Hitting a Moving Target: Simulation and Crystallography Study of ATAD2 Bromodomain Blockers

Article abstract of DOI:10.1021/acsmedchemlett.0c00080

Small molecule ligand binding to the ATAD2 bromodomain is investigated here through the synergistic combination of molecular dynamics and protein crystallography. A previously unexplored conformation of the binding pocket upon rearrangement of the gatekeeper residue Ile1074 has been found. Further, our investigations reveal how minor structural differences in the ligands result in binding with different plasticity of the ZA loop for this difficult-to-drug bromodomain.

Full text of DOI:10.1021/acsmedchemlett.0c00080

Subscriber access provided by CARLETON UNIVERSITY
Letter
Hitting a moving target: simulation and crystallography
study of ATAD2 bromodomain blockers
Aymeric Dolbois, Laurent Batiste, Lars Wiedmer, Jing Dong, Manuela Brütsch,
Danzhi Huang, Nicholas M. Deerain, Dimitrios Spiliotopoulos, Iván Cheng-
Sánchez, Eleen Laul, Cristina Nevado, Pawel Sledz, and Amedeo Caflisch
ACS Med. Chem. Lett., Just Accepted Manuscript • Publication Date (Web): 30 Jun 2020
Downloaded from pubs.acs.org on June 30, 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 28
ACS Medicinal Chemistry Letters
1
2
3
4
5
6
7
8
9
Hitting a moving target: simulation and
crystallography study of ATAD2 bromodomain
blockers
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
Aymeric Dolbois,1‡ Laurent Batiste,2‡ Lars Wiedmer,2 Jing Dong,2† Manuela Brütsch,1
Danzhi Huang,2 Nicholas M. Deerain,2 Dimitrios Spiliotopoulos,2 Iván Cheng-Sánchez,1
Eleen Laul,1 Cristina Nevado,1* Paweł Śledź,2* Amedeo Caflisch2*
Department of Chemistry1 and Department of Biochemistry,2 University of Zurich,
Winterthurerstrasse 190, 8057 Zurich, Switzerland
ABSTRACT: Small molecule ligand binding to ATAD2 bromodomain is investigated here
through the synergistic combination of molecular dynamics and protein crystallography.
A previously unexplored conformation of the binding pocket upon rearrangement of the
gatekeeper residue Ile1074 has been found. Further, our investigations reveal how minor
ACS Paragon Plus Environment
1

ACS Medicinal Chemistry Letters
Page 2 of 28
1
2
3
4
5
6
7
8
structural differences in the ligands result in binding with different plasticity of the ZA loop
for this difficult-to-drug bromodomain.
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
KEYWORDS: ATAD2 bromodomain, conformational selection, fuzzy binding, protein X-
ray crystallography, molecular dynamics
Contemporary drug discovery often relies on the availability of structural data to guide
the design of ligands and hit optimization.1 This task becomes challenging when the
protein target is flexible and its conformational changes need to be considered.2,3 Proper
understanding of protein flexibility and conformational selection in small molecule binding
events can be leveraged to the benefit of lead discovery campaigns, in particular with the
help of atomistic simulation approaches.4-14 Reader domains involved in epigenetic
regulation (regulation of gene expression without alterations to the DNA sequence), are
highly flexible and adaptive receptor targets.15-17 Given their potential implication in the
control of genetic information transfer within the cell, the quest for potent and selective
ACS Paragon Plus Environment
2

Page 3 of 28
ACS Medicinal Chemistry Letters
1
2
3
4
5
6
7
8
ligand binders that interact with these proteins has been the focus of intense efforts both
in academia and industry over the past decade.
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 particularly difficult-to-drug target in this series is the ATAD2 (ATPase family AAA
domain-containing 2) bromodomain, a highly flexible and challenging protein-protein
interaction domain implicated in multiple cancers.18-23 Despite its high therapeutic
potential, only two probes have been recently reported.24-27 ATAD2 is considered poorly
druggable mainly due to a shallower, more polar and flexible KAc binding pocket as
compared to other bromodomains. Here, we report a detailed structural investigation on
the binding interactions to the acetyl-lysine (KAc)’s reader ATAD2 through a combination
of molecular dynamics (MD) simulations and protein crystallography for several de novo
designed ligands. Our study sheds light on the multiple factors underlying the plasticity of
this bromodomain, including the local disorder of the hydrophobic side chains involved in
binding as well as the remarkable flexibility and adaptability of the ZA loop.
Our investigation commenced with acetylthiazole hit 1 (Table 1), which was identified
by protein structure-based pharmacophore search in the available commercial chemical
space.28 While several sub-classes of 2-amino thiazole (2-AT) are considered Pan Assay
ACS Paragon Plus Environment
3

