Evidence for the importance of high-valent Fe=O and of a diketone in the molecular mechanism of action of antimalarial trioxane analogs of artemisinin

Full text of DOI:10.1021/ja954131p

J. Am. Chem. Soc. 1996, 118, 3537-3538
3537
Scheme 1
Evidence for the Importance of High-Valent FedO
and of a Diketone in the Molecular Mechanism of
Action of Antimalarial Trioxane Analogs of
Artemisinin
Gary H. Posner,*,† Sheldon B. Park,† Llu¨ısa Gonza´lez,†
Dasong Wang,† Jared N. Cumming,†,‡ Donna Klinedinst,§
Theresa A. Shapiro,§ and Mario D. Bachi|
A
Department of Chemistry, School of Arts and
Sciences, The Johns Hopkins UniVersity
Baltimore, Maryland 21218
Department of Medicine, School of Medicine
The Johns Hopkins UniVersity
Baltimore, Maryland 21205
Department of Organic Chemistry
The Weizmann Institute of Science, RehoVot, Israel
ReceiVed December 8, 1995
B
Ancient Chinese folk medicine has led relatively recently to
chemical identification of the 1,2,4-trioxane sesquiterpene
artemisinin (qinghaosu, 1), representing a new class of fast-
acting, clinically useful, antimalarial drugs.1,2 This natural
endoperoxide and some related synthetic organic endoperoxides,
causing oxidative stress to malaria parasites,3 are reduced by
the iron-rich parasites to form cytotoxic radical intermediates.4
We have designed and synthesized some structurally simplified
1,2,4-trioxanes to determine whether they are reduced by ferrous
iron to form the same kinds of radical intermediates as formed
upon iron(II) reduction of artemisinin.5,6 The simplified triox-
anes we report here were formulated such that an intermediate
carbon-centered radical would be intercepted before â-scission
of Fe(III)-O• [TFe(IV)dO] either by a competing â-scission
of a better radical leaving group X• from radical intermediate
2a (Scheme 1A) or by a competing subsequent 1,5-hydrogen
atom shift in radical intermediate 3a (Scheme 1B); such
mechanistically competing reactions were expected to interrupt
formation of a high-valent iron-oxo species and thus to
diminish or to undermine completely the antimalarial activity
of these artemisinin analogs. Also, another analog was designed
to have improved and therefore medicinally more desirable
antimalarial potency by incorporating a structural feature that
would facilitate â-scission of a high-valent iron-oxo species.
We report here in Vitro antimalarial activities that support the
intermediacy of such high-valent iron-oxo species Via the
â-scission pathway in the antimalarial artemisinin analogs shown
in Table 1. We report evidence also of a simple diketone as an
unexpectedly antimalarial product formed upon iron(II) reduc-
tion of a trioxane.
Table 1. Chemical Structure-Antimalarial Activity Relationships
in Chloroquine-Sensitive Plasmodium falciparum (NF54) Parasites
in Vitroa
10-15
4-8
IC50
trioxane
X
R
ng/mL
3.0
86
>1000
98
nM
8.3
240
4
5
6
7
8
Ph
CH2CH2OH
Me3Si CH2CH2OH
Me3Sn CH2CH2OH
Ph
>2500
H
H
340
Me3Sn
>1000
>2500
10
11
12
13
14
15
Me
Et
220
170
46
35
11
960
710
160
110
46
(CH3)2CHCH2CH2
PhCH2CH2CH2
CH2dCH
Ph
11
38
artemisinin (1)
arteether
chloroquine
3.0
1.9
2.6
11
6.1
5.0
a Antimalarial activity was determined as reported previously.5 The
standard deviation for each set of quadruplicates was e45% of the
mean. R2 values for the curves were g0.993.
† Department of Chemistry, The Johns Hopkins University.
‡ Recipient of a Graduate Fellowship from the Organic Chemistry
Division of the American Chemical Society (1995-1996) sponsored by
Abbott Laboratories.
Trioxanes 5 and 6 differ by only one atom: silicon Vs tin.
