Direct NMR Probing of Hydration Shells of Protein Ligand Interfaces and Its Application to Drug Design

Article abstract of DOI:10.1021/acs.jmedchem.7b00845

Fragment-based drug design exploits initial screening of low molecular weight compounds and their concomitant affinity improvement. The multitude of possible chemical modifications highlights the necessity to obtain structural information about the binding mode of a fragment. Herein we describe a novel NMR methodology (LOGSY titration) that allows the determination of binding modes of low affinity binders in the protein-ligand interface and reveals suitable ligand positions for the addition of functional groups that either address or substitute protein-bound water, information of utmost importance for drug design. The particular benefit of the methodology and in contrast to conventional ligand-based methods is the independence of the molecular weight of the protein under study. The validity of the novel approach is demonstrated on two ligands interacting with bromodomain 1 of bromodomain containing protein 4, a prominent cancer target in pharmaceutical industry.

Full text of DOI:10.1021/acs.jmedchem.7b00845

Subscriber access provided by FLORIDA ATLANTIC UNIV
Article
Direct NMR Probing of Hydration Shells of Protein
Ligand Interfaces and its Application to Drug Design
Leonhard Geist, Moriz Mayer, Xiao-Ling Cockcroft, Bernhard Wolkerstorfer,
Dirk Kessler, Harald Engelhardt, Darryl B. McConnell, and Robert Konrat
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b00845 • Publication Date (Web): 14 Sep 2017
Downloaded from http://pubs.acs.org on September 15, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 31
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Direct NMR Probing of Hydration Shells of Protein
Ligand Interfaces and its Application to Drug
Design
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
Leonhard Geist†, Moriz Mayer‡, Xiao-Ling Cockcroft‡ , Bernhard Wolkerstorfer‡, Dirk Kessler‡ ,
Harald Engelhardt‡ , Darryl B. McConnell‡ and Robert Konrat*†
† Christian Doppler Laboratory for HighꢀContent Structural Biology and Biotechnology /
Department of Structural and Computational Biology, University of Vienna, Campus Vienna
Biocenter 5, Aꢀ 1030 Vienna, Austria
‡ Boehringer Ingelheim RCV GmbH & Co KG, Dr.ꢀBoehringerꢀGasse 5ꢀ11, Aꢀ1121 Vienna,
Austria
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 31
1
2
3
4
5
6
7
8
9
ABSTRACT Fragmentꢀbased drug design exploits initial screening of low molecular weight
compounds and their concomitant affinity improvement. The multitude of possible chemical
modifications highlights the necessity to obtain structural information about the binding mode of
a fragment. Herein we describe a novel NMR methodology – LOGSYꢀtitration – that allows the
determination of binding modes of low affinity binders in the proteinꢀligand interface and reveals
suitable ligand positions for the addition of functional groups that either address or substitute
proteinꢀbound water – information of utmost importance for drug design. The particular benefit
of the methodology and in contrast to conventional ligandꢀbased methods is the independence of
the molecular weight of the protein under study. The validity of the novel approach is
demonstrated on two ligands interacting with Bromodomain 1 of Bromodomain containing
protein 4, a prominent cancer target in pharmaceutical industry.
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
BRD4 • drug design • epitope mapping • NMR spectroscopy • proteinꢀbound water
INTRODUCTION
Rational drug design is an iterative process, in which 3D structural information about a
specific proteinꢀligand interaction is used to generate an inhibitor that specifically blocks a
targeted interaction. NMR spectroscopy provides wellꢀestablished methods for initial library
screening, as saturation transfer difference (STD),1,2 water ligand observed via gradient
spectroscopy (WaterLOGSY),3,4 fluorine NMR (19FꢀNMR),5 structureꢀactivity relationship by
ACS Paragon Plus Environment

Page 3 of 31
Journal of Medicinal Chemistry
1
2
NMR (SAR by NMR).6 Additionally, NMR methods were developed that allow the
characterization of the ligandꢀorientation in the proteinꢀbound state, e.g. adiabatic fast passage ꢀ
nuclear Overhauser effect spectroscopy (AFPꢀNOESY),7 interligand NOE (ILOE) for
pharmacophore mapping (INPHARMA),8 solvent accessibility, ligand binding, and mapping of
ligand orientation by NMR (SALMON),9 group epitope mapping (GEM) by STD,1 difference of
inversion recovery rate with and without target irradiation (DIRECTION),10 transferred nuclear
Overhauser effect (NOE) spectroscopy.11,12 Since Xꢀray coꢀcrystallographic structure
determination of proteinꢀligand complexes of initial hits is sometimes not feasible, independent
techniques are required that provide specific information about ligand epitopes, as well as the
protein surroundings of the bound ligand, to aid the systematic optimization of compounds.
Especially, knowledge of ligand positions that are close to proteinꢀbound water molecules will be
very valuable for the choice of chemical modification to be introduced at certain ligand
positions.
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
In our study we employ GEM by STD1 and a novel application of the WaterLOGSY4 method
called LOGSYꢀtitration to distinguish between buried and solventꢀaccessible parts of ligands, as
well as to map proteinꢀbound water molecules that are in proximity to the ligand. As published,
WaterLOGSY4 experiments also allow the extraction of information about solvent accessible
epitopes of a ligand. The SALMON approach uses the sign of a WaterLOGSY cross peak in the
spectrum of the ligand in presence of the protein as a measure of solvent accessibility.2,9
Negative cross peaks are observed only for highly solvent exposed ligand protons, whereas
positive cross peaks stem from buried ligand protons. We noticed that for weakly binding
ligands, especially in combination with small proteins, positive WaterLOGSY cross peaks are
hardly observed, which hampers the extraction of differential solvent accessibility scores with
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 31
1
2
3
4
5
6
7
8
9
the SALMON methodology. Therefore, we propose performing WaterLOGSY experiments at
increasing protein concentrations and using the titration slopes as a measure of solvent
accessibility. A major advantage of the LOGSYꢀtitration method is its applicability to small
proteins, as it probes the influence of the protein on the solvation of the ligand. The potential of
this new methodology is illustrated with an application to the Bromodomain 1 of Bromodomain
containing protein 4 (Brd4ꢀBD1).13 Brd4 is a member of the bromodomain and extraꢀterminal
(BET) family that contains two tandem bromodomains (BD1 and BD2) and binds to acetylated
lysines in histones.13,14 Inhibition of Brd4ꢀchromatin interaction is relevant for the treatment of
various forms of cancer and has been shown to have potent antiproliferative effects in a variety
of hematological cancers through the suppression of cꢀMYC and FOSL1 transcription.8,15ꢀ17
We successfully could map buried and solventꢀaccessible epitopes in two ligands interacting
with Brd4ꢀBD1, revealing an unexpected flip of a core aromatic scaffold shared by the two
ligands. Furthermore, we predicted ligand positions, which are in proximity to proteinꢀbound
water. Concomitant introduction of a secondary amine at a suggested position led to significant
improvement of affinity of ligands by replacing weakly interacting H2O and forming a hydrogen
bond to a nearby asparagine residue.
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
Determination of ligand binding modes with LOGSY-titrations:
Figure 1 illustrates the LOGSYꢀtitration experiment. For a reversibly binding ligand the
magnetization transferred by crossꢀrelaxation from irradiated or selectively inverted water to
ligand protons is a population weighted average of free (positive) and bound (negative) crossꢀ
relaxation contributions (Figure 1a). Increasing the amount of protein will shift the equilibrium
ACS Paragon Plus Environment

