Insights into the synthesis of steroidal A-ring olefins

Article abstract of DOI:10.1002/hlca.201300082

The classical synthesis, followed by purification of the steroidal A-ring Δ¹-olefin, 5α-androst-1-en-17-one (5), from the Δ¹-3-keto enone, (5α,17β)-3-oxo-5-androst-1-en-17-yl acetate (1), through a strategy involving the reaction of Δ¹- 3-hydroxy allylic alcohol, 3β-hydroxy-5α-androst-1-en-17β-yl acetate (2), with SOCl₂, was revisited in order to prepare and biologically evaluate 5 as aromatase inhibitor for breast cancer treatment. Surprisingly, the followed strategy also afforded the isomeric Δ²-olefin 6 as a by-product, which could only be detected on the basis of NMR analysis. Optimization of the purification and detection procedures allowed us to reach 96% purity required for biological assays of compound 5. The same synthetic strategy was applied, using the Δ⁴-3-keto enone, 3-oxoandrost-4-en-17β-yl acetate (8), as starting material, to prepare the potent aromatase inhibitor Δ⁴-olefin, androst-4-en-17-one (15). Unexpectedly, a different aromatase inhibitor, the Δ^3,5-diene, androst-3,5-dien-17-one (12), was formed. To overcome this drawback, another strategy was developed for the preparation of 15 from 8. The data now presented show the unequal reactivity of the two steroidal A-ring Δ¹- and Δ⁴-3- hydroxy allylic alcohol intermediates, 3β-hydroxy-5α-androst-1-en- 17β-yl acetate (2) and 3β-hydroxyandrost-4-en-17β-yl acetate (9), towards SOCl₂, and provides a new strategy for the preparation of the aromatase inhibitor 12. Additionally, a new pathway to prepare compound 15 was achieved, which avoids the formation of undesirable by-products. Copyright

Full text of DOI:10.1002/hlca.201300082

Helvetica Chimica Acta – Vol. 97 (2014)
39
Insights into the Synthesis of Steroidal A-Ring Olefins
by Carla L. Varelaa), Fernanda M. F. Roleira*a), Saul C. P. Costaa), Alexandra S. C. T. Pintoa), Ana I. O.
S. Martinsa), Rui A. Carvalhob), Nate´rcia A. Teixeirac)d), Georgina Correia-da-Silvac)d),
and Elisia´rio Tavares-da-Silva*a)
a) CEF, Center for Pharmaceutical Studies & Pharmaceutical Chemistry Group, Faculty of Pharmacy,
University of Coimbra, PT-3000-548 Coimbra
(phone: þ 351239488400; fax: þ 351239488503; e-mail: froleira@ff.uc.pt; etavares@ff.uc.pt)
b) Department of Life Sciences, Faculty of Science and Technology & Centre for Neuroscience and Cell
Biology (CNC), University of Coimbra, PT-3001-401 Coimbra
c) Biochemistry Laboratory, Faculty of Pharmacy, University of Porto, Rua Jorge Viterbo Ferreira, 228,
PT-4050-313 Porto
d) Institute for Molecular and Cellular Biology (IBMC), University of Porto, Rua do Campo Alegre, 823,
PT-4150-180 Porto
The classical synthesis, followed by purification of the steroidal A-ring D1-olefin, 5a-androst-1-en-17-
one (5), from the D1-3-keto enone, (5a,17b)-3-oxo-5-androst-1-en-17-yl acetate (1), through a strategy
involving the reaction of D1-3-hydroxy allylic alcohol, 3b-hydroxy-5a-androst-1-en-17b-yl acetate (2),
with SOCl2, was revisited in order to prepare and biologically evaluate 5 as aromatase inhibitor for breast
cancer treatment. Surprisingly, the followed strategy also afforded the isomeric D2-olefin 6 as a by-
product, which could only be detected on the basis of NMR analysis. Optimization of the purification and
detection procedures allowed us to reach 96% purity required for biological assays of compound 5.
The same synthetic strategy was applied, using the D4-3-keto enone, 3-oxoandrost-4-en-17b-yl acetate
(8), as starting material, to prepare the potent aromatase inhibitor D4-olefin, androst-4-en-17-one (15).
Unexpectedly, a different aromatase inhibitor, the D3,5-diene, androst-3,5-dien-17-one (12), was formed.
To overcome this drawback, another strategy was developed for the preparation of 15 from 8. The data
now presented show the unequal reactivity of the two steroidal A-ring D1- and D4-3-hydroxy allylic
alcohol intermediates, 3b-hydroxy-5a-androst-1-en-17b-yl acetate (2) and 3b-hydroxyandrost-4-en-17b-
yl acetate (9), towards SOCl2, and provides a new strategy for the preparation of the aromatase inhibitor
12. Additionally, a new pathway to prepare compound 15 was achieved, which avoids the formation of
undesirable by-products.
