ERK1/2 Regulates Intracellular ATP Levels through α-Enolase Expression in Cardiomyocytes Exposed to Ischemic Hypoxia and Reoxygenation*

Extracellular signal-regulated kinase 1/2 (ERK1/2) is known to function in cell survival in response to various stresses; however, the mechanism of cell survival by ERK1/2 remains poorly elucidated in ischemic heart. Here we applied functional proteomics by two-dimensional electrophoresis to identify a cellular target of ERK1/2 in response to ischemic hypoxia. Approximately 1500 spots were detected by Coomassie Brilliant Blue staining of a sample from unstimulated cells. The staining intensities of at least 50 spots increased at 6-h reoxygenation after 2-h ischemic hypoxia. Of the 50 spots that increased, at least 4 spots were inhibited in the presence of PD98059, a MEK inhibitor. A protein with a molecular mass of 52 kDa that is strongly induced by ERK1/2 activation in response to ischemic hypoxia and reoxygenation was identified as α-enolase, a rate-limiting enzyme in the glycolytic pathway, by liquid chromatography-mass spectrometry and amino acid sequencing. The expressions of the α-enolase mRNA and protein are inhibited during reoxygenation after ischemic hypoxia in the cells containing a dominant negative mutant of MEK1 and treated with a MEK inhibitor, PD98059, leading to a decrease in ATP levels. α-Enolase expression is also observed in rat heart subjected to ischemia-reperfusion. The induction of α-enolase by ERK1/2 appears to be mediated by c-Myc. The introduction of the α-enolase protein into the cells restores ATP levels and prevents cell death during ischemic hypoxia and reoxygenation in these cells. These results show that α-enolase expression by ERK1/2 participates in the production of ATP during reoxygenation after ischemic hypoxia, and a decrease in ATP induces apoptotic cell death. Furthermore, α-enolase improves the contractility of cardiomyocytes impaired by ischemic hypoxia. Our results reveal that ERK1/2 plays a role in the contractility of cardiomyocytes and cell survival through α-enolase expression during ischemic hypoxia and reoxygenation.

more, ␣-enolase improves the contractility of cardiomyocytes impaired by ischemic hypoxia. Our results reveal that ERK1/2 plays a role in the contractility of cardiomyocytes and cell survival through ␣-enolase expression during ischemic hypoxia and reoxygenation.
The extracellular signal-regulated kinase 1/2 (ERK1/2) 1 pathway, which comprises mitogen-activated protein kinase/ extracellular signal-regulated kinase kinase (MEK) and ERK1/2, is involved in various cellular processes such as cell proliferation, cell differentiation, cell death, cell survival, and cell motility in different cell types (1)(2)(3)(4). Of these diverse processes, cell survival and cell death play crucial roles in cardiomyocytes (5), since cardiomyocytes, which are terminally differentiated, cannot proliferate, even when they are damaged, the damage can lead to cell death in the case of serious diseases such as myocardial infarction and myocarditis (6). In human heart failure, apoptotic cell death has been observed to some extent, although the rate of apoptosis is controversial (1-35%) as samples are limited to post-mortem and heart transplantation specimens (7)(8)(9). Recent studies have identified low levels of myocyte apoptosis (about 1%) in failing human heart (9 -11). In heart failure, idiopathic dilated cardiomyopathy and ischemic cardiomyopathy are closely associated with apoptosis (12). Wencker et al. (13) demonstrated that very low levels of myocyte apoptosis (0.01%) are sufficient to cause a lethal, dialated cardiomyopathy.
We have shown that ischemia induces the translocation of ERK1/2 from the cytosol to the nucleus and that ERK1/2 is activated during reperfusion in rat heart (14). In cardiomyocytes subjected to ischemia-reperfusion, reactive oxygen species such as O 2 Ϫ , hydroxyl radical, and hydroperoxide are released from cardiomyocytes themselves dependent on the duration of both ischemia and reperfusion (15). Reactive oxygen species can activate various intracellular signal transduction pathways including the MAPK family and protein kinase C (PKC) isoforms (15,16). Noshita et al. (17) demonstrated that copper/zinc superoxide dismutase, an antioxidant, attenuates neuronal cell death by preventing ERK1/2 activation in focal brain ischemia using transgenic mice. In neonatal cardiomyocytes, activated ERK1/2 inhibits cell death by oxidative stress, indicating that the reactive oxygen species produced may activate the ERK1/2 pathway, including PKC isoforms, leading to cell survival in heart (18 -20). Indeed, transgenic mice that express an activated form of MEK1 show resistance to ischemia/reperfusion-induced apoptosis and a dramatic increase in cardiac function by echocardiography in isolated working heart preparations, without signs of decompensation (21). Two pathways for ERK1/2-induced cell survival, transcription-independent and -dependent pathways, have been reported (22,23). In the transcription-independent pathway, ERK1/2 activation by survival factors inhibits apoptosis through RSK, which inactivates the proapoptotic protein BAD via the phosphorylation of Ser 112 . The phosphorylation of BAD induces the association of BAD to 14 -3-3 from Bcl-XL, promoting cell survival (24). ERK1/2 activation also inhibits caspase-8 activity and the cleavage of Bid, leading to the inhibition of CD95/Fas-mediated cell death in T cells (25,26). On the other hand, in the transcription-dependent pathway, the mechanism by which ERK1/2 protects cells from apoptosis is complex and remains unknown.
