Structure-Activity Relationship in Pyrazolo[4,3- c]pyridines, First Inhibitors of PEX14-PEX5 Protein-Protein Interaction with Trypanocidal Activity

Article abstract of DOI:10.1021/acs.jmedchem.9b01876

Trypanosoma protists are pathogens leading to a spectrum of devastating infectious diseases. The range of available chemotherapeutics against Trypanosoma is limited, and the existing therapies are partially ineffective and cause serious adverse effects. Formation of the PEX14-PEX5 complex is essential for protein import into the parasites' glycosomes. This transport is critical for parasite metabolism and failure leads to mislocalization of glycosomal enzymes, with fatal consequences for the parasite. Hence, inhibiting the PEX14-PEX5 protein-protein interaction (PPI) is an attractive way to affect multiple metabolic pathways. Herein, we have used structure-guided computational screening and optimization to develop the first line of compounds that inhibit PEX14-PEX5 PPI. The optimization was driven by several X-ray structures, NMR binding data, and molecular dynamics simulations. Importantly, the developed compounds show significant cellular activity against Trypanosoma, including the human pathogen Trypanosoma brucei gambiense and Trypanosoma cruzi parasites.

Full text of DOI:10.1021/acs.jmedchem.9b01876

Subscriber access provided by University of Massachusetts Amherst Libraries
Article
Structure-activity relationship in pyrazolo[4,3-c]pyridines, first inhibitors
of PEX14-PEX5 Protein-Protein Interaction (PPI) with trypanocidal activity
Maciej Dawidowski, Vishal C. Kalel, Valeria Napolitano, Roberto Fino, Kenji Schorpp,
Leonidas Emmanouilidis, Dominik Lenhart, Michael Ostertag, Marcel Kaiser, Marta Kolonko,
Bettina Tippler, Wolfgang Schliebs, Grzegorz Dubin, Pascal Mäser, Igor Tetko, Kamyar
Hadian, Oliver Plettenburg, Ralf Erdmann, Michael Sattler, and Grzegorz M Popowicz
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.9b01876 • Publication Date (Web): 20 Dec 2019
Downloaded from pubs.acs.org on December 22, 2019
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 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 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.
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 125
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Mäser, Pascal; Swiss Tropical and Public Health Institute; University of
Basel
Tetko, Igor; Institute of Structural Biology, Helmholtz Zentrum München
Hadian, Kamyar; Helmholtz Zentrum Munchen Deutsches
Forschungszentrum fur Umwelt und Gesundheit, Assay Development and
Screening Platform
Plettenburg, Oliver; Helmholtz Zentrum Munchen Deutsches
Forschungszentrum fur Umwelt und Gesundheit, Institute of Medicinal
Chemistry; Institute of Organic Chemistry, Leibniz Universität Hannover
Erdmann, Ralf; Institute of Biochemistry and Pathobiochemistry,
Department of Systems Biochemistry, Faculty of Medicine, Ruhr-
University Bochum
Sattler, Michael; Institute of Structural Biology, Helmholtz Zentrum
München; Center for Integrated Protein Science Munich at Chair of
Biomolecular NMR, Department Chemie, Technische Universität München
Popowicz, Grzegorz; Helmholtz Zentrum München, Institute of Structural
Biology; Center for Integrated Protein Science Munich at Chair of
Biomolecular NMR, Department Chemie, Technische Universität München
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
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 125
1
2
3
4
5
6
7
8
9
Structure-activity relationship in pyrazolo[4,3-
c]pyridines, first inhibitors of PEX14-PEX5
Protein-Protein Interaction (PPI) with
trypanocidal activity
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
Maciej Dawidowski1,2,3*, Vishal C. Kalel4, Valeria Napolitano5,6, Roberto Fino1,2, Kenji
Schorpp7, Leonidas Emmanouilidis1,2†, Dominik Lenhart1,2,8, Michael Ostertag1,2, Marcel
Kaiser9,10, Marta Kolonko1,11, Bettina Tippler4, Wolfgang Schliebs4, Grzegorz Dubin6, Pascal
Mäser9,10, Igor V. Tetko1, Kamyar Hadian7, Oliver Plettenburg8,12, Ralf Erdmann4, Michael
Sattler1,2*, Grzegorz M. Popowicz1,2*
1Institute of Structural Biology, Helmholtz Zentrum München, Ingolstädter Landstrasse 1,
85764 Neuherberg, Germany.
2Center for Integrated Protein Science Munich at Chair of Biomolecular NMR, Department
Chemie, Technische Universität München, Lichtenbergstrasse 4, 85747 Garching, Germany.
3Department of Drug Technology and Pharmaceutical Biotechnology Medical University of
Warsaw, Banacha 1, 02-097 Warszawa, Poland
4Institute of Biochemistry and Pathobiochemistry, Department of Systems Biochemistry,
Faculty of Medicine, Ruhr-University Bochum, 44780 Bochum, Germany.
1
ACS Paragon Plus Environment

