Activation of p90RSK reduces CHIP Ub ligase activity induced by Ang II in vitro and after myocardial infarction in diabetic mice in vivo

ERK5-CHIP association is critical for the up-regulation of CHIP Ub ligase activity, leading to the ubiquitination and subsequent proteosomal degradation of ICER¹³. Since p90RSK activation was crucial for an Ang II-mediated increase in ICER levels, we investigated whether p90RSK could negatively regulate CHIP Ub ligase activity and thereby stabilize the ICER protein. To detect CHIP Ub ligase activity, we performed an in vitro ubiquitination assay using a GST-ICER fusion protein as a substrate¹³. CHIP Ub ligase activity was significantly decreased after Ang II stimulation, and transduction with Ad-DN-p90RSK significantly inhibited this phenomenon (Figs. 2A and 2B).

Open in a separate window

Figure 2

p90RSK inhibited CHIP Ub ligase activity and ERK5-CHIP association

(A) Ang II inhibited CHIP Ub ligase activity. Cardiomyocytes were transduced with either Ad-DN-p90RSK or Ad-LacZ for 24 h, followed by stimulation with Ang II (200 nM) for indicated times. Immunoprecipitated CHIP was subjected to an in vitro ubiquitination assay to determine CHIP Ub ligase activity using GST-fused ICER protein as a substrate (lane 1-4). The expression of Flag-tagged DN-p90RSK, CHIP, and tubulin was detected by Western blotting with each specific antibody. In vitro Ub ligase assay of GST-ICER with the Ub conjugation system and immunoprecipitates with IgG control (lane 7), and in vitro Ub ligase assay with GST only (lane 6) or without any GST substrate (lane 5) with the Ub conjugation system and immunoprecipitates with CHIP Ab (lane 1-6) were also shown. Immunobloting with anti-CHIP (2nd from top) and anti-ICER antibody (3rd and 4th from top) after in vitro CHIP Ub ligase assay suggested that the Ub bands we observed in this assay, especially above 120kDa, are mainly due to the ubiquitination of GST-ICER but not from GST, IgG or endogenous CHIP as described in the method. (B) Quantifications of relative CHIP Ub ligase activity from 3 independent experiments (mean±S.D., n=3, *p< 0.05 compared to the Ad-LacZ and untreated control). Results were expressed as the relative ratio of untreated cells in the LacZ control. (C) p90RSK activation inhibited ERK5-CHIP association. Cardiomyocytes were transduced with Ad-DN-p90RSK or Ad-LacZ for 24 h, followed by 45 min Ang II stimulation. Cell lysates were immunoprecipitated with anti-CHIP, and immunoblotted with anti-ERK5. Expression of ERK5, CHIP and Flag-tagged DN-p90RSK was detected by Western blotting using each specific antibody. (D) Relative co-immunoprecipitated ERK5 band intensitites from 3 independent experiments were quantified (mean±S.D., n=3, *p<0.05, compared to untreated control). Results are expressed as fold decrease in ERK5 band intensity in Ang II-treated cells that are expressing either Ad-LacZ or Ad-Flag-DN-p90RSK compared to untreated cells.

Next we investigated whether p90RSK activation could regulate ICER mRNA expression. As shown in Online Fig. IIIA, we found that Ang II increased ICER mRNA expression. However, we could not detect any inhibition of ICER mRNA expression by Ad-DN-p90RSK transduction. These data suggested that p90RSK activation regulates ICER levels at the post-transcriptional rather than the transcriptional level. Since CHIP Ub ligase activity requires ERK5-CHIP interaction¹³, and since Ang II regulates this process by activating p90RSK, we examined whether Ang II-mediated p90RSK activation could regulate ERK5-CHIP interaction. In quiescent cardiomyocytes, ERK5 activity regulates the association of ERK5 with CHIP ¹³. Indeed, we found that the ERK5-CHIP association was significantly decreased by stimulation of cardiomyocytes with Ang II ( s. 2C and 2D) and Ad-DN-p90RSK transduction blocked this decrease in cells treated with Ang II (Figs. 2C and 2D). These results indicated a key role for p90RSK in the reduction of ERK5-CHIP association, and subsequent inhibition of CHIP Ub ligase activity in cells stimulated with Ang II.