Despite this small structural difference within a complex organic
endoperoxide, C4â-silyl analog 5 has significant antimalarial
activity, whereas tin analog 6 is virtually inactive (Table 1).
Likewise, 8a-unsubstituted C4â-stannyl analog 8 has virtually
no antimalarial activity, whereas similar C4â-substituted analog
7 is active. These results are consistent with selective â-scission
of Me3Sn• [rather than Fe(III)-O•] from the intermediate C4-
carbon radical 2a (X ) Me3Sn, Scheme 1A). Also, the absence
(<0.1%) of hexamethyl Dewar benzene (HMDB) rearrangement
during FeBr2 reduction of C4â-stannyl trioxane 8 is consistent
with the absence of a high-valent iron-oxo intermediate.6 As
a control to show that the presence of a tin atom does not
necessarily destroy a trioxane’s antimalarial activity, tin-
containing dihydroartemisinin analog 9 was synthesized7 and
was found to have measurable in Vitro antimalarial activity (IC50
) 750 ng/mL, 1300 nM).
§ Department of Medicine, The Johns Hopkins University.
| The Weizmann Institute of Science.
(1) For reviews, see: (a) Klayman, D. L. Science 1985, 228, 1049. (b)
White, N. J. Trans. R. Soc. Trop. Med. Hyg. 1994, 88 (Suppl. 1), 1. (c)
Zhou, W.-S.; Xu, X.-X. Acc. Chem. Res. 1994, 27, 211. (d) Jung, M. Curr.
Med. Chem. 1994, 1, 45. (e) Meshnick, S. R.; Taylor, T. E.; Kamchon-
wongpaisan, S. Microbiol. ReV., in press. (f) Cumming, J. N.; Ploypradith,
P.; Posner, G. H. AdV. Pharmacol. 1996, 37.
(2) Avery, M. A.; Gao, F.; Chong, W. K. M.; Hendrickson, T. F.; Inman,
W. D.; Crews, P. Tetrahedron 1994, 50, 957 and references therein.
(3) Free Radicals in Tropical Diseases; Aruoma, O. I., Ed.; Harwood
Academic: Chur, Switzerland, 1993.
(4) (a) Meshnick, S. R.; Yang, Y.-Z.; Lima, V.; Kuypers, F.; Kamchon-
wongpaisan, S.; Yuthavong, Y. Antimicrob. Agents Chemother. 1993, 37,
1108. (c) Hong, Y.-L.; Yang, Y.-Z.; Meshnick, S. R. Mol. Biochem. Parsitol.
1994, 63, 121.
(5) Posner, G. H.; Wang, D.; Cumming, J. N.; Oh, C. H.; French, A. N.;
Bodley, A. L.; Shapiro, T. A. J. Med. Chem. 1995, 38, 2273. See also:
Bloodworth, A. J.; Shah, A. Tetrahedron Lett. 1995, 36, 755l.
(6) Posner, G. H.; Cumming, J. N.; Ploypradith, P.; Oh, C. H. J. Am.
Chem. Soc. 1995, 117, 5885.
(7) All new compounds were fully characterized spectroscopically and
by HRMS and/or combustion analysis.
0002-7863/96/1518-3537$12.00/0 © 1996 American Chemical Society

3538 J. Am. Chem. Soc., Vol. 118, No. 14, 1996
Communications to the Editor
3-phenyltrioxane 1513 also has high antimalarial activity, its
reduction by FeBr2 unexpectedly does not generate high-valent
FedO (i.e., no rearangement of HMDB) but does generate an
antimalarial product (diketone 17). The two major products of
FeBr2 reduction of 3-phenyltrioxane 15 are ring-contracted
tetrahydrofuran 16 and 1,5-diketone 17, formed plausibly Via
electron transfer to oxygen-1 of trioxane 15 (eq 2).7 Because
some dicarbonyl compounds are known to alkylate proteins,14
as does artemisinin,15 the in Vitro antimalarial activity of 1,5-
diketone 17 was determined (IC50 ) 330 ng/mL, 1400 nM).
