Computational Screening of Small-Molecule Inhibitors Targeting Viral Methyltransferases

1. Introduction

Viral RNA methyltransferases (MTases) are essential enzymes responsible for the 5' capping of viral messenger RNA (mRNA), a process that protects viral transcripts from host exonuclease degradation and facilitates efficient translation [1, 2]. In addition, the cap structure enables molecular mimicry that helps viruses evade host innate immune detection, particularly through the recognition of capped RNA by the host interferon system [3, 4]. For many RNA viruses, including coronaviruses and flaviviruses, the MTase activity resides in nonstructural proteins such as nsp16 (coronavirus 2'-O-MTase) and NS5 (flavivirus MTase) [5, 6]. These enzymes sequentially methylate the N7 and 2'-O positions of the viral RNA cap using S-adenosyl-L-methionine (SAM) as the methyl donor [3, 7].

Given the critical role of MTase activity in viral replication and immune evasion, these enzymes represent attractive targets for antiviral drug discovery [2, 8]. In veterinary medicine, coronaviruses (e.g., infectious bronchitis virus in poultry, porcine epidemic diarrhea virus, and bovine coronavirus) and flaviviruses (e.g., West Nile virus in horses and birds, Japanese encephalitis virus in swine) cause substantial economic losses and zoonotic threats [9, 7]. The high conservation of the MTase catalytic domain across viral genera, particularly the SAM-binding pocket, offers the potential for developing broad-spectrum inhibitors [3, 7]. Computational screening approaches, including molecular docking, pharmacophore modeling, and molecular dynamics simulations, have become indispensable tools for identifying and optimizing small-molecule inhibitors against these targets in a cost-effective manner [1, 4, 9].

This review provides an exhaustive technical overview of computational screening strategies targeting viral MTases, focusing on the biophysical principles of SAM-pocket interactions, docking workflows, binding free energy calculations, and visualization of ligand-binding grids. All discussions are grounded in the context of veterinary virology, with parallels drawn to animal coronaviruses and flaviviruses where applicable.

2. Viral Methyltransferases as Drug Targets

2.1 mRNA Capping and Methylation

The viral mRNA cap (m7GpppN) is formed through a series of enzymatic reactions. In coronaviruses, nsp16 forms a complex with its cofactor nsp10 to catalyze 2'-O-methylation of the first and second nucleotides of the cap, using SAM as the methyl donor [5, 9]. Similarly, flavivirus NS5 MTase performs both N7 and 2'-O methylation sequentially [3, 7]. The methylation steps are critical for viral RNA stability and for preventing recognition by host pattern recognition receptors such as RIG-I [2, 4].

2.2 Structural Conservation of the SAM-Binding Pocket

The SAM-binding pocket of viral MTases is evolutionarily conserved, comprising a deep cleft lined with residues that coordinate the methionine and adenine moieties of SAM [5, 10]. Key residues include aspartates, lysines, and phenylalanines that form hydrogen bonds and pi-stacking interactions [9, 11]. For example, in SARS-CoV-2 nsp16, residues Asp6912, Cys6913, Asp6897, and Asp6928 are critical for sinefungin binding [9]. In flavivirus NS5 MTase, the SAM pocket is similarly defined by conserved residues such as Lys61, Asp146, and Phe133 (numbering varies by virus) [10, 12]. This conservation enables the design of inhibitors that can bind across multiple viral species [3, 7].

2.3 Veterinary Relevance

Animal coronaviruses such as avian infectious bronchitis virus (IBV) and porcine epidemic diarrhea virus (PEDV) rely on nsp16/nsp10 complexes for cap methylation [1, 2]. Flaviviruses like West Nile virus (WNV) and Japanese encephalitis virus (JEV) utilize NS5 MTase [3, 13]. The high sequence identity of MTase domains between human and animal strains supports the hypothesis that inhibitors developed against human viral MTases may exhibit cross-species activity, a critical feature for veterinary applications where multiple host species are at risk [7, 13].

3. The SAM-Binding Pocket: A Druggable Target

The SAM-binding site presents several advantages for small-molecule inhibition. It is a well-defined cavity with predominantly hydrophobic and polar features, making it amenable to structure-based drug design [5, 10]. Importantly, the methyl donor SAM occupies a compact volume, and competitive inhibitors that mimic SAM or its byproduct S-adenosyl-L-homocysteine (SAH) can achieve high binding affinity [3, 7].

Sinefungin, a natural SAM analog, is a broad-spectrum MTase inhibitor and is frequently used as a positive control in screening assays [8, 9]. However, sinefungin is not suitable as a therapeutic due to toxicity and poor selectivity [5]. Thus, computational screens aim to identify non-nucleoside compounds with improved selectivity and pharmacokinetic profiles [4, 8].