Page 3 of 125
Journal of Medicinal Chemistry
1
2
3
4
5
6
5Faculty of Biochemistry, Biophysics and Biotechnology, Jagiellonian University,
Gronostajowa 7, Krakow, 30-387, Poland
7
8
9
6Małopolska Center of Biotechnology, Jagiellonian University in Kraków, Gronostajowa 7,
Kraków, 30-387, Poland
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
7Assay Development and Screening Platform, Institute of Molecular Toxicology and
Pharmacology, Helmholtz Zentrum München, Ingolstädter Landstrasse 1, 85764 Neuherberg,
Germany.
8Institute of Medicinal Chemistry, Helmholtz Zentrum München, Ingolstädter Landstr. 1,
Neuherberg, 85764, Germany
9Swiss Tropical and Public Health Institute, Socinstrasse 57, 4051 Basel, Switzerland.
10University of Basel, 4001 Basel, Switzerland.
11Department of Biochemistry, Faculty of Chemistry, Wrocław University of Science and
Technology, Wybrzeże Wyspiańskiego 27, 50-370, Wrocław, Poland.
12Institute of Organic Chemistry, Leibniz Universität Hannover, Schneiderberg 1b, Hannover,
30167, Germany
ABSTRACT
Trypanosoma protists are pathogens leading to a spectrum of devastating infectious diseases.
The range of available chemotherapeutics against Trypanosoma is limited and the existing
therapies are partially ineffective and cause serious adverse effects. Formation of the PEX14-
PEX5 complex is essential for protein import into the parasites’ glycosomes. This transport is
critical for parasite metabolism and failure leads to mislocalization of glycosomal enzymes,
with fatal consequences for the parasite. Hence, inhibiting the PEX14-PEX5 protein-protein
interaction (PPI) is an attractive way to affect multiple metabolic pathways. Herein, we have
used structure-guided computational screening and optimization to develop the first line of
2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 125
1
2
3
4
5
6
7
8
9
compounds that inhibit PEX14-PEX5 protein-protein interaction. The optimization was driven
by several X-ray structures, NMR binding data and molecular dynamics simulations.
Importantly, the developed compounds show significant cellular activity against Trypanosoma,
including the human pathogen T. brucei gambiense and T. cruzi parasites.
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
INTRODUCTION
Human African trypanosomiasis (HAT, sleeping sickness) and Chagas disease (American
trypanosomiasis) are among most devastating and deadliest parasitic diseases. They occur
predominantly in sub-Saharian Africa and Latin America, respectively, but Chagas disease has
spread globally due to international human migration. Both qualify as neglected tropical
diseases (NTDs) threatening millions of people in poor communities.1-4 The impact of
Trypanosoma protists on livestock is also threatening the well-being of multiple communities.5
HAT is caused by two protozoans: Trypanosoma brucei gambiense and Trypanosoma brucei
rhodesiense (accounting for 98% and 2% of HAT cases, respectively), transmitted by the bite
of an infected tsetse fly. The disease consists of two phases: the haemolymphatic stage (stage
I) with trypanosomes being present in the blood and lymphatic system and the
meningoencephalitic stage (stage II), when parasites have spread to the central nervous system
(CNS). Without appropriate diagnosis and treatment the disease can be fatal.3, 6 No effective
vaccine has so far been been developed because the parasites employ several immune evasion
strategies such as antigenic variation. Therefore, chemotherapy is the only therapeutic option
currently available (Figure 1).7, 8 Suramin and pentamidine are used in stage I of HAT, while
melarsoprol, eflornithine and nifurtimox-eflornithine combination therapy (NECT) are used in
stage II of the disease. Unfortunately, these drugs have severe side effects and require
continuous parenteral administration, which is not optimal, given that the patients live in rural
African areas with very limited health resources.6, 9 There is a better treatment option, since
3
ACS Paragon Plus Environment

Page 5 of 125
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
recently a first all-oral drug fexinidazole completed clinical trials in humans and was approved
for HAT treatment.10-13 The efficacy of this monotherapy is comparable to intravenous NECT
and the possibility of oral administration makes it easier to use in rural African realities.
Nevertheless, cross-resistance with nifurtimox is likely to develop due to the same mechanism
of action and no convenient polytherapy is available to overcome this issue.12, 14-16 Another
drug candidate, the benzoxaborole derivative SCYX-7158 (acoziborole) is currently in
advanced phase of clinical trials.12, 17 In spite of these promising developments, the Drugs for
Neglected Diseases initiative (DNDi) still requires new candidates for stage-2 HAT
treatment.15
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 situation is more dramatic for Chagas disease, caused by Trypanosoma cruzi. The parasite
is spread by triatomine bugs (‘kissing bugs’), by blood transfusions, congenitally, and even
orally, by contaminated food.18 It is considered to be the parasitic disease with the greatest
socioeconomic impact in Latin America and one of the leading causes of myocarditis. There
are two stages of the disease. In the acute phase, the symptoms vary between the individuals
and in most cases they decline spontaneously after 1-2 months. In the chronic phase, the
patients may develop potentially fatal cardiac and digestive pathologies. Currently, there are
two drugs for Chagas disease, benznidazole and nifurtimox (Figure 1). However, both have
serious side effects and are ineffective in the chronic phase of the disease.4, 19
Undeniably, there is an urgent need for novel drugs against Trypanosoma infections, especially
for Chagas disease, with high clinical efficiency, favorable pharmacological profiles and ease
of use. In an ideal setting, the pipeline would comprise several new agents that are suitable for
combination therapy, which, in turn, would allow for better efficacy and minimal risk of
resistance development.2, 16 This can be achieved inter alia by finding promising targets and
pathways essential for parasite survival. In the past, various molecular targets have been
4
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 125
1
2
3
4
5
6
7
8
9
proposed and successfully validated.16, 20, 21 It is intriguing that, although the trypanosomatids
share similarities in genetic sequence and biochemical pathways,22, 23 so far most drug
discovery campaigns aimed either at T. brucei or T. cruzi, with fewer seeking for new, broad-
spectrum agents which would affect both species.24-27
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
Glycosomes are peroxisome-related organelles unique to kinetoplastids, including
Trypanosoma spp. These organelles contain the enzymes required for glycolysis and for other
intermediary metabolic pathways.28 Hence, glycosomes are essential for parasite survival and
they have long been considered attractive targets for development of new drugs against
trypanosomatids.20, 29, 30
As glycosomes lack DNA, glycosomal matrix proteins encoded in the nuclear genome have to
be post‐translationally transported from the cytosol to the organelle. Cargo proteins are
imported by a cascade of interactions of specialized transport proteins, called peroxins (PEX#).
The PEX14-PEX5 protein-protein interaction (PPI) plays a key role in this process. The PEX5
import receptor is capable of binding short peroxisomal targeting sequences (PTS) of the cargo.
Next, the PEX5-cargo complex binds PEX14, which is localized at the glycosomal membrane
and is required for cargo translocation into the matrix of the organelle.30, 31 Many of the
glycosomal enzymes have been proposed and validated as potential drug targets.20 However,
targeting the glycosomal transport system should be a more attractive drug target because it
provides the opportunity to interfere with the function of many critical metabolic processes at
once. Furthermore, mislocalization to the cytosol provides the glycosomal kinases access to
the cytosolic ATP pool. Enzymes such as T. brucei hexokinase function at such a high rate that
they deplete ATP faster than it can be replenished by substrate-level phosphorylation, which
will be fatal to the trypanosomes.32
5
ACS Paragon Plus Environment

