Effect on the conformations of the spike protein of SARS‐CoV‐2 due to mutation

Abstract

The spike protein of SARS‐CoV‐2 mediates receptor binding and cell entry and is the key immunogenic target for virus neutralization and the present attention of many vaccine layouts. It exhibits significant conformational flexibility. We study the structural fluctuations of spike protein among the most common mutations that appeared in the variant of concerns (VOC). We report the thermodynamics of conformational changes in mutant spike protein with respect to the wild‐type from the distributions of the dihedral angles obtained from the equilibrium configurations generated via all‐atom molecular dynamics simulations. We find that the mutation causes the increase in distance between the N‐terminal domain and receptor binding domain, leading to an obtuse angle cosine θ distribution in the trimeric structure in spike protein. Thus, an increase in open state is conferred to the more infectious variants of SARS‐CoV‐2. The thermodynamically destabilized and disordered residues of receptor binding motif among the mutant variants of spike protein are proposed to serve as better binding sites for the host factor. We identify a short stretch of region connecting the N‐terminal domain and receptor binding domain forming a linker loop where many residues undergo stabilization in the open state compared to the closed one.

Abbreviations

SARS CoV‐2

Severe acute respiratory syndrome coronavirus 2

COVID‐19

Corona virus disease of 2019

ACE2

Angiotensin converting enzyme 2. It is the key modulator of the renin angiotensin system

NTD

N‐terminal domain

RBD

receptor binding domain

FP

fusion peptide

HR1

heptapeptide repeat sequence 1

HR2

heptapeptide repeat sequence 2

TM

trans membrane domain

WHO

World Health Organization

VoC

Variant of concern

VoI

Variant of interest

MD

Molecular Dynamics

PME

Particle‐mesh Ewald. it is method used for long ranged columbic interactions

LINCS

LINear Constraint Solver for molecular simulations with bond constraints

NVT

constant number of atoms, volume and temperature

NPT

constant number of atoms, pressure and temperature

fs

femto second

ps

pico second

nm

nano meter

HBM

histogram based method

ΔG

Equilibrium conformational changes in free energy

K_B

the Boltzmann's constant

TΔS

conformational entropy change

mt(K417N)‐spike

spike protein consisting K417N mutation

mt(L452R)‐spike

L452R spike mutation

mt(E484K)‐spike

E484K mutation in spike

mt(N501Y)‐spike

N501Y mutation in spike

mt(mult)‐spike

K417N, L452R, E484K and N501Y mutant residues together in spike is considered

mt(del)‐spike

deletion of H69 and V70 in spike

1. INTRODUCTION

Since the first documented cases of SARS‐CoV‐2 ¹ infection in Wuhan, China in late 2019, ² the COVID‐19 pandemic poses an unprecedented threat to global public health, with more than 286 million infections and over 5.4 million deaths around the world (https://www.who.int/emergencies/diseases/novel-coronavirus-2019/situation-reports/). Despite rapid development and emergency authorization of vaccines, immune escape mutants have emerged, and SARS‐CoV‐2 infections remain a concern for the global community. Although vaccination has significantly lowered the rates of hospitalization, severity, and death, ³ , ⁴ , ⁵ , ⁶ current vaccines do not confer absolute prevention of upper‐airway transmission of SARS‐CoV‐2. The number of vaccine breakthrough infections and re‐infections, consequently, have been continuously reported. ⁷ , ⁸ , ⁹ It is of vital importance to examine the impact of the mutant variants as soon as they are detected by genomic sequence analysis. In light of the crucial role of spike protein in virus infection and host immune evasion, studies have been prioritized on the emerging mutations of spike protein circulating SARS‐CoV‐2 strains and investigations on their biological significance. ¹⁰ The structural information is essential for the structure‐based design of vaccine immunogens and entry inhibitors of SARS‐CoV‐2.

Spike protein (180–200 kDa) of SARS‐CoV‐2 virus consists of a homo‐trimeric large clover‐shaped protrusion that mediates viral entry to the host cell through the human ACE2 receptor ¹¹ distributed mainly in the lung, intestine, heart, kidney, and alveolar epithelial type II cells. ¹² Each spike monomer (1273 aa) consists of a signal peptide located at the N‐terminus, the S1 subunit and the S2 subunit. The S1 subunit is responsible for the receptor binding region, which comprises an N‐terminal domain (NTD) and a receptor‐binding domain (RBD). A short stretch of amino acid residues connects the NTD arm with that of the RBD forming the linker. The S2 subunit comprises the fusion peptide (FP), heptapeptide repeat sequence 1 (HR1), heptapeptide repeat sequence 2 (HR2), transmembrane (TM) domain, and cytoplasm domain. The S2 domain entangles to create the stalk, transmembrane, and small intracellular domains. ¹³ Spike protein subsists in a metastable, prefusion conformation acting as an inactive precursor. However, when the virus interacts with the host cell, extensive structural rearrangement of the spike protein occurs, by cleaving it into S1 and S2 subunits, allowing the virus to fuse with the host cell membrane. RBD located in the S1 subunit interacts with the cell receptor ACE2 in the region of aminopeptidase N. Remarkable conformational heterogeneity can be found in the RBD region. Within a single protomer, the RBD could adopt a receptor inaccessible closed “down” state in which the RBD is buried at the interface between the protomers and be accessible to ACE2, or an open “up” state that enables exposure of the receptor‐binding motif that mediates interaction with ACE2. ¹⁴ , ¹⁵

In late 2020, the emergence of variants that posed an increased risk to global public health prompted WHO the characterization of specific variants of interest (VOIs) and variants of concern (VOCs) of SARS‐CoV‐2. Currently, there are five VOCs: alpha variant (B.1.1.7; RBD mutations: N501Y, A570D), beta variant (B.1.351; RBD mutations: K417N, E484K, and N501Y), gamma variant (P.1, B.1.1.28.1; RBD mutations: K417N/T, E484K, and N501Y), delta variant (B.1.617.2; RBD mutations: L452R, T478K), and omicron variant (B.1.1.529; multiple RBD mutations; new VOC designated on November 26, 2021). The mutated variants are characterized by the presence of genetic changes that are known to affect virus characteristics such as transmissibility, disease severity, immune escape, and diagnostic or therapeutic escape. ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ It has been observed that alpha, beta, and gamma variants contain some common mutations, like E484K and N501Y in the receptor binding motif of the RBD region. N501Y and E484K have been reported for the increase in ACE2 binding affinity ²⁴ and decrease in efficacy for antibody binding accountable for immune evasion. ²⁵ , ²⁶ Thus, particularly the RBD variants in SARS‐CoV‐2 are vital to recognize mutant viral strains with higher transmissibility and the potential to bring about immune invasion. Here, we study the effect of mutations in RBD on the conformations of the spike protein of SARS‐CoV‐2.

