Nucleophilic opening of oxabicyclic ring systems

Full text of DOI:10.1002/anie.200902046

Communications
DOI: 10.1002/anie.200902046
Synthetic Methods
Nucleophilic Opening of Oxabicyclic Ring Systems**
Corinna S. Schindler, Stefan Diethelm, and Erick M. Carreira*
The ease of access to functionalized oxabicyclic ring systems
renders them attractive starting points for chemical syn-
thesis.[1] Their strategic incorporation in a synthesis can lead
to considerable simplification of the route to a given target
and provide key building blocks for asymmetric synthesis. In
the context of our recent studies aimed at the synthesis of the
banyaside and suomilide core structures,[2,3] we noted an
unexpected transformation involving the opening of an
oxabicyclo[2.2.1]heptene intermediate. Further investigations
led us to identify unprecedented reactivity of oxabicyclic rings
(Scheme 1). Herein, we document that oxabicyclo-
In the course of our synthetic efforts towards the total
synthesis of banyasides A and B and suomilide, we used an
oxabicyclo[2.2.1]heptene in the rapid assembly of the
common stereochemically complex core. In the reported
strategy, the intramolecular olefin amination of 9 was
À
followed by remote C O bond cleavage to give 10
(Scheme 2), a process which may be described formally as
Scheme 2. Strategies for the cleavage of oxabicyclo[2.2.1]heptenes.
an allylic displacement reaction. The precursor oxabicyclo-
[2.2.1]heptene 9 proved to be readily accessible through the
implementation of a tandem Diels–Alder/aldol sequence.[4] In
the course of these studies, we explored numerous methods
for the direct conversion of 9 into 10. One particular line of
investigation included the examination of RhI catalysis. This
approach was inspired by studies by Lautens and co-workers,
who pioneered the use of metal catalysts and reagents for the
opening of oxabicyclic rings.[5,6] In initial experiments, we
observed that the treatment of 9 with a catalyst formed by the
combination of (R,S)-josiphos with RhI in the presence of
AgOTf and Bu4NI afforded the tricyclic product 11 in 65%
yield. Although 11 was an undesired product, we proceeded
to examine this transformation further, as it had not been
documented previously. Moreover, rhodium-mediated prod-
uct formation was unexpected on the basis of observations
with and the mechanistic constructs suggested for related
systems.[7]
Scheme 1. Ring-opening reaction leading to hexahydroindoles 3 and
octahydroquinolines 4, as well as perhydroindoles 7 and perhydroiqui-
nolines 8.
[2.2.1]heptenes 1 and 2 are converted into hexahydroindoles
3 and octahydroquinolines 4, respectively, in 50–95% yield
upon treatment with TMSOTf/NEt3. Astonishingly, the ring
opening of saturated oxabicyclo[2.2.1]heptanes 5 and 6 (Y=
H or OH) took place with equal ease. This unusual reactivity
opens up rapid access to functionalized reduced indoles and
quinolines.
Additional studies of the conversion of oxabicyclo-
[2.2.1]heptene 9 into 11 proved revealing (Table 1). The use
of the cationic catalyst [Rh(cod)2OTf][8] led to the isolation of
11 in slightly improved yield (67%) and to the hypothesis that
the Rh complex might be functioning as an electrophilic
activator. Consequently, we screened a number of Lewis
acids, including SnCl4, TiCl4, BF3·OEt2, MgBr2, BBr3, and silyl
triflates; only BBr3 and the silyl triflates gave leading results.
The treatment of 9 with BBr3 furnished 11 in 20% yield.
By contrast, the exposure of 9 to TMSOTf/Et3N afforded 11
in 95% yield. When the more hindered tert-butyldimethylsilyl
triflate was employed, some conversion was observed, but at a
considerably reduced rate, to give the product in 36% yield.
Interestingly, the use of a mixture of TfOH/NEt3 did not lead
[*] C. S. Schindler, S. Diethelm, Prof. Dr. E. M. Carreira
Laboratorium fꢀr Organische Chemie, ETH Zꢀrich, HCI H335
Wolfgang-Pauli-Strasse 10, 8093 Zꢀrich (Switzerland)
Fax: (+41)44-632-1328
E-mail: carreira@org.chem.ethz.ch
[**] C.S.S. thanks the Roche Research Foundation for a predoctoral
fellowship. S.D. thanks Novartis for a masters fellowship. The
research was also supported by a grant from the Swiss National
Science Foundation (NF5/2-7723408). We are grateful to Dr.
