Fused Tetracyclic quinoxalines from reactions of o-phenylenediamines in polyphosphoric acid

Article abstract of DOI:10.1071/C96170

The condensation of 2,3-diaminobenzoic acid and the 5-chloro derivative with o-hydroxyphenylglyoxylic acid, isatin and benzothiophen-2,3-dione in polyphosphoric acid leads to the appropriate tetracycles. Isomeric products are formed from these unsymmetric

Full text of DOI:10.1071/C96170

C S I R O P U B L I S H I N G
Australian Journal
of Chemistry
Volume 50, 1997
© CSIRO Australia 1997
A journal for the publication of original research
in all branches of chemistry and chemical technology
w w w. p u b l i s h . c s i ro . a u / j o u rn a l s / a j c
All enquiries and manuscripts should be directed to
The Managing Editor
Australian Journal of Chemistry
CSIRO PUBLISHING
PO Box 1139 (150 Oxford St)
Collingwood
Vic. 3066
Australia
Telephone: 61 3 9662 7630
Facsimile: 61 3 9662 7611
Email: john.zdysiewicz@publish.csiro.au
Published by CSIRO PUBLISHING
for CSIRO Australia and
the Australian Academy of Science

Aust. J. Chem., 1997, 50, 473–478
Fused Tetracyclic Quinoxalines from Reactions
of o-Phenylenediamines in Polyphosphoric Acid
Leslie W. Deady and Anthony J. Kaye
School of Chemistry, La Trobe University, Bundoora, Vic. 3083.
The condensation of 2,3-diaminobenzoic acid and the 5-chloro derivative with o-hydroxyphenylglyoxylic
acid, isatin and benzothiophen-2,3-dione in polyphosphoric acid leads to the appropriate tetracycles.
Isomeric products are formed from these unsymmetrical diamines, and methods of assigning particular
structures are described.
As part of continuing research into polycyclic nitrogen
heterocycles as DNA-intercalating antitumour agents,
we have prepared derivatives, namely compounds (3)
and (4), of systems in which a ꢀve-membered ring
containing O, S or NH is inserted into the phenazine
system (Scheme 1). The basic tetracycles are known;
the thia and aza examples have been prepared by
reaction of a phenylene-1,2-diamine with the appro-
priate dicarbonyl compound, while the oxa analogue
has been prepared by a quite diꢁerent route.1
We had a particular interest in the carboxy sub-
stituent, speciꢀcally in the position shown. This
introduced some complexity into the reaction. Also,
when the diamine is unsymmetrical, there is the task of
assigning any product to the isomeric (3) or (4) series.
This is an extension of a long-standing problem in
quinoxaline chemistry. In the tetracyclic compounds,
previous examples have either used a symmetrical
diamine or have left the structures unresolved. This
paper reports on our studies of a common path to
all three hetero series (Scheme 1) and methods for
establishing individual structures as (3) or (4).
O
Results and Discussion
Y
NH2
NH2
Precursors
O
Isatin (1c) was the aza dicarbonyl compound, and this
was converted into o-hydroxyphenylglyoxylic acid (5).2
This can be cyclized to benzofuran-2,3-dione (1a)2,3
but we successfully used (5) itself in the condensation
reaction. The sulfur compound (1b) was prepared
from benzenethiol and oxalyl chloride.4 A four-step
preparation of 2,3-diaminobenzoic acid (2a) used an
expensive starting material5 and a more recent synthesis
was also lengthy.6 An alternative preparation, starting
from the moderately priced 2-amino-5-chlorobenzoic
acid, was devised (Scheme 2). By suitable choice of
reducing agent in the ꢀnal step, both the dechloro
(2a) and chloro (2b) diamines were accessible.
X
CO2H
(1a) X = O
(1b) X = S
(1c) X = NH
(2a) Y = H
(2b) Y = Cl
Y
N
Y
N
X
R
X
R
N
X
N
CO2R
CO2R
Y
X
Y
Cl
Cl
NO2
(3a)
(3b)
O
O
H
Cl
H
H
H
H
Et
H
(4a)
(4b)
O
O
H
H
H
H
Et
H
(i) Ac2O/AcOH
Cl
(ii) conc. HNO3/conc. H2SO4
(iii) dil. HCl
(3c) NH
(3d) NH
(4c) NH
(4d) NH
H
H
NH2
NH2
H
CO2H
CO2H
(6)
(3e) NH Cl
(4e) NH Cl
(3f) NH Cl Et
(3g)
(4f) NH Cl Et
(4g)
S
H
H
S
H
H
H2/Ni
SnCl2/HCl
(2b)
Scheme 1
Scheme 2
(2a)
Manuscript received 20 September 1996
0004-9425/97/050473$05.00
10.1071/CH96170

474
L. W. Deady and A. J. Kaye
Condensations
Formation of the thieno system is simplest11 and
there are no reports of other types of reaction prod-
ucts. This was also true in the present work, where
both aqueous acetic acid and polyphosphoric acid were
investigated as condensation media with contrasting
results. Reaction of (1b) with (2a) gave good yields
of tetracyclic products; (3g) (5 : 1) was favoured in
aqueous acetic acid, while the change to polyphos-
phoric acid reversed the preference and (4g) (3 : 1)
predominated.
