Synthesis and pharmacological evaluation of 6-aminonicotinic acid analogues as novel GABAA receptor agonists

Article abstract of DOI:10.1016/j.ejmech.2014.07.039

A series of 6-aminonicotinic acid analogues have been synthesized and pharmacologically characterized at native and selected recombinant GABA _A receptors. 6-Aminonicotinic acid (3) as well as 2- and 4-alkylated analogues (9-11, 14-16) display low to mid-micromolar GABA_AR binding affinities to native GABA_A receptors (K_i 1.1-24 μM). The tetrahydropyridine analogue of 3 (22) shows low-nanomolar affinity (K _i 0.044 μM) and equipotency as an agonist to GABA itself as well as the standard GABA_A agonist isoguvacine. Cavities surrounding the core of the GABA binding pocket were predicted by molecular interaction field calculations and docking studies in a α₁β ₂γ₂ GABA_A receptor homology model, and were confirmed by affinities of substituted analogues of 3. The tight steric requirements observed for the remarkably few GABA_AR agonists reported to date is challenged by our findings. New openings for agonist design are proposed which potentially could facilitate the exploration of different pharmacological profiles within the GABA_AR area.

Full text of DOI:10.1016/j.ejmech.2014.07.039

European Journal of Medicinal Chemistry 84 (2014) 404e416
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Original article
Synthesis and pharmacological evaluation of 6-aminonicotinic acid
analogues as novel GABAA receptor agonists
Jette G. Petersen a, Troels Sørensen b, Maria Damgaard a, Birgitte Nielsen a,
,
*
Anders A. Jensen a, Thomas Balle b, Rikke Bergmann a, Bente Frølund a
a Department of Drug Design and Pharmacology, Faculty of Health and Medical Sciences, University of Copenhagen, Universitetsparken 2,
DK 2100 Copenhagen, Denmark
b Faculty of Pharmacy, The University of Sydney, Camperdown Campus, 2006 NSW, Australia
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of 6-aminonicotinic acid analogues have been synthesized and pharmacologically characterized
at native and selected recombinant GABAA receptors. 6-Aminonicotinic acid (3) as well as 2- and 4-
alkylated analogues (9e11, 14e16) display low to mid-micromolar GABAAR binding affinities to native
Received 1 May 2014
Received in revised form
10 July 2014
Accepted 11 July 2014
Available online 11 July 2014
GABAA receptors (Ki 1.1e24
mM). The tetrahydropyridine analogue of 3 (22) shows low-nanomolar af-
finity (Ki 0.044 M) and equipotency as an agonist to GABA itself as well as the standard GABAA agonist
m
isoguvacine. Cavities surrounding the core of the GABA binding pocket were predicted by molecular
interaction field calculations and docking studies in a a1b2g2 GABAA receptor homology model, and were
confirmed by affinities of substituted analogues of 3. The tight steric requirements observed for the
remarkably few GABAAR agonists reported to date is challenged by our findings. New openings for
agonist design are proposed which potentially could facilitate the exploration of different pharmaco-
logical profiles within the GABAAR area.
Keywords:
GABA-A receptor agonists
Structureeactivity studies
2-Aminonicotinic acid analogues
© 2014 Elsevier Masson SAS. All rights reserved.
1. Introduction
glycine receptors. The assembled receptor complex is a circular
arrangement of five subunits making up a chloride selective ion-
g
-Aminobutyric acid (GABA) is the major inhibitory neuro-
conducting channel. 19 different human GABAA subunits have
been identified; a1e6 b1e3 g1e3 , ε, , and r1e3, and these
subunits combine in different stoichiometries, the most com-
mon ones being 2 -2 -1 heteropentamers [3]. From the at
least 26 native GABAA receptors subtypes proposed, the pre-
dominant combinations are believed to be a1b2g2 a2b3g2, and
a3b3g2 [4]. GABAARs composed of subunits, also designated
GABACRs due to their distinct physiological and pharmacological
properties, assemble as homopentameric or pseudohomomeric
receptors [5].
Typical orthosteric GABAAR agonists contain both basic and
acidic functionalities positioned in a narrow distance range from
each other and are in general relatively small molecules [6]. These
GABAA agonists include the highly potent natural product com-
pound muscimol [7,8], 4,5,6,7-tetrahydroisoxazol [5,4-c]pyridin-3-
ol (THIP) [9], 1,2,3,6-tetrahydropyridine-4-carboxylic acid (iso-
guvacine), piperidine-4-carboxylic acid (isonipecotic acid) [8], and
5-(4-piperidyl)-3-isoxazolol (4-PIOL) [10,11] (Fig. 1). In general,
studies have shown that the introduction of even small sub-
stituents onto a GABAA agonist, such as muscimol and THIP, is
transmitter in the central nervous system (CNS) where it exerts its
effects through ionotropic (GABAA) receptors resulting in fast syn-
aptic inhibition and metabotropic (GABAB) receptors resulting in
slow, prolonged inhibitory signals. The GABA neurotransmitter
system is essential for the overall balance between neuronal exci-
tation and inhibition and known to be implicated in many disorders
such as anxiety, epilepsy, mood and cognitive disorders, insomnia,
and schizophrenia [1,2].
,
,
,
d
q, p
a
b
g
,
r
The GABAA receptors (GABAARs) belong to the Cys-loop su-
perfamily of ligand-gated ion channels also comprising the
nicotinic acetylcholine receptors, 5-HT3 serotonin receptors, and
Abbreviations: 4-PIOL, 5-(4-piperidyl)-3-isoxazol; DCVC, dry column vacuum
chromatography; ELIC, Erwinia ligand-gated ion channel; FC, flash chromatography;
FLIPR, fluorescent imaging plate reader; FMP, FLIPR membrane potential; GABA, g-
aminobutyric acid; GLIC, Gloeobacter ligand-gated ion channel; GluCl, glutamate-
gated ion channel; IFD, induced fit docking; RP-FC, reverse phase flash chroma-
tography; THIP, 4,5,6,7-tetrahydroisoxazol[5,4-c]pyridine-3-ol.
* Corresponding author.
E-mail addresses: bfr@sund.ku.dk, bfr@farma.ku.dk (B. Frølund).
http://dx.doi.org/10.1016/j.ejmech.2014.07.039
0223-5234/© 2014 Elsevier Masson SAS. All rights reserved.

J.G. Petersen et al. / European Journal of Medicinal Chemistry 84 (2014) 404e416
405
report on the synthesis and pharmacological characterization at the
GABAAR of a series of compounds containing 2-aminopyridine or
analogous amidine/guanidine ring systems as novel amine
bioisosteres.
2. Results
2.1. Chemistry
The series of 2-aminopyridines 1e8 with carboxylic acid sub-
stituents systematically placed around the pyridine ring were all
commercially available except for 2-(2-aminopyridin-4-yl)acetic
acid (6) which was synthesized from 4-methyl-2-aminopyridine
(Scheme 1). tert-Butoxycarbonyl protection of the amine affording
28 was followed by lithiation of the methyl group using lithium
diisopropylamide (LDA) and subsequent quenching with dieth-
ylcarbonate to give the ethyl ester 29. Deprotection under basic
conditions afforded the target compound 6.
The synthetic procedures for the 6-aminonicotinic acid ana-
logues (9e21) are depicted in Schemes 2e4. The methyl substituted
analogues 9, 14, 19 (Scheme 2) were prepared from methyl-3-
bromo-6-aminopyridines 30e32 using
a palladium catalyzed
cyanation [23] followed by acidic hydrolysis of the nitriles to the
corresponding carboxylic acids. The 2-substituted analogues 10e13
(Scheme 3) were prepared from 2,6-dichloropyridine using nucle-
ophilic aromatic substitution under high pressure to form 6-chloro-
2-aminopyridine (36) [24]. Subsequent regioselective iodination
with N-iodosuccinimide (NIS) [25] and palladium catalyzed cyan-
ation afforded the central intermediate 38 serving as cross-
coupling partner for the formation of 2-substituted analogues
39e41. Ethyl and butyl groups were introduced via Negishi cross
coupling with Et2Zn or BuZnBr [26], and the phenyl group was
introduced using Suzuki cross coupling with phenylboronic acid.
The 4-substituted analogues 15e17 (Scheme 3) were prepared from
commercially available 4-chloro-2-aminopyridine 42 in an analo-
gous manner to the 2-substituted analogues. The conversion of 44
to 45 and 46 was performed with iron-catalyzed cross coupling
with EtMgBr or BuMgBr [27] and to 47 by cross-coupling using
phenylboronic acid. The Negishi cross coupling conditions applied
for the synthesis of the 2-substituted 39 and 40 were unsuccessful
for the formation of the 4-substituted compounds. The 5-position
(Scheme 4) was accessed via iodination of commercially available
methyl 6-aminonicotinate (48) with iodine and silver sulphate
(Ag2SO4) [28]. The iodinated intermediate 49 was deprotected by
acidic hydrolysis affording compound 18 [28] or subjected to
Suzuki cross coupling with potassium vinyltrifluoroborate afford-
ing compound 50. Selective hydrogenation of the double bond
followed by deprotection by acidic hydrolysis afforded the 5-ethyl
analogue 20.
Fig. 1. Structures of GABAA agonists: GABA, muscimol, THIP, isoguvacine, isonipecotic
acid, 4-PIOL and general structures of the 6-aminonicotinic acid analogues 1e27.
detrimental for receptor activity indicating very limited space in
the areas surrounding the core scaffolds of these agonists [6].
Classical medicinal chemistry and pharmacophore models have
aided the insight into details of binding and function of these ag-
onists [12,13]. As a high-resolution 3D structure of a GABAAR has yet
to be published, rational ligand design is depending on homology
models created from related templates with low amino acid
sequence identities to the GABAAR. These include X-ray structures
of acetylcholine binding proteins (AChBPs) [14e16], full length re-
ceptor structures of the glutamate-gated chloride channel from
Caenorhabditis elegans (GluCl) [17], and of the bacterial ion chan-
nels ELIC [18] and GLIC [19] from Erwinia chrysanthemi and Gloeo-
bacter violaceus.
We have recently published homology models of the GABAA
a1b2g2 receptor based on these templates with both agonists [20]
and antagonists bound [21]. The most notable difference between
the agonist and antagonist models is the position of the hairpin
shaped C-loop which is more open in the antagonist model to
accommodate the larger antagonists. Furthermore, the functionally
important a1Arg66 is, as previously suggested [12,21], observed in
two distinct conformations in the two models. The models have
aided design of compounds and exploration of binding modes for a
number of agonists [20] and antagonists [21] and have led to a
better understanding of the architecture of the binding site,
including the strict steric requirements for agonist binding as
evident from the structureeactivity relationships (SARs) of 4-
aminomethyl-1-hydroxypyrazole analogues of muscimol [22].
However, the enhanced structural insight has not advanced the
field of subtype selective GABAAR agonists.
The aminotetrahydro -pyridine and -pyrimidine compounds 22
[29], 23 and 27 [29] were synthesized using catalytic hydrogena-
tion over a PtO2 catalyst starting from 6-aminonicotinic acid 3, 21,
and 26, respectively (Scheme 5). The synthesis of pyrimidine
compound 24 [30] was achieved from 5-bromopyrimidine by
Inspired by the abovementioned agonist model, we have
explored 6-aminonicotinic acid (3) as a lead structure and devel-
oped of a series of novel GABAAR agonists (9e27) in order to
challenge the limited structural diversity of GABAAR agonists and
facilitate the quest for subtype selective GABAAR agonists. We
Scheme 1. Reagents and conditions: (a) Boc2O, tert-BuOH, rt, (b) LDA,
THF, ꢀ70 ꢁCe0 ꢁC, then diethylcarbonate, 0 ꢁC, (c) 1 M aq. NaOH, MeOH, reflux.

