Nuclear PKM2 regulates the Warburg effect

Introduction

Most differentiated normal cells metabolize glucose to carbon dioxide by oxidating pyruvate in the mitochondrial tricarboxylic acid cycle under normoxic conditions. In contrast, most cancer cells have increased glucose uptake and metabolize glucose to lactate in the presence of sufficient oxygen, a phenomenon known as the Warburg effect or aerobic glycolysis.¹ Since the 1980s, detection of a high level of aerobic glycolysis by imaging uptake of 2-^{F-deoxyglucose with positron emission tomography has been used clinically to diagnose and monitor treatment responses of cancers.1} Despite Otto Warburg having described aerobic glycolysis 90 years ago¹ and the current widespread appreciation and clinical application of this phenomenon, the mechanisms underlying the Warburg effect were still poorly understood until recent progress was made in the study of cancer metabolism.

In the glycolysis pathway, pyruvate kinase (PK) is a rate-limiting glycolytic enzyme that catalyzes the conversion of phosphoenolpyruvate and ADP to pyruvate and ATP. PK has 4 isoforms in mammals: PKL, expressed in liver and kidney; PKR, expressed in erythrocytes; PKM1, predominantly expressed in adult muscle, brain, and bladder cells and in adult fibroblasts; and PKM2, expressed in most cells except adult muscle, brain, and liver. The PKM1 and PKM2 isoforms result from mutually exclusive alternative splicing of the PKM pre-mRNA that results in inclusion of either exon 9 (PKM1) or exon 10 (PKM2).²

PKM2 expression is upregulated in human cancer cells.² Increased PKM2 expression has also been detected in cancer-associated fibroblasts.³ In response to epidermal growth factor (EGF) receptor (EGFR) activation, PKM2 is upregulated by NFκB-mediated PKM gene transcription. Importantly, NFκB activation induced by EGFR is distinct from that induced by inflammatory responses and cyokine stimulation, in which polyubiquitination and TAK1-dependent phosphorylation of IKK are instrumental.⁴5 EGFR activation results in PLCγ1-dependent PKCε activation, leading to RINCK1 ubiquitin ligase-mediated monoubiquitylation of PKCε at K321. Monoubiquitylated PKCε interacts with a ubiquitin-binding domain in NEMO zinc finger and recruits the cytosolic IKK complex to the plasma membrane, where PKCε phosphorylates IKKβ at S177 and activates IKKβ. Activated RelA interacts with hypoxia-inducible factor 1α (HIF1α), which is required for RelA to bind the PKM promoter and activates PKM transcription.⁶ In turn, PTB, which is also upregulated by EGFR activation, splices PKM pre-mRNA into PKM2 mRNA to upregulate PKM2 expression.⁷ These results indicate that a coordinated regulation of increased PKM transcription and PTB-dependnent splicing of PKM pre-mRNA, rather than a shift from PKM1 to PKM2 expression, upregulates PKM2 expression in response to growth factor stimulation. ⁶8

In human lung cancer cells, replacing PKM2 with PKM1, an isoform with high constitutive activity, inhibits the Warburg effect and tumor formation in nude mouse xenografts.⁹ The tumor-specific functions of PKM2 are supported by the finding that oxidation of PKM2 C358 leads to inhibition of PKM2 and to diversion of glucose flux into the pentose phosphate pathway, thereby generating sufficient reducing potential for detoxification of reactive oxygen species.¹⁰ PKM2 can also be inhibited by PKM2 tyrosine phosphorylation at Y105, whereas serine can bind to and activate PKM2.¹¹12 The reduction in PKM2 activity in response to serine deprivation results in more pyruvate being diverted to the mitochondria and more glucose-derived carbon being channeled into serine biosynthesis to support cell proliferation.¹¹ These findings highlight the role of PKM2 in glycolysis and the contribution of cytosolic PKM2 regulation to the synthesis of amino acids, nucleotides, and NADPH production.¹³ Neverthless, the mechanism by which PKM2, but not PKM1, regulates glucose uptake and lactate production and whether PKM2 has a unique role in controlling cancer metabolism and cell proliferation were not clear until recent reports from our laboratory and others shed light on the essential function of nuclear PKM2 in promoting the Warburg effect and tumorigenesis.