Page 5 of 31
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
towards the proteinꢀbound state and eventually lead to a sign change of the observed cross peaks.
The amount of bound crossꢀrelaxation received via the protein depends on the peculiarities of the
binding pocket (nearby bound water molecules, proton density at the interaction site) as well as
the kinetics of the proteinꢀligand interaction. As individual ligand protons will experience a
different surrounding in the proteinꢀbound state, cross peaks of different ligand protons will
behave differently during the LOGSY titration (Figure 1b).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. Concept of the LOGSYꢀtitration experiment: (a) For reversibly binding ligands (mMꢀ
µM) magnetization received from water during the WaterLOGSY experiment is a combination
of crossꢀrelaxation obtained from bulk water and crossꢀrelaxation transferred either directly via
protein bound water molecules or indirectly via spinꢀdiffusion mediated by protein protons. (b)
LOGSYꢀtitration: Ilogsy/Iref is plotted as a function of the protein concentration. Intensities in the
WaterLOGSY are normalized to the 1H 1D reference spectrum measured at the same protein
concentration. The slopes are fitted to a linear function and normalized to the graph with the
strongest slope (LOGSYꢀfactor 100%).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 31
1
2
3
4
5
6
7
8
9
We used mixing times of τmix = 1s for the WaterLOGSY experiments, where the cross peak
intensities for our ligands have their maximum. Increased amounts of protein will lead to an
increase in macromolecular crossꢀrelaxation from water to the ligand, either directly via protein
bound water molecules or indirectly via spinꢀdiffusion mediated by protein protons. Ligand
protons that are close to the protein in the bound state will show stronger LOGSYꢀtitration slopes
than ligand protons that are solvent accessible in the bound state and not in close contact with the
protein. The fitted slopes are normalized to the signal with the strongest slope (100% LOGSYꢀ
factor) (Figure 1b). Figure S2 in the Supporting Information provides simulations of expected
LOGSYꢀtitration cross peak intensities, demonstrating that at low protein concentrations the
intensity changes approximately linearly with protein concentration.
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
Typically, values around 100% correspond to ligand protons that receive a maximum of
magnetization via the protein or proteinꢀbound water molecules, whereas values smaller than
20% delineate protons that are facing the solvent in the bound state. Intermediate values
correspond to protons positioned between protein and solvent. Importantly, since only the slope
of the titration curve is analyzed, relevant information can be obtained also in cases where no
zero crossing occurs (for example, in case of weak binders in combination with low molecular
weight proteins, as well as ligand signals for highly solvent accessible protons). This feature is
particularly attractive given that in early stages of the drug design process ligands tend to have
moderate affinities.
Figure 2 shows LOGSYꢀtitration factors and STD amplification factors (STDꢀAF0) of two
ligands 1 and 2, which reversibly bind to Brd4ꢀBD1 with IC50 (1) = 24 µM, and IC50 (2) = 2 µM,
respectively. Xꢀray coꢀcrystal structures of both ligands were available and could be compared to
ACS Paragon Plus Environment

Page 7 of 31
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
the NMR derived LOGSY and STD enhancements. In both ligands, protons of the triazoloꢀ
pyridazineꢀmoiety have the largest LOGSY and STDꢀAF0 values, showing that this part of the
ligand is deeply embedded in the proteinꢀbinding site. The different substituents, for example the
methoxy groups 2 and 3 and aromatic protons of the phenylꢀmoiety in ligand 1, as well as
aromatic protons of the phenyl group in ligand 2, display smaller LOGSY and STDꢀAF0 values,
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
Figure 2. LOGSYꢀfactors calculated from the slopes of the LOGSYꢀtitrations and normalized
STDꢀAF0s for (a, b) ligand 1 and (d, e) ligand 2. (c, f) Xꢀray coꢀcrystal structures of the
complexes Brd4ꢀBD1 with ligand 1 (5M39) and ligand 2 (5M3A), respectively.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 31
1
2
3
4
5
6
7
8
Still, while the distribution of STD values is nearly the same for both ligands, LOGSYꢀfactors
reveal significant differences for both ligands and thereby indicate different ligand binding
modes. Most important, while the STDꢀAF0s of the methyl groups 1 of the two ligands are
similar (99% and 100%), they have a significantly different LOGSYꢀfactor. In ligand 2 it is
clearly reduced (58% compared to 100%) and is indicative of a different hydration shell located
around the position of the methyl group (compare Figure 2a and 2d). The binding mode of the
triazoloꢀpyridazineꢀmoiety is indeed flipped, as confirmed in the Xꢀray structure, leading to a
higher solvent accessibility for this methyl group in ligand 2 as shown in Figure 2f. This flip
would not have been predicted from the STD NMR spectra, most likely due to the fact, that the
protein environment with regards to the nonꢀexchangeable protons is similar in the two binding
modes leading to similar STD enhancements. The NMR results are therefore in excellent general
agreement with the position of the ligand in the binding pocket obtained from the Xꢀray crystal
structure.
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
In sum, both STDꢀNMR and LOGSYꢀtitrations provide information about ligand binding
epitopes, however only the LOGSYꢀfactors revealed the flipped binding mode of the triazoloꢀ
pyridazineꢀringsystem of ligand 2. It can thus be concluded that in medicinal chemistry
applications STD data are valuable for the identification of ligand positions in close proximity to
the proteinꢀbinding interface, whereas additional information about their hydration shells can be
obtained using the LOGSYꢀtitration technique.
Determination of protein-bound water in proximity to ligand protons:
ACS Paragon Plus Environment

