Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2017 Mar;27(3):329-351.
doi: 10.1038/cr.2016.159. Epub 2016 Dec 30.

Mitochondrial PKM2 regulates oxidative stress-induced apoptosis by stabilizing Bcl2

Affiliations

Mitochondrial PKM2 regulates oxidative stress-induced apoptosis by stabilizing Bcl2

Ji Liang et al. Cell Res. 2017 Mar.

Abstract

Pyruvate kinase M2 isoform (PKM2) catalyzes the last step of glycolysis and plays an important role in tumor cell proliferation. Recent studies have reported that PKM2 also regulates apoptosis. However, the mechanisms underlying such a role of PKM2 remain elusive. Here we show that PKM2 translocates to mitochondria under oxidative stress. In the mitochondria, PKM2 interacts with and phosphorylates Bcl2 at threonine (T) 69. This phosphorylation prevents the binding of Cul3-based E3 ligase to Bcl2 and subsequent degradation of Bcl2. A chaperone protein, HSP90α1, is required for this function of PKM2. HSP90α1's ATPase activity launches a conformational change of PKM2 and facilitates interaction between PKM2 and Bcl2. Replacement of wild-type Bcl2 with phosphorylation-deficient Bcl2 T69A mutant sensitizes glioma cells to oxidative stress-induced apoptosis and impairs brain tumor formation in an orthotopic xenograft model. Notably, a peptide that is composed of the amino acid residues from 389 to 405 of PKM2, through which PKM2 binds to Bcl2, disrupts PKM2-Bcl2 interaction, promotes Bcl2 degradation and impairs brain tumor growth. In addition, levels of Bcl2 T69 phosphorylation, conformation-altered PKM2 and Bcl2 protein correlate with one another in specimens of human glioblastoma patients. Moreover, levels of Bcl2 T69 phosphorylation and conformation-altered PKM2 correlate with both grades and prognosis of glioma malignancy. Our findings uncover a novel mechanism through which mitochondrial PKM2 phosphorylates Bcl2 and inhibits apoptosis directly, highlight the essential role of PKM2 in ROS adaptation of cancer cells, and implicate HSP90-PKM2-Bcl2 axis as a potential target for therapeutic intervention in glioblastoma.

