Identification and SAR Evaluation of Hemozoin-Inhibiting Benzamides Active against Plasmodium falciparum

Article abstract of DOI:10.1021/acs.jmedchem.6b00719

Quinoline antimalarials target hemozoin formation causing a cytotoxic accumulation of ferriprotoporphyrin IX (Fe(III)PPIX). Well-developed SAR models exist for β-hematin inhibition, parasite activity, and cellular mechanisms for this compound class, but no comparably detailed investigations exist for other hemozoin inhibiting chemotypes. Here, benzamide analogues based on previous HTS hits have been purchased or synthesized. Only derivatives containing an electron deficient aromatic ring and capable of adopting flat conformations, optimal for π-π interactions with Fe(III)PPIX, inhibited β-hematin formation. The two most potent analogues showed nanomolar parasite activity, with little CQ cross-resistance, low cytotoxicity, and high in vitro microsomal stability. Selected analogues inhibited hemozoin formation in Plasmodium falciparum causing high levels of free heme. In contrast to quinolines, introduction of amine side chains did not lead to benzamide accumulation in the parasite. These data reveal complex relationships between heme binding, free heme levels, cellular accumulation, and in vitro activity of potential novel antimalarials.

Full text of DOI:10.1021/acs.jmedchem.6b00719

Subscriber access provided by UIC Library
Article
Identification and SAR evaluation of hemozoin-inhibiting
benzamides active against Plasmodium falciparum
Kathryn J. Wicht, Jill Michelle Combrinck, Peter J Smith, Roger Hunter, and Timothy J Egan
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b00719 • Publication Date (Web): 14 Jun 2016
Downloaded from http://pubs.acs.org on June 15, 2016
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
identification and SAR evaluation of hemozoinꢀinhibiting benzamides
active against Plasmodium falciparum
a
a,b
b
a
a,
Kathryn J. Wicht , Jill M. Combrinck , Peter J. Smith , Roger Hunter, Timothy J. Egan *
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
Department of Chemistry, University of Cape Town, Rondebosch 7701, South Africa.
b
Division of Pharmacology, Department of Medicine, Faculty of Health Sciences, University
of Cape Town, Observatory 7925, South Africa.
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ABSTRACT
Quinoline antimalarials target hemozoin formation causing a cytotoxic accumulation of
ferriprotoporphyrin IX (Fe(III)PPIX). Wellꢀdeveloped SAR models exist for βꢀhematin
inhibition, parasite activity and cellular mechanisms for this compound class, but no
comparably detailed investigations exist for other hemozoin inhibiting chemotypes. Here,
benzamide analogues based on previous HTS hits have been purchased or synthesized. Only
derivatives containing an electron deficient aromatic ring and capable of adopting flat
conformations, optimal for πꢀπ interactions with Fe(III)PPIX, inhibited βꢀhematin formation.
The two most potent analogues showed nanomolar parasite activity, with little CQ crossꢀ
resistance, low cytotoxicity and high in vitro microsomal stability. Selected analogues
inhibited hemozoin formation in Plasmodium falciparum causing high levels of free heme. In
contrast to quinolines, introduction of amine side chains did not lead to benzamide
accumulation in the parasite. These data reveal complex relationships between heme binding,
free heme levels, cellular accumulation and in vitro activity of potential novel antimalarials.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
ACS Paragon Plus Environment

Page 3 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
. INTRODUCTION
Despite encouraging recent developments, malaria remains a major global health
1
ꢀ3
burden. Crossꢀresistance has been reported among the still extensively used quinolines,
4
, 5
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
chloroquine (CQ), mefloquine (MQ), amodiaquine (AQ) and quinine (QN),
probably
because of their similar structures and modes of action. These drugs are believed to target the
hemozoin formation process in the digestive vacuole (DV) of the malaria parasite
6, 7
Plasmodium falciparum. Hemozoin (Hz) is an inert crystalline form of ferriprotoporphyrin
IX (Fe(III)PPIX), a cytotoxic byꢀproduct formed by autoxidation of heme released during
parasite digestion of host hemoglobin.
Though there is well established resistance to the quinoline antimalarials, the process of
Hz formation is unaltered within the parasite. Rather, mutations in digestive vacuole
membrane proteins, for example PfCRT or PfMDR1, are responsible for decreasing drug
5
, 8
accumulation at the site of action. Thus, the inhibition of Hz remains a viable drug target.
The quinolines are known to accumulate in P. falciparum, where CQ and AQ have been
9
directly shown to pH trap in the DV. Moreover, these weakly dibasic compounds
10
accumulate several thousandꢀfold so that they inhibit the target in the micromolar range,
despite their nanomolar activities against parasites. Quinolines have been thought to inhibit
the formation of Hz either via complexation with hematin or interaction with the surface of
11ꢀ13
the Hz crystal in the DV.
Synthetic Hz, known as βꢀhematin (βH), can be formed efficiently and reliably under
abiotic conditions which mimic the lipidꢀmediated process in the DV. Several assays to
identify βH inhibitors have been developed. Early examples did not include lipid mediators
1
4ꢀ17
and employed IR spectroscopy or radiolabeled heme for detection of inhibition.
More
18, 19
recently, neutral lipid blends (NLBs) have been used to mimic the natural process.
Huy et
20
al. described the use of a detergent, TWEEN 20, which is a cheap and convenient neutral
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
1
lipid mimetic, to mediate βH formation. Later Carter et al., demonstrated that the detergent
Nonidet Pꢀ40 (NPꢀ40) is the best available imitator of the NLB, resulting in the highest yield
of βH crystals and giving similar IC50 values for the inhibitors CQ and AQ. More importantly
this wellꢀestablished assay, which relies on UVꢀvis absorbance measurements for detection of
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
22
a bisꢀpyridylꢀheme complex, gave a tenꢀfold higher hit rate of parasiteꢀactive compounds
2
3ꢀ25
among the βH hits, than any previous methods used in screening efforts.
In addition, the
NP40 assay procedure lends itself well to HTS, as it incorporates nonꢀinterfering aqueous
acetone at pH 7.4 in the final plate development stage to handle poorly soluble compounds,
22
which had previously resulted in false negatives using the pyridine detection method. HTS
was conducted using this approach at Vanderbilt University with AQ and CQ as standards
(
βH IC50s of 21 and 53 ꢀM respectively) to discover novel βH inhibitors, active against the
25
D6 strain of P. falciparum. Of the 530 βH inhibiting hits, 73 showed parasite IC50s ≤ 5 ꢀM
and fourteen primary scaffolds were identified. Subsequent cell fractionation studies to
validate Hz as the target revealed doseꢀdependent increases in parasite free heme. In
particular, several compounds, representing two of the primary scaffolds, caused
extraordinarily high levels of free heme at 2.5 times the relevant IC50 (30ꢀ50%) compared to
7
, 26
the levels seen with CQ and other quinolineꢀcontaining scaffolds (8ꢀ14%).
Two of these
compounds contained either a pyridylbenzamide or phenylbenzamide moiety. The benzamide
chemotype was found to occur 33 times among the βH hits, with 13 compounds displaying
parasite growth inhibition. Of the fourteen scaffolds, this subset represented the second
highest hit rate (45%) for parasite activity of the βH hits after the quinolines.
Phenylbenzamides have been reported to show antimalarial activity against the CQꢀsensitive
2
7
3
D7 strain of P. falciparum in a screen published by GSK in 2012. One potent benzamide
analogue, a kinase inhibitor, was later optimized for solubility and parasite selectivity by
4
ACS Paragon Plus Environment

