I got this from rowan/boltz:
- NQO1 P187S + FAD: Cysteamine is the real structure winner. It has tight probe geometry (0.451 Å RMSD), tight FAD
geometry (0.221 Å RMSD), and a much closer recurring cofactor-probe distance (3.244 Å) with better complex_ipde
(0.643). Piceatannol is still the score winner (binding_confidence 0.623), but structurally it is loose (15.383 Å
probe RMSD, 7.139 Å cofactor-probe distance).- GSTM2 + GSH: NAC is still the better practical lead. Cysteinylglycine is slightly tighter geometrically (0.879 Å vs
1.404 Å probe RMSD), but NAC keeps the better score (0.570 vs 0.456) with similarly stable GSH geometry.So the clean current conclusion is:
- best defective NQO1 rescue-style candidate from this panel: Cysteamine
- best GSTM2 + GSH compensation candidate from this panel: NAC
If you want to keep moving, the next sensible step is not another broad screen. It’s to package the “current winners”
set and/or push just Cysteamine and NAC into the next stage.
Cysteamine is not “fixing NQO1” in the same way FAD rescue would. It is a small aminothiol, so if it shows up as the cleaner-geometry winner in NQO1 P187S + FAD, the most plausible read is:
- it is a better direct chemical trap / sacrificial nucleophile candidate
- it can fit near the stressed quinone-facing region without needing a big, fussy pose
- it may be more compatible with a damaged or floppy mutant holo-state than a bulkier polyphenol like piceatannol