PubMed Disclaimer

Figures

Figure 1
Figure 1
PKM2 translocates to the mitochondria and inhibits H2O2-induced Bcl2 degradation and apoptosis. Immunoprecipitation and immunoblotting analyses were performed with the indicated antibodies. Data are representative of at least three independent experiments. (A, B) U87 (top panel) or U251 (bottom panel) cells that stably express control shRNA (shNT) or shRNA against PKM2 (shPKM2) were treated with or without H2O2 (0.5 mM, 24 h). The cells were harvested for apoptosis analysis using Annexin V-Cy5 Apoptosis Kit (A) or caspase 3 activity analysis (B). Data represent the means ± SD of three independent experiments. CON, control. (C) U87 or U251 cells with or without PKM2 depletion were treated with H2O2 (0.5 mM, 24 h) and harvested for cytosolic (Cyto) and mitochondrial (Mito) fractionation. (D) U87 cells were treated with or without H2O2 (0.5 mM, 1 h). Immunofluorescence analyses were performed using anti-PKM2 or anti-VDAC antibodies. Representative images were shown in left panel. Co-localization of two proteins in 300 cells was quantified using the software of Volocity (PerkinElmer, right panel). (E) U87 or U251 cells were treated with or without H2O2 (0.5 mM, 1 h). Cytosolic and mitochondrial fractions of the cells were prepared. (F) U87/shNT or U87/shPKM2 cells were treated with or without H2O2 (0.5 mM, 4 h). (G) U87/shNT or U87/shPKM2 cells were treated with or without H2O2 (0.5 mM, 4 h). Quantitative RT-PCR was performed to examine mRNA levels of Bcl2. Data represent the means ± SD of three independent experiments. (H) U87/shPKM2 cells were pretreated with or without MG132 for 0.5 h, followed by treatment of H2O2 (0.5 mM) for 4 h. (I) U87/shNT or U87/shPKM2 cells were infected with a lentivirus expressing SFB-Bcl2. The cells were pretreated with MG132 for 0.5 h and then treated with or without H2O2 (0.5 mM, 4 h). Pull-down of SFB-Bcl2 was performed with streptavidin agarose beads. PD, pull-down.
Figure 2
Figure 2
PKM2 interacts with and stabilizes Bcl2. Immunoprecipitation and immunoblotting analyses were performed with the indicated antibodies. Data are representative of at least three independent experiments. (A) U87 or U251 cells were treated with or without H2O2 (0.5 mM, 1 h). Succinate dehydrogenase (SDHA) that has not been reported to interact with PKM2 or Bcl2 was used as a negative control. (B) U87 cells were treated with or without H2O2 (0.5 mM, 1 h). Immunofluorescence was performed using anti-PKM2 and anti-Bcl2 antibodies. Representative images were presented in left panel. Co-localization of two proteins in 300 cells was quantified using the software of Volocity (right panel). (C) U87 cells were infected with a lentivirus expressing Flag-PKM1 or Flag-PKM2 and then treated with or without H2O2 (0.5 mM, 1 h). (D) U87 cells were infected with a lentivirus expressing SFB-PKM2 WT or the indicated deletion mutants and then treated with or without H2O2 (0.5 mM, 1 h). Pull-down of SFB-PKM2 proteins was performed with streptavidin agarose beads. (E) U87 cells stably expressing SFB-PKM2 WT or ILLL were treated with or without H2O2 (0.5 mM, 1 h). Pull-down of SFB-PKM2 proteins was performed with streptavidin agarose beads. (F-H) U87 cells were depleted of endogenous PKM2 and then reconstituted with the expression of rPKM2 WT or rPKM2 ILLL (F). The cells were treated with or without H2O2 (0.5 mM) for 4 h (G, H). The cells were pretreated with MG132 for 0.5 h before H2O2 treatment (G). (I) U87 cells that stably express SFB-Bcl2 were treated with or without H2O2 (0.5 mM, 1 h). SFB-Bcl2 complex was precipitated using streptavidin agarose beads. A peptide competition assay was performed by mixing precipitated SFB-Bcl2 complex with 1.0 μg synthesized Flag or PKM2 389-405 peptides. (J, K) U87 cells that stably express SFB-Bcl2 were infected with a lentivirus expressing HA-tagged PKM2 389-405 peptide or Flag peptide and then treated with or without H2O2 (0.5 mM) for 1 (J) or 4 h (K). SFB-Bcl2 complex was precipitated using streptavidin agarose beads (J). Expression of Flag or PKM2 peptides was examined using dot blot assay with anti-HA antibody.
