Cell-specific Activation of the Atrial Natriuretic Factor Promoter by PITX2 and MEF2A*

The PITX2 homeodomain protein is mutated in patients with Axenfeld-Rieger syndrome and is involved in the development of multiple organ systems, including the heart. We have examined the interaction of PITX2 isoforms with myocyte-enhancing factor 2A (MEF2A), which is a known regulator of cardiac development. A direct interaction between PITX2a and MEF2A was demonstrated using yeast two-hybrid and GST pull-down assays. To study the functional significance of this interaction, we used the atrial natriuretic factor (ANF) promoter. Coexpression of MEF2A and PITX2a or Pitx2c resulted in a strong synergistic activation of the ANF promoter in LS8 oral epithelial cells but not in other cell lines (NIH/3T3, Chinese hamster ovary, or C2C12). The synergism was dependent on promoter context, because it required MEF2 binding sites and was not seen with two other PITX2 target promoters. DNA binding by MEF2A was required but not sufficient for synergism. Upstream activators of p38 MAP kinases, MKK3 and MKK6, increased PITX2a and Pitx2c activity to yield up to 90-fold activation of the ANF promoter in LS8 cells. Because Axenfeld-Rieger syndrome is autosomal dominant and

The PITX2 gene was cloned based on its linkage in Axenfeld-Rieger syndrome (ARS) 1 (1). ARS is an autosomal dominant human disorder characterized by ocular anterior chamber anomalies causing glaucoma in Ͼ50% of affected individuals as well as dental hypoplasia, mild craniofacial dysmorphism, and umbilical stump abnormalities (1)(2)(3). Other features associated with this syndrome include abnormal cardiac, limb, and pituitary development. PITX2 is a member of the bicoid/paired-like homeodomain family (3,4). There are three isoforms, PITX2a, b, c, that differ only in their amino-terminal regions. The Pitx2 knock-out mouse is lethal at embryonic day 15 because of cardiac abnormalities and shows defects in cardiac positioning, atrial septation, and outflow tract development (4). Pitx2 is a proliferation target in the Wnt/Dvl pathway, which, when disrupted, leads to cardiac outflow tract abnormalities (5). Pitx2 has also been shown to be a downstream target in left-right asymmetry pathways and organogenesis (6 -9). It is expressed asymmetrically on the left side of the lateral plate mesoderm and derivative organs, including the heart and the gut.
We have used the yeast two-hybrid assay to analyze candidate PITX2 interacting partners. With respect to the role of PITX2 in heart development, we were intrigued by the recent findings that PITX2 could activate the promoter of the atrial natriuretic factor (ANF) gene (10). ANF plays a key role in electrolyte and fluid homeostasis in the cardiovascular system by causing natriuresis, diuresis, vasorelaxation, and inhibition of renin and aldosterone secretion (11). ANF knock-out mice are hypertensive, and ANF overexpression leads to hypotension (12). However, although ANF was first identified as a hormone produced in the cardiac atrium, it is also expressed in extra-atrial tissue, suggesting that ANF might have autocrine and paracrine functions outside the cardiovascular system (13). Most notably, ANF has been found in epithelial cells of the kidney, adrenal, pancreas, salivary, and lacrimal glands (13)(14)(15).
Myocyte enhancing factors (MEF2A-D) are known regulators of neuronal and muscle development (16). MEF2 proteins contain a MADS box DNA binding domain as well as a MEF2 domain for dimerization and protein-protein interaction. MEF2A and MEF2C are highly expressed in skeletal and cardiac muscle. Mef 2C knock-out mice show a complete absence of ANF expression in the heart (17), and Mef 2A knock-out mice die between postnatal days 5-10 of sudden cardiac death and mitochondrial insufficiency (18). Pitx2 is expressed in myotomes, myoblasts, and muscles and may be involved in muscle patterning (19,20). Based on these phenotypic observations and the recent report by Amendt and colleagues (10) that PITX2 can activate the ANF promoter, we reasoned that PITX2 and MEF2 might co-regulate cardiac genes.
