New Fe(III)-mediated radical cascade reactions of cyclopropyl silyl ethers

Article abstract of DOI:10.1021/ol0341703

(Matrix presented) Fe(III)-mediated ring opening of cyclopropyl ethers bearing a phenyl-substituted butenyl side chain leads to the generation of β-keto radicals that undergo 5-exo cyclization followed by a novel cascade sequence resulting in the formatio

Full text of DOI:10.1021/ol0341703

ORGANIC
LETTERS
2003
Vol. 5, No. 7
1107-1109
New Fe(III)-Mediated Radical Cascade
Reactions of Cyclopropyl Silyl Ethers
,†
Kevin I. Booker-Milburn,* J. Leighton Jones,† Graham E. M. Sibley,†
Richard Cox,‡ and Jim Meadows‡
School of Chemistry, UniVersity of Bristol, Cantock’s Close, Bristol, BS8 1TS,
United Kingdom, and Chemical DeVelopment, Medicines Research Centre,
GlaxoSmithKline, Gunnels Wood Lane, SteVenage, Herts., SG1 2NY, United Kingdom
k.booker-milburn@bristol.ac.uk
Received January 30, 2003
ABSTRACT
Fe(III)-mediated ring opening of cyclopropyl ethers bearing a phenyl-substituted butenyl side chain leads to the generation of â-keto radicals
that undergo 5-exo cyclization followed by a novel cascade sequence resulting in the formation of tricyclic ethers.
Previously we have reported that both cyclopropylsilyl ethers1
(e.g., 1) and cyclopropanone ketene-acetals2 undergo Fe(III)-
mediated tandem ring opening/cyclization to yield mono- and
bicyclic products.3 The reaction is thought to proceed by
initial oxidative ring cleavage to give â-carbonyl radicals
(e.g., 2) that then undergo 5-exo cyclization to the products.
In the case of cyclopropylsilyl ethers, we had previously
discounted cyclization studies of simple variants such as 4
because of problems encountered with the cyclopropanation
of acyclic enol-ethers 3. However, recent observations in our
laboratory with cyclopropanation led us to reinvestigate
acylic-enol ethers, and herein we describe the unexpected
reactions of these cyclopropanes on treatment with Fe(III)
salts (Scheme 1).
best conditions for the selective cyclopropanation of silyl
enol-ethers in the presence of alkenes were those described
by Furukawa (Et2Zn, CH2I2, Et2O).4 Unfortunately, these
conditions were unsuccessful for related acyclic-enol ethers
we investigated in the past (very slow, incomplete reactions).
Use of the more reactive conditions described by Denmark5
resulted in biscyclopropanation. Recently, we found that the
use of Et2Zn/CH2I2 in toluene6 dramatically accelerated the
rate of cyclopropanation of cyclic enol-ethers. Following on
from these results, we found that 5 could be cyclopropanated
in 88% yield by using only the hexane present in Et2Zn (1
Scheme 1. Fe(III)-Mediated Cyclopropane Ring-Opening
Treatment of MVK with 3-butenylmagnesium chloride
under our previously developed conditions gave the enol
ether 5 in 86% yield. Previously, we1 had found that the
Radical Cylization Reactions
† University of Bristol.
‡ GlaxoSmithKline.
(1) Booker-Milburn, K. I.; Thompson, D. F. J. Chem. Soc., Perkin Trans.
1 1995, 2315.
(2) Booker-Milburn, K. I.; Barker, A.; Brailsford, W.; Cox, B.; Mansley,
T. E. Tetrahedron 1998, 54, 15321.
(3) Independent studies by Narasaka and co-workers showed that Mn-
(III) could be used to effect similar transformations with cyclopropanols:
Iwasawa, N.; Funahashi, M.; Hayakawa, S.; Ikeno, T.; Narasaka, K. Bull.
Chem. Soc. Jpn. 1999, 72, 85.
10.1021/ol0341703 CCC: $25.00 © 2003 American Chemical Society
Published on Web 03/06/2003

