Spike Protein Dynamics and Host Receptor Binding: A Computational Approach to Predicting Zoonotic Potential

Introduction

The capacity of an enveloped virus to cross species barriers and establish infection in a new host is fundamentally governed by the molecular interaction between its surface glycoprotein and a cognate cellular receptor on the target cell. For coronaviruses, the spike (S) protein mediates this critical first step of entry. The S protein is a trimeric class I fusion protein that undergoes extensive conformational rearrangements to facilitate membrane fusion [1]. Understanding the biophysical determinants of S protein receptor binding is therefore central to predicting zoonotic risk from animal reservoirs such as bats, birds, and swine [2, 3]. Computational virology offers a suite of tools to model these interactions at atomic resolution, enabling the prospective assessment of host range and spillover potential [4, 5].

This article reviews the computational methodologies employed to characterize spike protein dynamics and host receptor binding, with a focus on molecular dynamics (MD) simulations, binding free energy calculations, and machine learning (ML) predictors. The discussion emphasizes applications in veterinary medicine and wildlife surveillance, drawing on structural data from coronaviruses and other zoonotic agents.

Structural Biology of the Spike Protein

The coronavirus S protein is composed of two functional subunits: the N-terminal S1 subunit, which contains the receptor-binding domain (RBD), and the C-terminal S2 subunit, which drives membrane fusion [1, 6]. The RBD undergoes a hinge-like movement between a "closed" (receptor-inaccessible) and an "open" (receptor-accessible) conformation [1, 6]. This conformational equilibrium is modulated by mutations, glycosylation, and interactions with the host membrane environment [7, 8, 9].

Cryo-electron microscopy (cryo-EM) studies have resolved the trimeric S protein in multiple conformational states, revealing the cooperative nature of RBD opening [1]. The D614G substitution, for example, reshapes allosteric networks within the spike trimer, shifting the equilibrium toward the open conformation and enhancing receptor binding [6]. Such structural insights provide the foundation for computational models that simulate the dynamic behavior of the S protein under various conditions [10, 11].

Molecular Dynamics Simulations of Spike-Receptor Interactions

MD simulations solve Newton's equations of motion for a system of atoms over time, providing a trajectory of conformational states [11, 12]. For spike-receptor systems, MD simulations are used to study the stability of the RBD-receptor complex, the role of specific amino acid side chains, and the influence of solvent and ions on binding [13, 14, 12].

Simulations are typically performed using all-atom force fields (e.g., CHARMM, AMBER) in explicit solvent environments [11, 12]. The system is first energy-minimized, then equilibrated under constant temperature and pressure conditions before production runs that can extend from hundreds of nanoseconds to several microseconds [11]. Analysis of MD trajectories includes root-mean-square deviation (RMSD) and root-mean-square fluctuation (RMSF) calculations to assess structural stability and residue flexibility [14, 11].

For example, MD simulations of the SARS-CoV-2 spike RBD in complex with angiotensin-converting enzyme 2 (ACE2) from different species have revealed species-specific binding affinities [13, 5]. The N481K mutation in the RBD was shown to alter local hydrogen bonding networks and reduce binding affinity to human ACE2, suggesting a potential host range restriction [14]. Similarly, simulations of the transferrin receptor interaction with spike variants have provided mechanistic insights into alternative entry pathways [13].

Binding Free Energy Calculations

Quantifying the strength of the spike-receptor interaction is essential for predicting host tropism. Binding free energy (ΔG_bind) can be estimated using several computational approaches, including molecular mechanics generalized Born surface area (MM/GBSA), molecular mechanics Poisson-Boltzmann surface area (MM/PBSA), and free energy perturbation (FEP) [5, 12].

MM/GBSA and MM/PBSA methods combine molecular mechanics energies with continuum solvation models to estimate the free energy of binding from a single MD trajectory [5, 12]. These methods decompose the total binding energy into contributions from van der Waals interactions, electrostatic interactions, and solvation effects [5]. Per-residue decomposition identifies "hot spots" that contribute disproportionately to binding affinity [5, 12].

