Synthetic spirocyclic endoperoxides: New antimalarial scaffolds

Article abstract of DOI:10.1039/c4md00454j

Here we report the development of a straightforward synthetic procedure for the preparation of spirocyclic endoperoxides as synthetic analogues of the natural product dihydroplakortin. The peroxides presented here are more potent antiplasmodials than dihydroplakortin itself and we proved for the first time their antimalarial activity in vivo. This journal is

Full text of DOI:10.1039/c4md00454j

MedChemComm
View Article Online
View Journal
CONCISE ARTICLE
Synthetic spirocyclic endoperoxides: new
antimalarial scaffolds†
Cite this: DOI: 10.1039/c4md00454j
abc
Margherita Brindisi,abc Sandra Gemma,
Sanil Kunjir,abc Luisa Di Cerbo,abc
*
Simone Brogi,abc Silvia Parapini,acd Sarah D'Alessandro,acd Donatella Taramelli,acd
Annette Habluetzel,ce Sofia Tapanelli,ce Stefania Lamponi,ab Ettore Novellino,af
abc
*
Giuseppe Campiani
and Stefania Butiniabc
Received 8th October 2014
Accepted 11th November 2014
Here we report the development of a straightforward synthetic procedure for the preparation of spirocyclic
endoperoxides as synthetic analogues of the natural product dihydroplakortin. The peroxides presented
here are more potent antiplasmodials than dihydroplakortin itself and we proved for the first time their
antimalarial activity in vivo.
DOI: 10.1039/c4md00454j
www.rsc.org/medchemcomm
Cyclic peroxides such as 1,2-dioxolanes, 1,2,4-trioxanes and 1,2- properties.11 We also reported the development of novel endo-
dioxanes are a class of organic compounds with interesting peroxides as synthetic analogues of DHP characterized by a
pharmacological properties and widely represented in nature. bicyclic tetrahydrofuro[2,3-c][1,2]dioxane core. The synthetic
Artemisinin (ART) is an endoperoxide-based natural product, peroxide 2 (Fig. 1) was the most potent compound of the series
which is highly effective against clinically relevant P. falciparum and showed higher antiplasmodial potency than dihydro-
strains responsible for human malaria. Currently, the so-called plakortin. The major limitation of these analogues was their
artemisinin-based combination therapies (ACT) are employed synthesis as an inseparable mixture of diastereoisomers.7 The
as rst line treatment in most malaria endemic countries, structure–activity relationships (SARs) for this class of
adhering to WHO recommendations.1–3 However, lower compounds were rationalized on the basis of the mechanism of
susceptibility to artemisinins is being reported from highly action proposed for ART.
malaria endemic regions.4 So, novel peroxides characterized by
Although the mechanism of action of ART is still debated,12 it
different structural features could delay the potential selection seems that the peroxide bond is activated within the food
of P. falciparum resistant strains.5 Moreover, the cost associated vacuole of the parasite, a specialized organelle accommodating
with the extraction of this drug or with synthetic precursors the hydrolysis of hemoglobin, a necessary source of amino acids
from natural sources prompted researchers to develop synthetic for the parasites. During this process, free Fe(II)-heme is
peroxides as low-cost alternatives to ART.6
For several years we have been engaged in the design of
suitable synthetic strategies for the preparation of antimalarial
endoperoxides.7–10 In particular we performed the rst stereo-
selective synthesis of 9,10-dihydroplakortin (DHP, 1, Fig. 1),10
a
natural product endowed with interesting antiplasmodial
aEuropean Research Centre for Drug Discovery and Development (NatSynDrug),
University of Siena, via Aldo Moro 2, 53100, Siena, Italy. E-mail: gemma@unisi.it;
campiani@unisi.it
bDip. di Biotecnologie, Chimica e Farmacia, University of Siena, via Aldo Moro 2,
53100, Siena, Italy
cCentro Interuniversitario di Ricerche sulla Malaria (CIRM), University of Perugia,
Perugia, Italy
d
`
Dip. di Scienze Farmacologiche e Biomolecolari, Universita di Milano, Via Pascal 36,
20133 Milan, Italy
eSchool of Pharmacy, University of Camerino, Piazza Cavour 19F, 62032 Camerino,
Italy
fDip. di Farmacia, University of Naples Federico II, Via D. Montesano 49, 80131
Naples, Italy
† Electronic supplementary information (ESI) available: Scheme S1, Fig. S1 and
experimental details. See DOI: 10.1039/c4md00454j
Fig. 1 Structure of reference (1 and 2) and title (3–11) endoperoxides.
