Utilization of lithium amide in the synthesis of N-arylanthranilic acids and N-arylanthranilamides

Article abstract of DOI:10.1021/op0501242

A procedure for the preparation of N-arylanthranilic acids and N-arylanthranilamides was developed. Lithium amide promoted coupling of anilines with 2-fluorobenzoic acids or 2-fluorobenzamides lead to the desired compounds in good yield. Both primary and

Full text of DOI:10.1021/op0501242

Organic Process Research & Development 2005, 9, 843−846
Utilization of Lithium Amide in the Synthesis of N-Arylanthranilic Acids and
N-Arylanthranilamides
Edward M. Davis, Thomas N. Nanninga,* Howie I. Tjiong, and Derick D. Winkle
Pfizer Chemical Research and DeVelopment, 7000 Portage Road, Kalamazoo, MI 49001, U.S.A.
Scheme 1. Reaction scheme
Abstract:
A procedure for the preparation of N-arylanthranilic acids and
N-arylanthranilamides was developed. Lithium amide promoted
coupling of anilines with 2-fluorobenzoic acids or 2-fluoroben-
zamides lead to the desired compounds in good yield. Both
primary and N-substituted anilines were effective partners for
the reaction. In the reaction of 2,3,4-trifluorobenzoic acid with
various anilines, selectivity for ortho substitution was observed
exclusively.
Anthranilic acids and their derivatives represent an
important class of compounds in the pharmaceutical industry.
These compounds have been marketed as anti-inflammatory
agents for the treatment of dysmenorrhea and rheumatic
pain.1 Recently, N-arylanthranilic acids have been studied
for the treatment of cancer2 and Alzheimer’s disease.3 A
research program in our labs required a reproducible,
efficient, and scaleable approach to the synthesis of N-
arylanthranilic acids and their derivatives.
Traditionally, N-arylanthranilic acids have been prepared
by the copper mediated coupling of anilines with haloben-
zenes, a process known as the Ullmann reaction.4 Unfortu-
nately, the success of these reactions often requires strongly
activated aryl halides and forcing reaction conditions.
Furthermore, isolation of the products often requires tedious
steps to eliminate the significant quantities of copper
byproducts and adequately purify the product. Recently,
advances in homogeneous catalytic processes have led to
the development of catalytic N-arylation reactions,5 which
proceed with activated and unactivated aryl iodides, bro-
mides, and chlorides and utilize mild reaction conditions.6
Triarylbismuth reagents have also been successfully em-
ployed in the preparation of N-arylanthranilates as well as
other diarylamines.7 A catalyst-free substitution of 2-fluo-
robenzoic acids with lithium anilides prepared from the
corresponding anilines and lithium hexamethyldisilazide at
-78 °C has been demonstrated.8 While this catalyst-free
methodology was very attractive, the cryogenic reaction
conditions and large quantities of waste complicate the scale-
up of such a process.
We investigated the use of the strong base LDA to
promote the reaction of 2-chloro-4-iodoaniline (1a) with
2,3,4-trifluorobenzoic acid (2a). It was found that by adding
portions of LDA while the reaction was less than -20 °C
and warming the reaction mixture to room temperature
between additions, the reaction proceeded in high yield.8c
The portions of LDA added were in decreasing molar
equivalency amounts as such: 2, 0.5, 0.25, 0.13, 0.06, 0.03,
and 0.01 equiv. The addition of 3 equiv of LDA “all at once”
resulted in significant formation of undesired byproducts,
presumably due to the formation of benzyne intermediates.
A 3 equiv amount of base was required for the reaction to
go to completion since the product is rapidly deprotonated
under the reaction conditions (Scheme 1).
* Corresponding author. E-mail: Thomas.Nanninga@pfizer.com.
(1) (a) Scherrer, R. A.; Short, F. W. U.S. Patent 3,313,848, 1967.
