Impact of M and E protein mutations in SARS-CoV-2 Omicron variants on reduced antibody binding and increased structural stability: a bioinformatics study (2021–2023)

Abstract

Background

First identified in late 2021, the Omicron variant of SARS-CoV-2 accumulated substantially more mutations than previously circulating variants. This study investigated the genomic characteristics and structural consequences of mutations in the membrane (M) and envelope (E) proteins of dominant Omicron lineages circulating in southern Iran between March 2021 and March 2023.

Methods

A total of 528 clinical samples were analyzed using next-generation sequencing (NGS), Nextclade lineage assignment, and complementary bioinformatics approaches. The structural effects of selected mutations were further evaluated using protein–protein docking, PDBe PISA interface analysis, MM/GBSA binding free energy calculations, and 100-ns molecular dynamics simulations.

Results

Between 2021 and 2023, BA.5.2 accounted for 32.4% of sequenced isolates, whereas XBB.1.9.1 became the predominant lineage during the later phase of the study (14.2%). Structural analysis demonstrated that the interaction interface between the M protein dimer and the Fab fragment remained largely conserved across all investigated variants. However, MM/GBSA calculations revealed mutation-dependent differences in binding energetics, with the BA.5 (Q19E, A38S, A63T) variant exhibiting the least favorable binding free energy despite preservation of the overall interaction interface. Molecular dynamics simulations further showed that the investigated E protein variants maintained compact conformations with reduced conformational fluctuations relative to the wild-type protein throughout the simulation.

Conclusions

Combined genomic surveillance and structural analyses demonstrated that the investigated Omicron-associated mutations largely preserved the overall architecture of the M protein–Fab interaction interface while modulating residue-level energetic contributions and the conformational dynamics of the E protein. These findings indicate that the investigated mutations primarily affect interaction energetics and protein dynamics rather than inducing major structural rearrangements. The integrated computational framework presented in this study provides a useful approach for evaluating the structural consequences of newly emerging SARS-CoV-2 variants and prioritizing mutations for future experimental validation.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12864-026-13206-8.

Introduction

The subfamily Coronavirinae is comprised of four genera: Alphacoronavirus, Betacoronavirus, Gammacoronavirus, and Deltacoronavirus [1]. Phylogenetic analyses suggest that SARS-CoV-2 is closely related to beta-coronaviruses [2]. Similar to other coronaviruses, the genome of SARS-CoV-2 consists of positive-sense single-stranded RNA [(+) ssRNA], featuring a 3’-UTR poly (A) tail and a 5′-cap [3]. The SARS-CoV-2 genome is less than 30 kb in length and contains 14 open reading frames (ORFs) that encode structural proteins including spike (S), envelope (E), membrane (M), and nucleocapsid (N), along with accessory proteins and non-structural proteins (NSPs) that play roles in virus replication and assembly [4, 5]. Approximately 65% of the viral genome is located within the first ORF, which can be translated into polyproteins pp1a (representing non-structural proteins 1–11) or pp1ab (which includes non-structural proteins 1–16) [6]. Among these, non-structural proteins NSP3, NSP9, NSP10, NSP12, NSP15, and NSP16 are crucial for viral replication. Additional ORFs encode various structural and accessory proteins [7, 8].

Public health officials and healthcare providers need to understand the characteristics and implications of mutations in the Omicron variants in their specific context. A comprehensive characterization of this variant, including genetic analysis, transmission dynamics, impact on vaccine efficacy and disease severity, can guide decision-making and the implementation of effective control measures.

The M protein serves as a membrane protein for interaction with the nucleocapsid protein and is essential for the formation of the viral envelope [9]. It contributes to both viral assembly and budding, making it an important target for vaccine and antiviral drug development [10]. As the predominant structural protein in the viral envelope, the M protein also plays an essential role in recruiting the S protein to sites of viral assembly and budding. It is also required for genome packaging and nucleocapsid integration into the virion. The extracellular domain, which is located on the viral surface and is glycosylated, interacts with the S protein and plays a key role in virus assembly and stability [11, 12].

