New estrogen receptor antagonists. 3,20-Dihydroxy-19-norpregna-1,3,5(10)-trienes: Synthesis, molecular modeling, and biological evaluation

Article abstract of DOI:10.1016/j.ejmech.2017.11.042

New estrogen receptor α (ERα) antagonists – 3,20-dihydroxy-19-norpregna-1,3,5(10)-trienes containing an additional carbocyclic ring D′ at the 16α,17α positions – were synthesized. The effects of the new compounds on the MCF-7 breast cancer cells and ERα activation were investigated. All the steroids studied were synthesized starting from estrone methyl ether. The scaffold of compounds containing the six-membered ring D′ was constructed via the Diels–Alder reaction of butadiene with 3-methoxy-19-norpregna-1,3,5(10),16-tetraen-20-one 5. The hydrogenation of primary 16α,17α-cyclohexeno-adduct 7 followed by 3-demethylation (by HBr–AcOH) and reduction of 20-oxo group (by LiAlH₄) or in one step by DIBAH gave target mono- and dihydroxy steroids 9–11. The Corey–Chaykovsky reaction of the same 3-methoxy-19-norpregna-1,3,5(10),16-tetraen-20-one 5 gave 16α,17α-methylene-substituted compound. The reaction of the latter with DIBAH immediately yielded 3,20-dihydroxy-16α,17α-methyleno-19-norpregna-1,3,5(10)-triene 13. The same procedures using 3-methoxy-19-norpregna-1,3,5(10),16-tetraen-20-one 5 produced corresponding 3,20-dihydroxy-16,17-19-norpregna-1,3,5(10)-triene 16 and 3,20-dihydroxy-19-norpregna-1,3,5(10),16-tetraene 17. All compounds were fully characterized by 1D and 2D NMR, HRMS, and X-ray diffraction. The molecular docking showed that the target compounds can bind to ER, their binding mode being similar to that of natural estradiol. 16α,17α-Methylene- or unsubstituted compounds exhibit the highest cytotoxicity against MCF-7 cells, being simultaneously relatively weak ERα inhibitors. 3,20-Dihydroxy steroids containing the six-membered ring D′ proved to be the most effective ERα inhibitors. These compounds displayed moderate cytotoxicity comparable of that of tamoxifen and showed no toxic effect on MCF-10A normal, nontumorigenic epithelial cells. The new ER antagonists were found to be good candidates for further testing as agents for the treatment and prevention of ERα-positive breast cancers.

Full text of DOI:10.1016/j.ejmech.2017.11.042

Accepted Manuscript
New estrogen receptor antagonists. 3,20-Dihydroxy-19-norpregna-1,3,5(10)-trienes:
Synthesis, molecular modeling, and biological evaluation
Yury V. Kuznetsov, Inna S. Levina, Alexander M. Scherbakov, Olga E. Andreeva,
Irina V. Fedyushkina, Andrey S. Dmitrenok, Alexander S. Shashkov, Igor V. Zavarzin
PII:
S0223-5234(17)30945-5
10.1016/j.ejmech.2017.11.042
EJMECH 9917
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 19 July 2017
Revised Date: 30 October 2017
Accepted Date: 17 November 2017
Please cite this article as: Y.V. Kuznetsov, I.S. Levina, A.M. Scherbakov, O.E. Andreeva, I.V.
Fedyushkina, A.S. Dmitrenok, A.S. Shashkov, I.V. Zavarzin, New estrogen receptor antagonists. 3,20-
Dihydroxy-19-norpregna-1,3,5(10)-trienes: Synthesis, molecular modeling, and biological evaluation,
European Journal of Medicinal Chemistry (2017), doi: 10.1016/j.ejmech.2017.11.042.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
N
EW ESTROGEN RECEPTOR ANTAGONISTS. 3,20-DIHYDROXY-19-NORPREGNA-1,3,5(10)-
TRIENES
:
SYNTHESIS
,
MOLECULAR MODELING
,
AND BIOLOGICAL EVALUATION
a
a
b
b
Yury V. Kuznetsov *, Inna S. Levina , Alexander M. Scherbakov , Olga E. Andreeva , Irina
V. Fedyushkina c, Andrey S. Dmitrenok a, Alexander S. Shashkov a, Igor V. Zavarzin a
a
N. D. Zelinsky Institute of Organic Chemistry, Russian Academy of Sciences, Leninsky
prospect 47, Moscow 119991, Russia
b
Institute of Carcinogenesis, N.N. Blokhin Cancer Research Centre, Kashirskoye sh. 24,
Moscow 115478, Russia
c Orekhovich Institute of Biomedical Chemistry, Pogodinskaya ul. 10, Moscow, 119121 Russia
* Corresponding author. N. D. Zelinsky Institute of Organic Chemistry, RAS, Leninsky prospect
47, Moscow, 119991 Russia. Tel./Fax: +74991372944/+74991355328. E-mail address:
yukuv@mail.ru
Graphical abstract
1

ACCEPTED MANUSCRIPT
EW ESTROGEN RECEPTOR ANTAGONISTS. 3,20-DIHYDROXY-19-NORPREGNA-1,3,5(10)-
N
TRIENES: SYNTHESIS, MOLECULAR MODELING, AND BIOLOGICAL EVALUATION
Yury V. Kuznetsov a, Inna S. Levina a, Alexander M. Scherbakov b, Olga E. Andreeva b, Irina V.
Fedyushkina c, Andrey S. Dmitrenok a, Alexander S. Shashkov a, Igor V. Zavarzin a
a
N. D. Zelinsky Institute of Organic Chemistry, Russian Academy of Sciences, Leninsky
prospect 47, Moscow, 119991 Russia
b
Institute of Carcinogenesis, N.N. Blokhin Cancer Research Centre, Kashirskoye sh. 24,
Moscow 115478, Russia
c Orekhovich Institute of Biomedical Chemistry, Pogodinskaya ul. 10, Moscow, 119121 Russia
* Corresponding author. N. D. Zelinsky Institute of Organic Chemistry, RAS, Leninsky prospect
47, Moscow, 119991 Russia. Tel./Fax: +74991372944/+74991355328. E-mail address:
yukuv@mail.ru
ABSTRACT
New estrogen receptor α (ERα) antagonists – 3,20-dihydroxy-19-norpregna-1,3,5(10)-
trienes containing an additional carbocyclic ring D' at the 16α,17α positions – were synthesized.
The effects of the new compounds on the MCF-7 breast cancer cells and ERα activation were
investigated. All the steroids studied were synthesized starting from estrone methyl ether. The
scaffold of compounds containing the six-membered ring D' was constructed via the Diels–Alder
reaction of butadiene with 3-methoxy-19-norpregna-1,3,5(10),16-tetraen-20-one 5. The
hydrogenation of primary 16α,17α-cyclohexeno-adduct 7 followed by 3-demethylation (by HBr–
AcOH) and reduction of 20-oxo group (by LiAlH4) or in one step by DIBAH gave target mono-
and dihydroxy steroids 9 – 11. The Corey–Chaykovsky reaction of the same 3-methoxy-19-
norpregna-1,3,5(10),16-tetraen-20-one 5 gave 16α,17α-methylene-substituted compound. The
reaction of the latter with DIBAH immediately yielded 3,20-dihydroxy-16α,17α-methyleno-19-
norpregna-1,3,5(10)-triene 13. The same procedures using 3-methoxy-19-norpregna-
1,3,5(10),16-tetraen-20-one 5 produced corresponding 3,20-dihydroxy-16,17-19-norpregna-
1,3,5(10)-triene 16 and 3,20-dihydroxy-19-norpregna-1,3,5(10),16-tetraene 17. All compounds
2

ACCEPTED MANUSCRIPT
were fully characterized by 1D and 2D NMR, HRMS, and X-ray diffraction. The molecular
docking showed that the target compounds can bind to ER, their binding mode being similar to
that of natural estradiol. 16α,17α-Methylene- or unsubstituted compounds exhibit the highest
cytotoxicity against MCF-7 cells, being simultaneously relatively weak ERα inhibitors. 3,20-
Dihydroxy steroids containing the six-membered ring D' proved to be the most effective ERα
inhibitors. These compounds displayed moderate cytotoxicity comparable of that of tamoxifen
and showed no toxic effect on MCF-10A normal, nontumorigenic epithelial cells. The new ER
antagonists were found to be good candidates for further testing as agents for the treatment and
prevention of ERα-positive breast cancers.
Keywords: antiestrogens, estrogen receptor α, cytotoxicity, 19-norpregnatriene,
cycloaddition, NMR assignment, breast cancer.
1. Introduction
The development of effective drugs for the treatment of hormone-dependent cancer is
emerging as an important field of research in medicinal chemistry. For instance, the design of
new steroid hormone antagonists, in particular, estrogens that play a significant role in the
development of breast cancer, has received extensive attention.
Estrogens, with 17β-estradiol I (E2) being the predominant one, exert their biological
functions via estrogen receptors (ER), thereby playing an essential role in the female
reproductive system, and also control other important aspects of human physiology. Some
effects of estrogens are beneficial (protective effects against injuries in the cardiovascular and
central nervous systems, positive effects on bone health), while proliferative effects of estrogens
in some target organ tissues – targeted tissues – can be pathological (uncontrolled proliferation).
Breast cancer is one of the most common types of cancer and the most frequent cause of cancer
death in women worldwide [1].
The majority of breast malignant tumors are hormone-dependent, and the growth of (ER+)
cancer cells depends on estrogens [2].
Hence, the major strategies for hormonal therapy are to block the estrogens signaling
pathways. The mechanisms of action of hormonal drugs are based on blocking the biosynthesis
of estrogens by aromatase inhibitors [3] or by steroid sulfatase (STS) inhibitors [4], and, since
estrogens exert most of their functions via estrogen receptors (ER), - on inhibition of estrogen
3

