First Structure-Activity Relationship of 17β-Hydroxysteroid Dehydrogenase Type 14 Nonsteroidal Inhibitors and Crystal Structures in Complex with the Enzyme

Article abstract of DOI:10.1021/acs.jmedchem.6b01436

17β-HSD14 belongs to the SDR family and oxidizes the hydroxyl group at position 17 of estradiol and 5-androstenediol using NAD⁺ as cofactor. The goal of this study was to identify and optimize 17β-HSD14 nonsteroidal inhibitors as well as to disclose their structure-activity relationship. In a first screen, a library of 17β-HSD1 and 17β-HSD2 inhibitors, selected with respect to scaffold diversity, was tested for 17β-HSD14 inhibition. The most interesting hit was taken as starting point for chemical modification applying a ligand-based approach. The designed compounds were synthesized and tested for 17β-HSD14 inhibitory activity. The two best inhibitors identified in this study have a very high affinity to the enzyme with a K_i equal to 7 nM. The strong affinity of these inhibitors to the enzyme active site could be explained by crystallographic structure analysis, which highlighted the role of an extended H-bonding network in the stabilization process. The selectivity of the most potent compounds with respect to 17β-HSD1 and 17β-HSD2 is also addressed.

Full text of DOI:10.1021/acs.jmedchem.6b01436

Subscriber access provided by UNIV OF WATERLOO
Article
First Structure-Activity Relationship of 17#-Hydroxysteroid
Dehydrogenase Type 14 Nonsteroidal Inhibitors
and Crystal Structures in Complex with the Enzyme
Florian Braun, Nicole Bertoletti, Gabriele Möller, Jerzy Adamski, Torsten
Steinmetzer, Mohamed Salah, Ahmed Saad Abdelsamie, Chris J van
Koppen, Andreas Heine, Gerhard Klebe, and Sandrine Marchais-Oberwinkler
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b01436 • Publication Date (Web): 09 Nov 2016
Downloaded from http://pubs.acs.org on November 10, 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 82
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
First Structure-Activity Relationship of 17β-Hydroxysteroid Dehydrogenase Type 14
Nonsteroidal Inhibitors and Crystal Structures in Complex with the Enzyme
†∆
†∆
‡
‡§$
†
Florian Braun , Nicole Bertoletti , Gabriele Möller , Jerzy Adamski , Torsten Steinmetzer ,
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
#
#Θ
#
†
†
Mohamed Salah , Ahmed S. Abdelsamie , Chris J. van Koppen , Andreas Heine , Gerhard Klebe ,
†
*
Sandrine Marchais-Oberwinkler
†
Institute for Pharmaceutical Chemistry, Philipps University Marburg, 35032 Marburg, Germany
German Research Center for Environmental Health, Institute of Experimental Genetics, Genome
‡
Analysis Center, Helmholtz Zentrum München, 85764 Neuherberg, Germany
§
Chair of Experimental Genetics, Technical University Munich, 85350 Freising-Weihenstephan,
Germany
$
German Center for Diabetes Research (DZD), 85764 Neuherberg, Germany
#
ElexoPharm GmbH, Campus A1.2, 66123 Saarbrücken, Germany
Abstract
1
5
7β-HSD14 belongs to the SDR family and oxidizes the hydroxyl group at position 17 of estradiol and
+
-androstenediol using NAD as cofactor. The goal of this study was to identify and optimize 17β-
HSD14 nonsteroidal inhibitors as well as to disclose their structure-activity relationship. In a first
screen, a library of 17β-HSD1 and 17β-HSD2 inhibitors, selected with respect to scaffold diversity,
was tested for 17β-HSD14 inhibition. The most interesting hit was taken as starting point for chemical
modification applying a ligand-based approach. The designed compounds were synthesized and tested
for 17β-HSD14 inhibitory activity. The two best inhibitors identified in this study have a very high
affinity to the enzyme with a K equal to 7 nM. The strong affinity of these inhibitors to the enzyme
i
active site could be explained by crystallographic structure analysis, which highlighted the role of an
extended H-bonding network in the stabilization process. The selectivity of the most potent
compounds with respect to 17β-HSD1 and 17β-HSD2 is also addressed.
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 82
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
Introduction
Human 17β-hydroxysteroid dehydrogenase type 14 (17β-HSD14), also called DHRS10 and retSDR3,
1
,2
is an oxido-reductase belonging to the SDR (Short-chain Dehydrogenase-Reductase) family. In
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
vitro, the enzyme oxidizes the hydroxyl group at position 17 of estradiol (E2) and 5-androsten-3β,17β-
+
diol (5-diol) in presence of the cofactor NAD , however, in vivo, its natural substrate is still unknown.
3
While Sivik et al. described a broad distribution pattern of 17β-HSD14 across various tissues based
on immunohistochemistry studies, Northern blotting experiments showed that the enzyme is
2
4
predominantly expressed in the brain, liver and placenta as well as in the kidney. Furthermore,
2
immunofluorescence studies revealed cytosolic localization.
In order to understand the function of 17β-HSD14, the enzyme needs to be further characterized.
Inhibitors are useful chemical tools, which can be used not only to characterize the binding site of an
enzyme but also to get insight into the physiological role of the latter upon in vivo administration.
5
With the exception of compound 1, which we recently presented, no inhibitor has been reported for
this enzyme.
The crystal structure of the human 17β-HSD14 has been determined recently as the holo protein (PDB
ID: 5JS6 and 5JSF) and as the ternary complex with the nonsteroidal inhibitor 1 (PDB ID: 5ICM).
5
17β-HSD14 is a tetramer. The binding cavity was shown to be rather lipophilic with a conical shape.
The substrate active site is narrow in the vicinity of the catalytic triad and is solvent exposed at the
other end. Tyr253’ from the C-terminal chain of the adjacent unit in the tetramer reduces the size of
the active site.
6
Up to now 14 different 17β-HSD subtypes have been reported. 17β-HSD1 and 17β-HSD2, the two
best characterized subtypes, predominantly catalyze the oxidation and reduction of estrogens and
7
–20
androgens. Inhibitors of these two enzymes have already been reported.
While 17β-HSD1 is a
cytosolic enzyme and shows a reductive activity in vivo (activation of estrogens), 17β-HSD2 is
membrane-bound and catalyzes the oxidation of estrogens and androgens to their less potent analogues
similarly to 17β-HSD14. However, 17β-HSD2 and 17β-HSD14 differ in their tissue distribution
2
1
pattern since in contrary to 17β-HSD14, 17β-HSD2 is not present in the human brain temporal lobe.
2
ACS Paragon Plus Environment

