Molecular Dynamics Simulations of Viral Glycoproteins: Predicting Host Receptor Binding and Immune Escape

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

Key Takeaways

• Molecular Dynamics (MD) simulations are crucial for understanding viral glycoprotein behavior at atomic resolution, enabling prediction of host receptor binding affinity and identification of immune escape mutations.
• The MD simulation workflow involves preparing 3D structures, solvating them, parameterizing with force fields (e.g., CHARMM, AMBER), and running all-atom simulations for nanoseconds to microseconds.
• Analysis metrics like RMSD, RMSF, and PCA assess conformational stability and dynamics, while MM/GBSA and MM/PBSA methods estimate binding free energies, critical for predicting receptor tropism and cross-species transmission.
• MD simulations reveal how mutations in viral glycoproteins, such as SARS-CoV-2 spike protein's RBD or influenza HA, alter receptor interactions and antigenic properties, informing vaccine design and antiviral development.
• Glycosylation patterns on viral glycoproteins significantly influence immunogenicity and antibody accessibility, with MD simulations elucidating mechanisms of glycan shielding and guiding strategies for enhanced vaccine responses.
• These computational approaches are directly applicable to veterinary medicine for assessing zoonotic risk, designing vaccines against livestock pathogens (e.g., Newcastle disease virus), and developing antiviral therapies.

Molecular dynamics (MD) simulations have become an indispensable tool in computational virology for dissecting the biophysical properties of viral glycoproteins at atomic resolution. These simulations enable the prediction of host receptor binding affinity, the identification of conformational epitopes, and the characterization of escape mutations that arise under immune pressure [<a href="#ref-1">1</a>, <a href="#ref-2">2</a>]. By integrating structural biology data with all-atom force fields, researchers can model the dynamic behavior of glycoproteins from diverse viral families, including influenza A hemagglutinin (HA), paramyxovirus fusion (F) and attachment (G) proteins, coronavirus spike (S) proteins, and retrovirus envelope (Env) trimers [<a href="#ref-3">3</a>, <a href="#ref-4">4</a>]. This article provides a technical review of the MD simulation workflow, its application to predicting receptor tropism and immune evasion, and its relevance to veterinary medicine and zoonotic risk assessment.

Computational Workflow for MD Simulations of Viral Glycoproteins

The typical MD simulation pipeline for viral glycoproteins begins with preparation of a three-dimensional structure, sourced either from experimental methods such as X-ray crystallography or cryo-electron microscopy, or from de novo prediction using tools such as AlphaFold2 [<a href="#ref-5">5</a>, <a href="#ref-6">6</a>]. The structure is then placed in a solvated simulation box with explicit water molecules and ions, parameterized with a molecular mechanics force field such as CHARMM or AMBER [<a href="#ref-7">7</a>, <a href="#ref-8">8</a>]. All-atom MD simulations are run for trajectories spanning hundreds of nanoseconds to several microseconds, using integration time steps of 1 to 2 femtoseconds under periodic boundary conditions [<a href="#ref-9">9</a>]. Equilibration phases involve energy minimization, heating, and pressure coupling to mimic physiological conditions [<a href="#ref-4">4</a>, <a href="#ref-10">10</a>].

Post-simulation analysis focuses on metrics that quantify conformational stability and binding energetics. Root mean square deviation (RMSD) and root mean square fluctuation (RMSF) are used to assess overall structural drift and per-residue flexibility, respectively [<a href="#ref-11">11</a>, <a href="#ref-12">12</a>]. Principal component analysis (PCA) is applied to capture essential dynamics and identify dominant motions linked to receptor recognition or antibody evasion [<a href="#ref-13">13</a>]. Binding free energies are estimated using the molecular mechanics generalized Born surface area (MM/GBSA) or the molecular mechanics Poisson Boltzmann surface area (MM/PBSA) approaches [<a href="#ref-14">14</a>, <a href="#ref-15">15</a>]. Solvated interaction energy (SIE) calculations further prioritize ligand binding at protein interfaces [<a href="#ref-16">16</a>].

