Samarium diiodide mediated cascade radical cyclisations - Synthesis of the eudesmane tricyclic framework

Article abstract of DOI:10.1016/S0040-4039(00)01859-1

Radical cascade cyclisation of methylenecyclopropyl cyclohexanone adducts, using samarium diiodide to generate an initial ketyl radical, provide a short route to tricyclic ethers, and the stereochemical outcome can be influenced by the solvent used for the reaction. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4039(00)01859-1

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 41 (2000) 10341–10345
Samarium diiodide mediated cascade radical
cyclisations—synthesis of the eudesmane tricyclic framework
Faye C. Watson and Jeremy D. Kilburn*
Department of Chemistry, University of Southampton, Southampton SO17 1BJ, UK
Received 22 September 2000; accepted 19 October 2000
Abstract
Radical cascade cyclisation of methylenecyclopropyl cyclohexanone adducts, using samarium diiodide
to generate an initial ketyl radical, provide a short route to tricyclic ethers, and the stereochemical
outcome can be influenced by the solvent used for the reaction. © 2000 Elsevier Science Ltd. All rights
reserved.
Keywords: methylenecyclopropane; samarium diiodide; cascade radical cyclisation.
We recently described the highly stereoselective cascade cyclisation of methyelenecyclopropyl
ketone 1, mediated by samarium diiodide, which led to bicyclic ether 2, which in turn was used
1
as an intermediate in the synthesis of (±)-paeonilactone B. We wished to extend this methodology
to the synthesis of more complex natural products, and in particular, to tricyclic compounds such
as the eudesmane sesquiterpenoid lactones, which occur with a range of substitution patterns,
2
stereochemistries and degrees of saturation. In analogy to the preparation of 2 from 1, we
envisaged that the eudesmane skeleton could be accessed via a tricyclic intermediate 4, which, in
turn, could be prepared by cyclisation of a methylenecyclopropyl cyclohexanone adduct 3
(
Scheme 1). In this paper we describe our preliminary studies on cyclisations of such adducts.
Scheme 1.
*
Corresponding author. E-mail: jdk1@soton.ac.uk
0
040-4039/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(00)01859-1

10342
Radical cyclisation precursors were prepared by addition of lithiated methylenecyclopropane
to aldehyde 5 to give the two, diastereomeric alcohols 6 and 7 in a 2:1 ratio and in 82% yield.
After separation by chromatography, both alcohols could be alkylated to give the corresponding
allyl and propargyl ethers, although the alkylation with propargyl bromide required more
forcing conditions, reflecting the hindered nature of the alcohols 6 and 7. Ketal deprotection,
using TsOH/acetone/H O in each case, gave the corresponding ketones 8–13 (Scheme 2).
2
Scheme 2.
Initially we investigated the SmI mediated cyclisation of the free alcohols 8 and 11, which we
2
anticipated should cyclise with the stereochemistry controlled by coordination of the alcohol to
3
the samarium bound to the ketyl radical, in analogy to Matsuda’s results. Indeed slow addition
t
4
of ketone 8 to SmI , in the presence of BuOH (2 equiv.) and HMPA (10 equiv.) in THF at 0°C,
2
gave the expected cis diol 16, but in only 9% yield, and accompanied by the dimer 14 and
5
elimination product 15 (Scheme 3).
Scheme 3.
The formation of 15 can be readily understood since the reaction mechanism presumably
proceeds by cyclisation of the ketyl radical onto the methylenecyclopropane and opening of the
cyclopropylmethyl radical 17, to give 18. Reduction of 18 to the corresponding anion 19 is
t
−
followed by quenching with a proton from BuOH to give diol 16, or by elimination of HO ,

10343
assisted by chelation to Sm(III), which holds the alcohol in a favourable axial orientation (see
0). Dimerisation to give 14, while initially surprising, given the high concentration of SmI in
2
2
6
the reaction, is precedented by related cyclisations we have observed. The outcome of the
reaction, using HMPA/ BuOH, was largely unaffected by altering the temperature, and in the
absence of HMPA/ BuOH, using neat THF as solvent, treatment of 8 with SmI yielded diol 21
t
t
2
as the major product. However, using a MeOH/THF (1:4) mixture as solvent in place of
t
7
HMPA/ BuOH, as described by Procter, at −78°C, gave a much cleaner reaction and diol 16
was isolated in 62% yield along with a 35% yield of the eliminated product 15 (Scheme 4).
Scheme 4.
Cyclisation of alcohol 11 gave similar results, giving bicyclic diol 16 in 50% yield along with
1
5 in 12% yield, using SmI in MeOH/THF (1:4) at −78°C, but giving a poor reaction when
2
t
carried out using HMPA/ BuOH.
t
Cyclisation of propargyl ether 9 using SmI with BuOH/HMPA in THF (at −78°C) gave the
2
tricyclic ether 25 in 50% yield, and as a single diastereoisomer (Scheme 5). An improved yield
of 60% was obtained using the MeOH/THF solvent system.
Scheme 5.
Although chelation of the propargyl ether to the ketyl-bound samarium is not as effective as
with the free alcohol 8, the preferred conformation of the ketyl radical derived from 9 can still
be expected to place substituents on the starting cyclohexyl ring in an equatorial orientation as
in 22. After addition to the methylenecyclopropane and ring-opening, radical 23 cyclises on to
the propargyl unit to give a cis-fused cyclic ether.
Cyclisation of 12, however, under either set of conditions gave instead the bicyclic ether 28,
indicating that initial cyclisation of the ketyl radical onto the methylenecyclopropane is
disfavoured and a 6-exo cyclisation onto the alkyne is preferred instead (Scheme 6). Presumably,
while cyclisation of 9, via intermediate 22, places the cyclopropyl alkene bond in a satisfactory
environment, gauche to the ketyl CꢀO bond, the cyclisation of 12, via intermediate 26, would
require the cyclopropyl alkene bond to be eclipsed with the ketyl CꢀO bond, which would be
8
unfavourable as a result of electronic repulsion (Scheme 6). Cyclisation instead proceeds via
intermediate 27, although this does require breaking the (weak) chelate to the propargyl ether.

