Serum response factor mediates AP-1-dependent induction of the skeletal alpha-actin promoter in ventricular myocytes.

"Fetal" gene transcription, including activation of the skeletal alpha-actin (SkA) promoter, is provoked in cardiac myocytes by mechanical stress and trophic ligands. Induction of the promoter by transforming growth factor beta or norepinephrine requires serum response factor (SRF) and TEF-1; expression is inhibited by YY1. We and others postulated that immediate-early transcription factors might couple trophic signals to this fetal program. However, multiple Fos/Jun proteins exist, and the exact relationship between control by Fos/Jun versus SRF, TEF-1, and YY1 is unexplained. We therefore cotransfected ventricular myocytes with Fos, Jun, or JunB, and SkA reporter genes. SkA transcription was augmented by Jun, Fos/Jun, Fos/JunB, and Jun/JunB; Fos and JunB alone were neutral or inhibitory. Mutation of the SRF site, SRE1, impaired activation by Jun; YY1, TEF-1, and Sp1 sites were dispensable. SRE1 conferred Jun activation to a heterologous promoter, as did the c-fos SRE. Deletions of DNA binding, dimerization, or trans-activation domains of Jun and SRF abolished activation by Jun and synergy with SRF. Neither direct binding of Fos/Jun to SREs, nor physical interaction between Fos/Jun and SRF, was detected in mobility-shift assays. Thus, AP-1 factors activate a hypertrophy-associated gene via SRF, without detectable binding to the promoter or to SRF.

uretic factor (ANF)) that are highly expressed in embryonic but not normal adult ventricular myocardium, a phenomenon referred to as reinduction of a "fetal" phenotype (1)(2)(3). SkA expression has been associated with increased contractility in the rodent heart (4) and with impaired contractility in patients with heart failure (5). In addition to this set of genes, two others are induced, and to varying degrees have been implicated, in this response to mechanical load. First, "immediateearly" transcription factors including Fos and Jun potentially might couple trophic signals to long-term changes in growth and gene expression. Second, myocardial growth factors including angiotensin II and transforming growth factor ␤ (TGF-␤) evoke most aspects of the fetal/hypertrophic program in cultured cardiac muscle cells (1)(2)(3). Induction of TGF-␤ by load, passive stretch, ␣ 1 -adrenergic agonists, and angiotensin II suggests that TGF-␤ might participate in the onset, maintenance, or inhibition of cardiac hypertrophy, as an autocrine or paracrine factor (1). We recently demonstrated that induction of the SkA promoter by TGF-␤ requires the MADS box protein serum response factor (SRF) and the SV40 enhancer-binding protein, TEF-1 (6); both also mediate activation of this promoter by ␣ 1 -adrenergic agonists (7). Dominant-negative mutations of the type II and type I TGF-␤ receptor, which share related serine/ threonine kinase domains, suffice to disrupt TGF-␤-dependent transcription (8,9); however, the molecular circuit that confers signal from the TGF-␤ receptor complex to SkA promoterbinding proteins is unknown.
Among the secondary or tertiary messengers that might be involved in this signaling cascade, it is noteworthy that activation of nuclear oncogene transcription factors Fos, Jun, and JunB precedes growth and the up-regulation of "fetal" cardiac genes, in cultured myocytes (10,11) and intact animals (12)(13)(14). Fos and Jun proteins (Fos, FosB, Fra-1, Fra-2, Jun, JunB, and JunD) each possess a basic domain for DNA binding and a leucine heptad repeat (leucine zipper) as an interface for homoor heterodimerization (15). Each member of this AP-1 transcription factor family recognizes the 12-O-tetradecanoylphorbol-13-acetate response element (TRE: TGA(G/C)TCA), although noncanonical sites also are reported (16 -18). Three lines of evidence support the inference that Fos/Jun proteins might mediate TGF-␤ signal transduction: TGF-␤ up-regulates the expression of junB and c-fos in skeletal myocytes (19), cardiac myocytes, 2 and other cell types, and AP-1 sites mediate autoinduction of TGF-␤1 itself (20). Moreover, in skeletal muscle, forced expression of either Fos or Jun reproduces the suppressive effect of TGF-␤ on myogenic differentiation, although the issue of physical association between Fos/Jun with myogenic helix-loop-helix proteins is unresolved (21)(22)(23).
