Glycogen Synthase Kinase 3β Inhibits Myocardin-Dependent Transcription and Hypertrophy Induction Through Site-Specific Phosphorylation
Cardiomyocyte hypertrophy is transcriptionally controlled and inhibited by glycogen synthase kinase 3β (GSK3β). Myocardin is a muscle-specific transcription factor with yet unknown relation to hypertrophy. Therefore, we investigated whether myocardin is sufficient to induce cardiomyocyte hypertrophy and whether myocardin is regulated by GSK3β through site-specific phosphorylation. Adenoviral myocardin overexpression induced cardiomyocyte hypertrophy in neonatal rat cardiomyocytes, with increased cell size, total protein amount, and transcription of atrial natriuretic factor (ANF). In vitro and in vivo (HEK 293 cells) kinase assays with synthetic peptides and full-length myocardin demonstrated that myocardin was a “primed” GSK3β substrate, with serines 455 to 467 and 624 to 636 being the major GSK3β phosphorylation sites. Myocardin-induced ANF transcription and increase in total protein amount were enhanced by GSK3β blockade (10 mmol/L LiCl), indicating that GSK3β inhibits myocardin. A GSK3β phosphorylation-resistant myocardin mutant (8xA) activated ANF transcription twice as potently as wildtype myocardin under basal conditions with GSK3β being active. Conversely, a GSK3β phospho-mimetic myocardin mutant (8xD) was transcriptionally repressed after GSK3β blockade, indicating that GSK3β phosphorylation at the sites identified inhibits myocardin transcriptional activity. GAL4-myocardin fusion constructs demonstrated that GSK3β phosphorylation reduced the intrinsic myocardin transcriptional activity. A cell-permeable (Antennapedia protein transduction tag) peptide containing the mapped myocardin GSK3β motifs 624 to 636 induced hypertrophy of cultured cardiomyocytes, suggesting that the peptide acted as substrate-based GSK3β inhibitor in cardiomyocytes. Therefore, we conclude that the GSK3β–myocardin interaction constitutes a novel molecular control of cardiomyocyte hypertrophy. Phosphorylation by GSK3β comprises a novel post-transcriptional regulatory mechanism of myocardin.
Myocardin is a muscle-specific transcription factor of the SAP domain family of transcription factors and is predominantly expressed in cardiac and smooth muscle cells.1,2 Myocardin is required for myocardial cell differentiation in the developing Xenopus embryo and potently activates a cardiac muscle gene program, which includes transcription of atrial natriuretic factor (ANF), a molecular index of hypertrophy.1 For smooth muscle gene expression, myocardin displaces the ternary complex factor Elk-1 from serum response factor.3 The role and regulation of myocardin in cardiomyocyte hypertrophy signaling, however, is yet unknown. In particular, post-translational myocardin modification regulating its transcriptional activity is unclear.
