Identification of Ligand Binding Hot Spots of the Histamine H1 Receptor following Structure-Based Fragment Optimization

Article abstract of DOI:10.1021/acs.jmedchem.6b00981

Developments in G protein-coupled receptor (GPCR) structural biology provide insights into GPCR-ligand binding. Compound 1 (4-(2-benzylphenoxy)piperidine) with high ligand efficiency for the histamine H₁ receptor (H₁R) was used to design derivatives to investigate the roles of (i) the amine-binding region, (ii) the upper and lower aromatic region, and (iii) binding site solvation. SAR analysis showed that the amine-binding region serves as the primary binding hot spot, preferably binding small tertiary amines. In silico prediction of water network energetics and mutagenesis studies indicated that the displacement of a water molecule from the amine-binding region is most likely responsible for the increased affinity of the N-methylated analog of 1. Deconstruction of 1 showed that the lower aromatic region serves as a secondary binding hot spot. This study demonstrates that an X-ray structure in combination with tool compounds, assessment of water energetics, and mutagenesis studies enables SAR exploration to map GPCR-ligand binding hot spots.

Full text of DOI:10.1021/acs.jmedchem.6b00981

Subscriber access provided by CORNELL UNIVERSITY LIBRARY
Article
1
Identification of Ligand Binding Hot Spots of the Histamine H
Receptor Following Structure-Based Fragment Optimization
Sebastiaan Kuhne, Albert J. Kooistra, Reggie Bosma, Andrea Bortolato, Maikel Wijtmans,
Henry F. Vischer, Jonathan S. Mason, Chris de Graaf, Iwan J. P. De Esch, and Rob Leurs
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b00981 • Publication Date (Web): 19 Sep 2016
Downloaded from http://pubs.acs.org on September 21, 2016
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
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 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Identification of Ligand Binding Hot Spots of the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Histamine H Receptor Following StructureꢀBased
1
Fragment Optimization
†
†
†
‡
†
Sebastiaan Kuhne , Albert J. Kooistra , Reggie Bosma , Andrea Bortolato , Maikel Wijtmans ,
†
‡
†
†
†
Henry F. Vischer , Jonathan S. Mason , Chris de Graaf , Iwan J.P. de Esch , Rob Leurs *
†
Amsterdam Institute for Molecules, Medicines and Systems (AIMMS), Division of Medicinal
Chemistry, Faculty of Science, Vrije Universiteit Amsterdam, De Boelelaan 1108, 1081 HZ
Amsterdam, The Netherlands.
‡
Heptares Therapeutics Ltd., BioPark, Broadwater Road, Welwyn Garden City, Herts, AL7 3AX,
U.K.
KEYWORDS
Histamine H receptor (H R), structureꢀbased drug design, binding site mapping, computerꢀaided
1
1
drug design, siteꢀdirected mutagenesis
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ABSTRACT
Developments in G proteinꢀcoupled receptor (GPCR) structural biology provide insights into
GPCRꢀligand binding. Compound 1 (4ꢀ(2ꢀbenzylphenoxy)piperidine) with high ligand efficiency
for the histamine H receptor (H R) was used to design derivatives to investigate the roles of: i)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
1
the amineꢀbinding region, ii) the upper and lower aromatic region and iii) binding site solvation.
SAR analysis showed that the amineꢀbinding region serves as the primary binding hot spot,
preferably binding small tertiary amines. In silico prediction of water network energetics and
mutagenesis studies indicated that the displacement of a water molecule from the amineꢀbinding
region is most likely responsible for the increased affinity of the Nꢀmethylated analog of 1.
Deconstruction of 1 showed that the lower aromatic region serves as a secondary binding hot
spot. This study demonstrates that an Xꢀray structure in combination with tool compounds,
assessment of water energetics, and mutagenesis studies enables SAR exploration to map GPCRꢀ
ligand binding hot spots.
ACS Paragon Plus Environment

