key: cord-1046230-fnj2shvn
authors: Pourmand, Saeed; Zareei, Sara; Shahlaei, Mohsen; Moradi, Sajad
title: Inhibition of SARS-CoV-2 pathogenesis by potent peptides designed by the mutation of ACE2 binding region
date: 2022-05-17
journal: Comput Biol Med
DOI: 10.1016/j.compbiomed.2022.105625
sha: 30609186dd2b2131474e3e2b5af395ee11fda485
doc_id: 1046230
cord_uid: fnj2shvn

The outbreak of COVID-19 has resulted in millions of deaths. Despite all attempts that have been made to combat the pandemic, the re-emergence of new variants complicated SARS-CoV-2 eradication. The ongoing global spread of COVID-19 demands the incessant development of novel agents in vaccination, diagnosis, and therapeutics. Targeting receptor-binding domain (RBD) of spike protein by which the virus identifies host receptor, angiotensin-converting enzyme (ACE2), is a promising strategy for curbing viral infection. This study aims to discover novel peptide inhibitors against SARS-CoV-2 entry using computational approaches. The RBD binding domain of ACE2 was extracted and docked against the RBD. MMPBSA calculations revealed the binding energies of each residue in the template. The residues with unfavorable binding energies were considered as mutation spots by OSPREY. Binding energies of the residues in RBD-ACE2 interface was determined by molecular docking. Peptide inhibitors were designed by the mutation of RBD residues in the virus-receptors complex which had unfavorable energies. Peptide tendency for RBD binding, safety, and allergenicity were the criteria based on which the final hits were screened among the initial library. Molecular dynamics simulations also provided information on the mechanisms of inhibitory action in peptides. The results were finally validated by molecular docking simulations to make sure the peptides are capable of hindering virus-host interaction. Our results introduce three peptides P7 (RAWTFLDKFNHEAEDLRYQSSLASWN), P13 (RASTFLDKFNHEAEDLRYQSSLASWN), and P19 (RADTFLDKFNHEAEDLRYQSSLASWN) as potential effective inhibitors of SARS-CoV-2 entry which could be considered in drug development for COVID-19 treatment.

Peptides with therapeutic potential have been considered during the last decades and the number of FDA-64 approved peptide drugs has increased [1] . They offer several advantages like ease of synthesis, high 65 specificity, and limited accumulative behavior. These benefits have made peptides favorable agents in the 66 development of diagnostic approaches, vaccines, and drugs against highly infectious viruses including 67 influenza, acquired immunodeficiency syndrome, chronic hepatitis B, dengue virus, and coronavirus 68 disease 2019 [2, 3] . Moreover, antiviral peptides (AVPs) have the potential to block a virus' cycle at 69 different levels from viral attachment to the host cell to its replication. Some AVPs have natural origins 70

[4-7] while others are discovered or rationally designed by bioinformatics techniques machine learning 71 [8] [9] [10] [11] . Since the outbreak of COVID-19, which is caused by the SARS-CoV-2 virus, peptide inhibitors 72

have also been among the promising anti-covid agents from various resources [12] . While some studies 73

seek to find peptide anti-COVID-19 agents from natural resources [13] , others tried to rationally design 74 novel peptides [12] . 75

Due to the importance of the entrance step in the pathogenesis of viruses as obligate intracellular 76

parasites, it has been the main target for anti-viral development. This is also the case about SARS-CoV-2 77 whose pathogenesis depends on angiotensin-converting enzyme II as a receptor [14] [15] [16] to either directly 78 fuse its genetic materials into the cell or enter in endosome-based endocytosis [17] . Host receptor 79 recognition is mediated by the SARS-CoV-2 spike S1 subunit which is expressed on the viral envelope. 80

Four main domains make up S, one of which is located in the direct interaction with ACE2, named 81

receptor-binding domain (RBD) [15, 18] . On the receptor side, the N-terminal helix of ACE2 is 82 recognized by RBD and interacts with the virus [16, 19] . 83

