p204 Is Required for the Differentiation of P19 Murine Embryonal Carcinoma Cells to Beating Cardiac Myocytes

Among 10 adult mouse tissues tested, the p204 protein levels were highest in heart and skeletal muscle. We described previously that the MyoD-inducible p204 protein is required for the differentiation of cultured murine C2C12 skeletal muscle myoblasts to myotubes. Here we report that p204 was also required for the differentiation of cultured P19 murine embryonal carcinoma stem cells to beating cardiac myocytes. As shown by others, this process can be triggered by dimethyl sulfoxide (DMSO). We established that DMSO induced the formation of 204RNA and p204. Ectopic p204 could partially substitute for DMSO in inducing differentiation, whereas ectopic 204 antisense RNA inhibited the differentiation. Experiments with reporter constructs, including regulatory regions from the Ifi204 gene (encoding p204) in P19 cells and in cultured newborn rat cardiac myocytes, as well as chromatin coimmunoprecipitations with transcription factors, revealed that p204 expression was synergistically transactivated by the cardiac Gata4, Nkx2.5, and Tbx5 transcription factors. Furthermore, ectopic p204 triggered the expression of Gata4 and Nkx2.5 in P19 cells. p204 contains a nuclear export signal and was partially translocated to the cytoplasm during the differentiation. p204 from which the nuclear export signal was deleted was not translocated, and it did not induce differentiation. The various mechanisms by which p204 promoted the differentiation are reported in the accompanying article (Ding, B., Liu, C., Huang, Y., Yu, J., Kong, W., and Lengyel, P. (2006) J. Biol. Chem. 281, 14893-14906).

The interferons are vertebrate cytokines with antimicrobial, cell growth regulatory, and immunomodulatory activities that also affect differentiation (1)(2)(3). They function by modulating the expression of many genes, including those of the gene 200 cluster (4 -10). In the mouse this cluster consists of at least 10 genes, which encode the p200 family proteins (11). The human counterpart of the cluster is smaller; it consists apparently of only four genes (MNDA, IFI16, AIM2, and IFIX) that encode the Hin200 family proteins (12)(13)(14)(15).
Among the two best characterized members of the murine p200 family proteins is p202a (which was designated earlier as p202) (16 -26). p202a modulates transcription, cell proliferation, and apoptosis. It functions primarily by binding various sequence-specific transcription fac-tors and inhibiting their activity generally, but not exclusively, by binding them and blocking their sequence-specific binding to DNA (24). p202a overexpression was implicated in the susceptibility of mice to the autoimmune disease lupus (27).
The other much studied p200 family protein p204 is encoded by the Ifi204 gene (5). p204 is structurally related to p202a and is inducible by interferons, as is p202a. Depending on the cell type and the state of differentiation, p204 can be located in the nucleolus, the nucleoplasm, or the cytoplasm (28 -30). Its overexpression in cultured mammalian cells is growth-inhibitory (28,31). It can delay the progression of cells from the G 0 /G 1 to the S phase of the cell cycle (31). p204 can inhibit the transcription of ribosomal RNA by binding to the ribosomal DNAspecific UBF transcription factor and by inhibiting its binding to DNA (30). p204, similarly to p202a, contains the pRb-, p107-, and p130-binding motif LXCXE and can bind these pocket proteins (19,30,32).
Among 10 adult mouse tissues tested, the level of p204 is highest in the heart and second highest in skeletal muscle (29). During the differentiation of cultured C2C12 mouse skeletal muscle myoblasts to myotubes, the p204 (and p202a) levels increased severalfold as a consequence of the transactivation of the Ifi204 gene by the muscle-specific MyoD, myogenin, and E12/E47 transcription factors (29). This increase did not depend on interferon action (33). p204 is required for the differentiation of C2C12 myoblasts to myotubes. It enables the differentiation, at least in part, by overcoming the inhibition of the activities of MyoD, E12/E47, and other myogenic basic region-helix-loop-helix transcription factors by the inhibitor of differentiation (Id) 2 proteins Id1, Id2, and Id3 (33).
p204 is also involved in osteogenesis. It is expressed in osteoblasts of various tissues in embryonic and neonatal mice. It is induced in the course of the BMP-2-triggered differentiation of C2C12 cells to osteoblasts. p204 binds and acts as a cofactor of Cbfa-1 (Core binding factor ␣-1), which is an essential transcription factor in osteoblasts differentiation and bone formation (26).
The aim of this study has been to explore whether p204 is involved in and is required for cardiac myocyte differentiation and the basis of the high level of expression of p204 in mouse heart. As a model system for cardiac myocyte differentiation, we used cultured murine P19 embryonal carcinoma cells (34 -37). The P19 line of pluripotent embryonal stem cells was derived from a teratocarcinoma induced in C3H/HC mice (38). The cells can be maintained in culture in an undifferentiated state. When injected into early mouse embryos, * This work was supported by NIAID Research Grant AI-12320 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. P19 cells are capable of contributing to a variety of apparently normal tissues suggesting that the mechanisms of their differentiation are similar to those of normal embryonic cells (39). P19 cells can also be induced to differentiate in vitro into multiple cell types, including beating cardiac myocytes, skeletal myocytes, and myotubes, as well as neurons (34,36). The differentiation of P19 cells to cardiac and skeletal myocytes in vitro is initiated by plating P19 cells grown in tissue culture dishes and dissociated with trypsin into bacteriological culture dishes to which the cells do not attach and in which they form aggregates. The cultures are supplemented with DMSO. The aggregates are transferred to tissue culture dishes and further incubated in the presence of DMSO. Proliferating cells migrate out from the aggregates, including many that are differentiated to cardiac myocyte-type beating cells (and later also skeletal myocyte-type cells). In the absence of DMSO (whose mode of action remains to be clarified), there occurs little or no differentiation of P19 cells in vitro (34,36). DMSO causes an immediate, transient increase in Ca 2ϩ ions in the cells, and it was proposed that permeabilization and membrane penetration by DMSO may alter molecular interactions in the cells (40,41).
