New insights into the degradation of qinghaosu (Artemisinin) mediated by non-heme-iron chelates, and their relevance to the antimalarial mechanism

Article abstract of DOI:10.1002/1522-2675(20010418)84:4<928::AID-HLCA928>3.0.CO;2-I

Probing the degradation of qinghaosu (1) induced by cysteine-iron chelates by varying the acidity of the reaction medium led to interesting insights into the cleavage of 1, which help to elucidate the antimalarial action of the 1,2,4-trioxanes.

Full text of DOI:10.1002/1522-2675(20010418)84:4<928::AID-HLCA928>3.0.CO;2-I

928
Helvetica Chimica Acta ± Vol. 84 (2001)
New Insights into the Degradation of Qinghaosu (Artemisinin)
Mediated by Non-Heme-Iron Chelates, and Their Relevance to the
Antimalarial Mechanism
by Yikang Wu*, Zheng-Yu Yue1), and He-Hua Liu
State Key Laboratory of Bio-Organic Chemistry and Natural Products Chemistry, Shanghai Institute of
Organic Chemistry, Chinese Academy of Sciences, 354 Fenglin Road, Shanghai 200032, China
(e-mail: yikangwu@pub.sioc.ac.cn)
Probing the degradation of qinghaosu (1) induced by cysteine-iron chelates by varying the acidity of the
reaction medium led to interesting insights into the cleavage of 1, which help to elucidate the antimalarial action
of the 1,2,4-trioxanes.
With the wide-spread distribution of multi-drug-resistant malaria cases and
malariaꢀs reappearance in Europe [1][2], the trioxane-type drugs exemplified by
qinghaosu (QHS; 1) have attracted much attention [3][4] because of their novel mode
of action. Since the early 1990s, it is broadly believed that the intraparasitic free heme is
responsible for the cleavage preceding the parasiticidal action, because of its unique
occurrence and high reactivity. However, this notion has recently been challenged by
the observation [5] that in the presence of cysteine, even traces of non-heme iron may
cleave QHS rapidly.
Scheme 1. The Two C-Centered Radicals 4 and 5, Generated from QHS (1). For clarity, some ligands at FeIII are
not shown.
5
H
5
H
6
"Fe2+
(?)
"
3
2
O
1
O
3
5a
8a
12a
5a
.
1
O
FeO
8
O
O
13
2
12
12
H
H
H
O
10
O
.
H
3
1
9
11
(?)
HO
H
FeO
(?)
.
O
"Fe2+
H
"
(?)
H
FeO
AcO
H
5
O
3
FeO
1
H
5a
H
5
O
2
O
O
.
O
12
H
4
O
2
As a model for the intracellular species containing Fe S type redox centers, the
cysteine-iron chelate has provided remarkable new insights into the parasiticidal action
1
)
On leave from the Department of Chemistry, Heilongjiang University, Harbin 150080, China.

Helvetica Chimica Acta ± Vol. 84 (2001)
929
of QHS. For instance, the observed covalent bonding between the QHS-derived radical
and the S-atom provides a possible (and practically the only) molecular-level
explanation for the irreversible alkylation of non-heme proteins associated with the
cleavage of QHS. Furthermore, the ꢁintramolecularꢀ alkylation of the S-ligand strongly
suggests that the two C-centered radicals 4 and 5 (see Scheme 1) contribute differently
to the antimalarial action, with 4 being the most likely ꢁkillingꢀ species. In this
contribution, we wish to communicate some further insights gained from the cysteine-
iron model, which would not only help to elucidate the parasiticidal action but also
facilitate the rational design of future generations of trioxane-based antimalarials.
As illustrated in Schemes 2 and 3, there exist several ways [6] for 4 and 5 to evolve
further; some of the paths (e.g., Path c in Scheme 2) could never damage the
intracellular biomolecules, while others were suggested to be responsible for the lethal
effects of QHS. It thus becomes imperative to know how the C-centered radicals
choose among these possible pathways. To probe the elusive relations between the ꢁO-
1ꢀ and ꢁO-2ꢀ routes (i.e., via O(1) radical 3 and via O(2) radical 2, resp.; see Scheme 1) as
well as those between the H-abstractions and radical substitutions, we examined the
effects of added base on the cleavage in the cysteine-iron model system.
Scheme 2. Some Reaction Paths for the Primary Radical 4
RSH
HO2C
H
NH2
a
CH2
.
5a
12a
Path a
H
Path b
b
c
1
O
H
5
S
S
4
b
Fe
OH
O
O
O
O
O
c
NH2
H
H
O
H
O
5
4
H
6
"4"
HO2C
O
O
5a
8
Path c
AcO
8
12
8a
H
6
12a
H
H
H
O
1
O
10
9
7
O
O
Scheme 3. Some Reaction Paths for the Secondary Radical 5
RSH
b
a
S
H
a
O
.
H
Fe
NH2
c
O
HO
Path a
O2C
c
HO
O
O
H
Path b
HO
H
H
10
H
O
O
H
O
O
"5"
O
O
O
O
H
Path c
H
H
H
O
O
11
9

