Discovery of Fragment-Derived Small Molecules for in Vivo Inhibition of Ketohexokinase (KHK)

Article abstract of DOI:10.1021/acs.jmedchem.7b00947

Increased fructose consumption and its subsequent metabolism have been implicated in hepatic steatosis, dyslipidemia, obesity, and insulin resistance in humans. Since ketohexokinase (KHK) is the principal enzyme responsible for fructose metabolism, identification of a selective KHK inhibitor may help to further elucidate the effect of KHK inhibition on these metabolic disorders. Until now, studies on KHK inhibition with small molecules have been limited due to the lack of viable in vivo pharmacological tools. Herein we report the discovery of 12, a selective KHK inhibitor with potency and properties suitable for evaluating KHK inhibition in rat models. Key structural features interacting with KHK were discovered through fragment-based screening and subsequent optimization using structure-based drug design, and parallel medicinal chemistry led to the identification of pyridine 12.

Full text of DOI:10.1021/acs.jmedchem.7b00947

Subscriber access provided by UNIVERSITY OF ADELAIDE LIBRARIES
Article
Discovery of Fragment-Derived Small Molecules
for In Vivo Inhibition of Ketohexokinase (KHK)
Kim Huard, Kay Ahn, Paul A. Amor, David A. Beebe, Kris A Borzilleri, Boris A. Chrunyk, Steven
B. Coffey, Yang Cong, Edward L. Conn, Jeffrey S. Culp, Matthew S. Dowling, Matthew F.
Gorgoglione, Jemy A. Gutierrez, John D. Knafels, Erik A. Lachapelle, Jayvardhan Pandit, Kevin
D. Parris, Sylvie Perez, Jeffrey A. Pfefferkorn, David A. Price, Brian Raymer, Trenton T. Ross,
Andre Shavnya, Aaron C. Smith, Timothy A. Subashi, Gregory Tesz, Benjamin A. Thuma,
Meihua Tu, John D. Weaver, Yan Weng, Jane M. Withka, Gang Xing, and Thomas V Magee
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b00947 • Publication Date (Web): 30 Aug 2017
Downloaded from http://pubs.acs.org on August 30, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 52
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
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 52
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
Discovery of FragmentꢀDerived Small Molecules for
In Vivo Inhibition of Ketohexokinase (KHK)
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
◊*
┴†
┴
┴
║
║
Kim Huard, Kay Ahn, Paul Amor, David A. Beebe, Kris A. Borzilleri, Boris A. Chrunyk,
‡
‡
‡
‡
‡
Steven B. Coffey, Yang Cong, Edward L. Conn, Jeffrey S. Culp, Matthew S. Dowling,
┴
┴
║
‡
Matthew F. Gorgoglione, Jemy A. Gutierrez, John D. Knafels, Erik A. Lachapelle,
║
║
┴
┴
◊
Jayvardhan Pandit, Kevin D. Parris, Sylvie Perez, Jeffrey A. Pfefferkorn, David A. Price,
◊
┴
‡
‡
║
Brian Raymer, Trenton T. Ross, Andre Shavnya, Aaron C. Smith, Timothy A. Subashi,
┴
‡
◊
‡
◊
Gregory Tesz, Benjamin A. Thuma, Meihua Tu, John D. Weaver, Yan Weng, Jane M.
║
┴
┴
Withka, Gang Xing , Thomas V. Magee
◊
Pfizer Inc. Medicine Design, 1 Portland Street, Cambridge, Massachusetts USA 02139
┴
Pfizer Inc. Internal Medicine, 1 Portland Street, Cambridge, Massachusetts USA 02139
║
Pfizer Inc. Structural Biology and Biophysics, Eastern Point Road, Groton, Connecticut USA
06340
‡
Pfizer Inc. Medicine Design, Eastern Point Road, Groton, Connecticut USA 06340
Ketohexokinase, Fructokinase, Fragmentꢀbased drug design, Fructose metabolism, Hepatic
steatosis, Nonꢀalcoholic steatohepatitis
ACS Paragon Plus Environment

Page 3 of 52
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
ABSTRACT: Increased fructose consumption and its subsequent metabolism have been
implicated in hepatic steatosis, dyslipidemia, obesity and insulin resistance in humans. Since
ketohexokinase (KHK) is the principal enzyme responsible for fructose metabolism,
identification of a selective KHK inhibitor may help to further elucidate the effect of KHK
inhibition on these metabolic disorders. Until now, studies on KHK inhibition with small
molecules have been limited due to the lack of viable in vivo pharmacological tools. Herein we
report the discovery of 12, a selective KHK inhibitor with potency and properties suitable to
evaluate KHK inhibition in rat models. Key structural features interacting with KHK were
discovered through fragmentꢀbased screening and subsequent optimization using structureꢀbased
drug design and parallel medicinal chemistry led to the identification of pyridine 12.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
INTRODUCTION
Ketohexokinase (KHK) is the rateꢀlimiting enzyme involved in fructose metabolism and
1
catalyzes the phosphorylation of fructose to fructoseꢀ1ꢀphosphate (F1P). Human KHK (hKHK)
is expressed from two alternative mRNA splice variants, encoding the two distinct enzyme
isoforms KHKꢀA and KHKꢀC. The affinity and capacity of KHKꢀC for fructose phosphorylation
is much greater than KHKꢀA as evidenced by a much lower K (0.73 mM for KHKꢀC vs. 28.6
m
3
ꢀ1
2 ꢀ1
mM for KHKꢀA), greater k (3.14 x 10 s for KHKꢀC vs 3.04 x 10 s for KHKꢀA) and a 405ꢀ
cat
6
ꢀ1 ꢀ1
4
ꢀ1 ꢀ1
fold greater overall catalytic efficiency (4.30 x 10 M s for KHKꢀC vs. 1.06 x 10 M s for
2
KHKꢀA). While KHKꢀA is ubiquitously expressed, the expression of KHKꢀC is limited to the
3
liver, kidney, and intestine – the primary sites of fructose metabolism in the body.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 52
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
Fructose consumption, which has increased dramatically over the last century, may be a dietary
factor contributing to nonꢀalcoholic fatty liver disease (NAFLD), dyslipidemia, obesity and
4
insulin resistance in humans. In contrast to glucose, metabolism of fructose is not regulated by
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
feedback inhibition. As a consequence, fructose is preferentially metabolized relative to other
carbohydrates, and its metabolism generates various reactive and signaling metabolites that
5
contribute to metabolic disease progression. As an example, glycolytic metabolites generated
from fructose, such as glucoseꢀ6ꢀphosphate, activate the carbohydrate response element binding
protein (ChREBP), a highly lipogenic transcription factor that can promote both steatosis and
5c,
6
insulin resistance with carbohydrate overꢀfeeding.
Consequently, postꢀprandial
hypertriglyceridemia occurs in both rodents and humans following fructose feeding due to
3
, 7
decreased very low density lipoprotein (VLDL) clearance and enhanced de novo lipogenesis.
In clinical studies, high fructose consumption also increased insulin resistance and hepatic
8
steatosis. Additionally, correlation studies have revealed that consumption of fructoseꢀ
containing carbonated beverages is associated with NAFLD in patients who lack other features
of the metabolic syndrome, indicating that fructose intake independent of metabolic disease can
9
increase liver fat. Taken together, these data suggest that increased fructose consumption
promotes features of the metabolic syndrome, such as hepatic steatosis, insulin resistance and
obesity.
Studies in KHKꢀnull mice (i.e. lacking both KHKꢀA and ꢀC) have previously demonstrated that
KHK is required for the adverse metabolic effects of fructose. Mice lacking KHK were protected
from fructoseꢀinduced hypertriglyceridemia, insulin resistance, hepatic lipid accumulation,
3, 10
weight gain and adiposity compared to the wild type littermate controls.
Interestingly, KHKꢀ
ACS Paragon Plus Environment

