Practical syntheses of oxindole derivatives: Chemical development towards 2-(5-chloro-2-oxo-2,3-dihydroindol-1-yl)acetamide and (S)-2-(5-Chloro-2-oxo-2,3- dihydroindol-1-yl)propionamide

Article abstract of DOI:10.1021/op800203p

We describe development of scalable syntheses of novel oxindole-type SV2A ligands with improved potency towards seizure suppression.

Full text of DOI:10.1021/op800203p

Organic Process Research & Development 2009, 13, 442–449
Practical Syntheses of Oxindole Derivatives: Chemical Development towards
2-(5-Chloro-2-oxo-2,3-dihydroindol-1-yl)acetamide
and (S)-2-(5-Chloro-2-oxo-2,3-dihydroindol-1-yl)propionamide
Fabienne Broeders, Laurent Defre`re,* Marie-France Deltent, Frank Driessens, Fre´de´ric Gilson, Luc Grooters,
Xavier Ikonomakos, Fre´de´ric Limauge, Emmanuel Sergeef, and Natacha Verstraeten
Chemical Process Research and Early Scale-Up, UCB Pharma, Chemin du Foriest, B-1420 Braine-l’Alleud, Belgium
Abstract:
We describe development of scalable syntheses of novel
oxindole-type SV2A ligands with improved potency towards
seizure suppression.
Introduction
Figure 1. Progression of the acetam family within UCB.
Epilepsy is a chronic, often lifelong, neurological disorder
that affects ∼1% of the population worldwide and which is
characterized by recurrent involuntary seizures, caused by
disturbances in the normal electrical activity of the brain.
Epilepsy has a tremendous impact, not only on the individuals
afflicted with this disease but also on their families as well as
on society in general, because of the resultant personal and
financial burden.
Despite the recent introduction of second-generation anti-
convulsants, refractory epilepsy is still a significant clinical
problem. In a prospective study, Kwan and Brodie1 show that
fewer than 50% of patients with epilepsy become seizure-free
Figure 2. Discovery of the two oxindole-type lead molecules 3
after beginning an initial antiepileptic medication. In addition,
more than 30% of patients cannot be adequately treated with
existing antiepileptic drugs (AED), because of either lack of
efficacy or toxicity.
and (S)-4.
acetamide scaffold 2,6 whose structure is similar to that of the
acetam family, was undertaken in order to find a new ligand
with an improved affinity towards the SV2A protein.
This led to the discovery of two ligands 3 and (S)-4 which
were chosen as candidates for development (Figure 2).
In a range of secondary epilepsy models, these two com-
pounds displayed a more complete seizure suppression than
levetiracetam, 1b, itself along with an outstanding safety margin
(no marked effects on cardiovascular, respiratory, or CNS
functions).
At this stage of the project, we were asked to provide
kilograms of both compounds 3 and (S)-4 in order to progress
into the preclinical toxicological studies. Ultimately this research
work would serve as a framework for future process optimiza-
tion to allow for larger-scale manufacturing.
The synaptic vesicle protein SV2A has been identified2 as
the brain-specific binding site for levetiracetam, 1b (see Figure
1, Keppra, UCB Pharma), a structural analogue of piracetam,
1a (Nootropil, UCB Pharma). Interaction with SV2A is likely
to exert a major role in its antiepileptic properties. Using this
novel molecular target, medicinal chemists initiated a thorough
drug discovery program3 targeting the identification of ligands
of the so-called acetam family with an improved SV2A affinity.
Brivaracetam, 1c, was selected as the lead compound whose
clinical program (phase I and II studies) has been finalized.4
In parallel to these studies, another drug discovery program5
around the relatively unexplored 3-unsubstituted oxindole-N-
* To whom correspondence should be addressed. Telephone: +32 2 386 3315.
Fax: +32 2 386 2704. E-mail: laurent.defrere@ucb-group.com.
(1) Kwan, P.; Brodie, M. J. N. Engl. J. Med. 2000, 342, 314.
(2) Lynch, B. A.; Lambeng, N.; Nocka, K.; Kensel-Hammes, P.; Bajjalieh,
S. M.; Matagne, A.; Fuks, B. Proc. Natl. Acad. Sci. U.S.A. 2004, 101,
9861.
Discovery Synthesis
Prior to beginning our chemical development work the total
quantity of 3 and (S)-4 was around 5 g. While the synthetic
routes were capable of supplying these small amounts (see
(3) Kenda, B. M.; Matagne, A. C.; Talaga, P. E.; Pasau, P. M.; Differding,
E.; Lallemand, B. I.; Frycia, A. M.; Moureau, F. G.; Klitgaard, H. V.;
Gillard, M. R.; Fuks, B.; Michel, P. J. Med. Chem. 2004, 47, 530.
(4) Von Rosenstiel, P. Neurotherapeutics 2007, 4, 84.
(5) Starck, J.-P.; Kenda, B. Indolone-Acetamide Derivatives, Processes
for Preparing Them and Their Uses. (UCB Pharma). Patent number
WO/2004/087658 A1, 2004.
(6) Morozov, I. S.; Novikova, N. N.; Silenko, I. D.; Nerobkova, L. N.;
Zagorevskij, V. A. Amide-2-oxo-1-indolineacetic Acid Possessing
Anticonvulsive Activity. (Nii Farmakologii AMN SSSR). Patent
number SU841264, 1995.
