Fe(III)/Cu(II) mediated 5- and 6-exo oxidative ring expansion/cyclisation of cyclopropyl ethers: Studies towards dictyol C and α-eudesmol

Article abstract of DOI:10.1016/S0040-4039(00)00648-1

The outcome of oxidative ring expansion/cyclisation of cyclopropanes with a mixed Fe(III)/Cu(II) system was found to be dependent on the mode of cyclisation. When 6-exo cyclisation was attempted unusual products resulting from oxidation of ring opened primary radicals were obtained. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4039(00)00648-1

Tetrahedron Letters 41 (2000) 4651±4655
Fe(III)/Cu(II) mediated 5- and 6-exo oxidative ring
expansion/cyclisation of cyclopropyl ethers: studies towards
dictyol C and a-eudesmol
Kevin I. Booker-Milburn,a,* Brian Cox,b Mick Grady,a Frank Halleyc and
Shaun Marrisona
aSchool of Chemistry, University of Bristol, Cantock's Close, Bristol BS8 1TS, UK
bGlaxoWellcome, Medicines Research Centre, Gunnels Wood Road, Stevenage, Hertfordshire SG1 2NY, UK
cAventis Pharma Ltd., Rainham Road South, Dagenham, Essex RM10 7XS, UK
Received 9 March 2000; accepted 20 April 2000
Abstract
The outcome of oxidative ring expansion/cyclisation of cyclopropanes with a mixed Fe(III)/Cu(II) system
was found to be dependent on the mode of cyclisation. When 6-exo cyclisation was attempted unusual
products resulting from oxidation of ring opened primary radicals were obtained. # 2000 Elsevier Science
Ltd. All rights reserved.
For a number of years we have been studying the oxidative cyclisation of cyclopropanes with
Fe(III) salts and have developed this chemistry to provide a new method for the rapid synthesis of
[n.3.0] carbocycles1 (n=4,5,6) as well as highly functionalised cyclopentanes.2 Present studies
involve the evaluation of our tandem ring expansion/cyclisation methodology in the synthesis of
natural products. In this letter we discuss our preliminary results towards the natural products
dictyol C 13 and a-eudesmol 24 which possess antimicrobial and potent calcium channel blocking
activities, respectively. The schemes illustrate how our previous chemistry could be used to enable
rapid access to these systems (Scheme 1). An initial common strategy was to generate the exo-
cyclic products 5 and 8 by ring expansion of cyclopropanes 3 and 6, followed by 5-exo and 6-exo
radical cyclisation, respectively. Subsequent alkene isomerisation5 would then give the endocyclic
isomers with the correct geometry for the respective natural products.
Dictyol C: We had previously shown1 that the chemistry in Scheme 1 was suitable for the forma-
tion of 7,5-systems by ring expansion and 5-exo radical cyclisation onto both alkenes and alkynes.
Conjugate addition of both butenyl- and butynyl-magnesium bromides onto mono-protected
enone 9 gave the corresponding enol ethers, which, upon treatment with diethylzinc/diiodomethane,
* Corresponding author. Fax: +44(0)117 929 8611; e-mail: k.booker-milburn@bristol.ac.uk
0040-4039/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(00)00648-1

4652
Scheme 1.
gave cyclopropyl ethers 10 and 11 in good and low overall yields, respectively (4:1 ratio of diaster-
eomers). Oxidative cyclisation with Fe(NO3)3 using our previously optimised conditions gave
bicyclic-ketones 12 and 13 in moderate yields. It was pleasing to observe that the acid labile acetal
group remained intact upon exposure to the Lewis acidic conditions (Scheme 2).
Scheme 2. Reagents: (i) RMgBr, Me3SiCl, CuI, HMPA, THF; (ii) Et2Zn, CH2I2; (iii) Fe(NO3)3, 1,4-cyclohexadiene,
DMF
Unfortunately, the conversion of protected dione 9 to bicyclic ketone 13 was simply too inecient
to be used in a practical synthesis of dictyol C and an alternative method of introducing the
exocyclic alkene was sought. Our attention was drawn to the pioneering work of Kochi6 from the
early 1960s which involved the oxidation of primary radicals with Cu(II) salts to yield alkenes.
The synthetic potential of this chemistry has been realised by Snider7 in his elegant and compre-
hensive work with Mn(III)/Cu(II) mediated radical cyclisations. We were therefore pleased to
®nd that, in a model study, treatment of cyclopropane 14 with a Fe(NO3)3/Cu(OAc)2 mixture
(1.5:1) in DMF gave only the trans-fused exocyclic alkene 15, in good yield (Scheme 3). Present
studies are concerned with the total synthesis of dictyol C by the application of this strategy to
enone 9 and derivatives. The key bene®t of the Fe(III)/Cu(II) system is the reduction in the
number of steps in the synthesis of 5, as cumbersome manipulation of protected alkynyl side
chains will be unnecessary.