To explore the role of p90RSK activation in the reduction of CHIP Ub ligase activity in vivo, we utilized the cardiac-specific DN-p90RSK-Tg mice. These mice showed relative resilience to cardiac ischemia/reperfusion injury³¹, therefore we examined the effects of inhibiting p90RSK in diabetic mice after MI. The cardiac function of DN-p90RSK-Tg mice was unaltered ³¹, and the ICER levels in these transgenic were similar to that of non-transgenic littermate control (NLC) sham operated mice (Online Fig. IVA). Mice were rendered diabetic by intraperitoneal injection of streptozotocin (STZ). On day 7 after STZ injection, both NLC and DN-p90RSK-Tg mice showed elevated random blood sugar (BS) levels with similar body weight (Online Fig. VB). We have previously shown that preoperative insulin-treatment of STZ-injected mice stabilizes cardiac function after MI, suggesting myocardial derangement in diabetic mice after MI is a function of hyperglycemia rather than a toxicity artifact of STZ³². DN-p90RSK-Tg diabetic mice had improved survival after MI compared to that of NLC mice (Online Fig. VA). One week after MI, the LV weight/TL (tibial length) and lung weight/TL ratios were increased in NLC mice (Online Figs. VC and VD), whereas these parameters were reduced in DN-p90RSK-Tg mice (OnlineFig. VD). Echocardiography performed one week after MI showed increased left ventricular end diastolic diameter (LVEDd) and left ventricular end systolic diameter (LVESd) in diabetic NLC mice with a concomitant decrease in fractional shortening (FS) and ejection fraction (EF) (Online Fig. VE) compared to sham-operated non-MI controls. By contrast, we found that these MI-induced cardiac effects were attenuated in DN-p90RSK-Tg mice (Online Fig. VE).

The role of p90RSK activation in regulating CHIP Ub ligase activity, subsequent ICER levels and cardiomyocyte apoptosis in vivo were examined. As shown in Figs. 3A, 3D, and S6, CHIP Ub ligase activity was significantly decreased one week after MI in diabetic NLC mice but not in diabetic DN-p90RSK-Tg mice (Figs. 3A and 3D and Online Fig. VI). Since CHIP Ub ligase activity regulates ICER protein stability in heart failure ^{12, 13}, we examined the ICER protein level as well as the abundance of downstream ICER mediators, such as Bcl-2¹⁰ and SERCA2³³. In NLC mice, we found increased ICER levels and a concomitant reduction in Bcl-2 expression after MI in diabetic mice and this was rescued in DN-p90RSK-Tg diabetic mice after MI (Figs. 3A, 3B and 3D). SERCA2 expression was also reduced in NLC diabetic mice after MI. However, the reversal of this effect in DN-p90RSK diabetic mice after MI is less obvious (Figs. 3B and 3D). Interestingly, the level of ERK5 phosphorylation at T218 and Y220 was decreased slightly in DN-p90RSK diabetic mice after MI (Fig. 3C), suggesting a role for activated p90RSK in phosphorylating ERK5 T218/Y220-the same residues phosphorylated by MEK5 responsible for ERK5 activation.

Open in a separate window

Figure 3

CHIP Ub ligase activity and ICER expression after MI in diabetic DN-p90RSK-Tg mice

(A) Heart samples from sham control and DM+MI groups in NLC mice and DM+MI group in DN-p90RSK-Tg mice were collected and the CHIP Ub ligase activity assay was performed as described in the Methods (top). Western blots showing in (A) CHIP, ICER, and p90RSK, in (B) Bcl-2 and SERCA2, and in (C) p-(T218/Y220)ERK5, ERK5 and tubulin expression. (D) Quantifications of relative CHIP Ub ligase activity and ICER expression (upper) and the expression of Bcl-2 and SERCA2 (lower) (**p<0.01 *p< 0.05 compared to NLC sham operation control, mean±S.D., n=3,). Results are expressed as the relative ratio of mean value of NLC sham operation.

TUNEL-positive cells significantly increased in diabetic NLC mice after MI when compared with sham controls (Online Figs. VF and VG), supporting the known role of cellular apoptosis in the development of heart failure³⁴. Notably, reduced incidence of apoptosis was found in diabetic DN-p90RSK-Tg mice after MI (Online Figs. VF and VG), suggesting a functional consequence for p90RSK activation in promoting cardiac apoptosis after MI in diabetic animals.

p90RSK activation inhibits ERK5-CHIP association via competitive binding with ERK5 and direct phosphorylation of ERK5-S496

