aza-prins-pinacol approach to 7-azabicyclo[2.2.1]heptanes and ring expansion to [3.2.1]tropanes

Article abstract of DOI:10.1021/ol0501267

(Chemical Equation Presented) The 7-azabicyclo[2.2.1]heptane ring system can be rapidly accessed from 5-(1-hydroxyallyl)-2-alkoxy-N-tosylpyrrolidines via an unusual aza-Prins-pinacol reaction mediated by Lewis acid. The products can undergo ring expansion

Full text of DOI:10.1021/ol0501267

ORGANIC
LETTERS
2005
Vol. 7, No. 7
1335-1338
aza-Prins-pinacol Approach to
7-Azabicyclo[2.2.1]heptanes and Ring
Expansion to [3.2.1]Tropanes
Alan Armstrong* and Stephen E. Shanahan
Department of Chemistry, Imperial College London, South Kensington,
London SW7 2AZ, UK
a.armstrong@imperial.ac.uk
Received January 20, 2005
ABSTRACT
The 7-azabicyclo[2.2.1]heptane ring system can be rapidly accessed from 5-(1-hydroxyallyl)-2-alkoxy-N-tosylpyrrolidines via an unusual aza-
Prins-pinacol reaction mediated by Lewis acid. The products can undergo ring expansion to isomeric tropanones. These reactions show
promise for a concise entry to biologically relevant azabicyclic targets.
Nitrogen-bridged bicyclic ring systems have a long associa-
tion with natural product chemistry and pharmaceutical
products.1 The tropane nucleus (Figure 1) is common to both
tion of the analgesic natural product epibatidine 3 in 1992.3
Numerous syntheses of epibatidine and analogues have been
reported,4 although these are rarely suitable for the prepara-
tion of diverse analogues. An exception is the reductive Heck
coupling strategy reported by Regan.5 However, this is most
suited to varying the pyridyl fragment.
We recently reported6 that aminative rearrangement of
alkoxydihydropyrans 4 gives facile access to pyrrolidines 5
(Scheme 1). The method is attractive because the substrates
Scheme 1
Figure 1.
atropine 1 and cocaine 2 and continues to be exploited today.2
By contrast, interest in the 7-azabicyclo[2.2.1]heptane skel-
eton has increased considerably since the structural elucida-
4 can be readily accessed via inverse electron demand [4 +
2] cycloaddition between an enol ether and an enone.
(1) Silverman, R. B. The Organic Chemistry of Drug Design and Drug
Action; Academic Press: San Diego, 1992; p 2.
(2) Langlois, M.; Fischmeister, R. J. Med. Chem. 2003, 46, 319-344.
(3) Spande, T. F.; Garraffo, H. M.; Edwards, M. W.; Yeh, H. J. C.;
Pannell, L.; Daly, J. W. J. Am. Chem. Soc. 1992, 114, 3475-3478.
10.1021/ol0501267 CCC: $30.25
© 2005 American Chemical Society
Published on Web 03/05/2005

Additionally, where a substituent R is present, we have
demonstrated that the C4-C5 relative configuration of the
product can be predictably controlled by the choice of
appropriate aziridination conditions.6 Compounds 5 are
potentially bifunctional electrophiles: as well as containing
a carbonyl group, they are precursors to N-sulfonyl iminium
ions, which are competent electrophilic species.7 We sought
to exploit the bifunctional nature of pyrrolidines 5 in synthetic
applications. We conceived that addition of vinyl organo-
metallic reagents to the aldehyde group in 5a should give
access to allylic alcohols 6 (Scheme 2). Subsequent acidic
Table 1.
isolated yield
of 7a (%)
entry
acid
molar equiv
conditions
1
2
3
4
5
6
CSAa
TiCl4
Sc(OTf)3
SnCl4
SnCl4
SnCl4
0.08
1
0.1
2
1
0.1
80 °C, 16 h
0 °C, 8 min
25 °C, 17 h
0 °C, 8 min
0 °C, 8 min
25 °C, 15 h
0
31
0
96
96
nd
Scheme 2
a Acetonitrile solvent.
(Stereochemical configurations of compounds 7 were as-
signed on the basis of NOE studies and/or coupling constant
data. See Supporting Information for details.) Optimization
of the reaction identified preferred conditions (entry 5)
employing 1 equiv of SnCl4 at 0 °C for 8 min. Attempts
toward a catalytic process with substoichiometric levels of
SnCl4 were unsuccessful (entry 6). At no time was the
alternative exo stereoisomer observed. Interestingly, pro-
longed exposure to SnCl4 was found to bring about the clean
formation of unexpected alternative products. These were
identified as isomeric tropanes 8a and 9a (Table 3, entry 1).
activation of the masked N-sulfonyliminium function could
be predicted to promote a novel variation on the aza-Cope-
Mannich sequence pioneered by Overman.8 The expected
products would be 7-azabicyclo[2.2.1]heptanes 7, the relative
configuration of which was left as a question to address
experimentally. This plan would constitute a flexible syn-
thesis of the bicycle 7 from alkoxydihydropyrans 4 in only
three chemical steps.
Table 2.
An initial test substrate 6a was prepared, as a mixture of
diastereoisomers, via reaction of pyrrolidine aldehyde 5a with
the Grignard reagent derived from R-bromo-styrene (Scheme
3). Several Bronsted and Lewis acids were screened (Table
sub-
entry strate
temp yield
X
Y
Z
time
(°C)
(%)
1
2
3
4
5
6
7
6a
6b
6c
6d
6e
6f
Ph
H
H
H
Me
H
H
H
8 min
8 min
0
0
0
0
20
0
96
81
62
61
32
11
0
Me
Me
Me
H
Me 8 min
Scheme 3
H
H
H
H
8 min
14.5 h
1 min
8 min
SPh
cyclohexenyl
H
6g
0
By extension of the reaction time at ambient temperature, a
high total yield (76%) of these [3.2.1]bicyclic compounds
(4) Representative examples: Broka, C. A. Tetrahedron Lett. 1993, 34,
3251-3254. Fletcher, S. R.; Baker, R.; Chambers, M. S.; Herbert, R. H.;
Hobbs, S. C.; Thomas, S. R.; Verrier, H. M.; Watt, A. P.; Ball, R. G. J.
Org. Chem. 1994, 59, 1771-1778.
1), and gratifyingly SnCl4 was found to bring about the
desired transformation to azabicycloheptane 7a in high yield.
Only one diastereoisomer of 7a was observed, possessing
endo stereochemistry with respect to the phenyl group.
(5) Clayton, S. C.; Regan, A. C. Tetrahedron Lett. 1993, 34, 7493-
7496.
(6) Armstrong, A.; Cumming, G. R.; Pike, K. Chem. Commun. 2004,
812-813.
(7) Ahman, J.; Somfai, P. Tetrahedron Lett. 1992, 33, 3791-3794.
1336
Org. Lett., Vol. 7, No. 7, 2005

