Chemoselectivity in the reactions between ethyl 4,4,4-trifluoro-3-oxobutanoate and anilines: Improved synthesis of 2-trifluoromethyl-4- and 4-trifluoromethyl-2-quinolinones

Article abstract of DOI:10.1055/s-2003-41043

Chemoselectivity in the reactions between ethyl 4,4, 4-trifluoroacetoacetate (ethyl 4,4,4-trifluoro-3-oxobutanoate) and various anilines was systematically studied as a function of the reaction conditions used (solvent/temperature, catalyst). The results obtained allowed chemoselective (>90%) synthesis of the corresponding ethyl 3-arylamino-4,4,4-trifluorobut-2-enoates and N-aryl-4,4,4-trifluoro-3-oxobutyramides, which were cyclized to afford 2-trifluoromethyl-4-quinolinones and 4-trifluoromethyl-2-quinolinones, respectively.

Full text of DOI:10.1055/s-2003-41043

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
PAPER
2005
Chemoselectivity in the Reactions Between Ethyl 4,4,4-Trifluoro-3-oxobuta-
noate and Anilines: Improved Synthesis of 2-Trifluoromethyl-4- and 4-Tri-
fluoromethyl-2-quinolinones
S
ynthesis of 2-Trif
m
luoromethyl-4- and
i
4-triflu
t
orom
e
r
thyl-2-qu
i
inolin
iones O. Berbasov, Vadim A. Soloshonok*
Department of Chemistry and Biochemistry, University of Oklahoma, Norman, OK 73019, USA
Fax +1(405)3256111; E-mail: vadim@ou.edu
Received 21 April 2003; revised 14 May 2003
pounds (e.g. ethyl 4,4,4-trifluoroacetoacetate, ethyl 3,3,3-
Abstract: Chemoselectivity in the reactions between ethyl 4,4,4-
trifluoroacetoacetate (ethyl 4,4,4-trifluoro-3-oxobutanoate) and
various anilines was systematically studied as a function of the re-
action conditions used (solvent/temperature, catalyst). The results
trifluoropyruvate, and 1,1,1,5,5,5-hexafluoropentane-2,4-
dione) with benzylamine derivatives, we conducted a sys-
tematic study of these reactions and found proper condi-
obtained allowed chemoselective (>90%) synthesis of the corre- tions and reagents allowing for highly regio/
sponding ethyl 3-arylamino-4,4,4-trifluorobut-2-enoates and N-
aryl-4,4,4-trifluoro-3-oxobutyramides, which were cyclized to af-
ford 2-trifluoromethyl-4-quinolinones and 4-trifluoromethyl-2-
recent results9 and taking into account the obvious simi-
quinolinones, respectively.
chemoselective and efficient preparation of the target flu-
orinated imines/enamines. Drawing inspiration from our
larity in the nature of the problems faced by Schlosser’s1,2
and our groups, we decided to apply our approach to study
the regioselectivity in the reactions between ethyl 4,4,4-
Key words: chemoselectivity, enamino esters, keto amides, quino-
lines, fluorine
trifluoroacetoacetate (3) and anilines 4.
Here we report synthetically useful procedures for
Recently Schlosser’s group reported a series of publica-
chemoselective (>90%) preparation of the corresponding
ethyl 4,4,4-trifluoro-3-phenylaminobut-2-enoates 5 and
4,4,4-trifluoro-3-oxo-N-phenylbutyramides 6, which
were cyclized, using the literature methods, to afford 2-tri-
fluoromethyl-4-quinolinones 1 and 4-trifluoro-methyl-2-
quinolinones 2, respectively.