This journal is © The Royal Society of Chemistry 2014
Med. Chem. Commun.

View Article Online
MedChemComm
Concise Article
released and reacts with the peroxide bond of ART to form an endoperoxide 10 is reported in the ESI.† The synthesis of
O-centered radical, that probably undergoes a 1,5-H shi to intermediates 12a,d and 13c is described in Scheme 2. 1-Ada-
form a highly toxic C-centered radical,13,14 ultimately respon- mantylaldehyde 16a and cyclohexanone 16b were submitted to
sible for the antiplasmodial activity. The formation of C- the Horner–Wadsworth–Emmons olenation reaction to afford
centered radicals has also been proposed for several synthetic the a,b-unsaturated esters 17a and 17b. DIBAL-mediated
endoperoxides, comprising DHP.15–17 For our synthetic perox- reduction of the above mentioned esters 17a and 17b and of
ides, we found that subtle modications of the substituent at commercially available ester 17c afforded the corresponding
C3, affecting the distance between the C5 of the lateral chain primary alcohols that were subsequently converted into iodides
˚
and the O2 of the peroxide bridge (C5–O2 distance <3 A for 18a–c. While iodides 18a and 18c were obtained by treatment of
optimal activity), had a strong impact on the antiplasmodial alcohols with iodine, triphenylphosphine, and imidazole,
potency, in line with the mode of action of ART and DHP.7 As a iodide 18b could not be obtained by using the same protocol.
continuation of our previous work, we were interested in 18b was instead obtained by treatment of the alcohol precursor
developing suitable synthetic strategies aimed at modifying the with potassium iodide in the presence of boron triuoride.
lateral chain at C3 of the bicyclic skeleton (3–8) and the size of Iodides 18a–c were next used to alkylate lactone 19 using
the rings forming the bicyclic system (9–11) reducing the lithium hexamethyldisilylamide (LiHMDS) as the base to afford
number of chiral centers. In particular, we were interested in 12a,d and 13c. Intermediates 12b and 12c, bearing cyclohexyl
the synthesis of spiro-derivatives 6–8 that lack the stereogenic and cyclopentyl substituents, were prepared as reported in
center at C3 and represent a further simplication of this class Scheme 3. The synthesis of an analogue bearing an adamantyl-
of compounds.
radical at the same position was also attempted without
As described in Scheme 1, a three-step reaction protocol was success. Accordingly, cyclohexyl and 1-adamantyl derivatives
employed for the assembly of the bicyclic core starting from 20a,b were treated with cerium chloride as a Lewis acid and
olen intermediates 12a–d and 13a–c. The rst step of the sodium hypochlorite as the oxidizing agent to afford chlor-
protocol consisted of the Mukaiyama peroxysilylation reac- omethyl derivatives 21a,b in reasonable yields. Intermediates
tion18,19 of the double bond of intermediates 12a–d and 13a–c to 21a,b were in turn converted into the corresponding iodides
afford peroxides 14a–d and 15a–c, respectively. The Mukaiyama 22a,b. Intermediate 22c was synthesized starting from 25, in
reaction proceeds regioselectively at the more substituted turn prepared following a different synthetic approach by
carbon of the double bond. Subsequently the lactone func- reacting Grignard reagent 23, formed in situ, with propargyl
tionality was reduced to the corresponding lactol by using dii- alcohol 24.20 The resulting primary alcohol 25 was then con-
sobutylaluminum hydride (DIBAL). Finally, trimethylsilyl verted to 22c via a mesyl ester intermediate that was sonicated