Page 7 of 125
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
A series of RNAi studies has previously shown the importance of PEX14 function for glucose
metabolism in Trypanosoma.32-34 Recently, we reported a proof-of-concept study for targeting
the PEX14-PEX5 PPI in T. brucei and T. cruzi parasites by small molecules.35 Using a
structure-based drug discovery (SBDD) approach we developed a subset of prototypic
pyrazolo[4,3-c]pyridine derivatives – the first inhibitors of this PPI and the first compounds
that disrupt glycosomal import of matrix proteins and kill T. brucei and T. cruzi in vitro in
nanomolar concentrations. Here, we present the details on the medicinal chemistry program
that led to this group of active molecules along with the in-depth structure–activity relationship
(SAR) study.
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 1. Structures of drugs used for the treatment of Trypanosoma-related diseases and clinical candidates.
RESULTS AND DISCUSSION
Structural basis for PEX14-PEX5 PPI targeting. During the initial phase of the project,
the structural basis for the PEX14-PEX5 PPI interaction was known (Figure 2)36-38. The N-
terminal part of PEX5 contains several diaromatic WXXX(F/Y) motifs that address the
respective hydrophobic Trp and Phe/Tyr pockets on the surface of the N-terminal helical
6
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 125
1
2
3
4
5
6
7
8
9
domain (NTD) of PEX14. These two relatively shallow cavities are separated by two π-
stacked phenyl rings of PEX14 (Phe17 and Phe34, the explanation of residue numbering
among different species is given in Table S1) that protrude to the solvent. In addition to the
hydrophobic and aromatic binding of Trp and Phe/Tyr of WXXX(F/Y) motifs in PEX5 with
their respective hotspots on the PEX14 surface, the positively charged Lys38 and Arg22 side
chains of PEX14 plausibly facilitate the protein-protein complex formation by cation-π
interactions. Other hydrophilic residues that might be of an importance for TbPEX14 ligand
design are Asn13, Glu16 and Asp20. These three residues are different in T. cruzi and human
sequences, thus they might account for the ligands’ cross-species selectivity (Figure 2, right).
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 PEX14-PEX5 protein-protein interface poses several challenges for medicinal chemistry
that are typical for the PPI target class39. First, it is hydrophobic in nature, with only few water
exposed polar amino acid side chains in the spatial vicinity. Furthermore, the interface is
relatively flat, with only two shallow cavities in the highly solvent-exposed surface. The Phe
hotspot appeared as particularly difficult to address, being a ‘shelf’40-42 rather than a classical
PPI pocket.
We created a pharmacophore model that mimicked the native binding of the PEX5
WXXX(F/Y) motif to TbPEX14. Two pharmacophoric elements were set as hydrophobic,
aromatic rings addressing the respective Trp and Phe pockets. Further, we anticipated that the
presence of ‘protruding’ Phe17 and Phe34 side-chains can be an exploitable feature for ligand
design and we defined the third pharmacophore as another connecting aromatic ring that would
allow for additional π-stacking. The in silico screening cascade using a TbPEX14 homology
model prepared from the NMR structure of HsPEX14 NTD (PDB 2W84)36 identified several
15
compounds with convincing docking poses. We tested these hits for binding to N-labelled
1
TbPEX14 NTD using the H-15N HSQC NMR chemical shift perturbations (CSP) assay. A
7
ACS Paragon Plus Environment

Page 9 of 125
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
pyrazolo[4,3-c]pyridine derivative 1 was the most potent hit, with KD of 163 µM in the NMR
CSP analysis and was resynthesized. The compound disrupted the interaction of the PEX5-
derived peptide (ALSENWAQEFLA) with both TbPEX14 and TcPEX14, having EC50 of 265
µM and 539 µM, respectively, in the AlphaScreen43 assays (Figure 3 and Table 1).
Furthermore, 1 is characterized by a proper drug-likeness profile and was chemically tractable,
rendering this compound an attractive starting point for medicinal chemistry optimization.
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 docking results suggested that the phenyl residue of 1 addressed the Trp pocket on the
TbPEX14 surface, while the indole moiety filled the Phe hotspot (Figure 4, top). Such
placement of aromatic residues is in fact a mirror image of the native binding mode of PEX5
WXXX(F/Y) fragments to PEX14. However, we have previously shown that the Trp pocket of
PEX14 can effectively accommodate Phe residue of F/YXXXF motif in its PPI with another
peroxin, PEX19.36 Another hallmark of the docking-derived binding pose of 1 to PEX14 is that
the pyrazolo[4,3-c]pyridine central scaffold of the hit compound lays over the Phe17 and Phe34
residues, forming the favorable π-π interactions. We have also identified lower-scored poses,
in which the projection of the aromatic phenyl and indole residues of 1 in TbPEX14 cavities
was reversed, i.e. directly mimicking the native placement of the aromatic rings in the PEX14-
PEX5 PPI (Figure 4, bottom). Although this inverted binding pose seemed less plausible due
to the number of possible unfavorable interactions with the hydrophilic residues present in the
vicinity of the hotspots, it could not be completely disregarded because the overall shape of the
PEX14-PEX5 PPI interface is highly symmetrical.
8
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 125
1
2
3
4
5
6
7
8
9
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. Structural basis for PEX14-PEX5 PPI interface targeting. Left: Solution NMR
structure of the HsPEX14 NTD in complex with the di-aromatic WXXXF pentapeptide motif
of PEX5 (PDB code: 2W8436). PEX14 is shown in light blue, whereas PEX5 is shown in
magenta. Two hydrophobic cavities on PEX14 surface accommodating Trp103 and Phe107
residues of PEX5 WXXXF are separated by two π-stacked phenylalanine side chains (Phe17
and Phe34). Right: Superposition of the solution NMR structures of HsPEX14 (light blue,
PDB code: 2W84) and TbPEX14 (gold, PDB code: 5MMC35) NTDs. The polar residues
(represented as sticks) surrounding the binding pocket are significantly different in both
species. Bottom: Sequence alignment of HsPEX14, TbPEX14 and TcPEX14.
Figure 3. Structure, properties and proposed derivatization sites of the pyrazolo[4,3-
c]pyridine in silico hit 1.
9
ACS Paragon Plus Environment