Upon activation, ERK5 exerts its cardio-protective effect by promoting ERK5-CHIP binding, which increases CHIP Ub ligase activity¹³. Since there is a role for p90RSK activation in Ang II-mediated disruption of ERK5-CHIP binding and subsequently reduction of CHIP Ub ligase activity (Figs. 1 and ​and2),2), we investigated how p90RSK regulates ERK5-CHIP association. To determine the protein domain of ERK5 responsible for the p90RSK-ERK5 and ERK5-CHIP interaction, we generated three truncated ERK5 fragments and evaluated their association with p90RSK (Fig. 4A) and CHIP (Fig. 4B) using a mammalian two-hybrid assay. A plasmid containing GAL4-DBD (DNA binding domain) and the full-length p90RSK (Fig. 4A) or CHIP (Fig. 4B) were constructed in the pBIND vector; plasmids to express VP16-ERK5 (full-length or truncated forms) were constructed in the pACT vector. Constitutively active MEK5α (CA-MEK5α) induces forced activation of ERK5 and significantly increases ERK5-CHIP association by decreasing the inhibitory effect of the NH₂-terminal kinase domain of ERK5. This ERK5-CHIP association at the helical linker domain of CHIP is fundamental for increasing CHIP Ub ligase activity¹³. Therefore, we examined the effect of CA-MEK5α expression on the association of ERK5 and CHIP. The mammalian two-hybrid assay demonstrated that the NH₂-terminal region of ERK5 (aa1-148) and the COOH-terminal region of ERK5 (aa571-807) associate with CHIP (Figs. 4A and 4B). The latter site overlaps with the p90RSK binding site.

Open in a separate window

Figure 4

p90RSK kinase activation inhibits ERK5-CHIP association

(A, B) p90RSK-ERK5 and CHIP-ERK5 association were examined by a mammalian two-hybrid assay in Hela cells. The activation domain VP16 fused to full length or three truncated ERK5 proteins (A-B) and full length p90RSK (A) or CHIP (B) fused to the Gal4 binding domain were cotransfected with the Gal4-responsive luciferase reporter pG5-luc. The total transfected DNA amount was normalized with empty VP16 vector. Binding affinity was assayed by luciferase activity normalized by the renilla luciferase activity. Data are representative of three independent experiments. **p<0.01, mean±S.D., n=3. (C) Both wild type p90RSK (WT-p90RSK) and ERK5-Fr (aa571-807) but not DN-p90RSK disrupted ERK5-CHIP association. After plasmid transfection as indicated, immunoprecipitated ERK5 was immunoblotted with anti-CHIP. Protein expression was determined by immunoblotting with each specific antibody. The blots are representative of data obtained from at least three different blots. (D) Ang II increased p90RSK-ERK5 interaction via p90RSK activation. Cardiomyocytes were stimulated with Ang II after 3h of FMK-MEA pre-treatment. Immunoprecipitated ERK5 was immunoblotted with anti-p90RSK. Each protein expression was detected as described in C. (E) Relative p90RSK-ERK5 binding was quantified as described in Fig.2D. *p<0.05, mean±S.D., n=3.

To examine whether the wild type p90RSK and the ERK5 fragment (ERK5-Fr, aa571-807) alter ERK5-CHIP association in vitro, we utilized an over-expression system. Since CHIP binding site is within the ERK5 aa571-807 region, we focused on ERK5-Fr (aa571-807). ERK5-CHIP binding was increased by CA-MEK5α (Fig. 4C, lane 3). Over-expression of both wild type p90RSK and ERK5-Fr inhibited ERK5-CHIP association (Fig. 4C, lane 5-6 and Online Fig. VIIA, B), suggesting that the ability of WT p90RSK to inhibit ERK5-CHIP association was due to competition between p90RSK and ERK5 for the same binding site on CHIP (Fig. 8E).

Open in a separate window

Figure 8

CHIP Ub ligase activity and ICER expression after MI in NLC, WT-p90RSK-Tg, and Double-Tg mice

(A-C) Heart samples from control and MI groups in NLC, WT-p90RSK-Tg, and Double-Tg mice were collected and CHIP Ub ligase activity assay was performed as described in methods (top). Western blot shows (A) CHIP, ICER, p90RSK, p-ERK5, and ERK5 and (B) Bcl-2, SERCA2, and tubulin expression. (C) Quantifications of relative CHIP Ub ligase activity and ICER expression (upper), Bcl-2 and SERCA2 expression (lower) as described in Fig.2B and Fig.1B. (**p<0.01, * p<0.05, mean±S.D., n=3). (D) A model of DM + MI or Ang II-mediated p90RSK-ERK5-CHIP signal transduction pathway that regulates cardiac apoptosis and subsequent cardiac dysfunction. (E) A scheme depicting p90RSK-mediated regulation of the ERK5-CHIP module. At the basal level, inactive p90RSK inhibits the D-domain to bind with ERK5³⁹. p90RSK-free ERK5 associates with CHIP at its linker and U-box domain and maintains its CHIP Ub ligase activity to prevent ICER induction and subsequent apoptosis¹³. However, once p90RSK gets activated, the inhibition of the kinase domain is released³⁹, and the D-domain of p90RSK associates with the ERK5 COOH-terminal domain, leading to compete with ERK5-CHIP association and ERK5-S496 phosphorylation, which disrupts ERK5-CHIP interaction. The disruption of ERK5-CHIP interaction inhibits CHIP Ub ligase activity¹³, increases ICER induction, and induces apoptosis^{13, 10}.

