Stereoselective formal [3 + 3] cycloaddition approach to cis-1-azadecalins and synthesis of (-)-4a,8a-diepi-pumiliotoxin C. Evidence for the first highly stereoselective 6π-electron electrocyclic ring closures of 1-azatrienes

Article abstract of DOI:10.1021/ja020698b

Evidence is described here to support that a highly stereoselective 6π-electron electrocyclic ring closure of 1-azatrienes is a key step in formal [3 + 3] cycloaddition or annulation reactions of chiral vinylogous amides with α,β-unsaturated iminium salts

Full text of DOI:10.1021/ja020698b

Published on Web 08/13/2002
Stereoselective Formal [3 + 3] Cycloaddition Approach to
cis-1-Azadecalins and Synthesis of
(-)-4a,8a-diepi-Pumiliotoxin C. Evidence for the First Highly
Stereoselective 6π-Electron Electrocyclic Ring Closures of
1-Azatrienes
Heather M. Sklenicka, Richard P. Hsung,*,† Michael J. McLaughlin, Lin-li Wei,
Aleksey I. Gerasyuto, and William B. Brennessel
Contribution from the Department of Chemistry, UniVersity of Minnesota,
Minneapolis, Minnesota 55455
Received May 16, 2002
Abstract: Evidence is described here to support that a highly stereoselective 6π-electron electrocyclic
ring closure of 1-azatrienes is a key step in formal [3 + 3] cycloaddition or annulation reactions of chiral
vinylogous amides with R,â-unsaturated iminium salts. This would represent the first highly stereoselective
6π-electron electrocyclic ring closure of 1-azatrienes. We have also unambiguously demonstrated that these
specific ring closures are reversible, leading to the major diastereomer that is also thermodynamically more
stable, and that a rotation preference likely also plays a role. A synthetic application is illustrated here to
stereoselectively transform the resulting dihydropyridines to cis-1-azadecalins with unique anti relative
stereochemistry at C2 and C2a, leading to synthesis of epi isomers of (-)-pumiliotoxin C.
Introduction
Annulation reactions of vinylogous amides with R,â-unsatur-
ated carbonyl systems represent a powerful synthetic approach
for constructing nitrogen heterocycles and related nitrogen
alkaloids.1,2 Most of these reactions have led to pyridines, and
in some cases, 2-pyridones, 4-pyridones, and 1,4-dihydropy-
ridines have been reported (Figure 1).1 The attractive aspect of
these reactions is that the formation of the six-membered
nitrogen heterocycle ultimately constitutes a formal hetero [3
+ 3] cycloaddition.3,4 Two of the five carbons along with the
nitrogen atom in the resulting six-membered nitrogen hetero-
cycle come from the vinylogous amide with the remaining three
carbons from the R,â-unsaturated carbonyl system. Two possible
regiochemical alignments can be commonly found in these
Figure 1. Precedents in formal aza-[3 + 3].
reactions. The formation of pyridines or 4-pyridones mainly
results from a head-to-head (both carbonyls are in the same
direction) regiochemical alignment, and products such as
2-pyridones and 1,4-dihydropyridines are results from a head-
to-tail (carbonyls are in opposite directions) alignment. Despite
that a variety of different R,â-unsaturated carbonyl systems have
been employed in these reactions, R,â-unsaturated iminium salts
have not been used,5 nor have 1,2-dihydropyidines been found
in these reactions.
* To whom correspondence should be addressed. E-mail: hsung@
chem.umn.edu.
† A Recipient of 2001 Camille Dreyfus Teacher-Scholar Award.
(1) Kucklander, U. Enaminones as Synthons. In The Chemistry of Functional
Groups: The Chemistry of Enamines Part I; Rappoport, Z., Ed.; John Wiley
& Sons: New York, 1994; p 523.
(2) (a) Grundon, M. F. In The Alkaloids: Quinoline Alkaloids Related to
Anthranilic Acids; Academic Press: London, 1988; Vol. 32, p 341. (b)
Daly, J. W.: Garraffo, H. M.; Spande, T. F. In The Alkaloids; Cordell, G.
A., Ed.; Academic Press: New York, 1993; Vol. 43, p 185. (c) Michael,
J. P. Nat. Prod. Rep. 1999, 16, 675, 697. (d) Jones, T. H.; Gorman, J. S.
T.; Snelling, R. R.; Delabie, J. H. C.; Blum, M. S.; Garraffo, H. M.; Jain,
P.; Daly, J. W.; Spande, T. F. J. Chem. Ecol. 1999, 25, 1179. (e) Michael,
J. P. Nat. Prod. Rep. 2000, 17, 579. (f) Lewis, J. R. Nat. Prod. Rep. 2001,
18, 95.
(3) For the term “formal [3 + 3]” used to describe [3 + 3] carbo-cycloadditions,
see: (a) Seebach, D.; Missbach, M.; Calderari, G.; Eberle, M. J. Am. Chem.
Soc. 1990, 112, 7625. For earlier studies on [3 + 3] carbo-cycloadditions,
see: (b) Landesman, H. K.; Stork, G. J. Am. Chem. Soc. 1956, 78, 5129.
(c) Hendrickson, J. B.; Boeckman, R. K., Jr. J. Am. Chem. Soc. 1971, 93,
1307. (d) Alexakis, A.; Chapdelaine, M. J.; Posner, G. H. Tetrahedron
Lett. 1978, 19, 4209.