Flavivirus NS5 MTase has two binding sites: the SAM site and the RNA-binding site. The RNA-binding site is shallow and solvent-exposed, making the SAM site the preferred target for inhibitor development [10, 12]. For coronavirus nsp16, the SAM pocket is deeply buried within the nsp10-nsp16 interface, and inhibitors must accommodate the tight binding environment [5, 14].

4. Computational Screening Workflow

Computational screening typically follows a multi-step pipeline integrating several in silico methods [1, 11, 13]. The general workflow is illustrated by the Mermaid diagram below.

graph TD
    A[Structure of viral MTase target], > B[Pharmacophore modeling or grid generation]
    B, > C[Virtual screening of compound libraries]
    C, > D[Molecular docking (e.g., Glide, AutoDock)]
    D, > E[Scoring and ranking of hits]
    E, > F[Visual inspection and clustering]
    F, > G[Optional: MD simulations and binding free energy calculations]
    G, > H[Experimental validation (biochemical and cell-based assays)]
    H, > I[Lead optimization cycles]

4.1 Target Preparation

The starting point is a high-resolution crystal structure of the MTase, preferably in complex with a known ligand (e.g., SAM or sinefungin) [5, 9]. For nsp16, structures such as PDB 6WKQ (nsp16-sinefungin) have been used for pharmacophore modeling and docking [9]. For flavivirus NS5, structures from DENV and ZIKV are widely used [4, 10, 15]. The protein must be prepared by adding hydrogens, assigning protonation states, and optimizing hydrogen bonding networks [1].

4.2 Pharmacophore Modeling

Structure-based pharmacophore models can abstract key interactions between the target and known ligands [9]. For nsp16, pharmacophores derived from clusters of MD-simulated complexes have been used to identify common features such as hydrogen bond donors, acceptors, and hydrophobic regions [9]. Pharmacophore screening reduces the chemical space to be explored by docking [9].

4.3 Virtual Screening and Docking

Virtual screening involves docking large libraries (e.g., commercial or natural product collections) into the SAM-binding pocket using algorithms that evaluate ligand pose and binding energy [1, 8, 10]. Docking scores are used to rank compounds. For example, molecular docking against DENV NS5 MTase with a library of limonoids identified compounds with docking scores of approximately -9 to -10 kcal/mol [10]. Similarly, docking of natural product-inspired machaeriols against nsp16 yielded good inhibitory activity [8].

Table 1 summarizes selected computational screening studies from the provided literature.

Table 1. Selected computational screening studies targeting viral MTases

5. Molecular Docking: Algorithms and Scoring

Molecular docking predicts the preferred orientation of a ligand within the binding pocket and estimates binding affinity through scoring functions [1, 9, 15]. Docking algorithms can be categorized into rigid, semi-flexible, and fully flexible approaches. For MTase screening, semi-flexible docking (allowing ligand torsion but keeping protein rigid) is common, followed by refinement using induced-fit docking or MD [9, 11].

Scoring functions include force field-based (e.g., MM/GBSA), empirical, and knowledge-based methods [10]. Many studies use Glide SP/XP, AutoDock Vina, or similar programs [1, 15]. The docking score is not always directly correlated with experimental IC50, so careful validation is required [2, 11].

For the SAM pocket, key interactions include hydrogen bonds with catalytic aspartates and stacking with aromatic residues [5, 9]. For example, in nsp16, hydrogen bonds with Asp6897 and Asp6928 are critical [9]. In flavivirus NS5, interactions with Lys61 and Asp146 are essential [10].

6. Molecular Dynamics Simulations

Molecular dynamics (MD) simulations provide a dynamic view of protein-ligand complexes, assessing stability over time (typically 50-500 ns) [9, 10, 12]. MD is used to:

• Validate stable binding of candidate compounds
• Identify key water-mediated interactions
• Calculate binding free energies using MM/PBSA or MM/GBSA
• Evaluate conformational changes in the binding pocket [4, 9, 11]

In studies targeting nsp16, 150 ns MD simulations showed that compounds C1 and C2 maintained stable hydrogen bonds and salt bridges with key residues [9]. For flavivirus NS5, MD simulations of DIS with DENV MTase showed RMSD below 0.15 nm and 3-4 hydrogen bonds per nanosecond [10]. Similarly, for ZIKV NS5, MD guided optimization of hit compounds improved selectivity and activity [4].

Binding free energy calculations using MM/PBSA or MM/GBSA provide more accurate affinity estimates than docking scores, often correlating better with experimental results [9, 10]. Energy decomposition can identify residues contributing most to binding, guiding further optimizations [9].

7. Displaying Ligand-Binding Grids in 3D Viewers

Visualizing the binding mode is essential for rational design. In advanced computational platforms, a 3D viewer can display a ligand-binding grid that represents the spatial distribution of interaction potentials (e.g., steric, electrostatic, hydrophobic) around the active site [1, 12]. These grids are generated during docking calculations (e.g., Glide grids) and can be overlaid on the protein surface to guide the selection of compounds.

