Synthetic approaches for the synthesis of a cytostatic steroidal B-D bilactam

Article abstract of DOI:10.1016/S0039-128X(03)00095-3

A new synthetic procedure and a modification of the original method described in the literature for the synthesis of the steroidal B-D bilactam, 3β-hydroxy-7α,17α-diaza-B,D-dihomo-5-androsten-7,17-dione are reported. The key step in the modified method involved protection of the D-lactamic nitrogen atom of 3β-acetoxy-17α -aza-D-homo-5-androsten-17-one using a reagent of specific electrophilicity (due to the stereoelectronic properties of the cyclic amide), as Beckmann rearrangement of the B-steroidal ring was hindered, possibly via long range effects, by the presence of the unprotected D-lactamic moiety. Using the 3β-acetoxy-5-androsten-17-one as starting material, a new synthetic procedure was developed through ketalization of the 17-ketone and allylic oxidation to the 7-ketone, which was subsequently followed by Beckmann rearrangement of the B- and D-steroid rings. Both approaches resulted in 45 and 67% yields of the desired B,D-bilactam, respectively, in contrast to the 15% yield, which has been reported in the literature.

Full text of DOI:10.1016/S0039-128X(03)00095-3

Steroids 68 (2003) 659–666
Synthetic approaches for the synthesis of
a cytostatic steroidal B–D bilactam
Anna I. Koutsourea, Evaggelia S. Arsenou,
Manolis A. Fousteris, Sotiris S. Nikolaropoulos∗
Laboratory of Pharmaceutical Chemistry, School of Health Sciences, Department of Pharmacy,
University of Patras, 26500 Patras, Greece
Received 13 January 2003; received in revised form 3 April 2003; accepted 11 April 2003
Abstract
A new synthetic procedure and a modification of the original method described in the literature for the synthesis of the steroidal B–D
bilactam, 3␤-hydroxy-7␣,17␣-diaza-B,D-dihomo-5-androsten-7,17-dione are reported. The key step in the modified method involved
protection of the D-lactamic nitrogen atom of 3␤-acetoxy-17␣-aza-D-homo-5-androsten-17-one using a reagent of specific electrophilicity
(due to the stereoelectronic properties of the cyclic amide), as Beckmann rearrangement of the B-steroidal ring was hindered, possibly via
long range effects, by the presence of the unprotected D-lactamic moiety. Using the 3␤-acetoxy-5-androsten-17-one as starting material, a
new synthetic procedure was developed through ketalization of the 17-ketone and allylic oxidation to the 7-ketone, which was subsequently
followed by Beckmann rearrangement of the B- and D-steroid rings. Both approaches resulted in 45 and 67% yields of the desired
B,D-bilactam, respectively, in contrast to the 15% yield, which has been reported in the literature.
© 2003 Elsevier Inc. All rights reserved.
Keywords: Beckmann rearrangement; B,D-bilactam; Homo-aza-steroids; N-Benzoyl protection; Steroidal synthesis
1. Introduction
responding D-homo-aza-steroid 2 (Fig. 1) has been reported
to produce increased proliferation in acute non-lymphocytic
leukemia cells, which suggests an enhanced potential for
further differentiation and may, thus, contribute to clini-
cal remission of the disease [4]. Additionally, most of the
compounds that carry alkylating agents on either their A
or D rings are inactive against L1210 leukemia [5], while
homo-aza-steroidal esters of the same alkylating agents
proved much more effective in vitro and in vivo [6–8]. This
evidence led us to the suggestion that the lactam moiety
confers antileukemic activity to the esteric steroidal alky-
lating agents, which, additionally, were less toxic than the
alkylating agents themselves [9].
In order to investigate the lactam’s contribution to the fi-
nal antileukemic activity and to study the structure–activity
relationships for steroids of this category, we decided to
study compound 1. Another fact that contributed to this de-
cision was the observation that the introduction of a second
–NHCO– group at the 17-position on an A-lactamic steroid
resulted in derivative 3 (Fig. 1) with inherent cytostatic ac-
tivity. It was found that compound 3 inhibited DNA syn-
thesis reversibly, bound to nucleic acids, and increased life
span of treated P388-bearing animals by 48–58% [10].
Experimental, clinical, and epidemiological studies have
shown that steroid hormones, as well as their agonists and
antagonists, produce salutary and, in some cases, excellent
results in the prevention and treatment of cancer. Several of
these steroid hormones are in clinical use for the treatment
of hormone-depended tumors (breast, prostate, ovarian, en-
dometrium), as well as for the treatment of paraneoplas-
tic syndromes and cancer symptoms [1,2]. Moreover, it has
been reported that steroid hormones influence the growth of
many cancers, and the presence of tumor-associated recep-
tors for these hormones offers the opportunity for targeting
using drug–hormone conjugates [3].
In order to design, synthesize, and evaluate the antitumor
activity of new steroidal derivatives as a part of a project for
the synthesis of new drug–hormone conjugates, we decided
to study the modified androgenic skeleton carrying two
lactamic moieties at the B and D rings, 1 (Fig. 1). The cor-
∗
Corresponding author. Tel.: +30-2610-997723;
fax: +30-2610-992776.
E-mail address: snikolar@upatras.gr (S.S. Nikolaropoulos).