We are particularly interested in the stability of the mutated protein with respect to the wildtype. The relative stability of protein conformations has been extracted from the mean‐field description based on conformational thermodynamics data. ²⁷ , ²⁸ , ²⁹ , ³⁰ In this method, the changes in thermodynamics free energy and entropy of a protein in a conformation about a reference conformation are estimated from fluctuations of the dihedral angles in the two states over the simulated trajectories. Earlier studies based on conformational thermodynamics suggest that the destabilized and disordered residues of a protein in a particular conformation are the functional ones in that state, leading to binding specificity. ³⁰

We perform all‐atom MD simulations of the complete spike protein of SARS‐CoV‐2 in its wildtype and the several mutated VOC strains, such as K417N, L452R, E484K, and N501Y, using the GROMOS96 53a6 force‐field in the GROMACS 2018.6 package. We designate various proteins as follows: spike protein consisting of K417N mutation designated as mt(K417N)‐spike, L452R spike mutation as mt(L452R)‐spike, E484K mutation in spike as mt(E484K)‐spike, N501Y mutation as mt(N501Y)‐spike, mt(mult)‐spike include K417N, L452R, E484K, and N501Y mutant residues, and mt(del)‐spike consists of deletion of H69 and V70. We find that the spike protein prefers to be in an open state under mutations in the RBD, while a close conformation is preferred in the wild‐type. We calculate the thermodynamics cost of conformational changes associated with various spike protein mutations. The RDB residues in mt(K417N)‐spike, mt(L452R)‐spike, mt(E484K)‐spike, mt(N501Y)‐spike, and mt(mult)‐spike show destabilization and disorder in comparison to the wildtype conformation. However, the mt(del)‐spike shows order and stabilization of the RBD residues in comparison to the wildtype spike protein. On the other hand, the NTD region does not reveal much significant changes in stability and order. The linker loop reveals an increase in order and stability in the mutated variants. Our studies may shed important light on the hostile nature of the different SARS‐CoV‐2 variants.

2. MATERIALS AND METHODS

2.1. System preparation and simulation details

The cryo‐EM structure of SARS‐CoV‐2 spike trimer in a tightly closed state (7DF3) as well as in its open state (7DK3) are considered to study the conformational dynamics of spike protein and its effect on mutation. The mutations are obtained from the cryo‐EM structure. We have considered the most common mutations that appeared in VOC strains like K417N, L452R, E484K, and N501Y. We have also chosen a system that contains multiple mutations together designated as “mt(mult)‐spike” and a separate system that includes only H69del and V70del without any other spike mutations as “mt(del)‐spike.”

We perform 1 μs long all‐atom MD simulation using the standard protocol for isothermal isobaric ensemble (NPT) with 310 K and 1‐atm pressure in GROMACS ³¹ package. We use periodic boundary conditions, spc216 water model, and GROMOS9353a6 ³² force field for simulations in GROMACS 2018.6 package. Electro‐neutrality is maintained by adding mono‐valent ions Na⁺ and Cl^–. Long‐ranged columbic interactions are considered using the PME approach. ³³ LINCS algorithm ³⁴ is used to constraint the bonds, and leap‐frog integration is used to perform simulation. Minimization is done for 50,000 steps using the steepest descent algorithms. Equations of motion are integrated using a leap‐frog algorithm with an integration time step of 2 fs. Systems are equilibrated through two steps (NVT and NPT) using position restraints to heavy atoms. NVT and NPT equilibration is carried out at 300 K temperature and 1 bar pressure. We maintain the total number of particles (N = 166,251), pressure, and temperature same for all the systems to make the simulated ensembles equivalent.

Seven independent systems, including wildtype spike protein, designated as wt‐spike, are considered for MD simulations for 1 μs with different initial conditions: (i) open conformation and (ii) closed conformation. All runs are repeated three times with different speeds. The equilibrations of the simulated structures are assessed from the saturation of the root mean squared deviations (RMSD). All the data have been averaged over six independent trajectories for each system.

2.2. Conformational thermodynamics

Conformational thermodynamics changes for spike proteins and their mutated varieties at different conformations are estimated properly from equilibrium fluctuations of the dihedral angles using the histogram‐based method (HBM). ³⁰ Equilibrium conformational changes in free energy are defined by $\mathrm{\Delta }{G}_{(\zeta )}=-K_{\mathrm{B}}T\ln [\frac{H^{\mathrm{mt}}(\zeta )}{H^{\mathrm{wt}}(\zeta )}]$, where $H^{\mathrm{mt}}(\zeta )$ and $H^{\mathrm{wt}}(\zeta )$ signify peak value of normalized probability distribution of protein dihedral (ζ) of mutated and wildtype spike protein, respectively, and K _B, the Boltzmann's constant. Conformational entropy change associated with a particular dihedral (ζ) at a temperature T is calculated using$T\mathrm{\Delta }S=T(S^{\mathrm{mt}}(\zeta )-S^{\mathrm{wt}}(\zeta ))$, where $S^{\mathrm{mt}}(\zeta )$ and $S^{\mathrm{wt}}(\zeta )$ can be obtained using Gibbs entropy formula $S(\zeta )=-K_{\mathrm{B}}T{\sum }_i{H}_i(\zeta )\ln H_i(\zeta )$; the sum is taken over histogram bins.

