Cloning and Expression of a Novel Dominant-Negative-acting Estrogen Response Element-binding Protein in the Heterogeneous Nuclear Ribonucleoprotein Family*

Most genera of New World primates exhibit a compensated form of resistance to steroid hormones produced by the adrenal gland, gonads, and kidney. Estrogen resistance in New World primate cells is associated with the relative overexpression of a nonreceptor-related estrogen response element-binding protein (ERE-BP) that competes with estrogen receptor for ERE binding. Using the concatamerized ERE half-site (AGGTCAcag) in DNA affinity chromatography, we purified to homogeneity a 40–42-kDa ERE-BP. The affinity-purified ERE-BP bound specifically to either single- or double-stranded DNA bearing the consensus ERE half-site motif AGGTCA. Four distinct internal tryptic peptides from this protein were generated and shown to exhibit sequence similarity to proteins in the heterogeneous nuclear ribonucleoprotein family. These tryptic peptide fragments were used to generate a series of degenerate oligonucleotides that were successfully employed in isolating a full-length ERE-BP cDNA by polymerase chain reaction. Although a member of a family of proteins generally recognized for their ability to bind single strand RNA, the estrogen resistance-associated protein ERE-BP can effectively bind double strand DNA and competitively squelch estrogen receptor-directed transactivation.

receptors for hormone on a per cell basis, or some combination of the two events. More recently Brandon et al. 1 found that, when removed from the New World primate cell and transfected into an Old World primate cell, the ability of the New World primate estrogen receptor (ER) 2 to bind hormone and transactivate a reporter gene was not compromised. These data indicate that there exists in New World primate cells another protein(s) that interferes with the transactivating potential of the endogenous ER. We recently discovered such a protein that was greatly overexpressed in New World primate cells compared with Old World primate cells (9). This protein competes with ER for binding to the estrogen response element (ERE) and has been coined the "estrogen response element-binding protein" or ERE-BP. Here we report the successful purification and characterization of the 40 -42-kDa ERE-BP as a member of the heterogeneous nuclear ribonucleoprotein (hnRNP) family. We further demonstrate that endogenously expressed ERE-BP in estrogen-resistant primate cells as well as transient transfection of the ERE-BP cDNA in estrogen-responsive primate cells squelches estrogen-induced ER-ERE-directed transactivation.

EXPERIMENTAL PROCEDURES
Cells-The B95-8, owl monkey kidney (OMK), and Vero cell lines were obtained from the American Type Culture Collection (ATCC, Manassas, VA) and maintained as described previously (10). The B95-8 cell line was established from the common marmoset (Callithrix jacchus), a New World primate that exhibits gonadal steroid hormone resistance. The OMK cell line, also from a New World primate is estrogen-responsive. The Vero cell line originated from the kidney of an African green monkey, which is an Old World primate and estrogen-responsive.
Transient Transfection-A eukaryotic expression vector containing cDNA sequence encoding the full-length, wild-type human ER␣ and a reporter plasmid PERMLT-Luc containing a single ERE cloned upstream from the adenovirus major-late promoter sequence were generously provided by Dr. Elizabeth A. Allegretto (Ligand Pharmaceuticals, Inc., San Diego, CA) (11). 5 ϫ 10 5 B95-8 or OMK cells were seeded into wells of 6-well plates in phenol red-free medium containing 10% charcoal-stripped fetal calf serum and allowed to proliferate for 24 h. Transfections were performed in triplicate with the following DNA suspension: 6 g of ERE-BP cDNA in the 3.1 expression vector (Invitrogen, Carlsbad, CA) and/or 0.5 g of ER expression vector, 5 g of ERE luciferase reporter plasmid, and 5 g ␤-galactosidase plasmid with pGEM™-3Z vector as carrier DNA (Promega, Madison, WI) to a final concentration of 20 g DNA/ml LipoTAXI solution (Stratagene, La Jolla, CA). An equal volume of 20% fetal calf serum-supplemented, * This work was supported by National Institutes of Health Grants DK09093 (to H. C.) and AR37399 (to J. S. A.) and is dedicated to the memory Bayard D. Catherwood. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM  antibiotic-free medium was added to each well 5 h after transfection followed by the addition of 10 nM 17␤-estradiol. After an additional 48 h at 37°C, the cells were lysed, and luciferase and ␤-galactosidase activities were measured.
