Positive Trans-acting Factors Research Articles

Molecular Characterization of Three Ethylene Responsive Element Binding Factor Genes from Cotton

Ethylene-responsive factors (ERFs) are important regulators of plant gene expression. In this study, three novel ERF genes, GhERF2, GhERF3 and GhERF6, were isolated from cotton (Gossypium hirstum) using rapid amplification of cDNA ends-polymerase chain reaction. Transient expression analysis using GhERF-green fluorescent protein fusions showed that these three proteins were targeted to the nucleus. Fusion proteins consisting of GhERF2, GhERF3 or GhERF6 coupled to the GAL4 DNA binding domain strongly activated transcription in yeast. Furthermore, GhERF6 was shown to be able to bind specifically to GCC boxes using a particle bombardment assay in tobacco cells. Semi-quantitative reverse transcription-polymerase chain reaction revealed that GhERF2 and GhERF3 are constitutively expressed in all organs, while GhERF6 is only constitutively expressed in vegetative organs. When plants were treated with ethylene, abscisic acid, salt, cold and drought, the transcripts of GhERF2, GhERF3 and GhERF6 were rapidly induced to high levels. Promoter analysis also indicated that the 5' upstream regions of the three genes possess elements induced by these physiological and environmental factors. Collectively, our data suggest that GhERF2, GhERF3 and GhERF6 might function as positive trans-acting factors in the plant responses to ethylene, abscisic acid and other stresses and provide useful clues for further research into the mechanism of them in regulating cotton multiple stress responses.

The synergetic effects of two CCAAT boxes in Aspergillus niger glaA gene promoter on activation of PglaA transcription.

EMSA and footprinting analyses have revealed that the -489-- -414 bp and the -390-- -345 bp (designated DC and PC respectively) upstream of the Aspergillus nigerT21 glaA gene were bound by one protein factor in the A. nigerT21 whole cell extract. Both DC and PC contained CCAAT pentanucleotides. The functions of DC and PC in regulation of expression of glucoamylase (GLA) were studied. CCAAT pentanucleotides were replaced with CGTAA and the mutated DNA fragments DCm and PCm lost the binding activities of protein factors in vitro. In vivo when either DC or PC was mutated or the relative orientations between them were changed on the PglaA, the transcriptional activity of PglaA decreased to a basal level. Introduction of multi-copies of DC into the original site at the PglaA in A. nigerT21 decreased the expression of endogenous GLA expression and the exogenous reporter E. coli uidA gene introduced under the PglaA promoter, while having no effect on the uidA gene under the control of PgpdA. EMSA revealed that the levels of the specific DNA-binding protein factors in the transformants maintained the same meaning that introduction of multi-copies of DC caused the titration effect. AnghapC gene was cloned from A. nigerT21 cDNA and introduced into the DC multi-copied strains. The expression of AnghapC improved the expression of the endogenous GLA and the exogenous gene controlled by PglaA. These results showed that both the CCAAT pentanucleotides were necessary for DC and PC binding to the protein factors, and the simultaneous binding of DC and PC to the protein was necessary for promoting the transcriptional activity of PglaA. AngHapC was the specific positive trans-acting protein factor binding to DC.

Overexpression of the tobacco Tsi1 gene encoding an EREBP/AP2-type transcription factor enhances resistance against pathogen attack and osmotic stress in tobacco.

Using mRNA differential display analysis, we isolated a salt-induced transcript that showed a significant sequence homology with an EREBP/AP2 DNA binding motif from oilseed rape plants. With this cDNA fragment as a probe, cDNA clone Tsi1 (for Tobacco stress-induced gene1) was isolated from a tobacco cDNA library. RNA gel blot analysis indicated that transcripts homologous with Tsi1 were induced not only in NaCl-treated leaves but also in leaves treated with ethephon or salicylic acid. Transient expression analysis using a Tsi1::smGFP fusion gene in BY-2 cells indicated that the Tsi1 protein was targeted to the nucleus. Fusion protein of Tsi1 with GAL4 DNA binding domain strongly activated transcription in yeast, and the transactivating activity was localized to the 13 C-terminal amino acids of Tsi1. Electrophoretic mobility shift assays revealed that Tsi1 could bind specifically to the GCC and the DRE/CRT sequences, although the binding activity to the former was stronger than that to the latter. Furthermore, Agrobacterium-mediated transient expression and transgenic plants expressing Tsi1 demonstrated that overexpression of the Tsi1 gene induced expression of several pathogenesis-related genes under normal conditions, resulting in improved tolerance to salt and pathogens. These results suggest that Tsi1 might be involved as a positive trans-acting factor in two separate signal transduction pathways under abiotic and biotic stress.