Introduction. – Among several biological activities, steroidal A-ring olefins, with a
C(17)¼O group, particularly D3- and D4-olefins (Fig.) [1][2], were shown to be
interesting aromatase inhibitors (AIs). AIs block the biosynthesis of estrogens and
offer a therapeutic alternative for the treatment of estrogen-dependent cancers, namely
breast cancer [3 – 5]. This has been attributed to two different factors: the planarity
conferred by the C¼C bonds, which allows better fitting the enzyme receptor core, and
the H-bonding capacity of the C(17)¼O O-atom to a receptor residue [1][6][7].
As part of a project on new structureꢀactivity relationships (SAR) of steroidal AIs
[1][6][7], we are now interested in synthesizing D1- and D4-olefins 5 and 15,
respectively (Schemes 1, 2, and 3). Concerning 5, there are very few references
describing its preparation, and the most complete one is a 50 years old reporting the
ꢀ 2014 Verlag Helvetica Chimica Acta AG, Zꢁrich

40
Helvetica Chimica Acta – Vol. 97 (2014)
Figure Steroidal D3- and D4-olefin aromatase inhibitors
synthesis using the D1-3-keto enone 1 as starting material, and the allylic alcohol 2 as
intermediate [8]. In a recent work, we revisited this synthesis [7] and, in the present
work, we further explored new insights into the referred strategy. In this study, along
with the previously reported D1-olefin 5, the D2-isomer 6 was identified as a by-product
(Scheme 1).
Scheme 1. Synthesis of (5a)-Androst-1-en-17-one (5) from (5a,17b)-3-Oxoandrost-1-en-17-yl Acetate
(1)
i) Li(t-BuO)3AlH, anh. THF, reflux, 3 h; 94%. ii) SOCl2, benzene, 5 – 88, 1 – 2 h. iii) LiAlH4, Et2O, reflux,
11 h. iv) CrO3, H2SO4, acetone, 08; 73% of 5/6.
Furthermore, to achieve an alternative synthetic strategy for the preparation of the
D4-olefin 15 [2][7], that avoids the formation of the D4-17b-hydroxylated derivative as
well as the D3-olefin as by-products, we applied the same methodology using the D4-3-
keto enone 8 as starting material. Unexpectedly, in this case, instead of the desired 15,
another AI, the D3,5-diene 12 [9 – 16], was obtained as the only product (Scheme 2). To
overcome this problem, another synthetic strategy, using also 8 as starting material, was
developed for the preparation of 15 (Scheme 3).

Helvetica Chimica Acta – Vol. 97 (2014)
41
Scheme 2. Formation of Androsta-3,5-dien-17-one (12) from Testosterone (7)
i) Ac2O, dry pyridine, r.t., 21 h 25 min; 84%. ii) Li(t-BuO)3AlH, anh. THF, reflux, 3 h 30 min. iii) SOCl2,
benzene, 5 – 88, 5 h 30 min; 56%. iv) LiAlH4, Et2O, reflux, 8 h; 91%. v) CrO3, pyridine, r.t., 19 h; 38%.
Scheme 3. Preparation of Androst-4-en-17-one (15) from (17b)-3-Oxoandrost-4-en-17-yl Acetate (8)
i) NaBH4, CF3COOH, AcOH, MeCN, CH2Cl2, r.t., 3 h 30 min; 99%. ii) KOH, H2O, dioxane, r.t., 2 – 3 d;
99%. iii) CrO3, H2SO4, H2O, acetone, 08, 5 min; 75%.
1H-NMR Spectroscopy turned out to be the most adequate technique for the
detection of the D2-olefin 6, as well as of its precursors, and the 2D-COSY experiment
allowed us to unequivocally identify the precursor 10 of the D3,5-diene 12.
Results and Discussion. – The D1-olefin 5 was prepared as described in [7][8]
(Scheme 1). Briefly, reduction of enone 1 gave allylic alcohol 2, which afforded an

42
Helvetica Chimica Acta – Vol. 97 (2014)
untractable crude (TLC, NMR, and LC/MS control) after treatment with SOCl2.
Reaction of this crude with LiAlH4 and conventional workup furnished a mixture (one
TLC spot) of the isomers 3 and 4 in similar amounts (NMR analysis). To isolate the
desired compound 3, the isomeric mixture was subjected to column chromatography
using neutral alumina and hexane/CH2Cl2. However, the isolated fractions, subjected to
1H-NMR analysis, always turned out to be mixtures 3/4 with variable compositions
(from 1:3 to 9 :1). Given the impossibility of obtaining the pure compound 3, the next
oxidative step was performed with the the 9 :1 mixture of 3/4 using Jones reagent,
which allowed, after a laborious column chromatography purification process of the
obtained crude, isolation of the D1-olefin 5 in the required purity for further biological
studies (96% by LC/MS control) (Scheme 1).
