Deacetylation of serine hydroxymethyl-transferase 2 by SIRT3 promotes colorectal carcinogenesis

Abstract

The conversion of serine and glycine that is accomplished by serine hydroxymethyltransferase 2 (SHMT2) in mitochondria is significantly upregulated in various cancers to support cancer cell proliferation. In this study, we observed that SHMT2 is acetylated at K95 in colorectal cancer (CRC) cells. SIRT3, the major deacetylase in mitochondria, is responsible for SHMT2 deacetylation. SHMT2-K95-Ac disrupts its functional tetramer structure and inhibits its enzymatic activity. SHMT2-K95-Ac also promotes its degradation via the K63-ubiquitin-lysosome pathway in a glucose-dependent manner. TRIM21 acts as an E3 ubiquitin ligase for SHMT2. SHMT2-K95-Ac decreases CRC cell proliferation and tumor growth in vivo through attenuation of serine consumption and reduction in NADPH levels. Finally, SHMT2-K95-Ac is significantly decreased in human CRC samples and is inversely associated with increased SIRT3 expression, which is correlated with poorer postoperative overall survival. Our study reveals the unknown mechanism of SHMT2 regulation by acetylation which is involved in colorectal carcinogenesis.

Fig. 1

SHMT2 is acetylated at K95 and SIRT3 is the major deacetylase for SHMT2. a Western blot detection of acetylation levels of ectopically expressed SHMT2 in HeLa and U2OS cells after treated with 7.5 mM NAM for the duration indicated. AcK, pan-acetyl-lysine antibody. b Acetylated SHMT2 K95 was identified by a tandem mass spectrum. The identified peptide is shown. c K95 in SHMT2 is conserved. The sequences around SHMT2 K95 from different species were aligned. d K95 is the primary acetylation site of SHMT2. The indicated plasmids were transfected into HeLa cells, the acetylation levels were analysis by western blot. e Characterization of acetyl-SHMT2 (K95) antibody. Acetylation level of SHMT2-Flag or SHMT2-K95Q-Flag ectopically expressed in SHMT2-knockout HCT116 cells was measured by the site-specific K95 acetylation antibody (Ac-K95). f K95 acetylation of SHMT2 is broadly verified in different human cancer cells. Endogenous SHMT2 K95-Ac levels of indicated cells after NAM treatment. g SIRT3 interacts with SHMT2 in vivo. Whole HCT116 cell lysates were immunoprecipitated with control IgG, anti-SIRT3 or anti-SHMT2 antibodies, and precipitated proteins were detected by anti-SHMT2 or anti-SIRT3 antibodies, respectively. h SIRT3 deacetylates SHMT2 in vitro. Recombinant human (rh) SIRT3 deacetylates SHMT2 proteins from HCT116 cells. i The acetylation levels of IPed SHMT2 were determined by western blot in stable SIRT3-overexpressing HCT116 cells. j Recombinant SIRT3 wild-type or a catalytically inactive mutant H248Y and the site-specific K95-acetylated SHMT2 incubated with NAD⁺ or NAM, and acetylation level was determined by Ac-K95 and AcK. k SIRT3 knockout increases SHMT2 acetylation level and decreases SHMT2 activity. Corresponding SHMT2 activity was measured and normalized against protein levels. Error bars represent ± s.d. for triplicate experiments. ***P < 0.001 by Student’s t-test. l SHMT2 K95-Ac increases and protein level decreases in Sirt3^−/− MEFs. Western blot detection of endogenous SIRT3 and SHMT2 protein and K95-Ac. m NAM treatment or SIRT3 deficiency increases SHMT2-K95-Ac and decreases SHMT2 protein level in cytoplasm. Cytosolic and mitochondria fractions of HCT116 cells after NAM treatment for 6 h were obtained (left). Cytosolic and mitochondria fractions of SIRT3-knockout HCT116 cells were obtained (right). The levels of SHMT2 protein and SHMT2-K95-Ac were compared