Peptide inhibitors against SARS-CoV-2 entry have been proposed using three main strategies. Firstly, 84 many peptides focused on virus fusion. This process is made possible by the catalytic cleavage of spike 85 protein by the host proteases TMPRSS2 (transmembrane protease serine 2), furin, and cathepsin-L [20-86 22], is shown that can be blocked by peptides [23] [24] [25] [26] [27] . The second strategy is approached by receptor 87

antagonists. The peptides in this category are designed or shown to block virus entry by occupying the 88 host N-terminal helix of ACE2 [28, 29] . While fusion inhibitors and receptor antagonists must approach 89 the host cell to perform their anti-COVID-19 activity, in the last category, virus inactivators, peptides can 90

prevent the viral infection before the virus achieve ACE2 N-helix in blood. These inhibitors mostly 91 interact with the SARS-CoV-2 RBD domain and neutralize its binding affinity for ACE2 [30] [31] [32] . 92

As the computational methods provide the cost and time effective manner for investigating the 93 biomolecular interactions in detailed state [33, 34] , in this study, we applied computational approaches 94

with the mutation-based rational design of peptides to prevent host cell recognition mediated by the 95 SARS-CoV-2 RBD domain, thereby discovering peptides with potential competitive affinity for blocking 96 viral pathogenesis. 97 98

The overall diagram of the methods is illustrated in Fig. 1 . 100

ACE2 residues that embrace viral spike RBD were considered as an inspiration for inhibitor design due to 102 their ability to form a stable complex with the virus. Therefore, the peptides which can mimic the binding 103 J o u r n a l P r e -p r o o f behavior of the virus binding region may interfere with the viral attachments and subsequently, may 104 prevent viral fusion and pathogenesis. 105

The PDB structure of SARS-CoV2 RBD complexed with ACE2 (PDB ID: 6m0j) was obtained from the 106 RCSB protein data bank ( approach which is measured by the amount of free energy change following the removal of water 119 molecules from the protein interface. This parameter takes the desolvation contribution into account. W1-120 4 coefficients are weighted which is calculated for different types of docking problems [36] . 121

After the separate introduction of the receptor's and peptides' files to the server, each run was performed 122

for the receptor and a peptide with default parameters. Among the results, the conformations with the 123 lowest docking scores were selected for further analysis. 124

In molecular dynamics (MD) simulations, the candidate peptides-ACE2 complexes were put into a 100-ns 125 simulation using GROMACS package v. 2020 [38], Gromos96 54a7 [39] , and SPCE water model.

Gromos 96 was selected since it has a favorable agreement with the NMR data and the X-ray crystal 127 structure of the protein [40] . Each complex was centered in a cubic box with a minimum distance of 1.0 128 nm from the edges. The energy of all systems was minimized for 50000 steps followed by a thermal 129 equilibrium step using Berendsen thermostat at 310 K. The pressure was equilibrated for 1 ns to achieve 130 the pressure of 1 bar using Berendsen barostat, LINCS algorithm [41] for bond constrain, and PME mesh 131

[42] for the calculation of long-range electrostatic interactions. Moreover, the Fourier grid spacing and 132

Coulomb radius were set at 0.16 and 1.2 nm, respectively, and the Van der Waals interactions were 133 limited to 1.2 nm. Finally, the production states were performed under the leapfrog algorithm for 100 ns. 134

The resulted trajectories were then analyzed by built-in GROMACS utilities e.g. root mean squared 135 deviation (RMSD), root mean squared fluctuation (RMSF), radios of gyration(Rg), principle component 136 analysis (PCA) and solvent accessible surface area (SASA). 137

To identify how important is the contribution of each reside in spike binding, the resides of decoy 139

peptide was analyzed using g_mmpbsa program [43] , the definitions of which can be applied to our 140 system as 141

, and ∆ describe the total energy of spike-template peptide 143 complex, solution free energy of spike, and solution free energy of free template peptide, respectively. 144