Previous reactions of benzofuran-2,3-dione (1a) with
phenylene-1,2-diamine have resulted in the formation
of compound (7) rather than the desired tetracycle.7
We found that reaction in hot polyphosphoric acid was
quite successful and (7) was only a minor component.
As a model reaction, (8) was formed in 50% yield. A
single product (4a) was formed from (2a). This result
was exceptional, as inseparable isomeric mixtures were
the norm in the present work. Thus, a 1 : 1 mixture
of (4b) and (3b) was formed from (2b) (see below for
the assignments). These results can be rationalized if,
under these conditions, the 3-amino group of (2a) is
the more nucleophilic and, in the preferred reaction,
reacts with the more electophilic 3-carbonyl group of
(1a) [or the ketone C=O of (5)]. In (2b), the chlorine
meta to the 3-amino group reduces its nucleophilicity
relative to the 2-amino group so that more of the
other isomer is formed.
Structural Assignment
It was not possible to assign structures to the (3)
or (4) series from 1H or 13C n.m.r. spectra, but this
could be done for the oxa and aza compounds from
the chemical shifts produced by N -oxidation. We have
1
recently applied this technique with H and 13C n.m.r.
spectra12,13 to solve similar problems.
The systematic numbering of the tetracycles depends
on the isomer and the heteroatom, and comparisons are
therefore confusing. For the n.m.r. discussion, there-
fore, the system shown (Table 1) has been adopted.
In this, numbers refer to the A-ring, starting at the
acid-substituted position, letters apply to the D-ring,
starting ortho to the heteroatom, and ꢀ and ꢁ refer
to the ‘inner’ ring-junction carbons, with ꢀ being that
between the two heteroatoms. The diꢁerence between
(3) and (4) is therefore in the relative positions of ꢀ
and ꢁ with respect to the numbered positions. This
cannot be ascertained from the n.m.r. spectra although
the atoms can be identiꢀed, and this was a necessary
ꢀrst step. The aza ethyl ester [later established as
(4d)] was a conveniently soluble reference compound
for the 1H and 13C n.m.r. experiments required to
achieve this.
H
CO CO2H
N
O
OH
OH
(5)
N
(7)
N
N
O
7
N
4
10
NH2
1
O
(8)
N
H
Y
(9)
Z
1
14
3
15
The 1H, 13C, JC,H HETCOR
and JC,H HMBC
HN
spectra provided the data summarized in Table 1; these
allow the atoms to be assigned without recourse to
any reference spectra. For the carbons with attached
NH
N
8
O
2
N
3
4
N
protons, H 2 is the only one with JC,H coupling to the
H
Ph
3
carbonyl carbon. There is also JC,H coupling to one
Y
Z
CO2H
other hydrogen-bound carbon, which must therefore
be C 4. The remaining doublets for H a and H d can
(11)
(10a)
H
H
3
be distinguished by the extra JC,H coupling for the
(10b) Cl CO2H
latter. The upꢀeld triplet for H c is assigned from
3
The reaction of isatin with phenylene-1,2-diamine
has received considerably more study. The reaction is
complex and the Schiꢁ base (9) and spiro compound
(10a), as well as the indolo[2,3-b]quinoxaline, have
been isolated, depending on the conditions.8ꢀ10 In the
present work, reaction with (2b) in aqueous acetic acid
gave a good yield of the spiro compound (10b), but a
change to hot polyphosphoric acid brought about the
desired reaction, with only a trace of spiro contaminant.
Isomeric mixtures [(4c)-to-(3c) ratio 4 : 1, (4e)-to-(3e)
ratio 1 : 2] were obtained in both cases, with the chloro
substituent favouring (3), relative to hydrogen, as in
the furo analogues.
the JC,H coupling to C a. The close triplets for H b
3
and H 3 can be distinguished by the JC,H coupling
3
of the former to C d and the lack of JC,H coupling
between H 3 and any other hydrogen-bound carbon.
3
The quaternary carbons may be assigned from JC,H
coupling to protons already identiꢀed. This allows
distinction between C e (3JC,H to H a and H c) and
C f (3JC,H to H b and H d), C 5 (3JC,H only to H 3)
and C 6 (3JC,H to H 2 and H 4), and C ꢁ (3JC,H only
3
to H d) and C ꢀ (the only one with no JC,H).
