Identification of a Novel Serum Response Factor Cofactor in Cardiac Gene Regulation*

The transcription factor serum response factor (SRF) plays an important role in the regulation of a variety of cardiac genes during development and during adult aging. A novel SRF cofactor, herein called p49/STRAP, for SRF-dependent transcription regulation-associated protein, was recently identified in our laboratory. This protein interacted mainly with the transcriptional activation domain of the SRF protein and was found to bind to SRF or to the complex of SRF and another cofactor, such as myocardin or Nkx2.5. The expression of p49/STRAP affected the promoter activity of SRF target genes in a non-uniform manner. For example, p49 activated MLC2v and cardiac actin promoters when it was co-transfected with SRF, but it repressed atrial natriuretic factor promoter activity, which was strongly induced by myocardin. The p49/STRAP mRNA was observed to be highly expressed in fetal, adult, and senescent human hearts, and also in hearts of young adult and old mice, suggesting that p49/STRAP may be an important SRF cofactor in the transcriptional regulation of mammalian cardiac muscle genes throughout the life span.

It is well appreciated that the mammalian adult heart undergoes a number of changes with advancing age (1)(2)(3). Recent studies indicate that one of the key transcription factors in muscle and other tissues, serum response factor (SRF), 1 is implicated in the regulation of cardiac genes during development and during adult aging (4 -7). SRF is a member of the MADS (MCM1, Agamous, Deficiens, SRF) family of transcription factors that regulate a number of immediate early and muscle-specific genes, and also serves to regulate cell prolifer-ation, cell size, and cell survival (4 -11). SRF forms dimers and recruits SRF cofactors or SRF-binding proteins when it binds to the serum response element (SRE), which is located in the promoter region of each of its target genes (8 -12). SRF is highly expressed in the heart during embryonic and early postnatal development, and it is mildly increased by ϳ20% from postmaturational adulthood to senescence (4 -7, 12). The mRNA levels of a number of SRF target genes, including atrial natriuretic factor, ␣-myosin heavy chain, and sarcoplasmic reticular calcium ATPase have also been reported to undergo changes during postnatal cardiac development and during senescence (4 -7, 12-14). In a transgenic mouse model in which the human SRF gene was mildly overexpressed in the heart, cardiac changes resembling those that have been observed during adult aging in terms of myocardial function, morphology, and gene expression are observed in young adulthood (7). The mildly increased cardiac-specific SRF expression apparently up-regulates some SRF target genes, whereas it down-regulates others in the heart (7). This bidirectional pattern of altered gene expression following mild SRF up-regulation suggests that possibly other transcription regulators, including perhaps certain SRF cofactors, may pose either positive and/or negative modulatory effects on the activation of SRF target genes (7, 14 -18).
SRF has been reported to exhibit functional interactions with a number of SRF cofactors and/or binding proteins in the regulation of SRF target genes (15)(16)(17). These interactions likely modulate SRF function and may also enable SRF to mediate tissue-specific regulation at different developmental stages (18 -20). To date, a number of SRF cofactors, including the TCF family of proteins, the SAP protein myocardin, Nkx 2.5, and Hop, have been identified, and their various functions in cardiac development have been investigated (20 -23). Fewer studies have reported on the role of SRF cofactors in the regulation of cardiac genes during adult aging and senescence.
In an effort to identify potential SRF cofactors that may contribute to cardiac gene regulation during aging, we performed yeast two-hybrid screening using both SRF NH 2 -and COOH-terminal portions as bait. Here we report the identification of a novel transcription regulator, p49/STRAP (SRF-dependent transcription regulation associated protein), isolated with the SRF COOH-terminal bait, and propose a model of gene regulation by SRF, p49/STRAP, and other cofactors. The protein p49/STRAP displayed functional cooperation with SRF and myocardin, and repressed the atrial natriuretic factor promoter activity, which was strongly induced by myocardin. The p49/STRAP mRNA was highly expressed in fetal and postnatal hearts, and was increased by ϳ45% in old compared with young adult hearts. The age-and cardiac-specific changes of p49/ STRAP and other SRF cofactors in senescence may reflect a dynamic pattern of well regulated gene expression during the process of adult aging.