0039-128X/$ – see front matter © 2003 Elsevier Inc. All rights reserved.
doi:10.1016/S0039-128X(03)00095-3

660
A.I. Koutsourea et al. / Steroids 68 (2003) 659–666
2.2. 3β-Acetoxy-17α-aza-D-homo-N-benzoyl-5-
androsten-7,17-dione (5)
H
N
O
Benzoyl chloride (0.35 ml; 3 mmol) was added to a so-
lution of 3␤-acetoxy-17␣-aza-D-homo-5-androsten-7,17-
dione (4) (0.36 g, 1 mmol) in 10 ml of dry benzene. The
reaction mixture was refluxed under anhydrous condi-
tions for 18 h. Then, the solvent was removed in vacuum,
and the residue was diluted in DCM. The crude mix-
ture was chromatographed over silica gel/DCM, and the
product was eluted with DCM/methanol (99:1) to yield
3␤-acetoxy-17␣-aza-D-homo-N-benzoyl-5-androsten-7,17-
dione (0.39 g, 85%). The m.p. 232–235 ◦C (methanol). IR
NH
O
AcO
1
H
N
NHCOCH3
O
HN
AcO
H
O
2
3
17
7
ν (cm−1): 1732 (C O acetate), 1701 ( C O), 1674 ( C O
=
=
=
=
=
and C O benzoyl), 1599 (C C), 1240, 1032 (C–O acetate).
1H NMR (CDCl3) δ: 7.8 (d, 2H), 7.53 (t, 1H), 7.433 (t, 2H),
5.8 (s, 1H), 4.715 (m, 1H), 2.072 (s, 3H), 1.566 (s, 3H), 1.25
(s, 3H). Analysis calculated for C28H33NO5 (M = 463.57):
C, 72.55; H, 7.18; N, 3.02; found: C, 72.25; H, 7.12; N, 3.10.
Fig. 1. Homo-aza-steroids of biological interest.
This paper describes the results of our endeavors to im-
prove the synthetic procedure for the preparation of 1, as in
the literature there was only one reference for its synthesis
with poor yield (<15%) [11]. The observation that the syn-
thesis of the B-lactam was attainable in acceptable yields
(60–70%) when the configuration of the steroidal D-ring
was changed to a non-lactamic structure, as in cholestene,
[12,13], led us to the conclusion that the existence of the lac-
tamic D-ring hinders, as proved by control experiments, the
formation of the B-lactam mainly by intermolecular effects.
This suggestion resulted in a modification of the original
method found in the literature, which enabled us to prepare
1 in 45% yield. This modified method, as well as another
new synthetic approach, which afforded a better yield (67%)
of the B,D-bilactam, are presented and discussed.
2.3. 3β-Acetoxy-17α-aza-D-homo-N-benzoyl-5-
androsten-7,17-dione oxime (6)
3␤-Acetoxy-17␣-aza-D-homo-N-benzoyl-5-androsten-7,
17-dione (5) (0.39 g, 0.85 mmol) was diluted in 5 ml of
absolute ethanol and 3 ml of pyridine. Hydroxylamine hy-
drochloride (0.069 g; 1 mmol) was added to this solution,
and the reaction mixture was stirred at room temperature
for 36 h. After completion of the reaction, the mixture was
poured into ice–water, and the white precipitate was fil-
tered under vacuum, washed with water, and dried. The
oxime was obtained in 90% yield (0.36 g, 0.77 mmol). The
m.p. 162–164 ◦C (methanol). IR ν (cm−1): 3287 (N–OH),
1728 (C O acetate), 1705 ( C O), 1678 (C O benzoyl)
1242, 1032 (C–O acetate). H NMR (CDCl3) δ: 7.846 (d,
17
=
=
=
2. Experimental
1
2H), 7.547 (t, 1H), 7.447 (t, 2H), 6.59 (s, 1H), 4.716 (m,
1H), 2.062 (s, 3H), 1.591 (s, 3H), 1.15 (s, 3H). Analysis
calculated for C28H34N2O5 (M = 478.58): C, 70.27; H,
7.16; N, 5.85; found: C, 70.18; H, 7.09; N, 5.94.
2.1. General
Melting points were measured on a Kallenkamp melting
point apparatus in capillary tubes and are uncorrected. IR
spectra were recorded on a FT-IR Jasco spectrophotome-
1
2.4. 3β-Acetoxy-7α,17α-diaza-B,D-dihomo-
N17-benzoyl-5-androsten-7,17-dione (7)
ter, and H NMR spectra were recorded on a Brucker 400
spectrometer in CDCl3. The chemical shifts are given as δ
values (ppm) with tetramethylsilane as the internal standard.
Elemental analyses were performed on a Carlo Erba, CHN
Analyzer. Thin layer chromatography (TLC) was performed
on E. Merck precoated silica gel plates (Kieselgel 60 F254).
Visualization was accomplished by exposure to iodide va-
pors and/or under UV light (254 nm). Column chromatog-
raphy was performed with silica gel (E. Merck, 70–230
mesh).
3␤-Hydroxy-5-androsten-17-one (9) was purchased from
Merck, Germany. 3␤-Acetoxy-17␣-aza-D-homo-5-andro-
sten-17-one [14] and 3␤-acetoxy-17␣-aza-D-homo-5-
androsten-7,17-dione (4) [15] were prepared according to
procedures in the literature.
Freshly distilled thionylochloride (1.5 ml) in 3 ml of diox-
ane were added dropwise to a solution of 3␤-acetoxy-17␣-
aza-D-homo-N-benzoyl-5-androsten-7,17-dione oxime (6)
(0.36 g, 0.77 mmol) in 10 ml dry dioxane cooled to 0 ◦C.
