Skip to main content
ACS Pharmacology & Translational Science logoLink to ACS Pharmacology & Translational Science
. 2025 Jul 9;8(8):2526–2535. doi: 10.1021/acsptsci.5c00117

Benzestrol Isomer Stereochemistry Determines the Distinct Estrogenic Activities and Conformations of the Eight Isomers When Bound to Estrogen Receptor α

Yingwei Hou †, Jerome C Nwachukwu ‡, Kathryn Carlson †, Sung Hoon Kim †, Samantha Angle §, Eric N Jacobsen §, Toby Woods ∥, Sofia L Mateus ⊥, Sean W Fanning ⊥, Yvonne Ziegler #, Benita S Katzenellenbogen #, Kendall W Nettles ‡, John A Katzenellenbogen †,*
PMCID: PMC12340639  PMID: 40810159

Abstract

The nonsteroidal estrogen, benzestrol, has potent estrogenic activity, and through a recent stereocontrolled synthesis, we have obtained all eight of its constituent stereoisomers. We find that only one of them, RSS benzestrol, has very high binding affinity for estrogen receptor alpha (ERα); the other seven isomers have 60 to 600-fold lower affinity. We now show that the potencies of the isomers in two cell activity assays, proliferation of ER-positive breast cancer cells and stimulation of estrogenic gene activity, reflect their varying binding affinities for ERα. The crystal structure of the RSS isomer itself is consistent with its presumed absolute configuration and also reveals its conformational flexibility within the solid crystal lattice. We modeled each of the 8 benzestrol stereoisomers bound to ERα. Their calculated binding energies and internal torsional energies grouped with their experimentally measured binding affinities and biological activities, and the conformation for the highest affinity RSS isomer bound to ERα maps closely onto the conformation of the ERα-bound potent nonsteroidal estrogen, trans-diethylstilbestrol. Hence, we now provide a structural context for this congeneric series of benzestrol stereoisomers by proposing energy-based conformations they adopt when bound to ERα that underlie their effectiveness as estrogens.

Keywords: estrogen receptor, benzestrol isomers, congeneric ligands, nonsteroidal estrogens, conformationally flexible estrogens, computer modeling


graphic file with name pt5c00117_0012.jpg


graphic file with name pt5c00117_0009.jpg


The estrogen receptor alpha (ERα; gene name ESR1) controls important functions in its many target tissues, both reproductive and nonreproductive, including bone, skeletal muscle, and liver. ERα is also able to bind a wide range of compounds of diverse structure, and this binding has been termed “pervasive” or “promiscuous, but the term “eclectic” is more accurate, because ERα can distinguish with exquisite specificity between ligands of relatively similar structure, even ones that are flexible and differ only in stereochemistry.

Benzestrol, a potent, bisphenolic nonsteroidal estrogen first reported in the 1940s, has three adjacent sp3 carbon stereocenters linking the two phenols through their para positions. Early synthetic efforts in this nonsteroidal estrogen series explored various substitution patterns with H, Me, Et, and n-Pr groups on the adjacent carbon atoms, and found the highest estrogenic activity with the Me, Et, Et pattern. − Because the substituents on the three linking carbon atoms can have R or S absolute configurations, this congeneric series has 8 distinct stereoisomers as 4 diastereomeric pairs of enantiomers (Figure ). The synthetic methods available at the time of its discovery were not enantioselective, but by laborious fractional crystallization four distinct racemic isomers were isolated. By quantitative in vivo assays (Allen–Doisy tests), potent estrogenic biological activity was found in only one of these racemates; however, neither the relative nor absolute stereochemistry of any of these four isomers was known. The most active racemate, designated the B-2 isomer based on the fractionation scheme, was termed “benzestrol”. , Although the name benzestrol refers only to the B-2 isomer, the name “benzestrol isomers” refers collectively to all 8 isomers.

1.

1

Systematic stereochemical layout of the 8 isomers of benzestrol with their shorthand R,S stereochemical descriptors, which are used as their names in all figures and tables. To facilitate identification of the single high activity isomer, RSS (at the lower left), it is shown here and in other multicomponent figures and tables in enlarged, boldface font. The four compounds in the left column are in the 2R series and in the right column in the 2S series. The boldface uppercase letters A–D between the columns identify the four pairs of enantiomers. The numbers in square brackets are relative binding affinity values for ERα (with estradiol = 100). The numbers in parentheses were those used during the description of their synthesis in the recently published companion paper to facilitate cross-referencing between the two publications.

As reported recently by us, all eight benzestrol stereoisomers were obtained in a stereoselective way using a recently developed catalytic version of the Matteson rearrangement reaction, providing the first opportunity to study the activity of all eight of these stereoisomers. Here, we report the distinctive character of the benzestrol stereoisomers in terms of their binding affinities to ERα and their potencies and efficacies in stimulating estrogen-driven cell proliferation and gene activation. With only single bonds linking the two phenols through the three contiguous sp3 hybridized carbon atoms, the benzestrol isomers have many rotational degrees of freedom and hence great conformational mobility. We provide information on two conformations found in a small molecule X-ray structure of the highest affinity RSS isomer, and we use quantitative computational molecular modeling to identify the most likely conformations of all 8 benzestrol stereoisomers when bound to ERα. These reveal that the RSS isomer bound with the lowest calculated binding and internal torsional energies to provide a structural context for this congeneric series of benzestrol stereoisomers that underlies their varying estrogenic activities.

Results and Discussion

A Stereochemistry-Based Nomenclature Shorthand

To simplify stereoisomeric comparisons throughout this report, we named each benzestrol isomer uniquely by their stereochemical designations of R or S (but omitting their locant numbers). For example, the compound below, which is 2R,3S,4S-benzestrol, is given the shorthand name RSS.graphic file with name pt5c00117_0010.jpg

In Figure , the four isomers for one enantiomeric series (2R) are listed in the left column and for the other series (2S) in the right column. In this way, the four enantiomeric pairs, designated A–D, are adjacent to one another, left and right, in each of the four rows of this figure. To facilitate cross-referencing between the compound names in this paper and the designations given to them in the recently published chemical synthesis paper, we also include (in Figure and Table only) the numerical designations given to these compounds in the chemistry paper in parentheses (e.g., RSS, (19)).