ACCEPTED MANUSCRIPT
receptor by selective ER modulators (SERMs) [5], or degradation of estrogen receptor by
selective ER down-regulators (SERDs) [6, 7].
However, the efficacy of known steroidal and nonsteroidal antiestrogens is often limited by
drug resistance, insufficient tissue selectivity, and/or unwanted side effects [8]. Therefore, the
development of new antitumor agents for antiestrogen therapy is a challenge.
Previously, we described a new class of synthetic analogues (called progestins) of the
natural hormone progesterone containing the additional six-membered ring D' fused to the ring D
of the steroid scaffold (so-called pregna-D'-pentaranes) II. These synthetic progestins are
promising for practical application because they have high progestogenic activity in vivo and in
vitro and can inhibit the growth of cancer cells [9, 10].
The molecular docking of pregna-D'-pentaranes to the ligand-binding domain (LBD) of
progesterone receptor (PR), the main molecular target of progestins in cells, showed that, despite
a larger volume of these molecules compared to progesterone, they are well-accommodated in
the ligand-binding pocket of the receptor, the additional six-membered ring D' fitting well in the
α-region of the cavity that is present in this pocket near the ring D, thus increasing the
hydrophobic binding without distortion of the overall conformation of the ligand molecule [10].
It is well-known that ER LBD has a substantially larger volume than the endogenous ligand
and comprises several subpockets, on of which is at the 16α–17α region of the ring D of the
ligand [11, 12]. Therefore, we supposed that D'-pentaranes containing the aromatic ring A III
could bind to ER.
Based on the foregoing, we performed the synthesis and molecular modeling of 3,20-
dihydroxy-19-norpregna-1,3,5(10)-triene derivatives, which are steroids containing two hydroxyl
groups at both ends of the hydrophobic steroid nucleus characteristic of estrogens, and
investigated the effects of these compounds on breast cancer cells and estrogen receptor activity.
O
OH
OH
17
16
(CH2)n
D'
H
H
HO
HO
O
II
I
III:
n = 0, 1, 4
2. Results and discussion
2.1 Chemical synthesis
Key intermediate 5 was synthesized in several steps from estrone methyl ether 1 (Scheme
1). The reaction of the latter with trimethylsilyl cyanide in the presence of anhydrous zinc iodide
[13] gave silyl cyanohydrin 2 resulting from the attack on the less hindered α-side of the steroid
molecule by isomeric 17β-carbonitrile that is present as a minor impurity. The hydrolysis of silyl
4

ACCEPTED MANUSCRIPT
cyanohydrin 2 in aqueous hydrochloric acid in ethanol produced cyanohydrin 3 [14], which was
subjected to dehydration in the POCl3–pyridine system giving conjugated nitrile 4. There are
contradictory data on the conditions and yields for this dehydration reaction performed by
procedures described in the literature [13-18]. Thus, these procedures give unstable results. We
supposed that this reaction involves two steps – the formation of chlorophosphonic acid esters of
steroid cyanohydrin followed by their cleavage. We found that the completeness of the cleavage
of intermediate chlorophosphonic acid esters, which can be achieved by refluxing the reaction
mixture for a long period of time, is crucial for high yields of target product 4. An alternative
approach [13] via the direct conversion of silyl cyanohydrin 2 using tetrabutylammonium,
cesium, and potassium fluorides afforded conjugated nitrile 4 in low yield. The reaction of
conjugated nitrile 4 with methylmagnesium iodide [15, 16] in a toluene–diethyl ether mixture at
65°C followed by the decomposition of the resulting imine magnesium salt with acetic acid and
hydrolysis of the imine under reflux with 10% HCl produced conjugated ketone 5. In order to
obtain the latter in high yield, the temperature should be strictly maintained and the reaction
mixture should be thoroughly stirred during the decomposition of imine magnesium salt;
otherwise, the reaction gives considerable amounts of by-products. We isolated one of these by-
products and characterized it as “dimer” 6. The formation of this dimer is apparently attributed to
the reaction of the intermediate imine magnesium salt with products of its further hydrolysis
followed by the cyclization to the dihydropyridine derivative and aromatization of the latter (see
the Supplementary). When determining the structure of product 6, we also made the complete
assignment of the NMR spectra of compounds 4 and 5.
b
OTMS
CN
O
OH
NC
CN
c
d
a
O
O
O
O
4
2
1
3
O
N
e
+
O
O
O
5
6
Scheme 1.
Synthesis of key intermediate 5 and dimerization product 6. Reagents and
conditions: (a) TMSCN, ZnI2, CH2Cl2, reflux, 90%; (b) POCl3, KF, pyridine, reflux, 16 h, 48%;
(c) HCl (aq.), EtOH, reflux, 85%; (d) POCl3, pyridine, reflux, 12 h, 82%; (e) MeMgI, diethyl
ether, toluene, 65°C, 6 h, AcOH, H2O, 5°C, HCl (aq), reflux 2 h, 75% (two steps).
5

ACCEPTED MANUSCRIPT
The Lewis acid (AlCl3) catalyzed Diels–Alder reaction [19] of conjugated ketone 5 with
butadiene produced pentacyclic steroid 7 containing the additional 16α,17α-cyclohexane ring in
high yield. This steroid was hydrogenated under atmospheric pressure in the presence of 10%
palladium on activated carbon as the catalyst (Scheme 2). Hydrogenated adduct 8 was
demethylated with a mixture of 48% hydrobromic acid and sodium iodide in glacial acetic acid
to prepare 3-hydroxysteroid 9 in high yield. The addition of an equimolar amount of sodium
iodide as a source of a highly reactive nucleophile to a standard demethylation system allowed us
to substantially reduce the reaction time.
3,20-Diols 10a and 10b were synthesized by reducing the 20-keto group in hydroxysteroid
9 with lithium aluminum hydride in THF. In this case, we isolated only enantiomer 10a having
20(R) configuration. The heating of 3-methoxy-20-ketosteroid 8 with a toluene solution of
diisobutylaluminum hydride (DIBAH) under reflux allowed the synthesis of a diastereomeric
mixture of 3,20-diols by one-pot reduction–dementylation. Individual isomers 10a and 10b were
isolated from the mixture by preparative reversed-phase HPLC (RP-HPLC). The absolute
configuration of the new chiral center С20 was established by NMR spectroscopy. For this
1
13
purpose, the H and C chemical shift assignments (Tables 1a and 1b) and NOE experiments
were performed. The 2D NOESY spectra of 10a and 10b were not clear enough due to
consequences of the chemical shifts of protons of CH3-21 methyl with H22 and H25 protons. We
performed ge-1D NOESY experiments to reveal the differences in NMR for these isomers.
Irradiation of the hydroxyl proton at C20 in ge-1D NOESY experiments provided responses for
H20, CH3-21, H12 (δН1.81) for 10a (R-isomer), while responses for H20, CH3-21, H16 (δН
2.24), H22 (δН 1.73), H25 (δН 1.89) were observed on irradiation of the hydroxyl proton of 10b
(S-isomer) (Fig. 1). Besides, compound 10a was studied by X-ray diffraction (Fig. 2). The
configuration at C20 was assigned as R.
3-Methoxy-20-hydroxysteroid 11 was synthesized as a mixture of 20(R) and 20(S) isomers
by reduction of 20-ketosteroid 8 with lithium aluminum hydride or DIBAH at room temperature.
A comparative study of the 20-keto group reduction by these reagents showed that the ratio of
the 20(R) and 20(S) isomers is 4.3 : 1 for LAH and 2.3 : 1 for DIBAH (NMR data).
6

ACCEPTED MANUSCRIPT
e
O
O
O
HO
a
b
c
d
5
H
H
H
H
O
HO
O
HO
7
10a: 20(R)
10b: 20(S)
8
9
d
HO
H
O
11: 20(R,S)
Scheme 2.
Synthesis of pentacyclic steroids containing the additional 16α,17α-
cyclohexane ring and functional hydroxyl groups at positions 3 and 20. Reagents and conditions:
(a) butadiene, AlCl3, CH2Cl2, RT, 68%; (b) H2, 10% Pd/C, dioxane, 93%; (c) HBr (48%), AcOH,
NaI, reflux, 3.5 h, 77%; (d) LiAlH4, THF, RT, 48%; (e) DIBAH, toluene, reflux, 6 h.
H
H
H
H
O
21
H
21
H
H
20
H
O
20
H
H
H
H
H
12
H
12
22
22
H
H
B-ring
B-ring
H
16
H
16
25
25
B-ring
B-ring
S
R
Fig. 1. Structure assignment for compounds 10a and 10b using ge-1D NOESY. Most
significant interactions and fragments of the steroid scaffold are shown for clarity.
Fig. 2. General view of one independent molecules of compound 10a in the crystal
structure shown with displacement ellipsoids (p=50%).
To evaluate the size effect of the additional ring D' of the steroids on the biological
properties, we synthesized 3,20-dihydroxysteroids containing the additional cyclopropane ring D'
7

ACCEPTED MANUSCRIPT
annulated at positions 16α,17α (Scheme 3) and steroids 16 and 17 that lack the additional ring D'
(Scheme 4).
Compound 12 was synthesized by the Corey–Chaykovsky reaction of conjugated ketone 5
with trimethylsulfoxonium iodide. The corresponding 3,20-dihydroxysteroid 13 was synthesized
in one step from compound 12 by heating the latter under reflux with DIBAH. The cyclopropane
moiety, which can be decomposed during demethylation by strong acids, remains intact under
these conditions [20, 21]. Steroid 13 was isolated as a mixture of 20(R) and 20(S) isomers.
1
Nevertheless, the structure assignment for both isomers was carried out using 2D H/1H COSY,
1
TOCSY, NOESY and H/13C HSQC, HMBC techniques. The configurations of asymmetric
centers were established by the careful analysis of the NOESY spectra. The distance between
CH3-21 and H16 for the 20(R) and 20(S) isomers should be different. The hydrogen H16β is the
nearest neighbor of CH3-21 in the R isomer, while this methyl is far apart from H16β in the S
isomer (Fig. 3). The NOESY spectrum of a mixture of these isomers (see the Supplementary)
shows a strong CH3-21 (δН 0.84)/H16β (δН 0.96) cross-peak, but no cross-peaks are observed
between the corresponding signals CH3-21 (δН 1.09)/H16β (δН 1.23), which is evidence that the
former pair of signals is from the R isomer, whereas the latter one is from the S isomer.
O
HO
b
a
5
H
H
O
HO
12
13: 20(R,S)
Scheme 3.
Synthesis
of
3,20-dihydroxy-16α,17α-cyclopropano-19-norpregna-
1,3,5(10)-triene 13. Reagents and conditions: (a) Me3SOI, NaH, DMSO/THF, RT, 24 h, 59%; (b)
DIBAH, toluene, reflux, 6 h, 32%.
H
H
O
H
H
21
20
H
H
+
H
H
20
O
21
H
H
H
H
B-ring
B-ring
16
16
B-ring
B-ring
R
S
Fig. 3. Structure assignment for the 20(R) and 20(S) isomers of compound 13 using
NOESY. Most significant interactions and fragments of the steroid scaffold are shown for
clarity.
8

