Expedient synthesis of 17α,21-dihydroxy-9β,11β-epoxy- 16αmethylpregna- 1,4-diene-3,20-dione 21-acetate from prednisolone utilising a novel Mattox rearrangement

Article abstract of DOI:10.1016/j.steroids.2013.09.008

A six step transformation of prednisolone to 17a,21-dihydroxy-9b,11b-epoxy- 16a-amethylpregna-1,4- diene-3,20-dione 21-acetate has been achieved in 13% unoptimised yield. Novel conditions for effecting a Mattox rearrangement and double dehydration of prednisolone were identified. Enhanced knowledge on the oxidation of silyl D19,20-enol ethers and structural factors that impact the success of the oxidation are also presented.

Full text of DOI:10.1016/j.steroids.2013.09.008

Steroids 78 (2013) 1281–1287
Contents lists available at ScienceDirect
Steroids
journal homepage: www.elsevier.com/locate/steroids
Expedient synthesis of 17a,21-dihydroxy-9b,11b-epoxy-16a-
methylpregna-1,4-diene-3,20-dione 21-acetate from prednisolone
utilising a novel Mattox rearrangement
⇑
David G. Hulcoop, Peter D.P. Shapland
Second Generation API UK, Product Development, GlaxoSmithKline R&D, Gunnels Wood Road, Stevenage, Hertfordshire SG1 2NY, United Kingdom
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 28 June 2013
Received in revised form 10 September 2013
Accepted 15 September 2013
Available online 27 September 2013
A six step transformation of prednisolone to 17a,21-dihydroxy-9b,11b-epoxy-16a-amethylpregna-1,4-
diene-3,20-dione 21-acetate has been achieved in 13% unoptimised yield. Novel conditions for effecting
a Mattox rearrangement and double dehydration of prednisolone were identified. Enhanced knowledge
19,20
on the oxidation of silyl
are also presented.
D
-enol ethers and structural factors that impact the success of the oxidation
Ó 2013 Elsevier Inc. All rights reserved.
Keywords:
Mattox rearrangement
Conjugate addition
Oxidation of silyl enol ether
1
. Introduction
,21-Dihydroxy-9b,11b-epoxy-16
few synthetic steps and, potentially, quick introduction into the
supply chain.
17a
a-methylpregna-1,4-diene-
3,20-dione 21-acetate 1 is a key synthetic intermediate in the syn-
2
. Experimental
thesis of a range of medicinal glucocorticoids, such as mometasone
furoate 2 [1], dexamethasone 3 [2], flumethasone 4 and fluticasone
propionate 5, Fig. 1. These glucocorticoids see widespread use as
therapeutic agents for the treatment of inflammatory skin disorders
and respiratory diseases. As such, a robust and economical synthesis
of this key intermediate is important for delivering a resilient and
economically viable supply chain for these medicines. The vast
majority of bulk steroids are synthesized from steroidal sapogenins
in multi-step syntheses [3]. Quicker and cheaper routes to this key
synthetic intermediate would offer a substantial business advantage.
2
.1. General procedures
All reactions were carried out under a nitrogen atmosphere. Sol-
vents and reagents were of analytical grade purchased from com-
mercial sources and used without any purification or drying.
and C NMR spectra were acquired on a Bruker spectrometer in
CDCl or DMSO-d at frequencies of 400 and 100 MHz. Chemical
shifts are given as d ppm values relative to residual un-deuterated
solvent. High-resolution mass spectra were recorded on a linear
ion trap combined with a Fourier transform ion cyclotron reso-
nance mass spectrometer using an electrospray ionisation source
operated in positive ion mode. IR spectra were recorded as solids.
HPLC chromatograms were recorded on either: a C18(2) column
at 40 °C; eluent gradient: 100% 0.05% v/v TFA in water to 95%
1
H
13
3
6
As part of an assessment of improved routes to 17
a,21-dihy-
droxy-9b,11b-epoxy-16 -methylpregna-1,4-diene-3,20-dione 21-
a
acetate to support GlaxoSmithKline’s steroid portfolio, a route from
prednisolone (6) was considered. While prednisolone itself is typ-
ically derived from steroidal sapogenins [4], the routes to access it
and the required biotransformations are very well developed
resulting in its general availability from numerous suppliers [5].
Another advantage of utilizing prednisolone as the starting mate-
rial for this approach is that the synthesis is likely to be consider-
ably shorter compared to routes from sapogenins. This would
result in the requirement for process optimization of relatively
0
.05% v/v TFA in MeCN over 8 min; or a C18 column at 60 °C; elu-
ent gradient: 100% 0.05% v/v TFA in water to 95% 0.05% v/v TFA in
MeCN over 2.5 min.
2
2
.2. Chemical syntheses
.2.1. Mattox rearrangement product 9
⇑
Corresponding author. Tel.: +44 1438 763658.
E-mail address: peter.z.shapland@gsk.com (P.D.P. Shapland).
Prednisolone (2.0 g) was dissolved in tetrahydrofuran (20 mL)
and cooled to 0 °C in an ice bath. Thionyl chloride (3 equiv.) was
0
039-128X/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.steroids.2013.09.008

1
282
D.G. Hulcoop, P.D.P. Shapland / Steroids 78 (2013) 1281–1287
O
O
O
O
OAc
OH
Cl
O
OH
HO
HO
O
OH
O
H
H
H
H
Cl
H
F
H
O
O
O
(
1)
Mometasone furoate (2)
Dexamethasone (3)
This
F
O
O
work
S
O
OH
OH
O
HO
HO
O
OH
OH
HO
H
H
H
F
F
H
F
F
H
H
H
O
O
O
Prednisolone (6)
Flumethasone (4)
Fluticasone propionate (5)
Fig. 1. Key medicines synthesized via 17a,21-dihydroxy-9b,11b-epoxy-16a-methylpregna-1,4-diene-3,20-dione 21-acetate 1.
added slowly over 10 min and the mixture was stirred for 30 min
at 0 °C. Triethylamine (15.5 equiv.) was added drop wise maintain-
ing the reaction below 5 °C and the reaction was stirred for an
additional 15 min. The reaction was allowed to warm to ambient
temperature for 1 h. Saturated aqueous sodium chloride (20 mL)
was added followed by water (5 mL) to dissolve the precipitated
solid and generate a biphasic mixture. The upper organic layer
was concentrated to approximately 6 mL and water (20 mL) was
added drop wise at ambient temperature and the resultant slurry
was aged for 1 h at ambient temperature. The solid was collected
by filtration, washed twice with water (5 mL) and dried under vac-
34.3, 33.4, 31.9, 26.7, 20.5, 15.4; HRMS (ESI): m/z calcd for
(M+H)+, 367.1904; found, 367.1914.