1
The remaining H and 13C n.m.r. assignments listed
in Tables 2 and 3 were made by reference to those for
(4d); these were aided by additional proton-coupled

Fused Tetracyclic Quinoxalines
475
Table 1. C,H coupling for compound (4d) in (CD3)2SO
1
3
ꢀ denotes JC,H (HETCOR); + denotes JC,H (HMBC)
c
d
b
N
N
*
e
X
4
1
4
1
*
5
6
a
f
5
6
α
β
3
2
3
2
β
α
f
a
b
e
X
N
N
d
c
CO2H
CO2H
(3)
(4)
Carbon
H c
H a
H b
H 3
H 2
H d
H 4
7ꢁ37, t
7ꢁ55, d
7ꢁ72, t
7ꢁ74, t
8ꢁ00, d
8ꢁ35, d
8ꢁ38, d
CO
C ꢀ
C f
167ꢁ3
145ꢁ8
144ꢁ4
140ꢁ4
138ꢁ3
137ꢁ6
132ꢁ0
131ꢁ8
131ꢁ0
128ꢁ5
125ꢁ0
122ꢁ6
121ꢁ0
118ꢁ9
112ꢁ2
+
+
+
+
C ꢁ
C 5
C 6
C 4
C b
C 1
C 2
C 3
C d
C c
C e
C a
+
+
+
+
ꢀ
ꢀ
+
+
ꢀ
ꢀ
ꢀ
+
ꢀ
ꢀ
+
+
ꢀ
+
+
Table 2. 1H n.m.r. data in (CD3)2SO
13C and JC,H HETCOR experiments, and the eꢁect of
O relative to NH illustrated by spectra of dibenzofuran
and carbazole.16,17 Labelling as (3) or (4) required
additional data from N -oxidation experiments.
By using quinoline18,19 and our other work as
examples,12,13 N -oxidation results in downꢀeld shifts
in the 1H n.m.r. spectrum for ‘near’ hydrogens, for
example, H 4 and H d for reaction at Nꢁ in (4). In the
13C spectra, on the other hand, N -oxidation causes
marked upꢀeld shifts for quaternary carbons next to
the N–O function, while the more distant ring-junction
carbons are hardly aꢁected. In addition, quinoline N -
oxidation is also accompanied by a substantial upꢀeld
shift for the peri carbon; C 4 is the equivalent in the
tetracycles.
1
Cpd
H 2
H 3
H 4
H a
H b
H c
H d
(4d)
(4a)
(4a)B
(3b)
(4b)
8ꢁ00
8ꢁ18
8ꢁ21
8ꢁ15
8ꢁ15
7ꢁ74 8ꢁ38 7ꢁ55 7ꢁ72 7ꢁ37 8ꢁ35
A
A
A
A
A
A
A
D
A
A
A
A
A
D
8ꢁ44
8ꢁ73
8ꢁ35
8ꢁ48
7ꢁ61 8ꢁ36
7ꢁ60 8ꢁ47
7ꢁ61 8ꢁ34
7ꢁ61 8ꢁ34
7ꢁ60 8ꢁ44
—
—
—
(4b)B 8ꢁ21
8ꢁ66
D
D
(3c)
(4c)
(4c)B
(3e)
(4e)
(4e)B
(3g)
(4g)
8ꢁ37C 7ꢁ93
8ꢁ33C
8ꢁ44
8ꢁ44
8ꢁ23
8ꢁ11
8ꢁ28
7ꢁ85 8ꢁ51 7ꢁ63 7ꢁ77 7ꢁ43 8ꢁ37
7ꢁ78 8ꢁ79 7ꢁ55 7ꢁ69 7ꢁ39 8ꢁ54
—
—
—
8ꢁ52 7ꢁ61 7ꢁ77 7ꢁ42 8ꢁ35
8ꢁ33 7ꢁ61 7ꢁ77 7ꢁ41 8ꢁ37
8ꢁ76 7ꢁ60 7ꢁ77 7ꢁ42 8ꢁ58
8ꢁ45 8ꢁ19 7ꢁ82 7ꢁ69 8ꢁ35C
8ꢁ32C 8ꢁ0
8ꢁ24
7ꢁ96 8ꢁ50 8ꢁ19 7ꢁ82 7ꢁ69 8ꢁ43
A
Occurred as a multiplet at ꢂ 7ꢁ8–7ꢁ9.
B
C
The N -oxidation product.
Peaks bearing this superscript within a row may be inter-
There are four possible structures for the N -oxide
from a compound which could be either (3) or (4).