Page 11 of 125
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
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. Docking poses of 1 to TbPEX14-PEX5 PPI interface. Top: Phenyl residue of 1
occupies the Trp pocket while the indole is positioned in the Phe pocket on the PEX14 surface.
Bottom: The ‘reversed’ binding mode of 1 to TbPEX14-PEX5 PPI interface. Surface (left) and
cartoon (right) representation of TbPEX14 (gold).
Initial optimization trials. For optimization purposes we divided the initial hit 1 into four
regions: 1) the pyrazole N substituent, 2) the residue addressing Trp pocket, 3) the residue
addressing Phe pocket and 4) the central scaffold (Figure 3).
According to the preferred docking pose of 1, the methyl group on pyrazole N-1 atom points
to the solvent-exposed region of the interface, suggesting that substituents of different size
should be tolerable at this position. Despite the proximity of Asn13 (Ser13 in TcPEX14 and
Thr13 in HsPEX14) and Glu16 (Gln16 in TcPEX14 and Lys16 in HsPEX14), we expected that
modifications addressing this region of the interface may affect the ligand’s physicochemical
properties (e.g. solubility) rather than significantly influence the affinity to the target. We have
found that the binding to PEX14 depends on the position of the substituent in the pyrazole ring.
10
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 125
1
2
3
4
5
6
7
8
9
Contrary to the active hit 1, the N-2 regioisomer 2 showed modest inhibitory activity in the
AlphaScreen assay (Table 1) and did not bind TbPEX14 in a 1H-15N HSQC NMR experiment.
This was inconsistent with the anticipated binding mode of 1 to TbPEX14, since the distance
to the proximal Asn13 and Glu16 should allow the methyl group to be tolerated at the N-2
position of the pyrazole. Compound 3 lacking substituents at the N-1 atom of the pyrazole had
a slightly decreased ability to disrupt PEX5 interaction with both T. brucei and T. cruzi PEX14,
and it was inactive against the human protein. Likewise, the hydroxyethyl derivative 4 showed
no superior activity compared to 1. Carboxylate derivatives 5-7 proved inactive. In conclusion,
the initial SAR of the solvent-exposed region could not be fully rationalized on the basis of
molecular docking. We have chosen the hydroxyethyl residue in this position due to the
superior solubility of the compounds and sought further structural modifications that would
yield more potent compounds suitable for co-crystallization trials with TbPEX14.
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
Since the binding of the PEX5 WXXX(F/Y) motifs to PEX14 is hydrophobic and aromatic in
nature, we assumed that optimizing the residues addressing the Trp and Phe hotspots would
likely improve ligand affinity. Structural data show that the Trp pocket is quite shallow (Figures
2 and 4). As a result, the orientation of the indole ring in the native PEX14-PEX5 PPI is vertical
with respect to the surface of the interface, and the imidazole component of the fused system
slightly points out from the hotspot cavity, being surrounded by Lys38 and Asn13 (Ser13 in
TcPEX14 and Thr13 in HsPEX14). On the other hand, this pocket has an elongated shape and
the docking pose of 1 suggested that there is space that could be filled by a residue larger than
the unsubstituted phenyl. Hence, we first modified the substitution pattern on the phenyl ring
(Table 1). We synthesized derivatives 8-12 having various small lipophilic substituents in
meta- and para- position, none of which showed superior activity or species selectivity, when
compared to 1. In contrast to the flat SAR of the substituted phenyls, derivatives 13 and 14
bearing naphthalene moiety, were more effective in disrupting the TbPEX14-PEX5 PPI, with
11
ACS Paragon Plus Environment

Page 13 of 125
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
AlphaScreen-derived EC50 values 5 and 8-fold lower than the respective parent compounds 1
and 4. We have also observed the increased ability of these derivatives to interrupt the Tc and
HsPEX14-PEX5 complex formation.
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
As already mentioned, the Phe hotspot on the PE14-PEX5 PPI interface appeared more difficult
to target than the Trp pocket because this cavity is shallow and more exposed to solvent. The
docking pose of 1 in TbPEX14 showed that this region of the interface is well-filled by the
indole moiety of the inhibitor. In fact, replacement of the bicyclic aromatic system of 1 with
smaller phenyls yielded only very weak compounds 15-18 (Table 1). Interestingly, inhibitor 19
bearing a bulky naphthalene was inactive, while its 4-methoxy derivative 20, proved 10-fold
more potent in disrupting the TbPEX14-PEX5 PPI than 4.
Secondary optimization – Phe pocket. The above SAR was in large parts difficult to
rationalize without additional structural data. Fortunately, at this stage of the project we were
able to obtain a high-resolution X-ray structure of TbPEX14 bound to 13 (PDB code: 5L8735,
Figure 5) that guided further optimization. Overall, the crystallographic data were consistent
with the binding mode of 1 to TbPEX14 as predicted by docking. According to the TbPEX14-
13 complex structure, the double aromatic naphthalene system effectively filled the Trp
hotspot. However, unlike the vertically oriented Trp side chain in the native PEX5
WXXX(F/Y)-PEX14 PPI, the orientation of the naphthyl group in the Trp pocket is horizontal.
A very important feature of the X-ray structure that could not have been predicted on the basis
of docking is the presence of one structural water that mediated binding of 13 to the TbPEX14
Asn13 side chain. This has a pronounced influence on the ligand binding mode. Most of all,
the ligand is shifted towards Asn13 and Glu16 residues with respect to its orientation
speculated on the basis of docking. As a consequence, the indole residue of 13 is shifted out
towards the upper rim of the Phe pocket rather than being deeply inserted into this cavity as we
12
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 125
1
2
3
4
5
6
7
8
9
postulated previously. Thus, the filling of the hotspot by the indole can be considered shallow
and suboptimal. On the other hand, the aromatic residue of 13 is closer to the Arg22 side chain
of PEX14, which, in turn, can stabilize the ligand-protein complex by forming cation-π
interactions. The presence of the above-mentioned thermodynamically stable water molecule
may also explain the loss of affinity of 2, in which the methyl group attached to the N-2
pyrazole nitrogen atom interferes with the solvation shell of the protein-ligand complex.
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
Driven by the obtained high-resolution structural information we decided to systematically
exploit the SAR of the Phe hotspot, while keeping the Trp pocket occupied by the naphthyl
(Table 2). The overall optimization strategy was thus focused on preserving the orientation of
the aromatic portion of the substituents addressing the Phe hotspot close to its rim, while
simultaneously extending the ligand deeper towards its bottom. First, we synthesized
compounds 21-24 bearing various small substituents in the N-1 position of the indole system,
of which only the methyl and ethyl derivatives 21 and 22, respectively, showed potency
comparable to 14 in the AlphaScreen assay. On the other hand, compounds 23 and 24, with
larger, more polar pendant residues, were slightly less active in disrupting the PEX14-PEX5
PPI. As seen in the TbPEX14-13 complex structure, the Phe pocket of the interface is filled
tightly by the indole residue of the ligand. Hence, there is no possibility of rotation along the
N-CH2 bond that links the indole residue to the central scaffold, which would be required to
accommodate large indole N-1 substituents. Furthermore, as observed in the TbPEX14-13
complex structure, the ligand binds to the interface in its low-energy conformation. Hence, a
considerable entropic penalty would have to be paid for the N-1 substituted ligand adaptation
to the hotspot.
While indole N-1 substituents are in a steric conflict with the amino acid side chain forming
the Phe pocket and there is no possibility to compensate this by the adjustment of ligand’s
13
ACS Paragon Plus Environment