3. RESULTS

3.1. Effect of mutation on spike protein conformation

In the current study, we have considered the S1 subunit that interacts with the host cell receptor ACE2, consisting of the NTD (14–306 residues), the RBD (331–528 residues), and the linker (306–331). The crystal structure RBD of spike protein exists in both closed (Figure 1A) and open (Figure 1B) conformations. We show equilibrium snapshots for different cases in Figure 1. We have found that the wt‐spike protein attains a closed conformation in all cases (Figure 1C). However, in all the mutated systems, mt(K417N)‐spike, mt(L452R)‐spike, mt(E484K)‐spike, mt(N501Y)‐spike, and mt(mult)‐spike, the open conformation is observed (Figure 1D–H). On the other hand, we have observed that the mt(del)‐spike protein prefers close conformation (Figure 1I).

FIGURE 1.

Color online: Structure of trimeric spike protein of SARS‐CoV‐2 in its wildtype and mutated form at different conformation states. NTD is shown in blue, RBD in pink, and linker arm in brown. (A) Spike protein at its closed state (PDB id: 7DF3). (B) Open state of spike protein (PDB id: 7DY3). Snapshot at 1 μs time span of (C) wt‐spike, (D) mt(K417N)‐spike, (E) mt(L452R)‐spike, (F) mt(E484K)‐spike, (G) mt(N501Y)‐spike, (H) mt(mult)‐spike, and (I) mt(del)‐spike

To quantify the conformational changes, we consider the centers of mass for NTD and RBD regions, respectively, from the equilibrated trajectory of each of the systems and then the distance S between the centers of mass of NTD and the RBD arm has been calculated over simulated trajectories. We show the distribution of S, H(S) over the equilibrium trajectories for all the systems in Figure 2A. For the wt‐spike protein, H(S) shows a sharp peak at around 3 nm and mt(del)‐spike shows approximately around 3.5 nm. This distance is comparable to the crystal structure data. However, the rest of the mutant variants show a peak in the range of 4.5–6 nm. Thus, the distance between the NTD and RBD of the spike protein increases due to mutation, so the RBD prefers the open conformation to the closed one. We further consider the cosine of the angle θ between the vectors joining the center of mass of the NTD arm, those of the linker, and the RBD arm. We show the distribution of cosθ, H(cosθ) over the equilibrium trajectories for all the systems in Figure 2B. Both wt‐spike and mt(del)‐spike show a peak around θ= 90°. In mt(K417N)‐spike, there is a peak at θ= 110°, mt(N501Y)‐spike confers a peak at θ= 100°, and mt(E484K)‐spike shows a peak at θ= 105°. mt(L452R)‐spike shows a shift in peak for wt‐spike at θ= 180°, while in the mt(mult)‐spike maximum peak can be found at θ= 120°. These data also support open conformation in mutant variants of SARS‐CoV‐2.

FIGURE 2.

Color online: (A) Histogram distribution of the distance between the $C_{\alpha }$ atoms between the two center of mass of NTD arm to that of RBD arm calculated over time span. S, H(S) over the equilibrium trajectories for all the systems are represented by wt‐spike in black, mt(del)‐spike in grey, mt(K417N)‐spike in yellow, mt(L452R)‐spike in blue, mt(E484K)‐spike in green, mt(N501Y)‐spike in red, and mt(mult)‐spike in violet. (B) The distribution of cos θ, H(cos θ) over the equilibrium trajectories for all the systems (color demarcation are same as (A)). Mutation causes the RBD arm to be more distantly apart from NTD arm, causing an obtuse angle cosine θ distribution in the trimeric structure in spike protein.

3.2. Conformational thermodynamics due to mutation

The flexibility of the protein conformations is given in terms of the dihedral fluctuations. The dihedral angles φ, ψ of the backbone and χ ₁ of the side chain for a different region of the spike proteins in wt‐spike and variant systems are computed from the equilibrated portion of the trajectories. We denote the distribution of a dihedral angle θ (either of φ, ψ, and χ ₁) of the i‐th residue in wt‐spike by $H_i^{\mathrm{wt}}(\theta ),$and variant systems by $H_i^{\mathrm{K}417\mathrm{N}}(\theta )$, $H_i^{\mathrm{L}452\mathrm{R}}(\theta )$, $H_i^{\mathrm{E}484\mathrm{K}}(\theta )$, $H_i^{\mathrm{N}501\mathrm{Y}}(\theta )$, $H_i^{\mathrm{mult}}(\theta )$, and $H_i^{\mathrm{del}}(\theta )$, depending on the kind of modification over the wt‐spike protein. A broadened or multiple‐peaked distribution indicates enhanced flexibility in the given dihedral.

Let us first consider the critical residues of the RBD region playing a vital role in ACE2 interaction. We observe that $H_{\mathrm{N}487}^{\mathrm{wt}}(\phi )$ and $H_{\mathrm{N}487}^{\mathrm{del}}(\phi )$ exhibit a sharp peak (Figure 3A), while in the rest of the mutated system, an increase in flexibility can be found in this degree of freedom due to mutation. The same trend can be found in the backbone dihedral distribution (ψ) of this residue. Sharp unimodal distributions have been observed in $H_{\mathrm{N}487}^{\mathrm{wt}}(\psi )$ and $H_{\mathrm{N}487}^{\mathrm{del}}(\psi )$, while $H_{\mathrm{N}487}^{\mathrm{K}417\mathrm{N}}(\psi )$, $H_{\mathrm{N}487}^{\mathrm{L}452\mathrm{R}}(\psi )$, $H_{\mathrm{N}487}^{\mathrm{E}484\mathrm{K}}(\psi )$, and $H_{\mathrm{N}487}^{\mathrm{mult}}(\psi )$ are rather flat (Figure 3B). In N487, a huge increase in flexibility is noticed due to the side chain (χ ₁). The sharp peak of $H_{\mathrm{N}487}^{\mathrm{wt}}({}_1)$ changes into multimodal peaks due to mutations (Figure 3C). $H_{\mathrm{Y}505}^{\mathrm{wt}}(\phi )$ and $H_{\mathrm{Y}505}^{\mathrm{del}}(\phi )$ have unimodal distribution, while the mutant variants show relatively flatter distribution depicting enhanced flexibility at that region (Figure 3D). In Y505, $H_{\mathrm{Y}505}^{\mathrm{wt}}(\psi )$ shows a sharp peak, while in the rest of the mutated system a multimodal distribution is observed, showing an increase in flexibility in this region (Figure 3E). The side chain dihedral (₁) does not show much difference due to mutation (Figure 3F). The cases of the dihedral angles of the other residues interacting with the host factor of the receptor binding motif are shown in Figures S1–S6. Overall, the distribution of dihedral angles shows an increase in flexibility in the RBD region due to spike protein mutation.