Page 5 of 52
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
null mice are also protected from the deleterious effects of endogenously produced fructose, as
they display reduced weight gain, insulin resistance and hepatic steatosis when fed a high
glucose, fructoseꢀfree diet that promotes endogenous fructose synthesis, features of insulin
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
11
resistance and the metabolic syndrome. KHK loss of function mutations have been reported in
humans and result in a benign and asymptomatic syndrome known as essential fructosuria, in
1
2
which fructose from the diet results in fructose wasting in the urine. On the basis of the overall
data from human genetics and nonꢀclinical studies, KHK inhibition was proposed as a potentially
safe mechanism to treat metabolic diseases.
13
At the molecular level, KHKꢀC (KHK hereafter) exists as a dimer in solution and has been
1
4
observed to form a dimer in all published crystal structures. The main structural features of the
protein as well as its interactions with fructose and AMPꢀPNP, a nonꢀhydrolyzable
14a, 14d
imidophosphate analogue of ATP, have been described in detail elsewhere.
Key
interactions are illustrated in Figure 1, which shows two orthogonal views of AMPꢀPNP bound
to KHK. Three noteworthy features of this binding site, indicated in Figure 1, are exploited in
subsequent inhibitor designs: (1) hydrogen bond between the ribose 3’ꢀOH and the backbone
carbonyl oxygen of Gly229, part of a loop (A224ꢀA230) that forms the ‘roof’ of the binding
pocket in this representation; (2) water molecule that makes two hydrogen bonds with AMPꢀ
PNP, and with protein backbone atoms (backbone amide of Phe245 and backbone carbonyl of
Cys282); (3) section of the binding site made up of three consecutive proline residues (246ꢀ248)
that we termed the ‘prolineꢀwall’, forms the ‘southwest’ corner of the binding pocket. Three
crystal structures with bound fragments that separately interact with each of these elements are
shown in Figure 3.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 52
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
Relative to protein kinases, the flipped orientation of the purine ring of AMPꢀPNP in the binding
pocket is noteworthy as it offers the possibility of finding nonꢀhinge binding inhibitors, which
could avoid broad kinome activity. An unusual feature of the hKHKꢀC dimeric assembly is the
formation of a betaꢀclasp at the interface, in which residues 19ꢀ33 from one subunit (designated
A) form part of a beta sheet with βꢀstrands from the other subunit (B); hence it is a pseudoꢀ
homodimer. As a result, Asp27 from monomer B (Asp27(B)) extends into the active site of
monomer A within about 6Å of the bound AMPꢀPNP (illustrated in Figure S1), and is a feature
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
4bꢀf
that was exploited in inhibitor design by others (vide infra).
A
B
1
2
3
Figure 1. (A) hKHKꢀC complexed with AMPꢀPNP (PBD ID: 3NBV), showing the two
molecules in the crystallographic asymmetric unit. Molecule A (colored by secondary structure)
is in a closed conformation and Molecule B (yellow) is in an open conformation. Each molecule
is bound to AMPꢀPNP, and the binding site in the closed conformation is shown in a closeꢀup
view in 1B. (B) A closeꢀup view of the AMPꢀPNP binding site is shown in two approximately
ACS Paragon Plus Environment

Page 7 of 52
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
orthogonal views. Three key interactions described in the text are identified with numbers. A
semiꢀtransparent surface is drawn around the atoms constituting the binding site, colored
according to the hydrophobicity of the underlying atoms, with brown being more hydrophobic
than green. Hydrogen bonds are indicated with dotted lines.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Two chemical series of hKHK inhibitors have been reported in the literature (Figure 2). A
pyrimidinopyrimidine series, represented by 1, was identified from a high throughput screening
(HTS) campaign. This chemical series was optimized to provide the most potent inhibitors of
hKHK reported to date. However, only low exposure levels were achieved in rat
14b, 14c
pharmacokinetic (PK) studies due to high metabolic clearance.
A second series, based on
an indazole core was initially identified through Xꢀray crystallographic fragment screening and
further developed into analogs such as indazole 2 with moderate inhibitory activity against
14dꢀf
hKHK and a good rat PK profile.
The basic amine is a key functional group in both series
that specifically interact with Asp27 of monomer B as a way to improve potency in the hKHK in
1
4bꢀf
vitro assay.
To our knowledge, in vitro inhibition of rodent KHK was not reported for either
the pyrimidinopyrimidine or indazole series and, until now, there were no reports of in vivo
KHK inhibition. In our in vitro enzymatic assays, significantly decreased potency against the rat
KHK (rKHK) vs hKHK enzyme was observed for basic analogues interacting directly with
Asp27(B). Comparison of the human and rat KHK amino acid sequences revealed a difference at
residue 107: asparagine in human and serine in rat (Figure S1). Based on the Xꢀray structure,
such a difference is hypothesized to create a different environment around Asp27 and cause
potency shifts between species for basic compounds interacting directly with Asp27(B).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 52
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 2. Representative structures of KHK inhibitors from the pyrimidinopyrimidine and
indazole series reported in literature.
Given the limitation of the currently available KHK inhibitor tools for biological studies, we
sought to identify a novel series of inhibitors that are equipotent against the human and rat KHK
enzyme, with good PK properties. Fragmentꢀbased drug discovery (FBDD) is now a wellꢀ
1
5
established technology for generating new chemical leads and drugs. Screening of a Pfizer
1
6
proprietary fragment library using NMR ligandꢀbased methods has been useful in identification
of unique fragment hits for many internal targets. KHK was considered an excellent candidate
for FBDD due to relatively easy generation of high quality and large quantity of a biotin acceptor
peptide (BAP)ꢀtagged stable protein for development of NMR and SPR binding assays, the
precedent for robust protein crystallography, and suitable DMSO tolerance for high
1
7
concentration screening. Due to prior success with our fragment library for kinases, and the
enabling biophysical characteristics of KHK, we initiated a FBDD approach to identify novel
chemical matter. Herein we report the identification of new fragment hits from a primary ligandꢀ
based NMR screen, secondary confirmation by SPR and biochemical assays, followed by
determination of key interactions observed by Xꢀray protein crystallography, which were then
ACS Paragon Plus Environment

