SUCLG1 restricts POLRMT succinylation to enhance mitochondrial biogenesis and leukemia progression

Abstract

Mitochondria are cellular powerhouses that generate energy through the electron transport chain (ETC). The mitochondrial genome (mtDNA) encodes essential ETC proteins in a compartmentalized manner, however, the mechanism underlying metabolic regulation of mtDNA function remains unknown. Here, we report that expression of tricarboxylic acid cycle enzyme succinate-CoA ligase SUCLG1 strongly correlates with ETC genes across various TCGA cancer transcriptomes. Mechanistically, SUCLG1 restricts succinyl-CoA levels to suppress the succinylation of mitochondrial RNA polymerase (POLRMT). Lysine 622 succinylation disrupts the interaction of POLRMT with mtDNA and mitochondrial transcription factors. SUCLG1-mediated POLRMT hyposuccinylation maintains mtDNA transcription, mitochondrial biogenesis, and leukemia cell proliferation. Specifically, leukemia-promoting FMS-like tyrosine kinase 3 (FLT3) mutations modulate nuclear transcription and upregulate SUCLG1 expression to reduce succinyl-CoA and POLRMT succinylation, resulting in enhanced mitobiogenesis. In line, genetic depletion of POLRMT or SUCLG1 significantly delays disease progression in mouse and humanized leukemia models. Importantly, succinyl-CoA level and POLRMT succinylation are downregulated in FLT3-mutated clinical leukemia samples, linking enhanced mitobiogenesis to cancer progression. Together, SUCLG1 connects succinyl-CoA with POLRMT succinylation to modulate mitochondrial function and cancer development.

Keywords: FMS-like Tyrosine Kinase 3; Lysine Succinylation; Mitochondrial Biogenesis; Mitochondrial RNA Polymerase; Succinate-CoA Ligase.

Figure 1. SUCLG1 maintains mtDNA-encoded gene expression and mitochondrial mass.

(A) Spearman’s correlation of mitochondrial metabolic genes with ETC genes was analyzed in various TCGA transcriptomes. (B) List of top ten mitochondrial metabolic genes that positively correlate with ETC genes expression in pan-cancer analysis. Shown are Spearman’s correlation coefficients. (C) Scrambled control or shRNAs against SUCLG1 were transduced into human CD34⁺ cord blood (CB) cells or cancer cell lines as indicated. Mitochondrial mass was quantified by MTG staining. Shown is the relative MTG fluorescence intensity compared to scrambled control group. n = 3 independent biological replicates. (D, E) Total RNA was extracted from control or stable SUCLG1-knockdown HL60 and MV411 cells. Relative mRNA expression of mtDNA and nDNA-encoded genes was quantified by qPCR and displayed on a log₂ scale (D). n = 3 independent biological replicates. Protein expression was determined by western blotting (E). (F, G) Control and SUCLG1-knockdown MV411 cells were analyzed with transmission electron microscopy. Mitochondria numbers in 100 cells were determined (F). Shown were representative images (G), arrowheads indicate mitochondria. Scale bars: 1 μm. Data represent means ± SD, n = 100 different cells, t test. **P < 0.01. (H) Mitochondria isolated from control or SUCLG1-knockdown MV411 cells were subjected to ETC complex activity assay. Complex activities were normalized to total mitochondrial protein. Data represent means ± SD, n = 3 independent biological replicates, t test. **P < 0.01; *P < 0.05. (I–L) Wild-type SUCLG1 or its P170R (PR) and A209E (AE) mutants were re-expressed in SUCLG1-knockdown MV411 cells. Protein expression of mtDNA and nDNA-encoded genes was determined by western blotting (I). mtDNA abundance (J), mRNA expression levels (K), and oxygen consumption rates (L) of stable cells were determined. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; n.s. not significant. Source data are available online for this figure.

Figure 2. SUCLG1 reduces succinyl-CoA level to restrict POLRMT succinylation.

