Characterization of the Main Radical and Products Resulting from a Reductive Activation of the Antimalarial Arteflene (Ro 42-1611)

Article abstract of DOI:10.1021/jo990744z

The peroxide-containing antimalarial drug arteflene (Ro 42-1611) generates an alkyl radical after the reductive homolytic cleavage of the peroxide bond in the presence of a heme model Mn^II(TPP). This alkyl radical has been trapped by TEMPO, and the different products of the reduction activation of arteflene have been characterized. These data suggest that, in these experimental conditions, arteflene is not a significant alkylating agent compared to artemisinin, a trioxane-containing antimalarial drug.

Full text of DOI:10.1021/jo990744z

6776
J . Org. Chem. 1999, 64, 6776-6781
Ch a r a cter iza tion of th e Ma in Ra d ica l a n d P r od u cts Resu ltin g
fr om a Red u ctive Activa tion of th e An tim a la r ia l Ar teflen e
(Ro 42-1611)
J e´roˆme Cazelles, Anne Robert, and Bernard Meunier*
Laboratoire de Chimie de Coordination du CNRS, 205 route de Narbonne,
31077 Toulouse Cedex 4, France
Received May 4, 1999
The peroxide-containing antimalarial drug arteflene (Ro 42-1611) generates an alkyl radical after
the reductive homolytic cleavage of the peroxide bond in the presence of a heme model MnII(TPP).
This alkyl radical has been trapped by TEMPO, and the different products of the reduction activation
of arteflene have been characterized. These data suggest that, in these experimental conditions,
arteflene is not a significant alkylating agent compared to artemisinin, a trioxane-containing
antimalarial drug.
Sch em e 1. Str u ctu r e of Ar tem isin in 1,
Yin gzh a osu 2, a n d Ar teflen e 3
In tr od u ction
Artemisia a. is a plant with a long history of medical
use against fever and malaria in China.1 The active
molecule, artemisinin (1, Scheme 1), isolated in 1972, is
a sesquiterpene lactone containing an endoperoxide
function critical for its activity. This drug and its hemi-
synthetic derivatives artesunate and artemether are
currently being developed to treat severe and multidrug-
resistant malaria, including cerebral malaria.
A reasonable hypothesis on the mechanism of action
of artemisinin is the reductive activation of its endoper-
oxide bridge by FeII-heme resulting from the digestion
of hemoglobin by Plasmodium,2,3 which leads to dioxygen-
derived radicals responsible for an oxidative stress4
within infected erythrocytes or to C-centered radicals.5-10
We recently illustrated the alkylating properties of one
of these C-centered radicals by reporting the character-
ization of a covalent adduct between artemisinin and a
heme model based on meso-tetraphenylporphyrin.11,12 The
isolated chlorin-type adduct resulted from a C-alkylation
of a â-pyrrolic position of the porphyrin macrocycle by a
carbon radical at the C4 position of artemisinin produced
after reductive cleavage of the peroxide bridge. In similar
conditions, we observed the same alkylating behavior of
â-artemether, a modified analogue of artemisinin. We
also found that the synthetic antimalarial trioxane BO7
gave a chlorin adduct after reductive activation.10,12 These
results suggested that the alkylating ability of artemisi-
nin, which is responsible for the death of the parasite, is
not limited to this naturally occurring product but is a
general feature probably required for the antimalarial
activity of endoperoxide-containing drugs (for a recent
review on the alkylating ability of artemisinin and related
trioxanes, see ref 13).
(1) Klayman, D. L. Science 1985, 228, 1049-1055.
(2) Meshnick, S. R.; Taylor, T. E.; Kamchonwongpaisan, S. Microbiol.
Rev. 1996, 60, 301-315.
Unfortunately, the extraction and, therefore, the cost
of artemisinin derivatives is a limiting factor to the wide
use of these efficient drugs in tropical and subtropical
regions, where malaria is endemic and chloroquine-
resistant strains are spreading. There is a need for new
low-cost synthetic drugs with pharmacological properties
similar to those of artemisinin. Many artemisinin
models5,14-19 and other naturally occurring peroxide
compounds,20 different than artemisinin, with good an-
timalarial activity have been reported. Yingzhaosu A (2,
Scheme 1) is one of these natural peroxide compounds,
but little information is available about the biological
activity of this compound, probably because its isolation
from the plant source, a rare ornemental vine, Artabotrys
uncinatus (L.) Merr., was difficult and erratic and did
not provide sufficient amounts of material for a proper
(3) Kamchonwongpaisan, S.; Meshnick, S. R. Gen. Pharmacol. 1996,
27, 587-592.
(4) Krungkrai, S. R.; Yuthavong, Y. Trans. R. Soc. Trop. Med. Hyg.
1987, 81, 710-714.
(5) Cumming, J . N.; Ploypradith, P.; Posner, G. H. Adv. Pharmacol.
1997, 37, 253-297.
(6) Posner, G. H.; Oh, C. H.; Wang, D.; Gerena, L.; Milhous, W. K.;
Meshnick, S. R.; Asawamahasakda, W. J . Med. Chem. 1994, 37, 1256-
1258.
(7) Hong, Y.-L.; Yang, Y.-Z.; Meshnick, S. R. Mol. Biochem. Parasitol.
1994, 63, 121-128.