Page 9 of 52
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
used to rapidly provide a nonꢀbasic pharmacological tool compound for further study in rodent in
vivo models.
RESULTS AND DISCUSSION
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
During our search for KHK inhibitors, hKHK was screened against our previously described
fragment library containing 2592 fragments using a NMR saturation transfer difference (NMRꢀ
1
8
STD) ligandꢀbased method. The total hit rate was 17%, which is consistent with other inꢀhouse
NMR kinase fragment screens. Of these 451 hits, 286 (11%) were considered of high confidence
based on the intensity of the NMRꢀSTD signal. Although novel hits are generally observed for
each target screened, hKHK had a significant number of unique fragment hits (11 hits, 2.4%)
which were not hits in any of our 31 prior NMR fragment screens. Prior experience has taught us
that application of orthogonal biophysical and biochemical assays is critical for validating
fragment hits, prioritizing for Xꢀray crystallography, and building confidence for medicinal
chemistry efforts. To this end, 448 available primary fragment hits were evaluated in an SPR
binding assay using an 11 point dose response curve with a top concentration of 300 µM and
twoꢀfold dilution. Of these primary fragments hits, 114 had binding affinities (K ) <100 µM
D
while 16 fragments had K values <25 µM.
D
The fragment hits were evaluated for their physicochemical properties, inhibitory activity in the
KHK enzymatic assay and chemical tractability. Of the 179 SPRꢀvalidated hits, 23 were found to
be functionally active in the in vitro biochemical assay (IC values ranging from 46 to 439 µM)
5
0
and these 23 fragments were prioritized for xꢀray crystallography. Figure S2 shows that the
distribution of LogD and fraction Sp3 for the selected fragments is representative of the
distribution for all SPRꢀconfirmed hits. Additionally, Figure S2 highlights the fact that the
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 52
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
selected fragments were neutral, which is aligned with our goal of avoiding basic amines.
Finally, Figure S2 also illustrates the range of Kd covered by the selected 23 fragments, showing
that not only tight binders were progressed to crystallization. All 23 fragments were found to
bind in the ATP binding site with similar key interactions observed across multiple fragments.
Consistent with the flipped orientation of the purine ring of ATP in this pocket relative to protein
kinases, none of the hits contained a hingeꢀbinding motif. Table 1 and Figure 3 summarize the
profile and binding modes of representative hits. As mentioned previously, the hits from NMR
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
screening were confirmed by SPR, with K values ranging from 20 µM to 103 µM for the
D
selected examples. Compounds 3, 4 and 5 also showed measurable inhibitory activity in the
hKHK enzymatic assay (IC = 118 µM, 319 µM and 100 µM respectively).
50
Table 1. Representative hits from the NMR fragment screening
Compound
3
4
5
6
a
K (µM)
20
118
62
319
44
100
103
D
b
hKHK IC (µM)
>500
5
0
(pIC ± SD)
(3.92 ± 0.05)
246
(3.50 ± 0.07)
217
(4.0 ± 0.3)
214
5
0
c
MW
170
d
FQ
0.53
0.55
0.50
0.44
a
b
K values were determined by SPR (n=1). IC values for hKHK are reported as the mean of
D
50
c
d
3
replicates with pIC ± SD in parentheses. MW: molecular weight. FQ: fit quality is a scaled
50
1
9
measure of ligand efficiency that centers optimal binders near 1.0. The functional groups
ACS Paragon Plus Environment

Page 11 of 52
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
identified in blue, purple and pink respectively make key interactions with the ribose pocket,
conserved water molecule and proline wall as described in Figure 3.
Most of the confirmed hits contained a heteroaromatic core (observed in the narrow ATP binding
pocket) with multiple aliphatic substituents. As exemplified by compounds 3 and 4, saturated
heterocyclic rings with polar functionalities such as an alcohol or an ether were deeply embedded
in the ribose pocket (Figure 3A and B). Interestingly, a high fraction of the fragment hits
contained a nitrile group. In the examples selected for crystallization, the nitrile group was
found to accept a hydrogen bond with a conserved water molecule deep in the binding site
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 3B). Additionally, various lipophilic groups such as methyl or trifluoromethyl (in
compounds 4 and 5 respectively) made hydrophobic interactions with the proline wall (Figure
3B and C).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 52
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
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
B
C
Figure 3. Binding modes of fragment hits 3ꢀ5, illustrating the three key interactions in the
binding site highlighted in Figure 1. (A) Compound 3 contains a saturated ring with a polar
substitution in the ribose pocket, making a hydrogen bond with the backbone carbonyl oxygen of
Gly229 (PBD ID: 5WBM). (B) The nitrile group of 4 interacts with a conserved molecule of
water deep in the binding pocket (PBD ID: 5WBO). (C) The trifluoromethyl group of 5 makes
favorable hydrophobic interactions with the proline wall (PBD ID: 5WBP). A sulfate ion (drawn
as a ballꢀandꢀstick model) from the crystallization buffer is seen in all three structures to be
bound at the site previously occupied by the γ phosphate group of AMPꢀPNP.
ACS Paragon Plus Environment