Importantly, both PITX2 and MEF2 actions at the ANF promoter appear to be inherently weak. MEF2 activation of the ANF promoter is relatively mild and does not involve direct binding to DNA but rather appears to be dependent on synergism with other factors (21,22). Likewise, activation of the ANF promoter by PITX2c was enhanced by cotransfection with the Nkx2.5 homeodomain protein (10). These observations led us to reason that the ANF promoter might be a receptive target for detecting synergistic interactions between PITX2 and MEF2A proteins.
Our results show that PITX2a and MEF2A directly interact and have the ability to synergistically activate the ANF promoter. Interestingly, there was synergistic activation in the LS8 oral epithelial cells, but not in other cell lines. Furthermore, synergism was only observed with the ANF promoter and not other PITX2 responsive promoters, apparently because of a requirement for MEF2 DNA binding. The dominant negative PITX2-K88E mutant was able to decrease ANF promoter activity in C2C12 cells that express the endogenous Pitx2 gene as well as suppress PITX2a synergism with MEF2A in LS8 cells. These results support the prediction that PITX2a and MEF2A can coordinately regulate the ANF promoter in the oral epithelium.

EXPERIMENTAL PROCEDURES
Yeast Two-hybrid Analysis-The human PITX2a, C-HD, N-HD, HD, and C173 cDNAs were cloned into the Gal4 DNA binding domain bait vector (pGBKT7; Clontech) as described (23). The human MEF2A cDNA (GeneStorm plasmid U49020; Invitrogen) was cloned into the Gal4 AD vector (pGADT7; Clontech) EcoRI and ClaI restriction sites. The baits and preys were transformed into the AH109 and Y187 yeast strains, respectively. Mating was carried out according to the protocol provided with the Clontech MatchMaker3 system between MAT␣ Y187 cells carrying the prey vector and MATa AH109 cells expressing the bait vector. Yeast containing the bait and prey constructs were grown on selection plates. For the matings, one colony of each was picked and placed in 0.5 ml of 2ϫ YPDA (1% yeast extract, 2% peptone, 2% dextrose, and 0.003% adenine) medium (Clontech) without auxotrophic selection. The cultures were grown overnight at 30°C. 10 l of mated cultures were spread on synthetic dropout ϪLeu/ϪTrp plates. The plates were then incubated at 30°C for 3-5 days. The resulting mated yeast were then spotted onto synthetic dropout ϪLeu/ϪTrp/ϪHis/ϪAde auxotrophic selection plates.
␤-Galactosidase Assay-Individual colonies of the mated yeast were grown in 10 ml of ϪLeu/ϪTrp media to an A 600 of 0.4 -0.6. The yeast were collected by centrifugation and resuspended in 1 ml of buffer Z (Clontech). 100 l of the resuspended yeast were lysed by freezing in liquid nitrogen (30 -60 s) followed by thawing at 37°C (30 -60 s) four times. The ␤-galactosidase activity was assayed by measuring chemiluminescence from the lysate (25 l) in a 96-well plate luminometer (Dynex) using the Galacto-Light Plus kit (Tropix).
GST Pull-down-The TNT T7 coupled reticulocyte lysate system (Promega) was used to make [ 35 S]methionine-labeled PITX2a, MEF2A, and MEF2-DIVE (amino acids 1-86) proteins. The labeled proteins were made in a total volume of 50 l with 3 l of [ 35 S]methionine at 30°C for 90 min as recommended by the manufacturer. 5 l of labeled proteins were incubated with 50 l of GST, GST-PITX2a fragments, or GST-PITX2a full-length protein-coated beads with 300 ng of ethidium bromide per microliter for 1.5 h. The incubations and subsequent steps were carried out as described previously (23). The bound proteins were resolved by electrophoresis on a 10 or 15% SDS-polyacrylamide gel. The gel was dried and exposed to autoradiographic film.
