Synthesis of 17beta-N-substituted 19-Nor-10-azasteroids as inhibitors of human 5alpha-reductases I and II.

Article abstract of DOI:10.1016/S0968-0896(02)00254-7

The synthesis of 17beta-[N-(phenyl)methyl/phenyl-amido] substituted 10-azasteroids has been accomplished by either the TiCl4- or TMSOTf-catalysed reaction of carbamates 11 and 12 with Danishefsky's diene. The reaction provided 5alpha-H isomers 3a-5a and 5beta-H isomers 3b-5b depending on the reaction conditions. Both epimers of each compound were tested against human 5alpha-reductase types I and II. Unexpectedly, 5beta-H compounds were found more active than their 5alpha-H counterparts, the best inhibitors being 3b (IC50=279 and 2000 nM toward isoenzyme I and II, respectively) and 5b (IC50=913 and 247 nM toward isoenzymes I and II, respectively).

Full text of DOI:10.1016/S0968-0896(02)00254-7

Bioorganic & Medicinal Chemistry 10 (2002) 3455–3461
Synthesis of 17ꢀ-N-Substituted 19-Nor-10-azasteroids as
Inhibitors of Human 5ꢁ-Reductases I and II
Dina Scarpi,a Ernesto G. Occhiato,a Giovanna Danza,b
Mario Seriob and Antonio Guarnaa,*
a
`
Dipartimento di Chimica Organica ‘Ugo Schiff’, Universita di Firenze, Via della Lastruccia 13,
I-50019 Sesto Fiorentino, Florence, Italy
b
`
`
Dipartimento di Fisiopatologia Clinica, Unita di Endocrinologia, Universita di Firenze,
Viale G. Pieraccini 6, I-50134 Florence, Italy
Received 19 April 2002; accepted 21 June 2002
Abstract—The synthesis of 17b-[N-(phenyl)methyl/phenyl-amido] substituted 10-azasteroids has been accomplished by either the
TiCl4- or TMSOTf-catalysed reaction of carbamates 11 and 12 with Danishefsky’s diene. The reaction provided 5a-H isomers
3a–5a and 5b-H isomers 3b–5b depending on the reaction conditions. Both epimers of each compound were tested against human
5a-reductase types I and II. Unexpectedly, 5b-H compounds were found more active than their 5a-H counterparts, the best inhi-
bitors being 3b (IC50=279 and 2000 nM toward isoenzyme I and II, respectively) and 5b (IC50=913 and 247 nM toward iso-
enzymes I and II, respectively).
# 2002 Elsevier Science Ltd. All rights reserved.
Introduction
10-Azasteroids are potent in vitro inhibitors of human
5a-reductases I and II, with activity depending on the
degree of unsaturation of the A and B rings and sub-
stitution on the D ring. The highest activity is observed
for the 17b-(N-tert-butyl)carbamoyl derivative 1 (Fig.
1), which is a dual inhibitor toward type I (IC50 127
nM) and type II (IC50 37 nM) isoenzymes.
The enzyme 5a-reductase plays an important role in
androgen metabolism by converting testosterone (T)
into dihydrotestosterone (DHT). The enzyme exists in
two isoforms, named type I and II, that catalyze the
same reaction but in different tissues: type I is mainly
expressed in skin, and type II in human prostate. There
are many pathologies related to DHT formation,
among them the most important is certainly human
prostatic hyperplasia (BPH), whose incidence in the
aging male population is very high.1 Acne,2 alopecia3 in
man and hirsutism4 in woman are also disorders related
to DHT level. It is therefore important the development
of inhibitors of 5a-reductase that, lowering the DHT
level, can be used as drugs for the pharmacological
treatment of the above androgen-dependent pathologies.
Among all the inhibitors discovered in recent years,5
three classes were based on the testosterone skeleton,
modified by the introduction of a nitrogen atom in the
A ring (4-azasteroids),6,7 in the B ring (6-azasteroids)8
We first synthesized 10-azasteroids by a methodology
based on the thermal rearrangement of isoxazoline-
5-spirocyclopropanes.9a More recently we developed a
new synthetic route in which the key step, that is the
formation of the A ring, is a Lewis acid catalyzed hetero
Diels–Alder reaction of silyloxy dienes with N-(acyl-
oxy)-iminium ions generated in situ from the corre-
sponding 4-N-(tert-butoxycarbonyl) 3-ethoxy derivatives
(Scheme 1).10,11 The A ring can be further modified by
the introduction of a Á1(2) and/or Á4(5) double bond.
Moreover, we developed a synthetic methodology that
afforded either 5a-H or 5b-H epimer simply by chang-
ing the order of addition of the reactants in the A-ring
formation step.12
9
and at the bridgehead position 10(10-azasteroids).
Since the introduction of polar groups on C-17 is
proved to modulate and modify the biological activity
of these testosterone-derived inhibitors, we decided to
*Corresponding author. Tel.: +39-055-457-3481; fax: +39-055-457-
3531; e-mail: antonio.guarna@unifi.it
0968-0896/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(02)00254-7

3456
D. Scarpi et al. / Bioorg. Med. Chem. 10 (2002) 3455–3461
Imine 7 (Scheme 2) was quantitatively obtained from
610 by treatment with aniline and catalytic p-TsOH in
refluxing toluene over molecular sieves. Crude 7 was
reduced to 8 by an excess of NaBH3CN in MeOH at pH
6, followed by addition of concd HCl up to pH 2–3
(overall yield 96% after chromatographic purification).