However, in the present work, all N -oxides showed
changed.
D Peaks are obscured by the signals of the corresponding protons
of the major product (4c).
Table 3. 13C n.m.r. data in (CD3)2SO
C 5 C 6 C ꢀ C ꢁ C a C b
137ꢁ6 145ꢁ8 140ꢁ4
Cpd
C 1
C 2
C 3
C 4
C c
C d
C e
C f
CO
(4d)
(4c)
131ꢁ0 128ꢁ5
132ꢁ7 132ꢁ2
125ꢁ0 132ꢁ0 138ꢁ3
125ꢁ5 133ꢁ3 138ꢁ0
112ꢁ2 131ꢁ8
112ꢁ5 132ꢁ2
121ꢁ0 122ꢁ6 118ꢁ9 144ꢁ4
121ꢁ7 122ꢁ6 118ꢁ9 144ꢁ2
167ꢁ3
166ꢁ4
137ꢁ2 144ꢁ3 141ꢁ7
(4c)A 133ꢁ7 133ꢁ2B 125ꢁ1 123ꢁ6 125ꢁ0C 139ꢁ8D 148ꢁ0 126ꢁ0C 111ꢁ9 131ꢁ3B 121ꢁ8 122ꢁ4 114ꢁ8 140ꢁ6D 166ꢁ5
(4a)
131ꢁ6 130ꢁ3
127ꢁ9 132ꢁ0 140ꢁ4
136ꢁ4
138ꢁ1
155ꢁ1 141ꢁ1
158ꢁ1 124ꢁ8
113ꢁ1 133ꢁ5
125ꢁ1 123ꢁ1 120ꢁ6 158ꢁ3
165ꢁ8
165ꢁ8
(4a)A 131ꢁ5 131ꢁ8B 127ꢁ0 120ꢁ4 135ꢁ6
The N -oxidation product.
111ꢁ6 131ꢁ6B 124ꢁ6 123ꢁ0 116ꢁ5 154ꢁ3
A
B,C,D
Resonances bearing the same capital superscript within a row may be interchanged.

476
L. W. Deady and A. J. Kaye
appropriate shifts for C 4 and H 4. Thus, the steric
eꢁect of the 1-CO2H group restricted oxidation to
Nꢁ, so that the possibilities were now reduced to
the two structures shown and the expectations from
N -oxidation are as follows. For (4), upꢀeld shifts are
expected for C 4, C 5 and C ꢀ, whereas downꢀeld shifts
for H 4 and H d are expected. For (3), upꢀeld shifts
for C 4, C 5 and C ꢁ, and downꢀeld shifts for H 4, are
expected.
Experimental
N.m.r. spectra were recorded on Bruꢀker AM-300 (300ꢁ13
and 75ꢁ47 MHz for 1H and 13C, respectively) and Bruꢀker DRX-
400 (400ꢁ13 and 100ꢁ62 MHz) spectrometers, in (CD3)2SO
unless otherwise stated. The electrospray mass spectra were
obtained on a VG Bio-Q triple quadrupole mass spectrometer
by using a water/methanol/acetic acid (50 : 50 : 1) mobile phase.
Microanalyses were performed at the Campbell Microanalytical
Laboratory, University of Otago, New Zealand.
The (13C–1H) HETCOR experiment was performed by using
the pulse sequence described by Bax and Morris.14 The refo-
cusing delay was optimized to 160 Hz (3ꢁ0 ms). The spectrum
was acquired as 512ꢂ128 data points, zero-ꢁlled and subjected
to both Fourier transforms to aꢂord the 1024ꢂ1024 point data
matrix. The number of transients per t1 increment was 128.
Spectral widths were 4761 Hz in F1 (1H) and 14705 Hz in
Though some signal overlap occurred in the 1H n.m.r.
spectra, the most downꢀeld peaks, which included those
for H 4 and H d, were always distinguishable. As an
example, inspection of the results from Tables 2 and
3 for the single compound formed in the dechloro
oxa case indicates N -oxidation shifts only compatible
with this being (4a). This was true also for the aza
analogue (4c).
F2
(
(
13C). The 90ꢀ pulse widths were 14ꢁ0 (1H) and 13ꢁ5 ꢃs
13C).
The long-range proton-detected (three bond) (1H–13C) het-
eronuclear multiple bond correlation (HMBC) experiment used
the pulse sequence described by Bax and Summers.15 The low
pass J-ꢁlter portion of the experiment was set for an average
one-bond heteronuclear coupling of 150 Hz (3ꢁ3 ms). The
long range delay utilized to excite the heteronuclear multiple-
quantum coherence was set for 8ꢁ3 Hz (60 ms). The spectrum
was acquired as 2 Kꢂ267 data points, zero-ꢁlled and subjected
to both Fourier transforms to aꢂord the 1024ꢂ1024 point data
matrix. The number of transients per t1 increment was 32.