(4) For recent reviews on stepwise [3 + 3] carbo-cycloadditions using enamines,
enol ethers, or â-keto esters, see: (a) Filippini, M.-H.; Rodriguez, J. Chem.
ReV. 1999, 99, 27. (b) Rappoport, Z. The Chemistry of Enamines. In The
Chemistry of Functional Groups; John Wiley & Sons: New York, 1994.
(c) Filippini, M.-H.; Faure, R.; Rodriguez, J. J. Org. Chem. 1995, 60, 6872,
and see refs 21-33 cited therein. For reviews on metal mediated stepwise
[3 + 3] cycloaddition reactions, see: (d) Fru¨hauf, H.-W. Chem. ReV. 1997,
97, 523. (e) Lautens, M.; Klute, W.; Tam, W. Chem. ReV. 1996, 96, 49.
9
10.1021/ja020698b CCC: $22.00 © 2002 American Chemical Society
J. AM. CHEM. SOC. 2002, 124, 10435-10442
10435

A R T I C L E S
Sklenicka et al.
Scheme 1
those using 1,3-diketo systems but opposite of Hickmott-Stille’s
aza-annulation using acid anhydrides or chlorides.14 Thus,
mechanistically, these new reactions of vinylogous amides 1
were thought to involve a sequence that consists of a Knoev-
enagel condensation followed by a key 6π-electron electrocyclic
ring closure of 1-azatrienes 4.15 The net result of this sequential
anionic-pericyclic strategy16 is the formation of two σ-bonds
in addition to an important carbocenter adjacent to the nitrogen
atom that could be controlled stereochemically.17 It provides a
rapid construction of 1,2-dihydropyridines from simple and
accessible vinylogous amides. In this paper, we disclose full
details of this stereoselective formal [3 + 3] cycloaddition
reaction, mechanistic evidences for the first stereoselective ring
closure of 1-azatrienes, and applications of this methodology
in stereoselective constructions of cis-1-azadecalins.
Results and Discussions
(1) Stereoselectivity and Mechanism. (a) Stereoselective
Formal [3 + 3] Formal Cycloaddition Using Chiral Vinyl-
ogous Amides. Having established the synthetic feasibility in
accessing 1,2-dihydropyridines, a stereoselective variant of this
formal [3 + 3] cycloaddition using chiral vinylogous amides
such as 5 and 6 was developed to give 1,2-dihydropyridines 7
and 8, respectively, in high diastereoselectivity and good yields
when reacted with a range of different R,â-unsaturated iminiums
salts (Scheme 2).17 Given the vast number of nitrogen alkaloids
that are known to possess this important stereogenic center
adjacent to the nitrogen atom, this stereoselective reaction
represents an attractive and novel entry to natural products with
the 1-azadecalinic structural motif.2,18,19 Most significantly, this
stereoselective manifold potentially represents the first examples
Our interest in this area commenced with an investigation
involving condensations of R,â-unsaturated aldehydes with
4-hydroxy-2-pyrones and 1,3-cyclohexanediones (Scheme 1).
These reactions have been known for the almost 50 years since
Link’s first report using 4-hydroxycoumarins6 and have been
extensively examined by Moreno-Man˜as.7 However, these
reactions also suffer from problems caused by the competing
1,2- versus 1,4-additions and C- versus O-addition, thereby
providing a synthetically less useful process.7 Our initial
contribution was to recognize that instead of generating R,â-
unsaturated iminium ions in situ, the direct use of iminium salts
provided a highly efficient entry to 2H-pyrans in favor of the
head-to-head alignment (Scheme 1).8-12
(12) For other studies using 2-pyrones and 1,3-cyclohexanones, see: (a) Hua,
D. H.; Chen, Y.; Sin, H.-S.; Maroto, M. J.; Robinson, P. D.; Newell, S.
W.; Perchellet, E. M.; Ladesich, J. B.; Freeman, J. A.; Perchellet, J.-P.;
Chiang, P. K. J. Org. Chem. 1997, 62, 6888. (b) Moorhoff, C. M. Synthesis
1997, 685. (c) Jonassohn, M.; Sterner, O.; Anke, H. Tetrahedron 1996,
52, 1473. (d) Krasnaya, Z. A.; Bogdanov, V. S.; Burova, S. A.; Smirnova,
Y. V. Russ. Chem. 1995, 44, 2118. (e) Appendino, G.; Cravotto, G.;
Tagliapietra, S.; Nano, G. M.; Palmisano, G. HelV. Chim. Acta 1990, 73,
1865. (f) Schuda, P. F.; Price, W. A. J. Org. Chem. 1987, 52, 1972. (g)
Tietze, L. F.; v. Kiedrowski, G.; Berger, B. Synthesis 1982, 683. (h) de
Groot, A.; Jansen, B. J. M. Tetrahedron Lett. 1975, 16, 3407.
