Characterization of trypsin-treated forms of the estrogen receptor from rat and lamb uterus

Bivalent ligands as probes of estrogen receptor action

The estrogen receptor (ER) is a hormone-regulated transcription factor which is thought to bind to specific DNA sequences as a homodimer. In order to better understand structural requirements for dimerization and its functional role in ER action, we synthesized a series of bivalent ligands based on the non-steroidal estrogen hexestrol. These molecular probes join two hexestrol molecules of the erythro (E, active) configuration with either 4 or 8 carbon linkers (designated E-4-E and E-8-E series, respectively), or with longer linkers comprised of ethylene glycol units (E-eg-E series). Several other bi- and monovalent control compounds were prepared. The bivalent ligands bind to ER with a relative affinity 1-7% that of estradiol. While most of the ligands demonstrated normal monophasic displacement curves in competitive binding assays with [3H]estradiol, uncharacteristic biphasic competitive binding curves were seen for some of the ligands, indicating possible structure-specific, negative site-site interaction. In ER-deficient Chinese hamster ovary (CHO) cells transfected with an expression vector encoding ER, one series of bivalent ligands (E-4-E) had little stimulatory activity and inhibited transcription stimulated by hexestrol, as determined by a transient transfection assay using an estrogen-responsive reporter gene construct [(ERE)2-TATA-CAT, containing two estrogen response elements linked to a TATA promoter and the chloramphenicol acetyl transferase reporter gene]. Monovalent or control bivalent ligands failed to antagonize hexestrol-stimulated activity and were as fully active as hexestrol itself. Studies performed in MCF-7 human breast cancer cells, which contain endogenous ER, yielded similar bioactivity profiles for the E-4-E bivalent inhibitory ligands, showing them to be effective estrogen antagonists, when using either induction of progesterone receptor or (ERE)2-TATA-CAT transcriptional activation as the endpoint. The E-8-E ligand, however, acted as a partial agonist/antagonist of ERE-reporter gene transactivation and a full agonist of progesterone receptor induction in MCF-7 cells, thus showing cell- and response-specific differences in the effects of this bivalent ligand. These bivalent ligands for ER do not show enhanced potency or receptor binding affinity; however, some of them display binding properties that suggest the possibility of structure-specific negative site-site interaction, and some of them function as quite effective estrogen antagonists.

Oxohexestrol derivatives labeled with fluorine-18. Synthesis, receptor binding and in vivo distribution of two non-steroidal estrogens as potential breast tumor imaging agents

We have prepared two non-steroidal estrogens in the 2-oxohexestrol series labeled with the positron-emitting radionuclide fluorine-18, 1-fluoro-5-oxohexestrol (4) and 1-fluoro-2-oxohexesterol (5). We anticipated that the polar ketone function at the interior of these ligands would reduce their level of non-specific binding, which might increase the selectivity of their uptake in vivo. The two compounds were prepared by total synthesis: compound 4 was prepared in fluorine-18 labeled form by [18F]fluorine ion displacement on a suitably protected methanesulfonate precursor followed by deprotection under acidic hydrogenolytic conditions; the isomer 5 was prepared from a protected alpha-keto trifluoromethanesulfonate precursor with deprotection under basic conditions as the final step. The binding affinity of these hexestrol derivatives for the estrogen receptor was determined by competitive radiometric binding assays at 0 and 25 degrees C, and their lipophilicity (as octanol-water partition coefficients, log P values) and non-specific binding were estimated. The log P values determined by a reversed phase HPLC method were higher, relative to estradiol, than those calculated by the fragment method of Rekker. In tissue distribution studies in immature (50 g) rats, both of these compounds showed selective uptake in estrogen target tissues. At 1 h, activity in the uterus reached the level of 2.5-3.0% of the injected dose per gram tissue, with uterus-to-blood and uterus-to-muscle ratios of 14-20 and 8-14, respectively. The uptake efficiency and selectivity of these fluoro-oxohexestrols in principal estrogen target tissues is less than that of fluorine-18 labeled steroidal estrogens we have prepared previously, but their receptor-mediated uptake in certain secondary target tissues is substantial. The specific and non-specific components of target tissue uptake of these two compounds appear to be directly related to their non-specific binding and their binding selectivity.

