Probing the active site of aromatase with 2-methyl-substituted androstenedione analogs

Article abstract of DOI:10.1016/S0039-128X(03)00089-8

To gain insight into the spatial nature of the androstenedione (AD) binding (active) site of aromatase in relation to the catalytic function of the enzyme, we synthesized 2,2-dimethylAD (4), 2β- and 2α-methylADs (5 and 6), 19-oxygenated derivatives of compounds 4 and 6, and 2-methyleneAD (17), and we then tested their inhibitory activity as well as their aromatase reaction (aromatization for 2-methyl and 2-methylene analogs or 19-oxygenation for 2,2-dimethyl steroids) with human placental aromatase. 2-Methyl and 2-methylene steroids 5, 6, and 17 were good competitive inhibitors of aromatase (K_i=22-68nM), but less effective compared to the 2,2-dimethyl analog 4 (K_i=8.8nM), indicating that a combination of 2β- and 2α-methyl moieties is essential for the formation of a thermodynamically stable inhibitor-aromatase complex. A series of 2α-methyl steroids were good substrates for aromatase, whereas 2β-methyl steroid 5 was an extremely poor substrate, and a series of 2,2-dimethyl steroids did not serve as substrate, suggesting that a 2β-methyl moiety of the 2,2-dimethyl and 2β-methyl steroids would prevent the aromatase reaction probably due to steric hindrance in each case. The 2-methylene compound 17 was also aromatized to produce 2-methylestrogen with a low conversion rate where the 1,4-diene structure may have been created before the C₁₀-C₁₉ bond cleavage. Kinetic analysis of the aromatization of androgens revealed that a good substrate was not essentially a good inhibitor for aromatase.

Full text of DOI:10.1016/S0039-128X(03)00089-8

Steroids 68 (2003) 503–513
Probing the active site of aromatase with 2-methyl-substituted
androstenedione analogs
∗
Mitsuteru Numazawa , Yoko Watari, Keiko Yamada, Nao Umemura, Wakako Handa
Tohoku Pharmaceutical University, 4-1 Komatsushima-4-chome, Aobaku, Sendai 981-8558, Japan
Received 4 March 2003; received in revised form 28 April 2003; accepted 7 May 2003
Abstract
To gain insight into the spatial nature of the androstenedione (AD) binding (active) site of aromatase in relation to the catalytic function
of the enzyme, we synthesized 2,2-dimethylAD (4), 2␤- and 2␣-methylADs (5 and 6), 19-oxygenated derivatives of compounds 4 and 6,
and 2-methyleneAD (17), and we then tested their inhibitory activity as well as their aromatase reaction (aromatization for 2-methyl and
2
5
4
-methylene analogs or 19-oxygenation for 2,2-dimethyl steroids) with human placental aromatase. 2-Methyl and 2-methylene steroids
, 6, and 17 were good competitive inhibitors of aromatase (Ki = 22–68 nM), but less effective compared to the 2,2-dimethyl analog
(Ki = 8.8 nM), indicating that a combination of 2␤- and 2␣-methyl moieties is essential for the formation of a thermodynamically
stable inhibitor–aromatase complex. A series of 2␣-methyl steroids were good substrates for aromatase, whereas 2␤-methyl steroid 5 was
an extremely poor substrate, and a series of 2,2-dimethyl steroids did not serve as substrate, suggesting that a 2␤-methyl moiety of the
2,2-dimethyl and 2␤-methyl steroids would prevent the aromatase reaction probably due to steric hindrance in each case. The 2-methylene
compound 17 was also aromatized to produce 2-methylestrogen with a low conversion rate where the 1,4-diene structure may have been
created before the C10–C19 bond cleavage. Kinetic analysis of the aromatization of androgens revealed that a good substrate was not
essentially a good inhibitor for aromatase.
©
2003 Elsevier Inc. All rights reserved.
Keywords: 2-Methylandrostenedione; 2,2-Dimethylandrostenedione; Aromatase inhibition; Aromatization; 19-Oxygenation; Human placental microsomes
1
. Introduction
Furth et al. [11] previously reported that the 2,2-dimethyl
substrate analog 2,2-dimethylAD (4) is an excellent com-
petitive inhibitor of human placental aromatase and one
of the most powerful steroidal inhibitors (Ki = 2.3 nM,
Km for AD = 20 nM). In addition, it was also reported
that the 2,2-dimethyl analogs of the aromatase suicide
substrates 4-hydroxyAD [11] and 6-oxoAD [12] lost
their enzyme-inactivating ability via a suicide mechanism
without diminution of their competitive inhibitory po-
tency. This indicates that the 2,2-dimethyl group prevented
the mechanism-based inactivation reaction, namely, the
aromatase-catalyzed production of a reactive electrophile
that binds irreversibly to a nucleophilic residue of the ac-
tive site of the enzyme. The inactivation reaction caused by
6-oxoAD was previously established to involve two succes-
sive 19-oxygenations followed by the 4␤,5␤-epoxidation
[13,14], whereas the inactivation by 4-OHAD was
speculated to proceed through the enzyme-generated
19-oxygenated intermediates [15,16]. Thus, we were inter-
ested in whether or not the dimethyl function was really
essential for the tight binding of inhibitor 4 to aromatase and
whether or not the dimethyl function blocked the aromatase
reaction of AD as seen in the inactivation by the suicide
Aromatase is the enzyme responsible for catalyzing
the conversion of the androgens, androst-4-ene-3,17-dione
androstenedione, AD) and testosterone to the estrogens,
(
estrone and estradiol, respectively [1–3]. The aromatiza-
tion process appears to proceed with three oxygenations
of the androgens, followed by the eventual loss of the
angular methyl group at C19 and the elimination of the
1
␤,2␤-hydrogens, resulting in the aromatization of the
A-ring of the androgen molecule to form estrogen [4–6].
Two of the three oxygenations seem to occur at the C19 posi-
tion, giving the 19-gem-diol, a hydrated form of the 19-oxo
compound 3, through the 19-hydroxy intermediate 2. It
is currently unknown whether the third oxygenation takes
place at this position or at a separate site. Inhibitors of aro-
matase are useful in treating estrogen-dependent diseases,
such as breast cancer [7–10]. For this reason, several cate-
gories of steroidal aromatase inhibitors have been designed.
∗
Corresponding author. Tel.: +81-22-234-4181; fax: +81-22-275-2013.
E-mail address: numazawa@tohoku-pharm.ac.jp (M. Numazawa).
0
039-128X/$ – see front matter © 2003 Elsevier Inc. All rights reserved.
doi:10.1016/S0039-128X(03)00089-8

5
04
M. Numazawa et al. / Steroids 68 (2003) 503–513
substrates. This study focused on the aromatase inhibitory
activity and the aromatase-reaction of a series of 2␣- and
ture (900 mg) was dissolved in 95% EtOH (15 ml) and
p-toluensulfonic acid (400 mg, 2.32 mmol) was added to
this solution. The reaction mixture was refluxed for 1 h, di-
luted with EtOAc (100 ml), washed with 5% NaHCO3 and
water, and dried with Na2SO4. Evaporation of the solvent
yielded an oil, which was purified by column chromatog-
raphy to afford a mixture of 2,2-dimethyl-, 2␤-methyl-,
and 2␣-methyl-17␤-ols (11–13) (470 mg). The mixture of
2
␤-methylAD derivatives, 19-oxygenation for 2,2-dimethyl
steroid 4, and aromatization for 2␣- and 2␤-methyl analogs
and 6. The 2,2-dimethyl group was essential for the tight
binding to aromatase, in contrast, a 2␤-methyl moiety of the
5
2,2-dimethyl function would prevent the aromatase reaction.