Although the formation of the D2-isomer 6 has not been reported before, our results
revealed that migration of the C¼C bond from the C(1) to the more stable C(2)
occurred to a considerable extent. In fact, the D1- and D2-isomers 5 and 6, respectively,
possessed similar physico-chemical properties, exhibiting the same Rf values with
several chromatography solvents and similar crystallization conditions. Accordingly, it
is very difficult to distinguish the two isomers. Facing these difficulties, the complete
diagnosis of the C¼C bond position in this kind of compounds could only be achieved
by 1H-NMR spectroscopy. The D1-isomer 5 presents two signals at 5.52 ppm (HꢀC(2))
and 5.83 ppm (HꢀC(1)) for the olefinic H-atoms, whereas the D2-isomer 6 displays only
one common typical multiplet around 5.9 ppm for both olefinic H-atoms, allowing the
accurate identification of both compounds 5 and 6. Therefore, the NMR analysis
appears to be the most adequate technique to disclose the nature of the compounds
isolated after column chromatography.
To obtain the D4-olefin 15, the above-mentioned synthetic strategy was further
applied to the D4-3-keto enone 8 (prepared from 7 [7]) as starting material (Scheme 2).
In this case, instead of 15 we obtained the D3,5-diene 12. Treatment of 8 with Li(t-
BuO)3AlH, under the conditions previously described, led to D4-3b-hydroxy allylic
alcohol 9 with traces of its 3a-isomer. Treatment of this compound with SOCl2 in
benzene, contrarily to what was expected, gave the D3,5-diene precursor 10. Then, 10
was treated with LiAlH4 to afford the D3,5-diene derivative 11, which, after oxidation
with CrO3 in pyridine [10], led to 12. This approach opens a new way to prepare
compound 12, which is also an AI [9].
As the 1H-NMR did not allow unequivocal elucidation of the position of the diene
C¼C bonds in the A/B-ring system, a 2D-COSYanalysis of 10 was performed. The most
relevant signals in the COSY spectrum were those of the three olefinic H-atoms at 5.9,
5.6, and 5.4 ppm, of the HaꢀC(17) at 4.5 ppm, and of the Me(18) and Me(19) groups at
0.82 and 0.95 ppm, respectively. The HaꢀC(17) correlates with Me(18) (3J) which
resonates at 0.82 ppm; therefore, the resonance at 0.95 ppm is due to Me(19). Focusing
on the three olefinic H-atoms, of the D3,5-diene isomer 10, the HꢀC(4) will correlate
strongly with the HꢀC(3) and weakly with the HꢀC(6). In fact, we observed that an
olefinic H-atom absorbing at 5.9 ppm strongly correlates with an olefinic H-atom at
5.6 ppm (3J); the same olefinic H-atom (at 5.9 ppm) has a very weak correlation with an
olefinic H-atom absorbing at 5.4 ppm (4J), and there is no observable correlation
between the olefinic H-atom signals at 5.6 ppm and 5.4 ppm (5J). Therefore, the signal

Helvetica Chimica Acta – Vol. 97 (2014)
43
at 5.6 ppm corresponds to HꢀC(3), the signal at 5.9 ppm to HꢀC(4), and the signal at
5.4 ppm to HꢀC(6), which is in agreement with the structure of compound 10.
To obtain the D4-olefin 15, an alternative synthetic strategy was developed from the
same starting material 8 (Scheme 3). In this case, reaction of 8 with a mixture of NaBH4
in CF3COOH (TFA), glacial AcOH, and MeCN was performed in a controlled
environment [17], to yield compound 13 in 99% yield. This compound was then
submitted to a base-catalyzed hydrolysis to give quantitatively compound 14, which was
then subjected to Jones oxidation to furnish the desired 15 in 75% yield. The above-
mentioned strategy avoided the formation of the D4-17b-hydroxylated derivative as
well as the D3-olefin isomer as by-products.
In summary, the preparation of D1-olefin 5 from the D1-3-keto enone 1 according to
the revisited protocol takes place with the formation of the D2-isomer 6 as a by-product,
which can only be detected by NMR analysis. However, an adequate sequential
purification procedure by column chromatography, assisted by 1H-NMR control of the
separated fractions, allows the isolation of the AI 5 in 96% purity, which is an adequate
purity for compounds to proceed to biochemical assays.
