Asymmetric synthesis of (-)-paroxetine using PLE hydrolysis

Article abstract of DOI:10.1016/S0040-4039(00)00942-4

(-)-Paroxetine hydrochloride was produced asymmetrically in seven steps starting from 4-fluoro-benzaldehyde. The stereocenter at C-4 was initially set through desymmetrization of glutaric acid bis methyl ester 4 by PLE hydrolysis. A unique one-pot reduction-alkylation procedure was then developed to provide entry to lactam 2 with the appropriate absolute stereochemistry. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4039(00)00942-4

Tetrahedron Letters 41 (2000) 5647±5651
Asymmetric synthesis of (^)-paroxetine using PLE hydrolysis
Marvin S. Yu,* Ivan Lantos, Zhi-Qiang Peng, J. Yu and Thomas Cacchio
SmithKline Beecham Pharmaceuticals, Synthetic Chemistry Department, 709 Swedeland Road, PO Box 1539,
King of Prussia, PA 19406, USA
Received 13 April 2000; accepted 30 May 2000
Abstract
(^)-Paroxetine hydrochloride was produced asymmetrically in seven steps starting from 4-¯uoro-
benzaldehyde. The stereocenter at C-4 was initially set through desymmetrization of glutaric acid bis
methyl ester 4 by PLE hydrolysis. A unique one-pot reduction±alkylation procedure was then developed to
provide entry to lactam 2 with the appropriate absolute stereochemistry. # 2000 Elsevier Science Ltd. All
rights reserved.
Keywords: enzymes and enzyme reactions; piperidines; stereochemistry.
(^)-Paroxetine hydrochloride, 1, marketed as Paxil/Seroxat, is a selective serotonin reuptake
inhibitor used in the treatment of depression, obsessive compulsive disorder, and panic which
currently generates sales in excess of $1.5 billion/year. In our role as process chemists we desired a
new asymmetric synthesis1 which would be amenable to large-scale preparation. We report herein
a new synthesis of 1, whose key feature is a porcine liver esterase mediated asymmetric
desymmetrization2 followed by a novel reduction±alkylation sequence to provide the required
stereochemistry at the benzylic center of the piperidine. The retrosynthetic analysis shown in
Scheme 1, envisions paroxetine arising from optically pure lactam 2, which would be derived
from monomethyl glutaric acid 3. Enzymic hydrolysis of 4 would give optically enriched 3.
Scheme 1.
* Corresponding author. Tel: 610-270-6979; fax: 610-270-4022; e-mail: marvin_s_yu@sbphrd.com
0040-4039/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(00)00942-4

5648
The synthesis (Scheme 2) begins with the preparation of bis ester 4 in 75% yield by reaction of p-
¯uorobenzaldehyde with ethyl acetoacetate and NaOH, followed by esteri®cation.3 An extensive
study of numerous variables eventually found that hydrolysis of 4 with pig liver esterase (600U/
gram 4, room temperature, 10% aqueous acetone bu€ered to pH=7.0) consistently a€orded
optically active acid ester 3 in 86% yield and 95% ee.4 The absolute stereochemistry of 3 was
initially not determined, since using borane or LiBH4 could selectively reduce either the acid or
the ester, respectively.5 Therefore, regardless of the absolute stereochemical outcome of 3, the
desired enantiomer of paroxetine could, in theory, be obtained, and the optical rotation of the
material produced would then be compared to authentic drug substance. Reduction of the acid
functionality of 3 with borane provided 5 in 94% yield. Alcohol 5 was mesylated and treated with
benzylamine to provide lactam 6 in 82% yield. The crystalline lactam 6 obtained was found to be
>99% ee.6 Acylation a€orded 7 (88%). The relative stereochemistry of the aryl group and the C-
3 side chain was determined by examination of the 1H NMR coupling constants between the C-3
.
and C-4 protons (J=11.0 Hz). Reduction of 7 with either LAH (71%) or BH3 THF in re¯uxing
THF (92%) gave aminoalcohol 8. Etheri®cation with sesamol (80%) and hydrogenolysis of the
benzyl group (93%) completed the synthesis of paroxetine. Analysis of the ®nal product by circular
dichroism, however, revealed that undesired (+)-paroxetine was obtained. This work, however,
demonstrated that 3 was a useful intermediate to paroxetine and inversion of the stereochemistry
by reduction of the ester would provide the proper absolute stereochemistry.
Scheme 2.