Afterwards, the reaction mixture was stirred for 4 h under
a nitrogen atmosphere and room temperature. The mixture
was then poured into ice–water, neutralized with a solution
of NH3 (pH = 7), and extracted with DCM. The organic
layer was washed with water and dried over anhydrous
Na2SO4. Evaporation of the solvent gave a brown oily
residue, which was chromatographed over silica gel/DCM,
and the product was eluted with DCM/methanol (97:3) to

A.I. Koutsourea et al. / Steroids 68 (2003) 659–666
661
yield compound 7 (0.17 g, 45%). The m.p. 176–177 ◦C
and filtered through a short column of silica gel layered with
Celite. The column was eluted with the same solvent mix-
ture. The oily residue collected was chromatographed on a
column of silica gel with a mixture of DCM–methanol (9:1)
to give 3␤-acetoxy-5-androsten-7-one-17,17-ethylendioxy
(4 g, 80%). The m.p. 176–177 ◦C (methanol) ([16]
(methanol). IR ν (cm−1): 3271, 3182 (N–H), 1738 (C O
=
17
7
=
=
=
acetate), 1701 ( C O), 1660 ( C O and C O benzoyl),
1614 (C C), 1244, 1035 (C–O). H NMR (CDCl3) δ: 7.705
1
=
(d, 2H), 7.467 (t, 1H), 7.365 (t, 2H), 6.894 (s, 1H), 5.803
(s, 1H), 4.634 (m, 1H), 3.227 (t, 1H), 1.976 (s, 3H), 1.467
(s, 3H), 1.208 (s, 3H). Analysis calculated for C28H34N2O5
(M = 478.58): C, 70.27; H, 7.16; N, 5.85; found: C, 70.40;
H, 7.05; N, 5.99.
175–177 ◦C). IR ν (cm−1): 1732 (C O acetate), 1668
=
7
1
=
=
( C O), 1615 (C C), 1242, 1037 (C–O acetate). H NMR
(CDCl3) δ: 5.636 (s, 1H), 4.648 (m, 1H), 3.8 (m, 4H), 1.983
(s, 3H), 1.139 (s, 3H), 0.802 (s, 3H). Analysis calculated
for C23H32O5 (M = 388.50): C, 71.11; H, 8.33; found: C,
70.87; H, 8.52.
2.5. 3β-Hydroxy-7α,17α-diaza-B,D-dihomo-5-
androsten-7,17-dione (8)
3␤-Acetoxy-7␣,17␣-diaza-B,D-dihomo-N17-benzoyl-5-
androsten-7,17-dione (7) (0.17 g, 0.35 mmol) was diluted
in 5 ml methanol. Two ml of LiOH 1N was added to this
solution, and the mixture was stirred at room temperature
for 5 h. The reaction mixture was filtered under vacuum
to remove the remaining salt. Evaporation of the solvent
2.8. 3β-Acetoxy-5-androsten-7-one-17,17-
(ethylenedioxy) oxime (12)
Hydroxylamine hydrochloride (1.66 g, 23 mmol) was
added to a solution of 3␤-acetoxy-5-androsten-7-one-17,17-
ethylendioxy (6.7 g, 17.2 mmol) in absolute ethanol (50 ml)
and pyridine (18 ml). The mixture was stirred at room tem-
perature for 10 h. The reaction mixture was then poured into
a mixture of ice and water to give a white precipitate. Fil-
tration under vacuum and washing with water gave 12 as a
white solid (6.5 g, 94%). The m.p. 203–205 ◦C (methanol).
gave 0.1 g of 8 (90%). The m.p. 272–274 ◦C (methanol).
IR ν (cm−1): 3330 (OH), 3280, 3155 (NH), 1657 ( C O
7
=
17
1
=
=
and C O), 1604 (C C), 1063 (C–O). H NMR (CDCl3)
δ: 7.611 (s, 1H), 7.471 (s, 1H), 5.575 (s, 1H), 3.686 (m,
1H), 3.038 (t, 1H), 1.155 (s, 3H), 1.001 (s, 3H). Analysis
calculated for C19H28N2O3 (M = 332.44): C, 68.65; H,
8.49; N, 8.43; found: C, 68.51; H, 8.61; N, 8.25.
IR ν (cm−1): 3398 (N–OH), 1730 (C O acetate), 1641
=
(C C), 1255, 1033 (C–O acetate). 1H NMR (CDCl3) δ:
=
6.778 (s, 1H), 6.496 (s, 1H), 4.584 (m, 1H), 3.78 (m, 4H),
1.97 (s, 3H), 1.06 (s, 3H), 0.814 (s, 3H). Analysis calcu-
lated for C23H33NO5 (M = 403.51): C, 68.46; H, 8.24; N,
3.47; found: C, 68.81; H, 8.23; N, 3.12.
2.6. 3β-Acetoxy-5-androsten-17,17-(ethylenedioxy) (10)
p-Toluenesulfonic acid (0.2 g, 1.051 mmol) was added
to a solution of 3␤-acetoxy-5-androsten-17-one (10.5 g,
32 mmol) in triethylorthoformate (15 ml) and ethylene
glycol (6 ml), and the mixture was stirred at 90 ◦C for
1 h. The reaction mixture was then concentrated to about
6 ml by heating at 110–120 ◦C and diluted with methanol
(30 ml) and pyridine (2 ml). Water was added dropwise
to this solution until a white precipitate formed. Filtration
under vacuum gave 10 as a white solid (11.72 g, 98%).
