4-N-, 4-S-, and 4-O-chloroquine analogues: Influence of side chain length and quinolyl nitrogen pKa on activity vs chloroquine resistant malaria

Article abstract of DOI:10.1021/jm701478a

Using predictions from heme-quinoline antimalarial complex structures, previous modifications of chloroquine (CQ), and hypotheses for chloroquine resistance (CQR), we synthesize and assay CQ analogues that test structure-function principles. We vary side chain length for both monoethyl and diethyl 4-N CQ derivatives. We alter the pK_a of the quinolyl N by introducing alkylthio or alkoxy substituents into the 4 position and vary side chain length for these analogues. We introduce an additional titratable amino group to the side chain of 4-O analogues with promising CQR strain selectivity and increase activity while retaining selectivity. We solve atomic resolution structures for complexes formed between representative 4-N, 4-S, and 4-O derivatives vs μ-oxo dimeric heme, measure binding constants for monomeric vs dimeric heme, and quantify hemozoin (Hz) formation inhibition in vitro. The data provide additional insight for the design of CQ analogues with improved activity vs CQR malaria.

Full text of DOI:10.1021/jm701478a

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
3466
J. Med. Chem. 2008, 51, 3466–3479
4-N-, 4-S-, and 4-O-Chloroquine Analogues: Influence of Side Chain Length and Quinolyl
Nitrogen pKa on Activity vs Chloroquine Resistant Malaria
Jayakumar K. Natarajan,†,‡ John N. Alumasa,† Kimberly Yearick,† Kekeli A. Ekoue-Kovi,† Leah B. Casabianca,†
Angel C. de Dios,†,§ Christian Wolf,†,§ and Paul D. Roepe*,†,‡,§
Department of Chemistry, Department of Biochemistry and Cellular & Molecular Biology, Center for Infectious Disease, Georgetown
UniVersity, 37th and O Streets, Washington, D.C. 20057
ReceiVed NoVember 27, 2007
Using predictions from heme-quinoline antimalarial complex structures, previous modifications of
chloroquine (CQ), and hypotheses for chloroquine resistance (CQR), we synthesize and assay CQ analogues
that test structure-function principles. We vary side chain length for both monoethyl and diethyl 4-N CQ
derivatives. We alter the pKa of the quinolyl N by introducing alkylthio or alkoxy substituents into the 4
position and vary side chain length for these analogues. We introduce an additional titratable amino group
to the side chain of 4-O analogues with promising CQR strain selectivity and increase activity while retaining
selectivity. We solve atomic resolution structures for complexes formed between representative 4-N, 4-S,
and 4-O derivatives vs µ-oxo dimeric heme, measure binding constants for monomeric vs dimeric heme,
and quantify hemozoin (Hz) formation inhibition in vitro. The data provide additional insight for the design
of CQ analogues with improved activity vs CQR malaria.
Introduction
Three of the most successful antimalarial drugs ever used
(quinine (QN),achloroquine (CQ), and mefloquine (MQ)) are
quinoline derivatives, and numerous additional lead compounds
with improved activity vs CQ resistant (CQR) malaria have been
discovered via a variety of recent synthetic modifications of
these structures.9–13 It is important to note that even in the face
of widespread resistance to CQ, similar compounds (MQ, QN,
amodiaquine (AQ), and isoquine (IQ)) remain active against
laboratory strains of CQR parasites. This could be in part due
to the fact that the principle target of CQ and related quinolines
is believed to be one or more forms of uncrystallized heme
(ferriprotoporphyrin IX, FPIX) found within the digestive
vacuole of the intraerythrocytic malarial parasite.14 FPIX is
released from proteolyzed red blood cell hemoglobin (Hb) within
the parasite digestive vacuole (DV) as the parasite very rapidly
grows within the human red blood cell. As such, the principle
target (heme made by the host) cannot be mutated or alterna-
tively expressed by the parasite in order to confer resistance.