930
Helvetica Chimica Acta ± Vol. 84 (2001)
Table 1 shows that, with the increase in the initial amount of added base, more 6, 8,
9, and cystine were formed. The formation of 7 and 11 (via 10), however, was
suppressed. Grouping these end products according to the initial radical 2 or 3 reveals
that a larger excess of initial base resulted in an increase in reaction via the ꢁO-2ꢀ route
since the total ꢁO-2ꢀ products, i.e., 6 ± 8, all from the O(2) radical 2, increased
monotonically when the initial base amount was increased. Such a shift of the balance
between two routes associated with the increase in the basicity of the reaction medium
may reflect a faster elimination of 2 compared with 3, providing supporting evidence
for the postulate of interchangeable [7] (cf. [6] above) radical anions. This is based on
the following considerations: Oxidation of the cysteine SH releases a proton by the
.
‡
transformation RSH ‡ FeIII ! RS ‡ FeII ‡ H , which makes the reaction medium
increasingly acidic. The formation of cystein-iron chelates (more powerful reducing
agents than the free FeII ion), however, requires relatively high basicity. According to
Jameson et al. [8], the predominant complex (in H2O) at pH 2.7 ± 3.9, 5.5 ± 8, and 8.8 ±
‡
11.2 is [FeL] , [Fe(OH)L], and [Fe(OH)L2]2 , respectively, where L stands for
cysteine. Therefore, more ligands are expected to attach to the Fe2‡ ion at relatively
higher pHs. Addition of a larger excess of base at the beginning could relatively better
maintain the basicity of the reaction medium and consequently a higher concentration
of the complex(es) (giving a more-intense blue-purple color) containing more ligands.
Introduction of additional ligand(s) to a lower-order complex would inevitably increase
the steric crowding around the Fe-core and thus make the reducing species bulkier. In
another [9] context, we have shown that increasing the steric hindrance at C(9) of QHS
(1) led to greatly reduced overall cleavage rates, with the ꢁO-2ꢀ route being practically
shut off, which indicated that the attack at O(1) is more sensitive to steric crowding.
Similarly, a bulkier reducing agent would be less likely to attack O(1) than O(2).
Consequently, more 3 would have formed and reaction via the ꢁO-1ꢀ route would
have increased (rather than decreased as observed), if 2 and 3 were not interchange-
able.
It is even more interesting to note that the ratios for the H-abstraction to the radical
substitution for each route (i.e. the ꢁO-1ꢀ or ꢁO-2ꢀ route) also varied with regular trends
Table 1. The Effects of Excess Base on the Product Distributiona)
Excess base [mol-%]
0
5
10
20
Cystine [mmol]
6 [mmol]
7 [mmol]
8b) [mmol]
9 [mmol]
11 [mmol]
ªO-1ºc)
0.12
0.15
0.19
0.20
0.073
0.339
ca. 0
0.012
0.100
0.112
0.412
0.078
0.224
0.130
0.013
0.055
0.068
0.431
0.085
0.219
0.129
0.014
0.053
0.067
0.434
0.111
0.155
0.182
0.020
0.032
0.052
0.448
ªO-2ºd)
a) The reactions were carried out in parallel at 38 ± 408 (bath) under N2 in MeCN/H2O 1 : 1 (v/v) (10 ml) starting
with 0.50 mmol of QHS, 1.00 mmol of cysteine hydrochloride monohydrate, 5.0 ml of MeCN, 5.0 ml of doubly
distilled deaerated H2O, 5.0 ´ 10 4 mmol of FeII/FeIII ion, and the indicated amount of Et3N. The amount (in
mmol) of the products (except 8) was estimated from the amount of isolated 6 and the molar ratios measured
from the integrals for the acetal (or aldehyde) 1H-signals in the 1H-NMR spectra of the crude product mixtures
(QHS was not detected). b) Estimated from the missing mmol. c) (9 ‡ 11). d) (6 ± 8).