23 27 4
C H O
2.2.3. Alkene 11
Copper(II) chloride (0.051 g, 0.382 mmol) was slurried in tetra-
hydrofuran (10 mL) and cooled to <3 °C. A tetrahydrofuran solution
of methyl magnesium chloride (3 M, 0.332 mL, 0.996 mmol) was
added to the chilled slurry over 1 min keeping the contents
<3 °C. A solution of 7 (1.00 g, 2.73 mmol) in tetrahydrofuran
(
7.5 mL) was added over 2 min keeping the contents <3 °C. Tri-
methylchlorosilane (0.698 mL, 5.46 mmol) was added in one por-
tion followed by more solution of methyl magnesium chloride in
tetrahydrofuran (3 M, 1.137 mL, 3.41 mmol) which was added
drop wise over 23 min keeping the contents <2.0 °C. After
1
uum at 40 °C to yield 9 (1.14 g, 63%). H NMR (400 MHz, DMSO-
d
(
6
): d 9.63 (s, 1H), 7.87 (s, 1H), 7.45 (d, J = 10.2 Hz, 1H, 1-H), 6.20
dd, J = 10.1 and 1.9 Hz, 1H, 2-H), 6.04 (s, 1H, 4-H), 5.60 (d,
J = 5.8 Hz, 1H, 11-H), 3.00 (dd, J = 18.2 and 8.4 Hz, 1H), 2.81–2.60
m, 3H), 2.52–2.40 (m, 2H), 2.30–2.14 (m, 2H), 2.02–1.91 (m,
1
2 min, the reaction was quenched with triethylamine (0.856 mL,
(
6
.14 mmol) followed by saturated aqueous ammonium chloride
1
H), 1.52–1.05 (m, 4H), 1.44 (s, 3H, 19-CH
3
), 0.97 (s, 3H, 18-CH
): d 187.3, 185.1, 167.4, 155.5, 146.8,
43.2, 143.0, 126.4, 122.9, 120.6, 51.1, 45.7, 43.8, 37.7, 34.9, 34.5,
25 3
1.3, 26.3, 25.7, 25.6, 15.1; HRMS (ESI): m/z calcd for C21H O
3
);
(
(
5 mL). This addition was highly exothermic (0–23 °C). Water
3 mL) was added to dissolve the solids and the quenched mixture
1
3
C NMR (100 MHz, CDCl
3
1
3
was stirred vigorously for at least 30 min. The layers were sepa-
rated and the blue aqueous layer was extracted with heptane
(
M+H)+, 325.1804; found, 325.1793.
(
2 ꢀ 4 mL). The organic layers were dried through sodium sulfate
and concentrated to dryness to give crude silyl enol ether
(1.20 g). The crude silyl enol ether was dissolved in dichlorometh-
ane (10 mL) and cooled to <3 °C. A solution of peracetic acid in ace-
tic acid (32%, 0.566 mL, 2.73 mmol) was added drop wise over
4 min keeping the contents <4 °C. After 19 h, the reaction was
quenched with aqueous hydrochloric acid (2 N, 3 mL) and stirred
vigorously. The organic layer was washed with aqueous potassium
carbonate (25%, 4 mL) and concentrated to dryness to give a pale
yellow foam. tert-Butylmethyl ether (10 mL) was added and the
mixture was sonicated to cause crystallisation. The slurry was
heated to reflux and stirred to break down any larger lumps. The
material gummed out and so was concentrated to dryness to give
crude product as a pale yellow foam (1.1079 g). The crude material
was analysed by HPLC and LCMS. This analysis indicated a signifi-
cant impurity isomeric to protonation at C-17 but not the same as
for protonation of the silyl enol ether. Also present was a signifi-
cant amount of di-oxygenated material, potentially a 9,11-epoxide.