Since p90RSK inhibits ERK5-CHIP association via competitive binding with ERK5, we have investigated whether Ang II-mediated p90RSK activation is required for p90RSK-ERK5 association in cardiomyocytes. Ang II increased p90RSK-ERK5 binding, and this was significantly inhibited by FMK-MEA (Figs. 4D and 4E). As p90RSK activation induced by Ang II increased p90RSK-ERK5 binding, we determined the role of p90RSK activation in ERK5-CHIP binding. Wild type p90RSK but not DN-p90RSK completely inhibited ERK5-CHIP association (Fig.4C, lane 4-5, and Online Fig.VIIC and D), supporting the role of p90RSK activation in increased p90RSK-ERK5 association as well as decreased ERK5-CHIP association. Based on these results, we reasoned that since Ang II induced p90RSK, but not ERK5, activation (Online Fig. VIII), the regulation of ERK5-CHIP binding mediated by p90RSK activation most likely occurs by a process that requires direct interaction of the two without ERK5 enzymatic activity being activated (Fig. 8E).

Since p90RSK can associate with ERK5, we asked whether p90RSK directly phosphorylated ERK5 and as a consequence inhibited CHIP Ub ligase activity. To test this possibility, we performed an in vitro kinase assay by incubating recombinant p90RSK with GST-tagged-ERK5 fragments and found that p90RSK phosphorylated two ERK5 fragments (aa401-600 and aa601-816) (Fig. 5A). In addition, liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) analyses identified ERK5-S496 as being directly phosphorylated by p90RSK (Figs. 5B and 5C). To determine whether ERK5-S496 phosphorylation is involved in the Ang II-mediated reduction of CHIP Ub ligase activity in cardiomyocytes, we generated an adenovirus expressing the ERK5-S496A mutant (Ad-ERK5-S496A). Cardiomyocytes were transduced with Ad-ERK5-S496A, Ad-ERK5-WT, or Ad-LacZ, followed by Ang II stimulation. Cell lysates were then subjected to an in vitro CHIP ubiquitination assay using a GST-ICER fusion protein as a substrate. Ang II inhibited CHIP Ub ligase activity in cardiomyocytes transduced by Ad-LacZ or Ad-ERK5-WT. In contrast, Ad-ERK5-S496A transduction abolished this inhibitory effect on CHIP Ub ligase activity (Figs. 5D and 5E). In addition, the Ang II-mediated increase in ICER levels was significantly inhibited in cardiomyocytes overexpressing Ad-ERK5-S496A mutant, but not in those expressing the Ad-ERK5-WT or Ad-LacZ (Figs.5F and 5G). These findings indicate that p90RSK-mediated ERK5-S496 phosphorylation is required for inhibiting CHIP Ub ligase activity (Figs. 8D and 8E).

Open in a separate window

Figure 5

p90RSK phosphorylated ERK5-Ser496 and inhibited CHIP Ub ligase activity

(A) In vitro p90RSK kinase assay using 4 different GST-ERK5 fragments as substrates. An autoradiogram (A, upper panel) shows strong ³²P incorporation into a GST-ERK fragment (aa401-600), and weak incorporation into another fragment (aa601-816). Ponceau staining (A, lower panel) of the membrane demonstrates the position of proteins after separation by SDS-PAGE. (B) LC-MS/MS analysis identified p90RSK-mediated ERK5-S496 phosphorylation with less than 0.05 error. An LC-MS/MS plot of different peptide fragments shows S496 phosphorylation based on mass shift of phosphorylated serine with H₂O (+98 Da). The predicted mass of each b/y fragment and the actual value are shown above with the b6 fragment containing the phosphorylated S496 having a predicted mass of 706.4 Da versus an actual value of 706.31 Da. (C) An error plot of the LC-MS/MS data shows the error for calculating each b/y peptide fragment is less than 0.5 Da. (D) Ang II-mediated reduction of CHIP Ub ligase activity was inhibited by ERK5-S496A mutant. Cardiomyocytes were transduced with either Ad-ERK5-WT, Ad-ERK5-S496A mutant, or Ad-LacZ for 24 h followed by 2 h of Ang II stimulation as indicated. (E) Relative CHIP Ub ligase activity was quantified as described in Fig.2B. (**p<0.01, *p<0.05, mean±S.D., n=3). (F) Ang II-mediated ICER induction was inhibited by ERK5-S496A mutant. Cardiomyocytes were transduced as described in D followed by 24 h of Ang II stimulation. (G) Quantification of ICER expression as described in Fig.1B. (**p<0.01, *p<0.05, mean±S.D., n=3).

