Molecular Hybridization of Potent and Selective γ-Hydroxybutyric Acid (GHB) Ligands: Design, Synthesis, Binding Studies, and Molecular Modeling of Novel 3-Hydroxycyclopent-1-enecarboxylic Acid (HOCPCA) and trans-γ-Hydroxycrotonic Acid (T-HCA) Analogs

Article abstract of DOI:10.1021/acs.jmedchem.7b01351

γ-Hydroxybutyric acid (GHB) is a neuroactive substance with specific high-affinity binding sites. To facilitate target identification and ligand optimization, we herein report a comprehensive structure-affinity relationship study for novel ligands targeting these binding sites. A molecular hybridization strategy was used based on the conformationally restricted 3-hydroxycyclopent-1-enecarboxylic acid (HOCPCA) and the linear GHB analog trans-4-hydroxycrotonic acid (T-HCA). In general, all structural modifications performed on HOCPCA led to reduced affinity. In contrast, introduction of diaromatic substituents into the 4-position of T-HCA led to high-affinity analogs (medium nanomolar K_i) for the GHB high-affinity binding sites as the most high-affinity analogs reported to date. The SAR data formed the basis for a three-dimensional pharmacophore model for GHB ligands, which identified molecular features important for high-affinity binding, with high predictive validity. These findings will be valuable in the further processes of both target characterization and ligand identification for the high-affinity GHB binding sites.

Full text of DOI:10.1021/acs.jmedchem.7b01351

Subscriber access provided by Gothenburg University Library
Article
Molecular hybridization of potent and selective #-Hydroxybutyric
Acid (GHB) ligands; Design, synthesis, binding studies, and
molecular modelling of novel 3-hydroxycyclopent-1-enecarboxylic
Acid (HOCPCA) and trans-#-hydroxycrotonic acid (T-HCA) Analogs
Jacob Krall, Claus Hatt Jensen, Francesco Bavo, Christina Birkedal Falk-Petersen,
Anne Stahr Haugaard, Stine Byskov Vogensen, Yongsong Tian, Mia Nittegaard-Nielsen,
Sara Björk Sigurdardóttir, Jan Kehler, Kenneth Thermann Kongstad, David E. Gloriam,
Rasmus Prætorius Clausen, Kasper Harpsøe, Petrine Wellendorph, and Bente Frølund
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b01351 • Publication Date (Web): 13 Oct 2017
Downloaded from http://pubs.acs.org on October 15, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 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
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
Molecular hybridization of potent and selective γꢀ
Hydroxybutyric Acid (GHB) ligands; Design,
synthesis, binding studies, and molecular
modelling of novel 3ꢀhydroxycyclopentꢀ1ꢀ
enecarboxylic Acid (HOCPCA) and transꢀγꢀ
hydroxycrotonic acid (TꢀHCA) Analogs
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
†,‡
†,‡
†,ƒ,‡
†
Jacob Krall, Claus Hatt Jensen, Francesco Bavo,
Christina Birkedal FalkꢀPetersen,
†
†
†
†
Anne Stæhr Haugaard, Stine Byskov Vogensen, Yongsong Tian, Mia NittegaardꢀNielsen, Sara
†
§
†
†
Björk Sigurdardóttir, Jan Kehler, Kenneth Thermann Kongstad, David E. Gloriam, Rasmus
†
†
†
†,
Prætorius Clausen, Kasper Harpsøe, Petrine Wellendorph, and Bente Frølund *
†
Department of Drug Design and Pharmacology, Faculty of Health and Medical Sciences,
ƒ
University of Copenhagen, DKꢀ2100 Copenhagen, Denmark. Department of Pharmaceutical
ACS Paragon Plus Environment

