Novel role of C terminus of Hsc70-interacting protein (CHIP) ubiquitin ligase on inhibiting cardiac apoptosis and dysfunction via regulating ERK5-mediated degradation of inducible cAMP early repressor

Abstract

Growing evidence indicates a critical role of ubiquitin-proteosome system in apoptosis regulation. A cardioprotective effect of ubiquitin (Ub) ligase of the C terminus of Hsc70-interacting protein (CHIP) on myocytes has been reported. In the current study, we found that the cardioprotective effect of insulin growth factor-1 (IGF-1) was mediated by ERK5-CHIP signal module via inducible cAMP early repressor (ICER) destabilization. In vitro runoff assay and Ub assay showed ICER as a substrate of CHIP Ub ligase. Both disruption of ERK5-CHIP binding with inhibitory helical linker domain fragment (aa 101–200) of CHIP and the depletion of ERK5 by siRNA inhibited CHIP Ub ligase activity, which suggests an obligatory role of ERK5 on CHIP activation. Depletion of CHIP, using siRNA, inhibited IGF-1-mediated reduction of isoproterenol-mediated ICER induction and apoptosis. In diabetic mice subjected to myocardial infarction, the CHIP Ub ligase activity was decreased, with an increase in ICER expression. These changes were attenuated significantly in a cardiac-specific constitutively active form of MEK5α transgenic mice (CA-MEK5α-Tg) previously shown to have greater functional recovery. Furthermore, pressure overload-mediated ICER induction was enhanced in heterozygous CHIP^+/− mice. We identified ICER as a novel CHIP substrate and that the ERK5-CHIP complex plays an obligatory role in inhibition of ICER expression, cardiomyocyte apoptosis, and cardiac dysfunction.—Woo, C.-H., Le, N.-T., Shishido, T., Chang, E., Lee, H., Heo, K.-S., Mickelsen, D. M., Lu, Y., McClain, C., Spangenberg, T., Yan, C., Molina, C. A., Yang, J., Patterson, C., Abe, J.-I. Novel role of C terminus of Hsc70-interacting protein (CHIP) ubiquitin ligase on inhibiting cardiac apoptosis and dysfunction via regulating ERK5-mediated degradation of inducible cAMP early repressor.

Epidemiological studies strongly indicate that cardiomyocyte apoptosis is a key event in the development and progression of heart failure. It has been established that myocyte loss through apoptosis contributes to the transition from cardiac hypertrophy to heart failure. Inducible cAMP early repressor (ICER) has been shown to be an important inducer or mediator of apoptosis in cardiomyocytes (1). Previously, our group demonstrated that down-regulation of phosphodiesterase 3A (PDE3A) was associated with apoptosis and induction of ICER, which provides a mechanistic framework for how isoproterenol (ISO) and angiotensin II regulate myocyte apoptosis (2, 3). Sustained elevation of ICER favored apoptosis through inhibition of cAMP response element binding protein (CREB)-mediated transcription and down-regulation of Bcl-2, which is an important Bcl-2 family antiapoptotic protein in cardiomyocytes. This feedback regulatory mechanism between PDE3A and ICER was critical for sustained ICER induction and subsequent myocyte apoptosis. Of note, we found that ICER expression was significantly increased in failing hearts from ischemic heart disease and dilated cardiomyopathy (11), which also suggests the importance of ICER on the development of heart failure in humans.

Extracellular stress-regulated kinase 5 (ERK5) is a member of the ERK subfamily, which belongs to the mitogen-activated protein kinase (MAPK) family. ERK5 has an N-terminal kinase domain homologous to ERK1/2 but with a unique C-terminal region, which has a transactivation domain (4). MAPK kinase 5 (MEK5) has been identified as a direct upstream kinase, which can phosphorylate the TEY motif of ERK5. Genetically mutated mice with the ERK5 gene deleted have impaired cardiac and vascular development, which suggests the prosurvival role of ERK5 in the cardiovascular system. Cardiac-specific overexpression of CA-MEK5α (constitutively active form of MEK5α) in transgenic mice (αMHC-CA-MEK5α-Tg) inhibited ICER induction and myocyte apoptosis induced by pressure overload and myocardial infarction (MI; refs. 5, 6); the acceleration of apoptosis and cardiac dysfunction after thoracic aorta constriction (TAC) in cardiac-specific ERK5-knockout mice has been reported (7) as well. We have shown that the prosurvival insulin-like growth factor (IGF-1) inhibited myocyte apoptosis through ERK5 activation (5) and that a reduction of ERK5 with ERK5 siRNA or an adenovirus expressing the dominant negative form of ERK5 abolished the inhibitory effect of IGF-1 on ISO-mediated ICER induction. Clearly ERK5 regulates ICER induction and subsequent myocyte apoptosis (5), but the molecular mechanism linking ERK5 activation and ICER reduction remains largely unknown.

Ubiquitination is a post-translational modification that plays a critical role in protein quality control through proteosomal degradation. It has been reported that the chaperone-dependent E3 ubiquitin (Ub) ligase CHIP [carboxyl terminus of Hsp70-interacting protein; also known as STUB1 (STIP1 homology and U-Box containing protein 1)], has a strong cardioprotective effect, demonstrated by inhibition of apoptosis following ischemia/reperfusion injury (8). Compared with nontransgenic littermate control (NLC) mice, the CHIP-knockout mice showed increased infarct size and myocyte apoptosis after cardiac ischemia. A strong correlation appears between E3 Ub ligases and the specificity for their substrates in regulating ubiquitination-dependent protein degradation.

Since both ERK5 and CHIP have strong cardioprotective effects, we hypothesized that the CHIP Ub ligase played a role in the ERK5-dependent regulation of apoptosis via ICER ubiquitination and subsequent protein degradation. Here we report a novel signaling module of ERK5 kinase and CHIP E3 Ub ligase, which shows a prominent role in cardiomyocyte apoptosis and cardiac dysfunction in vivo.

