Evidence for the non-concerted [4+2]-cycloaddition of N-aryl imines when acting as both dienophiles and dienes under Lewis acid-catalysed conditions

Article abstract of DOI:10.1016/S0040-4039(02)02392-4

The ytterbium(III) triflate-catalysed reaction of a para-methoxyaniline, ethyl glyoxalate-derived imine, with a series of different dienes has resulted in products which initially suggest the operation of two possible modes of aza-Diels-Alder reaction, ho

Full text of DOI:10.1016/S0040-4039(02)02392-4

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 9633–9636
Evidence for the non-concerted [4+2]-cycloaddition of N-aryl
imines when acting as both dienophiles and dienes under
Lewis acid-catalysed conditions
a
b
b,
Stephen Hermitage, David A. Jay and Andrew Whiting *
a
GlaxoSmithKline Medicines Research Centre, Gunnels Wood Road, Stevenage, Herts SG1 2NY, UK
b
Department of Chemistry, Science Laboratories, University of Durham, South Road, Durham DH1 3LE, UK
Received 8 August 2002; revised 14 October 2002; accepted 24 October 2002
Abstract—The ytterbium(III) triflate-catalysed reaction of a para-methoxyaniline, ethyl glyoxalate-derived imine, with a series of
different dienes has resulted in products which initially suggest the operation of two possible modes of aza-Diels–Alder reaction,
however, a more likely explanation is that a common reaction mechanism is operating, involving a stepwise Lewis acid-catalysed
process and a common type of zwitterionic ytterbium complex. © 2002 Elsevier Science Ltd. All rights reserved.
There have been several recent reports of the applica-
tion of inverse electron-demand aza-Diels–Alder reac-
1
tions involving N-aryl imino dienes. However, many
(1)
groups have also reported that such imines react readily
2
as dienophiles with a range of dienes. Our interest in
this field and in particular the development of new
3
asymmetric aza-Diels–Alder reactions has led us to
consider this apparent dichotomy using para-
methoxyaniline imine 1 by studying its reactivity with a
range of dienes under Lewis acid conditions (Eq. (1)),
with a view to not only gaining a deeper understanding
of the aza-Diels–Alder process, but potentially develop-
ing new catalytic asymmetric aza-Diels–Alder pro-
with a range of different Lewis acids, different solvents
and in both the presence and absence of air (and hence
water, vide infra), in order to identify rapidly which
reactions could be cleanly catalysed. It is noteworthy
that imine 1 fails to react significantly with any of the
listed dienes in the absence of a Lewis acid catalyst at
ambient temperatures. Of the various catalysts and
solvents screened, ytterbium(III) triflate proved to be
the best general catalyst for all the reactions reported
herein. However, it became immediately obvious that
the reaction products were not all derived from imine 1
behaving solely as an imino dienophile. Indeed, one
reaction which immediately attracted our attention was
entry 2, Table 1, which required the presence of water
to provide the product 3b (under anhydrous reaction
conditions, the reaction produced a highly complex
mixture of products from which it was not possible to
isolate clean reaction products). In contrast, the
remaining dienes, i.e. 1-acetoxybutadiene, cyclopentadi-
ene and 1-methoxycyclohexadiene (entries 3–5) only
produced the corresponding tetrahydroquinolines 3c–e,
respectively, but it was noteworthy that the products in
3
cesses. In order to achieve this and to enable rational
planning of subsequent synthetic applications, it is nec-
essary to understand the chemoselectivity and therefore
mechanism operating in such processes. In this commu-
nication, we propose that a common, non-concerted
mechanism can explain the apparently dichotomous
behaviour of a single N-aryl imine with a range of
relatively electron-rich dienes under Lewis acid-
catalysed conditions.
Imine 1 and its methyl ester analogue have been
3
a,4
reported
to undergo facile asymmetric Lewis acid-
5
catalysed ‘cycloaddition’ with Danishefksy’s diene. The
reaction of 1 was also examined with several dienes (see
Table 1) under parallel reaction screening conditions
*
0
Corresponding author.
040-4039/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(02)02392-4

