![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,
Anton L. Bryantsev, and
Richard M. Cripps*
Department of Biology, University of New Mexico, Albuquerque, New Mexico 87131
Received 29 June 2007/ Returned for modification 17 August 2007/ Accepted 14 December 2007
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The transcription factors that mediate muscle cell differentiation are also broadly conserved. Foremost among these proteins are members of the myocyte enhancer factor 2 (MEF2) family of transcriptional regulators (reviewed in reference 6). MEF2 proteins dimerize and bind to DNA via conserved MADS and MEF2 protein domains, and MEF2 proteins recognize an AT-rich sequence (2), which is found in the promoters and enhancers of numerous muscle-specific genes. MEF2-encoding genes are found in all animal genomes, from single-gene copies in Caenorhabditis elegans (named CeMef2) and Drosophila melanogaster (named Mef2), to four genes in mammals (named mef2a to -d).
Consistent with the pervasiveness of MEF2 sites in muscle-specific genes and the conservation of this gene family during evolution, mutational studies in several animal models have established important roles for MEF2 factors in muscle differentiation. In Drosophila, inactivation of the single Mef2 gene resulted in a complete failure of differentiation for all muscle lineages (7, 27). In mice, inactivation of mef2c resulted in a failure of normal cardiac and visceral muscle development (5, 31, 30). Furthermore, inactivation of murine mef2a caused cardiac failure shortly after birth (38). Also, expression of dominant-negative isoforms of MEF2 predicted to inhibit all MEF2 function resulted in the inhibition of skeletal myogenesis in mammalian cells in vitro (40) and in interference in cardiac development in vivo (23). Finally, a point mutation in the human gene encoding MEF2A is associated with susceptibility to cardiac disease (19). These studies present a strong argument for the importance of MEF2 proteins in muscle development and disease.
Despite the critical requirement of MEF2 for muscle differentiation, MEF2 proteins are not themselves capable of activating the myogenic program in naive cells in tissue culture (36). Similarly in Drosophila, ectopic expression of Mef2 in the ectoderm caused the activation of some muscle-specific markers, and yet failed to induce myogenesis at high levels (31). These findings suggested that MEF2 must act with specific cofactors in order to control myogenesis. Much progress has subsequently been made in defining both positive and negative cofactors for MEF2, in a variety of muscle tissues. MEF2 collaborates positively with skeletal muscle-specific members of the basic helix-loop-helix family of transcriptional regulators, including MyoD and myogenin, to control skeletal myogenesis (36, 37). In addition, the SAP domain proteins Myocardin and MASTR also stimulate the transcriptional activity of MEF2 in cardiac and skeletal muscle tissue, respectively (11). MEF2 proteins also interact with ubiquitous factors, such as the p300 coactivator (46) and members of the histone deacetylase family of transcriptional repressors (32).
In Drosophila, there is relatively little direct data concerning the identification of cofactors that might collaborate with MEF2 to control myogenesis. Nevertheless, a number of enhancers for muscle-specific genes have been described, many of which contain MEF2 binding sites that are critical for full gene activation (3, 13, 29, 34, 42). Among these, we recently described the promoter for the Act57B actin gene in Drosophila (24). Act57B is the predominant embryonic muscle actin (15, 48), being expressed at high levels in all of the major muscle lineages and also being one of the earliest markers of muscle differentiation in the embryo (24). We found that activation of Act57B expression in all embryonic muscle lineages was controlled by a proximal 595-bp promoter element. Within this region, full transcriptional activation required the integrity of a single MEF2 binding site at position –209 relative to the transcriptional start site, validating the importance of MEF2 as a direct regulator of myogenesis in this system. This finding was also supported by the study of Sandmann et al. (44), who showed, using microarray analysis, that Act57B expression was strongly downregulated in Mef2 mutants. Furthermore, these authors utilized chromatin immunoprecipitation-microarray analyses to identify the MEF2 site at position –209 as an in vivo target of MEF2 (44).
Our studies also demonstrated that an 
300-bp region distal to the MEF2 site was critical for full Act57B activation in embryonic muscle tissue, suggesting the existence of a MEF2 cofactor that interacted with this distal sequence. This analysis permitted us to carry out a directed approach to the identification and functional characterization of a MEF2 collaborating factor in Drosophila. We delineate here sequences within the 300-bp distal region that are required for full gene activation in vivo, and we show that a nuclear factor binds to this region with the characteristics identical to that of the C2H2 zinc finger transcriptional regulator chorion factor 2 (CF2) (22). We further show that MEF2 and CF2 can each activate Act57B expression in vitro and in vivo and that these factors function synergistically to maintain high levels of actin expression in the Drosophila embryo. More globally, MEF2 and CF2 synergistically activate other muscle structural genes, and loss of CF2 function results in reductions in the expression levels of several muscle structural genes, including Act57B. Our findings describe the first collaborating factor for MEF2 in the Drosophila system and further delineate the network of transcriptional events required for normal muscle development.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
27 was also provided by T. Hsu and was described earlier (21a). To select Df(2L)27 homozygotes, this strain was balanced over a GFP-expressing CyO chromosome. P-element-mediated transformation was carried out as described by Rubin and Spradling (43) by injection of DNA mixtures into yw embryos. At least three transgenic lines of each construct were studied, and representative images of stained embryos are shown.
