Diversity-Oriented Synthesis as a Strategy for Fragment Evolution against GSK3β

Article abstract of DOI:10.1021/acsmedchemlett.6b00230

Traditional fragment-based drug discovery (FBDD) relies heavily on structural analysis of the hits bound to their targets. Herein, we present a complementary approach based on diversity-oriented synthesis (DOS). A DOS-based fragment collection was able to produce initial hit compounds against the target GSK3β, allow the systematic synthesis of related fragment analogues to explore fragment-level structure-activity relationship, and finally lead to the synthesis of a more potent compound.

Full text of DOI:10.1021/acsmedchemlett.6b00230

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Subscriber access provided by the University of Exeter
Letter
Diversity-oriented synthesis as a strategy
for fragment evolution against GSK3#
Yikai Wang, Jean-Yves Wach, Patrick Sheehan, Cheng Zhong, Chenyang Zhan, Richard
Harris, Steven C. Almo, Joshua A. Bishop, Stephen J Haggarty, Alexander Ramek, Kayla
Berry, Conor O'Herin, Angela N. Koehler, Alvin W Hung, and Damian Winston Young
ACS Med. Chem. Lett., Just Accepted Manuscript • DOI: 10.1021/acsmedchemlett.6b00230 • Publication Date (Web): 14 Jul 2016
Downloaded from http://pubs.acs.org on July 15, 2016
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.
ACS Medicinal Chemistry Letters 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.

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Page 1 of 6
ACS Medicinal Chemistry Letters
1
2
3
4
5
6
7
8
9
Diversityꢀoriented synthesis as a strategy for fragment evolution
against GSK3β.
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
Yikai Wang,[a] JeanꢀYves Wach,[a] Patrick Sheehan,[a] Cheng Zhong,[a] Chenyang Zhan,[b] Richard Harꢀ
ris,[b] Steven C. Almo,[b] Joshua Bishop,[a] Stephen J. Haggarty,[a] Alexander Ramek,[a] Kayla Berry,[a]
Conor O'Herin,[a] Angela N. Koehler,[a] Alvin W. Hung†*[a] and Damian W. Youngɪ*[a]
[a] Chemical Biology Program, The Broad Institute of Harvard and MIT, 415 Main St., Cambridge, MA, 02142.
[b] Department of Neurology & Psychiatry, Harvard Medical School and Massachusetts General Hospital, 185 Cambridge
St, Boston, MA, 02114.
[c] Department of Biochemistry, Albert Einstein College of Medicine 1300 Morris Park Avenue, Bronx, New York, 10461
Diversity Oriented Synthesis; Fragmentꢀbased drug discovery; Fragment growing; GSK3β
ABSTRACT: Traditional fragmentꢀbased drug discovery (FBDD) relies heavily on structural analysis of the hits bound to their
targets. Herein, we presented a complimentary approach based on diversityꢀoriented synthesis (DOS). This DOSꢀbased fragment
collection was able to produce initial hit compounds against the target GSK3β, allow the systematic synthesis of related fragment
analogues to explore fragmentꢀlevel SAR, and finally lead to the synthesis of a more potent hit compound.
The development of other methods to evolve and optimize
fragments13 without structural guidance would certainly exꢀ
Fragmentꢀbased drug discovery (FBDD) is an established