triate (TMSOTf)-promoted simultaneous deprotection and in the presence of lithium iodide to afford the desired iodide
cyclization of the resulting lactols afforded the cyclised endo- 22c in good overall yield. Iodides 22a–c were used as alkylating
peroxides 3–5, isolated as inseparable mixtures of diastereo- agents for the enolate of lactone 19. While 22a,c smoothly fur-
isomers, and spiroperoxides 6–9 obtained as racemates. nished the corresponding alkylation products, the same reac-
Following a slightly modied protocol, the synthesis of tion performed with the adamantyl derivative 22b failed,
probably due to its higher steric hindrance. For the synthesis of
intermediates 13a and 13b (Scheme 4), alcohol intermediates
27a and 27b, in turn obtained by aldol condensation of alde-
hydes 26a and 26b with 19 in the presence of lithium diiso-
propylamide (LDA), were dehydrated to the corresponding
Scheme
2 Synthesis of intermediates 12a,d and 13c. [Reaction
conditions: (a) triethyl 2-phosphonopropionate (for 17a), or triethyl
phosphonoacetate (for 17b), NaH, THF, 0 ꢀC, 1 h, then 25 ꢀC, 2 h; (b)
Scheme
1
General synthesis of endoperoxides 3–9. [Reaction DIBAL, DCM, ꢁ78 ꢀC, 1.5 h; (c) I2, PPh3, imidazole, 1 : 3 MeCN/Et2O,
conditions: (a) Co(thd)2, Et3SiH, O2, t-BuOOH (5 M in nonane), 1,2-
0
ꢀC, 30 min (from 17a and 17c) or KI, BF3$Et2O, dioxane, 25 ꢀC, 3 h
dichloroethane (1,2-DCE), 25 ꢀC, 4 h; (b) DIBAL, DCM, ꢁ78 ꢀC, 1.5 h; (c) (from 17b); (d) (i). 19, LiHMDS, THF, ꢁ78 ꢀC, 30 min; (ii). 17a–c, ꢁ78 ꢀC,
TMSOTf, DCM, ꢁ78 ꢀC, 5 min].
30 min, then 25 ꢀC, 2 h].
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2014

View Article Online
Concise Article
MedChemComm
soware.22,23 The endoperoxides reported here have been
produced and tested as mixtures of diastereoisomers (3–5) or as
racemates (6–11), while docking studies were performed for
each enantiomer. As expected, heme is unable to discriminate
between enantiomers24 and no differences in the binding mode
of each enantiomeric couple were observed. Based on our
molecular docking calculations, we found a good correlation
between the distance of oxygen atoms from the reactive heme
iron center and the antimalarial activity (Table 1). In fact, the
compounds that display a larger distance of the peroxide moiety
from the iron center such as compounds 6, 9–11 were found less
potent in vitro than compounds 3–5, 7, 8 exhibiting a shorter
distance. For example, peroxide 4 (one of the most active
compounds in the set) displays a very short distance between
the peroxide oxygen atoms and the iron center (Table 1) and
both atoms could be able to coordinate the heme iron. In
contrast, the inactive analogue 9 does not possess the appro-
priate distance to interact with heme iron.
Scheme 3 Synthesis of intermediates 12b and 12c. [Reaction condi-
tions: (a) CeCl3$7H2O, 10–13% NaOCl, 1 : 1 DCM/H2O, 25 ꢀC, 14 h; (b)
NaI, acetone, 25 ꢀC, 18 h; (c) CuI, THF, 25 ꢀC 24 h, 80 ꢀC, 5 h; (d) MsCl,
Et3N, Et2O, 0 ꢀC, 12 h; (e) LiI, Et2O, sonication, 25 ꢀC, 3 h; (f) (i). 19,
LiHMDS, THF, ꢁ78 ꢀC, 30 min; (ii) 22a and 22c, ꢁ78 ꢀC, 30 min, then 25
ꢀC, 2 h].
For comparison, distances calculated for the reference
compounds ART and DHP are also in good agreement with a
correct interaction with heme (Table 1 and Fig. S1 in the ESI†).
Also the free-binding energy21 (DGbind) calculated for all
complexes (Table 1), displayed a good correlation with anti-
plasmodial potency, being around ꢁ20 kcal molꢁ1 for inactive
compounds (9), comprised between ꢁ20 and ꢁ25 kcal molꢁ1 for
compounds in the micromolar range (6, 10, 11), and <ꢁ25 kcal
molꢁ1 for compounds in the low micromolar range (1, 3–5, 7, 8).