MATERIALS AND METHODS

Reagents and antibodies

Isoproterenol, IGF-1, forskolin, MG-132, cycloheximide, and streptozotocin were purchased from Sigma (St. Louis, MO, USA). Antibodies against phospho-ERK5 and ERK5 were purchased from Cell Signaling Technology (Danvers, MA, USA). Anti-Flag, anti-actinin and anti-tubulin antibodies were purchased from Sigma. Anti-HA, anti-Myc, anti-ubiquitin, and anti-CHIP antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Anti-HisG antibody was purchased from Invitrogen (Carlsbad, CA, USA).

Cell culture

Primary cultures of neonatal rat cardiomyocytes were performed as described previously (9). Briefly, neonatal myocytes were obtained by enzymatic dissociation of cardiac ventricles from 2- to 3-d-old Sprague-Dawley rat neonates. The ventricular tissue fragments were subjected to multiple rounds of enzymatic digestion with collagenase II (Worthington Biomedical Corp., Lakewood, NJ, USA). Cells were then collected by centrifugation at 800 rpm for 3 min at 4°C. Nonmyocytes were removed via 2 rounds of preplating on culture dishes. The enriched cardiomyocytes were cultured in low-glucose DMEM with 10% fetal bovine serum, 50 U/ml penicillin, 50 μg/ml streptomycin, and 10 μM cytosine 1-β-d-arabinofuranoside (Sigma), which was added to inhibit the growth of contaminating nonmyocytes. More than 90% of cells were cardiomyocytes (positive for α-actinin). CHO cells were maintained in F-12 medium (Life Technologies, Inc., Carlsbad, CA, USA) supplemented with 10% fetal bovine serum and antibiotics.

Plasmid and adenovirus vector construction

Mouse ERK5 and the constitutively active form of MEK5α (CA-MEK5α) were cloned as described previously (10). The plasmids encoding human CHIP and CHIP-H260Q were created as described in previous reports (11, 12). Flag-tagged ICER was created by PCR amplification into pCMV-Tag2B vector with BamHI and EcoRI. Gal4-CHIP and VP16-ERK5 were created by inserting the human CHIP and mouse ERK5 into BamH1 and NotI sites of the pBIND and pACT vectors, respectively. All constructs were verified by DNA sequencing. Adenovirus expressing ICER and CA-MEK5α were generated using ViraPower Adenoviral Expression System (Invitrogen).

Mammalian 2-hybrid analysis and transfection of cells

The protein–protein interaction between ERK5 and CHIP was determined by a mammalian 2-hybrid system (Promega, Madison, WI, USA) as described previously (13). Briefly, cells were transfected in Opti-MEM (Invitrogen) with lipofectamine mixture containing the pG5-luc vector, pACT-ERK5, and pBIND-CHIP in the presence or absence of pHA-CA-MEK5α. After 3 h, cells were washed, and fresh F12 or DMEM supplemented with 10% fetal bovine serum was added. Since pBIND also contains the Renilla luciferase gene, transfection efficiencies were normalized with the Renilla luciferase activity. Cells were collected 24 h after transfection, and then the luciferase activity was assayed with the dual luciferase kit (Promega) using a TD-20/20 Luminometer (Turner Designs, Sunnyvale, CA, USA). Transfections were performed in triplicate, and each experiment was repeated ≥3 times.

Immunoprecipitation (autoubiquitination assay) and Western blot analysis

Cells were collected in phosphate-buffered saline containing 10 mM N-ethylmaleimide (NEM), and cell extracts were prepared in modified radioimmunoprecipitation assay 1 (RIPA) buffer [50 mM Tris-HCl, pH 7.4; 150 mM NaCl; 1 mM EDTA; 1% Nonidet P-40; 0.1% sodium dodecyl sulfate (SDS); 1 mM dithiothreitol; 1:200-diluted protease inhibitor cocktail (Sigma); 1 mM PMSF; 10 mM NEM; and 0.1 mM iodoacetamide]. Mouse hearts were washed with 10 ml of cold PBS. Isolated mice hearts were frozen in liquid nitrogen and homogenized with 0.5 ml of same modified RIPA buffer as described above. Immunoprecipitation with a mouse monoclonal anti-Flag, HA, Xpress antibody, or polyclonal ERK5 antibody was performed as described previously (4). Bound proteins were released in 2× SDS sample buffer, resolved by SDS-polyacrylamide gel electrophoresis, transferred onto a Hybond enhanced chemiluminescence nitrocellulose membrane (GE Healthcare, Piscataway, NJ, USA), and visualized by using the enhanced chemiluminescence detection reagents (PerkinElmer, Shelton, CT, USA) according to the manufacturer's instructions. Immunoblotting with ICER antibody and PDE3A were performed as described previously (3).

In vitro ubiquitination assay with GST-ICER

Heart tissue and cell extracts were prepared in modified RIPA butter containing 10 mM NEM. Ubiquitination activity was determined by in vitro ubiquitination assay with GST-fused ICER protein as a substrate (Ubiquitin-Protein Conjugation kit, BostonBiochem, Cambridge, MA, USA). Briefly, the lysates were applied for immunoprecipitation with anti-CHIP antibody. Immunoprecipitated CHIP was incubated with recombinant proteins, including ubiquitin and E1/E2 enzyme mixture in energy buffer for 60 min at 37°C. Reaction was stopped by adding SDS loading buffer and then followed by Western blot analysis with anti-ubiquitin antibody.