FIGURE 3.

Color online: Histogram distribution of dihedral angles of wildtype and variant system at RBD. ACE2 binding motif indicating maximum perturbation is shown here. The color definitions are same as Figure 2. (A) $H_{\mathrm{N}487}^{\mathrm{wt}}(\phi )$ and $H_{\mathrm{N}487}^{\mathrm{del}}(\phi )$ exhibit a sharp peak than the rest of the mutated system, and increase in flexibility can be found at this degree of freedom due to mutation. (B) Sharp unimodal distribution has been observed in $H_{\mathrm{N}487}^{\mathrm{wt}}(\psi )$ and $H_{\mathrm{N}487}^{\mathrm{del}}(\psi )$. Increase in flexibility is observed due to the flat curves like in $H_{\mathrm{N}487}^{\mathrm{K}417\mathrm{N}}(\psi )$, $H_{\mathrm{N}487}^{\mathrm{L}452\mathrm{R}}(\psi )$, $H_{\mathrm{N}487}^{\mathrm{E}484\mathrm{K}}(\psi )$, and $H_{\mathrm{N}487}^{\mathrm{mult}}(\psi )$. (C) Enhance in flexibility can be noticed due to the side chain (χ ₁). The sharp peak of $H_{\mathrm{N}487}^{\mathrm{wt}}({}_1)$ changes into multimodal flattened peak at this degree of freedom due to mutation. (D) $H_{\mathrm{Y}505}^{\mathrm{wt}}(\phi )$ and $H_{\mathrm{Y}505}^{\mathrm{del}}(\phi )$ elicit a sharp unimodal distribution, and the mutant variants represent flat curved distribution conferring increase in flexibility at this region. (E) $H_{\mathrm{Y}505}^{\mathrm{wt}}(\psi )$ shows sharp peak, while rest of the mutated system elicits a multimodal distribution responsible for increase in flexibility. (F) Y505 (₁) does not show much remarkable difference due to mutation.

Let us now consider the dihedral angle distributions of certain residues in the NTD domain. It is noticed that the dihedral distribution φ of Y170 is similar for all the systems (Figure 4A). $H_{\mathrm{Y}170}^{\mathrm{wt}}(\psi )$ shows a sharper peak than the other systems, which depicts enhanced flexibility due to mutation (Figure 4B). $H_{\mathrm{Y}170}^{\mathrm{wt}}({}_1)$, $H_{\mathrm{Y}170}^{\mathrm{del}}({}_1)$, and $H_{\mathrm{Y}170}^{\mathrm{E}484\mathrm{K}}({}_1)$ are sharp unimodal, whereas $H_{\mathrm{Y}170}^{\mathrm{K}417\mathrm{N}}({}_1),H_{\mathrm{Y}170}^{\mathrm{L}452\mathrm{R}}({}_1),H_{\mathrm{Y}170}^{\mathrm{N}501\mathrm{Y}}({}_1)$, and $H_{\mathrm{Y}170}^{\mathrm{mult}}({}_1)$ show bimodal distribution (Figure 4C), suggesting an increase in flexibility due to such mutations. The cases of the dihedral angles of certain other residues from the NTD domain are shown in Figure S7. It is found that the impact of spike mutations is not so significant for the NTD of the S1 subunit.

FIGURE 4.

Color online: The dihedral distribution at NTD domain (A) (φ)distribution in Y170 is almost uniform for all the systems. (B) $H_{\mathrm{Y}170}^{\mathrm{wt}}(\psi )$ shows a sharp peak than the rest of the system, which depicts elevation of flexibility due to mutation in rest of the system. (C) $H_{\mathrm{Y}170}^{\mathrm{wt}}({}_1)$, $H_{\mathrm{Y}170}^{\mathrm{del}}({}_1)$, and $H_{\mathrm{Y}170}^{\mathrm{E}484\mathrm{K}}({}_1)$ elicit sharp unimodal peak, whereas $H_{\mathrm{Y}170}^{\mathrm{K}417\mathrm{N}}({}_1),H_{\mathrm{Y}170}^{\mathrm{L}452\mathrm{R}}({}_1),H_{\mathrm{Y}170}^{\mathrm{N}501\mathrm{Y}}({}_1)$, and $H_{\mathrm{Y}170}^{\mathrm{mult}}({}_1)$ show bimodal distribution. (D) $H_{\mathrm{A}263}^{\mathrm{E}484\mathrm{K}}(\phi )$ is bimodal, whereas the rest of the system show almost uniform unimodal distribution. (E) $H_{\mathrm{A}263}^{\mathrm{wt}}(\psi )$ is sharper than the rest of the system.

