Steroid receptor-nuclear matrix interactions. The role of DNA

The transportosome system as a model for the retrotransport of soluble proteins

The classic model of action of the glucocorticoid receptor (GR) sustains that its associated heat-shock protein of 90-kDa (HSP90) favours the cytoplasmic retention of the unliganded GR, whereas the binding of steroid triggers the dissociation of HSP90 allowing the passive nuclear accumulation of GR. In recent years, it was described a molecular machinery called transportosome that is responsible for the active retrograde transport of GR. The transportosome heterocomplex includes a dimer of HSP90, the stabilizer co-chaperone p23, and FKBP52 (FK506-binding protein of 52-kDa), an immunophilin that binds dynein/dynactin motor proteins. The model shows that upon steroid binding, FKBP52 is recruited to the GR allowing its active retrograde transport on cytoskeletal tracks. Then, the entire GR heterocomplex translocates through the nuclear pore complex. The HSP90-based heterocomplex is released in the nucleoplasm followed by receptor dimerization. Subsequent findings demonstrated that the transportosome is also responsible for the retrotransport of other soluble proteins. Importantly, the disruption of this molecular oligomer leads to several diseases. In this article, we discuss the relevance of this transport machinery in health and disease.

Androgen receptor activity is affected by both nuclear matrix localization and the phosphorylation status of the heterogeneous nuclear ribonucleoprotein K in anti-androgen-treated LNCaP cells

The androgen receptor (AR) plays a central role in the development and progression of prostate cancer (PCa) and anti-androgen therapy is a standard treatment. Unfortunately, after a few years, the majority of patients progress, developing androgen-independent PCa. AR-driven gene transcription recruits a large number of co-activator/co-repressor complexes; among these, the heterogeneous nuclear ribonucleoprotein K (hnRNP K) directly interacts with and regulates the AR translational apparatus. Here we examined AR and hnRNP K expression in response to the treatment of LNCaP cells with anti-androgen cyproterone acetate (CPA) or bicalutamide (BIC). AR and hnRNP K modulation and compartmentalization were studied by Western blot and confocal microscopy. Phosphate-affinity gel electrophoresis was employed to examine how anti-androgens modified hnRNP K phosphorylation. 10(-6) M CPA significantly stimulated LNCaP proliferation, whereas for 10(-4) M CPA or 10(-5) M BIC an antagonistic effect was observed. After anti-androgen treatment, AR expression was remarkably down-regulated within both the cytoplasm and the nucleus; however, when CPA had an agonist activity, the AR associated with the nuclear matrix (NM) increased approximately 2.5 times. This increase was synchronous with a higher PSA expression, indicating that the NM-associated AR represents the active complex. After BIC treatment, hnRNP K expression was significantly lower in the NM, the protein was hypophosphorylated and the co-localization of AR and hnRNP K decreased. In contrast, CPA as an agonist caused hnRNP K hyperphosphorylation and an increase in the co-localization of two proteins. These findings demonstrate that, in vitro, there is a strong relationship between NM-associated AR and both cell viability and PSA levels, indicating that AR transcriptional activity is critically dependent on its subnuclear localization. Moreover, the agonistic/antagonistic activity of anti-androgens is associated with modifications in hnRNP K phosphorylation, indicating an involvement of this protein in the AR transcriptional activity and likely in the onset of the androgen-independent phenotype.

Scaffold attachment factor B (SAFB)1 and SAFB2 cooperatively inhibit the intranuclear mobility and function of ERα

