Hydrogen sulfide coordinates glucose metabolism switch through destabilizing tetrameric pyruvate kinase M2

Abstract

Most cancer cells reprogram their glucose metabolic pathway from oxidative phosphorylation to aerobic glycolysis for energy production. By reducing enzyme activity of pyruvate kinase M2 (PKM2), cancer cells attain a greater fraction of glycolytic metabolites for macromolecule synthesis needed for rapid proliferation. Here we demonstrate that hydrogen sulfide (H₂S) destabilizes the PKM2 tetramer into monomer/dimer through sulfhydration at cysteines, notably at C326, leading to reduced PKM2 enzyme activity and increased PKM2-mediated transcriptional activation. Blocking PKM2 sulfhydration at C326 through amino acid mutation stabilizes the PKM2 tetramer and crystal structure further revealing the tetramer organization of PKM2-C326S. The PKM2-C326S mutant in cancer cells rewires glucose metabolism to mitochondrial respiration, significantly inhibiting tumor growth. In this work, we demonstrate that PKM2 sulfhydration by H₂S inactivates PKM2 activity to promote tumorigenesis and inhibiting this process could be a potential therapeutic approach for targeting cancer metabolism.

Fig. 1. H₂S destabilizes tetramerization of PKM2.

a Upper: MDA-MB-231 cell lysates were treated with 100 μM NaHS for 30 min at 37 °C and subjected with or without 1 mM DTT for 10 min. The biotin switch assay was then applied to precipitate sulfhydrated proteins. The biotin-labeled protein was analyzed by immunoblotting with anti-PKM2 antibody to detect sulfhydration of PKM2. Bottom: The SSH-labeled PKM2 was normalized with the level of total PKM2. Data are presented as the means ± SEM (n = 3 biological replicates). One-way ANOVA followed by Tukey’s multiple comparisons test was used for the statistical analysis (***p < 0.001; ****p < 0.0001). Immunoblotting experiments were repeated at least 3 times with similar results. b The PK activity on recombinant PKM2 in the presence of 100 μM NaHS for 30 min on ice and subject to 4 mM DTT for another 10 min on ice. Pyruvate kinase activities were then assayed by measuring the amount of pyruvate production. Data are presented as the means ± SD (n = 3 technical replicates). The results are consistent across two biological replicates. c Upper: Glycerol gradient ultracentrifugation profiles of purified recombinant PKM2 and the effects of FBP and H₂S on PKM2 oligomerization. Recombinant PKM2 (10 μg) was exposed to 100 μM FBP, 100 μM NaHS, or both. After centrifugation, fractions were collected and analyzed by immunoblotting with PKM2 antibody. The distributions of PKM2 tetramer and monomer/dimer were as indicated. Bottom: Quantitative analysis of PKM2 protein level. The histogram represents normalized means ± SEM (n = 3 biological replicates in FBP+NaHS group; n = 4 biological replicates in NaHS group; n = 5 biological replicates in control and FBP group). The two-tailed student’s t test was used for the statistical analysis (*p < 0.05, group of FBP compared to the other three groups). Immunoblotting experiments were repeated at least 3 times with similar results. d Left: MDA-MB-231 cells were exposed to 1 μM NaHS for 1 h. Subcellular localization of PKM2 was detected by immunocytochemistry. Nuclei were counterstained with DAPI. The representative images are shown. Scale bars: 10 μm. Right: The percentage of nuclear PKM2 is shown in a violin plot with individual points. Horizontal black dotted lines display the median and the percentage of cells with >20% nuclear localization of PKM2 were indicated (n = 65–70 individual cells, data were combined from three independent experiments). The two-tailed student’s t test was used for the statistical analysis (****p < 0.0001). e MDA-MB-231 cells were treated with 1 μM NaHS for 24 h. The relative expression of genes regulated by PKM2 was measured by qRT-PCR. Data are presented as the means ± SD (n = 3 biological replicates in CCND1 group and n = 5 biological replicates in all the other groups). The two-tailed student’s t test was used for the statistical analysis (*p < 0.05; **p < 0.01; ***p < 0.001). f Cell proliferation assays were performed in MDA-MB-231 cells with CBS and CTH knockdown by siRNA. Data are presented as the means ± SD (n = 3 biological replicates). Two-way ANOVA followed by Tukey’s multiple comparisons test was used for the statistical analysis (*p < 0.05, **p < 0.01, siCBS&CTH-1 or siCBS&CTH-2 compared to the siCon). g Western blot analysis of CBS, CTH, 3MST, and GAPDH expression in MCF-10A, MDA-MB-231(MB-231), MCF-7, and HCC-1395 cells. The samples derive from the same experiment but different gels for CTH and 3MST, and another for CBS and GAPDH were processed in parallel. GAPDH is the internal control. Immunoblotting experiments were repeated at least 3 times with similar results. h Upper: MCF-10A, MDA-MB-231, MCF-7, and HCC-1395 cells were lysed and subjected to the biotin switch assay to precipitate sulfhydrated proteins. The biotin-labeled protein was analyzed by immunoblotting with anti-PKM2 antibody to detect sulfhydration of PKM2. Bottom: The SSH-labeled PKM2 was normalized with the level of total PKM2. Data are presented as the means ± SEM (n = 4 biological replicates in the group of MCF-7 and HCC-1395; n = 6 biological replicates in the group of MCF-10A and MB-231). The One-way ANOVA followed by Tukey’s multiple comparisons test was used for the statistical analysis (*p < 0.05). Immunoblotting experiments were repeated at least 3 times with similar results. i Upper: MDA-MB-231 cells were under normal (20% O₂) or hypoxia incubator (1% O₂) for 48 h with or without 0.25 mM AOAA. Cells were then lysed and subjected to the biotin switch assay to precipitate sulfhydrated proteins. The biotin-labeled protein was analyzed by immunoblotting with anti-PKM2 antibody to detect sulfhydration of PKM2. Bottom: The SSH-labeled PKM2 was normalized with the level of total PKM2. Data are presented as the means ± SEM (n = 4 biological replicates). One-way ANOVA followed by Tukey’s multiple comparisons test was used for the statistical analysis (**p < 0.01). Immunoblotting experiments were repeated at least 3 times with similar results. Source data are provided as a Source Data file.