A synthesis of 7 alpha-substituted estradiols: synthesis and biological evaluation of a 7 alpha-pentyl-substituted BODIPY fluorescent conjugate and a fluorine-18-labeled 7 alpha-pentylestradiol analog

In an effort to assist in the preparation of ligands for the study of the estrogen receptor (ER), we have developed a new synthesis of 7 alpha-substituted estradiols. The key step in the synthesis involves a copper-catalyzed, alpha-selective, 1,6-conjugate addition of 4-pentenyl magnesium bromide to a suitably protected 6-dehydrotestosterone derivative. Desaturation and then reductive aromatization of the resulting 7 alpha-pentenyl androgen gave the 7 alpha-pentenylestradiol in good yields. The alpha-stereoselectivity of this addition in the testosterone series, compared with the 19-nortestosterone series, is significantly improved by the presence of the C-19 methyl group, which shields the beta face from attack. A key intermediate was functionalized further by substitution with fluorine-18 to provide a potential imaging agent for positron emission tomography, and by conjugation with a BODIPY (Molecular Probes Inc., Eugene, OR, USA) fluorophore to make a fluorescent probe for the estrogen receptor. The synthesis and biological evaluation of these analogs is presented, as well as a discussion of the improvements in the synthetic procedure.

Synthesis, radiolabeling and tissue distribution of 11 beta-fluoroalkyl- and 11 beta-fluoroalkoxy-substituted estrogens: target tissue uptake selectivity and defluorination of a homologous series of fluorine-18-labeled estrogens

We have synthesized six estrogens substituted at the 11 beta-position with a fluoroalkyl or fluoroalkoxy substituent. These compounds bind to the estrogen receptor with moderate to high affinity, with the fluoroalkyl analogs being higher affinity binders than the fluoroalkoxy ones. All of these fluorine-substituted estrogens were prepared in fluorine-18-labeled form, with high radiochemical purity and at effective specific activities (15.4-50.4 TBq/mmol; 415-1362 Ci/mmol) adequate for biodistribution studies. In immature female rats, five of the six fluoroestrogens showed selective uptake by the uterus, with uterine uptake as a percent of the injected dose per gram being 4-9% at 1 h, and uterus-to-blood or uterus-to-muscle ratios being 10-40. Selective uterine uptake was eliminated by co-administration of a blocking dose of unlabeled estradiol. The only compound that did not show selective uterine uptake was 11 beta-fluoropropoxyl estradiol; its rapid metabolism and its low affinity for the estrogen receptor, particularly at 25 degrees C, may account for its lack of specific uptake. The level of bone activity, an index of metabolic defluorination, shows that the defluorination rates of these six estrogens are a complex function of structure and functionality. Least prone to defluorination is 11 beta-(2-fluoroethoxy)estradiol and most prone is 11 beta-(2-fluoroethyl)estradiol. The extent of defluorination of the remaining compounds shows weak evidence for the protective effect of a heteroatom-substituted beta to the site of metabolism (the CH bonds on the fluorine-bearing carbon atom). The binding affinity, tissue distribution and metabolism of these 11 beta-fluoroalkyl- and fluoroalkoxy-substituted estrogens further our understanding of the behavior of fluorine-18-labeled estrogens as potential imaging agents for estrogen receptor-positive breast cancer.