Preparation of Cellular Extracts-Confluent cultures of cells were harvested, pelleted, and washed twice in ice-cold phosphate-buffered saline (PBS; 20 mM Na 2 HPO 4 and 150 mM NaCl, pH 7.2) and cell extracts prepared as described previously by us (9). Postnuclear extracts were microconcentrated and desalted through a Microcon-30 filter (30 kDa molecular mass cut-off; Amicon, Beverly, MA). The protein concentration of extracts was determined by the method of Bradford (12).
Electromobility Shift Assay (EMSA)-Oligonucleotide sequence for the consensus ERE probe was 5Ј-CTAGAAAGTCAGGTCACAGTGAC-CTGATCAAT-3Ј and 5Ј-CTAGAAAGTCAGGTCACAGGATCAAT-3Ј for the ERE half-site probe. The double strand oligonucleotides used as specific competitors in EMSAs were prepared by annealing complementary sequence. The above-mentioned single strand consensus ERE oligonucleotide was used to prepare both single-and double strand labeled probes. The ERE 32-mer, either annealed with complementary sequence or not, was radiolabeled with [ 32 P]ATP (NEN Life Science Products) by T4 Kinase (Life Technologies, Inc.) to a specific activity of 10 8 cpm/g of DNA. Extract (ϳ10 g of protein) was incubated on ice with 2 g of poly(dI⅐dC) (Boehringer-Mannheim), 20 mM HEPES (pH 7.9), 100 mM KCl, 5 mM MgCl 2 in 10% glycerol for 15 min. Radiolabeled probe (50 fmol), with or without a 50-fold molar excess of radioinert, competitive, single-or double-stranded oligonucleotide, was then added and the incubation continued at room temperature for 15 min. Aliquots of the reaction were subjected to electrophoresis in a 6% polyacrylamide, 0.5ϫ TBE gel in 0.5ϫ TBE running buffer at 100 V. The gels were dried and exposed to Kodak X-OMAT AR film.
Protein Sequencing-Desalted and microconcentrated fractions eluted from DNA affinity chromatography were loaded onto 10% SDSpolyacrylamide gel electrophoresis and visualized with copper staining (Bio-Rad). The dominant 40 -42-kDa band was excised from the gel, destained, and washed twice with 50% acetonitrile, 50% distilled water. The protein band was subjected to "in-gel" digestion with trypsin and extracted (14). The resultant peptide fragments were separated by narrow bore high performance liquid chromatography using a Vydac C18 2-ϫ 150-mm reverse phase column on a Hewlett-Packard 1090 chromatograph equipped with a 1040 diode array detector. Eluted peptides were screened for length and homogeneity by matrix-assisted laser desorption time-of-flight mass spectrometry on a Finnigan Lasermat 2000. Selected fractions were then submitted to automated Edman degradation and amino acid sequencing on an Applied Biosystems 477A or Hewlett Packard G1005 protein sequencer (Harvard Microchemistry Facility, Cambridge, MA) as described previously (15).
Southwestern and Western Blot Detection of ERE-BP-Extracts from estrogen-resistant B95-8 and estrogen-responsive OMK cells were suspended in 2ϫ Laemmli buffer and heated at 97°C for 5 min. Denatured samples were loaded onto 10% SDS-polyacrylamide gels (5 g of protein per lane), electrophoresed for 1.5 h at 100 V, and electroblotted onto nitrocellulose membrane overnight at 4°C. The nitrocellulose membrane was blocked in 5% nonfat dry milk, 50 mM Tris (pH 7.5), 50 mM NaCl, 1 mM EDTA, and 1 mM dithiothreitol at room temperature for 1 h, and then hybridized in binding buffer (20 mM Hepes (pH 7.9), 100 mM KCl, 5% glycerol, 1 mM EDTA, 1 mM dithiothreitol, and 15 g/ml salmon sperm DNA) containing 10 6 cpm/ml of the radiolabeled ERE half-site double strand probe at 23°C for 2 h. The membrane was washed in the binding buffer at room temperature for 20 min and autoradiographed with an intensifying screen at Ϫ80°C. For Western blot analysis nitrocellulose membranes prepared as just described were incubated with monoclonal mouse anti-ER (Santa Cruz Biotech, Santa Cruz, CA) or anti-hnRNP-C 4F4 monoclonal antibody (16), both diluted 1:1000, horseradish peroxidase-conjugated anti-mouse IgG antibody (Sigma), and chemiluminescence detection reagent (ECL, Amersham Pharmacia Biotech).
