Mutation sites on Omicron strain
The variant Omicron owns 60 mutations compared with the original Wuhan variant (Table 1 and Fig. 1). Among the mutations, 50 non-synonymous mutations, 8 synonymous, 2 non-coding mutations were detected on the 2019-nCov virus and 34 mutations were found to be distributed on the spike region. Interestingly, 3 small deletion mutations and 1 small insertion mutation and 15 single mutation are located in the 2019-nCov-Spike/ACE2 receptor-binding interface domain. It also carries some changes and deletions in other genomic regions. Interestingly, only one mutation was distributed in the Envelope domain. In addition, this variant has 3 mutations on the Membrane site. The ORF1b and Nucleocapsid region also have 11 and 4 mutation sites respectively [19].
Table 1
Distributions of mutation sites on Omicron variant
Region | Mutations |
Spike | A67V, Δ69–70, T95I, G142D, Δ143–145, Δ211, L212I, ins214EPE, G339D, S371L, S373P, S375F, K417N, N440K, G446S, S477N, T478K, E484A, Q493R, G496S, Q498R, N501Y, Y505H, T547K, D614G, H655Y, N679K, P681H, N764K, D796Y, N856K, Q954H, N969K, L981F |
ORF1ab | nsp3 (K38R, V1069I, Δ1265, L1266I, A1892T), nsp4 (T492I), nsp5 (P132H), nsp6 (Δ105–107, A189V), nsp12 (P323L), nsp14 (I42V) |
Envelope | T9I |
Membrane | D3G, Q19E, A63T |
Nucleocapsid | P13L, Δ31–33, R203K, G204R |
The Spike protein play a critical role in identifying and binding the host cell surface receptors, and mediating the fusion of viral envelope to the cell membrane. The spike protein is like a "key" and the ACE2 receptor on the cell is like a "lock". The key is locked for the virus to enter the cell. The main goal of developing the COVID-19 vaccine is also to prevent keys from opening locks to prevent the virus from infecting cells. Thus, learning mutation sites on Omicron spike proteins will supply an extremely important driving role for drug and vaccine development.
Currently, multiple crystal structures of the 2019-nCov-Spike/ACE2 complex have been resolved in the RCSB PDB database. In this paper, structure of the SARS-nCov-2 spike glycoprotein (closed state) with RCSB PDB ID: 6VXX [20] was selected, and excess elements including waters, ions and peptides were deleted for structural studies. Spike protein belongs to the trimer with the top region on each monomer, which was capable of tightly binding to ACE2. All amino acids on each protein between the 2019-nCov-Spike protein and ACE2 within the 15nm distance from each other were set as the binding interface.
As shown in Fig. 1, amino acid mutation sites on the Spike region were uniformly distributed over the Spike region. Among them, the mutation sites which were adjacent to the ACE2 binding interface, is more important for the stability of 2019-nCov/ACE2 complex due to the strong structural interference on the complex system. Here, we selected the 32 amino acid mutation sites on the Spike region for subsequent analyses.
Structure Constructions Of 34 Single-mutants And 1 Multiple-mutant
The 3D structures of the wild-type 2019-nCov/ACE2 complex system was directly extracted from the crystal structure with PDB ID: 6VXX. Subsequently, single-point mutants based on the wild-type Spike protein were constructed by using the mutated wizard module on Pymol software [21]. One total number of 34 single point mutations were obtained.
Binding Energies Of 35 Mutants
Mutation of an amino acid on protein often causes the variation of biological function. We firstly performed primary molecular simulations for 34 single-point mutant and 1 multiple mutant. We used the molecular dynamics software Amber 16 [22] to conduct structural optimization for each 2019-nCov-Spike/ACE2 mutant complex. The whole protein system was parameterized with gaff and AMBER ff99SB force fields. The whole protein complex was geometrical centered with a 10Å plus cubic water box, and electrically neutralized by adding Na + ions. The first step was heating balance process which the whole system was balanced by using the temperature control method of 100ps. The boost balancing process was then balanced for 100ps, and an isotropic Berendsen pressure control method was added. An unrestricted molecular dynamics simulations for free simulation phase was conducted. Temperature and pressure control method is the same as in the previous stage. One 10Å cut-cutoff distance between 2019-nCov and ACE2 protein was used for van der Waals and short-range electrostatic energy calculation, and the long-range electrostatic energy was calculated using the PME method. During molecular dynamics simulations, we force 1500 kcal/mol on all heavy atoms for each protein. Each positional optimization time was at least 6ns per system. Binding energy between the ACE2 protein and the 2019-nCov-Spike mutation system was calculated by using the MMPBSA module based on the 6ns molecular dynamic simulation, as shown in Table 2.
