Novel spiro and fused heterocycles from the allylation of indigo

Article abstract of DOI:10.1016/j.tetlet.2009.09.098

The allylation of indigo results in the one-step synthesis of two unique complex heterocyclic systems: a spiroindoline-pyridoindolone arising from the addition of three allyl moieties and a fused pyridoindolo-azepinoindolone generated from the addition and subsequent cyclisation of two allyl moieties. The structures of these novel heterocycles are assigned unambiguously using extensive NMR experiments and by X-ray crystallographic analysis. The distribution of the products is influenced by the use of thermal versus microwave heating. Crown Copyright

Full text of DOI:10.1016/j.tetlet.2009.09.098

Tetrahedron Letters 50 (2009) 6947–6950
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Novel spiro and fused heterocycles from the allylation of indigo
b
a,
Mohammed K. Abdel-Hamid a, John B. Bremner a, , Jonathan Coates , Paul A. Keller
,
*
*
Celia Miländer a, Yasmine S. Torkamani a, Brian W. Skelton c, Allan H. White c, Anthony C. Willis d
a School of Chemistry University of Wollongong, Wollongong, NSW 2522, Australia
b Avexa Limited, 576 Swan St, Richmond, Vic 3121, Australia
c School of Biomedical, Biomolecular and Chemical Sciences, University of Western Australia, Nedlands, WA 6009, Australia
d Research School of Chemistry, The Australian National University, Canberra, ACT 0200, Australia
a r t i c l e i n f o
a b s t r a c t
Article history:
The allylation of indigo results in the one-step synthesis of two unique complex heterocyclic systems: a
spiroindoline–pyridoindolone arising from the addition of three allyl moieties and a fused pyridoindolo-
azepinoindolone generated from the addition and subsequent cyclisation of two allyl moieties. The struc-
tures of these novel heterocycles are assigned unambiguously using extensive NMR experiments and by
X-ray crystallographic analysis. The distribution of the products is influenced by the use of thermal versus
microwave heating.
Received 3 August 2009
Revised 9 September 2009
Accepted 18 September 2009
Available online 24 September 2009
Keywords:
Indigo
Crown Copyright Ó 2009 Published by Elsevier Ltd. All rights reserved.
Domino reaction
Azepinoindolone
Pyridoindolone
Spiroindoline
Despite the venerable lineage of indigo and the extensive
knowledge of its chemical properties,1 N- and N,N0-alkylation and
subsequent reactions of this compound are not well developed.
This can be explained partly by the limited solubility of indigo it-
self, and also partly by early reports that indigo could not be alkyl-
ated2 or formed complex mixtures3 on such reactions. This last
perception was altered, however, by the later report4 of a near
quantitative formation of N,N0-dimethylindigo on reaction of indi-
go in DMF (with a trace of water) with iodomethane in the pres-
ence of barium oxide as a base. Other cis-N,N0-bridged indigo
derivatives, for example, the 1,2-ethano,5 1,3-propano5,6 and 1,4-
butano systems,5 together with a mono N-benzyl derivative,7,8
can also be accessed by base-induced (NaH) N-alkylation reactions.
Added impetus for further investigation arose from the relatively
recent discovery of the chlorinated indigo-N-glycosides, Akashin
A, B and C from a Streptomyces sp.;9 these compounds have signif-
icant anti-tumour activity against a range of human tumour cell
lines.10
2,20-bis-indolic system. In this context we investigated the allyla-
tion of indigo, as this was expected to provide derivatives with bet-
ter solubility profiles as well as to introduce useful functionality for
subsequent reactions, for example, metathesis reactions. The ally-
lations resulted in the serendipitous and unprecedented one-pot
formation of two unusual heterocyclic systems and the results
are reported in this Letter.
Initially, a suspension of indigo 1 and caesium carbonate in
anhydrous DMF at 70 °C was treated with allyl bromide (5 mol e-
quiv). After filtration and silica gel chromatographic separations,
the N-allylisatin 212 was isolated in 25% yield along with the 1-al-
lyl-50-allyloxy-30,40-dihydrospiroindoline-pyrido[1,2-a]indol-one 3
(32% yield) and the pyridoindolo-azepino[1,2-a]indol-11(7H)-one
4 (11% yield). Indigo was recovered (23%) giving an overall atom
return of 91% (Table 1, entry 1).13
Initial sonication of 1 in DMF ensuring maximal dissolution was
required for optimal yields. Our subsequent attempts at reaction
optimisation involved the use of a microwave reactor14 under sealed
reaction conditions. Surprisingly, this resulted in the seven-mem-
bered structure 4 being the major product and the spiro heterocycle
3 being the minor product (Table 1, entry 2). This inversion of major
product to minor product was also observed for entries 3 and 4. Fur-
ther extensions of the reaction involved the use of different allyl sub-
stituents, and the results are summarised in Table 1. In all cases, the
spiro structure B was formed with the exception of the gem-di-
methyl allyl examples (entries 9 and 10), where presumably steric
interactions prevented the cyclisation to either structure. Here, only
As part of a continuing programme investigating functionalised
biaryls as molecules with potential bioactivities and subsequent
therapeutic uses,11 we became interested in the N-substitution
reactions of the readily available dye, indigo, with its embedded
* Corresponding authors. Tel.: +61
2 4221 4692/4255; fax: +61 4221 4287
(P.A.K.).
E-mail addresses: jbremner@uow.edu.au (J.B. Bremner), keller@uow.edu.au (P.A.
Keller).
0040-4039/$ - see front matter Crown Copyright Ó 2009 Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2009.09.098