Page 3 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
INTRODUCTION
In the past eight years tremendous progress has been made in the crystallization of GPCRs,
leading to more than 150 structures of 33 receptors representing different GPCR families and
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
2
3
subfamilies , including aminergic GPCRs such as the histamine H receptor (H R). With the
1
1
2
, 4
emerging GPCR crystal structures a new opportunity has emerged to link ligandꢀbased data
with (protein) structureꢀbased information. Amongst all available Class A GPCR crystal
structures, the H R has the deepest ligandꢀbinding pocket, with doxepin as the coꢀcrystallized
1
ligand (Figure 1A–C). The H R antagonist binding site is divided into three regions: i) the
1
amineꢀbinding region, ii) the upper aromatic region and iii) the lower aromatic region (Figure
1
C), and the importance of the residues lining these regions for ligand binding is supported by
5
ꢀ10
3.32
several siteꢀdirected mutagenesis studies.
The amineꢀbinding region is defined by D107
,
6.48
6.51
7.39
7.43
W428 , Y431 , I454 , and Y458 (Figure 1A, C). The amine moiety of doxepin interacts
3.32
4, 11, 12
with the hallmark D107 that is present in all aminergic GPCRs.
In addition, the H R has
1
a cluster of aromatic residues in TM3, 4, 5 and 6 that accommodates the aromatic ring systems of
the H R antagonists (Figure 1A, C) and can be divided into the upper aromatic region (defined
1
3
.33
4.56
6.51
6.52
6.55
by Y108 , W158 , Y431 , F432
and F435 , Figure 1A, C) and the lower aromatic
5
.47
6.44
6.48
region (defined by F199 , F424 and W428 , positioned deep in the transmembrane helical
bundle, Figure 1A, C). Together, both the upper and lower aromatic region accommodate the
butterfly shape of the hydrophobic aromatic rings of the coꢀcrystallized ligand doxepin (Figure
1
A, C). It has been proposed that desolvation of these hydrophobic aromatic regions is an
1
3, 14
important determinant of GPCR ligand binding.
Previously, we used the H R Xꢀray structure for a highly successful virtual screening (73 % hit
1
rate) for new fragmentꢀlike H R antagonists, resulting in the identification of VUF14544 (4ꢀ(2ꢀ
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
15, 16
benzylphenoxy)piperidine, 1) with an H R affinity of 6 nM.
Docking studies with 1
1
indicated that the fragmentꢀlike molecule can be accommodated in the doxepin binding site
Figure 1A, C) and offers a good starting point to probe the various main binding pockets in the
(
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
orthosteric H R binding site. Ligand 1 benefits from a high ligand efficiency (LE, defined as the
1
binding energy divided by the number of heavy atoms of a ligand) and does not contain stereoꢀ
or regioisomers. This ligand is therefore used in the current study as a scaffold to design and
synthesize a set of 23 derivatives to systematically investigate the role of: 1) the amineꢀbinding
region, 2) the upper and lower aromatic binding region and 3) the impact of (de)solvation of the
14
binding site on the binding of H R antagonists. WaterFLAP analysis of H R binding site
1
1
solvation was for the first time applied to guide siteꢀspecific protein mutagenesis studies in order
to complement the structureꢀbased SAR analysis.
ACS Paragon Plus Environment

Page 5 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. (A) The H R crystal structure (PDBꢀcode: 3RZE) with the coꢀcrystallized ligand
1
3
doxepin (green carbon atoms) and the proposed binding mode of the structureꢀbased virtual
15
screening hit 1 (magenta carbon atoms). The yellow ribbons represent parts of the backbone of
transmembrane (TM) helices 2, 3, 5, 6 and 7. Selected binding site residues are depicted as ballꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
andꢀsticks with light gray carbon atoms. Oxygen, nitrogen, and hydrogen atoms are colored red,
blue, and cyan, respectively. Polar hydrogen atoms of the ligands are shown, but are omitted for
3
.33
the binding site residues. For clarity residue Y108 is omitted. Throughout this manuscript the
Uniprot numbers (first) and the BallesterosꢀWeinstein numbering (second in superscript) are
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
7
reported for each residue. (B) Structure of doxepin and H R affinity as determined in this
1
study. Ligand efficiency (LE) = (ꢁG)/N, where ꢁG = −RT ln(K ) and N is the number of nonꢀ
i
ꢀ1
ꢀ1
hydrogen atoms (HA), and R = (8.31447215 J K mol and 4184 J = 1 kcal), and T = 298.15 K.
(C) The H R binding site surface is shown and colored to indicate the different regions of the
1
binding site, i.e. the amineꢀbinding region and the upper and lower aromatic binding region. (D)
Schematic 2D representation of the structureꢀbased virtual screening hit 1 interacting with the
three examined regions.
ACS Paragon Plus Environment

