Repurposing a Library of Human Cathepsin L Ligands: Identification of Macrocyclic Lactams as Potent Rhodesain and Trypanosoma brucei Inhibitors

Article abstract of DOI:10.1021/acs.jmedchem.7b01869

Rhodesain (RD) is a parasitic, human cathepsin L (hCatL) like cysteine protease produced by Trypanosoma brucei (T. b.) species and a potential drug target for the treatment of human African trypanosomiasis (HAT). A library of hCatL inhibitors was screened, and macrocyclic lactams were identified as potent RD inhibitors (K_i < 10 nM), preventing the cell-growth of Trypanosoma brucei rhodesiense (IC₅₀ < 400 nM). SARs addressing the S2 and S3 pockets of RD were established. Three cocrystal structures with RD revealed a noncovalent binding mode of this ligand class due to oxidation of the catalytic Cys25 to a sulfenic acid (Cys-SOH) during crystallization. The P-glycoprotein efflux ratio was measured and the in vivo brain penetration in rats determined. When tested in vivo in acute HAT model, the compounds permitted up to 16.25 (vs 13.0 for untreated controls) mean days of survival.

Full text of DOI:10.1021/acs.jmedchem.7b01869

Subscriber access provided by UNIV OF DURHAM
Repurposing a Library of Human Cathepsin L Ligands: Identification of
Macrocyclic Lactams as Potent Rhodesain and Trypanosoma brucei Inhibitors
Maude Giroud, Uwe Dietzel, Lilli Anselm, David Banner, Andreas Kuglstatter, Jörg Benz, Jean-Baptiste
Blanc, Delphine Gaufreteau, Haixia Liu, Xianfeng Lin, August Stich, Bernd Kuhn, Franz Schuler,
Marcel Kaiser, Reto Brun, Tanja Schirmeister, Caroline Kisker, François Diederich, and Wolfgang Haap
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b01869 • Publication Date (Web): 29 Mar 2018
Downloaded from http://pubs.acs.org on March 29, 2018
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 82
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Repurposing a Library of Human Cathepsin L
Ligands: Identification of Macrocyclic Lactams as
Potent Rhodesain and Trypanosoma brucei
Inhibitors
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
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
†
,#
‡,#
°
°
°
Maude Giroud, Uwe Dietzel, Lilli Anselm, David Banner, Andreas Kuglstatter, Jörg Benz,
°
°
°
ꢀ
ꢀ
∆
JeanꢀBaptiste Blanc, Delphine Gaufreteau, Haixia Liu, Xianfeng Lin, August Stich, Bernd
Kuhn, Franz Schuler, Marcel Kaiser, Reto Brun, Tanja Schirmeister, Caroline Kisker,
°
°
§, ¥
§, ¥
||
‡
†
°
François Diederich, * Wolfgang Haap *
†
Laboratorium für Organische Chemie, ETH Zurich, VladimirꢀPrelogꢀWeg 3, 8093 Zurich,
Switzerland
‡
RudolfꢀVirchow Center for Experimental Biomedicine, University of Würzburg, Josefꢀ
SchneiderꢀStrasse 2, 97080 Würzburg, Germany
°
Roche Pharmaceutical Research and Early Development (pRED), Roche Innovation Center
Basel, F. HoffmannꢀLa Roche Ltd., Grenzacherstrasse 124, 4070 Basel, Switzerland
ꢀ
Roche Pharma Research and Early Development, Roche Innovation Center Shanghai, 720
Cailun Road, Pudong, Shanghai 201203, China
∆
Department of Tropical Medicine, Medical Mission Institute, Salvatorstrasse 7, 97074
Würzburg, Germany
§
Swiss Tropical and Public Health Institute, Socinstrasse 57, 4051 Basel, Switzerland
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 82
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
¥
University of Basel, Petersplatz 1, 4003 Basel, Switzerland
‖||
Institut für Pharmazie und Biochemie, Johannes GutenbergꢀUniversität Mainz, Staudinger
Weg 5, 55128 Mainz, Germany
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
KEYWORDS. macrocycle – rhodesain – trypanosome – sleeping sickness – cysteine protease –
human African trypanosomiasis – drug repurposing – target repurposing.
ACS Paragon Plus Environment
2

Page 3 of 82
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
ABSTRACT: Rhodesain (RD) is a parasitic, human cathepsin L (hCatL) ꢀlike cysteine
protease produced by Trypanosoma brucei (T. b.) species and a potential drug target for the
treatment of human African trypanosomiasis (HAT). A library of hCatL inhibitors was screened,
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
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
and macrocyclic lactams were identified as potent RD inhibitors (K < 10 nM), preventing the
i
cellꢀgrowth of T. b. rhodesiense (IC50 < 400 nM). SARs addressing the S2 and S3 pockets of RD
were established. Three cocrystal structures with RD revealed a nonꢀcovalent binding mode of
this ligand class, due to oxidation of the catalytic Cys25 to a sulfenic acid (Cys–SOH) during
crystallization. The Pꢀglycoprotein efflux ratio was measured and the in vivo brain penetration in
rats determined. When tested in vivo in acute HAT model, the compounds permitted up to 16.25
(vs. 13.0 for untreated controls) mean days of survival.
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 82
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
Human African trypanosomiasis (HAT or African sleeping sickness) is a protozoan infection
caused by two subspecies of Trypanosoma brucei (T. b.): T. b. gambiense, responsible for the
1
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
chronic form of HAT, and T. b. rhodesiense, responsible for the more acute form of the disease.
1
The third subspecies, T. b. brucei, causes Nagana disease in cattle. HAT can be divided into two
clinical stages, the haemolymphatic stage, which, if left untreated, progresses into stage 2, the
2
,3
neurological stage. To date, only five drugs against this pathogen are available, the most recent
development being the nifurtimoxꢀeflornithine (NECT) combination therapy, launched by the
4
Drugs for Neglected Diseases initiative (DNDi) in 2009, and recommended by WHO thereafter.
Because of the severe side effects of these drugs, research towards the development of novel
5–9
compounds is urgently required.
In the field of neglected tropical diseases (NTDs), it is
1
0,11
common to initiate industrial–academic partnerships,
where compounds or targets are
1
2
13,14
repurposed from previous drug programs. Such partnerships notably led to fexinidazole
and
1
5,16
oxaborole (SCYXꢀ7158),
which are currently in clinical trials and close to the verge of being
17
introduced in clinical use.
Parasitic cysteine proteases have emerged as attractive targets for drug discovery due to their
1
8,19
indispensable roles in disease propagation.
Bloodstream Trypanosoma brucei parasites
express two papain family cysteine proteases: the human cathepsin L (hCatL)ꢀlike rhodesain
2
0–22
(
RD) and an hCatBꢀtype enzyme TbCatB, with overlapping activity profiles.
RD is involved
2
3
in the parasite’s iron homeostasis,
in performing the turnover of variant surface
24,25
26
glycoproteins,
in degrading the host immunoglobulins, and might help the parasite to cross
2
6
27
the bloodꢀbrain barrier (BBB). RD is structurally highly similar to hCatL, a potential drug
ACS Paragon Plus Environment
4