Page 3 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
§
Sciences, University of Milan, Italy. Discovery Chemistry, H. Lundbeck A/S, Ottiliavej 9, DKꢀ
2
500 Valby, Denmark.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Keywords γꢀhydroxybutyric acid (GHB), 3ꢀhydroxycyclopentꢀ1ꢀenecarboxylic acid (HOCPCA),
transꢀ4ꢀhydroxycrotonic acid (TꢀHCA), highꢀaffinity GHB binding site, structureꢀaffinity
studies, pharmacophore model
Abstract γꢀhydroxybutyric acid (GHB) is a neuroactive substance with specific highꢀaffinity
binding sites. To facilitate target identification and ligand optimization, we herein report a
comprehensive structureꢀaffinity relationship study for novel ligands targeting these binding
sites. A molecular hybridization strategy was used based on the conformationally restricted 3ꢀ
hydroxycyclopentꢀ1ꢀenecarboxylic acid (HOCPCA) and the linear GHB analog, transꢀ4ꢀ
hydroxycrotonic acid (TꢀHCA). In general, all structural modifications performed on HOCPCA
led to reduced affinity. In contrast, introduction of diaromatic substituents into the 4ꢀposition of
TꢀHCA led to highꢀaffinity analogs (medium nanomolar K ) for the GHB highꢀaffinity binding
i
sites as the most highꢀaffinity analogs reported to date. The SAR data formed the basis for a 3ꢀ
dimensional pharmacophore model for GHB ligands, which identified molecular features
important for highꢀaffinity binding, with high predictive validity. These findings will be valuable
in the further processes of both target characterization and ligand identification for the highꢀ
affinity GHB binding sites.
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
Introduction
γꢀHydroxybutyric acid (GHB, Figure 1) is an endogenous neuroactive substance and a
proposed neurotransmitter found in micromolar concentrations in mammalian brains and
metabolically derived from the more potent inhibitory neurotransmitter γꢀaminobutyric acid
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
2
(GABA, Figure 1). GHB is clinically prescribed for treatment of alcohol dependence and
3
narcolepsy. Additionally, GHB is a drug of abuse (Fantasy or liquid ecstasy), producing mild
4
euphoria, sedation, and coma at high dosage. In the central nervous system, GHB displays a
complex neuropharmacology involving both lowꢀ and highꢀaffinity targets. The in vivo
pharmacological effects of GHB are largely mediated by weak agonism of the GABAB
5
receptors. In addition, GHB binds with nanomolar to micromolar affinity to another unknown
target preserved in GABAB(1) knockout mice and with distinct spatial distribution and
6
, 7
ontogenesis in the brain.
This emphasizes the existence of several highꢀaffinity GHB binding sites in distinct proteins,
1
all of which may be therapeutic targets. With the use of photoaffinity labelling, proteomic
analysis, and molecular pharmacology studies, we recently reported the α β δ ionotropic
4
1ꢀ3
8
GABA receptors as highꢀaffinity GHB targets. However, with a ∼40% reduction in the highꢀ
A
affinity GHB binding sites in α knockout mice, the remaining ∼60% of the highꢀaffinity binding
4
8
sites remain elusive. In fact, no direct agonist effect of GHB at extrasynaptic GABA receptors
A
9
could be demonstrated in rat brain slices in relevant regions. Hence, in order to guide the search
for the identity and localization of the highꢀaffinity GHB binding sites, and facilitate future drug
development in the field, more structureꢀaffinity information of ligands targeting these specific
sites is desirable.
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. Chemical structures of GHB, GABA, TꢀHCA, compound 1, HOCPCA, and
compounds 2a–c.
As one of the first GHB analogs reported, the semiꢀrigid transꢀ4ꢀhydroxycrotonic acid (Tꢀ
3
HCA) was found to displace [ H]GHB binding with a small (fourꢀfold) improved affinity
1
0
compared to GHB. In contrast, replacing the double bond with a triple bond to give the more
10
rigid and extended analog of GHB, compound 1, was reported to result in loss of affinity. We
have published two series of highꢀaffine and selective GHB analogs that showed markedly
1
0ꢀ13
improved binding compared to previously reported GHB analogs.
With the conformationally
restricted GHB analog 3ꢀhydroxycyclopentꢀ1ꢀenecarboxylic acid (HOCPCA, Figure 1, K 0.16
i
µM), displaying a 27ꢀfold higher affinity compared to that of GHB (K 4.3 µM), we hypothesized
i
that the structure of HOCPCA closely mimics the bioactive conformation of GHB at the highꢀ
1
3
affinity binding site. Additionally, 4ꢀsubstituted biꢀaromatic GHB analogs, represented by 4ꢀ(4ꢀ
(2ꢀiodobenzyloxy)phenyl)ꢀ4ꢀhydroxybutanoic acid (Figure 1, 2a), displayed a 71ꢀfold improved
affinity compared to GHB, which indicated the presence of a hydrophobic cavity near the GHB
12
scaffold able to accommodate large aromatic substituents. Furthermore, it was shown that the
GHB binding site interacts stereoselectively with substituted GHB analogs, e.g. 2a, as well as the
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
conformational restricted HOCPCA, as the (R)ꢀisomers displayed a 10ꢀ and 13ꢀfold, respectively,
1
2, 13
improved affinity compared to the (S)ꢀisomer.
The high selectivity of HOCPCA for the highꢀ
affinity GHB binding sites was later established, as HOCPCA did not show any affinity for 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
13, 14
GABA receptors and 45 other neurotargets.
Previous structureꢀaffinity relationship (SAR)
B
studies of GHB have shown that introduction of even small substituents into the 2ꢀposition is
highly detrimental for affinity and that only minor alkyl substituents are tolerated in the 3ꢀ
1
1
position of GHB. Conversely, several studies have reported improved binding affinity by
1
0ꢀ12
introduction of substituents in the 4ꢀposition of GHB.
A similar structural trend has been
10, 11
reported for smaller substituents in the corresponding position in TꢀHCA.
In addition to
conformationally constraining the main pharmacophore elements, the scaffold of HOCPCA
allows for introduction of substituents in wellꢀdefined positions suitable for addressing the
reported SAR of GHB. Thus, the high selectivity, conformational restriction, and stereoselective
interaction provides HOCPCA as a unique scaffold for further exploration of the highꢀaffinity
GHB binding sites.
In the present study, we merge the structural features of three classes of highꢀaffinity GHB
ligands and study the conformational determinants of the highꢀaffinity GHB binding sites. The
synthesis and binding affinity determination of this series of compounds are described along with
a pharmacophore model rationalizing the SAR data observed in the present study relative to
previously reported data. A reliable and wellꢀvalidated pharmacophore model has not previously
been published and will be valuable for future progress in the field.
Results
Design strategy
ACS Paragon Plus Environment

Page 7 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 2. General structures of target compounds.
Comparison of HOCPCA and compounds 2a–c by simply overlaying the hydroxyl and
carboxyl functionalities, indicates an area in the binding site in close vicinity to the hydroxyl
group of GHB and HOCPCA, which could be favorable for large hydrophobic and/or aromatic
moieties. To investigate and validate this hypothesis we chose to introduce substituents in the 2ꢀ
or 3ꢀposition of HOCPCA and the corresponding 4ꢀposition of TꢀHCA and compound 1 thereby
probing the suggested hydrophobic area. A set of substituents was chosen inspired by previously
reported 4ꢀsubstituted GHB analogs such as 2a. Due to challenges in the synthesis we were not
able to obtain the direct 3ꢀsubstituted HOCPCA analogs. However, since previous reports
demonstrate that the hydroxyl group of many GHB ligands is favorable, but not vital for
1
5, 16
obtaining highꢀaffinity ligands,
we decided to test if hydrophobic substituents in the 3ꢀ
position of HOCPCA could compensate for the absence of the hydroxyl group. Finally, by
varying the size and flexibility of the cyclic scaffold, the impact of the conformationally
restricted structure of the cyclopentene ring on the binding affinity of HOCPCA was addressed.
Chemistry
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
The HOCPCA and TꢀHCA analogs 1, 4a,b, and 7a were obtained from commercial suppliers,
while the developed HOCPCA and TꢀHCA analogs were prepared according to Schemes 1–4.
The cyclohexane analogs of HOCPCA (cisꢀ4c and transꢀ4c, Scheme 1) were obtained from
compound 10, which was converted into compound 11 by changing the ethyl ester to a benzyl
ester. Diastereomeric separation of racemic mixture into the cisꢀ and transꢀstereoisomers (cisꢀ11
and transꢀ11) followed by catalytic hydrogenation afforded compounds cisꢀ4c and transꢀ4c.
Replacement of the α,βꢀdouble bond of HOCPCA with a cyclopropane ring was performed by
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
17
a SimmonsꢀSmith cyclopropanation using the Furukawa zinc reagent on the alkene 14,
14
prepared as previously described (Scheme 1). Due to coordination between the zinc reagent and
1
8
the allylic alcohol, mainly the cisꢀstereoisomer with respect to the hydroxyl group and
1
9
cyclopropane ring was formed, thus only the cisꢀstereoisomer was isolated. Deprotection with
aqueous hexafluorosilicic acid furnished the cyclopropane analog 3.
a
Scheme 1
ACS Paragon Plus Environment

