Antimalarial Inhibitors Targeting Serine Hydroxymethyltransferase (SHMT) with in Vivo Efficacy and Analysis of their Binding Mode Based on X-ray Cocrystal Structures

Article abstract of DOI:10.1021/acs.jmedchem.7b00008

Target-based approaches toward new antimalarial treatments are highly valuable to prevent resistance development. We report several series of pyrazolopyran-based inhibitors targeting the enzyme serine hydroxymethyltransferase (SHMT), designed to improve microsomal metabolic stability and to identify suitable candidates for in vivo efficacy evaluation. The best ligands inhibited Plasmodium falciparum (Pf) and Arabidopsis thaliana (At) SHMT in target assays and PfNF54 strains in cell-based assays with values in the low nanomolar range (3.2-55 nM). A set of carboxylate derivatives demonstrated markedly improved in vitro metabolic stability (t_1/2 > 2 h). A selected ligand showed significant in vivo efficacy with 73% of parasitemia reduction in a mouse model. Five new cocrystal structures with PvSHMT were solved at 2.3-2.6 ? resolution, revealing a unique water-mediated interaction with Tyr63 at the end of the para-Aminobenzoate channel. They also displayed the high degree of conformational flexibility of the Cys364-loop lining this channel.

Full text of DOI:10.1021/acs.jmedchem.7b00008

Subscriber access provided by CORNELL UNIVERSITY LIBRARY
Article
Antimalarial Inhibitors Targeting Serine Hydroxymethyltransferase
SHMT) with In Vivo Efficacy and Analysis of their
(
Binding Mode Based on X-ray Cocrystal Structures
Geoffrey Schwertz, Matthias C. Witschel, Matthias Rottmann, Roger Bonnert, Ubolsree Leartsakulpanich,
Penchit Chitnumsub, Aritsara Jaruwat, Wanwipa Ittarat, Anja Schäfer, Raphael A. Aponte, Susan
A. Charman, Karen L. White, Abhijit Kundu, Surajit Sadhukhan, Mel Lloyd, Gail M. Freiberg,
Myron Srikumaran, Marc Siggel, Adrian Zwyssig, Pimchai Chaiyen, and François Diederich
J. Med. Chem., Just Accepted Manuscript • Publication Date (Web): 24 May 2017
Downloaded from http://pubs.acs.org on May 24, 2017
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 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Antimalarial Inhibitors Targeting Serine
Hydroxymethyltransferase (SHMT) with In Vivo
Efficacy and Analysis of their Binding Mode Based
on Xꢀray Cocrystal Structures
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
†
‡
§,ø
||
Geoffrey Schwertz , Matthias C. Witschel , Matthias Rottmann , Roger Bonnert , Ubolsree
§
,ø
⊥
⊥
⊥
⊥
Leartsakulpanich , Penchit Chitnumsub , Aritsara Jaruwat , Wanwipa Ittarat , Anja Schäfer ,
‡
∆
∆
◊
◊
Raphael A. Aponte , Susan A. Charman , Karen L. White , Abhijit Kundu , Surajit Sadhukhan ,
#
∞
∞
†
†
Mel Lloyd , Gail M. Freiberg , Myron Srikumaran , Marc Siggel , Adrian Zwyssig , Pimchai
†
,*
≡
Chaiyen , and François Diederich
†
Laboratorium für Organische Chemie, ETH Zurich, VladimirꢀPrelogꢀWeg 3, 8093 Zurich,
Switzerland
‡
BASF SE, CarlꢀBoschꢀStrasse 38, 67056 Ludwigshafen, Germany
§
Swiss Tropical and Public Health Institute (SwissTPH), Socinstrasse 57, 4051 Basel,
Switzerland
ø
Universität Basel, Petersplatz 1, 4003 Basel, Switzerland
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 79
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
||
Medicines for Malaria Venture, Route de PréꢀBois 20, CHꢀ1215 Geneva, Switzerland
⊥
National Center for Genetic Engineering and Biotechnology, 113 Thailand Science Park,
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
Phahonyothin Road, Pathumthni 12120, Thailand
∆
Centre for Drug Candidate Optimisation, Monash Institute of Pharmaceutical Sciences, Monash
University, 381 Royal Parade, Parkville, Victoria 3052, Australia
◊
TCG Lifesciences Private Limited, Block BN, Plot 7, Saltlake Electronics Complex, Sector V,
Kolkata 700091, West Bengal, India
#
Covance Laboratories Ltd., Otley Road, Harrogate, HG3 1PY, United Kingdom
∞
AbbVie, Molecular Characterization, Dept. R4AE, 1 North Waukegan Rd., N. Chicago, IL,
6
0064ꢀ6217
≡
Department of Biochemistry and Center for Excellence in Protein and Enzyme Technology,
Faculty of Science Mahidol University, 272 Rama VI Road, Bangkok 10400, Thailand
KEYWORDS
Malaria, Plasmodium, SHMT, drug target, structureꢀbased design, inhibitor, pyrazolopyran.
ACS Paragon Plus Environment
2

