A novel chiral resolving reagent, bis((s)-mandelic acid)-3-nitrophthalate, for amlodipine racemate resolution: Scalable synthesis and resolution process

Article abstract of DOI:10.1021/op900070c

A novel bis((S)-mandelic acid)-3-nitrophthalate (1), a chiral resolving reagent for the separation of (S)-(-)-isomers of amlodipine from the racemate thereof, is designed and synthesized. A simple three-step pilot-scale preparation of 1, along with the optimization of a resolution process on the racemate amlodipine, is reported.

Full text of DOI:10.1021/op900070c

Organic Process Research & Development 2009, 13, 1382–1386
A Novel Chiral Resolving Reagent, Bis((S)-Mandelic acid)-3-nitrophthalate, for
Amlodipine Racemate Resolution: Scalable Synthesis and Resolution Process
Hong Woo Lee,* Sung Jae Shin, Hosung Yu, Sung Kwon Kang, and Choong Leol Yoo*
Chemical Research Lab, Chong Kun Dang Research Institute, Cheonan P.O. Box 74, Cheonan 330-831, South Korea
Abstract:
A novel bis((S)-mandelic acid)-3-nitrophthalate (1), a chiral resolv-
ing reagent for the separation of (S)-(-)-isomers of amlodipine
from the racemate thereof, is designed and synthesized. A simple
three-step pilot-scale preparation of 1, along with the optimization
of a resolution process on the racemate amlodipine, is reported.
Figure 1. (R,S)-Amlodipine.
process. Furthermore, solvent recovery is inconvenient due to
its high boiling point. Method (iii) uses isopropanol as the main
crystallization solvent to overcome the problems of DMSO with
Introduction
Amlodipine (Figure 1) and its salts are long-acting calcium
the relatively expensive dibenzoyl derivative of D-tartaric acid,
channel blockers used in the treatment of angina, congestive
but we found in our test experiments that this diastereomeric
heart failure, and hypertension. It is being used in its enantio-
salt has wet, cakelike crystal properties not suitable for
meric mixture of salt forms, but the two enantiomers of
industrial-scale filtration. Method (iv) merely demonstrates less
amlodipine and their salts have different pharmacological
than a 10-g scale of (S)-amlodipine separation from its racemate
profiles. The (S)-(-)-isomer is the more potent calcium channel
which is not a sufficient demonstration for a large-scale
blocker showing about 2000-fold potency in in vitro evaluation
application; in addition the corresponding (S)-amlodipine-hemi-
in the rat aorta than the (R)-(+)-isomer.1 While a number of
(1R, 3S)-camphorate salt has hygroscopic properties and quickly
sources in the literature and patents have reported the separation
of (S)-(-)-amlodipine from its racemate,2 four selective dia-
stereomeric salt crystallization methods have showed promising
results for industrial application: (i) of (S)-amlodipine-hemi-D-
tartrate salt in the presence of DMSO,2d (ii) of (S)-amlodipine-
becomes quite wet after filtration. Therefore, it seemed worth-
while to develop and to optimize an alternative method. Our
initial approach was to test other resolving reagents commonly
used in industry with various solvent systems, but none of those
trials was successful. Thus, we decided to design and develop
a new resolving reagent. Two criteria were set for the desig of
new resolving reagents. The new reagents (1) must be readily
prepared from cheap commercial materials and (2) must be
hemiLtartrate salt via processing the filtered solution of (R)-
amlodipine-hemi-L-tartrate salt in DMSO,2e (iii) of (S)-amlodipine-
hemidibenzoyl-D-tartrate salt,2f (iv) of (S)-amlodipine-hemi-(1R,
3S)-camphorate salt.2g
useable with solvents more suitable for an industrial crystal-
However, these methods have some limitations and undesir-
lization process. Here we now report the scalable and com-
able aspects to be used in industrial-scale production. For
mercially viable preparation and resolution process using our
example, DMSO used in methods (i) and (ii) can cause rotten
new chiral resolving reagent.
cabbage odor problems by dimethyl sulfide in conventional
municipal wastewater treatment.3 On the manufacturing side,
a relatively high freezing point makes the manufacturer’s
handling unfavorable because at, or just below, room temper-
Results and Discussion
There are very few approaches to the rational design of
resolving reagents because it is difficult to forecast small and
ature it is a solid, which limits its efficacy in the crystallization
subtle differences in interactions within the diastereoisomeric
* To whom correspondence should be addressed. Telephone: +82 415293384.