Page 13 of 52
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Due to the identification of multiple fragment hits with similar binding modes, the Pfizer
screening library was searched for compounds possessing an aryl core densely functionalized
with substituents that could satisfy all three key interactions described above and illustrated in
Figure 3. This approach led to the identification of compounds 7 and 8 where an alcoholꢀ
substituted pyrrolidine or piperidine was positioned to fill the ribose pocket, a nitrile group forms
a hydrogen bound with the conserved water molecule and a trifluoromethyl group interacts with
the proline wall (Figure 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
HO
HO
N
N
N
N
N
H
N
N
F
F
N
F
F
F
N
OH
F
Pyrrolopyridine 7
hKHK IC50 = 1.48 ꢀM
Pyridine 8
hKHK IC50 = 16.79 ꢀM
Figure 4. Compounds 7 (blue in the crystal structure, PBD ID: 5WBQ) and 8 (pink in the crystal
structure, PBD ID: 5WBR) with improved potency against hKHK. The functional groups
identified in blue, purple and pink reproduce the key interactions described for the fragments hits
in Table 1 and Figure 3. In this figure, which shows the binding site in exactly the same view as
the right hand panel of Figure 3, all protein atoms have been hidden for clarity. The nitrile group
in both compounds interacts with a conserved water molecule, shown as a gold sphere. A second
water molecule, shown as a red sphere, is seen in the crystal structure of 7, which occupies the
same space as that filled by the hydroxyl group in 8.
As described in Table 2, pyrrolopyridine 7 is a hKHK inhibitor with moderate potency and low
2
0
metabolic stability in human liver microsome (HLM) assays. As observed in the crystal
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 52
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
structure (Figure 4), the ethyl substitution at the C2 position does not make any important
interactions and was therefore removed without loss of potency. This modification significantly
decreased LogD and resulted in much improved metabolic stability in the HLM assay (Table 2).
The methylpyrrolidinol group was found to complement the ribose pocket with the S enantiomer
9 found to be more potent than the R enantiomer 10.
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 2. Structureꢀactivity relationship (SAR) with the pyrrolopyridine core
a
c
1
2
hKHK IC50 (µM)
pIC ±SD)
b
HLM Clapp
Compound
R
R
SFLogD
2.8
(
(µL/min/mg)
5
0
1.48
7
Et
H
H
103
(5.8 ± 0.1)
0
.67
9
2.2
2.1
<8
<8
(
(
6.17 ± 0.08)
24.21
10
4.62 ± 0.01)
a
IC values for hKHK are reported as the mean of 3 replicates with pIC ± SD in parentheses.
5
0
50
b
LogD values were determined using the shakeꢀflask method with an octanol:water partition at
2
1
c
20
pH 7.4. Apparent clearance from human liver microsomes.
As depicted in Figure 5, pyrrolopyridine 9 was identified as a relatively potent and efficient
22
human and rat KHK inhibitor with adequate microsomal stability, permeability (P ) and
app
thermodynamic solubility. When evaluated in rat PK studies, a mean free Cmax of 1.3 µM with
ACS Paragon Plus Environment

Page 15 of 52
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
high interꢀanimal variability was observed following oral administration at a high dose of 300
mg/kg. With an IC of 0.64 µM against the rKHK, this exposure profile did not allow complete
50
inhibition of KHK activity over a timeframe suitable for studying metabolic endpoints in vivo,
thus further improvement of the series was pursued.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A
B
a
hKHK IC = 0.67 µM
5
0
a
rKHK IC = 0.64 µM
5
0
HLM Clapp <8 µL/min/mg
ꢀ6
b
Papp = 14 x10 cm/s
c
Therm. Sol. = 147 µM
d
9
Fu = 0.032
a
Figure 5. (A) In vitro profile of pyrrolopyridine 9. IC50 values for hKHK and rKHK are
reported as the mean of 3 replicates (pIC50 = 6.17 ± 0.08 and 6.19 ± 0.01, respectively).
b
22 c
Permeability (P ) was assessed in Ralph Russ canine kidney cells.
Thermodynamic
app
d
solubility was measured at pH 7.4. Fu = faction unbound in plasma as assessed by equilibrium
2
3
dialysis. (B) Free plasma level of 9 in rat PK studies. 300 mg/kg dose was administered orally
using 0.25% HPCꢀSL/0.05% DOSS as vehicle. Data presented as mean values (± standard
deviation) with n = 3 per data point. Solid line = exposure. Dashed line = in vitro rKHK IC .
5
0
Since the substituents of pyrrolopyridine 9 recapitulated all three important interactions observed
with the fragment hits (Figures 3 and 4), efforts to further improve the series were focused on the
heteroaromatic core. When comparing the crystal structures of pyrrolopyridine 7 and pyridine 8,
the two scaffolds overlapped well and both displayed all three critical interactions with the ribose
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 52
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
pocket, conserved water molecule and hydrophobic pocket of KHK (Figure 4). Although
pyridine 8 is a weaker KHK inhibitor, presumably due to suboptimal interaction of the
functionalized piperidine ring in the ribose pocket, it offered an additional and synthetically
enabled vector that was targeted for further optimization. As illustrated in Figure 6, merging
functionalities from compounds 9 and 8 provided compound 11, which displayed moderate KHK
inhibitory potency (hKHK IC = 1.03 µM) and microsomal stability (HLM Cl = 14
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
0
app
µL/min/mg) in a reasonable LogD space (SFLogD = 2.2).
Figure 6. Merging pharmacophores led to the identification of pyridine 11.
Using parallel chemistry, the piperazine vector of 11 was rapidly explored. In order to avoid
discrepancies between potency against human and rat KHK (vide supra), neutral analogs were
the key focus of optimization efforts. The optimal balance of potency, physical properties and
clearance was achieved with transꢀdihydroxy pyrrolidine analog 12 having a hKHK IC value of
5
0
0.45 µM and no observable turnover in the HLM assay (Figure 7). The two hydroxyl groups
form hydrogen bonds with solventꢀexposed polar side chains Glu227 and Asn107, as well as
with a bound water molecule. A complete map of interactions between compound 12 and hKHK
is reported in Figure S3.
ACS Paragon Plus Environment

