Adenylyl cyclase type VI gene transfer reduces phospholamban expression in cardiac myocytes via activating transcription factor 3.

Cardiac-directed expression of adenylyl cyclase type VI (AC(VI)) increases stimulated cAMP production, improves heart function, and increases survival in cardiomyopathy. In contrast, pharmacological agents that increase intracellular levels of cAMP have detrimental effects on cardiac function and survival. We wondered whether effects that are independent of cAMP might be responsible for these salutary outcomes associated with AC(VI) expression. We therefore conducted a series of experiments focused on how gene transcription is influenced by AC(VI) in cultured neonatal rat cardiac myocytes, with a particular focus on genes that might influence cardiac function. We found that overexpression of AC(VI) down-regulated mRNA and protein expression of phospholamban, an inhibitor of the sarcoplasmic reticulum Ca(2+)-ATPase. We determined that the cAMP-responsive-like element in the phospholamban (PLB) promoter was critical for down-regulation by AC(VI). Overexpression of AC(VI) did not alter the expression of CREB, CREM, ATF1, ATF2, or ATF4 proteins. In contrast, overexpression of AC(VI) increased expression of ATF3 protein, a suppressor of transcription. Following AC(VI) gene transfer, when cardiac myocytes were stimulated with isoproterenol or NKH477, a water-soluble forskolin analog that directly stimulates AC, expression of ATF3 protein was increased even more, which correlated with reduced expression of PLB. We then showed that AC(VI)-induced ATF3 protein binds to the cAMP-responsive-like element on the PLB promoter and that overexpression of ATF3 in cardiac myocytes inhibits PLB promoter activity. These findings indicate that AC(VI) has effects on gene transcription that are not directly dependent on cAMP generation.

Adenylyl cyclase (AC) 1 is the effector molecule in the ␤-adrenergic receptor-G-protein-AC signaling pathway in cardiac myocytes and other cells. Previous studies showed that the amount of AC sets a limit on the ability of cardiac myocytes to generate cAMP (1), and cardiac-directed expression of AC type VI (AC VI ) has pronounced favorable effects on cardiovascular function in normal and failing hearts (2)(3)(4)(5)(6).
The mechanisms explaining these favorable effects of AC VI are not precisely known. The most direct explanation is that the benefits stem from increased intracellular levels of cAMP; this explanation is contrary to current dogma in heart failure which asserts that inotropic agents that increase cAMP are bad for the heart. Indeed, pharmacological agents that stimulate the ␤-adrenergic receptor or decrease the breakdown of cAMP increase cardiac function but do not appear to prolong life (7)(8)(9). In contrast, AC VI , a dominant AC isoform in mammalian cardiac myocytes, improves global cardiac function, attenuates myocardial hypertrophy, and increases survival in murine cardiomyopathy (3,4). However, when cardiac-directed ␤-adrenergic receptor expression is used to treat this same model, life is shortened (10), underscoring the marked differences evoked by these signaling elements, both of which are associated with increased intracellular cAMP.
Because of the disparate results between ␤-adrenergic receptor and AC expression in heart failure, we postulated that the salutary effects of AC expression may depend on mechanisms not directly linked with cAMP production per se. A suitable initial approach, we reasoned, would be to determine whether AC VI gene transfer (independently of cAMP generation) alters transcription of genes important in contractile function.
To investigate cAMP-independent events in cardiac myocytes, we focused on how AC VI affects transcriptional regulation and to what extent that effect is different from the effect evoked by agents that directly stimulate cAMP production. We found that AC VI , in the absence of ␤-adrenergic receptor or AC stimulation, down-regulated the expression of genes containing the cAMP-responsive element (CRE). Specifically, phospholamban (PLB), an inhibitor of the sarcoplasmic reticulum Ca 2ϩ -ATPase, was down-regulated by AC VI gene transfer, a result of increased expression of activating transcription factor-3 (ATF3), a transcriptional suppressor in the CRE-binding protein (CREB)/ATF family (11). In contrast, direct stimulation of ␤-adrenergic receptors with isoproterenol or of AC with NKH477 increased PLB expression and did not alter expression of ATF3 protein.