Fax: +82 415583004. E-mail: hwlee@ckdpharm.com; clyoo@ckdpharm.com.
(1) Arrowsmith, J. E.; Campbell, S. F.; Cross, P. E.; Stubbs, J. K.; Burges,
R. A.; Gardiner, D. G.; Blackburn, K. G. J. Med. Chem. 1986, 29,
1696–1702.
salts and their effect on crystal structure. Furthermore, the
solvent plays a decisive role.4 Thus, we have begun with
the simple approach of incorporating additional functionality
around the stereocenter of lactic acid or mandelic acid. Although
there was no clear explanation, a literature reference suggested
that added functionalitysto increase interactionssshould en-
hance rigidity of the compound.4 We reasoned this is because
a flexible resolving reagent can interact indiscriminately with
both enantiomers by quickly adopting a structural change,
whereas a rigid one would prefer either one of the diastreomeric
(2) (a) Arrowsmith, J. E. Preparation of R-and S-Amlodipine. Eur. Pat.
Appl. S9 522 301, 1989. (b) Goldmann, S.; Stoltefuss, J.; Born, L.
J. Med. Chem. 1992, 35, 3341–3344. (c) Goldmann, S.; Stoltefuss, J.
Angew. Chem., Int. Ed. Engl. 1991, 30, 1559–1578. (d) Spargo, P. L.
Separation of the enantiomers of Amlodipine via their diastreomeric
tartrates. U.S. Patent 5,750,707, 1998. (e) Choung, Y. S.; Ha, M. J.
Processes for the Preparation of S-(-)-Amlodipine. U.S. Patent
7,202,365, 2007. (f) Kim, J. S.; Choi, J. Y.; Kim, N. H.; Lee, N. K.
Optical Resolution Method of Amlodipine. U.S. Pat. Appl. Publ. 2008/
0249314 A1, 2008. (g) Byun, I. S.; Kim, Y. Y.; Kim, W. J. Method
For Preparing an Optically Active Amlodipine. WO 2008/026838, 2008.
(3) Glindemann, D.; Novak, J.; Witherspoon, J. EnViron. Sci. Technol. 2006,
40 (1), 202–207.
(4) Bruggink, A. Rational Design in Resolutions. In Chirality in Industry
II; Collins, A. N., Sheldrake, G. N., Crosby, J., Eds.; John Wiley &
Sons Ltd: Chichester, 1997; pp 81-98.
1382
•
Vol. 13, No. 6, 2009 / Organic Process Research & Development
10.1021/op900070c CCC: $40.75  2009 American Chemical Society
Published on Web 10/02/2009

Table 2. Solvent effect on amlodipine resolutiona
yield of
ee of
(S)-amlodipine · (S)-amlodipine
d
entry
solvent
1/2 1 (%)b
(%)c
Sexp
1
2
3
4
5
6
7
8
9
DMSO
-
-
-
-
-
-
-
-
-
-
0.55
0.50
0.37
0.51
0.52
0.62
DMF
MeOH
-
-
acetone
-
-
IPA
-
-
IPA/ACN (9/1)
MIBK
-
-
86
62
46
65
59
65
64
81
80
79
88
96
MEK/MTBE (5/1)
MEK/MTBE (10/1)
10 MEK/IPE (5/1)
11 MEK/IPE (10/1)
12 MEK
Figure 2. New chiral resolving reagents bearing phenyl
functionality.
a Resolution experiments were conducted at temperatures between 15-20
°C with 0.25 equiv of 1. b Relative to the theoretical amount, i.e. half of the
starting racemic amlodipine. c Enantiomeric excess (determined by HPLC) of
the amlodipine liberated from the diastereomeric salt. d Experimental resolution
efficiency or experimental resolvability, calculated from the chemical yield of
the precipitated salt and the enantiomeric excess of the amlodipine liberated
from the salt.