RT-PCR-Approximately 200 ng of total RNA, isolated with total RNA isolation reagents (Life Technologies, Inc.), was used as template. Successful amplification of 351-base pair cDNA sequence was achieved with 30 cycles of reverse transcription-PCR using two of the primers, 5Ј-GTTTATGGAGGCCTTAGCTGGGACAC-3Ј and 5Ј-TCCATAGAGC-TCCCCATGGACAACAAG-3Ј. This PCR product was ligated into the PCR 2.1 (Invitrogen) and the cDNA sequence verified.
Amplification of a Full-length cDNA for ERE-BP-B95-8 poly(A)ϩ RNA (2.5 g) was used as a template to generate the 5Ј-and 3Ј-ends of the ERE-BP cDNA with the Marathon cDNA amplification kit (CLON-TECH Laboratories Inc., Palo Alto, CA). Second strand cDNA synthesis and adapter ligation were performed as instructed in the enclosed manual. The adapter-ligated cDNA was then used as a template for annealing adapter-and ERE-BP-specific primers for the RACE reaction: 5Ј-AGATCCCCACTGTTGCTGTTGCTG-3Ј and 5Ј-GGTGGCCTT-TCTCCAG-ATACACCT-3Ј for the 5Ј-and 3Ј-RACE reactions, respectively. Full-length cDNA was generated by end-to-end amplification using specific 5Ј and 3Ј primers. The amplified products were then separately subcloned into the PCR 3.1 expression vector and sequenced.

Squelching of Estrogen Receptor-Estrogen Response Element
Transactivation in New World Primate Cells-Most genera of New World primates are estrogen resistant in vivo (2) and crude New World primate cell extracts bearing ERE-BP have been shown to competitively inhibit binding of the ER to the ERE (9). However, resistance to estrogen action has not been previously shown in cells overexpressing the endogenous ERE-BP. Expression of a reporter gene bearing an ERE promoter region was reduced by 63% when transiently transfected into estrogen-resistant New World primate cells compared with similarly transfected, estrogen-responsive, wild-type cells (Fig.  1, panel A); the endogenous ER content of both cell lines was comparable ( Fig. 1 insert). This difference in basal reporter activity between hormone-resistant and hormone-responsive cells was markedly amplified when cells were co-transfected with the ERE-reporter and wild-type human ER constructs (Fig. 1, panel B). Moreover, basal reporter activity did not increase after incubation of transfected hormone-resistant B95-8 cells with the stimulatory ligand 17␤-estradiol as in estrogen-responsive OMK cells.
Affinity Purification and Functional Characterization of ERE-BP-The result of the above described experiment suggested that endogenous ERE-BP in the hormone-resistant New World primate cell line was competing with the estrogen-activated ER for binding to the transfected ERE and inhibiting transactivation. However, it did not exclude the possibility that there was more than a single protein overexpressed in New World primate cells that participates in the squelching of estrogen action. To rule out this possibility, we created a double strand DNA affinity support containing concatamers of the ERE half-site motif AGGTCAcag shown by us to greatly enrich New World primate cell extracts for ERE-binding proteins (9). Moreover, to avoid contamination of extracts with ERE-binding ER, we employed "low-salt", postnuclear extracts of B95-8 cells devoid of endogenous ER. An aliquot of each preparation of crude extract was examined for its ability to bind to the ERE in EMSA, and the specificity of ERE-BP-ERE binding was confirmed by the complete displacement of labeled ERE from the ERE-BP in the presence of a 50-fold excess of radioinert probe.
