Data source and screening

We used Sangerbox software to download mRNA data of patients with KIRC from the TCGA database (http://www.sangerbox.com/tool). The Limma R package was used to screen for genes that were differentially expressed between KIRC and healthy samples at the significance level of P < 0.05.

Pan-cancer zDHHC3 gene expression and survival analyses

We used the online tool TIMER 2.0 (Li et al. 2020) to analyze differentially expressed genes (DEGs) in various cancers using TCGA data (https://genemania.org/). We also applied GEPIA (Tang et al. 2017), which uses TCGA and CTEx mRNA sequencing data, to examine differential gene expression between tumor and healthy tissues, and to conduct patient survival and gene expression correlation analyses (http://gepia.cancer-pku.cn/index.html).

Confirmation of zDHHC3 expression in GSE152938 data

We selected two KIRC samples and one healthy kidney tissue sample from the GSE152938 dataset to characterize zDHHC3 expression at the single-cell level. The Seurat R package was used to implement standard single-cell sequencing data processing pipelines. Filtering was performed to identify cells with < 200 and > 5000 genes and a mitochondrial gene percentage > 10%. Then, the screening was performed to identify DEGs between KIRC and healthy samples using the FindMarkers function.

GSEA of zDHHC3-related genes

GSEA is a powerful tool for the analysis of genome-wide expression profiles using chip data. Unlike differential analysis, GSEA does not require manual screening for the identification of DEGs, which may result in the missing of critical information. It can be used to identify gene sets that are not very different from one another, but have consistent difference trends. We performed a GSEA of zDHHC3-related genes using the clusterProfiler package, with h.all.v7.3.symbols.gmt serving as the reference gene set. P values < 0.05 were considered statistically significant.

WGCNA of zDHHC3-related genes

WGCNA (Langfelder et al. 2008) is a valuable tool for the evaluation of pairwise correlations between gene expression profiles, the clustering of genes with synergistic changes into modules, and the exploration of associations between gene modules and disease phenotypes. Its first steps are the detection of sample and gene quality, the cluster analysis of samples based on the gene expression, and the deletion of outliers. Then, a similarity matrix is constructed by calculating correlations between all gene pairs and the soft threshold power of β, and determining whether a scale-free network can be established. This matrix is converted into a topological overlap matrix, a dissimilarity matrix is constructed, and co-expressed gene modules are identified by dynamic tree cutting. Finally, highly similar modules are merged based on module-level correlation, and correlations between each module and the disease phenotype are calculated. The hub genes in the most relevant disease module are then screened.

Gene ontology and Kyoto encyclopedia of genes and genomes analyses

The Database for Annotation, Visualization and Integrated Discovery (Sherman et al. 2021) is an online tool for the analysis of gene and protein functions (DAVID, https://david.ncifcrf.gov/). The main biological function categories analyzed are cell compartment (CC) and molecular function (MF), as well as biological processes (BP). Metascape (Zhou et al. 2019) is another online analytical tool that integrates multiple databases for the examination of biological processes, signal pathways, and protein–protein and protein–drug interactions. P value < 0.05 was considered statistically significant (https://metascape.org/gp/index.html).

Prediction of target genes of zDHHC3

GENEMANIA (Franz et al. 2018) is a website containing data on 166,691 genes and 660,443,499 protein interactions (https://genemania.org/). It can be used to verify various aspects of protein interactions, including co-localization, co-expression, physical interactions, prediction, shared protein domains, and genetic interactions. The data on zDHHC3 and the genes related most closely to it were uploaded to GENEMANIA for the construction of protein–protein interaction (PPI) networks and identification of potential target genes of zDHHC3.

Confirmation of target gene expression with GSE213324 data

We selected 60 healthy kidney tissues and 63 RCC tissues from the GSE213324 dataset to characterize the expression of zDHHC3 target genes. The Limma R package was used to screen DEGs between the KIRC and healthy samples using the significance threshold of p < 0.05.

