2025 Volume 73 Issue 4 Pages 412-418
Estrogen receptors (ERs) and their ligands regulate a variety of physiological processes in humans, and altered ER signaling is associated with serious disorders, including breast cancer. Estrogens also bind to other receptors such as G-protein-coupled receptor 30 (GPER), and so fluorescent estrogen ligands would be useful for various functional studies and for development of drug candidates. Here we describe fluorescent estrogen receptor ligands based on 3-aryl-7-hydroxycoumarin. Notably, these ligands also function as pH-dependent OFF-ON-OFF type fluorescent sensors, enabling the detection of specific ranges of pH.
Estrogens, including estrone, estradiol, and estriol, are sex hormones, whose binding to estrogen receptors (ERs) induces specific gene transcriptional activities.1–3) Altered ER signaling is associated with serious disorders, including breast cancer and its recurrence, and selective estrogen receptor modulators and down-regulators have been developed as therapeutic agents4–7) (Fig. 1). Estrogens also bind to other receptors, such as G-protein-coupled receptor 30 (GPER), which is expressed in various tissues, including heart, intestine, and the central nervous system (CNS).8–10) Therefore, it is important to understand the physiological roles of these receptors and the molecular mechanisms involved. For this purpose, selective ligands have been developed.11–15) In particular, fluorescent ligands would be useful to visualize the localization of the receptors by fluorescence microscopy. Fluorescent ligands consisting of an estradiol derivative with a fluorophore at the 17α position have been used to visualize intracellular ER and cell membrane-bound GPER.16–18) However, these ligands are relatively large, and the chemical properties of the linker and fluorophore may influence the localization and degradation of the receptors after ligand binding. On the contrary, ligands whose core structure is the fluorophore itself might overcome these limitations. For example, fluorescent false neurotransmitters (FFNs), which incorporate a coumarin fluorophore in a neurotransmitter structure, have been employed to visualize activity-dependent heterogeneity in dopamine release at the level of individual synapses.19,20) Coumesterol, which has a coumarin core structure, could be regarded as a similar type of molecule (Fig. 1). In this paper, we describe the development of fluorescent estrogen derivatives based on coumarin, and we show that they can also serve as sensors of pH change.
Estrogen receptor ligands include a pharmacophore containing two hydroxy groups, and several non-steroidal compounds such as diethylstilbestrol and raloxifene have been developed, in addition to coumarin-based compounds.21–24) Our group has constructed a library of coumarin derivatives utilizing Suzuki–Miyaura coupling and Husigen 1,3-cycloaddition reactions,25,26) and discovered a fluorescent progesterone receptor ligand (1),27,28) as well as various fluorescent sensors, including a ratiometric viscosity sensor (2)29) that exhibits two fluorescent peaks in low-viscosity solution, but only one in high-viscosity solution (Fig. 2).
From this library, we also discovered compounds 3 and 4, containing two hydroxy groups, as photofunctional compounds.30,31) Compound 3 shows an OFF-ON-OFF change of fluorescence intensity with increasing pH, and so could be utilized as a fluorescent sensor for a specific range of pH. This function is regulated by the sequential deprotonation of the two phenolic hydroxy groups, that is, the first deprotonation of the 7-hydroxy group affords a strongly fluorescent species, while the second deprotonation decreases the fluorescence. The detection range of pH could be modulated by introducing halogen substituents, which shift the pKa values of the two phenolic hydroxy groups. We found that 3a emits strong fluorescence at around pH 8–10, while 3b, a tetrachloro derivative, does so at around pH 6. Compound 4 has a photoremovable protective group, and provides selective functionality over a specific range of pH, where the monoanion form mainly exists.32)
Compounds 3 and 4 possess two hydroxy groups, but their positions and separation were thought to be unsuitable for estrogen receptor ligands, so we designed compounds 5 and 6 as new candidates (Fig. 3).