Figure 3
Figure 3
HSP90α1 alters PKM2 conformation and facilitates the interaction between PKM2 and Bcl2. Immunoprecipitation and immunoblotting analyses were performed with the indicated antibodies. Data are representative of at least three independent experiments. (A) GST pull-down assay was performed by mixing bacteria-purified recombinant GST-PKM2 and His-Bcl2, or GST-PKM2 and Histone H3 at 4 °C for 4 h. GST-PKM2 was precipitated using glutathione agarose beads. (B) U87 cells that stably express SFB-PKM2 were treated with or without H2O2 (0.5 mM, 1 h). SFB-PKM2 was pulled down using streptavidin agarose beads. (C) U87 cells that stably express shNT or shHSP90α1 were treated with or without H2O2 (0.5 mM, 1 h). (D) Flag-HSP90α1 WT and Flag-HSP90α1 deletion mutants, HSP90α1 ΔC(1-224), MD(1-544) and ΔN(545-731), used in the experiments are represented as black bars. (E) U87 cells were infected with a lentivirus expressing Flag-HSP90α1 WT or Flag-HSP90α1 deletion mutants and then treated with H2O2 (0.5 mM, 1 h). (F) U87 cells were infected with a lentivirus expressing Flag-HSP90α1 WT or Flag-HSP90α1 1-293 or Flag-HSP90α1 1-360 or Flag-HSP90α1 1-469 (bottom panel). Flag-HSP90α1 WT and Flag-HSP90α1 truncation mutants, HSP90α1 1-293, 1-360 and 1-469, used in the experiments are represented as black bars (top panel). (G) U87 cells were infected with a lentivirus expressing Flag-HSP90α1 WT or Flag-HSP90α1 Δ347-360. (H) Chaperone assay was performed by mixing GST-PKM2 and recombinant HSP90α1 WT or G97D. After the reaction, GST-PKM2 was pulled down using glutathione agarose beads. (I) U87 cells were depleted of endogenous HSP90α1 and then reconstituted with the expression of rHSP90α1 WT or G97D. The cells were further infected with a lentivirus expressing SFB-PKM2 and then treated with or without H2O2 (0.5 mM, 1 h). (J) HSP90α1-depleted U87 cells were reconstituted with the expression of rHSP90α1 WT or G97D and then treated with or without H2O2 (0.5 mM, 1 h). (K) Chaperone assay was performed by mixing recombinant HSP90α1 WT or G97D and GST-PKM2. After the reaction, GST-PKM2 were pulled down using glutathione agarose beads and incubated with His-Bcl2 at 4 °C for 4 h. (L) HSP90α1-depleted U87 cells that were reconstituted with the expression of rHSP90α1 WT or G97D were treated with or without H2O2 (0.5 mM, 4 h).
Figure 4
Figure 4
PKM2 phosphorylates Bcl2 at T69. Immunoprecipitation and immunoblotting analyses were performed with the indicated antibodies. Data are representative of at least three independent experiments. (A) U87 cells that stably express shNT or shPKM2 were infected with a lentivirus expressing SFB-Bcl2 and then treated with or without H2O2 (0.5 mM, 1 h). SFB-Bcl2 was pulled down using streptavidin agarose beads. (B) Chaperon assay was performed with recombinant HSP90α1 and GST-PKM2. After reaction, GST-PKM2 was pulled down using glutathione agarose beads. In vitro kinase assays were performed by mixing isomerized GST-PKM2 and His-Bcl2 WT or Bcl2 T69A or Bcl2 S70A. (C) Chaperon assay was performed with recombinant HSP90α1 and GST-PKM2 or GST-PKM1. GST-PKM2 or GST-PKM1 was pulled down using glutathione agarose beads. In vitro kinase assay was performed by mixing His-Bcl2 and isomerized GST-PKM2 or GST-PKM1. (D) Chaperon assay was performed with recombinant HSP90α1 and GST-PKM2. GST-PKM2 was pulled down using glutathione agarose beads. In vitro kinase assays were performed by mixing isomerized GST-PKM2 and His-Bcl2. After reaction, His-Bcl2 was pulled down using Nickel agarose beads and analyzed by mass spectrometry. Mass spectrometric analysis of a tryptic fragment at m/z 182.054 (mass error was 0.10 p.p.m.) matched to the doubly charged peptide 69-pho-tSPLQTPAAPGAAAGPALSPVPPVVHLTLR-98, suggesting