Advances in Bioinformatics

Advances in Bioinformatics / 2019 / Article

Research Article | Open Access

Volume 2019 |Article ID 6815105 |

Sabeena Mustafa, Hanan Balkhy, Musa Gabere, "Peptide-Protein Interaction Studies of Antimicrobial Peptides Targeting Middle East Respiratory Syndrome Coronavirus Spike Protein: An In Silico Approach", Advances in Bioinformatics, vol. 2019, Article ID 6815105, 16 pages, 2019.

Peptide-Protein Interaction Studies of Antimicrobial Peptides Targeting Middle East Respiratory Syndrome Coronavirus Spike Protein: An In Silico Approach

Academic Editor: Nurit Haspel
Received06 Jan 2019
Accepted08 May 2019
Published01 Jul 2019


There is no effective therapeutic or vaccine for Middle East Respiratory Syndrome and this study attempts to find therapy using peptide by establishing a basis for the peptide-protein interactions through in silico docking studies for the spike protein of MERS-CoV. The antimicrobial peptides (AMPs) were retrieved from the antimicrobial peptide database (APD3) and shortlisted based on certain important physicochemical properties. The binding mode of the shortlisted peptides was measured based on the number of clusters which forms in a protein-peptide docking using Piper. As a result, we identified a list of putative AMPs which binds to the spike protein of MERS-CoV, which may be crucial in providing the inhibitory action. It is observed that seven putative peptides have good binding score based on cluster size cutoff of 208. We conclude that seven peptides, namely, AP00225, AP00180, AP00549, AP00744, AP00729, AP00764, and AP00223, could possibly have binding with the active site of the MERS-CoV spike protein. These seven AMPs could serve as a therapeutic option for MERS and enhance its treatment outcome.

1. Introduction

Middle East Respiratory Syndrome-Coronavirus (MERS-CoV) was identified in Saudi Arabia in 2012 and it belongs to Coronaviridae family and mostly reported among the Middle Eastern people. This virus causes the respiratory illness called the Middle East Respiratory Syndrome (MERS) [1]. Phylogenetic studies show that bats are the reservoir of this virus and camel is the only host through which the virus spreads to humans [2].

According to the World Health Organization (WHO), at the end of November 2018, a total of 2274 laboratory-confirmed cases of Middle East Respiratory Syndrome (MERS), including 806 associated deaths, were reported globally, where the majority of these cases were reported from Saudi Arabia (1896 cases, including 732 related deaths). Although different classes of treatment trials are ongoing, no effective treatment or vaccine is available for this disease, which causes the necessity of the effective therapeutic treatments. In this scenario, peptides can serve as potential treatment option for MERS. It has been shown that peptides act as modulators in viral diseases. For example, Melnik et al. [3] shortlisted nine peptides based on the Wimley-White interfacial hydrophobicity scale (WWIHS), where four of these peptides (WWIHS = 3.5) had greater than 50% inhibition of human cytomegalovirus. In another study, it was shown that several peptides with WWIHS = 5.2 inhibited multiple strains of influenza with M [4]. By and large, other peptides with positive WWIHS values have been shown to inhibit various viruses such as Rift Valley Fever [5], Dengue, White Nile [6], and SARS [7] and the mechanism of action is by interfering with fusion of host cellular and viral glycoprotein membranes [8]. It is for these reasons we propose that antimicrobial peptides (AMPs) can be used as an effective therapeutic agents against MERS. Several peptides have been extensively studied and identified as anti-MERS-CoV peptides [912] and anti-MERS-CoV AMPs in the past few years [13].

In order to target MERS-CoV, the knowledge of structural and nonstructural proteins is important. In this work, we focus on the structural protein of MERS-CoV spike (S) protein. Understanding the S protein structure is useful in the drug discovery for developing anti-MERS-CoV components. The S protein consists of S1 and S2 regions and it is a type I transmembrane glycoprotein, which is located at the viral envelope surface in a trimer state. The S1 and S2 subunits play a role in viral entry, binding, and fusion [14]. The S1 subunit has a receptor-binding domain (RBD) which binds to the receptor dipeptidyl peptidase 4 (DPP4). The S2 subunit consists of heptad repeats 1 and 2 (HR1 and HR2), which forms a complex called the fusion core, and represents a key membrane fusion architecture [14]. During the process of membrane fusion, HR1 and HR2 regions form a six helix bundle core with a hydrophobic region being inserted into the host membrane and thereby fusion occurs [15].

Proteins interact with other proteins in order to perform cellular tasks and knowledge of this can facilitate the development of therapeutics. X-ray crystallography and mutagenesis are techniques used in determining protein complexes and consequently protein interfaces. However, these techniques are expensive; hence, an in silico approach in predicting protein interaction, protein-protein docking, and protein interface is needed. Protein-peptide docking methods can be divided into three categories: template-based docking; local docking; and global docking [16]. Example of protein-protein interactions includes the use of graphical models by predicting the binding site between two proteins [17]. Protein interface prediction involves determining a subset of residues on the protein surface which are involved in intermolecular interactions. Example of prediction of protein interfaces includes ComplexContact, which is a web server for determining interfacial residue-residue contact prediction of a putative protein complex. This is useful in deciphering how proteins form a complex by looking at how their residues interact [18]. The challenges faced in developing these computational methods are that no methods yield excellent results and there is no gold standard benchmark dataset that can be used to compare them [19]. In addition, there are other challenges like (i) modelling significant conformational changes of both peptide and protein molecules, (ii) selection of the highest accuracy structure out of many generated models, and (iii) integration of experimental data and computational predictions into the protein-peptide docking scheme [16].

In this study, we have considered a set of AMPs in order to identify their role as putative modulators for MERS-CoV proteins. More specifically, we aim at evaluating the inhibitory mechanism of a set of AMPs with specific physicochemical properties and by employing peptide-protein interaction in order to determine its accuracy in binding with spike fusion core of MERS-CoV. The motivation for this study is that we will use highly available and reusable data that would otherwise be costly to produce. In addition, the simplicity of the proposed method, once optimized, will make it easy to identify the most important peptides that act as therapeutic agents. Our study is purely based on the binding efficacy of antimicrobial peptides on MERS-CoV spike (S). The consequences of such effort are twofold: (i) it will eliminate or at the least minimize the cost of synthesizing countless numbers of peptides and (ii) accelerate drug discovery of MERS therapeutics.