Nonsteroidal estrogens bearing acyl azide functions: potential electrophilic and photoaffinity labeling agents for the estrogen receptor

In an effort to develop novel affinity labeling agents for the estrogen receptor, we have synthesized two nonsteroidal ligands, a 1-aroyl-2-aryl tetralin system (1) and a 2-aryl-3-aroylbenzo[b]thiophene system (2). These agents, patterned after the Lilly antiestrogens trioxifene and LY 117018, respectively, embody acyl azide functions as part of a benzoyl chromophore. The acyl azide group has weak acylating activity, suitable for electrophilic affinity labeling, but this function is also photoreactive and, in its particular embodiment within these ligands, it could provide an efficient photochemical route to the highly reactive singlet acyl nitrene. The tetralin system (1) was prepared in nine steps from 6-methoxy-1-tetralone, and the benzothiophene system (2) was prepared in four steps from a known substituted benzo[b]thiophene precursor. In competitive binding assays, both compounds show reasonable binding affinity for the rat and lamb uterine estrogen receptor: estradiol = 100%, 1 = 3%, and 2 = 12%. When assayed by indirect receptor consumption assays, both compounds appear to have substantial capacity for irreversible binding (electrophilic reaction) with the receptor. This reactivity, which suggests that acylation of the receptor has occurred, is photoreversible. The nature of this ligand-receptor interaction is being investigated further.

Significance of the 8S complex in oestrogen receptor recognition

Previous studies from this laboratory have drawn attention to discrepancies between enzyme-linked immunoassay (EIA) and steroid binding assay (SBA) in the analysis of oestrogen receptors (ER) in breast tumours. In particular, EIA values were at least 3-fold higher than SBA values in tumours which also contained progesterone receptors (PR) when both 4 and 8S isoforms of the ER are present. To test the influence of these isoforms on the two assay systems, the relationships between the oestrogen receptor (ER) values obtained by EIA and SBA were examined in tumour cytosols prepared in the presence of molybdate and protease inhibitors to prevent degradation of the 8S form. Under these conditions, values for ER were the same by EIA and SBA (slope = 1.08, r = 0.886, n = 25) when EIA was performed using low salt phosphate buffer instead of the high salt-containing Abbott-diluent provided with the kit. However, after disruption of the 8S assembly using high K+ concentration, the slope of the regression was 6.37, r = 0.865, n = 25. Using ER from rat uterus, EIA was also performed on intact 8S oligomers, on 8S ER dissociated by high salt, and on glycerol density gradient-fractionated 4S ER. The identity of the ER oligomers and components was confirmed by glycerol density gradient fractionation, and by isoelectric focussing. For the 4S ER, EIA gave similar values whether using low or high salt phosphate buffer. However EIA values for the 8S form were 2-fold higher when the supplied diluent was used than when the assay was performed in low salt buffer. The amount of oestradiol which could be extracted was affected by the different conditions used. Addition of KCl or trypsin to disrupt the 8S ER caused an increase in the amount of extractable oestradiol compared with control values (control = 52 +/- 4.0, high KCl = 91 +/- 4.4, trypsin = 152 +/- 7.5, pg oestradiol/mg protein). We conclude that further antibody binding sites are revealed from the 8S ER form after its disaggregation by high salt. The steroid extraction data also suggests the possibility that tightly bound steroid is retained within the 8S ER structure, and released by 8S disaggregation. Both of these may contribute to the differences between EIA and SBA values.

Efficient and selective photoaffinity labeling of the estrogen receptor using two nonsteroidal ligands that embody aryl azide or tetrafluoroaryl azide photoreactive functions