Immunohistochemistry and in situ hybridization. Antibody stains were performed essentially as described by Patel (41). The primary antibody was mouse anti-β-galactosidase (Promega Corp., Madison, WI) at a final concentration of 1:1,000. Antibody detection was performed by using a Vectastain Elite staining kit and diaminobenzidine reagent (Vector Laboratories, Burlingame, CA). In situ hybridization was performed according to the method of O'Neill and Bier (39): an antisense Act57B riboprobe labeled with digoxigenin (Roche Molecular Biologicals, Indianapolis, IN) was generated according to the method of Kelly et al. (24) and detected by using NBT/BCIP substrate (Vector Laboratories). For the ectopic expression assays, all embryos at stage 14 were assessed for activation of Act57B in the ventral ectoderm. Percentages in the text refer to a minimum of 14 embryos studied. Stained embryos were mounted in 80% glycerol and photographed by using an Olympus BX51 compound microscope with differential interference contrast optics and 35-mm film. Photographic slides were scanned and assembled by using Adobe Photoshop.
DNA methods. Act57B promoter fragments were generated by PCR and cloned into the pGEM-T Easy PCR cloning vector (Promega Corp.). Promoter fragments were then subcloned into the reporter construct pCaSpeR-hs-AUG-βgal (pCHAB) (47a) for generation of transgenic flies. Creation of –593/+2, –593/-245 and –270/+2 Act57B-lacZ constructs was described previously (24). Sequences of PCR primers for other constructs were as follows: for –521/+2, 57BC9-1/3 5'+ (5'-GGGAATTCATATAGCCGATATGGCCG-3') and 57B Ex-PCR II (5'-GGCTCGAGCTAAAGTATCGCCGCGTTGGT-3'); and for –390/+2, 57BC9-2/3 5'+ (5'-GGGAATTCGATCGTGAGCAGGCAGCC-3') and 57B Ex-PCR II. Underlined sequences indicate those introduced for cloning purposes.
To generate the Troponin I gene reporter, a 2.9-kb genomic stretch of DNA corresponding to the –700 to +2199 area of TnI (IRE2.9 [34]) was PCR amplified with forward 5'-GAAGGTCTCCAAATACGAAA-3' and reverse 5'-GCTGTTGTTGTTTATTGACTTC-3' primers. This product was inserted into pCHAB that had been linearized at the EcoRI site and subsequently blunt-ended with Klenow enzyme. The orientation of the insert in the final construct, pIRE2.9-CHAB, was confirmed with restriction digest analysis and sequencing.
The Mhc reporter construct, p
MHC-LacZ1, harboring the genomic sequence from positions –2759 to + 2124 of the Myosin heavy chain gene was kindly provided by S. I. Bernstein (San Diego State University, San Diego, CA) (21).
Mef2 and Cf2 expression constructs for cell culture were generated in similar manners. The cDNAs of each transcription factor was amplified from plasmid templates (pBluescript KS Mef2 [27]; pNB40-Cf2, [22], provided by T. Hsu, Medical University of South Carolina) using the following pairs of specific primers: forward (5'-ATGGGCCGCAAAAAAA-3') and reverse (5'-CTATGTGCCCCATCCGCCC-3') primers for Mef2 and forward (5'-ATGATAAAGTCCACCACG-3') and reverse (5'-CTAGAGCGGATGCAGCTTG-3') primers for Cf2. PCR products were blunt cloned into the pPac-Pl expression vector (Drosophila Genome Resource Center) at the EcoRV restriction site, and the correct orientation of inserts was verified by restriction digestion and sequencing.
Protein purification and synthesis. Nuclear extracts from 12 to 20 h Drosophila embryos were prepared as described by Lichter and Storti (26). A preparation of CF2 protein was generated by in vitro transcription and translation of the pNB40-Cf2 plasmid (22) using SP6 RNA polymerase and the TNT coupled reticulocyte lysate system (Promega Corp.).
DNA-binding assays. For electrophoretic mobility shift assays with nuclear extract, reactions contained 2 µg of poly(dI-dC), 2 µl of 5x buffer (20), 1 µl of nuclear extract (at 2 µg/µl), 1 µl of 32P-labeled probe DNA (50,000 cpm/µl), and competitor at either a 50x or a 300x molar ratio. Water was added to a final volume of 10 µl. For the binding reactions with CF2 lysate, reactions contained 1 µg of poly(dI-dC), 2 µl of 5x CF2 buffer (22), 3 µl of lysate, and probe and competitor DNAs as described above. All components except for the probe and competitor were mixed and incubated for 15 min on ice; the probe and competitor DNA were then added, and reactions were incubated at room temperature for 15 min. Samples were loaded onto a 3% nondenaturing polyacrylamide gel that was run at 4°C. The gels were then dried and exposed to autoradiography film overnight.
Large probes for DNA-binding assays (see Fig. 2 and 3C) were first generated by PCR and cloned into the pGEM-T Easy vector. DNA fragments were then excised from the plasmid using NotI and labeled by end filling the NotI overhang by the addition of dGTP, 32P-labeled dCTP, and Klenow enzyme (New England Biolabs, Beverley, MA). The PCR primers used for the generation of these fragments were as follows: –593/–500, 57B 5'+V (5'-GGGAATTCCGTAACGAACCGACC-3') and 57B EXI-IIXI (5'-GGCTCGAGCTATATCGGCCATATCGGCTATAT-3'); –521/–373, 57B conserved B++ (5'-GGATATAGCCGATATGGCCG-3') and 57B conserved B––(5'-GGGGCTGCCTGCTCACGATC-3'); and –390/–245, 57B Downstream B++ (5'-GGGATCGTGAGCAGGCAGCC-3') and 57B Ex-region BupII (5'-GGCTCGAGGGCGCACAAGCGAGAGCG-3').
