Phosphorylation of Bad at Thr-201 by JNK1 Promotes Glycolysis through Activation of Phosphofructokinase-1*

The mitogen-activated protein kinase JNK1 suppresses interleukin-3 withdrawal-induced cell death through phosphorylation of the BH3-only pro-apoptotic Bcl-2 family protein Bad at Thr-201. It is unknown whether JNK1 regulates glycolysis, an important metabolic process that is involved in cell survival, and if so, whether the regulation depends on Thr-201 phosphorylation of Bad. Here we report that phosphorylation of Bad by JNK1 is required for glycolysis through activation of phosphofructokinase-1 (PFK-1), one of the key enzymes that catalyze glycolysis. Genetic disruption of Jnk1 alleles or silencing of Jnk1 by small interfering RNA abrogates glycolysis induced by growth/survival factors such as serum or interleukin-3. Proteomic analysis identifies PFK-1 as a novel Bad-associated protein. Although the interaction between PFK-1 and Bad is independent of JNK1, Thr-201 phosphorylation of Bad by JNK1 is required for PFK-1 activation. Thus, our results provide a novel molecular mechanism by which JNK1 promotes glycolysis for cell survival.

The mitogen-activated protein kinase JNK1 suppresses interleukin-3 withdrawal-induced cell death through phosphorylation of the BH3-only pro-apoptotic Bcl-2 family protein Bad at Thr-201. It is unknown whether JNK1 regulates glycolysis, an important metabolic process that is involved in cell survival, and if so, whether the regulation depends on Thr-201 phosphorylation of Bad. Here we report that phosphorylation of Bad by JNK1 is required for glycolysis through activation of phosphofructokinase-1 (PFK-1), one of the key enzymes that catalyze glycolysis.
Bad is a BH3-only pro-apoptotic Bcl-2 family protein, having two splicing forms Bad s and Bad L (34). Bad plays a critical role in the cross-talk between the growth/survival factor signaling pathway and the intrinsic death machinery (35). The pro-apoptotic activity of Bad is regulated by extracellular stimuli via post-translational modifications including phosphorylation (35)(36)(37)(38). In response to the stimulation of growth/survival factors such as interleukin-3 (IL-3) and insulin-like growth factor-I (IGF-I), Bad is phosphorylated at the "regulatory serines" (Ser-112, Ser-136, and Ser-155) by a group of protein kinases such as cAMP-dependent protein kinase, Akt, Rsk, and p70 S6K (37)(38)(39)(40). The hyperphosphorylated form of Bad is inactive in terms of its pro-apoptotic function as it is sequestrated in the cytoplasm via its interaction with 14-3-3, a group of cytoplasmic anchorage proteins that interact with proteins through specific phospho-serine/threonine motifs (35). Withdrawal of growth/survival factors leads to dephosphorylation of Bad, which translocates to the mitochondria, where it binds prosurvival Bcl-2 family proteins such as Bcl-X L and thereby allows the oligomerization of Bax and Bak (41). This results in increased mitochondria permeability, release of cytochrome c, caspase activation, and ultimately, apoptotic cell death. Recently, it has been shown that JNK suppresses IL-3 withdrawal-induced apoptosis in hematopoietic pro-B FL-5.12 cells via phosphorylation of Bad at Thr-201 (42). Phosphorylation of Bad by JNK at Thr-201 is also required for its interaction with 14-3-3 (42).
In addition to regulation of apoptosis, Bad is involved in regulation of glycolysis, a metabolic process that is required for cell survival (43). It has been shown that inactivation of Bad by phosphorylation at the regulatory serines (Ser-112, Ser-136, and Ser-155) is required for the activity of mitochondrial glucokinase, one of the key enzymes that are responsible for glycolysis and ATP production (43). Thus, inactivation of Bad may be critical for suppression of apoptosis and maintaining of glycolysis. Here we report that JNK1-mediated phosphorylation of Bad at Thr-201 is required for maintaining normal glycolysis for cell survival through activation of phosphofructokinase-1 (PFK-1), which is another key enzyme in glycolysis.