1
7␤-ols 11–13 (460 mg) was dissolved in acetone (150 ml),
2
. Experimental
and Jones reagent was added to this solution until the orange
color remained at 0 C. The reaction mixture was stirred for
3
◦
2
.1. Materials and general methods
◦
min at 0 C. After the excess reagent was destroyed by
adding MeOH (2 ml), the reaction mixture was condensed
to about 30 ml below 40 C, diluted with EtOAc (300 ml),
ꢀ
◦
2
,2-DimethylAD (4) [11], testosterone 17␤-(2 -tetrahyd-
ropyranyl) or 17␤-tert-butyldimethylsilyl ether, compound
[17] or 14 [18], 17␤-(tert-butyldimethylsiloxy)-6␤,19-
epoxyandrost-4-en-3-one (18) [19], and 2-methylestradiol
20] were synthesized according to the methods previ-
washed with 5% NaHCO3 and water, dried with Na2SO4,
and evaporated to give a solid. Recrystallization of the solid
from acetone yielded 2␣-methyl-17-one (6); in contrast, pu-
rification of the mother liquor by column chromatography
7
[
3
3
ously reported. [1␤- H]AD (27.5 Ci/mmol; H-distribution,
4–79% at 1␤) was obtained from New England Nuclear
(hexane–EtOAc, 9:1 v/v) produced 2,2-dimethyl-17-one (4)
7
and 2␤-methyl-17-one (5) along with steroid 6.
Corp. (Boston, MA, USA), and reduced nicotinamide ade-
nine dinucleotide phosphate (NADPH) was purchased from
Kohjin Co. Ltd. (Tokyo, Japan).
Melting points were measured on a Yanagimoto melt-
ing point apparatus (Kyoto, Japan) and are uncorrected.
Infrared (IR) spectra were recorded in KBr pellets for the
solid products or in neat forms for the oily products on a
Perkin-Elmer FT-IR 1725X spectrophotometer (Norwalk,
CT, USA). Ultra violet (UV) spectra were determined in
2.3. 2,2-Dimethylandrost-4-ene-3,17-dione (4)
◦
◦
Yield: 5% from 7; m.p. 144–147 C ([11] 150–151 C).
H NMR δ: 0.92 (3H, s, 18-Me), 1.11 and 1.18 (3H each,
1
s, 2␤-Me and 2␣-Me), 1.30 (3H, s, 19-Me), 5.70 (1H, d,
J = 1.2 Hz, 4-H).
2.4. 2β-Methylandrost-4-ene-3,17-dione (5)
9
(
5% ethanol on a Hitachi 150-20 UV spectrophotometer
Tokyo, Japan). 1H-Nuclear magnetic resonance (NMR)
spectra were obtained in CDCl3 solutions with a JEOL EX
70 (270 MHz) spectrometer (Tokyo, Japan) using tetram-
◦
1
Yield: 2% from 7; m.p. 166–169 C. H NMR δ: 0.92
(3H, s, 18-Me), 1.11 (3H, d, J = 6.8 Hz, 2␤-Me), 1.18 (3H,
s, 19-Me), 5.75 (1H, d, J = 1.0 Hz, 4-H). FT-IR (KBr):
2
−
1
ethylsilane as an internal standard. Mass spectra (MS)
were determined with a JEOL JMS-DX 303 spectrometer.
Thin-layer chromatography (TLC) was performed on E.
Merck precoated TLC silica gel plates (silica gel 60F-254,
layer thickness 0.25 and 0.5 mm for analytical and prepara-
tive use, respectively; Darmstadt, Germany). Column chro-
matography was conducted with silica gel 60, 70–230 mesh
1736 and 1656 (C=O) cm . UV (EtOH) λ : 242 nm (ε =
max
+
15,600). MS m/z (relative intensity): 300 (M , 38), 258 (26),
244 (19), 41 (100). Analysis calculated for C H O : C,
79.95; H, 9.39. Found: C, 79.99; H, 9.34.
20
28
2
2.5. 2α-Methylandrost-4-ene-3,17-dione (6)
(
E. Merck, 70–230 mesh).
◦
◦
Yield: 40% from 7; m.p. 149–150 C ([21] 157–159 C).
H NMR δ: 0.92 (3H, s, 18-Me), 1.11 (3H, d, J = 6.6 Hz,
1
ꢀ
2
4
.2. Reaction of 17β-(2 -tetrapyranyloxy)androst-
-en-3-one (7) with methyliodide-tert-BuOK
2␣-Me), 1.24 (3H, s, 19-Me), 5.73 (1H, d, J = 1.5 Hz,
1
−
4-H). FT-IR (KBr): 1738 and 1664 (C=O) cm . UV λ
:
max
+
2
40 nm (ε = 15,200). MS m/z (relative intensity): 300 (M ,
A suspension of tert-BuOK (894 mg, 7.91 mmol) in
39), 258 (29), 244 (21), 41 (100).
THF (12 ml) was slowly added to a solution of 4-en-3-one
7
(1.57 g, 4.22 mmol) in THF (6 ml) containing CH3I
2.6. 2-Methylene-17β-(tert-butyldimethylsiloxy)androst-
4-en-3-one (15)
◦
(
3.75 ml, 60 mmol) with stirring at −60 C [17]. The mix-
◦
ture was stirred at −60 C for 3 h, diluted with EtOAc
(300 ml), washed with 5% HCl, 5% NaHCO3 solution,
A solution of bis(dimethylamino)methane (4.0 ml,
and water sequentially, and dried with Na2SO4. Evap-
oration of the solvent gave an oil that was purified by
column chromatography (hexane–EtOAc, 10:1 v/v), giv-
ing a mixture of 2,2-dimethyl, 2␤-methyl, and 2␣-methyl
products (8, 9, and 10) as an oil (907 mg). The oily mix-
29.2 mmol) in anhydrous ether (50 ml) was slowly added to
an ice-cold solution of acetyl chloride (2.0 ml, 21.9 mmol)
in anhydrous ether (28 ml), and the mixture was stirred for
30 min [22]. The resulting precipitate was rapidly filtered
and suspended in CH3CN (45 ml). A solution of 4-en-3-one

M. Numazawa et al. / Steroids 68 (2003) 503–513
505
1
4 (3.8 g, 9.45 mmol) in THF (26 ml) was added dropwise
of 2␤- and 2␣-methyl compounds 5 and 6. This mixture
was purified by column chromatography (hexane–EtOAc,
4:1 v/v) and subsequent high-performance liquid chro-
matography (HPLC), giving the 2-methyl compounds
as the analytical samples. HPLC conditions: pump, Wa-
ters M510 (Waters, Milford, MA, USA); mobile phase,
CH3CN–H2O (45:15 v/v, 4 ml/min); stationary phase, C18
column (5 ␮m, 250 mm × 10 mm, Mightysil, Kanto Chem-
ical Co. Inc., Tokyo, Japan); detector, Waters UV detec-
tor 486 (at 250 nm). 2␤-Methyl steroid 5 (retention time:
to this suspension, and the reaction mixture was heated for
4
to afford a solid that was suspended in THF (28 ml). Fil-
tration and evaporation provided a crude product, which
was purified by column chromatography (hexane–EtOAc,
◦
5 h at 80 C. After this time, the solvent was evaporated
15:1 v/v) to yield 2-methylene compound 15 (275 mg, 7%)
along with the recovered starting material 14 (1.53 g, 41%).
◦
1
Compound 15: m.p. 158–160 C (from acetone). H NMR
δ: 0.010 and 0.016 [3H each, s, 17␤-Si(Me)2], 0.75 (3H,
s, 18-Me), 0.88 (9H, s, 17␤-Si(Me)2C(Me)3), 1.11 (3H, s,
◦
36.4 min): m.p. 168–170 C. 2␣-Methyl steroid 6 (retention
◦
1
9-Me), 3.56 (1H, t, J = 8.2 Hz, 17␣-H), 5.22 and 5.83
time: 39.0 min): m.p. 149–152 C. The two compounds
(
1H each, s, 2-C=CH2), 5.94 (1H, t, J = 2.1 Hz, 4-H).
were identical to the corresponding samples synthesized
above.
−
1
FT-IR (KBr): 1666 (C=O) and 1624 (C=C) cm . UV λmax:
+
2
5
61 nm (ε = 16,400). MS m/z (relative intensity): 414 (M ,
), 357 (99), 281 (20), 75 (100). Analysis calculated for
2.10. Reaction of 17β-(tert-butyldimethylsiloxy)-6␤,19-
C H42O2Si: C, 75.30; H, 10.21. Found: C, 75.15; H, 10.41.