tions on synthetically useful elaboration of 2- and 4-(tri-
fluoromethyl)quinolinones
1
and
2
into various
polyfunctional trifluoromethyl-containing quinoline de-
rivatives.1–4 While the chemistry developed by the au-
thors’ features high regioselectivity and chemical yields,
the synthetic usefulness of the method as a whole is nulli-
fied by the poor availability of the starting quinolinones 1
and 2.1–4 In our own experience we faced the same logistic
problems of relatively low chemical yields of the corre-
sponding a-fluoroalkyl-containing imines/enamines,
starting materials for preparing fluorinated amines and
amino acids by the biomimetic transamination methodol-
ogy developed by us.5–8 To solve the problem of low
chemo/regioselectivity and unsatisfactory chemical yields
in the reactions between highly electrophilic and/or poly-
functional trifluoromethyl-containing carbonyl com-
The reaction dichotomy or chemodivergence can be an ir-
ritating synthetic obstacle or a pleasant synthetic bonus
provided by the chemistry of polyfunctional organic com-
pounds. However, search for reaction conditions allow-
ing, at will, selective synthesis of either of two or more
different target products, using the same starting com-
pounds, is not an ordinary task. For instance, ethyl triflu-
oroacetoacetate (3) react with aniline 4 (Scheme 1) to
afford 2-trifluoromethyl-4-quinolinone (1)1 or 4-trifluo-
romethyl-2-quinolinone (2),2 depending on the reaction
conditions used. While these methods feature attractive
H2N
F3C
O
Et
R
O
O
+
3
4
110 °C
polyphosphoric acid, 100 °C
R
O
CF3
75% H2SO4
90 °C
F3C
O
R
R
Et
150 °C
F3C
N
R
N
O
H
H
N
H
CF3
N
H
O
O
O
6
5
2
1
Yields: 5.5-17% (based on 3)
11-34% (based on 4)
Yields: 46-79%
Scheme 1
Synthesis 2003, No. 13, Print: 18 09 2003. Web: 28 08 2003.
Art Id.1437-210X,E;2003,0,13,2005,2010,ftx,en;M01103SS.pdf.
DOI: 10.1055/s-2003-41043
© Georg Thieme Verlag Stuttgart · New York

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
2006
D. O. Berbasov, V. A. Soloshonok
PAPER
operationally convenient conditions, their practical use is on searching for reaction conditions that would lead to the
plagued by the chemical yields. Thus, 2-trifluoromethyl- chemoselective preparation of key intermediates 5 and 6.
4-quinolinones 11 are available in moderate-to-low yields
by heating the starting compounds 3 and 4 in polyphos-
Brief analysis of the reaction conditions for the condensa-
tions of 3 and 4 (Scheme 1) allowed us to notice that the
phoric acid. On the other hand, preparation of 4-trifluo-
reactions conducted under the acidic conditions give a
romethyl-2-quinolinones 22 by the direct condensation of
preference for the formation of the intermediate enamino
keto ester 3 with anilines 4, followed by heating of the re-
ester 5, while under the neutral, or slightly basic (aniline)
sultant mixture in 75% sulfuric acid, gives disappointing-
conditions amide 6 is the favored product. This conclu-
sion is in good agreement with the trend of chemoselectiv-
ly low yields of the target products 2.2,10
There are several alternative approaches available in the ity we had observed before in the reactions of keto ester 3
literature for preparing quinolinones (quinolinols) 2. In with benzylamines.9 Therefore, we decided to try the re-
one of them,11 quinoline-4-carboxylic acid was first fluor- action between 3 and acetic acid salt of aniline 4a
inated with SF4 to give the corresponding trifluoromethyl (Scheme 2) using chloroform as a solvent, the conditions
derivative (69%) which was oxidized with 30% H2O2 to that allowed us to prepare N-benzylenamino ester 8 with
afford 4-(trifluoromethyl)quinoline 1-oxide (68.5%). virtually complete chemoselectivity.9 Under these condi-
Thermal isomerization of the N-oxide, thus prepared, in a tions the reaction proceeded at a low rate and, quite sur-
solution of Ac2O afforded the target 2 in 74% yield. An- prisingly, with the formation of noticeable amounts of
other method12 focused on preparation of amides 6 enamino amide 7a (Table 1, entry 1). Application of the
(Scheme 1), used as a starting material for further Knorr– trifluoroacetic acid salt of 4a in the reaction with 3 result-
Conrad–Limpach cyclization to prepare quinolinones ed in an improved ratio of enamino ester 5a which was
(quinolinols) 2. In this method a,a-dihydroperfluoroal- isolated in 90% yield (entry 2). To briefly assess the gen-
kanoic acid, a,a-dihydroperfluorobutanoic acid in partic- erality of the reaction, we conducted condensations of
ular, was converted to the corresponding N-arylamide keto ester 3 with the trifluoroacetates of p-methyl- 4b, p-
(PCC, >95%), which was dehydrofluorinated (Et3N, methoxy- 4c and p-fluoroanilines 4d (entries 3–5). In all
NaHCO3, >95%) to give N-aryl-3-fluoro-3-fluoroalkyl- cases the target compounds 5b–d were isolated with more
prop-2-enamides. These a,b-unsaturated amides were than 90% chemical yields rendering the reaction condi-
converted to the target N-aryl-3-oxoamides 6 using two- tions described here synthetically useful for chemoselec-
step procedure including Michael addition reaction with tive synthesis of enamino esters 5, key intermediates for
pyrrolidine and hydrolysis of the corresponding addition preparation of various derivatives of 2-trifluoromethyl-4-
products.
quinolinones 1.