For the SAM-binding pocket, grids often highlight favorable regions for hydrogen bond donors (near catalytic aspartates) and hydrophobic moieties (around the adenine-binding subpocket). Interactive 3D viewers allow medicinal chemists to visually assess whether a docked ligand occupies these favorable grid regions, thereby improving the likelihood of binding. The ability to rotate, zoom, and display protein-ligand contacts (e.g., hydrogen bonds, pi-cation interactions) is invaluable for hit-to-lead optimization.

8. Case Studies in Veterinary Context

8.1 Coronaviruses

SARS-CoV-2 nsp16 has been extensively studied as a prototype for animal coronaviruses. Virtual screening using pharmacophore models followed by docking and MD identified several novel inhibitors, such as C1 and C2, which target the SAM pocket [9]. These compounds form strong hydrogen bonds with Asp6897 and Asp6928 [9]. Another study identified macheriols RS-1 and RS-2 from natural product-inspired libraries, showing sub-micromolar IC50 values and distinct binding interactions in the SAM pocket [8]. Drug repurposing screens identified dolutegravir and bictegravir as potential 2'-O-MTase inhibitors, with binding energies of -9.4 and -8.4 kcal/mol, respectively [16, 17].

8.2 Flaviviruses

Flavivirus NS5 MTase is highly conserved across DENV, ZIKV, WNV, and YFV, making it suitable for broad-spectrum inhibitor design [3, 7]. A fluorescence polarization-based high-throughput screening assay, combined with virtual docking, identified NSC 111552 and NSC 288387 as low-micromolar inhibitors that bind directly to the SAM pocket [3]. These compounds reduced ZIKV replication in cell culture [3]. Limonoids from natural sources were screened against DENV NS5; DIS showed the most stable MD behavior and favorable ADMET profiles [10]. Marine fungus-derived chevalone E also showed high docking scores against DENV NS5 MTase [15].

8.3 Dual Targeting

Some inhibitors target both nsp16 (2'-O-MTase) and nsp14 (N7-MTase) of coronaviruses. Compounds SS148 and WZ16 were identified by cross-screening and co-crystallized with nsp10-nsp16, revealing RNA-dependent SAM-competitive inhibition [5]. This dual inhibition strategy may reduce the likelihood of resistance emergence [5].

9. Challenges and Future Directions

Despite successes, computational screening faces challenges: false positives due to assay interference (e.g., aggregation, redox activity), poor aqueous solubility, and off-target toxicity [2, 4]. Experimental validation remains essential. In veterinary drug development, additional considerations include species-specific pharmacokinetics, safety in food animals, and cost of goods.

Advances in machine learning and free energy perturbation (FEP) will improve prediction accuracy. Integration with cryo-EM structures (e.g., Relion/cryoSPARC workflows) can provide more accurate target conformations. The development of broad-spectrum inhibitors active against both coronaviruses and flaviviruses remains a high priority for pandemic preparedness in animals.

10. Conclusions

Computational screening of small-molecule inhibitors targeting viral MTases is a powerful strategy for discovering antivirals against important veterinary pathogens. By leveraging the conserved SAM-binding pocket, structure-based pharmacophore modeling, molecular docking, and molecular dynamics simulations, researchers can identify and optimize lead compounds with high efficiency. The integration of these methods, validated by biochemical assays, has already yielded promising inhibitors against coronavirus nsp16 and flavivirus NS5 MTases. Future work should expand veterinary-specific compound libraries and validate leads in relevant animal models.

References

[1] Bukhari, K., Alharbi, R., Alharbi, B.F., et al. Targeting the 2’-O-Methyltransferase Activity of SARS-CoV-2 nsp16 Through Discovery of Small Molecule Inhibitors Via an Integrated Computational Approach. Journal of Pharmaceutical Innovation, 2025. URL: https://www.semanticscholar.org/paper/34cf12715e26c4f634ab583fb10a2dbfcc01e9c4

[2] Cao, M., Trieshmann, C.W., Samrat, S., et al. Assay development and screening of inhibitors targeting the SARS-CoV-2 2′-O-methyltransferase NSP16. Pharmaceutical Science Advances, 2025. URL: https://www.semanticscholar.org/paper/a0fd7eda0ced2efe091986de03469cf326a59293

[3] Samrat, S., Bashir, Q., Huang, Y., et al. Broad-spectrum small-molecule inhibitors targeting the SAM-binding site of flavivirus NS5 methyltransferase. ACS Infectious Diseases, 2023. URL: https://www.semanticscholar.org/paper/872a3e4c7dc1d9da065362e2be789d9e16179800