Page 17 of 52
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
A
B
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
hKHKꢀC IC = 0.45 µM
C
5
0
b
hKHKꢀA IC = 0.39 µM
5
0
a
rKHK IC = 0.34 µM
5
0
SFLogD = 2.2
HLM Clapp <8 µL/min/mg
RLM Clapp <8 µL/min/mg
ꢀ6
Papp = 3.5 x10 cm/s
Therm. Sol. = 503 µM
7
.4
Fu = 0.13
a
Figure 7. (A) Structure and profile of pyridine 12 in in vitro assays. IC values for rat and
50
human KHKꢀC inhibition are reported as the mean of 19 replicates (pIC = 6.3 ± 0.2 and 6.5 ±
5
0
b
0.2 respectively). IC value for hKHKꢀA inhibition is reported as the mean of four replicates
5
0
(
pIC = 6.41 ± 0.09). RLM = rat liver microsomes. (B) The binding mode of 12 in the crystal
50
structure (PBD ID: 5WBZ) is reported with a view of the binding site that is identical to the right
panel of Figures 3 and 4. All protein atoms have been hidden for clarity. (C) Free plasma level of
1
2 in rat PK studies. A 150 mg/kg dose was administered orally using 1% polyvinylpyrrolidone /
0
.025% sodium lauryl sulfate / 1% pluronic (nanocrystal suspension) as vehicle. Data presented
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 52
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
as mean values (± standard deviation) with n = 4 per data point. Solid line = exposure. Dashed
line = in vitro rKHK IC50.
Pyridine 12 does not interact directly with Asp27(B) and inhibited rKHK and both human
isoforms with comparable potency in enzymatic assays. On the basis of its high metabolic
stability in rat liver microsomes (RLM), improved solubility and acceptable permeability (Figure
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
7A), compound 12 progressed to PK studies in rats. As determined with a 1 mg/kg intravenous
(i.v.) dose, 12 demonstrated moderate in vivo systemic clearance (25.9 mL/min/kg), moderate
volume of distribution (1.38 L/kg) and short effective halfꢀlife (1.2 h) in rats. When orally dosed
at 5 mg/kg using 0.5% methylcellulose as the vehicle, 12 exhibited an oral bioavailability of 45%
(Cmax = 2.94 ꢀg/mL, AUClast = 14.5 ꢀg·h/mL) in rats. Free plasma concentration equivalent to
multiples over the in vitro IC were achieved for an extended period of time (~14 hours) at an
5
0
oral dose of 150 mg/kg (Figure 7C).
Compound 12 showed high selectivity for KHK with no significant activity detected against a
broad panel of transporters, receptors, ion channels and enzymes, including the hERG channel
and major human CYP450s (Table S1). Overall, 12 had a clean in vitro safety profile with no
genetic toxicology risks identified in Ames mutagenicity and in vitro micronucleus assays, and
no toxicity was observed in HepG2 cells at concentrations up to 300 ꢀM with 72 hour
incubation. Supported by inhibitory potency against rKHK (Figure 7A), suitable exposure in rat
PK studies (Figure 7B) and clean in vitro safety profile, compound 12 was resynthesized on
larger scale to enable the evaluation of KHK inhibition in vivo.
Sprague Dawley rats were utilized to demonstrate KHK inhibition in vivo, focusing on the liver
and kidney which are primary sites of KHKꢀC expression and fructose metabolism. All
ACS Paragon Plus Environment

Page 19 of 52
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
procedures performed on these animals were in accordance with regulations and established
guidelines and were reviewed and approved by the Pfizer Institutional Animal Care and Use
Committee. Sprague Dawley rats were administered increasing oral doses of compound 12
followed by an i.v. bolus of fructose. Two minutes after fructose administration, animals were
sacrificed and the liver and kidney tissues were removed and snapꢀfrozen in liquid nitrogen. F1P
was directly quantified in liver and kidney tissues by liquid chromatography/mass spectrometry
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
(
LC/MS). Direct F1P quantitation is challenging due to highly abundant sugar monophosphate
isomers, such as glucoseꢀ6ꢀphosphate (G6P), fructoseꢀ6ꢀphosphate (F6P), and glucoseꢀ1ꢀ
phosphate (G1P). Because the isomers have one common m/z, selectivity may not be achieved in
the mass spectrometer; rather, they must be separated chromatographically. Widely adopted MSꢀ
compatible LC methods result in poor separation or coꢀelution of hexose monophosphate
isomers. Therefore, an inꢀhouse ion exchange liquid chromatography method, which maintains
compatibility with MS, was developed to address this problem. Using this method, doseꢀ
dependent decreases in both liver and kidney F1P concentrations were measured with ED50
values of 30.0 mg/kg and 15.7 mg/kg, respectively (Figure 8). The decrease in both liver and
kidney F1P levels indicates potent KHK inhibition in these tissues.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 52
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
Kidney
Liver
30.0
ED50 (mg/kg)
pED50 ± SE
15.7
1.20 ± 0.08
1.48 ± 0.09
Figure 8. Concentrations of hepatic and renal fructoseꢀ1ꢀphosphate (F1P) following an i.v. bolus
of fructose and increasing doses of compound 12 in Sprague Dawley rats. A nonꢀlinear dose
response curve was fitted to a variable slope (four parameter) using GraphPad Prism and the
calculated ED for kidney and liver F1P reduction are reported. Data presented as mean values
5
0
(
± standard error) with n = 7ꢀ8 per data point.
To determine the effect of compound 12 on systemic fructose metabolism, plasma fructose levels
were measured following an oral bolus of fructose. Sprague Dawley rats were treated with
increasing doses of compound 12 and then administered a 2 g/kg oral bolus of fructose. Because
KHK is the primary enzyme responsible for metabolic clearance of fructose, treatment with
compound 12 was anticipated to increase plasma fructose levels. Indeed, a doseꢀdependent
increase in plasma fructose levels was observed (Figure 9), suggesting that compound 12
inhibited systemic fructose metabolism. Taken together, these data support that oral
ACS Paragon Plus Environment