Mass Spectrometry-Bad-associated proteins were immunoprecipitated from 8 mg of cytosolic extracts of FL5.12/Bcl-X L cells or FL5.12 Bcl-X L /Bad stable cells. Proteins were separated by SDS-PAGE. Gels were stained with colloidal Coomassie Blue, and protein bands of interest were excised and in-gel digested. The resultant tryptic peptides were identified as described previously (45).
Protein Kinase Assays and Immunoblotting-Immune complex kinase assays and glutathione S-transferase (GST)-c-Jun pull down kinase assays were performed and quantitated as described previously (8). Briefly, kinase assays were carried out for 60 min at 30°C in 20 mM HEPES (pH 7.6), 20 mM MgCl 2 , 1 mM dithiothreitol, 10 M nonradioactive ATP, and 10 Ci of [␥-32 P]ATP. WT or mutant GST-Bad or GST-c-Jun (2-4 g) was used as substrate. Reactions were terminated by the addition of 4ϫ SDS sample buffer and heated at 95°C for 5 min. The proteins were separated by SDS-PAGE and visualized by Coomassie Brilliant Blue staining. Phosphorylated proteins were detected and quantitated by PhosphorImager (GE Healthcare). Active JNK1 was immunoprecipitated from the lysates prepared from cells stimulated by UV (60 J/m 2 , 2 min), using anti-JNK1 antibody (antibody 333). Immunoblotting was performed as described previously (46).
Measurement of Glycolysis-Glycolysis was measured by monitoring the conversion of [5-3 H]glucose to [ 3 H]H 2 O as described previously (48). Briefly, 10 6 cells were washed once in phosphate-buffered saline prior to resuspension in 1 ml of Krebs buffer followed by incubation for 30 min at 37°C. Cells were pelleted, resuspended in 0.5 ml of Krebs buffer containing 10 mM glucose. The reaction was started by the addition of 10 mCi of [5-3 H]glucose. After 1 h at 37°C, triplicate 50-l aliquots were transferred to uncapped PCR tubes containing 50 l of 0.2 N HCl. Each tube was transferred to a scintillation vial containing 0.5 ml of H 2 O such that the water in the vial and the contents of the PCR tube were not mixed with each other. The vials were sealed, and diffusion was allowed to occur for 3 days. The amounts of diffused and undiffused 3 H were determined by scintillation counting. Appropriate [ 3 H]glucose-only and 3 H 2 O-only controls were included for the calculation of 3 H 2 O in each sample that reflects the rate of glycolysis.
In Vitro Translation-The pCDNA3.1-based PFK-1 and Bcl-X L were translated in vitro using the rabbit reticulocyte lysate TNT-coupled transcription/translation system (Promega, Madison, WI) following the manufacturer's instructions.
siRNA Transfection and Retroviral Infection-siRNA specific for murine JNK1 or green fluorescent protein was designed by using the SiDESIGN center (Dharmacon, Lafayette, CO). The target sequence for Jnk1 is 5Ј-TGATTCAGATGGAGTTAGA-3Ј, and for Gfp, the sequence is 5Ј-GAAGAAGATGGT-GCGCTCC-3Ј. The siRNA was transfected into cells using Lipofectamine 2000 following the manufacturer's instructions. Retroviral particles were produced using the pSOS vector sys-tem, as described previously (49). The FL5.12 Bcl-X L /M2-Bad cells were infected with the retroviruses four consecutive times (with 6-h intervals). After 48 h, the infection procedure was repeated to enrich the frequency of infection. The bulk infected cells were collected for experiments 48 -72 h after the second round of infection.
Statistic Analysis-All statistic analyses were performed by using the Student's t test. A probability level of 0.05 or less was considered as significant. Data are presented as mean values Ϯ standard error.