1. Binding Affinity of the Benzestrol Isomers to Estrogen Receptor α.

C2,3,4 Cpd # RBA RBA Ratio [RSS/###] Enantiomeric pairs A–D
RRR ent-22 0.313 ± 0.01 495 A
RRS ent-16 2.41 ± 0.63 64 B
RSR 13 0.244 ± 0.07 635 C
RSS 19 155.0 ± 27 (1) D
SSS 22 2.03 ± 1.02 76 A
SSR 16 0.323 ± 0.06 480 B
SRS ent-13 1.66 ± 0.17 93 C
SRR ent-1 2.41 ± 0.65 64 D
DES   158 ± 25 0.98  
E2   100 1.55  
a

Compound number used in the chemical synthesis paper.

b

Relative binding affinities from 2 or more assays (E2 = 100; K d for E2 is 200 pM).

c

The binding affinity ratios with the high affinity isomer RSS.

d

The designation of enantiomeric pairs A–D refers to the four sets of paired structures in Figure .

Biological Characterization of the 8 Benzestrol Stereoisomers

ERα Binding Affinity

The ERα binding affinities of the benzestrol isomers were measured in a competitive radiometric assay, and they are expressed as Relative Binding Affinities (RBAs), which are referenced to that of estradiol (E2), with the RBA of E2 considered to be 100 (for representative binding curves, see Supporting Figure S1). The RBA values for each isomer are posted in Figure and listed in Table . This table is organized so that the effects of the absolute configurations at each of the three stereocenters on binding affinity are evident.

Notably, only 1 isomer, RSS benzestrol, has a very high binding affinity for ERα, and it is higher than that of estradiol and similar to that of trans-diethylstilbestrol (DES). Any change in its configuration causes a major reduction in its binding affinity (Table ). The binding of the high affinity enantiomer, RSS, is most reduced (ca. 500–600 fold) by inverting a single configuration at C-4 RSR, or double configurations that include C-4 (i.e., C-4 and C-2 SSR, or C-4 and C-3 RRR). There is a lesser effect on binding affinity (ca. 60–90 fold) from inverting configurations at C-2 SSS or at C-3 RRS, both C-2 and C-3 SRS, or all three positions, which produces the RSS enantiomer SRR. Looking at the enantiomer pairs (Table rightmost column), the three pairs A, B, and C have RBA values that differ by only 6.5–7.5 fold, whereas the enantiomer pair D, which includes RSS, have RBA values that differ by 64 fold, highlighting the singularly high affinity of the RSS stereoisomer.

Cellular Activity

We examined the estrogenic potency and efficacy of the benzestrol isomers and estradiol in two well-established assays of cellular activity. These are stimulation of the proliferation of ER-positive MCF7 human breast cancer cells (Figure A) and activation of an estrogen-responsive luciferase reporter gene (3xERE-Luc) (Figure B).

2.

2

(A) Dose–response curves and EC50 values in nM for the stimulation of MCF7 cell proliferation by E2 and the benzestrol isomers. Cells grown in media with charcoal-stripped fetal bovine serum were treated with the indicated concentrations of compounds, and after 5 days, cell numbers were compared by the CellTiter-Glo assay. (B) Dose–response curves and EC50 values in nM for transcriptional activation of a luciferase reporter gene (3xERE-Luc) by E2 and the benzestrol isomers at 24 h. Panels A and B are arranged with the 2R benzestrol isomers on the top and the 2S isomers on the bottom, following the order in the left and right columns in Figure . The potency (EC50 in nM) values for E2 and the benzestrol isomers are shown here, and they are also listed in Supporting Table S1 together with their binding affinities (K i in nM). Each data point represents the mean ± SD (n = 2 independent experiments × 4 wells).

When studied over a wide concentration range (10–15 to 10–6 M) in both assays, all compounds showed nice sigmoidal activity curves that reached full or nearly full activity, but they had a very broad spread of potencies (>25,000 fold), with EC50 values ranging from 4 pM–108 nM. Only estradiol and the highest-binding affinity RSS isomer showed very high potencies; with all of the others being markedly lower. As an additional check that the benzestrols can activate genes that are normally stimulated by estrogens, we assayed all of the isomers at 1 μM, and they all profiled as estrogen agonists in the transcriptional up-regulation of two well-known endogenous estrogen-regulated genes, GREB1 and TFF1/pS2 (Supporting Figure S2).

Relationships between Binding Affinities and Cellular Activities

To examine how closely the potencies (EC50 values) of the benzestrol isomers in the two cell-based assays track with their ERα binding affinities (K i values; listed in Supporting Table S1), we prepared two scatter plots of their binding (Log K i) and activity (Log EC50) values (Figure A,B). In both cases, only one benzestrol isomer, RSS, is close to E2 in having very high binding affinity and cellular potency. The other 7 isomers cluster in an area having values 2–3 log units lower in binding affinities and showing even greater reduction (4–5 logs) in cell stimulatory potencies from that for RSS benzestrol. Thus, only one of the 8 benzestrol stereoisomers (RSS) has high binding affinity (K i = 0.129 nM), comparable to that of E2 (K i = 0.2 nM), and it displays the highest potency in proliferation assays [(EC50 = 0.0096 nM), comparable to that of E2 (EC50 = 0.0032 nM)], and also in reporter gene activation [(EC50 = 0.0038 nM), comparable to that of E2 (EC50 = 0.015 nM)].

3.

3

Scatter plots of benzestrol isomer binding affinity (K i) (Supporting Table S1) (A) vs MCF7 cell proliferative potency (EC50) (Figure A), and (B) vs reporter gene activity potency (Log EC50) (Figure B).

