Fragment-based Differential Targeting of PPI Stabilizer Interfaces

Article abstract of DOI:10.1021/acs.jmedchem.9b01942

Stabilization of protein-protein interactions (PPIs) holds great potential for therapeutic agents, as illustrated by the successful drugs rapamycin and lenalidomide. However, how such interface-binding molecules can be created in a rational, bottom-up man

Full text of DOI:10.1021/acs.jmedchem.9b01942

Subscriber access provided by Uppsala universitetsbibliotek
Article
Fragment-based Differential Targeting of PPI Stabilizer Interfaces
Xavier Guillory, Madita Wolter, Seppe Leysen, Joao Filipe Neves, Ave Kuusk, Sylvia Genet,
Bente Somsen, John Morrow, Emma Rivers, Lotte van Beek, Joe Patel, Robert Goodnow, Heike
Schoenherr, Nathan Fuller, Qing Cao, Richard G. Doveston, Luc Brunsveld, Michelle R. Arkin,
M. Paola Castaldi, Helen Boyd, Isabelle Landrieu, Hongming Chen, and Christian Ottmann
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.9b01942 • Publication Date (Web): 05 Jun 2020
Downloaded from pubs.acs.org on June 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 46
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 46
1
2
3
4
5
6
7
8
9
Fragment-based Differential Targeting of PPI
Stabilizer Interfaces
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
Xavier Guillory1,†,‡, Madita Wolter1,‡, Seppe Leysen1,†,‡, João Filipe Neves2,3,†, Ave Kuusk1,4,
Sylvia Genet1, Bente Somsen1, John Kenneth Morrow5, Emma Rivers4, Lotte van Beek1, Joe
Patel6,†, Robert Goodnow6, Heike Schoenherr6, Nathan Fuller6, Qing Cao6, Richard G.
Doveston1,†, Luc Brunsveld1, Michelle R. Arkin5, Paola Castaldi6, Helen Boyd4, Isabelle
Landrieu2,3,*, Hongming Chen4,*, Christian Ottmann1,7*
1 Laboratory of Chemical Biology, Department of Biomedical Engineering and Institute for
Complex Molecular Systems, Eindhoven University of Technology, P. O. Box 513, 5600MB
Eindhoven, The Netherlands.
2 CNRS ERL9002 Integrative Structural Biology F-59000 Lille, France.
3 Univ. Lille, Inserm, CHU Lille, Institut Pasteur de Lille, U1167 - RID-AGE - Risk Factors and
Molecular Determinants of Aging-Related Diseases, F-59000 Lille, France
4 Hit Discovery, Discovery Sciences, Biopharmaceutical R&D, AstraZeneca, Gothenburg, 431
50 Mölndal, Sweden.
5 Department of Pharmaceutical Chemistry and Small Molecule Discovery Centre (SMDC),
University of California, San Francisco 94143, United States.
ACS Paragon Plus Environment
1

Page 3 of 46
Journal of Medicinal Chemistry
1
2
3
4
5
6
6 Oncology and Discovery Sciences, IMED Biotech Unit, AstraZeneca, Gatehouse Park,
Waltham, Massachusetts 02451, United States.
7
8
9
7 Department of Organic Chemistry, University of Duisburg-Essen, 47057 Duisburg, Germany.
KEYWORDS: FBDD, PPI stabilization, X-ray, 14-3-3, p53, TAZ
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
ABSTRACT. Stabilization of protein-protein interactions (PPIs) holds great potential for
therapeutic agents, as illustrated by the successful drugs rapamycin and lenalidomide. However,
how such interface-binding molecules can be created in a rational, bottom-up manner is a largely
unanswered question. We report here how a fragment-based approach can be used to identify
chemical starting points for the development of small-molecule stabilizers that differentiate
between two different PPI interfaces of the adapter protein 14-3-3. The fragments discriminately
bind to the interface of 14-3-3 with the recognition motif of either the tumor suppressor protein
p53 or the oncogenic transcription factor TAZ. This x-ray crystallography driven study shows that
the rim of the interface of individual 14-3-3 complexes can be targeted in a differential manner
with fragments that represent promising starting points for the development of specific 14-3-3 PPI
stabilizers.
ACS Paragon Plus Environment
2

Journal of Medicinal Chemistry
Page 4 of 46
1
2
3
4
INTRODUCTION
5
6
7
8
9
PPI modulation is one of the most promising and active fields in current drug discovery and
chemical biology.1 However, PPI modulation is in many instances understood synonymously with
PPI inhibition, disregarding the complementary strategy of stabilizing PPIs. This is despite the fact
that a multitude of natural products2 and a growing number of synthetic compounds3 validate the
feasibility and value of small-molecule PPI stabilization. Stabilizers of PPIs display a number of
advantages like their uncompetitive nature and their potential for higher specificity. They bind to
composite pockets at the interface of their target protein complexes that naturally can show a much
higher structural variability compared for example to active sites of enzymes or ligand binding
pockets of receptors. Consequently, the potential to develop selective PPI-stabilizing compounds
might be higher than with ligands that bind to conventional drug targets. An especially interesting
target for PPI modulation is the adapter protein 14-3-3. This “hub” protein, consisting of a family
of seven isoforms, interacts with several hundred partner proteins4 and regulates the activity of
many disease-related proteins like C-Raf5–7 (cancer), Tau8,9 (Alzheimer’s) and CFTR10,11 (cystic
fibrosis).
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
Here, we focused on the interaction of 14-3-3 with the transcriptional co-activator TAZ and the
tumor suppressor protein p53. TAZ interacts with 14-3-3 via a motif surrounding S89, which, as
part of the Hippo signaling pathway, is phosphorylated by LATS1/2 resulting in cytoplasmic
retention of TAZ and thus its functional deactivation.12 Loss of this regulation leads to abrogation
of contact inhibition which is one of the prerequisites of cancer formation and growth of solid
tumors. Binding of 14-3-3 to p53, following phosphorylation of T387 of p53 in response to DNA
damage, is protective against MDM2-mediated degradation.13 In addition, 14-3-3 binding to p53
facilitates p53 dimer-dimer interaction14 and stabilization of the functional p53 tetramer.15 14-3-3
ACS Paragon Plus Environment
3