(13) (a) Hsung, R. P.; Wei, L.-L.; Sklenicka, H. M.; Douglas, C. J.; McLaughlin,
M. J.; Mulder, J. A.; Yao, L. J. Org. Lett. 1999, 1, 509. For other related
approaches reported recently, see: (b) Hedley, S. J.; Moran, W. J.; Prenzel,
A. H. G. P.; Price, D. A.; Harrity, J. P. A. Synlett 2001, 1596. (c) Nemes,
P.; Bala´zs, B.; To´th, G.; Scheiber, P. Synlett 1999, 222. (d) Hua, D. H.;
Chen, Y.; Sin, H.-S.; Robinson, P. D.; Meyers, C. Y.; Perchellet, E. M.;
Perchellet, J.-P.; Chiang, P. K.; Biellmann, J.-P. Acta Crystallogr. 1999,
C55, 1698. (e) Heber, D.; Berghaus, Th. J. Heterocycl. Chem. 1994, 31,
1353.
(14) (a) Hickmott, P. W.; Sheppard, G. J. Chem. Soc. C 1971, 2112. (b)
Benovsky, P.; Stephenson, G. A.; Stille, J. R. J. Am. Chem. Soc. 1998,
120, 2493. (c) Paulvannan,; K.; Stille, J. R. Tetrahedron Lett. 1993, 34,
215, 6677. (d) Paulvannan, K.; Stille, J. R. J. Org. Chem. 1992, 57, 5319.
(e) Greenhill, J. V.; Mohamed, M. J. Chem. Soc., Perkin Trans. 1 1979,
1411. (f) Chaaban, J.; Greenhill, J. V.; Rauli, M. J. Chem. Soc., Perkin
Trans. 1 1981, 3120. (g) Greenhill, J. V.; Moten, A. J. Chem. Soc., Perkin
Trans. 1 1984, 287. (h) Heber, D.; Berghaus, Th. J. Heterocycl. Chem.
1994, 31, 1353.
(15) For leading references on electrocyclic ring-closures involving 1-azatrienes,
see: (a) Maynard, D. F.; Okamura, W. H. J. Org. Chem. 1995, 60, 1763.
(b) de Lera, A. R.; Reischl, W.; Okamura, W. H. J. Am. Chem. Soc. 1989,
111, 4051. (c) Okamura, W. H.; de Lera, A. R.; Reischl, W. J. Am. Chem.
Soc. 1988, 110, 4462. For an earlier account, see: (e) Oppolzer, V. W.
Angew. Chem. 1972, 22, 1108. For a recent account similar to Okamura’s
elegant studies, see: (f) Tanaka, K.; Mori, H.; Yamamoto, M.; Katsumura,
S. J. Org. Chem. 2001, 66, 3099.
(16) For a review on synthetic methods involving sequential transformations,
see: (a) Tietze, L. F.; Beifuss, U. Angew. Chem., Int. Ed. Engl. 1993, 32,
131. (b) Tietze, L. F. J. Heterocycl. Chem. 1990, 27, 47.
(17) Sklenicka, H. M.; Hsung, R. P.; Wei, L.-L.; McLaughlin, M. J.; Gerasyuto,
A. I.; Degen, S. J.; Mulder, J. A. Org. Lett. 2000, 2, 1161.
(18) For a recent review on preparations of piperidines, see: Laschat, S.; Dickner,
T. Synthesis 2000, 1781.
At the heel of the success in using R,â-unsaturated iminium
salts, first reactions of vinylogous amides 1 with R,â-unsaturated
iminium salts 2 were explored and found to be very efficient,
leading exclusively to 1,2-dihydropyridines 5 (Scheme 1).13 The
regiochemistry of these reactions is the same (head-to-head) as
(5) Vinamidiniums have received recent attention in syntheses of pyridines;
see: (a) Davies, I. W.; Marcoux, J.-F.; Reider, P. J. Org. Lett. 2001, 3,
209. (c) Davies, I. W.; Taylor, M.; Marcoux, J.-F.; Wu, J.; Dormer, P. G.;
Hughes, D.; Reider, P. J. J. Org. Chem. 2001, 66, 251.
(6) (a) Ikawa, M.; Stahmann, M. A.; Link, K. P. J. Am. Chem. Soc. 1944, 66,
902. (b) Seidmna, M.; Robertson, D. N.; Link, K. P. J. Am. Chem. Soc.
1950, 72, 5193. (c) Seidmna, M.; Link, K. P. J. Am. Chem. Soc. 1952, 74,
1885.
(7) de March, P.; Moreno-Man˜as, M.; Casado, J.; Pleixats, R.; Roca, J. L.;
Trius, A. J. Heterocycl. Chem. 1984, 21, 1369.
(8) Hsung, R. P.; Shen, H. C.; Douglas, C. J.; Morgan, C. D.; Degen, S. J.;
Yao, L. J. J. Org. Chem. 1999, 64, 690.
(9) The advantage of using R,â-unsaturated iminium salts has also recently
been noted; see: Cravotto, G.; Nano, G. M.; Tagliapietra, S. Synthesis 2001,
49.
(10) (a) McLaughlin, M. J.; Shen, H. C.; Hsung, R. P. Tetrahedron Lett. 2001,
42, 609. (b) Douglas C. J.; Sklenicka, H. M.; Shen, H. C.; Golding, G. M.;
Mathias, D. S.; Degen, S. J.; Morgan, C. D.; Shih, R. A.; Mueller, K. L.;
Seurer, L. M.; Johnson, E. W.; Hsung, R. P. Tetrahedron 1999, 55, 13683.
(11) For applications to natural product syntheses, see: (a) McLaughlin, M. J.;
Hsung, R. P.; Cole, K. C.; Hahn, J. M.; Wang, J. Org. Lett. 2002, 4, 2017.