The sample was analysed by TLC and an aliquot of foam (ꢁ205 mg)
was purified by prep-TLC (Macherey–Nagel SIL G-200 UV254
plates; 20 ꢀ 20 cm ꢀ 2 mm) eluted with 1:1 v/v ethyl acetate/hep-
tane. Bands corresponding to Spots 1, 2, 3 and 4 were scraped off
and put into a Bond Elut filter. The material was extracted with
ethyl acetate (25 mL) followed by 1:1 v/v ethyl acetate/dichloro-
2
.2.2. Tetraene 7
Mattox rearrangement product 9 (2.0 g) was slurried with N,N-
dimethylaminopyridine (1.1 equiv.) in dichloromethane (10 mL) at
ambient temperature. Acetic anhydride (1.1 equiv.) was added and
the mixture was stirred at ambient temperature for 30 min to pro-
vide a solution of acetate 10. The dichloromethane was removed
under vacuum and the residue was dissolved in tetrahydrofuran
(
10 mL). 1,8-Diazabicyclo[5.4.0]undec-7-ene (DBU, five drops, cat-
alytic) was added and the reaction mixture was heated to 60 °C for
h. The reaction mixture was cooled to ambient temperature,
3
washed twice with saturated sodium chloride (20 mL) and concen-
trated to approximately 4 mL. Water (40 mL) was added over
3
0 min and the resultant slurry was aged at ambient temperature
for 1 h then isolated by filtration and dried under vacuum to yield 7
1
(
1.92 g, 85%). H NMR (400 MHz, CDCl
3
): d 7.20 (d, J = 10.3 Hz, 1H,
1
6
-H), 6.75 (m, 1H, 16-H), 6.27 (dd, J = 10.3 and 1.7 Hz, 1H, 2-H),
.06 (s, 1H, 4-H), 5.55 (m, 1H, 11-H), 5.03 (d, J = 16.1 Hz, 1H, 21-
0
H), 4.88 (d, J = 15.9 Hz, 1H, 21-H ), 2.37–2.73 (m, 5H), 2.05–2.23
(
m, 3H), 2.16 (s, 3H, 21-OAc), 1.47–1.57 (m, 1H), 1.42 (s, 3H, 19-
13
CH
3
), 1.15–1.29 (m, 1H), 0.91 (s, 3H, 18-CH
): d 190.3, 186.3, 170.4, 166.3, 154.5, 150.2,
43.4, 143.4, 127.3, 123.9, 121.1, 65.6, 52.3, 46.0, 45.2, 37.4, 34.4,
3
);
C NMR
(
100 MHz, CDCl
3
1

D.G. Hulcoop, P.D.P. Shapland / Steroids 78 (2013) 1281–1287
1283
methane (25 mL). The filtrates were concentrated to dryness and
the residues were dissolved in dichloromethane, filtered through
cotton wool and re-concentrated to give:
(2 mL) and sucked dry to give 18 (1.982 g, 44%). 1H NMR
(400 MHz, CDCl ): d 7.15 (d, J = 10.3 Hz, 1H, 1-H), 6.24 (dd,
J = 10.3 and 1.7 Hz, 1H, 2-H), 6.03 (s, 1H, 4-H), 5.50 (d, J = 5.9 Hz,
H, 11-H), 2.95–3.13 (m, 1H), 2.87 (s, 1H), 2.64 (dd, J = 13.7 and
3
1
Band 1 (5.2 mg): Analysis indicated only 79% purity with some
contamination from Band 2 material. This material appears to
be consistent with protonation of the silyl enol ether from the
b-face, indicated by structure 14. While no strong NoE correla-
tions support this, the presence of two protonated species visi-
ble in the HPLC of the silyl enol ether indicate that
diastereoisomeric protonations can occur in the presence of
acid (i.e. 13 and 14). This is the minor of those two species
and so would be consistent with protonation from the most
hindered position. Conceivably, the material could be consistent
with addition of the methyl group to C-16 with a b-orientation.
Band 2 (24.2 mg): Analysis indicated a 61:32 mixture of 11 and
3.9 Hz, 1H), 2.52–2.61 (m, 1H), 2.04–2.43 (m, 3H), 2.23 (s, 3H,
21-H3), 1.86 (dd, J = 19.8 and 10.3 Hz, 1H), 1.35–1.78 (m, 3H),
1.38 (s, 3H, 19-CH
0.90 (d, J = 7.1 Hz, 3H, 16-CH
(100 MHz, CDCl ): d 211.5, 186.3, 166.8, 154.6, 142.6, 127.2,
3
), 1.18 (ddd, J = 26.2, 13.0 and 4.4 Hz, 1H),
13
3
3
), 0.78 (s, 3H, 18-CH ); C NMR
3
123.7, 120.6, 90.6, 48.0, 47.4, 45.9, 36.6, 36.4, 34.7, 33.5, 32.6,
32.2, 27.6, 26.7, 15.4, 14.5.
The liquors were analysed by TLC. Concentrated liquors
(ꢁ580 mg) were purified by prep-TLC (Macherey–Nagel SIL G-
200 UV254 plates; 20 ꢀ 20 cm ꢀ 2 mm) eluted with 1:1 v/v ethyl
acetate/heptane. Bands corresponding to Spots 3, 4 and 6 were
scraped off and put into a Bond Elut filter. For each band, the mate-
rial was extracted with ethyl acetate (25 mL) followed by 1:1 v/v
ethyl acetate/dichloromethane (25 mL). The filtrates were concen-
trated to dryness. The residues were dissolved in dichloromethane,
1
4.
Band 3 (47.8 mg): This material was still somewhat crude but
with one major component. Analysis indicated a m/z = 415
amu which indicated extra oxidation. The minor components
were m/z = 399 and isomeric with main product. The NMR data
filtered through cotton wool and re-concentrated to give 20
1
(79.1 mg) as a colourless foam. H NMR (400 MHz, CDCl
3
): d 7.15
indicated the extra oxygen to be present as a 9
The corresponding 9b,11b-epoxide (1) is known and the NMR
is not consistent with this. An NoE interaction with 7b-H indi-
a
,11
a
-epoxide.
(d, J = 10.3 Hz, 1H, 1-H), 6.25 (dd, J = 10.3 and 1.7 Hz, 1H, 2-H),
6.05 (s, 1H, 4-H), 5.47 (m, 1H, 11-H), 2.57–2.69 (m, 1H), 2.37–
2.45 (m, 1H), 2.10–2.37 (m, 4H), 2.21 (s, 3H, 21-H3), 1.96 (dd,
J = 19.8 and 9.8 Hz, 1H), 1.69–1.88 (m, 3H), 1.52–1.63 (m, 2H),
cated a b-orientation of 11-H and hence the
the oxygen implying structure 12.
a-orientation of
1.41 (s, 3H, 19-CH ), 0.95
3
), 1.12–1.25 (m, 1H), 0.99 (s, 3H, 18-CH
3 3
(d, J = 7.1 Hz, 3H, 16-CH ); C NMR (100 MHz, CDCl ): d 212.4,
3
1
3
Band 4 (11.5 mg): Analysis indicated a mixture of two com-
pounds in a ratio of 29:64 with corresponding masses of 373/
186.2, 166.5, 154.4, 143.7, 127.3, 123.9, 120.4, 92.9, 46.8, 46.1,
45.9, 37.4, 34.5, 34.3, 33.0, 32.1, 30.3, 29.7, 26.7, 14.8, 14.2.