3-(4-Azido-2,3,5,6-tetrafluorobenzoyl)-6-hydroxy-2-(4- hydroxyphenyl)benzo[b]thiophene 1 (tetrafluoroaryl azide, TFAA) and its protio analogue 3-(4-azidobenzoyl)-6- hydroxy-2-(4-hydroxyphenyl)benzo[b]thiophene 2 (protioaryl azide, PAA), photoaffinity labeling (PAL) reagents for the estrogen receptor (ER), have been prepared in high specific activity tritium-labeled form (19 Ci/mmol) and shown to undergo selective and efficient photocovalent attachment to ER from rat uterus. Both azides 1 and 2 demonstrate high binding affinity for ER as determined by both a competitive binding assay (relative binding affinities: estradiol = 100; TFAA = 9.3; PAA = 66) and a direct binding assay (Kd: estradiol = 0.24 nM; TFAA = 2.64 nM; PAA = 0.37 nM). When unlabeled TFAA and PAA are irradiated at greater than 315 nm, they demonstrate site-specific photoinactivation of ER that reaches 43% and 55%, respectively, by 30 min. Specific photocovalent attachment to ER can be effected by irradiation of the tritium-labeled azides; the covalent attachment efficiency is good (1 = 20-30%, 2 = ca. 25%) and the selectivity of ER labeling is high. Characterization of the photolabeled proteins by SDS-polyacrylamide gel electrophoresis shows specific labeling of a major component at Mr 60,000 and a minor species at Mr 46,000, the same two species that are labeled by [3H]tamoxifen aziridine, a well-characterized affinity label for ER. The ER-specific antibodies H222Sp gamma and D547Sp gamma show a clean precipitation of only these two species. In the MCF-7 human breast cancer cell line, PAA is a full estrogen agonist in terms of stimulation of cell proliferation and induction of progesterone receptor. These two azides provide the first system in which the photocovalent attachment efficiency of an aryl azide can be compared to its tetrafluorosubstituted aryl azide analogue in a complex biological receptor system. Azides 1 and 2 are the most efficient and selective PAL reagents prepared to date for ER, and they should be useful in further studies of the hormone-binding domain of this protein.

Mapping on the calf estrogen receptor of the binding domain for an antibody interfering with receptor activation

The localization on the calf estrogen receptor of the binding domain for B36 (an IgM antibody which prevents and reverses the effects of receptor activation) has been studied by means of controlled proteolysis of the receptor-estradiol complex using trypsin, chymotrypsin, and papain. We successively determined for intact and proteolyzed receptor-estradiol complex (i) the abilities of estradiol-binding species to aggregate in low salt medium, to bind to nonspecific DNA absorbed onto cellulose, and to interact with B36 antibody in sucrose gradients; (ii) the hydrodynamic properties of estradiol-binding species, by gel permeation chromatography and sucrose gradient centrifugation in high salt media and (iii) the molecular weights of B36-reactive species, by immunoblot analysis. Three tryptic receptor fragments of Mr 36,000, 34,000, and 33,000 and two chymotryptic fragments of Mr 36,000 and 33,000 included both the hormone- and B36-binding domains but did not interact with DNA, whereas at least two receptor fragments resulting from the action of chymotrypsin and papain bound estradiol with high affinity but interacted neither with DNA nor with B36. Taking into account these results and assuming that structure of the calf estrogen receptor is similar to those of sequenced estrogen receptors (which show a highly conserved organization with considerable homologies in the functional domains), we propose that the B36-binding domain is located either between the DNA- and hormone-binding domains (model I) or at the C-terminal end of the estrogen receptor (model II). The regions that include the main proteolytic cleavage sites of the receptor are also specified, and the abilities of the two models of the calf estrogen receptor to account for the effect of B36 on receptor activation are discussed.

Estrogen receptor interaction with immobilized metals: differential molecular recognition of Zn2+, Cu2+ and Ni2+ and separation of receptor isoforms