This is the first time that any diketone has been reported to
haVe antimalarial actiVity;16 the low antimalarial activity of
diketone 17 may be due to its relative difficulty in reaching the
malaria parasite inside the red blood cell. Thus, phenyltrioxane
15 may be acting in part as a prodrug for release of 1,5-diketone
17 inside the malaria parasite.1f
Further experimental support for these mechanistic conclu-
sions comes from the ferrous bromide induced transformation
of C4â-stannyl analog 8 into two major non-tin-containing
olefinic products (eq 1). Although formation of non-tin-
containing terminal olefin 8b might result from a concerted
radical fragmentation8 initiated by electron donation from Fe-
(II) to oxygen-1 in trioxane 8, formation of exocyclic olefin
8a, isolated in 15% yield, must arise from â-scission of Me3-
Sn• from its C4-radical precursor.9 Thus, at least 15% of the
Fe(II) reduction of trioxane 8 proceeds Via the antimalarially
crucial pathway involving C4-radical 2a (Scheme 1A).10
C3-substituted analogs 12 and 13, in which a 1,5-hydrogen
atom shift from the secondary C4-radical intermediate 3a
(Scheme 1B) could occur to form tertiary radical 3b (from
analog 12) or even considerably more stable11 secondary
benzylic radical 3b (from analog 13), both have comparable
and considerable antimalarial activity. Thus, competing sub-
sequent 1,5-hydrogen atom shifts12 apparently are not fast
enough to intercept the key C4-radical intermediates before
â-scission of Fe(III)-O• can occur. The antimalarial activities
of analogs 12 and 13, being much higher than those of C3-
methyl and C3-ethyl analogs13 10 and 11 in which no subsequent
1,5-H shift is possible (Table 1), may be due to the higher
lipophilicity of analogs 12 and 13 and therefore possibly also
to their better transport in biological systems. Both C3-
substituted analogs 12 and 13 react with ferrous bromide and
excess HMDB in THF to form hexamethylbenzene, a rear-
rangement characteristic of a high-valent iron-oxo intermedi-
ate.6 In a control reaction, HMDB in THF was not rearranged
when (t-BuO)2 was reduced by FeBr2.
In conclusion, the results summarized in eq 1 show that when
a group like Me3Sn, being a better radical leaving group9 than
Fe(III)-O•, is situated as in trioxane 8, then generation of Fe-
(III)-O• is precluded (cf. Scheme 1A). The absence of
antimalarial activity of tin-containing compounds 6 and 8, in
contrast to the considerable antimalarial activity of structurally
similar compounds5 4, 5, and 7, provides the first evidence
supporting the central role of a biologically relevant17 high-
valent iron-oxo species in the mode of action of antimalarial
analogs of the natural trioxane artemisinin.6 Also, the first
example is provided of any 1,5-diketone having antimalarial
activity.18,19
Acknowledgment. We thank the NIH (grants AI-34885 and NCRR,
OPD-GCRC RR00722) and the Burroughs Wellcome Fund for financial
support, the Canadian Government for an NSERC postdoctoral fel-
lowship to S.B.P., the Generalitat de Catalunya for a postdoctoral
fellowship to L.G., and the Hopkins-Weizmann Exchange Program for
support of Mario D. Bachi during a stay in Baltimore in 1994.
To facilitate â-scission of Fe(III)-O• from the C4-radical
intermediate, leading to formation of a C3-C4 carbon-carbon
double bond,9b a structural feature that would stabilize such a
new olefinic bond was incorporated. C3-vinyl analog 14,7 in
which the new olefinic bond can be stabilized by conjugation
with the unsaturated C3-substituent, is considerably more
antimalarially active than the corresponding C3-ethyl analog
11.13b Ferrous bromide induced reduction of peroxide 14 causes
rearrangement of HMDB into hexamethylbenzene.6 Although
Supporting Information Available: Spectra of 6, 8, 8a, 8b, 9,
12-14, 16, and 17 (20 pages). This material is contained in many
libraries on microfiche, immediately follows this article in the microfilm
version of the journal, can be ordered from the ACS, and can be
downloaded from the Internet; see any current masthead page for
ordering information and Internet access instructions.
