In vivo footprinting of the estrogen-inducible vitellogenin II gene from chicken

Ectopic Methylation of a Single Persistently Unmethylated CpG in the Promoter of the Vitellogenin Gene Abolishes Its Inducibility by Estrogen through Attenuation of Upstream Stimulating Factor Binding

The enhancer/promoter of the vitellogenin II gene (VTG) has been extensively studied as a model system of vertebrate transcriptional control. While deletion mutagenesis and in vivo footprinting identified the transcription factor (TF) binding sites governing its tissue specificity, DNase hypersensitivity and DNA methylation studies revealed the epigenetic changes accompanying its hormone-dependent activation. Moreover, upon induction with estrogen (E₂), the region flanking the estrogen-responsive element (ERE) was reported to undergo active DNA demethylation. We now show that although the VTG ERE is methylated in embryonic chicken liver and in LMH/2A hepatocytes, its induction by E₂ was not accompanied by extensive demethylation. In contrast, E₂ failed to activate a VTG enhancer/promoter-controlled luciferase reporter gene methylated by SssI. Surprisingly, this inducibility difference could be traced not to the ERE but rather to a single CpG in an E-box (CACGTG) sequence upstream of the VTG TATA box, which is unmethylated in vivo but methylated by SssI. We demonstrate that this E-box binds the upstream stimulating factor USF1/2. Selective methylation of the CpG within this binding site with an E-box-specific DNA methyltransferase, Eco72IM, was sufficient to attenuate USF1/2 binding in vitro and abolish the hormone-induced transcription of the VTG gene in the reporter system.

Phosphorylation of human estrogen receptor alpha by protein kinase A regulates dimerization

Phosphorylation provides an important mechanism by which transcription factor activity is regulated. Estrogen receptor alpha (ERalpha) is phosphorylated on multiple sites, and stimulation of a number of growth factor receptors and/or protein kinases leads to ligand-independent and/or synergistic increase in transcriptional activation by ERalpha in the presence of estrogen. Here we show that ERalpha is phosphorylated by protein kinase A (PKA) on serine-236 within the DNA binding domain. Mutation of serine-236 to glutamic acid prevents DNA binding by inhibiting dimerization by ERalpha, whereas mutation to alanine has little effect on DNA binding or dimerization. Furthermore, PKA overexpression or activation of endogenous PKA inhibits dimerization in the absence of ligand. This inhibition is overcome by the addition of 17beta-estradiol or the partial agonist 4-hydroxy tamoxifen. Interestingly, treatment with the complete antagonist ICI 182,780 does not overcome the inhibitory effect of PKA activation. Our results indicate that in the absence of ligand ERalpha forms dimers through interaction between DNA binding domains and that dimerization mediated by the ligand binding domain only occurs upon ligand binding but that the complete antagonist ICI 182,780 prevents dimerization through the ligand-binding domain. Heterodimer formation between ERalpha and ERbeta is similarly affected by PKA phosphorylation of serine 236 of ERalpha. However, 4-hydroxytamoxifen is unable to overcome inhibition of dimerization by PKA. Thus, phosphorylation of ERalpha in the DNA binding domain provides a mechanism by which dimerization and thereby DNA binding by the estrogen receptor is regulated.

Activation of vitellogenin II gene expression by steroid hormones in the old Japanese quail

Alterations in the basal transcription rates of eukaryotic genes are believed to involve the binding of trans-acting factor(s) with specific DNA sequences in the promoter. We show here two interrelated events for the VTGII gene of the old, non-egg laying Japanese quail: alterations in the structure of the chromatin encompassing the gene, and binding of trans-acting factors to the promoter of the gene. Estradiol/progesterone alone or together cause alterations in the conformation of the chromatin of the promoter region of the gene. This may allow free access of nuclear protein(s) to the cis-acting elements, ERE, PRE and NF1, in the promoter of the gene and cause activation of transcription.

