Interaction of qinghaosu (artemisinin) with cysteine sulfhydryl mediated by traces of non-heme iron

Article abstract of DOI:10.1002/(SICI)1521-3773(19990903)38:17<2580::AID-ANIE2580>3.0.CO;2-J

The antimalarial action of 1,2,4-trioxanes such as qinghaosu (QHS) may take place through the mechanism shown schematically: In the presence of cysteine traces of non-heme iron (FeSO₄) may cleave the peroxy bond of QHS rapidly, and the transien

Full text of DOI:10.1002/(SICI)1521-3773(19990903)38:17<2580::AID-ANIE2580>3.0.CO;2-J

COMMUNICATIONS
fail to drive the reaction to completion. Along with the fact
that heme is accumulated in the food vacuole[6] and the
extraordinarily low toxicity of QHS with respect to uninfected
cells (where there is practically no free heme), this leads to the
presently well accepted notion[2c, 4f] that the in vivo cleavage of
QHS (or 1,2,4-trioxanes in general) is exclusively induced by
the heme iron. As a consequence, essentially all the subse-
quent biological as well as mechanistic studies aimed at
identifying the target molecules of the 1,2,4-trioxanes are
confined to those events either involving the heme (or its
oxidized and polymerized form, hemozoin) directly[7] or
closely associated[2c, 8] with the heme iron. However, further
development of the theory has been rather difficult, because
there seem to be no logical links between the parasiteꢁs death
and the damage occurring to, for example, the free heme or
hemozoin (none of which is essential for the parasiteꢁs life).
Now we have found that heme is not the only iron-
containing species that can trigger the degradation of QHS in
vivo. Although they play only a negligible role under near
physiological conditions, free FeII ions may form one or more
complexes with, for example, amino acids and thus be greatly
activated. In the presence of cysteine, for instance, QHS can
be degraded very rapidly. In aqueous acetonitrile at 388C,
QHS is fully consumed instantly if one equivalent each (with
respect to QHS) of FeSO4 and cysteine is present. Reducing
the amount of added iron slows down the reaction accord-
ingly. However, even with as little as 10 4 equivalents of the
added iron, the cleavage can be practically finished in about
one day. In a typical run at 388C under N2 in deaerated
aqueous acetonitrile (1/1) containing two equivalents of
cysteine (0.1m), QHS (0.05m) is completely degraded by
10 3 equivalents of FeSO4 (5 Â 10 5 m) in 3 ± 4 h, giving a
product mixture (recovered from the organic phase after
workup) of compounds 2, 3, 4, and 5 (Scheme 1) in an
approximate molar ratio of 29:9:1:6, as shown by the 1H NMR
spectrum (300 MHz).
Apart from the tremendous rate acceleration, a striking
consequence of the presence of cysteine is that a previously
unknown aldehyde 5[9] is formed in a significant quantity
(Scheme 1; see also reference [3b]). It appears that this new
product can be formed only from radical 7 by abstraction of a
hydrogen atom from, for example, the SH group in cysteine
followed by elimination of an acetate as shown in Scheme 2.
The formation of 5 provides unequivocal evidence for the
involvement of the highly reactive primary radical 7, which
was not trapped in our earlier work,[3b] in the degradation
process. (However, Butler et al.[10] have reported concurrent
trapping of both 6 and 7 using different trapping agents.)
The concentration of free amino acids (probably due to the
catabolism of hemoglobin) is greater in infected than in
uninfected erythrocytes,[11] and malaria parasite has a high
concentration of the reduced glutathione (the main sulfur-
hydryl-containing reducing agent in physiological systems).[12]
Therefore, the non-heme iron mediated degradation mecha-
nism demonstrated in our model system may well be in
operation in physiological systems too. It should be noted that
this is the first piece of established evidence for intermolec-
ular interactionsÐabstraction of a hydrogen atom from
another molecule, a process that may cause damage to the
Interaction of Qinghaosu (Artemisinin) with
Cysteine Sulfhydryl Mediated by Traces of
Non-Heme Iron**
Yikang Wu,* Zheng-Yu Yue, and Yu-Lin Wu
Although it was once considered to be nearly eradicated in
the 1950s and 1960s, malaria, an infectious disease known
since ancient times, still poses a serious threat to the health of
mankind. It claims 1 ± 3 million lives[1] annually owing to the
emergence of multidrug-resistant strains. Urged by this
situation, the scientific community has been making great
efforts to develop novel antimalarial compounds since the
1960s. Qinghaosu (QHS, 1, also called artemisinin; see
Scheme 1), a natural 1,2,4-trioxane of herbal origin discovered
by Chinese scientists in the 1970s, appears to be one of the
most promising leads to such agents. While it is still unclear
how QHS kills malaria parasites, the active species resulting
from cleavage of the peroxy bond in QHS are now generally
believed[2±5] to be responsible for the parasiticidal activity, and
the intraerythrocytic heme iron is regarded as the trigger for
the in vivo cleavage reactions by nearly all workers in this
field. However, the commonly accepted opinions on an issue
that is not fully understood are not necessarily objective
descriptions of the genuine situation. Here we report that
traces of non-heme iron in the presence of cysteine can also
cleave QHS efficiently, and the transient carbon-centered
radical formed can covalently bond to the ligand at iron
through a sulfur atom. The implications of these new findings
may substantially change our current understanding and thus
greatly facilitate identification of the vital intracellular target
molecules of QHS.