Page 9 of 31
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
It has to be noted that in the LOGSY experiment the NOESY transfer between ligand and
protein also involves spin diffusion via protein protons in the binding cleft. We therefore
performed LOGSYꢀtitrations with perdeuterated protein to suppress spinꢀdiffusion via the
protein in order to detect direct NOEs from proteinꢀbound water molecules to the ligand (Figure
3a and 3c). Accordingly, large LOGSYꢀfactors are found for protons exhibiting NOEs to
interfacial water molecules (embedded in the interaction interface), while ligand protons
accessible to solvent water molecules display smaller values. Maximal LOGSYꢀfactors were
found for protons experiencing crossꢀrelaxation to ordered water molecules located at the bottom
of the binding cleft (Figure 3b and 3d). Specifically, strong (direct) water NOEs were observed
for methylꢀgroup 1 of ligand 1 and – due to the swapped orientation of the triazoloꢀpyridazineꢀ
ring – to the aromatic proton 8 of ligand 2. Most interestingly, however, a significant water NOE
was found for aromatic proton 8 of ligand 1 (Figure 3b).
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. LOGSYꢀfactors calculated from the slopes of the LOGSYꢀtitrations with perdeuterated
Brd4ꢀBD1 for (a) ligand 1 and (c) ligand 2. 100% corresponds to the signal with the strongest
slope. (b, d) Xꢀray coꢀcrystal structures of the complexes Brd4ꢀBD1 with ligand 1 (5M39) and
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 31
1
2
3
4
ligand 2 (5M3A), respectively. H2O molecules resolved in the electron density map and within a
5
6
distance of 6Å from the ligand (blue spheres).
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
It is instructive to compare our findings with available Xꢀray data from other ligandꢀBrd4
complexes available inꢀhouse. In total, eleven complex structures were examined, in which the
ligands do not contain functional groups corresponding to the position 8, which would interfere
with potential proteinꢀbound water at that site. In all eleven crystal structures highly ordered
water molecules were found in conserved positions of the binding pocket of Brd4ꢀBD1 and close
to methyl group 1 of ligand 1 (Figure 4, blue spheres). The interaction interface contains a
second water cluster although with higher positional variability (Figure 4, yellow spheres) as
well as water molecules in an exposed cluster (Figure 4, red spheres) which were detectable in
only 6 out of 11 complex structures, suggesting that water molecules in this part of the hydration
shell are less ordered and presumably part of a dynamic water cluster. Water molecules located
in this cluster are presumably responsible for the observed water NOE to the aromatic proton at
position 8 of ligand 1. Here it should be noted that there is ample evidence that Xꢀray protein
crystallography does not reliably detect loosely defined hydration water molecules.18 In contrast,
it is well established that NMR also identifies transiently bound or “disordered” water molecules
that are not always resolved in the electron density map.19 Our findings nicely demonstrate that
the LOGSY titration experiment reported here is a powerful and efficient technique to probe the
existence and location of transiently bound water molecules in protein ligand complexes. These
water molecules could be valid starting points for chemical replacement strategies to displace or
substitute bound water molecules by suitable functional groups.
ACS Paragon Plus Environment

Page 11 of 31
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 4. Superposition of H2O molecules in Brd4ꢀBD1 complex structures of 6 different ligand
coꢀcrystal structures (other ligands not shown here). Ligand 1 is shown as representative ligand
structure. Spheres represent H2O molecules found in the Xꢀray structures of the complexes (blue,
highlyꢀordered & consistently found in all complexes; yellow, low positional variation; red, large
positional variation and not conserved in all proteinꢀligand complexes).
We thus suggest exploiting the experiment with perdeuterated protein for the identification of
protons with large LOGSYꢀfactors as the primary candidates for substitution with groups that
either displace or address proteinꢀbound water molecules. Successful applications for this novel
strategy are shown in the supplementary material (Table S1). Overall, the replacement of the
aromatic proton 8 with an aminoꢀ or aminoꢀmethylꢀgroup in a series of ligands containing the
triazoloꢀpyridazineꢀring led up to a 100ꢀfold increase in binding affinity (Table S1). Consistently,
in all Xꢀray coꢀcrystal structures obtained with such a modification, the triazoloꢀpyridazineꢀring
binding mode of ligand 1 was found (methyl group pointing inwards), and the (substituted)
aminoꢀgroup was found to be hydrogenꢀbonded to the sideꢀchain of Asn140 of Brd4ꢀBD1.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 31
1
2
3
4
5
6
DISCUSSION AND CONCLUSIONS
7
8
9
Detailed information about the binding mode of a fragment to its target as well as its
environment in the bound state is indispensable for medicinal chemists to optimize early phase
lead candidates with respect to affinity and specificity. Although Xꢀray crystallography is the
method of choice to provide highꢀresolution structural information, it often fails in early stages
of the drug optimization process because of low binding affinity and concomitantly
compromised protein crystallizability. Additionally, it is well established that water molecules
are often found at the proteinꢀligand interaction interface and contribute to the binding affinity by
forming distinct hydrogen bonding patterns. Despite their crucial role in ligandꢀbinding events
they are sometimes undetectable by Xꢀray crystallography. To overcome this bottleneck we
present a novel ligandꢀbased NMR spectroscopy approach for the probing of ligand binding
modes and the characterization of water hydration shells in proteinꢀligand binding clefts.
Although our methodology shares some similarities with the already mentioned SALMON
approach using short mixing times (τmix = 150 ms) it is important to stress the fundamental
differences.2 The use of short mixing times was shown to be necessary for the extraction of
solvent accessible epitopes of very buried ligands.2 Unfortunately, due to the small size of Brd4ꢀ
BD1 (MW = 15 kDa), it was impossible to record positive WaterLOGSY cross peaks for any of
the ligands. Switching back to longer mixing times (τmix = 1s), we realized that the distribution of
positive and negative cross peaks strongly depends on the affinity of the ligand and on the
protein concentration.
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