W. Bernd Schweizer for the X-ray crystallographic analysis of the
derivative of 20b.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200902046.
6296
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 6296 –6299

Angewandte
Chemie
Table 2: Lewis acid mediated oxabicyclo[2.2.1]heptene[9] ring opening.[a]
Table 1: Screening of reaction conditions for the conversion of 9 into 11.
Entry
Substrate
Product
Yield [%]
Entry
Conditions
Yield [%]
1
2
3
4
5
6
josiphos,[a]·[{Rh(cod)Cl}2], AgOTf, Bu4NI
josiphos,[a]·[Rh(cod)2OTf]
BBr3
TMSOTf, Et3N
TBDMSOTf, Et3N
TfOH, Et3N
65
67
20
95
36
1
95
92
[b]
–
9
11
[a] Josiphos=(S)-1-[(RP)-2-(di-tert-butylphosphanyl)ferrocenyl]ethyl-
bis(2-methylphenyl)phosphane. [b] No reaction occurred; the starting
material was recovered. cod=1,5-cyclooctadiene, TBDMS=tert-butyldi-
methylsilyl, Tf=trifluoromethanesulfonyl, TMS=trimethylsilyl.
2
12a
12b
to product formation. This last result suggests that a Brønsted
acid is not the active agent, and implicates the Lewis acidic
silyl group as the catalyst.
3
4
5
67
We then proceeded to investigate the scope of the ring-
opening reaction of additional oxabicyclo[2.2.1]heptenes. The
reactions were conducted in CH2Cl2 at 238C in the presence
of TMSOTf/Et3N (Table 2). Like the free amine 9, which
formed the basis of our preliminary experiments, amide
analogues were converted into the desired products in good to
excellent yields (Table 2, entries 2 and 3). The transformation
of 13a demonstrated that the observed ring-opening reaction
is not unique to systems constrained by an oxazolidinone ring.
The amide 14a also underwent ring opening (Table 2,
entry 4). Interestingly, internal competition between two
amides led to preferential formation of the quinoline nucleus
15b (six-membered-ring formation) over the indole counter-
part (five-membered-ring formation; Table 2, entry 5). Only
under the reaction conditions with TMSOTf/Et3N were 12b–
15b obtained; the starting materials were recovered when Rh
catalysts were used.
13a
14a
15a
13b
14b
15b
73
50[b]
[a] Reaction conditions: substrate (0.07 mmol), TMSOTf (4.0 equiv),
Et3N (5.0 equiv), CH2Cl2 (3 mL), 238C. [b] TBDMSOTf (4.0 equiv), Et3N
(5.0 equiv), CH2Cl2 (3 mL), 238C; 50% based on recovered starting
material, 80% conversion, 20% starting material recovered.
Although the cyclizations detailed in Table 2 are unpre-
cedented, they can be rationalized readily as a consequence of
the electrophilic activation of an allylic leaving group.[10] To
examine the extent to which ring opening is facilitated by the
embedded 5,6-olefin, we investigated the ring opening
reaction of unactivated oxabicyclic substrates in which the
NMR spectroscopic studies to provide further confirmation
of the product structures.[13] We speculate that the energetic
driving force for this unprecedented transformation stems
from the strain energy, estimated at 6.5 kcalmolÀ1, of the
oxabicycloheptanes.[14]
Several features of oxabicyclo[2.2.1]heptenes and -hep-
tanes have previously been cleverly exploited in the elabo-
ration of both with predictable levels of regio- and stereose-
lectivity. Examples include the inherent facial bias of an
embedded olefin (endo versus exo) as well as the stereoelec-
tronic influences of remote groups.[15] The same aspects may
now be used in combination with the transformation we have
described to provide access routes to ring structures common
to a number of alkaloids, such as the Amaryllidaceae
alkaloids[16] galwesine, galasine, and galanthine.
À
C5 C6 bond is saturated. On the basis of the known reactivity
profile of these systems and ethers in general, we did not
anticipate the results observed with these substrates.