The E protein is a small membrane protein involved in viral assembly and release, functioning alongside the M protein to help stabilize the viral envelope. Although its precise role is not completely understood, it is thought to be implicated in the virus’s pathogenesis [13]. SARS-CoV-2’s envelope protein 2 is the smallest structural protein, ranging from 76 to 109 amino acids long, primarily located in the endoplasmic reticulum and Golgi apparatus of the host cell [14]. Here, it plays a role in viral assembly, pathogenesis, and release. The E protein features a topology of five helices forming a narrow pore and a bipartite channel [15]. While the amino acid sequences of E proteins can vary significantly, they maintain structural conservation across different beta-coronavirus genera, consisting of a hydrophilic ectodomain at the N-terminal end (7–12 amino acids), a hydrophobic transmembrane domain (TMD) of 25 amino acids, and a long, highly conserved hydrophilic C-terminal endodomain [16]. Several recent computational studies have explored the structural and functional consequences of SARS-CoV-2 evolution using predictive modeling approaches. These investigations demonstrated that mutations accumulated in circulating variants may alter protein stability, conformational dynamics, receptor interactions, and immune recognition patterns. Such computational frameworks provide important insights into viral adaptation and complement experimental observations. However, most available studies have primarily focused on the S protein, whereas comparatively limited attention has been directed toward structural alterations in the M protein and their potential impact on antibody interactions [17–20].

The transmissibility and pathogenicity of a SARS-CoV-2 strain can be assessed by considering the presence or absence of a specific mutation [21]. This research study analyses 528 complete SARS-CoV-2 genomes collected from Hormozgan province in southern Iran. These assessments can be very helpful in managing and responding to the spread of a variant type in an infectious disease epidemic. In this evaluation, the impact of mutations on the structural stability and functionality of the M and E proteins will also be analyzed. These findings will be valuable for understanding the virus’s evolutionary pathway and supporting public health responses based on key proteins, and may also provide some predictions regarding the impact of mutations on disease severity. The ongoing battle against COVID-19 entered a new phase in October 2021 with the emergence of the Omicron variant of SARS-CoV-2. Omicron, characterized by an unprecedented number of mutations, has raised significant concerns due to its increased transmissibility and ability to evade vaccine-induced immunity. To gain a full understanding of the nature of the mutations, research is needed to assess their impact on disease severity and resistance to existing vaccines. This study aimed to investigate the genetic and epidemiological characteristics of the omicron variant in Iran during the critical years of 2021 and 2023. The analysis focused on the genetic changes observed in 528 samples collected in Iran from March 2021 to March 2023. Using advanced sequencing methods including nanopore sequencing, the research team successfully obtained a comprehensive picture of the viral genome and identified changes in the key M and E proteins of SARS-CoV-2. Subsequent analyses, including molecular docking and dynamics studies, elucidated the impact of these mutations on the stability of the E protein structure and the interaction between the M protein and the fragment antigen-binding (Fab) receptor [22]. This research adds to the growing understanding of the behavior of the Omicron variant and provides important insights into how different variants interact with cellular receptors. Such knowledge is critical to developing targeted strategies to combat this evolving threat in the ongoing battle against COVID-19.

Materials and methods

Patients and detection of SARS-CoV-2 using RT-PCR and next-generation sequencing

This study used a subset of samples and genomic data from a previously described cohort in Hormozgan Province [23]. Although the methodologies for SARS-CoV-2 detection by RT-PCR and whole-genome sequencing using the Nanopore ARTIC protocol were identical to those described previously [24–26], the present study specifically focused on the M and E protein genes from 528 samples. This differs from our previous work, which characterized mutations in the receptor-binding domain (RBD) and ORF regions.

Phylogenetic analysis and genomic variation of SARS-CoV-2

Following sequencing, raw next-generation sequencing (NGS) data (FASTQ files) were filtered using NanoFilt to retain high-quality reads (Q ≥ 10). The filtered reads were then aligned to the SARS-CoV-2 Wuhan-Hu-1 reference genome (accession number NC_045512.2) using Minimap2. Mutation analysis and sequence comparison were performed using the GISAID database [27]. The GISAID platform and the Nextclade web application (https://clades.nextstrain.org/) were used for lineage assignment, comparison with the SARS-CoV-2 reference sequence, and phylogenetic placement. The COVID-19 Genome Reporter (http://giorgilab.unibo.it/coronannotator/) was used to identify mutations and annotate genomic variants.

Data retrieval and variant modeling

The crystal structure of the wild-type (WT) M protein dimer in complex with the Fab receptor (PDB ID: 7VGR) was retrieved from the RCSB Protein Data Bank (https://www.rcsb.org/). During structure preparation, water molecules and ligands were removed from the complex. The receptor was prepared for subsequent analyses using the DockPrep tool implemented in UCSF Chimera (version 1.10.1). The complete structure of the E protein was modeled using the I-TASSER web server [28]. Structural models of the SARS-CoV-2 M protein variants, including BA.2 (Q19E, A63T), BA.2 (Q19E, A38S, A63T), XBB.1.9.1 (L17F, Q19E, A63T), and BA.2 (Q19E, V23L, A63T), were generated from the WT structure using UCSF Chimera. Similarly, E protein variants, including BA.5.2.1 (T9I), BA.2.75.1.1 (T9I, T11A, D72G), and BN.1.2 (T9I, T11A, D72G), were generated based on the corresponding WT E protein structure [29–31].

