Identification of rad27 mutations that confer differential defects in mutation avoidance, repeat tract instability, and flap cleavage.

In eukaryotes, the nuclease activity of Rad27p (Fen1p) is thought to play a critical role in lagging-strand DNA replication by removing ribonucleotides present at the 5' ends of Okazaki fragments. Genetic analysis of Saccharomyces cerevisiae also has identified a role for Rad27p in mutation avoidance. rad27Delta mutants display both a repeat tract instability phenotype and a high rate of forward mutations to canavanine resistance that result primarily from duplications of DNA sequences that are flanked by direct repeats. These observations suggested that Rad27p activities in DNA replication and repair could be altered by mutagenesis and specifically assayed. To test this idea, we analyzed two rad27 alleles, rad27-G67S and rad27-G240D, that were identified in a screen for mutants that displayed repeat tract instability and mutator phenotypes. In chromosome stability assays, rad27-G67S strains displayed a higher frequency of repeat tract instabilities relative to CAN1 duplication events; in contrast, the rad27-G240D strains displayed the opposite phenotype. In biochemical assays, rad27-G67Sp displayed a weak exonuclease activity but significant single- and double-flap endonuclease activities. In contrast, rad27-G240Dp displayed a significant double-flap endonuclease activity but was devoid of exonuclease activity and showed only a weak single-flap endonuclease activity. Based on these observations, we hypothesize that the rad27-G67S mutant phenotypes resulted largely from specific defects in nuclease function that are important for degrading bubble intermediates, which can lead to DNA slippage events. The rad27-G240D mutant phenotypes were more difficult to reconcile to a specific biochemical defect, suggesting a structural role for Rad27p in DNA replication and repair. Since the mutants provide the means to relate nuclease functions in vitro to genetic characteristics in vivo, they are valuable tools for further analyses of the diverse biological roles of Rad27p.

Substrate specificity of the exonuclease associated with calf DNA polymerase.

The digestion mechanism and substrate specificity of the 3' to 5' exonuclease associated with calf thymus DNA polymerase epsilon have been examined. The use of single-molecule mismatched DNA substrates has allowed further characterization of the structural substrate requirements of the nonprocessive exonucleolytic activity of DNA polymerase epsilon. The digestion characteristics of these substrates demonstrated that a single-stranded segment 5' to the double-stranded complementary region is not a prerequisite for efficient exonucleolytic degradation of the mismatched single-stranded segment at the 3' end of the molecule. In contrast to the known inhibitory effect of DNA polymerase activity at moderate concentration of monovalent ions, the distribution of digestive products was virtually unaffected by the addition of 80 mM KCl to the reaction. Aphidicolin, an inhibitor of DNA polymerase activity, also inhibits exonucleolytic activity on substrates containing a terminal mismatch, while little effect is observed for the digestion of single-stranded DNA substrates. However, if a long terminal mismatched DNA substrate is used to mimic the structure of a single-stranded DNA molecule, the extent of digestion is significantly decreased by the addition of aphidicolin. Inhibition of the digestion of single-stranded DNA by aphidicolin is also observed if a double-stranded complementary region with a 3' single-stranded DNA segment is added to the reaction. These results indicate that aphidicolin inhibits exonuclease activity by sequestering the enzyme to a portion of mismatched DNA molecules away from the site where the exonuclease must act. Additionally, they demonstrate that, although the polymerase and exonuclease active sites are structurally linked, polymerase function is not a necessary requirement for exonuclease function.

Estrogen receptor binding to nuclei from normal and neoplastic rat mammary tissues in vitro.

To investigate the role of hormones in regulating growth of neoplastic mammary cells, we established a heterologous assay for studying interactions of partially purified calf uterine [3H]estradiol-charged estrogen receptor ([3H]ER) with rat tumor nuclei in vitro. This system displays saturable high affinity binding of [3H]ER which is time and salt dependent. Optimal assay conditions required for the heterologous system were identical to those we reported for the homologous calf nuclear binding system. Specificity of [3H]ER binding was demonstrated; 10-fold excess unlabeled estrogen-charged ER (EcR) competed for greater than 90% of the [3H]ER binding sites and binding of [3H]estradiol (not complexed with ER) was less than 1% of [3H]ER binding. Binding of [3H]ER displayed tissue specificity in decreasing order: R3230AC mammary tumor greater than lactating mammary gland = liver greater than kidney greater than lung. Scatchard analysis of saturation data provided estimates of binding affinity to nuclei from R3230AC mammary tumors [Kd, 2.0 +/- 0.3 (SE) nM); the number of binding sites per nucleus for R3230AC tumors was 95,000 +/- 13,800. [3H]ER binding to nuclei isolated from R3230AC rat mammary tumors grown in intact rats was 40% higher than that observed in tumors from ovariectomized animals. Results of administration of individual pharmacological doses of either progesterone or an estrogen to ovariectomized rats did not restore nuclear ER binding levels in R3230AC tumors to those detected in tumors from intact rats. These results suggest that the physiological levels of endogenous hormones produced by the ovaries are important in regulating the number of ER binding sites in nuclei from these mammary tumors.