The nucleosome: a powerful regulator of transcription

Nucleosomes provide the architectural framework for transcription. Histones, DNA elements, and transcription factors are organized into precise regulatory complexes. Positioned nucleosomes can facilitate or impede the transcription process. These structures are dynamic, reflecting the capacity of chromatin to adopt different functional states. Histones are mobile with respect to DNA sequence. Individual histone domains are targeted for posttranslational modifications. Histone acetylation promotes transcription factor access to nucleosomal DNA and relieves inhibitory effects on transcriptional initiation and elongation. The nucleosomal infrastructure emerges as powerful contributor to the regulation of gene activity.

Estrogen-dependent transcriptional activation and vitellogenin gene memory

The concept of hepatic memory suggests that a gene responds more rapidly to a second exposure of an inducer than it does during the initial activation. To determine how soon estrogen-dependent DNA/protein interactions occur during the primary response, in vivo dimethylsulfate footprinting was carried out using genomic DNA amplified by ligation-mediated PCR. When estrogen was added to disrupted cells from a hormone-naive liver, changes within and around the estrogen response elements occurred within seconds, indicating a direct and rapid effect on this estrogen-responsive promoter that had never before been activated. Because this effect was so rapid relative to the delayed onset of mRNA accumulation during the primary response, run-on transcription assays were used to determine the transcription profiles for four of the yolk protein genes during the primary and secondary responses to estrogen. As with the accumulation of mRNA, the onset of transcription was delayed for all of these genes after a primary exposure to estrogen. Interestingly, after the secondary exposure to estrogen, the vitellogenin I, vitellogenin II, and very low density apolipoprotein II genes displayed a more rapid onset of transcription, whereas the primary and secondary profiles of apolipoprotein B transcription in response to estrogen were identical. Because the apoB gene is constitutively expressed in the absence of estrogen, and the vitellogenins are quiescent before the administration of the hormone, hepatic memory most likely represents a relatively stable event in the transition to an active state of a gene that is committed for tissue-specific expression.

Chromatin remodeling regulated by steroid and nuclear receptors

Coactivators and corepressors regulate transcription by controlling interactions between sequence-specific transcription factors, the basal transcriptional machinery and the chromatin environment. This review consider the access of nuclear and steroid receptors to chromatin, their use of corepressors and coactivators to modify chromatin structure and the implications for transcriptional control. The assembly of specific nucleoprotein architectures and targeted histone modification emerge as central controlling elements for gene expression.

Direct study of DNA-protein interactions in repressed and active chromatin in living cells

Current methods for analysis of chromatin architecture are invasive, utilizing chemicals or nucleases that damage DNA, making detection of labile constituents and conclusions about true in vivo structure problematic. We describe a sensitive assay of chromatin structure which is performed in intact, living yeast. The approach utilizes expression of SssI DNA methyltransferase (MTase) in Saccharomyces cerevisiae to provide an order-of-magnitude increase in resolution over previously introduced MTases. Combining this resolution increase with the novel application of a PCR-based, positive chemical display of modified cytosines provides a significant advance in the direct study of DNA-protein interactions in growing cells that enables quantitative footprinting. The validity and efficacy of the strategy are demonstrated in mini-chromosomes, where positioned nucleosomes and a labile, operator-bound repressor are detected. Also, using a heterologous system to study gene activation, we show that in vivo hormone occupancy of the estrogen receptor is required for maximal site-specific DNA binding, whereas, at very high receptor-expression levels, hormone-independent partial occupancy of an estrogen-responsive element was observed. Receptor binding to a palindromic estrogen-responsive element leads to a footprint with strand-specific asymmetry, which is explicable by known structural information.