Transfection of the CHIP siRNA

The chip siRNA was purchased from Dharmacon (Lafayette, CO, USA). The rat and mouse specific chip target sequence was 5′-GGGAUGAUAUUCCUAGUGC-3′. A nonspecific control siRNA from Invitrogen was used as a negative control. Cardiomyocytes were transiently transfected with 40 nM of medium GC control RNA or chip siRNA using Lipofectamine 2000 and plus reagent (Invitrogen) following protocols provided by the manufacturer. The cells were harvested 36 to 48 h after siRNA transfection, and protein expression were measured by immunoblotting with antibodies against CHIP (Santa Cruz Biotechnology).

Analysis of apoptosis

Cardiomyocyte apoptosis was measured by the terminal deoxyribonucleotide transferase (TdT)-mediated dUTP nick-end labeling (TUNEL) detecting in situ DNA fragmentation. TUNEL staining was performed using the In Situ Cell Death Detection Kit (Roche, Indianapolis, IN, USA) as described previously (14). For the counterstaining of myocytes, cells were also stained with anti-cardiomyocyte-specific sarcomeric α-actinin antibody. Briefly, cells were washed 2 times with cold PBS, fixed with 4% paraformaldehyde in PBS for 20 min, and then permeabilized with 0.2% Triton X-100 for 5 min. Cells were incubated with blocking buffer (5% goat serum in PBS) for 60 min to block nonspecific binding and incubated with anti-α-actinin antibody (1:500 dilution in 1% goat serum/PBS) for 2 h at room temperature or overnight at 4°C. The cells were washed 3 times with PBS and incubated with secondary antibodies labeled with Alexa Fluor 546 dye against mouse IgG₁ (Molecular Probes, Eugene, OR, USA; 1:2000 dilution) for 90 min at room temperature. After washing, cells were applied to TUNEL staining following the manufacturer's instructions. All of the images were collected using a confocal laser-scanning microscope (Fluoview; Olympus, Center Valley, PA, USA) with a LUMPlanFl60× lens or a fluorescence microscope (Axiophot; Carl Zeiss MicroImaging, Inc., Thornwood, NY, USA) equipped with an Acroplan Water ×60 W lens. Quantification of the overlay image was done using Photoshop 8.0 (Adobe Systems, San Jose, CA, USA). An average of 200 anti-α-actinin antibody-positive cells from random fields were analyzed. All measurements were performed in a masked fashion, and ≥3 independent experiments were performed.

Animals

Mouse constitutively active form of MEK5alpha (CA-MEK5α, S311D/T315D) cDNA was subcloned into a pBluescript-based Tg vector between the 5.5-kb murine-αMHC promoter and 250-bp SV-40 polyadenylation sequences. We maintained 3 different lines of CA-MEK5α-Tg in FVB strains, and all 3 lines showed a similar phenotype and MEK5α expression as reported (10). The generation of haploinsufficient mice of stub1 has previously been reported (12). The number of animals is reported in the figures (see Figs. 7 and 8). All animal procedures were performed with the approval from the University of Rochester Committee on Animal Resources.

Figure 7.

ERK5 activation increased CHIP Ub ligase activity and decreased ICER induction after coronary artery ligation in diabetic mice. Heart tissues were harvested from NLC and CA-MEK5α-Tg mice without (A) or with coronary artery ligation (B). Tissue extracts were subjected to an ubiquitination assay for 60 min at 37°C with GST-ICER as a substrate. Reaction was stopped with an SDS sample buffer, followed by immunoblotting with anti-ubiquitin antibody. Amounts of protein expression were determined by Western blot analysis with specific antibodies as indicated. Relative ubiquitination of GST-ICER was measured by densitometric quantification and summarized in bar graphs Values are means ± sd; n=3–4. *P < 0.05.

Figure 8.

Pressure overload-mediated ICER induction, cleaved caspase 3 expression, and cardiac dysfunction were accelerated in CHIP^−/+ mice. A) After TAC, echocardiographically determined cardiac function (see Table 1) presented as percentage fractional shortening (% FS) at the indicated times. B) Heart weight 6 wk after TAC was increased in CHIP^−/+ mice compared with NLC mice (NLC, n=7; CHIP^−/+, n=6). C) At 6 wk after TAC or sham operation, heart tissues were collected from wild-type NLC (n=6; open bars in D, E) and CHIP^−/+ mice (n=6; solid bars in D, E). Tissue extracts were immunoblotted with anti-ICER, anti-Bcl-2, anti-cleaved caspase 3, and anti-tubulin antibodies. D, E) Bar graphs summarize expression levels of ICER (D) and Bcl-2 (E). F) Expression level of pERK5, ERK5, pERK1/2, ERK1/2, pAKT, AKT, CHIP, and tubulin in the heart 6 wk after TAC in NLC and CHIP^−/+. Six weeks after TAC or sham operation, heart tissues were collected from NLC and CHIP^−/+ mice. Tissue extracts were applied for Western blot analysis. Expression levels of pERK5, ERK5, pERK1/2, ERK1/2, pAKT, AKT, CHIP, and tubulin were determined by immunoblotting with specific antibodies. *P < 0.05; **P < 0.01.

Mouse model of TAC

The chronic pressure overload mouse model was created by performing TAC as described previously (5). Briefly, 10–12 wk-old FVB male mice (≈ 20g) were anesthetized with 1.5–2.0% isofluorane and placed on a ventilator. TAC was created via a left thoracotomy by placing a ligature securely around the trans part of the aorta and a 27-gauge needle and then removing the needle. Animals in the sham group (as control) underwent a similar procedure without constriction. The survival rate of this surgery was ∼90%.