Related co-transfection studies likewise support the premise that Fos/Jun mediates the induction of fetal cardiac genes by TGF-␤, as shown for ANF (24,25), ␤ myosin heavy chain, 3 and SkA (18). In the latter study, forced expression of Jun (or Fos plus Jun) up-regulated transcription of the human SkA promoter in cardiac myocytes from neonatal rats and in P19 teratocarcinoma cells. Deletion analysis of the SkA promoter indicated that nucleotides Ϫ153 to Ϫ36 were required for maximal trans-activation by Fos/Jun. Although no consensus AP-1 site was found within this region, sequence-specific binding to a noncanonical motif was believed to occur. Despite this suggestive information, the conclusion that Fos/Jun proteins augment the transcription of SkA and other fetal cardiac genes would be premature. Dichotomous results, repression (24) as well as activation by Fos/Jun (25), have been reported for ANF. Mechanical load, ischemia, and isoproterenol each highly induce JunB in myocardium in vivo (13,26,27), but few functional comparisons among Jun proteins are known for cardiac myocytes (24). Although prior results pointed to direct binding of Fos/Jun near the first SRE of the human SkA promoter, it has not been demonstrated whether point mutations of this construct that abolish AP-1 binding remain susceptible, or not, to induction by AP-1 factors. Finally, given the emerging importance of SRF in concert with TEF-1, and given the displacement of SRF at the first SRE by a GLI-Krü ppel protein, YY1, the relationship between control by these three factors and the induction by Fos/Jun merits study.
In the present report, we demonstrate that Jun, Fos plus Jun, Fos plus JunB, and Jun plus JunB all transactivate the SkA gene in cardiac myocytes, whereas JunB, like Fos, is ineffective individually. Notably, a SRE is necessary for AP-1 responsiveness of the SkA promoter, and suffices to confer induction by AP-1 to a heterologous promoter. Induction required full-length Jun protein, and did not involve measurable binding of AP-1 factors to the SkA SRE1, physical association of AP-1 and SRF, augmentation of SRF binding by AP-1, or potentiation of the SRF trans-activation domain, residues 266 -508 (SRF(266 -508)). Together, these results indicate that AP-1 factors can act through SRF to induce a hypertrophy associated gene, SkA, but trans-activate SRE reporter genes in the absence of direct binding to the promoter or DNA-bound SRF.
Cells were transfected by a modified DEAE-dextran method. DNA (2.5 g/ml reporter and 0 -10 g/ml expression vectors) was mixed with 150 g/ml DEAE-dextran (average molecular weight, 500,000; Sigma) in DF supplemented with 2.5% Cosmic calf serum (Hyclone). For all comparisons, DNA and promoter content were kept constant using equivalent amounts of vector. Cells were washed once, were incubated with 1 ml/35-mm dish of DNA-DEAE-dextran complex for 1 h, and were then shocked for 30 s with 10% dimethyl sulfoxide in DF. Cells were cultured overnight in DF supplemented with 5% horse serum, after which the medium was replaced by DF supplemented with 5 g/ml transferrin, 1 nM Na 2 SeO 3 , 1 nM LiCl, 25 g/mL ascorbic acid, and 100 g/ml bovine serum albumin (fatty acid free). Twenty-four h later, cells were harvested and assayed for luciferase activity (6) and for protein content using the Bradford assay (Pierce). Luciferase activity for each promoter was corrected for protein content of each extract and was normalized to the activity of each promoter in parallel cultures of control, vector-transfected cells. Co-transfection with a constitutive lacZ gene was omitted for three reasons. The brief half-life of luciferase compared to other reporter proteins can lead to misleading interpretation, especially for transcriptional repression (36); commonly used viral promoters contain functional AP-1 or SRF sites (37)(38)(39) and are upregulated by expression vectors used here 5 ; neutral core