406
J.G. Petersen et al. / European Journal of Medicinal Chemistry 84 (2014) 404e416
Scheme 2. Reagents and conditions; (a) Zn(CN)2, Pd2(dba)3, dppf, DMF/H2O 99:1,
120 ꢁC, (b) 6 M or conc. HCl, reflux.
Scheme 5. Reagents and conditions; (a) H2, PtO2, conc. HCl, H2O, rt, (b) nBuLi, THF,
EtOH, ꢀ110 ꢁC then dry ice, Et2O, ꢀ78 ꢁC to rt, (c) H2, Rh/C, H2O.
preparations. Functional characterization of selected compounds
was carried out at the human a1b2g2S and r1 GABAARs transiently
expressed in tsA201 cells using the FLIPR™ Membrane Potential
Blue Assay, performed as described previously [22].
The pharmacological data for the 2-aminopyridine series (1e8)
are given in Table 1 together with data for reference compounds
GABA, isoguvacine, and isonipecotic acid. A distinct preference for
the positioning of the carboxylic acid or acetic acid moieties in the
4- and 5-positions (R3 and R2 in Table 1) is evident, as analogues
3e6 display binding affinity in the low micromolar range while 3-
and 6-substituted analogues (1,2,7,8) are inactive. The parent
compounds, isonicotinic acid, nicotinic acid and 3- and 4-
aminobenzoic acid (not shown) were devoid of affinity at
GABAAR, highlighting importance of the amidine group and not its
individual aniline and/or pyridine components.
Scheme 3. Reagents and conditions; (a) NIS, DMF, rt, (b) 2.2 eq. nBuLi, THF/
Et2O, ꢀ110 ꢁC, then CO2 (gas), ꢀ110 ꢁC to ꢀ30 ꢁC, (c) Zn(CN)2, Pd2(dba)3, dppf, DMF/
H2O 99:1, 120 ꢁC, microwave irradiation, (d) Et2Zn, Pd2(PPh3)4, 1-methyl-2-
pyrrolidinone (NMP), 100 ꢁC, microwave irradiation, (e) BuZnBr, Pd2(PPh3)4, NMP, rt,
(f) PhB(OH)2, Pd2(PPh3)4, toluene/EtOH 10:1, 100 ꢁC microwave irradiation, (g) Conc.
HCl, reflux, (h) EtMgBr, Fe(acac)3, NMP/THF 10:1, rt, (i) BuMgBr, Fe(acac)3, NMP/THF
10:1, rt.
Table 1
Pharmacological data for reference compounds GABA, isoguvacine, and isonipecotic
acid, and 1e8.a
halogenemetal exchange and subsequent quenching with CO2 at a
low temperature. The tetrahydropyrimidine compound 25 was
obtained by catalytic hydrogenation over a Rh/C catalyst.
2.2. Pharmacology
Entry
R1
R2
R3
R4
[3H]muscimol
binding Ki (
M)b
[pKi SEM]
The binding affinities of the compounds at native GABAARs were
m
measured by displacement of [3H]muscimol in rat membrane
GABA
0.033c
Isoguvacine
Isonipecotic
0.055 [7.26 0.06]
0.51 [6.30 0.04]
acid
1
2
3
4
5
6
7
8
CO2H
CH2CO2H
H
H
H
H
H
H
H
H
H
H
H
H
>100
>100
H
H
H
H
H
H
CO2H
CH2CO2H
H
H
H
H
4.4 [5.39 0.09]
90 [4.05 0.06]
19 [4.72 0.02]
73 [4.14 0.02]
>100
CO2H
CH2CO2H
H
H
CO2H
CH2CO2H >100
a
GABAA receptor binding affinities at rat synaptic membranes.
IC50 values were calculated from inhibition curves and converted to Ki values.
b
Data are given as the mean [mean pKi
experiments.
SEM] of three to four independent
Scheme 4. Reagents and conditions; (a) I2, Ag2SO4, EtOH, (b) Conc. HCl, reflux, (c)
c
CH2CHBF3K, PdCl2(PPh3)2, K2CO3, toluene/H2O 10:1, 100 ꢁC (d) H2, Pd/C, THF, rt.
From Ref [58].

J.G. Petersen et al. / European Journal of Medicinal Chemistry 84 (2014) 404e416
407
Table 3
Compound 3 displayed the highest binding affinity of the
Pharmacological data for 3, 22e27.a
compounds initially tested and was chosen as the main scaffold for
further SAR studies. Introduction of a range of substituents in the
available positions of the pyridine ring was performed (Table 2).
Unbranched alkyl groups as substituents in the 2-position of 3 such
as methyl, ethyl and butyl are tolerated, compounds 9e11 exhib-
iting binding affinities for the GABAAR sites comparable to or
slightly higher than that of 3. However, substituents such as phenyl
and chloro (12,13) were detrimental for GABAAR binding. A similar
SAR was observed for the corresponding 4-substituted analogues
(14e17). In sharp contrast to this, the 5-position of 3 did not
tolerate any substitutions; iodide, methyl and ethyl substituted
compounds 18e20 displayed no significant binding at concentra-
tions up to 100 mM.
Entry
[3H]muscimol binding
Ki (
M)b [pKi SEM]
a1b2g2S EC50
(
m
M) [pEC50 SEM]
pKaHc
Conversion of 3 to the tetrahydropyridine compound 22 lead to
a 100-fold increase in binding affinity, thus displaying a binding
affinity similar to those of GABA and isoguvacine (Table 3). Analo-
gously to compound 3, introduction of a methyl substituent in the
amino nitrogen in 22 given rise to compound 23 lead to a signifi-
cant decrease in binding affinity.
The pyrimidine and 2-aminopyrimidine carboxylic acids 24 and
26 were devoid of binding affinity for the GABAAR at concentrations
up to 100 mM. However, analogously to 3 and the corresponding
m
3
4.4 [5.39 0.09]
0.044 [7.36 0.01]
49 [4.32 0.06]
>100
nd
6.2
12.4
nd
~1.4
12.2
22
23
24
25
2.6 [5.58 0.09]
>1000 [<3.0]
nd
Agonist,
Rmax 25 4.9 at 1 mM
nd
5.3 [5.29 0.08]
26
27
>100
0.54 [6.27 0.05]
1.9
12.7
14 [4.85 0.08]
a
GABAA receptor binding affinities at rat synaptic membranes and functional
characterization at the human a1b2g2S GABAA receptors transiently expressed in
tsA201 cells using the FLIPR™ Membrane Potential Blue assay.
tetrahydropyridine analogue 22, their non-aromatic counterparts
25 and 27 displayed binding affinities in the high nanomolar to low
micromolar range.
b
IC50 values were calculated from inhibition curves and converted to Ki values.
Data are given as the mean [mean pKi
SEM] of three to four independent
With respect to the functional properties of the analogous, 3 and
the 2-methylated analogue 9 were found to be weak agonists at
a1b2g2S and r1 GABAARs transiently expressed in tsA201 cells using
the FMP assay (Table 4). The agonist activity decreased upon
introduction of larger substituents leading to compounds that were
experiments.
c
pKaH determination by diode array detector or potentiometric detection [49].
nd: not determined.
inactive at concentrations up to 1000 mM. Interestingly, the non-
Table 2
Pharmacological data for reference compounds GABA, muscimol, isoguvacine, and isonipecotic acid, and 3, 9e21.a
Entry
R1
R2
R3
R4
[3H]muscimol binding Ki (
m
M)b [pKi SEM]
a1b2g2S EC50
(
mM) [pEC50 SEM]
r1 EC50 (mM) [pEC50 SEM]
GABA
Muscimol
Isoguvacine
Isonipecotic acid
0.033c
1.8 [5.75 0.05]
0.54 [6.27 0.06]
11 [4.95 0.09]
Agonist
0.27 [6.57 0.02]
0.70 [6.16 0.08]
39 [4.41 0.08]
Weak agonistc
0.006c
0.055 [7.26 0.06]
0.51 [6.30 0.04]
Rmax 66 5.9 at 1 mM
>300 [<3.5]
Weak agonistc
>1000 [<3.0]
>1000 [<3.0]
nd
3
9
H
H
H
H
H
H
H
Me
Et
Bu
Ph
H
H
H
H
H
H
H
H
H
H
H
H
H
H
I
H
H
H
H
H
H
H
H
H
H
H
H
H
Me
4.4 [5.39 0.09]
1.4 [5.87 0.06]
2.1 [5.68 0.02]
7.1 [5.15 0.03]
>100
Weak agonistc
Me
Et
Bu
Ph
Cl
H
H
H
H
H
Weak agonistc
10
11
12
13
14
15
16
17
18
19
20
21
>1000 [<3.0]
>1000 [<3.0]
nd
nd
>100
nd
3.6 [5.44 0.00]
9.3 [5.03 0.03]
24 [4.63 0.07]
>100
>100
>100
>1000 [<3.0]
>1000 [<3.0]
>1000 [<3.0]
nd
nd
nd
Weak agonistd
>1000 [<3.0]
>1000 [<3.0]
nd
nd
nd
nd
nd
H
H
H
Me
Et
H
>100
>100
nd
nd
a
GABAA receptor binding affinities at rat synaptic membranes and functional characterization at the human a1b2g2S and r1 GABAA receptors transiently expressed in
tsA201 cells using the FLIPR™ Membrane Potential Blue assay.
b
IC50 values were calculated from inhibition curves and converted to Ki values. Data are given as the mean [mean pKi SEM] of three to four independent experiments.
From Ref [58].
Weak agonist response at 1 mM. nd: not determined.
c
d