Coronary ligation surgery

Mice received a single dose of the analgesic buprenorphine (Buprene) at a dose of 0.05 mg/kg s.c. prior to surgery. If necessary, another dose was administered 6–12 h later. Also, beginning 1 d prior to surgery, mice received acetaminophen at a dose of 1.5 mg/ml of drinking water as analgesia. This allowed the mice to acclimate to the different taste in the water before surgery. Mice were initially anesthetized with 1.5–2.0% isoflurane in 100% O₂ in an animal induction box for the initial procedure. Maintenance anesthesia was 0.5 to 1.5% isoflurane.

MI was induced in the mouse by ligating the left anterior descending coronary artery. After intubation, a midline incision was made between the sternum and the left internal mammalian artery. The heart was exposed, and the left anterior descending coronary artery was ligated intramurally 2 mm from its origin with a 9-0 Ethilon suture. After confirming an absence of bleeding, the chest was closed in 2 layers: The ribs (inner layer) were closed with 5-0 coated vicryl sutures in an interrupted pattern, and the skin was closed using 6-0 nylon or silk sutures in a cuticular manner. The anesthesia was stopped and the mouse was allowed to recover for several minutes before the endotracheal tube is removed. A sham operation involved an identical procedure, except that the suture was passed through the myocardium without occlusion.

Echocardiographic analysis

Echocardiographic analysis with M mode was performed using the Vevo770 with the 707B scanhead echocardiography machine equipped with a 30 MHz frequency probe (VisualSonics, Toronto, ON, Canada). Echocardiography (M mode) was obtained in unanesthetized mice. LV function was measured in the short-axis view at the level of the papillary muscles. The percentage of fractional shortening (FS) was assessed by measurement of the end diastolic and end systolic diameter [(end diastolic diameter − end systolic diameter)/end diastolic diameter × 100%]. We collected and averaged the data from 3 beats from 1 trace, and 3 traces from each animal. The pooled data were analyzed for statistical significance.

Statistical analysis

Data are reported as means ± sd or means ± se as indicated. Statistical analysis was performed with the StatView 4.0 package (Abacus Concepts, Berkeley, CA, USA). Differences were analyzed with a 1-way or a 2-way repeated measure analysis of variance as appropriate, followed by Schéffe's correction for multiple comparisons. Values of P < 0.05 were considered significant; levels of significance are indicated in the figures.

RESULTS

ERK5 activation decreases ICER protein stability via ubiquitin-proteasome system (UPS)

Previously, we reported that ISO-mediated ICER induction inhibited PDE3A expression and formed a PDE3A-ICER feedback loop (3). ISO and forskolin (Fsk) induced PKA activation to stabilize the ICER protein (3). In contrast, IGF-1-mediated ERK5 activation played an important role in destabilizing the ICER protein through post-translational modification (5), which was opposite to the effect of PKA activation. The expression level of ICER is regulated by CREB-dependent ICER gene transcription as well as proteasome-dependent ICER protein ubiquitination and degradation (15). However, the involvement of ubiquitin-proteasome system on this ERK5-mediated ICER destabilization remains unclear.

To investigate the involvement of protein stability in ERK5-mediated ICER destruction, we used cycloheximide (CHX), which is an inhibitor of global protein synthesis. First, we transduced cardiomyocytes with Ad-LacZ or Ad-CA-MEK5α for 24 h and subsequently transduced with Ad-ICER for 12 h (this time allowed a sufficient expression of ICER, data not shown). After Ad-ICER transduction, cells were treated with CHX and Fsk as indicated in the schematic diagram (Fig. 1A). As shown in Fig. 1A, ICER expression was significantly decreased in the presence of Ad-CA-MEK5α, compared with Ad-LacZ control. Of note, since we would like to detect the role of CHIP on ICER stability, we performed Ad-ICER transduction. Expression of exogenous ICER by Ad-ICER transduction excludes potential confounds such as ERK5 activation interfering with PKA-mediated endogenous ICER mRNA induction. Since we pretreated the cells with CHX, the forskolin-induced increase in ICER expression is due solely to PKA-mediated increase of ICER protein stability but not due to the regulation of transcriptional machinery or RNA stability, as we described previously (3). Therefore, in this set of experiments, the decrease in ICER protein amount was likely due to destabilization of ICER by ERK5 activation.

Figure 1.

ERK5 activation inhibited ICER induction via the ubiquitin-proteosome system. A) Neonatal rat cardiomyocytes were cotransduced with Ad-ICER and Ad-LacZ or Ad-CA-MEK5α for 24 h and incubated with CHX for the indicated times (see Materials and Methods). Expression of ICER, CA-MEK5α, and tubulin was detected by Western blotting with the respective antibodies (middle panel) and the relative expression of ICER was summarized by a line graph from 3 independent experiments (right panel; n=3). B) Cardiomyocytes were cotransduced with Ad-ICER and Ad-CA-MEK5α or Ad-LacZ at 50 MOI for 24 h and additionally incubated with forskolin (Fsk; 10 μM) for 12 h. Cells were treated further with MG132 for 8 h and subjected to a Western blot analysis with anti-ICER, anti-HA, and anti-tubulin antibodies (left panel). C) Myocytes were transduced with Ad-LacZ or Ad-CA-MEK5 for 24 h and then exposed to 10 μM ISO for an additional 24 h. Ad-CA-MEK5α-transduced cells were treated with MG132 for 8 h, as indicated, and subjected to immunoblotting with the stated antibodies (left panel). Bar graph results are means ± sd from 3 independent experiments. *P < 0.05; ** P < 0.01.

Next, we examined the effect of the proteosomal inhibitor MG132 on ERK5-mediated ICER protein destabilization. MG132 treatment abolished the CA-MEK5α-induced decrease in ICER protein confirming the role of UPS in ERK5-mediated ICER reduction (Fig. 1B). To confirm this regulatory mechanism of endogenous ICER under physiological conditions further, we stimulated myocytes with ISO. As observed with Fsk-mediated ICER stabilization, MG132 treatment reversed ERK5-mediated inhibition of ICER induction by ISO (Fig. 1B). These data suggested that ERK5 destabilized the ICER protein through regulation of the UPS.

