Schischkiniin support studies: Synthetic access to 1,1′-bisindoles

Article abstract of DOI:10.1039/c2cc37241j

The first general synthetic approach to 1,1′-bisindoles is described. Two simultaneous, ambient temperature Mori-Ban reactions were used to effect bisindolization of several diallylated hydrazobenzenes with minimal cleavage of the N-N bond.

Full text of DOI:10.1039/c2cc37241j

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
Schischkiniin support studies: synthetic access
to 1,10-bisindoles†‡
Cite this: Chem. Commun., 2013,
49, 4349
Christy Wang and Jonathan Sperry*
Received 4th October 2012,
Accepted 26th October 2012
DOI: 10.1039/c2cc37241j
www.rsc.org/chemcomm
The first general synthetic approach to 1,10-bisindoles is described. to the lack of synthetic studies towards 1,10-bisindoles, the first
Two simultaneous, ambient temperature Mori–Ban reactions were general synthesis of this fascinating heterocyclic motif was sought.
used to effect bisindolization of several diallylated hydrazobenzenes
with minimal cleavage of the N–N bond.
Our main aim was that the methodology would target
3,30-disubstituted-1,10-bisindoles and thus be applicable to
the total synthesis of (1). The initial retrosynthesis is outlined
Schischkiniin (1) is a structurally remarkable alkaloid isolated from in Scheme 1 and is based around early-stage N–N bond formation
seeds of the thistle Centaurea schischkini in 2005 (Fig. 1).1 This and the late-stage construction of both heterocycles. It was
natural product possesses an exceedingly rare 1,10-bisindole motif2 envisaged two simultaneous Mori–Ban reactions10 could be used
embedded within a 14-membered macrocycle that is proposed to to construct a series of 1,10-bisindoles 2 from precursors 3. The
arise from dehydration of two Trp-Gly diketopiperazines followed bisindolization precursors 3 would be available from various allyl
by [2+2]-cycloaddition.1 Although the 1,10-bisindole moiety was bromides 4 and the common hydrazobenzene 5, which in turn is
observed in an hypothetical intermediate during the oxidation of available from 2-iodonitrobenzene 6. The construction of both
3-indoleacetic acid over 60 years ago,3,4 syntheses of this hetero- heterocycles in a single step by the ambitious transformation of 3
cyclic motif are sparse (4 synthetic examples).5–9 2,3-Diphenylindole to 2 requires the delicate N–N bond to remain intact during the
and N-hydroxyindole-3-carboxylic acid can be converted to their process.
respective 1,10-bisindoles with hydroiodic acid5,6 and a rare earth
The reductive dimerization of commercially available 2-iodo-
metal complex has been used to effect dimerization of indole to nitrobenzene 6 provided a straightforward synthesis of hydrazo-
1,10-bisindole.7 In the only example of a 1,10-bisindole that does not benzene 5 (Scheme 2). For the initial feasibility studies, we chose
rely on the dimerization of a preconstructed indole, 2-phenyl-1,10- to pursue the bisindolization of 7 derived from the diallylation of
bisindole has been synthesized by the palladium-catalyzed domino 5 with 3,3-dimethyallyl bromide. During the pursuit of 7, it was
reaction of a 2-bromophenylacetylene with 1-aminoindole.8,9 Due surprising to find a lack of literature precedent for the dialkylation
of hydrazobenzenes11 and despite numerous attempts, we were
unable to affect one-pot diallylation of hydrazobenzene 5 with 3,3-
dimethyallyl bromide. Thus, a two-step procedure was employed.
Fig. 1 Schischkiniin and the 1,10-bisindole motif.
School of Chemical Sciences, University of Auckland, 23 Symonds Street, Auckland,
New Zealand. E-mail: j.sperry@auckland.ac.nz; Fax: +64 9 373 7422;
Tel: +64 9 373 7599 ext 88269
† This article is part of the ChemComm ‘Emerging Investigators 2013’ themed issue.
