Selective inhibitors of cyclin G associated kinase (GAK) as anti-hepatitis C agents

Article abstract of DOI:10.1021/jm501759m

Cyclin G associated kinase (GAK) emerged as a promising drug target for the treatment of viral infections. However, no potent and selective GAK inhibitors have been reported in the literature to date. This paper describes the discovery of isothiazolo[5,4-

Full text of DOI:10.1021/jm501759m

Subscriber access provided by SETON HALL UNIV
Article
Selective inhibitors of Cyclin-G associated kinase (GAK) as anti-HCV agents
Sona Kovackova, Lei Chang, Elena Bekerman, Gregory Neveu, Rina Barouch-Bentov, Apirat Chaikuad,
Christina Heroven, Michal Sala, Steven Cesar De Jonghe, Stefan Knapp, Shirit Einav, and Piet Herdewijn
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/jm501759m • Publication Date (Web): 30 Mar 2015
Downloaded from http://pubs.acs.org on April 1, 2015
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 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Selective inhibitors of Cyclin-G associated kinase (GAK) as anti-HCV agents
†
,‡,#
†,‡,#
ǁ
ǁ
Sona Kovackova , Lei Chang , Elena Bekerman , Gregory Neveu , Rina Barouch-
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
ǁ
§
§
†,‡
†,‡
Bentov , Apirat Chaikuad , Christina Heroven , Michal Šála , Steven De Jonghe , Stefan
§
ǁ
†,‡,
Knapp , Shirit Einav , Piet Herdewijn *
†
KU Leuven, Rega Institute for Medical Research, Laboratory of Medicinal Chemistry,
Minderbroedersstraat 10, 3000 Leuven, Belgium.
‡
KU Leuven, Interface Valorisation Platform, Kapucijnenvoer 33, 3000 Leuven, Belgium.
ǁ
Department of Medicine, Division of Infectious Diseases and Geographic Medicine, and
Department of Microbiology and Immunology, Stanford University School of Medicine,
Stanford, California, USA.
§
University of Oxford, Target Discovery Institute (TDI) and Structural Genomics Consortium
(
SGC), Old Road Campus Research Building, Oxford OX3 7DQ, UK.
#
Both authors contributed equally
* Phone: +32 16 337387. Fax: +32 16 337340. E-mail: Piet.Herdewijn@rega.kuleuven.be
Abstract
Cyclin-G associated kinase (GAK) emerged as a promising drug target for the treatment of
viral infections. However, no potent and selective GAK inhibitors have been reported in the
literature to date. This paper describes the discovery of isothiazolo[5,4-b]pyridines as
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 66
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
selective GAK inhibitors, with the most potent congeners displaying low nanomolar binding
affinity for GAK. Co-crystallization experiments revealed that these compounds behaved as
classic type I ATP-competitive kinase inhibitors. In addition, we have demonstrated that these
compounds exhibit a potent activity against hepatitis C virus (HCV) by inhibiting two
temporally distinct steps in the HCV lifecycle (i.e. viral entry and assembly). Hence, these
GAK inhibitors represent chemical probes to study GAK function in different disease areas
where GAK has been implicated (including viral infection, cancer and Parkinson’s disease).
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
Introduction
Cyclin G associated kinase (GAK) was first identified in experiments investigating proteins
1
associated with cyclin G, a protein involved in cell cycle regulation. GAK (also known as
auxillin 2) is a 160 kDa serine/threonine protein kinase that belongs to the numb-associated
kinase (NAK) family, which also includes STK16/MPSK1 (Serine/threonine kinase
16/myristoylated and palmitoylated serine/threonine kinase 1), AAK1 (adaptor-associated
2
kinase) and BIKE (BMP-2 inducible kinase).
GAK is expressed ubiquitously and bears a strong homology (43%) to the neuronal-specific
protein auxilin, a heat shock cognate 70 (Hsc70) cochaperone with a role in uncoating clathrin
vesicles. GAK is a key regulator of clathrin-mediated trafficking both in the endocytic and
secretory pathways. It recruits clathrin and clathrin adaptor protein complex 2 (AP-2) to the
3
plasma membrane and phosphorylates a T156 residue within AP2M1, the µ subunit of AP-2,
thereby stimulating its binding to cargo proteins and enhancing cargo recruitment, vesicle
3
,4,5,6
assembly and efficient internalization.
Moreover, GAK regulates endocytosis of receptors
3
that is mediated by alternate clathrin adaptors and is implicated in later steps of endocytosis,
including regulation of clathrin-coated vesicles (CCVs) uncoating, which enables recycling of
3
,5
clathrin back to the cell surface. GAK is an important regulator of Epidermal Growth Factor
ACS Paragon Plus Environment

Page 3 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
3
Receptor (EGFR); it is known to promote EGF uptake and may also function in receptor
7
signaling. Last, GAK also plays an important role in regulating clathrin-mediated sorting
3,5
events in the trans-Golgi network.
Interestingly, GAK-dependent phosphorylation of clathrin adaptor proteins has been
implicated in the regulation of viruses. AP2M1 was shown to be recruited to the surface 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
8
lipid droplets by the HCV capsid protein, core. The interaction between HCV core and
8
AP2M1 was shown to be critical for HCV assembly. Notably, either overexpression of an
AP2M1 phosphorylation-site mutant or suppression of GAK expression disrupted core-
8
AP2M1 binding and HCV assembly. More recently, GAK was shown to regulate HCV entry
1
1
independently of its effect on HCV assembly, in part by activating AP2M1. Hence, GAK
represents a cellular host factor essential for regulation of HCV entry and assembly and a
potential target for antiviral strategies. Indeed, erlotinib, an approved anticancer drug that
9
,10
potently inhibits GAK (in addition to its known cancer target, EGFR ) inhibits HCV entry
as well as core-AP2M1 binding, thereby also disrupting HCV assembly, but not HCV RNA
8
,11
replication .
To the best of our knowledge, no potent and selective GAK inhibitors have been reported in
the literature to date. Like erlotinib, other approved kinase inhibitors, such as dasatinib,
gefitinib, and pelitinib, display a high affinity for GAK with Kd values in the low nanomolar
12
range (Chart 1). Similarly, pyridinyl imidazoles, such as SB203580 and SB201290 that have
13
been developed as p38 inhibitors, potently inhibit GAK. Nevertheless, since all these
compounds were designed to target other kinases, their inhibitory effect on GAK represents
an off-target effect, and their use is limited by significant toxicities resulting from lack of
selectivity. Moreover, while several compounds that bind GAK with an excellent ligand
efficiency (LE) of 0.51 kcal/mol (Figure 1) were discovered by a fragment-based screening
14
using weak affinity chromatography, their binding affinity was low (Kd value of 2 µM).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 66
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
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
Chart 1 : Known GAK inhibitors
Because of the potential for GAK to serve as an antiviral drug target and the lack of selective
small-molecule GAK inhibitors, we embarked on the synthesis and biological evaluation of a
novel series of GAK inhibitors. In addition to their potential as lead molecules for the
development of a novel antiviral strategy, these compounds represent useful chemical probes
to further investigate the function of GAK in aspects of general cell biology and other disease
15
16
conditions, such as cancer and Parkinson’s disease, where GAK plays an important role.
Screening – Hit discovery
In order to discover novel GAK inhibitors, a drug-like compound library of 150 analogues
was screened to identify potential ligands of GAK. This compound collection was made as
part of a program to synthesize novel compound libraries based on original and patentable
chemistry, which has not been explored before in drug discovery, and hence, represent
unexplored chemical space. This library consists mainly of compounds based on a bicyclic,
ACS Paragon Plus Environment

