BMP signaling regulates Nkx2-5 activity during cardiomyogenesis.

Nkx2-5 regulates the transcription of muscle-specific genes during cardiomyogenesis. Nkx2-5 expression can induce cardiomyogenesis in aggregated P19 cells but not in monolayer cultures. In order to investigate the mechanism by which cellular aggregation regulates Nkx2-5 function, we examined the role of bone morphogenetic protein 4 (BMP4). We showed that the expression of the BMP inhibitor, noggin, was sufficient to inhibit the induction of cardiomyogenesis by Nkx2-5 during cellular aggregation. Furthermore, soluble BMP4 could activate Nkx2-5 function in monolayer cultures, resulting in the formation of cardiomyocytes. Therefore, BMP signaling is necessary and sufficient for the regulation of Nkx2-5 activity during cardiomyogenesis in P19 cells.

Nkx2-5 activity is essential for cardiomyogenesis.

The homeobox transcription factor tinman is essential for heart vessel formation in Drosophila. In contrast, mice lacking the murine homologue Nkx2-5 are defective in cardiac looping but not in cardiac myocyte development. The lack of an essential role for Nkx2-5 in cardiomyogenesis in mammalian systems is most likely the result of genetic redundancy with family members. In this study, we used a dominant negative mutant of Nkx2-5, created by fusing the repressor domain of engrailed 2 to the Nkx2-5 homeodomain, termed Nkx/EnR. Expression of Nkx/EnR inhibited Me(2)SO-induced cardiomyogenesis in P19 cells but not skeletal myogenesis. Nkx/EnR inhibited expression of cardiomyoblast markers, such as GATA-4 and MEF2C, but not of mesoderm markers, such as Brachyury T and Wnt5b, or of skeletal lineage markers, such as MyoD and Mox1. To identify the minimal region of Nkx2-5 that can trigger cardiomyogenesis, we analyzed the activity of various Nkx2-5 deletion mutants. The C-terminal domain was not necessary for the ability of Nkx2-5 to induce cardiomyogenesis and loss of this domain did not enhance myogenesis. Therefore, Nkx2-5 function is essential for commitment of mesoderm into the cardiac muscle lineage, and the N-terminal region, together with the homeodomain, is sufficient for cardiomyogenesis in P19 cells.

Wnt signaling regulates the function of MyoD and myogenin.

The myogenic regulatory factors (MRFs), MyoD and myogenin, can induce myogenesis in a variety of cell lines but not efficiently in monolayer cultures of P19 embryonal carcinoma stem cells. Aggregation of cells expressing MRFs, termed P19[MRF] cells, results in an approximately 30-fold enhancement of myogenesis. Here we examine molecular events occurring during P19 cell aggregation to identify potential mechanisms regulating MRF activity. Although myogenin protein was continually present in the nuclei of >90% of P19[myogenin] cells, only a fraction of these cells differentiated. Consequently, it appears that post-translational regulation controls myogenin activity in a cell lineage-specific manner. A correlation was obtained between the expression of factors involved in somite patterning, including Wnt3a, Wnt5b, BMP-2/4, and Pax3, and the induction of myogenesis. Co-culturing P19[Wnt3a] cells with P19[MRF] cells in monolayer resulted in a 5- to 8-fold increase in myogenesis. Neither BMP-4 nor Pax3 was efficient in enhancing MRF activity in unaggregated P19 cultures. Furthermore, BMP-4 abrogated the enhanced myogenesis induced by Wnt signaling. Consequently, signaling events resulting from Wnt3a expression but not BMP-4 signaling or Pax3 expression, regulate MRF function. Therefore, the P19 cell culture system can be used to study the link between somite patterning events and myogenesis.

Myocyte enhancer factor 2C upregulates MASH-1 expression and induces neurogenesis in P19 cells.

MEF2C is a transcription factor expressed in neural lineages. After transient transfection, the MEF2 family of factors can act synergistically with the neural-specific transcription factor, MASH-1, and activate exogenous neural-specific promoters. To determine whether MEF2C is capable of modulating endogenous gene expression, P19 cell lines were analyzed that overexpressed MEF2C, termed P19[MEF2C] cells. Here we show that P19[MEF2C] cells differentiate into neurons when aggregated with ME(2)SO. MEF2C-induced neurons expressed neurofilament protein, the nuclear antigen NeuN, as well as MASH-1. Our results indicate that MEF2C can directly or indirectly activate the expression of MASH-1, leading to neurogenesis.

