Analysis of Effects of Spike Hotspot Residues on Binding Stability of SARS-CoV-2 Variants with ACE2

MA Xue-jing,LI Jun-fu,ZHANG Zhao-ying,BAI Jing,LI Mo

(1.Department of Life Science,Cangzhou Normal University,Cangzhou 061001,Hebei,China;2.Department of Agriculture and Animal Husbandry Engineering,Cangzhou Technical College,Cangzhou 061001,Hebei,China)

Abstract:The spike glycoprotein(S protein)on the surface of the severe acute respiratory syndrome coronavirus 2(SARS-CoV-2)recognizes and binds to the host cell surface receptor angiotensin-converting enzyme 2(ACE2)through its receptor binding domain(RBD),but the effect of mutations in the S protein RBD on its interaction with the receptor is unclear.Through molecular docking and protein binding analysis of S protein and ACE2,this study found that,compared with the wild-type,the strains of both 501Y and B.1.617 lineages displayed more stable receptor-binding ability.Among the variants,501Y.V1 was proved to be the most stable strain.At the RBD-ACE2 binding interface,18% of single amino acid mutations in RBD could enhance the stability of the RBD-ACE2 complex.Knowing about the receptor-binding information of SARS-CoV-2 spike variants would be helpful in designing neutralizing antibodies and vaccines against the viral infection.

Key words:spike glycoprotein;mutation;free energy;hotspot residue;molecular docking

Coronaviruses are large enveloped viruses with a positive-sense single-stranded RNA genome.They can be divided into four genera:α,β,γ,and δ.The severe acute respiratory syndrome coronavirus 2(SARS-CoV-2)belongs to the β-coronavirus genus[1].

The severe acute respiratory syndrome coronavirus(SARS-CoV)uses human angiotensin-converting enzyme 2(ACE2)as a receptor to infect human cells[2].Due to similar sequences and conservative functional motifs,SARS-CoV-2 may have a similar infection route[3].A study showed that,three days after SARS-CoV-2 infection,the viral RNA content reached a peak in the lungs of the human ACE2 transgenic mice while no virus was detected in the wild-type group[4].This indicates that ACE2 is necessary for the viral infection.

SARS-CoV-2 encodes at least 27 proteins,including 15 non-structural proteins(nsp1-nsp10 and nsp12-nsp16),4 structural proteins(S,E,M,and N)and 8 accessory proteins(3a,3b,p6,7a,7b,8b,9b,and orf14)[5].The S protein is a highly glycosylated homotrimer.Analysis of its structure found that the virus uses ACE2 to enter the host cells,and the binding ability of the S protein to ACE2 is significantly higher than that of the SARS-CoV S protein to the same receptor[6～7].When the virus enters a host cell,the homotrimer will undergo structural changes,which can have at least ten different structural states[8].Once the S1 subunit binds to the ACE2 receptor,the structural changes will take place.The S1 subunit falls off,and the S2 subunit forms a highly stable fusion structure[9].Structural analysis of the S protein before and after fusion has made us further understand the mechanism of SARS-CoV-2 infection[10].

The structure of the complex of S protein with the complete ACE2 protein has also been resolved.Compared with SARS-CoV receptor binding domain(RBD),SARS-CoV-2 RBD binds to human ACE2 in a tighter conformation.In addition,several residue changes in the SARS-CoV-2 RBD stabilized the two virus-binding hotspots on the ACE2-RBD interface[11].The overall binding mode of SARS-CoV-2 RBD to ACE2 is almost the same as that of SARSCoV RBD.Most of the residues in SARS-CoV-2 RBD that are critical for ACE2-binding are either highly conserved or have similar side chain properties to the residues in SARS-CoV RBD[12].

Mutations in SARS-CoV-2 S protein would change the viral infectivity.Since late February 2020,the D614G variant has quickly become a main epidemic strain.This mutation could increase the infection ability of SARS-CoV-2 by 9 times.Structural analysis of the mutant S protein showed that D614G actually reduces the binding affinity of the S protein to ACE2,but the mutant S protein has a more open conformation and therefore is easier to bind to ACE2[13].On December 14,2020,the United Kingdom reported to WHO a variant of SARS-CoV-2 named VOC 202012/01(501Y.V1),which proved 70% to 80% more infectious[14].On December 18,2020,a SARS-CoV-2 variant with the N501Y mutation was detected in South Africa,called the 501Y.V2 variant.The variant is related to a higher viral load and may increase infectivity,but there is no evidence that it could cause more serious diseases[15].On January 10,2021,a new SARS-CoV-2 variant from Brazil,501Y.V3,was detected[16～17].Up to now,there are three main lineages containing N501Y:501Y.V1 first discovered in the UK,501Y.V2 first discovered in South Africa,and 501Y.V3 first discovered in Brazil.The B.1.617 lineage,first appeared in India in October,2020,is considered to be one of the reasons for the explosive growth of new cases in India.At present,three subspecies have derived from the B.1.617 strain,namely B.1.617:A,B.1.617:B and B.1.617:C.There is another small cluster of B.1.617 without E484Q mutation.Preliminary analysis showed that the reproduction rates of both B.1.617:A and B.1.617:B are much higher than other mutant strains spreading in India[18].

