A Short Enantioselective Synthesis of (S)-Levetiracetam through Direct Palladium-Catalyzed Asymmetric N -Allylation of Methyl 4-Aminobutyrate

Article abstract of DOI:10.1055/a-1493-9078

An exceedingly short and enantioselective synthesis of the antiepileptic drug (S)-levetiracetam was elaborated. As the chirogenic key step, a Pd-catalyzed asymmetric N -allylation of methyl 4-aminobutyrate was achieved in the presence of only 1 mol% of a catalyst prepared in situ from [Pd(allyl)Cl] 2 and a tartaric acid-derived C 2 -symmetric diphosphine ligand.

Full text of DOI:10.1055/a-1493-9078

SYNLETT
0
9
3
6
-
5
2
1
4
1
4
3
7
-
2
0
9
6
Georg Thieme Verlag KG Rüdigerstraße 14, 70469 Stuttgart
2021, 32, 1089–1092
letter
1089
en
D. Albat et al.
Letter
Synlett
A Short Enantioselective Synthesis of (S)-Levetiracetam through
Direct Palladium-Catalyzed Asymmetric N-Allylation of Methyl
4-Aminobutyrate
Dominik Albat
Cl
Jörg-Martin Neudörfl
Hans-Günther Schmalz*
L*
H3N
CO2Me
OCO2Me
O
O
PdL*
0
0
0
0
-
0
0
0
3-0
4
8
9
-1
8
2
7
O
O
O
N
N
base
O
O
University of Cologne, Department of Chemistry,
Greinstraße 4, 50939 Köln, Germany
schmalz@uni-koeln.de
94% ee
NH2
(S)-levetiracetam
PPh2 Ph2P
HUeMnaniavsiCel-roGssrcüirhtneymstohpafeolzrCn@oSdlcuionhnggmin-Akeau,lozDtehlenop.rdaertment of Chemistry, Greinstraße 4, 50939 Köln, Germany
Received: 07.04.2021
Accepted after revision: 28.04.2021
Published online: 28.04.2021
In the course of our research into asymmetric Pd-cata-
lyzed N-allylation of amino acid derivatives,7 we envisioned
challenging (benchmarking) the developed methodology
for the synthesis of 1 according to the retrosynthetic analy-
sis shown in Scheme 1. Our main goal was to achieve the
direct N-allylation of methyl 4-aminobutanoate (4), as a
synthetic equivalent of pyrrolidone, with high enantiose-
lectivity by employing the racemic carbonate rac-3. Effi-
cient conversion of the cyclized product 2 into (S)-levetirac-
etam (1) should then be achieved through oxidative cleav-
DOI: 10.1055/a-1493-9078; Art ID: st-2021-b0130-l
Abstract An exceedingly short and enantioselective synthesis of the
antiepileptic drug (S)-levetiracetam was elaborated. As the chirogenic
key step, a Pd-catalyzed asymmetric N-allylation of methyl 4-aminobu-
tyrate was achieved in the presence of only 1 mol% of a catalyst pre-
pared in situ from [Pd(allyl)Cl]2 and a tartaric acid-derived C2-symmetric
diphosphine ligand.
Key words asymmetric catalysis, palladium catalysis, chiral ligands,
amino acid derivatives, ozonolysis, levetiracetam
age of the double bond and amide formation.