Predicting Host Receptor Binding and Tropism

One of the most powerful applications of MD simulations is the prediction of host receptor binding specificity and tropism. For coronaviruses, the interaction between the receptor-binding domain (RBD) of the spike protein and the host angiotensin-converting enzyme 2 (ACE2) receptor is a critical determinant of cross-species transmission [<a href="#ref-17">17</a>, <a href="#ref-18">18</a>, <a href="#ref-19">19</a>]. MD simulations have revealed how mutations in the RBD, such as N481K in SARS-CoV-2, alter hydrogen bonding networks and electrostatic complementarity at the interface [<a href="#ref-11">11</a>]. Similarly, the D614G substitution in the spike protein reshapes allosteric networks and modulates the opening mechanism of the RBD, increasing receptor accessibility [<a href="#ref-20">20</a>].

For influenza A viruses, MD simulations of HA in complex with sialic acid receptors distinguish avian-type alpha-2,3 linkages from mammalian-type alpha-2,6 linkages, providing a computational correlate of host range [<a href="#ref-7">7</a>, <a href="#ref-21">21</a>]. The convergent evolution of the N156K mutation in pandemic H1N1 HA has been shown through MD to alter antigenic drift while maintaining receptor binding affinity [<a href="#ref-7">7</a>]. In paramyxoviruses such as Nipah virus, simulations of the attachment glycoprotein (G) in complex with ephrin receptors reveal key residue contacts that mediate bat-to-human spillover [<a href="#ref-10">10</a>, <a href="#ref-22">22</a>]. For the tick-borne encephalitis virus, MD models of the E protein suggest molecular mimicry with host proteins, potentially linking viral entry to autoimmune sequelae [<a href="#ref-6">6</a>].

Non-enveloped viruses also benefit from MD insights. For white spot syndrome virus, simulations of envelope proteins VP28, VP26, and VP24 have identified multi-target inhibitors that block viral attachment in shrimp [<a href="#ref-23">23</a>]. The Oropouche virus glycoprotein and RNA-dependent RNA polymerase have been targeted in immunoinformatic vaccine design workflows that rely on MD validation of epitope stability [<a href="#ref-3">3</a>].

Identification of Immune Escape Mutations

MD simulations are uniquely suited to characterize the conformational basis of antibody escape. The HIV-1 envelope trimer exhibits high conformational variability, which facilitates immune evasion by cloaking conserved epitopes [<a href="#ref-1">1</a>]. Long-timescale MD simulations of the HIV-1 Env trimer on the virion surface have shown that N-glycan moieties modulate the tilting of the glycoprotein spikes, thereby altering accessibility to broadly neutralizing antibodies [<a href="#ref-24">24</a>]. For SARS-CoV-2, hierarchical mutational profiling combined with energy landscape analysis has identified distinct mechanisms of resistance against antibodies that target the RBD and the N-terminal domain [<a href="#ref-21">21</a>]. The glycosylation state of the spike protein plays a central role in immunogenicity; MD simulations demonstrate that glycan shielding can sterically hinder antibody binding, while specific glycosylation sites can be engineered to enhance vaccine responses [<a href="#ref-19">19</a>, <a href="#ref-25">25</a>].

MD simulations also inform the design of escape mutation prediction pipelines. For influenza A, deep mutational scanning data coupled with MD free energy calculations can rank HA mutations by their likelihood of escaping polyclonal sera [<a href="#ref-7">7</a>]. In the context of dengue virus, consensus physics-based and artificial intelligence driven screening of the pre-fusion envelope protein has identified potent inhibitors that stabilize the closed conformation, thereby reducing exposure of fusion loop epitopes [<a href="#ref-26">26</a>]. For Ebola virus, fragment-based drug design approaches guided by quantitative structure-activity relationships have identified small molecules that bind the glycoprotein and block entry [<a href="#ref-27">27</a>]. Similarly, natural products have been computationally screened against the Ebola glycoprotein to identify entry inhibitors [<a href="#ref-8">8</a>].