|
|
|
β-Galactosidase assays. To determine the β-galactosidase activity in transfected cells, a mammalian β-galactosidase assay kit (Pierce Biotechnology, Rockford, IL) was used in accordance with the manufacturer's recommendations. Cells were lysed in 100-µl/well M-PER extraction reagent for 15 min while rotating at 350 rpm. Aliquots (20 µl) of cell lysates were mixed with equal amounts of All-in-One β-galactosidase assay reagent in a 96-well plate, incubated at 37°C for 20 to 45 min, and stopped with 40 µl of stop solution/well. Samples were read in an OpsysMR multiplate reader (Dynex Technologies, Chantilly, VA) at 405 nm. Measured absorbance values for experimental samples were normalized for the absorbance of nontransfected cells (total background) and subsequently divided by total protein concentration (as detected by Bradford assay for each sample), thereby yielding normalized β-galactosidase activity in arbitrary units. Reporter activation was determined as the fold difference of normalized β-galactosidase activity in samples cotransfected with TF DNA compared to controls cotransfected with an empty vector. Each experiment represents results from three to five independent transfections ± the standard error of the mean.
RT-PCR.
Total RNA was harvested from embryos at stage 17 or transfected cells by using an RNeasy Mini kit (Qiagen, Valencia, CA). RNA concentrations were measured and adjusted to 140 ng/µl. Coupled reverse transcription-PCR (RT-PCR) from 1 µg of total RNA was performed with a SuperScript III One-Step RT-PCR system (Invitrogen) using the sets of specific primers and conditions summarized in Table 1. The number of cycles was adjusted so that it was within a linear range of the amplification kinetics. The primers designed for Mhc and TnI yielded RNA template-specific products of distinguishably different sizes from those of genomic DNA (Table 1). However, it was not possible to design such primers for Act57B and Act5C due to high conservation between these and other actin genes and a low frequency of introns in their structure. To avoid possible amplification artifacts, the RNA preparations were additionally treated with DNase I to eliminate traces of genomic DNA and, as a quality control, samples were run for the same number of cycles with reverse transcriptase omitted from the reaction mix. These precautions ensured that resulting Act57B- and Act5C-specific products were due to amplification from appropriate RNA molecules. The specific conditions for each reaction are indicated in Table 1.
|
27 deficiency eliminating the Cf2 gene, and an equal amount of their siblings, were harvested based on the balancer chromosome green fluorescent protein (GFP) fluorescence. RNA was isolated by using an RNeasy Mini kit (Qiagen). cDNA synthesis was performed by using SuperScript II kit (Invitrogen) with random primers and 1 µg of total isolated RNA per reaction. Subsequent quantitative PCR was carried out in 96-well plates in ABI Prism 7000 machine (Applied Biosystems). Template and primer concentrations were adjusted to the optimum. The primers used were as indicated in Table 1. The resulting products were also analyzed on agarose gels to verify the absence of multiple amplified products. The differences in the expression of muscle-specific genes were calculated by using the following formula: 
(x) = E(x)[CT(x27/27) – CT(x27/+)]/E(Act5C)[CT(Act5C27/27)–CT(Act5C27/+)], where x is the gene in question, Act5C is the reference gene, E is the primer efficiency calculated in separate calibration runs (typically, 1.98 to 1.99), and CT is the threshold cycle by the ABI Prism 7000 software. For graphical representation the remaining concentration (C) levels for muscle genes were calculated as follows: C(x) = [1 [– (x)]·100. For each gene, the average results from four separate runs (each set in pentaplicated samples) were pooled together, averaged, and used for the calculations of the standard error of the mean.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
70 bp to position –521 did not affect the initiation of Act57B expression at stage 11 (Fig. 1C), although reporter gene expression levels were consistently reduced at stage 16 (Fig. 1D). This was apparent by the lesser ease with which individual muscle fibers could be distinguished in stained embryos at higher magnifications (Fig. 1D'). A further 5' deletion of 130 bp to position –390/+2 showed no additional strong reduction in reporter gene expression (Fig. 1E, F, and F'). In contrast, the smallest –278/+2 Act57B-lacZ showed reduced reporter activity at stage 12 (Fig. 1G), and reporter levels also appeared to be further reduced at stage 16 compared to larger genomic fragments that were tested (Fig. 1H and H'). These studies demonstrated that more than one region of the Act57B upstream promoter was required for normal reporter gene expression in vivo. Although it is difficult to assign relative levels of expression in these immunohistochemical stains, two regions seem most critical to normal expression: the region between positions –593 and –521 is a genomic element required for full gene activation at later stages of development, and a more proximal region from positions –390 to –278 was important for the initiation of Act57B expression at stages 11 and 12, as well as for later sustained Act57B expression. Our subsequent studies therefore focused initially upon the identification of factors interacting with this more proximal region.
Identification of CF2 as a nuclear factor interacting specifically with the Act57B promoter. In order to determine whether factors existed that were capable of interacting with the Act57B promoter region, we carried out electrophoretic mobility shift assays using as a probe a –390/–245 region of the Act57B gene (shown diagrammatically in Fig. 2A). When this fragment was mixed with nuclear extracts prepared from 12- to 24-h-old Drosophila embryos, we consistently observed the formation of a slowly migrating complex, which we hypothesized corresponded to a transcription factor capable of regulating Act57B expression (Fig. 2B).
To more precisely locate the region of the DNA probe with which the nuclear factor was interacting, we next competed binding reactions with sets of double-stranded oligonucleotides spanning the entire probe sequence. The relative locations of these competitors are shown in Fig. 2A. We found that, whereas most of the oligonucleotide competitors failed to reduce the intensity of the bound complex, oligonucleotides 11 and 12 effectively competed for the formation of the complex (Fig. 2B).
