Utility of Japp-Klingemann reaction for the preparation of 5-carboxy-6-chloroindole via Fischer indole protocol

Article abstract of DOI:10.1016/j.tetlet.2007.01.148

5-Carboxy-6-chloroindole, a precursor for p38 kinase inhibitor, was prepared from 4-amino-2-chloro-3-iodobenzoicacid by following the Japp-Klingemann synthetic approach. The structures of the key intermediates were also confirmed by X-ray analyses. Computational analysis was helpful in understanding the importance of the substituents at the cyclization step of the synthesis.

Full text of DOI:10.1016/j.tetlet.2007.01.148

Tetrahedron Letters 48 (2007) 2353–2356
Utility of Japp–Klingemann reaction for the preparation of
5-carboxy-6-chloroindole via Fischer indole protocol
Yihui Chen,a Masayuki Shibata,b Manju Rajeswaran,c Thamarapu Srikrishnan,d
Sundeep Dugare and Ravindra K. Pandeya,*
aChemistry Division, PDT Centre, Roswell Park Cancer Institute, Buffalo, NY 14263, USA
bBiomedical Informatics, UMDNJ-School of Health Related Professions, Department of Health Informatics, Newark, NJ 07107, USA
cKodak Research and Development Laboratories, Eastman Kodak Company, Rochester, NY 14650-2106, USA
dTumor Biology, Roswell Park Cancer Institute, Buffalo, USA
eScios Pharmaceuticals, a subsidiary of Johnson and Johnson, Palo Alto, CA, USA
Received 26 December 2006; revised 23 January 2007; accepted 24 January 2007
Available online 1 February 2007
Abstract—5-Carboxy-6-chloroindole, a precursor for p38 kinase inhibitor, was prepared from 4-amino-2-chloro-3-iodobenzoicacid
by following the Japp–Klingemann synthetic approach. The structures of the key intermediates were also confirmed by X-ray anal-
yses. Computational analysis was helpful in understanding the importance of the substituents at the cyclization step of the synthesis.
Ó 2007 Elsevier Ltd. All rights reserved.
4
The indole nucleus is probably the most widely distrib-
uted heterocyclic ring system found in nature.1 Due to
the existence of a vast array of structurally diverse and
biologically active indoles, it is not surprising that the in-
dole nucleus is an important feature in many medicinal
agents. The synthesis and reactivity of indole derivatives
have been a topic of research interest for well over a cen-
tury. Reports of several thousand individual indole
derivatives appear annually in the chemical literature.
The primary reason for this sustained interest is the wide
range of biological activities found among indole ana-
logs.1 One of the most important biological activities
of indole-based compounds is that these compounds
can be used as inhibitors of p38 kinase.2 These inhibitors
are important in treating a variety of diseases: rheuma-
toid arthritis, osteoarthritis, sepsis, asthma, adult respi-
ratory distress syndrome, cardiovascular disorders, and
Alzheimer’s disease. Among the indoles synthesized so
far, compound 1 (Fig. 1) has also shown some promising
p38 kinase inhibitory activity.2
HO2C
3
5
2
1
6
NH
Cl
7
1
Figure 1. 6-Chloro-5-indole carboxylic acid.
o-Amino- or o-nitro-phenylacetylenes can be cyclized
to indoles by endo-dig addition.5 Reductive cyclization
of o-nitrostyrenes to indoles has also been reported.6
The Nenitzescu reaction is another approach for indole
synthesis.7 The Watanabe indole synthesis deals with an
oxidative cyclization.8 It is the metal-catalyzed indole
synthesis from anilines and glycols or ethanolamines
with subsequent intramolecular cyclization of o-amino-
phenethyl alcohols to indole.9 For all the metal-cata-
lyzed indole synthesis approaches, palladium is the
most widely used catalyst. The review articles by Gribble
and Wolfe provide a detailed description on Pd medi-
ated cyclization.9,10 Other metals reported to catalyze
indole synthesis include rhodium, ruthenium, titanium,
zirconium, and copper.9 The selection of the metal
depends on the nature of the substrate(s).
In recent years, several synthetic methodologies have
been reported for preparing indole analogs. For exam-
ple, the Leimgruber–Batcho approach starts from the
b-dialkylamino-o-nitrostyrene3 or o-, b-nitrostyrene.4
Among the synthetic methodologies reported so far for
the preparation of the indole analogs, the Fischer indole
synthesis still maintains its prominent role for large-
scale production.9 In particular, the Japp–Klingemann
*
Corresponding author. Tel.: +1 716 845 3203; fax: +1 716 845
8920; e-mail: ravindra.pandey@roswellpark.org
0040-4039/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2007.01.148