It is worth noting that ART presents a calculated DGbind well
below the value found for the spiroperoxides presented here,
consistent with its nanomolar antiplasmodial potency.
Compound 8, displaying an estimated DGbind of ꢁ30.69 kcal
molꢁ1, is the most potent plakortin-related endoperoxide
described to date.11 Accordingly, both epimers of compounds 4
and 5 (Fig. 2A–D) interact with heme in a similar fashion.
However, the larger steric hindrance of the cyclohexyl ring of 4
with respect to the cyclopentyl ring of 5 results in a stronger
hydrophobic interaction with heme for 4. For the couple of
analogues 6 and 7 (Fig. 3A and B), the steric hindrance of the
cyclohexyl ring physically hampers the accommodation of 6 for
a strong interaction with Fe(II)-heme (Fig. 2A), while the cyclo-
pentyl ring of 7 (Fig. 3B) has a reduced steric hindrance and is
well tolerated allowing both O2 and O1 of the peroxide system to
interact with Fe(II)-heme. Docking studies (Fig. 2 and 3) high-
lighted that subtle differences in the structure of the peroxides
and in particular of the size and shape of their side chains are
responsible for a ne-tuning of the antiplasmodial activity. On
the other hand, the bulky adamantane ring (8, Fig. 3C) is able to
maximize hydrophobic interactions with the planar protopor-
phyrin-IX ring while maintaining a correct orientation of both
the peroxide oxygens for their interaction with Fe(II)-heme, thus
resulting in an overall improved DGbind and, as a consequence,
higher antiplasmodial potency. Finally, the seven-membered
ring of compound 9 (Fig. 3D) is not tolerated since it hampers a
favourable conformation for reaching the right distance to the
iron atom contained in the protoporphyrin-IX ring (Fig. 3D).
Spiroperoxide 8, the most potent antiplasmodial agent of the
series, was submitted to a preliminary in vivo evaluation in the
Scheme 4 Synthesis of intermediates 13a and 13b. [Reaction condi-
tions: (a) 19, LDA, ꢁ78 ꢀC, 20 min, then 25 ꢀC, 8 h; (b) SOCl2, pyridine,
25 ꢀC, 2 h].
olens. Finally, dioxolane rac-11 was synthesized as described
in Scheme 5. The carbonyl group of the commercially available
ketone 28 was reduced using sodium borohydride and the
double bond was hydroperoxysilylated to 29 under Mukaiyama
reaction conditions. Treatment of 29 with methanesulfonyl
chloride in the presence of triethylamine directly afforded the
cyclization product rac-11.
The antiplasmodial activity of the novel endoperoxides 3–11
was tested against two laboratory P. falciparum strains, the
chloroquine-sensitive (CQ-S) D10 and the chloroquine-resistant
(CQ-R) W2, according to the described procedures.7 The
resulting IC50 values are reported in Table 1.
To understand the modulation of antiplasmodial potency
observed for the synthesized compounds, we in depth analyzed
the interaction of the peroxide moiety of compounds 4–11 with
Fe(II)-heme (which is the rst and the key step for the formation
of the toxic radical species) by molecular docking analysis
employing Glide soware, aer structure optimization was
performed by using ab initio calculation employing the Jaguar
Scheme 5 Synthesis of rac-11. [Reaction conditions: (a) NaBH4, EtOH,
0
ꢀC, 1 h, then 25 ꢀC, 12 h; (b) Co(thd)2, Et3SiH, O2, t-BuOOH (5 M in
nonane), 1,2-DCE, 25 ꢀC, 4 h; (c) MsCl, Et3N, DCM, 0 ꢀC, 2 h].
This journal is © The Royal Society of Chemistry 2014
Med. Chem. Commun.