Page 5 of 82
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
28
29–31
target against diabetes, atherosclerosis and abdominal aortic aneurysms, various cancers,
3
2
and Alzheimer’s disease.
Based on the structural similarities of hCatL and RD, a library of 49 hCatL inhibitors, designed
for a metabolic disease campaign at F. HoffmannꢀLa Roche Ltd., was screened against RD.
Herein, we report the identification of macrocyclic ligands, such as 1 and 2, as potent,
trypanocidal RD inhibitors, with already good drugꢀlike properties. Efforts towards their further
derivatization and the optimization of pharmacokinetic (PK) properties were deployed. Using
structureꢀbased drug design (SBDD), the macrocycles were optimized towards higher
selectivities for RD versus hCatL (the compounds tested so far did not inhibit other human
cathepsins, such as B, K, S, and V). In addition, three crystal structures of macrocycles in
complex with RD (PDB IDs: 6EX8, 6EXO, and 6EXQ), which add valuable structural
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
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
3
information to the two existing RD crystal structures (PDB IDs: 2P7U , 1.6 Å resolution and
22
2P86 , 1.16 Å resolution), are reported, together with 3 crystal structures of macrocycles in
34
complex with hCatL (PDB IDs: 6EZP, 6EZX, and 6F06). Brain penetration, which is crucial
for targeting stage 2 HAT, was evaluated in in vitro and in vivo assays. At last, selected
macrocycles were tested in acute in vivo HAT models.
RESULTS AND DISCUSSION
3
5
Hit Identification. A series of 49 structurally diverse hCatL inhibitors, provided by F.
3
6–
HoffmannꢀLa Roche Ltd., was tested in a fluorometric assay against the cysteine protease RD.
38
Out of 49 ligands, 44 hits for RD with inhibitory constants K < 0.5 ꢀM were identified, among
i
them 24 compounds with K < 0.01 ꢀM. These hits belong to two different ligand classes:
i
prolineꢀbased inhibitors, already reported in the context of halogen bonding in the S3 pocket of
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 82
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
39
35
hCatL, and macrocycles. The most potent hits were tested in a first screen against the parasite
®
40
T. b. brucei in an Alamar blue assay to evaluate their ability to inhibit in vitro cellꢀgrowth
3
5
(Figure S1). The macrocyclic series 1–18 (Table S1), among them 5 compounds with IC
5
0
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
values < 0.4 ꢀM and no observable cytotoxicity (rat myoblast Lꢀ6 cells, IC ꢀvalues > 50 ꢀM),
50
was superior over the proline series, in both the enzymatic and parasitic assays. They fulfill the
4
1
hit criteria of the WHO and were therefore chosen as starting point for further optimization.
Moreover, the polar surface area (PSA), an important molecular descriptor for brain
4
2
penetration, and thus a criterion for targeting the second, neuroencephalitic stage of HAT, was
significantly lower for the macrocyclic series than for the nonꢀmacrocyclic proline series (Figure
S2). In conformational analyses of the macrocycles in the unbound state, the lowestꢀenergy
43,44
conformers showed an intramolecular Hꢀbond,
which reduces the polar surface area by ~10
2
Å compared to the bound state and thus contributes to their high trypanocidal activity (see
Section S2 in the SI). The structures of the most potent hits 1 and 2 are shown in Figure 1.
Although not particularly selective against hCatL, 1 displayed excellent selectivity towards the
human cathepsins B (IC = 9000 nM), K (IC = > 25000 nM), S (IC = 2700 nM), and V (IC
5
0
50
50
50
= 1043 nM).
Figure 1. Structure of the two hits 1 and 2 and their binding affinities for RD and hCatL,
respectively, as well as the cellꢀgrowth inhibition of T. b. brucei.
ACS Paragon Plus Environment
6