Page 5 of 46
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
thus plays an important role in maintaining the tumor suppressor role of p53 by stimulating p53’s
transcriptional activity. Enhancing the binding of 14-3-3 to these two cancer-related proteins with
a small-molecule 14-3-3 PPI stabilizer could convey a positive therapeutic effect in many cancer
types.
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
Importantly, the potential “druggability” of 14-3-3 proteins has been validated by nature in the
form of the natural products class of the fusicoccanes. These diterpene glycosides have been shown
to stabilize a number of 14-3-3 PPIs and display biological activities that might be of therapeutic
benefit in cancer,5,16–20 cystic fibrosis10,11,21 and axon regeneration.22 The eponymous molecule –
the fungal toxin Fusicoccin A (FC-A) – was shown to stabilize the interaction of 14-3-3 proteins
with the regulatory C-terminus of the plant plasma ATPase, thereby activating this important
proton pump.23 Later it was shown that the related compound Cotylenin A displays a remarkable
anti-cancer activity while showing a very low general toxicity.24 One possible target for this
activity was suggested by us to be the negative regulatory complex between the N-terminus of the
oncogenic kinase C-Raf and 14-3-3 proteins.7 The crystal structure of Cotylenin A in complex
with 14-3-3ζ and the di-phospho peptide C-RafpS233pS259 showed that the natural product binds
to a gap in the interface pocket simultaneously establishing contacts to both 14-3-3 and the C-Raf-
derived peptide. Recently, an optimized semi-synthetic fusicoccane was shown to display
enhanced anti-proliferative activity against cancer cells.25 Furthermore, we reported that the
natural product FC-A displayed a beneficial effect for a potential therapy of breast cancer and
cystic fibrosis by stabilizing the inhibitory interaction of 14-3-3 with estrogen receptor α26 or the
chloride channel CFTR.11 Although the fusicoccanes are very useful proof-of-principle molecules,
their optimization is not without synthetic challenges. Therefore, additional approaches to identify
starting points for the development of PPI stabilizers are needed. Since traditional high throughput-
ACS Paragon Plus Environment
4

Journal of Medicinal Chemistry
Page 6 of 46
1
2
3
4
5
6
7
8
9
screening (HTS) were so far rewarded with only limited success,27 alternative methods like
fragment-based drug discovery (FBDD) holds the most promise to achieve this goal.
In this study we report the identification of fragments that bind to the distinct interfaces of 14-
3-3 with the partner proteins p53 and TAZ, hinting at the possibility to introduce specificity at an
early stage of fragment-based optimization of 14-3-3 PPI stabilizing ligands.
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
Screening for fragments that bind to the interface of 14-3-3 with TAZ and p53. For the
identification of interface-binding fragments that might be developed into PPI-stabilizing
compounds we employed a crystallography-based screen of fragment cocktails28. Here, crystals of
the binary 14-3-3 complexes were soaked in solutions of fragment mixtures (100 cocktails of 5
each, part of the X-ray Crystallography subset of the Astra Zeneca fragments collection) at a final
concentration of 10 mM per fragment. This screen identified two scaffolds that bound to a shallow
pocket that is located in the central binding channel of 14-3-3. Both fragment scaffolds contain an
amidine group (Figure 1 A,D) that engages in a salt-bridge with the side-chain carboxyl group of
glutamic acid 14 (E14) of 14-3-3 (Figure 1 C,F). The planar aromatic ring systems sit in a shallow
pocket that is comprised by 14-3-3 residues E39, N42 and L43 (Figure 1C). The first fragment
(AZ-001) can adopt two orientations with the chloride substituent either pointing towards or away
from the peptide-binding channel of 14-3-3. The second fragment (AZ-002) shows a more defined
orientation, with the benzyl substituent of the thiophene ring pointing away from the peptide,
resting on an uncharged cushion comprised by the hydrocarbon part of C38 and E39 of helix 3 of
14-3-3. Structural analogs of AZ-002, namely AZ-003 and AZ-004, were additionally selected
from the screen and observed to bind to 14-3-3 in the presence of the p53pT387, but not contacting
the peptide (Figure S1-2).
ACS Paragon Plus Environment
5

Page 7 of 46
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. Identification of fragments binding to the interface of 14-3-3σ with p53. (A) Fragment
AZ-001. (B) Crystal structure of AZ-001 (yellow sticks) in complex with 14-3-3σ (white surface)
and p53pT387 (orange sticks), PDB ID: 6S40. The final 2Fo-Fc electron density maps are shown
as blue mesh. (C) Detailed view of the binding pocket of AZ-001. The most prominent interaction
is a salt-bridge (dotted black line) between the amidine of AZ-001 and carboxyl group of E14 of
14-3-3σ. (D) Fragment AZ-002. (E) Crystal structure of AZ-002 (yellow sticks) in complex with
14-3-3σ (white surface) and p53pT387 (orange sticks), PDB ID: 6RWI. The final 2Fo-Fc electron
ACS Paragon Plus Environment
6

Journal of Medicinal Chemistry
Page 8 of 46
1
2
3
4
5
6
7
8
9
density maps are shown as blue mesh. (F) Detailed view of the binding pocket of AZ-002. Also
with AZ-002, the most prominent interaction is a salt-bridge (dotted black line) between the
amidine of AZ-002 and carboxyl group of E14 of 14-3-3σ.
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
AZ-003 interacts with both 14-3-3 and TAZ and induces a conformational change in the
TAZpS89 peptide. In contrast to the 14-3-3/p53pT387 complex, the binding site of AZ-003 lies
directly at the interface of 14-3-3 with the TAZpS89 peptide (Figure 2 A,B). When this fragment
was soaked into crystals of the binary complex, we observed a significant change in the
conformation of the TAZ peptide (Figure 2 C,D). Most obvious is the additional electron density
at the C-terminus of the peptide which allowed us to build three additional amino acids (L96-T98).
Secondly, a peptide flip is observed between L94 and Q95 which allows the main-chain carbonyl
oxygen of Q95 to establish a polar contact with the amine group of AZ-003 (Figure 2 D). Finally,
the carbonyl oxygen of L96 is interacting in the same manner with this amine substituent. It is
likely that these interactions are the reason for the observed changes in the conformation of the
TAZ peptide, since the similar structural analog of AZ-004 does not invoke comparable structural
differences (Figure S3). Complementary protein-based NMR experiments allowed us to confirm
the binding site of AZ-003 to 14-3-3σ and to the 14-3-3σ/TAZpS89 complex in solution (Figure
S4, S5). Further characterization of the effect of AZ-003 on the binary complex using fluorescence
polarization (Figure S6) revealed an increase in anisotropy in the control experiment (no protein),
a sign of interferences which make this assay unsuitable for this particular fragment. Even though
a stabilizing activity could not be confirmed by FP for AZ-003, based on the observed
conformational change induced by the newly formed interactions with two residues of TAZpS89,
we believe that fragment AZ-003 represents a promising chemical starting point that could be
ACS Paragon Plus Environment
7

Page 9 of 46
Journal of Medicinal Chemistry
1
2
3
4
5
6
developed into a more potent orthosteric stabilizer, for example by modulating the phenyl ring
attached to the thiophene C4 position.
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. AZ-003 bound to 14-3-3σ/TAZpS89. (A) Structure of fragment AZ-003. (B) Crystal
structure of AZ-003 (yellow sticks) bound to the complex of 14-3-3σ (white surface) and TAZpS89
(green sticks) PDB ID: 6RLHC. The final 2Fo-Fc electron density maps are shown as blue mesh.
(C) Contacts between AZ-003, 14-3-3σ and TAZpS89. The structure of TAZpS89 bound to 14-3-
3σ in the absence of AZ-003 is superimposed as magenta sticks. (D) Details of the changes binding
of AZ-003 induces in the conformation of the TAZpS89 peptide. The peptide backbone flip is
indicated by an orange arrow, the distance between the positions of the main-chain carbonyl
ACS Paragon Plus Environment
8