Page 3 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
ABSTRACT
Targetꢀbased approaches towards new antimalarial treatments are highly valuable to prevent
resistance development. We report several series of pyrazolopyranꢀbased inhibitors targeting the
enzyme serine hydroxymethyltransferase (SHMT), designed to improve microsomal metabolic
stability and to identify suitable candidates for in vivo efficacy evaluation. The best ligands
inhibited Plasmodium falciparum (Pf) and Arabidopsis thaliana (At) SHMT in target assays and
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
PfNF54 strains in cellꢀbased assays with values in the low nanomolar range (3.2–55 nM). A set
of carboxylate derivatives demonstrated markedly improved in vitro metabolic stability (t1/2 > 2
h). A selected ligand showed significant in vivo efficacy with 73% of parasitemia reduction in a
mouse model. Five new cocrystal structures with PvSHMT were solved at 2.3–2.6 Å resolution,
revealing a unique waterꢀmediated interaction with Tyr63 at the end of the paraꢀaminobenzoate
channel. They also displayed the high degree of conformational flexibility of the Cys364–loop
lining this channel.
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 79
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
INTRODUCTION
The discovery of artemisinin, the active substance of Artemisia annua L., by Chinese researchers
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
in the soꢀcalled Project 523 was a major breakthrough in the fight against Malaria in the past
1
century. This discovery was honored by the 2015 Nobel prize in Medicine awarded to Dr.
Youyou Tu who led the pioneering research. Artemisinin derivatives are still the most widely
used treatment, mainly in combination with other antimalarials, known as artemisininꢀbased
combination therapy (ACT). Over the past few years, however, emergence of resistance against
artemisinin and ACTs has been reported in Southeast Asia, threatening current eradication
2
–4
efforts.
Although mutation of the protein P. falciparum Kelch13 (Pf Kelch13) has been
5
–7
recently identified to be a potential key player involved in the resistance mechanism, the fact
that the mode of action and the exact target of artemisininꢀbased drugs are still not clearly
8
defined makes resistance a difficult hurdle to overcome. This, combined with the fact that
resistance is described for most other antimalarials currently in use, renders the discovery of new
9
targets and new treatments crucial for complete eradication of Malaria.
In this respect, new inhibitors also containing a pharmacophoric peroxide bond have been
1
0
11
developed, such as artemisone and OZ439 (Figure 1), which is in phase IIa clinical trials.
NITD609 (Figure 1), a spiroindolone derivative, holds great promise and cleared parasitemia in
1
2,13
adults in phase II clinical trial.
Figure 1), has just finished phase II, openꢀlabel, twoꢀpart clinical study trials after showing its
potential of acting on all stages of the life cycle of the parasite, blocking transmission of
Another lead based on an imidazolopiperazine core, KAF156
(
1
4–17
parasitemia and even prophylactic activity.
Lately, novel multistage antimalarial inhibitors
18
identified by phenotypic screening and derived from diversityꢀoriented synthesis have shown
ACS Paragon Plus Environment
4

Page 5 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
19
remarkable efficacy. In addition, a selective β2 proteasome inhibitor exhibited a high curing
potential with almost complete reduction of parasite burden after a single bolus dose of
ꢀ1 20
3
5 mg kg . A targetꢀbased approach led to the discovery of DSM265 (Figure 1), an inhibitor of
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 enzyme dihydroorotate dehydrogenase (DHODH), which is currently in phase IIb clinical
2
1
trials. Finally, P218 (Figure 1), an inhibitor of the dihydrofolate reductase (DHFR) enzyme
involved in the folate cycle, emerged as a viable substitute of pyrimethamine and cycloguanil,
2
2
and is currently undergoing preclinical testing. This portfolio of potentially new therapeutics is
encouraging, however, for some of them the molecular targets remain unknown, thus prevention
or identification of resistance mechanism is highly challenging.
Figure 1. Molecular structures of novel antimalarials currently investigated in clinical trials.
O
2
H N
N
Cl
N
O
F
N
O
O
F
NH
O
N
NH
O
N
Cl
H
N
O
H
OZ439
NITD609
F
KAF156
SF
5
OH
NH
N
2
HN
O
O
O
N
F
N
N
F
N
HCl
2
H N
N
DSM265
P218
We recently demonstrated that the pyridoxal 5’ꢀphosphate (PLP) dependent enzyme serine
2
3–36
37–39
hydroxymethyltransferase (SHMT),
from the folate cycle,
which remarkably had been
neglected so far, could be targeted by pyrazolopyranꢀbased ligands to efficiently reduce
40
parasitemia in vitro (Figure 2). PvSHMT has a homodimeric structure and two active sites per
homodimer, where both active sites were found to be populated by a ligand (SI, Section S4.1,
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 79
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
40
Figure S11). Here, we report the structureꢀbased design and synthesis of an extended library of
pyrazolopyranꢀbased ligands, targeting improved metabolic stability. The in vitro metabolic
4
0
stability, which was poor in the earlier reported series of ligands, was improved substantially
with the current ligands with in vitro microsomal halfꢀlives exceeding 2 h. At the same time, onꢀ
target activity on Plasmodium falciparum (Pf) SHMT and Arabidopsis thaliana (At) SHMT
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
(
used to refine the SAR) as well as in vitro activity against the PfNF54 sensitive strain were
maintained in the low nanomolar IC (half maximal inhibitory constant) and EC (half maximal
5
0
50
effective concentration) ranges, respectively. No apparent cytotoxicity of the improved leads was
observed. We show for the first time that SHMT inhibitors can be effective in vivo. One of the
most promising candidates has been advanced to an in vivo mouse model. Gratifyingly, a
significant reduction of parasitemia of 73% was measured compared to untreated mice at 4 × 50
mg/kg (p.o.). Additionally, five new cocrystal structures were solved, confirming the binding
mode of the pyrazolopyranꢀbased ligands. A detailed investigation of the molecular recognition
properties at the active site of PvSHMT was undertaken, revealing an interesting waterꢀmediated
interaction with Tyr63 at the end of the channel, where the paraꢀaminobenzoate side chain of the
substrate tetrahydrofolate (THF) binds. The particular conformational flexibility of the Cys364–
loop lining this channel was analyzed as well. It should be noted that there has also been recent
4
1–43
interest in human SHMT as a potential new cancer target.
ACS Paragon Plus Environment
6

