A practical and efficient synthesis of (3R,4S)-1-benzyl-4- phenylpyrrolidine-3- carboxylic acid via an aziridinium ion intermediate

Article abstract of DOI:10.1021/op900230r

A practical and efficient synthesis of (3R,4S)-1-benzyl4-phenylpyr- rolidine-3-carboxylic acid (1), a key chiral building block for synthesis of biologically active compounds was established by utilizing a stereospecific and regioselective chlorination of in situ generated aziridinium ion, followed by a nitrile anion cyclization. Starting from commercially available (R)-styrene oxide and 3-(benzylamino)propionitrile, the four-step synthesis features a through process without purification of intermediates until isolation of crystalline 1. The robust, chromatography-free and reproducible synthesis of 1 achieved an 84% overall yield from (R)-styrene oxide. This highly efficient process was successfully demonstrated at pilot scale with 17 kg output of 1.

Full text of DOI:10.1021/op900230r

Organic Process Research & Development 2010, 14, 127–132
A Practical and Efficient Synthesis of (3R,4S)-1-Benzyl-4-phenylpyrrolidine-3-
carboxylic acid via an Aziridinium Ion Intermediate
Atsushi Ohigashi,*,† Takashi Kikuchi,† and Shunsuke Goto‡
Process Chemistry Labs, Astellas Pharma Inc., 160-2, Akahama, Takahagi-shi, Ibaraki 318-0001, Japan
Abstract:
A practical and efficient synthesis of (3R,4S)-1-benzyl-4-phenylpyr-
rolidine-3-carboxylic acid (1), a key chiral building block for
synthesis of biologically active compounds was established by
utilizing a stereospecific and regioselective chlorination of in situ
generated aziridinium ion, followed by a nitrile anion cyclization.
Starting from commercially available (R)-styrene oxide and
3-(benzylamino)propionitrile, the four-step synthesis features a
through process without purification of intermediates until isola-
tion of crystalline 1. The robust, chromatography-free and
Figure 1. (3R,4S)-1-Benzyl-4-phenylpyrrolidine-3-carboxylic
acid (1).
During an ongoing project, we required multikilogram quantities
of (3R,4S)-1-benzyl-4-phenylpyrrolidine-3-carboxylic acid (1)
for the synthesis of pharmaceutical candidates to support
preclinical and clinical studies (Figure 1). Compound 1 is a
reproducible synthesis of 1 achieved an 84% overall yield from
(R)-styrene oxide. This highly efficient process was successfully
demonstrated at pilot scale with 17 kg output of 1.
trisubstituted N-benzyl pyrrolidine with a C3 carboxylic acid
and a C4 phenyl moiety in the trans-3R,4S configuration.
Unfortunately, this compound was unavailable from commercial
sources in multikilogram quantities.
Introduction
trans-3,4-Disubstituted pyrrolidines are key chiral building
blocks for the synthesis of various biologically active com-
pounds such as thrombin inhibitors,1 coagulation factor Xa
inhibitors2 and CCR5 receptor antagonists.3 There is often the
requirement for such trans-pyrrolidines to be enantiopure.
While a wide range of approaches to trans-3,4-disubstituted
pyrrolidines are reported, there are far fewer examples for their
preparation on a large scale. One of the most extensively studied
and widely used methods involves an asymmetric 1,3-dipolar
cycloaddition of nonstabilized azomethine ylides to olefinic
dipolarophiles.4a-e However, this chiral auxiliary-based approach
gave a mixture of diastereomeric 1,3,4-trisubstituted pyrrolidines
with moderate diastereoselectivity (up to 48% de).4b As the
medicinal chemistry group employed this approach for the
preparation of screening/optimization samples, we first evaluated
it for the multikilogram preparation of 1 by choosing com-
mercially available and inexpensive (S)-4-phenyl-2-oxazolidi-
none as a chiral auxiliary (Scheme 1). As a result, the desired
diastereomer 6a was obtained as colorless oil from trans-
cinnamoyl chloride (2) in 64% isolated yield (2 steps), and
moderate diastereoselectivity was consistent with that of the
literature. This strategy proved to have several drawbacks to
overcome for multikilogram production: (1) silica gel chroma-
tography was necessary for the separation of 6a and 6b; (2)
the ylide 5 precursor, N-benzyl-N-(methoxymethyl)trimethyl-
silylmethylamine (4), was not commercially available in mul-
tikilogram quantities. Therefore, this approach did not encourage
further investigation, and the development of a new practical
and efficient synthesis of 1 was desired.