Alternatively, the quantitative reduction of the imine
group in 7 is accomplished by 1.5 equiv NaBH4 in
MeOH followed, after isolation of the intermediate, by
C–C double bond reduction by NaBH3CN in MeOH at
pH 2 (96%). Even though high yields are obtained with
both methods, the former reaction is preferable since it
affords the final product 8 in one step. Analogously to
earlier observations,10 both reductions exclusively
occurred on the less hindered a face, leading to 8 as a
single diastereoisomer with the same stereochemistry of
the natural androgens. Lactam 8 was then protected as
N-Boc derivative 9, using an excess of Boc2O in reflux-
ing DCM and catalytic amount of DMAP. It must be
highlighted that 9 was obtained as a single product in
98% yield; no reaction at the N-phenyl amine group
occurred and 9 was used as common precursor for the
subsequent steps. Benzoylation of 9, by benzoyl chloride
in anhydrous THF at rt and using anhydrous K2CO3 as
a base, afforded pure 10 in 70% yield. Noteworthy,
benzoylation of 8 under the same reaction conditions
occurred on the N-phenyl amine group only. However,
in the latter reaction perfectly anhydrous conditions are
required, because even small traces of water cause ben-
zoylation of the amide group, probably due to the for-
mation of HCl by partial hydrolysis of benzoyl chloride
and subsequent protonation of the N-phenyl amine
group.
Figure 1.
evaluate whether, analogously to the 17b-carbamoyl
groups, the introduction of a 17b-[N-(aryl)alkyl/aryl-
amido] group determines important variations in the
activity of 10-azasteroids. A similar work has been
already carried out on 4-azasteroids, in which this kind
of substitution resulted in potent and selective type I or
dual inhibitors.13 In the present paper, we thus report
on the synthesis of six different 17b-[N-(phenyl)methyl/
phenyl-amido] substituted 10-azasteroids, as three pairs
of 5a-H/5b-H epimers (Fig. 1) and their biological
evaluation.
Formation of the N-Boc ethoxy derivatives 11 and 12
(starting from 10 and 9, respectively) was obtained by
reduction of the carbonyl group with 1.5 equiv LiBHEt3
in THF at ꢀ78 ꢁC, followed by treatment with 2 N HCl
in abs EtOH. These compounds represent the key
intermediates of the whole synthesis and were both
obtained in 57% yield as single diastereoisomers. As
already reported,10 this procedure provides the final
compounds in higher yields with respect to the NaBH4
reduction, and no traces of over-reduced products were
Furthermore, since only the 5a-H epimers, on the basis
of the reduction mechanism of 5a-reductase,9 could be
seen as product like inhibitors, we were interested also
in seeing how the 5-a/5-b stereochemistry at the A/B
ring junction affects the inhibitory potency.
Chemistry
1
In the synthesis reported by Li et al. the introduction of
the 17b-N-amido substituent was carried out on the
final 4-aza-5a-androstane-3,17-dione product by reduc-
tive amination, followed by further N-acylation.13 In
our synthesis (Schemes 2 and 3), the transformation
of the C-17 carbonyl group into the N-phenyl amine
group was performed in the initial steps, while the
further N-acylation was performed either before or
after the A-ring formation step (5a–5b and 4a–4b,
respectively).
found. In the H NMR spectra of 11 and 12, 3-H reso-
nates as a sharp doublet at 5.41 ppm (J=5.8 Hz) and
5.39 ppm (J=5.8 Hz), respectively. This suggests an
axial–equatorial relationship between 3-H and one of
the protons in C-2; the other vicinal J of 3-H is close to
zero and cannot be identified. These data are consistent
with the equatorial position of the ethoxy group.10
N-Boc ethoxy compound 12 was used in the next A-ring
formation step, with Danishefsky’s diene as a partner of
Scheme 1.

D. Scarpi et al. / Bioorg. Med. Chem. 10 (2002) 3455–3461
3457
Scheme 2. (a) PhNH2, toluene, reflux, 16 h; (b) NaBH3CN, MeOH, pH 6!2, 24 h; (c) Boc2O, Et3N, DMAP 0.1 equiv, DCM, reflux, 12 h;
(d) PhCOCl, K2CO3, THF, rt, 12 h; (e) (i) LiBHEt3, THF, ꢀ78 ꢁC, 15 min; (ii) 2 N HCl, EtOH, ꢀ78!0 ꢁC, 30min.
Scheme 3. (a) 1-Methoxy-3-[(trimethylsilyl)oxy]-1,3-butadiene, then TiCl4, 0!25 ꢁC, 1 h; then NaHCO3 satd, 25 ꢁC, 30min; (b) 1-methoxy-3-[(tri-
methylsilyl)oxy]-1,3-butadiene, Et3N, TMSOTf, DCM, 0!25 ꢁC, 45 min; then NaHCO3 satd, 25 ꢁC, 36 h; (c) Ac2O, pyridine, rt, 24 h.
the reaction (Scheme 3). As already reported,12 we ver-
ified that the stereochemical outcome at C-5 can be
chosen by using different reaction conditions. Addition
of TMSOTf to a mixture of diene, Et3N and 12 in
anhydrous DCM, followed by a 36 h NaHCO3 treat-
ment, results in 3b, with 5b-H stereochemistry, in 20%
yield after chromatographic purification. On the con-
trary, addition of Danishefsky’s diene to a mixture of 12
and TiCl4 affords 3a, in 13% yield but with opposite
5a-H stereochemistry. These yields are consistent with
those already reported for analogous reactions.12 Sur-
prisingly, applying to 11 the same reaction conditions
employed for the synthesis of 3b leads to 5 as a 1.5:1
5a-H (5a)/5b-H (5b) diastereoisomeric mixture. After
chromatography, the two diastereoisomers 5a and 5b
were obtained in 14 and 13% yield, respectively.
Further functionalisation at position 17 of 3a and 3b
was quantitatively achieved by acetylation with Ac2O in
pyridine and catalytic DMAP, affording pure 4a and 4b,
respectively. It must be noticed that acetylation of 9,
using the same reaction conditions, resulted in a mixture
of products composed by deprotected amide 13, mono-
acetylated compound 14 (analogue to 10) and diacetyl-
ated compound 15 (Scheme 4). On the other hand,
acetylation conditions analogue to those used for the

3458
D. Scarpi et al. / Bioorg. Med. Chem. 10 (2002) 3455–3461
Scheme 4. (a) Ac2O, pyridine, rt, 24 h; (b) AcCl, K2CO3, THF, rt, 12 h.
synthesis of 10, that is acetyl chloride in THF and
anhydrous K2CO3 as a base, afforded the expected 14.