(8) Asawamahasakda, W.; Ittarat, I.; Pu, Y.-M.; Ziffer, H.; Meshnick,
S. R. Antimicrob. Agents Chemother. 1994, 38, 1854-1858.
(9) Meshnick, S. R. Lancet 1994, 344, 1441-1442.
(10) J efford, C. W.; Kohmoto, S.; J aggi, D.; Timari, G.; Rossier, J .-
C.; Rudaz, M.; Barbuzzi, O.; Ge´rard, D.; Burger, U.; Kamalaprija, P.;
Mareda, J .; Bernardinelli, G.; Manzanares, I.; Canfield, C. J .; Fleck,
S. L.; Robinson, B. L.; Peters, W. Helv. Chim. Acta 1995, 78, 647-
662.
(11) Robert, A.; Meunier, B. J . Am. Chem. Soc. 1997, 119, 5968-
5969.
(12) Robert, A.; Meunier, B. Chem. Eur. J . 1998, 4, 1287-1296.
(13) Robert, A.; Meunier, B. Chem. Soc. Rev. 1998, 27, 273-279.
10.1021/jo990744z CCC: $18.00 © 1999 American Chemical Society
Published on Web 08/10/1999

Activation of the Antimalarial Arteflene (Ro 42-1611)
J . Org. Chem., Vol. 64, No. 18, 1999 6777
evaluation of its antimalarial activity.21,22 However, a
series of analogues have been prepared and evaluated
as antimalarial agents. In particular, molecules having
a keto group within the ring system and a lipophilic
substituent possessed activities comparable to those of
artemisinin. One of these models, arteflene (3, Scheme
1), has been taken through clinical trials.22-24 However,
this drug is associated with a certain degree of recru-
descence.3
The reactivity of arteflene in the presence of a heme
model, under experimental conditions described for ar-
temisinin and BO7, has been investigated. We found that
a manganese(II) porphyrin can activate the peroxide
function of 2 by electron transfer and induces the homo-
lytic cleavage of the C4-C5 bond. The resulting alkyl
radical centered at C5 was not in a suitable position to
alkylate the porphyrin ligand located in the vicinity via
an intramolecular process. The porphyrin ligand was
recovered unchanged, but trapping of the C5-centered
radical by TEMPO allowed us to isolate and characterize
the two fragments of arteflene. These data provide a
deeper insight into the mechanism of action of this
antimalarial agent.
the cis/trans ratio was 30/70. The cis olefin was purified by
column chromatography (SiO2, pentane/diethyl ether 95/5 to
90/10, v/v).26 [1H NMR, 250 MHz in CD2Cl2: δ, ppm 7.30-
7.60 (m, 5H), 6.88 and 6.21 (2 × d, 2 × 1H, J ) 12.6), 2.18 (s,
3H)]. Aluminum oxide 90, 70-230 mesh, activity II-III
(Merck) and silica 60, 70-200 µm (SDS, France) were used
for column chromatography.
Rea ction of Ar teflen e w ith Mn III(TP P )Cl in th e P r es-
en ce of Tetr a -n -bu tyla m m on iu m Bor oh yd r id e. Isola tion
of Com p ou n d s 6-11. MnIII(TPP)Cl (30 mg, 43 µmol, 1 equiv)
and arteflene (53 mg, 129 µmol, 3 equiv) were dissolved in
CH2Cl2 (5 mL). This solution was carefully degassed and kept
under a nitrogen atmosphere. Tetra-n-butylammonium boro-
hydride (110 mg, 430 µmol, 10 equiv) was then added as a
solid. The mixture was allowed to stand for 3.5 h at room
temperature with magnetic stirring under nitrogen. Demeta-
lation was performed as previously decribed12 by addition of
a degassed solution of cadmium(II) nitrate [(Cd(NO3)2‚4H2O,
263 mg, 860 mmol, 20 equiv) in DMF (2 mL)] to the reaction
mixture, and stirring was continued for 20 min. An aqueous
solution of acetic acid (10 vol %, 10 mL) was then added under
air. The organic layer was extracted with CH2Cl2, washed with
water, dried over sodium sulfate, and evaporated to dryness.
Purification of the crude product was performed by column
chromatography on silica gel using a hexane/dichloromethane
mixture (gradient from 60/40 to 10/90, v/v). The tetraphen-
ylporphyrin ligand was eluted at the solvent front followed by
a mixture of compounds 6, 7, and 8, which was recovered by
evaporation of the solvent (6/7/8 ratio was in the range of 45/
36/19 to 30/30/40, overall yield 28% with respect to arteflene).
NMR characterization was performed on the mixture of
compounds 6-8. For clarity, the spectra of the three com-
pounds are described separately. Compound 6: 1H NMR (δ,
CD2Cl2) 8.00-7.70 (m, 3H), 6.69 (br d, J ) 11.7, 1H), 5.91 (dd,
J ) 11.7 and 9.5, 1H), 4.36 (m, J ) 9.5, 6.2 and 1.0, 1H), 1.27
(d, J ) 6.2, 3H). Compound 7: 1H NMR (δ, CD2Cl2) 8.00-7.70
(m, 3H), 6.96 (br d, J ) 15.8, 1H), 6.39 (dd, J ) 15.8 and 5.8,
1H), 4.54 (m, J ) 6.5, 5.8 and 1.7, 1H), 1.37 (d, J ) 6.5, 3H).