2
6
epoxyandrost-4-en-3-one (18) with methyliodide-tert-BuOK
2
.7. 2-Methylene-17β-hydroxyandrost-4-en-3-one (16)
A solution of compound 18 (4.4 g, 10.6 mmol) and MeI
(
12 ml) in THF (45 ml) was added to a suspension of
A solution of 17␤-siloxy steroid 15 (260 mg, 0.62 mmol)
◦
tert-BuOK (2.05 g, 18.2 mol) in 140 ml of THF at −60 C,
while stirring under a stream of N2 gas, and the mixture was
stirred for 45 min. After this time, the mixture was diluted
with EtOAc (1000 ml), washed with 5% HCl, 5% NaHCO3
solution, and water, sequentially, and dried with Na2SO4.
Evaporation of the solvent gave an oily product that was pu-
rified by silica gel column chromatography (hexane–EtOAc,
in THF (6 ml) and propan-2-ol (6 ml) was treated with 1 M
HCl (3.5 ml) overnight at room temperature, and the result-
ing mixture was neutralized by adding NaHCO3, diluted
with EtOAc (200 ml), washed with water, and dried with
Na2SO4. Evaporation of the solvent gave a solid that was re-
crystallized from acetone to yield 17␤-ol 16 (170 mg, 92%).
◦
1
M.p. 149–151 C. H NMR δ: 0.80 (3H, s, 18-Me), 1.07
3H, s, 19-Me), 3.66 (1H, t, J = 8.4 Hz, 17␣-H), 5.22
1
2
5:1 v/v) to yield the 2,2-dimethyl product 19 (880 mg,
0%) and the 2␣-methyl product 20 (720 mg, 17%) along
(
and 5.84 (1H each, s, 2-C=CH2), 5.94 (1H, t, J = 2.1 Hz,
−
1
with the recovered starting material 18 (2.5 g, 61%).
4
2
2
-H). FT-IR: 1668 (C=O) and 1620 (C=C) cm . UV λmax:
+
59 nm (ε = 11,100). MS m/z (relative intensity): 300 (M ,
6), 147 (13), 134 (63), 41 (100). Analysis calculated for
2
.11. 2,2-Dimethyl-17β-(tert-butyldimethylsiloxy)-6␤,19-
C20H28O2: C, 79.95; H, 9.39. Found: C, 79.68; H, 9.50.
epoxyandrost-4-en-3-one (19)
2
.8. 2-Methyleneandrost-4-ene-3,17-dione (17)
◦
1
M.p. 73–76 C (from EtOAc). H NMR δ: 0.002 and 0.008
3H each, s, 17␤-OSi(Me)2C(Me)3], 0.80 (3H, s, 18-Me),
0
[
1
7␤-Hydroxy steroid 16 (70 mg, 0.23 mmol) in acetone
.88 [9H, s, 17␤-OSi(Me)2C(Me)3], 1.15 and 1.22 (3H each,
◦
(24 ml) was treated with Jones reagent for 3 min at 0 C as
s, 2␤-Me and 2␣-Me), 3.46 and 4.20 (1H each, d, J = 8 Hz,
described above. A crude product was obtained and recrys-
tallized from acetone to give 17-ketone 17 (60 mg, 90%).
M.p. 194–197 C. H NMR δ: 0.92 (3H, s, 18-Me), 1.13
1
5
9-CH2), 3.56 (1H, t, J = 8 Hz, 17␣-H), 4.66 (1H, d, J =
1
−
Hz, 6␣-H), 5.80 (1H, s, 4-H). FT-IR: 1675 (C=O) cm
.
◦
1
+
MS m/z (relative intensity): 447 (M , 27), 403 (13), 387
100), 357 (23). Analysis calculated for C27H44O3Si: C,
2.91; H, 9.97. Found: C, 73.16; H, 9.77.
(
3H, s, 19-Me), 5.24 and 5.86 (1H each, s, 2-C=CH2), 5.94
(
7
(
1
1H, t, J = 2.1 Hz, 4-H). FT-IR: 1738 and 1672 (C=O) and
−
1
621 (C=C) cm . UV λmax: 259 nm (ε = 13,200). MS m/z
+
(relative intensity): 298 (M , 82), 147 (19), 134 (100). Anal-
ysis calculated for C20H O2: C, 80.49; H, 10.78. Found:
C, 80.88; H, 10.91.
2.12. 2α-Methyl-17β-(tert-butyldimethylsiloxy)-6␤,19-
epoxyandrost-4-en-3-one (20)
26
◦
1
2
.9. Hydrogenation of 2-methylene steroid 17 with
M.p. 118–120 C (from MeOH). H NMR δ: 0.004 [6H,
s, 17␤-OSi(Me) C(Me) ], 0.80 (3H, s, 18-Me), 0.875 [9H,
H2-tris(triphenylphosphine)rhodium chloride
2
3
s, 17␤-OSi(Me)2C(Me)3], 1.20 (3H, d, J = 7.3 Hz, 2␣-Me),
A solution of 2-methylene compound 17 (180 mg,
.60 mmol) in THF and toluene (1:1) (20 ml) was hydro-
genated with the rhodium complex (200 mg, 0.22 mmol)
for 17 h, filtered, and evaporated to provide a 1:1 mixture
3.55 (2H, m, 19-Ha and 17␣-H), 4.13 (1H, d, J = 7.8 Hz,
0
19-H ), 4.61 (1H, d, J = 5.3 Hz, 6␣-H), 5.85 (1H, s, 4-H).
b
−
1
FT-IR: 1663 (C=O) cm . UV λmax: 239 nm (ε = 11,900).
+
MS m/z (relative intensity): 430 (M , 3), 373 (100), 343 (4),

5
06
M. Numazawa et al. / Steroids 68 (2003) 503–513
7
9
5 (40). Analysis calculated for C H42O3Si: C, 72.50; H,
.83. Found: C, 72.60; H, 9.85.
2.17. 2,2-Dimethyl-6β,19-epoxyandrost-4-ene-
3,17-dione (23)
26
◦
1
Yield: 94%; m.p. 174–176 C. H NMR δ: 0.97 (3H, s, 18-
Me), 1.16 and 1.21 (3H each, s, 2␤-Me and 2␣-Me), 3.51 and
2
.13. Treatment of 17β-(tert-butyldimethylsilyloxy) steroids
1
9 and 20 with diluted HCl
4
6
.20 (1H each, d, J = 8 Hz, 19-CH2), 4.72 (1H, d, J = 5 Hz,
−1
-H), 5.84 (1H, s, 4-H). IR: 1736 and 1672 (C=O) cm
.
1
M HCl (67 or 4.8 ml) was added to a solution of 2,2-
UV λmax: 236 nm (ε = 13,600). MS m/z (relative intensity):
dimethyl steroid 12 (4.5 g, 10.14 mmol) or 2␣-methyl-17␤-
tert-butyldimethylsilyloxy steroid 13 (325 mg, 0.756 mmol)
in propan-2-ol (138 or 10 ml) and THF (93 or 6.5 ml).
The mixture was stirred overnight at room temperature
and then diluted with EtOAc (500 or 100 ml), washed
with 5% NaHCO3 solution and water, and dried with an-
hydrous Na2SO4. Evaporation of the solvent gave 17␤-ol
+
3
28 (M , 60), 314 (6), 244 (100). Analysis calculated for
C21H28O3: C, 76.79; H, 8.59. Found: C, 76.64; H, 8.87.
2
.18. 2α-Methyl-6β,19-epoxyandrost-4-ene-3,17-dione
(24)
◦
1
Yield: 88%; m.p. 121–123 C (from acetone–hexane). H
2
(
1 or 22 that was purified by column chromatography
hexane–EtOAc, 3:1 v/v) and then recrystallized from an
appropriate solvent.