The analogous cyclization reaction proceeded quite read-
ily in the absence of the olefin (Table 3) to afford [4.3.0] and
[4.4.0] bicyclic products. The substrate scope is broad with
respect to the nature of the nucleophilic group: both amides
(Table 3, entries 1 and 3–4) and amines (entry 2) can be used,
as well as inverse amides (entries 5 and 6). Furthermore, the
cyclization proceeds well in the absence of entropic con-
straints (Table 3, entries 5 and 6), and the presence of an
additional electron-withdrawing substituent does not impair
the cyclization (Table 3, entry 4). The structure of the ring-
opening products was confirmed by X-ray crystallographic
analysis of the p-bromobenzoate ester of 20b[11,12] (Table 3,
entry 5). Having secured this structure, we were able to show
that the reduction of 14b (Table 2, entry 4; Pd/C, H2) afforded
20b. This structural correlation complemented extensive
In conclusion, we have documented novel access to highly
functionalized [4.3.0] and [4.4.0] bicyclic structures. This
approach stems from the unexpected observation of a Lewis
acid mediated ring-opening reaction of oxabicycloheptenes
Angew. Chem. Int. Ed. 2009, 48, 6296 –6299
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 6297

Communications
Table 3: Lewis acid mediated oxabicyclo[2.2.1]heptane[9] ring opening.[a]
mixture was purified by flash chromatography (CH2Cl2/MeOH 95:5)
to afford the corresponding product.
Entry Substrate
Product
Yield [%]
Received: April 16, 2009
Published online: July 15, 2009
1
89
Keywords: asymmetric synthesis · Lewis acids ·
perhydroindoles · perhydroquinolines · ring opening
.
16a
16b
17b
[1] a) P. Chiu, M. Lautens, Top. Curr. Chem. 1997, 190, 1; b) K. R.
Crawford, S. K. Bur, C. S. Straub, A. Padwa, Org. Lett. 2003, 5,
3337; c) A. Padwa, M. D. Weingarten, Chem. Rev. 1996, 96, 223.
[2] a) A. Pluotno, S. Carmeli, Tetrahedron 2005, 61, 575; b) K. Fujii,
K. Sivonen, K. Adachi, N. Kazuyoshi, Y. Shimizu, H. Sano, K.
Hirayama, M. Suzuki, K. Harada, Tetrahedron Lett. 1997, 38,
5529; c) for a recent review on the chemistry of the aeruginosins,
see: K. Ersmark, J. R. Del Valle, S. Hanessian, Angew. Chem.
2008, 120, 1220; Angew. Chem. Int. Ed. 2008, 47, 1202.
[3] C. S. Schindler, C. R. J. Stephenson, E. M. Carreira, Angew.
Chem. 2008, 120, 8984; Angew. Chem. Int. Ed. 2008, 47, 8852.
[4] E. J. Corey, T.-P. Loh, J. Am. Chem. Soc. 1991, 113, 8966.
[5] a) M. Lautens, K. Fagnou, Proc. Natl. Acad. Sci. USA 2004, 101,
5455; b) K. Fagnou, M. Lautens, Angew. Chem. 2002, 114, 26;
Angew. Chem. Int. Ed. 2002, 41, 26; c) M. Lautens, K. Fagnou, S.
Hiebert, Acc. Chem. Res. 2003, 36, 48, and references therein;
d) M. Lautens, K. Fagnou, T. Rovis, J. Am. Chem. Soc. 2000, 122,
5650.
2
89[b]
17a
3
91
18a
18b
4
88
80
78
[6] K. Fagnou in Modern Rhodium-Catalyzed Organic Reactions
(Ed.: P. A. Evans), Wiley-VCH, Weinheim, 2005, pp. 173.
[7] a) M. Lautens, K. Fagnou, J. Am. Chem. Soc. 2001, 123, 1770;
b) M. Lautens, K. Fagnou, D. Yang, J. Am. Chem. Soc. 2003, 125,
14884.
19a
19b
20b
21b
5
[8] R. Webster, C. Bꢀing, M. Lautens, J. Am. Chem. Soc. 2009, 131,
444.
[9] The oxabicyclic substrates were prepared by an enantioselective
Diels–Alder reaction of furan and bromoacrolein with the
tryptophan-derived oxazaborolidine catalyst described by Corey
and Loh (see references [3] and [4]). The l-tryptophan-derived
catalyst was used for the reactions in entries 3 and 4 of Table 2
and entries 5 and 6 of Table 3; the d-tryptophan-derived catalyst
was used for the reactions in entries 1 and 2 of Table 2 and
entries 1–4 of Table 3.