The potent AI D4-olefin 15 cannot be obtained from the D4-3-keto enone 8 by the
reported strategy. Instead, another important aromatase inhibitor, the D3,5-diene 12, is
formed. This achievement offers a new synthetic way to compound 12. From these data,
presented for the first time, to the best of our knowledge, it is possible to establish the
unequal reactivity of steroidal D1- and D4-3-hydroxy allylic alcohol intermediates 2 and
9, respectively, towards the elimination reactions by treatment with SOCl2 (Schemes 1
and 2). Indeed, while the D1-isomer 2 renders the mixture of olefins 3/4 by the expected
elimination process, followed by isomerization of the C¼C bond, the D4-isomer 9
embarks in an additional elimination of a HꢀC(6) H-atom, followed by the migration
of the C¼C bond with the formation of the D3,5-diene derivative 10.
The preparation of the AI 15 can be achieved by a new method, using also the D4-3-
keto enone 8 as starting material, by a sequence involving the reduction at C(3),
followed by the hydrolysis of the AcO group at C(17) of 13, and subsequent oxidation
of the resulting OH group, which precludes the formation of undesirable by-products.
Experimental Part
General. Testosterone (7) was purchased from Pharmacia & Upjohn Company, Kalamazoo,
Michigan (USA), and (5a,17b)-3-oxoandrost-1-en-17-yl acetate (1) and (5a)-androst-2-en-17-one (6)
were purchased from Steraloids, Inc. (Newport RI, USA). Other reagents and solvents were used as
obtained from the suppliers without further purification, with the exception of CH2Cl2, which was dried
through reflux and distilled from CaH2 [18]. M.p.: Reichert Thermopan hot-block apparatus;
uncorrected. IR Spectra: Jasco 420FT/IR spectrometer. 1H- and 13C-NMR spectra: Varian 600 MHz
spectrometer, using a 3-mm broadband NMR probe; chemical shifts in ppm downfield from TMS used as
an internal standard; all J values in Hz. ESI- and LC-MS: mass spectrometer QIT-MS Thermo Finningan,
model LCQ Advantage MAX, coupled to a liquid chromatograph of high performance Thermo
Finningan (column: C18; reversed phase (RP); H2O/MeCN 40 :60).
(3b,5a,17b)-3-Hydroxyandrost-1-en-17-yl Acetate (2). See [7].
(5a,17b)-Androst-1-en-17-ol (3) and (5a,17b)-Androst-2-en-17-ol (4). Prepared as described in [7].
1H-NMR Analysis of the obtained crude product revealed a mixture 3/4 1:1. The crude product was
purified by column chromatography (CC) (neutral Al2O3; hexane/CH2Cl2 80 :20) to give a white solid
(one TLC spot). 1H-NMR Analysis of this solid indicated an enriched mixture of 3 (90%) with 4 (10%).

44
Helvetica Chimica Acta – Vol. 97 (2014)
Data of 3. 1H-NMR ((D6)DMSO; selected signals): see [7].
1
Data of 4. H-NMR ((D6)DMSO; selected signals): 0.63 (s, Me(18)); 0.71 (s, Me(19)); 3.42 (ddd,
J(17a,16a) ¼ 9.0, J(17a,16b) ¼ 9.0, J(17a,OH) ¼ 5.0, HaꢀC(17)); 4.42 (d, J(OH,17a) ¼ 5.0, HObꢀC(17));
5.53 – 5.59 (m, HꢀC(2), HꢀC(3)).
(5a)-Androst-1-en-17-one (5) and (5a)-Androst-2-en-17-one (6). Prepared as described in [7]. The
obtained crude product was crystallized from MeOH/H2O to give white crystals (one TLC spot).
1H-NMR Analysis of these crystals revealed the presence of a mixture 5 (83%)/6 (17%). Purification by
CC (hexane/Et2O), followed by consecutive recrystallizations from MeOH, gave 5 in 96% purity (LC/
MS analysis) with a small amount of isomer 6.
Data of 5. 1H-NMR (CDCl3): see [7].
1
Data of 6. H-NMR (CDCl3; selected signals): 0.78 (s, Me(19)); 0.87 (s, Me(18)); 5.55 – 5.62 (m,
HꢀC(2), HꢀC(3)).
(17b)-3-Oxoandrost-4-en-17-yl Acetate (8). To a soln. of 7 (2.0 g, 6.93 mmol) in dry pyridine (48 ml),
Ac2O (7.9 ml, 83.9 mmol) was added, and the mixture was stirred for 21 h 25 min at r.t. (208), until all the
starting material was consumed (TLC control). Then, CH2Cl2 (250 ml) was added, and the org. layer was
washed with 10% NaHCO3 (3 ꢁ 150 ml), 10% HCl (3 ꢁ 150 ml), and H2O (3 ꢁ 150 ml), dried (anh.