Cardiomyocyte hypertrophy is a physiological process that occurs in response to biomechanical stress induced by pressure or volume overload in animals and humans. Pharmacologically, cardiomyocyte hypertrophy can be induced by pro-hypertrophic agonists such as angiotensin II or endothelin-1 (ET-1).4,5 Cardiomyocyte hypertrophy is regulated by multiple signaling cascades and transcriptional events.4,5 An important negative regulator of cardiomyocyte hypertrophy is glycogen synthase kinase 3β (GSK3β),6–8 an ubiquitously expressed serine/threonine kinase that, in contrast to many other kinases, is constitutively active and is inactivated by agonist-induced phosphorylation at serine residue 9.9–11 Studies in cultured cardiomyocytes and in transgenic mice have shown that GSK3β inactivation by phosphorylation is necessary and sufficient for the development of cardiomyocyte hypertrophy.6,7,12
One the best understood GSK3β targets in hypertrophic signaling is the family of nuclear factors of activated T cells (NFATs), which are phosphorylated by GSK3β at their nuclear localization signals under basal conditions. This shields the NFAT nuclear localization signal and retains NFATs in the cytosol, thus rendering them transcriptionally inactive. Pro-hypertrophic agonists inactivate GSK3β, thus reducing NFAT phosphorylation and allowing NFAT translocation to the nucleus for activation of a hypertrophic gene program.4,5 Although a number of further GSK3β substrates such as β-catenin, GATA4, c-myc, JunD, and CREB have been described, none of these known GSK3β targets fully explains the anti-hypertrophic properties of GSK3β.6–8,13
Because cardiomyocyte hypertrophy is regulated at the transcriptional level, we hypothesized that GSK3β may phosphorylate and thereby regulate additional cardiac transcription factor(s). Myocardin appeared as an ideal candidate for being a downstream target of GSK3β, as myocardin expression is limited to muscle tissue and, most importantly, possesses several highly conserved GSK3β phosphorylation motifs. Therefore, we investigated whether myocardin induces cardiomyocyte hypertrophy and whether myocardin function is regulated by GSK3β. Consequently, we identified myocardin as a novel modulator of cardiomyocyte hypertrophy being controlled by GSK3β through phosphorylation at specific serine motifs.
Materials and Methods
Full-length human S9AGSK3β-cDNA and the 3.7kB rat ANF promoter-driven luciferase reporter have been described.12 Full-length mouse myocardin cDNA (gene bank accession number AF384055) was amplified by reverse transcription polymerase chain reaction (Superscript System, Invitrogen) from neonatal mouse cardiomyocytes. All clones obtained contained the 48-amino-acid splice insert known from Myocardin A14 (gene bank accession number AF437877) yielding a cDNA of 983 amino acids. Correct clones were subcloned into pcDNA 3.1(+) myc-His-A (Invitrogen) and sequenced. To generate GAL4-myocardin, the GAL4 DNA binding domain was amplified from pFA-CMV (Stratagene) and replaced the N-terminal portion of myocardin up to the endogenous BamHI site as described.1 The pFR-Luciferase reporter was from Stratagene. Site-directed mutagenesis was performed as described.12
Full-length mouse myocardin, including a carboxy-terminal myc-tag, was inserted into the pAdTrack-CMV vector (AdEasy-system).15 This vector contains 2 cytomegalovirus (CMV) promoters independently driving green fluorescent protein (GFP) and the gene of interest. Viral particles from the supernatant of infected HEK293 cells were purified using the Adeno-X virus purification kit (BD Biosciences).
Cell Culture and Transfection
Neonatal ventricular cardiomyocytes were isolated from 1- to 2-day-old Sprague-Dawley rats (Charles River Laboratories, Boston, Mass) as described.12 The next day, cardiomyocytes were transfected using Effectene (Qiagen). Six hours after transfection, cells were serum-starved and stimulated for 48 hours with ET-1 (100 nmol/L) or LiCl (10 mmol/L). Luciferase activity was measured using a standard assay system (Promega). Each experiment contained triplicate wells. Human embryonic kidney (HEK 293) cells were transfected as described.16
Peptide Phosphorylation Assays
