Towards the design of multiepitope-based peptide vaccine candidate against SARS-CoV-2

Coronavirus disease 2019 is a current pandemic health threat especially for elderly patients with comorbidities. This respiratory disease is caused by a beta coronavirus known as severe acute respiratory syndrome coronavirus 2. The disease can progress into acute respiratory distress syndrome that can be fatal. Currently, no specific drug or vaccine are available to combat this pandemic outbreak. Social distancing and lockdown have been enforced in many places worldwide. The spike protein of coronavirus 2 is essential for viral entry into host target cells via interaction with angiotensin converting enzyme 2. This viral protein is considered a potential target for design and development of a drug or vaccine. Previously, we have reported several potential epitopes on coronavirus 2 spike protein with high antigenicity, low allergenicity and good stability against specified proteases. In the current study, we have constructed and evaluated a peptide vaccine from these potential epitopes by using in silico approach. This construct is predicted to have a protective immunogenicity, low allergenicity and good stability with minor structural flaws in model build. The population coverage of the used T-cells epitopes is believed to be high according to the employed restricted alleles. The vaccine construct can elicit efficient and long-lasting immune response as appeared through simulation analysis. This multiepitope-based peptide vaccine may represent a potential candidate against coronavirus 2. However, further in vitro and in vivo verification are required.


Background:
In December 2019, multiple pneumonia cases of unknown etiology were reported in Wuhan, China [1]. Later on, genomic analysis of samples collected from admitted patients had revealed that a novel beta-coronavirus was the causative pathogen [2]. This RNA virus was temporarily known as 2019 novel coronavirus (2019-nCoV), but was renamed later as severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) [3]. The disease caused by this virus is known as coronavirus disease 2019 (COVID- 19), it is usually characterized by fever, dry cough, fatigue and muscles pain. Additionally, dyspnea can be observed in some  patients and it may progress into acute respiratory distress syndrome (ARDS) [4]. The transmission of SARS-CoV-2 is largely dependent on respiratory droplets generated through sneezing or coughing [5]. The median incubation period of SARS-CoV-2 is estimated to be 5.1 days with 95% confidence interval of 4.5 to 5.8 days [6]. The outbreak of SARS-CoV-2 was recognized as a global pandemic threat on March 11, 2020 [7].
SARS-CoV-2 has a genomic sequence similarity of 79% with a previously known coronavirus,

SARS-CoV [8]. Both viruses can invade host alveolar cells through interaction of viral spike
protein with angiotensin converting enzyme 2 (ACE2). However, the binding affinity of SARS-CoV-2 to ACE2 seems to be 10-20 times higher [9]. It is believed that COVID-19 pathogenesis may involve blockade of ACE2 and subsequent imbalance between angiotensin-(1-7) and angiotensin II. This suggests that ACE2 may be further involved in  pathogenesis beyond viral entry point [10].
Similar to SARS-CoV, no specific treatment is currently available to combat COVID-19.
However, management of ARDS does involve the use of oxygen-based therapy and antibiotics against possible sepsis [11]. In attempts to fight COVID-19, clinical trials are keep going to assess the effect of several antiviral agents like Remdesivir [12]. Additionally, computational modelling attempts have proposed the repurposing of several FDA approved drugs against SARS-CoV-2 [13,14].
In the same time, no vaccine was ever developed for any coronavirus. However, advanced techniques had been employed to generate potential SARS-CoV-2 vaccine candidates like mRNA based vaccine, viral vector vaccine and viral subunit vaccine [15]. Recently, immunoinformatics tools had been used to screen SARS-CoV-2 proteins for potential B-cells and Tcells epitopes [16,17]. The prediction of these epitopes had enabled virtual design of peptide based vaccine against SARS-CoV-2 [18]. 3 Previously, we have predicted several potential epitopes within SARS-CoV-2 spike protein sequence. Of interest was the linear B-cells epitope with sequence 'GFNCYFPLQSYGF', this epitope is believed to be part of spike protein receptor binding protein (RBD) that is involved in interaction with ACE2 [17,19]. In the current study, we have constructed a peptide-based vaccine design by combining seven potential epitopes predicted from our previously published findings [17]. Then, we have evaluated this vaccine design by using several immunoinformatics tools to affirm its stability, safety and efficiency. The aim of this study is to present a multiepitope-based peptide vaccine design for potential use against SARS-CoV-2.

Setting up a research plane:
A flowchart that summarizes the steps of vaccine construction and evaluation study can be seen in Figure 1.

Selection of potential epitopes for peptide vaccine build:
We have selected seven potential epitopes for the construction of vaccine candidate. The sequence of these epitopes can be seen in Table 1, they have high antigenicity, low allergenicity and good stability against selected proteases as predicted by our previously published study [17]. Table 1: Linear epitopes previously predicted on SARS-CoV-2 spike protein crystal.