promoters that are unaffected by Fos/Jun, such as the c-fos Ϫ57/ϩ109 core promoter, are insufficiently active in cardiac myocytes for accurate quantitation of lacZ activity, 5 but can be utilized to drive luciferase in parallel cultures, as an ostensibly constitutive control. Protein content was not significantly altered by any of the inteventions, compared to the corresponding vehicle-treated, vector-transfected cells. Where indicated, serumfree medium was supplemented with 1 ng/ml of TGF-␤1 purified from duced in vitro using the SP6 TnT coupled reticulocyte lysate system (Promega). Gel mobility shift assays were performed as described previously (40): 6 l of the coupled transcription/translation reaction mixture was incubated with 20,000 cpm of 32 P-end-labeled DNA probe in 20 l of 20 mM HEPES, pH 7.9, 25 mM KCl, 0.1 mM EDTA, 10% glycerol, 10 mM MgCl 2 , 0.2 mM dithiothreitol, 0.025% Nonidet P-40, 50 g/ml of poly(dG-dC) or poly(dI-dC), 50 g/ml of bovine serum albumin. Poly(dG-dC) was used for assays involving the SkA SRE1-TATA probe and poly(dI-dC) for the TRE probe. DNA-protein complexes were incubated at room temperature for 45 min. The reaction mixtures were then loaded on a 4% low ionic strength polyacrylamide gel (acrylamide: bisacrylamide, 80:1) containing 45 mM Tris-borate, 1 mM EDTA, and 0.05% Nonidet P-40, and were electrophoresed in 45 mM Tris-borate, 1 mM EDTA at 4°C and 250 V for 2 h. The SkA SRE1-TATA probe, comprising SkA SRE1 (nucleotides Ϫ100 to Ϫ73) fused to mouse c-fos nucleotides Ϫ56 to Ϫ19, was generated by digesting the SkA SRE1-⌬56Fos luciferase expression vector with HindIII and PvuII. To ensure high specific activity, the probe was double labeled with the Klenow fragment of DNA polymerase I using [␣-32 P]dCTP and with polynucleotide kinase using [␥-32 P]dATP. Competing double-stranded oligonucleotides encompassing the MϪ94/Ϫ89 and MϪ81/Ϫ79 mutations of the SkA SRE1, which abolish SRF and YY1 binding respectively, were detailed previously (6). A double-stranded oligonucleotide containing a consensus TRE was subcloned into the SacI/NheI sites of pGL3 (5Ј-AGCTCGCTTGATGACTCAGCCGGAAGCTAG-3Ј); the sense strand is shown, and the consensus sequence is underlined. The probe was endlabeled using the Klenow fragment of DNA polymerase and [␣-32 P]dCTP. AP-1 consensus and mutant oligonucleotides were obtained from Santa Cruz Biotechnology (Santa Cruz, CA).
Statistics-For each promoter tested, luciferase activity is shown as the mean Ϯ S.E., relative to parallel cultures of vehicle-treated, vectortransfected cells. Data were compared by analysis of variance followed by the Student-Newman-Keuls multiple comparison test, using a significance level of p Ͻ 0.05.

AP-1 Factors Differentially Activate the Skeletal ␣-Actin
Promoter in Ventricular Myocytes-In preliminary Northern blot analyses, we observed a marked increase in c-fos and junB in ventricular myocytes treated with TGF-␤, with little or no change in c-jun, 5 as shown previously in skeletal muscle (19). To compare the potential for these three immediate-early genes to trans-activate the SkA promoter, ventricular myocytes were co-transfected with the SkA luciferase reporter gene and Fos, Jun, or JunB expression vectors, singly or in combination (Fig.  1). Whereas Fos and JunB alone were neutral, Jun activated transcription via the SkA promoter in a dose dependent fashion, up to 3.4-fold (p Ͻ 0.01). No further increase in Jun-dependent transcription was seen with a 5-fold larger, 2-kilobase pair segment of the promoter. 5 Although Fos and JunB singly had no effect, in tandem they transactivated the SkA promoter synergistically (2.2 Ϯ 0.16; p Յ 0.01). More than additive induction likewise was seen using Fos plus Jun (3.1 Ϯ 0.32; p Յ 0.01), but not Jun cotransfected with JunB. Thus, co-expression of Fos and JunB, AP-1 factors that are highly induced by TGF-␤, can suffice to increase transcription of the SkA promoter.