JA954131P
(14) Pyle, S. J.; Amaranth, V.; Graham, D. G.; Anthony, D. C. J.
Neuropathol. Exp. Neurol. 1992, 51, 451. See also: Fishwick, J.; McLean,
W. G.; Edwards, G.; Ward, S. A. Chem.-Biol. Interact. 1995, 96, 263.
(15) (a) Yang, Y.-Z.; Asawamahasakda, W.; Meshnick, S. R. Biochem.
Pharmacol. 1993, 46, 336. (b) Asawamahasakda, W.; Benakis, A.;
Meshnick, S. R. J. Lab. Clin. Med. 1994, 123, 757. (c) Asawamahasakda,
W.; Ittarat, I.; Pu, Y.-M.; Ziffer, H.; Meshnick, S. R. Antimicrob. Agents
Chemother. 1994, 38, 1854.
(16) A peroxidic 1,6-dicarbonyl compound has been reported to have
antimalarial activity: Baker, J. T.; McChesney, J. D.; Chi, H. T. Pharm.
Res. 1993, 10, 662.
(17) For some recent studies, see: (a) Dexter, A. F.; Hager, L. P. J. Am.
Chem. Soc. 1995, 117, 817. (b) Minisci, F.; Fontant, F.; Araneo, S.;
Recupero, F.; Banfi, S.; Quici, S. J. Am. Chem. Soc. 1995, 117, 226. (c)
Tian, Z.-Q.; Richards, J. L.; Traylor, T. G. J. Am. Chem. Soc. 1995, 117,
21.
(8) Curran, D. P.; van Elburg, P. A.; Giese, B.; Gilges, S. Tetrahedron
Lett. 1990, 31, 2861. These mechanistic details are being investigated further,
and results will be reported in a full paper in due course.
(9) (a) Kochi, J. K. J. Am. Chem. Soc. 1962, 84, 1193. (b) Motherwell,
W. B.; Crich, D. Free Radical Chain Reactions in Organic Synthesis;
Academic: New York, 1992; Chapter 4. (c) Davies, A. G.; Roberts, B. P.;
Tse, M.-W. J. Chem. Soc., Perkin Trans. 2 1978, 145.
(10) For example, FeBr2-induced reduction of the potent antimalarial
artemisinin leads to only 10-15% of its C4-hydroxylated product Via the
crucial C4-radical intermediate.6
(11) Approximate homolytic C-H bond dissociation energies (kcal/mol)
are as follows: primary, 98; secondary, 94.5; tertiary, 91; PhCH2-H, 85;
see: Solomons, T. W. G. Organic Chemistry, 5th ed.; John Wiley & Sons:
New York, 1992, p 266.
(12) Ceccherelli, P.; Curini, M.; Marcutullio, M. C.; Mylari, B. L.;
Wenkert, E. J. Org. Chem. 1986, 51, 1505. (b) Sejbal, J.; Klimot, J.; Vystrcil,
A. Coll. Czech. Chem. Commun. 1988, 53, 118.
(13) (a) Jefford, C. W.; Velarde, J. A.; Bernardinelli, G.; Bray, D. H.;
Warhurst, D.C.; Milhous, W. K. HelV. Chim. Acta 1993, 76, 2775. (b)
Posner, G. H.; Oh, C. H.; Gerena, L.; Milhous, W. K. Heteroat. Chem.
1995, 6, 105.
(18) Posner, G. H.; Wang, D.; Gonza´lez, L.; Tao X.; Cumming, J. N.;
Klinedinst, D.; Shapiro, T. A. Tetrahedron Lett. 1996, 37, 815. We have
recently found 1,3-dibenzoylpropane (the 1,5-diketone 9 in this reference)
to have some (albeit low) antimalarial activity.
(19) For a recent ionic mechanism discussion of artemisinin activation
by non-heme iron, see: Haynes, R. K.; Vonwiller, S. C. Tetrahedron Lett.
1996, 37, 257.