Orienting the heterocyclic periphery: A structural model for chloroquine's antimalarial activity

Article abstract of DOI:10.1039/c4cc05328a

The antimalarial drug chloroquine binds to gallium proto-porphyrin-IX in methanol and in the solid state and represents a unique drug-heme model.

Full text of DOI:10.1039/c4cc05328a

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
Orienting the heterocyclic periphery: a structural
model for chloroquine’s antimalarial activity†
Cite this: Chem. Commun., 2014,
50, 13765
Erin L. Dodd and D. Scott Bohle*
Received 10th July 2014,
Accepted 4th September 2014
DOI: 10.1039/c4cc05328a
www.rsc.org/chemcomm
The antimalarial drug chloroquine binds to gallium proto-porphyrin-IX magnetic resonance (NMR)9 and Raman spectroscopy studies.10
in methanol and in the solid state and represents a unique drug– A number of Q.M. studies have attempted to shed further light on
heme model.
these interactions,11 but, in the absence of a well defined
structural basis to begin this modeling, the results of these
Chloroquine, the potent antimalarial drug, has found new life as efforts have been ambiguous.
a chemotherapy agent.1 The origin of chloroquine’s remarkable
To solve these problems, we have developed two new soluble
antimalarial activity inspired an intense sustained medicinal models for hemozoin: the first being based on ferric meso- and
chemistry effort which has led to thousands of quinoline derivatives deutero-protoporphyrin-IX,12 and the second being a gallium
having been prepared and assayed for their antimalarial activity.2 protoporphyrin-IX model.13 Both form soluble and crystallographi-
Although chloroquine resistance is now widespread, and this limits cally characterized hemozoin-like propionate bridged dimers.
its use for single agent antimalarial therapy, its use in cancer Gallium(III) has a similar ionic radius as that of ferric iron, 0.62 Å
therapy highlights the continued uncertainty of its targets vs. 0.65 Å, respectively,14 and both share similar coordination
and of its drug action mechanisms. Unfortunately the effort chemistries of the trivalent oxidation state.15 Ga(III)/Fe(III) mimicry
to understand its anti-malarial pharmacology is haunted by has been used extensively16 to understand difficult heme and non-
1
continued absence of any precise structural data for its bio- heme biochemistry such as that in H NMR determination of the
chemical interactions.3 This is a poor position to begin efforts diastereomerically controlled axial ligation of the pyropheophorbide
to develop chloroquine and the related quinoline family of anti- A unit of chlorophyll.17 Herein we demonstrate that, in solution and
malarials as antineoplastic agents.
in the solid state, chloroquine forms a distinct complex with
Chloroquine’s antimalarial activity is widely attributed to its [Ga(III)(PPIX)]2 and that these provide the first experimental struc-
disruption of heme processing in the digestive vacuole of Plasmodia. tural model for this critical heme–drug interaction.
Warhurst’s original hypothesis4 of chloroquine activity has evolved
Chloroquine and Ga(PPIX) form a well-defined complex in
to a model of drug inhibition of heme crystallization into the solution as monitored by 1H NMR, Fig. 1 and 2, and Fig. S3–S5
ultimate product of heme processing, hemozoin or malaria (ESI†). In solution, this interaction is in dynamic equilibrium
pigment.5 The insolubility, paramagnetism, and nanocrystalline
character of hemozoin has caused considerable difficulties in
working with this unusual heme product6 and its adducts with
chloroquine. Direct observation of drug–hemozoin(or heme)
binding has not been possible and some of the best evidence
for its operation in vivo is the co-localization of radio-labeled drug
with the heme crystals.7 A heme–chloroquine complex has been
observed by UV spectroscopy,8 and binding mechanisms based
upon p–p complexation have been proposed based on nuclear
Department of Chemistry, McGill University, Montreal, Quebec, Canada.
