2 and elderly ≥ 65 years [
4]. Prevention of this considerable mortality rate on one side and progressing resistance of
S. pneumoniae to existing antimicrobials on the other emphasize the critical need for an efficient vaccine that protects high-risk groups from the majority of clinically invasive serotypes. Presently, there are two types of polysaccharide pneumococcal vaccines on the market, unconjugated (plain) polysaccharide vaccines (PPV) and protein-conjugated polysaccharide vaccines (PCV) [
5]. PPV comprising the T-cell independent polysaccharide antigens is ineffective in infants under two-year old who are at the highest risk for severe pneumococcal diseases. PCV protects children, however, it is costly, very difficult to manufacture, requires multiple injections, and needs refrigeration. Moreover, the PCV serotype coverage includes pneumococcal serotypes of developed countries [
6,
7]. Both existing vaccines are serotype-dependent and only elicit serotype-specific immunity [
7]. Because of the boundaries of available
S. pneumoniae vaccines, the development of a powerful, broad-spectrum, and cost-effective vaccine that can be effective in preventing infections caused via different pneumococcal serotypes is a significant concern of World Health Organization.
Protein-based vaccine, comprising multiple conserved pneumococcal protein antigens, can give an alternative to serotype-dependent vaccines [
8]. Understanding the different roles of virulence and colonization factors that induce pneumococcal infections is fundamental for developing the effective protein-based vaccines. Various pneumococcal proteins have been evaluated as possible vaccine candidates throughout the years, for instance, pneumococcal surface protein A and C (PspA/C) [
9,
10], the members of pneumococcal histidine triad (Pht) family A, B, D, and E (PhtA, B, D, and E) [
11,
12], and ATP-binding cassette (ABC) transporters such as PsaA, PiuA, and PiaA [
13,
14]. Each of the above mentioned virulence proteins could produce various levels of protection against immunological challenges with numerous pneumococcal serotypes in animal model [
7].
Pneumococcal surface protein C is one of the most important surface proteins of pneumococcus that is considered as a vaccine candidate [
15]. PspC has been shown to have notable roles in adhering of pneumococcus, invading, and evading complement [
10,
16]. Hence, it is also called factor H-binding inhibitor of complement (Hic) [
17], C3-binding protein A (PbcA) [
18],
S. pneumoniae-secretory IgA-binding protein (SpsA) [
19], and choline-binding protein A (CbpA) [
20]. PspC consists of three domains: an exposed N-terminal (N-ter) α-helical domain, a proline-rich domain, and C-terminal (C-ter) repeat region [
21]. This antigen shows variability among different serotypes and is classified into 11 groups [
22]. The protein PspC as a recombinant vaccine candidate has been indicated to induce antibody-mediated protection against pneumococci [
21,
23]. Investigations have revealed that anti-PspC3 (PspC from group 3) antibodies are capable to identify the majority of Pnc clinical isolates [
24]. Pht family, which is a group of surface-associated pneumococcal antigens, contains 5 conserved His triad motifs (His-xx-His-x-His) in PhtA, B, and D, as well as 6 motifs in PhtE [
12,
25]. A number of roles have been suggested for this family including involvement in the homeostasis of Zinc (Zn
2+) which is crucial for the colonization and invasion of pneumococcus [
26], and also complement deposition inhibition on the surface of pneumococcus via the recruitment of complement factor H [
27]. Among the members of this family, PhtD has been determined to be the most conserved across different strains [
28], and so to induce the broadest protection against pneumococcal diseases [
29]. Researches have indicated that the C-terminal of PhtD (PhtD-C), compared to the other areas of the protein, is more surface-exposed and could be a more immunogenic part [
28,
30]. Another protein candidate is pneumococcal surface adhesion A (PsaA) that has been assessed against Pnc infections with encouraging results in animal models and human clinical trials [
31]. The PsaA antigen, a conserved lipoprotein expressed by all pneumococcal strains, is an important virulence factor of pneumococci [
32,
33]. PsaA plays vital functions in manganese (Mn) transport [
34] and adherence to the cells of the host [
35]. In pneumococcus, the acquisition of Mn
2+ has been found to have significant roles in growth, proliferation, and virulence [
36].