(A) Schematic overview of reactions catalyzed by succinyl-CoA ligase complex. SDHA succinate dehydrogenase subunit A, OGDH oxoglutarate dehydrogenase, SUCLG2 succinate-CoA ligase [GDP-forming] subunit beta, Pi phosphate. (B) Metabolites were extracted from control or SUCLG1-knockdown MV411 cells, and quantified by mass spectrometry. The relative abundance of indicated metabolites was shown on a log₂ scale. n = 3 independent biological replicates. (C) Extracellular acidification rates of control or SUCLG1-knockdown MV411 cells were determined. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; n.s. not significant. (D, E) Glucose (Glc) consumption and lactate production were assayed in control and SUCLG1-knockdown MV411 cells. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01. (F) ATP abundance was quantified in control and SUCLG1-knockdown MV411 cells, and normalized to total protein. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01. (G) Glutamine (Gln) consumption was assayed in control and SUCLG1-knockdown MV411 cells. Data represent means ± SD, n = 3 independent biological replicates, t test. *P  <  0.05. (H) Metabolites were extracted from SUCLG1-knockdown and rescue MV411 cells, and quantified by mass spectrometry. The relative abundance of indicated metabolites was shown on a log₂ scale. n = 3 independent biological replicates. (I) Total RNA was extracted to determine the mRNA expression of succinate-related metabolic genes in control and SUCLG1-knockdown cells. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; *P  <  0.05; n.s. not significant. (J) Mitochondria were isolated from stable MV411 cells. Metabolites were extracted to determine succinyl-CoA abundance. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; n.s. not significant. (K) Endogenous SUCLG1 was immunoprecipitated from MV411 cells. The protein interaction of SUCLG1 was determined by western blotting. IgG was included as a negative control for immunoprecipitation. SUCLG1-knockdown cells were included to validate the specificity of SUCLG1 antibody. (L, M) Endogenous POLRMT was immunoprecipitated from control or SUCLG1-knockdown MV411 (L) and HL60 (M) cells. Lysine succinylation was determined by western blotting. IgG was included as a negative control for immunoprecipitation. (N) Immunopurified POLRMT-HA protein was incubated with succinyl-CoA (sucCoA) or succinate. Flag-tagged SUCLG1 protein was immunopurified and treated with or without heat-inactivation (95 °C for 10 min). POLRMT succinylation was determined after in vitro succinylation. (O) Endogenous POLRMT was immunoprecipitated from SUCLG1-knockdown and rescue MV411 cells. Lysine succinylation was determined by western blotting. IgG was included as a negative control for immunoprecipitation. (P–R) Scrambled control or SUCLG1-knockdown (sh-#1) MV411 cells were cultured with 1 mM pyruvate or 0.2 mM uridine for 24 h. Oxygen consumption rates were determined. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; *P < 0.05; n.s. not significant. (S–V) Scrambled control or SUCLG1-knockdown (sh-#1) MV411 cells were cultured with 1 mM pyruvate or 0.2 mM uridine. Cell proliferation was determined. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; n.s. not significant. Source data are available online for this figure.

Figure 3. K622 succinylation suppresses POLRMT activity.

(A) Structural illustration of POLRMT. Cyan indicates POLRMT protein, and the sugar-phosphate backbone of DNA-RNA hybrid was colored brown (PDB ID:4BOC). (B–E) HA-tagged POLRMT and its mutants were re-expressed in POLRMT-knockdown MV411 cells. Cells were further transduced with scrambled control or shRNA against SUCLG1. POLRMT-HA was immunopurified to determine its succinylation and protein interaction (B). MitoChIP was performed to evaluate mtDNA binding of POLRMT-HA (C). Mitochondria were isolated to evaluate in organello labeling of nascent RNA (D). UTP labeling was quantified and normalized to mtDNA (E). Data represent means ± SD, n = 3 independent biological replicates, one-way ANOVA with Dunnett’s multiple comparisons test. **P  <  0.01; n.s. not significant. (F–H) Relative mtDNA abundance (F), mRNA expression levels of mtDNA-encoded genes (G), and oxygen consumption rates (H) were determined in stable MV411 cells. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; *P  <  0.05; n.s. not significant. (I–L) Scrambled control or OGDH-knockdown stable MV411 cells were cultured with or without 50 μM dimethyl succinate (DMS). K622 succinylation of immunopurified endogenous POLRMT (I), mtDNA-encoded gene expression (J), MTG intensity (K), and oxygen consumption (L) were determined. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; *P  <  0.05; n.s. not significant. (M) Correlation analysis of MRPL45 and SUCLG1 mRNA levels in cancer cell line encyclopedia. (N) Relative ATP and GTP abundance in indicated cell lines were quantified and normalized to total protein. Data represent means ± SD, n = 3 independent biological replicates. (O) K622 succinylation of immunopurified endogenous POLRMT was determined in indicated cell lines. Source data are available online for this figure.