Compound 8: 1H NMR (δ, CD2Cl2) 8.00-7.70 (m, 2H), 7.50
(d, J ) 7.5, 1H), 3.87 (m, J ) 6.2 and 6.2, 1H), 2.95 (m, 2H),
1.73 (m, 2H), 1.23 (d, J ) 6.2, 3H).
Exp er im en ta l Section
Ma t er ia ls. (1S,4R,5R,8S)-4-[(Z)-2,4-Bis(trifluoromethyl)-
styryl-4,8-dimethyl-2,3-dioxabicyclo[3.3.1]nonan-7-one, namely,
arteflene (Ro 42-1611), was a gift from F. Hoffmann-La Roche.
Dichloromethane (stabilized with amylene) and hexane sup-
plied by Fluka were of low evaporation residue content
(e0.0005%). All other commercially available reagents and
solvents were obtained from Aldrich or Fluka. The tetra-n-
butylammonium borodeuteride was prepared by reaction of
tetra-n-butylammonium chloride with sodium borodeuteride
98%D: NaBD4 (100 mg, 2.4 mmol, 1.5 equiv) and NaOH (4
mg) were dissolved in deuterated water (1 mL). This solution
was mixed with a solution containing n-Bu4N+Cl- (443 mg,
1.6 mmol, 1 equiv) in D2O (1 mL) and stirred for 1 min. The
resulting tetra-n-butylammonium borodeuteride was extracted
with CH2Cl2 (5 mL). The organic layer was dried over sodium
sulfate, and the product was recovered by evaporation of the
solvent to dryness (yield: 80%; 95+ atom % D measured by
The mixture of compounds 6-8 (2.9 mg) was stirred at room
temperature with 4-(dimethylamino)pyridine (DMAP, 1.5 mg,
12 µmol, 1.2 equiv) and acetic anhydride (9.4 µL, 100 µmol,
10 equiv) in CH2Cl2 (500 µL). After 1 h of reaction, the
resulting products were passed through an alumina column
1H NMR). [IR (KBr) ν ) 1756, 1721, 1681 cm-1 (BD4-)]. MnIII
-
with CH2Cl2 as eluent. The mixture of acetylated derivatives
+
9-11 was recovered by evaporation of solvent. MS (DCI/NH3
)
(TPP)Cl was prepared by metalation of chlorin-free H2TPP
with MnII(OAc)2‚4H2O in DMF in the presence of 2,4,6-
collidine.25 The cis-benzylideneacetone was obtained by ir-
radiation of the commercially available trans-isomer with a
200 W tungsten lamp in acetonitrile under argon. After 48 h,
m/z 343(15), 344(100)[M+ + 18 for 9 and 10], 345(22), 346-
(64)[M+ + 18 for 11], 347(10). NMR characterization was
performed on the mixture of compounds 9-11, but for clarity,
the spectra of the three compounds are described separately.
Compound 9: 1H NMR (δ, CDCl3) 8.00-7.60 (m, 3H), 6.69 (br
d, J ) 11.8, 1H), 5.84 (dd, J ) 11.8 and 9.0, 1H), 5.42 (m, J )
9.0, 6.2 and 1.0, 1H), 1.99 (s, 3H), 1.25 (d, J ) 6.2, 3H).
Compound 10: 1H NMR (δ, CDCl3) 8.00-7.60 (m, 3H), 6.95
(br d, J ) 15.8, 1H), 6.26 (dd, J ) 15.8 and 5.8, 1H), 5.54 (m,
J ) 6.5, 5.8 and 1.7, 1H), 2.09 (s, 3H), 1.42 (d, J ) 6.5, 3H).
Compound 11: 1H NMR (δ, CDCl3) 8.00-7.60 (m, 2H), 7.45
(d, J ) 7.5, 1H), 4.97 (m, J ) 6.2 and 6.2, 1H), 2.86 (m, 2H),
2.04 (s, 3H), 1.88 (m, 2H), 1.27 (d, J ) 6.2, 3H).
(14) J efford, C. W. Adv. in Drug Res. 1997, 29, 271-325.
(15) Torok, D. S.; Ziffer, H.; Meshnick, S. R.; Pan, X.-Q. J . Med.
Chem. 1995, 38, 5045-5050.
(16) Lin, A. J .; Zikry, A. B.; Kyle, D. E. J . Med. Chem. 1997, 40,
1396-1400.
(17) Avery, M. A.; Fan, P.; Karle, J . M.; Miller, R.; Goins, K.
Tetrahedron Lett. 1995, 36, 3965-3968.
(18) Posner, G. H.; Cumming, J . N.; Woo, S.-H.; Ploypradith, P.; Xie,
S.; Shapiro, T. A. J . Med. Chem. 1998, 41, 940-951.
(19) Cumming, J . N.; Wang, D.; Park, S. B.; Shapiro, T. A.; Posner,
G. H. J . Med. Chem. 1998, 41, 952-964.
Isola tion of Com p ou n d 12. MnIII(TPP)Cl (12 mg, 16 µmol,
1 equiv), arteflene (20 mg, 49 µmol, 3 equiv), and 2,2,6,6-
tetramethylpiperidine 1-oxyl (TEMPO, 15 mg, 98 µmol, 6
equiv) were dissolved in CH2Cl2 (2 mL). This solution was
degassed and kept under argon. Tetra-n-butylammonium
borohydride (42 mg, 163 µmol, 10 equiv) was then added as a
solid. The mixture was stirred at room temperature for 3.5 h.