The mode of action is a long-pending problem in the
chemotherapy of malaria with QHS-type agents. The current
working theory[2c, 4f] is that the 1,2,4-trioxane, on contact with
heme iron, is cleaved at the peroxy bond into a radical anion,
which evolves further to give carbon-centered radicals, high
valent iron, and other reactive intermediates; all these may
then interact with so far unidentified vital intracellular target
molecules through undefined mechanisms and finally lead to
the death of the parasite. Following the pioneering work of
Meshnick et al.[2a] and Posner and Oh,[4a] the cleavage process
has been intensively investigated[3±5] in various model systems
with near stoichiometric amounts of free FeII ion. Substan-
tially reduced amounts of non-heme iron, such as in FeSO4,
[*] Prof. Dr. Y.-K. Wu, Prof. Z.-Y. Yue,[‡] Prof. Y.-L. Wu
State Key Laboratory of Bio-organic & Natural Products Chemistry
Shanghai Institute of Organic Chemistry
Chinese Academy of Sciences
354 Fenglin Road Shanghai 200032 (China)
Fax : (‡ 86)21-64166128
E-mail: yikangwu@pub.sioc.ac.cn
‡
[ ] On leave from:
Department of Chemistry, Heilongjiang University
Harbin 150080 (China)
[**] This work was financially assisted by the Chinese Academy of
Sciences (CAS, no. KJ 951-A1-504-04), the Life Science Special Fund
of CAS Supported by the Ministry of Finance (no. Stz 98-3-03), the
Ministry of Sciences and Technology (no. 970211006-6), and the
National Natural Science Foundation of China (no. 29832020).
2580
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
1433-7851/99/3817-2580 $ 17.50+.50/0
Angew. Chem. Int. Ed. 1999, 38, No. 17

COMMUNICATIONS
mixture by column chromatography on reverse-phase silica
gel eluting with water containing an increasing amount of
methanol gives a white solid, whose elementary analysis
suggests a formula of C16H27NO6S ´ 2H2O (the presence of
waters of crystallization is confirmed by the observations
made in the determination of the melting point).
Based on our mechanistic knowledge, we have envisaged
that the primary C-4 radical 7 may ªintramolecularlyº attack
the cysteine sulfur atom complexed to the FeIII ion and release
an FeII ion (Scheme 3). Because the conformation of the C-5a/
C-12a/O-1 linkage is relatively fixed in the trans arrange-
mentÐwhich, owing to the strain developed in the five-
membered transition state, is expected to slow down the
Scheme 1.
Scheme 3.
otherwise very facile radical substitution at O-1 leading to 2Ð
the seven-membered ring in the transition state is expected to
be much easier to attain than otherwise (e.g., in an unre-
stricted linear chain). The product resulting from such an
ªintramolecularº radical substitution at the sulfur atom
followed by a series of degradation steps at the QHS moiety
would lead to structure 10, which fits the formula C16H27NO6S.
The 1H NMR spectrum also shows a clear ABX system from
the cysteine residue and two distinct Me doublets from QHS,
but no sign of a Me triplet (which ensures that C-4 and C-5 do
not exist as an ethyl group). The remaining signals, however,
are seriously broadened, suggesting that we are dealing with
an exchanging system.
Scheme 2. RSH ˆ cysteine.
hydrogen-donor moleculeÐof the intermediate carbon-cen-
tered radicals involved in the degradation of QHS. Apart
from its evident importance to antimalarial studies, such a
mechanism might play a critical role in the understanding of
many other[13] bioactivities of QHS that do not seem to be
related to the intraerythrocytic heme iron.