2. Materials and Methods

We propose a two-stage computational approach to determine possible antimicrobial peptides (AMPs) that can target spike protein of MERS-CoV. The first stage involves database screening of AMPs from APD3 database ( [15] based on physicochemical properties. The second stage involves the structural bioinformatics studies to analyze the peptide-protein interaction complex of MERS-CoV S protein using the shortlisted AMPs and implement the docking studies in order to determine the interacting residues with greater affinity. The flowchart depicting the methodology employed in this study is shown in Figure 1. Detailed description of the steps involved is presented in the following subsections.

2.1. Retrieval of the Prefusion Structure of MERS-CoV spike Protein

We considered the S protein of MERS-CoV, which is illustrated in Figure 2; in particular, we selected the prefusion form. The rationale for targeting HR regions in the prefusion conformation is because the antimicrobial peptide can potentially prevent protein refolding and fusion. Hence, this will prevent the formation of the 6-HB (postfusion) and the entry of the virus into the host cell. It is for this reason we retrieved the cryo-EM structure of the MERS S spike protein from Protein Data Bank (PDB) with the PDB ID: 5X59 [20], which is a prefusion structure of MERS-CoV spike glycoprotein with threefold symmetry as shown in Figure 3(a). The postfusion of MERS-CoV spike protein is shown in Figure 3(b). The structure was elucidated by electron microscopy with a resolution of 3.7Å. This structure has total weight of 444204.84 and a sequence length of 3969 amino acid residues.

2.2. Database Screening of Antimicrobial Peptides

The set of AMPs were retrieved from the antimicrobial peptide database (version 3), APD3 [15]. This database contains a total of 2961 AMPs from six kingdoms, namely, bacteria, archaea, protists, fungi, plants, and animals. In particular, we selected a list of basic antimicrobial peptides that are broad-spectrum. The strategy employed is based on a similar database screening [21] with additional criteria. The extracted AMPs from APD3 were filtered according to the following criteria, namely,(a)20aa length 55aa: The longer the sequence, the better the antiviral activity [14],(b)basic residues should be abundant [13],(c)net charge because the virus membrane is negatively charged [21],(d)nontoxic to mammalian cells [21],(e)peptides with unknown anti-MERS-CoV activity,(f)not annotated as synthetic (i.e., man-made peptides) in the database [21],(g)interfacial activity [8] should be as follows:(i)Wimley-White interfacial hydrophobicity scale (WWIHS 0),(ii)interfacial helical hydrophobic moment (iHHM 0).

In addition to the list of AMPs, we selected a number of peptides from the literature, which have been verified experimentally to have anti-MERS-CoV activity. These peptides act as positive control in which their docking complexes with MERS-CoV will be compared with our predicted complexes, in particular, two positive controls, namely P9 [13] and HR2P [10]. P9 is a subsequence derived from mouse beta defensin (mBD4) while HR2P is a peptide from the HR2 region of MERS-CoV spike protein.

2.3. Ab Initio Modelling of the Shortlisted AMPs and Validation

The 3D structure of the shortlisted AMPs was predicted by submitting amino acid sequence into I-TASSER (, an online server which stands for Iterative Threading ASSembly Refinement [22]. This integrated platform works based on sequence-structure-function relation. This is an automated modelling server, which predicts a model based on confidence score (C-score) and build five models with confidence ranging from -5 to 2. C-score gives the estimate of accuracy of the prediction. If the C-score increases, the confidence of the model also increases. Based on C-score, best model of the peptide was selected for the study. The predicted AMPs structures obtained from I-TASSER were subjected for validation using Rampage server ( [23]. This validation recognizes errors in theoretical models of protein structures by performing statistical analysis of all available protein structures.

2.4. Protein Preparation of the Receptor (5X59) and the Ligand AMPs

The protein preparation wizard (PrepWizard) was used to prepare the structure of the modelled AMPs and also the crystal structure of the fusion core spike protein S2 (5X59). The aim of protein preparation was to optimize the molecule (Schrödinger Suite 2018 Protein Preparation Wizard, Schrödinger LLC, New York, NY, 2018). The protein chains were edited for missing hydrogen atoms; bond orders and hydrogen bonds were optimized. The preparation process of the protein continued until it attained a minimized state, which usually has a default Root Mean Square Deviation (RMSD) value of 0.30.

2.5. AMPs-Receptor Docking

To determine the binding mode of shortlisted AMPs with the spike protein receptor (5X59), 5X59 and the modelled AMPs were subjected to peptide-protein docking using Piper module of Schrödinger [24, 25]. Piper algorithm is based on fast Fourier transform and it consists of two steps, namely, conformational sampling and structural clustering. The conformational sampling involves performing exhaustive evaluation of an energy function given in Equation (1) [25] in a discretized space of mutual orientations of two proteins. On the other hand, the structural clustering aids in identifying and ranking likely docked protein poses.For peptide-protein docking, the AMPs were set as ligands and docked with receptor 5X59. The number of ligand rotation to probe was set for 10,000 rotations and, for each dock, five poses were retrieved. This was done in order to find large clusters of structures below a certain energy value. The shortlisted AMPs with the best Piper cluster size than experimentally validated peptides against MERS-CoV are considered as putative anti-MERS-CoV AMPs. In addition, we have used ClusPro 2.0 server [26] in determining protein-protein interaction. Briefly, ClusPro, rotates the ligand with 70,000 rotations and, for each rotation, it translates the ligand in , , and axis relative to the receptor on a grid. The ClusPro 2.0 server is based on Piper, but the method is extended to be used with pairwise interaction potentials [26].

2.6. Binding Mode of Docked Complexes

The docked complex structure output format was submitted into the Protein Interactions Calculator (PIC) webserver ( in order to map the interaction of the resulting docked complex [27]. The parameters such as number of hydrogen bonds, number of hydrophobic residues, and number of aromatic and ionic interactions were considered in interpreting the strength of the interaction.

