Process for preparing Ezetimibe intermediate by enantioselective CBS catalyzed ketone reduction with BH3-DEA prepared in situ

Article abstract of DOI:10.1016/j.tetlet.2007.01.118

The (S) alcohol in the benzylic position of compound 2, a key intermediate in the synthesis of the cholesterol lowering agent Ezetimibe, was introduced by the (R)-MeCBS catalyzed asymmetric carbonyl reduction of ketone 1 using borane diethylaniline complex (BDEA) as the reducing agent. The latter was prepared in situ from sodium borohydride (NaBH₄), diethylaniline (DEA) and dimethylsulfate (DMSO₄). BDEA prepared in situ offers considerable advantages from the industrialization standpoint (cost and stability on storage of the reagents) over commercial solutions of BH₃-THF (BTHF) or BH₃-DMS (BMS). The effect of critical reaction parameters such as addition mode of reagent, temperature, solvent, reaction quenching as well as LiCl addition on the selectivity has been examined. This reaction has been successfully applied in the process for the preparation of key intermediate 2 for Ezetimibe.

Full text of DOI:10.1016/j.tetlet.2007.01.118

Tetrahedron Letters 48 (2007) 2123–2125
Process for preparing Ezetimibe intermediate
by enantioselective CBS catalyzed ketone reduction
I
with BH –DEA prepared in situ
3
Blandine Bertrand, Sonia Durassier, St e´ phane Frein and Alain Burgos*
PPG-SIPSY, Z.I. La Croix Cadeau B.P.79, 49242 Avrill e´ Cedex, France
Received 28 November 2006; revised 22 January 2007; accepted 23 January 2007
Available online 27 January 2007
Abstract—The (S) alcohol in the benzylic position of compound 2, a key intermediate in the synthesis of the cholesterol lowering
agent Ezetimibe, was introduced by the (R)-MeCBS catalyzed asymmetric carbonyl reduction of ketone 1 using borane diethyl-
aniline complex (BDEA) as the reducing agent. The latter was prepared in situ from sodium borohydride (NaBH ), diethylaniline
4
(
(
DEA) and dimethylsulfate (DMSO
cost and stability on storage of the reagents) over commercial solutions of BH
4
). BDEA prepared in situ offers considerable advantages from the industrialization standpoint
–THF (BTHF) or BH –DMS (BMS). The effect of
3
3
critical reaction parameters such as addition mode of reagent, temperature, solvent, reaction quenching as well as LiCl addition on
the selectivity has been examined. This reaction has been successfully applied in the process for the preparation of key intermediate 2
for Ezetimibe.
Ó 2007 Elsevier Ltd. All rights reserved.
A process to prepare the (S) alcohol in the benzylic
position of compound 2 by reduction of ketone 1 with
BTHF enhanced by para-toluenesulfonic acid has been
ity as compared with BDMS or BDEA where only one
hydride can be used Scheme 1.
1
described in the literature. BTHF is known to be unsta-
BDMS presents an advantage in this process from the
industrialization standpoint over BTHF. Indeed, BDMS
is more stable than BTHF, thus the amide bond of com-
pound 2 is not easily reduced. A classic reagent addition
can be used for this reaction with this reagent. However,
ble on storage (THF ring opening) and explosions on
2
storage have been reported. Therefore, the use of
BTHF at industrial scale can be cumbersome from the
logistic and supply chain standpoint.
1
as pointed out by the authors, the use of BMS leads to
The reactivity of BTHF presents in this case a major
drawback from the industrialization standpoint. Indeed
to overcome the reduction of the amide bond of com-
pound 2 (compound 4), BTHF must be added to a solu-
tion of compound 1 in the presence of MeCBS and THF
environmental concerns due to release of methylsulfide.
BDEA is even more stable than BDMS. The dissocia-
tion of BDEA to give B H and DEA only occurs at
2
6
3
100 °C. One can imagine that using BDEA, a classic
addition process can be used for this process. Moreover,
DEA is easily recovered from the reaction and recycla-
ble. Unfortunately, commercial solutions of BDEA are
far more expensive than BTHF of BMS solutions.
(
reverse addition process). This requires a particular set-
up of the industrial equipment which does not fit easily
in multi products workshop. Moreover, BTHF is only
available at 1 M concentration which makes the process
productivity low from a production standpoint. The
only positive outcome of using BTHF is that two
hydrides over three theoretically available can be used
in the reduction process to obtain a good stereoselectiv-
In situ generation of BH has been investigated in depth
3
over the years to overcome the issues of cost and stability
of BTHF. Unfortunately, combinations like NaBH /
4
4b
4
a
4a
H SO , NaBH /BX (X@Cl, F), NaBH /TMSCl,
2
4
4
3
4
4
a
NaBH /MeI, do not give pure enough BTHF to be
4
q
used in the MeCBS catalyzed asymmetric carbonyl
reduction of ketone or worse the process is too hazardous
(NaBH /H SO ) from the industrialization standpoint.
This chemistry was disclosed in our patent WO 2005/035540.
*
Corresponding author. Tel.: +33 0 241 966 169; fax: +33 0 241 968
40; e-mail: aburgos@ppg.com
2
4
2
4
0040-4039/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2007.01.118

