Stereochemical and Protonation Artefacts in Docking a Flexible Xanthophyll to the NLRP3 MCC950 Cleft: A Case Study with a Non-Natural Fucoxanthin Stereoisomer and Neutral versus Anionic MCC950

Abstract Background: The NLRP3 inhibitor MCC950 is a compact, anionic sulfonylurea. Fucoxanthin, a large, flexible, neutral xanthophyll, has been reported to suppress NLRP3 activation in cells, but whether it can occupy the MCC950 cleft has not been tested structurally, and docking such ligands is prone to stereochemical and protonation artefacts. Methods: An exploratory in silico case study used rDock docking into the MCC950 cleft of NLRP3 (PDB 7PZC, 3.90 Å), with an independent test in the 2.84 Å structure 7ALV (template: the MCC950 analogue RM5), MM-GBSA on 15 short replicates per ligand, explicit-solvent simulations (including a four-system pilot with and without ADP) and anchor-geometry measurements in seven deposited NLRP3–inhibitor structures. Results: (1) Neutral MCC950 engaged only Arg578, whereas its anion restored the Arg351/Arg578 pairing in docking and in 11 of 15 short replicates (8 of 15 for neutral MCC950). (2) The fucoxanthin model, built from a SMILES string without stereochemistry, was a non-natural isomer (CIP multiset S4R1; three Z chain bonds); its best score (−21.76 rDock units) was not reproduced with regenerated conformer pools, and its MM-GBSA advantage over neutral MCC950 (−20.8 kcal/mol) fell to −13.5 kcal/mol against the anion, before entropy. It reproduced both anchors in 2 of 15 replicates. (3) In rigid docking no all-E structure, including the allene hand given by the CCD name, was placed without repulsion in the 7PZC cleft (0 of 260 poses). In 7ALV an all-E input (allene hand not audited) lay at the cavity rim (14 of 100 poses scored negative) in a template-defined cavity (407 ų; longest extent 20.2 Å) shorter than the ligand (25–31 Å). (4) Two-arginine contact is not general in this cleft: MCC950 lies at the contact boundary of Arg351 (3.88 Å). Starting pose and Arg351 rotamer shifted Arg351 distances by about 3 Å, and single-trajectory pilots from unmatched poses did not rank ligands, isomers or ADP states. Conclusions: This is a methodological case study, not a test of natural fucoxanthin, which was not simulated; whether it or fucoxanthinol binds NLRP3 remains open. In rigid docking and by cavity geometry, all-E fucoxanthin was not accommodated by the MCC950 cleft as modelled, which does not exclude induced fit or binding elsewhere. Protonation-validated benchmarks, audited stereochemistry, starting poses of matched quality and replicate trajectories are advisable for such evaluations (not validated here), and experimental binding data are required for any biological claim.