Page 5 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
8
Novartis. Moreover, two phenylbenzamides in the MMV Malaria Box have been reported
29
to be βH inhibitors in the NPꢀ40 assay.
While a number of antimalarial chemotypes have previously been reported to inhibit βH
2
5, 29ꢀ39
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
formation,
detailed systematic investigation of the relationships between structure and
βH inhibition, between βH inhibition and in vitro antimalarial activity and of cellular
mechanisms have so far been confined to quinolines. Previous studies on quinolines suggest
that suitable compounds containing electron deficient aromatic ring systems can interact
35, 40
strongly with Fe(III)PPIX resulting in inhibition of βH formation.
However, recent
evidence shows that βH inhibition likely occurs via binding of the inhibitor to the crystal
13
surface, rather than free Fe(III)PPIX in solution. This probably explains why no convincing
direct correlation has been observed between Fe(III)PPIX binding in solution and either βH
41
inhibition or parasite activity for the quinolines. Where Hzꢀinhibiting compounds can
access and accumulate in the parasite DV a buildꢀup of cytotoxic free Fe(III)PPIX occurs,
7, 41
resulting in parasite death.
Extension of such studies to nonꢀquinolines is a prerequisite for
a deeper and more general understanding of this pathway and its potential as a target for drug
discovery and development.
In this study, benzamide derivatives, starting from the scaffold core, Nꢀphenylbenzamide,
have been purchased or synthesized in order to explore structure activity relationships
(SARs). The study has been designed to probe the effect of electron deficient versus electron
rich aromatic rings on βH inhibition, the relationship between βH inhibition and in vitro
antimalarial activity and the effects of inhibitors on cellular Fe(III)PPIX levels in the parasite.
In addition, the hypothesis that introducing basic side chains should drastically improve
accumulation and hence increase potency in the parasite has been tested and the possible role
of Fe(III)PPIX association in accumulation has also been explored. It must be emphasized
that the primary aim of this study was to gain an understanding of the mechanistic aspects of
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
βH inhibitors in a nonꢀquinoline system, rather than to optimize the activity and medicinal
properties of the series. Nevertheless, preliminary in silico and in vitro pharmacological
studies have also been carried out on the most potent analogues to investigate the potential of
this series for drug optimization.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
. RESULTS AND DISCUSSION
2.1
Scaffold selection
One of the major challenges of drug development is selecting appropriate research
routes after HTS, particularly when it comes to identifying potential leads. In our case, HTS
revealed a monobenzamide (Figure 1a) and dibenzamide, 5ꢀnitroꢀN,N'ꢀdi(pyridinꢀ4ꢀ
yl)benzeneꢀ1,3ꢀdicarboxamide (Figure 1b), which inhibited
times the potency of CQ. The dibenzamide further showed a promising parasite activity of
.19 µM against the D6 strain, the third most potent compound of all 171 actives in the
βH formation with over three
0
screening effort. This compound was shown to be nonꢀcytotoxic against the Chinese hamster
ovarian (CHO) cell line with an IC50 >275 ꢁM. Besides potent activity and lack of
cytotoxicity, the benzamides were attractive from a synthetic point of view owing to the ease
with which derivatives could be prepared, as well as their structural difference from many of
the Hz inhibiting antimalarials that contain fused ring systems such as the quinolines and
30, 31, 42, 43
neocryptolepines.
Structural alerts for potential metabolic stability were confirmed
4
4
in silico using MetaSite, which identified the reduction of the nitro group, Nꢀoxidation of
the pyridyl moiety and amide hydroxylation as potential sites of metabolism in the liver.
45
However, VolSurf
predicted the benzamide derivatives from the HTS to have suitable
metabolic stability (within the 95% confidence boundary) when mapped relative to the
chemical space of known drugs. Furthermore, very successful drugs such as Lipitor,
6
ACS Paragon Plus Environment

