Fig 2: Knockdown of MBTPS2 and effects on key genes in cholesterol and fatty acid synthesis.a Differential expression of significant genes from RNA-Seq and/or qRT-PCR validation of RNA-Seq comparing LNCaP siMBTPS2 to LNCaP NTS. Blue arrow indicates decreased expression. ****P < 0.001, ***P < 0.005, **P < 0.01, *P < 0.05. b Correlation of MBTPS2 expression with regulators of cholesterol and fatty acid synthesis genes from CBioPortal using SU2C/PCF Dream Team metastatic cohort(31).

Fig 3: siRNA targeted knockdown of MBTPS2 downregulates the Cholesterol and Lipid synthesis pathways.a Gene set enrichment analysis of Hallmark fatty acid metabolism comparing LNCaP NTS vs LNCaP siMBTPS2 (n = 3 independent experiments, normalised enrichment score (NES) = -1.17). b ICC images representative of FASN staining after NTS, siMBTPS2 or siSREBP treatment in LNCaP cells (FASN shown in red, DAPI in blue). c Analysis of ICC imaging of FASN (n = 3 independent experiments with >40 cells imaged per experiment, error bars are ±SEM, ****p < 0.001, ANOVA with Tukey’s analysis. d Gene set enrichment analysis of GOBP regulation of cholesterol metabolic processes comparing LNCaP NTS vs LNCaP siMBTPS2 (n = 3 independent experiments, NES = -1.30). e Representative images of Filipin-based assay to detect cholesterol in LNCaP cells treated with NTS, siMBTPS2 or siSREBP siRNA. Cells were treated for 48 h and kept in either medium supplemented with cholesterol, or serum-free medium for 24 h prior to fixing. f Analysis of filipin staining (n = 3 independent experiments with >40 cells imaged per experiment, error bars are ±SEM, ***p < 0.005, ****p < 0.0001 ANOVA with Tukey’s analysis).

Fig 4: Identification of MBTPS2 by Sleeping Beauty transposon screen on a background of PtenNull-driven prostate cancer.a RT-PCR for Mbtps1 expression in prostate tumours from PtenNull, Mbpts2WT and Mbpts2INT mice (n = 4 mice for PtenNull, n = 3 mice for Mbtps2WT and Mbtps2INT, *p < 0.016; unpaired t-test). b Kaplan–Meier (log-rank) curve demonstrating reduced survival in the Mbtps2INT (n = 3) mice in comparison to Mbtps2WT (n = 18) mice, P < 0.0001; log-rank (Mantel–Cox) test. c Quantification of Ki67-positive cells in Mbtps2INT prostate tumours (n = 3 mice, minimum cell count of 100 cells per mouse) demonstrated increased levels of proliferation compared with tissue from Mbtps2WT tumours (n = 9, minimum cell count of 100 cells per mouse), **p = 0.0091; Mann–Whitney). d Percentages of mice with lung and lymph node metastases in Mbtps2INT mice compared to Mbtps2WT mice tested by Fisher’s exact test. e MBTPS2, SREBF1 and SREBF2 alterations visualised on CBioPortal using TCGA Firehose Legacy cohort for primary PC, and SU2C/PCF Dream Team(31) for the metastatic PC cohort. f Kaplan–Meier (log-rank) curve using TCGA Firehose Legacy cohort for primary PC, and SU2C/PCF Dream Team(31) showing patients with MBTPS2, SREBF1 or SREBF2 alterations (N = 31) and without these alterations (N = 594). P = 0.0002; log-rank (Mantel–Cox) test.

Fig 5: Diagrammatic representation of the ATF6 and SREBP activation by MBTPS2.ATF6 and SREBP are synthesised in their inactive precursor forms and are embedded in the endoplasmic reticulum. ER stress and lipid/cholesterol deprivation can trigger the release of ATF6 and SREBPs, respectively. These inactive proteins are trafficked to the Golgi where they are sequentially cleaved by S1P and MBTPS2. Cleavage results in the release of mature, transcriptionally active proteins which are then shuttled to the nucleus. Here, ATF6 can bind and activate expression of ER-stress response genes, whereas SREBP activate expression of lipid and cholesterol metabolism.