Page 7 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Figure 2. a) The folate cycle. b) Molecular structure of tetrahydrofolate (THF). c) Schematic
depiction of the binding mode of the initial lead (+)ꢀ1 as seen in a cocrystal structure analysis
4
0
(
PDB ID: 4TMR). Intermolecular Hꢀbonding are indicated by red dashed lines.
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
Serine hydroxymethyltransferase (SHMT) converts tetrahydrofolate (THF) to 5,10ꢀ
methylenetetrahydrofolate (5,10ꢀCH ꢀTHF), a crucial component for the conversion of
deoxyuridine monophosphate (dUMP) to the DNA precursor deoxythymidine monophosphate
dTMP) by thymidylate synthase (TS). Dihydrofolate (DHF) formed by this process is converted
back to THF by dihydrofolate reductase (DHFR). In Plasmodium, components of the cycle are
replenished by de novo synthesis: dihydropteroate synthase (DHPS) condenses pteridine
diphosphate and paraꢀaminobenzoic acid (pABA) to dihydropteroate (DHP), which is converted
to DHF by dihydrofolate synthase (DHFS).
2
(
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 79
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
RESULTS AND DISCUSSION
Ligand Design
The first generation ligands, such as (+)ꢀ1, showed high activity in cellꢀbased assays against
PfNF54 and various drugꢀresistant mutant strains, as well as very low cytotoxicity. Their
metabolic stability however was insufficient, and exposure was unlikely to be sufficient to
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
4
0
achieve in vivo efficacy. Improving metabolic stability was therefore necessary to advance this
program, and structural changes of the ligands was guided by structureꢀbased design based on
the two cocrystal structures that had been obtained. Figure 2c schematically shows the binding
mode of (+)ꢀ1 at the active site of PvSHMT (PDB ID: 4TMR). Two subꢀpockets are
distinguishable. The pteridine pocket, which binds the pyrazolopyran core and also the pteridine
core of THF by several Hꢀbonds to the protein (PDB ID: 4OYT), whereas an extended channel
hosts the biaryl motif of (+)ꢀ1 and the pABA side chain of THF. In the new ligand series, the
pyrazolopyran core was kept intact, whereas its alkyl substituents and in particular the biaryl
moiety attached to the stereogenic center were systematically altered based on computational
4
4
predictions using the modeling software MOLOC. These ligand alterations are reported in the
following.
Synthesis and Cell-based Activity of Thiophene Derivatives (±)-2–27
Methyl ester (±)ꢀ1, identified in a primary screening of ligands and initially developed as
45
potential herbicide targeting AtSHMT, was found to have a poor metabolic stability when
incubated with human liver microsomes (t = 3 min). As there was evidence of ester cleavage in
1/2
(±)ꢀ1, we first focused on the replacement of this moiety aiming at maintaining EC ꢀvalues
50
40
against PfNF54 strain in the single digit nanomolar range. The ligands (±)ꢀ2–8 (Scheme 1)
ACS Paragon Plus Environment
8

Page 9 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
were synthesized in 5–6 steps starting from 9/10. The key steps are a SuzukiꢀMiyaura crossꢀ
4
6
coupling to 11a–g, a Knoevenagel condensation to 12a–g, and a Michael addition of 3ꢀmethylꢀ
ꢀpyrazolinꢀ1ꢀone, followed by an intramolecular cyclization to provide (±)ꢀ2–8 (for more
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
details, see the Supporting Information Section S1).
a
Scheme 1. Representative Synthesis of Pyrazolopyran Ligands (±)ꢀ2–8.
a
Reagents and conditions: (a) bis(pinacolato)diboron, KOAc, [PdCl (dppf)⋅CH Cl ] (2.0 mol %),
2
2
2
toluene,
ꢀ
w, 150 °C, 6 min; (b) thienyl boronic acid or bromothiophene analogue, [PdCl (PPh ) ]
2
3 2
(5.0 mol %), Na CO , THF/H O 4:1, 60 °C, 4–16 h; (c) malononitrile, TiCl , pyridine, CHCl ,
2 3 2 4 3
6
3 °C, 48–72 h; (d) 3ꢀmethylꢀ2ꢀpyrazolinꢀ1ꢀone, piperidine, 1,4ꢀdioxane/EtOH 1:1, ꢀw, 65 °C,
3.5 h.
Ligands (±)ꢀ2, 3, and 5 displayed moderate affinities for PfSHMT, whereas oxadiazole (±)ꢀ6
and sulfonamide (±)ꢀ7 analogues gave better inhibition with IC values in the low three digit
5
0
nanomolar range (Table 1). Higher affinity was achieved with the methoxymethyl ether (±)ꢀ4
(IC = 42 n
M
). Based on these results, the benzothiophene analogue (±)ꢀ8 was designed with
5
0
47,48
the intention of maintaining the S···S interaction
to Cys364 and improving the hydrophobic
interactions with the surrounding residues as Tyr63 and Phe266 (SI, Section S3, Figure S9a).
The predicted interactions nicely led to an improved IC value of 60 n . Cocrystal structures of
M
5
0
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 79
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
35
(
±)ꢀ4 and (±)ꢀ7 with PvSHMT, which features a virtually identical active site to PfSHMT, were
obtained. The overall binding mode seen for complexed (+)ꢀ1 was confirmed. However, no good
explanation became apparent for the higher target affinity of (±)ꢀ4 as compared, for example, to
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
(
±)ꢀ3 (SI, Section S4.3, Figure S15 and S16). The methoxymethyl ether moiety is located at the
exit of the pABA channel at the periphery of the protein, it is turned out of the plan compared to
the ester group in (+)ꢀ1 (C–C–C–O torsion angle = 69°) and does not directly interact with any
protein residue. Regarding (±)ꢀ7, its sulfonamide moiety is found in the preferred staggered
4
9
conformation, with the Nꢀlone pair bisecting the O=S=O fragment and at a 3.7 Å distance of
Tyr63, is too far away to establish a strong polar interaction (SI, Section S4.3, Figure S17). On
the other hand, one of the Nꢀmethyl groups is at 3.1 Å to the edge of Phe134 leading to a short
C–H···π contact and two short O···H–C contacts are established between one S=O bond and
Val141 and Pro267 (SI, Section S4.3, Figure S18).
Table 1. Biological Activities of Thiophene Derivatives (±)-2–8.
EC Pf
IC Pf SHMT
50
5
0
Num.
R
a
NF54 [n
M
]
± SD [n
M
]
S
S
3
F C
(
(
±)ꢀ2
±)ꢀ3
594
388 ± 5
O
410
228
501
323 ± 43
42 ± 1
O
S
S
(±)ꢀ4
Cl
N
(
(
±)ꢀ5
±)ꢀ6
202 ± 9
S
S
N
212
151
197
133 ± 6
99 ± 2
60 ± 1
O
O
N
O
S
(
(
±)ꢀ7
±)ꢀ8
S
a
Standard deviations are given.
ACS Paragon Plus Environment
10