Now we consider residues from the linker loop. It is observed that $H_{\mathrm{K}310}^{\mathrm{K}417\mathrm{N}}(\phi ),H_{\mathrm{K}310}^{\mathrm{L}452\mathrm{R}}(\phi ),H_{\mathrm{K}310}^{\mathrm{N}501\mathrm{Y}}(\phi )$, and $H_{\mathrm{K}310}^{\mathrm{mult}}(\phi )$ have sharp unimodal peaks, while $H_{\mathrm{K}310}^{\mathrm{wt}}(\phi )\mathrm{and}{H}_{\mathrm{K}310}^{\mathrm{del}}(\phi )$ show an increase in flexibility (Figure 5A). Thus, the mutation causes a decrease in flexibility in this degree of freedom. Similarly, for ψ of K310, a decrease in flexibility is observed particularly in $H_{\mathrm{K}310}^{\mathrm{K}417\mathrm{N}}(\psi ),H_{\mathrm{K}310}^{\mathrm{L}452\mathrm{R}}(\psi )$, and $H_{\mathrm{K}310}^{\mathrm{N}501\mathrm{Y}}(\psi )$ (Figure 5B). The side chain dihedral (₁) of K310 shows more flexibility in $H_{\mathrm{K}310}^{\mathrm{del}}({}_1)$ and $H_{\mathrm{K}310}^{\mathrm{wt}}({}_1)$ than $H_{\mathrm{K}310}^{\mathrm{del}}({}_1)$ and $H_{\mathrm{K}310}^{\mathrm{wt}}({}_1)$. (Figure 5C). $H_{\mathrm{F}329}^{\mathrm{K}417\mathrm{N}}(\phi ),H_{\mathrm{F}329}^{\mathrm{E}484\mathrm{K}}(\phi )$, and $H_{\mathrm{F}329}^{\mathrm{mult}}(\phi )$ exhibit a sharp tall unimodal peak suggesting that the flexibility is lost after mutation (Figure 3D). The sharp peak of $H_{\mathrm{F}329}^{\mathrm{wt}}(\psi )$ depicts a decrease in flexibility due to such mutation, whereas in the rest of the system the increase in flexibility is prominent (Figure 3E). In the side chain dihedral (₁) of F329, the unimodal acute peak in $H_{\mathrm{F}329}^{\mathrm{K}417\mathrm{N}}({}_1)$ becomes compressed and bimodal distribution in different other systems, illustrating expansion of flexibility in this region (Figure 3F). Dihedral distributions of the rest of the residues from the region are shown in Figures S8–S10.

FIGURE 5.

Color online: Histogram distribution of dihedral angles of wildtype and variant system at linker loop. The color representations are followed from Figure 2. (A) Increase in flexibility is observed in $H_{\mathrm{K}310}^{\mathrm{del}}(\psi )$. (B) Decrease in flexibility can be observed in $H_{\mathrm{K}310}^{\mathrm{K}417\mathrm{N}}(\psi ),H_{\mathrm{K}310}^{\mathrm{L}452\mathrm{R}}(\psi )$, and $H_{\mathrm{K}310}^{\mathrm{N}501\mathrm{Y}}(\psi )$. (C) Rigidity in the dihedral angle distribution due to side chain (₁) is conferred in most of the mutant variants. (D) $H_{\mathrm{F}329}^{\mathrm{wt}}(\phi )$ and $H_{\mathrm{F}329}^{\mathrm{del}}(\phi )$ exhibit a bell‐shaped flattened peak than the tall sharper peak of the mutated systems. (E) The sharp peak of $H_{\mathrm{F}329}^{\mathrm{K}417\mathrm{Y}}(\psi )$ illustrates the decrease in flexibility over $H_{\mathrm{F}329}^{\mathrm{wt}}(\psi ).$ Distribution in rest of the system is almost uniform. (F) In the side chain dihedral (₁) of F329, the unimodal acute peak in $H_{\mathrm{F}329}^{\mathrm{Y}417\mathrm{N}}({}_1)$ depicts maximum rigidity due to this degree of freedom.

We account for changes in free energy △G_i ^conf and entropy T△S_i ^conf of conformational changes of the mutated systems with respect to wt‐spike from the distributions of the dihedral angles. Positive values of the changes in free energy and entropy indicate destabilization and disorder of the mutated system in comparison to the wt‐spike, while the negative values indicate stabilization and order in the mutated system. The overall changes in conformational thermodynamics of the various domains of spike protein are obtained by adding all the dihedral contributions from the residues of the particular region. We observe that (Table 1) NTD becomes energetically destabilized and disordered in the mutant variants than in the wt‐spike system. However, NTD from the mt(del)‐spike remains stabilized and more ordered with respect to the wt‐spike. The instability and disorder in the NTD arm are primarily dominated by backbone fluctuations (Table 1). In mt(del)‐spike, the side chain dihedral imparts maximum stability and order. The RBD residues of the spike protein undergo disorder and destabilization in the mutated system compared to the wt‐spike system (Table 2). We observe that the mt(del)‐spike system remains energetically stabilized and ordered with reference to the wt‐spike system. This trend is similar to that found in the NTD arm of the different systems of the spike protein. The major contributions to free energy and entropy changes of the RBD region come from backbone dihedrals. It has been found that the overall entropy and free energy changes of the linker loop remain marginal (Table 3), although the residues show more order and stabilization in the mutated case. This suggests that the linker residues play a role like a hinge to control the opening between RBD and NTD. It may be noted that in the mt(del)‐spike system, the linker shows disorder and destabilization where no hinge role is needed.

TABLE 1.

Comparative changes in the conformational thermodynamics (kJ/mol) of the N‐terminal region of spike protein in the different mutated systems to wildtype

	T△Si conf				△Gi conf
System	ϕ	ψ	χ 1	Total	ϕ	ψ	χ 1	Total
mt(K417N)‐spike	22.21	29.35	37.82	89.38	7.86	8.2	13.04	29.1
mt(L452R)‐spike	13.67	29.06	19.87	62.6	9.87	8.54	5.56	7.99
mt(E484K)‐spike	22.11	22.18	3.56	47.85	6.47	13.93	−1.12	19.28
mt(N501Y)‐spike	4.37	4.22	13.33	21.92	7.66	5.74	6.56	19.96
mt(mult)‐spike	24.07	28.63	19.66	72.36	21.31	10.71	−4.98	27.04
mt(del)‐spike	1.48	10.53	−15.44	−3.43	10.28	−1.96	−25.02	−16.7

TABLE 2.