Custom-synthesized peptides (Affina) were solubilized in kinase buffer16 at a final concentration of 200 μmol/L. In a 30 μL reaction volume, 500 U of GSK3β (Cell Systems) with 5 μCi of γ32P-ATP (Hartmann) were added. After 30 minutes at 30°C, reaction products were separated by reverse phase FPLC using a Source 5RPC column and an Äkta explorer HPLC system (both Amersham Pharmacia). The starting buffer was 10 mmol/L NH4-acetate, pH 7.0 in 2% acetonitrile, and bound peptide was eluted by a linear increase of the acetonitrile concentration.16
In Vitro Kinase Assay
Forty-eight hours after transfection of HEK 293 cells, myocardin constructs were immunoprecipitated in lysis buffer using an anti-myc antibody (clone 9E5, Santa Cruz) as described.12 After extensive washing with kinase buffer, 500 U of GSK3β together with 5 μCi of γ32P-ATP was added in a final volume of 30 μL. After 30 minutes at 30°C, proteins were separated by 10% SDS-PAGE and phosphorylation was detected by autoradiography.16
In Vivo Phosphorylation
Forty-eight hours after transfection, HEK 293 cells were switched to phosphate-free Dulbecco’s modified eagle’s medium-high glucose (ICN) supplemented with 10% phosphate-free fetal calf serum and 250 μCi 32P-orthophosphoric acid (Hartmann). After 3 hours, myocardin constructs were immunoprecipitated using anti-myc antibody as described above and phosphorylation was detected by SDS-PAGE and autoradiography.16
Total Protein Amount Measurements
Twenty-four hours after plating, cardiomyocytes were transduced with adenoviruses at the multiplicity of infection (MOI) indicated or peptides (final concentrations: 1 μmol/L peptide in 0.1% DMSO) and incubated with serum-free and leucine-free medium (ICN) for 36 hours in the presence of 1.5 μCi/mL 3H-leucine (Hartmann). After trichloroacetic precipitation, scintillation counting was performed as described.12
Immunostaining and Microscopy
After fixation and permeabilization, cells were stained with a mouse monoclonal anti-α sarcomeric actinin antibody (clone EA-53, Sigma) followed by Cy3-conjugated anti-mouse immunoglobulin G antibody (Dianova). Nuclei were stained with DAPI (Sigma). Cells were imaged using a fluorescence microscope as described.12
Peptide Transduction of Cardiomyocytes
Peptides were dissolved in dimethylsulfoxide and added (1.5 μmol/L final) to serum-starved cardiomyocytes. After the time points indicated, cells were trypsinized and resuspended in phosphate-buffered saline containing 1% bovine serum albumin (all Sigma), and 10 000 cells were analyzed for their green fluorescence using a flow cytometry system (FACScalibur cell sorter, BD Biosciences). Forty-eight hours after peptide transduction, cell size measurements were performed as described.12
Student t test was used for the comparison of independent groups. For multiple group comparisons, 1-way ANOVA testing was used.
Myocardin Overexpression Induces Cardiomyocyte Hypertrophy
We used rat neonatal cardiomyocytes as an established cell culture system for cardiomyocyte hypertrophy.4,5,12,17 To study the effects of myocardin on cardiomyocyte hypertrophy, we generated a replication-defective adenovirus15 that simultaneously expresses GFP (to identify successfully transduced cells) and full-length, myc-tagged mouse myocardin from 2 CMV promoter-driven expression cassettes. Transduction of cardiomyocytes with this adenovirus resulted in GFP as well as myocardin expression (data not shown). Virally transduced cardiomyocytes overexpressing myocardin were clearly enlarged and displayed enhanced sarcomeric organization in comparison to nontransduced cardiomyocytes (Figure 1A). Quantitative analysis revealed a significantly increased cell length and cell surface area of cardiomyocytes overexpressing myocardin in comparison to nontransduced cardiomyocytes (Figure 1B) or cardiomyocytes overexpressing only GFP (data not shown). Overexpression of myocardin, but not GFP, resulted in a significantly enhanced total protein amount similar to pharmacological induction of hypertrophy with the GSK3β inhibitor lithium chloride (Figure 1C). In addition, myocardin overexpression activated mRNA and protein expression of ANF as a molecular marker of cardiomyocyte hypertrophy (Figure 1D and data not shown). Taken together, these data demonstrate that myocardin overexpression is sufficient to induce hypertrophy of neonatal cardiomyocytes.