ICER is a novel substrate of CHIP Ub ligase

It is well established that molecular chaperones, including CHIP, have strong cardioprotective effects through the regulation of protein quality control (8, 16–18). To address whether CHIP as an Ub ligase promotes ubiquitination and degradation of ICER, we performed runoff reactions with CHX to stop general protein synthesis after transducing with Ad-CHIP or Ad-LacZ. Again, we performed Ad-ICER transduction for detecting the role of CHIP on ICER stability as described in Fig. 1. Ad-CHIP enhanced ICER destabilization compared with Ad-LacZ, indicating that CHIP degrades ICER (Fig. 2A). To determine the involvement of CHIP ligase activity in ICER degradation, cells were cotransfected with pFlag-ICER and pMyc-CHIP or pMyc-CHIP-H260Q, which is an Ub ligase-activity-dead mutant, with vehicle or MG132 treatments. As shown in Fig. 2B, wild-type CHIP significantly decreased ICER expression compared with CHIP-H260Q mutant, indicating that CHIP-mediated ICER degradation was dependent on Ub ligase activity. In addition, pretreatment with MG132 significantly inhibited CHIP-mediated down-regulation of ICER despite constant CHIP expression (Fig. 2B). Furthermore CHIP wild-type, but not CHIP-H260Q mutant, markedly increased ubiquitination of ICER (Fig. 2C). These data suggested that ICER was a substrate of CHIP leading to ubiquitination and subsequent proteosomal degradation of this proapoptotic protein.

Figure 2.

ICER is a novel substrate of CHIP Ub ligase. A) Neonatal rat cardiomyocytes were cotransduced with Ad-ICER and Ad-LacZ, or Ad-CHIP for 24 h and incubated with CHX for the indicated times. Top panel: expression of ICER and CHIP was detected by Western blotting with the respective antibodies. Botttom panel: line graph summary of the relative expression of ICER, determined by densitometry (n=3). B) Myocytes were transfected with pcDNA3, pMyc-CHIP, pMyc-CHIP-H260Q, and pFlag-ICER in the presence or absence of 10 μM MG132. Top panel: protein expression was determined by Western blotting. Bottom panel: bar graph summary from 3 independent experiments. C) Top panel: ICER ubiquitination was determined by immunoprecipitation with anti-Flag antibody followed by immunoblotting with anti-ubiquitin antibody. Bottom panel: bar graph summary from 3 independent experiments. Values are means ± sd. *P < 0.05; **P < 0.01.

Association of ERK5 with CHIP is critical for ERK5-mediated up-regulation of CHIP Ub ligase activity

Since we found that both ERK5 and CHIP were involved in ICER stability, first we examined whether ERK5 associated with CHIP. Cardiomyocytes were stimulated by IGF-1 for 14 h, or transduced with Ad-LacZ or Ad-CA-MEK5α. Coimmunoprecipitation study showed ERK5 and CHIP association increased by IGF-1 and Ad-CA-MEK5α (Fig. 3A). To confirm this ERK5-CHIP interaction further, we used a mammalian 2-hybrid system as described previously (13). The 2-hybrid assay showed that CA-MEK5α significantly increased the association between ERK5 and CHIP (Fig. 3B). These data demonstrated that CHIP is a new binding partner to ERK5 and that this protein–protein interaction is dependent on ERK5 kinase activation.

Figure 3.

ERK5-CHIP association was critical for the up-regulation of CHIP Ub ligase activity. A) Cardiomyocytes were transduced with Ad-CA-MEK5α or IGF-1 treatment to activate ERK5. Lysates were immunoprecipitated with anti-ERK5 antibody, followed by Western blotting with anti-CHIP antibody to determine ERK5-CHIP association. Ad-lacZ transduction served as a negative control. Expression of CHIP and CA-MEK5α was determined by immunoblotting with anti-CHIP and anti-HA antibodies, respectively. B) To confirm the binding specificity, a mammalian 2-hybrid assay was performed as described in Materials and Methods. CHO cells were cotransfected with pG5-Luc, pBIND-CHIP, and pACT-ERK5 or pACT in the presence or absence of pCA-MEK5α. Protein binding affinity was determined by luciferase activity using dual luciferase assay. C) CHO cells were cotransfected with reporter gene mixture and pBIND-CHIP fragments (CHIP-Fr1, Fr2, Fr3) in the presence or absence of pCA-MEK5α. Binding affinity was determined by luciferase activity. D) Cardiomyocytes were cotransfected with pFlag-ERK5, pMyc-CHIP, and pcDNA3 or pXp/His-CHIP fragments (CHIP-Fr1, Fr2, Fr3). After immunoprecipitation with anti-Flag antibody, ERK5-CHIP interaction was determined by immunoblotting with anti-Myc antibody. Expression of His-tagged CHIP fragments was determined by Western blotting with anti-HisG antibody (Invitrogen). E) Disruption of ERK5-CHIP association with CHIP-Fr2 fragment transfection inhibited CA-MEK5α-mediated CHIP ligase activation. Data are representative of 3 independent experiments. *P < 0.05; **P < 0.01.