EXPERIMENTAL PROCEDURES
Yeast Two-hybrid System-The bait construct containing the NH 2 terminus of the SRF protein (1-244 residues), and another containing the COOH terminus of the SRF protein (247-499 residues) were each constructed by fusing the SRF fragments to the GAL4 DNA-binding domain in the pGBT9 vector (Clontech). The constructs were then used to screen an EML cDNA library (24) and a human heart cDNA library (Clontech) with a method described by Zhang et al. (24). The cDNA clones representing potential SRF-interacting proteins were sequenced and were compared with the GenBank TM database by using Blast Search.
Cloning of Full-length Coding Region Sequence of p49/STRAP-Two independent cDNA clones (G65, G78), which were isolated with SRF COOH-terminal bait construct, matched a single gene in the GenBank TM database, which has not been previously characterized (we named it p49/STRAP). The full-length coding region of the mouse and human p49/STRAP gene were amplified by PCR using heart cDNA samples (Clontech). The rapid amplification of cDNA ends PCR was performed to confirm the full coverage of the p49 coding region sequence as provided in this report. The sequences have been submitted to GenBank TM with the accession numbers AY611629 and AY611630.
Antibodies and Plasmid Constructs-A polyclonal antibody against a peptide (KSKKGTEDALLKNQRRAQ) of the p49/STRAP protein was commercially generated by standard procedures (Genemed Synthesis Inc.). The p49/STRAP antibody was shown to be specific for p49/STRAP in whole-cell lysates by Western blotting with competing peptide (see Fig. 2B). Other antibodies that were employed include HA.11 (Covance), FLAG (Sigma), and SRF (Santa Cruz).
In Vitro Protein Interaction Assays-GST fusion proteins were purified with glutathione-conjugated agarose beads (Sigma). The p49/ STRAP protein was translated in vitro using a TNT quick coupled transcription/translation system (Promega) and pcDNA3-p49/STRAP plasmid in the presence of [ 35 S]methionine (Amersham Biosciences) according to the manufacturer's instruction. In the in vitro binding assay, 2 g of agarose-bound GST fusion proteins were incubated with [ 35 S]methionine-labeled p49/STRAP protein for 1 h at 4°C in NETN buffer (20 mM Tris⅐HCl, pH8.0, 1 mM EDTA, 1%Nonidet P-40, 150 mM NaCl, 0.5% glycerol, 1ϫ protease inhibitor mix). Beads were washed four times with NETN buffer and then analyzed on SDSpolyacrylamide gels, and the binding activity was detected by autoradiography.
Northern Blotting and Western Blotting-Healthy young adult (3month-old) and old (20-month-old) mice were obtained from colonies maintained by the NIA, the National Institutes of Health, under contractual agreement with Harlan Sprague-Dawley, Inc. (Harlan, IN). The human heart mRNA samples were obtained from Biochain Institute (Hayward, CA). The human tissue blot, human cardiovascular system blot, and mouse tissue blot were purchased from Clontech. The Northern blotting and Western blotting were performed as described (6,25). The studies were conducted with Institutional Review Board approval and in accordance with the NIH Guiding Principles for Research Involving Animals and Human Beings.
Transfection Assays-Transient transfections were carried out with FIG. 1. Alignment of mouse and human p49 protein sequences. The sequence underlined and in bold represents the overlapping sequence of the two yeast cDNA clones, which perfectly matches the p49/STRAP sequence.

p49/STRAP, a New SRF Cofactor
the Lipofectamine and Plus reagents as previously described (25). At ϳ4 h after the transfection was initiated, the NIH3T3 cells were placed in Dulbecco's modified Eagle's medium with 10% fetal calf serum and incubated overnight. The cells were then cultured in Dulbecco's modified Eagle's medium with 0.1% fetal calf serum for another 24 h and then placed in Dulbecco's modified Eagle's medium with 20% fetal calf serum for an additional 3.5 h. Firefly luciferase activity was measured as relative light units. To control for variability, the number of relative light units from individual transfection experiments was normalized by measuring Renilla luciferase activity expressed from a cytomegalovirus promoter-driven vector in the same samples. Individual transfection experiments were carried out in triplicate, and the results were reported as mean firefly luciferase/Renilla luciferase activity (mean Ϯ S.D.) from representative experiments.
Subcellular Localization-The expression plasmid (pLP-EGFP-p49/ STRAP) containing the EGFP-p49/STRAP fusion protein was generated using the Creator DNA cloning system (Clontech). At ϳ30 h after the transfection, the expression of EGFP-p49/STRAP fusion protein was examined by fluorescence microscopy using a Zeiss Deconvolution microscope with AxioVison version 3.1 software.
