Discovery and Development of Small-Molecule Inhibitors of Glycogen Synthase

Article abstract of DOI:10.1021/acs.jmedchem.9b01851

The overaccumulation of glycogen appears as a hallmark in various glycogen storage diseases (GSDs), including Pompe, Cori, Andersen, and Lafora disease. Accumulating evidence suggests that suppression of glycogen accumulation represents a potential therapeutic approach for treating these GSDs. Using a fluorescence polarization assay designed to screen for inhibitors of the key glycogen synthetic enzyme, glycogen synthase (GS), we identified a substituted imidazole, (rac)-2-methoxy-4-(1-(2-(1-methylpyrrolidin-2-yl)ethyl)-4-phenyl-1H-imidazol-5-yl)phenol (H23), as a first-in-class inhibitor for yeast GS 2 (yGsy2p). Data from X-ray crystallography at 2.85 ?, as well as kinetic data, revealed that H23 bound within the uridine diphosphate glucose binding pocket of yGsy2p. The high conservation of residues between human and yeast GS in direct contact with H23 informed the development of around 500 H23 analogs. These analogs produced a structure-activity relationship profile that led to the identification of a substituted pyrazole, 4-(4-(4-hydroxyphenyl)-3-(trifluoromethyl)-1H-pyrazol-5-yl)pyrogallol, with a 300-fold improved potency against human GS. These substituted pyrazoles possess a promising scaffold for drug development efforts targeting GS activity in GSDs associated with excess glycogen accumulation.

Full text of DOI:10.1021/acs.jmedchem.9b01851

Subscriber access provided by Murdoch University Library
Article
Discovery and development of small-molecule inhibitors of glycogen synthase
Buyun Tang, Mykhaylo S. Frasinyuk, Vimbai M Chikwana, Krishna K Mahalingan, Cynthia
A Morgan, Dyann M Segvich, Svitlana P. Bondarenko, Galyna P Mrug, Przemyslaw
Wyrebek, David S. Watt, Anna A DePaoli-Roach, Peter J. Roach, and Thomas D Hurley
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.9b01851 • Publication Date (Web): 05 Mar 2020
Downloaded from pubs.acs.org on March 6, 2020
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery and development of small-molecule
inhibitors of glycogen synthase
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
Buyun Tang1, Mykhaylo S. Frasinyuk2,3, Vimbai M. Chikwana1, Krishna K. Mahalingan1,
Cynthia A. Morgan1, Dyann M. Segvich1, Svitlana P. Bondarenko3, Galyna P. Mrug2,3,
Przemyslaw Wyrebek4, David S. Watt4-6, Anna A. DePaoli-Roach1, Peter J. Roach1 and Thomas
D. Hurley1*
1Department of Biochemistry and Molecular Biology, Indiana University School of Medicine,
Indianapolis, Indiana 46202, United States; 2V. P. Kukhar Institute of Bioorganic Chemistry and
Petrochemistry, NAS of Ukraine, Kyiv 02094, Ukraine; 3National University of Food
Technologies, Kyiv 01601, Ukraine; 4Department of Molecular and Cellular Biochemistry,
University of Kentucky, Lexington, Kentucky 40506, United States; 5Center for Pharmaceutical
Research and Innovation, College of Pharmacy, University of Kentucky, Lexington, Kentucky
40536, United States; 6Lucille Parker Markey Cancer Center, University of Kentucky,
Lexington, Kentucky 40536, United States
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 48
1
2
3
4
ABSTRACT
5
6
7
8
The over-accumulation of glycogen appears as a hallmark in various glycogen storage diseases
(GSDs), including Pompe, Cori, Andersen and Lafora disease. Accumulating evidence suggests
that suppression of glycogen accumulation represents a potential therapeutic approach for
treating these GSDs. Using a fluorescence polarization assay designed to screen for inhibitors of
the key glycogen synthetic enzyme, glycogen synthase (GS), we identified a substituted
imidazole, (rac)-2-methoxy-4-(1-(2-(1-methylpyrrolidin-2-yl)ethyl)-4-phenyl-1H-imidazol-5-
yl)phenol (H23), as a first-in-class inhibitor for yeast glycogen synthase 2 (yGsy2p). Data from
X-ray crystallography at 2.85 Å, as well as kinetic data, revealed that H23 bound within the
UDP-glucose binding pocket of yGsy2p. The high conservation of residues between human and
yeast GS in direct contact with H23 informed the development of around 500 H23 analogs.
These analogs produced a structure-activity relationship (SAR) profile that led to the
identification of a substituted pyrazole, 4-(4-(4-hydroxyphenyl)-3-(trifluoromethyl)-1H-pyrazol-
5-yl)pyrogallol, with 300-fold improved potency against human GS. These substituted pyrazoles
possess a promising scaffold for drug development efforts targeting GS activity in GSDs
associated with excess glycogen accumulation.
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
2
ACS Paragon Plus Environment

