This Article Abstract Full Text (PDF) Supplemental Data All Versions of this Article: 18/6/1411    most recent Author Manuscript (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager NURSA Molecule Pages Link Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Bourdeau, V. Articles by Mader, S. Search for Related Content PubMed PubMed Citation Articles by Bourdeau, V. Articles by Mader, S.

Genome-Wide Identification of High-Affinity Estrogen Response Elements in Human and Mouse

Département de Biochimie (V.B., J.D., D.N., S.M.), Université de Montréal, Montréal, Québec, Canada H3C 3J7; Departments of Physiology (V.B., J.H.W.), and Medicine (J.H.W., S.M.), McGill University, Montréal, Québec, Canada H3G 1Y6; McGill University and Genome Québec Innovation Centre (Y.N.), Montréal, Québec, Canada H3A 1A4; and European Molecular Biology Laboratories (R.M., N.B., F.G.), D-69117 Heidelberg, Germany

Address all correspondence and requests for reprints to: Sylvie Mader, Département de Biochimie, Université de Montréal, CP 6128 Succursale Centre-Ville, Montréal, Québec, Canada H3C 3J7. E-mail: sylvie.mader{at}umontreal.ca.

	ABSTRACT

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS Note Added in Proof REFERENCES

 
Although estrogen receptors (ERs) recognize 15-bp palindromic estrogen response elements (EREs) with maximal affinity in vitro, few near-consensus sequences have been characterized in estrogen target genes. Here we report the design of a genome-wide screen for high-affinity EREs and the identification of approximately 70,000 motifs in the human and mouse genomes. EREs are enriched in regions proximal to the transcriptional start sites, and approximately 1% of elements appear conserved in the flanking regions (–10 kb to +5 kb) of orthologous human and mouse genes. Conserved and nonconserved elements were also found, often in multiple occurrences, in more than 230 estrogen-stimulated human genes previously identified from expression studies. In genes containing known EREs, we also identified additional distal elements, sometimes with higher in vitro binding affinity and/or better conservation between the species considered. Chromatin immunoprecipitation experiments in breast cancer cell lines indicate that most novel elements present in responsive genes bind ER in vivo, including some EREs located up to approximately 10 kb from transcriptional start sites. Our results demonstrate that near-consensus EREs occur frequently in both genomes and that whereas chromatin structure likely modulates access to binding sites, far upstream elements can be evolutionarily conserved and bind ERs in vivo.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS Note Added in Proof REFERENCES

 
ESTROGENS, SUCH AS 17ß-estradiol (E2), are steroid hormones produced mainly by the ovary and act by endocrine, paracrine, and possibly autocrine mechanisms on a number of target tissues. Their actions on the reproductive, cardiovascular, and central nervous systems and on bone (1, 2, 3, 4) are mediated by two estrogen receptors, ER and ERß (4, 5), members of the nuclear receptor superfamily of ligand-inducible transcription factors (6, 7, 8, 9). A central DNA binding domain, which corresponds to the conserved C region (10, 11), is responsible for primary sequence recognition and cooperative binding by receptor dimers on elements containing two appropriately spaced copies of specific recognition motifs. The ligand binding domain contains a strong dimerization interface that stabilizes receptor homodimers or heterodimers (12, 13, 14) and enhances binding to imperfect motifs and even to single copies of recognition motifs (11, 15, 16, 17). However, ERs bind with highest affinity to 15-bp palindromes composed of PuGGTCA motifs separated by three variable bp (17, 18, 19). These estrogen response elements (EREs) are recognized by ERs with high specificity, because other nuclear receptors bind either different motifs or similar motifs with different spacing and/or orientation (7, 19, 20, 21). Studies using chromatin immunoprecipitation approaches have demonstrated that ER binding to EREs is hormone inducible in vivo (22, 23, 24, 25) and results in the ordered recruitment of a series of coactivator complexes, leading to histone acetylation, chromatin remodeling, and enhanced recruitment of the basal transcription machinery (26, 27, 28, 29, 30, 31, 32, 33).

Surprisingly, only a few near-consensus EREs have been characterized to date in the promoters of E2-regulated human or mouse genes. This probably reflects, in part, the fact that ERs can mediate estrogenic regulation through multiple mechanisms, including tethering to DNA via protein-protein interactions with other transcription factors (19, 34, 35), or so-called "nongenomic" effects, which affect gene expression by modulating the activity of upstream components of various intracellular signaling pathways (36, 37, 38). The paucity of known near-consensus EREs may also result from the fact that their presence is usually investigated only in relatively proximal sequences of E2-responsive genes. However, with sequences of the human and mouse genomes now available, genome-wide screening strategies can be used to expand the repertoire of potential high-affinity ER-binding sites.

In this study, we have sought to identify high-affinity EREs in the human and mouse genomes and to determine whether these elements represent bona fide ER binding sites in vivo. Indeed, little is known about the effect of chromatin structure or promoter context on ER binding in vivo. In addition, it is not clear how far from transcription start sites EREs can be positioned to recruit ERs and play a role in regulating gene expression. Our characterization of high-affinity EREs in the human and mouse genomes will greatly facilitate the study of the mechanisms modulating accessibility of ERs to their binding sites and of the subsequent events leading to transcriptional regulation by estrogen. This analysis will also provide a powerful resource for researchers analyzing the molecular events underlying the broad-ranging physiological actions of estrogens.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS Note Added in Proof REFERENCES

 
Identification of Sequences Corresponding to High-Affinity ER Binding Sites in Vitro
Palindromic response elements composed of two PuGGTCA motifs separated by 3 bp, such as the one found in the Xenopus vitellogenin A2 (VitA2) promoter, represent the highest affinity binding sites for ERs in vitro (Fig. 1A) (17, 18, 19). Surprisingly, only three perfect palindromic EREs have been identified to date in the vicinity of E2-regulated genes in the human genome. These genes, EBAG9, COX7A2L, and EFP/ZNF147, were cloned by screening of a CpG island genomic library for binding to ERs (39, 40, 41). In addition, a few near-consensus EREs have been characterized in human E2-responsive genes (see Table 1 for EREs diverging from consensus at one or two positions). A systematic analysis of single nucleotide substitutions in gel shift assays indicates that all elements with one variation still bind ERs in vitro (Fig. 1B). However, some replacements were more detrimental than others (–5C, –4C, –2A), reflecting steric or charge clashes between specific ER amino acids and bases of the variant EREs (Nguyen, D., G. Pesant, S. Cheinberg, and S. Mader, in preparation). Note that, as expected from the high degree of homology between the DNA binding domains of ER and ERß (90% in region C as defined in Ref.10), all mutations had a similar impact on complex formation by the two receptors. Introduction of symmetrical replacements in both arms of the palindrome generally affected ER binding much more drastically than when a unilateral change was introduced, although a few replacements were well tolerated even when present in both arms (Fig. 1C). Interestingly, examination of the known natural elements containing two replacements indicates that they all contain one change with mild effect on affinity (–6Py/+6Pu or –1G/+1C; Fig. 1, B and C and Table 1) combined with another replacement.