No.
Epitope sequence Epitope type

Population coverage of selected T-cells epitopes:
Major histocompatibility complex (MHC) molecule are highly polymorphic and expressed at

Peptide vaccine construction:
After population coverage analysis, these seven epitopes were then merged together in a sequential manner by using AAY linker. This linker is considered as a cleavage site of mammalians proteasomes and can increase epitopes presentation by enhancing the formation of natural epitopes and reducing the formation of junctional epitopes

Secondary and tertiary structures prediction of vaccine design:
The secondary structure of vaccine construct was predicted by using PSIPRED 4.0 tool [29].
While tertiary structure was modelled by using SPARKS-X, this web-based tool can recognize protein folding through sequence alignment [30]. For these prediction tools, the sequence of vaccine design was submitted as one letter code. To further improve vaccine tertiary structure, the generated PDB file was then submitted to Galaxy refine server. This refinement server applies molecular dynamics (MD) simulation to implement repeated cycles of structural perturbation with subsequent relaxation [31]. 6 Vaccine structure validation: The peptide vaccine structure was submitted as PDB file to RAMPAGE server for MD analysis was carried out for 10 nanoseconds by using the same protocol applied in our previously published article [14].

Simulation of immunogenic response to vaccine construct:
The ability of peptide vaccine to activate various components of host immune system was predicted by using C-ImmSim server. This modelling platform combines the dynamics of immune system together with genomic information [44]. At first, the vaccine sequence was

Results and discussion:
The selected six T-cells epitopes showed high predicted population coverage according to MHC restricted alleles. According to Figure 2, the worldwide coverage of these epitopes was  The primary sequence of peptide vaccine was constructed by merging the seven potential epitopes together by using AAY linker, while the adjuvant human beta-defensin 2 chain A was linked to the N-terminus of vaccine construct by using the rigid linker EAAAK. The primary structure of peptide vaccine can be seen in Figure 3 (A) as one letter format. Prediction of vaccine secondary structure as appeared in Figure 3 (B) revealed that the percentage of βstrand, α-helix and coil are 10, 53 and 37 respectively. The tertiary structure of vaccine design was modelled by using SPARKS-X server with a Z-score of 5.26, the model was further refined by using Galaxy refine server with a root mean square deviation (RMSD) of 0.496 from initial build. The refined tertiary structure of peptide vaccine can be seen in Figure 3 (C). and epitopes are also colored by gold, light blue and red respectively. 9 A summary of predicted physicochemical properties, immunogenicity and allergenicity of vaccine construct can be observed in Table 2. According to this table, the peptide   The validation results of vaccine refined tertiary structure can be observed in Figure 4. Based on analysis of Ramachandran plot in Figure 4 (A), 93.2% of residues are in favored position of torsional angles plot while 4.8% are in allowed region. Only 2.0% of the residues in the refined model are considered outliers. Then, the Z-score of the refined model was recorded to be -3.59 by using ProSA-web tool. This quality score is located within the Z-scores range calculated for native proteins of similar size as can be seen in Figure 4 (B). The model quality was also assessed, as presented in Figure 4 (C), by plotting predicting energy as a function of sequence position. The plot for window size of 40 residues is smoother than 10 residues size.
As obvious, our refined model may have some structural flaws as the predicted energy for window size of 40 residues jumps slightly higher than zero in part of the plot. Finally, the overall quality score of ERRAT server is reported to be 82.645. This score represents the percentage of protein sequence with predicted error value that falls below rejection limit. A plot of error value versus residues number can be seen in Figure 4 (D). All the above structural validation results predict that the refined construct has a good stability with minor flaws. candidate, the 99% and 95% lines represent confidence interval for rejection limit.

11
Molecular docking analysis of vaccine construct against TLR8 had generated many potential complexes. An overview for docking result of the first ranked complex can be seen in Figure   5. This vaccine-TLR8 complex has a geometric shape complementarity score of 17544 with approximate interface area of 2745.70. A cartoon representation for this complex is well presented in Figure 5 (A). According to analysis of protein-protein interface illustrated in Figure 5 (B), The peptide vaccine was able to form 5 hydrogen bonds with TLR8 interface residues. The anticipated length of these bonds is less than 3.5 Angstrom. and TLR8 can be seen in Figure 6. Based on total potential energy calculations in Figure 6 (A), the complex looks to be stable with minimal changes in potential energy predicted throughout simulation period. However, the vaccine construct seems to exhibit a higher 12 flexibility per residue than TLR8 as noted from root mean square fluctuation (RMSF) calculations in Figure 6 (B). The residues of peptide vaccine show a higher variation in position from its simulation time averaged position (reference position). The results of both molecular docking and dynamics simulation may indicate the preferential binding of vaccine design to Toll-like receptor 8. An RMSF value of zero refers to the absence of residue in the molecule under simulation. 13 Finally, the results of simulation for host immune response against injected vaccine can be seen in Figure 7. As predicted by simulation server, we found that three injections of vaccine construct with a time frame of 4 weeks between each two doses are necessary to produce loglasting and efficient immune response. According to Figure 7