Probing the binding pocket of the active site of aromatase with 2-phenylaliphatic androsta-1,4-diene-3,17-dione steroids

Article abstract of DOI:10.1016/j.steroids.2010.01.008

A series of 2-phenylaliphatic-substituted androsta-1,4-diene-3,17-diones (6) as well as their androstenedione derivatives (5) were synthesized as aromatase inhibitors to gain insights of structure-activity relationships of varying the alkyl moiety (C₁ to C₄) of the 2-phenylaliphatic substituents as well as introducing a methyl- or trifluoromethyl function to p-position of a phenethyl moiety to the inhibitory activity. The inhibitors examined showed a competitive type inhibition. The 2-phenpropylandrosta-1,4-diene 6c was the most powerful inhibitor (K_i: 16.1 nM) among them. Compounds 6c along with the phenethyl derivative 6b caused a time-dependent inactivation of aromatase (k_inact: 0.0293 and 0.0454 min^-1 for 6b and 6c, respectively). The inactivation was prevented by the substrate androstenedione, and no significant effect of l-cysteine on the inactivation was observed in each case. Molecular docking of the phenpropyl compound 6c to aromatase was conducted to demonstrate that the phenpropyl group orients to a hydrophobic binding pocket in the active site to result in the formation of thermodynamically stable enzyme-inhibitor complex.

Full text of DOI:10.1016/j.steroids.2010.01.008

Steroids 75 (2010) 330–337
Contents lists available at ScienceDirect
Steroids
journal homepage: www.elsevier.com/locate/steroids
Probing the binding pocket of the active site of aromatase with 2-phenylaliphatic
androsta-1,4-diene-3,17-dione steroids
Madoka Takahashi, Kouwa Yamashita, Mitsuteru Numazawa∗
Tohoku Pharmaceutical University, 4-1 Komatsushima-4-chome, Aobaku, Sendai 981-8558, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of 2-phenylaliphatic-substituted androsta-1,4-diene-3,17-diones (6) as well as their androstene-
dione derivatives (5) were synthesized as aromatase inhibitors to gain insights of structure–activity
relationships of varying the alkyl moiety (C1 to C4) of the 2-phenylaliphatic substituents as well as intro-
ducing a methyl- or trifluoromethyl function to p-position of a phenethyl moiety to the inhibitory activity.
The inhibitors examined showed a competitive type inhibition. The 2-phenpropylandrosta-1,4-diene 6c
was the most powerful inhibitor (Ki: 16.1 nM) among them. Compounds 6c along with the phenethyl
derivative 6b caused a time-dependent inactivation of aromatase (kinact: 0.0293 and 0.0454 min−1 for
6b and 6c, respectively). The inactivation was prevented by the substrate androstenedione, and no sig-
nificant effect of l-cysteine on the inactivation was observed in each case. Molecular docking of the
phenpropyl compound 6c to aromatase was conducted to demonstrate that the phenpropyl group ori-
ents to a hydrophobic binding pocket in the active site to result in the formation of thermodynamically
stable enzyme–inhibitor complex.
Received 18 November 2009
Received in revised form 7 January 2010
Accepted 14 January 2010
Available online 22 January 2010
Keywords:
Aromatase
Competitive inhibitor
Androsta-1,4-diene-3,17-dione
2-Phenylaliphatic substituent
Suicide substrate
Breast cancer
© 2010 Elsevier Inc. All rights reserved.
1. Introduction
suicide substrates for aromatase (Fig. 2). We have studied the
structure–activity relationships of 6-substituted (alkyl, pheny-
Aromatase is a cytochrome P-450 enzyme that catalyzes the
conversion of the androgens, androstenedione (AD) and testos-
terone, to the estrogens, estrone and estradiol, respectively [1–3].
The aromatase reaction is thought to proceed through three
sequential oxygenations at C-19 of the androgens [4–7] (Fig. 1). Aro-
matase is a potential therapeutic target for the lowering of estrogen
levels in patients with advanced estrogen-dependent breast cancer
[8–11]. The specific blockade of estrogen biosynthesis has been pur-
sued intensely with the goal of developing practical clinical drugs.
For this reason, a number of powerful aromatase inhibitors, which
are analogs of the natural substrate AD, have been described by
various laboratories.
laliphatic, alkoxy, and ester) ꢀ1-AD analogs [27–29] as well as
19-substituted ones [30] as aromatase inhibitors, some of which
are among the most potent competitive inhibitors and mechanism-
based inhibitors reported so far. The length (bulkiness) and the
stereochemistry of the 6-substituents as well as the electronic
effects of the 19-substituents play a critical role not only in the
binding of inhibitors to the active site of aromatase, but also in
the cause of the mechanism-based inactivation of aromatase by
the 1,4-dien-3-one steroid 1. Recently, we have also reported the
structure–activity relationships of the 2-alkyl or alkoxy substituted
ꢀ1-ADs as aromatase inhibitors [31]. Among the 2-substituted
derivatives, the 2-hexyl compound is the most powerful compet-
itive inhibitor and inactivates aromatase with the lowest KI value
in a time-dependent manner.
Androsta-1,4-diene-3,17-dione (ꢀ1-AD, 1), which is the 1-ene
derivative of AD, is one of the prototypical mechanism-based
inhibitors (suicide substrates); the C(1) C(2) double bond is the
structural feature responsible for aromatase inactivation, although
the inactivation mechanism is currently unknown [5,12–21]. Sev-
eral analogs of compound 1, having a D-ring lactone (testolactone)
[22,23], 1-methyl (atamestane) [24] or 6-methylene (exemes-
tane) [25,26] structure, are clinically evaluated as orally active
In the cause of our studies on aromatase inhibition by ꢀ1-
AD analog, we were interest in C-2 substitution of the ꢀ1-AD 1
by a phenylaliphatic group. Thus, we synthesized 2-substituted
(C6H5CHn, n = 1–4) ꢀ1-AD derivatives 6a–6d along with p-methyl
and p-trifluoromethylphenethyl analogs 6e and 6f, and tested their
ability to inhibit aromatase activity as well as their ability to
inactivate aromatase in a suicide manner. Furthermore, molecular
docking studies of the phenpropyl derivative 6c, the most power-
ful inhibitor among them, with human placental aromatase protein
were performed.
∗
Corresponding author. Tel.: +81 22 727 0138; fax: +81 22 727 0137.
E-mail address: numazawa@tohoku-pharm.ac.jp (M. Numazawa).
0039-128X/$ – see front matter © 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.steroids.2010.01.008

M. Takahashi et al. / Steroids 75 (2010) 330–337
331
resulting solutions washed with water, dilute H2SO4 and water,
and dried with Na2SO4. Evaporation of the solvent gave phenbutyl
iodide, p-methylphenethyl iodide and p-trifluoromethyl iodide.
(A) To a stirring solution of 160 ␮l (1.13 mmol) of diisopropylamine
in 1.6 ml of THF, cooled to 0 ◦C, was added BuLi (690 ␮L of 1.6 M
in hexane). After 20 min, the solution was cooled to −20 ◦C and
200 mg (0.50 mmol) of compound 2 in 1.6 ml THF was added
via cannula to the solution. After an additional 15 min, 170 ␮l
(0.98 mmol) of hexamethylphosphoric triamide (HMPA) was
added all at once and the mixture was allowed to stir for another
20 min, and then 2.02 mmol of benzyl bromide or phenbutyl
iodide was added dropwise to the mixture [34]. After 1 h at
−20 ◦C reaction, the reaction mixture was quenched with 1 M
aqueous HCl and diluted with EtOAc. The organic phase was
washed with brine and dried with Na2SO4. Evaporation of the
solvent gave an oil which was purified by silica gel (20 g) col-
umn chromatography with hexane–EtOAc (60:1) to yield an
oily substance. Then, the oil was subjected to preparative TLC
(hexane–EtOAc, 8:1, multiple developments) to give 2␣-benzyl
steroid 3a (64%) or 2␣-phenbutyl steroid 3d (22%).