View larger version (36K): [in this window] [in a new window]   Fig. 1. In vitro Binding of ERs to Consensus and Near-Consensus EREs A, Consensus ERE with base numbering used in this study. B, Binding of human ER or ERß to response elements derived from the vitellogenin A2 (VitA2) ERE carrying a single replacement in one of the two arms of the palindrome. C, Binding of human ER or ERß to response elements derived from the VitA2 ERE carrying two symmetrical replacements. The percentage of probe found in a specific complex with human ER or ERß was quantified by phosphor imager and is shown with the corresponding SEM values.

View larger version (26K): [in this window] [in a new window]   Fig. 2. Distribution of EREs in the Human and Mouse Genomes A, Distribution of consensus EREs on human chromosomes. The percentage of nucleotides in individual chromosomes is plotted, along with the percentage of total response elements found per chromosome. B, Distribution of consensus EREs on mouse chromosomes. C, Distribution of EREs found between –10 and +5 kb of mRNA 5'-ends. The ratio between the number of EREs found and the expected number assuming a random distribution of these elements in the –10 to +5 interval was calculated for consecutive 1-kb segments of flanking sequences in the human (dark bars) and mouse (gray bars) genomes. For the elements conserved between the two genomes (white bars), the distribution was characterized only between –8 to +3 kb in order not to introduce a bias in the 2-kb constraint in relative distances between the two species.

These data suggest that the distribution of near-consensus EREs reflects mostly a random distribution of these sequences in the different chromosomes, with enrichment in the close vicinity of transcriptional start sites. The large total number of EREs located in the vicinity (–10 to +5 kb) of transcriptional start sites may suggest that these elements can mediate regulation of a much larger fraction of the genome in different estrogen target tissues than previously recognized. However, some of these elements may not represent binding sites in vivo, possibly because of chromatin accessibility, and/or may not participate in transcriptional regulation.

Identification of EREs Conserved in Human and Mouse
Near-consensus EREs previously identified in human genes often have counterparts at similar positions in their mouse orthologs (Table 1). Because conservation of newly identified EREs between the two species may indicate a functional role of these elements, we searched for the presence of EREs located in known human and mouse orthologs with less than 2 kb difference in distance from their respective transcriptional start sites (see Materials and Methods). EREs in 660 different pairs of orthologs were thus identified (supplemental Table 1; published as supplemental data on The Endocrine Society’s Journals Online web site at http://mend.endojournals.org). As expected from the distribution of EREs in the vicinity of genes, conserved EREs were also more abundant in the 0 to +1-kb region after the transcription start site (Fig. 2C; note that this analysis was intentionally limited to the –8 to +3-kb region. The numbers of conserved EREs found between –10 and –8 kb or +3 and +5 kb would appear artificially low, because a conserved ERE found at a distance of 2 kb or less in the other species may not be included in the –10 to +5-kb window from the transcriptional start site). However, conserved EREs were proportionately more represented in the 0 to +2-kb region than bulk EREs, whereas elements in the –5 to –8-kb region were markedly less represented (Fig. 2C). Nevertheless, we note that a significant number of conserved EREs mapped between –5 and –10 kb of the transcriptional start sites (24.6%). It remains possible that some of these distal EREs may be located closer to another gene, or even within an open reading frame, accounting for their conservation.

To investigate whether location of EREs in coding sequences (CDS) may account for their conservation, we characterized the position of these elements for a subset of conserved EREs, i.e. those that were identical in sequence between the two orthologs (Table 2). A total of 41 elements were found in the vicinity of 47 genes. Of these elements, 14 were located fully and four were found partly in a CDS (Table 2 bottom and middle, respectively), whereas 23 were not found in any annotated open reading frames (Table 2, top). More than half of these (12 EREs) were located between –1 to +1, another seven between –1 and –10 kb, and another four between +1 and +5 kb. Thus these results indicate that EREs that are conserved independently of their presence in a CDS are more likely to be found close to the 5'-end of mRNAs, although some elements are still located at several kb kilobase pairs from the start sites. Furthermore, the strong conservation of these elements, which is unlikely to result from chance alone, may indicate a role in the regulation of the neighboring genes.

Proximal and Distal EREs in Known Estrogen-Responsive Genes
To assess the potential of EREs found in this screen to bind ERs, we first compared our list of elements lying between –10 and +5 kb of the 5'-ends of known genes to published data on E2-up-regulated genes identified through expression studies in different human target tissues. This generated a list of 345 elements close to 236 genes (Table 3). Several of these genes were identified from a gene array experiment performed in the presence of cycloheximide, suggesting direct regulation (54). Interestingly, several promoters already shown to contain functional EREs were found to include additional elements, sometimes located more distally (Table 3; see also selected elements in Table 1). When assayed for binding to ER in gel shift assays, several of these novel elements bound similarly or even better than those previously characterized (Fig. 3A, OVGP1, CTSD, and TERT). Of interest, the mouse ortholog of the CTSD gene did not contain a near-consensus element corresponding to the known human proximal element, but our search identified a new element located 8 kb upstream that is conserved between human and mouse. Another example of a gene containing conserved distal EREs is the EFP/ZNF147 gene. Our screen (performed in this case with a larger cutoff) identified a conserved imperfect element in the 3'-untranslated region of the mouse and the human orthologs, different from the consensus ERE previously reported in the human gene (Table 1). The presence of distal elements in E2-regulated genes suggests that they may participate in the estrogenic regulation of these genes.

View larger version (47K): [in this window] [in a new window]   Fig. 3. In Vitro Binding of ER to EREs Identified in This Study A, Relative binding of hER to known and newly characterized EREs in estrogen target genes. B, Binding of hER to the novel EREs characterized through genomic screening. Binding is expressed as the percentage radioactivity of specifically bound probe vs. bound plus free probe.