Page 11 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Overall there was a good correlation between the targetꢀbased and the PfNF54 cellꢀbased
assay. In this assay, the best inhibitor of the series was sulfonamide (±)ꢀ7 with an EC value of
50
1
51 n
Since the cellꢀbased affinity of this series of thiophene derivatives was still two orders of
magnitude lower than that of (±)ꢀ1 (EC value = 3 n ), modifications at other positions of the
molecule were evaluated to improve biological activity. Table 2 gives the EC values for the
M (Table 1).
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
M
5
0
5
0
cellꢀbased assay which vary by more than a factor of 600. The IC values in the PfSHMT target
5
0
assay, however, were all very similar, in the tripleꢀdigit nanomolar range, possibly due to a
limited dynamic range of the assay in particular at higher binding affinity. All IC values for
5
0
PfSHMT are given in Table S1 in the SI, together with the IC values of selected compounds
50
measured against AtSHMT, which correlate much better with the EC50 data. Arabidopsis
thaliana SHMT, the plant counterpart and PfSHMT are highly similar with 45% of sequence
identity, and only very few amino acid residues differ within each active site (SI, Section S4.2,
4
5
Figures S12, S13 and S14). Many of the SARs established in the following from the cellꢀbased
assay data are nicely supported by the AtSHMT target assay data.
For comparison, ligand (±)ꢀ5 (Table 1) was chosen as reference compound. We first
investigated changes of the meta substituent on the biaryl phenyl ring. Substitution of the nitrile
in (±)ꢀ5 by a chloride ((±)ꢀ13) or bromide ((±)ꢀ14) did not have significant effect (Table 2).
Introduction of a CF group ((±)ꢀ15) resulted in an improved in vitro potency by a factor of 2
3
(
EC = 226 nM).
50
ACS Paragon Plus Environment
11

Journal of Medicinal Chemistry
Page 12 of 79
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
3
4
Table 2. Structure-Activity Relationships for Positions R , R , and R .
R2
R1
R3
R4
EC Pf NF54
50
Num.
CN
NH
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
*
[
n ]
M
HN
N
O
2
1
2
3
4
R
R
R
R
S
S
S
S
S
S
S
Cl
Cl
Cl
Cl
Cl
Cl
Cl
(
±)ꢀ13
±)ꢀ14
Cl
Br
Me
Me
Me
Me
Me
Me
Me
485
892
226
(
(±)ꢀ15
(±)ꢀ16
CF3
OCF3
3760
7106
4511
5501
(
(
(
±)ꢀ17
±)ꢀ18
±)ꢀ19
CF
3
O
O
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
(
±)ꢀ20
CF3
CF3
CF3
CF3
CF3
CF3
CF3
CF3
Me
CD3
Et
18
11
N
O
N
O
O
O
O
O
O
O
(±)ꢀ21
O
N
(
(
(
±)ꢀ22
±)ꢀ23
±)ꢀ24
172
576
29
O
N
cyPr
H
O
N
O
N
(±)ꢀ25
Me
Me
Me
340
1202
498
O
N
(
(
±)ꢀ26
±)ꢀ27
O
N
With larger substituents, the potency was significantly reduced and EC values in the singleꢀ
5
0
digit micromolar range were measured ((±)ꢀ16–19). The larger groups interfere with the
ACS Paragon Plus Environment
12

Page 13 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
conformationally flexible Cys364–loop, which was only discovered later and is discussed below
based on new cocrystal structures.
Gratifyingly, by applying the CF ꢀsubstitution to the sulfonamide ligand (±)ꢀ7, a low n
M
cellꢀ
3
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
based activity could be reached, giving more than a 10ꢀfold boost in potency (EC (±)ꢀ20 =
5
0
3
4
18 n
core in the sulfonamide ligand series, preparing ligands (±)ꢀ21–27. In position 3, the methyl
substituent was optimal, and its potency only rivaled by a CD ꢀ (EC (±)ꢀ21 = 11 n ) or a Hꢀ
). Ethyl and cyclopropyl substituents in this position were
M). Encouraged by this result, we investigated positions R and R on the pyrazolopyran
M
3
50
substituent (EC50 (±)ꢀ24 = 29 n
M
detrimental to activity.
4
Study of the substituent R at the stereogenic center revealed that this position was not
amenable to broad changes. Substitution of the isopropyl in (±)ꢀ20 (EC = 18 n
M
) by ethyl,
5
0
cyclopropyl, and cyclobutyl ((±)ꢀ25–27) strongly reduced the cellꢀbased activity (Table 2). The
3
4
preference for R = Me and R = CH(CH ) undoubtedly are targetꢀbinding related as could be
3 2
validated by the set of 7 cocrystal structures obtained so far in the project. Although (±)ꢀ20 had
promising antiparasitic activity, it was not further pursued due to its poor liver microsomal
stability (t1/2 = 3 min; SI, Section S2.2, Table S2). Instead, identification of potential metabolites
was performed in order to guide the next round of optimization.
Identification of Metabolites from (±)-20
Metabolite identification performed on (±)ꢀ20 in NADPHꢀsupplemented human liver
microsomes revealed major degradation pathways via thiophene oxidation and Nꢀdemethylation
accounting for approximately 50 and 33% of total metabolites, respectively (SI, Section S2.4,
Tables S5 and S6). The results were confirmed by an isotopic labeling experiment using the
ACS Paragon Plus Environment
13

Journal of Medicinal Chemistry
Page 14 of 79
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
deuterated analogue (±)ꢀ21, enabling exclusion of potential metabolism on the methylꢀpyrazole
4
0
moiety (SI, Section S2.4, Tables S7 and S8). A previously reported ligand (±)ꢀ28, lacking the
thiophene moiety and featuring a 3,5ꢀdichlorophenyl substituent at the stereogenic center (for the
structure, see SI, Figure S5), was studied as well. For this compound, the main metabolite arises
from oxidation on the pyrazolopyran core, possibly by oxidation of the vinylogous cyanamide
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
(SI, Section S2.4, Figure S5 and Tables S9 and S10). These results suggested that metabolism is
compoundꢀspecific and that the vinylogous cyanamide moiety might be a liability. However, we
focused during the further optimization on the replacement of the substitutedꢀthiophene
peripheral part to prepare more stable ligands.
Heteroalicyclic Rings as Thiophene Replacements
In an effort to identify suitable thiophene replacements with the goal to achieve high in vitro
potency, a series of 13 inhibitors featuring Nꢀheteroalicycles of various sizes and different nature
were synthesized. The terminal rings in (±)ꢀ29–37 were selected to both maintain hydrophobic
interactions with the surrounding residues in the pABA channel and reduce the overall
lipophilicity of the molecule. Ligands (±)ꢀ29–37 were prepared by employing a Buchwaldꢀ
5
0
Hartwig crossꢀcoupling between the building block 38 and the corresponding secondary amines
to give 39a–i, followed by the classical twoꢀsteps sequence via 40a–i leading to the final
pyrazolopyran analogues (Scheme 2).
ACS Paragon Plus Environment
14