As shown in Fig. 3C, CHIP contains 3 distinguishable domains, consisting of N-terminal tetratricopeptide repeat domain (TPR), helical linker domain, and C-terminal U-box domain (containing the Ub ligase activity) (19). To investigate the binding regions of CHIP responsible for its interaction with ERK5, we generated 3 truncated mutants and evaluated its association with ERK5 in a mammalian 2-hybrid assay. The plasmids expressing the GAL4-DBD and the CHIP (full-length or truncated forms) were constructed with the pBIND vector. The plasmid expressing VP16-ERK5 was constructed with the pACT vector. The mammalian 2-hybrid assay demonstrated that both helical linker domain fragment (CHIP-Fr2, aa 101–200) and U-box domain (CHIP-Fr3, aa 201–303) interacted with ERK5 (Fig. 3C). In a separate experiment, we examined whether the fragments could interfere with the ERK5-CHIP association. This fragment interference approach is highly desirable since no mutations possibly having an unintentional effect on the CHIP protein structure are introduced. We cotransfected cells with wild-type ERK5 and CHIP in the presence of CHIP fragments. The CHIP-Fr2 fragment significantly inhibited the binding of ERK5 with CHIP even with the expression level relatively low compared with CHIP-Fr1 and CHIP-Fr3 fragments (Fig. 3D). Disruption of ERK5-CHIP association with the CHIP-Fr2 fragment significantly inhibited CA-MEK5α-mediated CHIP ligase activation, as determined by CHIP autoubiquitination (Fig. 3E). Thus, ERK5-CHIP association was dependent on ERK5 kinase activation, and this protein: protein interaction regulates the Ub ligase activity.

IGF-1 increased CHIP ligase activity via an ERK5 dependent pathway

The ERK5-CHIP association was critical for increasing the CHIP Ub ligase activity. Therefore, we investigated the role of IGF-1-mediated ERK5 activation on CHIP Ub E3 ligase activity. CHIP Ub ligase activity leads to increased substrate ubiquitination as well as autoubiquitination (20, 21). We first measured the CHIP E3 ligase activity using a CHIP autoubiquitination assay. Cells were cotransfected with pMyc-CHIP and pHA-ubiquitin in the presence or absence of CA-MEK5α, and autoubiquitination was determined by immunoprecipitation with anti-Myc antibody, followed by immunoblotting with antiubiquitin antibody. As shown in Fig. 4A, CA-MEK5α significantly increased the autoubiquitination of CHIP. However, we could not detect CHIP autoubiquitination with the ligase-activity-dead mutant of CHIP (CHIP-H260Q), supporting that the broad bands seen around 75–250 kDa were autoubiquitinated CHIP.

Figure 4.

ERK5 activation increased CHIP Ub ligase activity. A) To check the role of ERK5 activation in CHIP-mediated autoubiquitination, cells were cotransfected with pHA-Ubiquitin and pMyc-CHIP or pMyc-CHIP-H260Q in the presence of CA-MEK5α as indicated. Left panel: lysates were immunoprecipitated with anti-Myc antibody, followed by Western blotting with anti-ubiquitin antibody. Right panel: relative ubiquitination activity was summarized by a bar graph. B) CHIP Ub ligase activity was determined by an in vitro ubiquitination assay kit (BostonBiochem) with GST-fused ICER protein as a substrate. Cardiomyocytes were transfected with siRNA against rat CHIP or control and then followed by IGF-1 treatment and transduction of Ad-CA-MEK5α for 24 h. Immunoprecipitated CHIP was incubated with recombinant proteins, including ubiquitin and E1/E2 enzyme mixture, for 60 min at 37°C. Left panel: CHIP ligase activity was determined by immunoblotting with anti-ubiquitin antibody. Right panel: bar graph summary of relative ubiquitination activity from 3 independent experiments. Amounts of protein expression were determined by Western blot analysis with specific antibodies. Data are representative of 3 independent experiments. *P < 0.05; **P < 0.01.

Next, to determine the role of endogenous ERK5 on CHIP Ub ligase activity, we utilized ERK5 siRNA to deplete ERK5 expression and detected CHIP Ub ligase activity using in vitro ubiquitination assay with GST-fused ICER protein as a substrate. As shown in Fig. 4B, both IGF-1 and transduction with Ad-CA-MEK5α significantly increased the CHIP Ub ligase activity, which was inhibited by siRNA depletion of ERK5. These results indicated that IGF-1 increased CHIP Ub ligase activity via endogenous ERK5 activation.

Critical role of CHIP on ICER degradation and apoptosis in cardiomyocytes

We found that ICER can be a novel substrate of CHIP Ub ligase, the critical role of ERK5-CHIP association, and IGF-1-mediated ERK5 activation on CHIP Ub ligase activation. Therefore, we investigated whether CHIP is critical for IGF-1 or ERK5-mediated ICER degradation. Cardiomyocytes were transfected with control siRNA or CHIP siRNA for 24 h, then transduced with Ad-CA-MEK5α or Ad-LacZ. As shown in Fig. 5A, CA-MEK5α-mediated inhibition of ICER expression was decreased by CHIP depletion. Since IGF-1 inhibited ICER induction through ERK5 activation, we also determined whether CHIP depletion could block the IGF-1-mediated reduction of ICER expression. IGF-1 inhibited ISO-mediated ICER induction, and this reduction was also reduced by CHIP depletion (Fig. 5B). These results suggested that ICER was a physiological substrate of CHIP and that CHIP played an obligatory role in ERK5 activation-mediated ICER degradation in cardiomyocytes.

Figure 5.

CHIP was responsible for ERK5 activation-mediated inhibition of ICER degradation. Cardiomyocytes were transfected with control or rat CHIP specific si-RNA for 24 h, followed by transduction with Ad-CA-MEK5α or Ad-LacZ at 50 MOI for additional 24 h (A) or 20 ng/ml IGF-1 treatment for 14 h (B). Left panels: cells were further treated with ISO for 24 h, harvested, and subjected to immunoblotting with anti-ICER, anti-CHIP, anti-HA (HA-CA-MEK5α), or anti-tubulin antibodies. Right panels: bar graph summary of relative ICER expression from 3 independent experiments. *P < 0.05; **P < 0.01.