EXPERIMENTAL PROCEDURES
Detection of Gene Expression Using GeneArray System-The "mouse cAMP/Ca 2ϩ signaling pathway finder" membrane was purchased from SuperArray Inc. (Frederick, MD). Among the 96 genes spotted on the membrane, 52 of them contained CRE or CRE-like elements, and others contained elements that respond to calcium. Neonatal rat cardiac myocytes from the following four experimental conditions were used in these experiments: uninfected; uninfected but stimulated with the forskolin analog NKH477 (10 M); infected with adenovirus encoding AC VI (Ad.AC VI ) at 600 virus particles (vp) per cell; and adenovirus encoding enhanced green fluorescent protein (Ad.EGFP) at 600 vp/cell. Five micrograms of total RNA extracted from the cardiac myocytes from each condition was reverse-transcribed into cDNA in the presence of [␣-32 P]dCTP. The radioactive labeled cDNA probes were hybridized to the genes spotted on the GeneArray membrane.
Cloning of PLB Promoter and Site-directed Mutagenesis-The PLB promoter fragment from Ϫ156 to ϩ64 bp (12) was cloned from rat DNA by PCR. The primers used in the PCR contain restriction enzyme sites NheI and XhoI at each end and the following sequences: PLBF, 5Ј-CTAGCTAGCTGAAGCACAACATGTTACCG-3Ј; PLBR, 5Ј-CCGCTC-GAGTTAGTTGTGTGAAGTCTGGG-3Ј. The PCR was performed using a kit containing Pfu polymerase (Stratagene, San Diego, CA). The PCR product was subcloned into PCR 4 blunt-TOPO cloning vector (Invitrogen, Carlsbad, CA), and the sequences were confirmed by sequencing. The PLB promoter fragment was released from the TOPO vector by NheI and XhoI digestion and subcloned into pGL-3 vector (Promega, Madison, WI) in front of the luciferase gene to obtain the pPLB-luc reporter plasmid. To mutate the CRE-like site in the PLB promoter, site-directed mutagenesis was performed using the Quick-change kit (Stratagene, San Diego, CA). The primers used for mutating the CRElike element were 5Ј-p-GTTGGGATTCCTATGGACCATAGTAAGACC-TCCCTAGAATG-3Ј and 5Ј-p-CATTCTAGGGAGGTCTTACTATG-GTCCATAGGAATCCCAACC-3Ј.
Neonatal Rat Cardiac Myocyte Isolation, Gene Transfer, and Stimulation-Cardiac myocytes from neonatal rats were isolated (1). Cells were infected 1 day after attachment to the plate with recombinant adenovirus (600 vp/cell) encoding enhanced green fluorescence protein (EGFP) or AC VI with an AU1 tag (six amino acids, DTYRYI, derived from bovine papilloma virus-1) at the C terminus of the AC VI for protein detection. Stimulation with isoproterenol (10 M) or NKH477 (10 M) was performed 1 day after virus infection for 10 min (for change in phosphorylation) or 20 h (for gene expression).
Transfection and Luciferase Assay-Transfection of cardiac myocytes was performed using GeneShuttle-40 reagent (Qiagen, Carlsbad, CA). DNA and Geneshuttle-40 were diluted separately in Dulbecco's modified Eagle's medium without serum, and a 1:6 ratio of DNA to lipid was used. The diluted DNA and lipid were combined and incubated (23°C, 30 min), and the mixture was transferred onto cultured cardiac myocytes. The luciferase assay was performed 48 h after transfection.