Immunohistochemistry-Cells were grown on laminin-coated glass coverslips and fixed in 4% paraformaldehyde in phosphate buffered saline (PBS) for 20 min and then permeabilized with 0.3% Triton X-100 for 15 min. The permeabilized cells were washed twice in PBS and then blocked in 10% fetal bovine serum for 1 h. Cells were stained with anti-MEF2 (H-300) (Santa Cruz Biotechnology Inc.) (diluted 1:50 in PBS and 1% bovine serum albumin) overnight at 4°C. The cells were washed twice with PBS and then blocked again in 10% fetal bovine serum for 30 min followed by incubation with rhodamine-conjugated anti-rabbit IgG (Chemicon Inc.) (diluted 1:200) for 1 h. Unless indicated, all steps were at room temperature. The cells were then washed twice with PBS and mounted with Aqua-mount (Lerner Laboratories Inc.).
Cell Culture and Reporter Gene Assays-The human PITX2a, PITX2a-K88E, and mouse Pitx2c cDNAs were cloned into pcDNA3.1 MycHisC expression vector (Invitrogen) that contains the cytomegalovirus (CMV) promoter and an in-frame C-terminal c-Myc epitope (24). The CMV-MEF2A expression plasmid was obtained commercially (GeneStorm plasmid U49020; Invitrogen). The ANF-luciferase reporter plasmid containing 638 bp of rat ANF promoter sequence and the Gad1-luciferase reporter containing 506 bp of the mouse Gad1 promoter were kindly provided by Y. Lee (University of Wisconsin-Madi-son) (25) and B.G. Condie (University of Georgia) (26), respectively. The NP338-ANF (containing only 50 bp of the ANF promoter) and the NP613-ANF (containing three canonical MEF2 binding sites from the muscle creatine kinase promoter upstream of the 50-bp ANF promoter) luciferase reporter plasmids have been described (22). The 3ϫ MEF2-Gad1 plasmids were constructed by subcloning the muscle creatine kinase-MEF2 binding sites from the NP613-ANF plasmid into the KpnI site upstream of the Gad1 promoter fragment. A 345-bp fragment with KpnI ends was PCR-amplified from NP613. This fragment contains the three MEF2 sites and a flanking 251-bp region containing an SV40 polyadenylation signal. The Gad1 ϩ 3ϫ MEF2 (ϩ) plasmid has the MEF2 sites in the same orientation as in the 3ϫ MEF2 ANF plasmid (NP613). The Gad1 ϩ 3ϫ MEF2 (Ϫ) plasmid has the MEF2 sites in the opposite orientation, such that the polyadenylation region is between the MEF2 sites and the 5Ј-end of the Gad1 promoter. An additional plasmid (pGad-MEF2 mp3) containing two copies of the 3ϫ MEF2 fragment in the plus orientation was also generated. Plasmids encoding MKK3 and MKK6 were kindly provided by R. Davis (University of Massachusetts). The TK-Bic-luciferase reporter has been described (27).
The mouse oral epithelial cell line LS8 was kindly provided by Dr. Malcolm Snead (University of Southern California) and grown in Dulbecco's modified Eagle's medium, 10% fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin. The NIH/3T3 cells were grown in Dulbecco's modified Eagle's medium, 10% bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin. C2C12 cells were grown in Dulbecco's modified Eagle's medium, 10% fetal bovine serum, 300 g/ml glutamine, 100 units/ml penicillin, and 100 g/ml streptomycin. Chinese hamster ovary (CHO) cells were grown in ␣-minimum Eagle's medium, 10% fetal bovine serum, 300 g/ml glutamine, 100 units/ml penicillin, and 100 g/ml streptomycin. For LS8 cells, ϳ1 ϫ 10 5 cells were cultured in 12-well plates, transfected with a total of 2 g of DNA including 0.25 g of ANF-luciferase, 0.5 g of Gad1-luciferase, or TK-Bic-luciferase and 0.5 g of each expression vector (as indicated) using 4 l of LipofectAMINE 2000 (Invitrogen) and then harvested after 48 h. CHO and NIH/3T3 cells were transfected in the same way but were harvested at 16 -24 h. For C2C12 cells, ϳ6 ϫ 10 4 cells were cultured in 12-well plates, transfected with a total of 2 g of DNA including 0.5 g of ANF-luciferase and 0.5 g of each expression vector (as indicated) using 3 l of FuGENE-6 (Roche) and then harvested after 20 -24 h. An empty CMV-promoter vector pCMV5 was used to bring the total amount of DNA to 2 g. Unless otherwise noted, a CMV ␤-galactosidase plasmid (0.02 g of pCMV-␤gal) (Clontech) was used as a control for transfection efficiency. For all studies, the cells were incubated in antibiotic-free media with serum for 16 -48 h following transfection. SB203580 and SB202190 (Sigma) were prepared in dimethyl sulfoxide and added to the cells at a final concentration of 25 M for 6 h prior to harvest. The cells were washed, scraped, and assayed using luciferase reagents (Promega) and ␤-galactosidase reagents (Tropix Inc.). Luciferase activities were corrected for transfection efficiency. Expression of the Myc-tagged PITX2 proteins was confirmed by Western blots as described previously (24).