MgSO4), filtered, and concentrated to dryness. Crystallization of the obtained residue from AcOEt gave
pure 8 (1.92 g, 84%). M.p.: 141 – 1428 ([19]: 139 – 1408). IR (CHCl3): 3018 (¼CH), 1736 (C¼O), 1675
(C¼C), 1248 (CꢀO). 1H-NMR (CDCl3; selected signals): 0.82 (s, Me(18)); 1.18 (s, Me(19)); 2.03 (s,
MeCOO); 4.58 (dd, J(17a,16a) ¼ 9.0, J(17a,16b) ¼ 8.0, HaꢀC(17)); 5.71 (s, HꢀC(4)). 13C-NMR
(150 MHz, CDCl3): 12.0 (C(18)); 17.4 (C(19)); 20.5; 21.1; 23.4; 27.4; 31.4; 32.7; 33.9; 35.4; 35.7; 36.6;
38.6; 42.4; 50.2; 53.7; 82.4 (C(17)); 123.9 (C(4)); 170.9 (C(5)); 171.1 (OC¼O); 199.4 (C(3)).
(3b,17b)-3-Hydroxyandrost-4-en-17-yl Acetate (9). To a soln. of 8 (2.0 g, 6.05 mmol) in anh. THF
(75 ml) under N2, Li(t-BuO)3AlH (2.0 g, 7.86 mmol) was added, and the mixture was heated under reflux
for 2 h, then an excess of Li(t-BuO)3AlH (500.1 mg, 1.97 mmol) was added. The reaction proceeded until
complete transformation of the starting material (31ꢂ2 h). After removal of the solvent under vacuum,
H2O (200 ml) was added, and the aq. layer was extracted with CH2Cl2 (3 ꢁ 200 ml). The org. layer was
then washed with H2O (200 ml), dried (anh. Na2SO4), filtered, and concentrated to dryness to give 1.97 g
1
of a crude material mainly composed of 9. H-NMR ((D6)DMSO; selected signals): 0.76 (s, Me(18));
0.99 (s, Me(19)); 1.98 (s, MeCOO); 3.88 – 3.92 (m, HaꢀC(3)); 4.49 (dd, J(17a,16a) ¼ 9.0, J(17a,16b) ¼
8.0, HaꢀC(17)); 4.54 (d, J(OH,3a) ¼ 5.5, HObꢀC(3)); 5.19 (br. s, HꢀC(4)). 13C-NMR (150 MHz,
(D6)DMSO): 11.8 (C(18)); 18.4 (C(19)); 20.0; 20.8; 22.9; 27.0; 28.9; 31.4; 32.2; 35.1; 35.2; 36.2; 36.7; 41.9;
49.7; 53.8; 65.8 (C(3)); 81.8 (C(17)); 125.5 (C(4)); 143.9 (C(5)); 170.2 (OC¼O).
(17b)-Androsta-3,5-dien-17-yl Acetate (10). A soln. of crude 9 (500.4 mg) in benzene (10 ml) was
kept at 5 – 88 under N2, treated with SOCl2 (0.5 ml, 6.72 mmol), and the mixture was stirred for 2 h
25 min. Then, an excess of SOCl2 (0.1 ml, 1.38 mmol) was added. The reaction did not proceed to the
complete transformation of the starting material (51ꢂ2 h). Benzene was evaporated under vacuum at r.t.
giving an oily residue, to which solid NaHCO3 (500 mg) was added, followed by 10% NaHCO3 (100 ml).
The aq. layer was extracted with CH2Cl2 (3 ꢁ 100 ml), and the resulting org. layer was washed with H2O
(3 ꢁ 100 ml), dried (anh. MgSO4), filtered, and concentrated to dryness to give a white solid residue. This
residue was purified by CC (silica gel 60; PE (60 – 808)/AcOEt) to afford 263.4 mg of 10 as a white
crystalline residue in an overall yield of 56% from 8. Recrystallization from PE 60 – 808/AcOEt. M.p.
1
116 – 1198. IR (CHCl3): 3018 (¼CH), 1736 (C¼O ester), 1648 (C¼C), 1244 (CꢀO). H-NMR (CDCl3;
selected signals): 0.83 (s, Me(18)); 0.96 (s, Me(19)); 2.04 (s, MeCOO); 4.61 (dd, J(17a,16a) ¼ 9.0,
J(17a,16b) ¼ 8.0, HaꢀC(17)); 5.37 – 5.38 (m, HꢀC(6)); 5.58 – 5.60 (m, HꢀC(3)); 5.91 – 5.93 (m, HꢀC(4)).