Effects of host hormonal status on binding of activated estrogen receptor to nuclei from R3230AC and 7,12-dimethylbenz[a]anthracene-induced mammary tumors.

The effects of various hormonal perturbations that alter growth of two different rat mammary tumors in vivo were investigated by study of the interactions of [3H]estradiol-charged estrogen receptors ([3H]ER) with tumor nuclei in vitro. Nuclei from the transplantable R3230AC adenocarcinoma were isolated after ovariectomy, estrogen treatment, or progesterone treatment. Saturable specific binding of [3H]ER to nuclei was assayed in this in vivo-like system. Scatchard analysis of [3H]ER-nuclear binding data indicated that these perturbations did not affect affinity, which ranged from Kd 1.0 to 2.4 nM. However, the number of [3H]ER-binding sites/nucleus was altered according to the treatment: intact rats, 94,500 +/- 4,200; ovariectomy, 70,400 +/- 3,200; ovariectomy plus estradiol, 82,100 +/- 5,800; and ovariectomy plus progesterone, 73,900 +/- 2,500. Nuclei from primary tumors induced by 7,12-dimethylbenz(a)anthracene displayed similar affinities for [3H]ER, although these tumors had fewer binding sites per nucleus. Animals bearing 7,12-dimethylbenz(a)-anthracene-induced tumors were either ovariectomized or made diabetic by administration of streptozotocin, perturbations that cause regression of the majority of tumors. The number of [3H]ER binding sites per nucleus, in tumors classified according to growth characteristics in host animals subsequent to hormonal perturbation, was: intact growing 36,300 +/- 3,400; ovex regressing, 15,400 +/- 3,400; ovex, estrogen-treated growing, 28,100 +/- 2,700; diabetic regressing, 19,500 +/- 2,400; diabetic static, 32,100 and diabetic growing, 42,000 +/- 7,100. These results indicate that (a) the number of nuclear ER-binding sites can be reduced by hormonal interventions that cause tumor regression and (b) endogenous ovarian hormones may play a role in regulating nuclear ER binding.

Inhibition of mammalian DNA polymerases by hematoporphyrin derivative and photoradiation.

Hematoporphyrin derivative (HPD) plus photoradiation caused the inactivation of DNA polymerases from calf thymus and R3230AC rat mammary tumor. Photosensitization of purified DNA polymerase-alpha as well as two forms of DNA polymerase-delta (I and II) from calf thymus were evaluated. Although all polymerase enzyme forms were inactivated at 70 micrograms HPD/ml, DNA polymerase-delta II was the most sensitive, displaying a 90% inactivation under conditions that did not cause significant inactivation of the other polymerase forms. Unlike DNA polymerase-alpha, the delta-forms have an associated 3'- to 5'-exonuclease activity. The exonuclease associated with DNA polymerase-delta II was uniquely sensitive to a low level of HPD and light exposure. DNA polymerase-delta II can be distinguished from other polymerase forms in cell extracts by its relative insensitivity to the polymerase inhibitor N2-(p-n-butylphenyl)deoxyadenosine 5'-triphosphate. In cytosols prepared from calf thymus and R3230AC rat mammary tumors, DNA polymerase-delta II was preferentially inhibited by HPD plus light. Furthermore, in experiments in which tumor-bearing rats were administered HPD prior to preparation of tumor cytosols, DNA polymerase-delta II was specifically inactivated by exposure to light. These results are discussed in view of their possible role in cancer therapy, and the potential use of HPD as a specific inhibitory agent of DNA polymerase-delta II is suggested.

Cooperative estrogen receptor interaction with consensus or variant estrogen responsive elements in vitro.