The K i value for ligand binding to ERα represents a relatively simple physical interaction, whereas the EC50 value for ligand potency in cell activity, particularly proliferation, results from a multifaceted signal transduction pathway involving multiple protein interactions that occur after ligand binding. These other interactions affect the detailed relationship between ERα ligand binding affinities and potency in cell responses. , It is notable that the range of the proliferative potencies of the benzestrol isomers (>25,000) is 40-fold greater than the range of their binding affinities (ca. 600). Also, because the RSS isomer plays such a dominant role in these “1 + 7” scatter plots, detailed correlations among the other isomers are not warranted.

The results of these assays highlight the question of how the stereochemical features of the RSS isomer stabilize a conformation that uniquely supports these high binding and activity levels. And, of equal importance, how the conformers of the other seven isomers are differentiated by their low affinities and low potencies. Our efforts to address these questions by structural studies and computational modeling are given below.

Structural Studies

Small Molecule Crystal Structure of the Highest Affinity Benzestrol Stereoisomer RSS (19)

The absolute and relative configuration of the high affinity benzestrol stereoisomer was presumed to be RSS, based on the known absolute configuration of the initial α-chloro boronic ester-derived stereocenter and by assuming that the chiral catalyst exerted consistent absolute stereocontrol in the generation of each subsequent stereocenter, and these were consistent with the resultant chiral HPLC chromatograms and spectral data. We confirmed the absolute stereoconfiguration as RSS from the disordered single-crystal X-ray structure as depicted in Figure . The unit cell for this molecule is unusual by containing two RSS molecules in rather different conformations and orientations (Figure A): One RSS molecule is well ordered in a fixed position (RSS-1, gray), but the other molecule, which is oriented differently than RSS-1, appears to be the superposition of two molecules of RSS. One (RSS-3, blue) has the same conformation as that of the lone molecule (RSS-1, gray), but the other (RSS-2, green), with its long axis flipped over by 180°, has a different conformation so that its overall contour better matches that of RSS-3 in this flipped orientation (see Figure B,C for identification of RSS-1, RSS-2, and RSS-3).

4.

4

(A) A thermal ellipsoid model of the full unit cell showing the three RSS molecules colored separately: RSS-1 ( gray ), and the two overlapping ones, RSS-2 ( green ) and RSS-3 ( blue ). To give a clearer view, hydrogens are only shown on carbons 2, 3, and 4. Note how closely the two overlaid conformations of the differently oriented RSS-2 and RSS-3 molecules result in overlap of the phenolic hydroxy groups (red atoms) and rings. (B) The unit cell (63.7% existence) showing only 2 molecules, RSS-1 ( gray ) and RSS-2 ( green ). (C) The unit cell (36.3% existence) showing only 2 molecules, RSS-1 ( gray ) and RSS-3 ( blue ). In RSS-1 ( gray ), the green ellipsoid near the methyl of the C-3 ethyl substituent indicates that this methyl group can have slightly different orientations in (B and C).

Because of this, the unit cell shown in Figure A is actually the superposition of two different populations of unit cells, each with only two molecules. The one with RSS-1 ( gray ) and RSS-2 ( green ) (Figure B) has a 63.70(1)% population, and the other with RSS-1 ( gray-green ) and RSS-3 ( blue ) (Figure C) has a 36.30(1)% population. Once extracted, one sees that regardless of their different orientations, the stereochemistries of all three molecules in the structure (RSS-1, 2, 3) are consistent with the expected 2R,3S,4S configuration (Figure B,C). The detailed crystal data and refinement information are in Supporting Table S2.

Looking more closely at the region of the unit cell that is co-occupied by two RSS structures, it is seen that the long axes of the molecules are not only flipped by 180°, but the two RSS molecules also have different conformations. From the complete list of C2–C3 and C3–C4 dihedral angles of the three RSS molecules (Table ), we can see that of the three conformations, RSS-1 ( gray ) and RSS-3 ( blue ) have nearly identical dihedral angles (Figure C), each with an antiperiplanar dihedral between C2–C3 and a (+)-synclinal dihedral between C3–C4. By contrast, in RSS-2 ( green ) the locations of the two torsional angles are reversed (Figure B): C2–C3 is antiperiplanar and C3–C4 is synclinal, but in this case the second torsion angle is (−)-synclinal (Table ). These torsional angle differences enable RSS-2 and RSS-1 to occupy nearly the same space in the unit cell when they are oriented in opposite directions, with the two different phenol rings and the substituents on the three-carbon backbone situated in nearly the same space (Figure A). The dihedral angle differences of the poses of RSS suggest that beyond inherent torsional energy considerations (discussed later), different poses of RSS can provide significant populations for the formation of the crystal lattice.

2. Summary of Dihedral Angles and Internal Torsional Energy of 3 RSS Benzestrol Conformers in X-ray Crystal Structures and the Conformer from Modeling RSS-modeled .
  Dihedral Angles
 
  (ArC)-C2–C3-(C4) (C2)-C3–C4-(CAr) Internal Torsional Energy
RSS-1 ( gray ) Antiperiplanar −178.2° (+)-Synclinal 60.7° 34.551 kcal/mol
RSS-2 ( green ) (−)-Synclinal −68.5° Antiperiplanar 169.2° 26.716 kcal/mol
RSS-3 ( blue ) Antiperiplanar −178.7° (+)-Synclinal 70.4° 28.892 kcal/mol
RSS-modeled (+)-Synclinal 71.0° Antiperiplanar −169.0° 41.332 kcal/mol
a

We have defined the dihedral bond orientation using as reference points the larger of the two substituents on each of the two carbon atoms, as shown in parentheses in the headers in the two columns below the Dihedral Angles label.

b

The internal torsional binding energy is the torsional binding energy of the initial conformation from X-ray structural analysis or glide docking, minus the torsional binding energy after molecular mechanics minimization. It is discussed later (see Table ).

c

The fourth entry, RSS-modeled, comes from the modeling section, which is presented later.

The conformational flexibility of RSS benzestrol gives it an unusual almost “chameleon-like character”, whereby through adjustments of its internal torsional angles, one molecule can closely mimic the overall shape of same molecule when oriented in the opposite direction, much like how a chameleon can blend into different environments by adjusting its color. This adjustment of the shape feature of all the benzestrol isomers is further examined by computational modeling below.