13C-NMR (150 MHz, CDCl3): 12.0 (C(18)); 18.8 (C(19)); 20.4; 21.2; 22.9; 23.5; 27.5; 31.3; 31.6; 33.7; 35.2;
36.8; 42.5; 48.3; 51.2; 82.8 (C(17)); 122.6 (C(6)); 125.1 (C(3)); 128.8 (C(4)); 141.5 (C(5)); 171.2 (OC¼O).
ESI-MS: 315.1 (76, [M þ H]þ).
(17b)-Androsta-3,5-dien-17-ol (11). To a soln. of 10 (100.8 mg, 0.32 mmol) in Et2O (15 ml), LiAlH4
(76.2 mg, 2.01 mmol) was added cautiously under N2, and the mixture was heated under reflux for 8 h.
Then, a sat. soln. of sodium potassium tartrate (150 ml) was added, and the mixture was extracted with
Et2O (4 ꢁ 100 ml). The org. layer was then washed with H2O (4 ꢁ 100 ml), dried (anh. MgSO4), filtered,
and concentrated to dryness to give pure 11 (79.4 mg, 91%). White solid. Recrystallization from AcOEt/

Helvetica Chimica Acta – Vol. 97 (2014)
45
hexane. M.p. 140 – 1428. IR (CHCl3): 3302 (OH), 3021 (¼CH), 1646 (C¼C), 1054 (CꢀO). 1H-NMR
((D6)DMSO; selected signals): 0.67 (s, Me(18)); 0.89 (s, Me(19)); 3.45 (ddd, J(17a,OH) ¼ 5.0,
J(17a,16a) ¼ 9.0, J(17a,16b) ¼ 9.0, HaꢀC(17)); 4.45 (d, J(OH,17a) ¼ 5.0, HObꢀC(17)); 5.33 – 5.35 (m,
HꢀC(6)); 5.56 – 5.58 (m, HꢀC(3)); 5.87 – 5.89 (m, HꢀC(4)). 13C-NMR (150 MHz, (D6)DMSO): 11.2
(C(18)); 18.5 (C(9)); 20.1; 22.4; 22.9; 29.8; 30.8; 31.4; 33.2; 34.6; 36.3; 42.3; 48.0; 50.9; 79.9 (C(17)); 122.6
(C(6)); 124.5 (C(3)); 128.8 (C(4)); 140.8 (C(5)). ESI-MS: 271.2 (100, [M ꢀ H]þ).
Androsta-3,5-dien-17-one (12). To a soln. of 11 (62.0 mg, 0.23 mmol) in pyridine (3 ml), a pyridine
soln. (2.3 ml) of CrO3 (98.0 mg, 0.98 mmol) was added at 08. The mixture was stirred at r.t. for 19 h until
total transformation of the starting material (TLC control). The mixture was then diluted with Et2O
(150 ml) and poured into H2O (50 ml). The org. phase was washed with brine (6 ꢁ 150 ml) and H2O (3 ꢁ
200 ml), dried (anh. MgSO4), filtered, and concentrated to dryness giving a yellow residue which was
purified by CC (neutral Al2O3; PE (40 – 608)) to furnish pure 12 (9.0 mg, 38%). M.p. 81 – 838 ([20]: 80 –
1
828). IR (CHCl3): 3018 (¼CH), 1739 (C¼O), 1652 (C¼C). H-NMR (CDCl3; selected signals): 0.91 (s,
Me(19)); 0.97 (s, Me(18)); 5.39 – 5.41 (m, HꢀC(6)); 5.59 – 5.62 (m, HꢀC(3)); 5.92 – 5.94 (m, HꢀC(4)).
13C-NMR (150 MHz, CDCl3): 13.7 (C(19)); 18.8 (C(18)); 20.2; 21.8; 22.9; 30.6; 31.3; 31.4; 33.7; 35.3; 35.8;
47.7; 48.5; 51.9; 122.1 (C(6)); 125.3 (C(3)); 128.7 (C(4)); 141.6 (C(5)); 221.0 (C(17)). ESI-MS: 269.1 (99,
[M ꢀ H]þ).
(17b)-Androst-4-en-17-yl Acetate (13). NaBH4 (566.2 mg, 14.97 mmol) was added in small portions
under stirring and cooling to a previously cooled mixture of CF3COOH (3.5 ml), glacial AcOH (3.5 ml),
and MeCN (3.5 ml). A soln. of 8 (1.0 g, 3.03 mmol) in dry CH2Cl2 (18 ml) was added to the mixture.