ACS Medicinal Chemistry Letters
Page 4 of 28
1
2
3
4
5
6
7
8
9
INterference CompoundS (PAINS), additional experimental and computational evidence
summarized herein below made us confident in our choice of this fragment for further
optimization.29,30 Our key discovery hypothesis was the interaction between the
acetylthiazole and the conserved asparagine residue (Asn1064), as well as between the
primary ammonium group of the alanine moiety and Asp1071 (Figure 1). The stability of
these interactions was confirmed by MD simulations (for a movie of the MD simulation,
see SI), and the binding of compound 1 was validated both experimentally via TR-FRET
(336 µM, see SI) and crystallographically (Fig. 1A). As predicted, 1 occupies the KAc
binding pocket of ATAD2 surrounded by the hydrophobic residues Val1008, Val1013,
Val1018, Tyr1021, Tyr1063 and Ile1074. The amino group of the alanine moiety of 1
indeed reaches out to Asp1071, making a relatively weak salt bridge (3.6 Å).
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
Based on the structure of compound 1 structure, a series of derivatives was designed
and subsequently prepared to explore the binding landscape of this highly flexible and
thus challenging target. The compounds synthesized within this campaign have been
summarized in Table 1 (See experimental section. 2D structures, PDB codes for protein-
ligand complexes as well as IC50 or Kd values measured by TR-FRET or ITC can also be
ACS Paragon Plus Environment
4

Page 5 of 28
ACS Medicinal Chemistry Letters
1
2
3
4
5
6
7
8
9
found in Table 1 and in the SI). Within this compound series, we first set out to rigidify the
structure and improve the salt bridge observed for compound 1 by presenting the
secondary amino group to the Asp1071. To this end a compound bearing a piperazine
was synthesized (2). The X-Ray structure of this ligand in complex with ATAD2 showed
the desired interaction of the piperazine group with Asp1071, with its distal nitrogen atom
forming a higher quality interaction than the original fragment (2.6 Å – Fig. 1B). In contrast,
the proximal nitrogen assumed a close position to the amino group of 1 and did not
engage with the aspartate residue. A guanidinium derivative of 1 was also prepared (3),
which formed a network of hydrogen bonds between Tyr1063 and Asp1071 (Fig. 1C).
This led to rearrangement of the side chains of the gatekeeper residue Ile1074 as well as
that of Val1008, translating into slight remodeling of the binding pocket and tilting of the
thiazole ring to maintain hydrophobic interactions with these residues. Similar
rearrangement of the binding site was seen in the case of a close derivative 4 with an
additional methyl group introduced to improve hydrophobic interactions to Tyr1063 (Fig.
1D). While some interactions seemed to have improved with this approach, no significant
change in binding affinity could be observed for these compounds.
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
5

ACS Medicinal Chemistry Letters
Page 6 of 28
1
2
3
4
5
6
7
8
Table 1. Binders of ATAD2 bromodomain described in this study.
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
R1
S
NH
O
N
R2
O
Com-
poun
d n°
Com-
poun
d n°
1
1
IC50
PDB
IC50
PDB
R1
R2
R1
R2
(µM) code
(µM) code
1
2
3
4
H
H
H
H
336
5F36
10
11
12
13
> 150 6HI7
> 150 6HI8
> 150 6EPT
> 150 6HIA
NH2
NH2
NH2
N
NH
> 150 6EPU
140 6EPX
> 150 6HI3
NH2
HN
NH2
N
NH
NH
NH2
HN
HN
NH2
N
NH
HN
NH2
NH2
NH2
N
S
N
NH
O
5
6
7
> 150 6EPV
> 150 6EPJ
> 150 6HI4
14
15
16
80
6HIB
NH2
NH2
NH2
O
NH
N
O
> 150 6HIC
> 150 6HID
NH2
NH2
HN
HO
O
N
NH
HN
H3CO
N
O
NH
N
15
8
9
> 150 6HI5
17
18
H
6HIE
NH2
F
F
NH2
252
NH
N
N
N
95
6HI6
> 150
/
NH2
F
F
NH2
1 Measured in a Time-Resolved (TR)-FRET assay. 2 Measured by ITC.
ACS Paragon Plus Environment
6

