Protein kinase-A dependent phosphorylation of transcription enhancer factor-1 represses its DNA-binding activity but enhances its gene activation ability

Abstract

The cAMP-dependent signaling pathway has been implicated in cardiac cell growth/differentiation and muscle gene transcription. Previously, we have identified a cAMP-inducible E-box/M-CAT hybrid motif in the cardiac α-myosin heavy chain (α-MHC) gene promoter. The two factors, TEF-1 and Max, that bind to this motif are found to physically associate with each other and exert a positive cooperative effect for gene regulation. Here we show that TEF-1, but not Max, is a substrate for protein kinase-A (PK-A)-dependent phosphorylation. TEF-1 is phosphorylated by PK-A at residue serine-102. This post-translational modification of TEF-1 repressed its DNA-binding activity, but not its ability to interact with the Max protein. Replacement of serine-102 in TEF-1 by a neutral or a charged amino acid did not abolish its DNA-binding ability, suggesting that changing a charge at the 102 amino-acid position of TEF-1 was not sufficient to inhibit its DNA-binding activity. We also show that PK-A response of the α-MHC gene is stimulated by the presence of wild-type TEF-1 but not by mutant TEF-1 having serine-102 replaced by alanine, suggesting that phosphorylation at this residue accounts for the cAMP/PK-A response of the gene. Thus, these data demonstrate that TEF-1 is a direct target of cAMP/PK-A signaling in cardiac myocytes.

INTRODUCTION

The cAMP-dependent signaling pathway has been shown to elicit opposite effects in two types of striated muscle cells (skeletal versus cardiac). Whereas activators of cAMP levels lead to repression of cell differentiation and muscle-specific gene expression in skeletal-muscle cells (1), they are found to promote cell growth as well as muscle gene transcription in cardiac myocytes (2). The positive regulatory role of the cAMP-dependent pathway in the differentiation and growth of cardiac myocytes is also supported by the finding of clusters of adrenergic cardiac cells capable of synthesizing and releasing catecholamines that can activate β-adrenergic receptor/cAMP signaling pathway (3,4). These intra-cardiac adrenergic cells have been identified in the embryonic heart at a developmental stage that precedes the sympathetic innervation (3). Furthermore, targeted destruction of genes encoding the enzymes of catecholamine synthesis as well as those controlling the signaling of the β-adrenergic pathway have resulted in fetal death, with severe phenotypic anomalies of the heart muscle in mid-gastrulation stage (5). These studies document that, in addition to the well-established role of the β-adrenergic/cAMP pathway to heart contractile function, this pathway is also involved in the development and/or maintenance of the cardiac myogenic program (4). In skeletal-muscle cells, the inhibitory effect of cAMP on myoblast differentiation and muscle gene transcription has been ascribed to silencing of the activation potential of the myogenic basic helix–loop–helix class (MyoD family) of regulatory factors (1). However, no target factor and/or downstream mechanism for a positive regulatory role of cAMP in cardiac myocytes have been elucidated thus far.

The myosin heavy chain (MHC) gene, which encodes contractile proteins of the thick filament of the sarcomere has served as a model system for deciphering events of cardiac muscle gene transcription. In the mammalian heart, two MHC isoforms, α- and β-, are expressed. These two isoforms are differentially regulated, which has considerable physiologic relevance. While the high-ATPase activity in α-MHC accounts for greater velocity of cardiac muscle contraction, the low ATPase activity in β-MHC leads to a greater economy of force generation (6). β-MHC is also expressed in slow-twitch skeletal muscle fibers, but a high level of α-MHC expression remains restricted to cardiac myocytes (6). In the developing cardiac myocyte, α-MHC transcripts have been seen as early as gestational day 8–9, when other cardiac markers also become detectable. Recent gene ablation studies have revealed that the expression of both alleles of the α-MHC gene is essential not only for heart contractile function, but also for normal growth of the myocardium (7). Likewise, both in animals and humans, the loss of α-MHC transcripts has been found to be associated with dysfunction of the failing heart (8).

Regulation of the cardiac α-MHC gene expression is controlled primarily at the transcriptional level. Several stimuli, such as thyroid hormone, exercise, adrenergic stimulation, cell-contractile activity and hemodynamic overload have been shown to change the α-MHC gene expression. The DNA elements that confer responsiveness of the gene to these stimuli have been identified as well (9–11). In addition, cis-regulatory DNA elements that are involved in activation of the α-MHC gene expression in cardiac muscle cell background have been delineated. These include the binding sites of myocyte-specific factor-2 (MEF-2), GATA-4, and transcription enhancer factor-1 (TEF-1). The C(Ar)G box and an A-rich element that binds a factor immunologically related to the MADS family of proteins, serum response factor (SRF), have been also implicated in cardiac-specific expression of the α-MHC gene (12–14). Furthermore, an interplay between factors such as thyroid receptor (TR) and MEF-2 have been shown to be important for the induction of the α-MHC gene expression in cardiac muscle cells (15). Other factors that have been implicated in cardiac-specific gene activation include members of the CSX/Nkx2.5 and HAND families (16,17). Although the expression of none of these factors is totally restricted to cardiac myocytes a combinatorial interaction among them has been suggested to be controlling the cardiac-specific gene regulation (16). Since these factors constitute cell-specific transcription complexes, it is also believed that they may be potential targets for signaling mechanisms augmenting the cardiac gene expression. Recently, a TEF-1 isoform, RTEF-1 has been found to be involved in the α₁-adrenergic agonist induced expression of the β-MHC gene in primary culture of cardiac myocytes (18). However, potential targets for signals generated by stimuli which influence the expression of the α-MHC gene in cardiac myocytes are not yet known.