3. Results

3.1. Peptide Modelling Using I-TASSER and Validation

As we theorize that the spike protein of MERS-CoV represents the key receptor for our analysis, we focus on developing a theoretical model for the selected AMPs using I-TASSER server and its evaluation using Ramachandran plot. The Ramachandran plots of the theoretical models were developed and compared using a server, namely, Rampage. Out of the 37 models developed, most of the models had good quality score and backbone conformation which are considered as reliable. Supplementary Table 1 represents the number of residues in the favored region, allowed region, and outlier region in 13 models predicted.

3.2. Filtering of AMPs Based on Database Screening Criteria

The filtering process using the criteria mentioned in Section 2.2 resulted in 37 shortlisted AMPs as shown in Table 1, where majority of the AMPs belong to the defensin family from different species.

S.No Peptide APD3 ID Definition LengthWWIHS iHHM Net charge

1 KTCENLADTFRGPCFATSNCAP00532Lunatusin201.82.62 0
2 GLFVGVLAKVAAHVVPAIAEHFAP00260 Maculatin 1.1 22 1.29 3.26 1
3 GIGKFLHSAGKFGKAFVGEIMKSAP00771 Magainin 1 23 1.34 7.19 3
4 GIGKFLHSAKKFGKAFVGEIMNSAP00144Magainin 2 23 0.93 7.1 3
5 GEGFLGMLLHGVGHAIHGLIHGKAP02663Piscidins 23 1.26 4.72 0
6 GLRSKIWLWVLLMIWQESNKFKKMAP00764 Dermaseptin-S9 24 6.07 2.95 4
7 FLPVLAGIAAKVVPALFCKITKKCAP00074Brevinin-1 24 1.31 3.48 4
8 GWGSFFKKAAHVGKHVGKAALTHYLAP00166Pleurocidin 25 0.37 7.49 4
9 FFGWLIRGAIHAGKAIHGLIHRRRHAP00340Chrysophsin-2 25 0.02 5.97 0
10 ALWMTLLKKVLKAAAKAALNAVLVGANAAP00160Dermaseptin-S4 28 0.92 4.71 4
13 GWFKKAWRKVKNAGRRVLKGVGIHYGVGLIAP00692Hagfish cathelicidin 30 0.24 6.41 8
15 RRCICTTRTCRFPYRRLGTCIFQNRVYTFCCAP00174Guinea pig neutrophil 31 2.16 2.06 7
16 ACYCRIGACVSGERLTGACGLNGRIYRLCCRAP00225RatNP-4 rat defensin, 31 2.43 2.58 4
20 GSIPACGESCFKGKCYTPGCSCSKYPLCAKNAP01065Cycloviolacin O14 31 0.73 1.48 3
21 CGESCVFIPCITTVLGCSCSIKVCYKNGSIPAP02571Cycloviolacin VY1 31 3.15 2.77 1
22 ATCYCRTGRCATRESLSGVCEISGRLYRLCCRAP00180Human defensin 5 32 1.39 3.83 4
23 AFTCHCRRSCYSTEYSYGTCTVMGINHRFCCLAP00181Human defensin 6 32 3.82 2.56 2
24 VTCYCRRTRCGFRERLSGACGYRGRIYRLCCRAP00222RatNP-1 rat defensin 32 1.02 4.29 8
25 VTCYCRSTRCGFRERLSGACGYRGRIYRLCCRAP00223RatNP-2 rat defensin 32 1.7 4.01 7
27 VCSCRLVFCRRTELRVGNCLIGGVSFTYCCTRVAP00179Human neutrophil peptide-4333.28 1.54 4
28 DFASCHTNGGICLPNRCPGHMIQIGICFRPRVKCCRSWAP00036Bovine beta-defensin 138 0.22 4.72 4
34 VSFPWSCAALSGVCRQGACLPSELYFGPLGCGKGSLCCVSYFLAP02830Channel catfish beta defensin438.812.3 1

3.3. Protein Preparation, Docking, and Evaluation of Top Complexes

The PrepWizard prepared the structures by automatically adding missing hydrogen atoms and correcting bond order assignments, charge states, and orientation of various groups and performed restrained minimizations which allow hydrogen atoms to be freely minimized.

Further, AMP-MERS docking (docking of antimicrobial peptides with 5X59) was performed by using Piper algorithm and the pose with the best fit was selected for each peptide-protein complex based on cluster size. The resulting structures were grouped into clusters and ranked according to cluster size values from the largest to the smallest. Top ranked peptide and protein complex details are presented in Tables 2, 3, and 4.

Rank Peptide Length DefinitionSpecies Cluster

1 P930 Mouse Beta-defensin Mouse 328
2 AP0022531RatNP-4 (rat defensin) Rat 285
3 AP0018032Human defensin 5 (alpha defensin) Human277
4 AP0054940Plectasin (fungal defensin) Fungus 270
5 AP0074441AvBD-5, chicken avian beta defensin) Chicken253
6 AP0072929Kalata B1 (cyclotides) Plant 247
7 AP0076424Dermaseptin-S9 Frog 223
8 AP0022332RatNP-2 (rat alpha defensin) Rat 219
9 HR2P36 HR2 region of MERS-CoV Synthetic 208

10 AP0016028Dermaseptin-S4 Frog 200
11 AP0017431Guinea pig neutrophil cationic peptide 1 Guinea pig193
12 AP0073031Kalata B8 (cyclotides) Plant 175
13 AP0022232RatNP-1 (rat alpha defensin,) Rat 175
14 AP0266323Piscidins Fish 171
15 AP0106131Circulin D (cyclotides) Plant 163
16 AP0135646Cp-thionin II Plant 160
17 AP0053220Lunatusin Plant 157
18 AP0257131Cycloviolacin VY1 (cyclotides) Plant 155
19 AP0069230HFIAP-3 (Hagfish cathelicidin) Fish148
20 AP0026022Maculatin 1.1 Frog 146
21 AP0273355Piscidin Fish 144
22 AP0027531Circulin B (cyclotides) Plant 143
23 AP0164430Brevinin-2-RN1 Frog 143
24 AP0214850Apl-AvBD16 (Beta def) Bird 139
25 AP0106531Cycloviolacin 014 (cyclotides) Plant 136
26 AP0102231Cycloviolin A (cyclotides) Plant 134
27 AP0003638Bovine Beta-defensin 1 Bovine 127
28 AP0007424Brevinin-1 Frog 120
29 AP0034025Chrysophsin-2 Fish 120
30 AP0016625Pleurocidin Fish 118
31 AP0283043ccBD (Channel Catfish beta def) Fish 115
32 AP0018132Human defensin 6 Human 115
33 AP0178843Myticin C molluscs 111
34 AP0017933Human neutrophil peptide-4 (Alpha def) Human97
35 AP0064133Pardaxin 1 Fish 96
36 AP0077123Magainin 1 Frog 90
37 AP0074242Chicken AvBD6 (Beta def) Chicken 88
38 AP0014423Magainin 2 Frog 85
39 AP0084643Mundticin KS (Bacteriocin) Bacteria66