Page 13 of 31
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 5. Comparison between SALMON and LOGSYꢀtitration of ligand 2 interacting with
protonated Brd4ꢀBD1. (a) 1D spectrum of ligand 2 (1 mM) and WaterLOGSY spectra of ligand
2 with 24 µM Brd4ꢀBD1 (blue) and 45 µM Brd4ꢀBD1 (red). (b) LOGSYꢀtitration of ligand 2
interacting with Brd4ꢀBD1 (5ꢀ45 µM).
Figure 5a shows the 1D spectrum and two WaterLOGSY spectra of ligand 2 recorded at two
different protein concentrations (24 µM Brd4ꢀBD1 in blue, 45 µM Brd4ꢀBD1 in red). It is
important to note that the distribution of positive and negative signals is fairly different
depending on the protein concentration. As can be seen from the Figure 5a, although the
SALMON approach provides a rough assessment of the ligand binding epitope further details
cannot be concluded from a single point measurement. In contrast, Figure 5b clearly shows that
analysis of the corresponding LOGSYꢀtitration curves reveals far more details about the ligand
binding mode and internally bound water molecules. Most importantly, we could demonstrate
that the method also identifies transiently bound or “disordered” water molecules that are not
always resolved in the electron density map and that these water molecules are valid starting
points for chemical replacement strategies to substitute bound water molecules by suitable
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 31
1
2
3
4
5
6
7
8
9
functional groups. Several studies have shown that hydration water plays a pivotal role not only
for the folding of proteins but also for proteinꢀligand interaction.20 An example for the beneficial
impact of replacing water molecules to binding affinities were provided by Grzesiek and coꢀ
workers, who could show that a carbonyl group of the highꢀaffinity HIV protease inhibitor
DMP323 replaces a conserved H2O molecule, and thereby contributes to the total binding free
energy.21 The possibility of identifying proteinꢀbound water molecules close to the bound ligand
will clearly be of great value for the decision of medicinal chemists to incorporate additional
functional groups, which either address retainable water or replace substitutable water. It is
important to note, that this valuable information is provided without atomistic structural
information commonly provided by Xꢀray crystallography.
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 summary, we present a novel NMR approach for probing solvent accessibility and ligand
binding modes of proteinꢀligand complexes. Most importantly, the method allows the
identification of ligand positions for medicinal chemistry modifications to improve binding
affinities through substituting or addressing bound water molecules. Given the facile
applicability of the approach and the advantage of not needing a complete structure of the protein
or a proteinꢀligand complex, we anticipate wide usage of the method particularly at early stages
of drug discovery and especially for targets where obtaining Xꢀray coꢀcrystal structures is limited
or not possible.
EXPERIMENTAL SECTION
Synthetic methods
ACS Paragon Plus Environment

Page 15 of 31
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
General Remarks. Chemicals and reagents were obtained from commercial suppliers and were
used without further purification. Proton and Carbon NMR spectra were obtained on a Bruker
Avance 400 FTꢀNMR or Bruker Avance 500 FTꢀNMR instrument with chemical shifts (δ)
reported relative to tetramethylsilane as an internal standard. High resolution mass spectroscopy
data were obtained on a LTQ Orbitrap XL (Thermo Scientific) equipped with a NSI Source
(Advion Nanomate) in ESI positive mode. Analytical HPLCꢀMS analyses were conducted using
an Agilent 1100/1200 series LC/MSD system or UPLCꢀmicromass Triple quad. The analytic
method A1, A2 and A3 is defined in Table 1. Compound purities were calculated as the
percentage peak area of the analyzed compound by UV detection at 230ꢀ400 nm. If purity data is
not explicitly mentioned the compound displays a purity > 95%. Flash column chromatography
was carried out using hand packed silica gel 60 (100ꢀ200 mesh) and product was eluted under
medium pressure liquid chromatography. Preparative high performance chromatography was
carried out on a Gilson system (pump system: 333 & 334 prepꢀscale HPLC pump; fraction
collector: 215 liquid handler, detector: Gilson UV/VIS 155) using preꢀpacked reversed phase
silica gel columns from Waters. The methods for preparative high performance chromatography
P1_a, P1_b and P2 are defined in Table 1.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Selected Experimental Procedures.
Synthesis of ligand 1
Scheme S1 in the Supporting Information shows the structures of the compounds involved in
the synthesis of ligand 1.
6-(3,4-Dimethoxy-phenyl)-3-methyl-[1,2,4]triazolo[4,3-b]pyridazine (1)
6ꢀChloroꢀ3ꢀmethylꢀ[1,2,4]triazolo[4,3ꢀb]pyridazine
(30
mg;
0.178
mmol),
3,4ꢀ
Dimethoxyphenylboronic acid (35 mg; 0.19 mmol), caesium carbonate (139 mg; 0.43 mmol) and
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 31
1
2
3
4
5
6
7
8
9
1,1'ꢀ(bis(diphenylphosphino)ferrocenepalladium(II) dichloride, dichloromethane (15 mg; 0.02
mmol) was dissolved in 0.2 mL THF. The mixture stirred for 30min at 100 °C. The crude
product was purified using method P1_a, yielding 0.04 g (38%, 0.12 mmol) of the title
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
Purity by method A3: > 95%, RT = 0.82 min; MS (ESI+) m/z 271 (M+H)+; HRMS (m/z): [M]+
calcd. for C14H14N4O2, 270.1117; found, 270.1111
1H NMR (DMSOꢀd6, 500 MHz): δ = 8.46 (d, J = 9.8 Hz, 1H), 8.16 (d, J = 9.8 Hz, 1H), 7.78 (dd,
J = 8.5, 2.2 Hz, 1H), 7.72 (d, J = 2.2 Hz, 1H), 7.16 (d, J = 8.5 Hz, 1H), 3.90 (s, 3H), 3.86 (s, 3H),
2.82 ppm (s, 3H)
13C NMR (DMSOꢀd6, 125 MHz): δ = 153.3, 151.5, 149.1, 146.7, 142.2, 125.8, 123.8, 121.4,
120.9, 111.8, 110.1, 55.7, 55.6, 9.2 ppm
Synthesis of ligand 2
Scheme S2 in the Supporting Information shows the structures of the compounds involved in
the synthesis of ligand 2.
1-Methyl-5-phenoxy-1H-pyrazole-4-carboxylic acid ethyl ester (2)
Phenol (18.17 g; 193.08 mmol) was dissolved in 500 mL DMA, potassium carbonate (53.37 g;
386.16 mmol) was added in small portions and stirred for 10 min at room temperature.
Afterwards 5ꢀBromoꢀ1ꢀmethylꢀ1Hꢀpyrazoleꢀ4ꢀcarboxylic acid ethyl ester (30.00 g; 128.72
mmol) was added to the reaction mixture dropwise and stirred for 16 h at 140 °C. 10% citric acid
solution and dichloromethane was added. The organic layer was separated, dried and the solvent
ACS Paragon Plus Environment

