A new alternative synthesis of 5-cyanophthalide, a versatile intermediate in the preparation of the antidepressant drug citalopram

Article abstract of DOI:10.1016/S0014-827X(01)01108-9

An alternative versatile synthesis of 5-cyanophthalide, a key synthetic intermediate in the preparation of the antidepressant drug Citalopram, is presented. The synthesis reported here allows the preparation of this important intermediate in three steps, avoiding the manipulation of environmentally detrimental cyanides. Copyright

Full text of DOI:10.1016/S0014-827X(01)01108-9

www.elsevier.com/locate/farmac
Il Farmaco 56 (2001) 715–718
Short Communication
A new alternative synthesis of 5-cyanophthalide, a versatile
intermediate in the preparation of the antidepressant drug
citalopram
Fabrizio Micheli *, Luca Crippa, Daniele Donati, Romano Di Fabio, Colin Leslie
GlaxoSmithKline Group, GlaxoWellcome SpA, Medicines Research Centre, Via Fleming 4, 37135 Verona, Italy
Received 15 March 2001; accepted 30 March 2001
Abstract
An alternative versatile synthesis of 5-cyanophthalide, a key synthetic intermediate in the preparation of the antidepressant drug
Citalopram, is presented. The synthesis reported here allows the preparation of this important intermediate in three steps, avoiding
the manipulation of environmentally detrimental cyanides. © 2001 Elsevier Science S.A. All rights reserved.
Keywords: Antidepressant; Citalopram; Synthesis
1. Introduction
makes them environmentally unfriendly and their use is
best avoided where possible. Copper cyanide is much
less toxic than the corresponding sodium or potassium
salt, but it is also much less water soluble if not
5-Cyanophthalide (1, Fig. 1) (1-oxo-1,3-dihydro-
isobenzofuran-5-carbonitrile) is a key intermediate in
the synthesis of the antidepressant drug Citalopram (2,
Fig. 1), a well known Serotonin reuptake inhibitor. The
original approach in preparation of this drug required
the reaction of intermediate 1 with the appropriate
Grignard reagents to introduce the requisite side chains;
surprisingly, during these reactions, almost no nucle-
ophilic attack on the cyano moiety is observed and only
modest amounts of by-products are reported. It is not
surprising that, considering the high commercial value
of Citalopram, a number of patent applications cover-
ing alternative methodologies for its synthesis have
been filed subsequently [1–9].
Fig. 1.
Apparently, less attention was being paid to interme-
diate 1. Its first preparation was reported in 1931 by
Levy [10]. This procedure involves the reaction of the
corresponding diazonium salt with copper cyanide as
depicted in Scheme 1 (route a).
Cyanides are commonly employed reagents in many
industrial fields, but clearly their highly toxic nature
* Corresponding author.
E-mail address: fm20244@glaxowellcome.co.uk (F. Micheli).
Scheme 1.
0014-827X/01/$ - see front matter © 2001 Elsevier Science S.A. All rights reserved.
PII: S0014-827X(01)01108-9