4653
Scheme 3.
ꢀ-Eudesmol: Previously, we and others8a,c have found that Fe(III)- and Mn(III)-mediated ring
expansion chemistry could be used to form six-membered rings by 6-exo cyclisation; although, in
the one transannular example we studied, problems arose through competitive 1,5-hydrogen
atom abstraction.8b Our initial approach was to consider the 6-exo-dig radical cyclisation as
outlined in Scheme 1. However, in light of the good results obtained with the dictyol C synthesis,
it became apparent that the mixed Fe(III)/Cu(II) oxidative cyclisation onto an alkene should
provide a more practical synthesis of exocyclic alkene 8. Conjugate addition of pentenylmagnesium
bromide to the cyclopentenones 16a/b followed by cyclopropanation gave ethers 17a/b in good
overall yield (17b as a 2:1 mixture of diastereoisomers). Oxidative ring expansion±cyclisation with
Fe(NO3)3 alone gave mixed results. With R=H the 6-exo product 18a predominated with trace
amounts of the uncyclised products 19a and 20a formed from oxidation and hydrogen atom
abstraction of the uncyclised radicals, respectively. Where R=Me oxidation gave an equal mixture
of 18b and 20b. In both cases, the cyclised products were obtained as a complex mixture of
diastereoisomers, with two major isomers predominating in a ratio of 3:2. The higher amount of
uncyclised material observed with R=Me clearly indicated that the methyl group has an unfavour-
able in¯uence on the rate of 6-exo cyclisation (Scheme 4).
Scheme 4.
When the same reactions were repeated with the Fe(III)/Cu(II) mix described above a rather
startling result was obtained. Only trace amounts of the expected exocyclic alkenes 22a/b were
formed (>19:1 stereoselectivity) and the major products in both cases were the a-methylene
cyclopentanones 21a/b (Scheme 5).
Scheme 5.

4654
The formation of the a-methylene cyclopentanone can be explained by assuming that initial
oxidation of the cyclopropane gives a rapidly equilibrating mixture (via the cyclopropyl alkoxy
radical) of the ring opened and ring expanded radicals 24a/b and 23a/b. Reaction of the primary
radicals 24a/b with Cu(II) then gives the Cu(III) species 25a/b which undergoes b-elimination to
the alkene according to the mechanism of Kochi. The enones 19a/b are either formed similarly or are
simply a result of SET oxidation of the secondary radical with Fe(III). In the whole history of this
work this is the ®rst time that we have observed products resulting from reactions of primary
radicals such as 24a/b. This unusual result gives several important insights into the mechanism of
these ring expansions. Most importantly, this result clearly shows that the ring opened primary
radicals 24a/b are true reaction intermediates in equilibrium with the ring expanded secondary
radicals. Obviously, the slower rate of 6-exo cyclisation allows the other observed pathways to
proceed within the reaction timescale. Treatment of the cyclopropanes 17a/b with Cu(OAc)2 alone
gave no reaction. This clearly indicates that Cu(II) is only involved once the radical species have
been formed, and no oxidative cleavage of the cyclopropane ring can occur until Fe(III) is present.9
Finally, treatment of cyclopropane 26 with the Fe(III)/Cu(II) mix gave the 6,5-bicyclic ketone
27 with the exocyclic double bond, along with the a-methylene cyclopentanone 28 (Scheme 6). As
expected, due to the faster rate of cyclisation, the 5-exo product was formed. Surprisingly
perhaps, was the formation of 28, which would suggest that 5-exo cyclisation of the six-membered
b-keto radical is slower than the analogous seven-membered case (14!15), thus allowing time for
the intermediate primary radical to react with the Cu(II).
Scheme 6.
In summary, these results provide the ®rst evidence of the existence of equilibrating primary
radicals in the oxidation of cyclopropylsilyl ethers with Fe(III). Further progress on the synthesis
of dictyol C and a-eudesmol will be reported in due course.
Acknowledgements
We would like to thank Aventis Pharma and GlaxoWellcome for studentships as well as the
Engineering and Physical Sciences Research Council (EPSRC) for the provision of a research
grant (GR/K15985) and a CASE quota to S.M.

4655
References
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. Enoki, N.; Tsuzuki, K.; Omura, S.; Ishida, R. Chem. Lett. 1983, 1627.
4. (a) Humber, D. C.; Pinder, A. R. Tetrahedron Lett. 1966, 41, 4985; (b) Frey, B.; Hunig, S.; Koch, M.; Reissig, H.
Synlett 1991, 854.
5. Rigby, J. H.; Wilson, J. Z. J. Am. Chem. Soc. 1984, 106, 8217.
6. (a) de la Mare, H. E.; Kochi, J. K.; Rust, F. F. J. Am. Chem. Soc. 1963, 85, 1437; (b) Kochi, J. K. Acc. Chem. Res.
1974, 7, 351.
7. Snider, B. B.; Vo, N. H.; Foxman, B. M. Chem. Rev. 1996, 96, 339.
8. For previous examples of 6-exo cyclisations of cyclopropane derived radicals, see: (a) Iwasawa, N.; Funahashi,
M.; Hayakawa, S.; Narasaka, K. Chem. Lett. 1993, 545; (b) Booker-Milburn, K. I.; Dainty, R. F. Tetrahedron
Lett. 1998, 39, 5097; (c) Blake, A. J.; Highton, A. J.; Majid, T. N.; Simpkins, N. J. Org. Lett. 1999, 1, 1787.
9. Ryu has demonstrated that cyclopropylsilyl ethers undergo cleavage to b-Cu(II) ketones using Cu(BF4)2. Evidence
was given in support of a non-radical mechanism in these cases: Ryu, I.; Matsumoto, K.; Kameyama, Y.; Ando,
M.; Kusumoto, N.; Ogawa, A.; Kambe, N.; Murai, S.; Sonoda, N. J. Am. Chem. Soc. 1993, 115, 12330.