442
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development
10.1021/op800203p CCC: $40.75  2009 American Chemical Society
Published on Web 01/30/2009

Scheme 1. Medicinal chemistry route to the acetamide 3
Figure 3. Carom-CH2 disconnection.
philic reagents (see ref 5 for more examples) at a late stage
without concerning oneself with the structure of the starting
isatin.
For the second lead oxindole (S)-4, whose initial synthesis
was depicted in Scheme 2, the main problems encountered were
the low overall yield and the reliance upon a chromatographic
separation of enantiomers at the final stage. The conversion8
of indole rac-10 to oxindole rac-4 was particularly troublesome,
due to incomplete reaction, a range of impurities, and a tedious
workup.
Scheme 2. Medicinal chemistry route to the propionamide
(S)-4
In view of these shortcomings we found it desirable to seek
alternative high-yielding and more staightforward routes to the
target molecules 3 and (S)-4. Very clearly the challenge was
to identify a safe, robust, and rapid route to the title compounds
in order to make them available for the preclinical toxicology
studies, rather than identify the commercial synthesis.
We first initiated a literature survey that indicated that many
disconnections could possibly lead to the desired 3-unsubstituted
oxindoles. Among all the possibilities, we only considered the
routes based on readily available aromatic starting materials
already possessing the chlorine atom in the position para to the
nitrogen atom, thus avoiding the low-yielding NCS-mediated
chlorination step.
Schemes 1 and 2), they both suffer from several disadvantages
and thus were of limited use for scale-up without extensive
modification.
In the case of Scheme 1, the strategy was based on the
N-alkylation of the widely available and inexpensive isatin 5
with chloroacetamide, since the same transformation applied
to a 3-unsubstituted oxindole as starting product invariably leads
to an intractable mixture (N- vs O- vs C-alkylation among other
things). Reduction of the ketone function of 6 to the corre-
sponding methylene was achieved in a two-step sequence
involving first the reduction to the alcohol 7 and then to the
oxindole 2 Via the acetate 8. It was already shown that another
possibility7 consisted of preparing a dithioacetal from the isatin
and then reducing it to the oxindole with Raney nickel, but this
alternative was disfavoured as it involved the need to deal with
noxious thiol side products.
It was found that alcohol 7 is somewhat unstable towards
molecular oxygen present in air and can easily reoxidize back
to the isatin 6 if not trapped quickly as here by acetylation to
8. We noticed later that this particularly cumbersome behaviour
was exacerbated when electron-withdrawing groups were al-
ready present on the aromatic ring. This side reaction was the
most serious drawback of the preparation presented here,
together with the low yield (∼25%) of the final aromatic
electrophilic substitution with N-chlorosuccinimide (NCS).
Nevertheless, this strategy has allowed medicinal chemists to
introduce diversity at position 5 (and eventually 7) by electro-
Given the time constraints to deliver the active materials, it
was decided to study the most promising approaches in parallel
before testing the viability of the synthesis upon a preliminary
scale-up.
Results and Discussion
C-C Bond Formation from Aniline Derivatives. Histori-
cally this transformation is best represented as an intramolecular
Friedel-Craft alkylation of R-haloacetanilides, but this often
requires harsh conditions9 (excess aluminium trichloride, typi-
cally above 150 °C) often incompatible with more sensitive
functionalities. Nafion-H- or rhodium (II) salts-mediated cy-
clizations of R-diazoacetanilides originally reported by Doyle10
have been used on few occasions,11 as has the elegant ring
contraction of diazoquinolinediones (Wolff rearrangement)12
induced thermally or by rhodium (II) salts (not shown in Figure
3). However, those two methods suffer from modest yields in
many cases, and the diazo precursors are always prepared by
diazo transfer from sulfonyl azides that are known to be
(8) Marfat, A.; Carta, M. P. Tetrahedron Lett. 1987, 28, 4027.
(9) Protiva, M.; Sindelar, K.; Sedivy, Z.; Metysova, J. Collect. Czech.,
Chem. Commun. 1979, 44, 2108.
(10) Doyle, M. P.; Shanklin, M. S.; Pho, H. Q.; Mahapatro, S. N. J. Org.
Chem. 1988, 53, 1017.
(11) Venkatesan, H.; Davis, M. C.; Altas, Y.; Snyder, J. P.; Liotta, D. C.
J. Org. Chem. 2001, 66, 3653, and references cited therein.
(12) Lee, Y. R.; Suk, J. Y.; Kim, B. S. Tetrahedron Lett. 1999, 40, 8219.
(7) Wenkert, E.; Bringi, N. V. J. Am. Chem. Soc. 1958, 80, 5575.
Vol. 13, No. 3, 2009 / Organic Process Research & Development
•
443

Scheme 3. Radical-mediated access to oxindoles
Figure 4. Synthesis from nitroaromatics.
Scheme 5. Nucleophilic substitution approaches
Scheme 4. Palladium-mediated access to oxindoles
unstable13 or not commercially available14 and thus not ame-
nable to safe large-scale preparations.
the disadvantage of using moisture- and/or air-sensitive
catalytic systems. However, in our hands, compounds 15
(R ) CH2CO2Me, CH2CONH2, or CH2CO2H) did not
show any conversion to the expected oxindoles.
These poor results and other unfavorable consideration led
us to set this disconnection aside to explore others.
C-C Bond Formation from Nitroaromatic Derivatives.
In this approach (see Figure 4) the key compound is the 2-(5-
chloro-2-nitrophenyl)acetic acid 24b which can be accessed by
different methods (Scheme 5). It was obvious to first study the
incorporation of the acetic acid side chain Via malonate
chemistry facilitated by the presence of an ortho nitro function
on the aromatic ring starting from the cheap 2,4-dichloroni-
trobenzene 20 (aromatic nucleophilic substitution/pathway A)20
or from even cheaper 4-chloronitrobenzene 23 (vicarious
nucleophilic substitution of hydrogen or VNS/pathway B)21 (see
Scheme 5).