In preceding studies, we demonstrated that PKC␣, PKC␦, PKC⑀, and PKC isoforms translocate to distinct fractions during ischemia (27). In particular, PKC, which is expressed mainly in the rat heart, translocates from the cytoplasm to the nucleus during ischemia (28). We have indicated that during reperfusion after ischemia, an atypical PKC participates in the activation of nuclear ERK1/2 through MEK (14,28,29). The nuclear MEK/ERK1/2 pathway induces rapid gene expression under pathophysiological conditions to prevent cell death, since the expression of c-jun and c-fos mRNAs rapidly increases for only 10 min after ischemia (30). Recently, in an analysis of transcriptional products produced by ERK1/2 activation, Lewis et al. (31) revealed that the proteomics approach to monitoring changes in protein profiles during signaling events using twodimensional gel electrophoresis is a most useful approach. Twenty-five previously unrecognized proteins were identified as downstream targets of the ERK1/2 pathway by this approach (31). However, cell survival factors induced by ERK1/2 in response to ischemia remain unreported, presumably because ERK1/2 has cell-specific downstream targets in ischemic heart. Therefore, we applied the strategy of functional proteomics to identify targets of the ERK1/2-specific pathway involved in cell survival in response to ischemic hypoxia using cultured cardiomyocytes. In the present study, we identify for the first time that ␣-enolase, a rate-limiting enzyme in the glycolysis pathway, is a downstream factor of ERK1/2 in response to ischemic hypoxia. We also show that ERK1/2 is involved in the maintenance of intracellular ATP levels through ␣-enolase expression during reoxygenation after ischemic hypoxia, leading to cell survival and the recovery of contractility in cardiomyocytes.

EXPERIMENTAL PROCEDURES
Materials-Anti-␣-enolase antibody and anti-ERK1/2 antibody were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). ␣-Enolase protein was from Sigma and was purified on an anion exchange column (Resourse Q, Applied Biosystems). A vector of pEBB-HIF-1␣ was purchased from Novus Biologicals Inc. A vector of pcDNA3-c-myc was provided from Dr. Yoshiyuki Kuchino, and a vector of pSV-HA-PGC-1 was provided from Prof. Bruce M. Spiegelman. Water passed through a MilliQ filter was used in all experiments (synthesis A10, Millipore, Bedford, MA). All other chemicals were commercially available.
Cell Culture and Ischemic Cell Model-Primary cultures of 1-to 2-day-old neonatal rat ventricular myocytes were prepared as reported previously (32). The cardiomyocytes dissociated enzymatically were seeded on culture trays at a concentration of 3 ϫ 10 5 cells/cm 2 and then incubated in L-15 Medium (Worthington Biochemical Co.) supplemented with 5% fetal bovine serum at 37°C. The cell suspension was replated to reduce the fibroblast content. Cells were grown in a random orientation (isotropic growth) to make a confluent layer, and by day 5 they showed synchronized spontaneous beating with a regular frequency. The relative population of cardiomyocytes over nonmyocytes in the respective culture, which was estimated by immunolabeling against anti-actin antibody, was Ͼ97%. H9c2 cell culture was achieved as described previously (29,33,34). Simulated ischemia was achieved as described previously (29,33,34). Briefly, the cells were incubated in slightly hypotonic Hanks' balanced saline solution (1.3 mM CaCl 2 , 5 mM KCl, 0.3 mM KH 2 PO 4 , 0.5 mM MgCl 2 , 0.4 mM MgSO 4 , 69 mM NaCl, 4 mM NaHCO 3 , and 0.3 mM Na 2 HPO 4 ) without glucose or serum for 2 h at 37°C. Hypoxia was achieved using an air-tight incubator from which oxygen was removed by replacement with nitrogen. The oxygen concentration in the incubates was adjusted to 1%. After incubation under the conditions of ischemic hypoxia, the cells were incubated in medium without serum under normoxic conditions (20% O 2 , 5% CO 2 ) at 37°C for the indicated times.
Transfection with Vectors into Cells-Transfection of vectors into H9c2 cells was carried out as described previously (29,33). Briefly, H9c2 cells (typically 60% confluent in 60-mm dishes) were washed three times with phosphate-buffered saline. In the transfection of a pcDNA3 empty vector, a pSV-HA-PGC-1, a pcDNA3-c-myc, or a pEBB-HIF-1␣, appropriate dilutions of plasmid DNA in 800 l of serum-free Dulbecco's modified Eagle's medium including liposomes (Transfast, Promega Co., Madison, WI) were preincubated at room temperature for 10 min. The cells were incubated for 1 h at 37°C in the presence of 5% CO 2 . At the end of the incubation period, 4 ml of medium containing 10% fetal bovine serum was added. After incubation for 24 h, the cells were incubated in serum-free medium for 48 -96 h. To test the role of ERK1/2 activation in ischemic hypoxia and reoxygenation, we transfected MEK mutants with pMACS4.1 vector, which expresses a truncated CD4 surface marker, into H9c2 cells by the protocol as described above. The expressed cells were separated with magnetic microbead-bound antibody for the CD4 surface marker, and the cells were selected by G418 for 2 weeks. The cells were incubated in serum-free medium for 48 -96 h before the stimulation of ischemic hypoxia and reoxygenation. In transient transfection of H9c2 cells, we did not accurately determine the transfection efficiency, which was estimated at 20 -30% from the data by GFP expression vector. However, the cells introduced the exogenous genes became to ϳ100% by the selection of magnetic beads and G418 when transfection efficiency was measured using fluorescein isothiocyanate-labeled anti-CD4 antibody.