Page 5 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
hetero-aromatic flat scaffold. The different scaffolds are unrelated to each other, and, in
addition, they differ in their substitution pattern (the nature, as well as the spatial orientation
of the substituents). These flat core structures are typical for kinase inhibitors, since they
function as a central scaffold that binds to the ATP-binding site of the enzyme, and is
therefore predicted to yield a high hit rate in kinase assays. We opted for the KINOMEscan™
screening platform that employs an active site-directed competition binding assay to
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
12
quantitatively measure interactions between a test compound and a given kinase.
Compounds that bind the kinase active site and either directly (sterically) or indirectly
allosterically) prevent kinase binding to the immobilized ligand reduce the amount of kinase
(
captured on the solid support. Conversely, test molecules that do not bind the kinase have no
effect on the amount of kinase captured on the solid support. Hits are identified by measuring
the amount of kinase captured in test versus control samples using a quantitative, precise and
sensitive qPCR method that detects the associated DNA label. In the first round of screening,
the compound library was tested at a single concentration of 10 µM. The results are reported
as the percentage of kinase/phage remaining bound to the ligands/beads, relative to a control.
High affinity compounds have % of control values close to zero, while weaker binders have
higher % control values. For the most promising compounds, dose-response curves were
generated to determine binding constant (Kd) values. This initial screening led to the
discovery of a promising hit (Figure 1), with a good potency (%Ctrl = 0.3 ; Kd = 0.3 µM),
moderate lipophilicity (clogP = 2.44) and a favorable ligand efficiency (LE) of 0.494
kcal/mol. Overall, these parameters qualified this hit compound as an optimal starting point
for the discovery of novel GAK inhibitors. Notably, this hit compound is based on an
isothiazolo[4,3-b]pyridine scaffold, a skeleton that is unexplored in organic chemistry. This is
in contrast to the bicyclic hetero-aromatics (such as purines, quinazolines, benzimidazoles and
indoles), which are considered ‘privileged structures’ in kinase drug discovery and are widely
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 66
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
described in scientific and patent literature, making novelty difficult to achieve. Indeed,
1
7
SciFinder searches revealed only 19 known isothiazolo[4,3-b]pyridine derivatives, and,
hence, further elaboration of this scaffold gives access to a largely unexplored chemical space.
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 1 : Hit compound
Chemistry
The general approach for the synthesis of the isothiazolo[4,3-b]pyridine derivatives is shown
in Scheme 1. The synthesis started from 3-nitro-5-bromopyridine-2-carbonitrile 1. Reduction
of the nitro group to the corresponding amino group was achieved by treatment with iron
18
under acidic conditions. The major compound was the desired compound 2, however, it was
always accompanied by formation of the corresponding carboxamide, resulting from acidic
hydrolysis of the cyano group. This mixture was used as such in the subsequent reaction. The
aromatic thioamide 3 was obtained by treatment of the mixture with phosphorus pentasulfide
19
as thionation reagent. An oxidative ring closure using hydrogen peroxide yielded the 3-
20
amino-6-bromo-isothiazolo[4,3-b]pyridine 4.
Subsequent palladium-catalyzed Suzuki
21
coupling with appropriate arylboronic acids afforded a series of 3-amino-6-aryl-
isothiazolo[4,3-b]pyridines 5a-e. Alternatively, diazotation of the exocyclic amino group with
sodium nitrite, hydrogen bromide and CuBr, furnished the 3,6-dibromo-isothiazolo[4,3-
2
0
b]pyridine 6. Treatment of 6 with sodium methoxide or nitrogen-containing nucleophiles
yielded a series of 3-substituted-6-bromo-isothiazolo[4,3-b]pyridine analogues 7. Finally,
reaction of 7 with a number of arylboronic acids then yielded target compounds 8-13. These
Suzuki reactions were either conducted in a mixture of dioxane/water, using sodium carbonate
ACS Paragon Plus Environment