Page 7 of 28
ACS Medicinal Chemistry Letters
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. A-D – Crystal structures of the ATAD2 bromodomain in the complex with
compounds 1-4. 2Fo-Fc electron density maps are shown in blue. In all figures, conserved
water molecules and water molecules involved in bridging polar interactions (dashed
lines) between protein and ligand are shown (red spheres).
To better understand the conformational rearrangements within the pocket and their
potential impact on binding, we reanalyzed large datasets of all-atom simulations that we
previously collected for ATAD2 bromodomain, both of its apo form as well as of the one
ACS Paragon Plus Environment
7

ACS Medicinal Chemistry Letters
Page 8 of 28
1
2
3
4
5
6
7
8
9
bound to KAc-containing peptide.31 The rearranged conformations of residues Val1008
and Ile1074 (Fig. 2A) were populated in the runs with both apo and holo ATAD2 and
appeared independent of the movements of the ZA loop, hinting at a conformational
selection mechanism in ligand binding. Moreover, while both observed conformations of
Val1008 turned out to appear with similar frequency, there was a clear preference for the
conformation of the gatekeeper Ile1074 that opens a remodelled cavity (Fig. 2B). While
this preference was limited in terms of estimated energy difference (~1-2 kcal/mol), we
envisaged the possibility of attaining better ligands by binding to the more populated
conformation of the gatekeeper Ile1074.
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
8

Page 9 of 28
ACS Medicinal Chemistry Letters
1
2
3
4
5
6
7
8
9
A
C
B
Ile1074
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
Val1008
Figure 2. A. Superposition of structures of compounds 1 and 3 (shown in black). B.
Populations of different conformers of the residues Val1008 and Ile1074 according to the
MD simulations. Individual data points represent the conformational states in crystal
structures deposited in the PDB as of 01.03.2019. C. Co-crystal structure of compound 5
bound to ATAD2 bromodomain.
We decided to cyclize the original fragment 1 and ‘grow’ it deeper into the pocket to
gain stabilization of the acetyl group, whose high mobility has been observed in the MD
simulations. (see Supporting Information for details). Furthermore, most bromodomains,
ACS Paragon Plus Environment
9

ACS Medicinal Chemistry Letters
Page 10 of 28
1
2
3
4
5
6
7
8
9
like ATAD2, contain up to four water molecules in their KAc binding pocket. As shown by
others32 different acetyl mimic can perturb this water network and induce selectivity over
other bromodomains. Therefore, we hypothesized that improved selectivity could be
attained for a reshaped pocket by the cyclic acetyl mimic, as suggested by docking.
Compound 5 was synthesized confirming this hypothesis according to the X-Ray structure
shown in Fig. 2C. Still, no appreciable affinity improvement could be quantified for this
compound and thus alternative modifications of the acetylthiazole core were
subsequently explored.
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
Based on the initial exploration of the ligand blueprints and interactions within the
binding site, the molecules were grown in the direction of the ZA loop. Recently, our
groups have reported a novel molecular diversification software named AutoCouple.33
AutoCouple generates combinatorial virtual library of binder candidates that are novel and
synthetically accessible in one formal step by coupling a validated fragment with
commercially available building blocks. The library, fairly exhaustive in terms of easily
accessible chemical derivative space, is then evaluated for possible binding to the target
by the means of tethered docking, and candidates most likely to bind are shortlisted.
ACS Paragon Plus Environment
10

Page 11 of 28
ACS Medicinal Chemistry Letters
1
2
3
4
5
6
7
8
9
Guided by these predictions, we prepared a series of derivatives of compound 1 via Pd-
catalyzed Suzuki cross-coupling reactions (compounds 6-10) to pick up additional
interactions with the ZA loop while maintaining the key interaction cascade to Asp1071
and Asn1067.
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
This focused library provided insight into conformational changes upon ligand binding.
The same rearrangement of the side chains of the gatekeeper residue Ile1074 and
Val1008, as well as the increased size of the binding pocket and slight tilting of the
thiazole ring were observed for example for compound 6. In addition, the phenol moiety
engages a superficial solvent-exposed hydrogen bond to Arg1007 (Figure 3A).
Interestingly, compound 7, a close derivative of 6, lacks the stabilizing interaction with
Arg1007 due to the methylation of the hydroxyl group, and its aromatic moiety is
disordered in the corresponding co-crystal structure (Fig. S1A).
ACS Paragon Plus Environment
11