GSK3β Phosphorylates Primed Myocardin Peptides In Vitro
GSK3β can phosphorylate substrates with a (S/T-X3-S/T)n-motif. After initial phosphorylation of a serine or threonine residue (P+4) by a different kinase (“priming kinase”), GSK3β sequentially phosphorylates serine or threonine residues located 4 amino acids apart from one another in a carboxy-to-amino-terminal direction (P0, P−4, P−8). Many, if not all, GSK3β substrates require this priming to be phosphorylated by GSK3β. The residues between the phospho-acceptor sites are not conserved, but prolines are favored at the carboxy-terminal sites of the serines/threonines, characterizing GSK3β as a proline-directed kinase.9–11,13
Mouse myocardin contains 2 putative GSK3β motifs (amino acids [aa] 455 to 467 and aa624 to aa636) that are 100% conserved between mouse and human myocardin. Many of the serine residues are neighbored by carboxy-terminal prolines (Figure 2A). Secondary structure prediction indicated a high surface and accessibility probability for these motifs (data not shown).
We used synthetic peptides containing the putative GSK3β phosphorylation motifs with (primed peptides, phospho-467 and phospho-636, respectively; named after the P+4 residue) or without (nonprimed peptides 467 and 636, respectively) a phospho-serine in the position of the putative priming site (Figure 2A). Peptides were incubated in vitro with recombinant, purified GSK3β in the presence of γ32P-ATP, and reaction products were resolved by FPLC. Using the nonprimed peptides, no shift in the peptide retention time (data not shown) or incorporation of radioactivity (Figure 2D and 2E) was detected, indicating lack of phosphorylation by GSK3β. In contrast, both primed peptides were phosphorylated by GSK3β, as evidenced by both shifted peptide peaks (Figure 2B and 2C) and radioactivity incorporation in the fractions corresponding to the shifted peaks (Figure 2D and 2E). Thus, primed myocardin peptides constitute GSK3β substrates in vitro.
Mapping of the Major GSK3β Phosphorylation Sites in Myocardin
Next, we investigated if GSK3β phosphorylates the complete myocardin protein. Full-length mouse myocardin was expressed in HEK 293 cells and immunoprecipitated via a C-terminal myc tag for in vitro kinase assays. Consistent with the peptide results, GSK3β readily phosphorylated full-length wild-type mouse myocardin (Figure 3A, lane 2). Replacement of individual P0 (aa463 or aa632) or P+4 (aa467 or aa636) serine residues with alanines did not result in a significant decrease of myocardin phosphorylation by GSK3β (Figure 3A). Therefore, we generated myocardin mutants, where all 4 serines of a GSK3β phosphorylation motif were replaced by alanines (S455,459,463,467Amyocardin and S624,628,632,636Amyocardin), and a combination mutant (8xAmyocardin), in which all 8 serines in both GSK3β phosphorylation motifs (serines 455,459,463,467,624,628,632, and 636) were replaced by alanines. The mutants S455,459,463,467Amyocardin and S624,628,632,636Amyocardin both showed a trend toward a decreased phosphorylation, but only the combination mutant (8xAmyocardin) was significantly resistant toward phosphorylation by GSK3β (Figure 3B and 3C). These results indicate that the myocardin amino acid motifs 455 to 467 and 624 to 636 are the primary GSK3β phosphorylation sites in myocardin.
In Vivo Phosphorylation of Myocardin by GSK3β
To assess whether myocardin is a GSK3β substrate within intact cells, we performed in vivo phosphorylation assays in transiently transfected HEK 293 cells. Expression of wild-type mouse myocardin resulted in its phosphorylation (Figure 3D, lane 2). Coexpression of a dominant-active GSK3β mutant (S9AGSK3β) together with myocardin led to a marked increase of myocardin phosphorylation (Figure 3D, lane 3). Importantly, expression of 8xAmyocardin resulted in a substantially reduced phosphorylation in intact cells, both under basal conditions and after coexpression of dominant-active GSK3β (Figure 3D, lanes 4 and 5).