Considering these methods, one may agree that despite As mentioned above, formation of enamino amide 7a was
the synthetically useful chemical yields they have in com- rather unexpected since in the reaction between keto ester
mon one significant drawback, which is the methodologi- 3 and the acetate of benzylamine, conducted under the
cal deficiency of a multi-step procedure. From this point same conditions, the enamino ester 8 was obtained as the
of view, the direct condensation between the keto ester 3 sole product.9 Therefore we decided to study this intrigu-
and anilines 4 (Scheme 1) holds an apparent advantage of ing difference in reactivity in detail. Thus, use of benzene
a straightforward procedure, provided of course, that the as a solvent (higher reaction temperature) for the reaction
target products can be obtained in chemical yields compa- of 3 with trifluoroacetate of 4a resulted in a substantially
rable with the literature methods. Therefore, we decided increased formation of enamino amide 7a (entry 6 vs 2).
to study the reactions between keto ester 3 and anilines 4 Further increase in the reaction temperature (toluene as a
in detail to realize the methodological potential of this ap- solvent) and application of 4a salt with the weaker acetic
proach for preparing trifluoromethyl-containing quinoli- acid allowed the preparation of amide 7a in synthetically
nones (quinolinols) 1 and 2. Assuming that the Knorr– useful chemical yield (entry 7 vs 6 and 2).
Conrad–Limpach cyclization of enamino esters 5 and
amides 6 do not pose any synthetic problem, we focused
H
N
R
F3C
OEt
F3C
F3C
OEt
Acid
H3N
N
O
N
O
+
4a-d
H
H
R
O
O
3
Table 1
7a-d
R
5a-d
R
CH3COO
ref. [9]
H3N
OEt
Ar
H
F3C
OEt
F3C
N
4a
R = H (a),
Me (b),
MeO (c),
F (d)
F3C
N
O
N
O
R
Ar
N
O
8
H
5'a-d
7'a
R
R
> 99% chemoselectivity
Scheme 2
Synthesis 2003, No. 13, 2005–2010 © Thieme Stuttgart · New York

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
PAPER
Synthesis of 2-Trifluoromethyl-4- and 4-trifluoromethyl-2-quinolinones
2007
Table 1 Reactions of Keto Ester 3 with Salts 4a–d. Synthesis of Enamino Esters 5a–d and Enamino Amides 7a–d (Scheme 2)a
Entry
Solvent
Acid
Ratiob
Time
(h)
Conversion
(%)
Ratioc
7/5
Yield of
5a–cd (%)
1
CHCl3
CHCl3
CHCl3
CHCl3
CHCl3
benzene
toluene
toluene
AcOH
TFA
TFA
TFA
TFA
TFA
AcOH
–
1:1.5
1:1.5
1:1.5
1:1.5
1:1.5
1 :1.5
1:2
50
68
70
69
75
45
35
1.5
95
>99
99
15 (93:7)e/85 (78:22)f
9 (82:18)/91 (>99:1)
10 (>99:1)/90 (85:15)
1/>99 (90:10)
–
2
92
96
90
81
–
3g
4h
5i
6
99
98
4 (100:0)/96 (86:14)
60 (96:4)/40 (89:11)
87 (90:11)/13 (76:24)
88 (89:11)/12 (74:26)
66
7
>99
63
–
8
1:3.5
–
a All reactions were conducted under reflux in the indicated solvent.
b Ratio between keto ester 3 and the corresponding aniline 4a–d (Scheme 2).
c Determined by 19F NMR (300 MHz) analysis of the crude reaction mixtures.
d The enamino ester 5 was isolated as the major reaction product.
e Ratio of enamino amide 7a–d to imino amide 7¢a–d respectively (Scheme 2).
f Ratio of enamino ester 5a–d to imino ester 5¢a–d respectively (Scheme 2).
g Entry corresponding to p-toluidine 4b.
h Entry corresponding to p-anisidine 4c.
i Entry corresponding to p-fluoroaniline 4d.
We believe that the difference in the outcome of the reac- application as precursors for preparation of keto amides 6,
tions of keto ester 3 with benzylamine and aniline stems the key compounds in synthesis of 4-trifluoromethyl-2-
from a low stability of enamino esters 5a–d and lower ba- quinolinones 2. To this end we attempted chemoselective
sicity of aniline as compared with that of benzylamine. hydrolysis of derivative 7a to afford the target amides 6.