The Aspergillus nidulans multimeric CCAAT binding complex AnCF is negatively autoregulated via its hapB subunit gene

Cis-acting CCAAT elements are frequently found in eukaryotic promoter regions. Many of them are bound by conserved multimeric complexes. In the fungus Aspergillus nidulans the respective complex was designated AnCF ( A. nidulans CCAAT binding factor). AnCF is composed of at least three subunits designated HapB, HapC and HapE. Here, we show that the promoter regions of the hapB genes in both A. nidulans and Aspergillus oryzae contain two inversely oriented, conserved CCAAT boxes (box α and box β). Electrophoretic mobility shift assays (EMSAs) using both nuclear extracts and the purified, reconstituted AnCF complex indicated that AnCF binding in vitro to these boxes occurs in a non-mutually exclusive manner. Western and Northern blot analyses showed that steady-state levels of HapB protein as well as hapB mRNA were elevated in hapC and hapE deletion mutants, suggesting a repressing effect of AnCF on hapB expression. Consistently, in a hapB deletion background the hapB-lacZ expression level was elevated compared with the expression in the wild-type. This was further supported by overexpression of hapB using an inducible alcA-hapB construct. Induction of alcA-hapB expression strongly repressed the expression of a hapB-lacZ gene fusion. However, mutagenesis of box β led to a fivefold reduced expression of a hapB-lacZ gene fusion compared with the expression derived from a wild-type hapB-lacZ fusion. These results indicate that (i) box β is an important positive cis-acting element in hapB regulation, (ii) AnCF does not represent the corresponding positive trans-acting factor and (iii) that AnCF is involved in repression of hapB.

Cloning and characterization of the rat TIS11 gene.

The TIS11 gene is an immediate early gene that is induced rapidly and transiently by phorbol 12-myristate 13-acetate and various growth factors. To study transcriptional regulation of the gene, a genomic clone of rat TIS11 was isolated, and the organization of exon-intron structure and transcriptional initiation site were determined. The rat TIS11 gene consisted of 2 exons spanning approximately 2.5 kb. Several canonical sequences for binding of transcriptional factors were found in the 5'-flanking region. The 5.3 kb of the 5'-flanking region fused to a luciferase reporter gene showed promoter activity when introduced into rat pheochromocytoma PC12 cells. Analyses with serial 5'-deletion mutants suggested that the major positive regulatory region is located at the region of -241 to -76, and that the minimum promoter region is within the 76-bp upstream of the transcriptional initiation site. Gel mobility shift assays revealed that PC12 cell nuclear proteins specifically bind to the major positive regulatory region of the TIS11 gene. The identified nuclear protein components may act as the positive trans-acting factors in the basal expression of the TIS11 gene in PC12 cells.

Promoter regions involved in density-dependent regulation of basic fibroblast growth factor gene expression in human astrocytic cells.

Expression of mitogenic basic fibroblast growth factor (bFGF) in the central nervous system is inhibited by direct cell contact and is implicated in reactive and neoplastic transformation of astrocytes. The molecular mechanisms controlling expression of bFGF were examined in cultures of human astrocytes. Cell-density-dependent depletion of bFGF mRNA levels parallels changes in bFGF gene protein. Regulation of transcription of a bFGF luciferase reporter gene containing an upstream region (bp -1800 to +314) of the bFGF gene promoter mimicks the density-dependent regulation of the endogenous bFGF gene in transfected astrocytes. Deletion analysis has identified a fragment (bp -650 to -513) and sequences further downstream (bp -274 to +314) as the regions required for the regulation of bFGF gene activity by cell density. Unlike in astrocytes, changing the cell density of glioma cell cultures does not affect the levels of bFGF protein and mRNA. bFGF luciferase constructs were expressed at the same level in high- or low-density cultures of glioma cells, indicating altered regulation of the bFGF gene promoter. Electrophoretic mobility shift assays showed binding of nuclear proteins to a fragment of bFGF gene promoter from bp -650 to -453. This binding was abolished by a deletion of the upstream cell-density-responsive region (bp -650 to -512). Binding was observed with nuclear extracts from subconfluent astrocytes but was reduced in extracts from confluent astrocytes. Our results indicate that induction of bFGF in astrocytes upon reduction of cell density is mediated transcriptionally by positive trans-acting factors interacting with bFGF promoter. In contrast, nuclear proteins from glioma cells bind to the promoter region from bp -650 to -453 independent of cell density. Thus, the constitutive binding of trans-acting factor(s) to the region of the bFGF promoter from bp -650 to -453 may be responsible for the continuous expression of bFGF that leads to the uncontrolled growth of glioma cells.