Functional Analysis of Newly Identified EREs
To assess whether EREs selected among those found in the promoters of E2-responsive genes (Table 1, bottom) bind ER in vitro with the predicted high affinity, gel shift assays were performed with elements in their natural context (15 bp core with 6 bp of flanking sequences; Fig. 3B). Although flanking sequences slightly modulated binding compared with the VitA2-derived elements carrying the same core ERE sequence, all elements were found to bind ER efficiently. Chromatin immunoprecipitation (ChIP) experiments were performed next to monitor in vivo recruitment of hER on the newly identified EREs. These experiments were conducted in ER-positive MCF7 cells, ER-negative MDA-MB231 cells, or MDAMB231 cells stably transfected with hER (MDA::hER). As positive controls for ER recruitment, both the pS2/TFF1 and complement component 3 (C3) elements (Table 1) were used and, as expected, demonstrated E2-dependent recruitment of ER on their proximal promoters (Fig. 4). We investigated binding of ER to 14 EREs found in 10 genes known to be regulated by estrogen in breast cancer cells (ADORA1, CASP7, CTSD, GAPD, GREB1, IGFBP4, LY6E, NRIP1, SCNN1A, TERT; Table 1 and references within). In the two ER-expressing cell lines, E2-induced ER binding was observed on 13 of 14 of the elements (which include both conserved and nonconserved EREs; see Table 1), except for the GAPD ERE where binding was restricted to MCF7 cells (Fig. 4). No binding was observed in the parental MDA-MB231 cells, or with preimmune rabbit IgG, on any of the promoters tested. Consistent with the published regulation of these genes by estradiol in breast cancer cells, association of the fragments of genomic DNA around the transcriptional start sites with the TATA box binding protein (TBP) and the phosphorylated polymerase II (P-PolII) was induced in the presence of E2 (Fig. 4). Notably, in addition to the proximal elements (ERE2) found in the CTSD and TERT genes, both upstream elements (ERE1) bound ER. Binding of both the GREB1 EREs at –9.5 kb (ERE2) and at –1.5 kb (ERE3) was also observed, but the most upstream ERE (–21 kb, ERE1) did not bind despite its consensus ERE sequence. This may suggest that this element was not accessible for ER binding, possibly because of chromatin organization.

View larger version (34K): [in this window] [in a new window]   Fig. 4. In Vivo Binding of Human ER to EREs in ER-Positive Breast Cancer Cells Interaction of hER with selected EREs in the absence or presence of estrogen was investigated by chromatin-immunoprecipitation assay in ER-negative MDA-MB231 (MDA) cells, in ER-positive MCF7 cells, or in MDA-MB231 cells stably transfected with an expression vector for hER (MDA::hER) and treated or not with E2 for 2 h. Fragments containing putative EREs were immunoprecipitated from formaldehyde-fixed chromatin preparation with antibodies against hER, TBP, or phosphorylated polymerase II (P-PolII), or with preimmune rabbit IgG. A separate fragment spanning the transcription start site was used for PCR amplification when the start site was situated at more than 400 bp of the ERE. Whenever available, experimentally mapped transcription start sites were included in the +1 fragments. Experiments were performed two to three times and a representative result is shown.

Results from ChIP experiments demonstrate that it is possible to identify bona fide ER binding sites both in proximal and distal promoter sequences of E2 target genes, suggesting that the search for ER binding sites should not be limited to promoter sequences immediately upstream of the transcriptional start site. In addition, our results indicate that access to some high-affinity binding sites is restricted in vivo, possibly in a tissue-specific manner.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS Note Added in Proof REFERENCES

 
In this study, our goal has been to define high-affinity in vitro ER binding sequences and to characterize the occurrence of these motifs throughout the human and mouse genomes. This approach differs from those of previous studies, which have aimed at characterizing a larger array of potential binding sites for ERs or nuclear receptors in defined stretches of genomic sequences (60, 61). These approaches are based on nucleotide frequency matrices constructed via compilation of the relatively small number of known natural response elements and can identify many variant sequences. Although the sensitivity of the detection can be lowered for screening of genomic sequences, these types of detection programs cannot be easily used for genome-wide studies (the maximal length of input sequences is 100 kb for Dragon ERE Finder version 2.0 and 30 kb for the NUBIScan program). On the other hand, we chose to identify only high-affinity binding sites in a genome-wide approach to determine their distribution with respect to transcriptional start sites, their conservation between the mouse and human genomes, and their functionality as receptor binding sites in vivo. Our choice of elements was based both on functional validation of variant EREs by in vitro binding assays, and on the variations observed in known near-consensus sequences. Although not within the scope of this study, our functional characterization of EREs both in vitro and in vivo is also expected to contribute to the development of refined nucleotide frequency matrices for the detection of a wider range of elements in genomic sequences of interest.

Our screen identified a large number of EREs (71,119 in the human genome, or about one in every 43 kb of genomic DNA). Not surprisingly, the frequency of ERE occurrence was found to be highly dependent on their sequence. For instance, elements containing CG dinucleotides are drastically less represented than other EREs. CG dinucleotides are represented at a frequency of about 0.8% in the genome, which is five times less than the expected frequency based on the typical fraction of Cs and Gs (42). Elements containing two CG dinucleotides are found at only about 1–2% of the number of consensus EREs, which do not contain CG dinucleotides except possibly in the spacer. On the other hand, consensus EREs are slightly more represented than the expected frequency in random sequences (122% in human and 150% in mouse).

The sequence bias for some of the elements and the observation that some high-affinity EREs are found in tandem repeats precluded evaluation of which proportion of near-consensus EREs are under positive selective pressure, as could be expected if they mediate important regulatory roles. However, it was possible to evaluate the relative distribution of EREs in the vicinity of human or mouse genes. We have calculated that the number of EREs found in the human genome in the –1 to + 2-kb region around the 5'-end of mRNAs exceeds the number of elements expected from their frequency in the genome by 64%. It is unclear whether this represents conservation of elements found in the vicinity of transcriptional start sites, or bias for undetermined causes. A contributing factor may be the increased GC content in CpG islands, which are associated with the 5'-end of genes. Consensus EREs have an average GC content varying between 40 and 73% (depending on the identity of the variable bases). For most sequences, this is higher than the genome-wide average of 41%, but compatible with the 60–70% average in CpG islands (42). It is surprising, however, that the main peak of ERE representation is in the 0 to +1-kb region around the 5'-end of mRNAs both in human and mouse. Although the basis for this observation is unclear, it is worth noting that only the most 5'-mRNA start site for each gene was taken into consideration in this representation. Thus, some of these EREs may regulate promoters of downstream initiation sites.