An even more exciting finding from this work is that a
highly polar, water-soluble compound is detected in the
reaction mixture that may amount up to nearly half of the
starting QHS. With thin-layer chromatography (TLC) this
compound can be clearly differentiated from the starting
materials and all products soluble in organic solvents by the
use of nBuOH/AcOH/H2O (3/1/1) as eluent. Furthermore, it
can be visualized with ninhydrin, revealing the presence of an
NH2 group. Isolation of this component from the reaction
Nonetheless, up to this point the presence of a sulfur and a
nitrogen atom in 10, which can only come from cysteine under
the given conditions, and of the carbon framework from QHS
are proven beyond all doubt. To demonstrate that the sulfur
atom of the cysteine residue is indeed covalently bonded to
C-4 of the QHS moiety and to get around all the difficulties
caused by the high polarity and the exchangeable protons, we
treated 10 with acetic anhydride. The result is very gratifying:
The product of ªacetylationºÐacetic anhydride acts here as
an activating agent of the carboxylic acid moiety in QHS and
initiates cyclic acetal formation, which consequently gener-
ates a cyclic sulfonium ion as an excellent leaving group at the
b-carbon atom of the cysteine residueÐis easily characterized
1
as compound 11 based on the IR, H and 13C NMR, DEPT,
DQF-COSY, NOESY, HMQC, MS, and HR-MS data,[14]
Angew. Chem. Int. Ed. 1999, 38, No. 17
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
1433-7851/99/3817-2581 $ 17.50+.50/0
2581

COMMUNICATIONS
Acta 1996, 79, 1475 ± 1487; d) A. J. Bloodworth, A. Shah, Tetrahedron
Lett. 1995, 36, 7551 ± 7554; e) R. K. Haynes, S. C. Vonwiller, Tetrahe-
dron Lett. 1996, 37, 253 ± 256; f) R. K. Haynes, S. C. Vonwiller,
Tetrahedron Lett. 1996, 37, 257 ± 260.
which undoubtedly show that a s bond is present between the
sulfur and C-4 in 11 (and therefore 10).
The significance of the concatenation of QHS and cysteine
observed in this simple chemical model system lies in the
precise mechanistic basis it may offer at the molecular level
for the radical-mediated alkylation of, for example, pro-
teins[5a, 7b, 8] (especially in the absence of heme iron). Such a
basis has been missing in all documented investigations in
spite of its evident indispensability. To date, only addition of
QHS to the porphyrin in a substantially simplified heme
model has been elucidated,[7a] which unfortunately cannot
explain how QHS can possibly be irreversibly connected to
any of the biomolecules reported in the literature for a radical
reaction associated with cleavage induced by an FeII ion.
Sulfur± iron bonds (or more generally S M bonds, where M
is a transition metal with low-lying redox states suitable for
cleaving peroxides) similar to that studied in this work are
known to exist in many enzymes and functional proteins. The
potential structural difference between S Fe bonds in malaria
parasites and those in host cells might be responsible for the
extraordinarily high selective cytotoxicity of QHS. By pre-
senting the first model for the irreversible denaturing/
disabling of the redox center in these speciesÐwhich are
liable to be overlooked in, for example, radioactive isotope
labeling studies because of their low abundanceÐthe present
work may prompt further investigations into this so far largely
overlooked dimension of non-heme proteins. Furthermore,
the results along this line might practically reestablish our
understanding of the parasiticidal action of QHS.
[6] A. F. G. Slater, A. Cerami, Nature 1992, 355, 167 ± 169.
[7] a) A. Robert, B. Meunier, J. Am. Chem. Soc. 1997, 119, 5968 ± 5969;
b) S. Bharel, R. A. Vishwakarma, S. K. Jain, J. Chem. Soc. Perkin
Trans. 1 1998, 2163 ± 2166.
[8] a) W. Asawamahasakda, I. Ittarat, Y.-M. Pu, H. Ziffer, S. R. Meshnick,
Antimicrob. Agents Chemother. 1994, 38, 1854 ± 1858; b) Y.-Z. Yang,
B. Little, S. R. Meshnick, Biochem. Pharmacol. 1994, 48, 569 ± 573;
c) Y.-Z. Yang, W. Asawahasakda, S. R. Meshnick, Biochem. Pharma-
col. 1993, 46, 336 ± 339.