We have utilized iminodiacetate (IDA) gels with immobilized Zn2+, Cu2+ and Ni2+ ions to evaluate the metal binding properties of uterine estrogen receptor proteins. Soluble (cytosol) receptors labeled with [3H]estradiol were analyzed by immobilized metal affinity chromatography (IMAC) before as well as after (1) 3 M urea-induced transformation to the DNA-binding form, and (2) limited trypsin digestion to separate the steroid- and DNA-binding domains. Imidazole (2-200 mM) affinity elution and pH-dependent (pH 7-3.6) elution techniques were both evaluated and found to resolve several receptor isoforms differentially in both the presence and absence of 3 M urea. Individual receptor forms exhibited various affinities for immobilized Zn2+, Cu2+ and Ni2+ ions, but all intact receptor forms were strongly adsorbed to each of the immobilized metals (Ni2+ greater than Cu2+ much greater than Zn2+) at neutral pH. Generally, similar results were obtained with IDA-Cu2+ and IDA-Ni2+ in the absence of urea. Receptors were tightly bound and not eluted before 100 mM imidazole or pH 3.6. Different results were obtained using IDA-Zn2+; at least four receptor isoforms were resolved on IDA-Zn2+. Receptor-metal interaction heterogeneity and affinity for IDA-Zn2+ and IDA-Cu2+, but not IDA-Ni2+, were substantially decreased in the presence of 3 M urea. The receptor isoforms identified and separated by IDA-Zn2+ chromatography were not separable using high-performance size-exclusion chromatography, density gradient centrifugation, chromatofocusing or DNA-affinity chromatography. The affinity of trypsin-generated (mero)receptor forms for each of the immobilized metals was decreased relative to that of intact receptor. High-affinity metal-binding sites were mapped to the DNA-binding domain, but at least one of the metal-binding sites is located on the steroid-binding domain. Recovery of all receptor forms from the immobilized metal ion columns was routinely above 90%. These results demonstrate the differential utility of various immobilized metals to characterize and separate individual receptor isoforms and domain structures. Receptor-metal interactions warrant further investigation to establish their effects on receptor structure/function relationships. In addition to the biological implications, recognition of estrogen receptor proteins as metal-binding proteins suggests new and potentially powerful receptor immobilization and purification regimes previously unexplored by those in this field.

DNA and ribonucleotide binding characteristics of two forms of the androgen receptor from rat prostates

Androgen receptors (sedimentation value approximately 4S and Stokes radius 2.8 nm) present in the cytoplasmic fraction obtained from prostates of castrated rats bind to DNA-Sepharose and double stranded DNA. A receptor fragment (sedimentation value approximately 3S and Stokes radius 2.3 nm) obtained from rat prostates in the course of a purification procedure showed greatly diminished binding affinity for both DNA-Sepharose and soluble DNA. In contrast, both the 4S cytosol receptor and the 3S receptor form interacted with equal affinity with prostate RNA or poly(UG). These observations provide evidence that for DNA binding a different or additional part of the receptor molecule is required than for RNA and polyribonucleotide binding.

Modulation of progestin binding activity in cultured human breast carcinoma cells: the effect of serum type and concentration

Progesterone receptor levels in MCF-7 human breast cancer cells increase as a specific response to estrogen and to some nonsteroidal antiestrogens. In the present study we demonstrate that the type and quantity of serum present during culture of these cells modifies the level of progestin binding activity, but not the level of estradiol binding activity. MCF-7 cells maintained in media supplemented with 5% charcoal-dextran treated calf serum (CDCS) contain 0.3 - 0.4 pmol of cytosol progesterone receptor (PRc) per mg DNA. When cells previously maintained in 5% CDCS-media are shifted to media containing 5% charcoal-dextran treated fetal calf serum (CDFCS), the level of progestin binding increases after day 16, and stabilizes at 2 - 3 pmol/mg DNA at days 30 to 40. Shifting these cells back to 5% CDCS-media, reduces PRc to 0.2 - 0.4 pmol/mg DNA within 3 days. This reduction is dose dependent with a half-optimal decrease at 1% CDCS, and a full decrease at 2% CDCS (4d incubation). Nuclear progestin binding was uniformly low (0.2 - 0.4 pmol/mg DNA) and unaffected by type or concentration of serum, and no consistent change in cytosol or nuclear estrogen receptor levels was observed. These cytoplasmic progestin binding sites are translocated to the nucleus by progesterone, and are similar to estradiol (E2) induced sites by Scatchard binding and sucrose gradient analysis. Similar serum-dependent changes are also observed in the T47D human breast cancer cell line where growth in CDFCS-media results in 4-fold higher progestin binding levels than observed in CDCS-media. Our findings suggest the presence of non-dialyzable stimulatory factor(s) in CDFCS that influence the progestin receptor level and highlight the fact that serum components can alter dramatically the cellular progestin binding activity.