Previously we have identified an E-box/M-CAT hybrid motif (EM) of the α-MHC gene that participates both in the basal and cAMP-induced expression of the gene (11). We have also shown that a basic helix–loop–helix leucine zipper (HLH-LZ) protein, Max, and an M-CAT binding factor, TEF-1, bind to the EM motif, and that these two factors physically interact and positively cooperate for α-MHC gene expression (19). Here we show that TEF-1, but not Max, is a substrate for PK-A-dependent phosphorylation. TEF-1 is phosphorylated by PK-A at serine-102 which leads to a decrease in its DNA-binding activity; however, it does not change its ability to bind the Max protein. We also demonstrate that PK-A phosphorylation of TEF-1 potentiates its ability to activate α-MHC gene expression. These results demonstrate a possible mechanism of the cAMP/PK-A-mediated activation of cardiac muscle gene expression.

MATERIALS AND METHODS

Construction of plasmids

The bacterial expression plasmid pGEX-KG was utilized to direct overexpression of full-length glutathione-S-transferase (GST)-Max, or GST-TEF-1 (19). In order to generate GST-TEF-1 deletion mutants, a forward primer containing XbaI site: 5′-AATCTAGAGATTGAGCCCAGCAGCTGGAGCGGC-3′, and a reverse primer containing XhoI site: 5′-TTTCTCG AGTCACCAGGCAGGAACTGAGGGGGC-3′ for GST-TEF(1–210) and 5′-TTTCTCGAGTCACAGGGCCTTGTCCTTGGCAGT-3′ for GST-TEF(1–123) constructs were utilized. For GST-TEF(114–430) a forward primer, 5′-TCTAGAGATGGATCAGACTGCCAAGGAC-3′ and reverse primer, 5′-GGCCGGCTCGAGTCAGTCCTTCACAAGCCTGTAGATATGGTG-3′ were used. The forward and reverse primers were annealed with TEF-1 cDNA template and the desired fragment of TEF-1 was amplified by PCR using Promega cDNA amlification Kit (Promega Inc., Madison, WI). The resulting DNA fragment was digested with XbaI and XhoI enzymes and subcloned into the XbaI and XhoI sites of the pGEX-KG vector. The GST-TEF(1–133mt1) and GST-TEF(1–113mt2) constructs were synthesized using the same forward primer as for GST-TEF(1–210) and reverse primers starting at amino-acid position 113 of TEF-1 in which codon TCT for serine-102 was mutated to GCT and GAT, respectively (as shown in Fig. 3). The cloning ends of the constructs were confirmed by the dideoxy sequencing method. The plasmid pMP.EM.CAT contains a –256 to +30 bp fragment of the α-MHC gene linked immediately upstream to the CAT reporter gene in the promoter-less pGCATC vector. The CMVTEF-1 expression plasmid, kindly furnished by Paul Simpson (University of California, San Francisco, CA), contains rat TEF-1 cDNA under the control of the cytomegalovirus (CMV) promoter. The pBS.Max expression plasmid has been described elsewhere (20). The CMV.PK-A expression plasmid containing cDNA of the catalytic subunit of PK-A was purchased from Stratagene, Inc. (La Jolla, CA). The Flag-tagged-TEF-1 and Max expression plasmids were constructed by subcloning of the full-length cDNA of TEF-1 or Max into the pCMV vector as described previously (19). The point mutation in the plasmids was created by utilizing the Stratagene site-directed mutagenesis kit (Stratagene).

Figure 3.

Serine-102 of TEF-1 is phosphorylated by PK-A. (A) Schematic diagram showing wild-type (wt) GST-TEF(1–113) peptide and two other mutants in which serine-102 was substituted by either alanine or aspartic acid. (B, C) Proteins were subjected to PK-A phosphorylation using [γ-³²P]ATP and analyzed by SDS–PAGE followed by Coomassie blue staining (B) and autoradiography (C) of the same gel. (D) EMSA was carried out using an M-CAT oligo as a labeled probe and increasing amounts (from 1 to 3 µg) of the different proteins as indicated at the top of the gel. For the competition assay, a 100-fold excess of unlabeled M-CAT oligo was included in the EMSA-binding reaction.

Overexpression and purification of GST fusion proteins

The GST fusion proteins were expressed in bacteria and purified as described previously (19). Briefly, bacteria harboring plasmids GST-Max or GST-TEF-1 were grown overnight in LB-ampicillin medium. The next morning, cells were diluted 1:10 with fresh medium and induced with 0.1 mM IPTG to direct expression of fusion proteins. After 3–7 h of expression, cells were pelleted by centrifugation, resuspended in phosphate-buffered saline (PBS) containing protease inhibitors and sonicated to break open cell walls. The bacterial lysate was solubilized by addition of Triton X-100 to a final concentration of 1% and centrifuged to remove the insoluble material. The soluble fraction was incubated with glutathione–agarose beads for 30 min at 4°C. Beads were pelleted and GST fusion proteins bound to the glutathione–agarose beads were washed with PBS containing 0.1% Triton X-100 and analyzed by SDS–PAGE and Coomassie blue staining. Phosphorylation of proteins was done in 10 µl of a kinase buffer containing 17.5 mM Tris–HCl, pH 7.5, 2.5 mM MgCl₂, 40 mM ATP, plus 10 µCi [γ-³²P]ATP and 1–2 µg of GST-TEF-1. The reaction was started by addition of 10 U (unless indicated otherwise) of the catalytic subunit of PK-A (Sigma Inc., St Louis, MO) and the reaction mixture was incubated at 37°C for 15 min.