* To whom correspondence should be addressed. Telephone: +81-293-23-
4413. Fax: +81-293-24-2708. E-mail: atsushi.oohigashi@jp.astellas.com.
† Astellas Pharma Inc.
‡ Sankyo Kasei Co., Ltd.
(1) Fevig, J. M.; Abelman, M. M.; Brittelli, D. R.; Kettner, C. A.; Knabb,
R. M.; Weber, P. C. Bioorg. Med. Chem. Lett. 1996, 6, 295.
(2) Fevig, J. M.; Buriak, J., Jr.; Stouten, P. F. W.; Knabb, R. M.; Lam,
G. N.; Wong, P. C.; Wexler, R. R. Bioorg. Med. Chem. Lett. 1999, 9,
1195.
(3) Some examples: (a) Kim, D.; Wang, L.; Hale, J. J.; Lynch, C. L.;
Budhu, R. J.; MacCoss, M.; Mills, S. G.; Malkowitz, L.; Gould, S. L.;
DeMartino, J. A.; Springer, M. S.; Hazuda, D.; Miller, M.; Kessler,
J.; Hrin, R. C.; Carver, G.; Carella, A.; Henry, K.; Lineberger, J.;
Schleif, W. A.; Emini, E. A. Bioorg. Med. Chem. Lett. 2005, 15, 2129.
(b) Shen, D.-M.; Shu, M.; Mills, S. G.; Chapman, K. T.; Malkowitz,
L.; Springer, M. S.; Gould, S. L.; DeMartino, J. A.; Siciliano, S. J.;
Kwei, G. Y.; Carella, A.; Carver, G.; Holmes, K.; Schleif, W. A.;
Danzeisen, R.; Hazuda, D.; Kessler, J.; Lineberger, J.; Miller, M. D.;
Emini, E. A. Bioorg. Med. Chem. Lett. 2004, 14, 935. (c) Shen, D.-
M.; Shu, M.; Mills, S. G.; Willoughby, C. A.; Shah, S.; Lynch, C. L.;
Hale, J. J.; Mills, S. G.; Chapman, K. T.; Malkowitz, L.; Springer,
M. S.; Gould, S. L.; DeMartino, J. A.; Siciliano, S. J.; Lyons, K.;
Pivnichny, J. V.; Kwei, G. Y.; Carella, A.; Carver, G.; Holmes, K.;
Schleif, W. A.; Danzeisen, R.; Hazuda, D.; Kessler, J.; Lineberger,
J.; Miller, M. D.; Emini, E. A. Bioorg. Med. Chem. Lett. 2004, 14,
941. (d) Shu, M.; Loebach, J. L.; Parker, K. A.; Mills, S. G.; Chapman,
K. T.; Shen, D.-M.; Malkowitz, L.; Springer, M. S.; Gould, S. L.;
DeMartino, J. A.; Siciliano, S. J.; Salvo, J. D.; Lyons, K.; Pivnichny,
J. V.; Kwei, G. Y.; Carella, A.; Carver, G.; Holmes, K.; Schleif, W. A.;
Danzeisen, R.; Hazuda, D.; Kessler, J.; Lineberger, J.; Miller, M. D.;
Emini, E. A. Bioorg. Med. Chem. Lett. 2004, 14, 947. (e) Shankaran,
K.; Donnelly, K. L.; Shah, S. K.; Guthikonda, R. N.; MacCoss, M.;
Mills, S. G.; Gould, S. L.; Malkowitz, L.; Siciliano, S. J.; Springer,
M. S.; Carella, A.; Carver, G.; Hazuda, D.; Holmes, K.; Kessler, J.;
Lineberger, J.; Miller, M. D.; Emini, E. A.; Schleif, W. A. Bioorg.
Med. Chem. Lett. 2004, 14, 3419.
We surveyed reported synthetic methods for trans-3,4-
disubstituted pyrrolidines applicable to large-scale production.
(4) For some applications of this reaction, see: (a) Ma, Z.; Wang, S.;
Cooper, C. S.; Fung, A. K. L.; Lynch, J. K.; Plagge, F.; Chu, D. T. W.
Tetrahedron: Asymmetry 1997, 8, 883. (b) Karlsson, S.; Han, F.;
Ho¨gberg, H.-E.; Caldirola, P. Tetrahedron: Asymmetry 1999, 10, 2605.
(c) Ling, R.; Ekhato, V.; Rubin, J. R.; Wustrow, D. J. Tetrahedron
2001, 57, 6579. (d) Karlsson, S.; Ho¨gberg, H.-E. Tetrahedron:
Asymmetry 2001, 12, 1977. (e) Kotian, P. L.; Lin, T.-H.; El-Kattan,
Y.; Chand, P. Org. Process Res. DeV. 2005, 9, 193.