Cardiac-specific ERK5 knockout mice (ERK5-CKO) showed reduced CHIP Ub ligase activity after MI

To corroborate the necessity of ERK5 activation in increasing CHIP Ub ligase activity and subsequent reduction of ICER, we utilized ERK5-CKO. Since ERK5 activation is cardio-protective¹³ and ERK5-CKO mice displayed significantly increased levels of cardiac dysfunction after MI under the non-DM condition (Figs. 6A, 6B and 6C), we subjected the mice to MI only. No basal cardiac abnormality was found in ERK5-CKO mice as we reported previously³⁵, and no increase in ICER levels was observed (Online Fig. IVB). After MI, ERK5-CKO mice showed a significant reduction in CHIP Ub ligase activity, and an increase in ICER levels (Figs. 6F and 6G) compared to the NLC mice. As expected, the ERK5-CKO mouse hearts showed a significant increase in the number of TUNEL-positive cardiomyocytes compared with the NLC group (Figs. 6D and 6E). These data further confirmed the importance of ERK5 in regulating CHIP Ub ligase activity and subsequent ICER levels and apoptosis in vivo (Fig. 8D). Since a difference in the infarct size may correlate with post-MI cardiac dysfunction, we compared the permanent coronary ligation-mediated infarct size between NLC (C57BL/6) and ERK5-CKO mice. As shown in Online Fig. IX, we did not find any difference in the infarct size among these groups, although we found significant cardiac dysfunction and an increase in lung weight/TL in ERK5-CKO mice.

Open in a separate window

Figure 6

CHIP Ub ligase activity and ICER expression after MI in ERK5-CKO mice

(A) LV weight/tibial length (TL) and (B) lung weight/TL after MI in NLC and ERK5-CKO mice as indicated (* p<0.05, mean±S.D., n=8-9). (C) Echocardiographic data obtained after MI in NLC and ERK5-CKO mice as indicated. FS (fractional shortening), EF (ejection fraction). (** p<0.01, mean±S.D., n=8-9) (D) Cardiomyocyte apoptosis in remote areas was enhanced in ERK5-CKO mice. Animals were subjected to either MI or sham. TUNEL-positive cardiomyocytes were counted among more than 8,000 to 10,000 nuclei in the remote area of each mouse heart after one week of surgery. 40X objective lens. Scale bars, 40 μm. Representative pictures of TUNEL (top), 4’, 6’-diamidino-2-phenylindole (DAPI) (middle), and α-actinin merged with TUNEL and DAPI staining (bottom) of the remote area from NLC and ERK5-CKO mice subjected to MI or sham operation as indicated. (E) A bar graph showing TUNEL-positive cells (%) in various animals. (* p<0.05, mean±S.D., n=3-4). (F) Heart samples from sham and MI groups in NLC mice and in ERK5-CKO mice were collected and CHIP Ub ligase activity assay was performed as described in methods (top). CHIP, ERK5, ICER, and tubulin expressions were detected by Western blot. (G) Relative CHIP Ub ligase activity (left, n=9) and ICER expression (right, n=3) were quantified as described in Fig.2B and Fig.1B, respectively. (** p<0.01, * p<0.05, mean±S.D.).

The LC-MS/MS analysis was performed by the University of Rochester Proteomics Center (Dr. Fred K. Hagen, Ph.D). We thank Drs. Jane Sottile, Bradford C. Berk, and Scott Cameron for critical reading and suggestions for the manuscript.

Sources of funding

This work is supported by grants from the Mochida Memorial Foundation for Medical and Pharmaceutical Research to Dr. Tetsuro Shishido, the America Heart Association to Dr. Woo (SDG 0930360N), and from the National Institute of Health to Dr. Abe (HL102746-03, HL088637, and HL108551), Dr. Yan (HL077789 and HL088400), and Dr. Yang (HL088637 and HL108551). Drs. Abe and Yan are recipients of Established Investigator Awards of the American Heart Association (0740013N and 0740021N).