Regulation of the chicken ovalbumin gene by estrogen and corticosterone requires a novel DNA element that binds a labile protein, Chirp-1

Because induction of the chicken ovalbumin (Ov) gene by steroid hormones requires concomitant protein synthesis, efforts have focused on defining the binding site in the Ov gene for a labile transcription factor. Previous gel mobility shift studies identified one such site in the steroid-dependent regulatory element (SDRE) between -900 and -853. To ascertain whether estrogen and glucocorticoid affect the binding of this labile protein, genomic footprinting of the Ov gene was done by treating primary oviduct cell cultures with dimethyl sulfate. Several alterations that include steroid-dependent protection of guanine residues -889 and -885 and hypersensitivity of adenine residues -892 and -865 were observed. Of particular importance, the in vivo footprinting data are corroborated by two functional studies, one with linker-scanning mutations and the other with point mutations. Ten-base-pair linker-scanning mutations between -900 and -878 severely reduced the induction by estrogen and glucocorticoid. Likewise, point mutations of the protected guanine residues profoundly attenuated the response to these steroid hormones. In addition, in vitro binding activity correlated with in vivo functional activity. For example, mutant A4e shows no transcriptional activity in response to steroid hormones, and a corresponding oligomer does not bind protein in vitro. In contrast, mutant A4c is fully active in both contexts. These data support the contention that the ovalbumin gene is regulated by a steroid hormone-induced transcriptional cascade that culminates in the binding of chicken ovalbumin induced regulatory protein or protein complex (Chirp-I) to a DNA element from -891 to -878 in the SDRE.

Activation of the silent progesterone receptor gene by ectopic expression of estrogen receptors in a rat fibroblast cell line

We describe the construction and characterization of a novel estrogen (E2)-responsive cell line, Rat1+ER, which ectopically expresses estrogen receptor (ER). Human ER cDNA was introduced by retrovirus-mediated gene transfer into the Rat1 cell line, which does not express functional ER endogenously. Rat1+ER cells express functional ER based on radioreceptor assays, immunoblotting, and transient transfection experiments using E2-responsive reporter plasmids. The effects of this ectopic ER expression were studied on three endogenous E2-responsive genes, prolactin (PRL), progesterone receptor (PR), and epidermal growth factor receptor (EGFR). PRL, usually expressed in the lactotrophs of the pituitary, is not expressed at all in Rat1+ER cells, with or without E2 addition, and appears to require other factors for expression. In contrast, although PR is not expressed in Rat1 cells, it is induced in Rat1+ER cells upon the addition of E2. This induction appears to occur at the transcriptional level and is insensitive to cycloheximide treatment. This is one of the few examples where the expression of one gene activates an otherwise silent gene. Another contrasting observation is that, although EGFR is basally expressed in Rat1+ER cells, the addition of E2 has no effect. Our studies paint a complicated picture of E2 regulation of endogenous genes: the activation of the PR gene may only require the presence of E2 and ER, whereas EGFR and PRL genes require factors in addition to ER for basal as well as E2-regulated expression.

Characterization of liver-enriched proteins binding to a developmentally demethylated site flanking the avian apoVLDLII gene

Although the avian apoVLDLII gene is normally expressed exclusively in the liver of the laying hen, the gene can be activated by estrogen in birds of either sex beginning between days 7-9 of embryogenesis. Developmentally programmed demethylation of sites in the 5'- and 3'-flanking regions of the gene have been shown to occur during this period of embryogenesis, suggesting that they may reflect changes in protein-DNA interactions that are involved in the acquisition of competence to activate the apoVLDLII gene. We have detected specific protein interactions at one location approximately 2.6 kb upstream from the apoVLDLII gene, that includes an Msp I site whose methylation status changes between days 7 and 9 of embryogenesis. The sequence of this region bears significant similarity to binding sites of members of the bZIP family of liver-enriched or -specific factors such as C/EBP, DBP, and LAP, that are characteristically produced relatively late during liver development. In the studies described here, we demonstrate that proteins binding to the upstream apoVLDLII site do not correspond to previously identified liver-enriched or -specific factors. They also display a pattern of activity during development and in human and avian hepatoma cell lines indicating that their expression is increased in proliferating cells. Southwestern blotting and UV cross-linking studies indicate that two proteins of approximately 60 kD are capable of binding to the site and we describe the purification of these factors from crude nuclear protein extracts obtained from rooster liver.