Page 3 of 48
Journal of Medicinal Chemistry
1
2
3
4
INTRODUCTION
5
6
7
8
Glycogen is a branched polymer of glucose, composed of chains thirteen residues long on
average, arranged in layers to form a molecule of up 107 daltons, corresponding to 55,000
glucose units. The biosynthesis of glycogen is a highly conserved, complex and coordinated
process that involves two key events. Glycogen synthase (GS) lengthens linear chains through
the progressive addition of -1,4-linked glucose residues to the non-reducing end of the polymer,
and glycogen-branching enzyme (GBE) catalyzes the intramolecular transfer of seven glucose
residues from the end of a linear chain to a C-6 hydroxyl group to produce -1,6 branch points
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
for further elongation1,2,3,4
.
GS is the rate-limiting enzyme for glycogen biosynthesis5. Higher eukaryotes have two isoforms,
GYS1 and GYS26,7, that share approximately 70% sequence identity at the amino acid level and
display the greatest variation within the N- and C-terminal sequence extensions where regulatory
sites are located. The two GS isoenzymes are differentially expressed: GYS2 is exclusively
expressed in the liver, while expression of GYS1 is highest in skeletal muscle but is also
expressed in most other tissues. These two principal storage tissues for glycogen serve distinct
physiological roles. Glycogen in the liver plays an essential, regulatory role in glucose
homoeostasis, but glycogen in other tissues acts as an intracellular energy reserve such as during
muscle contraction8. Similar to higher eukaryotes, Saccharomyces cerevisiae has two genes
encoding GS, GSY1 and GSY2, with Gsy2p as the predominant isoform9. GS undergoes
activation allosterically through glucose-6-phosphate (G6P) binding and inactivation by
phosphorylation at a number of sites through the action of multiple protein kinases, including
glycogen synthase kinase-3 (GSK-3) in mammals10,11. After transport into cells, glucose
undergoes phosphorylation to G6P and enters one of several metabolic pathways. Increased
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 48
1
2
3
4
5
6
7
8
9
levels of intracellular G6P stimulates GS activity that in turn drives the incorporation of glucose
into glycogen in a feed-forward mechanism12. The dephosphorylation of GS is mediated by
forms of type 1 protein phosphatase (PP1), the catalytic subunit of which is held in close
proximity to the glycogen granule and GS by a family of glycogen targeting subunits, including
protein-targeting-to-glycogen (PTG), an indirect activator of GS13.
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
Glycogen metabolism constitutes a key pathway in living cells that regulates systemic carbon or
energy allocation14. A number of diseases are associated with abnormal glycogen metabolism
including type 2 diabetes (T2D) and glycogen storage diseases (GSDs)15,16,17. Defects in
enzymes directly involved in either glycogen synthesis or degradation18 are the basis for more
than 12 different GSDs, and glycogen over-accumulation is characteristic of most GSDs. In
Pompe disease (GSD2), a deficiency of the enzyme, acid--glucosidase (GAA), leads to
lysosomal glycogen accumulation in many tissues including skeletal, cardiac, and smooth
muscle19. In Cori disease (GSD3) and Andersen disease (GSD4), defects in the glycogen
debranching enzyme (AGL) and glycogen branching enzyme (GBE1), respectively, result in the
accumulation of glycogen with abnormal structure20,21. Lafora disease is a fatal progressive
myoclonus epilepsy accompanied by neurodegeneration for which the presence of abnormal
glycogen inclusions known as Lafora bodies22 are the hallmark. Lafora bodies consist of poorly
branched, hyperphosphorylated, and insoluble forms of glycogen that occur in neuronal, muscle
and other tissues23,24,25. Notably, impaired glycogen metabolism and consequent glycogen
accumulation was also found to be pathological for β-cell dysfunction in T2D16,17
.
A common finding with excess glycogen accumulation is an association with impaired
autophagy and dysregulated mitochondrial metabolism. These derangements often lead to cell
death and early-onset lethal, disease progression in GSD-affected patients16,26,27. Currently, there
4
ACS Paragon Plus Environment

