AT-rich Element Research Articles

Effect of DNA-binding Drugs on Early Growth Response Factor-1 and TATA Box-binding Protein Complex Formation with the Herpes Simplex Virus Latency Promoter

Adjacent binding sites for early growth response factor-1 (EGR1) and TATA box-binding protein (TBP) were identified on the herpes simplex virus latency promoter in previous work. The binding of EGR1 to the GC-rich region prevented TBP binding to the AT-rich region. With the simultaneous addition of both EGR1 and TBP, the intercalator nogalamycin prevented EGR1 complex formation, resulting in a dose-dependent increase of the TBP.DNA complex. The minor groove binder chromomycin A3 inhibited EGR1 complex formation but resulted in a smaller increase of the TBP complex. In contrast, an alkylating intercalator hedamycin strongly inhibited binding of both proteins. The ability of these GC-binding drugs to prevent EGR1.DNA complex formation was in the following order: hedamycin > nogalamycin > chromomycin A3, and the specificity was nogalamycin > chromomycin A3 > hedamycin. With transcription factor IIA (TFIIA) in the assay, TBP was able to bind the promoter whereas formation of the EGR1.DNA complex was reduced. An AT minor groove-binding drug, distamycin A, disrupted the TBP.TFIIA.DNA complex and restored the EGR1.DNA complex. We conclude that the binding motif and sequence preference of DNA-interactive drugs are manifested in their ability to inhibit the transcription factor-DNA complexes.

Human Cart-1: Structural Organization, Chromosomal Localization, and Functional Analysis of a Cartilage-Specific Homeodomain cDNA

Homeoproteins control cell fates during development, specifying pattern formation and the ontogeny of specific tissues and organs in embryogenesis. Cart-1 cDNA was recently cloned from a rat chondrosarcoma tumor and it encodes a protein containing a paired-like homeodomain that is selectively expressed in cartilage during early chondrocyte differentiation. Here we report the molecular cloning of the human Cart-1 cDNA from a HeLa cervical carcinoma cDNA library. The human Cart-1 cDNA sequence is 88% identical and the deduced amino acid sequence is 95% identical to the rat sequence, indicating that Cart-1 structure is highly conserved. Northern and reverse transcriptase polymerase chain reaction (RT-PCR) analysis revealed Cart-1 mRNA expression in HeLa cervical carcinoma cells and human cervical tissue, but Cart-1 mRNA was not detected in GH3 rat pituitary cells and murine 10T1/2 one-half fibroblast cells. The Cart-1 gene was localized to human chromosome 12 and regionally mapped to the 12q21.3-q22 by PCR analysis of rodent-X-human somatic cell hybrid DNA and the CEPH megabase-insert YAC DNA pools, respectively. The Holt-Oram syndrome, characterized by upper limb and atrial septal dysplasias, also maps to the 12q21.3-q22 region. Cotransfection studies show that Cart-1 inhibits the rat prolactin promoter and that this repression is mediated by footprint II, an AT-rich element that functions as an inhibitory site of prolactin gene expression in nonpituitary cells and which was used to clone Cart-1. Taken together, these data indicate that Cart-1 may also influence cervix development, identify a putative DNA binding site for Cart-1, and, begin to define its functional role as modulator of gene expression.

Upstream regulatory elements controlling expression of the Entamoeba histolytica lectin

Entamoeba histolytica genomic organization and putative promoter elements appear to be distinct from both metazoan and better characterized protozoan organisms. The recent development of DNA-mediated transfection for E. histolytica enabled characterization of cis-acting promoter elements required for gene expression. A deletion and replacement analysis was conducted on the promoter of an E. histolytica gene encoding the heavy subunit of the N-acetyl-β-d-galactosamine-specific adhesin (hgl5). Deletion of the DNA from −1000 bases to −272 bases upstream from the start of transcription of hgl5 did not decrease reporter gene expression. Subsequent nested deletions and 10-bp replacement mutagenesis identified four positive upstream regulatory elements between bases −219 to −200, −189 to −160, −69 to −60, and −49 to −40. A negative upstream regulatory element between bases −89 to −80 was conserved upstream of three other E. histolytica genes. Mutation of the previously unidentified ‘GAAC’ element conserved within the putative core promoter decreased reporter gene expression by 75%. Site directed mutagenesis of the putative TATA element decreased reporter gene expression by greater than 50%, while mutation of the putative initiator element resulted in a more modest decrease. This analysis suggests that E. histolytica promoters are unlike other protozoan promoters, with AT-rich upstream regulatory elements, a nonconsensus TATA element, the ‘GAAC’ element, and an unusual initiator element.