Altered DNA/protein complexes specific for the beta-interferon regulatory region observed in murine embryonal carcinoma F9 cells.

Murine embryonal carcinoma (EC) F9 cells do not produce interferon (IFN) at the protein or RNA level in response to inducing agents, while retinoic acid differentiated F9 cells do produce IFN. A probe was constructed spanning positions -104 to -39 of the human beta-IFN upstream regulatory region to examine this developmental control at the level of a transcriptional regulatory mechanism. Gel mobility shift analyses were used to examine this molecular mechanism to determine whether the differential expression of positive or negative trans-acting factors may act to control beta-IFN expression in undifferentiated EC cells. These analyses showed that while nuclear extracts from poly-I,C induced L929 cells, in the IFN producing cell line, showed two shifted bands, nuclear extracts from both induced and uninduced F9 cells showed only one shifted band using the -104/-39 probe. While this single shifted band co-migrated with the faster migrating species of L929 cell extracts, competition analysis revealed differences between the two complexes. An oligonucleotide representing the positive regulatory domain PRDII competed efficiently for the probe when induced F9 cell extracts were examined, but failed to compete when induced L929 cell extracts were examined. In contrast, an oligonucleotide representing the positive regulatory domain PRDI competed very well when induced L929 cell extracts were examined but had only a minimal effect when induced F9 cell extracts were examined. These data suggest the involvement of developmentally regulated transcriptional factor(s) which have yet to be characterized.

MNF1, a leaf tissue-specific DNA-binding protein of maize, interacts with the cauliflower mosaic virus 35S promoter as well as the C4 photosynthetic phosphoenolpyruvate carboxylase gene promoter.

When gel shift assays were performed with maize nuclear extract and a DNA fragment containing the cauliflower mosaic virus (CaMV) 35S promoter, three DNA-protein complexes were observed. Analyses with nuclear extracts prepared from green leaves, etiolated leaves, stems and roots showed that the complexes resulted from the existence of at least two nuclear factors. One of them is presumably a constitutive nuclear factor found in all tissues tested, and another is a leaf-specific factor present both in green and etiolated leaves. This leaf-specific nuclear factor seemed to be identical to MNF1, previously identified as a factor interacting with the promoter of the maize gene for phosphoenolpyruvate carboxylase involved in the C4 photosynthesis. Deletion analysis revealed that MNF1 binds to the sequence from -281 to -235 relative to the transcription start site of the CaMV 35S promoter. MNF1-like nuclear protein was also found in tobacco nuclear extracts. The possibility that MNF1 participates as a positive trans-acting factor in the expression of genes in maize leaves is discussed.

Regulatory elements that control the lineage-specific expression of myoD.

The molecular basis of skeletal muscle lineage determination was investigated by analyzing DNA control elements that regulate the myogenic determination gene myoD. A distal enhancer was identified that positively regulates expression of the human myoD gene. The myoD enhancer and promoter were active in myogenic and several nonmyogenic cell lines. In transgenic mouse embryos, however, the myoD enhancer and promoter together directed expression of a lacZ transgene specifically to the skeletal muscle lineage. These data suggest that during development myoD is regulated by mechanisms that restrict accessibility of myoD control elements to positive trans-acting factors.

MET4, a leucine zipper protein, and centromere-binding factor 1 are both required for transcriptional activation of sulfur metabolism in Saccharomyces cerevisiae.