GSK3β Inhibits Myocardin Transcriptional Activity and Hypertrophy Induction
The effects of GSK3β on the myocardin transcriptional activity were investigated in rat neonatal cardiomyocytes. Because myocardin potently activates ANF transcription, ANF promoter-driven luciferase activity was assessed as molecular index of cardiomyocyte hypertrophy.1,18 All experiments were performed in serum-free medium to ensure that GSK3β is catalytically active under control conditions. Endogenous GSK3β is inhibited by LiCl and ET-1, leading to a 3- to 5-fold activation of the ANF promoter in the absence of transfected myocardin (Figure 4A). Expression of wild-type myocardin (wtmyocardin) led to a 17-fold activation of the ANF promoter under serum-free conditions, which was further significantly enhanced by GSK3β inhibition through either LiCl or ET-1 (Figure 4A). These results were confirmed by detecting ANF mRNA and protein levels (data not shown). Thus, GSK3β negatively regulates myocardin transcriptional activity in cardiomyocytes.
We then investigated whether GSK3β modulates the hypertrophy induction caused by myocardin overexpression. To this end, 3H-leucine incorporation was quantified as metabolic hallmark of cardiomyocyte hypertrophy. Adenoviral overexpression of myocardin, but not GFP, led to a significant dose-dependent increase in total protein amount in rat neonatal cardiomyoyctes (Figure 4B). Inhibition of the endogenous GSK3β with LiCl led in itself (no adenovirus, MOI 0) to a significant increase in total protein amount. This was dose-dependently and significantly further augmented by myocardin overexpression to a “supra-physiological” level, until it leveled off at about a 2-fold increase in protein amount (Figure 4B). These data indicate that GSK3β serves as a molecular negative control of the myocardin-induced hypertrophic response.
GSK3β Phosphorylation at the Mapped Motifs Inhibits the Intrinsic Transcriptional Activity of Myocardin
Next, we investigated whether GSK3β phosphorylation at the motifs identified (serines 455, 459, 463, and 467 and serines 624, 628, 632, and 636) alters myocardin transcriptional activity. Therefore, we measured activation of ANF transcription after transfection of myocardin mutants in which these serine residues had been replaced either by nonphosphorylatable alanines or by phospho-mimetic aspartates. Whereas the transcriptional activities of myocardin mutants in which only 1 phosphorylation motif was eliminated (S455,459,463,467Amyocardin and S624,628,632,636Amyocardin) were not different from the wild-type molecule, the phosphorylation-resistant 8xAmyocardin mutant was significantly (P<0.05) more active under basal (serum-free) conditions in cultured cardiomyocytes (Figure 5A). The transcriptional activity of all constructs could be further enhanced by inhibition of the endogenous GSK3β using either LiCl or, to a lesser extent, ET-1 (Figure 5A). Moreover, when compared with wild-type myocardin, adenoviral overexpression (MOI 25) of 8xAmyocardin increased cardiomyocyte maximal cell length and cell surface area to a significantly higher level (data not shown). These results indicate that rendering myocardin resistant to GSK3β-mediated phosphorylation increases its baseline transcriptional activity and augments its hypertrophy induction. Of note, additional regulatory mechanism(s) influence myocardin activity, as the 8xAmyocardin mutant failed to exhibit a dominant-active or dominant-negative phenotype (further induction by LiCl or ET-1). Conversely, a phospho-mimetic myocardin mutant carrying 8 aspartates in place of the GSK3β-phosphorylatable serines (8xDmyocardin) was not different from the wild-type myocardin molecule under basal conditions (where GSK3β is active), but displayed a partially repressed transcriptional state after GSK3β inhibition with LiCl (Figure 5B).
Importantly, the protein stability of the various myocardin mutants were comparable, thus ruling out the confounding possibility that differences in protein turnover caused the observed differences of transcriptional activity (data not shown). Therefore, these results confirmed that GSK3β phosphorylation at the sites identified functionally inhibited myocardin transcriptional activity.