Computational Modeling

We tried to obtain crystal structures of the benzestrol isomers when bound to the ERα ligand binding domain (ER-LBD). Despite findings from a thermal shift assay that RSS benzestrol was binding to the ER-LBD (see Supporting Figure S3), we were unable to produce crystals suitable to obtain a structure, perhaps because of the high degree of conformational freedom that was evident from the small molecule crystal structure of RSS benzestrol (Figure ). As an alternative approach, we used computational modeling of the interaction of the 8 benzestrol stereoisomers with the ERα-LBD to predict the conformations they adopt when bound to ERα and to understand what favorable or unfavorable features of the energetics of its interaction are likely to be responsible for their very different binding affinities and estrogenic activities.

Modeling of Benzestrol Isomers in the Context of Related Non-Steroidal Estrogens

Benzestrol is a member of the class of bisphenolic nonsteroidal estrogens having two phenols connected through their para positions by a single chain with a variable number of sp3 or sp2 hybridized carbon atoms. , These include DES and hexestrol. The ERα binding affinity and estrogenic potency of these other estrogenic bisphenols are greatly affected by the hybridization, substitution, and stereochemistry of their linking carbon atoms, as shown in Figure .

5.

5

Examples of simple bisphenolic nonsteroidal estrogens related to benzestrols and their ERα binding affinities [RBAs, E2 = 100]. (trans and cis-DES binding is in mouse ERα and the value for trans-DES differs somewhat from the RBA for trans-DES we have obtained using human ERα (Table )). For benzestrol, we show only the high activity RSS isomer and one low activity isomer RSR.

Best known are the bisphenols, the cis and trans-diethylstilbestrols (DES), which are linked by two sp2-hybridized carbons, each bearing an ethyl group. trans-DES has very high binding affinity (RBA 286) and potency, and its two phenols match rather closely the location of the A- and D-rings of estradiol in crystal structures with ERα-LBD (PDB IDs: 3ERD and 1ERE); cis-DES binds much less well (RBA 0.8). , The saturated analogs of the stilbestrols, the hexestrols, exist as a higher binding meso-stereoisomer (RBA 300) and two lower binding dl-stereoisomers with similar low affinities (RBA 3.2). In an earlier computational modeling study with only the ligand itself, we found that the meso isomer has a preferred antiperiplanar conformation, which places its phenols nicely congruent with those of the high affinity trans-DES and is more stable than both synclinal conformers. By contrast, the dl isomers prefer enantiomeric synclinal conformations, which bear greater resemblance to the low-affinity cis-DES. Presumably, the lower affinities of the dl-hexestrols reflect the torsional energy penalties need for them to reach the antiperiplanar conformation, thereby mimicking trans-DES, which is preferred for binding to ERα.

Hexestrol has symmetry and only a one principal rotatable bond between its two intervening sp3-hybridized carbon atoms; so, it was relatively simple to relate the different torsional energy profiles of the hexestrol isomers to their conformational preferences and biological activities. By contrast, benzestrol has two principal rotatable bonds (between C2–C3 and C3–C4) and an unsymmetrical pattern of substitution (one methyl and two ethyl groups), which makes it much more complicated to link isomer stereochemistry with conformation, receptor binding, and biological activity. Nevertheless, stereoisomerism clearly plays a defining role, as only one of the 8 stereoisomers (RSS) has very high binding affinity and biological activity (Tables , S1). Furthermore, when extended, RSS benzestrol is overall longer than trans-DES; so, it is neither intuitively evident how RSS benzestrol is readily able to adopt a trans-DES-like conformation suitable for high affinity ER binding nor what this conformation might be. Similarly, it is not clear what conformational features of the other 7 isomers preclude their high affinity binding.

Computer Modeling of Benzestrol Isomer Fit into the ERα Ligand Binding Pocket (ER-LBP)

We undertook comparative computer modeling of the 8 benzestrol stereoisomers bound within the ERα-ligand binding pocket, and assessed how well their calculated binding energies match their experimentally measured binding affinities and to what degree their conformations and the spacing of their phenols match the fit of the known crystallographic structure of the high affinity bisphenolic nonsteroidal estrogen, trans-DES, when bound to the ERα-LBD (PDB ID: 4ZN7). For this work, we used Schrödinger Maestro’s glide docking feature, which compares multiple ligand binding poses using an iterative minimization scheme to improve the final predicted fit. In this and prior studies, we found our experimentally measured binding affinities aligned well with glide docking’s MM-GBSA dG Bind score (“calculated binding energy”, CBE). Because of various approximations, the CBE does not correspond directly to a true free energy of binding, but it provides a valuable index for making comparisons among the relative binding affinities of compounds with similar size and structure when bound to a protein. It works well with the benzestrol stereoisomers, which have differences only in stereochemistry.

After glide docking of the benzestrol stereoisomers in the ligand-binding pocket from the ERα LBD structure with trans-DES (PDB ID: 4ZN7), we selected the three conformations of each stereoisomer having the lowest CBEs (Table ). As an internal benchmark for our approach, we used the same glide dock strategy to obtain CBEs for meso- and dl-hexestrol stereoisomers we had studied before. The scatterplot of these data show a clear separation of the high affinity meso-hexestraol and RSS compounds from the other compounds, supporting the glide docking of the RSS stereoisomer.