Figure 4. Leukemia-derived FLT3 mutants upregulate SUCLG1 to boost mitobiogenesis.

(A) A panel of AML cells was subjected to western blotting to determine protein expression. CD34⁺ CB cells from three different donors were included as normal controls. K622 succinylation (K622sc) of immunopurified endogenous POLRMT was assayed. AML cell lines carrying FLT3 mutations were colored red. IgG was included as a negative control for immunoprecipitation. Ratios of K622sc and SUCLG1 were normalized to POLRMT protein and Actin, respectively. (B, C) Metabolites were extracted to determine cellular succinyl-CoA levels. The correlation of POLRMT^K622sc with SUCLG1 protein (B) and succinyl-CoA abundance (C) were analyzed. Pearson’s correlation. (D–H) HA-tagged FLT3 and its mutants were stably expressed in CD34⁺ CB cells. Endogenous POLRMT was immunopurified to determine K622 succinylation. Ratios of K622sc were normalized to POLRMT protein (D). Cellular succinyl-CoA levels (E), mtDNA abundance (F), mtDNA-encoded gene expression (G), and MTG intensity (H) were determined. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; n.s. not significant. (I–K) shRNA against SUCLG1 (#1) was stably expressed in control or FLT3^ITD-expressing cells (I). Oxygen consumption (J) and activities of ETC complexes (K) were determined. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; *P < 0.05. (L–P) shRNAs against FLT3 were stably expressed in MV411 cells. Endogenous POLRMT was immunopurified to determine K622 succinylation. Ratios of K622sc were normalized to POLRMT protein (L). Cellular succinyl-CoA levels (M), mtDNA abundance (N), mtDNA-encoded gene expression (O), and MTG intensity (P) were determined. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; *P < 0.05. (Q, R) Scrambled control or shRNAs targeting SUCLG1 (#1) or FLT3 (#1) were co-expressed in MV411 cells. Oxygen consumption rates (Q) and activities of ETC complexes (R) in stable cells were determined. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; *P < 0.05; n.s. not significant. Source data are available online for this figure.

Figure 5. FLT3 signaling modulates nuclear transcription to enhance mitochondrial respiration.

(A, B) Shown are differentially expressed mitochondrial genes in control and FLT3-knockdown cells. Data were extracted from GSE226461, n = 3 independent biological replicates, t test. (C) Venn diagram showing the numbers of significantly downregulated mitochondrial genes in MV411 and MOLM14 cells after depleting FLT3, n = 3 independent biological replicates, t test. (D) mRNA expression of 25 overlapped genes in (C) was validated in FLT3-knockdown MV411 cells. CS was included as a negative control. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; *P < 0.05; n.s. not significant. (E–G) Shown are the consensus sequences in the promoter regions of 25 mitochondrial genes regulated by FLT3. (H) shRNAs targeting FLT3-regulated transcription factors were transduced into MV411 cells. SUCLG1 mRNA levels were determined by qPCR. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; *P < 0.05; n.s. not significant. (I) Scrambled control or E2F1-knockdown MV411 cells were treated with or without FLT3 inhibitor (FLT3-IN-3). K622 succinylation of endogenous POLRMT was determined by western blotting. (J) Schematic overview of SUCLG1 promoter and TSS region. (K, L) Control IgG or E2F1 antibodies were used for chromatin immunoprecipitation. Scrambled control or E2F1-knockdown MV411 or MOLM14 cells were subjected to ChIP-qPCR assay. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01. (M) Scrambled control or E2F1-knockdown MV411 cells were treated with or without FLT3 inhibitor (FLT3-IN-3). Oxygen consumption rates were determined. Data represent means ± SD, n = 3 independent biological replicates, t test was performed between untreated and FLT3i-treated groups. **P  <  0.01; n.s. not significant. Source data are available online for this figure.

Figure 6. SUCLG1 suppresses POLRMT succinylation to support leukemic proliferation.

