Synthesis of 3-methyl-3-hydroxy-6-oxo-androstane derivatives

Article abstract of DOI:10.1016/j.steroids.2004.04.007

3α,17β-Dihydroxy-3β-methyl-5α-androstan-6-one (1) and 3β,17β-dihydroxy-3α-methyl-5α-androstan-6-one (13) were prepared by the reaction of methylmagnesium bromide with the 3-ketosteroids. Structures and configurations in position 3 were determined by NMR s

Full text of DOI:10.1016/j.steroids.2004.04.007

Steroids 69 (2004) 605–612
Synthesis of 3-methyl-3-hydroxy-6-oxo-androstane derivativesଝ
∗
´
ˇ ´
´
Hana Chodounská, Vladimır Pouzar , Miloš Budešınský, Barbora Slavıková, Ladislav Kohout
Institute of Organic Chemistry and Biochemistry, Academy of Sciences of the Czech Republic, 166 10 Prague 6, Czech Republic
Received 24 February 2004; received in revised form 20 April 2004; accepted 23 April 2004
Available online 17 June 2004
Abstract
3␣,17␤-Dihydroxy-3␤-methyl-5␣-androstan-6-one (1) and 3␤,17␤-dihydroxy-3␣-methyl-5␣-androstan-6-one (13) were prepared by
the reaction of methylmagnesium bromide with the 3-ketosteroids. Structures and configurations in position 3 were determined by NMR
spectra. Substitution in the position 6 influences the ratio of the products.
© 2004 Elsevier Inc. All rights reserved.
Keywords: Biologically active steroids; 3-Methyl-3-hydroxy-steroids
1. Introduction
6␣. We assumed the reversion of the product ratio formed
under these conditions.
Biologically active steroid compounds with 3␣-hydroxy
group can be easily found in nature [1–3]. Their physiolog-
ical properties present a great potential to pharmacological
industry and agriculture. A significant drawback for their
real life applications is often caused by their low stability in
living organisms; their fast metabolic deactivation, presum-
ably by conjugation of the 3-hydroxy group or its oxidation
to the corresponding ketone. The problem can be solved by
protection of 3-hydroxy group by introduction of geminal
methyl group in position 3. This can prevent the mentioned
deactivation [4].
The published data [5] dealing with the reaction of an
organometallic reagent with 3-oxo-5␣-steroid show, that in
the resulting mixture 3␤-hydroxyderivative prevails with the
ratio being (3:2). The opposite tendency can be observed by
using a more complex reaction sequence utilizing methyl-
sulfoxonium reagent. This reaction gives an oxirane inter-
mediate ((3R)-spiro[oxiran-2ꢀ,5␣-steroid]). The cleavage of
this oxirane with sodium iodide and acetic acid followed
with hydrogenolysis of 3␣-hydroxy-3␤-iodomethyl deriva-
tive give the target 3␣-hydroxy-3␤-methyl compound. In this
paper, we dealt with reaction of organometallic reagent with
3-oxo-5␣-steroid containing a bulky substituent in position
2. Experimental
2.1. General methods and equipment
Melting points were determined on a Koefler melting
point micro apparatus Boetius (Germany) and are uncor-
rected. Analytical samples were dried over phosphorus
pentoxide at 50 ◦C/100 Pa. Optical rotations were mea-
sured in chloroform using an Autopol IV (Rudolf Re-
search Analytical, Flanders, USA, [α]D values are given in
10−1 degree cm2 g−1). IR spectra were recorded on a Bruker
IFS 88 spectrometer in chloroform solutions, wavenumbers
are given in cm−1. Detailed NMR study of C(3)-epimeric
pairs (1, 13), (8, 9), and (11, 12) was done on Varian
UNITY-500 and Buker AVANCE-500 instruments (1H
at 500 MHz; 13C at 125.7 MHz). Proton NMR spectra of
other compounds were measured on spectrometers Varian
UNITY-200 (at 200 MHz) and/or Bruker AVANCE-400 (at
400 MHz) in CDCl3 with tetramethylsilane as internal refer-
ence. TAI stands for trichloroacetyl isocyanate. TAC stands
for trichloroacetyl carbamate. Chemical shifts are given in
ppm (δ-scale), coupling constants and widths of multiplets in
Hz. Unless otherwise stated, the data were interpreted as the
first-order spectra. Thin-layer chromatography (TLC) was
performed on silica gel (ICN Biochemicals). Preparative
TLC (PLC) was carried out on 200 mm × 200 mm plates
coated with a 0.7-mm thick layer of the same material. For
ଝ
Presented in part at the conference “Advances in Organic, Bioorganic
and Pharmaceutical Chemistry” (in Czech), Nymburk, Czech Republic,
N∗ovember 28–30, 2003.
Corresponding author. Fax: +420 220 183 578.
E-mail address: pouzar@uochb.cas.cz (V. Pouzar).
0039-128X/$ – see front matter © 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.steroids.2004.04.007

606
H. Chodounska´ et al. / Steroids 69 (2004) 605–612
column chromatography, 60–120 ␮m silica gel was used.
Methylmagnesium bromide, 3 M solution in diethyl ether
was purchased from Sigma-Aldrich (Prague, Czech Repub-
lic). Jones reagent was prepared by the dissolving of CrO3
(5.2 g) in concentrated H2SO4(4.6 ml) and water (4 ml) and
subsequently diluting with water (20 ml). Whenever aque-
ous solutions of hydrochloric acid was used, concentration
was 5%. Solvents were evaporated on a rotatory evaporator
in vacuo (0.25 kPa, bath temperature 40 ◦C).
(H-17); 3.89 m, 1 H, W = 23 (H-6ꢀb (THP)); 3.57 m, 1 H, W
= 32 (H-3); 3.44 m, 2 H (H-6 and H-6ꢀa (THP)); 2.22 m, 1 H
(H-5); 1.19 s, 9 H ((CH3)3 (Piv)); 0.82 s, 3 H (19-H3); 0.79 s,
3 H (18-H3). Analysis calculated for C29H48O5 (476.7):
73.07% C, 10.15% H; found: 73.15% C, 10.28% H.