Cell culture and transfection

Primary myocytes were cultured from 18-day-old fetal rat hearts in Ham’s F-12 medium (Gibco BRL Inc., Gaithersburg, MD) with 5% calf serum (11). Cultures generally consisted of >90% myocytes, as measured by immunocytofluorescence with anti-myosin antibody. Sol8 cells were grown in growth-medium containing Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 20% fetal bovine serum (FBS) in an atmosphere of 5% CO₂. Myogenic differentiation of cells was induced by exposure of confluent cultures to differentiation medium containing DMEM supplemented with 5% horse serum. All culture media contained penicillin (5 mg/ml), streptomycin (5 mg/ml) and neomycin (100 mg/ml).

Primary cultures of cardiac myocytes were transfected after 48 h in culture with 10 µg of DNA/plate by use of a lipotaxi reagent (Stratagene) according to the manufacturer’s protocol.

Preparation of nuclear extract and electromobility gel shift assay

Nuclear extract was prepared from neonatal rat hearts by the method of Dignam et al. (21), with slight modifications as described previously (11). For the electromobility gel shift assay (EMSA), double-stranded oligos were 5′-end-labeled with T4 polynucleotide kinase and [γ-³²P]ATP. The analytical binding reaction was carried out in a total volume of 25 µl containing ~10 000 c.p.m. (0.1–0.5 ng) of the labeled DNA probe, 2–5 µg of the nuclear extracts, and 1 µg of poly(dI–dC) (Sigma) as a non-specific competitor. The binding buffer consisted of 10 mM Tris–HCl pH 7.4, 100 mM NaCl, 0.1 mM EGTA [ethylene glycol-bis (β-aminoethylether)-N,N,N′,N′-tetraacetic acid], 0.5 mM dithiothreitol, 0.3 mM MgCl₂, 8% glycerol and 0.5 mM phenylmethylsulfonyl fluoride (PMSF). After incubation at room temperature for 20 min, the reaction mixtures were loaded on 5% native polyacrylamide gels, and electrophoresis was carried out at 150 V in a 0.5× TBE buffer in a cold room. Sense-strand sequences of double-stranded oligo probes used in this study are: E-box, 5′-GGGGCACGTGCC-3′, α-MHC EM, 5′-CAGCAGGCACGTGGAAGAGC-3′; and cardiac troponin-T M-CAT, 5′-AGTGTTGCATTCCTCTCTGG-3′.

In vivo phosphorylation of TEF-1

After 1 day in differentiation medium, Sol8 cells were transfected with pCMV expression plasmids encoding either the catalytic subunit of PK-A, Flag.TEF-1 or Flag.Max proteins. On the second day after transfection, cells were rinsed with phosphate-deficient DMEM (D-DMEM) and then incubated for 18 h at 37°C with D-DMEM containing 2% dialyzed FBS and 1 mCi/ml of ³²P (Amersham, Corp., Arlington Heights, IL). After labeling, cells were washed twice with PBS, harvested, and lysed in 0.6 ml of ice-cold radioimmunoprecipitation assay (RIPA) buffer (20 mM Tris–HCl pH 7.5, 1% vol/vol NP-40, 0.5% wt/vol deoxycholate, 1% SDS, 1 mM PMSF, 1% vol/vol aproteinin, 25 mM benzamidine, 2 mM Na₃VO₄, 50 mM NaF, 40 mM β-glycerophosphate and 25 mM pyrophosphate). The lysate was twice frozen and thawed, and centrifuged at 12 000 g for 15 min. The supernatant was transferred to a new tube and pre-cleaned with 100 µl of 10% protein-A–Sepharose beads which were washed twice in 500 µl of RIPA buffer. After 1 h incubation, the supernatant was transferred to a new tube. For immunoprecipitation of labeled Flag-tagged fusion proteins, the cell extract was incubated with 2 µl M-2 (anti-Flag) antibody conjugated to agarose beads (1.8 mg antibody/ml) in a total volume of 1 ml overnight with continuous rocking. The next morning, beads were pelleted and washed five times in the RIPA buffer, and the resultant proteins were suspended in 2× Laemlli buffer and subsequently resolved by SDS–PAGE, transferred to polyvinylidene difluoride (PVDF) membrane, and analyzed by autoradiography. Western-blot analysis was carried out using the ECL protein detection kit (Amersham) as described previously (19).

Characterization of TEF-1 binding to Max

The TNT-coupled rabbit reticulocyte lysate system (Promega), was used for translation of pBS-Max, pBS-MyoD and pBS-myogenin plasmids. After translation, the specific incorporation of [³⁵S]methionine into proteins was determined by TCA precipitation and the integrity of the translated proteins was checked by SDS–PAGE and autoradiography. For in vitro binding assay, ³⁵S-labeled proteins were incubated with 2–3 µg of GST or GST-fusion proteins on glutathione–agarose beads in 1× protein interaction buffer (PIB) (20 mM HEPES pH 7.5, 75 mM KCl, 1 mM EDTA, 2 mM MgCl₂, 2 mM DTT and 0.5% NP-40) for 2 h at 4°C with continuous rocking. The beads were pelleted and washed five times with 1× PIB. The bound proteins were eluted with Laemlli’s sample buffer and analyzed on SDS–PAGE.