Transfection of vectors into cardiomyocytes was carried out as below. Briefly, isolated cardiomyocytes (typically 90% confluent in 6-well plates) were cultured in L-15 with 5% serum for a day before the transfection. For each well of cells, 4.0 g of DNA and 8 l of liposomes (LipofectAMINE 2000, Invitrogen, CA) were diluted into 200 l of L-15 without serum, respectively, and incubated for 5 min at room temperature. The diluted DNA was combined with the diluted liposomes, and incubated for 20 min at room temperature. The DNA-liposome complex was directly added into the well and incubated for 24 h at 37°C in the presence of 5% CO 2 . After incubation for 24 h, the cells were incubated in serum-free medium for 48 h and were used for the assay, since the gene introduced was strongly expressed at 48 -72 h after the transfection, although the cardiomyocytes exogenous gene was expressed at 10 -20%.
Introduction of Purified ␣-Enolase into Cells-Purified ␣-enolase (1.0 g/dish, 35-mm dishes) was diluted in phosphate-buffered saline and incubated with diluted Chariot TM (Active Motif, Carlsbad, CA) for 30 min at room temperature. The ␣-enolase-Chariot complex was added to cells that had been washed in phosphate-buffered saline after removal from the medium (35,36). After the addition of Dulbecco's modified Eagle's medium (for H9c2 cells) or L-15 (for cardiomyocytes) without serum and incubation at 37°C for 1 h, complete growth medium was added, and the cells were incubated at 37°C for 20 h. The introduction of proteins was confirmed using IgG-labeled Alexa488, which was detected in the cells more than 70% in H9c2 cells and cardiomyocytes.
Detection of Apoptosis-Total DNA was extracted from cells (typically 80% confluent in 35-mm dishes) using a nucleic acid extraction kit (Apoptosis Ladder Detection Kit, Wako Pure Chemicals, Osaka, Japan). Annexin V-EGFP staining was carried out according to instruction manual (Apo Alert Annexin V-EGFP, Clontech).
Ischemia and Reperfusion Procedure in Animal Model-The experimental protocol was approved by the University of Tokyo committees on animal experiments. Male Sprague-Dawley rats (250 -300 g, SLC Ja-pan) were anesthetized with pentobarbital sodium (50 mg/kg), intubated, and mechanically ventilated with a rodent ventilator with room air. The left anterior descending coronary artery was ligated as described previously (37). The ligature was released for 6 h after left anterior descending coronary artery ligation for 20 min (20-min ischemia and 6-h reperfusion). The ventricular tissues were prepared from two different areas: the ischemic area in the center of the territory of the left anterior descending coronary artery and the nonischemic area in the posterior part of the left ventricle far from the ischemic area. The tissues were then washed with cold saline and stored in RNAlater (Qiagen) for determination of mRNA or frozen in liquid nitrogen for determination of protein amounts.
Reverse Transcription (RT)-PCR Method-RT-PCR was performed as described previously (33,34). Briefly, total RNA was isolated from H9c2 cells or rat heart using Nucleo Spin RNA II kits (Clontech). RT was performed with 3 g of total RNA and oligo(dT) 12-18 primer (Invitrogen) using SUPERSCRIPT II (Invitrogen). Total RNAs were isolated from H9c2 cells, and RT was performed with 3 g total RNA. RT products were amplified by 23 cycles of PCR at 94°C for 15 s, 63°C for 30 s, 68°C for 1 min using TGGGTGATGAGGGTGGATTC and CTTTGAGCAGGAGGCAGTTG primers for ␣-enolase or by 23 cycles of PCR at 95°C for 30 s, 50°C for 30 s, 72°C for 1 min using ACCACAGTCCATGCCATCAC and TCCACCACCCTGTTGCTGTA primers for G3PDH.
Two-dimensional Electrophoresis-Two-dimensional electrophoresis was carried out as described previously with slight modifications (38). Briefly, cells were trypsinized and washed twice with phosphate-buffered saline and lysed in 60 l of sample buffer (8 M urea, 2% Triton X-100, 40 mM Tris, 1% dithioerythritol, 2.5 mM EDTA, 2.5 mM EGTA, 0.5% Pharmalyte, pH 3-10) per cellular pellet corresponding to ϳ2 ϫ 10 7 cells. DNA was sheared using a sonicator and removed by centrifugation at 100,000 ϫ g for 60 min. Isoelectric focusing was performed using a Amersham Biosciences dry strip kit. Fifty microliter samples were applied to the acidic part of Immobiline strips with a nonlinear gradient from pH 4 to 6 using sample cup holders. Strips were reswollen in 8 M urea, 0.15% dithioerythritol, 2% Triton X-100, 2.5 mM EDTA, 1% Pharmalyte, pH 4 -6. The second dimension was standard SDS-PAGE in 12% gels. Apparent molecular weights were determined from the migration of marker proteins separated in parallel processed gels. The spots in each gel were detected by Coomassie Brilliant Blue staining.