Estrogen receptor alpha (ERα) plays a key role in physiological and pathophysiological processes as a ligand-activated transcriptional factor that is regulated by cofactors. ERα-mediated transcriptional regulation is closely correlated with the mobility of ERα in the nucleus in association with the nuclear matrix, the framework for nuclear events including transcription. However, the relationship between ERα mobility and the cofactors of ERα is unclear. Scaffold attachment factor B1 (SAFB1) and its paralog SAFB2 are nuclear matrix binding proteins that have been characterized as ERα corepressors. Here, using chimeric fluorescent proteins (FPs), we show that SAFB1 and SAFB2 colocalize with ERα in the nucleus of living cells after 17β-estradiol (E2) treatment. Co-immunoprecipitation (co-IP) experiments indicated that ERα interacts with both SAFB1 and SAFB2 in the presence of E2. Fluorescence recovery after photobleaching analysis revealed that SAFB1 and SAFB2 each decrease ERα mobility, and interestingly, coexpression of SAFB1 and SAFB2 causes a synergistic reduction in ERα dynamics under E2 treatment. In accordance with these mobility changes, ERα-mediated transcription and proliferation are cooperatively inhibited by SAFB1 and SAFB2. These results indicate that SAFB1 and SAFB2 are crucial repressors for ERα dynamics in association with the nuclear matrix and that their synergistic regulation of ERα mobility is sufficient for inhibiting ERα function.

Subnuclear compartmentalization of sequence-specific transcription factors and regulation of eukaryotic gene expression

To date, the majority of the research regarding eukaryotic transcription factors has focused on characterizing their function primarily through in vitro methods. These studies have revealed that transcription factors are essentially modular structures, containing separate regions that participate in such activities as DNA binding, protein-protein interaction, and transcriptional activation or repression. To fully comprehend the behavior of a given transcription factor, however, these domains must be analyzed in the context of the entire protein, and in certain cases the context of a multiprotein complex. Furthermore, it must be appreciated that transcription factors function in the nucleus, where they must contend with a variety of factors, including the nuclear architecture, chromatin domains, chromosome territories, and cell-cycle-associated processes. Recent examinations of transcription factors in the nucleus have clarified the behavior of these proteins in vivo and have increased our understanding of how gene expression is regulated in eukaryotes. Here, we review the current knowledge regarding sequence-specific transcription factor compartmentalization within the nucleus and discuss its impact on the regulation of such processes as activation or repression of gene expression and interaction with coregulatory factors.

Membrane restraint of estrogen receptor alpha enhances estrogen-dependent nuclear localization and genomic function

Estrogen receptor (ER) alpha localizes to both the nucleus and the plasma membrane, mediating estrogen-dependent genomic and nongenomic signaling, respectively. In some cells, ERalpha appears to be excluded from the nucleus, and it is unclear whether genomic signaling takes place. The purpose of this study was to determine whether membrane-associated ERalpha is capable of genomic signaling, or whether this pool of receptors strictly serves membrane-mediated signaling. ERalpha fused to the C-terminal cytoplasmic tail of bovine rhodopsin (Rh-ERalpha) activates ER response element-dependent transcription only in the presence of estrogen; the activity is antagonized by the estrogen antagonist ICI 182,780 and by the dominant-negative mutant of ERalpha and is unaffected by inhibitors of MAPKs and Akt signaling, indicating that this was due to direct genomic action. The activity of Rh-ERalpha containing the activating Y537S mutation was also estrogen dependent, suggesting that estrogen gated the entry of Rh-ERalpha into the nucleus. Indeed, cell fractionation studies demonstrated that Rh-ERalpha protein, in contrast to ERalpha that was nuclear at baseline, was excluded from the nucleus in the absence of hormone, and localized to the inner nuclear membrane on incubation with estrogen. These data demonstrate that membrane tethered ERalpha is capable of nuclear function and that its transcriptional activity is regulated by hormone-dependent entry into the inner nuclear membrane. Furthermore, these experiments provide evidence that under certain circumstances, membrane proteins are capable of nuclear function without detectable nucleoplasmic localization.

The binding of 3H-labelled androgen-receptor complexes to hypothalamic chromatin of neonatal mice: effect of sex and androgenization

The binding of 3H-labelled androgen-receptor complexes, prepared by (NH4)2SO4 precipitation from the 105,000 g supernatant of hypothalamic cytosol, to hypothalamic chromatin of neonatal mice covalently coupled to cellulose was measured in vitro. Saturation binding was also determined after extraction of histones and the masking of acidic proteins with high molarities of guanidine hydrochloride. This investigation showed the presence of high-affinity, low-capacity acceptor sites for [3H]-testosterone-receptor complexes in male hypothalamic chromatin (Kd value = 0.39 x 10(-10) M and binding sites of 41 fmol per mg of DNA). Acceptor activity seems to be associated with the acidic protein fraction of chromatin. No specific acceptor sites of similar nature were found in chromatin taken from the hypothalami of female mice. On the basis of these results, it is suggested that the androgen-unresponsiveness of female mice is related to the absence of acceptors for the androgen-receptor in female mice hypothalami.

