Response to dancer's comments on our article "substrate modification approach to achieve efficient resolution: Didesmethylcitalopram: A key intermediate for escitalopram" [Org. Process Res. Dev. 2007, 11, 289-292]

Article abstract of DOI:10.1021/op8002079

Recently, we published a synthesis of escitalopram (S-1) consisting of the resolution of didesmethylcitalopram (3) and subsequent methylation of S-didesmethylcitalopram (S-3) (Org. Process Res. De v. 2007, 11, 289-292). Some of our observations regarding

Full text of DOI:10.1021/op8002079

Organic Process Research & Development 2009, 13, 34–37
Response to Dancer’s Comments on Our Article “Substrate Modification Approach
to Achieve Efficient Resolution: Didesmethylcitalopram: A Key Intermediate for
Escitalopram” [Org. Process Res. Dev. 2007, 11, 289-292]
Chandrashekar R. Elati,* Naveenkumar Kolla, and Vijayavitthal T. Mathad*
Department of Research and DeVelopment, Dr. Reddy’s Laboratories Ltd., IPD, Unit-III, Bollaram,
Hyderabad - 502325, Andhrapradesh, India
Table 1. Chiral purities of the precipitate and solids
Abstract:
obtained from concentration of the filtrate
Recently, we published a synthesis of escitalopram (S-1) consisting
of the resolution of didesmethylcitalopram (3) and subsequent
methylation of S-didesmethylcitalopram (S-3) (Org. Process Res.
DeW. 2007, 11, 289-292). Some of our observations regarding
citalopram resolution and C-alkylation of a benzofuran analogue
(2) to produce didesmethylcitalopram (3) were disputed by Dr.
Dancer of H. Lundbeck (preceding article). A detailed response
to his comments regarding stabilization of the 3-chloroproylamine
free base by dilution with certain solvents, its storage and handling,
optimized experimental conditions for C-alkylation to prepare
didesmethylcitalopram, and a corrected process for citalopram
resolution are included.
precipitated solid
solid from filtrate
S-isomer R-isomer S-isomer R-isomer
isolation
(%)
(%)
(%)
(%)
first isolation
second isolation
third isolation
45.73
48.23
57.56
54.25
51.77
42.44
59.33
83.33
96.38
40.67
16.67
3.62
Results and Discussion
We agree with the findings of Dancer and Lopez De Diego2
wherein the resolution of citalopram with (-)-DPTTA as the
chiral resolving agent is not feasible in the manner cited in our
article, but we disagree that resolution of citalopram (1) is not
possible by other means. The desired isomer, S-citalopram, was
isolated from the mother liquor after repeated resolutions as
discussed in the Introduction. With respect to the process for
C-alkylation for the preparation of didesmethylcitalopram (3),
we disagree with their conclusions, and our response is
discussed in this section, Results and Discussion.
Introduction
Our recent article1 published in Org. Process Res. DeV.
described a three-step synthesis of escitalopram (S-1) consisting
of: (i) alkylation of 1-(4-fluorophenyl)-1,3-dihydro-isobenzo-
furan-5-carbonitrile (2) to produce didesmethylcitalopram (3),
(ii) diastereomeric salt resolution of didesmethylcitalopram (3)
using (-)-di-p-toluoyltartaric acid (-)-DPTTA as the chiral
acid, and (iii) Eschweiler-Clarke methylation of S-didesmeth-
ylcitalopram (S-3) using formic acid and formaldehyde to form
escitalopram (S-1) (Scheme 1). Unfortunately, the synthesis of
escitalopram (S-1) via the diastereomeric salt resolution of
citalopram (1) with DPTTA as a chiral acid was not feasible at
industrial scale as disclosed due to poor yields, low purities,
and a lengthy process. Since citalopram (1) was an unfavorable
substrate to achieve efficient resolution, our efforts were
redirected towards the synthesis of didesmethylcitalopram (3)
whose subsequent successful resolution enabled us to achieve
the synthesis of escitalopram (S-1). In summary, our manuscript
concerned the substrate modification approach that offered a
better molecule for resolution as compared to citalopram (1)
itself.