(2) (a) Sebolt-Leopold, J. S.; Dudley, D. T.; Herrera, R.; Van Becelaere, K.;
Wiland, A.; Gowen, R. C.; Tecle, H.; Barrett, S. D.; Bridges, A.;
Przybranowski, S.; Leopold, W. R.; Saltiel, A. R. Nature Medicine; New
York, 1999; Vol. 5, pp 810-816. (b) Barrett, S. D.; Bridges, A. J.; Doherty,
A. M.; Dudley, D. T.; Saltiel, A. R.; Tecle, H. PCT Int. Appl. WO
199901426. (c) Barrett, S. D.; Kaufman, M. D.; Milbank, J. B.; Rewcastle,
G. W.; Spicer, J. A.; Tecle, H. PCT Int. Appl. WO 2003062191.
(3) Augelli-Szafran, C. E.; Barvian, M. R.; Bigge, C. F.; Glase, S. A.; Hachiya,
S.; Keily, J. S.; Kimura, T.; Lai, Y.; Sakkab, A. T.; Suto, M. J.; Walker,
L. C.; Yasunaga, T.; Zhuang, N. PCT Int. Appl. WO 200076489.
(4) (a) Bunnett, J. F.; Zahler, R. E. Chem. ReV. 1951, 49, 273-412. (b) Paine,
A. J. J. Am. Chem. Soc. 1987, 109, 1496-1502.
We desired to have a simpler method with higher
throughput and less waste; thus we investigated the use of
alternative bases (Table 1). Both lithium hydride and lithium
amide were found to give good results, but we focused on
the use of lithium amide as it seemed safer to use and resulted
in higher conversion to product.9 Using 2 equiv of base gave
(5) Klapars, A.; Huang, X.; Buchwald, S. L. J. Am. Chem. Soc. 2002, 124,
7421-7428. Harris, M. C.; Huang, X.; Buchwald, S. L. Org. Lett. 2002,
4, 2885-2888. Kuwano, R.; Utsunomiya, M.; Hartwig, J. F. J. Org. Chem.
2002, 67, 6479-6486.
(6) For reviews see: (a) de Meijere, A.; Diederich, F., Eds. Metal Catalyzed
Cross-Coupling Reactions, 2nd ed. VCH: 2004. (b) Jiang, L.; Buchwald,
S. L. Vol. 2, Chapter 13, pp 699-760 in ref 6a and references therein. (c)
Schlummer, B.; Scholz, U. AdV. Synth. Catal. 2004, 346, 1599-1626 and
references therein.
(7) (a) Sorenson, R. J. J. Org. Chem. 2000, 65, 7747-7749. (b) Combes, S.;
Finet, J.-P. Tetrahedron 1998, 4313-4318.
(8) Chen, M. H.; Beylin, V. G.; Iakovleva, E.; Kesten, S. J.; Magano, J.; Vrieze,
D. Synth. Comm. 2002, 32, 411-417. (b) Chen, M. H.; Magano, J.; U.S.
Patent 6,686,499, Feb. 3, 2004. (c) Chen, M. H.; Davis, E. M.; Magano,
J.; Nanninga, T. N.; Winkle, D. D. WO 2002018319 A1.
(9) (a) Glamkowski, E. J.; Chiang, Y. J. Heterocycl. Chem. 1987, 24, 1599-
1604. (b) Kim, D. H. J. Heterocycl. Chem. 1981, 18, 287-291.
10.1021/op0501242 CCC: $30.25 © 2005 American Chemical Society
Published on Web 09/30/2005
Vol. 9, No. 6, 2005 / Organic Process Research & Development
•
843

Table 1. Base and solvent screen for the coupling of
2-chloro-4-iodoaniline (1a) with 2,3,4-trifluorobenzoic
acid (2a)
Table 3. Coupling reactions of N-substituted anilines with
2-fluorobenzoic acids in THF
time yielda
entry
R
R′
CH3
CH3
CH3
X
product (h)
(%)
temp
(°C)
24 h
entry
basea
solvent
conversionb (%)
1
2
3
4
5
6
7
8
H
H
H
H
H
F
3g
3g
3g
3h
3i
3j
3k
3l
3.5 87
Cl
Br
F
F
F
F
F
72
48
49
67
1
2
3
4
5
6
7
8
9
KOH
LiOtBu
LiOtBu
LiH
DMFc
DMFc
DMFc
DMAcc
MeCN
THF
MeCN
THF
THF
50
20
20
80
70
66
70
50
50
NR
7
sCH2Ph
Ph
4.5 77
19d
82
80
35
88
56e
95
72
68
0
sCHdCHs
sCH2CH2s
sCH2CH2CH2s
LiH
4
79
1.5 81b
LiH
LiNH2
LiNH2
LiNH2
a Yields are of pure products following crystallization. b LiNH2 (3.1 equiv)
was used.