‡ Electronic supplementary information (ESI) available: Experimental procedures,
1H and 13C NMR spectra of all new compounds. See DOI: 10.1039/c2cc37241j
Scheme 1 Approach to 1,10-bisindoles.
c
This journal is The Royal Society of Chemistry 2013
Chem. Commun., 2013, 49, 4349--4351 4349

View Article Online
Communication
ChemComm
Table 1 Double Mori–Ban cyclization of 7
Conditionsa,b,c
Yield (%)
HCO2Naf
Time 9
Solvent (h) (%) (%)
11
d
Pd(OAc)2 TBAClꢀ
Entry (mol%)
xH2Oe (eq.) (eq.)
1
2
3
4
5
6
7
8
9
10
5
1
20
20
20
20
20
20
20
3
3
3
3
3
3
3
3
3
3
1.2
0.6
0.12
2.2
DMF
DMF
DMF
DMF
72
72
72
48
48
72
1
0
0
27
0
0
Scheme 2 Synthesis of bisindolization precursor 7.
10 17
0
0
11 29
16 23
24 10
2.2 + Na2CO3 DMF
0
2.2
2.2
2.2
2.2
10
0
DMF
MeCN 48
DMA
THF
NMP
48
36
16
10
34
9
a
b
All reactions conducted on 0.05 mmol (30 mg) scale. All reactions
conducted at 27 1C. Heating the reaction above this temperature caused
c
N–N bond cleavage and 1,10-bisindole 9 was not observed. 9 and 11
d
are readily separable by flash chromatography. Pd(OAc)2 was superior
e
to PdCl2(MeCN)2. TBAClꢀxH2O was identified as the best ammonium
f
salt for this reaction. HCO2Na was superior to HCO2H and HCO2K.
reaction, conducting the bisindolization under reductive Heck
conditions was next considered. If successful, the resulting
1,10-bisindoline18 should undergo facile oxidation to the
desired 1,10-bisindole. Table 1 describes our attempts at effecting
Scheme 3 Failed attempts.
The first allylation was conducted in the presence of lithium double Mori–Ban cyclization of 7 under reductive conditions. The
diisopropylamide and the monoallylated product 8 subjected to a conditions used in entry 1 are based on literature precedent17c
second allylation in the presence of KOH/DMSO. This two-step and gratifyingly (albeit unexpectedly), a trace amount of the
allylation procedure facilitated the synthesis of bisindolization pre- 1,10-bisindole 9 was isolated, along with 3-isopropylindole (11)
cursor 7 in excellent yield (Scheme 2). Although both the mono- and arising from N–N bond cleavage.20 The isolation of the
diallylated hydrazobenzenes 8 and 7 degraded during attempted 1,10-bisindole 9 (the ‘normal’ Heck product from reductive reac-
purification, the desired product 7 was obtained in excellent purity tion conditions) infers that the b-hydride elimination occurs faster
over two steps (from 5) without need for chromatography.
than the reduction step.19 Aiming to build on this initially
With 7 to hand, we were keen to attempt the key bisindolization encouraging result, we embarked on several rounds of optimiza-
(Scheme 3). It became apparent that traditional Mori–Ban condi- tion of which the condensed results can be seen in the subsequent
tions (Conditions A, Scheme 3)10,12 caused extensive N–N bond entries in Table 1. Although reducing the quantity of catalyst
cleavage at temperatures above 25 1C, affording mixtures of 10 and (entries 2 and 3)21 had a detrimental effect on the reaction,
11 favouring the latter at higher temperature. No reaction was increasing the catalyst loading (entry 4)22 gave 10% yield of the
observed when employing these conditions at ambient tempera- 1,10-bisindole 9. The addition of base23 to the reaction mixture
ture.13 Similarly disappointing results were obtained using the caused extensive degradation of 7 and the production of small
Jeffery-inspired14 modified catalytic system employed by Larock15 quantities of the cleaved product 11 (entry 5).24 No conversion of
(Conditions B, Scheme 3). Sakamoto’s report16 describes the use substrate 7 was observed if the reaction was conducted in the
of Ag2CO3 as base to facilitate intramolecular Heck reaction of absence of sodium formate (entry 6), inferring it may play
N-allyl-2-iodoanilines at ambient temperature (Conditions C, an essential role in regenerating the active Pd(0)-catalyst.19
Scheme 3) but again, no reaction occurred under these conditions Entries 7–10 outline a solvent screen that was conducted, the
and extensive N–N bond cleavage was observed upon heating. results of which showed N-methylpyrrolidone gave the best overall
These initial results presented a significant conundrum: the yield of 9 (34%, entry 10). Although 34% seems moderate, it
indolization failed at ambient temperature but heating the represents B60% yield per heterocycle formation.
reaction above B25 1C cleaved the N–N bond. An alternative
The scope of this reaction was next investigated (Table 2).
strategy was clearly required.