To verify that induction of the SkA promoter by forced expression of Fos, Jun, and JunB is specific, a canonical AP1responsive element and a constitutive neutral promoter were examined, as positive and negative controls, respectively. Control of the SkA promoter by the various permutations of AP-1 factors accurately resembled activation of the human collagenase TRE1, although the TRE1 was more highly induced. By contrast, little or no activation was seen using the c-fos neutral core promoter, ⌬56Fos. Jun, Fos/Jun, and Fos/JunB up-regulated the ⌬56Fos reporter by no more than 40, 70, and 60%, respectively (p Յ 0.01). Indeed, slight inhibition was observed at high concentrations of Fos or JunB (p Յ 0.01).
Trans-activation of the ␣SkA by the AP-1 Factors Requires SRE1 in Concert with the TATA Box-To test the hypothesis that induction by Fos/Jun might map to one or both TGF-␤ response elements of the SkA promoter (i.e. SRE1 and a TEF-1 site), cardiac myocytes were co-transfected with 10 g of the Jun expression vector or the empty vector control, together with a luciferase reporter gene driven by the wild-type SkA promoter versus linker-scanning mutations as summarized in Fig. 2. Detailed previously (6), mutation of the YY1 site weakly activates basal activity; mutations of the SRF, TEF-1, Sp1, and TATA motifs inhibit basal activity of the promoter by up to 90%. Of the five linker-scanning mutations tested, three had no significant effect on induction by Jun. The mutation specific for YY1 (MϪ81/Ϫ79) lies immediately 3Ј to the proximal SRF binding site, and disrupts a potential TRE-like sequence (18). Mutation of the YY1 site did not alter trans-activation by Jun (5.09 Ϯ 0.42, relative to activity of this promoter in vectortransfected cells), nor did a mutation that destroys the binding site for TEF-1 (MϪ70/65; 4.97 Ϯ 0.76). The mutation that blocks the binding of Sp1 (MϪ52/Ϫ47) also had no significant effect (3.88 Ϯ 0.41). MϪ94/Ϫ89 is a substitution in the 5Ј arm of the SRE1 palindrome, which specifically disrupts SRF binding while sparing YY1: by contrast to activation of the wildtype promoter and these three mutations by Jun, mutation of the SRF binding site reduced Jun-dependent transcription of the promoter by one-half (2.34 Ϯ 0.40; p Յ 0.01). A similar decrease in induction by Jun resulted from mutation of the TATA box (MϪ28/Ϫ23; 2.00 Ϯ 0.07; p Յ 0.01). Together, these mutations suggest that SRF, but not TEF-1 or Sp1, is necessary for full trans-activation of the SkA promoter by Jun. That a mutation of the TATA box also inhibited trans-activation by Jun is intriguing, as both SRF (40, 41) and Jun (42) have been proven to associate with basal transcription factors.
A SRE Confers Jun-dependent Transcription to a Heterologous Promoter-Conversely, to establish whether the SkA SRE1 suffices for trans-activation by Jun, we co-transfected Jun with the SkA SRE1 upstream of the neutral promoter, ⌬56Fos (Fig. 3A). Luciferase expression driven by ⌬56Fos was not significantly changed by Jun (1.39 Ϯ 0.09). By contrast, Jun trans-activated the SRE1-⌬56Fos construct 5.02 Ϯ 0.48 fold (p Յ 0.01). To distinguish whether AP-1 dependent activation was specific to the SkA SRE1, this element was exchanged for the c-fos SRE: Jun induced the c-fos SRE-⌬56Fos reporter to the same extent (5.11 Ϯ 0.47; p Յ 0.01). By contrast, Jun decreased transcription of these isolated elements in parallel cultures of cardiac fibroblasts under the same transfection conditions; where indicated, cardiac fibroblasts were cultured in the presence of TGF-␤1, to demonstrate that the isolated SREs are functional in this cell background (Fig. 3B). Together with mutagenesis of the full-length SkA promoter, these findings indicate that SRF is both necessary and sufficient for activation of the SkA promoter in ventricular muscle cells by AP-1 transcription factors.