Page 5 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
are no effective treatments for GSDs that ameliorate all cellular and organ dysfunction. For
example, in Pompe disease, the administration of a genetically engineered enzyme (i.e., Enzyme
Replacement Therapy, ERT) that rescues the defective GAA in peripheral tissues fails to reverse
the neurological defects. Partial suppression of glycogen synthesis through inhibition of GS
activity represents an alternative, potentially effective strategy for treating various types of GSDs
where ERT may not provide a complete solution. For instance, mouse models of Lafora disease
that lack either of the causative genes, EPM2A or EPM2B, recapitulate aspects of the patient
phenotype, in that they accumulate polyglucosan bodies and misfolded proteins, display
increased endoplasmic reticulum stress, and show signs of neurodegeneration23,27. Such animal
models provided a venue for disrupting either PTG or GYS1 in these EPM2A(-/-) and EPM2B(-/-)
knock-out mice as a mechanism for diminishing glycogen accumulation, Lafora body formation
and the associated neurological and epileptic symptoms27,28,29,30. Inhibition of GS activity via
suppression of mTOR signaling also increases the effectiveness of treatments for Pompe disease
in conjunction with ERT31. Moreover, a recent report has shown that GYS2 inhibition with
RNAi prevents liver injury in mouse models of GSDs32. Although human clinical characteristics
may vary from animal phenotypes, the availability of mouse models and the appearance of
promising therapeutic approaches suggest the feasibility of using small-molecule interventions to
treat GSDs. We now report efforts to develop small-molecule inhibitors of human GS based on a
multicomponent study involving high-throughput screening (HTS), X-ray crystallography and
extensive SAR development work that led to substituted pyrazoles with low micromolar,
inhibitory IC50 values against human glycogen synthase 1 (hGYS1).
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
RESULTS
Development of fluorescence polarization assay for HTS
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 48
1
2
3
4
5
6
7
8
9
The standard radiochemical assay for GS activity utilizes a time-consuming 14C-glucose
incorporation assay (i.e., UDP-[U-14C]glucose and glycogen(n) affords [14C]glycogen(n+1) and
UDP) that involves appropriate safety measures for handling radioisotopes as well as multiple
transfer and washing steps to remove residual substrates33. The combination of laboratory effort,
cost, safety and waste disposal favored the development of a fluorescence polarization (FP) assay
for HTS as a sensitive, inexpensive, and rapid alternative34. We reserved the classical
radiochemical assay for subsequent studies to confirm the hits emerging from the FP assay. An
FP assay utilizes a fluorescently-labeled tracer molecule that binds to GS to form a slower
rotating complex than fluorophore alone and that leads to enhanced polarization of the emitted
light. Incubation of the GS-fluorophore complex with small-molecules that displace the
fluorescent tracer produces greater rotational motion of fluorophore than that of the GS-
fluorophore complex and causes relative depolarization of the emitted light. Using this assay
system as a vehicle for HTS, we identified small-molecules that reduce the fluorescence
polarization signal as GS-interacting agents.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The development of any FP assay involves the design and synthesis of a functional, fluorescent
probe. Although G6P is a well-known, allosteric activator of GS, the regioselective modification
of similarly reactive hydroxyl groups in G6P presented an unwelcome challenge. Because
glucosamine-6-phosphate (GlcN6P) also activated GS35 and has a single, reactive amine group,
its selection as a platform on which to incorporate a fluorophore was an attractive alternative to
G6P. Using the 14C-glucose incorporation assay, we confirmed that either G6P or GlcN6P
activated hGYS1 with AC50 values of 1.6
0.1 mM or 5.9
0.1 mM, respectively
±
±
(Supplemental Figure 1A). With direct evidence in hand that GlcN6P activated hGS, we
synthesized a fluorophore-modified GlcN6P (aka GlcN6P-fluorescein-5-Ex) using GlcN6P and
6
ACS Paragon Plus Environment

Page 7 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
the N-hydroxysuccinimidyl ester of fluorescein-5-Ex (Figure 1A), purified the fluorophore by
high-performance liquid chromatography (HPLC) and confirmed the expected molecular mass
by mass spectrometry in which it displayed a M+H+ peak at m/z 735.1267 consistent with
[C31H31N2O15PS + H]+ with calculated exact mass within acceptable limits (i.e., calculated m/z
735.1256) (Supplemental Figure 1B).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
We examined the binding of the fluorophore, GlcN6P-fluorescein-5-Ex, to GS by adding the
tracer to varying concentrations of yGsy2p to generate a saturation binding curve and calculate a
dissociation constant of 7.6
0.7 μM (Figure 1B). Furthermore, a competitive displacement by
±
G6P (Kd
70.9
3.6 µM) demonstrated that the tracer molecule bound to the G6P allosteric
±
=
site (Figure 1C). For comparison, we determined yGsy2p activation by G6P using the standard
14C-glucose incorporation assay which yielded an AC50 of 98.1
3.1 µM (Supplemental
±
Figure 1C). We also tested the binding of GlcN6P-fluorescein-5-Ex and displacement by G6P
using hGYS1. While the GlcN6P-coupled fluorophore and the fluorophore alone bound to
hGYS1, G6P was unable to displace either compound, a finding that suggested nonspecific
binding of the fluorophore to hGYS1. This reason, combined with the fact that production of
human GYS1 in insect cells results in a heavily phosphorylated enzyme that requires very high
concentrations of G6P for activity measurements, makes the full-length recombinant human
enzyme unsuitable for an assay designed to displace G6P. However, since the fluorescent probe
could be fully displaced from yGsy2p through competition with G6P, we used yGsy2p and not
hGYS1 for HTS. The high conservation of residues within the G6P allosteric site, the active site,
and overall protein sequence identity (~55%) between yGsy2p and hGYS1 further supported the
decision to use the yeast enzyme as a screening surrogate for the human enzyme (Supplemental
Figure 2). The robustness of the FP assay for HTS was assessed by determining the Z′-factor, a
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 48
1
2
3
4
5
6
7
8
9
parameter reflective of both the signal dynamic range and data variability36. While the ideal Z′-
factor is 1, a Z′-factor between 0.5 and 1 is considered excellent and suitable for screening
assays. The FP values for the positive controls (maximal binding, without G6P) and negative
controls (fully displaced tracer, with G6P) were 109.23 ± 1.97 and 26.48 ± 1.93 (mean ± SD)
from the 384-wells, respectively. The Z′-factor for this assay was determined as 0.86, indicating
that this assay is well suited for HTS (Figure 1D).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. Development of fluorescence polarization assay for HTS. (A) 2-(6-Hydroxy-3-oxo-
3H-xanthen-9-yl)-5-(2-((3-oxo-3-(((2R,3R,4R,5S,6R)-2,4,5-trihydroxy-6-
(hydroxymethyl)tetrahydro-2H-pyran-3-yl)amino)propyl)thio)acetamido)benzoic acid (β-anomer
of GlcN6P-fluorescein-5-Ex). (B) Saturation binding isotherm for GlcN6P-fluorescein-5-Ex
8
ACS Paragon Plus Environment