10.1021/op900230r  2010 American Chemical Society
Published on Web 11/30/2009
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
127

Scheme 1. Asymmetric 1,3-dipolar cycloaddition approach
methodology to the synthesis of 1. In particular, we needed to
address the problems associated with the regioselective ring-
opening of in situ generated epoxide 8: (1) a bulky nucleophile
such as tert-butylamine has a fundamental role in good
regioselectivity; (2) excess tert-butylamine (5 volumes) is
essential to improve the regioselectivity of the reaction, and
the unreacted tert-butylamine was removed by evaporation; (3)
crystallization to isolate the desired regioisomer 9a was neces-
sary for the separation from the undesired regioisomer 9b.
Inspired by this strategy, we envisioned a synthetic route
for 1 starting from commercially available (R)-styrene oxide
(13) and 3-(benzylamino)propionitrile (14) (Scheme 3). The
advantage of this synthetic plan did not rely upon stereocontrol
at the C3 and C4 positions. On the other hand, a common
problem associated with regioselective epoxide opening still
existed. Furthermore, removal of 14 by a simple evaporation
was impractical due to its high boiling point. With the above
in mind, we investigated the ring-opening reaction of 13 with
14. This reaction was completed in 42 h to afford an 89:11
mixture of the corresponding two isomeric amino alcohols 15a
and 15b as an oil with the best selectivity.7 However, separation
of 15a and 15b was difficult with any practical purification
methods. Furthermore, the excess 14 which was necessary to
attain a practical reaction time posed an issue of removing 14
from the desired 15a.
On the other hand, novel methods of hydroxyl group
activation at the benzylic position have been reported. On the
basis of these reports,8a-d we envisaged mesylation of the
regioisomeric mixture of amino alcohols 15a and 15b using
MsCl/Et3N should converge with a common aziridinium
intermediate 18 which would afford an isomerically and
enantiomerically pure ꢀ-chloroamine 19 as a result of spontane-
ous ring-opening (Scheme 4). If this sequence is realized, the
epoxide-opening regioselectivity will be of no consequence.
The mesylation hypothesis was confirmed by successful
transformation of isolated amino alcohols 15a and 15b to the
single ꢀ-chloroamine 19 in good yield. Base treatment of 19
induced an SN2-type intramolecular cyclization to afford the
desired mixture of pyrrolidine nitriles 16a and 16b which were
converted to a single isomeric pyrrolidine carboxylic acid 1 after
the epimerization/saponification.6b We report here a practical
and efficient synthesis of the single enantiomer of 1, utilizing
stereospecific and regioselective chlorination of the in situ
generated aziridinium ion intermediate 18.
to 1
Scheme 2. Synthesis of trans-pyrrolidine carboxylic acid 12
via a nitrile anion cyclization
There were two other reported methods for the construction of
trans-3,4-disubstituted pyrrolidines. One method involved the
Rh-catalyzed asymmetric 1,4-arylation of aryl boronic acid to
3-pyrroline esters,5 and the other is the enantioselective in-
tramolecular cyclization of the acyclic hydroxynitriles.6a,b The
former method gave the desired product in good enantioselec-
tivity with bulky esters, but poor conversion. Better yields were
obtained with smaller esters; however, the enantioselectivity was
significantly lower. The latter method highlighted an intramo-
lecular nitrile-anion cyclization which afforded an enantio-pure
trans-pyrrolidine carboxylic acid 12 as a result of consecutive
epimerization/saponification of a pyrrolidine nitrile 11 (Scheme
2).6b However, there were limitations in the application of this
Results and Discussion
We investigated the ring-opening reaction of (R)-styrene
oxide (13) with 3-(benzylamino)propionitrile (14) (Table 1). The
reaction in DMF did not occur at all (entry 1) and proceeded
sluggishly in protic solvents such as EtOH, i-PrOH, and
(7) The ring-opening of (R)-styrene oxide was conducted with 3-(benzyl-
amino)propionitrile (2.0 equiv) in i-PrOH (1 volume) at 90 °C for
42 h.
(8) For some applications of this reaction, see: (a) Chuang, T.-H.;
Sharpless, K. B. Org. Lett. 2000, 2, 3555. (b) Chuang, T.-H.; Sharpless,
K. B. HelV. Chim. Acta 2000, 83, 1734. (c) Frizzle, M. J.; Caille, S.;
Marshall, T. L.; McRae, K.; Nadeau, K.; Guo, G.; Wu, S.; Martinelli,
M. J.; Moniz, G. A. Org. Process Res. DeV. 2007, 11, 215. (d)
Anderson, S. R.; Ayers, J. T.; DeVries, K. M.; Ito, F.; Mendenhall,
D.; Vanderplas, B. C. Tetrahedron: Asymmetry 1999, 10, 2655.