Amino Acid Sequence-Samples separated by two-dimensional electrophoresis were transferred to polyvinylidene difluoride membranes (Applied Biosystems). The immobilized proteins indicated by the arrow in Fig. 1H were reduced, S-carboxymethylated, and subjected to in situ digestion with Achromobacter protease I and then to reverse phase-high performance liquid chromatography (Wakosil-II AR C18, Wako Pure Chemicals) as described previously (39 -41). Amino acid sequencing was performed with a gas phase sequencer (model PPSQ-10, Shimadzu, Kyoto, Japan), and the peptides indicated by arrows were identified as described in the legend to Fig. 2D.
Measurement of Contractility in Cardiomyocytes-Cells were placed on the stage of an inverted microscope (Nikon, Eclipse TS100) and stimulated with 30 mA at a frequency of 2.5 Hz (10-ms duration) using a pair of platinum wires connected to a stimulator (Star Medical). The cardiomyocytes were displayed on the computer monitor with an Ion-Optix MyoCam camera and analyzed for cell length during shortening and relengthening using soft-edge software (IonOptix, Milton, MA).

Cell Survival Induced by ERK1/2 during Ischemic Hypoxia
and Reoxygenation-First we observed cell conditions in the presence of 50 M PD98059, a MEK inhibitor, to clarify the role of ERK1/2 activation during ischemia and reperfusion. Marked cell death was observed by treatment with PD98059 after 24 h of reperfusion following 2 h of ischemia in the cell model (ischemic hypoxia and reoxygenation) (Fig. 1, A-C). Cell death in the presence of PD98059 increased significantly following 12-h reoxygenation after 2-h ischemic hypoxia in a dose-dependent manner (Fig. 1, D and E). These results are similar to those in the presence of U0126, another MEK inhibitor (data not shown). Treatment with 50 M PD98059 resulted in the inhibition of ERK1/2 activity to the basal level, while the phosphorylation of c-Jun N-terminal protein kinase, necessary for ac- tivation, was unaffected by PD98059 during reoxygenation (data not shown). Treatment with H89, a cAMP-dependent protein kinase inhibitor, had no effect on cell survival during ischemic hypoxia and reoxygenation (data not shown). Next, we evaluated cells subjected to ischemic hypoxia and reperfusion for internucleosomal cleavage by testing for DNA laddering, a hallmark of apoptosis. Consistent with the decrease in cell number in the presence of PD98059, DNA fragmentation was observed after 24 h of reoxygenation (Fig. 1F). In the absence of PD98059, no DNA fragmentation was detectable in agarose gels after 24 h of reoxygenation following ischemic hypoxia under the conditions used in this study (Fig. 1F). Treatment A-C, H9c2 cells were pretreated with a pcDNA3 empty vector or a pcDNA3-MEKDN mutant (KA97) in addition to pMACS4.1 coding a truncated human CD4 surface molecule containing liposomes, and cells containing the introduced vectors were selected by magnetic bead-bound anti-CD4 antibody and by G418. Cells were exposed to ischemic hypoxia and reoxygenation for the indicated time, extracted, and subjected to immunoblotting with anti-␣-enolase antibody (A) and anti-ERK1/2 antibody (B) as described previously. The figures show representative immunoblots obtained from five independent experiments. The levels of ␣-enolase and ERK1/2 proteins were determined from the immunoblots by densitometric analysis (mean Ϯ S.E., n ϭ 5; *, p Ͻ 0.05 versus control; #, p Ͻ 0.05 versus empty vector) (C). D and E, cardiomyocytes were pretreated with a pcDNA3 empty vector or a pcDNA3-MEKDN mutant (KA97). Cells were exposed to ischemic hypoxia and reoxygenation for the indicated time, extracted, and subjected to immunoblotting with anti-␣-enolase antibody (D) and anti-ERK1/2 antibody (E). The figures show representative immunoblots obtained from three independent experiments. F-H, total RNAs were isolated from H9c2 cells, and RT was performed with total RNA. RT products were amplified by PCR using primers for ␣-enolase (F) or by PCR using primers for G3PDH (G). The figures show representative data obtained from four independent experiments. The levels of ␣-enolase and G3PDH mRNAs were determined by densitometric analysis (mean Ϯ S.E., n ϭ 4; *, p Ͻ 0.05 versus control; #, p Ͻ 0.05 versus empty vector) (H). I and J, total RNAs were isolated from cardiomyocytes, and RT was performed with total RNA. RT products were amplified by PCR using primers for ␣-enolase (I) or by PCR using primers for G3PDH (J). The figures show representative data obtained from three independent experiments. with PD98059 (50 M) alone for 24 h had no effect on the decline in cell number and DNA laddering, demonstrating that cell death is unaffected by treatment with PD98059. In addition, the staining of Annexin V-EGFP, a marker of apoptosis, also showed similar results (Fig. 1, G-J). These findings indicate that the MEK/ERK1/2 pathway activates cell survival during reoxygenation after ischemic hypoxia.