FEP calculations provide more rigorous estimates of relative binding free energies by simulating the alchemical transformation of one ligand (or receptor) into another [15, 16]. FEP has been applied to predict the impact of RBD mutations on ACE2 binding, with good correlation to experimental measurements [15, 16]. These calculations are computationally expensive but offer high accuracy for ranking variant effects [15].

Machine Learning Predictors of Host Range

The high dimensionality of sequence and structural data has motivated the development of ML models to predict host range and zoonotic potential [4, 17]. These models are trained on features derived from spike protein sequences, structures, and evolutionary conservation patterns [4, 17].

Deep mutational scanning (DMS) experiments provide large-scale functional data on the effects of single amino acid substitutions on receptor binding and antibody escape [18, 4]. DMS data can be used to train ML models that predict the fitness landscape of the spike protein under different selective pressures [18, 4]. For example, models combining DMS data with iterative experimental validation have been used to forecast emerging variants of concern [4].

Biological foundation models, such as protein language models, learn representations of protein sequences from large unlabeled corpora and can be fine-tuned for specific prediction tasks [17]. These models capture evolutionary and structural information without requiring explicit 3D structures, enabling rapid screening of novel viral sequences [17]. Applications include predicting ACE2 binding affinity for bat coronavirus spike sequences and identifying mutations that enhance human receptor recognition [17].

Integrating Computational and Experimental Data

A robust computational pipeline for zoonotic risk assessment integrates multiple data sources and analytical methods. The workflow typically begins with sequence acquisition from public repositories such as NCBI or GISAID, followed by phylogenetic analysis to identify related viruses with known host ranges [19, 20]. Structural modeling, using homology modeling or AlphaFold, generates 3D coordinates for the spike protein of interest [5, 12]. MD simulations and binding free energy calculations then evaluate the stability and affinity of the spike-receptor complex [11, 5, 12]. Finally, ML models trained on DMS and structural data predict the likelihood of cross-species transmission [4, 17].

The following Mermaid diagram illustrates a typical computational workflow for predicting zoonotic potential from spike protein sequence data.

flowchart TD
    A[Viral Sequence Data], > B[Phylogenetic Analysis]
    B, > C[Identify Closest Relatives]
    C, > D[Structural Modeling]
    D, > E[Molecular Dynamics Simulations]
    E, > F[Binding Free Energy Calculations]
    F, > G[ML Prediction of Host Range]
    G, > H[Zoonotic Risk Assessment]
    H, > I[Surveillance Recommendations]

Case Studies in Veterinary Virology

Bat Coronaviruses

Bats are recognized as major reservoirs of coronaviruses with zoonotic potential [2, 3]. Computational modeling of bat coronavirus spike proteins has identified key residues in the RBD that determine compatibility with human ACE2 [13, 5]. MD simulations of the RaTG13 bat coronavirus spike in complex with human ACE2 revealed a lower binding affinity compared to SARS-CoV-2, consistent with experimental data [5]. Mutations at positions 493 and 501 were shown to enhance binding to human ACE2, highlighting potential evolutionary pathways for spillover [5].

Avian Influenza Viruses

Influenza A viruses circulate in wild waterfowl and poultry, with the hemagglutinin (HA) protein determining receptor specificity [19]. Avian influenza HA preferentially binds to α2,3-linked sialic acids, while human-adapted HA binds to α2,6-linked sialic acids [19]. MD simulations of HA-receptor complexes have been used to predict the impact of mutations on receptor binding specificity and to assess the pandemic potential of emerging strains [19].

Swine Coronaviruses

Porcine epidemic diarrhea virus (PEDV) and porcine hemagglutinating encephalomyelitis virus (PHEV) are coronaviruses that cause significant disease in swine [3, 21]. Computational analysis of the PEDV spike S1 domain has identified glycosylation patterns that influence receptor binding and immunogenicity [21]. MD simulations of PHEV spike protein interactions with sialic acid receptors have provided insights into the neurotropism of historical versus contemporary strains [3].

Limitations and Challenges

Despite significant advances, computational predictions of zoonotic potential face several limitations. Force field inaccuracies can lead to errors in binding free energy estimates [11, 12]. The conformational flexibility of glycans on the spike surface is difficult to model accurately, yet glycosylation plays a critical role in receptor binding and immune evasion [8, 9]. MD simulations are also limited by timescale; rare conformational events relevant to receptor binding may not be sampled within typical simulation lengths [11].