RESULTS

PK-A phosphorylation of TEF-1 represses its DNA-binding, but not its protein-interaction ability

Regulation of gene transcription by cAMP is frequently mediated through the activation of cAMP-dependent protein kinase-A (PK-A), where binding of cAMP to the regulatory subunit of the holoenzyme causes release and activation of the catalytic subunit (22). We have previously identified two factors, Max and TEF-1, that bind to the cAMP-inducible EM motif of the α-MHC gene (19). As an initial approach to exploring the mechanism of cAMP-dependent activation, we examined whether either of the factors i.e., TEF-1 and/or Max, could be a substrate for cAMP-dependent phosphorylation. Thus, GST recombinant Max or TEF-1 proteins were incubated with the catalytic subunit of PK-A as well as [³²P]ATP and analyzed on SDS–PAGE. The results showed that GST-TEF-1 is highly phosphorylated but GST-Max or GST alone are not (Fig. 1A). The two bands observed with GST-TEF-1 exhibited an apparent 81 kDa recombinant protein that includes ∼55 kDa TEF-1 and 26 kDa of the GST moiety. The lower band observed with GST-TEF-1 arises from proteolysis, and their relative amounts varied from batch to batch preparation of the fusion protein.

Figure 1.

PK-A phosphorylation of TEF-1 inhibits its DNA-binding activity, but not its ability to interact with Max. (A) GST, GST-Max and GST-TEF-1 proteins (1 µg of each) were incubated with 10 U of the PK-A catalytic subunit in a kinase reaction buffer containing [γ-³²P]ATP. Phosphorylation of the protein was visualized by SDS–PAGE followed by autoradiography. (B) EMSA was carried out using a ³²P-labeled EM probe (as described in Materials and Methods) and GST-TEF-1 incubated with varying concentrations (10, 20 and 50 U) of the PK-A catalytic subunit and cold ATP. (C) In vitro translated [³⁵S]methionine-labeled Max was incubated with glutathione–agarose beads bound to GST, GST-Max, GST-TEF-1 or PK-A-phosphorylated-GST-TEF-1 proteins. Proteins bound to beads were analyzed on SDS–polyacrylamide gels in the lanes indicated.

Because post-translational modification is known to change the biochemical characteristics of a protein, we next examined the effect of TEF-1 phosphorylation on its DNA-binding as well as its protein interaction ability. An EMSA was carried out in which a labeled EM probe was incubated with GST-TEF-1 that was subjected to phosphorylation using cold ATP and varying concentrations of the catalytic subunit of PK-A. A parallel reaction was also carried out by use of [³²P]ATP to ensure phosphorylation of the protein. The GST-TEF-1 protein that was incubated with PK-A in the absence of ATP was used as a negative control. Figure 1B shows that, as the concentration of PK-A in the reaction buffer was increased, phosphorylation of TEF-1 gradually reduced the DNA-binding activity of the protein. At higher concentrations of PK-A, when the majority of GST-TEF-1 molecules were phosphorylated, virtually no DNA-binding activity was observed. On the other hand, in an in vitro protein–protein interaction assay, the PK-A-phosphorylated-GST-TEF-1 (Fig. 1C, lane 4) that failed to bind DNA did retain the in vitro synthesized Max protein as efficiently as its non-phosphorylated counterpart (Fig. 1C, lane 3). Max is known to form a homodimer (19); therefore, in this assay a pull-down of [³⁵S]methionine-labeled Max by GST-Max but not by GST protein was used as positive and negative controls, respectively. These data documented that the phosphorylation of GST-TEF-1 by PK-A inhibits its DNA-binding activity, but not its ability to interact with the Max protein.

PK-A phosphorylation site of the TEF-1 protein

To identify the site for PK-A dependent phosphorylation of TEF-1, we generated three different deletion mutants in which a portion of the C- or N-terminal region of TEF-1 has been omitted (Fig. 2A). Each truncated TEF-1 peptide was synthesized as a GST recombinant protein and analyzed for PK-A phosphorylation ability. As shown in Figure 2C, equal phosphorylation by PK-A was observed in a full-length GST-TEF-1 as well as in two other mutants [GST-TEF(1–210) and GST-TEF(1–123)] in which the C-terminal region was deleted up to amino acid (aa) 123. However, the other construct that lacked the N-terminal 113 aa of TEF-1, GST-TEF(114–430), failed to incorporate ³²P under identical conditions. In order to see whether the inability of a GST fusion protein for phosphorylation was due to the degradation of the protein, the same gel was stained with Coomassie blue dye (Fig. 2B). As shown in Figure 2B, GST-TEF(114–430) that did not undergo phosphorylation was indeed synthesized as a full-length fusion peptide with an expected molecular weight of ∼60 kDa, thus documenting that the PK-A site of TEF-1 must be located within the first 113 aa of the protein. Deletion of the N-terminal 113 aa of TEF-1 removed the TEA/ATTS domain, which has been previously shown to be essential to the DNA-binding activity of the protein (23). To validate the DNA-binding characteristics of the GST-fusion proteins used in this assay, we also carried out an EMSA using the M-CAT oligo that has a high-affinity TEF-1 binding site, as a labeled probe. As shown in Figure 2D, a fusion protein in which the TEA/ATTS domain was omitted [GST-TEF(114–430)] failed to bind to the DNA probe, whereas two other fusion peptides that contained this region did bind successfully to DNA. These data confirm the importance of the TEA/ATTS domain for protein–DNA interaction, and validate the different truncated versions of the TEF-1 protein used in this study.