The m.p. 140–141 ◦C ([16] 140–142 ◦C) (methanol). IR
2.9. 3β-Acetoxy-7α-aza-B-homo-5-androsten-7,17-
dione (13)
Freshly distilled thionylochloride (0.71 ml) in 2.25 ml
THF was added dropwise to a solution of 3␤-acetoxy-5-
androsten-7-one oxime (0.5 g, 1.23 mmol) in 3 ml freshly
distilled THF cooled to 0 ◦C, while maintaining the tem-
perature at 0 ◦C. After this addition, the reaction mixture
was stirred at 0 ◦C for 8 h under a nitrogen atmosphere.
The mixture was then poured into ice–water, neutralized
with a solution of NH3 (pH = 7), and extracted with DCM.
The organic layer was washed with water and dried over
anhydrous Na2SO4. Evaporation of the solvent gave an oily
residue, which was chromatographed over silica gel/DCM,
and the product was eluted with DCM/methanol (97:3) to
give 13 (0.23 g, 52%) and 14 (0.074 g, 15%). Compound 14
was hydrolyzed with an equimolar quantity of aq. HCl so-
lution in methanol to give 13 (98%). The m.p. 219–220 ◦C
ν (cm−1): 1732 (C O acetate), 1619 (C C), 1248, 1035
=
=
1
(C–O acetate). H NMR (CDCl3) δ: 5.317 (s, 1H), 4.552
(m, 1H), 3.8 (m, 4H), 1.967 (s, 3H), 0.96 (s, 3H), 0.8 (s,
3H). Analysis calculated for C23H34O4 (M = 374.51): C,
73.76; H, 9.15; found: C, 73.61; H, 8.94.
2.7. 3β-Acetoxy-5-androsten-7-one-17,17-
(ethylenedioxy) (11)
Chromium trioxide (23 g, 0.23 mol) was suspended in dry
DCM (140 ml) at −20 ◦C, and dimethylpyrazole (22.84 g,
0.23 mol) was added as quickly as possible. After 15 min,
a solution of 3␤-acetoxy-5-androsten-17,17-ethylendioxy
(5 g, 13 mmol) in dichloromethane (40 ml) was added drop-
wise. Stirring was continued for 5 h while maintaining the
temperature between −10 and −20 ◦C. The reaction mixture
was then diluted with a mixture of toluene–ethyl acetate (7:3)
(methanol). IR ν (cm−1): 3200 (NH), 1736 (C O acetate),
=
17
7
=
=
=
1728 ( C O), 1662 ( C O), 1616 (C C), 1244, 1032
(C–O acetate). H NMR (CDCl3) δ: 7.308 (s, 1H), 5.905
1
(s, 1H), 4.762 (m, 1H), 3.526 (t, 1H), 2.088 (s, 3H), 1.342
(s, 3H), 0.933 (s, 3H). Analysis calculated for C21H29NO4
(M = 359.46): C, 70.17; H, 8.13; N, 3.90; found: C, 69.97;
H, 8.32; N, 3.79.

662
A.I. Koutsourea et al. / Steroids 68 (2003) 659–666
2.10. 3β-Acetoxy-7α-aza-B-homo-5-androsten-7-one-
17,17-(ethylenedioxy) (14)
2.13. 3β-Acetoxy-7α-aza-B-homo-5-androsten-7-
one-13,17-seco-13(18)-en-17-onitrile (16)
IR ν (cm−1): 3190 (NH), 2245 (C N), 1734 (C O ac-
The m.p. 225–226 (methanol). IR ν (cm−1): 2951
≡≡
=
7
7
=
=
etate), 1658 ( C O), 1612 (C C), 1244, 1033 (C–O acetate).
1H NMR (CDCl3) δ: 6.718 (s, 1H), 5.863 (s, 1H), 5.029 (d,
2H), 4.708 (m, 1H), 3.684 (t, 1H), 3.590 (t, 1H), 3.384 (t,
1H), 2.047 (s, 3H), 1.164 (s, 3H). Analysis calculated for
C21H28N2O3 (M = 356.46): C, 70.76; H, 7.92; N, 7.86;
found: C, 70.51; H, 8.14; N, 7.88.
=
=
=
(NH), 1720 (C O acetate), 1662 ( C O), 1618 (C C),
1240, 1037 (C–O acetate). 1H NMR (CDCl3) δ: 5.855
(s, 1H), 5.603 (s, 1H), 4.742 (m, 1H), 3.922 (m, 4H),
2.623 (t, 1H), 2.193 (s, 3H), 2.062 (s, 3H), 0.877 (s,
3H). Analysis calculated for C23H33NO5 (M = 403.51):
C, 68.46; H, 8.24; N, 3.47; found: C, 68.74; H, 8.33;
N, 3.52.
2.14. Control experiment for the establishment of the type
of interference of the unprotected D-ring lactam at the
rearrangement of the 7-oxime
2.11. 3β-Acetoxy-7α-aza-B-homo-5-androsten-
7,17-dione oxime (15)
Freshly distilled thionylochloride (1.5 ml) in 3 ml of diox-
ane was added dropwise to a solution of 3␤-acetoxy-17␣-
aza-D-homo-N-benzoyl-5-androsten-7,17-dione oxime (6)
(0.36 g, 0.77 mmol) and 3␤-acetoxy-17␣-aza-D-homo-5-
androsten-17-one (0.26 g, 0.77 mmol) in 10 ml dry dioxane
cooled to 0 ◦C. After this addition, the reaction mixture was
stirred for 4 h under a nitrogen atmosphere and room tem-
perature. The mixture was then poured into ice–water, neu-
tralized with a solution of NH3 (pH = 7), and extracted with
DCM. The organic layer was washed with water and dried
over anhydrous Na2SO4. Evaporation of the solvent gave a
brown oily residue, which was chromatographed over silica
gel/DCM, and the product was eluted with DCM/methanol
(97:3) to yield a compound (0.15 g, 42%), which proved
identical with the previous obtained compound 7, while
the 3␤-acetoxy-17␣-aza-D-homo-5-androsten-17-one was
quantitative recovered.