However, whereas 10 could be obtained in high yield, in
the case of 14 conversions were never quantitative
(independently on the acetyl chloride equivalents
employed) and the yields were lower than 50%. There-
fore, we preferred using the methodology reported in
Schemes 2 and 3. The stereochemistry at C-5 can be
In general, compounds 3–5 were less active than
17b-carbamoyl substituted 10-azasteroid 1. This is not
unexpected since we have already shown as Á4(5) unsat-
urated compounds, like 1, are by far more active than
Á
1(2) 19-nor-10-azasteroids, such as 3–5. In the series of
Á1(2) 5a-H derivatives 3a, 4a and 5a the introduction of
the 17b-N-acyl group decreases the activity toward
5aR-I: in fact, N-phenyl amine 3a (IC50=0.580 mM) is
nearly ten times more potent than 5a (IC50=4.7 mM),
that bears a benzoyl group, and more potent than 4a
(bearing an acetyl), which shows a 24% inhibition when
tested at 10 mM concentration. Different results can be
observed considering the activity of the cited com-
pounds against 5aR-II. In this case, only the introduc-
tion of the N-benzoyl moiety determines the maintenance
of the activity toward the enzyme: in fact, 5a (IC50=5.6
mM) proved to be a better inhibitor than both 3a and 4a,
that, at 10 mM concentration, inhibit the enzyme only by
31 and 9%, respectively. Furthermore, only compound 5a
showed a higher potency than the corresponding 5a-H
17-oxo 10-azasteroid 2 (5.6 mM vs 46 mM).
1
determined by analysis of the H NMR spectra. In fact
5-H resonates at different chemical shifts depending on
its a or b orientation. Analogously to the 10-azasteroids
already synthesized,9,10a the downfield shifted signals
can be assigned at a 5b-H.10a According to this, in 3b
5b-H resonates at 3.78 ppm, while in 3a 5a-H resonates
at 2.65 ppm. The same assignment applies to 5, where
5a-H resonates at 2.69 ppm and 5b-H at 3.76 ppm.
Obviously, since compounds 4a and 4b directly derive
from 3a and 3b by acetylation, the C-5 stereochemistry
1
is identical to that of their precursors. In the H NMR
spectrum, 5-H resonates at 2.60and 3.76 ppm for the a
(4a) and b (4b) epimer, respectively.
As for the Á1(2) 5b-H series, all compounds 3b, 4b and
5b resulted more active than their corresponding 5a-H
epimers, even though the behavior within the single
series is similar. In fact, looking at the 5aR-I inhibition
data, the less potent inhibitor resulted the N-acetyl
derivative 4b (65% inhibition at 10 mM concentration),
whereas the 17b-N-phenyl amino substituted 3b
(IC50=0.279 mM) was still more active than 5b
(IC50=0.913 mM), even if the difference in potency is
not as remarkable as for 3a and 5a.
Inhibition Activity
Compounds 3a–b, 4a–b and 5a–b were tested against
human recombinant 5aR-I and II, according to the repor-
ted procedures,14 and their inhibition data are reported as
IC50 in Table 1, including 1 and 2 (Fig. 1) for comparison.
For a few compounds, the value of IC50 could not be
determined. In these cases, we reported the data related to
the percentage of inhibition obtained when these com-
pounds were tested at the defined concentration of 10 mM.
In the case of type II enzyme, the data are consistent
with the behavior of the homologous a series. Com-
pound 5b (IC50=0.247 mM) is 10times more potent
than 3b (IC50=2.0 mM), and more potent than 4b, that
inhibits the enzyme by 45% at 10 mM concentration.
Table 1. In vitro activities of 19-nor-10-azasteroids 17b-[N-(phenyl)]
and 17b-[N-(phenyl)methyl/phenyl-amido] substituted
Inhibitor
IC50 (mM)
Human type I 5aR
Human type II 5aR
While azasteroid 1 is a dual inhibitor (IC50=0.127 and
0.037 mM against 5aR-I and 5aR-II, respectively),9a all
compounds tested in this work were more active against
5aR-I. The only exception is represented by 5b, that is
roughly 4 times more potent against isoenzyme II. Our
results are consistent with the data reported by Li et al.
for an homologous series of 4-azasteroids.13 In that
paper, almost all 17b-N-amido compounds were more
active against 5aR-I.
CHO 1827
CHO 1829
1
2
0.127ꢂ0.012a
0.037ꢂ0.007b
46.0ꢂ18.4b
31%d
n.d.c
3a
3b
4a
4b
5a
5b
0.580ꢂ0.115
0.279ꢂ0.083
24%d
2.000ꢂ0.646
9%d
65%d
45%d
4.7ꢂ1.1
5.6ꢂ1.2
0.913ꢂ0.284
0.247ꢂ0.078
aDetermined in DU-145 cells.8a
However, an unexpected result was obtained in this work:
for the first time it was found that, in the same series, the
5b-H isomers are more potent than the corresponding
5a-H compounds.
bDetermined in human prostate homogenate.8a
cNot determined.
dPercentage values refer to inhibition obtained by the specified com-
pound at 10 mM concentration.

D. Scarpi et al. / Bioorg. Med. Chem. 10 (2002) 3455–3461
3459
These data seem to be in contrast with the transition
state model previously reported,9a,15 which includes a
partial bond formation on the a face of the steroid by
the attack of NADPH on C-5. According to this model,
compounds 3a, 4a and 5a are product-like transition
state analogues and therefore were expected to be the
most potent inhibitors.