Compounds 5 and 6 were synthesized as shown in Chart 1. Suzuki–Miyaura coupling of 7 and 8 yielded compound 5. For the synthesis of 6, 9 was obtained by Pechmann condensation reaction between 2,4-dichlorooresorcinal and ethyl acetoacetate, and then bromination using N-bromosuccinimide (NBS) afforded 10. Suzuki–Miyaura coupling of 10 and 11 yielded compound 6.
The estrogen receptor binding affinities of 5 and 6 as well as coumesterol and estradiol were evaluated by fluorescence polarization assay utilizing PolarScreen ER Alpha Competitor Assay, Green (Thermo Fisher, Waltham, MA, U.S.A.; A15882). The enhanced fluorescence polarization of Fluormone ES Green bound to estrogen receptor was decreased by competitive binding of each compound. Fluorescence from Fluormone ES Green did not overlap those of 5, 6 or coumesterol (Fig. 4). Fluorescence from Fluormone ES Green did not overlap those of 5, 6 and coumesterol, respectively, so the fluorescence from the tested compound did not interfere with the measurement. The relative binding affinity of each compound to the estrogen receptor is shown as the IC50 value. Compounds 5 and 6 exhibited moderate affinity for estrogen receptor (IC50 1.2 and 64 μM, respectively), although their binding was weaker than that of estradiol (0.016 μM) or coumesterol (0.053 μM). Thus, the introduction of halogen atoms weakened the affinity.
Changes in fluorescence polarization of Fluormone ES Green bound to estrogen receptor α induced by the addition of each compound are shown as relative values.
Compounds 5 and 6 were expected to possess pH-dependent fluorescence, like compounds 3, so the absorption and fluorescent spectra of 5 at various pH values in the range of pH 3.0–12.0 were measured. As shown in Fig. 5, the absorption at around 360 nm and the fluorescence intensity at around 460 nm were increased from pH 3.0 to 9.0, while a decrease of fluorescence was observed at around pH 9.0. The two phenolic hydroxy groups are expected to be deprotonated as the pH is increased, and the first deprotonation would occur at the 7-hydroxy group, as judged from the reported acidity of the 7-hydroxycoumarin moiety33) and the correlation of the fluorescence decrease with the absorption increase at 360 nm.
Absorption spectra (a, c) and fluorescence spectra (b, d) excited at 360 nm were measured for 3 μM solutions of 5 in 10 mM sodium phosphate buffer containing 0.3% DMSO as a co-solvent at the indicated pH values.
In the case of compound 6, the absorption at around 360 nm and the fluorescence intensity at around 460 nm changed similarly, but the pH at which the changes occurred was shifted to a more acidic range, from around pH 4.0–6.0 (Fig. 6). The pH at which the decrease of fluorescence occurred was also shifted to a more acidic region, at around pH 7.0. These shifts presumably reflect the change in pKa due to the introduction of chloride and fluoride at the ortho positions of the two phenolic hydroxy groups. The pKa values of the two hydroxy groups estimated from the absorption and fluorescence intensity changes. First deprotonation at 7-hydroxy group of coumarin moiety showed characteristic absorbance at around 360 nm, so its pKa was obtained from this absorption and pH (Supplementary Materials). pKa of the hydroxy group other than 7-hydroxy one was determined by fluorescence intensity. The estimated pKa values of 5 and 6 appear to be 7.8 and 4.5 for the 7-hydroxy group, respectively, and 9.9 and 6.6 for the hydroxy group on the 3-phenyl group, respectively.
Absorption spectra (a, c) and fluorescence spectra (b, d) excited at 360 nm were measured for 3 μM solutions of 6 in 10 mM sodium phosphate buffer containing 0.3% DMSO as a co-solvent at the indicated pH values.