408
J.G. Petersen et al. / European Journal of Medicinal Chemistry 84 (2014) 404e416
Table 4
Functional characteristics of GABA, muscimol, isoguvacine, isonipecotic acid, 22 and 27 at the human a1b2g2S
expressed in tsA201 cells in the FLIPR™ Membrane Potential Blue assay.
, a2b2g2S, a3b2g2S, a5b2g2S and r1 GABAAR subtypes transiently
Entry
a1b2g2S EC50
(m
M)
a2b2g2S EC50
[pEC50 SEM] Rmax SEM (%)
(
m
M)
a3b2g2S EC50
[pEC50 SEM]
(
m
M)
a5b2g2S EC50
[pEC50 SEM]
(
m
M)
r1 EC50 (mM)
[pEC50 SEM]
[pEC50 SEM] Rmax SEMa (%)
Rmax SEM (%)
Rmax SEM (%)
Rmax SEM (%)
GABA
1.8 [5.75 0.05]
100
0.65 [6.18 0.06]
100
0.92 [6.04 0.04]
100
0.11 [6.97 0.12]
100
0.27 [6.57 0.02]
100
Muscimol
Isoguvacine
Isonipecotic acid
22
0.54 [6.27 0.06]
93 8.9
0.41 [6.39 0.14]
99 10
4.5 [5.35 0.11]
95 4.3
24 [4.62 0.04]
85 6.7
0.82 [6.08 0.12]
101 7.4
0.21 [6.67 0.10]
97 3.5
5.6 [5.25 0.10]
91 4.7
17 [4.77 0.09]
85 3.5
1.1 [5.98 0.07]
93 4.2
0.054 [7.27 0.11]
114 11
0.78 [6.10 0.09]
101 8.4
2.8 [5.54 0.06]
115 5.9
0.092 [7.04 0.09]
100 4.7
0.70 [6.16 0.08]
98 9.1
13 [4.89 0.06]
85 7.6
3.1 [5.51 0.06]
85 13
Agonistb
Weak agonistb
Rmax 66 5.9 at 1 mM
2.6 [5.58 0.09]
69 5.1
2.3 [5.91 0.05]
64 5.3
27
14 [4.85 0.08]
33 2.5
9.4 [5.03 0.06]
91 6.1
19 [4.71 0.10]
85 4.5
2.1 [5.68 0.07]
113 9.4
8.1 [5.80 0.06]
55 4.9
a
The Rmax values are given as percentage of Rmax of GABA.
Weak agonist response at 1 mM.
b
aromatic analogue 22 was shown to be equipotent with GABA and
isoguvacine as an agonist at the a1b2g2S GABAARs.
A subset of compounds was subjected to a broader functional
profiling at the human a1b2g2S, a2b2g2S, a3b2g2S, a5b2g2S, and r1
b2Ser156 and b2Tyr157. At the carboxylic acid end of the molecule,
an alternative orientation of the carboxylic acid moiety and Arg66
was observed (Fig. 3A). In this binding pose, the 2- and 4-positions
of compound 3 aligned with the observed upper and lower cavities,
respectively (Fig. 2). This orientation precluded interaction with
Thr129. An alternative binding pose fulfilling all interactions to
residues implied in GABA binding was excluded due to high in-
ternal energy (see Supporting Information). Analogues of 3 with
substituents in the 2- and 4-positions (9e12 and 14e16) were
docked and found to align well with the docked pose of 3 and with
substituents placed in the observed cavities as illustrated in Fig. 3B
and C (For docking scores please refer to Supporting Information).
GABAAR subtypes in the FMP assay. Analogously to the standard
agonists GABA, muscimol and isoguvacine, 22 and 27 displayed
agonist activities at all tested GABAAR subtypes with only minor
variations among the different subtypes (Table 4). A tendency to
lower EC50 values at the extrasynaptic a5b2g2S subtype is normally
seen for GABAA agonists and was also evident in this study. Inter-
estingly, compounds 22 and 27 showed a partial agonist profile at
the a1b2g2S and r1 GABAARs in the assay, displaying maximal re-
sponses in the range of 33e69% of the maximal response elicited by
GABA at the two receptors. In general compounds 22 and 27 were
less efficacious than the standard agonists included in this study,
whereas they were full agonists (100e113% of Rmax of GABA) at the
a5b2g2S. It is important to stress that whereas we have found the
FMP assay to be very predictive when it comes to the rank order of
agonist potencies, the efficacies displayed by GABAA and nAChR
agonists in the assay have differed substantially from those ob-
tained in at the respective receptors expressed in the Xenopus
oocyte expression system. Thus, we will refrain from drawing any
solid conclusions about the apparent different agonist efficacies
displayed by 22 based on the FMP data in this study.
2.3. Computational modelling
To aid in compound design, the surface of the GABA binding site
in the a1b2g2 GABAAR homology model [20] was mapped using the
program GRID (C3 probe) (Fig. 2).
GRID calculations revealed a hydrophobic cavity below the
agonist binding site (extending towards the membrane) and
further, slight alterations of rotameric states of a1Leu117 and
a1Arg119 exposed an additional cavity above the agonist binding
site (facing away from the membrane) which is in agreement with
similar cavities observed in an antagonist bound model (Fig. 2) [21].
In the homology model, GABA and muscimol were predicted to
form a salt bridge with a1Arg66 and further accepts hydrogen
bonds from the two threonines, a1Thr129 and b2Thr202 via the
carboxylic acid moiety and further to form a salt bridge and
hydrogen bonds with residues b2Glu155 and b2Ser156 as well as a
Fig. 2. Outline of the GABA agonist binding site. Molecular interaction field calcula-
tions performed with the program GRID using
a methyl (C3) probe (grey
mesh, ꢀ1 kcal/mol) reveals an extension to the pocket protruding downwards towards
the membrane. In the upward direction (away from the membrane) the site is capped
by a1Leu117 and a1Arg119 (sticks, wheat coloured carbons). Alteration of the rota-
meric state of these residues (sticks, red carbons) reveals an extension to the pocket
(red mesh, ꢀ1 kcal/mol). The binding pose for 3 (obtained by induced fit docking as
described for ligand docking in the experimental section) is superimposed on the
model to illustrate the position of the cavities relative to the template molecule. (For
interpretation of the references to colour in this figure legend, the reader is referred to
the web version of this article.)
pecation interaction with b2Tyr205 via the protonated amine. [20]
Using an induced fit docking protocol, a similar binding mode was
obtained for 3, with hydrogen bond interactions from the proton-
ated amidine group to b2Glu155 and the backbone carbonyls of

J.G. Petersen et al. / European Journal of Medicinal Chemistry 84 (2014) 404e416
409
Fig. 3. Predicted binding modes for compounds (A) 3 (green carbons), (B) 11 (purple carbons), (C) 16 (cyan carbons) and (D) 22 (yellow and slate coloured carbons, for the (R)- and
(S) enantiomer respectively). Grey mesh indicates the shape of the pocket determined by the steric GRID methyl probe visualized at energy level of ꢀ1 kcal/mol. (For interpretation
of the references to colour in this figure legend, the reader is referred to the web version of this article.)
The docked pose of the tetrahydropyridine analogue 22 was
observed to fulfil all interactions seen for GABA without resulting in
a high internal energy as it was observed for 3 (Fig. 3D and
Supporting Information).
bioisosteres within the GABAAR area. In addition to the bioisosteric
potential, the 2-aminopyridine functionality forms a structurally
restricted scaffold suitable for probing the GABAAR binding site
aiming for subtype selectivity.
The bioisosteric relationship was validated and the optimal
position of the carboxyl acid group on the 2-aminopyridine ring
was first investigated to focus the following SAR studies. A clear
preference for a carboxyl group in the 4- or 5-positions of the 2-
aminopyridine ring system was observed. Compounds 3 and 5
3. Discussion and conclusion
Bioisosteric replacement of the carboxylic acid group of GABA
has been widely explored [6,31,32], while the amino group at the
other end of the molecule has received much less attention. It has
been reported that alkylation of the amino group in GABA and
muscimol leads to almost complete loss of affinity for the GABA
receptor site [33]. However, if a secondary amine function is
incorporated in a cyclic structure, as in isoguvacine [8] and THIP [9],
GABAAR binding affinity is retained. Incorporation of the amino
group in an amidine system, such as that found for guanidinoacetic
acid [34] and guanidinopropionic acid [35] or even in cyclic ami-
dines like imidazole acetic acid and amino-2-thiazolin-4-ylacetic
acid [36], restores to some degree the affinity for the GABA recep-
tor site. Finally, a series of highly potent GABAAR antagonists with
the amine being part of an aminopyridazine ring system, including
the standard antagonist gabazine, have been reported [37].
displayed affinities in the low micromolar range (Ki 4.4 and 19 mM)
suggesting that these compounds were characterized by optimal
distances between the basic and acidic functionalities for the 2-
aminopyridine scaffold.
Based on molecular interaction fields and docking, 3 seemed
ideal as a template to probing the two pockets observed in the
homology model by substitutions in the 4-position or 2-position of
the 6-aminopyridine ring system.
In contrast to the limited substitution generally tolerated for
GABAA agonists, introduction of methyl, ethyl and even a butyl
group in the 2- or 4-position of the 6-aminopyridine ring of 3 (9e11,
14e16) was allowed and did not change binding affinity signifi-
cantly. Thereby, the identified binding mode of 3, 9e11 and 14e16
was supported. In contrast, introduction of methyl and ethyl sub-
stituents in the 5-position of 3 leads to loss of affinity. We have
In the present study we have introduced the 2-aminopyridine
and analogous amidine/guanidine ring systems as novel amine

410
J.G. Petersen et al. / European Journal of Medicinal Chemistry 84 (2014) 404e416
previously reported on a series of analogues of the low efficacy
GABAAR agonist 4-PIOL, where a similar SAR pattern was observed.
Introducing unbranched alkyl substituents into the 4-position of
the 3-isoxazolol ring of 4-PIOL, elicited a similar response as seen
for the 2- and 4-substituted variants of 3. Further probing of the
binding pocket by structurally similar 3-hydroxyisothiazole and 1-
hydroxypyrazole scaffolds also reflected these findings [38]. The
combined SAR data from these studies verify the performed GRID
calculations that imply spacious cavities in the GABAAR binding site
pocket above and below the central cavity where the core scaffolds
binds.
extrasynaptic a5-subunit containing GABAARs are proposed to be
involved in learning and memory [47]. Interestingly, 22 as well as
27 were shown to be full agonists at the extrasynaptically a5b2g2S
subtype and less efficacious at the synaptically localized GABAAR
subtypes (a1b2g2S, a2b2g2S, a3b2g2S, and r1). This observed ten-
dency towards functional selectivity for the a5b2g2S GABAAR sub-
type is also seen for isoguvacine and especially isonipecotic acid in
this study. Considering the coarse nature of the FMP assay these
findings have to be confirmed by electrophysiological recordings at
GABAARs expressed in e.g. Xenopus oocytes.
Incorporating the amino group of GABA in a pyrimidine or
pyrimidin-2-amine ring system, 24 and 26, led to compounds
devoid of affinity. Though mimicking the electron distribution in
the guanidinic/amidinic system, the significantly lowered pKaH
values (pKaH 1.4e1.9), with an increasing uncharged fraction of
compound at physiological pH, most likely result in loss of affinity.
In general, a significant increase in binding affinities was
observed for the tetrahydro analogues, 22, 25, 27, of 2-
aminopyridine, pyrimidine and 2-aminopyrimidine compared to
their parent compounds, 3, 24 and 26. For the 2-
aminotetrahydropyridine analogue of 3 (22), a 100-fold increase
in affinity was observed. At least two different effects could be held
accountable for this: 1) Increased conformational flexibility of 22
allowing it to form interactions identical to those of GABA as
opposed to 3 which has to assume a high energy conformation to
do the same (Supporting Information). 2) Significant changes in
pKaH values for the basic groups (Table 3). The more basic amidine
group in 22 is to a larger extent protonated and may form stronger
interactions in the binding pocket relative to 3.
To challenge the importance of the basic center, analogues to 3
containing only one nitrogen atom in the basic center; 3- and 4-
pyridine and 3- and 4-anilin, were included in the study. They
showed no affinity for the GABAARs in agreement with the need for
a positive ionisable group [39,40]. The slightly lower pKaH values of
the pyridine and aniline groups compared to the 2-aminopyridine
group likely influence the extent of protonation thereby affecting
the strength of interactions involving the basic center. Further,
introduction of a methyl group in the external amino group in 3
affording 21 and in the corresponding tetrahydro analogue 22
affording 23 abolished GABAAR affinity. A similar detrimental effect
has been reported for muscimol and the secondary amine ana-
logues isoguvacine and THIP [8,41,42]. These observations suggest
that ligands for the GABA receptors are sensitive to steric hindrance
in proximity of the cationic moiety of the molecule, and that the
primary or secondary structure of the amino group is of minor
importance.
Despite the obvious need for subtype selective GABA ligands
only few orthosteric GABAAR agonists have been systematically
characterized on different GABAAR subtypes [43,44]. In general,
only small differences in binding affinities and agonist activity were
observed reflecting the high degree of similarity between the in-
dividual GABAAR binding sites. However, THIP and allosteric mod-
ulators for the benzodiazepine binding site [45] show that
functional selectivity can be obtained for both orthosteric ligands
and modulators at some receptor subtypes in spite of a lack of
binding selectivity.
In the present study, a5b2g2S subtype preference was observed
for 22 based on potency, which is in line with observations on
GABAAR agonists in general. The a5b2g2S GABAAR subtype is located
largely at extrasynaptic sites, primarily in hippocampus, where
they are subject to activation of low ambient GABA concentration
mediating tonic inhibition distinct from the transient activation of
synaptic GABAARs leading to classical inhibitory postsynaptic
phasic currents [46]. The tonic inhibition mediated by
3.1. Conclusion
The study identified the 2-aminopyridine and 2-
aminotetrahydropyridine/pyrimidine scaffolds as bioisosteres for
the amino group in GABA at GABAA receptors. A series of 2-
aminopyridine/pyrimidine carboxylic acid (1e21, 24, 26) and cor-
responding tetrahydro analogues (22, 23, 25, 27) have been syn-
thesized and characterized pharmacologically at GABAARs. The 2-
aminopyridine analogue 6-aminonicotinic acid (3) was shown to
be a weak agonist, and SAR studies showed that substitutions in the
2- and 4-position were tolerated. The tetrahydro analogues (22, 25,
27) were shown to be moderate to potent agonists (Ki of
0.044e5.4 mM). Cavities surrounding the core of the GABA binding
pocket were predicted by molecular interaction field calculations
and docking studies, and were confirmed by affinities of substituted
analogues of 3. This is in line with previous SAR studies of the
significantly larger 4-PIOL analogues but has not previously been
reported in relation to agonists. The bioisosteric properties of the
novel scaffolds are supported by a common binding mode and
similar interactions for the tetrahydro analogues of 3 and GABA
itself, which is in agreement with the observed agonist profile.
These findings have prompted us to further elucidate the structural
features of this class of GABA analogues in regard to the identified
cavities and the observed agonist profile, subtype selectivity and
activity at other GABA related targets. New possibilities have been
revealed for ligand design for the orthosteric binding site in general
paving the way for diversity in both agonist structure and phar-
macological profile.
4. Experimental section
4.1. Chemistry
4.1.1. General procedures
Reagents and solvents were obtained from commercial sup-
pliers and used without further purification. Anhydrous CH2Cl2,
DMF and THF were obtained by using a solvent purification system,
and other anhydrous solvents were obtained by storage over 4 Å
molecular sieves. All reactions involving air- and/or moisture sen-
sitive reagents were performed under a nitrogen atmosphere using
syringe-septum cap techniques and/or with flame-dried glassware.
Analytical thin-layer chromatography (TLC) was carried out using
Merck silica gel 60 F254 plates and the compounds were visualized
with UV light (254 nm), ninhydrin or KMnO4 spraying agents. NMR
spectra were recorded on a 300 MHz Varian Mercury Plus spec-
trometer, a 300 MHz Varian Gemini 2000 spectrometer or a
400 MHz Bruker Avance spectrometer. Dry column vacuum chro-
matography (DCVC) [48] was performed using Fischer Scientific
silica gel 60 (20e45
formed using Merck silica gel 60 (40e63
(RP-FC) was performed using Merck LiChroprep RP-18 (40e63
Analytical HPLC (anal. HPLC) was performed on a Merck-Hitachi
HPLC system consisting of an L-7100 pump, an L-7200 autosam-
pler, and an L-7400 UV detector (254 nm), using an X-Terra column
m
m). Flash chromatography (FC) was per-
m). Reverse phase FC
m).
m
m