Page 7 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Lidoderm and paracetamol contain benzamide moieties and we hypothesized that replacing
the nitro and pyridyl moieties with other functionalities could overcome potential metabolic
weaknesses. Other models within VolSurf predicted acceptable CACO2 permeation, protein
binding, solubility and volume of distribution within the 95% confidence boundary of known
drug space.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a)
b)
Figure 1. Examples of parasiteꢀactive hits found in the HTS. a) The monoꢀpyridylbenzamide,
3,5ꢀdinitroꢀNꢀ(pyridinꢀ4ꢀyl)benzamide (VU0358764) and b) the diꢀpyridylbenzamide, 5ꢀnitroꢀ
25
N,N'ꢀdi(pyridinꢀ4ꢀyl)benzeneꢀ1,3ꢀdicarboxamide (VU0001281).
2.2 SAR analysis for βH inhibition of the monobenzamide derivatives
The scaffold was deconstructed down to the core, Nꢀphenylbenzamide and this as well
as simple derivatives were purchased in order to identify the smallest substructure for which
βH activity is observed (IC50 cutoff < 200 µM for activity). The purchased compounds 1 – 6
did not display activity against βH formation (IC50 > 5500 µM). These six compounds
contained either an unsubstituted ring A (1) or a substituent in the para position of ring A (2ꢀ
6
) as shown in Table 1. Purchased compounds 7 – 10, all possessing a halogen substituent in
meta and para positions demonstrated an interesting phenomenon at very high (mM)
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
concentrations where upon addition of Fe(III)PPIX, a pink precipitate rapidly formed,
indicative of low spin Fe(III)PPIX complex formation in a similar manner to that of pyridineꢀ
Fe(III)PPIX coordination in the βH assay. Coordination of these compounds to the central Fe
atom of Fe(III)PPIX was believed to be via the N of their pyridyl substituent which possesses
added nucleophilicity over that of pyridine owing to some amide resonance to the para
pyridyl N. The precipitation of this complex was probably a result of the lipophilic influence
of the halogen(s) with 9, containing two chloro substituents, showing the effect at the lowest
compound concentration and 7, with only one p-chloro, showing the effect at five times
higher concentration.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 1. Purchased benzamides for core structure analysis. All the compounds were inactive
and pyridylbenzamides 7ꢀ10 formed a pink precipitate upon monodentate coordination to the
Fe center at the indicated concentration range when added to a solution of 100 ꢀM
Fe(III)PPIX and 30.5 ꢀM NP40 at pH 4.8.
Concentration range at which
Code
precipitate forms (µM)
1 – 6
N/A (Inactive ≤ 5500 µM)
4500ꢀ5500
7
8
9
1300ꢀ2300
700ꢀ1300
1
0
1300ꢀ2300
8
ACS Paragon Plus Environment

Page 9 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The strategy for the synthesis of monobenzamides was aimed to determine the effect of an
ortho substituent on ring A and whether or not a phenyl ring was tolerated as ring B, leading
to generation of compounds 12aꢀe (Figure 2).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2: Synthesis of monoꢀbenzamide derivatives. *Triethylamine (Et
in the synthesis of the pyridylbenzamides (12cꢀe).
3
N) was only added
The SAR analysis revealed that, for the monoꢀbenzamides two nitro groups in the
meta positions, as opposed to ortho or para, were favorable for inhibition of βH formation
2
(Table 2) and that both phenyl or pyridyl for R were tolerated. The 4ꢀpyridyl analogue (12c)
was most active while the 2ꢀpyridyl (12e) was the least active.
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 2. βH inhibition activity for the synthesized monobenzamides
Compound
R1
R2
βH inhibition IC (µM)
50
1
2a
2b
2ꢀnitro
phenyl
phenyl
4ꢀpyridyl
3ꢀpyridyl
2ꢀpyridyl
ꢀ
>1000
22 ± 1
1
3,5ꢀdinitro
3,5ꢀdinitro
3,5ꢀdinitro
3,5ꢀdinitro
ꢀ
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
12c
2d
2e
CQ
10.9 ± 0.3
13 ± 1
1
1
134 ± 8
31.5 ± 0.5
The precise mechanism by which βH inhibitors inhibit crystal growth is still a matter
of debate. Over the past two decades it has been proposed that inhibition by quinoline
14, 16, 46
antimalarials is via direct drug complexation to hematin in solution,
by binding to the
11, 47, 48
fastest growing faces of the βH crystal,
or via a drugꢀheme complex capping the Hz
4
9
crystal to block sites of further crystal growth. Nevertheless, all the proposed mechanisms
involve an interaction between the drug compound and Fe(III)PPIX. Recently, de Villiers et
12
al. have reported crystal structures of quinidineꢀheme (QDꢀFe(III)PPIX) and QNꢀ
Fe(III)PPIX complexes, which demonstrate that three key interactions are involved in heme
binding by these particular drugs, namely coordination, hydrogen bonding and πꢀπ stacking.
For QD and QN, hydrogen bonding occurs between the protonated quinuclidine nitrogen and
the propionate group of Fe(III)PPIX. Another study showed that the interaction between
Fe(III)PPIX and dihydroxyxanthones features hydrogen bonding via their hydroxyl groups,
5
0
as well as πꢀπ stacking and carbonylꢀFe coordination.
Most of the benzamides possess
features which are capable of binding to Fe(III)PPIX via all three of these interactions,
specifically, a pyridyl N for coordination, an amide for hydrogen bonding and aromatic rings
for πꢀπ stacking. However, 12b does not possess a pyridyl moiety for coordination to the
10
ACS Paragon Plus Environment

