Molecular Dynamics Simulations of Viral Spike Glycoproteins: Insights into Host Receptor Binding and Antibody Escape

By Dr. Zubair Khalid, DVM, MS, PhD · 2026-05-10

Key Takeaways

• Molecular Dynamics (MD) simulations, employing force fields like AMBER and CHARMM, are crucial for dissecting the atomic-level dynamics of viral spike glycoproteins, revealing how sequence mutations influence host receptor binding and antibody recognition.
• MD simulations elucidate the conformational plasticity of spike proteins, demonstrating how mutations such as SARS-CoV-2's D614G stabilize the receptor-binding domain (RBD) in a more accessible conformation, thereby enhancing host cell entry.
• Computational analysis of spike glycoprotein dynamics, including techniques like Principal Component Analysis (PCA) and Markov State Models, identifies "escape hotspots" where mutations confer resistance to neutralizing antibodies by altering epitope accessibility and binding free energies.
• Glycan shielding, visualized through MD simulations of glycosylated spike models, plays a significant role in immune evasion by sterically hindering antibody access to conserved epitopes, with mutations affecting glycan processing impacting immunogenicity.
• The principles and computational workflows derived from MD studies of human coronaviruses are directly transferable to veterinary pathogens, aiding in the rational design of vaccines and therapeutics by predicting conserved epitopes and variant susceptibility.
• Integration of MD with emerging technologies like deep learning and AlphaFold2 structures accelerates the identification of escape mutations and the design of broadly neutralizing immunogens, crucial for rapid response to emerging viral threats in both human and animal health.

Introduction

Viral spike glycoproteins mediate host cell entry by binding to cognate receptors and driving membrane fusion. Their conformational plasticity enables immune evasion through epitope masking and rapid mutation, making them primary targets for vaccine design. Molecular dynamics (MD) simulations have become indispensable tools for dissecting the atomic-level motions of these glycoproteins, revealing how sequence changes alter receptor engagement and antibody recognition [<a href="#ref-1">1</a>, <a href="#ref-2">2</a>, <a href="#ref-3">3</a>, <a href="#ref-4">4</a>]. By integrating MD with experimental structures, researchers can map energetic landscapes of binding interfaces, predict escape mutations, and rationally design immunogens that elicit broadly neutralizing responses [<a href="#ref-5">5</a>, <a href="#ref-6">6</a>, <a href="#ref-7">7</a>]. This review examines the principles of MD simulation of viral spikes, their application to host receptor binding and antibody escape, and the translational value for veterinary vaccinology.

Fundamental Principles of MD Simulations Applied to Viral Glycoproteins

MD simulations numerically integrate Newton’s equations of motion for a system of atoms, generating trajectories that capture conformational fluctuations over time. For viral spike proteins, simulations typically employ all-atom force fields such as AMBER, CHARMM, or GROMOS, which parameterize bonded and nonbonded interactions [<a href="#ref-2">2</a>, <a href="#ref-8">8</a>, <a href="#ref-9">9</a>]. Coarse-grained models reduce computational cost for large glycoprotein complexes but sacrifice atomic detail [<a href="#ref-10">10</a>, <a href="#ref-11">11</a>]. Simulation timescales range from hundreds of nanoseconds to several microseconds, enabling observation of loop rearrangements, domain motions, and transient binding events [<a href="#ref-12">12</a>, <a href="#ref-13">13</a>, <a href="#ref-14">14</a>]. Analysis metrics include root‑mean‑square deviation (RMSD), root‑mean‑square fluctuation (RMSF), principal component analysis (PCA), and dynamic cross-correlation matrices (DCCM) to identify correlated motions and allosteric pathways [<a href="#ref-2">2</a>, <a href="#ref-3">3</a>, <a href="#ref-15">15</a>]. Binding free energies are estimated using molecular mechanics Poisson, Boltzmann surface area (MM‑PBSA) or thermodynamic integration, quantifying the impact of point mutations on receptor affinity [<a href="#ref-8">8</a>, <a href="#ref-9">9</a>, <a href="#ref-11">11</a>].

Key Simulation Parameters and Techniques

These computational frameworks allow researchers to simulate glycosylated spikes in explicit solvent and membrane environments, capturing the physical chemistry of host‑pathogen interfaces [<a href="#ref-18">18</a>, <a href="#ref-19">19</a>].

Conformational Dynamics and Host Receptor Binding

Spike glycoproteins exist in metastable prefusion states that undergo large conformational rearrangements upon receptor engagement. For coronaviruses, the receptor‑binding domain (RBD) of the S1 subunit switches between “up” (receptor‑accessible) and “down” (occluded) conformations. MD simulations have revealed how mutations such as D614G reshape the allosteric network of the spike trimer, stabilizing the RBD in a more open conformation and enhancing ACE2 binding [<a href="#ref-2">2</a>, <a href="#ref-15">15</a>]. Furin cleavage at the S1/S2 boundary further modulates the conformational landscape, lowering the energy barrier for the transition to the postfusion state [<a href="#ref-12">12</a>, <a href="#ref-15">15</a>].

