Skip to main content
Therapeutic lead

Enumeration of ddhCTP/ddhUTP prodrugs as flavivirus RdRp chain terminators

Bacterial viperins (Pfam Radical_SAM) convert CTP/UTP/GTP into 3'-dehydro analogues (ddhNTPs). Viral RdRps incorporate these as chain terminators. Chemically enumerate ddhNTP derivatives and their prodrugs to cross the membrane of eukaryotic cells infected by flaviviruses (Zika, Dengue, West Nile virus).

Content

Initial intuition

Bernheim et al. (Nature 2021) showed that bacterial viperins—a Radical-SAM enzyme family—produce 3’-dehydro nucleotide derivatives (ddhNTPs). They act as chain terminators against RNA-dependent RNA polymerases (RdRps). This is the same chemical strategy as the human viperin RSAD2.

More importantly, it is the chemical strategy of Sofosbuvir (modified UTP analogue, hepatitis C RdRp inhibitor, blockbuster). For the flavivirus family (Zika, Dengue, West Nile, Japanese encephalitis)—for which no approved antiviral exists in 2026—bacterial ddhNTPs represent a direct chemical starting point.

Hypothesis

A virtual library of ddhNTP structural analogues, designed as phosphoramidate prodrugs (ProTide) to cross the membrane of infected hepatocytes/neurons, should contain selective inhibitors of flavivirus NS5 RdRps.

Three enumeration axes:

  1. Base variation: ddhCTP, ddhUTP, ddhGTP, ddhATP—test incorporation selectivity by NS5 RdRps (which only use A, U, G, C).
  2. 2’ variation: O-methyl, F, OH, H—modulate affinity versus human Pol-II polymerase (viral selectivity).
  3. Prodrug variation: coupling with L-alaninate phosphoramidate (Sofosbuvir template) or diethyl phosphate (Tenofovir template).

The goal: ~120 virtual candidates, filtered by ADMET and docking scores on the Zika NS5 cryo-EM structure (PDB 6KOM), retaining a top-20 for partner synthesis.

In silico plan

  1. 3D model: retrieve AlphaFold structures for representative bacterial viperins (P0DW53, P0DW49, P0DW52) and human RSAD2. Characterize the Radical_SAM active site, model the enzyme-nucleotide-[4Fe-4S] cluster complexes.
  2. RDKit enumeration: ~120 ddhNTP analogues/prodrugs.
  3. Docking on Zika NS5 RdRp (PDB 6KOM) and Dengue NS5 (PDB 4V0R) via AutoDock Vina. Target: nucleotide incorporation pocket.
  4. Selectivity filter: docking score on human Pol-II (PDB 6BQF)—an acceptable hit must have a ΔG_viral / ΔG_human ratio < 0.7.

Limitations and risks

  • Endogenous human ddhNTPs (synthesized by RSAD2) are already partially characterized and have not shown systemic antiviral efficacy in vivo. The challenge is likely the intracellular concentration achieved, not the incorporation itself.
  • Mitochondrial toxicity is the cardinal risk of nucleotide analogues (cf. Telbivudine withdrawal, Fialuridine failure). The screen must include a mitochondrial DNA replication assay.
  • Incorporation by viral RdRps is sensitive to very fine variations (the 2’-α-F of Sofosbuvir vs 2’-α-H of other analogues determines selectivity). Modeling transition state conformations is notoriously difficult for short MDs.

V1 Family Bacterial Viperins, small molecule modality (the most immediate in the arsenal). Seminal paper Bernheim 2021 (Nature, DOI 10.1038/s41586-020-2762-2). Strategy aligned with the family sheet: enumerate ddhNTP derivatives, docking/MD on RdRps of flaviviruses, coronaviruses, orthomyxoviruses, and veterinary viruses.

Peer review

Local evaluation panel. Gesture e of the engagement report.

All leads Library