(b) Wang, J.; Cole, K. P.; Wei, L. L.; Zehnder, L. R.; Hsung, R. P.
Tetrahedron Lett. 2002, 43, 3337. (c) Cole, K. P.; Hsung, R. P.; Yang,
X.-F. Tetrahedron Lett. 2002, 43, 3341. (d) Wei, L.-L.; Hsung, R. P.;
Sklenicka, H. M.; Gerasyuto, A. I. Angew. Chem., Int. Ed. 2001, 40, 1516.
(f) McLaughlin, M. J.; Hsung, R. P. J. Org. Chem. 2001, 66, 1049. (g)
Zehnder, L. R.; Hsung, R. P.; Wang, J.-S.; Golding, G. M. Angew. Chem.,
Int. Ed. 2000, 39, 3876. (h) Zehnder, L. R.; Dahl, J. W.; Hsung, R. P.
Tetrahedron Lett. 2000, 41, 1901.
9
10436 J. AM. CHEM. SOC. VOL. 124, NO. 35, 2002

Cycloaddition Approach to cis-1-Azadecalins
A R T I C L E S
Scheme 2. Stereoselective Formal [3 + 3] Cycloadditon
closure of 1-azatrienes 9 after the Knoevenagel condensation.
Pathway b, the most likely alternative pathway, would involve
an N-1,4-addition followed by the Knoevenagel-type condensa-
tion via the intermediate 10.
A closer analysis of these two possibilities would reveal that
to distinguish them, one could attempt to detect and/or isolate
intermediates related to 9 and 16 (Scheme 3). The major
difference in the two possibilities would reside in the C-1,2-
addition step. For pathway a, the initial C-1,2-addition inter-
mediate would be 12. While an ensuing â-elimination could
lead to the 1-azatriene intermediate 9, a tautomerization could
first give 13 and a subsequent â-elimination would then provide
9. It is also noteworthy that the intermediate 13 itself could
Scheme 3
2
also undergo an intramolecular SN displacement to give the
dihydropyridine product 11. This would lead to a mechanistic
possibility that cannot be completely ruled out even if the
1-azatriene intermediate 9 can be detected.
On the other hand, for pathway b, the C-1,2-addition step
via 10 after the initial N-1,4-addition would lead to an iminium
type intermediate 14. While 14 could undergo â-elimination to
directly give the vinyl iminium intermediate 15 that could further
tautomerize to the dihydropyridine 11, simple intuitions suggest
that tautomerization of 14 could proceed faster, leading to the
neutral allylamine 16, and that â-elimination of the NR2 group
in this case could be slower. It could be further suggested that
allylic amines such as 16 could be isolated or trapped. Evidence
in differentiating these two pathways can lend support to whether
this reaction does involve a 6π-electron electrocyclic ring closure
of 1-azatrienes.
(2) Isolation of an Intermediate Related to 16. Although
we have never spectroscopically observed or isolated any
intermediates related to 16 in intermolecular reactions of
vinylogous amides under any conditions, this can only serve as
a negative support. However, an intramolecular Variant of this
annulation reaction using vinylogous amides tethered with R,â-
unsaturated iminium salt as in 17 provided an opportunity to
isolate an intermediate related to 16 (Scheme 4).11d
For the intramolecular reaction, the mechanism could also
proceed via two possibilities: Pathway c and pathway d with
distinctions similar to those described for pathways a and b.
However, intuitions would support pathway d (corresponding
to pathway b in intermolecular reactions) since it does not
require an initial counterintuitiVe formation of a macrocycle
18 as suggested in the pathway c (corresponding to pathway
a). If this intuitive assertion is valid, then the intermediate 21
in pathway d, related to 16 in pathway b, should be detectable
or attainable by isolation, thereby leading to positive support
for an N-1,4-/C-1,2-addition pathway.
To achieve this task, the vinylogous amide 24 tethered with
the enal was prepared as a 3:1 E/Z mixture from 22 using
standard conditions (Scheme 5).20 While the intramolecular
formal [3 + 3] cycloaddition reaction of 24 under the conditions
using piperidinium acetate salt20 at 150 °C gave the desired
product 25 in 55% yield as a single diastereomer, the intermedi-
ate 26 was isolated in 40% yield also as a single diastereomer
of stereoselective 6π-electron electrocyclic ring closure of
1-azatrienes.
(b) Evidence for the 6π-Electron Electrocyclic Ring
Closure of 1-Azatrienes. (1) Mechanistic Analysis. Although
the mechanism described in Scheme 1 provides a reasonable
pathway leading to the dihydropyridine 5 that is in agreement
with the assigned regiochemistry (head-to-head), there are other
possible mechanistic pathways that could also lead to such
conclusions. While it is not possible to completely rule out all
mechanistic possibilities, two pathways could be assigned as
the major possibilities (Scheme 3). One would involve pathway
a with the key step being the 6π-electron electrocyclic ring
(19) For recent approaches to dihydropyridines and related piperidines, see: (a)
Comins, D. L.; Sandelier, M. J.; Grillo, T. A. J. Org. Chem. 2001, 66,
6829. (b) Charette, A. B.; Grenon, M.; Lemire, A.; Pourashraf, M.; Martel,
J. J. Am. Chem. Soc. 2001, 123, 11829. (c) Brooks, C. A.; Comins, D. L.
Tetrahedron Lett. 2000, 41, 3551. (d) Matsumura, Y.; Nakamura, Y.; Maki,
T.; Onomura, O. Tetrahedron Lett. 2000, 41, 7685. (e) Comins, D. L.;
Zhang, Y.; Joseph, S. P. Org. Lett. 1999, 1, 657. (f) Comins, D. L.; Brooks,
C. A.; Al-awar, R. S.; Goehring, R. R. Org. Lett. 1999, 1, 229. (g) Comins,
D. L.; Kuethe, J. T.; Hong, H.; Lakner, F. J. J. Am. Chem. Soc. 1999, 121,
2651. (h) References 15 and 13b-e.