Upon exposure to air, the dark green solution was washed
with water (pH 5), dried over sodium sulfate, and evaporated
to dryness. The crude mixture was then acetylated by addi-
(20) Kamchonwongpaisan, S.; Nilanonta, C.; Tarnchompoo, B.;
Thebtaranonth, C.; Thebtaranonth, Y.; Yuthavong, Y.; Kongsaeree, P.;
Clardy J . Tetrahedron Lett. 1995, 36, 1821-1824.
(21) Zhang, L.; Zhou, W.-S.; Xu, X.-X. J . Chem. Soc., Chem. Com-
mun. 1988, 523-524.
(22) Hofheinz, W.; Bu¨rgin, H.; Gocke, E.; J aquet, C.; Masciadri, R.;
Schmid, G.; Stohler, H.; Urwyler, H. Trop. Med. Parasitol. 1994, 45,
261-265.
(23) J aquet, C.; Stohler, H. R.; Chollet, J .; Peters, W. Trop. Med.
Parasitol. 1994, 45, 266-271.
(24) Weidekamm, E.; Dumont, E.; J aquet, C. Trop. Med. Parasitol.
1994, 45, 278-283 and references therein.
(25) Hoffmann, P.; Robert, A.; Meunier, B. Bull. Soc. Chim. Fr. 1992,
129, 85-97.
(26) J acobsen, E. N.; Deng, L.; Furukawa, Y.; Martinez, L. E.
Tetrahedron 1994, 50, 4323-4334.

6778 J . Org. Chem., Vol. 64, No. 18, 1999
Cazelles et al.
tion of a solution of DMAP (16 mg, 130 µmol, 8 equiv) and
(CH3CO)2O (92 µL, 978 µmol, 60 equiv) in CH2Cl2 (2 mL). After
1 h of stirring at room temperature, the reaction mixture was
washed with water (successively at pH 5, pH 8, and pH 5) and
dried over sodium sulfate. The purification step was performed
using an alumina column with a mixture hexane/dichlo-
romethane (gradient from 100/0 to 0/100, v/v). A mixture of
the acetylated derivatives 9 and 11 was first eluted (ratio 9/11
) 100/0 to 85/15, v/v); the adduct 12 was then recovered in
6% yield with respect to arteflene. Rf ) 0.5 (SiO2, hexane/
Sch em e 2. Isola ted P r od u cts fr om th e Red u ctive
Activa tion of Ar teflen e by Zn /CH3COOH, F eCl2 or
Hem in Ch lor id e/N-Acetylcystein e28
1
dichloromethane, 60/40 v/v); H NMR (δ, CD2Cl2) 5.14 (m, J
) 3.5, 3.5 and 3.5, 1H), 4.74 (m, J ) 11.0, 11.0 and 4.0, 1H),
3.98 (m, J ) 12.0, 12.0, 4.0 and 4.0, 1H), 2.55 (m, J ) 11.4,
4.0, 4.0 and 2.0, 1H), 2.34 (m, J ) 13.8, 4.0, 3.5 and 2.0, 1H),
2.01 and 2.03 (2 × s, 6H), 1.72 (m, J ) 11.0, 7.0, and 3.5, 1H),
1.40-1.25 (m, 2H), 1.44 (m, 6H), 1.07 (m, 12H), 0.87 (d, J )
7.0, 3H); MS (DCI/NH3+) m/z 370(100)[M+ + 1], 371(24), 372-
(3). [R]365 ) +180 (c ) 290 × 10-6, CH2Cl2).
Rea ction of Ar teflen e w ith Mn III(TP P )Cl in th e P r es-
en ce of Tetr a -n -bu tyla m m on iu m Bor od eu ter id e. Isola -
tion of Com p ou n d s 9-d to 11-d 2. MnIII(TPP)Cl (17 mg, 24
µmol, 1 equiv) and arteflene (30 mg, 73 µmol, 3 equiv) were
dissolved in CH2Cl2 (1.5 mL). This solution was degassed and
kept under an argon atmosphere. Tetra-n-butylammonium
borodeuteride (63 mg, 244 µmol, 10 equiv) was then added as
a solid. After 1.5 h of stirring at room temperature, the reaction
mixture was washed with water (pH 5), dried over sodium
sulfate, and evaporated to dryness. The crude product was
purified using a silica column and hexane/dichloromethane
(gradient from 50/50 to 0/100, v/v). Acetylation was carried
out by addition of a solution of DMAP (8 mg, 655 µmol) and
(CH3CO)2O (24 µL, 254 µmol). After 1 h of stirring, the product
was washed with water (successively at pH 5, pH 8, and pH
5), dried over sodium sulfate, and purified on silica gel with
dichloromethane as eluent. After evaporation of the solvent,
the mixture of acetylated derivatives 9-d to 11-d 2 was recov-
ered. For clarity, NMR spectra are described separately.
3H). 4-P h en ylbu ta n -2-ol: 1H NMR (δ, CD2Cl2) 7.65-7.10 (m,
3H), 3.79 (m, 1H), 2.70 (m, 2H), 1.71 (m, 2H), 1.20 (d, J ) 6.0,
3H).