a All reactions were run using 3.5 equiv of base unless otherwise noted.
b Percent conversion was determined using reversed-phase HPLC with UV
detection. c Care should be taken when using strong bases in polar aprotic solvents
as thermal instability has been reported in such systems. d Diisopropylamine (1
equiv) was used as an additive. e Base (2 equiv) was used.
Table 4. Coupling reactions of anilines with
2-fluorobenzamides in THF
Table 2. Coupling reactions of primary anilines with
2-fluorobenzoic acids in THF
time yielda
entry
R
R1
R2
X
product (h)
(%)
1
2
3
4
5
CH3
H
H
sCH2CH2s
CH3
H
6-iPr
4-OCH3
H
iPr
H
H
H
H
H
H
H
F
5a
5b
5c
5d
5e
3
24
6
2
4
75
72b
74
62
54
entry
R
R′
X
product time (h) yield (%)a
H
1
2
3
4
5
6
Cl
Cl
iPr
H
CF3
Cl
4-I
H
F
3a
3b
3c
3d
3e
3f
4
22
23
1
92
75
85
83
0
H
H
H
H
F
a Yields are of pure products following crystallization. Unless otherwise noted,
3.1 equiv of lithium amide was used for reactions with primary anilines, and
2.1 equiv of lithium amide was used for reactions with N-substituted anilines.
b Base (4 equiv) were used.
6-iPr
4-OCH3
H
H
1
76
a Yields are of pure products following crystallization.
conditions (Table 3, entry 6). Diphenylamine was found to
react under these conditions resulting in a good yield of N,N-
diphenylanthranilic acid (Table 3, entry 5).
only a 56% conversion to product (Table 1, entry 8) as in
the case of using LDA.
This methodology was extended to the use of 2-fluo-
robenzamide as the coupling partner, providing direct access
to N-arylanthranilamides (Table 4). While there are numerous
reports in the literature of copper mediated coupling of
anilines with halobenzoic acids, there have only been a few
reported cases of coupling anilines with halobenzamides.9
Both primary and N-substituted anilines were found to
participate in the reaction, providing moderate to good yields
of the desired products. The reactions using N-substituted
anilines as the reactant were found to be faster than reactions
using primary anilines. This is likely due to the increased
electron density resulting from the alkyl substituent and
suggests that the rate-determining step is the coupling of
lithio anilide with the 2-fluorobenzoic acid salt, not the
deprotonation of the aniline. This is consistent with the
unreactive nature of deprotonated 2-trifluoromethylaniline
and deprotonated indole in the reaction mentioned above.
Primary anilines were found to participate well in the
reaction with 2-fluorobenzoic acid or 2,3,4-trifluorobenzoic
acid (Table 2). Even sterically hindered 2,6-diisopropylaniline
reacted with 2-fluorobenzoic acid producing product 3c in
high yield, although requiring a 23 h reaction time (Table 2,
entry 3). 2-Trifluoromethylaniline was found to be unreactive
under these conditions, possibly due to the strongly electron-
withdrawing nature of the trifluoromethyl substituent (Table
2, entry 5).