Previously we showed that benzylamine derived com- Since the minor reaction product enamino ester 5a was
pound 8 exists exclusively in enamino form stabilized by not expected to interfere with the hydrolysis, we used the
intramolecular hydrogen bond. We have also demonstrat- mixture of 7a and 5a (ratio = 88:12) as the starting mate-
ed that due to the low electrophilicity of the conjugated es- rial. The reaction was conducted under heterogeneous
ter function in 8, it did not react with excess of conditions by stirring an ethereal solution of 7a and 5a
benzylamine or its salts.9 By contrast, as it follows from with aqueous 3 N HCl. Upon completion (monitoring by
NMR data, enamino esters 5a–d exist as a mixture with 19F NMR spectroscopy), the organic layer was evaporated
the corresponding ketimines 5¢a–d in which the ester to give the target amide 6a (hydrate) in 60% yield
group is relatively electrophilic to react with a nucleo- (Scheme 3). Further attempts to increase the chemical
phile. Furthermore, taking into account the low basicity of yield by optimization of the reaction conditions did not
anilines 4a–d, we can assume that their salts, in particular give any improvement, probably due to noticeable solu-
acetates, might be unstable at high temperatures (benzene, bility of product 6a in the aqueous phase. Though success-
toluene) generating free bases 4a–d capable of reacting ful and operationally simple, as compared with the
with the ester group of imines 5¢a–d. A strong support for literature methods, this approach still cannot be rendered
this rationale comes from the reaction of keto ester 3 with synthetically efficient, affording the target compounds 6
free aniline 4a conducted in toluene which gave a mixture in moderate yields via two-step (condensation-hydrolysis)
of 5a and 7a in virtually the same ratio observed in the re- procedure. Therefore, we continued a search for conve-
action of 3 with the acetate of 4a (entry 8 vs 7).
nient and efficient reaction conditions for a more efficient
preparation of amides 6 and thus quinolinones 2.
Taking into account that enamino amides 7 can be ob-
tained in synthetically useful yields, we considered their
3N HCl
7a (7'a)
CF3
H
75% H2SO4
90 °C
R
F3C
HO OH
N
O
4a-d/NEt3
ref. [2]
N
H
O
R
3
Table 2
6'a-d
R = H (a), Me (b), MeO (c), F (d)
2a
Scheme 3
Synthesis 2003, No. 13, 2005–2010 © Thieme Stuttgart · New York

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
2008
D. O. Berbasov, V. A. Soloshonok
PAPER
Table 2 Synthesis of Amides 6a–d by the Reactions of Keto Ester
3 with Anilines 4a–d in the Presence of Triethylamine (Scheme 3)a
As discussed above, the target amides 6 are assumed to be
favored products in the direct condensation between keto
ester 3 and anilines 4 conducted under neutral or slightly
basic conditions. Therefore, we focused our efforts on us-
ing various amines as additives to influence the chemose-
lectivity of the reaction. After numerous attempts, we
found that the condensation between keto ester 3 and
aniline 4a conducted in the presence of triethylamine gave
rise to amide 6a with virtually complete (>99%) chemose-
lectivity (Scheme 3). Thus, heating a mixture of 3/4a/
Et3N in a ratio of 1:3.5:3 in toluene under reflux resulted
in relatively fast (2.5 h) and complete (>98%) conversion
of the starting compounds to afford amide 6a as the sole
product (Table 2, entry 1). A decrease in the ratio of both
4a and Et3N resulted in lower reaction rates but did not in-
fluence the chemical outcome (entries 2, 3). The reactions
conducted in benzene and chloroform showed the same
excellent chemoselectivity but proceeded at substantially
slower rates (entries 4, 5). Therefore, in our opinion, the
optimum reaction conditions presented in entry 3 were
used to check the generality of this procedure. The con-
densations of keto ester 3 with p-methyl- 4b, p-methoxy-
4c and p-fluoroanilines 4d (entries 6–8) proceeded at sim-
ilar reaction rates and regardless of the nature of the sub-
stituent on the phenyl ring of 4, affording the target
amides 6b–d as individual reaction products. On the other
hand, isolation of 6a–d posed some problems due to solu-
bility of the hydrates of amides 6a–d in aqueous solutions.
The highest isolated yields of 6a–d we obtained were in
the range of 55–65%. Therefore, taking into account that
under the Knorr–Conrad–Limpach conditions, the 20%
excess of aniline 4a–d and Et3N might not interfere with
the cyclization forming the corresponding water-soluble
salts, we tried the cyclization of 6a without prior isolation.