The effect of triiodothyronine on the association of the rat apolipoprotein A-I, C-III and A-IV genes with the nuclear matrix

The association of apolipoprotein genes with the nuclear matrix of rat liver has been studied. The proportion of the apolipoprotein (apo) A-I, A-IV, C-III and E genes and the albumin gene associated with the matrix in normal rat liver was approximately 11%. In hypothyroid rats, the association of the apo A-I and apo A-IV genes with the matrix was reduced to 1.3% and 2.1% respectively, while in hyperthyroid rats, the association increased to 59% and 39% respectively. In contrast, the association of the apo C-III gene, which is located between the A-I and A-IV genes on chromosome 11, was not significantly affected by thyroid status. The association of the apo E and albumin genes with the matrix was also unaltered by changes in thyroid status. The increased association of the apo A-I and A-IV genes with the nuclear matrix of hyperthyroid animals accompanies the transcriptional activation of these genes by thyroid hormones.

Estradiol receptor-nuclear interactions in aging mouse uteri: the role of DNA and nuclear matrix

We have studied the interactions of estradiol receptor (ER) with nuclear DNA or matrix (NM) in the uteri from young, middle-aged and old mice. Recombination studies using heat-activated ER and nuclear subfractions from various age groups suggested a 30% reduction for DNA binding (P less than 0.05) but not for NM with aging. Cross-incubation studies of heat-activated ER with nuclear subfractions from mice of all three age subgroups showed that this reduced DNA binding ability followed the onset of anestrous and did not result from proteolytic degradation of ER complexes.

Nuclear acceptor sites for estrogen-receptor complexes in the liver of the turtle, Chrysemys picta. I. Sexual differences, species specificity and hormonal dependency

Hepatic estrogen receptors (ERs) of the female turtle, Chrysemys picta, when complexed with [3H]estradiol ([3H]E2), were shown to bind specifically to liver chromatin isolated from the same species. The binding of the [3H]E2 receptor complex to chromatin requires both the steroid ligand and the receptor protein. Maximal binding occurred within 60-70 min of incubation at 4 degrees C in a Tris buffer containing 0.1 M KCl. The binding of the [3H]E2 receptor complex to intact chromatin was saturable, whereas the binding to turtle or calf thymus DNA remained linear. Scatchard analyses revealed more estrogen receptor binding sites on hepatic chromatin isolated from female turtles than that prepared from the males (binding capacities: female chromatin = 67.9 +/- 6.8 fmol/mg DNA equivalent; male chromatin = 28.5 +/- 2.5 fmol/mg DNA equivalent). Furthermore, the [3H]E2 receptor complex was bound with a higher affinity to female chromatin than to male chromatin (association constants: female chromatin = 11.7 +/- 2.7 X 10(10) M-1; male chromatin = 2.5 +/- 0.7 X 10(10) M-1). In contrast to turtle hepatic [3H]E2 receptors, ERs in rat liver or mouse uterine cytosol exhibited little binding affinity for hepatic chromatin isolated from the turtle. Tissue specificity was demonstrated in the interaction of the [3H]E2 receptor complex and chromatin; high affinity, saturable binding of the [3H]E2 receptor complex was only observed on chromatin isolated from the liver but not on those prepared from the heart, kidney and muscle. A 3- to 4-fold increase in the number of hepatic chromatin [3H]E2 receptor binding sites was observed in 21-day ovariectomized or hypophysectomized female (capacities = 209.3 +/- 6.1 and 270 +/- 10.1 fmol/mg DNA equivalent, respectively). It is postulated that [3H]E2 receptor binding sites on the chromatin of intact females are partially 'masked', and removal of a gonadal and/or pituitary factor(s) unveils additional binding sites on the female chromatin. This paper is first to report the presence of high affinity, species- and tissue-specific acceptor sites on the liver chromatin of a reptilian species. The fact that the levels and properties of these acceptor sites are dependent on the sex and hormonal state of the animal suggests that they may play a role in the regulation of hepatic estrogen responsiveness and vitellogenesis in this species.