Page 11 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Fe(III)PPIX center and yet retains a low IC50, ruling out coordination as a key interaction for
βH activity. Furthermore, the lack of hydrogen bonding substituents in 12bꢀe suggests that
this interaction can only occur via the amide. Since many of the compounds with amides,
including all the purchased benzamides, were inactive, an amide hydrogen bonding
interaction was considered very unlikely to be the crucial determinant of βH inhibition
activity. By a logical process of elimination, πꢀπ stacking was left as the major interaction
leading to inhibitory activity. Although the exact requirements for πꢀπ stacking are still not
clear, it was hypothesized that having larger planar aromatic molecular surfaces would
increase the strength of the interaction with either hematin or Hz. The ability of the molecule
to lie flat would therefore be expected to influence the βH inhibition activity. This hypothesis
was considered to be the only feasible explanation for the observed SARs.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The πꢀπ stacking mechanism was consistent with the structureꢀactivity data for
pyridyl derivatives 4 with 2,4ꢀdinitro groups (inactive) vs 12c with 3,5ꢀdinitro substituents
(active) and phenyl derivatives, 12a with 2ꢀnitro groups (inactive) vs 12b with 3,5ꢀdinitro
substituents (active). In silico internal energy calculations carried out using Gaussian
51
software demonstrated that the barrier to rotation around the bond linking the carbonyl
group to the phenyl ring for 4 possesses minimum energy conformers with torsion angles
about the C=CꢀC=O bond of 70° and ꢀ110°. For the molecule to be planar, the torsional angle
would be 0° or ±180°. The steric strain energies of these flat conformers are far greater than
the thermally accessible energy 3kT indicated in Figure 3a. If the molecule is only able to
bind to Fe(III)PPIX when in the planar conformation, this hypothesis would explain the lack
of βH inhibition activity for 4. This hypothesis is also consistent with the positive activity
data for 12c which is able exist in a planar conformation (±180°) below 3kT (Figure 3b)
owing to the low steric constraint caused by meta as opposed to orthoꢀnitro substituents. It
may also explain why 2e with a 2ꢀpyridyl group had lower activity than the 3ꢀpyridyl and 4ꢀ
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
pyridyl analogues due to the close proximity of the negative ortho nitrogen lone pair to the
electron cloud on the carbonyl oxygen, which is likely to affect the conformational preference
of 12e.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. Free energy diagrams for rotation around the indicated bond where 0° or ±180°
represents the flat conformation for (a) purchased compound 4 (βH inactive – shown in
green) compared to synthesized compound 12c (βH active – shown in blue) and (b)
compound 12c (βH active). The value 3kT represents the thermally accessible energy value at
37 °C. Note the difference in scale of the yꢀaxes (a) and (b).
A summary of the SARs for the monobenzamides is given in Figure 4. These results
demonstrate the minimum requirements for βH activity of the benzamides, which were then
applied in the synthesis of diꢀ and triꢀbenzamide derivatives. Most importantly, only
compounds with substituents in the meta position on ring A were chosen for further
investigation, since substituents in the ortho and para positions were not tolerated for βH
2
inhibitory activity. Furthermore, 2ꢀpyridyl was excluded for the R substituent since the
results for the monobenzamides suggested that the βH inhibition activity was substantially
12
ACS Paragon Plus Environment

Page 13 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
lower in this analogue. The effect of different R
investigated for the dibenzamides.
2
substituents on the SARs was further
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. βH inhibitory SARs for the monobenzamides.
2.3 SAR analysis for βH inhibition of the dibenzamide derivatives
Although two of the monobenzamides were able to inhibit βH inhibition with
comparable activity to AQ, the parasite data from the HTS indicated that dibenzamides were
more effective at preventing parasite growth than the monobenzamides. For this reason, a
focused library of dibenzamides was synthesized. In addition, the 4ꢀpyridyl parent compound
was resynthesized (15a) along with 3ꢀpyridyl (15b) and phenyl (15c) derivatives (Table 3). In
2 1
addition to varying the type of ring B (R substituent), the substituent on ring A (R ) was also
varied from the parent nitro to cyano, methyl, tꢀbutyl, hydrogen or methoxy, resulting in
derivatives 15b ꢀ 15n. These substituents were chosen for their range of electronꢀwithdrawing
vs releasing as well as hydrophobic vs hydrophilic properties, which excluded the halogens,
since the chloro derivative was shown to be inactive in the purchased monoꢀbenzamides. In
addition, replacing X = CH with X = N resulted in derivatives with a pyridyl central ring as
opposed to a phenyl (15o and 15p) all of which are defined in Figure 5, along with the
tribenzamide 20.
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
R1
R1
R1
SOCl2
reflux, 4 hrs Cl
OH
2 eq. R2-NH2, 2 eq. Et3N*
pyridine, RT, 2-8 hrs
HN
NH
Cl
HO
X
R2
X
R2
X
O
O
O
O
O
O
1
3a-n
14a-p
15
*
Only for pyridyl derivatives
1
1
1
1
1
1
1
5a: X = CH; R = NO ; R = 4-pyridyl 15h: X = CH; R = t-butyl; R = 4-pyridyl
5b: X = CH; R = NO ; R = 3-pyridyl 15i: X = CH; R = t-butyl; R = 3-pyridyl
5c: X = CH; R = NO ; R = phenyl
5d: X = CH; R = CN; R = 4-pyridyl 15k: X = CH; R = H; R = 4-pyridyl
5e: X = CH; R = CN; R = phenyl
5f: X = CH; R = CH ; R = 4-pyridyl 15m: X = CH; R = OMe; R = 4-pyridyl
5g: X = CH; R = CH ; R = phenyl
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1 2 2
1 2
1
2
2
1
2
15j: X = CH; R = t-butyl; R = phenyl
1 2
1
2
2
1
2
1
2
15l: X = CH; R = H; R = phenyl
1 2
1
2
1
3
2
1
2
15n: X = CH; R = OMe; R = phenyl
1 2
5o: X = N; R = H; R = 4-pyridyl
15p: X = N; R = H; R = phenyl
1
3
2
1
1 2
1
2
Figure 5. The synthesized dibenzamides, analogues of the hit compound 15a, and the
3
tribenzamide 20. *Triethylamine (Et N) was only added in the synthesis of the
pyridylbenzamides.
Dibenzamides 15a, 15b and 15c displayed similarly potent βH IC50 values (2.9ꢀ7.0
µM), four to ten times more active than CQ and at least three times the activity of the
monobenzamide analogues (Table 3). This improved activity is in agreement with the
hypothesis that larger planar surfaces result in improved βH inhibition due to increased πꢀπ
stacking ability. In addition, the IC50 reduction for the dibenzamides parallels the 2ꢀfold
improvement in βH inhibition activity seen for the diꢀquinolines, piperaquine and
5
2
hydroxypiperaquine compared to monoꢀquinoline CQ. The cyano derivatives (15d and 15e)
maintained excellent βH activity about six times that of CQ for both the pyridyl and phenyl
rings. However, for the methyl (15f and 15g), tꢀbutyl (15hꢀj), hydrogen (15k and 15l) and
2
methoxy derivatives (15m and 15n), only those with pyridyl rings at R showed activity,
while those with a phenyl ring displayed no activity below 1000 µM. Compounds 15o and
14
ACS Paragon Plus Environment