ACS Medicinal Chemistry Letters
Page 12 of 28
1
2
3
4
5
A
B
6
7
8
9
Val1013
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
Asp1014
Arg1007
C
Val1013
Asp1014
Figure 3. A-B – Crystal structures of the ATAD2 bromodomain in the complex with
compounds 6 and 8, respectively. C – Different binding modes seen in the all-atom MD
simulations of compound 8-ATAD2 bromodomain complex: left – van der Waals
interaction between the side chain of Val1013 and amino-pyridine, right – hydrogen bond
between the backbone NH of Asp1014 and nitrogen atom of pyridine ring.
Some of these molecules bound to the enlarged conformation of the pocket but without
engaging the ZA loop. In these cases, both the loop and aromatic moiety appear to be
ACS Paragon Plus Environment
12

Page 13 of 28
ACS Medicinal Chemistry Letters
1
2
3
4
5
6
7
8
9
disordered, as for pyridine-containing molecule 8 (Fig. 3B). Significant mobility of the ZA
loop is expected, based on previous MD simulations.31 To understand the interaction
mode of 8 and the higher mobility of the ZA loop, a series of all-atom simulations of the
complex were performed, revealing the larger disorder in the ZA loop, as compared to
simulations previously reported by us.31 While the key interactions of 8 in the binding
pocket remained stable for more than 150 ns (e.g. salt bridge to Asp1071), multiple
interaction modes of the aromatic substituent with the ZA loop emerge from the
simulations (Fig. 3C). These include the direct hydrogen bond between the nitrogen atom
of the pyridine and the main chain NH of Asp1014, and the hydrophobic contact between
the aromatic ring and the side chain of Val1013. These interactions engage and
disengage over the course of the simulation without a single major defined conformation,
which is reminiscent of fuzzy binding. Similarly to compound 8, compound 9 bears a
pyridine substituent that cannot be seen in the electron density in the co-crystal structure,
along with part of the ZA loop. In contrast, the ZA loop and the R1 substituent are ordered
for 10, which bears a 3-aminobenzene substituent, and shows a binding mode similar to
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
13

ACS Medicinal Chemistry Letters
Page 14 of 28
1
2
3
4
5
6
7
8
the phenol derivative 6 (Figure S1C). Thus, the promiscuous fuzzy binding behavior
seems to be induced by the pyridine moiety.
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
It should be noted that fuzzy or disordered binding is usually avoided in ligand design
to favor well-defined enthalpy-driven interactions, which provide potency and selectivity.
To study the relationship between the compound structure and the fuzzy binding mode
further, molecules containing an aminopyridine moiety were extended by merging them
with different motifs designed to better engage the Asp1071 residue. As these molecules
might be able to select different conformations of the binding pocket, we envisaged it
would be worth to recapitulate how this affects the fuzzy binding behavior. As seen before,
compound 4 selects for enlarged conformation of the pocket and would project the
aromatic group alongside the same vector as in 8. As expected, the aminopyridine-
bearing derivative 11 exhibits very similar fuzzy binding to that seen for 8 (Figure 4). In
contrast, both derivatives of 2 and 3 (12 and 13, respectively), show ordered binding of
the aminopyridine group as predicted in the original docking efforts. Importantly, 12 binds
to the same conformation of the pocket as 2, and introduction of the aromatic moiety does
not shift its conformational selectivity, unlike in the case of 1 and 8. It is clear from these
ACS Paragon Plus Environment
14

Page 15 of 28
ACS Medicinal Chemistry Letters
1
2
3
4
5
6
7
8
9
results that the fuzzy binding behavior arises as a result of very delicate interplay of
multiple factors, so that even small modifications to the ligand may disrupt such binding
mode, which further complicates the structure-based design against such a highly mobile
target. To slightly modulate interactions with Tyr1063, the alanine moiety was modified
with a highly strained cyclobutyl ring (14), which not only led to ordering of the
aminopyridine moiety, but also shifted the conformational equilibrium of the gatekeeper
residue Ile1074 and slightly improved binding affinity for this compound.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A
B
C
D
Figure 4. A-D – Crystal structures of the ATAD2 bromodomain in the complex with
compounds 11-14.
ACS Paragon Plus Environment
15