J.G. Petersen et al. / European Journal of Medicinal Chemistry 84 (2014) 404e416
411
(4.6 ꢂ 150 mm) eluted at a flow rate of 0.8 mL/min. A linear
d
8.37 (s, 1H), 6.82 (s, 1H), 2.56 (s, 3H). 13C NMR (D2O)
d
167.7, 157.2,
gradient elution was performed with eluent A (H2O-TFA 100:0.1)
containing 0% of solvent B (MeCNeH2OeTFA, 90:10:0.01) rising to
100% of B during 25 min. Data were acquired and processed using
the EZChrom Elite Software version 3.1.7 by Hitachi. All microwave
reactions were carried out in a glass reactor using a Biotage Initiator
instrument. Melting points were determined by OptiMelt from
Stanford Research Systems in open capillary tubes and are uncor-
rected. Elemental analysis was 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.
154.9, 140.7, 116.3, 114.1, 21.8. Anal. (C7H8N2O2$CF3COOH) C, H, N.
4.1.2.6. 6-Amino-4-ethylnicotinic acid hydrochloride (15).
Hydrolysis according to the general procedure using 45 (96 mg,
0.65 mmol), conc. HCl, microwave irradiation, 12 h. Recrystalliza-
tion from MeOH afforded the product as off-white solid (41 mg,
31%). mp (decomp.) >245 ꢁC. 1H NMR (MeOH-d4)
d 8.44 (d,
J ¼ 0.8 Hz, 1H), 6.88 (q, J ¼ 0.8 Hz, 1H), 3.11 (dq, J ¼ 0.75 Hz, J ¼ 7.30,
2H),1.26 (t, J ¼ 7.3 Hz, 3H). 13C NMR (MeOH-d4)
d 166.0,164.6,156.4,
141.7, 117.6, 113.7, 28.8, 14.3. Anal. (C8H10N2O2$HCl) C, H, N.
4.1.2.7. 6-Amino-4-butylnicotinic
acid
hydrochloride
(16).
4.1.2. General procedure for hydrolysis (9e12, 14e20)
The protected compounds (33e35, 39e41, 51) were dissolved in
6 M or conc. HCl and heated at reflux using conventional heating or
microwave irradiation for 12 h to 4 days followed by evaporation
under vacuum.
Hydrolysis according to the general procedure using 46 (83 mg,
0.38 mmol), conc. HCl, microwave irradiation, 12 h. The solvent was
decanted, and the precipitate dried under vacuum affording the
product as off-white crystals (70 mg, 79%). mp 195.3e198.0 ꢁC. 1H
NMR (D2O) d 8.36 (s,1H), 6.85 (s,1H), 2.96 (m, 2H),1.54 (m, 2H),1.35
(sxt, J ¼ 7.3 Hz, 2H), 0.89 (t, J ¼ 7.3 Hz, 3H). 13C NMR (D2O)
d 167.6,
161.6, 154.0, 139.7, 117.0, 113.2, 33.6, 31.5, 21.9, 13.0. Anal.
4.1.2.1. 6-Amino-2-methylnicotinic
acid
hydrochloride
(9).
(C10H14N2O2$1.05 HCl) C, H, N.
Hydrolysis according to the general procedure using 33 (200 mg,
1.50 mmol), 6 M HCl, conventional heating, 24 h. Recrystallization
from EtOH/Et2O afforded the product as brown solid (218 mg, 77%).
An analytical amount was washed with hot MeCN and used for
4.1.2.8. 6-Amino-4-phenylnicotinic
acid
hydrochloride
(17).
Hydrolysis according to the general procedure using 47 (120 mg,
61 mmol), conc. HCl, microwave irradiation, 12 h. Recrystallization
from MeOH/Et2O afforded the product as white solid (20 mg, 7%).
characterization and in pharmacological tests. 1H NMR (D2O):
d 8.31
(d, J ¼ 9.4 Hz, 1H), 6.91 (d, J ¼ 9.4 Hz, 1H), 2.77 (s, 3H). 13C NMR
mp (decomp.) >207 ꢁC. 1H NMR (MeOH-d4)
d
7.19 (d, J ¼ 0.8 Hz, 1H),
(D2O):
d 167.8, 154.9, 152.0, 145.0, 115.1, 110.3, 18.8. Anal.
6.19 (m, 3H), 6.12 (m, 2H), 5.65 (d, J ¼ 0.8 Hz, 1H). 13C NMR (MeOH-
(C7H8N2O2$1.5HCl) C, H, N.
d4)
d 166.4, 159.1, 156.1, 141.1, 138.8, 130.7, 129.5, 129.0, 118.5, 115.6.
Anal. HPLC purity 95% (254 nm).
4.1.2.2. 6-Amino-2-ethylnicotinic
acid
hydrochloride
(10).
Hydrolysis according to the general procedure using 39 (124 mg,
0.84 mmol), conc. HCl, conventional heating, 3 days. Purification by
RP-DCVC (0.1% TFA in H2O) followed by evaporation from conc. HCl,
and recrystallization from EtOH afforded the product as white
4.1.2.9. 6-Amino-5-iodonicotinic acid hydrochloride (18) [28].
Hydrolysis according to the general procedure using 49 (60 mg,
0.22 mmol), conc. HCl, conventional heating, 24 h. Recrystallization
from MeOH/Et2O afforded the product as white crystals (59 mg,
crystals (16 mg, 10%). mp 254.4e255.7 ꢁC. 1H NMR (D2O)
d
8.17 (m,
1H), 6.77 (m, 1H), 3.04 (m, 2H), 1.20 (m, 3H). 13C NMR (D2O)
d
167.7,
91%). mp. 280e284 ꢁC. 1H NMR (CD3OD)
d
8.82 (d, J ¼ 1.8 Hz, 1H),
8.50 (d, J ¼ 1.8 Hz, 1H). 13C NMR (CD3OD)
d 163.4, 155.9, 153.2, 139.7,
156.8, 155.1, 145.2, 114.4, 110.4, 25.5, 13.0. Anal. (C8H10N2O2$1.1HCl)
C, H, N.
118.1, 79.3. Anal. (C6H5IN2O2$0.8HCl): C, H, N.
4.1.2.3. 6-Amino-2-butylnicotinic
acid
hydrochloride
(11).
4.1.2.10. 6-Amino-5-methylnicotinic
acid hydrochloride (19).
Hydrolysis according to the general procedure using 40 (225 mg,
1.28 mmol), conc. HCl, conventional heating, 4 days. Purification by
RP-DCVC (0.1% TFA/MeCN 100:0 to 95:5) followed by evaporation
from conc. HCl, and recrystallization from EtOH afforded the
product as off-white crystals (18 mg, 6%). mp 212.8e215 ꢁC. 1H NMR
Hydrolysis according to the general procedure using 35 (200 mg,
1.50 mmol), 6 M HCl, conventional heating, 24 h. Recrystallization
from EtOH/Et2O afforded the product as grey solid (233 mg, 83%).
1H NMR (D2O) 8.31 (s, 1H), 8.09 (s, 1H), 2.21 (s, 3H). 13C NMR (D2O)
d
d
166.9, 154.9, 141.9, 137.0, 123.1, 166.6, 16.1. Anal. HPLC purity 99%
(D2O)
1.65e1.49 (m, 2H), 1.39e1.23 (m, 2H), 0.89e0.77 (m, 3H). 13C NMR
(D2O) 167.8, 155.7, 155.1, 145.1, 114.9, 110.3, 31.6, 31.2, 21.9, 12.9.
d 8.11e8.22 (m, 1H), 6.86e6.66 (m, 1H), 3.09e2.95 (m, 2H),
(254 nm).
d
4.1.2.11. 6-Amino-5-ethylnicotinic
acid
hydrochloride
(20).
Anal. (C10H14N2O2$1.1HCl) C, H, N.
Hydrolysis according to the general procedure using 51 (70 mg,
0.39 mmol), conc. HCl, conventional heating, 24 h. Recrystallization
from EtOH/Et2O afforded the product as white solid (71 mg, 91%).
4.1.2.4. 6-Amino-2-phenylnicotinic
acid
hydrochloride
(12).
mp. 265e268 ꢁC. 1H NMR (CDCl3)
d
8.59 (d, J ¼ 2.1 Hz, 1H), 7.85 (d,
Hydrolysis according to the general procedure using 41 (175 mg,
0.90 mmol), conc. HCl, conventional heating, 4d. Recrystallization
from MeOH afforded the product as light brown solid (23 mg, 10%).
J ¼ 2.1 Hz, 1H), 5.30 (b s, 2H), 2.45 (q, J ¼ 7.5 Hz, 2H), 1.28 (t,
J ¼ 7.5 Hz, 3H). 13C NMR (CDCl3)
d 167.0, 160.1, 149.0, 136.4, 121.3,
mp 205.7e207.4 ꢁC. 1H NMR (MeOH-d4)
d
8.41 (d, J ¼ 9.5 Hz, 1H),
116.9, 52.1, 23.7, 12.21. Anal. (C8H10N2O2$1.1 HCl) C, H, N.
7.54 (m, 5H), 7.05 (d, J ¼ 9.5, 1H). 13C NMR (MeOH-d4)
d 166.6, 157.1,
152.5, 146.4, 134.3, 131.9, 130.1, 129.7, 117.2. 113.2. Anal.
4.1.3. 2-(2-Aminopyridin-4-yl)acetic acid hydrochloride (6)
(C12H10N2O2$1.3 HCl) C, H, N.
Aqueous NaOH (1 M, 1.7 mL) was added to a solution of 29
(0.393 g, 1.4 mmol) in MeOH (8 mL). The reaction mixture was
heated to reflux for 20 min. Conc. HCl (0.17 mL) was added to pH 4
followed by evaporation under vacuum. Ion exchange and recrys-
tallisation from isopropanol/water afforded the product as colour-
4.1.2.5. 6-Amino-4-methylnicotinic acid trifluoroacetic acid (14).
Hydrolysis according to the general procedure using 34 (482 mg
crude), conc. HCl, conventional heating, 24 h. Purification by RP-
DCVC followed by recrystallization from H2O afforded the prod-
uct as white crystals (163 mg, 23% over 2 steps). 1H NMR (D2O)
less crystals (122 mg, 39%). 1H NMR (300 MHz, D2O):
d 7.64 (1H, d,
J ¼ 6.3 Hz), 6.82 (1H, s), 6.75 (1H, d, J ¼ 6.3 Hz), 3.65 (2H, s). 13C NMR