Page 15 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
15p with a central unsubstituted pyridyl ring and R
1
= H showed weak activity for the pyridyl
derivative and no activity below 1000 µM for the phenyl derivative. The tribenzamide 20
showed good activity, slighter better than that of CQ.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 3. βH inhibition activities for the diꢀbenzamides 15aꢀ15p and triꢀbenzamide 20.
Compound
5a
15b
5c
5d
5e
15f
5g
5h
R
1
substituent
βH inhibition IC50 (µM)
4.3 ± 0.1
4.7 ± 0.7
7.0 ± 0.2
6.34 ± 0.09
4.4 ± 0.2
8 ± 1
1
Nitro
1
1
Cyano
1
Methyl
1
>1000
1
13.3 ± 0.7
6.8 ± 0.1
>1000
1
5i
tertꢀButyl
1
5j
15k
5l
5m
5n
15o
5p
507 ± 33
>1000
H
1
1
22 ± 1
Methoxy
H (X = N)
1
> 1000
260 ± 7
1
>1000
2
0
N/A
N/A
21 ± 1
CQ
31.5 ± 0.5
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
In general, a compound with a flat, hydrophilic, electron withdrawing group such as
NO
2
(15aꢀc), CN (15d, e) or benzamide (20)) at R
1
was able to maintain activity with either a
pyridyl or phenyl ring at R
2
. Conversely, when R
1
was H or a more electron releasing group,
only the pyridyl derivative(s) were active and those with phenyl at R showed no ability to
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
inhibit βH formation even at a relatively high concentration. This may indicate that the
molecules have two possible orientations of interaction with Fe(III)PPIX giving rise to βH
inhibition. The first involves πꢀπ stacking with the central ring, which can only take place
with the flat, electron withdrawing nitro and cyano groups. The second involves an
interaction with the aromatic substituent R , which appears not to occur in the case where R
2
2
is phenyl as opposed to pyridyl, similar to that for naphthalene analogues of the 4ꢀ
5
3
aminoquinolines, which also lack βH inhibition activity. This may be as a result of the
relative electron richness of the phenyl as opposed to the pyridyl ring, with the latter being
electron deficient owing to the electronegative nitrogen atom. Interestingly, the high IC50 of
15o, the X = N derivative with R = pyridyl, can be explained in a similar manner to that of
2
1
2e. Since both these compounds have an ortho-pyridyl ring B, the influence of the ortho
nitrogen in close proximity to the amide results in a decreased probability of πꢀπ interaction,
either by causing the molecule to be less planar, or by adding too much electron density at
that particular site of the molecule. The SARs for the dibenzamides are shown in Figure 6a.
Note that the planar and electron withdrawing nitro and cyano substituents at R follow a
1
different SAR scheme to that of the methyl, tꢀbutyl, H and methoxy substituents. Although
the exact role of the R
1
substituent is not clear, it is possible that this substituent forces or
H were it binds,
directs the benzamide rings into the groove on the crystal face of
β
accounting for the strong inhibitory activity (Figure 6b). Previous studies on xanthones and
porphyrins as βH inhibitors also reported significant interactions between the planar aromatic
39, 54
rings of Fe(III)PPIX and the inhibitor.
However, additional interactions not present in the
16
ACS Paragon Plus Environment

Page 17 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
benzamides, involving carbonyl or hydroxyl groups to bind the iron of heme for enhanced
activity, were also described.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(
a)
b)
Figure 6. (a) βH inhibitory SARs for the dibenzamides and (b) a possible mode of interaction
to explain the observed βH inhibition activities. In this example, the dibenzamide, 15h
7
interacts with the fastest growing Hz crystal face with the R tꢀbutyl moiety directed
1
55
outwards. Graphics were supplied by R. Muller and generated using Materials Studio.
2.4 Activity of benzamides against P. falciparum in culture.
The synthesized benzamide derivatives were tested for inhibitory activity against the
CQꢀsensitive NF54 strain of P. falciparum. The monobenzamides 12bꢀe displayed a range of
IC50 values from 2.3 to 22 µM and 12a, which was not a βH inhibitor, showed no activity
below 41.3 µM, the maximum concentration tested (Table 4). A good correlation was
identified between the inverse of the βH inhibition IC50 and inverse parasite growth inhibition
2
IC50, with an r of 0.88 (Figure 7a).
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The original hit compound, 15a (Figure 1b), was resynthesized and tested in the
available CQꢀsensitive NF54 strain. The activity was found to be three times weaker than in
the D6 strain used in the HTS, with an IC50 of 0.7 µM. The 3ꢀpyridyl and phenyl analogues,
1
5b and 15c, exhibited a further 2ꢀfold lower activity. The other derivatives were also tested
in the NF54 and/or D10 strains of P. falciparum. Where comparisons were made, both strains
gave similar IC50 values for tested compounds (Table 4). Replacing the nitro substituent (R
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
)
of 15a with cyano caused a 3ꢀfold increase in IC50 to 1.9 µM for 15d, the pyridyl derivative.
A further IC50 increase to 5.2 µM was observed for the phenyl derivative 15e. A similar
reduction in activity was noticed when replacing the R pyridyl ring with a phenyl group for
2
the derivatives where R is methyl (15f and 15g), tꢀbutyl (15i and 5j), hydrogen (15k and 15l)
1
or methoxy (15m and 15n). Both compounds with X = N (15o and 15p) were inactive up to
315 µM in the parasite, most likely because they did not possess sufficient βH inhibition
activity. Only compound 15i, the tꢀbutyl derivative, showed a slight improvement in parasite
activity relative to the parent compound, with IC of 0.6 µM. The trend in activities for the
50
4ꢀpyridyl vs 3ꢀpyridyl vs phenyl analogues of the tꢀbutyl derivatives (15hꢀj) was different
from that of the nitro derivatives (15aꢀc), in that the order of potency between the 3ꢀpyridyl
vs 4ꢀpyridyl was reversed. In all cases, however, the phenyl derivatives were less active
against P. falciparum. The tribenzamide, 20, showed very poor parasite activity despite being
a good βH inhibitor. This may have been as a result of its excessive hydrophobicity
(predicted logP = 5.21).
18
ACS Paragon Plus Environment