approach for generating therapeutic lead molecules.1ꢀ4 Key to
pand the generality of the FBDD approach. Toward this goal,
we herein describe the application of diversityꢀoriented synꢀ
the optimization process is the application of structureꢀbased
thesis (DOS) to discover and optimize fragments to the kinase
methods to guide the synthetic decisionꢀmaking of growing
fragment hits into more potent compounds.5ꢀ8 However, the
strong reliance on highly resolved fragmentꢀtarget structures
can be limiting. Targets that do not readily lend themselves to
structural guided approaches provide little recourse to
FBDD.9,10 A strategy for evolving more potent compounds
from fragments that does not completely depend on structureꢀ
based methods would be a useful tool within the FBDD arena,
making the approach more broadly applicable.
GSK3β.
Figure 1. A logical chemical synthesis strategy related to the
fragments would impact optimization and modification of fragꢀ
ment hits. Fragments constructed using highly modular syntheses
would enable rapid substitutions around a fragment core. Thereꢀ
fore, the generation of fragmentꢀSAR using modular fragment
syntheses would aid the fragment evolution process as a compleꢀ
mentary approach to structureꢀguided approaches.
There have been recent discussions related to evolving
fragment scaffolds without structural guidance. For example,
Konrat and coworkers described a strategy for using protein
metaꢀanalysis and ligandꢀbased NMR techniques to generate
leads against two targets without 3D structures.11 Similarly,
Viola and coworkers reported a systematic chemistry approach
to explore fragmentꢀlevel structureꢀactivity relationships.12
DOS involves the generation of skeletal, stereochemical,
regiochemical and appendage variation using a common synꢀ
thesis scheme.14ꢀ15 The build/couple/pair (B/C/P) strategy in
DOS was developed to guide, in a modular fashion, the synꢀ
thetic planning processes to efficiently generate diverse comꢀ
pounds.16ꢀ18 This produces a broader degree of variation in
molecular structure in a set of compounds compared to other
approaches. Importantly, this molecular diversity is produced
under the aegis of one general synthesis pathway, which sigꢀ
nificantly simplifies the synthetic planning and development
[a]
Dr Yikai Wang, Dr JeanꢀYves Wach, Patrick Sheehan, Dr Chen
Zhong, Alexander Ramek, Kayla Berry, Conor O’Herin, Dr Angela N.
Koehler, Dr Alvin W. Hung* , Dr Damian W. Young*.
Chemical Biology Program, The Broad Institute of Harvard and MIT,
7 Cambridge Center, Cambridge MA 02142.
Eꢀmail: whung@etc.aꢀstar.edu.sg; damiany@bcm.edu
Dr Joshua Bishop, Dr Stephen J.Haggarty
Department of Neurology & Psychiatry, Harvard Medical School and
Massachusetts General Hospital, 185 Cambridge St, Boston, MA,
02114
[b]
[c]
Dr Richard Harris; Dr Steve C. Almo
Department of Biochemistry, Albert Einstein College of Medicine
1300 Morris Park Avenue, Bronx, New York, 10461
ACS Paragon Plus Environment
Supporting information for this article is given via a link at the end of
the document.