Specific binding of estradiol-liganded, partially purified calf uterine estrogen receptor (ER) to a 38-base pair estrogen responsive element (ERE) consensus sequence, containing the inverted repeat 5'-GGTCAnnnTGACC-3', was measured in vitro. The ERE sites were inserted as single or multiple tandem copies in a plasmid vector [p GEM-7Zf(+)]. Results showed that one dimeric ER can interact with one ERE, and steric constraints do not inhibit binding of ER to adjacent EREs. Molybdate-stabilized monomeric (4S) ER did not bind to EREs. ER bound to single and tandem double EREs with Kd values of 0.24 and 0.23 nM, respectively. When the plasmid contained three or more tandem copies of the ERE, ER bound in a cooperative manner, as indicated by convex Scatchard plots and Hill coefficients greater than 1.5. To determine those characteristics of the consensus sequence that are important for maximal high-affinity ER binding, ten variant ERE oligomer sequences were synthesized and cloned into pGEM-7Zf(+) as single copies or as four copies in tandem. ER binding affinity was maximal for the consensus ERE and was reduced for variants containing one or two nucleotide changes in the inverted repeat. The number of nucleotides separating the inverted repeat in the ERE was critical for high-affinity ER binding. Certain sequence-variant EREs when cloned as single copies bound less ER compared to the consensus ERE, yet when cloned as four tandem copies, ER binding displayed cooperativity by Scatchard and Hill analyses. Results demonstrate that cooperative interactions noted in vivo by others are present when measured in vitro. Results strongly imply that the number, spacing, and nucleotide sequence of EREs could precisely control the amount of ER binding to estrogen-responsive genes.

Duplex DNA binding activity of the avian myeloblastosis virus DNA polymerase.

Purified avian myeloblastosis virus DNA polymerase has a strong binding affinity for closed circular double-stranded DNA with no 3'-hydroxy termini. Because of this affinity the DNA polymerase can retain labeled native ColE1 DNA on nitrocellulose filters. When the reaction contains four enzyme molecules per ColE1 molecule about 50% of the DNA is retained. Higher enzyme: DNA ratios cause retention of nearly 100% of the DNA. The binding activity comigrates with DNA synthetic activity through ion-exchange chromatography and glycerol gradient centrifugation, in indication that it is an intrinsic activity of the DNA polymerase. Escherichia coli DNA polymerase I and bacteriophage T4 DNA polymerase do not show this binding activity, which suggests that it is not a common property of DNA polymerases. A novel application of enzyme kinetics using endonuclease-treated DNA reveals the relative quantities of enzyme molecules which are synthesizing at 3'-termini vs. molecules bound to double-stranded regions. Heat stability measurements indicate that the polymerizing activity of the enzyme can be almost completely eliminated while about half of the original binding activity is retained.

Pausing by retroviral DNA polymerases promotes strand transfer from internal regions of RNA donor templates to homopolymeric acceptor templates.

We have examined the ability of reverse transcriptases (RT) to catalyze strand transfer from internal regions of RNA templates, resulting in switching of a primer from one template to another. To study this phenomenon, we employed a system of donor and acceptor templates in which homologous strand transfer can occur from a homopolymeric sequence, positioned internally on the donor template. Our results indicate that reverse transcriptases from human immunodeficiency virus (HIV), avian myeloblastosis virus (AMV), and murine leukemia virus (MuLV) are all able to catalyze strand transfer from this sequence. Catalysis of this reaction is not dependent upon ribonuclease H (RNase H) activity, since an RNase H-deficient form of HIV-RT is able to catalyze the reaction efficiently. Additionally, N-ethylmaleimide, which inhibits RNase H but not polymerase activity, did not inhibit the template switching by either the native or RNase H-deficient forms of HIV-RT. Our data further indicate that template switching may be promoted by RT pausing at a specific site on the donor template. Conditions that increase RT pausing at this site also increase template switching. These results suggest that transient RT pausing at specific sites on the viral genome during reverse transcription may promote template switches that in turn lead to recombination.

Parameters that influence processive synthesis and site-specific termination by human immunodeficiency virus reverse transcriptase on RNA and DNA templates.