Monomeric PKM2 Translocates into the Nucleus

Activation of receptor tyrosine kinases, such as EGFR and platelet-derived growth factor receptor (PDGF-R), results in translocation of PKM2, but not PKM1, into the nucleus in glioblastoma cells, breast cancer cells, and prostate cancer cells.⁷14 Inhibition of ERK1/2 MAP kinase by its inhibitor or expression of its kinase-dead mutant blocks the translocation of PKM2 into the nucleus.⁷ The specific regulation of PKM2 by ERK1/2 was further revealed by the finding that exon 10 of the PKM gene, which encodes specifically for PKM2 (and not PKM1), contains a docking domain for binding to ERK1/2. Expression of the PKM2 I429R/L431R docking domain mutant and an ERK2 docking groove mutant with the mutation in a common docking domain (D316/319N) and the glutamic acid–aspartic acid pocket (T157/158E) was shown to block the interaction between PKM2 and ERK1/2.⁷ In addition, ERK1/2 phosphorylates S37 of PKM2 but not PKM1, which provides a binding motif for interaction with the peptidyl-proline isomerase protein interacting with never in mitosis gene A 1 (PIN1) and leads to cis–trans isomerization of PKM2.⁷ Importin α, which functions as an adaptor and links nuclear localization signal (NLS)-containing proteins to importin β, binds to the NLS of PKM2, which is also encoded by exon 10 of the PKM gene and transports PKM2 into the nucleus.⁷ The PKM2 S37A phosphorylation mutant and the PKM2 R399/400A (NLS) mutant lose their ability to interact with PIN1 and importin α, respectively, and fail to translocate into the nucleus.⁷

In the absence of EGF treatment, PKM2 is localized primarily in the cytosol.⁷ Thus, the PKM2 NLS may be inter- or intramolecularly masked, preventing it from being recognized by importins. This hypothesis is supported by the in vitro evidence that the phosphorylation-mimic PKM2 S37D mutant does not bind to importin α5.⁷ However, inclusion of PIN1, which promotes the cis–trans isomerization of PKM2 S37D, enables PKM2 S37D to bind to importin α5.⁷ In addition, in vivo experiments showed that PKM2 S37D, which is able to translocate into the nucleus in PIN1 ^{cells, fails to do so in} PIN1 ^cells.7 Furthermore, R399/400 in the NLS of PKM2 is required for PKM2 to bind to importin α5 and translocate into the nucleus in the presence of PIN1.⁷ These results strongly suggest that phosphorylation of PKM2 at S37 leads to cis–trans isomerization of PKM2 by PIN1 to expose the PKM2 NLS to importin α5. In combination with the finding that the recruitment of PIN1 to PKM2 is mediated by both the binding of the ERK2 docking groove to the PKM2-specific exon 10-encoded I429/L431 region and the phosphorylation of PKM2 at S37,⁷ these results highlight the significance of precise and sequential post-translational modifications of PKM2 and its interactions with ERK1/2, PIN1, and importin α5 in its nuclear translocation and subsequent compartment-dependent functions.

PKM2 exists in a highly active tetrameric form, which has a high affinity for its substrate PEP, and in less active dimeric and monomeric forms, with low affinities for PEP.¹⁵ Co-expression of HA-tagged wild-type (WT) PKM2 and Flag-tagged PKM2 S37A in 293T cells and immunoblotting of immunoprecipitated WT HA-PKM2 with anti-Flag antibody showed that PKM2 S37A interacted with WT PKM2 (Fig. 1A), suggesting that PKM2 S37A can form dimers or tetramers with WT PKM2. However, PKM2 S37A is unable to translocate into the nucleus.⁷ In line with these results, co-expression of HA-tagged WT PKM2 and Flag-tagged PKM2 S37D in 293T cells showed that the latter, but not the former, accumulated in the nucleus, indicating that PKM2 S37D was unable to translocate WT PKM2 into the nucleus (Fig. 1B). These results strongly suggest that monomeric PKM2 translocates into the nucleus.

Figure 1. Monomeric PKM2 translocates into the nucleus. (A) 293T cells were co-transfected with plasmids expressing HA-tagged WT PKM2 and Flag-tagged PKM2 S37A. Immunoprecipitation (IP) and western blotting (WB) analyses were performed with the indicated antibodies. (B) 293T cells were co-transfected with plasmids expressing HA-tagged WT PKM2 and Flag-tagged PKM2 S37D. Total cell lysates and nuclear fractions were prepared and immunoblotting analyses were performed with the indicated antibodies. (C) Bacterially purified recombinant His-tagged PKM2 S37D mutant was mixed in the absence of (left) or with (right) purified GST-tagged PIN1 and subjected to an in vitro cis–trans isomerization assay and size-exclusion chromatography. The fractions were collected and immunoblotting analyses were performed with an anti-PKM2 antibody. Co-elutions occurred on the same column under identical conditions and were determined with a chromatography MW calibration standard. (D) Chromatographic elution profile of the MW calibration kit proteins (carbonic anhydrase, bovine erythrocytes, 29 kD; Albumin, bovine serum, 66 kD; Alcohol Dehydrogenase, yeast, 150 kD; β-Amylase, sweet potato, 200 kD) on the HiPrep 16/60 Sephacryl S-200 HR column. Ve, elution volume of the protein of interest; Vo, void volume as determined according to the manufacturer’s instructions.