Page 7 of 82
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Motivated by these promising results, we were interested in further pursuing the optimization
and evaluation of this macrocyclic ligand class as RD inhibitors against HAT.
3
3
Ligand Design. Two crystal structures of RD, released in 2009 and 2010 (PDB IDs: 2P7U
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
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
2
and 2P86 ), respectively, and the coꢀcrystal structures of related macrocycles with hCatL (see
Section S7 in the SI) enabled the use of SBDD for further ligand optimization. Macrocycle 1
possesses a nitrile warhead, which is supposed to undergo reversibleꢀcovalent binding with the
3
9
45
catalytic Cys25 at the active site of RD. The formed thioimidate is stabilized by the oxyanion
46
hole of the protease, as seen in the modeled (with MOLOC, MAB force field) binding mode of
3
3
1
in RD (protein coordinates taken from PDB ID: 2P7U , 1.65 Å resolution) (insert of Figure
2A). In analogy to the interactions in hCatL, the ligand forms hydrogen bonds to the backbone
NH and C=O groups of Gly66 and Asp161, respectively, and fills the S2 and S3 pockets of RD,
thereby establishing multiple lipophilic contacts with the enzyme (Figure 2B). There is high
sequence conservation in the ligand binding site between RD and hCatL, however some amino
acid differences become apparent in the S2 and S3 pockets (Figure 2B), which were targeted to
improve the selectivity for RD. While this ligand class already contains many drugꢀlike
properties, modifications to increase the metabolic stability by changing the saturated
macrocyclic bridge and by introducing amide isosters, were undertaken. Compounds 1–18
(Table S1), tested in the initial screen, featured salicylic acid derivatives addressing the S3
pocket of RD. In a first step, the introduction of extended S3 pocket vectors was initiated, to
probe if these compounds will possess an increased selectivity against hCatL. In a second step,
aromatic heterocycles were introduced into the S3 pocket vector and later also on the phenyl ring
filling the S2 pocket.
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 82
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
A
B
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 2. (A) Predicted binding mode of macrocycle 1 in complex with RD (protein coordinates
3
3
taken from PDB ID: 2P7U , 1.65 Å resolution). Color code: gray C , green Cligand, red O, blue
RD
N, yellow S. Atomic surface shown. The insert shows the expected thioimidate formation by the
catalytic Cys25. (B) Schematic binding mode of 1, showing the putative covalent bond between
Cys25 and the nitrile head group and the stabilization of the formed thioimidate by the oxyanion
hole. The residues from RD are labeled in black and the residues from hCatL in magenta.
Hydrogen bonds are represented as red dashed lines in (A) and (B).
SAR Exploration in the S3 Pocket of RD. Synthesis. The syntheses of macrocycles 1, 2, and
1
9a–e relied on a convergent approach (Scheme 1). The amino building block 20 was prepared
from commercially available Nꢀ[(tertꢀbutoxycarbonyl]ꢀ3ꢀ(chloro)ꢀLꢀtyrosine
dicyclohexylammonium salt (21), which was doubly allylated to give 22 (Scheme 1a).
Saponification of 22 yielded 23, and amide coupling of 23 with 1ꢀaminocyclopropylcarbonitrile
hydrochloride in the presence of HATU and iPr NEt afforded Bocꢀprotected 24 (not shown). Nꢀ
2
Boc deprotection in neat formic acid, which avoided any decomposition or hydrolysis of the
ACS Paragon Plus Environment
8

Page 9 of 82
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
nitrile, gave 20. Similarly, salicylic acid derivatives 25a–f were doubly allylated (26a–f) and
saponified, yielding 27a–f (Scheme 1b). Amide coupling between 20 and 27a–f led to
nd
compounds 28a–e (Scheme 1c). Ringꢀclosing metathesis using either Grubbs catalyst 2
4
7
48
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
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
generation or GrubbsꢀHoveyda catalysts provided macrocycles 19a–e featuring a transꢀ
configured alkene as a mixture of atropisomers (see below). Reduction of the alkene in 19a,b
using H and Pd/C 10% afforded macrocycles 1 and 2 in moderate yields, because of ringꢀ
2
3
5
opening of the macrocycles.
35
35
a
Scheme 1. Synthesis of Macrocycles 1, 2, and 19a–e
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 82
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
Reagents and conditions: (i) allyl bromide, Cs CO , DMF, 16 h at 22 °C, 62%; (ii) LiOH,
2
3
THF/H O/MeOH 2:2:1, 1 h at 22 °C, 100%; (iii) 1ꢀaminocyclopropylcarbonitrile·HCl, HATU or
TBTU, iPr NEt, DMF or DMA, 4 h at 22 °C, 54%; (iv) HCOOH, 2 h at 22 °C, 100%; (v) allyl
bromide, Cs CO , DMF, 16 h at 22 °C, 89–100%; (vi) allyl bromide, K CO , acetone, 16 h at 22
°C, 89–100%; (vii) LiOH, THF/H O/MeOH 2:2:1, 1 h at 22 °C, 19–99%; (viii) HATU, iPr NEt,
DMF, 2 h at 22 °C, 22–100%; (ix) Grubbs catalyst 2 generation, CH Cl , 2 h at 50 °C; 43–
3%; (x) HoveydaꢀGrubbs 2 generation catalyst, toluene, 2 h at 90 °C, 43–53%; (xi) H , Pd/C
0%, EtOAc, 4 h at 23 °C, 50–57%. The substituent R represents the vector addressing the S3
pocket of RD and is defined in Table 1. DCHA = dicyclohexylamine; DMA = N,Nꢀ
dimethylacetamide;
bis(dimethylamino)methylene]ꢀ1Hꢀ1,2,3ꢀtriazolo[4,5ꢀb]pyridinium
hexafluorophosphate;
tetrafluoroborate; THF = tetrahydrofuran.
2
2
2
3
2
3
2
2
nd
2
2
nd
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
5
1
2
1
DMF
=
N,Nꢀdimethylformamide;
HATU
=
1ꢀ
[
3ꢀoxide
TBTU
=
Oꢀ(Benzotriazolꢀ1ꢀyl)ꢀN,N,N′,N′ꢀtetramethyluronium
3
5
In addition, macrocycle 8 could be oxidized in the presence of 2 equivalents of mCPBA,
giving 29 (Scheme 2).
a
Scheme 2. Derivatization of 29.
a
Reaction conditions: (i) mCPBA (2 eq.), CH Cl , 20 min. at 0 °C, 36%. mCPBA = 3ꢀ
2
2
chloroperbenzoic acid.
Enzyme Inhibition and Antiꢀparasitic Activities. The inhibition of RD and hCatL was
3
6–38
quantified in a fluorometric assay, as previously reported (Table 1).
A selectivity index (SI)
was defined as SI = K (hCatL)/K (RD). The cellꢀgrowth inhibition against T. b. rhodesiense was
i
i
®
40
determined in the Alamar blue assay and the cytotoxicity evaluated by determining the cellꢀ
growth inhibition of rat myoblast Lꢀ6 cells. All macrocycles were inhibiting RD and hCatL in the
ACS Paragon Plus Environment
10

