FOXO4 sequesters p53 in the nucleus of senescent cells, preventing p53-driven apoptosis; the FOXO4-DRI stapled peptide (Baar et al. Cell 2017) disrupts this interaction and clears senescent cells in vivo. The 2025 Kohoutova et al. Nature Communications structure defines the DBD:p53-TAD binding surface with atomic precision, including key contacts at FOXO4-DBD residues R19, H98, S100, H215, and the p53-TAD Phe19/Trp23/Leu26 hydrophobic triad. This plan targets that interface for RFdiffusion-conditioned miniprotein design (20-50 aa), followed by ProteinMPNN sequence design, AF3-Multimer forward-fold validation (ipTM > 0.6 threshold), ESM-2 perplexity filter (log-likelihood > -1.5 per residue), and Foldseek novelty screen vs PDB. Success criterion: at least one candidate with AF3-Multimer ipTM > 0.65, Foldseek TM-score < 0.5 to nearest PDB hit, and predicted displacement of p53-TAD as assessed by complex pAE < 10 Å. Planned wet-lab route: gene synthesis → E. coli expression (His-tag) → SPR against FOXO4-DBD (P98177 aa 157-269) → HCT116 cell senescence induction (doxorubicin) → SA-β-gal / γH2AX clearance assay.
Details
- disease
- aging / cellular senescence
- target_ref
- UniProt:P98177
- primary_endpoint
- ≥1 redesigned sequence with ESM-2 pseudo-log-likelihood ≥ wild-type FOXO4-DBD segment (baseline computed in same run)
- identification_strategy
- in_silico_KO
Raw fields (3)
- assay_spec
Fetch PDB 9IYE (NMR ensemble); extract FOXO4-DBD:p53-TAD interface contacts (≤4.5 Å heavy-atom cutoff, model 1); enumerate hotspot residues; run ProteinMPNN (T=0.1, 32 seqs) on FOXO4-DBD scaffold with p53-TAD contact residues fixed; score outputs with ESM-2 log-likelihood; rank top-5 sequences.; hotspot residue table: chain, resnum, resname, mean contact frequency across NMR ensemble, burial delta-SASA — planned; top-5 ProteinMPNN redesigns of FOXO4-DBD ranked by ESM-2 log-likelihood; FASTA + per-sequence score — planned; {'kind': 'claim', 'text_template': 'ProteinMPNN redesign FOXO4-DBD-v[N] achieves ESM-2 log-likelihood ≥ reference FOXO4-DRI peptide baseline — pending metric computation'}- hypothesis
Structure-informed de novo peptides targeting the FOXO4 forkhead DNA-binding domain (DBD) surface that contacts the p53 transactivation domain (TAD) will outperform the original FOXO4-DRI stapled peptide on in-silico ipTM / ipAE metrics (AF3-Multimer complex validation) and ESM-2 log-likelihood, providing a tractable senolytic lead with a defined binding epitope grounded in the 2025 Kohoutova et al. NMR/chemical-shift perturbation structural data (doi:10.1038/s41467-025-59106-5).
- kill_criteria
Abort or revise if: 9IYE NMR ensemble: p53-TAD chain may be a short disordered peptide — confirm chain ID and residue completeness before contact enumeration; ProteinMPNN GPU not available in sandbox — CPU inference will be slow; limit to 32 sequences to stay within 18-min budget; ESM-2 baseline requires WT FOXO4-DBD segment sequence; extract from PDB SEQRES or UniProt P98177 residues 156-256 (DBD boundaries)