Identification of the Target Protein Induced by ERK1/2-To clarify the survival factor induced by ERK1/2 activation, we performed two-dimensional electrophoresis to separate whole proteins extracted from cardiomyocytes exposed to ischemic hypoxia and reoxygenation in the presence or absence of 50 M PD98059. Approximately 1500 spots were observed in the gel by Coomassie Brilliant Blue staining of a sample from unstimulated cells ( Fig. 2A). At 6-h reoxygenation after 2-h ischemic hypoxia, the staining intensities of at least 50 spots increased as compared with those in unstimulated cells (Fig. 2B). Of the 50 spots induced by reperfusion, at least 4 were signif-icantly suppressed to control levels or below by 50 M PD98059 (Fig. 2C). We first tried to determine the amino acid sequence of the p52 protein, which was strongly induced by ERK1/2 activation. Twelve peptides were observed on reverse phasehigh performance liquid chromatography following treatment of the p52 protein with protease under the conditions described under "Experimental Procedures." The amino acid sequences of the two peptide fragments with peaks indicated by arrows in Fig. 2D were determined, and the amino acid sequences of peaks A and B were found to be (K)TIATALVSK and (K)YN-QILRIEEELGSK, respectively. These amino acid sequences are identical to deduced amino acid sequences of rat ␣-enolase. To confirm this result further, the molecular weights of the peptide fragments derived from the p52 protein were determined by liquid chromatography-mass spectrometry (LC-MS). The molecular weights of peaks separated by liquid chromatography were MHϩ: 810.8 (peak A), 1182.5 (peak A), 899.5 (peak B), 1007.5 (peak C), 1520.0 (peak D), 1636.0 (peak C), FIG. 4. Effect of transcription factors on induction of ␣-enolase during ischemic hypoxia and reoxygenation. H9c2 cells were pretreated with a pcDNA3 empty vector, a pSV-HA-PGC-1, pcDNA3-c-myc, or a pEBB-HIF-1␣ containing liposomes. Cells were exposed to ischemic hypoxia for 2 h and reoxygenation for 6 h, extracted, and subjected to immunoblotting with anti-␣-enolase antibody (A) and anti-ERK1/2 antibody (B). The figures show representative immunoblots obtained from three independent experiments. The levels of ␣-enolase and ERK1/2 proteins were determined from the immunoblots by densitometric analysis (mean Ϯ S.E., n ϭ 3; *, p Ͻ 0.05 versus empty vector; #, p Ͻ 0.05 versus (Ϫ)PD98059) (C). 1691.5 (peak E), 1464.0 (peak F), 1961.5 (peak G), 3043.0 (peak H), and 2449.0 (peak I). These molecular weights are similar to the theoretical molecular weights of peptide fragments derived by treating rat ␣-enolase with protease (Fig. 2E). Taking the data from the amino acid sequences and LC-MS analyses together, we identified the p52 protein as ␣-enolase.
Expression of ␣-Enolase by ERK1/2 during Ischemic Hypoxia and Reoxygenation-To confirm the expression of ␣-enolase by ERK1/2 during ischemic hypoxia and reoxygenation, we carried out immunoblotting using anti-␣-enolase antibody. The amount of ␣-enolase protein increased at 6-h reoxygenation after ischemic hypoxia for 2 h. In cells transfected with MEKDN, the amount of the ␣-enolase protein was reduced to control levels or below by ischemic hypoxia and reoxygenation in H9c2 cells (Fig. 3, A-C) and in cardiomyocytes (Fig. 3, D-E), consistent with the results of two-dimensional electrophoresis (Fig. 2, A-C), suggesting that the ␣-enolase protein is degraded during ischemic hypoxia and reoxygenation as well as synthesized. Treatment with MEKCA led to a weak increase in the amount of ␣-enolase protein induced by ischemic hypoxia and reoxygenation (data not shown). The amount of ␣-enolase protein might reach a maximum during ischemic hypoxia and reoxygenation, since ERK1/2 activity increased remarkably during ischemic hypoxia and reoxygenation. We observed the expression of the mRNA of ␣-enolase during ischemic hypoxia and reoxygenation using the RT-PCR method. The amount of the RT-PCR product for ␣-enolase increased linearly up to 27 cycles. The expression of ␣-enolase mRNA was significantly increased until 3 h of reoxygenation after ischemic hypoxia. The increase in expression was blocked in H9c2 cells transfected with MEKDN (Fig. 3, F-H). In cardiomyocytes, the increase of ␣-enolase mRNA was suppressed at 3 h of reoxygenation by transfection with MEKDN (Fig. 3, I and J). The results were consistent with those in the presence of PD98059 (data not shown). The expression of G3PDH mRNA remained nearly constant throughout ischemic hypoxia and reoxygenation (Fig.  3J). These findings demonstrate that ERK1/2 activation during ischemic hypoxia and reoxygenation induces the expression of ␣-enolase. To examine the pathway of the induction of ␣-enolase by ERK1/2 activation, we observed the amount of ␣-enolase in the cells transfected with various transcription factors at 12-h reoxygenation after 2-h ischemic hypoxia. Transfection with c-myc and hypoxia-inducible factor-1␣ (HIF-1␣) led to the induction of ␣-enolase by ischemic hypoxia and reoxygenation as compared with the cells transfected with the empty vector and PGC-1 (Fig. 4). The induction of ␣-enolase in the cells expressing c-myc was inhibited by PD98059, but PD98059 had no effect in the cells expressing HIF-1␣. These results suggest that the induction of ␣-enolase by ERK1/2 during ischemic hypoxia and reoxygenation may be mediated by c-Myc.