As described above, these compounds show an OFF-ON-OFF type change of fluorescence intensity with change in pH, similarly to coumesterol (Figs. 7a–7c). The range of pH within which the fluorescence intensity was strong depended on the introduction of halogen atoms, and compound 6 with four halogen atoms emitted fluorescence at around pH 6, similar to our previously developed compound 3b with four chlorine atoms at the same positions. These fluorescence changes are due to deprotonations of the hydroxy groups, that is, the first deprotonation at the 7-hydroxy group enhances the fluorescence, while the second deprotonation diminishes it (Fig. 7d). Thus, the two OFF states can be attributed to the appearance of neutral and dianionic forms as the pH changes. These OFF states would be due to photo-induced electron transfer (PeT) between 7-hydroxycoumarin and the phenolic moiety, as in the cases of compounds 3 and 4. In the neutral form, electron transfer from the phenolic moiety to hydroxycoumarin could occur in the excited state, resulting in non-fluorescence.34–36) In the monoanionic form generated by deprotonation of the 7-hydroxy group, electron transfer would not occur, resulting in fluorescence. In the dianionic form, the hydroxy group on the 3-phenyl group also exists in the anionic form, and this could facilitate electron transfer to the coumarin moiety by changing the oxidation potential and charge state of the phenolic moiety. In the dianionic form of 6, the fluorescence was not completely quenched, probably because the oxidation and reduction potentials of the structural component were not optimal for PeT to occur. This proposed mechanism is supported by the pKa values of the two hydroxy groups.
The fluorescence intensities were measured for 3 μM solution of each compound in 10 mM sodium phosphate buffer (pH 3.0–12.0) containing 0.3% DMSO as a co-solvent. (d) Proposed mechanism of the pH-dependent fluorescence changes in compound 5.
Although the binding affinities of compounds 5 and 6 were relatively weak compared with those of estradiol and coumesterol, the pH-dependent fluorescence of 5 and 6 would be potentially useful. The fluorescence of 5 and 6 in the presence of an estrogen receptor was relatively weak, probably because they would mainly exist in a non-deprotonated form in the ligand binding site of the estrogen receptor. Microenvironmental changes around the receptor protein could potentially be visualized; for example, during endocytosis, the pH gradually changes during transition from the cell membrane bound state (pH 7.4), to early endosome (pH 5.9–6.2) and late endosome/lysosome (pH 5.0–5.5).37–39) Recently, there has been increasing interest in the degradation processes of damaged or misfold proteins related to neurodegenerative diseases, including Huntington’s disease, Alzheimer’s disease, and amyotrophic lateral sclerosis (ALS)40–45) as well as cell membrane-bound receptor proteins, via ubiquitin-proteasome and/or lysosomal proteolytic pathways, where the microenvironment such as the pH could be changed.46,47) Our compound may be used for visualizing such processes, like other recently reported probes.48) With this in mind, we are working to develop compounds with stronger binding affinities to ER or GPER and different pH ranges of detection.
In conclusion, we have developed fluorescent estrogen receptor ligands based on 3-aryl-7-hydroxycoumarin, which also function as fluorescent sensors of pH change. Their OFF-ON-OFF mode of fluorescent change could be useful for the detection of specific ranges of pH.
All reagents were purchased from Sigma-Aldrich Co., LLC or Merck KGaA, Tokyo Chemical Industry Co., Ltd., FUJIFILM Wako Pure Chemical Corporation, and Kanto Kagaku. Silica gel for column chromatography was purchased from KANTO CHEMICAL CO., INC. NMR spectra were recorded on a Bruker Avance 400 or Bruker Avance 500 spectrometer. Mass spectral data were obtained on a Brucker Daltonics micro-TOF-2 focus or SHIMADZU LCMS-9030 in the positive and negative ion detection modes. UV spectra were recorded with JASCO V-550, and fluorescence spectra were recorded with JASCO FP-6600. In each measurement, 10 mM solutions of each compound in DMSO were prepared as the stock solution. After dilution by each pH buffer, the measurement was performed.