Page 11 of 82
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
single to doubleꢀdigit nanomolar K range. Macrocycle 19a inhibited the cellꢀgrowth of T. b.
i
rhodesiense with an IC of 4 nM, but increasing the size of the S3 substituent as in 19b–d
5
0
(Table 1) decreased the trypanocidal activity to the 300 nM IC range. Based on our structural
5
0
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
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
models, the high potency of macrocycle 19e (K (RD) = 1 nM, IC (T. b. rhodesiense) = 0.5 n )
M
i
50
might be accounted for by the parallelꢀdisplaced stacking interactions of the terminal ring of the
naphthyl substituent with Phe61 (d(C···CPhe61) = 3.4 Å) (Figure S4 in the SI). Except for 19c
(
IC (Lꢀ6) = 6.49 ꢀM), low to moderate cytotoxicities (IC (Lꢀ6) > 35 ꢀM) were observed in
5
0
50
this series and the compounds possess excellent selectivities relative to Lꢀ6 cells.
Table 1. Inhibition of RD and hCatL, Selectivity Indices (SI), Cell-Growth Inhibition of T.
brucei rhodesiense, and Cytotoxicity on L-6 cells of Selected Inhibitors.
[
c]
IC50 (T.
b. rhod.)
[ꢀM]
[e]
[a]
[a]
IC50 (L-6
Ki
Ki
[
b]
cells)
Compound
19a
R =
(RD)
[nM]
(hCatL)
[nM]
SI
[
d]
[ꢀM]
(
SI )
0.002
20.2
2.0
2.4
0.1
7.8
37.0
45.3
(
18500)
0
.347
1
9b
15.5
(131)
ACS Paragon Plus Environment
11

Journal of Medicinal Chemistry
Page 12 of 82
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
0
0
.338
1
9c
3.0
5.4
7.3
7.4
2.4
1.4
6.49
44.3
(
19)
.336
1
9d
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
(
132)
0
.0005
1
9e
1.0
6.1
1.2
3.7
1.2
0.6
43.8
(
87600)
0
.332
2
9
>100
[f]
(
n.a. )
[a] Average of two measurements, each performed in duplicate; standard deviations < 10%; [b]
SI = selectivity index = K (hCatL)/K (RD); [c] Average of at least 2 measurements, T. b.
rhodesiense strain STIB 900, bloodstream form (trypomastigotes); standard deviations < 10%;
d] Selectivity index relative to the Lꢀ6 cells; [e] Average of at least 2 measurements, rat skeletal
myoblast cell Lꢀ6 strain; standard deviations < 10%; [f] N.a. = not applicable.
i
i
[
SAR Exploration in the S2 Pocket of RD. Synthesis. Macrocycles 30a,d and 31a–d were
derived from Lꢀtyrosine 32, which was iodinated to give 33 according to known procedures
4
9
(
Scheme 3). Subsequent NꢀBoc protection (34, not shown) and bisꢀallylation with allyl bromide
and Cs CO afforded 35 in 78% yield, which was saponified with LiOH to give 36 in 94% yield.
2
3
Amide coupling of 36 with 1ꢀaminocyclopropanecarbonitrile hydrochloride gave 37 in 85%, and
deprotection in neat HCOOH at 23 °C provided amine 38 in quantitative yield. A second amide
coupling between 38 and carboxylic acids 27a or 27f afforded 39 and 40 in 33 and 70% yield,
nd
respectively. Ringꢀclosing metathesis of 39 and 40 using Grubbs catalyst 2 generation in
CH Cl at 50 °C gave the corresponding macrocycles 41 and 42 in 37 and 46% yield,
2
2
respectively. All macrocycles isolated were shown to possess a transꢀconfigured alkene.
Reduction of the olefin with the commonly used Pd/C 10% and H was initially attempted.
2
ACS Paragon Plus Environment
12

Page 13 of 82
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
However, under these conditions, no hydrogenation was observed and instead deꢀiodination
occurred. While treating 41 and 42 with Pt/C 10% and H led to ring opening at both C –O
2
sp2
positions, RaneyꢀNickel at 10 bars of H , 50 °C, overnight afforded 43 and 44 in 70 and 77%
2
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
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
yield, respectively. SuzukiꢀMiyaura crossꢀcoupling of 43 or 44 with potassium trifluoroborates
50
afforded macrocycles 30a,d and 31a–d in 70–77% yield. Reduction of the alkene functionality
prior to further derivatization was crucial, as all attempted crossꢀcouplings failed in its presence.
a
Scheme 3. Preparation of Macrocycles 30a,d and 31a–d
ACS Paragon Plus Environment
13

Journal of Medicinal Chemistry
Page 14 of 82
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
Reaction conditions: (i) I , NH OH, 1 h at 22 °C, 33%; (ii) Boc O, Et N, H O/1,4ꢀdioxane
2
4
2
3
2
1
:1, 13 h at 22 °C, quant.; (iii) allyl bromide, Cs CO , DMF, 5 h at 22 °C, 78%; (iv) LiOH,
THF/MeOH/H O 2:2:1, 1 h at 22 °C, 94%; (v) aminocyclopropylcarbonitrile hydrochloride,
HATU, iPr NEt, DMF, 3 h at 22 °C, 85%; (vi) HCOOH, 4 h at 22 °C, quant.; (vii) 27a or 27f,
HATU, iPr NEt, DMF, 2 h at 22 °C, 33–70%; (viii) Grubbs catalyst 2 generation, CH Cl , 1 h
2 3
2
2
nd
2
2
2
2
at 50 °C, 37–46%; (ix) H , RaꢀNi, EtOAc, 20 h at 50 C and 10 bars, 70–77%; (x) R BF K,
2
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
[
PdCl (dppf)·CH Cl ], Et N, Cs CO , 3:1:2 iPrOH/H O/CH Cl , 12 h at 90 °C, 13–85%; (xi)
2 2 2 3 2 3 2 2 2
2
R BF K, [Pd(OAc) ], RuPhos, Na CO , EtOH, 18 h at 85 °C, 13–85%. Dppf = 1,1'ꢀ
3
2
2
3
bis(diphenylphosphanyl)
diisopropoxybiphenyl.
ferrocene;
RuPhos
=
2ꢀDicyclohexylphosphinoꢀ2′,6′ꢀ
Biological Results. Macrocycles 41–44 inhibited RD (K = 20–89 nM) efficiently, with
i
however no selectivity against hCatL (K = 2.4–13.4 nM, SI = 0.1) (Table 2). The parasitic cell
i
growth was inhibited in the subꢀ to singleꢀdigit nanomolar range (IC = 0.3–3 nM). Ligands 41–
5
0
4
4 presented moderate cytotoxicity in the Lꢀ6 rat skeletal myoblast assay (IC = 37–56 ꢀM).
5
0
The presence of the trifluoromethyl group on the P3 substituent seems to consistently enhance
the trypanocidal activity as was also observed for the comparison between 30a,d and 31a–d.
Introducing a 5ꢀ or 6ꢀmembered heterocycle at the 3ꢀposition of the Lꢀtyrosine moiety as in
30a,d and 31a–d led to a substantial loss in binding affinity of 3–20ꢀfold in RD and hCatL,
which was reflected in the results of the cellꢀgrowth inhibition assays (IC (T. b. rhodesiense) =
5
0
0
.03–3.88 ꢀM). Some of the loss in affinity upon introduction of a 6ꢀmembered heterocycle at
the 3ꢀposition of the Lꢀtyrosine moiety may be rationalized by the fact that substantial ligand
strain is generated to fit the ligand into the S2 pocket of RD (Figure 3). We note that some
compounds potently inhibited trypanosome growth, despite having relatively poor affinity for
RD, and tentatively assign this to offꢀtarget effects of the novel biarylꢀtype pharmacophore.
ACS Paragon Plus Environment
14

