Reactions of Antimalarial Peroxides with Each of Leucomethylene Blue and Dihydroflavins: Flavin Reductase and the Cofactor Model Exemplified

Article abstract of DOI:10.1002/cmdc.201000508

Flavin adenine dinucleotide (FAD) is reduced by NADPH-E.coli flavin reductase (Fre) to FADH₂ in aqueous buffer at pH7.4 under argon. Under the same conditions, FADH₂ in turn cleanly reduces the antimalarial drug methylene blue (MB) to leucomethylene blue. The latter is rapidly re-oxidized by artemisinins, thus supporting the proposal that MB exerts its antimalarial activity, and synergizes the antimalarial action of artemisinins, by interfering with redox cycling involving NADPH reduction of flavin cofactors in parasite flavin disulfide reductases. Direct treatment of the FADH₂ generated from NADPH-Fre-FAD by artemisinins and antimalaria-active tetraoxane and trioxolane structural analogues under physiological conditions at pH7.4 results in rapid reduction of the artemisinins, and efficient conversion of the peroxide structural analogues into ketone products. Comparison of the relative rates of FADH₂ oxidation indicate optimal activity for the trioxolane. Therefore, the rate of intraparastic redox perturbation will be greatest for the trioxolane, and this may be significant in relation to its enhanced invitro antimalarial activities. ¹HNMR spectroscopic studies using the BNAH-riboflavin (RF) model system indicate that the tetraoxane is capable of using both peroxide units in oxidizing the RFH₂ generated insitu. Use of the NADPH-Fre-FAD catalytic system in the presence of artemisinin or tetraoxane confirms that the latter, in contrast to artemisinin, consumes two reducing equivalents of NADPH. None of the processes described herein requires the presence of ferrous iron. Ferric iron, given its propensity to oxidize reduced flavin cofactors, may play a role in enhancing oxidative stress within the malaria parasite, without requiring interaction with artemisinins or peroxide analogues. The NADPH-Fre-FAD system serves as a convenient mimic of flavin disulfide reductases that maintain redox homeostasis in the malaria parasite. Antimalarial peroxides and flavin reductase: NADPH-E.coli flavin reductase (Fre) reduces FAD to FADH₂, which in turn rapidly reduces artemisinins and antimalarial peroxides to deoxy or ketone products under physiological conditions. Thus, antimalarial activity is due to perturbation of intraparasitic redox homeostasis by oxidation of FADH₂ in critical flavoenzymes with consequent sequestration of NADPH. The tetraoxane uses both peroxide units in consuming two equivalents of NADPH in the NADPH-Fre-FAD system.

Full text of DOI:10.1002/cmdc.201000508

DOI: 10.1002/cmdc.201000508
Reactions of Antimalarial Peroxides with Each of
Leucomethylene Blue and Dihydroflavins: Flavin
Reductase and the Cofactor Model Exemplified
Richard K. Haynes,*[a] Kwan-Wing Cheu,[a] Maggie Mei-Ki Tang,[a] Min-Jiao Chen,[a]
Zu-Feng Guo,[a] Zhi-Hong Guo,*[a] Paolo Coghi,[b] and Diego Monti*[b]
Flavin adenine dinucleotide (FAD) is reduced by NADPH–E. coli
flavin reductase (Fre) to FADH2 in aqueous buffer at pH 7.4
under argon. Under the same conditions, FADH2 in turn cleanly
reduces the antimalarial drug methylene blue (MB) to leuco-
methylene blue. The latter is rapidly re-oxidized by artemisi-
nins, thus supporting the proposal that MB exerts its antimalar-
ial activity, and synergizes the antimalarial action of artemisi-
nins, by interfering with redox cycling involving NADPH reduc-
tion of flavin cofactors in parasite flavin disulfide reductases.
Direct treatment of the FADH2 generated from NADPH–Fre–
FAD by artemisinins and antimalaria-active tetraoxane and tri-
oxolane structural analogues under physiological conditions at
pH 7.4 results in rapid reduction of the artemisinins, and effi-
cient conversion of the peroxide structural analogues into
ketone products. Comparison of the relative rates of FADH2 ox-
idation indicate optimal activity for the trioxolane. Therefore,
the rate of intraparastic redox perturbation will be greatest for
the trioxolane, and this may be significant in relation to its en-
hanced in vitro antimalarial activities. 1H NMR spectroscopic
studies using the BNAH–riboflavin (RF) model system indicate
that the tetraoxane is capable of using both peroxide units in
oxidizing the RFH2 generated in situ. Use of the NADPH–Fre–
FAD catalytic system in the presence of artemisinin or tetraox-
ane confirms that the latter, in contrast to artemisinin, con-
sumes two reducing equivalents of NADPH. None of the pro-
cesses described herein requires the presence of ferrous iron.
Ferric iron, given its propensity to oxidize reduced flavin cofac-
tors, may play a role in enhancing oxidative stress within the
malaria parasite, without requiring interaction with artemisi-
nins or peroxide analogues. The NADPH–Fre–FAD system
serves as a convenient mimic of flavin disulfide reductases that
maintain redox homeostasis in the malaria parasite.
Introduction
Artemisinin 1, artesunate 2, artemisone 3,[1] and other deriva-
tives are transformed, through catalytic processes mediated by
the potent redox-active antimalarial drug methylene blue (MB)
7[2,3,4] in the presence of excess ascorbate or the NAD(P)H
model N-benzyl-1,4-dihydronicotinamide (BNAH) 13 in aque-
ous buffer at physiological pH, into products that arise via
single-electron transfer or two-electron reduction.[5] The two-
electron reduction involves the in situ generation of leuco-
methylene blue (LMB) 9, which is re-oxidized to MB by the ar-
temisinins, whilst the latter are reduced to deoxy products.
The observation explains the ability of artemisinins to syner-
gize the antimalarial activity of MB. MB acts as a subversive
substrate for the malaria parasite Plasmodium falciparum fla-
voenzymes glutathione reductase (GR) and thioredoxin reduc-
tase (TrxR), among others, thereby resulting in the sequestra-
tion of NADPH away from its role of reducing the essential co-
factor flavin adenine dinucleotide (FAD) required for ultimate
maintenance of reduced glutathione levels in the parasite. Oxi-
dation of LMB by artemisinins must speed up the cycle, thus
enhancing sequestration of NADPH.
the reduced conjugates RFH2 10, FMNH2, and FADH2 12 are
rapidly oxidized by the artemisinins. If yeast GR, used as a sur-
rogate for parasite GR, is treated with artemisinin in the pres-
ence of its substrate, oxidized glutathione (GSSG), accelerated
consumption of NADPH takes place, especially under aerobic
conditions.[5] Thus, incipient oxidation of FADH2 by artemisinin
results in formation of FAD, re-reduction by NADPH, and ensu-
ing autoxidation of the FADH2 by oxygen. This leads to en-
hanced consumption of NADPH; generation of reactive oxygen
species (ROS) must also occur, although no attempt has been
made to demonstrate this. Therefore, like MB, artemisinins per-
[a] Prof. Dr. R. K. Haynes, K.-W. Cheu, M. M.-K. Tang, M.-J. Chen, Dr. Z.-F. Guo,
Prof. Dr. Z.-H. Guo
Department of Chemistry
Institute of Molecular Technology for Drug Discovery and Synthesis
The Hong Kong University of Science and Technology
Clear Water Bay, Kowloon, Hong Kong (P.R. China)
Fax: (+852)2358-1594
E-mail: haynes@ust.hk
chguo@ust.hk
[b] Dr. P. Coghi, Prof. Dr. D. Monti
The two-electron reduction products are efficiently obtained
from the artemisinins and reduced flavins when these are gen-
erated in situ from catalytic amounts of the antimalaria-active
riboflavin (RF)[6,7] 9 or cofactors flavin mononucleotide (FMN)
and FAD 11 by reduction with BNAH 13 or NAD(P)H. Thereby,
Department of Organic and Industrial Chemistry
CNR-ISTM, Via G. Venezian 21, 20133 Milano (Italy)
E-mail: diego.monti@istm.cnr.it
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201000508.
ChemMedChem 2011, 6, 279 – 291
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
279

R. K. Haynes, Z.-H. Guo, D. Monti, et al.
MED
turb redox homeostasis in the malaria parasite by interfering
with redox cycling involving the flavin cofactor, and hence
NADPH, in the functioning of GR. They likely also affect TrxR,
lipoamide dehydrogenase, and others in the same way.[4]
This ‘cofactor model’ of artemisinin-induced cytotoxicity dif-
fers fundamentally from the ‘ferrous iron-reductive bio-activa-
tion’ model which posits a reaction of artemisinins with FeII to
provide carbon radicals as cytotoxic agents.[8] Model studies in-
dicate that both the FeII-mediated generation of carbon radi-
cals from artemisinins and their trapping under biologically
plausible conditions is difficult, as the radicals undergo facile
internal quenching.[9,10] Artemisinins susceptible to decomposi-
tion by hemoglobin– (Hb) or heme–FeII display enhanced activ-
ities against malaria parasites cultured under carbon monox-
ide, a reagent that blocks the reaction of Hb–FeII and heme–
FeII with the artemisinins through formation of stable FeII–CO
adducts. Thus, the much-publicized chemical reactions of arte-
misinins with heme represents an attrition pathway for artemi-
sinins in the intraparasitic environment.[11,12]
1,2,4,5-Tetraoxanes such as 4 and 5[13,14] and 1,2,4-trioxolanes
such as 6,[15,16] have potent antimalarial activities (Figure 1). Al-
though the reactivity of 6 with FeII was not assessed, other tri-
oxolanes are readily decomposed.[14,16,17] The arylsulfonamide
tetraoxane 5 is remarkably resistant to FeII under anhydrous
conditions (1.0 equiv FeBr2 in THF to 48 h).[14] Nevertheless, ‘fer-
rous iron-reductive bio-activation’ for the tetraoxanes is main-
tained by those that made this striking observation;[18] this re-
quires FeII, relatively transient under normoxic aqueous condi-
tions and likely bearing oxygen ligands within the intraparasitic
labile iron pool,[19] to be appreciably more active in vivo in gen-
erating carbon radicals than a pumped-up anhydrous FeII
halide, present at much higher concentrations in an organic
solvent in an inert atmosphere in a laboratory flask. As implicit
recognition of this problem, it is countered that ‘the parasite
death that ensues in the presence of artemisinin is more likely
to involve specific radicals and targets rather than nonspecific
cell damage caused by freely diffusing oxygen- and carbon-
centered radical species’.[20] This begs the questions as to the
nature of the radicals, the targets, and the putative role played
by FeIII chelators such as desferrioxamine (DFO).[10,21]
Figure 1. Artemisinins, peroxides, redox-active substrates, and ketone prod-
ucts discussed in this study. Antimalarial data for tetraoxanes 4 and 5 from
Ref. [18], trioxolane 6 from Ref. [16].
bind to an acceptor site by transferring hydride from
NAD(P)H.[23–27] Structures of isoforms that include the acceptor
binding site have been elucidated, and the manner in which
NAD(P)H reduces the bound acceptor has been discussed.[25,27]
Of interest is their ability to use the acceptors MB or RF admin-
istered during treatment of methemoglobinemia.[23,25] In
human erythrocytes, MB and RF are reduced by NADPH–flavin
reductase to LMB and RFH2, which in turn reduce methemo-
globin to Hb. The MB and RF thereby regenerated are reduced
by the NADPH–flavin reductase in continuation of the
cycle.[23,25] In a malaria-infected erythrocyte, it is not known if
the parasite uses flavin reductase for maintenance of redox ho-
meostasis. A functionally related ferredoxin–NADP+ reductase
occurs in the parasite apicoplast,[28] and a chorismate synthase
possessing flavin reductase activity has been detected in the
parasite cytosol.[29] MB is a subversive substrate for each of par-
asite GR, TrxR, and lipoamide dehydrogenase in which the
bound MB is reduced by the FADH2 generated by reduction of
We now extend the scope of the cofactor model by examin-
ing the behavior of the artemisinins and peroxides toward
LMB 8, RFH2 10, and FADH2 12 (Figure 1). Whilst previously
LMB, RFH2, and FADH2 were generated with sodium dithion-
ite,[5] we introduce E. coli flavin reductase (Fre)[22] for this pur-
pose. This has the special benefit of permitting both the gen-
eration of the reduced conjugates of cofactors under biomim-
etic conditions, and an examination of their reactivity with per-
oxides without potential interference from excess dithionite.
With dithionite, the reduced conjugate, in the case of LMB, has
to be extracted into an organic solvent prior to examining its
reaction with the artemisinin. Alternatively, in an aqueous
medium the dithionite may have to be decomposed by de-
creasing the pH, and then re-establishing neutral pH prior to
conducting the reactions of the of the artemisinins with the re-
duced conjugates.[5] NAD(P)H–flavin oxidoreductases do not
contain a flavin cofactor, but reduce endogenous flavins that
280
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 279 – 291