3. Calculated Binding Energies (CBEs) of the Benzestrol (and Hexestrol) Stereoisomers from Glide Docking Using the a Chain of the trans-DES-ER-LBD Co-Crystal Structure (PDB ID 4ZN7).
    Glide Docking Calculated Binding Energy (CBE, kcal/mol)
Benzestrol isomer Binding (Log K i) A B C AVG
RRR –7.20 –74.61 –65.50 –58.80 –66.30
RRS –8.15 –77.72 –70.11 –62.84 –70.22
RSR –7.17 –71.84 –66.19 –52.32 –63.45
RSS –9.94 –83.13 –76.02 –75.66 –78.27
SSS –8.23 –75.80 –75.42 –66.28 –72.50
SSR –7.26 –62.38 –61.74 –58.74 –60.95
SRS –7.95 –74.01 –70.57 –69.58 –71.39
SRR –8.16 –68.32 –62.08   –65.20
meso -Hexestrol –10.18 –79.64 –79.63 –79.60 –79.62
dl-Hexestrol –8.20 –71.58 –71.62 –71.53 –71.58
a

Only two reasonable conformations were found by glide docking.

We examined in greater detail how similar to one another the three conformations predicted by glide docking were for the highest affinity RSS benzestrol isomer and how well they matched the conformation of trans-DES when bound to the ERα-LBD (Figure ). Although the three RSS benzestrols bound by ERα-LBD have somewhat different CBEs (Table ), the three conformations overlap very neatly, with little observable differences (e.g., only minor alterations in the C4-ethyl group torsional angle, Figure a). What is notable about this conformation is that the benzestrol backbone is not fully extended but bends back on itself due to a (+)-synclinal dihedral conformation of the C2–C3 bond, the C3–C4 torsional remaining antiperiplanar. By bending in this manner, the distance between the two phenols in this conformation is reduced, thereby enabling the two phenolic hydroxyl groups of the RSS benzestrol model to interact with two key residues in the ERα LBP, His524 and Glu353, the same two phenols with which trans-DES binds (PDB ID: 4ZN7) (Figure b) and the ones frequently involved in stabilizing the binding of estrogens (e.g., estradiol (PDB ID: 1ERE).

7.

7

(a) Overlay of the three nearly identical RSS benzestrol modeled conformations (teal, cyan, and pink) having the lowest calculated binding energies (CBEs) from glide docking in the ERα ligand binding pocket (ERα-LBP) of the ERα-LBD complex with DES (PDB ID: 4ZN7). (b) Overlay of a composite of the RSS lowest CBE conformers (green) with DES (white) in this crystal structure.

A closer look at the RSS benzestrol model with the DES ER crystal structure (Figure b) shows that while the C-4 phenol from RSS benzestrol is neatly aligned with one DES phenol, with both ligands maintaining a good interaction with His524, the engagements of the second phenols with Glu353 are different. DES is fully extended when its second phenol is engaging with Glu353, whereas RSS benzestrol bends by using a synclinal twist at the C2–C3, a conformational adjustment that brings the second, C-2 phenol closer to the second phenol of DES. The second phenol in RSS benzestrol is angled and oriented somewhat differently than the second phenol in DES; however, the Glu353 residue can engage in a productive hydrogen bonding interaction with the second RSS phenol simply by a small 0.74 Å shift in the position of its carboxyl residue by minor adjustments in the torsional angles of its side chain. It is notable that this second ring in both DES and RSS benzestrol, which corresponds to the A-ring of estradiol, preserves a T-shape pi-stacking interaction with the Phe404 residue that is seen in ER crystal structures with many phenolic ligands and is considered to provide a positive contribution to the ligand binding energy. −

There is an interesting contrast between this RSS-modeled structure and the conformations of the RSS-1, 2, and 3 structures obtained from the small molecule X-ray structure determination (Figure and Table ). Like the RSS-modeled here, the conformation of the RSS-2 from the X-ray analysis also has a similar pattern of synclinal and antiperiplanar dihedral angles between C2–C3, and C3–C4, respectively, but in the RSS-modeled conformation it is a (+)-synclinal torsion whereas it is a (−)-synclinal torsion in RSS-2 in the X-ray structure. These differences give rather different overall molecular shapes for RSS-2 and RSS-modeled, differences that likely reflect the consequences of being imbedded in a crystal lattice (X-ray, RSS-2) vs being modeled within the confines of the ligand-binding pocket in the Y537S ERα-LBD (RSS-modeled).

We have attempted to explore further whether this apparently high-affinity conformation for the RSS isomer is shared or disfavored with the other seven low-affinity isomers. (For reference, the poses of all eight benzestrol stereoisomers after glide docking overlaid on trans-DES are shown in Supporting Figure S4). The dihedral angles of all eight benzestrol isomers are shown in Table . Only the SRR stereoisomer, the enantiomer of the high affinity RSS isomer, shares the (+)-synclinal-antiperiplanar conformation of the backbone, although its CBE is less than that of the RSS isomer (Table , Figure ), differences presumably due to their differential interactions with the surrounding residues of the LBP.

4. Ligand Binding Affinity, Key Dihedral Angles, Calculated Binding Energy (CBE), and Torsional Energy Strain of Benzestrol (and Hexestrol) Isomers When in the Conformation Bound to ERα-LBD.
Stereochemistry Binding (Log K i) (ArC)- C2–C3 -(C4) Dihedral Angle (C2)- C3–C4 -(CAr) Dihedral Angle CBE (kcal/mol) from Glide Docking Ligand torsional strain (kcal/mol) when bound
RRR –7.20 antiperiplanar 153.3° antiperiplanar −166.2° –66.30 66.31
RRS –8.15 (−)-anticlinal −116.3° (−)-synclinal −82.5° –70.22 64.00
RSR –7.17 (+)-synclinal 75.2° (+)-anticlinal 136.3° –63.45 72.29
RSS –9.94 (+)-synclinal 71.0° antiperiplanar −169.0° –78.27 52.68
SSS –8.23 (+)-anticlinal 140.9° (−)-synclinal −55.3° –72.50 69.92
SSR –7.26 (−)-synclinal –77.6° (−)-anticlinal –122.2° –60.95 69.75
SRS –7.95 antiperiplanar –164.7° (+)-synclinal 62.6° –71.39 66.40
SRR –8.16 (+)-synclinal 37.7° antiperiplanar –170.7° –65.20 57.60
  C(Ar)-C3–C4-C(Ar) Dihedral Angle  
meso -Hexestrol d –10.18 (−)-antiperiplanar –174.3° –79.62 32.38
dl-Hexestrol d –8.20 (−)-antiperiplanar –165.3° –71.58 39.46
a

The assignment of torsional angle names are best estimates and sometimes deviate somewhat from the ideal angles (synclinal: (±)­60°; anticlinal: (±)­120°; antiperiplanar: 180°).

b

The average CBE of three (or two) bound conformations after glide docking (see Table ).

c

The torsional energy of the bound conformation of each ligand extracted from the glide docking of the benzestrol isomers in DES ERα LBD cocrystal structure (PDB ID: 4ZN7), evaluated using Chem3D MM2 without energy minimization. (Because of its smaller size, the hexestrol isomer torsional energy values are on a very different scale than the benzestrol isomers and are omitted from the plot in Figure .).

