1216
J.-J. Wu et al. / Tetrahedron Letters 56 (2015) 1215–1217
26
O
H
F
H
LiBr, BF3•OEt2
CH2Cl2, 25 °C
E
O
O
H
H
Br
H
H
H
H
AcO
AcO
6 (58%)
diosgenin acetate (5)
+
DCM, reflux, 3 h
quantative
BF3•OEt2
O
‡
α-H elimination
2
H
1
1
2
3
6
7
26 OBF2
Br
Br
H
H
2
fast
Br-
O
1
H
H
16
3
AcO
Br-
slow
H
7 (16%)
11
1
2
3
nucleophilic attack at C26
α
nucleophilic attack at C16
-H (C20) elimination
optimum condition: LiBr, BF3•OEt2, DCM/MeCN (4/1), 25 °C, 3 h; workup; reflux, 3 h, 87%
O
Br
H
12
H
H
O
H
H
Br
O
H
Br
H
H
H
H
I
H
H
AcO
AcO
H
H
8 (66%)
9 (72%)
10 (85%)
Scheme 2. A mild halogenation-ring opening reaction of spiroketals.
Pseudodiosgenin-type products,
x
-halo enol ethers such as 6 in
etherate generates transition state 11 whose C16 and C26 are
exposed to nucleophilic attack of bromide ion and C20–H is easily
deprived of. Therefore, dibromide 7 is produced through nucleo-
Scheme 2 which undergo nucleophilic substitution and ring-
closing reaction conveniently, would be more suitable for this task.
Fuchs and co-workers have reported a concurrent ring opening and
philic substitutions at both C16 and C26, and
through a nucleophilic attack at C26 and an
11 or through an intramolecular nucleophilic substitution followed
by an -H elimination from 7. Dibromo ketone 10 is derived from
x
-bromo enol ether
iodination of E/F rings in hecogenin with PPh3ÁI2 to give
x
-iodo
a-H elimination from
derivatives in good yields.5 However, their method requires a
relatively high temperature and the products are frequently
contaminated by 26-chloride side products. Therefore, a milder
and cleaner reaction is still valuable.
a
dibromo-22-ketone like 7 when the C12 ketone is available as an
aldol acceptor (Fuchs also reported a similar result in Ref. 5b).
With such a mild halogenation-ring opening reaction in hand,
we then moved on to the synthesis of solasodine (1), as shown in
Scheme 3. Our preparation of azido derivative 12 was realized by
carrying out azide substitution right after the ring opening and
bromination of spiroketal. Diosgenin acetate 5 was treated with
LiBr and boron trifluoride etherate at ambient temperature, the
crude product was heated in DMF at 60 °C for two hours and
immediately treated with NaN3 for five hours to give 12 in 76%
yield at gram scale (purification was not necessary, the crude 12
could be used directly). Reduction of azide 12 with TMSCl/NaI sys-
tem in MeCN was followed by a spontaneous cyclization of the
We have previously realized a Lewis acid catalyzed mercaptol-
ysis of steroidal sapogenins to obtain 26-thioethers in good yields,
and found that boron trifluoride etherate is a superior activator for
C26–O and that external nucleophiles attack C26 to give 26-substi-
tuted derivatives.6 So we envisaged that the addition of halogen
nucleophiles might provide the desired
x-halo derivatives.
Indeed, by simply switching the thiols with lithium bromide
our original procedure smoothly produced the desired 26-bromo
derivative 6 in 58% yield, along with 16,26-dibromo-22-keone 7
in 16% yield. With an unstable c-bromo ketone unit, compound 7
was quantitatively converted into enol ether 6 upon stirring in
refluxing DCM or in DMF at 60 °C. Increasing the concentration
of bromide ion in reaction mixture, by adding a catalytic amount
of Bu4NBr (0.1 equiv) or by using MeCN as cosolvent, could accel-
erate ring opening reaction. Compound 6 was obtained in 87% yield
according to the optimized procedure: treating diosgenin acetate 2
with LiBr (or NaBr) and boron trifluoride etherate in DCM/MeCN
(4:1) at 25 °C for three hours, an aqueous workup, and heating a
solution of the crude product in DCM or DMF for another three
hours. It should be noted that lithium/sodium iodide were also
resultant
x-NH2 enol ether under the acidic condition to afford
the desired solasodine acetate 13 in 69% yield along with a less
polar compound 14 in 6–8% yield. Hydrolysis of 13 by 5 N NaOH
afforded solasodine (1) in excellent yields. The spectroscopic prop-
24.6
erties of the synthetic 1 ([
a]
D
À102 (c 1.0, CHCl3); reported: [
a]
D
À103 (c 0.70, CHCl3)7)are consistent with those reported, therefore
we achieved the synthesis of solasodine in three steps and 50%
overall yield form diosgenin acetate. When the synthesis was per-
formed without purifying 12 and 13, solasodine was isolated in
38–44% overall yield.
used as nucleophiles to give the
x-iodo derivative 8 in 66% yield.
The reaction produced 9 in 72% yield from tigogenin acetate (with-
out C5–C6 double bond), and gave dibromo ketone 10 in 85% yield
from hecogenin acetate (with a C12 ketone).
A plausible mechanism is outlined in Scheme 2. The selective
activation of the oxygen atom in F-ring with boron trifluoride
Because of the resemblances between 13 and 14 in NMR and MS
spectra, we initially assigned 14 as 22-epimer of 13. Again, hydro-
lysis of 14 by 5 N NaOH gave 15 in excellent yields. Careful recrys-
tallization of 15 from MeOH/acetone allowed us to obtain a crystal
suitable for X-ray analysis.8 We were surprised to find that 15