A
conceptually related approach had recently been described
by Stecko and co-workers, who initially employed an enan-
tioselective Pd-catalyzed hydrolysis of rac-3 to prepare the
(R)-allylic alcohol; this was further transformed into the
corresponding (S)-amine through a stereospecific [3,3]-re-
arrangement before the pyrrolidone unit was established
Epilepsy is the most common neurological disorder, af-
fecting about 50 million people all over the world.1 Treat-
ment consists primarily of the use of antiepileptic agents
such as levetiracetam (1), which shows no significant drug–
drug interactions and has favorable pharmacokinetic prop-
erties, making it one of the most commonly prescribed an-
ticonvulsants.2
Pd-catalyzed
asymmetric
allylic substitution
O
Despite its structural simplicity, the synthesis of 1 in
nonracemic form is not trivial, as reflected by the various
protocols that have been developed. Classical approaches
are based on the late construction of the pyrrolidone unit
from (S)-2-aminobutanamide (or the corresponding esters)
obtained by desulfurization of methionine, resolution, or
biocatalysis.3 Notably, the direct nucleophilic introduction
of the pyrrolidone ring in an SN2 fashion has been achieved
only with less readily available substrates.4 Asymmetric ap-
proaches toward 1 involve the use of Evans’s auxiliary in the
diastereoselective ethylation of a pyrrolidone-1-yl-acetic
acid derivative5 and the highly enantioselective hydrogena-
tion of the dehydroamino acid (enamide) derivative corre-
sponding to 1.6
O
O
N
N
O
OMe
CO2Me
NH2
+
NH2
O
rac-3
4
2
(S)-levetiracetam (1)
Previous work (ref. 8):
COCl
OH
NH2
H2O
rac-3
Cl
2
1
Pd*
S: 10 steps; 37% yield
This work:
CO2Me
NH2
O
N
rac-3
1
Pd*
S: 4 steps; 46% yield
2
Scheme 1 Retrosynthetic analysis of levetiracetam (1) based on an
asymmetric Pd-catalyzed allylic substitution
© 2021. Thieme. All rights reserved. Synlett 2021, 32, 1089–1092
Georg Thieme Verlag KG, Rüdigerstraße 14, 70469 Stuttgart, Germany

1090
D. Albat et al.
Letter
Synlett
by reaction of the amine with 4-chlorobutanoyl chloride.8
Although the Stecko synthesis provides the target 1 with a
respectable overall yield, its low step economy9 reflects the
difficulties associated with the direct introduction of suit-
ably functionalized amines in asymmetric allylic substitu-
tion reactions.10,11
Addressing this particular challenge, we recently suc-
ceeded in developing a powerful protocol for the Pd-cata-
lyzed asymmetric N-allylation of -amino acid esters by
employing a new class of tartaric acid-derived C2-symmet-
ric chiral diphosphine ligands, such as (S,S)-iPr-MediPhos
(L*),12 which not only gave rise to high enantioselectivities,
but also formed particularly active catalysts (Scheme 2).7
Here, we describe an adaptation of this methodology in an
exceedingly short and efficient synthesis of levetiracetam
(1), according to the strategy outlined in Scheme 1.
commercially available stable crystalline hydrochloride 7
and triethylamine) in the presence of dppe as an achiral li-
gand resulted in the clean formation of a mixture of the al-
lylic amine rac-8 and the corresponding cyclized pyrroli-
done derivative rac-2 (Table 1; entry 1). On using the di-
phosphine ligand (R,R)-iPr-MediPhos (ent-L*) in the N-
allylation of 4, full conversion of rac-3 was observed after
just 2.5 hours, giving a 9:1 mixture of ent-8 and ent-2. Upon
prolongation of the reaction time, the amount of the pyrro-
lidone product ent-2 steadily increased, indicating that the
cyclization of ent-8 occurs as a secondary process under the
reaction conditions. After 58 hours at room temperature,
the content of the desired product ent-2 was 93%. Determi-
nation of the enantiomeric excess proved to be difficult, as
the enantiomers of neither the amine rac-8 nor the lactam
rac-2 could be separated on the available GC columns. How-
ever, after conversion of lactam 2 into ester 9 (see below),
the enantiomeric excess could be determined by means of
chiral GC. In this way, the resulting sample of ent-2 (Table 1,
entry 4) was found to have an enantiomeric purity of 93%
ee. The expected (R)-configuration was confirmed by cor-
relation of the molecular rotation of ent-9 ([]D = 41° in CH-
Cl3) with the reported value.3b
H2N
CO2Me
O
O
5
O
[Pd(allyl)Cl]2/L*
Et3N, THF
HN
CO2Me
O
O
O
O
r.t., 17 h
89%
rac-3
6
PPh2 Ph2P
98% ee
L*
In contrast to our original reaction system (Scheme 2),7
the N-allylation of 7 showed only a small increase in
enantioselectivity upon lowering the concentration (Table
1, entries 4–7). Although the reaction proceeded with ≤94%
ee at a concentration of 0.50 M, the content of the desired
pyrrolidone derivative ent-2 in the mixture was only about
45% after 17 hours.