We report here that in P19 cells (i) DMSO induced p204 formation, (ii) ectopic 204AS RNA blocked the DMSO-induced differentiation (as well as p204 formation), and (iii) ectopic p204 could partially substitute for DMSO in inducing the differentiation of P19 cells to beating myocytes. p204 contains an NES, and in the course of differentiation to myocytes a portion of p204 was translocated from the nucleus to the cytoplasm of P19 cells (29). p204 lacking an NES or carrying a mutated NES was not translocated in P19 cultures treated with DMSO and could not substitute for p204 in triggering differentiation.
The 5Ј-flanking region of the gene encoding p204 (i.e. Ifi204) (11) contains several binding sites for the Gata4, Nkx2.5, and Tbx5 transcription factors. All these factors are important transcriptional regulators of cardiac gene expression and are members of protein families conserved in evolution from Drosophila to vertebrates (42)(43)(44)(45)(46)(47). Nkx2.5 is an NK-2 class homeodomain protein (48). Mouse Nkx2.5 is expressed throughout the development of the heart and is required for this process. In DNA it binds to the 5Ј-AAGTG-3Ј sequence (NKE element), among others. Gata4 is a zinc finger protein (49). It is expressed in the developing mouse heart almost simultaneously and spatially overlapping with Nkx2.5 (46, 50 -52). Gata4 binds in DNA to a sequence element containing a core 5Ј-GATA-3Ј motif. This motif occurs in the promoter region of many genes encoding cardiac proteins, including the atrial natriuretic factor gene (53). Tbx5 is a member of the family of transcription factors containing T box-type DNA binding domains (42). It is also expressed throughout heart development. Its consensus DNAbinding sequence is 5Ј-(A/G/T)GGTG(T/C)(C/T/G)(A/G) (43). This also occurs in the promoter regions of various cardiac-specific genes, including that encoding atrial natriuretic factor (43). Nkx2.5, Gata4, and Tbx5 can pairwise bind one another and can cooperatively or synergistically activate gene expression (52, 54 -56).
We report results showing that ectopic Gata4, Nkx2.5, and Tbx5 synergistically transactivated the expression of reporter constructs, including segments from the 5Ј-flanking region of the Ifi204 gene in murine 10T1/2 fibroblasts and in cardiac myocytes from rats, and also cooperatively transactivated the expression of endogenous p204 in 10T1/2 cells. Ectopic p204 (also DMSO) triggered P19 cells (to differentiate and) to increase their expression of the Gata4 and Nkx2.5 transcription factors. An examination of P19 cells differentiated to myocytes by chromatin coimmunoprecipitation assays with ␣Gata4 or ␣Nkx2.5 antibodies and PCR revealed the binding of endogenous Gata4 and Nkx2.5 to a segment of the 5Ј-flanking region of the Ifi204 gene, which contains a cluster of Gata4-and Nkx2.5-specific sequences. These findings indicate the following: (i) p204 is required for cardiac type myocyte differentiation from P19 embryonal stem cells, and (ii) the level of p204 increases during myocyte differentiation in consequence of the transactivation of the Ifi204 gene by cardiac transcription factors. The various mechanisms by which p204 enabled the differentiation of P19 cells to cardiac-type myocytes are reported in the accompanying article (76).
Cell Culture-P19 cells (a gift from M. W. McBurney) were derived from mouse teratocarcinoma induced in C3H/HC mice. The cells were cultured and induced to differentiate to cardiac-type myocytes as described previously (34) and in supplement M. Neonatal rat cardiac myocytes were isolated from hearts of 3-day-old rats as described previously (58,59). 10T1/2 cells, cloned murine embryonic fibroblasts (ATCC 226 CCL) (60), were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. For further details see supplement M.
Transfection-P19 cells were transfected as specified with expression plasmids encoding p204, p204ASRNA, p204MNES, p204⌬NES, Id3, p204 plus Id3, or p204⌬NES plus FLAGId3 using Lipofectamine 2000 in minimum essential medium-␣ (Invitrogen), supplemented with 2.5% fetal bovine serum (Invitrogen) and 7.5% donor calf serum (TerraCell). Stable cell lines were established by selection with G418. Clones were assayed to ascertain high level expression of the ectopic RNA or protein of interest by Northern or Western blotting, respectively. For further details see supplement M.
Assay of p204, Gata4, Nkx2.5, and MHC by Immunofluorescence Microscopy-P19 cells growing on 35-mm culture dishes were treated as described previously (34) and incubated as specified with mouse MF20, the Developmental Studies Hybridoma Bank, and M2 monoclonal ␣FLAG (Sigma), rabbit ␣p204 and rabbit ␣Id3 (Santa Cruz Biotechnology), and goat ␣Gata4 or ␣Nkx2.5 antibodies (Santa Cruz Biotechnology). As secondary antibodies, fluorescein isothiocyanate or Texas Red-linked goat anti-rabbit and anti-mouse IgG and donkey antigoat IgG (Santa Cruz Biotechnology) were used. Nuclei were stained with DAPI.