Resu lts a n d Discu ssion
The activation of the peroxide function of antimalarial
1,2,4-trioxanes and cyclic dialkylperoxides such as arte-
flene by heme resulting from the digestion of host
hemoglobin by Plasmodium is probably able to generate
in vivo alkylating species responsible for the parasite
death. Moreover, when [14C]-arteflene or [3H]-dihydroar-
temisinin were incubated with P. falciparum infected
erythrocytes, beside a small amount of low-molecular
heme-drug adducts (13-15% of the total parasite-
associated radioactivity), these drugs were able to specif-
ically alkylate the same and not particularly abundant
parasite proteins, one of which has a similar size to that
of the histidine-rich protein (HRP, 42 kDa)27 involved in
the polymerization of heme in infected erythrocytes (the
radioactivity associated with parasite proteins was ca.
65%).8 The reductive activation of arteflene by zinc in
acetic acid, by iron(II) chloride, or by hemin chloride
associated with N-acetylcysteine as reducing agent has
been recently investigated,28 in conditions similar to those
previously reported for artemisinin or synthetic trioxane
derivatives.29-31 In these conditions, the two identified
products (Scheme 2) of reaction of arteflene were the diol
13, resulting from the reductive cleavage of the O-O
bond followed by the reduction of the generated RO•, and
the R,â-unsaturated ketone 4 (this latter one is the major
one in the presence of hemin chloride/N-acetylcysteine).
The formation of the ketone 4 suggests the parallel
formation of a C-centered radical 5 (Scheme 3), but this
radical was not trapped in these reaction conditions.
Id en tifica tion of Com p ou n d s 6-8. In the present
study, we have been able to characterize different prod-
ucts resulting from the reductive activation of arteflene
3 by manganese(II) tetraphenylporphyrin. In the experi-
mental conditions leading to the alkylation of the por-
1
Compound 9-d : H NMR (δ, CD2Cl2) 8.00-7.50 (m, 3H), 6.72
(br d, J ) 11.7, 1H), 5.87 (d, J ) 11.7, 1H), 1.97 (s, 3H), 1.27
(s, 3H). The signal at 5.42 (H4) disappeared. Compound 10-d :
1H NMR (δ, CD2Cl2) 8.00-7.50 (m, 3H), 6.96 (br d, J ) 15.8,
1H), 6.32 (d, J ) 15.8, 1H), 2.08 (s, 3H), 1.40 (s, 3H). The sig-
nal at 5.54 (H4) disappeared. Compound 11-d 2: 1H NMR (δ,
CD2Cl2) 8.00-7.60 (m, 2H), 7.50 (d, J ) 7.5, 1H), 2.87 (m,
1.5H), 2.03 (s, 3H), 1.85 (m, 1.5H), 1.23 (s, 3H). The signal at
4.97 (H4) disappeared.
Isola tion of Com p ou n d 12-d . The synthesis and workup
were the same as for 12, but with tetra-n-butylammonium
borodeuteride as reducing agent [MnIII(TPP)Cl (12 mg, 16
µmol, 1 equiv), arteflene (20 mg, 49 µmol, 3 equiv), 2,2,6,6-
tetramethylpiperidine 1-oxyl (TEMPO, 15 mg, 98 µmol, 6
equiv), and tetra-n-butylammonium borodeuteride (43 mg, 163
µmol, 10 equiv) were dissolved in CH2Cl2 (2 mL)]. 1H NMR (δ,
CD2Cl2) signals were the same as for 12 except that the signal
at 4.74 (H7) disappeared. MS (DCI/NH3+) m/z 370(18), 371-
(100)[M+ + 1], 372(22), 373(4).
Rea ction of cis-Ben zylid en ea ceton e w ith Mn III(TP P )-
Cl in th e P r esen ce of Tetr a -n -bu tyla m m on iu m Bor oh y-
d r id e. A solution of MnIII(TPP)Cl (24 mg, 34 µmol, 1 equiv)
and cis-benzylideneacetone (15 mg, 103 µmol, 3 equiv) in
CH2Cl2 (3 mL) was degassed and kept under a nitrogen
atmosphere. Tetra-n-butylammonium borohydride (61 mg, 342
µmol, 10 equiv) was then added as a solid. After 4 h of stirring
at room temperature, the reaction mixture was washed with
water (pH 5), dried over sodium sulfate, and evaporated to
dryness. The crude product was purified using a silica column
with a hexane/dichloromethane gradient (from 60/40 to 0/100,
v/v) initially, followed by a dichloromethane/methanol mixture
(gradient from 100/0 to 97/3, v/v). After evaporation of the
solvent, a mixture of cis-4-phenyl-3-buten-2-ol and 4-phen-
ylbutan-2-ol (ratio 54/46) was recovered. For clarity, NMR
spectra are described separately. cis-4-P h en yl-3-bu ten -2-ol:
1H NMR (δ, CD2Cl2) 7.65-7.10 (m, 3H), 6.49 (d, J ) 11.9, 1H),
5.68 (dd, J ) 11.9 and 9.0, 1H), 4.75 (m, 1H), 1.33 (d, J ) 6.0,
(27) Sullivan, D. J ., J r.; Gluzman, I. Y.; Goldberg, D. E. Science 1996,
271, 219-222.
(28) O′Neill, P. M.; Bishop, L. P.; Searle, N. L.; Maggs, J . L.; Ward,
S. A.; Bray, P. G.; Storr, R. C.; Park, B. K. Tetrahedron Lett. 1997, 38,
4263-4266.