Figure 2.

The PK-A phosphorylation site of TEF-1 is located within the first 113 aa of the protein. (A) Schematic diagram of TEF-1 deletion mutants and summary of their PK-A phosphorylation ability. (B, C) Different GST-TEF-1 proteins were subjected to PK-A phosphorylation using [γ-³²P]ATP. Proteins were resolved by SDS–PAGE and visualized by Coomassie blue staining (B) and autoradiography (C) of the same gel. (D) EMSA was carried out with an M-CAT oligo as a labeled probe and increasing amounts (from 1 to 3 µg) of different GST-TEF-1 proteins as indicated at the top of the gel.

To delineate the exact phosphoamino acid that served as a substrate for PK-A phosphorylation of TEF-1, another construct was generated that synthesized a GST fusion protein having the N-terminal 113 aa of TEF-1, GST-TEF(1–113) (Fig. 3A). Upon incubation of this recombinant protein with PK-A, a single band of ∼40 kDa was observed which is the expected molecular weight of the GST-TEF(1–113) fusion peptide (Fig. 3C). Comparison of bands between the Coomassie blue-stained gel and the ³²P-incorporated gel autoradiogram (Figs 2 and 3) provided evidence for a possible region for a PK-A site in the TEF-1 protein. Although, following preparation of the GST-TEF(1–113) fusion protein, different sizes of truncated peptides were made, none of the bands that were less than 40 kDa became phosphorylated (Fig. 3B and C). This indicated that the PK-A site of TEF-1 must be at the C-terminal end of the 1–113 aa segment of the protein. Inspection of the amino acids of this region of TEF-1 revealed that serine-102 was contained within a region that resembled the consensus sequence for PK-A phosphorylation (RRXS). To determine that this serine represented the actual phosphoamino acid, we mutated the codon of serine-102 to encode either alanine or aspartic acid [GST-TEF(1–113mt-1) and GST-TEF(1–113mt-2), respectively]. Each mutated peptide was synthesized as a GST-fusion protein and tested for its PK-A phosphorylation ability. As shown in Figure 3C, these mutations completely abolished the PK-A phosphorylation of GST-TEF(1–113), thus confirming that serine-102 is indeed a single PK-A site of TEF-1. Because phosphorylation of GST-TEF(1–113) inhibits the DNA-binding activity of the protein (data not shown), we were interested in finding out whether changing serine-102 to aspartic acid that mimics a negative charge on the protein, or to alanine that adds on to a neutral amino acid, would have an effect on the DNA-binding activity of the protein. The results showed that both GST-TEF-1 mutants (mt-1 and mt-2) retained their ability to bind to the M-CAT DNA binding site, albeit with a lesser affinity than that of the wild-type protein (Fig. 3D). These data suggest that (i) serine-102 is important, but not critical, for TEF-1 DNA-binding activity, and (ii) mere addition of a negative charge on the protein is not sufficient to result in loss of DNA-binding ability. Rather, phosphorylation of serine-102 may have additional effects on the protein-configuration leading to inhibition of the TEF-1 DNA-binding activity.

In vivo phosphorylation of TEF-1 by PK-A

Because specific antibodies against different TEF-1 isoforms are not available, we transfected cells with a pCMVFlag.TEF-1 expression plasmid encoding the full-length Flag.TEF-1 fusion protein in the presence or absence of PK-A expression vectors to determine in vivo TEF-1 phosphorylation by PK-A. In a parallel set of plates cells were also transfected with a plasmid expressing the Flag.Max fusion protein for serving as a negative control. After labeling of cells with ³²P, Flag recombinant proteins were immunoprecipitated from the nuclear extract with an anti-Flag antibody and analyzed by SDS–PAGE and subsequently by western blot analysis. In the presence of PK-A, two major peptides of Mr ∼35 and 55 kDa were found to be highly phosphorylated as compared to controls that received no PK-A. However, no phosphorylation of the Flag.Max protein was seen (Fig. 4A). To validate these recombinant proteins further, the same membrane was probed with an anti.Flag antibody and subjected to a western blot analysis after 1 week when the ³²P activity of membrane was diminished. As shown in Figure 4B, both 35 and 55 kDa phosphorylated bands were indeed Flag.TEF-1 fusion proteins, and Flag.Max was synthesized as a full-length ∼22 kDa protein. We also tested whether in vivo PK-A phosphorylation of the endogenous TEF-1 protein could modulate its DNA-binding activity. An EMSA was performed with the TEF-1-binding oligo (M-CAT or EM) as a labeled probe and nuclear extract obtained from cells transfected with or without PK-A. As shown in Figure 4C, the intensity of the TEF-1/M-CAT complex was markedly reduced due to transfection of cells with the PK-A expression plasmid. Similar results were obtained using nuclear extracts from cells treated with forskolin, an adenyl-cyclase activator. However, there was no difference in the intensity of the E-box/protein complexes generated from the same nuclear extracts. These results clearly demonstrate in vivo phosphorylation of the TEF-1 protein by PK-A, leading to repression of its DNA-binding activity.

Figure 4.