4
15 amu. The more massive (major component) is isomeric
with 12 and so was tentatively assigned as the C-17 epimer,
5. The minor component is concordant with and tentatively
1
Also isolated was crude 19 (19.9 mg) as a colourless oil. H NMR
1
(400 MHz, CDCl
J = 10.3 Hz, 1H, 1-H), 6.27 (m, 1H, 2-H), 6.05 (s, 1H, 4-H), 5.50 (d,
J = 5.6 Hz, 1H, 11-H), 2.11 (s, 3H, 21-H3), 1.38 (s, 3H, 19-CH ),
3
, relevant peaks for identification only): d 7.16 (d,
assigned as 16.
3
2.2.4. Alkene analogue 18
3 3
0.94 (d, J = 6.8 Hz, 3H, 16-CH ), 0.64 (s, 3H, 18-CH ).
Copper(II) chloride (0.246 g, 1.829 mmol) was slurried in tetra-
hydrofuran (40.0 mL) and cooled to <3 °C. A tetrahydrofuran solu-
tion of methyl magnesium chloride (3 M, 1.590 mL, 4.77 mmol)
was added to the chilled slurry over 3 min keeping the contents
2.2.5. Epoxide analogue 22
Copper(II) chloride (24.08 mg, 0.179 mmol) was slurried in tet-
rahydrofuran (4.5 mL) and cooled to <3 °C. A tetrahydrofuran solu-
tion of methyl magnesium chloride (3 M, 0.156 mL, 0.467 mmol)
was added to the chilled slurry over 3 min keeping the contents
<3 °C. A slurry of 21 (415.1 mg, 1.279 mmol) in tetrahydrofuran
(3.38 mL) was added over 1 min keeping the contents <3.5 °C. Tri-
methylchlorosilane (0.327 mL, 2.56 mmol) was added in one por-
tion. More tetrahydrofuran solution of methyl magnesium
chloride (3 M, 0.533 mL, 1.599 mmol) was added drop wise over
24 min keeping the contents <3.5 °C. After a further 21 min the
reaction was quenched with triethylamine (0.401 mL, 2.88 mmol)
followed by saturated aqueous ammonium chloride (2 mL). Water
(1 mL) was added to dissolve the solids and the quenched mixture
was stirred vigorously for at least 30 min. The layers were sepa-
rated and the blue aqueous layer was extracted with heptanes
(2 ꢀ 3 mL). The organic layers were dried through sodium sulfate
and concentrated to dryness to give intermediate silyl enol ether
<
3 °C. The slurry was cooled to ꢂ5 °C. A slurry of 17 (4.03 g,
3.07 mmol) in tetrahydrofuran (30.0 mL) was added over 1 min
keeping the contents <1 °C. Trimethylchlorosilane (3.34 mL,
6.1 mmol) was added in one portion. More tetrahydrofuran solu-
1
2
tion of methyl magnesium chloride (3 M, 5.44 mL, 16.33 mmol)
was added drop wise over 28 min keeping the contents <2.5 °C.
After 25 min, the reaction was quenched with triethylamine
(
4.10 mL, 29.4 mmol) followed by saturated aqueous ammonium
chloride (20 mL). This addition was highly exothermic (0–23 °C).
Water (10 mL) was added to dissolve the solids and the quenched
mixture was stirred vigorously for at least 30 min. The layers were
separated and the blue aqueous layer was extracted with heptane
(
2 ꢀ 12 mL). The organic layers were dried through sodium sulfate
and concentrated to dryness to give intermediate silyl enol ether
(
1
5.55 g). H NMR analysis indicated one major component that ap-
1
peared to be consistent with the silyl enol ether. The crude silyl
enol ether was dissolved in dichloromethane (40.0 mL) and cooled
to <3 °C. A solution of peracetic acid in acetic acid (32%, 2.71 mL,
(513.3 mg). H NMR analysis indicated one major component that
appeared to be consistent with the silyl enol ether. The crude silyl
enol ether was dissolved in dichloromethane (4.50 mL) and cooled
to <3 °C. A solution of peracetic acid in acetic acid (32%, 0.269 mL,
1.279 mmol) was added drop wise over 10 min keeping the con-
tents <3.5 °C. After 1 h, more peracetic acid solution in acetic acid
(32%, 0.054 mL, 0.813 mmol) was added in one portion and the
mixture was allowed to warm to ambient temperature. After
20 min the reaction was quenched with aqueous hydrochloric acid
(2 N, 7 mL) and stirred vigorously for 20 min. The organic layer was
washed with aqueous potassium carbonate solution (25%, 10 mL)
and concentrated to dryness to give a pale yellow solid. The solid
was slurried in acetone (2 mL) and heated to reflux. The resultant
1
<
(
3.07 mmol) was added drop wise over 4 min keeping the contents
3.5 °C. After 50 min, more peracetic acid solution in acetic acid
32%, 0.678 mL, 3.27 mmol) was added in one portion and the mix-
ture was allowed to warm to ambient temperature. After 105 min
the reaction was quenched with hydrochloric acid (2 N, 12 mL) and
stirred vigorously for 15 min. The organic layer was washed with
2
5% aqueous potassium carbonate solution (12 mL) and concen-
trated to dryness to give a pale yellow gum. The gum was tritu-
rated with acetone (8 mL) and stirred at ambient temperature.