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
ACS Medicinal Chemistry Letters
Page 2 of 6
throughout the discovery and lead generation phases. The merꢀ
its of the B/C/P approach have been demonstrated in various
highꢀthroughput screening (HTS) applications.19ꢀ21 Here, we
extend the concept of DOS towards the application of fragꢀ
mentꢀbaseddrug discovery (Figure 1).
Based on the B/C/P pathway described above, a series of
1
2
3
4
5
6
7
8
structurally similar compounds to parent fragment 1S was
synthesized and evaluated in DSF experiments. In parallel,
using ADPꢀGloTM assays, single point biochemical inhibitions
were also performed to rank compounds based on their inhibiꢀ
tion against GSK3β (at 1 mM). As shown in Figure 3b, fragꢀ
ment 1R (antipode of the initial hit 1S) showed a larger ꢁT
and inhibition than the original fragment hit 1S (1.7 °C and
46% inhibition at 1 mM, Figure 3b). The variation in thermal
stabilizations and inhibitions of both fragment enantiomers
against GSK3β provided early validation of fragment binding.
A set of 86 fragments was prepared and screened against
glycogen synthase kinase 3β (GSK3β), a serine threonine kiꢀ
nase which is aberrantly overexpressed in cancer and Alzꢀ
heimer’s disease.22ꢀ25 These fragments were derived from three
different pathways26ꢀ28 using the DOS B/C/P approach (Figure
2a). Accordingly, these modular routes could be exploited to
generate fragmentꢀlevel SAR from an identified hit.
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
Some preliminary SAR could also be observed from the set of
piperazinone compounds synthesized. Fused pyridyl rings to
the piperazinone scaffolds were found to stabilize GSK3β
more favorably than their phenyl counterparts. As shown in
Figure 3b, both enantiomeric compounds 4S and 4R had no
measurable effect in the DSF. It was also found that the posiꢀ
tion of the nitrogen in the pyridyl rings affected the thermal
stabilization of GSK3β. Importantly, depending on the posiꢀ
tion of the pyridineꢀN, the extent of variation in activity and
thermal stabilizations of the corresponding enantiomers was
different. We hypothesize that these regioisomers bind to
GSK3β through nonꢀidentical binding conformations, accountꢀ
ing for differences in stabilization between antipodes. Based
on the qualitative DSF and point inhibition results, fragment
1R was prioritized for further optimization.
Figure 2. a. Three different build/couple/pair pathways leading to
a library of skeletally and stereochemically diverse fragments. b.
Fragment hits from a DSF screening at 2.5 mM against GSK3β.
Compounds displaying a thermal shift ≥ 0.5 °C are shown.
As shown in Figure 2b, differential scanning fluorimetry
(DSF) was used as the primary screening method for detecting
fragment binding against GSK3β. Interestingly, three comꢀ
pounds from the same synthetic pathway (pathway III) disꢀ
played weak thermal stabilization (ꢁT) of 0.5 °C or greater
(Figure 2b).
The detailed synthesis of 1S is shown in Figure 3a.26 Here, the
build phase consisted of procuring commercially available
arylnitrofluorides and αꢀamino esters. The appropriate arꢀ
ylnitrofluorides were then coupled with amino esters using
SNAr reactions to yield 1S’. Subsequent reduction of the nitro
group under mild conditions afforded the aniline, which unꢀ
derwent concomitant cyclization (the pairing phase) to the
desired product 1S. The enantiomeric excess of compounds
produced by this synthetic pathway was verified by chiral
HPLC analysis to be greater than 95% in all cases, allowing
for activity comparison using compounds with equivalent enꢀ
antiopurity.
Figure 3. a. Synthesis of 1S as a representation of the B/C/P
pathway. b. DSF and biochemical assay results of analogues of
1S. c. Chemical modifications on 1R to identify key binding inꢀ
teractions and probe potential pocket for fragment growing. DSF
assays were performed at 2.5 mM unless otherwise indicated;
biochemical assays were performed at 1 mM.
ACS Paragon Plus Environment