Luciferase activity of the reporter co-transfected with control plasmid was defined as 100%. In the experiments whose results are displayed in Fig. 3 and Fig. 7, the control plasmids (pAd and pCG) carrying the cytomegalovirus (CMV) promoters were included to ensure that each transfection mixture contained the same amount of CMV promoter.
Immunofluorescence Staining-Cardiac myocytes were fixed with 10% formalin solution containing 4% formaldehyde (Sigma) for 15 min at 23°C. Fixed cells were washed four times with phosphate-buffered saline (PBS). Cells were blocked with normal rabbit serum for 1 h and incubated with primary antibodies including anti-AU1 (Babco, Berkeley, CA) for detection of AC VI at 1:1000 dilution, and anti-ATF3 (Santa Cruz Biotechnology, Santa Cruz, CA) at 1:50 dilution for 1 h at 23°C. The cells were washed four times with PBS and then incubated for another hour with a secondary antibody that was conjugated with fluorescein isothiocyanate or rhodamine. Stained cells were imaged by fluorescence microscopy.
Western Blot and Northern Blot Analysis-Western blotting of cardiac myocyte extracts was performed according to the protocol developed by Hagiwara et al. (13). The CREB/ATF family proteins were detected using individual antiserum or phosphoserine 133-specific CREB antiserum. PLB protein was detected using a monoclonal anti-PLB antibody (ABR Affinity BioReagents, Golden, CO) at a dilution of 1:500. Northern blot analysis was performed as described previously (1). Twenty micrograms of total RNA obtained from cells from each condition was separated on a 1% agarose gel; PLB-specific mRNA was detected by using [␣-32 P]dCTP-labeled PLB probe that contained rat PLB cDNA.
Nuclear Extracts-Nuclear extracts were prepared as described previously (14). Cardiac myocytes were washed in PBS, and the cell pellets were resuspended in cell lysis buffer A containing 10 mM HEPES (pH 8.0), 1.5 mM MgCl 2 , 10 mM KCl, 0.5 mM dithiothreitol, 300 mM sucrose, 0.1% Nonidet P-40, and protease inhibitors mixture tablets (1 tablet dissolved in 50 ml of buffer A) (Roche Applied Science). They were homogenized 3-4 times on ice in a glass homogenizer by 10 strokes with pestle B. The homogenates were transferred into tubes and centrifuged in a Microfuge at 3000 rpm for 10 min. The nuclei pellets were washed once by adding buffer A without Nonidet P-40 and collected as above. Nuclei were lysed by adding buffer B containing 20 mM HEPES (pH 7.9), 420 mM NaCl, 1.5 mM MgCl 2 , 0.2 mM EDTA, 25% glycerol, and protease inhibitor mixture. Nuclei were homogenized as above, incubated on ice for 10 min, and centrifuged in a Microfuge at 13,000 rpm for 15 min at 4°C. The supernatant was collected and stored in small aliquots at Ϫ80°C.
Electrophoretic Mobility Shift Assay (EMSA)-This assay was performed as described previously (14). The oligonucleotides (oligos) used as probes in EMSA contained the following sequences: plbCRE, 5Ј-GATTCCTATGTGATGTCATAAGACCT-3Ј; and plbCREm, 5Ј-GATTC-CTATGGACCATAGTAAGACCT-3Ј. Double-stranded oligos were labeled by filling in the 3Ј-end C of the complementary strand with [␣-32 P]dCTP using Klenow fragment of DNA polymerase in the DECAprime II kit (Ambion, Austin, TX). Nuclear extracts (10 g) were incubated with the radioactive labeled probe (10 5 cpm) and 1 g of poly-(dI-dC) in binding buffer containing 50 mM KCl, 20 mM HEPES (pH 7.9), 0.5 mM EDTA, 5% glycerol, 1 mM dithiothreitol, 1 mg/ml bovine serum albumin, and 0.1% Nonidet P-40 for overnight at 4°C. The mixtures were then loaded onto 4% native polyacrylamide gels and electrophoresed in 0.5ϫ TBE buffer (6 h, 23°C). The gel was transferred onto Whatman No. 3MM paper, wrapped in a seal wrap, and exposed to an x-ray film at Ϫ80°C. Competition with unlabeled oligos was performed by incubating cold oligos with nuclear extracts for 10 min before adding the probe. Supershifts with antibodies were performed by either incubating nuclear extracts with antibody overnight before adding the probe or by incubating nuclear extracts with the probe overnight and then adding antibodies for 30 -60 min.