NMR δ: 0.88 (3H, s, 18-Me), 1.18 and 1.20 (3H, d, J =
7
1
.1 Hz, 2␣-Me), 3.62 and 4.13 (1H each, d, J = 8 Hz,
9-CH2), 4.67 (1H, d, J = 5 Hz, 6-H), 5.89 (1H, s, 4-H).
−
1
IR: 1736 and 1677 (C=O) cm . UV λmax: 237 nm (ε =
2
.14. 2,2-Dimethyl-17β-hydroxy-6β,19-epoxyandrost-4-
+
1
1,800). MS m/z (relative intensity): 314 (M , 55), 244
en-17-one (21)
(
100), 150 (20). Analysis calculated for C20H O3: C, 76.40;
26
H, 8.34. Found: C, 76.50; H, 8.35.
◦
1
Yield: 94%; m.p. 181–184 C (from EtOAc). H NMR δ:
.84 (3H, s, 18-Me), 1.15 and 1.22 (3H each, 2␤-Me and
␣-Me), 3.48 and 4.20 (1H each, d, J = 8 Hz, 19-CH2),
.66 (1H, m, 17␣-H), 4.67 (1H, d, J = 5 Hz, 6␣-H), and
0
2
3
5
2.19. Zinc powder reduction of 6β,19-epoxides 23 and 24
A mixture of 6␤,19-epoxides 23 or 24 (500 mg each,
−
1
.81 (1H, s, 4-H). FT-IR: 3204 (OH) and 1667 (C=O) cm
.
about 1.5 mmol), propan-2-ol (70 ml), CH3COOH (21 ml),
and Zn powder (8.5 g) was stirred under reflux for 5 h. The
resulting solid material was removed by filtration, and the
filtrate was condensed to about 20 ml, diluted with EtOAc
UV λmax: 237 nm (ε = 12,900). MS m/z (relative intensity):
+
3
30 (M , 81), 316 (8), and 246 (100). Analysis calculated
for C21H30O3: C, 76.32; H, 9.15. Found: C, 76.40; H, 9.47.
(
200 ml), washed with 5% NaHCO3 and water, and dried
2
3
.15. 2α-Methyl-17β-hydroxy-6β,19-epoxyandrost-4-en-
-one (22)
with anhydrous Na2SO4. Evaporation of the solvent gave
a solid, which was purified by column chromatography
(
hexane–EtOAc, 1:1 v/v) and subsequently recrystallized to
◦
1
Yield: 92%; m.p. 59–60 C (from EtOAc). H NMR δ:
.84 (3H, s, 18-Me), 1.20 (3H, d, J = 7.3 Hz, 2␣-Me), 3.57
afford a 19-ol 25 or 26.
0
and 4.14 (1H each, d, J = 7.8 Hz, 19-CH2), 3.65 (1H, m,
2
.20. 2,2-Dimethyl-19-hydroxyandrost-4-ene-3,17-dione
1
7␣-H), 4.62 (1H, d, J = 5.3 Hz, 6␣-H), 5.86 (1H, s, 4-H).
1
(25)
−
FT-IR: 3504 (OH) and 1672 (C=O) cm . UV λmax: 239 nm
+
(
ε = 10,800). MS m/z (relative intensity): 316 (M , 57),
◦
1
Yield: 72%; m.p. 194–197 C. H NMR δ: 0.94 (3H, s,
9-Me), 1.16 and 1.19 (3H each, s, 2␤-Me and 2␣-Me),
.83–3.98 (2H, m, 19-CH2), 5.84 (1H, s, 4-H). IR: 3496
2
46 (100). Analysis calculated for C20H28O3: C, 75.91; H,
.92. Found: C, 75.63; H, 9.10.
1
3
(
8
−
1
OH), 1731 and 1662 (C=O) cm . UV λmax: 242 nm (ε =
+
2
.16. Oxidation of 17β-hydroxy steroids 21 and 22
14,000). MS m/z (relative intensity): 330 (M , 8), 300 (100),
with Jones reagent
274 (8), 257 (5). Analysis calculated for C H O : C,
21
30
3
76.32; H, 9.15. Found: C, 76.00; H, 9.49.
Jones reagent (4.3 or 2.0 ml) was added dropwise to a so-
lution of 17␤-ol 21 (2.0 g, 6.06 ml) or 22 (1.0 g, 3.10 mmol)
2.21. 2α-Methyl-19-hydroxyandrost-4-ene-3,17-dione (26)
◦
in acetone (600 or 300 ml) at 0 C, and the reaction mixture
◦
◦
1
was stirred for 3 min at 0 C. After addition of MeOH (5 or
Yield: 75%; m.p. 209–212 C (from acetone). H NMR
δ: 0.92 (3H, s, 19-Me), 1.09 (3H, d, J = 6.6 Hz, 2␣-Me),
2.83 (1H, m, 2␤-H) 3.97 and 4.08 (1H each, d, J = 11 Hz,
19-CH2), 5.91 (1H, s, 4-H). IR: 3467 (OH), 1734 and 1662
3
ml), the solvent was evaporated to about 200 or 100 ml un-
◦
der reduced pressure at 40 C, diluted with EtOAc (500 or
3
00 ml), washed with 5% NaHCO3 solution and water, and
−
1
dried with anhydrous Na2SO4. The solvent was then evap-
orated to give a solid. Recrystallization of this solid product
from acetone gave 17-keto steroid 23 or 24.
(C=O) cm . UV λmax: 242 nm (ε = 12,800). MS m/z (rela-
+
tive intensity): 316 (M , 29), 286 (10). Analysis calculated
for C20H28O3: C, 75.91; H, 8.92. Found: C, 75.73; H, 9.02.

M. Numazawa et al. / Steroids 68 (2003) 503–513
507
2
.22. Oxidation of 19-hydroxy steroids 25 and 26 with
and 600 ␮M NADPH, one-tenth of the preincubation mix-
pyridium dichromate (PDC)
ture was used to assay the remaining aromatase activity. Ap-
parent Km and K values were calculated using non-linear
i
PDC (440 mg, 1.17 mmol) was added to a solution of 19-ol
regression analysis with GraFit software [26].
2
5 or 26 (250 mg, about 0.26 mmol) in CH2Cl2 (35 ml), and
the mixture was stirred for 5 h at room temperature. The
solid material was filtered off, and the filtrate was diluted
with EtOAc (150 ml), washed with water, and dried with
anhydrous Na2SO4. Evaporation of the solvent yielded a
solid, which was purified by silica gel column chromatogra-
phy and recrystallized in an appropriate solvent to produce
a 19-al 27 or 28.
2.27. 19-Oxygenation and aromatization studies with gas
chromatography–mass spectrometry (GC–MS)
Incubation experiments were carried out principally ac-
cording to the previous methods [27,28]. Briefly, a mixture
of 4 ␮M of substrate, 200 ␮M of NADPH, 500 ␮g protein
of placental microsomes, and 100 ␮l of MeOH in 2.0 ml of
◦
6
7 mM phosphate buffer (pH 7.4) was incubated at 37 C.
After 30 min of incubation, 6␣-methylestradiol (10 ng)
was added to each incubation mixture and extracted with
EtOAc. The extract was placed in a Sep-Pak C18 cartridge
2
.23. 2,2-Dimethylandrost-4-ene-3,17,19-trione (27)
◦
1
Yield: 78%; m.p. 206–209 C. H NMR δ: 0.85 and 1.16
(Waters) in the experiments with 2,2-dimethyl steroids as
(3H each, s, 2␤-Me and 2␣-Me), 0.97 (3H, s, 18-Me), 5.88
substrate, while in the experiments with 2␣-methyl steroids
as substrate, the extract was reduced with NaBH4 and then
placed in the cartridge. The 19-oxygenated or aromatized
product was obtained in the 80% MeOH fraction in each
experiment. The recovery rates for the 19-oxygenated prod-
ucts and 2-methylestradiol with these procedures were in
the range of 75–85%.