412
J.G. Petersen et al. / European Journal of Medicinal Chemistry 84 (2014) 404e416
(75 MHz, D2O): 174.9, 154.0, 153.4, 135.1, 115.1, 114.1, 67.2. Anal.
(C7H8N2,0.8 HCl) C, H, N.
precooled dry Et2O (50 mL). The reaction mixture was poured into
this mixture and allowed to heat to room temperature. The ether
was extracted with water (3 ꢂ 50 mL) and 1 N NaOH (50 mL), and
the combined water phases were acidified with 1 N HCl. The water
was reduced to a quarter volume by freezedrying resulting in
precipitation of a white solid. The solid was separated by filtration
to give 0.564 g (56%) of product as a white solid. mp. (decomp.)
4.1.4. 6-Amino-2-chloronicotinic acid (13)
A solution of 37 (200 mg, 0.79 mmol) in anhydrous THF/Et2O
(1:1, 12.5 mL) was cooled to ꢀ110 ꢁC on an EtOH/liquid N2 bath. n-
BuLi (2.5 M in hexanes, 0.69 mL, 1.73 mmol) was added dropwise
and after 15 min, CO2 (g) was bubbled through the solution for
35 min while the temperature was allowed to rise to ꢀ30 ꢁC. H2O
(10 mL) was added, and the phases were separated. The organic
phase was extracted with H2O (4 ꢂ 10 mL) and 1 M NaOH (10 mL).
The combined aqueous layer was acidified with 6 M HCl till pH 2
and evaporated partly under vacuum until precipitation occurred.
The precipitate was isolated by filtration and dried under vacuum
affording the product as brown solid (30 mg, 22%). mp (decomp.)
>256 ꢁC. 1H NMR (DMSO-d6)
1H). 13C NMR (DMSO-d6)
(C5H4N2O2,0.2HCl) C, H, N.
d
9.39 (s, 1H), 9.21 (s, 2H) 3.34 (br s.,
164.8, 161.1, 157.7, 124.9. Anal.
d
4.1.9. 1,4,5,6-Tetrahydropyrimidine-5-carboxylic acid (25)
24 (0.250 g, 1.90 mmol) was dissolved in water (25 mL). Rh/C
(0.235 g, 5%, 0.114 mmol Rh) was added and the reaction mixture
was flushed with H2 three times. The reaction mixture was stirred
under H2 over night. The mixture was filtered through cotton wool
and evaporated to dryness to give the crude product as a white solid
(203 mg, 84%). The product was recrystallized from MeOH and Et2O
to give 25 (120 mg; 49%) as white crystals. mp. (decomp.) >294 ꢁC.
>264 ꢁC. 1H NMR (DMSO-d6)
d
12.57 (b s, 1H), 7.88 (d, J ¼ 8.5 Hz,
1H), 7.05 (b s, 2H), 6.39 (d, J ¼ 8.5 Hz, 1H). 13C NMR (DMSO-d6)
d
165.3, 161.0, 149.4, 141.8, 111.9, 106.0. Anal. (C6H5Cl N2O2) C, H, N.
4.1.5. 6-(Methylamino)nicotinic acid hydrochloride (21)
1H NMR (D2O)
d
7.94 (s, 1H), 3.65 (dd, J ¼ 13.3, 4.8 Hz, 2H), 3.52 (dd,
A solution of 6-chloronicotinic acid (1.00 g, 6.35 mmol) in pyr-
idine (2.0 mL) and methylamine (40% aq. solution, 2.7 mL,
32.0 mmol) was heated using microwave irradiation at 150 ꢁC for
12 h. H2O (4 mL) was added, and 1 M HCl was added until pH ~ 3,
then a few drops of 5 M NaOH was added, and precipitation
occurred. The mixture was stored for 3 days at 5 ꢁC, and the pre-
cipitate was collected by filtration. The precipitate was washed
with cold H2O and dried under vacuum affording the product as
white solid (307 mg, 32%). mp 253.2e254.8 ꢁC. 1H NMR (DMSO-d6)
J ¼ 13.4, 7.8 Hz, 2H), 2.90 (m, 1H). 13C NMR (D2O)
36.5. Anal. (C5H8N2O2,0.05HCl) C, H, N.
d 177.4, 151.1, 40.4,
4.1.10. 2-Amino-1,4,5,6-tetrahydropyrimidine-5-carboxylic acid
hydrochloride (27) [29]
2-Aminopyrimidine-5-carboxylic acid (26) (0.200 g, 1.44 mmol)
was dissolved in water (20 mL) and conc. HCl (1.2 mL). PtO2
(0.040 g, 0.176 mmol) was added and the reaction mixture was
flushed with H2 three times and stirred under H2 for 48 h. The
mixture was filtered through cotton wool and evaporated to dry-
ness to give the crude product as a white solid (248 mg, 96%).
100 mg crude product was recrystallized from EtOH to give 41 mg
d
12.35 (b s, 1H), 8.55 (dd, J ¼ 2.3 Hz, J ¼ 0.5 Hz, 1H), 7.80 (dd,
J ¼ 8.8 Hz, J ¼ 2.3 Hz, 1H), 7.32 (d, J ¼ 4.0, 1H), 6.47 (dd, J ¼ 8.8,
J ¼ 0.5 Hz, 1H), 2.82 (d, J ¼ 4.0 Hz, 3H). 13C NMR (DMSO-d6)
d 166.8,
161.3, 150.8, 150.6, 137.3, 113.8, 27.8. Anal. (C7H8N2O2$0.05HCl) C, H,
(41%) of white crystals. mp. 258.8e260.5 ꢁC. 1H NMR (D2O)
d
3.60
N.
(m, 4H), 3.17 (quintet, J ¼ 5.3 Hz, 1H). 13C NMR (D2O)
39.4, 35.5. Anal. (C5H9N3O2,HCl) C, H, N.
d 174.9, 153.7,
4.1.6. 2-Amino-3,4,5,6-tetrahydropyridine-5-carboxylic acid
hydrochloride (22) [29]
4.1.11. tert-Butyl 4-methylpyridin-2-ylcarbamate (28)
3 (100 mg, 0.72 mmol) was dissolved in water (10 mL) and conc.
HCl (0.6 mL). PtO2 (20 mg, 0.08 mmol) was added and the reaction
mixture was hydrogenated at rt for 17 h. The mixture was filtered
through cotton wool and evaporated under vacuum. Recrystalli-
zation from EtOH afforded the product as pale yellow crystals
(50 mg, 39%). mp. (decomp) > 200 ꢁC. 1H NMR (D2O) 3.63 (m, 2H),
4-Methylpyridine-2-amine (10.0 g, 92 mmol) was added to a
solution of di-tert-butyl dicarbonate (22.2 g, 102 mmol) in tert-
butanol (600 mL). The reaction mixture was stirred for 10 h at room
temperature followed by evaporation under vacuum. Recrystalli-
sation from isopropanol afforded the product as colourless crystals
(19.3 g, 88%): 1H NMR (300 MHz, CDCl3):
d 9.04 (1H, b s), 8.14 (1H, d,
3.03 (m, 1H), 2.75 (t, J ¼ 6.7 Hz, 2H), 2.20 (m, 1H), 2.05 (m, 1H). 13
C
J ¼ 4.8 Hz), 7.83 (1H, s), 6.77 (1H, d, J ¼ 4.8 Hz), 2.35 (3H, s), 1.54 (9H,
NMR (D2O) 176.4, 166.1, 41.9, 36.7, 24.1, 20.1. Anal.
s).
(C6H10N2O2,1.1HCl) C, H, N.
4.1.12. Ethyl 2-(2-(tert-butoxycarbonylamino)pyridin-4-yl)acetate
(29)
4.1.7. 6-(Methylamino)-2,3,4,5-tetrahydropyridine-3-carboxylic
acid hydrochloride (23)
Prepared as described for 22 using 21 (100 mg, 0.66 mmol), H2O
(6.5 mL), conc. HCl (0.15 mL), and PtO2 (20 mg, 0.09 mmol) for 3
days. Recrystallization from glacial AcOH/EtOAc afforded the
product as white solid (48 mg, 38%). mp 201.5e204.7 ꢁC. 1H NMR
A solution of 28 (1.0 g, 4.8 mmol) in THF (10 mL) was added to a
solution of LDA (11 mmol) in THF (5 mL) at ꢀ70 ꢁC. Stirring was
continued for 1 h at ꢀ70 ꢁC and 0 ꢁC for 1 h. Diethylcarbonate
(567 mL, 4.8 mmol) was added to the reaction mixture and stirring
was continued for 1 h at 0 ꢁC. H2O (60 mL) was added and the
reaction mixture was extracted using Et2O (3 ꢂ 100 mL). The
combined organic phase was dried and evaporated under vacuum.
Purification by FC (petroleum ether/EtOAc 4:1) afforded the prod-
(D2O)
J ¼ 7.5 Hz, 1H), 2.99 (m, 1H), 2.85 (s, 3H), 2.68 (t, J ¼ 6.5 Hz, 2H), 2.17
(m, 1H), 2.00 (m, 1H). 13C NMR (D2O)
176.7, 163.6, 42.1, 37.2, 27.5,
d
3.66 (dd, J ¼ 13.3 Hz, J ¼ 5.5 Hz, 1H), 3.58 (dd, J ¼ 13.3 Hz,
d
24.6, 20.4. Anal. (C7H12N2O2$1.6HCl) C, H, N.
uct (400 mg, 30%). 1H NMR (300 MHz, CDCl3):
d 8.52 (1H, b s), 8.19
(1H, d, J ¼ 5.1 Hz), 7.94 (1H, s), 6.92 (1H, d, J ¼ 5.1), 4.16 (2H, q,
4.1.8. Pyrimidine-5-carboxylic acid (24) [30]
J ¼ 7.2 Hz), 3.62 (2H, s), 1.53 (9H, s), 1.26 (3H, t, 7.2 Hz).
A solution of 5-bromopyrimidine (1.00 g, 6.29 mmol) in dry Et2O
and dry THF (1:1, 60 mL) was cooled to between ꢀ105 ꢁC
and ꢀ110 ꢁC on an EtOH/liquid N2 bath and stirred for 15 min. nBuLi
(4.10 mL, 2.3 M, 9.44 mmol), cooled to ꢀ78 ꢁC, was added dropwise
under stirring at the same temperature, and the mixture was stir-
red for additional 10 min. Pellets of dry ice was covered with
4.1.13. 6-Amino-2-methylniconitrile (33)
A mixture of 30 (500 mg, 2.67 mmol), Zn(CN)2 (188 mg,
1.60 mmol), Pd2(dba)3 (122 mg, 0.13 mmol), and dppf (148 mg,
0.27 mmol) was added degassed DMF/H2O (99:1, 3.1 mL). The re-
action mixture was heated using conventional heating at 120 ꢁC for