Page 9 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Reagents and conditions: (a) LiOH, H O/THF, rt., (b) Cs CO , BnBr, DMF, rt., (c) H , 10%
2
2
3
2
t
Pd/C, MeOH, rt., (d) i) 12, BuLi, THF, –78 °C, ii) CO , THF, –50 °C, (e) Et Zn, CH I , CH Cl ,
molecular sieves, 0 °C to rt., (f) H SiF , MeCN, rt., (g) 13, BuLi, R ꢀX, THF, –78 °C, (h) CrO ,
Ac O, AcOH, CH Cl , rt., (i) NaBH , CeCl , MeOH, 0 °C, (j) 3,4ꢀdihydroꢀ2Hꢀpyran, HCl, rt.,
k) Pd(OAc) , N(Bu) Cl, NaHCO , R ꢀI, DMF, 4Å molecular sieves, DMF, 60 °C, (l) Pyridinium
2
2
2 2
2
2
n
2
2
6
3
2
2
2
4
2
3
(
2
4
3
pꢀtoluenesulfonate, MeOH, 50 °C.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
0
Inspired by a procedure reported by Luparia et al., a series of 2ꢀalkyl substituted analogs
20
(
5a–e) were synthesized from compound 13 (Scheme 1). Regioselective iodineꢀlithium
n
exchange performed with BuLi followed by subsequently quenching with an appropriate alkyl
halide provided the 2ꢀsubstituted bromocyclopentenes intermediates 15a–e. Carboxylation of the
t
bromides 15a–e with BuLi and carbon dioxide followed by cleavage of the tertꢀ
butyldimethylsilyl (TBS) group with aqueous hexafluorosilicic acid gave the 2ꢀalkyl substituted
analogs 5a–e.
The 2ꢀaryl substituted analogs of HOCPCA (compound 5f–j) were synthesized via allylic
oxidation of the methyl ester 16. Reduction of the ketone by a Luche reduction and
tetrahydropyranyl (THP) protection of the formed alcohol gave the intermediate 17. The aryl
substituents were introduced via a Heck cross coupling procedure with palladium acetate and
21
tetrabutylammonium chloride to give compounds 18f–j. Subsequently cleavage of the THPꢀ
group and hydrolysis of the methyl ester using pyridinium pꢀtoluenesulfonate and lithium
hydroxide, successively, provided 2ꢀaryl substituted HOCPCA analogs 5f–j.
a
Scheme 2
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
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
3
3
Reagents and conditions: (a) CrO , Ac O, AcOH, CH Cl , rt., (b) R ꢀMgX or R ꢀLi, THF, 0
3
2
2
2
°
2
C, –20 °, or –78 °C, (c) PhMgBr, THF, –78 °C, (d) i) 4ꢀBrꢀBiPh, Mg, I (cat), THF, reflux, ii)
0, THF, –78 °C, (e) NBS, AIBN, CH Cl , reflux, (f) Pd(OAc) , KF, R ꢀB(OH) , 1,4ꢀdioxane,
rt., (g) LiOH, H O/THF, rt.
2
3
2
2
2
2
2
It was initially envisaged that a series of 3ꢀsubstituted analogs of HOCPCA could be
synthesized directly by organometallic addition to ketone 19, obtained from allylic oxidation of
1
4
methyl ester 16 with chromium trioxide as previously reported (Scheme 2). However, all
attempts to perform a nucleophilic addition to the α,βꢀunsaturated ketone 19 with either
organolithium or Grignard reagents (PhLi, PhMgBr or BnMgCl), with or without addition of
hexamethylphosphoramide (HMPA) or 1,3ꢀdimethylꢀ3,4,5,6ꢀtetrahydroꢀ2(1H)ꢀpyrimidinone
(
DMPU), and at varying temperatures (0 °C, –20 °C or –78 °C) resulted in rapid elimination of
the formed alcohol or decomposition of the molecule. Therefore, no further efforts were put into
the synthesis of the 3ꢀsubstituted analogs of HOCPCA. Instead, two saturated 3ꢀarylꢀsubstituted
analogs (6a,b) of HOCPCA were synthesized from the corresponding commercially available
ketone 20. NMR analysis of the crude mixtures revealed a 9:1 ratio between the cisꢀ and transꢀ
stereoisomers, with respect to the hydroxyl and carboxyl groups, of both 6a and 6b. Due to the
minor formation of the transꢀstereoismer, only the cisꢀstereoisomers could be obtained pure after
purification.
ACS Paragon Plus Environment

Page 11 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
To probe the mutual importance of the hydroxyl group and the presence of lipohilic groups in
the 3ꢀposition, the dehydroxylated analogs of HOCPCA were pursued (compounds 7b–e,
Scheme 2). WolfꢀZiegler bromination of the α,βꢀunsaturated carboxylic acid ester 16, followed
by Suzuki cross coupling and hydrolysis of the methyl ester provided the dehydroxylated 3ꢀarylꢀ
substituted series of HOCPCA analogs 7b–e. Due to instability of the allyl bromide 21, the
Suzuki cross coupling was performed immediately upon isolation of compound 21.
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
Scheme 3
a
n
Reagents and conditions: (a) 22, BuLi, appropriate ketone, THF, –78 °C to –40°C, (b) 23,
n
BuLi, appropriate ketone, THF, –78 °C to –40°C, (c) LiOH, H O/THF, rt., (d) RedꢀAl, THF, –
2
78 °C.
The 4ꢀspirocyclicꢀ and 4ꢀarylꢀsubstituted alkynyl and 4ꢀspirocyclicꢀsubstituted TꢀHCA analogs
(8a–c, 8d,e, and 9a–c, respectively) were synthesized as described in Scheme 3 starting from
either ethyl propiolate (22) or propiolic acid (23). Formation of the acetylide of 22 or 23
2
2, 23
followed by subsequently quenching with appropriate ketones
provided compounds 24a–c or
8a–e, respectively, and 24a–c were then converted into 8a–c under basic conditions. The 4ꢀ
spirocyclicꢀsubstituted TꢀHCA analogs 9a–c were obtained from 24a–c by a selective reduction
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
of the alkynes into the transꢀalkenes 25a–c following a procedure described by Meta and
2
4
Koiide using RedꢀAl followed by basic hydrolysis of the ethyl ester.
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
Scheme 4
a
Reagents and conditions: (a) 1,1’ꢀBiPh, PhCH Ph or Ph(CH ) Ph, AlCl , CH Cl , rt., (b) 26
2
2 2
3
2
2
or 27, Glyoxylic acid, NaOH, H O/EtOH, reflux, (d) CeCl , NaBH , MeOH, 0 °C to rt., (e)
2
3
4
n
Chiral HPLC separation: Colomn: ChiralPak IF, mobile phase: heptane/2ꢀPrOH/TFA (95/5/0.1)
The 4ꢀarylꢀsubstituted TꢀHCA analogs 9d–h were synthesized as described in Scheme 4. The
α,βꢀunsaturated ketone intermediates 25d–h were synthesized by either a FriedelꢀCraft acylation
of 1,1’ꢀbiphenyl, diphenylmethane or bibenzyl with maleic acid anhydride (25d–f) following a
2
2
procedure described by Kameo et al. or an aldol condensation between glyoxylic acid and
25
either 1ꢀ(4ꢀ(benzyloxy)phenyl)ethanone (26) or 1ꢀ(4ꢀ((4ꢀbromobenzyl)oxy)phenyl)ethanone
2
5
(
27) (25g and 25h, respectively). A Luche reduction of the ketone intermediates (25d–h)
provided the 4ꢀaryl substituted TꢀHCA analogs 9d–h. The enantiomers 9e, (+)ꢀ9e (98% ee) and
–)ꢀ9e (98% ee), were obtained in excellent ee’s by chiral preparative HPLC using a Chiralpak IF
column.
(
Structure-affinity relationship studies
ACS Paragon Plus Environment