(20) See Supporting Information for preparations of all relevant new compounds
and details of their full characterizations along with their 1H NMR spectra.
These reactions are best carried out in toluene and/or EtOAc using
piperidine/Ac2O. L-Proline does not work well for this reaction contrary to
literature accounts due to low solubility in EtOAc. We can also use
piperidinium salts as well as other amine salts.
9
J. AM. CHEM. SOC. VOL. 124, NO. 35, 2002 10437

A R T I C L E S
Sklenicka et al.
Scheme 4
Scheme 6
closure readily.15,21 We were, however, able to isolate success-
fully in one case the 1-azaztriene 29 partially containing the
ring-closed product 30 in ∼50-60% yield from reactions of
the aminopyrone 27 with the R,â-unsaturated iminium salt 28
at 80 °C (Scheme 6). Although the stereochemistry of C3-C4
olefin was not vigorously assigned, there appeared to be mainly
1
just one isomer of 29, as suggested in H NMR and LCMS.20
On the basis of literature precedents of related 1-heterot-
rienes,8,15,21 this assignment of the 1-azatriene 29 is well-
substantiated, although attempts to separate 29 and 30 for further
characterizations failed. In addition, the mixture of 29 and 30
could be resubjected to reaction conditions at higher temperature
(150 °C) and/or with a longer reaction time (48 h) to give 30
quantitatively (an overall yield of 60% from 27).
Scheme 5
This collective evidence suggests that, while not completely
ruling out other mechanistic pathways, this formal [3 + 3]
cycloaddition of vinylogous amides proceeds via a Knoevenagel
condensation followed by an electrocyclic ring closure of
1-azatrienes analogous to those described by Link and Moreno-
Man˜as using 4-hydroxy-2-pyrones. Although 6π-electron elec-
trocyclic ring closure of 1-azatrienes has been elegantly studied
by Okamura,15 and although Okamura has also made the first
attempts in effecting these ring closures stereoselectively,22 the
diastereoselectivity observed using chiral vinylogous amides
represents the first highly stereoselective 6π-electron electro-
cyclic ring closure of 1-azatrienes.
(c) Rotational Preference and Reversibility of the Ring
Closure. (1) Rotational Preference. Having established evi-
dence to support that the stereoselective formal [3 + 3]
cycloaddition reaction of vinylogous amides involves a key
stereoselective ring closure of 1-azatrienes, the observed high
diastereoselectivity could be proposed as a result of a key
rotational preference in the 1-azatriene 31 (Figure 2). Rotational
preferences leading to stereoselective constructions of sp2 and
sp3 hybirdized stereocenters have excellent precedents in various
pericyclic processes, and such stereoselectivity has been ap-
propriately termed as torquoselectivity.23,24
when the reaction was carried out at 85 °C. The intermediate
26 was fully characterized, although stereochemistry at C4 was
not vigorously assigned. Further heating of 26 at higher
temperature and/or for a longer reaction time, the tricycle 25
was isolated in comparable yield as a single diastereomer in
favor of the same major isomer.
This experiment suggests that if the mechanistic pathway
would involve an N-1,4-addition/C-1,2-addition as described for
pathway d, 21 is a viable reaction intermediate that could be
isolated. Thus, inability to observe or attain the proposed
intermediate 16, corresponding to 21 in pathway d, under any
reaction conditions suggests that an N-1,4-addition/C-1,2-
addition as described in pathway b is likely not involved in
intermolecular formal [3 +3] cycloadditions.
Rotations of the vinyl strand in 31 along in directions a and
b would lead to diastereomers 32a and 32b, respectively. It was
(21) For leading references on electrocyclic ring-closures involving 1-oxatrienes,
see: (a) Kametani, T.; Kajiwara, M.; Fukumoto, K. Tetrahedron 1974,
30, 1053. (b) Shishido, K.; Ito, M.; Shimada, S.-I.; Fukumoto, K.; Kametani,
T. Chem. Lett. 1984, 1943. (c) Shishido, K.; Shitara, E.; Fukumoto, K. J.
Am. Chem. Soc. 1985, 107, 5810. (d) Shishido, K.; Hiroya, K.; Fukumoto,
K.; Kametani, T. Tetrahedron Lett. 1986, 27, 971.
(22) Professor Bill Okumura performed the first attempts at effecting these ring
closures of 1-azatrienes stereoselectively using optically active phenyl-
ethylamine: A personal communication.
(3) Identification of the 1-Azatriene Intermediate Related
to 9. Finding and/or isolating 1-azatriene intermediates such as
9 were not trivial, although we have on occasions isolated and
trapped 1-oxatriene intermediates in related reactions.8 These
1-heterotrienes are usually very reactive and undergo ring
9
10438 J. AM. CHEM. SOC. VOL. 124, NO. 35, 2002

Cycloaddition Approach to cis-1-Azadecalins
A R T I C L E S
Chart 1
Figure 2. Stereoselective ring closure of 1-azatriene.