Table 1. Representative prescreening result of resolving
reagents on amlodipine resolutiona
ee of
yield of
(S)-amlodipine · saltb
(S)-amlodipine
d
compound
solvent
(%)c
Sexp
1
IPA/MTBE/DCM
41
70
0.29
(1/25/5)
2e
3
-
ACN/IPA (1/9)
-
MEK/MTBE (4/1)
MEK/MTBE (5/1)
-
-
5
-
10
52
-
-
10
-
87
20
-
-
0.005
-
0.09
0.10
-
diastereomeric salt in good yields but with only moderate
enantiomeric excesses. Finally, methyl ethyl ketone (MEK) gave
diastereomeric salt in good yield with satisfactory resolution
result (65%, 96% ee, entry 12). Furthermore, this diastereomeric
salt showed relevant characteristics suitable for an industrial
application: (1) nonhygroscopic and stable under relatively high
humidity conditions (i.e., at relative humidities above about 60%
and up to about 85%), (2) good filtration property for fast and
easy collection of resulting salt. Upon reslurry of this diaster-
eomeric salt in the mixed solvent system MEK/hexane (2/1)
enrichment of enantiomeric exess to 97.5% was observed with
a 92% recovery of yield. While the enantiomeric exess was
improved up to >99% by the MEK/EtOH (5/1) slurry method,
the recovery yield was only 80%; thus, this method was
discarded. Reslurry of the diastereomeric salt was necessary to
consistently obtain the final (S)-amlodipine besylate, which is
the final ingredient for drug tablet, with >99% ee (Scheme 2).
Although theoretical amount required for the resolution is 0.25
equiv, optimal amount of compound 1 was verified. Our
experiment confirmed this theoretical amount as it is shown in
Table 3. Comparable yield of diastereomeric salt was obtained
when 0.5 equiv of compound 1 was used, but still the best
resolution result was achieved again with 0.25 equiv of 1 (entry
2 in Table 3). The temperature factor was also investigated with
the optimized amount of 1 in MEK single solvent (Table 4).
Resolutions were conducted several times at the given temper-
atures to ensure the results were consistent. Temperatures below
10 °C (entry 1 and 2) or above 24 °C (entry 4), have not
exceeded the result previously obtained at the temperature range
between 15-20 °C (entry 3).
4e
5
6
7e
a Resolution experiments were conducted at temperatures between 15-20
°C. b Relative to the theoretical amount, i.e. half of the starting racemic
amlodipine. c Enantiomeric excess (determined by HPLC) of the amlodipine
liberated from the diastereomeric salt. d Experimental resolution efficiency or
experimental resolvability, calculated from the chemical yield of the precip-
itated salt and the enantiomeric excess of the amlodipine liberated from the
salt. e Precipitation was not observed in various solvents and solvent mixture
systems.
salts for converse reasons. Phenyl functionality seemed fairly
reasonable, considering the availability of various derivatives
at low prices. Furthermore, it can add rigidity to a resulting
diastereomeric salt. Seven candidates have been prepared by
generally known procedures or following the literature prepara-
tion protocol (Figure 2),5 and careful prescreening of these
compounds revealed compound 1 as the best candidate (Table
1).6
Proton NMR analysis of this isolated salt revealed the ratio
of 2:1, salt of (S)-amlodipine to compound 1. Thus, optimization
of resolution experiments with compound 1 was conducted in
various solvent systems at the temperature range of 15-20 °C.