Journal of Medicinal Chemistry
Page 10 of 46
1
2
3
4
5
6
oxygen is shown. Polar contacts between the amino group of AZ-003 and main-chain carbonyls
of TAZpS89 are indicated as dotted, black lines.
7
8
9
Preliminary fragment extension confers selectivity towards 14-3-3σ in complex with
p53pT387 over TAZpS89. Because our crystallography-based fragment screen identified
fragments that bind to both complexes, additional steps to achieve specificity were needed. Based
on the structural data presented, fragments were designed and custom synthesized (Scheme S1-
S4), leading to the identification of fragments (Figure 3A) that were compatible with binding to
the 14-3-3σ/p53pT387 complex, but not to the 14-3-3σ/TAZpS387 complex. Several structural
features of the 14-3-3σ/p53pT387 complex have to be taken into consideration for the prospect of
achieving specificity in stabilization. First, the sharp U-turn that the peptide is showing C-terminal
of E388 is enabled by the consecutive presence of both a glycine (G389) and a proline (P390).
This architecture is further supported by the unique feature that the entire protein ends three amino
acids further with the sequence DSD where both the side chain of the ultimate aspartate and the
free C-terminus engage the guanidinium group of R60 in a salt bridge. Together with the rather
unusual glutamate at position +1 (E388) after the phosphorylated threonine (pT387), this makes
quite a unique interface (Figure 1B,E) which might allow for the development of a specific
orthosteric 14-3-3σ/p53pT387 PPI stabilizer. With these requirements in mind, synthesized
fragments AZ-005, AZ-006, and AZ-007 (Figure 3A) represent useful chemical starting points for
structural optimizations towards p53pT387-specific 14-3-3 PPI stabilizers. The crystal structures
of these fragments soaked into binary crystals of 14-3-3σ with the p53 peptide (Figure 3B) showed
that the nitrogen-bearing extensions of the core scaffold point towards the p53 peptide and reach
into the water network between 14-3-3 and p53 (Figure 3C). Since the TAZ peptide occupies the
entire length of the 14-3-3 binding channel, these binding modes are expected to be incompatible
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
9

Page 11 of 46
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
with the presence of the TAZ peptide (Figure S7). Consequently, we could not identify additional
electron density after soaking these fragments into 14-3-3σ/TAZpS89 crystals, with the exception
of AZ-006 which can form a disulfide bond with C38 and is therefore able to compete with
TAZpS89 for binding to 14-3-3, resulting in a clear loss of electron density of the TAZ peptide,
but of parts of the fragment, too (Figure S8).
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 3. AZ fragments that reach into the water network between 14-3-3σ and p53pT387. (A)
Fragments AZ-005, AZ-006, AZ-007. (B) Crystal structures of the fragments bound to the
complex of 14-3-3σ (white surface) and p53pT387 (orange sticks), PDB IDs: 6RX2, 6R5L, and
6RWH, respectively. The final 2Fo-Fc electron density maps are shown as blue mesh. (C) Contacts
between the fragments and the water network between 14-3-3σ and p53pT387.
AZ-008 binds to and shows weak stabilization of the 14-3-3/p53pT387 complex. For the goal
of developing an orthosteric PPI stabilizer of the 14-3-3/p53pT387 complex it seems necessary
ACS Paragon Plus Environment
10

Journal of Medicinal Chemistry
Page 12 of 46
1
2
3
4
5
6
7
8
9
that an extended fragment based on AZ-005 – AZ-007 has to reach the p53 peptide and
preferentially engage the side chain of E388. These considerations led to the design of AZ-008
(Figure 4A) which in a docking calculation is predicted to contact the carboxyl group of E388 with
its terminal amino group (Figure 4B). We tried to soak this compound into pre-formed 14-3-
3/p53pT387 crystals and also undertook extensive co-crystallization trials, but unfortunately were
not able to detect any additional electron density by X-ray crystallography for compound AZ-008.
We thus used WaterLOGSY NMR experiments to detect the potential interaction of AZ-008 to
14-3-3σ. Positive 1H signals of the ligand in the presence of the protein are indicative of binding.
Positive NMR signals in the WaterLOGSY spectra of AZ-008 mixed to 14-3-3σ, both in the
presence or absence of the p53pT387 peptide, confirmed the interactions (Figure S9). The two-
dimensional spectrum (1H-15N TROSY-HSQC) of 14-3-3σ (15N2H labeled) was next used as
reporter to further characterize the binding. Because the resonances in this spectrum are sensitive
to the chemical environment of the corresponding residues, a modification in the presence of a
ligand of a resonance chemical shift value, and/or intensity, reports an interaction.29 In addition,
chemical shift assignments of 14-3-3σ have been performed, allowing to link a resonance in the
spectrum to a specific amino acid residue in the 14-3-3σ sequence. The addition of AZ-008 to
14-3-3σ or 14-3-3σ in the presence of the p53pT387 peptide resulted in detectable broadening
(intensity decrease) and chemical shift value perturbation (CSP) of a few resonances in the 1H-15N
2D spectrum of the protein (Figures S10-S11, Figure 4 C,D, Figure 5A). Based on the signal
assignments, we concluded that addition of AZ-008 affected mainly the intensity of resonances of
residues R41, N42, E115 and F119 of 14-3-3σ (Figure 4 C,D, Figure S11 A,B), with no variation
in the AZ-008 binding pocket mapping in the presence of the p53pT387 peptide (Figure 4D, Figure
5A, Figure S10). These experiments indicated that binding of AZ-008 occurs in the same pocket
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
11

Page 13 of 46
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
in both cases, with or without the p53pT387 peptide. Visualizing the corresponding amino acid
perturbation induced by AZ-008 on the 14-3-3σ structure clearly highlighted a binding site (Figure
4C), which can be matched to data from the co-crystal structure of 14-3-3σ and p53pT387 in the
presence of its analogs AZ-001/002 (Figure 1). The binding pose resulting from docking AZ-008
to the 14-3-3σ/p53pT387 complex (Figure 4B) is in accordance with both the NMR data and the
bound-conformation adopted by other AZ fragments based on this scaffold.
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. AZ-008 binds to the 14-3-3σ/p53pT387 complex. (A) Structure of AZ-008. (B) Docking
of AZ-008 (yellow sticks) bound to the 14-3-3σ (white sticks) in complex with p53pT387 (orange
sticks). (C) Mapping on the crystal structure of 14-3-3σ (PDB ID: 1YZ5) of the amino acid
residues corresponding to the most affected resonances by the presence of AZ-008. Residues
whose corresponding correlation peak intensities were the most affected are colored in red and an
ACS Paragon Plus Environment
12