Implications for Veterinary Vaccine Design and Antiviral Development

The veterinary applications of MD simulations of viral glycoproteins are extensive. For Newcastle disease virus (NDV), computational and experimental studies have demonstrated that polyene macrolides can disrupt the F protein mediated membrane fusion, offering a basis for antiviral therapy in poultry [<a href="#ref-5">5</a>]. The measles virus, which also infects non-human primates, has been studied with long-timescale MD to identify cannabichromevarin as a stabilizer of the prefusion F protein, a strategy that could be applied to paramyxovirus vaccine antigen design for livestock [<a href="#ref-4">4</a>]. For the Middle East respiratory syndrome coronavirus (MERS-CoV), in silico exploration of the heptad repeat 2 domain of the spike fusion machinery has led to the design of antiviral peptides that inhibit membrane fusion [<a href="#ref-28">28</a>].

MD simulations are also integrated with immunoinformatics to design multi-epitope vaccines. For infectious hematopoietic necrosis virus (IHNV) in fish, a multiple-epitope vaccine was designed using MD to confirm the structural stability of the epitopes displayed on a carrier protein [<a href="#ref-22">22</a>]. For the monkeypox virus, a multi-epitope peptide vaccine targeting the A35R glycoprotein and E8L membrane protein was refined through MD to ensure proper folding and antigen presentation [<a href="#ref-29">29</a>]. The same approach has been applied to the Oropouche virus, where the glycoprotein and RdRp were used as targets for a multi-epitope precision vaccine [<a href="#ref-3">3</a>].

The role of pH in glycoprotein dynamics is also relevant for veterinary receptors. MD simulations have shown that microenvironmental pH influences the structural architecture of junctional adhesion molecules, which serve as receptors for reoviruses and other pathogens, suggesting that tissue specific pH changes can affect viral entry efficiency [<a href="#ref-18">18</a>]. For the SARS-like bat coronavirus spike, structural dynamics and allosteric communication have been mapped using MD to predict which bat species maintain spikes with high human ACE2 binding affinity [<a href="#ref-30">30</a>]. These models are critical for assessing zoonotic spillover risk from bat reservoirs [<a href="#ref-17">17</a>].

Integrated Workflow Diagram

A typical integrated workflow for using MD simulations to predict receptor binding and immune escape is depicted in the following diagram:

flowchart TD
 A["Structural data: X-ray, Cryo-EM, AlphaFold2"] --> B["Structure preparation: solvation, ionization, force field assignment"]
 B --> C["All-atom MD simulation: equilibration, production runs"]
 C --> D["Trajectory analysis: RMSD, RMSF, PCA"]
 D --> E["Binding free energy calculation: MM/GBSA, SIE"]
 D --> F["Mutation scanning: in silico mutagenesis, free energy perturbation"]
 E --> G["Rank receptor binding affinity predictions"]
 F --> H["Identify escape mutations by antibody pressure"]
 G --> I["Predict host tropism and spillover risk"]
 H --> J["Design stable vaccine antigens or escape resistant therapeutics"]
 I --> K["Experimental validation: surface plasmon resonance, pseudovirus entry assays"]
 J --> K

Key Studies Using MD Simulations of Viral Glycoproteins

Future Directions

The continued development of coarse-grained MD models will enable simulations of larger glycoprotein complexes, such as entire virion surfaces, over physiologically relevant timescales [<a href="#ref-9">9</a>, <a href="#ref-12">12</a>]. Integration of MD with deep mutational scanning and machine learning will further automate the prediction of escape mutations under defined antibody pressure [<a href="#ref-16">16</a>, <a href="#ref-21">21</a>]. For veterinary virology, these approaches can be directly applied to pathogens of livestock, poultry, and wildlife to inform surveillance and vaccine antigen design [<a href="#ref-5">5</a>, <a href="#ref-22">22</a>]. The combination of MD with structural prediction methods like AlphaFold2 will allow rapid modeling of glycoproteins from newly emerged viruses without experimental structures [<a href="#ref-6">6</a>, <a href="#ref-30">30</a>]. As computational power increases, routine MD screening of variant effects on receptor binding and immune escape will become a standard component of outbreak preparedness programs [<a href="#ref-13">13</a>, <a href="#ref-18">18</a>].

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