As it is shown in entries 1-5 in Table 2, single polar solvent
systems which usually provide good solubility for various
organic molecules failed to form precipitates. Mixed solvent
systems consisting of solvent and antisolvent generally gave
(5) Gorobets, E.; McDonald, R.; Keay, B. A. Org. Lett. 2006, 8, 1483–
1485.
(6) For each of new chiral resolving reagent in Figure 2, about 20 sets of
prescreening were run with the solvents chosen from ICH class 2 or
class 3 solvents.
Since the preparation of compound 1 was done by a lab
preparation procedure, we had to develop a scalable preparation
Vol. 13, No. 6, 2009 / Organic Process Research & Development
•
1383

Table 3. Optimization of compound 1 equivalenta
Table 5. Coupling reagent test for the preparation of
compound 9
yield of
(S)-amlodipine ·
1/2 1 (%)b
ee of
(S)-amlodipine
(%)c
d
entry equivalent
Sexp
1
2
3
4
0.2
55
65
59
65
93
96
93
85
0.51
0.62
0.55
0.55
0.25
0.35
0.5
a Resolution experiments were conducted in MEK at temperatures between
15-20 °C. b Relative to the theoretical amount, i.e. half of the starting racemic
amlodipine. c Enantiomeric excess (determined by HPLC) of the amlodipine
liberated from the salt. d Experimental resolution efficiency or experimental
resolvability, calculated from the chemical yield of the precipitated salt and the
enantiomeric excess of the amlodipine liberated from the salt.
entry
coupling reagenta
% yieldb
1
2
3
4
5
6
(EtO)2P(O)Cl
(EtO)2P(S)Cl
(MeO)2P(S)Cl
(PhO)2P(O)Cl
ethyl chloroformate
DCC
45
42
Table 4. Optimization of temperaturea
10
45
no rxn
58
yield of
ee of
(S)-amlodipine· (S)-amlodipine
d
entry temperature
1/2 1 (%)b
(%)c
Sexp
1
2
3
4
-2–5 °C
5–10 °C
15–20 °C
24–26 °C
85
76
61
54
0
71
96
90
0
a DMAP (0.3 equiv), DCM (g/10 mL), 12 h, room temperature. b Isolation
yield.
0.32
0.59
0.49
such as alkoxy chlorophosphate or thiophosphate, have been
employed and gave a range of low to moderate yields (entry
1-4), leading us to employ the DCC method (entry 6). DCC
activation consistently gave 10% higher yield than other
methods, and removal of dicyclohexyl urea (DCU) was
unexpectedly easy. Unlike the conventional DCC activation at
0 °C, activation was conducted at room temperature and stirred
for 2 h prior to the addition of alcohol compound 8 to complete
the activation of sterically congested diacids. After the reaction,
DCU was filtered off, and residual DCU in the crude was further
removed by hexane slurry followed by filtration. Crystallization
of the crude mixture in MeOH gave the desired product in 58%
(>99% purity) yield. Deprotection of PMB on compound 9 with
TFA/DCM went smoothly, giving compound 1 in 92% as pale,
yellowish solid after simple workup procedure (Scheme 1).
These reaction steps repeatedly gave desired quality of product
up to a kilogram scale in our pilot reactor without further
complications. Pilot-scale resolution of racemic amlodipine with
compound 1 also proceeded smoothly, giving consistent quality
and quantity of (S)-amlodipine diastereomeric salt (Scheme 2).