Journal of Medicinal Chemistry
Page 14 of 46
1
2
3
4
5
6
7
8
9
additional 5-8 are colored in brown. (D) Plot of the intensity ratios I/I0 of corresponding 1H-15N
correlation peaks in the 2D 14-3-3σ spectra corresponding to the bound 14-3-3σ (100 µM 14-3-3σ
in the presence of 2000 µM AZ-008, I) and free 14-3-3σ (100 µM 14-3-3σ I0) protein (y axis)
versus 14-3-3σ amino acid sequence (x axis). A total of 133 correlation peak intensities are shown.
The x axis is not proportional to the sequence length. The helices of 14-3-3σ are identified below
the x axis as blue cartoons, while disordered regions are represented by green 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
1
Interestingly, addition of AZ-008 also induced chemical shift perturbations of the H NMR
signals of the bound peptide (Figure 5B, Figure S12). In particular, the 1H signals (Figure 5B) of
the Hδ methyl protons of L383 of the p53pT387 peptide are affected by AZ-008 binding, despite
the fact that they are not close to the fragment binding site. Perturbation of these signals thus did
not correspond to a direct effect due to the local change of environment. This data rather suggests
that the binding of the fragment impacted the global binding of the peptide to 14-3-3σ. Moreover,
the extent of decrease of the intensity of specific resonances in the 2D spectrum of 14-3-3σ in the
presence of AZ-008 reports on the amount of complex present in the solution. The intensity of
resonances corresponding to amino acid residues located in, or close to, the p53pT387 binding site
specifically decreased with the addition of AZ-008 to a solution containing the 14-3-3σ/p53pT387
complex (Figure 5C,D) and, reversely, the extent of CSP increased (Fig. S11C). Combined with
the results of other methods, this observation indicates that the amount of 14-3-3σ/p53pT387
complex present in solution was increased when both ligands were present, suggesting that the
complex was therefore stabilized in the presence of AZ-008. Further experiments using
Fluorescence Polarization (Figure 5E, Figure S13) and Surface Plasmon Resonance (Figure 5F,
Figure S14) also suggested that AZ-008 is able to stabilize the interaction between 14-3-3σ and a
p53pT387 peptide (32mer) containing the complete C-terminal domain. The enthalpically
ACS Paragon Plus Environment
13

Page 15 of 46
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
favorable contact between the amino substituent of AZ-008 and E388 of p53pT387 observed in
the docking complex described above helps to explain the observed effect. Although this effect is
objectively small, it appears to be tangible as it was confirmed quantitatively by two separate
methods, but also qualitatively by NMR. Moreover, the two-fold effect observed here is in
accordance with the change in activity that small, non-covalent molecular fragments are expected
to induce. Most importantly, these results hint at the possibility of transforming inactive fragments
initially identified in X-ray crystallography and/or NMR studies into active molecules through
rational design of structural extensions.
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
14

Journal of Medicinal Chemistry
Page 16 of 46
1
2
3
4
5
6
7
8
9
Figure 5. AZ-008 stabilizes the 14-3-3σ/p53pT387 complex. (A) Plot of the intensity ratios I/I0
(y-axis) of corresponding 1H-15N correlation peaks in the 2D 14-3-3σΔC17 spectra of, in red: the
peptide-bound 14-3-3σΔC17 (100 µM 14-3-3σ in the presence of 500 µM p53pT387, I) and free
14-3-3σ (100 µM 14-3-3σ I0), and in blue: corresponding to the peptide and fragment-bound 14-
3-3σ (100 µM 14-3-3σ with 500 µM p53pT387 and 2000 µM AZ-008 , I) and free 14-3-3σ (100
µM 14-3-3σ, I0) protein. Note that the boxed regions correspond to the same areas affected in
Figure 4 D and therefore show that the binding site of the fragment is the same when the peptide
is present or not. (B) 1H NMR signals of the spectrum of 100 µM 14-3-3σ (in black), 100 µM 14-
3-3σ in the presence of 500 µM p53pT387 (in red) and 100 µM 14-3-3σ in the simultaneous
presence of 500 µM p53pT387 and 2000 µM AZ-008 (in blue). The NMR signals correspond to
the Hδ protons of L383 of the p53pT387 peptide. CSP in the peptide signals can be observed when
AZ-008 is added to the solution containing the 14-3-3σ/p53pT387 complex. (C,D) Plot of the
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
1
ratios of the bound (I)/free (I0) H-15N TROSY-HSQC correlation peak intensities of individual
resonances corresponding to residues of 100 µM 14-3-3σ, remote from the interaction site (C) or
close to the p53pT387 interaction site of (D), in the presence of 500 µM p53pT387 (in red) and
100 µM 14-3-3σ in the simultaneous presence of 500 µM p53pT387 and 2000 µM AZ-008 (in
blue). Error bars of I/I0 ratios were calculated based on the maximum noise estimate of each
spectrum. Note that the resonances corresponding to regions affected by the p53pT387 peptide
binding were considerably more affected in the presence of AZ-008. (E) Titration of 14-3-3σ to a
fluorescently labelled p53pT387 peptide in the absence and presence of 1 mM of AZ-008. Shown
are average and standard deviation of three independent experiments. (F) SPR measurement of 14-
3-3σ binding to immobilized p53pT387 peptide in the absence and presence of 1 mM AZ-008.
Shown are average and standard deviation of three independent experiments.
ACS Paragon Plus Environment
15

Page 17 of 46
Journal of Medicinal Chemistry
1
2
3
4
DISCUSSION
5
6
7
8
Targeted stabilization of PPIs is an underexplored concept in drug discovery despite numerous
examples from natural products and synthetic molecules advocating for this principal strategy.2,3,30
Successful drugs that operate by stabilizing PPIs have almost exclusively only been recognized as
PPI stabilizers after their therapeutic relevance was shown and in many cases even long after FDA
approval. One famous example of such drugs is thalidomide and its derivatives, for example
lenalidomide,31 which are collectively called IMiDs (immunomodulatory drugs). Lenalidomide
acts as a “molecular glue” that hijacks a ubiquitin ligase for the targeted degradation of a neo-
substrate protein.32 Another example is the history of the immunosuppressants rapamycin and
FK506 which puzzled the community for a long time, how these natural products actually convey
their therapeutic effect. Eventually it was shown that although they initially bind to the same
primary receptor protein – FKBP12 – their actual functional targets are the phosphatase
Calcineurin (FK506) and the protein kinase mTOR (rapamycin) that were “glued” to FKBP12 by
the natural products and thus inactivated.33,34
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
A third, recent success story of the PPI induction and stabilization concept is PROTACs
(PROteolysis TArgeting Chimeras), a technology employing heterobifunctional small molecules
that facilitate the interaction of a target protein with a ubiquitin E3 ligase.35 This approach has not
only attracted a lot of attention in the chemical biology field but is currently explored by many
pharma companies and at least half a dozen start-up companies. Together with the aforementioned,
rather serendipity-driven success examples of IMiDs and “molecular glue” natural products,
PROTACs may define a new conceptional era of pharmacologic intervention where an
enhancement or induction of a native regulatory process (targeted protein degradation) instead of
a direct protein inhibition is the underlying principle.
ACS Paragon Plus Environment
16