716
F. Micheli et al. / Il Farmaco 56 (2001) 715–718
After having evaluated the commercially available
starting materials in which the cyano moiety was al-
ready present and analysing the target from a retro-syn-
thetic point of view, two possible alternatives for the
generation of the lactone ring were identified, which are
illustrated in Scheme 2:
1. Start from a halogenated benzonitrile, introduce the
carboxylic moiety, generate an appropriate leaving
group and build the lactone.
Scheme 2.
2. Start from 4-cyanobenzoic acid, introduce a hydrox-
ymethyl equivalent and build the lactone.
2. Results and discussion
According to the hypothesised path a in Scheme 2,
we decided to start from commercially available 4-
bromo-3-methyl benzonitrile (3) as depicted in Scheme
3.
Scheme 3.
Metalꢀhalogen exchange was performed using t-BuLi
at low temperature and the resulting anion was trapped
with solid carbon dioxide. The resulting benzoic acid
(4) was reacted under radical conditions to introduce a
leaving group on the methyl group and a number of
cyclisation reactions were attempted on this transient
intermediate.
Various bases such as NaOH, K2CO3 and NaHCO3
were used to promote the bromo-lactonisation, unfor-
tunately the desired product was never identified. Prob-
ably, the cyano moiety was so sensitive to the radical
conditions to originate a number of side reactions
which, in the end, led to complex reaction mixtures and
extensive degradation.
Due to the poor results obtained following path a, we
moved to the proposed path b starting from 4-
cyanobenzoyl chloride as depicted in Scheme 4.
The benzoyl chloride was transformed into the ortho-
directing phenyl amide (y=92%) using equimolar
quantities of the reactants and a slight excess of pyri-
dine in dichloromethane at r.t. The use of t-BuLi to
generate the ortho-anion was not possible because of
the reactivity of the cyano group; in all of the condi-
tions used the undesired t-butyl ketone derivative (7)
was isolated. To limit this side reaction we decided to
use a more hindered base. Lithium tetramethylpipe-
ridine (LiTMP), freshly prepared in THF at −20°C,
was therefore used to ortho-metalate the intermediate
cyano phenyl amide (5) (−78°C, 30 min). The hydrox-
ymethyl group was introduced in its masked form as a
formyl moiety (DMF, −78°C, 1.5 h followed by
overnight stirring at r.t.). Acid quenching of the reac-
tion mixture (1 N HCl was added dropwise until pH 1)
followed by phase extraction with AcOEt and chro-
matographic purification (cyclohexane/AcOEt 70:30)
consequently allowed the isolation of the desired hemi-
aminal (6) in 68% yield.
Scheme 4.
prepared freshly in situ or complexed using a large
excess of the water-soluble sodium or potassium
cyanides. Another preparation of intermediate 1 involv-
ing cyanides was reported in 1951 by Tirouflet [11].
Two recent patents [12,13] report the preparation of
intermediate 1 avoiding the need of cyanides and start-
ing from 5-carboxyphtalide. The first methodology
(Scheme 1, route b) involves the preparation of an
amide which is dehydrated to the corresponding cyano
derivative, while the second method (Scheme 1, route c)
involves the reaction of the acid itself with a dehydrat-
ing agent in the presence of an appropriate
sulphonamide.
In our continuing effort to develop new synthetic
methodologies, we were interested in finding alternative
methods for the preparation of 5-cyanophthalide via a
simple method, free of the need of using cyanides, and
starting
intermediates.
from
readily
available
commercial