Page 19 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 4. CQꢀsensitive parasite activities for the synthesized monoꢀ, diꢀ and triꢀbenzamides.
NF54
D10 parasite
Compound Derivative info
parasite IC50 (µM) IC50 (µM)
1
2a
2b
>41.3 ± ND
5 ± 1
ND
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
ND
12c
Monobenzamides
2.3 ± 0.3
5 ± 1
2.4 ± ND
2.3 ±ND
24 ± 5
0.9 ± 0.1
ND
1
2d
2e
1
22 ± 4
15a
0.7 ± 0.2
1.32 ± 0.04
1.4 ± 0.2
1.9 ± 0.2
ND
1
5b
5c
R = Nitro
1
1
ND
15d
2.0 ± 0.2
5.2 ± ND
ND
R
R
1
1
= Cyano
= Methyl
1
1
1
1
1
1
5e
5f
5g
5h
5i
2.5 ± 0.6
146 ± 1
1.6 ± 0.2
0.6 ± 0.1
8.1 ± 0.5
9 ± 4
ND
ND
R
1
= Tertꢀbutyl
0.7 ± 0.1
ND
5j
15k
ND
R
R
R
1
1
1
= H
15l
>314
ND
15m
3.0 ± 0.5
37 ± 13
>315
ND
= Methoxy
= H (X = N)
15n
ND
1
5o
5p
ND
1
>315
ND
Triꢀ
20
81 ± 12
>42
phenylbenzamide
Standard
CQ
0.008 ± 0.002
0.021 ± 0.002
As with the monobenzamides alone, inverse activity of the βH and NF54 parasite
IC50s were correlated when all the benzamides were taken into account, in contrast to the 4ꢀ
3
3, 35
aminoquinolines which do not show convincing correlations between these activities.
The relationship was statistically significant with P < 0.0001. Furthermore, a moderately
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
2
good r value of 0.68 was observed for the correlation, which increased to r = 0.84 upon
removal of the point corresponding to compound 15i, which was an apparent outlier (Figure
7
b). This compound showed a better parasite activity than expected based on its βH inhibition
IC50. Possible reasons for this could be that its properties differ markedly from the other
compounds, that there is favorable uptake in the parasite, or that the βH IC50 measured for
this compound in the NPꢀ40 assay does not accurately represent its Hz inhibitory activity in
the parasite. Nevertheless, the correlations in Figure 7 suggest that the benzamides do indeed
exert their antiparasitic action via inhibition of intracellular Hz formation.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a)
b)
0.6
0
.4
.2
0
0.0
.000
0
0.025
0.050
0.075
0.100
1
/βH IC50 (ꢀM)
Figure 7. A linear correlation between the inverse of the βH and parasite growth IC50 values
2
for the CQꢀsensitive NF54 strain for (a) the monobenzamides (r = 0.875; P = 0.0195) and (b)
2
the monoꢀ, diꢀ and triꢀbenzamides combined (r = 0.68, P < 0.0001). The statistics improve
2
when derivative 15i (red circle) is excluded (r = 0.84, P < 0.0001). Measurements of βH
inhibition were in triplicate and parasite growth inhibition in at least duplicate.
This mechanistic hypothesis was validated using a cellular fractionation assay to
convincingly demonstrate that the benzamides cause concentration dependent increases in
free heme and corresponding decreases in Hz within whole cell parasite cultures. In this
2
6
assay, for which full validation and details have been reported previously, parasitized red
20
ACS Paragon Plus Environment

Page 21 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
cells are incubated with the test compound for 32 h, the trophozoites are isolated, lysed and
spun down to separate the Hbꢀcontaining supernatant. The resulting precipitate consists of
two fractions; firstly, the nonꢀHz heme, consisting of free heme or heme associated with
drug, lipids or proteins, which is readily solubilized in SDS buffer and pyridine for detection;
and secondly, the Hz fraction which remains completely insoluble until dissolved in NaOH
and SDS, and which is then neutralized before being independently measured via the pyridine
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
22
detection procedure. Three benzamides were analyzed using this cell fractionation assay.
Two of the selected compounds (12d and 15i) showed excellent βH inhibition activity in the
NPꢀ40 assay, while one compound was not a βH inhibitor (15j). The latter served as a
negative control. The monobenzamide 12d and the dibenzamide 15i were chosen because
they have relatively weak and strong parasite growth inhibition activity respectively. This
was desirable so as to observe whether free heme levels had an influence on this activity. No
significant change in the free heme or Hz levels with increasing concentration was observed
for 15j (Figure 8a). This was in agreement with the NPꢀ40 assay for βH inhibition and
provided strong support for the technique for finding Hz inhibiting compounds used in this
study. It would appear that this compound acts weakly on an alternate parasite target. By
contrast, 12d and 15i showed a large increase in the amount of nonꢀHz heme (ꢁheme ∼40
fg/cell at 2.5 times the IC50) with a corresponding dramatic decrease in Hz (ꢁHz ∼40 fg/cell
at 2.5 times the IC ) within the parasite relative to the control (Figure 8b, c). The significant
50
changes in free heme and Hz levels for these benzamides suggested that the precise manner
by which they cause hemeꢀrelated toxicity differs from that of CQ, since much smaller
changes in free heme and Hz levels by ~5 fg/cell and ~15 fg/cell respectively were observed
2
6
at 2.5 times the CQ IC . Other quinoline antimalarials, such as AQ, QN and mefloquine
50
7
showed even smaller changes than CQ. This finding demonstrates that higher levels of free
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
heme are required for disrupting parasite growth in the case of the benzamides compared to
CQ and other quinoline antimalarials.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a)
b)
22
ACS Paragon Plus Environment

