Mitochondrial KMT9 methylates DLAT to control pyruvate dehydrogenase activity and prostate cancer growth

Abstract

Prostate cancer (PCa) growth depends on de novo lipogenesis controlled by the mitochondrial pyruvate dehydrogenase complex (PDC). In this study, we identify lysine methyltransferase (KMT)9 as a regulator of PDC activity. KMT9 is localized in mitochondria of PCa cells, but not in mitochondria of other tumor cell types. Mitochondrial KMT9 regulates PDC activity by monomethylation of its subunit dihydrolipoamide transacetylase (DLAT) at lysine 596. Depletion of KMT9 compromises PDC activity, de novo lipogenesis, and PCa cell proliferation, both in vitro and in a PCa mouse model. Finally, in human patients, levels of mitochondrial KMT9 and DLAT K596me1 correlate with Gleason grade. Together, we present a mechanism of PDC regulation and an example of a histone methyltransferase with nuclear and mitochondrial functions. The dependency of PCa cells on mitochondrial KMT9 allows to develop therapeutic strategies to selectively fight PCa.

Fig. 1. KMT9 is localized in mitochondria and regulates PDC activity in PCa cells.

a Subcellular fractionation analysis of PC-3M or DU145 cells with quantification of mitochondrial KMT9α percentage. b Immunofluorescence staining of PC-3M cells showing KMT9α, mitochondria (MitoTracker), and nuclei (DAPI). Scale bar: 5 μm. c Proteinase K protection assay for mitochondria isolated from PC-3M cells. TFAM (mitochondrial matrix, MM), CYCS (intermembrane space, IMS), and TOMM20 (outer mitochondrial membrane, OMM) served as controls for compartmental digestion. d Import of [³⁵S]-labeled KMT9α and KMT9β into mitochondria isolated from PC-3M cells. Δψ, membrane potential; PK, proteinase K. e Presence of KMT9α and KMT9β in mitochondrial extracts of diverse cancer and non-cancer cell lines was revealed by Western blot analysis using the indicated antibodies. f Schematic illustration of KMT9α interactome in PC-3M cells. g Co-immunoprecipitation using antibodies against either the N-terminus (N) or the C-terminus (C) of KMT9α showing the interaction between DLAT and KMT9α in mitochondrial lysates. (Input: 5% of total extract). h Schematic illustration of global metabolomics analyses in siCtrl or siKMT9α-treated PC-3M cells. i–l Relative abundance of pyruvate (i), NAD⁺ (j), lactate (k), and acetyl-CoA (l) identified by global metabolomics analyses in siCtrl or siKMT9α-treated PC-3M cells. m Schematic showing pyruvate conversion to acetyl-CoA by PDC in mitochondria. n PDC activity in PC-3M cells expressing different KMT9α variants with or without endogenous KMT9α depletion. o Restoration of PDC activity in vitro. Extracts of PC-3M cells transfected with siCtrl or siKMT9α were supplemented with recombinant wildtype KMT9 heterodimer (rKMT9α/KMT9β) or catalytically inactive heterodimer (rKMT9α (N122A)/KMT9β). i–l, n, o Data are shown as box plots (i–l n = 6 biological replicates; center line indicates median, box bounds show 25th and 75th percentiles, whiskers represent minimum and maximum values within 1.5 times the interquartile range) or mean + SD (n, o n = 4 biological replicates). Statistical significance was determined by two-sided Welch’s two-sample t-test with false discovery rate (FDR) correction (i–l) or two-sided Student’s t-test (n, o). (a–e, g, n, o). All experiments were independently repeated at least three times with similar results. Source data are provided as a Source Data file.

Fig. 2. Mitochondrial KMT9 regulates PDC activity by monomethylation of DLAT at K596.

a Western blot analysis of PDC immunocaptured from mitochondrial fractions of siCtrl- or siKMT9α-treated PC-3M cells using anti-pan-methyllysine antibody. b Detection of DLAT K596me1 in mitochondrial fractions of DU145 cells treated with siCtrl or siKMT9α and transfected with expression plasmid for LacZ, KMT9α, or catalytically inactive KMT9α (N122A) by Western blot. Membranes were decorated with the indicated antibodies. TOMM20 served as loading control. c In vitro methylation of DLAT. PDC immunocaptured from DU145 cells was incubated with recombinant (r) KMT9α/β or rKMT9α (N122A)/KMT9β and S-adenosyl methionine (SAM). Methylation reactions were analyzed by Western blot using anti-DLAT K596me1 or anti-DLAT antibody. d Detection of DLAT K596me1, KMT9α, and KMT9β in mitochondrial extracts of indicated human specimens by Western blot. e Detection of DLAT K596me1 in nuclear, mitochondrial, or cytosolic fractions of DU145 cells transfected with siCtrl or siKMT9α and expression plasmid for LacZ, MTS-KMT9α, or NLS-KMT9α by Western blot. LMNA, TOMM20, and α-Tubulin served as controls for nuclear, mitochondrial, and cytosolic fractions, respectively. f, g Activity of PDC immunocaptured from extracts of PC-3M cells transfected with expression plasmid for LacZ, DLAT, DLAT (K596R), or DLAT (K547R) in combination with siCtrl or siDLAT. Knockdown efficiency and expression of exogenous DLAT proteins was verified by Western blot (f). Data represent mean + SD from 4 biological replicates. Statistical significance was determined by two-sided Student’s t-test. (a–g) All experiments were independently repeated at least three times with similar results. Source data are provided as a Source Data file.

