Synthesis, X-ray crystal structures and biological activity of 16-amino-17-substituted-D-homo steroid derivatives

Article abstract of DOI:10.1016/S0039-128X(03)00097-7

D-Homo derivatives in the androstane and estrane series, 12-19, were synthesized by a fragmentation-cyclization reaction of 16-oximino-17-hydroxy-17-substituted derivatives 3-9, or by cyclization of the corresponding D-seco derivatives 20-26. The structur

Full text of DOI:10.1016/S0039-128X(03)00097-7

Steroids 68 (2003) 667–676
Synthesis, X-ray crystal structures and biological activity of
16-amino-17-substituted-D-homo steroid derivatives
Katarina M. Penov Gašia,∗, Dušan A. Miljkovic ,
a
´
a
a
´
ˇ
´
´
Ljubica D. Medic Mijacevic , Evgenija A. Djurendic ,
a
a
´
ˇ a
Srdjan Z. Stojanovic , Marija N. Sakac , Maja Dj. Djurendic ,
´
b
b
´
Slobodanka M. Stankovic , Dušan Lazar ,
c
c
´
ˇ
´
Silvana Andric , Radmila Kovacevic
a
Department of Chemistry, Faculty of Sciences, University of Novi Sad, 21000 Novi Sad, Trg Dositeja Obradovic´a 3,
Serbia and Montenegro, Yugoslavia
b
Department of Physics, Faculty of Sciences, University of Novi Sad, 21000 Novi Sad, Trg Dositeja Obradovic´a 4,
Serbia and Montenegro, Yugoslavia
c
Department of Biology, Faculty of Sciences, University of Novi Sad, 21000 Novi Sad, Trg Dositeja Obradovic´a 2,
Serbia and Montenegro, Yugoslavia
Received 30 August 2002; received in revised form 24 March 2003; accepted 31 March 2003
Abstract
D-Homo derivatives in the androstane and estrane series, 12–19, were synthesized by a fragmentation-cyclization reaction of 16-oximino-
17-hydroxy-17-substituted derivatives 3–9, or by cyclization of the corresponding D-seco derivatives 20–26. The structures were con-
firmed by X-ray analysis of compounds 12 and 16. Preliminary assessment of inhibitory effects of D-homo derivatives from androstane
series towards aromatase, 3␤-hydroxysteroid dehydrogenase (3␤-HSD), 17␣-hydroxylase/C17-20 lyase (P450c17) and 17␤-HSD indi-
cated much lower inhibitory potential compared to previously tested activity of another type of D-modified steroids, namely D-seco
derivatives. Also, assessment of potential antiestrogenic activity of derivatives from estrane series showed absence of such an
activity.
© 2003 Elsevier Inc. All rights reserved.
Keywords: D-Homo steroids; Synthesis; Inhibition of steroidogenic enzymes; X-ray analysis
1. Introduction
tert-butyl alcohol. We also describe the results of X-ray
diffraction analysis of D-homo derivatives, 12 and 16.
We previously reported the synthesis and anti-aromatase
activity of some D-modified steroids, i.e. several D-seco
derivatives of 5-androstene and 16-oximino-17-hydroxy-17-
substituted steroids [1]. With the aim of studying biological
activity of some other D-modified steroids, D-homo deriva-
tives possessing 16-amino functionality were prepared
using the fragmentation-cyclization reaction of steroidal
␣-hydroxy oximes [2], whose mechanism was investigated
in more detail in this paper. The objective of this work
was to prepare this type of D-homo steroids from the cor-
responding D-seco derivatives with potassium hydroxide
in boiling ethylene glycol and potassium tert-butoxide in
2. Experimental
2.1. General
Melting points were determined using a Büchi SMP 20
apparatus and are uncorrected. IR spectra were recorded on
a Perkin-Elmer 457 and Perkin-Elmer FT-IR 1725 X spec-
trometers. NMR spectra were taken on a Bruker AC 250E
spectrometer operating at 250 MHz (proton) and 62.9 MHz
(carbon), using standard Bruker software; the tetramethyl-
silane peak (δ 0.00) was used as reference in CDCl3 and
1
DMSO-d6 for H NMR, whereas the central carbon line of
∗
chloroform-d was set at 77.0 ppm, or that of DMSO-d6, at
Corresponding author. Tel.: +381-21-350-122; fax: +381-21-54-065.
E-mail address: sazvon@eunet.yu (K.M. Penov Gaši).
39.5 ppm, for 13C NMR. Mass spectra were recorded on a
0039-128X/$ – see front matter © 2003 Elsevier Inc. All rights reserved.
doi:10.1016/S0039-128X(03)00097-7

668
K.M. Penov Gaši et al. / Steroids 68 (2003) 667–676
Finnigan MAT 8230 (EI 70 eV), and a Kratos 80 instruments.
All reagents used were analytical grade commercially avail-
able substances.
NOH); 7.26 (5H, s, Ph). MS m/z: 391 (M+ − H2O), 272,
254, 213, 91. Anal. calcd. for C26H35NO3: C, 76.24; H,
8.61; found: C, 76.09; H, 8.46.
2.2.1.4. 3-Methoxy-16-oximino-17α-benzyl-17β-hydroxye-
stra-1,3,5(10)-triene (6). White crystals (0.88 g, 68%,
mp 202–204 ◦C). IR νmax (KBr) cm−1: 3550, 3380, 3080,
3020, 1665, 1575, 1050, 1030, 790, 770, 735, 700; 1H
NMR (DMSO-d6): 0.75 (3H, s, H-18); 3.69 (3H, s, CH3O);
2.80 (2H, ABq, J = 14.3 Hz, Bn); 4.15 (1H, s, OH); 6.60
(1H, d, J = 2.6 Hz, H-4); 7.12 (1H, dd, J = 8.4 Hz,
2.6 Hz, H-2); 7.16 (1H, d, J = 8.4 Hz, H-1); 7.17 (2H,
d, J = 7.5 Hz, H-2ꢀ, 6ꢀ, Ph); 7.22 (1H, t, J = 7.2 Hz,
H-4ꢀ, Ph); 7.30 (2H, t, J = 7.5 Hz, H-3ꢀ, 5ꢀ, Ph). 13C
NMR (DMSO-d6): 162.9 (C-16); 157.5 (C-3); 137.9; 132.5;
126.4; 125.9; 113.9; 83.2 (C-17); 55.3 (OCH3); 14.9 (C-18).
MS m/z: 405 (M+), 387 (M+ − H2O), 314, 268, 91. Anal.
calcd. for C26H31NO3: C, 77.00; H, 7.71; found: C, 76.62;
H, 7.52.
2.2. Chemical synthesis
2.2.1. General procedure for preparation of
compounds 3–6
To a suspension of metallic lithium (0.50 g, 72.04 mmol)
in absolute ether (20 ml) bromobenzene (9 ml, 86 mmol) in
dry THF (5 ml) was added with stirring at room tempera-
ture. The reaction mixture was vigorously stirred at room
temperature in a nitrogen atmosphere. When the reaction
was completed (after 2 h), ␣-picoline (9 ml, 94 mmol) in dry
THF (30 ml) was added and the content stirred at room tem-
perature for 1 h.