[9] Physical and spectroscopic data for 5 (1H and 13C NMR signals
assigned based on the DQF-COSY and HMQC spectra): M.p. 45.5 ±
46.58C; [a]9D ˆ ‡54.5 (c ˆ 1.0 in CHCl3); 1H NMR (600 MHz, CDCl3):
d ˆ 10.10 (s, 1H), 2.68 (quint., 1H, J ˆ 8.0 Hz; H-9), 2.41 (m, 1H;
H-8a), 2.07 (m, 1H; H-6), 2.04 (m, 1H; H-7), 1.84 (dq, J ˆ 3.4, 13.0 Hz,
1H; H-8b), 1.68 (brd, J ˆ 13.2 Hz, 1H; H-8a), 1.63 (m, 1H; H-5), 1.52
(m, 1H; H-5a), 1.35 (m, 1H; H-5), 1.33 (m, 1H; H-7), 1.03 (d, J ˆ
6.4 Hz, 3H; Me at C-6), 1.02 (d, J ˆ 6.5 Hz, 3H; Me at C-9), 0.99 (t,
J ˆ 7.3 Hz, 3H, H-4); 13C NMR (150 MHz, CDCl3): d ˆ 204.09 (C-12),
179.17 (C-10), 91,73 (C-12a), 53.96 (C-5a), 52.84 (C-8a), 37.56 (C-9),
35.31 (C-7), 33.68 (C-6), 22.19 (C-5), 20.65 (C-8), 20.22 (Me at C-6),
13.34 (C-4), 11.42 (Me at C-9); IR (film): nÄ ˆ 1795, 1735 cm 1; MS m/z
‡
(%): 225 ([M‡1] , 56.3), 207 (58.3), 195 (100), 179 (12.7), 167 (0.4),
151 (17.3), 121 (81.3), 97 (29), 83 (91), 69 (65), 55 (91), 43 (18);
elemental analysis calcd for C13H20O3: C 69.60, H 8.99; found: C 69.68,
H 9.14.
[10] A. R. Butler, B. C. Gilbert, P. Hulme, L. Irvine, L. Renton, A. C.
Whitwood, Free Radicals Res. 1998, 28, 471 ± 476.
[11] P. J. Rosenthal, S. R. Meshnick, Mol. Biochem. Parasitol. 1996, 83,
131 ± 139
[12] H. Atamna, H. Ginsburg, Eur. J. Biochem. 1997, 250, 670 ± 679.
[13] See for example a) A. C. Beekman, H. J. Woerdenbag, W. van Uden,
N. Pras, A. W. T. Konings, H. V. Wikstrom, J. Pharm. Pharmacol. 1997,
49, 1254 ± 1258; b) Q. Y. Zheng, L. G. Darbie (Hauser Chemical
Research), WO-A 9701548 A1, 1997 [Chem. Abstr. 1997, 126, 157663];
c) H. C.-Y. Lai, N. P. Singh (University of Washington), WO-
A 9634602 A1, 1996 [Chem. Abstr. 1997, 126, 26821]; d) S. O. Duke,
F. E. Dayan, A. Hernandez, M. V. Duke, H. K. Abbas, Brighton Crop
Prot. Conf. Weeds 1997, 2, 579 ± 586 [Chem. Abstr. 1998, 128, 98854];
e) M. Jung, R. F. Schinazi, Bioorg. Med. Chem. Lett. 1994, 4, 931 ± 4;
f) R. L. Berens, E. C. Krug, P. B. Nash, T. J. Curiel, J. Infect. Dis. 1998,
177, 1128 ± 1131.
Received: April 6, 1999
Revised version: May 12, 1999 [Z13239IE]
German version: Angew. Chem. 1999, 111, 2730 ± 2733
Keywords: alkylations ´ antimalarial agents ´ chelates ´
S ligands ´ radicals
[14] Physical and spectroscopic data for 11 (1H and 13C NMR signals
assigned based on the DQF-COSY and HMQC spectra): M.p. 148.5 ±
[1] F. H. Collins, S. M. Pasewitz, Annu. Rev. Entomol. 1995, 40, 195 ± 219.
[2] a) S. R. Meshnick, A. Thomas, A. Ranz, C. M. Xu, H. P. Pan, Mol.