Adjuvants are another component of vaccines that could significantly enhance the efficiency of vaccines and have been employed for more than a century [
37]. In the past few decades, toll-like receptor (TLR) agonists have been extensively studied as possible adjuvants for vaccines [
38]. Most of the TLR agonists currently being investigated are non-protein microbial products such as lipopeptides, oligonucleotides, and lipopolysaccharides [
39]. On the other side, a developing number of studies have revealed there are various bacterial proteins that can activate TLR signaling and immune responses. Bacterial proteins that show one or more of the following features can be considered as potential adjuvant candidates: (1) induction of pro-inflammatory cytokines, (2) up-regulation of the costimulatory molecule expression on antigen presenting cells (APCs), (3) activation of the TLR2/TLR4 signaling, (4) induction of antibody- or cell-mediated immunity [
39]. Microbial protein TLR agonists present unique characteristics that are not provided by non-protein TLR agonists, including the ability to modify the structure/function as needed for favorable immunogenicity and minimum toxicity [
39]. To ensure the co-delivery of adjuvant and antigen into the target cells, protein TLR agonists could be genetically joined to protein antigens, leading to powerful stimulation of innate and acquired immune response [
39].
S. pneumoniae endopeptidase O (PepO) is a TLR agonist that is able to induce a strong natural immune response in a TLR2/TLR4-dependent way [
40,
41]. This protein which share homology with M13 peptidase family consists of two conserved domains: M13_N domain (residues 3–383) and M13 domain (residues 437–627). It has shown that the TLRs 2 and 4 bind to the Pnc endopeptidase O through the M13_N domain [
40,
41].
Recent advances in cost- and time-efficient approaches of immunoinformatics and the availability of a broad array of tools for vaccine designing have notably allowed to develop new stable and accurate epitope vaccines [
42‐
44]. Considering the feasibility and benefits of vaccines generated via immunoinformatics methods, the aim of this research was to employ these approaches to design an epitope-based vaccine against
streptococcus pneumoniae. Herein, we used several virulence proteins of pneumococcus, i.e. PspC, PhtD, and PsaA, to predict B-cell and helper T-cell epitopes from them. The M13_N domain of PepO was subjected to molecular docking with TLR2/4 to determine the potential regions involved in the PepO-TLR interaction. Furthermore, molecular docking, molecular dynamics simulation, and in silico cloning were done on the final construct. The novelty of this study is the construction of an epitope-based vaccine containing more effective antigenic epitope-rich domains of three pneumococcal proteins, PspC, PhtD, and PsaA, which are capable of production of more diverse and robust responses. Another novelty of our study lies in the use of a new truncated domain of PepO as a TLR agonist adjuvant candidate for promoting the immunogenicity of the vaccine.
Discussion
Streptococcus pneumoniae is a powerful pathogen owing to its ability to adhere to cells, invasion to host tissues, and escape the immune system of the host [
82], and can cause fatal diseases such as sepsis, pneumonia, and meningitis with a significant death rate worldwide [
83]. The currently existing pneumococcal vaccines have several limitations including serotype dependence [
7,
84] and so they are not capable of protecting subjects against all Pnc serotypes. A hopeful alternative can be the utilization of proteins and peptide fragments, some of which have been investigated as vaccine candidates in pre-clinical trials, either alone or in combination with each other. Pneumococcal proteins commonly expressed across all serotypes have been regarded as a potential candidate for developing serotype-independent Pnc vaccines [
7]. In many biological subjects, in silico techniques could be highly useful to decrease the cost and time of experimental researches [
37,
85,
86]. Epitope identification by immunoinformatics tools could be quite helpful in the different applications within the field of epitope mapping, such as designing peptide-based vaccines, identifying immunological processes, predicting epitopes used in the disease diagnosis, and so on [
87‐
90]. Peptide-based vaccines are one of the alternative and appealing approaches to develop vaccines, which only include peptide fragments with the highest antigenicity and the most ability to induce immune responses [
91,
92]. In reality, the immune responses are only formed against immunogenic epitopes, and other parts of the protein virulence factor causing unwanted reactions can be eliminated [
93]. The combining of several highly-conserved peptides can be employed to induce both native and acquired immune responses, and prevent various stages of Pnc infection, as well as present broader serotype coverage [
7].