Inactivation of the centromere-binding factor 1 (CBF1) gene results in yeast strains that require methionine for growth. This auxotrophy is due to the inability of such strains to concentrate and assimilate sulfate from the medium. Northern (RNA) blot experiments reveal that the CBF1 protein is required for full induction of MET25 and MET16 gene transcription. However, we show that induction of the sulfate assimilation pathway is not achieved solely by CBF1. This induction also requires the integrity of a positive trans-acting factor, encoded by the MET4 gene. The MET4 gene was cloned, and its sequence reveals that it encodes a protein related to the family of the bZIP transcriptional activators. Evidence that MET4 is a transcriptional activator was provided by demonstrating that DNA-bound LexA-MET4 fusion proteins stimulate expression of a nearby promoter. The use of LexA-MET4 fusion proteins also reveals that the leucine zipper of MET4 is required for the recognition of the MET25 promoter. Moreover, an 18-bp fragment of the MET25 5' upstream region was found to confer S-adenosylmethionine-dependent regulation of a fusion gene. This regulation was shown to depend on both MET4 and CBF1. The obtained results suggest that the binding of CBF1 to its cognate sequences increases the ability of MET4 to stimulate transcription of the MET genes.

Relationship of eukaryotic DNA replication to committed gene expression: general theory for gene control

The historic arguments for the participation of eukaryotic DNA replication in the control of gene expression are reconsidered along with more recent evidence. An earlier view in which gene commitment was achieved with stable chromatin structures which required DNA replication to reset expression potential (D. D. Brown, Cell 37:359-365, 1984) is further considered. The participation of nonspecific stable repressor of gene activity (histones and other chromatin proteins), as previously proposed, is reexamined. The possible function of positive trans-acting factors is now further developed by considering evidence from DNA virus models. It is proposed that these positive factors act to control the initiation of replicon-specific DNA synthesis in the S phase (early or late replication timing). Stable chromatin assembles during replication into potentially active (early S) or inactive (late S) states with prevailing trans-acting factors (early) or repressing factors (late) and may asymmetrically commit daughter templates. This suggests logical schemes for programming differentiation based on replicons and trans-acting initiators. This proposal requires that DNA replication precede major changes in gene commitment. Prior evidence against a role for DNA replication during terminal differentiation is reexamined along with other results from terminal differentiation of lower eukaryotes. This leads to a proposal that DNA replication may yet underlie terminal gene commitment, but that for it to do so there must exist two distinct modes of replication control. In one mode (mitotic replication) replicon initiation is tightly linked to the cell cycle, whereas the other mode (terminal replication) initiation is not cell cycle restricted, is replicon specific, and can lead to a terminally differentiated state. Aberrant control of mitotic and terminal modes of DNA replication may underlie the transformed state. Implications of a replicon basis for chromatin structure-function and the evolution of metazoan organisms are considered.

Relationship of eukaryotic DNA replication to committed gene expression: general theory for gene control.

The historic arguments for the participation of eukaryotic DNA replication in the control of gene expression are reconsidered along with more recent evidence. An earlier view in which gene commitment was achieved with stable chromatin structures which required DNA replication to reset expression potential (D. D. Brown, Cell 37:359-365, 1984) is further considered. The participation of nonspecific stable repressor of gene activity (histones and other chromatin proteins), as previously proposed, is reexamined. The possible function of positive trans-acting factors is now further developed by considering evidence from DNA virus models. It is proposed that these positive factors act to control the initiation of replicon-specific DNA synthesis in the S phase (early or late replication timing). Stable chromatin assembles during replication into potentially active (early S) or inactive (late S) states with prevailing trans-acting factors (early) or repressing factors (late) and may asymmetrically commit daughter templates. This suggests logical schemes for programming differentiation based on replicons and trans-acting initiators. This proposal requires that DNA replication precede major changes in gene commitment. Prior evidence against a role for DNA replication during terminal differentiation is reexamined along with other results from terminal differentiation of lower eukaryotes. This leads to a proposal that DNA replication may yet underlie terminal gene commitment, but that for it to do so there must exist two distinct modes of replication control. In one mode (mitotic replication) replicon initiation is tightly linked to the cell cycle, whereas the other mode (terminal replication) initiation is not cell cycle restricted, is replicon specific, and can lead to a terminally differentiated state. Aberrant control of mitotic and terminal modes of DNA replication may underlie the transformed state. Implications of a replicon basis for chromatin structure-function and the evolution of metazoan organisms are considered.