2354
Y. Chen et al. / Tetrahedron Letters 48 (2007) 2353–2356
reaction11,12 is quite efficient and deals with the prepara-
tion of intermediate hydrazones from the aryl diazo-
nium salts derived from arylamines and b-keto-acids
(or b-keto-esters). We were interested to investigate the
utility of this approach for the synthesis of the desired
5-carboxy-6-chloroindole. The amino group in 3 was
converted into the corresponding aryldiazonium analog
4. It was treated with ethyl 2-methyl-acetoacetate 5 un-
der basic conditions to produce intermediate 6, which on
subsequent acid rearrangement gave indole 8 (instead of
9), formed by the intramolecular cyclization at the
ortho-position of the chlorine atom. The formation of
8 without producing indole 9 even in traces was surpris-
ing. If steric hindrance is considered to be a determining
factor for the cyclization of intermediate 4, then com-
pared to the ortho-position of the chlorine atom, the
para-position is less sterically hindered, and one can ex-
pect that the cyclization would occur at the less hindered
para-position. However, the cyclization in fact occurred
at the more sterically hindered position (ortho-position).
This observation suggests that at the cyclization step the
steric hindrance is not a dictating factor (Scheme 1).
Figure 2. Single crystal X-ray structure of intermediate 7.
Among all the intermediates, the unstable intermediate
12 seems to be responsible for the formation of indole
8 over 9. On the basis of the charge distribution study
on compound 12 obtained with ab initio MO at MP2
level using the 6-31+G* basis set, the position 5 was found
to be more electron dense (À0.32) than position 3
(À0.24) (Fig. 3). This characteristic could be responsible
for the formation of indole 8. In order to further confirm
the rationality of the charge distribution at the intramo-
lecular cyclization step, the chloro-substituent in com-
pound 12 was replaced with a hydroxy group and the
electron density of the possible intermediate 17 was cal-
culated at the same level as chloro derivative discussed
above. As expected, in contrast to compound 12, inter-
mediate 17 had a higher electron density at position 3
(À0.46) than at position 5 (À0.34). If charge density
had been the determining factor the cyclization should
have preferentially occurred at position 5 (instead of po-
sition 3) and indeed the position 5 cyclized indole com-
pound 16 was isolated as a sole product (see Scheme 3).
The other possibility that we looked at was the energy
level for the formation of both indoles 8 and 9 calculated
by semi-empirical MO, PM3, as well as more sophisti-
cated ab initio RHF/6-31G**. However, the calculated
values for the formation of both the indoles were quite
close and did not help in determining the factors respon-
sible for generating the undesired indole (see Supplemen-
tary data for details). In another set of experiments,
intermediate 7 was isolated and characterized by X-ray
analysis before subjecting it to cyclization. As can be seen
from Figure 2, in compound 7, the acetyl group that was
present in 6 is gone and the nitrogen–nitrogen bond is
now linked with a single bond, while the adjacent nitro-
gen–carbon single bond was converted into a double
bond. A sigmatropic rearrangement of 7 under acid-cat-
alyzed conditions afforded compound 8 as a sole product
(Scheme 2).
In another approach, compound 19 in which position 3
was blocked by an iodo-group was first synthesized. It
was then subjected to the reaction sequences depicted
in Scheme 4 and the desired indole 1 was obtained in
a moderate yield.
H3COOC
HOOC
Cl
H3COOC
0.5% acetyl
chloride
NH2
Cl
NaNO2/HCl
Cl
NH2
-
+
Cl
Cl
Cl
N
N
H3CO2C
H
2
2
4
H
3
H3CO2C
H
H
3
H
N
H
H3COOC
1
6
+
4
3
+
C
C
H
C
C
+
5
5
6
2
4
N
N
H
N
H
B1
NH
CO2C2H5
A
CO2C2H5
CO2C2H5
H
NaOEt
Cl
7
10
CH3
7
9
H
CH3CO-CH-CO2C2H5
H
Cl
Cl
5
H
2
H3CO2C
-H+
C
C
H3CO2C
+ 2H+
H
H3COOC
1
6
3
C
C
+
+
5
CO2C2H5
CH3
4
N
H
N
H
N
N
?
CO2C2H5
H2
H2
HCl
Cl
Cl
N=N C CO2C2H5
CH3 C=O
11
7
6
12
Cl
H3CO2C
H3COOC
4
3
CH
5
6
2
B
CO2C2H5
A
1
NH
CO2C2H5
N
H
7
8
8
Scheme 1. Preparation of indole 8 from 2-chloro-4-aminobenzoic acid
2.
Scheme 2. A possible mechanism for the formation of indole 8 from
intermediate 7.