PspC protein which is almost found in all pneumococcal strains is one of the major and surface-exposed virulence factors [
21]. A study showed that the survival time for mice vaccinated with a combination of PspC and a genetic toxoid derivative of pneumolysin (PdB) was considerably longer than that for mice treated only with PdB, but was not remarkably different from that for mice immunized only with PspC [
94]. In another study in 2014, the PspC peptides which participate in adherence and invasion were genetically fused to L460D pneumolysoid protein. The fusion construct was more immunogenic than L460D pneumolysoid alone, and anti-fusion antibodies were active against pneumococcus. Immunization of mice with the fusion construct protected them from the Pnc carriage, pneumonia, meningitis, bacteremia, and sepsis [
95]. PhtD and PhtE antigens, two members of the conserved Pht family, have been found to be surface-exposed and expressed almost in all pneumococcal strains [
96]. It has been exhibited that both antigens induce efficient protection against colonization, pneumonia and sepsis in animal models [
97]. Godfroid et al. indicated that mice vaccinated with this family were protected more effective against pneumonia compared to those immunized with other Pnc vaccine candidates such as PsaA and PspA [
96]. In the research of Verhoeven et al., the efficacy of a multivalent vaccine containing PhtD, choline-binding protein A (PcpA), and detoxified pneumolysin (PlyD1) was evaluated and it was shown that the vaccine protected the infant mouse model against serotypes 3 and 6A pneumococcus [
98]. Plumptre et al., demonstrated the truncated derivatives of PhtD as vaccine antigens in animal models were more effective than the whole PhtD protein [
12]. Pneumococcal surface adhesion A is another conserved and immunogenic antigen that is expressed by all Pnc strains [
99]. Lin et al. reported that mice vaccinated with the 23-valent capsular polysaccharide (CPS)–rPsaA conjugate showed rapid pneumococcal clearance from blood and could provide the best protection against Pnc challenge [
100]. Lu et al., prepared a fusion conjugate consisting of a fusion protein of pneumolysin nontoxic derivative (PdT) with PsaA coupled to cell wall polysaccharide (CWPS). The fusion conjugate protected mice against colonization, while mice vaccinated with the mixture of the three antigens were not protected [
101]. In the study of Lu et al., in 2015, the fusion protein PspA-PsaA was assessed in an animal model against fatal challenges with pneumococci. The results showed that levels of antibodies against both proteins increased in mice vaccinated by the fusion protein. Also following fatal challenge, immunized mice revealed decreased levels of pneumococci in the lungs and blood compared with the control group and were protected against pneumococcal challenge [
32]. Another study showed that PsaA antigen prevented pneumococcal carriage more than PspA [
102].
The principal drawback of epitope vaccines is that they could be degraded quickly by proteinase in the body, making them hard to be recognized by immune cell receptors [
103,
104]. One of the approaches recommended for enhancing the immune responses elicited by epitope vaccines is to utilize adjuvants in the vaccine constructs [
104]. In recent decades, TLR agonists have been broadly examined as candidate adjuvants for vaccine constructs [
38]. There are various TLR agonists with different sources which a group of them are bacterial proteins capable of inducing TLR signaling and native immune responses [
39]. Pneumococcal endopeptidase O is one of the bacterial TLR agonists that can be considered as a potential candidate adjuvant [
39]. Shu and colleagues identified the Pnc endopeptidase O to be a TLR2/TLR4 bi-ligand. To distinguish the binding site of TLRs in PepO, they prepared two recombinant truncated forms of it, PepO 1–430 and PepO 431–630, according to two conserved domains of PepO named M13_N domain and M13 domain, respectively. Their result revealed that TLR2 and 4 bind to the M13_N domain of this protein [
41]. In this regard, in the present study, to identify the main residues potentially involved in the interaction between the N-PepO and TLRs, we employed molecular docking using ClusPro 2.0. Following docking, the best complexes were chosen with the lowest energy docking mode -859.4 (for N-PepO-TLR2) and -967.1 (for N-PepO-TLR4). The DIMPLOT from LigPlot + v2.2.4 was used to analyze the chosen complexes (see Figs.
2 and
3). The N-PepO residues Asp
34, Arg
3, Gln
5, Asp
6, Phe
8, Tyr
9, Asn
13, Glu
15, Arg
354, and Lys
357, and likewise, the N-PepO amino acids Tyr
4, Gln
5, Ala
11, Ile
12, Asn
13, Glu
15, Thr
19, and Glu
21 were predicted to play roles in the interactions with TLR2 and TLR4, respectively (see Table
1). Ultimately, to maintain the domain structure, a truncated fragment from N-PepO (residues 1–112) was selected as a potential TLR2/4 agonist candidate in this research.