ACCEPTED MANUSCRIPT
The hydrogenation of conjugated ketone 5 gave saturated ketone 14. The demethylation of
the latter with HBr/NaI/AcOH followed by the reduction of 3-hydroxy-20-ketosteroid 15 that
formed with lithium aluminum hydride produced 3,20-dihydroxysteroid 16 which was isolated
as 20(R) isomer after column chromatography and recrystallization. In the NOESY spectrum of
16, the CH3-21 protons (δН 1.01) show (apart from trivial CH3-21, H-20, and OH-20 mutual
signals) a strong cross-peak with H16β (δН 1.63), a weak cross-peak with H16α (δН 1.14), and a
medium cross-peak with H17 (δН 1.30), while only a cross-peak with H17 appears for the
hydroxyl proton at C20 (δН 4.09). This assignment was confirmed by X-ray analysis (Fig. 4).
O
O
HO
b
c
a
HO
O
HO
14
15
16: 20(R)
5
HO
d
HO
17: 20(R,S)
Scheme 4.
Synthesis of 3,20-dihydroxy-19-norpregna-1,3,5(10)-tri- and 1,3,5(10),16-
tetraenes. Reagents and conditions: (a) H2, 10% Pd/C, dioxane, 91%; (b) HBr (48%), AcOH,
NaI, reflux, 3.5 h, 53%; (c) LiAlH4, THF, RT, 40%; (d) DIBAH, toluene, reflux, 6 h, 47%.
Fig. 4. General view of one of the two independent molecules in the crystal structure of
compound 16 shown with displacement ellipsoids for non-hydrogen atoms (p=50%).
A comparative study of the reduction of 16,17-unsubstituted 3-methoxy ketone 14 with
lithium aluminum hydride and DIBAL demonstrated that in this case (cf. compound 11), the
stereoselectivity of the reduction of the 20-keto group is somewhat higher, while the difference
in the ratio of the 20(R)/(S) isomers obtained using these two reducing agents is smaller (4.6 : 1
and 3.9 : 1 for LAH and DIBAL, respectively). Thus, the initially lower content and the complex
9

ACCEPTED MANUSCRIPT
isolation procedure for the product were considered to explain the depletion of the 20(S) isomer
of compound 16.
3,20-Dihydroxysteroid 17 containing a double bond at position 16 was synthesized by
reduction–demethylation of conjugated ketone 5 with DIBAL. Since an almost equimolar
mixture of the 20(R) and 20(S) isomers of 17 was obtained, the configuration at C20 was
assigned by comparing the intensities of the cross-peaks in the NOESY spectrum (see the
Supplementary) corresponding to interactions of the side chain protons and the nearest protons of
the rigid steroid scaffold. The intensities of the significant cross-peaks H20/CH3-18, OH-20/H16,
and CH3-21/H16 are comparable for both isomers, whereas the intensities of the CH3-21/CH3-18
cross-peaks differ clearly between the isomers. The stronger peak (δН 1.22/ δН 0.81) should
correspond to the R configuration because the distance between the 18- and 21-methyl protons in
this configuration is shorter than that in the S configuration (Fig. 5).
As opposed to the above-described examples, the formation of an equimolar mixture of the
20(R) and 20(S) isomers is apparently attributed to the fact that the 20-keto group that is reduced
is far apart from the stereocenter at C13, resulting in lower asymmetric induction.
The 1H and 13C NMR signal assignments of compounds 10a, 10b, 13, 16, and 17 are
summarized in Tables 1a,b, respectively.
H
H
H
O
H
H
18
H
H
H
21
18
H
H
21
H
H
H
C-ring
O H
H
H
C-ring
H
H
16
16
C-ring
C-ring
R
S
Fig. 5. Structure assignment for the 20(R) and 20(S) isomers of compound 17 using
NOESY. Most significant interactions and fragments of the steroid scaffold are shown for
clarity.
Table 1a.
and 17.
1H NMR (DMSO-d6, 600 MHz) spectra of compounds 10a, 10b, 13, 16,
Compound
10a
R
10b
S
13
16
R
17
Atom n.
R
S
R
S
1
2
3
4
5
6
7
8
6.99
6.48
-
6.42
-
6.99
6.48
-
6.42
-
7.03
6.51
-
6.42
-
7.02
6.50
-
6.44
-
2.69
2.70
1.72;1.23
1.36
2.70
1.75; 1.25
1.25
2.72
1.82; 1.33
1.45
1.77; 1.27 1.76;1.25
1.26
10

ACCEPTED MANUSCRIPT2.05
9
10
2.01
2.03
2.03
-
2.14
-
-
-
11
2.10; 1.29 2.14; 1.27
2.20; 1.30
2.15; 1.28
2.27; 1.43
12
13
14
15
16
17
18
19
1.81; 1.52
1.61
1.50
1.71
2.00; 1.36 1.77; 1.36 2.12; 1.27 1.93; 1.60 1.88; 1.50
-
-
-
-
1.01
1.03
1.16
1.52
2.08; 1.86
5.50 5.52
1.47
1.81; 1.60 1.56; 1.30 1.53; 1.33 1.53; 1.28 1.60; 1.12
2.02
0.86
2.24
0.75
0.96
0.91
1.23
1.63; 1.14
1.30
-
-
-
0.87
-
0.70
0.81
0.84
20
21
22
3.81
1.15
1.28
3.85
1.10
4.15
0.84
4.07
1.09
3.51
1.01
-
4.25
1.22
4.20
1.20
1.73; 1.25 0.68; 0.40 0.54; 0.44
-
23
24
25
3-OH
20-OH
1.45; 1.14 1.71; 1.45
1.47 1.42; 1.31
1.59; 1.31 1.89; 1.42
8.93
-
-
-
-
-
-
-
-
-
8.94
8.92
4.10
3.95
4.10
4.12
4.02
4.08
4.44
4.49
Table 1b.
and 17.
Compound
Atom n.
13C NMR (DMSO-d6, 150 MHz) spectra of compounds 10a, 10b, 13, 16,
10a
R
10b
S
13
16
R
17
R
S
R
S
1
2
3
4
5
6
7
8
125.8
112.6
154.8
114.9
137.1
29.2
125.6
112.6
154.9
114.9
137.0
29.0
125.9
112.6
154.8
114.8
137.0
29.2
27.5
38.5
43.4
130.6
26.3
39.4
42.3
54.5
25.3
23.8
57.9
12.0
125.6
112.6
154.9
114.9
137.1
29.0
27.4
37.0
43.8
130.6
26.1
27.7
38.9
43.2
130.7
26.1
32.5
49.2
48.2
32.9
35.4
49.1
14.4
27.6
38.7
43.1
130.6
25.8
31.7
47.3
48.6
32.3
40.1
49.2
14.1
27.6
37.1
43.8
27.5
37.0
43.7
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
130.7
130.5
26.1
35.3
42.1
46.1
26.0
19.5
41.0
16.9
35.1
41.0
46.7
25.7
18.1
39.9
17.1
-
34.7
45.9
56.6
56.2
30.1
121.2
159.5
16.4
121.7
160.1
16.5
71.1
22.2
23.8
20.2
29.5
33.1
74.1
20.2
24.6
23.3
20.6
29.8
65.4
20.6
6.4
62.5
23.2
4.6
68.3
21.7
-
-
-
-
63.6
23.7
63.4
23.6
-
-
-
-
-
-
-
2.2
Molecular modeling
To investigate the possible binding mode and ligand–receptor interactions, we performed
molecular docking of the compounds into the ligand binding site of the nuclear receptor ERα
(PDB: 1QKU) followed by energy calculations by the MM-PBSA method in order to determine
the individual characteristics of these interactions.
11

ACCEPTED MANUSCRIPT
The basic requirements for high-affinity ligand binding to ERα are well known due to an
extensive structure–activity relationship studies and X-ray crystallographic data for ligand–ER
LBD complexes [22-24]. According to the X-ray structure of the estradiol complex with ER
LBD [22], two appropriately spaced hydroxyl groups are located at both ends of the nearly
planar hydrophobic scaffold in such a way that the phenolic hydroxyl forms hydrogen bonds
with the amino acid residues glutamate (Glu353) and arginine (Arg394) and also with a
conserved water molecule, while another hydroxyl (17β-OH) is recognized by a single histidine
(His524). All other contacts in the complex are hydrophobic [22].
The docking results showed that all the compounds under study (8-11, 13, 16, and 17) can
bind at the ligand-binding pocket (LBP) of estrogen receptor α (ER-α).
In particular, as can be seen in Fig. 7, compound 10a interacts with similar protein residues
of LBD, like natural estradiol in the X-ray structure of E2 in complex with ERα LBD [22], and
can be well-accommodated in the ligand-binding pocket despite the larger volume of steroid 10a
compared to estradiol. The additional ring D', which enhances hydrophobic binding, fits well in
α-region of the cavity near the ring D.
The analysis of the binding of the compounds under study demonstrates that all of them
can bind to ER LBD with similar energies that are comparable to or slightly higher than the
value estimated for the estradiol molecule (Table 2).
Fig. 7. Interactions (hydrogen bonds) calculated for compound 10a and the key ligand-
binding pocket residues of human estrogen receptor α (left) and the arrangement of ligand 10a in
the ligand-binding pocket (right) (calculated based on the crystal structure of the ERα LBD–E2
complex, PDB:1QKU). The surfaces are colored by lipophilicity, ranging from brown (most
lipophilic) to blue (least lipophilic).
Table 2.
Estimated binding energies of compounds 8–11, 13, and 15–17 to the human
estrogen receptor alpha ligand-binding domain.
Compound
∆E (GBTOT),
12
STD