Page 15 of 82
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
A
B
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
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. (A) Predicted binding mode of macrocycle 31b in complex with RD (protein
2
2
coordinates taken from PDB ID: 2P86 , 1.16 Å resolution), highlighting the unfavorable torsion
angle = –4 ° in the bound state; (B) Macrocycle 31b optimized in the free state, showing a
τ
probable hydrophobic collapse of the ligand in solution. The torsion angle
τ = 42° in the
51
unbound state is ideal for such biaryl systems. Color code: green Cligand, gray Cprotein, red O,
blue N, yellow S, cyan F. Distances given in Å.
Table 2. Inhibition of RD and hCatL, Selectivity Indices (SI), Cell-Growth Inhibition of T.
brucei rhodesiense, and Cytotoxicity on L-6 cells of Selected Inhibitors.
[c]
[e]
IC50 (T.
b. rhod.)
[ꢀM]
IC50
[
a]
[a]
Ki
Ki
(L-6
cells)
1
2
[b]
Compound
R
R
(RD)
[nM]
(hCatL)
[nM]
SI
[
d]
(SI )
[ꢀM]
ACS Paragon Plus Environment
15

Journal of Medicinal Chemistry
Page 16 of 82
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
[f]
4
4
1
2
H
I
I
2.5
8.7
n.d.
n.d.
0.6
n.d.
.3
(101)
n.d.
0
CF
5.6
30.15
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
43
H
I
I
n.d.
n.d.
5.8
n.d.
0.1
n.d.
.03
1892)
.93
n.d.
0
4
4
CF
3
57.4
56.75
(
1
30a
31a
31b
H
CF
CF
CF
H
837
751
997
824
1.2
1.1
n.d.
0.8
1.0
n.d.
>100
72.2
[
g]
N
N
N
(n.a. )
0.018
3
3
3
(4011)
0
.509
N
>20000
870
6393.3
676
>100
>100
86.5
[g]
(n.a. )
0.019
N
31c
N
[g]
(n.a. )
1
.23
70)
0.14
3
3
0d
1d
1085.9
1105.7
1364.2
HN
HN
N
N
(
4
5% inh.
at 20
CF
>100
3
[g]
ꢀM
(n.a. )
[
a] Average of two measurements, each performed in duplicate, standard deviations < 10%; [b]
SI = selectivity index = K (hCatL)/K (RD); [c] Average of at least 2 measurements, T. b.
rhodesiense strain STIB 900, bloodstream form (trypomastigotes), standard deviations < 10%;
d] Selectivity index relative to the Lꢀ6 cells; [e] Average of at least 2 measurements, rat skeletal
myoblast cell Lꢀ6 strain, standard deviations < 10%; [f] n.d. = not determined; [g] n.a. = not
applicable.
i
i
[
Atropisomerism and Racemization of Selected Macrocycles. Atropisomerism, i.e. the
5
2
hindered rotation about single bonds leading to planar chirality, is an often overlooked source
ACS Paragon Plus Environment
16

Page 17 of 82
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
53–56
of stereoisomerism leading to the formation of diastereoisomers or enantiomers.
Discrete,
stable atropisomers may feature greatly differing binding affinities as well as physicoꢀchemical
5
5
and ADME properties.
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
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
Atropisomers have been classified by LaPlante et al. in three classes I–III, based on their halfꢀ
54
lives of interconversion (t ). Class I atropisomers feature rapid rotation and short halfꢀlives for
1
/2
‡
–1
isomerization (ꢁG < 20 kcal mol ). In this case, atropisomerism is not relevant for the
‡
stereochemical integrity of a drug. Class II atropisomers have higher barriers to rotation (ꢁG =
–
1
2
0–30 kcal mol ), and the halfꢀlifes t of the atropisomers are now in the range of minutes to
1/2
several hours, relevant for the stereochemical integrity. It is therefore important for drug
development to analyze Class II atropisomer interconversion in greater detail. Exposure to the
5
6
active stereoisomer will be controlled by the interconversion rate. Also, halfꢀlives of the
isomers can become longer than the halfꢀlifes of their in vivo elimination, leading to in vivo
enrichment of one isomer. A method used for detecting class II atropisomers is HPLC analysis
on chiral supports, as exemplified by the determination of the racemization halfꢀlife of
5
6
‡
–1
anthranilic acid HCV polymerase inhibitors. Class III atropisomers (ꢁG > 30 kcal mol ) have
halfꢀlifes t1/2 of years, and the individual stereoisomers can be separated, isolated, and safely
stored.
5
2
Most of the macrocycles reported herein possess planar chirality and were isolated as 1:1 or
56
3
:2 atropisomeric mixtures. We applied a method similar to the one reported by LaPlante et al
to determine the halfꢀlives for isomerization of macrocycle 41 (Figure 4, Section S5 in the SI).
The atropisomeric mixture of 41 was separated by HPLC on a chiral column (heptane/EtOH 7:3
for 20 min) at 22 °C. Upon injection on the column, two peaks were observed at a 57:43 ratio,
corresponding to the (R ,S) and (S ,S) (where p = planar chiral) diastereoisomers (not assigned).
p
p
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 82
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
For the analysis, one peak (B) was collected, and reꢀinjected on the same HPLC column
–
1
(
injection volume = 10
points t = 5, 25, 45, 65, 90, and 200 min (Section S5 in the SI). After 200 min, integration of the
peaks was performed at = 220 nm and the initial atropisomeric mixture was recovered.
ꢀL, flow rate = 1 mL min ) under the same conditions at various time
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
λ
Figure 4. Slow equilibrium between the two diastereoisomeric atropisomers A and B of
macrocycle 41. The drawing of compound A depicts the (R ,S) configuration and the drawing of
p
5
2
B the (S ,S) configuration, assigned according to the rules for planar chirality but the
p
configurations were not experimentally assigned. Xꢀray crystal structures later showed that the
(Sp,S) diastereoisomer of the macrocycles is bound exclusively (see below).
A halfꢀlife for isomerization of τ1/2 = 37.5 min was extrapolated by curveꢀfitting (Figure S20 in
the SI), suggesting that 41 lies within class II at 22 °C. However upon repeating the experiment
at 37 °C, immediate reꢀequilibration was observed. Therefore, the exposure of macrocycle 41
should not be controlled by its interconversion halfꢀlife.
Structural Analysis in the Presence of RD. The binding mode of the macrocyclic ligands
was determined for RD and hCatL by Xꢀray crystallography in order to better understand their
molecular recognition pattern (Figures 5–6, Sections S6 (RD) and S7 (hCatL) in the SI). Three
3
5
35
crystal structures of ligands 1, 3 , and 7 in complex with RD were obtained (1: PDB ID:
ACS Paragon Plus Environment
18