Journal of Medicinal Chemistry
Page 18 of 46
1
2
3
4
5
6
7
8
9
Our approach to enhance the regulatory function of 14-3-3 PPIs with different partners like p53
or TAZ is an innovative extension of this concept with the added advantage that a physiological,
already implemented regulatory network is addressed with small molecules and also that function-
enhancing PPIs, like in the case of 14-3-3/p53, can be promoted. We and others have shown in the
past that especially the natural product class of the fusicoccanes can be employed as orthosteric
stabilizers of 14-3-3 PPIs.7,11,21,22,26,36–39 However, with only supramolecular ligands40,41 that are
hard to tract by medicinal chemistry, and two screening hits of 14-3-3 PPIs27,42, there is a clear
need to expand the chemical diversity of 14-3-3 PPI stabilizers. A very useful approach to achieve
this goal is fragment-based drug discovery (FBDD).28 We have shown that fragments binding to
14-3-3 protein complexes can be identified43,44 and in particular covalent fragment discovery is a
fruitful approach,45 as can be seen in the context of the current work, too (Figure 3 and S15).
Especially in the new field of targeting rim-of-the interface pockets with orthosteric stabilizers,
FBDD is extremely valuable since it circumvents the bias of traditional compound collections used
for high throughput-screening and allows to efficiently probe the chemical space to identify
starting points for the development of PPI stabilizers. However, one potential drawback of FBDD
in regards to the discovery of PPI stabilizers might be that most identified fragments are silent in
classical biophysical assays used for high-throughput screening, as perfectly exemplified within
this study. X-ray crystallography and NMR studies therefore appear as the methods of choice, to
first find fragment scaffolds capable of binding to the target protein(s), and then develop their
activities, a goal that has not yet been successfully accomplished and that represents a considerable
challenge.
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
17

Page 19 of 46
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
As our example has shown, specificity for two different, but closely related PPI interfaces can
be aimed for even at the very early stage of extending an initial fragment hit. Through a relatively
limited series of protein x-ray crystallography identified fragments (see also Figures S16-S34) we
were able to identify promising chemical starting points for the future development of active
stabilizers that discriminate between the 14-3-3 complexes with p53pT387 and TAZpS89. This is
important for the prospect to make the widespread 14-3-3 interactome accessible for drug
discovery, but also bears implications for other PPIs. An analysis by the Skolnick group has shown
that in the crystal structures of 1,611 protein-protein complexes, the majority of “ligandable”
pockets are located with a 6 Å distance from the protein interfaces,46 which is well exemplified by
the fragment-binding site of our study. This area, possibly privileged for small-molecule
accommodation, is the region where orthosteric PPI stabilizers are supposed to bind. Together with
the overall high structural diversity of PPI interfaces this observation nourishes the prospect that
for many regulatory protein complexes a specific, rim-of-the interface targeting, orthosteric
stabilizer might be developed.
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
EXPERIMENTAL SECTION
Peptide Synthesis. FITC labeled and acetylated 13mer TAZpS89 were synthesized via Fmoc
solid-phase peptide synthesis (SPPS) using an automated Intavis MultiPepRSi peptide synthesizer
and a rink amide AM resin (Novabiochem; 0.59 mmol/g loading). The peptides used for NMR
measurements and for crystallization studies were acetylated at the N terminus (1:1:3
Ac2O/pyridine/NMP) before resin cleavage. The peptides used for FP assays were manually
labeled with FITC (Sigma-Aldrich) attached to a β-Alanine linker introduced via Fmoc-β-Ala-OH
ACS Paragon Plus Environment
18

Journal of Medicinal Chemistry
Page 20 of 46
1
2
3
4
5
6
(Iris Biotech) under standard SPPS conditions. All peptides were purified by preparative reversed-
phase HPLC with MS detection.
7
8
9
TAMRA labeled 32-mer p53 peptide and acetylated 12-mer p53 were commercially synthesized
by Anaspec and Caslo, respectively.
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
14-3-3 Expression. His6-tagged 14-3-3σ full-length and ΔC (truncated after T231; for
crystallography) were expressed in BL21(DE3) competent cells with a pPROEX HTb plasmid and
purified with a nickel column. The His6-tag was cleaved off using TEV protease, and a second
purification was performed by size-exclusion chromatography (Superdex75). The proteins were
dialyzed against FP or crystallization buffers before use.
Crystallography. 14-3-3σΔC/TAZpS89 crystals: 14-3-3σΔC and acetylated TAZpS89 peptide,
both in crystallization buffer (25 mM HEPES pH 7.5, 100 mM NaCl, 10 mM MgCl2 and 2 mM -
mercaptoethanol), were combined in a 1:2 14-3-3σΔc/TAZpS89 molar ratio with a final protein
concentration of 12 mg/mL. After overnight complexation, hanging-drop crystallization wells
were set-up using a 1:1 ratio of precipitation buffer (95 mM HEPES pH 7.1, 0.19 M CaCl2, 5%
glycerol and 27% PEG400) and complexation mixture for a total drop volume of 2 µL. Crystals
were grown within 10 days at 4°C.
14-3-3σΔC/p53pT387 crystals: 14-3-3σΔc and acetylated p53pT387 peptide, both in
crystallization buffer (25 mM HEPES pH 7.5, 100 mM NaCl, 10 mM MgCl2 and 2 mM -
mercaptoethanol), were combined in a 1:2 14-3-3σΔc/TAZ molar ratio with a final protein
concentration of 12 mg/mL. After overnight complexation, hanging-drop crystallization wells
were set-up using a 1:1 ratio of precipitation buffer (0.1 M HEPES pH 7.5, 0.2 M CaCl2, 5%
glycerol, 2 mM -mercaptoethanol and 28% PEG400) and complexation mixture for a total drop
volume of 2 µL. Crystals were grown within 10 days at 4°C.
ACS Paragon Plus Environment
19