RESULTS
Direct Binding of PITX2a and MEF2A Proteins-We used the yeast two-hybrid assay to test whether PITX2a could physically interact with MEF2A. An interaction between a PITX2a-Gal4 DNA binding domain protein and a MEF2A-Gal4 activation domain protein was detected using the lacZ reporter gene and two auxotrophic selectable markers. There was 5-fold greater ␤-galactosidase activity with the PITX2a and MEF2A fusion proteins compared with the vector control (Fig. 1A). Likewise, growth was observed in the absence of histidine with the fusion proteins, but not with vector controls (Fig. 1B). The inhibitor 3-aminotriazole was included to increase stringency and eliminate background growth from the PITX2a bait (23). Similar results were seen using adenine as the selectable marker (not shown). The magnitude of the ␤-galactosidase activation and resistance to 3-aminotriazole are comparable with our previous measurements of PITX2a homodimerization (23). In the yeast two-hybrid assay, MEF2A was able to interact with full-length PITX2a and, to a lesser degree, with a fragment containing the homeodomain and C-terminal region but not with other PITX2a fragments (Fig. 1).
The interaction was confirmed by affinity chromatography using GST-PITX2a fusion proteins and in vitro synthesized MEF2A proteins ( Fig. 2A). For comparison, PITX2a homodimerization was also measured. The MEF2A protein bound to GST-PITX2a at levels comparable with those of PITX2a ( Characterization of Myoblast and Oral Epithelial Cell Lines-As model systems for examining the functional significance of the interaction between PITX2a and MEF2A, we selected cell lines representing muscle and epithelial cell lineages that express the endogenous Pitx2 gene. The C2C12 myoblastlike cells are known to express both the endogenous Pitx2 and MEF2 genes (5,28,29). The LS8 oral epithelial cells have been reported to express the endogenous Pitx2 gene (30). For comparison with heterologous cells that do not express Pitx2, we used the CHO cell line (30). MEF2 immunohistochemical staining was detected in the nuclei of C2C12 and LS8 cells but not in those of CHO cells (Fig. 3). Thus, these cells appear to express endogenous MEF2 proteins. Because the antiserum recognizes multiple members of the MEF2 family, we cannot identify the MEF2 family member (e.g. MEF2A or MEF2C). As a control, staining was detected in a subset of CHO cells after transfection with the MEF2A expression vector (Fig. 3).
Activation of the ANF Promoter by PITX2a and MEF2A in C2C12 Myoblasts, CHO Cells, and NIH/3T3 Cells-As PITX2 and MEF2 are both involved in heart development, we first wanted to test their activity on the ANF promoter in the myoblast-like C2C12 cell line. Transfection of a PITX2a expression vector yielded only weak activation of the ANF promoter in these cells (Fig. 4A). As reported previously in HeLa cells, MEF2A had only a minimal effect on the ANF promoter (Ͻ2fold) (22). Cotransfection of MEF2A and PITX2a yielded essentially the same weak promoter activity as seen with PITX2a alone (Fig. 4A). Similar results were seen with the Pitx2c expression vector with and without MEF2A (not shown). As a control, we confirmed that the comparable levels of Myc-tagged exogenous PITX2a were expressed by Western blots (Fig. 4B).