We have examined the parameters that determine the length and distribution of products synthesized processively by the human immunodeficiency virus reverse transcriptase (HIV-RT). On native or homopolymer templates, the overall length distribution of processively synthesized products is increased by increased temperature or deoxynucleoside triphosphate concentration, or decreased ionic strength. Specific terminations of processive synthesis on either native DNA or RNA templates occur most frequently at positions where the reverse transcriptase (RT) pauses during synthesis. These sites correlate with the template sequence 3'-(A/U)(A/U)(G/C)-5', particularly when this sequence is predicted to be base paired with another region of the template in a secondary structure. Many positions of termination are in similar positions on DNA or RNA templates. Notable exceptions are runs of A residues, which promote termination on DNA but not RNA templates. Termination intensities vary when different RTs are used demonstrating an influence of RT structure.

Characterization of an RNase H deficient mutant of human immunodeficiency virus-1 reverse transcriptase having an aspartate to asparagine change at position 498.

It has been proposed that Asp-443, Glu-478, and Asp-498 are important in RNase H mediated catalysis by human immunodeficiency virus-1 reverse transcriptase (Davies J.F., Hostomska, Z. Hostomsky, Z., Jordan, S.R. and Matthews, D.A. (1991) Science 251, 88-95; Mizrahi, V., Usdin, M.T., Harington, A. and Dudding, L.R. (1990) Nucleic Acids Res. 18, 5359-5363). Single point mutations at either position 443 (Mizrahi, V., Usdin, M.T., Harington, A. and Dudding, L.R. (1990) Nucleic Acids Res. 18, 5359-5363) or 478 (Schatz, O., Cromme, F.V., Grüninger-Leitch, F. and Le Grice, S.F.J. (1989) FEBS Lett. 257, 311-314) severely inhibit RNase H activity but have only small effects on polymerase activity. We show here that a single mutation at position 498 of Asp to Asn (mutant D498N) results in a stable enzyme with a 20-fold reduction in the ratio of RNase H to polymerase activity. The mutant and wild type enzymes were equally processive, paused in the same locations, and extended primers at the same rate during DNA synthesis on a heteropolymeric RNA template. The rate of elongation on the homopolymeric template poly(rA) was also the same. The results indicate that the mutation does not affect normally measured catalytic parameters of the polymerase function of the enzyme. D498N catalyzed strand transfer synthesis on homopolymeric, but not heteropolymeric templates, indicating that RNase H activity is not required for the former activity, but is for the latter.

Effects of multiple estrogen responsive elements, their spacing, and location on estrogen response of reporter genes.

Most highly estrogen-responsive genes possess multiple estrogen-responsive elements (EREs) that act synergistically to activate expression. Synergism between EREs appears to depend on structural features of the EREs and the promoter. To examine the activation process, we cloned single or multiple tandem copies of the consensus ERE into reporter plasmids. These plasmids contained either a chloramphenicol acetyl transferase reporter gene driven by a minimal promoter or a luciferase reporter gene driven by the Simian virus 40 (SV40) promoter. Using MCF-7 human breast cancer cells, we demonstrate that synergism among EREs depends on the number of EREs, their spacing, and the distance of the EREs from the promoter. The induction capacity of EREs falls off slowly with distance from the promoter. Remarkably, multiple EREs can induce effectively and synergize even when they are located more than 2000 nucleotides from the promoter. For EREs located immediately upstream of the promoter, both the distance separating the EREs and the distance to the promoter have to be optimal for synergy. Altering either distance changes the response from synergistic to additive. For distant EREs, presumed to interact by a looping mechanism at the promoter, the length of DNA between the EREs and the promoter is not critical. Synergy among closely spaced EREs that are far from the promoter only requires an optimal distance separating the ERE centers of symmetry. Interestingly, very widely separated EREs can also synergize, presumably also because of their ability to interact by looping. The estrogen response from single or multiple tandem copies of ERE half-palindromes near the SV40 promoter was also tested. The negligible induction capacity of a single half-site was not significantly increased in multiple sites. The biological role of half-EREs is not apparent in the system employed here.

An explanation for observed estrogen receptor binding to single-stranded estrogen-responsive element DNA.

Estrogen-inducible genes contain an enhancer called the estrogen response element (ERE), a double-stranded inverted repeat. The estrogen receptor (ER) is generally thought to bind to the double-stranded ERE. However, some reports provide evidence that an ER homodimer can bind a single strand of the ERE and suggest that single-stranded ERE binding is the preferred binding mode for ER. Since these two models describe quite different mechanisms of receptor action, we have attempted to reconcile the observations. Analyzing DNA structure by nuclease sensitivity, we found that two identical molecules of a single strand of DNA containing the ERE sequence can partially anneal in an antiparallel manner. Bimolecular annealing produces double-stranded inverted repeats, with adjacent unannealed tails. The amount of annealing correlates exactly with the ability of ER to bind bimolecular EREs. Either strand of an ERE could anneal to itself in a way that would bind ER. We conclude that ER binds only the annealed double-stranded ERE both in vitro and in vivo.