Unfortunately, the reactions between the nitro derivatives
20 and 23 and the anions of a malonate ester and a bromoac-
etate, respectively, failed to be totally regioselective, with partial
displacement of the chlorine atom in the para position always
accompanying the formation of the desired product (this has
already been reported by a group20a from Pfizer). These routes
were therefore discarded.
The classical two-step sequence (see pathway C)22 involved
the condensation of the deprotonated nitrotoluene with an oxalic
acid ester23 followed by oxidative cleavage to the corresponding
phenylacetic acid by hydrogen peroxide in the presence of a
It was therefore decided to explore the possibility of carrying
out the ring closure by radical cyclization from xanthate
derivatives, which is generally a very powerful carbon-carbon
bond-forming process.15 The R-xanthyl-acetanilide precursor 16
was easily and efficiently prepared without any optimization
as depicted in Scheme 3, but despite a promising precedent in
the literature,16 cyclization onto the electron-poor aromatic ring
remained very sluggish. The conversion was incomplete after
3 days at 135 °C even though fresh lauroyl peroxide was added
portionwise every few hours (∼5 equiv in all). Moreover, the
final product 17 was always accompanied by the reduction
product 18 as already pointed out earlier for a similar radical-
mediated oxindole formation,17 and even in a separate reaction
conducted at higher dilution.
We also considered the possibility of preparing our oxindoles
by a palladium-catalyzed process (see Scheme 4), such as
cylization of an R-chloro-acetanilide 15 using Buchwald meth-
odology18 or by R-arylation19 of the appropriate 2-halo-N-
acetylaniline 19 although both of these methods suffer from
(13) Urben, P. G. Bretherick’s Handbook of ReactiVe Chemical Hazards,
5th ed.; Butterworth-Heinemann: Oxford, 1995; Vol. 1, p 902.
(14) Goddard-Borger, E. D.; Stick, R. V. Org. Lett. 2007, 9, 3797.
(15) We thank Professor Zard for the helpful discussions during his visit
here in Braine-l’Alleud. Quiclet-Sire, B.; Zard, S. Z. Chem. Eur. J.
2006, 12, 6002.
(16) Axon, J.; Boiteau, L.; Boivin, J.; Forbes, J. E.; Zard, S. Z. Tetrahedron
Lett. 1994, 35, 1719.
(17) Nishio, T.; Iseki, K.; Araki, N.; Miyazaki, T. HelV. Chim. Acta 2005,
88, 35.
444
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development

Scheme 6. Study of the intramolecular coupling
Figure 5. Disconnection Wia a C-N coupling reaction.
base (typically sodium hydroxide). This approach was not
followed up as a viable option given the expense of the starting
2-nitro-5-chlorotoluene 26.
C-N Coupling Approach. When ideas were put down on
paper, it was obvious that a carbon-nitrogen coupling reaction
between an amino acid derivative and the appropriate halo-
aromatic would be a considerable improvement of the discovery
synthesis (see Figure 5). This was particularly true for the chiral
structure (S)-4 since coupling of an L-alanine derivative would
allow a straightforward access to the expected structures and
most notably with the right absolute configuration therefore
avoiding any chiral separation, resolution or asymmetric
synthesis. It was then discovered that 2-bromo-5-chloro-
phenylacetic acid 27 was commercially available and was an
ideal halogenated substrate to study this particular C-N
coupling reaction.
The transition metal-catalyzed C-N bond formation of
aromatic compounds has made tremendous advances in the past
decade. The palladium version of this reaction has been
reviewed a couple of times these last years.24 In this field most
of the papers disclose the preparation of oxindoles by the
intramolecular coupling of the corresponding amide, but this
transformation seems to be restricted to N-alkyl-phenylaceta-
mide derivatives.25 Moreover, palladium catalysis is limited in
many cases due to the air and moisture sensitivity as well as
the higher cost of palladium and the associated ligands. For
this reason we decided to first investigate the use of copper
catalysis.26 At the outset of our work, only a few methods were
available for the clean and efficient copper-mediated N-arylation
of amino acids.27 In this area a major breakthrough was made
in 1998 by Ma’s group27b followed by numerous variants, and
his work served as a starting point for our investigations.
We first chose to study the intramolecular version of this
coupling reaction. The starting 2-bromo-5-chlorophenylaceta-
mides 28 were efficiently prepared as depicted in Scheme 6.
Activation of the starting phenylacetic acid 27 was performed
with carbonyl diimidazole (CDI) under standard conditions (1.2
equiv CDI, dichloromethane, 20 °C), and the activated ester
was then reacted with the hydrochloride salt of glycine- or
L-alanine methyl ester (1 equiv) at the same temperature. The
corresponding phenyl acetylamino-acetic and -propionic acid
compounds 28 were isolated in 96% and 93% yields, respectively.
A quick screening of conditions (copper source, solvent,
base, and temperature) revealed that significant N-arylation took
place with copper (I) iodide in dimethylsulfoxide or N,N-
dimethylformamide (>120 °C) in the presence of cesium- or
potassium carbonate with no additional ligands. However,
oxindoles 29 formed under these conditions are partly consumed
and oxidized to the corresponding isatins 30 which can further
condense with oxindoles 29 still present in the medium to form
indigo¨ıd derivatives28 as detected by mass spectroscopy. Deg-
radation of oxindoles under similar conditions has already been
reported,29 while it has been pointed out that copper salts30a or
molecular oxygen30b can catalyze oxidation of oxindoles to
isatins.