Page 21 of 46
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The compounds screened were part of the X-ray Crystallography subset of the AstraZeneca
fragments collection. Fragment soaking was performed on crystals of 10 days and older by adding
0.2 μL of a 100 mM stock solution in dimethyl sulfoxide to the 2 µL drop for a final concentration
of 20 mM. Crystals were fished after a minimum of 2 days and flash-cooled in liquid nitrogen.
Data sets were collected either on an in-house Rigaku Micromax-003 (Rigaku Europe, Kemsing
Sevenoaks, UK) sealed tube X-ray source and a Dectris Pilatus 200K detector (DECTRIS Ltd.,
Baden-Daettwil, Switzerland) or at the DESY PETRA III synchrotron beamline P11 (DESY,
Hamburg, Germany). All datasets were indexed, integrated and scaled using either iMosflm or
DIALS followed by AIMLESS. Phasing was performed using molecular replacement in Molrep
using PDB ID 5N75 (14-3-3σ/TAZ) or 5MHC (14-3-3σ/p53) as search model. Phaser, Coot and
phenix.refine were thereafter used in alternating cycles of model building and refinement.
Production of 15N2H labeled 14-3-3σΔC for NMR spectroscopy. The 15N2H labeled 14-3-3σ
(ΔC, cleaved after T231) for NMR studies was expressed in E. coli BL21 (DE3) cells transformed
with a pProExHtb vector carrying the cDNA to express an N-terminally His6-tagged human 14-3-
3σ. Briefly, the cells were grown in 1 L of deuterated M9 minimal medium supplemented with 2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
g/L 12C6 H7 Glucose, 1 g/L 15N Ammonium Chloride, 0.4 g/L Isogro 15N12C2H Powder – Growth
Medium (Sigma Aldrich) and 100 µg/mL ampicillin and the recombinant protein was then purified
by affinity chromatography using a Ni-NTA column (GE Healthcare). The His6-tag was further
cleaved by the TEV protease and the protein was further dialyzed overnight at 4°C against NMR
buffer (100 mM Sodium Phosphate, pH 6.8, 50 mM NaCl), concentrated, aliquoted, flash frozen
and stored at -80°C. A detailed protocol can be found at.29
1
15N-1H TROSY-HSQC NMR spectroscopy. H-15N TROSY-HSQC (Transverse Relaxation
Optimized Spectroscopy- Heteronuclear Single Quantum Coherence Spectroscopy) spectra were
ACS Paragon Plus Environment
20

Journal of Medicinal Chemistry
Page 22 of 46
1
2
3
4
5
6
7
8
9
acquired in 3 mm tubes (sample volume 200 μL) using a 900 MHz Bruker Avance Neo
spectrometer, equipped with a cryoprobe. The spectra were recorded at 32 °C, in a buffer
containing 100 mM sodium phosphate, 50 mM NaCl, pH 6.8, 4% (v/v) DMSO-d6, 1 mM DTT,
EDTA-free protease inhibitor cocktail (Roche, Basel, Switzerland) and 10% (v/v) D2O. The
experiments were recorded with 3072 complex data points in the direct dimension and 128
complex data points in the indirect dimension, with 184 scans per increment. For the evaluation of
the binding of AZ-008 to the 14-3-3σ/p53 complex, spectra of 15N2H labeled 14-3-3σΔC 100 µM
were recorded in the presence and absence of 2000 µM AZ-008, 500 µM p53pT387 peptide, and
simultaneously 500 µM p53pT387 peptide and 2000 µM AZ-008. For the evaluation of the binding
of AZ-003 to the 14-3-3σ/TAZ complex, spectra of 15N2H labeled 14-3-3σ 100 µM were recorded
in the presence and absence of 2000 µM AZ-003, 125 µM TAZpS89 peptide, and simultaneously
125 µM TAZpS89 peptide and 2000 µM AZ-003. Assignments of the backbone resonances of
15N2H labeled 14-3-3σΔC were previously reported.29 A 1H spectrum with water-suppression was
additionally recorded for each sample for monitoring the peptide signals. The reference for the 1H
chemical shift was relative to DSS (4,4-dimethyl-4-silapentane-1-sulfonic acid) while 15N
chemical shift values were referenced indirectly. Spectra were collected and processed with
Topspin 4.0 (Bruker Biospin, Karlsruhe, Germany) and analyzed with Sparky 3.12 (T. D. Goddard
and D. G. Kneller, SPARKY 3, University of California, San Francisco). CSPs verified as
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
1
chemical shift positions (in ppm) on the H-15N TROSY-HSQC were calculated using the
following equation:
Δ훿(1퐻)2 + [0.14 ∗ Δ훿(15푁)2
Δ훿 =
]
ACS Paragon Plus Environment
21

Page 23 of 46
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
WaterLOGSY NMR experiments. WaterLOGSY spectra were acquired in 5 mm tubes
(sample volume 530 µL) using a 600 MHz Bruker Avance I spectrometer equipped with a CPQCI
cryogenic probe. The spectra were recorded with 32768 complex data points, with 1024 scans per
increment and with a mixing time of 1.7s (acquisition time of 94 minutes). The spectra were
acquired at 16 °C, in a buffer containing 100 mM sodium phosphate, 50 mM NaCl, pH 6.8 and
10% (v/v) D2O. The final concentration of DMSO-d6 was 2% (v/v) and was kept constant for all
the experiments. For the evaluation of the binding of AZ-008 to the 14-3-3σ/p53 complex,
WaterLOGSY spectra were recorded on solutions containing 14-3-3σ 25 µM in the presence and
absence of 500 µM AZ-008, 50 µM p53pT387 peptide, and simultaneously 50 µM p53pT387
peptide and 500 µM AZ-008. An additional control experiment consisted of the acquisition of a
waterLOGSY spectrum in the presence of 500 µM AZ-008 and in the absence of protein and
peptide. A 1H spectrum with water-suppression was additionally recorded for each sample. Spectra
were collected, processed and analyzed with Topspin 4.0 (Bruker Biospin, Karlsruhe, Germany).
In silico docking of AZ-008 in 14-3-3/p53 complex. Docking experiments of fragments
libraries were performed using Schrödinger Drug Discovery suite (Schrödinger Release 2019-1:
Schrödinger, LLC, New York, NY, 2019).
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 Schrödinger protein preparation wizard was used to prepare the crystal structure. Hydrogen
atoms were added and possible Metal binding states generated. The protonation and tautomeric
states of Asp, Glu, Arg, Lys and His were adjusted to match a pH of 7.5 and possible orientations
of Asn and Gln residues were generated. Hydrogen bond sampling with adjustment of active site
water molecule orientations was performed using PROPKA. Water molecules with fewer than two
hydrogen bonds to non-waters were deleted. Finally, the protein-ligand complexes were subjected
to geometry refinements using the OPLS3 force field in restrained minimizations. The ligands
ACS Paragon Plus Environment
22