Cloning and characterization of the gene encoding pyruvate phosphate dikinase from Giardia duodenalis

This paper reports the cloning and molecular characterization of the gene encoding pyruvate phosphate dikinase (PPDK) from Giardia. The ORF is 2652 nucleotide residues in length and not interrupted by introns. The gene appears to exist as a single copy in the genome and predicts a 97 629 Da protein containing 884 amino acid residues. Comparison of the deduced Giardia PPDK sequence with those of homologous enzymes from other organisms revealed high sequence similarities and the presence of various conserved domains known to be essential for substrate binding and catalysis. Analysis of the ppdk gene and 19 other protein-coding genes from the protist revealed no typical TATA boxes, positioned at around −30, but the presence of two novel consensus sequence motifs in the 5′ flanking regions. One is an AT-rich element immediately preceding the translation initiation codon and the other a 14-bp box centered at −30. These shared consensus sequence patterns present in the 5′ flanking region of Giardia genes are suggested to play a role in the control of transcription initiation.

Analysis of muscle creatine kinase gene regulatory elements in skeletal and cardiac muscles of transgenic mice.

Regulatory regions of the mouse muscle creatine kinase (MCK) gene, previously discovered by analysis in cultured muscle cells, were analyzed in transgenic mice. The 206-bp MCK enhancer at nt-1256 was required for high-level expression of MCK-chloramphenicol acetyltransferase fusion genes in skeletal and cardiac muscle; however, unlike its behavior in cell culture, inclusion of the 1-kb region of DNA between the enhancer and the basal promoter produced a 100-fold increase in skeletal muscle activity. Analysis of enhancer control elements also indicated major differences between their properties in transgenic muscles and in cultured muscle cells. Transgenes in which the enhancer right E box or CArG element were mutated exhibited expression levels that were indistinguishable from the wild-type transgene. Mutation of three conserved E boxes in the MCK 1,256-bp 5' region also had no effect on transgene expression in thigh skeletal muscle expression. All these mutations significantly reduced activity in cultured skeletal myocytes. However, the enhancer AT-rich element at nt - 1195 was critical for expression in transgenic skeletal muscle. Mutation of this site reduced skeletal muscle expression to the same level as transgenes lacking the 206-bp enhancer, although mutation of the AT-rich site did not affect cardiac muscle expression. These results demonstrate clear differences between the activity of MCK regulatory regions in cultured muscles cells and in whole adult transgenic muscle. This suggests that there are alternative mechanism of regulating the MCK gene in skeletal and cardiac muscle under different physiological states.

Identification of a Minimum Enhancer Sequence for the Type II Collagen Gene Reveals Several Core Sequence Motifs in Common with the Link Protein Gene