(29) Posner, G. H.; Cumming, J . N.; Ploypradith, P.; Oh, C. H. J .
Am. Chem. Soc. 1995, 117, 5885-5886.
(30) J efford, C. W.; Vicente, M. G. H.; J acquier, Y.; Favarger, F.;
Mareda, J .; Millasson-Schmidt, P.; Brunner, G.; Burger, U. Helv. Chim.
Acta 1996, 79, 1475-1487.
(31) Haynes, H. K.; Vonwiller, S. C. Tetrahedron Lett. 1996, 37, 257-
260.

Activation of the Antimalarial Arteflene (Ro 42-1611)
J . Org. Chem., Vol. 64, No. 18, 1999 6779
Sch em e 3. Mech a n ism of Red u ctive Activa tion of Ar teflen e by Mn II(TP P ) in th e P r esen ce of Bor oh yd r id e;
R ) 2,4-Bis(tr iflu or om eth yl)p h en yl
phyrin ligand by artemisinin, â-artemether, and BO7 at
a â-pyrrolic position,11,12 arteflene was not able to alkylate
the porphyrin ligand (MnIITPP being generated in situ
346 (M+ + 18 for 11), whereas no molecular peaks could
be detected under similar conditions for the alcohols 6-8.
Isola tion of Com p ou n d 12. The stable free radical
2,2,6,6-tetramethylpiperidine 1-oxyl (TEMPO) is a spe-
cific trap for alkyl radicals (rate constant for trapping R•
radicals ) 109 M-1 s-1; it reacts 106 times slower with
alkoxyl radicals such as t-BuO• and not at all with
alkylperoxyl radicals). Furthermore, the reaction between
TEMPO and R• gives only the coupled product TEMPO-
R, and TEMPO does not abstract hydrogen atoms that
are bound to the carbon adjacent to the radical center in
-
from MnIII(TPP)Cl and BH4 in CH2Cl2). After the reac-
tion of MnIITPP with arteflene, the crude reaction
mixture was treated to demetalate the metalloporphyrin
via a transmetalation from manganese(II) to cadmium
(II).12,13 This mild demetalation procedure avoids possible
modification of chlorin or porphyrin adducts due to
strongly acidic conditions that are usually required for
the demetalation of manganese porphyrins. The unmodi-
fied free-base tetraphenylporphyrin was recovered after
chromatography, but no macrocycle modified by arteflene
alkylation was detected in the reaction mixture.
However, the products 6-8, resulting from arteflene
fragmentation, were isolated in a 28% yield with respect
to the initial amount of arteflene and characterized by
NMR as a mixture (see Scheme 3 for structures of 6-8).
The H4 protons were detected at 4.36 and 4.54 ppm for
the allylic alcohols 6 and 7, respectively, and the values
of the coupling constants between H10 and H11 (11.8 and
15.8 Hz in compounds 6 and 7, respectively, and 13.3 Hz
in arteflene) indicated the isomerization of the C10-C11
double bond of 4 to produce the thermodynamically more
stable trans-isomer during the reduction.
-
R•.32,33 Arteflene was reacted with MnIII(TPP)Cl and BH4
in the presence of TEMPO (2 equiv versus arteflene). The
manganese porphyrin was not demetalated (it was easier
to remove Mn(TPP)Cl than the free-base porphyrin
ligand by chromatography). The crude reaction mixture
was directly acetylated, and compound 12, resulting from
the addition of the cyclohexyl radical 5 to TEMPO, was
recovered by chromatography (yield 6%). The TEMPO
1
adduct 12 was completely characterized by its H NMR
spectrum. The presence of two acetyl resonances at 2.01
and 2.03 ppm indicated the derivatization of two hydroxyl
functions at positions C1 and C7, the second one due to
the reduction of the ketone function of arteflene by
borohydride (δ ) 4.74 ppm for H7). At C7, the introduc-
tion of H- was stereospecifically achieved in an axial
position,34 as confirmed by the large coupling constants
of H7 with the axial protons H8 and H6a (both being
equal to 11.0 Hz). By using borodeuteride as reducing
agent (compound 12-d ), the complete introduction of D-
at position C7 resulted in the extinction of the resonance
at 4.74 ppm. The analysis of the 1H-1H coupling con-
stants also indicates that the configurations of C1 and
C8 were unchanged with respect to arteflene and that
the TEMPO moiety was introduced at C5 in an equatorial
position, as expected for a bulky substituent.
It should be mentioned that, by activation of arteflene,
a significant amount of product 8 resulting from the
hydrogenation of the C10-C11 double bond was also
recovered. The H4 proton was detected at 3.87 ppm, and
a multiplet at 2.95 was attributed to the benzylic CH2
group.
The acetylation of the mixture of 6-8 with acetic
anhydride in the presence of DMAP afforded quantita-
1
tively the derivatized alcohols 9-11. In H NMR spectra,
the acetyl groups were detected at 1.99, 2.09, and 2.04
ppm for compounds 9, 10, and 11, respectively, as proofs
for the presence of hydroxyl functions in compounds 6-8.
The signals of the protons H4 for 9, 10, and 11 were
deshielded by 1.0-1.1 ppm compared to corresponding
protons in alcohols 6, 7, and 8. The other signals were
not significantly modified by acetylation (∆δ < 0.15 ppm).
