Electrochemical deacetoxylation: Synthesis of 11-ketotigogenin

Full text of DOI:10.1021/jo9513469

J . Org. Chem. 1996, 61, 405-407
405
A polarographic study4 of a number of steroidal R-ke-
tols indicated two separate and distinct reduction waves
at -2.0 and -2.45 V (vs a saturated calomel electrode);
these were ascribed to the cleavage of the hydroxyl (or
acetoxyl) group and the reduction of the carbonyl group,
respectively. On the basis of this result, a number of
preparative electrolyses were attempted5 using a mercury
pool cathode in a divided cell, with aqueous ethanol as
solvent. However, in most of the cases studied, cleavage
of the R-substituent could not be accomplished without
concomitant reduction of the ketone carbonyl group;
compound 3b gave diol 4 as the only isolated product, in
a yield of over 90%.
Electr och em ica l Dea cetoxyla tion :
Syn th esis of 11-Ketotigogen in
D. J . Mazur,*,† P. M. Kendall,† C. W. Murtiashaw,‡
P. Dunn,‡ S. L. Pezzullo,‡ S. W. Walinsky,‡ and
J . B. Zung‡
The Electrosynthesis Company, Inc.,
Lancaster, New York 14086-9779, and Pfizer, Inc.,
Central Research Division, Groton, Connecticut 06340
Received J uly 24, 1995
In tr od u ction
11-Ketotigogenin ((3â,5R,25R)-3-hydroxyspirostan-11-
one, 1a ) is a useful potential intermediate for the
synthesis of corticosteroids and related compounds. It
has traditionally been prepared from naturally-occurring
hecogenin (2), with a variety of synthetic schemes1,2 being
used to establish the desired functionality at C-11. The
best results have been obtained by reduction of readily
available3 11-ketorockogenin diacetate (3b) with calcium
in liquid ammonia.2 Although the calcium/ammonia
reduction proceeds in good yield, it has features which
might make it unattractive for large-scale production.
These include the capital investment associated with the
necessary equipment for low-temperature reactions, as
well as safety and environmental factors associated with
the handling and recovery of liquid ammonia and the
disposal of residual calcium. For these reasons, an
alternative route from 3b to 1a was sought. An electro-
chemical approach seemed particularly attractive from
safety and environmental standpoints. In addition to the
advantage of providing mild, easily scalable reaction
conditions, the electrochemical approach could conceiv-
ably offer improved chemical selectivity.
Dissolving metal reductions of steroidal R-ketols have
been observed2,6 to be profoundly sensitive to reaction
conditions, particularly to the nature of the proton source
used. It was hypothesized that similar considerations
might apply to the electrochemical reaction and that the
desired selectivity of reduction might be achieved in a
different electrolysis medium, using high hydrogen over-
potential electrodes such as lead or carbon-based materi-
als such as vitreous carbon or high surface area graphite.7
Use of these electrode materials in place of mercury
would also be desirable from an environmental stand-
point.
Resu lts a n d Discu ssion
Initial experiments on the electroreduction of 3b were
carried out in a divided glass H-cell equipped with a
Nafion 417 cation exchange membrane, using tetraalkyl-
ammonium salts as supporting electrolytes. The cathode
materials examined included lead, vitreous carbon, and
high surface area graphite felt. In protic media such as
methanol or aqueous acetonitrile, compound 3b was
rapidly and efficiently hydrolyzed to 3a , which was
relatively inert to further reaction under the conditions
used. This hydrolysis is believed to be caused by elec-
trogenerated base and could be completely suppressed
by using a pH 5 tetraethylammonium acetate/acetic acid
buffer as the electrolyte; in experiments involving a
buffered catholyte, no reaction of 3b was observed.
In contrast to the above results, it was found that 3b
was readily reduced in aprotic media with a lead or
graphite felt cathode. Electrolyses were typically carried
out in N-methyl-2-pyrrolidinone (NMP) or NMP/THF
mixtures,8 with tetrabutylammonium tetrafluoroborate
(3) Djerassi, C.; Ringold, H. J .; Rosenkranz, G. J . Am. Chem. Soc.