2124
B. Bertrand et al. / Tetrahedron Letters 48 (2007) 2123–2125
O
O
O
OH
O
O
OH
O
O
OH
OH
O
N
O
B
, THF
N
O
N
O
F
N
O
F
N
O
CH3
BH –DEA
F
3
F
DME
1
2
3
4
NaBH4 + Me SO + DEA
2
4
Scheme 1. (R)-MeCBS catalyzed reduction of 1.
A process to generate BTHF in situ from NaBH and
expected we found trace amounts (<1%) of compound
4 using the normal reagent addition. However, the de
(vs compound 3) was slightly less than the de’s reported
with BTHF or BDMS (Table 1, entries 1 and 2). We
found that when using THF as co-solvent during the
asymmetric carbonyl reduction of compound 1, the de
was comparable to the de reported with BTHF or
BDMS (Table 1, entry 4). We also showed that the opti-
mum temperature for the reduction from the de stand-
point is 20 °C (Table 1 entries 3–5).
4
5
DMSO is described in the literature. DMSO is an
4
4
industrial reagent at PPG-SIPSY already used in several
large scale processes. So, we decided to investigate the
generation in situ of BDEA from NaBH , DEA and
4
DMSO , Eq. 1.
4
Me
2
SO
4
+ NaBH
4
+ DEA ! BH
3
–DEA + MeOSO
3
Na
ð1Þ
Because BH forms a stable complex with DEA, the
3
choice of the solvent for the generation reaction was
not limited to THF. We found that monoglyme
Encouraged by the above results, we decided to tele-
scope the two processes to assess the quality of our
BDEA generated in situ.
(
DME) gave the best results from the yield standpoint
and from the stirring point of view, presumably because
of the higher solubility of NaBH in DME than in THF.
Using 5% of MeCBS, the de obtained was slightly less
than the de obtained with commercial BDEA (Table 1,
entries 4 and 6). A carefully examination of the reaction
mixture showed that N-methyl-diphenylprolinol was
present in the crude product. N-Methyl diphenylprolinol
was attributed to the reaction of MeCBS with the resid-
4
The reaction proceeds smoothly at 40 °C and a small
excess of DMSO is enough to drive the reaction to
4
completion. The volume of methane released during
the course of the reaction indicates that the yield is
nearly quantitative. Because of the stability of BH3–
DEA complex, B H does not escape from the reaction
ual DMSO (Scheme 2).
2
6
4
mixture even under the methane stream.
Our hypothesis was that this side reaction of MeCBS
with DMSO reduces the amount of catalyst available
In parallel, we studied the reduction of compound 1
4
with commercially BDEA with MeCBS in DME. As
for the asymmetric reaction and thus reduces the de of
Table 1.
b
Entry
Borane source
Solvent
Reaction conditions
Normal addition
de (%)
Compound 4 (%)
Chemical purity (area %)
Toluene area not integrated
1
1
2
BDMS
>98
93.5
<1
0.9
BDEA Commercial solution
DME
BDEA Commercial solution
THF/DME
BDEA Commercial solution
THF/DME
BDEA Commercial solution
THF/DME
BDEA in situ
THF/DME
BDEA in situ
THF/DME
Normal addition 20 °C
Normal addition 0 °C
Normal addition 20 °C
Normal addition 40 °C
Normal addition 20 °C
>95
>95
>95
89
3
4
5
6
7
94.1
98.9
94.9
97.6
98.2
98.3
94
0.2
0.7
6.6
2.8
4.9
2.3
2.9
>95
>90
95
Normal addition 20 °C
LiCl
Normal addition 20 °C
c
8
BDEA in situ 1 equiv
THF/DME
BDEA in situ
a
9
Normal addition 20 °C
>94
THF/DME
All the reactions were carried out with 5% MeCBS catalysis.
a
Reaction was carried out with 5% HCBS catalysis.
de were determined by HPLC on a Chiralcel ODH column.
See Ref. 7 for the experimental procedure.
b
c