ACCEPTED MANUSCRIPT
(configuration of C20)
kkal/mol
-66.09
-74.22
-75.72
-79.14
-81.53
-76.09
-82.08
-69.87
-66.38
-67.72
-71.15
-66.12
-67.63
-70.43
±2.91
±1.60
±2.62
±1.71
±2.12
±2.09
±2.70
±3.50
±3.54
±3.08
±2.53
±2.03
±2.20
±2.81
E2
8
9
10a (R)
10b (S)
11 (R)
11 (S)
13 (R)
13 (S)
15
16 (R)
16 (S)
17 (R)
17 (S)
2.3
Biological evaluation
Сytotoxic activity
2.3.1
Compounds 8-11, 13, and 15-17 were tested in vitro for cytotoxic activity against the
estrogen receptor-positive MCF-7 breast cancer cell line. Cytotoxicity was evaluated by the
MTT test based on the accumulation of the MTT reagent (3-[4,5-dimethylthiazol-2-yl]-2,5-
diphenyltetrazolium bromide) by living cells. Compounds 13 and 16 displayed the highest
activity. They exhibited the cytotoxic effect at submicromolar level. Three steroids – 10a, 10b,
and 17 – showed activity comparable to that of the SERM tamoxifen. Only weak effects in
MCF-7 cells were detected for compound 9, while 8, 11, and 15 exhibited no activity in MCF-7
breast cancer cells. To elucidate possible toxicity of prepared compounds against normal
epithelium MCF-10A non-tumorigenic epithelial cells were tested. The MCF10A human
mammary epithelial cell line is a widely used in vitro model for studying normal breast cell
function and transformation.
As can be seen in Table 3, compounds 13 and 16 displayed high toxicity towards MCF-
10A cells, while 10b has an insignificant effect. The other compounds, as well as the reference
drug tamoxifen, exhibited no activity (no toxic effect) against nontumorigenic cells.
Table 3.
The IC50 (half-maximal inhibitory concentration) of compounds. The
cells were grown for 72 h and then the inhibitory activity of the
compounds was evaluated by the MTT test.
IC50, µM
Compound
8
MCF-7 breast cancer cells
MCF-10A normal cells
NA
NA
13

ACCEPTED MANUSCRIPT
15.8 ± 1.4
NA
NA
23.0 ± 2.5
NA
9
10a
10b
6.8 ± 0.7
4.6 ± 0.5
NA
11*
0.21 ± 0.03
NA
3.1 ± 0.4
NA
13*
15
0.15 ± 0.03
4.6 ± 0.4
5.3 ± 0.6
5.9 ± 0.5
19.1 ± 1.8
NA
16
17*
tamoxifen
NA – does not reach 50% growth inhibition at concentrations lower than 25 µM
* tested as a mixture of 20(R,S) epimers
2.3.2
ERα-mediated luciferase reporter gene assay
The ERα transcriptional activity was evaluated using the ERα luciferase reporter gene
assay. The ability of the compounds to inhibit ERα activity was analyzed in MCF-7 cells after
ERα luciferase reporter plasmid transfection. The β-galactosidase plasmid construct was used to
normalize transfection efficiency. Compounds 8, 11, 15, and 17 displayed no antiestrogenic
activity, while steroids 9, 10b, 13, and 16 partially inhibited ERα (Fig. 8). Compound 10a
exhibited the highest ERα antagonist potency. The treatment with this steroid resulted in strong
down-regulation of ERα activity. The incubation with 5 µM 10a led to ca. 70% inhibition of
ERα activity in MCF-7 cells, demonstrating the efficacy of 10a being comparable to that of the
SERM drug tamoxifen.
As can be seen from the results of the ERα transcriptional activity assay (see Fig. 8), only
five compounds (9, 10a, 10b, 13, and 16) of the nine tested compounds have pronounced ER
antagonist activity.
In both biological assays, compounds containing the hydrophobic steroid scaffold with
the aromatic ring A and the 3- and 20-hydroxyl groups at both ends exhibited the highest
activity. It should be noted that the spatial arrangement of the 20-hydroxyl group has apparently
a weaker influence on the effects of the compounds (cf. 10a and 10b). Meanwhile, the phenolic
hydroxyl is essential for the manifestation of inhibitory activity. Thus, the steroid containing the
20-hydroxyl and the 3-OMe group instead of the second hydroxyl (11) does not exhibit these
properties, whereas the steroid containing 20-keto and 3-OH groups (9) exerts some activity in
both assays. The introduction of the additional six-membered ring D' at 16α,17α positions into
the steroid containing 3-phenol and 20-hydroxyl groups proved to be an effective modification of
the steroid scaffold for the design of agents exhibiting antiestrogenic activity. It should be
emphasized that both the compounds containing the 16α,17α-methylene substituent and the
compounds containing the ring D devoid of substituents retain this activity to a lesser extent.
14

ACCEPTED MANUSCRIPT
140
120
100
80
60
40
20
0
cont
8
9
10a
10b
11
13
15
16
17 tamo
Fig. 8. Estrogen receptor α activity in MCF-7 cells. MCF-7 human breast cancer cells
were transfected with the ERE-TK-LUC plasmid containing the luciferase reporter gene under
control of the estrogen responsive element (ERE) and cotransfected with the β-galactosidase
plasmid. The medium was removed 24 h after the transfection. The synthesized steroids, the
SERM drug tamoxifen at 5 µM concentration, or the vehicle control (cont) were added to
phenol-free DMEM supplemented with 5% DCC serum (Hyclone) and 10 nM 17β-estradiol. The
luciferase and β-galactosidase activities were determined after 24 h as described in Section 4.4.2
using the ERα luciferase reporter gene assay.
3. Conclusion
A series of 3,20-dihydroxy-19-norpregna-1,3,5(10)-triene derivatives – steroids containing
the additional carbocyclic ring D' and two hydroxyl groups at both ends of the hydrophobic
steroid nucleus characteristic of estrogens – were synthesized. Docking studies showed that these
compounds can bind to ERα, their binding mode is similar to that of natural estradiol. Most of
these compounds exhibit cytotoxic activity against breast cancer cells and inhibit ER
transcriptional activity in the ERα luciferase reporter gene assay. The 16α,17α-methylene-
substituted or unsubstituted compounds display the highest cytotoxicity, being simultaneously
relatively weak ERα inhibitors. 3,20-Dihydroxysteroids containing the six-membered ring D'
proved to be the most effective estrogen receptor α inhibitors. These compounds exhibit
moderate cytotoxicity comparable to that of tamoxifen, while having no toxic effect on normal
(nontumorigenic) cells. The results of this study showed that the new ER antagonists are good
candidates for further testing as agents for the treatment and prevention of ERα-positive breast
cancers.
4. Materials and methods
4.1
General
15

ACCEPTED MANUSCRIPT
All reagents were purchased from Acros Organics and were of reagent or analytical grade.
Solvents were purified according to standard procedures [25]. The analytical TLC was carried
out on silica gel 60 F254 plates (Merck) using dichloromethane, a 40 : 1 toluene–acetone
mixture, or a 50 : 1 dichloromethane–acetone mixture as eluents. The spots were visualized by
UV light or by spraying with aqueous KMnO4. The preparative separation was carried out by
column chromatography on silica gel 60 (0.063–0.100mm) (Merck) at a compound–sorbent ratio
of 1 : 40. Preparative RP-HPLC was performed on a Sunfire C18 (19 mm x 250 mm) preparative
column with elution at a flow rate of 20 ml/min using MeCN/water (70 : 30) and UV detection at
280 nm. The melting points were determined on a Boetius micro-melting point apparatus
(Germany). All yields are given for purified compounds.
One- and two-dimensional NMR spectra were recorded on Bruker AM300 (300.13 MHz
1
13
1
13
for H and 75.5 MHz for C) and Bruker AV-6 (600.13 MHz for H and 150.9 MHz for C)
spectrometers using standard Bruker software. Chemical shifts are given in δ (ppm) relative to
the residual solvent peaks: δH 7.27 and δC 77.0 for CHCl3; δH 2.50 and δC 39.5 for DMSO-d5.
The mixing time for TOCSY and NOESY experiments were 100 ms and 600 ms, respectively,
spin-lock time for ROESY experiments was 150 ms. The 1H/13C and 1H/15N HMBC experiments
were optimized for the spin-spin coupling constant of 8 Hz. High-resolution mass spectra
(HRMS) were measured on a Bruker micrOTOF II instrument using electrospray ionization
(ESI) in positive ion mode (interface capillary voltage 4500 V); the mass range from m/z 50 to
m/z 3000 Da; external or internal calibration was performed using Electrospray Calibrant
Solution (Fluka). Solutions in acetonitrile or methanol were injected with a syringe (flow rate 3
µL/min). Nitrogen was applied as a dry gas; interface temperature was set at 180 °C. X-ray
diffraction data were collected on a Bruker APEX DUO diffractometer. Frames were integrated
using the Bruker SAINT software package with a narrow-frame algorithm. A semiempirical
absorption correction was applied with the TWINABS or SADABS programs based on
intensities of equivalent reflections. The structures were solved by direct methods and refined by
16