10344
Scheme 6.
Finally, cyclisation of the allyl ethers 10 and 13 introduces an additional chiral centre (Scheme
). Cyclisation of 13 gave the bicyclic ether 29, following a 6-exo cyclisation onto the allyl ether
7
in preference to the methylenecyclopropane (cf. 12). However, cyclisation of 10 using HMPA/
THF gave predominantly 31 (30:31=1:10), whereas using MeOH/THF gave predominantly 30
9
(
30:31=3.5:1), and in very good yield. This unexpected reversal of stereoselectivity can be
1
0
11
understood with reference to Beckwith’s and RajanBabu’s detailed studies of the cyclisation
of cyclohexyl radicals onto pendant butenyl (or allyloxy) groups. Assuming that 10 cyclises via
a weakly chelated intermediate 32, then the resulting cyclohexyl radical intermediate 33 will have
the allyloxy substituent in an axial position, which should, according to the Beckwith/Rajan-
Babu model, cyclise to give an anti-methyl substituent as in 30. Flipping of conformation 33
(
which involves breaking the chelation from the allyl ether oxygen to the samarium metal, and
which may be favoured in the presence of HMPA) gives 34 with the allyloxy substituent in an
equatorial position, which should cyclise to give a syn-methyl substituent as in 31.
Scheme 7.
Thus samarium diiodide mediated cyclisation of methylenecyclopropyl cyclohexanone
adducts, under suitable solvent conditions, provides a stereoselective route to tricyclic ethers.
Extension of this methodology to the synthesis of eudesmane natural products is ongoing in our
research group.

10345
References
1
2
3
4
5
. Boffey, R. J.; Santagostino, M.; Whittingham, W.; Kilburn J. D. Chem. Commun. 1998, 1875.
. Banerjee, A. K.; Vera, W. J.; Gonzalez, N. C. Tetrahedron 1993, 49, 4761.
. Kito, M.; Sakai, T.; Yamada, K.; Matsuda, F.; Shirahama, H. Synlett 1993, 158.
. Molander, G. A.; McKie, J. A. J. Org. Chem. 1995, 60, 872.
. All new compounds were characterised by IR, MS, H and C NMR, with H– H and H– C correlation
spectra, where necessary, to aid the assignments. The stereochemistry of all compounds was established
unambiguously by consideration of coupling constants, detailed NOE studies or by X-ray crystallographic
analysis. Full details will be reported in due course.
1
13
1
1
1
13
6. Boffey, R. J.; Whittingham, W.; Kilburn, J. D. Tetrahedron Lett. 1999, 40, 5625.
7. Johnston, D.; McCusker, C. M.; Procter, D. J. Tetrahedron Lett. 1999, 40, 4913. Note that while a few
equivalents of MeOH have been routinely used as a proton donor source in SmI mediated cyclisations, here it
2
is used in large excess as a co-solvent for the reaction.
8
. Electronic repulsion between the ketyl oxygen functionality and the alkene p-system is generally accepted as being
a major factor determining stereoselectivity in SmI mediated cyclisations of unactivated olefinic ketones: see Ref.
2
4
and Beckwith, A. L. J. Tetrahedron 1981, 37, 3073.
9
. For other examples where the stereochemical outcome of a SmI -mediated cyclisation is altered by choice of
2
solvent or additives, see Refs. 1 and 3 and: (a) Fevig, T. L.; Elliott, R. L.; Curran, D. P. J. Am. Chem. Soc. 1988,
1
10, 5064; (b) Molander, G. A.; McKie, J. A. J. Org. Chem. 1992, 57, 3132; Johnston, D.; McCusker, C. M.;
Muir, K.; Procter, D. J. J. Chem. Soc., Perkin Trans. 1 2000, 681.
1
1
0. Beckwith, A. L. J.; Page, D. M. Tetrahedron 1999, 55, 3246, and references cited therein.
1. RajanBabu, T. V. Acc. Chem. Res. 1991, 24, 139, and references cited therein.
.