Page 7 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
RESULTS
Probing the H R binding site. To explore the three different H R binding pockets (Figure 1C)
1
1
we designed and synthesized a variety of analogs of 1. By making subtle chemical changes
Figure 2) we probed each of the three binding pockets in order to define H R binding hot spots
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
(Figure 3).
Figure 2. Scaffold exploration. The different molecules designed to revisit the molecular
determinants of H R binding. The color of the molecules corresponds to the color coding in
1
Figure 6.
Targeting the amine-binding region. Many high affinity H R reference compounds contain a
1
tertiary amine, e.g. doxepin (Figure 1B). To investigate the effect of Nꢀmethyl substitution on
H R binding affinity, secondary amine 1, tertiary amine 2, and quaternary ammonium cation 3
1
were designed and synthesized (Figure 2). Opening of the piperidine ring of 1 allowed the
comparison of a primary (4a, 5a), secondary (4b, 5b) and tertiary (4c, 5c) amine. In addition, the
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
steric and hydrophobic fit of the tertiary amine was investigated by amines incorporated in a
series of aliphatic rings (4d–g and 5d–g), also seen in H R reference compounds such as
1
1
0
triprolidine. In addition, the linear analogs (4a–g and 5a–g) bearing a two or three methylene
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
spacer between the phenylether and the basic amine, allowed for the investigation of the
3
.32
sensitivity of the distance between the basic amine interacting with D107
pharmacophoric features.
and the aromatic
The compounds probing the amine binding region were synthesized as outlined in Scheme 1.
Commercially available phenol 13 was used in a Mitsunobu reaction to obtain both the Bocꢀ
protected ligand 14 and amine 2 in moderate yield. Compound 14 was deprotected using TFA to
afford ligand 1. Methylation of amine 2 with methyl iodide gave quaternary ammonium cation 3.
Alkylation of phenol 13 with 1ꢀbromoꢀ2ꢀchloroethane and 1ꢀbromoꢀ3ꢀchloropropane afforded 15
and 16, respectively. Subsequent nucleophilic aliphatic substitution using microwave assisted
heating afforded 4a–g and 5a,b,d–g in moderate to good yield. Compound 5c was obtained by
reacting 16 with dimethylamine at room temperature.
ACS Paragon Plus Environment

Page 9 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Scheme 1. Synthesis of ligands to target the amineꢀbinding region.
a
c
2
O
OH
O
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
3
I
N
N
R1
d
1
2
1
R = Me
3
1
4 R = Boc
b
1
1 R = H
e
O
O
n
R2
n
Cl
15 n = 2
16 n = 3
4a-g (n = 2)
5a-g (n = 3)
2
R =
NH2
NH
N
N
N
N
N
a
b
c
d
e
f
g
a
Reagents and conditions: (a) 14: tertꢀbutyl 4ꢀhydroxypiperidineꢀ1ꢀcarboxylate, PPh , DEAD
40 wt% in toluene), THF, 0 °C to rt, 5 h. 2: 1ꢀmethylpiperidinꢀ4ꢀol, PPh , DEAD (40 wt% in
3
(
3
toluene), THF, 0 °C to rt, 4 h. (b) TFA, DCM, 0 °C to rt, 4 h. (c) MeI, DCM, rt, 5.5 h. (d) 15:
Br(CH ) Cl , K CO , (CH ) CO, reflux, 16 h. 16: Br(CH ) Cl , K CO , (CH ) CO, reflux, 25 h.
2
2
2
3
3 2
2 3
2
3
3 2
(
e) 4a: NH (7 M in MeOH), MW (130 °C, 6 h). 5a: NH (7 M in MeOH), MW (125 °C, 2 h). 4b
and 5b: CH NH (33 wt% in EtOH), MW (130 °C, 35 min). 4c: (CH ) NH (5.6 M in EtOH),
MW (150 °C, 25 min). 5c: (CH ) NH (5.6 M in EtOH), rt, 11 days. 4d and 5d: azetidine,
3
3
3
2
3 2
3
2
NaHCO , DMF, MW (120 °C, 30 min and 140 °C, 30 min). 4e–g and 5e–g: amine, NaHCO ,
3
3
NaI, MeCN, MW (120–160 °C, 20–50 min). Compound 1, 2, 4a–d and 5a–d were obtained as
hemi)fumaric acid salts.
(
Targeting the upper and lower aromatic region. The relative importance of the two aromatic
rings in ligand 1 was investigated by: i) ligand deconstruction (6 and 7), ii) different vectors (8, 9
and 10) and iii) chloro substitution (11 and 12). The compounds probing the aromatic region
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
were synthesized as outlined in Scheme 2. Amine 7 was obtained via a reductive amination with
secondary ammonium salt 6 and formaldehyde in the presence of acetic acid and sodium
triacetoxyborohydride as a reducing agent. Due to the moderate yields and tedious purification in
the Mitsunobu reactions described above, a second strategy using mesylate 18, which was
synthesized by reacting methanesulfonyl chloride with alcohol 17 in the presence of
triethylamine, was used to obtain compound 19. Significant amounts of elimination product were
formed during this reaction, but after treatment with TFA and column purification, amine 8 was
obtained in sufficient overall yield and good purity. The chloroꢀsubstituted analogs 11 and 12
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
18
were synthesized using phenol 24, which was synthesized according to a patent procedure.
First, 2ꢀbromophenol (20) was protected using benzylbromide to provide benzylether 21. Bromoꢀ
lithium exchange followed by the addition of 4ꢀchlorobenzaldehyde gave alcohol 22. Reduction
of alcohol 22 using BF .Et O and triethylsilane afforded benzylether 23, which was subsequently
3
2
deprotected using BF .Et O and dimethylsulfide to obtain phenol 24 in good overall yield. The
3
2
same synthetic strategies used for 19, 8, 7 were applied to obtain the compounds 25, 11, 12,
respectively.
ACS Paragon Plus Environment

