Oxidative fragmentation of pregna-14,16-dien-20-ones to 14 beta-hydroxyandrost-15-en-17-ones.

Article abstract of DOI:10.1021/jo011175+

Two methods have been developed for efficient conversion of pregna-14,16-dien-20-ones into 14 beta-hydroxyandrost-15-en-17-ones. One procedure consists of treatment of the ring-D dienone successively with sodium borohydride and singlet oxygen. The reaction is illustrated by the conversion of pregna-14,16-dien-20-one 1 into 14 beta-hydroxyandrost-15-en-17-one 3, via the corresponding allylic alcohol 2. Although this two-step procedure is simple, it provides 3 in relatively low yield, accompanied by a smaller amount of the isomeric 14 alpha-hydroxyandrost-15-en-17-one 6. An alternative one-step conversion is achieved by treatment of dienone 1 with a peroxyacid in the presence of a strong protic acid. This process is illustrated by the two-step conversion of dienone 1 into hydroxy ketone 11 in 51% overall yield (Scheme 5) and by the analogous conversion of dienone 13 into hydroxy ketone 24 in 61% overall yield (Scheme 11).

Full text of DOI:10.1021/jo011175+

Oxid a tive F r a gm en ta tion of P r egn a -14,16-d ien -20-on es to
14â-Hyd r oxya n d r ost-15-en -17-on es
J ennifer D. Fell and Clayton H. Heathcock*
Center for New Directions in Organic Synthesis,† Department of Chemistry, University of California,
Berkeley, California 94720
heathcock@cchem.berkeley.edu
Received December 20, 2001
Two methods have been developed for efficient conversion of pregna-14,16-dien-20-ones into 14â-
hydroxyandrost-15-en-17-ones. One procedure consists of treatment of the ring-D dienone succes-
sively with sodium borohydride and singlet oxygen. The reaction is illustrated by the conversion of
pregna-14,16-dien-20-one 1 into 14â-hydroxyandrost-15-en-17-one 3, via the corresponding allylic
alcohol 2. Although this two-step procedure is simple, it provides 3 in relatively low yield,
accompanied by a smaller amount of the isomeric 14R-hydroxyandrost-15-en-17-one 6. An alternative
one-step conversion is achieved by treatment of dienone 1 with a peroxyacid in the presence of a
strong protic acid. This process is illustrated by the two-step conversion of dienone 1 into hydroxy
ketone 11 in 51% overall yield (Scheme 5) and by the analogous conversion of dienone 13 into
hydroxy ketone 24 in 61% overall yield (Scheme 11).
SCHEME 1
For several years, we have been investigating ap-
proaches to the synthesis of cephalostatins, a group of
bis-steroidal pyrazines isolated from the marine tube
worm Cephalodiscus gilchristi.1 During the course of our
investigations, we examined dienes 1 and 2 as possible
precursors to hydroxy-enone 3 (Scheme 1). Although we
initially thought the conversions would take several
steps, we found that 3 could be generated from both
dienes, in one-pot processes through unusual oxidation-
fragmentation reactions. These results are described
herein.
Keto diene 1 was synthesized from the steroid heco-
genin in four steps following published procedures.2 Re-
duction of 1 under Luche conditions3 provided the C20
alcohol (2) (Scheme 2) as a single diastereomer4 as deter-
mined by analysis of the O-methyl mandelate esters.5
Synthesis of 3 would require the oxidative cleavage of
the C17 side chain and the installation of the C14 hy-
droxy group. We envisioned achieving both transforma-
tions by the addition of singlet oxygen to the diene. The
peroxide generated from the reaction could then be re-
SCHEME 2
† The Center for New Directions in Organic Synthesis is supported
by Bristol-Myers Squibb as Sponsoring Member and Novartis as a
Supporting Member.
(1) (a) Pettit, G. R.; Inoue, M.; Kamano, Y.; Herald, D. L.; Arm, C.;
Dufresne, C.; Christie, N.; Schmidt, J . M.; Doubek, D. L.; Krupa, T. S.
J . Am. Chem. Soc. 1988, 110, 2006. (b) Pettit, G. R.; Tan, R.; Xu, J .;
Ichihara, Y.; Williams, M. D.; Boyd, M. R. J . Nat. Prod. 1998, 61, 955.
(2) Micovic, I. V.; Ivanovic, M. D.; Piatak, D. M. Synthesis 1990,
591.
(3) Gemal, A. L.; Luche, J .-L. J . Am. Chem. Soc. 1981, 103, 5454.
(4) The modest yield of the reaction is due to the loss of material
upon chromatography of the product. Typically the reaction proceeded
well enough that no purification was necessary.
(5) Trost, B. M.; Belletire, J . L.; Godleski, S.; McDougal, P. G.;
Balkovec, J . M.; Baldwin, J . J .; Christy, M. E.; Ponticello, G. S.; Varga,
S. L.; Springer, J . P. J . Org. Chem. 1986, 51, 2370.
duced and the resulting C17,C20 diol oxidized to the
ketone.
In the event, treatment of 2 with singlet oxygen did
provide an expected peroxide (4). However, the reaction
also provided hydroxy enone 3. When the cycloaddition
reaction was performed on keto diene 1, none of the
desired peroxide was generated. The direct formation of
hydroxy enone 3 was an unexpected but most fortuitous
result. The singlet oxygen reaction produces two diaster-
eomeric endoperoxides, the R isomer 4 and the â isomer
10.1021/jo011175+ CCC: $22.00 © 2002 American Chemical Society
Published on Web 05/31/2002
4742
J . Org. Chem. 2002, 67, 4742-4746