The type II collagen gene (Col2a1) is expressed primarily in chondrocytes. Transcription of Col2a1 is mediated by cell-specific regulatory elements located within the promoter and first intron. Here, we map a minimal enhancer and identify elements that determine cartilage-specific Col2a1 expression by analyzing the activity of a series of chimeric genes consisting of rat Col2a1 first intron deletion mutants ligated to the chloramphenicol acetyltransferase reporter gene. We show that a 100-base pair (bp) segment within the first intron is the minimum size necessary for high level, cell type-specific expression of Col2a1. Sequence analysis of this 100-bp Col2a1 enhancer revealed several sequence motifs similar to motifs present within the regulatory region of the link protein gene, another cartilage gene. These motifs include an AT-rich element, a C1 motif and a C3 motif. Deletion of any of these elements reduced Col2a1 enhancer activity in chick embryo chondrocytes. We also tested enhancer-mediated activity in CFK2 cells which differentiate to a chondrogenic phenotype and begin to express type II collagen mRNA after extended culture. In stably transfected CFK2 cells, constructs containing the 100-bp enhancer were activated during the transition from prechondrogenic to chondrogenic cell populations and deletions within the enhancer strongly down-regulated activity. Chondrocyte-specific DNA-protein complexes were identified using nuclear extracts prepared from chick embryo chondrocytes and 32P-labeled oligonucleotides from these regions of the first intron. These results suggest that interaction of chondrocyte specific nuclear factors with multiple core elements from a small region within the first intron are important for cell-type specific Col2a1 enhancer activity.

Structural distortions induced by integration host factor (IHF) at the H' site of phage lambda probed by (+)-CC-1065, pluramycin, and KMnO4 and by DNA cyclization studies.

Integration Host Factor (IHF) is a sequence-specific DNA-bending protein that is proposed to interact with DNA primarily through the minor groove. We have used various chemical probes [(+)-CC-1065, a minor-groove-specific agent that alkylates N3 of adenine and traps bends into the minor groove; pluramycin, a minor-major-groove threading intercalator that alkylates N7 of guanine; KMnO4, which reacts more strongly with bases in denatured DNA] to gain more information on the interaction of IHF with the H' site of phage lambda. In addition to the 13-bp core consensus recognition element present at all IHF binding sites, the H' site also has an upstream AT-rich element that increases the affinity of IHF for this site. Our results reveal new details of IHF-DNA interaction at this site. Results with (+)-CC-1065 modification suggest that IHF interacts with the adenines on the 3'-side of the AT-rich element and likely induces a minor-groove bend in its vicinity, which in turn stabilizes the interaction. Pluramycin modification experiments suggest the presence of both short- and long-range structural perturbations (possible DNA unwinding events) on either side of the IHF contact region. Although IHF is known to induce a large bend in DNA at the H' site, no separation of base pairs was detected when the bent DNA was probed with KMnO4. DNA cyclization studies indicate a large magnitude (approximately 180 degrees) for the IHF-induced bend at the H' site, consistent with > 140 degrees bend estimated by gel electrophoresis methods. These studies suggest that IHF-induced DNA bending is accompanied by the introduction of a DNA node, DNA unwinding, and/or by some other DNA distortion. An enhanced binding and stability of IHF was observed on small circular DNA.

Organization of the Saccharomyces cerevisiae actin gene UAS: functional significance of reiterated REB1 binding sites and AT-rich elements.

The upstream activation sequence (UAS) in the Saccharomyces cerevisiae actin gene promoter contains three different motifs, specifically two AT-rich tracts, two binding sites for the yeast protein REB1, and an Mlul site. Synthetic UAS elements containing individual motifs, or combinations of them, were inserted in place of the natural UAS, and assayed using a lacZ reporter gene. The REB1 binding sites were found to be essential for, and sufficient to restore partial, UAS activity. AT-rich tracts alone were inactive. Multimerization of a REB1 binding site created a UAS that in galactose is more active, but in glucose less active, than a UAS having a single REB1 site with one AT-rich tract. In general, transcription during growth in galactose or glycerol/lactate responds more to multimerization of motifs. The results suggest that the natural actin promoter UAS retains activity on these alternative carbon sources because of reiteration of sequence elements within it; the additional elements appear to be redundant when cells are grown on glucose. The Mlul site, which is present upstream of a number of yeast genes involved in DNA synthesis and confers cell cycle periodicity to those genes, contributes to the activity of the synthetic UAS elements, but not in a cell-cycle-dependent manner.

Functional interaction between the POU domain protein Tst-1/Oct-6 and the high-mobility-group protein HMG-I/Y.