In order to further study the effects of S protein mutations on SARS-CoV-2 infection ability,nearly 4 000 different mutations that may change the ability of SARS-CoV-2 to bind to human cells were cataloged,revealing how a single mutation may affect the virus behavior[19].Furthermore,multiple RBD mutations have already appeared among SARS-CoV-2 pandemic isolates,but their impacts on receptor recognition remain largely uncharacterized.Comprehensive knowledge of how multiple mutations impact the SARS-CoV-2 RBD would be helpful to understand viral evolution and guide the design of vaccines and other countermeasures.To address this need,we analyzed how amino acid mutations in SARSCoV-2 RBD would affect binding energy and residue interactions between the viral RBD and human ACE2.

1 Materials and methods

1.1 Protein preparation

The model of S protein and ACE2(21～615)was predicted by SWISS MODEL(PDB ID:7CN8.1.A;PDB ID:6M18.1.B)[20]and docked to form a complex by ZDOCK[21].On this basis,the S protein models for 501Y and B.1.617 lineages with ACE2(21～615)were constructed by mutation.The amino acids to be mutated were selected and the mutations were performed by Rotamers in Structure Editing tool,Chimera 1.15[22].The structure of ACE2-RBD complex was obtained from Protein Data Bank(PDB ID:6MJ0)[12].The RBD models for single amino acid mutants were prepared by Chimera 1.15.The most likely structure,generally the first one that its mutated amino acid would not collide with others,was chosen.The model of ACE2-RBD complex for multisite mutant was prepared by Chimera 1.15.

1.2 Molecular docking

ZDOCK was used for molecular docking.The model of S-ACE2(21～615)complex was predicted with binding residues[12]as a restriction.The first prediction was downloaded and used for further analysis.

1.3 Protein binding analysis

Protein binding analysis was conducted on PDBePISA[23].The first complex structure of ZDOCK prediction result was uploaded.The interface area and the free energy change can be viewed in the Interface List.The interface residues can be viewed in Details.The results were visualized by Chimera X 1.2[24].

2 Results

2.1 Effects of S protein mutations in 501Y lineage on ACE2 binding

The S protein sequences of 501Y lineage were compared with that of the original SARS-CoV-2.The distribution of mutational sites over the entire S protein monomer is shown in Fig.1A.RBD consists of a core region and a flexible extension,known as receptor binding motif(RBM,residues 437～508)which directly interacts with the peptidase domain of ACE2[25].The binding free energy change(ΔG)of ACE2 and RBD in 501Y.V1,501Y.V2,and 501Y.V3 has been explored using protein binding analy-sis.The RBD in 501Y lineage displayed increased binding stability to ACE2 in comparison to the wildtype SARS-CoV-2 RBD(Fig.1B).Therefore,the overall impacts of substitutions or deletions in S protein of 501Y lineage are favorable for ACE2 binding.

Fig.1 Mutations in the S proteins of 501Y lineage and their effects on ACE2 binding(A)Schematic diagram of mutations(marked by sticks)in the 501Y lineage S proteins;(B)Scatter plot of the calculated ACE2-RBD interfacial area and the binding free energy for the wild-type and spike variants.Complexes of ACE2-RBD for wild-type SARS-CoV-2 and 501Y lineage were submitted to PDBePISA for binding free energy calculation.The change of binding free energy was calculated as ΔG.The binding stability of 501Y lineage increased,and 501Y.V1 increased the most.