Scheme 2 Highly enantioselective Pd-catalyzed N-allylation of methyl
glycinate (5) by using the chiral diphosphine (S,S)-iPr-MediPhos (L*)
Although our initial attempts to employ pyrrolidin-2-
one directly as a nucleophile in a Pd-catalyzed reaction
with rac-3 failed due to a lack of conversion, the use of
methyl 4-aminobutyrate (4) (generated in situ from the
Table 1 Optimization of the Asymmetric N-Allylation of Methyl 4-Aminobutyrate (4)a
CO2Me
Cl
1.0 mol% [Pd(allyl)Cl]2
2.4 mol% ent-L*
1.3 equiv Et3N
H3N
7
CO2Me
+
O
O
HN
N
THF, r.t.
O
O
ent-8
ent-2
rac-3
Entry
Ligand
Conc.b (mol/L)
Time (h)
Conv.c (%)
Product ratioc (%)
eed (%)
ent-8
ent-2
1
2
3
4
5
6
7
dppe
2.00
2.00
2.00
2.00
1.00
0.50
0.25
13
2.5
17
58
17
17
19
100
100
100
100
100
100
≤5
54
89
59
7
46
11
41
93
48
45
–
(rac)
n.d.e
n.d.
93
ent-L*
ent-L*
ent-L*
ent-L*
ent-L*
ent-L*
52
55
–
93
94
–
a Reactions were performed on a 0.5 mmol scale using 1.3 equivalents of 7.
b Concentration of rac-3.
c Product ratio as determined by GC-MS.
d Determined at the stage of 9 (after ozonolysis of ent-2) by GC on a chiral stationary phase.
e n.d. = not determined.
© 2021. Thieme. All rights reserved. Synlett 2021, 32, 1089–1092

1091
D. Albat et al.
Letter
Synlett
To synthesize the (S)-form of the key intermediate 2,
the N-allylation of 7 was performed under the optimized
conditions by employing just 1 mol% of the Pd catalyst
prepared from the (S,S)-biphosphine ligand L*. After re-
moval of the metal catalyst by filtering the solution
through a pad of Celite and evaporating all the volatiles,
the crude 1:1 mixture of 8 and 2 was treated with Et3N in
ethyl acetate for 16 hours at 80 °C to ensure complete cy-
clization of 8 to 2. In this way, the desired -lactam 2 was
obtained in a total yield of 72% over the two steps (Scheme
3) in an operationally straightforward manner.
Because the synthesized sample of (S)-levetiracetam (1)
was obtained in pure crystalline form, we were able to con-
firm its absolute configuration additionally by means of X-
ray crystallography (Figure 1),15 thereby confirming all pre-
vious configurational assignments.
In conclusion, we have elaborated a short and efficient
synthesis of the antiepileptic drug (S)-levetiracetam (1) by
adapting our recently developed protocol for the enantiose-
lective Pd-catalyzed N-allylation of amino acid esters as the
key step.16 We also demonstrated that methyl 4-aminobu-
tyrate (4), liberated in situ from the corresponding hydro-
chloride 7, is a useful synthetic equivalent for pyrrolidone
in such reactions. In addition, we identified an improved
two-step protocol for the oxidative conversion of the allyl-
amine derivative 2 into the amide 1. The resulting overall
synthesis of 1 proceeds in only four steps with 46% overall
yield.
CO2Me
HN
0.5 mol% [Pd(allyl)Cl]2
1.2 mol% L*
7
Et3N, EtOAc
80 °C, 16 h
1.3 equiv Et3N
THF, r.t.