Peptide library was built by OSPREY v. 3.0 (Open-Source Protein Redesign for You) [44] python-based 146 script in which the input pdb files of the template was introduced as a strand. The mutation hotspots were 147 then defined by strand. Flexibility and the probable residues of Ile, Trp, Ser, Thr, and Asp were 148 determined by setLibraryRotamers scripts, respectively. The defined strand was used to make a conf 149 space and the osprey forcefield parameters were selected. The remained scripts were performed as with 150

default parameters according to the OSPREY documentation. 151

The toxicity and allergenicity properties of screened peptides were predicted by ToxinPred 

It is well known that the SARS-COV-2 virus is equipped with a glycoprotein anchor, Spike (S), which 160

guarantees host cell recognition and viral entry by interacting with angiotensin-converting enzyme 2 161

(ACE2) which is expressed on the surface of human cells [1, 2] . Spike protein has been the initial main 162 target for drug design. It is formed by S1 and S2 subunits, both of which are comprised of various 163

domains. While the S1 subunit is known to be responsible for cell recognition, the S2 subunit plays a vital 164 role in membrane fusion, following which the viral entry occurs. However, it is following the direct 165 establishment of interaction between the S1 and ACE2 peptidase domain that the spike undergoes 166 proteolytic cleavage, and the viral infection initiates [47] . S1 includes a domain called receptor-binding 167 domain (RBD) (residues 333-526) with two main subdomains: five anti-parallel β-sheets (e.g., β1, β2, β3, 168 β4, and β7) and connecting loops. α4, β5, β6, and α5 form a region named receptor-binding motif (RBM) 169

and it includes most of the residues responsible for receptor binding [16, 19] Whether the SARS-CoV-2's genetic material enters the cytoplasm directly or the whole virus makes use 173 of endosomes for cell penetration, both mechanisms require ACE2 recognition by its RBD domain. 174

Therefore, RBD binders may inhibit viral attachment to the receptor and subsequently perform as a virus 175

inactivator. In the present study, ACE2-derived peptides were computationally designed based on the 176 RBD-binding residues of ACE2 and their potential propensities were predicted using computer-aided 177 approaches. 178

To construct a peptide inhibitor library, the interactions that bind SARS-CoV-2 to the host ACE2 were 179 praised using the experimentally approved PDB structure of their complex. Among the residues shown in 180 Fig. 2 , ACE2 residues which are targeted by the virus were Gln24, Thr27, Phe28, Asp30, Lys31, His34, 181

Glu37, Asp38, Tyr41, Gln42, Leu79, Met82, Tyr83, Asn330, Lys353, Gly354, Asp355, and Arg357. The selection of this region is supported by alanine scanning results that showed the critical role of 189 residues 22-57 for S1 attachment [53] . The template was then docked against spike protein to validate its 190 binding potential. The docking result showed that it chose the RBM motif for binding, as the source PDB, 191

with the binding score of -700.6. The designed peptide, therefore, must have greater negative energy than 192 the template and a higher affinity for RBD than the receptor to effectively prohibit the virus-receptor 193

interaction. 194

The template-receptor complex was put into 100ns of MD simulation to reveal the spots where their 195

interactions Following the mutation-based peptide design, a peptide library of 216 unique sequences was constructed 203

using OSPREY (Table S1 ). Several studies introduced peptide inhibitors against RBD. However, they 204 used different approaches to design peptides [54] [55] [56] [57] [58] [59] [60] , and therefore, their results may seem incomparable 205 to the present study. In the next step, 11 peptides having the greatest OSPREY scores were chosen and 206 their probable allergenicity and toxicity were evaluated. As shown in 

The results showed that except for peptide P181 (-673.1), other designed peptides had greater binding 214 scores compared to template peptides (-761.6, -726.6, and -730 .1 for P7, P13, and P19) ( Table 3 ). This 215

suggests that peptides P7, P13, and P19 had a higher affinity for spike protein, with peptide P7 having the 216 highest affinity. 217

The higher binding potential of P7 may be explained by its highest number of hydrogen bonds (five) and 218 the highest number of spike residues involved in its binding (Table 3) . Among the viral residues, Arg403 219 J o u r n a l P r e -p r o o f was only connected with P7 while others formed interactions with at least two peptides (Fig. 3) . 220