Further treatment of this diastereomeric salt with NaOH
followed by salt formation with benzenesulfonic acid gave (S)-
amlodipine besylate dihydrate, which is the active pharmaceuti-
cal ingredient (API) used in final dosage form, in 85% (>99%
ee). About 90% of compound 1 was recovered by acidification
of corresponding disodium salt followed by back extraction of
aqueous layer with DCM. The purity of compound 1 after DCM
extraction was the same as that of the initially used compound
1, showing over 99% purity with the retention of chiral purity
(>97% ee). Recovered compound 1 was subjected to reusability
test. As it is expected by recovered compound’s purity, the
results showed no differences in quality and quantity of (S)-
amlodipine as well as recovered compound 1. The cycles of
recovery and reuse of compound 1 were done three times
without introducing new impurities while maintaining those
impurities below 0.1%. More than three times of cycles
increased existing impurities to just over 0.1%. Since our goal
a Resolution experiments were conducted 3 times at the given temperature
in MEK with 0.25 equiv of 1, and calculated mean values were recorded.
b Relative to the theoretical amount, i.e. half of the starting racemic
amlodipine. c Enantiomeric excess (determined by HPLC) of the amlodipine
liberated from the salt. d Experimental resolution efficiency or experimental
resolvability, calculated from the chemical yield of the precipitated salt and the
enantiomeric excess of the amlodipine liberated from the salt.
procedure for compound 1. Synthesis of our chiral resolving
reagent began with (S)-mandelic acid protection with the
p-methoxybenzyl (PMB) group. Initial preparation of compound
8 was done by using MeOH/water (5/1) cosolvent system
following the literature protocol. While this preparation method
was acceptable for lab-scale preparation, a significant amount
of byproduct due to a PMB alkylation on the secondary alcohol
of both starting compound and compound 8 were problematic
for scale up. The cause of this byproduct formation was
attributed to the unwanted cesium salt formation on the
secondary alcohol; thus, we decided to remove water in order
to take advantage of cesium carbonate’s poor solubility in
MeOH solvent. This slight change has made a significant
difference in the reaction, and we were successfully able to
obtain compound 8 without the byproduct problem in 89%
yield. Next, the coupling reaction between compound 8 and
3-nitrophthalic acid was examined. Compound 9 was initially
prepared by converting 3-nitrophthalic acid to the corresponding
reactive acid chloride followed by the coupling reaction of this
with compound 8 in the presence of weak bases, such as Et3N
or pyridine, and column purification. This preparation method
was also problematic because the phthalic anhydride formation
took place prior to the coupling reaction. More than half of
3-nitrophthalic acid was converted to the corresponding phthalic
anhydride. The nitro group on the 3-position would probably
be responsible for this significant byproduct formation by
increasing the reactivity of carboxylic acid chloride on the
2-position. Thus, we have decided to try less reactive carboxylic
acid activating reagents (Table 5).
As can be seen in entry 5, mixed anhydride activation using
ethyl chloroformate clearly gave disappointing results, offering
no product after the reaction. Other mixed anhydride activations,
1384
•
Vol. 13, No. 6, 2009 / Organic Process Research & Development

Scheme 1. Optimal process for compound 1 synthesis
Scheme 2. Chiral resolution process on (R,S)-amlodipine with compound 1
was to control all unknown impurities below 0.1%, more than
three times of recovery and reuse was not desirable.
mol) in MeOH (12 L) and stirred for 30 min. The solution was
concentrated under reduced pressure, charged with DMF (8 L),
and p-methoxybenzylchloride (2.68 kg, 17.12 mol) was added
while keeping the temperature of the reaction no more than 30
°C. The reaction mixture was stirred for 21 h at between 20-25
°C. Insoluble inorganic salts were filtered off, and the filtrate
was transferred in ethyl acetate/hexane (16 L/4 L). Organic layer
was washed with aq 10% NH4Cl (20 L × 3) followed by 10%
NaHCO3 (8 L) and concentrated under reduced pressure.
Hexane (40 L) slurry of this concentrate at between 0-5 °C
and filtration gave compound 8 in 89% (3.2 kg, 11.72 mol) as
white solid. 1H NMR (CDCl3, 400 MHz) δ 7.36-7.38 (m, 5H),
7.27. (d, J ) 1.5 Hz, 2H), 6.87 (d, J ) 1.7 Hz, 2H), 5.11 (s,
2H), 4.26 (m, 1H), 3.76 (s, 3H). Anal. Calcd for C16H16O4: C,
70.57; H, 5.92. Found: C, 70.21; H, 5.82.