Comparative changes in the conformational thermodynamics (kJ/mol) of the receptor binding domain of spike protein in the different mutated systems to wildtype

	T△Si conf				△Gi conf
System	ϕ	ψ	χ 1	Total	ϕ	ψ	χ 1	Total
mt(K417N)‐spike	4.3	4.58	11.64	20.52	14.03	5.45	5.79	25.27
mt(L452R)‐spike	2.92	−4.57	3.83	2.18	8.8	−0.81	0.35	8.34
mt(E484K)‐spike	23.22	19.81	21.62	64.65	0.8	9.95	13.8	24.55
mt(N501Y)‐spike	19.02	20.73	19.34	59.09	3.59	6.57	8.2	18.36
mt(mult)‐spike	1.62	11.7	11.47	24.79	10.44	4.93	3.42	18.79
mt(del)‐spike	−25.3	−28.09	−52.46	−105.85	−16.12	−1.61	−8.36	−26.09

TABLE 3.

Comparative changes in the conformational thermodynamics (kJ/mol) of the linker loop of spike protein in the different mutated systems to wildtype

	T△Si conf				△Gi conf
System	ϕ	ψ	χ 1	Total	ϕ	ψ	χ 1	Total
mt(K417N)‐spike	0.6	−0.64	−0.41	−0.45	0.36	−0.28	−0.2	−0.12
mt(L452R)‐spike	−0.52	‐3.65	−0.17	‐4.34	0.39	−0.56	0.32	0.15
mt(E484K)‐spike	−1.65	−2.82	−0.86	−5.33	−2.69	−0.88	3.11	−0.46
mt(N501Y)‐spike	−0.52	−0.17	0.17	−0.52	−0.05	−0.12	−0.2	−0.37
mt(mult)‐spike	−0.93	−0.17	0.32	−0.78	−0.83	−0.15	−0.03	−0.65
mt(del)‐spike	1.43	1.28	2.89	5.6	0.32	0.37	0.42	1.11

Both the residue‐wise and the domain‐wise free energy and entropy costs for conformational changes in the mutated protein with respect to the wt‐type protein in the RBD and the linker regions are shown in Tables S1–S5. The overall change in entropy in the residues of RBD of mt(K417N)‐spike is 5.62 kJ/mol, where the major changes are in the backbone dihedrals. The total change in free energy is 7.28 kJ/mol to which the backbone dihedrals contribute the most (Table S1). In case of another point mutation mt(L452R)‐spike, the total change in entropy of the region is 11.86 kJ/mol, where the backbone dihedrals together account for most of the changes (Table S2). The changes in the stability are marginal in all cases except Y449 having the highest change in stability. The total change in free energy is 4.8 kJ/mol (Table S2). Altogether change in entropy of mt(E484K)‐spike is 1.63 kJ/mol, where the backbone dihedrals are the primary contributors. The backbone dihedral (φ) is the key factor for changes in free energy of mt(E484K)‐spike in this region, the overall change of free energy being 1.66 kJ/mol (Table S3). The total changes in entropy of the region for mt(N501Y)‐spike is 8.59 kJ/mol, which is exhibited by the total changes in the backbone dihedrals (Table S4). The change in free energy is 5.35 kJ/mol, where (φ) backbone dihedral acts as pivotal for the changes (Table S4). The backbone dihedrals contribute to major changes in the region, the total changes in entropy and free energy in mt(mult)‐spike being 9.09 and 6.71 kJ/mol, respectively (Table S5). The overall changes in entropy and free energy in the mt(del)‐spike are −10.55 and −7.03 kJ/mol, where the backbone dihedrals are the key factors for such changes in the system (Table S6). The details of conformational thermodynamics data for certain residues of the NTD domain are given in Tables S7–S12. The dihedral fluctuations show conformational free energy destabilization. Similar increases are observed in conformational entropy. The maximum destabilization and disorder are through the backbone fluctuations. The mt(del)‐spike shows similar thermodynamic changes as the other cases in NTD. The thermodynamic data for certain linker residues responsible for critical conformation changes are shown in Tables S13–S18. The overall changes in the entropy and free energy in mt(K417N)‐spike are −9 and −4.34 kJ/mol, the maximum being for F329 (Table S13). The backbone dihedral distribution accounts for maximum changes in this system. In mt(L452R)‐spike, the entropy change and free energy change are −9.85 and −4.99 kJ/mol, respectively (Table S14). Such ordering and stability in free energy are governed by backbone dihedral fluctuations (Table S14). The total changes in entropy for K310, G311, Y313, F329, and P330 in mt(E484K)‐spike is −6.81 kJ/mol, primarily due to the backbone dihedral distributions. Besides, the total change in free energy is −3.23 kJ/mol due to the backbone fluctuations. We find that P330 imparts a slight increase in entropy (Table S15). In mt(N501Y)‐spike, −6.92 kJ/mol is the total change in entropy and the total change in free energy is −4.7 kJ/mol (Table S16). However, the changes in entropy and free energy found due to such residues of linker loop in mt(mult)‐spike are marginal, −0.53 and −1.87 kJ/mol, respectively (Table S17). The mt(del)‐spike system shows that the total change in entropy is 14.24 kJ/mol and the change in free energy is 3.47 kJ/mol (Table S18).