Page 7 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
as base and Pd(dppf)Cl as a catalyst, or alternatively, in a mixture of dimethoxyethane/water
2
with potassium carbonate as a base and Pd(PPh ) as a catalyst.
3
4
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
Reagents and conditions : a) Fe, CH
, CH OH, rt ; d) R B(OH) , Na
HBr, NaNO ,H O, 0°C to rt ; f) R H, EtOH, 75°C; g) R
dioxane/water, 100 °C or R B(OH) , K CO , Pd(PPh ) , H O, DME.
3
COOH, 0°C to rt ; b) P
CO , Pd(dppf)Cl , dioxane/water, 100 °C ; e) CuBr,
2
B(OH) , Na CO , Pd(dppf)Cl ,
2 5
S , EtOH, 75°C; c) 30% aq.
H
2
O
2
3
6
2
2
3
2
2
2
3
6
2
2
3
6
2
2
3
3 4
2
Scheme 1 : Synthesis of isothiazolo[4,3-b]pyridine analogues
Structure-activity relationship studies
The isothiazolo[4,3-b]pyridine scaffold offers two sites for structural variation. Therefore, the
structure-activity relationship (SAR) study started with the synthesis of a small library matrix
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 66
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 isothiazolo[4,3-b]pyridines with structural variation at positions 3 (amino, methoxy,
ethanolamine, morpholine, and N-Me-piperazine) and 6 (substituted aryl and heteroaryl
groups). Similarly to our primary screening strategy, all compounds were tested at a 10 µM
concentration and the most promising ones were then subjected to determination of Kd
values. The matrix depicted in Table 1, demonstrates that compounds 9c and 9d, resulting
from the combination of a 3,4-dimethoxyphenyl or a 3-thienyl substituent at position 6 with
an ethanolamino or a morpholino moiety at position 3 have a strong affinity for GAK (Kd
values of 52 nM and 42 nM, respectively).
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
R
R
3
6
9a; %Ctrl:0,8;
5
a; %Ctrl:68
8a; %Ctrl:18
8b; %Ctrl:2.2
10a; %Ctrl:76 11a; %Ctrl:18
10b; %Ctrl:72 11b; %Ctrl:37
Kd = 0,5 µM
5b; %Ctrl:55
5c; %Ctrl:70
9b; %Ctrl:11
8c; %Ctrl:0.4;
9c; %Ctrl:0;
10c; % Ctrl:19 11c; %Ctrl:19
Kd = 0,12 µM
Kd= 0,052 µM
S
8d; % Ctrl 0,3;
9d; %Ctrl:0.1;
5
d; %Ctrl:61
10d; %Ctrl:7.4 11d; %Ctrl:22
10e; %Ctrl:88 11e; %Ctrl: 93
Kd = 0,3 µM
Kd= 0,042 µM
5e; %Ctrl:80
8e; % Ctrl:45
9e; %Ctrl:25
ACS Paragon Plus Environment

Page 9 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Table 1 : GAK affinity data of an isothiazolo[4,3-b]pyridine library. Values represent the
average of two independent experiments.
A dual strategy was then used for further optimization. To probe the optimal substitution
pattern at position 6, the morpholino substituent at position 3 was fixed, and a wide range of
substituted aryl or heteroaryl groups was evaluated. As it became clear that a 3,4-
dimethoxyphenyl or a 3-thienyl substituents were optimal for GAK binding, the SAR focused
on closely related derivatives of both of these aryl groups (Table 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
Replacing the 3-thienyl group of compound 9d by a 2-thienyl group (compound 12a) led to
equipotent compounds. On the other hand, the presence of a 2-furanyl moiety led to a 5-fold
decrease in GAK affinity, yielding compound 12b with a Kd value of 0.2 µM. The SAR of
the dimethoxyphenyl ring was investigated by the synthesis of a number of regio-isomeric
dimethoxyphenyl derivatives. Compared to the original 3,4-dimethoxyphenyl analogue 9c
(
Kd = 0.052 µM), the 2,4-dimethoxyphenyl analogue (compound 12c) lacks any affinity for
the GAK enzyme, the 2,5-dimethoxyphenyl derivative 12d shows intermediate affinity (Kd =
.13 µM), whereas the 3,5-dimethoxyphenyl derivative 12e demonstrates a similar GAK
0
affinity (Kd = 0.072 µM). The dioxolane analogue 12f (a ring-closed analogue of the 3,4-
dimethoxyphenyl moiety) is completely inactive. The insertion of an additional methoxy
group yielded the 3,4,5-trimethoxyphenyl analogue 12g, displaying very strong binding
affinity for the GAK enzyme (Kd = 0.0083 µM). In another round of SAR, the 3-methoxy
group was kept intact and the 4-methoxy moiety was replaced by an ester moiety (compound
1
2h), an amino group (compound 12i) and a hydroxyl function (compound 12j), generating
compounds with a strong affinity for the GAK enzyme, displaying Kd values of 0.018 µM,
0.0089 µM and 0.018 µM, respectively.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 66
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
a
Cpmd #
12a
Structure
% Ctrl @10 µM
Kd (µM)
0.05
0.047
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
12b
0.15
0.2
12c
6.3
0.7
ND
12d
0.13
1
2e
2f
0.2
0.072
1
51
0
ND
12g
0.0083
12h
0
0
0.018
12i
0.0089
ACS Paragon Plus Environment

Page 11 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
12j
0
0.018
a
Table 2 : SAR of the aryl moiety. Values represent the average of two independent
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
experiments. ND = not determined.
Since the first library screening demonstrated that a 3,4-dimethoxyphenyl moiety at position 6
was favorable for GAK binding, this substituent was kept intact, and a variety of amines were
introduced at position 3 (Table 3). Although a free amino group at position 3 of the
isothiazolo[5,4-b]pyridine scaffold is not tolerated for GAK binding (Table 1, compound 5c),
the presence of small aliphatic amines (primary as well as secondary), such as dimethylamino
(compound 13a), methoxyethylamino (compound 13b), diethanolamino (13c), gives rise to
compounds displaying Kd values in the range of 0.14-0.3 µM. Cyclo-aliphatic amines, such
as a pyrrolidino (compound 13d) and a cyclopropylmethylamino (compound 13e) yielded
compounds with a higher potency, with Kd values of 100 nM or 27 nM, for compounds 13d
and 13e, respectively.
The isosteric replacement of the oxygen of the morpholine moiety by a carbon (resulting in
compound 13f) or a sulfur (resulting in compound 13g) leads to a 5-fold drop in GAK
affinity, when compared to the corresponding morpholine analogue 9c. To better mimic the
morpholine substituent, a 4-hydroxy-piperidino and a 4-amino-pyrane moiety were
introduced, giving rise to isothiazolo[4,3-b]pyridines 13i and 13j, both exhibiting a strong
affinity for the GAK enzyme (Kd values of 63 nM and 88 nM, respectively). Overall, the data
in Table 3 indicate quite high tolerance for structural variation at position 3, with only an
aromatic phenethylamino group (compound 13k) completely lacking GAK affinity.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 66
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
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
Cmpd#
R
3
% Ctrl (10 µM)
Kd (µM)
13a
0.1
0.19
13b
0.1
4
0.14
0.32
13c
1
3d
3e
0
0
0.1
N
1
0.027
13f
2.6
0.8
0.1
0.27
0.23
0.11
13g
13h
ACS Paragon Plus Environment

