Expedient synthesis of 6-acylindolo[1,2- a ]quinoxalines

Article abstract of DOI:10.1055/s-0034-1380515

novel route leading to 6-acylindolo[1,2-a]quinoxalines involving condensation of N-(2-iodoaryl)-2-nitrosoanilines with β-diketones followed by Heck cyclization is described.

Full text of DOI:10.1055/s-0034-1380515

SYNLETT
0
9
3
6
-
5
2
1
4
1
4
3
7
-
2
0
9
6
© Georg Thieme Verlag Stuttgart · New York
2015, 26, 1352–1356
letter
1352
A. Trawczyński et al.
Letter
Synlett
Expedient Synthesis of 6-Acylindolo[1,2-a]quinoxalines
Adam Trawczyńskia
Magdalena Telegab
Zbigniew Wróbel*a
O
R3
O
I
H
N
N
R4
R2
1. t-BuNH2, MeCN, r.t.
O
O
2. Heck–Jeffery conditions
N
N
+
a Institute of Organic Chemistry, Polish Academy of Sciences,
Kasprzaka 44, 01-224 Warsaw, Poland
zbigniew.wrobel@icho.edu.pl
b Faculty of Chemistry, Adam Mickiewicz University,
Umultowska 89b, 61-614 Poznań, Poland
R4
R3
R1 = Cl, Ph
up to 92% overall yield
R2
R1
R1
R
2 = H, Cl
R3 = H, F, Cl, Me
R4 = Me, Ph
Received: 27.01.2015
Accepted after revision: 08.03.2015
Published online: 10.04.2015
R4
R4
R2
HO
R3
DOI: 10.1055/s-0034-1380515; Art ID: st-2015-b0060-l
R1
R3
Abstract A novel route leading to 6-acylindolo[1,2-a]quinoxalines
involving condensation of N-(2-iodoaryl)-2-nitrosoanilines with β-di-
ketones followed by Heck cyclization is described.
N
N
X = H
ref. 2a
Key words cyclization, condensation, heterocycles, annulation, Heck
reaction, indoles
R2
R2
CO2R
R1
CO2R
R1
The indolo[1,2-a]quinoxaline motif is present in many
biologically active compounds exhibiting kinase receptor
activity or antifungal activity.1 Only a few strategies for the
synthesis of such compounds have been reported (Scheme
1). All of them employ a suitably functionalized N-arylin-
dole scaffold which in inter- or intramolecular mode under-
goes annulation with the formation of the quinoxaline moi-
ety.1,2
N
HN
X
N
Ts
N
X = Ts
ref. 2b
HOCAr
Ar
R1
FeCl3•6H2O
X = H
2 = H
N
NH
R
This approach requires more or less sophisticated indole
derivatives as starting materials, and their synthesis from
simpler substrates is not always short or easy. Particularly
in the last example, a synthesis of indoloacetylene, the
starting material for the synthesis of 6-aryoylindolo[1,2-a]-
quinoxalines,2c requires at least five steps from available
simple compounds. According to the literature data it is the
only way to synthesize 6-aryloylindolo[1,2-a]quinoxalines.
An alternative approach for the synthesis of the title
compounds which comprises formation of the indole five-
membered ring by annulation of the suitable quinoxaline
derivative can be, however, proposed (Scheme 2). The route
to the latter can be founded on the basis of the results of our
earlier works.
ref. 1b
R2
O
N
NaN3
Ar
R1
R1
Ar
CuI, O2
ref. 2c
N
N
NO2
R3
R3
Scheme 1 Reported strategies for the synthesis of indolo[1,2-a]qui-
noxalines
In 2007 a simple and efficient procedure for the synthe-
sis of N-aryl-2-nitrosoanilines was found in our laboratory,3
and its application in the synthesis of a variety of quinoxal-
ine derivatives has been demonstrated.4 We have also found
lately that 2-methylene-1-aryl-1,2-dihydroquinoxalines,
such as 3 in Scheme 3, can be synthesized directly from
© Georg Thieme Verlag Stuttgart · New York — Synlett 2015, 26, 1352–1356

1353
A. Trawczyński et al.
Letter
Synlett
in any solvent but trifluoroacetic acid, which made its isola-
tion and purification by column chromatography very in-
convenient. On the other hand, due to its low solubility it
could be easily isolated from the reaction mixture by simple
filtering off the precipitate and washing it to yield practical-
ly pure product.