Page 11 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Scheme 2. Synthesis of ligands to target the aromatic region.
O
S
OH
O
O
b
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
O
O
N
N
N
Boc
17
Boc
N
H
18
HCl
7
6
c
OH
O
d
1
3
N
R1
1
1
9 R = Boc
1
8 R = H
OH
O
Br
Br
O
e
f
g
h
Cl
O
Cl
OH
2
0
21
22
23
i
j
OH
Cl
O
Cl
O
Cl
2
4
N
N
R2
2
Boc
2
5
1
1
1 R = H
k
2
2 R = Me
a
Reagents and conditions: (a) CH O (37 wt% in water), AcOH, Na(CH COO) BH,
DCM:MeOH 2:1, rt, 6 h. (b) CH SO Cl, TEA, DCM, 0 °C to rt. (c) 18, Cs CO , DMF, 65 °C, 26
h. (d) TFA, DCM, 0 °C to rt, 16 h. (e) 1. K CO , (CH ) O, rt, 1 h, 2. PhCH Br, reflux, 2 h. (f) 1.
nꢀBuLi, −78 °C, 30 min, 2. 4ꢀClꢀPhCHO, THF, −78 °C to 0 °C, 2 h. (g) Et SiH, BF .Et O,
2
3
3
3
2
2
3
2
3
3 2
2
3
3
2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
MeCN, −40 °C to rt, 3 h. (h) Me S, BF .Et O, DCM, 0 °C to rt, 28 h. (i) 18, Cs CO , DMF, 65
2
3
2
2
3
°
C, 26 h. (j) TFA, DCM, 0 °C to rt, 15 h. (k) CH O (37 wt% in water), AcOH,
2
Na(CH COO) BH, DCM:MeOH 2:1, rt, 2 h. Compound 7, 8, 11 and 12 were obtained as
3
3
fumaric acid salts.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Pharmacological evaluation. Binding affinities (K values) for the human H R proteins (wild
i
1
type and H R mutants) of test compounds described in Table 1 and 2 were evaluated using
1
3
19
radioligand displacement experiments, with [ H]ꢀmepyramine as a radioligand. Receptor
2
0
proteins were transiently expressed in HEK293T cells as described previously.
Structure-activity relationship for human H R binding. All ligands are evaluated based on
1
2
1
22
their LE and group efficiency (GE) . GE is used to dissect the contributions of different parts
22
23
of the ligand to its binding affinity , using e.g. a ligand deconstruction approach. The GE
values were furthermore plotted on the protein surface in order to visualize binding hot spots
(Figure 3). The SAR of the compounds targeting the amineꢀbinding region and the aromatic
regions are discussed below in detail.
Efficiency of ligand binding in the amine-binding region. An 8ꢀfold increase in affinity upon
methylation of secondary amine 1 to obtain tertiary amine 2 was observed (Table 1). Subsequent
methylation to obtain quaternary ammonium salt 3 shows a 12ꢀfold decrease in affinity. Table 1
shows that each subsequent methyl substitution in both linear analog series (4a–c and 5a–c) lead
to a roughly 10ꢀfold increase in binding affinity and an increase of 0.1 in LE in total. Thus, going
from a primary amine to a tertiary amine a 100ꢀfold increase in H R affinity is obtained. The
1
affinities for the ligands with a three methylene spacer (5a–c) are 3 to 5ꢀfold higher compared to
the two methylene spacer series (4a–c). Interestingly, the affinities of ligand 2 and its flexible
analog 5c are similar. Further extending the side chain by the introduction of aliphatic rings (4d–
g, 5d–g) on the basic amine, and hence increasing the lipophilicity of the ligand, generally shows
ACS Paragon Plus Environment

Page 13 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a decrease in affinity and therefore a drop in LE and GE. This clearly shows that small tertiary
amines (2, 4c and 5c) are optimal.
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
Table 1. Pharmacological evaluation of ligands targeting the amineꢀbinding region
a
pK H R ± SEM
i
1
b
c
#
n
R
LE
GE
WT
Mepyramine
8.9 ± 0.1
7.8 ± 0.1
0.58
0.53
ꢀ
ꢀ
1
8
.7 ± 0.1
0.57
1.23
2
O
7
.6 ± 0.1
0.47
0.47
−0.14
I
N
3
4a
2
5.8 ± 0.0
ꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
4b
2
2
7.0 ± 0.0
7.9 ± 0.1
0.53
0.57
1.64
1.43
4
c
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4d
2
2
8.2 ± 0.0
7.8 ± 0.1
0.56
0.51
1.09
0.68
4
e
4f
2
2
7.6 ± 0.0
7.2 ± 0.1
0.47
0.43
0.49
0.32
N
4
g
5
5
a
3
3
6.5 ± 0.0
7.5 ± 0.1
0.49
0.54
ꢀ
b
1.36
5c
3
3
8.5 ± 0.0
8.1 ± 0.1
0.58
0.53
1.36
0.73
5
d
5
e
3
3
8.0 ± 0.1
7.9 ± 0.1
0.50
0.47
0.51
0.38
5f
N
5
g
3
7.7 ± 0.1
0.44
0.27
ACS Paragon Plus Environment