The POU domain protein Tst-1/Oct-6 is a transcriptional activator of human papovavirus JC virus in transient transfections. Because of its endogenous expression in myelinating glia, Tst-1/Oct-6 might also be an important determinant for the glia specificity of JC virus in vivo. Activation of viral early and late genes depends on the ability of Tst-1/Oct-6 to interact with an AT-rich element within the viral regulatory region. Here, we show that this element not only is bound by Tst-1/Oct-6 but, in addition, serves as a binding site for the high-mobility-group protein HMG-I/Y. In the presence of HMG-I/Y, Tst-1/Oct-6 exhibited an increased affinity for this AT-rich element. The specificity of this effect was evident from the fact that no stimulation of Tst-1/Oct-6 binding was observed on a site that did not allow binding of HMG-I/Y. In addition, both proteins interacted with each other in solution. Direct contacts were identified between the POU domain of Tst-1/Oct-6 and a short stretch of 10 amino acids in the central portion of HMG-I/Y. These results point to an accessory role for HMG-I/Y in the activation of JC viral gene expression by the POU domain protein Tst-1/Oct-6. In agreement with such a role, HMG-Y synergistically supported the function of Tst-1/Oct-6 in transient transfections, measured on the early promoter of JC virus or on an artificial promoter consisting of only a TATA box and the common binding element for Tst-1 and HMG-I/Y.

Mutually Exclusive Interactions between Factors Binding to Adjacent Sp1 and AT-rich Elements Regulate Gastrin Gene Transcription in Insulinoma Cells

The gastrin gene is transiently expressed in fetal pancreatic islets during islet neogenesis but then switched off after birth when islet cells become fully differentiated. Previous studies identified a cis-regulatory sequence between -109 and -75 in the human gastrin promoter which binds islet cell-specific activators and a nonspecific repressor and thus may act as a molecular switch. The present study identified another cis-regulatory sequence (-163ACACTAAATGAAAGGGCGGGGCAG-140) which bound two islet nuclear proteins in a mutually exclusive manner, as defined by gel shift competition, methylation interference, and DNase I foot-printing assays. The general transactivator Sp1 recognized the downstream GGGCGGGG sequence, but Sp1 binding was prevented when another islet factor bound to the adjacent AT-rich sequence (CTAAATGA). This gastrin AT-rich element is nearly identical to the binding site (ATAAATGA) for the islet-specific transcription factor beta TF-1. However, the gastrin AT-binding factor appeared to differ from beta TF-1 in its gel mobility shift pattern. Transfections of rat insulinoma cells revealed that mutations which blocked binding to the AT-rich element but allowed Sp1 binding up-regulated transcriptional activity. These results suggest that the gastrin AT-binding factor blocks transactivation by Sp1 and may have a role in the repression of gastrin transcription seen at the end of islet differentiation.

Identification and Characterization of a Novel Cytokine-inducible Nuclear Protein from Human Endothelial Cells

Vascular endothelial cells undergo profound changes upon cellular activation including expression of a spectrum of cell activation-associated genes. These changes play important roles in many physiological and pathological events. By differential screening of a cDNA library prepared from interleukin-1 alpha and tumor necrosis factor-alpha-stimulated human dermal microvascular endothelial cells, we have identified a novel cytokine-inducible gene, designated as C-193. The compiled cDNA sequence of C-193 is 1901 base pairs long and shows no significant homology with any known gene sequence. Genomic DNA analysis revealed that C-193 is encoded by a single gene, which is conserved in different mammalian species. The C-193 gene was localized to human chromosome 10 by Southern blot analysis of somatic cell hybrids. Multiple AT-rich mRNA decay elements were identified in the 3'-untranslated region. C-193 mRNA expression was rapidly and transiently induced by treatment with interleukin-1 alpha or tumor necrosis factor-alpha, reached a peak of expression about 16 h post tumor necrosis factor-alpha stimulation, and the induction of C-193 was protein synthesis independent. Lipopolysaccharide and cycloheximide were also potent inducers of C-193 mRNA. Therefore, C-193 represents a new addition to the primary response gene family. In vitro translation of C-193 yielded a 36-kDa protein product, consistent with the predicted open reading frame of 318 amino acids and a calculated molecular mass of 36 kDa for C-193 protein. The predicted protein sequence contains a basic amino acid cluster similar to a nuclear localization signal, four tandem repeats of ankyrin-like sequence, and multiple consensus protein phosphorylation sites. C-193 was engineered with a FLAG tag at its carboxyl terminus and transiently expressed in COS cells. Consistent with the presence of a putative nuclear localization signal, the C-193-FLAG protein was localized to the nucleus of transfected COS cells by indirect immunofluorescence microscopy. C-193-FLAG prepared in vitro was capable of binding DNA cellulose. These results indicate that C-193 protein may play an important role in endothelial cell activation.