E-mail: scott.bohle@mcgill.ca; Tel: +1 514-398-7409
† Electronic supplementary information (ESI) available. CCDC 904065. For ESI
Fig. 1 Reaction of Ga(PPIX)(OH) with chloroquine free base to give
[Ga(PPIX)(OMe)(CQ)]2 (1). For numbering scheme see Fig. S1 (ESI†).
and crystallographic data in CIF or other electronic format see DOI: 10.1039/
c4cc05328a
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun., 2014, 50, 13765--13768 | 13765

View Article Online
Communication
ChemComm
upfield shift and broadening of the signal of the methine
proton H(20) of the porphyrin ring, which rests between the
propionate groups, and a lesser shift and further broadening of
the signals of the methylene protons of the propionic acid
groups themselves. In the dimerized form observed crystallo-
graphically (see below), one of these propionic acid groups
becomes a bridging propionate and also interacts with the
terminal N of the bound chloroquine. A Job plot analysis,
Fig. S5 (ESI†), fits well to either a 2 : 2 or 1 : 1 stoichiometry
with an apparent binding constant of chloroquine to Ga(PPIX)
of Keq = 1.48(5) Â 104 MÀ1, assuming the 1 : 1 stoichiometry and
ignoring dimerization and axial ligand exchange. However, this
is at best an estimate of what is a multi-step and possibly
cooperative series of equilibria.
Upon standing or concentration, solutions of gallium(III)
protoporphyrin IX dimer and chloroquine crystallize as a 2 : 2
metalloporphyrin/chloroquine ensemble which preserves the
solution interactions. Needle-shaped crystals of the drug–dimer
complex suitable for X-ray diffraction grow well in methanol
solutions containing ratios of two or more molecules of racemic
free base chloroquine per molecule of Ga(PPIX)(OH). X-ray
diffraction results in the model shown in Fig. 3 where views
of the asymmetric unit and the key drug–porphyrin interactions
are shown. As in malaria pigment, there is an inversion center
of symmetry relating the two metalloporphyrin units, with the
two enantiomers of chloroquine selectively bound to either one
Fig. 2 Above, 1H NMR titration for CQ added to Ga(PPIX)(OH); below,
Dd of CQ quinoline ring peaks with increasing Ga(PPIX) mole fraction
corresponding to the stacked spectra shown.
that is fast on the NMR timescale, and the peaks observed are of the two chiral faces of dimer. The planes of the quinoline
the average of those of all species. Large upfield peak shifts and porphyrin rings are oblique by 14.171, and there is little
occur for the protons on the N-edge of the quinoline ring of the overlap when viewed orthogonally (Fig. 3B). Chloroquine binds
chloroquine and the protons near the terminus of the side to gallium protoporphyrin IX with three E–H drug bonds
chain show dramatic shifts as well. There is also a very large oriented to the macrocycle’s p-bonds over the N–C bonds of
Fig. 3 Crystal structure of [Ga(PPIX)(OMe)(CQ)]2. A Propionate bridged dimer generated by inversion symmetry with the two enantiomers of
chloroquine hydrogen bound to the propionate carboxylate, N(7)–O(2) and methanol solvate, N(5)–O(5). B View down the Ga–O bond perpendicular
to the porphyrin plane. Key metric parameters (Å) include: Ga–O(1) 2.010(6), Ga–O(5) 2.079(6), Ga–N(1) 2.002(6), Ga–N(2) 2.018(7), Ga–N(3) 2.011(7),
Ga–N(4) 2.027(7), O(1)–C(23) 1.259(9), O(2)–C(23) 1.254(9), O(3)–C(34) 1.244(10), O(4)–C(34) 1.240(10). R1 = 0.0642. Data measured at 112 K, thermal
ellipsoids correspond to 30% probability, hydrogens omitted for clarity.
13766 | Chem. Commun., 2014, 50, 13765--13768
This journal is ©The Royal Society of Chemistry 2014

View Article Online
ChemComm
Communication
Fig. 4 (A) Substrate as chloroquine–alkoxide complex. (B) Configuration
at the key binding site in the gallium complex, and proposed configuration
of the analogous iron(III) protoporphyrin IX hydroxide complex.