The two-step allylation of 5 with various allyl bromides gave a
Literature searches reveal that reductive Heck reactions series of diallylated hydrazobenzenes that were subjected to
show precedent for proceeding at ambient temperature.17 As the optimized bisindolization conditions (entry 10, Table 1).
the Mori–Ban cyclization is essentially an intramolecular Heck Pleasingly, this new methodology can be used to access various
c
4350 Chem. Commun., 2013, 49, 4349--4351
This journal is The Royal Society of Chemistry 2013

View Article Online
ChemComm
Communication
Table 2 Synthesis of 1,10-bisindoles
Notes and references
1 M. Shoeb, S. Celik, M. Jaspars, Y. Kumarasamy, S. M. MacManus,
L. Nahar, P. K. Thoo-Lin and S. D. Sarker, Tetrahedron, 2005,
61, 9001.
2 For the isolation of three 1,10-bisindole dimers of agroclavine-1 and
epoxyagroclavine-1, see: N. F. Zelenkova, N. G. Vinokurova and
M. U. Arinbasarov, Appl. Biochem. Microbiol., 2003, 39, 44.
3 W. H. Houff, O. N. Hinsvark, L. E. Weller, S. H. Wittwer and
H. M. Sell, J. Am. Chem. Soc., 1954, 76, 5654.
4 A similar proposal has been put forward to explain the oxidation
chemistry of indole-3-methanol. R. N. Goyal, A. Kumar and P. Gupta,
J. Chem. Soc., Perkin Trans. 2, 2001, 618.
5 V. Askam and R. H. L. Deeks, J. Chem. Soc. C, 1968, 1243.
6 V. Dave, Can. J. Chem., 1972, 50, 3397.
Entry 1,10-Bisindole (yield)
Entry 1,10-Bisindole (yield)
7 L. Zhang, J. Xia, Q. Li, X. Li and S. Wang, Organometallics, 2011,
30, 375.
´
8 N. Halland, M. Nazare, J. Alonso, O. R’kyek and A. Lindenschmidt,
1
2
Chem. Commun., 2011, 47, 1042.
9 The pyrolysis of azo(2-methylindoline) affords 1,10-bis(2-methyl)-
indoline, although this product was not characterised. L. Peyrot,
M. Elkhatib, J. R. Vignalou, R. Metz, F. Elomar and H. Delalu,
J. Heterocycl. Chem., 2001, 38, 885.
10 (a) M. Mori, K. Chiba and Y. Ban, Tetrahedron Lett., 1977, 18,
1037; (b) R. Odle, B. Blevins, M. Ratcliff and L. S. Hegedus,
J. Org. Chem., 1980, 45, 2709; (c) A review on the intramolecular
Heck reaction contains an excellent section on indole synthesis, see:
S. E. Gibson and R. J. Middleton, Contemp. Org. Synth., 1996, 3,
447.
3
4
¨
11 (a) A. Bredihhin and U. Maeorg, Tetrahedron, 2008, 64, 6788;
(b) Y. Zhang, Q. Tang and M. Luo, Org. Biomol. Chem., 2011, 9,
4977.
12 (a) K. M. Aubart and C. H. Heathcock, J. Org. Chem., 1999, 64, 16;
´
(b) J. M. Mejıa-Oneto and A. Padwa, Org. Lett., 2006, 8, 3275;
(c) J. Ma, W. Yin, H. Zhou, X. Liao and J. M. Cook, J. Org. Chem.,
2009, 74, 264.