Transmembrane region prediction

The membrane topology of the SARS-CoV-2 M and E proteins was predicted using the TMHMM v2.0 server (https://services.healthtech.dtu.dk/services/TMHMM-2.0/) to map the identified amino acid substitutions onto transmembrane and non-transmembrane regions [32]. The amino acid sequences of the M protein (PDB ID: 8CTK) and E protein (PDB ID: 7K3G) were analyzed using the default parameters, and the predicted topology was compared with available experimental structural data.

Molecular docking of wild-type and mutant M proteins and YN7756_1 Fab

Molecular docking was performed to evaluate the effects of WT and mutant M protein dimers on their interactions with the Fab receptor. Protein–protein docking was carried out using the ClusPro 2.0 web server [33], which predicts protein complexes through rigid-body docking followed by clustering and energy-based ranking of the generated conformations. The WT M protein and four mutant variants identified in Iran, namely BA.2 (Q19E, A63T), BA.2 (Q19E, A38S, A63T), XBB.1.9.1 (L17F, Q19E, A63T), and BA.2 (Q19E, V23L, A63T), were individually docked with the Fab receptor. For each M protein–Fab system, ten docking poses were generated and evaluated based on interaction quality and structural consistency at the predicted binding interface. The PRODIGY server was then used to estimate the binding affinity of the WT and mutant complexes by calculating the binding free energy (ΔG) and dissociation constant (Kd) for the top-ranked docked models [34]. The selected complexes from the ClusPro 2.0 clustering results were subsequently visualized using PyMOL (version 4.6.0). Protein–protein interfaces were further characterized using the PDBe PISA server (https://www.ebi.ac.uk/pdbe/pisa/) [35]. Hydrogen bonds, salt bridges, hydrophobic contacts, and interface residues were analyzed to compare the interaction patterns between the WT and mutant M protein–Fab complexes.

Binding free energy calculation using MM/GBSA

Binding free energies of the receptor–Fab complexes were estimated using the MM/GBSA method based on representative conformations extracted from the molecular dynamics trajectories [36]. The binding free energy (ΔG_bind) was calculated as:

where , , , and represent the van der Waals, electrostatic, polar solvation, and nonpolar solvation energy terms, respectively. More negative ΔG_bind values indicate more favorable binding.

Simulation of molecular dynamics

The OPLS-AA force field was employed in molecular dynamics (MD) simulations to evaluate the stability of the structures of both wild-type E protein and its mutants, as well as the wild-type and mutant M-Fab protein complexes. These simulations were conducted using the GROMACS 2022 software. The molecular systems were placed in a solvent box filled with single point charge (SPC) water molecules at a distance of 10 Å. Sufficient Na + and Cl- ions were added to achieve charge neutrality. The energy minimization procedure for all MD simulations comprised two stages: first, equilibration under constant Number of Particles, Volume, and Temperature (NVT) for 100 picoseconds (ps), followed by equilibration under constant Number of Particles, Pressure, and Temperature (NPT). Throughout the equilibration process, a Parrinello-Rahman barostat maintained the temperature and pressure at 300 K and 1.0 bar, respectively. To manage long-range electrostatic interactions, the Particle Mesh Ewald (PME) method was utilized, with a cutoff distance of 10 Å and a grid spacing of 0.16 nm. Van der Waals (VDW) interactions were computed with a threshold of 1 nm. Each simulation system operated with a time increment of two femtoseconds (fs) during the 100 ns MD simulations. The stability of the systems after the simulations was evaluated using root mean square deviation (RMSD) and radius of gyration (Rg) metrics. Additionally, the fluctuations of amino acids throughout the simulations were analyzed using root mean square fluctuations (RMSF).

Results

Identification of mutations in the genetic variation of SARS-CoV-2

As described in our previous study, real-time RT-PCR was performed on RNA extracted from clinical samples to confirm active viral replication. High-quality genome sequences were generated, submitted, and analyzed from 528 samples obtained from individuals who tested positive for SARS-CoV-2. Sequences of Iranian SARS-CoV-2 isolates have been deposited in the GISAID database and are accessible to registered users via https://www.epicov.org/.

