Computational Design of Broad-Spectrum Antiviral Peptides Targeting Viral Fusion Proteins

1. Introduction

Viral entry into host cells is a critical step in the infection cycle of enveloped viruses, and the fusion process mediated by viral fusion proteins represents an attractive target for antiviral intervention. Antiviral peptides (AVPs) that mimic conserved regions of fusion proteins can competitively inhibit the conformational rearrangements required for membrane fusion, thereby blocking infection at an early stage [1]. The emergence of resistant viral strains and the limitations of conventional antivirals have spurred interest in computational methods to design peptides with broad-spectrum activity against multiple virus species or variants [1]. This article reviews the computational pipeline used to design such peptides, with emphasis on targeting viral fusion proteins through molecular docking, molecular dynamics (MD) simulations, and free energy calculations. The discussion centers on class I fusion proteins, including those of influenza hemagglutinin, HIV gp41, and coronavirus spike (S) protein, and uses recent in silico studies on coronavirus fusion machinery as primary case studies [2, 3].

2. Biological Basis of Viral Fusion Inhibition

Viral fusion proteins are classified by their structural architecture. Class I fusion proteins, present in orthomyxoviruses, retroviruses, and coronaviruses, are characterized by a central trimeric coiled-coil core formed by heptad repeat (HR) regions [2]. During fusion, the protein extends to insert a fusion peptide into the host membrane, then folds back to bring the viral and host membranes into close apposition. This refolding is driven by the interaction between HR1 and HR2 domains, which assemble into a six-helix bundle (6HB) [3]. Peptides derived from the HR2 region can bind to the HR1 trimer, preventing 6HB formation and thus blocking fusion [2, 3]. Broad-spectrum activity can be achieved by targeting highly conserved elements within the fusion machinery, such as the fusion peptide, the stem helix, or the HR1-HR2 interface [3]. Natural and synthetic AVPs have demonstrated activity against a range of enveloped viruses, and computational strategies are increasingly used to identify and optimize these sequences [1].

3. Computational Design Pipeline

The rational design of broad-spectrum antiviral peptides follows a structured computational workflow, as depicted in Figure 1.

flowchart TD
    A[Target Identification: Conserved regions of fusion protein], > B[Structure Retrieval: X-ray, cryo-EM, or homology models]
    B, > C[Molecular Docking: Peptide library screening against target pocket]
    C, > D[Molecular Dynamics Simulations: Refinement and stability assessment]
    D, > E[Binding Free Energy Calculations: MM-PBSA/GBSA or alchemical methods]
    E, > F[Peptide Optimization: Sequence modification to improve affinity and pharmacokinetics]
    F, > G[Experimental Validation: In vitro fusion inhibition and antiviral assays]
    G, >|Feedback| A

Figure 1. Computational workflow for designing broad-spectrum antiviral peptides targeting viral fusion proteins.

3.1 Target Selection and Structure Preparation

The first step is identification of conserved, functionally critical regions of the fusion protein. For class I fusion proteins, the HR1 domain, the fusion peptide, and the stem helix are often targeted because they are less tolerant to mutation [3]. High-resolution structures obtained from X-ray crystallography or cryo-electron microscopy (cryo-EM) are retrieved from public databases. For proteins without experimentally determined structures, AlphaFold2 homology models can be used after validation. The target pocket (e.g., the HR1 hydrophobic groove) is prepared by removing water molecules, adding missing hydrogen atoms, and assigning appropriate protonation states for physiological pH.

3.2 Molecular Docking

Peptide ligands are constructed in silico, either as linear sequences derived from the viral HR2 region or as modified sequences generated by computational mutagenesis. Docking algorithms, such as those using refined empirical force fields, sample peptide conformations and orientations within the target binding site. Scoring functions estimate binding affinity based on van der Waals, electrostatic, and desolvation terms [1]. Docking is used to rank candidate peptides and to identify key residue interactions. For broad-spectrum design, docking against multiple target structures from different virus strains or species can identify peptides that bind conserved features [2, 3].

3.3 Molecular Dynamics Simulations

Docked complexes are subjected to all-atom MD simulations to evaluate the stability of the peptide-protein interaction under physiological conditions. Explicit solvent models and appropriate lipid membrane bilayers are used if the target is a transmembrane or membrane-proximal region [2]. Simulations typically run for 50 to 500 ns; longer simulations (microsecond scale) may capture conformational changes and peptide unbinding events. Trajectories are analyzed for root-mean-square deviation (RMSD), root-mean-square fluctuation (RMSF), and hydrogen bond occupancy to confirm stable binding [2, 3]. Principal component analysis and free energy landscapes reveal essential binding modes.

3.4 Free Energy Calculations

Binding free energies are computed using methods such as Molecular Mechanics Poisson-Boltzmann Surface Area (MM-PBSA) or Molecular Mechanics Generalized Born Surface Area (MM-GBSA). Alchemical approaches (e.g., thermodynamic integration or free energy perturbation) offer higher accuracy but higher computational cost. Relative binding affinities between wildtype and mutant peptides guide optimization [1]. Energy decomposition per residue identifies hotspots for modification.