Fig. 3. Mitochondrial KMT9 fuels PCa cell proliferation via regulation of de novo lipogenesis.

a–c Proliferation of DU145 cells transfected with siCtrl or siDLAT in combination with expression plasmid for LacZ (a, b), DLAT (a), or DLAT (K596R) (b) as indicated. Knockdown efficiency and expression of exogenous DLAT proteins was verified by Western blot (c). α-Tubulin served as loading control. d–i Proliferation of DU145 cells transfected with siCtrl or siKMT9α in combination with an expression plasmid for LacZ (d–h), KMT9α (d), MTS-KMT9α (e, g), NLS-KMT9α (f, g), or MTS-KMT9α (N122A), and NLS-KMT9α (N122A) (h). Knockdown efficiency and expression of exogenous KMT9α proteins was verified by Western blot (i). α-Tubulin served as loading control. j Proposed mechanism for KMT9-mediated control of mitochondrial functions in PCa cells. KMT9 monomethylates DLAT at K596, which is required for PDC activity and mitochondrial metabolism. KMT9 depletion abolishes DLAT K596 monomethylation, thereby impairing PDC activity and mitochondrial metabolism in PCa cells. k, l Free fatty acid (FFA) (k) and triglyceride (l) levels in DU145 cells transfected with siCtrl or siKMT9α in combination with expression plasmid for LacZ, NLS-KMT9α, MTS-KMT9α, or MTS-KMT9α (N122A). m, n Proliferation of DU145 (m) or HepG2 (n) cells transfected with sub-optimal concentrations of siCtrl or siKMT9α in the presence or absence of 100 µM SB204990 to detect potential effects of co-treatment. a, b, d–h, k–n, Data are presented as mean ± SD (a, b, d–h, m, n, n = 4 biological replicates) or mean + SD (k, l, n = 3 biological replicates). Statistical significance was determined by a two-sided Student’s t-test (a, b, d–h, k–n). All experiments were independently repeated at least three times with similar results. Source data are provided as a Source Data file.

Fig. 4. Mitochondrial KMT9 controls PDC activity and de novo lipogenesis in prostate tumors in mice.

a, b Representative mouse prostates (a) and prostate/body weight ratio (mg/g) from Ctrl (n = 6 mice), Pten/Smad4 KO (n = 4 mice), and Pten/Smad4/Kmt9α KO (n = 5 mice) mice (b). c, d Multiplexed immunofluorescence showing MKI67 and PCNA (c) or DLAT, TOMM20, KMT9α and DLAT K596me1 (d) in prostate tissue from indicated genotypes. Scale bar: 50 μm. e, f Quantification of mitochondrial KMT9α/TOMM20 (e) and DLAT K596me1/DLAT (f) ratios in prostate sections (Ctrl n = 179 cells, Pten/Smad4 KO n = 404 cells, Pten/Smad4/Kmt9α KO n = 182 cells). g, h Pearson correlation between DLAT K596me1/DLAT ratio and mitochondrial KMT9α/TOMM20 ratio (g) or nuclear KMT9α/DAPI ratio (h) in Pten/Smad4 KO prostate sections. R, Pearson correlation coefficient. i–k Analysis of PDC activity (i), free fatty acid levels (j), and triglyceride levels (k) in prostate tissue of Ctrl, Pten/Smad4 KO, and Pten/Smad4/Kmt9α KO mice (n = 6 mice per group). l, m H-score ratios of DLAT K596me1/DLAT (l) and mitochondrial KMT9α/TOMM20 (m) in normal (n = 16 tissue microarray cores), normal adjacent tissue (NAT, n = 16 tissue microarray cores), and PCa tissues of Gleason grades 1-5 (n = 18, 27, 10, 48, 52 tissue microarray cores respectively). n, o Pearson correlation between H-score ratios of DLAT K596me1/DLAT and mitochondrial KMT9α/TOMM20 (n) or nuclear KMT9α/DAPI (o) in prostate tissue sections of human patients. R, Pearson correlation coefficient. Original immunofluorescence images used for the quantification are shown in Supplementary Fig. 12b. b, e, f, i–m, Data represent mean ± SD in bar graphs (b, i–k) or box plots (e, f, l, m; center line: median; box bounds: 25th and 75th percentiles; whiskers: minimum and maximum values within 1.5 times the interquartile range). b, e–o Statistical significance was determined using two-sided Student’s t-test (b, e, f, I, i–m) or Pearson correlation analysis with two-sided p-values calculated based on the t-distribution with n-2 degrees of freedom (g, h, n, o). All experiments were independently repeated at least three times with similar results. Source data are provided as a Source Data file.