Page 9 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
binding to yGsy2p. The binding affinity was determined by adding GlcN6P-fluorescein-5-Ex to
a final concentration of 20 nM in the presence of varying yGsy2p concentrations (0 to 50 µM).
7
8
9
Kd= 7.6
0.7 µM. Averages of triplicate assays ± SEM are shown. (C) Displacement 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
GlcN6P-fluorescein-5-Ex binding to yGsy2p by G6P. A mixture of yGsy2p and GlcN6P-
fluorescein-Ex was added to various G6P concentrations (0.68 µM to 40 mM). Kd= 70.9
3.6
±
µM. Averages of triplicate assays ± SEM are shown. (D) Z′-factor determination: fluorescence
polarization values of the bound (-G6P), and the free (+2 mM G6P) GlcN6P-fluorescein-5-Ex
are shown, for the 16 rows with each row containing 12 columns of either the bound or free state.
The graphs depict a representative experiment from at least three independent experiments.
Hits identification and validation
We used the aforementioned FP displacement assay to screen the 50K ChemBridge Diversity
library at a final assay concentration of 10 μM against the yGsy2p enzyme. The HTS was
adapted to the 384-well plate format using three columns for the DMSO negative controls and
one column for unlabeled G6P as the positive control, resulting in a total of 157 screening plates
(Supplemental Table 1 and 2). We used a Z’-score threshold of -1, corresponding to a
separation of 3 standard deviations between μc (means of the control DMSO signal) and μs
(means of the library sample signal) as a signal cutoff (Supplemental Table 2). These standards
produced 117 hits and an overall 0.23% hit rate. Initial stock availability led to re-purchasing of
110 compounds (designated H1-H110) from ChemBridge and validation of their activity using
the standard 14C-glucose incorporation assay at 100 μM concentration. This approach validated
16 hits with greater than 40% inhibition of yGsy2p activity (Figure 2A).
Since the yeast yGsy2p enzyme was used as a surrogate for the human hGYS1 enzyme, we next
examined if the 16 validated compounds were also active as inhibitors of hGYS1 activity. To
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 48
1
2
3
4
5
6
7
8
9
accomplish this necessary confirmation, we redesigned the hGYS1 construct to create a
constitutively active enzyme through deletion of the C-terminal phosphorylation domain
(residues 635-737) and through the substitution of the two N-terminal phosphorylation sites
(Ser8, Ser11) with Asn residues (designated hGYS1634S8,11N) (Figure 2B). Purification of
this truncated form of GS in significant quantity led to an enzyme that exhibited an activation
state of ~0.2, a ratio of GS activity in the absence over the presence of G6P as an index of the
phosphorylation state of the enzyme37. Unlike the heavily phosphorylated, full-length enzyme
produced in insect cells that had an activation state of <0.0138, this truncated form of GS was
well suited for validation assays designed to look for inhibitors because the enzyme was neither
overly inhibited by phosphorylation nor rendered insensitive to the effects of G6P. At a test
concentration of 100 μM, only 1 hit, namely a substituted imidazole, (rac)-2-methoxy-4-(1-(2-
(1-methylpyrrolidin-2-yl)ethyl)-4-phenyl-1H-imidazol-5-yl)phenol (H23), demonstrated greater
than 20% inhibition using this truncated hGYS1 (Figure 2C and 2D).
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
10
ACS Paragon Plus Environment

Page 11 of 48
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
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 48
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Hits identification and validation. (A) In the FP assay, setting a Z’-score threshold of
-1 gave 117 initial hits, out of which 110 (designated H1-H110) were re-purchased and tested.
The blue column represents the percentage of fluorescence polarization signal to control. The
hits were screened at 10 μM in single measurement. The orange column is the percentage of
yGsy2p activity to control measured through 14C-glucose incorporation assay. The hits were
tested at 100 μM. Averages of triplicate assays ± SEM are shown. (B) Schematic representation
of yGsy2p and hGYS1 protein. Phosphorylation residues are shown in red. To generate
hGYS1Δ634S8,11N mutant, hGSY1 was truncated at position 634 with mutations of Ser8/11 to
Asn to eliminate potential phosphorylation sites. (C) Percentage of hGYS1Δ634S8,11N activity
to control measured through 14C-glucose incorporation assay. The hits were tested at 100 μM.
Averages of triplicate assays ± SEM are shown. (D) Screening triaging strategy.
12
ACS Paragon Plus Environment