Moreover, it is evident from Fig. 3 (Fig. 3) . RBD was in line with that of template peptide. This suggests that the conformational changes in RBD 232

induced by P7 and P19 were more severe than those of other system (Fig. 4) . results obtained in SASA analysis are shown in figure 4C and as can be seen there is no significant 241 changes in the final values of SASA in all studied systems (Fig. 4C ). This is a sign for the systems that did 242 not undergo high values of opening or compactness which is in agreement with the results of Rg. 243 RMSF analysis also gained information about the volatility of each RBD residue during the simulation. It 244 is evident from Fig. 4D that the greatest amino acid fluctuation is seen in the RBD when it is complexed 245 with P7. This in line with the results obtained from other MD analysis which indicated that this peptide 246

forces the protein to possess higher mobility which can lead to instability in its structure. Definition 247

Secondary Structure of Protein (DSSP) gives information about the frequency of each secondary structure 248 in protein's conformation. DSSP analysis also revealed no significant changes in various types of 249 secondary structures over the span of simulation time (Table 4 ). Principal Component Analysis (PCA) 250

provides the main components of protein motion during the simulation [63] . The 2D diagram of the RBD 251 movements for different systems was obtained using the projection of the first two principal components. 252

As can be seen in Figure 5 , the P7-RBD pattern in more propagated over the diagram plane indicating the 253 higher protein flexibility and movements in this system which is in good agreement with other results. 

The number of hydrogen bonds was also investigated since they are the main interactions that stabilize a 264

complex. The number of H-bonds during 100 ns simulations varied for each peptide. In comparison with 265 other two peptides the results indicated that P7 formed more stable hydrogen bonds with the RBD (Fig.  266  6) . 267

Furthermore, the free non-bonded binding energies (e.g., Van der Waals, electrostatic, polar solvation, 268 and SASA energy) of the final systems were calculated using MM/PBSA method. The binding energies 269 between spike RBD domain and peptides during the whole simulation revealed that P19 had a stronger 270

interaction with the spike however other binding energies are also potent for complex formation (Table  271 4). These results are also in good agreement with that of docking results. 272 273 J o u r n a l P r e -p r o o f 

Finally the resulted RBD-peptide complexes were docked against host receptor to validate their inhibitory 275

activity. As shown in Figure 7 , the RBD domain proved divergence from the position it must take to 276 infect the host cell. As it can be seen, binding the peptides P7 and P13 to the protein resulted in non-277

proper binding of the spike RBD with the ACE2. The amino acids involve in RBD-receptor binding are 278 listed in Table 6 for all complexes. In previous studies the vital role of RBD Lys417 is approved in the 279 virus binding affinity, transmission, and immune escape by mutation analysis [64] . Accordingly, P7 and 280 P13 might be preferable since they successfully limited this residue. Moreover, the energetic assessment 281 of residues lying in the RBD-ACE2 complex indicated the stabilizing impact of Tyr449, Leu455, Phe456, 282

Ala475, Phe486, Glu493, Gly496, Gln498, Thr500, Asn501, Gly502, and Tyr505 in the virus-host 283 complex formation [65] . Also it can be seen that P7 let only 3 RBD residues (Phe456, Asn501, and 284

Tyr505) to access the receptor. However, this number increased to 4 and 6 residues for P13 (Leu455, 285

Phe456, Gln493, and Gly496) and P19 (Tyr449, Leu455, Phe456, Ala475, Gln493, and Gln498). 286

Regarding ACE2, X-ray diffraction experiments demonstrated residues Gln24, Lys31, Tyr41, Gln42, 287

Leu79, Met82, Tyr83, and Lys353 as essential elements for viral recognition [16] . In control docking, it is 288 evident that ACE2 could involve its vital residues the most when bound to P7 with Lys31, Tyr41, Gln42, 289

Tyr83, and Lys353 but fewer host residues were involved due to the inhibition of P13 (ACE2 Lys31, and 290

