Natural and synthetic DNA elements with the CArG motif differ in expression and protein-binding properties.

DNA elements with the CC(A/T)6GG, or CArG, motif occur in promoters that are under different regulatory controls. CArG elements from the skeletal actin, c-fos, and myogenin genes were tested for their abilities to confer tissue-specific expression on reporter genes when the individual elements were situated immediately upstream from a TATA element. The c-fos CArG element, also referred to as the serum response element (SRE), conferred basal, constitutive expression on the test promoter. The CArG motif from the myogenin gene was inactive. The skeletal actin CArG motif functioned as a muscle regulatory element (MRE) in that basal expression was detected only in muscle cultures. Muscle-specific expression from the 28-bp MRE and the 2.3-kb skeletal actin promoter was trans repressed by the Fos and Jun proteins. The expression and factor-binding properties of a series of synthetic CArG elements were analyzed. Muscle-specific expression was conferred by perfect 28-bp palindromes on the left and right halves of the skeletal actin MRE. Chimeric elements of the skeletal actin MRE and the c-fos SRE differed in their expression properties. Muscle-specific expression was observed when the left half of the MRE was fused to the right half of the SRE. Constitutive expression was conferred by a chimera with the right half of the MRE fused to the left half of the SRE and by chimeras which exchanged the central CC(A/T)6GG sequences. At least three distinct proteins specifically bound to these CArG elements. The natural and synthetic CArG elements differed in their affinities for these proteins; however, muscle-specific expression could not be attributed to differences in the binding of a single protein. Furthermore, the MRE did not bind MyoD or the myogenin-E12 heterodimer, indicating that muscle-specific expression from this element does not involve a direct interaction with these helix-loop-helix proteins. These data demonstrate that the conserved CArG motifs form the core of a family of functionally different DNA regulatory elements that may contribute to the tissue-specific expression properties of their cognate promoters.     

Distinct enhancers regulate skeletal and cardiac muscle-specific expression programs of the cardiac alpha-actin gene in Xenopus embryos

During vertebrate embryonic development, cardiac and skeletal muscle originates from distinct precursor populations. Despite the profound structural and functional differences in the striated muscle tissue they eventually form, such progenitors share many features such as components of contractile apparatus. In vertebrate embryos, the alpha-cardiac actin gene encodes a major component of the myofibril in both skeletal and cardiac muscle. Here, we show that expression of Xenopus cardiac alpha-actin in the myotomes and developing heart tube of the tadpole requires distinct enhancers within its proximal promoter. Using transgenic embryos, we find that mutations in the promoter-proximal CArG box and 5 bp downstream of it specifically eliminate expression of a GFP transgene within the developing heart, while high levels of expression in somitic muscle are maintained. This sequence is insufficient on its own to limit expression solely to the myocardium, such restriction requiring multiple elements within the proximal promoter. Two additional enhancers are active in skeletal muscle of the embryo, either one of which has to interact with the proximal CArG box for correct expression to be established. Transgenic reporters containing multimerised copies of CArG box 1 faithfully detect most sites of SRF expression in the developing embryo as do equivalent reporters containing the SRF binding site from the c-fos promoter. Significantly, while these motifs possess a different A/T core within the CC(A/T)(6)GG consensus and show no similarity in flanking sequence, each can interact with a myotome-specific distal enhancer of cardiac alpha-actin promoter, to confer appropriate cardiac alpha-actin-specific regulation of transgene expression. Together, these results suggest that the role of CArG box 1 in the cardiac alpha-actin gene promoter is to act solely as a high-affinity SRF binding site.     

Functional serum response elements upstream of the growth factor-inducible gene zif268.

The zif268 gene, which encodes a protein with three typical zinc finger sequences, is induced in mouse 3T3 cells by serum, phorbol 12-myristate 13-acetate platelet-derived growth factor, and fibroblast growth factor. The induction is coordinate with that of c-fos. The 5'-flanking region of zif268 contains sequences that resemble known regulatory elements, including four CC(A or T)6GG sequences similar to the core serum response elements (SREs) found upstream of c-fos and actin genes. To determine whether the zif268 SRE-like elements mediate induction, CAT (chloramphenicol acetyltransferase) plasmids with different lengths of zif268 upstream sequences were tested for inducibility in 3T3 cells by serum, platelet-derived growth factor, or phorbol 12-myristate 13-acetate. In addition, double-stranded oligonucleotides corresponding to each of the four zif268 putative SREs were tested individually for responsiveness when placed upstream of a thymidine kinase gene promoter. Each of the four SREs conferred inducibility by the agents tested, and multiple SREs resulted in greater inducibility than did a single element. Each of the zif268 SREs also competed with the c-fos SRE for binding by serum response factor present in HeLa cell nuclear extract. We conclude that the zif268 SRE-like sequences are functional and probably account for the coordinate induction of zif268 and c-fos.     