Cis-Acting components of human papillomavirus (HPV) DNA replication: linker substitution analysis of the HPV type 11 origin

Papillomavirus DNA replication requires the viral trans-acting factors E1 and E2 in addition to the host cell's general replication machinery. The origins of DNA replication in bovine and human papillomavirus genomes have been localized to a specific part of the upstream regulatory region (URR) which includes recognition sites for E1 and E2 proteins. To fine map cis-acting elements influencing human papillomavirus type 11 (HPV-11) DNA replication and to determine the relative contributions of such sites, we engineered consecutive linker substitution mutations across a region of 158 bp in the HPV-11 origin and tested mutant origins for replication function in a cell-based transient replication assay. Our results both confirm and extend the findings of others. E2 binding sites are the major cis components of HPV-11 DNA replication, and there is evidence for synergy between these sites. Differential capacity of the three E2 binding sites within the origin to affect replication may be attributed, at least in part, to context. At least one E2 binding site is essential for replication. The imperfect AT-rich palindrome of the E1 helicase binding site is not essential since replication occurs even in the absence of this sequence. However, replication is enhanced by the presence of the palindromic sequence in the HPV-11 origin. Sequence components adjacent to the E1 and E2 binding sites, comprising AT-rich and purine-rich elements and the consensus TATA box sequence, probably contribute to the overall efficiency of replication, though they are nonessential. None of the other cis elements of the HPV-11 origin region analyzed seems to influence replication significantly in the system described. The HPV-11 origin of DNA replication therefore differs from those of the other papovaviruses, simian virus 40 and polyomavirus, inasmuch as an intact helicase binding site and adjacent AT-rich components, while influential, are not absolutely essential.

Characterization of a glucocorticoid responsive element and identification of an AT-rich element that regulate the link protein gene.

The cartilage matrix is composed of characteristic components including type II collagen, aggrecan and link protein. In this paper, we report two DNA elements that regulate the link protein gene. Using transient transfection assays with link protein gene constructs in chondrocytes, chloramphenicol acetyl transferase (CAT) assays were used to measure the transcriptional activity of the link protein gene. Previously, we identified an enhancer-like activity within the first intron of the gene. In this paper, we report an active 34 bp (+1390 to +1424) fragment within this region that contains a glucocorticoid-like response element (GRE). Both deletion of, and site-specific mutations within this sequence motif reduced the dexamethasone-inducible activity. The GRE-like sequence from the rat link protein gene, or the homologous sequence from the human link protein gene were included in vectors containing the thymidine kinase promoter linked to the CAT gene (tkCAT). Both human and rat elements transferred the ability to respond to dexamethasone and hydrocortisone with a > 10-fold induction. Deletions through the promoter from -923 to -900 identified a second site required for both glucocorticoid and serum responsiveness. A four base substitution at this site resulted in a loss of serum responsiveness. This region contains an AT-rich element, similar to the AT-rich elements involved in homeotic protein regulation of the growth hormone gene and the muscle creatine kinase gene. Southwestern analysis using oligonucleotides containing the AT-rich element from the link protein gene or the muscle creatine kinase gene, identified a 32 kDa protein band from nuclear extracts of chick chondrocytes. Using these AT-rich oligonucleotides in band-shift analyses, nuclear extracts of chick sternal muscle, rat chondrosarcoma and chick sternal chondrocytes each showed formation of different complexes suggesting cell specificity. AT-rich elements have been identified as binding sites for homeodomain-containing proteins and can contribute to gene regulation by serum response factors. The identification of an AT-rich element in the link protein gene suggests similar functions for this element.