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Page 3 of 6
ACS Medicinal Chemistry Letters
Pyridinyl amides are known hinge binding motifs against
point, we were equipped with a rational plan to evolve the
fragment into a more potent compound.
kinases.29 We postulated that this interaction is also present for
fragment 1R (Figure 3c). This hypothesis was further corroboꢀ
rated with the synthesis of compounds 8R and 9R. The deleꢀ
tion of the amide carbonyl in 1R abolished thermal stabilizaꢀ
tion of fragment 8R. Additionally, removal of the hydrogen
bond donor capacity by Nꢀmethylation of the amide nitrogen
(compound 9R) resulted in significantly lower stabilization of
GSK3β.
1
2
3
4
5
6
7
8
The B/C/P pathway was once again utilized to generate anaꢀ
logues of 1R with larger groups from the corresponding amino
ester starting materials (optimization phase). Both enantiomers
were synthesized to provide sideꢀbyꢀside comparisons of the
two stereogenic growth vectors. DSF and single point inhibiꢀ
tion assays were used to evaluate their binding interaction and
inhibitory activity, respectively.
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
H
N
H
N
H
N
H
N
H
N
NH
N
N
H
O
N
N
H
O
N
N
H
O
N
N
H
O
OH
N
N
H
O
1S
12S
ꢀT = 0±0.1 °C
25±4%
13S
14S
15S
ꢀT = 0±0.1 °C
19±1%
ꢀT = 0±0.1 °C
55±1%
ꢀT = 2.1±0.1 °C
ꢀT = 1.7±0.1 °C
78±1%
76±1%
H
N
H
N
H
N
H
N
H
N
NH
N
N
H
1R
O
N
N
H
12R
O
N
N
O
N
N
O
OH
N
N
O
H
H
H
13R
14R
15R
ꢀT = 1.0±0.1 °C
46±2%
ꢀT = 1.6±0.1 °C
57±3%
ꢀT = 1.6±0.1 °C
63±2%
ꢀT = 2.9±0.1 °C
ꢀT = 4.5±0.1 °C
88±1%
94±1%
Figure 5. A small set of compounds were synthesized, extending
from a growth vector identified from previous SAR. For ranking
purposes, DSF assays were performed at 1.25 mM for comparison
and point inhibition biochemical assays were performed at 1 mM.
K = 610 ±=
d
61ꢁM
As shown in Figure 5, the binding of larger aromatic groups
was accommodated in GSK3β similarly to aliphatic groups in
compounds 1R and 12R. Interestingly, an additional hydroxyl
group on the phenyl ring (compound 14R) led to a significant
increase in potency, leading us to surmise that a larger, elecꢀ
tronꢀrich aromatic group could be tolerated. Satisfyingly,
compound 15R with an indolyl substitution resulted in a
4.5 °C thermal stabilization when tested at 1.25 mM. Results
from the biochemical assay conducted at a single concentraꢀ
tion (1 mM) correlated well with the DSF results.
Figure 4 a. Saturation transfer differential (STD) NMR specꢀ
trum of 1R against GSK3β b. WaterLOGSY NMR spectrum
of 1R against GSK3β. Both NMR techniques further validated
binding of fragment 1R against GSK3β. c. Using ITC, 1R was
determined to bind against GSK3β with Kd = 610 µM (LE =
0.37).
At this stage, Saturation Transfer Differential (STD) and Waꢀ
terLOGSY NMR experiments30 were also performed to valiꢀ
date binding of 1R. As shown in Figure 4, both ligandꢀbased
NMR experiments show compound 1R binding unambiguousꢀ
ly towards GSK3β, corroborating thermal shifts results. With
the results in hand, Isothermal Titration Calorimetry (ITC)
was then used to quantitatively evaluate the equilibrium dissoꢀ
ciation constant (Kd) of fragment 1R. Compound 1R was
found to bind to GSK3β with Kd = 0.6 mM with a desirable
ligand efficiency (LE) of 0.37 (Figure 4c).
Finally, ITC was used to quantify the binding of 15R, the
compound which displayed the greatest thermal stabilization
against GSK3β. As shown in Figure 6a compound 15R bound
to GSK3β with Kd = 9 ꢁM, a greater than 60ꢀfold improveꢀ
ment in binding over 1R. Additionally, 15R was found to
inhibit GSK3β with an IC50 = 18 µM. The enantiomer 15S was
also tested and was ~18ꢀfold less active than 15R (IC50 = 322
µM, Figure 6b), a result that is reflected by the low thermal
shifts observed by DSF,
In an effort to evaluate the importance of the chiral center
residing in 1R, three analogues were prepared (Figure 3c).
Oxidation of 1R afforded compound 10, therefore removing
the chirality of 1R. This planar fragment resulted in diminꢀ
ished activity and low thermal stability as compared to 1R.
Installation of a gemꢀdimethyl group produced compound 11,
which did not provide thermal shifts equivalent to 1R. This
indicated that the binding pocket cannot easily accommodate
two hydrophobic groups on both faces of the fragment. Finalꢀ
ly, to probe the size of the binding pocket, a larger isopropylꢀ
containing analogue 12R was synthesized which showed a ꢁT
of 2.3 °C (0.6 °C more stable than 1R). Given this last obserꢀ
vation, we concluded that the chiral center, in particular with
the R configuration, was a potential growth vector to increase
potency. Therefore, from the fragment SAR generated at this
As shown in Figure 6c, 15R demonstrates 50ꢀfold selectivity
against GSK3β over the closely related kinase Cyclinꢀ
dependent Kinase 5 (CDK5). The selectivity profile of the
chiral compound 15R against GSK3β over CDK5 is an attracꢀ
tive starting point for the preparation of additional ligands
with more enhanced selectivity toward GSK3β. It is also
noteworthy that compound 15R has a low molecular weight of
278 g/mol (LE of 0.33) and this provides further growth opꢀ
portunities towards achieving greater potency.
In most FBDD efforts, Xꢀray analysis is performed at the
earliest stage to guide optimization chemistry. However, it
was our strategy to demonstrate that the modular B/C/P chemꢀ
istry would be advantageous towards optimizing a fragment
ACS Paragon Plus Environment