We performed next a direct comparison of elements found in human and mouse orthologs to assess the fraction of total EREs that is conserved in the two species. Of the 9944 known orthologs (Mouse Genome Informatics database), 660 contained one or several conserved EREs at similar relative positions in the 5'-flanking sequences, i.e. at 2 kb from each other (708 conserved elements in total). This corresponds to about 1% of total elements in the human or mouse genome. Note that our criteria for conservation are relatively stringent, including both limited sequence variations (no more than two differences from consensus) and positional constraints (relative positions of EREs not further apart than 2 kb). Other possible sources of under-representation of conserved EREs are the incomplete identification of human/mouse orthologs, and the lack of systematic annotation of the mRNAs 5'-ends in the human and mouse genomes. Nevertheless, we have estimated that there is an approximately 74% overrepresentation of EREs compared with chance occurrence of these elements in both the human and mouse orthologs within 2 kb of each other (708 vs. 407 elements; see Materials and Methods). This overrepresentation suggests functional conservation of EREs throughout evolution, although we cannot rule out that functions unrelated to recruitment of ERs may contribute to this conservation (see below). The distribution of the conserved EREs was similar to that of total EREs in the vicinity of human or mouse genes but displayed a more marked concentration in the vicinity of the transcriptional start sites (see Fig. 2C). Interestingly, the percentage of conserved elements dropped markedly upstream of –5 kb, suggesting that upstream EREs are less likely to be conserved than elements closer to the initiation site, albeit these EREs still represent 24% of the total number of conserved elements between –10 and +5.

Apart from conservation due to functional importance, the most likely reason for preservation of an ERE is its location within the coding sequence of a gene. This parameter might contribute to the overrepresentation of EREs located downstream of the initiation start sites. In addition, some upstream elements may also be found in the coding sequences of other genes. The possible contribution of coding sequences in preserving ERE motifs in the human and mouse genomes was examined in the extreme case of EREs found to be totally conserved in sequence between the two species, including the Pu/Py and spacer base pair. Of 41 distinct EREs flanking 47 genes, 14 were found in coding sequences in both species, and four additional elements overlapped a CDS. It is unclear whether these elements could function as binding sites in vivo. Binding of yeast transcription factors such as Gcn4, Sbf, Mbf, and Rap1 to elements within open reading frames occurs reportedly less frequently than in promoter sequences (62, 63, 64) and, in the case of Gcn4, results in recruitment of histone acetyltransferase and SWI-SNF coactivator complexes, but not of the Mediator complex (64). Finally, 23 EREs were not found in coding sequences in either species. Interestingly, more than half of these EREs (12 elements) were located between –1 and +1 kb of the transcriptional start site, including eight elements between 0 and +1 kb. Therefore, although the presence of EREs in coding sequences may account, in part, for element conservation, it does not explain the large proportion of elements found immediately after the annotated transcriptional start site. Of interest in this regard, a recent publication indicated that an important fraction (40%) of NFB in vivo binding sites in chromosome 22 is located in intronic sequences (65).

Even though EREs appear less represented and/or conserved when located farther upstream of the transcriptional start site (–1 to –10 kb), our results suggest that they may contribute to estrogen target gene regulation. Indeed, some of these distal EREs are conserved between the human and mouse genomes and bind ER in vivo, as demonstrated by our ChIP experiments. These observations validate the choice of the –10 to +5-kb window around transcriptional start sites used in this study to identify potential ER binding sites. However, the frequency of near-consensus EREs found within this range strongly suggests that not all of these elements mediate transcriptional regulation of neighboring genes in a given cell context. It is unclear at this point whether access to some of these elements is restricted, or whether bound receptors may be unable to transactivate neighboring genes. The number of EREs in bulk genomic DNA also indicates that if all elements were accessible, binding sites would likely outnumber the molecules of receptor dimers in estrogen target cells. Indeed, the number of estrogen binding proteins in rat uterine cells was estimated via either biochemical or autoradiographical methods to vary between 5,000 and 30,000 molecules per cell (66, 67, 68, 69). Large-scale studies of transcription factor binding sites using ChIP assays have also indicated that the number of in vivo binding sites for different transcription factors in the genome is high. For instance, in ChIP experiments performed in HeLa S3 cells, 15% of the genes in chromosome 22 were found to contain functional NFB binding sites within 10 kb (65). In addition, 11% of the promoters were found to contain one or several high-affinity c-myc binding sites, suggesting competition between target sites in chromatin for limiting amounts of Myc protein levels (70). Alternatively, another possibility that arises from these observations is that stochastic and/or dynamic binding of transcription factors may allow usage of a larger number of binding sites.

We are aware that the total number of potential ER binding sites may be much higher than reported here. Elements not included in our search may represent good binding sites in vivo, including those that display a relatively low affinity in vitro. Whereas the compiled list of EREs is clearly not exhaustive, our ChIP experiments support the notion that the near-consensus EREs identified in estrogen-target genes represent likely binding sites in vivo (13 of 14 elements bound). Nevertheless, access to some of these elements may be restricted irrespective of their strength as ER binding sites in vitro. Neither a perfect ERE located within about 20 kb of the transcriptionally active GREB1 estrogen target gene, nor a near-consensus ERE located in the nontranscribed GAD2 gene were bound in ChIP experiments. Although results obtained in two breast cancer cell lines were very similar, it will be of interest in future studies to examine further how binding site recognition and coactivator recruitment are affected by cellular context. The future availability of large-scale gene expression studies performed in different tissues will allow us to expand the list of known E2 target genes and to determine whether the near-consensus EREs in their flanking sequences are bound in a tissue-specific manner. In addition, our data will facilitate future studies comparing the patterns of coactivators recruited by distal/proximal elements. For instance, it is possible that upstream EREs would recruit histone acetyltransferase and/or SWI/SNF complexes, resulting in long-range opening of chromatin and facilitating access of enhancer proteins to far upstream flanking sequences, whereas only proximal promoter sequences may be able to recruit mediator complexes. Alternatively, chromatin loops may allow upstream elements to participate in the recruitment of the basal machinery on the transcriptional start site. It is our hope that the database of high-affinity EREs derived from this study (URL: http://mapageweb.umontreal.ca/maders/eredatabase/) will prove a useful tool for the characterization of primary E2-regulated genes in various human and mouse target tissues and will ultimately enhance our understanding of the molecular mechanisms underlying the physiological actions of estrogens.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS Note Added in Proof REFERENCES