JNK1 Is Essential for IL-3-or Serum-induced Glycolysis-
Maintenance of normal glycolysis is important for cell survival (50 -52). The observations that phosphorylation of Bad at the regulatory serines (Ser-112, Ser-136, and Ser-155) is required for glycolysis (43), and JNK-mediated phosphorylation of Bad at Thr-201 for cell survival (42) prompted us to test whether JNK1 regulates glycolysis, using hematopoietic pro-B FL5.12 cells stably expressing Bcl-X L and Bad (for simplicity, referred to as FL-5.12 cells hereinafter unless indicated otherwise) (36,42). Cells were subjected to IL-3 withdrawal for 2 h followed by the readdition of IL-3 for 15 min in the presence or absence of the specific JNK inhibitor SP600125. Under this condition, JNK1 activity was significantly reduced upon IL-3 withdrawal, stimulated by IL-3 readdition, and inhibited by SP600125 (Fig.  1A), consistent with the previous report (42). Metabolically labeling the cells with [5-3 H]glucose revealed that the glycolytic rate was reduced following IL-3 withdrawal but became robust upon IL-3 readdition (Fig. 1A), consistent with the previous report that IL-3 up-regulates glycolysis in parental FL5.12 cells (53). The JNK inhibitor SP600125 significantly blocked IL-3 readdition-stimulated glycolysis (Fig. 1A). Similar results were obtained with FL-5.12 cells expressing specific retroviral  shRNA of JNK1 but not control shRNA (Fig. 1B). The stimulation of glycolysis by IL-3 readdition was modest (Fig. 1B). This is most likely due to the high basal level of glycolysis that may be independent on IL-3.
This notion was further supported by the observation that in immortalized MEFs, silencing of JNK1 by siRNA or genetic disruption of Jnk1 alleles abolished fetal bovine serum (FBS)-induced glycolysis (Fig. 1, C and D). Thus, JNK1 is required for growth factor-induced glycolysis. Given the fact that up-regulation of glycolysis is a direct effect of IL-3, rather than a compensatory result of growth factor stimulation of cell growth and proliferation (53), IL-3-induced JNK1-dependent higher glycolytic rate may be beneficial to cell survival under stress conditions. JNK Regulates Activation of PFK-1-Hexokinase and PFK-1 are two rate-limiting enzymes in the glycolytic process (54 -56). Previous studies have shown that phosphorylation of Bad at Ser-112, Ser-136, and Ser-155 by several protein kinases (cAMP-dependent protein kinase, Akt, and Rsk) is required for activation of glucokinase, which is one of the hexokinase isoforms (57,58), thereby regulating the glycolytic process (43). Since Bad is also phosphorylated at Thr-201 by JNK1, we were curious about whether JNK1 regulates glycolysis through modulation of the activity of PFK-1 and hexokinase. In IL-3dependent FL-5.12 cells, PFK-1 had high basal activity, as measured by PFK-1 assay (Fig. 2A). The activity of PFK-1 was partially reduced by IL-3 withdrawal and further enhanced by IL-3 readdition ( Fig.  2A). This suggests that PFK-1 activity is likely controlled by IL-3dependent and -independent mechanisms. However, the augmented PFK-1 activity by IL-3 readdition was completely abrogated by the JNK inhibitor SP600125 ( Fig. 2A) or retroviral shRNA of JNK1 (Fig. 2B). Thus, the stimulation of PFK-1 by IL-3 depends on JNK1 activity. Consistent with this conclusion, FBS induced significant increase of PFK-1 activity in MEFs but was unable to do so in MEFs expressing siRNA of JNK1 or in JNK1 null MEFs (Fig. 2, C and D). Similar results were obtained with hexokinase, although to a less extent (data not shown). Taken together, JNK1 may regulate IL-3-or FBS-stimulated glycolysis by, at least in part, modulating the activity of PFK-1.