Oxidative Fragmentation of Pregna-14,16-dien-20-ones
SCHEME 3
synthesis of 3. Since there is some precedent in the
literature for performing a Baeyer-Villiger reaction in
preference to epoxidation,7,8 we envisioned a Baeyer-
Villiger reaction on keto-diene 1 as a means of installing
the C17 ketone via an enol acetate. This strategy also
produced surprising results. Treatment of 1 with anhy-
drous m-CPBA and concentrated sulfuric acid in chloro-
form gave the desired hydroxy enone 3 in 30% yield,
accompanied by 15% of the unexpected diketone 7.
The formation of hydroxy enone 3 is dependent on both
the use of anhydrous m-CPBA and the presence of a
strong acid. Treatment of 1 with commercially available
75% m-CPBA and no added acid resulted in exclusive
formation of the monoepoxide 8. Moreover, treatment of
1 with 100% m-CPBA in the presence of H2O also
resulted solely in the formation of 8. This result strongly
indicates that the water present in commercial m-CPBA
is responsible for the marked reactivity difference. The
reason may be that, in the presence of water, the carbonyl
group of the dienone is less likely to be protonated,
thereby retarding the acid-catalyzed Baeyer-Villiger
reaction, and epoxidation predominates. The fact that
omission of the acid catalyst (and the use of anhydrous
m-CPBA) also results exclusively in monoepoxide forma-
tion demonstrates that protonation of the carbonyl group
is essential for the Baeyer-Villiger reaction to take place.
Additionally, it was found that monoepoxide 8 does not
convert to desired enone 3, but rather reacts to produce
the side-product diketone 7. This transformation mostly
likely takes place though a Baeyer-Villiger reaction to
produce the enol acetate, which undergoes an acyl
transfer reaction similar to a Fries rearrangement (Scheme
4) followed by the â elimination of the epoxide.
Though these experiments, it was made clear that to
favor formation of 3, the Baeyer-Villiger reaction must
occur prior to epoxidation. Since the presence of H2O
promotes epoxidation, the use of anhydrous reaction
conditions is essential. Therefore, in the optimized reac-
tion conditions, H2SO4 was replaced with anhydrous
MeSO3H. Moreover, the use of excess m-CPBA increased
the yield. This is largely a result of the fact that excess
m-CPBA results in shorter reaction times and decreased
exposure of the tertiary allyic alcohol present in the
product to the very acidic reaction conditions. Also, the
yield was improved when 3 was not isolated, but was
hydrogenated prior to purification. These conditions
(Scheme 5) provide the desired hydroxy-ketone (11) in a
synthetically useful 51% yield.
TABLE 1. Rea ction of Dien on e 2 w ith Sin glet Oxygen
temp, yield of 3, yield of 4,
run
sensitizer
Rose Bengal EtOH
TPP EtOH
Rose Bengal MeOH
TPP CH2Cl2
Rose Bengal CH2Cl2/MeOH
Rose Bengal, CH2Cl2
TEA salt
solvent
°C
%
%
1
2
3
4
5
6
23
23
0
0
0
23
21
54
16
21
22
13
12
30
39
40
20
0
5. We believe that isomer 5 undergoes facile Grob
fragmentation, as illustrated in Scheme 3, giving rise to
hydroxy enone 3. Diastereomer 4 is more stable because
the necessary anti-periplanar conformation is disfavored
by steric repulsion between the side-chain methyl group
and the equatorial acetoxy group at C12 (see Scheme 3).
However, fragmentation of 4 is induced by heating in
methanol, giving the diastereomeric R-hydroxy enone 6.
The singlet oxygen reaction was examined in more
detail (Table 1). Factors such as sensitizer, solvent, and
temperature were studied to determine their effect on
product ratios and yield. Cooling the reaction to 0 °C (run
3) provided an increase in yield of both products, but upon
lowering the temperature further, only starting material
was recovered. A significant effect on the product ratio
was observed when the solvent was changed from an
alcohol (runs 1-3) to CH2Cl2 (runs 4-6). Since hydroxy
groups can exert a directing effect on singlet oxygen
addition reactions6 and the use of a non-hydrogen bond-
ing solvent such as CH2Cl2 should amplify this effect,
these results indicate that the R peroxide 4 arises from
the directed addition of the singlet oxygen to the more
hindered a face of the diene by the C20 alcohol. In an
alcoholic solvent, the C20 alcohol is hydrogen-bonded to
the solvent molecules and is less likely to exert its
directing influence on the singlet oxygen. Thus, more of
the â peroxide is formed, which subsequently fragments
to hydroxy-enone 3.
While the singlet oxygen reaction is an interesting
transformation from a mechanistic standpoint, the m-
CPBA methodology proved to be the most synthetically
useful route to hydroxy ketone 11 because it is easily
scaleable and a single diastereomer of the product is
obtained in acceptable yield. For these reasons, it ap-
peared to be the method of choice if advanced intermedi-
ate 3 was to be used as part of our cephalostatin 1 project.
However, since the methodology is well suited to prepar-
ing the C14 hydroxylated androstane steroid skeleton,
it also has the potential to provide more immediate
utility. 14-Hydroxy-androstane intermediates have been
used extensively in the synthesis of cardiac glycosides,
While this singlet oxygen methodology can be used to
generate hydroxy-enones 3 and 6, the drawbacks, includ-
ing irreproducible yields upon scale-up, difficulties in
purification of the products, and generation of diastere-
omeric mixtures, led us to investigate other routes for
(6) Adam, W.; Peters, E. M.; Peters, K.; Prein, M.; von Schering, H.
G. J . Am. Chem. Soc. 1995, 117, 6686.
(7) Montury, M.; Gore´, J . Tetrahedron 1977, 33, 2819.
(8) Krow, G. R. Org. React. 1993, 43, 300.
J . Org. Chem, Vol. 67, No. 14, 2002 4743