Page 13 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
13i
0.1
0.063
0.088
ND
13j
0.1
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
13k
81
a
Table 3 : SAR of the amine moiety. Values represent the average of two independent
experiments. ND = not determined
Although improving binding affinity is important in an optimization campaign, an
overemphasis on potency has often resulted in molecules with unsuitable physicochemical
properties for further development for in vivo use. We therefore complemented our lead
optimization by monitoring a number of in silico parameters (Table 4). The optimized
compounds had a low-molecular weight, which in combination with their potent binding
affinity, yielded very ‘efficient’ LE values exceeding 0.4 kcal/mol. Moreover, the
optimization campaign did not result in any gain in lipophilicity, as the most potent GAK
ligands, 12g and 12i, displayed cLogP values of 2.49 and 2.28, respectively.
Cmpd # MW
Kd (µM)
LE (kcal/mol) cLogP
6d (Hit) 277.37 0.3
0.494
0.408
0.458
2.44
2.49
2.28
1
1
2g
2i
387.45 0.0083
342.42 0.0089
Table 4 : In silico parameters of hit and optimized compounds
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 66
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
Selectivity profiling
Most kinase inhibitors are ATP-competitive and bind to the ATP domain, which is highly
conserved across the kinome. Selectivity screens usually involve close analogues of the target
kinase harboring similar ATP binding sites. However, we opted for a kinome-wide selectivity
screen, using a panel of 456 kinases available at DiscoverX. As a representative example,
12g, the most potent congener, was selected for the selectivity profiling at a single
concentration of 10 µM. Remarkably, no affinity of 12g was observed for any of the other
members of the NAK family (AAK1 : 90% ; STK16 : 92% ; BIKE : 92% compared to the
DMSO control) (Figure 2). Moreover, only 7 additional kinases appeared to interact with 12g,
with a binding activity of less than 10%, when compared to DMSO control. The exact Kd
values for these kinases were determined (Table 5). This large scale kinase profiling confirms
that the identified lead compound is selective. Compound 12g displays the strongest affinity
for GAK (Kd = 8.3 nM). A 3-fold level of selectivity is observed when compared with the
affinity for KIT (Kd = 29 nM), whereas for the other kinases (CLK2, CSF1R, FLT3, MEK5,
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
PDGFRA and PDGFRB), at least a 8-fold level of selectivity is achieved.
a
Kinase
GAK
Kd (nM)
8.3
CLK2
CSF1R
FLT3
KIT
710
320
110
29
MEK5
150
PDGFRA 220
PDGFRB 70
ACS Paragon Plus Environment

Page 15 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Table 5 : Kd values for off-target kinases affected by 12g. Values represent the average of
two independent experiments.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2 : Specificity profile of compound 12g
X-ray crystallography
To shed a light on the binding mode of the compounds to GAK, co-crystallization
experiments were initiated. We have recently solved the structure of GAK in complex with a
single chain antibody (nanobody), which resulted in reproducible crystallization conditions
22
for the GAK kinase domain. The nanobody bound distal to the ATP binding site interacting
with the lower kinase lobe (Figure 3). The lead compound 12i was co-crystallized and bound
as expected to the ATP binding site of GAK. The binding mode of the inhibitor was well
defined by electron density and the structure was refined to 2.1 Å resolution. Data collection
and refinement statistics are provided in supplemental Table 2 (see Supporting Information).
Two kinase domain:nanobody complexes were present in the asymmetric unit that interacted
22
via the extended activation loops of GAK as described previously. The upper lobe of the
kinase domain was flexible as indicated by high B-factors in particular in the N-terminal
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 66
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
region (Supplemental Figure 1). The helix αC and the DFG motive were in an active
conformation compatible with the type I binding mode of the inhibitor (Figure 3). The
nitrogen of the isothiazolo moiety of the inhibitor formed a hydrogen bond with the backbone
amide of Cys126. Interactions with small gatekeepers, which are found in only a small subset
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
23
of kinases, have been explored for the development of selective kinase inhibitors. Therefore,
our observation of this contact in the case of 12i may explain its excellent selectivity for
GAK. The 4-amino-3-methoxy-phenyl substituent of 12i is oriented towards the solvent
exposed area forming a number of hydrophobic contacts, in particular with Leu46 and a long
range polar interaction with Arg44. Overall, 12i exhibits a good shape complementarity with
the GAK ATP binding site.
Figure 3 : Structure of GAK in complex with 12i.A) Overview of GAK:NbGAK_4 complex
with inhibitor 12i bound. The inset shows │F
│-│F │omitted map contoured at 3σ for the
O
C
bound inhibitor. B) Detailed interactions of 12i within the GAK ATP binding site.
Antiviral activity
We have recently shown that GAK is a regulator of HCV entry and assembly and therefore
represents a potential target for anti-HCV treatment. Moreover, erlotinib, an approved
ACS Paragon Plus Environment