Coimmunoprecipitation

Caki-2 and RCC23 cells were lysed with binding buffer containing a protein inhibitor cocktail (1:100), and the protein concentration was determined after centrifugation at 14,000×g at 4 ℃ for 10 min. Protein samples (500 μg) were mixed with 3 μg mouse zDHHC3 monoclonal antibody (sc-377378; Santa Cruz) overnight at 4 °C. Then, the antibody–protein complex was mixed with 50 μl protein A/G magnetic beads for 2 h, and washed three times with washing buffer to separate unbound antibodies and proteins. Binding and protein loading buffers (1:4) were then added and the samples were heated at 95 ℃ for 5 min before analysis by western blotting.

Short hairpin RNA design and transfection

The short hairpin RNA (shRNA) of SLC9A2 and zDHHC3 were synthesized by WZBIO (Wuhan, China). The sequences of SLC9A2 and zDHHC3 were as follows, SLC9A2: 5’-CGCCCATTCTTTGAGAACATT-3’, zDHHC3:5’-CCCAAAGGAAATGCCACTAAA-3’, and non-targeted control shRNA served as the negative control (NC). Cells were cultured in Dulbecco’s modified Eagle medium in a 6-cm cell culture dish, and 10 ul shRNA was added at 60–80% confluence. The mixture was incubated at 37 ℃ for 48 h, and the cells were harvested for further analysis.

Kidney tissue collection

Kidney tissues were collected from seven healthy individuals and seven patients with first-diagnosis KIRC aged 18–70 years at Remnin Hospital of Wuhan University between December 2022 and February 2023. KIRC diagnoses were confirmed by clinical and imaging examinations and histopathological analysis of biopsy or surgical samples. The patients did not receive radiotherapy, chemotherapy, or other anti-tumor treatment before admission. Patients with incomplete clinical data were excluded.

Immunofluorescence

Paraffin sections of kidney tissues were roasted at 65 ℃ for 45 min to remove wax, and then antigen repair was performed. Sealed with 1% bovine serum albumin solution at room temperature for 1 h, sections were incubated with antibodies (zDHHC3,sc-377378; SLC9A2, 46,256–1; 1:100 dilution) at 4 ℃ overnight and then incubated with fluorescent-labeled secondary antibodies for 1 h. Finally, the nucleus was stained with DAPI for 10 min. The sections were sealed with an anti-fluorescent quencher and stored at 4 °C.

Acyl–biotin exchange assay

Briefly, protein was enriched by antibody and protein A/G magnetic beads (B23202; Bimake), then incubated with 50 mM N-ethylmaleimide (N80860; Macklin) at room temperature for 4 h. Samples were divided into two equal parts and added 1 M Hydroxylamine (H828371; Macklin) and NaCl, respectively. Next, the samples were incubated with Biotin-HPDP (5 μM, A8008; Apexbio) at room temperature for 1 h, then added 1 × loading buffer (without mercaptoethanol) and heated at 95 ℃ for 5 min, and subsequent western blot analysis was performed.

Western blot

The concentration of protein extracted from cells was determined by BCA assay kit, then sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) was used to separate proteins. The proteins were electrotransferred to polyvinylidene difluoride (PVDF) membranes and then blocked with 5% skim milk at room temperature for 1 h. Next, membranes were incubated with primary antibody at 4℃ overnight, and the secondary antibody was incubated at room temperature for 1 h. The blots were visualized and analyzed with chemiluminescent reagents and ImageJ software. The following primary antibodies were used: zDHHC3 (sc-377378, 1:500; Santa), P21-activated kinase 7 (PAK7; 35,292–1, 1:500; SAB), solute carrier family 9 member A2 (SLC9A2, also sodium–hydrogen exchanger 2; 46,256–1, 1:500; SAB), caspase 3 (19,677–1-AP, 1:500; Proteintech), and β-actin (66,009–1-LG, 1:10,000; Proteintech).

Statistical analysis

SPSS 22.0 software was used for analysis and graphpad prism 8.0 software was used for plotting. The data in box plots are presented as means ± standard deviations and were analyzed using the unpaired Tukey test. All experiments were repeated successfully five times and yielded consistent results. The significance level was set to p < 0.05.