These findings were important for two reasons. First, the observation that some oligonucleotides were capable of competition, whereas others were not, indicated that the interaction of the nuclear factor with the Act57B probe was sequence specific. Second, our competition assays identified more precisely the region of the Act57B gene that was interacting with the nuclear factor.
Observation of the DNA sequence corresponding to the locations of oligonucleotides 11 and 12 revealed several repeats of a 5'-TATA-3' motif, which lies at the core of the recognition sequence of the transcription factor CF2 (17). Since CF2 is a C2H2 zinc finger protein that regulates gene expression in the chorion cells of the maturing egg in Drosophila (17, 18), and since embryonic expression of CF2 is exclusive to the three muscle lineages of the mesoderm (4), we reasoned that CF2 might be interacting specifically with the Act57B enhancer.
To determine whether CF2 could interact with the Act57B gene enhancer, we next carried out an electrophoretic mobility shift assay using as a probe the –390/–245 region and using the CF2 protein generated in vitro for binding. We found that there was a robust interaction between CF2 and the –390/–245 region and that this interaction was competed for with excess nonradioactive –390/–245 sequence (Fig. 2C). More importantly, when we competed the interaction of CF2 and –390/–245 probe using the oligonucleotides spanning the region, CF2 binding was competed for only by oligonucleotides 11 and 12.
These findings firstly demonstrated that CF2 was capable of interacting with the Act57B promoter. This is likely to be a relevant interaction given the broad expression of CF2 in the embryonic mesoderm. Second, our results showing identical responses to binding competition by a nuclear factor and by CF2 protein strongly suggested that CF2 was the interacting factor identified in nuclear extracts. We note that when CF2 protein is synthesized in vitro the electrophoretic mobility shift assays show two shifted bands of very similar mobilities. This probably arises from the synthesis of two CF2 isoforms in the in vitro expression system (data not shown). Whether this arises from posttranslational modification of the CF2 protein or from internal translation initiation is not clear. Nevertheless, there is a strong and specific interaction between CF2 protein and the Act57B promoter region.
In order to more precisely define the interaction of CF2 with the Act57B promoter, we next sought to determine by using electrophoretic mobility shift assays whether CF2 could bind to the smaller oligonucleotides 11 and 12. In fact, CF2 was incapable of interacting with these sequences (data not shown).
As an alternative strategy, we next designed an oligonucleotide probe, termed 2XCF2, which spanned both of the competitor oligonucleotides. Examination of the 2XCF2 sequence identified a number of motifs that approximately matched the CF2 consensus site. These motifs were found throughout the sequence overlapped by oligonucleotides 11 and 12 (Fig. 3A).
In DNA-binding assays with 2XCF2, we found a strong interaction between the probe and CF2 protein, and this interaction was competed for by oligonucleotides 11 and 12 (Fig. 3B, left panel). This result indicated that CF2 protein could effectively interact with the critical promoter region of Act57B. However, it was unusual that CF2 was not able to bind to oligonucleotides 11 and 12, despite their being able to compete very effectively for CF2 binding to longer probes. One possible explanation for this result is that CF2 can only transiently interact with the smaller oligonucleotides, either because they are physically shorter or because they comprise only one or a few binding sites. Such a transient interaction might be sufficient to interfere with CF2 binding to longer sequences but might not be sufficient for the sustained binding that is required in an electrophoretic mobility shift assay.
To define specific sequences required for CF2 interaction with Act57B, we next altered the 2XCF2 sequence to mutate one or both of the regions containing CF2 consensus motifs. These mutated oligonucleotides were used as competitors for interaction of CF2 protein with the wild-type target sequence (Fig. 3B, right panel). We found that mutation of the leftmost sequence did not affect the ability of the oligonucleotide to compete protein-probe interactions; however, mutation of the rightmost consensus sites reduced the degree of competition for binding, although competition was still evident. These results suggest that CF2 can interact more readily with the sequence corresponding to the larger, rightmost, cluster of consensus sites.
In order to confirm the specificity of the initial interaction of CF2 with the –390/–245 region that we had observed in Fig. 2, we next determined whether the 2XCF2 and 2XCF2-mutant sequences could compete for interaction with the entire –390/–245 sequence. This was indeed the case: a –390/–245 competitor reduced the intensity of the complex between CF2 and the labeled –390/–245 region; the intensity of the complex was also strongly reduced by 2XCF2 competitor, and the intensity of the complex was not reduced by using a 2XCF2-mutant competitor (Fig. 3C). These studies established the specificity of interaction of CF2 with the Act57B promoter region being tested. We note that the shorter oligonucleotide (2XCF2) acts more effectively as a competitor compared to the longer –390/–245 competitor, even when present in equimolar concentrations as shown here. The reason for this observation is not clear but was consistently observed for these competition assays, as well as for the additional studies described below.
Taken together, these findings defined a potentially important interaction between CF2 and the Act57B promoter in a region that we have shown to be required for full activation of actin gene expression.
CF2 also interacts specifically with additional Act57B promoter sequences. Since the deletion analyses shown in Fig. 1 also identified a more distal region between positions –593 and –521 as being important for full activation of Act57B expression, we tested for the ability of CF2 to interact with other Act57B promoter sequences that had been addressed in the deletion analyses. We observed the formation of a shifted complex comprising CF2 and the –593/–500 region; in contrast, we did not observe an interaction between CF2 and the –521/–373 Act57B region (Fig. 3D). These studies indicated that CF2 could also bind to some additional regions of Act57B, and yet not all promoter regions contained CF2 binding sites. Furthermore, by examination of this sequence several 5'-TATA sequences were observed that might function as CF2 binding sites (Fig. 4A).