Page 17 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
anticancer drug known to target GAK significantly inhibited binding of HCV core to AP2M1,
8
,11
HCV entry, and assembly.
Nevertheless, while erlotinib binds GAK with a high affinity
(
Kd = 3.4 nM), it binds EGFR, its primary anticancer target, with a comparable affinity (Kd =
9
,10
1
nM), and several other kinases (albeit at a lower affinity). The use of erlotinib as a
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
chemical tool to probe the role of GAK in HCV infection is therefore somewhat limited,
particularly, since EGFR has also been recognized as an essential host factor for HCV
24
infection. Moreover, the limited selectivity contributes to erlotinib’s side effects, which may
reduce its potential as an antiviral agent, particularly in long duration regimens as those
required for the treatment of HCV. We therefore hypothesized that since the isothiazolo[4,3-
b]pyridines 12g and 12i are structurally unrelated to erlotinib and lack anti-EGFR activity,
they represent attractive chemical tools to study the role of GAK in HCV infection, and that
their strong potency, promising selectivity profile and favorable drug-like properties make
them a potential novel class of antiviral agents.
To further validate GAK as an antiviral target and determine the antiviral effect of these
compounds on HCV infection, Huh-7.5 (human hepatoma) cells were infected with cell
culture grown J6/JFH(p7-Rluc2A) HCV (HCVcc), a Renilla luciferase-containing reporter
25
virus that replicates and produces high viral titers in Huh-7.5 cells. Infected cells were
treated with various concentrations of 12g, 12i, or DMSO. Drug-containing media was
replenished every 24 hours. Antiviral activity and cellular viability were measured by
luciferase and alamarBlue-based assays, respectively, 72 hours postinfection. As shown in
Figure 4A, treatment with either 12g or 12i resulted in a dose-dependent inhibition of viral
replication. Half maximal effective concentration (EC ) values were 2.55±0.43µM
50
(
p=0.0002) and 2.81±0.8µM (p=0.009) for 12g and 12i, respectively. The half maximal
cytotoxic concentration (CC50) values were 23.27±3.4 µM (p=0.000012) for 12g and
.92±1.43 µM (p= 0.000044) for 12i.
8
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 66
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
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 4: Selective GAK inhibitors inhibit HCV infection and AP2M1 phosphorylation. Dose
response curves of 12g and 12i effects on infection of Huh-7.5 cells with cell culture grown
HCV (HCVcc) (A), entry of pseudoparticles of HCV (HCVpp) (B), HCV RNA replication 72
hours postelectroporation with J6/JFH(p7-Rluc2A) RNA (C), and intra- (D) and extra-cellular
(E) infectivity in naive cells inoculated with cell lysates or supernatants derived from the
electroporated cells, respectively. Plotted in black (left y-axes) are relative luciferase values
normalized to DMSO treated controls. Corresponding cellular viability, as measured by
alamarBlue-based assays, are plotted in grey (right y-axes). Data reflect means and s.d. (error
ACS Paragon Plus Environment

Page 19 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
bars). F. The effect of the inhibitors on AP2M1 phosphorylation by Western analysis in
lysates derived from Huh-7.5 cells. A representative membrane blotted with anti-phospho-
AP2M1 (p-AP2M1) and anti-AP2M1 antibodies is shown. Numbers represent relative p-
AP2M1/total AP2M1 protein ratio normalized to DMSO controls.
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
Next, we sought to pinpoint the step of the viral lifecycle that is disrupted by these GAK
inhibitors. To study the effect of the GAK inhibitors on HCV entry, we measured the entry of
pseudoparticles of HCV (HCVpp) (lentiviral vectors that incorporate the HCV glycoproteins
2
6
on the viral envelope ) upon a 4 hour treatment with various concentrations of 12g and 12i
using luciferase reporter-based assays. Huh-7.5 cells were infected with HCVpp for 1 hour on
ice followed by a temperature shift to 37°C, 4 hour treatment with various concentrations of
the compounds or DMSO, and medium replacement for removal of residual drugs and
unbound virus. These compounds inhibited HCVpp entry in a dose-dependent manner, with
half maximal effective concentrations (EC50’s) of 3.6±1.2µM (p=0.03) and 2.05±0.36µM
(p=0.0025), respectively, and a minimal effect on cellular viability (Figure 4B).
To determine the effect of the compounds on later steps of the HCV lifecycle and distinguish
between a defect in viral RNA replication, assembly and/or release, Huh-7.5 cells were
24
electroporated with in vitro transcribed luciferase reporter J6/JFH(p7-Rluc2A) HCV RNA
and treated daily for 72 hours with either 12g, 12i or DMSO. HCV RNA replication was
measured by luciferase assays 72 hours postelectroporation and intra- and extra-cellular
infectivity were measured by luciferase assays in naïve cells inoculated with either clarified
cell lysates or supernatants derived from the electroporated cells, respectively. Treatment with
12g and 12i did not affect HCV RNA replication (Figure 4C). Nevertheless, 12g and 12i
demonstrated a dose-dependent effect on intracellular infectivity (Figure 4D), with EC50
values of 1.64±0.272 µM (p=0.009) and 2.43±0.72 µM (p=0.0281), respectively. Moreover,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 66
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
12g and 12i significantly inhibited extracellular infectivity (Figure 4E), with EC values of
50
2.13±0.72 µM (p=0.04) and 2.47±0.9 µM (p=0.04), respectively. These data indicate that the
two selective GAK inhibitors tested disrupt HCV assembly and infectious virus production
without affecting HCV RNA replication.
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
8
,11
Similar to small interfering RNAs (siRNAs) that target GAK as well as erlotinib,
these
compounds thus inhibit two temporally distinct steps in the HCV lifecycle: entry and
assembly. The EC values for the antiviral effect of these compounds range from ~1.5-3 µM,
50
depending on the assay used. These EC50 values are slightly higher or comparable to those of
8
,11
erlotinib (0.5-1.5 µM)
likely reflecting the absence of EGFR inhibition exhibited by
erlotinib.
Taken together, these results provide a pharmacological validation of the requirement for
GAK in the regulation of HCV entry and assembly. Moreover, the selective GAK inhibitors
represent candidate compounds to target two steps of the HCV lifecycle: viral entry and
assembly (Figure 5).
To validate the mechanism of action of these compounds, we studied their effect on AP2M1
phosphorylation. Huh-7.5 cells were treated with various concentrations of the compounds or
DMSO. Since AP2M1 phosphorylation is transient (due to the activity of the phosphatase
2
7
PP2A) , to allow capturing of the phosphorylated AP2M1 state, these cells were incubated for
1 hour in the presence of the PP2A inhibitor, calyculin A, prior to lysis. The ratio of
phosphorylated AP2M1 (p-AP2M1) to total AP2M1 was measured by quantitating Western
blot band intensity. p-AP2M1 to AP2M1 ratios were reduced by either 12g or 12i in a dose-
dependent manner and were significantly lower than the ratios measured in the DMSO control
(Fig. 4F). These results indicate that 12g and 12i modulate AP2M1 phosphorylation, as
predicted for compounds that target GAK.
ACS Paragon Plus Environment