Page 15 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Scheme 2. Synthesis of Ligands (±)-29–37 Employing a Buchwald-Hartwig Cross-
a
Coupling.
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
Reagents and conditions: (a) corresponding secondary amine, Cs CO , [Pd (dba) ] (2.0 mol %),
2
3
2
3
XꢀPhos (8.0 mol %), 1,4ꢀdioxane, 110 °C, 16 h; (b) malononitrile, TiCl , pyridine, CHCl , 63 °C,
4
4
3
8–72 h; (c) 3ꢀmethylꢀ2ꢀpyrazolinꢀ1ꢀone, piperidine, 1,4ꢀdioxane/EtOH 1:1,
ꢀw, 65 °C, 3.5 h.
As the Nꢀheterocyclic derivatives (±)ꢀ29–37 lack functional groups able to interact with
4
7
Cys364, sultam (±)ꢀ41 was designed to establish chalcogen interactions (S···O) with this
residue. The synthesis started with the conversion of 38 to Bocꢀprotected 42 and aniline 43.
Subsequent treatment with 4ꢀchloroꢀbutaneꢀ1ꢀsulfonyl chloride afforded the sultam precursor 44,
which underwent Knoevenagel condensation to 45 and core formation, yielding (±)ꢀ41 (Scheme
3
).
ACS Paragon Plus Environment
15

Journal of Medicinal Chemistry
Page 16 of 79
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
Scheme 3. Synthesis of Sultam (±)-41.
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
Reagents and conditions: (a) tertꢀbutyl carbamate, Cs CO , [Pd (dba) ] (2.0 mol %), XꢀPhos
8.0 mol %), 1,4ꢀdioxane, 110 °C, 16 h, 40%; (b) TFA, CH Cl , 25 °C, 3 h, 71%; (c) 4ꢀchloroꢀ
2
3
2
3
(
2 2
butaneꢀ1ꢀsulfonyl chloride, Et N, CH Cl , 25 °C, 4 days, 47%; (d) malononitrile, TiCl , pyridine,
3
2
2
4
CHCl , 63 °C, 48 h, 66%; (e) 3ꢀmethylꢀ2ꢀpyrazolinꢀ1ꢀone, piperidine, 1,4ꢀdioxane/EtOH 1:1,
3
ꢀw, 65 °C, 3.5 h, 39%.
Another sulfoneꢀcontaining inhibitor (±)ꢀ46 with a thiadiazinane scaffold as a more polar
derivative of (±)ꢀ41 was also synthesized. The sulfamoyl carbamate was introduced by reacting
43 with the isocyanate 47 to give 48, which was transformed into 49 by a diꢀNꢀalkylation of 48
with 1,3ꢀdibromopropane/K CO (Scheme 4). Bocꢀcleavage followed by Nꢀmethylation gave 50,
2
3
which was transformed via 51 into (±)ꢀ46. Nꢀmethylpyrimidone analogue (±)ꢀ52 was also
prepared with the intention to establish a contact with Cys364, as well as increasing the solubility
of the inhibitor. It was synthesized following the synthetic route shown in Scheme 1.
ACS Paragon Plus Environment
16

Page 17 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Scheme 4. Synthesis of Thiadiazinane (±)-46.
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
Reagents and conditions: (a) tertꢀBuOH, CH Cl , 0 °C, 30 min, then 43, Et N, CH Cl , 25 °C,
2
2
3
2
2
16 h, 80%; (b) 1,3ꢀdibromopropane, K CO , CH CN, reflux, 16 h, 55%; (c) TFA, CH Cl , 25 °C,
2
3
3
2
2
3
6
3
h, 79%; (d) MeI, K CO , DMF, 25 °C, 3 h, 92%; (e) malononitrile, TiCl , pyridine, CHCl ,
2
3
4
3
3 °C, 48 h, 55%; (f) 3ꢀmethylꢀ2ꢀpyrazolinꢀ1ꢀone, piperidine, 1,4ꢀdioxane/EtOH 1:1,
.5 h, 29%.
ꢀw, 65 °C,
Ligand (±)ꢀ53 was designed to interact with Tyr63 via its sulfonamide moiety. Its synthesis
started from the commercially available Bocꢀprotected iodoazetidine 54, which was transformed
51
into organoꢀzinc derivative 55 in situ, prior to a Negishi crossꢀcoupling reaction reaction with
38 (Scheme 5) to give 56. The desired ligand (±)ꢀ53 was subsequently obtained via sulfonamides
5
7 and 58.
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 79
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
Scheme 5. Synthesis of the (Methanesulfonyl)azetidine (±)-53.
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
Reagents and conditions: (a) Zn, 1,2ꢀdibromoethane, DMA, 70 °C, 30 min, then
chlorotrimethylsilane, DMA, 25 °C, 2 h, and 50 °C, 2 h; (b) 38, [Pd (dba) ] (2.0 mol %), tri(2ꢀ
furyl)phosphine (4.0 mol %), DMA, 70 °C, 20 h, 56%; (c) conc. HCl, MeOH, 25 °C, 3 h, then
methanesulfonyl chloride, CH Cl , 25 °C, 12 h, 35% (over 2 steps); (d) malononitrile, TiCl ,
pyridine, CHCl , 63 °C, 48 h, 59%; (e) 3ꢀmethylꢀ2ꢀpyrazolinꢀ1ꢀone, piperidine, 1,4ꢀ
2
3
2
2
4
3
dioxane/EtOH 1:1, ꢀw, 65 °C, 3.5 h, 30%.
The inhibitors of this series had EC values in the cellular assay in the 25–610 nanomolar
50
range (Table 3). The measured IC values against PfSHMT varied between 84–398 n
M
, and
5
0
again the IC values obtained for selected inhibitors against AtSHMT correlated much better
5
0
with the data from the cellꢀbased assay (SI, Section S2.1, Table S1). The highest cellular potency
was measured for the fused Nꢀheterocyclic derivatives (±)ꢀ30–35, suggesting the establishment
of good hydrophobic interactions with the apolar surface at the exit of the pABA channel,
formed by Leu124, Leu130, Phe134, and Val141 within the active site as illustrated for the
difluoroazetidine derivative (±)ꢀ30 in Figure S9b (see SI, Section S3). In the AtSHMT assay,
compound (±)ꢀ33 was the most potent with an IC value of 17.7 nM.
50
Attempts to gain affinity by interacting with Cys364 apparently remained unsuccessful ((±)ꢀ41,
(
±)ꢀ46, and (±)ꢀ52) according to both EC and IC values (SI, Section S2.1, Table S1). Sulfone
50 50
ACS Paragon Plus Environment
18