To test what domains of Jun might be necessary or sufficient for these effects, we co-transfected cardiac myocytes with 10 g of the Jun expression vector, SRE reporter genes, and 0 -100 ng of an SRF expression vector (Fig. 3, A and C). This concentration of SRF by itself did not significantly activate any of the three reporter genes. However, exogenous SRF was synergistic with Jun, up-regulating both the SkA SRE1-and c-fos SRE-⌬56Fos constructs, up to 14-fold, nearly three times the level of induction produced by Jun alone. When wild-type SRF was co-transfected with Jun mutants, no trans-activation was observed with a deletion of the DNA-binding domain (Jun⌬RK), deletion of the dimerization domain (Jun⌬LZ), or partial deletion of the trans-activation domain (TAM67). Each of the three Jun mutants also failed to transactivate, by themselves, the full-length SkA reporter gene. 5 Conversely, in cells cotransfected with Jun, the synergistic effect of wild-type SRF was abolished with SRFpm1, a point mutation of SRF which cannot bind DNA, and was markedly impaired with a mutation that deletes a trans-activation domain, SRF(1-338); the remaining cooperative effect is consistent with the residual activation domain of this mutation (43). As the high levels of endogenous Jun and SRF in myocardium (6,44), together with the limited efficiency for transfection of ventricular myocytes, would confound efforts to compare the levels of each protein in vivo, we synthesized each Jun and SRF mutation in vitro in the presence of [ 35 S]methionine and verified that the translational efficiency and inherent stability of each protein were equivalent to that of wild-type Jun and SRF. 5 Thus, functional DNAbinding, dimerization, and trans-activation domains were each necessary for up-regulation of the SkA promoter by Jun. The necessity for full-length Jun contrasts with effects shown for the isolated Jun trans-activation domain in repression of myogenin activity (22).
Activation of SRF Transcription by Jun-A requirement for full-length Jun would be anticipated by either of two contrasting models: activation through direct binding of Jun, suggested previously (18), or indirect activation by a Jun-induced protein.
Given the requirement for SRF binding to SRE1 (Fig. 2), Jun might thus up-regulate the SkA promoter indirectly, by augmenting SRF expression or activity. Although the small proportion of cardiac myocytes that take up foreign genes during transient transfection precludes a direct test of whether Jun induces SRF in this background, we co-transfected ventricular muscle cells with Jun together with a murine SRF luciferase reporter gene. Forced expression of Jun increased transcription of the SRF promoter (2.29 Ϯ 0.19; p Յ 0.01). However, it is implausible that changes in SRF abundance alone could explain induction of the SkA promoter by Fos/Jun. First, the magnitude of SRF induction was modest. More importantly, overexpression of SRF at up to 100 ng/culture did not increase transcription of the isolated SREs (Fig. 3), but causes squelching at several higher concentrations 6 ; thus, SRF is not limiting in ventricular muscle cells. Therefore, it is necessary to consider that AP-1 factors might modulate activity of SRF, or promote transcription via SREs, by other mechanisms.
To test whether Jun could augment gene expression via the trans-activation domain of SRF, we employed a GAL4 fusion protein comprising the DNA-binding domain of yeast GAL4 fused to SRF amino acids 266 -508, GAL4SRF(266 -508). Five g of GAL4SRF(266 -508) were co-transfected with or without the Jun expression vector, using a GAL4-dependent luciferase reporter (Fig. 4). In the absence of exogenous Jun, the GAL4/ SRF fusion protein increased transcription of the reporter 8.39 Ϯ 0.42 (p Ͻ 0.01); transcription via the SRF activation domain was not augmented significantly by the addition of exogenous Jun. Thus, Jun did not increase transcription, when the SRF activation domain was tethered to DNA by a heterologous DNA-binding domain. This also argues against a generalized activation of transcription by Jun.