(18) Hennessy, E. J.; Buchwald, S. L. J. Am. Chem. Soc. 2003, 125, 12084.
(19) (a) Shaughnessy, K. H.; Hamann, B. C.; Hartwig, J. F. J. Org. Chem.
1998, 63, 6546. (b) Lee, S.; Hartwig, J. F. J. Org. Chem. 2001, 66,
3402. (c) Culkin, D. A.; Hartwig, J. F. Acc. Chem. Res. 2003, 36,
234. (d) Zhang, T. Y.; Zhang, H. Tetrahedron Lett. 2002, 43, 1363.
(e) Ku¨ndig, E. P.; Seidel, T. M.; Jia, Y.; Bernardinelli, G. Angew.
Chem., Int. Ed. 2007, 46, 8484.
(20) (a) Quallich, G. J.; Morrissey, P. M. Synthesis 1993, 51. (b) Forbes,
I. T. Tetrahedron Lett. 2001, 42, 6943.
On the other side, the intermolecular version of the C-N
coupling reaction would furnish the 2-aminophenylacetic acid
derivatives 31 which logically should be less oxidation prone.
Results are shown in Scheme 7are and listed in Table 1.
The coupling of L-alanine was first studied, and after a quick
screen it was found that again polar aprotic solvents were
markedly better for this transformation. No reactions occurred
in toluene (entries 1 and 2), while reactions in dioxane, tert-
(21) (a) Makosza, M.; Winiarski, J. Acc. Chem. Res. 1987, 20, 282. (b)
Lawrence, N. J.; Liddle, J.; Jackson, D. A. Synlett 1996, 1, 55. (c)
Makosza, M.; Wojciechowski, K. Liebigs Ann. 1997, 9, 1805. (d)
Makosza, M.; Kwast, A. Eur. J. Org. Chem. 2004, 10, 2125.
(22) (a) Walker, J.; Daisley, R. W.; Beckett, A. H. J. Med. Chem. 1970,
13, 983. (b) Hardtmann, G. E. Oxindoles as Sleep-Inducers. (Sandoz).
U.S. Patent 4,160,032, 1979. (c) Namil, A.; Benoit-Guyod, M.; Leclerc,
G. Eur. J. Med. Chem. 1995, 30, 973.
(23) (a) Noland, W. E.; Baude, F. J. Organic Syntheses; Wiley: New York,
1973; Collect. Vol. V, p 567. (b) Modi, S. P.; Oglesby, R. C.; Archer,
S. Organic Syntheses ; 1998; Collect. Vol. IX, p 601.
(24) (a) Prim, D.; Campagne, J.-M.; Joseph, D.; Andrioletti, B. Tetrahedron
2002, 58, 2041. (b) Schlummer, B.; Scholz, U. AdV. Synth. Catal.
2004, 346, 1599. (c) Buchwald, S. L.; Mauger, C.; Mignani, G.; Scholz,
U. AdV. Synth. Catal. 2006, 348, 23. (d) Corbet, J.-P.; Mignani, G.
Chem. ReV. 2006, 106, 2651.
(26) Ley, S. V.; Thomas, A. W. Angew. Chem., Int. Ed. 2003, 42, 5400.
(27) (a) Ma, D.; Yao, J. Tetrahedron: Asymmetry 1996, 7, 3075. (b) Ma,
D.; Zhang, Y.; Yao, J.; Wu, S.; Tao, F. J. Am. Chem. Soc. 1998, 120,
12459. (c) Ma, D.; Xia, C. Org. Lett. 2001, 3, 2583.
(28) Sassatelli, M.; Saab, E.; Anizon, F.; Prudhomme, M.; Moreau, P.
Tetrahedron Lett. 2004, 4827.
(25) (a) Wolfe, J. P.; Rennels, R. A.; Buchwald, S. L. Tetrahedron 1996,
52, 7525. (b) Yang, B. H.; Buchwald, S. L. Org. Lett. 1999, 1, 35. (c)
Poondra, R. R.; Turner, N. J. Org. Lett. 2005, 7, 863.
(29) van den Hoogenband, A.; Lange, J. H. M.; den Hartog, J. A. J.; Henzen,
R.; Terpstra, J. W. Tetrahedron Lett. 2007, 48, 4460.