Characterization and developmental expression of Xenopus C/EBP gene

The Xenopus homolog of the transcription factor C/EBP (CCAAT/enhancer core binding protein), cloned from an adult Xenopus liver cDNA library, encodes a protein whose sequence is 67% homologous to that of rat C/EBP at the amino acid level, with virtually identical sequence of the basic-zipper region at the carboxyl terminus. As determined by gel electrophoretic mobility shift assays, the protein synthesized from xC/EBP cDNA bound specifically to the consensus binding site for C/EBP-like proteins. Northern blotting and RNase protection revealed a single species of xC/EBP mRNA of 2.7 kb which was most abundant in adult Xenopus liver, with smaller amounts in spleen, kidney, oviduct and brain and undetectable in heart and skeletal muscle. Although a small amount of this transcript could be detected in unfertilized eggs and early embryos, its accumulation rose sharply at the onset of metamorphosis (stage 55/56), and continued to increase through metamorphic climax to reach its highest level in stage 66 froglet liver, but thereafter declining in adult liver. In situ hybridization revealed a uniform pattern of distribution of xC/EBP mRNA in the liver and fat body throughout metamorphosis. Towards the end of metamorphosis, high levels of xC/EBP mRNA were detected in epithelial cells of the digestive tract. However, the spatial pattern of cells expressing the transcript changed markedly in the developing kidney. Our results suggest that xC/EBP may be involved as a transcription factor in the establishment of the adult phenotype during post-embryonic development of Xenopus.

Mutational studies reveal a complex set of positive and negative control elements within the chicken vitellogenin II promoter

The endogenous chicken vitellogenin II (VTGII) gene is transcribed exclusively in hepatocytes in response to estrogen. We previously identified two estrogen response elements (EREs) upstream of this gene. We now present an analysis of the VTGII promoter activated by these EREs in response to estrogen. Chimeric VTGII-CAT genes were cotransfected into LMH chicken hepatoma cells along with an estrogen receptor expression vector, and transient CAT expression was assayed after culturing the cells in the absence or presence of estrogen. An analysis of constructs bearing deletions downstream of the more proximal ERE indicated that promoter elements relevant to transcription in LMH cells extend to between -113 and -96. The relative importance of sequences within the VTGII promoter was examined by using 10 contiguous linker scanner mutations spanning the region from -117 to -24. Although most of these mutations compromised VTGII promoter function, one dramatically increased expression in LMH cells and also rendered the VTGII promoter capable of being activated by cis-linked EREs in fibroblasts cotransfected with an estrogen receptor expression vector. Gel retardation and DNase I footprinting assays revealed four factor-binding sites within this promoter. We demonstrate that three of these sites bind C/EBP, SP1, and USF (or related factors), respectively; the fourth site binds a factor that we denote TF-V beta. The biological relevance of these findings is suggested by the fact that three of these binding sites map to sites previously shown to be occupied in vivo in response to estrogen.

Mutational studies reveal a complex set of positive and negative control elements within the chicken vitellogenin II promoter

The endogenous chicken vitellogenin II (VTGII) gene is transcribed exclusively in hepatocytes in response to estrogen. We previously identified two estrogen response elements (EREs) upstream of this gene. We now present an analysis of the VTGII promoter activated by these EREs in response to estrogen. Chimeric VTGII-CAT genes were cotransfected into LMH chicken hepatoma cells along with an estrogen receptor expression vector, and transient CAT expression was assayed after culturing the cells in the absence or presence of estrogen. An analysis of constructs bearing deletions downstream of the more proximal ERE indicated that promoter elements relevant to transcription in LMH cells extend to between -113 and -96. The relative importance of sequences within the VTGII promoter was examined by using 10 contiguous linker scanner mutations spanning the region from -117 to -24. Although most of these mutations compromised VTGII promoter function, one dramatically increased expression in LMH cells and also rendered the VTGII promoter capable of being activated by cis-linked EREs in fibroblasts cotransfected with an estrogen receptor expression vector. Gel retardation and DNase I footprinting assays revealed four factor-binding sites within this promoter. We demonstrate that three of these sites bind C/EBP, SP1, and USF (or related factors), respectively; the fourth site binds a factor that we denote TF-V beta. The biological relevance of these findings is suggested by the fact that three of these binding sites map to sites previously shown to be occupied in vivo in response to estrogen.