Expression of ␣-Enolase during Ischemia and Reperfusion in Rat Heart-We examined the amount of ␣-enolase protein at 6-h reperfusion after 20-min ischemia in rat heart. The amount of ␣-enolase protein in the ischemic area increased slightly compared with that in the nonischemic area, although no significant difference was detected. The amount of ERK1/2 protein, used as an internal control, remained almost constant (Fig. 5, A-C). Next, we observed the expression of ␣-enolase mRNA by RT-PCR in ischemic heart and found it to increase significantly in ischemic area at 6-h reperfusion, whereas the level of G3PDH mRNA, used as an internal control, remained almost unchanged (Fig. 5, D-F). These results indicate that ␣-enolase expression was induced during ischemia and reperfusion in rat heart.
Physiological Role of ␣-Enolase during Ischemic Hypoxia and Reoxygenation-␣-Enolase is a glycolytic enzyme that catalyzes the formation of phosphoenolpyruvate from 2-phosphoglycerate, a high energy intermediate that produces ATP upon glycolysis. In the working heart, at least in part, the control of ATP generation is regulated by ␣-enolase and pyruvate kinase FIG. 5. Expression of ␣-enolase during ischemia and reperfusion in rat heart. The samples were prepared from the nonischemic and ischemic areas of the heart subjected to release for 6 h after left anterior descending coronary artery ligation for 20 min (20-min ischemia and 6-h reperfusion), extracted, and subjected to immunoblotting with anti-␣-enolase antibody (A) and anti-ERK1/2 antibody (B). The figures show representative immunoblots obtained from four independent experiments. The levels of ␣-enolase and ERK1/2 proteins were determined from the immunoblots by densitometric analysis (mean Ϯ S.E., n ϭ 4) (C). Total RNAs were isolated from nonischemic area and ischemic area of hearts, and RT was performed with total RNA. RT products were amplified by PCR using primers for ␣-enolase (D) or by PCR using primers for G3PDH (E). The figures show representative data obtained from four independent experiments. The levels of ␣-enolase and G3PDH mRNA were determined from the gels by densitometric analysis (mean Ϯ S.E., n ϭ 4; *, p Ͻ 0.05 versus nonischemic area) (F).
in the glucose group (42). Therefore, we measured intracellular ATP levels during ischemic hypoxia and reoxygenation using the luciferin-luciferase assay. The intracellular ATP levels in H9c2 cells treated with PD98059 were lower at each time point until 6-h reoxygenation after 2-h ischemic hypoxia as compared with control cells (Fig. 6A), while no differences in cell viability were observed (Fig. 6B). Similar results were also observed in the cardiomyocytes (Fig. 6, C and D). Treatment with PD98059 (50 M) alone for 24 h had no effect on the cell death in cardiomyocytes (data not shown). In MEKDN-transfected H9c2 cells and cardiomyocytes, the observed ATP levels were almost the same as those obtained with PD98059 (Fig. 7, A and C). At 24-h reoxygenation after 2-h ischemic hypoxia, the number of MEKDN cells was significantly decreased in H9c2 cells compared with empty vector-transfected cells (Fig. 7B). In the MEKDN-transfected cardiomyocytes, the number of cells was decreased by ischemic hypoxia and reoxygenation. The significant difference between control and MEKDN cardiomyocytes was not observed, since the cells expressing MEKDN may be much lower than those in the H9c2 cells (Fig. 7D). The inhibition of the glycolytic reaction in mitochondria resulted in a significant decrease in intracellular ATP levels, and this decrease finally induced cell death (Fig. 7, E and F), suggesting that ATP levels are closely related to cardiac cell death after ischemic hypoxia. MEKCA cells experienced a greater increase in ATP levels than cells transfected with empty vector, in contrast to the data for MEKDN (KA97) cells (Fig. 8) and MEKDN (SA218/SA222) cells (data not shown). To confirm the role of ␣-enolase in ATP generation and cell survival, we introduced the ␣-enolase protein into cells transfected with MEKDN using the Chariot protein. The efficiency of introduction of IgG-labeled Cy2 by the Chariot was more than 80% of the cells observed by microscopy (data not shown). The ␣-enolase protein was induced by ischemic hypoxia and reoxygenation, and the amount of protein was significantly reduced to the control levels or below by transfection of MEKDN as in the case of the data shown in Fig. 2, A-C (Fig. 9A). The introduction of ␣-enolase using Chariot into H9c2 cells led to the recovery of the ␣-enolase protein in cells transfected with MEKDN at 6-h reoxygenation (Fig. 9A), although ERK1/2 protein levels were almost the same in each lane (Fig. 9B). The similar results were observed in cardiomyocytes (data not shown). Furthermore, we observed the localization of ␣-enolase by immunocytochemistry using anti-␣-enolase antibody. The staining of ␣-enolase was mainly detected in the cytoplasm in control cells (Fig. 9, C and D). After the introduction of exogenous ␣-enolase, the staining was also observed in the cytoplasm (Fig. 9, E and  F). The data suggest that exogenous ␣-enolase localized in the FIG. 6. ATP levels and cell death during ischemic hypoxia and reoxygenation