Note
DOI: 10.1002/bkcs.11627
BULLETIN OF THE
H. Jun et al.
KOREAN CHEMICAL SOCIETY
Synthesis of (S,S)-Reboxetine
*
Hyeyeon Jun, Min Lee Yu, and Soo Y. Ko
Department of Chemistry, Ewha Womans University, Seoul 03760, South Korea.
*E-mail: sooyko@ewha.ac.kr
Received August 5, 2018, Accepted October 9, 2018
Keywords: Reboxetine, Diastereomer, Tandem reaction, One-pot
Advents of asymmetric reactions employing asymmetric
Our previous work on the synthesis of (R,S)-reboxetine
started with Si-protected trans-cinnamyl alcohol. The AD
was the asymmetric reagent and the resulting diol was
simultaneously activated in the form of the cyclic sulfate. A
series of tandem reactions transposed the activation, first to
C-1 (via cyclic sulfate rearrangement),9 then to C-3 (via
epoxide ring-closure),10 allowing the sequential introduc-
tions of nucleophiles there (Nu1=N3 at C-1,
Nu3=2-ethoxyphenoxide at C-3). During this process, an
inversion of configuration at C-2 and a retention (double-
inversion) of configuration at C-3 took place, so that the
AD product, a syn-diol, was transformed to the anti-1-Nu1-
3-Nu3-2-ol product. The whole sequence was conducted in
a step- and pot-economic manner (Scheme 1, top).
In our efforts to develop a synthetic sequence for (S,S)-
reboxetine, which would be comparable to our own syn-
thetic pathway for the anti-diastereomer, we were pre-
scribed to use the same starting material (trans-cinnamyl
alcohol), same key processes (AD, cyclic sulfate activation,
activation transfer, etc.), same disconnection strategy
(Nu1=N3, Nu3=2-ethoxyphenyloxy), and all these in a com-
parably efficient and economic process.11
The two synthetic pathways, one for (R,S)-reboxetine,
and the other for (S,S)-reboxetine, must deviate from each
other in the stereochemical courses, as the latter required
the AD product, a syn-diol, to be transformed to the syn-
1-Nu1–3-Nu3-2-ol product. In practice, it may be realized
with one less (or one more) step of configurational inver-
sion at C-2 or C-3 than in the (R,S)-reboxetine pathway.
We envisaged that a simple change in the order of opera-
tions would accomplish the task. Thus, a Nu3-substitution
at C-3 preceding the activation transfer to C-1 would result
in one less configurational inversion at C-3 and fulfill the
stereochemical requirements needed for the synthesis of
syn-reboxetine (Scheme 1, bottom).
Thus, trans-cinnamyl alcohol (1) was Si-protected (2),
then subjected to the AD protocol. One less configurational
inversion at C-3 meant that the (R,R)-syn-diol (3) was the
right enantiomer leading to (S,S)-reboxetine, which dictated
the use of AD-mix-β as the appropriate choice of the asym-
metric reagent (Scheme 2).12 The two hydroxyl groups of
the AD product (3) were both activated in the form of
cyclic sulfate (4). 2-Ethoxyphenoxide nucleophile was then
reagents or catalysts are regarded as a paradigm shift in the
field of stereoselective synthesis. In this paradigm, choices
of the appropriate asymmetric reagents/catalysts control the
configurations of newly created stereocenters (reagent-con-
trol).1 This is in contrast to the traditional substrate-control,
in which a pre-existing stereocenter in the starting material
induces the configurations of newly created stereocenters in
the product.2 The archetype of the reagent-control strategy
is probably the hexose synthesis by Sharpless–Masamu-
ne.1b,c Here, each of the diastereomeric hexose products
was synthesized from a common starting material, follow-
ing a common synthetic sequence.
In reality, asymmetric reagents/catalysts do not always
grant an equal access to diastereomers. They are often
effective in producing one diastereomer (in either enantio-
meric form), but not so for other diastereomers, for which
different and often lengthier synthetic sequences may have
to be devised.3 The syntheses of reboxetine offer a good
example. Reboxetine is a potent selective norepinephrine
reuptake inhibitor and is effective in treating depression
and attention deficit/hyperactivity disorder (ADHD).4 Initial
interests were directed mainly on the syn-diastereomers
[(S,S) and (R,R)], the racemic mixture of which is currently
on the market. More recently, the anti-diastereomers [(R,S)
and (S,R)] began to gain attentions as some anti-derivatives
exhibited biological activities, which are comparable to,
and at the same time, distinct from those of (S,S)-reboxetine
(Figure 1).5
A survey of the synthetic sequences of reboxetines
reveals that the pathways for reboxetine diastereomers are
not parallel to each other, as in the hexose synthesis, but
require different starting materials or different strategies.6,7
For example, following the asymmetric dihydroxylation
(AD) process, syn-reboxetine was synthesized via O–Aryl
disconnection strategy, while anti-diastereomer required an
O–Benzyl disconnection strategy. The latter is a contribu-
tion from our laboratories.8 Having previously achieved a
synthesis of (R,S)-reboxetine, we embarked on a synthetic
study of (S,S)-reboxetine. Our aim was to develop a syn-
thetic sequence for (S,S)-reboxetine, which would be paral-
lel, as closely as possible, to our own synthetic pathway for
the anti-diastereomer.
Bull. Korean Chem. Soc. 2018
© 2018 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Wiley Online Library
1