Page 21 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5
The proposed model for the mechanism by which the selective GAK inhibitors, 12g and 12i
inhibit HCV infection. By directly inhibiting GAK kinase activity, 12g and 12i interfere with
clathrin/AP-2-mediated endocytosis of HCV (step 1) as well as AP-2 binding to core in HCV
assembly (step 5).
Conclusion
We have previously reported that GAK is an essential host factor for HCV entry and assembly
and that erlotinib, an approved anticancer drug with inhibitory activity against GAK,
8,11
effectively disrupts HCV entry and assembly.
Here, for the first time, we report on a
specific optimization campaign towards the identification of novel GAK inhibitors. This
endeavor resulted in the identification of novel and highly selective GAK inhibitors, based on
an isothiazolo[4,3-b]pyridine scaffold, an unexplored chemical space. Their drug-like
properties, as well as their selectivity profile, make these compounds an attractive tool to
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 66
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
study regulatory mechanisms of intracellular traffic in cell biology and disease conditions,
where GAK is known to be implicated. We used these compounds to better understand the
role of GAK in HCV infection and further validate this host factor as an antiviral target. We
have demonstrated that compounds 12g and 12i exhibit a potent in vitro anti-HCV activity
and that their antiviral effect correlates with a decrease in AP2M1 phosphorylation. Although
to the best of our knowledge, KIT has not been reported to have a role in HCV infection, it is
bound by these compounds, albeit at a lower affinity (~30nM). We therefore cannot exclude
the possibility that inhibition of KIT or other cellular kinases, in addition to GAK, contributes
to the antiviral effect of these compounds.
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
Unlike approved direct acting antivirals (DAA), which target HCV RNA replication, these
compounds inhibit two distinct steps of the HCV lifecycle: entry and assembly. By displaying
limited off-target binding to other kinases, these more selective GAK inhibitors have the
potential to reduce drug toxicity. Moreover, inhibiting a host rather than a viral target is more
likely to provide a higher genetic barrier for drug resistance than most DAAs and a
comprehensive protection against all HCV genotypes. Hence, selective GAK inhibitors may
potentially represent a novel class of antivirals for inclusion in future combination drug
regimens for treating HCV. Last, since multiple other viruses hijack clathrin adaptor proteins-
28,29,30
mediated pathways either for viral entry or late steps of their viral lifecycle,
we predict
that the requirement for GAK is broadly conserved among RNA viruses, and therefore that
selective GAK inhibitors may display a broad-spectrum antiviral activity.
Acknowledgments
This research was supported by a grant from the IWT (Agentschap voor Innovatie door
Wetenschap en Technologie-Vlaanderen ; Grant IWT-SBO 100014 ; to P.H). Mass
ACS Paragon Plus Environment

Page 23 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
spectrometry was made possible by the support of the Hercules Foundation of the Flemish
Government (grant 20100225–7).
This work was supported in part by grants from Stanford Clinical and Translational Science
Award (CTSA) to Spectrum (UL1 TR001085), Stanford SPARK program, Stanford Bio-X,
Stanford Translational Research and Applied Medicine Program, grant 2013100 from the
Doris Duke Charitable Foundation, and grant RSG-14-11 0-0 1-MPC from the American
Cancer Society to S.E.
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
G.N. is supported by the Child Health Research Institute, Luclie Packard Foundation for
Children’s Health, as well as the Stanford CSTA (grant number UL1 TR000093). We thank
Dr. Charles M. Rice for kindly providing us the pFL-J6/JFH, pFL-J6/JFH(p7-Rluc2A)
plasmid, Apath LLC (St. Louis, MO) for providing Huh-7.5 cells, and Dr. Shoshana Levy for
providing the plasmids used in the HCVpp entry assays (pNL4-3.Luc.R-E, pcDM8 and
pcDM8-E1E2).
SK is supported by the SGC, a registered charity (number 1097737) that receives funds from
AbbVie, Bayer, Boehringer Ingelheim, the Canada Foundation for Innovation, the Canadian
Institutes for Health Research, Genome Canada, GlaxoSmithKline, Janssen, Lilly Canada, the
Novartis Research Foundation, the Ontario Ministry of Economic Development and
Innovation, Pfizer, Takeda, and the Wellcome Trust [092809/Z/10/Z]. AC is supported by the
European Union FP7 Grant No. 278568 “PRIMES” (Protein interaction machines in
oncogenic EGF receptor signalling).
The funders had no role in study design, data collection and analysis, decision to publish, or
preparation of the manuscript.
Experimental section
General
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 66
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 all reactions, analytical grade solvents were used. All moisture-sensitive reactions were
1
13
carried out in oven-dried glass-ware (135 °C). H and C NMR spectra: Bruker Avance 300
1
13
1
13
Mhz ( H-NMR: 300 MHz, C-NMR: 75 MHz), 500 MHz ( H-NMR: 500 MHz, C-NMR:
1
13
1
25 MHz) and 600 MHz ( H-NMR: 600 MHz, C-NMR: 150 MHz) using tetramethylsilane
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
as internal standard for H-NMR spectra and (D
6
)-DMSO (39.5 ppm) or CDCl
3
(77.2 ppm)
1
3
and CD OD (49.0 ppm) for C-NMR spectra. Abbreviations used are: s = singlet, d =
3
doublet, t = triplet, q = quartet, m = multiplet, br = broad. Coupling constants are expressed in
Hz. Mass spectra are obtained with a Finnigan LCQ Advantage Max (ion trap) mass
spectrophotometer from Thermo Finnigan, San Jose, CA, USA. High resolution mass
spectrometry spectra were acquired on a quadrupole orthogonal acceleration time-of-flight
mass spectrometer (Synapt G2 HDMS, Waters, Milford, MA). Samples were infused at 3
µL/min and spectra were obtained in positive (or negative) ionization mode with a resolution
of 15000 (FWHM) using leucine enkephalin as lock mass. Precoated aluminum sheets (Fluka
Silica gel/TLC-cards, 254 nm) were used for TLC. Column chromatography (CC) was
performed on ICN silica gel 63-200, 60 Å. Purity of final compounds was determined by
analytical RP-HPLC analysis on a XBridge column (C-18, 5 ꢀm, 4.6 mm × 150 mm) in
combination with a Waters 600 HPLC system and a Waters 2996 photodiode array detector
from Waters, Milford, Massachusetts, USA. Elution was done using a gradient mixture of
H
2
O containing 0.2% (vol) of TFA (A) and Acetonitrile (B) (Supporting information). All
compounds were at least 95% pure, with the exception of compound 11b (90.32% purity).
3-Amino-5-bromopyridine-2-carbonitrile (2)
A solution of iron powder (3.36 g, 60 mmol) in acetic acid (15 ml) was stirred at 0 °C. To this
solution was added dropwise a solution of 3-nitro-5-bromopyridine-2-carbonitrile (2.51 g, 11
mmol) In acetic acid (15 ml). The reaction mixture was stirred at room temperature for two
ACS Paragon Plus Environment