Vol. 13, No. 3, 2009 / Organic Process Research & Development
•
445

Table 1. Screening of conditions for the intermolecular C-N coupling reaction
entry
R
conditions
oxindole 32a (%)
comments
1
Me toluene, CuI 10 mol %, 1.5 equiv K2CO3, 100 °C, 15 h
Me toluene, CuI 10 mol %, 1.5 equiv K3PO4, 100 °C, 15 h
Me dioxane, CuI 10 mol %, 1.5 equiv K2CO3, 100 °C, 15 h
Me tBuOH, CuI 10 mol %, 1.5 equiv K2CO3, 80 °C, 15 h
Me H2O, CuI 10 mol %, 1.5 equiv K2CO3, 100 °C, 15 h
Me DMF, CuI 10 mol %, 1.5 equiv K2CO3, 90 °C, 15 h
Me DMF, CuI 10 mol %, 1.5 equiv K2CO3, 110 °C, 1.5 h
Me DMF, CuI 10 mol %, 2.2 equiv K2CO3, 110 °C, 1.5 h
Me DMF, CuI 10 mol %, 1.5 equiv NaHCO3, 110 °C, 10 h
Me DMF, CuI 10 mol %, 1.5 equiv K2CO3, 140 °C, 1 h
Me DMF, CuI 5 mol %, 1.5 equiv K2CO3, 140 °C, 0.7 h
Me DMF, CuI 1 mol %, 1.5 equiv K2CO3, 140 °C, 1 h
0
0
no reaction
no reaction
2
3
20
19
29
73
89
77
13
88
91
61
very slow, many impurities
very slow
4
5
57% of phenol 33b
6
9% starting material, ee ) 82%c
ee ) 80%c
7
8
ee ) 64%c
9
reaction stalled
10
11
12
ee ) 76%c
ee ) 87.4%c
incomplete conversion and formation of
an unidentified impurity (17%)
5% of the same unidentified impurity
>4 kg scaled, ee ) 80%c, 60% yield after
crystallization (ee unchanged)
no reaction
13
14
Me DMF, CuI 3 mol %, 1.5 equiv K2CO3, 140 °C, 0.7 h
Me DMF, CuI 5 mol %, 1.5 equiv K2CO3, 140 °C, 3 h
68
84
15
16
17
H
H
H
DMF, CuI 10 mol %, 1.5 equiv K2CO3, 140 °C, 10 h
DMF, CuI 10 mol %, 2 equiv K3PO4, 140 °C, 10 h
DMF, CuI 10 mol %, 2 equiv Cs2CO3, 140 °C, 10 h
0
0
0
no reaction
no reaction
a HPLC area in the crude reaction mixture.
b
c Measured after cyclization to oxindole 32b in HCl.
d See Experimental Section for more details.
since.31 We thus had to revise our work to find an alternative
Scheme 7. Study of the intermolecular coupling
route to 3 (see next part and Figure 6).
Figure 6. Disconnection with 5-chlorooxindole 34 as starting
material.
butanol, or water were sluggish (20-30% yield on average after
15 h) and accompanied by the formation of many different
impurities (mainly phenol 33 in water, entry 5). On one hand,
73% (HPLC area %) of 2-aminophenylacetic acid derivative
31b was obtained in N,N-dimethylformamide at 90 °C (entry
6). Cyclization to the oxindole 32b was not observed under
these conditions, while significant racemization took place (but
did not intensify upon prolonged reaction time). On the other
hand racemization was found to be more significant when more
base was added (entry 8), while a temperature rise up to 140
°C completed the reaction in 30-40 min on a 0.5 g scale and
improved the yield and quality of the final product (entry 10).
The quantity of copper (I) iodide was also reduced successfully
down to 5 mol % (entry 11), but a drop to 1% (entry 12)
extended the reaction time, and reduced the yield while an
unidentified impurity was generated (17 HPLC area %). It was
therefore decided to stick to the best conditions found to scale
the C-N coupling reaction up to 4-5 kg of starting material
(entry 14, see Experimental Section for more details).
Aqueous hydrochloric acid was added to cyclize the 2-ami-
nophenylacetic acid 31b to the oxindole 32b. At that stage the
acidic pH is brought back to 7.0 with sodium bicarbonate to
break an oxindole-copper complex (first detected by the signals
broadening in the 1H NMR spectrum caused by paramagnetic
copper species), then the final uncomplexed oxindole 32b was
extracted and crystallized from methanol/water (2/1 v/v) in 60%
yield. This crystallization did not alter the enantiomeric excess
of the crude coupling reaction mixture.
The acid 32b obtained in this way was transformed to the
final primary amide (S)-4 using EDCI (1.1 equiv), HOBt (1
equiv), and NH4OHaq (2 equiv) in N,N-dimethylformamide (5
vol) at 0 °C for 2 h. This allowed us to isolate a crude material
by extraction with an unchanged ee of 80%. In parallel to this,
we found that the racemate amide rac-4 belongs to the class of
racemic compounds32 with a eutectic composition of 99.5/0.5.
The solubility of this eutectic composition has been evaluated
in ethyl acetate to be close to 20 mg/mL. We therefore decided
Unfortunately we were very disappointed to see that under
similar conditions, N-arylation of glycine did not work at all
(entries 15-17). This has already been disclosed in the
literature27b although efficient methods have been developed
(30) (a) Rossiter, S. Tetrahedron Lett. 2002, 43, 4671. (b) Bonanomi, M.;
Baiocchi, L. Gazz. Chim. Ital. 1989, 119, 445.
(31) Lu, Z.; Twieg, R. J. Tetrahedron Lett. 2005, 46, 2997.
(32) Jacques, J.; Collet, A.; Wilen, S. H. Enantiomers, Racemates and
Resolutions; Krieger: Malabar, FL, U.S.A., 1994; pp 88 and 192.
446
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development

Scheme 8. N-Alkylation attempts from 5-chlorooxindole 34
Scheme 9. Reductive amination from (5-chloro-2-aminophenyl)acetic acid (sodium salt) in aqueous solution
to capitalize on this situation and to use the minimum amount
of ethyl acetate to dissolve a 99.5/0.5 composition (eutectic)
from the crude 90/10, thus leaving a chemically pure solid with
a nearly racemic composition (55/45 as measured by chiral
HPLC, see Experimental Section). The solution containing the
expected amide (S)-4 was then stripped under vacuum and the
residue slurried in isopropanol (6 vol) to lead to (S)-4 in 67%
isolated yield (99% ee).