View Article Online
MedChemComm
Table 1 Antiplasmodial activity, calculated O1, O2/Fe(II)-heme distancesb, and DGbind (kcal molꢁ1 c
Concise Article
)
D10
W2
IC50 (mM)a
O2–Feb
O1–Feb
DGbind (kcal molꢁ1
)
c
Cpd
Structure
2.27 (E1)d
2.21 (E2)e
2.13 (E1)
2.16 (E2)
ꢁ29.33
ꢁ30.48
3
0.85
0.39
0.39
2.25 (E1)
2.20 (E2)
2.22 (E1)
2.22 (E2)
ꢁ30.28
ꢁ29.25
4
5
0.14
0.8
2.31 (E1)
2.34 (E2)
2.24 (E1)
2.32 (E2)
ꢁ27.30
ꢁ26.55
1.2
6
7
5.3
1.1
3.89
2.81
3.03
2.24
ꢁ23.82
ꢁ28.43
0.84
0.47
8
0.61
>20
0.16
2.18
3.36
2.40
2.16
3.90
2.99
ꢁ30.69
ꢁ19.76
ꢁ24.46
9
>17.0
10
8.5
>25
11
7.2
3.2
2.34
3.22
ꢁ22.64
1 (ref. 7)
ART
CQ
0.86
0.029
0.015
0.44
0.012
0.228
2.20
2.14
—
2.27
2.10
—
ꢁ31.04
ꢁ43.13
—
a
b
IC50 values are the mean of at least three determinations in duplicate. Standard errors were all within 10% of the mean. The distances were
˚
calculated and reported in A by means of the measurement tool implemented in the Maestro graphical interface from the poses obtained by
c
d
e
docking calculations. The binding energy was calculated by means of Prime MM-GSBA.21 E1 ¼ epimer 1: 3R*, 4aS*, 7aR*. E2 ¼ epimer 2:
3S*, 4aS*, 7aR*.
P. berghei mouse model of malaria using the Peters 4 day test.26 strains. Being heme an achiral target, compounds were tested
Aer administration of 8 at a daily dose of 100 mg kgꢁ1 (i.p.), as racemates since no difference on antiplasmodial activity
mice showed a 72.5% reduction in parasitaemia. Notably, for each enantiomer could be expected.24 Molecular model-
compound 8 did not elicit, at the concentration used in the in ling studies highlighted key molecular features responsible
vivo study, any sign of toxicity in mice, consistently with its for activity and potency, and will be useful for further opti-
determined in vitro toxicity against mouse broblasts NIH3T3 mizing their in vitro potency. Moreover, we have herein for the
(TC50 ¼ 150 mM).
rst time provided evidence that plakortin-derived synthetic
In conclusion, we herein presented the development of a analogues do have antimalarial activity in vivo. The straight-
straightforward synthetic procedure for the preparation of forward synthetic approach to spirocyclic endoperoxides may
spiroperoxides as racemic mixtures endowed with anti- pave the way to the development of more potent analogues
plasmodial activity against P. falciparum CQ-S and CQ-R in vivo.
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2014

View Article Online
Concise Article
MedChemComm
2 M. Njoroge, N. M. Njuguna, P. Mutai, D. S. Ongarora,
P. W. Smith and K. Chibale, Chem. Rev., 2014, DOI:
10.1021/cr500098f.
3 http://www.who.int/malaria/publications/atoz/9789241500838/
en/, access date: 28/07/2014.
4 E. A. Ashley, M. Dhorda, R. M. Fairhurst, C. Amaratunga,
P. Lim, S. Suon, S. Sreng, J. M. Anderson, S. Mao, B. Sam,
C. Sopha, C. M. Chuor, C. Nguon, S. Sovannaroth,
S. Pukrittayakamee, P. Jittamala, K. Chotivanich,
K. Chutasmit, C. Suchatsoonthorn, R. Runcharoen, et al.,
N. Engl. J. Med., 2014, 371, 411–423.
5 F. Ariey, B. Witkowski, C. Amaratunga, J. Beghain,
A. C. Langlois, N. Khim, S. Kim, V. Duru, C. Bouchier,
L. Ma, P. Lim, R. Leang, S. Duong, S. Sreng, S. Suon,
C. M. Chuor, D. M. Bout, S. Menard, W. O. Rogers,
B. Genton, et al., Nature, 2014, 505, 50–55.
6 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. Padmanilayam, J. Santo Tomas,
C. Scheurer, B. Scorneaux, Y. Tang, H. Urwyler, S. Wittlin
and W. N. Charman, Nature, 2004, 430, 900–904.
7 S. Gemma, S. Kunjir, S. S. Coccone, M. Brindisi, V. Moretti,
S. Brogi, E. Novellino, N. Basilico, S. Parapini, D. Taramelli,
G. Campiani and S. Butini, J. Med. Chem., 2011, 54, 5949–
5953.