Antimalarial Peroxides
the FAD cofactor by NADPH.[4,30] Therefore, irrespective of the
direct use of flavin reductase as such by the parasite, it will
serve as a mimic of parasite flavin disulfide reductases under
biologically plausible conditions.
figure S3 Supporting Information). Treatment of this solution
with MB rapidly produced LMB and FAD. This confirms that
FADH2 in GR or TrxR is capable of reducing MB.[4] The LMB so-
lution obtained from NADPH–Fre–MB was treated with an
excess of each of artemisinin 1 and tetraoxane 4, resulting in
reappearance of the absorptions due to MB. Thus, both rapidly
oxidize LMB under physiological conditions. These experiments
are more easily performed than experiments that involve the
use of sodium dithionite to generate LMB by reduction of MB.
Results
Peroxides, NADPH–Fre–MB and NADPH–Fre–FAD–MB
Addition of MB to NADPH in the presence of E. coli flavin re-
ductase (Fre)[22] in aqueous buffer (pH 7.4) under argon result-
ed in rapid disappearance of the absorption at lmax 660 nm as
LMB was produced (Figure 2). Without Fre, partial reduction of
MB by NADPH to a steady-state concentration consisting of
~25% LMB took place; the extent of reduction is about the
same as that observed with NADH.[5] Thus, as previously re-
ported,[23,25] MB serves as a substrate for Fre. Treatment of the
NADPH–Fre–LMB solution with an excess (10 equiv) of each of
artemisinin 1 and tetraoxane 4 resulted in rapid reappearance
of the absorption at lmax 660 nm as LMB was oxidized to MB
(Figure 2 and figure S2 Supporting Information).
Peroxides, NADPH–Fre–FAD and BNAH–RF
FADH2 was generated from NADPH–Fre–FAD in aqueous buffer
at pH 7.4 under argon. The mixture was then treated directly
with an excess of each of artemisinin 1, artemisone 3, tetraox-
ane 4, and trioxolane 6 in pH 7.4 aqueous buffer/acetonitrile or
THF, and with artesunate 2 in the aqueous buffer alone. In all
cases, rapid re-oxidation of the FADH2 to FAD occurred
(Figure 4 and figure S4 Supporting Information). Whilst the
rates of oxidation of FADH2 by artemisinin and the tetraoxane
were similar, it is apparent that the trioxolane 6 more rapidly
oxidizes the reduced flavin than does artemisinin under the
same conditions. The significance of this observation is dis-
cussed below.
The suitability of Fre to serve as an in vitro mimic of the re-
duction of MB by flavoenzyme disulfide reductases was as-
sessed. A mixture of excess NADPH with Fre was treated with
FAD 11, resulting in rapid disappearance of FAD absorptions at
l 370 and 445 nm as FADH2 12 was generated (Figure 3 and
The tetraoxane 4 and trioxolane 6 were converted, by cata-
lytic RF 9 (0.2 equiv) and excess BNAH 13 in 1:1 aqueous
pH 7.4 buffer/CH3CN or buffer/
THF under argon[5] for 3 h, into
the ketones 16 and 17 (Figure 1)
from which they were prepared.
No other products were isolated.
A 1:1 mixture of artemisinin 1
and tetraoxane 4 (1 equiv each,
2 equiv total) with catalytic RF
(0.4 equiv)–BNAH (4.0 equiv) pro-
vided the two-electron reduction
products 14 and 15 from 1[5]
and the ketones 16 and 17 from
4 (Figure 1). The relative rates of
these reactions according to
1H NMR spectroscopic examina-
tion of the reaction[5] are similar
(Figure 5).
The tetraoxane 4 possesses
two peroxide units, and there-
fore its oxidizing capacity should
be twice that of artemisinin 1 or
trioxolane 6. A 1:1 mixture of 1
Figure 2. a) Reduction of MB by NADPH–Fre under Ar. MB (44 nmol) in pH 7.4 aqueous buffer at 228C was treated and 4 (2 equiv total) was treated
with Fre (4 nmol) and NADPH (230 nmol). Absorptions at lmax 660 nm were monitored every 18 s until reduction
with catalytic RF (0.2 equiv)–
to LMB had taken place (within 27 min). b) Partial reduction of MB by NADPH in the absence of Fre under Ar: MB
BNAH (2.0 equiv). If the two per-
(44 nmol) in pH 7.4 buffer at 228C was treated with NADPH (230 nmol). Absorptions at lmax 660 nm were moni-
oxide units in the tetraoxane
react with the RFH2 produced by
the reduction of RF in situ by
the BNAH, this will result overall
in the consumption of 2 equiv of
BNAH, and at least half of the te-
tored every 18 s until no further reduction was observed to occur (within 19 min) (initial and last traces shown).
c) Comparison of relative rates of reduction of MB by NADPH–Fre and NADPH under Ar under the foregoing con-
ditions obtained by monitoring decrease in absorbance of MB at lmax 660 nm. d) LMB solutions obtained accord-
ing to panel a) above from NADPH (110 nmol), Fre (8 nmol), and MB (64 nmol) are respectively treated with each
of artemisinin 1 (600 nmol, ~10 equiv) and tetraoxane 4 (601 nmol, ~10 equiv) in degassed CH3CN. Absorptions at
lmax 660 nm were monitored every 1.8 s until complete oxidation had taken place (within 16 min); rate of increase
of lmax at 660 nm is plotted vs. time.
ChemMedChem 2011, 6, 279 – 291
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
281

R. K. Haynes, Z.-H. Guo, D. Monti, et al.
MED
Discussion
The NADPH–Fre–FAD system may be used as a mimic of the
effect of artemisinin on NADPH consumption in the yeast GR
process.[5] Its use here greatly facilitates the generation of re-
duced flavin cofactors, and the results described herein con-
firm and extend those previously obtained by using sodium di-
thionite, and the model ascorbate–MB, BNAH–MB and –flavin
systems.[5] Reduction of MB by NADPH alone is likely to be too
slow to contribute effectively to redox cycling of MB (cf.
Figure 2), as was previously assumed.[5] As gauged by the ease
of reduction of MB to LMB by NADPH–Fre and by FADH2 gen-
erated from NADPH–Fre–FAD (Figure 2), MB will undergo rapid
redox cycling mediated by the flavoenzyme disulfide reductas-
es GR, TrxR, and others. This is in accord with the demonstra-
tion that MB acts as a subversive substrate for these reductas-
es, resulting in enhanced consumption of NADPH.[4,30] As arte-
misinins rapidly oxidize LMB, they will act in synergy with MB
by accelerating NADPH consumption.
Reductive cleavage via hydride transfer from the reduced
flavin to the peroxide bridge in the artemisinins has been dis-
cussed previously.[5] For the tetraoxane 4, initial hydride trans-
fer from the reduced flavin to the less-hindered oxygen of one
of the two peroxide units occurs (Scheme 1). As tetraoxane 4
is capable of using the two peroxide units, it is proposed that
intermediate 18 undergoes a second hydride-mediated cleav-
age. The resulting ketone hydrates then provide the ketones.
Hydride transfer to the peroxide in trioxolane 6, more facile
than in the case of tetraoxane 4 (see below), generates hemia-
cetal 19, which decomposes to the ketones. There is evidence
Figure 3. a) Reduction of FAD by NADPH–Fre under Ar. A mixture of NADPH
(314.2 nmol) and Fre (4 nmol) in aqueous buffer at pH 7.4 at 228C was treat-
ed with FAD (205 nmol) and absorptions at lmax 450 nm were monitored
every 18 s until reduction to FADH2 had taken place (within 28 min). b) Oxi-
dation of FADH2 by MB: the baseline was set to zero, and the foregoing so-
lution was treated with MB (154 nmol). Absorptions at 450 nm were moni-
tored every 30 s until maximum absorption was reached (within 9 min). Suf-
ficient MB was added to ensure that residual MB was apparent, thus verify-
ing that all NADPH had been consumed. c) The baseline was set to zero,
and the preceding mixture was treated with artemisinin (599 nmol) in
CH3CN. There resulted a rapid initial oxidation of the LMB; absorptions at
lmax 660 nm were monitored every 18 s until the absorbance reached the
maximum (within 28 min).
1
adduced from H NMR spectroscopy that the formation of pi-
peridone 17 is not synchronous with the formation of adaman-
tanone 16. The latter forms more slowly, presumably due to
slow collapse of the hydrate 16h under the neutral reaction
conditions. The alternative for all peroxides is that incipient for-
mation of an adduct takes place via reaction through C4a of
FADH2, which collapses to FAD and the reduction products.
This has been discussed for artemisinins, and the possibility
should be borne in mind when the peroxides act as antimalari-
al agents in vivo.[5] The catalytic cycling of reduced flavin,
either RFH2 in the BNAH–RF or FADH2 in the NADPH–Fre–FAD
experiments, results in net consumption of BNAH or NADPH as
the peroxide is reduced. Overall, the smooth reduction of tet-
raoxane 4 by LMB or FADH2 stands in notable contrast to the
inertness of tetraoxane 5 toward FeII.
troxane will remain unconsumed. This turns out to be the
case; the final mixture contains approximately 45% unreacted
1 and 64% unreacted 4 (Figure 5b). However, it was not possi-
ble to obtain reproducible rates of consumption of BNAH
through monitoring reactions by NMR spectroscopy. Further-
more, an examination of diluted reaction mixtures of BNAH–
RF–peroxide by UV spectroscopy was not possible because of
the very slow reaction. Therefore, the enzyme-catalyzed reac-
tions of each of artemisinin 1 and tetraoxane 4 in the presence
NADPH–Fre–FAD were examined. Mixtures of NADPH (5 equiv),
FAD (0.2 equiv), and 1 and 4 (each 1 equiv) in pH 7.4 buffer
were treated with Fre, and the decrease in absorption due to
NADPH at l 340 nm was monitored. The results are shown in
Figure 5c. It is apparent that the tetraoxane consumes approxi-
mately twice the amount of NADPH.
Conclusions
The results described herein and in our previous work suggest
that an appreciation of the mechanism of action of antimalarial
peroxides should realign with their logical abilities to oxidize
susceptible biomolecules. In spite of intensive investigations,
the identity of putative targets associated with the presumed
ferrous iron activation of artemisinins remains elusive. Iron in
its own right contributes to oxidative stress, and this is exacer-
bated within the malaria parasite with its catabolism of Hb.[31]
However, there arises the possibility that ferric iron oxidizes re-
282
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 279 – 291