Page 3 of 82
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The presence of 17β-HSD14 in the brain might indicate that this enzyme is involved in the regulation
2
of active estrogens and androgens in this organ.
The goal of this work was to identify 17β-HSD14 nonsteroidal inhibitors and to optimize their
structures, which led to highly active compounds. The inhibitor optimization was performed following
a ligand-based approach. The synthesis and biological evaluation of highly active compounds with a
nonsteroidal scaffold together with four new crystal structures of the ternary complexes are reported.
Analysis of the crystal structures of the ternary complexes revealed the location of the inhibitor
binding sites as well as the resulting protein-inhibitor interactions and a complex pattern of hydrogen
bonds (H-bonds) contributing to the strong affinity of these compounds to the enzyme. The
physicochemical properties of the new inhibitors as well as selectivity considerations were also
addressed.
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
Design of 17β-HSD14 inhibitor candidates
Although 17β-HSDs belong to the same superfamily they share a low overall sequence identity.
Nonetheless, considering the fact that 17β-HSD1 and 17β-HSD2 catalyze the same reaction as
1
7β-HSD14, the substrate binding site of the three enzymes should exhibit a high structural similarity.
Based on this idea, it was assumed that some inhibitors developed for 17β-HSD1 and 17β-HSD2
should also bind to 17β-HSD14 and that a common scaffold could be used as starting point, to
optimize them for 17β-HSD14 binding. In a first screen, a small library of 34 17β-HSD1 and 17β-
HSD2 inhibitors, chosen on the basis of structural diversity (Figure 1), was tested for 17β-HSD14
inhibitory activity using a radioactive displacement assay. This assay was performed with the
recombinantly expressed enzyme in a bacterial suspension since the pure protein was not available at
that time. Thereby sets of active and inactive compounds were identified.
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 82
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
Figure 1: Scaffold of inhibitors from a 17β-HSD1 and 17β-HSD2 library tested for 17β-HSD14
inhibitory activity.
R
3
O
OH
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
S
R
OH
R
1
N
1
R
2
OH
A
B
Y
OH
OH
X
X = C, Y = S
X = S, Y = N
HO
O
R
1
C
D
HO
F
X
OH
R
2
OH
R
X = C, N
1
E
While the series of tested naphthalenes A and thiophene amides B contained mostly inactive
compounds, the dihydroxyphenylbenzenes C, -thiophenes and -thiazoles D showed examples of low to
moderate inhibitory activity against 17β-HSD14 (between 10% and 45% inhibition at 1 µM). In
addition, some of the latter derivatives were also reported to possess very high potency for 17β-HSD1
and/or 17β-HSD2 (IC50 in the low nM range), which might lead to difficulties achieving high
selectivity for 17β-HSD14. These parent scaffolds were therefore not considered for further
optimization. Some members of the pyridine ketone class E also showed remarkable inhibitory
activity for 17β-HSD14, paralleled by rather low or moderate activity against 17β-HSD1 (Table 1,
compound 2), which rendered this class as promising scaffold for further investigations. They were
therefore chosen as a starting point for the development of new 17β-HSD14 inhibitors.
4
ACS Paragon Plus Environment

Page 5 of 82
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 1: Most interesting compounds identified in a first screen.
Inhibition of
Inhibition of
Inhibition of
Cmpd
Structure
17β-HSD14
17β-HSD1
17β-HSD2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
b
c
%
Inh. @ 1 µM
IC (µM)
IC (µM)
50
50
HO
Me
A
A
A
5
F
2
3
4
5
B
C
62%
1.27
5.48
0.10
0.26
0.04
0.26
2
N
OH
O
HO
Me
5
B
C
0%
2
N
OH
O
HO
Me
F
B
C
19%
32%
0.29
OH
O
B
C
6
2
N
OH
A
19.65
O
HO
Me
a
3
+
Recombinant 17β-HSD14 enzyme, bacterial suspension, substrate [ H]-E2 [18.3 nM], NAD [7.5 mM], mean value of 3
determinations; standard deviation < 10 %.
b
3
Placental 17β-HSD1 enzyme, cytosolic fraction, substrate [ H]-E1 + E1 [500 nM], NADH [0.5 mM], mean value of at least
3 determinations; standard deviation < 20 %.
c
3
+
Placental 17β-HSD2 enzyme, microsomal fraction, substrate [ H]-E2 + E2 [500 nM], NAD [1.5 mM], mean value of at
least 3 determinations; standard deviation < 20 %.
The most interesting hits identified in the preliminary screen are listed in Table 1. The compounds can
be categorized into 2,5- (2, 3) and 2,6-substituted (5) pyridine ketones and the 1,4-substituted phenyl
(
4). In the 2,5-compound class, 2 and 3 (62% inhibition at 1 µM vs 0%) differ by the presence/absence
of a fluorine atom in ortho position to the OH group of the C-ring, suggesting the importance of this
atom for inhibitory activity. Furthermore, comparison of 2 with 4 (62% inhibition at 1 µM vs 19%)
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 82
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
shows that the pyridine core B is more potent than the phenyl analog, pointing toward the importance
of the nitrogen atom in the B-ring for activity. Concerning the 2,6-compound class, comparison of 5
with 2,5-substituted 3 (32% inhibition at 1 µM vs 0%) shows that moving the A-ring from 5- to 6-
position leads to a gain in activity.
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
Furthermore, considering the selectivity aspect, the poor inhibitory activity of the 2,6-substituted 5
toward 17β-HSD1 (IC = 19.65 µM), compared to the 2,5-substituted analog 3 (IC = 5.48 µM)
5
0
50
suggests that the 2,6-substitution pattern might improve selectivity for 17β-HSD14 over 17β-HSD1.
Selectivity against 17β-HSD2 does not become obvious with the set of studied test compounds.
Consequently, the 2,6-pyridine ketone class was selected for optimization and the derivatives, with
modification at the A-ring as well as at the C-ring by substituents with different properties were
synthesized. Special attention was paid to the physicochemical properties of the designed compounds
in order to focus on compounds which should have a promising bioavailability profile according to the
2
2
23
Veber rules and the Lipinski rules of 5.
Chemistry
The synthesis of compounds 1, 6, 8, 9, 11-23, 25, 29 and 33 was achieved in four steps starting from
the dibromopyridine derivatives 1d and 33d and analogous benzene 29d, respectively (Scheme 1).
2
4
Lithiation with n-butyl lithium and nucleophilic addition to the appropriate aldehyde provided the
alcohol intermediates 1c, 11c-15c, 29c and 33c. For the synthesis of compound 23 (Scheme 1),
arylbromide 23e was used to prepare the corresponding Grignard reagent. Nucleophilic addition to the
carbaldehyde 23d afforded the alcohol intermediate 23c. Oxidation with 2-iodoxybenzoic acid resulted
2
5
in the corresponding ketones 1b, 11b-15b, 23b, 29b and 33b. Subsequently, Suzuki couplings with
different arylboronic acids afforded compounds 1a, 6a, 8a, 9a, 11a-23a, 25a, 29a and 33a using either
standard conditions (Cs CO , 80 °C, overnight) or microwave irradiations (Na CO , 150 °C, 60 min,
2
3
2
3
1
50 W). Cleavage of the methoxy groups with boron tribromide gave the final compounds 1, 6, 8, 9,
1-23, 25, 29 and 33.
1
6
ACS Paragon Plus Environment