In vivo and in vitro protein-DNA interactions at the distal oestrogen response element of the chicken vitellogenin gene: evidence for the same protein binding to this sequence in hen and rooster liver

The major egg white protein, vitellogenin, is synthesized in a tissue specific and oestradiol dependent manner in the liver of egg-laying hens. In this paper, we describe a detailed study of the protein-DNA interactions at the distal oestrogen response element (ERED) located 600 bp upstream of the start of transcription. In vivo footprinting of hepatocytes from adult hens and roosters with 0.5-0.0005% dimethylsulphate (DMS) revealed, at critical concentrations of DMS, protection of distinct guanosine residues within the ERED and adjacent downstream sequence in both cases. From this, it was concluded that there were proteins present in both tissues binding to this region in vivo. In vitro studies using missing base contact probing and proteolytic clipping band shift assays with hen and rooster liver nuclear extracts identified the ERE binding protein to be the same or very closely related in both tissues. Furthermore, the protein from rooster nuclear extracts bound to the ERE sequence even when the DNA was methylated at CpG dinucleotides, u.v. cross-linking experiments performed with bromodeoxyuridine substituted ERE, revealed that a nuclear protein with Mr of about 75,000-80,000 bound specifically to this sequence. These studies demonstrate that apart from the oestrogen receptor, at least one other protein can interact specifically with the chicken vitellogenin ERE, independently of hormonal expression of the gene.

Oestrogen facilitates the binding of ubiquitous and liver-enriched nuclear proteins to the apoVLDL II promoter in vivo

Using genomic and in vitro DNasel footprinting, we have analyzed protein-DNA interactions within the promoter region of the oestrogen-inducible gene encoding chicken apoVLDL II. The footprints coincide with previously detected guanosine-protein contacts in vivo. All footprints identified are present in the apoVLDL II-expressing liver exclusively and absent in hormone-naive liver, spleen and oviduct. They comprise recognition sites for the oestrogen receptor, the ubiquitous COUP-transcription factor, the liver-enriched C/EBP and/or DBP and the liver-specific LF-A1. In vitro, binding of protein to the oestrogen response element (ERE) is excluded by the prior binding of a protein, possibly C/EBP or DBP, to an adjacent element. The recognition sequence of the COUP-TF is also a target for LF-A1. The results suggests that oestrogen-dependent liver specific activation of the apoVLDL II promoter is established by the binding of the oestrogen receptor to EREs and multiple liver-enriched factors (C/EBP, DBP and LF-A1) to their nearby recognition sequences. Apparently, several DNA binding nuclear proteins cooperate to keep the promoter in a state that is accessible for the RNA polymerase complex.

In vivo monitoring of a cAMP-stimulated DNA-binding activity

The transcriptional activity of the tyrosine aminotransferase (TAT) gene is influenced by two major signal transduction pathways, by glucocorticoids and by glucagon acting via cAMP. We analyzed the effect of cAMP on protein-DNA interactions in vivo and on the transcription rate of the TAT gene. We demonstrate that a cAMP-responsive element (CRE) is located in a tissue-specific DNase I-hypersensitive region, 3.6 kb upstream of the start site of transcription. By using the genomic footprinting technique, we show that this sequence is occupied by protein in uninduced cells and that the in vivo footprint is transiently increased upon cAMP induction. Protein binding at the TAT-CRE correlates with the rate of transcription of the TAT gene. Cycloheximide treatment reveals that the genomic footprint is subject to rapid turnover; however, subsequent cAMP induction in the continued presence of cycloheximide restores the footprint partially. We conclude that as a part of the signal transduction pathway, a cAMP-dependent, post-translational modification increases the DNA-binding activity of a protein to the TAT-CRE and thereby stimulates the transcription rate of the TAT gene.