Preparation of Compounds Preparation of 57 (98 mg, 0.38 mmol), 8 (0.15 g, 0.70 mmol), 2M Na2CO3 (5.6 mL), and tetrakis(triphenylphosphine)palladium (23 mg, 0.020 mmol) were added to DME (7.5 mL). The reaction mixture was stirred and heated at 60 °C for 6 h under atmosphere of argon. Then it was cooled to room temperature, and saturated aqueous NH4Cl was added to it. The mixture was extracted with AcOEt. Combined organic layer was washed with brine, dried over Na2SO4, and concentrated in vacuo. The resulting crude was purified by column chromatography (silica gel, AcOEt : n-Hexane = 1 : 1) and 5 (59 mg, 58%) was obtained as a white powder. A small amount of this compound was recrystallized from MeOH to give an analytical sample. 1H-NMR: (500 MHz, DMSO-d6) δ: 7.63 (1H, d, J = 9.0 Hz), 7.07 (2H, d, J = 9.0 Hz), 6.81 – 6.78 (m, 3H), 6.07 (1H, d, J = 2.5 Hz), 2.20(s, 3H); 13C-NMR: (125 MHz, DMSO-d6) δ: 160.6, 148.1, 148.0, 143.8, 131.5, 127.1, 125.1, 123.9, 122.3, 114.8, 112.9, 112.5, 101.9,16.4; HRMS (ESI+): Calcd for C16H12O4Na [M + Na+]; 291.0628, Found; 291.0637. Anal. Calcd for C16H12O4·0.2MeOH: C, 70.84; H, 4.70. Found: C, 70.96; H, 4.93.
Preparation of 9A mixture of 2,4-dichlororesorcinal (0.20 g, 1.1 mmol) and ethyl acetoacetate (0.16 g, 1.2 mmol) in conc. H2SO4 (2.0 mL) was stirred at room temperature for 3 h. The reaction mixture was poured into iced water to give a yellow precipitate. After filtration, the residue was washed with water and dried to give 8 as a yellow solid (0.11 g, 41%). 8 was used for the next reaction without further purification. 1H-NMR: (500 MHz, DMSO-d6) δ: 7.76(1H, s), 6.26(1H, s), 2.38(3H, s).
Preparation of 10Benzoyl peroxide (8.0 mg, 0.041 mmol) was added to a solution of 9 (0.11 g, 0.47 mmol) and N-bromosuccinimide (84 mg, 0.47 mmol) in dry CH3CN (15 mL). After being stirred for 13 h at 40 °C, the mixture was filtered to obtain 10 (54 mg, 35%) as an orange powder. 10 was used for the next reaction without further purification. 1H-NMR: (500 MHz, DMSO-d6) δ: 7.90(1H, s), 2.60(3H, s).
Preparation of 610 (30 mg, 0.093 mmol), 11 (51 mg, 0.20 mmol), 2M Na2CO3 (1.5 mL), and tetrakis(triphenylphosphine)palladium (8.0 mg, 6.9 μmol) were added to DME (2.5 mL). The reaction mixture was stirred and heated at 60 °C for 3 h under atmosphere of argon. Then it was cooled to room temperature, and saturated aqueous NH4Cl was added to it. The mixture was extracted with dichloromethane : MeOH (10 : 1). Combined organic layer was washed with brine, dried over Na2SO4, and concentrated in vacuo. The residue was diluted with MeOH, and the precipitate was filtrated off. The filtrate solution was concentrated in vacuo. The residue was diluted with AcOEt, and the precipitate was filtrated off. The filtrate solution was concentrated in vacuo. The residue was diluted with DCM, and the precipitate was collected to obtain 6 (3.5 mg, 11%) as a white powder. 1H-NMR: (400 MHz, CD3OD): δ7.72 (1H, s), 6.91 (2H, dd, J = 7.2 Hz, J = 1.2 Hz), 2.31 (3H, s); 13C-NMR: (125 MHz, CD3OD) δ: 162.1, 150.4, 150.0, 133.8, 133.1, 133.0, 130.0, 129.9, 125.1, 115.0, 114.9, 114.9, 114.8, 16.9; HRMS (ESI−) Calcd for C16H8Cl2F2O4 [M−H]−; 370.9768, Found; 370.9762.
The work described in this paper was supported by Grants-in-Aid for Scientific Research from The Ministry of Education, Culture, Sports, Science and Technology, Japan (Grant No. 21K06464). TH also gratefully acknowledges funding from The Cosmetology Research Foundation. A part of this research is a Cooperative Research Project of the Research Center for Biomedical Engineering.
The authors declare no conflict of interest.
This article contains supplementary materials.