(ꢀ)-1,2,4,4aꢁ,5,6a,7,8,9aꢁ,9bꢀ-Dodecahydro-6aꢀ-methyl-
7ꢀ-(N-phenylamino)-(3H)-cyclopenta[ f ]quinoline-3-one
(8). Enelactame 6 (3.00 g, 13.7 mmol), aniline (12.5
mL, 137 mmol) and p-TsOH (24 mg, 0.14 mmol) were
dissolved in toluene (90mL) in a Dean–Stark appara-
tus, and the resulting solution refluxed for 16 h. After
removal of the solvent, crude imine 7 was obtained and
1
used without purification in the next step. 7: H NMR
Therefore, it is surprising that the 5b-H compounds
were up to 20times more active than the corresponding
azasteroids having 5a stereochemistry. Currently, we are
unable to explain this result. However, we have already
shown9b that the presence of the bridgehead nitrogen
atom at 10position confers a high degree of flexibility to
the A and B ring moiety of the azasteroids, thus mak-
ing the cis fusion in the 5b epimers less drastically
different from the normal trans fusion of a steroid-
like structure. To the best of our knowledge, com-
pounds 3b and 5b represent the first examples of
potent aza-steroidal inhibitors of the enzyme
5a-reductase having 5b-H stereochemistry.
(CDCl3, d ppm): 7.78 (s, 1H), 7.17–7.09 (m, 2H), 6.68–
6.59 (m, 3H), 4.81 (m, 1H), 3.55 (m, 2H), 2.53–1.87 (m,
7H), 1.66–1.38 (m, 4H), 0.81 (s, 3H).
Crude imine 7 was suspended in MeOH (160mL) and
the pH adjusted to 6 by adding concd HCl. After cool-
ing to 0 ꢁC, NaBH3CN (2.66 g, 41.1 mmol) was then
slowly added and the resulting solution stirred for 30
min. Concd HCl was then added (pH 2) and further
NaBH3CN (0.98 g, 15.1 mmol) was slowly added and
the solution left under stirring at rt. After 12 h, the sus-
pension was poured into water (160mL), NaOH (satd)
was added up to pH 10and the product extracted with
DCM (3ꢃ200 mL). The organic layer was first washed
with brine and then dried over Na2SO4. After filtration
and evaporation of the solvent, crude 8 was obtained as
a red oil. Purification by flash chromatography, eluting
first with CH2Cl2 and then with CH2Cl2–MeOH, 10:1
(Rf=0.47), afforded pure 8 (3.93 g, 96%) as a white
solid. 8: mp 186–187 ꢁC; [a]D25 ꢀ9.99 (c 0.66, CHCl3); 1H
NMR (CDCl3, d ppm): 7.20–7.12 (m, 2H), 6.80–6.60
(m, 3H), 5.82 (s, 1H), 3.40(m, 1H), 2.97–2.82 (m, 1H),
2.60–2.35 (m, 2H), 2.35–1.07 (m, 13H), 0.88 (s, 3H); 13C
NMR (CDCl3, d ppm): 172.3 (s), 147.9 (s), 129.1 (d, 2
C), 117.1 (d), 113.3 (d, 2 C), 63.1 (d), 58.6 (d), 49.1 (d),
44.3 (d), 38.4 (t), 35.7 (d), 31.2 (t), 29.9 (t), 28.7 (t), 25.1
(t), 22.5 (t), 12.2 (q); MS m/z (%): 298 (M+, 23), 132
(100), 86 (60), 84 (91); IR (CDCl3): 2942, 2870, 2244,
Finally, it is noteworthy that, within the 5b-H series, the
selectivity of the inhibitors against 5aR-I and 5aR-II
can be reversed by the introduction of the benzoyl
group, being the N-phenyl amino substituted compound
3b more potent toward 5aR-I and its N-benzoyl deriva-
tive 5b more potent toward 5aR-II.
Conclusions
In this paper we have described the synthesis of
novel 17b-[N-(phenyl)-methyl/phenyl]-amido substituted
19-nor-10-azasteroids and their evaluation as inhibitors
against both the isoforms of human 5a-reductase. We
applied a hetero Diels–Alder reaction based methodol-
ogy in order to obtain each compound as 5a and 5b
epimer. We found that, surprisingly, the 5b-H com-
pounds were more active than their transition state
analogue counterparts 5a-H. These are the first exam-
ples of 5b-H inhibitors having steroidal structure of the
enzyme 5a-reductase. Furthermore, the presence of the
substituent at C-17 can modulate the selectivity toward
5a-reductase type I and II.
1653, 1599, 1501 cmꢀ1
. Elem. Anal. calcd for
C19H26N2O: C, 76.45; H, 8.78; N, 9.43. Found: C,
76.42; H, 9.09; N, 9.04.
(+)-1,2,4,4aꢁ,5,6a,7,8,9aꢁ,9bꢀ-Dodecahydro-4-N-(tert-
butoxycarbonyl)-6aꢀ-methyl-7ꢀ-(N-phenylamino)-(3H)-
cyclopenta[f]quinoline-3-one (9). To lactame 8 (500 mg,
1.68 mmol), dissolved in anhydrous CH2Cl2 (10mL),
were added Et3N (259 mL, 1.85 mmol), Boc2O (1.48 g,
6.72 mmol) and DMAP (30mg, 0.15 mmol), and the
resulting solution refluxed for 12 h. Water (10mL) was
then added and, after separation of the phases, the
organic layer was washed with 1 M KHSO4 (10mL),
satd NaHCO3 (10mL), brine and then dried over
Na2SO4. After removal of the solvent and purification
of the crude oil by flash chromatography (eluant: first
CH2Cl2 and then CH2Cl2–MeOH, 40:1, Rf=0.53), pure
9 (6.56 g, 98%) was obtained as a white solid. 9: mp 67–
70 ꢁC; [a]D25 +11.9 (c 1.02, CHCl3). 1H NMR (CDCl3, d
ppm): 7.15–7.07 (m, 2H), 6.67–6.58 (m, 3H), 3.50–3.41
(m, 1H), 3.27 (td, J=10.3, 3.7 Hz, 1H), 2.58–2.39 (m,
2H), 2.37–2.25 (m, 1H), 1.90–1.55 (m, 6H), 1.51–1.23
(m, 6H), 1.50(s, 9H), 0.83 (s, 3H); 13C NMR (CDCl3, d
ppm): 170.6 (s), 154.0 (s), 148.1 (s), 129.1 (d, 2 C), 117.0
(d), 113.4 (d, 2 C), 83.8 (s), 62.9 (d), 62.6 (d), 49.4 (d),
44.0(d), 38.1 (t), 36.0(t), 33.2 (t), 30.0(t), 27.6 (q, 3
C), 26.8 (t), 24.6 (t), 22.6 (t), 12.1 (q); MS m/z (%): 398
Experimental
Melting points are uncorrected. Chromatographic
separations were performed under pressure on silica
gel using flash-column techniques; Rf values refer to
TLC carried out on 25-mm silica gel plates (Merck
F254), with the same eluant indicated for column
chromatography. 1H and 13C NMR spectra were
recorded at 200 and 50.33 MHz, respectively. EI mass
spectra were carried out at 70eV ionizing voltage. THF
was distilled from Na/benzophenone. CH2Cl2 was dis-
tilled from CaH2. All reactions requiring anhydrous
conditions were performed in flame-dried glassware.