Helvetica Chimica Acta ± Vol. 84 (2001)
931
(Table 2). For example, increasing the amount of base added led to an increase in the
amount of 6 (i.e., 6/(6 ± 8)) as well as that of 8 and a decrease in the weight of 7 in the
total ꢁO-2ꢀ products. Similarly, the amount of 9 in the total ꢁO-1ꢀ products increased
with increasing the initial excess base. These trends reflect the influence of the
increased steric crowding at the Fe-core caused by higher basicity on the relative rates
of the corresponding reaction paths. The H-abstraction from the cysteine SH by 4 (Path
a in Scheme 2) would be least affected, because the primary radical is not so close to the
Fe chelate. Path c, however, would be affected much more due to the repulsion between
H
C(5a), H C(9), etc. and the bulky chelate in the conformation required for the
back-side radical attack. That is why the amount of 7 in the total ꢁO-2ꢀ products
remarkably decreased. Formation of 8 does not require the chelate moiety to approach
the lactone ring as in Path c. Therefore, the rate would not be lowered so much as in
Path c. There may also be a statistically favorable factor: in the bulkier chelate, i.e.,
[Fe(OH)L2]2 , there are two S-atoms available for the primary radical to attack.
Secondary radicals are more sensitive to steric hindrance than primary ones. It is, hence,
not surprising that the substitutions with 5 (e.g., Path a in Scheme 3) would be
significantly suppressed when the reducing agent became bulkier. Note that no
ꢁintramolecularꢀ attack at the S-atom by the secondary radical in 5 was observed.
Table 2. The Effects of Excess Base on the Ratio Abstraction/Substitution
Excess base [mol-%]
0
5
10
20
6/(6 ± 8) [%]
7/(6 ± 8) [%]
8/(6 ± 8) [%]
9/(9 ‡ 11) [%]
11/(9 ‡ 11) [%]
18
82
18
52
30
19
81
20
51
30
21
79
25
35
41
38
62
ꢀ 0
11
89
Through these simple experiments, some valuable insights into the parasiticidal
action of QHS are obtained: 1) Random intermolecular attacks by the C-centered
radicals are not likely to be responsible for the parasiticidal activities of QHS (1), when
one takes into account the quenching by the abundant intracellular glutathione (GSH,
a good H-donor similar to cysteine in this work). Besides, there exists some built-in
ꢁself-quenchingꢀ mechanism2) (e.g., Path c in Scheme 2) that may compete even more
keenly than GSH with the intermolecular damaging processes, especially when the
cleavage is mediated by small reductants. 2) When cleaved by intracellular Fe S type
redox centers containing multi-S-ligands (i.e., bulkier chelates), the ꢁintramolecularꢀ
alkylation of the S-ligands can be quite effective; the inactivation of the redox center
due to alkylation is thus very possible. 3) Contrary to the broadly accepted notion, the
secondary C-centered radical 5 probably does not contribute to the antimalarial activity
at all, because of its inability to deliver ꢁintramolecularꢀ attacks.
This work was supported by the Chinese Academy of Sciences (CAS ꢁKnowledge Innovationꢀ Project), the
Life Science Special Fund of CAS (supported by the Ministry of Finance of China (Stz 98-3-03)), the National
2
)
Although these reactions are well-known, no one seems to have realized that they constitute radical self-
quenching devices and may be responsible for the extraordinarily low toxicity of QHS to host cells.