Fig. 5 Major fluorescence emission peak of chloroquine (373 nm)
decreases in intensity upon addition of Ga(PPIX)(OH). Minor peak
(417 nm) does not change. Ga(PPIX)(OH) peaks are observed due to direct
excitation of the porphyrin at the excitation wavelength.
the porphyrin’s pyrrole rings. This combination of C–H, and
N–H aromatic interactions, with long ring–ring separations,
(3.40–3.63 Å) results in a unique tilted but oriented edge
interaction with relatively minor and weak electron donor/
acceptor arene p-stacking. The quinoline ring nitrogen of the
chloroquine hydrogen bonds to a coordinated methanol on a porphyrin. Titration of Ga(PPIX)(OH) or Ga(OEP)(OMe) against
six-coordinate gallium, and there is an extensive hydrogen chloroquine in methanol gives a dramatic reduction in intensity
bonding and solvation network, Fig. S6 and S7 (ESI†).
of the 365 nm emission of chloroquine, Fig. 5, that is not evident
The inclusion of a hydrogen bonded methanol or water in titration against acetic acid alone. A weaker peak at 417 nm,
molecule in the coordination sphere of the gallium, giving it previously obscured, remains at constant intensity throughout.
an in-plane six coordinate geometry, is distinct from the solid The absence of any change in quantum yield on addition of acid
state structures of hematin anhydride (b-hematin)18 and discounts simple pH effects on the quantum yield of the chloro-
malaria pigment19 which are out of plane and five coordinate. quine in the ranges observed.
Although Ga(PPIX) forms a condensed phase that is analogous
High-spin iron(III) porphyrins are, in general, fluorescence
to malaria pigment,16f in the presence of CQ it does not form. quenchers, while gallium porphyrins are highly fluorescent
In general, the monomer and 6-coordinate species of Ga(PPIX) molecules themselves, and are currently being developed for
are considerably more soluble. In the case of 1 chloroquine use as photosensitizers in photodynamic therapy.24 Any
hydrogen bonding to the methanol will generate a stronger quenching of quinoline fluorescence emission that takes place
Ga–OMe linkage which in turn stabilizes a planar six coordinate must be due to close-range interactions between the drug and
gallium. We propose that by analogy with hematin in water, the porphyrin molecules. The reaction between the drug and the
six coordinate hydroxide-like complex in Fig. 4B would be porphyrin is slower than the excitation/emission pathway, and
stabilized by the a high field ligand driving the metal to a lower the quenching observed is therefore directly related to the
spin state. Ferric heme proteins with coordinated hydroxides are amount of drug which is complexed to metalloporphyrin in
often S = 1/2 and six coordinate.20 Alkoxide and phenoxide the solution. This effect is readily quantifiable by fitting the
antimalarials also have a high affinity for iron(III), as seen in a data to a simple linear Stern–Volmer plot, after adjusting for
halofantrine-heme structure reported recently which have Fe–O concentration of the drug and the small absorption by the
bonds between the heme and the drug.21 The geometry for Ga(PPIX)(OH) (Fig. S9, ESI†).
[Ga(PPIX)(OMe)(CQ)]2 in Fig. 3 could represent a drug–substrate
From these data we can determine an approximation of the
interaction for heme in solution, possibly as [Fe(PPIX)(H2O:CQ)]2 drug binding constant using the fluorescence intensities to be
(Fig. 4), prior to crystallization. Such an interaction would inhibit Kassociation = 6.67 Â 104 based on an assumption of 1 : 1
the growth of hemozoin, and thus account for the drug action of interaction. Photoexcitation is known to increase the basicity
chloroquine.22 This structure poises both the diethylamine and of the quinoline ring N via promotion of the stability of the
quinoline ring nitrogens in positions with hydrogen bonds to amidine tautomer, whose pKa is significantly higher,25 which
suitable acceptors. This corresponds to the expected protonation could account for the discrepancy between the NMR and
state of these antimalarials in the digestive vacuole and nicely fluorescence results. Regardless of the origin of these excited
solves the conundrum poised by the theoretical prediction of state dynamics, the practical implications for utilizing these
drug binding to the surface of the (001) growing face of malarial models in high throughput antimalarial drug discovery screens
pigment without a second proton acceptor for the quinoline ring is compelling.
nitrogen hydrogen bond.22a,23 In solution, and with a solvent
bound or associated with the iron, this is no longer a problem.
To conclude, we have determined the unambiguous structure
of the bound chloroquine–gallium(III) protoporphyrin IX reciprocal
In an effort to expand upon our solution observations to dimer complex by crystallography and established that key aspects
include biomimetic concentrations, we explored the electronic of this structure are maintained in solution. The structure includes
interactions of the species in solution, using the nascent multiple sites of binding interactions of the drug to the metallo-
fluorescent properties of both chloroquine and the gallium protoporphyrin IX species, with quinoline–porphyrin stacking
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun., 2014, 50, 13765--13768 | 13767

View Article Online
Communication
ChemComm
8 (a) T. J. Egan, W. W. Mavuso, D. C. Ross and H. M. Marques, J. Inorg.