Page 15 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Affinity values (pK ) were determined using a competition binding experiment for the H R
i
1
3
between the respective ligand and [ H]ꢀmepyramine. Values represent the mean ± SEM of N ≥ 3
b
c
experiments. LE calculated as described in Figure 1. Group efficiency = −ꢁꢁG/ꢁN, where
ꢁꢁG = ꢁG (molecule B) − ꢁG (molecule A) and ꢁN = HA (molecule B) – HA (molecule A).
Compound 1 was used as molecule A to calculate the GE for compound 2 and 3. Compound 4a
was used to calculate the GEs for compound 4b–g. Compound 5a was used to calculate the GEs
for compound 5b–g.
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
Efficiency of ligand binding in the aromatic regions. Deconstruction of ligands 1 and 2 lead to
compound 6 and 7, respectively. Both ligands lacking the benzyl moiety show a ~100ꢀfold
decrease in affinity, but the LE is increased due to the reduced heavy atom count (Table 2).
Ligand 8 and 9, with the benzyl group on the meta and para position, show a 100 and 500ꢀfold
decrease in affinity, respectively compared to 1. Direct attachment of the phenyl ring at the ortho
position (10) shows only a 10ꢀfold decrease in affinity compared to 1. Chloroꢀsubstitution on the
benzyl ring of secondary amine 1 to afford 11 shows a slight increase in affinity, while chloroꢀ
substitution of tertiary amine 2 shows comparable affinity.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 2. Pharmacological evaluation of ligands targeting the aromatic regions
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
pK H R ± SEM
i
1
1
2
b
c
#
R
R
LE
GE
WT
mepyramine
8.9 ± 0.1
7.8 ± 0.1
0.58
0.53
ꢀ
1
2
H
0.47
0.41
Me
8.7 ± 0.1
0.57
6
7
H
H
H
5.4 ± 0.2
6.6 ± 0.1
0.57
0.64
ꢀ
Me
1.64
8
9
H
H
5.8 ± 0.1
5.0 ± 0.1
0.40
0.34
0.08
−0.08
1
0
1
2
H
H
6.5 ± 0.2
8.2 ± 0.2
8.6 ± 0.1
0.47
0.53
0.53
0.25
0.55
1
1
Me
−0.14
a
Affinity values (pK ) were determined using a competition binding experiment for the H R
i
1
3
between the respective ligand and [ H]ꢀmepyramine. Values represent the mean ± SEM of N ≥ 3
experiments. LE calculated as described in Figure 1. GE calculated as described in Table 1.
b
c
Compound 6 was used as molecule A to calculate the GE for compound 1 and 7–10. Compound
ACS Paragon Plus Environment

Page 17 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
7
was used as molecule A to calculate the GE for compound 2. Compound 1 was used as
molecule A to calculate the GE for compound 11. Compound 2 was used as molecule A to
calculate the GE for compound 12.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
H R binding affinity hot spot analysis. All ligands described in Table 1 and 2 were docked
1
2
4, 25
3
with PLANTS
in the H R crystal structure in order to project the GEs of the ligands on the
1
protein surface to visualize the binding hot spots (Figure 3). The most energetically favorable
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3.32
pose that forms an ionic and Hꢀbond interaction with D107 was selected (i.e. the pose with the
2
4
lowest PLANTS ChemPLP score , Supplementary Table S1) and subsequently, the GEs of all
ligands were projected on the protein surface resulting in different hot spots (Figure 3). The most
prominent hot spot is in the amineꢀbinding region, followed by the lower aromatic region and the
upper aromatic region. All tested ligands described in Table 1 and 2 have LEs in the range of
0
.34–0.64, and GEs in the range of −0.14–1.64. All ligands, except 9, have a similar or higher
2
6
LE compared to the mean LE (0.4) reported for ligands targeting aminergic GPCRs. The
changes in substituents for all compounds, except 3,9 and 12, have a positive GE value (Table 1
and 2). Positive GE values indicate a beneficial contribution of a substituent to the binding
affinity, i.e. the higher the GE value the more efficient probing of a certain binding region by that
2
2, 27, 28
particular substituent. Taking GE values for several case studies
into account we consider
our values to be satisfactory. The concentrated primary binding hot spot in the amine binding
region is a result of the small amines 2, 4b,c and 5b,c, having the highest GEs, ranging from
1
.36–1.64. The secondary, but weaker, binding hot spot in the lower aromatic region is the result
of the (substituted) benzylꢀsubstituents of the ligands 1, 2, and 11 probing this pocket with GEs
in the range of 0.41–0.55. The upper aromatic region is the weakest binding hot spot (GEs is
−
0.08) due to the fact that the (substituted) benzylꢀsubstituents are probing the lower aromatic
region instead (Supplementary Figure S1A).
ACS Paragon Plus Environment