N-Alkylation Route. In parallel to this, the disconnection
involving the readily available 5-chlorooxindole 34 as advanced
starting material was also explored. However, it had been shown
previously that deprotonation of such substrates followed by
alkylation does not proceed cleanly (see Scheme 8/pathway A).
For these reasons we investigated the possibility of making
the C-N bond of 3 by reductive amination with glyoxylic acid.
However Scheme 8(pathway B) shows that even if amidation
of glyoxylic acid is possible (though always incomplete),
reduction of the iminium species after addition of sodium
triacetoxyborohydride to lead to the acetic acid compound 32a
was never observed. Instead of that the intermediate hemiaminal
smoothly loses a molecule of water to lead to dehydration
products.
In a similar manner the “benzotriazole” methodology
developed by Katritzky33 was briefly investigated, but the
reaction of 5-chlorooxindole 34 with benzotriazol-1-yl-
hydroxyacetic acid 3534 (pathway C) led to the same
disappointing results.
A paper in the literature reporting the preparation of tenidap35
(structure based on the 5-chlorooxindole moiety) mentions the
possibility of carrying out the ring opening of the oxindole 34
with aqueous sodium hydroxide and then further functionalizing
the resulting aniline with an electrophile. This prompted us to
explore the possibility of incorporating the amino acid residue
Via a reductive amination protocol after the ring opening of
oxindole 34 (Scheme 9). The cleavage of the amide bond of
the oxindole 34 was reproduced with sodium hydroxide 4 M
(5 equiv) at 100 °C overnight followed by neutralization to pH
6, and extraction into ethyl acetate allowed us to isolate the
phenylacetic acid 38 in 74% yield. Unfortunately the latter
(33) Katritsky, A. R.; Yao, G.; Lan, X.; Zhao, X. J. Org. Chem. 1993, 58,
2086.
(34) Katritzky, A. R.; Rachwal, S.; Rachwal, B. J. Chem. Soc. Perkin Trans.
I 1987, 791.
(35) Kumar, P. R.; Goud, P. S.; Raju, S.; Sailaja, M.; Sarma, M. R.; Reddy,
G. O. Org. Process Res. DeV. 2001, 5, 61.
Vol. 13, No. 3, 2009 / Organic Process Research & Development
•
447

recyclizes in solution to the oxindole 34 too quickly to be
considered as an isolable and viable intermediate.
but proved to be uneffective with glycine as aminoacid. Thus,
another strategy involving a one-pot “oxindole ring opening/
reductive amination/ring closure” was chosen for the second
lead oxindole.
Overall the processes were safe, reproducible, reliable,
robust, and high yielding. Isolation of final products was very
convenient, and the purity and impurity profiles were strictly
controlled and reproducible.
We also looked at the possibility to isolate a carboxylate
salt of this amino acid as intermediate. All attempts to isolate
the sodium or potassium salt failed, but following a previous
experience in another related transformation we discovered that
oxindole 34 is slowly cleaved in aqueous barium hydroxide
(0.5 equiv, 100 °C, 20 h), while the barium carboxylate 39
precipitates spontaneously from the reaction mixture in 72%
isolated yield. Unfortunately salt 39 turned out to be too poorly
soluble to further react significantly in a reductive amination
process.
It was then discovered that the reductive amination protocol
from the sodium salt could be performed in the original basic
aqueous solution. The previously obtained solution of 2-ami-
nophenylacetate was cooled to 20 °C, and aqueous glyoxylic
acid (2 equiv) was added in tetrahydrofuran. Glacial acetic acid
was added until pH adjusted to 5, followed by the portionwise
addition of 1.5 equiv of solid sodium triacetoxyborohydride
while the reaction mixture was kept at 15 °C. The temperature
was maintained below 20 °C in order to avoid the ring-clo-
sure of the unreacted 2-aminophenylacetate to the oxindole 34.
After 17 h at this temperature, the complete conversion to
the expected aniline was observed. Concentrated hydrochloric
acid was then added to adjust the pH to 2 which allowed the
cyclization and the precipitation of the acetic acid derivative
32a in 85% yield (98% purity HPLC). Ultimately this method
proved very efficient and successful upon scaling-up (>2.5 kg
scale, see Experimental Section).
The last goal, quite trivial at first sight, was to prepare the
primary amide 3 from the freshly prepared acid 32a. This one-
step transformation turned out to be much more difficult than
expected. Indeed, the use of classical coupling agents (carbonyl
diimidazole CDI, carbodiimide EDCI/HOBt, uronium salts
TBTU/HOBt), and different ammonia equivalents (ammonium-
hydroxide, -chloride, or -bicarbonate) led only to partial con-
versions together with the formation of a dimeric compound
(often as major product) whose structure was not determined.
Finally it was decided to perform this transformation in two
steps as explained below. Indeed acid 32a is cleanly transformed
to the corresponding methyl ester 40 under standard conditions
(2 equiv thionyl chloride in methanol, 0° to 25 °C overnight)
which is subsequently crystallized from dichloromethane/
methanol in 73% isolated yield. This ester 40 is then smoothly
converted to the expected amide 3 in aqueous ammonia
(NH4OH 28% w/w, 6 mL/g of ester 40, 24 h, 20 °C). Under
these conditions, conversion of the ester 40 is always greater
than 98% although the reaction is accompanied by a reproduc-
ible ∼5% quantity of hydrolysis product 32a. The final product
3 is easily purified and isolated by crystallization from acetone/
water (3:1) at 60 °C in 70% yield (43% overall from 34).