8
+
O
N
72%
(two steps)
rac-3
2
O
N
93% ee
Conflict of Interest
2
The authors declare no conflict of interest.
Scheme 3 Procedure for the synthesis of desired pyrrolidine derivative 2
Funding Information
To complete the synthesis of levetiracetam (1), we used
the ozonolysis conditions of Marshall et al.13 to induce oxi-
dative degradation of the allylic amine 2 to form the methyl
ester 9 directly (Scheme 4). The final conversion of 9 into
the corresponding amide 1 was then efficiently achieved by
treatment with aqueous ammonia (Scheme 4).14
This work was supported by the University of Cologne and the Fonds
der Chemischen Industrie.
U
n
i
vers
i
ät
t
z
u
K
ö
l
n(
)
Supporting Information
Supporting information for this article is available online at
O3
https://doi.org/10.1055/a-1493-9078.
S
u
p
p
orti
n
gI
n
f
orm
a
ti
o
n
S
u
p
p
orti
n
gI
n
f
orm
a
ti
o
n
2.5 M NaOH
in MeOH
O
O
60 vol% NH3
O
N
N
N
OMe
NH2
H2O
0 °C, 17 h
CH2Cl2
–78 °C, 70 min
References and Notes
O
O
73%
88%
2
9
1
(1) (a) Sander, J. W. A. S.; Shorvon, S. D. J. Neurol., Neurosurg. Psychi-
atry 1996, 61, 433. (b) Kwan, P.; Brodie, M. J. Expert Rev. Neu-
rother. 2006, 6, 397.
Scheme 4 Completion of the synthesis of (S)-levetiracetam (1)
(2) (a) Ben-Menachem, E. Expert Opin. Pharmacother. 2003, 4, 2079.
(b) Leeman, B. A.; Cole, A. J. Annu. Rev. Med. 2008, 59, 503.
(c) Ulloa, C. M.; Towfigh, A.; Safdieh, J. Neuropsychiatr. Dis. Treat.
2009, 5, 467.
(3) (a) Cossement, E.; Motte, G.; Geerts, J.-P.; Gobert, J. GB 2225322,
1992. (b) Das Sarma, K.; Zhang, J.; Huang, Y.; Davidson, J. G. Eur.
J. Org. Chem. 2006, 3730. (c) Tucker, J. L.; Xu, L.; Yu, W.; Scott, R.
W.; Zhao, L.; Ran, N. WO 2009009117, 2009. (d) Tang, X.-L.; Lu,
X.-F.; Wu, Z.-M.; Zheng, R.-C.; Zheng, Y. G. J. Biotechnol. 2018,
266, 20. (e) Mylavarapu, R.; Anand, R. V.; Kondaiah, G. C. M.;
Reddy, L. A.; Reddy, G. S.; Roy, A.; Bhattacharya, A.; Mukkanti,
K.; Bandichhor, R. Green Chem. Lett. Rev. 2010, 3, 225.
(4) (a) Boschi, F.; Camps, P.; Comes-Franchini, M.; Muñoz-Torrero,
D.; Ricci, A.; Sánchez, L. Tetrahedron: Asymmetry 2005, 16, 3739.
(b) Kotkar, S. P.; Sudalai, A. Tetrahedron Lett. 2006, 47, 6813.
(5) (a) Chandra Babu, K.; Buchi Reddy, R.; Mukkanti, K.; Suresh, K.;
Madhusudhan, G.; Nigam, S. J. Chem. 2013, 176512 DOI:
10.1155/2013/176512. (b) Raju, V.; Somaiah, S.; Sashikanth, S.;
Laxminarayana, E.; Mukkanti, K. Drug Invent. Today 2014, 6, 32.