Next, we examined whether CHIP mediates the antiapoptotic effect of ERK5. Apoptosis was determined by TUNEL detection of in situ DNA fragmentation in myocytes after ISO treatment. ISO-induced apoptosis was significantly inhibited by Ad-CA-MEK5α and IGF-1 treatments (Fig. 6A, B), and siRNA depletion of CHIP abrogated this cytoprotective effect. Taken together CHIP appeared obligatory in the ERK5-mediated ICER degradation and subsequent inhibition of cardiomyocyte apoptosis.

Figure 6.

CHIP was critical for ERK5 activation-mediated inhibition of myocyte apoptosis. A, B) Cardiomyocytes were transfected with control or rat CHIP specific si-RNA for 24 h, followed by transduction with Ad-CA-MEK5α or Ad-LacZ at 50 MOI (A) or IGF-1 treatment (B) for an additional 24 h. After further treatment with ISO for 48 h, cells were assayed by TUNEL staining using the In Situ Cell Death Detection Kit (Roche). Myocytes were counter stained with anti-cardiomyocyte-specific sarcomeric α-actinin antibody. Bar graphs present percentage of TUNEL positive cells from total myocytes counted. **P < 0.01. C) Representative photomicrographs of the above experiment showing TUNEL (green), actin (red) and DAPI (blue) signals and their merged images.

ERK5 activation up-regulates CHIP Ub ligase activity and inhibits ICER expression after MI in diabetic mice (DM)

Cardiac-specific expression of CA-MEK5α inhibited ICER up-regulation and inhibited myocyte apoptosis as well as cardiac dysfunction after MI in streptozotocin-treated DM in vivo (DM+MI; refs. 5, 6). We utilized this model and determined the CHIP Ub ligase activity in the hearts. First, the pre-MI CHIP Ub ligase activity was increased in αMHC-CA-MEK5α-Tg samples compared to NLC (Fig. 7A). Second, the CHIP Ub activity decreased in NLC group after DM + MI, but this decrease was not observed in the αMHC-CA-MEK5α-Tg group. Finally, the up-regulation of ICER after DM + MI was markedly decreased in αMHC-CA-MEK5α-Tg compared with NLC (Fig. 7B). The ERK5 regulation of CHIP Ub ligase activity and subsequent ICER expression appeared operational in vivo. These findings provide a mechanistic framework to our previously reported observation that DM + MI-mediated apoptosis was decreased in αMHC-CA-MEK5α-Tg (6).

Depletion of CHIP accelerated pressure overload-mediated ICER induction in the heart

To confirm whether ICER is a functional substrate of CHIP Ub ligase in vivo, we utilized CHIP^−/+mice and implemented the pressure overload-derived heart failure model with TAC. No difference was detected in baseline cardiac function between CHIP^−/− and their littermates, with no obvious anatomical abnormalities in the null-mutant hearts, whereas CHIP depletion accelerated ischemia/reperfusion injury of the heart with severe cardiomyocyte apoptosis (8). Here we used haploinsufficient stub1 mice because CHIP^−/− mice exhibited partial perinatal lethality and decreased longevity (12, 22). CHIP^−/+ mice developed cardiac dysfunction (Fig. 8A, B and Table 1) after 6 wk of TAC accompanied by enhanced ICER induction (Fig. 8D) and inhibition of Bcl-2 expression as described previously (6) (Fig. 8E). We did not find any difference in ERK5, ERK1/2, and Akt activation between CHIP^−/+and CHIP^+/+ mice (Fig. 8F). These data support our findings in vitro and indicate that ICER is also a CHIP substrate in vivo.

Table 1.

Physiological parameters and echocardiogram measurements after 6 wk of surgery

Parameter	Group
	NLC, n = 7	CHIP−/+, n = 6
BW (g)	28.2 ± 3.3	27.5 ± 2.6
HW (mg)	142 ± 9	158 ± 16*
LVW (mg)	110 ± 6	120 ± 15
LW (mg)	155 ± 7	159 ± 10
HW/BW (mg/g)	5.05 ± 0.36	5.82 ± 0.97*
LVW/BW (mg/g)	3.94 ± 0.36	4.43 ± 0.89
LW/BW (mg/g)	5.57 ± 0.88	5.86 ± 0.78
HR (bpm)	684 ± 43	683 ± 46
AWd (mm)	1.28 ± 0.13	1.37 ± 0.13
PWd (mm)	1.19 ± 0.1	1.11 ± 0.21
LVEDd (mm)	2.92 ± 0.36	3.12 ± 0.32
LVESd (mm)	1.46 ± 0.27	1.81 ± 0.26*
FS (%)	50.4 ± 4.5	42.3 ± 4.4**
EF (%)	84.4 ± 4.5	75.8 ± 5.2**

BW, body weight; HW, heart rate; LVW, left ventricle (LV) weight; LW, lung weight; HR, heart rate; AWd, LV anterior wall end-diastolic dimension; PWd, LV posterior wall end-diastolic dimension; LVEDd, LV end-diastolic dimension; LVESd, LV end-systolic dimension; FS, fractional shortening; EF, ejection fraction. Values are means ± sd.

^*P < 0.05,

^**P < 0.01 vs. NLC.