2.4. 6α-(tert-Butyldimethylsilyloxy)-3β-(2-tetrahydro-
pyranyloxy)-5α-androstan-17β-yl pivalate (5)
tert-Butyldimethylsilyl chloride (1.56 g, 10.4 mmol)
was added at 0 ◦C to a solution of hydroxyderivative 4
(2.25 g, 4.72 mmol) and imidazole (1.26 g, 18.5 mmol) in
N,N-dimethylformamide (50 ml). The reaction mixture was
allowed to stand at room temperature overnight, diluted
with ether (200 ml), and washed successively with 5%
aqueous citric acid (three times), water (three times), satu-
rated aqueous sodium hydrogen carbonate (three times) and
water (twice). Combined organic extracts were dried over
the anhydrous sodium sulfate. The solvent was evaporated,
and the residue was chromatographed on a column of silica
gel (80 g) with a mixture of light petroleum/ether (98:2 to
90:10). The yield of 5 was 2.56 g (92%), mp 140–142 ◦C
2.2. 3β-(2-Tetrahydropyranyloxy)-androst-5-en-17β-yl
pivalate (3)
To a stirring solution of the alcohol 2 [6] (4.5 g,
12.0 mmol) in pyridine (10 ml) at 0 ◦C was added pival-
oyl chloride (1.2 eq., 1.7 ml, 14.4 mmol). The mixture was
warmed to 30 ◦C and let stand overnight. Water was added
and the mixture was extracted with ethyl acetate (3 ×
30 ml) and the combined extracts were washed with sat-
urated aqueous sodium hydrogen carbonate. The organic
layer was dried over anhydrous sodium sulfate, filtered
and concentrated under reduced pressure (4.9 g). Chro-
matography on the column of silica gel (150 g, elution with
light petroleum/acetone, 19:1) and crystallization (ethyl
acetate/light petroleum) provided 3.3 g (60%) of protected
=
(ether), [α]D +61.8 (c 0.27). IR: 1717 (C O); 1174, 1160
1
(C–O); 1077 (C–OSi). H NMR: 4.73 dd, 1H, J1 = 3.1, J2
=7 (H-2ꢀ (THP)); 4.56 dd, 1H, J1 = 7.7, J2 = 8.9 (H-17);
3.89 m, 1 H, W = 25.3 (H-6ꢀb (THP)); 3.46 m, 2 H (H-3 and
H-6ꢀa (THP)); 3.38 dt, 1 H, J1 = 4.5, J2 = 9.8 (H-6); 1.19 s,
9 H ((CH3)3 (Piv)); 0.88 s, 9 H ((CH3)3CSi); 0.82 s, 3 H
(19-H3); 0.79 s, 3 H (18-H3); 0.05 s, 3 H (CH3Si); 0.03 s,
3 H (CH3Si). Analysis calculated for C35H62O5Si (591.0):
71.14% C, 10.57% H; found: 70.81% C, 10.61% H.
derivative 3, mp 173–174 ◦C, [α]D +51.7 (c 0.51). IR: 1717
1
=
=
(C O); 1673 (C C); 1173, 1160 (C–O); 1077. H NMR:
5.30 m, 1 H (H-6); 4.62 dd, 1 H, J1 = 3, J2 = 7 (H-2ꢀ
(THP)); 4.56 dd, 1 H, J = 9.1 and 7.7 (H-17); 3.89 m, 1 H,
W = 23 (H-6ꢀb (THP)); 3.57 m, 1 H, W = 32 (H-3); 3.44 m,
1 H (H-6ꢀa (THP)); 2.22 m, 1 H (H-5); 1.19 s, 9 H ((CH3)3
(Piv)); 1.02 s, 3 H (19-H3); 0.82 s, 3 H (18-H3). Analysis
calculated for C29H46O4 (458.7): 75.94% C, 10.11% H;
found: 75.97% C, 10.28% H.
2.5. 6α-(tert-Butyldimethylsilyloxy)-3β-hydroxy-5α-
androstan-17β-yl pivalate (6)
2.3. 6α-Hydroxy-3β-(2-tetrahydropyranyloxy)-5α-
androstan-17β-yl pivalate (4)
Tetrahydropyranyl derivative 5 (1 g, 1.69 mmol) was
added at 0 ◦C to a solution of magnesium bromide [7,8]
(1.3 g, 7.0 mmol) in the mixture of absolute ether (5.4 ml)
and absolute benzene (0.6 ml). The reaction mixture was
allowed to stand at room temperature overnight, diluted
with ether (20 ml), and saturated aqueous ammonium chlo-
ride solution (20 ml). Ethereal layer was separated and
water phase extracted with ether (three times). Combined
organic extracts were washed with saturated aqueous am-
monium chloride solution (three times), water (twice) and
dried over the anhydrous sodium sulfate. Evaporation of
the solvent and crystallization of the residue from light
petroleum afforded 790 mg (94%) of 6, mp 170–172 ◦C
To a suspension of sodium borohydride (0.5 g, 13.2 mmol)
in THF (20 ml) a solution of boron trifluoride etherate (2 ml,
12.5 mmol) in THF (10 ml) was added via syringe through
septum under argon at 0 ◦C. The resultant mixture was stirred
for 30 min. Then the solution of olefin 3 (4.0 g, 8.7 mmol) in
THF (30 ml) was added and the mixture was stirred at 0 ◦C
for 9 h. The mixture was treated with aqueous solution of
potassium hydroxide (0.5 g in 5 ml of water) and hydrogen
peroxide (30%, 3 ml) and stirred again for 80 min at 0 ◦C.