Page 13 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The binding affinities of the compounds for the highꢀaffinity GHB binding sites were
3
measured by displacement of [ H]NCSꢀ382 in rat brain synaptosomal membranes. At the applied
2
6
concentration (16 nM), NCSꢀ382 is selective for the highꢀaffinity GHB binding sites. The
obtained data are summarized in Tables 1–4.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The affinity data for the conformationally restricted cyclic analogs of GHB (compounds 3 and
4
a–c) are presented in Table 1 alongside with data for GHB and HOCPCA. In general, the
reduced planarity and increased structural flexibility compared to HOCPCA led to lower affinity.
Installation of a cyclopropylꢀring as an isostere for the double bond in HOCPCA (3) led to loss
of affinity in concentrations up to 100 µM. Likewise, the cyclobutane analog 4a showed a
markedly reduced binding affinity. A similar trend was shown for the less planar conformations
of GHB represented in the cyclopentane and cyclohexane analogs, 4b and 4c, respectively,
which showed affinity in the midꢀmicromolar range. Interestingly, the cisꢀ and transꢀ
stereoisomers of 4c were shown to be equally efficient in binding to the GHB binding site
indicating some flexibility of the site around the sixꢀmembered ring.
Table 1. Pharmacological data for GHB, HOCPCA, compound 3, and compounds 4a–c: Binding
3
affinities at the highꢀaffinity GHB binding site at rat brain cortical homogenate in the [ H]NCSꢀ
3
82 binding assay.
3
[
H]NCSꢀ382 binding
a
K (µM) [pK ± SEM]
i
i
GHB
3.6 [5.44 ± 0.07]
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
HOCPCA
0.13 [6.89 ± 0.04]
b
(S)ꢀHOCPCA
1.4
b
(
R)ꢀHOCPCA
0.11
3
4
4
>100
a
70 [4.16 ± 0.04]
20 [4.71± 0.05]
35 [4.45 ± 0.05]
28 [4.55 ± 0.01]
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
b
cisꢀ4c
transꢀ4c
a
IC values were calculated from inhibition curves and converted to K values. Data are given as
50
i
b
mean [mean pK ± SEM] of three to four independent experiments. From Wellendorph et al.,
i
1
3
2005.
Introduction of substituents in the 2ꢀposition of HOCPCA was in general detrimental for
binding affinity (Table 2). Unbranched alkyl groups, such as an ethyl or allyl led to loss of
affinity at concentrations up to 300 µM, whereas larger aromatic substituents in the 2ꢀposition of
HOCPCA showed binding affinities in the midꢀmicromolar range.
Table 2. Pharmacological data for GHB, HOCPCA, and compounds 5a–j: Binding affinities at
3
the highꢀaffinity GHB binding site at rat brain cortical homogenate in the [ H]NCSꢀ382 binding
assay.
3
[
H]NCSꢀ382 binding
2
R
a
K (µM) [pK ± SEM]
i
i
GHB
3.6 [5.44 ± 0.07]
0.13 [6.89 ± 0.04]
HOCPCA
5a
Me
Et
80 [4.10 ± 0.02]
>300
5
b
5c
Allyl
Bn
>300
5
d
38 [4.43 ± 0.03]
ACS Paragon Plus Environment

Page 15 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
5
e
3ꢀMeOꢀBn
Ph
46 [4.43 ± 0.02]
48 [4.32 ± 0.02]
70 [4.16 ± 0.04]
157 [3.81 ± 0.03]
>300
5f
5
g
4ꢀClꢀPh
4ꢀMeꢀPh
4ꢀMeOꢀPh
3,4ꢀClꢀPh
5h
5
i
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
5j
62 [4.21 ± 0.02]
a
IC values were calculated from inhibition curves and converted to K values. Data are given as
5
0
i
mean [mean pK ± SEM] of three to four independent experiments.
i
The binding affinities for the 3ꢀsubstituted analogs (6a,b and 7a–e) are presented in Table 3.
As indicated previously, absence of the hydroxyl group is detrimental for affinity as shown by a
~200ꢀfold affinity decrease of 7a versus HOCPCA. The addition of a phenyl ring (7b) in the 3ꢀ
position did not compensate for the removal of the hydroxyl group with only an approximately
twoꢀfold increased affinity relative to 7a. Further introduction of a hydroxyl group in the 2ꢀ
position of the phenyl ring (7c) led to a slight enhancement in affinity as compared to 7b.
Introduction of an additional 4ꢀphenyl (7d) or 4ꢀbenzyloxy ring (7e) displayed a fiveꢀ and tenꢀ
fold increase, respectively, in binding affinities compared to that of 7b. A similar trend was
observed for the saturated 3ꢀsubstituted analogs (6a,b), with the phenyl substituted analog (6a)
displaying an approximately threeꢀfold decrease in binding affinity relative to 4b. Interestingly,
an additional 4ꢀphenyl ring (6b) led to an affinity increase by a factor of 100 compared to that of
6
a, thus displaying a binding affinity only threeꢀfold lower than that of HOCPCA.
Table 3. Pharmacological data for GHB, HOCPCA, and compounds 6a,b and 7a–e: Binding
3
affinities at the highꢀaffinity GHB binding site at rat brain cortical homogenate in the [ H]NCSꢀ
3
82 binding assay.
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
HO
R
3
R3
COOH
COOH
6a,b
7a–e
3
[
H]NCSꢀ382 binding
R3
a
K (µM) [pK ± SEM]
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
GHB
3.6 [5.44 ± 0.07]
0.13 [6.89 ± 0.04]
53 [4.28 ± 0.01]
0.51 [6.29 ± 0.10]
31 [4.51 ± 0.04]
16 [4.80 ± 0.10]
4.6 [5.34 ± 0.03]
3.2 [5.49 ± 0.10]
1.6 [5.77 ± 0.05]
HOCPCA
6a
Ph
4ꢀbiPh
H
6
b
7a
7
b
Ph
7c
2ꢀOHꢀPh
4ꢀbiPh
4ꢀBnOPh
7
d
7e
a
IC values were calculated from inhibition curves and converted to K values. Data are given
as mean [mean pK ± SEM] of three to four independent experiments.
5
0
i
i
1
0
In contrast to previous reports, we found that the triple bond analog of GHB, 1, displayed
affinity for the highꢀaffinity GHB binding sites equivalent to GHB (Table 4). Only slight or no
enhancement in affinity was achieved by introduction of spirocylic aliphatic substituents in the
4
ꢀposition of 1 (8a–c). A similar effect was observed for the corresponding analogs of TꢀHCA
(
9a–c), the double bond analog of GHB. Likewise, introduction of larger aromatic substituents in
the 4ꢀposition in 1 did not affect the binding affinity (8d,e). However, introduction of larger
aromatic substituents in the 4ꢀposition of TꢀHCA provided a series of very highꢀaffinity TꢀHCA
analogs (9d–h) that were able to displace [3H]NCSꢀ382 binding with K values up till six times
i
1
2
lower than that of the corresponding GHB analogs in the study and previously reported. The 4ꢀ
bromobenxyloxyphenyl substituted TꢀHCA analog 9h, together with the previously reported (R)ꢀ
2
a, thus represents the most highꢀaffinity ligand reported to date (Figure 3). Interestingly, a 36ꢀ
fold preference for the (+)ꢀisomer of the 4ꢀbenzylphenyl TꢀHCA analog 9e was shown, which is
a markedly higher stereochemical preference than the tenꢀfold difference reported for the
ACS Paragon Plus Environment