Rank Peptide Length DefinitionSpecies Representative Member

1 AP0016625Pleurocidin Fish Center 134
2 AP0064133Pardaxin 1 Fish Center 134
3 AP0014423Magainin 2 Frog Center 117
4 AP0077123Magainin 1 Frog Center 117
5 AP0164430Brevinin-2-RN1 Frog Center 117
6 AP0076424Dermaseptin-S9 Frog Center 110
7 AP0257131Cycloviolacin VY1 (cyclotides) Plant Center 110
8 AP0027531Circulin B (cyclotides) Plant Center 107
9 AP0102231Cycloviolin A (cyclotides) Plant Center 107
10 AP0106131Circulin D (cyclotides) Plant Center 107
11 AP0054940Plectasin (fungal defensin) Fungus Center 101
12 AP0072929Kalata B1 (cyclotides) Plant Center 101
13 AP0073031Kalata B8 (cyclotides) Plant Center 101
14 AP0106531Cycloviolacin 014 (cyclotides) Plant Center 101
15 AP0178843Myticin C molluscs Center 97
16 AP0135646Cp-thionin II Plant Center 93
17 AP0074242Chicken AvBD6 (Beta def) Chicken Center 87
18 AP0214850Apl-AvBD16 (Beta def) Bird Center 87
19 AP0084643Mundticin KS (Bacteriocin) Bacteria Center 83
20 AP0053220Lunatusin Plant Center 78
21 P930Mouse Beta-defensin Mouse Center 68
22 AP0069230HFIAP-3 (Hagfish cathelicidin) Fish Center 67
23 AP0003638Bovine Beta-defensin 1 Bovine Center 66
24 AP0007424Brevinin-1 Frog Center 66
25 AP0074441AvBD-5, chicken avian beta defensin) Chicken Center 66
26 AP0266323Piscidins Fish Center 66
27 AP0017933Human neutrophil peptide-4 (Alpha def) Human Center 57
28 AP0017431Guinea pig neutrophil cationic peptide 1 Guinea pig Center 49
29 AP0273355Piscidin Fish Center 43
30 AP0016028Dermaseptin-S4 Frog Center 40
31 AP0018032Human defensin 5 (alpha defensin) Human Center 40
32 AP0022232RatNP-1 (rat alpha defensin,) Rat Center 40
33 AP0022332RatNP-2 (rat alpha defensin) Rat Center 40
34 HR2P36HR2 region of MERS-CoVSynthetic Center 39
35 AP0034025Chrysophsin-2 Fish Center 38
36 AP0018132Human defensin 6 Human Center 37
37 AP0026022Maculatin 1.1 Frog Center 37
38 AP0283043ccBD (Channel Catfish beta def) Fish Center 33
39 AP0022531RatNP-4 (rat defensin) Rat Center 31

Rank Peptide Length DefinitionSpecies Representative Energy

1 AP0026022Maculatin 1.1 Frog Center -1692.0
2 AP0273355Piscidin Fish Center -1581.2
3 AP0017933Human neutrophil peptide-4 (Alpha def) Human Center -1498.5
4 AP0034025Chrysophsin-2 Fish Center -1488.6
5 AP0018132Human defensin 6 Human Center -1399.5
6 AP0018032Human defensin 5 (alpha defensin) Human Center -1340.8
7 AP0022232RatNP-1 (rat alpha defensin,) Rat Center -1340.8
8 AP0022332RatNP-2 (rat alpha defensin) Rat Center -1340.8
9 AP0076424Dermaseptin-S9 Frog Center -1338.8
10 AP0074242Chicken AvBD6 (Beta def) Chicken Center -1264.8
11 AP0214850Apl-AvBD16 (Beta def) Bird Center -1264.8
12 HR2P36HR2 region of MERS-CoV Synthetic Center -1256.9
13 AP0017431Guinea pig neutrophil cationic peptide 1 Guinea pig Center -1223.0
14 AP0178843Myticin C molluscs Center -1202.7
15 AP0283043ccBD (Channel Catfish beta def) Fish Center -1184.3
16 AP0007424Brevinin-1 Frog Center -1184.2
17 AP0266323Piscidins Fish Center -1184.2
18 AP0135646Cp-thionin II Plant Center -1139.1
19 AP0016625Pleurocidin Fish Center -1137.1
20 AP0022531RatNP-4 (rat defensin) Rat Center -1103.8
21 AP0003638Bovine Beta-defensin 1 Bovine Center -1103.7
22 AP0074441AvBD-5, chicken avian beta defensin) Chicken Center -1103.7
23 AP0016028Dermaseptin-S4 Frog Center -1097.0
24 AP0064133Pardaxin 1 Fish Center -1050.1
25 AP0054940Plectasin (fungal defensin) Fungus Center -994.2
26 AP0027531Circulin B (cyclotides) Plant Center -993.9
27 AP0102231Cycloviolin A (cyclotides) Plant Center -993.9
28 AP0106131Circulin D (cyclotides) Plant Center -993.9
29 AP0084643Mundticin KS (Bacteriocin) Bacteria Center -937.1
30 P930Mouse Beta-defensin Mouse Center -925.5
31 AP0053220Lunatusin Plant Center -921.9
32 AP0257131Cycloviolacin VY1 (cyclotides) Plant Center -897.7
33 AP0014423Magainin 2 Frog Center -868.2
34 AP0077123Magainin 1 Frog Center -868.2
35 AP0164430Brevinin-2-RN1 Frog Center -868.2
36 AP0069230HFIAP-3 (Hagfish cathelicidin) Fish Center -821.3
37 AP0072929Kalata B1 (cyclotides) Plant Center -805.8
38 AP0073031Kalata B8 (cyclotides) Plant Center -805.8
39 AP0106531Cycloviolacin 014 (cyclotides) Plant Center -805.8

