🧬 biology

Structure-Guided Repurposing of SARS-CoV-2 Compounds Identifies Putative Inhibitors of Mpox Viral Enzymes

This study employs a structure-guided repurposing workflow to identify and prioritize specific SARS-CoV-2 antiviral candidates, such as J08_67 and J08_51, as potential inhibitors of key Mpox viral enzymes through rigorous computational docking, molecular dynamics simulations, and pharmacokinetic profiling.

Original authors: Abubakar S. Abdullahi, Mathieu J.M. Tjegbe, Clovis S. Metuge, Sunday N. Okafor, Fidele Ntie-Kang

Published 2026-08-13

📖 7 min read🧠 Deep dive

CC BY 4.0

Original authors: Abubakar S. Abdullahi, Mathieu J.M. Tjegbe, Clovis S. Metuge, Sunday N. Okafor, Fidele Ntie-Kang

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

Imagine the world of viruses as a bustling, chaotic city where tiny invaders try to hijack our cells to build more of themselves. To stop them, scientists act like master locksmiths, trying to design keys that fit perfectly into the specific locks (proteins) viruses use to function. Sometimes, these locks look surprisingly similar across different virus families, much like how a door handle on a house might look very similar to one on a castle. This similarity gives scientists a clever shortcut: instead of forging a brand-new key from scratch for every new virus, they can try out keys they've already made for one virus to see if they fit the locks of another. This is called "drug repurposing." In this story, the researchers are looking at the Mpox virus, a troublemaker that has been causing outbreaks again. They want to find new ways to stop it, specifically by jamming three critical machines inside the virus: a pair of scissors (protease) that cuts viral parts, a stamping machine (methyltransferase) that labels viral messages, and a shredder (nuclease) that destroys our immune system's warning signals.

The researchers decided to test a massive collection of keys originally designed to fight SARS-CoV-2 (the virus behind COVID-19) against these Mpox machines. They didn't just guess; they used powerful computer simulations to see how well these keys fit. Think of it like a high-tech video game where they drop thousands of virtual keys into virtual locks to see which ones turn smoothly. They found that while some keys looked great at first glance, they actually got stuck or fell apart when the simulation got more realistic. However, they did identify a few standout keys that seemed to fit the Mpox locks very well, suggesting they could be the start of new medicines. But here's the catch: these are just computer predictions. The keys haven't been tested in a real lab or on living cells yet, so they are promising "what-ifs" rather than guaranteed cures.

The Great Key Hunt: Finding Mpox Stoppers

In this study, a team of scientists went on a digital treasure hunt. Their goal was to see if they could borrow a library of compounds (chemical keys) that were previously studied for fighting SARS-CoV-2 and use them to stop the Mpox virus. They focused on three specific targets inside the Mpox virus, which are essential for the virus to survive and spread:

The Core Protease: Imagine this as the virus's internal scissors. It cuts long chains of viral proteins into smaller, usable pieces. If you jam these scissors, the virus can't build itself.
The VP39 Methyltransferase: This is like a stamping machine. It puts a special "cap" on the virus's genetic messages so the host cell reads them as its own instructions. Stop the stamp, and the cell ignores the virus.
The Poxin Nuclease: Think of this as a shredder. It destroys the warning signals our immune system sends out when it detects an infection. If you block the shredder, the immune system stays alert and fights back.

The team took a library of 1,238 different chemical compounds—some natural, some synthetic—that had been looked at for SARS-CoV-2. They ran these through a series of computer tests, starting with a fast, broad sweep (High-Throughput Screening) to find the most promising candidates, and then zooming in for a detailed look (Molecular Docking) to see exactly how they fit.

The Results: Winners, Losers, and "Almost" Winners

The computer simulations revealed some fascinating results, but they also showed why you can't just trust the first thing that looks good.

The Protease (The Scissors) Findings:
At first, a compound named 78097010 looked like the superstar. It had the best "docking score," which is like getting the highest grade on a test. However, when the scientists ran a more realistic simulation that accounted for water and how the molecule bends (called MM-GBSA), this compound fell apart. It turned out that while it fit the lock, it cost too much energy to get there, like trying to force a key into a lock that's too tight.
Instead, two other compounds emerged as the real contenders:

J08_67: This one had the most stable energy profile. It fit well and stayed put. In the simulations, it showed a binding energy of -37.03 kcal/mol, which is better than the reference compound they used for comparison. It also looked like it would be easy for the body to absorb.
J02_11: This one was a bit less stable in energy but was incredibly steady over time. In a 100-nanosecond simulation (which is like watching a movie of the molecule interacting with the virus), J02_11 held onto a specific part of the virus (a hydrogen bond with Lys126) for 85% of the time. It was a very reliable partner.