The solids were collected by filtration and washed with acetone

1
284
D.G. Hulcoop, P.D.P. Shapland / Steroids 78 (2013) 1281–1287
slurry was allowed to cool to ambient temperature and the solids
<3 °C. A solution of epoxide 24 (0.4996 g, 1.306 mmol) in tetrahy-
drofuran (3.75 mL) was added to the resulting mixture over 2 min
keeping the contents <3 °C. Trimethylchlorosilane (0.334 mL,
2.61 mmol) was added in one portion followed by more tetrahy-
drofuran solution of methyl magnesium chloride (3 M, 0.544 mL,
1.633 mmol), which was added drop wise over 30 min keeping
the contents <2.5 °C. After 75 min, the reaction was quenched with
triethylamine (0.41 mL, 2.94 mmol) followed by saturated ammo-
nium chloride (2.5 mL). This addition was highly exothermic (0–
23 °C). Water (1.5 mL) was added to dissolve the colourless solids
and the quenched mixture was stirred vigorously for at least
were collected by filtration, washed with acetone (0.5 mL) and
1
sucked dry to give 22 (262.3 mg, 57%). H NMR (400 MHz, CDCl
3
):
d 6.58 (d, J = 10.0 Hz, 1H, 1-H), 6.17 (dd, J = 10.0 and 1.7 Hz, 1H, 2-
H), 6.13 (s, 1H, 4-H), 3.19 (s, 1H, 11-H), 2.95–3.18 (m, 1H), 2.90 (s,
1
1
H), 2.60–2.71 (m, 1H), 2.17–2.54 (m, 4H), 2.22 (s, 3H, 21-H3),
.76–1.88 (m, 1H), 1.28–1.70 (m, 4H), 1.43 (s, 3H, 19-CH
3
), 1.00
13
(s, 3H, 18-CH
100 MHz, CDCl
3
), 0.86 (d, J = 7.1 Hz, 3H, 16-CH
3
).
C NMR
(
3
): d 211.1, 185.9, 164.8, 152.0, 128.0, 125.1, 90.3,
6
2
6.2, 62.9, 48.0, 47.9, 44.1, 35.3, 34.1, 33.6, 30.5, 29.7, 29.1, 27.7,
3.7, 18.3, 14.6.
3
0 min. The layers were separated and the blue aqueous layer
2
.2.6. Bromoformate 23
Tetraene 7 (5.08 g, 13.86 mmol) was slurried in N,N-dimethyl-
was extracted with heptane (2 ꢀ 2 mL). The organic layers were
dried over sodium sulfate and concentrated to dryness to give
1
formamide (17.5 mL). Perchloric acid (70%, 0.536 mL, 6.24 mmol)
was added and an exotherm from ca 20 to 27 °C was noted. The
mixture was allowed to cool to ambient temperature again. 1,3-Di-
bromo-5,5-dimethylhydantoin (DBH, 3.37 g, 11.78 mmol) was
added in one portion. After an initiation period of ca 30 s, an exo-
therm to 35 °C occurred with dissolution of the solids. After stirring
overnight, additional 1,3-dibromo-5,5-dimethylhydantoin (0.24 g,
intermediate crude silyl enol ether (0.6768 g). H NMR analysis
indicated one major component that was consistent with the silyl
enol ether. The crude silyl enol ether was dissolved in dichloro-
methane (5 mL) and cooled to <3 °C. A solution of peracetic acid
in acetic acid (32%, 0.271 mL, 1.306 mmol) was added drop wise
over 7 min keeping the contents <4 °C and the mixture was al-
lowed to warm to ambient temperature. After stirring for 18 h
the reaction was quenched with aqueous hydrochloric acid (2 N,
1.5 mL) and stirred vigorously for 15 min. Stirring was stopped
and the layers were separated. The organic layer was washed with
aqueous potassium carbonate (25%, 2.0 mL) and concentrated to
dryness to give a yellow foam. tert-Butylmethyl ether (3 mL) was
added and the mixture was sonicated to cause crystallisation.
The slurry was aged at ambient temperature for 1 h. The solids
were collected by filtration and washed with tert-butylmethyl
0
.839 mmol) was added in one portion. After stirring for an addi-
tional 6 h, water (50 mL) was added drop wise over 26 min and
the generated slurry was aged overnight. The solids were collected
by filtration, washed twice with water (50 mL) and dried under
1
vacuum overnight to give 23 (6.72 g, 99% yield).
H NMR
(
400 MHz, CDCl
3
): d 8.12 (s, 1H), 6.78 (d, J = 10.0 Hz, 1H, 1-H),
6
.73 (m, 1H, 16-H), 6.31 (dd, J = 10.0 and 1.7 Hz, 1H, 2-H), 6.07
(
s, 1H, 4-H), 5.87 (s, 1H, 11-H), 5.03 (d, J = 16.1 Hz, 1H, 21-H),
0
ether (1 mL) and sucked dry to give 1 (254.5 mg, 47% yield). 1
NMR (400 MHz, CDCl ): d 6.59 (d, J = 10.0 Hz, 1H, 1-H), 6.19 (dd,
4
.83 (d, J = 16.1 Hz, 1H, 21-H ), 2.20–2.70 (8H, m), 2.15 (s, 3H, 21-
H
OAc), 1.68–1.95 (m, 2H), 1.60 (s, 3H, 19-CH
CH
1
4
3
), 1.16 (s, 3H, 18-
): d 190.0, 185.6, 170.3, 164.2,
3
1
3
3
); C NMR (100 MHz, CDCl
3
J = 10.0 and 1.7 Hz, 1H, 2-H), 6.14 (s, 1H, 4-H), 5.02 (d,
0
58.8, 152.2, 150.5, 143.0, 129.8, 125.4, 83.3, 75.0, 65.4, 49.7,
9.5, 46.1, 36.8, 35.1, 32.1, 30.4, 28.3, 24.7, 20.4, 18.5; HRMS
J = 17.1 Hz, 1H, 21-H), 4.83 (d, J = 17.1 Hz, 1H, 21-H ), 3.21 (s, 1H),
2.88–3.10 (m, 1H), 2.60–2.71 (m, 1H), 2.43–2.53 m, 1H), 2.28–
2.39 (m, 2H), 2.18–2.27 (m, 1H), 2.15 (s, 3H, 21-OAc), 1.76–1.88
(
ESI): m/z calcd for
C
24
H
27BrO
6
(M+H)+, 491.1064; found,
4
91.1056.