A microtiter well assay for quantitative measurement of estrogen receptor binding to estrogen-responsive elements.

Reproducible, rapid measurement of estrogen receptor (ER) binding to DNA was accomplished in microtiter wells treated so that ER-DNA complexes or DNA bound in preference to free ER. Mixtures of 35S-labeled DNA and [3H]estrogen-charged ER ([3H]ER), incubated to equilibrium in microfuge tubes, were transferred to microtiter wells previously treated with histone followed by gelatin. After binding of the DNA or ER-DNA complex to the treated wells, free ER was removed by washing. Radioactivity retained in each well was measured by placing individual wells from snap-apart microtiter plates directly in scintillation fluid. Binding of DNA was saturable, and ER-DNA complex binding was complete within 2 h at 4 C. The use of 35S-labeled DNA and [3H]ER allowed stoichiometric determination of ER bound to DNA. The amount of ER specifically bound to a consensus estrogen-responsive element (ERE) containing the inverted repeat GGTCAgagTGACC was determined by comparing ER bound to plasmid containing or lacking the ERE. At saturating concentrations of ER, plasmids bearing one, two, and four EREs in tandem bound approximately one, two, and four dimeric ER molecules, respectively. Scatchard analysis of saturation binding data revealed a Kd of 0.15 nM for specific ER binding to a single ERE site. Thus, the assay detects ER retaining both DNA-binding and estrogen-binding functions. ER complexed with DNA in the well was also detected using a monoclonal antibody specific for the receptor. Simple modifications of this method would allow study of other DNA-protein interactions.

Interactions of estrogen-receptor and antiestrogen-receptor complexes with nuclei in vitro.

Interactions of the estrogen-receptor complex (ER) with nuclei in vitro were shown to be dose, time, temperature, and tissue dependent. Specificity was demonstrated by the ability of ER charged with unlabeled 17 beta-estradiol to compete with [3H]ER for the nuclear acceptor sites. The ionic environment affected [3H]ER nuclear interactions; [3H]ER binding varied inversely with ionic strength, and apparent nuclear saturation was observed in the presence of 0.1 M KCl. Nuclear interactions of estrogen receptor charged with 17 beta-estradiol (ER) or monohydroxytamoxifen (AER) were compared. While both ER and AER (nonradiolabeled) were efficient competitors for [3H]ER nuclear binding, differences were observed when 3H-labeled ligands were used for saturation analysis of the nuclear acceptor sites. Scatchard analysis of the data revealed similar apparent Kd values for [3H]ER and [3H]AER binding to the nuclear sites (mean +/- SD, 1.2 +/- 0.5 X 10(-9) and 2.6 +/- 0.2 X 10(-9) M, respectively). However, the relative number of nuclear binding events consistently differed, with 28,000 +/- 7,300 sites/nucleus (mean +/- SD) for ER vs. 17,800 +/- 6,300 sites/nucleus for AER. Treatment of nuclei with 0-300 unit-min/ml DNAase I before incubation with receptor complexes resulted in a parallel percent decrease in the number of ER and AER nuclear binding sites. Saturation analysis performed with nuclei previously digested with 0-30 unit-min/ml DNAase I demonstrated that the apparent affinities of the receptor complexes for nuclear sites remained unchanged. Therefore, we suggest that both AER and ER bind to acceptor sites in the small portion of the chromatin hypersensitive to DNase I, but that fewer AER than ER can bind at least some of the ER nuclear acceptor sites. The lower binding capacity for AER may result in a pattern of gene expression that produces the agonist/antagonist effects observed with antiestrogens.

Characterization of a cytosolic inhibitor of calf estrogen receptor binding to nuclei.