The RBD‑ACE2 interface is characterized by extensive hydrogen‑bond networks and hydrophobic contacts. Simulation studies combined with Markov state models have delineated the stepwise binding pathway, identifying intermediate states that may be targeted by inhibitors [<a href="#ref-13">13</a>, <a href="#ref-14">14</a>]. Electrostatic complementarity between the RBD and ACE2 evolves under selective pressure, as visualized by interfacial electrostatic potential maps [<a href="#ref-16">16</a>]. For bat‑borne coronaviruses, MD has illuminated how structural differences in the RBD of WIV1 and other sarbecoviruses determine host range and zoonotic potential [<a href="#ref-3">3</a>, <a href="#ref-4">4</a>]. A comparative study of locked spike structures from bat SARS‑like coronavirus WIV1 revealed species‑specific flexibility that correlates with receptor tropism [<a href="#ref-4">4</a>]. Temperature‑dependent adaptation via intra‑host recombination has also been shown to promote epistatic interactions that stabilize the spike in certain environments, as observed in experimental evolution combined with MD [<a href="#ref-12">12</a>].

For influenza A virus, hemagglutinin (HA) undergoes pH‑induced conformational changes in the fusion peptide region after receptor binding. Although less explored in the current computational literature, similar all‑atom and coarse‑grained simulations of HA have been used to probe the effects of glycosylation on receptor avidity and antibody accessibility [<a href="#ref-18">18</a>]. The principles gleaned from coronavirus MD studies are increasingly applied to other class I fusion proteins.

Antibody Escape Mechanisms

Neutralizing antibodies typically target the RBD or other exposed epitopes on the spike. MD simulations provide dynamic views of how mutations alter epitope conformation, solvent accessibility, and antibody‑antigen binding free energies. Deep mutational scanning data integrated with energy landscape analysis have identified “escape hotspots” where single amino acid substitutions confer resistance to multiple antibody classes [<a href="#ref-5">5</a>, <a href="#ref-7">7</a>, <a href="#ref-20">20</a>]. For example, the N481K mutation in the SARS‑CoV‑2 RBD was functionally and structurally characterized using MD, demonstrating reduced antibody affinity while maintaining ACE2 binding [<a href="#ref-1">1</a>]. Simulation‑guided mutational profiling of class I and class IV antibodies showed that escape mutations often arise at positions with high conformational frustration, where small perturbations destabilize the antibody‑antigen interface without compromising receptor engagement [<a href="#ref-5">5</a>, <a href="#ref-7">7</a>].

Glycan shielding also contributes to immune evasion by sterically blocking antibody access. MD simulations of fully glycosylated spike models revealed that glycans at specific N‑linked sites (e.g., N165, N234) can dynamically occlude conserved epitopes, and mutations that alter glycan processing affect immunogenicity [<a href="#ref-18">18</a>]. Liquid‑liquid phase separation of the RBD, driven by intrinsic disorder, has been observed in vitro and proposed as an additional mechanism to sequester antibodies away from functional binding sites [<a href="#ref-19">19</a>].

Several studies have employed computational screening to identify peptide or small‑molecule inhibitors that block the RBD‑ACE2 interface, using MD to validate binding modes and estimate affinities [<a href="#ref-9">9</a>, <a href="#ref-21">21</a>, <a href="#ref-22">22</a>, <a href="#ref-23">23</a>, <a href="#ref-24">24</a>, <a href="#ref-25">25</a>, <a href="#ref-26">26</a>, <a href="#ref-27">27</a>, <a href="#ref-28">28</a>, <a href="#ref-29">29</a>]. These approaches have yielded candidate molecules from natural product libraries and probiotic‑derived peptides, some of which show variant‑spanning activity in simulations [<a href="#ref-9">9</a>, <a href="#ref-27">27</a>]. While primarily developed for human coronaviruses, the methodology is transferable to veterinary coronaviruses such as canine coronavirus, for which an HRC‑derived peptide inhibitor was designed and characterized using similar MD protocols [<a href="#ref-30">30</a>].

Implications for Veterinary Vaccine Design

Veterinary vaccines against coronaviruses (e.g., porcine epidemic diarrhea virus, transmissible gastroenteritis virus, canine coronavirus) and influenza viruses must contend with antigenic drift and host range differences. MD simulations can accelerate vaccine design by predicting which epitopes are structurally conserved and likely to elicit cross‑protective immunity. For example, immunoinformatics approaches have leveraged spike mutation data to design multi‑epitope vaccines that include both receptor‑binding and conserved fusion peptide regions [<a href="#ref-31">31</a>]. Simulation‑based assessment of glycosylation site mutations informs the engineering of stabilized prefusion spike antigens that retain native antigenicity [<a href="#ref-14">14</a>, <a href="#ref-18">18</a>].