Biochem. Parasitol. 1991, 49, 181 ± 190; b) S. R. Meshnick, Y.-Z. Yang,
V. Lima, F. Kuypers, S. Kamchonwongpaisan, Y. Yuthavong, Anti-
microb. Agents Chemother. 1993, 37, 1108 ± 1114; c) J. Bhisutthibhan,
X.-Q. Pan, P. A. Hossler, D. J. Walker, C. A. Yowell, J. Carlton, J. B.
Dame, S. R. Meshnick, J. Biol. Chem. 1998, 273, 16192 ± 16198.
[3] a) W.-M. Wu, Z.-J. Yao, Y.-L. Wu, K. Jiang, Y.-F. Wang, H.-B. Chen, F.
Shan, Y. Li, Chem. Commun. 1996, 2213 ± 2214; b) W.-M. Wu, Y.-K.
Wu, Y.-L. Wu, Z.-J. Yao, C.-M. Zhou, Y. Li, F. Shan, J. Am. Chem. Soc.
1998, 120, 3316 ± 3325.
[4] a) G. H. Posner, C. H. Oh, J. Am. Chem. Soc. 1992, 114, 8328 ± 8329;
b) G. H. Posner, C. H. Oh, D.-S. Wang, L. Gerena, W. K. Milhous, S.
Meshnick, W. Asawamahasadka, J. Med. Chem. 1994, 37, 1256 ± 1258;
c) G. H. Posner, J. N. Cumming, P. Ploypradith, C. H. Oh, J. Am.
Chem. Soc. 1995, 117, 5885 ± 5886; d) G. H. Posner, S. B. Park, L.
Gonzalez, D.-S. Wang, J. N. Cumming, D. Klinedinst, T. A. Shapiro,
M. D. Balci, J. Am. Chem. Soc. 1996, 118, 3537 ± 3538; e) J. N.
Cumming, P. Ploypradith, G. H. Posner, Adv. Pharmacol. 1997, 37,
253 ± 297; f) G. H. Posner, M. H. Parker, J. Northrop, F. S. Elias, P.
Ploypradith, S. Xie, T. A. Shapiro J. Med. Chem. 1999, 42, 300 ± 304.
[5] a) M. A. Avery, P. C. Fan, J. M. Karle, J. D. Bonk, R. Miller, D. K.
Goins, J. Med. Chem. 1996, 39, 1885 ± 1897; b) C. W. Jefford, F.
Favarger, M. G. H. Vicente, Y. Jacruier, Helv. Chim. Acta 1995, 78,
452 ± 458; c) C. W. Jefford, M. G. H. Vicente, Y. Jacquier, F. Favarger,
J. Mareda, P. Millason-Schmidt, G. Brunner, U. Burger, Helv. Chim.
149.58C; [a]9D
ˆ
64.9 (c ˆ 0.53 in CHCl3); 1H NMR (CDCl3): d ˆ
5.95 (s, 1H; H-12), 3.46 (dq, J ˆ 1.2, 7.2 Hz, 1H; H-9), 2.84 (dt, J ˆ 3.0,
14 Hz, 1H; H-4a), 2.39 (dt, J ˆ 14, 3.0 Hz, 1H; H-4b), 2.12 (tt, J ˆ 15,
4.2 Hz, 1H; H-5b), 2.02 (m, 1H; H-5a), 2.00 (m, 1H; H-8a), 1.92 (m,
1H; H-7b), 1.89 (m, 1H; H-8a), 1.73 (m, 1H; H-6), 1.47 (dt, J ˆ 12,
2.4 Hz, 1H; H-5a), 1.24 (m, 1H; H-8b), 1.19 (t, J ˆ 6.6 Hz, 3H; Me at
C-9), 1.12 (dq, J ˆ 3, 13 Hz, 1H; 7a), 0.91 (d, J ˆ 6.0 Hz, 3H; Me at
C-6); 13H NMR (CDCl3): d ˆ 172.33 (C-10), 80.60 (C-12), 67.76 (C-
12a), 50.58 (C-5a), 44.70 (C-8a), 35.27 (C-9), 34.38 (C-7), 29.91 (C-6),
24.15 (C-4), 23.74 (C-8), 23.62 (C-5), 20.08 (Me at C-9), 12.78 (Me at
‡
C-6); IR (KBr): nÄ ˆ 3501, 1735 cm 1; MS m/z (%) 256 (M , 35), 238
(100); HR-MS calcd for C13H20O3S: 256.1133; found: 256.1132.
2582
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
1433-7851/99/3817-2582 $ 17.50+.50/0
Angew. Chem. Int. Ed. 1999, 38, No. 17