N-Substituted anilines were found to react nicely with
2-fluorobenzoic acid and 2.1 equiv of lithium amide provid-
ing the trisubstituted aniline products (Table 3). Utilization
of either 2-chlorobenzoic acid or 2-bromobenzoic acid
required longer reaction times and resulted in diminished
yields. Tetrahydroquiniline and indoline formed the desired
anthranilic acids, while indole did not react under these
844
•
Vol. 9, No. 6, 2005 / Organic Process Research & Development

Scheme 2. Comparison of reactivities of 2-fluorobenzoic
acid and 4-fluorobenzoic acid
irritant. MSDSs list lithium amide powder as a white/grey
powder with a melting point of 380-400 °C and decomposi-
tion above 400 °C. It has unlimited shelf life if stored dry
and inerted, but exposure to air could result in the formation
of dangerous peroxides. Colored (yellow) material should
be discarded in a safe manner by technically qualified
personnel. For pilot plant use the material was purchased in
1 kg bags which were opened and emptied inside an inerted
reaction vessel. Lithium amide or hydride powder was
purchased from Aldrich, FMC, and Chemetall.
Scheme 3. Comparison of regioselectivities using lithium
amide or sodium amide
Commercially available solvents and reagents were used
as received without further purification. All reactions were
performed under a nitrogen atmosphere. Gas evolution was
1
noted by use of an oil bubbler or flow meter. H and 13C
spectra were determined using a Varian INOVA or Brucker
Ultrashield 400 spectrometer. HPLC analysis was per-
formed using a YMC ODS AQ RP18 column, 250 mm ×
4.5 mm, 5 µ eluting with a 25/75 water/acetonitrile mixture
containing 0.1% formic acid at 1.0 mL/min and monitoring
at 220 nm.
2-(Indolin-1-yl)benzoic Acid (3k): To a 250 mL flask
were added indoline (5 g, 42.0 mmol), 2-fluorobenzoic acid
(6.2 g, 44.1 mmol), and THF (140 mL). To this solution
was added lithium amide (2.0 g, 88.2 mmol) in two portions
over 5 min. This mixture was heated to 50 °C under nitrogen
for about 4 h and cooled to room temperature. The reaction
was quenched with water (25 mL), concentrated HCl (10
mL), and MTBE (25 mL). The aqueous layer was removed,
and the organic layer was washed with water (25 mL).
Following removal of solvent in vacuo, the yellow solid was
dissolved in IPA (40 mL) at 65 °C, and water (48 mL) was
added slowly. Following slow cooling to 3 °C, the product
was filtered and washed with 40% IPA in water (2 × 10
mL). The off-white solid was dried in a vacuum oven at
∼50 °C. A 7.9 g (79% yield) amount of product 3k was
obtained. 1H NMR (400 MHz, CDCl3) δ 8.41 (d, 1H), 7.66
(t, 1H), 7.51 (t, 1H), 7.33 (t, 2H), 7.14 (t, 1H), 7.05 (t, 1H),
6.47(d, 1H), 3.29 (m, 4H); 13C NMR (100 MHz, CDCl3)
166.2, 149.5, 147.8, 134.9, 132.4, 131.9, 128.2, 127.8, 127.2,
125.9, 125.1, 123.3, 113.4, 59.3, 29.3; HPLC: 99.5% (by
area %).
2-(2-Chloro-4-iodophenylamino)-3,4-difluorobenzoic
Acid (3a).12 To an inerted 400 L Hastelloy vessel were added
13 kg (51.3 mol) of 2-chloro-4-iodoaniline (CIA), 10 kg (56.8
mol) of 2,3,4-triflurobenzoic acid (TFBA), and 35 L of THF.