Kluyveromyces marxianus small DNA fragments contain both autonomous replicative and centromeric elements that also function in Kluyveromyces lactis

Two fragments containing both an autonomous replicating sequence (ARS) and a centromere have been isolated and sequenced from the yeast Kluyveromyces marxianus. The ARS and centromeric core sequences are only 500 bp apart, but ARS activity could be separated from the centromeric sequences. Centromeric sequences are organized in a similar way to those of budding yeasts: two well-conserved elements: CDEI (5' TCACGTG 3') and CDEIII (5' TNTTCCGAAAGTWAAA 3'), are separated by a 165 bp AT-rich (+/- 90%) CDEII element whose length is twice that of Saccharomyces cerevisiae CDEII but almost identical to that of K. lactis. The ARS-core consensus sequence (5' TTTATTGTT 3') is also similar to that of K. lactis. Both ARS and centromeric elements function in this strain, albeit inefficiently, but not in S. cerevisiae. A third ARS-containing fragment with a different organization has been isolated and sequenced.

Nuclear protein 780BP from cauliflower binds an element in the 780 gene promoter of T-DNA

A 16 bp site of protein binding has been identified in the promoter of the 780 gene of T-DNA. Specific DNA-protein interactions were demonstrated between a double-stranded oligonucleotide containing this element (5'-TTGAAAAATCAACGCT-3') and a protein isolated from nuclear extracts of cauliflower inflorescences. Specific bases required for this binding activity (780 binding protein; 780BP) were defined by kinetic competition studies with mutated oligonucleotides, methylation interference assays and DNAse I footprinting. 780BP binding was not competed with up to 1000-fold excess of previously characterized plant regulatory elements such as as-1, the LRE, and the ocs, G-box, and AT-rich elements. In addition, 780BP was shown to bind sequences overlapping a mammalian hormone receptor element with greater affinity than the 780 element.

Unwound Regions in Yeast Centromere IV DNA

The DNA of the centromere of chromosome IV (CEN4) of Saccharomyces cerevisiae is found to be sensitive to single-strand specific nuclease P 1 when inserted into a negatively supercoiled plasmid. Fine mapping identifies two P 1-sensitive segments: one segment maps to essential centromere element CDEI and bordering CDEII bases, and the other segment is located in element CDEIII. The AT-rich element CDEII, which is expected to be early melting, is for the most part resistant to nuclease P 1. Cleavage is inhibited by NaCl, MgCl 2 and polyamines. The cleavage rate is only weakly dependent on P 1 concentration in the range of 0·5 to 20 munits/μl. The two P 1-sensitive segments are also modified by the DNA-confirmation-specific reagent KMnO 4. Negative superhelicity is required for all modifications. Two-dimensional topoisomer analysis indicates the unwinding of 80(±10) bases within the negatively supercoiled CEN4-containing plasmid. The data best fit a model in which the DNA of the CEN4 region undergoes a transition into a paranemic intermediate in which each strand is folded into an RNA-like foldback structure.

A GBF-binding site and a novel AT element define the minimal sequences sufficient to direct prespore-specific expression in Dictyostelium discoideum.

In order to better understand the molecular mechanisms of cellular differentiation in Dictyostelium discoideum, we have identified the minimum regulatory sequences of the prespore-specific gene SP60/cotC that are sufficient to confer cell-type-specific expression on a heterologous promoter. This region includes at least two essential cis-acting elements: a novel AT-rich element (or elements) and CAE3. The essential function of the AT element is confirmed through point mutations that decrease expression below the level of detection. CAE3 is one of three CA-rich elements (CAEs) required for the induction of SP60/cotC during development or in response to extracellular cyclic AMP. The CAEs have differential affinities for a specific developmentally induced nuclear activity (CAE1 > CAE2 >> CAE3). Here, we identify this activity as G-box-binding factor (GBF) and show that in vitro-transcribed and -translated GBF binds all three SP60/cotC CAEs in a sequence-specific manner. Previous studies have suggested that GBF mediates the induction of some prestalk genes, and these results demonstrate that it also has a specific role in prespore gene activation.