The results indicate that, out of 37 AMPs, 8 AMPs had a cluster size greater than 200 which is an indication of the binding of the peptides to MERS-CoV spike protein as shown in Table 2. These AMPs are derived from various sources such as fungus, plants, and fish. In addition, it was found that the seven AMPs had higher cluster size value than the positive control (HR2P: cluster value of 208), but lower cluster size value than the positive control (P9: cluster value of 328), where HR2P and P9 have been experimentally demonstrated to possess potent anti-MERS-CoV activity [10, 13]. In particular, we considered a cutoff value of 208. The positive control P9 was the best and had a cluster size value of 328 as shown in Table 2. Peptide AP00225 showed a very strong binding affinity score with a cluster size (binding affinity score) of 285 compared to all other putative peptides. Other top ranked putative AMPs include AP00180, AP00549, AP00744, AP00729, AP00764, and AP00223 with cluster size values of 277, 270, 253, 247, 223, and 219, respectively. These confirm the probability of these five peptides to be putative anti-MERS-CoV peptides. For further analyses, we selected four peptides, namely, AP00225, AP00180, AP00549, and AP00744, belonging to family of defensin. During validation of these four putative defensin peptides, AP00225 Ramachandran plot (Psi-Phi) pairs had 79.3% of residues in most favored regions, 6.9% core residues in allowed regions, and 13.8% residues in outlier regions. AP00180 Ramachandran plot (Psi-Phi) pairs had 84.3% of residues in most favored regions, 10.0% core residues in allowed regions, and 6.7% residues in outlier regions. AP00549 Ramachandran plot (Psi-Phi) pairs had 81.6% of residues in most favored regions, 7.9% core residues in allowed regions, and 10.5% residues in outlier regions. AP00744 Ramachandran plot (Psi-Phi) pairs had 87.2% of residues in most favored regions, 7.7% core residues in allowed regions, and 5.1% residues in outlier regions (Supplementary Table 1).

In addition, the results show that 20 AMPs and 12 AMPs were ranked higher than P9 and HR2P, respectively, shown in Table 3. However, the ranking based on energy scores shows that 11 AMPs and 17 AMPs were ranked higher than HR2P and P9 as tabulated in Table 4. We have used the results given by Piper (Table 2) and have used cluster size because, in ClusPro 2.0 documentation, they mentioned that the best way to rank models is by cluster size and not by energy scores.

3.4. Evaluation of Peptide-Protein Complex and Its Interactions Analysis

Once we observed that the AMPs could potentially bind to spike protein, the next step was to know the binding mode. In particular, we have used Protein Interactions Calculator (PIC) to recognize the interactions within the bound complexes. In structural bioinformatics, predicting protein-protein interactions which stabilize the tertiary and quaternary structures is an important task. For the top best four AMPs-MERS-CoV complexes with the best cluster size were subjected to PIC server and the binding mode (interactions) of each peptide are given in Table 5. PIC identified interactions such as hydrophobic residues interactions, ionic interactions, hydrogen bonds, aromatic-aromatic interactions and aromatic–sulphur interactions within the peptide-protein complexes. According to the PIC server results as shown in Table 5, AP00225 forms hydrophobic interactions with Val790, Tyr1142, Phe764, Leu731, Ile768, Pro1143, Pro767, and Val770; hydrogen bond interactions with Pro730; and ionic interactions with Gln792 and Ser734 as shown in Table 5. AP00180 forms hydrophobic interactions with Ala1007, Val790, Leu731, Pro767, Ile768, and Tyr1142; hydrogen bond interactions with Gly789 and Pro730; and ionic interactions with Glu1017 and Asp740 as shown in Table 5. AP00549 forms hydrophobic interactions with Ala1049, Pro59, Tyr64, Tyr928, Val929, Ala930, Ala920, Ile69, and Tyr71; hydrogen bond interactions with Ala1049 and Gly61; and ionic interactions with Arg1057, Arg62, and Asp922 as shown in Table 5. AP00744 forms hydrophobic interactions with Leu1200, Pro767, Val1168, Ile1180, Leu780, Phe778, Pro1143, Val983, and Ile985; hydrogen bond interactions with Ala1206; and ionic interactions with Asp771 as shown in Table 5. These residues may be considered as critical residues. AP00549 have overlapping residues with experimentally validated anti-MERS-CoV peptide P9, as highlighted in bold (Table 5) and Figure 4(a), while AP00225, AP00180, and AP00744 have common residues with HR2P as highlighted in italic (Table 5) and Figure 4(b). The binding of the peptide AP00549 has the same binding region as P9; AP00225, AP00180, and AP00744 have the same binding region as shown in Figures 5 and 6. The binding of the peptides to the receptor spike and ligands includes AP00179, AP00260, AP00340, AP02733, P9, and HR2P; see Figure 7.

APD3 IDHydrophobic interactions Main chain Hydrogen bond interactions Side chain Hydrogen bond interactions Ionic interactions Aromatic-aromatic interactions Aromatic-sulphur interactions

P9Tyr64, Ile69, Tyr809, Ala920, Tyr928, Val929, Tyr932, Ala1037Asn812, Ser919, Asp922, Asn1042, Asn812, Ser1038Tyr932Glu1039Tyr928, Tyr64 Cys925

AP00549 Ala1049, Pro59, Tyr64, Tyr928, Val929, Ala930, Ala920, Ile69, Tyr71Ala1049, Gly61Gln60, Gln1056, Cys925 Arg1057, Arg62, Asp922Tyr928, Tyr71 -

AP00225Val790, Tyr1142, Phe764, Leu731, Ile768, Pro1143, Pro767, Val770 Pro730 Thr791, Gln1119, Tyr1141, Leu729, Asn765, Leu731, His1146, Asn765, His766 Gln792, Ser734Glu1017Tyr1142