O
N
O
c
I
I
R4
R4
a
O
b
NH
N
N
N
N
c
As the resulting yield was not fully satisfying we tried to
isolate and purify the intermediate 3ba. It was, however,
moderately stable towards column chromatography and its
efficient purification was very difficult. Thus, we tried to
optimize the Heck-type cyclization step starting from
crude 3aa (X = Br) and 3ba (X = I). Several attempts were
made to optimize the cyclization step using PdCl2, Pd(OAc)2
with or without Ph3P, with various amounts of Et3N, Bu3N,
and DIPEA in MeCN, toluene, or DMF at 80–130 °C, varying
the amount of the catalyst from 1–5%. In all cases, however,
the results of the reaction were not reproducible. Finally,
we tried the Jeffery protocol for the Heck reaction,7 that is,
relatively big load of Pd(OAc)2 (10%), tetrabutylammonium
chloride hydrate (TBACl·H2O), and KOAc as a base in DMF.
Under these conditions applied for crude 3ba, which gave
usually better results than 3aa, the temperature of the reac-
tion could be lowered to 60 °C, and the results became re-
producible achieving 58% yield of 4ba (Table 1, entry 1, col-
umn A).
a, b - reported strategies of the annulation
c - this work approach
Scheme 2 Possible retrosynthetic routes to the title compounds
N-aryl-2-nitrosoanilines and 1,3-diketones.5 The reaction
proceeds efficiently only with nitrosoanilines bearing at
least one ortho substituent in the N-aryl moiety. For the
sake of this work Br or I at that position would provide the
intermediate 3 suitable for subsequent Heck cyclization.
The palladium-catalyzed arylation of olefins is widely used
for the carbon–carbon bond formation in organic synthesis.
The intramolecular variant of this reaction employing
enamine or enamide double bonds has found numerous ap-
plications for the synthesis of fused heterocycles, principal-
ly indole derivatives.6 To our knowledge the indolo[1,2-a]-
quinoxaline scaffold has never been constructed in such a
way.
The first step of the reaction sequence was optimized
earlier.5 Following on those results we decided to perform it
in the t-BuNH2/MeCN system; this turned out to be not the
best regarding yields but easy for workup as it involved only
volatile components.
The above conditions (method A) were then applied for
the second step of the complete sequence 1 → 4 starting
from several N-(2-iodoaryl)-2-nitrosoanilines 1b–g, pre-
pared analogously to the previously described procedures
from nitroarenes and 2-iodoanilines.3,8 The condensation
step (1 → 3) was carried out by stirring the reagents in the
t-BuNH2/MeCN system until complete conversion was
achieved (TLC monitoring), then the mixture was evaporat-
ed and directly subjected to the cyclization (3 → 4) in the
same reaction vessel.9 After the reaction was finished, the
pure products were easily isolated by filtration as it was
done in the test experiment. Contrary to the similar 6-
acylindolo[1,2-a]quinoxalines described in the other work2c
we found most of the products insoluble in almost any sol-
vent. In the cases of the partially soluble products (Table 1,
entries 15–17) column chromatography was used for their
isolation. The analytical samples of 4 were obtained by suc-
ceeding recrystallization from DMF or hexane–EtOAc.
In the test experiment N-(2-bromo-4-chlorophenyl)-5-
chloro-2-nitrosoaniline (1a, R1 = Cl, R3 = Cl, X = Br) and ben-
zoyl acetone (2a) in dry MeCN were treated with t-BuNH2
and stirred at room temperature in a stoppered flask for
four days. After that time the TLC control showed almost
complete conversion of 1 and solely one product. The crude
product 3aa was heated in an ampoule in dry MeCN with
addition of Pd(OAc)2, Ph3P, and Et3N at 120 °C overnight.
The orange solid of 6-benzoyl-2,9-dichloroindolo[1,2-a]-
quinoxaline 4aa was obtained in 48% overall yield from 1a.