When benzyllithium was used as reagent, metallic lithium
(2.10 g, 302.5 mmol) was suspended in dry THF (100 ml) in
a nitrogen atmosphere and cooled to −10 ◦C. To the vigor-
ously stirred suspension dibenzylether (10.35 g, 52.2 mmol)
was added and reaction mixture stirred at −10 ◦C for 1 h.
To a cooled solution (−10 ◦C) of ␣-picolyllithium (i.e.
benzyllithium), prepared as above, a solution of compound
1 (1 g, 3.15 mmol), i.e. 2 (1 g, 3.2 mmol) in dry THF (30 ml)
was added gradually at cooling. The reaction mixture was
vigorously stirred at −10 ◦C for another 2 h and then poured
into water (300 ml). After several days the crude products
3–6 were filtered, washed with cold ether and dried. Pu-
rification of crude products was achieved by crystallization
from MeOH.
2.2.1.5. 3-Methoxy-16-oximino-17α-methyl-17β-hydroxye-
stra-1,3,5(10)-triene (8). 3-Methoxy-16-oximino-17␣-me-
thyl-17␤-hydroxyestra-1,3,5(10)-triene (8) was prepared
from compound 2 (1 g, 3.2 mmol), following the same pro-
cedure as for compounds 7 and 9 whose syntheses were
described earlier [1]. The crude product 8 (0.83 g, 78%) was
purified by crystallization from MeOH to give pure com-
pound 8 (0.80 g, 75.5%, mp 230–232 ◦C). IR νmax (KBr)
cm−1: 3600, 3100, 1615, 1575, 785. 1H NMR (DMSO-d6):
0.72 (3H, s, H-18); 1.20 (3H, s, C17-CH3); 2.77 (2H, m,
C-6); 3.68 (3H, s, OCH3); 4.45 (1H, s, OH); 6.60 (1H, d,
J = 2.5 Hz, H-4); 6.68 (1H, dd, J = 8.5 Hz, 2.5 Hz, H-2);
7.17 (1H, d, J = 8.5 Hz, H-1); 10.34 (1H, s, NOH). 13C
NMR (DMSO-d6): 166.5 (C-16); 157.5 (C-3); 137.9; 132.5;
126.4; 113.9; 80.6 (C-17); 55.3 (OCH3); 14.4 (C-18). MS
m/z: 329 (M+), 313 (M+ − H2O), 268, 228, 211, 179, 147.
Anal. calcd. for C20H27NO3: C, 72.92; H, 8.26; found: C,
73.44; H, 8.27.
2.2.1.1. 3β,17β-Dihydroxy-16-oximino-17α-picolyl-5-
androstene (3). White crystals (0.9 g, 70%, mp 235–
236 ◦C; lit. [3] yield 39%, mp 235 ◦C).
2.2.1.2. 3-Methoxy-16-oximino-17α-picolyl-17β-hydroxye-
stra-1,3,5(10)-triene (4). White crystals (1 g, 83%, mp
203–204 ◦C). IR νmax (KBr) cm−1: 3500, 3200, 1625, 1605,
1571, 1445, 1280, 1250, 1050, 980, 772, 757. 1H NMR
(CDCl3): 0.93 (3H, s, H-18); 2.98 (2H, ABq, J = 8.3 Hz,
CH2Py); 3.78 (3H, s, CH3O); 6.63 (1H, d, J = 2.6 Hz,
H-4); 6.72 (1H, dd, J = 8.5 Hz, 2.6 Hz, H-2); 7.17 (1H,
d, J = 8.5 Hz, H-1); 7.25–8.07 (4H, several signals, Py);
8.15 (1H, bs, NOH); 13C NMR (CDCl3): 165.9 (C-16);
157.60 (C-3); 137.7; 136.5; 132.2; 126.1; 124.6; 113.9;
83.5 (C-17); 55.2 (OCH3); 13.9 (C-18). MS m/z: 406 (M+);
388 (M+ − H2O). Anal. calcd. for C25H30N2O3: C, 73.86;
H, 7.44; found: C, 74.15; H, 7.55.
2.2.2. General procedure for preparation of D-homo
derivatives 12–14 and 16 from oximes 3–6 or from seco
derivatives 20 and 22 using KOH in ethylene glycol
Oximes 3–6 (1 mmol) or seco derivatives 20 and 22
(1 mmol) were dissolved in ethylene glycol (60 mmol) and
KOH (2 mmol) was added. The reaction mixture was vig-
orously stirred at the boiling temperature in a nitrogen
atmosphere. When the reaction was completed (1–3 h), the
mixture was diluted with water (200 ml), acidified with
2 M HCl (except for compounds 12 and 13) to pH 2–3
and extracted with ether (for compounds 12 and 14) or
filtered (for compounds 13 and 16). The organic phase
was washed with water, dried (Na2SO4) and the solvent
was removed under reduced pressure. The crude products
were purified by crystallization from MeOH (for 13, 14
and 16) or by silica gel column (for 12: benzene–AcOEt,
4:1).
2.2.1.3. 3β,17β-Dihydroxy-16-oximino-17α-benzyl-5-and-
rostene (5). White crystals (0.84 g, 65%, mp 253–
255 ◦C). IR νmax (KBr) cm−1: 3490–3240, 3025, 1605,
1
793, 745, 727, 699. H NMR (CDCl3): 0.89 (3H, s, H-18);
1.07 (3H, s, H-19); 2.81 (2H, ABq, J = 14.3 Hz, Bn); 3.55
(1H, m, H-3); 5.40 (1H, d, J = 5.1 Hz, H-6); 7.12 (1H, bs,

K.M. Penov Gaši et al. / Steroids 68 (2003) 667–676
669
2.2.2.1. 3β-Hydroxy-D-homo-16-amino-17-(2ꢀ-pyridyl)-5,
16-androstadien-17a-one (12). White crystals (52%, mp
205 ◦C after crystallization from MeOH). IR (KBr): 3550,
at room temperature for 20 h, and then at stirring poured
in icy water acidified with 2 M HCl. The precipitate was
separated by filtration, washed with water and dried. The
crude product (0.92 g) was crystallized from MeOH, afford-
ing pure compound 15 (0.86 g, 73%, mp 300–302 ◦C). IR
(KBr): 3450, 3325, 1732, 1606, 1563, 1492, 1468, 1412. 1H
NMR (CDCl3): 1.04 (3H, s, H-18); 1.12 (3H, s, H-19); 2.16
(3H, s, Ac); 4.40 (2H, bs, NH2); 4.60 (1H, m, H-3); 5.39
(1H, d, J = 6.1 Hz, H-6); 7.12 (2H, d, J = 6.76 Hz, H-2ꢀ,
H-6ꢀ, Ph); 7.18 (1H, t, J = 7.37 Hz, H-4ꢀ, Ph), 7.32 (2H,
t, J = 7.5 Hz, H-3ꢀ, H-5ꢀ, Ph). 13C NMR (CDCl3): 200.7
(C-17a); 157.3; 130.8; 126.7; 121.5 (C-6); 27.6 (Ac); 19.2
(C-19); 16.4 (C-18). MS m/z: 433 (M+), 373 (M+ −AcOH),
200, 159, 131, 43. Anal. calcd. for C28H35NO3: C, 77.56;
H, 8.14; found: C, 77.30; H, 8.43.