Page 19 of 70
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
3
Figure 3. Binding hot spots. The H R crystal structure (PDBꢀcode: 3RZE) with the predicted
1
binding pose of tertiary amine 2 and the binding hot spot map with two views. The color coding
of the surface reflects the GE of all ligands described in Table 1 and Table 2 and is projected on
the protein atoms within 4 Å distance. The binding pocket surface was colored according to the
maximum GE values for each protein atom. Selected binding site residues are depicted as ballꢀ
andꢀsticks with light gray carbon atoms. Oxygen, nitrogen, and hydrogen atoms are colored red,
blue, and cyan, respectively. Polar hydrogen atoms of the ligand are shown, but are omitted for
the binding site residues.
WaterFLAP analysis. To explain the observed SAR, we investigated the role of water
14
molecules in the binding of H R ligands using the WaterFLAP method. This method can be
1
used to build water networks in binding sites, to predict the relative free energy of water
molecules, and to analyze the perturbation of an explicit water network as a consequence of
ligand binding. The WaterFLAP method combines a short molecular dynamic simulation to
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
assess the degrees of freedom of the water molecules (entropy) with a molecular interaction field
(GRID) analysis of the hydrophobic and apolar interaction field around the water molecules in
1
4, 29, 30
order to identify energetically favorable (‘happy’) and unfavorable (‘unhappy’) waters.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Previous WaterFLAP analyses of the H R crystal structure without doxepin in the binding site
1
(apo) predicted that there are ‘unhappy’ water molecules located in the lower aromatic region
1
3
and near the amine binding region (Figure 4A). Complementary statistical analysis of hydration
1
4, 31, 32
sites observed in molecular dynamics simulations using WaterMap
identified unhappy
water molecules in the same binding site regions (Supplementary Figure S2). WaterFLAP
analyses of ligand bound wildꢀtype H R indicated that the ‘unhappy’ water molecules in the
1
lower aromatic region are displaced by doxepin (Figure 4B), ligand 1 (Figure 4C), and ligand 2
(Figure 4D). The increase in affinity for ligand 2 compared to ligand 1 results in a binding
3.32
efficiency hot spot near D107 (Figure 3), but there are no specific interactions made by the
additional methyl group. However, an ‘unhappy’ water molecule is trapped between the basic
amine of ligand 1 (Figure 4C) and the protein and hence has limited degrees of freedom. The
‘unhappy’ water molecule is displaced upon the introduction of the extra methyl group (i.e.
ligand 2, Figure 4D) and this observation could explain the increased affinity for ligand 2. The
6
.51
7.39
7.43
residues F431 , I454 , and Y458
form a narrow and hydrophobic subpocket around the
6.51
7.43
‘unhappy’ water molecule. While F/Y
and Y
are highly conserved amongst aminergic
GPCRs, residue 7.39 is variable in terms of polarity and size. To investigate H R specific
1
determinants of ligandꢀdependent binding site desolvation we therefore focused on the role of the
7
.39
H R specific residue I454
1
in trapping the identified ‘unhappy’ water molecules in the amine
binding region. WaterFLAP analysis was for the first time applied to guide prospective protein
7.39
mutagenesis studies by the systematic in silico substitution of I454
to all natural occurring
ACS Paragon Plus Environment

Page 21 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
amino acids and the computational assessment of the water network energetics of the binding
sites of the resulting H R mutants. Based on these WaterFLAP calculations an isoleucine
1
(present in the wild type receptor) and phenylalanine are optimal for a maximal increase in
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
7.39
7.39
affinity upon Nꢀmethylation. Both the wild type H R (I454 ) and the H R I454 F mutant
1
1
receptor show ‘unhappy’ waters in the immediate vicinity of the basic amine that will be
displaced upon methylation of ligand 1 (Figure 4C, D, F). For the threonine mutant H R
1
7
.39
I454 T a reduced hydrophobic interaction field as well as an absence of ‘unhappy’ waters was
observed in the vicinity of the amine. Consequently, this WaterFLAP analysis suggested that a
7.39
I454 T mutation would reduce the effect of Nꢀmethylation (Figure 4E).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
7.39
Figure 4. WaterFLAP calculations of the binding site of WT H R (I454 ), (A) without ligand,
1
7
.39
in complex with (B) doxepin, (C) ligand 1, and (D) ligand 2, (E) H R I454 T mutant in
1
ACS Paragon Plus Environment