RT-PCR-P19 cells in 35-mm tissue culture dishes were lysed in 1 ml of RNA STAT-60 TM (Tel-Test Inc.), and total RNA was isolated following the instructions of the manufacturer and subjected to RT-PCR using the Omniscript RT kit (Qiagen). Sequences of primers are shown in supplement M.
Western Blotting-P19 cells were lysed as described previously (61), and 100 g of total protein was subjected to 4 -20% gradient PAGE and transferred to a polyvinylidene difluoride membrane and blotted with rabbit anti-p204 antibody (30). The blots were visualized using ECL (Pierce, SuperSignal Western Blotting kit). Rat anti-HA antibodies (Santa Cruz Biotechnology) and mouse anti-␤-tubulin antibody (the Developmental Studies Hybridoma Bank) were used to detect HA-tagged Tbx5 and ␤-tubulin, respectively. ␣Gata4 and ␣Nkx2.5 antibodies were used for detection of Gata4 and Nkx2.5, respectively.
Reporter Gene Assay-Synergistic transactivation of p204 reporter constructs was assayed as indicated by cotransfection with pCGNN-kx2.5 (3 g) and/or pCGNGata4 (3 g) and/or pCMVTbx5 (3 g) with pSVGal control plasmid (1 g) in 10T1/2 cells (62) by Lipofectamine 2000 (Invitrogen) (2 l/g DNA) in 35-mm dishes as described previously (62). pCGN and/or pCMV vector was used to bring the total amount of DNA transfected to 8 g. After a 36-h incubation, the cultures were harvested. Luciferase and ␤-galactosidase were assayed using the luciferase reporter gene assay (Roche Applied Science) and the ␤-galactosidase assay system (Promega), respectively. The experiments were repeated 3-4 times.
Sequence Analysis-To search for Gata4, Nkx2.5, and Tbx5 recognition sequences in the Ifi204 gene 5Ј-flanking region (11) (GenBank TM accession number AC006944), the MatInspector, version 2.1, data base was used, and the parameter selected for both core similarity and matrix was 0.8.
Chromatin Immunoprecipitation-The chromatin immunoprecipitation assay kit (Upstate) was used following the instructions of the manufacturer. Published procedures (63) were followed in the treatment of the cells with formaldehyde to cross-link the DNA protein interactions, the preparation of the nuclear extracts, the shearing of the chromatin by sonication, and the immunoprecipitation of DNA-protein complexes using immobilized antibodies (i.e. ␣Gata4 and ␣Nkx2.5). The nucleotide sequence of the 5Ј-flanking region of the Ifi204 gene was reported (Gen-Bank TM accession number AC006944). The sequences of the primers used in the PCR analysis of the DNA segments immunoprecipitated are available upon request.  The % of beating P19 cell aggregates on day 12 of differentiation was as follows: curve 1, ectopic p204 and treatment with 0.8% DMSO (39%); curve 2, treatment with 0.8% DMSO (38%); curve 3, 0.8% DMSO and transfection with the pCMV vector decreased this to 29%; curve 4, ectopic p204 and treatment with 0.1% DMSO (16%); curve 5, ectopic p204 and no DMSO (10%). No beating cultures arose in the absence of DMSO (curve 6) or in the presence of the pCMV vector (curve 7), and even in the presence of 0.8% DMSO if ectopic 204ASRNA was present (curve 8). No beating cultures arose (in the absence of DMSO) when p204 lacking a nuclear export signal (p204⌬NES) (curve 9) or p204 with a mutated nuclear export signal (p204MNES; see Fig. 3) was substituted for p204 (curve 10). No beating cultures arose in the presence of 0.1% DMSO and pCMV vector (curve 11). The standard deviations are indicated. For further details, see "Experimental Procedures." p204 in Cardiac Myocyte Differentiation clusters of beating cardiac myocytes by DMSO, we grew the cells in tissue culture dishes, shifted them (on day 0) to bacteriological Petri dishes in the presence of DMSO where they form aggregates, and shifted the aggregates to tissue culture dishes (on day 5). Within two days, many clusters, composed mainly of cells with the shape of cardiac myocytes, started beating (day 8; data not shown).

RESULTS AND DISCUSSION
As shown in the Western blots in Fig. 1A in P19 cells proliferating in tissue culture dishes (in the absence or presence of DMSO), no p204 was detected (day 0). Cells aggregated in Petri dishes without DMSO contained very low levels, if any, of p204, which increased slightly after their shift to tissue culture dishes (between day 6 and day 8). Cells aggregated in the presence of DMSO contained very low levels of p204 (on days 2-4) but much higher levels after their shift to tissue culture dishes (between day 6 and day 8). Thus, DMSO induced p204 in differentiating P19 cultures.
The immunofluorescence microscopic pictures (day 8;  Fig. 1C represent the time course of the increase in the percent of cell aggregates with beating myocytes in P19 cultures exposed to various conditions. The cultures that were transfected with expression plasmids were selected in the presence of G418 to eliminate untransfected cells. The conditions (Fig. 1C, curves [1][2][3][4][5][6][7][8][9][10][11] are specified in the figure. Curve 2, by day 12 of the differentiation process, 38% of the P19 cell aggregates exposed to 0.8% DMSO included beating myocytes. Transfection of pCMV204 did not increase significantly the yield of beating aggregates (39%) in a culture exposed to 0.8% DMSO (curve 1). Curve 6, no beating cultures arose in the absence of DMSO or (curve 8) in the presence of 0.8% DMSO if pCMV204AS (a plasmid encoding 204AS RNA) was transfected. Curve 5, transfection of pCMV204 resulted in the formation of aggregates, including beating myocytes (even) in the absence of DMSO. However, the percent of such aggregates (10%) was lower than in the corresponding culture (29%) (curve 3), which was transfected with pCMV, the control vector, and was incubated in the presence of DMSO. Curve 4, DMSO at a concentration as low as 0.1% increased the yield of aggregates with beating cells in a culture expressing ectopic p204 (16%), whereas as already noted (curve 5) ectopic p204 alone resulted only in a lower yield (10%). Curve 11, 0.1% DMSO treatment of cells expressing pCMV vector produced no beating culture.