Gln42) and P19 (ACE2 Lys 31). These data suggest that although P19 and P13 let several important 291 pathogenic RBD residues free, they hindered RBD to accommodate the ACE2 recognition region. In 292 contrast, P7 stifled the RBD-ACE2 complex formation by preventing most of RBD determinant residues. 293 294 Table 6 -the comparison of interactions between ACE2 host receptor and spike RBD domain when the final designed peptides interfere with their contact. The residues involved in H-binding are indicated in bold. In the present study, a library of 216 peptides was designed to investigate their inhibitory potential in 299 complex formation of RBD domain of SARS-Cov-2 and its main receptor the ACE2. At first using the 300 pre docking and MD simulation analysis the interaction of ACE2 derived peptide and the RBD was 301 evaluated. Among all the obtained peptides, based on the binding scores and biological factors such as 302 immunogenicity and stability, number of 3 peptides was selected for the rest of calculations. the results of 303 molecular dynamic simulation indicated that the P3 peptide causes more instability in protein dynamic.

This is extracted from the higher values of RMSD and residue RMSF during the simulation. Also the more 305 sever fluctuation in the Rg and more distribution in 2D PCA diagrams of the RBD confirmed the noted 306

conclusion. Also the results illustrated that in all systems the h-bonding interactions are a part of complex 307 stabilization forces between the RBD and designed peptide. The results of non-bonded interactions 308

showed that the P19 peptide performed the more strong interaction with the RBD however others also 309 possess a tight binding to the protein. Finally the docking results confirmed that among the three studied 310 peptides both the P7 and P13 have more adverse effects on binding of the RBD to its receptor. From all 311

the results obtained in this study it can be concluded that the as designed P7 peptide is capable of being 312

promising blocker of SARS-CoV-2 host cell recognition with high affinity. 313

THE NEW DRUGS of 2019 The 48 medicines represent another highly productive year 316 for the pharmaceutical industry, with cancer and rare-disease drugs again dominating the list

Peptides as therapeutic agents for dengue virus. 320 International journal of medical sciences

Human antimicrobial peptides as therapeutics for viral infections. Viruses

Antiviral peptides as promising therapeutic drugs. Cellular and Molecular 325 Life Sciences

Prioritization of potential vaccine candidates and designing a 327 multiepitope-based subunit vaccine against multidrug-resistant Salmonella Typhi str. CT18: A 328 subtractive proteomics and immunoinformatics approach

A) P7, B) P13, and C) P19) bound to the RBD (gray) domain of the SARS-CoV-2 with the best docking scores. The surface regions indicate RBD position in receptor recognition. 8. Pang

AVPpred: collection and prediction of highly effective 334 antiviral peptides

Analysis and prediction of highly effective antiviral peptides based on 336 random forests

DeepAVP: a dual-channel deep neural network for identifying variable-length 338 antiviral peptides

Putative inhibitors of SARS-CoV-2 main protease from a library of marine 343 natural products: a virtual screening and molecular modeling study. Marine drugs

Structural and functional basis of SARS-CoV-2 entry by using human ACE2. Cell

Structure of the SARS-CoV-2 spike receptor-binding domain bound to the ACE2 350 receptor

Mechanisms of SARS-CoV-2 entry into cells

Cryo-EM structure of the 2019-nCoV spike in the prefusion conformation

Structural basis of receptor recognition by SARS-CoV-2

The role of furin cleavage site in SARS-CoV-2 spike protein-mediated membrane 358 fusion in the presence or absence of trypsin. Signal transduction and targeted therapy

Cathepsin L plays a key role in SARS-CoV-2 infection in humans and 361 humanized mice and is a promising target for new drug development. Signal transduction and 362 targeted therapy

SARS-CoV-2 cell entry depends on ACE2 and TMPRSS2 and is blocked by a 364 clinically proven protease inhibitor. cell

TMPRSS2 and furin are both essential for proteolytic activation of SARS-CoV-2 in 366 human airway cells. Life science alliance

A novel peptide with potent and broad-spectrum antiviral activities against 368 multiple respiratory viruses