ML models are sensitive to the quality and diversity of training data [4, 17]. Models trained primarily on SARS-CoV-2 data may not generalize well to distantly related coronaviruses [4]. Furthermore, the absence of experimental validation for many bat and avian viruses limits the ability to benchmark computational predictions [2, 19].

Future Directions

The integration of cryo-EM with MD simulations offers a powerful approach to capture the full conformational landscape of the spike protein [1, 6]. Enhanced sampling techniques, such as metadynamics and replica exchange MD, can accelerate the exploration of rare events relevant to receptor binding [11, 6]. The development of more accurate force fields for glycans and lipids will improve the realism of membrane-embedded spike simulations [7, 22].

Advances in deep learning, including graph neural networks and transformer architectures, are expected to improve the accuracy of binding affinity predictions [4, 17]. The combination of DMS data with ML models trained on large sequence databases will enable real-time surveillance of emerging variants [18, 4]. These tools can be deployed in veterinary diagnostic laboratories to assess the zoonotic risk of novel viruses detected in animal populations [2, 3].

Conclusion

Computational approaches to modeling spike protein dynamics and host receptor binding provide a quantitative framework for predicting zoonotic potential. MD simulations, binding free energy calculations, and ML models each contribute unique insights into the molecular determinants of cross-species transmission. When integrated with experimental data and phylogenetic surveillance, these computational tools enable proactive risk assessment for emerging zoonotic threats from animal reservoirs. Continued development of force fields, sampling methods, and ML architectures will further enhance the predictive power of these approaches in veterinary virology.

References

[1] Wang Q, Cholak S, Woollard G, et al. Cryo-EM analysis of cooperative conformational changes in the SARS-CoV-2 spike protein trimer. IUCrJ. 2026. https://pubmed.ncbi.nlm.nih.gov/42345087/

[2] Porta EOJ, AlKharboush DF, Jackson L, et al. Targeting SARS-CoV-2 Structural and Accessory Proteins: Emerging Opportunities for Small-Molecule Coronavirus Antivirals. Pharmaceutics. 2026. https://pubmed.ncbi.nlm.nih.gov/42357322/

[3] Aslam A, Saleem W, Zarak I, et al. Different replication behavior of a contemporary porcine hemagglutinating encephalomyelitis virus strain Gent/PS412 compared with the historical neurotropic reference strain VW572. Vet Res. 2026. https://pubmed.ncbi.nlm.nih.gov/42343363/

[4] Sheffield T, Bruneau RC, Won S, et al. Combining machine learning and iterative experiments to keep pace with emerging viral variants of concern. PLoS Comput Biol. 2026. https://pubmed.ncbi.nlm.nih.gov/42308256/

[5] Delgado-Maldonado T, Vargas-Salas G, Bandyopadhyay D, et al. Ligand-Based Pharmacophore Modeling and Structure-Based Virtual Screening for Identifying Potential Therapeutic Agents Targeting the SARS-CoV-2 Spike Protein-ACE2 Receptor Interaction. ChemMedChem. 2026. https://pubmed.ncbi.nlm.nih.gov/42174382/

[6] Kearns FL, Bogetti AT, Calvó-Tusell C, et al. D614G reshapes allosteric networks and opening mechanisms of SARS-CoV-2 spikes. Proc Natl Acad Sci U S A. 2026. https://pubmed.ncbi.nlm.nih.gov/42101997/

[7] Urano R, Yoshimoto S, Hori K, et al. Synergistic Protein-Protein and Protein-Lipid Interactions Drive SARS-CoV-2 Envelope Assembly. J Chem Inf Model. 2026. https://pubmed.ncbi.nlm.nih.gov/42262910/

[8] Wang H, Li X, He P, et al. Investigating the effects and underlying mechanisms of glycosylation sites on the immunogenicity of COVID-19 vaccines. Signal Transduct Target Ther. 2026. https://pubmed.ncbi.nlm.nih.gov/42129141/

[9] Kumar A, Yadav AJ, Tripathi T, et al. Glycan shielding and epitope reorganization drive sotrovimab resistance in SARS-CoV-2 Omicron variants. Arch Biochem Biophys. 2026. https://pubmed.ncbi.nlm.nih.gov/42128042/