The data in Fig. 1, A-C, are in line with the following conclusions: (i) DMSO induced p204 expression in aggregated P19 cells; (ii) p204 was required for the induction of differentiation by DMSO; and (iii) p204 could partially substitute for DMSO in inducing the differentiation.
The RT-PCR assays in Fig. 2A reveal that DMSO did not trigger the transcription of the Ifi204 gene to 204 RNA in proliferating P19 cells (day 0). DMSO did, however, trigger the transcription to 204 RNA in differentiating P19 cells, and the 204 RNA was present from day 6 at least until day 12, the last day when it was tested. Transfection of pCMV204 (but not of pCMV) into P19 cells (without or after elimination of the untransfected cells) resulted in 204 RNA expression even in proliferating cells (Fig. 2B). This expression persisted during the differentiation of the cells at least until day 12 (data not shown). The ectopic p204 level was determined by Western blotting to confirm that 204 RNA in P19 cells was translated into p204 (Fig.  2C). Transfection of a pCMV204AS plasmid resulted in the expression of a high level of 204AS RNA in the cells (Fig. 2D). This ectopic 204AS RNA (but not ectopic pCMV vector) decreased the amount of p204 induced by DMSO to a barely detectable level (Fig. 2E). These results suggest that the inhibition of the DMSO-induced differentiation of P19 cells to beating myocytes by 204AS RNA was a consequence of the inhibition by 204AS RNA of the accumulation of p204 as induced by DMSO.
The  (64) amino acid sequence in p204, and the sequence of nucleotides encoding it in Ifi204. Fig. 3A-(2) is a schematic representation of p204 lacking an NES (p204⌬NES). The generation of p204⌬NES involved the deletion of a 12-amino acid segment from p204. It is possible that this might have altered the structure of p204 resulting in the loss of some of its activities. Therefore, as a control we also generated p204 with a mutated NES (p204MNES), in which only three amino acid residues were replaced ( Fig. 3A-(3)). This shows the altered amino acid sequence of p204MNES in which three Leu residues from the NES were replaced by Ala residues. The mutated nucleotide sequence encoding the altered amino acid sequence is also shown. The Western blots in Fig. 3, A-(4), A- (5), and A- (6), established that the pCMV plasmids encoding p204, p204⌬NES, and p204MNES transfected into P19 cells resulted in the expression of similar amounts of protein.
The fluorescence microscopic analysis in Fig. 3B involved the use of two types of fusion proteins as follows: (i) green fluorescent pro-  MAY 26, 2006 • VOLUME 281 • NUMBER 21 tein-204 (GFP-p204) and (ii) GFP linked to a p204 derivative from which the NES oligopeptide segment was lacking (GFP-p204⌬NES). GFP serving as a control was shown to be distributed between the nuclei and the cytoplasm in both proliferating (Fig. 3B, PROLIF.) and differentiated (DIFF.) P19 cells (Fig. 3B, bottom panels). GFP-p204 was nuclear in proliferating cells, but a portion of GFP-p204 was  A-(3), in p204MNES, the sequence of NES is altered; three Leu residues are replaced by Ala residues. These Ala residues and the nucleotides encoding Leu residues or Ala residues replacing Leu residues are in boldface letters. A-(4), A- (5), and A- (6), the levels of expression of ectopic p204, p204⌬NES, and p204MNES were very similar in differentiating P19 cells transfected with pCMV204, pCMV204⌬NES, or pCMV204MNES (as determined by Western blotting on day 8 of the differentiation). Con, transfected by vector; UC, untransfected control. B, distribution of ectopic green fluorescent protein, GFP-p204, and GFP-p204⌬NES between the nucleus and the cytoplasm in proliferating and differentiating P19 cells. P19 cells were transfected, as indicated, with plasmids encoding green fluorescent protein or the fusion proteins GFP-p204 or GFP-p204⌬NES. Two cultures from each of the three types of transfectants were incubated in tissue culture dishes (without DMSO). After a 36-h incubation, one of each two of the proliferating cultures was examined by fluorescence microscopy (in panel PROLIF). Second cultures from each of the three types of transfectants were shifted after 36 h to Petri dishes and supplemented with 0.8% DMSO (in panel DIFF). Four days later, these were transferred in the same medium to tissue culture dishes to allow differentiation during the next 48 -72 h. This was followed by fluorescence microscopy. Bar is 30 m. C, subcellular location of endogenous p204 and of ectopic FLAG-p204, FLAG-p204⌬NES, and FLAG-p204MNES in proliferating P19 cells and in differentiating P19 cells (tested on D6-D7 of the differentiation). Fluorescence microscopy, using ␣FLAG (top and bottom panels) and, as secondary antibodies, anti-mouse IgG tagged with fluorescein (top panel) or with rhodamine (bottom panel) is shown. In the middle panel, to reveal endogenous p204, ␣p204, and as secondary antibody, anti-rabbit IgG, tagged with fluorescein were used. Nuclei were stained with DAPI. The bar corresponds to 30 m. For further details see "Experimental Procedures." p204 in Cardiac Myocyte Differentiation translocated from the nucleus to the cytoplasm in differentiated cells (Fig. 3B, top panels). GFP-p204⌬NES, however, was nuclear in proliferating P19 cells and remained nuclear also in differentiated P19 cells (Fig. 3B, middle panels).