B. Bertrand et al. / Tetrahedron Letters 48 (2007) 2123–2125
2125
any residual active hydrides (Table 1, entry 8) before
the hydrolysis.
1
/ Me SO
2 4
In summary, a process to produce in situ BDEA com-
patible with MeCBS asymmetric carbonyl reduction of
ketone has developed and successfully applied to a key
intermediate in the synthesis of the cholesterol lowering
agent Ezetimibe. This process uses cheap, raw, storable
and recyclable (DEA) reagents. The process has been
extended to HCBS (Table 1, entry 9).
N
O
2/ H O
B
2
N
OH
CH3
CH3
Scheme 2. N-Methylation of MeCBS.
compound 2. Indeed, the relationship between ee and
catalyst ratio is well established, due to the competition
with non-stereoselective reduction with the borane
source.
References and notes
In order to suppress this side reaction, we needed to find
1
. Fu, Xiaoyong; McAllister, Timothy L.; Thiruvengadam,
T. K.; Tann, Chou-Hong; Su, Dan Tetrahedron Lett. 2003,
44, 801–804.
a way to remove the excess of DMSO . The excess of
4
DMSO is mandatory to avoid having residual NaBH
4
4
which has a negative impact on the de as mentioned in
the literature.
2. Web site: www.callery.com
1
3
4
. Internal study.
. (a) Zweifel, G.; Brown, H. C. Organic Reactions; John
Wiley & Sons: New York, Vol. 13; p 27; Abiko, A.;
Masamune, S. Tetrahedron Lett. 1992, 33, 5517; (b)
Sengupta, Sreela; Sahu Devi, P.; Chatterjee Sunil, K.
Indian J. Chem. 1994, 33B, 285–287; Bolm, C.; Seger, A.;
Felder, M. Tetrahedron Lett. 1993, 34, 8079–8080.
. Lane, C. Chem. Rev. 1976, 76, 773–799.
DMSO reacts with a variety of nucleophiles. The choice
4
was guided by the compatibility of the nucleophile with
BH , that is, which does not give a stable complex with
3
BH . A chloride, Eq. 2, rapidly appeared to be the best
3
choice even if the literature mentions a negative impact
5
6
on the ee. We anticipated that methylchloride is not a
6. Xu, J.; Wie, T.; Zhang, Q. J. Org. Chem. 2004, 69, 6860–
6866.
7. Procedure (Table 1, entry 8): Under nitrogen, 1.28 g
strong enough electrophile to react with MeCBS and
would anyway escape from the reaction mixture at
(
33.8 mmol) of NaBH was suspended in 24 ml of DME;
4
4
0 °C.
5
.4 g (35.5 mmol) of DEA was then added. The mixture was
was
ꢀ
ꢀ
Me
2
SO
4
þ Cl ! MeCl þ MeOSO3
ð2Þ
heated up to 40 °C then 4.47 g (35.5 mmol) of DMSO
added slowly to the previous suspension. The reaction was
4
Adding LiCl to the reaction mixture before the addition
of MeCBS did produce methylchloride gas and gave
compound 2 with a de in line with the de obtained with
commercial BDEA (Table 1, entries 4 and 7).
maintained at 40 °C until a clear solution was obtained.
2
86 mg (6.8 mmol) of LiCl was then added. The reaction
mixture was then cooled to room temperature (20 °C) then
1
.7 ml of a 1 M (R)-MeCBS in toluene (1.7 mmol) was
added. Ketone 1 [12 g (33.8 mmol)] was dissolved in 24 ml
of THF and added over a 1 h period to the previous
reaction mixture. The reaction was then maintained at
In order to ensure a good scaleup of the process, we
looked at the reaction mixture stability. We found that
on aging, the amount of compound 4 increases signifi-
cantly. This increase was a concern because the reaction
mixture is quenched by adding the reaction to an aque-
ous solution in order to control the release of hydrogen.
We expected the quenching time to increase with the
scale. We got around this problem by reducing the theo-
retical amount of BDEA to one strict equivalent against
compound 1 (i.e., using one strict equivalent of NaBH4)
and by quenching the reaction with acetone to destroy
2
0 °C for 10 min then 6 g (100 mmol) of acetone was
rapidly added. 6.07 g (43.9 mmol) of potassium carbonate
was dissolved in 72 ml of water. The reaction mixture was
then added to the carbonate solution. The aqueous phase
was discarded. The organic phase was partially concen-
trated and taken up in 72 ml of toluene. The organic phase
was washed with 36 ml of water, twice with 30 ml of 10%
sulfuric acid (50 mmol) then with 36 ml of water. The
organic phase was concentrated under vacuum to dryness
to give 13 g of a viscous oil, compound 2.