Because the C2C12 cells express the endogenous Pitx2 and Mef 2 genes, one possibility was that these proteins might not be rate-limiting, which might account for the rather weak activation by exogenous PITX2a and MEF2A and would, in effect, mask our ability to detect synergism between these factors. To determine whether endogenous Pitx2 contributed to ANF promoter activity, we expressed a dominant negative PITX2a protein. The PITX2a-K88E mutant has been shown to repress wild type PITX2a activity (23,24). We observed a decrease in ANF promoter activity upon the transfection of PITX2a-K88E, which is consistent with the expectation that the dominant negative protein would repress the endogenous Pitx2 (Fig. 4A). Upon cotransfection of MEF2A there was no increase in activation. Cotransfection of wild type PITX2a with PITX2a-K88E and MEF2A rescued ANF promoter activity back to baseline (Fig. 4A). These results suggest that endogenous Pitx2 is acting on the ANF promoter in C2C12 myoblasts.
In light of the potential complications from endogenous Pitx2 and MEF2 proteins in C2C12 cells, we turned to a heterologous system. The CHO cell line does not appear to express either Pitx2 (30) or MEF2 proteins (Fig. 3). Transfection of PITX2a yielded a somewhat greater activation of the ANF promoter than that seen in C2C12 cells (Fig. 4C). However, similar to our observations in the C2C12 cells, there was only an additive effect of PITX2a with MEF2A on ANF promoter activity (Fig.  4C). Likewise we observed only weak activation of the ANF promoter by the combination of PITX2a and MEF2A in the NIH/3T3 fibroblast cell line (Fig. 4D). Although greater than an additive increase in activity and statistically significant (p Ͻ 0.05, Student's t test), the activation by the combination of PITX2a and MEF2A was only 1.6-fold greater than that by PITX2a alone in the NIH/3T3 cells.
PITX2 and MEF2 Synergistically Activate the ANF Promoter in LS8 Cells-Based on the literature, the ANF promoter is active in nonmuscle cells, including epithelial cells of the oral region (13, 14, 53). Because Pitx2 is expressed in the oral epithelium and MEF2 proteins also play roles in non-muscle lineages, we decided to test PITX2a and MEF2A activation of the ANF promoter in the LS8 oral epithelial cell line. The LS8 cells express Pitx2, and we have shown previously that exogenous PITX2a can activate target promoters in these cells (24,30).
The transfection of PITX2a activated the ANF promoter 5-fold in LS8 cells (Fig. 5A). MEF2A expression on its own had little or no effect on promoter activity (ϳ1.5-fold activation). However, cotransfection of PITX2a with MEF2A resulted in a significantly higher 15-20-fold activation of the ANF promoter (Fig. 5A). As a control, we confirmed that comparable levels of the Myc-tagged exogenous PITX2a were expressed by Western blots (Fig. 5B). These results demonstrate that PITX2a and MEF2A can synergistically activate the ANF promoter in LS8 cells.
The mutant PITX2a-K88E protein could not activate the ANF promoter or synergize with MEF2A in the LS8 cells (Fig.  5A). The expression of PITX2a-K88E did not lower ANF promoter activity below baseline. The lack of basal repression by PITX2a-K88E in the LS8 cells suggests that, in contrast to the C2C12 cells, the endogenous Pitx2 protein levels may be too low to activate the ANF promoter. Importantly, the synergism between wild type PITX2a and MEF2A was abolished by coexpression of the dominant negative PITX2a-K88E mutant (Fig. 5A).