Fig. 1. Aromatization sequence of AD with human placental aromatase.
2. Experimental
2.1. Materials and general methods
17␤-tert-Butyldimethylsiloxyandrost-4-en-3-one (2) [32] was
synthesized according to the previous method. [1␤-3H]AD (specific
activity 27.5 Ci/mmol; 3H-distribution 74–79%) was obtained from
New England Nuclear (Boston, MA, USA). NADPH was purchased
from Sigma–Aldrich (St. Louise, MO, USA).
Melting points were measured on a Yanagimoto melting point
apparatus (Kyoto, Japan) and are uncorrected. Infra red (IR) spectra
were recorded on a Perkin-Elmer FT-IR 1725X spectrophotometer
in a KBr pellet or nujol form, and ultra violet (UV) spectra were
determined in 95% EtOH on a Hitachi 150-20 spectrophotometer
(Tokyo, Japan). 1H nuclear magnetic resonance (NMR) spectra were
obtained in CDCl3 solution with a JEOL LA 400 (400 MHz) and JEOL
LA 600 (600 MHz) spectrometers (Tokyo, Japan) using tetramethyl-
silane as an internal standard, and mass (MS) spectra (electron
impact, EI mode) and high resolution (HR)-MS with a JEOL JMS-DX
303 spectrometer. Thin-layer chromatography (TLC) was per-
formed on E. Merck precoated silica gel plates (silica gel 60F-254,
Darmstadt, Germany). Column chromatography was conducted
with silica gel 60, 70–230 mesh (E. Merck). High-performance liq-
uid chromatography (HPLC) was carried out using a Waters 600
pump and a UV detector (240 nm) (Milford, MA, USA) where Sym-
metry column (300 mm × 7.8 mm i.d., Waters) was employed for
preparative use and Puresil column (150 mm × 4.6 mm i.d., Waters)
for purity determination of compounds.
(B) LDA (1.13 mmol) was prepared similar to the above at 0 ◦C.
A solution of 1 g (2.48 mmol) of compound 2 in 8.00 ml
THF was added via cannula to the LDA solution at 0 ◦C.
After the same work-up as the above (78 mmol of HMPA),
13.79 mmol of phenethyl iodide, p-methylphenethyl iodide or
p-trifluoromethyl iodide was added dropwise at 0 ◦C. After 1.5 h
reaction, the reaction mixture was quenched with 1 M aqueous
HCl and diluted with EtOAc. The organic phase was washed with
brine and dried with Na2SO4. Evaporation of the solvent gave an
oil which was purified bysilicagel (100 g)columnchromatogra-
phy with hexane–EtOAc (30:1) to yield an oily substance. Then,
the oil was subjected to preparative TLC (hexane–EtOAc, 5:1)
to give 2␣-phenethyl steroid 3b, 2␣-p-methylphenethyl steroid
3e or 2␣-p-trifluoromethyl steroid 3f (11–22%).
(C) A mixture of 17␤-tert-butyldimethylsiloxyandrost-4-en-3-one
(2) (700 mg, 1.74 mmol), 1-iodo-3-phenyl propane (3.6 ml,
23.16 mmol), THF (1.5 ml) and HMPA (0.28 ml) was cooled
−60 ◦C and then tert-BuOK (390 mg, 3.48 mmol) in THF
(1.5 ml)–HMPA (0.28 ml) was added to this solution [35]. The
reaction mixture was stirred at −60 ◦C for 1.5 h and the temper-
ature was gradually elevated to room temperature. The mixture
was extracted of with EtOAc, washed with saturated NaHCO3
solution and saturated NaCl, and dried with Na2SO4. Evapora-
tion of the solvent gave an oil which was purified by silica gel
(70 g) column chromatography with hexane–EtOAc (60:1) to
yield an oily substance. Then, the oil was subjected to prepara-
tive TLC (hexane–EtOAc, 20:1, multiple developments) to give
2␣-phenpropyl steroid 3c (193 mg, 22%).
2.2. 2˛-Phenalkyl-17ˇ-tert-butyldimethylsiloxyandrost-
4-en-3-ones (3)
Phenbutyl bromide, p-methylphenethyl bromide and p-
trifluoromethyl bromide (Sigma–Aldrich. Co.) (5.84 mmol) was
refluxed in acetone (15.6 ml) containing NaI (2.3 g) for 6 h [33],
and the solvent removed in a vacuum. Hexane was added, and the
The purities of the oily substances were determined by analyti-
cal HPLC (CH3CN:H2O = 100:0, v/v; flow rate, 1.0 ml/min) to be more
than 98%.
3a: oil; tR = 19.4 min; IR (KBr): 1672 cm−1 (C O); UV ꢁmax
242 nm (ε = 14,700); 1H NMR ı: 0.01 (6H, s, Si–(CH3)2), 0.71 (3H, s,
18-Me), 0.88 (9H, s, Si–C(CH3)3), 1.10 (3H, s, 19-Me), 2.45 (1H, dd,
J = 9.2 and 13.9 Hz, phenyl-CHa), 3.48 (1H, dd, J = 3.8 and 14.0 Hz,
phenyl-CHb), 3.55 (1H, t, J = 8.2 Hz, 17␣-H), 5.76 (1H d, J = 1.5 Hz,
4-H), 7.18–7.31 (5H, m, aromatic protons); MS m/z: 492 (M+, 14),
435 (100), 359 (12), 91 (16). HR-MS for C32H48O2Si (M+): calcd
492.3424, found 492.3436.
3b: oil; tR = 24.9 min; IR (KBr): 1672 cm−1 (C O); UV ꢁmax
242 nm (ε = 14,400); 1H NMR ı: 0.01 (6H, s, Si–(CH3)2), 0.75 (3H,
s, 18-Me), 0.88 (9H, s, Si–C(CH3)3), 1.16 (3H, s, 19-Me), 2.68 (2H,
m, phenyl-CH2), 3.56 (1H, m, 17␣-H), 5.69 (1H, d, J = 1.2 Hz, 4-H),
7.17–7.29 (5H, m, aromatic protons); MS m/z: 506 (M+, 1), 499 (68),
Fig. 2. Structures of suicide substrates of aromatase having a 1,4-dien-3-one struc-
ture.

332
M. Takahashi et al. / Steroids 75 (2010) 330–337
402 (100), 373 (9). HR-MS for C33H50O2Si (M+): calcd 506.3580,
found 506.3576.
(3H, s, 19-Me), 2.63 (2H, m, phenyl-CH2), 3.65 (1H, m, 17␣-H), 5.70
(1H, d, J = 1.5 Hz, 4-H), 7.15–7.29 (5H, m, aromatic protons); MS m/z:
420 (M+, 100), 301 (38), 288 (98), 91 (55). HR-MS for C29H40O2:
(M+): calcd 420.3028, found 420.3044.
3c: oil; tR = 24.2 min; IR (neat): 1669 cm−1 (C O); UV ꢁmax
242 nm (ε = 8200); 1H NMR ı: 0.01 (6H, s, Si–(CH3)2), 0.74 (3H, s,
18-Me), 0.88 (9H, s, Si–C(CH3)3), 1.17 (3H, s, 19-Me), 2.63 (2H, m,
phenyl-CH2), 3.54 (1H, t, J = 8.3 Hz, 17␣-H), 5.69 (1H, d, J = 1.5 Hz,
4-H), 7.16–7.32 (5H, m, aromatic protons); MS m/z: 520 (M+, 86),
463 (100), 402 (20), 387 (14). HR-MS for C34H52O2Si (M+): calcd
520.3737, found 520.3748.