1
3250, 1590, 1565, 1530, 1470, 1420, 1050, 857, 800. H
NMR (CDCl3): 1.02 (3H, s, H-18); 1.10 (3H, s, H-19);
3.50 (1H, m, H-3); 5.36 (1H, d, H-6); 7.08–8.50 (4H, sev-
eral signals, Py). 13C NMR (CDCl3): 195.9 (C-17a); 158.8;
137.8 (C-16); 136.5; 132.3; 126.1; 124.7; 13.9 (CH3). MS
m/z: 392 (M+), 377, 132, 105, 71, 57, 44. Anal. calcd.
for C25H32N2O2: C, 76.49; H, 8.22; found: C, 76.58; H,
8.16.
2.2.2.2. 3-Methoxy-D-homo-16-amino-17-(2ꢀ-pyridyl)estra-
1,3,5(10),16-tetraen-17a-one (13). White crystals (79%
from compound 4, 75% from compound 20, mp 238–240 ◦C).
IR (KBr): 3550–3250; 2925, 2840, 1590, 1560, 1500, 1475,
1430, 1380, 1325, 1280, 1260, 960, 790. 1H NMR (CDCl3):
1.12 (3H, s, H-18); 3.78 (3H, s, OCH3); 2.92 (2H, bs, NH2);
6.64 (1H, d, J = 2.6 Hz, H-4); 6.72 (1H, dd, J = 8.5 Hz,
2.6 Hz, H-2); 7.24 (1H, d, J = 8.5 Hz, H-1); 7.05–8.43 (4H,
several signals, Py). 13C NMR (CDCl3): 200.7 (C-17a);
157.7 (C-3); 137.4; 126.4; 125.6; 55.2 (OCH3); 16.7 (C-18).
MS m/z: 388 (M+), 297 (M+ − CH3), 228, 160, 132, 105,
93. Anal. calcd. for C25H28N2O2: C, 77.29; H, 7.26; found:
C, 76.74; H, 7.11.
2.2.3. General procedure for preparation of D-homo
derivatives 17–19 from oximes 7–9 or seco derivatives
23–26 using tert-BuOK in tert-BuOH
Potassium (18–25 mmol) was dissolved in dry tert-butyl
alcohol (50–80 ml) at stirring in a nitrogen atmosphere.
When potassium-tert-butoxide was formed the corre-
sponding oxime (1 mmol) or seco derivative (1 mmol) in
tert-butyl alcohol (10 ml) was added. The reaction mixture
was vigorously stirred at reflux in a nitrogen atmosphere
for 2–4 h. After cooling, the mixture was diluted with wa-
ter (200 ml), neutralized with 6 M HCl and extracted with
CH2Cl2 or ether. The organic phase was washed with wa-
ter and dried (Na2SO4). The solvent was removed under
reduced pressure and the crude product was purified by
crystallization from MeOH (for compounds 17 and 18) or
by silica gel column for compound 19 (toluene–AcOEt,
15:1, 6:1).
2.2.2.3. 3β-Hydroxy-D-homo-16-amino-17-phenyl-5,16-an-
drostadien-17a-one (14). White crystals (84% from com-
pound 5, 75% from compound 22, mp 317–318 ◦C). IR
(KBr): 3450, 3225, 1630, 1600, 1540, 1420, 1060, 710,
650. 1H NMR (DMSO-d6): 1.03 (3H, s, H-18); 1.14 (3H, s,
H-19); 3.34 (1H, m, H-3); 5.32 (1H, m, H-6); 6.00 (2H, bs,
NH2); 7.01 (2H, d, J = 7.2 Hz, H-2ꢀ, H-6ꢀ, Ph); 7.17 (1H,
t, J = 7.2 Hz, H-4ꢀ, Ph); 7.30 (2H, t, J = 7.5 Hz, H-3ꢀ,
H-5ꢀ, Ph). 13C NMR (DMSO-d6): 198.3 (C-17a); 159.2;
136.2 (C-16); 131.1; 128.0; 125.7; 120.1 (C-6); 69.9 (C-3);
19.8 (C-19). MS m/z: 391 (M+), 376 (M+ − CH3), 200,
159, 131, 91. Anal. calcd. for C26H33NO2·H2O: C, 76.25;
H, 8.61; found: C, 76.13; H, 8.50.
2.2.3.1. 3β-Hydroxy-D-homo-16-amino-5,16-androstadien-
17a-one (17). White crystals (73% from compound 7,
88% from compound 23, mp 248–250 ◦C after crystal-
lization from MeOH). IR (KBr): 3413, 3215, 2960, 1646,
1
1549, 1463, 1365, 1198. H NMR (DMSO-d6): 0.96 (3H,
s, H-18); 1.03 (3H, s, H-19); 4.70 (2H, bs, NH2); 5.36 (1H,
d, J = 6.1 Hz, H-6). MS m/z: 315 (M+), 297 (M+ − H2O),
259, 69, 43. Anal. calcd. for C20H29NO2: C, 76,15; H, 9.27;
found: C, 75.67; H, 9.68.
2.2.2.4. 3-Methoxy-D-homo-16-amino-17-phenylestra-1,3,
5(10),16-tetraen-17a-one (16). White crystals (84%, mp
249–250 ◦C). IR (KBr): 3460, 3300, 3200, 1615, 1565,
1500, 1410, 1260, 1240, 1040, 780, 760, 705. 1H NMR
(CDCl3): 1.14 (3H, s, H-18); 3.78 (3H, s, CH3O); 2.96 (2H,
m, H-6); 4.46 (2H, bs, NH2). 13C NMR (CDCl3): 200.8
(C-17a); 137.4; 132.4; 126.4; 157.6 (C-3); 55.2 (OCH3);
16.6 (C-18). MS m/z: 387 (M+), 372, 227, 200, 159, 131.
Anal. calcd. for C26H29NO2: C, 80.58; H, 7.54; found: C,
80.87; H, 7.79.
2.2.3.2. 3-Methoxy-D-homo-16-aminoestra-1,3,5(10),16-te-
traen-17a-one (18). White crystals (77% from compound
8, 81% from compound 24, mp 236–240 ◦C after crystal-
lization from MeOH). IR (KBr): 3400, 3200, 1610, 1570,
1
1560, 1500, 1260, 1040. H NMR (CDCl3): 1.05 (3H, s,
H-18); 3.78 (3H, s, CH3O); 4.49 (2H, bs, NH2); 5.97 (1H,
s, H-17); 6.62 (1H, d, J = 2.7 Hz, H-4); 6.71 (1H, dd,
J = 9.4 Hz, 2.7 Hz, H-2); 7.72 (1H, d, J = 9.4 Hz, H-1).
13C NMR (CDCl3): 219.4 (C-17a); 157.3 (C-3); 137.5;
131.6; 126.6; 113.8; 55.1 (OCH3); 14.8 (C-18). MS m/z:
311 (M+). Anal. calcd. for C20H25NO2·H2O: C, 72.92; H,
8.26; found: C, 73.31; H, 8.21.