Page 17 of 31
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
was removed under reduced pressure. The crude product was purified using silica gel flash
column chromatography with a solvent mixture of ethyl acetate and hexane for elution. The
solvent of the corresponding fractions was evaporated under reduced pressure, yielding 18 g
(57%, 73.0 mmol) of the title 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
Purity by method A2: >95%; RT = 1.07 min, MS (ESI+) m/z 247 (M+H)+
1H NMR (400 MHz, CDCl3) δ (ppm) 7.92 (s, 1H), 7.28ꢀ7.22 (m, 2H), 7.14ꢀ7.05 (m, 1H), 6.92ꢀ
6.86 (d, 2H), 4.15ꢀ4.02 (m, 2H), 3.71 (s, 3H), 1.02ꢀ1.10 (m, 3H)
1-Methyl-5-phenoxy-1H-pyrazole-4-carboxylic acid (3)
1ꢀMethylꢀ5ꢀphenoxyꢀ1Hꢀpyrazoleꢀ4ꢀcarboxylic acid ethyl ester 3 (0.45 g; 1.83 mmol) was
dissolved in 10 mL of a 1 : 1 mixture of methanol and THF. Lithium hydroxide (0.15 g; 3.66
mmol) in 5 mL water was added and the reaction mixture was stirred for 16 h at 20 °C. Ethyl
acetate was added to the mixture, the aqueous layer was acidified with 1 M aqueous hydrochloric
acid. The organic layer was removed under reduced pressure. The crude product yields 0.25 g
(63%, 1.15 mmol) of the title compound.
Purity by method A1: >95%; RT = 2.92 min, MS (ESI+) m/z 219 (M+H)+
1H NMR (400 MHz, CDCl3) δ (ppm) 9.80ꢀ12.20 (br, 1H), 7.92 (s, 1H), 7.34ꢀ7.22 (m, 2H), 7.12
(m, 1H), 6.92ꢀ6.85 (m, 2H), 2.65 (s, 3H)
4-Iodo-1-methyl-5-phenoxy-1H-pyrazole (5)
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 31
1
2
3
4
5
6
7
8
9
Under an athmosphere of Argon 1ꢀMethylꢀ5ꢀphenoxyꢀ1Hꢀpyrazoleꢀ4ꢀcarboxylic acid 4 (20.00
g; 91.66 mmol) were suspended in 6.40 mL tertꢀbutanol and 160 mL 1,4ꢀdioxane. DIPEA (45.56
m; 274.97 mmol), diphenyl phosphorazidate (52.97 g; 192.48 mmol) were added to the mixture
and stirred for 10 min at 20°C. Afterwards the mixture was stirred for 3 h at 110 °C. The solvent
was removed under reduced pressure. The residue was purified using silica gel flash column
chromatography with a solvent mixture of ethyl acetate and hexane for elution. The solvent of
the corresponding fractions was evaporated under reduced pressure. This intermediate was
dissolved in 4 M hydrochloric acid in 1,4ꢀdioxane (3.20 mL; 12.80 mmol) and stirred for 2 days
at 20 °C. The solvent was removed under reduced pressure. Aqueous sodium hydrogen carbonate
solution and ethyl acetate were added to the residue. The organic layer was separated, dried and
the solvent was removed under reduced pressure. The crude product 1ꢀMethylꢀ5ꢀphenoxyꢀ1Hꢀ
pyrazolꢀ4ꢀylamine 5 (4 g, 21.0 mmol, 23%) was dissolved in aqueous hydrochloric acid (162
mL; 648 mmol) and cooled down to 0 °C. Afterwards sodium nitrite (1.9 g; 2.7 mmol) dissolved
in water was added and the mixture was stirred for 1 h at 0 °C. Then potassium iodide (10.5 g;
63 mmol) was added whilst vigorous stirring and within 30 min the mixture warmed to 20 °C. 2
ml of water were added to the reaction mixture and neutralized with aqueous sodium hydrogen
carbonate solution. Dichloromethane was added and the organic layer was separated, dried and
the solvent was removed under reduced pressure. The crude product was purified using silica gel
flash column chromatography with a solvent mixture of ethyl acetate and petroleum ether for
elution. The solvent of the corresponding fractions was evaporated under reduced pressure,
yielding 3.8 g (61%; 12.7 mmol) of the title 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
Purity by method A2: >95%; RT = 1.93 min, MS (ESI+) m/z 301 (M+H)+
ACS Paragon Plus Environment

Page 19 of 31
Journal of Medicinal Chemistry
1
2
1H NMR (400 MHz, CDCl3) δ (ppm) 7.46 (s, 1H), 7.34ꢀ7.30 (m, 2H), 7.10 (m, 1H), 6.89ꢀ6.87
(m, 2H), 3.69 (s, 3H)
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
(1-Methyl-5-phenoxy-1H-pyrazol-4-yl)boronic acid (6)
4ꢀIodoꢀ1ꢀmethylꢀ5ꢀphenoxyꢀpyrazole 6 (0.5 g; 1.67 mmol) was dissolved in THF and cooled
down at ꢀ 70 °C. Afterwards Triisopropyl borate (0.35 g; 1.83 mmol) and 2.5 M nꢀButyllithium
(1.06 ml; 2.67 mmol) were added to the solution and warmed overnight to 20 °C. The crude
product was purified using silica gel flash column chromatography with a solvent mixture of
ethyl acetate and petroleum ether for elution. The solvent of the corresponding fractions was
evaporated under reduced pressure, yielding 0.15 (41%, 0.69 mmol) of the title compound.
Purity by method A2>95%; RT = 2.84 min, MS (ESI+) m/z 219 (M+H)+
1H NMR (400 MHz, CDCl3) δ (ppm) 7.67 (s, 1H), 7.38ꢀ7.34 (m, 2H), 7.12 (m, 1H), 6.93ꢀ6.91
(m, 2H), 3.67 (s, 3H)
3-Methyl-6-(1-methyl-5-phenoxy-1H-pyrazol-4-yl)-[1,2,4]triazolo[4,3-b]pyridazine (7)
6ꢀChloroꢀ3ꢀmethyl[1,2,4]triazolo[4,3ꢀb]pyridazine (70.00 mg; 0.42 mmol), (1ꢀMethylꢀ5ꢀ
phenoxyꢀ1Hꢀpyrazolꢀ4ꢀyl)boronic acid (99.58 mg; 0.46 mmol) and sodium carbonate (88.02 mg;
0.83 mmol) were suspended in water (0.35 mL; 19.12 mmol) and N,NꢀDimethylformamid (4.00
mL; 51.98 mmol). The reaction mixture was flushed with nitrogen, afterwards 1,1'ꢀ
bis(diphenylphosphino)ferrocenepalladium(II) dichloride, dichloromethane (16.95 mg; 0.02
mmol) were added and the mixture were stirred in the microwave for 90 min at 120 °C. The
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 31
1
2
3
4
5
6
7
8
9
reaction mixture was diluted with water and ethyl acetate. The organic layer was separated,
washed with brine, dried and the solvent was removed under reduced pressure. The crude
product was purified using method P1_a, yielding 15 mg (12%, 0.049 mmol) of the title
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
Purity by method A3: > 95%, RT = 0.91 min; MS (ESI+) m/z 307 (M+H)+; HRMS (m/z): [M]+
calcd. for C16H14N6O, 306.1229; found, 306.1223
1H NMR (DMSOꢀd6, 500 MHz): δ = 8.33 (s, 1H), 8.24 (d, J = 9.8 Hz, 1H), 7.60 (d, J = 9.8 Hz,
1H), 7.37 (br t, J = 7.7 Hz, 2H), 7.12 (br t, J = 7.4 Hz, 1H), 7.00 (br d, J = 8.2 Hz, 2H), 3.73 (s,
3H), 2.28 ppm (s, 3H)
13C NMR (DMSOꢀd6, 126 MHz): δ = 156.3, 147.3, 146.7, 146.4, 142.9, 138.8, 130.6, 125.2,
124.0, 119.3, 115.5, 104.9, 35.1, 9.3 ppm
Table 1: Chromatography methods.
Column a
Solvent A
Solvent B
flow rat
[ml/min]
0.4
Gradient b
A1
A2
A3
Aquity UPLC BEH,
C18, 1.7 µm, 2,1x100
mm, 40°C
water with
0.1 % formic with 0.1%
acid
Acetonitrile
05 ꢀ> 100%
7 min
formic acid
Acetonitrile
with 0.02%
TFA
Venusil XBPꢀC18,
5µm, 2.1x50mm, 50°C 0.04% TFA
water with
1.0
1
10 ꢀ> 80%, 3
min
Waters XBridge, C18,
2.5 µm, 2.1x20mm,
60°C
water pH 9.2 Acetonitrile
(20mM
NH4HCO3)
?
P1_a Waters XBridge, C18,
10 µm, 30x100 mm
water pH 8
(buffer: NH3,
NH4HCO3)
water pH 8
Acetonitrile
Acetonitrile
100
100
05 ꢀ> 60%,
6 min
P1_b Waters XBridge, C18,
05 ꢀ> 95%,
ACS Paragon Plus Environment