5649
The reduction of the ester functionality of 3 was not, however, as straightforward as reduction
of the acid for several reasons. First, in the case of glutaric acid derivatives spontaneous
lactonization of the resultant d-hydroxy acid occurs. Jones, in fact, reports that they were unable
to ®nd conditions that a€orded the acid alcohol exclusively.5c Although lactone 10 could be
converted to the desired lactam, the yields were unacceptably low and the transformation added
to the length of the synthesis. We also found that 10 su€ered from instability making it unsuitable
for large-scale work. A third problem was that the chemoselectivity of the reduction was
exceptionally sensitive to the quality of the LiBH4. Undesired reduction of the acid was often seen
when using previously opened bottles of LiBH4, presumably due to the presence of borane
from decomposition of the reagent. These problems were circumvented, however, by a novel
deprotonation±reduction±alkylation sequence.
As shown in Scheme 3, the conversion of acid ester 3 to d-hydroxy ester 12 begins with
deprotonation using LiH in re¯uxing THF in order to protect the carboxylic acid as its carboxylate.
Reaction with LiBH4 then selectively reduced the ester with no reaction at the lithium carboxylate,
even with samples of LiBH4 which previously gave mixtures, to provide a boronate intermediate
of general structure 11. DMF was added to aid solubility followed by 1.6 equiv. of dimethyl
sulfate to alkylate the carboxylate. After 4 h at room temperature, the reaction was quenched
with MeOH resulting in an 86% yield of 12 from acid ester 3.7 Less than 1% of lactone could be
1
detected in the product by HPLC and H NMR. Deprotonation of 3 with LiH, therefore, not
only protects the acid from undesired reduction, but also allows for methylation of the
carboxylate avoiding lactone formation. The overall result is net inversion of the C-3 stereocenter
and provides a true complement to the borane reduction of the acid. This was converted to
desired (^)-paroxetine, which was identical to that of an authentic sample by all spectral and
physical characterization methods, by the route described previously. Subsequent experimentation
found that far less expensive NaBH4 could be used in place of LiBH4. After deprotonation with
LiH, addition of NaBH4 resulted in reduction of the ester as e€ectively, and in the same rate, as
using LiBH4. Since NaBH4 does not reduce esters under normal conditions,8 it is assumed that a
cation exchange between the borohydride and the carboxylate takes place, forming LiBH4 in situ.9
Scheme 3.
In summary, we have described a new asymmetric synthesis of (^)-paroxetine through the
enzymatic desymmetrization of a meso-diester. In addition, a novel method for the transformation