Table 2
Binding energies (Kcal/mol) of 35 mutants 2019-nCov-Spike/ACE2 complex.
Mutations | Energy | Mutations | Energy | Mutations | Energy | Mutations | Energy |
Q498R | -47.26 | G142D | -52.09 | N679K | -56.03 | Δ143–145 | -57.23 |
G446S | -47.98 | N969K | -52.45 | A67V | -56.03 | P681H | -58.25 |
D796Y | -49.06 | T95I | -53.11 | L212I | -56.08 | T478K | -60.09 |
E484A | -50.37 | S477N | -53.21 | T547K | -56.24 | N440K | -60.40 |
N501Y | -50.69 | Q954H | -54.89 | ins214EPE | -56.38 | Q493R | -61.08 |
K417N | -50.82 | Δ211 | -55.10 | N856K | -56.38 | S375F | -61.72 |
Y505H | -51.27 | Wild | -55.18 | H655Y | -56.80 | G339D | -62.21 |
S373P | -51.28 | Δ69–70 | -55.21 | D614G | -57.10 | G496S | -65.31 |
S371L | -51.32 | L981F | -55.77 | N764K | -57.19 | 34-mutations | -83.29 |
6 Mutations Triggered Lower Binding Affinity Of 2019-ncov-spike To Ace2
The binding free energy of wild-type 2019-nCov-Spike/ACE2 protein complex was − 55.18 kcal/mol. Among the 34 single point mutant systems, a total of 16 single point mutations with energy higher than >-55.18 kcal/mol were T95I, G142D, Δ211, S371L, S373P, K417N, G446S, S477N, E484A, Q498R, N501Y, Y505H, D796Y, Q954H and N969K. This result indicated that mot all mutation appeared in Spike region on 2019-nCov virus can lead to increased binding capacity with human ACE2 protein. For 12 mutations, the energy is almost unchanged compared with that of the wild-type complex, indicating that these mutation sites, do not affect much of the binding patterns between the virus and human beings. In addition, 6 complexes (G339D, Q493R, G496S, S375F, N440K and T478K) owns lower that − 60.00 kcal/mol, accounting for 17.14%, indicating that the Omicron COVID-19 mutant have a high frequency that improve the viral infectivity. Notably, for both mutations at G496S and G339D, the binding energies of simulated system were − 65.31 kcal/mol and − 62.21 kcal/mol respectively. Crucially, the multiple mutant system with 32 mutations on the Spike protein has a minimum energy of -83.29 kcal/mol, which is completely consistent with the expected results, because the Omicron virus has an extremely strong infectious capacity.
It can be seen from the Fig. 2 that there being basic rules of the predicted energy distribution in the protein structure. Among them, the mutation sites possessing lower energies were almost close to the 2019-nCov-Spike/ACE2 binding interface. Mutation sites owning similar binding energies with that of wild-type Spike proteins are generally far away from the binding interface. More importantly, the 6 mutation sites with lower energy than − 60.00 kcal/mol were all located on the Spike region, spread from G339 to Y505. Among which, the Q493R and G496S were directly involved in forming binding interface of 2019-nCov-Spike/ACE2 complex. All these results prompted us one deeper insights into the effects of mutations on the entire Spike protein structure and binding mode difference of 6 mutants (<-60.00 kcal/mol) and the multiple mutant.