(5) Belyk, K. M.; Beguin, C. D.; Palucki, M.; Grinberg, N.; DaSilva, J.;
Askin, D.; Yasuda, N. Tetrahedron Lett. 2004, 45, 3265.
(6) (a) Achini, R. HelV. Chim. Acta 1981, 64, 2203. (b) Chung, J. Y. L.;
Cvetovich, R.; Amato, J.; McWilliams, J. C.; Reamer, R.; DiMichele,
L. J. Org. Chem. 2005, 70, 3592.
128
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

Scheme 3. Synthetic strategy towards 1
Scheme 4. Regio- and stereoselective chlorination of the in situ generated aziridinium ion intermediate 18
Table 1. Effect of solvents and additives in the ring-opening of 13 with 14
HPLC area %
15a + 15b
ratioc
15a:15b
entry
14 (equiv)
solventa
additiveb
temperature (°C)
time (h)
d
1
2
3
4
5
6
7
8
9
1.5
1.5
1.5
1.5
1.2
1.2
1.2
1.2
1.2
DMF
EtOH
none
75
75
75
14
15
42
21
6
-
none
92
91
88
91e
82:18
85:15
82:18
76:24
i-PrOH
n-BuOH
60% EtOH
EtOH
none
none
100
80
none
f
MsOH
AcOH
PhOH
ZnI2
85
2
-
EtOH
85
2
78
95
66
57:43
68:32
45:55
EtOH
85
6
toluene
30
2
a 5 volumes in entries 1-5 and 1 volume in entries 6-9. b 1.0 equiv in entries 6-8 and 0.1 equiv in entry 9. c A ratio of each HPLC peak areas among 15a and 15b. d Not
reacted. e Unacceptable amounts of impurities formed. f Decomposition of 13.
n-BuOH to afford the mixture of regioisomers 15a and 15b
(entries 2-4). The ring-opening reaction in aqueous 60% EtOH9
proceeded very quickly but resulted in the formation of
unacceptable amounts of byproducts (entry 5). Among the
solvents tested, EtOH seemed to be the best from the viewpoints
of reaction rate and product quality (entry 2).
The serious disadvantage of this reaction (entry 2) was that
it required excess 14 and long reaction times. To address these
issues, various additives such as MsOH, AcOH, PhOH, and
ZnI2 were investigated.10a-e The use of MsOH gave a disap-
pointing result because of the decomposition of 13 (entry 6).
Except for MsOH, the additives enhanced the reaction rates
dramatically and reduced the amount of 14 required for the
aminolysis of 13 (entries 7-9). On the other hand, these
additives not only decreased the selectivity for 15a but also
tended to increase the amount of impurities. Since the reaction
of the mixture with MsCl/Et3N converts both regioisomers 15a
and 15b into a single ꢀ-chloroamine 19 via the common
aziridinium ion intermediate 18, the regioselectivity of ring-
opening of 13 could be addressed. During the development of
a drug candidate, we have encountered similar problems as-
sociated with the aminolysis of styrene oxide. Among many
acidic additives, PhOH gave a best result in terms of reactivity
and selectivity (unpublished results). Thus, we tried PhOH as
an additive and obtained a best result using 1.0 equiv of PhOH
in 1 volume of EtOH (entry 8). With this condition, 1.2 equiv
of 14 was sufficient for reaction completion. As described
above, when AcOH or ZnI2 was used, the reaction proceeded
very quickly but suffered from the formation of unacceptable
amounts of byproducts (entries 7 and 9). After complete reaction
(entry 8), toluene was added to the reaction mixture, and
(9) Azizi, N.; Saidi, M. Org. Lett. 2005, 7, 3649.
(10) For recent examples of various methods for activating epoxides, see:
(a) Heydari, A.; Mehrdad, M.; Maleki, A.; Ahmadi, N. Synthesis 2004,
1563. (b) Chakraborti, A. K.; Rudrawar, S.; Kondaskar, A. Eur. J.
Org. Chem. 2004, 3597. (c) Pachon, L. D.; Gamez, P.; van Brussel,
J. J. M.; Reedijk, J. Tetrahedron Lett. 2003, 44, 6025. (d) Placzek,
A. T.; Donelson, J. L.; Trivedi, R.; Gibbs, R. A.; De, S. K. Tetrahedron
Lett. 2005, 46, 9026. (e) Shivani; Pujala, B.; Chakraborti, A. K. J.