Resolution of Citalopram. We appreciate the efforts made
by Dancer and Lopez De Diego regarding the evaluation of
our results for citalopram resolution.1 We agree with them to
the extent that the diastereomeric salt of S-citalopram cannot
be crystallized by reaction with DPTTA; however, we discov-
ered that harvesting S-citalopram salt from the filtrate of the
salt mixture was feasible. This was conducted by producing a
solution of citalopram and (+)-DPTTA monohydrate in metha-
nolic acetonitrile at 70-75 °C and gradually cooling to 25-30
°C. Following a hold period, the precipitated solid was removed
by filtration, and the resulting filtrate was concentrated to obtain
a solid. The chiral purities of filtered solid and filtrate-derived
residue were 54.25:45.73 and 40.67:59.33 for the R- and
S-isomers, respectively. The residue was hydrolyzed with 10%
aqueous sodium hydroxide solution, and following workup in
toluene, the free base was attained. This was subjected to the
above process twice more to produce S-(+)-1-(+)-DPTTA salt
from the filtrate in 96.4% chiral purity4 (Table 1) in 11.0%
overall yield. This finding is in agreement with example 11 cited
in our patent.3 During the preparation of our article, we
inadvertently missed incorporating a few words in the text, e.g.
After publication of our article,1 Dancer and Lopez De Diego
have published observations2 concerning the nonreproducibility
of (a) the resolution of citalopram (1) in the manner detailed in
our article and (b) the method for the C-alkylation step cited in
our patent.3
* Corresponding authors. E-mail: drvtmathad@yahoo.co.in; chandrashekarerr@
drreddys.com.
(3) Sundaram, V.; Mathad, V. T.; Elati, R. R. C.; Kolla, N.; Vankawala,
P. J.; Govindan, S.; Chalamala, S.; Gangula, S. WO Patent 047274;
Chem. Abstr. 2005, 142, 481937.
(4) Enantiopurity was estimated by chiral HPLC analysis with Chiralcel
ODH, 250 mm × 4.6 mm, 5 µ; mobile phase: 2-propanol, n-hexane,
and diethylamine in the ratio of 50:950:2.0 (v/v); 0.8 mL/min; 240
nm.
(1) Elati, C. R.; Kolla, N.; Vankawala, P. J.; Gangula, S.; Chalamala, S.;
Sundaram, V.; Bhattacharya, A.; Vurimidi, H.; Mathad, V. T. Org.
Process Res. DeV. 2007, 11, 289–292.
(2) Dancer, R. J.; Lopez de Diego, H. Org. Process Res. DeV. 2008, 13,
23–33.
34
•
Vol. 13, No. 1, 2009 / Organic Process Research & Development
10.1021/op8002079 CCC: $40.75
2009 American Chemical Society
Published on Web 11/21/2008

Scheme 1. Synthesis of escitalopram via preparation and resolution of didesmethylcitalopram
Scheme 2. Self polymerization of 3-chloropropylamine
“repetitive diastereomeric salt formation of the product collected
from the concentration of the filtrate”, which led to the
nonreproducibility of the example cited in our article. The chiral
acid we listed: (-)-DPTTA should have been (+)-DPTTA. We
sincerely apologize for this unintentional error.
Table 2. Stability study results of 3-CPA free base in
different solvents
C-Alkylation of the Benzofuran Analogue. We did not
discuss the synthesis of didesmethylcitalopram (3) in detail in
our previous publication as our intention was to focus on the
resolution.1 The findings of Dancer and Lopez De Diego2 with
respect to C-alkylation are based solely on the presumption that
the free base of chloropropylamine is unstable, which is further
based on the disclosure in the literature5 that the “separation of
3-chloropropylamine and 2-chloropropylamine by distillation
appeared hopeless, because of the instability of the chloropro-
pylamines.” It is evident that the separation of 3-chloropropy-
lamine and 2-chloropropylamine is difficult but not the collec-
tion or extraction of 3-chloropropylamine free base from
commercially available 3-chloropropylamine hydrochloride.
