Fig 2: β-catenin palmitoylation at C466 coordinates an immunosuppressive program by conferring resistance to immunogenic ferroptosis and upregulating PD-L1(A) Uniform manifold approximation and projection plot of CRC single-cell RNA sequencing data (GSE132465), showing the clustering patterns of cells. ZDHHC5 expression is represented by color intensity, with a scale from low to high.(B) Significant positive correlation between ZDHHC5 and PD-L1 expression. Correlation analysis of ZDHHC5 and PD-L1 expression at both average (left) and relative (right) expression levels, based on (A).(C) Significant negative correlation between ZDHHC5 and CD8+ T cell infiltration. Correlation analysis of CD8A and ZDHHC5 expression at both average (left) and relative (right) expression levels, based on (A).(D) Significant negative correlation between ZDHHC5 expression in tumor epithelial cells and HLA-A expression in DCs. Correlation analysis was performed using the average expression levels of ZDHHC5 in tumor epithelial cells and HLA-A in DCs, based on (A).(E) β-catenin C466 palmitoylation is crucial for PD-L1 transactivation and suppression of released HMGB1. CTNNB1-knockout HCT116 cells were re-expressed with WT or C466A-mutant β-catenin, followed by qPCR analysis (left) and ELISA of culture supernatants (right).(F) PD-L1 expression is dependent on the β-catenin/TCF4 complex. CTNNB1-knockout HCT116 cells were re-expressed with WT or C466A-mutant β-catenin, treated with iCRT4 (100 μM) for 24 h, and analyzed by qPCR.(G) HMGB1 release is required for the efficacy of PD-L1 blockade in the context of β-catenin signaling suppression. Ctnnb1-knockdown MC38 cells reconstituted with C466A β-catenin or both C466A β-catenin and Hmgb1 knockdown were injected into immunocompetent C57BL/6J mice, and tumor volume was monitored after administration of 100 μg anti-PD-L1 on days 14 and 18.(H) Quantification of tumor weight from (G).(I) Relative levels of HMGB1 in plasma from C57BL/6 mice bearing MC38-derived xenografts shown in (G), measured using a mouse HMGB1 ELISA kit.Data are presented as mean ± SD; statistical significance was determined by Student’s t test; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; ns, not significant.

Fig 3: Competitive palmitoylation at C466 displaces S-nitrosylation to potentiate β-catenin transcriptional activity(A) Nuclear accumulation of β-catenin requires palmitoylation at C466. WT β-catenin or its C466A mutant was re-expressed in CTNNB1-knockout HCT116 cells, followed by immunofluorescence assays (left). Quantification of nuclear β-catenin intensity (right). Scale bars, 10 μm.(B) Schematic of competitive PTM crosstalk. Model depicting palmitoylation and S-nitrosylation competition at β-catenin C466.(C) β-Catenin palmitoylation inhibits its S-nitrosylation. Stable ZDHHC5 knockdown in HEK293T cells was performed, and the cells were evenly divided into two groups. One group underwent ABE analysis, and the other group was subjected to S-nitrosylation detection.(D) Palmitoylation of β-catenin at C466 inhibits S-nitrosylation at the same site. ZDHHC5 was co-expressed with WT β-catenin or C466A mutant in HEK293T cells, followed by ABE and S-nitrosylation detection assays.(E) A computational model of the β-catenin/TCF4 complex with palmitoylation modeled at β-catenin C466. The structural model is derived from the experimental co-crystal structure (PDB ID: 1JPW). (Top) Surface representation of the overall complex, with β-catenin in blue and TCF4 in orange. A black box highlights the location of the C466 residue. (Bottom) Close-up view of the C466 site. The palmitoyl moiety (green sticks) was modeled onto CYS 466 (yellow sticks) and is shown occupying a pocket at the β-catenin/TCF4 interface. Key surrounding amino acid residues from β-catenin (e.g., LYS 508) and TCF4 (e.g., GLU 17, PHE 21) are displayed as sticks. See Figure S4 for details.(F) CoIP analysis of the β-catenin/TCF4 interaction following ZDHHC5 knockdown. HA-β-catenin, FLAG-TCF4, and shZDHHC5 were co-expressed into HEK293T cells, which were treated with DMSO or iCRT4 (100 μM), followed by coIP assays.Data are presented as mean ± SD; statistical significance was determined by Student’s t test; ∗p < 0.05.

Fig 4: β-cat-oxazole, a first-in-class inhibitor of the ZDHHC5-β-catenin interaction, suppresses colorectal tumorigenesis(A) Virtual screening process to identify compounds targeting the ZDHHC5/β-catenin complex.(B) Cell viability of HCT116 cells after treatment with virtual screening candidates in combination with RSL3 (0.05 μM) for 5 days. Large dots represent the mean.(C) Chemical structure of β-cat-oxazole.(D) β-cat-oxazole induces cell death in a β-catenin-dependent manner. WT or CTNNB1-knockout HCT116 cells were treated with 10 μM β-cat-oxazole for 2 days, followed by trypan blue staining.(E and F) Ferrostatin-1 rescues β-cat-oxazole-induced cell death and HMGB1 release. HCT116 cells were treated with 5 μM β-cat-oxazole and 10 μM ferrostatin-1 for 3 days, followed by trypan blue staining (E) and measurement of HMGB1 levels in culture supernatants by ELISA (F).(G) SPR analysis of β-cat-oxazole binding to β-catenin.(H) β-cat-oxazole dose dependently disrupts the ZDHHC5-β-catenin interaction. HEK293T cells transfected with FLAG-ZDHHC5 were treated with varying concentrations of β-cat-oxazole (0, 5, 10, and 20 μM) for 2 days, followed by coIP analysis.(I) β-cat-oxazole inhibits β-catenin palmitoylation. HEK293T cells were treated with 10 μM β-cat-oxazole for 2 days, followed by ABE assay (left). Quantification of β-catenin palmitoylation from the left panel (right).(J) Schematic of the AOM/DSS-induced CRC mouse model. Mice were treated with 20 mg/kg β-cat-oxazole intraperitoneally every 3 days for 8 doses.(K–M) Representative macroscopic morphologies (K), tumor numbers (L), and tumor volume (M) in DMSO- or β-cat-oxazole-treated mice.(N) β-cat-oxazole treatment promotes CD8+ T cell infiltration in mouse CRC. AOM/DSS-induced CRC tumors in mice treated with β-cat-oxazole were subjected to immunofluorescence analysis.Data are presented as mean ± SD; statistical significance was determined by Student’s t test; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.