To better understand the requirements for PITX2a-MEF2A synergy, we tested a minimal ANF promoter fragment with and without MEF2 binding sites. For comparison, the 638-bp ANF promoter, which includes a non-consensus MEF2 site (22), was included. The minimal ANF promoter, which includes only 50-bp of upstream sequence, was weakly activated by PITX2a and MEF2A, alone or in combination (Fig. 6A). There was no significant difference in activation between PITX2a and the combination of PITX2a and MEF2A. However, the addition of three canonical MEF2 binding sites upstream of the minimal ANF promoter did result in a striking 35-fold synergistic activation by PITX2a and MEF2A. As expected, this promoter was activated to a greater extent than the minimal or full-length ANF promoters by MEF2A. PITX2a activation was comparable for all three promoters. Interestingly, the addition of MEF2 binding sites resulted in a greater synergistic activation of promoter activity than that seen with the native ANF promoter (Fig. 6A). These results suggest that binding of MEF2A to sites in the ANF promoter is necessary for synergism with PITX2a.
To identify the domains of MEF2A required for synergism, we tested the MEF2A fragment containing only the MADS box and MEF domain, which was shown to be sufficient for binding PITX2a (Fig. 2). These domains are also sufficient for binding to DNA. Co-transfection of this MEF2A fragment (MEF2-DIVE) decreased basal activity of the ANF promoter and did not synergize with cotransfected PITX2a (Fig. 6B). This finding indicates that binding of MEF2A to DNA is required but probably not sufficient for synergism with PITX2a.
PITX2a-MEF2A Synergism Is Promoter-dependent-To determine the promoter specificity of PITX2a-MEF2A synergism, we tested activation of two other promoters. Both the synthetic TK-Bic reporter containing four bicoid DNA binding sites (TA-ATCC) and the Gad1 reporter containing 506 bp of the promoter sequence are stimulated by PITX2 (26,27). MEF2A only weakly activated the TK-Bic promoter and did not synergistically increase the PITX2a activation (Fig. 7A). Likewise, MEF2A did not synergize with PITX2a to activate the Gad1 promoter (Fig. 7B). We also tested the effect of inserting multimerized MEF2 binding sites upstream of the Gad1 promoter (Fig. 7B). Although these sites were able to increase MEF2A responsiveness on the minimal ANF promoter (Fig. 6A), we did not observe any increase in MEF2A activation or detectable synergism with PITX2a on the Gad1 promoter. These data indicate that PITX2a and MEF2A synergism is selective for the ANF promoter.
PITX2 Activation Is Enhanced by p38 MAP Kinase-The p38 MAP kinases are known to activate MEF2 proteins (31,32). We wanted to know if activation of p38 MAP kinases would have an effect on PITX2 activation of the ANF promoter in the presence and absence of MEF2. To do this we used the MKK3 and MKK6 MAP kinase kinases that activate p38 MAP kinases. Upon cotransfection of MKK3 with PITX2 there was increased ANF activation with both the PITX2a and Pitx2c isoforms (Fig. 8A). We confirmed that the MKK3-enhanced activity of PITX2 is due to p38 MAP kinases by using the p38 MAP kinase inhibitors SB203580 (Fig. 8A) and SB202190 (data not shown). SB203580 treatment attenuated activation by MKK3 without significantly affecting basal ANF promoter activity. This result confirms that PITX2 is a downstream target of p38 MAP kinases.