The growing availability of cryo‑electron microscopy (cryo‑EM) structures for veterinary coronaviruses, such as the bat SARS‑like virus WIV1, provides high‑resolution templates for MD [<a href="#ref-3">3</a>, <a href="#ref-4">4</a>]. Coarse‑grained MD can rapidly evaluate the impact of multiple mutations across diverse viral lineages, prioritizing variants for experimental testing [<a href="#ref-10">10</a>, <a href="#ref-32">32</a>]. Studies on SARS‑CoV‑2 evolution in nonhuman primates have demonstrated that host adaptation involves specific spike changes that can be recapitulated in computational models, offering a framework for predicting cross‑species transmission risk [<a href="#ref-32">32</a>, <a href="#ref-33">33</a>].

Case Studies and Computational Workflows

A typical MD‑based workflow for spike glycoprotein analysis is depicted below. The pipeline integrates structural data from the Protein Data Bank (PDB), homology modeling when applicable, all‑atom or coarse‑grained simulation, free energy calculations, and machine‑learning‑driven mutation effect prediction.

flowchart TD
 A["Structural data: PDB / Cryo‑EM"] --> B["System preparation: solvation, ionization, glycan patching"]
 B --> C["Energy minimization and equilibration"]
 C --> D["Production MD simulation: NPT ensemble, 100 ns , 10 μs"]
 D --> E["Trajectory analysis: RMSD, RMSF, PCA, DCCM"]
 E --> F["Binding free energy: MM‑PBSA / FEP for RBD‑receptor complex"]
 F --> G["Identify key residues and mutation effects"]
 G --> H["Predict antibody escape mutations / vaccine epitope selection"]
 H --> I["Experimental validation: mutagenesis, binding assays"]

Several recent studies exemplify this pipeline. One investigation used MD‐guided pharmacophore modeling to screen virtual libraries for ACE2‑spike interface blockers, followed by free energy calculations to prioritize compounds [<a href="#ref-21">21</a>]. Another combined solvated interaction energy with MD to identify cannabinoids as variant‑spanning RBD blockers [<a href="#ref-9">9</a>]. Machine‑learning‑guided rational engineering of ACE2‑derived peptides employed MD to confirm binding poses after directed evolution predictions [<a href="#ref-22">22</a>]. The influence of allosteric mutations on spike opening was dissected through long‑scale simulations (microsecond) and PCA, revealing that D614G alters the free energy surface of RBD transitions [<a href="#ref-2">2</a>, <a href="#ref-15">15</a>].

Future Directions and Integration with Emerging Technologies

The fusion of MD with deep learning and large language models promises to accelerate the discovery of escape mutations and broadening antibody responses. Markov state models can cluster MD trajectories into conformational states and compute transition rates, providing a mesoscopic view of spike dynamics that is directly comparable to single‑molecule experiments [<a href="#ref-13">13</a>]. Integration with AlphaFold2 structures enables simulation of viral glycoproteins even in the absence of experimental templates [<a href="#ref-3">3</a>, <a href="#ref-4">4</a>, <a href="#ref-12">12</a>]. Deep mutational scanning datasets, when combined with MD‑derived energy landscapes, improve the accuracy of computational fitness landscapes for spike evolution [<a href="#ref-5">5</a>, <a href="#ref-6">6</a>, <a href="#ref-7">7</a>, <a href="#ref-20">20</a>].

For veterinary medicine, the application of MD to zoonotic coronavirus spikes from bat, pangolin, and mink origins will help assess spillover risk. Coarse‑grained simulations are well suited to scan large sequence spaces and identify mutations that enhance receptor binding across species barriers [<a href="#ref-3">3</a>, <a href="#ref-4">4</a>, <a href="#ref-32">32</a>, <a href="#ref-33">33</a>]. The development of high‑throughput computational pipelines that automatically simulate emerging variants and report susceptibility to vaccine‑induced antibodies will be critical for rapid response.

Conclusion

Molecular dynamics simulations have matured into a central technique for understanding the structural dynamics of viral spike glycoproteins. They reveal how receptor binding triggers conformational changes, how single mutations drive antibody escape, and how glycan shields modulate immunogenicity. The insights gained from systems such as SARS‑CoV‑2 and related coronaviruses are directly transferable to veterinary pathogens, including influenza, PRRSV, and coronavirus diseases of livestock and companion animals. Continued integration with experimental structural biology, deep learning, and high‑performance computing will further empower the rational design of vaccines and therapeutics for veterinary use.

References

Related Clinical & Scientific Guides

• A Practical Guide to Detecting Antimicrobial Resistance Genes in Shotgun Metagenomic Data
• Computational Immunology: Modeling the Immune System
• How to Set Hard Filters for Germline Variant Calling: A Practical Guide to GATK Best Practices