The CQR mechanism is therefore unique, complex, and has
taken decades to appear on a large scale even in the face of
massive CQ use.15
Quinolines, as well as acridines, xanthones, and other classes
of antimalarial drugs are believed to exert some component of
their toxic action via preventing the crystallization of FPIX to
hemozoin (Hz).14,16,17 A detailed molecular mechanism for this
process remains elusive, but likely includes binding of the drugs
to either (or both) precrystalline soluble heme or (and) growing
faces of the crystal.16,18,19 In both cases, multiple modes of
interaction are possible. For example, precrystalline FPIX exists
in multiple chemical forms within the DV (various monomeric
and dimeric forms) that are in pH-dependent equilibrium at
currently unknown ratios. Yet, it is known that these multiple
forms are each capable of binding CQ and other quinoline-based
drugs. The relative affinities for FPIX monomer vs dimer are
not the same for various quinoline antimalarials, and these
Promising progress on live sporozoite-based vaccines
notwithstanding,1,2 new antimalarials active against drug resis-
tant malaria are urgently needed. Moreover, due to economic
issues, such compounds need to be relatively inexpensive to
produce and distribute. Ideally, the compounds should be stable
without refrigeration, act on multiple stages of the malarial
parasite, have prophylactic as well as therapeutic utility, and
be equally active vs multiple species of Plasmodia (e.g., P.
falciparum and P. ViVax).
Mono and combination therapies that include natural, syn-
thetic, and semisynthetic drugs based on the iron-activated 1,2,4-
trioxane artemisinin show significant promise in this regard.3–7
A spiroadamantane trioxolane, N-(2-amino-2-methylpropyl)-3-
adamantyl-1,2,4-trioxaspiro[4.5]decane-8-carboxamide, (com-
pound OZ277) optimized by Vennerstrom and colleagues3 is
now in advanced stages of development, and other promising
derivatives based on artemisinin4–7 are also emerging options
for improved drug therapy. However, compared to other
antimalarial pharmacophores, most of these syntheses are
relatively expensive and resistance to compounds exhibiting the
endoperoxide pharmacophore of artemisinin has already been
noted.8 Unfortunately, resistance to artemisinin derivatives will
likely continue to evolve and spread, so until effective vaccines
are developed and implemented, a long-term strategy to rapidly
provide new, inexpensive drugs and drug combinations active
against current and emerging drug resistant malaria is critical.
In this regard, simultaneous further development of multiple
successful pharmacophores provides the best strategy for
“staying ahead of the resistance curve” over the next decade or
more.
* To whom correspondence should be addressed. Phone: 202 687 7300.
Fax: 202 687 7186. E-mail: roepep@georgetown.edu.
†
a Abbreviations: CQ, chloroquine; CQS(R), chloroquine sensitive (re-
sistant); DV, digestive vacuole; Hz, hemozoin; MQ, mefloquine; QN,
quinine; t-BuOK, potassium tert-butoxide; VAR, vacuolar accumulation
ratio.
Department of Chemistry, Georgetown University.
Department of Biochemistry and Cellular & Molecular Biology,
‡
Georgetown University.
§
Center for Infectious Disease, Georgetown University.
10.1021/jm701478a CCC: $40.75
2008 American Chemical Society
Published on Web 05/31/2008

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
4-N-, 4-S-, and 4-O-Chloroquine Analogues
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 12 3467
affinities also differ strongly depending on pH and other
variables. The point being that even subtle modifications to CQ
can confer preference for one chemical form of FPIX vs another
(i.e., Fe[III] monomer vs µ-oxo dimer or vice versa), making
these compounds more or less active than CQ in the presence
of a given FPIX composition. Changes in FPIX composition
are a key prediction of the altered DV physiology known to
exist for CQR parasites.20,21
modifications of the CQ structure. These include that simulta-
neously fine-tuning both basicity of the quinolyl N and the length
of the CQ side chain may be important for optimizing
interactions with FPIX and that basicity of the tertiary aliphatic
N for CQ is important for accumulation within the parasite DV
but not for previously predicted ionic stabilization of CQ-FPIX
structures. Along with these principles, previous studies25,26 have
demonstrated that desethyl CQ has similar activity relative to
CQ for CQS strains but lower activity vs some CQR strains of
P. falciparum. In addition, shortening or lengthening the
aliphatic side chain of CQ has in general been shown to have
little effect on the activity vs CQS strains but to increase activity
vs CQR strains.9,11,26 However, these two modifications have
not previously been systematically varied in tandem, which
might result in additive or opposing effects. Below we report
data that test the above structure-function predictions for CQ
analogues.