6948
M. K. Abdel-Hamid et al. / Tetrahedron Letters 50 (2009) 6947–6950
Table 1
The allylation of indigo to produce allylisatins (A), spiroindoline–pyridoindolones (B),
and pyridoindolo-azepinoindolones (C)
O
R3
O
H
O
N
R2
N
1. DMF, Cs2CO3, ))) 2.
Br
R1
N
H
O
R3
1
2
R1
O
R2
R1
O
O
O
N
N
+
+
R1
R2
R1
N
N
O
6
N
R3
R2
R2
2'''
2''
1
R1
R3
R3
O
R1
9'
7'
C
10'
4
A
B
N
3
Entry R1
R2
R3
D
1 %
Thermal 23
wave 16
Thermal 18
A %
B %
C %
O
N
H
1
2
3
4
5
6
7
8
9
H
H
H
H
H
H
H
H
H
H
H
H
H
H
2
25
14
22
12
14
6
3
3
6
6
9
9
32
11
36
10
42
12
4
4
7
7
11
38
9
Me
4'
3'
l
2
5
5
8
8
9
Me
Me
H
l
wave
Thermal
wave
15
6
17
35
Me
Me
Ph
Ph
H
H
H
H
l
a
Thermal 21
—
10 12b
10 9b
a
lwave 15
—
N
Me Me Thermal 63
Me Me wave 58
11 14
11
10
H
l
3
N
a
O
(—) indicates that this product was not isolated from the reaction.
b
4
Approximately 5% of an isomer probably arising from olefin migration is
present as indicated by 1H NMR analysis.
Figure 1. X-ray structures of N-allylisatin 2, spiroindoline–pyridoindol-one 9 and
the pyridoindole-azepinoindolone 4.15
the allylisatin 11 and unreacted indigo were isolated. In the cases of
mono-terminal substitution of the allyl derivatives (entries 5–8),
steric hindrance also appears to be playing a role, with the formation
of the corresponding spirocyclic structures 9 and 10 but with no evi-
dence of the fused heterocycles.
A proposed mechanism for the formation of 3 and 4, as models
for compounds of types B and C, respectively, is outlined in Scheme
1. After the initial formation of the monoallylated indigo intermedi-
ate I from allylation of the indigo mono anion, deprotonation with
subsequent enolisation allows for further O-allylation (Path A). A
third allylation on the imine nitrogen in II might then take place
providing a powerful electrophilic site for internal double bond at-
tack and subsequent proton loss to give the spirocyclic derivative 3
of the type B. Alternatively, base-induced isomerisation of N-ally-
lindigo I could take place to give the 3-hydroxyindolic tautomer
in equilibrium with its 3-indolone tautomer (Path B). Quaternisa-
tion of the imine nitrogen in either tautomer by reaction with allyl
bromide could then give the intermediate III which is set up for
internal nucleophilic attack by the electron-rich alkene on the elec-
tron-deficient iminium ion to afford the spiro system IV and then 3
(Path C). Alternatively, III could form the stabilised allylic zwitter-
ion V followed by intramolecular nucleophilic addition on the indo-
lone carbonyl. Subsequent loss of water and final cyclisation would
realise the azepino-indolone 4 of the type C. The diastereoselectiv-
ity observed with the spiro compounds 9 and 10 presumably arises
from the necessarily constrained approach of the nucleophilic
(E)-allylic double bond from both the top and the bottom face in
the relevant intermediates of type III or VI. This would afford the
20R,2S:20S,2R enantiomeric pair of isomers in the case of 9, and
the 20S,2S:20R,2R pair for 10. However, the reasons for the differ-
ences in yields with microwave versus thermal heating are not
readily apparent at this stage.
The structure of 3 was initially determined using an array of
NMR experiments. Foremost was the differentiation between the
possible positions of the double bond within the pyrido ring. Con-
firmation of the presence of the enamine arose from the observa-
tion of key correlations including the diastereotopic protons H2A0
and H20B coupling to the spiro carbon as indicated by analysis of
the gHMBC spectrum; the latter carbon was assigned as quaternary
from the DEPT spectrum analysis. Final structural confirmation of
the products came from X-ray crystallographic analysis of N-allyli-
satin 2, the spiroindolone 9 and the azepino-indolone 4 (Fig. 1).
For the spiro derivative 9, with two stereogenic atoms at posi-
tions 20 and 2, potential for diastereomers existed. However, in
the 13C NMR spectrum of 9 no doubling of peaks was apparent
and therefore it is likely that only one pair of enantiomers is pres-
ent. The crystal structure confirmed the relative stereochemistry
with the C20 proton substituent cis to the indolone N-allyl substitu-
ent and this disposition was reflected in the observation of an NOE
interaction between H20 and the allyl methylene group. Similarly,
with the spiro derivative 10 only one diastereomer was isolated.
Further support for the structure 3, rather than the alternative
adduct where the double bond is in the 20,30-position, came from
analysis of the calculated minimum energies of each.16 The conju-
gated enamine 3 had a minimum energy of 124.086 kcal/mol ver-
sus 127.621 kcal/mol for the non-conjugated analog.
The heterocyclic systems embodied in compounds of types B
and C have not been described previously. However, in the case
of B, the pyrido[1,2-a]indolic moiety is well known and derivatives
are of considerable pharmacological interest.17 There are no rela-
tives of system C, although a recent paper described a weakly cyto-
toxic indolic azafulvene marine natural product, pseudocerosine,
with partial similarities.18
Formation of the N-allylisatin 2 presumably arises from oxida-
tive cleavage of the N-allylindigo intermediate and/or from N,N0-
diallylindigo which could be readily formed from the former under
basic conditions with allyl bromide. Oxidative cleavage of N,N-dial-
kylindigo derivatives has been reported previously although an
oxidant (nitric acid) was used.3 Oxygen from the air may be an
oxidant in the current reactions although when the reaction with