ACS Medicinal Chemistry Letters
Page 16 of 28
1
2
3
4
5
6
7
8
9
A second AutoCouple campaign was launched to design derivatives that would engage
the ZA loop (see Supp. Fig. 2 for co-crystal structures of ATAD2 bromodomain with
compounds 15 and 16). Unfortunately, no significant gain in affinity was observed, which
showcased the intrinsic difficulty associated with ordering this highly flexible loop (i.e.
overcoming a high entropic penalty).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Next, we concentrated on further extending these molecules so that they could pick up
interactions with Arg1077, in line with improved affinities observed for compounds
previously reported by others (see supporting information for IC50 in our TR-FRET
assay).24 The proximal nitrogen atom in compound 2, which does not engage Asp1071
was alkylated with an ethyl-linked cyclohexyl ring scaffold yielding ligand 17. This
derivative forms a favorable charge-dipole interaction with Arg1077 through its CF2 group
(Figure 5). The rigidity of the ethyl difluorocyclohexyl group also allowed to maintain the
Ile1074 conformation with a smaller opening of the pocket. This molecule is also the most
potent derivative in this series, with IC50 of 15 µM in our TR-FRET assay and Kd of 25 µM
in ITC. In contrast to observations made for the more potent previously reported chemical
probe,24 further modifications tackling the interactions with the ZA loop by introduction of
ACS Paragon Plus Environment
16

Page 17 of 28
ACS Medicinal Chemistry Letters
1
2
3
4
5
6
7
8
9
an aminopyridine moiety (compound 18) substantially lowered the potency, thus
highlighting how not only the extensions, but the molecule’s core need to synergize to
achieve substantial gain in affinity for the ATAD2 bromodomain.
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
In conclusion, we have extensively explored conformational selection of the ATAD2
bromodomain. The MD trajectories and crystal structures allow us to formulate the
following conclusions. First, the key hydrophobic residues in the KAc binding site, i.e., the
gatekeeper Ile1074, Val1008, and Val1013, can assume multiple orientations for different
ligands and even for the same ligand. The local disorder of the hydrophobic side chains
involved in binding is observed in the simulations (carried out with compounds 5 and 8)
and the crystal structures with ligands 7, 8, 9, and 11. Such disorder is congruent with
sub-optimal interactions of high micromolar affinity binders. Second, the intrinsic plasticity
of the ZA loop results in fuzzy binding or well-defined packing onto the ligand depending
on fine details of the latter. As an example, compounds 8 and 10 differ by a single atomic
element and show fuzzy binding and single well-defined pose, respectively. Note that this
variability is not a consequence of differences in packing between protein molecules in
the crystal as all structures belong to the same space group. While the multiple
ACS Paragon Plus Environment
17

ACS Medicinal Chemistry Letters
Page 18 of 28
1
2
3
4
5
6
7
8
9
orientations of the hydrophobic side chains have not been reported previously, the
formation of a fuzzy complex and adaptable interface of binding, which is mainly due to
the flexibility of the ZA loop, have been observed in atomistic simulations of the histone
H4 peptide/ATAD2 bromodomain complex.31 The coupling of the ZA-loop rigidification
and the rearrangement of the hydrophobic side chains involved in binding can be explored
next as a potential strategy to attain potent and selective ligands for this highly mobile
target.
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
18

Page 19 of 28
ACS Medicinal Chemistry Letters
1
2
3
4
5
6
7
8
9
Figure 5. Crystal structure of the ATAD2 bromodomain in the complex with compound 17
(A) and the GSK8814 chemical probe (B). C – Superposition of compound 17 and the
GSK8814 chemical probe (shown in black) in the binding site. D – 2D structure of
GSK8814 and binding affinity determined in house via TR-FRET assay.
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
ASSOCIATED CONTENT
Supporting Information
The Supporting Information is available free of charge on the ACS Publications website.
Experimental procedures, crystallographic and biophysical data, as well as details of the
organic compound characterization and molecular dynamics simulations (PDF)
AUTHOR INFORMATION
Corresponding Author
*Phone: +41 44 635 55 21. E-mail: caflisch@bioc.uzh.ch.
*Phone: +41 44 635 39 45. E-mail: cristina.nevado@chem.uzh.
ACS Paragon Plus Environment
19