Fell and Heathcock
SCHEME 4
SCHEME 5
SCHEME 7
SCHEME 6
but their preparation is somewhat lengthy.9,10 We there-
fore decided to explore the application of the m-CPBA
oxidation-fragmentation route to the cardiac glycoside
steroid skeleton. Our attention therefore turned to the
synthesis of the requisite keto-diene 13 (Scheme 6).
The preparation of dienone 13 by a route analogous to
the preparation of dienone 1 has been reported several
times in the literature and in each report significant
amounts of enone 13 have been recovered after the
bromination-elimination process.11 To determine why
this is the case, the reaction sequence was examined more
closely. Examination of the bromination reaction in more
detail showed that treatment of 12 with NBS under
typical free radical conditions does not result in a simple
allylic bromination reaction, but instead gives three
products (Scheme 7). One of these, obtained as a crystal-
line solid, was found to be dibromide 14, resulting from
the addition of bromine to the double bond of enone 12.
The other two products were an inseparable mixture of
dibromide 15 and tribromide 16. Although products 15
and 16 cocrystallized, their structures were elucidated
by single-crystal X-ray analysis of the crystalline mixture.
In this structure determination, atom Br3, attached to
C21, had a refined occupancy of 8%.
Dibromide 14 arises from addition of bromine to the
enone olefin, while dibromide 15 and tribromide 16 are
derived from addition of bromine to the desired diene 13.
The C21 bromine of 16 most likely comes from a bromi-
nation. When more equivalents of NBS are used in the
reaction, more of 16 is obtained. When treated with base,
15 and 16 converted cleanly to the desired diene 13, while
14 reverted back to the starting enone.
With dienone 13 in hand our attention turned once
again to the oxidation-fragmentation reaction. Initial
reactions proved to be problematic. When 13 was treated
with purified m-CPBA and MeSO3H, the conditions that
were optimized for dienone 1, three products were
obtained (Scheme 8). The desired hydroxy enone 17 was
(9) Daniewski, A. R.; Kabat, M. M.; Masnyk, M.; Wicha, J .;
Wojciechowska, W. J . Org. Chem. 1988, 53, 4855.
(10) Groszek, G.; Kabat, M. M.; Kurek, A.; Masnyk, M.; Wicha, J .
Bull. Pol. Acad. Sci., Chem. 1986, 34, 313.
(11) (a) Bach, G.; Capitaine, J .; Engel, C. R. Can. J . Chem. 1968,
46, 733. (b) Ferland, J . M.; Lefebvre, Y. Can. J . Chem. 1984, 62, 315.
(c) Pettit, G. R.; Piatak, D. M. Can. J . Chem. 1966, 44, 844.
4744 J . Org. Chem., Vol. 67, No. 14, 2002