Page 7 of 82
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Scheme 1. Synthesis of compounds 1, 6, 8, 9, 11-23, 25, 29 and 33.
R
R
X
R
1
1
R
2
1
R
R
2
2
a
c
R
3
Br
R
Br
R
3
+
+
3
OMe
X
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
OMe
OMe
Br
X
Br
O
OH
O
1
d, 29d, 33d
b
1c, 11c-15c, 23c, 29c, 33c
1b, 11b-15b, 23b, 29b, 33b
d or e
R
1
R
1
Br
R
2
R
2
F
OMe
f
R
R
R
R
Ar
3
Ar
3
N
Br
X
X
OH
OMe
O
23d
23e
O
O
1, 6, 8, 9, 11-23, 25, 29, 33
1a,6a, 8a, 9a, 11a-23a, 25a, 29a, 33a
Cmpd
RAr
X
R1
R2
R3
OH-Pos
1
6
8
9
6-(3,4-dihydroxyphenyl)
6-(4-hydroxy-3-methylphenyl)
6-(3-fluoro-4-hydroxyphenyl)
6-(2-fluoro-3-hydroxyphenyl)
6-(2-fluoro-3-hydroxyphenyl)
6-(2-fluoro-3-hydroxyphenyl)
6-(2-fluoro-3-hydroxyphenyl)
6-(2-fluoro-3-hydroxyphenyl)
6-(2-fluoro-3-hydroxyphenyl)
6-phenyl
6-(3-N,N-dimethylamino)-phenyl)
6-(3-hydroxyphenyl)
6-(3-hydroxyphenyl)
6-(3-hydroxy-4-methylphenyl)
6-(4-fluorophenyl)
6-(3-chloro-4-fluorophenyl)
5-(3-fluoro-4-hydroxyphenyl)
6-(thiophen-3-yl)
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
CH
N
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
4-F
4-F
4-F
4-F
2-F 4-F
2-OH H
H
H
H
H
3
3
3
3
3
3
2
2
2
3
3
3
3
3
3
3
3
3
3
3
1
1
1
2
13
H
H
14
15
16
3-OH 4-OH
3-OH 6-OH
4-F
4-F
4-F
H
H
H
17
18
19
20
2-OH H
4-F
4-F
4-F
4-F
4-F
H
H
H
H
H
21
22
23
25
29
33
3-(2-fluoro-3-hydroxyphenyl)
6-(2-fluoro-3-hydroxyphenyl)
2-OH H
CH OH 4-F
H
2
Reagents and conditions: a) n-BuLi, anhydrous THF, -80 °C to room temperature, 1 h; b) Mg,
anhydrous THF, 60 °C, 2 h, 80 °C, 5 h; c) 2-iodoxybenzoic acid, anhydrous THF, 60 °C, 2-4 h; d)
Cs CO , Pd(PPh ) , R B(OH) , DME/water (2:1), 80 °C, overnight, for compounds 1a, 6a, 8a, 9a,
2
3
3 4
Ar
2
1
2a-15a, 17a-20a, 22a, 23a, 25a, 29a and 33a; e) Na CO , Pd(PPh ) , R B(OH) , DME/water (2:1),
2 3 3 4 Ar 2
microwave irradiation (60 min, 150 W, 150 °C), for compounds 11a, 16a and 21a; f), BBr3,
CH Cl , -80 °C to room temperature, overnight.
2
2
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 82
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
The precursor 33d was prepared starting from citrazinic acid, which was first brominated with
26
commercial phosphorus oxybromide and then reacted with methanol. Reduction of the formed ester
and protection of the primary alcohol 33e with TBSCl afforded compound 33d (Scheme 2).
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
Scheme 2. Synthesis of precursor 33d.
O
OH
OH
OTBS
Br
a,b
c
HO
N
OH
Br
N
Br
Br
N
33e
33d
Reagents and conditions: a) i) POBr , 130 °C, 0.5 h, 150 °C, 1.5 h, ii) MeOH, room temperature,
3
overnight; b) NaBH , EtOH, 85 °C, 3 h; c) TBSCl, imidazole, DMF, room temperature, 18 h.
4
The synthesis of the pyridine N-oxide 30 was achieved in four steps starting from the previously
described alcohol intermediate 12c. The Suzuki coupling reaction with 2-fluoro-3-
methoxyphenylboronic acid under standard condition afforded compound 30c. Oxidation into the
pyridine N-oxide 30b was performed by means of meta-chloroperoxybenzoic acid and further
oxidation with 2-iodoxybenzoic acid resulting in ketone 30a. Cleavage of the methoxy groups with
boron trifluoride dimethylsulfide complex afforded the final pyridine N-oxide 30 (Scheme 3).
8
ACS Paragon Plus Environment