The product was extracted with ether (300 ml), and the ethe-
real phase was washed with water (1 l in four portions) and
dried over anhydrous sodium sulfate. After evaporation of
the solvent, the crude product was purified by column chro-
matography (silica gel, 80 g). Elution with the mixture of
light petr◦oleum/ether (95:5) afforded 3.07 g (74%) of 4, mp
=
(light petroleum), [α]D +37.3 (c 0.21). IR: 1717 (C O);
1294, 1286, 1172 (C–O); 1077 (C–OSi). 1H NMR: 4.58 dd,
1H, J1 = 9.1, J2 = 7.7 (H-17); 3.58 m, 1 H, W = 32 (H-3);
3.38 m, W = 32, 1 H (H-6); 1.19 s, 9 H ((CH3)3 (Piv));
0.88 s, 9 H ((CH3)3CSi); 0.82 s, 3 H (19-H3); 0.79 s, 3 H
(18-H3); 0.40 s, 3 H (CH3Si); 0.34 s, 3 H (CH3Si). Analysis
calculated for C30H54O4Si (506.8): 71.09% C, 10.74% H;
found: 71.50% C, 10.85% H.
=
205–207 C (ethanol), [α]D +25.8 (c 0.68). IR: 1717 (C O);
1
1174, 1294, 1026 (C–O); 1077. H NMR: 4.62 dd, 1 H, J1
= 3, J2 = 7 (H-2ꢀ (THP)); 4.56 dd, 1 H, J1 = 9.1, J2 = 7.7

H. Chodounska´ et al. / Steroids 69 (2004) 605–612
607
Table 1
2.6. 6α-(tert-Butyldimethylsilyloxy)-3-oxo-5α-
Proton NMR data of compounds 1, 8, 9, 11, 13, and 12 in CDCl3
androstan-17β-yl pivalate (7)
Proton
1
8
11
13
9
12
To a stirring solution of hydroxyderivative 6 (400 mg,
0.79 mmol) in the absolute dichloromethane (10 ml) was
added pyridinium chlorochromate (100 mg, 0.46 mmol)
and the mixture was stirred for 1 h and then left at room
temperature over night. The mixture was filtered over a
small column of silica gel. Crystallization from the mix-
ture of acetone/heptane gave 384 mg (96%) of the ketone
H-1␣
H-1␤
H-2␣
H-2␤
Me-3
H-4␣
H-4␤
H-5
1.5–1.6 ∼1.34
1.5–1.6 ∼1.50
∼1.50
∼1.50
1.30
1.74
1.5–1.6 ∼1.60
1.06
1.62
1.31
1.73
∼1.58
∼1.58
1.213 s
1.51
1.51
∼1.56
∼1.50
∼1.53
1.5–1.6 ∼1.60
∼1.60
1.267 s 1.225 s 1.268 s 1.210 s 1.229 s
∼1.58 1.84
∼1.58
1.6–1.7 1.90 ddd 1.6–1.7
1.6–1.7 ∼1.26
∼1.58 ∼1.18
∼1.58
1.6–1.7
2.25 ddd
–
1.98 dt
2.32 dd
1.87 m
1.30
1.67
1.34
1.26
1.78
2.68 td 1.43 dd 2.69 ddd 2.24 bdd 1.11
7, mp 158–159 ◦C, [α]D +47.5 (c 1.9). IR: 1710 (C O);
=
H-6
–
3.36 dt
–
–
3.37 dt
0.93 dd
H-7␣
H-7␤
H-8
1.98 dt 0.97 dt 2.01 dt
2.31 dd 1.86 dt 2.31 dd 2.32 dd 1.86 dt
1.86 m ∼1.47
1.95 dt
1294, 1171, 1155 (C–O); 1092, 1078 (C–O–Si); 858, 838
(Si(CH3)2). 1H NMR: 4.58 dd, 1H, J1 = 8.0, J2 = 9.0
(H-17); 3.45 dt, 1 H, J5,6 = J6,7␣ = 10.4, J6,7␤ = 4.5
(H-6); 1.19 s, 9 H ((CH3)3 (Piv)); 1.02 s, 3 H (19-H3);
0.87 s, 9 H ((CH3)3CSi); 0.82 s, 3 H (18-H3); 0.27 s, 6 H
(2 × CH3Si). Analysis calculated for C30H52O4Si (504.8):
71.38% C, 10.38% H; found: 71.21% C, 10.54% H.
1.86 m
1.31
1.85 m
1.21
1.72
1.38
1.16
1.88
1.27
1.56
1.28
2.09
1.46
1.47
0.71
∼1.54
∼1.24
1.15
H-9
1.21
1.74
1.36
1.15
1.87
1.36
1.56
1.27
2.08
1.45
0.78
H-11␣
H-11␤
H-12␣
H-12␤
H-14
H-15␣
H-15␤
H-16␣
H-16␤
H-17
∼1.56
∼1.70
1.21
1.31
∼1.15
∼1.71
1.10
1.26
1.77
1.39
1.61
1.71
1.08
1.30
1.61
1.32
∼1.66
∼1.66
∼1.32
1.32
∼1.31
2.7. 6α-(tert-Butyldimethylsilyloxy)-3α-hydroxy-3β-
methyl-5α-androstan-17β-yl pivalate (8) and
6α-(tert-butyldimethylsilyloxy)-3β-hydroxy-3α-
methyl-5α-androstan-17β-yl pivalate (9)
2.17 ddt 2.17
2.17 dddd 2.17
∼1.46
1.48
∼1.46
1.49
3.68 bt 4.57 dd 4.62 dd 3.68 bt
4.57 dd
4.62 dd
0.808 s
0.772 s
–
3H-18
3H-19
TBDMS
0.748 s 0.793 s 0.806 s 0.751 s 0.793 s
0.720 s 0.757 s 0.727 s 0.776 s 0.822 s
–
0.023 s
0.026 s
0.879 s
–
–
0.025 s
0.031 s
0.879 s
1.191 s
Solution of methylmagnesium bromide in diethyl ether
(3 M, 0.5 ml) was added over 20 min dropwise to the stirring
solution of ketone 7 (360 mg, 0.71 mmol) in THF (23 ml)
cooled to −78 ◦C, under argon atmosphere. The reaction
mixture was stirred at −50 ◦C for 6 h and then allowed
to warm to room temperature. Reaction mixture was di-
luted with a saturated aqueous ammonium chloride solution
(5 ml), concentrated in vacuo on rotatory evaporator and
filtered under suction. Chromatography of obtained crude
product (380 mg) on a silica gel column (20 g) in light
petroleum (gradient to 2% of acetone) afforded isomers 8
and 9.
