A study of the epoxidation of cycloolefins by the t-buoh copper-permanganate system

Article abstract of DOI:10.1021/ol802923n

Evidence is presented that Cu(MnO₄)₂ effectively epoxidizes trisubstituted steroid olefins by a nonconcerted pathway.

Full text of DOI:10.1021/ol802923n

ORGANIC
LETTERS
2009
Vol. 11, No. 4
959-961
A Study of the Epoxidation of
Cycloolefins by the t-BuOH
Copper-Permanganate System
Younis Baqi, Simon Giroux, and E. J. Corey*
Department of Chemistry and Chemical Biology, HarVard UniVersity,
12 Oxford Street, Cambridge, Massachusetts 02138
corey@chemistry.harVard.edu
Received December 19, 2008
ABSTRACT
Evidence is presented that Cu(MnO4)2 effectively epoxidizes trisubstituted steroid olefins by a nonconcerted pathway.
It was reported several years ago that the reaction of various
steroidal olefins with a mixture of tert-butyl alcohol-CH2Cl2,
powdered KMnO4, and CuSO4 (wet) (Parish reagent) affords
good yields of epoxides invariably with the ꢀ-orientation of
the oxygen, even with substrates that yield predominately
R-epoxides with peroxy acids.1 We became curious about
this process for several reasons, including the possible use
in projected studies of steroids related to withaferin A (1).2
Compound 1 is of special interest because of its AyurVedic
history, antitumor activity, and potent inhibition of angio-
genesis (IC50 ) 12 nM) against human umbilical vein
endothelial cell sprouting.3,4 Specifically, 1 binds covalently
to the intermediate filament protein vimentin causing ag-
gregation in vivo so as to block angiogenesis.4 The action
could be synergistic with inhibition of the kinase vascular
endothelial growth factor (VEGF).
that epoxidation with m-chloroperbenzoic acid gave a modest
predominance of 5,6-R-epoxycholesteryl benzoate (ratio ꢀ/R
) 4:6). During the epoxidations with the Parish oxidizing
mixture there is considerable evolution of oxygen, and so
an excess of permanganate (ca. 8-9 equiv) must be
employed for complete conversion. Without copper(II), little
oxygen is evolved and little or no epoxide is formed. There
are at least two obvious explanations for the role of
copper(II): (1) t-BuOO• radical, which has been reported to
be formed by metal-ion oxidation of t-BuOOH and to effect
epoxidation of olefins, is somehow generated, or (2) the
effective epoxidation reagent is Cu(MnO4)2. This matter is
addressed below.
We found that the oxidation of the ∆5-olefin cholesteryl
benzoate with wet CuSO4/KMnO4 and 10 equiv of t-BuOH
in CH2Cl2 at reflux produced mainly the 5,6-ꢀ-epoxide 2
(ratio ꢀ/R ) 9:1), as reported.1 In comparison, we found
(1) Parish, E. J.; Li, H.; Li, S. Synth. Commun. 1995, 25, 927–940.
(2) Kupchan, S. M.; Doskotch, R. W.; Bollinger, P.; McPhail, A. T.;
Sim, G. A.; Renauld, J. A. S. J. Am. Chem. Soc. 1965, 87, 5805–5806.
(3) Mohan, R.; Hammers, H. J.; Bargagna-Mohan, P.; Zhan, X. H.;
Herbstritt, C. J.; Ruiz, A.; Zhang, L.; Hanson, A. D.; Conner, B. P.; Rougas,
Figure 1. Structures of withaferin A (1) and 2.
J.; Pribluda, V. S. Angiogenesis 2004, 7, 115–122
(4) Bargagna-Mohan, P.; Hamza, A.; Kim, Y.; Ho, Y. K.; Mor-Vaknin,
N.; Wendschlag, N.; Liu, J.; Evans, R. M.; Markovitz, D. M.; Zhan, C.-G.;
.
It has been reported that the reaction of ∆7-unstaturated
steroids with the Parish reagent leads selectively to the 7,8-
Kim, K. B.; Mohan, R. Chem. Biol. 2007, 14, 623–634
.
10.1021/ol802923n CCC: $40.75
Published on Web 01/30/2009
2009 American Chemical Society