M. K. Abdel-Hamid et al. / Tetrahedron Letters 50 (2009) 6947–6950
6949
O
N
OH
O
N
H
N
1. Cs2CO3
2. HCO3
H
N
N
allylation
indigo
N
O
O
Path B
O
I
Br
Path A
Br
O
O
H
N
O
N
O
Cs2CO3
-H
H
N
N
N
O
N
O
III
V
Path C -H
1. Nu Addn
2. HCO3
O
N
N
O
OH
O
HO
N
H
N
O
N
N
IV
II
1. Cs2CO3
2.
Br
1. -H2O
2. Cyclisation
3. -H
O
N
O
N
N
O
-H
O
N
O
N
)
N
3
B
(type
VI
4
C
)
(type
Scheme 1. Proposed mechanistic scheme for the formation of compounds of types B and C.
allyl bromide was repeated under nitrogen no change in the yield
of 2 was seen.
rad Operations Pty Ltd., and Avexa Ltd. for support. The award of
a University of Wollongong URC Matching Scholarship (Y. Torka-
mani) and the award of a University of Wollongong Postgraduate
Award (M. Abdel-Hamid) are also gratefully acknowledged.
In support of the above-mentioned mechanistic proposals, mass
spectrometric evidence for the presence of low yields of N-allyl- and
N,N0-diallylindigo was seen in the crude reaction mixture. Also, reac-
tion of N-allylindigo separately with allyl bromide and caesium car-
bonate in DMF at the lower temperature of 60 °C resulted in the
formation of N,N0-diallylindigo and compounds 2, 3 and 4.
In summary, a new illustration of atom efficiency and domino
reaction-based conciseness in heterocyclic synthesis has been re-
vealed through the preparation of two unique, structurally com-
plex and potentially biologically interesting heterocyclic systems
from the one-step reaction of the readily available dye, indigo, with
allylic bromides in the presence of base. Given the structural com-
plexity of the heterocycles of types B and C, any attempts at a ret-
rosynthetic analysis would have been unlikely to have proposed
indigo as the starting material in such a one-pot process. It is the
facile synthetic access to these systems, coupled with their associ-
ated functional groups, which provide scope for a range of further
transformations including cyclisation reactions to produce more
complex, and potentially useful heterocycles. Scope also exists to
use other readily available indigo derivatives with different ring
substituent groups, for example, indigo carmine, to broaden the
product properties, a feature likely to be of importance in future
biological activity studies.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.tetlet.2009.09.098.
References and notes
1. Johnson, A. J. In The Chemistry of Synthetic Dyes and Pigments; Lubs, H. A., Ed.;
ACS Monograph Series No. 127; Reinhold: New York, 1955; pp 551–576.
2. Baeyer, A. Chem. Ber. 1883, 16, 2188–2204.
3. Pummerer, R.; Meininger, F. Liebigs Ann. Chem. 1954, 590, 173–194; See also:
Pummerer, R.; Meininger, F.; Schrott, G.; Wagner, H. Liebigs Ann. Chem. 1954,
590, 195–215.
4. Kuhn, R.; Trischmann, H. Chem. Ber. 1961, 94, 2258–2263.
5. Setsune, J.-i.; Fujiwara, T.; Kitao, T. Chem. Express 1986, 1, 299–302.
6. Setsune, J.-i.; Matsura, T.; Fujiwara, T.; Kitao, T. Chem. Lett. 1984, 1755–1758.
7. Hein, M.; Michalik, D.; Langer, P. Synthesis 2005, 20, 3531–3534.
8. Meng, J.-B.; Li, Y.-Z.; Xu, L. L.; Wang, Y.-M. Gaodeng Xuexiao Huaxue Xuebao
2001, 22, 63. Chem. Abstr. 2002, 137, 179362.
9. (a) Maskey, R. P.; Gruen-Wollny, I.; Laatsch, H. Nat. Prod. Res. 2005, 19, 137–
142; For associated analog synthetic studies, see: (b) Hein, M.; Phuong, N. T. B.;
Michalik, D.; Goerls, H.; Lalk, M.; Langer, P. Tetrahedron Lett. 2006, 47, 5741–
5745.
10. Maskey, R. P.; Gruen-Wollny, I.; Laatsch, H. H. Angew. Chem., Int. Ed. 2002, 41,
597–599.
11. (a) Keller, P. A.; Yepuri, N. R.; Kelso, M. J.; Mariani, M.; Skelton, B. W.; White, A.
Tetrahedron 2008, 64, 7787–7795; (b) Bremner, J. B.; Coates, J. A.; Coghlan, D.
R.; David, D. M.; Keller, P. A.; Pyne, S. G. New J. Chem. 2002, 26, 1549–1551; (c)
Keller, P. A.; Birch, C.; Leach, S. P.; Tyssen, D.; Griffith, R. J. Mol. Graphics Modell.
2003, 21, 365–373.
Acknowledgements
We thank the University of Wollongong, the University of Wes-
tern Australia, the Australian National University, the former Am-