 
Bioinformatics
The algorithms developed (Nagai, Y., V. Bourdeau, and S. Mader, in preparation) enable a search of the NCBI fasta and gbs files of the Human Genome (Built 33; June 6, 2003) or of the Mouse Genome (Built 30; March 11, 2003) for a specified group of sequences and extract the positions of matching motifs in the genome contigs as well as the coordinates of the surrounding genes, mRNAs, and CDS within a preset cutoff distance of each motif. The programs, written in C, were run on an SGI Origin 2000 with 32 CPUs IP 27, R10000, 300-MHz processor (16 Go) using UNIX IRIX 6.5. Results presented in this article were generated using a cutoff of –10 to +5 kb of the mRNA 5'-ends (database available at http://mapageweb.umontreal.ca/maders/eredatabase/). Statistics for the total number of EREs occurring in the genomes were derived from lists of elements generated before the cutoff was applied. The total number of consensus EREs in the human genome was 891, and the number in the mouse genome was 923. Expected frequency in random DNA sequences was calculated as the total number of base pairs in the genome divided by the frequency of occurrence of a sequence with specified base pairs at 10 positions and two base pair choices at two positions (3,069334246/4¹¹ = 732 EREs in the human genome; 2,578250392/4¹¹ = 615 EREs in the mouse genome).

Conserved EREs were identified as elements present in both human and mouse gene orthologs (listed at the Mouse Genome Informatics database) at distances comprised between –10 to +5 kb from their respective mRNAs 5'-ends, and differing by less than +/– 2 kb. Because in numerous instances transcriptional start sites are mapped at the ATG codon in at least one species, or at the 5'-end of the gene for alternative upstream start sites, we have included EREs present in orthologs that were distant from each other by more than 2 kb when positions were calculated with respect to mRNAs 5'-ends, but less than 2 kb with respect to gene 5'-ends or initiator ATGs to minimize the underrepresentation of conserved EREs due to differential annotation in the human and mouse genomes. For example, the EBAG9 gene element, which is perfectly conserved in sequence, would not be selected on the basis of the distances to the 5'-end of mRNA because the mouse transcriptional start site was annotated at the ATG, but was included due to similar location with respect to the annotated 5'-end of the genes. Note that the 5'-end of the mRNA now coincides with the 5'-end of the gene in the most recent version of the mouse genome (Built 32).

The probability of an ERE to be found by chance in both the human and mouse orthologs was calculated by multiplying the probability of finding an element within –10 to +5 kb of the start site of one gene (14,074 elements in that window of the human genome/34,699 total genes) by that of finding an ERE in a 4-kb window (±2 kb) of corresponding sequence in the other species (12,828 elements x 4 kb/33,914 total mouse genes x 15 kb) and by the total number of orthologs (9,944 gene pairs) giving 407 expected elements.

Cell Culture
Hela cells were maintained in DMEM (Wisent, St-Bruno, Québec, Canada) supplemented with 5% fetal bovine serum (FBS, Sigma, Oakville, Ontario, Canada). Cells were switched 3 d before experiments to medium without phenol red containing charcoal-stripped serum. For gel shift assays, Hela cells were electroporated (10⁷ cells, 0.24 kV, 950 µF in a Bio-Rad Gene Pulser II apparatus; Bio-Rad Laboratories, Mississauga, Ontario, Canada) with 80 µg expression vector (pSG5-ER, pSG5-ERß or parental vector alone). Note that pSG5-ERß was generated by subcloning the open reading frame of hERß (71) from pCMVSport-ERß (a kind gift from Dr. T. Willson, Glaxo Wellcome, Inc., Research Triangle Park, NC) into the BamH1 site of pSG5 (72) by PCR amplification. Cells were treated with 25 nM E2 (Sigma Chemical Co., St. Louis, MO) 1–2 h before harvesting (48 h post transfection). Whole-cell extracts were prepared by three freeze-thaw cycles in gel retardation buffer as previously described (73).

For ChIP experiments, MCF-7 and MDA-MB231 cells were grown in DMEM (Sigma) supplemented with 10% fetal calf serum (Sigma). The medium was changed to phenol-red free DMEM supplemented with 2.5% dextran-charcoal treated fetal calf serum 48 h before hormone addition and replaced each 24 h. The MDA::hER cell line stably expressing hER was generated from hER-negative MDA-MB231 cells by transfection of pCDNA3.1/Hygro-hER (25).

Gel Shift and ChIP Assays
For gel shift assays, whole-cell extracts expressing ER or ERß or control extracts from cells transfected with the parental pSG5 vector were diluted to 120 mM KCl and assayed for binding to ³²P-labeled, double-stranded oligonucleotide probes (50,000 cpm/sample) as described previously (73). Radioactivity associated with bound or free probe was quantified using a Molecular Imager FX with the Quantity One software (Bio-Rad).

For ChIP assays, chromatin was cross-linked using 1.5% formaldehyde for 5 min at 37 C and fragmented by sonication as previously reported (25, 33), yielding fragments of average size approximately 350 bp. Antibodies against a C-terminal epitope of hER (HC20) and against TBP were purchased from Santa Cruz Biotechnology, Inc. (Heidelberg, Germany), and the antibody against phosphorylated polymerase II was purchased from Upstate Biotechnology, Inc. (Buckingham, UK). The sequences of the primers used in ChIP assays (synthesized by MWG GmbH, Ebersberg, Germany) are available upon request.

	Note Added in Proof

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS Note Added in Proof REFERENCES

 
A database of estrogen-responsive genes was published while this article was in revision (ERGDB, Ref.134). This database, compiled from a literature review of experimentally identified estrogen target genes, includes putative ERE sequences found in some of these genes through screening of their promoter-proximal sequences with Dragon ERE Finder version 2.0. Our database includes all ERGDB EREs that correspond to sequences used in our screen (although positions may vary, as our study used a later release of the human and mouse genomes), but also ERE sequences found beyond the cutoff distance used for the ERGDB (–4500 to +500 bp from transcriptional start sites), or located in genes not currently known to respond to estrogen.