The strategies applied in this investigation were the use of derivatives of three important virulence proteins PspC, PhtD, and PsaA into the vaccine construct which could help at preventing infection by pneumococci, and the identification of epitope-rich domains that could provoke both robust B-and T-cell immune responses. An efficient epitope vaccine should be engineered to include B and T cell epitopes capable of providing effective reactions to a specific infection [
105]. Since the recognition of putative B and T cell epitopes using the wet-lab approach requires experimental screening of a large number of active epitopes against inactive epitopes, the computational approach can act as a cost-effective, rapid, reliable, and accurate alternative [
106,
107]. In this regard, the use of a consensus prediction strategy is more reliable and robust providing better results than any of the individual prediction methods [
108]. Hence in this study, the B- and helper T-cell epitopes were predicted through various reliable databases and servers including ABCpred, LBtope, IEDB, Ellipro, DiscoTope, and NetMHCIIpan servers to achieve more reliable results. Finally, three epitope-rich regions containing high-ranked and shared epitopes, i.e. residues 196–256, 100–187, and 276–363 from PhtD-C, PsaA, and PspC, respectively, were chosen (Table
2). According to IEBD database, the selected region of PspC was consistent with two significant epitopes EDRRNYPSNT and EAKEPRNEEKVKQAK experimentally reported as B cell epitopes by Vadesilho et al. The outcomes of recent experimental study could support our predictions. The physical–chemical features and 3-D structure of the epitope-based vaccines correlate with the nature of the epitopes, linkers, and adjuvants, as well as their place and order in the vaccine construct [
109]. Herein, the chosen regions of PhtD, PsaA, and PspC were joined together in an appropriate pattern using suitable linkers. In a similar manner to Tian et al., the GGSSGG flexible linkers were applied to keep the independent folding of domains and their immunological activities [
63]. In addition, the EAAAK helical linker was employed for connecting the truncated N-PepO, as an adjuvant candidate, to the N-terminal of the above-quoted sequence to boost the stability and immunogenicity of the construct [
64,
65]. None of the selected three domains (196–256 of PhtD-C, 100–187 of PsaA, and 276–363 of PspC) has been used before and also the truncated fragment of PepO (residues 1 to 112) was not introduced earlier as a TLR agonist adjuvant candidate, showing the novelty of our study. The final engineered vaccine construct (ODAC) was assessed for toxicity, antigenicity, and allergenicity. It was antigenic, non-toxic, and non-allergic, which indicated its effectiveness in evoking sturdy immune responses without no harmful reactions. According to SOLpro results, the solubility of vaccine upon overexpression in
E. coli host, which is one of the fundamental requirements of most functional and biochemical investigations, was 0.893 [
110,
111]. The theoretical pI of the ODAC vaccine was 5.29 which indicates the ODAC is basic in nature. The molecular weight of the construct was 41.93 kDa, which is ideal because the purification of protein constructs with Mws fewer than 110 kDa is faster and easier [
104,
112]. The half-life (T1/2) of the engineered vaccine was computed to be 30 h, > 20 h, and more than 10 h, in mammalian reticulocytes, yeast, and
E. coli, respectively. The half-life is the time which takes for 50% of the protein amount inside a cell to eliminate following its synthesis in the cell. ProtParam uses the "N-end rule" which links the protein's half-life to the identity of its amino-terminal residues that applies to both prokaryotic and eukaryotic organisms [
113]. The aliphatic index of the designed vaccine was computed to be 68.01, implying that it could be a thermostable vaccine [
114]. The GRAVY index of ODAC vaccine was -0.954, and the negative value of this parameter demonstrates that the construct is hydrophilic and could well interact with molecules of water [
115]. The instability index is assessed to estimate whether a vaccine is (un)stable, if this index of a protein is below (above) 40 then it’s called stable (unstable) [
72,
115]. The instability index of the ODAC construct was computed to be 36.83, and as the value is < 40, the engineered vaccine is deemed stable (see Table
3).
The initial tertiary structure of the engineered vaccine was modeled via the Robetta server. Following this step, the refining process was utilized to promote the quality of the 3D structure, bringing it closer to the native structure. The qualities of the unrefined and refined models were compared through the R-plot, ProSA Z-score, and ERRAT score. The results showed that all scores related to the structure of the refined model improved well (Fig.