Page 9 of 82
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Scheme 3. Synthesis of N-oxide 30.
a
N
OMe
Br
N
OMe
OH OMe
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
F
OH OMe
OMe
1
2c
30c
+
+
+
+
N
b
c
OMe
N
R
-
-
O
OH OMe
O
F
O
R
F
OMe
R
30b
30a R = OMe
R = OH
d
30
Reagents and conditions: a) Cs CO , Pd(PPh ) , 2-fluoro-3-methoxyphenylboronic acid, DME/water
2
3
3 4
(
2:1), 80 °C, overnight; b) m-CPBA (77%), CH Cl , 0 °C to room temperature, overnight; c)
2 2
2
-iodoxybenzoic acid, anhydrous THF, 60 °C, 2 h; d) BF ·SMe , CH Cl , room temperature,
3 2 2 2
overnight.
The synthesis of compound 24, lacking the A-ring, was performed by coupling of the pyridine
carbaldehyde 24c with the arylbromide 23e followed by the oxidation to ketone 24a as previously
described (Scheme 4). The synthesis of compounds 26 and 27 was realized by amination of 1b with
2
7
different nonaromatic N-heterocycles following the procedure described by Bollinger et al. to obtain
compounds 26a and 27a. Compound 28, with the A-ring linked via an ether bridge to the pyridine
moiety, was synthesized by reacting the 6-bromopyridine derivative 1b in a copper-catalyzed
2
8
O-arylation using picolinic acid as ligand, resulting in compound 28a. All methoxy groups were
cleaved with boron tribromide to provide compounds 24 and 26-28 (Scheme 4).
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 82
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
Scheme 4. Synthesis of compounds 24 and 26-28.
F
Br
N
OMe
O
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
1b
N
c or d
O
F
24c
F
a,b
e
+
R
N
OH
R
N
OMe
F
O
O
2
4 R = H
26 R = piperidin-1-yl
24a, 26a-28a
Br
OMe
2
2
7 R = 4-methylpiperazin-1-yl
23e
8 R = O-Ph
Reagents and conditions: a) Mg, anhydrous THF, 60 °C, 2 h, 80 °C, 5 h; b) 2-iodoxybenzoic acid,
anhydrous THF, 60 °C, 3-4 h; c) N-heterocycles , K PO , 1,4-dioxane, 100 °C, 2-4 d, for compounds
3
4
2
6a and 27a; d) Cu(I)I, picolinic acid, K PO , DMSO, 80 °C, 3 d, for compound 28a; e) BBr
3 4 3,
CH Cl , -80 °C to room temperature, overnight.
2
2
The synthesis of the trisubstituted C-ring derivative 10 was achieved in four steps starting from the 6-
bromopyridine-2-carbaldehyde 10d (Scheme 5). Suzuki coupling with 2-fluoro-3-
methoxyphenylboronic acid afforded the 6-substitued carbaldehyde 10c. Nucleophilic addition of the
in situ formed Grignard reagent of arylbromide 10e, oxidation of the formed alcohol with 2-
iodoxybenzoic acid and cleavage of the methoxy groups with boron tribromide resulted in the desired
compound 10.
The arylbromide 10e was obtained following a three-step procedure starting from 2-bromo-5-
fluorophenol 10h. First, a modified Casiraghi-formylation reaction using paraformaldehyde in the
2
9
presence of magnesium chloride and triethylamine gave the intermediate 10g. Subsequent Dakin
3
0
reaction with hydrogen peroxide and sodium hydroxide followed by protection of the formed
dihydroxy compound 10f with methyliodide afforded compound 10e.
1
0
ACS Paragon Plus Environment

Page 11 of 82
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The synthesis of trisubstituted pyridines 31 and 32 was achieved in seven and eight steps, respectively,
starting from compound 10d (Scheme 5). Protection of the aldehyde group with trimethyl
orthoformate gave the 2-dimethoxy derivative 31f. Iodination with lithium diisopropylamide and
31
iodine afforded the 5-iodo-6-bromo derivative 31e in moderate yield. For compound 31, subsequent
Suzuki coupling with 2.4 equivalents of 4-methoxy-3-methylphenylboronic acid provided the 5,6-
bisubstituted pyridine 31d. The reaction with 2-fluoro-3-methoxyphenylboronic acid resulted, under
the same conditions, selectively in the mono-coupled derivative 32e, still brominated in 6-position.
Further, a second Suzuki coupling with 4-methoxy-3-methylphenylboronic acid performed on 32e
yielded the asymmetric pyridine 32d. Both, 31d and 32d were deprotected with acetic acid to form the
aldehydes 31c and 32c. For 31, a nucleophilic reaction with the in situ formed Grignard reagent of
compound 23e resulted in the alcohol intermediate 31b. For 32 the corresponding alcohol 32b was
obtained after lithiation of arylbromide 23e with butyl lithium and subsequent nucleophilic reaction.
Oxidation with 2-iodoxybenzoic acid and ether cleavage with boron tribromide gave the final
compounds 31 and 32.
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
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 82
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
Scheme 5. Synthesis of compounds 10, 31 and 32.
OMe
I
F
h
j
a
3
2e
OMe
OMe
OMe
OMe
Br
N
Br
N
Br
N
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
O
OMe
OMe
1
0d
31f
3
1e
Br
N
a
a
a
R
2
R
1
F
F
g
N
R
3
N
+ Br
OMe
O
Br
R
HO
F
OMe
OH
OMe
10c
10e
10h R = H
31d R = 4-OMe, R = 3-Me, R = CH(OMe)
e
f
1
2
3
2
k
k
1
1
0g R = CHO
0f R = OH
31c
R
= 4-OMe, R
= 3-Me, R
= CHO
1
2
3
32d R = 3-OMe, R = 2-F, R = CH(OMe)
2
1
2
3
32c R = 3-OMe, R = 2-F, R = CHO
1
2
3
b,c,d
b (for 31b),
m (for 32b
)
R
R
2
2
R
R
1
F
1
F
F
c,d
N
OH
N
OH
N
OMe
O
OH
O
OH
F
HO
HO
10
3
3
1
2
R = 4-OH, R = 3-Me
31b R = 4-OMe, R = 3-Me
1 2
1
2
OH
R = 3-OH, R = 2-F
32b R = 3-OMe, R = 2-F
1 2
1
2
Reagents and conditions: a) Cs CO , Pd(PPh ) , R B(OH) , DME/water (2:1), 80 °C, overnight; b)
2
3
3 4
Ar
2
Mg, anhydrous THF, 60 °C, 2 h, 80 °C, 5 h; c) 2-iodoxybenzoic acid, anhydrous THF, 60 °C, 3-4 h; d)
BBr CH Cl , -80 °C to room temperature, overnight; e) (CHO) , MgCl , Et N, anhydrous THF,
3
,
2
2
n
2
3
8
0 °C, 5 h; f) 1 M NaOH, H O , THF, room temperature, 4 h; g) MeI, K CO , DMF, room
2 2 2 3
temperature, overnight; h) HC(OMe) , p-TsOH, anhydrous MeOH, reflux, 4 h; j) LDA, anhydrous
3
THF -80 °C, 3 h, I , anhydrous THF, -80 °C to room temperature; k) acetic acid/water (2:3), reflux,
2
2
h; m) n-BuLi, anhydrous THF, -80 °C to room temperature, 1 h.
1
2
ACS Paragon Plus Environment