We map the changes in conformational free energy △G_i ^conf and entropy T△S_i ^conf of individual residues of the mutated systems with respect to wt‐spike. Here, we show free energetically stabilized and ordered residues in green and the destabilized and disordered ones in red. Careful examination of the ACE2 interacting residues of RBD of mt(K417N)‐spike shows that G446, Y449, N487, Y489, T500, and Y505 impart enhanced disorder, the maximum being in Y489. An increase in order is observed in Q493 and G502 (Figure 6A). It is found that Y449 and Y505 confer a major decrease in stability, whereas in the rest of them the free energy change is marginal. In case of another point mutation mt(L452R)‐spike, most of the interface residues that form crucial interaction with host factor ACE2 get disordered (Figure 6B). G446 and Y449 undergo a maximum decrease in order. Q493 shows a marginal increase in order. Y449 shows maximum destabilization in free energy of the region of mt(L452R)‐spike. G446, N487, Y489, Q493, T500, and Y505 of mt(L452R)‐spike show a slight decrease in stability due to free energy change. G502 of mt(L452R)‐spike account for a minor increase in stability. The RBD residues of the spike protein of mt(E484K)‐spike show enhanced disorder in G446, N487, Y489, and T500, Y489 having the maximum disorder. On the other hand, Q493, G502, and Y505 exhibit an increase in order (Figure 6C). It has been found that Y449, N487, Y489, and T500 are responsible for the decrease in stability in mt(E484K)‐spike, the highest destabilized residue being Y449. On the other hand, G446, Q493, G502, and Y505 are responsible for the marginal increase in stability in mt(E484K)‐spike. Almost all the residues of the region in mt(N501Y)‐spike protein show an increase in order, except Q493 and G502 (Figure 6D). Y505 is responsible for the maximum disorder. The changes in the stability of these residues are marginal for the mutation N501Y except Y449, which imparts a maximum increase in destabilization (Figure 6D). We find enhanced stability and order in Q493 and G502 in mt(N501Y)‐spike. In mt(mult)‐spike, Y449, N487, Y489, G502, and Y505 show an increase in disorder and disability, the maximum disorder being in Y489. Y449 imparts the largest increase in destabilization (Figure 6E). G446 and Q493 show only a marginal increase in order and stability. All the RBD residues interacting with ACE2 in the mt(del)‐spike remain ordered and stabilized, except N487. Y489 shows a maximum increase in order and Y449 imparts a maximum change in stability (Figure 6F).

FIGURE 6.

Color online: Illustration of the conformational thermodynamic changes of RBD domain on the average structure of wildtype and mutant variant of spike protein. The destabilized and disordered residues are marked in red (ball‐like model). The residues stabilized and ordered are shown in green (ball‐like model). The crucial residues forming vital interaction with host factor are only highlighted: (A) mt(K417N)‐spike, (B) mt(L452R)‐spike, (C) mt(E484K)‐spike, (D) mt(N501Y)‐spike, (E) mt(mult)‐spike, and (F) mt(del)‐spike.

Let us now consider the case of linker residues. In mt(K417N)‐spike, all the residues are ordered and stabilized (Figure 7A). In mt(L452R)‐spike, K310, G311, Y313, F329, and P330 are ordered as well as stabilized (Figure 7B). K310 imparts maximum increase in order and stability. Figure 7C illustrates the five linker loop residues of mt(E484K)‐spike, which undergo ordering and stabilization compared to wt‐spike, where K310 accounts for the largest change in entropy. The same five residues from the linker loop in mt(N501Y)‐spike remain ordered and stabilized with respect to wt‐spike (Figure 7D), where G311 of mt(N501Y)‐spike accounts for maximum changes. We observe that the dihedral distribution of K310, G311, Y313, F329, and P330 in mt(mult)‐spike is ordered and stabilized (Figure 7E). The residues in mt(del)‐spike show disorder, Y313 being the maximum. Y313, F329, and P330 illustrate instability, where F329 accounts for maximum stability, although K310 and G311 show a slight increase in stability (Figure 7F). Overall, the dihedral distributions in the linker loop show loss of flexibility in the mutated variants. The decrease in flexibility of the linker region in the mutant variants acts as a hinge to maintain articulation of the opening and closing movement between NTD and RBD. It has been found that mt(del)‐spike preferred to attain the close conformation in the equilibrated state.

FIGURE 7.

Color online: Close view of the conformational thermodynamic changes of linker region acting as a hinge in spike protein conformation state variation on the average structure of wildtype and mutant variant of spike protein. The destabilized and disordered residues are marked in red (ball‐like model). The residues stabilized and ordered are shown in green (ball‐like model). The residues of the linker responsible for maximum perturbation are shown here: (A) mt(K417N)‐spike, (B) mt(L452R)‐spike, (C) mt(E484K)‐spike, (D) mt(N501Y)‐spike, (E) mt(mult)‐spike, and (F) mt(del)‐spike.

4. DISCUSSIONS

The most striking aspect of our simulation results is the mutation‐induced transition from a closed state to an open state via the linker loop between NTD and RBD. This transition appears to be present across different mutant variants we consider in our studies. One can expect that among the two discrete conformational states, the “down state” in close condition shields from receptor binding, whereas the “up state” in the open condition is receptor accessible. As the mutant variants of VoC stains attain up state, they are more suitable to bind the host receptor pertaining to increased infectivity than that of the wildtype variant. Considerable flexibility and dynamics in the S1 domain, especially around the RBD, revealed from our simulations, can play a vital role in the virulence of SARS‐CoV‐2 variants. Flexibility in the structural context of spike mutations can cause efficient binding to ACE2‐like host factor. It can also be noted that the hinge‐like linker movement is much restricted in the mutant variants preferring to adopt open conformational state. These can be important for designing therapeutic intervention, vaccine, and drug design.

To study the effect of the mutations on ACE2 recognition, we dock the native spike protein wt‐spike and each of the mutated spike protein mt(del)‐spike with ACE2, where the disordered and destabilized residues are taken to be active residues. ³⁰ The docked complexes are shown in Figure S11. The docking results are shown in Table S19. The binding free energy ΔG and the dissociation constant (K _d) at 25°C are reported from the docking studies. Upon complexed with ACE2, the mutant variants of spike protein show comparatively better interaction than wt‐spike and mt(del)‐spike. The order of binding affinity of ACE2 is as follows: in mt(E484K)‐spike > mt(L452R)‐spike > mt(mult)‐spike > mt(N501Y)‐spike > mt(K417N)‐spike compared to that in wt‐spike (△G = −9.4 Kcal/mol) and mt(del)‐spike (△G = −9.0 Kcal/mol) cases. The number of interactions at the RBD–ACE2 interface in different systems is shown in Table S20. The mutant RBD forms more interfacial hydrogen bonds, imparting better binding with ACE2. While the stable and ordered linker residues help in stabilizing the open conformation in the mutated protein, the instability and disorder in the RBD facilitate the binding to ACE2.