In order to further evaluate the interaction of CF2 with this upstream region, we generated a double-stranded oligonucleotide spanning the four putative CF2 sites within the –593/–500 sequence (termed UPCF2) and a mutant version of this oligonucleotide in which the four sites were mutated (termed UPCF2mut). These sequences are shown in Fig. 4A.
In DNA-binding assays, the formation of a CF2-DNA complex with the –593/–500 region was effectively competed by excess –593/–500 sequence, as well as by the UPCF2 double-stranded competitor. In contrast, the UPCF2mut competitor did not effectively compete for DNA binding to the –593/–500 labeled probe (Fig. 4B). These studies confirmed that interaction of CF2 with this upstream region was sequence specific. Furthermore, such an interaction might account for the requirement of the –593/–521 sequence for full activation of the Act57B-lacZ reporter construct demonstrated in Fig. 1.
To assess the ability of the different clusters of CF2 sites to compete with each other for interaction with CF2, we also carried out cross-competition assays. With the –390/–245 region as a probe, we observed effective competition for binding when we used –390/–245, –593/–500, UPCF2, and 2XCF2 as competitors. No competition was observed with the mutant versions of the oligonucleotides (Fig. 4D). We saw a similar pattern of competition and/or noncompetition when we used the –593/–500 DNA as a probe and the same set of competitors (Fig. 4E). Since some residual bound complex was observed in the lanes with 2XCF2 as a competitor (Fig. 4C, lane 7), it can be argued that the interaction of CF2 with the upstream site is slightly stronger than its interaction with the 2XCF2 region. However, for the purposes of these studies we showed that two Act57B promoter regions can bind effectively to the CF2 transcription factor.
MEF2 and CF2 are potent transcriptional activators of Act57B in vitro and in vivo. Clearly, the C2H2 zinc finger transcription factor CF2 was able to bind potently to regions of the Act57B promoter that were critical for full levels of gene expression, suggesting that it might be a positive regulator of Act57B transcription. To test the hypothesis that CF2 is an Act57B activator, we initiated a series of cell culture experiments aimed at determining whether MEF2 or CF2 could activate Act57B promoter-lacZ constructs. First, to validate our earlier studies (24), we cotransfected Drosophila S2 cells with the full-length –593/+2 promoter-lacZ, along with a MEF2 expression plasmid. This resulted in strong and reproducible activation of the reporter (Fig. 5). In contrast, a point mutation of the putative MEF2-binding site within the –593/+2 promoter-lacZ construct dramatically impaired MEF2-dependent reporter responsiveness. These results confirmed the importance of the sole MEF2 site in this construct for Act57B gene activation.
|
Taken together, the results presented in Fig. 1 to 5 provide strong support for the identification of CF2 as an activator of Act57B: CF2 is capable of binding to two regions of the promoter, which were shown to be important for full reporter gene expression in vivo; furthermore, CF2 is identified as a strong transcriptional activator of the Act57B promoter via these same sequences in tissue culture.
We next sought to determine whether such a situation exists in vivo. We ectopically expressed either of the transcription factors in the ectoderm of transgenic embryos using the Gal4-UAS system (8), and we monitored the expression of the endogenous Act57B. As described by Lin et al. (28), this effect can be best observed in the cells of the ventral ectoderm and midline, where no underlying mesodermal expression hinders the observation of the ectopic stain. In order to maximize Gal4 activity in these embryos, crosses were performed at 29°C since Gal4 is more active at this temperature (35).
In contrast to the typical mesoderm-restricted pattern of expression for Act57B in control embryos at stage 14 (Fig. 6A), 89% of stage 14 embryos expressing high levels of ectodermal Mef2 displayed dramatically elevated expression of Act57B (Fig. 6B). Similarly, ectopic ectodermal Cf2 expression also led to Act57B activation in the ventral ectoderm in 86% of embryos (Fig. 6C), albeit at lower levels compared to the efficacy of MEF2. Since ectoderm does not normally activate muscle structural gene expression, these findings argue in favor of MEF2 and CF2 each being independent transcriptional activators of the Act57B gene in vivo.
|
Under these circumstances of reduced temperature, ectodermal expression of either Mef2 or Cf2 rarely resulted in activation of Act57B in the ventral ectoderm (22% of embryos for Mef2; 0% of embryos for Cf2; Fig. 6D and E). Remarkably, when Mef2 and Cf2 were coexpressed at 25°C, it resulted in strong activation of Act57B in 68% of embryos studied (Fig. 6F). This effect of increased Act57B expression was the first evidence of a cooperative functional interplay between MEF2 and CF2.
To further characterize the nature of the collaboration between MEF2 and CF2 and to assess this result more quantitatively, we resumed our in vitro cell culture assay using the –593/+2 Act57B reporter. Cotransfection of increased amounts of the transcription factors resulted in a dose-dependent linear activation of the reporter. When MEF2 and CF2 were cotransfected together (at 1:1 ratio), it resulted in significantly steeper activation of the Act57B reporter (Fig. 7A).
|
|
Transfection studies in cell culture indicated that CF2, as well as MEF2, could individually be effective transcriptional activators for both TnI and Mhc reporters (Fig. 7B and C and data not shown). This is the first direct evidence in support of the roles of MEF2 and CF2 upon these structural genes. More importantly, we found that CF2 and MEF2 also synergistically acted upon the promoter-lacZ constructs for both TnI and Mhc (Fig. 7B and C). These findings supported a general role for the collaboration of CF2 and MEF2 in the control of embryonic muscle development in Drosophila. It was interesting that the titration analyses of the TnI and Mhc reporters showed different degrees of synergism between MEF2 and CF2 (Fig. 7B and C and Table 2): whereas dual activation of the TnI reporter with MEF2 and CF2 just slightly exceeded the predicted additive limit, the Mhc reporter showed a remarkable boost of activation conferred by the combination of these two factors (Table 2). Possible reasons for these differences are addressed in the Discussion.