ACCEPTED MANUSCRIPT
the full-matrix least-squares technique against F2 with anisotropic displacement parameters. The
hydrogen atoms of hydroxyl groups were located from difference Fourier maps; all other
hydrogen atoms were positioned geometrically. All calculations were performed with the
SHELX software package. Atomic coordinates, bond lengths and angles, and thermal parameters
were deposited at the Cambridge Crystallographic Data Center (CCDC), the reference numbers
are given in the relevant sections of the Experimental section. These data can be obtained, free of
charge,
via
http://www.ccdc.cam.ac.uk/data_request/cif,
or
by
e-mailing
data_request@ccdc.cam.ac.uk, or by contacting the Cambridge Crystallographic Data Centre, 12
Union Road, Cambridge CB2 1EZ, UK; fax: +44(0)1223-336033.
4.2
Chemical synthesis
4.2.1
3-Methoxy-17β-trimethylsiloxyestra-1,3,5(10)-triene-17α-carbonitrile 2
Estrone 3-methyl ether (1, 30.1 g, 106 mmol) was taken in dry dichloromethane (250 ml).
To this solution, anhydrous zinc iodide (1.00 g, 3 mmol) and trimethylsilyl cyanide (20 ml, 14.9
g, 150 mmol) were added, and the mixture was refluxed with stirring for 2 h until the starting
material disappeared on TLC. Then the reaction mixture was concentrated in vacuo, and the
solid residue was dissolved in a boiling mixture of toluene and hexane (1 : 2). The hot solution
was filtered through a silica gel pad (5 g) and allowed to cool in a refrigerator to 8°C. The
precipitate was collected, washed with cold hexane, and dried to obtain 3-methoxy-17β-
trimethylsiloxyestra-1,3,5(10)-triene-17α-carbonitrile 2.
1
Yield: 90%, m.p. 135–136˚C, H NMR (CDCl3, 300 MHz): δ 0.26 (s, 9H, Si(CH3)3), 0.85
(s, 3H, 18-CH3), 1.35–2.58 (m, 13H, rest of the 5 x CH2 and 3 x CH of steroid ring), 2.87 (bs,
2H, 6-CH2), 3.79 (s, 3H, OCH3), 6.65 (s, 1H, 4-CH), 6.74 (d, 1H, 2-CH, J = 8.4 Hz), 7.22 (d,
13
1H, 1-CH, J = 8.4 Hz); C NMR (CDCl3, 75 MHz): δ 1.3, 12.3, 23.1, 26.3, 27.2, 29.8, 33.3,
38.1, 39.4, 43.4, 48.2, 48.4, 55.3, 111.6, 113.9, 122.4, 126.4, 132.2, 137.9, 157.6; HRMS: m/z
406.2175 [M + Na]+ (calcd for C23H33NNaO2Si, 406.2173).
4.2.2
3-Methoxy-17β-hydroxyestra-1,3,5(10)-triene-17α-carbonitrile 3
17

ACCEPTED MANUSCRIPT
Estrone 3-methyl ether (1, 70.2 g, 250 mmol) was taken in dry dichloromethane (600 ml).
To this solution, anhydrous zinc iodide (2.4 g, 7.5 mmol) and trimethylsilyl cyanide (43 ml, 31.7
g, 320 mmol) were added. The mixture was refluxed with stirring for 2 h until the starting
material disappeared on TLC and then concentrated in vacuo. The solid residue was dissolved in
a mixture of ethanol (600 ml), water (90 ml), and conc. hydrochloric acid (5 ml) and refluxed for
1 h until the starting material disappeared on TLC. Then distilled water (350 ml) pre-warmed to
80˚C was added, and the mixture was allowed to cool to ambient temperature. The precipitate
was collected, washed with distilled water, and dried to obtain 3-methoxy-17β-hydroxyestra-
1,3,5(10)-triene-17α-carbonitrile 3.
1
Yield: 85%, m.p. 158–161˚C with decomposition (cf. lit. m.p. 158.5˚C [26]), H NMR
(DMSO-d6, 300 MHz): δ 0.79 (s, 3H, 18-CH3), 1.22–2.44 (m, 13H, rest of the 5 x CH2 and 3 x
CH of steroid ring), 2.77 (bs(m), 2H, 6-CH2), 3.69 (s, 3H, OCH3), 6.56 (bs, 1H, OH), 6.61 (s,
13
1H, 4-CH), 6.68 (d, 1H, 2-CH, J = 8.3 Hz), 7.17 (d, 1H, 1-CH, J = 8.3 Hz); C NMR (CDCl3,
75 MHz): δ 11.9, 122.3, 25.8, 26.7, 29.1, 33.2, 36.6, 38.7, 43.0, 46.8, 49.4, 54.8, 79.8, 111.5,
113.4, 123.0, 126.2, 131.6, 137.3, 157.1; HRMS: m/z 312.1963 [M + H]+ (calcd for C20H26NO2,
312.1958).
4.2.3
3-Methoxyestra-1,3,5(10),16-tetraene-17-carbonitrile 4
a) Starting from 3-methoxy-17β-trimethylsiloxyestra-1,3,5(10)-triene-17α-carbonitrile 2.
Cyanohydrin silyl ether (2, 10 g, 26 mmol) and potassium fluoride (2.1 g, 36 mmol) were taken
in dry pyridine (20 ml) and then POCl3 (5 ml, 8.23 g, 54 mmol) was added dropwise with
stirring. The reaction mixture was incubated for 30 min, refluxed with stirring for 16 h, cooled to
ca. 50˚C, poured into cold water acidified with conc. hydrochloric acid to pH 1-2, and extracted
with CHCl3 (4 x 100 ml). The organic layer was washed with water and brine, dried over
anhydrous Na2SO4, filtered through a silica gel pad (10 g), and concentrated in vacuo. The
resulting yellow oil was recrystallized from a mixture of toluene and hexane (3 : 1) to obtain 3-
methoxyestra-1,3,5(10),16-tetraene-17-carbonitrile 4.
18

ACCEPTED MANUSCRIPT
Yield: 48%, m.p. 174–175˚C (cf. lit. m.p. 168–170˚C [15]), 1H NMR (CDCl3, 300 MHz): δ
0.97 (s, 3H, 18-CH3), 1.22–2.53 (m, 11H, rest of the 4 x CH2 and 3 x CH of steroid ring), 2.91
(m, 2H, 6-CH2), 3.79 (s, 3H, OCH3), 6.66 (bs, 1H, 4-CH), 6.68 (bs, 1H, 16-CH), 6.74 (d, 1H, 2-
CH, J = 8.3 Hz), 7.22 (d, 1H, 1-CH, J = 8.3 Hz); 13C NMR (CDCl3, 75 MHz): δ 16.3, 26.1, 27.6,
29.5, 32.6, 34.0, 37.1, 44.1, 48.4, 55.2, 111.5, 113.9, 115.9, 126.0, 127.5, 132.0, 137.5, 147.3,
157.6; HRMS: m/z 311.2122 [M + NH4]+ (calcd for C20H27N2O, 311.2118).
b) Starting from 3-methoxy-17β-hydroxyestra-1,3,5(10)-triene-17α-carbonitrile 3.
Cyanohydrin (3, 18.4 g, 59 mmol) was taken in dry pyridine (50 ml) and then POCl3 (11 ml, 18.1
g, 118 mmol) was added dropwise with stirring. The reaction mixture was incubated for 30 min
and refluxed with stirring for 12 h until the starting and intermediate compounds disappeared on
TLC. The completion of the reaction was confirmed by the easy and clear layer separation as a
result of the mixing of a sample of the reaction mixture with CHCl3 and water acidified with
HCl. Then the reaction mixture was cooled to ca. 50˚C, poured cold water acidified with conc.
hydrochloric acid to pH 1-2, and extracted with CHCl3 (3 x 100 ml). The organic layer was
washed with water and brine, dried over anhydrous Na2SO4, and concentrated in vacuo to obtain
a dark brown tarry mass, which was dissolved in a boiling mixture of toluene and hexane (150
ml and 60 ml). The hot solution was filtered through a silica gel pad (20 g) and allowed to cool
to room temperature and then in a refrigerator to 8°C. The precipitate that formed was collected,
washed with hexane, and dried to obtain 3-methoxyestra-1,3,5(10),16-tetraene-17-carbonitrile 4
in 82% yield.
4.2.4
3-Methoxy-19-norpregna-1,3,5(10),16-tetraen-20-one 5
3-Methoxyestra-1,3,5(10),16-tetraene-17-carbonitrile (4, 10.0 g, 34.1 mmol) was taken in
a mixture of toluene (160 ml) and dry diethyl ether (100 ml) under argon. To this solution, a 3.0
M methylmagnesium iodide solution in diethyl ether (25 ml, 75 mmol) was added. The reaction
mixture was incubated with stirring at 60–65˚C for ca. 5 h until the starting nitrile disappeared
on TLC. Then the reaction mixture was cooled to 5˚C, and glacial acetic acid (50 ml) was added
19

ACCEPTED MANUSCRIPT
dropwise with cooling and vigorous stirring in such a way as to maintain the internal temperature
below 15˚C. Then glacial acetic acid (40 ml) and water (10 ml) were added to the reaction
mixture, which was refluxed for 1 h. Next, water (10 ml) and conc. HCl (3 ml) were added to the
boiling reaction mixture and the mixture was refluxed for 1 h. After cooling, the organic layer
was separated and washed with water and brine. All aqueous layers were pooled and extracted
with CHCl3 (3 x 30 ml). The organic fractions were worked-up in the same way, pooled, dried
with anhydrous Na2SO4, and concentrated in vacuo to obtain a solid yellow residue, which was
dissolved in a boiling mixture of toluene and hexane (3 : 1). The hot solution was filtered
through a silica gel pad (5 g) and allowed to cool in a refrigerator to 8˚C. The crystalline
precipitate that formed was separated, washed with hexane, and dried to obtain 3-methoxy-19-
norpregna-1,3,5(10),16-tetraen-20-one 5.
Yield: 75%, m.p. 194–195˚C (cf. lit. m.p. 192–194˚C [15]), 1H NMR (CDCl3, 300 MHz): δ
0.93 (s, 3H, 18-CH3), 1.22–2.53 (m, 11H, rest of the 4 x CH2 and 3 x CH of steroid ring), 2.30 (s,
3H, 21-CH3), 2.91 (m, 2H, 6-CH2), 3.79 (s, 3H, OCH3), 6.65 (bs, 1H, 4-CH), 6.71 (d, 1H, 2-CH,
J = 8.9 Hz), 6.75 (bs, 1H, 16-CH), 7.22 (d, 1H, 1-CH, J = 8.9 Hz); 13C NMR (CDCl3, 75 MHz):
δ 15.9, 26.4, 27.1, 27.7, 29.6, 31.9, 34.7, 36.9, 44.2, 46.4, 55.2, 55.5, 111.3, 113.8, 126.1, 126.1,
132.7, 137.7, 144.3, 155.5, 157.4, 196.8; HRMS: m/z 311.2015 [M + H]+ (calcd for C21H27O2,
311.2006).
4.2.5
“Dimer” 6
3-Methoxyestra-1,3,5(10),16-tetraene-17-carbonitrile (4, 7.47 g, 25.5 mmol) was taken in
toluene (120 ml) under argon. To this solution, a 3.0 M methylmagnesium iodide solution in
diethyl ether (14 ml, 42 mmol) was added. The reaction mixture was incubated with stirring at
ca. 65˚C for ca. 5 h until the starting nitrile disappeared on TLC and then cooled to 5˚C followed
by the addition of glacial acetic acid (10 ml), which was accompanied by an increase in the
internal temperature to ca. 35˚C. Then the reaction mixture was cooled to ambient temperature,
water (25 ml) and conc. hydrochloric acid (1 ml) were added, and the mixture was refluxed for 1
20