(+)- 1,2,4,6,6a,7,8,9,9aa,9bb-Decahydro-6ab-methyl -
(3H)-cyclopenta[ f ]quinoline-3,7-dione (6) was prepared
as already described.10

3460
D. Scarpi et al. / Bioorg. Med. Chem. 10 (2002) 3455–3461
(M+, 4), 298 (9), 297 (3), 132 (100), 106 (14), 77 (8); IR
(CDCl3): 2876, 1740, 1662, 1601, 1147 cmꢀ1. Elem.
Anal. calcd for C24H34N2O3: C, 72.33; H, 8.60; N,
7.03. Found: C, 71.87; H, 8.47; N, 6.85.
(ꢀ)-17ꢀ-(N-phenylamino)-(5ꢁ)-10-azaestr-1-en-3-one (3a).
To a solution of 12 (429 mg, 1.0mmol) and 1-meth-
oxy-3-[(trimethylsilyl)oxy]-1,3-butadiene (585 mL, 3.0
mmol) in anhydrous CH2Cl2 (5 mL) cooled at 0 ꢁC,
TiCl4 (330 mL, 3.0mmol) was added dropwise and the
solution left under stirring 1 h at 25 ꢁC. Satd NaHCO3
(5 mL) was then added and the resulting mixture stirred
vigorously for 30min. After separation of the phases,
the product was extracted with CH2Cl2 (3ꢃ5 mL) and
the combined organic layers dried over Na2SO4. Filtra-
tion and evaporation of the solvent afforded crude 3a,
which was purified by flash chromatography, eluting
with acetone–petroleum ether, 1:1. Pure 3a was
obtained (Rf=0.50) as a brown solid (46 mg, 13%). 3a:
(+)-1,2,4,4aꢁ,5,6a,7,8,9aꢁ,9bꢀ-Dodecahydro-4-N-(tert-
butoxycarbonyl)-6aꢀ-methyl-7ꢀ-[(N-phenyl)benzamido]-
(3H)-cyclopenta[ f ]quinoline-3-one (10). To a mixture of
9 (2.00 g, 5.02 mmol) and anhydrous K2CO3 (1.53 g,
11.0mmol) in anhydrous THF (37 mL), benzoyl chlo-
ride (700 mL, 6.02 mmol) was added dropwise and the
suspension left under stirring in nitrogen atmosphere at
rt for 24 h. The precipitate was then filtered off and the
solvent evaporated. Satd NaHCO3 (30mL) was added
to the residue and the product extracted with CH2Cl2
(3ꢃ30mL). The combined organic phases were dried
over Na2SO4. After removal of the solvent and pur-
ification of the crude oil by flash chromatography
(eluant: CH2Cl2–MeOH, 30:1, Rf=0.38), pure 10 (1.77
g, 70%) was obtained as a yellowish solid. 10: [a]D25
+55.8 (c 1.02, CHCl3). 1H NMR (CDCl3, d ppm):
7.19–6.95 (m, 10H), 4.78 (m, 1H), 3.36–3.14 (m, 1H),
2.60–2.30 (m, 3H), 1.95–1.18 (m, 11H), 1.51 (s, 9H),
0.89 (s, 3H); 13C NMR (CDCl3, d ppm): 172.8 (s),
170.4 (s), 153.4 (s), 141.7 (s), 137.5 (s), 131.1 (d, 2 C),
128.7 (d), 128.4 (d, 2 C), 127.7 (d, 2 C), 127.4 (d, 2 C),
127.1 (d), 83.6 (t), 65.4 (d), 62.5 (d), 48.8 (d), 45.7 (d),
37.7 (d), 36.0(t), 33.0(t), 27.5 (q, 3 C), 26.8 (t), 25.0(t),
24.4 (t), 22.1 (t), 13.8 (q); MS m/z (%): 403 (M+-101,
0.27), 297 (5), 197 (5), 105 (100), 77 (33); IR (CDCl3):
25
mp 69–70 ꢁC; [a]D ꢀ112.3 (c 0.26, CHCl3). H NMR
(CDCl3, d ppm): 7.19–7.08 (m, 3H), 6.68–6.58 (m, 3H),
4.96 (d, J=7.6 Hz, 1H, 2-H), 3.43 (t, J=8.2 Hz, 2H),
2.65 (m, 1H, 5-H), 2.64 (dd, J=17.2, 6.6 Hz, 1H) 2.30
(m, 3H), 1.97–1.55 (m, 6H), 1.55–1.11 (m, 6H), 0.79 (s,
3H); 13C NMR (CDCl3, d ppm): 192.0(s), 149.5 (d),
148.2 (s), 129.8 (d, 2 C), 117.5 (d), 113.6 (d, 2 C), 98.0
(d), 66.2 (d), 62.8 (d), 59.0(d), 51.0(d), 43.5 (s), 42.3 (d),
41.0(t), 36.1 (t), 29.8 (t), 28.8 (t), 28.2 (t), 24.6 (t), 22.1
(t), 11.9 (q); MS m/z (%): 350(M +, 15), 236 (42), 218
(36), 137 (32), 132 (100), 91 (98); IR (CDCl3): 3001,
1
1629, 1601, 1579 cmꢀ1
.