[10] Mansur M, Hong H, Singh S, et al. SARS-CoV-2 Spike Protein's Structural Dynamics Affect the Activity of the Bebtelovimab Antibody. J Chem Inf Model. 2026. https://pubmed.ncbi.nlm.nih.gov/42202234/

[11] Navhaya LT, Monama MZ, Matsebatlela TM, et al. In-Depth Molecular Dynamics Simulations Reveal Ligand-Induced Modulations of the HSPA8-SARS-CoV-2 Spike Protein Interaction. Int J Mol Sci. 2026. https://pubmed.ncbi.nlm.nih.gov/42196270/

[12] Kongtaworn N, Sinsulpsiri S, Hanpaibool C, et al. Molecular Dynamics and Solvated Interaction Energy Prioritize Cannabidiol and Cannabinol as Variant-Spanning SARS-CoV-2 RBD-ACE2 Interface Blockers. Molecules. 2026. https://pubmed.ncbi.nlm.nih.gov/42075932/ *** Disclaimer: This article is for educational and informational purposes only. It is not intended to substitute for professional veterinary advice, diagnosis, treatment, or regulatory guidance. Always consult a licensed veterinarian or qualified specialist regarding animal health, disease diagnosis, and therapeutic decisions.

[13] Wu J, Li W, Wu P, et al. Mechanistic insights into the interaction of emerging SARS-CoV-2 spike variants with the transferrin receptor. J Biomol Struct Dyn. 2026. https://pubmed.ncbi.nlm.nih.gov/42251700/

[14] Smatti MK, Al-Khatib H, Suleman M, et al. Functional and structural characterization of the SARS-CoV-2 spike N481K mutation. Arch Virol. 2026. https://pubmed.ncbi.nlm.nih.gov/42243594/

[15] Chiu C, Jawaid MZ, Cox DL. Computational Study of Antibody Binding to SARS-CoV-2 Variants. Antibodies (Basel). 2026. https://pubmed.ncbi.nlm.nih.gov/42345847/

[16] Alshahrani M, Gatlin W, Ludwick M, et al. Mechanisms of Binding and Immune Escape Resistance for Broadly Neutralizing Antibodies Targeting Distinct Conserved SARS-CoV-2 Spike Epitopes: A Hierarchical Approach Integrating Mutational Profiling and Energy Landscape Analysis. Int J Mol Sci. 2026. https://pubmed.ncbi.nlm.nih.gov/42123600/

[17] He Y, Yang Z, Kuo YA, et al. A High-Throughput Platform for Rapid Adaptation of DNA Aptamers to SARS-CoV-2 Evolution. bioRxiv. 2026. https://pubmed.ncbi.nlm.nih.gov/42094361/

[18] Taylor AL, Starr TN. Deep mutational scanning of recent SARS-CoV-2 variants highlights changing amino acid preferences within epistatic hotspot residues. PLoS Pathog. 2026. https://pubmed.ncbi.nlm.nih.gov/42330076/

[19] Williman J, Tomas G, Vagnozzi A, et al. Genome-Wide Analysis of Serial Passage of the Infectious Bronchitis Virus Reveals Evolutionary Dynamics Underlying Attenuation and Immunogenicity. Vaccines (Basel). 2026. https://pubmed.ncbi.nlm.nih.gov/42347589/

[20] Liu J, Zhang Y, Yang L, et al. Genomic insights into SARS-CoV-2 evolution in Yantai, China after the adjustment of the dynamic zero-COVID policy in 2023. PeerJ. 2026. https://pubmed.ncbi.nlm.nih.gov/42079734/

[21] Uhlemann EM, Patel R, Thivierge B, et al. Efficient design and production of glycosylated PEDV spike S1 protein. Sci Rep. 2026. https://pubmed.ncbi.nlm.nih.gov/42236859/

[22] Pantarotto PV, Dos Santos KF, Silva ALT, et al. Heparin interactions with phospholipid and cholesterol monolayers mimicking SARS-CoV-2 envelope and host membranes: A Langmuir film approach. Biochim Biophys Acta Biomembr. 2026. https://pubmed.ncbi.nlm.nih.gov/42106139/