2.2.2.5. 3β-Acetoxy-D-homo-16-amino-17-phenyl-5,16-an-
drostadien-17a-one (15). To the solution of 3␤-hydroxy-
D-homo-16-amino-17-phenyl-5,16-androstadien-17a-one
(14, 1.00 g, 2.55 mmol) in dry pyridine (20 ml) acetic an-
hydride (10 ml) was added. The reaction mixture was left

670
K.M. Penov Gaši et al. / Steroids 68 (2003) 667–676
(C-16); 73.4 (C-3). MS m/z: 373 (M+ − AcOH), 282, 254,
237, 213, 185, 91. Anal. calcd. for C28H35NO3·EtOH: C,
75.12; H, 8.62; found: C, 75.35; H, 8.62.
2.2.3.3. 3β-Hydroxy-D-homo-16-amino-17-methyl-5,16-an-
drostadien-17a-one (19). White crystals (89% from com-
pound 9, 87% from compound 25, 89% from compound
26, mp 241 ◦C after crystallization from MeOH). IR (KBr):
1
3500, 3300, 2930, 1630, 1550, 1200, 1000, 790. H NMR
2.2.4.3. 3β-Hydroxy-16-cyano-17-benzyl-16,17-seco-5-and-
rosten-17-one (22). 3␤-Acetoxy derivative 21 (0.03 g,
0.07 mmol) was dissolved in methanol (1 ml) and KOH
(0.016 g, 0.29 mmol) was added. The reaction mixture was
heated at reflux for 1 h, and then poured in water, neutral-
ized with 6 M HCl and extracted with CH2Cl2. The organic
layer was dried with anhydrous Na2SO4 and after removal
of the drying agent evaporated to dryness. The obtained
crude product was chromatographed on a column of silica
gel (1 g, n-hexane–AcOEt, 9:1) affording pure compound
(DMSO-d6): 0.85 (3H, s, H-18); 0.93 (3H, s, H-19); 1.49
(3H, s, CH3, C-17); 3.25 (1H, m, H-3); 4.64 (1H, bs,
OH); 5.28 (1H, bs, H-6); 6.17 (2H, bs, NH2). 13C NMR
(DMSO-d6): 199.3 (C-17a); 159.1 (C-16); 141.05 (C-5);
120.1 (C-6); 99.2 (C-17); 69.9 (C-3); 19.1 (C-19); 16.4
(C-18); 8.5 (CH3, C-17). MS m/z: 330 (M+ +1), 329 (M+),
312 (M+ + 1 − H2O), 311 (M+ − H2O). Anal. calcd. for
C21H31NO2: C, 76.55; H, 9.48; found: C, 76.69; H, 9.79.
22 (0.027 g, 99.5%) as a colorless oil. IR νmax (film) cm−1
:
2.2.4. Synthesis of seco derivatives 20, 21, 22,
24, 27 and 28
3500–3250, 2250, 1720, 1610, 1500, 1270, 1000, 800,
1
730. H NMR (CDCl3): 0.92 (3H, s, H-18); 1.42 (3H, s,
2.2.4.1. 3-Methoxy-16-cyano-17-picolyl-16,17-secoestra-
1,3,5(10)-trien-17-one (20). 3-Methoxy-16-oximino-17␣-
picolyl-17␤-hydroxyestra-1,3,5(10)-triene (4, 1.00 g, 2.45
mmol), was dissolved in dry pyridine (10 ml) and TsCl
(3.00 g, 16.78 mmol) was added. The reaction mixture was
left at room temperature for 16 h and then poured in an
ice–water mixture (30 ml). The precipitate was separated
by filtration, washed with water and dried. The crude prod-
uct (0.90 g) was crystallized from MeOH, affording pure
compound 20 (0.81 g, 85%, mp 63–65 ◦C). IR νmax (KBr)
cm−1: 2251, 1783, 1609, 1573, 1467, 1288, 1257, 1094,
967, 793. 1H NMR (CDCl3): 1.25 (3H, s, H-18); 1.39 (3H,
s, H-18 from enol form); 3.78 (3H, s, CH3O); 4.09 (2H,
ABq, J = 7 Hz, CH2Py); 5.58 (1H, s, OH-17); 6.71 (1H, d,
J = 2.6 Hz, H-4); 7.00 (1H, dd, J = 8.5 Hz, 2.6 Hz, H-2);
7.24 (1H, d, J = 8.5 Hz, H-1); 7.07–7.60 (4H, several sig-
H-19); 2.25 (1H, s, OH); 3.81 (2H, ABq, J = 14.3 Hz,
Bn); 7.19–7.29 (5H, m, Bn). 13C NMR (CDCl3): 212.21
(C-17); 140.26; 134.11; 128.51; 126.87; 120.32; 119.03
(C-16); 77.54 (C-3); 18.60 (C-18); 15.36 (C-19). MS m/z:
391 (M+), 373 (M+ − H2O), 358 (M+ − H2O − CH3).
2.2.4.4. 3-Methoxy-16-cyano-17-methyl-16,17-secoestra-1,
3,5(10)-trien-17-one (24). From compound 8 (1.00 g,
3.03 mmol), after following the same procedure as for
compound 20, extraction with ether yielded crude product
of 24 (0.93 g). Crystallization from MeOH afforded pure
compound 24 (0.76 g, 81%, mp 89–91 ◦C). IR νmax (KBr)
1
cm−1: 2250, 1740, 1611, 1281, 1074, 967, 786. H NMR
(CDCl3): 1.20 (3H, s, C17-CH3); 2.98 (2H, m, H-6); 3.78
(3H, s, OCH3); 6.64 (1H, d, J = 2.7 Hz, H-4); 6.71 (1H,
dd, J = 8.0 Hz, 2.7 Hz, H-2); 7.20 (1H, d, J = 8.0 Hz,
H-1). MS m/z: 311 (M+), 293 (M+ − H2O), 268, 253, 228.
Anal. calcd. for C20H25NO2: C, 77.13; H, 8.09; found: C,
77.03; H, 8.11.
nals, Py); 8.15 (1H, d, H-3ꢀ, Py); 8.50 (1H, t, H-6ꢀ, Py). 13
C
NMR (CDCl3): 211.7 (C-17); 158.2 (C-3); 137.5; 131.0;
126.3; 121.2; 118.4 (C-16); 55.2 (OCH3); 15.2 (C-18).
MS m/z: 388 (M+); 250, 213, 110, 93. Anal. calcd. for
C25H28N2O2: C, 77.29; H, 7.26; found: C, 76.77; H, 7.27.
2.2.4.5. 3β-Hydroxy-16-cyano-17-phenyl-16,17-seco-5-an-
drosten-17-one (27). Potassium (0.085 g, 2.17 mmol) was
dissolved in dry tert-butyl alcohol (5 ml) at stirring in a ni-
trogen atmosphere, and compound 11 (0.030 g, 0.076 mmol)
was added. The reaction mixture was vigorously stirred at
reflux in a nitrogen atmosphere for 2 h. After cooling, the
mixture was diluted with water (30 ml), neutralized with
6 M HCl and extracted with CH2Cl2. The organic phase was
dried (Na2SO4), the solvent was removed under reduced
pressure and the crude product was purified by silica gel
column (toluene–AcOEt, 6:1) affording pure compound 27
(0.024 g, 84%, mp 144 ◦C after crystallization from MeOH;
lit. [1] yield 86%, mp 144 ◦C).