To examine whether PIN1-dependent cis–trans isomerization regulates the intermolecular interaction of PKM2, we mixed bacterially purified recombinant PKM2 S37D mutant with purified PIN1 and performed an in vitro cis–trans isomerization assay followed by size-exclusion chromatography and immunoblotting analyses with an anti-PKM2 antibody. PKM2 in the absence of PIN1 was mostly in tetramer form; few dimers or monomers were detected (Fig. 1C, left). According to the molecular weight (MW) calibration standard (Fig. 1D), volumes 40, 42, and 44 (containing PKM2 tetramers) co-eluted with a MW near 240 kDa, volumes 50, 52, and 54 (dimers) near 120 kDa, and volumes 60, 62, and 64 (monomers) near 60 kDa (Fig. 1C, left). The presence of PIN1 reduced the total amount of PKM2 in tetramer form and increased the amount of PKM2 in monomer form (Fig. 1C, right). These results strongly suggest that PIN1-dependent cis–trans isomerization of PKM2 alters PKM2 structure and promotes the conversion of tetrameric PKM2 to monomeric PKM2.

The ability of R399 in PKM2 NLS to form stable charge–charge interactions with PKM2 residues E418 and E396 (located on the other dimer of the tetramer PKM2) and its location in the interface are important for PKM2 tetramer formation.¹⁶ Inter- and intramolecular PIN1-dependent structural changes and PKM2 monomer formation expose the sterically buried PKM2 NLS and make the NLS accessible for interaction with importin α. Considering that a portion of PKM2 also exists as monomers before ERK1/2-dependent phosphorylation, the possibility that PIN1 also acts as a cis–trans isomerase for monomeric PKM2 cannot be excluded. Taken together, these findings strongly suggest that monomeric PKM2 translocates into the nucleus.

Nuclear PKM2 Regulates Cell Cycle Progression and the Warburg Effect

Besides interacting with HIF1α and STAT3 to regulate their downstream gene expression,¹⁷16 nuclear PKM2 binds to c-Src-phosphorylated β-catenin at Y333, which is required for β-catenin activation. Because β-catenin binds to WT PKM2 but not to the PKM2 K433E mutant (which loses its ability to bind tyrosine-phosphorylated proteins),¹⁴ β-catenin very likely associates with only monomeric PKM2 to regulate gene expression. The interaction between PKM2 and β-catenin is required for both proteins, which complex with TCF4 to bind to histone H3 at the CCND1 (encoding for cyclin D1) and MYC promoters.¹⁴18 An important finding was that PKM2 functions as a protein kinase and phosphorylates histone H3 at T11, which dissociates histone deacetylase 3 from the promoter regions, leading to subsequent histone H3-K9 acetylation and gene transcription.¹⁹

Histone H3 is coded by several genes in the human genome and expresses as H3.1, H3.2, and H3.3 (H3.3 having H3.3A, H3.3B, and H3.3C isoforms encoded by H3F3A, H3F3B, and H3F3C, respectively), which have highly conserved sequences that differ by only a few amino acids.²⁰21 Real-time quantitative polymerase chain reaction (PCR) analysis showed that H3.3 is the primary histone H3 in some tumor cells, including U251 (Fig. 2) and U87 (data not shown) human glioblastoma cells; this finding was supported by a large reduction in total H3 expression induced by H3.3 shRNA, as detected by an antibody that recognizes all H3 variants.¹⁹ Reconstituted expression of the RNAi-resistant H3.3B-T11A mutant in tumor cells with depleted endogenous H3 expression blocks EGF-induced histone H3-K9 acetylation, expression of genes, including CCND1 and MYC, and cyclin D1-dependent cell cycle progression.¹⁹ c-Myc expression, in turn, upregulates the expression of PTB and glycolytic enzymes including GLUT1 and LDHA, and PTB, in turn, splices PKM pre-mRNA into PKM2 mRNA for subsequent upregulation of PKM2 expression. In this way, nuclear PKM2-dependent expression of these glycolytic enzymes promotes the Warburg effect.⁷ Of note, expression of PKM2 S37A mutant, which does not affect its glycolytic enzymatic activity and fails to translocate into the nucleus, blocks EGFR activation-promoted Warburg effect, cell cycle progression, and tumorigenesis.⁷