ꢀ-epoxide.1 We have reevaluated this claim using 3ꢀ-
acetoxy-∆7-cholestene and the Parish reagent under the
recommended conditions1 since it seemed difficult to ratio-
nalize the ꢀ-selectivity with this substrate. We found, as we
had suspected, that the product was the 7,8-R-epoxide (3,
7:1 selectivity) and was identical to that obtained by
epoxidation of the ∆7-linkage by m-chloroperbenzoic acid.
The formation of the 5,6-ꢀ-epoxide from cholesteryl
benzoate (or acetate) using copper permanganate clearly
supports the possibility that this is the actual oxidant in the
Parish reagent. Further evidence is provided by additional
examples. First, we found that the epoxidation of ∆3,5
cholestadiene using the Parish reagent produced the corre-
sponding 5,6-ꢀ-epoxide (4) with a ꢀ/R selectivity of 9:1,
consistent with the ꢀ-preference for epoxidation of the ∆5-
linkage of cholesteryl benzoate. The structure of the epoxide
4 was confirmed by hydrogenation to the 6-ꢀ-alcohol which
was identified by acetylation and comparison with the
previously reported acetate 5.7 Using either Cu(MnO4)2 or
the Parish mixture, the epoxidation of stigmasteryl acetate
was also position selective for the ∆5-olefinic linkage and
afforded mainly the corresponding 5,6-ꢀ-epoxide (6) (6:1
5,6-ꢀ/5,6-R, 99% yield) (Scheme 1). The oxidation of 3-ꢀ-
1
The 7,8-R and 7,8-ꢀ-epoxides are readily identified by H
NMR analysis.6
Figure 2. 7,8-R-Epoxide 3.
Scheme 1
To follow up on these results, we prepared Cu(MnO4)2
by a new and convenient procedure from aqueous Cu(BF4)2
and KMnO4 in concentrated solution. Filtration of the
precipitate of KBF4 and removal of water under reduced
pressure gave a hydrate of Cu(MnO4)2 as a dark solid, soluble
not only in H2O, but also in mixtures of CH2Cl2, t-BuOH,
and HOAc. Although stable in aqueous or acetic acid
solution, the reagent decomposes fairly rapidly in CH3CN
or t-BuOH-CH2Cl2 solution. We found that a solution in
CH2Cl2-HOAc (98:2) decomposes at 23 °C in about 20 min
to yield approximately 1.2 equiv of O2 and a copious
precipitate of MnO2. Some Cu(MnO4)2 is entrained by the
precipitating MnO2 and thus removed from solution. The
instability of copper permanganate and its tendency to
coprecipitate with MnO2 explains the need to use a substan-
tial excess of permanganate in the oxidations described by
Parish et al.1 It is also possible that the manganese dioxide
just formed catalyzes the decomposition of Cu(MnO4)2.
acetoxy-∆7-cholestene with 1.5 equiv of Cu(MnO4)2 in 3:1
t-BuOH/CH2Cl2 at 23 °C for 1 h gave predominantly (4.4:
1) the 7,8-R-epoxide 3, as was the case when the Parish
reagent was employed.
It is probably not a coincidence in the several examples
outlined above that the Parish reagent and Cu(MnO4)2 led
to the same predominating diastereomeric epoxide. In
addition, there did not seem to be significant amounts of 1,2-
diol, despite the fact that vicinal dioxygenation is the
customary pathway for olefin oxygenation by sodium or
potassium permanganate in water-containing reaction mix-
tures.
The epoxidation reactions of Cu(MnO4)2 appear to differ
mechanistically from the much-studied epoxidation of olefins
with peroxycarboxylic acids, which clearly are concerted 1,2-
cycloaddition processes. One argument in support of this
surmise is the opposite stereochemical course observed with
The reaction of cholesteryl benzoate with 1.5 equiv of
Cu(MnO4)2 in 3:1 t-BuOH-CH2Cl2 solution at 23 °C for
1 h gave clean epoxidation of the double bond (>99% yield)
and mainly the ꢀ-epoxide (ꢀ/R ) 6:1). The reaction
proceeded much faster upon addition of HOAc, but the ꢀ/R
ratio was markedly diminished. In 2:1 HOAc-CH2Cl2 at
-20 °C, the reaction went to completion in 15 min and gave
a 1.5:1 mixture of ꢀ/R epoxides in >98% yield. It is possible
that the acceleration of the epoxidation by HOAc is due to
-
its H-bonding interaction with MnO4 .
(5) (a) Yu, J.-Q.; Corey, E. J. J. Am. Chem. Soc. 2003, 125, 3233–
3233. (b) Yu, J.-Q.; Corey, E. J. Org. Lett. 2002, 4, 2727–2729. (c) Bartlett,
P. D.; Gu¨nther, P. J. Am. Chem. Soc. 1966, 88, 3288–3294. (d) Barlett,
P. D.; Guaraldi, G. J. Am. Chem. Soc. 1967, 89, 4799–4801. For the
oxidative formation of peroxy radicals from hydroperoxides, see: (e) Haynes,
R. K.; Vonwiller, S. C. J. Chem. Soc., Chem. Commun. 1990, 1102–1104.
(f) Kharasch, M. S.; Pauson, P.; Nudenberg, W. J. Org. Chem. 1953, 18,
322–327. (g) Araneo, S.; Fontana, F.; Minisci, F.; Recupero, F.; Serri, A.
J. Chem. Soc., Chem. Commun. 1995, 1399–1400.
(6) Eguchi, S.; Yamaguchi, S.; Furuya, M.; Morisaki, M. Chem. Pharm.
Bull. 1988, 36, 2813–2818.
(7) Jones, D. N.; Shoppee, C. W.; Summers, G. H. R. J. Chem. Soc.
1955, 2876–2887.
960
Org. Lett., Vol. 11, No. 4, 2009