For kinetic analysis, the incubation mixture contained
various concentrations of each steroid, 200 ␮g of the mi-
crosomal protein, 200 ␮M NADPH, 100 ␮l of MeOH, and
(
1
1H, d, J = 2 Hz, 4-H), 10.01 (1H, s, 19-CH). IR: 1737 and
−
1
704 (C=O) cm . UV λmax: 246 nm (ε = 12,000). MS m/z
+
(
(
relative intensity): 328 (M , 65), 299 (44), 281 (15), 272
100), 257 (9), 244 (30), 228 (9). Analysis calculated for
C21H28O3: C, 76.79; H, 8.59. Found: C, 76.66; H, 8.56.
2
.24. 2α-Methylandrost-4-ene-3,17,19-trione (28)
◦
1
Yield: 80%; m.p. 150–154 C. H NMR δ: 0.89 (3H, s,
9-Me), 1.09 (3H, s, 2␣-Me), 5.97 (1H, d, J = 1.8 Hz,
-H), 9.98 (1H, s, 19-CHO). IR: 1737, 1708 and 1667
1
4
(
9
00 ␮l of the phosphate buffer. The mixture was incubated
◦
−1
at 37 C for 30 min, and the aromatized product was puri-
fied as descried above. Kinetic data for 2␣-methyl steroid
aromatization were obtained using 6␣-methylestradiol as an
internal standard according to the method [27] previously
reported.
C=O) cm . UV λmax: 245 nm (ε = 11,200). MS m/z (rel-
+
ative intensity): 314 (M , 80), 285 (100), 267 (37), 244
(
Found: C, 76.36; H, 8.52.
20). Analysis calculated for C20H O3: C, 76.40; H, 8.34.
26
The aromatized product 2-methylestradiol was converted
to the bistrimethylsilyl ether with bistrimethylsilyl-
trifluoroacetamide (BSTFA), and the 19-oxygenated prod-
ucts 2,2-dimethyl-19-ol (18) and 2,2-dimethyl-19-al (20)
were converted to the trimethylsilyl-bismethoxime and the
trismethoxime, respectively, by treatment with methoxime
hydrochloride followed by the silylating reagent [28]. These
derivatives were then subjected to GC–MS analysis under
the conditions previously reported [27,28] using a Finnigan
MAT SSQ GC–MS instrument (San Jose, CA, USA). Reten-
tion times: 15.53 min for 2-methylestradiol-bistrimethylsilyl
ether, 14.97 min for 6␣-methylestradiol-bistrimethylsilyl
ether, 15.73 min for 19-hydroxy-2,2-dimethyl steroid
18 bismethoxime-trimethylsilyl ether, and 16.35 min of
19-oxo-2,2-dimethyl steroid 20 trismethoxime.
2
.25. Enzyme preparation
Human placental microsomes (sedimenting at 105,000 ×
g for 60 min) were obtained as described by Ryan [23].
These were washed once with 0.05 mM dithiothreitol solu-
◦
tion, lyophilized, and stored at −20 C. No significant loss
of activity occurred over the period (2 months) of this study.
2
.26. Aromatase assay procedure
Aromatase activity was measured according to the method
of Siiteri and Thompson [24] in which tritiated water re-
leased from [1␤- H]AD into the incubation mixture during
3
aromatization was used as an index of enzyme activity. The
reversible inhibition assay and time-dependent assay proce-
dures used were principally the same as the assay methods
described in our previous work [25]. Briefly, 20 ␮g protein
from lyophilized microsomes and a 5 or 20 min incubation
time for the reversible assay were used in this study, and the
assay was carried out in 67 mM phosphate buffer, pH 7.5,
in the presence of NADPH under air. In the time-dependent
inactivation experiment using 200 ␮g microsomal protein
3. Results
3.1. Chemistry
Reaction of testosterone 17␤-tetrahydropyranyl ether (7)
◦
with limited amounts of CH3I and tert-BuOK at −60 C

5
08
M. Numazawa et al. / Steroids 68 (2003) 503–513
O
O
H C OTBDMS
OR
H3C
H C
3
H C
3
3
H C
H C
3
3
R
H3C
CH2 [(CH ) N]2
3 2
H C
2
R
R2
1
CH3COCl, ether
O
O
O
O
14
1
5; R=TBDMS
6; R=H
1
1
; R=CH (androstenedione, AD)
4; R =R =CH
1 2 3
3
Jones reagent,
acetone
2; R=CH OH
5; R =CH , R =H
1 3 2
2
3; R=CHO
6; R =H, R =CH
1 2 3
O
O
H C
3
H C
3
Fig. 1. Structures of androstenedione and its 19-oxygenated and 2-methy-
lated derivatives.
H C
H C
3
3
H , (PH P) RhCl, THF, toluene
H C
2
3
3
H C
3
2
O
O
5
and 6
17
gave a mixture of 2␤- and 2␣-methyl steroids 9 and 10
and 2,2-dimethyl steroid 8 (Figs. 1 and 2). The mixture
was, without further purification, hydrolyzed with diluted
HCl, and subsequent Jones oxidation produced a mixture
of 2␤- and 2␣-methylADs (5 and 6) and 2,2-dimethylAD
Fig. 3. Synthesis of 2-methylated steroids via a 2-methylene intermediate.
1
6
2
7␤-hydroxy group, and Zn powder reduction of the
␤,17-epoxy ring, sequentially, yielding 2,2-dimethyl- and
␣-methyl-19-ols 25 and 26, respectively. PDC oxidation of
(
4). The products were successfully separated by crystal-
lization followed by column chromatography. In contrast,
treatment of testosterone 17␤-tert-butyldimethylsilyl ether
the 19-ols produced the corresponding 19-oxo derivatives
7 and 28.
(14) with bis(dimethylamino)methane and acetyl chloride
2
yielded 2-methylene steroid 15 (Fig. 3). After removal of the
1
1
7-silyl group with diluted HCl, the 17␤-hydroxy product
6 obtained was treated with Jones reagent followed by hy-
3.2. Aromatase inhibition
drogenation with H2-tris(triphenylphosphine)rhodium chlo-
ride, yielding a mixture of 2␤- and 2␣-methyl products 5
and 6. HPLC purification gave these two compounds in their
pure forms. The stereochemistry of the 2␤-methyl group of
compound 5 was determined on the basis of H NMR spec-
troscopy. Thus, the correlation between the 19-methyl (δ:
Inhibition of aromatase activity in human placen-
tal microsomes by 2␤- and 2␣-methylADs (5 and 6),
2-methyleneAD (17), and 19-oxygenated 2␣-methyl- and
2,2-dimethyl-derivatives (25–28) was examined in vitro by
enzyme kinetics. The results are given in Table 1. Aro-
matase activity in placental microsomes was determined
using a radiometric assay [24] in which tritiated water re-
1
1
.18 ppm) and the 2␤-methyl (δ: 1.11 ppm) was observed by
the NOESY analysis.
␤,19-Epoxy compound 18 was treated with CH3I and
3
6
leased from [1␤- H]AD (1) into the incubation medium
tert-BuOK similarly as described above, producing a mix-
ture of 2,2-dimethyl and 2␣-methyl products 19 and 20
during aromatization was measured. To characterize the
nature of inhibitor binding to the active site of aromatase,
aromatization was measured at several inhibitor and sub-
strate concentrations. The results of these studies were
plotted on a typical Lineweaver–Burk plot. In these studies,
the apparent Km and Vmax for AD (1) were about 33 nM and
(Fig. 4). After separation of the two products by column
chromatography, steroids 19 and 20 were separately sub-
jected to hydrolysis of their 17␤-tert-butyldimethylsilyl
ether group with diluted HCl, Jones oxidation of the
OTHP
OR3
H3C
H C
3
Jones
CH I, tert-BuOK, THF
H3C
oxidation
H C
3
3
4, 5, and 6
R1
R2
O
–
60 ˚C
O
7
8; R = R =CH , R =THP
1 2 3 3
9
; R =CH , R =H, R =THP
1
3
2
3
1
0; R =H, R =CH , R =THP
1 2 3 3
1
1
1
1; R = R =CH , R =H
1 2 3 3
2; R =CH , R =H, R =H
1
3
2
3
3; R =H, R =CH , R =H
1
2
3
3
Fig. 2. Synthesis of 2-methylated steroids via a direct methylation.