(+)-17ꢀ-(N-phenylamino)-(5ꢀ)-10-azaestr-1-en-3-one (3b).
To a solution of 12 (570mg, 1.33 mmol), 1-methoxy-
3-[(trimethylsilyl)oxy]-1,3-butadiene (518 mL, 2.66 mmol)
and Et3N (408 mL, 2.93 mmol) in anhydrous CH2Cl2 (9
mL) cooled at 0 ꢁC was added dropwise TMSOTf (411
mL, 2.13 mmol). The solution was allowed to warm to rt
and left under stirring for 45 min. Satd NaHCO3 (9 mL)
was then added and the mixture vigorously stirred for
36 h. After separation of the phases, the product was
extracted with CH2Cl2 (3ꢃ9 mL) and the combined
organic layers dried over Na2SO4. Filtration and eva-
poration of the solvent afforded crude 3b, which was
purified by flash chromatography, eluting with CH2Cl2–
MeOH, 30:1. Pure 3b was obtained (Rf=0.69) as a pale
yellow solid (94 mg, 20%). 3b: mp 98–100 ꢁC; [a]D25
+103.2 (c 0.76, CHCl3). 1H NMR (CDCl3, d ppm):
7.22–7.09 (m, 3H), 6.70–6.61 (m, 3H), 4.91 (d, J=7.4
Hz, 1H, 2-H), 3.78 (m, 1H, 5-H), 3.55 (s br, 1H), 3.43 (t,
J=8.4 Hz, 1H), 3.23 (m, 1H), 2.30(m, 1H), 2.03 (m,
3001, 2984, 1758, 1642, 1347, 1246, 1146 cmꢀ1
.
1,2,4,4aꢁ,5,6a,7,8,9aꢁ,9bꢀ-Dodecahydro-4-N-(tert-but-
oxycarbonyl)-3ꢀ-ethoxy-6aꢀ-methyl-7ꢀ-[(N-phenyl)benz-
amido]-(3H)-cyclopenta[ f ]quinoline-3-one (11). To
a
solution of 10 (1.67 g, 3.32 mmol) in anhydrous THF
(13 mL), cooled at ꢀ78 ꢁC, was added dropwise a 1 M
LiBHEt3 solution in THF (5.0mL, 5.0mmol). After
15 min, 2 N HCl in abs EtOH was added up to pH 3,
immediately followed by further abs EtOH (10mL) and
the solution warmed up to 0 ꢁC and left under stirring
for 30min. After dilution with CH Cl2 (130mL), the
2
organic phase was washed with satd NaHCO3 (100
mL), H2O (100 mL), brine (100 mL) and dried over
Na2SO4. Filtration and evaporation of the solvent
afforded crude 11 (1.01 g, 57%), in sufficiently pure form
to be used in the next reaction without further purifica-
tion. 11: 1H NMR (CDCl3, d ppm): 7.20–7.04 (m, 10H),
5.41 (d, J=5.8 Hz, 1H), 4.83–4.70(m, 2H), 3.60–3.48 (m,
1H), 3.45–3.33 (m, 1H), 3.13–3.02 (m, 1H), 2.23–2.12 (m,
1H), 1.84–1.13 (m, 15H), 1.45 (s, 9H), 0.87 (s, 3H).
2H), 1.95–1.10(m, 13H), .079 (s, 3H);
13C NMR
(CDCl3, d ppm): 191.1 (s), 149.5 (s), 148.1 (d), 129.1 (d,
2 C), 117.0(d), 113.1 (d, 2 C), 96.5 (d), 62.7 (d), 59.6 (d),
53.1 (d), 52.2 (d), 43.3 (t), 42.3 (t), 36.0(t), 35.7 (d), 29.9
(t), 26.2 (t), 25.6 (t), 23.7 (t), 22.6 (t), 11.9 (q); MS m/z
(%): 350(M +, 38), 284 (8), 132 (100), 105 (11), 77 (21);
IR (CDCl3): 3695, 2930, 2855, 1600, 1253 cmꢀ1. Elem.
Anal. calcd for C23H30N2O: C, 78.82; H, 8.63; N, 7.99.
Found: C, 78.50; H, 9.12; N, 7.26.
1,2,4,4aꢁ,5,6a,7,8,9aꢁ,9bꢀ-Dodecahydro-4-N-(tert-bu-
toxycarbonyl)-3ꢀ-ethoxy-6aꢀ-methyl-7ꢀ-(N-phenyl-
amino)-(3H)-cyclopenta[ f ]quinoline-3-one (12). Prepared
as described for 11, starting from 9 (2.00 g, 5.02 mmol)
and obtaining crude 12 (1.23 g, 57%) in sufficiently pure
form to be used in the next reaction without further
17ꢀ-[(N-phenyl)benzamido]-10-azaestr-1-en-3-one
(5).
Prepared as described for 3b, starting from 11 (799 mg,
1.5 mmol) but obtaining 5 as a 1.5:1 mixture of 5a(5a)/
5b(5b) diastereoisomers. Purification of the mixture by
flash chromatography (eluant: acetone) afforded pure
5a (Rf=0.58; 95 mg, 14%) and 5b (Rf=0.29; 89 mg,
13%). 5a: mp 109–110 ꢁC; [a]D25 ꢀ20.6 (c 0.04, CHCl3).
1
purification. 12: H NMR (CDCl3, d ppm): 7.19–7.05
(m, 2H), 6.70–6.55 (m, 3H), 5.39 (d, J=5.8 Hz, 1H),
3.63–3.47 (m, 2H), 3.44–3.29 (m, 2H), 3.68 (m, 1H),
2.35–2.10(m, 2H), 1.95–1.50(m, 6H), 1.44 (s, 9H),
1.35–1.10(m, 5H), 0.84 (s, 3H).