Among the 528 analyzed cases, 32.4% were classified as the BA.5.2 lineage, making it the most prevalent strain. The next most frequently detected lineages were XBB.1.9.1 (14.2%) and BA.5.2.1 (11.2%). Collectively, these findings illustrate the distribution and prevalence of SARS-CoV-2 genotypic variants circulating in Iran relative to the reference SARS-CoV-2 genome [23]. Table 1 presents the prevalent mutations in the M and E proteins of SARS-CoV-2 variants detected in Iran, compared with the Wuhan Hu-1 reference genome (GenBank accession: NC_045512).

Table 1.

Mutational changes of the SARS-CoV-2 genotypes identified in Iran compared to the reference genome of the SARS-CoV-2 Wuhan-Hu-1 isolate (NC_045512)

Variant	Protein	Amino acid substitution(s)
BA.2	M	Q19E, A63T
BA.2	M	Q19E, A38S, A63T
XBB.1.9.1	M	L17F, Q19E, A63T
BA.2	M	Q19E, V23L, A63T
BA.5.2	E	T9I
BA.2.75.1.1	E	T9I, T11A
BN.1.2	E	T9I, T11A, D72G

M and E proteins transmembrane topology

TMHMM predicted that the SARS-CoV-2 M protein contains three TMDs spanning residues 20–39 (TM1), 51–73 (TM2), and 78–100 (TM3) (Fig. 1a; Table 2). The N- and C-termini were both predicted to face the cytoplasmic side. Mutation mapping showed that L17F and Q19E are located in the N-terminal cytoplasmic region adjacent to TM1, V23L and A38S within TM1, and A63T within TM2.

Fig. 1.

Predicted transmembrane topology of the SARS-CoV-2 (a) M and (b) E proteins generated using TMHMM v2.0. Identified mutations are mapped onto the predicted transmembrane and extramembrane regions

Table 2.

Transmembrane topology and mutation mapping for SARS-CoV-2 M and E proteins

Protein	Number of TM helices	Predicted TM region(s)	Mutation localization
M	3	TM1: 20–39; TM2: 51–73; TM3: 78–100	L17F and Q19E (N-terminal cytoplasmic region); V23L and A38S (TM1); A63T (TM2)
E	1	TM: 12–34	T9I (N-terminal region); T11A (TM boundary); D72G (C-terminal domain)

In contrast, the SARS-CoV-2 E protein was predicted to contain a single transmembrane helix spanning residues 12–34 (Fig. 1b; Table 2). The N-terminus (residues 1–11) was predicted to be cytoplasmic, whereas the C-terminal region (residues 35–75) was oriented toward the lumen/extracellular side. Mutation mapping indicated that T9I is located immediately upstream of the transmembrane helix, T11A at the transmembrane boundary, and D72G within the C-terminal domain.

Docking of SARS-CoV-2 M protein dimer WT and mutants in complex with YN7756_1 Fab and MM/GBSA binding free energy analysis

Chimera was used to generate the WT and mutant M protein dimers for subsequent interaction analyses with the Fab fragment. The investigated variants included BA.2 (Q19E, A63T), BA.2 (Q19E, V23L, A63T), BA.5 (Q19E, A38S, A63T), and XBB.1.9.1 (L17F, Q19E, A63T). Protein–protein docking was then performed using the ClusPro 2.0 server to evaluate the binding of each M protein dimer to the Fab fragment. The docking scores and predicted binding affinities are summarized in (Fig. 2; Table 3).

Fig. 2.

The rational docking model of the Fab receptor and M protein dimer mutations is shown in the following figures: (a) M protein dimer WT Fab, (b) BA.2 (Q19E, A63T)-Fab, (c) BA.2 (Q19E, V23L, A63T)-Fab, (d) BA.5 (Q19E, A38S, A63T)-Fab, and (e) XBB.1.9.1 (L17F, Q19E, A63T)-Fab

Table 3.