4e: oil; tR = 7.0 min; IR (KBr): 1668 cm−1 (C O); UV ꢁmax: 242 nm
(ε = 10,600); 1H NMR ı: 0.80 (3H, s, 18-Me), 1.17 (3H, s, 19-Me), 2.31
(3H, s, p-Me-phenethyl), 2.63 (2H, m, p-Me-phenyl-CH2), 3.66 (1H,
m, 17␣-H), 5.69 (1H, d, J = 1.5 Hz, 4-H), 7.07–7.11 (4H, m, aromatic
protons); MS m/z: 406 (M+, 2), 288 (100), 273 (13), 122 (28). HR-MS
for C28H38O2 (M+): calcd 406.2872, found 406.2883.
3d: oil; tR = 24.1 min; IR (KBr): 1675 cm−1 (C O); UV ꢁmax
242 nm (ε = 13,200); 1H NMR ı: 0.01 (6H, s, Si–(CH3)2), 0.74 (3H,
s, 18-Me), 0.88 (9H, s, Si–C(CH3)3), 1.18 (3H, s, 19-Me), 2.62 (2H,
m, phenyl-CH2), 3.55 (1H, t, J = 8.4 Hz, 17␣-H), 5.69 (1H, s, 4-
H), 7.16–7.29 (5H, m, aromatic protons); MS m/z: 534 (M+, 48),
477 (100), 402 (35), 75 (41). HR-MS for C35H54O2Si (M+): calcd
534.3893, found 534.3885.
4f: oil; tR = 6.8 min; IR (KBr): 1665 cm−1 (C O); UV ꢁmax: 242 nm
(ε = 9800); 1H NMR ı: 0.80 (3H, s, 18-Me), 1.22 (3H, s, 19-Me), 2.75
(2H, t, J = 8.0 Hz, p-CF3-phenyl-CH2), 3.64 (1H, m, 17␣-H), 5.71 (1H,
d, J = 1.5 Hz, 4-H), 7.32 and 7.53 (2H each, d, J = 8.1 Hz, aromatic pro-
tons); MS m/z: 460 (M+, 1), 345 (7), 288 (100), 273 (11). HR-MS for
C
28H35F3O2 (M+): calcd 460.2589, found 460.2596.
3e: oil; tR = 25.3 min; IR (neat): 1675 cm−1 (C O); UV ꢁmax
242 nm (ε = 11,000); 1H NMR ı: 0.01 (6H, s, Si–(CH3)2), 0.75 (3H,
s, 18-Me), 0.88 (9H, s, Si–C(CH3)3), 1.16 (3H, s, 19-Me), 2.31 (3H,
m, p-Me-phenethyl), 2.63 (2H, m, p-Me-phenyl-CH2), 3.55 (1H, t,
J = 8.3 Hz, 17␣-H), 5.69 (1H, d, J = 1.5 Hz, 4-H), 7.05–7.11 (4H, m, aro-
matic protons); MS m/z: 520 (M+, 2), 463 (18), 402 (100), 345 (32).
HR-MS for C34H52O2Si (M+): calcd 520.3737, found 520.3732.
3f: oil; tR = 22.4 min; IR (neat): 1672 cm−1 (C O); UV ꢁmax
242 nm (ε = 10,700); 1H NMR ı: 0.01 (6H, s, Si–(CH3)2), 0.75 (3H,
s, 18-Me), 0.88 (9H, s, Si–C(CH3)3), 1.17 (3H, s, 19-Me), 2.75 (2H,
t, J = 7.8 Hz, p-CF3-phenyl-CH2), 3.55 (1H, m, 17␣-H), 5.70 (1H, d,
J = 1.2 Hz, 4-H), 7.32 and 7.53 (2H each, d, J = 8.1 Hz, aromatic pro-
tons); MS m/z: 574 (M+, 1), 517 (18), 161 (54), 153 (100). HR-MS
for C34H49F3O2Si (M+): calcd 574.3453, found 574.3435.
2.4. 2˛-phenalkylandrost-4-ene-3,17-diones (5)
Jones reagent (11 drops) was added to a solution of the 17␤-
hydroxide 4 (0.13mmol) in acetone (20 ml) under ice-cooling and
the mixture was stirred for 3 min. After this time, MeOH (0.1 ml)
was added to the mixture and then extracted with EtOAc (10 ml),
washed with 5% NaHCO3 solution and water, dried with Na2SO4.
Evaporation of the solvent gave an oil which was purified by
silica gel (5 g) column chromatography (hexane–EtOAc = 7:1) to
yield an oily substance. Then, the oily substance was subjected
to preparative TLC (hexane–EtOAc, 4:1, multiple developments) or
preparative HPLC (CH3CN:H2O = 60:40, v/v; flow rate 5 ml/min) to
obtained compound 5a–5f (49–61%). The purities of oil substances
were obtained by analytical HPLC as described for those of com-
pounds 4 were more than 97%.
2.3. 2˛-Phenalkyl-17ˇ-hydroxyandrost-4-en-3-ones (4)
5a: mp 164–166◦C. IR (KBr): 1664 and 1733 cm−1 (C O); UV
ꢁmax: 241 nm (ε = 13,300); 1H NMR ı: 0.87 (3H, s, 18-Me), 1.11 (3H,
s, 19-Me), 2.45 (1H, m, phenyl-CHa), 3.48 (1H, dd, J = 3.8 and 14.1 Hz,
phenyl-CHb), 5.78 (1H, d, J = 1.8 Hz, 4-H), 7.17–7.30 (5H, m, aromatic
protons); MS m/z: 376 (M+, 100), 361 (7), 121 (40), 91 (78). Anal.
Calcd for C26H32O2: C, 82.94; H, 8.57. Found C, 83.01; H, 8.80.
5b: mp 129–131 ◦C. IR (KBr): 1670 and 1742 cm−1 (C O); UV
ꢁmax: 240 nm (ε = 15,200); 1H NMR ı: 0.93 (3H, s, 18-Me), 1.18 (3H,
s, 19-Me), 2.70 (2H, m, phenyl-CH2), 5.72 (1H, d, J = 1.5 Hz, 4-H),
7.17–7.29 (5H, m, aromatic protons); 13C NMR ı: 13.7, 17.4, 20.2,
21.7, 30.4, 30.7, 31.3, 32.1, 33.0, 35.0, 35.7, 39.2, 41.1, 42.0, 47.5,
50.8, 54.2, 124.0, 125.8, 128.3, 128.4, 142.0, 168.6, 200.8, 220.4;
MS m/z: 390 (M+, 2), 286 (100), 271 (16), 122 (27). Anal. Calcd for
5% HCl (1.1 ml) was added to a solution of the 17␤-siloxy com-
pounds 3 (0.26 mmol) in THF (2 ml) and 2-propanol (2 ml) and
the reaction mixture was stand at room temperature for 6 h. After
adding NaHCO3, the mixture was extracted with EtOAc (50 ml),
washed with 5% NaHCO3 solution and water, and dried with
Na2SO4. Evaporation of the solvent gave a 17␤-hydroxy product
which was purified by silica gel (10 g) column chromatography
(hexane–EtOAc = 5:1) followed by recrystallization from acetone
or trituration from ether to yield 17-hydroxy compounds 4a–4d
(35–68%). The purities of the oily substances (4e and 4f) were
determined by analytical HPLC (CH3CN:H2O = 80:20, v/v; flow rate,
1.0 ml/min) to be more than 98%.
4a: mp 252–254 ◦C. IR (KBr): 1672 cm−1 (C O); UV ꢁmax: 242 nm
(ε = 10,300); 1H NMR ı: 0.74 (3H, s, 18-Me), 1.10 (3H, s, 19-Me), 2.43
(1H, dd, J = 9.3 and 14.1 Hz, phenyl-CHa), 3.47 (1H, dd, J = 3.8 and
14.0 Hz, phenyl-CHb), 3.63 (1H, m, 17␣-H), 5.75 (1H, d, J = 1.5 Hz,
4-H), 7.16–7.30 (5H, m, aromatic protons); MS m/z: 378 (M+, 100),
363 (5), 121 (31), 91 (50). Anal. Calcd for C26H34O2: C, 82.49; H,
9.05. Found C, 82.60; H, 9.27.
C
27H34O2: C, 83.03; H, 8.77. Found C, 83.34 H, 8.58.