3␤-Acetoxy-7␣-aza-B-homo-5-androsten-7,17-dione (14)
(0.6 g, 1.66 mmol) was diluted in 5 ml of absolute ethanol
and 3 ml of pyridine. Hydroxylamine hydrochloride (0.15 g;
2.15 mmol) was added to this solution, and the reaction mix-
ture was refluxed for 2 h. After completion of the reaction,
the mixture was poured into ice–water, and the white pre-
cipitate was filtered under vacuum, washed with water, and
dried. The oxime 15 was obtained in 80% yield (0.48 g). The
m.p. 252–255 ◦C (methanol). IR ν (cm−1): 3314 (N–OH),
7
=
=
=
3198 (NH), 1736 (C O acetate), 1651 ( C O), 1602 (C C),
1
1236, 1024 (C–O acetate). H NMR (CDCl3) δ: 7.204 (s,
1H), 5.881 (s, 1H), 5.7 (s, 1H), 4.738 (m, 1H), 3.442 (t, 1H),
2.068 (s, 3H), 1.319 (s, 3H), 0.959 (s, 3H). Analysis calcu-
lated for C21H30N2O4 (M = 374.47): C, 67.35; H, 8.07; N,
7.48; found: C, 67.13; H, 8.36; N, 7.59.
2.12. 3β-Acetoxy-7α,17α-diaza-B,D-dihomo-5-
androsten-7,17-dione (1)
3. Results and discussion
Freshly distilled thionylochloride (1 ml) in 1 ml of dio-
xane was added dropwise to a solution of 3␤-acetoxy-7␣-aza-
B-homo-5-androsten-7,17-dione oxime (0.27 g, 0.72 mmol)
in 2.5 ml dry dioxane cooled to 0 ◦C, while maintaining the
temperature at 0 ◦C. After this addition, the reaction mixture
was stirred at room temperature for 5 h under a nitrogen
atmosphere. The mixture was then poured into ice–water,
neutralized with a solution of NH3 (pH = 7), and extracted
with DCM. The organic layer was washed with water and
dried over anhydrous Na2SO4. Evaporation of the solvent
gave a brown, oily residue, which was chromatographed
over silica gel/DCM, and the product was eluted with
DCM/methanol (95:5) to yield 1 (60%, 0,16 g) and 16
(14%, 0,035 g) as an oily residue. 1 was recrystallized from
methanol. The m.p. 293–295 (methanol). IR ν (cm−1):
In a previous report [11], steroid 1 was obtained in 13.4%
yield by applying the Schmidt reaction on steroid 4 (Fig. 4i).
Despite the low yield of the procedure, there have been no
further efforts to improve the synthesis of 1. As a part of our
on-going research directed towards the chemical synthesis
and antineoplasmatic evaluation of new, modified steroidal
conjugates with alkylating agents, we investigated alterna-
tive methods to achieve the lactamic enlargement of the
B-ring of steroid 4. Consequently, the following two im-
proved synthetic procedures were developed (Figs. 2 and 3).
Fig. 2 outlines the first synthetic procedure developed,
which is a modification of the original method. From a series
of unsuccessful experiments conducted in our laboratory for
the preparation of B,D-bilactam 1 via Beckmann rearrange-
ment, it was concluded that Beckmann rearrangement on the
B-ring of 3␤-acetate-17␣-aza-D-homo-5-androsten-7-one
oxime as well as 3␤-acetate-5-androsten-7,17-dioxime was
unattainable (Fig. 4iii). On the contrary, it was known that
preparation of the lactamic B-ring of cholestene via Beck-
mann rearrangement has been obtained in sufficient yields
(60–70%) [12,13]. All the above-mentioned data led us to
17
7
=
=
=
3219 (NH), 1732 (C O acetate), 1658 ( C O and C O),
1
=
1612 (C C), 1251, 1039 (C–O acetate). H NMR (CDCl3)
δ: 6.775 (s, 1H), 6.369 (s, 1H), 5.851 (s, 1H), 4.7 (m, 1H),
3.19 (t, 1H), 2.051 (s, 3H), 1.273 (s, 3H), 1.170 (s, 3H).
Analysis calculated for C21H30N2O4 (M = 374.47): C,
67.35; H, 8.07; N, 7.48; found: C, 67.44; H, 8.52; N, 7.69.

A.I. Koutsourea et al. / Steroids 68 (2003) 659–666
663
O
H
N
O
N
O
i
ii
AcO
O
AcO
O
4
5
O
O
N
O
N
O
iii
iv
N H
O
AcO
AcO
N
HO
6
7
H
O
N
N H
HO
O
8
Fig. 2. Modified method for the synthesis of B,D-bilactam. Conditions: (i) Bz-Cl/dry benzene, reflux, 18 h; (ii) H2NOH·HCl/C5H5N, EtOH, RT, 36 h;
(iii) SOCl2/dioxane, RT, 4 h; (iv) LiOH/MeOH, 1N, 5 h.
the conclusion that the Beckmann rearrangement of the
B-ring of the steroid 4 was hindered by the presence of the
D-lactamic ring either by intermolecular or intramolecular
effects.
chloride was significantly superior for the protection of the
D-lactamic nitrogen.