Page 13 of 48
Journal of Medicinal Chemistry
1
2
3
4
Crystal structure of the H23-yGsy2p complex
5
6
7
8
Although there are no mammalian GS crystal structures, the structures of two eukaryotic
enzymes, namely yGsy2p from Saccharomyces cerevisiae10 and GS from Caenorhabditis
elegans39 encouraged our undertaking crystallization experiments. As previously described, yeast
Gsy2p existed as tetramer in the crystal structure, and each subunit contained two Rossmann-fold
domains with the catalytic site in the interdomain cleft10. In the crystal packing environment in
crystals of the activated form of GS, one subunit appeared 13.3º more closed than the other three
“open” subunits40. We successfully determined the crystal structure of the H23-yGsy2p complex
to a resolution of 2.85 Å (Table 1), and we observed H23 binding in three of the four subunits,
all of which corresponded to the “open” domain positions (Figure 3B). The structure showed
H23 bound within the active site of GS in a location that overlapped with the binding site for
UDP-glucose (UDPG)40 (Figure 3C and 3D). Structural alignment of the R589/592A2·UDP
complex to the WT·H23 complex using their Cα carbons generated an overall root mean square
deviation (rmsd) of 0.36 Å, indicating a high degree of similarity. The binding of H23 to yGsy2p
was mediated by hydrogen-bond formation between the phenolic hydroxyl group of H23 and the
nitrogen backbone of Leu481; hydrophobic interactions of the 2-methoxyphenol sandwiched
between Tyr492 and Phe480; van der Waals interactions of the benzene group with the side
chains of Phe480 and Arg320; and additional van der Waals interactions of the N-
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
methylpyrrolidine moiety with Tyr492, Thr514, and Glu517 (Figure 3D).
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 48
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
14
ACS Paragon Plus Environment

Page 15 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 3. Crystal structure of the H23-yGsy2p complex (PDB: 6U77). (A) Chemical structure
of H23. (B) Ribbon diagram representation of the crystal structure of H23-yGsy2p complex.
H23 is represented by space-filling models in cyan, and binds to three subunits of yGsy2p which
are colored differently. (C) The electron density for H23 prior to the inclusion of the ligand in
refinement. The map shown is the original unbiased 2Fo-Fc map contoured at 1 standard
deviation. (D) Stick representation of the superposed UDP (purple) and H23 (yellow) structures
and their interactions with the surrounding amino acids.
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
Kinetic characterization of H23
In a study of the inhibitory potential for H23 against yeast and human GS using the standard 14C-
glucose incorporation assay, H23 exhibited IC50 values of 280 µM and 263 µM in the absence
and presence, respectively, of G6P for yGsy2p (Figure 4A). The similarity in IC50 values
indicated H23 was not in direct competition with G6P. Under subsaturating G6P concentrations,
the IC50 values of H23 against either hGYS1634S8,11N or wild-type hGYS1 were 161 μM and
875 µM, respectively (Figure 4B). Even though the presence or absence of G6P did not impact
H23 potency, the activity state of the human GS enzyme had a five-fold effect on the binding of
H23.
To understand H23 mode of inhibition, we performed co-variation experiments by
simultaneously varying the concentrations of UDPG at different fixed concentrations of H23 and
fitting the kinetic data against the competitive, non-competitive and uncompetitive inhibition
equations. For yGsy2p, the inhibition data with H23 was consistent with a competitive mode of
inhibition with respect to varied UDPG, with Ki values of 370 M in the presence of G6P, and
290 M in the absence of G6P. UDP displayed a similar competitive mode of inhibition with a
Ki value of 350 M (Figure 4C).
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 48
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. Kinetic characterization of H23. (A) Inhibition of H23 to yGsy2p in the absence or
presence of G6P. (B) Inhibition of H23 to hGYS1Δ634S8,11N and hGYS1 wild type. All IC50
curves represent one of three experiments performed using triplicate measurements for each
condition, with mean ± SEM shown. (C) Michaelis-Menten curves fit to the competitive
inhibition equation for varied H23 versus UDPG. The values of goodness of fit (R2) are shown
for all equations evaluated. H23 has a Ki of 370 ± 30 M in the presence of G6P and a Ki of 290
± 20 M in the absence of G6P. UDP is used as a positive control for competitive inhibition
against varied UDPG for yGsy2p, which displayed a Ki of 350 ± 10 M. The reported Ki values
are the mean ± SEM from three independent experiments in duplicate.
16
ACS Paragon Plus Environment

Page 17 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Development of structure-activity relationships for analogs of H23 toward hGYS1
The overall protein sequence alignment demonstrates 55% sequence identity between yGsy2p
and hGYS1. Because the amino acids in the binding site for H23, including Arg320, Phe480,
Leu481, Tyr492, and Glu517 in yGsy2p, are completely conserved across yeast and human
species (Supplemental Figure 2), the structural information derived from H23-yGsy2p crystal
structure provided a useful guide for structure-activity studies focused on hGYS1. We examined
a total of 491 analogs that shared at least 50% structural similarity with H23, and we tested their
activities against yGsy2p, hGYS1634S8,11N, and wild-type hGYS1 using the 14C-glucose
incorporation assay. Our initial kinetic studies showed that a five-membered, heteroaryl (HA)
core with a phenyl group at the A1 position, and a second, vicinal phenyl group were essential
elements of H23 (Table 2). Consequently, analog development modified the HA core, A1 and
substructures at R1-R7 positions shown in the structure in Table 2. In this study, all H23 analogs
possessed one of the following HA cores: imidazole (designated as HA1), pyrrole (designated as
HA2) or pyrazole (designated as HA3). Active compounds appeared in structures with any of
these three cores, but the most potent compounds had a pyrazole (HA3) scaffold. Commercial
libraries were the source of HA1 and HA2 compounds in Table 2, and a three-step synthesis
provided the HA3 compounds as outlined in Scheme 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
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 48
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
Scheme 1. Synthesis of pyrazoles 11-30. Reagents and conditions: (a) (i) arylacetonitrile,
BF3Et2O, HCl (gas), rt, 6-8 h, (ii) H+/H2O, 0.5-2 h; (b) (i) arylacetic acid, BF3Et2O, 80-90 0C, 2
h, (ii) H2O; (c) (i) DMF, BF3Et2O, POCl3, 50-60 0C, 2 h, (ii) H2O; (d) (i) Ac2O, KOAc, reflux, 8
h, (ii) H2SO4, EtOH, reflux, 0.5 h; (e) (i) (CF3CO)2O, pyridine, rt, 48-120 h, (ii) H2O; (f)
N2H4H2O, reflux, 0.5-6 h.
A step-by-step program of modifying positions in H23 involved changes designed either to
augment or diminish interactions with substituents in H23 and the adjoining residues in yGsy2p.
Briefly, starting from the imidazole core (HA1), we determined that the R1 hydroxyl group that
formed a hydrogen bond with the peptide nitrogen of Leu481 was essential to retain good
inhibitory activity (i.e., compare H23 versus 1 in Table 2). The addition of an extra methylene
18
ACS Paragon Plus Environment