This method of structure assignment could still be
used where mixtures of the isomers (3) and (4) were
present. It was beneꢀcial in fact to have both iso-
mers available for reaction under identical conditions,
and only the isomer (4) reacted. So, for example,
the mixture in the chloro oxa case after oxidation
showed distinguishable peaks for the N -oxide of (4b)
and unchanged (3b). This selectivity conforms with
the knowledge that 2-methoxyquinoxaline undergoes
oxidation at N 4.20
Spectral widths were 25062 Hz in F1 (
13C) and 5841 Hz in
F2 (1H). The 90ꢀ pulse widths were 9ꢁ33 (1H) and 10ꢁ4 ꢃs
(
13C) and a 1 s interpulse delay was employed.
The N -oxidation approach failed for the sulfur case.
While there are examples of N -oxidation in compounds
with adjacent nitrogen- and sulfur-containing rings,21
no success was achieved with the product [(3g) or
(4g)] of the acetic acid condensation described above.
Proton n.m.r. spectra were complex but always showed
upꢀeld shifts, which suggested breakdown of the ring
system. An alternative approach made use of the
ready desulfurization of many compounds, including
dibenzothiophen, with Raney nickel.22 The sulfur tetra-
cycle, with deactivated catalyst and careful control
of conditions (it over-reacted very easily), aꢁorded a
phenylquinoxalinecarboxylic acid. This was character-
ized by the appearance of the hetero-ring CH singlet
in the 1H n.m.r. spectrum, and the proton-coupled
13C spectrum allowed it to be identiꢀed as (11). In
particular, C 4a (138ꢀ3 ppm) and C 8a (140ꢀ8) were
assigned by reference to quinoxalines.23 The latter was
a well resolved triplet (3JC,H2,7 10ꢀ0 Hz), the former
less so, and this multiplicity is incompatible with the
alternative 2-phenyl isomer (C 8a should be a doublet).
Thus the major tetracyclic product from the acetic
acid condensation is (3g).
2-Amino-5-chloro-3-nitrobenzoic Acid (6)
2-Amino-5-chlorobenzoic acid (Maybridge Chemical Com-
pany; 5 g) was acetylated with a 1 : 1 mixture of acetic anhydride
(10 ml) and glacial acetic acid (10 ml) (reꢃuxed for 0ꢁ5 h
and poured into water) and the resulting amide had a m.p. of
190–192ꢀ. Nitration was carried out by adding the amide (2ꢁ0
g) in small portions over 45 min to a solution of concentrated
nitric acid (4 ml) and concentrated sulfuric acid (4 ml) at 0ꢀ,
and the mixture was stirred for a further 1 h then poured
into 100 ml of ice-water. With continued stirring, the initially
sticky precipitate gave a yellow solid. This was ꢁltered oꢂ to
give the nitro amide as a pale yellow powder (2ꢁ0 g), m.p.
195–197ꢀ. Hydrolysis of the amide was carried out by heating
with concentrated hydrochloric acid/water (1 : 4; 50 ml) for 1 h,
during which time a solid separated. The cooled solution was
ꢁltered to give (6) as yellow needles (1ꢁ40 g), m.p. 238–240ꢀ
[from light petroleum (b.p. 90–110ꢀ)/chloroform] (Found: C,
38ꢁ9; H, 2ꢁ4; N, 12ꢁ8. C7H5CIN2O4 requires C, 38ꢁ8; H, 2ꢁ3;
N, 12ꢁ9%). 1H n.m.r. [(CD3)2SO/CDCl3] ꢂ 8ꢁ10, s, 1H; 8ꢁ19,
s, 1H.
2,3-Diamino-5-chlorobenzoic Acid (2b)
The work described in this paper shows that polyphos-
phoric acid is a convenient medium to bring about
condensation of o-phenylenediamines with isatin and
its oxa and thia analogues. In addition, a detailed
n.m.r. analysis of the products provides a useful data
base, and convenient methods have been found which
enable isomeric structures of the resulting tetracycles
to be determined.
A mixture of compound (6) (2ꢁ3 g), stannous chloride
dihydrate (9ꢁ0 g) and concentrated hydrochloric acid (30 ml)
was stirred and heated at 100ꢀ until the clear solution was no
longer yellow. The white precipitate which formed on cooling
was collected by ꢁltration to give a hydrochloride salt of the
product (2b) (1ꢁ8 g), m.p. 244–246ꢀ (from hydrochloric acid)
(Found: C, 37ꢁ8; H, 3ꢁ9; N, 12ꢁ7. C7H7ClN2O2.HCl requires
C, 37ꢁ7; H, 3ꢁ6; N, 12ꢁ6%). 1H n.m.r. ꢂ 7ꢁ23, d, J 2ꢁ4 Hz,
1H; 7ꢁ47, d, J 2ꢁ4 Hz, 1H.