M, Aldrich) as a solvent. The reaction was complete within
20 min at 0 °C and for the first time allowed access to
multigram quantities of 6. Fe(III)-mediated ring opening/
radical cyclization of 6 proceeded well using three different
radical traps2 and produced the cyclopentanes 7-9 as 7:1
mixtures of cis:trans isomers (Scheme 2).
Scheme 3 a
Scheme 2 a
aReagents and conditions: (a) 3-butenylmagnesium bromide,
TMSCl, HMPA, THF, -78 °C, 84%; (b) Et2Zn, CH2I2, 0 °C, 91%;
(c) Fe(NO3)3, N-chlorosuccinimide, DMF, 50% 12; (d) Fe(NO3)3,
PhSSPh, DMF, 58% 13 (9:1 with minor diastereoisomer); (e)
Fe(NO3)3, 1,4-cyclohexadiene, DMF, 60% 14b (9:1 with minor
diastereoisomer) and 31% 15b.
suggesting that termination to 14b was caused by hydrogen
atom abstraction from DMF. Repeating the reaction in
dimethylacetamide (DMA) improved matters somewhat, and
the ratio of 14b:15b was almost 1:1. These optimized
conditions were then used for the other cyclopropanes
described in Table 1. Cyclization of 11a (R ) H) gave only
a Reagents and conditions: (a) 3-butenylmagnesium bromide,
TMSCl, HMPA, THF, -78 °C, 86%; (b) Et2Zn, CH2I2, 0 °C, 88%;
(c) Fe(NO3)3, 1,4-cyclohexadiene, DMF, 2.5 h, 57% (7a:7b ) 7:1);
(d) Fe(NO3)3, N-chlorosuccinimide, DMF, 65% (8a:8b ) 7:1); (e)
Fe(NO3)3, PhSSPh, DMF, 60% (9a:9b ) 7:1).
We then sought to explore the stereochemical influence
of a substituent in the cyclization as we had done previously
for cyclopropanone acetals.2 The phenyl-substituted enol-
ether 10 was synthesized and cyclopropanated as before to
yield the cyclization substrate 11b (R ) Me) in high overall
yield. Fe(III)-mediated cyclization using either NCS or
PhSSPh as radical traps furnished the cyclized products 12
(sole product) and 13 in reasonable yield, with minor
amounts (9:1) of another diastereomer of 13 obtained. The
relative stereochemistry of 13 was obtained by X-ray
crystallography of the corresponding 2,4-dinitrophenylhy-
drazone.7 To our surprise, however, when 11b was treated
with Fe(NO3)3 and 1,4-cyclohexadiene as a H-atom donor,
the expected cyclization product 14b was formed along with
a very unusual tricyclic ether product 15b (Scheme 3).
We then sought to probe the scope of this likely cascade-
type sequence by synthesizing a range of substituted cyclo-
propylsilyl ethers bearing a phenyl group in the side chain.
All the cyclopropanes 11a-e were synthesized from known
enones by the conjugate addition-cyclopropanation sequence
used above. Initially, we sought to increase the proportion
of 15b by slowing down the rate of hydrogen atom
abstraction from radical intermediates leading to 14b.
Surprisingly, carrying out the reaction without 1,4-cyclo-
hexadiene made little difference in the ratios of 14b and 15b,
Table 1. a
11
R
15 (%)
14 (%)
a
b
c
d
e
H
Me
Et
nBu
iPr
56
45
40
37
33
50
34
29
23
a Other minor diastereomers were formed in ratios of 9:1, 5:1, and 3:1
for compounds 11b, 11c, and 11d,e, respectively. The stereochemistry not
assigned.
the tricyclic ether 15a. Interestingly, the relative size of the
R group was found to have a significant effect on both the
yield and the stereochemistry of the products. As the size of
R increased, the yield of 15 decreased. This was matched
by a decrease in the observed stereoselectivity of the
monocyclization products 14c-e. Only the major isomers
of 14b-e have been depicted, as it was not possible to assign
the relative stereochemistry of the minor products. Attempts
to synthesize phenyl derivatives of 11 (R ) Ph and
pMeOC6H4) were unsuccessful as the corresponding enol-
ethers failed to undergo cyclopropanation.
(4) Furukawa, J.; Kawabata, N.; Nishimura, J. Tetrahedron 1968, 24,
53.
(5) Denmark, S. E.; Edwards, J. P. J. Org. Chem. 1991, 56, 6974.
(6) Jenkins, H. Unpublished results.
(7) See Supporting Information.
1108
Org. Lett., Vol. 5, No. 7, 2003

believe that this radical then undergoes oxidation to the cation
20 by electron transfer to the ferric nitrate [Fe(III)fFe(II)].
Finally, this cation undergoes ionic ring closure to form the
tricyclic ether product. It would appear that the larger R
groups affect the cis selectivity of the cyclization of 16 to
17, and presumably therefore a diastereomer of 17 is
generated that cannot undergo cyclization to 18. This would
account for the lower stereoselectivity observed with the
monocyclized products 14c-e.
Scheme 4. Mechanism of Tricyclic Ether 15 Formation
In summary, access to previously unobtainable cyclopro-
pylsilyl ethers has uncovered a novel Fe(III)-mediated
oxidative radical cascade process resulting in the formation
of complex tricyclic ethers.
Acknowledgment. We thank the EPSRC (Postdoctoral
Grant GR/M83810) and GlaxoSmithKline (CASE award to
J.L.J.) for generous funding of this work.
The formation of 15 is the result of a previously unreported
radical cascade process, the mechanism of which is proposed
in Scheme 4. It is likely that on oxidation of 11 with Fe-
(III), the â-keto radical 16 is formed and undergoes 5-exo
cyclization to the cyclopentylmethyl radical 17. This then
undergoes two different termination pathways. First, hydro-
gen atom abstraction from the solvent leads to the observed
cyclopentanes 14a-e. Obviously, that process is relatively
slow and the radical is able to undergo a second 5-exo
cyclization onto the newly formed ketone carbonyl, resulting
in the alkoxyl radical 18.8 If the stereochemistry of 18 is as
shown, it would then undergo a facile 1,5 H-atom abstraction
leading to the formation of the benzylic radical 19. We
Supporting Information Available: Experimental pro-
cedures and characterization for all new compounds. This
material is available free of charge via the Internet at
http://pubs.acs.org.
OL0341703
(8) (a) Beckwith, A. L. J.; Hay, B. P. J. Am. Chem. Soc. 1989, 111,
2674. (b) For an annulation sequence involving abstraction/cyclization
reactions of cyclopropylamines, see: Ha, J. D.; Lee, J.; Blackstock, S. C.;
Cha, J. K. J. Org. Chem. 1998, 63, 8510.
Org. Lett., Vol. 5, No. 7, 2003
1109