5c: oil; tR = 7.2 min; IR (neat): 1671 and 1739 cm−1 (C O); UV
ꢁmax: 241 nm (ε = 15,100); 1H NMR ı: 0.91 (3H, s, 18-Me), 1.20
(3H, s, 19-Me), 2.63 (2H, m, phenyl-CH2), 5.72 (1H, d, J = 1.5 Hz, 4-
H), 7.16–7.29 (5H, m, aromatic protons); 13C NMR ı: 13.7, 17.4,
20.2, 21.7, 28.7, 29.0, 30.7, 31.2, 32.1, 35.0, 35.7, 36.3, 39.1, 41.8,
41.9, 47.5, 50.8, 54.2, 124.0, 125.7, 128.3, 128.4, 142.5, 168.7, 200.8,
220.5; MS m/z: 404 (M+, 100), 300 (72), 286 (85), 271 (18). HR-MS
for C28H36O2 (M+): calcd 404.2715, found 404.2715.
4b: mp 143–145 ◦C. IR (KBr): 1668 cm−1 (C O); UV ꢁmax
:
242 nm (ε = 11,200); 1H NMR ı: 0.80 (3H, s, 18-Me), 1.17 (3H, s,
19-Me), 2.69 (2H, m, phenyl-CH2), 3.65 (1H, m, 17␣-H), 5.70 (1H, d,
J = 1.5 Hz, 4-H), 7.16–7.29 (5H, m, aromatic protons); MS m/z: 392
(M+, 2), 288 (100), 273 (8), 122 (25). Anal. Calcd for C27H36O2: C,
82.61; H, 9.24. Found C, 82.70; H, 9.37.
5d: semi-crystalline; mp 117–119 ◦C. IR (KBr): 1665 and
1743 cm−1 (C O); UV ꢁmax: 241 nm (ε = 13,800); 1H NMR ı: 0.92
(3H, s, 18-Me), 1.21 (3H, s, 19-Me), 2.63 (2H, m, phenyl-CH2), 5.72
(1H, d, J = 1.5 Hz, 4-H), 7.15–7.29 (5H, m, aromatic protons); MS m/z:
418 (M+, 100), 299 (30), 286 (85), 271 (13). HR-MS for C29H38O2
(M+): calcd 418.2872, found 418.2859.
4c: mp 199–201 ◦C. IR (KBr): 1667 cm−1 (C O); UV ꢁmax: 241 nm
(ε = 9000); 1H NMR ı: 0.79 (3H, s, 18-Me), 1.18 (3H, s, 19-Me), 2.63
(2H, m, phenyl-CH2), 3.64 (1H, m, 17␣-H), 5.69 (1H, d, J = 1.4 Hz,
4-H), 7.15–7.29 (5H, m, aromatic protons); MS m/z: 406 (M+, 100),
301 (30), 288 (53), 273 (6). Anal. Calcd for C28H38O2: C, 82.71; H,
9.42. Found C, 82.55; H, 9.54.
5e: oil; tR = 7.7 min; IR (KBr): 1671 and 1739 cm−1 (C O); UV
ꢁmax: 241 nm (ε = 12,000); 1H NMR ı: 0.93 (3H, s, 18-Me), 1.19
(3H, s, 19-Me), 2.31 (3H, s, p-Me-phenethyl), 2.65 (2H, m, p-Me-
phenyl-CH2), 5.72 (1H, d, J = 1.5 Hz, 4-H), 7.07–7.11 (4H, m, aromatic
protons). 13C NMR ı: 13.7, 17.4, 20.2, 21.0, 21.7, 30.5, 31.3, 32.0,
32.5, 35.0, 35.7, 39.2, 41.1, 42.0, 47.5, 50.8, 54.2, 124.0, 128.3, 129.0,
4d: semi-crystalline; mp 87–89 ◦C. IR (KBr): 1669 cm−1 (C O);
UV ꢁmax: 242 nm (ε = 15,300); 1H NMR ı: 0.80 (3H, s, 18-Me), 1.19

M. Takahashi et al. / Steroids 75 (2010) 330–337
333
135.2, 138.9, 168.5, 200.8, 220.4; MS m/z: 404 (M+, 2), 286 (100),
271 (14), 122 (21). HR-MS for C28H36O2 (M+): calcd 404.2715, found
404.2722.
55), 297 (73), 284 (64), 105 (100). HR-MS for C28H34O2 (M+): calcd
402.2559, found 402.2556.
6f: oil; tR = 5.2 min; IR (KBr): 1630 cm−1 (C C), 1664 and
1740 cm−1 (C O); UV ꢁmax: 250 nm (ε = 11,000); 1H NMR ı: 0.90
(3H, s, 18-Me), 1.08 (3H, s, 19-Me), 2.87 (2H, m, p-CF3-phenyl-
CH2), 6.09 (1H, d, J = 1.0 Hz, 4-H), 6.49 (1H, s, 1-H), 7.23 and 7.50
(2H each, d, J = 8.1 Hz, aromatic protons). 13C NMR ı: 13.7, 18.7,
21.9, 22.0, 31.1, 31.4, 32.1, 32.4, 34.2, 35.0, 35.6, 43.1, 47.6, 50.4,
52.6, 123.9, 125.00, 125.03, 125.1, 129.1, 135.7 145.8, 151.8, 167.9,
186.0, 219.8; MS m/z: 456 (M+, 62), 438 (6), 297 (57), 284 (100).
HR-MS for C28H31F3O2 (M+): calcd 456.2276, found 456.2278.
5f: oil; tR = 7.4 min; IR (KBr): 1669 and 1739 cm−1 (C O); UV
ꢁmax: 241 nm (ε = 11,000); 1H NMR ı: 0.93 (3H, s, 18-Me), 1.20
(3H, s, 19-Me), 2.75 (2H, t, J = 8.0 Hz, p-CF3-phenyl-CH2), 5.74 (1H,
d, J = 1.2 Hz, 4-H), 7.32 and 7.53 (2H each, d, J = 7.8 Hz, aromatic pro-
tons). 13C NMR ı: 13.7, 17.4, 20.2, 21.7, 30.4, 30.7, 31.3, 32.1, 32.9,
35.0, 35.7, 39.3, 41.1, 42.1, 47.5, 50.8, 54.2, 124.0, 125.21, 125.24,
125.3, 128.7, 146.3, 168.8, 200.4, 220.3; MS m/z: 458 (M+, 1), 439 (4),
286 (100), 271 (48). HR-MS for C28H33 F3O2 (M+): calcd 458.2433,
found 458.2430.
2.6. Enzyme preparation
2.5. 2-Alkylandrosta-1,4-diene-3,17-diones (6)
Human placental microsomes (particles sedimenting at
105,000 × g for 60 min) were obtained using the method reported
by Ryan [12]. They were washed once with 0.05 mM dithiothreitol
solution, lyophilized, and stored at −80 ◦C. No significant loss of
activity occurred during this study (2 months).
DDQ (0.15 mmol) was added to a solution of the 4-ene-
3,17-diones 5 (0.10 mmol) in dioxane (2 ml) and the mixture
was heated under reflux for 19 h. After this time, the mixture
was treated with Al2O3 (1 g) column (EtOAc) and the steroidal
material was submitted to purification with preparative TLC
(hexane–EtOAc = 4:1, multiple developments) or preparative HPLC
(CH3CN:H2O = 60:40 (v/v); flow rate 5 ml/min) to obtained com-
pound 6a–6f (29–49%). The purities of oil substances were obtained
by analytical HPLC as described for those of compounds 4 to be more
than 98%.