The experiment was repeated starting from iodo derivative
1b with similar final result of 4ba = 4aa. The product
turned out to be practically insoluble at room temperature
R5
O
N
R5
O
N
O
R3
X
X
N
R4
R4
Heck
conditions
H
N
R2
O
O
t-BuNH2
R3
N
N
R5
+
R4
MeCN, r.t.
R3
R2
R2
R1
R1
R1
2a: R4 = Ph, R5 = H
2b: R4 = Me, R5 = H
2c: R4 = i-Pr, R5 = H
1
X = Br, I
3
4
Scheme 3 One-pot, two-step synthesis of 6-acylindolo[1,2-a]quinoxalines 4
© Georg Thieme Verlag Stuttgart · New York — Synlett 2015, 26, 1352–1356

1354
A. Trawczyński et al.
Letter
Synlett
Table 1 One-Pot, Two-Step Bisannulation of 1 (Scheme 3, X = I)9,10
Entry
R1
Cl
R2
R3
Cl
Compd 1
R4
R5
Compd 4
Method A:
Method B:
total yield (%)a
total yield (%)a
1
2
H
1b
Ph
H
4ba
4bb
58
36
0
78
81
0
Me
Et
H
3
Me
H
4
Cl
H
H
1c
Ph
Me
Ph
Ph
Me
i-Pr
Et
4ca
4cb
4da
4ea
4eb
4ec
23
20
56
38
41
23b
0
92
–
5
H
6
Ph
Cl
H
H
Cl
1d
1e
H
–
7
Me
H
82
80
–
8
H
9
H
10
11
12
13
14
15
16
17
Me
H
–
Cl
Cl
Cl
H
H
Cl
H
H
F
1f
Ph
Me
Ph
Me
Ph
Me
Ph
4fa
40
25
26
28
74b
73b
61b
67
–
H
4fb
4ga
4gb
4ha
4hb
4ia
Me
t-Bu
H
1g
1h
1i
H
62
64
82b
–
H
H
H
H
72b
a Isolated yield of products separated by filtration.
b Isolated by column chromatography.
For relatively more stable 3, the Heck cyclization
seemed to be high-yielding. It was shown in the reaction of
partially purified (of ca. 80% purity) 3ba which, following
method A, gave 4ba in almost 80% yield, thus it appeared to
be nearly quantitative. However, most of the intermediate
3, which could not be isolated due to their instability, gave
rather low yields of final products also in the one pot 1 → 4
sequence. Since it might be caused by some amount of wa-
ter coming into the reaction mixture along with TBACl·H2O,
we decided to apply a more restricted procedure using an-
hydrous, nonhygroscopic bromide (TBABr, method B) in-
stead. The overall yields of the two-step sequence grew up
significantly up to 92% (Table 1, method B), most notably in
cases when the condensation step was efficient.
The obtained results do not allow for discussing the de-
tailed mechanistic scheme of this Heck-type cyclization. Al-
though the generally preferred way of the five-membered
ring formation in this type of reactions is the exo-mode cy-
clization, the intramolecular 5-endo-trig ring closure has
been reported, particularly in the synthesis of indole skele-
ton.6
action does not seem obvious. Any rotation of the interme-
diate A in order to reach syn conformation of the [Pd] sub-
stituent and β-hydrogen, requested for their elimination, is
impossible due to the rigid ring configuration. Besides,
there are no other suitable hydrogen atoms, neither at the
nitrogen atom nor at the side chain/ring, which could play
the role. This reasoning may vote for route B as the most
likely. On the other hand, one can argue that conformation
of the vicinal cis substituents in the pyrrolidine ring of A is
in fact near to that of syn conformation in the open-chain
intermediate, thus, one of the hydrogens of the methylene
group can be eliminated without difficulty.