Page 19 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
unit to the R7 chain eliminated inhibitory potential (4 versus 3). In a similar fashion, the most
potent of the pyrrole HA2 compounds also had a hydroxyl group at R1 position (7 versus 9 or 10).
Interestingly, substitution of a carboxyl group at R7 position improved the potency to hGYS1 by
2-fold (IC50 values for H23 versus 7), a finding that suggested different conformational binding
modes for H23 and 7 for hGYS1.
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
Because none of the analogs within the HA1 or HA2 series generated noteworthy improvements
in potency, we turned to the pyrazole (HA3) series. Within the HA3 series, SAR studies
demonstrated the importance of a hydrogen bond acceptor at the R1 position as displayed in the
following order of potency: -OH > -NO2 > -CN > -F > -H > -Cl (i.e., 23 versus 27, 25, 19, 21 and
15). When a chlorine is present at the R1 position, adding a second chlorine substituent at the R2
position improved potency (12 versus 11 or 14 versus 13 or 17 versus 15). The most potent
compounds in this series had a pyrrogallol group at the R4-6 positions, and the hydroxyl groups
on this substituent conferred inhibition to GS even if the R1 position was non-optimal. For
instance, 11 was 16-fold more potent than H23. Strikingly, the meta-hydroxyl group (R5) was
critical to binding. Substitution of this hydroxyl groups with just hydrogen decreased potency
significantly (16 versus 15 or 28 versus 27). SAR studies also indicated that substituents at the
R7 position had an effect on activity in the following order of -CF3 > -H > -CH3 (15 or 13 versus
11; 17 or 14 versus 12). In summary, analog development initiated from the substituted
imidazole H23 led to a substituted pyrazole, namely 4-(4-(4-hydroxyphenyl)-3-(trifluoromethyl)-
1H-pyrazol-5-yl)pyrogallol (23) that had an in vitro IC50 value of 2.75 µM, an improvement in
potency toward hGYS1 of more than 300-fold.
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 48
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
20
ACS Paragon Plus Environment

Page 21 of 48
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
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 48
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
Kinetic characterization of compound 23
In order to understand the molecular features of our most potent analog, compound 23, we
examined the kinetic characteristics of its interaction with GS (Figure 5A). Compound 23 has a
competitive mode of inhibition for wild-type hGYS1, with a Ki value of 1.31 ± 0.14 µM (Figure
5B). In light of the SAR and our kinetic results, we hypothesized that the binding of 23 to the
active site of yGsy2p resembled the binding of H23 in the active site. In this model, the hydroxyl
group at R1 position formed a hydrogen bond with the nitrogen backbone of Leu481. The three
hydroxyl groups in the pyrogallol subunit formed hydrogen bonds with Thr514 or Glu517.
Additional structural flexibility of hydrogen-bond formation depended on the relative position of
23 in the binding pocket. As our SAR study showed, the hydroxyl group at R5 position was
important to confer inhibition, and an additional hydrogen-bond formation between Arg320 and
the pyrazole ring may also strengthen binding, particularly when an electron-withdrawing
22
ACS Paragon Plus Environment