ACS Medicinal Chemistry Letters
Page 20 of 28
1
2
3
4
5
*Phone: +41 44 635 55 87. E-mail: p.sledz@bioc.uzh.ch.
6
7
8
9
Present Addresses
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
†Present address: Astex Therapeutics, Cambridge, UK
Author Contributions
‡These authors contributed equally.
Funding Sources
This work has been funded by the Swiss National Science Foundation (Project Funding
to AC) and Swiss Cancer League - Krebsliga (to CN and AC). PŚ is grateful to the UZH
BioEntrepreneurship program for financial support.
Notes
The authors declare no competing financial interests.
ACKNOWLEDGMENT
The use of beamlines at the Swiss Light Source (Paul Scherrer Institut, Villigen,
Switzerland) is gratefully acknowledged. We thank Cassiano Langini for help with
ACS Paragon Plus Environment
20

Page 21 of 28
ACS Medicinal Chemistry Letters
1
2
3
4
5
6
7
8
9
analysis of the MD data, and Maurus Mathis and Johanne Ling for their help with the
synthesis and characterization of the compounds on the outset of the project. We are
grateful to ChemAxon for providing academic licence for Marvin version 16.2.15.0
(2016), which was used for drawing, displaying and characterizing chemical structures
and protonation states, as well as generating conformers (www.chemaxon.com).
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
ABBREVIATIONS
ATAD2, ATPase family AAA domain-containing 2
PDB CODES
The PDB codes of ATAD2 in complex with the following compounds are listed herein
below:
1 5F36; 2 6EPU; 3 6EPX; 4 6HI3; 5 6EPV; 6 6EPJ; 7 6HI4; 8 6HI5; 9 6HI6; 10 6HI7;
11 6HI8; 12 6EPT; 13 6HIA; 14 6HIB; 15 6HIC; 16 6HID; 17 6HIE.
Authors will release the atomic coordinates and experimental data upon article
publication.
REFERENCES
(1)
Blundell, T. L. Structure-Based Drug Design. Nature 1996, 384, 23–26.
ACS Paragon Plus Environment
21

ACS Medicinal Chemistry Letters
Page 22 of 28
1
2
3
4
5
(2)
Arkin, M. R.; Randal, M.; DeLano, W. L.; Hyde, J.; Luong, T. N.; Oslob, J. D.;
6
7
8
Raphael, D. R.; Taylor, L.; Wang, J.; McDowell, R. S.; et al. Binding of Small Molecules
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
to an Adaptive Protein-Protein Interface. Proc. Natl. Acad. Sci. 2003, 100, 1603–1608.
(3)
Wells, J. A.; McClendon, C. L. Reaching for High-Hanging Fruit in Drug
Discovery at Protein-Protein Interfaces. Nature. 2007, 450, 1001–1009.
(4)
Śledź, P.; Caflisch, A. Protein Structure-Based Drug Design: From Docking to
Molecular Dynamics. Curr. Opin. Struct. Biol. 2018, 48, 93–102.
(5)
Durrant, J. D.; McCammon, J. A. Molecular Dynamics Simulations and Drug
Discovery. BMC Biol. 2011, 9, 71.
(6)
Ganesan, A.; Coote, M. L.; Barakat, K. Molecular Dynamics-Driven Drug
Discovery: Leaping Forward with Confidence. Drug Discovery Today 2017, 22, 249–
269.
(7)
Zhao, H.; Caflisch, A. Molecular Dynamics in Drug Design. Eur. J. Med. Chem.
2015, 91, 4–14.
ACS Paragon Plus Environment
22