Page 19 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(
±)ꢀ37 and sulfonamide (±)ꢀ53 gave both promising EC (65 nM) and IC values (PfSHMT:
5
0
50
1
27 and 84 nM; AtSHMT: 43 and 31 nM) and we took this as a lead for establishing Hꢀbonding
interactions with Tyr63, which had been suggested by the modeling (see the carboxylic acid
ligand series below).
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. Biological Activities of Inhibitors (±)-29–37, (±)-41, (±)-46, and (±)-52,53.
EC Pf
5
0
Num.
R
NF54 [n
M
]
(
±)ꢀ29
±)ꢀ30
350
(
44
56
59
(±)ꢀ31
(
(
(
(
±)ꢀ32
±)ꢀ33
±)ꢀ34
±)ꢀ35
25
49
26
(
(
±)ꢀ36
±)ꢀ37
193
65
(
(
±)ꢀ41
±)ꢀ46
454
523
O
N
N
(
(
±)ꢀ52
±)ꢀ53
610
65
O
O
S
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 79
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
Aromatic N-Heterocyclic Ligands
In parallel to the previous series of inhibitors, and in a permanent effort towards reaching singleꢀ
digit nanomolar activity, we investigated a larger set of aromatic Nꢀheterocyclic derivatives ((±)ꢀ
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
46
59–71). The scaffolds were synthesized by a SuzukiꢀMiyaura crossꢀcoupling between building
block 38 (or its borylated analogue) and the respective aromatic Nꢀheterocycles as shown in
Scheme 1. The cellꢀbased activities for selected ligands are summarized in Table 4. The
complete biological characterization for all aromatic Nꢀheterocyclic ligands is provided in the
Supporting Information (Section S2.1, Table S1). The in vitro activities of many of the prepared
compounds gave values in the low nanomolar range. Gratifyingly, five of the prepared pyridine
derivatives gave the desired EC values in the singleꢀdigit nanomolar range ((±)ꢀ60,61 and (±)ꢀ
50
67–69) (Table 4) and similar activity was also measured in the AtSHMT assay (SI, Section S2.1,
Table S1).
Table 4. Biological Activities of Selected N-Heterocyclic Ligands (±)-59–71.
EC Pf NF54
50
Num.
R
[nM]
(
±)ꢀ59
±)ꢀ60
±)ꢀ61
115
4.8
3.2
10
(
(
(±)ꢀ62
(±)ꢀ63
27
125
18
(
(
(
±)ꢀ64
±)ꢀ65
±)ꢀ66
52
(±)ꢀ67
4.2
ACS Paragon Plus Environment
20

Page 21 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(
±)ꢀ68
5.1
6.4
(±)ꢀ69
(±)ꢀ70
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
14
66
(
±)ꢀ71
No apparent cytotoxicity for (±)ꢀ60 and (±)ꢀ61 was measured on rat myoblast cell line (L6)
and human HepG2 cell line, with IC values only in the high micromolar range (SI, Sections
50
S2.7 and S2.8, Tables S23 and S24). hERG inhibition was also assessed for (±)ꢀ60, resulting in
an IC of 29 ꢀM which highlights once again the safeness of this series (SI, Section S2.8, Table
5
0
S24). The potential of these ligands however was greatly diminished by their very short halfꢀlife
when incubated with human liver microsomes (< 10 min) (SI, Section S2.2, Table S2). Despite
this liability, the high antiparasitic efficacy of (±)ꢀ61 prompted us to further study its
antiplasmodial activity in a P. falciparum SCID (severe combined immunodeficiency) mouse
52
model at 4 × 50 mg/kg (p.o.). The rather low 25% parasitemia reduction is possibly the result
of poor exposure due to high metabolic instability of (±)ꢀ61 (SI, Section S2.9, Figure S7, Tables
S25 and S26).
Carboxylic Acid Series
4
0
While measuring the metabolic stability of selected ligands previously reported, carboxylic
acid derivative (±)ꢀ72 drew our attention as it was reasonably stable in the microsomal test
system (t1/2 = 137 min), but had only moderate potency (EC Pf NF54 = 340 nM) (Table 5).
50
Subsequently, we put our effort onto improving the antiparasitic efficacy of this carboxylic acid
analogue, while maintaining a high microsomal stability, using the knowledge learned from the
ACS Paragon Plus Environment
21

Journal of Medicinal Chemistry
Page 22 of 79
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
previous optimization cycles by replacing the metaꢀCN on the phenyl ring at the stereogenic
center with a metaꢀCF group or inserting an Nꢀheterocycle for instance.
3
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 5. Biological Activities, logD7.4 Values, and Microsomal Half-Lives of Carboxylate-
Based Ligands (±)-72–79.
R
IC Pf
*
CN
NH
50
EC Pf
IC At
SHMT [n
b
c
50
50
Num.
HN
SHMT ± SD
logD7.4
t1/2 [min]
137
41
NF54 [n
M
]
a
M
]
N
O
2
[nM]
R
(
±)ꢀ72
±)ꢀ73
±)ꢀ74
±)ꢀ75
340
220 ± 70
189 ± 7
119 ± 6
77 ± 1
0.088
1.4
(
(
(
54
19.1
1.8
d
632
822
n.d.
n.d.
1.7
64
n.d.
59
(
(
(
±)ꢀ76
±)ꢀ77
±)ꢀ78
172
116
194
268 ± 15
312 ± 10
56 ± 1
n.d.
41.9
27.8
n.d.
1.6
1.5
70
55
157
ACS Paragon Plus Environment
22