M. Numazawa et al. / Steroids 68 (2003) 503–513
509
OTBDMS
OR
H C
3
3
H C
3
CH I, tert-BuOK, THF
R
R
3
1
2
O
O
-78˚C
O
O
1
8
1
2
2
2
9; R =R =CH , R =TBDMS
1 2 3 3
1
M HCl,
propan-2-ol,
THF
0; R =H, R =CH , R =TBDMS
1
2
3
3
1; R =R =CH , R =H
1
2
3
3
2; R =H, R =CH , R =H
1
2
3
3
Jones reagent,
acetone, 0˚C
O
O
H C
3
H C
3
HOH C
2
Zn, acetic acid,
propan-2-ol
R
R1
1
R
R
O
2
2
O
O
2
5; R =R =CH
23; R =R =CH
1 2 3
1
2
3
2
6; R =H, R =CH
24; R =H, R =CH
1 2 3
1
2
3
PDC, CH Cl
2
2
O
H C
3
H
C
O
R
1
R
2
O
27; R =R =CH
1 2 3
2
8; R =H, R =CH
TBDMS: tert-butyldimethylsilyl
1
2
3
Fig. 4. Synthesis of 19-oxygenated 2-methyl steroids.
1
25 (pmol/min)/mg protein, respectively. All of the steroids
3.3. Aromatase-catalyzed reaction
studied exhibited clear competitive-type inhibition. The
apparent inhibition constants (K ) were obtained by Dixon
plots. The Lineweaver–Burk plot of aromatase inhibition
by 2␤-methylAD (5) is shown in Fig. 5.
Aromatase-catalyzed 19-oxygenation studies of 2,2-
dimethyl steroid 4 and its 19-hydroxy derivative 25 were
initially carried out. The two steroids were separately
i
Table 1
In vitro aromatase inhibition by 2,2-dimethyl-, 2␤-methyl-, and 2␣-methylAD analogs
Substrate
IC50 (␮M)a
Ki (nM)b
Inhibitionc
2
2
2
2
2
2
2
2
,2-Dimethyl (4)
0.089 ± 0.007
5.1 ± 0.26
0.60 ± 0.02
0.25 ± 0.02
0.69 ± 0.09
11 ± 0.8
8.8 ± 0.33
670 ± 74
66 ± 4.1
Competitive
Competitive
Competitive
Competitive
Competitive
Competitive
Competitive
Competitive
,2-Dimethyl-19-ol (25)
,2-Dimethyl-19-one (27)
␤-Methyl (5)
22 ± 2.1
␣-Methyl (6)
55 ± 2.4
␣-Methyl-19-ol (26)
␣-Methyl-19-one (28)
-Methylene (17)
1200 ± 90
1690 ± 146
68 ± 3.7
12 ± 2.4
0.55 ± 0.02
a
3
A concentration of 300 nM of [1␤- H]AD and 20 ␮g of protein from human placental microsomes were used.
b
c
Inhibition constant was obtained by Dixon plot. In these experiments, the apparent Km for the substrate androstenedione was found to be about 33 nM.
Inhibition type was determined by Lineweaver–Burk plot. Results are shown as the means ± S.E.

5
10
M. Numazawa et al. / Steroids 68 (2003) 503–513
GC–MS. The 19-hydroxy product 25 was determined as the
bismethoxime-trimethylsilyl ether and the 19-oxo product
7 as the trismethoxime. In each experiment, the production
of 19-oxygenated metabolite 25 or 27 was not observed
detection limit: 0.2 or 0.5 (pmol/min)/mg protein for 19-ol
2
(
25 or 19-al 27).
Aromatization of 2␤-methylAD (5), 2␣-methylAD (6),
its 19-hydroxy and 19-oxo analogs 26 and 28, and 2-
methyleneAD (17) with human placental aromatase was also
explored. It was presumed that the aromatization product
would be a mixture of 2-methylestrone and its 17␤-reduced
estrogen, 2-methylestradiol, produced by the action of
17␤-hydroxysteroid dehydrogenase [29,30]. Thus, the
aromatization product was converted into 2-methylestradiol
by treatment with NaBH4 in each experiment. This product
was then purified and analyzed as the bistrimethylsilyl ether
by GC–MS. The mass spectra and retention times of the
aromatized products of all of the androgens were identical
with those of the authentic 2-methylestradiol, respectively
Fig. 5. Lineweaver–Burk plot of inhibition of human placental aromatase
by 2␤-methylAD (5) with AD as a substrate. Each point represents the
mean of two determinations, which varied by less than 5% of the mean.
The inhibition experiments performed with all of the other steroids gave
essentially similar plots (data not shown).
(
Fig. 6). The mass spectrum of the bistrimethylsilyl ether
+
incubated with human placental aromatase in the pres-
ence of NADPH under air. The incubation products were
purified and subjected to reaction with methoxyamine, fol-
lowed by silylation with BSTFA, and then analyzed by
of 2-methylestradiol showed a molecular ion (M ) as the
+
base peak with a characteristic ion of M − 31.
The aromatization rate of 2␣-methylAD (6) at a concen-
tration of 4 ␮M increased linearly with increasing amounts
Fig. 6. Mass spectra of the bistrimethylsilyl derivative of the authentic 2-methylestradiol (A) and that of the aromatized product obtained by incubation
of 2␣-methylAD (6) with human placental microsomes (B).

M. Numazawa et al. / Steroids 68 (2003) 503–513
511
Table 2
9-Oxygenation of 2,2-dimethylAD or aromatization of 2-methylAD
1
analogs with human placental microsomes
Substratea
19-Oxygenation
((pmol/min)/mg protein)
2
2
,2-Dimethyl (4)b
,2-Dimethyl-19-ol (25)
<0.2
<0.5
c
Aromatizationd
((pmol/min)/mg protein)
2
2
2
2
2
␤-Methyl (5)
␣-Methyl (6)
-Methylene (17)
␣-Methyl-19-ol (26)
␣-Methyl-19-one (28)
0.075 ± 0.009
78.9 ± 4.6
0.31 ± 0.021
127 ± 0.36
186 ± 6.9
a
All substrates (4 ␮M) were incubated with human placental micro-
Fig. 7. Kinetic analysis of aromatization of 2␣-methyl steroid 6 with
human placental aromatase. Various concentrations (0.025, 0.05, 0.1, 0.2,
0.5, 1, 2 ␮M) of substrate 6 were separately incubated with placental
◦
somes (500 ␮g of protein) at 37 C for 30 min in the presence of NADPH
(
200 ␮M) in 2.1 ml of 67 mM phosphate buffer, pH 7.4.
b
◦
The conversion of 2,2-dimethyl steroid 5 into its 19-hydroxy metabo-
microsomes (200 ␮g protein) for 30 min at 37 C, and the aromatization
lite 25 was determined by GC–MS with 19-hydroxyandrostenedione (2)
rate was analyzed by GC–MS with an internal standard as described in
Section 2. Each experiment was carried out in duplicate. The variation
between two determinations was less than 10%.
as an internal standard.
c
The conversion of 2,2-dimethyl-19-ol (25) into its 19-oxo metabolite
2
7 was determined by GC–MS with compound 2 as an internal standard.
d
The aromatized product 2-methylestradiol was analyzed as the
bistrimethylsilyl derivative by GC–MS with 6␣-methylestradiol as an in-
ternal standard in each experiment. Results are the means± S.D. (n = 4).
of placental microsomes (up to 500 ␮g of protein); further-
more, the conversion of substrate 6 increased linearly with
incubation time (up to 45 min). Thus, the aromatization rates
of all the substrates were first determined in experiments
using a 30 min incubation time and 500 ␮g of placental mi-
crosomes; the conversions were less than 13% (Table 2).
The aromatization rate of AD to estradiol, obtained by the
3
radiometric assay with [1␤- H]AD as a substrate [24], was
1
25 (pmol/min)/mg protein.