Fig. 2

K95 acetylation impairs the formation of tetrameric SHMT2 and inhibits its enzymatic activity. a K95 mutation decreases SHMT2 enzyme activity. The enzyme activity of SHMT2 wild-type and K95R/Q mutants purified by immunoprecipitation from HeLa cells was measured and normalized against protein level. b Inhibition of deacetylases decreases SHMT2 enzyme activity. The enzyme activity of SHMT2 purified by immunoprecipitation from HeLa cells after NAM treatment was measured and normalized against protein level. The protein levels and acetylation levels were determined by western blot. c Kinetic comparison of wild-type SHMT2 and variants. Recombinant SHMT2 proteins were purified, and steady-state kinetic analyses for DL-β-phenylserine were performed. Comparison of WT (blue line) and variants K464Q (green line, negative control), K95R (violet line) and K95Q (red line) shows that the K95 is important in catalysis activity. N.A. represents not detected effectively. d The site-specific K95-acetylated SHMT2 is a totally no activity protein.SHMT2 and K95Ac-SHMT2 were recombinant expressed and detected in total lysates by western blot or purified by nickel affinity chromatography and performed activity assay. e Molecular modeling of acetylation of K95 in SHMT2. f K95 mutation prevents SHMT2 tetramerization. Recombinant wild-type and K96R/Q mutant SHMT2 proteins were purified by nickel affinity chromatography and separated by gel filtration, followed by western analysis. Dotted box denotes K95R/Q mutant has less tetrameric formation fractions. g K95 mutation attenuates the interaction of SHMT2 monomers. Flag-tagged and Myc-tagged SHMT2 WT or K95R/Q mutants were expressed in HEK293T cells, and the proteins were immunoprecipitated before being subjected to western blot. h K95Q mutation or loss of SIRT3 disturbs SHMT2 tetramerization. Cell lysates of SHMT2-knockout HCT116 cells expressing Flag-tagged WT or K95Q mutant SHMT2 (up) and cell lysates of SIRT3-knockout or control HCT116 cells (down) were treated with or without 0.025% glutaraldehyde and analyzed by western blot with anti-SHMT2 antibody. Tetramer and monomer of SHMT2 were indicated. For a, b, and d, mean values ± s.d. of relative enzyme activity of triplicate experiments are presented. ***P < 0.001

Fig. 3

K95 acetylation promotes SHMT2 degradation through macroautophagy. a Inhibition of deacetylases reduces endogenous SHMT2 protein level. Western blot detection of K95-Ac and protein level of endogenous SHMT2 in SW620 cells after NAM treatment for the duration indicated. b SHMT2 is a stable protein and NAM treatment accelerates SHMT2 degradation. SW480 cells were treated with protein synthesis inhibitor cycloheximide (CHX) (75 μg ml⁻¹) with or without NAM treatment for duration indicated. Endogenous SHMT2 protein and K95-Ac level was analyzed by western blot. c High glucose increases SHMT2 K95-Ac. SHMT2-Flag was transfected 24 h and then cells were treated with indicated glucose concentration. d High glucose decreases SHMT2 protein level. Cells were cultured with different concentration of glucose. The steady-state level and K95-Ac of endogenous SHMT2 were analyzed by western blot. e Low glucose promotes the interaction between SIRT3 and SHMT2. The indicated plasmids were co-transfected 24 h, followed by indicated glucose concentration treatment, and the interaction between SIRT3 and SHMT2 was examined by immunoprecipitation. f SHMT2 is degraded by the autophagic pathways. Cells were treated with CHX for 10 h and meanwhile treated with autophagic pathways inhibitor NH₄Cl for 12h. NH₄Cl treatment, ‘‘ + ” represent 10 mM and ‘‘ + + ” represent 20 mM. g 3-MA increases SHMT2 accumulation. Cells were treated with 3-MA (2 mM) for duration indicated. h ATG5-knockdown increases SHMT2 protein level. HCT116 cells were transfected with siATG5 48 h before harvest. The levels of SHMT2 and LC3B were analyzed by western blot. The decrease of LC3BII/I indicates autophagic activity suppressed. i K95Q mutation increases the K63-polyUb binding of SHMT2. Flag-tagged SHMT2 WT and K95Q were co-transfected with HA-Ub WT, K48-only, and K63-only mutants in HEK293T cells. j NH₄Cl increases the ubiquitin binding of SHMT2. Plasmids transfection for 18 h then cells were treated with 15 mM NH₄Cl for 24 h before harvest. k NAM treatment increases SHMT2 K63 poly-Ub chains. SHMT2-Flag and HA-Ub-K63-only were co-transfected for 36 h and then cells were treated 7.5 mM NAM for 4 h before harvest. For a, b, d mean values of quantitation ± s.d. are reported. *P < 0.05; **P < 0.001; ***P < 0.001. Representative western blot results and quantitation (herein after) of triplicated western blot are shown

Fig. 4

TRIM21 is the E3 Ligase for SHMT2. a Identification of SHMT2-interacting proteins. SHMT2-interacting proteins involved in ubiquitination are shown. b TRIM21 interacts SHMT2. Interactions between ectopically expressed SHMT2 and TRIM21 in HEK293T cells were determined by western blot. c Wild-type TRIM21 not the ligase-dead (LD) mutant TRIM21 increases SHMT2 ubiquitylation. HA-Ub and SHMT2-Flag were co-transfected with Myc-TRIM21 WT or Myc-TRIM21-LD mutant in HEK293T cells. d K95Q mutant binds more TRIM21. Immunoprecipitation was performed from SHMT2-knockdown HeLa cells with either WT or K95Q mutant SHMT2-Flag stably rescued before being subjected to western blot with endogenous TRIM21 antibody. e TRIM21 increases K63-Ub chains of SHMT2. The indicated plasmids were transfected into HEK293T cells. f TRIM21 knockdown decreases K63-only-Ub not K63R-Ub chains of SHMT2. The indicated plasmids were transfected into TRIM21 knockdown or control HEK293T cells. The knockdown efficiency of TRIM21 was probed by Real-time PCR. g High glucose increases SHMT2 K63 poly-Ub chains. SHMT2-Flag and HA-Ub-K63-only were co-transfected into HCT116 cells for 36 h and then medium was replaced with containing 0 mM or 25 mM glucose medium for 6 h before harvest. h TRIM21 exacerbates SHMT2 degradation under high glucose. Vector or Myc-TRIM21 was transfected into HCT116 cells for 24 h and then medium was replaced with containing 0 mM or 25 mM glucose medium for 12 h before harvest. i TRIM21 knockdown increases endogenous SHMT2 protein. TRIM21 was stably knocked down in HCT116 cells by shRNA. The knockdown efficiency and SHMT2 protein level were determined by western blot. The mRNA level change of TRIM21 and SHMT2 was determined by Real-time PCR. For f and i, Mean values ± s.d. of relative mRNA expression of triplicate experiments are presented. ***P < 0.001