Conclusion
In summary, we have designed and synthesized a new chiral
resolving reagent specifically aimed at the separation of (S)-
amlodipine from the enantiomeric mixture of amlodipine. A
scalable and commercially viable preparation of compound 1
and a robust resolution process proved its efficacy by providing
more than 99% pure (S)-amlodipine besylate salt, which is the
final ingredient for the drug tablet, with over 99% ee. Further-
more, up to three times of recovery and reuse of compound 1
in resolution eased our concerns over the cost of production.
Experimental Section
Bis((S)-2-(4-methoxybenzyloxy)-2-oxo-1-phenylethyl)-3-
nitrophthalate (9). 3-Nitrophthalic acid (1 kg, 4.8 mol), in
DCM (21.2 L) was treated with DCC (2.56 kg, 12.4 mol), and
stirred for 2 h at room temperature. To this solution was added
compound 8 (2.84 kg, 10.4 mol) followed by DMAP (0.24 kg,
1.96 mol) and stirred for 3 h. Reaction mixture was filtered,
and the filter cake was washed with hexane (4.24 L). Filtrate
was stirred for 1 h, and insoluble solids were filtered again.
Filtrate was concentrated and crystallized in MeOH (5.28 L)
to give compound 9 in 55% yield (1.92 kg, 2.64 mol). 1H NMR
(CDCl3, 400 MHz) δ 8.55 (d, J ) 5.2 Hz, 1H), 8.03 (d, J )
7.2 Hz, 1H), 7.64 (t, J ) 8.0 Hz, 1H), 7.51-7.44 (m, 5H),
7.26-7.40 (m, 5H), 7.20-7.02 (m, 4H), 6.94-6.80 (m, 4H),
6.65 (s, 2H), 6.60 (s, 2H), 6.22 (s, 1H), 6.01 (s, 1H), 3.80 (s,
3H), 3.76 (s, 3H). Anal. Calcd for C40H33NO12: C, 66.75; H,
4.62; N, 1.95 Found: C, 66.71; H, 4.65; N, 1.92.
General. Melting points were determined on an open
1
capillary apparatus and are uncorrected. H spectra and 13C
spectra were obtained in the solvent indicated on a Bruker DPX
400 spectrometer. All purity values were obtained by HPLC
analysis.
HPLC methods. (A) Chiral HPLC method. Ultron Es-
OVM, 4.6 mm × 250 mm, isocratic elution with 78:22 pH 7
buffer solution/ACN over 20 min, 1.0 mL/min flow at 25 °C
with detection at 237 nm.
(B) Novapack C18, 4.6 mm × 150 mm, isocratic elution
with 50:35:15 pH 3.0 buffer solution/MeOH/ACN over 40 min,
1.0 mL/min flow at 25 °C with detection at 237 nm.
(C) Chiral HPLC method. ChiralCel OD, 4.6 mm × 250
mm, isocratic elution with 90:10 hexane/IPA containing 0.4%
TFA over 40 min, 2.0 mL/min at 25 °C with detection at 230
nm.
Bis((S)-mandelic acid)-3-nitrophthalate (1). Compound 9
(1.76 kg, 2.44 mol) in DCM (9 L) was treated with TFA (2.84
kg, 24.8 mol) and stirred for 2 h. Reaction mixture was
concentrated under reduced pressure and dissolved again in
DCM (9 L). DCM layer was basified with 10% NaHCO3 (16
L), and the aqueous phase was separated. The aqueous layer
(D) Novapack C18, 4.6 mm × 150 mm, isocratic elution
with 70:30 pH 3.0 buffer solution/ACN over 40 min, 1.0 mL/
min flow at 25 °C with detection at 230 nm.