in cells treated with PD98059. A and B, H9c2 cells were incubated for 30 min in the absence or presence of PD98059. After the cells were exposed to ischemic hypoxia and reoxygenation, ATP levels were determined by the luciferin-luciferase method at the indicated times (mean Ϯ S.E., n ϭ 8; *, p Ͻ 0.05 versus control) (A), in addition, cell viability was determined at the indicated times (mean Ϯ S.E., n ϭ 8) (B). C and D, cardiomyocytes were incubated for 30 min in the absence or presence of PD98059. After the cells were exposed to ischemic hypoxia for 2 h and reoxygenation for 6 h, ATP levels were determined by the luciferinluciferase method (mean Ϯ S.E., n ϭ 8; *, p Ͻ 0.05 versus control) (C), in addition, cell viability was determined at 6-h reoxygenation after 2-h ischemic hypoxia (mean Ϯ S.E., n ϭ 8) (D). cytoplasm as well as endogenous ␣-enolase. The introduction of ␣-enolase caused ATP levels to rise in cells transfected with the empty vector and prevented most of the cell death observed during ischemic hypoxia and reoxygenation (Fig. 10, A-C). ATP levels were significantly reduced during post-ischemic reoxygenation in MEKDN cells but were almost fully restored to the control level by the introduction of ␣-enolase (Fig. 10A). Consistent with the changes in ATP levels, the survival of cells exposed to ischemic hypoxia was dramatically improved with increasing levels of the ␣-enolase protein in H9c2 cells (Fig. 10, B-E) and cardiomyocytes (Fig. 10, F-I). IgG-labeled Cy2, a control protein, had no effect in ATP levels (data not shown). It is known that oxidative stress that occurs during ischemia injures mitochondria (43,44) and that this leads to a decrease in ATP production by the glycolysis system. The key enzyme in glycolysis, ␣-enolase, might be induced to compensate for a drop in ATP levels.
To determine whether the observed decrease in ATP levels affects cardiac contractility, we examined the mechanical properties of isolated cardiomyocytes. In cardiomyocytes exposed to ischemic hypoxia, the peak height showing cell shortening (Fig.  11, A and B) decreased significantly compared with that in untreated cells. The decrease in ischemic cells was further enhanced by treatment with PD98059. The contraction impaired by MEK inhibitor during ischemic hypoxia was dramatically restored by the introduction of ␣-enolase into cells (Fig.  11, C and D). The rate of cell shortening (ϪdL/dt) and the rate of cell relengthening (dL/dt), as well as the peak height, also reached control levels or above following treatment with ␣-enolase (Fig. 11, E and F). There were no significant differences in times to peak or the times of 50% regression from the peak among cells (Fig. 11, G and H). These findings suggest that ␣-enolase can restore contractility of cardiomyocytes impaired by ischemic hypoxia. DISCUSSION We have used functional proteomics, a method for monitoring downstream factors following ERK1/2 activation (31), to identify ␣-enolase as a downstream factor of ERK1/2 in response to ischemic hypoxia. There are two possible mechanisms for the expression of ␣-enolase by ERK1/2. One possibility is that a c-Myc oncogenic transcription factor may be involved in the ERK1/2-induced expression of ␣-enolase. As indicated by our study, ␣-enolase was induced by ischemic hypoxia and reoxygenation in the cells transfected with c-Myc cDNA, and the induction was blocked by PD98059. These find- ings suggest that ERK1/2 may induce ␣-enolase through c-Myc during ischemic hypoxia and reoxygenation. In previous studies, ischemia in the myocardium and the metabolic inhibition of cardiomyocytes induce c-myc mRNA expression (45,46). The c-myc promoter has binding sites for Ets family transcription factors, which can be activated by the ERK1/2 pathway (47-49). Using a dominant negative mutant of ERK1/2 and the inhibitor, it was shown that the MEK/ERK1/2 pathway is necessary for c-myc expression in response to growth factors (50). The promoter of the ␣-enolase gene has two perfect Myc-Max binding motifs (CACGTG) (51), and c-Myc directly transactivates genes encoding ␣-enolase, which leads to an increase in glucose uptake in Rat1 fibroblasts (52). The other possibility is that HIF-1␣ mediates the expression of ␣-enolase in response to ischemia and reperfusion of heart. HIF-1␣ is a transcription factor that regulates the expression of genes involved in the cellular utilization of oxygen and glucose in ischemia and in tumors (53,54). The genes encoding the glycolytic enzymes, aldose A, lactate dehydrogenase A, and ␣-enolase, contain hypoxia response elements that can bind to HIF-1␣. Indeed, in cells exposed to hypoxia, HIF-1␣ is involved in the transcriptional activation of glycolytic enzymes including ␣-enolase (55). Recently, it was reported that the ERK1/2 pathway directly phosphorylates HIF-1␣ and up-regulates its activity through p300/CBP (56,57). In our experiments, significant induction of ␣-enolase by postischemic reoxygenation was observed in cells transfected with HIF-1␣ cDNA (data not shown), however, the induction was unaffected by PD98059. Other factor(s) than ERK1/2 may be involved in the induction of ␣-enolase by HIF-1␣ during ischemia and reperfusion.