2.7. Aromatase assay procedure
Aromatase activity was measured according to the procedure of
Siiteri and Thompson [36]. The screening assay for determination
of IC50 value, the kinetic assay, and the time-dependent assay were
carried out essentially according to the assay methods described
in our previous work. Briefly, 20 ␮g of protein of the lyophilized
microsomes and 20-min incubation time for the screening assay,
and 20 ␮g of protein of the microsomes and 5-min incubation time
for the kinetic assay, respectively, were employed in this study,
and the assays were carried out in 67 mM phosphate buffer in the
presence of NADPH in air [27]. In the time-dependent inactivation
experiment, 1/10 of the incubation mixture was used for assays of
the remaining aromatase activity.
6a: mp 165–168 ◦C. IR (KBr): 1633 cm−1 (C C), 1662 and
1739 cm−1 (C O); UV ꢁmax: 250 nm (ε = 15,200); 1H NMR ı: 0.91
(3H, s, 18-Me), 1.19 (3H, s, 19-Me), 3.65 (2H, m, phenyl-CH2), 6.10
(1H, d, J = 1.1 Hz, 4-H), 6.60 (1H, s, 1-H), 7.18–7.30 (5H, m, aromatic
protons); 13C NMR ı: 13.8, 18.8, 22.0, 22.2, 31.2, 32.1, 32.4, 35.1,
35.2, 35.7, 43.3, 47.7, 50.4, 52.6, 124.0, 126.2, 128.4, 129.1, 137.5,
139.5, 151.6, 167.7, 185.8, 220.0; MS m/z: 374 (M+, 80), 359 (36),
223 (19), 210 (100). Anal. Calcd for C26H30O2: C, 83.38; H, 8.07.
Found C, 83,13; H, 8.28.
6b: mp 197–199◦C; IR (KBr): 1629 cm−1 (C C), 1664 and
1739 cm−1 (C O); UV ꢁmax: 250 nm (ε = 14,300); 1H NMR ı: 0.89
(3H, s, 18-Me), 1.09 (3H, s, 19-Me), 2.77 (2H, m, phenyl-CH2), 6.08
(1H, d, J = 1.0 Hz, 4-H), 6.40 (1H, s, 1-H), 7.08–7.24 (5H, m, aromatic
protons); 13C NMR ı: 13.7, 18.6, 21.8, 21.9, 31.1, 31.4, 32.1, 32.4,
34.2, 35.0, 35.6, 43.0, 47.7, 50.4, 52.6, 124.0, 125.7, 128.1, 128.9,
135.8, 141.5, 151.8, 167.6, 186.1, 220.2; MS m/z: 388 (M+, 73), 371
(16), 297 (73), 284 (100). Anal. Calcd for C27H32O2: C, 83.46; H, 8.30.
Found: C, 83.43: H, 8.32.
2.8. Molecular modeling
The crystal structure of human placental aromatase was down-
loaded from the Protein Data Bank (Accession Code, 3EQM).
Molecular modelings, 3D structures, of ligands were built by
Spartan’08 software (Wavefunction, Inc., Irvine, CA, USA) using
Hartree-Fock 6-31G* calculations. Molecular docking was per-
formed by AutoDock Vina (The Scripps Research Institute, CA, USA)
[37]. Docking to aromatase was carried out to propose docked
structures with similar calculated affinities (−3.1 to −5.7 kcal/mol).
The primary criterion used in choosing best docked structure was
the position of the ꢀ1-AD derivative relative to the active site
aspartate (Asp 309), arginine (Arg 115) and methionine (Met 374),
with reference to the bound conformation of AD. The program was
able to reproduce the docked conformation of compound 6c in
aromatase with a reasonable rmsd. Graphics were generated with
PyMOL software (DeLano Scientific LLC, Palo Alto, CA, USA).
6c: oil; tR = 5.5 min; IR (neat): 1635 cm−1 (C C), 1665 and
1733 cm−1 (C O); UV ꢁmax: 250 nm (ε = 14,400); 1H NMR ı: 0.94
(3H, s, 18-Me), 1.21 (3H, s, 19-Me), 2.64 (2H, t, J = 7.9 Hz, phenyl-
CH2), 6.08 (1H, d, J = 1.5 Hz, 4-H), 6.75 (1H, s, 1-H), 7.16–7.29 (5H,
m, aromatic protons); 13C NMR ı: 13.8, 19.0, 21.9, 22.2, 29.0, 30.1,
31.2, 32.1, 32.4, 35.1, 35.7, 43.1, 47.7, 50.5, 52.6, 124.1, 125.7, 128.3,
128.4, 128.6, 137.4, 142.3, 150.3, 167.5, 186.2, 220.1; MS m/z: 402
(M+, 94), 298 (100), 283 (15), 134 (96). HR-MS for C28H34O2 (M+):
calcd 402.2559, found 402.2567.
6d: mp 140–142 ◦C; IR (KBr): 1626 cm−1 (C C), 1661 and
1742 cm−1 (C O); UV ꢁmax: 249 nm (ε = 14,900); 1H NMR ı: 0.94
(3H, s, 18-Me), 1.21 (3H, s, 19-Me), 2.62 (2H, t, J = 8.0 Hz, phenyl-
CH2), 6.07 (1H, d, J = 1.2 Hz, 4-H), 6.73 (1H, s, 1-H), 7.15–7.29 (5H,
m, aromatic protons); 13C NMR ı: 13.8, 18.9, 21.9, 22.2, 27.9, 29.0,
31.1, 31.2, 32.1, 32.4, 35.1, 35.6, 35.7, 43.1, 47.7, 50.4, 52.6, 124.1,
125.6, 128.2, 128.4, 137.5, 142.6, 150.2, 167.4, 186.2, 220.0; MS m/z:
416 (M+, 100), 299 (79), 297 (31), 91 (71). Anal. Calcd for C29H36O2:
C, 83.61; H, 8.71. Found C, 83.75 H, 8.80.
3. Results
3.1. Chemistry
2-Phenyaliphatic-substituted ꢀ1-ADs (6) were synthesized
principally according to reaction sequence of the prepara-
tion of 2-alkyl-substituted ꢀ1-AD derivatives. Various pheny-
laliphatic iodide (alkyl: methyl to butyl) and p-methyl- and
p-trifluoromethyl-phenethyl iodide subjected to reaction with
testosterone tert-butyldimethylsilyl ether (2) in presence of tert-
BuOK or lithium diisopropylamide (LDA) as a base under cooling
as a key reaction (Fig. 3). Conditions using tert-BuOK at −60 ◦C
was effective only for the formation of 2␣-phenpropyl compound
6e: mp 181–182 ◦C; IR (KBr): 1631 cm−1 (C C), 1664 and
1738 cm−1 (C O); UV ꢁmax: 250 nm (ε = 13,800); 1H NMR ı: 0.90
(3H, s, 18-Me), 1.10 (3H, s, 19-Me), 2.29 (3H, s, p-Me-phenethyl),
2.71 (2H, m, p-Me-phenyl-CH2), 6.08 (1H, d, J = 1.1 Hz, 4-H), 6.43
(1H, s, 1-H), 6.98–7.04 (4H, m, aromatic protons). MS m/z: 402 (M+,

334
M. Takahashi et al. / Steroids 75 (2010) 330–337
Fig. 4. Lineweaver–Burk plots of inhibition of human placental aromatase by 2-
phenpropyl-ꢀ1-AD (6c) with AD as the substrate. Concentrations of inhibitor: (ꢀ)
control (0 nM); (᭹) 12.5 nM; (ꢀ) 25 nM; (ꢁ) 50 nM. The inhibition experiments with
all the other steroids examined gave essentially similar plots to Fig. 4 (data not
shown).
steroids, the inhibitory activities of the phenethyl derivative 6b and
its p-methyl compound 6e and the phenpropyl derivative 6c were
relatively high (apparent Ki: 206 nM for 6b, 16.1 nM for 6c, and
348 nM for 6e) whereas the other compounds were poor inhibitors
(Ki: more than 2600 nM).