Figure 1 Structure of synthetic (S)-levetiracetam in the crystalline state
© 2021. Thieme. All rights reserved. Synlett 2021, 32, 1089–1092

1092
D. Albat et al.
Letter
Synlett
(6) (a) Surtees, J.; Marmon, V.; Differding, E.; Zimmermann, V. WO
2001064637, 2001. (b) Friedfeld, M. R.; Zhong, H.; Ruck, R. T.;
Shevlin, M.; Chirik, P. J. Science 2018, 360, 888. (c) Zhong, H.;
Friedfeld, M. R.; Camacho-Bunquin, J.; Sohn, H.; Yang, C.;
Delferro, M.; Chirik, P. J. Organometallics 2019, 38, 149.
(7) (a) Dohmen, S.; Reiher, M.; Albat, D.; Akyol, S.; Barone, M.;
Neudörfl, J. M.; Kühne, R.; Schmalz, H.-G. Chem. Eur. J. 2020, 26,
3049. (b) Albat, D.; Neudörfl, J.-M.; Schmalz, H.-G. Eur. J. Org.
Chem. 2021, 2099.
(8) Narczyk, A.; Mrozowicz, M.; Stecko, S. Org. Biomol. Chem. 2019,
17, 2770.
(9) (a) Wender, P. A.; Verma, V. A.; Paxton, T. J.; Pillow, T. H. Acc.
Chem. Res. 2008, 41, 40. (b) Wender, P. A.; Miller, B. L. Nature
2009, 460, 197.
then added and the solution was stirred for 20 min at r.t. before
carbonate rac-3 (103 mg, 0.60 mmol) was added neat from a
syringe. After 20 min, methyl 4-aminobutyrate hydrochloride
(7; 116 mg, 0.76 mmol) and Et3N (0.11 mL, 0.79 mmol) were
added, and stirring was continued for 16 h. QuadraSil AP (~50
mg) was then added to capture Pd, and the mixture was stirred
for 1 h then filtered through a short pad of Celite with EtOAc.
After removal of the solvent under reduced pressure, the crude
product mixture of 2 and 8 was dissolved in EtOAc (1 mL) and
Et3N (0.11 mL, 0.79 mmol), then heated at 80 °C for 20 h. The
solvent was removed under reduced pressure and the crude
product was purified by column chromatography [silica gel,
cyclohexane–EtOAc (1:1)] to give a yellowish oil; yield: 78 mg
20
(0.43 mmol; 72%); [] (c = 0.24, CHCl3): []365 –364.6, []436
(10) For selected reviews, see: (a) Trost, B. M.; Van Vranken, D. L.
Chem. Rev. 1996, 96, 395. (b) Trost, B. M.; Crawley, M. L. Chem.
Rev. 2003, 103, 2921. (c) Graening, T.; Schmalz, H.-G. Angew.
Chem. Int. Ed. 2003, 42, 2580. (d) Grange, R. L.; Clizbe, E. A.;
Evans, A. A. Synthesis 2016, 48, 2911; and references cited
therein.
–206.7 °, []546 –111.1, []579 –96.0, []589 –92.5°.
1H NMR (400 MHz, CDCl3):  = 5.64 (dtd, 3J = 15.5 Hz, 3J = 6.3 Hz,
4J = 1.3 Hz, 1 H), 5.34 (ddt, 3J = 15.5 Hz, 3J = 6.6 Hz, 4J = 1.6 Hz, 1
H), 4.48 (Ψq, 3J = 7.0 Hz, 1 H), 3.27 (Ψqt, 3J = 9.6 Hz, 3J = 7.0 Hz, 2
H), 2.45–2.35 (m, 2 H), 2.10–1.92 (m, 4 H), 1.68–1.46 (m, 2 H),
3
0.98 (t, J = 7.5 Hz, 3 H), 0.86 (t, 3J = 7.4 Hz, 3 H). 13C NMR (100
(11) (a) Trost, B. M.; Calkins, T. L.; Oertelt, C.; Zambrano, J. Tetrahe-
dron Lett. 1998, 39, 1713. (b) Humphries, M. E.; Clark, B. P.;
Williams, J. M. Tetrahedron: Asymmetry 1998, 9, 749. (c) Tosatti,
P.; Horn, J.; Campbell, A. J.; House, D.; Nelson, A.; Marsden, S. P.