Page 17 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
12
stereoisomers of 2ꢀiodobenzyloxyphenyl substituted analog 2a of GHB. Functional studies on
the compounds are awaiting.
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. Pharmacological data for GHB, HOCPCA, TꢀHCA, and compounds 1, 2a–c, 8a–e, and
9a–h: Binding affinities at the highꢀaffinity GHB binding site at rat brain cortical homogenate in
3
the [ H]NCSꢀ382 binding assay.
3
R1
R2 [ H]NCSꢀ382 binding
a
K (µM) [pK ± SEM]
i
i
GHB
3.6 [5.44 ± 0.07]
HOCPCA
TꢀHCA
1
0.13 [6.89 ± 0.04]
b
H
H
H
H
H
H
H
1.1
H
4.3 [5.50 ± 0.04]
c
(
(
S)ꢀ2a
R)ꢀ2a
2ꢀIꢀBnOPh
2ꢀIꢀBnOPh
4ꢀBnOPh
4ꢀBrꢀBnOPh
0.22
c
0.022
c
2
2
8
8
8
8
8
9
9
9
9
9
b
c
0.11
c
0.034
a
b
c
–(CH
–(CH
–(CH
Bn
2
2
2
)
)
)
3
4
5
–
–
–
1.6 [5.78 ± 0.04]
2.7 [5.57 ± 0.05]
7.7 [5.12 ± 0.06]
3.5 [5.45 ± 0.04]
1.6 [5.80 ±0.03]
d
e
H
H
4ꢀBnOPh
a
b
c
–(CH
–(CH
2
)
)
)
3
4
5
–
–
–
3.5 [5.47 ± 0.06]
3.1 [5.51 ± 0.03]
4.3 [5.37 ± 0.01]
0.075 [7.14 ± 0.08]
0.042 [7.38 ± 0.02]
0.030 [7.47 ± 0.06]
1.16 [5.97 ± 0.02]
0.042 [7.39 ± 0.02]
0.043 [7.43 ± 0.03]
2
2
–(CH
d
e
4ꢀbiPh
4ꢀBnPh
4ꢀBnPh
4ꢀBnPh
H
H
H
H
H
H
(
(
+)ꢀ9e
–)ꢀ9e
9
9
f
4ꢀPh(CH
) Ph
2 2
g
4ꢀBnOPh
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
9
h
a
4ꢀBrꢀBnOPh
H
0.023 [7.64 ± 0.08]
IC values were calculated from inhibition curves and converted to K values. Data are given
5
0
i
b
as mean [mean pK ± SEM] of three to four independent experiments. From Wellendorph et al.,
i
13 c
12
2
005. From Høg et al., 2008.
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
Figure 3. Concentrationꢀdependent inhibition of [ H]NCSꢀ382 binding to rat brain cortical
homogenate by TꢀHCA analogs (+)ꢀ9e, (–)ꢀ9e, and 9h. The affinities were determined as
described under Experimental Section. Results are given as mean ± SD of a single representative
experiment performed in triplicate. Two additional experiments gave similar results (data
summarized in Table 4).
Pharmacophore modeling
To improve understanding of the SAR of the developed GHB ligands we constructed a
pharmacophore model, which will aid in future design of new ligands targeting the highꢀaffinity
GHB binding site. The pharmacophore model (Figure 4) was based on HOCPCA and
compounds 2b, 3, 5b, 5i, 6a, 8e, 9d, and 9e divided into high affinity ligands (pK > 6),
i
intermediate binders (4 < pK < 6), and compounds with weak or no affinity (pK < 4).
i
i
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
Figure 4. Pharmacophore model for the GHB highꢀaffinity binding site displaying the proposed
binding mode of compound (R)ꢀ9e (cyan carbon atoms). The yellow spheres indicate
unfavorable positions of substituents (exclusion volumes) in the highꢀaffinity binding sites while
the other spheres indicate positions favorable for: aromatic ring systems (purple), hydrogen bond
acceptor (orange), hydrogen bond donor (blue), and negatively charged group (red). The arrows
show the direction of hydrogen bond acceptor lone pairs and donor hydrogen.
The best scoring pharmacophore model, comprised of the hydroxyl group as both a hydrogen
bond acceptor and donor, the negatively charged carboxylic acid and an aromatic ring, was
refined by adding exclusion volumes based on the inactive compounds 3 and 5b. Requiring a
match of three out of the four pharmacophore elements identified both enantiomers of all ten
highꢀaffinity ligands, 18 out of the 27 intermediate affinity compounds, and discarded both
enantiomers of the five inactive compounds (see Supplementary Table 1 for the complete
performance of the pharmacophore model). The model ranked the compounds with four
matching pharmacophore elements as “best fits” and thus, qualitatively matching the fact that
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
compounds with the distal aromatic moiety in the 4ꢀposition of GHB display the highest affinity.
All 2ꢀsubstituted HOCPCA analogs (5a–j), with the exception of 5a having the smallest 2ꢀ
substituent, were discarded due to the exclusion volumes based on 5b. HOCPCA analogs 7a–e,
which lack the hydroxyl group, were only recognized if asking for two out of four matching
pharmacophore elements (not shown). With the exception of transꢀ4a, (S)ꢀcisꢀ4c and (S)ꢀ5a, the
model recognized both stereoisomer (all four for 4b) of the identified high and intermediate
affinity compounds and, with the exception of transꢀ4c, 7c, 8d, 8e, 9h and 9f, the (R)ꢀenantiomer
were ranked higher than the (S)ꢀenantiomer. For 8d,e the (S)ꢀenantiomer corresponds to the (R)ꢀ
enantiomer of the other analogs due to the adjacent triple bond. These observations are consistent
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
2, 13
with the experimental data for HOCPCA, 2a and 9e.
Discussion
Despite notable interest in the field especially to the enigmatic GHB highꢀaffinity binding sites
and their possible implication in neurological disorders, the physiological function of the highꢀ
1
affinity GHB binding sites distinct from the GABA receptors are still elusive. Most of the
B
observed pharmacological effects of GHB have been demonstrated to involve GABAB
6
, 27, 28
receptors,
yet several GHB ligands that show no affinity for the GABA receptors display
B
2
9ꢀ31
pharmacological effects in vitro, putatively mediated by the highꢀaffinity GHB binding sites.
Interestingly, in the clinic, GHB is preferably used over the GABA agonist baclofen in both
B
2, 3
narcolepsy and alcoholism. To achieve a more detailed understanding of the molecular basis
for GHB actions and to provide potential future drug candidates, new selective ligands for the
specific GHB highꢀaffinity binding sites with a broad structural diversity are needed.
ACS Paragon Plus Environment