9
634
S. Hermitage et al. / Tetrahedron Letters 43 (2002) 9633–9636
†
entries 3 and 5 (Table 1) were obtained as mixtures of
diastereoisomers.
ent reaction mechanisms could be operating to explain
the observed products in Table 1. These are: (1) normal
electron-demand imino-dienophile–Danishefsky’s diene
Diels–Alder reaction (entry 1); (2) nucleophilic addition
At first sight, it was tempting to infer that three differ-
Table 1.
a
All yields are isolated after silica gel chromatography.
b
Reaction conditions: dry MeCN, 5 mol% Yb(OTf) , Ar, 5 h.
3
c
Reaction conditions: MeCN, 100 mol% H O, 5 mol% Yb(OTf) , Ar, 5 h.
2
3
d
†
3
c is obtained as the major diastereoisomer, together with two other minor stereoisomers. This compound is also unstable and aromatises to
the corresponding quinoline on silica gel and in air in EtOAc solution (half-life ca. 3 days, rt).
e
f
1
Obtained as a single diastereoisomer, stereochemistry assigned according to the H NMR spectrum and literature precedent (see Ref. 8).
1
Obtained as a mixture of two major diastereoisomers (2:1 ratio) according to the H NMR spectrum.
†
Compound 3c was obtained together with small amounts (ca. 19 and 17%, respectively, with respect to 3c) of the two further diastereoisomers,
tentatively assigned structures A and B, respectively, on the basis of the H NMR spectrum of the mixture.
1

