A selective Cu(II)/Fe(III)-mediated hydrogenation of steroidal haloalkenes in the presence of hydrazine

Article abstract of DOI:10.1016/S0040-4039(02)02875-7

A new system for hydrogenation of haloalkenes is reported. Cu(II)/Fe(III)-mediated selective hydrogenation of steroidal haloalkenes in the presence of hydrazine proves to be a very efficient method for the synthesis of 17β-halosteroids, potential candidates as antiestrogens or androgen receptor-mediated imaging agents. The reaction stereospecifically affords β-haloalkanes without any concomitant formation of dehalogenation products.

Full text of DOI:10.1016/S0040-4039(02)02875-7

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 44 (2003) 1351–1354
A selective Cu(II)/Fe(III)-mediated hydrogenation of steroidal
haloalkenes in the presence of hydrazine
Kai-Hsuan Chang,a Malikah N. Jenkins,b Hsin-Ru Wua and Wen-Shan Lia,*
aInstitute of Chemistry, Academia Sinica, Taipei 115, Taiwan
bDepartment of Chemistry, Purdue University, West Lafayette, IN 47907, USA
Received 12 November 2002; revised 17 December 2002; accepted 20 December 2002
Abstract—A new system for hydrogenation of haloalkenes is reported. Cu(II)/Fe(III)-mediated selective hydrogenation of
steroidal haloalkenes in the presence of hydrazine proves to be a very efficient method for the synthesis of 17b-halosteroids,
potential candidates as antiestrogens or androgen receptor-mediated imaging agents. The reaction stereospecifically affords
b-haloalkanes without any concomitant formation of dehalogenation products. © 2003 Elsevier Science Ltd. All rights reserved.
There are significant therapeutic interests associated
with steroidal haloalkanes because of their demon-
strated value as antiestrogens for the treatment of
breast cancer1,2 and their binding affinity as imaging
agents3,4 for the androgen receptor. Katzenellenbogen
and co-workers have also shown that steroidal
haloalkanes, especially at the 16a-position from halo-
genation of the D ring of the steroid skeleton, display
great binding affinities for the lamb and rat uterine
estrogen receptors.5,6 Likewise, steroids bearing a halo-
gen at C-17b in the D ring have the potential to act as
imaging agents or new halo-steroidal antiestrogens.
However, only the replacement of an oxygen function
at C-17b by fluorine has been studied,4 because fluorine
substitution at C-17b is readily accessible by synthesis.
The preparation of other C-17 halo compounds, e.g.
17b-chloro, 17b-bromo and 17b-iodo, is quite trouble-
some. Furthermore, recently we have utilized 17a-
gen moiety8 and alkene formation from elimination
instead of substitution of the 17a-triflate. Several other
methods, such as substitution of the 17b-acetoxy
group9 and hypoiodite reaction of 20-hydroxy
steroids,10–12 are reported to give 17b-bromo and 17b-
iodo steroids. However, these methods incorporate very
tedious chiral separation of diastereomeric intermedi-
ates or provide no information on yields. Diimide,
produced from hydrazine oxidation, is a well-known
cis-specific reducing agent. Although significant pro-
gress has been made on diimide reduction of olefins to
alkanes,13–24 the reduction of vinyl bromide with
diimide is inevitably retarded.25 Despite numerous
attempts to overcome these drawbacks, to the best of
our knowledge, no versatile synthesis of 17b-chloro,
17b-bromo and 17b-iodo steroids has yet been
reported. This paper discloses a new alternative syn-
thetic protocol to allow the production of highly
desired 17b-halo steroids from steroidal haloalkenes.