In example 1 of our patent,3 the alkylation was performed
in DMSO, using 3-chloropropylamine free base (3-CPA free
base) obtained by free basing 3-chloropropylamine hydrochlo-
ride (3-CPA HCl). In all three examples in our patent, we have
specifically mentioned that the rapid addition of 3-chloropropyl
amine to the carbanion of the benzofuran derivative (2) is crucial
to the success of the alkylation. Any delay in adding 3-chlo-
ropropyl amine would lead to the failure of intended C-
alkylation. In example 2,3 the alkylation was performed in
acetone. It is known that carbonyl compounds under basic
conditions often undergo condensation reactions. In comparison
to example 1, this transformation was not as smooth and did
result in a low yield.3 Not much discussion was devoted to this
particular example as our aim was to isolate the desired product
instead of the aldol product.
time for
temp (°C) polymerizationb (h)
entry
storage conditionsa
1
2
3
4
5
6
7
8
9
3-CPA + DMSO
3-CPA + DCM
3-CPAc
3-CPA + DMSO
3-CPA + DCM
3-CPAc
3-CPA + DCM + DMSO
3-CPA + DCM
3-CPAc
0-5
0-5
∼36.0
∼36.0
∼18.0
∼12.0
∼12.0
∼8.0
0-5
25-30
25-30
25-30
0-5
30-35
40-45
25-30
25-30
25-30
∼24.0
∼5.0
∼0.25
∼72.0
∼2.0
10 3-CPA + toluene
11 3-CPA + MTBE
12 3-CPA + 1,4-dioxane
∼2.0
a 1.0 g of 3-CPA free base/10 mL of respective solvent. b Polymer is confirmed
based on the appearance of a precipitate. c 3-CPA was obtained after the complete
distillation of dichloromethane.
separated and added directly to the reaction mass containing
the precursor of didesmethylcitalopram (3). This successful
alkylation procedure is based on the following study which was
not discussed in that article.
During the process development of C-alkylation, handling
and storing 3-CPA free base was a major challenge as it is
unstable. Storing 3-CPA for >8 h resulted in the slow ap-
pearance of insoluble solids and white vapor, which could be
due to self-polymerization of 3-chloropropylamine (Scheme 2).
In an attempt to circumvent this problem, we designed a
protocol to study and establish the stability of 3-CPA free base
in different solvents at different temperature conditions (Table
2). Factors which were important to improve the robustness of
3-CPA free base were solvent dilution and temperature. While
3-CPA free base polymerized at 25-45 °C in the absence of
solvent, 3-CPA free base extracted into DCM, MTBE, or
In example 13,3 3-CPA HCl was free based in a biphasic
medium of water and toluene using sodium hydroxide as a base
in order to obtain 3-CPA free base. The organic layer was
(5) Kharasch, M. S.; Fuchs, C. F. J. Org. Chem. 1945, 10, 159–169.
Vol. 13, No. 1, 2009 / Organic Process Research & Development
•
35

Table 3. Extraction of 3-CPA free base from 3-CPA HCl
toluene were relatively stable. Hence, 3-CPA free base diluted
in a nonpolar aprotic solvent such as toluene ameliorated the
self-polymerization reaction, and hence, its use for C-alkylation
reaction was successful.
and its subsequent use for the preparation of 3
HPLC purity6
solvent
for
extraction
3-CPA
addition
modea
yield
of 3
(%)
2.84
RRT
3 (%) impurity
In light of these results, 3-CPA free base was extracted into
either toluene or dichloromethane, and this solution was used
directly for the C-alkylation step. We found that the direct
addition of a solution of 3-CPA free base in dichloromethane
to the benzofuran anion resulted predominantly in the formation
of an unknown impurity (Table 3, entry 2). In contrast, addition
of a solution of 3-CPA in toluene to a solution of the benzofuran
analogue (2) in DMSO/KOtBu led to clean alkylation and
furnished didesmethylcitalopram (3) in 69% yield. The reaction
conditions in terms of addition temperature, addition time of
3-CPA, reaction time, and mole ratio of 3-CPA were studied
and led to our optimized reaction conditions described as
example 13 in our patent.