Page 23 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
7.39
complex with ligand 1, and (F) H R I454 F mutant in complex with ligand 1. Waters are color
1
coded to show the most ‘unhappy’ as red (>3.5 kcal/mol), then yellow (2.2–3.5 kcal/mol), bulk
solvent as gray (−1 to 2.2 kcal/mol) and ‘happy’ as blue (<−1 kcal/mol). ‘Unhappy’ waters are
energetically unfavorable and prefer to be in the bulk solvent. GRID maps are contoured
(transparent solid) and colored in the following manner: CH3 methyl group probe in gray at 1
kcal/mol defining the pocket surface and C1= probe (lipophilic) in yellow at −2.7 kcal/mol. The
amine and aromatic subpockets are indicated in panel 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
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Site-directed mutagenesis. To test the WaterFLAP hypothesis, the binding affinities (K ) of
i
7
.39
compound 1 and 2 were evaluated on the wild type H R (WT) and the mutants H R I454 T and
1
1
7
.39
3
3
H R I454 F using radioligand displacement of [ H]ꢀmepyramine. The radioligand [ H]ꢀ
1
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
mepyramine binds all three receptors with high, nanomolar affinity and only shows a slight
7
.39
7.39
decrease in affinity for the I454 T and I454 F mutants (pK 8.4 ± 0.1 and 8.5 ± 0.1,
D
respectively) compared to WT (pK 8.6 ± 0.1; Supplementary Table S2). The expression levels
D
7
.39
7.39
of I454 T and I454 F were high, but compared to wildꢀtype this was decreased 2ꢀ and 10ꢀ
fold, respectively (Supplementary Table S2). The total amount of membranes was therefore
3
adjusted to bind similar levels of [ H]ꢀmepyramine (Supplementary Figure S3). As can be seen
7
.39
in Figure 5A and 5C, at both the WT H R and the H R I454 F, methylation of secondary
1
1
amine 1 to give tertiary amine 2 results in an almost 10ꢀfold increase in affinity. Moreover, in
agreement with the WaterFLAP analysis the difference in affinity between the ligand 1 and 2 is
7
.39
lost upon the I454 T mutation (Figure 5B).
ACS Paragon Plus Environment

Page 25 of 70
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
3
WT
7.39
Figure 5. Displacement of [ H]ꢀmepyramine (mep) from the (A) H R , (B) H R I454 T and
1
1
7.39
(
C) H R I454 F by the unlabeled ligands 1 and 2. HEK293T cells expressing the Nꢀterminal
1
HAꢀtagged H R or mutant receptors were homogenized. Cell homogenate was then incubated for
1
3
4
h with [ H]ꢀmepyramine and increasing concentrations of either 1 or 2 in 50 mM
Na HPO /KH PO pH 7.4 at 25 °C. The binding reaction was terminated and bound radioligand
2
4
2
4
quantified by liquid scintillation counting. Representative graphs are shown of N≥3 with mean
and SEM of triplicate values. All binding affinity values ± SEM can be found in Supplementary
Table S3.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
DISCUSSION
A diverse set of 23 analogs based on the fragmentꢀlike H R antagonist 1 has been used to
1
further unravel the molecular determinants of H R binding using receptorꢀligand crystal structure
1
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
information, ligand SAR exploration, computational assessment of water network energetics
with WaterFLAP and prospective mutagenesis studies.
Figure 6. SAR exploration. LEs vs number of heavy atoms for the compounds described in
Figure 2. The pharmacological data can be found in Table 1 and 2.
Probing the amine-binding region. SAR (Figure 6) and H R binding site analysis (Figure 3)
1
demonstrates the preference for small tertiary basic amines in the amine binding region, resulting
3
.32
6.48
in a primary binding hot spot around the conserved D107 and its microenvironment W428
,
6
.51
7.39
7.43
Y431 , I454
and Y458 . The increase in H R affinity upon methylation going from a
1
primary (4a, 5a) to a secondary (1, 4b, 5b) to a tertiary (2, 4c, 5c) amine (Figure 6) is in line with
33
the results reported by Shah et al. The effect of simple methylation of a ligand on
affinity/potency has been investigated in several studies. However, these studies mainly focus on
ACS Paragon Plus Environment