Multiple cis- and trans-Acting Elements Involved in Regulation of the neu Gene

A 2.4-kb rat neu genomic DNA fragment that hybridized to the 5'-most coding sequence of the rat neu cDNA was cloned. S1 nuclease mapping identified multiple transcriptional initiation sites. DNA sequence analysis revealed that this fragment contained 64 bp of the first intron, 81 bp of the first exon, and the upstream noncoding sequence of the neu gene. The sequence immediately upstream of the translation start site was G + C rich (greater than 75%) and contained a consensus CCAAT sequence despite the absence of a TATA box. An Sp1-binding site was found, in addition to various sequence motifs common to the promoters of the human neu gene (erbB2), the epidermal growth factor receptor gene, and the simian virus 40 enhancer. A 2.2-kb EcoRI-Narl fragment containing sequences upstream from the 3'-most transcriptional start site was fused to the bacterial chloramphenicol acetyltransferase reporter gene and shown to promote transcription efficiently. A series of promoter deletion constructs was made, and results from transfection and subsequent chloramphenicol acetyltransferase assays suggested the presence of multiple cis-acting elements that contributed either positively or negatively to the transcription activity. Cotransfection competition experiments using subcloned cis-acting elements confirmed the existence of trans-acting factors interacting with these DNA fragments. In addition, a gel retardation assay was performed to demonstrate the physical binding of nuclear factors to certain fragments. The results complemented those of the deletion studies and led us to conclude that transcriptional regulation of the neu proto-oncogene involves at least one negative and three positive trans-acting factors interacting with different cis-acting elements along the neu gene promoter.

Multiple cis- and trans-acting elements involved in regulation of the neu gene.

A 2.4-kb rat neu genomic DNA fragment that hybridized to the 5'-most coding sequence of the rat neu cDNA was cloned. S1 nuclease mapping identified multiple transcriptional initiation sites. DNA sequence analysis revealed that this fragment contained 64 bp of the first intron, 81 bp of the first exon, and the upstream noncoding sequence of the neu gene. The sequence immediately upstream of the translation start site was G + C rich (greater than 75%) and contained a consensus CCAAT sequence despite the absence of a TATA box. An Sp1-binding site was found, in addition to various sequence motifs common to the promoters of the human neu gene (erbB2), the epidermal growth factor receptor gene, and the simian virus 40 enhancer. A 2.2-kb EcoRI-Narl fragment containing sequences upstream from the 3'-most transcriptional start site was fused to the bacterial chloramphenicol acetyltransferase reporter gene and shown to promote transcription efficiently. A series of promoter deletion constructs was made, and results from transfection and subsequent chloramphenicol acetyltransferase assays suggested the presence of multiple cis-acting elements that contributed either positively or negatively to the transcription activity. Cotransfection competition experiments using subcloned cis-acting elements confirmed the existence of trans-acting factors interacting with these DNA fragments. In addition, a gel retardation assay was performed to demonstrate the physical binding of nuclear factors to certain fragments. The results complemented those of the deletion studies and led us to conclude that transcriptional regulation of the neu proto-oncogene involves at least one negative and three positive trans-acting factors interacting with different cis-acting elements along the neu gene promoter.

Two genetically defined trans-acting loci coordinately regulate overlapping sets of liver-specific genes

Mice homozygous for deletions around the albino locus fail to activate expression of a set of neonatal liver functions and die shortly after birth. This phenotype is thought to result from the loss of a positive trans-acting factor, denoted alf, in deletion homozygotes. Using differential cDNA screening, we isolated and characterized genes whose cell type-specific transcription is affected by alf and found as a common feature that expression of these genes is induced by glucocorticoids and cAMP. Surprisingly, a subset of these alf-responsive genes is negatively controlled by the tissue-specific extinguisher locus Tse-1. Administration of glucocorticoids and cAMP leads to reversal of Tse-1—mediated extinction of these genes. These results show that two trans-acting factors coordinately regulate expression of overlapping sets of liver-specific genes. We suggest that both the lethal phenotype and the extinguished state result from interference with hormone signal transduction.

Negative trans-acting factors extinguish Ia expression in B cell-L 929 somatic cell hybrids

Numerous studies have implicated trans-acting factors in the regulation of MHC class II gene expression. Some of these factors have been shown to act by inducing the expression of class II genes while others have been demonstrated to downregulate such expression. These reports have dealt almost exclusively with the role of trans-acting factors in the regulation of class II gene expression in hematopoietic-derived cells. We decided to extend these studies to the role trans-acting factors play in nonhematopoietic-derived (NHD) cells. In order to address this question we made somatic cell hybrids between the NHD Ltk − cell line and normal B cells to determine if the existence of positive trans-acting factors from the B cell would lead to the expression of Ltk − class II genes in the resultant hybrid. Our results clearly indicate that not only was there no induction of Ltk − class II gene expression in the hybrids, but there was a loss of B cell class II gene expression as well. These results suggest that Ltk − cells possess negative trans-acting factors that appear to predominate over the positive trans-acting factors possessed by B cells. We have further extended these studies to test the MHC-inducing activity of IFN-γ and IL-4 on these hybrids. Our results indicate that the hybrids responded to IFN-γ with an increase in class I but not class II expression for both fusion partners. Furthermore, neither B cell nor L cell class II genes were induced by IL-4. Taken together, these results indicate that Ltk − cells possess negative trans-acting factors that not only maintain the Ia − phenotype of these cells, but also block the action of positive trans-acting factors from B cells.

