Romo et al. analyzed the Diels-Alder reaction of N,N′-
disubstituted 2-imidazolones with unsaturated lactames.8
Furthermore, 2-silylated9 and 2-sulfanylated 5-alkenylimi-
dazoles have been studied, the latter leading Ohta et al.10 to
the synthesis of 15,15′-dimethylageliferin.
Scheme 1. Model Study on the Diels-Alder Reaction of
2-Amino-4-alkenylimidazoles with N-Phenylmaleimide
However, no [4 + 2] cycloaddition has been reported for
oroidin itself or for any related 4(5)-alkenylimidazole bearing
an amino substituent in the imidazole 2-position, although
this may be relevant for the chemical understanding of the
pyrrole-imidazole alkaloids.11
The 2-amino group renders electron density to the alkenyl-
imidazole system. Al Mourabit et al. discussed the ambiva-
lent reactivity of various tautomers of 2-amino-4(5)-alkenyl-
imidazoles, which, according to ab initio calculations, should
be energetically almost equal.12
We started with model studies on dienes 3,13 7, and 9
representing the amino component of the amide oroidin
(“eastern section”, Scheme 1). Compound 7 is accessible
from 1-benzyl-4-iodoimidazole via Sonogashira coupling
with Boc-protected propargylic amine, followed by azidation
at C-2 and reduction with Red-Al. Diene 9 was obtained
via chemoselective detritylation of 3 on treatment with
MeOH/HOAc (10:1) under reflux (12 h, 95%) retaining the
Boc group. For reasons of comparability with earlier stud-
ies,5,6 we chose N-phenylmaleimide and maleimide as
reactive dienophiles.
Conversions were almost quantitative by TLC. Cycload-
duct 5 was obtained at room temperature on reaction of 3
with N-phenylmaleimide. Already during workup of 5 we
observed some conversion to the rearranged product 6 with
the trityl group now attached to the terminal amino group.
Prolonged standing in CDCl3 led to complete conversion of
5 to 6. However, cycloadduct 8 obtained from the benzyl
analogue 7 was stable. Cycloaddition also occurred starting
from diene 9 in the absence of an imidazole protecting group
providing 10 in 58% yield after workup. The polar character
of the unprotected 2-amino imidazole moiety led to some
loss of material during chromatography on silica. We did
not observe any of the initial Diels-Alder products with the
double bond in the exocyclic position of the imidazole ring,
probably because rapid tautomerization occurs via the NH
groups of the 2-aminoimidazole moiety.
MM2 and PM3 calculations indicate that the energy
difference between the endo and exo products is smaller than
2 kcal/mol. However, NOESY spectra of 5, 6, and 8 indicate
that the endo products were formed. In particular, correlations
between 5-H/5a-H and 5a-H/8a-H were observed. The endo
selectivity of the cycloaddition is in agreement with earlier
observations by Romo et al.8 and Lovely et al.6 The endo
product should be kinetically preferred.5 On treatment with
[D4]MeOH, we observed exchange of 8a-H. However, no
diastereomeric product was formed, even on prolonged
reaction times. Reprotonation at C-8a appears to always occur
from the same side.
We did not observe any Diels-Alder homodimerization
of the 2-aminoimidazoles competing with the reaction of
maleimide. Heating of 3 or 9 in the absence of maleimide
did not yield any isolable product either. Even if two eastern
sections of oroidin were covalently connected to each other
by a urea linkage (conversion of compound 1113 to 12;
Scheme 2), we were not able to thermally induce a [4 + 2]
cycloaddition. We had been encouraged to that experiment
by the fact that N,N′-disubstituted ureas can be cyclized to
seven-membered rings, for instance, via intramolecular Heck
reaction.14
(5) Walters, M. A.; Lee, M. D. Tetrahedron Lett. 1994, 35, 8307-8310.
(6) (a) Lovely, C. J.; Du, H.; Dias, H. V. R. Org. Lett. 2001, 3, 1319-
1322. (b) Lovely, C. J.; Du, H.; Dias, H. V. R. Heterocycles 2003, 50,
1-7. (c) Lovely, C. J.; Du, H.; He, Y.; Dias, H. V. R. Org. Lett. 2004, 6,
735-738.
(7) He, Y.; Chen, Y.; Wu, H.; Lovely, C. J. Org. Lett. 2003, 5, 3623-
3626.
In which manner would the natural product oroidin itself
behave? Reflecting the reactivity of our model compounds,
oroidin (1‚HCO2H, obtained by total synthesis as the
formate15) starts to react with N-phenylmaleimide and
maleimide already at room temperature. Addition of Y(OTf)3
(20 mol %) led to acceleration of the Diels-Alder reactions
(8) (a) Dilley, A. S.; Romo, D. Org. Lett. 2001, 3, 1535-1538. (b)
Dransfield, P. J.; Wang, S.; Dilley, A.; Romo, D. Org. Lett. 2005, 7, 1679-
1682.
(9) Deghati, P. Y. F.; Wanner, M. J.; Koomen, G. J. Tetrahedron Lett.
1998, 39, 4561-4564.
(10) Kawasaki, I.; Sakaguchi, N.; Fukushima, N.; Fujioka, N.; Nikaido,
F.; Yamashita, M.; Ohta, S. Tetrahedron Lett. 2002, 43, 4377-4380.
(11) The amino component of the amide oroidin has been dimerized with
concomitant oxidation: Olofson, A.; Yakushijin, K.; Horne, D. A. J. Org.
Chem. 1997, 62, 7918-7919.
(12) Abou-Jneid, R.; Ghoulami, S.; Martin, M.-T.; Dau, E. T. H.; Travert,
N.; Al-Mourabit, A. Org. Lett. 2004, 6, 3933-3936.
(13) Breckle, G.; Polborn, K.; Lindel, T. Z. Naturforsch., B: Chem. Sci.
2003, 58, 451-456.
(14) Hayashi, M.; Sai, H.; Horikawa, H. Heterocycles 1998, 48, 1331-
1335. A discussion on the estimated energetics of the putative cyclization
of 12 is included in the Supporting Information.
(15) Oroidinium formate was obtained by refluxing N-trityloroidin
(synthesized via compound 3, ref 13) in CHCl3-MeOH-HCO2H (2:2:1).
820
Org. Lett., Vol. 8, No. 5, 2006