3.5 Sequence Optimization and Experimental Validation

Based on docking and MD results, peptide sequences are iteratively modified to enhance affinity, proteolytic stability, and solubility. Incorporation of non-natural amino acids, D-amino acids, or cyclization can improve properties. The refined candidates are then synthesized and tested in cell-based fusion inhibition assays and antiviral efficacy studies [1]. The workflow cycles back to computational refinement if initial experimental results are suboptimal.

4. Case Studies in Computational Design

4.1 Targeting the Heptad Repeat 2 Domain of MERS-CoV Spike

The spike protein of Middle East respiratory syndrome coronavirus (MERS-CoV) is a class I fusion protein. Alotaiq et al. explored the HR2 domain as a source of fusion inhibitory peptides using in silico methods [2]. The HR2 region is highly conserved and forms the outer layer of the 6HB bundle. Peptides corresponding to HR2 were docked onto the HR1 trimer derived from the MERS-CoV S2 subunit. Molecular dynamics simulations of the peptide-HR1 complex showed stable binding through hydrophobic and electrostatic interactions. Binding free energy calculations using MM-PBSA confirmed favorable association. The study demonstrated that computational screening of HR2 variants could identify peptides with potential pan-coronavirus activity, as the HR1 groove is structurally similar across betacoronaviruses [2].

4.2 Targetable Elements in the S2 Subunit for Pan-Coronavirus Inhibitors

Guo et al. reviewed the conserved S2 elements among coronaviruses, including the fusion peptide, stem helix, and HR1-HR2 bundle [3]. The high conservation of these regions, in contrast to the variable receptor-binding domain (RBD), makes them ideal targets for broad-spectrum peptide inhibitors. Computational modeling of the stem helix, which folds into a hydrophobic pocket after fusion, has guided the design of peptides that prevent the final stages of membrane fusion. The study emphasized that targeting multiple conserved elements simultaneously could yield peptides with activity against diverse coronavirus genera. Structure-based design using docking and MD simulations has been central to identifying peptide sequences that bind to the prehairpin intermediate conformation of the S2 subunit, blocking refolding [3].

5. Broad-Spectrum Activity and Conservation

A key advantage of targeting fusion machinery is the sequence and structural conservation of critical regions. Table 1 summarizes conserved target regions in class I fusion proteins from different virus families.

Table 1. Conserved target regions in viral class I fusion proteins suitable for broad-spectrum peptide design.

Note: Orthomyxovirus and retrovirus examples are based on general virology and are not directly cited from the provided literature; the computational principles apply equally.

6. Challenges and Future Directions

Despite the promise of computational design, several challenges remain. Antiviral peptides are often susceptible to proteolytic degradation, have short half-lives, and may exhibit poor bioavailability [1]. Computational approaches are now integrating predictions of pharmacokinetics and immunogenicity to address these issues. Advances in machine learning, including deep learning for protein-peptide interaction prediction and generative models for sequence design, are expected to accelerate the discovery of broad-spectrum candidates. The integration of MD simulations with enhanced sampling methods can better capture the energetics of peptide binding to dynamic fusion intermediates. Cross-linking strategies, such as linking the article on Structure-Guided Design of Broad-Spectrum Viral Fusion Inhibitors and In Silico Design of Peptide-Based Viral Entry Inhibitors Targeting Class I Fusion Proteins, provide additional context for these methodologies. The use of the 3D Protein Viewer to visualize peptide-protein interactions can greatly aid in the selection of optimal design candidates.

7. Conclusion

Computational design of broad-spectrum antiviral peptides targeting viral fusion proteins is a rigorous, multiscale endeavor that combines structural biology, molecular modeling, and biophysical analysis. By focusing on conserved elements of the fusion machinery such as HR domains and fusion peptides, peptides can be engineered to inhibit a range of related viruses. The case studies on coronavirus S2 subunits demonstrate the utility of molecular docking, MD simulations, and free energy calculations in identifying stable, high-affinity peptide inhibitors [2, 3]. Continued refinement of computational methods and integration with experimental validation will be essential to translate these designs into effective veterinary antiviral therapies [1].

References

[1] Raj A, Sharmin S, Ahmed Z, et al. Harnessing Antiviral Peptides: From Molecular Mechanisms to Clinical Translation. Current Research in Pharmacology and Drug Discovery. URL: https://www.semanticscholar.org/paper/85a3f00139b3cedb0c0812f7926a54a2701259a5

[2] 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. URL: https://pubmed.ncbi.nlm.nih.gov/42082899/

[3] Guo L, Lin S, Chen Z, et al. Targetable elements in SARS-CoV-2 S2 subunit for the design of pan-coronavirus fusion inhibitors and vaccines. Signal Transduction and Targeted Therapy. URL: https://www.semanticscholar.org/paper/c6e192376f5d9256ced702e3f95e6dcc1109142a *** 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.