20a
6
21a
[10] To the best of our knowledge, there is only one example in the
literature of the opening of an oxabicycloheptene by displace-
[a] Reaction conditions: substrate (0.07 mmol), TMSOTf (4.0 equiv),
Et3N (5.0 equiv), CH2Cl2 (3 mL), 238C. [b] The product was isolated as
the hydrochloride salt.
À
ment of a bridging C O bond, namely, the ring opening of 2-
methylthio-7-oxabicyclo[2.2.1]heptene by silyl enol ethers in the
presence of TBDMSOTf: I. Yamamoto, K. Narasaka, Chem.
Lett. 1995, 1129. The ring opening proceeds by cleavage of the
and heptanes. The ease with which these previously unex-
plored transformations take place is remarkable and opens
new possibilities for strategic planning in complex-molecule
synthesis.
À
C O bond, and net substitution occurs with retention of
configuration. The stereochemical result underscores the fact
À
that the reaction proceeds by heterolytic cleavage of the C O
bond to give a secondary carbocation, which is stabilized
significantly by the vinyl sulfide.
[11] Crystallographic data for the p-bromobenzoate ester of 20b:
C23H24BrNO4, Mr = 458.35, orthorhombic, space group P212121,
a = 9.7647 (2), b = 10.6347 (3), c = 19.7861 (5) ꢁ, V= 2054.68
(9) ꢁ3, 1calcd = 1.482 MgmÀ3, T= 223 K, reflections collected:
18793, independent reflections: 4693 (R(int) = 0.15), R(all) =
0.0678, wR(gt) = 0.1315. CCDC 721256 contains the supplemen-
tary crystallographic data for this paper. These data can be
obtained free of charge from The Cambridge Crystallographic
Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
Experimental Section
The oxabicycle (0.07 mmol) was dissolved in anhydrous CH2Cl2
(3 mL) at 238C under argon. Triethylamine (51 mL, 0.4 mmol,
5 equiv) was added, followed by freshly distilled TMSOTf (53 mL,
0.3 mmol, 4 equiv, 95%, Aldrich). The resulting solution was stirred
at 238C, and the reaction was monitored by TLC. After completion
(2–5 h), the reaction mixture was filtered through a plug of silica, and
the filtrate was evaporated. The residue was dissolved in CH2Cl2
(5 mL) and treated with a solution of HCl in MeOH (4 equiv, 1.0m,
Fluka). After 30 min, the solution was evaporated, and the crude
[12] ORTEP representation of the p-bromobenzoate ester of 20b
(probability ellipsoids at 50%; Table 3, entry 5).
6298 www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 6296 –6299

Angewandte
Chemie
[14] A. F. Bedford, A. E. Beezer, C. T. Mortimer, H. D. Springall, J.
Chem. Soc. 1963, 3823.
[15] a) R. R. Schmidt, C. Beitzke, A. K. Forrest, J. Chem. Soc. Chem.
Commun. 1982, 909; b) C. Le Drian, E. Vieira, P. Vogel, Helv.
Chim. Acta 1989, 72, 338; c) C. Le Drian, P. Vogel, Helv. Chim.
Acta 1987, 70, 1703; d) Y. Auberson, P. Vogel, Helv. Chim. Acta
1989, 72, 278; e) C. Nativi, J. L. Reymond, P. Vogel, Helv. Chim.
Acta 1989, 72, 882; f) A. Warm, P. Vogel, J. Org. Chem. 1986, 51,
5348; g) T. Takahashi, A. Iyobe, Y. Arai, T. Koizumi, Synthesis
1989, 189; h) H. C. Brown, J. V. N. Vara Prasad, J. Org. Chem.
1985, 50, 3002; i) A. V. Rama Rao, J. S. Yadav, V. Vidyasagar, J.
Chem. Soc. Chem. Commun. 1985, 55.
[13] We considered the mechanistic possibility of pinacol-type
rearrangement sequences. However, the formation of the
aminal product inevitably positions the tertiary hydroxy group
at the other ring-fusion carbon atom. X-ray crystallographic
analysis of the p-bromobenzoate ester of 20b proved critical in
ruling out this option.
[16] A. Latvala, M. A. Oenur, T. Goezler, A. Linden, B. Kivcak, M.
Hesse, Phytochemistry 1995, 39, 1229.
Angew. Chem. Int. Ed. 2009, 48, 6296 –6299
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 6299