J.G. Petersen et al. / European Journal of Medicinal Chemistry 84 (2014) 404e416
413
24 h. The resulting mixture was cooled to rt, and sat. aq. NH4Cl/
conc. NH3/H2O (4:1:4, 10 mL) was added resulting in precipitation.
The mixture was cooled to 0 ꢁC and filtered. The precipitate was
washed with sat. aq. NH4Cl/conc. NH3/H2O (4:1:4, 10 mL) and H2O
at 5 ꢁC and dried under vacuum affording the product as dark or-
ange solid (279 mg, 78%). mp. 162e164 ꢁC. IR 2211 cmꢀ1. 1H NMR
solid (333 mg, 53% over 2 steps). 1H NMR (CDCl3)
d
7.56 (d,
J ¼ 8.5 Hz, 1H), 6.34 (d, J ¼ 8.5 Hz, 1H), 4.99 (b s, 2H), 2.86 (q,
J ¼ 7.6 Hz, 2H),1.31 (t, J ¼ 7.6 Hz, 3H). 13C NMR (CDCl3)
d 167.4, 159.8,
141.4, 118.5, 105.7, 96.6, 30.4, 13.5.
4.1.19. 6-Amino-2-butylnicotinonitrile (40)
(CDCl3)
d
7.54 (d, J ¼ 8.4 Hz, 1H), 6.33 (d, J ¼ 8.4 Hz, 1H), 5.06 (b s,
A mixture of NMP (8 mL), I2 (82.6 mg, 0.33 mmol), and Zn
(809 mg, 12.4 mmol) was stirred at rt until the red colour of I2
disappeared (ca. 4 min). Butyl-1-bromide (0.88 mL, 8.14 mmol) was
added and the mixture was stirred at 80 ꢁC for 3 h. The mixture was
cooled to rt and 38 (1.00 g crude, max. 4.30 mmol) and Pd2(PPh3)4
(150 mg, 0.13 mmol) were added, and the reaction mixture was
stirred at rt for 1 h. Sat. aq. NH4Cl (120 mL) was added, and the
mixture was extracted with EtOAc (2 ꢂ 120 mL). The combined
organic phase was washed with brine (120 mL), dried over Mg2SO4,
and evaporated under vacuum. Purification by FC (heptane/EtOAc
100:0 to 0:100) afforded the product with small impurities as light
2H), 2.55 (s, 3H). 13C NMR (CDCl3)
97.6, 23.8.
d 162.8, 159.9, 141.5, 118.9, 105.8,
4.1.14. 6-Amino-4-methylniconitrile (34)
Prepared as described for 33 using 31 (500 mg, 2.67 mmol),
Zn(CN)2 (190 mg, 1.60 mmol), Pd2(dba)3 (122 mg, 0.13 mmol), dppf
(12 mg, 0.21 mmol), and degassed DMF/H2O (99:1, 3 mL) affording
the crude product as brown solid (482 mg, 136% crude yield). The
product was used without further purification for preparation of
34. 1H NMR (CDCl3)
3H). 13C NMR (CDCl3)
d
8.25 (s, 1H), 6.34 (s, 1H), 4.90 (b s, 2H), 2.39 (s,
160.9, 153.7, 151.6, 117.8, 110.8, 100.4, 20.53.
d
yellow solid (225 mg, 29% over 2 steps). 1H NMR (CDCl3)
d 7.56 (d,
J ¼ 8.5 Hz, 1H), 6.35 (d, J ¼ 8.5 Hz, 1H), 5.00 (b s, 2H), 2.83 (t,
J ¼ 8.8 Hz, 2H), 1.66e1.79 (m, 2H), 1.43 (sxt, J ¼ 7.3 Hz, 2H),
1.20e1.35 (m, 3H, impurity), 0.97 (t, J ¼ 7.3 Hz, 3H), 0.90 (t,
4.1.15. 6-Amino-5-methylniconitrile (35)
Prepared as described for 33 using 32 (500 mg, 2.67 mmol),
Zn(CN)2 (188 mg, 1.60 mmol), Pd2(dba)3 (6 mg, 0.13 mmol), dppf
(14 mg, 0.27 mmol), and degassed DMF/H2O (99:1, 3.1 mL). Pre-
cipitation afforded the product as red-brown solid (281 mg, 79%).
J ¼ 6.7 Hz, 2H, impurity). 13C NMR (CDCl3)
d 166.3, 159.7, 141.1, 135.1
(impurity), 127.6 (impurity), 118.5, 105.5, 96.6, 36.7, 31.8 (impurity),
31.4, 29.0 (impurity), 22.6 (impurity), 22.4, 14.1 (impurity), 13.8.
mp.184e188 ꢁC. 1H NMR (CDCl3)
d
8.03 (s,1H), 7.04 (s, 1H), 4.94 (b s,
162.1, 148.7, 140.3, 117.5, 113.8,
2H) 1.94 (s, 3H). 13C NMR (CDCl3)
98.3, 16.5.
d
4.1.20. 6-Amino-2-phenylnicotinonitrile (41)
38 (500 mg, 3.25 mmol), phenylboronic acid (595 mg,
4.88 mmol), K3PO4 (1.73 g, 8.14 mmol) and Pd(PPh3)4 (0.16 mmol)
were added to a microwave vial together with toluene/EtOH (10:1,
11 mL). The vial was flushed with N2, sealed and heated under
microwave irradiation for 1 h at 100 ꢁC. The reaction mixture was
cooled to rt, 1 M NaOH (30 mL) was added, and the mixture was
extracted with EtOAc (2 ꢂ 30 mL). The combined organic phase was
dried over Mg2SO4 and evaporated under vacuum. Purification by
DCVC (heptane/EtOAc 100:0 to 60:40) afforded the product as
yellow solid (416 mg, 65%). mp 204.1e204.9 ꢁC. 1H NMR (CDCl3)
4.1.16. 6-Chloro-5-iodopyridin-2-amine (37) [24]
36 (1.00 g, 7.78 mmol) was dissolved in DMF (40 mL), the re-
action flask was covered with aluminium foil, and NIS (970 mg,
4.28 mmol) was added. The reaction mixture was stirred for 19 h
before additional NIS (970 mg, 4.28 mmol) was added, and the
reaction mixture was stirred for another 24 h. H2O (200 mL) was
added, and the mixture was extracted with EtOAc (2 ꢂ 250 mL). The
combined organic layer was washed with H2O (2 ꢂ 200 mL) and
brine (200 mL), dried over Na2SO4, and evaporated under vacuum.
Recrystallization from EtOH followed by recrystallization of the
mother liquor afforded the product as orange, needle-shaped
d
7.89e7.83 (m, 2H), 7.75 (d, J ¼ 8.8 Hz, 1H), 7.54e7.47 (m, 3H), 6.50
(d, J ¼ 8.8 Hz, 1H). 13C NMR (DMSO-d6)
d 161.0, 141.6, 138.0, 129.5,
128.4, 128.2, 119.7, 106.6, 92.2.
crystals (1.45 g, 73%). mp 154.3e155.9. 1H NMR (CDCl3)
d
7.75 (d,
157.6,
J ¼ 8.2 Hz, 1H), 6.22 (d, J ¼ 8.2 Hz, 1H). 13C NMR (CDCl3)
d
4.1.21. 4-Chloro-5-iodopyridin-2-amine (43)
151.6, 149.4, 108.8.
As described for 37 using 4-chloropyridin-2-amine 42 (4.93 g,
38.3 mmol), NIS (11.22 g, 49.9 mmol), and DMF (80 mL). Purification
by DCVC afforded the product as orange solid (8.74 g, 90%). 1H NMR
4.1.17. 6-Amino-2-chloronicotinonitrile (38)
Prepared as described for 33 using 37 (4.00 g, 15.7 mmol),
Zn(CN)2 (1.11 g, 9.43 mmol), dppf (871 mg, 1.57 mmol), Pd2(dba)3
(720 mg, 0.78 mmol), and degassed DMF/H2O (99:1, 20 mL). The
reaction mixture was heated at 120 ꢁC using microwave irradiation
for 1 h. Precipitation afforded the crude product as a dark green
solid (3.66 g, 152% crude yield). 200 mg crude product was purified
for characterization by DCVC (heptane/EtOAc/Et3N 100:0:1 to
60:40:1) affording the product as white solid (62 mg). mp
(decomp.) >213 ꢁC. 1H NMR (DMSO-d6) 7.76 (d, J ¼ 8.5 Hz, 1H), 7.52
(CDCl3)
158.6, 156.2, 148.2, 109.3, 82.0.
d
8.24 (s, 1H), 6.60 (s, 1H), 4.52 (bs, 1H). 13C NMR (CDCl3)
d
4.1.22. 6-Amino-4-chloronicotinonitrile (44)
As described for 38 using 43 (8.58 g, 33.7 mmol), Zn(CN2)
(2.38 g, 20.2 mmol), Pd2(dba)3 (1.54 g, 1.69 mmol), dppf (1.87 g,
3.37 mmol), and DMF/H2O 99:1 (45 mL). Purification was attemp-
ted using DCVC (heptane/EtOAc/Et3N 70:30:1 to 0:100:1 to EtOAc/
MeOH 50:50) but the product is poorly soluble, and it was eluted
from the column using MeOH. Some pure fractions of product were
collected for characterization. Yield of impure product (6.86 g,
(s, 2H), 6.44 (d, J ¼ 8.5 Hz, 1H). 13C NMR (DMSO-d6)
d 161.5, 151.5,
142.3, 117.0, 106.9, 93,5.
>99%). mp (decomp.) >91 ꢁC. 1H NMR (DMSO-d6)
d
8.37 (s, 1H), 7.37
4.1.18. 6-Amino-2-ethylnicotinonitrile (39)
(b s, 2H), 6.60 (s,1H). 13C NMR (DMSO-d6)
95.5.
d 154.9,143.4,116.2,107.1,
38 (1.00 g crude, max 4.30 mmol), diethylzinc (1.0 M in hexane,
7.81 mmol, 7.8 mL), and Pd2(PPh3)4 (150 mg, 0.13 mmol) were
added to a microwave vial along with NMP (8 mL). The vial was
flushed with N2, sealed and heated using microwave irradiation at
100 ꢁC for 30 min. The contents of the vial were cooled to rt and
washed with sat. aq. NaHCO3 (120 mL), and H2O (120 mL), dried
over Na2SO4, and evaporated under vacuum. Purification by FC
(heptane/EtOAc 80:20 to 50:50) afforded the product as brown
4.1.23. 6-Amino-4-ethylnicotinonitrile (45)
A flame-dried flask was charged with 44 (500 mg crude, max.
2.46 mmol), THF/NMP (10:1, 22 mL), and Fe(acac) (115 mg,
0.33 mmol). EtMgBr (3.0 M in Et2O, 2.71 mL, 8.10 mmol) was added
dropwise resulting in a colour change from red to brown to red/
violet. The reaction mixture was stirred for 2 h at rt, then sat. aq.