Page 21 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
In the absence of knowledge about the molecular nature of the highꢀaffinity GHB binding site,
most molecular research in the field has been focused on developing tool compounds, primarily
based on the structure of GHB. However, since GHB is a small and flexible molecule it is not
ideal for establishing a binding mode and thereby providing more precise information on the
topography of the binding site. In the present study, we have expanded the investigation of the
SAR for ligands targeting the highꢀaffinity GHB binding sites and comprised a pharmacophore
model highlighting the molecular features important for high affinity binding. We have been
using HOCPCA and the conformationally less flexible linear double or triple bond containing
GHB analogs, TꢀHCA, and 1 as lead structures in combination with structural fragments
previously described for a series of linear GHB analogs displaying high affinity.
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
Being a small low molecular weight compound displaying high binding efficiency, HOCPCA
constitutes an ideal lead structure for further development. The replacement of the cyclopentene
ring of HOCPCA with saturated cycloalkanes (4a–c) was detrimental for the binding affinity
supporting the hypothesis that the cyclopentene ring of HOCPCA represents the bioactive
conformation of GHB (Figure 5A). The pharmacophore model confirms that the relative
orientation/location of the carboxylic acid and hydroxyl group of HOCPCA is important for
binding and explains the favoring of the almost planar cyclopentene ring of HOCPCA with disꢀ
allowed areas. The poor tolerance of outꢀofꢀplane bulk is best exemplified by the lack of binding
for 3 (K >100 µM) but is likely also part of the explanation for the 175ꢀfold lower affinity of
i
transꢀ4c relative to HOCPCA (Figure 5B). Additionally, saturation of the cyclopentene ring
8
increase the entropic penalty of binding with the conformational flexibility. The reverse
structural modification is reported for GHB and the corresponding unsaturated analog TꢀHCA, a
32
naturally occurring compound in the human brain that binds to the highꢀaffinity GHB site,
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
increasing the planarity and decreasing the flexibility while retaining a perfect fit to the
pharmacophore model (Figure 5A). Interestingly, replacement of the double bond in TꢀHCA by a
triple bond to give 1, representing a rigid and slightly extended conformation of GHB, was
shown to display an affinity similar to GHB itself. Altogether this reflects limited steric and
directional tolerability for the highꢀaffinity GHB binding.
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
Figure 5. A. The pharmacophore model predicts an almost perfect overlay of the functional
groups of the conformationally restricted TꢀHCA (purple stick) and (R)ꢀHOCPCA (green stick)
with the believed active conformation of the flexible GHB (black stick). B. Overlay of (R)ꢀ
HOCPCA (green stick) and (1R,3R)ꢀtransꢀ4c (violetꢀpurple stick) from the pharmacophore
model showing the poor tolerance for ligands deviating from the nearꢀplanar conformation of
HOCPCA.
1
1
Bourguignon et al., have demonstrated that no substitutions are allowed in the 2ꢀposition of
GHB and only small substituents are allowed in the 3ꢀposition. The pharmacophore model
presented here shows that substituents in the 2ꢀposition of GHB, like it is the case for 4, will
violate the steric requirements for planarity close to the carboxylic acid. Additionally, we find
that the 2ꢀposition of HOCPCA is analogous to the 3ꢀposition in GHB (Figure 5) and we have
demonstrated a comparable trend, where alkyl and aromatic substituents led to significant
ACS Paragon Plus Environment