2.2.4.2. 3β-Acetoxy-16-cyano-17-benzyl-16,17-seco-5-and-
rosten-17-one (21). 3␤,17␤-Dihydroxy-16-oximino-17␣-
benzyl-5-androstene (5, 0.10 g, 0.24 mmol) was dissolved in
dry pyridine (2.4 ml) and Ac2O (1.20 ml) was added. The re-
action mixture was heated at reflux for 1 h, and then poured
in water (10 ml), acidified with 2 M HCl and extracted with
CH2Cl2. The organic layer was dried (Na2SO4) and after
removal of the drying agent evaporated to dryness. The ob-
tained crude product was chromatographed on a column of
silica gel (3 g, benzene–AcOEt, 15:1), and after crystalliza-
tion from EtOH afforded pure compound 21 (0.066 g, 63%,
mp 122 ◦C). IR νmax (KBr) cm−1: 2240, 1730, 1710, 1255,
1030, 735, 700. 1H NMR (CDCl3): 1.06 (3H, s, H-19);
1.39 (3H, s, H-18); 2.04 (3H, s, Ac); 3.89 (2H, ABq, J =
14.3 Hz, Bn); 4.62 (1H, m, H-3); 5.40 (1H, d, J = 6.1 Hz,
H-6); 7.13–7.39 (5H, m, Bn). 13C NMR (CDCl3): 212.0
2.2.4.6. 3-Methoxy-16-cyano-17-benzyl-16,17-secoestra-1,
3,5(10)-trien-17-one (28). From compound 6 (1.00 g,
2.46 mmol), following the procedure described for com-
pound 20, crude compound 28 (0.91 g) was obtained.
Crystallization from MeOH afforded pure compound 28
=
(C-17); 170.4 (C O, Ac); 139.1 (C-5); 121.2 (C-6); 118.9

K.M. Penov Gaši et al. / Steroids 68 (2003) 667–676
671
(0.80 g, 84%, mp 141–142 ◦C). IR νmax (KBr) cm−1: 2241,
1730, 1498, 1254, 1045, 872, 762, 701, 632. 1H NMR
(CDCl3): 1.37 (3H, s, H-18); 3.80 (3H, s, OCH3); 3.82
(2H, m, Bn); 6.65 (1H, d, J = 2.7 Hz, H-4); 6.74 (1H, dd,
J = 8.0 Hz, 2.7 Hz, H-2); 7.20–7.40 (6H, m, H-1 and 5H
from Bn). 13C NMR (CDCl3): 212.05 (C-17); 157.78 (C-3);
137.45; 134.10; 128.57; 126.94; 126.24; 119.09 (C-16);
113.46; 55.19; 15.37. MS m/z: 387 (M+), 296, 268, 227,
173, 91. Anal. calcd. for C26H29NO2: C, 80.56; H, 7.54;
found: C, 80.29; H, 7.83.
and treated isotropically in the final refinement as “riding
groups”. Weighted full-matrix least-squares refinement con-
verged at R = 0.042 for 1825 observed reflections; Rw
=
ꢀ
ꢀ
0.047, where Rw
=
(||F0| − |Fc|| · w1/2
/
(|F0| · w1/2),
w = 2.2010/[σ2(F0) + 0.000664F2]; 296 parameters re-
fined; maximum and minimum peak0height in final ꢀ␳ map
is 0.12 and −0.08 eÅ−3
.
2.3.2. 3-Methoxy-D-homo-16-amino-17-phenylestra-
1,3,5(10),16-tetraen-17a-one (16)
2.3. X-ray crystal studies of 12 and 16
2.3.2.1. Crystal data. C26H29NO2, M = 387.5, mono-
clinic, space group P21, a = 18.559(6) Å, b = 8.419(2) Å,
c = 6.862(4) Å, V = 1056.2(7) Å3, Z = 2, Dcalcd
=
2.3.1. 3β-Hydroxy-D-homo-16-amino-17-(2ꢀ-pyridyl)-5,
16-androstadien-17a-one (12)
1.218 g cm−3, µ(Cu K␣ radiation, λ
=
1.54178 Å)
= 5.9 cm−1, F(000) = 416. The crystal structure has been
deposited at the Cambridge Crystallographic Data Centre
(CCDC 204016).
2.3.1.1. Crystal data. C25H32N2O2 + CH4O, M
=
424.58, orthorhombic, space group P212121, a = 31.323(9)
Å, b = 9.768(1) Å, c = 7.535(1) Å, V = 2305.4(8) Å3,
Z = 4, Dcalcd = 1.223 g cm−3, µ(Cu K␣ radiation,
λ = 1.54178 Å) = 5.3 cm−1, F(000) = 920. The asym-
metric unit contains 1:1 steroid and methanol molecules.
The crystal structure has been deposited at the Cambridge
Crystallographic Data Centre (CCDC 204015).
2.3.2.2. Crystallographic measurements.
A
transpar-
ent colorless crystal of the dimensions 0.32 mm ×
0.10 mm × 0.15 mm, was mounted on a PHILIPS PW
1100 four-circle diffractometer. Data were collected using
graphite-monochromated Cu K␣ radiation. The system-
atic absences correspond to those of the monoclinic space
group P21. The refined lattice parameters were obtained
by a least-squares procedure from the setting angles of 20
reflections with 15◦ < θ < 19.5◦. The intensities were col-
lected in the range 1◦ < θ < 69.99◦; ω/2θ scan method; h:
−22 to 22, k: 0–10, l: 0–8; three standard reflections were
measured every hour. Of 1629 independent non-zero reflec-
tions, 1618 with F > 2σ(F) were considered as observed
and used in calculations. Data were corrected for Lorentz
and polarization effects and for absorption.
2.3.1.2. Crystallographic measurements. A transparent
colorless crystal crystallized from methanol, of the dimen-
sions 0.32 mm × 0.10 mm × 0.40 mm, was mounted on an
Enraf-Nonius CAD-4 diffractometer. Data were collected
using graphite-monochromated Cu K␣ radiation. The sys-
tematic absences (h 0 0 when h = 2n, 0 k 0 when k = 2n,
0 0 l when l = 2n) correspond to those of the orthorhom-
bic space group P212121. The refined lattice parameters
were obtained by a least-squares procedure from the setting
angles of 25 reflections with 24◦ < θ < 52◦. The inten-
sities were collected in the range 1◦ < θ < 65◦; ω − θ
scan method; h: 0–36, k: 0–11, l: 0–8; three standard re-
flections were measured every hour, variation of intensities
2%. Of 2275 independent non-zero reflections, 1825 with
F > 3σ(F) were considered as observed and used in cal-
culations. Data were corrected for Lorentz and polarization
effects and for absorption.
2.3.2.3. Structure analysis. The structure was solved by
direct methods using SHELXS86 [5] and refined using
SHELXL97 [6]. All non-H atoms of steroid molecule were
found in E-map. H atoms were generated in idealized posi-
tions and were then refined either as riding with the atoms
to which they are bonded (CH, CH2, NH2) or as parts of
rigid CH3 groups. The isotropic temperature factors of the
H atoms were fixed at either 1.2 × Ueq (CH, CH2) and
1.3 × Ueq (NH2) or 1.5 × Ueq (CH3) of their carrier atoms.
