Electrochemical aerobic oxidation of aminocyclopropanes to endoperoxides

Article abstract of DOI:10.1002/anie.200702903

(Chemical Equation Presented) Electrical synthesizer: Results of cyclic voltammetric studies on a range of aminocyclopropanes in the absence or in the presence of oxygen are in full agreement with the currently accepted autocatalytic mechanism for the aer

Full text of DOI:10.1002/anie.200702903

Communications
DOI: 10.1002/anie.200702903
Electrochemical Oxidation
Electrochemical Aerobic Oxidation of Aminocyclopropanes to
Endoperoxides**
Claire Madelaine, Yvan Six,* and Olivier Buriez*
Dedicated to Professor William B. Motherwell on the occasion of his 60th birthday
The oxidation of cyclopropanes to 1,2-dioxolanes has been
documented for almost 40 years and has been performed on
many substrates under various conditions.[1–3] Among these,
very few examples involve aminocyclopropanes. The trans-
formation of secondary cyclopropylamines into a-amino
endoperoxides by autocatalytic aerobic oxidation was
reported by Wimalasena et al. a few years ago.[4] It was
found that a catalytic amount of tris(1,10-phenantroline)iron-
(III) hexafluorophosphate, or hydrogen-abstracting agents
such as benzoyl peroxide or tert-butyl peroxide under UV
irradiation, could increase the reaction rate. A similar trans-
formation was observed in our laboratory, spontaneously
occurring in the presence of air and silica gel from electron-
rich tertiary arylcyclopropylamines.[5] In this case, as well as in
the iron-catalyzed process, the mechanism probably involves
a one-electron oxidation of the substrate to a cation radical
intermediate. After cyclopropane ring opening, reaction with
oxygen, and ring closure, the newly produced cation radical
would be able to function as an oxidant itself and thus close
the catalytic cycle (Scheme 1).[4,5]
Although interesting, this reaction suffers from serious
limitations: a) the transformation appears to be facile only in
the special cases where the nitrogen atom bears a hydrogen
atom or an electron-rich aromatic ring; b) the use of
peroxides under UV irradiation is by nature limited to
secondary amines because it involves hydrogen abstraction
from the nitrogen atom; c) the iron(III) catalyst mentioned
above is not commercially available; and d) occasionally
experiencing problems in isolating the expected peroxides, we
were led to suspect that some of them might be unstable.[6] We
anticipated that electrochemical methods such as cyclic
Scheme 1. Autocatalytic aerobic oxidation of aminocyclopropanes to a-
amino endoperoxides.
voltammetry and preparative electrolyses would be tools of
choice to gain better insight in this reaction and could lead to
the design of a more general and environmentally friendly
experimental procedure.[7]
For the purpose of this study, aminocyclopropanes 1a–e,
2a,b, 3, and 4 were prepared by using intra- or intermolecular
Kulinkovich–de Meijere reactions,[8] following a modification
Table 1: Synthesis of aminocyclopropanes 1a–e, 2a,b, and 3 through
intramolecular Kulinkovich–de Meijere reactions.[a]
[*] C. Madelaine, Dr. Y. Six
Institut de Chimie des Substances Naturelles
UPR 2301 du CNRS
Av. de la Terrasse, 91198 Gif-sur-Yvette Cedex (France)
Fax: (+33)1-6907-7247
R1
R2
R3
n
1
Yield [%]
4-nBuOC6H4
4-nBuOC6H4
4-nBuOC6H4
Ph
Bn
Ph
Ph
H
H
H
H
H
1
1
2
1
1
1
1
1a[b]
1b
1c
1d
1e
42
100
44
90
75
Me
Me
Me
Me
Me
Me
E-mail: yvan.six@icsn.cnrs-gif.fr
Dr. O. Buriez
Ecole Normale SupØrieure, DØpartement de Chimie
UMR CNRS-ENS-UPMC 8640 “Pasteur”
24 rue Lhomond, 75231 Paris Cedex 05 (France)
Fax: (+33)1-4432-3863
H
(CH2)2OTBS
(CH2)2OBn
2a
2b
38
53
E-mail: olivier.buriez@ens.fr
[a] Bn=benzyl, TBS=tert-butyldimethylsilyl. [b] By-product 3 isolated in
32% yield as the same time as compound 1a.
[**] We warmly thank Dr. Sophie Susplugas and Prof. Philippe Grellier
from the MusØum National d’Histoire Naturelle (Paris, France) for
the evaluation of the in vitro antimalarial activity of our compounds.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
8046
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 8046 –8049