Page 21 of 31
Journal of Medicinal Chemistry
1
2
3
4
5
10 µm, 50x100mm
(buffer: NH3,
NH4HCO3)
12 min
6
7
8
9
P2
Waters Sunfire, C18,
10 µm, 63*100 mm
water with
0.2 % formic
acid
Acetonitrile
100
02 ꢀ> 30 %,
6min
a
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
company, column name, kind of particle, particle size, column dimension, column
temperature;
b
% of solvent B at gradient start ꢀ> % of solvent B at gradient end, gradient time;
Protein Expression and Purification
Recombinant human Brd4ꢀBD1 (bromodomain 1 of Bromodomain containing protein 4 was
expressed in E. coli BL21(DE3) and contains an Nꢀterminal TEVꢀcleavable His6ꢀtag (The
15
expression plasmid was kindly provided by Boehringer Ingelheim). N labeled H6ꢀTEVꢀBrd4ꢀ
BD1 was expressed following the expression protocol for efficient isotopic labeling of
recombinant proteins using a fourfold cell concentration in isotopically labeled minimal
medium.22 After resuspension in minimal medium, cells were put for 5 min on ice and 30 min at
18 °C before induction with 0.4 mM IPTG (isopropylꢀβꢀDꢀ1ꢀthiogalactopyranoside).
Perdeuterated H6ꢀTEVꢀBrd4ꢀBD1 was expressed by transferring small volumes of E. coli
BL21(DE3) cells carrying the expression plasmid from M9 medium (no labeled material, 100 %
H2O) to M950%D2O medium (no labeled material, 50 % D2O, 50 % H2O), and finally to 800 mL
M9100%D2O (no labeled material, 100 % D2O). After reaching an OD600 = 0.7ꢀ0.9, cells were
centrifuged at 4000 g for 5 minutes at 20 °C, washed once with a small amount of M9100%D2O (no
labeled material, 100 % D2O) and again centrifuged at 4000 g for 5 minutes at 20 °C. The pellet
was resuspended in 200 mL M9100%D2O (containing 1 g/L NH4Cl and Dꢀglucoseꢀ1,2,3,4,5,6,6ꢀ
15
d7). Cells were put for 5 min on ice and 1 hour at 18 °C before induction with 0.4 mM IPTG
(isopropylꢀβꢀDꢀ1ꢀthiogalactopyranoside). Cells were collected after 18 hours of expression at
18°C and the pellet resuspended in 40ml of buffer containing 20mM sodium phosphate, 500mM
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 31
1
2
3
4
5
6
7
8
9
NaCl, 20mM imidazole (pH7.5), 5% glycerole and protease inhibitor (Roche cOmplete Mini,
EDTA free). Bacteria were lysed by sonication. Proteins were purified by Ni2+ affinity
chromatography (HiTrap Chelating HP, 5ml, GE Healthcare), incorporating a high salt wash
(buffer containing 1.5M NaCl), and cleaved with TEV protease overnight at 4°C (20mM sodium
phosphate, 500mM NaCl, 2.5% glycerol, 2mM βꢀmercaptoethanol). The cleaved protein was
again loaded onto a Ni2+ column to bind the cleaved His6ꢀtag and the His6ꢀtagged TEV protease
and the flowꢀthrough containing Brd4ꢀBD1 concentrated and loaded onto a gel filtration column
equilibrated in 10mM sodium phosphate, 100mM NaCl (pH7.5) (HiLoad 16/60 Superdex 75pg,
GE Healthcare). Brd4ꢀBD1 containing fractions were concentrated in an Amicon Ultraꢀ15
centrifugal filter device 3K NMWL and stored at ꢀ20°C.
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
NMR spectroscopy:
1D, STD and WaterLOGSY experiments were acquired at 277 K and 298 K on a Varian Inova
1
500 MHz spectrometer using a standard 5 mm Hꢀ13Cꢀ15N tripleꢀresonance probehead. NMR
spectra for the characterization of small molecules were collected at 298 K on a Bruker Avance
III HD 500 MHz spectrometer. The system was equipped with a 5 mm TCI cryoprobe.
1D: 1D spectra were recorded with a double WATERGATE23 suppression element.
Water-LOGSY: The WaterꢀLOGSY pulse sequence has been described before.4,24 Selective
water inversion was achieved with a double excitation sculpting element employing water
selective refocusing pulses (REꢀBURP25, inversionꢀbandwidth = 0.5 ppm). The mixing time was
set to 1 s, incorporating a hard 180˚ pulse in the middle of the mixing time and a water selective
flipꢀback pulse at the beginning of the double WATERGATE water suppression module. The
acquisition time and interscan delay were set to 1 s and 2 s, respectively. The number of scans
ACS Paragon Plus Environment