To ascertain that the presence of GFP in the fusion proteins did not affect the translocation (in Fig. 3B), the tests were also performed using ectopic FLAG-p204, FLAG-p204⌬NES, and FLAG-p204MNES. Confirming the observations in Fig. 3B, the immunofluorescence microscopic analysis (in Fig. 3C, bottom panel) revealed that in proliferating P19 cells FLAG-p204 was nuclear (and so were FLAG-p204⌬NES and FLAG-p204MNES (not shown)). In P19 cells whose differentiation to cardiac myocytes cells was triggered by DMSO, much of the FLAG-p204 was translocated to the cytoplasm; however, FLAG-p204⌬NES and FLAG-p204MNES, which lacked an intact NES, remained in the nuclei (Fig. 3C, upper panels). The level of endogenous p204 in proliferating P19 cells is too low for immunofluorescent detection. However, we have reported earlier that in proliferating C2C12 myoblasts, in which the endogenous p204 level is higher and thus detectable, p204 is nuclear (29). In differentiating P19 cells the endogenous 204 was distributed between the nuclei and the cytoplasm (Fig. 3C, middle panel).
The findings in Fig. 1C indicate that (in addition to controlling the subcellular location of p204, as shown in Fig. 3, B and C) the NES was also required for the differentiation of P19 cells to beating cardiac-type myocytes. Thus, although the transfection of pCMV204 triggered the differentiation of P19 cells (even in the absence of DMSO) (Fig. 1C (curve 5)), the transfection of pCMV204⌬NES or pCMV204MNES did not trigger the formation of beating myocytes from P19 cells in the absence of DMSO (Fig. 1C (curves 9 and 10)). Furthermore, ectopic p204⌬NES or p204MNES did not substitute for ectopic p204 in triggering MHC protein formation in aggregated P19 cells in the absence of DMSO (data not shown).
p204, but Not p204 Lacking the NES, Could Substitute for DMSO in the Triggering of the Expression of the Gata4 and Nkx2.5 Transcription Factors in Aggregated P19 Cultures-As reported earlier (49,65), two of the transcription factors expressed early in the course of differentiation of aggregated P19 cultures to cardiac-type myocytes in the presence of DMSO (and also in the course of murine heart development) are Gata4 and Nkx2.5. Moreover, in the absence of DMSO, the P19 aggregates (even if transfected with an empty vector) contained primarily undifferentiated carcinoma cells, and neither Gata4 nor Nkx2.5 expression could be detected (65,66).
As shown in the Western blots (Fig. 4A), aggregated P19 cells, incubated with DMSO or incubated without DMSO but transfected with pCMV204 and thus expressing p204 (p204), expressed both Gata4 and Nkx2.5 (left panel). Aggregated cells incubated without DMSO and transfected with empty vector (Fig. 4A, Con), however, did not express either Gata4 or Nkx2.5 (left panel). Moreover, the expression of ectopic 204ASRNA strongly inhibited the expression of Nkx2.5 and Gata4 as induced by DMSO (Fig. 4A, middle and   4A, left panel). This was the case although the level of p204⌬NES expressed was similar to the level of p204 in cells transfected with pCMV204 (see Fig. 3, A-(4) and A-(5)). These results are in line with the observations that p204⌬NES or p204MNES could not substitute for wild type p204 in inducing MHC either (data not shown). All these findings are in accord with the results in Fig. 1C in demonstrating that p204 lacking an active NES was unable to trigger the differentiation of P19 cells to beating cardiac myocytes.
The immunofluorescence microscopic pictures in Fig. 4B are in accord with the Western blots in Fig. 4A in showing the pronounced induction of Gata4 and Nkx2.5 in differentiating P19 cells by ectopic p204, but little if any, if it was lacking the NES. In the pictures of cultures with ectopic p204 (but not in those with ectopic p204⌬NES) (see Fig.  4B), some of the p204 has spread into the cytoplasm. This (as noted earlier, see Fig. 3C) is in consequence of the differentiation of the P19 cells induced by p204 (but not by p204⌬NES).
The RT-PCR patterns in Fig. 4C revealed that the induction of MHC protein by DMSO or ectopic p204 in differentiating P19 cells (Fig.  1B) was (i) the consequence of the induction of ␣and ␤MHCRNA, and (ii) the ectopic 204ASRNA inhibited the induction of the transcription of the MHC genes by DMSO.
The Expression of the Ifi204 Gene Could Be Synergistically Transactivated in 10T1/2 Cells, Rat Cardiac Myocytes, and P19 Cells Differentiating to Myocytes by the Gata4, Nkx2.5, and Tbx5 Cardiac Transcription Factors-Searching for cardiac transcription factor recognition sequences in the 5Ј-flanking region of the murine Ifi204 gene (11), we noted the occurrence of numerous recognition sequences for Gata4, Nkx2.5, and Tbx5 (Fig. 5A).