Angewandte
Chemie
of the procedure reported by Cha and Lee[9] (Table 1,
Scheme 2).
The study of their electrochemical behaviour was carried
out by cyclic voltammetry, and a typical voltammogram is
pare 1b vs 1d and 1d vs 2a,b), while the N-benzyl derivative
1e is more difficult to oxidize. In all cases, the process
generates the corresponding aminium cation radicals, which
are unstable at low scan rates in agreement with the
irreversibility of Ox1. In some cases (e.g. 1a and 3), however,
the reduction of these intermediates can be observed if the
potential range is swept sufficiently rapidly (v > 20 VsÀ1), and
the cyclic voltammograms become reversible.
When the experiments are carried out under aerobic
conditions, two main changes are observed in the voltammo-
grams: 1) the intensity of Ox1 decreases, and 2) a new
oxidation wave Ox2 appears at a higher potential value on
the same timescale. The first point is characteristic of an
electron catalytic process in which the starting compound is
oxidized by an intermediate species but not by the electrode
itself,[10] which supports the mechanism displayed in
Scheme 1. The second point clearly indicates the formation
of a new compound that can be assigned to the corresponding
endoperoxide as verified with an authentic sample (see
below). In the cases of 2a and 2b, a third oxidation wave
Ox3 appears to emerge between Ox1 and Ox2, probably owing
to the presence of the OTBS and OBn groups. A structure–
reactivity relationship is observed for the endoperoxides:
although aminocyclopropanes 1b and 1c are oxidized at
similar potential values, it is not the case for the correspond-
ing peroxides 5b and 5c, the latter of which features a six-
membered ring that is oxidized at a higher potential (0.89 vs
0.78 V). In the special case of 1d, the wave Ox2 is only
observed on the timescale of a preparative electrolysis (see
below), indicating a very slow formation of the endoperoxide
product. Importantly, the higher potential of Ox2 as compared
to Ox1 in all cases fully supports the hypothesis of an
autocatalytic cycle where the cation radical of the product can
oxidize the cyclopropylamine starting material (Scheme 1).
The results obtained on the cyclic voltammetry timescale
led us to attempt the electrosynthesis of the endoperoxides.
Preparative-scale electrolyses were carried out in a divided
cell at constant potential values corresponding roughly to the
peak potential of Ox1. The aminium cation radicals could thus
be generated cleanly and selectively to react with oxygen
which was introduced by bubbling air through the anodic
compartment. The electrolyses were followed by in situ cyclic
voltammetry to ascertain the consumption of the amino-
cyclopropane along with the formation of the product, and to
verify its stability on this timescale. Figure 2 shows typical
voltammograms obtained before and after the experiment.
Interestingly, a very ill-defined wave Ox2 was observed in the
cyclic voltammogram recorded before running the electrol-
ysis. Under these conditions, the amount of dissolved oxygen
is small compared to the amount of aminocyclopropane
initially introduced in the cell. However, and as expected, the
intensity ratio I(Ox2)/I(Ox1) increased as the electrolysis
proceeded.
Scheme 2. Synthesis of aminocyclopropane 4 through an intermolecu-
lar Kulinkovich–de Meijere reaction.
presented in Figure 1. Under argon, the aminocyclopropanes
are oxidized in a single irreversible wave Ox1 located between
Figure 1. Cyclic voltammograms of 1b (2.1 mm) in MeCN with TBABF4
(0.1m) recorded at a platinum electrode (0.5 mm diameter) at
0.2 VsÀ1 in the absence (Ox2) and in the presence of air (Ox1).
+ 0.5 and + 1.1 V (versus the saturated calomel electrode
(SCE)) depending on their chemical structures (Table 2). As
expected, the presence of an electron-donating group either
at the para position of the aromatic ring or on the cyclo-
propane moiety renders the oxidation process easier (com-
Table 2: Oxidation wave peak potentials of cyclopropylamines 1a–e,
2a,b, 3, and 4.
Cyclopropylamine
E(Ox1) [V][a]
E(Ox2) [V][a]
E(Ox3) [V][a]
1a
1b
1c
1d
1e
2a
2b
3
+0.54[b]
+0.54
+0.52
+0.79
+1.10[b]
+0.67
+0.70
+0.55
+0.63[b]
+0.50
+0.88[b]
+0.78
+0.89
+1.17
–
+1.08
+1.10
+0.80
+0.84[b]
+0.82
–
–
–
–
–
+0.82
+0.84
The experiments were stopped after the quasi-total
consumption of the aminocyclopropanes. In all cases (1b–d,
2a, 3, and 4), the last voltammograms showed the formation
of the corresponding endoperoxides (wave Ox2) and demon-
strated their relative stability on the timescale of the reactions
(about 1 h depending on the amount of starting material and
–
–
–
3
4
[a] Oxidation potential versus SCE, measured by cyclic voltammetry at
0.2 VsÀ1 in MeCN as the solvent unless otherwise stated. [b] Measured
in N,N-dimethylformamide.
Angew. Chem. Int. Ed. 2007, 46, 8046 –8049ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 8047