ACCEPTED MANUSCRIPT
h. After cooling, the organic layer was separated and washed with water and brine. All aqueous
layers were pooled and extracted with CHCl3 (3 x 30 ml). The organic fractions were worked-up
in the same way, pooled, dried with anhydrous Na2SO4, and concentrated in vacuo to obtain a
yellow viscous residue , which was chromatographed on silica gel using, sequentially, a mixture
of dichloromethane and hexane (2 : 1) and a mixture of dichloromethane, hexane, and acetone (4
: 1 : 1). Then the relevant fractions were pooled and concentrated. Ketone 5 and “dimer” 6 were
obtained in yields of 60% and 5%, respectively. .
“Dimer”: m.p. 266–269˚C (toluene), 1H NMR (CDCl3, 300 MHz): δ 1.06 (s, 3H, 18-CH3),
1.16 (s, 3H, 18’-CH3), 2.37 (s, 3H, 21-CH3), 1.43–2.52 (m, 24H, rest of the 7 x CH2 and 6 x CH
of both steroid rings), 2.72 (t, 1H, 15-CH), 2.93 (m, 5H, 6-CH2, 6’-CH2, 15-CH), 3.81 (s, 6H, 2 x
3-OCH3), 6.42 (bs, 1H, 16’-CH), 6.68 (bs, 2H, 4-CH, 4’-CH), 6.75 (m, 2H, 2-CH, 2’-CH), 6.95
(s, 1H, 20-CH), 7.24 (d, 2H, 1-CH, 1’-CH, J = 8.1 Hz); HRMS: m/z 600.3819 [M + H]+ (calcd
for C42H50NO2, 600.3836); a sample for X-ray diffraction was additionally recrystallized from
toluene, CCDC ID: 1495043.
4.2.6
3-Methoxy-16α,17α-cyclohex-3’,4’-eno-19-norpregna-1,3,5(10)-trien-20-one 7
3-Methoxy-19-norpregna-1,3,5(10),16-tetraen-20-one (5, 13.5 g, 44 mmol) and anhydrous
aluminum chloride (0.94 g, 7 mmol) were taken in dry dichloromethane (180 ml) under argon.
Dry butadiene (6.4 g, 119 mmol) was slowly purged with stirring at ambient temperature through
the resulting dark beet-red solution. The mixture was further stirred for 18 h in a sealed round-
bottom flask until the starting material disappeared on TLC (toluene– acetone). Then the reaction
was quenched with methanol (5 ml) and saturated NaHCO3 (10 ml), after which the solution
turned pale yellow. The organic layer was separated, washed with water, dried over anhydrous
Na2SO4, and concentrated in vacuo. The oily residue was recrystallized from hexane to obtain 3-
methoxy-16α,17α-cyclohex-3’,4’-eno-19-norpregna-1,3,5(10)-trien-20-one 7.
Yield: 68%, m.p. 135–137˚C (cf. lit. m.p. 136–137.5˚C [19]), 1H NMR (CDCl3, 300 MHz):
δ 0.76 (s, 3H, 18-CH3), 1.30–2.48 (m, 15H, rest of the 6 x CH2 and 3 x CH of steroid and
21

ACCEPTED MANUSCRIPT
additional rings), 2.15 (s, 3H, 21-CH3), 2.86 (m, 2H, 6-CH2), 3.14 (m, 1H, 16-CH), 3.79 (s, 3H,
OCH3), 5.83 (m, 2H, 3’,4’-CH=CH-), 6.64 (s, 1H, 4-CH), 6.72 (d, 1H, 2-CH, J = 8.1 Hz), 7.21
(d, 1H, 1-CH, J = 8.1 Hz); 13C NMR (CDCl3, 75 MHz): δ 16.9, 26.3, 27.1, 27.8, 28.5, 29.8, 29.9,
32,8, 33.8, 34.3, 38.7, 43.5, 46.3, 50.0, 55.2, 66.6, 111.5, 113.8, 126.0, 126.1, 129.4, 132.4,
157.5, 211.1; HRMS: m/z 387.2283 [M + Na]+ (calcd for C25H32NaO2, 387.2295).
4.2.7
3-Methoxy-16α,17α-cyclohexano-19-norpregna-1,3,5(10)-trien-20-one 8
3-Methoxy-16α,17α-cyclohex-3’,4’-eno-19-norpregna-1,3,5(10)-trien-20-one (7, 9.2 g, 25
mmol) and 10% Pd/C (0.45 g) as the catalyst were taken in dioxane (150 ml). The mixture was
hydrogenated at atmospheric pressure until the starting material disappeared on TLC
(visualization by spraying with aqueous KMnO4). Then the catalyst was filtered off, the solvent
was evaporated in vacuo, and the residue was recrystallized from a hexane–toluene mixture (5 :
1) to obtain 3-methoxy-16α,17α-cyclohexano-19-norpregna-1,3,5(10)-trien-20-one 8.
Yield: 93%, m.p. 138–139˚C (cf. lit. m.p. 139.5–142˚C (methanol) [27]), 1H NMR (CDCl3,
300 MHz): δ 0.73 (s, 3H, 18-CH3), 0.81–2.41 (m, 19H, rest of the 8 x CH2 and 3 x CH of steroid
and additional rings), 2.17 (s, 3H, 21-CH3), 2.87 (m, 2H, 6-CH2), 3.02 (m, 1H, 16-CH), 3.79 (s,
3H, OCH3), 6.64 (s, 1H, 4-CH), 6.72 (dd, 1H, 2-CH, J1 = 8.1 Hz, J2 = 2.3 Hz), 7.19 (d, 1H, 1-
13
CH, J = 8.8 Hz); C NMR (CDCl3, 75 MHz): δ 16.1, 21.2, 22.5, 26.4, 27.2, 27.3, 27.7, 28.0,
29.8, 29.9, 32.3, 34.2, 38.9, 43.9, 47.3, 49.5, 55.3, 64.4, 111.6, 113.9, 126.1, 132.6, 138.0, 157.6,
212.5; HRMS: m/z 389.2438 [M + Na]+ (calcd for C25H34NaO2, 389.2451).
4.2.8
3-Hydroxy-16α,17α-cyclohexano-19-norpregna-1,3,5(10)-trien-20-one 9.
A mixture of 3-methoxy-16α,17α-cyclohexano-19-norpregna-1,3,5(10)-trien-20-one (8,
0.37 g, 1.0 mmol), sodium iodide (0.15 g, 1 mmol), glacial acetic acid (5 ml), and conc.
hydrobromic acid (3 ml) was refluxed for 3.5 h and then poured into cold water. The precipitate
that formed was filtered off, dried, and recrystallized from aqueous methanol to obtain 3-
hydroxy-16α,17α-cyclohexano-19-norpregna-1,3,5(10)-trien-20-one 9.
22

ACCEPTED MANUSCRIPT
1
Yield: 77%, m.p. 239–240˚C, H NMR (DMSO- d6, 300 MHz): δ 0.62 (s, 3H, 18-CH3),
0.71–2.33 (m, 19H, rest of the 8 x CH2 and 3 x CH of steroid and additional rings), 2.09 (s, 3H,
21-CH3), 2.70 (bs(m), 2H, 6-CH2), 2.87 (m, 1H, 16-CH), 6.43 (s, 1H, 4-CH), 6.50 (d, 1H, 2-CH,
13
J = 7.7 Hz), 7.01 (d, 1H, 1-CH, J = 7.7 Hz); 8.97 (bs, 1H, 3-OH); C NMR (DMSO- d6, 75
MHz): δ 15.9, 20.9, 22.2, 26.2, 26.6, 27.6, 27.7, 27.9, 29.4, 29.6, 31.8, 33.8, 38.7, 43.4, 47.0,
48.8, 63.8, 113.0, 115.2, 126.0, 130.5, 137.3, 155.2, 211.7; HRMS: m/z 353.2469 [M + H]+
(calcd for C24H33O2, 353.2475).
4.2.9
20(R)-3,20-Dihydroxy-16α,17α-cyclohexano-19-norpregna-1,3,5(10)-triene 10a
3-Hydroxy-16α,17α-cyclohexano-19-norpregna-1,3,5(10)-trien-20-one (9, 0.25 g, 0.7
mmol) was taken in anhydrous tetrahydrofuran (15 ml). To this solution, LiAlH4 (0.07 g, 1.8
mmol) was carefully added. The reaction mixture was stirred for 48 h. Then methanol (5 ml),
saturated NH4Cl (10 ml), and conc. hydrochloric acid (2 ml) were sequentially added dropwise.
The organic layer was separated, and the aqueous layer was extracted with tetrahydrofuran (2 x
10 ml) and chloroform (2 x 10 ml). Each organic fraction was washed separately with saturated
NH4Cl. Then all extracts were pooled, dried over anhydrous Na2SO4, and concentrated in vacuo.
The resulting oily mass was purified by column chromatography (silica gel, dichloromethane–
acetone, 20 : 1) followed by recrystallization from aqueous ethanol to obtain 20(R)-3,20-
dihydroxy-16α,17α-cyclohexano-19-norpregna-1,3,5(10)-triene 10a.
1
Yield: 48%, m.p. 196–197˚C, H NMR (DMSO-d6, 300 MHz): δ 0.86 (s, 3H, 18-CH3),
1.05–2.15 (m, 20H, rest of the 8 x CH2 and 4 x CH of steroid and additional rings), 1.14 (d, 3H,
21-CH3, J = 5.8 Hz), 2.70 (m, 2H, 6-CH2), 3.82 (m, 1H, 20-CH), 4.12 (d, 1H, 20-OH), 6.42 (s,
1H, 4-CH), 6.48 (d, 1H, 2-CH, J = 8.1 Hz), 6.99 (d, 1H, 1-CH, J = 8.1 Hz), 8.95 (bs, 1H, 3-OH);
13C NMR, see the Table 1b in text; HRMS: m/z 355.2639 [M + H]+ (calcd for C24H35O2,
355.2632); a sample for X-ray diffraction was obtained from aqueous methanol, CCDC ID:
1495042.
4.2.10 20(S)-3,20-Dihydroxy-16α,17α-cyclohexano-19-norpregna-1,3,5(10)-triene 10b
23