6.

6

Binding affinity (Log K i) of the 8 benzestrol stereoisomers vs their average calculated binding energy (CBE) obtained using glide docking with MM-GBSA binding energy minimization (filled dots). Values for the hexestrol isomers (meso and dl) are shown as stars.

In our earlier study of the hexestrols, as well as our reanalysis here (Table ), the lower affinity of the dl-isomers appears to arise from internal torsional strain required for these isomers to rotate the central bond to reach an antiperiplanar conformation of the two phenols required for high affinity binding. In our glide docking of the 8 benzestrol stereoisomers, we imagine that isomers will develop varying degrees of internal torsional strain energy as they reach suitable conformations best suited for binding to the ERα-LBD. To assess such internal torsional strain, we extracted the bound conformations with the lowest CBE from each ligand–receptor complex, freeing it from ligand-protein interactions. Then, without further minimization, we determined its internal torsional strain in the bound conformation by a simple molecular mechanics analysis (Chem3D MM2). The results are given in Table , demonstrating that the high affinity RSS stereoisomer has low internal torsional energy relative to the other isomers (Figure ). Hence, the benzestrol stereoisomers, like the hexestrol stereoisomers, have the lowest CBE when they adopt a conformation that matches well with contours and steric constraints of the ligand binding pocket while also maintaining low internal torsional energy. The relationship between the calculated binding energies values and internal torsional energies among the various isomers, nevertheless, is very similar. In each case there is about a 4-fold difference between changes in torsional energies and log K i values.

8.

8

Plot of Log K i vs internal torsional energy in the conformation bound to the ERα-LBD crystal structure with DES (PDB ID: 4ZN7), calculated using Chem3D MM2 without energy minimization (Table ). (Because of their smaller sizes, the torsional energies of the hexestrol isomers are not commensurate with those of the benzestrols and hence are omitted from this plot.)

The best conformation of the RSS stereoisomer, obtained from the glide docking (cf., Figure a), shows substantially greater calculated binding energy and has less internal torsional strain energy than the other 7 isomers (Table ). It also can nicely overlay onto the pose of DES bound to ERα-LBD (Figure b). We consider the structures shown in Figure to be good representations for the actual ERα-bound conformation of the high affinity RSS benzestrol stereoisomer.

Conclusion

By studying all 8 stereoisomers of benzestrol, we found that only one, RSS benzestrol, has very high ERα binding affinity, with the remaining seven having 60–600 fold lower affinities. Their estrogenic potencies in two cell activity assays reflect the wide range of differences in their ERα binding affinities. The crystal structure of RSS itself revealed that it has conformational flexibility, enabling it to adopt within the solid crystal lattice two different favorable poses, one of which was similar to the structure docked into the receptor. Using computer modeling, we explored how the 8 stereoisomers might best bind to ER, and the bound conformation for the high affinity RSS isomer maps nicely onto the conformation of the potent nonsteroidal estrogen trans-diethylstilbestrol when bound to ERα. Therefore, we now provide both a biological characterization of the 8 benzestrol stereoisomers, as well as a structural context for this congeneric series of benzestrol stereoisomers by proposing energy-based conformations they adopt when bound to ERα that underlie their varying effectiveness as estrogens and highlight the conformation that the high affinity benzestrol RSS stereoisomer is likely to adopt that is responsible for its high estrogenic activity.

Methods

Biological Methods

ERα Binding Affinity

Relative binding affinities (RBA) were determined by a competitive radiometric ligand binding assay with 1 nM full-length, purified human ERα (Thermo Fisher Scientific, cat. no. A15674), and 2 nM [3H]­E2 (Revvity NET317) as tracer. , Incubations were at 0 °C for 18–24 h. Hydroxyapatite was used to adsorb the receptor–ligand complex, and unbound ligand was washed away. The determination of the RBA values is reproducible in independent experiments with a CV of 0.3, and the values shown represent the average ± range or SD of two or more separate determinations.

Cell Culture Methods

MCF7 and HEK-293T cells were maintained in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) (Millipore Sigma, cat. no. F0926) at 37 °C in a 5% CO2 incubator. Cell lines were cultured with 1× penicillin-streptomycin-neomycin antibiotics mixture (Thermo Fisher Scientific, cat. no. 15640055), 1× MEM nonessential amino acids (Thermo Fisher Scientific, cat no. 11140-050), 1× GlutaMAX (Gibco by Thermo Fisher Scientific, cat. no. 35050061), and 2.5 μg/mL plasmocin (InvivoGen, cat. no. ant-mpp).

Cell Proliferation Assay

MCF7 cells were suspended in steroid-free media supplemented with 10% charcoal-stripped FBS and placed in 384-well plates. Cells were treated with compounds and assessed for cell growth 5 days later using CellTiterGlo (Promega), as previously described.

Luciferase Reporter Gene Assay

Steroid-deprived 293T cells in 10 cm dishes were transfected with 5.0 μg of 3xERE-luciferase reporter and 0.5 μg of ERα expression plasmids using FuGENE HD reagent (Promega, cat no. E2311). The next day, the cells were transferred to 384-well plates and returned to the incubator. The following day, the cells were treated with compounds for 24 h. Luciferase activity was measured using the Britelite plus reporter gene assay system (PerkinElmer, cat no. 6066761).