6950
M. K. Abdel-Hamid et al. / Tetrahedron Letters 50 (2009) 6947–6950
12. Aboul-Fadi, T.; Abdel-Hamid, F.; Hassan, E. A. Arch. Pharmacol. Res. 2003, 26,
778–784.
C28H28N2O2. Mr = 424.5. Monoclinic, P21/n, a = 13.3380(3), b = 12.2113(3),
c = 14.9625(4) Å, b = 107.903(1)° (T ca. 200 K), Z = 4. R1, wR2 = 0.046, 0.134 for
13. Despite the excellent atom recovery, a number of very minor unidentified
coloured products were present.
14. Microwave reactor used was a CEM Discover. Individual reaction conditions
are reported in the Supplementary data.
5314
(3648 > 2
r
(I))
CCD
reflections.
Conformational
disorder
(0.685(7):0.315(7)) is found in the CH:CHCH3 terminus of the OC4H7 string.
CCDC 723875.
16. Spartan 08 (Wavefunction): Calculations performed at semi-empirical AM1
level.
15. Structure
b = 9.055(4), c = 14.852(6) Å (T ca. 150 K), Z = 4. R1, wR2 = 0.036, 0.101 for
1335 unique (1269 > 2 (I)) CCD reflections. CCDC 721339. Structure
C22H16N2O. Mr = 324.4. Triclinic, P1, a = 8.2014(4), b = 13.4942(10),
c = 16.5073(9) Å, = 107.853(3), b = 91.354(4), = 91.958(3)° (T ca. 200 K),
Z = 4. R1, wR2 = 0.045, 0.121 for 6158 (3647 > 2 (I)) CCD reflections. Unresolved
lattice solvent was modelled using SQUEEZE. CCDC 723874. Structure
2 C11H9NO2, Mr = 187.2. Orthorhombic, P212121, a = 6.745(3),
17. (a) Taylor, D. L.; Ahmed, P. S.; Chambers, P.; Tyms, A. S.; Bedard, J.; Duchaine, J.;
Falardeau, G.; Lavallee, J. F.; Brown, W.; Rando, R. F.; Bowlin, T. Antiviral Chem.
Chemother. 1999, 10, 79–86; (b) Kato, M.; Ito, K.; Nishino, S.; Yamakuni, H.;
Takasugi, H. Chem. Pharm. Bull. 1994, 42, 2546–2555.
18. Schupp, P. J.; Kohert-Schupp, C.; Yoshida, W. Y.; Hemscheidt, T. K. Org. Lett.
2009, 11, 1111–1114.
r
4
a
c
r
9