5
6
13 K. Koerber-Ple and G. Massiot, Synlett, 1994, 759.
14 (a) T. J. Jeffery, J. Chem. Soc., Chem. Commun., 1984, 1287;
(b) T. J. Jeffery, Tetrahedron Lett., 1985, 26, 2667; (c) T. J. Jeffery,
Synthesis, 1987, 70.
15 (a) R. C. Larock and S. Babu, Tetrahedron Lett., 1987, 28, 5291;
(b) W. Hong, L. Chen, C. Zhong and Z. Yao, Org. Lett., 2006, 8, 4919.
16 T. Sakamoto, Y. Kondo, M. Uchiyama and H. Yamanaka, J. Chem.
Soc., Perkin Trans. 1, 1993, 1941. See also ref. 12c.
a
b
3-Isopropylindole also isolated (9%). 3-Methylindole also isolated
c
d
(18%). 3-Ethylindole also isolated (22%). 3-Cyclohexylindole also
isolated (18%). 3-Benzylindole also isolated (8%). 3-Methylindole
and 3-isopropylindole also isolated (21%).
e
f
17 (a) M. Ichikawa, M. Takahashi, S. Aoyagi and C. Kibayashi, J. Am.
Chem. Soc., 2004, 126, 16553; (b) G. K. Jana and S. Sinha, Tetrahedron
Lett., 2012, 53, 1671; (c) P. Gao and S. P. Cook, Org. Lett., 2012,
14, 3340.
18 The palladium-catalyzed cyclization of 2-iodo-N-(prop-2-ynyl)-
anilines under reductive conditions (formic acid) affords 3-methyl-
eneindolines. See: B. Burns, R. Grigg, V. Sridharan and T. Worakun,
Tetrahedron Lett., 1988, 29, 4325.
19 The isolation of ‘normal’ Heck products under reductive conditions
has been reported previously, see: P. Liu, L. Huang, Y. Lu,
M. Dilmeghani, J. Baum, T. Xiang, J. Adams, A. Tasker, R. Larsen
and M. Faul, Tetrahedron Lett., 2007, 48, 2307.
20 We are currently investigating if Mori–Ban cyclization occurs pre- or
post N–N bond cleavage.
21 Ligand-free Heck reactions can proceed in higher yield by lowering
the catalyst loading, see: M. T. Reetz and J. G. de Vries, Chem.
Commun., 2004, 1559.
22 At higher catalyst loadings (>20 mol%) the reaction stalls and
formation of palladium black is clearly visible. Also see
ref. 21.
23 N. Charrier, E. Demont, R. Dunsdon, G. Maile, A. Naylor, A. O’Brien,
S. Redshaw, P. Theobold, D. Vesey and D. Walter, Synthesis, 2006,
3467.
24 Considering the diallylated hydrazobenzenes are exposed to basic
conditions during their synthesis, we can offer no conclusive
explanation as to why the presence of base causes extensive degra-
dation during the attempted bisindolization reaction.
3,30-dialkyl-1,10-bisindoles (9 and 12–14, entries 1–4), 3,30-
dibenzyl-1,10-bisindole (15, entry 5) and an unsymmetrical
1,10-bisindole (16, entry 6) in synthetically useful yields.
In conclusion, the first general synthesis of 1,10-bisindoles has
been accomplished. This new methodology relies on several dially-
lated hydrazobenzenes undergoing two simultaneous Mori–Ban
cyclizations to construct both heterocycles in a single step with
minimal N–N bond cleavage. Despite the bisindolization being
conducted under reductive conditions, the ‘normal’ Heck products
were obtained inferring that the b-hydride elimination occurs faster
than the reduction step. To the best of our knowledge, the propensity
for a class of substrates to all undergo the exclusive transformation
to ‘normal’ Heck products under base-free, reductive Heck condi-
tions is unprecedented. These exceedingly mild indolization condi-
tions reported herein should provide a valuable addition to current
indole syntheses. Application of this methodology towards the total
synthesis of the alkaloid 1 is currently in progress.
We are indebted to the RSNZ Marsden Fund for financial
support.
c
This journal is The Royal Society of Chemistry 2013
Chem. Commun., 2013, 49, 4349--4351 4351