Fig. 2. Replacement of PKM2 sulfhydration site to serine at C326 increases tetrameric oligomerization.

a The PKM2 monomer shows the positions of ten cysteines in the folded structure. b Purified recombinant PKM2 was treated with 100 μM NaHS and then subjected to trypsin digestion. LC-MS/MS was performed. A side-by-side comparison of the thiol (-SH) and persulfide (-SSH) detected by MS on the same cysteine residues was shown. Dots with a solid orange color are depicted as positive detection by LC-MS/MS; the rest are depicted in hollow circles. c MDA-MB-231 cells were treated with 10 mM MMTS in the cell culture medium for 20 min to block free thiols. Cells were then lyzed and endogenous PKM2 was immunoprecipitated, treated with IAM to block sulfhydrated cysteine, and then subjected for LC-MS/MS analysis. A side-by-side comparison of the thiol (-SH) and persulfide (-SSH) detected by MS on the same cysteine residues is shown. Dots with a solid orange color are depicted as positive detection by LC-MS/MS; the rest are depicted in hollow circles. d Gel filtration analysis of PKM2 recombinant proteins (wild-type or C326S mutant) in the absence of FBP. The molecular mass markers, tetramer, dimer, and monomer are as indicated. e Upper: Glycerol gradient ultracentrifugation profiles of MDA-MB-231 cells transfected with V5-tagged PKM2 or PKM2^C326S mutant. After centrifugation, fractions were collected and analyzed by immunoblotting. Bottom: Quantitative analysis of PKM2 protein level. The histogram represents normalized means ± SEM (n = 4 biological replicates). The one-tailed student’s t test was used for the statistical analysis (*p < 0.05; **p < 0.01). Immunoblotting experiments were repeated at least 3 times with similar results. Source data are provided as a Source Data file.

Fig. 3. PKM2^C326S exhibits a unique conformation.

a Crystal structure of the PKM2 C326S mutant. b Structural overlay of the C326S mutant on the T-state conformation (in complex with Phe, PDB: 4FXJ). The effector loop and α14-α15 regions are indicated by asterisk and star respectively. c Structural overlay of PKM2^C326S on the R-state conformation (in complex with FBP and Ser, PDB: 4B2D). PKM2^C326S is in blue, T-state in pink, and R-state in green. d Hydrogen bonding of residues 326 to Q310 in PKM2^C326S and the R-state conformation. e, f Zoom-in view of the effector loop region at the C-C interface. Side chains of W482 and W515 are shown in the stick presentation. Protomers are indicated in parentheses. g, h Zoom-in view of the α14-α15 region at the C-C interface. Hydrogen bonds are shown as dashed lines.