Page 23 of 31
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
was set to 2048. Sample concentrations were typically 1 mM ligand and 0ꢀ45 µM Brd4ꢀBD1 in a
buffer (10 mM sodium phosphate, 100 mM NaCl, pH7.5) containing 10% D2O. Intensity ratios
Ilogsy/I1D are derived from scaling of the 1D intensity to the corresponding LOGSY intensity with
the Bruker TopSpin 3.5pl5 software.
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
STD: STD experiments were performed in H2O. STDꢀNMR spectra were acquired as a
pseudoꢀ2D version, interleaving onꢀ and offꢀresonance saturation. Selective saturation was
achieved by a cascade of gaussian pulses (bandwidth = 0.5 ppm, pulseꢀlength = 7.5 ms, power =
15 dB). Onꢀresonance saturation was applied at 0 ppm and offꢀresonance saturation at ꢀ42 ppm.
A T1ρꢀfilter of 50 ms was applied to eliminate protein resonances from the spectrum. The
acquisition time and interscan delay were set to 0.5 s and 4 s, respectively. The number of scans
for either onꢀ or offꢀresonance saturation was set to 256. Sample concentrations were 1 mM
ligand and 45 µM Brd4ꢀBD1 in a buffer (10 mM sodium phosphate, 100 mM NaCl, pH7.5)
containing 10% D2O. STD amplification factors (STDꢀAF) were calculated as described before.1
(IrefꢀIsat)/Iref needed for STDꢀAF calculation was derived from scaling of the Iref intensity to the
corresponding STD intensity (Iref ꢀ Isat) with the Bruker TopSpin 3.5pl5 software. Initial growth
rate STD amplification factors (STDꢀAF0) were obtained by fitting STDꢀAFs measured at
different saturation times (tsat = 0.5, 1, 2, 3, 4 s) and calculated as described before.26
Small molecule NMR spectroscopy: Samples were dissolved in 600 µL DMSOꢀd6 and TMS
was added as an internal standard. 1D 1H spectra were acquired with 30° excitation pulses and an
13
interpulse delay of 4.2sec with 64k datapoints and 20ppm sweep width. 1D C spectra were
acquired with broadband composite pulse decoupling (WALTZ16) and an interpulse delay of
3.3sec with 64k datapoints and a sweep width of 240ppm. Spectra were processed with Topspin
3.1 software from Bruker and analyzed with ACDlabs NMR workbook 2016.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 31
1
2
3
4
5
6
7
8
9
Co-Crystallization, Processing and Refinement of BRD4-BD1 ligand structures
Recombinant protein of Brd4ꢀBD1 for crystallographic studies was produced as formerly
described.27 The protein was concentrated to a final concentration of 11 mg/ml in a buffer
containing 10 mM HEPES at pH 7.5 and 100mM sodium chloride.
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
Brd4ꢀBD1 protein was incubated with 2mM ligand and coꢀcrystals in complex with ligand 1
and ligand 2 were grown by mixing 1 µl protein solution with 1 µl reservoir solution (29 % PEG
3350 w/v, 200 mM diꢀsodium malonate and HEPES pH 7.2) using the hanging drop vapor
diffusion method at 293 K. Crystals appeared after 4 days. Before flash freezing in liquid
nitrogen the crystals were cryo protected by the addition of ethylene glycole in the crystallization
drop to a final concentration of 25 % (v/v).
Images were processed with autoPROC.28 The resolution limits were set using default
autoPROC settings. The structures were solved by molecular replacement using the Brd4ꢀBD1
structure 2OSS as a search model. Subsequent model building and refinement was done using
standard protocols using CCP4,29 COOT30 and autoBUSTER.31 Crystallographic data collection
statistics and refinement statistics can be found in the supporting information (Supplementary
Table S2 and S3).
Compound 1 (Space group=P 1, unit cell: a = 30.38 Å, b = 39.51 Å, c = 57.97 Å, α= 99,71°,
β= 105,05°, γ = 89,90°, resolution = 1.38 Å) was refined to R/Rfree = 22.2/24.2% with 100 % of
the residues in Ramachandran allowed regions.
ACS Paragon Plus Environment

Page 25 of 31
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Compound 2 (Space group=P 21 21 21, unit cell: a = 37.26 Å, b = 44.18 Å, c = 78.49 Å, α=
β= γ = 90°, resolution = 1.65 Å) was refined to R/Rfree = 18.7/21.4% with 100 % of the residues
in Ramachandran allowed regions.
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 following files are available free of charge.
Description of the LOGSYꢀtitration methodology and the extraction of LOGSY factors.
Simulation of LOGSY titration intensities. Comparison of LOGSYꢀtitrations with protonated and
deuterated protein. Table of compounds that show an affinity increase after introduction of a
secondary amine. Schemes for the synthesis of ligand 1 and ligand 2. Crystallographic data
collection and refinement statistics. Molecular Formula Strings for all ligands (PDF).
Accession Codes.
Atomic coordinates of Brd4ꢀBD1 bound to ligand 1 (5M39) and ligand 2 (5M3A) have been
deposited in the Protein Data Bank. Authors will release the atomic coordinates and experimental
data upon article publication.
AUTHOR INFORMATION
Corresponding Author
*Email: robert.konrat@univie.ac.at.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 31
1
2
3
4
Phone: ++43ꢀ(0)1ꢀ4277ꢀ52202
5
6
7
Author Contributions
8
9
L.G. performed the NMR experiments. L.G., M.M. and R.K. wrote the manuscript. M.M. and
H.E. supervised the medicinal chemistry team. H.E. designed the synthetic strategies. D.K. and
B.W. determined and analyzed the coꢀcrystal structures. X.ꢀL.C. performed the analysis on coꢀ
crystal structures. R.K. and D.B.M. supervised the study. All authors have given approval to the
final version of the manuscript.
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
Funding Sources
Funding was provided by Boehringer Ingelheim and the Christian Doppler Laboratory for Highꢀ
Content Structural Biology and Biotechnology, Austria.
ACKNOWLEDGMENT
L.G. is funded by the Christian Doppler Laboratory for HighꢀContent Structural Biology and
Biotechnology, Austria. The authors want to thank Dr. Andreas Zoephel for providing plasmids
and protocols for Brd4ꢀBD1 expression and purification, as well as Teresa Adacker for writing
the experimental section on the synthesis of the compounds.
ABBREVIATIONS
AFPꢀNOESY, adiabatic fast passage ꢀ nuclear Overhauser effect spectroscopy; BD,
bromodomain; BET, bromodomain and extraꢀterminal; Brd4ꢀBD1, bromodomain 1 of
bromodomain containing protein 4; DIRECTION, difference of inversion recovery rate with and
without target irradiation; GEM by STD, group epitope mapping; INPHARMA, interligand NOE
(ILOE) for pharmacophore mapping; SALMON, solvent accessibility, ligand binding, and
ACS Paragon Plus Environment