Fused Tetracyclic Quinoxalines
477
2,3-Diaminobenzoic Acid (2a)
reduced pressure to 4 ml, diluted with water to 30 ml and the
pH adjusted to 3 with 10% sodium hydroxide. The solid which
separated was ꢁltered oꢂ to give the ethyl esters (4d) and (3d)
(>19 : 1 by 1H n.m.r.) (0ꢁ22 g), m.p. 228–230ꢀ (from ethanol)
(Found: C, 70ꢁ2; H, 4ꢁ2; N, 14ꢁ2. C17H13N3O2 requires C,
70ꢁ1; H, 4ꢁ5; N, 14ꢁ4%).
A mixture of compound (2b) (0ꢁ5 g) and freshly prepared
Raney nickel24 (2 g of ethanol-wet material) in 0ꢁ2 M potassium
hydroxide in ethanol (50 ml) was hydrogenated at atmospheric
pressure. The catalyst was removed by ꢁltration through Celite,
the ꢁltrate was concentrated to 5 ml, diluted to 30 ml with
water and acidiꢁed to pH 2 with concentrated hydrochloric
acid. The solvent was removed at reduced pressure to give the
product (2a) as a hydrochloride salt in a mixture (1ꢁ6 g) with
potassium chloride. 1H n.m.r. (D2O) ꢂ 6ꢁ74, t, H 5; 7ꢁ40, d, J
7ꢁ9 Hz, H 4; 7ꢁ81, d, J 8ꢁ2 Hz, H 6. This mixture was used
in the condensation reactions given below.
2-Chloro-6 H-indolo[2,3-b]quinoxaline-4-carboxylic Acid
(4e) and 3-Chloro-6 H-indolo[2,3-b]quinoxaline-1-carboxylic
Acid (3e)
Reaction of the hydrochloride of the diamine (2b) (0ꢁ5 g)
with isatin (0ꢁ6 g) and polyphosphoric acid (15 g) as for (4c)
gave 0ꢁ6 g of a dark solid, which contained a 2 : 1 mixture of
(3e) and (4e), after the same base/acid treatment. Extraction
with hot tetrahydrofuran gave a yellow solid (0ꢁ22 g), m.p.
>300ꢀ, with the same composition.
This was esteriꢁed as for (4c) to give a mixture of the
ethyl esters as a tan solid, m.p. 220–225ꢀ (from ethanol/water)
(Found: C, 62ꢁ4; H, 3ꢁ5; N, 12ꢁ7. C17H12ClN3O2 requires C,
62ꢁ7; H, 3ꢁ7; N, 12ꢁ9%).
Benzofuro[2,3-b]quinoxaline (8)
Phenylene-1,2-diamine (0ꢁ5 g) was combined witho-hydroxy-
phenylglyoxylic acid2 (0ꢁ5 g) and polyphosphoric acid (15 g).
The mixture was stirred and heated at 110ꢀ for 5 h, then
cooled to room temperature and thoroughly mixed with water
(200 ml). The yellow solid which separated was collected by
ꢁltration and recrystallized from ethanol to give the product
(8) (0ꢁ4 g, 50%), m.p. 169–170ꢀ (lit.1 171–172ꢀ). 1H n.m.r. ꢂ
7ꢁ59, t, H 2; 7ꢁ81–7ꢁ94, m, H 3,4,8,9; 8ꢁ14, d, H 7; 8ꢁ30, d,
H 10; 8ꢁ35, d, H 1.
[1]Benzothieno[2,3-b]quinoxaline-10-carboxylic Acid (3g)
and [1]Benzothieno[2,3-b]quinoxaline-7-carboxylic Acid (4g)
(A) A warm solution of 2,3-diaminobenzoic acid/potassium
chloride (1ꢁ6 g) in acetic acid/water (1 : 1; 10 ml) was added,
with stirring, to a hot solution of benzothiophen-2,3-dione4
(0ꢁ52 g) in glacial acetic acid (20 ml). The reaction mixture
was warmed and stirred for 15 min and the resultant precipitate
was ꢁltered oꢂ to yield a khaki solid (0ꢁ34 g, 38%). 1H n.m.r.
analysis showed this to contain (3g) and (4g) (5 : 1), with m.p.
range of 240–261ꢀ after recrystallization from ethanol (Found:
C, 64ꢁ4; H, 3ꢁ1; N, 10ꢁ3. C15H8N2O2S requires C, 64ꢁ3; H,
2ꢁ9; N, 10ꢁ0%).