bromo-3a-(triphenylsilyloxy)-5a-androstan-6-one
to
confirm the capability of the steroid skeleton to facili-
tate relatively long-range through-bond intramolecular
energy transfer.7 Thus, the analogous 17b-halo sub-
strates are the next interesting targets to study. Practi-
cally, 17b-fluoro and 17a-halo derivatives are obtained
by either hydrogenation of the vinyl fluorides using
Pd/C as a catalyst or displacement of the activated
17b-triflate with halogen ion. However, employment of
either method to prepare 17b-chloro, 17b-bromo and
17b-iodo suffers from low yields and side reactions; for
example, palladium-catalyzed elimination of the halo-
We have recently explored the selective reduction of the
vinyl bromide group in 17-bromo-3b-hydroxy-5a-
androstan-5,16-diene (1b) to 17b-bromoalkane 2b,
keeping the 5-ene intact, by a combination of cop-
per(II) acetate and hexacyanoferrate(III) complex in the
presence of hydrazine and air under mild conditions.
The reaction of 1b with Cu(II)/Fe(III) complex was
carried out under various reaction conditions to
confirm the optimum conditions (Table 1).
As can be seen by the results summarized in Table 1,
the copper complex has higher catalytic activity for the
present reaction (entry 1), while Fe(III) complex was
inert (entry 2). Upon addition of Cu(OAc)2 in 0.1
* Corresponding author. Tel.: +886-(2)-278-98662; fax: +886-(2)-278-
31237; e-mail: wenshan@chem.sinica.edu.tw
0040-4039/03/$ - see front matter © 2003 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(02)02875-7

1352
K.-H. Chang et al. / Tetrahedron Letters 44 (2003) 1351–1354
Table 1. Selective hydrogenation of 1b by Cu(II)/Fe(III) complexa
Entry
Metal oxidant
Ratio (mmol)
Reaction time (h)
Yield (%)c
3b5
2b
1
2
3
4
5
6
Cu(OAc)2
K3Fe(CN)6
Cu(OAc)2/K3Fe(CN)6
Cu(OAc)2/K3Fe(CN)6
Cu(OAc)2/K3Fe(CN)6
Cu(OAc)2/K3Fe(CN)6
0.35
5.2
0.035/5.2
0.035/7.0
0.035/7.0
0.035/7.0
24
24
24
24
48
72
9
3
28
36
77
96
43
0
Traceb
0
0
0
a Compound 1b (3.5 mmol) was allowed to react with Cu(II)/Fe(III) complex in the presence of hydrazine hydrate and air in methanol (100 mL)
at room temperature.
b Less than 1%.
c Isolated yields.
equiv, the target compound 2b was obtained only in
low yield (9%)(entry 1). In addition, the debromination
product 3bOH-5-ene 3b5, was also formed in 43% yield.
A similar result was reported by Cowell et al.8 when Pd
complex was used as a metal catalyst. In the case of the
reaction with Cu(OAc)2/K3Fe(CN)6 (the molar ratio is
0.035/5.2), 17b-bromoalkane 2b was obtained in 28%
yield with a trace amount of the debromination com-
pound 3b5 (entry 3). Therefore, we decided to decrease
the molar ratio of Cu(OAc)2/K3Fe(CN)6 to 0.035/7.0.
In this event, only 2b was isolated and no other prod-
ucts were observed (entry 4). Elongation of the reaction
time to 48 or 72 h26 produced 17b-bromoalkane 2b in
77% or 96% isolated yield. On the basis of these results,
the metal oxidant Cu(OAc)2/K3Fe(CN)6 at a molar
ratio of 0.035/7.0 was deemed to be the most effective
complex for use with hydrazine hydrate and air in
methanol, and this combination was used in further
hydrogenation experiments (Table 2).27
Under the hydrogenation protocol, a series of 17-
steroidal haloalkenes (Fig. 1, 1a–c) have been similarly
reduced without any formation of dehydrogenation
products (Table 2, entry 1). Even with electron-with-
drawing halids, high yields (2a–c, 94–96%) could be
obtained in a reasonable amount of time (24–72 h).