Several examples from laboratory and pilot-plant prepara-
tions of didesmethylcitalopram (3) are depicted in Table 4. The
results of these experiments represent the successful synthesis
of 3, and hence we can completely rule out the perception2 that
we utilized the hydrochloride salt of 3-chloropropyl amine, as
under these conditions the hydrochloride salt would be highly
insoluble in toluene.
entry
1
2
3
4
DCM
A
B
C
D
50
b
87.19
6.46
-
0.07
80.06
-
DCM
no solvent
toluene
c
69
92.68
ND
a A ) 3-CPA extracted into DCM, concentrated and diluted with DMSO. B )
3-CPA extracted in DCM and directly added to reaction mass. C ) Direct addition
of 3-chloropropylamine hydrochloride to the reaction by dissolving it in DMSO. D
) 3-CPA extracted in toluene and directly added to reaction. b No attempts for
isolation were made, as an impurity formed. c No product formed.
Table 4. Results of experiments carried out under final
experimental conditions for the preparation of 3
mole
ratio
solvent
for
HPLC
KOtBu yield purity6
entry 2 (g) of 3-CPA extraction (equiv) % of 3 of 3
1
2
3
200
500
6000
1.75
1.75
1.75
toluene
toluene
toluene
1.5
1.5
1.5
71.0
72.0
67.0
91.6
94.3
92.1
Second Isolation. To a solution of free base 1 obtained above
(15.0 g, 0.046 mol) in acetonitrile (75.0 mL) was added the
(+)-DPTTA monohydrate (18.7 g, 0.046 mol) solution in
acetonitrile (75.0 mL) and was stirred for 60 min at room
temperature. The mixture was slowly heated to 70-75 °C at
which point methanol (12.0 mL) was added to obtain a clear
solution. The resulting clear solution was slowly cooled to
25-30 °C and stirred for 1.5-2.0 h. The precipitated solid was
separated by filtration, and the resulting filtrate was distilled
off completely to obtain a thick residue. The resulting residue
(12.0 g) was hydrolyzed with 10% aqueous NaOH solution (120
mL), and free base was extracted into toluene (2 × 120 mL).
The combined organic layers were dried over sodium sulfate
and concentrated to afford 5.0 g of free base.
Conclusions
We conclude the following: (a) the resolution of citalopram
(1) is feasible, and the nonreproducibility of the example cited
in our article was due to an inadvertent error during the
compilation of our manuscript; (b) we stand by our findings
on the C-alkylation step: 3-CPA hydrochloride salt was
hydrolyzed in a mixture of toluene-water-sodium hydroxide,
and 3-CPA free base was extracted into toluene for a successful
subsequent C-alkylation if conducted in a rapid manner; and
(c) 3-CPA free base diluted in toluene is sufficiently stable for
these purposes.
Third Isolation. To a solution of the above free base 1 (5.0
g, 0.015 mol) in acetonitrile (25.0 mL) was added the (+)-
DPTTA monohydrate (6.06 g, 0.015 mol) solution in acetonitrile
(25.0 mL) and was stirred for 60 min at room temperature. The
mixture was slowly heated to 70-75 °C at which point
methanol (4.0 mL) was added to obtain a clear solution. The
resulting clear solution was slowly cooled to 25-30 °C and
stirred for 1.5-2.0 h. The precipitated solid was separated by
filtration, and the resulting filtrate was concentrated to afford
6.0 g of S-(+)-1-(+)-DPTTA salt. Yield (%): 11.0 (calculated
relative to theoretical which is half of the starting racemate);
chiral purity:4 96.4%.
Preparation of 1-(3-Aminopropyl)-1-(4-fluorophenyl)-1,3-
dihydro-2-benzofuran-5-carbonitrile (3). (a) Preparation of
3-Chloropropan-1-amine Solution in Toluene. A mixture of
aqueous NaOH solution (182 g in 1100 mL of water), 3-CPA
hydrochloride (470 g, 3.61 mol), and toluene (2750 mL) were
stirred for 10-15 min, and the organic layer was separated.