is due to the rapid acceleration of ring expansion to unstable
tropanes observed with this substrate, making halting the
reaction at the azabicycloheptane stage difficult. For all
substrates, only one diastereoisomer of 7 was observed, exo
with respect to the aldehyde substituent, despite precursors
6 being diastereoisomeric mixtures. Substrates 6c and 6d
(entries 3 and 4) demonstrate that the reaction is stereo-
specific with respect to olefin geometry in 6, because they
exclusively afforded products 7c and 7d, respectively.
Substrate 6g, possessing a cyclic alkenyl fragment, failed to
rearrange, giving a complex product mixture (entry 7).
Formation of tropanes 8 or 9 (Table 3) was also tolerant
of substituent X being changed from phenyl to methyl
(entries 2-4). However, employing an unsubstituted vinyl
group (entry 5) halted the reaction at azabicycloheptane 7e
as previously described. Given that ring expansion to
tropanes 8/9 is observed to be a slower process than
formation of 7, this result is not altogether surprising in view
of the already slow rate of azabicyclo[2.2.1]heptane forma-
tion observed with this substrate. Olefin geometrical isomers
6c and 6d again produced different isomeric tropane products
(entries 3 and 4). This is the expected outcome of the
proposed ring expansion mechanism for isomerization of 7
to 8/9 (Scheme 4): connectivity of the bicyclic ring system
is maintained throughout, maintaining the syn or anti
disposition of substituents Y/Z relative to the nitrogen bridge.
(Proposed structures for 8c, 9c, and 8d are corroborated by
key spectral features. See Supporting Information for details.)
The complete diastereoselectivity and stereospecificity
(with regard to alkene geometry) of formation of products 7
prompted us to consider alternative mechanistic explanations
for their generation. Two pathways were prominent in our
consideration (Scheme 5): an aza-Cope-Mannich process
Table 3.
entry
substrate
yield 8 (%)
yield 9 (%)
1
2
3
4
5
6a
6b
6c
6d
6e
60
37
49
42
0a
16
17
trace
0
0
a No ring expansion observed after formation of 7e as outlined in Table
2.
was obtained. 8a and 9a proved to be readily separable by
chromatography. It was demonstrated that these tropanes
arose via the intermediacy of azabicycloheptane 7a: a pure
sample of 7a was resubjected to the reaction conditions and
smoothly isomerized to the ring-expanded system. This novel
interconversion of the two ring systems can be postulated
to be mechanistically analogous (Scheme 4) to the recently
Scheme 4
Scheme 5
observed carbocyclic variant reported by Davies et al.9 The
ca. 4:1 isolated ratio of 8a:9a partly reflects epimerization
during chromatography; the initial ratio according to NMR
analysis of the crude reaction mixture was close to 1:1.
With preferred reaction conditions identified to access
either the [2.2.1] or expanded [3.2.1] ring systems, we
synthesized a range of alternative rearrangement precursors
6 (Tables 2 and 3). We hoped to probe the generality of the
rearrangement and to shed further light on the mechanisms
of conversion to products 7-9. The strategy employed for
preparation of 6b-g was again addition of the appropriate
vinyl Grignard reagent to aldehyde 5a.
With respect to the formation of azabicycloheptanes 7
(Table 2), replacement of the phenyl vinyl substituent of 6a
with an aliphatic group (methyl) is well tolerated (entries
2-4). However, the presence of an unsubstituted vinyl group
(entry 5) dramatically slowed the formation of 7e. The low
yield of 7f from electron-rich vinyl compound 6f (entry 6)
as originally planned (pathway I) and an aza-Prins-pinacol
reaction (pathway II). In pathway II, the carbocation
intermediate B must undergo rapid pinacol rearrangement
to give the more strained [2.2.1]azabicycle 7 as the kinetic
product; prolonged exposure to Lewis acid would lead to
the thermodynamically more stable, less strained [3.2.1]-
system 8/9 via reversion to B and H-migration/enol formation
(cf. Scheme 4). In both pathways I and II, the observed exo
(8) Overman, L. E.; Kakimoto, M. J. Am. Chem. Soc. 1979, 101, 1310.
For a review, see: Overman, L. E. Acc. Chem. Res. 1992, 25, 352-359.
(9) Davies, H. M. L.; Dai, X. J. Am. Chem. Soc. 2004, 126, 2692-
2693. Niess, B.; Hoffmann, H. M. R. Angew. Chem., Int. Ed. 2005, 44,
26-29.
Org. Lett., Vol. 7, No. 7, 2005
1337