(A, B) SUCLG1-knockdown and rescue MV411 cells were subjected to cell proliferation assay (A). Cells were further treated with POLRMT inhibitor (POLRMTi) to determine viability (B). Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; n.s. not significant. (C, D) POLRMT-knockdown and rescue MV411 cells were subjected to cell proliferation assay (C). shRNA targeting SUCLG1 (#1) was stably expressed to determine cell proliferation (D). Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; n.s. not significant. (E, F) Leukemia cells or CD34⁺ CB cells were cultured with increasing doses of rotenone as indicated. Cell viability was determined after three days of culture. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; *P < 0.05; n.s. not significant. (G) Schematic overview showing the establishment of mouse MLL-AF9 + FLT3^ITD leukemia. (H, I) MLL-AF9 bone marrow cells were transduced with vector control or HA-tagged FLT3^ITD. Scrambled control or shRNAs targeting Suclg1 was stably expressed. Protein expression was determined by western blotting (H). Animal survival was assayed after transplanting leukemic bone marrow (I). n = 9 mice, log-rank (Mantel–Cox) test. **P  <  0.01. (J) POLRMT-knockdown and rescue MV411 cells were transduced with scrambled control or shRNA against SUCLG1. Cells were transplanted into sublethally irradiated NSG mice. Animal survival was determined. n = 5 mice, log-rank (Mantel–Cox) test. **P  <  0.01. (K–M) Human clinical AML bone marrow samples were collected. Metabolites were extracted to determine succinyl-CoA abundance (K). K622 succinylation of immunopurified POLRMT (L) and SUCLG1 protein expression (M) were determined by western blotting. n = 5–10, t test. **P  <  0.01; *P  <  0.05; n.s. not significant. (N–P) mRNA expression of POLRMT (N) and SUCLG1 (O) in FLT3^WT (n = 116) and FLT3^TKD/ITD (n = 38) TCGA AML samples were analyzed. t test. *P  <  0.05; n.s. not significant. TCGA AML patients were grouped by FLT3 mutation status and SUCLG1 mRNA expression level. Overall survival was determined (P). log-rank (Mantel–Cox) test. *P  <  0.05; n.s. not significant. (Q) Working model of SUCLG1-mediated succinyl-CoA signaling and FLT3-induced remodeling of mitobiogenesis. Source data are available online for this figure.

Figure EV1. SUCLG1 maintains mtDNA-encoded gene expression and mitochondrial mass.

(A) Schematic overview of TCA cycle and the reaction catalyzed by SUCLG1. (B) Gene set enrichment analysis of the relation between SUCLG1 and ETC, also known as oxidative phosphorylation, in TCGA AML and two other AML transcriptomic studies (GSE227839 and GSE185824), Kolmogorov–Smirnov test. (C) The top ten mitochondrial metabolic genes that positively correlate with ETC genes in TCGA AML. Shown are Spearman’s correlation coefficients. (D) Scrambled control or shRNAs against SUCLG1 were transduced into human CD34⁺ CB cells or cancer cell lines as indicated. Total DNA was extracted to quantify mtDNA. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01. (E) Mitochondria were isolated from control or SUCLG1-knockdown MV411 cells. Whole-cell lysate and mitochondrial fractions were subject to western blotting. Histone H3, G6PD and HSP60 serve as markers for nucleus, cytosol and mitochondria, respectively. (F) Shown is the correlation between SUCLG1 and ETC-related genes in various TCGA cancers. (G, H) mRNA expression of ETC-correlated genes was determined in control and SUCLG1-knockdown MV411 or HL60 cells. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; *P < 0.05; n.s. not significant. (I, J) Control and SUCLG1-knockdown MV411 cells were stained with CellROX Green (I). Cellular ROS levels were determined (J). Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01. (K, L) Apoptosis of control and SUCLG1-knockdown MV411 cells was determined. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01. (M) Stable MV411 cells were treated with 0.1 μM doxorubicin for 4 days, senescence was visualized by β-galactosidase staining. Scale bars: 50 μm. (N) BrdU incorporation was assayed in control and SUCLG1-knockdown MV411 cells. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01. (O) Distribution of stable MV411 cells in cell cycle was determined. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; *P < 0.05. (P) Multiple alignments of amino acid sequences adjacent to P170 and A209 of SUCLG1. (Q, R) The structural locations of P170 and A209 residues were analyzed (PDB ID: 6XRU). The catalytic center contains succinate, Mg²⁺, and desulfo-CoA (succinyl-CoA analog). (S) Flag-tagged SUCLG1 and disease-derived mutants (P170R and A209E) were expressed in HEK293T cells. SUCLG1 enzymes were immunopurified and subjected to catalytic activity assay. Enzyme activities were normalized to Flag-tagged protein. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01. Source data are available online for this figure.

Figure EV2. SUCLG1 reduces succinyl-CoA level to restrict POLRMT succinylation.