Page 23 of 28
ACS Medicinal Chemistry Letters
1
2
3
4
(8)
Su, J.; Liu, X.; Zhang, S.; Yan, F.; Zhang, Q.; Chen, J. Insight Into Selective Mechanism
5
6
7
8
of Class of I‐BRD9 Inhibitors Toward BRD9 Based on Molecular Dynamics Simulations. Chem.
Biol. Drug Des. 2019, 93, 163–176.
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
(9)
Su, J.; Liu, X.; Zhang, S.; Yan, F.; Zhang, Q.; Chen, J. A Theoretical Insight Into Selectivity
of Inhibitors Toward Two Domains of Bromodomain‐Containing Protein 4 Using Molecular
Dynamics Simulations. Chem. Biol. Drug Des. 2018, 91, 828–840.
(10) Huang, D.; Rossini, E.; Steiner, S.; Caflisch, A. Structured Water Molecules in the Binding
Site of Bromodomains Can Be Displaced by Cosolvent. ChemMedChem 2014, 9, 573–579.
(11) Magno, A.; Steiner, S.; Caflisch, A. Mechanism and Kinetics of Acetyl-Lysine Binding to
Bromodomains. J. Chem. Theory Comput. 2013, 9, 4225–4232.
(12) Steiner, S.; Magno, A.; Huang, D.; Caflisch, A. Does Bromodomain Flexibility Influence
Histone Recognition? FEBS Letters 2013, 587, 2158–2163.
(13) Bacci, M.; Langini, C.; Vymětal, J.; Caflisch, A.; Vitalis, A. Focused Conformational
Sampling in Proteins. J. Chem. Phys. 2017, 147, 195102.
(14) Gay, J. C.; Eckenroth, B. E.; Evans, C. M.; Langini, C.; Carlson, S.; Lloyd, J. T.; Caflisch,
A.; Glass, K. C. Disulfide Bridge Formation Influences Ligand Recognition by the ATAD2
Bromodomain. Proteins 2019, 87, 157–167.
(15) Jain, A. K.; Barton, M. C. Bromodomain Histone Readers and Cancer. J. Mol.
Biol. 2017, 429, 2003–2010.
ACS Paragon Plus Environment
23

ACS Medicinal Chemistry Letters
Page 24 of 28
1
2
3
4
5
(16) Muller, S.; Filippakopoulos, P.; Knapp, S. Bromodomains as Therapeutic
6
7
8
Targets. Expert Rev. Mol. Med. 2011, 13, e29.
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
(17) Ferri, E.; Petosa, C.; McKenna, C. E. Bromodomains: Structure, Function and
Pharmacology of Inhibition. Biochem. Pharmacol. 2016, 106, 1–18.
(18) Hwang, H. W.; Ha, S. Y.; Bang, H.; Park, C. K. ATAD2 as a Poor Prognostic
Marker for Hepatocellular Carcinoma after Curative Resection. Cancer Res. Treat.
2015, 47, 853–861.
(19) Morozumi, Y.; Boussouar, F.; Tan, M.; Chaikuad, A.; Jamshidikia, M.; Colak, G.;
He, H.; Nie, L.; Petosa, C.; De Dieuleveult, M.; et al. ATAD2 Is a Generalist Facilitator of
Chromatin Dynamics in Embryonic Stem Cells. J. Mol. Cell Biol. 2016, 8, 349–362.
(20) Ciró, M.; Prosperini, E.; Quarto, M.; Grazini, U.; Walfridsson, J.; McBlane, F.;
Nucifero, P.; Pacchiana, G.; Capra, M.; Christensen, J.; et al. ATAD2 Is a Novel
Cofactor for MYC, Overexpressed and Amplified in Aggressive Tumors. Cancer Res.
2009, 69, 8491–8498.
ACS Paragon Plus Environment
24

Page 25 of 28
ACS Medicinal Chemistry Letters
1
2
3
4
5
6
7
8
9
(21) Zhang, M. J.; Zhang, C. Z.; Du, W. J.; Yang, X. Z.; Chen, Z. P. ATAD2 Is
Overexpressed in Gastric Cancer and Serves as an Independent Poor Prognostic
Biomarker. Clin. Transl. Oncol. 2016, 18, 776–781.
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
(22) Hou, M.; Huang, R.; Song, Y.; Feng, D.; Jiang, Y.; Liu, M. ATAD2
Overexpression Is Associated with Progression and Prognosis in Colorectal Cancer.
Jpn. J. Clin. Oncol. 2016, 46, 222–227.
(23) Caron, C.; Lestrat, C.; Marsal, S.; Escoffier, E.; Curtet, S.; Virolle, V.; Barbry, P.;
Debernardi, A.; Brambilla, C.; Brambilla, E.; et al. Functional Characterization of ATAD2
as a New Cancer/Testis Factor and a Predictor of Poor Prognosis in Breast and Lung
Cancers. Oncogene 2010, 29, 5171–5181.
(24) Bamborough, P.; Chung, C. wa; Demont, E. H.; Furze, R. C.; Bannister, A. J.;
Che, K. H.; Diallo, H.; Douault, C.; Grandi, P.; Kouzarides, T.; et al. A Chemical Probe
for the ATAD2 Bromodomain. Angew. Chemie. Int. Ed. 2016, 55, 11382–11386.
ACS Paragon Plus Environment
25