Org. Chem. 2007, 72, 3713.
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
129

Scheme 5. Alternative synthetic route for 1 starting from
(S)-2-phenylglycinol (20)
aqueous acid-base extractions were used to remove PhOH and
the unreacted amine 14. Thus, the oily mixture of 15a and 15b
(68:32)11 was obtained and used without further purification
for the subsequent chlorination.
In the chlorination step, the mixture of regioisomers 15a and
15b was quantitatively converted via the aziridinium ion
intermediate 18 into single ꢀ-chloroamine 19 by treatment with
MsCl/Et3N in toluene at room temperature. Although we did
not isolate and identify 18, the fact that a single regioisomer
19 was obtained from the mixture of 15a and 15b implies the
existence of 18. Interestingly, chlorination of the mixture of
15a and 15b with thionyl chloride instead of MsCl/Et3N
afforded only racemic 19 due to the participation of benzylic
cation under highly acidic conditions. Therefore, the MsCl/Et3N
system is essential for the stereospecific chlorination. After
completion, the reaction mixture was washed with water,
aqueous sodium bicarbonate, and sodium chloride and concen-
trated under reduced pressure to give 19 as an oil. Fortunately,
19 was sufficiently stable under these conditions and was used
without any further purification in the next cyclization step.
In the cyclization step, ꢀ-chloroamine 19 was treated with
1.1 equiv of sodium hexamethyldisilazide (NaHMDS) in toluene
at -30 to -20 °C. The reaction was complete within 10 min,
and after an aqueous workup yielded an 89/11 trans/cis mixture
of pyrrolidine nitriles (16a and 16b)12 as an oil. The obtained
mixture was used without further purification for the hydrolysis
since the mixture of 16a and 16b was successfully converted
into a single trans-pyrrolidine carboxylic acid 1 by hydrolysis,
as described later. The attempts to combine the chlorination
and cyclization steps in a one-pot fashion resulted in the large
amount of unreacted 19.
The final step in the synthesis involves a basic epimerization/
saponification of trans-/cis-pyrrolidine nitriles 16a and 16b to
trans-pyrrolidine carboxylic acid 1. On the basis of Chung
methodology,6b the hydrolysis of the mixture of 16a and 16b
with NaOH in aqueous EtOH at reflux temperature was
investigated. This reaction proceeded through an intermediate
and was completed in 7 h. The intermediate was not identified,
but it is likely to be the corresponding amide. Adjustment of
the pH to around 4 by using 35% hydrochloric acid and
subsequent addition of water afforded single 1 as a crystalline
solid with an overall yield of 84% (four steps). These procedures
produced 1 with an optical purity of 99.6% ee13 starting from
(R)-styrene oxide (13) with >99% ee. This observation indicates
that there was no chirality leakage in the whole process. In
addition, this highly efficient process was successfully demon-
strated at pilot scale with 17 kg output of 1.
ꢀ-chloroamine 19 via the in situ aziridinium ion intermediate
18, we designed an alternative synthetic route for 1 starting from
commercially available and cheaper (S)-2-phenylglycinol (20)
(Scheme 5). We investigated the conjugate addition of N-benzyl-
(S)-2-phenylglycinol (21) with acrylonitrile since the method
for preparing 21 from 20 had been reported.14a,b Treatment of
21 with acrylonitrile and AcOH gave 15b in 88% isolated yield
(unoptimized). In the chlorination and cyclization steps, the
same reaction conditions as above were employed. Amino
alcohol 15b was converted into the mixture of 16a and 16b
via 19 in 67% yield (unoptimized). We think with further
optimization this alternative route may be more useful for the
synthesis of 1.
Conclusions
In this contribution, a practical and efficient synthesis of
(3R,4S)-1-benzyl-4-phenylpyrrolidine-3-carboxylic acid (1), a
key intermediate for biologically active compounds, is de-
scribed. The advantage over the original methods is as follows:
(1) access to 1 from commercially available raw materials, (2)
improvement of overall yield and cost-efficiency, (3) avoidance
of column chromatographic purifications throughout the whole
process. Furthermore, this synthesis has permitted the prepara-
tion of multikilogram quantities of 1.
Experimental Section
1H NMR spectra were recorded on a JEOL JNM-AL400
using TMS as an internal standard. IR spectra were recorded
on a Horiba FT-720 Fourier transform infrared spectrometer.
MS spectra were recorded on a Hewlett-Packard 1100LC/MSD
mass spectrometer using API-ES ionization. HPLC was per-
formed using Shimazu LC-10AT systems. All reagents and
solvents were commercially available and used without further
purification. All reactions were performed under an atmosphere
of nitrogen.