ACCEPTED MANUSCRIPT
3-Methoxy-16α,17α-cyclohexano-19-norpregna-1,3,5(10)-trien-20-one (8, 0.37 g, 1.0
mmol) was taken in anhydrous toluene (15 ml). The system was flushed with argon, and a 1.2 M
DIBAH solution in toluene (5 ml, 6 mmol) was added to the reaction mixture under argon. Then
the mixture was refluxed with stirring for 6 h, after which methanol (2 ml), water (10 ml), and
conc. hydrochloric acid (2 ml) were sequentially added with vigorous stirring. The organic layer
was separated, and the aqueous layer was extracted with ethyl acetate (4 x 10 ml). The pooled
organic fractions were washed with water and saturated NH4Cl, dried over anhydrous Na2SO4,
and concentrated in vacuo. The resulting oily mass was purified by preparative HPLC to obtain
20(R)-3,20-dihydroxy-16α,17α-cyclohexano-19-norpregna-1,3,5(10)-triene 10a and 20(S)-3,20-
dihydroxy-16α,17α-cyclohexano-19-norpregna-1,3,5(10)-triene 10b as amorphous solids. After
recrystallization from aqueous ethanol, these products were obtained in yields of 44% and 14%,
respectively.
20(S)-3,20-Dihydroxy-16α,17α-cyclohexano-19-norpregna-1,3,5(10)-triene 10b: 14%,
m.p. 203–204˚C (at 136˚C the phase transition was observed for the initially formed crystals), 1H
NMR (DMSO-d6, 300 MHz): δ 0.75 (s, 3H, 18-CH3), 1.07–2.29 (m, 20H, rest of the 8 x CH2 and
4 x CH of steroid and additional rings), 1.10 (d, 3H, 21-CH3, J = 5.5 Hz), 2.69 (m, 2H, 6-CH2),
3.85 (m, 1H, 20-CH), 4.13 (d, 1H, 20-OH), 6.42 (bs, 1H, 4-CH, J = 2.2 Hz), 6.48 (dd, 1H, 2-CH,
13
J1 = 8.8 Hz, J2 = 2.2 Hz), 6.99 (d, 1H, 1-CH, J = 8.1 Hz), 8.96 (bs, 1H, 3-OH); C NMR, see
the Table 1b in text; HRMS: m/z 377.2552 [M + Na]+ (calcd for C24H34NaO2, 377.2451).
4.2.11 20(R,S)-3-Methoxy-20-hydroxy-16α,17α-cyclohexano-19-norpregna-1,3,5(10)-
triene 11
3-Methoxy-16α,17α-cyclohexano-19-norpregna-1,3,5(10)-trien-20-one (8, 0.37 g, 1.0
mmol) was taken in anhydrous tetrahydrofuran (15 ml). To this solution, LiAlH4 (0.08 g, 2.1
mmol) was carefully added, and then the reaction mixture was stirred for 48 h, after which
methanol (5 ml), saturated NH4Cl (10 ml), and conc. hydrochloric acid were sequentially added
dropwise. The organic layer was separated, and the aqueous layer was extracted with
24

ACCEPTED MANUSCRIPT
tetrahydrofuran (2 x 10 ml) and chloroform (2 x 10 ml). Each organic fraction was washed
separately with saturated NH4Cl. Then all extracts were pooled, dried over anhydrous Na2SO4,
and concentrated in vacuo. The resulting oily mass was recrystallized from a 1 : 1 toluene–
hexane mixture to obtain 20(R,S)-3-methoxy-20-hydroxy-16α,17α-cyclohexano-19-norpregna-
1,3,5(10)-triene 11.
1
Yield: 62%, m.p. 145–147˚C, H NMR (CDCl3, 300 MHz): δ 0.85 + 0.99 (s, 3H, 18-CH3,
R+S), 1.16–2.43 (m, 21H, rest of the 8 x CH2 and 4 x CH of steroid and additional rings and 20-
OH), 1.29 + 1.33 (d, 3H, 21-CH3, J = 6.61 Hz, R+S), 2.89 (m, 2H, 6-CH2), 3.81 (s, 3H, OCH3),
4.08 (q, 1H, 20-CH, J1 = 6.6 Hz, J2 = 5.9 Hz), 6.67 (s, 1H, 4-CH), 6.48 (dd, 1H, 2-CH, J1 = 8.1
13
Hz, J2 = 2.2), 7.22 (d, 1H, 1-CH, J = 8.1 Hz); C NMR (CDCl3, 75 MHz): δ (14.4), 14.9, 20.5
(20.1), (20.8), 22.2, 24.2, 24.3, (24.8), (26.2), 26.5, 28.1, 29.9, 30.0, (30.8), (32.2), (33.0), 33.2,
33.5, 36.2, (38.9), 39.1, (40.7), 43.7, (48.1), 48.8, (49.2), 49.6, (49.8), 49.9, 55.2, 73.5, (76.9),
111.4, 113.8, (126.2), 126.3, (132.9), 133.0, 138.1, 157.4; HRMS: m/z 369.2779 [M + H]+ (calcd
for C25H37O2, 369.2788).
4.2.12 Comparative study of the 20-keto group reduction of 3-methoxy-16α,17α-
cyclohexano-19-norpregna-1,3,5(10)-trien-20-one 8 and 3-methoxy-19-norpregna-1,3,5(10)-
trien-20-one 14 with lithium aluminum hydride and DIBAH.
a) LAH: Ketone (1 mmol) was taken in anhydrous THF (10 ml). To this solution, LiAlH4
(0.05 g, 1.3 mmol) was carefully added, and the reaction mixture was stirred for 4 h, after which
methanol (5 ml), saturated NH4Cl (10 ml), and conc. hydrochloric acid (1 ml) were sequentially
added dropwise with vigorous stirring. The organic layer was separated, and the aqueous layer
was extracted with ethyl acetate (4 x 10 ml) and chloroform (3 x 10 ml). The organic fractions
were washed with saturated NH4Cl, pooled, dried over anhydrous Na2SO4, and concentrated in
1
vacuo. The crude product was analyzed by H NMR (CDCl3, 300 MHz) comparing the
intensities of major and minor picks at δН 0.98 and δН 0.84 in the case of compound 8, and at δН
0.81 and δН 0.72 in the case of compound 14.
25

ACCEPTED MANUSCRIPT
b) DIBAH: Ketone (1 mmol) was taken in anhydrous toluene (10 ml). The system was
flushed with argon, and a 1.2 M DIBAH solution in toluene (2 ml, 2.4 mmol) was added to the
reaction mixture under argon. Then the mixture was stirred for 4 h, after which methanol (2 ml),
water (10 ml), ethyl acetate (10 ml), and conc. hydrochloric acid (1 ml) were sequentially added
dropwise with vigorous stirring. Further processing was the same as in the example (a) above.
4.2.13 3-Methoxy-16α,17α-cyclopropano-19-norpregna-1,3,5(10)-trien-20-one 12
Trimethylsulfoxonium iodide (1.55 g, 7.0 mmol) and sodium hydride (60 % suspension in
oil, 0.4 g, 10.0 mmol) were taken in a mixture of anhydrous DMSO (30 ml) and THF (10 ml),
and the solution was stirred until hydrogen evolution ceased. Then a solution of 3-methoxy-19-
norpregna-1,3,5(10),16-tetraen-20-one (5, 1.55 g, 5 mmol) in anhydrous THF (50 ml) was added.
The reaction mixture was stirred for 24 h at room temperature until the starting material
disappeared on TLC, poured into water (400 ml), and extracted with chloroform (3 x 50 ml). The
organic layers were pooled, washed with water and brine, dried over anhydrous Na2SO4, and
concentrated in vacuo. The resulting sticky solid was recrystallized from methanol with an
addition of a small amount of THF to obtain 3-methoxy-16α,17α-cyclopropano-19-norpregna-
1,3,5(10)-trien-20-one 12.
1
Yield: 59%, m.p. 138–140˚C (cf. lit. m.p. 137–139˚C (methanol) [28], H NMR (CDCl3,
300 MHz): δ 0.91 (m, 1H, 22α-CH), 1.01 (s, 3H, 18-CH3), 1.04 (m, 1H, 22β-CH), 1.22–2.47 (m,
12H, rest of the 4 x CH2 and 4 x CH of steroid ring), 2.00 (s, 3H, 21-CH3), 2.88 (m, 2H, 6-CH2),
3.80 (s, 3H, OCH3), 6.65 (s, 1H, 4-CH), 6.73 (d, 1H, 2-CH, J1 = 8.1 Hz), 7.21 (d, 1H, 1-CH, J =
8.1 Hz); 13C NMR (CDCl3, 75 MHz): δ 15.0, 17.2, 23.9, 25.7, 26.4, 27.1, 27.8, 29.7, 34.6, 37.0,
41.3, 44.3, 46.6, 47.7, 55.3, 111.5, 113.9, 126.2, 132.7, 137.8, 157.6, 208.0; HRMS: m/z
347.1985 [M + Na]+ (calcd for C22H28NaO2, 347.1982).
4.2.14 20(R,S)-3,20-Dihydroxy-16α,17α-cyclopropano-19-norpregna-1,3,5(10)-triene 13
20(R,S)-3,20-Dihydroxy-16α,17α-cyclopropano-19-norpregna-1,3,5(10)-triene 13 was
prepared from 3-methoxy-16α,17α-cyclopropano-19-norpregna-1,3,5(10)-trien-20-one (1, 0.32
26