Page 23 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(±)ꢀ79
154
75 ± 1
26.2
1.2
226
a
b
c
Standard deviations are given. Measured by HPLC, for the method see the SI, Section S2.2. In
vitro metabolic stability measured in human liver microsomes. n.d. = not determined.
d
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 prepared a series of metaꢀCF derivatives (±)ꢀ73–77 (Table 5) and as expected, substitution
3
of the cyano moiety in (±)ꢀ72 by a CF group (EC (±)ꢀ73 = 54 nM) helped to improve cellꢀ
3
50
based potency, though at the expense of its metabolic stability (Table 5). All other prepared
derivatives, with either a terminal aromatic ring or an Nꢀlinked piperidine, gave mixed results,
when considering all parameters including target affinity, cellꢀbased affinity and microsomal
halfꢀlife. Also the piperidine derivatives showed good solubility (for solubility measurements of
selected ligands, see the SI, Section S2.3, Table S3). Examining the logD7.4 value of (±)ꢀ72 and
(±)ꢀ73 (1.4 and 1.8, respectively, Table 5), we reasoned that a compound having lipophilicity in
this range would be likely to lead to a good balance between potency and stability. Indeed, the
metaꢀfluoro derivative (±)ꢀ78 has a moderate activity on PfNF54 (EC = 194 n
M
), but to our
delight, it surpassed all ligands studied so far with a microsomal halfꢀlife of 157 min. Ligand (±)ꢀ
8 was also incubated in rat liver microsomes resulting in a comparable halfꢀlife of 95 min (SI,
5
0
7
Section S2.2, Table S2). This step forward in the pursuit of metabolically stable inhibitor was
further confirmed in an in vivo rat model, in which an apparent in vivo halfꢀlife of 3 h was
measured (SI, Section S2.5). The difference between in vitro and in vivo halfꢀlives can possibly
be attributed to plasma protein or microsome binding. Noticeably, applying the CF →F
3
replacement strategy to piperidine derivative (±)ꢀ77 led to a slight potency loss, which was
compensated by a large stability gain ((±)ꢀ79). In addition (±)ꢀ78 and (±)ꢀ79 had high target
affinity in both PfSHMT and AtSHMT assays.
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 79
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
We subsequently prepared ligands (±)ꢀ80–87 with carboxylate side chains of larger length
(
Table 6). The cocrystal structure of the natural substrate THF bound to SHMT (PDB ID: 4OYT,
2
.4 Å) shows the two carboxylates of its terminal glutamate located at the exit of the pABA
channel at the periphery of the binding pocket. Overlaying this previously reported cocrystal
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
3
6
structure of THF with PvSHMT and the modeled carboxylic acid derivative (+)ꢀ85 clearly
suggested that extension of the carboxylate chain might allow to better mimic the THF side
chain, thus eventually improving the potency (SI, Section S3, Figure S10).
Table 6 indeed shows that extension of the carboxylate side chain has a beneficial effect on
cellular potency in both CF ꢀ and Fꢀseries (e.g.: EC values (±)ꢀ77 (116 n
M
) > (±)ꢀ82 (84 n
M) >
3
50
(
±)ꢀ83 (17 n
M
); or (±)ꢀ79 (154 n
M
) > (±)ꢀ86 (55 n
M
) > (±)ꢀ87 (17 n
M
)). Activity gain came
along with a loss in metabolic stability, which became unacceptable in the CF series.
3
Table 6. Biological Activities and Microsomal Half-Lives of Ligands (±)-80–87 with
Extended Carboxylate Side Chains.
R
IC50
PfSHMT ±
IC50
AtSHMT
[nM]
*
CN
NH
EC50
PfNF54
b
Num.
HN
t1/2 [min]
a
N
O
2
[nM
]
SD [nM
]
R
O
CF
CF
3
HO
(
±)ꢀ80
22
186 ± 9
59.0
54
O
3
2
HO
(
(
±)ꢀ81
±)ꢀ82
10
84
165 ± 7
186 ± 7
27.0
13.5
31
36
O
CF
3
HO
N
ACS Paragon Plus Environment
24

Page 25 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
O
CF
3
HO
N
(±)ꢀ83
17
52
168 ± 11
172 ± 10
18.2
22.4
12
2
O
F
HO
(
(
±)ꢀ84
±)ꢀ85
159
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
O
F
2
HO
19
90 ± 4
33.6
95
O
F
F
HO
N
(±)ꢀ86
55
17
97 ± 1
37.1
13.8
194
50
O
HO
N
(
±)ꢀ87
130 ± 4
2
a
b
Standard deviations are given. In vitro metabolic stability measured in human liver
microsomes.
53
As acidic ligands tend to have a high plasma protein binding, this parameter was measured
via ultracentrifugation in mouse plasma for the highly potent ligand (±)ꢀ81 and the most stable
analogue (±)ꢀ86. Significant differences were observed as (±)ꢀ81 was highly bound to plasma
proteins (99.2% bound), whereas (±)ꢀ86 had much lower binding (82.9% bound) (SI, Section
S2.10, Table S29). Taking all parameters into account, (±)ꢀ86 was clearly the compound with the
most attractive profile and the best balance between potency and metabolic stability.
Encouraged by these results we investigated the pure enantiomers of (±)ꢀ86. For this purpose,
the enantiomers of the direct precursor to (±)ꢀ86 were separated by chiral HPLC and subsequent
hydrogenation afforded pure (+)ꢀ86 and (–)ꢀ86 (see Scheme 6 in the Experimental Section and
HPLC traces in the SI, Section 1.10). As observed previously with enantiomerically pure
4
0
pyrazolopyranꢀbased ligands, (+)ꢀ86 is considerably more potent than (–)ꢀ86 in all in vitro
assays (Table 7). This large discrepancy can be easily rationalized by the preferred binding mode
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 79
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
of pyrazolopyranꢀbased ligands; as for all analogues cocrystallized with PvSHMT, exclusively
the (+)ꢀenantiomers where found to be bound to the enzyme as detailed below. Additionally, (+)ꢀ
86 is roughly twoꢀfold more potent than the racemic mixture. Remarkably, (+)ꢀ86 also has an
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
improved metabolic stability with a in vitro halfꢀlife greater than 255 min.
Table 7. Biological Activities and Microsomal Half-Lives of Enantiopure Ligands (+)-86
and (–)-86.
b
EC PfNF54
IC50 PfSHMT
IC AtSHMT
t1/2 [min]
50
50
Num.
a
[
nM
]
± SD [n
M
]
[nM]
(±)ꢀ86
(+)ꢀ86
(–)ꢀ86
55
35
6143
97 ± 1
110 ± 2
1581 ± 49
37.1
12.6
977
194
>255
206
c
a
b
Standard deviations are given. In vitro metabolic stability measured in human liver
microsomes. This compound showed minimal degradation (<15%) over the course of the
incubation.
c
Cytotoxicity and Mutagenicity Evaluation
Cytotoxicity of selected compounds from each series studied above was determined against rat
myoblast L6ꢀcells in a 72 h cytotoxicity assay (SI, Section S2.7, Table S23). All ligands showed
excellent selectivity indexes, with IC values ranging from 9.9 to >200
ꢀM
. Also, (±)ꢀ60 was
5
0
tested in the human HepG2 cell line resulting in an IC of 18.9 ꢀM (SI, Section S2.8, Table
2
0
S24).
54
Mutagenicity was addressed via an AMES assay employing (±)ꢀ78 as reference compound,
with and without metabolic activation. Out of the 5 strains examined, reproducible evidence of
mutagenicity (≥2ꢀfold increase in revertant colony numbers over concurrent vehicle controls)
was observed only in strain TA98 without metabolic activation at 5000 ꢀg/plate in Experiment 1
ACS Paragon Plus Environment
26