D. Scarpi et al. / Bioorg. Med. Chem. 10 (2002) 3455–3461
3461
1H NMR (CDCl3, d ppm): 7.22–6.95 (m, 11H), 4.97 (d,
Acknowledgement
J=7.8 Hz, 1H, 2-H), 4.78 (m, 1H), 3.40(m, 1H, 9-H),
2.69 (m, 1H, 5-H), 2.56 (dd, J=17.2, 6.2 Hz, 1H), 2.29
(dd, J=16.2, 8.8 Hz, 1H), 2.02–1.19 (m, 14H), 0.87 (s,
3H); 13C NMR (CDCl3, d ppm): 191.8 (s), 173.1 (s),
148.8 (d), 141.8 (d), 137.7 (s), 131.3 (d, 2 C), 128.9 (d),
128.6 (d, 2 C), 127.9 (d, 2 C), 127.6 (d, 2 C), 127.3 (d),
98.6 (d), 66.5 (d), 65.4 (d), 59.0(d), 50.2 (d), 45.4 (s),
42.6 (d), 40.9 (t), 36.3 (t), 30.5 (t), 28.7 (t), 25.3 (t), 24.8
(t), 22.2 (t), 13.7 (q); MS m/z (%): 454 (M+, 2), 349 (3),
105 (65), 84 (100), 77 (22); IR (CDCl3): 1708, 1636,
1579, 1466 cmꢀ1. 5b: mp 92–93 ꢁC; [a]2D5 +73.4 (c 0.03,
CHCl3). 1H NMR (CDCl3, d ppm): 7.25–6.95 (m, 11H),
4.91 (d, J=7.6 Hz, 1H, 2-H), 4.78 (t, J=9.9 Hz, 1H),
3.76 (m, 1H, 5-H), 3.26 (td, J=10.6, 4.4 Hz, 1H, 9-H),
2.50–2.20 (m, 2H), 2.10–1.90 (m, 2H), 1.90–1.00 (m,
12H), 0.82 (s, 3H); 13C NMR (CDCl3, d ppm): 198.0(s),
173.1 (s), 149.6 (d), 141.8 (s), 135.7 (s), 131.3 (d, 2 C),
128.9 (d), 128.6 (d, 2 C), 128.0(d, 2 C), 127.6 (d, 2 C),
127.3 (d), 96.6 (d), 65.3 (d), 59.5 (d), 53.2 (d), 51.8 (d),
45.3 (s), 42.5 (d), 36.3 (t), 35.5 (t), 30.9 (t), 26.3 (t), 25.7
(t), 23.6 (t), 22.2 (t), 13.7 (q); MS m/z (%): 454 (M+,
12), 105 (100), 77 (32); IR (CDCl3): 1709, 1631, 1571,
The authors thank Ministero dell’Universita e della
Ricerca Scientifica (MURST-COFIN 2000–2002), CNR
(Target Project on Biotechnology, Grants 01.00752.PF49).
References and Notes
1. Geller, J. J. Am. Geriatr. Soc. 1991, 39, 1208.
2. Smith, L. S.; Tegler, J. J. Annu. Rep. Med. Chem. 1989, 24,
177.
3. Dallob, A.; Sadick, N. S.; Unger, W.; Lipert, S.; Geissler,
L. A.; Gregoire, S. L.; Nguyen, H. H.; Moore, E. C.; Tanaka,
W. K. J. Clin. Endocrinol. Metab. 1994, 79, 703.
4. Brooks, J. R. Clin. Endocrinol. Metab. 1986, 15, 391.
5. (a) Frye, S. V. Curr. Pharm. Des. 1996, 2, 59. (b) Machetti,
F.; Guarna, A. Expert Opin. Ther. Pat. 2002, 12, 201.
6. (a) Geller, J. J. Clin. Endocrinol. Metab. 1990, 71, 1552. (b)
Stoner, E. J. Steroid Biochem. Molec. Biol. 1990, 37, 375. (c)
Gormley, G. J.et al. New Engl. J. Med. 1992, 327, 1185.
7. (a) Rasmusson, G. H.; Reynolds, G. F.; Utne, T.; Jobson,
R. B.; Primka, R. L.; Berman, C.; Brooks, J. R. J. Med. Chem.
1984, 27, 1690. (b) Rasmusson, G. H.; Reynolds, G. F.;
Steinberg, N. G.; Walton, E.; Patel, G. F.; Liang, T.; Cascieri,
M. A.; Cheung, A. H.; Brooks, J. R.; Berman, C. J. Med.
Chem. 1986, 29, 2298. (c) Mellin, T. N.; Busch, R. D.; Ras-
musson, G. H. J. Steroid Biochem. Mol. Biol. 1993, 44, 121. (d)
Bakshi, R. K.; Patel, G. F.; Rasmusson, G. H.; Baginsky,
W. F.; Cimis, G.; Ellsworth, K.; Chang, B.; Bull, H.; Tolman,
R. L.; Harris, G. S. J. Med. Chem. 1994, 37, 3871.
8. (a) Frye, S. V.; Haffner, C. D.; Maloney, P. R.; Mook, R. A.,
Jr.; Dorsey, G. F., Jr.; Hiner, R. N.; Batchelor, K. W.; Bram-
son, H. N.; Stuart, J. D.; Schweiker, S. L.; Van Arnold, J.;
Bickett, D. M.; Moss, M. L.; Tian, G.; Unwalla, R. J.; Lee,
F. W.; Tippin, T. K.; James, M. K.; Grizzle, M. K.; Long, J. E.;
Schuster, S. V. J. Med. Chem. 1993, 36, 4313. (b) Frye, S. V.;
Haffner, C. D.; Maloney, P. R.; Mook, R. A., Jr.; Dorsey,
G. F., Jr.; Hiner, R. N.; Cribbs, C. M.; Wheleer, T. N.; Ray,
J. A.; Andrews, R. C.; Batchelor, K. W.; Bramson, H. N.;
Stuart, J. D.; Schweiker, S. L.; Van Arnold, J.; Croom, S.;
Bickett, D. M.; Moss, M. L.; Tian, G.; Unwalla, R. J.; Lee,
F. W.; Tippin, T. K.; James, M. K.; Grizzle, M. K.; Long, J. E.;
Schuster, S. V. J. Med. Chem. 1994, 37, 2352.