Then, the mixture was let to react at r.t. under magnetic stirring and N2, until consumption of all the
starting material (31ꢂ2 h; TLC control). The mixture was then neutralized with a soln. of 10% NaHCO3
and extracted with CH2Cl2 (4 ꢁ 100 ml). The org. layer was washed with H2O (4 ꢁ 100 ml), dried (anh.
MgSO4), filtered, and concentrated to dryness to give 13 (945.0 mg, 99%). White solid. Recrystallization
from CH2Cl2/hexane/EtOH. M.p. 95 – 998. IR (CHCl3): 3024 (¼CH), 1737 (C¼O), 1663 (C¼C), 1043
1
(CꢀO). H-NMR (CDCl3; selected signals): 0.80 (s, Me(18)); 1.01 (s, Me(19)); 2.03 (s, MeCOO); 4.58
(dd, J(17a,16a) ¼ 9, J(17a,16b) ¼ 8, HaꢀC(17)); 5.29 – 5.30 (m, HꢀC(4)). 13C-NMR (150 MHz, CDCl3):
12.0 (C(18)); 19.2 (C(19)); 19.4; 20.9; 21.2; 23.5; 25.7; 32.4; 32.8; 35.8; 36.9; 37.1; 37.8; 42.5; 50.5; 54.4; 82.8
(C(17)); 119.3 (C(4)); 144.7 (C(5)); 171.2 (OC¼O).
(17b)-Androst-4-en-17-ol (14). Compound 13 (945.0 mg, 2.99 mmol) was added to a mixture
dioxane/H2O 85 :15 (90 ml) with 2% NaOH (18 ml), at r.t. The mixture was let to react until total
transformation of the starting material (52 h; TLC control) and then neutralized with an aq. soln. of 5%
HCl. The dioxane was evaporated under vacuum leading to a white solid residue that was diluted with
H2O (200 ml) and extracted with AcOEt (4 ꢁ 100 ml). The org. layer was then washed with H2O (4 ꢁ
100 ml), dried (anh. MgSO4), filtered, and concentrated to dryness to afford 14 (819.0 mg, 99%).
White solid. Recrystallization from hexane/AcOEt. M.p. 143 – 1468. IR (CHCl3): 3278 (OH), 2959
(¼CH), 1651 (C¼C). 1H-NMR ((D6)DMSO): 0.65 (s, Me(18)); 0.97 (s, Me(19)); 3.43 (ddd, J(17a,16a) ¼
9.0, J(17a,16b) ¼ 9.0, J(17a,OH) ¼ 5.0, HaꢀC(17)); 4.43 (d, J(OH,17a) ¼ 5.0, HObꢀC(17)); 5.23 – 5.25
(m, HꢀC(4)). 13C-NMR (150 MHz, (D6)DMSO): 11.2 (C(18)); 18.8 (C(19)); 19.1; 20.6; 23.0; 25.1; 29.8;
31.9; 32.5; 35.5; 36.4; 36.5; 37.2; 42.4; 50.3; 54.1; 79.9 (C(17)); 118.7 (C(4)); 144.2 (C(5)).
Androst-4-en-17-one (15). Jones reagent (2.7 ml) was added dropwise to a soln. of 14 (839.8 mg,
3.06 mmol) in acetone/dioxane 60 :10 (190 ml), at 08 under magnetic stirring, until a persistent brown
coloration was obtained. Then, the excess of the oxidant was destroyed with the addition of iPrOH until
the soln. turned greenish. The dioxane and acetone were evaporated under vacuum. To the remaining
residue, H2O (200 ml) was added, and the aq. phase was extracted with AcOEt (4 ꢁ 100 ml). The org.
layer was then washed with 10% NaHCO3 (3 ꢁ 100 ml) and H2O (3 ꢁ 100 ml), dried (anh. MgSO4),
filtered, and concentrated to dryness to give a white solid residue (129.3 mg) after the addition of some
drops of Et2O. This residue was then purified by CC (SiO2; hexane/AcOEt 97:03) to afford pure 15
(624.2 mg, 75%). White crystalline solid. Recrystallization from hexane/AcOEt. M.p. 74 – 768. IR
(CHCl3): 3018 (¼CH), 1738 (C¼O), 1657 (C¼C). 1H-NMR (CDCl3; selected signals): 0.88 (s, Me(18));
1.03 (s, Me(19)); 5.31 – 5.33 (m, HꢀC(4)). 13C-NMR (150 MHz, CDCl3): 13.7 (C(18)); 19.2 (C(19)); 19.4;
20.6; 21.8; 25.7; 31.5; 32.1; 32.3; 35.5; 35.8; 37.1; 37.8; 47.7; 51.2; 54.5; 119.6 (C(4)); 144.3 (C(5)); 221.3
(C(17)). ESI-MS: 271.0 ([M ꢀ H]þ, 50%).