Until very recently, design and synthesis of additional CQ
analogues guided by experimentally determined, atomic level
resolution drug-drug target structures has not been possible.
Many previous modifications have been addressed by combi-
natorial synthetic strategies, but how those modifications might
influence formation of a drug-target complex has been difficult
to rationalize or predict. Also, comparison between activity for
many of the CQ derivatives that have been made is difficult
because different strains of malaria have been used, and activity
3
has been assayed in multiple ways (e.g., via H hypoxanthine
Compounds 1-10. Compounds 1-10 were designed to
systematically explore the relationship between mono- vs diethyl
substitution at the terminal aliphatic N and the length of the
aliphatic side chain vs activity against CQS and CQR parasites
(Table 1). These compounds were prepared in two steps from
4,7-dichloroquinoline and a series of R,ω-diamines. The ami-
nation reaction proceeded at elevated temperatures with high
yields and the subsequent alkylation with ethyl bromide gave a
mixture of approximately 50% of the desired secondary and
tertiary amine, leaving about 50% of remaining starting materials
that were recovered in all cases (see Scheme 1). Aminoquino-
lines 1-10 were then analyzed for activity vs two CQS and
two CQR laboratory strains of P. falciparum using a new semi-
high-throughput SYBR Green I based assay. This assay was
developed independently in two laboratories,22,23 is easily
standardized, and was recently validated vs a large collection
of antimalarial compounds by the Walter Reed Army Institute.24
As described below and in the Discussion section, standardiza-
tion of the activity of candidate antimalarials against different
strains and species of Plasmodium is essential for future
progress, and the SYBR Green I assay offers one inexpensive
route that should be accessible to most laboratories engaged in
malaria research.
incorporation, lactate dehydrogenase activity, DNA staining,
etc.). These assays measure growth inhibition and/or death of
parasite cultures in different ways and thus do not report
identical IC50 data. Furthermore, most published modifications
to CQ have not included systematic, subtle alterations (e.g.,
iterative addition or substitution of single atoms) that can provide
clues to important structure-function principles.
In this study, we have analyzed a series of specific modifica-
tions to CQ that are predicted to be relevant for CQ activity vs
CQR parasites based on four considerations: (1) the geometry
of quinoline vs FPIX µ-oxo dimer structures we have recently
solved,16,18 (2) the existence of coordinate CQ-FPIX monomer
complexes under conditions that mimic those of the DV,19 (3)
differences in DV pH that have been deduced for CQS and CQR
parasites,20,21 (4) previous observations that the CQ side chain
length alters selectivity for CQR vs CQS parasites.9,11 We have
measured the activity of these rationally designed compounds
vs two CQ sensitive (CQS) and two CQ resistant (CQR) strains
using a standardized, easily validated, and inexpensive new
assay based on SYBR Green I intercalation that has recently
been adopted and validated by a number of laboratories.22–24
Takentogether,ourresultssuggestimportantnewstructure-function
principles for quinoline antimalarial drug design based on
chloroquine, including (1) replacement of the terminal tertiary
amino function by a secondary moiety reduces the potency vs
CQR strains, which suggests the side chain amino group is
recognized by the CQ resistance mechanism, (2) substitution
of S or O for N at position 4 significantly alters the quinolyl N
basicity and lowers the antimalarial potency while improving
the selectivity index (defined as the ratio of CQR strain IC50/
CQS strain IC50), (3) introduction of an additional basic amino
group to the side chain of 4-O CQ derivatives can improve the
potency while retaining an improved selectivity index, and (4)
surprisingly, no straightforward relationships between the ability
to bind FPIX µ-oxo dimer vs inhibition of Hz formation and
antimalarial potency exists for this series of CQ derivatives.