9. (a) Guarna, A.; Belle, C.; Machetti, F.; Occhiato, E. G.;
Payne, A. H.; Cassiani, C.; Comerci, A.; Danza, G.; De Bellis, A.;
Dini, S.; Marrucci, A.; Serio, M. J. Med. Chem. 1997, 40, 1112.
(b) Guarna, A.; Occhiato, E. G.; Machetti, F.; Marrucci, A.;
Danza, G.; Serio, M.; Paoli, P. J. Med. Chem. 1997, 40, 3466.
10. (a) Guarna, A.; Occhiato, E. G.; Machetti, F.; Scarpi, D.
J. Org. Chem. 1998, 63, 4111. (b) Guarna, A.; Occhiato, E. G.;
Machetti, F.; Giacomelli, V. J. Org. Chem. 1999, 64, 4985.
11. For recent publications on hetero Diels–Alder reactions of
silyloxy dienes with acyliminium ions, see: Waldmann, H.
Synthesis 1994, 535. Also: Rutjes, F.P.J.T.; Veerman, J.J.N.;
Meester, W.J.N.; Hiemstra, H.; Schoemaker, H.E. Eur. J. Org.
Chem. 1999, 5, 1127.
1355 cmꢀ1
.
(ꢀ)-17ꢀ-[(N-phenyl)acetamido]-(5ꢁ)-10-azaestr-1-en-3-
one (4a). To a solution of 3a (15 mg, 0.04 mmol) in
pyridine (750 mL) cooled at 0 ꢁC was added dropwise
Ac2O (30 mL, 0.27 mmol). The solution was allowed to
warm to rt and left under stirring for 18 h. Water was
added (5 mL) and the product extracted with CH2Cl2
(2ꢃ5 mL). The combined organic layers were then
washed with 5% citric acid (10mL), 10% NaHCO (10
3
mL), H2O (10mL) and dried over Na SO4. Filtration
2
and evaporation of the solvent afforded 4a (15 mg,
95%). 4a: mp 72–74 ꢁC; [a]D25 ꢀ114.3 (c 0.21, CHCl3).
1H NMR (CDCl3, d ppm): 7.35 (m, 3H), 7.18 (d, J=7.9
Hz, 1H, 1-H), 6.98 (m, 2H), 4.96 (d, J=7.9 Hz, 1H, 2-
H), 4.59 (t, J=9.7 Hz, 1H), 3.40(m, 1H, 9-H), 2.60(m,
1H, 5-H), 2.63 (dd, J=16.5, 6.6 Hz, 1H), 2.28 (dd,
J=16.5, 8.8 Hz, 1H), 1.94 (m, 1H), 1.80–1.00 (m, 13H),
1.71 (s, 3H), 0.70 (s, 3H); 13C NMR (CDCl3, d ppm):
191.8 (s), 172.4 (s), 148.8 (d), 141.7 (s), 131.2 (d), 130.5
(d), 129.2 (d, 2 C), 128.1 (d), 98.5 (d), 66.1 (d), 64.2 (d),
59.0(d), 50.1 (d), 45.2 (s), 42.6 (d), 41.3 (t), 36.5 (t), 34.1
(t), 29.7 (t), 25.5 (t), 22.6 (t), 22.3 (q), 22.1 (t), 14.0(q);
MS m/z (%): 392 (M+, 34), 132 (100), 83 (44); IR
(CDCl3): 1715, 1640, 1585 cmꢀ1
.
(+)-17ꢀ-[(N-phenyl)acetamido]-(5ꢀ)-10-azaestr-1-en-3-
one (4b). As described for 4a, starting from 3b (7 mg,
0.02 mmol) and obtaining 4b (8 mg, quantitative) as a
pale yellow solid. 4b: mp 95–96 ꢁC; [a]D25 +150.0 (c 0.27,
1
12. Occhiato, E. G.; Scarpi, D.; Machetti, F.; Guarna, A.
Tetrahedron 1998, 54, 11589.
13. Li, X.; Singh, S. M.; Labrie, F. J. Med. Chem. 1995, 38, 1158.
14. Guarna, A.; Machetti, F.; Occhiato, E. G.; Scarpi, D.;
Comerci, A.; Danza, G.; Mancina, R.; Serio, M.; Hardy, K.
J. Med. Chem. 2000, 43, 3718.
15. (a) Holt, D.A.; Levy, M.A.; Metcalf, B.W. Inhibition of
Steroid 5a-Reductase. In Advances in Medicinal Chemistry;
Maryanoff, B.E.; Maryanoff, C.A., Eds.; JAI: Greenwich, CT,
1993; Vol. 2, p 1. (b) Abell, A. D.; Henderson, B. R. Curr.
Med. Chem. 1995, 2, 583.
CHCl3). H NMR (CDCl3, d ppm): 7.35–6.95 (m, 6H),
4.91 (d, J=7.3 Hz, 1H, 2-H), 4.58 (t, J=9.8 Hz, 1H),
3.76 (m, 1H, 5-H), 3.20 (m, 1H, 9-H), 2.40–2.00 (m,
4H), 1.71 (s, 3H), 1.90–0.75 (m, 12H), 0.66 (s, 3H); 13C
NMR (CDCl3, d ppm): 195.6 (s), 172.4 (s), 149.6 (d),
141.6 (s), 130.5 (d, 2 C), 129.2 (d, 2 C), 128.1 (d), 96.5
(d), 64.1 (d), 59.4 (d), 53.2 (d), 51.7 (d), 43.3 (s), 42.4 (d),
36.1 (t), 35.4 (t), 29.7 (t), 26.1 (t), 25.4 (t), 23.6 (t), 22.1
(t), 19.1 (q), 13.3 (q); MS m/z (%): 392 (M+, 1), 149 (8),
86 (59), 84 (100); IR (CDCl3): 3692, 1639, 1574 cmꢀ1
.