MAN1, an inner nuclear membrane protein that shares the LEM domain with lamina-associated polypeptide 2 and emerin.

The "MAN antigens" are polypeptides recognized by autoantibodies from a patient with a collagen vascular disease and localized to the nuclear envelope. We now show that one of the human MAN antigens termed MAN1 is a 82.3-kDa protein with an amino-terminal domain followed by two hydrophobic segments and a carboxyl-terminal tail. The MAN1 gene contains seven protein-coding exons and is assigned to human chromosome 12q14. Its mRNA is approximately 5.5 kilobases and is detected in several different cell types that were examined. Cell extraction experiments show that MAN1 is an integral membrane protein. When expressed in transfected cells, MAN1 is exclusively targeted to the nuclear envelope, consistent with an inner nuclear membrane localization. Protein sequence analysis reveals that MAN1 shares a conserved globular domain of approximately 40 amino acids, which we term the LEM module, with inner nuclear membrane proteins lamina-associated polypeptide 2 and emerin. The LEM module is also present in two proteins of Caenorhabditis elegans. These results show that MAN1 is an integral protein of the inner nuclear membrane that shares the LEM module with other proteins of this subcellular localization.

Myocyte enhancer factor 2C and myogenin up-regulate each other's expression and induce the development of skeletal muscle in P19 cells.

Two families of transcription factors, myogenic regulatory factors (MRFs) and myocyte enhancer factor 2 (MEF2), function synergistically to regulate myogenesis. In addition to activating structural muscle-specific genes, MRFs and MEF2 activate each other's expression. The MRF, myogenin, can activate MEF2 DNA binding activity when transfected into fibroblasts and, in turn, the myogenin promoter contains essential MEF2 DNA binding elements. To determine which MEF2 is involved in this regulation, P19 cells stably expressing MyoD and myogenin were compared for their ability to activate the expression of MEF2 family members. There was very little cross-activation of MyoD expression by myogenin and vice versa. Myogenin expression, and not MyoD, was found to up-regulate MEF2C expression. MEF2A, -B, and -D expression levels were not up-regulated by overexpression of either MyoD or myogenin. To examine whether MEF2C can differentially regulate MyoD or myogenin expression, P19 cell lines overexpressing MEF2C were analyzed. MEF2C induced myogenesis in P19 cells and up-regulated the expression of myogenin with 25-fold greater efficiency than that of MyoD. Therefore, myogenin and MEF2C participate in a regulatory loop in differentiating stem cells. This positive regulation does not extend to MyoD or the other MEF2 family members. Consequently, MEF2C appears to play a specific role in early events of myogenesis.

Cloning, tissue distribution, subcellular localization and overexpression of murine histidine-rich Ca2+ binding protein.

The histidine-rich Ca2+ binding protein (HRC) resides in the sarcoplasmic reticulum of muscle and binds Ca2+. Since Ca2+ concentrations can regulate gene expression via calcineurin, the mouse homologue of HRC (mHRC) was isolated and characterized. mHRC was detected in muscle progenitor cells, in primary clonal thymic tumors and a tumor cell line, suggesting a broader role for mHRC than in Ca2+ storage during muscle contraction. mHRC was present in the perinuclear region of myoblasts. To examine if it can regulate gene expression, mHRC was overexpressed in cells differentiating into cardiac and skeletal muscle. mHRC had no effect on cardiogenesis or myogenesis. Therefore, if mHRC plays a role in the regulation of gene expression during cellular differentiation, it does not appear to be either rate-limiting or inhibitory.

Cardiac and skeletal muscle development in P19 embryonal carcinoma cells.

Mouse P19 embryonal carcinoma cells are pluripotent stem cells that can be maintained in culture in an undifferentiated state or can be induced to differentiate in vitro into multiple cell types. P19 cells aggregated in the presence of dimethylsulfoxide differentiate into spontaneously beating cardiomyocytes and bipolar skeletal myocytes that exhibit the biochemical and physiologic properties of their embryonic equivalents. P19 cells can be readily manipulated genetically, resulting in the loss or over-expression of a gene of interest. Because of this versatility, the P19 system is suited for examining the molecular mechanisms controlling the developmental decisions of stem cells differentiating into the skeletal or cardiac muscle lineage.