Molecular docking results of the M protein dimer- Fab WT and mutations complexes

Complex	ClusPro score (kcal/mol)	Predicted ΔG (kcal/mol)	Kd (M, 37 °C)	Structural interactions (PDBe PISA)	Key MM/GBSA hotspot residues
Wild type	-1565.6	-53.7	4.2 × 10⁻⁴⁰	Conserved hydrogen-bond interaction network involving Arg44, Asn43, Ser94, Ser108, Arg131 and Glu135.	Phe45, Phe37, Leu101, Met109 (major energetic hotspots)
BA.2 (Q19E, A63T)	-1581.4	-53.7	4.1 × 10⁻⁴⁰	Conserved interaction interface	Same hotspot residues as WT
XBB.1.9.1 (L17F, Q19E, A63T)	-1568.5	-51.9	8.2 × 10⁻³⁹	Conserved interaction interface	WT hotspots retained; Trp92 showed increased contribution
BA.2 (Q19E, V23L, A63T)	-1596.4	-54.1	1.9 × 10⁻⁴⁰	Conserved interaction interface	WT hotspots retained with minor redistribution of residue contributions
BA.5 (Q19E, A38S, A63T)	-1596.4	-50.6	7.1 × 10⁻³⁸	Conserved interaction interface	Reduced contributions of Phe45 and Phe37; additional contributions from Asn57, Tyr53, and Phe96

Structural analysis using PDBe PISA demonstrated that the overall interaction interface between the M protein and the Fab fragment was largely conserved across all complexes. The principal hydrogen-bond interaction network, involving residues Arg44, Asn43, Ser94, Ser108, Arg131, and Glu135, was largely preserved in both the WT and mutant complexes (Table 3).

To determine whether the conserved interaction interfaces were associated with similar energetic stabilization, MM/GBSA binding free energy calculations were performed. All complexes exhibited negative binding free energies, indicating thermodynamically favorable M protein–Fab association (Table 4). Among the investigated systems, BA.2 (Q19E, A63T) showed the most favorable binding free energy (− 120.74 kcal/mol), followed by BA.2 (Q19E, V23L, A63T) (− 120.21 kcal/mol). These values were slightly more favorable than that of the WT complex (− 118.29 kcal/mol). The XBB.1.9.1 (L17F, Q19E, A63T) variant exhibited a binding free energy (− 118.86 kcal/mol) comparable to the WT complex, whereas BA.5 (Q19E, A38S, A63T) displayed the least favorable binding free energy (− 116.01 kcal/mol).

Table 4.

MM/GBSA binding free energy decomposition of the wild-type and mutant M protein–Fab complexes

Variant	VDW	ELE	GB	SA	TOTAL (ΔGbind)
WILD	-227.44	-1516.58	1652.37	-26.65	-118.29
BA.2 (Q19E, A63T)	-237.51	-1422.27	1566.98	-27.95	-120.74
BA.2 (Q19E, V23L, A63T)	-234.90	-1421.29	1563.70	-27.71	-120.21
XBB.1.9.1 (L17F, Q19E, A63T)	-234.86	-1415.62	1559.11	-27.48	-118.86
BA.5 (Q19E, A38S, A63T)	-235.21	-1391.07	1537.78	-27.52	-116.01

Analysis of the MM/GBSA energy components showed that the WT complex possessed the strongest electrostatic interaction (ELE = − 1516.58 kcal/mol); however, this favorable contribution was largely offset by the highest polar solvation penalty (GB = + 1652.37 kcal/mol). In contrast, the BA.2-derived variants exhibited moderately weaker electrostatic interactions but substantially lower desolvation penalties, resulting in slightly more favorable overall binding free energies. The XBB.1.9.1 variant displayed a balanced energetic profile that was comparable to the WT complex. In the BA.5 (Q19E, A38S, A63T) complex, the less favorable binding free energy was primarily associated with reduced electrostatic stabilization despite favorable van der Waals interactions.

To identify the molecular basis of these energetic differences, MM/GBSA per-residue decomposition analysis was performed (Table 4). The WT complex was primarily stabilized by energetic hotspot residues including Phe45, Phe37, Leu101, and Met109. Comparable hotspot profiles were observed for the BA.2 (Q19E, A63T) and BA.2 (Q19E, V23L, A63T) variants, whereas the XBB.1.9.1 variant showed an increased energetic contribution from Trp92. In contrast, the BA.5 (Q19E, A38S, A63T) variant exhibited reduced energetic contributions from Phe45 and Phe37. Although Asn57, Tyr53, and Phe96 contributed favorably to the interaction, these additional contributions did not fully compensate for the reduced stabilization provided by the conserved hotspot residues, which is consistent with the least favorable MM/GBSA binding free energy observed for this complex.

Molecular dynamics of SARS-CoV-2 M protein dimer WT and mutants in complex with YN7756_1 Fab

Structural dynamics of the WT and mutant M protein–Fab complexes were investigated using 100-ns molecular dynamics simulations to evaluate their structural stability, compactness, and residue flexibility. The stability of each complex was assessed by monitoring the RMSD of the protein backbone throughout the simulation (Fig. 3A). All complexes reached stable trajectories following the initial equilibration phase and exhibited only minor structural fluctuations during the remainder of the simulation, indicating that the overall conformations were well maintained (Fig. 3A).