Preparations of B95-8 cell extracts with documented specific ERE binding activity were pooled and subjected to microconcentration (molecular mass cut-off of 30 kDa). The retentate was subjected to DNA affinity chromatography and eluted through a stepwise linear gradient of KCl. As previously reported (9), a symmetrical peak of ERE binding activity, unaffected by anti-ER antibody or by pre-exposure to 100 nM 17␤estradiol, eluted from the central portion of the gradient with most specific activity measured in the 0.5 M KCl fraction. Response element-binding protein(s) present in DNA affinitypurified extracts of B95-8 cells was further characterized for its ability to compete with equimolar concentrations of human ER for binding to the ERE (Fig. 2, panel A). EMSA of affinitypurified ERE-BP and wild-type human ER binding to human ERE probe, through increasing concentrations of radioinert competitive ERE, yielded very similar displacement plots. An aliquot of the affinity-purified ERE-BP was committed to poly-acrylamide gel electrophoresis, and the presence of a 40 -42-kDa species was revealed with silver staining (Fig. 2, panel B).
Using the labeled "half-sited" double strand ERE oligonucleotide for detection, Southwestern blot analysis of proteins in crude (Fig. 2, panel C) as well as in affinity-purified New World primate extracts showed that a 40 -42-kDa peptide(s) was responsible for ERE binding. ERE-BP, which cross-reacts with anti-hnRNP-C antibody (16), was detected by Western blot analysis and confirmed the presence of a plentiful 40 -42-kDa protein in the postnuclear and nuclear extracts of the estrogenresistant New World primate cell line B95-8 (Fig. 2, panel D). By contrast, ERE-BP was absent by polyacrylamide gel electrophoresis, Southwestern, and Western blot analysis of postnuclear extracts of any of the hormone-responsive primate cell lines.
Amino Acid Sequencing of ERE-BP Tryptic Peptides-Gels containing the ERE-BP were collected and the 40 -42-kDa bands excised from the gels. The gel slices were then subjected to in-gel proteolytic digestion and peptides were generated for amino acid sequencing. Four tryptic peptides from the parent protein were reproducibly recovered from the digest (Table I). All retained a high degree of sequence homology (range 57-100%) with at least four members of the large hnRNP family, hnRNP-A, -C, -D, and UP2 (17,18). Three of the highly conserved RNA binding motifs characteristic of hnRNPs (17) were identified in three of our peptides. The RNP-bearing tryptic peptides also possessed sequence identity with a single strand DNA-binding protein in the hnRNP-A/B family, called single strand D-box binding factor (19).
DNA Binding Properties of Affinity-purified ERE-BP-ERE-BP retarded a double strand DNA probe containing the AGGTCA half-site motif (Fig. 3, left, panel A). Addition of either excess, unlabeled single or double strand oligonucleotide competed away ERE-BP binding to double strand probe. If ERE-BP is similar in structure to single strand D-box binding factor, then it is reasonable to assume that it can also bind to  specific cis-acting sequences in single strand DNA; this was the case (Fig. 3, right, panel A). ERE-BP bound to a single strand probe containing the consensus ERE and was competed away with the addition of excess unlabeled of either single strand or double strand oligonucleotides containing one or two copies of the ERE half-site motif AGGTCA. In fact, unlabeled single strand probe bearing only a single ERE half-site appeared to be most effective in competing away ERE-BP binding to single strand probe; no competitive displacement of ERE-ERE-BP binding was observed with single strand oligonucleotides of similar size lacking the AGGTCA motif of similar size (Fig. 3,  panel B).
ERE-BP cDNA Cloning-The fact that the 40 -42-kDa ERE-BP was a member of the hnRNP family was confirmed by cloning a cDNA that codes for the entire open reading frame of the protein. An initial 351-base pair cDNA was cloned by RT-PCR using degenerate oligonucleotides based on the amino acid sequence of two of our tryptic peptides and total B95-8 cell RNA as template (see Fig. 4). This cDNA was then employed to develop 5Ј-and 3Ј-nested primers for use in cloning by 5Ј-and 3Ј-RACE a full-length cDNA for ERE-BP (Fig. 4). Compared with other members of the hnRNP family, including those known to bind to double strand DNA and influence transcription (19 -22), the deduced amino acid sequence for ERE-BP retained a high degree of sequence conservation in the two RNA binding domains, RBD-1 and RBD-2 (Fig. 5). However, outside of the highly conserved RNA binding domains, ERE-BP sequence diverged considerably from that of other members of the hnRNP family, particularly in the C-terminal extent of the protein.