1951, 73, 5513.
(4) Kabasakalian, P.; McGlotten, J . Anal. Chem. 1959, 31, 1091.
(5) Kabasakalian, P.; McGlotten, J .; Basch, A.; Yudis, M. D. J . Org.
Chem. 1961, 26, 1738.
(6) Rothman, E. S.; Wall, M. E. J . Am. Chem. Soc. 1957, 79, 3228.
(7) Tilak, B. V.; Sarangapani, S.; Weinberg, N. L. In Technique of
Electroorganic Synthesis: Weinberg, N. L., Tilak, B. V., Eds.; J ohn
Wiley & Sons: New York, 1982; Vol. III, Chapter IV.
(8) The lack of solubility of the starting material in solvents such
as DMF or acetonitrile was a limiting factor, but the material is quite
soluble in THF, and the use of this cosolvent gave the best overall
results.
† The Electrosynthesis Co., Inc.
‡ Pfizer, Inc.
(1) Djerassi, C.; Ringold, H. J .; Rosenkranz, G. J . Am. Chem. Soc.
1954, 76, 5533.
(2) Chapman, J . H.; Elks, J .; Phillipps, G. H.; Wyman, L. J . J . Chem.
Soc. 1956, 4344.
0022-3263/96/1961-0405$12.00/0 © 1996 American Chemical Society

406 J . Org. Chem., Vol. 61, No. 1, 1996
Notes
or acetate as the electrolyte. Under these conditions,
complete consumption of the starting material was
observed after passage of 2-4 F/mol of starting material
added; at this point the catholyte contained primarily a
mixture of the desired product 1a and its corresponding
acetate 1b, along with a small amount (5-10%) of the
deacetylated material 3a . Although 1b was slowly
converted to 1a upon continued electrolysis,9 it proved
more convenient to stop the electrolysis as soon as the
starting material was consumed and hydrolyze the
acetate in situ by addition of a catalytic amount of
potassium hydroxide in methanol to the catholyte. After
neutralization of the mixture, the steroidal materials
were separated from residual solvent/electrolyte by re-
moving the majority of the volatile solvents under
reduced pressure and diluting the residue with several
volumes of water. In this way, 74-97% yields of crude
product assaying at 65-80% by HPLC were typically
obtained.
conventional flag anode, and compound 1b was formed
smoothly. The hydrolysis of 1b to 1a was somewhat
complicated by the presence of magnesium salts in the
mixture, and a substantial excess of potassium hydroxide
(2.25 equiv based on 3b) was required to accomplish this
transformation. However, this fact would not preclude
the possible use of a sacrificial anode with other sub-
strates.
The electrochemical methodology utilized in this study
is straightforward, and it is felt that it could easily be
extended to other R-ketol acetates and possibly also to
ketones containing other R-substituents (e.g. halogen).
It provides a selective method of removing the substitu-
ent under mild, neutral conditions at near ambient
temperature and as such could be a potentially useful
tool for organic synthesis.
Ta ble 1. Electr och em ica l Dea cetoxyla tion of
11-Ketor ock ogen in Dia ceta te (3b) to 11-Ketotigogen in
(1a )
The nonaqueous electrolysis process was initially car-
ried out in a divided glass cell and was also successfully
scaled up to a small (10 cm2 electrode area) filter-press
type electrochemical flow cell with no loss of current
efficiency. In the course of these experiments it was
observed that a polymeric coating was gradually formed
on the surface of the anode, except when an easily
oxidizable species such as acetate ion was present in the
anode compartment; the polymerization problem was
most easily avoided by use of tetrabutylammonium
acetate as the electrolyte. This observation also sug-
gested the possibility of carrying out the reaction in an
undivided cell. The anodic oxidation of acetate can take
place in a variety of ways, depending on the conditions;10
two possible anodic processes are shown below.