Compounds 6b, 6c, and 6e with the high affinity for aromatase
were then tested for their ability to cause a time-dependent inacti-
vation of aromatase. The phenethyl and phenpropyl compounds
6b and 6c showed the time-dependent inactivation only in the
presence of NADPH under aerobic conditions. Pseudo-first-order
kinetics were obtained during the first 12 min of inactivation with
the inhibitors, and the kinetic data were analyzed according to the
method of Kitz and Wilson [38] (Fig. 5). Double-reciprocal plots of
Kobs vs. inhibitor concentration gave kinact and KI values, respec-
tively (Table 2). The KI and kinact values for compound 6b and 6c
Fig. 3. Synthesis of 2-phenylaliphatic-substituted ꢀ1-ADs (6).
3c (22% yield). Conditions using LDA at −20 ◦C were used for the
productions of 2␣-benzyl- and 2␣-phenbutyl- compound 3a and
3d (64% for 3a and 22% for 3d), whereas 2␣-phenethyl- (3b), 2␣-
p-methylphenethyl- (3e) and 2␣-p-trifluoromethylphenethyl- (3f)
compounds were produced under conditions employing LDA at 0 ◦C
(yields: 11–22%). In the reactions, 2␣-phenylaliphatic 4-en-3-ones
3 were produced as major products while the 2␤-phenyaliphatic
isomers, minor products were not isolated. After the purifica-
tion with silica gel column chromatography or preparative TLC
of compounds 3, these compounds were subjected sequentially
to deprotection of the 17␤-silyl ethers 3 with diluted HCl in a
mixture of THF and 2-propanol and oxidation of the deprotected
product 17␤-ols 4 with Jones reagent in acetone, giving the 2␣-
phenylaliphatic 17-keto compounds 5. Finally, an introduction of a
double bond at C-1 of the 17-keto compounds 5, the AD derivatives
with DDQ in dioxane gave 2-phenylaliphatic ꢀ1-AD analogs 6.
The structures of the compounds synthesized were confirmed
by the spectrometric analysis and HR-MS, elemental analysis or
HPLC analysis.
Table 1
In vitro aromatase inhibition by 2␣-phenylaliphatic-ADs (5) and 2-phenylaliphatic-
ꢀ1-ADs (6).
Compound
IC50, ␮Ma
Apparent Ki, nMb
2␣-phenylaliphatic-ADs (5)
5a, benzyl
5b, phenethyl
5c, phenpropyl
5d, phenbutyl
466
126
50.2
222
38.8
46.6
–
–
5110 260
–
4095 191
5142 288
3.2. Aromatase inhibition
5e, p-methylphenethyl
5f, p-trifluoromethylphenethyl
Reversible inhibition of aromatase activity in human placen-
tal microsomes by 2-aliphatic-substituted ADs 5 and their ꢀ1-AD
derivatives 6 was initially studied in vitro by enzyme kinetics.
The amount of aromatase activity was measured by determinat-
ing the amount of 3H2O released from [1␤-3H] androstenedione,
which is an index of estrogen formation [36]. IC50 values were
first obtained, and then, to characterize the nature of their bind-
ing to the active site of aromatase, aromatization was measured
in the presence of several concentrations of the inhibitors and
androstenedione. The results of these studies were plotted on typ-
ical Lineweaver–Burk plots. All of the steroids examined exhibited
clear-cut competitive inhibition. The Lineweaver–Burk plots for the
phenpropyl-1,4-dien-3-one 6c are shown in Fig. 4. The apparent
inhibition constants (Ki) were determined by analysis of the Dixon
plot and shown in Table 1. In these studies, the apparent Km and
Vmax for androstenedione was about 40 nM and 120 pmol/min/mg
protein, respectively. All the AD derivatives 5 were poor inhibitors
of aromatase (IC50: more than 50 ␮M). In series of the ꢀ1-AD
2-phenylaliphatic-ꢀ1-ADs (6)
6a, benzyl
6b, phenethyl
6c, phenpropyl
6d, phenbutyl
86.4
2.57
0.145
34.2
6.36
34.12
4500 290
206 8.2
16.1 0.58
2690 190
348 21.7
4002 354
6e, p-methylphenethyl
6f, p-trifluoromethylphenethyl
For comparison
ꢀ1-ADc
0.830
90 6.4
a
300 nM of [1␤-3H]androstenedione and 20 ␮g of protein from human placental
microsomes were used.
b
Apparent inhibition constant (Ki) was obtained by Dixon plot. 20 ␮g of protein
from human placental microsomes and 5 min of incubation were employed. All of
the inhibitors examined showed a competitive type of inhibition based on analysis
of the Lineweaver–Burk plot. The results were means S.E. Apparent Km and Vmax
values for AD were about 40 nM and 120 pmol/min/mg protein, respectively, in this
study.
c
The Ki values of ꢀ1-AD were previously obtained to be 43 3.0 nM [27] and
65 3.5 nM [28].

M. Takahashi et al. / Steroids 75 (2010) 330–337
335
Fig. 6. Inactivation of human placental aromatase by 2-phenpropyl-ꢀ1-AD (6c)
under various conditions: (ꢀ) control; (᭹) inhibitor (600 nM); (ꢀ) control + AD
(5.4 ␮M) or inhibitor (600 nM) + AD (5.4 ␮M); (ꢂ) control + l-cystein (0.5 mM);
(ꢁ) inhibitor (600 nM) + l-cystein (0.5 mM); (×) inhibitor (600 nM) in the absence
NADPH.
hydrophobic binding pocket, extends to towards the C-2 posi-
tion of compound 6c and a bulky phenyl group of the phenpropyl
group orients to the accessible volume of the binging pocket. The
hydrophilic amino acids such as Ile 132, Ile 133, Ile 305, Phe 148,
Met 303, and Ala 306 are involved in the binding pocket and may
play a critical role in the binding to the active site (Fig. 7).
Fig. 5. Time-dependence (a) and concentration-dependence (b) of inactivation of
human placental aromatase by 2-phenpropyl-ꢀ1-AD (6c) in the presence of NADPH
in air. Concentrations of the inhibitor: (ꢀ) control (0 nM); (᭹) 300 nM; (ꢀ) 600 nM;
(ꢁ) 900 nM; (ϑ) 1200 nM. The time-dependent inactivation experiments with 2-
phenethyl-ꢀ1-AD (6b) gave essentially similar plots to Fig. 5 (data not shown).
4. Discussion
The 2-phenylaliphatic ꢀ1-ADs (6a–6d), which have various
lengths of methylene units from a methyl to a butyl as well
as the p-methylphenyl- and p-trifluoromethylphenylethyl deriva-
tives 6e and 6f were synthesized. The effect of introducing the
phenylalkyl chain at the 2-position of ꢀ1-AD were examined with
respect to affinity for aromatase and rate of inactivation of the
enzyme. The 2-phenpropyl derivative 6c was very powerful com-
petitive inhibitor of aromatase in human placental microsomes
with apparent Ki value of 16.1 0.58 nM where the Km/Ki ratio of
compound 6c is about 2.5. This corresponds well to those obtained
for 6␤-alkyl- and 6␤-phenylaliphatic-ꢀ1-ADs which are among
the most powerful inhibitors of aromatase [27,28]. Compound 6c
inactivated aromatase in a mechanism-based manner. The rate
of inactivation decreased when the substrate AD was included
in the incubation mixture while l-cysteine, a nucleophile, failed
to protect aromatase significantly from the inactivation by the
inactivator. Thus, covalent-bound formation between aromatase
and the reactive intermediate appears to occur at the active site,
therefore, preventing diffusion of the active inhibitor, a reactive
electrophile, in the surrounding media [39]. Recently, we suggested
that the mechanism-based inactivator ꢀ1-AD, a parent compound
of the phenpropyl steroid 6c, may inactivate aromatase through
oxygenation of 19-oxo function derived from two sequential oxy-
genations at the C-19 position of the parent compound [30,40]. The
aromatase-bound metabolite may be produce during the cleav-
age of the 19-oxo group, although the exact mechanism is clear
so far. Covey [5] has suggested previously that aromatase inac-
tivates itself because the mechanism-based aromatase inhibitor
induces the enzyme to autoxidize itself. Since we have no evi-
dence for covalent modification of aromatase by the compound 6c,
a substrate analog-induced autoxidation mechanism for inactiva-
Table 2
Kinetic analysis of time-dependent inactivation of aromatase caused by 2-
phenylaliphatic-ꢀ1-ADsa (6).