Piv
–
1.190 s 1.196 s
–
1.196 s
1294, 1171, 1155 (C–O). 1H NMR: 4.57 dd, 1 H, J1 = 7.8,
J2 = 9.1 (H-17); 3.38 dt, 1 H, J1 = 4.5, J2 = 9.8 (H-6);
2.16 m, 1 H, W = 45 (H–16); 1.25 s, 1 H (C(3)-CH3); 0.80 s,
3 H (19-H3); 0.77 s, 3 H (18-H3). Analysis calculated for
C25H42O4 (406.6): 73.85% C, 10.41% H; found: 73.61% C,
10.51% H.
The yield of compound 9 was 38 mg (10%): mp
2.9. 3α-Hydroxy-3β-methyl-6-oxo-5α-androstan-17β-yl
pivalate (11)
157–159 ◦C (ether), [α]D +49.9 (c 0.43). IR: 3603 (OH);
=
1717 (C O); 1294, 1171 (C–O); 1092 (C–OSi); 860, 838
1
(Si(CH3)2); H and 13C NMR data (see Tables 1 and 2).
To a solution of hydroxyderivative 10 (120 mg, 0.3 mmol)
in chloroform (10 ml) at 0 ◦C Jones reagent (10 drops) were
added. After 30 min the reaction mixture was filtered through
the column of silica gel (5 g) in a mixture of chloroform/ether
(5:1). Solvents were evaporated and crystallization of the
residue from acetone/heptane afforded 110 mg (92%) of ke-
The yield of compound 8 was 270 mg (73%). 1H and 13
C
NMR data (see Tables 1 and 2). This crude oily product was
used directly for the next step.
2.8. 3α,6α-Dihydroxy-3β-methyl-5α-androstan-17β-yl
pivalate (10)
tone 11, mp 205–206 ◦C, [α]D −7.5 (c 0.25). IR: 1715, 1703
1
(C O); 1293, 1287, 1172, 1155 (C–O). H and 13C NMR
=
data (see Tables 1 and 2). Analysis calculated for C25H40O4
(404.6): 74.22% C, 9.97% H; found: 74.10% C, 10.05% H.
To a solution of silyl derivative 8 (200 mg, 0.39 mmol)
in acetone (10 ml), three drops of concentrated hydrochlo-
ric acid were added. After 30 min standing at room temper-
ature reaction mixture was diluted with water (30 ml) and
neutralized with aqueous ammonia solution (25%, 0.5 ml).
Crystalline product 10 (136 mg, 87%) was collected by fil-
tration and crystallized from a mixture of acetone/heptane:
2.10. 3α,17β-Dihydroxy-3β-methyl-5α-androstan-6-one
(1)
To a solution of the pivaloyl derivative 11 (60 mg,
0.16 mmol) in methanol potassium hydroxide (0.1 g,
mp 226–228 ◦C, [α]D +32.0 (c 0.39). IR: 1710 (C O);
=

608
H. Chodounska´ et al. / Steroids 69 (2004) 605–612
Table 2
2.12. 3β,17β-Dihydroxy-3α-methyl-5α-androstan-6-one
(13)
Carbon-13 chemical shifts of compounds 1, 8, 9, 11, 13, and 12 in CDCl3
Carbon
1
8
11
13
9
12
Ketone 12 (20 mg, 0.05 mmol) was dissolved in methanol
(4 ml). Potassium hydroxide (0.02 g, 0.36 mmol) was added
and the mixture was heated to 60 ◦C for 4 h under argon at-
mosphere and then poured on ice. Precipitate (15 mg) was
separated by filtration and dried. Crystallization from ether
afforded methyl derivative 13 (8 mg, 51%), mp 213–215 ◦C,
C-1
C-2
C-3
C-4
C-5
C-6
C-7
C-8
33.55
33.63
69.21
34.24
53.70
34.41
34.25
69.23
36.87
48.05
33.59
36.19
37.00
36.03
71.52
38.06
51.21
36.17
33.52
69.17
34.21
53.65
35.57
71.02
35.06
56.46
35.54
71.01
35.04
56.42
212.21
46.38
38.08
51.52
41.02
20.97
36.38
43.46
53.93
23.20
30.43
81.60
11.09
12.36
31.84
–
70.34 212.14 210.59
70.13 210.47
41.48
34.13
53.66
36.37
20.40
36.91
42.86
50.58
23.53
27.50
82.32
12.14
12.44
31.85
−4.00
−4.60
18.15
25.96
178.48 178.14
38.86
27.22
46.33
37.78
51.26
40.99
20.83
36.65
43.26
53.74
23.35
27.39
81.92
12.11
12.34
31.82
–
46.24
37.96
51.47
41.29
21.17
36.33
43.46
54.26
23.19
30.41
81.57
11.10
12.84
25.90
–
41.50
34.13
53.89
36.82
20.63
36.91
42.88
50.60
23.54
27.50
82.28
12.14
13.07
26.62
46.20
37.69
51.25
41.26
21.06
36.62
43.29
54.09
23.35
27.38
81.87
12.11
12.83
25.93
–
=
[α]D −4.2 (c 0.17). IR: 3434 (OH); 1701 (C O); 1125, 1051
(C–O). 1H and 13C NMR data (see Tables 1 and 2). Analy-
sis calculated for C20H32O3 (320.5): 74.96% C, 10.06% H;
found: 74.82% C, 10.15% H.