that T69 was phosphorylated. The Sequest score for this match was Xcorr = 2.18; Mascot scores were 27, and expectation value was 5.1 × 10−4. The Best score evidence ID was 460, and site probability was 78.4%. (E) PKM2-depleted U87 cells that were reconstituted with the expression of rPKM2 WT or ILLL were infected with a lentivirus expressing SFB-Bcl2 and then treated with or without H2O2 (0.5 mM, 1 h). (F, G) U87 cells that stably express SFB-Bcl2 WT or T69A were treated with or without H2O2 (0.5 mM, 4 h). The cells were pretreated with MG132 for 0.5 h before H2O2 treatment (F).
Figure 5
Figure 5
PKM2-dependent Bcl2 T69 phosphorylation prevents the binding of Cul3-based E3 ligase to Bcl2. Immunoprecipitation and immunoblotting analyses were performed with the indicated antibodies. Data are representative of at least three independent experiments. (A) PKM2-depleted U87 cells were reconstituted with the expression of rPKM2 WT or ILLL and then treated with or without H2O2 (0.5 mM, 1 h). (B) U87 cells that stably express SFB-Bcl2 WT or T69A were treated with or without H2O2 (0.5 mM, 1 h). SFB-Bcl2 was pulled down using streptavidin agarose beads. (C) PKM2-depleted U87 cells were infected with a lentivirus expressing SFB-Bcl2 WT or T69E and then treated with or without H2O2 (0.5 mM, 1 h). SFB-Bcl2 was pulled down using streptavidin agarose beads. (D, F) PKM2-depleted U87 cells were reconstituted with the expression of rPKM2 ILLL and then transfected with or without siRNA against CUL3 (siCul3). The cells were treated with or without H2O2 (0.5 mM, 4 h). The cells were pretreated with MG132 for 0.5 h (D). (E, G) U87 cells that stably express SFB-Bcl2 T69A were transfected with or without siCul3 and then treated with or without H2O2 (0.5 mM, 4 h). The cells were pretreated with MG132 before H2O2 treatment (E).
Figure 6
Figure 6
Phosphorylation of Bcl2 T69 by PKM2 enhances the resistance of tumor cells to oxidative stress and promotes gliomagenesis. Immunoblotting analyses were performed with the indicated antibodies. Data are representative of at least three independent experiments. (A) U87 cells were depleted of endogenous PKM2 and then reconstituted with the expression of rPKM2 WT or ILLL. The cells were treated with or without H2O2 (0.5 mM, 24 h). The cells were harvested for analyses of apoptosis using Annexin V staining (top panel) or caspase 3 activity (bottom panel). Data represent the means ± SD of three independent experiments. (B) U87 cells were depleted of endogenous Bcl2 and then reconstituted with the expression of rBcl2 WT or T69A. The cells were treated with or without H2O2 (0.5 mM, 24 h). The cells were harvested for analyses of apoptosis using Annexin V staining (top panel) or caspase 3 activity (bottom panel). Data represent the means ± SD of three independent experiments. (C, D) U87 cells were depleted of endogenous Bcl2 and reconstituted with the expression of rBcl2 WT or T69A (C). U87 cells were depleted of endogenous PKM2 and reconstituted with expression of rPKM2 WT or ILLL (D). Endogenous PKM2-depleted U87 cells with reconstituted expression of rPKM2 ILLL were further infected with the lentivirus expressing Flag-Bcl2 WT or T69E (D). All genetically-modified U87 cells above were harvested and intracranially injected into randomized athymic nude mice (seven mice per group). After 40 days, the mice were euthanized and tumor growth was examined. H&E-stained coronal brain sections show representative tumor xenografts (top panel). Representative images of tumor boundaries were presented in middle panel with 200× magnification. Tumor volumes were measured using length (a) and width (b) and calculated using the equation: V = ab/2 (bottom panel). Data represent the means ± SD of seven mice. (E) Kaplan-Meier survival curves for the mice intracranially injected with Bcl2-depleted U87 cells reconstitutively expressing rBcl2 WT or T69A, or PKM2-depleted U87 cells