Communications
radical is secondary rather than primary. Besides this, the
complete regioselectivity of the cyclopropane ring-opening of
1 and 2, leading exclusively to [3.3.0]rather than [3.2.1]
systems, is noteworthy. This can be accounted for by the
geometrical alignment of the bond being broken with the
orbital of the aminyl cation radical.
Finally, the analogy of the stable endoperoxide 6a with
some antimalarial drugs led us to examine its antiplasmodial
activity.[12] Indeed, the endoperoxide moiety of antimalarial
trioxanes such as artemisinin has been shown to be directly
involved in the mechanism leading to the destruction of the
parasite.[13] The two diastereoisomers of 6a were thus
separated by preparative HPLC, and their in vitro antima-
larial activities were evaluated against the chloroquine-
resistant FcB1 strain of Plasmodium falciparum. Both diaste-
reoisomers exhibited moderate but interesting antiplasmodial
activities, with IC50 = 13 mm for the cis diastereoisomer and
IC50 = 4.4 mm for the trans diastereoisomer.[14,15]
Figure 2. Cyclic voltammograms of 1b (17.2 mm) in MeCN with
TBABF4 (0.1m) recorded at a platinum electrode (0.5 mm diameter) at
0.2 VsÀ1, under aerobic conditions, before (Ox1) and after the prepara-
tive electrolysis (Ox2).
In summary, cyclic voltammetry performed on bicyclic
aminocyclopropanes 1a–e, 2a,b, 3, and 4 provides new insight
on the influence of the substitution pattern of these com-
pounds on their aerobic oxidation, and is fully consistent with
the currently accepted autocatalytic mechanism. A simple,
environmentally friendly, and efficient procedure for the
electrosynthesis of the corresponding endoperoxides was also
developed. The instability of some of these molecules was
confirmed, and conditions were found for the preparation of
stable compounds. We believe that these findings could open
the route to the design of similar oxidation methods starting
from a broader range of cyclopropane derivatives. Moreover,
reaction partners other than molecular oxygen could be
envisaged, which will be the subject of our future work in this
area, as well as the quest for new antimalarial compounds.
the potential applied). Interestingly, in some cases Ox2 is
reversible, indicating the stability of the cation radicals
obtained by oxidation of the endoperoxides.
After workup, the endoperoxide 5c prepared from 1c
decomposed too rapidly to be analysed by NMR spectrosco-
Experimental Section
Typically, the procedure for preparative electrolysis was as follows:
Aminocyclopropane 2a (65.4 mg; 197 mmol) was introduced in the
anodic compartment of a divided cell containing the solvent MeCN
(22 mL) and the supporting electrolyte TBABF4 (0.1 molLÀ1). A
potential value of + 0.60 V/SCE was then applied between the
reference and the working electrodes while bubbling air into the cell.
After 1.5 h, the crude solution contained in the anodic compartment
was concentrated under vacuum, extracted with diethyl ether,
filtered, and concentrated under reduced pressure to afford an
orange oil, which was analysed by NMR spectroscopy. Purification by
flash column chromatography on silica gel (ethyl acetate/heptane
10%) led to a 50:50 mixture of the two possible diastereoisomers of
pure compound 6a (33.0 mg, 90.7 mmol, 46%) as a yellow oil.
For further experimental details, see the Supporting Information.
py. The five-membered ring analogue 5b (from 1b) proved to
be somewhat more stable, and 1H and 13C NMR spectra could
be obtained. This compound had nevertheless totally decom-
posed the next day. Similar observations were made with 5d,
although its formation was slower compared to the other
products, in agreement with the absence of Ox2 on the
timescale of the cyclic voltammetry performed on 1d. This
highlights the influence of an alkoxy group at the para
position of the aromatic ring on the reactivity of the
intermediates involved in the chemical reactions following
the electron transfer. Endoperoxide 7 was easily obtained
from 3 but decomposed slowly, while 8 (from 4) was too
unstable to be analysed even just after its preparation. Finally,
6a (from 2a) was the only truly stable endoperoxide obtained,
showing that the additional substitution on the dioxolane ring
plays a major role in its stabilization.[11] Another effect of the
higher degree of substitution of the cyclopropane ring in 2a is
a faster conversion into the peroxide 6a as compared to the
reaction of 1d. This can be rationalized by invoking the
equilibrium between the cyclopropyl aminyl cation radical
and the ring-opened iminium radical shown in Scheme 1, with
the latter species being stabilized when the carbon-centered
Received: June 29, 2007
Published online: September 17, 2007
Keywords: cyclic voltammetry · electrochemical oxidation ·
.
peroxides · small ring systems · titanium
[1]D. H. Gibson, C. H. DePuy, Tetrahedron Lett. 1969, 10, 2203 –
2206.
8048 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 8046 –8049