Selective suppression of endogenous β-globin gene expression by transferred β-globin/TK chimeric gene in murine erythroleukemia cells

To examine the effect of gene transfer on expression of the endogenous β-globin gene in murine erythroleukemia (MEL) cells, a β-globin/TK chimeric gene was introduced into MEL cells. In some of the transformants, expression of the endogenous β-globin gene was only weakly induced with the addition of DMSO, while expression of the introduced β-globin/TK chimeric gene was well induced. Suppression of the endogenous β-globin gene is selective for β-globin gene, since expression of α-globin gene was only weakly affected in its induction in the transformants. Analysis of nine individual transformants indicated that suppression of the endogenous gene correlates more closely with the transcriptional activity than the copy number of the exogenous gene. Thus, the selective suppression can be explained by competition of positive trans-acting factor(s), present in limiting amounts, with high copy number of the exogenous gene the conformation of which is active.

A common liver‐specific factor binds to the rat albumin and α‐foetoprotein promoters in vitro and acts as a positive trans‐acting factor in vivo

Rat liver nuclear extracts were tested for the presence of factors which might be common to the transcriptional regulation of both the albumin and alpha-foetoprotein genes. Gel shift assay showed the formation of three complexes (I, II and III) with the albumin probe. Two of them (I and III) could be displaced by the alpha-foetoprotein promoter. Analysis of nuclear extracts from liver, kidney, spleen and brain and competition experiments using several oligonucleotides covering regions from the albumin and alpha-foetoprotein promoters showed that complex III results from the binding of the ubiquitous nuclear factor 1, while complex II involves a CCAAT-box-binding protein also detected in brain and spleen extracts. Complex I is formed upon binding of a liver-specific factor to a proximal element of the rat albumin promoter. This factor also binds to a similar sequence in the alpha-foetoprotein promoter and is closely related to the hepatocyte nuclear factor 1, as shown by competition experiments using an oligonucleotide covering its target sequence on the beta-fibrinogen promoter. Transfection competition experiments indicated that, in vivo, this factor acts as a positive trans-acting element in the expression of both the rat albumin and alpha-foetoprotein genes.

Identification of sequences responsible for acute-phase induction of human C-reactive protein.

Human C-Reactive protein (CRP) is inducible in liver cells during acute inflammation. Around 90 bp from the 5' flanking region of the human CRP gene contain, as shown here, information to induce the expression of a linked bacterial CAT gene specifically in human hepatoma (Hep3B) cells. The promoter is induced rapidly, faithfully and at high efficiency when transfected cells are exposed to conditioned medium from lipopolysaccharide stimulated peripheral monocytes. The sequences required for inducibility are located immediately upstream to the TATA element. A DNA segment from base -121 to -50 is capable of inducing transcription from the heterologous SV40 early promoter. Induction of CRP expression is probably exerted via the binding of at least one positive trans-acting factor.

CArG boxes in the human cardiac alpha-actin gene are core binding sites for positive trans-acting regulatory factors.

Positively acting, rate-limiting regulatory factors that influence tissue-specific expression of the human cardiac alpha-actin gene in a mouse muscle cell line are shown by in vivo competition and gel mobility-shift assays to bind to upstream regions of its promoter but to neither vector DNA nor a beta-globin promoter. Although the two binding regions are distinctly separated, each corresponds to a cis region required for muscle-specific transcriptional stimulation, and each contains a core CC(A + T-rich)6GG sequence (designated CArG box), which is found in the promoter regions of several muscle-associated genes. Each site has an apparently different binding affinity for trans-acting factors, which may explain the different transcriptional stimulation activities of the two cis regions. Therefore, we conclude that the two CArG box regions are responsible for muscle-specific transcriptional activity of the cardiac alpha-actin gene through a mechanism that involves their binding of a positive trans-acting factor in muscle cells.