The VP39 (The Stamping Machine) Findings:
Here, the results were even clearer. A compound named JBB43_0001 had the best initial score, but again, the detailed energy check told a different story. The real winner was J08_51.

J08_51 showed a massive binding energy of -73.61 kcal/mol in the initial check, and even after running the 100-nanosecond simulation, it stayed strong with an average energy of -68.38 ± 10.25 kcal/mol.
It formed a very strong, lasting bond with a part of the virus called Asp95 (holding on for 85% of the simulation time) and made new friends with Arg114 and Asp152.
However, the scientists noted that J08_51 might be a bit "clunky" for a real drug. It has a low predicted ability to be absorbed by the gut and is complex to make in a lab, meaning it would need some tweaking to become a medicine.

The Poxin (The Shredder) Findings:
This target was the most cautious. The top-scoring compound from the initial screen, SA40_0002, turned out to be a false alarm. It had a huge "solvation penalty," meaning it hated being in water, which would make it useless in the body.

The best candidate here was J02_14. It had a balanced profile and didn't fall apart like the others. However, it wasn't as strong as the reference compound the scientists used, and because the team didn't run the long 100-nanosecond simulation for this one, it remains a "maybe" rather than a confirmed lead.

What This Means (And What It Doesn't)

The big takeaway from this paper is that cross-viral repurposing works, but you have to be careful. Just because a key fits the lock on a computer screen doesn't mean it will turn the lock in real life. The study showed that looking at the "energy cost" of the fit (using MM-GBSA) and watching how the molecule behaves over time (using Molecular Dynamics) is crucial. Without these extra steps, the scientists would have picked the wrong keys (like 78097010 and SA40_0002).

The authors are very clear about the limits of their work. They did not prove that these compounds kill the virus. They did not test them in cells or animals. They did not show that they are safe for humans. All they have done is use computers to find a few "putative inhibitors"—meaning "suspected inhibitors"—that look promising enough to be tested in a real lab.

The paper suggests that J08_67 and J02_11 are the best starting points for the protease, and J08_51 is the best starting point for the VP39 enzyme. These compounds are now the "shortlist" for the next stage of research, where real scientists will mix them with real virus parts to see if they actually stop the virus from working. Until then, they remain fascinating computer-generated clues in the ongoing battle against Mpox.

Technical Summary: Structure-Guided Repurposing of SARS-CoV-2 Compounds Identifies Putative Inhibitors of Mpox Viral Enzymes

Problem Statement
Mpox (formerly monkeypox) has re-emerged as a global public health concern, yet therapeutic options remain limited. The primary antiviral, tecovirimat, targets the viral VP37 protein and faces challenges regarding resistance, limited efficacy in diverse demographics, and an inability to fully limit viral reproduction. There is an urgent need for alternative therapeutic strategies targeting distinct viral mechanisms. While de novo drug discovery is resource-intensive and slow, drug repurposing offers a rapid pathway. However, many viral enzymes share conserved structural and mechanistic features across virus families, suggesting that compounds optimized for one virus (e.g., SARS-CoV-2) may retain activity against related targets in another (e.g., Mpox). This study addresses the gap in identifying specific, potent inhibitors for three essential Mpox viral enzymes: the I7L core protease (essential for virion maturation), the VP39 mRNA cap methyltransferase (essential for mRNA capping), and the poxin nuclease (involved in immune evasion).

Methodology
The study employed a hierarchical, structure-guided computational framework to screen a manually curated library of 1,238 SARS-CoV-2 antiviral compounds (the UB-CeDD library) against the three Mpox targets. The workflow proceeded as follows:

Target Preparation: High-resolution crystal structures were retrieved from the Protein Data Bank (PDB IDs: 9LIL for core protease, 8B07 for VP39, and 8ORV for poxin). Structures were prepared using Schrödinger's Maestro, including protonation state assignment (pH 7.0), removal of artifacts, and energy minimization.
Protocol Validation: The docking protocol was validated by re-docking co-crystallized ligands into their respective binding sites, yielding Root-Mean-Square Deviation (RMSD) values of 1.46 Å (protease), 0.79 Å (VP39), and 1.83 Å (poxin), confirming the reliability of the binding site definitions.
Virtual Screening: A multi-stage screening process was applied:
- High-Throughput Virtual Screening (HTVS): The entire library was screened to identify initial hits.
- Standard Precision (SP) and Extra Precision (XP) Docking: Top hits were re-docked with increasing precision. Ligand interaction analysis was performed on compounds scoring ≤ -6.00 kcal/mol.
- MM-GBSA Rescoring: Binding free energies ( $\Delta G_{bind}$ ) were calculated using Molecular Mechanics Generalized Born Surface Area (MM-GBSA) to refine rankings and account for solvation and strain effects.
Molecular Dynamics (MD) Simulations: The top-scoring complexes for the core protease and VP39 underwent 100 ns MD simulations using the Desmond system. Stability was assessed via Root-Mean-Square Deviation (RMSD), Root-Mean-Square Fluctuation (RMSF), and protein-ligand interaction persistence.
Pharmacokinetic Profiling: Top candidates were evaluated for ADMET properties (Absorption, Distribution, Metabolism, Excretion, Toxicity) using SwissADME and pKCSM.

Key Results
The study identified specific lead candidates for each target, with rankings shifting significantly after MM-GBSA rescoring and dynamic analysis:

Mpox Core Protease (I7L):
- J08_67 emerged as the top energetic candidate with the most favorable static MM-GBSA binding energy ( $\Delta G_{bind} = -37.03$ kcal/mol), outperforming the co-crystallized ligand. It exhibited a balanced pharmacokinetic profile with high gastrointestinal absorption and a bioavailability score of 0.55.
- J02_11 demonstrated superior dynamic stability during 100 ns simulations, maintaining a stable receptor complex and forming a persistent hydrogen bond with Lys126 (85% occupancy). Its post-MD MM-GBSA energy was comparable to the reference ligand.
- Notably, the initial top docking hit (78097010) was deprioritized due to high solvation penalties and ligand strain energy in the MM-GBSA analysis.
VP39 Methyltransferase:
- J08_51 was identified as the principal candidate. While JBB43_0001 had the best initial docking score, J08_51 showed the strongest MM-GBSA binding energy ($-73.61$ kcal/mol) and maintained favorable interactions with Asp95, Arg114, and Asp152 throughout the simulation.
- However, J08_51 showed lower predicted gastrointestinal absorption (bioavailability score 0.17) and higher synthetic accessibility complexity, suggesting it requires medicinal chemistry optimization rather than direct repurposing.
Poxin Nuclease:
- J02_14 showed the most coherent non-native energetic profile among the hits, though it remained weaker than the co-crystallized ligand. The initial top docking hit (SA40_0002) was deprioritized due to a positive $\Delta G_{bind}$ driven by large solvation penalties and ligand strain.
- No MD simulations or pharmacokinetic analyses were performed for poxin candidates due to their lower confidence scores relative to the other targets.
Safety and Developability:
- The prioritized compounds (J08_67, J02_11, J08_51) showed no predicted Ames mutagenicity, hERG-I inhibition, or hepatotoxicity.
- The study highlighted that relying solely on docking scores can be misleading; MM-GBSA rescoring was critical for filtering out compounds with unfavorable solvation or strain energies.

Significance and Claims
The paper claims that cross-viral chemical-space repurposing is a viable strategy for generating target-specific hypotheses for Mpox. The study does not claim to have discovered a clinical drug but rather "putative inhibitors" and "lead hypotheses."

The primary significance lies in the demonstration that integrating solvation effects, ligand strain, and dynamic stability (via MD) significantly improves candidate prioritization beyond static docking scores. The authors assert that their integrated workflow successfully identified chemically distinct starting points for three mechanistically different Mpox enzymes. Specifically, they propose:

J08_67 as the most balanced candidate for the core protease based on energetics and developability.
J02_11 as a dynamically supported alternative for the core protease.
J08_51 as the strongest lead for VP39, pending structural optimization to improve bioavailability.
J02_14 as a lower-confidence probe for poxin.

The authors emphasize that these findings warrant biochemical validation (enzyme inhibition assays, cell-based antiviral testing) and further medicinal chemistry optimization before any therapeutic application can be considered. The study positions itself as a computational foundation for future experimental work rather than a definitive therapeutic solution.

The Great Key Hunt: Finding Mpox Stoppers

The Results: Winners, Losers, and "Almost" Winners

What This Means (And What It Doesn't)

More like this