AP00180 Ala1007, Val790, Leu731, Pro767, Ile768, Tyr1142 Gly789, Pro730 Gln1119, Tyr1142, Leu731, Leu729, Pro730, Asn765 Aln1007Glu1017, Asp740 - Cys1142

AP00744 Leu1200, Pro767, Val1168, Ile1180, Leu780, Phe778, Pro1143, Val983, Ile985 Ala1206 Cys1164, Val770, Tyr1153, Ser781 Asp771Tyr1142 -

HR2P Tyr1153, Ile1165, Val1168, Ile1180, Val1181, Leu780, Val770, Pro767, Pro1143, Tyr1142, Val790, Leu729, Pro730, Leu731-Asn1169His1146, His766- -

4. Discussion

Computational and structural biology methods have accelerated the discovery of novel drugs used to treat viral diseases [21, 28]. We followed the structural biology aspects which focus on the availability and retrieval of an S protein receptor structure from PDB which was resolved using cryo-EM structure method. We have applied the docking technique not only to predict the binding mode of AMPs to spike protein but also to study the peptide-protein interactions. The receptor used is the prefusion state of the S protein, because it is a type I fusion protein, which undergoes a nonreversible conformational change that results in the postfusion form of the protein. In postfusion, the protein has refolded and the membranes have undergone fusion, or the spike protein has been spent. Therefore, it is reasonable to target the S protein in the prefusion conformation as peptides bind to the prefusion conformation. This can potentially prevent protein refolding and entry of the virus into the host cell, as it has been shown for diverse Type I proteins like HIV gp41 and paramyxovirus F proteins. Once the 6HB fusion core is formed, it will not accommodate any peptide within it and any peptide related therapeutic intervention would have to take place either at the prefusion stage, or at the intermediate hairpin stages [29].

The model refinement has improved its quality of the theoretical models developed for selected AMPs. Further, Ramachandran plots from Rampage server determined the stereochemical quality of best scored four defensin models. The aim of model refinement using Rampage Ramachandran plot was to determine whether the theoretical model acquired the quality and side chain configuration [30]. This program helps researchers to evaluate the accuracy of the predicted models [31].

Based on filtering of potential peptides acting against S protein, we employed interfacial activity based on WWIHS and iHMM scales. It has been hypothesized that positive interfacial hydrophobicity of a peptide increases the chances for membrane binding, that is, interacting with the viral hydrophobic surfaces and hence inhibiting fusion and entry of the virus [8]. Therefore, since the best five putative AMPs have higher values of interfacial hydrophobicity, we can infer that their mechanism of action is by inhibiting fusion and hence blocking entry of the virus into the host cell. The mechanism of action for the positive controls, namely, P9 [13] and HR2P [10] has been experimentally validated to inhibit viral fusion. Nevertheless, these five putative anti-MERS-CoV AMPs have same values as positive control peptides in terms of WWHIS and iHMM values and also on an average sequence length.

In order to characterize binding properties of AMPs with the spike protein S2, we used Piper, which involves peptide-protein interaction, in which it determines the best-fit orientation of ligand with receptor [25]. The binding affinity is determined by Piper cluster size and not scores or probability. The docking score together with probability can give confidence on the binding and which are lacking in our study since we used cluster size as criteria for ranking best poses. However, it has been hypothesized that the best way to rank docking is by cluster size, which can be useful in informing experimental approaches [25]. In this study, we selected five peptides which had higher cluster value than experimentally validated anti-MERS-CoV peptide, P9 [13]. Nevertheless, these peptides have good binding with MERS-CoV spike protein (S2) in terms of hydrogen bonds and hydrophobic interactions.

Our results may indicate that AP00225 (Rat defensin), AP00180 (Human alpha defensin), AP00549 (Plectasin), AP00744 (Chicken beta defensin), AP00729 (Cyclotides), AP00764 (Dermaseptin-S9), and AP00223 (Rat alpha defensin) were the best docked peptides with high cluster size values, which is an indicative of strong binding affinity. Most of these putative peptides are defensin, similar to P9, which is a mouse beta defensin. On comparison of P9 and AP00549, we found that five residues, namely, Tyr64, Ile69, Ala920, Tyr928, and Val929, were common in the binding. On the same breath, HR2P have nine common residues (Leu780, Pro767, Tyr1142, Leu729, Leu731, Val1168, Val770, Pro1143, Val790) with AP00225, AP00180, and AP00744. These residues may be considered as key or critical and may play a major role in the protein protein-interaction [32] and might inhibit the formation of the six-helical bundle (6-HB). Further studies may help to understand the role of these residues in drug binding mechanism.

Interestingly, these seven best peptides have been experimentally validated to have activity against various microorganisms. For instance, AP00225 and AP00223 are rat defensin and this peptide has activity against Escherichia coli ML-35, Acinetobacter calcoaceticus HON-1, Staphylococcus aureus 502A, and Candida albicans 820 in vitro [33]. AP00180 is a Human alpha defensin and it inhibits nonenveloped BK virus infection [34]. AP00549 is fungal plectasin which belongs to the family of defensins and primary source is Pseudoplectania nigrella and this peptide is an effective antiviral against dengue virus [31]. AP00744 (Chicken avian beta defensin) has antimicrobial activity against Salmonella serovars [35]. AP00729 is kalata B1 which is generally known as plant cyclotides acting as a stable component for drug discovery [36]. Cyclotides possess many biological activities such as anti-HIV, antimicrobial, and hemolytic. Some cyclotides show cell penetrating properties. AP00764 (dermaseptin) induces the migration of immune cells. Dermaseptin shows inhibition against the microorganisms and cancer cell lines [37]. This peptide has innate immunity properties [38]; however, it has not been shown to have antiviral activity against MERS-CoV. This could suggest that mechanism of action of this and other peptides could be varied.

Finally, lack of wet-lab validation is a drawback in our research and we expect computational biology analysis and its integration with wet-lab data can be productive in the determination of potential anti-MERS-CoV components.