Weighted full-matrix least-squares refinement converged at
Rw = 0.0435 for 1618 observed reflections; Rw = 0.0439
for all reflections. Weighting scheme w = 1/[σ2(F02) +
(0.0862P)2+0.0335P], where P = (F02+2Fc2)/3. Two hun-
dred ninety-six parameters refined; maximum and minimum
2.3.1.3. Structure analysis. The structure was solved by
direct methods using SHELXS86 [5] and refined using
SHELXL97 [6]. All non-H atoms of steroid molecule
were found in E-map. The isotropic refinement gave R =
0.21. Since the R-factor did not converge satisfactorily
(anisotropic refinement gave 0.18), the investigation of
ꢀF-map revealed two unexpected peaks, which were at-
tributed to the oxygen and carbon from methanol molecule.
The repeated refinement, anisotropic for all non-H atoms
from steroid and O from methanol, and isotropic for C from
methanol gave a satisfactory R-factor of 0.092. In this stage
of refinement the H atoms were generated, except those
bound to N(1), O(1), and O. They were found in ꢀF-map
peak height in final ꢀ␳ map are 0.152 and −0.240 eÅ−3
.
2.4. Biological tests
For a preliminary assessment of the enzyme inhibition ef-
fects, compounds 12, 14 and 17 were tested towards possible
inhibition of the activities of aromatase, 3␤-hydroxysteroid

672
K.M. Penov Gaši et al. / Steroids 68 (2003) 667–676
dehydrogenase (HSD), 17␣-hydroxylase/C17-20 lyase
(P450c17) and 17␤-HSD.
conversion of ꢀ4-androstenedione to testosterone was un-
changed when compared to the corresponding control sam-
ples, the obtained changes in the conversion of progesterone
to testosterone were attributed to the activity of P450c17 in
the investigated samples [8]. Aliquots (106 cells/0.3 ml) of
crude suspensions of Leydig cells were added to plastic tubes
containing 0.1 ml of steroid substrates (progesterone or an-
drostenedione, 2 ␮M) and 0.1 ml M199 contained 0.1% BSA
(M199-BSA) so that final incubation volume was 0.5 ml.
Following incubation for 2 h in a shaking water bath (34 ◦C)
oscillating at 100 cycles/min under an atmosphere of 95%
O2–5% CO2 all tubes were centrifuged for 10 min at 1500×
g and individual samples of supernatants were stored at
−20 ◦C prior to measurement of testosterone (T) by RIA.
2.4.1. Chemicals
Antiestradiol serum no. 244, antiprogesterone serum no.
337, and antitestosterone serum no. 250 were kindly supplied
by Dr. G.D. Niswender (Colorado State University, CO,
USA). Pregnant mares’ serum gonadotrophin was obtained
from the Veterinary Institute Subotica (Serbia and Montene-
gro), [1,2,6,7-3H(N)]estradiol, [1,2,6,7-3H(N)]progesterone,
and [1,2,6,7-3H(N)]testosterone were obtained from New
England Nuclear (Belgium), NADPH, NAD, testosterone,
estradiol, progesterone, ꢀ4-androstenedione, collagenase
(type I), trypan blue, bovine serum albumin (BSA, fraction
V) were obtained from Sigma (St. Louis, MO). All other
reagents were of analytical grade.
2.4.4. Hormone assays
Estradiol, progesterone and testosterone were estimated
by the corresponding RIA. Each experiment was run in
a single assay. RIA for estradiol had the following spec-
ification: sensitivity 4 pg/tube, intra-assay and inter-assay
coefficients of variation were 8.5 and 11.3%, respectively.
Cross-reactivity (defined as percent of the ratio between the
dose of unlabeled estradiol and the dose of heterologous
steroid required to produce 50% inhibition of the binding
of 3H estradiol tracer) of antiestradiol serum no. 244 to
testosterone was 0.0068%, while to estrone was 0.28%. We
assessed also the possible cross-reactivity of antiestradiol
serum to 5-androstene derivatives investigated in our study.
We found that cross-reactivity averaged 0.004%. RIA for
progesterone had the following characteristics: sensitivity,
6 pg/tube; intra- and inter-assay coefficients of variation
were 6.8 and 10.7%, respectively. Since antiprogesterone
serum showed cross-reactivity with pregnenolone (relative
binding affinity was 1.4%) additional tubes without suspen-
sion of Leydig cells were also incubated and progesterone
production was calculated after substraction of these blanks
for pregnenolone. Antitestosterone serum used in androgen
assay showed a high cross-reactivity with dihydrotestos-
terone (DHT), and therefore assay values are referred to as
T + DHT levels. RIA for androgen had the following char-
acteristics: sensitivity, 6 pg/tube; the intra- and inter-assay
coefficients of variation were 5.8 and 7.5%, respectively.
For estimation of progesterone and androgen production
by RIA, samples were diluted 152 times and in such con-
dition there was no cross-reactivity of tested compounds
with antiprogesterone of antitestosterone serum used in the
corresponding RIA.
2.4.2. Aromatase activity
Anti-aromatase activity of the synthesized compounds
was tested according to the procedure described earlier [1].
In brief, aromatase activity was determined in denucleated
rat ovarian fraction using testosterone as a substrate and mea-
suring estradiol concentration by radioimmunoassay (RIA).
The given compound was added in a single concentration
(50 ␮M) to the incubation mixture containing denucleated
ovarian fraction, 1 mmol NADPH, 0.1 M phosphate buffer
and 50 or 500 nM of testosterone as substrate (sub-saturated
and saturated concentration, respectively; the estimated Km
for testosterone was 49.17 nM and Vmax 5.76 pmol/min mg
protein).
2.4.3. Enzyme activities of Leydig cells
The activities of 3␤-HSD, P450c17 and 17␤-HSD were
measured by incubating a crude suspension of Leydig cells
in the presence of corresponding substrate (pregnenolone,
progesterone, or ꢀ4-androstenedione) and the given com-
pound added in a single concentration (20 ␮M). Details con-
cerning preparation of crude suspension of Leydig cells and
assay procedures were published elsewhere [7,8]. Briefly,
adult male rats were sacrificed, testes were quickly removed,
decapsulated and dissociated by collagenase (1.2 mg/ml) as
previously described [7]. Trypan blue exclusion was used to
assess cell viability. The 3␤-HSD activity was estimated by
following the conversion of pregnenolone to progesterone
[8]. The incubation in the final volume of 0.5 ml was con-
ducted in the presence of crude suspension of Leydig cells
(106 cells/0.3 ml M199 contained 0.1% BSA) and 25 ␮M of
pregnenolone. All samples were incubated in triplicate. The
contents of tubes were gassed with 95% O2–5% CO2 and in-
cubation was carried out under shaking at 34 ◦C for 10 min.
The tubes were centrifuged 10 min at 1500 × g at 4 ◦C and
the supernatant was assayed for progesterone by RIA. In
such conditions, the estimated Km for pregnenolone was
9.0 ␮M. P450c17 and 17␤-HSD activities were estimated
by following the conversion of progesterone to testosterone
and ꢀ4-androstenedione to testosterone, respectively. If the
2.4.5. Uterotrophic assay
Immature Wistar strain female rats (21–23 days old) were
randomly divided into groups of six to eight animals each.