[4] Fiorucci, D., Meaccini, M., Poli, G., et al. Identification of Novel Non-Nucleoside Inhibitors of Zika Virus NS5 Protein Targeting MTase Activity. International Journal of Molecular Sciences, 2024. URL: https://www.semanticscholar.org/paper/62c55e6c4331a7530ca72c601a2dd41f1258368b

[5] Klíma, M., Khalili Yazdi, A., Li, F., et al. Crystal structure of SARS‐CoV‐2 nsp10–nsp16 in complex with small molecule inhibitors, SS148 and WZ16. Protein Science, 2022. URL: https://www.semanticscholar.org/paper/28fa82405b3c2c813cb472a6ed5d082620f87a95

[6] Meyer, C., Garzia, A., Miller, M., et al. Small-molecule inhibition of SARS-CoV-2 NSP14 RNA cap methyltransferase. Nature, 2024. URL: https://www.semanticscholar.org/paper/6d51917289f42ef86e7938288e25eedcd50bfcfb

[7] Brecher, M., Chen, H., Liu, B., et al. Novel Broad Spectrum Inhibitors Targeting the Flavivirus Methyltransferase. PLoS ONE, 2015. URL: https://www.semanticscholar.org/paper/f4541402a7182e28479601a2cd9790d047244da2

[8] Hanna, G.S., Benjamin, M.M., Choo, Y., et al. Informatics and Computational Approaches for the Discovery and Optimization of Natural Product-Inspired Inhibitors of the SARS-CoV-2 2′-O-Methyltransferase. Journal of Natural Products, 2024. URL: https://www.semanticscholar.org/paper/d37fc06481e4f1842380e799342bdef234b86d7d

[9] Shi, L., Wen, Z., Song, Y.-Y., et al. Computational investigation of potent inhibitors against SARS-CoV-2 2′-O-methyltransferase (nsp16): Structure-based pharmacophore modeling, molecular docking, molecular dynamics simulations and binding free energy calculations. Journal of Molecular Graphics and Modelling, 2022. URL: https://www.semanticscholar.org/paper/21e5b888d632c09c24ec2456f0cbdea4f92e9ea5

[10] Pal, P.C., Das, B. Targeting the dengue virus NS5-Methyltransferase SAM binding site with limonoids: Molecular docking, dynamics simulation, DFT and ADMET analysis. Computational Biology and Medicine, 2025. URL: https://www.semanticscholar.org/paper/d103cefc50b5d19c4df18466585b5a4ee5209861

[11] Nigam, A., Hurley, M.F.D., Li, F., et al. Drug Discovery in Low Data Regimes: Leveraging a Computational Pipeline for the Discovery of Novel SARS-CoV-2 Nsp14-MTase Inhibitors. bioRxiv, 2023. URL: https://www.semanticscholar.org/paper/971a751cf93b329f4628ef6fda80b86e7d7e9430

[12] Schmidt, T. Computational approaches for investigating protein-ligand interactions - towards an in-depth understanding of the dengue virus methyltransferase. Journal, 2013. URL: https://www.semanticscholar.org/paper/7f7aacb25b766989f5577f7fc427e8559300eff6 *** 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] Thomas, J., Kumar, R., Satija, J. Virtual reality and cheminformatics-driven discovery of a potential broad-spectrum natural antagonist against flaviviral methyltransferases and its cytotoxicity evaluation. Journal of Computer-Aided Molecular Design, 2025. URL: https://www.semanticscholar.org/paper/6ed715ae9ea1180e9c5e7789852007de4927d9b9

[14] Sulimov, A., Kutov, D., Ilin, I., et al. Novel Inhibitors of 2'-O-Methyltransferase of the SARS-CoV-2 Coronavirus. Molecules, 2022. URL: https://pubmed.ncbi.nlm.nih.gov/35566072/

[15] Hasan, M., Mia, M., Munna, S.U., et al. Seawater fungi-derived compound screening to identify novel small molecules against dengue virus NS5 methyltransferase and NS2B/NS3 protease. Informatics in Medicine Unlocked, 2022. URL: https://www.semanticscholar.org/paper/cd68a0413db2176361585a5cab894b9b08bb2f13

[16] Khan, R.J., Jha, R., Amera, G., et al. Targeting SARS-CoV-2: a systematic drug repurposing approach to identify promising inhibitors against 3C-like proteinase and 2′-O-ribose methyltransferase. Journal of Biomolecular Structure and Dynamics, 2020. URL: https://www.semanticscholar.org/paper/632fc7e04aa3ce56d270afe3768b860856706aaa

[17] Khan, R.J., Jha, R., Amera, G., et al. Targeting Novel Coronavirus 2019: A Systematic Drug Repurposing Approach to Identify Promising Inhibitors Against 3C-like Proteinase and 2'-O-Ribose Methyltransferase. Journal, 2020. URL: https://www.semanticscholar.org/paper/02bc1f9238f6f26f899075d7776de0afcdad9452