C-9
C-10
C-11
C-12
C-13
C-14
C-15
C-16
C-17
C-18
C-19
Me-3
TBDMS
2.13. 3,6-Dioxoandrost-4-en-17β-yl benzoate (15)
To a solution of hydroxyderivative 14 [9,10] (500 mg,
1.27 mmol) in acetone (35 ml) Jones reagent (1 ml) was
added. The mixture was vigorously stirred in the open
flask on the sun light at room temperature for 2 h and then
poured into the water (500 ml). Solid material was sepa-
rated by filtration and dried on air. This crude product was
dissolved in chloroform (20 ml), filtrated through alumina
column and the solvent was evaporated. Crystallization of
the residue from methanol gave 410 mg (80%) of diketone
15, mp 182–185 ◦C (decomposition), [α]D +20.2 (c 0.21).
−3.95
−4.61
18.13
25.95
–
–
Piv
–
178.51 178.52
a
a
38.88
27.22
–
–
27.23
27.22
=
=
a The signal was not detected.
IR: 1713, 1693 (C O); 1634 (C C); 1603, 1585 (arom);
1
1279 (C–O). H NMR: 8.04 m, 2 H, (H-2 and H-6 (OBz));
7.57 m, 1 H (H-4 (OBz)); 7.45 m, 2 H (H-3 and H-5 (OBz));
6.20 s, 1 H (H-4); 4.91 dd, 1 H, J1 = 7.6, J2 = 9.3 (H-17);
1.20 s, 3 H (19-H3); 1.00 s, 3 H (18-H3). Analysis calcu-
lated for C26H30O4 (406.5): 76.82% C, 7.44% H; found:
74.70% C, 10.05% H.
1.8 mmol) was added and the mixture was heated to 60 ◦C
for 4 h under argon atmosphere. Then the reaction mix-
ture was poured on ice. Precipitate (45 mg) was separated
by filtration and dried. Crystallization from ether afforded
methyl derivative 1 (38 mg, 80%), mp 224–225 ◦C, [α]D
2.14. 3,6-Dioxo-5α-androstan-17β-yl benzoate (16)
=
−8.5 (c 0.21). IR: 3607, 3457 (OH); 1700 (C O); 1168,
1
1155, 1050 (C–O). H and 13C NMR data (see Tables 1
To a solution of diketone 15 (300 mg, 0.74 mmol) in ethyl
acetate (4 ml) and ethanol (1 ml) palladium on carbon (5%,
30 mg) was added and the mixture was vigorously stirred in
the slight overpressure of hydrogen at room temperature for
6 h. Then the catalyst was filtered off and solvents were evap-
orated. Crystallization of the residue from methanol gave
compound 16 (180 mg, 60%), mp 213–216 ◦C (decomposi-
and 2). Analysis calculated for C20H32O3 (320.5): 74.96%
C, 10.06% H; found: 74.70% C, 10.05% H.
2.11. 3β-Hydroxy-3α-methyl-6-oxo-5α-androstan-17β-yl
pivalate (12)
=
=
tion), [α]D +28.8 (c 0.49). IR: 1712 (C O); 1634 (C C);
1603, 1585 (arom); 1280 (C–O). 1H NMR: 8.03 m, 2 H (H-2
and H-6 (OBz)); 7.56 m, 1 H (H-4 (OBz)); 7.44 m, 2 H (H-3
and H-5 (OBz)); 4. 91 dd, 1 H, J1 = 7.7, J2 = 9.1 (H-17);
0.99 s, 3 H (19-H3); 0.97 s, 3 H (18-H3).
To a solution of silyl derivative 9 (30 mg, 0.06 mmol) in
acetone (5 ml) 1 drop of concentrated hydrochloric acid was
added. After 30 min standing at room temperature the reac-
tion mixture was diluted with water (5 ml) and neutralized
with aqueous ammonia solution (25%, 0.2 ml). Crystalline
product was separated by filtration and dried. Then it was
dissolved in chloroform (2 ml) and the Jones reagent (five
drops) was added at 0 ◦C. After 30 min the reaction mixture
was filtered through the column of silica gel in a mixture
of chloroform/ether (5:1). Evaporation of solvents afforded
2.15. 3α,17β-Dihydroxy-3β-methyl-5α-androstan-6-one
(1) and 3β,17β-dihydroxy-3α-methyl-5α-androstan-6-one
(13) from 3,6-diketone 16
A solution of methylmagnesium bromide in diethyl
ether (3 M, 0.5 ml) was added through septum over 20 min
dropwise to the stirring solution of diketone 16 (400 mg,
1
solid compound 12 (22 mg, 95%). H and 13C NMR data
(see Tables 1 and 2).

H. Chodounska´ et al. / Steroids 69 (2004) 605–612
609
0.98 mmol) in THF (20 ml) cooled to −78 ◦C, under argon
atmosphere. The reaction mixture was stirred at −50 ◦C for
1 h and then diluted with a saturated aqueous ammonium
chloride solution (5 ml), concentrated in vacuo and filtered
under suction. Crude product (220 mg) column chromatog-
raphy (silica gel, elution with a gradient from 100% light
petroleum to 2% light petroleum/acetone) afforded 38 mg
(10%) of starting material and 350 mg (84%) of the mix-
ture of 3-methyl derivatives 17 and 18 with small amounts
of other products. This mixture was dissolved in methanol
(20 ml). Potassium hydroxide (0.30 g, 5.34 mmol) was added
and the mixture was heated to 60 ◦C for 4 h under argon
atmosphere. Then the reaction mixture was poured on ice.
Precipitate (310 mg) was separated by filtration and dried.