Page 25 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
hours. Then, ethyl acetate (300 ml) was added and the mixture was filtered (paper filter). The
filter cake was washed with ethylacetate. The filtrate was evaporated and partitioned between
ethylacetate (500 ml) and water (250 ml). The organic phase was washed with a 1 N NaOH
solution (ca. 200 ml). The combined organic phases were dried and evaporated in vacuo,
yielding a mixture of two compounds, i.e. 3-amino-5-bromopyridine-2-carbonitrile (major
compound) and 3-amino-5-bromopyridine-2-carboxamide (minor compound). This mixture
was used as such in the next reaction.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3-Amino-5-bromo-2-pyridinecarbothioamide (3)
To a solution of a mixture of 3-amino-5-bromopyridine-2-carbonitrile and 3-amino-5-
bromopyridine-2-carboxamide (crude) in ethanol (25 ml) was added phosphorus pentasulfide
(2 eq; 4.84 g). The mixture was heated overnight at 75°C. The solvents were evaporated and
the crude residue was purified by flash chromatography on silica, using a mixture of
cyclohexane/ethyl acetate (in a ratio of 7:1) as mobile phase, yielding the title compound
1
(3.36 gram crude). H-NMR (300 MHz, CDCl
3
):
δ
= 6.95 (bs, 2H, NH
2
), 7.27 (d, J = 1.95 Hz,
1
H, arom H), 7.89 (d, J = 1.95 Hz, 1H, arom H), 9.37 (bs, 2H, NH ) ppm.
2
3-amino-6-bromo-isothiazolo[4,3-b]pyridine (4)
To a solution of 3-amino-5-bromo-2-pyridinecarbothioamide in methanol (50 ml) was added
dropwise a 30% H solution in water (3.5 ml) at 0°C. The reaction mixture was stirred
2
O
2
overnight at room temperature and then cooled again to 0°C. The crystals were filtered off
1
and washed with cold methanol yielding the title compound (1.6 g, 63%). H-NMR (300
MHz, DMSO-d
Hz, 1H, arom H) ppm. C-NMR (75 MHz, DMSO-d
(C ), 143.69 (C ), 153.58 (CH), 173.34 (C ) ppm.
6
):
δ
= 8.03 (d, J = 2.01 Hz, 1H, arom H), 8.08 (bs, 2H, NH
2
), 8.29 (d, J = 2.01
13
6
): δ = 120.56 (C ), 129.39 (CH), 133.23
q
q
q
q
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 66
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
Synthesis of 3-amino-6-aryl-isothiazolo[4,3-b]pyridines (5a-e)
General procedure
To a solution of 3-amino-6-bromo-isothiazolo[4,3-b]pyridine (166 mg, 0.72 mmol) in 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
mixture of dioxane (10 ml) and water (1.5 ml) was added an appropriate boronic acid (2 eq),
sodium carbonate (2 eq, 153 mg) and Pd(dppf)Cl (0.1 eq, 59 mg) The reaction mixture was
2
stirred overnight at 100 °C. The reaction mixture was cooled down to room temperature and
the reaction was partioned between ethylacetate (60 ml) and brine (30 ml). The aqueous phase
was then extracted with ethyl acetate (40 ml). The combined organic phases were dried over
4
MgSO and evaporated in vacuo. The residue was purified by flash chromatography on
silicagel yielding the pure title compounds.
The following compounds were made according to this procedure:
3-Amino-6-phenyl-isothiazolo[4,3-b]pyridine (5a)
This compound was obtained using phenylboronic acid and the crude residue was purified by
flash chromatography using a mixture of cyclohexane and ethylacetate (in a ratio of 4:1) as
1
mobile phase, affording the title compound in 77% yield (125 mg, 0.55 mmol). H-NMR (300
3 2
MHz, CDCl ): δ = 5.84 (s, 2H, NH ), 7.52 (m, 3H, arom H), 7.66 (m, 2H, arom H), 7.95 (d, J
13
= 1.98 Hz, 1H, arom H), 8.68 (d, J = 1.95 Hz, 1H, arom H) ppm. C-NMR (150 MHz,
DMSO-d
6
):
δ
= 124.37 (CH), 127.38 (CH), 128.40 (CH), 129.18 (CH), 134.06 (C
q
), 135.37
+
(
C ), 137.25 (C ), 143.27 (CH), 153.67 (C ), 172.38 (C ) ppm. HR-MS [M+H] found
q
q
q
q
228.0594, calculated 228.0551.
3-Amino-6-(4-fluorophenyl)-isothiazolo[4,3-
b
]pyridine (5b)
ACS Paragon Plus Environment