Electrophoretic Mobility Shift Assays-Electrophoretic mobility shift assays were performed as described by using the SRE consensus oligonucleotide, which is derived from the c-fos promoter (5Ј-GGATGTC-CATATTAGGACATCT-3Ј) (6). The in vitro translated SRF and p49/ STRAP, as well as the protein from NIH3T3 cells transfected with pAd-Track-CMV-SRF and pcDNA3-HA-p49/STRAP plasmids, were employed for electrophoretic mobility shift assays.

P49/STRAP Is a Novel SRF-binding Protein-Sequencing
analysis and Blast Search against the GenBank TM database revealed that two independent yeast cDNA clones, which were isolated with the SRF COOH-terminal bait, matched a single gene in the GenBank TM database, the function of which remains uncharacterized.
The full-length coding region sequence of this gene was amplified from mouse cardiac cDNA by PCR. This gene encoded a 441-amino acid protein with a predicted mass of 49-kDa; therefore we named it p49/STRAP. The overlapping sequence of the two cDNA clones covered 132 amino acids of the p49/STRAP protein (Fig. 1), indicating that this domain was important for its binding to the SRF protein. The human p49/STRAP gene was also amplified from human cardiac cDNA. Sequence align- p49/STRAP, a New SRF Cofactor ment revealed 67% homology between human and mouse p49/ STRAP (Fig. 1).
To determine whether p49/STRAP processes conserved protein domain or motif, the p49/STRAP sequence was compared with the NCBI Conserved Domain database, Pfam Protein Family database, and SWISS-PRO Protein database. However, no existing domain or motif matched the p49/STRAP gene, indicating that p49/STRAP may belong to a new class of as yet uncharacterized proteins.
p49/STRAP Is Expressed in the Heart and Other Tissues-Northern blotting revealed that three p49 isoforms were detected in mouse tissues. Among the tissues tested, mouse heart, liver, kidney, and testes had a high level of p49 expression ( Fig. 2A).
Western blotting using mouse tissue lysates demonstrated that the p49/STRAP antibody recognized a 49-kDa protein (Fig.  2, A and B). This antibody-protein binding could be blocked by the p49/STRAP peptide (Fig. 2B), indicating that the 49-kDa protein is the main protein product of this gene.
Among human tissues, human heart and skeletal muscle had the highest levels of p49/STRAP mRNA, whereas the brain and lungs had the lowest levels of p49/STRAP mRNA (Fig. 2D). Unlike the mouse tissue, human tissue had only one major p49/STRAP transcript, at ϳ2.2 kb. In the human cardiovascular system, the p49 mRNA level is higher in the adult than in the fetal heart (Fig. 2E).
p49/STRAP Is Increased in Expression in the Aging Heart and in the Heart of Cardiomyopathy-To determine whether there might be an age-related change of p49/STRAP expression in the heart, the expression of p49/STRAP was determined in the myocardium of young adult and old mice. Western blotting analysis revealed that the cardiac p49/STRAP protein level in 20month-old mice was ϳ45% higher than in 3-month-old mice (Fig.  3B). The p49/STRAP mRNA was also increased in the heart of the 20-month-old compared with 3-month-old mice (Fig. 3A). In humans, the age-related change was also observed, as shown in Fig. 3C, p49/STRAP mRNA was increased in 77-year-old individual compared with 30-year-old individual.
To test whether the expression of p49/STRAP might be associated with a pathological condition, the level of p49/STRAP expression was examined in the myocardium of SRF transgenic mice that suffered from cardiomyopathy (6). Northern blotting revealed a 4-fold increase of p49/STRAP in the heart of adult transgenic compared with that of wild-type mice (Fig. 3D).
p49/STRAP Interacts with SRF in Vitro and in Vivo-To confirm the physical interaction between p49/STRAP and SRF proteins, we first transformed both the SRF bait plasmid containing SRF COOH terminus and the yeast plasmids containing p49/STRAP protein back into yeast cells. The transformants grew on the -Trp/-Leu/-His triple dropout plates, indicating that the two proteins interact in the yeast cells. Then we tested whether p49/STRAP binds to SRF in vitro. As shown in Fig. 4A, 35 S-labeled in vitro translated p49/STRAP protein bound to both immobilized GST-SRF (wild-type) and GST-dmSRF protein (a double mutant form of SRF) (25) but not to GST protein alone, indicating that p49/STRAP interacts with both the wild-type form and the mutant form of SRF proteins and that point mutations within the DNA binding domain of SRF did not affect the interaction between SRF and p49/STRAP.