400ns Md Simulations For 6 Single-mutant And 1 Multiple-mutant
For the selected 6 single-point mutations (T478K, N440K, Q493R, S375F, G339D and G496S) and 34-point mutation system, one 400ns long time molecular dynamics simulations were conducted. Molecular dynamics were carried out by using the software Amber 16. Binding analysis and binding free energies were calculated for protein-protein 2019-nCov-Spike/ACE2 complex after equilibrium phase based on the molecular dynamics trajectories for each system.
The root mean variance (RMSD) represents the dispersion of centroid coordinates means, which can reflect the structural changes of the protein. During the molecular dynamics process, the RMSD trend between initial structure and each time are monitored in real time, and the molecular dynamics simulation is stopped until the RMSD value is stable.
As shown in Fig. 3, RMSD values for all the six single-point mutants exhibited large and drastic fluctuations, indicating the instability of virus to the human receptor ACE2. Each mutation can cause the apparently initial structural changes for 2019-nCov/Spike complex within the starting 1 ~ 30ns time range, and the RMSD vales keeps floating up and down, indicating the conformation changes when the virus binds to the human ACE2 protein. Within the 30 ~ 190ns range, RMSD values for 6 single-point mutations (T478K, N440K, Q493R, S375F, G339D and G496S) maintained smoothly between 10Å and 14Å, indicating that the relatively stable binding modes between Spike and the human ACE2 protein. Notably, compared with 6 single point mutant systems, the multiple-mutant Omicron possessed relatively lower RMSD value. RMSD value for the multiple-mutant with 34-point mutations finally fluctuated from 6Å to 8Å. Relatively lower RMSD value, meaning the smaller structural alteration, also indicates the weaker conformation change of multiple-mutant system itself. All results imply the smaller swing of multiple-mutant system 2019-nCov-Spike/ACE2. It’s interestingly to note that the co-existing 34 mutations for Omicron strain does not trigger larger conformation change, but made itself relatively more stable to bind with human ACE2.
Table 3 Binding free energies (Kcal/mol) between each mutant and ACE2 protein.
System
|
∆EVDWAALS
|
∆Electronic
|
∆EGB
|
∆ESURFACE
|
∆EGAS
|
∆ESOL
|
∆ETOTAL
|
T478K
|
-95.2880
|
-1308.6568
|
1351.4505
|
-13.8726
|
-1403.9448
|
1337.5779
|
-66.3669
|
N440K
|
-90.7503
|
-1536.1082
|
1573.4960
|
-13.4663
|
-1626.8585
|
1560.0297
|
-66.8288
|
Q493R
|
-88.6793
|
-1088.8846
|
1123.6100
|
-14.0337
|
-1177.5639
|
1109.5763
|
-67.9876
|
S375F
|
-89.2105
|
-857.6900
|
893.8887
|
-13.3961
|
-946.9005
|
880.4927
|
-66.4079
|
G339D
|
-87.3446
|
-1143.4603
|
1177.7766
|
-13.7657
|
-1230.8049
|
1164.0109
|
-66.7940
|
G496S
|
-95.4827
|
-770.1586
|
812.1509
|
-13.6003
|
-865.6413
|
798.5506
|
-67.0907
|
34-mutations
|
-98.7460
|
-962.5140
|
990.7028
|
-15.2413
|
-1061.2599
|
975.4614
|
-85.7985
|
VDWAALS = van der waals contribution from MM.
EEL = electrostatic energy as calculated by the MM force field.
EGB = the electrostatic contribution to the solvation free energy calculated by GB respectively.
ECAVITY = nonpolar contribution to the solvation free energy calculated by an empirical model.