These transcription factors that function in cardiac differentiation can bind each other and are capable of synergistically transactivating gene expression (51, 52, 54 -56, 67). We assayed the activity of the three factors to transactivate Ifi204 gene expression in transfected 10T1/2 murine embryonic fibroblasts (Figs. 5B and 6), cultured myocytes from newborn rats (Fig. 7), and proliferating and differentiating P19 cultures (Fig. 8). In the bulk of the assays we used reporter constructs, including various segments from the 5Ј-flanking region of the Ifi204 gene driving luciferase expression.
One of the short Ifi204 segments (of 160 nucleotides) used in a reporter construct (Fig. 5A) included parts of exon 1 and intron 1 with three recognition sites for Gata4 (designated as GATA), one for Nkx2.5 (designated as NKE), and two for Tbx5 (designated as TBX). It should be noted that a cluster of several short (about 6 -8 bp) transcription factor

p204 in Cardiac Myocyte Differentiation
recognition sites, such as in this region, is rare in the genome, although single recognition sites are common. Multiple sequence-specific factors in such a cluster typically function synergistically (68).
The expression of the reporter construct, including the Ifi204 segment in Fig. 5A in transfected 10T1/2 cells, is shown in Fig. 5B (wild type  panel). The low level of expression in control cells (lane 1) (taken as 1 to serve as the standard for comparison) was only slightly increased by ectopic Gata4, Nkx2.5, or Tbx5 (lanes 2-4). Any two of these transcription factors activated the expression synergistically ( lanes 5-7 ), and all three factors activated it even further (25-fold) (lane 8). Mutation in the reporter of any one of the two TBX sequences and especially of both strongly decreased the synergistic boost of the expression by TBX5 but affected the boosts by Gata4 and Nkx2.5 less, if at all (Fig. 5B, panels Mutated TBX (1), Mutated TBX (2), and Doubly mutated TBX ). These results are in accord with the finding that synergistic transactivation by Tbx5 (with Nkx2.5 or Gata4) requires the binding of Tbx5 to DNA (54 -56).
The importance of the GATA and NKE sequences in the control of the transactivation by ectopic Gata4 and Nkx2.5 in 10T1/2 cells was demonstrated (see supplemental Fig. S1). The results (i) confirmed that Gata4 and Nkx2.5 can synergize by protein-protein interaction even if only one of the factors was bound to DNA (50,51,69), and (ii) they demonstrated that the deletion or mutation of a GATA or NKE sequence may decrease the synergistic transactivation even in reporters with two or more binding sites for the above transcription factors.
The effects of ectopic Nkx2.5, Gata4, and Tbx5 on the expression of endogenous p204 in 10T1/2 cells, together with the levels of expression of the three ectopic transfection factors, are shown in the Western blots in Fig. 6. The numbers below the p204 panel in Fig. 6 reveal the cooperativity or in some cases the slight synergy between any two of the three ectopic transcription factors. (The lack of a further increase in the level of endogenous p204 upon addition of the third transcription factor may only be apparent because it may be due, at least in part, to the fact that in this case in order to keep the total DNA concentration constant less DNA was transfected for each factor (2 g) than in all other cases (3 g).) The lesser increase in the level of p204 protein (Fig. 6) than of p204 RNA in 10T1/2 cells (Fig. 5B) by ectopic Gata4, Nkx2.5, and Tbx5 is likely to be the consequence of, at least in part, a post-transcriptional control of p204 expression. Such control was reported in the case of the expression of p202b, another protein of the p200 family (17,23).
To establish whether the increase in the expression of p204 by ectopic Nkx2.5, Gata4, and Tbx5 in 10T1/2 cells may be relevant for cardiac myocyte differentiation, we also tested the effect of these transcription factors on p204 expression in cultures of primary cardiac myocytes from newborn rats (Fig. 7). Both the endogenous level of Nkx2.5 protein and the endogenous Gata4 activity were reported to be very low in cultured newborn rat myocytes (70,71). We transfected newborn rat myocytes in culture with a reporter construct, including a sequence from the Ifi204 gene 5Ј-flanking segment (shown in Fig. 5A). The Gata4, Nkx2.5, and Tbx5 transcription factor binding sites in the construct are shown in Fig. 7. Transfection of an expression plasmid encoding any one of the above listed transcription factors transactivated p204 expression weakly (less than 2-fold). Transfection of any two of them had a synergistic effect (an over 8-fold increase). Transfection of all three had an even stronger synergistic effect (an over 14-fold increase). These results establish that ectopic Gata4, Nkx2.5, and Tbx5 can synergistically transactivate FIGURE 6. Ectopic Gata4, Nkx2.5, and Tbx5 (actually HA-tagged Tbx5 was used) cooperatively promoted the expression of endogenous p204 in cultures of 10T1/2 cells. The fold increase is indicated below the Western blot of endogenous p204. The protein levels of each of the ectopic transcription factors in cultures transfected with the appropriate expression plasmids are shown, together with the level of endogenous tubulin serving as an internal protein control. Con, culture transfected with empty vector. 36 -40 h after transfection the cultures were lysed, and p204 was detected using p204 antibodies and the ECL system. Thereafter the membrane was stripped and blotted using antibodies against Nkx2.5, Gata4, and Tbx5 (in the case of Tbx5 antibodies to HA were used) and tubulin (each after stripping). The number 3.05 is in parentheses to indicate that less DNA was transfected of each transcription factor when using Gata4/Nxk2.5 and Tbx5 than in all other cases in which only 1 or 2 transcription factors were used. See also the text. The standard deviations are indicated. For further details see "Experimental Procedures." FIGURE 7. Synergistic transactivation of the expression of an Ifi204 reporter construct by ectopic Gata4, Nkx2. 5, and Tbx5 transcription factors in rat cardiac myocytes. Cardiac myocytes isolated from 3-day-old rats were seeded on collagen-coated dishes. The Ifi204 reporter construct inserted into a pGL3 vector and driving luciferase expression included a segment from the 5Ј-flanking region of Ifi204 extending, as indicated, from nucleotide Ϫ98 to ϩ47. The binding sites for Gata4, Nkx2.5, or Tbx5 are indicated in the figure. The Ifi204 reporter constructs were cotransfected, as indicated, with the pCGN vector (Con), and pCGNGata4, pCGNNkx2.5, pCMVTbx5, as well as with pSVGal. The activities of ␤-galactosidase and luciferase were assayed in the extracts after 36 h. The luciferase activities were normalized to ␤-galactosidase activities, and the normalized activities of the wild type and mutated reporters in extracts from cells transfected with the plasmids encoding transcription factors were related to the activity of the extract from cells transfected with vector only. The standard deviations are indicated. For further details see "Experimental Procedures." p204 expression not only in 10T1/2 cells but also in cardiac myocytes.