(m, 2H), 1.25–1.67 (m, 4H), 1.43 (s, 3H, 19-CH
3
), 0.93 (s, 3H, 18-
1
3
3 3
CH ), 0.88 (d, J = 7.3 Hz, 3H, 16-CH ); C NMR (100 MHz, CDCl ):
3
2.2.7. Epoxide 24
d 204.5, 186.0, 170.6, 164.8, 152.1, 128.0, 125.1, 90.5, 67.4, 66.0,
Bromoformate 23 (1.50 g, 3.05 mmol) was slurried in acetone
62.7, 48.1, 47.6, 44.1, 35.3, 34.2, 33.2, 30.4, 29.9, 29.1, 23.7, 20.4,
(
22.5 mL) and the mixture was heated to 46 °C. Aqueous potassium
17.3, 14.4; HRMS (ESI): m/z calcd for
C
24
H
31
O
6
(M+H)+,
carbonate (25%, 6.75 mL, 12.21 mmol) was added and a small
endotherm to 40 °C occurred. The stirrer was stopped and the
two layers were allowed to settle. The lower aqueous layer was re-
moved by pipette and then the organic layer was allowed to cool to
ambient temperature with stirring. When the contents had
reached ambient temperature a fine precipitate had formed. The
slurry was concentrated at ambient temperature under reduced
pressure to a volume of about 5 mL. The solids were collected by
415.2115; found, 415.2110.
3. Results and discussion
We envisaged a retrosynthesis of 1 via key tetraene 7 back to
prednisolone 6, Scheme 1. In a forward sense, the conversion of 6
to 7 requires simple dehydration of the two alcohol groups at C-
11 and C-17 to generate alkenes. Although this conversion appears
trivial, we have found no reports of the direct double dehydration
filtration, washed with acetone (3 mL) and sucked dry to give 24
1
(
0.62 g, 53% yield).
3
H NMR (400 MHz, CDCl ): d 6.66 (d,
1
,4,9(11),16
J = 1.5 Hz, 1H, 16-H), 6.60 (d, J = 10.0 Hz, 1H, 1-H), 6.17 (dd,
J = 10.0 and 1.7 Hz, 1H, 2-H), 6.13 (s, 1H, 4-H), 4.95 (d,
of 6, or a C-21 protected analogue, to give a
D
-tetraene
steroid. The forward conversion of 7 to 1 requires formation of a
0
9,11
J = 16.1 Hz, 1H, 21-H), 4.86 (d, J = 16.0 Hz, 1H, 21-H ), 3.12 (s, 1H,
D
-epoxide and conjugate addition of a methyl group followed
1
1-H), 2.60–2.80 (m, 2H), 2.30–2.55 (m, 4H), 2.15 (s, 3H, 21-
OAc), 2.01–2.14 (m, 1H), 1.33–1.70 (m, 3H), 1.46 (s, 3H, 19-CH ),
): d 190.0, 186.1,
by oxidation at C-17. This sequence could conceivably be per-
formed in two different orders.
3
1
3
1
1
6
.09 (s, 3H, 18-CH
3
); C NMR (100 MHz, CDCl
3
Dehydration of steroidal alcohols is typically achieved by a di-
rect dehydration step or a two-step sequence of activation fol-
lowed by elimination. The former can be achieved by treatment
with thionyl chloride or phosphorus oxychloride in combination
70.4, 164.7, 152.3, 152.0, 141.9, 127.9, 125.2, 77.4, 66.9, 65.5,
1.7, 45.0, 44.2, 35.3, 33.3, 32.4, 29.3, 28.8, 23.8, 20.5, 18.6; HRMS
(
27 5
ESI): m/z calcd for C23H O (M+H)+, 383.1853; found, 383.1849.
0
with pyridine [6] or N,N -sulfinyl diimidazole [7]. The second ap-
2
3
.2.8. 17
,20-dione 21-acetate 1
Copper(II) chloride (0.025 g, 0.183 mmol) was slurried in tetra-
a
,21-Dihydroxy-9b,11b-epoxy-16
a
-methylpregna-1,4-diene-
proach is usually initiated by activation with methanesulfonyl
chloride or p-tolunesulfonyl chloride followed by base induced
elimination [8]. Additionally, acetates of steroidal alcohols at C-
17 can be treated with potassium acetate in DMF at high temper-
hydrofuran (5.0 mL) and cooled to <3 °C. A tetrahydrofuran solu-
tion of methyl magnesium chloride (3 M, 0.159 mL, 0.477 mmol)
was added to the chilled slurry over 1 min keeping the contents
1
6,17
atures to provide the required
D
-alkene [9]. Of critical impor-
tance for our synthesis of 1 was the necessity for the alcohol

D.G. Hulcoop, P.D.P. Shapland / Steroids 78 (2013) 1281–1287
1285
O
O
O
OAc
OH
OAc
OH
OH
HO
O
H
H
H
H
H
H
H
O
O
O
(1)
Tetraene (7)
Prednisolone (6)
Scheme 1. Retrosynthesis of 17a,21-dihydroxy-9b,11b-epoxy-16a-methylpregna-1,4-diene-3,20-dione 21-acetate 1 to Prednisolone 6.
functionality at C-21 to be retained throughout the sequence. It
was anticipated that this would require suitable use of a protecting
group during the synthetic sequence.
We found that activation of the C-11 hydroxyl and elimination
of the activated group was very facile but that activation of the C-
proposed a number of possible mechanisms all relying on enolisa-
tions and eliminations of water although he was unable to deter-
mine the precise order of events [10]. Li and co-workers showed
that both esters were required in their work and proposed an
enolisation-hydrolysis-elimination pathway for the rearrangement
in basic conditions [11]. The thionyl chloride-triethylamine rear-
rangement described above offers an additional potential mecha-
nism, Scheme 3, whereby the cyclic sulfite undergoes a pericyclic
rearrangement with extrusion of sulfur dioxide. This potential
mechanism is precluded from the other reported rearrangements.