Our attention then turned to the investigation of an alternative
approach starting from a cheaper chiral starting material than
(R)-styrene oxide (13). As amino alcohol 15b was successfully
converted into the isomerically and enantiomerically pure
Mixture of 3-{Benzyl[(2R)-2-hydroxy-2-phenylethyl]am-
ino}propanenitrile (15a) and 3-{Benzyl[(1S)-2-hydroxy-1-phe-
nylethyl]amino}propanenitrile (15b). (R)-(+)-Styrene oxide
(13) (8.47 kg, >99% ee), 3-(benzylamino)propionitrile (14)
(14.40 kg), PhOH (7.00 kg), and EtOH (1.8 L) were combined
in a reaction vessel, and the mixture was stirred at 86-89 °C
for 8 h. After cooling to 20-30 °C, toluene (32 L), and water
(45 L) were added to the reaction mixture. The mixture was
adjusted to pH 4.0 using 35% hydrochloric acid (1.6 L). The
(11) The ratio of each HPLC area % among 15a and 15b, see Experimental
Section.
(12) The ratio of each HPLC area % among 16a and 16b, see Experimental
Section.
(13) Enantiomeric purity of 1 was determined by chiral HPLC assay of
the corresponding pyrrolidine methanol, see Experimental Section.
(14) (a) Monbaliu, J.-C.; Tinant, B.; Marchand-Brynaert, J. Heterocycles
2008, 75, 2459. (b) Blanchet, J.; Bonin, M.; Chiaroni, A.; Micouin,
L.; Riche, C.; Husson, H.-P. Tetrahedron Lett. 1999, 40, 2935.
130
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

organic phase was separated and washed with water (18 L).
To the organic phase was added water (45 L) and the mixture
was adjusted to pH 12.0 using 24% aqueous sodium hydroxide
solution (6.6 L), and then the organic phase was separated. To
the organic phase was added water (45 L) and the mixture was
adjusted to pH 12.1 using 24% aqueous sodium hydroxide
solution (0.6 L). The organic phase was separated, and then
washed with water (27 L) and aqueous sodium chloride solution
[sodium chloride (5.4 kg) and water (22 L)] successively. The
organic phase was concentrated under vacuum at 40-50 °C
until approximately a 25 L of total volume was obtained.
Toluene (17 L) was added, and the concentration was continued
under vacuum at 40-50 °C to give a mixture of 15a and 15b
(21.50 kg, overweight) as a yellow oil, and it was used without
further purification for the next step. HPLC analysis of the
mixture of regioisomers 15a and 15b (HPLC condition A,
column: TSKgel ODS-80Tm 4.6 mm × 150 mm, eluent: 0.05
M KH2PO4 aq/MeCN, 40:60, flow rate: 1.0 mL/min, column
temperature: 40 °C, wavelength: 210 nm) showed a 68:32 ratio
of 15a:15b with tR ) 5.23 min and tR ) 4.70 min, respectively.
Analytically pure 15a and 15b were obtained by silica gel
chromatography. 3-{Benzyl[(2R)-2-hydroxy-2-phenylethyl]-
amino}propanenitrile (15a): 1H NMR (400 MHz, CDCl3) δ
7.39-7.24 (m, 10H), 4.76-4.69 (m, 1H), 3.92 (d, J ) 13.6
Hz, 1H), 3.62 (d, J ) 13.6 Hz, 1H), 3.47 (d, J ) 0.8 Hz, 1H),
3.02-2.95 (m, 1H), 2.85-2.75 (m, 2H), 2.66-2.60 (m, 1H),
2.47-2.43 (m, 2H); IR (ATR): 3417, 3030, 2834, 2243,
1603, 1492, 1451, 1375, 1342 cm-1; MS (API-ES+): 303 [M
+ Na]+. 3-{Benzyl[(1S)-2-hydroxy-1-phenylethyl]amino}pro-
panenitrile (15b): 1H NMR (400 MHz, CDCl3) δ 7.42-7.23
(m, 10H), 4.14-4.07 (m, 1H), 3.96-3.92 (m, 1H), 3.86 (d, J
) 13.6 Hz, 1H), 3.75-3.69 (m, 1H), 3.38 (d, J ) 13.6 Hz,
1H), 3.09-3.02 (m, 1H), 2.70 (dd, J ) 2.0, 9.2 Hz, 1H),
2.63-2.57 (m, 1H), 2.46-2.23(m, 2H); IR (ATR): 3480,
3060, 3028, 2933, 2844, 2247, 1494, 1452 cm-1; MS (API-
ES+): 303 [M + Na]+.