Statistics-Data are reported as mean Ϯ S.E. unless otherwise stated. Comparisons between two group means were made by using Student's t test for unpaired data (two-tailed). Multiple group comparisons were conducted using ANOVA, and post hoc testing was conducted (only when overall p Ͻ 0.05) by Student's t test for unpaired data (two-tailed), with Bonferonni correction for multiple tests. The null hypothesis was rejected when p Ͻ 0.05.

Regulation of Transcription by AC VI Gene Transfer Versus
Pharmacological Stimulation of AC-To determine whether overexpression of AC VI versus pharmacological stimulation of AC affected endogenous gene expression differently in cardiac myocytes, we used gene transcript profiling. The results of this FIG. 1. AC VI gene transfer versus AC stimulation on mRNA expression, initial screen. Total RNA from neonatal rat cardiac myocytes was reverse-transcribed into cDNA in the presence of [␣-32 P]dCTP. The radioactive labeled probes were used to hybridize with the cDNA on GeneArray membranes (the cAMP/Ca 2ϩ pathway finder, see "Experimental Procedures"). This initial screen suggested that PLB expression was reduced by AC VI gene transfer but increased by pharmacological stimulation of AC. Atrial natriuretic factor (ANF) and ␤ 1 -adrenergic receptor (␤ 1 AR) expression also appeared to be increased by AC stimulation but decreased by AC VI gene transfer. Control, neonatal rat cardiac myocytes alone; Ad.EGFP, after infection with adenovirus encoding EGFP; Ad.AC VI , after infection with adenovirus encoding AC VI ; NKH, after incubation with NKH447 (10 M), a watersoluble forskolin analog that directly stimulates adenylyl cyclase. initial screen of selected genes indicated that AC VI gene transfer was associated with reduced expression of PLB, ␤ 1 -adrenergic receptor, and atrial natriuretic factor. In contrast, expression of these genes was increased by direct AC stimulation with NKH477 ( Fig. 1).
AC VI Reduces PLB Expression-To confirm the findings obtained from the gene transcript profiling experiments, we selected PLB for further studies. We analyzed PLB expression by Northern blotting and found that PLB mRNA was reduced by AC VI gene transfer into cardiac myocytes ( Fig. 2A) and that stimulation with NKH477 was associated with further reduction in PLB mRNA (Fig. 2B). Gene transfer of an AC that lacks the transmembrane domain had no effect, suggesting that the effects on PLB were AC VI -specific. Reduced PLB expression by AC VI gene transfer was viral dose-related; increased vp/cell ratio was associated with a greater reduction in PLB mRNA ( Fig. 2A). In addition to reduced PLB mRNA, PLB protein was also reduced by overexpression of AC VI (Fig. 2C). In contrast, NKH477 increased PLB mRNA expression (Fig. 2B).