Page 23 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
trifluoromethyl group was present, and diminish binding when electron-donating methyl group
was present in the same position in the pyrazole ring (Figure 5C). To validate whether
compound 23 bound within the active site as modelled, we generated mutation within the active
site where Y513 was mutated to L513. This mutation, Y513L, did not abolish GS activity.
Unlike other active site mutations that almost completely eliminated enzyme activity, the Y513L
mutant decreased GS catalytic activity by only 10-fold. Consistent with our modelled mode of
binding, we found that inhibition by 23 toward Y513L was compromised compared to wild type
(Figure 5D).
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
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 48
1
2
3
4
5
6
7
8
9
Figure 5. Kinetic characterization of 23. (A) Dose response curves for parent compound H23
and the most potent analog 23 against wild-type hGYS1. (B) Michaelis-Menten curve of
competitive inhibition for 23 versus UDPG. The values of goodness of fit (R2) are shown for all
equations. 23 shows a Ki of 1.31 ± 0.14 µM under the tested condition. Ki value for 23 is the
mean ± SEM from three independent experiments performed using duplicate measurements for
each condition. (C) Hypothetical binding model of 23 to GS in the active site. The hydroxyl
group at R1 position forms a hydrogen bond with the nitrogen backbone of Leu481. The three
hydroxyl groups flanking the benzene form hydrogen bonds with Tyr514 or Glu517. An
additional hydrogen bond is formed between Arg320 and the pyrazole ring, which is
strengthened when electron-withdrawing group (like CF3) is present and diminished when
electron-donating group (like CH3) is present. Manual docking of 23 in the active site was
performed based on SAR and mode of inhibition studies. (D) Validation of 23 binding in the
active site. Compound 23 showed decreased potency against the active-site mutant Y513L
compared with wild-type yGsy2p. All IC50 curves represent one of three experiments performed
using triplicate measurements for each condition, with mean ± SEM shown.
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
Inhibition of glycogen synthase activity in cell lysates
We next examined these substituted pyrazoles as inhibitors of GS activity in cell lysates.
Cultured cells normally do not accumulate large amounts of glycogen, largely due to the high
levels of phosphorylation and the resulting low activity state of GS26. Lysates from two cell lines
were prepared for GS activity measurement: the HEK293-PTG overexpressing PTG, a regulatory
subunit of protein phosphatase 1, that recruits the phosphatase to glycogen where it promotes the
dephosphorylation and activation of GS41,42, and glycogen accumulation; and the Rat-1
fibroblasts which have low GS activity and glycogen as previously described43,44. The HEK293-
24
ACS Paragon Plus Environment

Page 25 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
PTG cells have a seven-fold increase in the GS activity ratio from 0.02 to 0.15. To determine GS
activity in lysates and optimize the conditions for measurement, we measured the incorporation
of 14C-glucose into glycogen at 0.2 mM UDPG and 1 mM G6P initially as a function of different
lysate concentrations. We observed a linear increase in GS activity within the range of 0.075-1.5
mg/ml lysate. Under the conditions of the assay, the lysate from the HEK293-PTG cells had 10-
fold more activity than the Rat-1 cell lysate (Figure 6A). To limit substrate utilization to under
10%, we used 0.15 mg/mL HEK293-PTG lysate and 0.75 mg/mL Rat-1 lysate with imidazole
H23, pyrrole 7 and pyrazoles 15, 19, 27 and 23. When tested at 100 μM, H23 and 7 did not
significantly inhibit GS activity in lysates. However, the remaining four substituted pyrazoles,
namely 15, 19, 27 and 23, reduced synthase activity in both HEK293-PTG and Rat-1 cell lysates
by >30%. Consistent with its greatest potency toward purified enzyme, pyrazole 23 exhibited
almost complete inhibition of synthase activity in lysates (Figure 6B). In summary, these
analogs targeted GS activity in the context of the glycogen particles present in cellular lysates
with potencies similar to those observed in purified enzyme preparations.
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 6. Inhibition of glycogen synthase activity in cell lysates. (A) Synthase activity in
HEK293-PTG and Rat-1 cell lysates in the presence of 0.2 mM UDPG and 1 mM G6P. The data
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 48
1
2
3
4
5
6
7
8
9
was fit to linear regression line with equations showing synthase activity rate (nmole/min, y-
value) under various lysate concentrations (mg/ml, x-value). The data represent averages of
triplicate assays ± SEM. (B) Percent of 14C-glucose incorporation to control (DMSO) by cell
lysates and recombinant hGYS1 in the presence of H23 and its analogs. For cell lysates and
recombinant hGYS1, 100 μM and 20 μM compounds were used respectively. Averages of
triplicate assays ± SEM are shown.
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
DISCUSSION
The suppression of glycogen accumulation emerged as an attractive therapeutic approach for
GSDs whose etiology derived from excessive glycogen storage. Animal models supported this
approach in which genetic or chemical depletion of glycogen alleviated disease symptoms in
models of Lafora disease27,28,29,30,45, Pompe disease31, and Cori disease32. Interestingly, recent
studies suggested that type 2 diabetes (T2D) might also be within the GSD spectrum, as
glycogen accumulation in pancreatic β cells under hyperglycemia contributed to the pathology of
β-cell dysfunction16,17
.
Only a few studies identified molecules that targeted glycogen accumulation, including the
widely-used glucose lowering T2D drug metformin46,47, mTORC1 inhibitor rapamycin that
indirectly suppress GS activity through signaling regulation31, GYS2 RNAi that mediates
enzyme reduction32, and the newly developed antibody-enzyme fusion that can degrade
polyglucosan45. However, these molecules either targeted glycogen synthesis in an indirect
manner, or present a challenge for the delivery of therapeutics to the central nervous system. In
this study, we adapted a high-throughput fluorescence polarization assay to screen directly for
small-molecule modulators of GS. A HTS of a commercial 50K chemical library identified 117
primary hits against yGsy2p, and through 14C-glucose incorporation assay only 1 hit was
26
ACS Paragon Plus Environment