Increase in the estrogen binding capacity of breast cancer cytosols following limited proteolysis with trypsin

When small amounts of trypsin were added to prelabelled estrogen receptors in 24 human breast cancer cytosols there was a substantial increase in the binding capacity [79 +/- 11 (SE)%]. At the same time the affinity of the hormone receptor interaction was maintained at a very high level or even increased. This finding is discussed in relation to previous results where a diisopropylfluorophosphate (DFP) inhibitable protease activity was shown to cause a similar augmentation of estrogen binding sites in human myometrial cytosols. Addition of sodiummolybdate at or immediately after homogenization led to a similar increase in estrogen binding sites. Because these two effects were not additive we propose that the limited trypsin treatment reactivates the binding sites previously inactivated through a mechanism which can be inhibited by sodiummolybdate.

Rapid analysis of estrogen and progesterone receptors using gel-exclusion high-performance liquid chromatography

Estrogen and progesterone receptors prepared from mouse, rat, and human uteri, as well as from human breast cancers, have been characterized by gel-exclusion high-performance liquid chromatography. The qualitative relationships previously established by sedimentation analysis between the cytoplasmic [aggregated (approximately 8S), deaggregated (approximately 4S), and trypsinized (approximately 3.6S)] and nuclear (approximately 5S) forms of the rat uterine estrogen receptor were maintained by this technique. Differences in the partition of estrogen and progesterone receptors from the same species as well as interspecies differences in these receptors were reproducibly observed. Multiple forms of human estrogen and progesterone receptors could clearly be resolved in a single analysis and were distinct from serum steroid binding tissue contaminants. Separation analyses, performed at flow rates up to 2 mL min-1, were capable of resolving all receptor forms in 10--12 min with the column returning to base line in 25 min. With this exclusion gel column (TSK-G3000SW) as a background upon which to reference different receptor forms, eight distinct partitions or elution positions have been enumerated. This approach has considerable promise for the rapid characterization of different forms of steroid-receptor proteins. Moreover, it should provide a critical advantage in minimizing the opportunities for receptor modification during separation analysis and in maximizing the opportunity to study short-lived interactions between receptors and physiologic or pharmacologic ligands.

Activation of mammalian cyclic AMP phosphodiesterases by trypsin

BHK fibroblasts contain two forms of cyclic AMP phosphodiesterase 3':5'-cyclic nucleotide 5'-nucleotidohydrolase EC 3.1.4.17) as analyzed by linear sucrose gradient fractionation; a 3.6-S form (peak I) and a 6.7-S form (peak II). Peak I is specific for cyclic AMP as substrate and displays Michaelis-Menten kinetics with an apparent Km of 2--3 micrometer. Peak II hydrolyzes cyclic GMP and displays anomalous kinetics for cyclic AMP hydrolysis. The activity of isolated peak II for cyclic AMP is increased by storage at 4 degrees C, treatment with trypsin, or treatment with rat brain and BHK fibroblast activator proteins. The activity of isolated peak I is unaffected by these conditions. Linear sucrose gradient fractionation demonstrates that activation of peak II by trypsin leads to the formation of a 3.6-S cyclic AMP-specific enzyme form, possibly peak I. In contrast to BHK fibroblasts (and most other mammalian tissues), rat uterus contains only one form of cyclic nucleotide phosphodiesterase on linear sucrose gradients, a 7-S form capable of hydrolyzing both cyclic AMP and cyclic GMP. Treatment of rat uterine supernatant with trypsin leads to the appearance of a 4-S, cyclic AMP-specific form with properties similar to that of BHK peak I. These data suggest that the kinetically complex, higher molecular weight cyclic nucleotide phosphodiesterases may consist of more than one catalytically active site and that multiple forms of the enzyme arise through dissociative mechanisms, possibly as a means of in vivo regulation.