Abstract / Summary
Abstract Background. p62/SQSTM1 activates NRF2 by sequestering KEAP1 through its KEAP1-interaction region (KIR) motif, which binds the KEAP1 Kelch domain and competes with NRF2 for KEAP1. In KRAS-mutant colorectal cancer (CRC), p62-mediated KEAP1 sequestration contributes to constitutive NRF2 activity and ferroptosis resistance. A small molecule that blocked the p62 KIR site could free KEAP1 to ubiquitinate NRF2, restore ferroptosis sensitivity, and complement SLC7A11 and GPX4 inhibition. Methods. We docked 38 compounds against the p62 KIR site on the KEAP1 Kelch domain (PDB 3ADE, 2.2 Å, chain A) and counter-screened against the NRF2 ETGE site (PDB 2FLU). The set spanned seven pharmacophore categories: KIR peptide mimetics, arginine-anchoring scaffolds, tyrosine-stacking compounds, fragment probes, reference Kelch inhibitors, de novo PPI scaffolds, and cross-project ferroptosis candidates. Seventeen p62 KIR contact residues (Tyr334 to Ser602) were mapped from the 3ADE co-crystal. All compounds were characterised for Lipinski compliance, PAINS status and QED. Results. Top compounds scored well at the p62 KIR site (− 7.10 to − 10.20 kcal/mol) but scored almost identically at the NRF2 ETGE counter-screen: 34 of 38 compounds gave |Δ| < 1.0 kcal/mol, mean Δ − 0.38 kcal/mol. The two Kelch peptide-binding surfaces are therefore hard to tell apart with small molecules, which is consistent with their shared contact residues but is also consistent with the overlapping, unequal docking boxes used here (see Discussion). We read the result as pan-Kelch binding rather than site-selective blockade, and note that a pan-Kelch compound would displace both p62 and NRF2. One compound, Arg_Anchor_004 (aminobenzothiazole-chlorobenzamide; p62 − 8.40, NRF2 − 7.30, Δ − 1.10 kcal/mol; Lipinski PASS, PAINS-clean, QED 0.757), was the only one to reach Δ ≤ −1.0. Cross-project candidate JKE-1674 scored − 8.90 kcal/mol at the p62 KIR site. Conclusions. Against this library the KEAP1 Kelch β-propeller does not distinguish the p62 KIR and NRF2 ETGE surfaces, so the tractable target here is a pan-Kelch inhibitor rather than a p62-selective one. Arg_Anchor_004 is the only drug-like compound with even modest p62 preference and is a fragment-based starting point. The pan-Kelch reading has a coherent rationale in KRAS-mutant CRC, where blocking both surfaces would suppress NRF2 by two routes at once, but the docking asymmetry means the selectivity conclusion needs a matched-box re-run and experimental confirmation.