ACS Medicinal Chemistry Letters
Page 26 of 28
1
2
3
4
5
6
7
8
9
(25) Fernández-Montalván, A. E.; Berger, M.; Kuropka, B.; Koo, S. J.; Badock, V.;
Weiske, J.; Puetter, V.; Holton, S. J.; Stöckigt, D.; Ter Laak, A.; et al. Isoform-Selective
ATAD2 Chemical Probe with Novel Chemical Structure and Unusual Mode of Action.
ACS Chem. Biol. 2017, 12, 2730–2736.
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
(26) Bamborough, P.; Chung, C. W.; Furze, R. C.; Grandi, P.; Michon, A. M.; Watson,
R. J.; Mitchell, D. J.; Barnett, H.; Prinjha, R. K.; Rau, C.; et al. Aiming to Miss a Moving
Target: Bromo and Extra Terminal Domain (BET) Selectivity in Constrained ATAD2
Inhibitors. J. Med. Chem. 2018, 61, 8321–8336.
(27) Bamborough, P.; Chung, C. W.; Demont, E. H.; Bridges, A. M.; Craggs, P. D.;
Dixon, D. P.; Francis, P.; Furze, R. C.; Grandi, P.; Jones, E. J.; et al. A Qualified Success:
Discovery of a New Series of ATAD2 Bromodomain Inhibitors with a Novel Binding Mode
Using High-Throughput Screening and Hit Qualification. J. Med. Chem. 2019, 62, 7506–
7525.
(28) Sunseri, J.; Koes, D. R. Pharmit: Interactive Exploration of Chemical Space.
Nucleic Acids Res. 2016, 44, W442–W448.
ACS Paragon Plus Environment
26

Page 27 of 28
ACS Medicinal Chemistry Letters
1
2
3
4
5
6
7
8
9
(29) Baell, J. B.; Holloway, G. A. New Substructure Filters for Removal of Pan Assay
Interference Compounds (PAINS) From Screening Libraries and for Their Exclusion in
Bioassays. J. Med. Chem. 2010, 53, 2719–2740.
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
(30). A recent study on 2-AT containing fragments and HTS hits revealed that not all
2-AT molecules are problematic and that acylation of the 2-amino group, as in our case,
reduces substantially the risk of pan hitting: Devine, S. M.; Mulcair, M. D.; Debono, C.
O.; Leung, E. W. W.;Nissink, J. W. M.; Lim, S. S.; Chandrashekaran, I. R.; Vazirani,
M.;Mohanty, B.; Simpson, J. S.; et al. Promiscuous 2-Aminothiazoles (PrATs): A
FrequentHitting Scaffold. J. Med. Chem. 2015, 58, 1205−1214.
(31) Langini, C.; Caflisch, A.; Vitalis, A. The ATAD2 Bromodomain Binds Different
Acetylation Marks on the Histone H4 in Similar Fuzzy Complexes. J. Biol. Chem. 2017,
292, 16734–16745.
(32) Crawford, T. D.; Tsui, V.; Flynn, E. M.; Wang, S.; Taylor, A. M.; Côté, A.; Audia,
J. E.; Beresini, M. H.; Burdick, D. J.; Cummings, R.; Dakin, L. A.; Duplessis, M.; Good, A.
C.; Hewitt, M. C.; Huang, H.-R.; Jayaram, H.; Kiefer, J. R.; Jiang, Y.; Murray, J.;
ACS Paragon Plus Environment
27

ACS Medicinal Chemistry Letters
Page 28 of 28
1
2
3
4
5
6
7
8
9
Nasveschuk, C. G.; Pardo, E.; Poy, F.; Romero, F. A.; Tang Y.; Wang, J.; Xu, Z.;
Zawadzke, L. E.; Zhu, X.; Albrecht, B. K.; Magnuson, S. R.; Bellon, S.; Cochran, A. G.
Diving into the Water: Inducible Binding Conformations for BRD4, TAF1(2), BRD9, and
CECR2 Bromodomains. J. Med. Chem. 2016, 59, 5391–5402.
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
(33) Batiste, L.; Unzue, A.; Dolbois, A.; Hassler, F.; Wang, X.; Deerain, N.; Zhu, J.;
Spiliotopoulos, D.; Nevado, C.; Caflisch, A. Chemical Space Expansion of
Bromodomain Ligands Guided by in Silico Virtual Couplings (AutoCouple). ACS Cent.
Sci. 2018, 4, 180–188.
Table of contents graphics
ACS Paragon Plus Environment
28