Compound
KI, nM
kinact, min−1
6b, phenethyl
6c, phenpropyl
6e, p-methylphenethyl
8124
370
0.0293
0.0454
−
NTb
For comparison
ꢀ1-AD [27]
952
0.059
a
The time-dependent inactivation of compound 6a, 6d and 6f were not deter-
mined because of their poor affinity for aromatase. Apparent KI and kinact were
obtained by Kitz–Wilson plot [38].
b
The time-dependent inactivation was not observed.
were 8124 nM and 0.0293 min−1 for compound 6b and 370 nM and
0.0454 min−1 for compound 6c. The p-methylphenethyl compound
6e did not cause the time-dependent inactivation.
The substrate AD significantly blocked the inactivation by the
1,4-diene inhibitors, while a nucleophile, l-cysteine, had no signif-
icant effect (Fig. 6).
3.3. Molecular docking
Phenpropyl-ꢀ1-AD steroid (6c) was selected for docking to the
active site of human placental aromatase. As shown in Fig. 7, the
compound 6c positions in the active site of the enzyme with hydro-
gen bondings of the C-3 and C-17 carbonyl groups with polar amino
acid residues Asp 309 and Arg 115 or Met 374, respectively, simi-
larly to the natural substrate AD. Interestingly, accessible volume,

336
M. Takahashi et al. / Steroids 75 (2010) 330–337
substituted ADs have high affinity to the binding site of aromatase;
the 2␣-methyl- and 2␣-ethyl compounds were powerful inhibitors
of the enzyme and elongation of the 2␣-subsituent up to the hexyl
derivative still showed high affinity to the enzyme [31]. The results
implied that the binding geometry of the 2-phenylaliphatic steroids
5 in the active site of aromatase would be different from those of
the alkyl steroids. The phenyl moiety would prevent the formation
of thermodynamically stable inhibitor–enzyme complex.
The crystal structure of human placental aromatase has recently
been reported by Ghosh et al. [41]. A molecule of exemestane, one
of ꢀ1-AD analogs, is build into the active site by using the AD back-
bone. AD and exemestane superimposed quite well accepts the
difference in puckering of the A-ring. Thus exemestane remains
tightly bound in the pocket through hydrogen bondings of the
C-3 and C-17 carbonyl groups to Asp 309 and Arg 115 or Met
374, respectively, and van der Waals interaction between C-6
methylidene and The 310 is observed. Docking experiment of
phenpropyl-ꢀ1-AD (6c), the most powerful aromatase inhibitor
and inactivator, to the crystal structure of the enzyme showed
that this compound was bound to the active site in the anchoring
effect of hydrogen bondings of the C-3 and C-17 carbonyl groups, as
observed in a exemestane molecule, and the lipophilic amino acid
residues, Ile 132, Ile 133, Ile 305, Phe 148, Met 303, and Ala 306, in
the vicinity of the 2-phenpropyl moiety would also play a critical
role in binding of the active site.
5. Conclusion
Fig. 7. 2-Phenpropyl-ꢀ1-AD (6c) is docked into the active site of the human pla-
cental aromatase. Views from the heme iron (a) and from the C-2 phenpropyl group
(b) are shown in selected active site residues of amino acids which are displayed
for gray or white for carbon, blue for nitrogen, red for oxygen and yellow for sulfur.
Compound 6c is colored magenta. The C-3 and C-17 carbonyl groups of compound
6c positions to hydrogen bonding with Asp 309 and Arg 115 or Met 137, respectively,
and the phenpropyl group is oriented in the lipophilic binding pocket of the active
site to produce thermodynamically stable enzyme–steroid complex. (For interpre-
tation of the references to color in this figure legend, the reader is referred to the
web version of the article.)
2-Phenylaliphtic ꢀ1-AD (6) were synthesized from testosterone
17-tert-butyldimethylsilyl ether (2) using phenylaliphatic halides
in the presence of LDA or KOC(CH3)3 as a key reaction as the aro-
matase inhibitors. The phenpropyl compound 6c was the most
powerful inhibitor and this inhibitor inactivated aromatase in a
mechanism-based manner. Docking experiments demonstrate that
compound 6c is oriented to the active site of aromatase, binding
pocket, though the hydrogen bondings of the 3- and 17-carbonyl
groups and the lipophilic interactions of the phenpropyl group to
produce thermodynamically stable enzyme–steroid complex.
tion may be considered as an alternate explanation for the observed
time-dependent inactivation.
Acknowledgments
The aromatase inhibitory activities of the phenethyl and p-
methylphenethyl compounds 6b and 6e were less powerful than
the phenpropyl derivative 6c but still good inhibitors for the
enzyme. The other compounds, the benzyl, phenbutyl, and p-
trifluoromethylphenethyl-1,4-dienes 6a, 6d, and 6f, showed poor
activities of the aromatase inhibition. Only the phenethyl com-
pound 6b among them caused the time-dependent inactivation of
aromatase. Compound 6e has a p-methylphenethyl moiety at the
2-position of ꢀ1-AD (1), which is similar length of the C-2 sub-
stituent to a phenpropyl moiety. This suggests that the terminal
phenyl group would play a critical role in the suicide inactivation of
aromatase and a steric reason would be operative in the aromatase-
catalyzed activation reaction of the phenethyl derivative 6e. All of
the time-dependent activators 6b and 6c showed that the KI val-
ues obtained from the inactivation experiments are more than 20
times higher than the corresponding apparent Ki values from the
competitive inhibition experiments. The similar tendency has pre-
viously been reported in the inactivation experiments using other
ꢀ1-AD steroid derivatives [27–31]. When the observed inactiva-
tion is based on the aromatase reaction per se, this relation of the KI
value to the Ki value suggests that binding of the activated inhibitor
to the nucleophilic residue of the active site rather than activation of
the inhibitor becomes rate determining or partial rate determining
[39].
This work was supported in part by a High Technology Research
Center Project from the Ministry of Education, Culture, Sports, and
Technology of Japan. Human term placenta was kindly donated by
Dr. Michihiro Yuki of Yuki Lady’s Clinic.
References
[1] Thompson Jr EA, Siiteri PK. The involvement of human placental microsomal
cytochrome P-450 in aromatization. J Biol Chem 1974;249:5373–8.
[2] Kellis Jr J, Vickery LE. Purification and characterization of human placental
aromatase cytochrome P-450. J Biol Chem 1987;262:4413–20.
[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] Thompson Jr EA, Siiteri PK. Utilization of oxygen and reduced nicotinamide
adenine dinucleotide phosphate by human placental microsomes during arom-
atization of androstenedione. J Biol Chem 1974;249:5364–72.
[5] Covey DF. Aromatase inhibitors: specific inhibitors of estrogen biosynthesis.
In: Berg D, Plempel M, editors. Sterol biosynthesis inhibitors: pharmaceutical
and agrochemical aspects. Chichester: Eliss Horwood Ltd.; 1988. p. 534–71.
[6] Oh SS, Robinson CH. Mechanism of human placental aromatase: a new active
site model. J Steroid Biochem Mol Biol 1993;44:389–97.