F. Micheli et al. / Il Farmaco 56 (2001) 715–718
717
The reduction of this intermediate was performed
with sodium borohydride (NaBH4, 14 equiv. in MeOH,
24 h, r.t.), the solvent was evaporated and the crude
was submitted to acid cyclisation (1 N HCl, 48 h, r.t.).
After extraction (CH2Cl2) and chromatographic purifi-
cation (cycloexane/AcOEt 9:1) the desired product (1)
was obtained in acceptable yield.
Direct introduction of the hydroxymethyl moiety (8)
was also attempted (step iv, Scheme 4), unfortunately
the use of gaseous formaldehyde proved unsuccessful.
To conclude, we have described a novel three-step
synthesis of 5-cyanophthalide (1) starting form com-
mercially available intermediates, which does not in-
volve the manipulation of toxic cyanides.
carefully added dropwise. The reaction was stirred for
10 min at −20°C, warmed to r.t. and additionally
stirred for 30 min.
This solution was added dropwise to a solution of
intermediate 5 (150 mg, 0.67 mmol) in dry THF (5 ml)
at −78°C. The resulting mixture was stirred at −78°C
for 30 min and subsequently DMF (0.115 ml, 1.48
mmol) was added. The reaction was then stirred for 1.5
h at −78°C and overnight at r.t.
The mixture was quenched by adding water (20 ml)
and 1 N HCl was added dropwise until pH 1. The
mixture was extracted with AcOEt (20 ml×3). The
organic phase was dried over sodium sulphate and
evaporated under reduced pressure. The crude was
purified by chromatography (cyclohexane/AcOEt 7:3)
to give 115 mg of intermediate 6 as a yellow oil
(y=68%).
3. Experimental
IR (nujol, cm−1): w(NH) 3340, w(CꢁN) 2231, w(CꢂO)
1
Infrared spectra were recorded on a Bruker IFS 48
spectrometer. 1H NMR spectra were recorded on a
Varian Unity 400 (400 MHz); the data are reported as
follows: chemical shift (in ppm) from the Me4Si line as
external standard, multiplicity (b=broad, s=singlet,
d=doublet, t=triplet, q=quartet, m=multiplet).
Chromatography was carried out by use of Merck
silica gel 60 (230–400 mesh). Elemental analyses were
determined by an EA 1108 Carlo Erba elemental analy-
ser and the analyses of the reported compounds are in
agreement with the calculated values. All the reactions
were carried out under a controlled atmosphere in
flame dried glassware. Anhydrous solvents were pur-
chased from Aldrich. Reactions were monitored by
analytical thin-layer chromatography (TLC) using
Merck silica gel 60 F-254 glass plates (0.25 mm).
1674. H NMR (DMSO-d6): l 8.16 (dd, 1H), 8.05 (dd,
1H), 7.96 (d, 1H), 7.86 (dd, 2H), 7.46 (td, 2H), 7.25 (t,
1H), 6.70 (d, 1H), 6.02 (d, 1H).
3.3. 1-Oxo-1,3-dihydro-isobenzofuran-5-carbonitrile (1)
Intermediate 6 (10 mg, 0.04 mmol) was dissolved in
MeOH (20 ml) and NaBH4 (21.16 mg, 0.56 mmol) was
added. The mixture was stirred for 24 h at r.t., the
solvent was evaporated and water (20 ml) was added.
The solution was treated with 1 N HCl until pH 1 and
the resulting mixture was stirred for 48 h at r.t. The
aqueous phase was extracted with CH2Cl2 (20 ml×5).
The organic phase was dried over sodium sulphate and
evaporated under reduced pressure. The crude was
purified by chromatography (cycloexane/AcOEt 9:1) to
give 2.5 mg of the desired product (1) (y=35%). IR
and NMR spectra in accordance to those reported in
the literature.
3.1. 4-Cyano-N-phenylbenzamide (5)
4-Cyanobenzoylchloride (4.0 g, 24.2 mmol), aniline
(2.2 g, 24.2 mmol) and pyridine (2.64 ml, 26.6 mmol)
were dissolved carefully in methylene chloride (20 ml)
in a round bottom flask. The reaction was stirred
overnight at r.t. The mixture was then washed with 0.1
N HCl and with a saturated solution of K2CO3. The
organic layer was then dried over sodium sulphate and
evaporated at reduced pressure to give 5.0 g of interme-
diate 5 as a white solid (y=92%).
Acknowledgements
We would like to thank Dr. Carla Marchioro and the
Analytical Dept. for their precious collaboration.
References
IR (nujol, cm−1): w(NH) 3354, w(CꢁN) 2232, w(CꢂO)
[1] K.P Bøgesø, A.S. Toft, Anti-depressive substituted 1-dimethy-
laminopropyl-1-phenyl phtalans. U.S. Patent 4,136,193, 1979.
[2] K.P. Bøgesø, Intermediate in the preparation of 1-(3-dimethyl-
aminopropyl)-1-(4%-fluorophenyl)-1,3-dihydroisobenzofuran-5-
carbonitrile. EP 171943, 1986.
1
1656. H NMR (DMSO-d6): l 10.47 (bs, 1H), 8.10 (m,
2H), 8.03 (m, 2H), 7.77 (m, 2H), 7.37 (m, 2H), 7.13 (m,
1H).
[3] K.P. Bøgesø, J. Perregaard, Preparation and isolation of antide-
pressant drug citalopram enantiomers and their pharmaceutical
compositions. EP 347066, 1989.
[4] H. Petersen, K.P. Bøgesø, M. Bech Sommer, Method for the
preparation of citalopram. WO 9819511, 1998.
3.2. 1-Oxo-3-hydroxy-1,3-dihydro-N-phenyl-
isoindole-5-carbonitrile (6)
HTMP (0.250 ml, 1.48 mmol) was dissolved in dry
THF (5 ml) and cooled to −20°C in a round bottom
flask and 1.6 M BuLi (0.925 ml, 1.48 mmol) was
[5] H. Petersen, P. Bregnedal, K.P. Bøgesø, Method for the prepara-
tion of citalopram. WO 9819512, 1998.

718
F. Micheli et al. / Il Farmaco 56 (2001) 715–718
[6] H. Petersen, Bregnedal, K.P. Bøgesø, Method for the prepara-
[10] S. Levy, J. Chem. Soc. (1931) 867–870.
[11] J. Tirouflet, Bull. Soc. Sci. Bretagne (1951) volume 26, Sonder-
heft S.39.
[12] H. Petersen, P. Dahlberg Nielsen, Esterification, amidation and
dehydration method for the preparation of 5-cyanophthalide.
WO 0039112, 2000.
tion of citalopram. WO 9819513, 1998.
[7] H. Petersen, Method for the preparation of citalopram. WO
9930548, 1999.
[8] M.H. Rock, H. Petersen, P. Ellegaard, Method for the prepara-
tion of citalopram. WO 0012044, 2000.
[9] H. Petersen, M.H. Rock, H. Svane, Method for the preparation
of citalopram. WO 0013648, 2000.
[13] B. Weber, Manufacture of 5-cyanophthalide by dehydration of
5-carboxyphthalide. WO 0044738, 2000.
.