DISCUSSION

ICER has been recognized as an important regulator of cardiomyocyte apoptosis through negative regulation of Bcl-2 expression (1–3). Previously, we found that ERK5 was cardioprotective in a variety of cardiac injury models via the down-regulation of ICER (5). However, the molecular mechanism by which ERK5 reduced ICER expression remained to be addressed. The main finding of the current study was that ERK5 activation regulated CHIP-mediated ICER ubiquitination and subsequent protein degradation leading to the inhibition of myocyte apoptosis and cardiac dysfunction. IGF-1-mediated ERK5 activation increased CHIP Ub ligase activity through protein–protein interaction, which inhibited ISO-mediated ICER induction and subsequent apoptosis. Finally, our results using αMHC-CA-MEK5α-Tg and haploinsufficient mice of stub1 (CHIP^−/+) demonstrated that ERK5-CHIP cascade was required for inhibition of myocyte apoptosis and cardiac dysfunction via ICER degradation. These observations revealed ERK5-CHIP complex as a novel antiapoptotic signal module in the heart.

cAMP-mediated ICER induction is regulated not only by transcriptional activity but also by a post-translational modification via UPS in cancer cell lines (1, 15, 23). Yehia et al. (15) reported that 8-Br-cAMP attenuates ICER ubiquitination and degradation, probably via a cAMP-dependent inhibition of ERK1/2 that phosphorylates and targets ICER to ubiquitin-mediated destruction. In their biochemical study, PD98059, an ERK1/2 inhibitor, prevented ICER destruction and identified the ERK1/2-mediated serine phosphorylation site of ICER using 2-D gel and point-mutation analysis. The phosphorylation mutant of ICER (ICER-S41A) had a half-life 4–5 times longer than wild-type counterpart, suggesting that phosphorylation is a key event in proteasome-dependent ICER degradation. However, how the Ub ligase was involved in the regulation of ICER expression after cardiac injury remained unknown. We found that CHIP degraded both wild-type ICER and phosphorylation mutant ICER (ICER-S41A) without any significant difference, suggesting that CHIP-mediated ICER ubiquitination was not regulated by ERK1/2-dependent ICER phosphorylation (data not shown). It remains unclear exactly how ERK5 specifically regulates CHIP-mediated ICER ubiquitination. Given the fact that ERK5 associates with CHIP via CHIP-Fr2 and CHIP-Fr3 fragments, it will be interesting to determine whether these regions of CHIP have a functional role in ICER ubiquitination. Interestingly, CHIP-Fr3 fragment contains a U-box domain, which is important for ubiquitination activity suggesting that ERK5 binding to CHIP may affect CHIP enzyme activity through a conformational change.

It has been recognized that chaperone-dependent protein quality control is important for protection against cardiac injury (16–18). In particular, molecular chaperone machinery removes damaged, unfolded, and aggregated proteins synthesized during ischemic injury and prevents cardiac dysfunction (24–26). Ballinger et al. (27) and Connell et al. (28) identified an Hsp70- and Hsp90-binding protein, also called CHIP, that has a dual function as a cochaperone and a Ub ligase. They found that their CHIP was highly expressed in the heart and regulated chaperone activity and protein quality control through ubiquitination-dependent proteosomal degradation. Compared with NLC, CHIP^−/− showed increased infarct size and myocyte apoptosis after cardiac ischemia (8), suggesting that chaperone function might be involved in the cardioprotective effects of CHIP. Recently, it has been reported that Hsp90 binds to activated ERK5 and increases ERK5 transcriptional activity (29). Since Hsp70 and Hsp90 are important regulators of CHIP chaperone function (30, 31), it will be intriguing to investigate whether ERK5 activation affects the chaperone function of CHIP under pathological conditions of the heart. Further investigation is necessary to clarify this issue.

While we focus on ERK5 activation, we do not believe this to be the only cardioprotective mechanism of IGF-1. As previously reported (32–34) Akt and ERK1/2 activations are likely mechanisms of IGF-1-induced cardioprotection. We propose that the ERK5-CHIP module is an additional critical cardioprotective mechanism mediating the IGF-1 action. This mechanism appears to be independent from Akt and ERK1/2, since we did not observe any differences in Akt and ERK1/2 activation in CA-MEK5α-Tg (5) and CHIP^−/+ (Fig. 8F) mice. The relative contributions and relationship between ERK5, Akt, and ERK1/2 in IGF-1-mediated cardioprotection should be investigated further.

Several apoptosis regulating molecules have been recognized as key regulators in cardiac dysfunction under pathological conditions. For example, mortality rate and incidence of left ventricular rupture after MI were significantly reduced and accompanied with reduced myocyte apoptosis in Trp53 targeted-deletion mice (35). Recently Liu et al. (36) reported that cardiac-specific overexpression of apoptosis signal-regulating kinase 1 develops cardiomyopathy accompanied with greater myocyte apoptosis following pressure-overload stimulation. Since both molecules are known substrates of CHIP (37, 38), it is interesting to speculate on the potential role of ERK5-CHIP cascade in protein degradation of apoptosis signal-regulating kinase 1 and p53 as it relates to myocyte apoptosis and cardiac dysfunction.

In summary, we have shown that ERK5 activation by IGF-1 enhanced the association of ERK5 with CHIP and increased CHIP Ub ligase activity. IGF-1 increased the association of ERK5 with CHIP linker and U-box domain, and the disruption of ERK5-CHIP binding with CHIP linker domain fragment significantly inhibited CHIP Ub ligase activity suggesting the critical role of ERK5-CHIP association on IGF-1-mediated CHIP Ub activity. IGF-1-mediated CHIP Ub activity was decreased by the depletion of ERK5 using ERK5 siRNA. The depletion of CHIP using CHIP siRNA in cardiomyocytes prevented IGF-1-mediated ICER protein destabilization and antiapoptotic effect. Increase in cardiac CHIP Ub ligase activity was observed in αMHC-CA-MEK5α-Tg mice, whereas acceleration of ICER induction and cardiac dysfunction were observed in heterozygous CHIP ^−/+ mice, which suggests the critical role of ERK5-CHIP complex in ICER ubiquitination and subsequent antiapoptotic effect in vivo.