AC VI Reduces Transcription through CRE or CRE-like Element-The rat minimal PLB promoter activity resides in a fragment from Ϫ159 to ϩ64 bp (12). Although there is no consensus CRE element (TGACGTCA) in that region, there is a CRE-like element TGATGTCA located at Ϫ58 bp in the PLB promoter. By using a luciferase reporter assay system, we studied the effect of AC VI on the activity of the PLB promoter. In a transient transfection assay, the minimal PLB promoter fragment showed relatively high activity when co-transfected with the pAd vector carrying the cytomegalovirus (CMV) promoter. However, co-transfection of the reporter with pAd.AC VI reduced its activity (p ϭ 0.024) (Fig. 3A). To determine the importance of the CRE-like element, the CRE-like site was mutated into GACCATAG (Active Motif, Carlsbad, CA) through site-directed mutagenesis, and the mutant promoter fragment was linked to the luciferase reporter gene and used in transient transfection assays as above. Mutation of CRE-like element was associated with marked reduction of basal PLB activity, and the suppressive effect of AC VI on the promoter was absent (Fig. 3A). These results suggest that the CRE-like element in the PLB promoter plays an important role in downregulation of PLB by AC VI .
In contrast to the inhibitory effect of pAd.AC VI on reporter FIG. 3. AC VI inhibits transcription of the PLB promoter through a CRE-like site. A, the normal PLB promoter or a mutant PLB promoter (mPLB) was linked to a luciferase reporter and co-transfected with pAd.AC VI (AC VI ) or control plasmid pAd (Con) into cardiac myocytes. AC VI inhibited minimal PLB promoter activity. Mutation of the CRE-like element in the mini-PLB promoter greatly reduced the PLB promoter activity, and the suppressive effect of AC VI was diminished. B, the normal PLB promoter luciferase reporter was co-transfected with each of three expression vectors: control (pAd), AC VI , and CREB. Transfected cells were stimulated with NKH477 (10 M) 1 day after transfection. The results showed the following: 1) the suppressive effect on the PLB promoter was specifically mediated by AC VI because the CREB construct did not have such effect; 2) the minimal PLB promoter cannot be activated by NKH477. Statistical analysis (3 ϫ 2 ANOVA) showed that for each of the three constructs, there was no increase in activity with NKH477 stimulation compared with basal (B) activity. There were also no differences in activity between control and CREB (basal or stimulated). In contrast, the AC VI expression vector showed marked reduction in basal (p Ͻ 0.01) and stimulated (p Ͻ 0.01) activities compared with control and CREB vectors. C, the somatostatin promoter-luciferase reporter was co-transfected with pAd into cardiac myocytes, and transfected cells were stimulated with NKH477 (10 M). NKH477 activates the minimal somatostatin promoter.
FIG. 2. AC VI reduces PLB expression. A, PLB mRNA was reduced by overexpression of AC VI . Neonatal rat cardiac myocytes were infected with different amounts of adenovirus encoding AC VI or EGFP. Twenty micrograms of total RNA isolated from neonatal rat cardiac myocytes was loaded in each lane and hybridized with PLB cDNA probe. B, PLB mRNA was down-regulated by overexpression of AC VI but increased by AC stimulation with NKH447. All adenoviruses were used at 600 vp/cell. NKH477 (10 M) was added 20 h after gene transfer. Ad.mAC, a mutant form of AC that lacks the transmembrane domain, had no effect, suggesting that the effects on PLB were AC VI -specific. AC VI , Ad.AC VI ; GFP, Ad.EGFP; Mini AC, Ad.mAC. C, PLB protein expression was reduced by AC VI . Twenty micrograms of protein from lysates of neonatal rat cardiac myocytes were loaded per lane and separated by SDS-PAGE. PLB protein was detected using an anti-PLB antibody. activity (Fig. 3A), co-transfection of the reporter with a CREB expression plasmid (pcDNA3flagCREB; a gift from Dr. Marc Montminy, Salk Research Institute) had no effect (Fig. 3B). NKH477 stimulation of cardiac myocytes transfected with the PLB promoter-luciferase reporter construct did not affect minimal PLB promoter activity, even in the presence of CREB (Fig.  3B). Statistical analysis (3 ϫ 2 ANOVA) showed that for each of the constructs there was no increase in activity with NKH477 stimulation. There were also no differences in activity between control and CREB (basal or stimulated). In contrast, the AC VI expression vector showed marked reduction in basal (p Ͻ 0.01) and stimulated (p Ͻ 0.01) activities compared with the control construct. Finally, the AC VI expression vector showed marked reduction in basal (p Ͻ 0.01) and stimulated (p Ͻ 0.01) activities compared with CREB as well.