Page 15 of 125
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
orientation, the analysis of the TbPEX14-13 complex structure showed that there is enough
space in the hotspot to accommodate substituents in the C-7 position. Therefore, in the next
optimization round we synthesized compounds 25-27 having small residues attached to this
carbon atom (Table 2). Gratifyingly, these proved superior to 14 in disrupting the TbPEX14-
PEX5 complex formation, each being 2 times more potent than the parent compound in the
AlphaScreen assay. As of Tc and HsPEX14-PEX5 PPI inhibition, the results varied. Compound
25 had a comparable activity to the parent 14, while its N-1 methyl derivative 26 was 2-times
more active. On the other hand, 27 bearing a methyl ester function on C-7 of indole was 2-fold
less potent in inhibiting the T. cruzi protein. We have also synthesized 28, in which indole was
attached to the central scaffold through C-2 carbon atom, allowing for a deeper insertion of the
aromatic residue into the pocket (Table 2). As a result, 28 was more active than its C-3
regioisomer 14 and roughly equipotent to the isomeric methoxy indole derivative 26.
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 different activities of the undecorated indole derivative 14 and 26 resulting from the
unequally deep occupation of the Phe pocket recalled the observations made previously for
naphthalenes 19 and 20 during the initial optimization trials. With the aim to improve binding
of the ligands and to further investigate SAR in the Phe pocket, we solved the crystal structure
of the methoxynaphthalene derivative 20 in complex with TbPEX14 (PDB code: 5L8A35,
Figure 6). Overall, the binding mode of 20 to TbPEX14-PEX5 PPI interface closely resembled
the one observed for 13 in the TbPEX14-13 complex. Thus, the phenyl ring of the inhibitor
was buried in the Trp pocket, while the naphthalene system filled the upper part of the Phe
hotspot, with the methoxy residue reaching the bottom part of the cavity. Importantly, the
orientation of the pyrazolo[4,3-c]pyridine central scaffold in 13 and 20 was identical, which
was a clear indication that a more potent PEX14-PEX5 PPI inhibitor can be obtained by
merging those two ligands. Indeed, the hybrid molecule 29 showed a superior activity in the
AlphaScreen TbPEX14-PEX5 PPI inhibition assay, when compared to both parent compounds
14
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 125
1
2
3
4
5
6
7
8
9
(Table 2). The methoxy group present at the C-4 atom of the naphthalene in 29 was crucial for
the high-affinity interaction with TbPEX14, since derivatives 30 and 31 lacking this residue
proved significantly less active. Next, we synthesized 32, in which one of the fused benzenes
of the naphthalene system was saturated. The compound was 3-times less potent than the parent
29, most likely due to its inability to form cation-π and π-π interactions with the Arg22 and
Phe17 residues, respectively. Next, we investigated compounds 33-35 having small
substituents other than methoxy at C-4 carbon of the naphthalene. Switching to a larger
methylthiol group in 33 resulted in a 3-times stronger inhibition of the TcPEX14-PEX5
complex formation when compared to 29. Sulfoxide 34, an expected metabolite of 33, was
significantly less potent. The dimethylamino derivative 35 was as active as the parent
compound 29 in inhibiting the TbPEX14-PEX5 PPI, and significantly more potent against Tc
and HsPEX14.
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
Finally, we synthesized a series of phenyl derivatives 36-45 having small lipophilic substituents
in para- or (and) meta- positions (Table 2). By this, we wanted to verify whether a single
aromatic system can outcompete, or at least be as active as, the bulkier and highly lipophilic
indoles and naphthalenes. By changing the substitution patterns on the phenyl, we have
observed that a gain of potency can be achieved by expanding the ligand towards the bottom
of the pocket, similar to what had been observed for the bicyclic aromatic systems in 14 and
21-35. Several phenyl derivatives were equipotent to 14 in disrupting the TbPEX14-PEX5
complex formation, however, none of them could outcompete the more potent 25-28 nor the
C-4 substituted naphthalenes 29 and 33-35.
15
ACS Paragon Plus Environment