[7] Akhtar M, Lee-Robichaud P, Akhtar ME, Wright JN. The impact of aromatase
mechanism on other P450s. J Steroid Biochem Mol Biol 1997;61:127–32.
[8] Bossche HV. Inhibitors of P450-dependent steroid biosynthesis: from research
to medical treatment. J Steroid Biochem Mol Biol 1992;43:1003–21.
[9] Brodie AM, Santen RJ. Aromatase and its inhibitors in breast cancer
treatment—overview and perspective. Breast Cancer Res Treat 1994;30:1–6.
[10] Johnston JO. Aromatase inhibitors. Crit Rev Biochem Mol Biol
1998;33:375–405.
All of the AD derivatives 5 were poor inhibitors of aromatase (Ki:
more than 4 ␮M). The previous findings demonstrated that 2-alkyl-

M. Takahashi et al. / Steroids 75 (2010) 330–337
337
[11] Harper-Wynne C, Dowsett M. Recent advances in the clinical application of
aromatase inhibitors. J Steroid Biochem Mol Biol 2001;76:179–86.
[12] Ryan KJ. Biological aromatization of steroids. J Biol Chem 1959;234:268–72.
[13] Gual C, Morato T, Hayano M, Gut M, Dorfman RI. Biosynthesis of estrogens.
Endocrinology 1962;71:920–5.
[14] Meigs RA, Ryan KJ. Enzymatic aromatization of steroids. I. Effects of oxygen
and carbon monoxide on the intermediate steps of estrogen biosynthesis. J
Biol Chem 1971;246:83–7.
[27] Numazawa M, Oshibe M, Yamaguchi S, Tachibana M. Time-dependent inactiva-
tion of aromatase by 6-alkylandrosta-1,4-diene-3,17-diones. Effects of length
and configuration of 6-alkyl group. J Med Chem 1996;39:1033–8.
[28] Numazawa M, Yamaguchi S. Synthesis and structure–activity relationships
of 6-phenylaliphatic-substituted C19 steroids having a 1,4-diene, 4,6-diene,
or 1,4,6-triene structure as aromatase inhibitors. Steroids 1999;64:187–
96.
[29] Numazawa M, Shelangouski M, Nagasaka M. Probing the binding pocket of
the active site of aromatase with 6-ether or 6-ester substituted androst-4-
ene-3,17-diones and their diene and triene analogs. Steroids 2000;65:871–
82.
[30] Numazawa M, Nagaoka M, Handa W, Ogawa Y, Matsuoka S. Studies directed
towards a mechanistic evaluation of inactivation of aromatase by the suicide
substrates androta-1,4-diene-3,17-diones and its 6-ene derivatives. Aromatase
inactivation by the19-substitutedderivatives and their enzymic aromatization.
J Steroid Biochem Mol Biol 2007;107:211–9.
[15] Milewich L, Bradfield DJ, Coe LD, Masters BSS, MacDonald P. Metabolism
of 1,4-androstadiene-3,17-dione by human placental microsomes. Enzyme
properties and kinetic parameters in the formation of estrogens and 17␤-
hydroxy-1,4-androstadien-3-one. J Steroid Biochem 1981;14:1115–25.
[16] Numazawa M, Yoshimura A, Oshibe M. Enzymic aromatization of 6-alkyl-
substituted androgens, potent competitive and mechanism-based inhibitors
of aromatase. Biochem J 1998;329:151–6.
[17] Numazawa M, Yoshimura A. Biological aromatization of ꢀ4,6- and ꢀ1,4,6
-
androgens and their 6-alkyl analogs, potent inhibitors of aromatase. J Steroid
Biochem Mol Biol 1999;70:189–96.
[18] Schwarzel WC, Kruggel WG, Brodie HJ. Studies on the mechanism of estro-
gen biosynthesis. VIII. The development of inhibitors of the enzyme system in
human placenta. Endocrinology 1973;92:866–80.
[19] Covey DF, Hood WF. Enzyme-generated intermediates derived from 4-
androstene-3,6,17-trione and 1,4,6-androstatriene-3,17-dione cause a time-
dependent decrease in human placental aromatase activity. Endocrinology
1981;108:1597–9.
[31] Takahashi M, Handa W, Umeta H, Ishikawa S, Yamashita K, Numazawa M.
Aromatase inactivation by 2-substituted derivatives of the suicide substrate
androsta-1,4-diene-3,17-dione. J Steroid Biochem Mol Biol 2009;116:191–9.
[32] Hosoda H, Yamashita K, Sagae H, Nambara T, Steroids CI. Dimethyl-tert-
butylsilyl ethers of steroids. Chem Pharm Bull 1975;23:2118–22.
[33] Abraham EP, Smith JC. Addition of hydrogen chloride and iodide to olefins.
Undecenoic acid. J Chem Soc 1936:1605–7.
[34] Roel B, Joannes W, Aede G. A novel route to the marasmane skeleton via
a tandem rearrangement-cyclopropanation reaction. Total synthesis of (+)-
isovelleral. J Org Chem 2001;66:2350–7.
[20] Covey DF, Hood WF.
A new hypothesis based on suicide substrate
inhibitor studies for the mechanism of action of aromatase. Cancer Res
1982;42(Suppl.):3327s–33s.
[35] Nedelec L, Gasc JC, Bucourt R. The kinetically controlled methylation of conju-
gated polycyclic ketones. Tetrahedron 1974;30:3263–8.
[21] Marsh DA, Brodie HJ, Garrett W, Tsai-Morris CH, Brodie AMH. Aromatase
inhibitors. Synthesis and biological activity of androstenedione derivatives. J
Med Chem 1985;28:788–95.
[36] Siiteri PK, Thompson EA. Human placental aromatase.
1975;6:317–22.
[37] Trott A, Olson J. Autodock Vina: improving the speed and accuracy of dock-
ing with a new scoring function, efficient optimization and multithreading. J
Comput Chem 2009, available online.
[38] Kitz R, Wilson IB. Effects of methanesulfonic acid as irreversible inhibitors of
acetylcholine esterase. J Biol Chem 1962;237:3245–9.
[39] Silverman RB. Mechanism-based enzyme inactivation, vol. 1. Boca Raton, FL:
CRC press; 1988. Chapter 1, p. 3–30.
[40] Numazawa M, Yamashita K, Kimura N, Takahashi M. Chemical aromatization
of 19-hydroxyandrosta-1,4-diene-3,17-dione with acid or alkaline: elimina-
tion of the 19-hydroxy methyl group as formaldehyde. Steroids 2009;74:208–
11.
[41] Ghosh D, Griswold J, Erman M, Pangborn W. Structural basis for androgen speci-
ficity and oestrogen synthesis in human aromatase. Nature 2009;457:219–
23.
J Steroid Biochem
[22] Segaloff A, Weeth JB, Rongone EL, Murison PJ, Bowers CY. Hormonal therapy in
cancer of the breast. XVI. The effect of ꢀ1-testololactone on clinical course and
hormonal excretion. Cancer 1960;13:1017–20.
[23] Lerner LJ, Bianchi A, Borman A. ꢀ1-Testololactone, a nonadrogenic augmentor
and inhibitor of androgens. Cancer 1960;13:1201–5.
[24] Henderson D, Norbisrath G, Kerb U. 1-Methyl-1,4-androstadiene-3,17-dione
(SH 489): characterization of an irreversible inhibitor of estrogen biosynthesis.
J Steroid Biochem 1986;24:303–6.
[25] Salle ED, Briatico G, Giudici D, Ornati G, Zaccheo T. Aromatase inhibition and
experimental antitumor activity of FCE 24304, MDL 18962 and SH 489. J Steroid
Biochem 1989;34:431–4.
[26] Salle ED, Ornati G, Giudici D, Lassus M, Evans RJ, Coombes RC. Exemestane
(FCE 24304) a new steroidal aromatase inhibitor. J Steroid Biochem Mol Biol
1992;43:137–43.