PFK-1 Is a Novel Bad-associated Protein-The observations that JNK1 inhibits IL-3 withdrawal-induced cell death through phosphorylation and inactivation of Bad (42) and that JNK1 A, cell extracts of FL5.12/Bcl-X L and FL-5.12 were subjected to immunoprecipitation with anti-M2 antibody, followed by SDS-PAGE, and stained with colloidal Coomassie Brilliant Blue (CBB). Transfected M2-Bad was analyzed by immunoblotting using anti-M2 antibody. The arrow indicates the major Bad-associated protein, which was identified as PFK-1 (see panel B below). H, IgG heavy chains; L, IgG light chains. B, the major Bad-associated protein identified in panel A was excised and sequenced by nano-HPLC electrospray ion trap mass spectrometry, as described under "Materials and Methods." C, COS-1 cells were transfected with mammalian expression vectors encoding PFK-1 or M2-Bad (2 g of each), without or with empty vector (2 g). After 40 h, M2-Bad was immunoprecipitated (IP) with anti-M2 antibody, and M2-Bad-associated PFK-1 was immunoblotted (IB) with anti-Xpress antibody. Total M2-Bad or Xpress-PFK-1 proteins in the cell extracts were analyzed by immunoblotting with anti-Bad or anti-Xpress antibody, respectively. D, M2-Bad was immunoprecipitated from stable FL5.12/Bcl-X L or FL-5.12 cells with anti-M2 antibody, and M2-Bad-associated PFK-1 was analyzed by immunoblotting using anti-PFK-1 antibody. The amounts of immunoprecipitated M2-Bad proteins and total PFK-1 proteins were examined by immunoblotting with anti-Bad and anti-PFK-1 antibody, respectively. E, purified GST-Bad proteins were incubated without or with active JNK1 in a kinase reaction buffer in the presence of [␥-32 P] ATP (top two panels, KA for kinase assays) or 17 M nonradioactive ATP (low two panels). GST-Bad proteins modified with nonradioactive ATP were mixed with cell extracts of COS-1 cells that have been transfected with mammalian expression vector encoding Xpress-PFK-1. Xpress-PFK-1 was immunoprecipitated with anti-Xpress antibody, and Xpress-PFK-1associated GST-Bad proteins were analyzed by immunoblotting using anti-Bad antibody. The amount of Xpress-PFK-1 in the immune complex was analyzed by immunoblotting with anti-Xpress antibody. F, GST-Bad proteins modified with nonradioactive ATP (as described in panel E) were mixed with in vitro translated 35 Slabeled PFK-1 or 35 S-labeled Bcl-X L for 12 h and extensively washed. GST-Bad-associated 35 S-PFK-1 or 35 S-Bcl-X L was visualized by radioautography. One-tenth of the input 35 S-PFK-1 or 35 S-Bcl-X L was analyzed by radioautography. The amount of GST-Bad proteins was examined by Coomassie Brilliant Blue staining. ns, nonspecific. JULY 25, 2008 • VOLUME 283 • NUMBER 30 regulates the rate of glycolysis and PFK-1 activity prompted us to examine the possible connection between PFK-1 and Bad. Using nano-HPLC electrospray ion trap mass spectrometry, we identified several proteins that were co-immunoprecipitated with Bad from FL-5.12 cells. One of Bad-associated proteins was PFK-1 (Fig. 3, A and B). Consistently, co-immunoprecipitation in combination of immunoblotting revealed that ectopically expressed Xpress-tagged PFK-1 or endogenous PFK-1 proteins were co-immunoprecipitated with M2-Bad proteins (Fig. 3, C and D). However, the interaction between Bad and PFK-1 was not affected by IL-3 (data not shown). To verify this observation, we employed an in vitro binding assay in which purified GST-Bad proteins were first phosphorylated by active JNK1 in the presence of either [␥-32 P]ATP (Fig. 3E, top panel, positive control) or nonradioactive ATP. GST-Bad proteins phosphorylated with nonradioactive ATP were isolated by GSH beads and mixed with COS-1 cell extracts containing Xpress-PFK-1. Immunoprecipitation using anti-Xpress antibody in combination with immunoblotting using anti-Bad antibody revealed that both phosphorylated and nonphosphorylated GST-Bad proteins that were run at the position of 50 kDa in SDS-gel interacted with PFK-1 (Fig. 3E). In a parallel experiment, GST-Bad proteins phosphorylated with or without nonradioactive ATP were incubated with in vitro-translated 35 S-labeled PFK-1 or the positive control [ 35 S]-Bcl-X L . GST pull-down assays revealed that phosphorylation by active JNK1 significantly reduced the binding of GST-Bad to [ 35 S]-Bcl-X L (Fig. 3F, low panel), consistent with the previous report (42). However, phosphorylation by active JNK1 had no detectable effect on the binding of GST-Bad to 35 S-labeled PFK-1 (Fig. 3F,  top panel). Thus, Bad directly interacts with the second ratelimiting enzyme of glycolysis, PFK-1, and the interaction is independent of its phosphorylation by JNK1. This is in analogy to the previous report that the interaction between Bad and glucokinase is also independent of Bad phosphorylation at the regulatory serines (43). It has been shown that Bad is involved in the assembly of the protein complex containing. The catalytic subunit of PP1 and cAMP-dependent protein kinase, glucokinase, and WAVE-1 (43). Future work is needed to determine whether PFK-1 is also a component of this complex.