By mass spectrometry (DCI/NH3+), the molecular peaks
were identified at m/z ) 344 (M+ + 18 for 9 and 10) and
Mech a n ism of th e Activa tion of a r teflen e by
Red u ced Meta llop or p h yr in s. The formation of the
(32) Chateauneuf, J .; Lusztyk, J .; Ingold, K. U. J . Org. Chem. 1988,
53, 1629-1632.
(33) Root, K. S.; Hill, C. H.; Lawrence, L. M.; Whitesides, G. M. J .
Am. Chem. Soc. 1989, 111, 5405-5412.
(34) Che´rest, M.; Felkin, H. Tetrahedron Lett. 1968, 2205-2208

6780 J . Org. Chem., Vol. 64, No. 18, 1999
Cazelles et al.
alcohols 6-8 can be explained by the reductive activation
of the peroxide function of 3 by the manganese(II)
porphyrin, resulting in the homolytic cleavage of the O-O
bond (Scheme 3). The alkoxyl radical formed on O3
quickly induces the â-scission of the C4-C5 bond,
producing the R,â-unsaturated ketone 4 and the cyclo-
hexyl radical 5. (a Priori, the O-O bond cleavage might
also produce an alkoxyl radical centered on O2, able to
isomerize by â-scission to other alkyl radicals; although
this pathway was not identified until now, it cannot be
completely ruled out owing to the moderate yield of the
identified products. However, the interaction between the
manganese(II) porphyrin and arteflene may also favor
the formation of the radical on O3 if the docking of
arteflene with MnIITPP favors the approach of O2 on the
metal center). Compound 4 is not stable under the
reaction conditions. In the presence of an excess of
borohydride, the reduction of the ketone to give the
alcohol 6 was expected. The R,â-unsaturated ketone 4 can
also be reduced by addition of borohydride to the conju-
gated double bond to give the saturated ketone which,
after reduction of the carbonyl, produces the saturated
alcohol 8. It has been well-documented that the reduction
of conjugated ketones by NaBH4 can follow either 1,2-
addition to the carbonyl bond or 1,4-addition to the
conjugated double bond, leading in general to substantial
amounts of fully saturated alcohols (for example, the
reduction of substituted cis- and trans-cinnamates to give
dihydrocinnamates was reported in ref 34). When the
model compound cis-benzylideneacetone was submitted
to the same reaction conditions [MnIII(TPP)Cl + BH4-],
we observed, beside the reduction of the carbonyl, the
reduction of the conjugated double bond with a 25% yield,
the products of the reaction being cis-4-phenyl-3-buten-
2-ol and 4-phenylbutan-2-ol (ratio 54/46). When trans-
benzylideneacetone was reacted in the same reaction
conditions, the ratio of trans-allylic alcohol to the fully
saturated alcohol was 70/30, close to that obtained in the
case of cis-benzylideneacetone.
The production of the trans-allylic alcohol 7, obtained
after reductive activation of arteflene, which contains a
cis double bond, is difficult to explain. It has to be
mentioned that this product is not a minor one, the ratio
6/7 being close to 1/1. In addition, when cis-benzylide-
neacetone was used as substrate, no significant amount
of the corresponding trans-allylic alcohol could be ob-
tained. It is noteworthy that no traces of the trans-allylic
alcohol 10 were obtained after acetylation of the reaction
mixture when the activation was performed in the
presence of the alkyl radical trap TEMPO. This result
suggests that the isomerization of the C10-C11 double
bond of the olefin 4 may occur via a radical species
(probably an allyl radical) involving a reduced-metal
species.
When borohydride was replaced in the reaction mix-
ture by borodeuteride, the reduction of ketone at position
C4 was achieved by introduction of D-, leading to
complete disappearance of the signals due to H4 at 5.42,
5.54, and 4.97 ppm, in compounds 9-d , 10-d , and 11-d 2.
In compound 11-d 2, a second deuterium atom was
introduced in the molecule; the resonances of CH2 groups
at C10 and C11 were detected as two multiplets at 1.85
and 2.87 ppm, respectively, accounting for 1.5 proton
each, indicating the introduction of one deuterium atom
mediated 1,4-addition to conjugated double bonds was
reported in the case of R-phenylcinnamates.35
The reductive activation of arteflene reported in this
article is an illustration that the mechanism recently
proposed for the reductive activation of artemisinin and
related trioxanes by a Mn(II) porphyrin,11,12 involving the
formation of an O-acetal radical that is thermodynami-
cally driven to an ester and a C-centered radical, is
operating for many different peroxidic antimalarial drugs.
In the case of arteflene, an alkoxyl radical is formed
instead of an acetal radical, and it is driven to a ketone,
consequently releasing an alkyl radical in the medium.
Arteflene activated by a manganese(II) porphyrin pro-
vided the cyclohexyl radical 5 but did not produce any
species able to alkylate the porphyrin ligand. The forma-
tion of a manganese-O2 bond concomitant with the
formation of the alkoxyl radical on O3, followed by the
â-fragmentation between C4 and C5, generates the
alkylating carbon-centered radical (5) that does not
alkylate the porphyrin cycle, contrary to what has been
observed with the primary alkyl radical produced by
activation of artemisinin.11 In principle, the rate of
nucleophilic attack of an alkyl radical to alkene is
controlled by steric and polar effects: on a monosubsti-
tuted olefin, the rate of addition of the cyclohexyl radical
was found to be at least five times that of ethyl radical,
consistent with the frontier orbital theory.36 The absence
of alkylation of the porphyrin by 5 is therefore probably
due to steric hindrance factors and not to electronic ones.