concn
of
charge
passed
yield
of
electrolytea
3b (M) (F/mol) 1a (%)b
0.2 M Bu4NBF4 in NMPc
0.075
0.15
0.2
0.2
0.2
3.0d
2.6d
3.25g
4.0g
3.5j
50
79
54
76
70
0.4 M Bu4NBF4 in NMP/THF (1:1)e
0.3 M Bu4NOAc in NMP/THF (1:1)f
0.3 M Bu4NOAc in NMP/THF (1:1)h
0.1 M Bu4NBr in NMP/THF (1:1)i
Graphite felt cathode. b Based on HPLC assay of crude isolated
a
product. c Divided glass cell; catholyte as specified, anolyte 0.2 M
d
Bu4NOAc in NMP. The electrolysis was continued until 1b was
completely converted to 1a . e Divided glass cell; catholyte as
specified, anolyte 0.4 M Bu4NOAc in NMP/THF. f Divided flow cell.
g
The product at the end of the electrolysis was primarily 1b, so
the catholyte was treated with 0.25 equiv of KOH in methanol.
Undivided glass cell. Undivided glass cell with Mg anode. j The
h
i
product at the end of the electrolysis was primarily 1b, so the
catholyte was treated with 1.25 equiv of KOH in methanol.
2AcO- f 2CO2 + CH3CH3 + 2e-
2AcO- f CO2 + CH3OAc + 2e-
Exp er im en ta l Section
Gen er a l. Melting points are uncorrected. HPLC analyses
were performed using a 3.9 × 300 mm C18 (4 µm) column with
refractive index detection. Most preparative electrolyses were
carried out in either an undivided “beaker-type” glass cell of
about 100 mL capacity or a divided “H-cell” having a working
electrode compartment of about 100 mL capacity and a counter
electrode compartment of about 50 mL capacity. The flow cell
experiment was conducted in a Micro Flow Cell12 from Electro-
Cell Systems AB, Sweden. The divided cell experiments utilized
The cathodic transformation of 3b to 1b produces an
equimolar amount of acetate in a two-electron process,
and would thus help to minimize loss of acetate. On a
larger scale, replenishment of acetate could be ac-
complished by pH-controlled addition of acetic acid. The
transformation of 3b to 1b (and ultimately to 1a ) in an
undivided cell using tetrabutylammonium acetate as the
electrolyte was successfully carried out, with no evidence
of anodic degradation of the products. No anode fouling
or increase in cell voltage was observed over the course
of the run.
Although the undivided cell process utilizing tetrabu-
tylammonium acetate is both efficient and simple to
operate, it suffers from the drawback of requiring a
relatively expensive electrolyte. A possible alternative
to this is the use of a sacrificial anode,11 which has been
employed for a number of otherwise difficult nonaqueous
electroorganic reductions; this approach requires only a
minimal amount of added electrolyte. The reduction of
3b was carried out using a magnesium rod in place of a
a Nafion 417 cation exchange membrane as a separator.
A
platinum flag anode was used for most glass cell experiments,
and a platinum-clad niobium sheet was used as an anode in the
flow cell. High surface area graphite felt electrodes were
prepared from 0.6 cm thick graphitized polyacrylonitrile fiber12
(surface area 0.5 m2/g).
Ma ter ia ls. Commercial quaternary ammonium salts and
reagent grade solvents were used without further purification.
P r ep a r a tion of 11-Ketotigogen in (1a ) in a n Un d ivid ed
Cell. An undivided glass cell was assembled with a platinum
flag anode (5 cm2) and a graphite felt flag cathode (5 cm2 × 0.6
cm thick). A solution of 6.33 g of tetrabutylammonium acetate
and 7.43 g (14 mmol) of keto ester 3b in 70 mL of THF/NMP
(1:1 by volume) was placed in the cell. The solution was
electrolyzed, with no external heating or cooling, at a constant
current of 125 mA (25 mA/cm2). Over the course of the run the
temperature of the solution rose to about 40 °C. The run was
continued until a total charge of 4.1 F/mol had been passed; by
this time not more than a trace of starting material was noted
by TLC. The electrolyte was diluted with a solution of 596 mg
of potassium hydroxide in 120 mL of methanol, and the resultant
(9) The relative amounts of 1a and 1b varied greatly from one
experiment to another; inasmuch as the hydrolysis of 1b to 1a is
presumably caused by cathodically-generated base, it could be strongly
dependent on the amount of adventitious moisture present in the
system.