Page 27 of 31
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
mapping of ligand orientation by NMR; SAR by NMR, structureꢀactivity relationship by NMR;
STD, saturation transfer difference; STDꢀAF0, initial growth rate STD amplification factors;
transferred NOE, transferred nuclear Overhauser effect; WaterLOGSY, Water ligand observed
via gradient spectroscopy
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
REFERENCES
1. Mayer, M.; Meyer, B. Group epitope mapping by saturation transfer difference NMR to
identify segments of a ligand in direct contact with a protein receptor. J. Am. Chem. Soc. 2001,
123, 6108ꢀ6117.
2. Ludwig, C.; Michiels, P. J.; Lodi, A.; Ride, J.; Bunce, C.; Gunther, U. L. Evaluation of
solvent accessibility epitopes for different dehydrogenase inhibitors. ChemMedChem 2008, 3,
1371ꢀ1376.
3. Dalvit, C.; Fogliatto, G.; Stewart, A.; Veronesi, M.; Stockman, B. WaterLOGSY as a
method for primary NMR screening: practical aspects and range of applicability. J. Biomol. NMR
2001, 21, 349ꢀ359.
4. Dalvit, C.; Pevarello, P.; Tatò, M.; Veronesi, M.; Vulpetti, A.; Sundström, M.
Identification of compounds with binding affinity to proteins via magnetization transfer from
bulk water. J. Biomol. NMR 2000, 18, 65ꢀ68.
5. Dalvit, C.; Fagerness, P. E.; Hadden, D. T.; Sarver, R. W.; Stockman, B. J. Fluorineꢀ
NMR experiments for highꢀthroughput screening: theoretical aspects, practical considerations,
and range of applicability. J. Am. Chem. Soc. 2003, 125, 7696ꢀ7703.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 31
1
2
3
4
6. Shuker, S. B.; Hajduk, P. J.; Meadows, R. P.; Fesik, S. W. Discovering highꢀaffinity
5
6
ligands for proteins: SAR by NMR. Science 1996, 274, 1531ꢀ1534.
7
8
9
7. Auer, R.; Kloiber, K.; Vavrinska, A.; Geist, L.; Coudevylle, N.; Konrat, R.
Pharmacophore mapping via crossꢀrelaxation during adiabatic fast passage. J. Am. Chem. Soc.
2010, 132, 1480ꢀ1481.
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
8. Villicana, C.; Cruz, G.; Zurita, M. The basal transcription machinery as a target for
cancer therapy. Cancer Cell Int. 2014, 14, 18.
9. Ludwig, C.; Michiels, P. J.; Wu, X.; Kavanagh, K. L.; Pilka, E.; Jansson, A.; Oppermann,
U.; Gunther, U. L. SALMON: solvent accessibility, ligand binding, and mapping of ligand
orientation by NMR spectroscopy. J. Med. Chem. 2008, 51, 1ꢀ3.
10. Mizukoshi, Y.; Abe, A.; Takizawa, T.; Hanzawa, H.; Fukunishi, Y.; Shimada, I.;
Takahashi, H. An accurate pharmacophore mapping method by NMR spectroscopy. Angew.
Chem. Int. Ed. Engl. 2012, 51, 1362ꢀ1365.
11. Ni, F.; Zhu, Y. Accounting for ligandꢀprotein interactions in the relaxationꢀmatrix
analysis of transferred nuclear Overhauser effects. J. Magn. Reson. B. 1994, 103, 180ꢀ184.
12. Balaram, P.; BothnerꢀBy, A. A.; Dadok, J. Negative nuclear Overhauser effects as probes
of macromolecular structure. J. Am. Chem. Soc. 1972, 94, 4015ꢀ4017.
13. Zeng, L.; Zhou, M. M. Bromodomain: an acetylꢀlysine binding domain. FEBS Lett. 2002,
513, 124ꢀ128.
14. Wu, S. Y.; Chiang, C. M. The double bromodomainꢀcontaining chromatin adaptor Brd4
and transcriptional regulation. J. Biol. Chem. 2007, 282, 13141ꢀ13145.
ACS Paragon Plus Environment

Page 29 of 31
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
15. Delmore, J. E.; Issa, G. C.; Lemieux, M. E.; Rahl, P. B.; Shi, J.; Jacobs, H. M.; Kastritis,
E.; Gilpatrick, T.; Paranal, R. M.; Qi, J.; Chesi, M.; Schinzel, A. C.; McKeown, M. R.;
Heffernan, T. P.; Vakoc, C. R.; Bergsagel, P. L.; Ghobrial, I. M.; Richardson, P. G.; Young, R.
A.; Hahn, W. C.; Anderson, K. C.; Kung, A. L.; Bradner, J. E.; Mitsiades, C. S. BET
bromodomain inhibition as a therapeutic strategy to target cꢀMyc. Cell 2011, 146, 904ꢀ917.
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
16. Zuber, J.; Shi, J.; Wang, E.; Rappaport, A. R.; Herrmann, H.; Sison, E. A.; Magoon, D.;
Qi, J.; Blatt, K.; Wunderlich, M.; Taylor, M. J.; Johns, C.; Chicas, A.; Mulloy, J. C.; Kogan, S.
C.; Brown, P.; Valent, P.; Bradner, J. E.; Lowe, S. W.; Vakoc, C. R. RNAi screen identifies Brd4
as a therapeutic target in acute myeloid leukaemia. Nature 2011, 478, 524ꢀ528.
17. Lockwood, W. W.; Zejnullahu, K.; Bradner, J. E.; Varmus, H. Sensitivity of human lung
adenocarcinoma cell lines to targeted inhibition of BET epigenetic signaling proteins. Proc. Natl.
Acad. Sci. U. S. A. 2012, 109, 19408ꢀ19413.
18. Halle, B. Protein hydration dynamics in solution: a critical survey. Philos. Trans. R. Soc.
Lond. Ser. B 2004, 359, 1207ꢀ1223; discussion 1223ꢀ1224, 1323ꢀ1328.
19. Ernst, J. A.; Clubb, R. T.; Zhou, H. X.; Gronenborn, A. M.; Clore, G. M. Demonstration
of positionally disordered water within a protein hydrophobic cavity by NMR. Science 1995,
267, 1813ꢀ1817.
20. Otting, G. NMR studies of water bound to biological molecules. Prog. Nucl. Magn.
Reson. Spectrosc. 1997, 31, 259ꢀ285.
21. Grzesiek, S.; Bax, A.; Nicholson, L. K.; Yamazaki, T.; Wingfield, P.; Stahl, S. J.;
Eyermann, C. J.; Torchia, D. A.; Hodge, C. N. NMR evidence for the displacement of a
ACS Paragon Plus Environment