Biochem., 1997, 68, 137–145; (b) C. H. Kaschula, T. J. Egan,
R. Hunter, N. Basilico, S. Parapini, D. Taramelli, E. Pasini and
D. Monti, J. Med. Chem., 2002, 45, 3531–3539.
9 (a) S. Moreau, B. Perly, C. Chachaty and C. Deleuze, Biochim.
Biophys. Acta, 1985, 840, 107–116; (b) I. Constantinidis and
J. D. Satterlee, J. Am. Chem. Soc., 1988, 110, 4391–4395; (c) A. Leed,
K. DuBay, L. M. B. Ursos, D. Sears, A. C. de Dios and P. D. Roepe,
Biochemistry, 2002, 41, 10245–10255.
interactions and two sites of hydrogen bonding interactions
between each drug–porphyrin subunit leading to a very stable
structure in which Van der Waals interactions with the porphyrin
itself, rather than the central metal, dominate the interactions
between the heme model and the drug. The structure is consistent
with many known structure activity relationships for chloroquine:
either enantiomer alone is active,26 while changing the length or
bulk of the side chain reduces activity.27
10 G. T. Webster, D. McNaughton and B. R. Wood, J. Phys. Chem. B,
2009, 113, 6910–6916.
Recent years have seen leaps and bounds in the improvement
of our understanding of the quinoline family of anti-malarial
agents and their interactions with free heme.2a,3,21 With these
results it is clear that chloroquine may bind to heme in a manner
distinct from that of the quinoline alkoxides such as quinine or
quinidine which directly bind heme through the drug oxygen.21
Thus the different quinoline sub-classes may target heme
detoxification in different ways.2a This is excellent news, as it
exemplifies the fragility of the hemozoin formation pathway in
the parasite and its susceptibility to many kinds of interruption
and opens us to the possibilities of exploring the diverse
mechanisms of activities of each of these mini-classes of drugs
to branch out in the development of new antimalarials into a
much more diverse pool of compounds, taking advantage of
these different pathways.
11 C. Portela, C. M. M. Afonso, M. M. M. Pinto and M. Joao Ramos,
Bioorg. Med. Chem. Lett., 2004, 12, 3313–3321.
12 D. S. Bohle, L. Dodd Erin, J. Kosar Aaron, L. Sharma, W. Stephens
Peter, L. Suarez and D. Tazoo, Angew. Chem., Int. Ed., 2011, 50,
6151–6154.
13 D. S. Bohle and E. L. Dodd, Inorg. Chem., 2012, 51, 4411–4414.
14 R. D. Shannon, Acta Crystallogr., Sect. A: Cryst. Phys., Diffr., Theor.
Gen. Crystallogr., 1976, 32, 751–767.
15 R. B. Martin, Met. Ions Biol. Syst., 1988, 24, 1–57.
16 (a) E. Vo, H. C. Wang and J. P. Germanas, J. Am. Chem. Soc., 1997, 119,
1934–1940; (b) G. Kubal, A. B. Mason, S. U. Patel, P. J. Sadler and
R. C. Woodworth, Biochemistry, 1993, 32, 3387–3395; (c) W. R. Harris
and V. L. Pecoraro, Biochemistry, 1983, 22, 292–299; (d) M. Merkx and
B. A. Averill, Biochemistry, 1998, 37, 8490–8497; (e) E. A. Fadeev,
M. Luo and J. T. Groves, J. Am. Chem. Soc., 2004, 126, 12065–12075;
( f ) D. S. Bohle, E. L. Dodd, T. B. J. Pinter and M. J. Stillman, Inorg.
Chem., 2012, 51, 10747–10761.
17 S. Sasaki, T. Mizoguchi and H. Tamiaki, Bioorg. Med. Chem. Lett.,
2006, 16, 1168–1171.
18 D. S. Bohle, E. L. Dodd and P. W. Stephens, Chem. Biodiversity, 2012,
9, 1891–1902.
This work was supported by NSERC, CRC, and the FQRNT. We
thank Dr X. Ottenwaelder and M. S. Askari of Concordia Uni-
versity for the use of their diffractometer, and Dr D. Thompson
and B. Myron of Memorial University of Newfoundland for help
with obtaining fluorescence lifetime data.