We also show that ␣-enolase expression participates in the maintenance of intracellular ATP levels in cardiomyocytes exposed to ischemic hypoxia and that this is closely related to cell survival. It is well known that ␣-enolase is the glycolytic enzyme that catalyzes the production of phosphoenolpyruvate from 2-phosphoglycerate, the second of the two high energy intermediates in the process of ATP production (58). In working perfused heart, the control of glucose utilization in the pathway from 3-phosphoglycerate to pyruvate is shared by ␣-enolase and pyruvate kinase, although major control is divided between the glucose transporter and hexokinase (42). It is rea- . Transfected cells were exposed to 6-h reoxygenation following 2-h ischemic hypoxia, after which ATP levels were determined by the luciferin-luciferase method (mean Ϯ S.E., n ϭ 8; *, p Ͻ 0.05 versus empty vector).
FIG. 9. Expression and introduction of the ␣-enolase protein in H9c2 cells. The ␣-enolase protein was introduced using the Chariot protein alone as a negative control (A-D) or a Chariot-␣-enolase complex (A, B, E, and F). The expression of the ␣-enolase protein in cells at 6-h reoxygenation after 2-h ischemic hypoxia reoxygenation was detected by immunoblotting using anti-␣-enolase (A) and anti-ERK1/2 antibodies (B) and observed through by bright field (C and E) or by immunocytochemistry using anti-␣-enolase antibody (D and F). IR, ischemic hypoxia and reoxygenation. Final magnification, ϫ400. sonable that the decrease in ␣-enolase expression leads to the suppression of ATP production in heart ischemia as shown in the report. Partial ATP depletion induces Fas-and caspasemediated apoptosis in Madin-Darby canine kidney cells (59,60), while complete ATP deletion blocks Fas-induced apoptosis (61) and leads to necrosis mediated by poly(ADP-ribose) polymerase in fibroblast cells (62). It is most important, not only for cell survival but also for the function of the heart, to maintain ATP levels in cardiomyocytes, because the heart requires ATP to contract through the myosin-actin complex and for the activation of several channels, such as Na ϩ -K ϩ ATPase, to create a membrane potential (63). The production of ATP is also required for ERK1/2 activation during ischemic hypoxia and reoxygenation in cardiomyocytes, suggesting that ERK1/2 activation initially leads to a recovery of ATP levels, which may support ERK1/2 activation in the late phase (64). FIG. 11. Effect of ␣-enolase on the contractility of isolated cardiomyocytes exposed to ischemic hypoxia and reoxygenation. Cardiomyocytes were isolated from 12 newborn rats and cultured. ␣-Enolase was introduced into the cardiomyocytes, which were then cultured in L15 medium including 0.5% fetal calf serum for 2 days. The cells were exposed to 2-h ischemic hypoxia and 12-h reoxygenation in the presence or absence of PD98059, and contractility was determined using a video-based edge-detection system (A). The contractility parameters were determined by the soft-edge software (B), and a representative record obtained from three independent experiments is shown (C). The parameters were determined from 10 records in each experiment carried out three times (mean Ϯ S.E., n ϭ 3; *, p Ͻ 0.05 versus control; #, p Ͻ 0.05 versus (Ϫ)␣-enolase) (D-H). IR, ischemic hypoxia and reoxygenation.
In our studies, the introduction of ␣-enolase appeared to lead to increases in cell survival and cardiomyocyte contractility greater than expected from the maintenance of ATP levels alone. ␣-Enolase may also have other functions besides the maintenance of ATP levels in ischemic heart. Recently, ␣-enolase was reported to act as a multifunctional protein, such as heat shock protein (65), the transcriptional repressor of the c-myc protooncogene (58), and the binding protein of the cytoskelton, chromatin structure, and plasminogen (66). These findings suggest that ␣-enolase is essential for protecting heart from ischemic injury. As suggested by our data, ␣-enolase is degraded during ischemic hypoxia and reoxygenation, and the transcription of ␣-enolase should be activated for survival. ERK1/2 plays a role in the transcriptional regulation, which may have implications in several diseases related to ERK1/2 activity (67). In preliminary data, some patients in heart failure show reduced amounts of ERK1/2 proteins. The expression of ␣-enolase by ERK1/2 in response to ischemia may be relevant to a local contractile defect observed in patients with chronic myocardial ischemia. Our investigation suggests that ERK1/2 has a novel function, ATP generation, in diseases related to ischemia.