NKH477 stimulation of cardiac myocytes transfected with the somatostatin promoter-luciferase reporter construct (a gift from Dr. Marc Montminy, Salk Research Institute) significantly activated somatostatin promoter activity (p Ͻ 0.019) (Fig. 3C), which may have resulted from activated CREB that bound to the consensus CRE site (TGACGTCA) in this promoter. Our results suggest that the suppressive effect of AC VI is mediated through the CRE-like element in the PLB pro-moter, but activation of endogenous PLB transcription by NKH477 is not mediated through the CRE-like element in the minimal PLB promoter.
Overexpression of AC VI Increases ATF3 Expression-To determine the mechanism by which AC VI down-regulates PLB expression, we measured the expression of transcription factors in the CREB/ATF families in response to AC VI gene transfer. We found that Ad.AC VI (600 vp/cell) did not alter the expression of CREB, CREM, ATF1, ATF2, and ATF4 proteins. However, the expression of ATF3 was increased (Fig. 4A, 1st  and 2nd lanes versus 5th lane). Following AC VI gene transfer, ATF3 protein expression was further increased by isoprotere- nol or NKH477 stimulation (Fig. 4A, 6th and 7th lanes; Fig.  4B), which correlated with a further reduction in PLB expression (Fig. 2C). ATF3 expression was maximal at 600 vp/cell (Fig. 4B). In contrast, stimulation of Ad.EGFP-infected cardiac myocytes with isoproterenol or NKH477 had no effect on ATF3 expression (Fig. 4A, 1st and 2nd lanes versus 3rd and 4th  lanes). Phosphorylation of CREB was mildly increased by AC VI gene transfer (Fig. 4A, 5th lane) but increased to a much greater degree upon stimulation with isoproterenol or NKH447 (6th and 7th lanes).
To establish that increased ATF3 protein was present in the nuclei, we used anti-ATF3 antibody to detect ATF3 in cardiac myocytes infected by Ad.AC VI . We detected nuclear presence of ATF3 following Ad.AC VI gene transfer but much less in uninfected cells (Fig. 5).
The Effects of AC VI Gene Transfer on PLB Expression Involve ATF3-Our results suggest that AC VI gene transfer reduces PLB gene expression via increased expression of ATF3. To confirm this, using EMSA analysis we examined whether ATF3 binds to a CRE-like motif derived from the PLB promoter. Upon incubation of [␣-32 P]dCTP-labeled oligos containing the CRE-like motif with nuclear extracts from Ad.AC VI -infected cardiac myocytes, multiple DNA-protein complexes were formed (Fig. 6, lane 1). A subset of these complexes was found to supershift when an anti-ATF3 antibody was added to the mixture after complexes had formed (Fig. 6, lane 2). A shift was also seen when the anti-ATF3 antibody was incubated (overnight) with the nuclear extract prior to adding the labeled probe (Fig. 6, lane 4). On the other hand, we did not detect supershifted bands after adding an anti-CREM antibody (a gift from Dr. Marc Montminy, Salk Research Institute) under these same conditions (Fig. 6, lanes 3 and 5). CREM can act as either a transcriptional activator or suppressor by interacting with CRE or CRE-like element. A transcriptional suppressor of CREM was identified in cardiac myocytes (15). Likewise, under these two conditions, we did not detect supershifted complexes after adding the anti-ATF3 antibody complexes formed with nuclear extracts from Ad.EGFP-infected cardiac myocytes (Fig.  6, lanes 9 and 11).
