J. R. Hitchin et al. / Tetrahedron Letters 53 (2012) 2868–2872
2869
O
O
H
O
O
H
H
H
H
H
H
H
H
H
H
H
H
H
H2N
H2N
H2N
H2N
H
H
H
H
H
H
H
H
1
2
3
4
Figure 1. Structures of 3-aminoandrostanones and 3-aminopregnanones.
two examples of phthalimide based Mitsunobu chemistry being
successfully applied to the synthesis of diastereomerically pure
3-aminocholestanes,9,10 to the best of our knowledge, this method-
ology has not been reported for the stereospecific synthesis of
Sequential cleavage of the phthalimide and acetal protecting
groups was completed by refluxing with hydrazine in ethanol fol-
lowed by treatment with aqueous hydrochloric acid in THF, respec-
tively, affording the desired 3b-amino-5a-androstan-17-one (1a)
either 3
amino-5
Herein, we describe the stereoselective synthesis of 3
amino-5 -androstan-17-ones and 3 - and 3b-amino-5
a
a
- and 3b-amino-5
-pregnan-20-ones.
a
-androstan-17-ones or 3
a
- and 3b-
as the hydrochloride salt. These two deprotection stages were tele-
scoped effectively into a single step. Typically, during the hydra-
zine-mediated phthalimide cleavage, the phthalizine side-
product precipitated from the reaction mixture and the amino
product remained in solution. The product solution was collected
and distilled to dryness to provide the crude acetal protected 3b-
aminoandrostane, which was subsequently dissolved in a 2 N
hydrogen chloride/THF/acetone mixture for acetal deprotection.
The reaction mixture was then concentrated to a low volume and
subsequently diluted with ether leading to precipitation of the
a
a
- and 3b-
-pregnan-
a
a
20-ones using a phthalimide based Mitsunobu reaction sequence.
Linear reaction schemes of six steps delivered both 3b-amino-5
androstan-17-one (1) and 3b-amino-5 -pregnan-20-one (3) as
hydrochloride salts from the corresponding 3b-hydroxy species 5
and 11, respectively. The related 3 -amino-compounds, 3 -ami-
no-5 -androstan-17-one (2) and 3 -amino-5 -pregnan-20-one
a
-
a
a
a
a
a
a
(4), were prepared in two steps from the corresponding acetal pro-
tected 3b-hydroxy species, 6 and 12, respectively. In all four exam-
ples, no chromatographic purification was required and the
intermediates and products were invariably isolated from the reac-
tion mixtures in high chemical yields as single diastereoisomers. In
each case the stereochemistry was confirmed by comparison with
the 1H NMR data reported in the literature.6–8 Furthermore, values
for the line-width at half height (w/2) of the key H-3 resonances
were determined from the 1H NMR data to consolidate the stereo-
specificity of the approach.11
3b-amino-5a-androstan-17-one (1a) product as the hydrochloride
salt. Pleasingly, the diastereoselectivity of the product was con-
firmed by preparation of the corresponding acetic acid salt (see
Supplementary data) and comparison with the 1H NMR data re-
ported in the literature.7 Furthermore, the orientation of the H-3
proton was confirmed to be axial (3a-H) from the magnitude of
the line-width at half height (w/2 = 23 Hz), which is consistent
with that expected for an axial proton in these systems.
The overall yields for the two six-step sequences from 3b-hy-
droxy-5a-androstan-17-one (5) were 37% and 28% for paths A
The key focus of our synthetic strategy was accessing the acetal
protected 3a-hydroxyandrostane (8) stereoselectively to enable
the subsequent Mitsunobu reaction with phthalimide to introduce
the required 3b-amine. Our initial synthetic strategy involved the
oxidation/reduction cycle depicted in Scheme 1, path A. Thus, pro-
and B, respectively, with the synthetic methodology performing
well on scale-up, successfully affording multi-gram quantities of
the desired product. It is also important to recognize that these
yields do not represent optimized processes and reflect only the ro-
bust and efficient nature of the overall strategy, in that generic
methodologies responded well when applied to the current syn-
thetic scheme.
tection of the 17-keto group of 3b-hydroxy-5a-androstan-17-one
(5) afforded the acetal 6 in excellent yield.12 The stereochemistry
of the 3b-hydroxy group was inverted by a sequential oxidation/
reduction13 process using Dess Martin periodinane (DMP) and K-
Synthesis of the corresponding 3
(2) was conducted as depicted in Scheme 2. 3b-Hydroxy-5
stane (6) was subjected to a Mitsunobu reaction with phthalimide
affording the 3 -phthalimide 10 in high yield as a single diastereo-
isomer. As before with the 3b-phthalimide 9, the 3 -isomer 10
a
-amino-5
a
-androstan-17-one
a-andro-
SelectrideÒ, respectively, to give the 3
a-hydroxyandrostane (8),
stereoselectively. However, this path left some 3b-hydroxy starting
material (<5%) representing a significant issue to the proposed
telescoping strategy, so the alternative route shown in Scheme 1,
path B was developed. The acetal protected 3b-hydroxyandrostane
(6) was subjected to a Mitsunobu reaction with benzoic acid,
inverting the stereochemistry of the 3-hydroxy group cleanly to af-
ford the corresponding benzoate in high yield. It is worthy of note
that the benzoate precipitated from the reaction mixture as it
formed and was subsequently collected by filtration as a single dia-
stereoisomer with high chemical purity (>98%) by 1H NMR analy-
sis. Subsequent sodium methoxide mediated cleavage of the
a
a
precipitated cleanly from the methanol during the course of the
reaction work-up and was collected by filtration. Sequential cleav-
age of the phthalimide and acetal groups was achieved by initially
refluxing the 3
drate in ethanol followed by treatment with an aqueous 2 N
hydrogen chloride solution in THF, to give the desired 3 -amino-
-androstan-17-one (2a) as the hydrochloride salt in an overall
a-phalimidoandrostane 10 with hydrazine monohy-
a
5a
yield of 53%. Once again, the diastereoselectivity of the product
was confirmed by direct comparison with the 1H NMR data re-
ported for this compound in the literature.8 Furthermore, the ori-
entation of the H-3 proton was confirmed to be equatorial (3b-H)
from the value of the line-width at half height (w/2 = 9 Hz), which
is typical for an equatorial proton and considerably smaller than
that expected for the corresponding axial proton.
benzoate group provided the 3a-hydroxyandrostane (8) in high
yield as a single diastereoisomer.
The amine was then introduced stereoselectively at the 3-posi-
tion by a Mitsunobu reaction with phthalimide, providing the
3b-phthalimide 9 in high yield. It is important to note that this
compound precipitated from methanol during the reaction work-
up as a single diastereoisomer and was collected by filtration. The
chemical purity was >98% by 1H NMR analysis, with no evidence
The synthesis of 3b-amino-5
tigated using the same strategy that had been successfully
developed for the synthesis of 3b-amino-5 -androstan-17-one
(1a) as depicted in Scheme 3. Protection of the 20-keto group in
a-pregnan-20-one (3a) was inves-
a
of any 3a-diastereomer.