We then examined PITX2 activation of the ANF promoter and synergism with MEF2A in LS8 cells in the absence and presence of MKK6 (Fig. 8B). As seen with MKK3, both PITX2a and Pitx2c activation of the ANF promoter were enhanced by cotransfection of MKK6. A corresponding increase in activation by the PITX2 and MEF2A proteins was also seen in the presence of MKK6 (Fig. 8B). The Pitx2c isoform showed greater activation of the ANF promoter than PITX2a both in the absence and presence of MEF2A. These data suggest a possible role of p38 MAP kinases in up-regulating PITX2 activation of the ANF promoter. DISCUSSION In this study we report that PITX2a can interact and synergize with the MEF2A protein. Thus, MEF2A can be added to the growing list of proteins shown to interact with the PITX1 and PITX2 proteins (33)(34)(35)(36). Based on the in vitro binding data, the interaction involves the PITX2 homeodomain and the MEF2A MADS and MEF2 domains, although in the yeast two-hybrid assay an interaction with just the PITX2 homeodomain was not observed. This difference may reflect the higher in vitro protein concentrations or folding requirements, as suggested previously by our PITX2a homodimerization study (23). To test the functional significance of the PITX2a-MEF2A interaction, we used the ANF promoter. Cotransfection of PITX2a and MEF2A resulted in a synergistic activation that appears to be specific for the ANF promoter. PITX2a was able to activate the Gad1 and TK-Bic promoters, but there was no synergism seen upon addition of MEF2A. The synergism was seen with both the human PITX2a and mouse Pitx2c isoforms. This suggests the potential for a complex combinatorial code of activation because the PITX2 isoforms can have either similar or different activities, possibly due to promoter and cellular context (10,(37)(38)(39).
The basis of the ANF promoter specificity appears to reside within the ability of MEF2A to bind the promoter. We initially did not suspect the involvement of MEF2A sites because the ANF, Gad1, and TK-Bic promoters do not have consensus MEF2 binding sites and there was little or no activation upon the expression of just MEF2A. In addition, MEF2A binding to DNA was not required for synergism with GATA factors on the ANF promoter (22). However, the ANF promoter does have a non-consensus, low-affinity MEF2 site (22). To investigate a possible requirement for MEF2 binding, we tested a minimal ANF promoter that does not include the low affinity site. It should be noted that although the minimal promoter does not contain the published PITX2 binding site at Ϫ305 bp (10), we still observed activation by PITX2a, presumably from cryptic site(s). Upon cotransfection of both PITX2a and MEF2A, no synergy was observed on the minimal ANF promoter. Interestingly, when we tested the minimal ANF promoter with three canonical MEF2 binding sites there was a very robust 35-fold synergistic activation by PITX2a and MEF2A. This synergism was even greater than that seen with the natural ANF promoter. Therefore, in contrast to MEF2A synergy with GATA factors, MEF2A binding appears to be required for synergistic activation of the ANF promoter by PITX2a. DNA binding alone is not sufficient for synergism, because a fragment of MEF2A that could still bind DNA and PITX2a, but lacks transactivation activity, did not synergize with PITX2a. Furthermore, insertion of the canonical MEF2 sites upstream of the Gad1 promoter was also not sufficient for synergistic activation by PITX2a and MEF2A. These data point to MEF2A DNA binding as being a crucial requirement but not sufficient for the promoter-specific synergism.
Surprisingly, a strong PITX2-MEF2 synergism was seen in the LS8 oral epithelial cell line, but not in other lines tested (NIH3T3 fibroblast cells, myoblast-like C2C12 cells, and CHO epithelial cells). In C2C12 cells, the observation that the dom- inant negative PITX2a-K88E protein lowered ANF promoter activity at first suggested that Pitx2 and MEF2 proteins might not be rate-limiting because of the endogenous proteins. However, this alone cannot account for the lack of synergism, because we also did not observe synergism in cells that do not express detectable endogenous Pitx2 or MEF2 proteins (CHO cells). This observation suggests that synergism may require epithelial cell-specific factors. This possibility is supported by the findings of Morin et al., who raised the possibility that specific GATA factors may be required for MEF2 activity on the ANF promoter in nonmuscle cell types (22).
ARS patients have one wild type and one mutated PITX2 allele. We tested the effect of the PITX2a-K88E mutant on wild type PITX2a synergism at the ANF promoter. Previous studies have shown that the PITX2a-K88E mutant does not bind or transactivate DNA but does repress wild type PITX2a activity and synergism with the Pit-1 transcription factor (23). This dominant negative phenotype is caused in part by increased dimerization between the PITX2a-K88E and wild type PITX2a proteins (24). We found that PITX2a-K88E repressed synergism between wild type PITX2a and MEF2A on the ANF promoter. The dominant negative PITX2a-K88E protein was also a useful tool for testing the contribution of endogenous Pitx2 proteins on the ANF promoter as described above. The ability of the PITX2a-K88E mutant to suppress wild type PITX2a synergism with MEF2A provides another piece of insight toward understanding the molecular basis of the syndrome.