Experimental Section
Unless specified otherwise, all reagents and solvents were
used as supplied by the manufacturer. All reactions were
conducted under an inert nitrogen atmosphere. 1H and 13C NMR
spectra were recorded in DMSO-d6 with TMS as internal
standard using a Bruker AV 400 spectrometer at 400 and 100
MHz respectively. HPLC was performed on a Waters Alliance
2695 mounted with a X-Terra C18, 5 µm, 150 mm × 4.6 mm
column, using a water/acetonitrile/NH4CO3/NH4OH gradient
method at a 2.5 mL/min at 215 nm or with an Inertsil ODS3,
5 µm, 150 mm × 4.6 mm column, using a water/acetonitrile/
TFA gradient at a 2.5 mL/min at 215 nm. Chiral purity was
determined by HPLC using a Prontosil Chiral AX-QN 1, 150
mm × 4 mm column, methanol/formic acid: 99/1 at 1 mL/min
(Conditions A) or a Chiralcel OD-H, 250 mm × 4.6 mm
column, ethanol/isohexane/diethylamine: 50/50/0.1 at 1 mL/
min (Conditions B). LC/MS analysis was performed on a ZQ
Waters single quadrupole spectrometer equipped with a ESI
source and HPLC Waters 2795 quaternary pump with diode
array. Data were acquired in a full MS scan from m/z 90 to
1000 in positive mode with an acidic elution and in both positive
and negative modes with a basic elution. Specific rotation was
recorded on a Perkin-Elmer MC341 polarimeter. DSC were
recorded on a Perkin-Elmer DSC 7 with a heating ramp of
10 °C/min.
(5-Chloro-2-oxo-2,3-dihydroindol-1-yl)acetic acid, 32a. A
70 L vessel, charged with aqueous 4 M NaOH (20.59 L, 82.36
mol), was heated to 60 °C, and 5-chlorooxindole 34 (2.817 kg,
16.81 mol) was added. The reaction mixture was heated at 100
°C for 19 h. Upon completion of the hydrolysis, the mixture
was cooled down to 15 °C, aqueous 50 wt % glyoxylic acid
(4.878 kg, 32.94 mol) diluted with 5.6 L of THF was charged,
and the pH of the reaction mixture was adjusted to 5 by addition
of acetic acid (3.5 L). Solid sodium triacetoxyborohydride
(5.398 kg, 25.47 mol) was added portionwise, maintaining mass
temperature below 20 °C, and the mixture was stirred for 17 h
at 15 °C. After addition of HCl, 37%, until the pH ) 2 (7.5 L),
the solid was filtered, washed with water (2 × 8.5 L), and dried
under vacuum at 70 °C for 48 h. Yield: 3.224 kg (85%), HPLC
purity: 98.3% (area %); 1H NMR (DMSO-d6) δ 7.34 (s, 1 H),
7.30 (d, J ) 8.3 Hz, 1 H), 6.99 (d, J ) 8.3 Hz, 1 H), 4.44 (s,
2 H), 3.66 (s, 2 H); MS m/z: 226/228 (MH+).
(5-Chloro-2-oxo-2,3-dihydroindol-1-yl)acetic acid methyl
ester, 42. A 40 L vessel was charged with (5-chloro-2-oxo-
2,3-dihydro-indol-1-yl)acetic acid 32a (3.220 kg, 14.27 mol)
and 25 L of methanol, and the mixture was cooled to 0 °C.
Thionyl chloride (3.47 kg, 29.17 mol) was then added dropwise,
maintaining mass temperature below 10 °C, and the reaction
mixture was stirred for 19 h at 25 °C. The suspension was
Summary and Conclusions
During the development of two new ligands for the SV2A
protein, the requirement for bulk (multikilogram) drug supplies
within a strict time frame were met using two distinct routes.
The first relied on a copper-mediated C-N coupling reaction
of L-alanine to indroduce the chiral (S)-propionamide side chain
448
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development

cooled to 0 °C and stirred for 1 h. The solid was filtered, washed
with 12.5 L of methanol at 0 °C, and dried under vacuum at
60 °C for 18 h to afford 3.256 kg of crude 42 as a yellow solid.
The crude material was dissolved in dichloromethane (12.6 L)
and stirred with 320 g of activated charcoal at 25 °C for 1 h.
The suspension was filtered on a Celite pad (Hyflo Super Cel)
and concentrated to dryness, and the residue was taken up with
methanol (29.3 L). The mixture was stirred at 0 °C for 1 h,
and the solid was filtered, washed with cold methanol (0 °C)
(6.4 L), and dried under vacuum at 60 °C for 18 h to afford 42
as a white solid. Yield: 2.520 kg (73%), HPLC purity: 99.7%
and water (4.67 L) was added dropwise. The suspension was
then stirred at 20 °C for 16 h and the solid was filtered, washed
with a 2:1 mixture of methanol/water (1.5 L), and dried under
vacuum at 45 °C for 18 h to afford chemically pure 32b. Yield:
2.472 kg (60%), HPLC purity: 99.2% (area %); ee ) 80%
determined by chiral HPLC upon Conditions A: 7.1 min (90.1%,
1
S-isomer) and 8.5 min (9.9%, R-isomer); H NMR (DMSO-
d6) δ 13.01 (broad, 1 H), 7.35 (s, 1 H), 7.30 (d, J ) 8.3 Hz, 1
H), 6.94 (d, J ) 8.3 Hz, 1 H), 5.04 (q, J ) 7.3 Hz, 1 H), 3.63
(s, 2 H), 1.47 (d, J ) 7.3 Hz, 3 H); MS m/z: 240/242 (MH+).