Thus, crude 6a, obtained by evaporation of the toluene so-
lution, was treated with 75% sulfuric acid according to the
literature procedures.2,12 The result was rather satisfactory
as the target 4-trifluoromethyl-2-quinolinone (2a) was
isolated in 73% (based on 3) chemical yield. Accordingly,
Entry Solvent
Ratio
3/4a–d/Et3N
Time (h) Conversionb Yield of
(%)
6a–dc (%)
1
toluene
toluene
toluene
benzene
CHCl3
1:3.5:3
1:2:1.5
1:1.2:2
1:1.2:2
1:1.2:2
1:1.2:2
1:1.2:2
1:1.2:2
2.5
3
98
60
–
2
96
3
7
96
60
–
4
5 days 92
5
14 days
90
96
95
92
–
6d
7e
8f
toluene
toluene
toluene
7
7
7
69
55
66
a All reactions were conducted under reflux in the indicated solvent
using designated ratio of 3, 4a–d and Et3N.
b Determined by 19F NMR spectroscopy (300 MHz) of the crude reac-
tion mixtures.
c Isolated yield of pure products 6a–d.
d Entry corresponding to p-toluidine 4b.
e Entry corresponding to p-anisidine 4c.
f Entry corresponding to p-fluoroaniline 4d.
Unless otherwise noted, all reagents and solvents were obtained
from commercial suppliers and used without further purification.
All the reactions were carried out without any special caution to ex-
clude air. Unless indicated, 1H, 19F and 13C NMR spectra, were tak-
en in CDCl3 solutions at 299.95, 282.24 and 75.42 MHz,
respectively, on the NMR instrument in the University of Oklahoma
NMR Spectroscopy Laboratory. Chemical shifts refer to TMS and
CFCl3 as the internal standards.
Yields refer to isolated yields of products of greater than 95% purity
as estimated by 1H and 19F NMR spectroscopy. All new compounds
were characterized by 1H, 19F, 13C NMR and mass spectroscopy.
4,4,4-Trifluoro-3-(phenylamino)but-2-enoic Acid Ethyl Ester
(5a); Typical Procedure
the condensation of keto ester 3 with anilines 4, conducted (Scheme 2, Table 1, entry 2)
in the presence of Et3N, and further cyclization of the in-
termediate 6 could be recommended as a general, conve-
nient, economical and operationally simple method for
preparing 4-trifluoromethyl-2-quinolinones 2 in reason-
able chemical yields.
To a solution of aniline 4a (0.381 g, 4.07 mmol) and trifluoroacetic
acid (0.465 g, 4.07 mmol) in CHCl3 (2 mL) was added a solution of
keto ester 3 (0.505 g, 2.75 mmol) in CHCl3 (3 mL) at r.t. The result-
ant mixture was refluxed until the reaction was completed (68 h,
monitored by 19F NMR spectroscopy). The CHCl3 solution was
washed with H2O (3 × 10 mL), dried (MgSO4) and evaporated to af-
ford the enamino ester 5a as a viscous oil (0.651 g, 92%); Rf 0.84
(hexanes–EtOAc, 5:1); exists as a mixture of enamino ester 5a and
imino ester 5¢a in a ratio of 4.4:1.
In summary, a systematic study of the reactions between
keto ester 3 and anilines 4 allowed us to find suitable re-
action conditions for highly chemoselective synthesis of
either enamino esters 5 (Table 1) or amides 613 (Table 2),
key intermediates for preparation of 2-trifluoromethyl-4-
quinolinones 1 and 4-trifluoromethyl-2-quinolinones 2.
Simplicity and operational convenience of the experimen-
tal procedures as well as reasonable chemical yields and
ratios of the starting compounds, compared to the litera-
ture methods, render the procedures developed in this
study synthetically useful.
5a
1H NMR (CDCl3): d = 9.83 (s, 1 H), 7.15–7.35 (m, 5 H), 5.33 (s, 1
H), 4.20 (q, J = 7.2 Hz, 2 H), 3.41 (s, 2 H), 1.30 (t, J = 7.2 Hz, 3 H).
13C NMR (CDCl3): d = 169.8, 147.3 (q, J = 31.4 Hz), 138.6, 129.1,
126.8, 126.2, 120.5 (q, J = 276.7 Hz), 88.8 (q, J = 5.5 Hz), 60.4,
14.6.
19F NMR (CDCl3): d = –63.39 (s).
Synthesis 2003, No. 13, 2005–2010 © Thieme Stuttgart · New York