Chemical and photochemical probing of DNA complexes

An overview of the chemical and photochemical probes which over the past ten years have been used in studies of DNA/ligand complexes and of non-B-form DNA conformations is presented with emphasis on the chemical reactions of the probes with DNA and on their present 'use-profile'. The chemical probes include: dimethyl sulfate, ethyl nitroso urea, diethyl pyrocarbonate, osmium tetroxide, permanganate, aldehydes, methidiumpropyl-EDTA-Fell (MPE), phenanthroline metal complexes and EDTA/FeII. The photochemical probes that have been used include: psoralens, UVB, acridines and uranyl salts. The biological systems analysed by use of these probes are reviewed by tabulation.

A nuclear factor I-like activity and a liver-specific repressor govern estrogen-regulated in vitro transcription from the Xenopus laevis vitellogenin B1 promoter

A hormone-controlled in vitro transcription system derived from Xenopus liver nuclear extracts was exploited to identify novel cis-acting elements within the vitellogenin gene B1 promoter region. In addition to the already well-documented estrogen-responsive element (ERE), two elements were found within the 140 base pairs upstream of the transcription initiation site. One of them, a negative regulatory element, is responsible for the lack of promoter activity in the absence of the hormone and, as demonstrated by DNA-binding assays, interacts with a liver-specific transcription factor. The second is required in association with the estrogen-responsive element to mediate hormonal induction and is recognized by the Xenopus liver homolog of nuclear factor I.

A nuclear factor I-like activity and a liver-specific repressor govern estrogen-regulated in vitro transcription from the Xenopus laevis vitellogenin B1 promoter

A hormone-controlled in vitro transcription system derived from Xenopus liver nuclear extracts was exploited to identify novel cis-acting elements within the vitellogenin gene B1 promoter region. In addition to the already well-documented estrogen-responsive element (ERE), two elements were found within the 140 base pairs upstream of the transcription initiation site. One of them, a negative regulatory element, is responsible for the lack of promoter activity in the absence of the hormone and, as demonstrated by DNA-binding assays, interacts with a liver-specific transcription factor. The second is required in association with the estrogen-responsive element to mediate hormonal induction and is recognized by the Xenopus liver homolog of nuclear factor I.

cis- and trans-acting elements of the estrogen-regulated vitellogenin gene B1 of Xenopus laevis

Vitellogenin genes are expressed under strict estrogen control in the liver of female oviparous vertebrates. Gene transfer experiments using estrogen-responsive cells have shown that the 13 bp perfect palindromic element GGTCACTGTGACC found upstream of the Xenopus laevis vitellogenin gene A2 promoter mediates hormonal stimulation and thus, was called the estrogen-responsive element (ERE). In the Xenopus vitellogenin genes B1 and B2 there are two closely adjacent EREs with one or more base substitutions when compared to the consensus ERE GGTCANNNTGACC. On their own, these degenerated elements have only a low or no regulatory capacity at all but act together synergistically to form an estrogen-responsive unit (ERU) with the same strength as the perfect palindromic 13 bp element. Analysis of estrogen receptor binding to the gene B1 ERU revealed a cooperative interaction of receptor dimers to the two adjacent imperfect EREs which most likely explains the synergistic stimulation observed in vivo. Furthermore, a promoter activator element located between positions --113 and --42 of the gene B1 and functional in the human MCF-7 and the Xenopus B3.2 cells has been identified and shown to be involved in the high level of induced transcription activity when the ERE is placed at a distance from the promoter. Finally, a hormone-controlled in vitro transcription system derived from Xenopus liver nuclear extracts was exploited to characterize two additional novel cis-acting elements within the vitellogenin gene B1 promoter. One of them, a negative regulatory element (NRE), is responsible for repression of promoter activity in the absence of hormone. The second is related to the NF-I binding site and is required, together with the ERE, to mediate hormonal induction. Moreover, we detected three trans-acting activities in Xenopus liver nuclear extracts that interact with these regions and demonstrated that they participate in the regulation of the expression of the vitellogenin promoter in vitro.