Fig. 2 Docked poses of both epimers of 4 (panels A and B in gray
sticks) and 5 (panels C and D) in complex with heme (green balls and
sticks), charged iron was colored cyan and represented by the CPK
model. The potential ligand–metal coordination bonds are reported as
black-green dotted lines. The picture was generated by Maestro.25
8 S. Gemma, F. Marti, E. Gabellieri, G. Campiani, E. Novellino
and S. Butini, Tetrahedron Lett., 2009, 50, 5719–5722.
9 S. Gemma, S. Kunjir, M. Brindisi, E. Novellino, G. Campiani
and S. Butini, Tetrahedron Lett., 2013, 54, 1233–1235.
10 S. Gemma, E. Gabellieri, S. S. Coccone, F. Marti,
O. Taglialatela-Scafati, E. Novellino, G. Campiani and
S. Butini, J. Org. Chem., 2010, 75, 2333–2340.
11 E. Fattorusso, S. Parapini, C. Campagnuolo, N. Basilico,
O. Taglialatela-Scafati and D. Taramelli, J. Antimicrob.
Chemother., 2002, 50, 883–888.
12 S. Krishna, S. Pulcini, C. M. Moore, B. H. Teo and
H. M. Staines, Trends Pharmacol. Sci., 2014, 35, 4–11.
13 J. N. Cumming, P. Ploypradith and G. H. Posner, Adv.
Pharmacol., 1997, 37, 253–297.
14 G. H. Posner, J. N. Cumming, P. Ploypradith and
H. O. Chang, J. Am. Chem. Soc., 1995, 117, 5885–5886.
15 C. W. Jefford, Curr.Top. Med. Chem., 2012, 12, 373–399.
16 O. Taglialatela-Scafati, E. Fattorusso, A. Romano, F. Scala,
V. Barone, P. Cimino, E. Stendardo, B. Catalanotti,
M. Persico and C. Fattorusso, Org. Biomol. Chem., 2010, 8,
846–856.
Fig. 3 Docked poses of 6–9 (gray sticks; A, B, C, and D, respectively) in
complex with heme (green balls and sticks), charged iron was colored
cyan and represented by the CPK model. The potential ligand–metal
coordination bonds are reported as black-green dotted lines. The
picture was generated by Maestro.25
17 F. Bousejra-El Garah, M. H. Wong, R. K. Amewu,
S. Muangnoicharoen, J. L. Maggs, J. L. Stigliani, B. K. Park,
J. Chadwick, S. A. Ward and P. M. O'Neill, J. Med. Chem.,
2011, 54, 6443–6455.
Acknowledgements
The authors thank MIUR for nancial support. S.B. thanks S-IN
Soluzioni Informatiche for technical support.
18 S. Isayama and T. Mukaiyama, Chem. Lett., 1989, 573–576.
19 P. M. O'Neill, S. Hindley, M. D. Pugh, J. Davies, P. G. Bray,
B. K. Park, D. S. Kapu, S. A. Ward and P. A. Stocks,
Tetrahedron Lett., 2003, 44, 8135–8138.
20 F. Barile, M. Bassetti, A. D'Annibale, R. Gerometta and
M. Palazzi, Eur. J. Org. Chem., 2011, 6519–6526.
Notes and references
1 World Health Organization, World Malaria Report, 2013,
Geneva.
This journal is © The Royal Society of Chemistry 2014
Med. Chem. Commun.

View Article Online
MedChemComm
Concise Article
21 P. D. Lyne, M. L. Lamb and J. C. Saeh, J. Med. Chem., 2006, 24 C. W. Jefford, Drug Discovery Today, 2007, 12, 487–495.
¨
49, 4805–4808.
22 Glide, version 5.7, Schrodinger, LLC, New York, NY, 2011.
25 Maestro, version 9.2, Schrodinger, LLC, New York, NY, 2011.
¨
26 D. J. Knight and W. Peters, Ann. Trop. Med. Parasitol., 1980,
74, 393–404.
¨
23 Jaguar, version 7.8, Schrodinger, LLC, New York, NY, 2011.
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2014