All mutations in 8xAmyocardin lie in the regulatory domain of myocardin but outside of the binding region for serum response factor.1 To explore whether 8xAmyocardin has an intrinsically higher transcriptional activity, we fused myocardin to the DNA binding domain of the yeast GAL4 transcription factor as previously described.1 GAL4-myocardin potently activated a GAL4 responsive element-dependent luciferase reporter in transfected HEK 293 cells (Figure 5C). Because the experiments in the HEK 293 cells were performed in the presence of 10% fetal calf serum (which inhibits GSK3β), there were no differences between wtmyocardin and 8xAmyocardin at baseline. In contrast, coexpression of dominant-active S9AGSK3β significantly and specifically reduced the transcriptional activity of wtmyocardin but not 8xAmyocardin (Figure 5C).
Next, we determined the effects of GAL4-wtmyocardin and GAL4-8xAmyocardin on the GAL4 responsive element-dependent luciferase activity in cardiomyocytes. As described in Figure 4, cardiomyocytes were cultured in serum-free medium to ensure that GSK3β is active under baseline conditions. Interestingly, when compared with GAL4-wtmyocardin, GAL4-8xAmyocardin had a significantly enhanced basal transcriptional activity in rat neonatal cardiomyocytes (Figure 5D) because of its reduced phosphorylation by GSK3β. Similar to the results shown in Figure 4A and Figure 5A, however, GAL4-8xAmyocardin transcriptional activity was further inducible by LiCl and ET-1 (Figure 5D). These data indicate that phosphorylation by GSK3β reduces the intrinsic myocardin transcriptional activity.
A Myocardin-Based Peptide Containing the Mapped GSK3β Phosphorylation Site Induces Cardiomyocyte Hypertrophy
Finally, we investigated whether the identified myocardin phosphorylation motifs can be recognized by endogenous GSK3β in cultured cardiomyocytes and whether they might act as substrate analogue GSK3β inhibitors. On the basis of the known anti-hypertrophic properties of GSK3β, we hypothesized that such a mode of action would result in cardiomyocyte hypertrophy.
Therefore, synthetic peptides containing the mapped GSK3β phosphorylation sites, a phospho-serine at the P+4 position (to generate primed peptides recognizable by GSK3β), and a tag derived from the Antennapedia (Ap) protein transduction domain were generated. The Ap tag was included to allow cellular uptake of the peptide. Additionally, fluorescein was N-terminally coupled to the peptide to monitor peptide uptake into living cardiomyocytes. The design of the peptides is shown in Figure 6A. After incubation of cardiomyocytes with 1.5 μmol/L peptide, green fluorescence of the cardiomyocytes was quantitated using flow cytometry.
As shown in Figure 6B, peptides Ap-myocardinp467 and Ap-myocardinp636 were both initially (6 hours) taken up by more than 99% of cardiomyocytes. Importantly, addition of peptide Ap-myocardinp636 resulted in a significantly (P<0.01) higher mean fluorescence at 6 hours (2497 versus 1025 relative light units). Whereas the percentage of cells positive for peptide Ap-myocardinp467 decreased to 83% at 48 hours, virtually all (99.7%) cells were still positive for peptide Ap-myocardinp636. Again, transduction of cardiomyocytes with peptide Ap-myocardinp636 resulted in a significantly (P<0.01) higher mean relative light units at 48 hours. These data indicate that peptide Ap-myocardinp636 efficiently transduced virtually all cardiomyocytes and is relatively stable up to 48 hours after transduction, whereas cardiomyocyte transduction using peptide Ap-myocardinp467 was significantly less efficient.