Binding of the androgen receptor to the nuclear matrix of human foreskin

The nuclear matrix (NM) is a salt and nuclease-resistant nuclear substructure. It is associated with active DNA transcription and has been shown to contain acceptor sites for steroid receptors in a number of specific target tissues. We have investigated the presence of acceptor sites for the androgen receptor (AR) in the NM of human newborn foreskin. The NM was prepared from the 800 g pellet by successive treatments with detergent, DNase and high salt extraction. It contained 13 +/- 7% of total proteins and 10 +/- 6% of total DNA. After extensive washing, the NM spheres were incubated in the presence of cytosol and [3H]methyltrienolone +/- 200-fold excess of unlabeled steroid. Maximal binding of the AR to NM was reached in 30 min and decreased slightly thereafter to reach an equilibrium which was maintained for 18 h. Binding was saturable. In the absence of AR, the steroid did not bind to NM. When Scatchard analysis was performed on cytosol previously incubated with NM, cytosolic binding capacity significantly decreased relative to preincubation values (3.6 +/- 1.9 to 1.3 +/- 1.2 fmol/mg protein, P less than 0.05, n = 6). In contrast, apparent binding affinity was not changed. 0.8 mg of NM protein could bind AR from 2.4 mg of cytosol protein. In conclusion, NM from human foreskin binds the AR with high affinity. This binding is rapid and is maintained for at least 18 h. This is consistent with a potential role of NM in the mechanism of action of androgens in their target tissues.

Multiple thyroid hormone binding sites on male rat liver nuclear matrices

Equilibrium binding of T3 to nuclear matrices isolated from male rat liver occurred after incubation for 3h at 20 degrees C. Two binding sites, having KD's of 6 and 95 nM, were revealed by Scatchard analysis. T3 and Triac competed for the binding of [125I]T3 to the high affinity site whereas only T3 competed for binding to the lower affinity site. Reverse T3 (rT3) did not compete for the binding of T3 to either class of binding sites. The binding sites were highly DNAse-sensitive, and less sensitive to protease treatment. The effect of binding of T3 to nuclear matrices by ATP, DTT and EDTA indicated that the sites are dissimilar to previously identified cytosolic binding sites. The higher affinity site resembles the T3 receptor in affinity and thyroid hormone specificity. The second site represents a new class of thyroid hormone binding sites. Its role in the regulation of thyroid hormone action warrants further investigation.

Steroid hormone receptor localization in the nuclear matrix: interaction with acceptor sites

The nuclear matrix is a conceptually attractive candidate for the site in the nucleus where steroid hormone-receptor complexes might interact to modulate DNA structure and function. We have demonstrated that in sex steroid target tissues a major proportion (50-100%) of the high affinity and steroid-specific receptors that become associated with the nucleus following hormonal stimulation are localized in the nuclear matrix. Direct cell-free binding assays confirm that this localization is due to the presence of specific acceptor sites in the matrix to which steroid-receptor complexes bind with high affinity and tissue specificity, and is not the result of spurious binding. The nuclear matrix appears to be a major site of hormone receptor binding in the nucleus, and this situation is consistent with the known ability of steroid hormones to stimulate gene transcription, a process which also appears to occur in association with the nuclear matrix.

Structure and function of the glucocorticoid receptor

The glucocorticoid receptor protein is a transcriptional regulatory protein that interacts with specific enhancer sequences. A stoichiometric analysis of the interaction indicates an equimolar relationship between the receptor protein and the specific enhancer sequences. The activity of the receptor protein is itself regulated by the binding of glucocorticoids. The two functional domains (DNA-binding and steroid-binding) are adjacent and lie within the C-terminal half of the receptor protein. The N-terminal half of the protein appears to modulate the interaction with DNA but does not have any role in the binding of the steroid. The functional domains have also been defined at the sequence level.