Page 17 of 125
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
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 5. Crystal structure of TbPEX14 in complex with 13 (PDB code: 5L8735). Left:
surface representation of TbPEX14 (gold). A highly conserved water molecule is represented
as a sphere. Right: cartoon representation of TbPEX14; relevant residues are shown as sticks.
A water-mediated hydrogen bond between PEX14 N13 side chain and the amide nitrogen of
13 is presented (gray dashed line).
Figure 6. Top: Crystal structure of TbPEX14 in complex with 20 (PDB code: 5L8A35).
Bottom: overlay of the binding poses of 13 (purple) and 20 (green). Surface (left) and cartoon
(right) representation of TbPEX14 (gold).
16
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 125
1
2
3
4
5
6
7
8
9
Secondary optimization – around Trp pocket. To further expand the SAR information in
PEX14-PEX5 PPI targeting we turned our attention to the region of the interface near the
entrance to the Trp hotspot. The crystallographic structures of TbPEX14 in complex with 13
and 20 show that the amide moiety that projects the aromatic residue of the ligand into the Trp
pocket is in a low-energy conformation, i.e. coplanar with the pyrazole ring. Hence, we posed
the question whether an entropic gain from freezing this favorable conformation can be
achieved and we synthesized 46 and 47, tricyclic derivatives of 29 (Table 3). Both compounds
were less potent than the parent inhibitor in the AlphaScreen assay. As previously mentioned,
the TbPEX14-13 complex structure shows an important interaction of the amide NH proton
with Asn13 of the protein, mediated through a water molecule. Hence, it is possible that the
gain in entropy in 46 and 47 from freezing the molecules in their low-energy conformations
cannot compensate the loss in binding affinity caused by the inability of forming the mentioned
important connection through water. Intriguingly, this structural change proved deleterious for
the ability of the inhibitors to bind Tc and HsPEX14 but did not abolish its interaction with the
T. brucei protein. As a result, 46 and 47 are highly selective in inhibiting the latter PPI. To
further explore the importance of the amide linkage in this region of the ligand, we synthesized
48, a derivative of 29 in which carbonyl oxygen was removed (Table 3). The obtained inhibitor
projecting the naphthalene into Trp pocket through alkylamino linker was significantly weaker
in inhibiting the PEX14-PEX5 PPI than the parent molecule. We attributed this to the repulsive
interactions with Lys38, weaker hydrogen bond accepting properties of the amine and to the
increased conformational freedom of the linker when compared to the amide. With the aim to
explore the possible polar interactions with Asn13 and Lys38 surrounding the entrance to the
Trp pocket, we synthesized sulfonamide 49 (Table 3). The compound failed to inhibit the
TbPEX14-PEX5 PPI, likely due to the inability of the relatively narrow Trp pocket to
accommodate the sulfonamide moiety.
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
17
ACS Paragon Plus Environment

Page 19 of 125
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Secondary optimization – amines in N-1 pyrazolo substituent. Our initial optimization trials
showed that the substitution pattern at N-1 of pyrazole does not significantly alter binding of
the ligands to the PEX14-PEX5 PPI interface. However, a closer inspection of the TbPEX14-
20 complex structure revealed that the pendant hydroxyalkyl chain can adopt a conformation
in which the hydroxyl lies in a spatial proximity of the carboxylic amino acid residues, Glu16,
Asp20 in the solvent-exposed region. Therefore, we posed a question if a switch to an amine
can lead to a further improvement in the binding affinity originating from additional
electrostatic contacts. Indeed, a simple –OH to –NH2 transition resulted in primary amine 50
which was 6-fold more active in disrupting the TbPEX14-PEX5 PPI and 5-fold more potent
against TcPEX14, respectively, than the parent 29 (Table 4). Encouraged by this result, we
pursued a series of derivatives 51-61 with various alkylamine residues at the N-1 pyrazolo
position. Secondary and tertiary amines were tolerated. An important observation was that
binding to PEX14 correlates positively with the basicity of the substituents (e.g. 50 vs 58, 54
vs 57, 55 vs 56, 59 vs 60) rather than with their hydrogen bond donor or acceptor capabilities
(e.g. 50 vs 58).
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
We solved high-resolution crystal structures of 50 and 61 in complex with the TbPEX14 N-
terminal domain (PDB codes: 5N8V35 and 6SPT, Figures 7 and 8, respectively). The TbPEX14-
50 complex structure provides a snapshot of the pendant amino group interacting with Glu16
(Figure 7). However, this interaction is not seen in the TbPEX14-61 complex structure (Figure
8), which reflects the flexibility of the hydroxyalkyl chain. As X-ray crystallography may not
always precisely reveal the flexible ligand-protein interactions due to the crystal symmetry-
related contacts, we analyzed the binding of the amino groups in 50 and 61 and hydroxyl of 29
with the carboxylic residues of Glu16 and Asp20 of TbPEX14 by performing molecular
dynamics (MD) simulations (Figure 9 and Figure S2 – see Supporting Information). The
computations clearly show the high occurrence of interactions of the amino groups with both
18
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 125
1
2
3
4
5
6
7
8
9
residues. In contrast, the MD simulation did not show significant interaction of the pendant
hydroxyethyl group of derivative 29. Here, the occurrence of the interaction during the MD run
is much lower.
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 negatively charged Glu16 residue is present in the TbPEX14 amino acid sequence, but not
in TcPEX14 (Gln16) or HsPEX14 (Lys16). Furthermore, in the HsPEX14 sequence Asp20 is
replaced by the neutral Asn20. In this context, it is intriguing that the ligands bearing positive
charge (i.e. amines 50-55, 59 and 61) show enhanced inhibitory activity in the latter two PPIs.
On the other hand, it has been shown that the binding affinity and thermodynamics in solvent-
exposed protein cavities, such as those present on PEX14 surface, is not only affected by the
direct interactions of ligand atoms with protein residues, but can also depend significantly on
the alterations of the water envelope around the ligand-protein complex.44 Indeed, both
TbPEX14-50 and TbPEX14-61 complex X-ray structures document the involvement of the
flexible alkylamines in the water network around the ligand. Similar engagement can be
expected in Tc and HsPEX14, which might explain the enhancement in inhibitory activity of
the ligands against the PPIs of those proteins with PEX5.
Overall, the sum of the effects described above provided a significant improvement of the
activity of the ligands against all three tested PEX14 proteins. This enhancement was most
pronounced in the case of TbPEX14-PEX5 PPI, which resulted in the higher selectivity of
compounds towards this PPI.
19
ACS Paragon Plus Environment