The resulting aqueous layer was re-extracted twice with toluene
(2 × 1100 mL) to ensure the complete extraction of 3-CPA
Experimental Section
Preparation of S-(+)-1-(3-Dimethylamino-propyl)-1-(4-
fluorophenyl)-1,3-dihydro-isobenzofuran-5-carbonitrile [S-(+)-
1-(+)-DPTTA]. First Isolation. To a stirred solution of 1 (50
g, 0.154 mol) in acetonitrile (250 mL) was added a (+)-DPTTA
monohydrate (62.8 g, 0.14 mol) solution in acetonitrile (250
mL) and was stirred for 60 min at room temperature. The
mixture was slowly heated to 70-75 °C, and methanol (40 mL)
was added to obtain a clear solution at 70-75 °C. The resulting
clear solution was slowly cooled to 25-30 °C and stirred for
1.5 -2.0 h. The precipitated solid was separated by filtration,
and the resulting filtrate was distilled off completely to obtain
a thick residue (23.0 g). The enantiomer ratios of filtered solid
(70.0 g) and residue (23.0 g) obtained through stripping the
filtrates were found to be 45.73:54.25 and 59.33:40.67 for the
S and R isomers, respectively. The resulting residue (23 g) was
free based with 10% aqueous NaOH solution (230 mL), and
the free base was extracted into toluene (2 × 230 mL). The
combined organic layers were dried over sodium sulfate and
concentrated to afford 15.0 g of free base.
36
•
Vol. 13, No. 1, 2009 / Organic Process Research & Development

free base (TLC). The combined organic layers were used as
such in the alkylation step as soon as possible.
separated, and the resulting aqueous layer was extracted with
toluene (3 × 1100 mL). The combined organic layers were
extracted with aqueous HCl solution (240 mL of concentrated
HCl in 5000 mL of water) and washed with toluene (4 × 2500
mL). The pH of the aqueous layer was adjusted to 10.5 with
aqueous NaOH solution (150 mL), and the mixture was
extracted with toluene (3 × 2500 mL). The resulting toluene
layer was washed with hot water (3 × 1500 mL) and
concentrated to obtain 447.0 g of 3.
(b) Coupling. Dimethylsulfoxide (2500 mL) and potassium-
tert-butoxide (355 g, 3.16 mol) were heated at 65-70 °C for
60 min, and a solution of 2 (500 g, 2.09 mol) in DMSO (1500
mL) was added to the flask over 10-15 min at 30-35 °C. To
the resulting purple colored solution was added in over 10 min
the solution of 3-CPA free base in toluene prepared above, and
the resulting mixture was maintained at 40-45 °C for 1 h. The
reaction mixture was cooled to 25-35 °C and slowly quenched
into cold water (5.0 L) at 10-15 °C. The organic layer was
1
Yield: 72%; HPLC purity:6 94.3%; H NMR (400 MHz,
DMSO-d6) of (-)-DPTTA salt of (()-3: δ 1.31-1.44 (m, 2H),
2.18-2.34 (m, 2H), 2.34 (s, 6H), 2.72 (t, J ) 7.6 2H), 5.12 (q,
J ) 3.2, 2H), 5.59 (s, 2H), 7.14 (t, J ) 8.8, 21H), 7.26 (d, J )
8.0, 4H), 7.56 (d, 2H), 7.71-7.81 (m, 4H); MS (APCI) m/z
297 (M+ + 1).
(6) Purity of 3 was estimated by HPLC analysis with Inertsil ODS 3V,
250 mm × 4.6 mm, 5 µ; mobile phase A: mixture of buffer and
acetonitrile in the ratio of 800:200; mobile phase B: mixture of
acetonitrile and buffer in the ratio of 700:300; run time: 60 min, 240
nm. Buffer preparation: to 1000 mL of water was added 0.5 mL of
phosphoric acid, and the pH was adjusted to 3.5 with triethylamine.
Gradient profile
Acknowledgment
time (min)
flow
A (%)
B (%)
We thank the management of Dr. Reddy’s Laboratories Ltd.
for supporting this work. Cooperation from the project col-
leagues is highly appreciated. We thank Dancer and Lopez De
Diego for their insightful experiments. We are grateful to Dr.
Trevor Laird and Dr. Jaan Pesti for providing us an opportunity
to revisit our work and respond to the comments made by Dr.
Dancer.
1
2
3
4
5
6
7
8
0.01
5.00
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
100
90
90
80
70
50
40
30
20
0.0
10.0
10.0
20.0
30.0
50.0
60.0
70.0
80.0
80.0
0.0
10.00
15.00
20.00
25.00
30.00
35.00
40.00
45.00
50.00
60.00
9
10
11
12
20
100
100
Received for review August 25, 2008.
OP8002079
0.0
Vol. 13, No. 1, 2009 / Organic Process Research & Development
•
37