Oncogenic Cysteines Create Peroxide-Sensitive Regulatory States That Impair GAP-Mediated Shutoff in KRAS

Abstract Covalent KRAS-G12C inhibitors exploit an acquired cysteine in the nucleotide-binding P-loop, and prior work established that oxidation can limit inhibitor engagement. Here we asked whether peroxide-dependent cysteine chemistry also affects GAP-mediated KRAS shutoff. Using purified KRAS variants, chemoselective sulfenic- and sulfinic-acid probes, intact-protein mass spectrometry, nucleotide-displacement and exchange assays, multiple-turnover phosphate-release assays, and multireplica molecular dynamics simulations, we compared WT-KRAS, G12C, G13C, and selected C118S backgrounds. G12C-containing variants accumulated strong BTD-trappable sulfenic-acid signal that persisted after C118 substitution, supporting C12 as its principal source. DiaAlk profiling identified a substantial C118-associated sulfination component and residual C118-independent labeling in G13C. GDP-bound simulations identified differences in phosphate proximity and local ion association, providing structural context for the distinct chemical profiles. Nucleotide-displacement responses were variant- and treatment-dependent, and detectable SOS1-mediated exchange persisted. For the functional assay, KRAS peroxide pretreatment was followed by complete catalase-mediated peroxide removal before GAP-334 addition. The GAP-dependent apparent phosphate-release slope increment was markedly reduced in G12C and G13C, whereas WT retained a substantial response. This differential response persisted after normalization to each variant’s own untreated GAP increment. The simplest interpretation is that cysteine oxidation and accompanying structural changes impair GAP responsiveness in these oncogenic variants. These findings identify a peroxide-sensitive dimension of KRAS regulation without assigning the functional effect to a single oxidation species or elementary catalytic step.