Preparative TLC in a mixture of light petroleum/acetone
(4:1) afforded 1 (96 mg, 32%) and 13 (136 mg, 45%) ac-
cording to mp, IR, and NMR spectra identical with samples
mentioned above.
hydroxyl group in compound 2 was further on protected with
dihydropyrane giving compound 3. 6␣-Hydroxy group was
introduced by hydroboration reaction of the double bond
5,6 and new hydroxyl group was subsequently protected
by tert-butyldimethylsilyl group. The individual protecting
groups were chosen in a way to take advantage of their dif-
ferent reactivity for the following selective deprotection re-
actions. Tetrahydropyranyl was cleaved off with magnesium
bromide [7,8]. The 3-hydroxy derivative 6 yielded by this
reaction was oxidized with pyridinium chlorochromate to
give ketone 7.
The reaction of 3-oxo derivative 7 with methylmagne-
sium bromide yielded a mixture of methylcarbinols 8 and 9
which were separated on a silica gel column. The amount
of the less polar crystalline product 9 was 10%. Conditions
to crystallize the more abundant (73%) product 8 were not
found either by choosing different solvents, or by lower-
ing the temperature, or prolonging the time of crystalliza-
tion. The analysis of NMR experiments, which will be de-
scribed in a latter paragraph, proved the product 8 to have
the ␣-configuration of hydroxyl group.
3. Results and discussion
Selectively protected androstenediol 2 was used for the
synthesis of the 3␣-methylcarbinol 1 (Scheme 1). The free
The product ratio 8:9 could be explained by the presence
of the bulky substituent (6␣-(tert-butyldimethylsilyloxy)
OH
OPiv
OPiv
i
ii
THPO
THPO
H
THPO
2
3
OH
iii
4
OPiv
iv
OPiv
OPiv
v
HO
THPO
O
H
H
6
H
7
5
TBDMSO
TBDMSO
TBDMSO
vi
OPiv
OPiv
OR
HO
1. vii
2.viii
HO
HO
H
H
8
9
H
12 R = Piv
13 R = H
O
ix
TBDMSO
TBDMSO
vii
OPiv
viii
OH
OPiv
ix
HO
HO
HO
H
H
H
OH
O
O
10
11
1
Scheme 1. (i) Pivaloyl chloride, pyridine; (ii) NaBH4, BF3-etherate, H2O2, KOH, THF; (iii) TBDMSCl, Im, DMF; (iv) MgBr2, ether, benzene; (v) PCC,
CH2Cl2; (vi) MeMgBr, −50 ◦C, THF; (vii) HCl, acetone; (viii) Jones reagent, CHCl3; (ix) KOH, MeOH.

610
H. Chodounska´ et al. / Steroids 69 (2004) 605–612
OBz
OBz
OBz
ii
i
HO
O
O
H
15
14
16
O
O
OBz
OBz
+
iii
HO
iv
HO
H
H
O
18
17
O
OH
OH
+
HO
HO
H
H
13
O
1
O
Scheme 2. (i) Jones reagent, acetone, air, sunlight; (ii) H2, Pd/C, EtOH, EtOAc; (iii) MeMgBr, −50 ◦C, THF; (iv) KOH, MeOH.
group) on the ␣-side of the molecule. In the case of selec-
tive reaction of 3,6-diketone 16 (lacking a bulky substituent
on ␣-side) under similar conditions the 3␣-methyl deriva-
tive prevails, see below. The product ratio changed with
the reaction temperature. At −10 ◦C the desired product 8
dominated only slightly in the reaction mixture, the ratio
␣:␤ being 3:2. This ratio increased significantly in favor of
the desired product 8 when the reaction temperature was
lowered to −50 ◦C. The reaction was quantitative, no side
products were isolated.
The reaction sequence was terminated by the deprotection
of the hydroxy group in position 6. This group was oxidized
with the Jones reagent, as there were no acid labile protecting
groups in the molecule any more, to afford ketone 11. The
last step was basic hydrolysis of the pivaloyl ester group in
compound 11. This reaction needs to be carefully controlled,
as the isomerization in position 5 could take place and would
lead to the change of annulation of rings A and B. From
the reaction mixture only product 1 was isolated having the
5␣-configuration.
The selective reaction of 3,6-diketone with methylmag-
nesium bromide presents a shorter way to 3-methylcarbinol
1. It starts with 3␤-hydroxyandrost-5-en-17␤-yl benzoate
14 [9,10]. This compound was oxidized with Jones reagent
[11] to give the 3,6-diketone 15 under conditions when the
oxidizing agent itself was Cr(IV) formed in situ. Hydrogena-
tion of this unsaturated diketone 15 in ethyl acetate/ethanol
catalysed by paladium on carbon gives a mixture of 5␤-
and 5␣-dihydroderivatives where the 5␣ one predominates.
Diketone 16 on reaction with methylmagnesium bromide at
−50 ◦C yielded quickly a reaction mixture that contained a
small amount of the starting compound and chromatograph-
ically unseparable mixture of 3␣-hydroxy-3␤-methyl- and
3␤-hydroxy-3␣-methyl-derivatives (identified with NMR)
accompanied by a small amounts of other products. The
selectivity of the reaction in respect to the two keto groups
in the molecule is caused by different steric hindrance.
The isomeric 3-methylderivatives were successfully sep-
arated after basic hydrolysis of the benzoyl protecting group
at C-17 to give compounds 1 and 13, identical with the
derivatives prepared via the reaction sequence mentioned
above (Scheme 1).
The reaction sequence in Scheme 1 has nine steps starting
with the 17-protected derivative 2 and the over all yield of
the desired product 1 being 17%. The shorter alternative
showed in Scheme 2 has four steps, starting also with the
17-protected derivative 14 and the yield of the isomer 1 was
only 13%, while the yield of the unwanted isomer 13 was
21%.
Although the reaction sequence 1 is more time and ma-
terial consuming, it was necessary to take this path. The
orthogonal protection of individual functional groups al-
lowed us to unambiguously identify and characterize the
obtained products, and to prepare a sample of the desired
3␣-hydroxyderivative. We found out the best reaction con-
ditions (low temperature, relatively short time, and excess of
reagent and solvent) for the key step—the reaction of 3-oxo
group with the Grignard reagent. It is important to realize,
that the reaction of the diketone 16 with methylmagnesium
bromide can theoretically yield two isomeric methyl deriva-
tives for each position 3 and 6, as well as various combina-
tions of 3,6-dimethyl derivatives. More complications arise
from the possibility of isomerization in position 5. The con-
ditions typical for Grignard reaction could even potentially
cause partial cleavage of the protecting group in position 17.