Antimalarial Peroxides
Figure 4. Oxidation of FADH2 generated from NADPH–Fre–FAD by each of artemisinin 1, tetraoxane 4, and trioxolane 6. a) NADPH (397.5 nmol, 1.90 equiv)–
Fre (4 nmol) in aqueous buffer at pH 7.4 at 228C under Ar was treated with FAD (200 nmol) and absorptions at lmax 450 nm were monitored every 3 s until
absorptions due to FAD had decreased to a minimum. b) Artemisone 3 (600 nmol, 3 equiv) in CH3CN under Ar was added, resulting in rapid oxidation of the
FADH2 to FAD; absorptions at lmax 339 and 450 nm were monitored to convergence (~30 min). c) Oxidation of FADH2 by artemisinin 1: FAD (209 nmol) was
added to NADPH (397.5 nmol)–Fre (4 nmol) in pH 7.4 aqueous buffer and decay of FAD was followed by monitoring decrease in absorption at lmax 450 nm
every 3 s until the FAD absorption no longer decreased (0-33 min). The FADH2 solution was treated at 33 min with artemisinin 1 (599 nmol, 3 equiv) in CH3CN.
The rate of oxidation of FADH2 to FAD was monitored at lmax 450 nm until constant (~38 min). d) Oxidation of FADH2 by tetraoxane 4: the solution in panel c)
was treated at 33 min with tetraoxane (601 nmol, 3 equiv) in CH3CN and increase of absorption at lmax 450 nm was monitored until constant (~38 min).
e) Comparison of relative rates of oxidation of FADH2 by artemisinin 1 and tetraoxane 4 (each 3 equiv with respect to FAD). f) Oxidation of FADH2 by trioxo-
lane 6: FAD (200 nmol) was added to NADPH (258 nmol)–Fre (4 nmol) in pH 7.4 aqueous buffer and decay of FAD was followed by monitoring decrease in ab-
sorption at lmax 450 nm every 3 s until the FAD absorption no longer decreased (0–32 min). The FADH2 solution was treated at 32 min with artemisinin 1
(600 nmol, 3 equiv) or trioxolane 6 (600 nmol) in THF. The rate of oxidation of FADH2 to FAD was monitored at lmax 450 nm until constant (~38 min); the
curve for trioxolane 6 is shown. g) Comparison of relative rates of oxidation of FADH2 by artemisinin 1 and trioxolane 6 (each 3 equiv with respect to FAD) de-
scribed in panel f).
duced flavin cofactors such as FMNH2 and FADH2 generated by
Fre, GR, or other flavoenzymes in vivo.[23,32] The rate greatly ex-
ceeds that of FeIII oxidation of thiols, including reduced gluta-
thione GSH, in a biological medium.[23,32] Therefore, a link to
the peroxide cofactor model may be provided by FeIII through
its assistance of the antimalarial action of peroxides, not by re-
acting with them, but rather by accelerating turnover of flavin
cofactors which enhances the consumption of NADPH. Thus, in
ChemMedChem 2011, 6, 279 – 291
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
283

R. K. Haynes, Z.-H. Guo, D. Monti, et al.
MED
Figure 5. Comparison of reactions of artemisinin 1 and tetraoxane 4. a) 1 and 4 and BNAH–RF: aliquots were withdrawn at 1 min, then at 30 min, and at
30 min intervals thereafter from 1:1 CH3CN/buffer solution (pH 7.4) containing 1 (186 mmol) and 4 (186 mmol), RF 9 (74.4 mmol, 0.4 equiv), BNAH 13 (745 mmol,
1
4 equiv), and the internal standard 1,3,5-trimethoxybenzene (25.6 mmol) and quenched with 1m NaHSO4 prior to analysis by H NMR spectroscopy. In the
1H NMR spectrum signals at d=5.86 (H-12, 1), 2.96 (Me of -NSO2Et in 4) and 6.09 (standard) were used for calculating conversion. b) Experiment as in panel a)
was repeated with 1 (252 mmol), 4 (252 mmol), RF (50.4 mmol, 0.2 equiv), BNAH (504 mmol, 2 equiv), and 1,3,5-trimethoxybenzene (33 mmol). Unreacted 1 and
4 at 180 min are 45 and 64%, respectively. c) NADPH consumption by 1 and 4 in the NADPH–Fre–FAD system under Ar: a solution of Fre (4 nmol) and FAD
(41 nmol) in aqueous buffer at pH 7.4 was treated with NADPH (238.4 nmol) and then treated with 1 (40.4 nmol) and decay of absorption due to NADPH at
l 340 nm monitored each 3.9 s for 83 min. The experiment was also conducted with NADPH (236.3 nmol), Fre (4 nmol), and FAD (41 nmol) with tetraoxane 4
(40.4 nmol). The control experiment involved the use of Fre (4 nmol)–FAD (41 nmol)–NADPH (238.4 nmol) in aqueous buffer at pH 7.4. An initially rapid con-
sumption of NADPH (0–6 min) is followed by a slower decay. A slow background consumption of NADPH occurs for the control, and at a slightly faster rate
for experiments involving 1 and 4. Although this may be due to slow leakage of air into the cuvette throughout the experiment resulting in oxidation of
FADH2 and reduction of the resulting FAD by NADPH–Fre, it is notable that NADPH attrition in the NADPH–Fre system is greater after treatment with each of
1 and 4 (cf. Ref. [5]). d) Relative amounts of NADPH consumed by each of 1 and 4 in the foregoing experiment.
the summary of the cofactor model as presented in Scheme 2,
oxidation of reduced flavin cofactors by FeIII is presented for
consideration. Whilst there is debate over the origin of the
effect of labile-FeIII scavengers such as DFO in antagonizing an-
timalarial action of artemisinins in vitro,[10,21] the proposal is
reconcilable with such data. The summary in Scheme 2 is
based largely on incisive commentaries on effects of redox-
active antimalarial drugs such as MB on parasite redox homeo-
stasis.[4] Although it is not yet known if critical parasite redox
flavoenzymes other than GR, especially TrxR and lipoamide de-
hydrogenase, interact with antimalarial peroxides, these, as
gauged by interaction of with MB,[4] are likely to be affected.[5]
Given its functional relationship with Fre described here, it is
also apparent that the flavoenzyme ferredoxin–NADP+ reduc-
tase, which is present in the parasite apicoplast,[28] must also
be susceptible. Such susceptibility may underpin the enhanced
in vitro biological activities of artemisinins usually observed
toward apicomplexan parasites, such as Plasmodium, Toxoplas-
ma, and Babesia species as compared with activities toward
non-apicomplexan parasites.[33]
sulfide reductases more rapidly than tetraoxanes or artemisi-
nins. Thus, the resulting rapid incipient perturbation of redox
homeostasis may translate into a greater parasiticidal activity.
Although differences in permeability and competing, but
futile, decomposition pathways (such as with heme), may tend
to mask intrinsic differences in antimalarial activities between
artemisinins and trioxolanes, it is clear that the latter have
more pronounced in vitro activities than structurally related
tetraoxanes (cf. Figure 1). Nevertheless, the second peroxide
unit in the tetraoxane has the potential to confer an extra di-
mension to biological activity. However, the peroxide units are
incorporated within a six-membered ring, and may be less re-
active, because of lower ring strain and greater steric shielding,
than the peroxide within the five-membered trioxolane.[34] By
flanking the tetraoxane nucleus by carbocyclic ring systems
that are more strained and less hindered than those in 4, and
at the same time by using functional groups that improve per-
meability, it may be possible to enhance the ability of tetraox-
anes to perturb redox homeostasis and permit application to
other targets[35] as well.
Clearly in an intraparasitic environment, trioxolanes are capa-
ble of oxidizing the reduced cofactor FADH2 of flavoenzyme di-
284
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 279 – 291