414
J.G. Petersen et al. / European Journal of Medicinal Chemistry 84 (2014) 404e416
NH4Cl (50 mL) and EtOAc (50 mL) were added. The organic phase
was separated, and the aqueous phase was extracted with EtOAc
(50 mL). The combined organic phases were washed with brine
(50 mL), dried over Mg2SO4, and evaporated under vacuum. Puri-
fication by DCVC (heptane/EtOAc 100:0 to 70:30) followed by
recrystallization from MeOH afforded the product as white crystals
(30 mg, 8% over two steps). mp 157.5e158.9 ꢁC. 1H NMR (CDCl3)
3.89 (s, 3H). 13C NMR (CDCl3)
117.7, 117.0, 52.2.
d 166.6, 159.0, 150.4, 136.2, 131.1, 118.9,
4.1.28. Methyl 6-amino-5-ethylnicotinate (51)
50 (100 mg, 0.56 mmol) was dissolved in THF (6 mL), 10% Pd/C
(50 mg) was added, and the solution was hydrogenated at rt for
16 h. The mixture was filtered through Celite, and the filtercake was
washed with MeOH. The organic phase was evaporated, and puri-
fication by DCVC (EtOAc/heptane 1:2) afforded the product as white
d
8.30 (s, 1H), 6.37 (s, 1H), 4.89 (b s, 1H), 2.73 (q, J ¼ 7.8 Hz, 2H), 1.28
(t, J ¼ 7.8, 3H). 13C NMR (CDCl3)
d 160.4, 157.1, 153.6, 117.4, 106.7,
99.4, 27.1, 13.5.
solid (94 mg, 93%). mp. 140e143 ꢁC. 1H NMR (CDCl3)
d 8.59 (d,
J ¼ 2.1 Hz, 1H), 7.87 (d, J ¼ 2.1 Hz, 1H), 4.97 (s, 2H), 3.87 (s, 3H), 2.46
(q, J ¼ 7 Hz, 2H),1.29 (t, J ¼ 7 Hz, 3H). 13C NMR (CDCl3)
d 165.7, 155.6,
4.1.24. 6-Amino-4-butylnicotinonitrile (46)
145.5, 133.4, 124.3, 113.3, 52.1, 23.7, 14.2.
Preparation of Grignard reagent: To a two-necked flask con-
taining Mg turnings (188 mg, 7.75 mmol) and THF (8.0 mL), BuBr
(0.88 mL, 8.15 mmol) was added dropwise. The reaction did not
start spontaneously upon addition of a few drops of BuBr, so a few
iodine crystals were added, and the reaction mixture was heated
using a heatgun. After the red colour of iodine had disappeared, the
remaining BuBr was added dropwise. The reaction mixture was
heated at 40e50 ꢁC for 30 min then cooled to rt. Cross coupling was
performed as described for 45 using 44 (1.00 g crude, max
4.91 mmol), Fe(acac) (230 mg, 0.65 mmol), and 2 consecutive ad-
ditions of BuMgBr (1.0 M in THF, 16.3 mL, 16.3 mmol and 3.26 mL,
3.26 mmol). In the work-up the order of addition of sat. aq. NH4Cl
and EtOAC was reversed. Purification by DCVC (heptane/EtOAc
80:20) afforded the product as white solid (213 mg, 21% over two
4.2. Methods for pKa determination of ionization constants
Titration was performed on a Sirius GLpKa Autotitrator (Sirius
Analytical Instruments Ltd, East Sussex, UK). Depending on the
structure two different detection techniques were used. If a com-
pound changes UVeVis absorption spectrum by the change of
charge, a Diode array detector was used (compounds 3,24,26). If no
change in the UVeVis absorption spectrum, traditional potentio-
metric detection was used (compounds 22,25,27).
4.2.1. pKa-determination by diode array detector [49]
The pKa values were determined by a series of three titrations on
1
ꢁC and ion
steps). mp 86.8e87.5 ꢁC. 1H NMR (CDCl3)
d 8.46 (s, 1H), 8.22 (s, 1H),
a dilution of 50
mL 10
mM compound stock at 24
strength of 0.17 M using methanol as co-solvent. Methanol con-
centrations in the range 23e48% were used in the three titrations
on each compound. The titrator was used in the mode with a diode-
array-detector. This detector recorded the UVeVis spectrum of the
solution in-line, and thus the titration can be mapped via com-
parison to the spectra of the protonated and unprotonated species.
The real aqueous pKa-value was determined by extrapolation to
zero-methanol content using a YasudaeShedlovsky plot.
8.05 (br s, 1H), 2.83 (m, 2H), 2.25 (s, 3H), 1.69 (m, 2H), 1.42 (sxt,
J ¼ 7.5 Hz, 2H), 0.97 (t, J ¼ 7.5 Hz, 3H). 13C NMR (CDCl3)
d 168.9,
158.2, 153.6, 151.8, 116.2, 113.2, 34.3, 32.0, 24.9, 22.3, 13.7.
4.1.25. 6-Amino-4-phenylniconitrile (47)
Prepared as described for 41 using 44 (600 mg crude, max
2.95 mmol), phenylboronic acid (715 mg, 5.86 mmol), K3PO4
(2.07 g, 9.80 mmol), and Pd(PPh3)4 (226 mg, 0.20 mmol). The re-
action mixture was heated at 100 ꢁC for 2 h using microwave
irradiation. Purification by DCVC (heptane/EtOAc 60:40 to 40:60)
afforded the product as yellow solid (269 mg, 47% over two steps).
4.2.2. pKa-determination by potentiometric detection [49]
The pKa values were determined by a series of three titration on
approx. 2.5 mg compound at 24
1
ꢁC and ion strength of 0.17 M
mp 180.8e183.0 ꢁC. 1H NMR (CDCl3)
d
8.45 (s, 1H), 7.61e7.46 (m,
161.32, 154.04,
using methanol as co-solvent. Methanol concentrations in the
range 23e48% were used in three titrations. A difference curve was
created from each of these titrations by blank subtraction, and from
these difference curves, methanol concentration dependent pKa
values were determined. The real aqueous pKa-value was deter-
mined by extrapolation to zero-methanol content using a Yasu-
daeShedlovsky plot.
5H), 6.55 (s, 1H), 5.02 (b s, 2H). 13C NMR (CDCl3)
d
151.19, 136.29, 128.79, 128.17, 127.46, 118.22, 107.05.
4.1.26. Methyl 6-amino-5-iodonicotinate (49)
Compound 48 (200 mg, 1.31 mmol) was dissolved in abs. EtOH
(15 mL), I2 (460 mg, 1.83 mmol) and AgSO4 (572 mg, 1.83 mmol)
were added in one portion. The reaction mixture was stirred at rt
for 24 h then evaporated under vacuum. Purification by DCVC
(CH2Cl2) afforded the product as off-white solid (208 mg, 57%). mp.
4.3. Pharmacology
161e164 ꢁC. 1H NMR (CDCl3)
d
8.65 (d, J ¼ 2.1 Hz, 1H), 8.44 (d,
Characterization of 1e27 in muscimol binding: The binding
assay was performed using rat brain synaptic membranes of cortex
and the central hemispheres from male SPRD rats with tissue
preparation as described in the literature [50]. On the day of the
experiment, the membrane preparation was quickly thawed, ho-
mogenized in 50 volumes of ice-cold buffer (50 mM TriseHCl
buffer, pH 7.4), and centrifuged at 48,000g for 10 min at 4 ꢁC. This
washing step was repeated four times and the final pellet was re-
suspended in buffer. The assay was carried out in 96-wells plates,
J ¼ 2.1 Hz, 1H), 5.45 (b s, 2H), 3.87 (s, 3H). 13C NMR (CDCl3)
d 165.1,
160.5, 151.1, 149.8, 148.3, 118.1, 52.4.
4.1.27. Methyl 6-amino-5-vinylnicotinate (50)
49 (291 mg, 1.04 mmol), potassium vinyltrifluoroborate
(280 mg, 2.08 mmol), K2CO3 (3 M, 0.7 mL, 2.10 mmol), and PdCl2
(PPh3)2 (40 mg, 0.05 mmol) were added degassed toluene/H2O
(10:1, 55 mL). The reaction mixture was heated at 100 ꢁC for 20 h
then cooled to rt. Et2O (100 mL) was added, and the organic phase
was washed with H2O (100 mL), sat. aq. NaHCO3 (100 mL), and H2O
(100 mL). The organic phase was dried over Mg2SO4 and evaporated
under vacuum. Purification by DCVC (heptane/EtOAc 4:1 to 2:1)
afforded the product as white solid (75 mg, 40%). 1H NMR (CDCl3)
by incubation of membranes (70ꢀ80
m
g protein) in 200
m
L buffer,
mL test
25
m
L [3H]muscimol (5 nM final concentration), and 25
substance in various concentrations, for 60 min at 0 ꢁC. The reaction
was terminated by rapid filtration through GF/C filters (Perkin
Elmer Life Sciences), using a 96 well Packard FilterMate cell-
d
8.65 (d, J ¼ 2.4 Hz,1H), 8.06 (d, J ¼ 2.4 Hz,1H), 6.58 (dd, J ¼ 17.1 Hz,
harvester, followed by washing with 3 ꢂ 250
mL of ice-cold buffer.
The dried filters were added Microscint scintillation fluid (Perkin
J ¼ 10.8 Hz, 1H), 5.74 (d, J ¼ 17.1 Hz, 1H), 5.47 (d, J ¼ 10.8 Hz, 1H),