Phosphorylation of TEF-1 by PK-A in vivo. (A) Sol8 cells in growth medium were transiently transfected with pCMVFlag.TEF-1 (lanes 1 and 2) or pCMVFlag.Max (lane 3) with and without PK-A expression plasmid. Eighteen hours after transfection, cells were transferred to differentiation medium, after which they were labeled with ³²P (1 mCi/ml) for another 18 h. Cells were harvested, and Flag. recombinant proteins were immunoprecipitated from the cell lysate and resolved by SDS–PAGE. Proteins were transferred to PVDF membrane and visualized by autoradiography. (B) After 1 week, when ³²P activity of the membrane was reduced, it was probed with anti-Flag antibody and subsequently analyzed by western blot analysis. (C, D, E) EMSA was performed using different probes (as described in Materials and Methods) and nuclear proteins (2, 4 and 8 µg) obtained from cells transfected with or without PK-A expression plasmid, or treated with 10 µM forskolin.

Functional significance of TEF-1 PK-A phosphorylation in α-MHC gene regulation

To examine the functional significance of TEF-1 phosphorylation, we cotransfected cardiac myocytes with expression plasmids encoding the catalytic subunit of PK-A, wild-type TEF-1 or full-length TEF-1(mt-1) with the serine-102 to alanine mutation, and with a reporter plasmid (pMHC/CAT). Overexpression of full-length TEF-1 resulted in a concentration-dependent biphasic effect on the expression of the reporter plasmid (Fig. 5A). A modest (50%) activation at low concentration was followed by repression (70%) at higher concentrations of the expression plasmid. As shown in Figure 5A, a similar concentration-dependent effect was also observed by overexpression of the mutant TEF-1(mt-1) protein, indicating that mutation of serine-102 does not abrogate its trans-regulatory effect. As shown in Figure 5B, overexpression of the catalytic subunit of PK-A activated the α-MHC/CAT gene expression almost 5-fold in a linear concentration-dependent manner, which is consistent with data obtained by treatment of cardiac myocytes with 8BrcAMP (11). To determine whether TEF-1 is involved in the PK-A response of the α-MHC gene, when cells were cotransfected with the PK-A and TEF-1 expression plasmids together, we found a further activation (3-fold) of the reporter gene expression. However, no such activation of the reporter gene was noted when PK-A was combined with the TEF-1(mt-1) plasmid (Fig. 5B). In a separate set of experiments we also determined whether sufficient amount of TEF-1(mt-1) protein was synthesized in cardiac myocytes. For this experiment cells were transfected with plasmids encoding Flag-tagged either TEF-1 or TEF-1(mt-1) proteins and expression of proteins was determined by a western blot analysis using anti-Flag antibodies. Results indicated that, as with wild-type TEF-1, a significant amount of TEF(mt-1) protein was also synthesized in cardiac myocytes (data not shown). Thus, together these results suggest a role of TEF-1 serine-102 phosphorylation in the PK-A dependent activation of the α-MHC gene expression.

Figure 5.

Involvement of TEF-1 in PK-A dependent activation of α-MHC gene expression. Primary cultures of fetal rat cardiac myocytes were cotransfected with 5 µg of the α-MHC/CAT reporter plasmid and varying amounts of expression vectors as shown in each panel. (A) Effect of different concentrations of pCMVTEF-1 and pCMVTEFmt-1 vectors on the expression of MHC/CAT reporter construct. (B) Effect of varying concentrations of PK-A expression vector alone, and when combined with TEF-1 expression vectors. CAT activity (mean ± S.E.) was normalized for β-galactosidase activity and expressed relative to the activity of pMHC/CAT in the vehicle-treated ventricular cells. Results are derived from 10–15 different cultures in which different plasmid preparations were used.

Serine-102 mutation does not abrogate ability of TEF-1 to bind to Max

To examine whether serine-102 participates in the TEF-1/Max interaction, we next examined the binding ability of TEF-1(mt-1) with Max protein. An in vitro translation reaction was programmed in the presence of [³⁵S]methionine with a plasmid encoding a full-length Max. After translation, glutathione–agarose beads that were bound with GST or GST-TEF-1 fusion proteins were incubated with the translated proteins. The beads were removed, washed extensively, and analyzed by SDS–PAGE. As shown in Figure 6, [³⁵S]methionine-labeled Max was retained on the beads that contained GST-TEF-1, GST-TEF-1(mt-1) and GST-TEF-1(mt-2) in which serine-102 was replaced by alanine and aspartic acid, respectively, but not on the beads with GST alone, thus confirming that serine-102 is not crucial for the physical interaction between Max and TEF-1. To map the region of TEF-1 that was required for binding to Max, we also analyzed several TEF-1 deletion mutants for their ability to interact with in vitro translated Max (Fig. 6). Each TEF-1 mutant was synthesized as a GST recombinant protein, and only those preparations that showed >50% of the expected size of GST fusion protein molecules were used for further study. As shown in Figure 6, GST-TEF-1 mutants in which the C-terminal region was deleted successfully retained translated Max in vitro. However, when the 113 aa N-terminal region of TEF-1 was omitted, Max binding was not observed. To narrow down the Max binding region of TEF-1 further, we generated another deletion mutant in which the first N-terminal 27 aa were deleted. This GST-TEF(27–113) mutant was also unable to retain in vitro translated Max. Together, the data indicate that the N-terminal first 27 aa segment of TEF-1, which comprises a highly acidic domain of the TEF-1 protein, is required for TEF-1/Max association. Since Max contains a b-HLH-LZ configuration, we were also interested to see whether TEF-1 could also bind to b-HLH myogenic factors; however, our results showed that GST-TEF-1 beads could not retain in vitro synthesized MyoD or myogenin, which indicated that under these assay conditions, TEF-1 protein interacted specifically with Max.