Additionally, the N-benzoyl protected lactam 5 was suc-
cessfully converted to its oxime 6 under mild conditions
(room temperature), avoiding deprotection of the lactamic
nitrogen, as was observed when the reaction mixture was
refluxed (Fig. 2).
Beckmann rearrangement of the oxime 6 to 7 was success-
fully achieved in a yield of 45%. This result consolidated the
hypothesis that the presence of the unprotected D-lactamic
nitrogen hindered the formation of the B-lactamic ring, while
the Beckmann rearrangement proved to be superior to the
Schmidt reaction reported previously. Apparently, protec-
tion of the lactamic nitrogen altered the physicochemical na-
ture of the molecule, resulting in a diminution of the effects
caused by the presence of the 17N–H in the rearrangement
of the B-steroidal ring.
In order to investigate the possible influence of the
D-lactamic ring on the rearrangement of 6-oxime, it was
decided that the nitrogen atom of the D-lactamic ring of 4
had to be protected. We initially tried to protect this atom
using already known procedures, such as Ac2O/C5H5N,
NaH/Ac2O/DMAP, TBDMSCl/NEt3 [17] (Fig. 5) but all
of these efforts proved fruitless due to the existence of the
19-CH3 and the weak nucleophilic nature of the lactamic
nitrogen. The only compound observed in TLC analysis
during the reaction was isolated and identified by NMR and
IR analysis as the initial steroidal skeleton. Based on these
facts we assumed that the N-protection was not attainable
when drastic conditions were employed.
From this point of view, five acyl chlorides were used in
order to determine their reactivity towards the nitrogen atom
of the D-lactamic ring. As shown in Fig. 5, acetyl chloride,
pivaloyl chloride, phenyl acetyl chloride, and cinammoyl
chloride failed to protect the lactamic nitrogen (1–5% yield)
in contrast to benzoyl chloride, which gave the desired prod-
uct in good yield (90%). Consequently, the use of benzoyl
The nature of these effects, inter- or intramolecular, was
figured out by one control experiment where an equimolar
amount of a known unprotected lactam (3␤-acetoxy-17␣-
aza-D-homo-5-androsten-17-one) was added in the reaction
mixture of the Beckmann rearrangement of oxime 6. The
product obtained proved identical to bilactam 7, while the
yield of the Beckmann rearrangement was not significantly

664
A.I. Koutsourea et al. / Steroids 68 (2003) 659–666
suggests a new approach in which the D-ring of the
O
O
O
molecule was converted to the corresponding lactamic
ring in the presence of the B-lactam. The readily available
3␤-hydroxy-5-androsten-17-one was used as starting mate-
rial. It was initially converted to the acetoxy derivative 9 in
which the 17-keto group was protected as the 17-ketal 10
in order to shield it from the following synthetic steps.
The oxidation methods described in the literature for
the allylic oxidation of ketal 10 to 11 include the use of
CrO3/TBHP [18], N-hydroxy pthalimide/1,1ꢀ-azobis (cy-
clohexane carbonitrile) [19], PCC/CrO3 Py2 or tert-butyl
chromate [20], 3,5-dimethylpyrazolium fluorochromate
[21], sodium periodate/TBHP [22], and CrO3/3,5 DMP
[23,24]. Among them, CrO3/3,5 DMP proved to be the
most effective as the risk of deprotection of the 17-ketal
was minimized since the reaction was carried out at a low
temperature (−10 to −20 ◦C), and it produced the desired
ketone 11 in 78% yield.
i
ii
AcO
AcO
AcO
AcO
10
9
O
O
O
O
iii
O
N
HO
11
12
iv
O
O
O
v
N
O
H
N
O
H
The corresponding oxime 12 was prepared under mild
conditions (stirring at room temperature), knowing the sen-
sitivity of the ketal moiety to reflux conditions.
AcO
AcO
13
vi
14
N OH
The crucial step in this procedure was the rearrangement
of 3␤-acetate-5-androsten-7-one-17,17-ethylendioxy oxime
(12) to the B-lactamic derivative. The classic conditions
described in the literature [25,26] for the Beckmann rear-
rangement of the B-ring of the steroidal skeleton involved
dropwise addition of freshly distilled SOCl2 in dioxane and
stirring at room temperature. When we tried to perform
these conditions for the formation of the B-lactam described
herein, we encountered major difficulties. Primarily, the
∼30% yield was lower than expected. Additionally, we ob-
served the formation of a significant amount of undesirable
by-products, and purification of the desired lactam was
difficult. In order to overcome this problem, we decided to
lower the temperature to 0 ◦C (to minimize the risk of the
side reaction product formation) and to change the reaction
solvent. THF was the most convenient solvent in this case
because of its low freezing point, its capacity to dilute the
oxime, and its inactivity. These conditions proved very ef-
fective for the Beckmann rearrangement of the B-ring of
the oxime 12 in contrast to the classic conditions [25,26],
as the yield of the reaction was significantly increased to
67%, and formation of the undesirable by-products was
subsequently reduced.