To ascertain whether PITX2 plays a role in p38 MAP kinasedependent activation of the ANF promoter, we overexpressed upstream kinases in the p38 pathway. The p38 MAP kinases are known to regulate MEF2 proteins in myocytes and are required for cardiac hypertrophy and ANF expression in hypertrophic hearts (16,40). Overexpression of MKK3 in the C2C12 cells increased ANF promoter activity even in the absence of exogenous PITX2 or MEF2A proteins (data not shown), possibly because of the endogenous proteins. In contrast, the p38 MAP kinases did stimulate PITX2a and Pitx2c activity in the LS8 oral epithelial line. Unexpectedly, these kinases had little effect on MEF2A activity in the LS8 cells but did increase the activity if PITX2 was also cotransfected. The combination of Pitx2c, MEF2A, and MKK6 yielded an almost 100-fold activation of the ANF promoter. Thus, PITX2 is a downstream target of p38 MAP kinases, which suggests a role for PITX2 in p38 MAP kinase activation of ANF expression.
What might be the significance of PITX2-MEF2 synergism? Genetic and epigenetic studies with Pitx2 (5-9, 41) and MEF2 (17,18) support the prediction that these factors may act together in cardiac tissue to regulate ANF and other genes. What about in the oral epithelium? In the oral epithelium it is clear that PITX2 is expressed early and plays an important role in Procedures"). The activity is shown as mean fold activation compared with the reporter alone Ϯ S.E. from at least three independent experiments for each panel.
FIG. 8. PITX2 activation of the ANF promoter with MEF2A is enhanced by p38 MAP kinases. A, LS8 cells were transfected with the ANF promoter reporter DNA alone or with PITX2a or Pitx2c in the presence or absence of MKK3 (0.5 g) as indicated. Six hours prior to harvest the cells were treated with either vehicle (Me 2 SO) or the p38 MAP kinase inhibitor SB203580 (SB). The activity is shown as mean fold activation compared with reporter alone Ϯ S.E. from three independent experiments. B, LS8 cells were transfected with the ANF promoter reporter DNA alone or with PITX2a, Pitx2c, or MEF2A in the presence and absence of MKK6 (0.5 g each) as indicated. The activity is shown as mean fold activation compared with reporter alone Ϯ S.E. from three independent experiments. tooth development (42)(43)(44). However, the expression and function of MEF2 expression in the epithelium is not clear. MEF2 RNAs are ubiquitously expressed, although the protein levels are much higher in muscle and neuronal lineages because of translational control (45)(46)(47). Nonetheless, functional MEF2 protein is present in the nuclei of nonmuscle cell lines, which indicates that the ubiquitous MEF2 RNAs can be translated in various cells (48). This agrees with our observation of MEF2 immunoreactive material in the LS8 oral epithelial cells. Whether MEF2 proteins are indeed coexpressed with PITX2 in the oral epithelium remains to be determined.
What then might be the consequence of PITX2 and MEF2A regulation of ANF in the oral epithelium? It is known that ANF and ANF receptors are expressed in oral epithelial cells (13,14,49). Indeed, it has actually been suggested for over a decade that ANF might have noncardiovascular autocrine and paracrine functions (13). Epithelial ANF may have a physiological role in the regulation of fluid secretion of the lacrimal and submaxillary glands and ducts analogous to its cardiovascular fluid homeostasis role (15,50,51). Thus, regulation of ANF expression by PITX2 and MEF2A may contribute to oral epithelial physiology. In addition, ANF has been suggested to play a role in oral epithelial cell proliferation (49), which is consistent with the role of PITX2 in proliferation (5). Future studies using tissue-specific and compound heterozygous PITX2 and MEF2A knock-out mice should help resolve these issues.