Y. Chen et al. / Tetrahedron Letters 48 (2007) 2353–2356
2355
H3CO2C
H3CO2C
ICl/Acetic Acid
I
H3CO2C
Cl
+
Cl
NH2
Cl
NH2
NH2
I
18
3
19
H3CO2C
H3CO2C
I
H3CO2C
Cl
I
I
NaNO2
HCl
H2O
-
Cl
+
N
Cl
NH
NH2
Cl
N
NCl
21
18
20
OH
H3CO2C
5 NaOEt
I
23
intermediate
CO2C2H5
HCl
CH3
20
Cl
NH
N
C
22
Cl
2
H
H
H3CO2C
C
C
1
3
HCl
Figure 3. Charge distribution of compound 12.
18
6
CO2C2H5
5
+
+
4
N
N
H2
H2
I
H3CO2C
0.5% acetyl
chloride in MeOH
23
H3CO2C
NaNO2/HCl
HO2C
OH
Cl
H3CO2C
+
N
-
Cl
N
NH2
NH2
OH
CO2C2H5
OH
14
15
13
N H
CH3
I
24
NaOEt
CH3-CO-CH-CO2C2H5
5
HCl
H
H3CO2C
NaNO2
H3CO2C
HCl
NaOEt
5
CO2C2H5
19
H
-
Cl
HCl
+
N H
6
H3CO2C
Cl
C
C
Cl
N
N
H3CO2C
3
5
4
I
1
2
2
5
I
B
A
6
26
1
+
+
7
CO2C2H5
4
25
N H
N
N
CO2C2H5
OH
3
OH
H2/Raney Ni
H2
H2
16
17, the intermediate
H3CO2C
NaOH
Cl
HO2C
Cl
HCl
CO2C2H5
Scheme 3. Preparation of indole 16.
N H
N H
1
9
For the preparation of iodinated derivative 19, com-
pound 3 was reacted with periodic acid and iodine,13
which resulted in the formation of an undesired 5-iodo-
substituted analog 18 as a sole product. However, reac-
tion of 3 with iodine monochloride (ICl)14 in acetic acid
as the solvent gave a mixture of 3- and 5-iodinated
derivatives 18 and 19, respectively. The yields of the
individual isomers were found to be temperature depen-
dent. For example, at high temperature (>100 °C) 18
was isolated as a sole product, whereas at room temper-
ature, both 18 and 19 were obtained, in the ratio of 2.5
to 1. Due to their close Rf values (similar lipophilicity), it
was extremely difficult to separate the individual isomers
in a reasonably large quantity by column chromatogra-
phy. Although the Rf values of 18 and 19 were quite
close, their NMR spectra were significantly different
and it was easy to characterize the individual isomers.
For compounds 18 and 19, the resonance peak of amino
group (–NH2) was observed at 4.563 ppm and
4.743 ppm, respectively. The chemical shifts for the
two aromatic protons were also quite different. For 18,
the resonances were observed at 8.231 and 6.733 ppm
as singlets. On the other hand, for 19 these resonances
as doublets were observed at 7.692 and 6.611 ppm,
respectively, both protons ortho- to each other showed
a strong interaction. The observed coupling constant
value was 8.4 Hz. Both isomers on subjecting to cycliza-
tion gave some interesting results. For example, isomer
Scheme 4. Synthesis of indole 1.
18 did not produce the expected indole 24 and the start-
ing material was recovered quantitatively. However, the
X-ray structure of intermediate 22 (Fig. 4) was found to
be similar to intermediate 7 (Fig. 2). These results sug-
gest that at the cyclization stage, due to a high steric hin-
drance caused by the substituents at positions 2 and 5
(Cl–, I–, respectively), the intermediate compound 23
was unable to cyclize and under acidic conditions
decomposed to the starting material 18.
After careful purification of all the compounds before
HCl treatment, besides compound 22 a small amount
of compound 21 was also isolated, which was possibly
formed from 20 under aqueous reaction conditions. By
following a similar synthetic approach, isomer 19 was
successfully converted to indole 26 (Scheme 4). This re-
sult introduces another possibility that in addition to the
charge distribution of the key intermediate discussed
above, other factors such as HOMO/LUMO character-
istics may be playing a significant role at the cyclization
step of the synthesis. For example, a HOMO of com-
pound 12 clearly showed that the C3 and C5 positions
in the ring have an opposite polarity and a LUMO of
the same compound shows a large population on the
ethylene carbon (C@C) (see the Supplementary data