Fig. 5

SHMT2 acetylation inhibits cell proliferation and tumor growth. a SHMT2-K95Q is compromised to support cell proliferation. SW480 cells stably knockdown SHMT2 with re-express the shRNA-resistant wild-type or K95Q mutant were established. SHMT2 knockdown efficiency and re-expression were determined by western blot. SHMT2 WT or SHMT2 K95Q cells were seeded in 12-well plates. Cell numbers were counted every 24 h. b SHMT2 K95Q increases intracellular serine level and serine/glycine ratio. Intracellular serine and glycine amount were measured by GC-MS. c SHMT2 K95Q is defective in supporting tumor growth in vivo. Xenograft was performed using the SHMT2-knockdown SW480 cells with reexpressing wild-type or K95Q mutant SHMT2 as indicated. ***P < 0.001. d SHMT2 K95Q decreases cellular NADPH level. SHMT2 WT or K95Q-rescued HCT116-SHMT2-knockout cells were established and confirmed by western blot. Intracellular NADPH was determined by using NAD(P)H-Glo Detection System. e SHMT2 K95Q increases cellular ROS level. SHMT2 WT or K95Q-rescued HCT116-SHMT2-knockout cells were seeded in confocal dishes and ROS level was measured by adding 10 μM H2DCF-DA. Fluorescent strength per unit area was quantified using the ImageJ software, followed by statistical analysis. f Sirt3-knockout mice developed less tumors in the AOM-DSS CRC mouse models. Sirt3 WT and KO mice were subjected to AOM-DSS colitis-associated cancer model. The representative images of colon tumors are shown. Tumor number of each group were shown. n = 8 (WT), n = 9 (Sirt3_KO), P = 0.0027 by student’s t-test. The expression of Sirt3 and Shmt2 in the colon tumors from WT and Sirt3_KO mice were detected by IHC. For a, d, and e, mean values ± s.d. of triplicate experiments are reported. *P < 0.05; **P < 0.001; ***P < 0.001

Fig. 6

K95-acetylation of SHMT2 is downregulated in CRC. a In total, 35 pairs of tumor tissues (T) and adjacent normal tissues (N) were lysed. Protein levels of SHMT2, SHMT2-K95-Ac, SIRT3 and TRIM21 were determined by direct western blot. Relative protein levels were normalized by β-actin. Shown are seven pairs of samples. See Supplementary Fig. 6 for the other 28 pairs of samples. b Quantification of relative SHMT2, SHMT2-K95-Ac, SIRT3, and TRIM21 protein levels in the 35 pairs of samples tested. The intensities of indicated proteins were quantified using the ImageJ software, followed by statistical analysis. *P < 0.05; **P < 0.001; ***P < 0.001; n.s. not significant. c SHMT2 protein levels show negative correlation with SHMT2-K95-Ac. Correlation between SHMT2 protein levels and SHMT2-K95-Ac levels in the tested 35 pairs of samples. Statistical analyses were performed with F-test. d Positive correlation between SHMT2 and SIRT3 staining patterns in colorectal cancer. Three-hundred and nine colorectal cancer specimens were immunohistochemically stained with indicated antibodies. Representative photos of tumors were shown (left). Case 1, low SHMT2 expression in a low-SIRT3-expression specimen; Case 2, high SHMT2 expression in a high-SIRT3-expreesion specimen. Scale bars: 100 μm. Pearson correlation test was used to assess the statistical significance using R statistical program (right). Note that some of the dots on the graphs represent more than one specimen (i.e., some scores overlapped). e The patients with high SIRT3 expression (n = 190) have poorer overall survival compared with low-SIRT3 expression (n = 119). Significance was determined using Kaplan–Meier analyses. f Acetylation at K95 under high glucose inhibits SHMT2 enzyme activity and promotes its lysosomal degradation via macrophage. In CRC, low glucose promotes deacetylation of SHMT2 to stabilize SHMT2 and maintain its high activity, increasing cell proliferation and tumor growth