The pyrrolic double bond of the metalloporphyrin, the
target of the alkyl radical, is not a simple olefin but a
conjugated one with two bulky phenyl groups on both
sides of these alkylable positions. Consequently, the
cyclohexyl radical 5 derived from arteflene is probably
too hindered to attack a â-pyrrolic position. Furthermore,
the close interaction of the peroxide function of arteflene
with the central ion of the metalloporphyrin suggests that
5 is probably formed away from any alkylable position
of the macrocycle, precluding its alkylation. It should be
noted that when [14C]-radiolabeled arteflene (at C11
position) was incubated with P. falciparum infected
erythrocytes, the radioactive center was recovered in the
covalent adducts between the drug and parasitic proteins.
Because C11 is not present in 5, this radical cannot be
involved in the arteflene-derived adducts observed in vivo
with the radiolabeled drug and is probably not respon-
sible for the parasite death. In fact, 5 can react with
molecular oxygen to generate the corresponding cyclo-
hexylperoxyl radical, which can dimerize and produce,
via a Russel reaction, the corresponding ketone and the
secondary alcohol at position C5. This cyclohexylperoxyl
radical might also be responsible for lethal radical chain
autoxidation processes. At high concentration, artemisi-
nin derivatives induce nonspecific oxidative damage
leading to the destruction of the parasite, but it should
be noted that this is observed at concentrations 103-105
times higher than in vitro effective antimalarial concen-
trations.37,38
In contrast, the C11 atom, incorporated in the radioac-
tive drug-protein adducts, is present in the R,â-unsatur-
(35) Schauble, J . H.; Walter, G. J .; Morin, J . G. J . Org. Chem. 1974,
39, 755-760.
(36) Giese, B. Angew. Chem., Int. Ed. Engl. 1983, 22, 753-764.
(37) Krungkrai, S. R.; Yuthavong, Y. Trans. R. Soc. Trop. Med. Hyg.
1987, 81, 710-714.
without selectivity between the positions C10 and C11.
(38) Scott, M. D.; Mesnick, S. R.; Williams, R. A.; Chiu, D. T.-Y. J .
Lab. Clin. Med. 1989, 114, 401-406.
-
The statistic incorporation of deuterium during the BD4
-

Activation of the Antimalarial Arteflene (Ro 42-1611)
J . Org. Chem., Vol. 64, No. 18, 1999 6781
ated ketone 4, which is a primary fragment generated
by reductive activation of arteflene (the isolated alcohols
6-8 are derived from 4 because of the presence of
borohydride in the presently used reaction conditions).
In vivo, the enone 4 may undergo a 1,4-addition from
nucleophilic side chains of amino acid residues, namely,
amine or thiol functions, giving rise to covalent addition
to the protein chain through an amine or thioether link,
respectively. Another possibility during the reductive
activation of arteflene by the reduced heme is the
formation of an alkoxyl radical on O2, instead of O3 as
observed when arteflene is activated by MnIITPP. In that
case, the â-scission of the RO• radical will induce the
cleavage of the C1-C9 bond, giving rise to a primary
alkyl radical centered at C9 still containing the whole
structure of arteflene, and therefore the radiolabeled C11
position will be retained. Subsequent alkylation of para-
sitic proteins by this nonsterically hindered alkyl radical
may be responsible for the parasite death. However, such
selectivity in the docking of arteflene onto the metal
center during the reductive peroxide opening by metal-
loporphyrins is rather unexpected.
the presently used reaction conditions and to a cyclohexyl
radical 5, which has been trapped by TEMPO. No
arteflene-macrocycle adducts have been obtained, con-
trary to what has been observed with artemisinin deriva-
tives. These data suggest that the C-centered radical
generated in the reductive activation of arteflene is
probably not responsible for the alkylation of parasite
proteins, in contrast to the behavior of trioxane-contain-
ing antimalarial drugs such as artemisinin.
Ack n ow led gm en t. This work was supported by
CNRS (Program “Physique et Chimie du Vivant”) and
by a grant from UNDP/World Bank/WHO Special Pro-
gram for Research and Training in Tropical Diseases
(Director’s Initiative Fund). F. Hoffmann-La Roche
(Basel, Switzerland) is gratefully acknowledged for a gift
of arteflene (Ro 42-1611).
Su p p or tin g In for m a tion Ava ila ble: IR spectrum of
tetra-n-butylammonium borodeuteride. 1H NMR spectra, with
peak assignment, of cis-benzylideneacetone, mixture of com-
pounds 6/7/8, mixture of compounds 9/10/11, compound 12,
mixture of compounds 9-d /10-d /11-d 2, compound 12-d , and
mixture of cis-4-phenyl-3-buten-2-ol and 4-phenylbutan-2-ol.
MS spectra (DCI/NH3+) of mixture of compounds 9/10/11,
compound 12, and compound 12-d . This material is available
free of charge via the Internet at http://pubs.acs.org.
Con clu sion
The reductive activation of the peroxide bridge of
arteflene by a metalloporphyrin generates the R,â-
unsaturated ketone 4, leading to alcohol derivatives in
J O990744Z