46
Helvetica Chimica Acta – Vol. 97 (2014)
ˆ
The authors thank FCT (FundaÅa˜o para a Ciencia e Tecnologia) for the strategic project PEst-OE/
SAU/UI0177/2011. Carla Varela also thanks FCT for a Ph.D. fellowship (SFRH/BD/44872/2008). We
also acknowledge the Rede Nacional de RMN for access to the facilities. The Varian VNMRS 600 MHz
spectrometer is part of the National NMR Network and was purchased with in the framework of the
National Programme for Scientific Re-equipment, contract REDE/1517/RMN/2005, with funds from
POCI 2010 (FEDER) and (FCT). The work was also funded by FEDER Funds through the Operational
ˆ
Competitiveness Program-COMPETE and by National Funds through FCT- FundaÅa˜o para a Ciencia e
Tecnologia under the project FCOMP-01 – 0124-FEDER-020970 (PTDC/QUI-BIQ/120319/2010). The
authors are also grateful to Laboratory of Mass Spectrometry (LEM) of the Node CEF/UC, integrated in
the National Mass Spectrometry Network (RNEM) of Portugal, for the MS analyses.
REFERENCES
[1] M. M. D. S. Cepa, E. J. Tavares da Silva, G. Correia-da-Silva, F. M. F. Roleira, N. A. A. Teixeira,
Steroids 2008, 73, 1409.
[2] M. Numazawa, A. Mutsumi, K. Hoshi, R. Koike, Biochem. Biophys. Res. Commun. 1989, 160, 1009.
[3] E. P. Winer, C. Hudis, H. J. Burstein, A. C. Wolff, K. I. Pritchard, J. N. Ingle, R. T. Chlebowski, R.
Gelber, S. B. Edge, J. Gralow, M. A. Cobleigh, E. P. Mamounas, L. J. Goldstein, T. J. Whelan, T. J.
Powles, J. Bryant, C. Perkins, J. Perotti, S. Braun, A. S. Langer, G. P. Browman, M. R. Somerfield, J.
Clin. Oncol. 2005, 23, 619.
[4] E. A. Perez, Oncologist 2006, 11, 1058.
[5] V. C. Jordan, A. M. H. Brodie, Steroids 2007, 72, 7.
[6] M. D. S. Cepa, E. J. Tavares da Silva, G. Correia-da-Silva, F. M. F. Roleira, N. A. A. Teixeira, J. Med.
Chem. 2005, 48, 6379.
[7] C. Varela, E. J. Tavares da Silva, C. Amaral, G. Correia da Silva, T. Batista, S. Alcaro, G. Costa,
R. A. Carvalho, N. A. A. Teixeira, F. M. F. Roleira, J. Med. Chem. 2012, 55, 3992.
[8] A. Bowers, A. D. Cross, J. A. Edwards, H. Carpio, M. C. Calzada, E. Denot, J. Med. Chem. 1963, 6,
156.
[9] M. Numazawa, A. Mutsumi, M. Tachibana, K. Hoshi, J. Med. Chem. 1994, 37, 2198.
[10] G. Phillipou, R. F. Seamark, Steroids 1975, 25, 673.
[11] T. Nambara, M. Kato¯, Chem. Pharm. Bull. 1965, 13, 1435.
[12] J. A. Boynton, J. R. Hanson, M. D. Liman, J. Chem. Res., Miniprint 1998, 1616.
[13] R. C. Cambie, C. M. Read, P. S. Rutledge, G. J. Walker, P. D. Woodgate, J. R. Hanson, J. Chem. Soc.,
Perkin Trans. 1 1980, 2581.
[14] J. R. Hanson, I. Kiran, J. Chem. Res., Synop. 1999, 594.
[15] H. Burrows, J. W. Cook, E. M. Roe, F. L. Warren, Biochem. J. 1937, 31, 950.
[16] S. Lieberman, K. Dobriner, B. R. Hill, L. F. Fieser, C. P. Rhoads, J. Biol. Chem. 1948, 172, 263.
[17] J. R. Hanson, P. B. Hitchcock, M. D. Liman, S. Nagaratnam, J. Chem. Soc., Perkin Trans. 1 1995,
2183.
[18] D. D. Perrin, W. L. F. Armarego, ꢃPurification of Laboratory Chemicalsꢄ, Pergamon Press, Oxford,
1988.
[19] W. Neudert, H. Rçpke, ꢃAtlas of Steroid Spectraꢄ, Springer-Verlag, Berlin, 1965.
[20] C. Djerassi, G. Rosenkranz, T. Romeo, US Pat. 2,698,854, 1955.
Received February 12, 2013