Aminoquinolines 3-5 are novel and have not previously been
analyzed vs malarial parasites, whereas 1, 2 and 6-10 have
been synthesized previously27–30 using similar but not identical
methods (Scheme 1) and tested vs less commonly used
laboratory strains of P. falciparum. Assessment of the activities
of all of these related CQ analogues has not previously been
standardized using the same strains, culture conditions, and
malarial growth inhibition assays. HB3 (CQS, Honduras) and
Dd2 (CQR, Indochina) are parents of a genetic cross that
produced a collection of progeny (GC03 [CQS] being one) for
which a very large amount of data has been collected regarding
thebiochemistryandgeneticsofchloroquinedrugresistance.15,31,32
Strain FCB (CQR, SE Asia) expresses similar CQR-causing
PfCRT mutations relative to Dd2,15 yet in most laboratories
shows 50-100% higher levels of CQR relative to Dd2. As such,
these strains are valuable reference points for future quinoline-
based antimalarial drug design guided by ongoing elucidation
of the CQR mechanism(s).
Results
Recently this consortium was the first to experimentally define
atomic level resolution structures for CQ, QN, quinidine (QD),
and AQ vs µ-oxo dimer FPIX,16,18 as well as the existence of
a coordinate CQ-monomeric FPIX complex under acidic
aqueous conditions.19 Other data suggest that DV pH may differ
for CQS vs CQR parasites.20 As previously suggested,16,18–20
these data led to several structure-function predictions for CQ
analogues that can now be systematically tested via strategic
We measured similar, but not identical, IC50 for 1, 2 vs CQS
(HB3, GC03) and CQR (Dd2, FCB) parasites, consistent with
earlier work that assayed CQS strains NF54 and Haiti 135 or
CQR strains K1 and Indochina I.26 Differences in precise IC50
are likely due to strain variation, our use of synchronized culture
vs asynchronous culture by others, the use of 3H hypoxanthine

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
3468 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 12
Natarajan et al.
Table 1. IC50 Values for 4-Amino-, 4-Alkoxy- and 4-Alkylthioquinoline Derivatives
experimental IC50 (µM)
previously reported IC50 (µM)
compound
HB3
Dd2
SIa
GCO3
FCB
SIa
NF5426
K126
SIa
Hait´ı 13529
Indochina I29
SIa
CQ
1
2
3
4
5
6
7
8
0.011
0.012
0.007
0.018
0.021
0.019
0.013
0.006
0.012
0.029
0.097
5.800
5.100
2.920
1.870
3.060
9.530
5.600
7.390
2.330
6.020
1.490
0.094
0.073
0.116
0.081
0.115
0.649
0.786
0.982
0.026
0.026
0.199
0.357
0.085
4.570
2.800
1.180
1.010
4.160
5.500
2.660
4.440
1.080
4.110
1.290
0.409
0.316
10.50
6.750
16.40
36.10
37.40
31.30
2.000
4.330
16.60
12.30
0.876
0.788
0.549
0.404
0.540
1.360
0.577
0.475
0.601
0.464
0.683
0.866
4.350
4.310
0.010
0.012
0.020
0.026
0.062
0.052
0.013
0.007
0.049
0.066
0.016
5.290
2.340
1.660
1.490
2.270
5.740
3.460
5.630
1.610
4.450
2.250
0.074
0.045
0.157
0.089
0.085
0.854
1.270
0.987
0.030
0.029
0.181
0.509
0.102
5.080
2.330
1.100
0.930
1.190
6.020
1.610
3.500
0.814
3.300
3.830
0.352
0.604
15.70
7.420
4.250
32.80
20.50
10.00
2.300
4.140
3.690
7.170
6.380
0.960
0.996
0.663
0.624
0.524
1.040
0.465
0.622
0.506
0.742
1.700
4.750
13.40
0.016
0.021
0.015
0.315
0.133
0.365
19.60
1.250
24.30
0.007
0.095
13.60
0.021
0.018
0.049
0.060
2.330
3.330
0.007
0.005
0.005
0.006
0.005
0.005
0.006
0.051
0.058
0.056
0.714
1.200
10.20
9.670
11.20
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
a Selectivity index; ratio of the IC50 for a given drug shown by a CQ-resistant strain vs IC50 for the companion CQ-sensitive strain. Column 4 shows SI
computed as Dd2 IC50/HB3 IC50, column 7 shows FCB IC50/GCO3 IC50, column 10 is K1/NF54, and column 13 is IndoI/Haiti 135.