Page 27 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
and at 1250
ꢀg/plate and above in Experiment 2, using the plate incorporation method (SI,
Section S2.6).
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
Animal Model
The promising properties of ligand (+)ꢀ86 prompted us to study its in vivo efficacy in a P.
52
falciparum SCID mouse model at 4 × 50 mg/kg (p.o.). Administration of (+)ꢀ86 on days 3, 4, 5,
and 6 after infection led to a significant reduction of parasitemia of 73% (vs. untreated mice) on
day 7 after infection (SI, Section S2.9, Figure S8, Tables S27 and S28). Albeit complete
clearance was not reached, this positive result provides a true proof of concept and validates
SHMT a viable in vivo antimalarial target. Encouraged by this outcome, we plan to further
develop this series of inhibitors in order to completely reduce parasitemia in vivo. Complete
determination of the pharmacokinetic profile and measurement of the in vivo exposure will give
valuable insight for improving efficacy.
Crystal Structure Determination of PvSHMT with (±)-4, (±)-7, (±)-80, (±)-81, and (±)-85.
PvSHMT was crystallized from the ternary complex of PvSHMT, glycine, and either (±)ꢀ4,
(±)ꢀ7, (±)ꢀ80, (±)ꢀ81, or (±)ꢀ85 using the microbatch method. The cocrystals diffracted to 2.6,
2.3, 2.3, 2.3, and 2.3 Å resolution, respectively, and belong to the C2 space group. The structures
were solved by molecular replacement using the coordinates of a chain A protomer of PvSHMT
36
(PDB code 4OYT) as the template. Despite using a racemic mixture of ligands for
cocrystallization, only the (+)ꢀenantiomer was present in the structure.
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 79
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
Figure 3. Cocrystal structure (PDB ID: 5GVN, 2.3 Å resolution) showing the proteinꢀligand
interactions of pyrazolopyran (+)ꢀ85 (lime) and PvSHMT (grey). a) Occupancy of the active site
by (+)ꢀ85, the waterꢀmediated interaction with Tyr63, and the cofactor (pyridoxal 5’ꢀphosphate,
PLP). The mesh spans the volume of the binding pocket. b) Polar interactions between (+)ꢀ85
and the protein; PLP is omitted for clarity. The water molecules (W1, W2, and W3) are
represented as red spheres. Distances are given in Å. Color code: Cprotein grey, CLigand lime, CPLP
yellow, F light cyan, N blue, O red, P orange, S yellow.
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
W2
a)
b)
3
.2
2
3
.0
2
.8
Tyr63
.7
W1
3.0
W3
W2
Cys364
3.2
2.7
3
2.8
.0
Tyr63
W1
W1
Asn356
3.5
Glu56
3.0
Asp136
Lys138
Thr357
2
.9
W3
3
.1
Glu56
3.7
Cys125
Lys139
2.9
2.8
3.0
Leu124
Thr183
PLP
Gly128
Binding Mode of (+)-85
Ligand (+)ꢀ85 nicely occupies the substrateꢀbinding site of PvSHMT (Figure 3). As seen in all
40
cocrystal structures reported herein and previously, both active sites from the PvSHMT
homodimer are occupied by (+)ꢀ85. From the cocrystal structure, it can be derived that (+)ꢀ85 (as
well as the other pyrazolopyran ligands) is (R)ꢀconfigured at the stereogenic center. The overall
binding mode matches well the one from a NꢀformylꢀTHF derivative (Figure 4) in PvSHMT
ACS Paragon Plus Environment
28

Page 29 of 79
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
36
(
PDB Code: 4OYT, 2.4 Å). The cysteine bridge formed by Cys125 and Cys364 is found in its
reduced form. The pyrazolopyran core is deeply buried into the pteridine pocket and is in close
proximity to the cofactor PLP (Figure 3a). The core is anchored into the active site via a strong
Hꢀbond network, establishing an array of polar interactions similar to the ones seen in the two
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
4
0
previously obtained cocrystal structures (Figure 3b). The N(2) atom of the pyrazole ring is
protonated and forms a short ionic Hꢀbond with the side chain of Glu56. The protonation of N(2)
is also seen in the unbound ligand state, as evidenced by the three smallꢀmolecule Xꢀray crystal
structures recorded for (±)ꢀ4, (±)ꢀ70, and (±)ꢀ88 (SI, Section S4.4, the structure of (±)ꢀ88 is
shown there). N(2)ꢀprotonation is supported by both the bond length in the pyrazole ring as well
as by Hꢀbonding interactions between neighboring molecules in the crystal lattice (SI, Section
4.4.2, Figure S28).
The biaryl moiety is located in the pABA channel and directs the carboxylate side chain
towards the periphery of the protein where Tyr63 adopts a conserved position in all structures
(
Figure 3a). We had hypothesized that extension of the carboxylate chain in (+)ꢀ85 might allow
interaction with Lys138 or Lys139 via a saltꢀbridge, consequently improving the binding affinity.
This interaction was suggested by molecular dynamics simulation where Lys139 was found to be
able to swing towards the ligand (SI, Section S5, Figure S30). However, this was not observed in
the cocrystals with either (+)ꢀ85, or with (+)ꢀ80 or (+)ꢀ81 (Figure 3a and SI, Section S4.3,
Figures S20 and S21). The carboxylate of (+)ꢀ85 in a gauche conformation of the side chain is
–
actually engaged in a waterꢀmediated interaction with Tyr63 via W1 (d(O ···O ) = 2.7 Å and
W1
d(O ···O ) = 2.8 Å). Two additional water molecules (W2 and W3) participate in solvating
W1
Tyr
the carboxylate; no other protein residues are in their close proximity. Various short apolar
proteinꢀligand contacts are depicted in the SI, Section 4.3, Figure S19. The isopropyl and methyl
ACS Paragon Plus Environment
29