Page 19 of 82
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
35
35
6
EX8, 1.6 Å resolution; 7 : PDB ID: 6EXO, 1.9 Å resolution, and 3 : PDB ID: 6EXQ, 2.5 Å
resolution, ligand structures are shown in Figure 5D).
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
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
D
Figure 5. (A) F ꢀF Electron density map shown as a blue mesh of ligand 1 and the active site
o
c
cysteine of RD (PDB ID: 6EX8, 1.6 Å resolution). The enzyme is shown in gray and ligand 1 is
shown in green. (B) Electron density map of ligand 7 and the active site cysteine of RD (PDB
ID: 6EXO, 1.9 Å resolution), the color code is the same as in A, ligand 7 is shown in orange; (C)
Electron density map of ligand 3 and the active site cysteine of RD (PDB ID: 6EXQ, 2.5 Å
resolution), the color code was maintained, and ligand 3 is shown in violet; (D) Chemical
structures of ligands 1, 3, and 7. The F ꢀF maps were calculated in the absence of the inhibitor
o
c
and the active site cysteine and are contoured at 2.5 σ. Refinement for Cys25ꢀSOH in (A) was
carried out in two different orientations at 50% occupancy each. In (C), the oxidation of Cys25 is
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 82
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
not visible due to the lower resolution of the dataset but can be assumed because of the same
crystallization conditions for all three structures.
The crystal structure of ligand 1 bound to RD at 1.6 Å resolution is discussed here in more
detail (PDB ID: 6EX8, Figure 6A). The complex crystallized in space group P1, with two chains
of mature enzyme in the asymmetric unit. The observed binding mode of ligand 1 is in
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
accordance with our design (Figure 2). In all three crystal structures, the (S ,S)ꢀconfigured
p
atropisomer binds exclusively to the protein. However, to our surprise, the nitrile warhead, which
was systematically covalently bound to the nucleophilic Cys25 in the hCatL complexes with
39
38,57
58
prolineꢀbased inhibitors, triazine nitriles,
and dipeptidyl nitriles, did not form a covalent
bond with Cys25 (Figure 6A). Instead, Cys25 was oxidized to a sulfenic acid (Cys25ꢀSOH),
5
9
which is most likely due to the extended time until crystals formed. In complex 1, the C≡N lies
at a distance of d(N≡C···S) = 3.3 Å from the Cys25ꢀSOH. This nitrile is further stabilized by a
N···HNGln19 contact (d(N···HNGln19) = 3.3 Å) with the side chain of Gln19 being in the oxyanion
hole. Our enzyme kinetics studies showed that the inhibitors bind in a reversibleꢀcovalent
37,38
manner.
It is possible, that the inhibitors formed a covalent bond and reverted over time to
the nonꢀcovalent state prior to crystallization, together with the oxidation of Cys25, which is also
6
0
reversible.
The S2 pocket is filled with the aromatic ring of the 3ꢀchloroꢀLꢀtyrosine moiety, which is
sandwiched between Leu67 and Leu160 and undergoes C–H···π interactions with Leu67
(d(CLeu67···C) = 3.6 Å) and Leu160 (d(CLeu160···C) = 3.9 Å). The chloro substituent is in the
vicinity of Met68 (d(Cl···SMet68) = 4.1 Å) and forms a short contact to the terminal methyl group
of Met68 (d(Cl···CMet68) = 3.6 Å) (contacts not shown). In the S3 pocket of RD, the
trifluorophenyl ring lies on top of the Gly66–Leu67 amide bond plane at the entrance of the
ACS Paragon Plus Environment
20

Page 21 of 82
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
pocket, with an interplanar angle of ~45º (d(C···C=OGly66) = 3.7 Å). The fluorine atoms of the
trifluoromethyl group display short, potentially favorable contact distances with the phenyl ring
of Phe61 (d(CPhe61···F) = 3.5 Å), and the backbone NH of Asn70 (d(NAsn70···F) = 3.2 Å). At
d(C=OAsp60···F) = 4.2 Å, the third Fꢀatom avoids repulsive interactions with the side chain
carboxylate of Asp60 (not shown). An overlay of the structures of macrocycle 7 in RD (PDB ID:
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
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
6
EXO) and hCatL (PDB ID: 6EZP) shows two similar binding modes despite the fact that 7
binds covalently to hCatL and not to RD (Figure 6B). All three macrocycles bind in a similar
way to RD, as shown by the overlay of the three coꢀcrystal structures (Figure 6C). An additional
Xꢀray structure of 14 bound to hCatL (PDB ID: 6F06) reveals as well a binding mode of 14 in a
nonꢀcovalent fashion (Figure S24). In this case, the sulfur atom of Cys25 forms a complex with a
2+
–
Zn ion together with a Cl ion and His162 making the sulfur atom unavailable for thioimidate
formation.
ACS Paragon Plus Environment
21