Page 21 of 52
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
administration of compound 12 inhibits hepatic and renal fructose metabolism in rats, resulting
in decreased formation of F1P in these organs and increased fructose concentration in plasma.
Future studies will be informative as to the degree of KHK inhibition required for metabolic
benefit since no clinical and nonꢀclinical data has been reported with a KHK inhibitor.
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 9. Plasma fructose levels in Sprague Dawley rats following an oral fructose bolus and
treatment with the indicated dose of compound 12. Animals were administered compound 12
orally one hour prior to an oral bolus of fructose (2 g/kg). Data presented as mean values (±
standard error) with n = 6ꢀ8 per data point.
SYNTHESIS
The synthesis of the pyrrolopyridine series began with construction of pyrrole 15 (Scheme 1).
Cyanogen bromide was first reacted with phenol and subsequent S 2 displacement with 1ꢀbutyne
N
led to nitrile 14. Condensation with diethyl aminomalonate led to formation of pyrrole 15.
Pyrrole 16 was commercially available. Reaction of these aminopyrroles with tosylate 13 and
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 52
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
subsequent DBUꢀmediated ring closure led to pyrrolopyridine cores 17 and 18. Chlorination and
subsequent S Ar reactions led to the formation of 7, 9, and 10.
N
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
Scheme 1. Synthesis of pyrrolopyridines 7, 9, and 10
Pyridine 11 was constructed in a twoꢀstep S Ar sequence (Scheme 2). Starting from
N
commercially available dichloropyridine 19, amine addition at the 6ꢀposition occurred with good
selectivity at ambient temperature to form pyridine 20. Subsequent amine addition at the 2ꢀ
position occurred at elevated temperatures to form pyridine 11. Pyridine 8 was synthesized in an
analogous manner.
ACS Paragon Plus Environment

Page 23 of 52
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Scheme 2. Synthesis of pyridine 11
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
To provide sufficient quantities for the planned in vivo experiments, access to enantiopure 12
was required (Scheme 3). Addition of methyl magnesium bromide to Nꢀbenzylꢀpyrrolidinꢀ3ꢀone
(21) proceeded in good yield to provide the corresponding tertiary alcohol. Benzyl deprotection
by hydrogenolysis led to alcohol 22. Chiral resolution by preferential crystallization with Lꢀ
dihydroorotic acid from ethanol followed by recrystallization from methanol produced the
desired pure (S)ꢀenantiomer. S Ar reaction of the core pyridine 19 with transꢀ
N
dihydroxypyrrolidine 23 proceeded smoothly and isolation of the desired major regioisomer was
achieved in good yield after crystallization. The final S Ar proceeded smoothly and 132 g of
N
pyridine 12 was isolated by crystallization from MTBE in high yield.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 52
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 3. Synthesis of pyridine 12
O
HN
NH
O
OH
O
1. MeMgBr
THF, 78%
HO
O
HO
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
O
HN
NH
OH
N
1. EtOH
2. MeOH, 40%
2. Pd/C, H2
MeOH, 81%
N
N
O
H
H
Bn
O
22
21
1
:
2
HN
OH
OH
HO
2
3
Cl
1. NaHCO3
N
N
N
EtOH, 65%
N
N
2
. 22, NaHCO3
EtOH, 98%
F3C
Cl
F3C
N
OH
3. MTBE, 97%
19
1
2
OH
CONCLUSION
In this report we described the identification of fragments interacting with KHK which were
subsequently optimized into moderately potent KHK inhibitors that could be used to modulate
KHK activity in vivo. The fragment hits were identified via an NMR screen and confirmed by
SPR. Selected fragments were coꢀcrystallized with hKHK and a number of key interactions were
observed with multiple fragments. In order to avoid discrepancy between the potencies against
rat and human KHK, basic amines were avoided during optimization efforts. This led to the
identification of pyridine 12, a nonꢀbasic moderately potent and highly selective KHK inhibitor
with favorable ADME and safety profiles and good exposure in rats. Compound 12 was
synthesized on a scale needed to support in vivo experiments and inhibition of KHK and
metabolism of fructose was demonstrated in rats. Compound 12 can facilitate the study of KHK
as a potential therapeutic target.
ACS Paragon Plus Environment