Fig. 4. Blockade of PKM2 sulfhydration at C326 increases PK activity, resulting in enhanced mitochondrial oxidative phosphorylation and reduced nuclear translocation.

a Upper Left: Western blot analysis of the expression of V5-tagged PKM2 (vector, wildtype, or C326S mutant) in MDA-MB-231 cells. The samples derive from the same experiment but different gels for V5 and GAPDH, and another for PKM2 were processed in parallel. Upper Right: MDA-MB-231 cells with stable expression were lysed and subjected to the biotin switch assay to precipitate sulfhydrated proteins. The biotin-labeled protein was analyzed by immunoblotting with anti-PKM2 antibody to detect sulfhydration of PKM2. Bottom: The PKM2-V5 was normalized with GAPDH and the SSH-labeled PKM2 was normalized with the level of total PKM2. Data are presented as the means ± SEM (n = 3 biological replicates). The two-tailed student’s t test (Left) or one-way ANOVA followed by Tukey’s multiple comparisons test (Right) was used for the statistical analysis (*p < 0.05; **p < 0.01; ***p < 0.001). Immunoblotting experiments were repeated at least 3 times with similar results. b Pyruvate kinase activities were assayed by measuring the amount of pyruvate production (nmol) in MDA-MB-231 cells expressing vector control, PKM2^wt, or PKM2^C326S. Data are presented as means ± SD (n = 4 biological replicates). One-way ANOVA followed by Tukey’s multiple comparisons test was used for the statistical analysis (*p < 0.05). c The oxygen consumption rate (OCR) curves in MDA-MB-231 expressing vector alone, PKM2^wt, or PKM2^C326S. Cells were treated with oligomycin, FCCP, and rotenone/antimycin A, respectively Data are presented as means ± SD (n = 3 biological replicates). The one-tailed student’s t test was used for the statistical analysis (*p < 0.05 C326S compared to the vector or PKM2 group). d The level of basal OCR and maximal OCR normalized to the cell numbers in MDA-MB-231 expressing vector alone, PKM2^wt or PKM2^C326S Data are presented as means ± SEM (n = 3 biological replicates). One-way ANOVA followed by Tukey’s multiple comparisons test was used for the statistical analysis (*p < 0.05; ns: not significant). e The ECAR curves in MDA-MB-231 expressing vector alone, PKM2^wt or PKM2^C326S. Cells were treated with glucose, oligomycin, and 2-DG, respectively. Data are presented as means ± SD (n = 3 biological replicates). f  Glycolysis normalized to the cell numbers in MDA-MB-231 cells expressing vector alone, PKM2^wt or PKM2^C326S. Data are presented as means ± SEM (n = 3 biological replicates). One-way ANOVA followed by Tukey’s multiple comparisons test was used for the statistical analysis (ns: not significant). g The relative expression of PKM2-responded genes was measured by qRT-PCR from MDA-MB-231 cells with PKM2^wt or PKM2^C326S. Data are presented as normalized means ± SD (n = 3 biological replicates). The one-tailed Student t test was used for the statistical analysis (*p < 0.05; **p < 0.01; ****p < 0.0001). h Metabolite analysis was conducted in MDA-MB-231 cells with PKM2^wt or PKM2^C326S expression. The level of glycolysis intermediates, including Glucose, Glucose-6-phosphate, Fructose-1,6-Bisphosphate, Ribose-5-phosphate, Glyceraldehyde-3-phosphate, Phosphoenolpyruvate, Pyruvate, and Lactate was determined by mass spectrometry. Data are presented as means ± SD (n = 4 in the group of Glucose-6-phosphate; n = 6 in all the other groups; data were combined from two independent experiments). The two-tailed student’s t test was used for the statistical analysis (*p < 0.05; **p < 0.01; ***p < 0.001, ****p < 0.0001). i Left: MDA-MB-231 cells expressing PKM2^wt or PKM2^C326S were exposed to 1 μM NaHS for 1 h. Subcellular localization of PKM2 was detected by immunocytochemistry. Nuclei were counterstained with DAPI. The representative images are shown. Scale bars: 10 μm. Right: The percentage of nuclear PKM2 is shown in a violin plot with individual points. Horizontal black dotted lines display the median and the percentage of cells with >20% nuclear localization of PKM2 were indicated (n = 41–47 individual cells, data were combined from three independent experiments). One-way ANOVA followed by Tukey’s multiple comparisons test was used for the statistical analysis (****p < 0.0001). Source data are provided as a Source Data file.