5. CONCLUSIONS

In conclusion, we have performed a detailed in silico analysis of the stability and order of the spike protein of SARS‐CoV‐2, primarily responsible for interaction with human cell receptor ACE2. We find an open conformation of the mutated protein, whereas a closed conformation of the wildtype protein. The RBD shows instability and disordered residues, while the linker region shows stability and order under mutation compared to the wildtype variant. This may help the ACE2 binding, which leads to higher infectivity of the mutated species. From a theoretical perspective, the wide control over the RBD “up”/“down” distribution available to the VoC strains of SARS‐CoV‐2 suggests that precise control of the RBD orientation helps to understand the role of conformational dynamics from the perspective of vaccine and drug development.

Within the RBD, the positions at which amino acid substitutions are present at the highest frequency are located close to the RBD–ACE2 interface. Such substitutions may facilitate immune escape by increasing receptor‐binding affinity, independent of any effect that they may have on antibody recognition of epitopes. ³⁵ For enhancing the fitness during virus–host evolution, it is expected that mutations of the RBD residues, especially residues in the receptor binding motif for both ACE2 binding and antibody recognition, need to be balanced to keep or increase ACE2 binding while disrupting antibody recognition at the same time. This aspect needs further microscopic studies similar to ours. Further studies are required to know the full range of the consequences emerging from the mutations or deletions. For example, the mutations may affect trimer formation or trimer stability. It is recently shown that the D614G mutation increases the spike protein trimer stability and halts the premature release of the S1 subunit. ³⁶

AUTHOR CONTRIBUTIONS

Aayatti Mallick Gupta curated, analyzed, and interpreted the data and wrote the manuscript. Jaydeb Chakrabarti interpreted the data and reviewed and edited the manuscript.

CONFLICT OF INTEREST

The authors declare that they have no conflicts of interest.

Supporting information

Table S1 The changes in conformational thermodynamics (kJ/mol) in the residues of RBM of mt(K417N)‐spike with respect to wt‐spike

Table S2 The changes in conformational thermodynamics (kJ/mol) in the residues of RBM of mt(L452R)‐spike with respect to wt‐spike

Table S3 The changes in conformational thermodynamics (kJ/mol) in the residues of RBM of mt(E484K)‐spike with respect to wt‐spike

Table S4 The changes in conformational thermodynamics (kJ/mol) in the residues of RBM of mt(N501Y)‐spike with respect to wt‐spike

Table S5 The changes in conformational thermodynamics (kJ/mol) in the residues of RBM of mt(mult)‐spike with respect to wt‐spike

Table S6 The changes in conformational thermodynamics (kJ/mol) in the residues of RBM of mt(del)‐spike with respect to wt‐spike

Table S7 The changes in conformational thermodynamics (kJ/mol) in the residues of NTD of mt(K417N)‐spike with respect to wt‐spike

Table S8 The changes in conformational thermodynamics (kJ/mol) in the residues of NTD of mt(L452R)‐spike with respect to wt‐spike

Table S9 The changes in conformational thermodynamics (kJ/mol) in the residues of NTD of mt(E484K)‐spike with respect to wt‐spike

Table S10 The changes in conformational thermodynamics (kJ/mol) in the residues of NTD of mt(N501Y)‐spike with respect to wt‐spike

Table S11 The changes in conformational thermodynamics (kJ/mol) in the residues of NTD of mt(mult)‐spike with respect to wt‐spike

Table S12 The changes in conformational thermodynamics (kJ/mol) in the residues of NTD of mt(del)‐spike with respect to wt‐spike

Table S13 The changes in conformational thermodynamics (kJ/mol) in the residues of linker of mt(K417N)‐spike with respect to wt‐spike

Table S14 The changes in conformational thermodynamics (kJ/mol) in the residues of linker of mt(L452R)‐spike with respect to wt‐spike

Table S15 The changes in conformational thermodynamics (kJ/mol) in the residues of linker of mt(E484K)‐spike with respect to wt‐spike

Table S16 The changes in conformational thermodynamics (kJ/mol) in the residues of linker of mt(N501Y)‐spike with respect to wt‐spike

Table S17 The changes in conformational thermodynamics (kJ/mol) in the residues of linker of mt(mult)‐spike with respect to wt‐spike

Table S18 The changes in conformational thermodynamics (kJ/mol) in the residues of linker of mt(del)‐spike with respect to wt‐spike

Table S19 Docking results of ACE2 binding due to mutations of RBD of spike protein

Table S20 Intermolecular interactions in RBD–ACE2 interface in different mutant variants

Figure S1 Histogram distribution of dihedral angles of wildtype and variant system due to RBD residue G446

Figure S2 Histogram distribution of dihedral angles of wildtype and variant system due to RBD residue Y449

Figure S3 Histogram distribution of dihedral angles of wildtype and variant system due to RBD residue Y489

Figure S4 Histogram distribution of dihedral angles of wildtype and variant system due to RBD residue Q493

Figure S5 Histogram distribution of dihedral angles of wildtype and variant system due to RBD residue T500

Figure S6 Histogram distribution of dihedral angles of wildtype and variant system due to RBD residue G502

Figure S7 Histogram distribution of dihedral angles of wildtype and variant system due to NTD residue A263

Figure S8 Histogram distribution of dihedral angles of wildtype and variant system due to linker residue G311

Figure S9 Histogram distribution of dihedral angles of wildtype and variant system due to linker residue Y313

Figure S10 Histogram distribution of dihedral angles of wildtype and variant system due to linker residue P330

Figure S11 Docked complex of spike protein (pink) and ACE2 (green)

Gupta AM, Chakrabarti J. Effect on the conformations of the spike protein of SARS‐CoV‐2 due to mutation. Biotechnol Appl Biochem. 2022;1–13. 10.1002/bab.2413