Page 13 of 82
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Calculation of physicochemical parameters
For each synthesized compound, the molecular weight (MW) was calculated to be in the range of 300
to 400 g/mol, aside from the trisubstituted compounds 31-33 with a slightly higher MW. The log P
3
2
was calculated in silico (using Molinspiration ) and turned out to be below 5 except for 31-33. 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
total polar surface area (TPSA), the number of rotational bonds as well as the number of H-bond
2
2
23
donors and acceptors fulfill the Veber rules and the Lipinski rules of 5. In addition, considering the
potential role of the enzyme in the brain, the capability of the inhibitors to cross the BBB should also
be taken into account. The physicochemical properties to be met by compounds showing a good BBB
3
3
penetration are described by Pajouhesh and Lenz. These criteria are matched for most of the
synthesized compounds especially for 17 and 28 as examples. The solubility range of most of the
compounds was also determined by mixing several concentrations of the studied inhibitors in 100 mM
phosphate buffer at pH 7.4 and analyzing its precipitation status at different time points (0, 1, 2 and
2
4 h). A table summarizing the physicochemical parameters can be found in the Supporting
Information (Table S3). The pKa values of the OH groups at the A- and at the C-ring were determined
in silico for all compounds (using Marvin Sketch), showing that the introduction of a fluorine group in
ortho position to a OH group decreases the pKa value by about one unit (e.g. pKa OH/C-ring: 7.8 for
8, 8.8 for 7).
Biological evaluation
Inhibition of 17β-HSD14 determined with a fluorimetric assay
A fluorimetric assay, quantifying the NADH fluorescence built up during the catalytic reaction, was
used to evaluate the inhibitory activity of the synthesized compounds. In the assay, the purified
+
recombinantly expressed human enzyme, E2 as substrate, NAD as cofactor, and the inhibitor were
5
used as already reported. Due to the low sensitivity of the assay, a high enzyme concentration
(between 3.0 µM and 3.5 µM) and a high concentration of substrate E2 (32 µM) was necessary. The
results are expressed as percent of inhibition measured at an inhibitor concentration of 2 µM. The
inhibition constant K was experimentally determined using an inhibitor concentration ranging from
i
1
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 82
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
2
.6 µM to 100 µM or 260 nM to 10 µM, depending on the inhibitor potency. As the inhibitor and
34,35
protein concentrations were in the same range, no classical kinetic analysis could be applied.
The
results were analyzed applying the quadratic Morrison equation for tight binding (see Supporting
3
6
Information). When the inhibitor was not sufficiently soluble at the required concentration, no K
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
i
could be determined and the results were expressed as percent inhibition at the highest soluble
concentration of the inhibitor. The results are shown in Tables 2-5. Compounds showing less than 10%
inhibition at a concentration of 100 µM were considered to be inactive.
Aggregation
It was verified that the compounds did not aggregate or induce aggregation of the protein by testing
the compounds with and without detergent (0.5% v/v Tween 20). No significant changes in the
inhibitor activity could be noticed at the different conditions.
17β-HSD14 inhibitory activity
Starting from the hit pyridine ketone 5, modifications were undertaken at the C-ring and A-ring by
introduction of different substituents (R, R’, Chart 1). These substituents were selected to represent
different electronic properties: electron donating or electron withdrawing, H-bond donor, H-bond
acceptor, lipophilic and hydrophilic. Changes were also performed at the central B-ring by
replacement of the nitrogen by a carbon or a N-oxide moiety. An additional phenyl ring (D) with
various substituents (R’’) were introduced in 5-position as well as a hydroxymethyl group (R’’’) in
4-position leading to trisubstituted derivatives.
Chart 1: Modifications undertaken on the hit compound 5.
R´´´
4
R´´
R´
D
A
B
C
6
2
5
N
OH
C
B
X
R
6
2
A
O
HO
O
Me
5
1
4
ACS Paragon Plus Environment

Page 15 of 82
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Substituent variations on the C-ring
In the first inhibitor screen it could be shown, that in the 2,5-pyridine class (Table 1) the addition of a
fluorine atom in ortho position to the hydroxyl group at the C-ring (2 compared to 3) resulted in a
notable increase in potency of the inhibitor. The analogous fluorinated compound 6 (2,6-substituted
pyridine) was therefore synthesized and proved to bind 10 times more potent than the reference
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
compound 5 (K = 245 nM for 6 vs 26 nM for 5, Table 2). The affinity enhancement caused by the 4-F
i
substituent at the C-ring could also be observed using another substitution pattern at the A-ring: 3-F/4-
OH (K = 467 nM for 7 vs 36 nM for 8, Table 2).
i
The influence of the substituent pattern used for the C-ring was subsequently studied in more detail
with compounds containing a 2-F/3-OH phenyl A-ring motif (compounds 9-15, Table 2). Addition of
a 2-OH or a 2-F to the 3-OH/4-F at the C-ring resulted in equipotent compounds (K = 13 nM; 11 nM
i
and 9 nM, respectively for 9, 10 and 11). Replacing the 3-OH/4-F at the C-ring (9) by a 2-OH/3-OH
motif (12; K = 64 nM) led to a slight decrease in affinity. Addition of a 4-OH group (14, K = 405 nM)
i
i
or of a 6-OH group at the C-ring (15, K = 796 nM) resulted in a strong decrease in activity.
i
The presence of the 3-OMe group at the C-ring (16a, inactive at a concentration of 100 µM) was
detrimental for the inhibitory activity compared to the 3-OH analogue (16, K = 63 nM).
i
1
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 82
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: 17β-HSD14 inhibitory activity and binding constant (K ) of 2,6-pyridine derivatives with
i
different substituents at the C-ring
R
5
4
B
N
C
R
4
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
2
6
3
2
R
3
R
A
1
4
2
O
R
2 3
5
-16
17β-HSD14
a
Compound
R1
R2
R3
R4
R5
% inhibition
K (nM)
i
a
@
2 µM
5
6
7
8
9
3-Me 4-OH
3-Me 4-OH
H
H
H
H
H
3-OH
3-OH
3-OH
3-OH
3-OH
H
4-F
H
34
60
16
67
72
76
72
65
66
25
11
245±21
26±3
3-F
3-F
2-F
2-F
2-F
2-F
2-F
2-F
2-F
4-OH
4-OH
3-OH
467±91
36±5
4-F
4-F
4-F
4-F
H
13±5
1
1
1
0
1
2
3-OH 2-OH 3-OH
3-OH 2-F 3-OH
3-OH 2-OH 3-OH
3-OH 2-OH
11±3
9±3
64±4
13
H
H
135±2
405±177
796±122
1
1
4
5
3-OH 2-OH 3-OH
3-OH 2-OH 3-OH
4-OH
6-OH
1
6a
H
H
H
H
H
H
3-OMe
3-OH
4-F
4-F
ni
nd
16
57
63±3
a
+
Recombinantly expressed purified 17β-HSD14 enzyme, fluorimetric assay, substrate E2 [32 µM], NAD [1.2
mM], 25°C, mean value of at least two independent experiments each with three technical repeats; ni: no inhibition
<10% inhibition at 100 µM), nd: not determined.
(
1
6
ACS Paragon Plus Environment