S. Hermitage et al. / Tetrahedron Letters 43 (2002) 9633–9636
9635
of 1-trimethylsilyloxybutadiene to an electron-deficient
imine (Mannich-like reaction) (entry 2); and (3) inverse
terium incorporation in the CH CHꢀCHCHO position
2
of product 3b, i.e. did not produce 6.
electron-demand aniline–imine–diene+alkene dienophile
Diels–Alder reaction. However, a more reasonable
explanation is that all the observed products are
derived from the same initial imine activation mecha-
nism, i.e. Lewis acid complexation of the imine 1,
followed by nucleophilic addition of the different dienes
to form the corresponding intermediate zwitterionic
ytterbium complexes (4a–e). The final products are then
derived from two alternative regioselective cyclisation
reactions or direct hydrolysis of the zwitterion. The fate
of each of the zwitterions 4, in terms of which mode of
reaction occurs, i.e. cyclisation versus hydrolysis, can be
readily explained by their different substituents and
therefore relative stabilities. In the case of entry 1, the
cationic part of zwitterion 4a is particularly stabilised
by resonance and presumably cyclises relatively slowly,
subsequent hydrolysis then provides cycloadduct 3a. In
contrast, zwitterion 4b is less stable and tends to pro-
duce several unidentified products in the absence of an
external nucleophile. However, the presence of water in
the reaction mixture allows rapid hydrolysis of the
zwitterion 4b to provide addition product 3b. In the
cases of entries 3–5, each of the zwitterions have inter-
mediate stability and cyclise via an electrophilic aro-
matic substitution mechanism to provide each of the
products 3c–e.
In summary, it has long been considered that aza-
Diels–Alder reactions can proceed through either an
unsymmetric, yet concerted cycloaddition mechanism,
6
or via stepwise processes. However, it is very likely
that in certain cases of Lewis acid-catalysed reactions of
N-aryl imines which are seemingly acting as either
dienes or dienophiles, that in fact the observed
chemoselectivity is controlled by a metal-activated,
acyclic reaction mechanism, i.e. an addition–cyclisation
process. On occasion, the intermediate zwitterionic
addition products can be intercepted, adding support to
this hypothesis. This analysis may remove the necessity
to invoke the intervention of independent mechanisms
7
operating in systems which appear to be remarkably
similar.
Acknowledgements
We thank EPSRC and GlaxoSmithKline for an indus-
trial CASE studentship (to DAJ) (Ref. no. 99314731X).
References
It is possible to gain mechanistic insight into the pro-
cesses operating in the reactions outlined in Table 1 by
closely examining the origin of product 3b. A normal
electron-demand Diels–Alder reaction of imine 1 and
1
. (a) Ishitani, H.; Kobayashi, S. Tetrahedron Lett. 1996, 37,
357–7360; (b) Stevenson, P. J.; Nieuwenhuyzen, M.;
Osborne, D. Chem. Commun. 2002, 444–445.
. (a) Kobayashi, S.; Komiyama, S.; Ishitani, H. Angew.
Chem., Int. Ed. Engl. 1998, 37, 979–981; (b) Kobayashi, S.;
Kusakabe, K.; Ishitani, I. Org. Lett. 2000, 2, 1225–1228;
c) Kobayashi, S.; Kusakabe, K.; Komiyama, S.; Ishitani,
I. J. Org. Chem. 1999, 64, 4220–4221.
. (a) Bromidge, S.; Wilson, P.; Whiting, A. Tetrahedron
Lett. 1998, 39, 8905–8908; (b) McFarlane, A. K.; Thomas,
G.; Whiting, A. J. Chem. Soc., Perkin Trans. 1 1995,
7
1-trimethylsiloxybutadiene would produce Diels–Alder
adduct 5, which after hydrolysis of the trimethylsilyloxy
hemi-aminal function and Z to E isomerisation of the
unsaturated aldehyde could result in the formation of
the observed product 3b. Hence, if the reaction were
2
(
carried out in the presence of D O, hydrolysis of adduct
2
5
would be accompanied by deuteration to give the
deuterated analogue of 3b, i.e. E-unsaturated aldehyde
. The alternative reaction mechanism would involve an
3
6
inverse electron-demand Diels–Alder reaction between
the aryl imine 1 and the non-oxygenated terminal olefin
of the 1-trimethylsilyloxybutadiene, which would ini-
tially produce 7a, then 7b after hydrolysis of the
trimethylsilyl ether. Products resulting from either of
these pathways were not identified in the crude reaction
mixture. In addition, support for the fact that zwitte-
rion 4b is indeed involved in the formation of product
2
803–2808; (c) McFarlane, A. K.; Thomas, G.; Whiting,
A. Tetrahedron Lett. 1993, 34, 2379–2382.
. Yao, S.; Saaby, S.; Hazell, R. G.; Jørgensen, K. A. Chem.
Eur. J. 2000, 6, 2435–2448.
4
5
. Danishefsky, S.; Kitahara, T. J. Am. Chem. Soc. 1974, 96,
7
807–7808.
6. (a) Boger, D. L.; Weinreb, S. Hetero Diels–Alder Method-
ology in Organic Synthesis; Academic Press: San Diego,
1987; (b) Jung, M. E.; Shishido, K.; Light, L.; Davis, L.
Tetrahedron Lett. 1981, 22, 4607–4610; (c) Stanetty, P.;
3
b is demonstrated by the fact that use of anhydrous
Yb(OTf) , dry MeCN, but using D O in place of water
3
2
to quench the intermediate in situ, produced no deu-

9
636
S. Hermitage et al. / Tetrahedron Letters 43 (2002) 9633–9636
Mihovilovic, M. D. Chem. Lett. 1997, 849–850; (d)
McCarrick, M. A.; Wu, Y.-D.; Houk, K. N. J. Org. Chem.
993, 58, 3330–3345; (e) Whiting, A.; Windsor, C. M.
Tetrahedron 1998, 54, 6035–6050.
7. Stanetty, P.; Mihovilovic, M. D.; Mereiter, K.; V o¨ llenkle,
H.; Renz, F. Tetrahedron 1998, 54, 875–894.
8. Bertilsson, S. K.; Ekegren, J. K.; Modin, S. A.; Andersson,
1
P. G. Tetrahedron 2001, 57, 6399–6406.