Page 23 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
c)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 8. Hz and free heme levels from cultured P. falciparum treated with varying
concentrations of benzamides a) 15j, a negative control showing no statistically significant
difference in Hz or free heme as a function of dose, b) 12d, a monobenzamide and c) 15i, a
dibenzamide exhibit strong dose dependent changes. Asterisks indicate statistical significance
relative to no drug (2ꢀtailed tꢀtest): *P < 0.05; **P < 0.01; ***P < 0.001, n = 3.
In the case of the βH inhibitors, the free heme curves and the parasite survival curves with
varying test compound concentrations crossed over close to the D10 IC50 of 12d (Figure 9a)
and 15i (Figure 9b) respectively, strongly supporting the hypothesis that they target the Hz
formation process. Direct evidence that these compounds disrupt the process of Hz
crystallization in the parasite is provided by the transmission electron microscopy (TEM)
images in Figure 10, demonstrating clear morphological changes in parasites treated with 15i
relative to the untreated cells. When dosed at 2.5 times the IC50 and incubated for 32 h,
intracellular Hz crystals appeared within swollen DVs as irregularly scattered and shapeless
clumps surrounded by a distinct electron dense penumbra, in contrast to the wellꢀformed,
tightlyꢀpacked and sharply delineated crystals observed in the control. Swelling of the DV
and a lack of crystal uniformity have previously also been observed in TEM images of CQ
7, 56, 57
treated cells.
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a)
b)
1
10
Non-Hz heme
100
Non-Hz heme
5
0
100
Parasite survival
Parasite survival
9
8
7
6
0
0
0
0
3
0
0
80
40
6
0
30
50
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
4
3
2
0
0
0
4
0
20
20
0
10
0
10
10
-10
0
-20
-
3
-2
-1
0
1
2
-3
-2
-1
0
1
2
log [12d(ꢀM)]
log [15i(ꢀM)]
Figure 9. Dose response curves for the nonꢀHz heme and the % parasite survival, which
cross near the parasite survival IC for compounds a) 12d and b) 15i. Scales adjusted so that
5
0
the IC for parasite survival coincides with 50% of the final value for increase in free heme
50
(
indicated by the dotted line).
Figure 10. TEM images of human erythrocytes infected with P. falciparum showing a) an
untreated control and b) two examples of cells treated with 15i. The thick white arrows show
the swollen DVs of treated cells. The brown arrows indicate the Hz crystal in the DV which
has been enlarged and adjusted (20% brightness, 20% contrast in each) for clarity to readily
discern the difference between crystal forms in untreated vs treated samples.
24
ACS Paragon Plus Environment

Page 25 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2.5 Attempts to improve accumulation through pH trapping
58
Fitch et al. showed that the accumulation of CQ is sensitive to the pH gradient
between the parasitized erythrocyte and the medium by varying the pH of the latter. Yayon et
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
5
9
al. reported high pHꢀdependent accumulation levels of CQ in the DV, investigated by flow
6
0
dialysis, in which diffusion across a membrane can be measured as well as the tracing of
1
4
61
CQ via fluorescence or a C labelled analogue. Millimolar concentrations of CQ were
predicted to exist in the DV at therapeutic doses, accounting for the high activity of the drug.
In the 1990s, more sophisticated methods allowed the proton gradients to be varied, by
decreasing the extravacuolar pH with the acidifying agent NH
4
Cl or via inhibitors of vacuolar
62,63
proton pumps.
Accumulation measurements under these conditions showed that lowering
the external pH results in less accumulation and hence these studies concluded that CQ
accumulates in the acidic DV as a result of its weak base properties. This pHꢀdependent
accumulation can be explained via the phenomenon of pH trapping, whereby a neutral
species passively diffuses into the acidic DV and is protonated to form a less lipophilic
(
lower logD) charged species, which is then unable to cross the membrane. Designing a
compound with greater basicity may be expected to enhance the pH trapping effect and result
in improved parasite activity. Indeed, upon addition of basic amino side chains to xanthones,
increased activities in the D6 parasite strain as well as heme affinities were previously
64
reported. This is also believed to be the case for the 4ꢀaminoquinolines AQ and CQ. It has
previously been reported that 4ꢀaminoꢀ7ꢀchloroꢀquinoline is able to inhibit βH formation and
41
shows a D10 parasite IC50 of 3.8 µM. Upon attachment of a variety of tertiary amino
containing side chains at the 4ꢀamino position, βH inhibition was maintained and parasite
activity was greatly enhanced by approximately 100ꢀfold. According to the study, this was
probably as a result of improved accumulation in the DV. In addition, it was proposed that
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
rational design could be employed to develop novel antimalarials from scaffolds that form πꢀ
41
π complexes with Fe(III)PPIX and inhibit βH formation.
A benzamide derivative where the R substituents were each N,Nꢀdimethylꢀ
2
phenylmethanamine (17) was expected to possess suitable properties for this study, namely, a
molecular weight below 500 g/mol and a positive clogD value of 1.12 at neutral pH (Figure
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
11). This compound was synthesized in order to probe the effect of greater basicity and pH
trapping for the benzamide series.
Figure 11. Preparation of 17 via reductive amination from 16.
Compound 17 was tested for both βH inhibition and parasite activity. In the NP40
assay, the compound maintained potent βH activity with an IC50 of 7.9 ± 0.2 µM, appearing
to be a stronger inhibitor than AQ or CQ. However, upon assessing the activity against P.
falciparum of this more basic derivative, which was expected to accumulate to a greater
extent in the parasite DV than 15a and hence possess enhanced activity, no improvement was
found. An NF54 parasite IC of 1.5 ± 0.2 µM was obtained. This activity was the same as
50
that of the phenyl derivative, 15c, within the error margin, suggesting that no significant
26
ACS Paragon Plus Environment