Myocyte enhancer factor 2C and Nkx2-5 up-regulate each other's expression and initiate cardiomyogenesis in P19 cells.

The Nkx2-5 homeodomain protein plays a key role in cardiomyogenesis. Ectopic expression in frog and zebrafish embryos results in an enlarged myocardium; however, expression of Nkx2-5 in fibroblasts was not able to trigger the development of beating cardiac muscle. In order to examine the ability of Nkx2-5 to modulate endogenous cardiac specific gene expression in cells undergoing early stages of differentiation, P19 cell lines overexpressing Nkx2-5 were differentiated in the absence of Me2SO. Nkx2-5 expression induced cardiomyogenesis in these cultures aggregated without Me2SO. During differentiation into cardiac muscle, Nkx2-5 expression resulted in the activation of myocyte enhancer factor 2C (MEF2C), but not MEF2A, -B, or -D. In order to compare the abilities of Nkx2-5 and MEF2C to induce cellular differentiation, P19 cells overexpressing MEF2C were aggregated in the absence of Me2SO. Similar to Nkx2-5, MEF2C expression initiated cardiomyogenesis, resulting in the up-regulation of Brachyury T, bone morphogenetic protein-4, Nkx2-5, GATA-4, cardiac alpha-actin, and myosin heavy chain expression. These findings indicate the presence of a positive regulatory network between Nkx2-5 and MEF2C and show that both factors can direct early stages of cell differentiation into a cardiomyogenic pathway.

A splice variant of the ITF-2 transcript encodes a transcription factor that inhibits MyoD activity.

Proteins of the basic helix-loop-helix (bHLH) family are transcription factors that bind DNA containing the E box motif (CANNTG) found in the promoters of many muscle-specific genes. ITF-2 is a bHLH protein with widespread expression that is thought to form active heterodimers with MyoD, a muscle-specific bHLH transcription factor. We have isolated cDNAs derived from two alternatively spliced forms of mouse ITF-2, termed MITF-2A and -2B. These proteins differ in their N termini. Neither MITF-2A nor -2B transactivated the cardiac alpha-actin promoter, which contains an E box, when transfected into nonmuscle cells. In fact, MITF-2B inhibited MyoD activation of the cardiac alpha-actin promoter. This inhibitory activity required the N-terminal 83 amino acids since MITF-2A showed no inhibitory activity, and a mutant MITF-2B with deletion of the N-terminal 83 amino acids failed to inhibit MyoD-mediated transcriptional activation. MyoD activity was also inhibited by Id, a HLH protein, and this inhibition was reversed by the addition of excess E12 or MITF-2A. However, the inhibition of MyoD activity by MITF-2B was not reversed with E12 or MITF-2A. While Id is thought to inhibit MyoD by binding and sequestering potential dimerization partners, MITF-2B appears to inhibit MyoD activity by forming an inactive heterodimer with MyoD. Thus, differentially spliced transcripts of mouse ITF-2 encode different proteins that appear to dimerize with MyoD and activate or repress transcription.

Cells differentiating into neuroectoderm undergo apoptosis in the absence of functional retinoblastoma family proteins.

The retinoblastoma (RB) protein is present at low levels in early mouse embryos and in pluripotent P19 embryonal carcinoma cells; however, the levels of RB rise dramatically in neuroectoderm formed both in embryos and in differentiating cultures of P19 cells. To investigate the effect of inactivating RB and related proteins p107 and p130, we transfected P19 cells with genes encoding mutated versions of the adenovirus E1A protein that bind RB and related proteins. When these E1A-expressing P19 cells were induced to differentiate into neuroectoderm, there was a striking rise in the expression of c-fos and extensive cell death. The ultrastructural and biochemical characteristics of the dying cells were indicative of apoptosis. The dying cells were those committed to the neural lineages because neurons and astrocytes were lost from differentiating cultures. Cell death was dependent on the ability of the E1A protein to bind RB and related proteins. Our results suggest that proteins of the RB family are essential for the development of the neural lineages and that the absence of functional RB activity triggers apoptosis of differentiating neuroectodermal cells.

Cellular aggregation enhances MyoD-directed skeletal myogenesis in embryonal carcinoma cells.