Expression of Transiently Transfected ERE-BP-If the actions of the ERE-BP in vitro to compete with the ER for ERE binding are operative in vivo, then squelching of ER-EREdirected transactivation should take place in estrogen-responsive, wild-type cells that have been induced to transiently overexpress the protein. This finding was confirmed by the results FIG. 3. The ERE-BP binds specifically to either double or single strand DNA. The ERE-BP binds to either double strand (panel A, on the left) or single strand (panel A, on the right) oligonucleotide probes bearing the ERE consensus half-site. The labeled double and single strand probe was competitively displaced by a 50-fold molar excess of radioinert single strand probe containing a single copy of the AGGTCA motif. Labeled double strand probe was more effectively displaced than was labeled single strand probe by excess double strand oligonucleotides bearing either two or single copies of the AGGTCA motif. Panel B shows displaceable binding of ERE-BP only to radioinert oligonucleotides probes ERE half-site motif AGGTCA. Nonspecific single strand oligonucleotide sequences employed are 5Ј-TTAGGGTTAGGGTTAGGGTTAGGG-3Ј and 5Ј-CCTTTG-GCATGCTGCCAATATG-3Ј, for ssTTAGGG and ssCTF/NF, respectively.

FIG. 4. Sequence of the full-length ERE-BP cDNA.
The sequence of the fulllength ERE-BP cDNA and its deduced amino acid sequence is shown. The tryptic peptide fragments for creation of synthetic oligonucleotides used to clone to initial ERE-BP cDNA are underlined. The termination codon is indicated by an asterisk.
of the experiment shown in Fig. 6. Hormone-responsive OMK cells and hormone-resistant B95-8 New World primate cells were equivalently co-transfected with ERE-BP cDNA, human ER cDNA and ERE reporter constructs. Reporter activity was reduced 79% in estrogen-stimulated, ERE-transfected wildtype cells to levels close to that observed in hormone-resistant cells. DISCUSSION We recently reported that steroid hormone-and vitamin D-resistant New World primate cells overexpress a protein that has the capacity to compete with the human ER for binding to the consensus ERE (9). We showed that this protein, which we call the ERE-BP, was distinct from endogenous ER and, unlike the ER, was not dependent upon the presence of the standard tandem half-site motif for DNA binding. Rather, it appeared that the ERE-BP was bound most avidly by the single ERE half-site motif, AGGTCA. Taking advantage of this fact we constructed an affinity support bearing concatamers of this motif for purification of ERE-BP. Here we show that through the use of this affinity support we were able to purify in a single chromatographic step a 40 -42-kDa protein(s) that binds to the ERE. Generation and amino acid sequencing of four different tryptic peptides from the native protein showed ERE-BP to be a member of the large family of single strand nucleic acidbinding proteins whose most prominent members are in the subfamily of RNA-binding proteins termed hnRNPs.
There are now at least twenty different hnRNPs, designated hnRNP-A through -U, that are known to reside principally in the nucleus of vertebrate cells (17). In general, hnRNPs are considered RNA-binding proteins designed specifically for the metabolism and transport of pre-mRNA (23,24). They are among the most abundant nuclear proteins, usually present in quantities several orders of magnitude greater than that of most DNA-binding proteins including transcription factors. Recently, proteins in the hnRNP family have also been shown to exist outside the nuclear compartment of the cell, being capable of shuttling back and forth between the nucleus and the cytoplasm (23). hnRNPs have also been shown not to be specific for binding only RNA (24). Tomonaga and Levens (25,26) have provided evidence that hnRNPs are capable of binding single strand DNA; this suggests that the cis-acting elements do not require uracil as a component of the nucleic acid binding motif. In this report we show that ERE-BP is an example of an hnRNP-related protein that can interact specifically with either single or double strand DNA (Fig. 3) containing the AG-GTCA motif.