Adv. Synth. Catal. 2010, 352, 3153.
(12) Dindaroğlu, M.; Akyol Dinçer, S.; Schmalz, H.-G. Eur. J. Org.
Chem. 2014, 4315.
(13) Marshall, J. A.; Garofalo, A. W.; Sedrani, R. C. Synlett 1992, 643.
(14) Ates, C.; Surtess, J.; Burteau, A.-C.; Marmon, V.; Cavoy, E. US
2004/0204476, 2004.
(15) CCDC 2074847 contains the supplementary crystallographic
data for (S)-levetiracetam (1). The data can be obtained free of
charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/structures. For earlier X-ray crystal struc-
tures of 1 determined at room temperature, see: (a) Song, J.;
Lou, K.-X.; Li, X.-J.; Wua, X.-P.; Feng, R.-X. Acta Crystallogr., Sect.
E: Struct. Rep. Online 2003, 59, o1772. (b) Bebiano, S. S.; ter
Horst, J. H.; Oswald, I. D. H. Cryst. Growth Des. 2020, 20, 6731;
(Structure determined at high pressure to investigate pressure-
dependent polymorphism).
MHz, CDCl3):  = 174.7, 135.0, 126.6, 54.1, 42.5, 31.6, 25.5, 24.9,
18.3, 13.7, 10.8. HRMS (ESI): m/z [M + H]+ calcd for C11H20NO:
182.1539; found: 182.1540.
Methyl 2-(2-Oxopyrrolidin-1-yl)butanoate (9)
A round-bottomed flask was charged with a solution of 2 (105
mg, 0.58 mmol) in CH2Cl2 (40 mL) and a 2.50 M solution of
NaOH in MeOH (3.0 mL, 7.5 mmol). The clear solution was
cooled to –78 °C and a weak stream of ozone was introduced
until the solution turned blue (70 min). Excess ozone was
removed by introducing a weak stream of oxygen for 10 min
before the solution was allowed to warm to r.t. H2O (50 mL) was
added, and the aqueous phase was extracted with CH2Cl2 (3 ×
100 mL). The combined organic phases were dried (MgSO4), fil-
tered, and concentrated under reduced pressure. The crude
product was purified by column chromatography [silica gel,
cyclohexane–EtOAc (1:1)] to give a yellow oil; yield: 78 mg
20
(0.42 mmol; 73%); [] (c = 0.26, CHCl3): []365 –152.9, []436
–87.1, []546 –46.1, []579 –40.0, []589 –38.6.
1H NMR (500 MHz, CDCl3):  = 4.69 (Ψdd, 3J = 10.7 Hz, 4J = 5.2
Hz, 1 H), 3.71 (s, 3 H), 3.52 (td, 3J = 8.7, 4J = 6.1 Hz, 1 H), 3.35 (td,
4
3
(16) Detailed experimental procedures and characterization data are
given in the Supporting Information.
3J = 8.8 Hz, J = 5.5 Hz, 1 H), 2.44 (t, J = 8.1 Hz, 2 H), 2.15–1.96
(m, 3 H), 1.69 (ddq, 3J = 14.6 Hz, 3J = 10.7 Hz, 3J = 7.3 Hz, 1 H),
0.92 (t, 3J = 7.4 Hz, 3 H). 13C NMR (125 MHz, CDCl3):  = 176.0,
171.7, 55.2, 52.2, 43.6, 31.0, 22.3, 18.4, 10.9. HRMS (ESI): m/z =
[M + H]+ calcd for C9H16NO3: 186.1124; found: 186.1127.
1-[(2E)-1-Ethylpent-2-en-1-yl]pyrrolidin-2-one (2)
Under an atmosphere of argon, a Schlenk flask was charged
with [Pd(allyl)Cl]2 (1.10 mg, 3.01 mol, 0.50 mol%) and ligand
L* (5.40 mg, 7.04 mol, 1.17 mol%). Anhyd THF (0.6 mL) was
© 2021. Thieme. All rights reserved. Synlett 2021, 32, 1089–1092