Page 27 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
This compound was obtained using 4-fluorophenylboronic acid and the crude residue was
purified by flash chromatography using a mixture of cyclohexane and ethylacetate (in a ratio
of 4:1) as mobile phase, affording the title compound in 76% yield (134 mg, 0.54 mmol).
1
H-NMR (300 MHz, DMSO-d
.1 Hz, 1H, arom H), 7.71 (m, 2H, arom H), 8.17 (d, J = 1.92 Hz, 1H, arom H) ppm. C-
NMR (150 MHz, DMSO-d6): = 115.94 (CH), 116.23 (CH), 124.43 (CH), 129.53 (CH),
29.65 (CH), 133.76 (C ), 134.43 (C ), 143.19 (CH), 153.65 (C ), 160.88 (C ), 164.24 (C ),
6 2
): δ = 6.36 (br s, 2H, NH ), 7.34 (m, 2H, arom H), 7.38 (d, J =
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
13
3
δ
1
1
q
q
q
q
q
+
q
72.49 (C ) ppm. HR-MS [M+H] found 246.0497, calculated 243.0457.
3
-Amino-6-(3,4-dimethoxyphenyl)-isothiazolo[4,3-b]pyridine (5c)
This compound was obtained using 3,4-dimethoxyphenylboronic acid and the crude residue
was purified by flash chromatography using a mixture of methanol and dichloromethane (in a
ratio of 1:40) as mobile phase, affording the title compound in 57% yield (117 mg, 0.41
1
mmol). H-NMR (300 MHz, DMSO-d
6
):
δ
= 3.81 (s, 3H, OCH
3
), 3.88 (s, 3H, OCH
), 7.89 (d, J = 1.92 Hz,
H, arom H), 8.63 (d, J = 1.95 Hz, 1H, arom H) ppm. C-NMR (150 MHz, DMSO-d ):
3
), 7.07 (d,
J = 8.89 Hz, 1H, arom H), 7.35 (m, 2H, arom H), 7.84 (br s, 2H, NH
2
13
1
5
δ
=
6
5.76 (CH
3
), 55.82 (CH
3
), 111.00 (CH), 112.34 (CH), 119.76 (CH), 123.56 (CH), 129.82
), 143.53 (CH), 149.38 (2 x C ), 153.96 (C ), 172.20 (C ) ppm.
HR-MS [M+H] found 288.0800, calculated 288.0762.
(C
q
), 133.77 (C
q
), 135.31 (C
q
q
q
q
+
3-Amino-6-(3-thienyl)-isothiazolo[4,3-b]pyridine (5d)
This compound was obtained using 3-thienylboronic acid and the crude residue was purified
by flash chromatography using a mixture of cyclohexane and ethylacetate (in a ratio of 3:2) as
1
mobile phase, affording the title compound in 69% yield (115 mg, 0.49 mmol). H-NMR (300
MHz, DMSO-d
6
):
δ
2
= 7.70 (m, 2H, arom H), 7.84 (br s, 2H, NH ), 7.98 (d, J = 1.89 Hz, 1H,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 66
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
1
3
arom H), 8.17 (q, J = 1.32 Hz, 1H, arom H), 8.74 (d, J = 1.95 Hz, 1H, arom H) ppm. C-
NMR (150 MHz, DMSO-d6): = 122.99 (CH), 123.20 (CH), 126.60 (CH), 127.65 (CH),
), 138.38 (C ), 143.17 (CH), 153.91 (C ), 172.22 (C ) ppm. HR-MS
δ
130.35 (C
q
), 133.71 (C
q
q
q
q
+
[
M+H] found 243.0157, calculated 243.0115.
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-amino-6-(3-pyridyl)-isothiazolo[4,3-b]pyridine (5e)
This compound was obtained using 3-pyridylboronic acid and the crude residue was purified
by flash chromatography using a mixture of methanol and dichloromethane (in a ratio of 1:20)
1
as mobile phase, affording the title compound in 75% yield (123 mg, 0.54 mmol). H-NMR
(300 MHz, DMSO-d
6 2
): δ = 7.52 (m, 1H, arom H), 7.95 (br s, 2H, NH ), 8.02 (d, J = 1.98 Hz,
1
H, arom H), 8.24 (m, 1H, arom H), 8.66 (m, 2H, arom H), 9.03 (d, J = 1.98 Hz, 1H, arom H)
13
ppm. C-NMR (150 MHz, DMSO-d ):
δ = 124.06 (CH), 125.16 (CH), 132.54 (C ), 133.01
q
6
q q q
(C ), 134.92 (CH), 142.85 (CH), 148.26 (CH), 149.37 (CH), 153.46 (C ), 173.63 (C ) ppm.
+
HR-MS [M+H] found 229.0537, calculated 229.0503.
3,6-di-Bromo-isothiazolo[4,3-b]pyridine (6)
A solution of 3-amino-6-bromo-isothiazolo[4,3-b]pyridine (1.1 g, 4.78 mmol) in HBr (50 ml)
was stirred for 10 minutes at room temperature and then CuBr was added in one portion (1.37
g, 9.56 mmol, 2.0 eq). The resulting mixture was cooled to 0°C. A solution of sodium nitrite
(
2
0.99 g, 14.34 mmol, 3.0 eq) in H O (15 ml) was added dropwise to the mixture over a period
of 30 minutes. The reaction mixture was stirred for 2 hours at 0°C and then overnight at room
temperature. Then, the mixture was cooled to 0°C and carefully neutralized with solid
potassium carbonate and extracted with ethyl acetate (2 x 50 ml). The combined organic
4
phases were dried over MgSO and evaporated in vacuo. The crude residue was purified by
silicagel flash chromatography, using the mixture of cyclohexane and ethylacetate (in a ratio
of 95:5) as mobile phase. The title compound was obtained in 78% yield (1.1 g, 3.75 mmol).
ACS Paragon Plus Environment

Page 29 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
H-NMR (300 MHz, DMSO-d6):
δ
= 8.70 (d, J = 2.01 Hz, 1H, arom H), 8.92 (d, J = 2.04 Hz,
H, arom H) ppm. C-NMR (75 MHz, DMSO-d6): = 121.12 (C ), 131.29 (CH), 137.79
), 144.54 (C ), 153.31 (CH), 154.58 (C ) ppm.
13
1
δ
q
(C
q
q
q
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
6
-Bromo-3-substituted-isothiazolo[4,3-b]pyridines (7a-c)
General procedure
To a solution of 3,6-di-bromo-isothiazolo[4,3-b]pyridine (150 mg, 0.51 mmol) in ethanol (10
ml) was added an appropriate nitrogen nucleophile (3 eq). The reaction was stirred overnight
at 75°C. The solvent was evaporated in vacuo and the crude residue was purified by silicagel
flash chromatography, the mobile phase being a mixture of cyclohexane and ethylacetate (in a
ratio of 4:1), yielding the pure title compounds.
The following compounds were made according to this procedure :
6-Bromo-3-morpholino-isothiazolo[4,3-b]pyridine (7a)
This compound was made using morpholine as nucleophile and the crude residue was purified
using a mixture of cyclohexane and ethylacetate (in a ratio of 4:1) as mobile phase, affording
1
the title compound in 92% yield (140 mg, 0.46 mmol). H-NMR (300 MHz, DMSO-d ):
δ =
6
3.84 (br s, 4H, 2 x NCH
2 2
), 3.87 (br s, 4 H, 2 x OCH ), 8.15 (s, 1H, arom H), 8.37 (s, 1H, arom
H) ppm.
6-Bromo-6-ethanolamino-isothiazolo[4,3-b]pyridine (7b)
This compound was made using ethanolamine as nucleophile and the crude residue was
purified using a mixture of cyclohexane and ethylacetate (in a ratio gradually ranging from
1:1 to 1:4) as mobile phase, affording the title compound in 94% yield (131 mg, 0.47 mmol).
ACS Paragon Plus Environment