To further investigate whether CF2 and MEF2 were equally effective in activating the endogenous TnI and Mhc genes, we conducted a semiquantitative RT-PCR-based assay. To validate our detection methods, we showed that in stage 17 embryos there were detectable expression levels for Act57B, TnI, Mhc, and the control cytoplasmic actin gene Act5C (Fig. 7D, lane 1). In contrast, S2 cells transfected with an empty vector retained their nondifferentiated state characterized by the absence of muscle-specific Act57B, TnI, and Mhc expression (Fig. 7D, lane 2). Consistent with our earlier observations on flies, ectopically expressed MEF2 (Fig. 7D, lane 3) and CF2 (Fig. 7D, lane 4) could each activate expression of the endogenous Act57B in cells. Moreover, the combined expression of these two factors resulted in a greater activation of Act57B (Fig. 7D, lane 5), further supporting the synergistic collaboration between MEF2 and CF2.
When we performed the same experiments and assayed for the activation of TnI and Mhc gene expression, MEF2 and CF2 individually failed to activate expression of the endogenous gene in S2 cells at high levels (Fig. 7D, lanes 3 and 4). However, the combined expression of these factors led to significant transcriptional stimulation of the target structural genes (Fig. 7D, lane 5).
Overall, we conclude from our studies that CF2 is an important transcriptional activator of different structural muscle genes in the Drosophila embryo and that it activates muscle gene transcription via collaboration with the myogenic factor MEF2.
Requirement of CF2 for normal muscle structural gene expression in vivo.
Given our demonstrations that CF2 could activate Act57B gene expression in vivo (Fig. 6) and in tissue culture (Fig. 7D), we next tested whether the loss of CF2 function resulted in a reduction or loss of expression of muscle structural genes. A Cf2 mutant named Df(2L)27, described by Bagni et al. (4), comprises a deletion of the Cf2 transcribed region and is homozygous larval lethal (21a). In order to determine whether the loss of Cf2 function affects muscle structural gene expression, we collected homozygous mutant embryos (at stage 16-17) or first-instar larvae and assayed them for the accumulation of the muscle-specific Act57B, TnI, and Mhc transcripts relative to the expression of the housekeeping Act5C cytoplasmic actin gene by using quantitative RT-PCR. As controls, we collected heterozygous siblings at embryonic or larval stages, which were assayed for expression of the same genes.
In embryos, there was an ca. 10% reduction in the levels of expression of all three muscle structural genes assayed relative to the Act5C control expression levels (Fig. 8). While this change in expression was relatively small, it must be borne in mind that removal of the CF2 sites from the Act57B-lacZ reporter construct results in similarly moderate effects upon reporter gene expression (see Fig. 1), even though these two separate experiments cannot be directly compared. This point is discussed in more detail below. More strikingly, at the larval stage the reductions in the levels of muscle gene expression were much more severe, suggesting that a major function of CF2 might be to maintain high levels of expression in fully differentiated tissues. This reduction in the expression of the three genes tested was highly reproducible.
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In Drosophila embryos, there is a profound loss of muscle differentiation in the absence of MEF2 function (7, 27), and MEF2 has been identified as a critical direct regulator of many structural genes in this organism (3, 13, 29, 34, 42). However, relatively few additional activators of Drosophila muscle genes have been identified, suggesting either that MEF2 functions more autonomously in the activation of the myogenic program in flies compared to vertebrates or that significant MEF2 cofactors have yet to be defined.
Here, we identify the first known collaborative factor for MEF2 function in Drosophila. CF2, a transcriptional regulator first identified as an activator of chorion protein genes in the female ovary (47), is expressed in all three muscle lineages in the embryonic mesoderm (4). Our data indicate that CF2 interacts with the Act57B promoter both in the context of embryonic nuclear extracts and when expressed in vitro, and CF2 binding sites correspond precisely to the genomic regions required for full Act57B-lacZ activation. Furthermore, overexpression studies in vitro and in vivo establish a synergistic relationship between MEF2 and CF2 in the activation of Act57B. These studies define the first known embryonic function for this zinc finger transcriptional regulator. Nevertheless, we should also note that even in the absence of CF2 sites, in the –270/+2 Act57B-lacZ construct there is significant muscle-specific expression. This indicates that CF2 contributes to high levels of structural gene expression but that it is not essential for the activation of Act57B.
We note that removing the genomic regions containing CF2 binding sites had clear effects upon Act57B-lacZ expression at stage 16. In contrast, removal of CF2 function from the embryos in the mutant analysis resulted in a more modest reduction in Act57B transcription. Although these two experiments utilize different approaches and different readouts, what might be the cause of this apparent discrepancy? The most reasonable explanation is that there are probably additional regulators of Act57B that bind to the genomic regions affected in our deletion analyses; thus, the deletions remove the influence not only of CF2 but also of other muscle-specific activators. In contrast, deletion of the CF2 gene might remove only the influence of CF2 from actin gene expression. Future deletion and point mutant analyses, carried out in the manner we applied here, would test this hypothesis and could ultimately shed light upon the identity of such additional factors.