When introduced into P19 embryonal carcinoma cells, recombinant genes encoding MyoD converted only a small percentage (< 3%) of the transfected cells into skeletal muscle. We isolated stably transfected cells that expressed the MyoD transcript. These P19[MyoD] cells continued to express markers characteristic of undifferentiated stem cells but also expressed myf-5 and the myotonic dystrophy kinase, transcripts normally present in myoblasts but absent from P19 cells. Aggregation of P19[MyoD] cells induced the expression of myogenin, desmin, and the retinoblastoma protein and resulted in the rapid and abundant development of skeletal muscle. Both the embryonic and the slow isoforms of myosin heavy chain were present in this muscle, indicating that it resembled skeletal muscle formed from primary myoblasts. Since aggregation of P19 cells normally results in inefficient differentiation and the development of only low levels of cardiac muscle but no skeletal muscle, we conclude that MyoD imposes the skeletal muscle program on P19 cells and that the differentiation of these cells requires inductive events provided by cell aggregation.

The E box is essential for activity of the cardiac actin promoter in skeletal but not in cardiac muscle.

The sequence CANNTG (E box) is frequently found in the promoters of muscle-specific genes and binds members of the basic helix-loop-helix (bHLH) family of transcription factors. We compared the need for the E box in the expression of a muscle-specific promoter normally expressed in both cardiac and skeletal muscle. The E box was mutated in a construct of the cardiac alpha-actin promoter driving the Escherichia coli lacZ gene. The wild-type and mutant constructs were transfected and stably integrated into the genomes of P19 embryonal carcinoma cells. The wild-type promoter was expressed in both cardiac and skeletal myocytes. The promoter lacking an E box was expressed in cardiac but not in skeletal muscle. Neither promoter was active in nonmuscle cells. Thus the E box is not necessary for the cardiac actin promoter activity in P19-derived cardiac muscle but is essential for its activity in skeletal muscle. This result is consistent with our inability to detect cardiac muscle-specific members of the MyoD family of bHLH transcription factors.

Deletion of NH2- and COOH-terminal sequences destroys function of the Ca2+ ATPase of rabbit fast-twitch skeletal muscle sarcoplasmic reticulum.

Deletion mutants of the Ca2+ ATPase of rabbit fast-twitch skeletal muscle sarcoplasmic reticulum (SERCA1a) were constructed and expressed in COS-1 cells. The mutants were expressed at levels 7- to 15-fold lower than the wild-type and were inactive. In vitro transcription-translation-insertion experiments showed that deletion of transmembrane sequences M1 and M2, but not of M8, M9, M10 or the NH2-terminal 30 amino acids inhibited the stable insertion of the enzyme into the membrane. Thus there was no correlation between loss of function and membrane insertion. A signal sequence for membrane insertion may exist in M1 and M2.

Mutation of glutamate 309 to glutamine alters one Ca(2+)-binding site in the Ca(2+)-ATPase of sarcoplasmic reticulum expressed in Sf9 cells.

Sf9 cells infected with a baculovirus vector containing SERCA1 cDNA expressed immunoreactive rabbit fast-twitch muscle Ca(2+)-ATPase at levels up to 3 mg/liter. The microsomal fraction isolated from infected Sf9 cells catalyzed Ca2+ transport at rates 6-fold above control values. To obtain direct evidence for the postulate (Clarke, D. M., Loo, T. W., Inesi, G., and MacLennan, D. H., et al. (1989) Nature 339, 476-478) that Glu309 contributes to a Ca(2+)-binding site in the transmembrane sector of the Ca(2+)-ATPase, Ca2+ binding to wild type and mutant (Glu309 to Gln) Ca(2+)-ATPases was measured. The wild type Ca(2+)-ATPase, expressed in Sf9 cells and purified using a monoclonal antibody bound to Sepharose beads, bound approximately 1.6-1.7 mol Ca2+/mol of enzyme at 2 microM Ca2+ in a buffer favoring the E1 conformation of the enzyme and at 10 microM Ca2+ in a buffer favoring the E2 conformation. Under identical conditions, the mutant Ca(2+)-ATPase bound less than 0.1 mol of Ca2+/mol of enzyme in E1 buffer, but 0.8 mol Ca2+/mol in the E2 buffer. In spite of the ability of the Glu309 to Gln mutant enzyme to bind about 1 mol of Ca2+/mol of enzyme, E2P formation was not inhibited by up to 100 microM Ca2+, and E1P formation from ATP and Ca2+ was not observed with up to 100 microM Ca2+ in intact microsomal vesicles from Sf9 cells. Nevertheless, with detergent-solubilized and purified mutant Ca(2+)-ATPases, E2P formation was inhibited with a K0.5 of 2 microM Ca2+. These observations are consistent with the view that a single intact Ca(2+)-binding site is present in the mutant Ca(2+)-ATPase, which is accessible to Ca2+ only from the lumenal side and only in the E2 conformation. Transition from E2 to E1-Ca2+ may occur during or following Ca2+ binding, accounting for the relatively high Ca2+ affinity and inhibition by Ca2+ of phosphorylation from Pi.