Fig. 3.

A comparison of molecular dynamics parameters between the wild-type Fab complex and four mutant SARS-CoV-2 spike protein complexes: BA.2 (Q19E, A63T), BA.2 (Q19E, V23L, A63T), XBB.1.9.1 (L17F, Q19E, A63T), and BA.2 (Q19E, A38S, A63T). A RMSD. B Rg. C RMSF

The average RMSD values for the WT, BA.2 (Q19E, A63T), BA.2 (Q19E, V23L, A63T), BA.5 (Q19E, A38S, A63T), and XBB.1.9.1 (L17F, Q19E, A63T) complexes were approximately 0.60, 0.59, 0.61, 0.63, and 0.47 nm, respectively. Although the XBB.1.9.1 complex exhibited the lowest average RMSD, the differences among the investigated systems were small, indicating comparable structural stability across all WT and mutant complexes.

The structural compactness of the complexes was further evaluated by calculating the radius of Rg throughout the simulation (Fig. 3B). The Rg values remained nearly constant for all systems, demonstrating that the overall compactness of the complexes was preserved during the 100-ns simulation. Collectively, the RMSD and Rg analyses indicate that both the WT and mutant M protein–Fab complexes maintained stable conformations after complex formation.

Since all complexes remained compact and structurally stable throughout the simulation, the Fab fragment remained associated with the M protein during the 100-ns simulation. RMSF analysis was subsequently performed to evaluate residue-level flexibility (Fig. 3C). Overall, the fluctuation profiles were similar among all investigated complexes, indicating that the mutations did not substantially alter local structural dynamics. The XBB.1.9.1 (L17F, Q19E, A63T) and BA.5 (Q19E, A38S, A63T) complexes exhibited slightly lower residue fluctuations than the remaining variants, whereas the WT and BA.2-derived complexes showed comparable RMSF profiles. Collectively, these findings suggest that the investigated mutations preserved the overall structural stability of the M protein–Fab complexes while producing only modest differences in residue-level flexibility during the molecular dynamics simulations.

Molecular dynamics simulation to evaluate the stability of E protein WT and mutant structures

The three-dimensional structure of the WT E protein was predicted using the I-TASSER server. Point mutations corresponding to the investigated variants were subsequently introduced into the WT model using the Rotamers and Minimize Structure tools implemented in UCSF Chimera, generating mutant structures for molecular dynamics simulations (Fig. 4). Each system was simulated for 100 ns to evaluate structural stability, compactness, and residue-level flexibility.

Fig. 4.

Three-dimensional structures of the SARS-CoV-2 E protein highlighting the identified mutations. A Wild-type (WT) E protein. B BA.5.2.1 variant carrying the T9I substitution. C BA.2.75.1.1 variant carrying the T9I, T11A, and D72G substitutions. D BN.1.2 variant carrying the T9I, T11A, and D72G substitutions. Mutated residues are shown in red and labeled accordingly

Backbone RMSD analysis showed that all systems reached stable trajectories during the simulation (Fig. 5A). The WT protein exhibited the largest conformational fluctuations, with an average RMSD of 1.10 nm, whereas the BA.5.2.1 (T9I), BA.2.75.1.1 (T9I, T11A, D72G), and BN.1.2 (T9I, T11A, D72G) variants displayed lower average RMSD values of 0.88, 0.70, and 0.68 nm, respectively. These findings indicate that the mutant proteins experienced smaller structural deviations from their initial conformations than the WT protein during the simulation.

Fig. 5.

Molecular dynamics simulation analysis of the WT and mutant SARS-CoV-2 E proteins. A RMSD profiles showing structural stability throughout the 100-ns simulation. B Rg profiles indicating the structural compactness of the proteins. C RMSF profiles illustrating residue-level flexibility of the WT and mutant E proteins

The structural compactness of each system was assessed using the Rg (Fig. 5B). The average Rg values were 2.70 nm for the WT protein and 2.48, 2.42, and 1.54 nm for BA.5.2.1, BA.2.75.1.1, and BN.1.2, respectively. The relatively constant Rg profiles throughout the simulation indicate that all systems maintained compact conformations.

Residue-level flexibility was further evaluated using RMSF analysis (Fig. 5C). The WT protein exhibited greater fluctuations across several regions compared with the mutant proteins. In particular, the BA.2.75.1.1 and BN.1.2 variants, which share the T11A and D72G substitutions, displayed the lowest residue fluctuations throughout most of the sequence. Reduced flexibility was observed not only at the mutation sites but also in adjacent and distal regions, suggesting that these substitutions influence the overall dynamic behavior of the E protein.