2.2 Binding analysis of RBDs in 501Y lineage to ACE2

The crystal structures of ACE2-RBD complexes were prepared using SWISS-MODEL,ZDOCK and Chimera through a series of steps described in the methodology part.All the crystal structures were subjected to PDBePISA to obtain the interface information of the complexes.The interfaces of complex structures were generated to show the binding region.The interface structures were subjected to H bond analysis application in Chimera X to investigate the interactions occurring between the receptor and RBD(Fig.2).The main aim of this interaction profile study was to elucidate the residues(hotspots)in both the receptor and RBD in their interactions.The residues involved in hydrogen bond formation in both ACE2 receptor and RBD were identified.Compared with the wild-type,all three mutants were found to lose the hydrogen bonds between A386(ACE2)and Y505(RBD),and between T324(ACE2)and V503(RBD).In 501Y.V1,Y83(ACE2)forms an extra hydrogen bond with N487(RBD),and K31(ACE2)only forms hydrogen bond with E484(RBD),not with C488(RBD);in addition,new hydrogen bonds are formed between K353(ACE2)and G502(RBD),and between G326(ACE2)and T500(RBD).501Y.V2 and 501Y.V3 strains have the same hydrogen bonds with ACE2.Unlike 501Y.V1,they lose the hydrogen bond between K31(ACE2)and E484(RBD),and a new hydrogen bond is formed between Y83(ACE2)and F486(RBD),which could result in lower binding stability.Therefore,E484 might be important for ACE2 binding.

Fig.2 Binding analysis of ACE2-RBD complexes for wild-type SARS-CoV-2 and 501Y lineageBinding analysis of ACE2-RBD complexes for wild-type SARS-CoV-2(A),501Y.V1(B),501Y.V2(C)and 501Y.V3(D)was performed.The chain in cyan represents S protein,while the chain in blue represents ACE2.The atoms of interacting residues between ACE2 and RBD are shown.The hydrogen bonds are highlighted by red lines.The interactions were analyzed and visualized by Chimera X.

2.3 Effects of S protein mutations in B.1.617 lineage on ACE2 binding

The S protein sequences of B.1.617 lineage were compared.The distribution of mutational sites over the entire S protein monomer is shown in Fig.3A.The binding free energy change(ΔG)of ACE2 and RBD in B.1.617,B.1.617:A,B.1.617:B and B.1.617:C has been explored using protein binding analysis.Three B.1.617 subspecies and the cluster of B.1.617 without E484Q mutation displayed higher binding stability to ACE2 but smaller interface area in comparison to the wild-type(Fig.3B).Although their binding stabilities showed no significant difference,they all increased compared with that of the wild-type.Therefore,the substitutions in S proteins of B.1.617 subspecies and B.1.617 without E484Q mutation seem favorable for ACE2 binding.

Fig.3 Mutations in the S proteins of B.1.617 lineage and their effects on ACE2 binding(A)Schematic diagram of mutations(marked by sticks)in B.1.617 lineage S proteins;(B)Scatter plot of the calculated ACE2-RBD interfacial area and the binding free energy for spike variants.

2.4 Binding analysis of RBDs in B.1.617 lineage to ACE2

The crystal structures of ACE2-RBD complexes were subjected to PDBePISA to obtain the interface information of the complexes.The binding region and the interactions occurring between the receptor and RBD were explored(Fig.4).Compared with the wild-type,B.1.617 has a new hydrogen bond formed between Y83(ACE2)and F486(RBD),and other hydrogen bonds are the same with those in 501Y.V1.The same hydrogen bonds are formed in three B.1.617 subspecies as in 501Y.V2 and 501Y.V3.Therefore,the only difference between B.1.617 and the other three subspecies is the hydrogen bond between K31(ACE2)and E484(RBD),proving again that the establishment of interaction between K31(ACE2)and E484(RBD)is important for ACE2 binding.

Fig.4 Binding analysis of ACE2-RBD complexes for B.1.617 lineageBinding analysis of ACE2-RBD complexes for B.1.617(A),B.1.617:A(B),B.1.617:B(C)and B.1.617:C(D)was performed.The chain in cyan represents S protein,while the chain in blue represents ACE2.The atoms of interaction residues in ACE2 and RBD are shown.The hydrogen bonds are highlighted by red lines.The interactions were analyzed and visualized by Chimera X.

2.5 Mutational scanning of amino acid mutations in the SARS-CoV-2 RBD and binding analysis of mutated RBD with ACE2

To determine how amino acid mutations in the SARS-CoV-2 RBD impact RBD-ACE2 binding stability,a mutant library of the contact sites in RBD was constructed using a software-based mutagenesis method,through which all 19 mutant amino acids at each position can be introduced.The library covered 288 possible RBD mutants with a single amino acid change.The wild-type RBD was provided as internal standard for the measurements.The binding stability of these RBD variants to the ACE2 has been analyzed by binding free energy calculations.To measure ACE2-binding stability,the ACE2-RBD complexes of the mutants were submitted to PDBe-PISA for calculating free energy,which was represented as ΔG relative to the wild-type SARS-CoV-2 RBD,with positive values indicating stronger binding stability.Fig.5A is a heatmap showing how each mutation affects ACE2-binding affinity.Those consistent with previously published experimental data[19]were selected and analyzed.The results showed that the RBD mutants containing a single amino acid change exhibited a wide range of ACE2-binding abilities,and that the RBD possessed considerable mutational tolerance.Among them,3.6% of the mutants had ACE2-binding stability as high as the wild-type RBD.Approximately 18.0% exhibited higher binding stability than the wild-type,while 78.4% exhibited lower,suggesting that most mutations might be deleterious to protein function.Six mutants,including Y453F,L455M,Q493M,Q498F,N501W,and Y505W,displayed a higher binding stability towards human ACE2 in comparison to the wild-type(Fig.5B).Overall,we obtained binding measurements for all 288 RBD mutants with a single amino acid change.