Page 27 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
validated as hGYS1 inhibitor (Figure 2). The low translation between these two assays might lie
in the nature of these two systems, since fluorescence polarization assay is affinity-based
whereas the radiochemical assay is activity-based. In addition, the kinetic aspects of the two
enzyme systems differ as the impact of regulatory input on the enzyme results in different
outcomes where regulation of yeast GS by G6P primarily impacts kcat, whereas in mammalian
systems the impact is primarily on Km for substrate48,49. There is no evidence that the kinetic
steps in catalysis differ between the two enzymes, but it is clear that the manner in which the
rates for those individual steps are impacted by regulatory input does differ. A more sensitive
assay directly targeting hGYS1 is likely to offer better screening outcome. Nonetheless, our FP
assay provided a novel HTS assay as a starting point, and did indeed identify a competitive
inhibitor of UDPG, namely (rac)-2-methoxy-4-(1-(2-(1-methylpyrrolidin-2-yl)ethyl)-4-phenyl-
1H-imidazol-5-yl)phenol (H23), that bound within the active site of yeast yGsy2p. This outcome
was supported by both an X-ray structure and by kinetic studies using H23 and analogs
developed in a subsequent SAR study. The site of interaction was surprising because the assay
identified small-molecule inhibitors that displaced G6P binding. However, a close inspection of
the known, eukaryotic GS structures demonstrated that the UDPG and G6P binding sites resided
on opposing ends of the same alpha-helix, with Tyr492 stacked against the uracil ring of UDP
and His500 forming a hydrogen bond with the phosphate moiety of G6P (Supplemental Figure
3). Considering this structural proximity and the cooperative nature of the structural transitions in
GS, it was not surprising that binding at one site could transmit structural information to the
other site and that under the subsaturating conditions of our HTS assay, binding of compounds
within the active site could promote displacement of the fluorophore from the G6P 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
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 48
1
2
3
4
5
6
7
8
9
Humans have two isoforms of GS, namely hGYS1 and hGYS2. The fact that hGYS1 is
universally expressed in most tissues while hGYS2 is restrictively expressed in liver prompted us
to identify inhibitors against hGYS1 in the hope of diminishing brain glycogen stores, aberrantly
accumulated in some GSDs such as Lafora disease26. The high sequence homology between
hGYS1 and hGYS2 enzymes could make it difficult to develop isoform-specific inhibitors.
Nonetheless, inhibition of hGYS2 might aid in the therapeutic efficacy of GSDs with excessive
glycogen accumulation in liver that would ultimately lead to liver damage. Recently, Pursell et
al. showed that GYS2 inhibition with RNAi prevents liver injury in mouse models of Cori
disease and had no adverse effects32.
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
Mutations in GYS150 or GYS251 in rare GSDs lead to decreased glycogen in muscle and liver
respectively. Disruption of the mouse GYS1 gene resulted in 90% perinatal lethality, likely due
to cardiac developmental problems during embryogenesis, but the surviving mice were
ostensibly normal and lived normal lifetimes52. Disruption of GYS2 in mice largely mimicked
the phenotype of human GSD0 patients, namely tendencies to post-prandial hyperglycemia and
to hypoglycemia upon fasting but compatible with a relatively normal life51,53. Given that small-
molecule inhibition will not have the penetrance of genetic defects, inhibition of GS activity by
small-molecules is unlikely to elicit extreme phenotypes, and small-molecule inhibition remains
as a potentially valuable means of treating these devastating diseases.
CONCLUSIONS
In summary, this study described a strategy for developing a high-throughput FP assay for the
screening of small-molecule inhibitors of GS, the binding mode of a leading imidazole H23
validated through both X-ray crystallographic and kinetic data; and an SAR study leading to
analogs, such as pyrazole 23, with low micromolar potency. These outcomes suggest that
28
ACS Paragon Plus Environment

Page 29 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
targeting glycogen synthase with small-molecule inhibitors represents an attractive approach for
developing new therapeutics for diseases in the GSD family. Unlike other enzyme-based
strategies where complete inhibition is the ultimate objective, a partial reduction of GS activity
may be sufficient to alleviate unwanted and damaging levels of glycogen deposition in the neural
tissue of patients suffering from Lafora disease27,28,30. The challenges of developing any
therapeutic, nevertheless, remain the same, and the gulf that stretches between our identification
of a pyrazole inhibitor 23 with activity in the low micromolar range and a therapeutic candidate
is wide and deep. Initial efforts will involve microsomal studies to evaluate anticipated
pharmacokinetic concerns about unwanted redox reactions of pyrazole 23 to 4-(4H-pyrazol-4-
ylidene)- or 2,3-dihydroxy-4-(3H-pyrazol-3-ylidene)cyclohexa-2,5-dien-1-ones. We will use a
combination of synthesis (i.e., SAR studies) and computational modeling to identify analogs that
avoid this concern and retain desired physiochemical properties (i.e., water solubility;
bioavailability); we will evaluate any promising, new leading structures for potential toxicity
issues (e.g., hERG studies); and we will explore biotinylated analogs to confirm the specificity of
these pyrazole analogs for the desired target. These initial, encouraging results bode well for the
future development of small-molecule strategies to study and potentially treat GSDs.
EXPERIMENTAL SECTION
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
Materials
The 50K compound Diversity set library, the primary hits H1-H110, and analogs 1-4, 7-9 were
purchased from ChemBridge Corporation (San Diego, CA). Analogs 5, 6 and 10 were purchased
from Vitas-M Laboratory (Champaign, IL). The purity of these purchased compounds were
>
95% based on the spectra (either LC/MS or NMR) provided by the vendors. All pyrazoles were
characterized and validated by both LC/MS and NMR analyses.
29
ACS Paragon Plus Environment