Glycopeptide antibiotics potently inhibit cathepsin l in the late 370 endosome/lysosome and block the entry of ebola virus, middle east respiratory syndrome 371 coronavirus (MERS-CoV), and severe acute respiratory syndrome coronavirus (SARS-CoV)

Cross-linking peptide and repurposed drugs inhibit both entry pathways of SARS-374

A broad-spectrum virus-and host-targeting antiviral peptide against SARS-CoV-2 376 and other respiratory viruses

Peptide modelling and screening against human ACE2 and spike glycoprotein 378 RBD of SARS-CoV-2. In silico pharmacology

A natural food preservative peptide nisin can interact with the SARS-CoV-380 2 spike protein receptor human ACE2. Virology

Pharmacophore-based peptide biologics neutralize SARS-384

CoV-2 S1 and deter S1-ACE2 interaction in vitro. bioRxiv, 2020

Computational design of ACE2-based peptide inhibitors of SARS-CoV-2. ACS 386 nano

Atomistic details on the mechanism of organophosphates resistance in insects: 388 insights from homology modeling, docking and molecular dynamic simulation

Cholesterol-lowering drugs the simvastatin and atorvastatin change the 391 protease activity of pepsin: An experimental and computational study

ClusPro PeptiDock: efficient global docking of peptide recognition motifs using 398 FFT

GROMACS: High performance molecular simulations through multi-level 400 parallelism from laptops to supercomputers. SoftwareX

Definition and testing of the GROMOS force-field versions 54A7 and 54B7

Molecular dynamics simulation of hen egg white lysozyme: 404 a test of the GROMOS96 force field against nuclear magnetic resonance data

LINCS: a linear constraint solver for molecular simulations

Particle mesh Ewald: An N⋅ log (N) method for Ewald sums 409 in large systems

g_mmpbsa A GROMACS tool for high-throughput MM-PBSA calculations. 411 Journal of chemical information and modeling

OSPREY 3.0: Open-source protein redesign for you, with powerful new 413 features

In silico approach for predicting toxicity of peptides and proteins

AllerTOP v. 2-a server for in silico prediction of allergens

Structural basis for the recognition of SARS-CoV-2 by full-length human ACE2

Potential inhibitors of the interaction between 421 ACE2 and SARS-CoV-2 (RBD), to develop a drug

A phytochemical-based medication search for the SARS-CoV-2 infection by 425 molecular docking models towards spike glycoproteins and main proteases

An overview of the safety, clinical application and antiviral 431 research of the COVID-19 therapeutics

Identification of critical determinants on ACE2 for 433 SARS-CoV entry and development of a potent entry inhibitor

Computational design of ultrashort peptide inhibitors of the receptor-binding 435 domain of the SARS-CoV-2 S protein

ACE2 Peptide Fragment Interaction with Different S1 Protein Sites. 437 International journal of peptide research and therapeutics

Inhibition of SARS-CoV-2 pseudovirus invasion by ACE2 protecting and Spike 442 neutralizing peptides: An alternative approach to COVID19 prevention and therapy

In silico design of peptides with binding to the receptor binding 445 domain (RBD) of the SARS-CoV-2 and their utility in bio-sensor development for SARS-CoV-2 446 detection

Computational Design of Stapled Peptide Inhibitor against SARS-CoV-2 450 Receptor Binding Domain

Truncated human angiotensin converting enzyme 2; a 452 potential inhibitor of SARS-CoV-2 spike glycoprotein and potent COVID-19 therapeutic agent

A molecular dynamics study on using of naturally occurring polymers for 455 structural stabilization of erythropoietin at high temperature

In silico investigation on the inhibitory effect of fungal secondary 458 metabolites on RNA dependent RNA polymerase of SARS-CoV-II: A docking and molecular 459 dynamic simulation study. Computers in biology and medicine

Effects of common mutations in the SARS-CoV-2 Spike RBD and its ligand, the 461 human ACE2 receptor on binding affinity and kinetics. Elife

Structure, interface stability and hot-spots identification for RBD (SARS-CoV-463 2): hACE2 complex formation