Site-directed mutagenesis of the Ca2+ ATPase of sarcoplasmic reticulum.

We have cloned cDNA encoding the Ca2+ ATPase from fast-twitch skeletal muscle and, on the basis of its amino acid sequence and immunological studies of its topology, have made deductions concerning its secondary structure and active sites. These deductions have led us to test models for Ca2+ transport through expression of the protein in functional form in COS-1 cells, mutagenesis, and measurement of altered function. Mutation of about 250 of the 1000 amino acids making up the Ca2+ pump has indicated that the sites of high affinity Ca2+ binding are located in the center of the transmembrane domain and are made up from residues located in transmembrane sequences M4, M5, M6 and M8. The ATP binding site appears to be located in the headpiece and is made up from a series of loop sequences connecting alternating alpha helices and beta strands. Sites of conformational interaction appear in all domains throughout the Ca2+ pump. In our present model, Ca2+ transport occurs through binding to high affinity sites accessible to the cytoplasm in the E1 conformation, followed by release to the lumen from low affinity sites which form during the ATP-induced transition of the protein from the E1 to the E2 conformation.

Import of precursor proteins into mitochondria: site of polypeptide unfolding.

The matrix-targeting signal of mitochondrial preornithine carbamyl transferase has been fused to either murine dihydrofolate reductase (pODHFR) or bacterial chloramphenicol acetyltransferase (pOCAT). Loosening of the tightly folded "native" structure of the two proteins following their synthesis in a rabbit reticulocyte lysate was assayed by the acquisition of protease sensitivity (pODHFR and pOCAT) or by the loss of enzyme activity (pOCAT). By these criteria, the bulk population of both precursor proteins was tightly folded following release from the ribosome, even in the presence of ATP and excess reticulocyte lysate. Neither protein unfolded as a consequence of binding to the surfaces of anionic liposomes or intact mitochondria. However, a non-native form of full-length pOCAT, exhibiting a loss of enzymatic activity and an enhanced protease sensitivity, was detected in association with a submitochondrial fraction that banded between the inner and outer mitochondrial membrane fractions on sucrose density gradients. Delivery of the precursor molecule to this position required ATP and a proteinaceous component on the surface of the organelle.

Identification of hydrophobic residues in the signal sequence of mitochondrial preornithine carbamyltransferase that enhance the rate of precursor import.

Previous studies employing circular dichroism and resonance energy transfer techniques have demonstrated that the signal peptide of mitochondrial preornithine carbamyltransferase (pOCT) has the potential to interact with the surface of an anionic phospholipid membrane via a short amphiphilic helical domain. Here we have used predictive secondary structure computations as a guide to localize the putative membrane binding region in the pOCT signal sequence and demonstrate that replacement of leucine residues at positions 5, 8, and 9 with the less hydrophobic residue, alanine, significantly reduces the rate of precursor import (4-5-fold compared to wild type); the amino acid substitutions had little effect, however, on the ability of a mitochondrial matrix extract to process the mutant precursor polypeptide. The mutant precursor bound to anionic liposomes with a lower affinity compared to wild-type pOCT and was inhibited to a lesser extent than pOCT during import into mitochondria in the presence of varying concentrations of liposomes. Taken together, the results suggest that this small region of the pOCT signal sequence, containing a limited number of critical hydrophobic residues, contributes to the optimal rate of precursor import, perhaps by functioning as a membrane surface-seeking entity.