To address two further possibilities, that Jun might activate the SRE by physical association with native SRF, or facilitate DNA binding by SRF, gel mobility-shift assays were performed using SRF, Fos, and Jun, produced by in vitro transcription and translation (Fig. 5A). A HindIII-PvuII DNA fragment of the SkA SRE1-⌬56Fos luciferase reporter gene was used as probe, encompassing the SkA SRE1 plus all sequences of the core c-fos promoter including the TATA box. A double-stranded oligonucleotide containing the consensus AP-1 binding site was used, for comparison. Recombinant SRF bound the SRE1-⌬56 Fos probe, was displaced by the competitor that binds SRF but not YY1 (MϪ81/Ϫ79), and was displaced poorly by the reciprocal mutation, which disrupts SRF binding (MϪ94/Ϫ89). Jun, Fos, and co-translated Fos/Jun showed no direct binding to the probe, did not form a higher order complex with DNA-bound SRF, and did not alter binding of SRF to DNA in any fashion. To ensure that the lack of protein-protein interaction was not due to inadequate expression of Jun or Fos, parallel experiments were performed using the AP-1 consensus probe. Jun and co-translated Fos/Jun bound the TRE, were displaced by excess unlabeled TRE, and were unaffected by the mutated TRE. Thus, both AP-1 factors were expressed as stable proteins with the expected DNA-binding and dimerization properties. To exclude direct association between Fos/Jun proteins and SRF, we demonstrated no binding of recombinant SRF to Jun homodimers or Fos/Jun heterodimers, and no effect of SRF on the binding of AP-1 factors to the TRE probe (Fig. 5B). DISCUSSION The present investigations show that transcription of SkA, a "fetal" cardiac gene associated with myocardial hypertrophy, can be augmented by AP-1 transcription factors (Jun, Fos plus Jun, Fos plus JunB, or Jun plus JunB) in ventricular myocytes from neonatal rats. Despite marked differences in the constructs and procedures used, this corroborates and extends the report of Bishopric et al. (18), that forced expression of Jun or Fos plus Jun up-regulates the human SkA promoter in cardiac cells. Because dichotomous results, both induction and repression, were reported for the ANF gene (27,28), consensus regarding the functional role of Fos/Jun factors in hypertrophy has been lacking. Although JunB, like Fos and Jun, is highly induced by both mechanical load (13,14) and trophic factors that up-regulate the endogenous SkA gene (TGF-␤, catecholamines, and angiotensin II 2 (18,45)), even less functional evidence has been available for JunB in cardiac muscle (24). Whereas Fos proteins do not form homodimers, Jun and JunB form both homodimers and heterodimers with Fos/Jun proteins and more distantly related factors, via the leucine zipper. Our results demonstrate functional synergy between JunB and Fos in cardiac myocytes; either alone had no effect or was inhibitory, while Fos plus JunB activate the SkA promoter. Permutations of Fos, Jun, and JunB that activate the SkA gene correspond to those that induced the human collagenase TRE, a canonical AP-1-responsive element.
Our results diverge from previous findings, however, on mechanisms to explain trans-activation of SkA by AP-1. Here, SRE1 was both necessary and sufficient. A molecular basis for activation of the SkA promoter by SRF in concert with TEF-1 has been proposed (6, 7), with virtually identical conclusions for the avian and rat promoters. Cooperation of SRF and TEF-1 also was implicated in two distinct transduction pathways for hypertrophy, TGF-␤ and ␣ 1 -adrenergic agonists. By contrast, the TEF-1 site is dispensable for full augmentation of the SkA promoter by Jun. In agreement, nested deletions of the human SkA promoter suggested that nucleotides encompassing the first SRE (Ϫ153 to Ϫ87) were required for maximal transactivation by Fos plus Jun, whereas more proximal sequences including the TEF-1 site at nucleotides Ϫ71 to Ϫ65 do not mediate AP-1 responsiveness (18). While direct binding of Jun and Fos/Jun to a noncanonical site near SRE1 was reported for the human SkA promoter, our results point instead toward an indirect mechanism, mapped to SRE1 itself. First, Jun can induce the avian SkA promoter in the absence of TEF-1, Sp1, or YY1 binding, yet the SRF binding site and TATA box are indispensable. This is intriguing, given that SRF contacts the RAP74 subunit of transcription factor IIF (39, 41), while Fos and Jun contact other basal factors, transcription factor IIB and TATA box-binding protein (42,46). Second, the isolated SkA SRE1 is sufficient for AP-1-dependent expression and is interchangeable, in this respect, with the c-fos SRE. Third, using recombinant Fos and Jun produced in reticulocyte lysates, we detected no binding to the SkA SRE1. This discrepancy with direct binding of AP-1 factors reported for the human SkA promoter (18)  in the earlier study, proteins were produced in E. coli, truncated Fos and Jun were used, and five times more protein was used for the SkA probe versus the authentic TRE. Alternatively, sequence dissimilarities may be germane. Among the characterized vertebrate SkA genes, only the human promoter matches the TGACTCA consensus TRE at five positions that include both cytosine residues.