Benzofuro[2,3-b]quinoxaline-7-carboxylic Acid (4a)
A mixture (1ꢁ6 g) of 2,3-diaminobenzoic acid/potassium
chloride [from 0ꢁ5 g of (6)] and o-hydroxyphenylglyoxylic acid2
(0ꢁ5 g) in polyphosphoric acid (15 g) was stirred and heated at
110ꢀ for 5 h, then cooled to room temperature and thoroughly
mixed with water (100 ml). The green solid (0ꢁ39 g) which
separated was ꢁltered oꢂ, washed with water and recrystallized
from acetonitrile to give the product (4a) (0ꢁ2 g). For micro-
analysis, a sample was stirred with acetic acid/water (1 : 1),
which removed the small amount of (7), ꢁltered and again
recrystallized from acetonitrile to give the acid (4a), m.p. c.
270ꢀ (after shrinking and darkening) (Found: C, 68ꢁ1; H, 2ꢁ9;
N, 10ꢁ6. C15H8N2O3 requires C, 68ꢁ2; H, 3ꢁ1; N, 10ꢁ6%).
(B) The conditions in polyphosphoric acid were as for (4a)
and gave a 75% yield of a mixture of (4g) and (3g) (3 : 1) as a
green solid, m.p. 237–240ꢀ.
3-Phenylquinoxaline-5-carboxylic Acid
8-Chlorobenzofuro[2,3-b]quinoxaline-10-carboxylic Acid (3b)
and 9-Chlorobenzofuro[2,3-b]quinoxaline-7-carboxylic Acid (4b)
A mixture of (3g) and (4g) [5 : 1; from the acetic acid
preparation (A) above] (0ꢁ05 g), sodium carbonate (0ꢁ1 g)
and deactivated Raney nickel25 (5 g, acetone-wet) in water
(4 ml) was stirred vigorously and heated at 70ꢀ for 3 h.
The catalyst was removed by ꢁltration through Celite, the
ꢁltrate was acidiꢁed to pH 2 with concentrated hydrochloric
acid and the solid which formed was ꢁltered oꢂ. This was
recrystallized from ethanol to give a little of the unchanged
tetracycle, and removal of the ethanol from the ꢁltrate gave the
title product (0ꢁ015 g) as pale yellow needles, m.p. 206–208ꢀ
(from methanol). 1H n.m.r. ꢂ 7ꢁ65–7ꢁ69, m, 4H; 7ꢁ97, t, 1H;
8ꢁ31–8ꢁ38, m, 4H; 9ꢁ74, s, H 2. 13C n.m.r. ꢂ 127ꢁ5, CH; 129ꢁ1,
CH; 129ꢁ3, CH; 130ꢁ9, CH; 132ꢁ7, CH; 134ꢁ7, C; 138ꢁ3, C;
140ꢁ8, C; 144ꢁ4, CH; 150ꢁ1, C; 165ꢁ9, C. Electrospray mass
spectrum: m/z 251 (M+1).
The hydrochloride (0ꢁ44 g) of the diamine (2b) and o-
hydroxyphenylglyoxylic acid2 (0ꢁ5 g) in polyphosphoric acid
(15 g) were reacted as for (4a) to give a light green solid (0ꢁ24
g, 27%), m.p. 237–240ꢀ (from acetonitrile). This was shown by
1H n.m.r. to be a c. 1 : 1 mixture of (3b) and (4b), containing a
trace of the ring-opened product which could not be removed.
Electrospray mass spectrum: m/z 299 (100%), 300 (16), 301
(36); all (M+1).
6 H-Indolo[2,3-b]quinoxaline-4-carboxylic Acid (4c)
2,3-Diaminobenzoic acid/potassium chloride (1ꢁ6 g) [from
0ꢁ5 g of (6)] was combined with isatin (0ꢁ6 g) and polyphos-
phoric acid (15 g). The mixture was stirred and heated at 140ꢀ
for 5 h, then cooled to room temperature and thoroughly mixed
with water (200 ml). The solid which separated was ꢁltered
oꢂ and washed thoroughly with water, to give a black glass
(0ꢁ75 g). This was treated with 1 M sodium hydroxide (50 ml),
ꢁltered, and the ꢁltrate acidiꢁed to pH 2 with concentrated
hydrochloric acid. The product was collected by ꢁltration to
give 0ꢁ6 g of a dark solid. 1H n.m.r. analysis suggested that
this was largely product but it could be further puriꢁed by
Soxhlet extraction with tetrahydrofuran to give a yellow solid,
m.p. >300ꢀ. This contained the title compound (4c) and the
isomeric (3c) in a 4 : 1 ratio, with trace impurities.