19 N. Klonis, R. Dilanian, E. Hanssen, C. Darmanin, V. Streltsov,
S. Deed, H. Quiney and L. Tilley, Biochemistry, 2010, 49, 6804–6811.
20 A. Feis, M. P. Marzocchi, M. Paoli and G. Smulevich, Biochemistry,
1994, 33, 4577–4583.
21 K. A. de Villiers, H. M. Marques and T. J. Egan, J. Inorg. Biochem.,
2008, 102, 1660–1667.
22 (a) R. Buller, M. L. Peterson, O. Almarsson and L. Leiserowitz, Cryst.
Growth Des., 2002, 2, 553–562; (b) I. Solomonov, M. Osipova,
Y. Feldman, C. Baehtz, K. Kjaer, I. K. Robinson, G. T. Webster,
D. McNaughton, B. R. Wood, I. Weissbuch and L. Leiserowitz, J. Am.
Chem. Soc., 2007, 129, 2615–2627.
Notes and references
1 V. R. Solomon and H. Lee, Eur. J. Pharmacol., 2009, 625, 220–233.
2 (a) A. P. Gorka, A. de Dios and P. D. Roepe, J. Med. Chem., 2013, 56,
5231–5246; (b) P. E. Thompson and L. M. Werbel, Antimalarial Agents, 23 K. Y. Fong and D. W. Wright, Future Med. Chem., 2013, 5, 1437–1450.
¨
Academic Press, New York, 1972, p. 395; (c) G. R. Coatney, 24 (a) C. Litwinski, S. Tannert, A. Jesorka, M. Katterle and B. Roder,
W. C. Cooper, N. B. Eddy and J. Greenberg, Survey of Antimalarial
Agents, US Government Printing Office, Washington, 1952, p. 323.
3 T. J. Egan, Targets, 2003, 2, 115–124.
4 D. C. Warhurst, C. A. Homewood, W. Peters and V. C. Baggaley, Proc.
Helminthol. Soc. Wash., 1972, 39, 271–278.
Chem. Phys. Lett., 2006, 418, 355–358; (b) Y. Nakae, E.-i. Fukusaki,
S.-i. Kajiyama, A. Kobayashi, S. Nakajima and I. Sakata, J. Photochem.
Photobiol., A, 2005, 172, 55–61; (c) B. C. Robinson, I. M. Leitch,
S. Greene and S. Rychnovsky, WO Pat., 20022096366A2, To Miravant
Pharmaceuticals, Inc, USA, 2002.
5 A. C. Chou, R. Chevli and C. D. Fitch, Biochemistry, 1980, 400, 1543. 25 (a) N. Mataga, Y. Kaifu and M. Koizumi, Bull. Chem. Soc. Jpn., 1956,
6 (a) D. S. Bohle, B. J. Conklin, D. Cox, S. K. Madsen, S. Paulson,
P. W. Stephens and G. T. Yee, ACS Symp. Ser., 1994, 572, 497–515;
(b) A. F. G. Slater, W. J. Swiggard, B. R. Orton, W. D. Flitter,
29, 373–379; (b) P. J. Kovi, A. C. Capomacchia and S. G. Schulman,
Anal. Chem., 1972, 44, 1611–1615; (c) S. G. Schulman, R. M. Threatte,
A. C. Capomacchia and W. L. Paul, J. Pharm. Sci., 1974, 63, 876–880.
D. E. Goldberg, A. Cerami and G. B. Henderson, Proc. Natl. Acad. 26 P. Augustijns and N. Verbeke, Clin. Pharmacokinet., 1993, 24, 259.
Sci. U. S. A., 1991, 88, 325–329.
7 D. J. Sullivan, I. Y. Gluzman, D. G. Russell and D. E. Goldberg, Proc.
Natl. Acad. Sci. U. S. A., 1996, 93, 11865–11870.
27 R. T. Berliner and T. Butler, in Summary of data on the durgs tested in
man, ed. F. Y. Wiselogle and J. V. Edwards, U. Michigan, Ann Arbor,
1946, pp. 221–390.
13768 | Chem. Commun., 2014, 50, 13765--13768
This journal is ©The Royal Society of Chemistry 2014