Figure 6.

Identification of TEF-1 interactive domain with Max. (A) Schematic diagram of the different GST-TEF-1 mutants and a summary of their observed binding activity with Max. In the GST-TEF(1–113mt-1) and GST-TEF(1–113mt-2) peptides, serine-102 was changed to alanine and aspartic acid, respectively. (B) In the rabbit reticulocyte lysate pBS-MyoD, -myogenin and -Max plasmids were transcribed and translated with [³⁵S]methionine. (C, D) The ³⁵S-labeled Max (C), or MyoD (D) were incubated with GST or GST-TEF-1 fusion peptides on beads as indicated above each lane, and proteins bound to beads were analyzed on SDS–polyacrylamide gels.

DISCUSSION

The cAMP-dependent signaling pathway has long been documented to be involved in cell growth, differentiation and cell-specific gene expression (24). Furthermore, in two types of striated muscle cells, activation of this pathway has been shown to have opposite effects. It inhibits cell growth/differentiation and muscle gene expression in skeletal muscle cells. In contrast, it stimulates these events in cardiac myocytes (1–3). We had previously demonstrated that agents which elevate the intracellular level of cAMP activate cardiac α-MHC gene expression, involving an EM DNA motif (11). This motif was also noted to play a role in the contractile activity-mediated activation of α-MHC gene expression in cardiac myocytes (10). In addition, we have shown that two factors, Max and TEF-1 associate physically with each other as they bind to the EM motif and have a positive cooperative effect on α-MHC gene regulation (19). In this report, we show that TEF-1 protein is a substrate for cAMP/PK-A dependent phosphorylation. TEF-1 is phosphorylated by PK-A at serine-102 leading to repression of TEF-1 DNA-binding activity; however, it interacts with Max as efficiently as does its non-phosphorylated counterpart. Moreover, in this study, we have presented evidence showing that phosphorylation of serine-102 of TEF-1 accounts for PK-A induced expression of the cardiac α-MHC gene.

TEF-1 is a prototype member of a new family of transcription enhancer factors (TEFs) delineated by a highly conserved homologous DNA-binding domain, TEA/ATTS (reviewed in 25). In humans, at least four different TEF genes have been identified which encode hTEF-1, hTEF-3, hTEF-4 and hTEF-5 isoforms (25). Homologs of these isoforms have also been isolated from other species such as mice (26) and chicks (27). Although the structural configuration of the TEA domain has not yet been established, the amino acids of this region have been predicted to form a configuration of three α-helices, or one α-helix and two β-sheets (23,25). Through mutagenesis, amino acids of helix 1 and helix 3 have been documented to be essential for protein–DNA contact; however, the role of the helix 2 amino acids remains unknown (23). In the present study, by using deletion and amino-acid substitution analysis, we have shown that TEF-1 contains a single PK-A site (RRXS) that is located immediately downstream of helix 3 of the DNA-binding domain of the protein (Fig. 7). Phosphorylation of serine-102 of this site by PK-A repressed the DNA binding activity of TEF-1 to the cognate DNA binding site. There are several examples in which phosphorylation has been shown to modify the DNA-binding activity of the factor either positively or negatively (28 and references therein). When a phosphorylation site is located within or near the DNA-binding domain of the protein, phosphorylation of the protein often results in the repression of DNA-binding activity. As is the case with TEF-1, the hepatocyte nuclear factor HNF4 and the nuclear receptor NGFIβ also contain phosphorylation sites located within or near the DNA-binding domain of the protein. In both instances PK-A phosphorylation exerts a negative effect on the DNA-binding activity of the protein (28,29). Previously, some evidence had suggested that amino acids of the C-terminal region (STY-rich) of TEF-1 could also positively regulate its DNA-binding activity (23), thus raising the possibility that PK-A phosphorylation could have antagonized the influence of the C-terminal region, leading to repression of TEF-1 DNA binding activity. This possibility, however, seems unlikely, because the full-length TEF-1 as well as the truncated version of TEF-1 that lacks most of the C-region responded identically to PK-A phosphorylation. Thus, the negative effect of phosphorylation on DNA-binding activity seems to be more of a local effect, rather than involving a change in the effect of the C-region of the protein.

Figure 7.

The PK-A site (RRXS) of TEF-1 is not conserved among different TEF isoforms. Top panel: schematic diagram showing location of different regions of TEF-1. Lower panel: amino acids of the third helix of the TEA/ATTS DNA-binding domain and the PK-A site (underlined) of TEF-1 is shown and compared with other TEF isoforms. Four different human TEF isoforms, as a prototype for each group, are shown in bold letters.