(A–C) ETC correlation scores of succinyl-CoA-metabolizing enzymes, SUCLA2 (A), SUCLG2 (B), and OGDH (C), were determined in various TCGA cancers. Shown are Spearman’s correlation coefficients. (D) OGDH protein expression in stable HL60 and MV411 cells was determined by western blotting. (E) Extracellular acidification rates of control or SUCLG1-knockdown HL60 cells were determined. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; n.s. not significant. (F, G) Glucose (Glc) consumption and lactate production were assayed in control and SUCLG1-knockdown HL60 cells. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01. (H) ATP abundance was quantified in control and SUCLG1-knockdown HL60 cells, and normalized to total protein. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; *P < 0.05. (I) Schematic overview of metabolic enzymes and transporters involved in succinate metabolism. (J) Mitochondria were isolated from stable SUCLG1-knockdown and rescue MV411 cells. Whole-cell lysate and mitochondrial fractions were subject to western blotting. Histone H3, G6PD, and HSP60 were included as markers for the nucleus, cytosol, and mitochondria, respectively. (K) HA-tagged POLRMT was expressed in scrambled control or SUCLG1-knockdown MV411 cells. POLRMT-HA was immunoprecipitated from stable MV411 cells and subjected to western blotting to determine lysine succinylation (suc-Lys), lysine methylation (me-Lys), lysine acetylation (ac-Lys), and tyrosine phosphorylation (p-Tyr). (L) Myc-tagged POLRMT and HA-tagged ubiquitin were co-expressed in stable MV411 cells. POLRMT was immunopurified with Myc antibody to determine its ubiquitination by western blotting. (M) Endogenous POLRMT was immunoprecipitated from control or SUCLG1-knockdown MV411 cells. Ubiquitination was determined by western blotting. (N) Bacterially expressed His-tagged POLRMT was purified and incubated with succinyl-CoA (sucCoA) or succinate in vitro. Lysine succinylation levels were determined by western blotting and normalized to POLRMT-His protein (ratio). (O) Scrambled control or SUCLG1-knockdown (sh-#1) MV411 cells were cultured with 1 mM pyruvate or 0.2 mM uridine. Colony formation was determined. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; n.s. not significant. Source data are available online for this figure.

Figure EV3. K622 succinylation suppresses POLRMT activity.

(A) Multiple alignments of amino acid sequences corresponding to K622 of POLRMT from various model organisms. (B) Dot blotting was performed to test the specificity of K622 succinylation antibody. Unmodified peptide was included as the negative control. (C) HA-tagged POLRMT and its K622 mutants were expressed in MV411 cells. Cells were treated with SIRT5 inhibitor (SIRT5i). POLRMT was immunopurified with HA antibody to determine K622 succinylation. K622sc signal was normalized to POLRMT-HA protein (ratio). (D, E) HA-tagged POLRMT and its K622 mutants were re-expressed in POLRMT-knockdown MV411 and HL60 cells at physiologically relevant levels. (F, G) Control or shRNA against SUCLG1 was expressed in POLRMT-knockdown and rescue HL60 cells. POLRMT-HA was immunopurified to assay protein interaction and K622 succinylation. K622sc signal was normalized to POLRMT-HA protein (ratio) (F). MitoChIP was performed to determine mtDNA binding of POLRMT-HA (G). Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; n.s. not significant. (H) Endogenous POLRMT was immunoprecipitated from control or SDHA-knockdown MV411 cells. K622 succinylation was determined by western blotting. (I) Metabolites were extracted from control and SDHA-knockdown MV411 cells. Succinate and succinyl-CoA levels were quantified by mass spectrometry and normalized to total protein. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01. (J, K) mtDNA abundance and mitochondrial mass were determined by qPCR and MTG staining, respectively. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01. (L, M) Scrambled control or OGDH-knockdown stable MV411 cells were cultured with or without 50 μM dimethyl succinate (DMS). Succinyl-CoA and succinate levels were quantified by mass spectrometry and normalized to total protein. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; *P < 0.05. Source data are available online for this figure.

Figure EV4. Leukemia-derived FLT3 mutants upregulate SUCLG1 to boost mitobiogenesis.