CN
N
O
H
N H
O
AcO
AcO
vii
15
16
H
N
H
N
O
O
viii
N H
N H
O
AcO
HO
O
1
8
Fig. 3. New procedure for the synthesis of B,D-bilactam. Conditions:
(i) p-TsOH/TEOF/ethylene glycol, 90 ◦C, 1 h; (ii) CrO3/3,5-DMP/dry
dichloromethane, −20 ◦C, 5 h; (iii) H2NOH·HCl/C5H5N, EtOH, RT,
10 h; (iv) SOCl2/THF, 0 ◦C, 8 h; (v) aq. HCl, RT, 1 h 30 min; (vi)
H2NOH·HCl/C5H5N, EtOH, reflux, 2 h; (vii) SOCl2/dioxane, RT, 5 h;
(viii) LiOH/MeOH, 1N, RT, 1 h.
altered (43%). Obviously the external unprotected lactam
does not hinder this rearrangement, leading us to conclude
that the formation of the B-lactamic ring is mainly hindered
by intermolecular effects due to the existence of the 17N–H
and not by intramolecular interactions between oxime 6 and
the external 17N–H moiety.
Deprotection of the lactamic amide 7 to the correspond-
ing alcohol 8 was accomplished using mild basic conditions.
A significant advantage was the concurrent deprotection of
the 3␤-acetoxy group. This resulted in omission of the hy-
drolysis step, which was necessary for the formation of the
3␤-hydroxy steroidal derivative in order to use it for further
functionalization with molecules of biological interest.
We describe herein another strategy (Fig. 3) devel-
oped for the synthesis of the target molecule 1. This
As shown in Fig. 3, Beckmann rearrangement of com-
pound 12 resulted in 13 as the main product (52%) and 14
as a second product (16%). The latter was easily hydrolyzed
to 13 with an equimolar quantity of aq. HCl solution in
methanol, increasing the final total yield to 67%. Ketone 13
was then converted to the corresponding D-lactam 1 using
the classic conditions described in the literature [14]. The
final B,D-bilactamic derivative 1 was hydrolyzed to 8 under
mild, basic conditions.
The B,D-bilactamic derivative 8 was the only product,
isolated and identified by NMR, IR, and elemental analysis,
in both synthetic procedures described. Therefore, it was

A.I. Koutsourea et al. / Steroids 68 (2003) 659–666
665
H
H
N
N
O
O
i
13.4%
N H
AcO
O
AcO
O
95%
ii
iii
iii
H
N OH
N
O
iii
AcO
N
AcO
N
HO
HO
Fig. 4. Rearrangement of the B-steroidal ring via Schmidt reaction. Conditions: (i) NaN3/PPA, 50–60 ◦C, 10 h; (ii) H2NOH·HCl/C5H5N, EtOH, reflux,
2 h; (iii) SOCl2/dioxane, RT, 1 h.
R
N
O
O
i or ii or iii
CH3CO-
= R
AcO
AcO
O
O
TBDMS-
H
N
R'
N
O
iv
CH 3CO-
(CH3)3CCO-
PhCH2CO-
AcO
O
= R'
O
PhCH=CHCO-
N
O
v
90%
AcO
O
Fig. 5. Synthetic attempts for the protection of the lactamic –NH– group. Conditions: (i) Ac2O/C5H5N, reflux, 20 h; (ii) NaH/Ac2O/DMAP, THF, 50 ◦C,
6 h; (iii) TBDMSCl/NEt3, RT, 17 h; (iv) corresponding chloride/dry benzene, reflux, 24–48 h; (v) Bz-Cl/dry benzene, reflux, 2 h.
concluded that only one of the oxime isomers was stable both
in the B and D rings. Due to steric and electronic effects,
this oxime was considered to be the syn-oxime in the case
of the B-ring and the anti-oxime in the case of the D-ring.
Consequently, the above-described synthetic approaches
for preparation of 1 via the Beckmann rearrangement
showed that the crucial step of the procedures was the
formation of the B-lactamic ring. Key reactions was ei-
ther the N-protection of the D-ring or the formation of the
B-lactamic ring as the initial step so as the negative impact
of the D-ring on Beckmann rearrangement of the B-ring
could be avoided.
yields (four- and six-fold increases, respectively) than the
reported Schmidt reaction. Additionally, these methods ren-
dered preparation of the B-lactamic ring attainable via the
previously restricted Beckmann rearrangement.
The biological evaluation of compound 1 with respect to
its antileukemic activity is currently under investigation.
Acknowledgements
This research was supported by the “K. Karatheodoris”
Grant No. 3013, Research Committee of the University of
Patras, Greece. The NMR spectra and the elemental analyses
were carried out in the Center of Instrumental Analysis of
Moreover, these two synthetic approaches enabled us to
prepare the desired B,D-bilactam 1 in significantly better

666
A.I. Koutsourea et al. / Steroids 68 (2003) 659–666
the University of Patras, and the authors are indebted to Dr.
D. Vachliotis for performing these measurements.
[12] Shoppee CW, Akhtar MI, Lack RE. Aza-steroids. Part VIII.
7␣-Aza-B-homo-5␣-cholestane and 7␣-aza-B-homocholest-5-ene. J
Chem Soc 1964:3392–5.
[13] Singh H, Padmanabhan S. 3,7␣-Diaza-A,B-bishomocholest-4␣-
ene-4,7-dione. Tetrahedron Lett 1967;38:3689–92.
[14] Regan BM, Hayes FN. 17 and 17␣-aza-D-homo-steroids. J Am Chem
Soc 1956;78:634–43.
[15] Salvador JAR, Sá e Melo ML, Neves ASC. Copper-catalyzed allylic
oxidation of ꢀ5-steroids by t-butyl hydroperoxide. Tetrahedron Lett
1997;38:119–22.