Page 17 of 82
Journal of Medicinal Chemistry
1
2
3
4
5
6
Substituent variations on the A-ring
2
,6-Pyridine derivatives containing the C-ring motif (3-OH/4-F or 2-OH/3-OH) were synthesized with
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
different substituents at varying positions of the A-ring (Table 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
Compounds with one substituent in the 3- or 4-position (17: 3-NMe , K = 7 nM; 18: 3-OH, K = 7 nM;
2
i
i
1
9: 3-OH, K = 44 nM and 21: 4-F, K = 221 nM) showed that their substitution with a polar moiety at
i i
3-position (17, 18, 19) led to stronger binding compared to the one with a lipophilic group at 4-
position (21). This effect is confirmed in case the A-ring is disubstituted (22) with the 3-Cl and 4-F
substituents, which led to a compound with a similar binding constant as the mono 4-F derivative 21
(
K = 190 nM and 221 nM for 22 and 21, respectively). These lipophilic groups exerted a detrimental
i
effect on the binding affinity, which was lower compared to the unsubstituted phenyl (16, K = 63 nM).
i
The compounds with two substituents in 2/3- or in 3/4-positions of the A-ring (1: 3-OH/4-OH, K = 7
i
nM; 6: 3-Me/4-OH, K = 26 nM; 8: 3-F/4-OH, K = 36 nM; 9: 2-F/3-OH, K = 13 nM) had similar
i
i
i
binding constants with the exception of 20 (3-OH/4-Me, K = 47 nM) with a slightly decreased affinity.
i
No significant difference in activity could be observed between mono- and disubstituted compounds at
the A-ring as long as a 3-OH or a 4-OH moiety was present. In summary, the best affinities were
achieved in the presence of a 3-OH or a 3-NMe moiety at the A-ring.
2
In the 2,5-pyridine class, the affinity of 23 (K = 17 nM) with a 3-F/4-OH substitution pattern at the A-
i
ring was similar to that of compound 2 (K = 24 nM) with a 3-Me/4-OH substitution pattern at the A-
i
ring and fell into the same range of compounds in the 2,6-class. Furthermore the methoxy derivative
2
3a was less active compared to the hydroxylated analogue 23 as similarly observed for compound
6a.
1
1
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 82
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 3: 17β-HSD14 inhibitory activity and binding constant (K ) of pyridine derivatives with
i
different substituents at the A-ring
4
R
A
4
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
3
5
B
N
C
R
4
3
4
6
2
R
2 2
B
N
C
R
4
2
2
R
3
3
R
A
1
2
R
3
4
2
O
R
2 3
O
1
, 6, 8, 9, 12, 16-22
2, 23a, 23
17β-HSD14
Position
A-ring
a
Compound
R1
R2
R3
R4
% inhibition
K (nM)
i
a
@
2 µM
69
60
60
67
72
65
57
57
65
51
64
37
44
12
61
1
2
6
8
9
6
5
6
6
6
6
6
6
6
6
6
6
6
5
5
3-OH
3-Me
3-Me
3-F
4-OH
4-OH
4-OH
4-OH
3-OH
3-OH
H
3-OH
3-OH
3-OH
3-OH
3-OH
2-OH
3-OH
3-OH
3-OH
2-OH
3-OH
3-OH
3-OH
4-F
4-F
7±1
24±9
26±3
36±5
13±5
64±4
63±3
7±1
4-F
4-F
2-F
4-F
12
16
17
18
2-F
3-OH
4-F
H
3-NMe
3-OH
H
H
4-F
2
H
4-F
7±2
19
3-OH
4-Me
4-F
3-OH
4-F
44±3
47±7
221±46
190±45
20
21
22
3-OH
H
4-F
3-Cl
3-F
4-F
4-F
b
2
3a
4-OMe 3-OMe
4-OH 3-OH
4-F
57%±6
17±5
23
3-F
4-F
a
+
Recombinantly expressed purified 17β-HSD14 enzyme, fluorimetric assay, substrate E2 [32 µM], NAD [1.2
b
mM], 25°C mean value of at least two independent experiments each with three technical repeats; % inhibition @
2
2,2 µM
1
8
ACS Paragon Plus Environment

Page 19 of 82
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Variation of the A-ring
In order to investigate the role of the phenyl A-ring, pyridine derivatives lacking the A-ring (24) or
derivatives decorated with different heterocycles in 6-position of the B-ring (25-28) were designed
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 4). Comparison of compound 24 (K = 1541 nM) with 16 (K = 63 nM) showed a decrease in
i i
activity. The nonaromatic piperidine 26 (K = 407 nM) was also a weaker binder, while the 4-
i
methylpiperazine 27 (K = 190 nM) showed a similar inhibitory activity, both compared to 16.
i
Table 4: 17β-HSD14 inhibitory activity and binding constant (K ) of 2,6-pyridine derivatives with
i
different substituents in 6-position of the pyridine ring (different A-rings)
F
B
N
C
6
2
R
OH
O
16, 24-28
17β-HSD14
a
Compound
R
% inhibition
K (nM)
i
a
@ 2 µM
1
2
2
2
6
4
5
6
phenyl
H
57
13
50
43
31
55
63±3
1541±146
97±28
thiophen-3-yl
piperidin-1-yl
407±16
190±21
50±0
27
4-methylpiperazin-1-yl
28
-O-phenyl
a
+
Recombinantly expressed purified 17β-HSD14 enzyme, fluorimetric assay, substrate E2 [32 µM], NAD [1.2
mM], 25°C mean value of at least two independent experiments each with three technical repeats
1
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 82
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
Variations on the B-ring
In the design section it was reported that in the 2,5-pyridine class the central ring played a crucial role
in affinity. This aspect was also investigated in the 2,6-class by the synthesis of the phenyl analogue
2
9 and the corresponding N-oxide 30 (Table 5). Comparison of the biological data of these compounds
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
with the pyridine analogue 12 (K = 64 nM) showed that both, the phenyl derivative 29 (K = 21 nM)
i
i
and the N-oxide derivative 30 (K = 132 nM) had a similar affinity as the pyridine 12.
i
Table 5: 17β-HSD14 inhibitory activity and binding constant (K ) for compounds with different B-
i
rings
B
X
C
6
2
OH
A
O
OH
F
OH
12, 29, 30
1
7β-HSD14
a
Compound
X
% inhibition
K (nM)
i
a
@
2 µM
77
12
N
C
64±4
21±2
2
9
0
60
+
-
3
N -O
47
132±13
a
+
Recombinantly expressed purified 17β-HSD14 enzyme, fluorimetric assay, substrate E2 [32 µM], NAD [1.2
mM], mean value of at least two independent experiments each with three technical repeats
Trisubstituted pyridines
Based on the hypothesis that the C-ring and the carbonyl group of both the 2,5- and 2,6-pyridines
achieve the same interaction as a consequence of binding in the same area, the similarity in affinity of
the 2,5-derivative 23 (K = 17 nM) and the 2,6-derivative 8 (K = 36 nM, Table 3) suggested that the
i
i
2
0
ACS Paragon Plus Environment