Page 21 of 125
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
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 7. High resolution crystal structure of N-terminal domain of TbPEX14 in complex with
compound 50 (PDB code: 5N8V35). Surface (left) and cartoon (right) representation of
TbPEX14 (gold).
Figure 8. High resolution crystal structure of N-terminal domain of TbPEX14 in complex
with compound 61 (1.2 Å, PDB code: 6SPT). Surface (left) and cartoon (right) representation
of TbPEX14 (gold).
20
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 125
1
2
3
4
5
6
7
8
9
Figure 9. Ligand interaction diagrams of the final molecular dynamics simulations snapshots.
The MDs were performed for TbPEX14-29 (A), TbPEX14-50 (B) and TbPEX14-61 (C)
complexes. Blue lines represent H-bonding interactions, red cation-π and green π –π
interactions. Introduction of alkylamine residues at N-1 pyrazolo position leads to formation
of direct and water-mediated interaction pattern significantly improving compound affinity.
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
Secondary optimization – central scaffold. Finally, we decided to obtain some SAR
information of the central bicyclic scaffold. Quaternization of the tertiary amine in 29 resulted
in a permanently charged derivative 62, which did not show superior binding to PEX14 (Table
5). Decreasing the size of the aliphatic ring of 29 by one methylene unit produced the
pyrrolo[3,4-c]pyrazole 63, which was 2-fold more active than the parent compound in
inhibiting the TbPEX14-PEX5 PPI, and 3-fold more potent against TcPEX14 and HsPEX14.
The amino derivative of alcohol 63, compound 64, showed the expected enhancement in
PEX14 binding, with respect to its 6-membered ring homologue 50. Finally, the amino alcohol
65 was the most potent PEX14 binder, representing a first inhibitor of TbPEX14-PEX5 PPI
with submicromolar EC50 value in the AlphaScreen assay.
Cellular activity of the PEX14-PEX5 PPI inhibitors. The synthesized compounds were
evaluated for their trypanocidal effect against T. brucei brucei bloodstream form (Lister 427,
MITat 1.2) in the resazurin-based cell survival assay. T. b. brucei is the causative agent of
animal trypanosomiasis and was used as a model organism for the primary screening of
trypanocidal activities of the inhibitors. The obtained EC50 values are presented in Tables 1-5.
Besides several outliers, the trypanocidal activity of the tested derivatives correlates well with
the results of the AlphaScreen-based TbPEX14-PEX5 PPI inhibition assay. We have also
observed a remarkable enhancement of the antiparasitic activity in compounds bearing an
aminoalkyl chain attached to the N-1 position of the pyrazole, resulting in submicromolar EC50
values in the resazurin-based assay (Tables 4 and 5). We attribute this to the greater inhibitory
21
ACS Paragon Plus Environment

Page 23 of 125
Journal of Medicinal Chemistry
1
2
3
4
potency of those compounds against the TbPEX14-PEX5 PPI. Compounds 50-52 were most
5
6
active from the series, having EC50 values below 100 nM (Table 4).
7
8
9
Additionally, the investigated PEX14-PEX5 PPI inhibitors were assayed for their cytotoxicity
in the human-derived HepG2 cell line (Tables 1-5). In general, the cytotoxicity of some of the
compounds followed their inhibitory activity against HsPEX14-PEX5 PPI. This was most
pronounced for many of the indole derivatives (eg. 2-8, 14, 21-28). On the other hand, some
other active HsPEX14 ligands did not follow this trend (eg. naphthalenes 20, 29, 35). Overall,
the selectivity index for trypanocidal activity against T. b. brucei improved along with
increased efficacy of compounds in disrupting PEX14-PEX5 PPI. This was most pronounced
in the alkylamine derivatives (Tables 4 and 5), many of which displayed favorable selectivity
profiles (eg. 50, 51, 53, 54, 61, 64, 65).
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 PEX14-PEX5 PPI inhibitors showing high trypanocidal efficacy against T. b. brucei were
further evaluated activity against T. b. rhodesiense STIB900, a human pathogen derived from
an African (United Republic of Tanzania) patient (Table 6). In general, the results followed the
SAR pattern observed in the primary T. b. brucei screening. The T. b. rhodesiense STIB900
strain was even more susceptible to the PEX14-PEX5 PPI inhibitors, which resulted in more
preferable selectivity indices. The high activity of the compounds against the human pathogen
may have practical implications and merit further development of the PEX14-PEX5 PPI
inhibitors as potential treatments against HAT.
Unlike T. brucei, T. cruzi parasites do not use glycolysis as the main energy source for the cell.
However, they still use glycosomes to compartmentalize other essential metabolic processes.
Hence, we wanted to check whether impairing the function of the organelle by PEX14-PEX5
PPI inhibition is also lethal to T. cruzi. When assayed for their trypanocidal activity against the
amastigote forms of the T. cruzi Tulahuen strain C2C4 grown in rat L-6 myoblasts (Table 6),
22
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 125
1
2
3
4
5
6
7
8
9
the tested PEX14-PEX5 PPI inhibitors demonstrated varied efficiency. The IC50 values were
higher than those obtained for T. brucei and thus the selectivity indices determined with
uninfected L-6 cells were narrower. The aminoalkyl derivatives 50, 61 and 64 displayed highest
potency, showing sub-micromolar trypanocidal activities and fair therapeutic indices. A
general observation was that the correlation between the ability to break the PEX14-PEX5 PPI
and trypanocidal efficacy is less clear in T. cruzi than in T. brucei. Presumably, this reflects the
characteristics of the assay, in which the T. cruzi parasites are located within myoblasts and
therefore the compounds have to pass multiple cellular membranes and environments to reach
the target.
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 1. Initial optimization of 1.
R1
O
HN
N
R2
N
N
R3
TbPEX14 TcPEX14 HsPEX14 T. brucei
HepG2
#
1
R1
R2
R3
SIc
EC50 [µM]a EC50 [µM]a EC50 [µM]a EC50 [µM]b EC50 [µM]b
18.5
265±37
>1000
539±75
>1000
223±30
>1000
>1000
>1000
>1000
>50
>2.7
CH3
(15.2-22.6)
NH
NH
NH
NH
NH
18.6
2d
>50
>2.7
>5.7
>2.4
nd
CH3
(13.7-25.2)
5.16
29.4
3
441±47
421±19
>1000
768±150
828±127
>1000
H
(3.02-7.30) (20.8-41.6)
21.0
4
>50
OH
(17.8-24.7)
COOH
5
>50
>50
23
ACS Paragon Plus Environment