Journal of Medicinal Chemistry
Page 22 of 82
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 6. (A) Binding mode of 1 in complex with RD (PDB ID: 6EX8, 1.6 Å resolution), color
code: green Cligand, gray Cenzyme; (B) Overlay of the crystal structures of 7 in complex with RD
(PDB ID: 6EXO, 1.9 Å resolution) and with hCatL (PDB ID: 6EZP, 2.34 Å resolution), color
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
code: green Cligand,RD, orange Cligand,hCatL, gray C , violet ChCatL; (C) Overlay of all three crystal
RD
structures of macrocycles 1 (in pink, PDB ID: 6EX8, 1.6 Å resolution), 7 (in gray, PDB ID:
6
EXO, 1.9 Å resolution) and 3 (in violet, PDB ID: 6EXQ, 2.5 Å resolution). Common color
code: red O, blue N, yellow S. Distances given in Å and represented as dashed black lines.
Drug Metabolism and Pharmacokinetics. Macrocycle 1 features reasonable solubility at pH
6
.5, and a logD7.4 value of 3.7 was determined (Table 3). In addition, 1 did not inhibit the major
cytochromes P450 significantly (IC (3A4) = 43 ꢀM, IC (2D6) = 11 ꢀM, IC (2C9) = 16 ꢀM).
5
0
50
50
Table 3. In Vitro Data of 1.
IC [ꢀM]
50
[a]
[b]
[c]
LYSA [ꢀg/mL]
logD7.4
PAMPA P
eff
CYPs 3A4; 2D6; 2C9
1
2
3.7
4
43/11/16
[
a] Solubility determined by lyophilisation solubility assays (LYSA) at pH 6.5; [b] logD7.4
=
intrinsic distribution coefficient between octanol and aqueous buffer (pH 7.4), measured in a
©
–1
CAMDIS assay; [c] Membrane permeability [nm s ] derived from the PAMPA assay.
The PK profile of 1 was characterized in mice, following intravenous administration (Table 4).
ACS Paragon Plus Environment
22

Page 23 of 82
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 4. Mouse Pharmacokinetic Profile of 1.
–1
–1 [a]
[b]
–1
[c]
[d]
CL (mL min kg )
Vss [L kg ]
t1/2 [h]
B/P
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
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
79
7.2
2.7
< 0.5
[a] Clearance; [b] Volume of distribution; [c] Halfꢀlife of compound in blood; [d] Brainꢀoverꢀ
plasma ratio, parameters calculated by nonꢀcompartmental analysis of plasma concentrationꢀtime
profiles.
Metabolic Stability Studies. Macrocycles 1 and 7 at 2 ꢀM concentration in DMSO were
incubated for 7 min in human and rat liver microsomes (see Section S9 in the SI). More than
9
8% of the parent compound 7 disappeared after the 7 min incubation. The results of mass
spectrometry experiments showed that the linker was a site of metabolism and was readily
oxidized ([M + 16] and [M + 32]) (Section S9). Thus, the modification of this linker, aiming at
removing at least one of the ether bridges, was undertaken.
Linker Modification. Synthesis. Initial efforts to stabilize the linker focused on fluorinating at
different positions, however all remained unsuccessful (see Section S10 in the SI). Using a
similar synthetic strategy to the one depicted in Scheme 1, at least one of the ether bridges could
be removed (Schemes 4 and 5). NꢀAlkylation of pyrazole derivatives 45a–d with 4ꢀbromobutꢀ1ꢀ
ene gave 46a–f. Subsequent saponification provided carboxylic acids 47a–f, which were coupled
with 20 to give 48a–f. Ringꢀclosing metathesis afforded macrocycles 49a–f.
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 82
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
Scheme 4. Preparation of Ligands 49a–d Described in Table 5.
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
Reaction conditions: (i) 4ꢀbromobutꢀ1ꢀene, Cs CO , DMF, 22 °C, 34–82%; (ii) LiOH,
2
3
THF/H O/MeOH 2:2:1, 1 h at 22 °C, 100%; (iii) 20, HATU, iPr NEt, DMF, 22 °C, 21–67%; (iv)
2
2
nd
HoveydaꢀGrubbs 2 generation catalyst, toluene, 2 h at 90 °C, 16–22%.
Similarly, NꢀAlkylation of pyridone derivatives 50a,b with 4ꢀbromobutꢀ1ꢀene gave 51a,b.
Subsequent saponification provided carboxylic acids 52a,b, which were coupled with 20 to give
53a,b. Ringꢀclosing metathesis afforded macrocycles 54a,b (Scheme 5).
Scheme 5. Preparation of Ligands 54a,b Described in Table 5.
a
Reaction conditions: (i) allyl bromide, Cs CO , DMF, 22 °C, 72–100%; (ii) LiOH,
2
3
THF/H O/MeOH 2:2:1, 1 h at 22 °C, 100%; (iii) 20, HATU, iPr NEt, DMF, 22 °C, 65–67%; (iv)
2
2
nd
HoveydaꢀGrubbs 2 generation catalyst, toluene, 2 h at 90 °C, 12–26%.
ACS Paragon Plus Environment
24