Figure 2. H3.3 is the major H3 isoform expressed in U251 human glioblastoma cells. Total mRNA of U251 cells was extracted, and the level the histone H3 isoforms was determined using quantitative real-time PCR with specific primers of histone H3.1/3.2 and histone H3.3. Data represent the mean ± SD of 3 independent experiments.

Summary

These findings described here have established that PKM2 acts as a glycolytic enzyme that regulates glycolysis in the cytosol but translocates into the nucleus upon receptor tyrosine kinase activation. In the nucleus, PKM2 couples with transcriptional factors and functions as a protein kinase that phosphorylates histone H3 for gene transcription.¹⁸1922 In addition, PKM2 directly regulates the expression of cyclin D1 for cell cycle progression. Moreover, PKM2-regulated upregulation of c-Myc expression promotes the expression of glycolytic enzyme genes, thereby promoting aerobic glycolysis in a feedback loop¹⁹; this nuclear function of PKM2 is essential to the Warburg effect, cell cycle progression, and tumorigenesis.

Materials and Methods

Materials

Rabbit polyclonal antibodies recognizing PKM2 and HA were obtained from Signalway Antibody. EGF and mouse monoclonal antibody for Flag were purchased from Sigma. HyFect transfection reagents were from Denville Scientific.

Cells and cell culture conditions

U87/EGFR GBM cells, U251 GBM cells, and human embryonic kidney 293T cells were maintained in Dulbecco modified Eagle medium supplemented with 10% bovine calf serum (HyClone). Cell cultures were made quiescent by growing them to confluence, and the medium was replaced with fresh medium containing 0.5% serum for 1 day. EGF at a final concentration of 100 ng/ml was used for cell stimulation.

Immunoprecipitation and immunoblotting analyses

Extraction of proteins with a modified buffer from cultured cells was followed by immunoprecipitation and immunoblotting with corresponding antibodies, as described previously.²³

DNA constructs and mutagenesis

PCR-amplified human PKM2 was cloned into pcDNA3.1/hygro(+) or pCEP4-HA vector. pcDNA 3.1/hygro(+)-PKM2 S37A and S37D were made by using the QuickChange site-directed mutagenesis kit (Stratagene).

Purification of recombinant proteins

WT and mutant His-tagged PKM2 and GST-tagged PIN1 proteins were expressed in bacteria and purified, as described previously.²⁴

Quantitative real-time PCR

Total RNA was extracted with use of an RNA High-Purity Total RNA Rapid Extraction Kit (Signalway Biotechnology). cDNA was prepared by using oligonucleotide (dT), random primers, and a Thermo Reverse Transcription Kit (Signalway Biotechnology). Quantitative real-time PCR analysis was performed using 2× SIBR Real-Time PCR Premixture (Signalway Biotechnology) under the following conditions: 5 min at 95 °C followed by 40 cycles at 95 °C for 30 s, 55 °C for 40 s, and 72 °C for 1 min using an ABI Prism 7700 sequence detection system. Data were normalized to expression of a control gene (β-actin) for each experiment.

The following primer pairs were used for quantitative real-time PCR: histone H3.1/3.2, 5′-AGCTCGGCCGTGATG-3′ (forward) and 5′-ATGCCCAAGGACATCCAGC-3′ (reverse); histone H3.3, 5′-AGCGCAGCCATCGGT-3′ (forward) and 5′-ATGCCCAAGGACATCCAGC-3′ (reverse).

Size-exclusion chromatography

Size-exclusion chromatography was performed with a HiPrep 16/60 Sephacryl S-200 HR Column (GE Healthcare Life Sciences). The samples of bacterially purified PKM2 or the reaction mix of bacterially purified PKM2 and PIN1 was loaded on the column and eluted with elution buffer (50 mM phosphate, 0.15 M NaCl [pH 7.2]). The fractions of each 1 ml were collected, and 25 μl of each fraction was analyzed by immunoblotting analyses. The elution profiles were compared with that of a size-exclusion chromatography calibration kit (Sigma) under identical conditions.