Collectively, the molecular dynamics simulations indicate that the investigated mutations are associated with reduced conformational fluctuations while preserving the overall structural integrity of the E protein throughout the 100-ns simulation.

Discussion

Structural proteins of SARS-CoV-2 play essential roles in viral assembly, morphogenesis, and host immune recognition [13–17]. Although the E protein is among the most conserved structural proteins, mutations in both the E and M proteins may alter protein conformation, intermolecular interactions, and antigen recognition [37–39]. In the present study, we investigated the structural consequences of selected Omicron-associated mutations using an integrated computational workflow combining molecular dynamics simulations, protein–protein docking, PDBe PISA interface analysis, and MM/GBSA calculations.

Previous studies have shown that the E protein exhibits a remarkably low mutation frequency compared with the other structural proteins of SARS-CoV-2, highlighting its functional importance [37, 39]. Experimental studies further demonstrated that deletion of the E protein markedly attenuates coronavirus replication and pathogenicity, emphasizing its critical contribution to virus assembly and morphogenesis [40]. Nevertheless, several Omicron-associated substitutions, particularly T9I, have been reported to modify the ion channel properties of the E protein and reduce channel selectivity and pH sensitivity. Previous experimental studies have also demonstrated that conserved residues within the E protein are essential for maintaining ion channel activity and overall protein function [41–43]. These observations suggest that even highly conserved structural proteins can undergo mutations capable of altering their structural and functional properties without disrupting the overall architecture of the protein. Interestingly, the three E protein substitutions were distributed across different structural regions of the protein, with T9I located immediately upstream of the transmembrane helix, T11A at the helix boundary, and D72G within the C-terminal domain [39, 42, 43]. This distribution suggests that the observed changes in conformational dynamics may result from cumulative effects on both membrane-associated and extramembrane regions rather than from alterations confined to the transmembrane segment.

Consistent with these observations, molecular dynamics simulations performed in the present study demonstrated that the investigated E protein variants exhibited lower RMSD and RMSF values than the WT protein while maintaining stable radius of gyration profiles throughout the 100-ns simulation. These findings indicate reduced conformational fluctuations and preserved structural compactness of the mutant proteins during the simulation. However, these computational observations should not be interpreted as direct evidence of enhanced biological function, since molecular dynamics simulations primarily describe structural behavior rather than protein activity. Experimental validation will therefore be required to determine whether these mutations influence ion channel function or viral fitness.

An important finding of this study was the distinction between binding affinity and structural stability of the M protein–Fab complexes. Protein–protein docking and PRODIGY analyses predicted modest differences in binding affinity among the investigated variants, whereas molecular dynamics simulations demonstrated that all complexes remained structurally stable and compact following complex formation. These findings indicate that thermodynamic affinity and conformational stability represent complementary, rather than equivalent, properties of protein–protein interactions. Accordingly, preservation of structural stability should not be interpreted as evidence of stronger antibody binding.

Structural interface analysis using PDBe PISA further demonstrated that the overall interaction interface between the M protein dimer and the Fab fragment remained largely conserved among the WT and mutant complexes. The principal interface residues participated in comparable intermolecular interactions across all investigated variants, with only a slight reduction in the number of interface contacts observed for the BA.5 (Q19E, A38S, A63T) complex. These observations indicate that the investigated mutations did not substantially remodel the overall antibody-binding interface.

Although the structural interaction interface remained largely conserved, MM/GBSA calculations revealed mutation-dependent differences in energetic stabilization. BA.2 (Q19E, A63T) and BA.2 (Q19E, V23L, A63T) exhibited the most favorable binding free energies, whereas the BA.5 (Q19E, A38S, A63T) complex showed the least favorable energetic profile. Per-residue energy decomposition further demonstrated that the reduced energetic contributions of the conserved hotspot residues Phe45 and Phe37 in the BA.5 variant were not completely compensated by favorable contributions from Asn57, Tyr53, and Phe96, resulting in weaker energetic stabilization despite preservation of the overall interaction interface. These findings emphasize that mutation-associated energetic redistribution, rather than large-scale structural rearrangement, is the primary factor underlying the observed differences in predicted binding affinity.