(S)-2-(5-Chloro-2-oxo-2,3-dihydroindol-1-yl)propiona-
mide, (S)-4. A 40 L vessel was charged with (S)-2-(5-chloro-
2-oxo-2,3-dihydroindol-1-yl)propionic acid 32b (2.465 kg, 10.28
mol, ee ) 80%) and DMF (12.3 L). The solution was cooled
down at 0 °C, EDCI (2.170 kg, 11.32 mol) and HOBT
monohydrate (1.575 kg, 10.28 mol) were added, and the mixture
was stirred at 0 °C for 2 h. Aqueous ammonia 28% (1.6 L)
was added over a period of 15 min, maintaining mass temper-
ature below 5 °C. After stirring at 0 °C for 1 h and addition of
water (12.5 L), the mixture was extracted with dichloromethane
(2 × 12.5 L). Dichloromethane was stripped off, and the residue
was taken up with enough ethyl acetate (26 L) to dissolve the
entire eutectic mixture. The resulting solid (55/45 mixture of
(S)-4 and its enantiomer) was filtered, and the filtrate was
washed with water (3 × 10 L) and concentrated under vacuum
to afford crude (S)-4 as a brownish solid (1.802 kg). This solid
was then suspended in isopropanol (11 L), stirred at 40 °C for
1 h, filtered, washed with isopropanol (2 × 0.6 L), and dried
under vacuum at 40 °C for 18 h to afford pure (S)-4. Yield:
1.645 kg (67%), HPLC purity: 99.0% (area %); ee ) 99%
determined by chiral HPLC upon Conditions B: 6.6 min (0.7%,
R-isomer) and 9.4 min (99.3%, S-isomer); [R]25D ) -59.5° (c
) 1, MeOH); 1H NMR (DMSO-d6) δ 7.51 (s, 1 H), 7.35 (s, 1
H), 7.29 (d, J ) 8.6 Hz, 1 H), 7.24 (s, 1 H), 6.84 (d, J ) 8.6
Hz, 1 H), 4.97 (q, J ) 7.3 Hz, 1 H), 3.59 (s, 2 H), 1.43 (d, J
) 7.3 Hz, 3 H). MS m/z: 239/241 (MH+).
1
(area %); H NMR (DMSO-d6) δ 7.35 (s, 1 H), 7.31 (d, J )
8.3 Hz, 1 H), 7.01 (d, J ) 8.3 Hz, 1 H), 4.56 (s, 2 H), 3.68 (2
× s, 2 H + 3 H); MS m/z: 240/242 (MH+).
2-(5-Chloro-2-oxo-2,3-dihydroindol-1-yl)acetamide, 3. A
40 L vessel was charged with (5-chloro-2-oxo-2,3-dihydroindol-
1-yl)acetic acid methyl ester 42 (2.518 kg, 10.50 mol) and 15.1
L of aqueous ammonia 28% w/w. After stirring at 20 °C for
21 h, the solid was filtered, washed with water (3 × 5 L), and
dried under vacuum at 60 °C for 48 h to afford 2.160 kg of
crude 3. The crude material was suspended in a 3/1 mixture of
acetone/water (44 L) and heated at 60 °C for 1 h. The mixture
was cooled down to 0 °C and stirred for 30 min; the crystals
were filtered, washed with cold acetone (0 °C) (3 × 3.6 L),
and dried under vacuum at 60 °C for 18 h to afford pure 3 as
an off-white solid. Yield: 1.628 kg (69%), HPLC purity: 100%
(area %); 1H NMR (DMSO-d6) δ 7.59 (s, 1 H), 7.33 (s, 1 H),
7.29 (d, J ) 8.3 Hz, 1 H), 7.22 (s, 1 H), 6.85 (d, J ) 8.3 Hz,
1 H), 4.24 (s, 2 H), 3.60 (s, 2 H); MS m/z: 225/227 (MH+);
DSC (onset): 236.18 °C.
(S)-2-(5-Chloro-2-oxo-2,3-dihydroindol-1-yl)propionic acid,
32b. A 70 L vessel was charged with (2-bromo-5-chlorophe-
nyl)acetic acid 27 (4.247 kg, 17.02 mol), (S)-alanine (1.518 kg,
17.04 mol), potassium carbonate (3.531 kg, 25.55 mol), CuI
(0.162 kg, 0.85 mol), and DMF (20 L). The mixture was then
heated to 140 °C and stirred for 45 min. After cooling down at
20 °C, 2 N HCl (22 L) was added, and the mixture was stirred
for 1 h and extracted with ethyl acetate (2 × 30 L). The
combined organic layers were then washed with water (2 ×
16 L), and after addition of water (22 L), the pH was adjusted
to 7 with 1 N NaOH (15 L). The two layers were separated,
ethyl acetate (46 L) was added to the aqueous layer, and the
pH was adjusted to 2 by addition of 2 N HCl (8 L). After
decantation, the organic layer was washed with water (10 L)
and evaporated to dryness to afford 3.98 kg of crude 32b as a
brownish solid. HPLC purity: 95% (area %); ee: 80%. This
crude material was dissolved in methanol (9.33 L) at 60 °C,
Acknowledgment
We thank Alain Fauconnier, Jehan Claessens, Christelle
Derwa, Ariane Descamps (mass spectrometry analyses), Reiner
Dieden and Benoˆıt Mathieu (NMR spectroscopy). We also
thank the reviewers for proofreading carefully the first version
of the manuscript.
Received for review August 21, 2008.
OP800203P
Vol. 13, No. 3, 2009 / Organic Process Research & Development
•
449