More broadly, MEF2 and CF2 can activate other muscle structural genes in Drosophila. Both of the enhancers that we tested have predicted binding sites for MEF2 and CF2 proteins (TnI [34] and Mhc [21]); however, our studies are the first direct functional tests of whether MEF2 and CF2 can transcriptionally activate these enhancers. Our findings suggest strongly that the collaboration of MEF2 and CF2 during fly embryonic muscle development is a general and important phenomenon.
We have used a simple tissue culture assay to demonstrate that MEF2 and CF2 collaborate to activate endogenous genes in the context of the intact genome, in addition to transiently transfected reporter constructs. Our assay closely supports the findings from other experiments presented here and is somewhat more quantifiable than ectopic expression assays in embryos. This is because embryos already express significant levels of the targets genes in the mesoderm, whereas in vitro the baseline from nontransfected cells is essentially zero. Similarly, the lacZ reporter vectors that we use frequently show leaky promoter activity, which might ultimately result in an underestimation of the true activation of target promoter-lacZ constructs. In this sense, the activation of endogenous genes in S2 cells is highly informative.
Presumably, a more complete set of MEF2/CF2 targets could be generated by cotransfecting S2 cells with expression plasmids for CF2 and MEF2 and then performing an array analysis of transcripts upregulated after this treatment. Such a data set would significantly overlap with the analysis of mesoderm differentiation described by Sandmann et al. (44), although it might have the advantage of indicating which genes are direct targets of combinatorial activation by MEF2/CF2.
Along these lines, our studies also indicate that there are subtle, yet potentially important, differences between the responsiveness of different target genes to activation by MEF2 and CF2 alone. In tissue culture, Act57B is readily activated by MEF2 and CF2 individually, and yet TnI and Mhc induction are detected far less readily. This might be a simple result of different ease of detection for different targets. Alternatively, there might be a more mechanistic basis for these differences. One explanation might be related to the proximity of binding sites of the respective factors to the target promoter. For Act57B, both MEF2 and CF2 sites that we have mapped are within 600 bp of the transcription start site. In contrast, for Mhc most of the candidate sites are located in the first intron, at least 800 bp away from the transcription start site (21). On the other hand, putative binding sites for MEF2 and CF2 are in relatively close proximity to the TnI transcription start site (34). Thus, it is not clear whether binding site proximity plays a role in the activation of genes in vitro, although it should be noted that for TnI and Mhc many of the sites have yet to be validated by DNA-binding assays.
The data of Bagni et al. (4) indicate that Cf2 expression is dependent upon the function of the Mef2 gene in Drosophila embryos; these authors also show that MEF2 levels are unaffected in Cf2 mutants, at least at the embryonic stage. It is therefore most likely that the reduction in muscle structural gene transcripts in Cf2 mutants arises directly from a loss of CF2 protein rather than from indirect effects upon MEF2 levels. Clearly, one of the means by which MEF2 activates myogenesis at high levels is via the activation of its own collaborative factors. There is some precedent for this mechanism, both in muscle and other systems. In Drosophila cardiac development, the homeodomain transcription factor Tinman is a direct activator of the GATA factor pannier (16), after which Tin and Pnr can collaborate to activate structural genes such as Sulfonylurea receptor (1). The activation of, and then collaboration with, a cofactor appears to be a commonly used genomic regulatory mechanism.
We and others have previously observed that critical regulatory genomic sequences are strongly conserved during evolution in the Drosophila group (see for example, reference 10), and the MEF2 site regulating Act57B transcription is conserved as distantly as Drosophila virilis (24). In contrast, the CF2-interacting sequences are less well conserved in the D. virilis ortholog of Act57B, although a number of AT-rich sequences can be discerned (data not shown). We hypothesize that the reason for the lesser conservation of CF2 versus MEF2 sites between D. melanogaster and D. virilis might arise from the presence of multiple CF2-interacting regions in the Act57B promoter. The presence of multiple sites might reduce the selective pressure to maintain the sequence of individual transcription factor binding sites.
CF2 is one of a large number of putative zinc finger transcription factors encoded by the Drosophila genome, and other members of this family are known to act as regulators of muscle development in Drosophila (see for example references 14 and 45). Mammalian genomes also encode multiple zinc finger proteins, and several of these contribute to the muscle phenotype (for a review, see reference 25). Major challenges in the future are to define how members of this extended protein family function during development and to elucidate why this family of proteins in particular have become so highly diverged. In light of the findings presented here, zinc finger factors may function broadly as cofactors for transcriptional activation.
|
ACKNOWLEDGMENTS |
|---|
This research was supported by GM61738 from the National Institutes of Health (NIH) to R.M.C. and by a predoctoral fellowship from the American Heart Association Pacific Mountain Affiliate to K.K.K.T. We acknowledge technical support from the Department of Biology's Molecular Biology Facility, supported by NIH grant number 1P20RR18754 from the Institute Development Award (IdeA) Program of the National Center for Research Resources.
|
FOOTNOTES |
|---|
Published ahead of print on 26 December 2007.
K.K.K.T. and A.L.B. contributed equally to this study.
Present address: Department of Molecular Cellular and Developmental Biology, 347 UCB University of Colorado, Boulder, CO 80309-0347.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2. Andres, V., M. Cervera, and V. Mahdavi. 1995. Determination of the consensus binding site for MEF2 expressed in muscle and brain reveals tissue-specific sequence constraints. J. Biol. Chem. 270:23246-23249.
3. Arredondo, J. J., R. M. Ferreres, M. Maroto, R. M. Cripps, R. Marco, S. I. Bernstein, and M. Cervera. 2001. Control of Drosophila paramyosin/miniparamyosin gene expression: differential regulatory mechanisms for muscle-specific transcription. J. Biol. Chem. 276:8278-8287.