It could be proposed that the 1-azatriene 33 also assumes
such a conformational preference prior to the ring closure. This
assumption would then render 33 with a conformation where
the C2′ phenyl is almost π-stacked with the C5-C6 alkene,
and there can be two possible rotations for the C5-C6 vinyl
strand in 33 during the ring closure. The rotation a should be
favored, leading to isomer 35 that is the observed major
diastereomer, while rotation b is less favored for it leads to the
severe steric interaction between the R and the C2′-phenyl group.
This assertion also implies that the level of diastereoselectivity
should be a function of the size of the R group. The diastere-
omeric ratios of dihydropyridines 34 and 37-39 (Chart 1), with
R groups being Me, n-Pr, -(CH2)4, and cyclohexyl (or primary,
secondary, and tertiary carbons), respectively, reveal that this
trend is in part true with the sole exception of 39, suggesting
that there are other factors. In addition, the diastereoselectivity
for 40 is much lower despite the R group being a large phenyl
ring. This is actually in agreement with the aforementioned
conformational assumption and rotation preference. An extended
π-π interaction could exist with the C2′ phenyl ring, the C5-
C6 vinyl strand, and the R ()Ph) group in the 1-azatriene
precursor corresponding to the dihydropyridine 40, thereby
providing less steric interaction and lower diastereoselectivity.
The proposed conformation for the 1-azatriene intermediate
33 is key to the rotational preference illustrated above in
rationalizing the observed high diastereoselectivity. In addition
to the chemical shift data, the observation that any perturbations
of this conformational preference lead to loss of stereoselectivity
also lends support to this conformation. The steric interaction
(pseudo A1,3) between the methylene at C10 and the entire
substituent on the nitrogen atom appears to be significant.
Presumably, to alleviate this interaction, the substituent on the
nitrogen atom moves or rotates away, but as a result, a closer
proximity of the C2′ phenyl to the C5-C6 vinyl strand is gained.
This assertion is supported by the observation that the diaste-
reoselectivity is significantly reduced when the steric presence
of the C10 carbon is reduced to a sp2 carbon as in the 1-azatriene
precursor corresponding to the dihydropyridine 41 (Chart 1).
(2) Reversibility of the Ring Closure. While the rotational
preference provides one explanation for the observed diaste-
reoselectivity in the ring closure of 1-azatrienes, the reversibility
of ring closure likely plays a significant role because there is
ample precedent regarding the reversibility in the ring closure
Figure 3. Mechanistic proposal.
not readily apparent how such rotational preferences could be
evoked and what chirality inducing groups on the nitrogen atom
should be employed. However, on the basis of much trial and
error, it became clear that the protected 1-hydroxy-1,2-bisphe-
nylethyl group is unique in this stereoselective ring closure.
On the basis of the X-ray structure of the desilylated
dihydropyridine 34, a mechanistic model is proposed in Figure
3 to describe such rotational preferences. X-ray structure of 34
reveals that the C2′-N and C1′-OH are anti to one another
and that the two phenyl groups are completely anti to one
another with the C2′ phenyl group being in close proximity to
C6. The correlation of this solid-phase conformation to the one
possibly present in the solution phase would be that the chemical
shift of the C11 methylene in 34 is absolutely upfield shifted,
suggesting an anisotropic effect from the proximal C2′ phenyl
ring. This upfield shift was also observed in other dihydropy-
ridine substrates with a range of -0.11 to +0.27 ppm for the δ
values.
(23) For a review on rotational preferences leading to diastereomeric induction
during a 6π-electron electrocyclic ring closure, see: Okamura, W. H.; de
Lera, A. R. In ComprehensiVe Organic Synthesis; Trost, B. M., Fleming,
I., Eds.; Paquette, L. A., Vol. Ed.; Pergamon Press: New York, 1991; Vol.
5, pp 699-750.
(24) For some examples involving torquoselective processes, see: (a) Giese,
S.; Kastrup, L.; Stiens, D.; West, F. G. Angew. Chem., Int. Ed. Engl. 2000,
39, 1970. (b) Hsung, R. P.; Quinn, J. F.; Weisenberg, B. A.; Wulff, W. D.;
Yap, G. P. A.; Rheingold, A. L. J. Chem. Soc., Chem. Commun. 1997,
615.
9
J. AM. CHEM. SOC. VOL. 124, NO. 35, 2002 10439

A R T I C L E S
Sklenicka et al.
Scheme 7
Scheme 8
Scheme 9
of 1-heterotrienes. In addition, since the rotational preference
leads to the major product with the least amount of steric
congestion between the R and C2′ phenyl group, the major
isomer of 35 is the thermodynamically more stable one. ∆E
values from AM1 calculations (Spartan) for the major isomers
of 37-39 are shown in Scheme 7, and the ∆E value for 34 is
1.66 kcal mol-1 also in favor of the substituent at C2 being â.
We pursued the following studies to establish such reversibility,
thereby suggesting that the overall stereoselectivity is a result
of thermodynamic equilibration.