Page 21 of 82
Journal of Medicinal Chemistry
1
2
3
4
5
available space for binding the additional ring must be large, even tolerating an A-ring attached to the
5
- or 6-position. Trisubstituted pyridines combining substitutions at the 5- and 6-positions were
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
therefore synthesized (31 and 32, Table 6). These trisubstituted compounds turned out to be equipotent
(
31, K = 9 nM; 32, K = 15 nM) with a similar affinity compared to the disubstituted derivatives 6 and
i 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
23.
Introducing a hydroxymethyl group at the 4-position of the B-ring led to a slight decrease in activity
33 K = 86 nM compared to 9 K = 13 nM, Table 6).
(
i
i
Table 6: 17β-HSD14 inhibitory activity and binding constant (K ) of pyridine derivatives with an
i
additional D-ring or a substituent in 4-position
OH
4
4
R
D
F
1
5
F
3
B
N
C
2
R
2 2
6
B
N
C
6
2
OH
OH
A
O
A
O
F
HO
OH
Me
31, 32
33
17β-HSD14
a
Compound
R1
R2
% inhibition
K (nM)
i
a
@
2 µM
60
6
26±3
17±5
9±4
2
3
3
1
61
3-Me
2-F
-
4-OH
3-OH
-
71
32
33
63
15±4
86±7
47
a
+
Recombinantly expressed purified 17β-HSD14 enzyme, fluorimetric assay, substrate E2 [32 µM], NAD [1.2
mM], mean value of at least two independent experiments each with three technical repeats
2
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 82
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
3
7
Pan Assay Interference Compounds
All the biologically evaluated compounds were tested in silico for nonspecific binding in order to
3
8
identify false positives using the Pains-remover computer tool. From the compounds analyzed, seven
did not pass the filter, including 10, 12, 14, 19, 20, 29, 30 suggested as nonspecific binders. They all
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
39
share as common characteristic a catechol moiety. It is known that catechols can be toxic, however,
the inhibitory data of these compounds are presented as they are useful to establish a better structure-
activity relationship comprehension. In case these compounds turn out to be highly interesting, further
assays should be performed to characterize their toxicity.
Crystal structures determination
The inhibitors with the highest binding affinity were selected for crystal structure determination in
order to get insight into their binding mode. Crystal structures could be obtained for four different
inhibitors by co-crystallization of the protein in complex with cofactor and ligands (6, PDB ID: 5L7T;
9
, PDB ID: 5L7Y; 10, PDB ID: 5L7W; 12, PDB ID: 5EN4). The data collection, processing, and
refinement statistics details are reported in the Supporting Information. The crystals were obtained by
two different conditions, however, all crystal structures show the same tetragonal space group (I422)
with only one monomer present in the asymmetric unit. The crystal structures disclosed, that the
5
protein is a homotetramer, in accordance with a previous study. The structures obtained have a
resolution ranging from 1.52 Å to 2.02 Å. The conformation of the protein in complex with inhibitor 1
5
(
PDB ID: 5ICM) has already been published and will not be the focus here. It will however be
included in the structural comparison.
Description of the inhibitor binding site
The superimposition of all five ternary complexes reveals that binding of the different ligands does not
induce conformational changes of the overall geometry of the protein (Figure 2), showing a mean
RMSD of 0.15 ±0.04 Å between the alignment of the Cα atoms of the structures, as calculated with
2
2
ACS Paragon Plus Environment

Page 23 of 82
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
4
0
COOT. The cofactor interacts with the Rossman fold region and experiences similar interactions as
5
already observed for the 17β-HSD14 holo structure and in complex with inhibitor 1. The inhibitor
binds into the substrate binding site, which is restricted by two α-helixes from the flexible loop (αFG1
and αFG2, residues 189-212, in orange), a portion of the αF helix (containing the catalytic Tyr154,
Lys158, in yellow), the short αEF helix (residue 142-146, in cyan), the segment that connects αE with
βD (from Asn89 to Arg98, in green), and the C-terminal tail (in red, Figure 3). Furthermore, Tyr253’
from the adjacent monomer (in gray on the right hand side, Figure 3) is pointing toward the inhibitor
binding site, reducing the volume of the active site cleft. The annotation of the different helices and β-
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
sheets follows the nomenclature described by Lukacik et al.. The flexible loop is in a conformation
that closes the binding pocket, reducing the size of the substrate binding site. Furthermore, the
inhibitor binding site is predominantly hydrophobic, with two main hydrophilic regions: The first one
corresponds to the two residues of the catalytic triad Tyr154 and Ser141, and the second one is formed
by His93 and Gln148 (Figure 4). This second region shapes the binding site in a peculiar form and
could be relevant for the achievement of selectivity considering that no other human SDR 17β-HSDs
presents a histidine at this position.
Figure 2. Superimposition of the crystal structures of 17β-HSD14 obtained in ternary complexes with
five inhibitors: 1, 6, 9, 10, 12. (A) The enzyme 17β-HSD14 is shown as ribbon model (5ICM in
orange, 5L7T in pink, 5L7Y in purple blue, 5L7W in ocher and 5EN4 in green); inhibitors are shown
+
as stick models. The cofactor NAD is shown as thin line. (B) Close-up view on the binding pocket.
The protein 17β-HSD14 is displayed by use of the solvent accessible surface. The carbon atoms of the
2
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 82
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
inhibitors are shown for 1 in orange 6 in pink, 9 in purple blue, 10 in ocher and 12 in green. Inhibitors
are shown as stick models and cofactor as thin line. The water molecules W1 and W2 are represented
in the same color as the corresponding inhibitor of the individual structures. W1 corresponds to water
molecule 472 in 5ICM, 518 in 5L7T, 508 in 5L7Y, 496 in 5L7W, 450 in 5EN4; W2 corresponds to
water molecule 530 in 5ICM and 502 in 5L7T, in the respective crystal structure. All structural
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
4
1
representations were prepared with PyMOL.
Figure 3. Ribbon representation of inhibitor 6 in complex with the protein 17β-HSD14 and cofactor
+
NAD . The inhibitor binding site is delimited by αFG1 and αFG2 (orange), αF (the helix containing
the catalytic Tyr154 and Lys158, yellow), αEF (cyan), αE and βD (green) and C-terminal tail (red).
Inhibitor 6 is shown as stick model and its carbon atoms are colored in pink. The symmetry equivalent
molecule containing Tyr253’ is shown in gray on the right hand side. The cofactor and Tyr253’ are
shown as thin lines.
2
4
ACS Paragon Plus Environment