References
[1] Erlichman C, Loprinzi CL. Hormonal therapies. In: CANCER
principles and practice of oncology. 5th ed. Philadelphia: Lippincott-
Raven Publication; 1997. p. 395.
[2] Lupulescu AP. Hormones, vitamins and growth factors in cancer
treatment and prevention. Cancer 1996;78:2264–80.
[3] Wasan H, Waxman J. In: Pusztai LL, Lewis CE, Yap E, editors.
Cell proliferation cancer. Oxford: Oxford University Press; 1996.
p. 260–81.
[4] Trafalis DTP, Sambani C, Kapsimali V, Economidou J, Politis
G, Catsoulacos P. Effects of homo-aza-steroids on acute non-
lymphocytic leukaemia cell proliferation in vitro. Br J Haematol
1995;91:907–14.
[5] Wall ME, Abernethy Jr GS, Caroll FJ, Taylor DJ. The effect of
some steroidal alkylating agents on experimental animal mammary
tumor and leukaemia systems. J Med Chem 1969;12:810–8.
[6] Catsoulacos P, Politis D, Wampler GL. A new steroidal alkylating
agent with improved activity in advanced murine leukaemias. Cancer
Chemother Pharmacol 1979;3:67–70.
[7] Catsoulacos P, Wampler GL. Activity of 3␤-hydroxy-13␣-amino-
13,17-seco-5␣-androstan-17-oic-13,17-lactam-p-bis(2-chloroethyl)-
aminophenylacetate. Oncology 1982;39:109–12.
[8] Athanasiou K, Demopoulos N, Catsoulacos P. Chromosome damage
and SCE induced by the cytostatic factor homo-aza-steroidal ester of
p-bis(2-chloroethyl)amino phenyl acetic acid in CHO cells in culture.
Environ Mutagen 1983;5:279–83.
[9] Nikolaropoulos SS, Tsavdaridis D, Arsenou ES, Papageorgiou A,
Karaberis E, Mourelatos D. Synergistic antineoplastic and cytogenetic
effects by the combined action of two homo-aza-steroidal esters of
nitrogen mustards on P388 and L1210 leukemias in vivo and in
vitro. Anticancer Res 2000;20:2745–52.
[16] Fieser L. Preparation of ethylenethioketals.
1954;76:1945–7.
J Am Chem Soc
[17] Hiemstra H, Klaver WG, Speckamp WN. Regioselective synthesis
of ( )-gabaculine. Tetrahedron Lett 1986;27(12):1411.
[18] Muzart J. Synthesis of unsaturated carbonyl compounds via a
chromium-mediated allylic oxidation by 70% tert-butyl hydro-
peroxide. Tetrahedron Lett 1987;28:4665–8.
[19] Padma M, Thoden JB, Powell DR, Lardy HA. Steroidal allylic
fluorination using diethylaminosulfur trifluoride: a convenient method
for the synthesis of 3␤-acetoxy-7␣- and 7␤-fluoroandrost-5-en-17-
one. Steroids 1996;61:453–60.
[20] Wender PA, Eissenstat MA, Filosa MP. A general methology for
pseudoguaiane synthesis: total synthesis of ( )-damsinic acid and
(
)-confertin. JACS 1979;101(8):2196–8.
[21] Bora U, Chaudhuri MK, Dey D, Kalita D, Kharmawphlang
W, Mandal GC. 3,5-Dimethylpyrazolium fluorochromate (VI),
C5H8N2H[CrO3F], (Dmpz HFC): a convenient new reagent for the
oxidation of organic substrates. Tetrahedron 2001;57:2445–8.
[22] Marwah P, Marwah A, Kneer N, Lardy H. Ergosteroids IV:
synthesis and biological activity of steroid glucuronosides, ethers
and alkylcarbonates. Steroids 2001;66:581–95.
[23] Pouzar V, Slav´ıkovà T, Èerný I. Synthesis of (19E)-3␤,7␣-
dihydroxy-17-oxoandrost-5-en-19-al-19-(O-carboxymethyl)oxime, a
new hapten for 7␣-hydroxydehydroepiandrosterone (3␤,7␣-dihy-
droxy-androst-5-en-17-one). Steroids 1998;63:454–8.
[24] Salmond WG, Barta MA, Havens JL. Allylic oxidation with 3,5-
dimethylpyrazole chromium trioxide complex. Steroidal ꢀ5-7-
ketones. JOC 1978;43(10):2057–9.
[25] Singh H, Padmanabhan S. Steroids. Part X. Some 5-en-7-ones and
7␣-aza-B-homo-5-en-7-ones in spirostan, androstan and pregnane
series. Indian J Chem 1969;7:1084–7.
[10] Athanasiou C, Catsoulacos P, Papageorgiou A, Athanasiou K.
Response of transplantable tumors in mice of macromolecular
synthesis to 17␤-acetamido-3-aza-A-homo-4␣-androsten-4-one.
Cancer Invest 1987;5:301–7.
[11] Singh H, Gupta SK, Padmanablan DP, Bhardwaj TR. Steroids
and related studies. Part XXXVI. 7␣-17␣-Diaza-7,17-dioxo-B,D-
dihomo-5-androsten-3␤-yl acetate. Indian J Chem 1977;15B:101–2.
[26] Gawley RE. The Beckmann reactions: rearrangements, elimination–
additions, fragmentations and rearrangement–cyclizations. Org React
1988;35:1–420.