orientation of substituent Z in compounds 7 requires a
chairlike topography for whichever process consumes the
intermediate iminium ion A. Overman and co-workers
proposed that pathway I is generally operative for systems
containing the more usual N-alkyl iminium ions.10 They
proposed that pathway II can occur for electronically biased
substrates such as vinyl ethers10 and showed that an
analogous mechanism is the norm for the related Prins-
pinacol reaction of oxonium ions.11 Key evidence presented
by Overman for pathway I is the observation that aza-Cope-
Mannich reactions tolerate a sulfone vinyl substituent10
(Scheme 5, substituent X). The formation of a carbocation
R to a sulfone group X, required by pathway II, is unlikely.12
As a test of mechanism and direct comparison with the work
of Overman, we prepared sulfone 6h and attempted its
rearrangement under our preferred reaction conditions for
the formation of azabicycloheptanes (Scheme 6). No re-
secondary rather than tertiary carbocation would be invoked.
The aza-Prins-pinacol mechanism is an attractive explanation
for the formation of azabicycloheptanes 7, since it readily
explains the observed stereochemical configuration of the
products and the lack of dependence on the configuration
of substrates 6. Should an aza-Cope-Mannich process operate
(Scheme 5, pathway I), stereoconvergence to a single isomer
of product 7 would require enol intermediate C to react
exclusively in the conformation shown, regardless of its E/Z
geometry. The pinacol process would be syn stereospecific,
irrespective of the configuration at the alcohol stereocenter
in carbocation intermediate B (Scheme 5, pathway II). Hence,
a single isomer of product 7 (aldehyde exo) would be formed.
In summary, we report the stereocontrolled formation of
the 7-azabicyclo[2.2.1]heptane nucleus via an unusual aza-
Prins-pinacol reaction. This mechanistic rationale is lent
credence by both the influence of substrate electronics and
the observed stereoselectivity and stereospecificity of the
reaction. The aldehyde products can undergo a novel ring
expansion to isomeric tropanes. Both ring systems are of
considerable industrial importance and can be accessed
concisely with multiple opportunities for structural diversity.
The ready availability of the starting materials 4 bearing a
wide range of substituents, including the possibility of
accessing them in enantiomerically pure form using catalytic
asymmetric hetero-Diels-Alder chemistry,13 is a highly
attractive feature. Synthetic applications of this methodology
to natural product synthesis are under investigation in our
laboratory.
Scheme 6
arrangement product was observed, the sulfonyl iminium
intermediate merely being trapped by the alcohol function
to give isolated product 10. We therefore suggest that the
formation of products 7 occurs via an aza-Prins-pinacol
process (Scheme 5, pathway II). Similar argument around
the need to form carbocationic intermediate B in an aza-
Prins-pinacol process can explain the observed dramatic rate
deceleration for substrate 6e (Table 2, entry 5), where a
Acknowledgment. We gratefully acknowledge support
from the EPSRC (studentship for S.E.S.) and unrestricted
funding from Bristol-Myers Squibb, Merck, and Pfizer.
Supporting Information Available: Full experimental
and characterization for compounds 7-10. This material is
available free of charge via the Internet at http://pubs.acs.org.
(10) Jacobsen, E. J.; Levin, J.; Overman, L. E. J. Am. Chem. Soc. 1988,
110, 4329-4336.
OL0501267
(11) Hopkins, M. H.; Overman, L. E.; Rishton, G. M. J. Am. Chem.
Soc. 1991, 113, 5354-5365.
(12) Creary, X.; Mahrsheikh-Mohammadi, E.; Eggers, M. D. J. Am.
Chem. Soc. 1987, 109, 2435-2442.
(13) Gademann, K.; Chavez, D. E.; Jacobsen, E. N. Angew. Chem., Int.
Ed. 2002, 41, 3059-3061.
1338
Org. Lett., Vol. 7, No. 7, 2005