Antimalarial Peroxides
Merck silica gel 60 (0.04–0.063 mm). UV/Vis spectra were recorded
1
on an Agilent 8453 instrument, and H and 13C NMR spectra were
obtained as solutions in CDCl3 on a Bruker AV 400 spectrometer
operating at 400 and 100 MHz, respectively. CDCl3 was used as
solvent unless otherwise stated. Melting points were carried out
on a Leica Hot Stage DME E compound Microscope and are cor-
rected. MS data were obtained on API QSTAR high-performance
triple quadrupole time-of-flight mass spectrometer with electro-
spray ionization, and on a Waters Micromass GCT premier ToF
high-resolution mass spectrometer (CI+, methane).
Tetraoxane 4: 1-(Ethanesulfonyl)piperidin-4-one 17. Ethanesulfonyl
chloride (1.7 mL, 18.6 mmol, 1.5 equiv) was added dropwise to a
stirred solution of 4-piperidone monohydrate HCl (1.9 g,
12.4 mmol) and K2CO3 (4.45 g, 32 mmol) in acetone (20 mL) and
H2O (10 mL) at 08C. The mixture was stirred overnight at room
temperature. A saturated solution of NaHCO3 (10 mL) was added
to quench the reaction. The mixture was extracted with CH2Cl2 (3ꢁ
15 mL) and dried over MgSO4. After filtration, the solvent was re-
moved by evaporation under reduced pressure to leave a colorless
crystalline residue that was recrystallized from CH2Cl2/hexanes to
give the product as colorless plates (2.0 g, 89%); mp: 68–698C;
1H NMR (400 MHz): d=1.39 (3H, t, J=7.6 Hz), 2.57 (4H, dd, J=6.4,
6.6 Hz), 3.05 (2H, q, J=7.6 Hz), 3.64 (4H, dd, J=6.0. 6.0 Hz).[36]
Ethereal H2O2. An aqueous solution of H2O2 (30%, 62 mL) was satu-
rated with NaCl and extracted with Et2O (25 mL) in four portions
(4ꢁ15 mL). The organic layer was separated followed by drying
over MgSO4 and filtered. The ethereal H2O2 was stored in a refriger-
ator at +48C.[37]
Tetraoxane 4. According to a published procedure for preparing
gem-dihydroperoxides,[37] ethereal H2O2 (1.44m, 7 mL, 10 equiv)
was added to a stirred mixture of 1-(ethylsulfonyl)piperidin-4-one
(383 mg, 2 mmol) and phosphomolybdic acid (73 mg, 2 mol%),
and the reaction mixture was stirred for 4 h at room temperature.
Et2O (15 mL) and H2O (5 mL) were added to dilute the reaction
mixture. The organic layer was separated and the aqueous solution
was extracted with EtOAc (3ꢁ10 mL). The combined organic layer
was washed with brine (3ꢁ8 mL) and then dried over MgSO4. The
solution was filtered and concentrated by evaporation under re-
duced pressure. The residue consisting of the crude gem-dihydro-
peroxide was dissolved in CH2Cl2 (2 mL) and CH3CN (1 mL) and
then added to a stirred solution of 2-adamantanone 16 (502 mg,
3.34 mmol) and rhenium(VII) oxide (32.4 mg, 0.0669 mmol,
3 mol%) in CH2Cl2 (3 mL).[38] The reaction mixture was stirred over-
night at room temperature. The solution was concentrated by
evaporation under reduced pressure, and H2O (10 mL) was added.
The mixture was extracted with CH2Cl2 (3ꢁ8 mL). The combined
organic layer was washed with brine (3ꢁ8 mL) followed by drying
over MgSO4. After filtration, the solution was concentrated by
evaporation under reduced pressure to leave the crude product as
a gum. This was submitted to column chromatography with
EtOAc/hexane 1:4 to give the tetraoxane as a fine white microcrys-
talline solid (376 mg, 50%); mp: 123–1248C; 1H NMR (400 MHz):
d=1.37 (3H, t, J=7.6 Hz), 1.56–2.04 (14H, m), 2.54 (2H, s), 2.96
(2H, q, J=7.6 Hz), 3.40 (4H, s); MS (ESI) calcd for C17H27NO6SNa+
396.1457, found 396.1482. The data are in agreement with those
reported in the literature.[14]
Scheme 1. Reduction of peroxides by reduced flavins at pH 7.4. a) Tet-
raoxane 4 is proposed to undergo incipient cleavage by hydride transfer
from reduced flavin to provide peroxy-hemiacetal 18, which undergoes a
second hydride-mediated cleavage. The alternative is for 18 to collapse via
extrusion of a ketone to a hydroperoxy-hemiacetal that may likewise be re-
duced. b) Hydride cleavage of trioxolane 6 is proposed to generate hemiace-
tal 19, which decomposes to the ketones. The latter reaction may proceed
via hydrate 16h of 2-adamantanone. The oxidation of two equivalents of re-
duced flavin by the tetroxane requires consumption of two reducing equiva-
lents of BNAH or NADPH.
Experimental Section
Materials and Methods
Artemisinin 1 was supplied by the Kunming Pharmaceutical Corpo-
ration (China), artesunate 2 was supplied by Dr. Robert Carter
(Abbott–Knoll AG, Basel, Switzerland), and artemisone 3 was pre-
pared according to a published procedure.[1] 2-Adamantanone
(99%, Aldrich), methylene blue (certified, BSC, Sigma–Aldrich), nico-
tinamide (Aldrich, ꢀ99%), riboflavin (98%, Acros), FAD (ꢀ95%,
Sigma), NADPH (~90%, Fluka) were used as received. N-Benzyl-1,4-
dihydronicotinamide (BNAH) was prepared as previously de-
scribed.[5] All reactions were carried out under an atmosphere of air
or Ar as indicated. Pentane, CH2Cl2, EtOAc, and hexane were dis-
tilled and dried prior to use. TLC was performed with Merck Kiesel-
gel 60 F254 plates and visualized with UV light (l 254 nm) and/or
heating after treatment with 5% ammonium molybdate in 10%
concentrated H2SO4. Column chromatography was performed with
Trioxolane 6: O-methyl-2-adamantanone oxime. A mixture of 2-ada-
mantanone 16 (2.03 g, 13.52 mmol) and methoxylamine HCl
(1.23 g, 14.71 mmol) in EtOH (30 mL) and pyridine (1.18 mL,
14.66 mmol) was heated at gentle reflux at 908C (bath tempera-
ture) under N2 for 1.5 h. The mixture was cooled to room tempera-
ChemMedChem 2011, 6, 279 – 291
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
285

R. K. Haynes, Z.-H. Guo, D. Monti, et al.
MED
CAT GGG Aat gac aac ctt aag ctg
taa ag-5’ and 3’-ACC GCT CGA Gga
taa atg caa acg cat cgc c-5’ as pri-
mers. The PCR product was digest-
ed by restriction enzymes and in-
serted into the pET-28a(+) vector
(Novagen) between the NcoI and
XhoI sites. The Fre gene was con-
firmed not to contain any spurious
mutations by full-length sequenc-
ing of the cloned DNA insert. The
gene product was expressed in
Luria broth containing 0.2 mm
IPTG at 168C for 24 h. Fre was pu-
rified from the crude extract pre-
pared from the harvested cells,
first by metal chelating chroma-
tography using a 5 mL HiTrap Che-
lating HP column (GE Healthcare)
and then by gel filtration using a
Sephacryl S-100 column (GE
Healthcare). It was established by
SDS-PAGE that the tagged protein
so obtained was purified to
Scheme 2. Proposed basis for oxidative stress cascade mediated by peroxidic antimalarials, as derived from com-
mentary of redox drug action in Ref. [4] (GR=glutathione reductase, ox=FAD cofactor, red=FADH2 cofactor;
Trx =thioredoxin, SH=reduced, S2 =oxidized; TrxR=thioredoxin reductase, ox=FAD, red=FADH2; Grx=glutare-
doxin): Maintenance of GSH and reduced thioredoxin (Trx) requires NADPH provided by the hexose monophos-
phate shunt (HMS), the rate-limiting enzyme of which is glucose-6-phosphate dehydrogenase (G6PD). HMS activi-
ty in P. falciparum-infected erythrocytes (IE) is up to 78-fold higher than that of uninfected erythrocytes; HMS ac-
tivity of the parasite contributes 80% of the total observed in the intact IE, and HMS activity of the host cell is in-
creased ~24-fold when compared with uninfected erythrocytes. The major portion of parasite HMS activity and
the increased HMS activity of the host cell counteract oxidative stress. Treatment of yeast GR, a surrogate for para-
site GR, with artemisinin 1 under aerobic conditions results in enhanced consumption of NADPH and restricted re- >95%. The purified protein was
duction of GSSG. That is, FADH2 becomes susceptible to autoxidation after exposure of GR to the artemisinin. The
putative effect of FeIII is proposed as it rapidly oxidizes FMNH2 and FADH2 generated by NADPH–Fre, NADPH–GR,
or NADPH–flavin diaphorases (Refs. [23,32]). Possibly, exposure to artemisinin may render reduced flavin cofactor
in GR amenable to oxidation by FeIII. It is not yet known if other critical parasite redox flavoenzymes such as TrxR,
lipoamide dehydrogenase, and the ferredoxin–NADP+ reductase occurring in the parasite apicoplast (Ref. [28]), in-
teract with antimalarial peroxides, but these are predicted to be affected.
quantified by a Coomassie Blue
protein assay kit (Pierce) and
stored
in
phosphate-buffered
saline (pH 7.4) containing 10%
glycerol at À208C until use.
1
ture, treated with brine (20 mL), and then extracted with CH2Cl2
(3ꢁ15 mL). The extracts were combined, and the solution was
dried over MgSO4 and then filtered. The solvent was evaporated
under reduced pressure to leave the O-methyl oxime as a white
solid; mp: 71.3–71.98C (lit.[39] 70–718C) (2.227 g, 92%); 1H NMR
(400 MHz): d=1.79–1.98 (12H, m), 2.54 (1H, s), 3.46 (1H, s), 3.81
(3H, s). The sample was dried thoroughly under high vacuum.
UV/Vis absorption, H NMR spectroscopy, and decomposition
experiments
Reduction of MB by NADPH–Fre, and reaction of LMB with artemi-
sinin 1 and tetraoxane 6 (Figure 2 and figure S2 Supporting Infor-
mation): NADPH (83.9 mg, 0.1007 mmol) was dissolved in degassed
pH 7.4 buffer (5 mL) to generate a solution containing 0.0201m.
The NADPH solution was purged with Ar and was then stored at
+48C for use. The NADPH solution (10 mL) was treated with pH 7.4
degassed buffer (1.6 mL) in an Ar-purged UV cuvette (d=1 cm) at
228C. The absorbance of NADPH measured at lmax 339 nm was
0.4949 AU. As the absorption coefficient of NADPH at l 339 nm is
Trioxolane 6. For the following preparation to succeed, all reagents
and solvents had to be scrupulously dry. Ozone was produced by
using a Triogen Lab 2B Laboratory O3 generator at a flow rate of
0.6 LminÀ1 O2 at a discharge voltage of 50 V and bubbled through
6220mÀ1 cmÀ1 [41]
the actual concentration of NADPH solution was
,
a
stirred solution of dried O-methyl-2-adamantanone oxime
(1.08 g, 6.0 mmol) and dried 1-(ethylsulfonyl)piperidin-4-one
(1.15 g, 6.0 mmol) in dry distilled pentane (30 mL) and dry distilled
CH2Cl2 (15 mL) at À788C for 2.5 h. After completion of the reaction,
the solution was flushed with O2 for 10 min and then warmed to
room temperature. It was then concentrated by evaporation under
reduced pressure at room temperature. The crude product was
submitted to column chromatography with EtOAc/hexane 1:6 to
give the product (1.223 g, 57.5%), as colorless rods; mp: 108.3–
0.0128m (~0.064 mmol). MB (12.8 mg, 0.0342 mmol) was dissolved
in degassed pH 7.4 buffer (5 mL) to generate a solution containing
0.00685m; the MB solution was purged under Ar and was stored
at +48C for use.
a. Reduction of MB by NADPH–Fre: The MB solution (10 mL) was
treated with pH 7.4 degassed buffer (1.9 mL) in an Ar-purged UV
cuvette (d=1 cm) at 228C. The absorbance of MB at lmax 660 nm
was 1.63 AU. The absorption coefficient of MB at lmax 660 nm is
1
108.48C (lit.[16] 110–1128C); H NMR (400 MHz): d=1.37 (3H, t, J=
71547mÀ1 cmÀ1 [42]
therefore, the actual concentration of the MB
;
7.6), 1.69–2.00 (18H, m), 2.93–2.30 (2H, q, J=7.6), 3.34–3.40 (2H,
m), 3.45–3.51 (2H, m); MS (ESI) calcd for C17H27NO6SNa+ 380.1508,
found 380.1528.
solution is 0.0044m (~0.022 mmol). Fre (200 mL, 400 mm) was dilut-
ed to 1 mL by additional of degassed pH 7.4 aqueous buffer to
generate a solution containing 80 mm. The purged MB solution
(10 mL, 0.044 mmol) and degassed pH 7.4 buffer (1.85 mL) were
added to an Ar-purged UV cuvette (d=1 cm) at 228C with moni-
toring of the absorption at lmax 660 nm until maximized. The MB
solution in the cuvette was treated with Fre (50 mL, 4 nmol) and
the NADPH solution (18 mL, 230.4 nmol). The reduction of MB was
Preparation of E. coli flavin reductase Fre: The E. coli flavin reduc-
tase Fre was overexpressed with a C-terminal hexahistidine tag
and purified to homogeneity as described previously.[40] Briefly, the
Fre gene was amplified by Phusion High-Fidelity DNA Polymerase
(Finnzymes) from the genomic DNA of E. coli K12 using 3’-ACA TGC
286
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 279 – 291