Oxidative Fragmentation of Pregna-14,16-dien-20-ones
SCHEME 8
SCHEME 10
SCHEME 11
SCHEME 9
stable cis configuration. In the case of monoepoxide 8 the
C12 acetoxy group inductively disfavors formation of the
tertiary carbocation 22 because of electrostatic inter-
action with the dipole of the C12 substituent. Therefore,
this substrate instead undergoes Baeyer-Villiger reac-
tion and subsequent Fries rearrangement as described
previously.
obtained in only 10% yield, along with an inseparable
mixture of 18 and 19.12 Diene 18 arises by dehydration
of the tertiary, allylic alcohol in 17, and this transforma-
tion was confirmed by treatment of pure 17 with meth-
anesulfonic acid in chloroform.
Formation of the unwanted side-product 19 epoxide
product was minimized by changing the reaction condi-
tions from m-CPBA in CHCl3 to magnesium monoper-
oxyphthalate (MMPP) in ether. Reduction of the initially
formed hydroxy enone gave the saturated hydroxy ketone
24 in 61% overall yield (Scheme 11). Initially the per-
phthalic acid was generated by reaction of MMPP with
methanesulfonic acid in ether and this solution was
added to an ether solution of dienone 13, as suggested
by Bo¨hme.14 However, preparation of the oxidizing agent
in this manner tended to be inconsistent, and better
results were obtained when the perphthalic acid was
generated in situ from magnesium monoperoxyphthalic
acid (MMPP) and MeSO3H.
Diketone 19 appears to arise from monoepoxide 20, as
shown in Scheme 9. Treatment of dienone 13 with
commercially available 75% m-CPBA in the presence of
acid furnished epoxide 20 in quantitative yield. Subjec-
tion of 20 to the previous reaction conditions (anhydrous
m-CPBA and methanesulfonic acid) gave enedione 19 in
43% yield. The structure of 19 was confirmed by conver-
sion to the known diketone 21.13
Interestingly, although monoepoxide 20 and monoep-
oxide 8 are similar in structure, they display very
different reactivity. The reason for this discrepancy
become apparent when one considers the likely mecha-
nism of the formation of diketone 19 (Scheme 10). In the
case of 20, protonation of the epoxide leads to formation
of tertiary cation 22 which undergoes a hydride shift to
provide diketone 23. Under the acidic conditions of the
reaction, the trans ring fusion isomerizes to the more
By using the revised reaction conditions, the desired
hydroxy ketone can be generated in good yield on this
substrate, demonstrating that the oxidation-fragmenta-
(12) These two products did partially separate upon crystallization
from ether.
(13) Ferland, J . M.; Lefebvre, Y. Can. J . Chem. 1984, 62, 315.
(14) Bo¨hme, H. Organic Syntheses; Wiley: New York, 1955; Collect.
Vol. III, p 619.
J . Org. Chem, Vol. 67, No. 14, 2002 4745

Fell and Heathcock
tion reaction methodology is a useful means of generating
important intermediates in the synthesis of cardiac
glycosides.
Ack n ow led gm en t. This work was supported by a
research grant from the United States Public Health
Service (GM46057).
In summary, two novel oxidatative fragmentation
reactions have been discovered that provide a concise
entry into the C14-hydroxylated androstane steroid
skeleton. The Baeyer-Villiger peracid route is syntheti-
cally useful and has been demonstrated to work on dienes
incorporating the steroid skeleton that is common to
many of the cardiac glycosides.
Su p p or tin g In for m a tion Ava ila ble: Experimental pro-
cedures and characterization for all new compounds reported
in this manuscript. This material is available free of charge
via the Internet at http://pubs.acs.org.
J O011175+
4746 J . Org. Chem., Vol. 67, No. 14, 2002