All of the 2␣-methyl steroids were good substrates for
Fig. 8. Inhibition of aromatization of 2␣-methylAD (6) by the natural
substrate AD. Steroid 6 (4.0 ␮M) was incubated for 30 min with human
placental microsomes (500 ␮g of protein), NADPH (240 ␮M), and 4.0 or
aromatase, while 2␤-methyl and 2-methylene analogs 5
and 17 were extremely poor substrates. The aromatization
of the good substrates was further studied to characterize
the affinity (Km) for the enzyme. Three 2␣-methyl sub-
strates 6, 26, and 28 showed a typical saturation curve with
increasing substrate concentration in each case (Fig. 7),
and the Lineweaver–Burk plot gave the apparent Km
and Vmax values shown in Table 3. The aromatization of
8
.0 ␮M concentrations of AD. The results are presented as the mean± S.D.
(n = 3). (*) P < 0.001, compared with control (0 ␮M).
2
␣-methyl steroid 6 was effectively blocked in the presence
of the natural substrate AD in a dose-dependent manner
Fig. 8).
(
4. Discussion
Table 3
Kinetic analysis of aromatization of 2␣-methylAD with human placental
a
microsomes
In order to define the effect of the C methyl substitution
2
of AD (1) on both the activity of aromatase inhibition and
the aromatase reaction (aromatization or 19-oxygenation),
a series of 2␣- and 2␤-methyl and 2,2-dimethylAD analogs
were synthesized and tested. The 2␤- and 2␣-methyl
steroids 5 and 6 and 2-methylene compound 17 were good
competitive inhibitors of aromatase in human placental
Substrate
Apparent Km (nM) Vmax ((pmol/min)/mg protein)
1
1
1
9-Methyl (6)
83 ± 15
82 ± 4.1
143 ± 9.0
197 ± 15
9-Hydroxy (26) 286 ± 11
9-oxo (28)
333 ± 38
a
◦
All incubations were performed at 37 C for 30 min with various
concentrations of the 2␣-methyl steroid and human placental microsomes
250 ␮g of protein) in the presence of NADPH (200 ␮M) as described in
the Section 2. Results are the means ± S.E.
microsomes (apparent K : 22 nM for 5, 55 nM for 6, and
i
(
68 nM for 17). The 2␤-methyl 5 was the most powerful

5
12
M. Numazawa et al. / Steroids 68 (2003) 503–513
among them, but its inhibitory activity was less than half
that of the 2,2-dimethyl analog 4, one of the most powerful
Enz
H+ H H C
H2C
H
O
H C
3
steroidal inhibitors [11] (K for 4: 8.8 nM). Introduction of
H2C
O
i
2
O , 2NADPH
2
a 2␤-methyl or 2␣-methyl group to AD did not change,
to a significant extent, the affinity for aromatase. Based
on these findings, a combination of 2␤- and 2␣-methyl
groups is essential for tight binding of inhibitor 4 to aro-
matase. The formation of such a thermodynamically stable
enzyme–inhibitor complex indicates that there is an accessi-
ble volume and hydrophobic binding pocket, corresponding
to the 2,2-dimethyl function of AD in the active site of
aromatase.
O
1
7
isomerization
H
O
C
H C
H C
3
3
O , NADPH
2
HO
O
Introduction of a hydroxy group to the C19 angular methyl
caused, to a greater extent, a decrease of affinity to aromatase
in both the 2,2-dimethyl and 2␣-methyl steroid series com-
pared to that observed in the series of the natural substrates
AD [2] and 16␣-hydroxyAD [29] (1/22 or 1/31 for the 2␣-
or 2,2-dimethyl series versus about 1:1/5 for the natural sub-
strates series). In addition, the conversion of the 19-hydroxy
group into the 19-oxo group increased the affinity in the
Fig. 9. Proposed mechanism for the conversion of 2-methylene steroid
7 into 2-methylestrogen.
1
positioning of the heme on the steroid nucleus resulting in
a failure of the oxygen delivery to the 19-position of the
steroids.
The aromatizations of steroids 6, 26, and 28 were ana-
lyzed kinetically (Table 3). The apparent Km and Vmax values
for the parent steroid 6 were 83 nM and 82 (pmol/min)/mg
protein, respectively. The 19-hydroxy and 19-oxo group
substitutions at C19 of steroid 6 increased the Vmax of the
production of 2-methylestrogen by about 1.7 or 2.4 times;
in contrast, these substitutions lowered, to a great extent,
the Km as seen in the AD series [2]. The Km for 19-methyl
2,2-dimethyl series in contrast to the natural substrates se-
ries [2,29], and the same conversion of the 2␣-methyl series
did not significantly affect the affinity. Based on these find-
ings, it is implied that the electronic nature of the binding
site in the vicinity of the C19 position of 2,2-dimethyl and
2
␣-methyl steroids 4 and 6 may be different from that of the
natural substrates.
9-Oxygenation of 2,2-dimethylAD (4) and aromatization
1
steroid 6 was similar to the K obtained by the inhibition
i
of 2-methyl and 2-methylene analogs 5, 6, and 17 with hu-
study, whereas the Km’s for 19-ol 26 and 19-al 28 were
man placental aromatase was next determined by GC–MS
much lower than the K ’s, respectively (Km/K : 0.24 for 26
i i
(
Table 2). Introduction of a 2,2-dimethyl group to AD com-
pletely blocked the aromatase reaction; production of the
9-hydroxy and/or 19-oxo compounds 26 and/or 27 was not
and 0.20 for 28). Thus, there was no significant correlation
between the ability to serve as a substrate and the ability
to serve as an inhibitor in the 2␣-methyl androgen series,
1
1
6
observed under the conditions employed (Table 1). The in-
troduction of a 2␤-methyl or 2-methylene group markedly
decreased the capacity to serve as a substrate of aromatase
where the production rate of 2-methylestrogen was less than
0
(
as seen in series of 6-alkylADs and their ꢀ -, ꢀ -, and
1,6
ꢀ
-analogs [26,31]. Based on the present results as well
as the previous ones, it is established that there is a marked
difference between a binding suitable for an inhibitor and
a binding suitable for a substrate in the active site of
aromatase.
.3% of the aromatization rate of the parent substrate AD
0.079 and 0.31 (pmol/min)/mg protein for 5 and 17, respec-
tively). In contrast, 2␣-methyl steroid 6 was a good sub-
strate for aromatase with about two-thirds the aromatization
rate of AD. The fact that the 2␣-methyl steroid aromatiza-
tion was effectively prevented by the natural substrate AD
indicated that these aromatization reactions were catalyzed
by aromatase enzyme. These results further suggest that a
2-Methylene steroid 17 was converted into 2-methyle-
strogen by aromatase, although the conversion rate was
very low. It is currently thought that the last step of sub-
strate AD aromatization, 1␤-hydrogen elimination and the
C10–C19 bond cleavage, proceeds through a 2,4-enolized
version of 19-oxo AD (3) [35]. A 2-methylene function
structurally blocks the formation of such a 2,4-enolized
version in the case of steroid 17. On the other hand, a
2-methylene-4-en-3-one compound can isomerize to yield
a 2-methyl-1,4-diene-3-one tautomer in a chemical sense.
Thus, it is implied that the 1,4-diene version may be cre-
ated before the cleavage of the C10–C19 bond of steroid 17
2
␤-methyl moiety of 2,2-dimethyl and 2␤-methyl steroids 4
and 5 would prevent their binding to aromatase in a proper
orientation probably through steric hindrance. It has been
predicted that the heme group of aromatase is positioned
so that the distance of the heme-iron atom from atoms C1,
C2 and C10 of AD will be about 7 Å, and the angle be-
tween the heme and the steroid A-ring mean planes will
1
and then aromatized via the sequence involved in ꢀ -AD
◦
be about 45 [32,33]. Oxygen delivery to the substrate is
aromatization; however, there is no direct evidence for this
(Fig. 9). The present results will be valuable for understand-
ing both the steric aspects of aromatase inhibitor binding
and the spatial structure of the active site of aromatase.
believed to occur from the out-of-ring position [34]. The
2
␤-methyl function that is oriented in a 1,3-diaxial rela-
tion to a 19-methyl function would sterically change the

M. Numazawa et al. / Steroids 68 (2003) 503–513
513
Acknowledgements
[16] Abul-Hajj YJ. Studies on the mechanism of action of the aromatase
inhibitor, 4-hydroxyandrostenedione. Steroids 1983;41:783–90.