The results collected in Table 1 reveal scope and also
some limitations of the reaction. The major one is its sensi-
tivity to steric hindrance caused by the diketone substitu-
ents. In fact, only unsubstituted acetyl moieties (R5 = H)
were reactive enough to condense with the aniline nitro-
gen, thus to involve in the cyclization–dehydration pro-
cess.5 When both methyl groups of the diketone were sub-
stituted the reaction did not occur (Table 1, entries 3 and
10). As a result, the products collection is narrowed to those
with 3-unsubstituted indole moiety. By contrast, the other
published method2c for the synthesis of 6-acylindolo[1,2-
a]quinoxalines (shown in Scheme 1), is apparently limited
to compounds with R4 = aryl. Moreover, the procedure re-
ported there turned out to be tricky. In our hands it was dif-
ficult to reproduce, at least in the tested case of 4ia. After
some repetitions and modifications of the described proce-
In fact, this process is often described as ‘formal 5-endo-
trig’ cyclization as the real course of the reaction is rather
obscured. In brief, possible routes of the cyclization of 3 un-
der the Heck reaction conditions can be depicted as in
Scheme 4. Parallel routes have been considered in numer-
ous papers, although without firm conclusions.6j–l Path a
which follows the typical course of intermolecular Heck re-
© Georg Thieme Verlag Stuttgart · New York — Synlett 2015, 26, 1352–1356

1355
A. Trawczyński et al.
Letter
Synlett
Supporting Information
Ln
H
H
Pd+
Supporting information providing general experimental remarks and
analytical data for the remaining new products, is available online at
COR
N
http://dx.doi.org/10.1055/s-0034-1380515.
S
u
p
p
o
nrtIo
g
f
rmoaitn
S
u
p
p
ortiInfogrmoaitn
N
β-hydride
elimination
Ln
a
Pd(OAc)2
KOAc
TBABr
Pd+
References and Notes
5-endo-trig A
3
4
COR
(1) (a) Lin, P.-T.; Salunke, D. B.; Chen, L. H.; Sun, Ch. M. Org. Biomol.
Chem. 2011, 9, 2925. (b) Xu, H.; Fan, L. L. Eur. J. Med. Chem. 2011,
46, 1919.
(2) (a) Patil, N. T.; Kavthe, R. D.; Shinde, V. S.; Sridhar, B. J. Org.
Chem. 2010, 75, 3371. (b) Wang, L.; Guo, W.; Zhang, X.-X.; Xia,
X.-D.; Xiao, W.-J. Org. Lett. 2012, 14, 740. (c) Samala, S.; Arigela,
R. K.; Kant, R.; Kundu, B. J. Org. Chem. 2014, 79, 2491.
N
Ln
N
reductive
elimination
Pd
b
COR
N
N
(3) (a) Wróbel, Z.; Kwast, A. Synlett 2007, 1525. (b) Wróbel, Z.;
Kwast, A. Synthesis 2010, 3865.
palladacycle B
intermediate
(4) (a) Wróbel, Z.; Stachowska, K.; Kwast, A.; Gościk, A.;
Królikiewicz, M.; Pawłowski, R.; Turska, I. Helv. Chim. Acta 2013,
96, 956. (b) Królikiewicz, M.; Cmoch, P.; Wróbel, Z. Synlett 2013,
24, 973. (c) Wróbel, Z.; Królikiewicz, M. J. Heterocycl. Chem.
2014, 51, 123. (d) Królikiewicz, M.; Błaziak, K.; Danikiewicz, W.;
Wróbel, Z. Synlett 2013, 24, 1945.
Scheme 4 Possible routes for the Heck-type cyclization
dure the yield varied from traces up to 37% [CuI (3 equiv =
300 mol%), 48 h] instead of reported 74% [CuI (5% mol), 16
h].2c
(5) Trawczyński, A.; Wróbel, Z. Synlett 2014, 25, 2773.
(6) For representative intramolecular Heck reactions of enamines
leading to indoles, see: (a) Baran, P. S.; Hafensteiner, B. D.;
Ambhaikar, N. B.; Guerrero, C. A.; Gallagher, J. D. J. Am. Chem.
Soc. 2006, 128, 8678. (b) Jia, J.; Zhu, J. J. Org. Chem. 2006, 71,
7826. (c) Fuwa, H.; Sasaki, M. Org. Lett. 2007, 9, 3347.
(d) Ackermann, L.; Kaspar, L. T.; Gschrei, C. J. Chem. Commun.
2004, 2824. (e) Kasahara, A.; Izumi, T.; Murakami, S.; Yanai, H.;
Takatori, M. Bull. Chem. Soc. Jpn. 1986, 59, 927. (f) Garcia-
Cuadrado, D.; de Mendoza, P.; Braga, A. A. C.; Maseras, F.;
Echavarren, A. M. J. Am. Chem. Soc. 2007, 129, 6880. (g) Michael,
J. P.; Chang, S.-F.; Wilson, C. Tetrahedron Lett. 1993, 8365.
(h) Barluenga, J.; Fernandez, M. A.; Aznar, F.; Valdes, C. Chem.