Page 27 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
accumulation had taken place. This lack of effect was especially striking when compared
with the predicted accumulation that would have been expected to take place due to pH
trapping assuming achievement of equilibrium. The predicted VAR due to pH trapping alone
was calculated using the Henderson–Hasselbach equation (Eq. 1) at an external pH of 7.4
(pH ) and intravacuolar pH of 4.8 (pH ) for a compound (B) with two protonatable sites.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
e
v
(ꢊꢋꢊꢌꢍꢎꢏꢐꢑꢍꢒꢈꢋꢊꢌꢍꢎꢏꢓꢔꢍꢎꢏꢐꢑꢐꢍꢒꢈꢃ)
(ꢊꢋꢊꢌꢍꢎꢏꢐꢑꢍꢒꢉꢋꢊꢌꢍꢎꢏꢓꢔꢍꢎꢏꢐꢑꢐꢍꢒꢉꢃ)
ꢄꢅꢆꢇ,ꢈ
ꢀ
ꢁꢂꢃ =ꢃꢄꢅꢆꢇ,ꢉ =ꢃ
Eq. 1
a
Using the predicted pK values of 9.12 and 8.52 for compound 17, the expected VAR
was estimated at 147,000. This indicated that the derivative should have been about 147,000
~
times more concentrated in the DV compared with 15c, which is not able to pH trap. Since
the compounds possessed the same activity, two hypotheses were suggested; either
compound 17 does not effectively accumulate, resulting in a higher required concentration to
inhibit parasite growth, or compound 17 accumulates but is not able to inhibit Hz formation
in the DV. The latter hypothesis was ruled out upon observing an increase in nonꢀHz heme in
the cellular fractionation assay from 7% to 41% at the IC50 of 17, with a simultaneous
decrease in the % Hz from 89% for the control to 53% (see supporting information S4.2).
With the biological target for the basic analogue strongly supported, the ability of 17 to
accumulate in parasitized red cells relative to the hit compound was investigated by
measuring the inoculum effect. The effect is only detected when accumulation of the drug
6
5
occurs in the organism, in this case, the parasitized erythrocyte. An apparent decrease in
activity with greater parasitemia can be rationalized by the depletion of total drug
concentration in the extracellular medium upon incubation with infected red blood cells, due
66, 67
to a significant uptake by the parasite.
By measuring the changes in IC50 with varying
initial parasitemia levels, the amount of drug accumulation can be compared for the pyridyl
derivative 15a versus the weak base derivative, 17. An insignificant inoculum effect and
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
cellular accumulation was observed for this derivative, since the IC50 of the compound
remained essentially constant over the range of tested parasitemias and the slope was not
statistically significant. On the other hand, the pyridyl derivative 15a showed larger
accumulation ratios in the cell as demonstrated by the slopes of the linear fits in Figure 12.
The lack of correlation between the calculated pH trapping ratio and the observed
accumulation ratio for the benzamides suggested that the mechanism of drug accumulation is
much more complex, at least in the CQꢀsensitive D10 strain, than the simple pH trapping
model suggests. These studies further showed that the parasite activity of Hz inhibitors
cannot necessarily be improved by adding basic moieties.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
Figure 12. Comparison of inoculum measurements for compounds 15a (r = 0.952; P =
2
0.0045) and 17 (r = 0.828; P > 0.05; not significant).
2.6 Possible role of Fe(III)PPIX association in cellular accumulation.
Although pH trapping could not directly account for the lower accumulation of the
weakly basic benzamide relative to the pyridylbenzamide, some studies have suggested that
despite the lack of correlation between Fe(III)PPIX interaction strength (K) and activity, the
28
ACS Paragon Plus Environment

Page 29 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
total cellular uptake of aminoquinolines and bisꢀquaternary ammonium compounds
6
8ꢀ71
nonetheless occurs via binding to Fe(III)PPIX.
affinity and D6 parasite activity
diethylaminoalkoxyxanthones, showing a dramatic improvement in activity (IC s from 2.2 to
Furthermore, a correlation between heme
was reported for 6ꢀbisꢀωꢀN, Nꢀ
5
0
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3
9
0.07 ꢀM) with a small change in logK (logKs from 5.9 to 5.5). While there is no consensus
as to the validity of the conclusions of these studies, the extent of Fe(III)PPIX binding for
selected benzamides was investigated in an attempt to determine whether there is a
correlation with cellular accumulation. This involved measurement of the association
constants via titration of the compounds into a solution of 7.5 µM monomeric hematin in
4
0% v/v DMSO at pH 7.5. Upon titration of the test compound solution into the hematin
solution, a hypochromism or convergent decrease in intensity of the characteristic monomeric
hematin spectrum at 401 nm (Soret band) was observed owing to significant Fe(III)PPIX
binding. The information gained from this experiment could be extended by monitoring the
maximum absorbance at 405 nm as a function of test compound concentration. This allowed
the association constant (K) to be calculated from a nonꢀlinear least squares fitting model for
a 1:1 complex. A representative example is shown in Figure 13. It should be noted that the
benzamide solutions were made up in 100% DMSO owing to their low solubility in 40%
DMSO. This resulted in a small increase in total percent DMSO in the working solution over
the course of the titration. However, this did not appear to affect the absorbance, and
isosbestic points were observed as expected, confirming a lack of intermediates. In addition,
the fit to the titration curve (Figure 13b) was consistent with 1:1 stoichiometry.
29
ACS Paragon Plus Environment