The association of acute phase protein genes with the nuclear matrix of rat liver during experimental inflammation

The association of acute phase protein genes with the nuclear matrix in livers from healthy rats and rats suffering from inflammation was studied. alpha 1-Acid glycoprotein and transthyretin are synthesized at low levels in normal liver and no matrix association of their genes was observed. Albumin, transferrin and the beta-chain of fibrinogen are synthesized at much higher levels in normal liver and their genes were found to be associated with the nuclear matrix. An inflammation induced increase in synthesis of alpha 1-acid glycoprotein and the beta-chain of fibrinogen resulted in stronger matrix association of their genes. However, inflammation induced decrease in the synthesis of albumin did not influence matrix association of its gene.

The nuclear matrix is the site of glucocorticoid receptor complex action in the nucleus

Binding of highly purified glucocorticoid receptor complexes to nuclear matrix was evaluated. Extraction of purified nuclei with 2M potassium chloride and brief deoxyribonuclease digestion leaves a matrix structure containing 1% of nuclear DNA and 6-12% of nuclear proteins. The nuclear matrix retained two binding sites for receptor complexes, a high affinity, low capacity site and a low affinity, high capacity site. These sites have affinities and capacities consistent with those reported for binding of these complexes to intact nuclei. More extensive deoxyribonuclease treatment of the matrix resulted in a marked reduction of high affinity complex binding. Furthermore, the DNA binding form of the receptor complex but not the unactivated receptor complex bound to DNA fibers anchored to nuclear matrix as visualized by 18 nm gold particle receptor complexes. The data suggest that the nuclear matrix is the major site for coordinating glucocorticoid hormone action in the nucleus.

Interaction of the calf uterine estrogen receptor with acceptor sites in heterologous chicken target cell nuclei

Estrogen receptor (ER) from chicken liver and calf uterus were used to study the capacity and the characteristics of the receptor binding sites (acceptor sites) in chicken target cell nuclei. Binding studies were performed at a physiological salt concentration of 0.15 M KCl. Binding of liver ER to liver nuclei was temperature-dependent, showing a 9-fold increase between 0 and 28 degrees C. The maximal number of acceptor sites measured in this cell-free system (280 sites/nucleus) was considerably lower than measured in nuclei after in vivo administration of estrogen (820 sites/nucleus). Moreover incubation of nuclei with the liver ER preparation resulted in a substantial breakdown of nuclear DNA, making this ER less suitable for DNA binding studies. The temperature-activated calf uterine receptor bound to liver nuclei at 0 degrees C, at which temperature no DNA degradation was measured. To all chicken cell nuclei tested, the receptor bound with a high affinity (Kd = 0.4-1.0 nM). Nuclear binding displayed tissue specificity: oviduct greater than heart, liver greater than spleen greater than erythrocytes and was salt dependent. Calf uterine ER binding in liver nuclei ranged from 3000-6000 acceptor sites per nucleus when assayed under conditions of a constant protein or a constant DNA concentration. Nuclei isolated from estrogen-treated cockerels bound a 2-fold lower number of calf uterine ER complexes when compared to control nuclei. Incubation of nuclei with a fixed concentration of [3H]ER from liver and increasing concentrations of uterine non-radioactive-ER also resulted in a reduced binding of the liver receptor. Both types of experiments suggest that liver and uterine ER compete for a common nuclear acceptor site. Our data demonstrate that the ER from calf uterus is very useful as a probe to examine the nature of the acceptor sites in heterologous chicken target cell nuclei. The assay system functions at 0 degrees C, a temperature at which no DNA degradation occurs.

Correlations between prostate chromatin structure and transcriptional activity and acceptor site distribution

Androgen-receptor complexes are intimately involved in the maintenance of rat ventral prostate chromatin in a transcriptionally active structure. The presence or absence of androgens influenced the quantity of transcriptionally active chromatin and the distribution of acceptor sites for androgen-receptor complexes as probed by endonucleolytic cleavage. After castration, changes in the sedimentation rates of nucleosome oligomers were consistent with the absence of androgen-receptor complexes and elongating polyribonucleotide chains. These changes were accompanied by decreases in the ability of chromatin released under conditions of minimal nuclease digestion to bind androgen-receptor complexes and to support incorporation of RNA precursors responsive to androgenic stimulation. Saturation analyses of chromatin and nuclear fractions with partially purified androgen-receptor complexes revealed two affinity classes of acceptor sites. After castration, alterations in the intrachromatin distribution of acceptor sites were consistent with their redeployment into areas of decreased nuclease sensitivity, as previously shown for androgen-responsive genes.

