FOXO4-DRI (D-retro-inverso peptide of FOXO4 residues 150-180) disrupts the FOXO4–p53 interaction required to retain p53 in the cytoplasm of senescent cells, releasing p53 to trigger apoptosis selectively in senescent cells. The original Baar et al. 2017 Cell paper demonstrated in vivo efficacy in mouse aging models, but FOXO4-DRI is a conformationally flexible peptide with moderate potency and delivery challenges. This plan designs improved binders via: (1) FOXO4 ID-domain hotspot mapping from PDB 4IYF (FOXO4 Interaction Domain crystal structure); (2) BindCraft de novo binder design conditioned on FOXO4 ID hotspot residues 160-166 and 209-212; (3) ProteinMPNN sequence optimization on RFdiffusion-generated backbones; (4) AF2-Multimer forward-fold validation of top designs against FOXO4 ID domain (independent of RFdiffusion design oracle); (5) Foldseek novelty screen vs PDB to confirm no prior-art scaffold overlap; (6) ESM-2 perplexity filter (target perplexity < 8) on top 20 designs; (7) Rosetta ddG estimation for binding affinity ranking. Success criteria: ≥1 design with AF2-Multimer ipTM > 0.75, pAE_interaction < 10 Å, ESM-2 perplexity < 8, Foldseek TM-score < 0.7 vs any PDB structure, and predicted improvement over FOXO4-DRI reference. Target tissue context: fibroblasts, hepatocytes, endothelial cells — all documented senescent cell types responsive to FOXO4-DRI. Paralogs to discriminate: FOXO1 (UniProt Q12778), FOXO3 (O43524) — ID domain sequence identity ~55% — selectivity will be assessed by AF2 cross-docking.
Details
- disease
- aging, cellular senescence
- target_ref
- UniProt:Q16236
- primary_endpoint
- 4IYF fetches successfully and contains ≥2 chains (FOXO4 + p53); if absent fall back to AlphaFold multimer model AF-Q13516-F1 cross-docked with p53 DBD (P04637)
- identification_strategy
- in_silico_KO
Raw fields (3)
- assay_spec
Structural hotspot mapping of FOXO4–p53 interface from PDB 4IYF (FOXO4 ID domain / p53 DBD crystal) + literature triangulation → define minimal pharmacophore → design 3 retro-inverso or stapled helix mimetic sequences as FASTA placeholders for next-tick RFdiffusion/BindCraft execution; Hotspot residue table: FOXO4 CR3 residues (L422, L425, L428, D478 candidate set) ranked by buried SASA contribution at p53 DBD interface, computed via BioPython SASA on 4IYF — PLANNED; Contact map heatmap (Cα–Cα < 8 Å) of FOXO4 CR3 vs p53 DBD chains from 4IYF — PLANNED; {'kind': 'claim', 'text_template': 'The FOXO4–p53 interface buries ≥X Ų of SASA across ≤Y hotspot residues, anchored by a leucine-rich helix segment at FOXO4 [residue range]'}; 3 candidate FASTA sequences: (1) retro-inverso FOXO4 CR3 helix mimetic, (2) stapled-helix variant with Aib substitutions at i,i+4, (3) truncated 18-mer covering top 5 hotspot contacts — PLANNED, ready for BindCraft input next tick- hypothesis
A de novo miniprotein binder targeting the FOXO4 interaction domain (ID) — specifically residues mediating contact with the p53 proline-rich domain — will outcompete FOXO4-DRI (estimated Kd ~1-10 µM) by ≥10-fold in silico affinity, achieving predicted Kd < 100 nM, while retaining selectivity over FOXO1 and FOXO3 paralogs, thereby representing an improved senolytic candidate for selective elimination of senescent cells in aging tissues.
- kill_criteria
Abort or revise if: 4IYF may not be the canonical FOXO4/p53 co-crystal — verify chain annotations on fetch; fall back to AF3 multimer prediction if no experimental complex exists; Retro-inverso sequences require d-amino acid synthesis cost flagged for wet-lab triage note; Budget gate: BindCraft execution deferred to next tick — this plan outputs sequences only, not predicted structures