Searching for the desired products in this complex mixture
of compounds with similar Rf without any standards was
quite difficult.

H. Chodounska´ et al. / Steroids 69 (2004) 605–612
611
CH3
CH3
HO
H3C
CH3
OH
3α-OH-3β-Me-
3β-OH-3α-Me-
(A)
(B)
Fig. 1. The NOE-contacts of C(3)-methyl group in 3␣-OH-3␤-Me- (A) and 3␤-OH-3␣-Me-derivatives (B).
3.1. Configuration of C(3)-methyl group in the epimeric
carbinols
and 12 the induced shifts for all protons in vicinal posi-
tion to OTAC group are approximately the same (+0.40 to
+0.50 ppm). These observations are in a good agreement
with stereochemical conclusions made from NOEs since
for 3␣-OH derivatives 1, 8, and 11 we should expect larger
induced acylation shifts for gauche-oriented hydrogens
H(2␣) and H(4␣) than for trans-oriented H(2␤) and H(4␤)
(see Fig. 2A). On the contrary for 3␤-OH derivatives 13,
9 and 12 all vicinal protons adopt gauche-orientation to
OTAC group and therefore show similar induced acylation
shifts (see Fig. 2B). In case of diols 1 and 13 the addi-
tional TAI-acylation of 17-OH group results in the induced
acylation shifts observed for protons on ring D and Me-18
(the largest 1.08 ppm at H(17)). In 13C NMR spectra the
largest positive TAI-acylation shifts (around +17 ppm)
were observed in whole series at C(3), while for carbons
in ␤-positions (C(2), C(4), and C(3)-Me) significant nega-
tive induced shifts (−2.4 to −5.7 ppm) were found. Much
Determination of the configuration at position 3 in
epimeric pairs of carbinols (1, 13), (8, 9), and (11, 12)
from NMR spectra is not straightforward due to the ab-
sence of any proper interproton coupling. To solve this
problem we have combined the proton 2D-NOE spectra
1
with in situ TAI-acylation shifts observed in both H and
13C NMR spectra. The first necessary step was the struc-
tural assignment of proton and carbon signals. It has been
achieved using 2D NMR methods—mainly H,H-COSY and
H,C-HSQC spectra—together with signal multiplicity and
chemical shift arguments. Geminal CH2 protons at ␣- and
␤-side of steroid skeleton were distinguished by NOE con-
tacts with protons of known stereochemistry (mainly 18-
and 19-Me-groups). The additional NOE contacts observed
between C(3)-methyl protons and axial hydrogen H(1␣)
and H(5␣) in compound 9 indicate an axial ␣-orientation
of C(3)-methyl group (see Fig. 1B). On the other hand
C(3)-methyl group in the epimer 8 showed the NOE con-
tacts only to the hydrogens in vicinal positions 2 and 4 in
agreement with the situation expected from models (see
Fig. 1A). The chemical relation between compounds 8, 11,
and 1 similarly as between compounds 9, 12, and 13 al-
lowed to assign the configuration of C(3)-methyl groups for
all studied carbinols. 1H and 13C NMR data are summarized
in Tables 1 and 2.
Table 3
TAI-acylation shifts (in ppm) of selected proton and carbon atoms in
carbinols 1, 8, 9, 11–13
3␣-OH-3␤-Me-derivatives
3␤-OH-3␣-Me-derivatives
13 12
1
8
11
9
Proton
H-1␣
H-1␤
H-2␣
H-2␤
3-Me
H-4␣
H-4␤
H-5
∼0
∼0
−0.08
∼0
−0.02
0.11
0.87
−0.07
0.362
0.64
0.05
0.04
0.06
0.50
0.43
0.345
0.40
0.43
0.05
0.052
0.05
0.07
0.40
0.40
0.379
0.40
0.40
0.06
0.054
0.05
0.47
0.47
0.374
0.40
0.40
0.05
0.049
0.91
0.83
The in situ TAI-acylation [12,13] was used to pre-
pare 3-OTAC derivatives of compounds 8, 9, 11, 12 and
3,17-diOTAC derivatives of diols 1, 13. All proton and car-
bon signals were structurally assigned in their NMR spectra
similarly as described above. The observed TAI-acylation
shifts of selected protons and carbons in the vicinity of
−0.03
0.368
0.65
−0.04
0.375
0.66
−0.04
−0.05
0.044
0.02
0.01
−0.03
−0.06
3H-19
0.044
0.030
1
Carbon
C-1
C-2
C-3
Me-3
C-4
C-5
C-10
C-19
reaction site (at C(3)) are summarized in Table 3. In H
−0.17
−2.97
16.84
−5.72
−2.49
−0.50
−0.40
0.31
−0.19
−2.81
16.76
−5.39
−2.48
−0.20
−0.34
0.44
−0.18
−2.88
16.88
−5.70
−2.42
−0.45
−0.33
0.31
0.13
−4.10
16.11
−3.48
−3.81
−1.00
−0.15
0.01
−0.62
−4.65
16.99
−3.54
−4.03
−0.57
−0.07
0.04
0.34
−4.03
16.23
−3.50
−3.73
−0.95
−0.06
0.03
NMR spectra the TAI-acylation shift of C(3)-Me group is
roughly same for whole series but the characteristic differ-
ences are observed for geminal protons in position 2 and 4.