Binding of the glucocorticoid receptor complex to the nucleosomal core in the P1798 mouse lymphosarcoma

Binding of the glucocorticoid receptor complex to nucleosomes has been studied using the mouse P1798 lymphosarcoma. Cells were incubated with [3H]triamcinolone acetonide (TA), and nuclei prepared and digested with 3 different concentrations of micrococcal nuclease. After fractionation with EDTA and NaCl, it was observed that [3H]TA bound with similar specific radioactivity to mononucleosomes containing both core and linker DNA, of 183 +/- 5, and 168 +/- 4 base pair lengths, respectively, as well as to core size DNA, of 148 +/- 3 base pair length, suggesting that the glucocorticoid receptor bound to the core portion of the nucleosome. Steroid binding was found to be associated with regions of the nucleosome that were depleted in histone H1 and enriched in high mobility group (HMG) proteins 1 and 2; only negligible binding was noted in nucleosomes enriched in histone H1 and depleted in HMG proteins. In addition to binding to core nucleosomes, the glucocorticoid receptor complex was also shown to bind to a fraction sedimenting at 5-6 S on sucrose gradients characterized by subnucleosome and mononucleosome size DNA, as well as by core histones. While binding of the steroid receptor complex to linker regions of the nucleosome cannot be ruled out, this data would appear to present the first concrete evidence that glucocorticoid binding, at least in the P1798 lymphosarcoma, is to core nucleosomes. Some caution in interpretation of the results is indicated, however, on 2 points: (1) receptor redistribution during nuclease digestion cannot be ruled out; (2) only the binding of a small proportion of the steroid receptor complex may be physiologically relevant.

A subset of non-histone nuclear proteins reversibly stabilized by the sulfhydryl cross-linking reagent tetrathionate. Polypeptides of the internal nuclear matrix

When rat liver nuclei are isolated in the presence of the irreversible sulfhydryl-blocking reagent iodoacetamide, digested with DNase I and RNase A, and extracted with 1.6 M NaCl, nuclear envelope (NE) spheres depleted of intranuclear material, as analysed by thin-section electron microscopy, are obtained. Two-dimensional isoelectric focusing (IEF)/SDS-PAGE and non-equilibrium pH gradient electrophoresis (NEPHGE)/SDS-PAGE reveal that the predominant polypeptides are lamins A, B and C. Nuclei isolated in the absence of sulfhydryl blocking reagents yield salt- and nuclease-resistant structures which contain sparse but demonstrable intranuclear material. A number of non-histone polypeptides are seen in addition to the lamins. Nuclei treated with the sulfhydryl cross-linking reagent sodium tetrathionate (NaTT) yield, after exposure to nucleases and 1.6 M NaCl, nuclear matrix-like structures containing an extensive intranuclear network and components of the nucleolus in addition to the NE. Increased amounts of the non-lamin, non-histone polypeptides are recovered with these structures. Subsequent treatment of these NaTT-cross-linked structures with reducing agents in 1.0 M NaCl selectively solubilizes the intranuclear components but leaves the nuclear envelope apparently intact. The lamins remain sedimentable and are virtually absent from the soluble (intranuclear) material. Instead, the major solubilized polypeptides are (a) 68 and 63 kD polypeptides which migrate in the vicinity of lamins B and C, respectively, but are distinguishable from the lamins by immunoblotting and by uni-dimensional peptide mapping; (b) a series of basic 60-70 kD polypeptides (pI greater than 8.0) which are not recognized by anti-lamin antisera; (c) an acidic (pI 5.3) 38 kD polypeptide; and (d) a number of high molecular mass (greater than 100 kD) polypeptides. These observations not only suggest a convenient method for fractionating matrix structures from rat liver nuclei into biochemically and morphologically discrete components, but also identify a subset of major non-lamin, non-histone nuclear polypeptides (comprising approx. 20% of the total nuclear protein) whose intermolecular interactions can be reversibly stabilized apparently by intermolecular disulfide bond formation by NaTT.

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.