Antimalarial Peroxides
followed by monitoring decrease of absorption at lmax 660 nm
until complete reduction to LMB had taken place.
Reduction of MB by NADPH–Fre–FAD, and reaction of LMB with
artemisinin 1 and tetraoxane 4 (Figure 3 and figure S3): NADPH
(17.2 mg, 0.0206 mmol) was dissolved in degassed pH 7.4 buffer
(1 mL) to generate a solution containing 0.0206m. The solution
was purged with Ar and was stored at +48C for use. The NADPH
solution (10 mL) was treated with pH 7.4 degassed buffer (1.9 mL)
in an Ar-purged UV cuvette (d=1 cm) at 228C. The absorbance of
the NADPH solution at lmax 339 nm was measured as 0.445 AU. The
absorption coefficient of NADPH at l 339 nm of 6220mÀ1 cmÀ1[41]
gives an actual concentration of 0.0137m (~0.01366 mmol). MB
(12.8 mg, 0.0342 mmol) was dissolved in degassed pH 7.4 buffer
(5 mL) to generate a solution containing 0.00685m. The MB solu-
tion was purged with Ar and was stored at +48C for use. The MB
solution (10 mL) was treated with pH 7.4 degassed buffer (1.9 mL)
in an Ar-purged UV cuvette (d=1 cm) at 228C. The absorbance of
the MB solution at lmax 660 was measured as 1.63 AU. The absorp-
b. Control: reduction of MB by NADPH: The MB solution in the cuv-
ette was treated with the NADPH solution (18 mL, 230.4 nmol) only.
The reduction of MB was followed by monitoring decrease of ab-
sorption at lmax 660 nm until no further reduction was observed to
occur.
c. Comparison of NADPH–Fre and NADPH control reductions of MB:
The relative rates of reduction of MB in the presence of NADPH–
Fre and in the absence of Fre under the foregoing conditions are
compared by monitoring relative rates of decrease of lmax at
660 nm.
tion coefficient of MB at l 660 nm is 71547mÀ1 cmÀ1 [42]
therefore,
;
d. Comparison of rates of oxidation of LMB by each of artemisinin 1
and tetraoxane 4: For the following experiment, the stock solutions
were prepared as follows: Artemisinin stock solution (10 mL in
CH3CN), Mr =282.33 Da, 0.5637 g, 1.997ꢁ10À3 mol, concentration:
1.997ꢁ10À3 mol per 0.01 L=0.1997m; tetraoxane stock solution
(5 mL in CH3CN), Mr =373.44 Da, 0.1869 g, 5.005ꢁ10À4 mol, con-
centration: 5.005ꢁ10À4 mol per 0.005 L=0.1001m; NADPH stock
solution (1 mL in pH 7.4 phosphate buffer), absorbance: 0.445, e=
6220mÀ1 cmÀ1 at l 339 nm, concentration in cuvette (10 mL made
the actual concentration of MB solution is 0.0044m (~0.022 mmol).
Fre (400 mL, 1000 mm) was diluted to 1 mL by addition of degassed
pH 7.4 aqueous buffer to generate a solution containing 400 mm
(0.4 mmol). FAD (17.4 mg, 0.0205 mmol) was dissolved in degassed
pH 7.4 buffer (1 mL) to generate a solution containing 0.0205m.
The FAD solution was purged under Ar and was stored at +48C
for use.
6220mÀ1 cmÀ1)=
a. Reduction of FAD by NADPH–Fre: The purged NADPH solution
(23 mL, 314.2 nmol), Fre (10 mL, 4 nmol) and degassed pH 7.4 buffer
(1.9 mL) were added to an Ar-purged UV cuvette (d=1 cm) at
228C and treated with the FAD solution (10 mL, 205 nmol). The re-
duction of FAD was followed by monitoring the decrease of ab-
sorption at lmax 450 nm until reduction had taken place.
up to 1.91 mL in buffer)=0.445/(1 cm
ꢁ
0.0715 mm, concentration 7.15ꢁ10À5 m ꢁ 0.00191 L/10ꢁ10À6 L=
0.0137m; MB stock solution (5 mL in pH 7.4 phosphate buffer), ab-
sorbance: 1.57, e=71547mÀ1 cmÀ1 at l 660 nm, concentration in
cuvette (20 mL made up to 2.92 mL in buffer)=1.57/(1 cm ꢁ
71547mÀ1 cmÀ1)=0.0193 mm, concentration: 1.93
ꢁ m ꢁ
10À5
0.00161 L/5ꢁ10À6 L=0.0032m, Fre: 80 mm in pH 7.4 buffer. Artemi-
sinin 1 (0.5637 g, 1.997ꢁ10À3 mol) was dissolved in degassed
CH3CN (10 mL) to give a final concentration of 0.1997m. The MB
solution was prepared from MB (7.2 mg, 0.0193 mmol) and de-
gassed pH 7.4 aqueous buffer (5 mL) to give a final concentration
of 0.00385m. The tetraoxane 4 (186.9 mg, 0.5005 mmol) and de-
gassed CH3CN (5 mL) were mixed to give a final concentration of
0.1001m. By measuring UV absorption at l 660 nm, the actual con-
centration was calculated to be 0.0032m as described above.
NADPH (17.2 mg, 0.0206 mmol) was dissolved in degassed pH 7.4
aqueous buffer (1 mL) to give a final concentration of 0.0206m; as
above, the actual concentration by measurement of UV absorption
at l 339 nm was determined to be 0.0137m. The Fre was prepared
in pH 7 buffer to give a final concentration of 400 mm. The de-
gassed pH 7.4 buffer (2900 mL) and MB solution (20 mL, 0.064 mmol)
were added to an Ar-purged UV cuvette (d=1 cm) at 228C. Once
the solution had equilibrated to maximum absorbance at
l 660 nm, the solution was then treated with Fre (20 mL, 8 nmol)
and the baseline was set to zero. The NADPH solution (8 mL,
110 nmol, 1.7 equiv) was added with mixing. Three experiments
were carried out: 1. Control, which measured rate of decrease of
MB until the reduction was complete. 2. Measuring the effect of
addition of artemisinin (3 mL, 0.6000 mmol, 10 equiv) to the same
cuvette on the rate of increase of MB absorbance; scans were re-
corded at intervals of 1.8 s until convergence. 3. Measuring the
effect of addition of tetraoxane (6 mL, 600.6 nmol, 10 equiv) to the
LMB solution, and scanning every 1.8 s until convergence was
reached. In both cases inhomogeneity of the mixture, probably
due to formation of fine particles of MB, took place after addition
of artemisinin or tetraoxane in the CH3CN, and this affected the
quality of the spectra recorded. In Figure 2d, the relative differen-
ces in rates oxidation of the LMB by each of artemisinin 1 and tet-
raoxane 4 are recorded.
b. Reduction of MB by FADH2: The above mixture was treated with
the purged MB solution (30 mL, 131 nmol) by monitoring the in-
crease of absorption at lmax 450 nm until the oxidation of FAD had
taken place; the appearance of an absorption at lmax 660 nm indi-
cated that there was no residual NADPH capable of initiating Fre-
catalyzed reduction of MB.
c. Oxidation of LMB by artemisinin: The baseline was recorded, and
the reaction mixture in the cuvette was treated with the purged
artemisinin solution (3 mL, 0.599 mmol) prepared from artemisinin
(563.7 mg, 1.9973 mmol) and degassed CH3CN (10 mL) under Ar.
The initial oxidation of the LMB was very fast, and it was not possi-
ble to record incipient absorption changes. Recording of spectra
was continued until the maximum was reached (~28 min). It was
noted that inhomogeneity of the mixture, probably due to forma-
tion of fine particles of MB, resulted after addition of artemisinin in
the CH3CN, and this affected the quality of the spectra recorded.
Figure S3: The tetraoxane solution was prepared from tetraoxane 4
(186.9 mg, 0.5005 mmol) and degassed CH3CN (5 mL) to give a
final concentration of 0.1001m. The methylene blue (MB) solution
was prepared from MB (12.8 mg, 0.0342 mmol) and degassed
pH 7.4 aqueous buffer (5 mL) to give a final concentration of
0.00385m. The concentration of MB as 0.00435m was re-deter-
mined by UV absorption at l 660 nm as described above. The
NADPH solution was prepared from NADPH (17.2 mg,
0.0206 mmol) and degassed pH 7.4 aqueous buffer (1 mL) to give a
final concentration of 0.0206m. The concentration of NADPH as
0.0137m was re-determined by UV absorption at l 339 nm as de-
scribed above. Fre was prepared in pH 7 buffer to give a final con-
centration of 400 mm. The degassed pH 7.4 buffer (2900 mL), the
NADPH solution (23 mL, 0.315 mmol, 1.51 equiv), were added to a
UV cuvette (d=1 cm) at 228C and treated with Fre (10 mL, 4 nmol).
ChemMedChem 2011, 6, 279 – 291
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
287