To test whether the interaction between p49/STRAP and SRF could occur in vivo within mammalian cells, we transfected NIH3T3 cells with plasmid constructs containing wildtype and two mutants of p49/STRAP tagged with HA epitope, and a plasmid construct containing Flag-SRF. As shown in Fig.  4B, the wild-type and both p49/STRAP mutants bound to SRF, indicating that p49/STRAP indeed interacted with SRF in vivo.
Analysis of the previously isolated two yeast cDNA clones that interacted with the SRF bait plasmid suggested that a domain from residue 53 to 185 in the p49/STRAP protein is important for its interaction with the SRF protein. However, the in vivo p49/STRAP-SRF interaction data indicated that the protein fragment from residue 1 to 91 in the p49/STRAP protein may possibly also be sufficient for such an interaction.
SRF Simultaneously Binds to p49/STRAP and Other Cofactors-Because p49/STRAP binds mainly to the COOH terminus of SRF, we hypothesized that other SRF cofactor(s) that interact with the NH 2 terminus of the SRF protein could potentially also bind to SRF at the same time. To test this hypothesis, we performed a parallel transfection assay. One set of cells was transfected with p49/STRAP, SRF, and myocardin expression plasmids; the other set of cells was transfected with p49/STRAP, SRF, and Nkx2.5 expression plasmids. As shown in Fig. 4, C and D, a three-protein complex was precipitated in each case, indicating that SRF could simultaneously interact with both p49/ STRAP and another cofactor, either myocardin or Nkx2.5.
p49/STRAP Is a Nuclear Protein That Does Not Bind to SRE-SRF is a nuclear protein with three nuclear localization signal sequences in its coding region (27). To identify whether p49/STRAP also localizes to the nucleus, an expression plasmid containing the EGFP-p49/STRAP fusion protein was transfected into NIH3T3 cells. The fusion protein was localized within the nucleus in the vast majority of cells that were examined (Fig. 5). Because p49 was isolated based on its ability to interact with SRF COOH terminus, which contains the SRF transcription activation domain, it was not expected that p49/ STRAP would form a ternary complex with SRF at the site of SRE. Electrophoretic mobility shift assays using proteins from both cell lysate and in vitro translated p49/STRAP and SRF revealed that no additional band was shifted by anti-SRF or anti-HA antibodies (data not shown) in the presence of a DNA fragment corresponding to c-fos promoter that contains SRE, thus confirming that p49/STRAP did not form the ternary complex with SRF at the site of SRE.
p49/STRAP Modulates the Transcriptional Activation of Cardiac Genes-To explore the biological effect of p49/STRAP and the consequences of an elevation of p49/STRAP expression on cardiac gene expression, the promoter activities of c-fos SRE, MLC2v, cardiac actin, and atrial natriuretic factor were utilized as indicators in cell transfection assays. As shown in Fig.  6A-C, p49/STRAP activated SRE-luciferase, MLC2v, and cardiac actin promoter activity, respectively, mainly in cooperation with SRF. However, p49 apparently effectively repressed atrial natriuretic factor promoter activity, which was strongly induced by myocardin (Fig. 6D). DISCUSSION In the present study, we report the identification and characterization of a new gene, p49/STRAP, as a novel SRF-dependent transcription regulator. This gene was initially isolated from yeast two-hybrid screening based on its ability to bind to the SRF COOH terminus. The subsequent protein-protein binding assays further confirmed that p49/STRAP is a SRFbinding protein. In addition, we observed that p49/STRAP could form a protein complex with SRF and also with other SRF cofactor(s), such as myocardin or Nkx2.5. The interaction of p49/STRAP with SRF alone or with both SRF and other cofactors can clearly affect the activation of cardiac gene promoters in diverse ways. Both the p49/STRAP mRNA and protein are highly expressed in the mouse and human heart, and their expression levels increase with advancing age, suggesting that p49/STRAP may play a significant role in the regulation of cardiac genes during adult aging.
The sequence of the p49/STRAP protein is conserved between human and mouse. However, to date the p49/STRAP protein sequence does not match any known conserved protein domain or known motif that has been deposited in several public databases, including the NCBI conserved domain database and Pfam Protein Family database. This suggests that p49/STRAP may belong to a new class of proteins that are yet to be determined. Inasmuch as the p49/STRAP protein does not bind to DNA and does not form a protein complex with SRF at the SRE site, it is likely that p49/STRAP modulates SRF function primarily through its interactions with SRF in the SRF transcriptional activation domain.