3-{Benzyl[(2R)-2-chloro-2-phenylethyl]amino}propaneni-
trile (19). The mixture of 15a and 15b (21.50 kg), toluene (84
L), Et3N (10.60 kg) were combined in a reaction vessel. After
cooling to below 20 °C, MsCl (10.30 kg) was added dropwise
at 8-20 °C, and the mixture was stirred at 20-27 °C for 2.5 h.
To the reaction mixture was added water (63 L) slowly at below
30 °C. The organic phase was separated and washed with
aqueous sodium bicarbonate solution [sodium bicarbonate
(10.50 kg) and water (63 L)] and aqueous sodium chloride
solution [sodium chloride (12.60 kg) and water (63 L)]
successively. The organic phase was concentrated under vacuum
at 35-45 °C until ∼40 L total volume was obtained. Toluene
(42 L) was added, and the concentration was continued under
vacuum at 35-45 °C to give 19 (41.14 kg, overweight) as an
orange oil. It was used without further purification for the next
step.
Mixture of (3R,4S)-1-Benzyl-4-phenylpyrrolidine-3-car-
bonitrile (16a) and (3S,4S)-1-Benzyl-4-phenylpyrrolidine-
3-carbonitrile (16b). Compound 19 (41.14 kg) and toluene (112
L) were combined in a reaction vessel. After cooling to -30
to -20 °C, 1 M NaHMDS in THF solution (75.17 kg) was
added slowly to the mixture at this temperature. The mixture
was stirred at -30 to -20 °C for 10 min and quenched with
water (4.5 L) at -30 °C. After warming to 23 °C, water (85 L)
was added to the mixture. The organic phase was separated
and washed with aqueous sodium chloride solution [sodium
chloride (13.40 kg) and water (67 L)]. The organic phase was
concentrated under vacuum at 35-45 °C until HPLC analysis
of the analytical sample showed residual toluene at 20-25
HPLC area % (HPLC condition A) to give a mixture of 16a
and 16b. It was used without isolation for the next step. HPLC
analysis of the mixture of diastereomers 16a and 16b (HPLC
condition A) showed an 89:11 ratio of 16a:16b with tR ) 8.34
min and tR ) 6.79 min, respectively.
Analytically pure 16a and 16b were obtained by silica gel
chromatography. (3R,4S)-1-Benzyl-4-phenylpyrrolidine-3-
carbonitrile (16a): 1H NMR (400 MHz, CDCl3) δ 7.37-7.24
(m, 10H), 3.74 (d, J ) 12.8 Hz, 1H), 3.69 (d, J ) 12.8 Hz,
1H), 3.14-3.03 (m, 3H), 2.93 (dd, J ) 8.8, 6.9 Hz, 1H), 2.91
(dd, J ) 8.8, 6.9 Hz, 1H), 2.78 (dd, J ) 9.6, 6.4 Hz, 1H); IR
(ATR): 3062, 3029, 2961, 2918, 2239, 1603, 1494, 1453 cm-1;
MS (API-ES+): 263 [M + H]+. (3S,4S)-1-Benzyl-4-phe-
1
nylpyrrolidine-3-carbonitrile (16b): H NMR (400 MHz,
CDCl3) δ 7.38-7.25 (m, 10H), 3.76 (s, 2H), 3.66 (dd, J ) 8.6,
7.7 Hz, 1H), 3.47 (ddd, J ) 9.6, 7.5, 6.2 Hz, 1H), 3.22-3.12
(m, 2H), 2.98 (dd, J ) 9.5, 6.2 Hz, 1H), 2.89 (dd, J ) 9.5, 8.0
Hz, 1H); IR (ATR): 3031, 2805, 2790, 2240, 1494, 1452, 1377,
1346 cm-1; MS (API-ES+): 263 [M + H]+.
(3R,4S)-1-Benzyl-4-phenylpyrrolidine-3-carboxylic Acid
(1). To the mixture of 16a and 16b in the reaction vessel was
added EtOH (99 L). After cooling to 0 °C, aqueous sodium
hydroxide solution [sodium hydroxide (9.00 kg) and water (20
L)] was added to the mixture at 0-13 °C. The mixture was
heated to reflux at 77-80 °C for 7 h and then allowed to cool
to 0 °C. The mixture was adjusted to pH 4.0 using 35%
hydrochloric acid (23.5 L) at 26 °C for crystallization. Water
(197 L) was added dropwise at 10-30 °C, and the mixture
was adjusted to pH 4.0 using 24% aqueous sodium hydroxide
solution (2.4 L) at 23 °C. The resulting slurry was cooled at 1
°C, filtered, and washed with aqueous EtOH solution [EtOH
(20 L) and water (39 L)] and then toluene (59 L). The wet
crystal was dried under vacuum at 43 °C to give 16.66 kg of 1
as a crystalline white solid [84% yield based on (R)-styrene
oxide]. 1H NMR (400 MHz, NaOD + D2O) δ 7.16-7.02 (m,
10H), 3.48 (d, J ) 12.4 Hz, 1H), 3.37 (d, J ) 12.8 Hz, 1H),
3.30 (q, J ) 8.4 Hz, 1H), 2.87 (t, J ) 7.6 Hz, 1H),
2.77-2.66(m, 3H), 2.39 (t, J ) 8.8 Hz, 1H); IR (ATR): 3440,
3396, 3306, 1673, 1602, 1580, 1496, 1451, 1398, 1377 cm-1;
MS (API-ES+): 282 [M + H]+.