2356
Y. Chen et al. / Tetrahedron Letters 48 (2007) 2353–2356
role but other factors such as conformational flexibility
and HOMO/LUMO may play additional roles at the
cyclization step of the synthesis.
Acknowledgment
Support from the shared resources of the RPCI
(P30CA16056) is appreciated.
Supplementary data
Experimental procedures, characterization data of com-
pounds 1, 3, 7–9, 14, 16, 18, 19, 21, 22, and 26. Crystal-
lographic file (cif) for compounds 7 and 22. Molecular
modeling calculations. Supplementary data associated
with this article can be found, in the online version, at
doi:10.1016/j.tetlet.2007.01.148.
Figure 4. Single X-ray structure of compound 22.
for HOMO/LUMO representation). Limited conforma-
tional flexibility through three bonds connecting the eth-
ylene carbon and the ring may prevent the ethylene
carbon for approaching the ring in the right orientation
for cyclization to occur at C5 position. This provides a
reasonable explanation for the quantitative recovery of
the starting material 18, without forming any cyclized
product.
References and notes
1. Kuethe, J. T.; Wong, A.; Qu, C.; Smitrovich, J.; Davies,
I. W.; Hughes, D. L. J. Org. Chem. 2005, 70, 504, and
references cited therein.
2. Luedtke, G.; Tester R.; Dugar, S.; Lu, Q.; Perumattam, J.;
Tan, X. United States Patent: US 200501711 83A1, 2005.
3. Clark, R. D.; Repke, D. B. Heterocycles 1984, 22, 195.
4. Sinhababu, A. K.; Borchardt, R. T. J. Org. Chem. 1983,
48, 3347.
5. Rudisill, D. E.; Stille, J. K. J. Org. Chem. 1989, 54, 5856.
6. Akazome, M.; Kondo, T.; Watanabe, Y. J. Org. Chem.
1994, 59, 3375.
7. Landwehr, J.; Troschutz, R. Synthesis 2005, 14, 2414.
8. Shim, S. C.; Youn, Y. Z.; Lee, D. Y.; Kim, T. J.; Cho,
C. S.; Uemura, S.; Watanabe, Y. Synth. Commun. 1996,
26, 1349.
9. Gribble, G. W. J. Chem. Soc., Perkin Trans. 1 2000, 1045.
10. Wolfe, J. P.; Thomas, J. S. Curr. Org. Chem. 2005, 9, 625.
11. Phillips, R. R. Org. React. 1959, 10, 143.
12. Gorlitzer, K.; Fabian, J.; Frohberg, P.; Drutkowski, G.
Pharmazia 2002, 57, 243.
13. Panetta, C. A.; Fang, Z.; Mattern, D. L. J. Org. Chem.
1995, 60, 7953.
14. Hubig, S. M.; Jung, W.; Kochi, J. K. J. Org. Chem. 1994,
59, 6233.
For a large-scale preparation of indole 1, we started with
a mixture of iodo-analogs 18 and 19. Two major
products isolated from the reaction mixture were
characterized as the desired intermediate indole 26
and an unreacted isomer 18. From the mixture, these
compounds can easily be separated by column chromato-
graphy. Our next step was to selectively remove the
iodo-group on 26 without removing the chloro substitu-
ent. Several approaches were tried to produce the
desired compound and finally the hydrogenation with
Raney nickel afforded 9 in a quantitative yield. At the
final step of the synthesis, the ester functionalities were
hydrolyzed on stirring in aqueous sodium hydroxide
solution. The corresponding carboxylic acid on stirring
in hot HCl preferentially decarboxylated the 2-carbox-
ylic acid functionality and the desired indole 1 was
obtained in 60% yield.15
15. Experimental: In brief, compound 2 was converted into 3
in 90% yield by refluxing in methanol with catalytic
In conclusion, starting from the readily available start-
ing materials, the preparation of the desired 5-car-
boxy-6-chloroindole, a useful precursor for p38 kinase
inhibitor, is discussed. Our study indicates that in
Japp–Klingemann reaction, the position and nature of
the substituents in the intermediate product(s) play an
important role at the cyclization step of the synthesis.
However, in our present study the reaction conditions
for the preparation of the intermediates and the final
products were not optimized.
amount of acetyl chloride. Compound
3 was then
dissolved in acetic acid and reacted with iodine mono-
chloride to afford a mixture of compounds 18 and 19 (the
ratio of 18 to 19 was roughly 2.5 to 1) in 75% yield. At low
temperature (<5 °C), compound 25 obtained by reacting
18 and 19 with sodium nitrite was treated with the sodium
ethyl 2-methylacetoacetate (obtained by reacting 5 with
sodium ethoxide). The resulting intermediate was reacted
with HCl gas-saturated ethanol to give 26 in 19% yield,
which on hydrogenolysis (Raney nickle/H2) produced 9 in
quantitative yield. At the final step of the synthesis, 9 was
reacted with sodium hydroxide and the resulting carbox-
ylic acid analog without any further purification was
decarboxylated under acidic conditions (concd HCl) to
produce the desired indole 1 in 60% overall yield (for
details see the ‘Supplementary data’).
The structures of the key synthetic intermediates were
also confirmed by X-ray analysis. The computer model-
ing study suggested that the distribution of the electron
density in the intermediate molecules plays an important