RNA Isolation and Real-Time PCR for Gene Transcription Analysis

Total RNA was isolated using RNeasy kit (QIAGEN, cat no. 74106), and reverse transcribed using the MultiScribe enzyme (Thermo Fisher Scientific, cat no. 4311235). Real-time PCR was performed using the TaqMan Fast Advanced Master Mix (Thermo Fisher Scientific, cat no. 4444557), and normalization of mRNA levels of GREB1 and TIFF1/pS2 genes were done as described. Results are the average ± SD from at least two independent experiments carried out in triplicate.

Structural Methods

Single Crystal Formation of RSS Benzestrol

2.5 mg of RSS benzestrol was dissolved in 100 μL of a 10:1 ethanol-to-deuterated chloroform mixture in a 3 mL vial. This vial was then placed inside a larger 24 mL vial containing 300 μL of deuterated chloroform. The solvent was permitted to diffuse and evaporate slowly over 1 day, resulting in the formation of a single crystal that was used for the X-ray analysis.

Thermal Shift Assay and Methods

We used a thermal shift assay to determine whether the benzestrol compounds bound to the receptor ligand binding domain (ERα-LBD). This assay measures the temperature dependence of protein unfolding with a dye, sypro orange, which selectively fluoresces when hydrophobic surfaces are exposed upon thermal protein denaturation; small molecules that bind to a protein stabilize it and increase the denaturation temperature (T M). , We incubated ERα-LBD with individual benzestrol isomers and measured the T M compared to DMSO (vehicle negative control) and 17β-estradiol (E2, positive control). Only the high-binding isomer RSS benzestrol had any impact on the melting temperature (T M) (for details see Supporting Information and Supporting Figure S3).

Computational Modeling

Ligand docking models were generated using the LigPrep package in Schrödinger Maestro 13.5 (https://www.schrodinger.com/platform/products/maestro/). The PDB file of ERα-LBD (Y537S) in complex with DES (PDB ID: 4ZN7) was selected as the docking structure, after choosing Chain A of the dimer and filling in missing side chain and loops. We used the Glide module to identify the DES ligand A-ring phenol with Glu353 interaction and D-ring phenol with His524 interaction, and then docked the benzestrol ligands with the same H-bond constraints. During docking, the van der Waals radius scaling factor was set as 0.80 and the partial charge cutoff as 0.15, and we followed standard precision in flexible ligand sampling, using at most 3 poses per ligand. For potential binding energy calculations of the complexes, we used the MM-GBSA function in Prime module with the solvation model: VSGB, force field: OPLS4, protein flexibility: 0.0 Å distance from ligand.

Supplementary Material

pt5c00117_si_001.pdf (452KB, pdf)

Acknowledgments

We are grateful to Dr. Christopher G. Mayne for his helpful comments.

Glossary

Abbreviations

CBE

calculated binding energy

DES

diethylstilbestrol

E2

estradiol

ER

estrogen receptor

ERα

estrogen receptor alpha

ERβ

estrogen receptor beta

LBD

ligand binding domain

LBP

ligand binding pocket

RBA

relative binding affinity.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsptsci.5c00117.

  • Methods for RSS benzestrol crystallization and X-ray structure determination, ERα-LBD thermal shift assay; tables of ERα binding affinities, potencies of proliferation and reporter gene assays and crystallography data; figures of representative competitive binding assay curves, transcriptional activity of estrogen-regulated genes; thermal shift assay curves, minimum energy conformations of all 8 benzestrol stereoisomers superimposed on DES (PDF)

Y.H., E.N.J., B.S.K., J.A.K., K.W.N., J.C.N. designed research; S.R.A., K.E.C., S.W.F., Y.H., J.A.K., S.H.K., J.C.N., T.W., Y.Z. performed research; S.H.K., S.L.M., J.C.N. contributed new reagents/analytic tools; S.W.F., Y.H., E.N.J., B.S.K., J.A.K., S.H.K., K.W.N., J.C.N. analyzed data; and S.R.A., K.E.C., S.W.F., Y.H., E.N.J., B.S.K., J.A.K., S.H.K., K.W.N. J.C.N.,T.W. wrote the paper. All authors reviewed and approved the paper.

This work was supported by grants from the NIH (R01CA275142 to K.W.N. and R01CA220284 to K.W.N., B.S.K., and J.A.K. and R37CA279341 to S.W.F.), and MIRA GM149244 to E.N.J.) and the Breast Cancer Research Foundation (BCRF-083 to B.S.K. and BCRF-084 to J.A.K. and B.S.K.).

The authors declare no competing financial interest.