(10) Torii, S.; Tanaka, H. In Organic Electrochemistry, 3rd ed.; Lund,
H., Baizer, M. M., Eds.: Marcel Dekker: New York, 1991; Chapter
14.
(12) High surface area graphite felt (ESC Grade GF-S6) and
ElectroCell electrochemical cells are available from the Electrosyn-
thesis Co., Inc.
(11) Chaussard, J .; Folest, J . C.; Nedelec, J . Y.; Perichon, J .; Sibille,
S.; Troupel, M. Synthesis 1990, 369 and references cited therein.

Notes
J . Org. Chem., Vol. 61, No. 1, 1996 407
mixture was stirred at ambient temperature until all of 1b was
converted to 1a (about 4 h). The mixture was then neutralized
by addition of 2 mL of acetic acid and concentrated under
reduced pressure, and the residue was added over about 15 min
to 350 mL of vigorously-stirred water containing 2 mL of acetic
acid. The resultant suspension was stirred overnight and
filtered, and the product was washed with water and dried under
vacuum at 40 °C to yield 5.87 g of a crude product mixture
consisting of 78% of the desired product 1a , together with 9% of
3a and 3% of residual 1b. Purification of a 2 g portion of the
crude material by flash chromatography13 gave 1.28 g (62%
yield14) of material of 97% HPLC assay and a mp 215-219 °C
(lit.1 mp 223-226 °C), identical to authentic 1a prepared by the
literature method.2
P r ep a r a tion of 11-Ketotigogen in (1a ) Usin g a Sa cr ificia l
An od e. An undivided glass cell was assembled with a magne-
sium rod anode (1.5 cm diameter) surrounded by a ring (2.5 cm
internal diameter) of graphite felt supported by a cylindrical
piece of stainless steel mesh. Next, a solution of 2.90 g (9 mmol)
of tetrabutylammonium bromide and 9.55 g (18 mmol) of keto
ester 3b in 80 mL of THF/NMP (1:1 by volume) was placed in
the cell. The solution was electrolyzed at a constant current of
690 mA (25 mA/cm2 cathode current density); the cell was
immersed in a cold water bath in order to keep the mixture
temperature below about 45 °C. After passage of 2.5 F/mol, the
current was reduced to 412 mA (15 mA/cm2), and the run was
continued until another 1.0 F/mol had been passed; by this time
not more than a few percent of starting material remained by
TLC. The electrolyte was diluted with 90 mL of methanol and
treated with 1.79 g (33 mmol) of potassium hydroxide; the
mixture became somewhat viscous (presumably due to gelati-
nous insoluble magnesium salts). The resultant mixture was
stirred at ambient temperature overnight, by which time the
hydrolysis of 1b to 1a was complete. The mixture was neutral-
ized by addition of 5 mL of acetic acid; upon addition of the acid
most of the magnesium salts appeared to dissolve. The mixture
was concentrated under reduced pressure and diluted over about
1 h with 350 mL of water. The resultant suspension was stirred
for 2 h and filtered, and the product was washed with water
and dried under vacuum at 40 °C to yield 7.16 g of crude material
containing 70% of the desired product 1a . The HPLC profile of
this material was similar to that of the material obtained using
a conventional anode. Table 1 summarizes the conditions used
for electrolyses and the yields obtained for 1a .
(13) Still, W. C.; Kahn, M.; Mitra, A. J . Org. Chem. 1978, 43, 2923.
(14) The yield has not been optimized; in particular, we believe that
the amount of 3a could very likely be reduced by more rigorous drying
of the solvents and reagents.
J O9513469