Proliferating P19 cultures are lacking endogenous Nkx2.5 as well as Gata4 and Tbx5 transcription factor activity (51,65,66,72), whereas P19 cultures in the course of differentiation to myocytes express Nkx2.5 and Gata4. Thus, we expected that Ifi204 reporter constructs with binding sites only for Gata4, Nkx2.5, and/or Tbx5 would be expressed only in differentiating but not in proliferating P19 cultures. The data in Fig. 8,  (a and b) demonstrate high level expression of 204 reporter constructs in differentiating and not, however, in proliferating P19 cells, thus proving the validity of this expectation. The data also indicate that the expression of the p204 reporter constructs decreases if the sequence of a Gata4-binding site (Fig. 8a) or an Nkx2.5-binding site (Fig. 8b) is mutated.
Coimmunoprecipitation of Chromatin Segments from the 5Ј-Flanking Region of the Ifi204 Gene Containing Gata4-and Nkx2.5-specific Sequences by Antibodies to Gata4 and Nkx2.5-As indicated earlier, undifferentiated proliferating P19 cells do not contain detectable levels of Gata4 and Nkx2.5 transcription factors (65,66,72) and express the Ifi204 gene very weakly, if at all ( Fig. 2A). P19 cells differentiating to myocytes, however, contain pronounced levels of Gata4 and Nkx2.5 (73) and strongly express the Ifi204 gene (Figs. 1A and 2B). These findings prompted us to test if antibodies to Gata4 and Nkx2.5 (but not control IgG) can coimmunoprecipitate chromatin segments (generated by sonication (Fig. 9A, panel b)) containing Gata4-and Nkx2.5-specific sequences from the 5Ј-flanking region of the Ifi204 gene in nuclear extracts from differentiating P19 cells but not from proliferating P19 cells. The data in Fig. 9A, panel a, indicate that this was the case.
As determined by PCR, of the five chromatin segments tested (i.e. four from the 5Ј-flanking region and one from the coding region, including the DNA with the last exon of p204 RNA), two segments, segment 1 (extending from nucleotide Ϫ5124 to Ϫ4705) and segment 4 (extending from nucleotide Ϫ184 to ϩ93), were coimmunoprecipitated by antibodies to Gata4 and to Nkx2.5 (but not by IgG), but only in a nuclear extract from differentiating P19 cells and not in that from proliferating P19 cells (Fig. 9A, panel a). (As noted earlier, the first nucleotide of intron 1 of the gene Ifi204 was taken as ϩ1.) Segments 2 and 3, located between segments 1 and 4, were not coimmunoprecipitated despite the occurrence in them of potential Gata4-and Nkx2.5-specific sequences. This indicates that the pres-ence of such a sequence does not ensure the binding of the appropriate transcription factor.
The coimmunoprecipitable segment 4 included a sequence (Fig. 5A) that was used (i) in a reporter construct whose expression in transfected cells was synergistically transactivated by Gata4 and Nkx2.5) (Fig. 5B), and (ii) to which Gata4 and Nkx2.5 bound in an electrophoretic mobility shift assay in vitro (not shown).
The reporter construct (in Fig. 9B) contained an ϳ5.5-kb sequence from the Ifi204 5Ј-flanking region. This included all the chromatin segments tested in the chromatin immunoprecipitation in Fig. 9A. The diagram in Fig. 9B reveals that the expression of this construct was cooperatively transactivated by Gata4 and Nkx2.5.
The results of the chromatin coimmunoprecipitation assays indicated that Gata4 and Nkx2.5 were bound in vivo to the chromatin from the 5Ј-flanking region of the Ifi204 gene in P19 cells differentiating to myocytes but not in proliferating P19 cells. These findings are in accord with other findings in this study by suggesting that the expression of p204 in P19 cells differentiating to myocytes is the consequence, at least in part, of the transactivation of the gene by Gata4 and Nkx2.5. Chromatin segment 1, which was also coimmunoprecipitated by antibodies to Gata4 and to Nkx2.5, also contained Gata4-and Nkx2.5-specific sequences. This segment was part of the 3Ј-terminal transcribed but not translated region of the Ifi203a gene, the 5Ј-terminal neighbor of the Ifi204 gene (11). According to a recent report (74), about a third of the transcription factor binding sites (as examined in human chromosomes 21 and 22) are located within or immediately 3Ј to well characterized genes and are significantly correlated with noncoding RNAs. Furthermore, overlapping pairs of protein coding and noncoding RNAs are often coregulated. Consequently, it will be interesting to see whether this segment 1 is a site in which a noncoding RNA transcript is initiated, and if so whether the transcript affects the expression of the Ifi203a and/or Ifi204 genes.