This observation is in keeping with the reactivities
predicted on the basis of the electronic effect of chloro,
bromo, and iodo groups. Similarly, 3a–c, like 1a–c,
reacted with Cu(OAc)2/K3Fe(CN)6 complex under the
same conditions to afford 4a–c in excellent yield (96–
97%) with the epoxide moiety intact (entry 2). Obvi-
ously, the combination of copper(II) and iron(III) salts
are apparently not sufficient Lewis acids to promote the
ring opening of the epoxide. 17-Halo-3b-hydroxy-5a-
androstan-16-ene (5a–c) gave the corresponding 17b-
halosteroids, 6a–c, in 95–98% yield (entry 3). In the
case
of
17-bromo-3a-(dimethylphenylsiloxy)-5a-
androstan-16-ene (7), reduction took place exclusively
Table 2. Reaction of various haloalkenes by Cu(II)/Fe(III) complexa
Entry
1
Substrate
Products
X
Time (h)
Yield (%)b
1a
1b
1c
3a
3b
3c
5a
5b
5c
7
2a
2b
2c
4a
4b
4c
6a
6b
6c
8
Cl
Br
I
Cl
Br
I
Cl
Br
I
Br
H
Br
72
48
24
72
48
24
72
48
24
48
96
96
94
96
95
96
97
97
95
97
98
90
98
80
2
3
4
5
6
9
11
10
12
a Reaction conditions: Substrate (3.5 mmol) was allowed to react with Cu(II)/Fe(III) complex in the presence of hydrazine hydrate and air in
methanol (100 mL) at room temperature.
b Isolated yield.

K.-H. Chang et al. / Tetrahedron Letters 44 (2003) 1351–1354
1353
References
1. Levesque, C.; Merand, Y.; Dufour, J.-M.; Labrie, C.;
Labrie, F. J. Med. Chem. 1991, 34, 1624.
2. Schwede, W.; Cleve, A.; Neef, G.; Ottow, E.; Stocke-
mann, K.; Wiechert, R. Steroids 1994, 59, 176.
3. Labaree, D. C.; Hoyte, R. M.; Nazareth, L. V.; Weigel,
N. L.; Hochberg, R. B. J. Med. Chem. 1999, 42, 2021.
4. Liu, A.; Carlson, K. E.; Katzenellenbogen, J. A. J. Med.
Chem. 1992, 35, 2113.
5. Heiman, D. F.; Stephen, G.; Senderoff, J.; Katzenellen-
bogen, J. A. J. Med. Chem. 1980, 23, 994.
6. Fevig, T. L.; Mao, M. K.; Katzenellenbogen, J. A.
Steroids 1988, 51, 471.
7. Li, W.-S.; Morrison, H. Org. Lett. 2000, 2, 15.
8. Cowell, D. B.; Davis, A. K.; Mathieson, D. W.; Nicklin,
P. D. J. Chem. Soc., Perkin Trans. 1 1974, 1505 and cited
references for previous work in the area.
9. Libman, J.; Mazur, Y. J. Chem. Soc., Chem. Commun.
1971, 730.
10. Biollaz, M.; Kalvoda, J. Helv. Chim. Acta 1972, 55, 366.
11. Habermehl, G.; Swidersky, K.-P. Justus Liebigs Ann.
Chem. 1978, 405.
Figure 1. 17-Steroidal haloalkenes and bicyclic bromoalkene
used as substrates in the selective hydrogenation.
12. Meystre, C.; Heusler, K.; Kalvoda, J.; Wieland, P. Helv.
on the 16-ene in 90% yield, despite the presence of the
labile dimethylphenylsiloxy moiety (entry 4). It is note-
worthy that 3a-benzoyl-5a-androstan-16-ene (9) reacted
slowly with Cu(OAc)2/K3Fe(CN)6 complex to afford
unsaturated 3a-benzoyl-5a-androstane (10) in excellent
yield (98%) in 96 h (entry 5), although 9 exhibits a
lower ionization potential than 5a–c.22,28 Moreover, the
present hydrogenation system was found to be applica-
ble to the synthesis of bicyclic a-bromoalkane (12) from
Cu(OAc)2/K3Fe(CN)6 complex and bicyclic bro-
moalkene (entry 6). The stereochemistry of C-1 substi-
tution was established by 1H NMR and molecular
mechanics structure modeling. From the knowledge of
the known stereochemical specificity of reactions on the
D ring of steroids, the incoming diimide29 in the nucle-
ophilic addition at the C-1 position, prefers to attack
the b-side of the five-membered ring, because of the
steric bulk of the a-orientated 7-CH3 moiety.