5. Conclusion

In the present work, we have used a docking and scoring algorithm for combined peptide-protein binding mode search using a list of AMPs. We named the method as AMP-MERS docking as it is a novel application of AMPs to MERS-CoV spike protein. Our computational study confirms that four AMPs were able to bind clearly to the specific binding site of S protein (5X59). From our results, it may confirm that these AMPs may be suitable for inhibiting MERS-CoV virus entry into the host cell by binding and preventing fusion. However, the results are preliminary and certainly need experimental confirmation using in vitro and in vivo experiments essential to validate them. Special assays studies are needed to confirm the mechanism of action. Considering all the structural aspects and binding affinity studies of the four AMPs may possibly be the first choice as an anti-MERS-CoV AMPs which could be exploited to design potential inhibitors for treating MERS. We conclude that molecular docking studies aid in deciphering the antiviral activity of molecules by determining the inhibition score and binding energy.

Data Availability

The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest

The authors declare no conflicts of interest.


The authors acknowledge the contribution of Dr. Mohamed Hussein, who assisted in installing Schrödinger software. This research was funded by King Abdullah International Medical Research Center grant number RC16/089.

Supplementary Materials

Supplementary Table 1 shows the model evaluation of modelled AMPs using Rampage Server. (Supplementary Materials)


  1. S. Milne-Price, K. L. Miazgowicz, and V. J. Munster, “The emergence of the Middle East Respiratory Syndrome coronavirus,” Pathogens and Disease, vol. 71, no. 2, pp. 121–136, 2014. View at: Publisher Site | Google Scholar
  2. H. A. Mohd, J. A. Al-Tawfiq, and Z. A. Memish, “Middle east respiratory syndrome coronavirus (MERS-CoV) origin and animal reservoir Susanna Lau,” Virology Journal, vol. 13, no. 1, article 87, 2016. View at: Google Scholar
  3. L. I. Melnik, R. F. Garry, and C. A. Morris, “Peptide inhibition of human cytomegalovirus infection,” Virology Journal, vol. 8, article 76, 2011. View at: Publisher Site | Google Scholar
  4. “Safety study of flufirvitide-3 nasal spray in healthy subjects,” View at: Google Scholar
  5. J. W. Koehler, J. M. Smith, D. R. Ripoll et al., “A fusion-inhibiting peptide against rift valley fever virus inhibits multiple, diverse viruses,” PLOS Neglected Tropical Diseases, vol. 7, no. 9, pp. 1–11, 2013. View at: Google Scholar
  6. Y. M. Hrobowski, R. F. Garry, and S. F. Michael, “Peptide inhibitors of dengue virus and West Nile virus infectivity,” Virology Journal, vol. 2, no. 1, article 49, 2005. View at: Publisher Site | Google Scholar
  7. B. Sainz, E. C. Mossel, W. R. Gallaher et al., “Inhibition of severe acute respiratory syndrome-associated coronavirus (SARS-CoV) infectivity by peptides analogous to the viral spike protein,” Virus Research, vol. 120, no. 1-2, pp. 146–155, 2006. View at: Publisher Site | Google Scholar
  8. H. Badani, R. F. Garry, and W. C. Wimley, “Peptide entry inhibitors of enveloped viruses: The importance of interfacial hydrophobicity,” Biochimica et Biophysica Acta (BBA) - Biomembranes, vol. 1838, no. 9, pp. 2180–2197, 2014. View at: Publisher Site | Google Scholar
  9. T.-J. Chia, Y.-C. Wu, J.-Y. Chen, and S.-C. Chi, “Antimicrobial peptides (AMP) with antiviral activity against fish nodavirus,” Fish and Shellfish Immunology, vol. 28, no. 3, pp. 434–439, 2010. View at: Publisher Site | Google Scholar
  10. L. Lu, Q. Liu, Y. Zhu et al., “Structure-based discovery of Middle East respiratory syndrome coronavirus fusion inhibitor,” Nature Communications, vol. 5, article 3067, 2014. View at: Publisher Site | Google Scholar
  11. R. Channappanavar, L. Lu, S. Xia et al., “Protective effect of intranasal regimens containing peptidic middle east respiratory syndrome coronavirus fusion inhibitor against MERS-CoV infection,” The Journal of Infectious Diseases, vol. 212, no. 12, pp. 1894–1903, 2015. View at: Publisher Site | Google Scholar
  12. J. Gao, G. Lu, J. Qi et al., “Structure of the fusion core and inhibition of fusion by a heptad repeat peptide derived from the s protein of middle east respiratory syndrome coronavirus,” Journal of Virology, vol. 87, no. 24, pp. 13134–13140, 2013. View at: Publisher Site | Google Scholar
  13. H. Zhao, J. Zhou, K. Zhang et al., “A novel peptide with potent and broad-spectrum antiviral activities against multiple respiratory viruses,” Scientific Reports, vol. 6, article 22008, 2016. View at: Google Scholar
  14. L. Du, Y. Yang, Y. Zhou et al., “MERS-CoV spike protein: a key target for antivirals,” Expert Opinion on Therapeutic Targets, vol. 21, no. 2, pp. 131–143, 2017. View at: Publisher Site | Google Scholar
  15. G. Wang, X. Li, and Z. Wang, “APD3: the antimicrobial peptide database as a tool for research and education,” Nucleic Acids Research, vol. 44, no. D1, pp. D1087–D1093, 2016. View at: Publisher Site | Google Scholar
  16. M. Ciemny, M. Kurcinski, K. Kamel et al., “Protein–peptide docking: opportunities and challenges,” Drug Discovery Therapy, vol. 23, no. 8, pp. 1530–1537, 2018. View at: Publisher Site | Google Scholar
  17. A. V. Hoojghan, Application of graphical models in protein-protein interactions and dynamics [Ph.D. thesis], University of Massachusetts Boston.
  18. F. Zhao, H. Zeng, X. Li et al., “Complexcontact: a web server for inter-protein contact prediction using deep learning,” Nucleic Acids Research, vol. 46, no. W1, pp. W432–W437, 2018. View at: Google Scholar