We then measured total protein amount in peptide-transduced cardiomyocytes as a metabolic marker of hypertrophy induction and indirect evidence of GSK3β inhibition. Pharmacological GSK3β inhibition with LiCl resulted in an increased total protein amount compared with vehicle (DMSO), similar to the results shown in Figure 1 and Figure 4 (Figure 6C). Importantly, an Ap-scrambled peptide (containing a myocardin-unrelated sequence) did not result in increased total protein amount despite successful transduction (data not shown), thus serving as a negative control (Figure 6C). In contrast, the myocardin-based peptides Ap-myocardin636 resulted in a significant (P<0.05) increase in total protein amount (Figure 6C). Furthermore, cardiomyocyte transduction with the peptide Ap-myocardinp636 but not Ap-myocardinp467 led to an increased cardiomyocyte length and surface area, 2 morphological hallmarks of cardiomyocyte hypertrophy (Figure 6D). Taken together, the cell-permeable peptide containing the mapped myocardin motif 624 to 636 induces cardiomyocyte hypertrophy potentially through GSK3β inhibition.
The data presented here show that myocardin overexpression is sufficient to induce cardiomyocyte hypertrophy in cultured cardiomyocytes. Myocardin is phosphorylated by GSK3β at multiple serine residues, thus negatively regulating the intrinsic myocardin transcriptional activity and hypertrophy induction in cultured cardiomyocytes. Additionally, a cell-permeable myocardin-based peptide induced cardiomyocyte hypertrophy.
Whereas we have conclusively demonstrated that myocardin overexpression is sufficient to evoke a hypertrophic phenotype as demonstrated by morphological, biochemical, and molecular criteria, it is yet unknown whether myocardin is critical for the development of cardiomyocyte hypertrophy in vivo. Importantly, GSK3β inhibited myocardin-dependent ANF transcription and cardiomyocyte total protein amount. Thus, GSK3β is a negative myocardin regulator, and this signaling pathway may contribute to the anti-hypertrophic effects of GSK3β. Interestingly, simultaneous myocardin overexpression and GSK3β inhibition lead to a “supra-physiological” increase in cardiomyocyte total protein amount. This indicates that maximal myocardin activation is able to override an intrinsic barrier of cardiomyocyte hypertrophy. On the basis of these results, we propose a model for the GSK3β-myocardin interaction (Figure 7): Under basal conditions, GSK3β and an elusive priming kinase phosphorylate myocardin, thus reducing its transcriptional activity. Pro-hypertrophic agents induce a PI3-kinase– and Akt-dependent GSK3β inactivation via phosphorylation at serine 9. This releases the myocardin inhibition and allows myocardin to activate a hypertrophic gene program. Of course, GSK3β inhibition has simultaneously profound effects on other GSK3β targets such as NFATs, and these molecules may act in concert to evoke the agonist-induced hypertrophic phenotype. In fact, the finding that the 8xAmyocardin transcriptional activity could still be enhanced by pro-hypertrophic agonists and that 8xDmyocardin was only partially repressed after GSK3β inhibition supports the concept that GSK3β inhibition acts through multiple signaling mechanisms, including pathways independent of myocardin phosphorylation.
Phosphorylation assays with synthetic myocardin peptides indicated that myocardin is a primed GSK3β substrate, a phenomenon described for many, if not all, GSK3β targets.10 Further experiments are needed to identify the responsible priming kinase. Although our data clearly indicate that the amino acid motifs 455 to 467 and 624 to 636 are the primary GSK3β phosphorylation sites in myocardin in vitro and within living cells, the phosphorylation-resistant 8xAmyocardin mutant showed some residual phosphorylation. This suggests that there may be additional GSK3β phosphorylation motifs in myocardin or that GSK3β activates other kinases, leading to myocardin phosphorylation.
Whereas we have not yet shown that endogenous myocardin is phosphorylated by GSK3β in intact cardiomyocytes, our results using a myocardin-based, cell-permeable peptide indicate that the mapped myocardin motif aa624–636 is recognized by and may inhibit GSK3β in intact cardiomyocytes, resulting in cardiomyocyte hypertrophy. These data suggest that the peptide Ap-myocardinp636 may behave as a substrate-analogue GSK3β inhibitor.