	ACKNOWLEDGMENTS

 
We thank Thomas Bédouin for assistance with initial screening for EREs; Robert Sladek, Amy Hauth, and Franz Lang for helpful discussions; and Marie Pageau (Biochemistry, University of Montréal) for help with access to computers.

	FOOTNOTES

 
This work was supported by Canadian Institutes of Health Research (CIHR) and Natural Sciences and Engineering Research Council (NSERC) grants (to S.M.); funding from the Fonds de la Recherche en Santé du Québec (FRSQ) and from Genome Canada and Genome Quebec for the Steroid Atlas project (to S.M. and J.H.W.); and by European Molecular Biology Organization (EMBO) and European Molecular Biology Laboratory (EMBL) grants (to F.G.). V.B. acknowledges support from a CIHR Strategic Program to the Montreal Center for Experimental Therapeutics in Cancer and from a CIHR postdoctoral fellowship. J.D. is recipient of a studentship from the FRSQ, and R.M. was supported by a long-term fellowship from EMBO. J.H.W. and S.M. are chercheurs-boursier of FRSQ, and S.M. holds the Canadian Imperial Bank of Commerce Breast Cancer Research Chair at Université de Montréal.

Present address for R.M.: Equipe EMR; Unité Mixte de Recherche 6026, Centre National de la Recherche Scientifique, Campus de Beaulieu, 35046 Rennes Cedex, France.

Abbreviations: ADORA1, Adenosine A1 receptor; C3, complement component 3; CASP7, caspase7, apoptosis-related cysteine protease; CDS, coding sequence; ChIP; chromatin immunoprecipitation; COX7A2L, cytochrome c oxidase subunit VIIa polypeptide 2 like; CTSD, cathepsin D; E2, 17ß-estradiol; EBAG9, estrogen receptor binding site associated, antigen 9; EFP, estrogen-responsive finger protein; ER, estrogen receptor; ERE, estrogen response element; FGF, fibroblast growth factor; GAD2, glutamate decarboxylase 2; GAPD, glyceraldehydes-3-phosphate dehydrogenase; GREB1, gene regulated by estrogen in breast cancer protein; IGFBP4, IGF binding protein 4; KHDRBS3 (T-STAR/ETOILE), KH domain containing, RNA binding, signal transduction associated 3; LY6E, lymphocyte antigen 6 complex, locus E; NRIP1 (RIP140), nuclear receptor interacting protein 1; OVGP1, oviductal glycoprotein 1; P-PolII, phosphorylated RNA polymerase II; RBM, RNA-binding motif; RNF14, ring finger protein 14; SCNN1A, sodium channel, nonvoltage-gated 1; TBP, TATA box binding protein; TERT, telomerase reverse transcriptase; TESK1, testis-specific kinase 1; TFF1/pS2, trefoil factor 1; TSHB, TSHß; VitA2, Xenopus vitellogenin A2; ZNF147, zinc finger protein 147.

Received for publication November 13, 2003. Accepted for publication February 26, 2004.

	REFERENCES

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS Note Added in Proof REFERENCES

 