(S)-4-Methoxybenzyl 2-hydroxy-2-phenyl Acetate (8). To
a cooled (0-5 °C) solution of (S)-mandelic acid (2 kg, 13.12
mol) in MeOH (4 L) was transferred Cs2CO3 (2.12 kg, 6.48
Vol. 13, No. 6, 2009 / Organic Process Research & Development
•
1385

was further washed with DCM (4 L), and pH was adjusted to
1.1-1.2 with conc. HCl. Back-extraction of the aqueous phase
with DCM (10.8 L) followed by concentration of the organic
phase gave compound 1 in 92% (1.08 kg, 2.24 mol) as pale,
yellowish solid (>99% pure by HPLC method D, >97% ee by
chiral HPLC method C). HPLC retention time: 1 ) 12.29 min,
Chiral HPLC retention times: (R)-1 ) 13.57 min, (S)-1 ) 19.17
min. 1: mp 60-62 °C; 1H NMR (CDCl3, 400 MHz) δ 8.32 (d,
J ) 5.2 Hz, 1H), 8.31 (d, J ) 7.2 Hz, 1H), 7.64 (t, J ) 8.0 Hz,
1H), 7.51-7.44 (m, 5H), 7.26-7.40 (m, 5H), 6.36 (s, 1H), 6.12
(s, 1H); 13C NMR (CDCl3, 100 MHz) δ 173.59, 173.25, 163.84,
163.03, 146.49, 135.85, 132.58, 132.41, 130.83, 129.91, 129.73,
129.68, 129.40, 129.08, 128.82, 128.71, 128.08, 127.90, 75.80,
75.75. Anal. Calcd for C24H17NO10: C, 60.13; H, 3.57; N, 2.92.
Found: C, 59.73; H, 3.55; N, 2.95.
Chiral Resolution of (R,S)-Amlodipine. Compound 1 (293
g) in MEK (1.25 L) was gradually added to (R,S)-amlodipine
(1 kg) suspension in MEK (1.25 L), and heated to 55 °C over
20 min. The mixture was cooled to 15-20 °C and stirred for
63 h. The precipitate formed was filtered, washed with tert-
butyl methyl ether (1 L) to give (S)-amlodipine·0.5 compound
1 diastereomeric salt as yellow solid in 65% yield (519 g, 96.4%
ee). Reslurry of the solid in MEK/hexane (1.5 L/0.75 L) for
16 h, filtration followed by filter cake wash with MEK/hexane
(1.5 L/0.75 L) gave solid (S)-amlodipine·0.5 compound 1
diastereomeric salt as pale-yellow solid in 61% yield (475 g,
97.5% de by chiral HPLC method A) after drying in Vacuo for
16 h. Chiral HPLC retention times: (R)-amlodipine ) 5.22 min,
(S)-amlodipine ) 7.42 min. 1H NMR (400 MHz, DMSO-d6) δ
8.82 (bs, 1H), 8.33 (d, J ) 8.04 Hz, 2H), 7.86 (t, J ) 8.04 Hz,
1H), 7.47 (d, J ) 6.76 Hz, 2H), 7.36 (d, J ) 6.76 Hz, 2H),
7.17-7.30 (m, 13H), 7.06-7.10 (m, 2H), 5.83 (s, 1H), 5.69
(s, 1H), 5.25 (s, 2H), 4.44-4.60 (dt, J ) 32.49 Hz, 4H),
3.91-3.96 (m, 4H), 3.55-3.56 (m, 5H), 3.46-3.47 (m, 5H),
2.94 (t, J ) 5.48 Hz, 4H), 2.20 (s, J ) 4.56 Hz, 6H), 1.07 (t,
J ) 7.12 Hz, 6H). Anal. Calcd for C64H67Cl2N5O20: C, 59.26;
H, 5.21; Cl, 5.47; N, 5.40. Found: C, 59.21; H, 5.18; Cl, 5.51;
N, 5.43.