Scheme 1. Synthesis of Chloroquine Derivatives 1-10
incorporation assays vs the present SYBR Green I approach,
or some combination. Regardless, we expanded this analysis
to include compounds bearing 4-6 methylene groups between
the two amino moieties (compounds 3-5) to explore the role
of deethylation (as occurs as a consequence of human metabo-
lism) vs side chain length could be analyzed in more depth
(compare the structure of compounds 1-5 vs 6-10).
that both longer and shorter side chain analogues are less well
recognized by the resistance mechanism (for example, drug
binding to the mutated PfCRT protein responsible for CQR33
could be weaker for long and short chain CQ analogues). If
this is indeed the case, then these data suggest that removal of
one alkyl group reverses this effect quite dramatically for longer
chain analogues. For example, the selectivity index (SI, defined
in Table 1 caption) obtained for 3-5 is 3 fold higher than for
CQ, while IC50 of 3-5 remains near that seen for CQ in CQS
strains. That is, the basic tertiary side chain amino group likely
contributes to recognition by the CQ resistance mechanism (see
Discussion).
Previously, Krogstad and colleagues9 as well as Ridley et
al.26 observed that the desethyl CQ derivatives 1 or 2 still exhibit
high IC50 vs CQR strains, whereas the diethyl analogues 6 or 7
show substantially lower IC50. One hypothesis for the trend in
the diethyl side chain series that has been offered previously is

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
4-N-, 4-S-, and 4-O-Chloroquine Analogues
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 12 3469
Scheme 2. Synthesis of CQ-Derived Ethers
Scheme 3. Synthesis of CQ-Derived Sulfides
Compounds 11-20. Compounds 11-15 were synthesized
from 4,7-dichloroquinoline and R,ω-alkanediols via consecutive
nucleophilic displacements (Scheme 2). Compound 16 was
synthesized in a single step from 4,7-dichloroquinoline and
2-diethylaminoethanethiol (Scheme 3), and compounds 17-20
were prepared from 7-chloroquinolyl-4-thiol and R,ω-dibro-
moalkanes via two consecutive SN2 displacements (Scheme 3).
These compounds were synthesized in order to inspect the
combined effects of heteroatom substitution at the 4 position
and the side chain modifications described above. This strategy
was pursued (in part) because recent solid state NMR studies
have shown that CQ may form a dative coordinate complex
with monomeric FPIX (via a heme Fe-quinoline N bond19)
under acidic conditions that mimic those of the parasite digestive
vacuole (DV). Thus, assuming other structural features remain
constant, altering the nucleophilicity of the quinolyl N (as
predicted for 4-O and 4-S substitutions) would influence
reactivity of the drug vs monomeric heme without necessarily
altering the structure required for noncovalent association to
µ-oxo dimer heme16 and hence preference for drug-monomer
vs drug-heme dimer complexes.16,18 Yet, none of these
compounds showed heightened activity relative to CQ and in
fact exhibited only modest antimalarial activity, with IC50 values
in the µM range (Table 1). However, we note that the selectivity
index (SI, cf. Table 1) is substantially improved for several of
these compounds. Thus, the 4-S and 4-O CQ analogue structures
are valid starting points for quinoline-based antimalarial drug
design wherein the goal is improved activity vs CQR strains
(e.g., lower SI, see results for compounds 21-23, below).
To further probe the molecular basis of these trends in relative
activity and selectivity index, we analyzed other features of
quinoline-based drugs that are believed to be critical with regard
to their antimalarial potency. We examined in detail a structur-
ally related set that best mimics the overall structure of CQ
(namely, the members of this set include those compounds that
contain side chains of similar length relative to CQ: 8, 13, 18).
The pKa of titratable N were calculated and measured (Table
2), binding constants for µ-oxo dimeric and monomeric heme
were measured in aqueous and 40% DMSO solutions, respec-
tively (Table 3), and the ability to inhibit Hz formation in vitro
at DV pH measured for CQS (5.6) and CQR (5.2) parasites are
tabulated (Table 4). In addition, we performed inversion-recovery
experiments at varied drug: dimer heme ratios and solved the
atomic level structures of complexes formed between these drugs
and µ-oxo dimer heme (Figure 1A,B).