Journal of Medicinal Chemistry
Page 24 of 46
1
2
3
4
5
6
7
8
9
were prepared using LigPrep in the Schrödinger suite. The OPLS3 force field was used in all ligand
preparation steps. Possible protonation and ionization states were assigned to each ligand using
EpiK at pH 7.5. Possible stereoisomers, tautomeric states and metal binding sites were generated.
Surface plasmon resonance (SPR). SPR analysis was performed using a Biacore 3000
instrument. CTD Thr387 phosphorylated 32-mer peptide of p53 (AnaSpec) was immobilized ontoa
CMD200M sensor chip (Xantec bioanalytics) by using standard NHS/EDC amine-coupling
chemistry. Peptide immobilization was carried out at a flow rate of 20 μL/min and protein titration
experiments at 10 μL/min. All measurements were conducted at 25°C in HBSP+ buffer (0.01 M
HEPES pH 7.4, 0.15 M NaCl, 0.05% Surfactant P20, 1% DMSO). 14-3-3σ was injected at
indicated concentrations in the presence of 1 mM AZ-008 or 1% DMSO control. KD values were
obtained by first fitting the non-normalized data using the “log(agonist) vs. response (three
parameters)” equation of GraphPad Prism, then taking the fitted upper plateau for data
normalization, subsequently using the “Sigmoidal dose-response (variable slope)” equation with a
constrain to 100% activity.
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
Fluorescence polarization (FP) assay. Fluorescence polarization measurements were
conducted in FP-buffer (10 mM HEPES pH 7.4, 150 mM NaCl, 0.01% TWEEN-20, 1.0 mg/mL
BSA). Protein titrations: dilution series of 14-3-3σ were made in triplet in low binding Corning
Black Round Bottom 384-well plates in FP-buffer containing a fixed concentration of labeled
32mer peptide (10 nM TAMRA-p53) and 1% DMSO or 1 mM AZ-008. Plate was incubated at
room temperature protected from light for 1 h and polarization was measured using PHERAstar
(BMG Labtech) microplate reader (λex:535 nm, λem:590 nm). KD values were obtained by first
fitting the non-normalized data using the “log(agonist) vs. response (three parameters)” equation
ACS Paragon Plus Environment
23

Page 25 of 46
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
of GraphPad Prism, then taking the fitted upper plateau for data normalization, subsequently using
the “Sigmoidal dose-response (variable slope)” equation with a constrain to 100% activity
Compound titrations: dilution series of the fragments dissolved in DMSO were made with fixed
concentration of labeled peptides (FITC-TAZpS89: 10 nM; TAMRA-p53pT387: 10 nM) and
protein (40 nM and 10 μM 14-3-3 for TAZ and p53 respectively) (Control: DMSO concentration
1% v/v). The Corning 384-well plates (black, low volume, low binding, round bottom) were
incubated at 4C protected from light overnight and polarization was measured using Tecan Infinite
F500 plate reader (Thermo Fisher) using appropriate excitation and emission wavelength for FITC
(λex:485 nm, λem:535 nm) and TAMRA (λex:535 nm, λem:590 nm)
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
Synthesis and Characterization of fragments.
General Chemistry Statement. All solvents and chemicals used were of reagent grade. Purity and
characterization of compounds were established by a combination of LC − MS and NMR
techniques. LC-MS spectra were obtained on a Shimadzu UFLC fitted with a Shimadzu LCMS-
2020 mass spectrometer and a Waters BEH C18 (50x2.1 1.7µm) or Shim-pack XR-ODS (50x3.0
2.2µm) or Phenomenex Gemini –NX 3u C18 110A (50x3.0 3µm) column at a flow rate of
1.2ml/min 95% A to 95% B over 2.0 min with a 0.6 min hold, where A = 0.1% Formic acid or
0.05% TFA in water and B = Acetonitrile. purity was determined by UV absorption at a
wavelength of 254 nm, and the mass ion was determined by electrospray ionization (ESI,
Micromass Instrument). All final compounds were found to be of 95% purity or greater, as
1
1
assessed by LC−MS and H NMR. H NMR spectra were recorded on a 300 MHz or 400 MHz
using Bruker spectrometers and using DMSO-d6, CDCl3 or MeOD-d4 as solvent depending on
solubility. Data for 1H NMR were reported as chemical shifts (in parts per million from internal
standard tetramethylsilane), multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, h = heptet,
ACS Paragon Plus Environment
24

Journal of Medicinal Chemistry
Page 26 of 46
1
2
3
4
5
6
7
8
9
m = multiplet, br = broad), and integration. Preparative high-performance liquid chromatography
(HPLC) was performed on a Waters or Phenomenex column using decreasingly polar mixtures of
water (containing 1% formic acid or 1% aqueous NH4OH) and acetonitrile. Purification by flash
column chromatography (FCC) was typically performed using silica gel (Merck 7734 grade), and
solvent mixtures and gradients are recorded herein. All reactions were performed under nitrogen
at room temperature (rt) unless otherwise stated.
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
Synthesis pathway to 7-[(2-amino-1-methyl-ethyl)amino]benzothiophene-2-carboxamidine (AZ-
005) (Scheme S1).
Step a: Methyl 7-[[2-(tert-butoxycarbonylamino)-1-methyl-ethyl]amino]benzothiophene-2-
carboxylate (1). Xantphos (341 mg, 0.59 mmol) and Pd2(dba)3 (270 mg, 0.30 mmol) were added
to Cs2CO3 (2884 mg, 8.85 mmol), methyl 7-bromobenzo[b]thiophene-2-carboxylate (800 mg, 2.95
mmol) and tert-butyl (2-aminopropyl)carbamate (617 mg, 3.54 mmol) in toluene (20 mL) at 20 °C
under nitrogen. The resulting solution was stirred at 130 °C for 4 hours. The crude product was
purified by flash silica chromatography, elution gradient 0 to 10% EtOAc in petroleum ether. Pure
fractions were evaporated to dryness to afford the title compound (1000 mg, 93 %) as a gum. 1H
NMR (300 MHz, DMSO-d6) δ 1.08 – 1.23 (d, J = 6.3 Hz, 3H), 1.40 (s, 9H), 2.92 – 3.08 (m, 1H),
3.16 – 3.31 (m, 1H), 3.62 – 3.79 (m, 1H), 3.89 (s, 3H), 5.38 – 5.50 (m, 1H), 6.73 – 6.84 (m, 1H),
7.03 – 7.13 (m, 1H), 7.27 – 7.31 (m, 2H), 8.10 (s, 1H). MS (ESI) [M+H]+ = 365.
Step b: methyl 7-[(2-amino-1-methyl-ethyl)amino]benzothiophene-2-carboxylate (2). TFA (0.106
mL, 1.37 mmol) was added to methyl 7-((1-((tert-butoxycarbonyl)amino)propan-2-
yl)amino)benzo[b]thiophene-2-carboxylate (500 mg, 1.37 mmol) in DCM (20 mL) at 20 °C. The
resulting solution was stirred at rt for 1 hour. The crude product was purified by flash silica
chromatography, elution gradient 0 to 10% MeOH in DCM. Pure fractions were evaporated to
ACS Paragon Plus Environment
25