  19. A. Valencia, I. Ezkurdia, L. Bartoli, M. L. Tress, P. Fariselli, and R. Casadio, “Progress and challenges in predicting proteinprotein interaction sites,” Briefings in Bioinformatics, vol. 10, no. 3, pp. 233–246, 2009. View at: Google Scholar
  20. H. M. Berman, J. Westbrook, Z. Feng et al., “The protein data bank,” Nucleic Acids Research, vol. 28, no. 1, pp. 235–242, 2000. View at: Publisher Site | Google Scholar
  21. G. Wang, “Natural antimicrobial peptides as promising anti-hiv candidates,” Current Topics in Peptide and Protein Research, vol. 13, pp. 93–110, 2012. View at: Google Scholar
  22. A. Roy, A. Kucukural, and Y. Zhang, “I-TASSER: a unified platform for automated protein structure and function prediction,” Nature Protocols, vol. 5, no. 4, pp. 725–738, 2010. View at: Publisher Site | Google Scholar
  23. S. C. Lovell, I. W. Davis, W. B. Arendall et al., “Structure validation by cα geometry: ϕ, ψ and cβ deviation,” Proteins: Structure, Function, and Bioinformatics, vol. 50, no. 3, pp. 437–450, 2003. View at: Publisher Site | Google Scholar
  24. G. Chuang, D. Kozakov, R. Brenke, S. R. Comeau, and S. Vajda, “DARS (decoys as the reference state) potentials for protein-protein docking,” Biophysical Journal, vol. 95, no. 9, pp. 4217–4227, 2008. View at: Publisher Site | Google Scholar
  25. D. Kozakov, R. Brenke, S. R. Comeau, and S. Vajda, “PIPER: an FFT-based protein docking program with pairwise potentials,” Proteins: Structure, Function, and Genetics, vol. 65, no. 2, pp. 392–406, 2006. View at: Publisher Site | Google Scholar
  26. D. Kozakov, D. R. Hall, B. Xia et al., “The ClusPro web server for protein–protein docking,” Nature Protocols, vol. 12, no. 2, pp. 255–278, 2017. View at: Publisher Site | Google Scholar
  27. K. G. Tina, R. Bhadra, and N. Srinivasan, “PIC: protein interactions calculator,” Nucleic Acids Research, vol. 35, no. 2, pp. W473–W476, 2007. View at: Publisher Site | Google Scholar
  28. A. O. Adedeji, K. Singh, A. Kassim et al., “Evaluation of SSYA10-001 as a replication inhibitor of severe acute respiratory syndrome, mouse hepatitis, and middle east respiratory syndrome coronaviruses,” Antimicrobial Agents and Chemotherapy, vol. 58, no. 8, pp. 4894–4898, 2014. View at: Publisher Site | Google Scholar
  29. E. Teissier, F. Penin, and E.-I. Pécheur, “Targeting cell entry of enveloped viruses as an antiviral strategy,” Molecules, vol. 16, no. 1, pp. 221–250, 2011. View at: Publisher Site | Google Scholar
  30. G. N. Ramachandran, C. Ramakrishnan, and V. Sasisekharan, “Stereochemistry of polypeptide chain configurations,” Journal of Molecular Biology, vol. 7, pp. 95–99, 1963. View at: Publisher Site | Google Scholar
  31. H. A. Rothan, Z. Mohamed, A. M. Suhaeb, N. A. Rahman, and R. Yusof, “Antiviral cationic peptides as a strategy for innovation in global health therapeutics for dengue virus: high yield production of the biologically active recombinant plectasin peptide,” OMICS: A Journal of Integrative Biology, vol. 17, no. 11, pp. 560–567, 2013. View at: Publisher Site | Google Scholar
  32. N. Haspel and F. Jagodzinski, Methods for Detecting Critical Residues in Proteins, Springer, New York, NY, USA, 2017.
  33. P. B. Eisenhauer, S. S. L. Harwig, D. Szklarek, T. Ganz, M. E. Selsted, and R. I. Lehrer, “Purification and antimicrobial properties of three defensins from rat neutrophils,” Infection and Immunity, vol. 57, no. 7, pp. 2021–2027, 1989. View at: Google Scholar
  34. A. S. Dugan, M. S. Maginnis, J. A. Jordan et al., “Human α-defensins inhibit BK virus infection by aggregating virions and blocking binding to host cells,” The Journal of Biological Chemistry, vol. 283, no. 45, pp. 31125–31132, 2008. View at: Publisher Site | Google Scholar
  35. P. Milona, C. L. Townes, R. M. Bevan, and J. Hall, “The chicken host peptides, gallinacins 4, 7, and 9 have antimicrobial activity against Salmonella serovars,” Biochemical and Biophysical Research Communications, vol. 356, no. 1, pp. 169–174, 2007. View at: Publisher Site | Google Scholar
  36. A. Gould, Y. Ji, T. L. Aboye, and J. A. Camarero, “Cyclotides, a novel ultrastable polypeptide scaffold for drug discovery,” Current Pharmaceutical Design, vol. 17, no. 38, pp. 4294–4307, 2011. View at: Publisher Site | Google Scholar
  37. D. Shi, X. Hou, L. Wang et al., “Two novel dermaseptin-like antimicrobial peptides with anticancer activities from the skin secretion of Pachymedusa dacnicolor,” Toxins, vol. 8, no. 5, 2016. View at: Google Scholar
  38. N. Iijima, N. Tanimoto, Y. Emoto et al., “Purification and characterization of three isoforms of chrysophsin, a novel antimicrobial peptide in the gills of the red sea bream, Chrysophrys major,” European Journal of Biochemistry, vol. 270, no. 4, pp. 675–686, 2003. View at: Publisher Site | Google Scholar
  39. S. Mustafa, H. Balkhy, and M. N. Gabere, “Current treatment options and the role of peptides as potential therapeutic components for Middle East Respiratory Syndrome (MERS): a review,” Journal of Infection and Public Health, vol. 11, no. 1, pp. 9–17, 2018. View at: Publisher Site | Google Scholar

Copyright © 2019 Sabeena Mustafa et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

More related articles

 PDF Download Citation Citation
 Download other formatsMore
 Order printed copiesOrder

Related articles

We are committed to sharing findings related to COVID-19 as quickly as possible. We will be providing unlimited waivers of publication charges for accepted research articles as well as case reports and case series related to COVID-19. Review articles are excluded from this waiver policy. Sign up here as a reviewer to help fast-track new submissions.