Various Inducers and Transcription Factors Can Promote the Tissuespecific Expression of p204-p204 was discovered as an interferoninducible protein (5). Studies on the differentiation of skeletal muscle revealed that in the course of this process p204 expression is activated by the skeletal muscle-specific MyoD and myogenin transcription factors (29). The results presented in this study established that in cardiac myocyte differentiation, p204 expression is synergis- The wild type or mutated binding sites for the Gata4, Nkx2.5, and Tbx5 transcription factors in the various Ifi204 reporters are indicated. In the mutated binding site in (a), the GATA sequence (in the 3Ј to 5Ј direction) TATC was mutated to CGTC, and in (b), the NKE sequence (in the 3Ј to 5Ј direction) CAC T T was mutated to CAAGG. A pSVGal internal control plasmid was cotransfected. Proliferating cells were transfected in the absence of DMSO and differentiating cells in the presence of DMSO. ␤-Galactosidase and luciferase activities were determined 36 h after transfection. Relative luciferase activity (R.L.A.) was calculated by comparing the normalized luciferase activities of extracts from cells transfected with the wild type or mutated reporter constructs to that of an extract from cells transfected with vector. The standard deviations are indicated. For further details see "Experimental Procedures." p204 in Cardiac Myocyte Differentiation tically activated by the cardiac Gata4, Nkx2.5, and Tbx5 transcription factors. In all the above cases, distinct transcription factors and cis-acting sequences in the Ifi204 gene mediated p204 expression. Furthermore, the expression of p204 in differentiating T cells was reported to be activated by Notch 1 (75); however, the mediators of the activation remain to be identified. Thus the expression of p204 in different tissues can be activated by a variety of tissue-specific mechanisms, i.e. by distinct inducers, transcription factors, and cis-acting sequences.
Indications for the Existence of a Positive Feedback Loop Including the p204, Gata4, and Nkx2.5 Proteins-The results presented reveal that in P19 cells differentiating to cardiac myocytes, (i) p204 (but not if lacking its NES) induced the expression of the Gata4 and Nkx2.5 transcription factors, and (ii) Gata4 and Nkx2.5 synergistically activated the expression of p204. These findings indicate the existence of a positive feedback loop, including the p204, Gata4, and Nkx2.5 proteins. The mechanisms of action of p204 in this feedback loop, together with the other mechanisms by which p204 and its NES FIGURE 9. A, antibodies to Nkx2. 5 and Gata4 (but not IgG controls) coimmunoprecipitate some of the chromatin segments from the 5Ј-flanking region of the Ifi204 gene containing Nkx2.5-and Gata4-specific binding sites. This is the case in nuclear extracts from P19 cells differentiating to myocytes but not from undifferentiated, proliferating P19 cells. (a), cultures of P19 cells proliferating in the absence of DMSO (PROLIF ) and cultures of P19 cells undergoing differentiation to cardiac myocytes in the presence of DMSO (DIFF ) on day 6 to day 7 of the process were incubated with formaldehyde to cross-link the transcription factors to DNA. Nuclear extracts were prepared, and the chromatin was sheared by sonication to 200 -1000-bp segments (see the gel electrophoretic pattern of the DNA segments in (b). The suspension was incubated with protein A beads without or after precoating with control IgG or antibodies to Gata4 (␣G) or Nkx2.5 (␣N). (The specificities of the antibodies (Ab) used were verified.) The DNA-protein complexes were eluted from the beads. The cross-linking was reversed, and the DNA was recovered by phenol/chloroform extraction and was assayed by PCR using pairs of primers for various segments of the 5Ј-flanking region of Ifi204 containing GATA or NKE sequences. The amounts of PCR products from the input DNA (Input) and from the DNA coimmunoprecipitated with ␣G or ␣N from the indicated segments of the Ifi204 gene 5Ј-flanking region (specified by the numbering of their 5Ј-and 3Ј-terminal nucleotides, taking as nucleotide ϩ1 the first nucleotide of intron 1 of the Ifi204 gene) were visualized by ethidium bromide-agarose gel electrophoresis. (c) is a schematic drawing of the Ifi204 gene showing the exons (numbered 1-9) as well as the segments S1 to S5 whose presence in the coimmunoprecipitates obtained using ␣G or ␣N was examined by PCR. Of the three negative control segments that were not coimmunoprecipitated by antibodies ␣G or ␣N, two, i.e. S2 and S3, were from the 5Ј-flanking region of the Ifi204 gene and are located between the S1 and S4 segments that were coimmunoprecipitated with ␣G and ␣N. The third negative control segment, S5, corresponds to the translated 3Ј-terminal exon of the Ifi204 gene. Control IgG did not coimmunoprecipitate any of the DNA segments tested. B, the reporter construct driving luciferase expression was generated by inserting into the pGL3 vector a segment from the 5Ј-flanking region of the Ifi204 gene extending from nucleotide Ϫ5482 to ϩ78. Ectopic Nxk2.5 and Gata4 cooperatively boosted the expression of the construct in 10T1/2 cells. For further details see "Experimental Procedures." enable the differentiation of P19 embryonal carcinoma stem cells to cardiac myocytes, are the topics of the accompanying article (76).