5). The refined 3D structure was subjected to discontinuous B-cell epitopes identification and molecular docking process. A proper vaccine ought to include B-cell epitopes, besides to T-cell epitopes, with the purpose of ensuring the induction of humoral immunity [
116]. The conformational B-cell epitopes were predicted on the basis of the interaction of vaccine-antibody through the Ellipro server, and the epitopes with a score of > 0.5 were picked (see Table
4). The refined model had a plurality of discontinuous B-cell epitopes, demonstrating that the engineered vaccine is quite potent in evoking humoral immunity. The docking processes between the ODAC vaccine and TLRs 2 and 4 were performed by Cluspro 2.0. The best complex models were considered with the lowest energy docking mode -922.1 and -798.3 for ODAC-TLR2 and -TLR4, respectively. The DIMPLOT was utilized for analyzing the chosen models (Figs.
7 and
8). The ODAC construct residues Met
1, Thr
2, Asp
6, Tyr
9, Asp
10, Asn
13Asp
34, Asp
36, Gln
37, Glu
40, Trp
50, and Glu
58, and likewise, the ODAC vaccine amino acids Thr
2, Arg
3, Asp
10, Glu
15, Trp
16, Asp
59, Arg
70, His
368, His
368, His
369, and His
372 were predicted to perform roles in the interactions with TLR2 and TLR4, respectively (see Table
5). The outcomes of the molecular docking showed that the truncated N-PepO of designed vaccine interacted favorably with TLRs, indicating it could act as a possible adjuvant for TLR2/TLR4 activation. With the purpose of assessing the molecular stability of ODAC-TLR2/4 docked complexes, simulations of MD were conducted through GROMACS. The structural fluctuations of the ODAC-TLR2 and ODAC-TLR4 reached the range of 0.6–0.8 nm and 0.57–0.72 nm after 20 and 15 ns, respectively (Fig.
9). The outcomes revealed that the binding of the designed vaccine to the TLR receptors was stable during simulation times. The truncated region of N-PepO as the TLR agonist was capable to keep well these connections. RMS Fluctuation is the other tool for computing the dynamic stability of complexes. The findings of RMSF showed that the residues of ODAC, except the truncated N-PepO residues, were more flexible than those of TLR2, while the engineered vaccine and TLR4 amino acids showed no notable flexibility (Fig.
10). Conforming to the RMSF analyses, the truncated N-PepO had the lowest fluctuations in the areas having the most interactions with the TLR2/4 receptors. The results from the C-ImmSim server were consistent with actual immune responses previously reported in pneumococcal infections [
117]. The Induction of B and T cell memory responses is deemed as one of the criteria for being an effective candidate vaccine [
118]. It was found that populations of memory B and T cells were increased by each injection of ODAC and these levels were maintained after the third injection (Fig.
11), indicating our construct as a suitable one. Another observation of interest was that the level of IFN-γ was elevated following the first injection and remained at the peaks after repeated the injections. This shows a high concentration of T helper cells and thus the efficient production of immunoglobulin, which supports humoral responses [
119]. To enhance the expression of the designed vaccine in
E. coli K-12, codon adaptation was conducted through JCat tool. The ODAC construct possessed the CAI value of 0.99 and GC content of 46.41%. Since CAI values > 0.8 and CG content > 30 and < 70% are considered to be desirable for expression [
104,
115], our outcomes are satisfactory.
In numerous recent studies, the reliable computational procedures have been used to select efficient epitopes and design novel vaccines against various pathogens such as
Mycobacterium tuberculosis [
118],
Helicobacter Pylori [
42],
Onchocerca volvulus [
111],
Plasmodium [
120], and Crimean-Congo hemorrhagic fever virus [
119]. In this respect, the present study recommends a final peptide construct as the best epitope vaccine based on the strategies used in the research to be an attractive intervention against the
Streptococcus pneumoniae pathogen. Furthermore, antigenic epitope-rich portions from each of the proteins involved in the construct could also be employed in future studies to develop the other new subunit vaccines consisting of other virulence proteins. In addition, the new truncated domain of PepO could play as a new adjuvant candidate in other studies to increase the potency of vaccines against pneumococci or other pathogens.