Angewandte
Chemie
[2]A. P. Schaap, L. Lopez, S. D. Anderson, S. D. Gagnon, Tetrahe-
dron Lett. 1982, 23, 5493 – 5496.
[8]a) V. Chaplinski, A. de Meijere, Angew. Chem. 1996, 108, 491 –
492; Angew. Chem. Int. Ed. Engl. 1996, 35, 413 – 414; review
articles: b) O. G. Kulinkovich, A. de Meijere, Chem. Rev. 2000,
100, 2789 – 2834; c) A. de Meijere, S. I. Kozhushkov, A. I.
Savchenko in Titanium and Zirconium in Organic Synthesis
(Ed.: I. Marek), Wiley-VCH, Weinheim, 2002, pp. 390 – 434;
d) A. de Meijere, S. I. Kozhushkov, A. I. Savchenko, J. Organo-
met. Chem. 2004, 689, 2033 – 2055.
[3]Recent examples: a) O. G. Kulinkovich, D. A. Astashko, V. I.
Tyvorskii, N. A. Ilyina, Synthesis 2001, 1453 – 1455; b) T.
Tokuyasu, S. Kunikawa, A. Masuyama, M. Nojima, Org. Lett.
2002, 4, 3595 – 3598; c) H. Ikeda, K. Akiyama, Y. Takahashi, T.
Nakamura, S. Ishizaki, Y. Shiratori, H. Ohaku, J. L. Goodman,
A. Houmam, D. D. M. Wayner, S. Tero-Kubota, T. Miyashi, J.
Am. Chem. Soc. 2003, 125, 9147 – 9157; d) H. Shimizu, S.
Onitsuka, H. Egami, T. Katsuki, J. Am. Chem. Soc. 2005, 127,
5396 – 5413; e) M. Kirihara, H. Kakuda, M. Ichinose, Y. Ochiai,
S. Takizawa, A. Mokuya, K. Okubo, A. Hatano, M. Shiro,
Tetrahedron 2005, 61, 4831 – 4839.
[4]K. Wimalasena, H. B. Wickman, M. P. D. Mahindaratne, Eur. J.
Org. Chem. 2001, 3811 – 3817.
[5]N. Ouhamou, Y. Six, Org. Biomol. Chem. 2003, 1, 3007 – 3009.
[6]J. A. Kowalska-Six, C. Madelaine, Y. Six, B. Crousse, unpub-
lished results.
[9]J. Lee, J. K. Cha, J. Org. Chem. 1997, 62, 1584 – 1585.
[10]A. J. Bard, L. R. Faulkner, Electrochemical Methods, Wiley, New
York, 1980.
[11]The stable endoperoxides previously synthesized in our labo-
ratory also featured this extra subsitution on the dioxolane
system. See Ref. [5].
[12]For a review covering new antimalarial drugs, see: J. Wiesner, R.
Ortmann, H. Jomaa, M. Schlitzer, Angew. Chem. 2003, 115,
5432 – 5451; Angew. Chem. Int. Ed. 2003, 42, 5274 – 5293.
[13]A. Robert, F. Benoit-Vical, B. Meunier, Coord. Chem. Rev. 2005,
249, 1927 – 1936, and references therein.
[7]To the best of our knowledge, there is only one report in the
literature showing cyclic voltammetry performed on cyclopro-
pylamines: H. B. Lee, M. J. Sung, S. C. Blackstock, J. K. Cha, J.
Am. Chem. Soc. 2001, 123, 11322 – 11324; see also Supporting
Information.
[14]Chloroquine: IC 50 = 0.11 mm; IC90 = 0.21 mm.
[15]The cis and trans diastereoisomers are defined according to the
relative configurations of the amine and the 2-(tert-butyldime-
thylsilyloxy)ethyl substituents of the dioxolane ring.
Angew. Chem. Int. Ed. 2007, 46, 8046 –8049ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 8049