References

  1. Katzenellenbogen J. A.. Commentary: The structural pervasiveness of estrogenic activity. Environ. Health Perspect. 1995;103(suppl 7):99–101. doi: 10.2307/3432517. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Nettles K. W., Bruning J. B., Gil G., O’Neill E. E., Nowak J., Hughs A., Kim Y., DeSombre E. R., Dilis R., Hanson R. N.. et al. Structural plasticity in the oestrogen receptor ligand-binding domain. EMBO Rep. 2007;8(6):563–568. doi: 10.1038/sj.embor.7400963. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Blanchard E., Stuart A., Tallman R.. Studies on a new series of synthetic estrogenic substances. Endocrinol. 1943;32(4):307–309. doi: 10.1210/endo-32-4-307. [DOI] [Google Scholar]
  4. Grundy J.. Artificial estrogens. Chem. Rev. 1957;57(2):281–415. doi: 10.1021/cr50014a002. [DOI] [Google Scholar]
  5. Shukis A., Ritter J. J.. A New Synthesis of Benzestrol1. J. Am. Chem. Soc. 1950;72(4):1488–1489. doi: 10.1021/ja01160a017. [DOI] [Google Scholar]
  6. Stuart A. H., Shukis A. J., Tallman R. C., McCann C., Treves G. R.. Synthetic Estrogenic Compounds. III. Trialkyl Derivatives of 1, 3-Di-(p-hydroxyphenyl)-propane. Benzestrol. J. Am. Chem. Soc. 1946;68(5):729–733. doi: 10.1021/ja01209a001. [DOI] [PubMed] [Google Scholar]
  7. Angle S. R., Sharma H. A., Choi C. K., Carlson K. E., Hou Y., Nwachukwu J. C., Kim S. H., Katzenellenbogen B. S., Nettles K. W., Katzenellenbogen J. A.. Iterative Catalyst-Controlled Diastereoselective Matteson Homologations Enable the Selective Synthesis of Benzestrol Isomers. J. Am. Chem. Soc. 2024;146(45):30771–30777. doi: 10.1021/jacs.4c12857. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Carlson K. E., Choi I., Gee A., Katzenellenbogen B. S., Katzenellenbogen J. A.. Altered ligand binding properties and enhanced stability of a constitutively active estrogen receptor: evidence that an open pocket conformation is required for ligand interaction. Biochem. 1997;36(48):14897–14905. doi: 10.1021/bi971746l. [DOI] [PubMed] [Google Scholar]
  9. Min C. K., Nwachukwu J. C., Hou Y., Russo R. J., Papa A., Min J., Peng R., Kim S. H., Ziegler Y., Rangarajan E. S.. et al. Asymmetric allostery in estrogen receptor-α homodimers drives responses to the ensemble of estrogens in the hormonal milieu. Proc. Natl. Acad. Sci. U. S. A. 2024;121(24):e2321344121. doi: 10.1073/pnas.2321344121. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Li Y., Coons L. A., Houtman R., Carlson K. E., Martin T. A., Mayne C. G., Melchers D., Jefferson T. B., Ramsey J. T., Katzenellenbogen J. A.. et al. A mutant form of ERα associated with estrogen insensitivity affects the coupling between ligand binding and coactivator recruitment. Sci. Signal. 2020;13(650):eaaw4653. doi: 10.1126/scisignal.aaw4653. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Dodge, J. A. ; Jones, C. D. . Non-steroidal Estrogens. In Estrogens and Antiestrogens I: Physiology and Mechanisms of Action of Estrogens and Antiestrogens; Springer: 1999, pp. 43–53. [Google Scholar]
  12. Sadler B. R., Cho S. J., Ishaq K. S., Chae K., Korach K. S.. Three-dimensional quantitative structure– activity relationship study of nonsteroidal estrogen receptor ligands using the comparative molecular field analysis/cross-validated r 2-guided region selection approach. J. Med. Chem. 1998;41(13):2261–2267. doi: 10.1021/jm9705521. [DOI] [PubMed] [Google Scholar]
  13. Korach K. S.. Biochemical and estrogenic activity of some diethylstilbestrol metabolites and analogs in the mouse uterus. Adv Exp Med Biol. 1982;138:39–62. doi: 10.1007/978-1-4615-7192-6_3. [DOI] [PubMed] [Google Scholar]
  14. Kilbourn M. R., Arduengo A., Park J. T., Katzenellenbogen J. A.. Conformational analysis of nonsteroidal estrogens: The effect of conformer populations on the binding affinity of meso-and dl-hexestrol to the estrogen receptor. Mol. Pharmacol. 1981;19(3):388–398. doi: 10.1016/S0026-895X(25)13698-5. [DOI] [PubMed] [Google Scholar]
  15. Kingston B., Pearson A., Herrera-Abreu M. T., Sim L.-X., Cutts R. J., Shah H., Moretti L., Kilburn L. S., Johnson H., Macpherson I. R.. et al. ESR1 F404 Mutations and Acquired Resistance to Fulvestrant in ESR1-Mutant Breast Cancer. Cancer Discov. 2023;14(2):274–289. doi: 10.1158/2159-8290.CD-22-1387. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Lee S., Barron M. G.. Structure-based understanding of binding affinity and mode of estrogen receptor α agonists and antagonists. PLoS One. 2017;12(1):e0169607. doi: 10.1371/journal.pone.0169607. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Pike A. C., Brzozowski A. M., Hubbard R. E.. A structural biologist’s view of the oestrogen receptor. J. Steroid Biochem. Mol. Biol. 2000;74(5):261–268. doi: 10.1016/S0960-0760(00)00102-3. [DOI] [PubMed] [Google Scholar]
  18. Katzenellenbogen J. A., Myers H. N., Johnson H. J. Jr. Reagents for photoaffinity labeling of estrogen binding proteins. Synthesis of some azide and diazo derivatives of estradiol, estrone, and hexestrol. J. Org. Chem. 1973;38(20):3525–3533. doi: 10.1021/jo00960a020. [DOI] [PubMed] [Google Scholar]
  19. Norris J. D., Paige L. A., Christensen D. J., Chang C.-Y., Huacani M. R., Fan D., Hamilton P. T., Fowlkes D. M., McDonnell D. P.. Peptide antagonists of the human estrogen receptor. Science. 1999;285(5428):744–746. doi: 10.1126/science.285.5428.744. [DOI] [PubMed] [Google Scholar]
  20. Hancock G., Young K., Hosfield D., Joiner C., Sullivan E., Yildiz Y., Lainé M., Greene G., Fanning S.. Unconventional isoquinoline-based SERMs elicit fulvestrant-like transcriptional programs in ER+ breast cancer cells. Npj Breast Cancer. 2022;8(1):130. doi: 10.1038/s41523-022-00497-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Bai N., Roder H., Dickson A., Karanicolas J.. Isothermal analysis of ThermoFluor data can readily provide quantitative binding affinities. Sci. Rep. 2019;9(1):2650. doi: 10.1038/s41598-018-37072-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Gee A. C., Katzenellenbogen J. A.. Probing conformational changes in the estrogen receptor: evidence for a partially unfolded intermediate facilitating ligand binding and release. Mol. Endocrinol. 2001;15(3):421–428. doi: 10.1210/mend.15.3.0602. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

pt5c00117_si_001.pdf (452KB, pdf)

Articles from ACS Pharmacology & Translational Science are provided here courtesy of American Chemical Society

RESOURCES