ACCEPTED MANUSCRIPT
g, 1.0 mmol) and a 1.2 M DIBAH solution in toluene (4 ml, 4.8 mmol) as described for 20(S)-
3,20-dihydroxy-16α,17α-cyclohexano-19-norpregna-1,3,5(10)-triene 10b.
1
Yield: 32%, m.p. 210–217˚C*, H NMR (DMSO-d6, 300 MHz): δ 0.43 (m, 1H, 22-CH,
R+S), 0.53 + 0.68 (m, 1H, 22-CH, R+S), 0.83 + 1.08 (d, 3H, 21-CH3, R+S, J = 6.6 Hz, J = 5.5
Hz), 0.87 + 0.91 (s, 3H, 18-CH3, R+S), 0.90 – 2.27 (m, 12H, rest of the 4 x CH2 and 4 x CH of
steroid ring), 2.69 (m, 2H, 6-CH2), 4.07 (m, 2H, 20-CH and 20-OH), 6.42 (s, 1H, 4-CH), 6.49 (d,
1H, 2-CH, J = 8.7 Hz), 6.99 (d, 1H, 1-CH, J = 8.8 Hz), 8.99 (bs, 1H, 3-OH); 13C NMR, see the
Table 1b in text; HRMS: m/z 335.1978 [M + Na]+ (calcd for C21H28NaO2, 335.1982).
4.2.15 3-Methoxy-19-norpregna-1,3,5(10)-trien-20-one 14
3-Methoxy-19-norpregna-1,3,5(10),16-tetraen-20-one (5, 0.71 g, 2.3 mmol) and 10% Pd/C
(0.15 g) as the catalyst were taken in dioxane (30 ml) and hydrogenated at atmospheric pressure
until the starting material disappeared on TLC (visualization by spraying with aqueous KMnO4).
Then the catalyst was filtered off, the solvent was evaporated in vacuo, and the residue was
recrystallized from a 2 : 1 hexane–toluene mixture to obtain 3-methoxy-19-norpregna-1,3,5(10)-
trien-20-one 14.
Yield: 91%, m.p. 135˚C (cf. lit. m.p. 134–136˚C [29]), 1H NMR (CDCl3, 300 MHz): δ 0.68
(s, 3H, 18-CH3), 1.25–2.45 (m, 13H, rest of the 5 x CH2 and 3 x CH of steroid ring), 2.18 (s, 3H,
21-CH3), 2.64 (m, 1H, 17-CH), 2.87 (m, 2H, 6-CH2), 3.80 (s, 3H, OCH3), 6.66 (s, 1H, 4-CH),
6.74 (dd, 1H, 2-CH, J1 = 8.1 Hz, J2 = 2.3 Hz), 7.23 (d, 1H, 1-CH, J = 8.8 Hz); 13C NMR (CDCl3,
75 MHz): δ 13.5, 23.0, 24.2, 26.7, 27.8, 29.9, 31.5, 38.8, 39.1, 43.7, 44.5, 55.2, 55.8, 63.9, 111.6,
113.9, 126.3, 132.5, 138.0, 157.6, 209.4; HRMS: m/z 335.1969 [M + Na]+ (calcd for
C21H28NaO2, 335.1982).
4.2.16 3-Hydroxy-19-norpregna-1,3,5(10)-trien-20-one 15
A mixture of 3-methoxy-19-norpregna-1,3,5(10)-trien-20-one (14, 0.66 g, 2.2 mmol),
sodium iodide (0.3 g, 2.0 mmol), glacial acetic acid (10 ml), and conc. hydrobromic acid (5 ml)
was refluxed for 3.5 h and then poured into cold water. The precipitate that formed was filtered
27

ACCEPTED MANUSCRIPT
off, dried, dissolved in boiling toluene, and filtered hot through silica gel (0.5 g). After cooling,
the precipitate that formed was filtered off and recrystallized from methanol to obtain 3-hydroxy-
19-norpregna-1,3,5(10)-trien-20-one 15 as yellowish prisms.
1
Yield: 53%, m.p. 247–248˚C (cf. lit. m.p. 244–247˚C [30]), H NMR (DMSO- d6, 300
MHz): δ 0.57 (s, 3H, 18-CH3), 1.15–2.36 (m, 13H, rest of the 5 x CH2 and 3 x CH of steroid
ring), 2.10 (s, 3H, 21-CH3), 2.65 (m, 1H, 17-CH), 2.72 (m, 2H, 6-CH2), 6.46 (s, 1H, 4-CH), 6.53
(d, 1H, 2-CH, J = 8.1 Hz), 7.06 (d, 1H, 1-CH, J = 8.1 Hz), 8.96 (bs, 1H, 3-OH); 13C NMR
(DMSO- d6, 75 MHz): δ 13.1, 22.3, 23.6, 26.3, 27.3, 29.1, 31.1, 38.1, 38.4, 43.1, 43.6, 54.8,
62.7, 112.7, 114.9, 125.9, 130.1, 137.0, 154.9, 208.5; HRMS: m/z 299.2013 [M + H]+ (calcd for
C20H27O2, 299.2006).
4.2.17 20(R)-3,20-Dihydroxy-19-norpregna-1,3,5(10)-triene 16
20(R)-3,20-Dihydroxy-19-norpregna-1,3,5(10)-triene 16 was prepared from 3-hydroxy-19-
norpregna-1,3,5(10)-trien-20-one (15, 0.25 g, 0.93 mmol) and LiAlH4 (0.08 g, 2.1 mmol) as
described for 20(R)-3,20-dihydroxy-16α,17α-cyclohexano-19-norpregna-1,3,5(10)-triene 10a.
1
Yield: 40%, m.p. 229–230˚C (cf. lit. m.p. 232–233˚C [31]), H NMR (DMSO- d6, 300
MHz): δ 0.71 (s, 3H, 18-CH3), 1.02 (d, 3H, 21-CH3, J = 6.6 Hz), 1.05–2.30 (m, 14H, rest of the
5 x CH2 and 4 x CH of steroid ring), 2.70 (m, 2H, 6-CH2), 3.52 (m, 1H, 20-CH), 4.10 (d, 1H, 20-
OH, J = 5.5 Hz), 6.43 (s, 1H, 4-CH), 6.49 (dd, 1H, 2-CH, J1 = 8.8 Hz, J2 = 2.2 Hz), 7.03 (d, 1H,
1-CH, J = 8.8 Hz), 8.93 (bs, 1H, 3-OH); 13C NMR, see the Table 1b in text; HRMS: m/z
301.2166 [M + H]+ (calcd for C20H29O2, 301.2162); a sample for X-ray diffraction was obtained
from aqueous methanol, CCDC ID: 1495041.
4.2.18 20(R,S)-3,20-Dihydroxy-19-norpregna-1,3,5(10),16-tetraene 17
20(R,S)-3,20-Dihydroxy-19-norpregna-1,3,5(10),16-tetraene 17 was prepared from 3-
methoxy-19-norpregna-1,3,5(10),16-tetraen-20-one (5, 0.37 g, 1.2 mmol) and a 1.2 M DIBAH
solution in toluene (4 ml, 4.8 mmol) as described for 20(S)-3,20-dihydroxy-16α,17α-
cyclohexano-19-norpregna-1,3,5(10)-triene 10b but without a chromatographic step. The sticky
28

ACCEPTED MANUSCRIPT
solid that was obtained after the usual work-up was purified by recrystallization from aqueous
methanol to obtain 20(R,S)-3,20-dihydroxy-19-norpregna-1,3,5(10),16-tetraene 17.
1
Yield: 47%, m.p. 195–198˚C, H NMR (DMSO-d6, 300 MHz): δ 0.81 + 0.84 (s, 3H, 18-
CH3, R+S), 1.20–2.33 (m, 11H, rest of the 4 x CH2 and 3 x CH of steroid ring), 1.20 + 1.22 (d,
3H, 21-CH3, S+R, J = 6.6 Hz), 2.73 (m, 2H, 6-CH2), 4.22 (m, 1H, 20-CH), 4.44 + 4.49 (d, 1H,
20-OH, R+S, J = 5.3 Hz), 5.50 + 5.52 (bs, 1H, 16-CH, R+S), 6.44 (d, 1H, 4-CH, J = 2.2 Hz),
6.50 (dd, 1H, 2-CH, J1 = 8.1 Hz, J2 = 2.2 Hz), 7.02 (d, 1H, 1-CH, J = 8.8 Hz), 8.96 (bs, 1H, 3-
OH); 13C NMR, see the Table 1b in text; HRMS: m/z 321.1822 [M + Na]+ (calcd for
C20H28NaO2, 321.1825).
4.3
Molecular modeling
The three-dimensional structure of the human ER ligand-binding domain (LBD) was taken
from the Protein Data Bank (PDB: 1QKU). The structures of compounds 8-11, 13, and 15-were
generated using the SYBYL8.1 software suite [32]. The structures of the compounds and the
protein molecule were optimized in vacuum using the Powell energy minimization algorithm
with the Tripos force field [33]. The partial atomic charges were calculated by the Gasteiger–
Hückel method. Interactions of the steroids with the ER LBD binding site were modeled using
the molecular docking program DOCK 6.5 [34]. The modeling of the ER complex with the
natural ligand E2 was used as the control. No more than three candidates (most probable
complexes) were selected using the Dock scoring function. The further optimization of virtual
complexes was performed using the Amber 9.0 package [35] with the general AMBER99 force
field (GAFF) as described previously [36]. The optimization included the explicit solvation of
the complex in aqueous environment (TIP3P model) and productive dynamics at 300 K for 10 ps
with periodic boundary conditions (NTP ensemble). This computational molecular dynamics
experiment was used for the structure optimization and for the assessment of the LBD–ligand
binding energies by means of MM-PBSA [35]. The averaging was performed over 10
observations saved at regular time intervals during simulation. The solvation component was
calculated as a mean value estimated by the Poisson–Boltzmann method. The final structure was
selected based on the minimum value of the complex Gibbs energy calculated by the MM-PBSA
method taking into account electrostatic interactions, changes in van der Waals interactions, and
solvation effects.
4.4
Biology
29