reconstitutively expressing rPKM2 WT or ILLL, or Bcl2 WT- or T69E-expressing U87 cells that were depleted of PKM2 and reconstituted with the expression of rPKM2 ILLL (10 mice per group). (F, G) Tumor tissues derived from the mice injected with PKM2-depleted U87 cells reconstitutively expressing rPKM2 WT or ILLL (F), or Bcl2-depleted U87 cells reconstitutively expressing rBcl2 WT or T69A (G) were dissected for immunoblotting analyses. (H, J) U87 cells were intracranially injected into randomized athymic nude mice. After 15 days, 7 mice were sacrificed to examine tumor growth; the remaining mice were intratumorally injected with the lentivirus expressing Flag or PKM2 (389-405) peptides, and then sacrificed on day 40. H&E-stained coronal brain sections show representative tumor xenografts (H, top panel). Representative images of tumor boundaries were presented in middle panel with 200× magnification. Tumor volumes were calculated (H, bottom panel). Data represent the means ± SD of 7 mice. Tumor tissues derived from the mice injected with the lentivirus expressing Flag or PKM2 389-405 peptides were dissected for immunoblotting analyses (J). (I) U87 cells were intracranially injected into randomized athymic nude mice. After 15 days, the mice were intracranially injected with the lentivirus expressing Flag or PKM2 (389-405) peptides. Kaplan-Meier survival curves were then plotted.
Figure 7
Figure 7
Bcl2 pT69 levels positively correlates with levels of conformation-altered PKM2 and grades of glioma malignancy and prognosis. (A-C) Immunohistochemical analyses of 40 specimens from GBM patients using anti-PKM2-389, anti-Bcl2 pT69 and anti-Bcl2 antibodies were performed. Representative images of two GBM specimens (A). Semi-quantitative scoring (using a scale from 0 to 8) was carried out (Pearson product moment correlation test; B, r = 0.76, P < 0.001; C, r = 0.81, P < 0.001). (D, E) Survival of 80 patients with low (0-4.5 staining scores, blue curve) versus high (4.6-8 staining scores, red curve) conformation-altered PKM2 levels (low, 22 patients; high, 58 patients) was analyzed (D). Survival of 80 patients with low (0-3.9 staining scores, blue curve) versus high (4.0-8 staining scores, red curve) Bcl2 T69 phosphorylation levels (low, 28 patients; high, 52 patients) was compared (E). Empty circles represent the deceased patients, and filled circles represent the censored (alive at last clinical follow-up) patients. (F, G) Forty diffuse astrocytoma specimens were immunohistochemically stained using PKM2-389 antibody (F) or anti-Bcl2 pT69 antibody (G). Staining scores of the specimens were compared with 40 (F) or 80 (G) stained GBM specimens (Student's t-test, two tailed, P < 0.001). (H) Schematic representation of the PKM2-mediated Bcl2 stabilization and apoptosis resistance. PKM2 translocates to the mitochondria under oxidative stress, where PKM2 interacts with and phosphorylates Bcl2 at T69. Bcl2 T69 phosphorylation that requires HSP90α1-mediated PKM2 isomerization prevents the binding of Cul3-based E3 ligase to Bcl2, which stabilizes Bcl2 and therefore inhibits apoptosis and promotes tumorigenesis.

References

    1. Reuter S, Gupta SC, Chaturvedi MM, Aggarwal BB. Oxidative stress, inflammation, and cancer: how are they linked? Free Radic Biol Med 2010; 49:1603–1616. - PMC - PubMed
    1. Visconti R, Grieco D. New insights on oxidative stress in cancer. Curr Opin Drug Discov Devel 2009; 12:240–245. - PubMed
    1. Waris G, Ahsan H. Reactive oxygen species: role in the development of cancer and various chronic conditions. J Carcinog 2006; 5:14. - PMC - PubMed
    1. Trachootham D, Alexandre J, Huang P. Targeting cancer cells by ROS-mediated mechanisms: a radical therapeutic approach? Nat Rev Drug Discov 2009; 8:579–591. - PubMed
    1. Schneider BL, Kulesz-Martin M. Destructive cycles: the role of genomic instability and adaptation in carcinogenesis. Carcinogenesis 2004; 25:2033–2044. - PubMed

MeSH terms