Page 25 of 125
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
16.7
COOEt
6
7
8
9
550±144
>1000
>1000
>1000
384±69
505±42
>1000
>50
>3.0
>9.7
>3.5
nd
(13.6-20.5)
NH
NH
NH
NH
3.05
29.7
>1000
COOEt
(2.83-3.27) (22.5-36.9)
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
5.74
740±167
559±33
555±120
509±35
>20
OH
(4.57-7.22)
Cl
11.5
nd
OH
OH
(6.75-19.7)
F
5.19
Cl
10
381±60
632±83
218±15
(3.04-7.34) nd
nd
NH
11.8
34.8
CH3
11
12
233±25
422±29
625±94
244±21
427±54
2.9
nd
OH
OH
(3.9-35.7) (31.3-38.7)
NH
NH
4.25
CF3
879±140
nd
(3.84-4.71)
2.96
14.4
13
14
51.3±2.1
53.7±4.8
99.5±25.7 51.1±4.1
4.9
2.7
CH3
(2.84-3.08) (12.8-16.8)
NH
NH
4.50
7.87
68.0±4.3
57.9±3.3
OH
(3.87-5.23) (7.28-8.44)
2.15
15
16
>1000
>1000
>1000
nd
>1000
nd
>50
>23.2
nd
OH
OH
(1.42-3.26)
Cl
nd
nd
nd
nd
17
18
>1000
nd
nd
nd
OH
OH
Cl
1.74
627±76
>1000
>1000
>50
>28.7
(0.82-3.60)
OCH3
24
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 125
1
2
3
4
5
6
7
8
9
19
20
>1000
nd
nd
nd
nd
nd
OH
OH
9.33
34.9
46.3±0.40 704±449
135±15
>3.6
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
(8.43-10.3) (28.3-41.5)
OCH3
a EC50 values were calculated as a Hill curve fit to 12 point titration. SD represents fitting
error. nd = not determined. bEC50 values are shown as mean (n=4). Values in parentheses are
95% confidence intervals. cSelectivity index is calculated as HepG2 EC50 [µM]/T. brucei
EC50 [µM]. dN-2 regioisomer of 1.
25
ACS Paragon Plus Environment

Page 27 of 125
Journal of Medicinal Chemistry
1
2
3
4
Table 2. Secondary optimization of the residue addressing the Phe pocket.
5
6
7
8
9
O
HN
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
N
R
N
N
HO
TbPEX14
TcPEX14
HsPEX14
T. brucei
HepG2
#
R
SIc
EC50 [µM]a EC50 [µM]a EC50 [µM]a EC50 [µM]b EC50 [µM]b
4.50
7.87
14
53.7±4.8
68.0±4.3
73.2±4.6
57.9±3.3
43.9
1.7
3.3
(3.87-5.23) (7.28-8.44)
NH
N
2.62
8.60
21
22
40.2±2.1
49.4±4.8
(2.42-2.85) (7.25-10.14)
2.86
68.5±3.5
166±15
63.4
141
nd
nd
(2.18-3.76)
N
N
5.70
22.7
23
24
86.7±10.1
80.9±19.5
4.0
(2.82-11.70) (18.0-27.4)
O
6.09
13.3
118±12
176
2.2
N
O
O
(5.66-6.56) (9.9-17.5)
S
4.38
13.8
25
26
27
24.4±0.1
25.4±1.6
28.9±1.5
65.1±11.1
37.3±2.0
102±38
51.9±2.0
23.9±1.9
59.5±3.8
3.2
2.2
2.7
OCH3
OCH3
(4.08-4.70) (11.9-15.6)
NH
N
3.36
7.34
(2.90-3.90) (6.26-8.42)
5.46
12.9
COOMe
(4.59-6.22) (13.2-16.6)
NH
26
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 125
1
2
3
4
5
6
7
8
9
5.41
9.39
28
29
20.3±2.6
14.8±1.9
43.6±4.4
104±40
35.5±1.6
39.7±5.0
1.7
HN
(4.29-6.84) (8.64-10.21)
3.56
>50
>14.0
(3.29-3.85)
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
OCH3
7.95
9.11
30
31
454±143
565±369
>1000
>1000
>1000
>1000
1.5
nd
(6.74-9.38) (8.93-9.24)
>20
nd
>50
nd
OH
32
33
38.3±1.7
19.0±4.4
142±49
60.8±4.7
nd
nd
OCH3
3.92
38.0±11.8
32.4±14.8
nd
(2.78-5.51)
SCH3
8.40
29.2
34
35
122±27
168±11.1
21.0±2.2
437±137
16.4±2.3
3.5
(7.53-9.36) (25.9-32.5)
S
O
5.70
16.2±5.9
>50
>8.8
(4.81-6.76)
N
13.1
38.2
36
37
38
200±25
75.1±2.9
55.7
>1000
387±178
nd
503±55
181±19
nd
2.9
nd
nd
(11.3-14.9) (30.6-47.1)
7.26
nd
(6.07-8.45)
7.62
nd
(6.68-8.72)
3.29
39
56.2±5.6
126±51
149±30
nd
nd
(3.16-3.44)
27
ACS Paragon Plus Environment

Page 29 of 125
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
6.22
40
41
42
43
44
45
51.7±4.5
53.2±7.0
95.0±14.8
50.9±4.8
44.7±5.4
137±63
369±201
229±44
280±54
287±64
354±173
259±160
113
nd
nd
3.5
nd
3.9
nd
nd
(5.90-6.52)
3.98
O
14.1
86.1±8.6
116±15
138±13
95.5±9.8
422±342
(3.83-4.14) (12.3-16.1)
Cl
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
5.64
nd
(5.08-6.21)
OCF3
OCH3
OCH3
10.4
40.8
(8.3-12.9)
(34.7-46.9)
Cl
Cl
4.29
nd
(3.36-5.22)
13.8
nd
(12.3-15.4)
aEC50 values were calculated as a Hill curve fit to 12 point titration. SD represents fitting
error. nd = not determined. bEC50 values are shown as mean (n=4). Values in parentheses are
95% confidence intervals. cSelectivity index is calculated as HepG2 EC50 [µM]/T. brucei
EC50 [µM].
Table 3. Secondary optimization around the Trp pocket.
TbPEX14
TcPEX14
HsPEX14
T. brucei
HepG2
#
R
SIc
EC50 [µM]a EC50 [µM]a EC50 [µM]a EC50 [µM]b EC50 [µM]b
6.42
O
46
49.2±19.5
463±203
448±408
>1000
>1000
>50
>50
>7.8
>8.9
N
(5.61-7.20)
N
N
N
N
OCH3
5.61
47
103±17
N
(5.13-6.11)
N
O
N
OCH3
28
ACS Paragon Plus Environment