We fused myocardin constructs to the DNA binding domain of the yeast transcription factor GAL4, thus creating a heterologous reporter system in cardiomyocytes independent of confounding effects exerted by lithium chloride. Indeed, lithium chloride in itself did not activate GAL4-driven luciferase (in contrast to ANF promoter-driven luciferase). Results obtained with the GAL4-myocardin constructs suggest that GSK3β phosphorylation reduces the intrinsic transcriptional activity of myocardin. This could possibly be explained through an alteration of the secondary structure of the myocardin regulatory domain (harboring the GSK3β phosphorylation sites). At present, however, the exact mechanism by which GSK3β phosphorylation inhibits myocardin remains to be determined. Moreover, our data do not exclude the possibility that GSK3β phosphorylation also regulates the binding of cofactors such as serum response factor or HOP to myocardin.19 Importantly, the effects of GSK3β phosphorylation on myocardin and NFAT are clearly different: Whereas GSK3β phosphorylation determines NFAT subcellular localization, wtmyocardin, 8xAmyocardin, and 8xDmyocardin both exhibit an exclusively nuclear localization (data not shown) but differ in their transcriptional activity.
In summary, site-specific phosphorylation by GSK3β identifies a novel post-translational regulatory mechanism of myocardin and this molecular interaction may contribute to the control of cardiomyocyte hypertrophy by GSK3β and myocardin.
This study was supported by the Deutsche Forschungsgemeinschaft (Ba1668/3-3 to C.B.). We thank Heike Aranda and Sarah-Anna Jainski for expert technical assistance.
↵*Both authors contributed equally to this study.
Original received July 26, 2004; revision received November 29, 2004; resubmission received August 17, 2005; accepted August 23, 2005.
Antos CL, McKinsey TA, Frey N, Kutschke W, McAnally J, Shelton JM, Richardson JA, Hill JA, Olson EN. Activated glycogen synthase-3 beta suppresses cardiac hypertrophy in vivo. Proc Natl Acad Sci U S A. 2002; 99: 907–912.
Haq S, Choukroun G, Kang ZB, Ranu H, Matsui T, Rosenzweig A, Molkentin JD, Alessandrini A, Woodgett J, Hajjar R, Michael A, Force T. Glycogen synthase kinase-3beta is a negative regulator of cardiomyocyte hypertrophy. J Cell Biol. 2000; 151: 117–130.
Hardt SE, Sadoshima J. Glycogen synthase kinase-3beta: a novel regulator of cardiac hypertrophy and development. Circ Res. 2002; 90: 1055–1063.
Haq S, Michael A, Andreucci M, Bhattacharya K, Dotto P, Walters B, Woodgett J, Kilter H, Force T. Stabilization of beta-catenin by a Wnt-independent mechanism regulates cardiomyocyte growth. Proc Natl Acad Sci U S A. 2003; 100: 4610–4615.
Wang DZ, Li S, Hockemeyer D, Sutherland L, Wang Z, Schratt G, Richardson JA, Nordheim A, Olson EN. Potentiation of serum response factor activity by a family of myocardin-related transcription factors. Proc Natl Acad Sci U S A. 2002; 99: 14855–14860.
He TC, Zhou S, da Costa LT, Yu J, Kinzler KW, Vogelstein B. A simplified system for generating recombinant adenoviruses. Proc Natl Acad Sci U S A. 1998; 95: 2509–2514.
Rossig L, Jadidi AS, Urbich C, Badorff C, Zeiher AM, Dimmeler S. Akt-dependent phosphorylation of p21(Cip1) regulates PCNA binding and proliferation of endothelial cells. Mol Cell Biol. 2001; 21: 5644–5657.
Yasukawa H, Hoshijima M, Gu Y, Nakamura T, Pradervand S, Hanada T, Hanakawa Y, Yoshimura A, Ross J Jr, Chien KR. Suppressor of cytokine signaling-3 is a biomechanical stress-inducible gene that suppresses gp130-mediated cardiac myocyte hypertrophy and survival pathways. J Clin Invest. 2001; 108: 1459–1467.