Phosphorylation of Bad by JNK1 Promotes Glycolysis
Phosphorylation of Bad at Thr-201 by JNK1 Is Required for PFK-1 Activation and Glycolysis-Although the interaction between Bad and PFK-1 is independent of Bad phosphorylation by JNK1, it is possible that the phosphorylation is required for Bad to regulate PFK-1 activity. To test this hypothesis, FL5.12/ Bcl-X L cells were transiently transfected with WT M2-Bad or M2-Bad(T201V) in which threonine 201 has been replaced by valine. Although WT Bad was significantly phosphorylated by active JNK1, Bad(T201V) mutant was not, as measured by in vitro kinase assays (Fig. 4A). IL-3-withdrawal partially reduced PFK-1 activity, and its readdition significantly stimulated PFK-1 activity in cells expressing WT M2-Bad (Fig. 4B), consistent with the results obtained by using stable FL-5.12 cells (Fig. 2). In contrast, IL-3-withdrawal or readdition only slightly affected the activity of PFK-1 in cells expressing M2-Bad(T201V) mutant (Fig. 4B). Similarly, the effect of IL-3 on the rate of glycolysis was also greatly reduced in cells expressing M2-Bad(T201V) mutant when compared with that in cells expressing WT M2-Bad (Fig. 4C). The difference in IL-3-induced PFK-1 activity or glycolysis between FL5.12 Bcl-X L /Bad and FL5.12 Bcl-X L /Bad(T201V) cells was not the result of the difference in expression levels of M2-Bad and M2-Bad(T201V), as analyzed by immunoblotting using anti-M2 antibody (Fig. 4D). These results indicate that Thr-201 phosphorylation of Bad is required for the regulation of PFK-1 activity and glycolysis. This is in analogy to the previous report that the phosphorylation of Bad at the regulatory serines is required for activation of glucokinase (43). Future work is needed to determine how phosphorylation of Bad at Thr-201 by JNK1 or the regulatory serines regulates the activity of PFK-1 or glucokinase, respectively.
It is possible that JNK1 may regulate PFK-1 activity and glycolysis in Bad-dependent and -independent mechanisms. We reasoned that if JNK1 regulates PFK-1 activity and glycolysis only through phosphorylation of Bad at Thr-201, inactivation of JNK1 should not further affect PFK-1 activity and glycolysis in cells expressing M2-Bad(T201V) mutant. To test this hypothesis, Bad null MEFs were transiently transfected with WT M2-Bad, M2-Bad(T201V), or empty vector, with or without siRNA of JNK1 or control siRNA. Immunoblotting analysis revealed that expression of JNK1 was significantly silenced by its specific siRNA but not the control siRNA (Fig. 5A), whereas the expression of M2-Bad and M2-Bad(T201V) were at similar levels (Fig. 5A). PFK-1 assay showed that in Bad null cells expressing WT M2-Bad, FBS significantly stimulated the activity of PFK-1, and the stimulation was reduced by siRNA of JNK1 (Fig. 5B). However, in Bad null cells expressing M2-Bad(T201V) mutant, FBS only slightly stimulated PFK-1 activity, and more importantly, siRNA of JNK1 had no detectable effect on PFK-1 activity (Fig. 5B). Similar results were obtained when the glycolytic rate was analyzed (Fig. 5C). Thus, JNK1 regulates PFK-1 activation and glycolysis through phosphorylation of Bad at Thr-201 in response to IL-3 or FBS.
In summary, our results suggest that nonphosphorylated Bad is ready to interact with PFK-1, but only JNK1-phosphorylated Bad is able to regulate PKF-1 activity. Thus, in addition to inhibition of the pro-apoptotic function of Bad, JNK1-mediated phosphorylation of Bad at Thr-201 is required for regulating the activity of PFK-1, which in turn promotes glycolysis for cell survival (Fig. 5D).