Compounds 1, 8, and 11 show large positive shifts for pro-
tons H(2␣) ∼0.85 and H(4␣) ∼0.65 while their ␤-oriented
partners H(2␤) and H(4␤) show very small negative shifts
(around −0.05). On the other hand in compounds 13, 9,

612
H. Chodounska´ et al. / Steroids 69 (2004) 605–612
(+0.45)
H( β)
2
(+0.40)
H( β)
4
(-0.05)
H( β)
4
(-0.05)
H( β)
2
C( )
2
C( )
2
TACO
CH
C( )
4
OTA
C
H C
3
C( )
4
3
H( α)
(+0.65)
C( )
1
H( α)
(+0.40)
C( )
1
4
4
C(5)
H(2α)
(+0.48)
C(5)
H(2α)
(+0.85)
OTAC
CH
3
CH
3
OTAC
3α-OTAC-3β-Me-
3β-OTAC-3α-Me-
(A)
(B)
Fig. 2. The Newman projection along C(3)–C(2) and C(3)–C(4) bond in: (A) 3␣-OTAC-3␤-Me-, and (B) 3␤-OTAC-3␣-Me-derivatives. The characteristic
observed TAI-acylation shifts of hydrogens at positions 2 and 4 are given in parentheses.
smaller induced shifts are observed in ␥- and ␦-positions.
Although some characteristic difference of acylation shifts
between both series are obvious (see Table 3) their interpre-
tation in the sense of configuration at C(3) is not straight-
forward. Diols 1 and 13 show the additional remarkable
TAI-acylation shifts for carbons of ring D and 3H(18) (ca
+4.25 ppm at C(17), −3.23 ppm at C(16), −0.17 ppm at
C(13), and +1.05 ppm at C(18)).
supported by research project Z4 055 905 and grants GACR
203/01/0084 and GA AS CR IAA4055305.
References
[1] Mellon SH, Griffin LD, Compagnone NA. Biosynthesis and action
of neurosteroids. Brain Res Rev 2001;37:3–12.
[2] Plassart-Schiess E, Baulieu E. Neurosteroids: recent findings. Brain
Res Rev 2001;37:133–40.
The key step of the reaction sequence is the methyla-
tion of 3-ketone with the Grignard reagent. The reaction of
Grignard reagents with 3-oxo-5␣-steroids yields dominantly
3␣-methyl-3␤-hydroxy derivatives [1]. A similar analogue
3␣-hydroxy-3␤-methyl derivative 17 was unfortunately the
minor product of the shorter reaction sequence, where a re-
action of 3,6-diketone 16 with methylmagnesium bromide
yields more of the undesired 3␤-hydroxy-3␣-methyl deriva-
tive 18.
[3] Bajguz A, Tretyn A. The chemical structures and occurence
of brassinosteroids in plants. In: Hayat S, Ahmad A, editors.
Brassinosteroids. Bioactivity and crop productivity. Dordrecht, the
Netherlands: Kluwer Academic Publishers; 2003. p. 1–44.
[4] Beekman M, Ungard JT, Gasior M, Carter RB, Dijkstra D, Goldberg
SR, et al. Reversal of behavioral effects of pentylenetetrazol
by the neuroactive steroid ganaxolone. J Pharmacol Exp Ther
1998;284:868–77.
[5] Hogenkamp DJ, Tahir SH, Hawkison JE, Upasani RB, Alauddin
M, Kimborough CL, et al. Synthesis and in vitro activity of
3␤-substituted-3␣-hydroxypregnan-20-ones: allosteric modulators of
the GABAA receptor. J Med Chem 1997;40:61–72.
[6] Pouzar V, Schneiderová L, Drašar P, Štrouf O, Havel M. Synthesis of
2-propynyl ethers of steroid alcohols. Collect Czech Chem Commun
1989;54:1888–902.
[7] Kim S, Park JH. Selective removal of tetrahydropyranyl ethers in the
presence of tert-butyldimethylsilyl ethers with magnesium bromide
in ether. Tetrahedron Lett 1987;28:439–40.
[8] Pouzar V, Sameš D, Havel M. Preparation of selectively protected
derivatives of 21-nor-5-pregnene-3␤,20-diol. Collect Czech Chem
Commun 1990;55:499–551.
[9] Chodounská H, Kasal A, Šaman D, Ubik K. One-pot deuteration
and reduction of ketones in the synthesis of [16,16,17-2H3]
epitestosterone. Collect Czech Chem Commun 1996;61:1037–45.
[10] Beyerman HC, Heiszwolf GJ. Reaction of steroidal alcohols with
isobutene: usefulness of t-butyl as hydroxyl-protecting group in a
synthesis of testosterone. J Chem Soc 1963;755–6.
The synthetic methods described above can be used
for the synthesis of 3␣-hydroxy-3␤-methyl derivative 1 in
relatively simple way and reasonable yield. The longer reac-
tion sequence using orthogonal protection of the functional
groups in the intermediates, therefore leading to the most
straightforward course of all the reaction steps involved
in the synthesis, along with NMR spectroscopy, allowed
unambiguous identification and structure as well as con-
figuration determination of products formed. The second
approach offers a faster solution to synthesize the desired
products.
Acknowledgements
[11] Šolaja BA, Milic´ DR, Došen-Mic´ovic´ LI. Oxidation of steroidal
5-en-3␤-ols with Jones reagent in ether. Steroids 1994;59:330–4.
[12] Goodlett VW. Use of in situ reactions for characterization of alcohols
and glycols by nuclear magnetic resonance. Anal Chem 1965;37:431–
2.
[13] Samek Z, Budeˇš´ınský M. In situ reactions with trichloroacetyl
isocyanate and their application to structural assignment of hydroxy
compounds by 1H NMR spectroscopy. A general comment. Collect
Czech Chem Commun 1979;44:558–88.
We thank Mrs. D. Hybšová for skillful technical
assistance. We are indebted to Dr. P. Fiedler for taking and
interpretation of IR spectra. Our thanks are due to Mrs.
M. Snopková and Dr. R. Pohl for the 200 and 400 MHz
1H NMR spectra measurements. Elemental analyses and
optical rotation measurements were carried out in the An-
alytical Laboratory (Dr. L. Holasová, head). The work was