An analytical sample of 19 was purified by silica gel
1
chromatography. H NMR (400 MHz, CDCl3) δ 7.37-7.23
Enantiomeric purity of 1 was determined by chiral HPLC
assay of the corresponding pyrrolidine methanol, [(3R,4S)-1-
benzyl-4-phenylpyrrolidin-3-yl]methanol, as follows: 70% Red-
Al in toluene (8 equiv) was added slowly to 1 (1 equiv) in
toluene (5 volumes) at 0-15 °C. The reaction mixture was
(m, 10H), 4.80 (t, J ) 6.8 Hz, 1H), 3.74 (d, J ) 13.6 Hz, 1H),
3.68 (d, J ) 13.6 Hz, 1H), 3.21-3.08 (m, 2H), 2.85 (t, J ) 6.8
Hz, 2H), 2.30-2.25 (m, 2H); IR (ATR): 3062, 3029, 2960,
2248, 1494, 1453 cm-1; MS (API-ES+): 263 [M - Cl]+.
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
131

stirred at room temperature for 8 h. The mixture was slowly
added to aqueous sodium hydroxide solution [sodium hydroxide
(12 equiv) and water (7 volumes)] at 0-20 °C; subsequently,
water (3 volumes) was slowly added at 10-30 °C. The organic
phase was separated and then washed with water (3 volumes)
and aqueous sodium chloride solution [sodium chloride (0.6
weight) and water (3 volumes)] successively. The organic phase
was concentrated under vacuum to dryness to afford a white
solid, identified as [(3R,4S)-1-benzyl-4-phenylpyrrolidin-3-
yl]methanol by comparison to authentic samples.15 The chiral
assay (HPLC condition B, column: CHIRALCEL OD-H 4.6
mm × 250 mm, eluent: n-hexane/i-PrOH/diethyamine, 90:10:
0.1, flow rate: 1.0 mL/min, column temperature: 40 °C,
wavelength: 220 nm) gave a 99.8:0.2 ratio of (3R,4S):(3S,4R)
enantiomers (99.6% ee) with tR ) 7.81 min and tR ) 6.28 min,
respectively.
concentrated under vacuum at 40 °C, and purification by silica
gel chromatography (SiO2: 20 g, elution: AcOEt/n-heptane, 1:4)
gave 2.18 g of 15b as a colorless oil (88% yield). It was
identified as 15b by comparison to an authentic sample. A crude
mixture of 16a and 16b was obtained from 15b (0.988 mg)
using the experimental procedure as described above (690 mg,
68% yield). The yield was estimated from the peak area of the
product on HPLC (HPLC condition A). HPLC analysis of the
mixture of 16a and 16b (HPLC condition A) showed a 92:8
ratio of 16a:16b.
Acknowledgment
We are grateful to Dr. S. Hachiya for helpful discussions
about the medicinal synthetic route to 1 and Mr. K. Ito for his
assistance in developing the HPLC assays used to characterize
the chiral pyrrolidines.
Mixture of (3R,4S)-1-Benzyl-4-phenylpyrrolidine-3-car-
bonitrile (16a) and (3S,4S)-1-Benzyl-4-phenylpyrrolidine-
3-carbonitrile (16b) starting from N-Benzyl-(S)-2-phenylg-
lycinol (21). Acetic acid (528 mg) was added to (S)-N-benzyl-
2-phenylglycinol (21) (2.00 g) in acrylonitrile (8.40 g) at room
temperature. The reaction mixture was refluxed for 41 h and
then allowed to cool to room temperature. The mixture was
Received for review August 30, 2009.
OP900230R
(15) The authentic samples of [(3R,4S)-1-benzyl-4-phenylpyrrolidin-3-
yl]methanol and its enantiomer were synthesized respectively in our
laboratories according to the literature 4b.
132
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development