[23] Ferrara F, De Berardinis F, Scognamiglio M, et al. The Shifting Paradigm of Monoclonal Antibodies in COVID-19 Management: From Early Triumphs to Viral Resistance and Future Perspectives. Antibodies (Basel). 2026. https://pubmed.ncbi.nlm.nih.gov/42345852/

[24] Ma X, Zhang Z, Yang D, et al. Beyond ACE2: unveiling the receptor-mediated endocytic dynamic mechanism of SARS-CoV-2 at the single-particle level. J Mater Chem B. 2026. https://pubmed.ncbi.nlm.nih.gov/42300082/

[25] Fu JYL, Lim SK, Jalalonmuhali M, et al. Immune Pressure From Plasma From Immunocompromised and Healthy Individuals Drives Differential SARS-CoV-2 Evolution and Escape Mutation Emergence. J Med Virol. 2026. https://pubmed.ncbi.nlm.nih.gov/42272263/

[26] Rodríguez-Castillo A, Postigo T, Pérez-García F, et al. Dynamic ACE(2)-Functionalized Magnetic Micromotors for Electrochemical Protein S Detection from SARS-CoV-2: Toward Diagnosis after Virus Mutation. ACS Sens. 2026. https://pubmed.ncbi.nlm.nih.gov/42269600/

[27] Stavridis D, Wacker M, Safarov R, et al. Deficit in blood MAIT cells, altered cytokines and T cell dynamics in patients with coronary artery disease. Cell Immunol. 2026. https://pubmed.ncbi.nlm.nih.gov/42235192/

[28] Sugumar M, Pragalath S, Kugarajah V. Immunoinformatics-based epitope prediction in MERS-CoV and docking analysis of Catharanthus roseus compounds for drug discovery. Comput Biol Chem. 2026. https://pubmed.ncbi.nlm.nih.gov/42218862/

[29] Altaf M, Kin Cheng MT, González-Vázquez LD, et al. SARS-CoV-2 intra-host recombination promotes epistatic spike interactions and temperature-dependent adaptation. Cell Rep. 2026. https://pubmed.ncbi.nlm.nih.gov/42213788/

[30] Demirel Ç, Kibar F, Dağlıoğlu G, et al. Long-Term Durability and Dynamics of Anti-S-RBD IgG Response in Healthcare Workers: A Comparative Analysis of Homologous and Heterologous SARS-CoV-2 Vaccination Schedules in a 1-Year Serial Cross-Sectional Study. Med Sci Monit. 2026. https://pubmed.ncbi.nlm.nih.gov/42198805/

[31] Bischof ML, Irrgang P, Tenbusch M, et al. Simultaneous quantification of anti-Spike IgG subclasses against various SARS-CoV-2 variants using a multiplex serological assay. Med Microbiol Immunol. 2026. https://pubmed.ncbi.nlm.nih.gov/42171795/

[32] Wauer NA, Calvó-Tusell C, Dommer AC, et al. Condensate-Like Organization in Respiratory Aerosols Modulates the Dynamics of an Airborne Virus. bioRxiv. 2026. https://pubmed.ncbi.nlm.nih.gov/42146628/

[33] Ullah S, Ullah A, Ahsan Halim S, et al. Selective Targeting of SARS-CoV-2 Spike Protein by Isatin-Hydrazide Hybrids: Synergistic In vitro and Molecular Dynamics Approaches. Curr Top Med Chem. 2026. https://pubmed.ncbi.nlm.nih.gov/42136454/

[34] Kleber Rodrigues Lima W, Monteiro Ferreira G, Zeneida Gomes Parente Alves Lima C, et al. Antiviral Potential of 3,4-Dimethoxychalcone Against SARS-CoV-2: A Promising Candidate. Curr Med Chem. 2026. https://pubmed.ncbi.nlm.nih.gov/42083366/

[35] Alotaiq N, Dermawan D, Chtita S. Targeting Middle East Respiratory Syndrome Coronavirus Spike Fusion Machinery With Antiviral Peptides: In Silico Exploration of the Heptad Repeat 2 Domain. Microbiologyopen. 2026. https://pubmed.ncbi.nlm.nih.gov/42082899/