To determine the specificity of oligonucleotide binding of the complexes, unlabeled oligos as well as oligos with the mutated CRE site were used in the EMSA. All DNA-protein complexes were abolished with an excess amount of unlabeled wild type oligos but not with the same amount of unlabeled mutant oligos (Fig. 6, lanes 6 versus 7 and lanes 13 versus 14). Moreover, no protein was found bound to the mutated CRE probe (Fig. 6,  lanes 15 and 16). Taken together, these results offer strong support that ATF3 specifically binds to the CRE-like motif found in the PLB promoter.
Next, we examined whether ATF3 directly alters PLB promoter activity. In a transient transfection assay, the minimal PLB promoter-luciferase reporter construct was co-transfected into cardiac myocytes with an ATF3 expression plasmid (16). ATF3 inhibited PLB promoter activity by Ͼ80% (p Ͻ 0.0006). Furthermore, when the CRE-like element was mutated, ATF3 no longer suppressed PLB promoter activity (Fig. 7). These results confirmed our prediction that ATF3 not only bound to the CRE-like element but also suppressed the minimal PLB promoter activity. DISCUSSION AC VI Alters the Expression of PLB and ATF3-In the present study, we have determined the effects of AC VI gene transfer on CRE-mediated gene transcription. We found that AC VI reduces PLB expression and increases the expression of ATF3, a transcriptional suppressor in the CREB/ATF3 family. Upon stimulation with isoproterenol or NKH477, PLB expression is fur-ther reduced, and ATF3 expression is further increased. We also demonstrated that ATF3 binds to the CRE-like element in the minimal promoter of PLB and reduces the activity of the PLB promoter. In contrast, stimulation through the ␤-adrenergic receptor (isoproterenol) or direct stimulation of AC (NKH447) had directionally opposite effects on PLB expression, and no effect on ATF3 expression.
Expression and Regulation of PLB Gene-The 5Ј-flanking promoter sequences of the PLB gene have been partially characterized from chicken (17), mouse (18), human (19), and rat (12,20). The minimal promoter activity of rat PLB is in a region from Ϫ159 to ϩ64 bp (Ϫ159/ϩ64) (12). Analysis of rat PLB 5Ј-untranslated region revealed a conserved phospholamban promoter element 1 (Ϫ159/Ϫ125), GATA (Ϫ101/Ϫ96), CAAT box (Ϫ82/Ϫ78), M-CAT-like (Ϫ67/Ϫ62), TATA-like box (Ϫ51/ Ϫ47), and E box (Ϫ11/Ϫ6). However, the CRE-like element (TGATGTCA) at Ϫ50/Ϫ57 was not characterized previously. Two CRE-like elements in the chicken PLB promoter were recognized (17), but their roles were not characterized. We have identified and characterized the CRE-like element in the rat PLB promoter, and we showed that it is critical for the basal activity of the PLB promoter. Mutation of this element leads to loss of PLB promoter activity. We also showed that AC VI gene transfer reduces the expression of PLB via increased expression of ATF3 with subsequent binding to the CRE-like element of PLB.
Significance of Reduced PLB Expression by AC VI -PLB plays an important role in the regulation of sarcoplasmic reticulum calcium cycling by binding to sarcoplasmic reticulum Ca 2ϩ -ATPase, thereby inhibiting its activity (21). Reduction of PLB through antisense suppression or genetic ablation increased sarcoplasmic reticulum calcium cycling and improved cardiac myocyte survival and cardiac function (22,23). In contrast, overexpression of PLB through an increase in PLB promoter activity was associated with hypertrophic cardiomyopathy (24), and cardiac-directed expression of PLB resulted in depressed contractility and cardiomyopathy after ␤-adrenergic receptor stimulation (25). Reduced PLB expression resulting from AC VI expression would be expected to have a favorable impact on cardiac function and may be an important mechanism by which AC, in contrast to other elements that increase intracellular cAMP, has a favorable effect in the failing heart.