Control of the SkA SRE1 by Jun and synergy with SRF both required full-length Jun protein with intact DNA binding, dimerization, and trans-activation domains. As no binding of AP-1 factors was detected to SRE1, mechanisms alternative to direct association must be considered. Conceivably, AP-1 factors might increase transcriptional activity of an SRF binding site through protein-protein interactions with SRF, increasing SRF abundance, or affecting transcriptional activity of SRF. Our results provide no support for a ternary or quarternary complex of Fos/Jun with DNA-bound SRF. No physical interaction was seen between SRF and AP-1 in gel mobility-shift assays, nor did AP-1 factors augment DNA binding by SRF. However, measurements of DNA-protein interaction might overlook low affinity binding or interactions that require a co-activator. Using GAL4 fusion proteins to overcome limitations of both gel retardation assays and endogenous SRF, we found that Jun could not potentiate the C-terminal activation domain of SRF. It is a formal possibility that Jun might interact (with affinity too low to be stable in vitro) only with native SRF or the native SRF⅐SRE complex. Increased SRF abundance is unlikely to explain AP-1 dependent SkA transcription, since SRF was not limiting in cardiac myocytes. Our findings are more consistent with the alternative, that Fos/Jun transcription factors increase, instead, the transcriptional activity of SRF. In principle, this could be contingent on altered expression of an autocrine or paracrine factor (23), a protein kinase modulating SRF activity (47), or a co-activator. Whereas SRF accessory proteins include most obviously the ternary complex factors Elk-1/TCF and SAP-1 (48), association and synergy with SRF both were demonstrated (49) for the cardiac-restricted homeodomain protein, Nkx-2.5, vertebrate homologue of the Drosophila tinman gene (50,51).
In summary, our studies reveal a novel AP-1-dependent pathway for gene induction in cardiac myocytes, via indirect activation of SREs. Hence, AP-1 factors might plausibly be involved in the up-regulation of genes containing SREs, including skeletal and smooth muscle ␣-actin and the immediateearly gene c-fos, in the setting of cardiac hypertrophy. It is unknown whether the greater induction of "fetal" ␣-actin transcripts relative to cardiac ␣-actin reflects inherent differences among SREs, contextual sequences, or elements elsewhere in the respective promoters. The SRE has been implicated in numerous settings as a pivotal regulatory element for cardiac gene expression during hypertrophy, for activation of the SkA promoter by TGF-␤ (6), basic FGF (52), and ␣ 1 -adrenergic agonists (7), up-regulation of ANF by ␣ 1 -adrenergic signals (53), and induction of c-fos by load (54), passive stretch (55), or angiotensin II (56). Control of SRE-dependent transcription by AP-1 factors was selective, as the neutral core promoter was unaffected by Jun, contrasting with the global increase in transcription provoked by Ras under similar conditions (36). Conversely, TGF-␤ selectively up-regulates fetal cardiac genes, with little change in RNA or protein content (57). Hence, the overall increase in cell RNA and protein can be dissociated from the "fetal/hypertrophic" program. Analogously, rapamycin blocks the increase in total cell protein and p70 S6 kinase activity in angiotensin II-treated cardiac myocytes, while not affecting induction of c-fos or fetal cardiac genes (58). The latter report highlights other data pointing to the possibility, with which the present study and our Ras results concur (36), of distinguishable signaling cascades for these components of the hypertrophic phenotype, that the global increase of cell protein is mediated by p70 S6 kinase, while the fetal program might be mediated, at least in part, by mitogen-activated protein kinase induction of Fos, and Jun N-terminal kinase acting though Jun (58). FIG. 5. Fos/Jun proteins do not bind the SkA SRE1 or DNA-bound SRF in vitro. SRE1-⌬56 Fos (A) and TRE (B) probes were end-labeled with 32 P, incubated with Fos, Jun, and SRF proteins produced in vitro, and analyzed by the electrophoretic mobility shift assay in the presence or absence of a 100-fold excess of the indicated unlabeled competitor. The amount of Fos, Jun, and SRF designates microliters of the programmed reticulocyte lysate. The total quantity of reticulocyte lysate per lane was maintained at 6 l using lysate programmed with the empty SV40 vector. ns, nonspecific binding produced by the control reticulocyte lysate.