To a separate inerted 400 L Hastelloy vessel were added
4.0 kg (174 mol) of lithium amide along with 20 L THF. 10
L of the solution of CIA and TFBA were transferred to the
slurry of lithium amide, and the reaction was stirred at 35 (
10 °C until a noticeable change in color was observed and
a significant evolution of gas had occurred. The remaining
solution was then transferred maintaining a temperature 35
( 10 °C. A 20 L rinse of THF was transferred into the
reaction. The reaction was heated to 50 ( 5 °C for about 4
h, then diluted with 50 L acetonitrile, and heated to 50 ( 5
°C for about 1 h. A quench solution was prepared by mixing
100 L of water with 20 kg of 35% aqueous HCl and
transferred to the reaction solution while maintaining the
In the substitutions with 2,3,4-trifluorobenzoic acid using
lithium amide, only ortho substitution was observed, while
coupling at the 3- or 4-positions was not observed. A
comparison of the reactivity of 2-fluorobenzoic acid with
4-fluorobenzoic acid demonstrates the difference in reactivity
at the 2- and 4-positions (Scheme 2). When these substrates
were treated with N-methylaniline and lithium amide in THF
at 50 °C, we found that an 87% yield of the 2-substituted
product was obtained, while the 4-fluorobenzoic acid did not
react. Additionally, if lithium amide was replaced with
sodium amide as the base in the trifluorobenzoic acid case,
a 94:6 (ortho:para) substitution mixture was observed
(Scheme 3). Apparently, the lithium cation plays an important
role in controlling the regioselectivity of the reaction,
directing substitution to the 2-position. In a related system,
Miyano has proposed that coordination of the metal cation
with the carboxyl group and the leaving group facilitates the
elimination process, thereby resulting in a preference for
substitution at the 2-position.10 Alternatively, the lithium
carboxylate may coordinate the incoming nucleophile, thereby
promoting substitution at the 2-position.11
In conclusion, an efficient and scaleable coupling of
anilines with 2-fluorobenzoic acids has been developed. This
method provided the anthranilic acid products in good to
high yield. The methodology was extended to the use of
2-fluorobenzamides and gave N-arylanthranilamide products.
Exclusive selectivity for the 2-position was observed when
2,3,4-trifluorobenzoic acid was used as the electrophile.
Experimental Procedure
General Remarks. Lithium amide (or lithium hydride)
is an extremely hazardous substance which is an extreme
(10) Hattori, T.; Satoh, S.; Miyano, S. Bull. Chem. Soc. Jpn. 1993, 66, 3840-
3842.
(11) For a related system, see Bahmanyar, S.; Borer, B. C.; Kim, Y. M.; Kurtz,
D. M.; Yu, S. Organic Lett. 2005, 7, 1011-1014.
Vol. 9, No. 6, 2005 / Organic Process Research & Development
•
845

temperature at 50 ( 5 °C. The solution was then cooled
slowly to 0 ( 5 °C, the solid product was collected by
(12) The reagents, reaction mixture, and the quenched reaction mixture were
evaluated by DSC (sealed, 35-250 °C @ 10 °C/min) with the following
results.
filtration, and the cake was washed with a mixture of 40 kg
of acetonitrile and 10 L of water, followed by 50 L of water.
The aqueous filtrates were neutralized by the addition of
sodium hydroxide and sent for incineration. The
wet cake was dried in a vacuum oven at 55 ( 5 °C re-
sulting in 18.9 kg (90%) of 2-(2-chloro-4-iodophenylamino)-
3,4-difluorobenzoic acid (CIPFA, 3a). HPLC: 99.2% (by
area %).
1H NMR (400 MHz, d6-DMSO) δ 9.2 (bs, 1H), 7.9 (m,
2H), 7.58 (m, 1H), 7.2 (m, 1H), 6.78 (m, 1H), 3.4 (bs, 1H);
melt 230-231 C.
Acknowledgment
The authors would like to thank Mr. Joseph M. Fritsch,
Dr. Kevin M. Schaab, and Dr. Arthur D. Besteman for
analytical assistance and Mr. Gary Dozeman for hazards
analysis.
In addition, calorimetric evaluation was performed (specific heats were
estimated). The addition of the TFBA/CIA/THF/MeCN solution gave a
calculated 293 °C adiabatic heat rise while being held at 35 °C (between
500 and 600 kJ/kg of product formed was calculated). The reaction resulted
in 164 L of gas/kg of CIA used being observed. This compares to 376 L
of gas/kg of CIA used being generated in theory. It seems that much of
the ammonia is solvated and stays in the vessel. Both the gas and heat
evolution were dose controlled and did not pose a problem upon scale up.
The HCl quench gave a calculated 49 °C adiabatic heat rise (20 °C heat
rise observed in the equipment) with no significant gas evolution.
Received for review July 20, 2005.
OP0501242
846
•
Vol. 9, No. 6, 2005 / Organic Process Research & Development