The deduced amino acid sequence of ERE-BP indicates the presence of two sets of the two consensus nucleic acid binding motifs, RNP-1 and RNP-2 (Fig. 5), which are characteristic of the hnRNP family of proteins (17). Although the central RNPcontaining domain of ERE-BP bears a high degree of sequence similarity to other hnRNPs, it is not yet known whether these same RNA binding sites are also responsible for the binding of ERE-BP to DNA. Regardless of the primary structural similarities among the hnRNPs and ERE-BP, there are at least three ways in which ERE-BP differs from the classical profile of an hnRNP. First, ERE-BP is not confined to the nuclear compartment of the cell. We isolated the protein in low-salt, postnuclear extracts of New World primate cells as well as from both the cytoplasmic and nuclear compartments of estrogenresistant cells. Second, it is not ubiquitously present in all vertebrate cells. As we have previously shown (9) and confirm FIG. 5. Comparison of the amino acid sequence in the RNA binding domains (RBD-1 and RBD-2) among ERE-BP and other members of the hnRNP family. Residues constituting the two RNA binding domains common to all proteins are numbered on the left. Conserved RNA-binding motifs RNP-1 and RNP-2 are boxed. GenBank TM accession numbers for hnRNPUP2, hnRNPD, SSdbf, hnRNPA1, and hnRNP-C are P07029, D55674, X80340, P09651, and P17132.
FIG. 6. Ability of transfected ERE-BP to squelch ER-directed ERE transactivation in wild-type cells. Co-transfection of plasmids bearing the human ERE-BP cDNA, human estrogen receptor-␣ cDNA, and an estrogen receptor-responsive promoter driving a luciferase reporter, into estrogen-responsive OMK cells (wt) and hormone-resistant B95-8 cells (rst) not treated or treated with 10 nM 17␤-estradiol (E2). Data are the mean of duplicate determinations of luciferase activity. here (Fig. 2), ERE-BP is expressed at very low to nondetectable levels in cells from hormone-responsive primates, including man. And third, ERE-BP appears to be versatile in its ability to bind nucleic acid (Fig. 3); ERE-BP can bind to single-or doublestranded DNA and, although yet unproved, the presence of the conserved RNA binding motifs in ERE-BP suggests that it may also interact with RNA.
The survival advantage conferred on primates of the New World by overexpression of the ERE-BP and the development of relative estrogen resistance remains an intriguing question.
Part of the answer may lie in the fact that there is overexpression of another family of proteins in New World primate cells that also has the potential to alter cellular responsiveness to 17␤-estradiol (27). These proteins are structurally related to proteins in the hsp-70 family of chaperone proteins and to the recently discovered intracellular vitamin D-binding protein (27,28). The latter is a high capacity, relatively low affinity (compared with the vitamin D receptor), largely cytoplasmic 25-hydroxylated vitamin D sterol-binding protein of unknown function up to now. We do not know whether one or more of these nonreceptor-related ligand-binding proteins acts in concert with ERE-BP to legislate estrogen resistance, or whether it acts to counter the dominant-negative actions of the ERE-BP by providing an easily available intracellular source of ligand for the ER, so that ER can compete more successfully with the ERE-BP for ERE binding. We do know that the state of relative estrogen resistance that exists in New World primates is not lethal and can be overcome by gonadal hypertrophy and maintenance of elevated serum 17␤-estradiol levels (4). Such is not the case in the vitamin D hormone system, where failure to maintain high serum levels of the vitamin D prohormone 25hydroxyvitamin D and vitamin D hormone 1,25-dihydroxyvitamin D results in morbid rachitic bone disease and death of New World primates during their adolescent growth period (29).
In summary, we have cloned and expressed the cDNA for an hnRNP-related, double strand DNA-binding protein, which is constitutively overexpressed in estrogen-resistant New World primate cells. We hypothesize that overexpression of nonreceptor-related and nonreceptor-associated proteins like the ERE-BP plays an important role in the dominant-negative regulation of steroid hormone-induced transactivation of hormone-responsive genes in almost all genera of New World primates. Preliminary data 3 indicate that there may be a subfamily of such DNA-binding hnRNPs responsible for a dominant-negative influence on the other gonadal steroid-, glucocor-ticoid-, and vitamin D hormone-response element-regulated promoters, and that these proteins, when expressed at much lower levels, may play a role in the normal transcription of human hormone-regulated genes in man. Absolute proof of these hypotheses awaits cloning and expression of the Old World primate (including Homo sapiens) homologs of these New World primate proteins.