[
17] Nedelec L, Gasc JC, Bucourt R. Kinetically controlled methylation
of conjugated polycyclic ketones. Tetrahedron 1974;30:3263–8.
18] Hosoda H, Yamashita K, Sagae H, Nambara T. Steroids. CI.
Dimethyl-tert-butylsilyl ethers of steroids. Chem Pharm Bull
1975;23:2118–22.
We are grateful to Dr. Hideo Imaizumi of Imaizumi
Hospital, Sendai, Japan, for supplying us with human
term placenta. This work was supported, in part, by a
Grant-in-Aid for Scientific Research from the Ministry of
Education, Science, Sports and Culture of Japan.
[
[19] Heusler K, Kalvoda J, Meystre Ch, Ueberwasser H, Wieland P,
Anner G, et al. A new preparation of 19-norsteroids. Experientia
1962;18:464–6.
[
20] Cushman M, He HM, Katzenellenbogen JA, Lin CM, Hamel E.
Synthesis, antitubulin and antimitotic activity, and cytotoxicity of
analogs of 2-methoxyestradiol, an endogenous mammalian metabolite
of estradiol that inhibits tubulin polymerization by binding to the
colchicine binding site. J Med Chem 1995;38:2041–9.
References
[
[
[
[
1] Thompson Jr EA, Siiteri PK. The involvement of human placental
microsomal cytochrome P-450 in aromatization. J Biol Chem
1
974;249:5373–8.
[
21] Birch AJ, Brown JM, Subba Rao GSR. Hydroaromatic steroid
hormones. X. Conversion of estrone into androst-4-ene-3,17-dione.
J Chem Soc 1964;3309–12.
2] Kellis Jr JT, Vickery LE. Purification and characterization of human
placental aromatase cytochrome P-450. J Biol Chem 1987;262:
4
413–20.
[
22] Nickisch K, Bittler D, Casals-Stenzel J, Laurent H, Nickolson R,
Nishino Y, et al. Aldosterone antagonists. 1. Synthesis and activities
of 6␤,7␤:15␤,16␤-dimethylene steroidal spirolactones. J Med Chem
3] Yoshida N, Osawa Y. Purification of human placental aromatase
cytochrome P-450 with monoclonal antibody and its characterization.
Biochemistry 1991;30:3003–10.
4] Covey DF. Aromatase inhibitors: specific inhibitors of estrogen
biosynthesis. In: Berg D, Plemel M, editors. Sterol biosynthesis
inhibitor: pharmaceutical and agrochemical aspects. Chichester, UK:
Ellis Horwood Ltd.; 1988. p. 534–71.
1
985;28:546–50.
23] Ryan KJ. Biological aromatization of steroids.
959;234:268–72.
[
[
[
J Biol Chem
1
24] Siiteri PK, Thompson Jr EA. Studies of human placental aromatase.
J Steroid Biochem 1975;6:317–22.
25] Numazawa M, Mutsumi A, Hoshi K, Oshibe M, Ishikawa E,
Kigawa H. Synthesis and biochemical studies of 16- or 19-substituted
androst-4-enes as aromatase inhibitors. J Med Chem 1991;34:2496–
[
5] Oh SS, Robinson CH. Mechanism of human placental aromatase: a
new active site model. J Steroid Biochem Mol Biol 1993;44:389–97.
6] Akhtar M, Lee-Robichaud P, Akhtar ME, Wright JN. The impact of
aromatase mechanism on other P450s. J Steroid Biochem Mol Biol
[
504.
1
997;61:127–32.
[
[
26] GraFit, version 5.0.3. Surrey, UK: Erithacus Software Limited; 2001.
27] Numazawa M, Yoshimura A, Oshibe M. Enzymic aromatization of
[
[
[
7] Vanden Bossche H. Inhibitors of P450-dependent steroid
biosynthesis: from research to medical treatment. J Steroid Biochem
Mol Biol 1992;43:1003–21.
8] Brodie AM, Santen RJ. Aromatase and its inhibitors in breast
cancer treatment-overview and perspective. Breast Cancer Res Treat
6-alkyl-substituted androgens, potent competitive and mechanism-
based inhibitors of aromatase. Biochem J 1998;329:151–6.
28] Numazawa M, Nagaoka M, Morio M, Kamiyama T. 19-Oxygenations
of 3-deoxy androgens, potent competitive inhibitors of estrogen
biosynthesis, with human placental aromatase. J Steroid Biochem
Mol Biol 1999;71:173–9.
[
1
994;30:1–6.
9] Brodie AMH, Njar VCO. Aromatase inhibitors in advanced breast
cancer: mechanism of action and clinical implications. J Steroid
Biochem Mol Biol 1998;66:1–10.
[
29] Numazawa M, Osada R, Tsuji M, Osawa Y. High-performance liquid
chromatographic determination of aromatization of 16␣-hydroxylated
androgens with human placental microsomes. Anal Biochem
[
[
[
[
10] Harper-Wynne C, Dowsett M. Recent advances in the clinical
application of aromatase inhibitors. J Steroid Biochem Mol Biol
1985;146:75–81.
2
001;76:179–86.
11] Furth PS, Rosenberger J, Marcotte PA, Robinson CH. Synthesis of
,2-dimethyl-4-hydroxy-4-androstene-3,17-dione as an inhibitor of
[
[
[
30] Ryan KJ, Placental synthesis of steroid hormone. In: Tulchinsky D,
Ryan KJ, editors. Maternal–fetal endocrinology. Philadelphia, USA:
2
W.B. Sanders; 1980. p. 3–16.
aromatase. J Enzyme Inhibit 1990;4:131–5.
31] Numazawa M, Yoshimura A. Biological aromatization of ꢀ4,6-
12] Numazawa M, Tachibana M. A- or B-ring-substituted derivatives of
androst-4-ene-3,6,17-trione as aromatase inhibitors. Structure-activity
relationships. Steroids 1994;59:579–85.
1,4,6
and ꢀ
-androgens and their 6-alkyl analogs, potent inhibitors of
aromatase. J Steroid Biochem Mol Biol 1999;70:189–96.
32] Kellis Jr JT, Vickery LE. The active site of aromatase cytochrome
P-450. Differential effects of cyanide provide evidence for proximity
of heme-iron and carbon-19 in the enzyme–substrate complex. J Biol
Chem 1987;262:8840–4.
33] Laughton CA, Zvelebil MJJM, Neidle S. A detailed molecular model
for human aromatase. J Steroid Biochem Mol Biol 1993;44:399–407.
34] Duax WL, Osawa Y. Steroid structure and function. VI. The
molecular conformation of 19-hydroxy-4-androstene-3,17-dione, as
13] Numazawa M, Midzuhashi K, Nagaoka M. Metabolic aspects of the
1␤-proton and the 19-methyl group of androst-4-ene-3,6,17-trione
during aromatization by placental microsomes and inactivationof
aromatase. Biochem Pharmacol 1994;47:717–26.
[
[14] Numazawa M, Mutsumi A, Tachibana M. Mechanism for aromatase
inactivation by a suicide substrate, androst-4-ene-3,6,17-trione. The
[
4␤,5␤-epoxy-19-oxo derivative as a reactive electrophile irreversibly
binding to the active site. Biochem Pharmacol 1996;52:1253–9.
15] Covey DF, Hood WF. Aromatase enzyme catalysis is involved in
the potent inhibition of estrogen biosynthesis caused by 4-acetoxy-
and 4-hydroxy-4-androstene-3,17-dione. Mol Pharmacol 1982;21:
an intermediate for estrogen synthetase.
980;13:383–6.
35] Cole PA, Robinson CH. Mechanism and inhibition of cytochrome
J Steroid Biochem
[
1
[
P-450 aromatase. J Med Chem 1990;33:2933–44.
1
73–80.