An inhibitory component that diminishes estrogen receptor (ER) binding to nuclei in vitro is present in cytosol prepared from calf uterus. The inhibitor is heat stable and resistant to enzymatic treatment with trypsin, chymotrypsin, proteinase K, deoxyribonuclease I, or ribonucleases A, T1, and U2. Results of chromatography on DEAE-cellulose and Sephadex G-150 indicate that the factor is a negatively charged macromolecule. Inhibitory activity is sensitive to sequential digestion with chondroitinase ABC, hyaluronidase, and heparinase. Approximately 70% of the inhibitory activity is destroyed by treatment with heparinase alone. Heparitinase destroys only 30% of this activity. Furthermore, the addition of pure hyaluronic acid or chondroitin sulfate to the ER-nuclei binding assay results in little inhibition, whereas addition of heparin inhibits 75% of receptor binding. Overall, these results indicate that glycosaminoglycans, present in bovine uterine cytosol, are capable of inhibiting ER-nuclei interactions. The most potent inhibitory glycosaminoglycan displays heparin-like characteristics.

Bovine estrogen receptor binds chromatin at pre-existing nuclease hypersensitive sites.

Partially purified estrogen receptor prepared from heifer uterine cytosol, and labeled in vitro with tritiated estradiol, was used to locate receptor binding sites in target and non-target nuclei from various bovine tissues. Nuclei were digested to various extents with bovine pancreatic deoxyribonuclease I, micrococcal nuclease or endogenous nuclease and then assessed for their ability to bind charged estrogen receptor. After very brief digestion with DNAase I, such that only hypersensitive sites were cleaved, calf uterus nuclei were no longer able to bind estrogen receptor. Brief digests with micrococcal nuclease or endogenous nuclease, such that most DNA was still of polynucleosomal length, eliminated the binding ability of both calf and heifer uterus nuclei. These results suggest that estrogen receptor binds to pre-existing nuclease hypersensitive sites. Interestingly, nuclei digested by HaeIII restriction endonuclease, which cleaves at specific sequences, demonstrated no loss of labeled estrogen receptor binding, even though digestion products were of similar size to those obtained from nuclei after treatment with the other nucleases. Since nuclease hypersensitive sites occur in regulatory regions of actively transcribed genes, including estrogen-inducible genes, binding of estrogen receptor at these sites, in vivo, may be part of the mechanism by which transcription is induced.

Comparison of tamoxifen ligands on estrogen receptor interaction with estrogen response elements.

The estrogen receptor (ER) is a ligand-activated transcription factor that binds to specific DNA sequences, estrogen response elements (EREs). Estradiol-liganded ER (E2-ER) binds cooperatively to stereoaligned EREs that are surrounded by naturally-occurring AT-rich sequences with a stoichiometry of one E2-ER dimer per ERE. When ER is bound by 4-hydroxytamoxifen (4-OHT), the active metabolite of the widely used therapeutic antiestrogen tamoxifen (TAM), the receptor binds to EREs with high affinity. However, one molecule of 4-OHT ligand dissociates from the ER dimer apparently during the process of binding to DNA, yielding a stoichiometry of one [3H]4-OHT molecule per ERE. To determine whether DNA-binding induced ligand dissociation is a general property of type I antiestrogens that are not covalently attached to the ER, we examined the interaction of ER liganded by tamoxifen (TAM) with EREs. We demonstrate that TAM-ER binds EREs with lower affinity than E2-ER, 4-OHT-ER, or ER liganded by the covalent antiestrogen tamoxifen aziridine. Unlike E2-ER, both TAM and 4-OHT-ER bind EREs non-cooperatively. Like 4-OHT, TAM appears to dissociate from the liganded ER as the receptor binds EREs. Additionally, partial proteolysis of ERE-bound ER by trypsin revealed different cleavage patterns for E2 versus 4-OHT and TAM. These findings indicate that the behavior of the ER liganded by TAM is generally similar to that of the antiestrogen 4-OHT.

Fusion estrogen receptor proteins: toward the development of receptor-based agonists and antagonists.

Estrogen-induced signaling mediated by estrogen receptors (ERs) is also affected by aberrant ERs that act as constitutively active or dominant negative modulators. Variant ERs can contribute to carcinogenesis and to the loss of estrogen responsiveness, rendering antiestrogen therapy ineffective. Determining target gene response during co-synthesis of different ER species is difficult, because dimers formed in the presence of more than one ER species are a heterogenous population of homo- or heterodimers. We engineered a homofusion ERalpha as a prototype single-chain receptor by genetically conjugating two ER monomers into a covalently fused single-chain protein to obtain a homogeneous population. This permits analysis of symmetrical or asymmetrical mutations that simulate variant homo- and heterodimers. Although a monomer, the homofusion receptor exhibited similar biochemical and functional properties to the dimeric ERalpha. We used activation function-2 (AF2) defective mutants as a model in either one or both receptor domains for a dominant-negative phenotype by suppressing the reporter activity induced by the WT receptor. When co-expressed with ERalpha, the fusion variant deficient in both AF2 functions suppressed the reporter activity effectively induced by ERalpha. These results show the utility of fusion receptors as models for generation of receptor-based agonists and antagonists.