Nucleotide sequence of the chicken cardiac alpha actin gene: absence of strong homologies in the promoter and 3'-untranslated regions with the skeletal alpha actin sequence

The entire nucleotide sequence of the chicken cardiac alpha-actin (CC alpha A) gene has been determined. This is the first complete sequence of a cardiac actin gene that includes the promoter region, cap site, all the introns, and the polyadenylation site. The gene contains six introns, five of which interrupt the coding region at amino acids (aa) 41, 150, 204, 267, and 327. The first intron is in the 5'-noncoding region and is 438 bp in length. The CC alpha A gene encodes an mRNA of approx. 1400 bp with 5'- and 3'-untranslated region of 59 and 184 nucleotides (nt), respectively. Like the chicken skeletal alpha-actin gene, the CC alpha A gene has the codon for the aa cysteine between the initiator ATG and the codon for the N-terminal aspartic acid residue of the mature protein. There are no strong homologies (less than 13 consecutive nt) in the promoter or 3'-untranslated regions between the CC alpha A and chicken skeletal alpha-actin genes even though both are expressed in skeletal muscle during development. However, the 3'-untranslated region of the CC alpha A gene demonstrates significant sequence homology (76% over a 200-nt region) with the same region in the partial sequence of the human cardiac gene. The conservation of these sequence homologies between identical isoforms rather than the different alpha actin genes suggests these conserved regions may have a role in regulation rather than tissue-specific expression, as previously proposed.     

Expression of the 53 kD forked protein rescues F-actin bundle formation and mutant bristle phenotypes in Drosophila.

forked mutations affect bristle development in Drosophila pupae, resulting in short, thick, gnarled bristles in the adult. The forked proteins are components of 200-300-microm-long actin fiber bundles that are present transiently during pupal development [Petersen et al., 1994: Genetics 136:173-182]. These bundles are composed of segments of 3-10 microm long, and forked protein is localized along the actin fiber bundle segments and accumulates at the junctions connecting them longitudinally. In the forked mutants, f(36a) and f(hd), F-actin bundles are greatly reduced in number and size, and bundle segmentation is absent. The p-element, P[w(+), falter] contains a 5.3-kb fragment of the forked gene that encodes the 53-kD forked protein [Lankenau et al., 1996: Mol Cell Biol 16:3535-3544]. Expression of only the 53-kD forked protein is sufficient to rescue the actin bundle and bristle phenotypes of f(36a) and f(hd) mutant flies. The 5.3-kb forked sequence, although smaller than the 13-kb region previously shown to rescue forked mutants [Petersen et al., 1994: Genetics 136:173-182], does contain the core forked sequence that encodes actin binding and bundling domains in cultured mammalian cells [Grieshaber and Petersen, 1999: J Cell Sci 112:2203-2211]. These data show that the 53-kD forked protein is sufficient for normal bristle development and that the domains shown previously to be important for actin bundling in cell culture may be all that are required for normal actin bundle formation in developing Drosophila bristles.     

Two cellular proteins bind specifically to a purine-rich sequence necessary for the destabilization function of a c-fos protein-coding region determinant of mRNA instability.

The c-fos proto-oncogene mRNA is rapidly degraded within minutes after its appearance in the cytoplasm of growth factor-stimulated mammalian fibroblasts. At least two functionally independent sequence elements are responsible for the lability of c-fos mRNA. One of these determinants is located within a 0.32-kb sequence present in the protein-coding region. We demonstrate by gel mobility shift experiments and UV cross-linking that at least two protein factors specifically interact with a 56-nucleotide purine-rich sequence located at the 5' end of the 0.32-kb coding region determinant of mRNA instability (CRDI). One protein is predominantly associated with the polysomes, while the other is detected in the post-ribosomal supernatant. Sequence comparison of members of the fos gene family revealed that the high purine content of the protein-binding region is conserved through evolution. Deletion of this region from the 0.32-kb CRDI severely impedes its function as an RNA-destabilizing element. Our results suggest that binding of the two proteins to the purine-rich sequence may participate in the rapid mRNA decay mediated by this 0.32-kb c-fos CRDI.     

Multiple CArG boxes in the human cardiac actin gene promoter required for expression in embryonic cardiac muscle cells developing in vitro from embryonal carcinoma cells.

Chimeric genes composed of the human cardiac actin promoter driving the Escherichia coli lacZ reporter gene were constructed, transfected, and stably integrated into genomes of P19 embryonal carcinoma cells. The transfected constructs were expressed actively in cardiac myocytes formed following dimethyl sulfoxide (DMSO)-induced cell differentiation but poorly in undifferentiated cultures and in cultures treated with retinoic acid to develop into derivatives of the neuroectoderm. A number of deletions of the promoter were constructed and tested. Three regions required for efficient expression in P19-derived cardiac muscle were identified, each containing sequences referred to as CArG boxes (CC[AT-rich]6GG). This analysis indicated that regulatory sequences important for expression in cardiac muscle were present upstream of the core promoter identified previously by transient assays in skeletal myoblasts. Expression of the cardiac actin promoter was enhanced 10-fold in undifferentiated P19 cells in the presence of the myoD protein. The promoter regions important for expression in P19-derived cardiocytes were similar to those important for myoD-induced enhancement, a result we interpret to be consistent with the idea that cardiac muscle might contain a myoD-like activity.     

Regulation of the human beta-actin promoter by upstream and intron domains.

We have identified three regulatory domains of the complex human beta-actin gene promoter. They span a region of about 3000 bases, from not more than -2011 bases upstream of the mRNA cap site to within the 5' intron (832 bases long). A distal upstream domain contains at least one enhancer-like element. A proximal upstream domain, with a CArG [for CC(A + T rich)6GG] motif found in all known mammalian actin genes, seems to confer serum, but not growth factor, inducibility. The third domain is within the evolutionarily conserved 3' region of the first intron and contains a 13 base-pair sequence, identical to the upstream sequence with the CArG motif. This domain also contains sequences that are both serum and fibroblast growth factor inducible.