932
Helvetica Chimica Acta ± Vol. 84 (2001)
Natural Science Foundation of China (29832020, 20025207), and the Major State Basic Research Development
Program (G2000077502). The authors also wish to thank Prof. Yu-Lin Wu for encouragement.
REFERENCES
[1] D. Mackenzie, New Sci. 1999, 163, 2204.
[2] D. Butler, Nature (London) 1997, 386, 535.
[3] For recent reviews, see J. N. Cumming, P. Ploypradith, G. H. Posner, Adv. Pharmacol. (San Diego) 1997, 37,
253; J. A. Vroman, M. Alvim, M. A. Avery, Curr. Pharm. Des. 1999, 5, 101; A. K. Bhattacharya, R. P.
Sharma, Heterocycles 1999, 51, 1681; H. Ziffer, R. J. Highet, D. L. Klayman, Prog. Chem. Org. Nat. Prod.
1997, 72, 121.
[4] For some very recent reports, see P. M. OꢀNeill, L. P. D. Bishop, N. L. Searle, J. L. Maggs, R. C. Storr, S. A.
Ward, B. K. Park, F. Mabbs, J. Org. Chem. 2000, 65, 1578; D. J. Walker, J. L. Pitsch, M. M. Peng, B. L.
Robinson, W. Peters, J. Bhisutthibhan, S. R. Meshnick, Antimicrob. Agents Chemother. 2000, 44, 344; C. W.
Jefford, J.-C. Rossier, W. K. Milhous, Heterocycles 2000, 52, 1345; K. J. McCullough, J. K. Wood, A. K.
Bhattacharjee, Y.-X. Dong, D. E. Kyle, W. K. Milhous, J. L. Vennerstrom, J. Med. Chem. 2000, 43, 1246; W.-
M. Wu, Y.-L. Wu, J. Chem. Soc., Perkin Trans. 1 2000, 4279; D.-Y. Wang, Y.-L. Wu, J. Chem. Soc., Chem.
Commun. 2000, 2193; G. H. Posner, P. Ploypradith, M. H. Parker, H. OꢀDowd, S.-H. Woo, J. Northrop, M.
Krasavin, P. Dolan, T. W. Kensler, S. Xie, T. A. Shapiro, J. Med. Chem. 1999, 42, 4275; H.-S. Kim, K.
Tsuchiya, Y. Shibata, Y. Wataya, Y. Ushigoe, A. Masuyama, M. Norjima, K. J. McCullough, J. Chem. Soc.,
Perkin Trans. 1 1999, 1867; F. Benoit-Vical, A. Robert, B. Meunier, Antimicrob. Agents Chemother. 1999, 43,
2555; R. K. Haynes, H. H. O. Pai, A. Voerste, Tetrahedron Lett. 1999, 40, 4715.
[5] Y.-K. Wu, Z.-Y. Yue, Y.-L. Wu, Angew. Chem., Int. Ed. 1999, 38, 2580.
[6] W.-M. Wu, Y.-K. Wu, Y.-L. Wu, Z.-J. Yao, C.-M. Zhou, Y. Li, F. Shan, J. Am. Chem. Soc. 1998, 121, 3316.
[7] G. H. Posner, D.-S. Wang, J. N. Cumming, C. H. Oh, A. N. French, A. L. Bodley, T. A. Shapiro, J. Med.
Chem. 1995, 38, 2273; C. W. Jefford, M. G. H. Vicente, Y. Jacquier, F. Favarger, J. Mareda, P. Millasson-
Schmidt, G. Brunner, U. Burger, Helv. Chim. Acta 1996, 79, 1475.
[8] R. F. Jameson, W. Linert, A. Tschinkowitz, J. Chem. Soc., Dalton Trans. 1988, 2109.
[9] D.-Y. Wang, Y.-K. Wu, Y.-L. Wu, Y. Li, F. Shan, J. Chem. Soc., Perkin Trans. 1 1999, 1827.
Received January 18, 2001