The first study was motivated by the observation that the
diastereomeric ratio (dr) of the dihydropyridine 39 did not follow
suit as suggested by the proposed rotational preference. Reheat-
ing a sample of 39 with a dr of 84:16 at 150 °C for 48 h in
toluene-d8 led to no significant loss of materials but a noticeably
improved dr of 93:7 in favor of the same major isomer (Scheme
7). This improved ratio provides a better agreement for the
assertion that the level of diastereoselectivity should be a
function of the size of the R group in lieu of the rotation
preference. It is also in good agreement in terms of the energetic
difference of major and minor isomers of 39 relative to those
of 37 or 38 (relative ∆E values).
The second study involved heating a sample of 40 that has a
ratio of 74:26 at 250 °C for 48 h. This in fact led to no
improvement of the dr, suggesting that having the R group being
a phenyl (sp2) indeed led to lower ratios than when R contained
an sp3 carbon and that it is actually lower than the ratio we had
obtained initially but reflects well of a potential prediction from
the ∆E value.
thermodynamic stability of the two diastereomers, a rotational
preference during the ring closure of 1-azatrienes also likely
plays a role in setting up an initial ratio close to the thermo-
dynamic ratio.
(2) Applications in Constructing cis-1-Azadecalin. Having
achieved a mechanistic understanding of this annulation reaction
of chiral vinylogous amides with R,â-unsaturated iminium salts,
we pursued a synthetic application to illustrate the utility of
dihydropyridines such as 7 in constructing cis-1-azadecalin
related to the dendrobatid alkaloid pumiliotoxin C (Figure
4).25-27
Such an approach represents an attractive and stereoselective
entry to cis-1-azadecalins because of its ability to achieve high
diastereomeric control of the stereogenic center at C2. To
achieve this task, two major issues need to be addressed. They
To further verify the reversibility of this ring closure involving
these 1-azatrienes, we heated samples of pure major isomer of
42 (g95:5 major to minor) at 150 °C for an overall of 60 h in
(25) (a) Daly, J. W.; Spande, T. F. In Alkaloids: Chemical and Biological
PerspectiVes; Pelletier, S. W., Ed.; Wiley-Interscience: New York, 1983;
Vol. 4, pp 124-162. (b) Garaffo, H. M.; Simon, L. D.; Daly, J. W.; Spande,
T. F.; Jones, T. H. Tetrahedron 1994, 50, 11329.
(26) For earlier total syntheses, see: (a) Oppolzer, W.; Frostl, W. HelV. Chim.
Acta 1975, 58, 593. (b) Oppolzer, W.; Fehr, C.; Warneke, J. HelV. Chim.
Acta 1977, 60, 48. (c) Oppolzer, W.; Flaskamp, E. HelV. Chim. Acta 1977,
60, 204. (d) Overman, L. E.; Jessup, P. J. J. Am. Chem. Soc. 1978, 100,
179.
(27) For recent approaches and total syntheses, see: (a) Akashi, M.; Sato, Y.;
Mori, M. J. Org. Chem. 2001, 66, 7873. (b) Oppolzer W.; laskamp E.;
Bieber L. W. HelV. Chim. Acta 2001, 84, 141. (c) Padwa, A.; Heidelbaugh,
T. M.; Kuethe, J. T. J. Org. Chem. 2000, 65, 2368. (d) Shieh, Y. S.; Yeh,
M. C. P.; Rao, U. N. J. Chin. Chem. Soc. (Taipei) 2000, 47, 283. (e) Kunz,
H.; Weymann, M.; Follmann, M.; Allef, P.; Oertel, K.; Schultz-Kukula,
M.; Hofmeister, A. Pol. J. Chem. 1999, 73, 15. (f) Riechers, T.; Krebs, H.
C.; Wartchow, R.; Habermehl, G. Eur. J. Org. Chem. 1946, 11, 2641. (g)
Back, T. G.; Nakajima, K. J. Org. Chem. 1998, 63, 6566. (h) Toyota, M.;
Asoh, T.; Fukumoto, K.; Heterocycles 1997, 45, 147. (i) Naruse, M.;
Aoyagi, S.; Kibayashi, C. J. Chem. Soc., Perkin Trans. 1 1996, 11, 2007.
(j) Davies, S. G.; Bhalay, G. Tetrahedron: Asymmetry 1996, 7, 1595. (k)
Mehta, G.; Praveen, M. J. Org. Chem. 1995, 60, 279. (l) Comins, D. L.;
Dehghani, A. J. Chem. Soc., Chem. Commun. 1993, 1838. (m) Paulvannan,
K.; Stille, J. R. Tetrahedron Lett. 1993, 34, 6673.
1
toluene-d8. The reaction was monitored using H NMR, and
ratios as a function of time are displayed in Scheme 8. Although
the rate of equilibration through a sequence of ring opening
and ring closure appeared to be slow, it is unambiguous that
the ring closure is reversible and ultimately led to 42a/b with
a ratio befitting of its original thermodynamic distributions. A
similar study was carried out using an enriched minor isomer
of 34 (28:72 major to minor) in toluene-d8 for 66 h, but
temperature was varied (Scheme 9). Although the final isomeric
ratio of 34 eventually matched that observed from the reaction,
the rate of equilibration from the minor isomer to the major
isomer appeared to slow especially at a temperature well below
that of 150 °C.
These studies suggest that the reversibility of ring closure
ultimately provides the diastereomeric ratio based on the
9
10440 J. AM. CHEM. SOC. VOL. 124, NO. 35, 2002

Cycloaddition Approach to cis-1-Azadecalins
A R T I C L E S
Scheme 11
Figure 4. Constructions of cis-1-azadecalins.