With Mattox rearrangement product 9 in hand we were able to
convert this rapidly to the desired tetraene 7, Scheme 4. Acylation
of 9 provided the enolacetate 10 which was immediately rear-
ranged under nucleophilic catalysis to 7 in 85% yield.
17a-hydroxyl was more problematic. Placement of a mesyl group
on the hydroxyl was not achievable using forcing conditions. We
were able to repeat the work of Salce and co-workers [9] through
preparation of prednisolone triacetate and elimination of the
16,17
1
7a-acetate to provide a
D
-alkene. No elimination of the
1
1b-acetate was observed so this method was not suitable for ra-
pid formation of 7. It is likely that the 11b-acetate is unable to
achieve the correct orientation to enable a syn-elimination.
We found that treatment of six with three equivalents of thio-
nyl chloride led to formation of two disatereoisomers of a cyclic
sulfite 8, as judged by LCMS. It was not clear whether the 11b-hy-
droxyl group had been eliminated under the reaction conditions
or under the conditions of the LCMS analysis. Upon treatment
with 15.5 equivalents of triethylamine, these rapidly provided
the Mattox rearrangement product 9 in a moderate, but unopti-
mised yield, Scheme 2. The Mattox rearrangement is typically
carried out under acidic conditions upon the parent dihydroxyac-
etone structure of the corticosteroid [10]. The same rearrange-
ment product has also been produced from 17,21-diesters of
corticosteroids under basic conditions [11]. While providing a
versatile method of functionalizing the D-ring of steroids and
the side chain attached to C-17, the Mattox rearrangement has
not been extensively used in synthesis.
At this junction, progression of 7 to 1 requires three fundamen-
9
,11
tal changes: epoxidation of the
D
-olefin, conjugate addition of a
methyl group to C-16 and oxidation of C-17. The latter two of these
are commonly combined into a single reaction sequence. The pre-
cise ordering of the transformations was something that offered
flexibility to the synthesis although the optimum way forward
was not clear at the outset. As a consequence, we elected to study
both approaches to 1 to identify the most efficient pathway.
We initially chose to examine the copper-catalysed addition of
a methyl Grignard reagent to 7, followed by trapping of the resul-
tant enolate as the trimethylsilylenol ether. Unfortunately, perace-
tic acid mediated oxidation of the silylenol ether derived from 7
gave an intractable mixture of compounds containing only small
amounts of 11. Attempted purification of the mixture failed to gen-
erate any single purified compound, yet tentative structural assign-
ments could be made which indicated significant over oxidation to
The proposed mechanisms for the classical Mattox rearrange-
ment of dihydroxyacetone structures and the base-mediated
rearrangement of di-esters of these structures differ subtly. Mattox
compounds containing a 9
a,11a-epoxide, 12 and 16. An additional
O
O
O
O
S
HO
OH
OH
O
HO
Et N,
THF
H
3
H
SOCl2,
THF
O
H
H
63%
H
H
H
H
O
O
O
Prednisolone (6)
(
9)
(
8)
Scheme 2. One pot Mattox rearrangement and elimination.
O
O
O
O
O
O
S
S
O
O
H
Et N,
THF
3
-SO2
O
O
O
H
H
H
H
H
H
O
O
Scheme 3. Potential pericyclic mechanism for Mattox rearrangement of cyclic sulfites.

1
286
D.G. Hulcoop, P.D.P. Shapland / Steroids 78 (2013) 1281–1287
O
O
HO
H
O
O
O
OAc
H
DBU,
H
Ac
DMAP,
DCM
2
O,
H
H
EtOAc
H
H
85%
O
O
H
(
10)
(7)
(9)
O
Scheme 4. Rearrangement of Mattox product 9 to tetraene 7.
O
O
O
O
OAc
OAc
OAc
OH
Cu(II)Cl , MeMgBr,
TMSCl, THF; then
peracetic acid, DCM
2
OH
H
H
+
O
H
H
H
H
O
O
O
O
(
11)
(12)
(
7)
+
O
O
OAc
+
OAc
H
H
H
H
H
H
O
O
(
(
13)
+
(14)
O
OAc
OH
OH
OH
H
O
+
H
H
H
O
15)
(16)
Scheme 5. Unsuccessful conversion of 7 to alkene 11.
O
O
O
Cu(II)Cl , MeMgBr,
2
TMSCl, THF; then
peracetic acid, DCM
R
OH
H
H
H
H
H
H
O
O
O
(17)
(
20); 1.8%
(
18) R=OH; 44%
19) R=H; 0.4%
(
O
O
Cu(II)Cl , MeMgBr,
TMSCl, THF; then
peraceticacid, DCM
2
OH
O
O
H
H
H
H
O
O
(21)
(22); 57%
Scheme 6. Conjugate addition–oxidation of simpler analogues 17 and 21.
C-17-epimer of this product 15 was also detectable and products
that were protonated at C-17 rather than oxidized, both on the
corresponding trimethylsilyl enol ether as a single diastereomer.
For this analogous silyl enol ether the rate of oxidation with per-
acetic acid was more rapid, compared to 7, due to the lack of steric
crowding of the silyl enol ether moiety by the C-21 acetoxy group.
Again, though, clean oxidation was not observed resulting in a
number of products, including the expected product 18 along with
a- and b-faces could be detected, 13 and 14, Scheme 5.