Fig. 5. Blockade of PKM2 sulfhydration at C326 causes defects in cell cycle progression and cytokinesis failure.

a, b The cell cycle of MDA-MB-231 cells expressing vector only, PKM2^wt, or PKM2^C326S was examined by the PI staining method and measured by flow cytometry. a The representative image is shown. b The percentage of cells at different cell cycle phases was analyzed. Data are presented as means ± SD (n = 4 biological replicates). One-way ANOVA followed by Tukey’s multiple comparisons test was used for the statistical analysis (*p < 0.05; **p < 0.01). c, d Immunostaining of actin filaments (Red), α-tubulin (Green), and DNA counterstaining with DAPI (Blue) of MDA-MB-231 cells expressing vector only, PKM2^wt, or PKM2^C326S. c The representative images are shown. Scale Bar: 50 μm. d The percentage of poly-nuclei cells (n ≧ 2) was analyzed. Data are presented as means ± SD (n = 3 biological replicates). One-way ANOVA followed by Tukey’s multiple comparisons test was used for the statistical analysis (*P < 0.05; **P < 0.01; ***P < 0.001). e, f Real-time live imaging of MDA-MB-231 cells expressing PKM2^wt or PKM2^C326S expression was performed to monitor cytokinesis failure. e Snapshots of a representative cell expressing PKM2^wt or PKM2^C326S undergoing division are shown. The red arrows point to cells undergoing mitotic division. Time stamps denote h:m of elapsed time. See also Movies S1 and S2. f Mitotic events with cytokinesis failure were recorded by time-lapse live cell microscopy. Results are expressed as means ± SD (n = 3 biological replicates). The one-tailed student t test was used for the statistical analysis (*p < 0.05). Source data are provided as a Source Data file.

Fig. 6. Blockade of PKM2 sulfhydration at C326 reduces cancer cell proliferation and tumor growth in the mouse xenograft model.

a Comparison of cell proliferation rate between MDA-MB-231 cells expressing vector, PKM2^wt, or PKM2^C326S. Cell proliferation assay was performed and measured by MTS reagents. Data are the means ± SD (n = 3 biological replicates). Two-way ANOVA followed by Tukey’s multiple comparisons test was used for the statistical analysis (#p < 0.05, C326S compared to the PKM2 or the vector group). b–e MDA-MB-231 cells expressing vector, PKM2^wt or PKM2^C326S were injected into the mouse mammary fat pad. b The tumor volumes were monitored every week. Data are means ± SD (n = 3 in the group of Vector; n = 9 in the group of PKM2; n = 10 in the group of C326S); data were combined from two independent experiments. Two-way ANOVA followed by Tukey’s multiple comparisons test was used for the statistical analysis (**p < 0.01, ***p < 0.001, C326S compared to the vector group; ##p < 0.01, ###p < 0.001, C326S compared to the PKM2 group). c The mean of mouse weight ± SD (n = 3 in the group of Vector; n = 9 in the group of PKM2; n = 10 in the group of C326S; data were combined from two independent experiments). d Representative bioluminescence images (BLI) at week 7 after implantation of respective cancer cells are shown. e The kinetics of individual tumors was monitored by IVIS. Data are means ± SD (n = n = 3 in the group of Vector; n = 9 in the group of PKM2; n = 10 in the group of C326S; data were combined from two independent experiments). One-way ANOVA followed by Tukey’s multiple comparisons test was used for the statistical analysis (*p < 0.05; ****p < 0.0001). f A schematic illustration revealing H₂S modulates glucose metabolism switch through destabilizing PKM2 oligomerization. Left: H₂S promotes dissociation of PKM2 tetramer to monomer/dimer through protein sulfhydration. Inhibition of PKM2 activity results in the accumulation of metabolic intermediates required for the biosynthesis. Meanwhile, the PKM2 monomers or dimers translocate into the nucleus to facilitate multiple gene expressions to promote cancer progression. Right: Blockade of PKM2 sulfhydration at cysteine 326 by mutation stabilizes PKM2 tetramer to maintain high PK activity, resulting in high energy generation, low biosynthesis, and inhibition of tumor growth. Source data are provided as a Source Data file.