DELTA G binding = final estimated binding free energy calculated from the terms above. (kCal/mol).
|
Difference Analysis Of Binding Modes
In order to elucidate the binding patterns of 2019-nCov-Spike to ACE2 protein in different mutants (6 single point mutants and 34-point mutant), we performed a comparative analysis of their structural difference especially the binding interfaces. Assessing the binding surface areas between these possible interfaces indicates that wild-type 2019-nCov-Spike wrap over ACE2 on the lowest level. The predicted interface here was consistent with the crystal structure of 2019-nCov-Spike/ACE2 (PDB ID 6VXX). According to the combined surface area values between 2019-nCov-Spike and human ACE2 protein, the wild-type mutant owns the minimum contacting surface area 206Å3 (ACE2) and 164Å3 (Spike), which was in corresponding with the lowest binding free energies of -55.18 kcal/mol. Whereas for 6 single-point mutants T478K, N440K, Q493R, S375F, G339D and G496S, the contacting surface areas were 179Å3 (ACE2)-237Å3 (Spike), 180Å3 (ACE2)-240Å3 (Spike), 179Å3 (ACE2)-228Å3 (Spike), 179Å3 (ACE2)-226Å3 (Spike), 186Å3 (ACE2)-232Å3 (Spike), 179Å3 (ACE2)-237Å3 (Spike) and 202Å3 (ACE2)-260Å3 (Spike). Interestingly, the six single-point mutants possessed the lower binding energies within − 66.3669~-67.9876 kcal/mol. Residues of the single-point mutant 2019-nCov-Spike protein forms more intensive binding interface than that of wild-type virus. What’s more, among all mutants, the Spike protein of Omicron has the lowest binding capacity to the human ACE2 at -85.7985 kcal/mol. As shown in Fig. 4, we extracted two surface area difference for both Spike and ACE2 the binding patterns when compared to humans.
Binding Modes Difference Between Human Ace2 And Mutants
We next investigated how the mutation sites influence the regional structure of 2019-nCov-Spike to the human ACE2 protein, and analyzed the important residues of 7 mutants which disturb the physicochemical properties of binding interface. The 6 single-point mutants were divided into two categories according to the distance from binding interface: (1) mutation sites not directly form the binding interface (G339D, S375F and N440K); (2) mutation sites formed direct interactions with ACE2 (T478K Q493R and G496S).
As shown in Fig. 5, the mutation sites G339D, S375F and N440K were located far from the binding interface, and weakly changes of mutation resulted in the conformation fluctuation near the mutation site. In mutant G339D, the system possessed relatively stronger binding capacity − 66.7940 kcal/mol than that of wild-type complex with − 55.18 kcal/mol. The binding modes of ACE2 and 2019-nCoV-Spike were shown in Fig. 5A. Typical hydrogen bond interactions including N343-R509 and N343-D339 play significant roles in maintaining structural stability for mutant G339D. Figure 6 mapped the conformation differences at the same position S375/F375 between wild-type and mutant S375D. From Fig. 6, we can see that one weak hydrogen bond interaction was formed between the hydroxyl group of Y508 to S375 on Spike protein. Binding affinities for mutant S375F decrease obviously was − 66.4079 kcal/mol. This obviously conformation variation resulted the intensive contacting between mutant Spike and human ACE2 protein. A relatively higher energy variation (N440K: -66.8288, wild-type: -55.18 kcal/mol) occurred due to the different binding modes for residues near mutation N440K. Binding modes data indicated that atoms on chemical group amin -NH3 from residue K440 make new hydrogen bonds with residue N437, as shown in Fig. 5C.
Similar results were also observed for mutants T478K, Q493R and G496S. These three mutation site are all located at the binding interface, and directly structural perturbations on the protein structure were observed in all 3 mutants. In mutant G496S, amino acid S496 of Spike region formed good hydrogen bonds with the adjacent residues K353 and D38 on ACE2. In wild-type coronavirus, only normal hydrogen bonds were formed between Q493(Spike) and E35-K31(ACE2). However, mutated residue R493 from the mutant Q493R formed stronger salt-bridging interaction with opposite E35 amino acid. The same situations were also detected in mutant T478K, as new polar interactions were formed between the mutated K478 and Q42-S19 on human ACE2 protein. The key point was that the residues K478, R493 and S496 on 34-points’ multiple mutant Omicron strain simultaneously formed the same interactions as shown in single point mutants T478K, Q493R and G496S. Thus, the new mutated residues K478, R493 and S496 formed intensive polar interactions significantly affect the structural stability of the mutant. Thus, we can infer that these three amino acids K478, R493 and S496 are crucial for high viral infection rates for mutant Omicron strain.