The interaction between the SARS-CoV-2 M protein and antibody-derived Fab fragments should be interpreted primarily in the context of antigen recognition rather than direct viral neutralization. Although the spike protein remains the principal target of neutralizing antibodies, accumulating evidence suggests that additional structural proteins, including the M protein, may contribute to immune recognition and antigen presentation [44–45]. Consequently, mutation-induced alterations in the energetic properties of the M protein–Fab interface may influence epitope accessibility and patterns of antibody recognition without substantially altering the global structure of the complex.

Collectively, the docking, PDBe PISA, molecular dynamics, and MM/GBSA analyses consistently indicate that the investigated Omicron-associated mutations largely preserve the structural integrity of both the E protein and the M protein–Fab complexes while producing measurable changes in conformational dynamics and residue-level energetic contributions. These findings highlight the importance of integrating structural dynamics with energetic analyses when evaluating the potential effects of emerging SARS-CoV-2 variants.

The continuous emergence of new SARS-CoV-2 lineages emphasizes the importance of ongoing structural surveillance and predictive computational approaches [46–48]. Although the present study focused on selected Omicron subvariants, the computational workflow employed here provides a useful framework for assessing newly emerging variants and prioritizing mutations for subsequent experimental validation of their structural and functional consequences.

Conclusion

Between March 2021 and March 2023, genomic surveillance of SARS-CoV-2 isolates from southern Iran identified the emergence of multiple Omicron-associated mutations in the E and M proteins, which were subsequently evaluated using structural bioinformatics approaches. This study combined genomic surveillance with structural bioinformatics to investigate the effects of selected Omicron-associated mutations in the SARS-CoV-2 E and M proteins. Molecular dynamics simulations demonstrated that the investigated E protein variants maintained compact conformations with reduced structural fluctuations relative to the wild-type protein. Similarly, analyses of the M protein–Fab complexes showed that the overall interaction interface was largely conserved across the investigated variants according to PDBe PISA, despite mutation-dependent differences in binding energetics revealed by MM/GBSA calculations. In particular, the BA.5 (Q19E, A38S, A63T) variant exhibited the least favorable energetic stabilization, whereas BA.2-derived variants showed slightly more favorable binding free energies. These findings indicate that Omicron-associated mutations primarily modulate residue-level energetic contributions rather than inducing major structural rearrangements of the interaction interface. Collectively, this integrated computational workflow provides a useful framework for evaluating newly emerging SARS-CoV-2 variants and prioritizing mutations for future experimental validation and therapeutic investigation.

Abbreviations

M

Membrane

E

Enveolpe

S

Spike

MD

Molecular dynamics

ORFs

Open reading frames

N

Nucleocapsid

NSPs

Non-structural proteins

TMD

Transmembrane domain

Fab

Fragment antigen-binding

NGS

Next-generation sequencing

RBD

Receptor-binding domain

WT

Wild-Type

RMSD

Root mean square deviation

Rg

Radius of gyration

RMSF

Root mean square fluctuations

PBM

Protein-binding motif

MHV

Murine hepatitis virus

Authors’ contributions

M.S.P: Conceptualization, Methodology, Software, Formal analysis, Investigation, Data curation, Writing – original draft, Visualization. B.S.: Conceptualization, Methodology, Formal analysis, Investigation, Data curation, Writing – original draft, Visualization. L.M.: Data curation, Software, Validation, Writing – review & editing. N.A.: Investigation, Validation, Formal analysis, Writing – review & editing. H.G.: Conceptualization, Methodology, Supervision, Project administration, Validation, Writing – review & editing. KH.A.: Conceptualization, Methodology, Supervision, Project administration, Validation, Writing – review & editing. All authors have participated in the (1) conception and design of the study or acquisition and interpretation of data; (2) drafting the article or revising it critically for important intellectual content; (3) final approval of the version to be submitted; and (4) agreement to be accountable for all aspects of the work.

Funding

This research received no specific grant from any funding agency.

Data availability

A total of 528 complete genome sequences were retrieved from the GISAID database ( https://gisaid.org/). The accession numbers for all sequences are listed in Supplementary Table 1 [19]. External Resources for Structural Modeling and Docking: The structure of the Fab receptor was obtained from the RCSB PDB (ID: 7DHG) (https://www.rcsb.org/). Homology modeling of the E protein was conducted via the I-TASSER server ( https://aideepmed.com/I-TASSER/).Protein-ligand docking studies were executed using the ClusPro 2.0 server ( https://cluspro.org/login.php?redir=/home.php ).The datasets used and analyzed in this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study is a secondary analysis of viral genome sequences from the public GISAID database. It does not involve direct use of human subject data (clinical samples or personal information). Therefore, the specific ethics approval and consent questions in this form are not applicable to this work.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.