Expression and Function of ATF3-Expression of ATF3 is FIG. 7. ATF3 inhibits transcription of the PLB promoter through a CRE-like site. The normal PLB promoter or a mutant PLB promoter (mPLB) was linked to a luciferase reporter and co-transfected with pCGATF3 (ATF3) or pCG (Con) into cardiac myocytes. ATF3 had a suppressive effect on minimal PLB promoter activity. Mutation of the CRE-like element in the mini-PLB promoter was associated with loss of PLB promoter activity, and the suppressive effect of ATF3 was diminished.
induced by growth-stimulating factors (fibroblast growth factor and epidermal growth factor), cytokines (interleukin-4), and stress signals (ionizing radiation, ultraviolet light, and methyl methanesulfonate) (11). In human neuroblastoma SK-N-MC cells, it was found that expression of ATF3 was induced by forskolin (26). However, forskolin did not induce the expression of ATF3 in cardiac myocytes. Expression of ATF3 can also be induced by adenovirus 12S-E1A protein (27). Because the E1 region encoding E1AB proteins was deleted from the recombinant adenovirus vector used in the present study, there is no expression of E1 proteins from Ad.AC VI or Ad.EGFP. Furthermore, our findings were specific to the vector expressing AC VI and were not seen with Ad.EGFP, confirming that the effect is transgene-specific.
Precise details regarding the biological roles of ATF3 is an evolving area, with some insights provided from experiments in which ATF3 is overexpressed or deleted. For example, overexpression of ATF3 in neuronal cells inhibited c-Jun N-terminal kinase-induced neuronal death and induced neurite elongation (28) and co-expression of ATF3 with c-Jun increased neurite sprouting (29). In cardiac myocytes, ATF3 overexpression was associated with reduced doxorubicin-induced apoptosis (30), an effect also observed in endothelial cells subjected to tumor necrosis factor-␣-induced apoptosis in endothelial cells (31). However, ATF3 accelerates camptothecin-induced apoptosis in human epithelioid carcinoma cells and fibrosarcoma cells (32). When ATF3 is expressed in cardiac myocytes chronically (transgenic mice), atrial and ventricular hypertrophy result (33). Constitutive expression of ATF3 in murine pancreatic ␤ cells was associated with abnormal islet formation and ␤-cell deficiency (34), and targeted deletion of ATF3 provided partial protection from cytokine-induced ␤-cell apoptosis.
Thus, increased ATF3 may be antiapoptotic and, when expressed chronically in all cells in the heart, is associated with myocardial hypertrophy, histological abnormalities, and reduced contractile function. However, in AC VI transgenic mice of three different lines, examined for up to 20 months of age, we do not find myocardial hypertrophy or any deleterious effects (2)(3)(4)6), underscoring the differences between ATF3 overexpression per se versus AC VI overexpression.
What Are the Endogenous Targets of AC VI ?-AC VI increases the expression of ATF3 and increases phosphorylation of CREB when stimulated with isoproterenol or NKH477. Both transcriptional factors regulate gene transcription through CRE or CRE-like elements. ATF3 has more flexibility in binding to CRE or CRE-like elements (16). It will be of interest to determine how ATF3, a suppressor, and phospho-CREB, an activator, differentially regulate their targets in response to AC VI. This will require identifying the endogenous targets of AC VI . Chromatin immunoprecipitation (CHIP) and CHIP-chip are powerful means for determining endogenous targets of transcriptional factors (35,36). By applying these methods in combination with anti-ATF3 or anti-CREB antibody, we will identify true target genes of AC VI.
In conclusion, overexpression of AC VI in neonatal rat cardiac myocytes selectively reduces CRE-mediated gene transcription in a cAMP-independent manner via increased expression of ATF3 protein, a transcriptional suppressor. This effect of AC VI on CRE-mediated transcription is opposite the effects from direct activation of ␤-adrenergic receptor or adenylyl cyclase using pharmacological agents. These findings indicate that AC VI has effects on gene transcription that are not directly dependent on cAMP generation.