R. K. Haynes, Z.-H. Guo, D. Monti, et al.
MED
A background was scanned from l 200–800 nm. The FAD solution
(10 mL, 0.2088 mmol) was added with mixing. Three experiments
were carried out: 1. Control, which measured rate of decrease of
FAD until the reaction was finished. 2. Effect of immediate addition
of MB solution (30 mL, 0.131 mmol, 0.623 equiv) to the same cuvette
on the rate of increase of FAD. 3. Effect of addition of tetraoxane
(6 mL, 601 nmol, 4.58 equiv) to the same cuvette on the rate of in-
crease of MB. Scanning was carried out every 18 s until the conver-
gence was reached.
at l 339 nm. The FAD solution was prepared from FAD (17.7 mg,
2.088ꢁ10À5 mol) and degassed pH 7.4 aqueous buffer (1 mL) to
give a final concentration of 0.0200m. The Fre was prepared in
pH 7 buffer to give a final concentration of 400 mm. The degassed
pH 7.4 buffer (2000 mL), the NADPH solution (25 mL, 0.3975 mmol,
1.90 equiv), were added to a UV cuvette (d=1 cm) at ambient tem-
perature and treated with the Fre (10 mL, 4 nmol). A background
was scanned from l 220 to 600 nm. The FAD solution (10 mL,
0.200 mmol) was added with mixing, and the decrease of FAD at
lmax 450 nm were recorded at 3 s intervals until the absorption was
constant (33 min). The tetraoxane solution (6 mL, 0.6006 mmol,
3 equiv) was added, and the rate of increase of FAD absorption at
lmax 450 nm was monitored for 38 min. Panel e) plots the relative
rates of oxidation of FADH2 by artemisinin 1 and tetraoxane 4 by
measuring increase absorption at lmax 450 nm.
Reduction of FAD by NADPH–Fre, and reaction of FADH2 with ar-
temisinins 1 and 3, tetraoxane 4, and trioxolane 6 (Figure 4 and
figure S4): The NADPH solution was prepared from NADPH
(83.9 mg, 1.007ꢁ10À4 mol) and degassed pH 7.4 aqueous buffer
(5 mL) to give a final concentration of 0.02014m. Concentration of
NADPH 0.0128m was re-determined by UV absorption at l 339 nm
as described above. The FAD solution was prepared from FAD
(17.7 mg, 2.088ꢁ10À5 mol) and degassed pH 7.4 aqueous buffer
(1 mL) to give a final concentration of 0.0200m. The Fre solution
was prepared in pH 7 buffer to give a final concentration of
400 mm. The degassed pH 7.4 buffer (2000 mL), the NADPH solution
(25 mL, 0.3975 mmol, 1.90 equiv) were added to a UV cuvette (d=
1 cm) at ambient temperature and treated with the Fre (10 mL,
4 nmol). A background was scanned from l 220 to 600 nm.
f and g. Comparison of oxidation of FADH2 by artemisinin 1 and tri-
oxolane 6: Artemisinin 1 (141.2 mg, 5.001ꢁ10À4 mol) was dissolved
in degassed THF (5 mL) to give a final concentration of 1.997m.
The NADPH solution was prepared from NADPH (83.9 mg, 1.007ꢁ
10À4 mol) and degassed pH 7.4 aqueous buffer (5 mL) to give a
final concentration of 0.02014m. The concentration of NADPH as
0.0172m was re-determined by measurement of the UV absorption
at l 339 nm. FAD (17.4 mg, 2.05ꢁ10À5 mol) was dissolved in pH 7.4
aqueous buffer (1 mL) to give a final concentration of 0.0205m.
The Fre was prepared in pH 7 buffer to give a final concentration
of 400 mm. The degassed pH 7.4 buffer (2000 mL), the NADPH solu-
tion (15 mL, 0.258 mmol, 1.26 equiv), were each added to a UV cuv-
ette (d=1 cm) at ambient temperature and treated with Fre
(10 mL, 4 nmol). A background was scanned from l 220 to 600 nm.
The FAD solution (10 mL, 0.205 mmol, 1.0 equiv) was added with
mixing. Two experiments were carried out: 1. Control, which mea-
sured rate of decrease of FAD until the reaction was finished; scan-
ning at lmax 450 nm was conducted at 3 s intervals until absorption
due to FAD no longer decreased. 2. Addition of artemisinin solu-
tion (6 mL, 0.60 mmol, 3 equiv) to the same cuvette on the rate of
increase of FAD. The rate of turnover of FAD was followed through
monitoring increase of absorption at lmax 450 nm over 38 min.
a. Reduction of FAD by Fre: The FAD solution (10 mL, 0.200 mmol)
was added to the cuvette with mixing. Decrease of absorptions
due to FAD was monitored by scanning at lmax 450 nm at 3 s inter-
vals until reduction was completed.
b. Oxidation of FADH2 by artemisone 3: The artemisone solution
was prepared from artemisone (0.1607 g, 4.002ꢁ10À4 mol) and de-
gassed CH3CN (10 mL) to give a final concentration of 0.0400m.
The effect of addition of the artemisone solution (15 mL,
0.600 mmol, 3 equiv) to the same cuvette on the rate of oxidation
of FADH2 was monitored lmax 450 nm over 30 min.
c. Oxidation of FADH2 by artemisinin 1: A solution was prepared
from 1 (563.7 mg, 1.997ꢁ10À3 mol) and degassed CH3CN (10 mL)
to give a final concentration of 1.997m. The NADPH solution was
prepared from NADPH (83.9 mg, 1.007ꢁ10À4 mol) and degassed
pH 7.4 aqueous buffer (5 mL) to give a final concentration of
0.02014m. The concentration of NADPH as 0.0128m was re-deter-
mined from the UV absorption at l 339 nm, as described above.
The FAD solution was prepared from FAD (17.7 mg, 2.088ꢁ10À5
mol) and degassed pH 7.4 aqueous buffer (1 mL) to give a final
concentration of 0.0200m. The Fre was prepared in pH 7.4 buffer
to give a final concentration of 400 mm. The degassed pH 7.4
buffer (2000 mL), the NADPH solution (25 mL, 0.3975 mmol,
1.90 equiv) were added to an Ar-purged UV cuvette (d=1 cm) at
ambient temperature and treated with Fre (10 mL, 4 nmol). A back-
ground scan was taken from l 220 to 600 nm. The FAD solution
(10 mL, 0.209 mmol) was added with mixing, and the decrease of
FAD absorption at lmax 450 nm were recorded at 3 s intervals until
the absorption was constant (33 min). Then artemisinin solution
(3 mL, 0.5991 mmol, 3 equiv) was added, and the rate of increase of
FAD by monitoring absorption at lmax 450 nm was carried out over
38 min.
The trioxolane solution was prepared from 6 (0.0357 g, 9.987ꢁ10À5
mol) and degassed THF (1 mL) to give a final concentration of
0.100m. The NADPH solution was prepared from NADPH (0.0839 g,
1.007ꢁ10À4 mol) and degassed pH 7.4 aqueous buffer (5 mL) to
give a final concentration of 0.02014m. The concentration of
NADPH as 0.0172m was re-determined by measurement of the UV
absorption at l 339 nm. FAD (0.0174 g, 2.05ꢁ10À5 mol) was dis-
solved in degassed pH 7.4 aqueous buffer (1 mL) to give a final
concentration of 0.0205m. Fre was prepared in pH 7 buffer to give
a final concentration of 400 mm. The degassed pH 7.4 buffer
(2000 mL), the NADPH solution (15 mL, 0.258 mmol, 1.26 equiv),
were each added to a UV cuvette (d=1 cm) at ambient tempera-
ture and treated with Fre (10 mL, 4 nmol). A background was
scanned from l 220 to 600 nm. The FAD solution (10 mL,
0.205 mmol, 1.0 equiv) was added with mixing. Two experiments
were carried out: 1. Control, which measured rate of decrease of
FAD until the reaction was finished; scanning at lmax 450 nm was
conducted at 3 s intervals until the absorption due to FAD no
longer decreased. 2. Addition of trioxolane solution (6 mL,
0.60 mmol, 3 equiv) and examining effect on the rate of increase of
FAD. The rate of turnover of FAD was followed by monitoring in-
crease of absorption at lmax 450 nm over 70 min.
d and e. Comparison of oxidation of FADH2 by artemisinin 1 and tet-
raoxane 4: Compound 4 (186.7 mg, 4.999ꢁ10À4 mol) was dissolved
in degassed CH3CN (5 mL) to give a final concentration of 1.997m.
The NADPH solution was prepared from NADPH (83.9 mg, 1.007ꢁ
10À4 mol) and degassed pH 7.4 aqueous buffer (5 mL) to give a
final concentration of 0.02014m. The concentration of NADPH as
0.0128m was re-determined by measurement of the UV absorption
Reaction of artemisinin 1 and tetraoxane 4 with BNAH and
NADPH–Fre–FAD (Figure 5): a. 1. BNAH–RF catalytic system. Artemi-
sinin 1 (52.6 mg, 0.186 mmol), tetraoxane 4 (69.5 mg, 0.186 mmol),
288
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 279 – 291