N-Oxidation Procedure
A sample of the acid (c. 0ꢁ08 g) was dissolved in hot glacial
acetic acid (4 ml). Hydrogen peroxide solution (30% v/v; 1ꢁ25
ml) was added and the reaction mixture was heated under
reꢃux for the time indicated below, when substantial separation
of solid had occurred. The mixture was cooled, and the solid
ꢁltered oꢂ and analysed by n.m.r. The following results were
obtained:
Acid
Time
Products
For microanalysis, a sample of the dark solid (0ꢁ35 g) in
ethanol (12 ml) and concentrated sulfuric acid (2 ml) was heated
under reꢃux for 2ꢁ5 h, then cooled, and the insoluble material
was removed by ꢁltration. The ꢁltrate was concentrated at
(3c)/(4c)
(3e)/(4e)
(4a)
2 h Mixture of the N-oxide of (4c) and unchanged (3c)
2 h Mixture of the N-oxide of (4e) and unchanged (3e)
2 h N-oxide of (4a) as yellow needles, m.p. 269–270ꢀ
(3b)/(4b) 48 h Mixture of the N-oxide of (4b) and unchanged (3b)

478
L. W. Deady and A. J. Kaye
10
11
Popp, F. D., J. Heterocycl. Chem., 1969, 6, 125.
Banerji, K. D., Mazumdar, A. K. D., and Sen, K. K., J.
Ind. Chem. Soc., 1973, 50, 268.
Deady, L. W., Desneves, J., and Ross, A. C., Tetrahedron,
1993, 49, 9823.
Deady, L. W., and Quazi, N. H., Spectrosc. Lett., 1995, 28,
1033.
Bax, A., and Morris, G., J. Magn. Reson., 1981, 42, 501.
Bax, A., and Summers, M. F., J. Am. Chem. Soc., 1986,
108, 2093.
Black, P. J., and Heꢂernan, M. L., Aust. J. Chem., 1965,
18, 353.
Giraud, J., and Marzin, C., Org. Magn. Reson., 1979, 12,
647.
Barbieri, G., Benassi, R., Lazzeretti, P., Schenetti, L., and
Taddei, F., Org. Magn. Reson., 1975, 7, 451.
Su, J.-A., Siew, E., Brown, E. V., and Smith, S. L., Org.
Magn. Reson., 1978, 11, 565.
Cheeseman, G. W. H., J. Chem. Soc., 1961, 1246.
Klemm, L. H., Mathur, S. B., Zell, R., and Merrill, R. E.,
J. Heterocycl. Chem., 1971, 8, 931.
Blicke, F. F., and Sheets, D. G., J. Am. Chem. Soc., 1949,
71, 4010.
McNab, H., J. Chem. Soc., Perkin Trans. 1, 1982, 357.
Pavlic, A. A., and Adkins, H., J. Am. Chem. Soc., 1946,
68, 1471.
Spero, G. B., McIntosh, A. V., and Levin, R. H., J. Am.
Chem. Soc., 1948, 70, 1907.
Acknowledgments
We thank Dr M. Sadek and Mr I. Thomas for recording
the HMBC spectrum and mass spectra, respectively, and
Mr D. Pummeroy and Mr G. L. Butt for contributions
to the synthesis of compound (6).
12
13
14
15
References
1
16
17
18
19
Zamet, J.-J., Haddadin, M. J., and Issidorides, C. H., J.
Chem. Soc., Perkin Trans. 1, 1974, 1687.
Huntress, E. H., and Hearon, W. M., J. Am. Chem. Soc.,
1941, 63, 2762.
2
3
Russell, G. A., Myers, C. L., Bruni, P., Neugebauer, F. A.,
and Blankespoor, R., J. Am. Chem. Soc., 1970, 92, 2762.
Papa, D., Schwenk, E., and Ginsberg, H. F., J. Org. Chem.,
1949, 14, 723.
Jones, J. B., and Taylor, K. E., Can. J. Chem., 1977, 55,
1653.
Denny, W. A., Rewcastle, G. W., and Baguley, B. C., J.
Med. Chem., 1990, 33, 818.
Logemann, W., Cavagna, G., and Tosolini, G., Chem. Ber.,
1963, 96, 2248.
Niume, K., Kurosawa, S., Toda, F., Hasegawa, M., and
4
5
20
21
6
22
7
23
24
8
Iwakura, Y., Bull. Chem. Soc. Jpn, 1982, 55, 2293.
Ivaschenko, A. V., Drushlyak, A. G., and Titov, V. V.,
Chem. Heterocycl. Compd., 1984, 537.
9
25