Page 27 of 46
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
dryness to afford the title compound (360 mg, 99 %) as a gum. 1H NMR (400 MHz, DMSO-d6) δ
1.20 – 1.26 (d, J = 6.4 Hz, 3H), 2.94 – 3.11 (m, 2H), 3.90 (s, 3H), 3.92 – 4.01 (m, 1H), 5.35 – 5.78
(br s, 1H), 6.75 – 6.81 (d, J = 6.5 Hz, 1H), 7.28 – 7.40 (m, 2H), 7.87 (s, 2H), 8.13 (s, 1H). MS
(ESI) [M+H]+ = 265
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
Step c: 7-[(2-amino-1-methyl-ethyl)amino]benzothiophene-2-carboxamidine (AZ-005). methyl 7-
((1-aminopropan-2-yl)amino)benzo[b]thiophene-2-carboxylate (300 mg, 1.13 mmol) was added
to ammonium chloride (243 mg, 4.54 mmol) and Trimethylaluminum (327 mg, 4.54 mmol) in
toluene (50 mL) at 20 °C under nitrogen. The resulting solution was stirred at 110 °C for 4 hours.
The crude product was purified by preparative Column: XBridge Prep OBD C18 Column
19*250mm,5um;Mobile Phase A:Water(0.1%FA), Mobile Phase B: MeCN; Flow rate: 20
mL/min; Gradient: 2% B to 30% B. Fractions containing the desired compound were evaporated
to dryness to afford the title compound (30.0 mg, 7.72 %) as a solid. 1H NMR (400 MHz,
Methanol-d4) δ 1.29 – 1.46 (d, J = 6.4 Hz, 3H), 3.10 – 3.24 (m, 2H), 4.03 – 4.17 (m, 1H), 6.89 –
7.00 (dd, J = 6.7, 1.9 Hz, 1H), 7.39 – 7.49 (m, 2H), 8.20 (s, 1H), 8.57 (s, 2H). MS (ESI) [M+H]+
= 249.
Synthesis pathway to 4-[(2-amino-1-methyl-ethyl)amino]-6-(sulfanylmethyl)benzothiophene-2-
carboxamidine (AZ-006) (Scheme S2).
Step a: methyl 4-bromo-6-methyl-benzothiophene-2-carboxylate. methyl 2-mercaptoacetate (2.69
g, 25.34 mmol) was added to 2-bromo-6-fluoro-4-methylbenzaldehyde (5 g, 23.04 mmol) and
K2CO3 (6.37 g, 46.08 mmol) in DMF (20 mL) at 25 °C. The resulting suspension was stirred at
100 °C for 2 hours. The reaction mixture was diluted with EtOAc (250 mL) and washed
sequentially with saturated brine (200 mL x 3). The organic layer was dried over Na2SO4, filtered
ACS Paragon Plus Environment
26

Journal of Medicinal Chemistry
Page 28 of 46
1
2
3
4
5
6
and evaporated to afford the title compound (5.0g, 76%) as a solid. Taken through to the next step
without further purification.
7
8
9
Step b: 4-bromo-6-methyl-benzothiophene-2-carboxamide. A solution of methyl 4-bromo-6-
methylbenzo[b]thiophene-2-carboxylate (6.5 g, 22.79 mmol) in NH3 in MeOH (7 M) (75 mL) was
stirred at 110 °C overnight. The mixture was concentrated to obtain the crude product which was
used in next step directly.
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
Step c: 4-bromo-6-methyl-benzothiophene-2-carbonitrile (3). TFA (10.98 mL, 77.74 mmol) was
added to a solution of 4-bromo-6-methylbenzo[b]thiophene-2-carboxamide (7 g, 25.91 mmol) and
Pyridine (10.48 mL, 129.56 mmol) in DCM (30 mL). The mixture was stirred at 25 °C for 0.5
hour. The reaction mixture was diluted with DCM (200 mL), and washed sequentially with 0.1M
HCl (100 mL x 2). The organic layer was dried over Na2SO4, filtered and evaporated to afford the
title compound (6 g) as a solid. Taken through to the next step without further purification.
Step d: 4-bromo-6-(bromomethyl)benzothiophene-2-carbonitrile. NBS (1.853 g, 10.41 mmol) was
added in one portion to 4-bromo-6-methylbenzo[b]thiophene-2-carbonitrile (2.5 g, 9.92 mmol)
and BPO (0.240 g, 0.99 mmol) in CCl4 (150 mL) at 90 °C. The resulting solution was stirred at 90
°C for 2 hours. The reaction mixture was diluted with EtOAc (200 mL) and washed sequentially
with saturated brine (150 mL x 2). The organic layer was dried over Na2SO4, filtered and
evaporated to afford crude product. The crude product was purified by flash silica chromatography,
elution gradient 0 to 2% EtOAc in petroleum ether. Pure fractions were evaporated to dryness to
afford the title compound (2.70 g, 82 %) as a white solid. 1H NMR (300 MHz, DMSO-d6) δ 4.89
(s, 2H), 7.91 (s, 1H), 8.31 (s, 1H), 8.39 (s, 1H).
ACS Paragon Plus Environment
27

Page 29 of 46
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Step e: (4-bromo-2-cyano-benzothiophen-6-yl)methyl acetate (4). Silver acetate (1.639 g, 9.82
mmol) was added to 4-bromo-6-(bromomethyl)benzo[b]thiophene-2-carbonitrile (2.5 g, 7.55
mmol) in DMF (12 mL) at 20 °C under nitrogen. The resulting mixture was stirred at 60 °C for 1
hour. The reaction mixture was filtered through celite. The crude product was purified by flash
C18-chromatography, elution gradient 40 to 50% MeCN in water. Pure fractions were evaporated
to dryness to afford the title compound (2.100 g, 90 %) as a solid. 1H NMR (300 MHz, DMSO-d6)
δ 2.13 (s, 3H), 5.23 (s, 2H), 7.82 (s, 1H), 8.22 (s, 1H), 8.38 (s, 1H).
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
Step f: [4-[[2-(tert-butoxycarbonylamino)-1-methyl-ethyl]amino]-2-cyano-benzothiophen-6-
yl]methyl acetate (5). Pd2(dba)3 (0.768 g, 0.84 mmol) was added to (4-bromo-2-
cyanobenzo[b]thiophen-6-yl)methyl acetate (2.6 g, 8.38 mmol), tert-butyl (2-
aminopropyl)carbamate (1.753 g, 10.06 mmol), Cs2CO3 (3.43g,10.52mmol) and Xantphos (0.970
g, 1.68 mmol) in toluene (25 mL) at 25 °C under nitrogen. The resulting solution was stirred at
130 °C for 3 hours. The reaction mixture was filtered using a Buchner funnel and the solvent was
removed under reduced pressure. The crude product was purified by flash silica chromatography,
elution gradient 1 to 10% EtOAc in petroleum ether. Pure fractions were evaporated to dryness to
afford the title compound (1.700 g, 50.3 %) as a solid. MS (ESI) [M+H]+ = 404.
Step g: tert-butyl N-[2-[[2-(N-hydroxycarbamimidoyl)-6-(hydroxymethyl)benzothiophene-4-
yl]amino]propyl]carbamate. Hydroxylamine hydrochloride (0.861 g, 12.39 mmol) was added to
(4-((1-((tert-butoxycarbonyl)amino)propan-2-yl)amino)-2-cyanobenzo[b]thiophen-6-yl)methyl
acetate (1 g, 2.48 mmol), K2CO3 (2.055 g, 14.87 mmol) in MeOH (20 mL) at 25 °C under
nitrogen. The resulting solution was stirred at 50 °C for 1 hour. The reaction mixture was filtered
through celite. The crude product was purified by flash silica chromatography, elution gradient 0
ACS Paragon Plus Environment
28