Page 25 of 52
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
EXPERIMENTAL SECTION
General. Unless otherwise noted, the reported experiments followed these general guidelines:
All chemicals, reagents, and solvents were purchased from commercial sources when available
and used without further purification. Compound 4 is commercially available from Enamine or
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
2
4
previously reported in literature. Compound 5 is commercially available from Enamine or
2
5
26
previously reported in literature. Compound 6 is previously reported in literature. Airꢀ and
moistureꢀsensitive reactions were carried out under an inert atmosphere of nitrogen, magnetically
stirred, and monitored by thin layer chromatography (TLC), high performance liquid
chromatography (HPLC), or UPLC (ultraꢀperformance liquid chromatography). Reactions were
carried out at room temperature (~23 °C). Silica gel chromatography was performed using an
automated system with preꢀpackaged columns. Concentration under reduced pressure (in vacuo)
was performed by rotary evaporation at 25–45 °C. Purified compounds were further dried under
high vacuum and/or heat to remove residual solvent. Yields refer to dried purified compounds.
NMR spectra were recorded with either a spectrometer at 600 MHz or with a spectrometer at 400
1
1
MHz for H acquisitions. Chemical shifts were referenced to the residual H solvent signals
(CDCl , δ 7.27; DMSO–d , δ 2.50; CD OD, δ 3.31). Signals are listed as follows: chemical shift
3
6
3
in ppm; multiplicity identified as s = singlet, br = broad, d = doublet, t = triplet, m = multiplet;
coupling constants in Hz. Liquid chromatographyꢀmass spectrometry (LCMS) was performed
via atmospheric pressure chemical ionization (AP) or electrospray ionization (ES) sources. Purity
is reported by relative UV area and is >95% unless noted. Based on their structures, the
compounds reported in this work are not predicted to be PAINS. Specific interaction with hKHK
is supported by data from orthogonal assays (SPR, NMR, functional) and coꢀcrystal structures
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 52
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
with hKHK. Additionally, compound 12 does not show activity in a number of assay formats as
reported in the selectivity Table S1.
Synthesis of analogues and intermediates. (R)ꢀ(1ꢀ(5ꢀMethylꢀ7Hꢀpyrrolo[2,3ꢀd]pyrimidinꢀ4ꢀ
yl)piperidinꢀ3ꢀyl)methanol (3). A vial was charged with 4ꢀchloroꢀ5ꢀmethylꢀ7Hꢀpyrrolo[2,3ꢀ
d]pyrimidine (100 mg, 0.60 mmol), (R)ꢀpiperidinꢀ3ꢀylmethanol (82.5 mg, 0.72 mmol), Hünig’s
base (0.21 mL, 1.2 mmol), and acetonitrile (3 mL). The vial was sealed and heated to 60 °C for
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
15 h. The reaction was then cooled and poured onto halfꢀsaturated ammonium chloride and ethyl
acetate. The layers were separated and organic layer was washed with saturated sodium
bicarbonate solution, dried over sodium sulfate, filtered, and concentrated in vacuo. Flash
column chromatography (20% ethyl acetate/heptane to 100% ethyl acetate to 10%
methanol/ethyl acetate), taking the pure fraction to provide 3 (20 mg, 14%). LCMS (AP+) calc.
1
for C H N O: 247.1; found: 247.3. H NMR (400 MHz, MeOHꢀd ) δ 8.18 (s, 1H), 6.99 (d,
1
3
18
4
4
J=1.2 Hz, 1H), 4.19–4.11 (m, 1H), 4.02–3.92 (m, 1H), 3.55 (dd, J=5.5, 10.9 Hz, 1H), 3.46 (dd,
J=7.8, 11.3 Hz, 1H), 3.01 (ddd, J=2.7, 11.3, 12.8 Hz, 1H), 2.78 (dd, J=10.1, 12.9 Hz, 1H), 2.44
(d, J=1.2 Hz, 3H), 2.01–1.82 (m, 3H), 1.82–1.66 (m, 1H), 1.33–1.17 (m, 1H).
2ꢀEthylꢀ7ꢀ(3ꢀhydroxyꢀ3ꢀmethylpyrrolidinꢀ1ꢀyl)ꢀ5ꢀ(trifluoromethyl)ꢀ1Hꢀpyrrolo[3,2ꢀb]pyridineꢀ6ꢀ
carbonitrile (7). Step 1: A round bottomed flask was charged with 17 (25.0 g, 98.0 mol) and
POCl (40 mL) and the reaction was heated to reflux temperature for 2 h. The reaction was
3
cooled to rt and the mixture was added dropwise to ice water. The mixture was then extracted
with EtOAc (200 mL x 2) and the layers separated. The combined organic phases were washed
with brine (150 mL), dried over Na SO and concentrated in vacuo to give 7ꢀchloroꢀ2ꢀethylꢀ5ꢀ
2
4
(trifluoromethyl)ꢀ1Hꢀpyrrolo[3,2ꢀb]pyridineꢀ6ꢀcarbonitrile (19 g, 71%) as a grey solid. LCMS
ACS Paragon Plus Environment

Page 27 of 52
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
(
ES−) calc. for C H ClF N (M−H): 272.0; found: 272.0. H NMR (400 MHz, CDCl ) δ 9.10
1
1
7
3
3
3
(
br. s, 1H), 6.73 (s, 1H), 2.97 (q, J=7.5 Hz, 2H), 1.45 (t, J=7.5 Hz, 3H).
Step 2: To a solution of 7ꢀchloroꢀ2ꢀethylꢀ5ꢀ(trifluoromethyl)ꢀ1Hꢀpyrrolo[3,2ꢀb]pyridineꢀ6ꢀ
carbonitrile (100 mg, 0.37 mol) and 3ꢀmethylpyrrolidinꢀ3ꢀol (75 mg, 0.55 mmol) in DMF (2 mL)
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
was added Et N (110 mg, 1.09 mmol). The mixture was stirred for 16 hours at 80 °C. The
3
mixture was purified by preparative TLC to give the 7 (72 mg, 59%) as a yellow solid. LCMS
1
(
ES+) calc. for C H F N O (M+H): 339.1; found: 339.1. H NMR (400 MHz, CD OD) δ 6.43
16 17
3
4
3
(
s, 1H), 4.52–4.42 (m, 1H), 4.16 (d, J=10.5 Hz, 1H), 4.13–4.05 (m, 1H), 3.86 (d, J=10.5 Hz, 1H),
.87 (q, J=7.5 Hz, 2H), 2.19–1.96 (m, 2H), 1.53 (s, 3H), 1.35 (t, J=7.5 Hz, 3H).
2
6ꢀ(4ꢀ(2ꢀHydroxyethyl)piperazinꢀ1ꢀyl)ꢀ2ꢀ(3ꢀ(hydroxymethyl)piperidinꢀ1ꢀyl)ꢀ4ꢀ
trifluoromethyl)nicotinonitrile (8). A vial was charged with 20 (73 mg, 0.22 mmol), DMF (0.25
mL), piperidinꢀ3ꢀylmethanol (25 mg, 0.22 mmol), and K CO (35 mg, 0.25 mmol), and stirred at
(
2
3
100 °C for 2 h. The reaction was cooled to rt and diluted with ethyl acetate (3 mL) and halfꢀ
saturated NaCl solution (3 mL). The layers were separated and the organic was concentrated in
vacuo and purified via preparatory reverseꢀphase HPLC to provide 8 (48 mg, 53%). LCMS
1
(
ES+) calc. for C H F N O (M+H): 413.2; found: 414.2. H NMR (400 MHz, MeODꢀd ) δ
1
9
26
3
5
2
4
6
.54 (s, 1H), 4.34–4.27 (m, 1H), 4.24–4.13 (m, 1H), 3.79–3.74 (m, 4H), 3.72 (t, J=5.9 Hz, 2H),
3.55–3.39 (m, 2H), 3.06 (ddd, J=3.1, 11.7, 13.0 Hz, 1H), 2.82 (dd, J=9.4, 12.9 Hz, 1H), 2.64–
.60 (m, 4H), 2.59 (t, J=5.9 Hz, 2H), 1.93–1.74 (m, 3H), 1.74–1.57 (m, 1H), 1.36–1.22 (m, 1H).
2
(
S)ꢀ7ꢀ(3ꢀHydroxyꢀ3ꢀmethylpyrrolidinꢀ1ꢀyl)ꢀ5ꢀ(trifluoromethyl)ꢀ1Hꢀpyrrolo[3,2ꢀb]pyridineꢀ6ꢀ
carbonitrile (9) and (R)ꢀ7ꢀ(3ꢀhydroxyꢀ3ꢀmethylpyrrolidinꢀ1ꢀyl)ꢀ5ꢀ(trifluoromethyl)ꢀ1Hꢀ
pyrrolo[3,2ꢀb]pyridineꢀ6ꢀcarbonitrile (10). Step 1: Compound 18 (200 g, 0.88 mol) was
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 52
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
dissolved in CH CN (1.5 L). Then POCl (673 g, 4.4 mol) was added carefully. After the
3
3
addition, the resulting mixture was heated to 60 °C overnight. The reaction mixture was then
cooled to rt and concentrated in vacuo. The residue was dissolved in ethyl acetate and slowly
poured onto water (1.0 L). The layers were separated and the organic layer was washed with
10% K CO , dried over Na SO , filtered, and concentrated in vacuo. Flash column
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
3
2
4
chromatography was then used to provide 7ꢀchloroꢀ5ꢀ(trifluoromethyl)ꢀ1Hꢀpyrrolo[3,2ꢀ
b]pyridineꢀ6ꢀcarbonitrile (105 g, 49%). LCMS calc. for C H ClF N (M+H): 246.0; found: 245.7
9
3
3
3
1
(M+H). H NMR (400 MHz, DMSOꢀd ) δ 13.11 (br. s, 1H), 8.29 (t, J=2.8 Hz, 1H), 7.04 (dd,
6
J=3.0, 2.0 Hz, 1H).
Step 2 (milligram synthesis): To a solution of 7ꢀchloroꢀ5ꢀ(trifluoromethyl)ꢀ1Hꢀpyrrolo[3,2ꢀ
b]pyridineꢀ6ꢀcarbonitrile (400 mg, 1.64 mmol) and 3ꢀmethylpyrrolidinꢀ3ꢀol (320 mg, 2.34 mmol)
in DMF (5 mL) was added Et N (500 mg, 4.95 mmol). The mixture was then stirred at 80 °C for
3
2
h. The mixture was then poured onto water (40 mL) and extracted with EtOAc (40 mL x 3).
The organic layer was washed with H O (30 mL x 3), brine (30 mL x 3), dried over Na SO ,
2
2
4
filtered, and concentrated in vacuo. The material was then purified by flash chromatography (0%
to 70% EtOAc/petroleum ether to provide clean racemic material (360 mg, 71%) as yellow solid.
The enantiomers were then separated by chiral SFC (column: Chiralpak ICꢀ3 150×4.6 mm I.D., 3
um; mobile phase: ethanol (0.05% DEA) in CO from 5% to 40%; flow rate: 2.35 mL/min) to
2
provide the two separated enantiomers:
9
: 150 mg, retention time: 4.123; LCMS (ES+) calc. for C H F N O (M+H): 311.1; found:
14 13 3 4
1
3
10.9. H NMR (400 MHz, MeOD) δ 7.54 (d, J=3.0 Hz, 1H), 6.62 (d, J=3.0 Hz, 1H), 4.48–4.38
2
0
(
m, 1H), 4.16–4.03 (m, 2H), 3.90–3.81 (m, 1H), 2.12–1.97 (m, 2H), 1.49 (s, 3H). [α] +19.2 (c
D
0
.64, CH OH).10: 160 mg; er: 98.9:1.1; retention time: 4.414; LCMS (ES+) calc. for
3
ACS Paragon Plus Environment