Fig.5 Mutational scanning to determine how a single amino acid mutation in the SARS-CoV-2 RBD impact ACE2-binding stability and binding analysis of ACE2-RBD complex for the multi-site mutant(A)The heatmap illustrating how single amino acid mutations affect RBD binding stability with ACE2.Squares are colored by mutational effect according to scale bars on the right,with red indicating deleterious mutations.The amino acids of the wild-type RBD are marked with“×”;(B)The selected single mutations of amino acids causing the maximum increase in binding stability.These mutants had the highest binding stability with ACE2;(C)Binding region of ACE2-RBD complex for the multi-site mutant.The chain in cyan represents S protein,while the chain in blue represents ACE2.The atoms of interacting residues in ACE2 and RBD are shown.The hydrogen bonds are highlighted by red lines.The interactions were analyzed and visualized by Chimera X.

The six sites displaying enhanced stability when mutated separately were mutated simultaneously to determine the effects of multiple mutations on ACE2-binding stability.The complex of multi-site mutant RBD and ACE2 was studied by PDBePISA.The amino acid residues involved in the hydrogen bonding with the receptor were identified.The important hot spots identified in ACE2 receptor and S protein RBD are displayed in Fig.5C.Based on the analysis,except for the hydrogen bonds between Y83(ACE2)and N487(RBD),and Q42(ACE2)and Y449(RBD),the other 8 hydrogen bonds were totally different from those of the wild-type.K417,G496,G502 and G446 in RBD are the new residues involved in hydrogen bond formation.

3 Discussion

Considering the reported mutations in SARSCoV-2 and the unclear interactions between the receptor ACE2 and the S protein RBD,this study evaluated the ACE2-binding stability of SARS-CoV-2 spike variants emerged during the pandemic,and screened for potential spike variants with stronger infectious ability.

The receptor-binding stability of RBD is enhanced in strains of the 501Y lineage.The results of single amino acid mutations showed that N501Y substitution could increase ACE2-binding stability.This result seems consistent with the previous in silico investigation of the 501Y.V1 and 501Y.V2,which showed that N501Y replacement should be favorable for the interaction between RBD and ACE2,while the K417N and E484K substitutions seem neutral or even unfavorable[26].Therefore,it could be concluded that N501Y replacement is favorable for ACE2 binding.In order to elucidate the hotspot residues between RBD and ACE2,protein binding analysis was conducted to obtain the interface information of the protein complex.In 501Y lineage,all mutants had a lower ΔG which suggested the mutations should have an overall effect on the enhancement of the binding stability to the receptor.The hydrogen bond between K31(ACE2)and E484(RBD)is important for ACE2 binding.

Using the same method,we investigated the receptor-binding stability and interacting residues of B.1.617 lineage.The RBD structures of B.1.617 subspecies and the cluster of B.1.617 without E484Q mutation displayed higher binding stability.Analysis of hotspot residues showed that B.1.617 subspecies lose the hydrogen bond between K31(ACE2)and E484(RBD)compared with the strain B.1.617,which indicated the difference of binding interface area between B.1.617 and its subspecies might be caused by this specific hydrogen bond.

Next,we explored whether there would be mutants binding to the receptor more strongly than the existing mutants.Unlike the previous study[19],we only selected the residues at the RBD-ACE2 interface for mutation.The RBD variants were constructed by Chimera,and the optimal structures were selected from a variety of possible predicted models.We found that 18% mutants with a single amino acid change strongly increased the receptor binding stability of the RBD,which is highly consistent with the results from previous experimental studies.Hotspot analysis of the multi-site mutant showed that more residues could be involved in the interaction than those of the wild-type.Therefore,RBD variants with multi-site mutations may increase the transmissibility and/or pathogenicity of the virus.

Overall,the mutations in S protein RBD would lead to structural changes and facilitate specific hydrogen bond formation between the ACE2 receptor and S protein RBD.Based on this hotspot amino acid information,blockers could be designed for inhibiting the binding of S protein RBD to human ACE2 receptor and thereby preventing the virus entry into host cells.