Page 23 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
decrease or absence of affinity in general (Table 2) represented by disꢀallowed areas in the
pharmacophore model.
The most highꢀaffinity GHB analogs reported so far, including HOCPCA, all possess a
hydroxyl group in the equivalent position and we prove the importance of this functional group
by the 193ꢀfold drop in affinity for 7a relative to HOCPCA. There is no clear evidence from the
included or previously reported compounds to whether the hydroxyl group acts as a hydrogen
bond donor, an acceptor or both. In our pharmacophore model, the dual role, as both hydrogen
bond donor and acceptor, has been included with good performance, indicating the presence of
suitable positioned interaction partners for both in the binding pocket. Another structural
characteristic for highꢀaffinity GHB ligands is an aromatic substituent in a location equivalent to
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
2
the 4ꢀposition of GHB. With the aim of investigating this in an equivalent position of 2a–c, we
synthesized a series of compounds addressing the aromatic feature.
For both the saturated and unsaturated analog of HOCPCA, 4b and 7a, respectively, (Table 3)
a single aromatic substituent in the 3ꢀposition, as for 6a and 7b, is tolerated with a slight drop in
affinity (twoꢀ to threeꢀfold) compared to the parent compounds. However, the affinity improved
by adding an additional aromatic moiety (Table 3) and especially the 100ꢀfold affinity increase
of 6b versus 6a proves a very favorable additional interaction between the distal aromatic moiety
and the highꢀaffinity binding site. This corresponds to what has been previously reported for the
1
2
corresponding GHB analogs, where the 4ꢀbiphenylꢀGHB analog shows equivalent affinity to
that of 6b, indicating an area near the 4ꢀposition of GHB and the equivalent position of
HOCPCA that is favorable for aromatic moieties. Furthermore, analogs of GHB, TꢀHCA, and
less so compound 1 (Table 4) display a similar effect upon introduction of aromatic 4ꢀsustituents
of varying length and flexibility (e.g. 9e–g). As for compounds 6a,b and 7a–e, the
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
pharmacophore model shows that the first aromatic ring of the 4ꢀsubstituent in 9e–g is not fitting
into a well defined pocket. In contrast, the model suggests the distal aromatic ring to be
positioned in a very well defined position (Figure 6, purple sphere) where favorable interactions
between the distal aromatic ring and the highꢀaffinity bindings can take place.
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
Figure 6. The proposed binding modes from the pharmacophore model of compounds (R)ꢀ9e
(cyan stick), (R)ꢀ9g (dark green stick), (R)ꢀ9f (bright orange stick), and (R)ꢀ6b (dark blue stick).
The spheres indicate positions favorable for: aromatic ring systems (purple), hydrogen bond
acceptor (orange), hydrogen bond donor (blue), and negative charged group (red). The arrows
show the direction of hydrogen bond acceptor lone pairs and donor hydrogen. The
pharmacophore model shows a wellꢀdefined pocket for the distal phenyl ring system while the
first phenyl ring is unimportant for the binding affinity of the developed GHB and HOCPCA
analogs.
Introducing substituents in the 4ꢀposition of GHB, TꢀHCA, and compound 1 introduces
chirality with a 39ꢀfold difference in affinity between the enantiomers of 9e (Table 4). In
contrast, the enantiomers of HOCPCA only display a 13ꢀfold difference in affinity with (R)ꢀ
ACS Paragon Plus Environment

Page 25 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
HOCPCA having the highest affinity (Table 1, K 0.11 µM). Interestingly, our pharmacophore
i
model can accommodate both enantiomers of HOCPCA and 9e (Figure 7) and, in accordance
with the affinity, ranks (R)ꢀHOCPCA higher than (S)ꢀHOCPCA. Additionally, the (R)ꢀ
enantiomer of 9e and similar ligands were ranked higher than the (S)ꢀenantiomer, i.e. predicted
as the higher affinity configuration.
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. The pharmacophore model generally identifies both enantiomers of the highꢀaffinity
ligands and intermediate binders as hits illustrated by the overlays of, A. (R)ꢀ9e (cyan stick) and
(S)ꢀ9e (yellow stick) and B. (R)ꢀHOCPCA (green stick) and (S)ꢀ9e (skyblue stick).
In conclusion, by use of a molecular hybridization strategy based on structural scaffolds and
fragments from three classes of potent and selective ligands for the highꢀaffinity GHB binding
sites, we have generated a series of ligands exploring the core scaffold and substitutions in the 2ꢀ
and 3ꢀposition of HOCPCA as well as the 4ꢀposition of GHB, TꢀHCA, and compound 1.
Importantly, these compounds guided the development of the first reported 3Dꢀpharmacophore
model for the highꢀaffinity GHB binding site. This model identified the acid, the hydroxyl, and
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
the distal aromatic moiety in the 4ꢀposition of GHB as the structural elements important for
affinity, and which can rationalize the obtained SAR. In general, all structural modifications
performed in this study for the HOCPCA core scaffold led to reduced affinity indicating that for
this specific structural scaffold, HOCPCA represents the optimal conformation and fit for the
specific binding sites. In contrast, using TꢀHCA as a core scaffold led to some of the most potent
ligands for the GHB highꢀaffinity binding sites reported to date. However, some general trends
for limited steric and directional tolerability of the scaffold, accommodation of larger aromatic
substituents in the 4ꢀposition of GHB and conformational preference were exposed. In the
absence of knowledge about the binding site and thus the structural details of the binding sites in
the form of a model of the protein, further insights into the structural determinants for binding
are valuable. The pharmacophore model presented here, will aid in the continued efforts to
design and synthesize even better GHB ligands for the molecular dissection and potentially
future drug development.
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
Experimental Section
Chemistry
General procedures: Compounds 1, 3a,b, and 7a were obtained from the commercial
suppliers: 1, Combi Blocks, Inc. (USA); 3a, Combi Blocks, Inc. (USA); 3b, Enamine Ltd.
1
4
20
14
20
33
23
(Latvia); 7a, Combi Blocks, Inc. (USA). Compounds 12, 13, 14, 15a,c, 19, 24b, 25d–
22
25
25
23
f, 26, and 27 were synthesized as described in the literature, while compounds 24a,c and
2
4
2
5c were synthesized after modified literature procedures. The spectroscopic data for
34
35
36
compounds 24a, 24c, and 25c is in agreement with data previously reported in the literature.
All reagents and solvents (reagent or chromatography grade) were obtained from commercial
ACS Paragon Plus Environment