Scheme 10
(OH)2 could be carried out to afford 44 in 70% overall yield
from 7, and the stereointegrity at C2 of 44 did not appear to
suffer.
Attempts to hydrogenate 44 using various conditions failed
but led to the speculation that the nitrogen atom may need to
be protected. Capping the nitrogen atom could be achieved using
either Ac2O or TFAA to give 49a and 49b (Scheme 11). In the
case of using Ac2O, some enol acetate 48a was also found but
could be readily hydrolyzed to 49a under mild acidic conditions.
Having protected the vinylogous amide nitrogen, high-pressure
hydrogenation of 49a and 49b proceeded smoothly to give a
mixture of alcohols 50a and 50b and ketones 51a and 51b,
respectively. Further Dess-Martin periodinane (DMP) oxidation
of the respective mixtures gave 51a in 70% overall yield from
49a and 51b in 68% overall yield from 49b. Both 51a and 51b
were isolated as single diastereomers, and no detectable amount
of the corresponding minor isomers was observed in 1H NMR.
This completes a highly stereoselective transformation of the
dihydropyridine 7 to cis-1-azadecalins 51a and 51b.
The stereochemical outcome deserves some comments since
the relative stereochemistry based on nOe experiments (see
Supporting Information) revealed that it is anti at C2 and C8a
in 51a and 51b. Evidently, out of four possible conformations
of 49a (or 49b), the conformer D clearly predominated to
provide the observed stereochemistry in the hydrogenation
(Figure 5). Conformer D is favored (Spartan AM1 calcula-
tions: a minimum of 2.01 kcal mol-1 over conformers A-C)
because it alleviates much of the well-known pseudo A1,2 strain
(or the gauche interaction) between the N-acyl group and the
n-Pr group at C2.30 This presents a rare and highly stereose-
lective entry to such anti relative stereochemistry at C2 and C8a
of cis-1-azadecalins.
entail (1) removal of the chiral inducing group that was not
trivial and (2) a stereoselective reduction of the C4a-C8a
tetrasubstituted olefin.
As shown in Scheme 10, the dihydropyridine 7 could be
readily brought up in large quantities, and an ensuing hydro-
genation would give 45 in 94% yield. Tebbe’s olefination28 was
attempted on the vinylogous amide carbonyl of 45 but failed to
give any olefination product 46. Removal of the TBS group
followed by hydrogenation using Pd(OH)2 and NH4OCHO29 at
140 °C successfully gave the key 1-azadecalinone 44 in 68%
overall yield from 7. The C3-C4 olefin in 7 was quickly
removed at first due to the fear of ring opening under various
reaction conditions, leading to loss of stereointegrity at C2.
However, a more efficent sequence via TBAF desilylation of 7
followed by a single hydrogenation using NH4OCHO and Pd-
cis-1-Azadecalinones 51a and 51b could be further elaborated
to provide (-)-4a,8a-diepi-pumiliotoxin along with 2-epi-(+)-
(28) (a) Cannizzo, L. F.; Grubbs, R. H. J. Org. Chem. 1985, 50, 2316. (b) Pine,
S. H.; Zahler, R.; Evans, D. A.; Grubbs, R. H. J. Am. Chem. Soc. 1980,
109, 3270.
(30) Kolocouris, A.; Outeirino, J. G.; Anderson, J. E.; Fytas, G.; Foscolos, G.
(29) Ram, S.; Spicer, L. D. Tetrahedron Lett. 1987, 28, 515.
B.; Kolocouris N. J. Org. Chem. 2001, 66, 4989.
9
J. AM. CHEM. SOC. VOL. 124, NO. 35, 2002 10441

A R T I C L E S
Sklenicka et al.
Conclusion
We have described here mechanistic evidence for the first
highly stereoselective 6π-electron electrocyclic ring closure of
1-azatrienes, a key step in formal [3 + 3] cycloaddition or
annulation reactions of chiral vinylogous amides with R,â-
unsaturated iminium salts. We have unambiguously demon-
strated that the 6π-electron electrocyclic ring closure is revers-
ible, leading to the thermodynamically more stable major isomer
and that a rotation preference also plays a role in this
stereoselective ring closure. A synthetic application is illustrated
here to transform the resulting dihydropyridines to cis-1-
azadecalins with unique anti relative stereochemistry at C2 and
C2a, leading to an efficient synthesis of epi isomers of (-)-
pumiliotoxin C.
Acknowledgment. R.P.H. thanks NIH (Grant NS38049) and
American Chemical Society Petroleum Research Fund (Type-
G) for financial support, and the McKnight Foundation for a
McKnight Fellowship. R.P.H. would like to thank Professor Bill
Okamura for insightful discussions, for teaching us the potential
significance of ring closure of 1-azatrienes, and for encouraging
us to pursue this study.
Supporting Information Available: Experimental procedures
and characterization data for all new compounds and copies of
1H NMR (PDF). This material is available free of charge via
the Internet at http://pubs.acs.org.
Figure 5. Proposed hydrogenation mechanism.
pumiliotoxin C,20 thereby illustrating synthetic potential of this
formal [3 + 3] cycloaddition methodology in constructing cis-
1-azadecalin related natural products.
JA020698B
9
10442 J. AM. CHEM. SOC. VOL. 124, NO. 35, 2002