Eur. J. 2005, 11, 2276. (i) Watanabe, T.; Arai, S.; Nishida, A.
Synlett 2004, 907. (j) Maruyama, J.; Yamashita, H.; Watanabe, T.;
Arai, S.; Nishida, A. Tetrahedron 2009, 65, 1327. (k) Latham, E. J.;
Stanforth, S. P. J. Chem. Soc., Perkin Trans. 1 1997, 2059. (l) Iida,
H.; Yuasa, Y.; Kibayashi, C. J. Org. Chem. 1980, 45, 2938.
(7) (a) Jeffery, T. Tetrahedron Lett. 1985, 26, 2667. (b) Jeffery, T. Tet-
rahedron 1996, 52, 10113.
Taking the above into account, our method remains at
the moment the best way to synthesize the title com-
pounds. Moreover, after the reaction, most of the products
could be separated without chromatography by simple fil-
tration which is the great advantage of the protocol.
In conclusion, we have described a new, short synthesis
of 6-acylindolo[1,2-a]quinoxalines 4. Together with the ef-
ficient preparation of N-(2-iodoaryl)-2-nitrosoanilines
from nitroarenes and 2-iodoanilines, the three-step meth-
od provides the title compounds in good overall yield, start-
ing from mostly commercially available starting materials
(Scheme 5).
O
R3
NO2
R4
NH2
R2
I
(8) General Procedure for the Synthesis of N-(2-Iodoaryl)-2-
nitrosoanilines 1a–i
O
O
N
N
+
+
To a cooled solution of KOt-Bu (3,7 g, 30 mmol) in DMF (50 mL)
was added dropwise at –60 °C a solution of the appropriate 2-
iodoaniline (9.6 mmol) in DMF (3 mL) and the nitroarene (9.6
mmol) in DMF (8 mL). The mixture was stirred at –60 °C for 0.5
h then the temperature was raised slowly to –30 °C, and the
reaction was continued for an additional 1 h, poured into sat.
NH4Cl solution (200 mL) and extracted with EtOAc. The extract
was washed with H2O, brine, and dried with Na2SO4. After evap-
oration, the crude product mixture was subjected to column
chromatography (SiO2, hexane–toluene). For analytical data, see
the Supporting Information.
R2
R3
R1
R4
R1
Scheme 5 Three-step way for the synthesis of 6-acylindolo[1,2-a]qui-
noxalines from simple and available starting materials
Acknowledgment
The work was financed by Polish Ministry of Science and Higher Edu-
cation, Diamond Grant DI 2011 014641.
(9) General Procedure for the Preparation of Compounds 4
To a solution of the N-(2-iodoaryl)-2-nitrosobenzamine 1 (2
mmol) and an appropriate 1,3-diketone 2 (2 mmol) in dry
MeCN (10 mL) was added t-BuNH2 (4 mmol). The mixture was
stirred at r.t. for 1–4 d (see the Supporting Information). The
solvent and other volatile materials were then evaporated in
© Georg Thieme Verlag Stuttgart · New York — Synlett 2015, 26, 1352–1356

1356
A. Trawczyński et al.
Letter
Synlett
vacuo, and the crude product was used in the next step without
purification. The reaction vial containing the crude product was
charged with Pd(OAc)2 (45 mg, 0.2 mmol), n-Bu4NCl·H2O
(method A; 1050 mg, 3.5 mmol) or n-Bu4NBr (method B; 1127
mg, 3.5 mmol) and KOAc (1000 mg, 10.2 mmol). The flask was
purged with argon, then DMF (10 mL) was added, and argon
was continuously bubbled via the reaction mixture for 20 min.
The mixture was then heated under a positive argon pressure in
an oil bath at 60–65 °C for 24 h. After cooling down some
amount of H2O (ca. 5 mL) was added. In the cases when a solid
product precipitated it was filtered off and washed with EtOH,
H2O, and EtOAc repeatedly, to yield the product pure on TLC.
Analytical sample was obtained by recrystallization from DMF.