The animals were treated by subcutaneous injection once a
day for 3 days with 0.1 ml of a solution of the test com-
pound in olive oil, either solely or in combination with estra-
diol benzoate (EB). The control group obtained the vehicle
only. The total administered amounts of tested compounds

K.M. Penov Gaši et al. / Steroids 68 (2003) 667–676
673
were 5 or 25 mg/kg body weight, whereas the EB dose was
0.03 mg/kg body weight. The animals were killed on the
fourth day. The uteri were removed, dissected free of adher-
ing fat and blotted dry after expulsion of uterine fluid and
the wet weight was recorded.
Percent agonist and antagonist activity in immature rat
uterine weight assays were calculated from the ratio of val-
ues recorded in treated and control animals thus:
imino ketones 1 or 2, whereupon a good yield of additional
products 3–9 was obtained (Scheme 1) [1,9]. In this paper
we describe a modified procedure for preparation of com-
pound 3 (from 1), whereupon a better yield (70%) was ob-
tained (39%) [3]. Compounds 4–6 were fully characterized
by the corresponding spectral data.
The second step was a “one-pot” fragmentation-cyclization
reaction of the oximino-alcohols 3–6 with potassium hy-
droxide in refluxing ethylene glycol. D-Homo derivatives
12–14 and 16 were obtained starting from compounds 3
to 6 in high yields and for the first time characterized by
spectral data (Scheme 1). However, 17-methyl substituted
derivatives 7 and 8, as well as 17-ethyl substituted derivative
9, by the reaction with potassium tert-butoxide in tert-butyl
alcohol at reflux temperature afforded the corresponding
novel D-homo derivatives 17, 18 and 19 in high yields.
The compound 14 was selectively O-acetylated to give the
corresponding 3␤-acetoxy derivative 15.
(C − A)
% agonism =
× 100
(B − A)
(B − D)
% antagonism =
× 100
(B − A)
where A, B, C and D are uterine wet weights, corrected for
differences in body weight, i.e. (mg/100 g body weight) for
vehicle alone, EB, test compound alone, or test compound
plus EB groups, respectively.
To prove the postulated mechanism (Scheme 2) [2], i.e.
the assumption that D-homo compounds are formed from
the intermediate seco-cyano ketones, we attempted the syn-
thesis of the mentioned D-homo derivatives starting from
the corresponding D-seco compounds 20–27.
3. Results and discussion
3.1. Chemistry
The preparation of D-seco derivatives 22, 23, 25 and
26 from 5, 7 and 9 has been described earlier [1,4,9]. In
this paper we synthesized D-seco derivative 22 in a better
yield than in previous paper [9] by using acetic anhydride
as fragmentation reagent followed by basic hydrolysis of
The first step in the synthesis of the D-homo steroid
derivatives 12–19 consisted of a stereospecific addition of
␣-picolyllithium, benzyllithium, methylmagnesiumiodide or
ethylmagnesiumiodide to the C-17 carbonyl group of ox-
20 R2 series, R3= CH2Py
21, R1 series, R3= Bn (3β-OAc)
22, R1 series, R3=Bn (3β-OH)
23, R1 series, R3= Me (3β−OTs)
24, R2 series, R3= Me
O
j
R3
C
C
N
R1 or R2
25, R1 series, R3=Et
26, R1 series, R3=Et (3β-OTs)
27, R1 series, R3=Ph
28, R2 series, R3=Bn
O
D
OH
R3
O
R3
a or b or c or d
C
C
C
D
D
N
NH2
N
H
OH
R1 or R2
R1 or R2
R1 or R2
3, R1 series, R3= CH2Py
4, R2 series, R3= CH2Py
5, R1 series, R3= Bn
6, R2 series, R3= Bn
7, R1 series, R3= Me
8, R2 series, R3= Me
9, R1series, R3= Et
12, R1 series, R3= Py
13, R2 series, R3= Py
14, R1 series, R3= Ph
15, R1 series, R3=Ph (3β-OAc)
16, R2 series, R3=Ph
17, R1 series, R3=H
1, R1= The residual part of
3β-hydroxy-16-oximino-
5-androstene-17-one
i
2, R2= The residual part of
3-methoxy-16-oximino-
estra-1,3,5(10)-triene-17-one
18, R2 series, R3=H
10, R1 series, R3=H
11, R1 series, R3=Ph
19, R1 series, R3=Me
Scheme 1. Reagents: (a) ␣-PyCH2Li, THF; (b) BnLi, THF; (c) MeMgI, ether, THF; (d) EtMgI, ether, THF; (e) KOH, (CH2OH)2; (f) tert-BuOK,
tert-BuOH; (g) TsCl, Py; (h) Ac2O, Py; EtONa, EtOH; (i) Ac2O, Py; (j) KOH, MeOH.

674
K.M. Penov Gaši et al. / Steroids 68 (2003) 667–676
O
O
B
B
3- 9
CH2R
CHR
C
N
C
N
20 - 26, 28
O
H
O
H
C
C
H2O
R
R
12 - 19
C
C
NH
N
R = Ph, Py, Me or H
Scheme 2. Proposed mechanism for the fragmentation-cyclization reaction of 3–9.
the obtained 3␤-acetoxy derivative 21. In the present work
fragmentation of oximino alcohol 6 with p-toluenesulphonyl
chloride in pyridine was carried out to afford 28 in a signif-
icantly better yield (84%). Novel D-seco derivatives 20 and
24 were obtained by fragmentation of 4 and 8 with the aid of
p-toluenesulphonyl chloride. As could be expected, D-seco
compounds 20 and 22 in boiling ethylene glycol in the pres-
ence of potassium hydroxide afforded really the correspond-
ing D-homo derivatives 13 and 14 (Scheme 1). However,
compounds 23 and 24 with potassium tert-butoxide in boil-
ing tert-butyl alcohol gave high yields of the corresponding
D-homo derivatives 17 and 18, which were isolated by di-
rect crystallization. Also, the reaction of the tert-butanolic
solution of compound 25 with potassium tert-butoxide at the
boiling temperature afforded the desired D-homo derivative
19, which was also formed under the identical treatment of
3␤-p-toluenesulphonyloxy D-seco derivative 26.
The assumption that for the formation of a D-homo deriva-
tive is needed the presence of the ␣-hydrogen atom at the
C-17 substituent was proved in this paper by the fact that
no D-homo derivatives were formed when compounds 11
and 27 (17␣-phenyl; Scheme 1), synthesized in our previ-
ous paper [1], were used under the same reaction conditions.
Namely, using tert-butoxide anion compound 11 was trans-
formed into D-seco derivative 27 (84%), but compound 27
failed to react further under these reaction conditions. In
our previous paper [1] the compound 27 (yield 86%) was
prepared from oximino alcohol 11 using TiCl3 in acidic
medium. Finally, when compound 10 (having no C-17 sub-
stituent) [10] was used, no reaction took place either.
The structures of compounds 12–19 were proposed on
the basis of chemical and spectral evidences. The X-ray
structural analysis of 12 (androstene series) and 16 (es-
trane series) unambiguously proved the structure of these
new types of D-homo steroids. Fig. 1 shows the stereo-
scopic view of compounds 12 and 16, respectively. The dis-
tances and bond angles have the expected values for these
kind of steroids. Packing arrangement shows the presence
of relatively strong hydrogen bonds in both compounds
(Table 1). The intramolecular N1 · · · N2 hydrogen bond in
the 12 reduces the potential for rotation about the C17–C20
bond. The rotation is hindered by rather high energy barrier
of ∼600 kJ/mol, as showed by the method of “rigid rotor
approximation” [11] which may have some bearing on the
observed inhibitory effects of this compound.