93% yield (278 g, 99% pure by HPLC method B, optical purity
98.3% ee by HPLC method A). HPLC retention time: amlo-
dipine ) 6.9 min. Chiral HPLC retention time: (R)-amlodipine
) 5.22 min. (S)-amlodipine ) 7.42 min. 1H NMR (400 MHz,
MeOD) δ 7.41 (dd, J ) 7.76 Hz, 1H), 7.24 (dd, J ) 7.92 Hz,
1H), 7.15-7.17 (m, 1H), 4.07-7.10 (m, 1H), 5.42 (s, 1H), 7.69
(dd, J ) 47.53 Hz, 2H), 4.02-4.05 (m, 2H), 3.57-3.60 (m,
5H), 2.88 (t, J ) 4.72 Hz, 3H), 2.35 (s, 3H), 1.17 (t, J ) 7.12
Hz, 3H).
Recovery of Compound 1. Aqueous phase was acidified
to pH 1.0-1.5 using conc. HCl (150 mL) and extracted with
DCM (1 L). Organic phase was separated and dried over
MgSO4. The suspension was filtered, concentrated by evaporator
to give compound 1 in 90% yield (177 g, 99% pure by HPLC
method D, >97% ee by chiral HPLC method C). HPLC
retention time: 1 ) 12.29 min, Chiral HPLC retention times:
(R)-1 ) 13.57 min, (S)-1 ) 19.46 min.
Synthesis of (S)-Amlodipine Besylate. (S)-Amlodipine (270
g) was dissolved in EtOH (1.25 L) at 20 °C and cooled to 0-5
°C. Benzenesulfonic acid in H2O (0.6 L) was transferred
gradually, and the reaction mixture was further diluted with H2O
(8.7 L). Reaction mixture was stirred for 22 h at 20 °C.
Precipitate was collected by filtration, washed with cold H2O
(4 L) to give (S)-amlodipine besylate ·2H2O as white solid in
85% yield (340 g, 99% pure by HPLC method B, >99% ee by
HPLC method A, moisture 6.2%). HPLC retention time:
amlodipine ) 6.92. Chiral HPLC retention time: (R)-amlodipine
1
) 5.22 min; (S)-amlodipine ) 7.42 min; mp 68-70 °C; H
NMR (400 MHz, DMSO-d6) δ 8.35 (s, 1H), 7.70-7.90 (bs,
3H), 7.39-7.51 (m, 2H), 7.22-7.37 (m, 6H), 7.05-7.12 (m,
1H), 5.27 (s, 1H), 4.60 (q, J ) 51.57 Hz, 2H), 3.88-4.00 (m,
2H), 3.68 (s, 2H), 3.47 (s, 2H), 3.05 (t, J ) 5.44 Hz, 2H), 2.11
(s, 3H), 1.07 (t, J ) 7.12 Hz, 3H); 13C NMR (DMSO-d6, 100
MHz) δ 167.22, 165.47, 149.62, 145.37, 143.24, 132.50, 131.10,
130.04, 128.75, 128.14, 127.64, 127.10, 126.76, 126.43, 104.01,
100.64, 73.22, 30.26, 61.41, 52.32, 41.52, 38.83, 19.04, 14.40.
Supporting Information Available
(S)-Amlodipine. (S)-Amlodipine·compound 1 diastereo-
meric salt (475 g, 97.5% ee) was suspended in DCM (2.4 L).
One M NaOH (1.10 L) was added and stirred for 30 min at 20
°C. Organic phase was separated, washed with H2O (1.10 L),
and dried over MgSO4. The MgSO4 was filtered off and further
washed with DCM (1 L). Organic phase was concentrated and
dried at 50 °C in Vacuo to give (S)-amlodipine free base in
Experimental details of the synthesis of compounds 2-7,
1H NMR data for compounds 2-7. This material is available
free of charge via the Internet at http://pubs.acs.org.
Received for review March 25, 2009.
OP900070C
1386
•
Vol. 13, No. 6, 2009 / Organic Process Research & Development