SRF target genes are regulated in a complex manner that is partly because of the participation of multiple SRF-binding proteins in the co-regulation of SRF target genes. SRF-binding proteins include the TCF family proteins (26 -30) and other transcription factors/regulators. With the increasing number of SRF cofactors being identified, one emerging question is how SRF cofactors might be recruited by SRF for the SRF-dependent transcriptional regulation. Molecular dissection of the functional domain of SRF has revealed that SRF has two major parts. The NH 2 terminus has the DNA-binding domain and the dimerization domain, while the COOH terminus has the transcriptional activation domain (31,32). Each part constitutes approximately half of the protein. Most of the SRF cofactors (except ATF6) that have been identified apparently predominantly bind to the SRF NH 2 terminus (18,33,34). However, we and others have observed that a mutant form of SRF, which substitutes amino acids in the DNA binding domain and thus prevents the proteins from binding to DNA, can still significantly affect the expression of SRF target genes, suggesting that the SRF transcriptional activation domain plays a critical role in the regulation of SRF target genes (25).
Our finding of the co-immunoprecipitation of two protein complexes containing three proteins, "p49/STRAP-SRF-myocardin" and "p49/STRAP-SRF-Nkx2.5," indicates that p49/ STRAP is able to interact with SRF in the presence of another cofactor while they are both bound to SRF. It is likely that multiple cofactors may interact with SRF at any given time within a cell, as shown in Fig. 7. This intriguing model of the dynamic pattern of the interactions among SRF and its cofactors warrants further study. On the NH 2 terminus of SRF, many cofactors may competitively bind to SRF, including the TCF family proteins Elk1, SAP-1, and SAP-2 (22,28), the SAP family protein myocardin (18), the GATA family protein GATA4 (16), Nkx2.5 (35), and Hop (36). On the COOH terminus of the SRF protein, it is plausible that p49/STRAP and ATF6 may also modulate the binding of one another to SRF (37). The concept of multiprotein functional complexes and regulation of intracellular and intercellular processes has been evolving (38,39). The participation by multiple cofactors and the exchange of the cofactors for each other at any given time may potentially account for the complex patterns of SRF target gene expression in different tissues, at different development stages, and under different stress conditions (40). The dynamic interactions among SRF, p49/STRAP, and other cofactors may  6. A, effect of p49 and SRF expression on the SRE-luciferase activity. B, regulation of cardiac actin promoter activity by p49 and SRF expression. C, regulation of MLC2v promoter activity by p49 and SRF. D, p49 represses the atrial natriuretic factor promoter activity, which is strongly induced by myocardin.
p49/STRAP, a New SRF Cofactor help to determine whether and how much SRF activates or represses its target gene(s).
It has been documented that the RNAs of certain SRF cofactors, such as TCF family proteins (Elk-1, SAP-1, and SAP-2), are present at similar relative levels in many different tissues, suggesting that TCF proteins may serve as universal instead of tissue-specific cofactors (21)(22)(23)33). However, some other cofactors, including myocardin and Nkx2.5, are highly expressed in the heart, and therefore may serve as tissue-specific cofactors. It is plausible that the recently identified Hop protein, which also modulates SRF activity, may also have a similar role (21,36). In addition, SRF and myocardin have been reported to be well expressed in the heart during embryogenesis and postnatal development (18,19,35). These data indicate that p49/STRAP, SRF, and myocardin are well expressed in the heart during different developmental stages and suggest that p49/STRAP may also contribute significantly to cardiac gene expression.
We have observed that the cardiac expression of p49/STRAP and myocardin are both increased during adult aging. We previously reported that SRF expression was increased by ϳ20% in the heart of the senescent compared with young adult rodents (5,7). In contrast, some other transcription factors, which are SRF cofactors, including Nkx2.5 and GATA4, have been reported to be decreased during adult aging (41). The ageand cardiac-specific increase of some SRF cofactors and decrease of other cofactors are unlikely to be a coincidence and rather may reflect the dynamic pattern of precisely regulated gene expression during the process of adult aging. Recent studies using sophisticated molecular methods have revealed dynamic patterns of gene expression during aging in animals across the species and support the concept that the change in gene expression during adult aging is likely because of "selective gene regulation" rather than random passive decline (42).