3.2. Biological properties
Screening assay procedures were used to assess the
potential inhibitory effects of D-homo derivatives of an-
drostane series towards aromatase, 3␤-HSD, P450c17 and
17␤-HSD activities. The results presented in Table 2 show
Table 1
Hydrogen bond network for compounds 12 and 16
Compound
D–H · · · A
D–H (Å)
H · · · A (Å)
D · · · A (Å)
D–H–A (◦)
Symmetrical operations
12
N1–H1 · · · N2
N1–H2 · · · O2
O1–H · · · O3
O3–H · · · O1
0.844(3)
0.857(2)
0.821(2)
0.830(3)
2.057(3)
2.248(2)
1.853(4)
1.843(3)
2.691(4)
3.033(3)
2.664(5)
2.659(4)
131.4(2)
152.1(2)
169.1(2)
167.1(3)
x, y, z
x, y, 1 + z
x, y, z
−1/2 − x, −y, 1/2 + z
16
N–H2 · · · O2
0.810(3)
2.170(4)
2.929(5)
156.0(3)
x, y, −1 + z

12
16
Fig. 1. Perspective view of the molecular structure of compounds 12 and 16.

676
K.M. Penov Gaši et al. / Steroids 68 (2003) 667–676
Table 2
Table 4
Inhibitory effects of 5-androstene derivatives on the aromatase activity in
the denucleated fraction of ovaries from PMSG pretreated rats
Agonistic and antagonistic effects of estratriene derivatives
Compound
Dose (mg/kg)
Agonistic effect (%)
Antagonistic
effect (%)
Compound (50 ␮M)
Aromatase activity
6
5
25
15.15
74.98
13.79
−11.27
0.05 ␮M testosterone
0.5 ␮M testosterone
Percent of inhibition vs. control
16
20
24
28
5
25
20.40
30.79
−63.89
12
14
17
39.98
18.42
2.22
5.81∗∗
1.81∗
0.61
29.84
16.23
3.61
0.42∗∗
4.61
0.41
−118.77
5
25
0.28
7.04
12.20
14.69
Results are expressed as the percentage of inhibition vs. control. Numbers
represent mean S.E.M. of six replicates. Significance: ∗∗P < 0.005;
∗P < 0.05 vs. control (Mann–Whitney non-parametric test).
5
25
10.69
41.03
2.72
0.03
5
25
−7.92
−1.46
4.67
17.79
that compound 12 expressed weak inhibitory effect toward
aromatase. However, anti-aromatase activity of this com-
pound was much lower than the activity of the previously
tested D-seco and 16-oximino derivatives of 5-androstene
[1] which are intermediate compounds in the synthesis of
D-homo derivatives described in this study. Also, the results
presented in Table 3, show that compound 12 induced signif-
icant inhibition of the activities of 3␤-HSD and 17␤-HSD.
Inhibitory effect on the conversion of progesterone to testos-
terone cannot be taken as an index of depressed P450c17
activity, since simultaneously strong inhibition of 17␤-HSD
occurred. However, such inhibition cannot be excluded. On
the other hand, compound 14 expressed inhibitory effect
only towards P450c17 (Table 3), while the other enzymes
were not affected. Furthermore, compound 17 expressed no
activity toward tested steroidogenic enzymes.
The representatives of the newly synthesized derivatives
of estrane were subjected to biological testing. In these ex-
periments, typical representatives of all involved classes of
synthesized compounds were chosen (Table 4).
Selected compounds were subjected to the determina-
tion of residual estrone activity, by use of the uterotrophic
method, in which case the average weight of the uterus has
been reduced by D-seco compounds 20 and 28. By the an-
tiuterotrophic method possible antiestrogen properties were
tested, in which case compounds 20 and 28 demonstrated
weak antagonistic effects.
Acknowledgements
This research was supported by the Ministry of Science,
Technologies, and Development of the Republic of Serbia
(grant no. 1896).
References
[1] Penov Gaši K, Stankovic´ S, Csanadi J, Djurendic´ E, Sakacˇ M,
Medic´-Mijacˇevic´ Lj, et al. New D-modified androstane derivatives
as aromatase inhibitors. Steroids 2001;66:645–53.
[2] Penov Gaši K, Miljkovic´ D, Medic´-Mijacˇevic´ Lj, Djurendic´ E,
Petrovic´ J, Pejanovic´ V, et al. A novel fragmentation-cyclization
reaction of steroidal ␣-hydroxy oximes. Tetrahedron Lett
1998;39:9759–60.
[3] Miljkovic´ D, Gaši K. 3␤-Acetoxy-17-pycolinylidene-5-androsten-
16-one, the key intermediate in the 21,27-bisnorsolanidine synthesis.
Bull Chem Soc Beograd 1981;46:263–8.
[4] Penov Gaši K, Stojanovic´ S, Medic´-Mijacˇevic´ Lj, Molnar Gabor D,
Pejanovic´ V, Sakacˇ M. The modified methods for obtaining some
biologically active D-seco derivatives of 5-androstene. Acta Period
Technol 2001;32:139–44.
[5] Sheldrick GM. Acta Crystallogr 1990;A46:467–73.
[6] Sheldrick GM. SHELXL 97. University of Göttingen, Germany,
1997.
[7] Kovacˇevic´ R, Saracˇ M. Bromocriptine-induced inhibition of
hydroxylase/lyase activity of adult rat Leydig cells. J Steroid Biochem
Mol Biol 1993;46:841–5.
[8] Maric´ D, Kostic´ T, Kovacˇevic´ R. Effect of acute chronic
immobilization stress on rat Leydig cells steroid genesis. J Steroid
Biochem Mol Biol 1996;58:351–5.
[9] Miljkovic´ D, Penov Gaši K, Djurendic´ E, Sakacˇ M, Medic´-Mijacˇevic´
Lj, Pejanovic´ V, et al. A novel rearrangement of steroidal ␣-hydroxy
oximes. Tetrahedron Lett 1997;38:4683–4.
[10] Miljkovic´ D, Petrovic´ J, Stajic´ M, Miljkovic´ M. The Beckmann
fragmentation reaction of some ␣-hydroxy ketoximes. J Org Chem
1973;38:3585–8.
Table 3
Inhibitory effects of 5-androstene derivatives on the activity of steroido-
genic enzymes in crude suspension of adult rat Leydig cells in vitro
Compound 3␤-HSD activity
(20 ␮M)
P450c17 activity
17␤-HSD activity
Percent of inhibition vs. control
12
14
17
29.77
15.50
–
3.49∗
3.63
34.49
31.72
14.22
2.28∗
4.05∗
4.00
72.89
16.88
–
3.92∗
1.67
Results are expressed as the percentage of inhibition vs. control. Numbers
represent mean S.E.M. of 6–18 replicates. Significance: ∗P < 0.005 vs.
control (Mann–Whitney non-parametric test).
[11] Serena Software. PCMODEL, Version 4.0. Serena Software,
Bloomington, IN, USA, 1989.