Page 27 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
34ꢀ36
CꢀH methylation.
There are several possible explanations for the increase in H R affinity
1
upon methylation: i) placement of the methyl in a hydrophobic pocket; ii) a conformational
change of the ligand induced by the methyl group; iii) a decrease in the free energy of
desolvation upon methylation (i.e. water molecules bound to the ligand need to be stripped when
going from an aqueous surrounding and entering a hydrophobic pocket); iv) displacement of
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
4, 34ꢀ36
‘
unhappy’ waters in the protein binding site, or combinations thereof.
The quaternary
ammonium salt 3 has a decreased affinity compared to the tertiary amine 2 (Figure 6), which can
3
.32
possibly be attributed to the fact that it is not able to form an Hꢀbond with D107 but only an
ionic interaction. Furthermore, the additional hydrophobic methyl group in 3 may cause a nonꢀ
3
.32
optimal steric and electronic fit when it is directed to the polar D107 . The necessity of an
optimal steric and electronic fit is further illustrated by the aliphatic head groups (4a–g and 5a–g,
Table 1). The small tertiary amines 4c and 5c have an optimal fit (Table 1, Figure 6), whereas a
further increase of the size and lipophilicity (4d–g and 5d–g) results in a decreased binding
affinity and hence LE and GE (Table 1, Figure 6).
Combined WaterFLAP calculations (Figure 4) and siteꢀdirected mutagenesis studies (Figure 5)
indicated that optimal binding of the amine binding region is correlated with the presence of a
trapped ‘unhappy’ water and depends on the character of the residue on position 7.39, a residue
3
7,
that is highly variable in aminergic GPCRs and has been previously related to stereoꢀselective
3
8
39, 40
and subtypeꢀselective binding.
The additional Nꢀmethyl group of the tertiary amine 2
3.32
41
shields the Hꢀbond between 2 and D107 from bulk water and displaces the ‘unhappy’ water
3
.32
7.39
that is trapped between D107
and I454
in secondary amine 1 bound H R (Figure 4C, D).
1
7
.39
7.39
The role of I454
was corroborated by the fact that WT and I454 F mutant H R had a
1
7.39
similarly increased affinity for 2 compared to 1, whereas the I454 T mutant does not
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
distinguish between 1 and 2 (Figure 5). A 10ꢀfold affinity increase in H R and alpha adrenergic
1
7
.39
α2B receptor (F409 ) has also been reported for the tricyclic amitriptyline vs its Nꢀdemethylated
7
.39
analog nortriptyline, whereas their affinity is similar in the dopamine D receptor (T369 ),
3
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
further supporting the role of residue 7.39 as a determinant of specific Nꢀmethyl effects in amine
4
2
ligand binding to aminergic GPCRs (bioactivity data obtained from ChEMBL , document ID:
CHEMBL1909046 DrugMatrix).
Probing the aromatic binding regions. Binding site analysis (Figure 3) shows that the
secondary binding hot spot is positioned in the lower aromatic region, but is less pronounced
compared to the primary binding hot spot in the amine binding region. This is the result of the
(
substituted) benzyl group of ligands 1, 2, and 11 pointing with positive GEs in the lower
aromatic binding region. Moreover, the lower aromatic binding region also contains a cluster of
unhappy’ water molecules which is displaced upon ligand binding (Figure 4B–D), further
‘
supporting that this pocket serves as a secondary binding hot spot. The highly ligand efficient
deconstructed ligands 6 and 7 (Figure 6), lacking the benzyl moiety, are therefore also predicted
to bind in the lower aromatic region with their aromatic moieties (Supplementary Figure S1B,
C). However, growing the small fragment 6 from different vectors (1, 8, 9 and 10) illustrates that
the ability of the ligand to adopt a butterfly shape and optimally place both its amine and
aromatic pharmacophoric features in the corresponding H R binding site regions is essential
1
(Figure 6, Supplementary Figure S1D, E and F). To further probe the secondary binding hot spot,
i.e. lower aromatic binding region, a chloro moiety was introduced (11 and 12). Several known
H R ligands such as desloratadine contain a chloro substituent on one of the aromatic rings and it
1
is hypothesized that those chloro moieties bind in the lower aromatic region between residues
5.47
6.44
6.48
F199 , F424
and W428
(Supplementary Figure S4). Due to the introduction of a chloro
ACS Paragon Plus Environment

Page 29 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
substituent, the aromatic rings in 11 and 12 are shifted more towards the upper aromatic region,
resulting in a suboptimal positioning of the aromatic rings. Simultaneously targeting the amine
binding region and the lower aromatic binding region with an additional chloro substituent does
therefore not give an additional increase in affinity, suggesting that there is an optimum in
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
targeting the H R binding site (Figure 3 and 6). It should be noted that a phosphate ion was
1
3
identified in the H R Xꢀray which could be an additional binding region to target in future
1
studies.
In conclusion, multiple binding hot spots were identified and the amine binding region was
found to be the primary binding hot spot, preferably binding small tertiary amines. WaterFLAP
guided siteꢀdirected mutagenesis studies showed that the displacement of an ‘unhappy’ water
3
.32
7.39
molecule in between D107 and I454 is probably the reason for the increased affinity upon
Nꢀmethylation of ligand 1. Molecular docking studies in combination with water displacement in
the lower aromatic region reveals the secondary binding hot spot. Both identified binding hot
spots can be simultaneously targeted, but there is a binding optimum after which additional
substituents probing these hot spots do not improve ligand binding anymore (Figure 6). This
study illustrates that a crystal structure, in combination with carefully designed tool compounds,
computational assessment of water energetics, and siteꢀdirected mutagenesis enables subtle SAR
exploration that can be a general approach to explore binding sites in GPCRs for which structural
data is available.
ACS Paragon Plus Environment