1. Couse JF, Korach KS 1999 Estrogen receptor null mice: what have we learned and where will they lead us? Endocr Rev 20:358–417[Abstract/Free Full Text]
2. McEwen BS, Alves SE 1999 Estrogen actions in the central nervous system. Endocr Rev 20:279–307[Abstract/Free Full Text]
3. Jordan VC 2001 Estrogen, selective estrogen receptor modulation, and coronary heart disease: something or nothing. J Natl Cancer Inst 93:2–4[Free Full Text]
4. Nilsson S, Makela S, Treuter E, Tujague M, Thomsen J, Andersson G, Enmark E, Pettersson K, Warner M, Gustafsson JA 2001 Mechanisms of estrogen action. Physiol Rev 81:1535–1565[Abstract/Free Full Text]
5. Green S, Chambon P 1988 Nuclear receptors enhance our understanding of transcription regulation. Trends Genet 4:309–314[CrossRef][Medline]
6. Beato M, Herrlich P, Schütz G 1995 Steroid hormone receptors: many actors in search of a plot. Cell 83:851–857[CrossRef][Medline]
7. Mangelsdorf DJ, Thummel C, Beato M, Herrlich P, Schutz G, Umesono K, Blumberg B, Kastner P, Mark M, Chambon P, Evans RM 1995 The nuclear receptor family: the second decade. Cell 83:835–839[CrossRef][Medline]
8. Chawla A, Repa JJ, Evans RM, Mangelsdorf DJ 2001 Nuclear receptors and lipid physiology: opening the X-files. Science 294:1866–1870[Abstract/Free Full Text]
9. Robinson-Rechavi M, Escriva Garcia H, Laudet V 2003 The nuclear receptor superfamily. J Cell Sci 116:585–586[Free Full Text]
10. Krust A, Green S, Argos P, Kumar V, Walter P, Bornert JM, Chambon P 1986 The chicken oestrogen receptor sequence: homology with v-erbA and the human oestrogen and glucocorticoid receptors. EMBO J 5:891–897[Medline]
11. Mader S, Chambon P, White JH 1993 Defining a minimal estrogen receptor DNA binding domain. Nucleic Acid Res 21:1125–1132[Abstract/Free Full Text]
12. Kumar V, Chambon P 1988 The estrogen receptor binds tightly to its responsive element as a ligand-induced homodimer. Cell 55:145–156[CrossRef][Medline]
13. Fawell SE, Lees JA, White R, Parker MG 1990 Characterization and colocalization of steroid binding and dimerization activities in the mouse estrogen receptor. Cell 60:953–962[CrossRef][Medline]
14. Cowley SM, Hoare S, Mosselman S, Parker MG 1997 Estrogen receptors and ß form heterodimers on DNA. J Biol Chem 272:19858–19862[Abstract/Free Full Text]
15. Kato S, Tora L, Yamauchi J, Masushige S, Bellard M, Chambon P 1992 A far upstream estrogen response element of the ovalbumin gene contains several half-palindromic 5'-TGACC-3' motifs acting synergistically. Cell 68:731–742[CrossRef][Medline]
16. Vanacker J-M, Petterson K, Gustafsson J-A, Laudet V 1999 Transcriptional targets shared by estrogen receptor-related (ERRs) and estrogen receptor (ER), but not ERß. EMBO J 18:4270–4279[CrossRef][Medline]
17. Klinge CM 2001 Estrogen receptor interaction with estrogen response elements. Nucleic Acids Res 29:2905–2919[Abstract/Free Full Text]
18. Klein-Hitpass L, Schorpp M, Wagner U, Ryffel GU 1986 An estrogen-responsive element derived from the 5' flanking region of the Xenopus vitellogenin A2 gene functions in transfected human cells. Cell 46:1053–1061[CrossRef][Medline]
19. Sanchez R, Nguyen D, Rocha W, White JH, Mader S 2002 Diversity in the mechanisms of gene regulation by estrogen receptors. Bioessays 24:244–254[CrossRef][Medline]
20. Glass CK 1994 Differential recognition of target genes by nuclear receptor monomers, dimers and heterodimers. Endocr Rev 15:391–407[CrossRef][Medline]
21. Khorasanizadeh S, Rastinejad F 2001 Nuclear-receptor interactions on DNA-response elements. Trends Biochem Sci 26:384–390[CrossRef][Medline]
22. Chen H, Lin RJ, Xie W, Wilpitz D, Evans RM 1999 Regulation of hormone-induced histone hyperacetylation and gene activation via acetylation of an acetylase. Cell 98:675–686[CrossRef][Medline]
23. Burakov D, Crofts LA, Chang CP, Freedman LP 2002 Reciprocal recruitment of DRIP/mediator and p160 coactivator complexes in vivo by estrogen receptor. J Biol Chem 277:14359–14362[Abstract/Free Full Text]
24. Shang Y, Hu X, DiRenzo J, Lazar MA, Brown M 2000 Cofactor dynamics and sufficiency in estrogen receptor-regulated transcription. Cell 103:843–852[CrossRef][Medline]
25. Reid G, Hubner MR, Metivier R, Brand H, Kos M, Gannon F 2003 Cyclic, proteasome-mediated turnover of unliganded and liganded ER on responsive promoters is an integral feature of estrogen signaling. Mol Cell 11:695–707[CrossRef][Medline]
26. Leo C, Chen JD 2000 The SRC family of nuclear receptor coactivators. Gene 245:1–11[CrossRef][Medline]
27. Robyr D, Wolffe A, Wahli W 2000 Nuclear hormone receptor coregulators in action: diversity for shared tasks. Mol Endocrinol 14:329–346[Free Full Text]
28. Rosenfeld MG, Glass CK 2001 Coregulator codes of transcriptional regulation by nuclear receptors. J Biol Chem 276:36865–36868[Free Full Text]
29. Rachez C, Freedman LP 2001 Mediator complexes and transcription. Curr Opin Cell Biol 13:274–280[CrossRef][Medline]
30. Dilworth FJ, Chambon P 2001 Nuclear receptors coordinate the activities of chromatin remodeling complexes and coactivators to facilitate initiation of transcription. Oncogene 20:3047–3054[CrossRef][Medline]
31. McDonnell DP, Norris JD 2002 Connections and regulation of the human estrogen receptor. Science 296:1642–1644[Abstract/Free Full Text]
32. Belandia B, Parker MG 2003 Nuclear receptors: a rendezvous for chromatin remodeling factors. Cell 114:277–280[CrossRef][Medline]
33. Métivier R, Penot G, Hubner MR, Reid G, Brand H, Kos M, Gannon F 2003 Estrogen receptor- directs ordered, cyclical, and combinatorial recruitment of cofactors on a natural target promoter. Cell 115:751–763[CrossRef][Medline]
34. Kushner PJ, Agard DA, Greene GL, Scanlan TS, Shiau AK, Uht RM, Webb P 2000 Estrogen receptor pathways and AP-1. J Steroid Biochem Mol Biol 74:311–317[CrossRef][Medline]
35. Safe S 2001 Transcriptional activation of genes by 17ß-estradiol through estrogen receptor-Sp1 interactions. Vitam Horm 62:231–252[Medline]
36. Kelly MJ, Levin ER 2001 Rapid actions of plasma membrane estrogen receptors. Trends Endocrinol Metab 12:152–156[CrossRef][Medline]
37. McEwen B, Akama K, Alves S, Brake WG, Bulloch K, Lee S, Li C, Yuen G, Milner TA 2001 Tracking the estrogen receptor in neurons: implications for estrogen-induced synapse formation. Proc Natl Acad Sci USA 98:7093–7100[Abstract/Free Full Text]
38. Moggs JG, Orphanides G 2001 Estrogen receptors: orchestrators of pleiotropic cellular responses. EMBO Rep 2:775–781[CrossRef][Medline]