Page 25 of 82
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. Surface representation of 17β-HSD14; color coded according to the Eisenberg
hydrophobicity scale (from dark red for highly hydrophobic amino acids to white for highly
4
2
+
hydrophilic amino acids). The cofactor NAD and amino acids are shown as stick models. The amino
acid of the symmetry equivalent molecule is referred as prime (’).
Description of the binding mode of inhibitors in complex with 17β-HSD14.
In the surface representation, it is obvious that the V-shape of the inhibitor scaffold matches well with
the geometry of the active site (Figure 2B). For all five crystallized inhibitors, additional water
molecules are observed in the binding pockets: Water W1 is found for all inhibitors in the same
position. W1 is localized between the B- and C-ring, at an approximate 4 Å distance from the B-ring.
W1 establishes an H-bond interaction with the side-chain of Asn186 (d≈ 2.7 Å). Water W2 is only
observed for 1 and 6 (Figure 2B) and found above the plane of the A-ring. W2 interacts with W1 (H-
bond contact d≈ 2.6 Å). In case of the other three inhibitors, the orientation of the A-ring plane is
shifted, inducing the displacement of W2.
Due to close similarity, only the interactions of 6 and 12, as representatives of the series, will be
described in detail. The structures and details of the interactions of all inhibitors are shown in Figure 5
(
A: 6, B: 12; C: 9; D: 10). Inhibitors 6 and 12 have the same B-ring scaffold (2,6-pyridine) differing in
2
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 82
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
the nature of their A- and C-ring substituents. At the C-ring, the keto group and the pyridine ring of 6
and 12 bind exactly at the same position. The angle between the keto group and the phenyl C-ring is
identical in 6 and 12 independent of the presence or absence of the 2-OH group at the C-ring.
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
The 3-OH groups at the C-ring of 6 and 12 interact via remarkably short H-bond interactions with the
side chain of Tyr154 (6, d= 2.4 Å, 12, d= 2.5 Å) and the side chain of Ser141 (6, d= 2.5 Å, 12, d= 2.5
Å) from the catalytic triad. The 4-F group at the C-ring of 6 is not involved in any specific interaction.
The phenyl C-ring of 6 and 12 is stabilized by van der Waals contacts with the nicotinamide moiety of
+
NAD . The carbonyl group is not involved in any direct interaction. The central pyridine B-ring is
anchored by van der Waals contacts with Trp192 and Leu195, which wrap around the top part of the
pyridine ring. No close contacts are observed with His93.
For 6, an H-bond interaction is formed between the 4-OH at the A-ring with the carbonyl backbone of
Ala149 (d= 2.7 Å). The 3-Me group at the A-ring is not involved in any interaction. The aromatic A-
ring is not stabilized by any π-stacking interactions, however, van der Waals interactions with Pro96
are observed. No water mediated H-bond interactions are observed; nonetheless, it is remarkable that
W1 remains present in this lipophilic environment. In total, 91 van der Waals contacts are achieved by
the ligand and its surface is buried to 94.0 % (considering only one monomer of the protein, the
number of van der Waals contacts achieved are summed up to 86 and the buried surface is 87.4 %).
In 12, a rotation of the A-ring plane is observed. Nevertheless, the altered orientation still allows the 3-
OH group at the A-ring to be at H-bond distance to Ala149-CO (d= 2.8 Å, as observed with 6). The
aromatic A-ring can also establish a van der Waals interaction with the amino acid Pro96. In summary,
82 van der Waals contacts are observed for 12 with 93.0 % of its surface buried in the protein binding
pocket (considering only one protein monomer, the ligand achieves 80 van der Waals contacts and its
surface is buried to 88.2 %).
2
6
ACS Paragon Plus Environment

Page 27 of 82
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. Crystal structures of 17β-HSD14 in complex with cofactor NAD , and inhibitors 6 (in pink,
A), 12 (in green, B), 9 (in purple blue, C) and 10 (in ocher, D). The inhibitors are shown as stick
models. The amino acids, within a distance of 5 Å, and the cofactor are shown as thin lines. H-bonds
are depicted as dotted lines. Distances are given in Å.
Interestingly, the crystal structure of 9 in ternary complex with 17β-HSD14 shows the presence of a
second inhibitor molecule at the interface between two tetramers (Figure 6A). Close inspection of this
interface binding site (Figure 6B) highlights that the inhibitor is stabilized through an H-bond
interaction between the 3-OH group at the A-ring to the hydroxyl group of Ser44 side chain (d= 3.2 Å)
and with a water molecule W405 (d= 2.6 Å), which is stabilized by Arg19 (d= 3.0 Å). The keto group
2
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 82
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
interacts with the other tetramer through a water molecule W434 (d= 2.7 Å), which is also bound to
the NH from the backbone of Leu83´ (d= 3.0 Å) and the carbonyl group of the backbone of Tyr128´
(
d= 2.9 Å). The copy of 9 binding to the interface is placed in a rather hydrophilic environment. The
overall geometry of the interface ligand differs from that of the active site ligand: The dihedral angles,
for the ligand in the active site and for the ligand binding at the interface, between the keto group and
C-ring are -29° and 4° respectively, and between the keto group and the B-ring are 129° and -133°
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
respectively (considering the plane through the keto group as 0°). The dihedral angle between C at the
2
A-ring and the nitrogen at the B-ring is 131° for the ligand in the binding pocket and -60° for the
ligand present at the interface. A superimposition of the interface and active site compounds can be
seen on Figure 7.
Figure 6. (A). Overall view of the crystal structure of two 17β-HSD14 tetramers in complex with 9.
The protein monomers are shown as ribbon models and colored in gray. The inhibitors are shown as
sphere models. The inhibitors located in the substrate binding site are colored in purple blue while the
inhibitors located between the interfaces of the tetramers are shown in yellow. (B). Close-up view of
the second ligand binding site of 9 located at the interface between two tetramers. The enzyme is
displayed by use of the solvent accessible surface. Inhibitor 9 is shown as stick model. The amino
acids are shown as thin lines. The amino acids of the symmetry equivalent molecule are referred to
prime (’). H-bonds are depicted as black dotted lines. Distances are given in Å.
2
8
ACS Paragon Plus Environment

Page 29 of 82
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 7. Superimposition of compound 9 based on the pyridine B-rings of the ligand at the interface
(in yellow) and active site (in purple blue).
Comparison of the 17β-HSD1, 17β-HSD2 and 17β-HSD14 structures
+
The existing crystal structure of 17β-HSD1 in ternary complex with the cofactor NADP and E2 (PDB
ID: 1FDT) allows the direct comparison with the 17β-HSD14 structure in complex with the cofactor
+
NAD and estrone (E1, PDB ID: 5HS6). The superimposition reveals a structural conservation of the
4
0
enzymes only in some regions (22% of sequence identity calculated with COOT, Figure 8). Small
+
+
differences in the NAD /NADP binding site can be observed between type 14 and type 1 (RMSD of
40
1.6 Å calculated with COOT based on Cα alignment). This result was expected since both enzymes
+
+
bind a different cofactor (NAD for 17β-HSD14 vs NADP for 17β-HSD1). The flexible loop in 17β-
HSD1 restricts the end of the binding cavity, while the corresponding loop in 17β-HSD14 is shifted
upwards leaving the binding site widely open. This results in a smaller binding pocket for the type 1
enzyme. Furthermore, while the catalytic triad is conserved in both enzymes, the steroids
accommodate in the binding sites with different orientations and achieve distinct interactions.
As no crystal structure is available for 17β-HSD2, the comparison between the structures of type 2 and
type 14 is not possible.
2
9
ACS Paragon Plus Environment