(A) HA-tagged FLT3 and its mutants were stably expressed in CD34⁺ CB cells. Oxygen consumption rates were determined by Seahorse experiments. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; *P  <  0.05. (B, C) Enzymatic activities of SDH and FH in stable CD34⁺ CB cells were determined and normalized to total protein. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; n.s. not significant. (D–F) Stable CD34⁺ CB cells were treated with or without SIRT5 inhibitor. K622 succinylation (D), mtDNA binding (E), and mtDNA-encoded gene expression (F) were determined. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01. (G) Endogenous POLG was immunoprecipitated from control or FLT3^ITD-expressing CD34⁺ CB cells. Lysine succinylation was determined by western blotting. (H, I) MV411 cells were treated with FLT3-IN-3, a chemical inhibitor of FLT3, for 24 h. Endogenous POLRMT was immunopurified to determine K622 succinylation. K622sc levels were normalized to POLRMT protein (ratio) (H). Cellular succinyl-CoA levels were determined and normalized to total protein (I). Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01. (J, K) Endogenous POLRMT was immunopurified from scrambled control or FLT3-knockdown MOLM14 cells to determine K622 succinylation. K622sc levels were normalized to POLRMT protein (ratio) (J). Cellular succinyl-CoA levels were determined and normalized to total protein (K). Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01. (L, M) MOLM14 cells were treated with FLT3-IN-3 for 24 h. Endogenous POLRMT was immunopurified to determine K622 succinylation. K622sc levels were normalized to POLRMT protein (ratio) (L). Cellular succinyl-CoA levels were determined and normalized to total protein (M). Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01. (N) mRNA expression of mitochondrial genes in TCGA AML dataset as indicated was extracted from cBioportal. Gene expression levels in FLT3^WT (n = 116) and FLT3^TKD/ITD (n = 39) samples were compared. t test. *P  <  0.05. (O–R) MV411 cells were treated with FLT3 inhibitors. Protein expression of indicated genes was determined by western blotting (O). Enzyme activities of SDH (P) and FH (Q) were assayed and normalized to total protein. Endogenous POLG was immunoprecipitated to evaluate lysine succinylation (R). Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; n.s. not significant. (S) Scrambled control or shRNAs targeting FLT3 (#1) and SUCLG1 (#1) were co-expressed in MV411 cells as indicated. Knockdown efficiency was tested by western blotting. Source data are available online for this figure.

Figure EV5. FLT3 signaling modulates nuclear transcription to enhance mitochondrial respiration.

(A–C) Multiple alignments of consensus sequences in FLT3-regulated mitochondrial genes. Numbers indicate the distance (bp) to TSS. (D, E) Scrambled control or shRNAs targeting E2F1 were expressed in MOLM14 cells. The knockdown efficiency was validated with western blotting (D). Total RNA was extracted to determine the mRNA expression of SUCLG1 (E). Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01. Source data are available online for this figure.

Figure EV6. SUCLG1 suppresses POLRMT succinylation to support leukemic proliferation.

(A) SUCLG1-knockdown and rescue MV411 cells were treated with POLRMT inhibitor (POLRMTi). Colony formation abilities were assayed. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; *P  <  0.05. (B, C) SUCLG1-knockdown and rescue MV411 cells were treated with FLT3 inhibitor FLT-IN-3. Cell proliferation (B) and colony formation (C) were assayed. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; *P  <  0.05; n.s. not significant. (D) Scrambled control or shRNA against SUCLG1 (#1) were stably expressed in POLRMT-knockdown and rescue MV411 cells. Colony formation was determined. Data represent means ± SD, n = 3 independent biological replicates, t test. **P  <  0.01; n.s. not significant. (E, F) MLL-AF9 bone marrow cells were transduced with vector control or HA-tagged FLT3^ITD. Scrambled control or shRNAs targeting Polrmt was stably expressed. Protein expression was determined by western blotting (E). Animal survival was assayed after bone marrow transplantation (F). n = 9 mice, log-rank (Mantel–Cox) test. **P  <  0.01. (G) POLRMT-knockdown and rescue MV411 cells were transplanted into sublethally irradiated NSG mice. Mice were treated with or without FLT3-IN-3, animal survival was determined. n = 5 mice, log-rank (Mantel–Cox) test. **P  <  0.01. (H–K) Human clinical AML bone marrow samples were collected. K622 succinylation of immunopurified POLRMT was determined by western blotting (H). Protein expression of POLRMT (I), mtDNA abundance (J), and MT-CO1 mRNA expression (K) were determined. n = 5–10, t test. *P  <  0.05; n.s. not significant. Source data are available online for this figure.