Chim. Acta 1962, 45, 1317.
13. Hunig, S.; Muller, H. R.; Thier, W. Angew. Chem., Int.
Ed. Engl. 1965, 4, 271 and cited references for previous
work in the area.
14. Garbisch, E. W.; Schildcrout, S. M.; Patterson, D. B.;
Sprecher, C. M. J. Am. Chem. Soc. 1965, 87, 2932.
15. Back, R. Rev. Chem. Intermed. 1984, 5, 293.
16. Johnstone, R. A. W.; Wilby, A. H. Chem. Rev. 1985, 85,
129.
17. Furst, A.; Berlo, R. C.; Hooton, S. Chem. Rev. 1965, 65,
51.
18. Freeman, F. Chem. Rev. 1975, 75, 439.
19. Corey, E. J.; Mock, W. L.; Pasto, D. J. Tetrahedron Lett.
1961, 2, 347.
20. Hunig, S.; Muller, H. R.; Thier, W. Tetrahedron Lett.
1961, 2, 353.
21. Squillacote, M.; Felippis, J. D.; Lai, Y. L. Tetrahedron
Lett. 1993, 34, 4137.
22. Nelson, D. J.; Henley, R. L.; Yao, Z.; Smith, T. D.
In summary, we have successfully developed a mild and
selective method for the production of 17b-halosteroids.
The methodology, which avoids several side reactions
such as dehalogenation and competing substitution, is
the first example of Cu(II)/Fe(III)-mediated synthesis of
17b-chloro, 17b-bromo and 17b-iodo steroids. These
are potential candidates as antiestrogens or androgen
receptor-mediated imaging agents. Application of this
method to the selective synthesis of acyclic haloalkanes
is underway.
Tetrahedron Lett. 1993, 34, 5835.
23. Tang, H. R.; McKee, M. L.; Stanbury, D. M. J. Am.
Chem. Soc. 1995, 117, 8967.
24. Lacombe, P.; Castagner, B.; Gareau, Y.; Ruel, R. Tetra-
hedron Lett. 1998, 39, 6785.
25. van Yamelen, E. E.; Davis, M.; Deem, M. F. J. Chem.
Soc., Chem. Commun. 1965, 71.
26. The reaction time can totally be shortened to 24 h or less
when hydrazine hydrate was frequently added into solu-
tion to facilatate the reaction.
27. General procedure for the synthesis of steroidal haloalkenes
(2a–c, 4a–c, 6a–c, 8, 10, 12): To a stirred solution of vinyl
bromide (3.50 mmol) and N2H4×H2O (14 mL) in
methanol (100 mL) under an atmosphere of air at room
temperature was added K3Fe(CN)6 (2.3 g, 7.00 mmol),
then Cu(OAc)2 (7.0 mg, 0.035 mmol). N2H4×H2O (3–5
mL) was added every 12 h until TLC analysis showed
that starting material was gone. The precipitate was
filtered and washed thoroughly with methanol. The
organic layer was evaporated and the remaining aqueous
Acknowledgements
This research was supported by the NSC and Academia
Sinica. M.N.J. thanks the Sloan Foundation, and the
Purdue University Henry Hass Endowment for fellow-
ships. We thank Professor H. Morrison (Purdue) for
many helpful discussions and advice. We also thank Dr.
Yu-Ju Chen for the mass spectroscopic analyses.