A GBF-Binding Site and a Novel AT Element Define the Minimal Sequences Sufficient To Direct Prespore-Specific Expression in Dictyostelium discoideum

In order to better understand the molecular mechanisms of cellular differentiation in Dictyostelium discoideum, we have identified the minimum regulatory sequences of the prespore-specific gene SP60/cotC that are sufficient to confer cell-type-specific expression on a heterologous promoter. This region includes at least two essential cis-acting elements: a novel AT-rich element (or elements) and CAE3. The essential function of the AT element is confirmed through point mutations that decrease expression below the level of detection. CAE3 is one of three CA-rich elements (CAEs) required for the induction of SP60/cotC during development or in response to extracellular cyclic AMP. The CAEs have differential affinities for a specific developmentally induced nuclear activity (CAE1 > CAE2 >> CAE3). Here, we identify this activity as G-box-binding factor (GBF) and show that in vitro-transcribed and -translated GBF binds all three SP60/cotC CAEs in a sequence-specific manner. Previous studies have suggested that GBF mediates the induction of some prestalk genes, and these results demonstrate that it also has a specific role in prespore gene activation.

Conserved regulatory elements in the type I DNA topoisomerase gene promoters of mouse and man

The gene for mammalian type I DNA topoisomerase is constitutively expressed, but also regulated by a number of external stimuli. We compared the nucleotide sequences of the human and the mouse topoisomerase I gene promoters because promoter elements, essential for basic as well as regulated gene expression, should be conserved during evolution. We found that proximal upstream sequences are highly conserved and include potential binding sites for ubiquitous transcription factors, a regulatory CRE site as well as two novel promoter elements that have been shown to be important for the expression of the human gene. The more distal parts of the upstream sequences are less well conserved but include two regions that are almost identical in the human and the mouse gene. One of these regions contains a binding site for a basic-helix-loop-helix/leucine-zipper protein, and the other contains an AT-rich element with the potential for DNA bending.

ATBF1, a multiple-homeodomain zinc finger protein, selectively down-regulates AT-rich elements of the human alpha-fetoprotein gene.

ATBF1 is a 306-kDa protein containing four homeodomains, 17 zinc finger motifs, and several segments potentially involved in transcriptional regulation (T. Morinaga, H. Yasuda, T. Hashimoto, K. Higashio, and T. Tamaoki, Mol. Cell. Biol. 11:6041-6049, 1991). At least one of the homeodomains of ATBF1 binds to an AT-rich element in the human alpha-fetoprotein (AFP) enhancer (enhancer AT motif). In the present work, we analyzed the transcriptional regulatory activity of ATBF1 with respect to the enhancer AT motif and similar AT-rich elements in the human AFP promoter and the human albumin promoter and enhancer. Gel retardation assays showed that ATBF1 binds to the AFP enhancer AT motif efficiently; however, it binds weakly or not at all to other AT-rich elements in the AFP and albumin regulatory regions studied. Alterations of the enhancer AT motif by site-specific mutagenesis resulted in the loss of binding of ATBF1. Cotransfection experiments with an ATBF1 expression plasmid and the chloramphenicol acetyltransferase (CAT) gene fused to AFP promoter or enhancer fragments showed that ATBF1 suppressed the activity of AFP enhancer and promoter regions containing AT-rich elements. This suppression was reduced when the mutated AT motifs with low affinity to ATBF1 were linked to the CAT gene. The ATBF1 suppression of AFP promoter and enhancer activities appeared to be due, at least in part, to competition between ATBF1 and HNF1 for the same binding site. In contrast to the AFP promoter and enhancer, the albumin promoter and enhancer were not affected by ATBF1, although they contain homologous AT-rich elements. These results show that ATBF1 is able to distinguish AFP and albumin AT-rich elements, leading to selective suppression of the AFP promoter and enhancer activities.