Antimalarial Peroxides
riboflavin 9 (28.0 mg, 0.0744 mmol, 0.20 equiv with respect to total
peroxide), BNAH 13 (159.7 mg, 0.745 mmol, 2.0 equiv with respect
to total peroxide), and 1,3,5-trimethoxybenzene (4.3 mg,
0.0256 mmol) were stirred in a 1:1 mixture of CH3CN/phosphate
buffer (5.5 mL) under Ar at room temperature. At the first minute,
0.5 mL of the reaction mixture was withdrawn and added to 1m
NaHSO4 solution (1 mL) immediately. Et2O (1 mL) was added to
dilute the reaction mixture. The organic layer was separated and
the aqueous layer was extracted with EtOAc (3ꢁ1 mL). The com-
bined organic layer was washed with 1m NaHSO4 solution (2 mL)
followed by brine (1 mL), dried over MgSO4, filtered and concen-
trated under reduced pressure. The residue was taken into CDCl3,
containing 40 mm. The purged FAD solution (10 mL, 41 nmol), Fre
(10 mL, 4 nmol) and degassed pH 7.4 buffer (2 mL) were added to
an Ar-purged UV cuvette (d=1 cm) at 228C. The reaction mixture
was treated with NADPH (17 mL, 236.3 nmol). The oxidation of
NADPH was followed by monitoring decrease of absorption at
lmax 340 nm until oxidation had taken place. 100 mL of a solution
prepared from tetraoxane 4 (186.9 mg, 0.5005 mmol) in degassed
CH3CN (5 mL) was diluted to 5 mL with degassed CH3CN in a 5 mL
volumetric flask. 2 mL of this solution, containing 40.44 nmol tet-
raoxane 4 was added to the cuvette, and the decrease in concen-
tration of the NADPH was followed by monitoring decrease of ab-
sorption at lmax 340 nm. For the control experiment, pH 7.4 buffer
(2 mL) was added to the cuvette, and absorption at lmax 340 nm
was monitored. For the comparison with artemisinin, NADPH
(17.2 mg, 20.6 mmol) was dissolved in degassed pH 7.4 buffer
(1 mL) to generate a solution containing 0.0206m. The solution
was stored under Ar at +48C prior to use. The NADPH solution
(10 mL) was treated with pH 7.4 degassed buffer (2.0 mL) in an Ar-
purged UV cuvette (d=1 cm) at 228C. The absorbance of NADPH
at lmax 339 was 0.459 AU. The absorption coefficient of NADPH at
l 339 nm is 6220mÀ1 cmÀ1; therefore, the actual concentration of
the NADPH solution was 0.0148m (~0.0148 mmol). FAD (17.4 mg,
20.5 mmol) was dissolved in degassed pH 7.4 buffer (1 mL) to gen-
erate a solution containing 0.0205m. Fre (10 mL, 400 mm) was dilut-
ed to 100 mL by the addition of degassed pH 7.4 aqueous buffer to
generate a solution containing 40 mm. The purged FAD solution
(10 mL, 41 nmol), Fre (10 mL, 4 nmol) and degassed pH 7.4 buffer
(2 mL) were added to an Ar-purged UV cuvette (d=1 cm) at 228C.
The mixture was treated with NADPH (17 mL, 236.3 mmol). The oxi-
dation of NADPH was followed by monitoring decrease of absorp-
tion at lmax 340 nm. Artemisinin (57 mg, 202 mmol) was dissolved in
degassed CH3CN (10 mL). 2 mL of this solution, containing
40.4 nmol artemisinin was added to the cuvette, and the decrease
of absorption due to NADPH at lmax 340 nm was monitored.
1
and the H NMR spectrum was obtained. Steps were repeated at
30, 60, 90, 120, 150, and 180 min. Peak of internal standard taken
for calibration: 6.09 ppm, integral calibrated to 3; peak of 1 taken
for calculation d=5.864 ppm (singlet, 1H); peak of tetraoxane
taken for calculation: d=2.96 ppm (quartet, 2H). Results were cal-
culated as follows: no. of mol 1 added=0.0526 g/282.33 gmolÀ1
1.863ꢁ10À4 mol; no. of mol of
added=0.0695 g/
=
4
373.44 gmolÀ1 =1.861ꢁ10À4 mol;[1] =[no. of mol of 1 calculated/
no. of mol of 1 at 0 min ꢁ 1.863ꢁ10À4 mol]/0.0055 L;[4] =[no. of
mol of 4 calculated/no. of mol of 4 at 0 min ꢁ 1.861ꢁ10À4 mol]/
0.0055 L. Data are plotted in Figure 5.
b. Competitive reaction of artemisinin 1 and tetraoxane 4 with RFH2
generated in the RF–BNAH catalytic system. Artemisinin (71.1 mg,
252 mmol), tetraoxane (94.1 mg, 252 mmol), riboflavin (19 mg,
50.4 mmol, 0.2 equiv), BNAH (108 mg, 504 mmol, 2 equiv) and 1,3,5-
trimethoxybenzene (5.5 mg, 32.7 mmol) were stirred in a 1:1 mix-
ture of CH3CN/phosphate buffer (7.5 mL) under Ar at room temper-
ature. At the first minute, 0.7 mL reaction mixture was withdrawn
and added to 1m NaHSO4 solution (1 mL) immediately. Et2O (1 mL)
was added to dilute the reaction mixture. The organic layer was
separated and aqueous layer was extracted with EtOAc (3ꢁ1 mL).
The combined organic layer was washed with 1m NaHSO4 solution
(2 mL) followed by brine (1 mL), dried over MgSO4, filtered and
concentrated by evaporation under reduced pressure. The residue
was taken into CDCl3 and the 1H NMR spectrum was obtained.
Steps were repeated at 30, 60, 90, 120, 150, and 180 min. The sig-
nals used were at d=6.09 ppm (ArH, standard, integral calibrated
to 30, artemisinin 1 5.864 ppm (singlet, 1H), 2-deoxyartemisinin 15
5.698 ppm (singlet, 1H), ring-opened acid 14 5.462 ppm (singlet,
1H), tetraoxane 4 2.96 ppm (quartet, J=7.6 Hz, 2H), and 1-(ethyl-
sulfonyl)piperidin-4-one 17 3.05 ppm (quartet, J=7.6 Hz, 2H). The
concentrations were adjusted as follows: no. of mol. of 1 added=
0.0711 g/282.33 gmolÀ1 =2.518ꢁ10À4 mol; no. of mol of 4 added=
0.0955 g/373.44 gmolÀ1 =2.557ꢁ10À4 mol;[1] =[no. of mol of 1 cal-
culated ꢁ 2.518ꢁ10À4 mol/(no. of mol of 1 + no. of mol of 14 +
no. of mol of 15) ]/0.0075 L;[4] =[no. of mol of 4 calculated ꢁ
2.557ꢁ10À4 mol/(no. of mol of 4 + no. of mol of 17)]/0.0075 L.
Decomposition of tetraoxane 4 and trioxolane 6: Tetroxane 4. A
mixture of tetraoxane (67.7 mg, 0.181 mmol), riboflavin (13.7 mg,
0.0364 mmol, 0.2 equiv), BNAH (116.8 mg, 0.543 mmol, 3 equiv) in
a 1:1 mixture of CH3CN/phosphate buffer (pH 7.4, 5 mL) was stirred
under Ar for 3 h at room temperature. Solid NaHSO4·H2O (1.09 g)
was added to quench the reaction, followed by 1,3,5-trimethoxy-
benzene (6.2 mg, 0.0369 mmol) as an internal standard. The result-
ing mixture was diluted with Et2O (10 mL) and the organic layer
was separated. The aqueous layer was extracted with EtOAc (3ꢁ
5 mL). The combined organic layer was washed with aqueous
NaHSO4 (1m, 8 mL) followed by brine (8 mL) and then dried over
MgSO4. After filtration, the solution was concentrated by evapora-
tion under reduced pressure to leave the crude product mixture
that was analyzed by 1H NMR spectroscopy.[5] This showed the
presence of each of unreacted tetraoxane (15%), 2-adamantanone
(89%, based on reacting tetraoxane) and piperidone (82%, based
on reacting tetraoxane). The ketones were isolated from product
mixtures from several reactions by chromatography with EtOAc/
hexane (1:4) to give firstly the standard, then unreacted tetraoxane
and adamantanone, MS (ToF) calcd for C10H15O+ 151.1117, found
151.0629. Next EtOAc/hexane (2:1) was used to elute 1-(ethanesul-
fonyl)piperidin-4-one, MS (ToF) calcd for C7H14NO3SH+ 192.0689,
found 191.0624. Samples were matched with authentic samples.
c and d. Consumption of NADPH in the NADPH–Fre–FAD system by
artemisinin 1 and tetraoxane 4. NADPH (17.2 mg, 20.6 mmol) was
dissolved in degassed pH 7.4 buffer (1 mL) to generate a solution
containing 0.0206m; the NADPH solution was stored under Ar at
+48C prior to use. The NADPH solution (13 mL) was treated with
pH 7.4 degassed buffer (2.0 mL) in an Ar-purged UV cuvette (d=
1 cm) at 228C. The absorbance of NADPH at lmax 339 nm was
0.56 AU. The absorption coefficient of NADPH at l 339 nm is
6220mÀ1 cmÀ1; therefore, the actual concentration of NADPH solu-
tion was 0.01394m (~13.941 mmol). FAD (17.4 mg, 20.5 mmol) was
dissolved in degassed pH 7.4 buffer (1 mL) to generate a solution
containing 0.0205m. Fre (10 mL, 400 mm) was diluted to 100 mL by
addition of degassed pH 7.4 aqueous buffer, to generate a solution
Trioxolane 6. A mixture of trioxolane (64.8 mg, 0.181 mmol), ribofla-
vin (13.8 mg, 0.0366 mmol, 0.2 equiv), BNAH (77.6 mg, 0.362 mmol,
2 equiv) in a 1:1 mixture of THF/phosphate buffer (pH 7.4, 5 mL)
was stirred under Ar for 3 h at room temperature. Solid NaH-
SO4·H2O (0.73 g) was added to quench the reaction, followed by
ChemMedChem 2011, 6, 279 – 291
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
289