This unexpected difficulty prompted study of a simpler model
system using 9,11-alkene analogue 17, which underwent clean
methyl Grignard addition to C-16 and trapping of the

D.G. Hulcoop, P.D.P. Shapland / Steroids 78 (2013) 1281–1287
1287
O
O
O
OAc
OAc K CO ,water
H
O
2
3
H
DBH, HClO
DMF
4
acetone
H
H
53%
Br
H
9
9%
O
O
(
7)
(23)
DBH= 1,3-dibromo-5,5-dimethylhydantoin
O
O
Cu(II)Cl , MeMgBr,
2
TMSCl, THF; then
peracetic acid, DCM
OAc
OH
OAc
O
O
4
7%
H
H
H
H
O
O
(
24)
(1)
Scheme 7. Completion of the synthesis of 1.
small quantities of the C-17 protonated compound 19, the 17,21-
dihydroxylated product [12] and, surprisingly, the C-17 b -hydrox-
ylated product 20, Scheme 6.
At this point, we turned our attention to the functionalisation of
steroids bearing a 9,11-epoxide. Initially we looked at the reaction
of 21. This 9,11-epoxide analogue underwent clean methyl Grig-
nard addition to C-16 and trapping of the corresponding trimethyl-
silyl enol ether followed by clean Rubottom oxidation to provide
Acknowledgements
The authors are thankful to the leadership of Second Generation
API UK for help and support. This work was financially supported
by GlaxoSmithKline.
Appendix A. Supplementary data
2
2, Scheme 6. This pleasing result was in stark contrast to the re-
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.steroids.2013
sults seen with 7 and 17.
Finally we looked at application to the system that would pre-
pare 1. Formation of epoxide 24 was achieved via the intermediacy
of bromoformate 23, which was isolated as a damp filter cake,
Scheme 7. Copper-catalysed Grignard addition was followed by
trapping of the resultant enolate as the trimethylsilylenol ether,
which underwent clean oxidation with peracetic acid providing 1
without other impurities and allowing isolation by crystallisation
in modest yield, Scheme 7.
.
09.008.
References
[1] Draper RW, Hu B, McPhail AT, Puar MS, Vater EJ, Weber L. Tetrahedron
1999;55:3355–64.
[
[
2] (a) Oliveto EP, Rausser R, Herzog HL, Hershberg EB, Tolksdorf S, Eisler M, et al.
Am Chem Soc 1958;80:6687–8(b) Scherico Novel Steroids and processes for
the manufacture thereof GB901092, 1958.
3] (a) Marker RE, Rohrmann EJ. Am Chem Soc 1940;62:518–20;
This striking difference in reactivity between the 9,11-alkene
substrates 7 and 17 and 9,11-epoxide substrates 21 and 24 was
unexpected at the start of this endeavour. Molecular models do
not provide much insight into the divergent reactivity seen upon
(
(
b) Marker REJ. Am Chem Soc 1940;62:2543–7;
c) Djerassi C, Romo J, Rosenkranz GJ. Org Chem 1951;16:754–60.
[4] Vardanyan RS, Hruby VJ. In: Synthesis of essential drugs. Elsevier; 2006. p.
51–6.
3
9,11
[5] A search of Available Chemicals Directory (ACD) on 11 February 2013
identified 81 different suppliers of prednisolone. ACD version 2011.12,
Accelrys. 2011.
silyl enol ether oxidation for the
D
-alkenes and 9b,11b-epox-
ides. The introduction of the additional ring of the epoxides would
be anticipated to introduce additional strain as a complicating fac-
tor. Despite the ambiguity of the origin of the divergent behavior, it
is readily apparent that addition of substituents to C-16 and oxida-
tion of C-17 is more favourable for the epoxide substrates. The cop-
per catalysed Grignard addition-C-17 oxidation sequence has been
described in the patent literature previously for a range of sub-
strates [13], although in these instances the oxidizing agent was
[6] (a) Allen WS, Bernstein SJ. Am Chem Soc 1955;77:1028–32;
(
b) Bernstein S, Littell R, Williams JHJ. Am Chem Soc 1953;75:4830–2.
7] Sólyom S, Szilágyi K, Toldy L. fuer Praktische Chemie (Leipzig)
988;330:309–12.
[
J
1
[8] Djerassi C, editor. Steroid reactions, an outline for organic chemists, San
Fransisco, Holden-Day Inc., 1963. p. 235.
[
9] Salce L, Hazen GG, Schoenewaldt EFJ. Org Chem 1970;35:1681–2.
10] Mattox VRJ. Am Chem Soc 1952;74:4340–7.
[11] Li M, Chen B, Lin M, Chen T-M, Fu X, Rustum A. Tetrahedron Lett
007;48:3901–5.
12] (a) Horiguchi Y, Nakamura E, Kuwajima I. Tet Lett 1989;30:3323–6;
b) Horiguchi Y, Nakamura E, Kuwajima IJ. Org Chem 1986;51:4323–5.
[13] (a) Preparation of non-hormonal steroid modulators of NF- B for treatment of
disease full text by McCall, John M. et al. From PCT Int. Appl., 2011127048,
011.(b) Preparation of non-hormonal steroids as modulators of NF- B for the
[
2
3
-chloroperbenzoic acid.
[
In conclusion, we have demonstrated a rapid synthesis of 1
(
from prednisolone in six steps and 13% yield without process
optimisation. Highlights include the development of novel condi-
tions for effecting a Mattox rearrangement of prednisolone and
extensive knowledge generation on substituent effects on C-17
oxidation of silyl enol ethers. The Mattox rearrangement condi-
tions constitute the first example of double dehydration of
prednisolone.
j
2
j
treatment of disease Full Text By McCall, John M. et al. From PCT Int Appl,
2009155056, 2009.(c) Method for preparation of dexamethasone and its
derivatives from 1,4,9,16-pregnatetraene-3,20-dione full text by Wang, Fujun
et al. From Faming Zhuanli Shenqing Gongkai Shuomingshu, 101397320,
2
009.(d) Method for preparation of Betamethasone and its derivative full text
by Deng, Lei et al. from Faming Zhuanli Shenqing Gongkai Shuomingshu,
01397319, 01 2009.
1