Phosphorylation of a factor could interfere with its DNA-binding activity by exerting electrostatic repulsion between the phosphate groups of the protein and of the DNA. Alternatively, phosphorylation could inhibit contact between the protein and the DNA by steric hindrance. In order to see whether phosphorylation of the protein could have exerted electrostatic repulsion between the DNA and the protein due to the addition of a negative charge, we substituted serine-102 with aspartic acid so as to mimic a negative charge on the TEF-1 protein. The DNA-binding activity was then analyzed by mobility gel-shift analysis, and the results revealed that this mutant, in fact, retained its DNA-binding activity. These findings indicate that the presence of a negative charge downstream of helix 3 was not sufficient to interfere with the ability of TEF-1 to bind to DNA. Furthermore, substitution of serine-102 of the PK-A phosphorylation site with either a neutral or an acidic amino acid had exactly the same effect: both mutant TEF-1 proteins showed the same DNA-binding activity to the M-CAT element, but were insensitive to PK-A action. Thus, these results indicate that the addition of a phosphate group could have repressed the DNA-binding activity more by a steric hindrance and/or conformational modification than by electrostatic repulsion between DNA and protein molecules.

Whatever the mechanism by which phosphorylation of serine-102 of a PK-A site regulates the DNA-binding affinity of TEF-1, this phenomenon does not appear to be identical among different TEF isoforms. As shown in Figure 7, the PK-A phosphorylation site near the third helix of the TEA domain is not conserved among different TEF isoforms, being clearly absent in hTEF-3 and hTEF-5 but present in the hTEF-1 and hTEF-4 isoforms. A similar variation of the PK-A site is also seen among different interspecies of the TEF homolog. Previously, TEF isoforms have been shown to be phosphorylated in vivo (30), and based on sequence conservation, it has been suggested that they are targets of many intracellular signaling kinases such as PK-C, ERK/MAPK and casein kinase-II (18,31). Thus, it appears that TEF isoforms could integrate different types of extra- and intracellular cues mediated by signaling pathways that could then regulate the expression of different genes during developmental and/or various pathophysiologic situations. Consistent with this hypothesis are recent observations that, unlike the role of TEF-1 in cAMP-induced expression of the α-MHC gene, the α₁-adrenergic/PK-C-dependent activation of the β-MHC as well as skeletal α-actin genes is controlled by another TEF isoform, RTEF-1 (18,31). As shown in Figure 7, the RTEF-1 isoform does not contain a PK-A phosphorylation site similar to TEF-1. Rather, it contains a PK-C site towards the C-terminal region of the RTEF-1 which is responsive to α₁-adrenergic stimulation (31). Thus, it appears that in order to mediate signals through the M-CAT sites of different cardiac genes, these two isoforms (TEF-1 and RTEF-1) could be subject to different post-translational mechanisms.

Results obtained from transfection experiments have demonstrated that phosphorylation of serine-102 of TEF-1 plays a role in cAMP/PK-A-induced expression of the α-MHC gene. This observation argues that, though PK-A phosphorylation of TEF-1 inhibits its DNA-binding activity, it enhances its gene activation potential. Previously, in order to explain participation of TEF-1 from the SV40 late promoter GT-IIc and SPH sites two different modes of action have been described, an activator mode and a repressor mode. In the activator mode, binding of TEF-1 to its motif in the SV40 late promoter activates transcription, perhaps by making contacts with other cofactors and components of the pre-initiation complex (32,33). In contrast, in the repressor mode, binding of TEF-1 to its site in the late promoter represses transcription, and coactivators such as T-antigen, by interacting with TEF-1, prevent its binding to DNA, however, that activates gene transcription (34,35). There are other examples where the squelching effect of TEF-1 has been implicated in its repression effect. This negative phenotype of TEF-1 could be alleviated by overexpression of subunits of TATA-box binding proteins, indicating that inhibition of transcription is likely due to disruption of pre-initiation complex formation (36). For regulation of α-MHC gene expression we have shown before that TEF-1 elicits a concentration-dependent effect at the EM site, a modest activation (50%) at very low concentration is followed by repression at higher concentrations of TEF-1. This repression effect of TEF-1 at the EM site can be alleviated by overexpression of Max, perhaps by antagonizing the squelching effect of TEF-1 (19). From the data presented here it appears that the activation of α-MHC gene transcription by PK-A phosphorylation of TEF-1 also results from removal of its repression effect at the EM site. Although the exact mechanism of gene regulation by PK-A-phosphorylated-TEF-1 remains yet to be understood at least three possibilities could be envisioned. First, a weak or transient binding of PK-A-phosphorylated-TEF-1 to its site may allow more rapid transcriptional initiation to take place. Previously, a difference in the affinity of TEF-1 to its binding sites, GT-IIc and SPH, has been implicated in its trans-activation effect from the SV40 late promoter (32,34). Second, displacement of TEF-1 from the EM site may allow its DNA-binding domain to become accessible for binding to other gene activators. In this scenario, as PK-A-phosphorylated-TEF-1 can still bind to Max, it could remain in contact with the α-MHC EM site and form a higher order protein complex that lead to gene activation. This situation would be analogous to the interaction of T-antigen with TEF-1, where TEF-1 DNA-binding activity is inhibited but its gene activation potential is enhanced. Third, once TEF-1 is displaced from the DNA-binding site, Max protein may recruit other activator(s) at the EM site so as to create a stable protein complex that may activate gene transcription.

In summary, the results of this study identify TEF-1 as a target of the cAMP-dependent signaling pathway for the activation of cardiac-muscle gene expression. It is worth emphasizing that the cAMP response of all the muscle-specific genes studied to date is controlled by changes in the activation potential of muscle-specific regulators, such as MyoD and myogenin in skeletal muscle cells, and TEF-1 in cardiac myocytes. Participation of TEF-1 in the cAMP/PK-A response of the α-MHC gene also suggests a role of this signaling pathway in the development and/or maintenance of the cardiac myogenic program.