4. Bagni, C., S. Bray, J. A. Gogos, F. C. Kafatos, and T. Hsu. 2002. The Drosophila zinc finger transcription factor CF2 is a myogenic marker downstream of MEF2 during muscle development. Mech. Dev. 117:265-268.[CrossRef][Medline]
5. Bi, W., C. J. Drake, and J. J. Schwarz. 1999. The transcription factor MEF2C-null mouse exhibits complex vascular malformations and reduced cardiac expression of angiopoietin 1 and VEGF. Dev. Biol. 211:255-267.[CrossRef][Medline]
6. Black, B. L., and E. N. Olson. 1998. Transcriptional control of muscle development by myocyte enhancer factor-2 (MEF2) proteins. Annu. Rev. Cell Dev. Biol. 14:167-196.[CrossRef][Medline]
7. Bour, B. A., M. A. O'Brien, W. L. Lockwood, E. S. Goldstein, R. Bodmer, P. H. Taghert, S. M. Abmayr, and H. T. Nguyen. 1995. Drosophila MEF2, a transcription factor that is essential for myogenesis. Genes Dev. 9:730-741.
8. Brand, A. H., and N. Perrimon. 1993. Targeted gene expression as a means of altering cell fates and generating dominant phenotypes. Development 118:401-415.[Abstract]
9. Carpenter, J. M. 1950. A new semisynthetic food medium for Drosophila, p. 96-97, vol. 24. Drosophila Information Service, Norman, OK.
10. Carroll, S. B., J. P. Klaiss, and L. M. Olds. 2005. Endless forms most beautiful: the new science of evo devo and the making of the animal kingdom. Norton, New York, NY.
11. Creemers, E. E., L. B. Sutherland, J. Oh, A. C. Barbosa, and E. N. Olson. 2006. Coactivation of MEF2 by the SAP domain proteins myocardin and MASTR. Mol. Cell 23:83-96.[CrossRef][Medline]
12. Cripps, R. M., T. L. Lovato, and E. N. Olson. 2004. Positive autoregulation of the myocyte enhancer factor-2 myogenic control gene during somatic muscle development in Drosophila. Dev. Biol. 267:536-547.[CrossRef][Medline]
13. Damm, C., A. Wolk, D. Buttgereit, K. Loher, E. Wagner, B. Lilly, E. N. Olson, K. Hasenpusch-Theil, and R. Renkawitz-Pohl. 1998. Independent regulatory elements in the upstream region of the Drosophila beta 3 tubulin gene (beta Tub60D) guide expression in the dorsal vessel and the somatic muscles. Dev. Biol. 199:138-149.[CrossRef][Medline]
14. Duan, H., J. B. Skeath, and H. T. Nguyen. 2001. Drosophila Lame duck, a novel member of the Gli superfamily, acts as a key regulator of myogenesis by controlling fusion-competent myoblast development. Development 128:4489-4500.[Medline]
15. Fyrberg, E. A., J. W. Mahaffey, B. J. Bond, and N. Davidson. 1983. Transcripts of the six Drosophila actin genes accumulate in a stage- and tissue-specific manner. Cell 33:115-123.[CrossRef][Medline]
16. Gajewski, K., N. Fossett, J. D. Molkentin, and R. A. Schulz. 1999. The zinc finger proteins Pannier and GATA4 function as cardiogenic factors in drosophila. Development 126:5679-5688.[Abstract]
17. Gogos, J. A., T. Hsu, J. Bolton, and F. C. Kafatos. 1992. Sequence discrimination by alternatively spliced isoforms of a DNA binding zinc finger domain. Science 257:1951-1955.
18. Gogos, J. A., J. Jin, H. Wan, M. Kokkinidis, and F. C. Kafatos. 1996. Recognition of diverse sequences by class I zinc fingers: asymmetries and indirect effects on specificity in the interaction between CF2II and A+T-rich elements. Proc. Natl. Acad. Sci. USA 93:2159-2164.
19. Gonzalez, P., M. Garcia-Castro, J. R. Reguero, A. Batalla, A. G. Ordonez, R. L. Palop, I. Lozano, M. Montes, V. Alvarez, and E. Coto. 2006. The Pro279Leu variant in the transcription factor MEF2A is associated with myocardial infarction. J. Med. Genet. 43:167-169.
20. Gossett, L. A., D. J. Kelvin, E. A. Sternberg, and E. N. Olson. 1989. A new myocyte-specific enhancer-binding factor that recognizes a conserved element associated with multiple muscle-specific genes. Mol. Cell. Biol. 9:5022-5033.
21. Hess, N. K., P. A. Singer, K. Trinh, M. Nikkhoy, and S. I. Bernstein. 2007. Transcriptional regulation of the Drosophila melanogaster muscle myosin heavy-chain gene. Gene Expr. Patterns. 7:413-422.[CrossRef][Medline]
21. Hsu, T., C. Bagni, J. D. Sutherland, and F. C. Kafatos. 1996. The transcriptional factor CF2 is a mediator of EGF-R-activated dorsoventral patterning in Drosophila oogenesis. Genes Dev. 10:1411-1421.
22. Hsu, T., J. A. Gogos, S. A. Kirsh, and F. C. Kafatos. 1992. Multiple zinc finger forms resulting from developmentally regulated alternative splicing of a transcription factor gene. Science 257:1946-1950.
23. Karamboulas, C., G. D. Dakubo, J. Liu, Y. De Repentigny, K. Yutzey, V. A. Wallace, R. Kothary, and I. S. Skerjanc. 2006. Disruption of MEF2 activity in cardiomyoblasts inhibits cardiomyogenesis. J. Cell Sci. 119:4315-4321.