39. Inoue S, Orimo A, Hosoi T, Kondo S, Toyoshima H, Kondo T, Ikegami A, Ouchi Y, Orimo H, Muramatsu M 1993 Genomic binding-site cloning reveals an estrogen-responsive gene that encodes a RING finger protein. Proc Natl Acad Sci USA 90:11117–11121[Abstract/Free Full Text]
40. Watanabe T, Inoue S, Hiroi H, Orimo A, Kawashima H, Muramatsu M 1998 Isolation of estrogen-responsive genes with a CpG island library. Mol Cell Biol 18:442–449[Abstract/Free Full Text]
41. Ikeda K, Sato M, Tsutsumi O, Tsuchiya F, Tsuneizumi M, Emi M, Imoto I, Inazawa J, Muramatsu M, Inoue S 2000 Promoter analysis and chromosomal mapping of human EBAG9 gene. Biochem Biophys Res Commun 273:654–660[CrossRef][Medline]
42. Lander ES, Linton LM, Birren B, Nusbaum C, Zody MC, Baldwin JC, Devon K, Dewar K, Doyle M, FitzHugh W, Funke R, Gage D, Harris K, Heaford A, Howland J, Kanne L, Lehoczky J, LeVine R, McEwan P, Mckernan K, Meldrim J, Mesirov JP, Miranda C, Morris W, Naylor J, et al 2001 Initial sequencing and analysis of the human genome. Nature 409:860–921[CrossRef][Medline]
43. Tsai S-J, Wu M-H, Chen H-M, Chuang P-C, Wing L-YC 2002 Fibroblast growth factor-9 is an endometrial stromal growth factor. Endocrinology 143:2715–2721[Abstract/Free Full Text]
44. Cohen RI, Chandross KJ 2000 Fibroblast growth factor-9 modulates the expression of myelin related proteins and multiple fibroblast growth factor receptors in developing oligodendrocytes. J Neurosci Res 61:273–287[CrossRef][Medline]
45. Colvin JS, Green RP, Schmahl J, Capel B, Ornitz DM 2001 Male-to-female sex reversal in mice lacking fibroblast growth factor 9. Cell 104:875–889[CrossRef][Medline]
46. Toshima J, Ohashi K, Okano I, Nunoue K, Kishioka M, Kuma K, Miyata T, Hirai M, Baba T, Mizuno K 1995 Identification and characterization of a novel protein kinase, TESK1, specifically expressed in testicular germ cells. J Biol Chem 270:31331–31337[Abstract/Free Full Text]
47. Toshima J, Koji T, Mizuno K 1998 Stage-specific expression of testis-specific protein kinase 1 (TESK1) in rat spermatogenic cells. Biochem Biophys Res Commun 249:107–112[CrossRef][Medline]
48. Venables JP, Vernet C, Chew SL, Elliott DJ, Cowmeadow RB, Wu J, Cooke HJ, Artzt K, Eperon IC 1999 T-STAR/ETOILE: a novel relative of SAM68 that interacts with an RNA-binding protein implicated in spermatogenesis. Hum Mol Genet 8:959–969[Abstract/Free Full Text]
49. Valve E, Penttila TL, Paranko J, Harkonen P 1997 FGF-8 is expressed during specific phases of rodent oocyte and spermatogonium development. Biochem Biophys Res Commun 232:173–177[CrossRef][Medline]
50. Kang HY, Yeh S, Fujimoto N, Chang C 1999 Cloning and characterization of human prostate coactivator ARA54, a novel protein that associates with the androgen receptor. J Biol Chem 274:8570–8576[Abstract/Free Full Text]
51. Couse JE, Mahato D, Eddy EM, Korach KS 2001 Molecular mechanism of estrogen action in the male: insights from the estrogen receptor null mice. Reprod Fertil Dev 13:211–219[CrossRef][Medline]
52. Murata Y, Robertson KM, Jones ME, Simpson ER 2002 Effect of estrogen deficiency in the male: the ArKO mouse model. Mol Cell Endocrinol 193:7–12[CrossRef][Medline]
53. Carreau S 2003 Estrogens—male hormones? Folia Histochem Cytobiol 41:107–111[Medline]
54. Soulez M, Parker MG 2001 Identification of novel oestrogen receptor target genes in human ZR75–1 breast cancer cells by expression profiling. J Mol Endocrinol 27:259–274[Abstract]
55. Mally MI, Cirulli V, Otonkoski T, Soto G, Hayek A 1996 Ontogeny and tissue distribution of human GAD expression. Diabetes 45:496–501[Abstract]
56. McCarthy MM, Kaufman LC, Brooks PJ, Pfaff DW, Schwartz-Giblin S 1995 Estrogen modulation of mRNA levels for the two forms of glutamic acid decarboxylase (GAD) in female rat brain. J Comp Neurol 360:685–697[CrossRef][Medline]
57. Bosma PT, Blazquez M, Fraser EJ, Schulz RW, Docherty K, Trudeau VL 2001 Sex steroid regulation of glutamate decarboxylase mRNA expression in goldfish brain is sexually dimorphic. J Neurochem 76:945–956[CrossRef][Medline]
58. Wondisford FE, Radovick S, Moates JM, Usala SJ, Weintraub BD 1988 Isolation and characterization of the human thyrotropin ß-subunit gene. Differences in gene structure and promoter function from murine species. J Biol Chem 263:12538–12542[Abstract/Free Full Text]
59. Gurr JA, Vrontakis ME, Athanasian EA, Wagner CR, Kourides IA 1986 Hormonal regulation of thyrotropin and ß subunit mRNAs. Horm Metab Res 18:382–385[Medline]
60. Podvinec M, Kaufmann MR, Handschin C, Meyer UA 2002 NUBIScan, an in silico approach for prediction of nuclear receptor response elements. Mol Endocrinol 16:1269–1279[Abstract/Free Full Text]
61. Bajic VB, Tan SL, Chong A, Tang S, Strom A, Gustafsson JA, Lin CY, Liu ET 2003 Dragon ERE Finder version 2: a tool for accurate detection and analysis of estrogen response elements in vertebrate genomes. Nucleic Acids Res 31:3605–3607[Abstract/Free Full Text]
62. Iyer VR, Horak CE, Scafe CS, Botstein D, Snyder M, Brown PO 2001 Genomic binding sites of the yeast cell-cycle transcription factors SBF and MBF. Nature 409:533–538[CrossRef][Medline]
63. Lieb JD, Liu X, Botstein D, Brown PO 2001 Promoter-specific binding of Rap1 revealed by genome-wide maps of protein-DNA association. Nat Genet 28:327–334[CrossRef][Medline]
64. Topalidou I, Thireos G 2003 Gcn4 occupancy of open reading frame regions results in the recruitment of chromatin-modifying complexes but not the mediator complex. EMBO Rep 4:872–876[CrossRef][Medline]
65. Martone R, Euskirchen G, Bertone P, Hartman S, Royce TE, Luscombe NM, Rinn JL, Nelson FK, Miller P, Gerstein M, Weissman S, Snyder M 2003 Distribution of NF-B-binding sites across human chromosome 22. Proc Natl Acad Sci USA 100:12247–12252[Abstract/Free Full Text]
66. Toft D, Gorski J 1966 A receptor molecule for estrogens: isolation from the rat uterus and preliminary characterization. Proc Natl Acad Sci USA 55:1574–1581[Free Full Text]
67. Clark JH, Gorski J 1970 Ontogeny of the estrogen receptor during early uterine development. Science 169:76–78[Abstract/Free Full Text]
68. Notides AC 1970 The binding affinity and specificity of the estrogen receptor of the rat uterus and anterior pituitary. Endocrinology 87:987–992[Medline]
69. Stumpf WE, Sar M, Zuber TJ, Soini E, Tuohimaa P 1981 Quantitative assessment of steroid hormone binding sites by thaw-mount autoradiography. J Histochem Cytochem 29:201–206
70. Fernandez PC, Frank SR, Wang L, Schroeder M, Liu S, Greene J, Cocito A, Amati B 2003 Genomic targets of the human c-Myc protein. Genes Dev 17:1115–1129[Abstract/Free Full Text]<