Page 25 of 82
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
35
In addition, treatment of 8 with 3 equivalents of mCPBA afforded epoxide 55 as a 3:2
mixture of atropisomers (Scheme 6). The diastereoselectivity of the epoxidation was tentatively
assigned considering the conformational restriction of the alkene and by 2D NMR.
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
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
35 a
Scheme 6. Derivatization of 8 .
a
Reaction conditions: (i) mCPBA (3 eq.), CH Cl , 55%.
2
2
61
Although alkyne metathesis remained unsuccessful, an alkyne bridge could be introduced to
yield macrocycles 56a and 56b (Scheme 7). Protection of salicylic acid derivative 57 gave 58,
which was Oꢀalkylated with 1,4ꢀdibromobutꢀ2ꢀyne to afford 59. NꢀBocꢀprotected 3ꢀchloroꢀLꢀ
tyrosine (21) was methylated to give 60, which was Oꢀalkylated with 59 to provide 61.
Saponification of 61 (giving 62, not shown) and subsequent amide coupling with 1ꢀ
aminocyclopropylcarbonitrile hydrochloride yielded 63. Deprotection in neat formic acid yielded
6
4, which was cyclized to yield atropisomers 56a and 56b. The structure of 56a was assigned by
2D NMR.
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 82
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 7. Preparation of 56a and 56b
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
Reaction conditions: (i) methylation; (ii) 1. CDI, DMF, 2 h at 40 °C; 2. DBU, tertꢀBuOH,
DMF, 14 h at 40 °C, 86%; (iii) 1,4ꢀdibromobutꢀ2ꢀyne, K CO , KI, acetone, 23 °C, 77%; (iv)
K CO , KI, acetone, 23 °C, 59%; (v) LiOH, 2:1 THF/H O, 23 °C; (vi)
aminocyclopropylcarbonitrile hydrochloride, HATU, CH Cl /DMF, 23 °C, 70%; (vii) HCOOH,
23 °C (viii) HATU, CH Cl , 23 °C, 68%. CDI = 1,1'ꢀcarbonyldiimidazole; DBU =
2
3
2
3
2
2
2
2
2
2
,3,4,6,7,8,9,10ꢀoctahydropyrimido[1,2ꢀa]azepine.
Enzyme Inhibition and Antiꢀparasitic Assays. The pyrazole series 54a,b led to potent inhibitors
of RD and hCatL and inhibited the cellꢀgrowth of T. b. rhodesiense in the nanomolar IC range
5
0
(
Table 5). Although no general selectivity was achieved with these compounds, 49d showed a
10ꢀfold selectivity for RD, which is the best measured for the entire macrocyclic ligand class.
Pyridone derivative 54a led to a decrease in binding affinity and cellꢀgrowth inhibition, however
5
4b was potent in the cellꢀbased assay for which a good explanation is not yet available.
Introducing an epoxide onto the bridge as in 55 significantly reduced the trypanocidal activities.
ACS Paragon Plus Environment
26

Page 27 of 82
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
On the other hand, compounds 56a,b with an alkyne linker were similarly potent to macrocycles
1
or 2. These compounds (except epoxide 55a) led to highly selective compounds for T. b.
rhodesiense relative to Lꢀ6 cells.
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
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. Inhibition of RD and hCatL, Selectivity Indices (SI), Growth Inhibition of T. b.
rhodesiense, and Cytotoxicity on L-6 Cells of Inhibitors 49a–d, 54a,b, 55, and 56a,b.
[
c]
IC50
[a]
[a]
(T. b.
rhod.)
[e]
Ki
(RD)
nM]
Ki
IC50
[b]
Compound
Structure
(hCatL)
[nM]
SI
(L-6)
[ꢀM]
[
ꢀM]
[
[
d]
(
SI )
0
.006
4
9a
52.4
17.2
40.3
2.5
0.8
>100
60.6
(
n.a.)
0
.005
4
9b
0.1
0.1
(12120)
0
.0006
49c
49d
18
2.3
38.5
40.0
(64167)
0
.0006
5.2
55.7
11
(66667)
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 82
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
4
7.8
> 20
[
f]
5
4a
> 20
ꢀ
M
n.a.
>100
ꢀM
(
n.a.)
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
0
.008
5
4b
299.0
273.5
0.9
16.2
(
2025)
O
O
S
O
1
6.9
O
5
5
O
H
HN
135.1
4.6
0.03
84.7
N
H
N
(5)
H
O
O
Cl
0.008
5700)
56a
80.6
24.1
0.3
0.3
45.6
47.1
(
0
.004
5
6b
4.8
1.2
(11775)
[
a] Average of two measurements, each performed in duplicate, standard deviations < 10%; [b]
SI = selectivity index = K (hCatL)/K (RD); [c] Average of at least 2 measurements, T. b.
rhodesiense strain STIB 900, bloodstream form (trypomastigotes), standard deviations < 10%;
[d] Selectivity index relative to the Lꢀ6 cells; [e] Average of at least 2 measurements, rat skeletal
myoblast cell Lꢀ6 strain, standard deviations < 10%; [f] N.a. = not applicable.
i
i
Selectivity of the Macrocycles Against Other Parasites. The in vitro cellꢀgrowth inhibitions
of selected macrocycles were determined for a panel of other protozoan parasites consisting of T.
cruzi, Leishmania donovani, and Plasmodium falciparum (Section S11 in the SI), the causative
ACS Paragon Plus Environment
28

Page 29 of 82
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
agents of Chagas’ disease, leishmaniasis, and malaria tropica, respectively. The inhibition by the
corresponding typical antiprotozoal drugs is reported for comparison. All compounds presented
good to excellent selectivity against the other parasites, suggesting that our inhibitors selectively
6
2
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
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
target RD, despite the structural similarities to the parasitic enzymes cruzain (from T. cruzi),
63
36
the leishmanial cysteine proteases, and falcipainꢀ2 (from P. falciparum). In addition, they all
possess moderate to low cytotoxicity, as demonstrated by the IC50 values against Lꢀ6 rat
myoblast cells.
P-gp Efflux Ratio and Brain Penetration. In order to be effective in stage 2 HAT, a drug
3
4
needs to be able to cross the blood brain barrier (BBB). One of the parameters used to evaluate
brain penetration is the Pꢀglycoprotein (Pꢀgp) efflux ratio (ER), which inversely correlates with
4
2,64
brain penetration for cellꢀpermeable compounds.
The human and mouse ER were determined
for macrocycles 1 and 2 in a transwell assay using Pꢀgp (MDR1) overexpressing LLCPK1 cells
(Table 6). Both macrocycles present a high Pꢀgp liability, with an ER in human of 26. With two
amide bonds and a nitrile in their structure, 1 and 2 are good Pꢀgp substrates, despite the putative
intramolecular Hꢀbond that should engage at least one of these amides. The PSAs (see Section
2
2
S2 in the SI for a full discussion) of 1 (80.2 Å ) and 2 (78.2 Å ) are however well within the
2
2 42
ideal range for CNS penetration (40 Å < PSA < 90 Å ).
ACS Paragon Plus Environment
29