R. K. Haynes, Z.-H. Guo, D. Monti, et al.
MED
Chem. 2005, 117, 2100–2101; Angew. Chem. Int. Ed. 2005, 44, 2064–
2065.
[10] R. K. Haynes, W.- C. Chan, C.- M. Lung, A. C. Uhlemann, U. Eckstein, D.
Taramelli, S. Parapini, D. Monti, S. Krishna, ChemMedChem 2007, 2,
1480–1497.
1,3,5-trimethoxybenzene (5.3 mg, 0.0315 mmol), which was added
as an internal standard. The resulting mixture was diluted with
Et2O (10 mL) and the organic layer was separated. The aqueous
layer was extracted with EtOAc (3ꢁ5 mL). The combined organic
layer was washed with aqueous NaHSO4 (1m, 8 mL) followed by
brine (8 mL) and then dried over MgSO4. After filtration, the solu-
tion was concentrated by evaporation under reduced pressure to
leave the crude product mixture that was analyzed by 1H NMR
spectroscopy.[5] This showed the presence of unreacted trioxolane
(11%), 2-adamantanone (56%), and piperidone (86%). Products
from several reactions were isolated by chromatography, and
matched with authentic samples of the ketones as described
above.
[11] P. Coghi, N. Basilico, D. Taramelli, W.- C. Chan, R. K. Haynes, D. Monti,
ChemMedChem 2009, 4, 2045–2053.
[12] It is proposed that the artemisinin–heme adducts contribute to oxida-
tive stress, although it is difficult to visualize how important this contri-
bution may be, as the adducts cannot form in significant amounts
under physiological conditions (Ref. [9]) nor indeed with CO under the
same conditions (Ref. [11]). The deposition ‘Although it is not a conven-
tional biological target, heme is the masterpiece of the mechanism of
action of peroxide-containing antimalarial drugs’ represents an idꢂe
fixe: B. Meunier, A. Robert, Acc. Chem. Res. 2010, 43, 1444–1451.
[13] J. L. Vennerstrom, H.-N. Fu, W. Y. Ellis, A. L. Ager, J. K. Wood, S. L. Ander-
sen, L. Gerena, W. K. Milhous, J. Med. Chem. 1992, 35, 3023–3027.
[14] G. L. Ellis, R. Amewu, S. Sabbani, P. A. Stocks, A. Shone, D. Stanford, P.
Gibbons, J. Davies, L. Vivas, S. Charnaud, E. Bongard, C. Hall, K. Rimmer,
S. Lozanom, M. Jesffls, D. Gargallo, S. A. Ward, P. M. O’Neill, J. Med. Chem.
2008, 51, 2170–2177.
[15] M. Padmanilayam, B. Scorneaux, Y.- X. Dong, J. Chollet, H. Matile, S. A.
Charman, D. J. Creek, W. N. Charman, J. S. Tomas, C. Scheurer, S. Wittlin,
R. Brun, J. L. Vennerstrom, Bioorg. Med. Chem. Lett. 2006, 16, 5542–
5545.
[16] a) J. L. Vennerstrom, S. Arbe-Barnes, R. Brun, S. A. Charman, F. C. Chiu, J.
Chollet, Y. Dong, A. Dorn, D. Hunziker, H. Matile, K. McIntosh, M. Padma-
nilayam, J. Santo Tomas, C. Scheurer, B. Scorneaux, Y. Tang, H. Urwyler,
S. Wittlin, W. N. Charman, Nature 2004, 430, 900–904; b) Y.- Q. Tang, Y.-
X. Dong, J. M. Karle, C. A. DiTusa, J. L. Vennerstrom, J. Org. Chem. 2004,
69, 6470–6473.
Acknowledgements
Work at HKUST was carried out in the Open Laboratory of Chem-
ical Biology of the Institute of Molecular Technology for Drug Dis-
covery and Synthesis with financial support from the Government
of the HKSAR University Grants Committee Areas of Excellence
Fund, Projects No. AoE P/10-01/01-02-I, AOE/P-10/01-2-II, and the
University Grants Council Grants No. HKUST 6493/06M and
600507 (R.K.H.) and 601108 (Z.H.G.). The Milan work was con-
ducted partly in the context of the AntiMal project, funded under
the 6th Framework Programme of the European Community
(Contract No. IP-018834) to D.M. The financial support of the Uni-
versity of Milan (FIRB reti 2005RBPR05NWWC) is also acknowl-
edged.
[17] X.- F. Wang, Y.- X. Dong, S. Wittlin, D. Creek, J. Chollet, S. A. Charman,
J. S. Tomas, C. Scheurer, C. Snyder, J. L. Vennerstrom, J. Med. Chem.
2007, 50, 5840–5847.
[18] P. M. O’Neill, R. K. Amewu, G. L. Nixon, F. B. El Garah, M. Mungthin, J.
Chadwick, A. E. Shone, L. Vivas, H. Lander, V. Barton, S. Muangnoichar-
oen, P. G. Bray, J. Davies, B. K. Park, S. Wittlin, R. Brun, M. Preschel, K.
Zhang, S. A. Ward, Angew. Chem. Int. Ed. 2010, 49, 5693–5697.
[19] a) Z. I. Cabantchik, O. Kakhlon, S. Epsztejn, G. Zanninelli, W. Breuer, (In-
tracellular and Extracellular Labile Iron Pools) in Iron Chelation Therapy
(Ed.: C. Hershko), Kluwer Academic, New York, 2002, pp. 55–75; b) O.
Kakhlon, Z. I. Cabantchik, Free Radical Biol. Med. 2002, 33, 1037–1046;
c) P. F. Scholl, A. K. Tripathi, D. J. Sullivan, Curr. Top. Microbiol. Immunol.
Keywords: artemisinins · flavoenzymes · malaria · peroxides ·
oxidative stress
[1] R. K. Haynes, B. Fugmann, J. Stetter, K. Rieckmann, H.- D. Heilmann, H.-
W. Chan, M.- K. Cheung, W.- L. Lam, H.- N. Wong, S. L. Croft, L. Vivas, L.
Rattray, L. Stewart, W. Peters, B. L. Robinson, M. D. Edstein, B. Kotecka,
D. E. Kyle, B. Beckermann, M. Gerisch, M. Radtke, G. Schmuck, W.
Steinke, U. Wollborn, K. Schmeer, A. Rçmer, Angew. Chem. Int. Ed. 2006,
45, 2082–2088.
[2] a) J. L. Vennerstrom, M. T. Makler, C. K. Angerhofer, J. A. Williams, Antimi-
crob. Agents Chemother. 1995, 39, 2671–2677; b) H. Atamna, M. Kru-
gliak, G. Shalmiev, E. Deharo, G. Pescarmona, H. Ginsburg, Biochem.
Pharmacol. 1996, 51, 693–700.
ˇ
ˇ
˚
2005, 295, 293–324; d) Mladenka, T. Simunek, M. Hübl, R. Hrdina, Free
Radical Res. 2006, 40, 263–272.
[20] B. Pacorel, S. C. Leung, A. V. Stachulski, J. Davies, L. Vivas, H. Lander,
S. A. Ward, M. Kaiser, R. Brun, P. M. O’Neill, J. Med. Chem. 2010, 53, 633–
640.
[21] P. A. Stocks, P. G. Bray, V. E. Barton, M. Al-Helal, M. Jones, N. C. Araujo, P.
Gibbons, S. A. Ward, R. H. Hughes, G. A. Biagini, J. Davies, R. Amewu,
A. E. Mercer, G. Ellis, P. M. O’Neill, Angew. Chem. 2007, 119, 6394–6399;
Angew. Chem. Int. Ed. 2007, 46, 6278–6283.
[22] Z. T. Campbell, T. O. Baldwin, J. Biol. Chem. 2009, 284, 8322–8328.
[23] A. Kinoshita, Y. Nakayama, T. Kitayama, M. Tomita, FEBS J. 2007, 274,
1449–1458.
[24] a) F. Xu, K. S. Quandt, D. E. Hultquist, Proc. Natl. Acad. Sci. USA 1992, 89,
2130–2134; b) K. S. Quandt, D. E. Hultquist, Proc. Natl. Acad. Sci. USA
1994, 91, 9322–9326.
[25] a) F. Fieschi, V. Nivi›re, C. Frier, J. L. Dꢂcout, M. Fontecave, J. Biol. Chem.
1995, 270, 30392–30400; b) M. Ingelman, S. Ramaswamy, V. Nivi›re, M.
Fontecave, H. Eklund, Biochemistry 1999, 38, 7040–7049.
[26] M. Laclau, F. Lu, M. J. MacDonald, Mol. Cell. Biochem. 2001, 225, 151–
160.
[27] L. J. Smith, S. Browne, A. J. Mulholland, T. J. Mantle, Biochem. J. 2008,
411, 475–484.
[28] E. Balconi, A. Pennati, D. Crobu, V. Pandini, R. Cerutti, G. Zanetti, A.
Aliverti, FEBS J. 2009, 276, 4249–4260.
ˇ
ˇ
ˇ
˙
[3] P. Grellier, J. Sarlauskas, Z. Anusevicius, A. Maroziene, C. Houee-Levin, J.
ˇ
˙
Schrevel, N. Cenas, Arch. Biochem. Biophys. 2001, 393, 199–206.
[4] a) K. Becker, S. Rahlfs, C. Nickel, R. H. Schirmer, Biol. Chem. 2003, 384,
551–566; b) R. L. Krauth-Siegel, H. Bauer, R. H. Schirmer, Angew. Chem.
2005, 117, 698–724; Angew. Chem. Int. Ed. 2005, 44, 690–715; c) K.
Buchholz, R. H. Schirmer, J. K. Eubel, M. B. Akoachere, T. Dandekar, K.
Becker, S. Gromer, Antimicrob. Agents Chemother. 2008, 52, 183–191.
[5] R. K. Haynes, W.-Chi Chan, H.-N. Wong, K.-Y. Li, W.-K. Wu, K.-M. Fan,
H. H. Y. Sung, I. D. Williams, D. Prosperi, S. Melato, P. Coghi, D. Monti,
ChemMedChem 2010, 5, 1282–1299.
[6] G. N. Sarma, S. N. Savvides, K. Becker, M. Schirmer, R. H. Schirmer, P. A.
Karplus, J. Mol. Biol. 2003, 328, 893–907.
[7] T. Akompong, N. Ghori, K. Haldar, Antimicrob. Agents Chemother. 2000,
44, 88–96.
[8] a) R. T. Eastman, D. A. H. Fidock, Nat. Rev. Microbiol. 2009, 7, 864–874;
b) T. N. C. Wells, P. L. Alonso, W. E. Gutteridge, Nat. Rev. Drug Discovery
2009, 8, 879–891; c) P. M. O’Neill, V. E. Barton, S. A. Ward, Molecules
2010, 15, 1705–1721.
[9] a) R. K. Haynes, W.-Y. Ho, H.-W. Chan, B. Fugmann, J. Stetter, S. L. Croft,
L. Vivas, W. Peters, B. L. Robinson, Angew. Chem. 2004, 116, 1405–1409;
Angew. Chem. Int. Ed. 2004, 43, 1381–1385; b) R. K. Haynes, Angew.
[29] T. Fitzpatrick, S. Ricken, M. Lanzer, N. Amrhein, P. Macheroux, B. Kappes,
Mol. Microbiol. 2001, 40, 65–75.
[30] K. Buchholz, M. A. Comini, D. Wissenbach, R. H. Schirmer, R. L. Krauth-
Siegel, S. Gromer, Mol. Biochem. Parasitol. 2008, 160, 65–69.
290
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 279 – 291

Antimalarial Peroxides
[31] N. Fisher, P. G. Bray, S. A. Ward, G. A. Biagini, Trends Parasitol. 2007, 23,
305–310.
[36] A Thomas, E. J. Wanner, J. Org. Chem. 2000, 65, 2444–2457.
[37] Y. Li, H. D. Hao, Q. Zhang, Y. Wu, Org. Lett. 2009, 11, 1615–1618.
[38] P. Ghorai, P. H. Dussault, Org. Lett. 2009, 11, 213–216.
[39] J. L. Vennerstrom, Y. X. Dong, J. Chollet, H. Matile, Spiro- and Dispiro-
1,2,4-trioxalane Antimalarials, US patent 2002, 6, 486, 199.
[40] Z. T. Campbell, T. O. Baldwin, J. Biol. Chem. 2009, 284, 8322–8328.
[41] R. B. Dawson, Data for Biochemical Research, 3rd Ed., Oxford, Clarendon
Press, 1985, p. 122 (ISBN: 0-19-855358-7).
[32] a) A. N. Woodmansee, J. A Imlay, J. Biol. Chem. 2002, 277, 34055–34066;
b) F. Petrat, S. Paluch, E. Dogruçz, P. Dçrfler, M. Kirsch, H.-G. Korth, R.
Sustmann, H. de Groot, J. Biol. Chem. 2003, 278, 46403–46413.
[33] a) S. Krishna, L. Bustamante, R. K. Haynes, H. M. Staines, Trends Pharma-
col. Sci. 2008, 29, 520–527; b) I. R. Dunay, R. K. Haynes, W.-C. Chan, L. D.
Sibley, Antimicrob. Agents Chemother. 2009, 53, 4450–4456.
[34] Minimized geometry structures (AM1, Spartan ‘08) indicate that the
chair-like tetraoxane nucleus (cf. Ref. [18]) enables attenuation of O
electron lone-pair repulsion that is not attainable in 6. In addition, the
1,3-disposition of substituents about the trioxolane nucleus in 6 results
in a slightly more exposed environment about the peroxide than is the
case for 4. However, ground-state energy for 6 is calculated to be
116 kJmolÀ1 lower than that of 4.
[42] Tabulated Molar Extinction Coefficient for Methylene Blue in Water:
http://omlc.ogi.edu/spectra/mb/mb-water.html (accessed December 1,
2010).
´
ˇ ˇ
[35] I. N. Cvijetic, Z. P. zizak, T. P. Stanojkovic, Z. D. Juranic, N. Terzic, I. M. Op-
´
´
´
Received: November 25, 2010
´
´
senica, D. M. Opsenica, I. O. Juranic, B. J. Drakulic, Eur. J. Med. Chem.
2010, 45, 4570–4577.
Published online on December 23, 2010
ChemMedChem 2011, 6, 279 – 291
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
291