Antiestrogen-liganded estrogen receptor interaction with estrogen responsive element DNA in vitro.

The mechanism whereby antiestrogens alter the ability of the estrogen receptor (ER) to enhance transcription of estrogen-regulated genes is largely unknown. The effect that selected estrogenic and antiestrogenic ligands have on binding of ER to specific DNA sequences, estrogen responsive elements (EREs) has been quantitated. No differences in purification properties of calf uterine ER liganded with 4-hydroxytamoxifen (4-OHT-ER), ICI 164,384 (ICI 164,384-ER) or estradiol (E2-ER) were detected. A microtiter well plate assay was employed in which liganded ER bound to plasmid DNA is preferentially retained compared to free liganded ER. Binding of E2-ER, 4-OHT-ER, or ICI 164,384-ER was measured to plasmids containing or lacking a 38bp consensus ERE in vitro. The EREs tested contain an inverted repeat (5'-CAGGTCAGAGTGACCTG-3'). Both E2-ER and 4-OHT-ER showed similar high affinity specific binding (Kd = 0.24 and 0.16 nM, respectively) to one copy of the ERE. ICI 164,384-ER did not bind to plasmids containing one ERE. At saturation, however, 4-OHT-ER binding was about 50% of that observed for E2-ER. When the plasmid contained 3 or 4 tandem copies of the ERE, binding of E2-ER, 4-OHT-ER, and ICI 164,384-ER binding was measurable. E2-ER bound in a cooperative manner as suggested by convex Scatchard plots and Hill coefficients > 1.5. In contrast, 4-OHT-ER binding displayed much reduced cooperativity, and ICI 164,384-ER did not display cooperative binding. From these results, we propose that the conformation of ER induced by 4-OHT reduces its binding capacity to this consensus ERE without altering its affinity of binding. Furthermore, higher order protein-protein interactions between antiestrogen-liganded ER bound to DNA differ from those of E2-ER bound to ERE.

Differential impact of flanking sequences on estradiol- vs 4-hydroxytamoxifen-liganded estrogen receptor binding to estrogen responsive element DNA.

The mechanism by which antiestrogens antagonize the ability of estrogen receptor (ER) to induce the transcription of estrogen-regulated genes is only partially understood. To examine the effect of estrogen responsive element (ERE) stereoalignment and flanking sequences on estradiol-liganded ER (E2-ER)-ERE and antiestrogen-liganded ER (4-hydroxytamoxifen-liganded ER or 4-OHT-ER)-ERE binding, several dimeric EREs, containing a perfect inverted repeat (5'-GGTCAgagTGACC-3') but lacking the AT-rich flanking sequences typical of highly estrogen-responsive promoters, were cloned into a plasmid vector. The ERE centers of symmetry were spaced 1.5, 2.0, 3.0, 6.4 and 6.7 helical turns apart. E2-ER and 4-OHT-ER binding to these constructs was specific and saturable, but orientation-independent and, in contrast to our earlier work with E2-ER binding to AT-rich EREs, not cooperative. The affinity of E2-ER binding decreased as the distance between adjacent EREs was increased, suggesting that E2-ER binding to closely spaced EREs is more stable (Kd = 0.38, 0.58, 0.83, 1.23, and 0.96 nM, respectively, for the above spacings). In contrast, the affinity of 4-OHT-ER binding increased with increased ERE spacing (Kd = 2.90, 4.79, 1.39, 1.77, and 0.92 nM, respectively). The presence of AT-rich sequences flanking the ERE increased the binding affinity of E2-ER and 4-OHT-ER, an increase reflected in slower dissociation rates of ER from these EREs. The AT-rich sequence also enhanced the binding capacity of E2-ER but not 4-OHT-ER. Since the binding capacity of 4-OHT-ER is identical with or without an AT-rich region, we suggest that flanking sequences are more important in stabilizing E2-ER binding and may be critical for cooperative binding to stereoaligned EREs.