the ∆5-steroid substrates: strong 5,6-ꢀ preference for
Cu(MnO4)2 vs 5,6-R preference for peroxy acids. An
important difference was shown with the substrate cis-
stilbene. Although this olefin affords the cis-epoxide with
peroxyacids, a 1:1 mixture of cis- and trans-epoxides was
observed with Cu(MnO4)2. This result argues that Cu(MnO4)2
behaves more like an oxy radical reagent than a peroxy acid
and that it adds to olefins to form a carbon radical that
cyclizes to epoxide in a distinct, later step. Such a pathway
is depicted in Scheme 2 for the reaction of cis-stilbene with
Figure 3. Intermediate 7 stabilized by hyperconjugation.
-
stabilized by hyperconjugation of MnO4 (stereoelectroni-
cally preferred axial attack).8 The same argument can account
for 7,8-R-epoxidation of ∆7-sterol derivatives. The epoxi-
dation by Cu(MnO4)2 that is described herein is either very
slow or negligible with 1,2-disubstituted or monosubstituted
olefins, a result that is consistent with the mechanistic
pathway outlined in Scheme 2 and that also indicates a
modest level of reactivity for this form of permanganate. In
contrast, quaternary ammonium permanganate reagents rap-
idly react with most olefins by a vicinal dioxygenation
pathway which can involve concomitant C-C cleavage.9
In conclusion, the data in this paper support the view that:
(1) epoxidations with the Parish reagent probably involve
Cu(MnO)4 as the reactive species; (2) the epoxidation of ∆7-
unsaturated sterol systems leads predominately to 7-R-
epoxidation rather than 7-ꢀ-epoxidation as claimed in the
literature;1 (3) the pathway of epoxidation by Cu(MnO4)2
occurs in two or possibly three steps and is nonconcerted;
(4) the stereochemical preferences for the oxidation of
unsaturated steroids can be explained by the operation of
stereoelectronic effects (axial attack by O) and a noncon-
certed pathway.
Scheme 2. Oxidation of cis-Stilbene by Cu(MnO4)2
Cu(MnO4)2. Another possibility is that Cu(MnO4)2 attacks
the olefin electrophilically so that a cationic intermediate such
as B (Scheme 2) is formed directly.
As depicted, it is also possible that Cu(II) plays a role in
the epoxidation by converting the intermediate benzylic
radical A to a cation B, which might facilitate ring closure
to form the epoxide. Obviously, since rapid rotation is
possible around the central single bond, the observed mixture
of cis- and trans-epoxides would be expected for the
oxidation mechanism shown in Scheme 2. Another reason
why the intermediate A might be favored with Cu(MnO4)2
is the possibility that the negative charges spread out over
Acknowledgment. Y.B. is indebted to Deutscher Akade-
mischer Austausch Dienst (DAAD) and Deutsche Fors-
chungsgemeinschaft (DFG, GRK804) for fellowship support;
S.G. is grateful to NSERC (Canada) for a postdoctoral
fellowship.
Supporting Information Available: Experimental pro-
cedures and characterization data for all new compounds.
This material is available free of charge via the Internet at
http://pubs.acs.org.
-
the three oxygens of the OMnO3 moiety of A might be
stabilized by interaction with associated Cu(II).
The two-step pathway also provides a simple explanation
for the stereochemical preference in Cu(MnO4)2 epoxidation.
For ∆5-steroid epoxidation, a radical-like Cu(MnO4)2 would
be expected to attach to C(6) by an axial bond to form the
intermediate 7 (radical or cationic form) which can be
OL802923N
(8) (a) Corey, E. J. Experientia 1953, 9, 329–331. (b) Corey, E. J.; Sneen,
R. A. J. Am. Chem. Soc. 1956, 78, 6269–6278.
(9) Sala, T.; Sargent, M. V. Chem. Commun. 1978, 253–254.
Org. Lett., Vol. 11, No. 4, 2009
961