1354
K.-H. Chang et al. / Tetrahedron Letters 44 (2003) 1351–1354
phase was added dilute aqueous HCl to adjust the pH to
(m, 2H), 0.73 (s, 3H), 0.56 (m, 1H); 13C NMR (100 MHz,
CDCl3): l 68.8, 63.2, 62.9, 61.5, 51.0, 50.5, 43.5, 41.7,
37.0, 36.0, 34.78, 32.2, 32.0, 30.6, 30.5, 24.2, 21.3, 16.9,
13.7; LRMS (FAB) m/z: 371 (M+H)+, 369 (M+H)+, 353
(M+H−H2O)+, 351 (M+H−H2O)+; HRMS (FAB) calcd
for C19H29BrO2 (M)+ 368.1351; found, 368.1352.
3h-Hydroxy-5i-benzoyl-7h-methyloctahydroindene-1h-
bromide (13): A benzoyl derivative of product 12; 1H
NMR (400 MHz, CDCl3): l 8.04 ppm (d, 2H), 7.59 (t,
1H), 7.47 (t, 2H), 5.05–4.99 (m, 1H), 4.42 (t, 1H, J=6.9
Hz), 1.50–1.26 (m, 10H), 1.13 (s, 3H); 13C NMR (100
MHz, CDCl3): l 165.9, 132.9, 130.3, 129.6, 128.4, 80.1,
71.4, 61.4, 47.6, 39.4, 35.3, 30.6, 29.6, 27.0, 15.8; LRMS
(FAB) m/z: 355 (M+H)+, 353 (M+H)+; HRMS calcd for
C17H21O3Br (M+H)+, 353.0752; found, 353.0758.
28. The controversy was also reported by Nelson, D. J., et al.
This shows that hydrogen reactions are not exclusively
controlled by both electronic and steric effects in the
present conditions. Torsional strain and bond angle
bending strain may contribute to the observed reactivity
differences.
8–8.5. The resulting water solution was extracted with
CH2Cl2 (3–15 mL), and the combined organic phases
were dried (Na2SO4). The solvent was evaporated and the
pale yellowish solid chromatographed on silica gel in
EtOAc/Hexane to afford corresponding compounds. All
the new products gave satisfactory spectral data in accord
with the assigned stereochemistry.
17i-Iodo-3i-hydroxyandrostan-5-ene (2c): Colourless
solid; mp 166–167°C; 1H NMR (400 MHz, CDCl3): l
5.35 (m, 1H), 3.76 (t, 1H, J=9.3 Hz), 3.53 (m, 1H),
2.38–2.19 (m, 3H), 2.18–1.97 (m, 2H), 1.90–1.80 (m, 2H),
1.78–1.60 (m, 3H), 1.60–1.40 (m, 3H), 1.37–1.20 (m, 2H),
1.17–0.99 (m, 7H), 0.83 (3H); 13C NMR (100 MHz,
CDCl3): l 140.9, 121.1, 71.6, 50.1, 43.8, 42.1, 41.7, 37.2,
36.6, 34.1, 33.2, 31.7, 31.5, 25.4, 20.7, 19.4, 16.8; LRMS
(FAB) m/z: 400 (M)+, 399 (M−H)+, 383 (M+H−H2O)+,
255 (M−H2O−I)+; HRMS calcd for C19H29IO (M)+
400.1263; found, 400.1269.
17i-Bromo-5i,6i-epoxy-3i-hydroxy-5h-androstane (4b):
Colourless solid; mp 144–145°C; 1H NMR (400 MHz,
CDCl3): l 3.65 (t, 1H, J=9.3 Hz), 3.59 (m, 1H), 3.00 (d,
1H, J=1.9 Hz), 2.19 (m, 1H), 2.06–1.82 (m, 4H), 1.78–
1.54 (m, 3H), 1.52–1.11 (m, 9H), 0.93 (s, 3H), 0.91–0.77
29. It should be noted that hydrogenation of the double
bond is not caused by hydrazine itself but by a oxidation
product, diimide, of hydrazine.