Page 29 of 52
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
C H F N O (M+H): 311.1; found: 310.9. H NMR (400 MHz, MeOD) δ 7.54 (d, J=3.0 Hz,
14
13
3
4
1
H), 6.62 (d, J=3.0 Hz, 1H), 4.48–4.38 (m, 1H), 4.16–4.03 (m, 2H), 3.90–3.81 (m, 1H), 2.12–
20
1
.97 (m, 2H), 1.49 (s, 3H). [α] –22.2 (c 1.00, CH OH).
D
3
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(
S)ꢀ7ꢀ(3ꢀHydroxyꢀ3ꢀmethylpyrrolidinꢀ1ꢀyl)ꢀ5ꢀ(trifluoromethyl)ꢀ1Hꢀpyrrolo[3,2ꢀb]pyridineꢀ6ꢀ
carbonitrile (9, multiꢀgram synthesis)
Step 1: A threeꢀneck flask equipped with an addition funnel and thermometer was charged with
methyl magnesium bromide (3.0 M in ether, 30 mL, 90 mmol) and tetrahydrofuran (100 mL) and
cooled to ꢀ78 °C. A solution of benzyl 3ꢀoxopyrrolidineꢀ1ꢀcarboxylate (9.80 g, 44.7 mmol) in
tetrahydrofuran (100 mL) was then added dropwise, keeping the internal temperature < ꢀ70 °C.
The reaction was then warmed to rt for 1 h then cooled to 0 °C and quenched with saturated
ammonium chloride (150 mL) with a slight exotherm. The mixture was then extracted with
MTBE (3 x 100 mL). The combined organic layers were dried over magnesium sulfate, filtered,
and concentrated in vacuo. The crude product was purified by flash column chromatography (0–
1
00% ethyl acetate/heptane) to provide crude racemic product. Chiral SFC was then used to
isolate benzyl (S)ꢀ3ꢀhydroxyꢀ3ꢀmethylpyrrolidineꢀ1ꢀcarboxylate (3.0 g, 29%) as peak 1 (column:
Chiral Tech ID 250 mm x 21.2 mm 5 um; isocratic, 90% CO , 9.8% MeOH, 0.2% NH OH,
2
4
detection: 210 nm; flow rate: 80.0 mL/min; backpressure: 120 Bar). LCMS (AP+) calc. for
1
C H NO (M+H): 236.1; found: 236.0. H NMR (400 MHz, CDCl ) δ 7.45–7.25 (m, 5H),
13
17
3
3
5.25–5.03 (m, 2H), 3.67–3.43 (m, 3H), 3.38–3.22 (m, 1H), 2.00–1.73 (m, 3H), 1.42 (s, 3H).
Step 2: A Parr reactor was charged with benzyl (S)ꢀ3ꢀhydroxyꢀ3ꢀmethylpyrrolidineꢀ1ꢀ
carboxylate (3.50 g mg, 14.9 mmol), Pd/C (wet, 10 wt%, 792 mg, 0.025 mmol) and ethanol (50
mL) then charged with 30 psi hydrogen and stirred for 5 h (recharging H pressure after 2 h). The
2
ACS Paragon Plus Environment