In the cases of soluble products (4ec,ha,hb,ia) the mixture was
diluted with H2O and extracted with EtOAc. The extract was
thoroughly washed with H2O, dried (Na2SO4), and the solvent
was evaporated. The residue was separated by column chroma-
tography (SiO2, hexane–EtOAc). Analytical sample was recrys-
tallized from hexane–EtOAc.
130.44, 130.62, 131.44, 132.93, 132.97, 134.90, 137.61, 140.74,
147.55, 161.34, 185.62. MS (EI): m/z (%) = 394 (13), 393 (18),
392 (68), 391 (36), 390 (100) [M+•], 389 (18), 364 (15), 363 (23),
362 (23), 261 (28), 357 (10), 355 (29), 327 (15), 250 (14), 177
(12). HRMS (EI): m/z calcd for C22H12N2OCl2: 390.0327; found:
390.0324.
Compound 4fa (condensation time 1 d): orange crystals; mp
252–254 °C (DMF). 1H NMR (500 MHz, CF3CO2D): δ = 7.54–7.94
(m, 7 H), 7.96–8.12 (m, 2 H), 8.14–8.26 (m, 1 H), 8.54–8.68 (m, 1
H), 8.77 (s, 1 H). 13C NMR (125 MHz, CF3CO2D): δ =107.88 (d, JCF
= 24 Hz), 116.21, 121.92, 122.44 (d, JCF = 28 Hz), 123.81, 125.91,
125.97, 127.56, 129.45, 130.83, 131.56, 131.65, 131.74, 131.78
(d, JCF = 255 MHz), 133.63, 137.63, 140.71, 147.26, 185.80 (one C
invisible). MS (EI): m/z (%) = 376 (37), 374 (100) [M+•], 373 (31),
348 (10), 347 (19), 346 (31), 345 (38), 234 (12), 105 (43), 77
(41). HRMS (EI): m/z calcd for C22H12N2OClF: 374.0622; found:
374.0618.
Compound 4hb (condensation time 1 d): orange crystals; mp
185–187 °C (hexane–EtOAc). 1H NMR (400 MHz, CDCl3): δ =
1.46 (s, 9 H), 2.79 (s, 3 H), 7.33 (dd, J = 8.8, 2.0 Hz, 1 H), 7.65 (dd,
J = 9.2, 2.0 Hz, 1 H), 7.86–7.91 (m, 2 H), 7.92 (d, J = 2.0 Hz, 1 H),
8.17 (d, J = 9.2 Hz, 1 H), 8.30 (d, J = 2.0 Hz, 1 H). 13C NMR (100
MHz, CDCl3): δ = 26.3, 31.7, 35.0, 104.4, 113.78, 114.8, 119.0,
123.9, 124.3, 126.0, 130.4, 130.4, 132.0, 132.6, 133.2, 136.6,
146.6, 148.6, 199.4. MS (EI): m/z (%) = 352 (35), 351 (24), 350
(100) [M+•], 337 (30), 336 (20), 335 (87), 44 (32). HRMS (EI): m/z
calcd for C21H19N2OCl: 350.1186; found: 350.1190.
(10) Analytical Data for the Selected 6-Acylindolo[1,2-a]quinoxal-
ines 4
Compound 4ba (condensation time 4 d): orange crystals; mp
273–275 °C (DMF). 1H NMR (500 MHz, CF3CO2D): δ = 7.59–7.64
(m, 2 H), 7.72 (d, J = 8.7 Hz, 1 H), 7.78 (s, 1 H), 7.84–7.92 (m, 2
H), 8.00–8.04 (m, 3 H), 8.10 (d, J = 8.7 Hz, 1 H), 8.52 (d, J = 9.4 Hz,
1 H), 8.74 (s, 1 H). 13C NMR (125 MHz, CF3CO2D): δ = 115.17,
115.51, 116.38, 121.91, 123.23, 123.65, 125.72, 127.67, 129.44,
© Georg Thieme Verlag Stuttgart · New York — Synlett 2015, 26, 1352–1356

Copyright of Synlett is the property of Georg Thieme Verlag Stuttgart and its content may not
be copied or emailed to multiple sites or posted to a listserv without the copyright holder's
express written permission. However, users may print, download, or email articles for
individual use.