Page 27 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
suppliers and used without further purification. Airꢀ and/or moistureꢀsensitive reactions were
performed under a nitrogen or argon atmosphere using syringeꢀseptum cap techniques and with
the use of flame or oven dried glassware. Anhydrous solvents were obtained by using a solvent
purification system (THF, CH Cl ) or by storage over 3 or 4 Å molecular sieves. Thin layer
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
2
chromatography (TLC) was carried out using Merck silica gel 60 F254 plates, and compounds
were visualized using UV (254 and 366 nm) and KMnO , Ninhydrin or Anisaldehyde spray
4
reagent. Flash Chromatography was carried out according to standard procedures using Merck
silica gel 60 (0.040–0.063 mm). Melting points were recorded on a SRS OptiMelt apparatus in
1
13
open capillary tubes and are uncorrected. H and C NMR data were recorded on a Bruker
1
19
Avance 400 MHz spectrometer equipped with a 5 mm PABBO BB( H, F) ZꢀGRD probe or a
Bruker Avance 600 MHz spectrometer equipped with a cryogenically cooled 5 mm CPDCH
13
1
C( H) ZꢀGRD probe, at 300 K. Data are tabulated in the following order: chemical shift (δ)
multiplicity (b, broad; s, singlet; d, doublet; t, triplet; q, quartet; m, multiplet), coupling
[
constant(s) J (Hz), number of protons]. The solvent residual peak was used as internal
37
reference. Analytical High performance liquid chromatography (HPLC) was performed on a
MerckꢀHitachi HPLC system consisting of an Lꢀ7100 pump, an Lꢀ7200 autosampler, and an Lꢀ
7400 UV detector (210 or 254 nm), using a Chromolith SpeedROD RPꢀ18 column (4.6 × 50
mm). A linear gradient elution was performed with eluent A (H O/TFA 100:0.1) containing 0%
2
of solvent B (MeCN/H O/TFA, 90:10:0.01) rising to 100% of B during 5 minutes with a flow
2
ꢀ1
rate of 4.0 mLmin . Data were acquired and processed using the EZChrom Elite Software
version 3.1.7 by Hitachi. The purity of the analyzed compounds are ≥95%, unless otherwise
stated. Determination of enantiomeric excess (ee) and purity for compounds (+)ꢀ9e and (–)ꢀ9e
were performed using an Ultimate 3000 HPLC system with a PGꢀ3200 pump, a Rheodyne
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
injector 9725i, a 10 mL loop, and an MWDꢀ3000SD detector (254 nm) using a Chiralpak IF (5
µm, 4.6 × 250 mm) column and a mobile phase consisting of nꢀheptane/2ꢀPrOH/TFA (95:5:0.1)
ꢀ1
at a flow rate of 1 mLmin . For HPLC control, data collection, and data handling Chromeleon
software ver. 6.80 was used. The ee and purity of the analyzed compounds are ≥95%, unless
otherwise stated. Preparative reversedꢀphase HPLC was carried out on an Ultimate 3000 HPLC
system (Dionex) with a multiꢀwavelength UV detector (210 or 254 nm) and 10 mL loop using a
preparative RP Phenomenex Gemini NXꢀC column (5 mm, 21.2 × 250 mm) using eluent A
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
8
ꢀ1
(H O/TFA, 100:0.1) and eluent B (MeCN/H O/TFA, 90:10:0.1) at a flow rate of 20 mLmin For
2
2
HPLC control, data collection, and data handling, Chromeleon Software ver. 6.80 was used.
Chiral preparative HPLC was carried out on an Ultimate 3000 HPLC system with a PGꢀ3200
pump, a Rheodyne injector 9725i, a 10 mL loop, and an MWDꢀ3000SD detector (254 nm) using
a Chiralpak IF (5 µm, 10 × 250 mm) column and a mobile phase consisting of nꢀheptane/2ꢀ
ꢀ
1
PrOH/TFA (95:5:0.1) at a flow rate of 6 mLmin . For HPLC control, data collection, and data
handling Chromeleon software ver. 6.80 was used. Optical rotation was recorded on an Anton
Paar Modular Circular Polarimeter MCP300 with the temperature set to 25 °C. The calculated
specific optical rotations were the average of three separate readings. Elemental analyses were
performed by Mr. J. Theiner, Department of Physical Chemistry, University of Vienna, Austria,
and are within ±0.4% of the calculated values, unless otherwise stated.
cis-4ꢀhydroxybicyclo[3.1.0]hexaneꢀ1ꢀcarboxylic acid (3). Under a nitrogen atmosphere, a
mixture of 3Å molecular sieves and Et Zn (1.1M in toluene, 24.7 mL, 27.2 mmol) in anhydrous
2
CH Cl (20 mL) at 0 °C was added CH I (2.19 mL, 27.2 mmol) drop wise. The mixture was
2
2
2 2
1
4
stirred for 10 min before a solution of compound 14 (1.1 g, 4.5 mmol) in anhydrous CH Cl (40
2
2
ACS Paragon Plus Environment

Page 29 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
mL) with 3 Å molecular sieves was added rapidly at 0 °C. The reaction mixture was allowed to
reach rt. and stirred for 20 h. H O (50 mL) and brine (20 mL) were added and the mixture was
2
acidified with HCl (aq. 1M) before it was extracted with EtOAc (4 × 100 mL). The combined
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
organic phases were dried over anhydrous MgSO , filtered, and evaporated in vacuo. Purification
4
by column chromatography (Heptane/EtOAc 3:1 + 1% AcOH) afforded 4ꢀ((tertꢀ
butyldimethylsilyl)oxy)bicyclo[3.1.0]hexaneꢀ1ꢀcarboxylic acid (295 g, 25%) as white solid: mp
1
9
7 °C. H NMR (400 MHz, CDCl ): δ 4.56–4.48 (m, 1H), 2.15–2.01 (m, 2H), 1.92–1.83 (m,
3
2H), 1.42–1.36 (m, 1H), 1.31–1.26 (m, 1H), 1.26–1.17 (m, 1H), 0.90 (s, 9H), 0.09 (s, 3H), 0.07
13
(
s, 3H). C NMR (150 MHz, CDCl ): δ 180.7, 72.8, 34.4, 29.7, 29.4, 25.9, 24.3, 18.2, 14.7, –
3
4.6, –4.8. In a polypropylene vial, 4ꢀ((tertꢀbutyldimethylsilyl)oxy)bicyclo[3.1.0]hexaneꢀ1ꢀ
carboxylic acid (286 mg, 1.1 mmol) was dissolved in MeCN (16 mL) and aq. H SiF (20–25%
2
6
w/w, 0.66 mL) was added. The solution was stirred for 1 h at rt. before H O (20 mL) was added
2
and the aqueous phase was extracted with Et O (3 × 20 mL) and EtOAc (2 × 20 mL). The
2
combined organic phases were dried over anhydrous MgSO , filtered, and evaporated.
4
Purification by column chromatography (Heptane/EtOAc 1:1 + 1% AcOH) afforded 3 (112 g,
1
71%) as clear colorless oil. H NMR (400 MHz, DMSOꢀd ): δ 11.99 (s, 1H), 4.61 (s, 1H), 4.35–
6
13
4
.27 (m, 1H), 2.01–1.90 (m, 1H), 1.82–1.66 (m, 3H), 1.12–0.98 (m, 3H). C NMR (150 MHz,
DMSOꢀd ): δ 175.2, 70.8, 32.5, 29.0, 28.7, 24.8, 13.0. The free acid of 3 (112 mg, 0.8 mmol)
6
was converted into the sodium salt by dissolving in EtOH (5 mL) and addition of NaOH (aq.
0.4935M, 1.598 mL, 0.8 mmol). The mixture was stirred for 15 min at rt. before it was
evaporated in vacuo to afford the sodium salt of 3 (126 mg, 97%) as white solid: mp > 240 °C.
1
H NMR (600 MHz, CD OD): δ 4.47–4.40 (m, 1H), 2.15–2.06 (m, 1H), 1.90–1.78 (m, 3H),
3
ACS Paragon Plus Environment