5650
of glutaric acid monoesters to d-hydroxy esters with opposite stereocon®guration to that obtained
from borane reduction has been developed. This one-pot reaction sequence should be applicable
to other monoester dicarboxylates where unwanted lactonization needs to be minimized or for
the formation of esters which would otherwise be inappropriate substrates for PLE. With this
new method, the desymmetrization may be optimized for enzyme and substrate without regard to
stereochemical outcome, since either d-hydroxy ester can be realized without the involvement of
other chemical intermediates. Future activities in these areas are planned and will be disclosed in
due course.
Acknowledgements
The authors would like to thank R. Lee Webb for conducting the CD experiments, which
determined that (+)-paroxetine had been synthesized.
References
1. For other routes to paroxetine, see: (a) Gledhill, L.; Martin, C. PCT Int. Appl., WO 9802556 A2 980122. (b)
Kreidl, J. et al. PCT INT. Appl., WO9801424 A1 980115. (c) Sugi, K. Eur. Patent Appl., EP 182827 A1 971217.
(d) Keverline-Frantz, K. I.; Boja, J. W.; Kuhar, M. J.; Abraham, P.; Burgess, J. P.; Lewin, A. H.; Carroll, F. I.
J. Med. Chem. 1988, 41, 247. (e) Wang, S.; Okazee, T.; Matsumura, Y. Eur. Pat. Appl., EP 819224 A1 971203. (f)
Adger, B. M.; Potter, G. A.; Fox, M. E. PCT Int. Appl., WO 9724323 A1 970710. (g) Engelstoft, M.; Hansen, J. B.
Acta Chem. Scand. 1996, 50, 164. (h) Zepp, C. M.; Gao, Y.; Heefner, D. L. US 25258517 A 931102. (i) Lundbeck,
J. M.; Everland, P.; Treppendahl, S.; Jakobsen, P. Eur. Pat. Appl., EP 374674 A2 900627. (g) Amat, M.; Hidalgo,
J.; Bosch, J. Tetrahedron: Asymmetry 1996, 7, 1591.
2. (a) Toone, E. J.; Werth, M. J.; Jones, J. B. J. Am. Chem. Soc. 1990, 112, 4969. (b) Lam, L. K.-P.; Jones, J. B. Can.
J. Chem. 1988, 66, 1422. (c) Francis, C. J.; Jones, J. B. J.C.S. Chem. Comm. 1984, 579. (d) Ohno, M.; Otsuka, M.
Org. React. 1989, 37, 1-55.
3. Perregaard, J.; Moltzen, E. K.; Meier, E.; Sanchez, C. J. Med. Chem. 1995, 38, 1998.
4. Enantiomeric purity of 3 determined by HPLC. Column: Chiracel OB-H, 25 cmÂ4.6 mm. Mobile phase: 95:5
hexane:ethanol with 0.1% TFA. Detection: UV absorbance at 210 nm. Flow rate: 1 mL/min.
5. (a) Huang, F.-C.; Lee, L. F. H.; Mittal, R. S. D.; Ravikumar, P. R.; Chan, J. A.; Sih, C. J.; Caspi, E.; Eck, C. R.
J. Am. Chem. Soc. 1975, 97, 4144. (b) Chen, C.-S.; Fujimoto, Y.; Girdaukas, G.; Sih, C. J. J. Am. Chem. Soc.
1982, 104, 7294. (c) Francis, C. J.; Jones, J. B. J. Chem. Soc., Chem. Commun. 1984, 579.
6. Enantiomeric purity of 6 determined by HPLC. Column: Regis (S,S) Whelk O-1, 25 cmÂ4.6 mm. Mobile phase:
95:5 hexane:ethanol. Detection: UV Absorbance at 210 nm. Flow rate: 1 mL/min.
7. Procedure for synthesis of 12: Acid ester 3 (5.0 g, 20.8 mmol) was dissolved in 50 mL THF in a 250 mL round-
bottomed ¯ask equipped with spinbar and condenser. LiH (200 mg, 25 mmol, 1.2 equiv.) was added and the
mixture heated at re¯ux and stirred for 1 h. LiBH4 (300 mg, 13.5 mmol, 0.65 equiv.) was added and stirring at
re¯ux continued for 10 h. The mixture was cooled to ambient temperature and 25 mL DMF added followed by
dimethylsulfate (3.30 mL, 34.5 mmol, 1.6 equiv.). After 4 h at ambient temperature the mixture had turned into a
clear pale yellow and reaction was complete as monitored by HPLC. MeOH (5 mL) and SiO2 (5 g) were added.
After 45 min the mixture was ®ltered and ®ltrate taken up in 100 mL t-butyl methyl ether and washed twice with
10% aq. NH4Cl. The organic layer was rotary evaporated and dried in vacuo to give 4.65 g of 12 as a pale yellow
oil, which was used without further puri®cation. Crude yield=99%. HPLC assay of this material showed it to be
1
86% by weight. H NMR (300 MHz, CDCl3, TMS/ppm) ꢀ 7.15 (dd, 2H, J=2.8, 3.3 Hz), 7.02 (apparent t, 2H,
J=2.8, 6.3 Hz), 3.56 (s, 3H), 3.50 (m, 2H), 3.31 (m, 1H), 2.61 (m, 2H), 1.86 (m, 2H), 1.62 (br s, 1H).
8. Examples of NaBH4 reduction of esters include: (a)Yang, C.; Pittmanm Jr., C. U. Synth. Commun. 1998, 28, 2027.
(b) Prasad, A. S. Bhanu; Kanth, J. V. Bhaskar; Periasamy, M. Tetrahedron 1992, 48, 4623. (c) Yamakawa, T.;

5651
Masaki, M.; Nohira, H. Bull. Chem. Soc. Jpn. 1991, 64, 2730. (d) Bianco, A.; Passacantilli, P.; Righi, G. Synth.
Commun. 1988, 18, 1765. (e) Santaniello, E.; Ferraboschi, P.; Fiecchi, A.; Grisenti, P.; Manzocchi, A. J. Org.
Chem. 1987, 52, 671. (f) Soai, K.; Oyamada, H.; Takase, M.; Ookawa, A. Bull. Chem. Soc Jpn. 1984, 57, 1948.
9. (a) Brown, H. C.; Narasimhan, S.; Choi, Y. M. J. Org. Chem. 1982, 47, 4702. (b) Sodium borohydride has been
converted to lithium borohydride through cation exchange with lihium halides. Brown, H. C.; Narasimhan, S.;
Choi, Y. M. Inorg. Chem. 1981, 20, 4454.