J.G. Petersen et al. / European Journal of Medicinal Chemistry 84 (2014) 404e416
415
Elmer Life Sciences), and the amount of filterbound radioactivity
was quantified in a Packard TopCount microplate scintillator
counter. The experiments were performed in triplicate at least
three times for each compound. Non-specific binding was deter-
mined using 1.0 mM GABA. The binding data was analyzed by a
non-linear regression curve-fitting procedure using GraphPad
Prism v. 6.00 (GraphPad Software, CA, USA). IC50 values were
calculated from inhibition curves and converted to Ki values using
the modified ChengePrusoff equation [51].
Functional characterization in the FMP assay: The functional
characterization of compounds 3, 9e11, 14e16, 22, 23, 25, 27 at
GABAARs in the FLIPR™ Membrane Potential Blue assay was per-
formed essentially as described previously [52]. 8 ꢂ 105 tsA-
201 cells were split into a 6 cm tissue culture plate and transfected
file format using the included k2f utility and visualized in PyMOL
[56].
4.4.3. Molecular mechanics energy calculations
The relative potential energy of the torsional angle between the
carboxylate group and the pyridine/piperidine ring of 3 and 22 was
estimated based on a coordinate scan performed using Macro-
Model v.10.1 [57], the OPLS_2005 force field, and water solvent
model. The energy scan increment was set to 10ꢁ and otherwise
default settings were used.
Acknowledgements
We thank Shahrokh Pradah for technical laboratory assistance
and H. Lundbeck A/S for pKa determination. J.G.P. was supported by
the Lundbeck and Carlsberg Foundations, T.S. was supported by the
Danish Research Council, R.B. was supported by the Velux Foun-
dation. A.A.J. was supported by the Novo Nordisk Foundation.
the following day with a total of 5
Hilden, Germany). Cells were co-transfected with 1
pcDNA3.1, 1 b2-pcDNA3.1 and 3 g2S-pCDNA3.1, or transfected
with 5 r1-pcDNA3. The following day, cells were split into poly-
-lysine-coated black 96-wells plates with clear bottom (BD Bio-
sciences, Bedford, MA). 16e24 h later the medium was aspirated,
and the cells were washed with 100 L Krebs buffer [140 mM NaCl/
4.7 mM KCl/2.5 mM CaCl2/1.2 mM MgCl2/11 mM HEPES/10 mM
Glucose, pH 7.4]. 50 L Krebs buffer was added to the wells (in the
antagonist experiments, various concentrations of the antagonist
were dissolved in the buffer) and then an additional 50 L Krebs
mg cDNA using Polyfect (Qiagen,
mg
a1,2,3,5
-
mg
mg
mg
D
Appendix A. Supplementary data
m
D-
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.ejmech.2014.07.039.
m
m
References
buffer supplemented with assay dye (1 mg/mL) was added to each
well. Then the plate was incubated at 37 ꢁC in a humidified 5% CO2
incubator for 30 min and assayed in a NOVOstar™ plate reader
(BMG Labtechnologies, Offenburg, Germany) measuring emission
[in fluorescence units (FU)] at 560 nm caused by excitation at
[1] A.C. Foster, J.A. Kemp, Glutamate- and GABA-based CNS therapeutics, Curr.
Opin. Pharmacol. 6 (2006) 7e17.
[2] G.A.R. Johnston, GABAA receptor channel pharmacology, Curr. Pharm. Des. 11
(2005) 1867.
[3] W. Sieghart, G. Sperk, Subunit composition, distribution and function of
GABAA receptor subtypes, Curr. Top. Med. Chem. 2 (2002) 795e816.
[4] W. Paul J., GABAA receptor subtypes in the brain: a paradigm for CNS drug
discovery? Drug Discov. Today 8 (2003) 445e450.
[5] A.R.J. Graham, C. Mary, R.H. Jane, N.M. Kenneth, GABAC receptors as drug
targets, Curr. Drug Targets: CNS Neurol. Disord. 2 (2003) 260e268.
[6] J.G. Petersen, R. Bergmann, P. Krogsgaard-Larsen, T. Balle, B. Frølund, Probing
the orthosteric binding site of GABAA receptors with heterocyclic GABA car-
boxylic acid bioisosteres, Neurochem. Res. 39 (2014) 1005e1015.
[7] G.A.R. Johnston, D.R. Curtis, W.C. de Groat, A.W. Duggan, Central actions of
ibotenic acid and muscimol, Biochem. Pharmacol. 17 (1968) 2488e2489.
[8] P. Krogsgaard-Larsen, G.A.R. Johnston, Structure-activity studies on the inhi-
bition of GABA binding to rat brain membranes by muscimol and related
compounds, J. Neurochem. 30 (1978) 1377e1382.
530 nm before and up to 1 min after addition of 33 mL agonist so-
lution. The experiments were performed in duplicate at least three
times for each compound at each receptor. EC70eEC90 concentra-
tions of GABA were used as agonist in the antagonist experiments.
Concentrationeresponse curves for agonists and concen-
trationeinhibition curves for antagonists were constructed based
in the difference in the fluorescence units (△FU) between the
maximal fluorescence recording made before and after addition of
agonist obtained for different concentrations of the respective li-
gands. The curves were generated by nonweighted least-squares
fits using the program KaleidaGraph 3.6 (Synergy Software).
[9] P. Krogsgaard-Larsen, G.A.R. Johnston, D. Lodge, D.R. Curtis, A new class of
GABA agonist, Nature 268 (1977) 53e55.
[10] U. Kristiansen, J.D.C. Lambert, E. Falch, P. Krogsgaard-Larsen, Electrophysio-
logical studies of the GABAA receptor ligand, 4-PIOL, on cultured hippocampal
neurones, Br. J. Pharmacol. 104 (1991) 85e90.
[11] J.R. Byberg, I.M. Labouta, E. Falch, H. Hjeds, P. Krogsgaard-Larsen, Synthesis
and biological activity of a GABAA agonist which has no effect on benzodi-
azepine binding and of structurally related glycine antagonists, Drug Des.
Deliv. 1 (1987) 261e274.
[12] B. Frølund, A.T. Jørgensen, L. Tagmose, T.B. Stensbøl, H.T. Vestergaard,
C. Engblom, U. Kristiansen, C. Sanchez, P. Krogsgaard-Larsen, T. Liljefors, Novel
class of potent 4-arylalkyl substituted 3-isoxazolol GABAA antagonists: syn-
thesis, pharmacology, and molecular modeling, J. Med. Chem. 45 (2002)
2454e2468.
[13] B. Frølund, L. Tagmose, T. Liljefors, T.B. Stensbøll, C. Engblom, U. Kristiansen,
P. Krogsgaard-Larsen, A novel class of potent 3-isoxazolol GABAA antagonists:
design, synthesis, and pharmacology, J. Med. Chem. 43 (2000) 4930e4933.
[14] K. Brejc, W.J. van Dijk, R.V. Klaassen, M. Schuurmans, J. van der Oost, A.B. Smit,
T.K. Sixma, Crystal structure of an ACh-binding protein reveals the ligand-
binding domain of nicotinic receptors, Nature 411 (2001) 269e276.
[15] S.B. Hansen, T.T. Talley, Z. Radic, P. Taylor, Structural and ligand recognition
characteristics of an acetylcholine-binding protein from Aplysia californica,
J. Biol. Chem. 279 (2004) 24197e24202.
[16] P.H.N. Celie, R.V. Klaassen, S.E. van Rossum-Fikkert, R. van Elk, P. van Nierop,
A.B. Smit, T.K. Sixma, Crystal structure of acetylcholine-binding protein from
Bulinus truncatus reveals the conserved structural scaffold and sites of vari-
ation in nicotinic acetylcholine receptors, J. Biol. Chem. 280 (2005)
26457e26466.
[17] R.E. Hibbs, E. Gouaux, Principles of activation and permeation in an anion-
selective Cys-loop receptor, Nature 474 (2011) 54e60.
[18] R.J.C. Hilf, R. Dutzler, X-ray structure of a prokaryotic pentameric ligand-gated
ion channel, Nature 452 (2008) 375e379.
4.4. Computational chemistry
4.4.1. Ligand docking
The full-length homology model of the a1b2g2 GABA receptor
developed by Bergmann et al. [20] was used for generating binding
mode hypotheses. Compounds 3, 11e13, 15e16, 22, 25 and 27 were
docked into the GABA binding site between chains A and B using
the Induced Fit Docking protocol in Schrodinger's Maestro v9.5
[53]. Amino acid residues in an 8 Å distance from the docked li-
gands were sampled and otherwise default settings were used.
Poses were filtered on the basis of having a GABA like hydrogen-
bonding pattern, with hydrogen bonds from the carboxylate
group to a1Arg66, a1Thr129 and b2Thr202, and hydrogen bonds
from the amidine group to b2Glu155, and the backbone carbonyl of
b2Ser156 and b2Tyr157. A table showing docking scores is provided
as supporting information (SI Table S.1).
4.4.2. GRID
Molecular interaction fields were calculated for the homology
model using the methyl (C3) probe and 3 planes per Angstrom
using the Molecular Discovery's GRID v22c program [54,55]. The
obtained molecular interaction fields were converted to the .xplor

416
J.G. Petersen et al. / European Journal of Medicinal Chemistry 84 (2014) 404e416
[19] N. Bocquet, H. Nury, M. Baaden, C. Le Poupon, J.-P. Changeux, M. Delarue, P.-
J. Corringer, X-ray structure of a pentameric ligand-gated ion channel in an
apparently open conformation, Nature 457 (2009) 111e114.
[20] R. Bergmann, K. Kongsbak, P.L. Sørensen, T. Sander, T. Balle, A unified model of
the GABAA receptor comprising agonist and benzodiazepine binding sites,
PLoS One 8 (2013) e52323.
g-aminobutyric acid acting as selective GABAA antagonists, J. Med. Chem. 30
(1987) 239e249.
[38] D. Krehan, S. í Storustovu, T. Liljefors, B. Ebert, B. Nielsen, P. Krogsgaard-
Larsen, B. Frølund, Potent 4-arylalkyl-substituted 3-isothiazolol GABAA
competitive/noncompetitive antagonists: synthesis and pharmacology,
J. Med. Chem. 49 (2006) 1388e1396.
[21] T. Sander, B. Frølund, A.T. Bruun, I. Ivanov, J.A. McCammon, T. Balle, New in-
sights into the GABAA receptor structure and orthosteric ligand binding: re-
ceptor modeling guided by experimental data, Proteins Struct. Funct. Bioinf.
79 (2011) 1458e1477.
[39] C.L. Padgett, A.P. Hanek, H.A. Lester, D.A. Dougherty, S.C. Lummis, Unnatural
amino acid mutagenesis of the GABA(A) receptor binding site residues reveals
a novel cation-pi interaction between GABA and beta 2Tyr97, J. Neurosci. 27
(2007) 886e892.
[22] J.G. Petersen, R. Bergmann, H.A. Møller, C.G. Jørgensen, B. Nielsen, J. Kehler,
K. Frydenvang, J. Kristensen, T. Balle, A.A. Jensen, U. Kristiansen, B. Frølund,
Synthesis and biological evaluation of 4-(aminomethyl)-1-hydroxypyrazole
[40] J.G. Newell, R.A. McDevitt, C. Czajkowski, Mutation of glutamate 155 of the
GABAA receptor beta2 subunit produces a spontaneously open channel: a
trigger for channel activation, J. Neurosci. 24 (2004) 11226e11235.
[41] W. Haefliger, L. Revesz, R. Maurer, D. Romer, H.-H. Büscher, Analgesic GABA
agonists. Synthesis ans structure-activity studies on analogues and derivatives
of muscimol and THIP, Eur. J. Med. Chem. 19 (1984) 149e156.
[42] E. Falch, P. Jacobsen, K.-L. P., Synthesis, GABA agonist activity and effects on
GABA and diazepam binding of some N-substituted analogues of GABA 20
(1985) 447e453.
[43] B. Ebert, S.A. Thompson, K. Saounatsou, R. McKernan, P. Krogsgaard-Larsen,
K.A. Wafford, Differences in agonist/antagonist binding affinity and receptor
transduction using recombinant human g-aminobutyric acid type a receptors,
Mol. Pharmacol. 52 (1997) 1150e1156.
[44] B. Ebert, K.A. Wafford, P.J. Whiting, P. Krogsgaard-Larsen, J.A. Kemp, Molecular
pharmacology of gamma-aminobutyric acid type A receptor agonists and
partial agonists in oocytes injected with different alpha, beta, and gamma
receptor subunit combinations, Mol. Pharmacol. 46 (1994) 957e963.
ꢀ ꢀ
€
analogues of muscimol as
g-aminobutyric acid(A) receptor agonists, J. Med.
Chem. 56 (2013) 993e1006.
[23] P.E. Maligres, M.S. Waters, F. Fleitz, D. Askin, A highly catalytic robust palla-
dium catalyzed cyanation of aryl bromides, Tetrahedron Lett. 40 (1999)
8193e8195.
[24] J. Magano, A. Acciacca, A. Akin, B.M. Collman, B. Conway, M. Waldo, M.H. Chen,
K.E. Mennen, The synthesis of a dopamine D2 partial agonist for the treatment
of schizophrenia, Org. Process Res. Dev. 13 (2009) 555e566.
[25] P. Eastwood, J. Gonzalez, S. Paredes, A. Nueda, T. Domenech, J. Alberti, B. Vidal,
Discovery of N-(5,6-diarylpyridin-2-yl)amide derivatives as potent and se-
lective A2B adenosine receptor antagonists, Bioorg. Med. Chem. Lett. 20,
1697e1700.
[26] J.L. Bolliger, M. Oberholzer, C.M. Frech, Access to 2-aminopyridines e com-
pounds of great biological and chemical significance, Adv. Synth. Catal. 353,
945e954.
[27] A. Fürstner, A. Leitner, Iron-catalyzed cross-coupling reactions of alkyl-
grignard reagents with aryl chlorides, tosylates, and triflates, Angew. Chem.
Int. Ed. 41 (2002) 609e612.
[45] J.R. Atack, GABAA receptor subtype-selective efficacy: TPA023, an
lective non-sedating anxiolytic and 5IA, an 5 selective cognition enhancer,
CNS Neurosci. Ther. 14 (2008) 25e35.
a2/a3 se-
a
a
[46] M. Farrant, Z. Nusser, Variations on an inhibitory theme: phasic and tonic
activation of GABA(A) receptors, Nat. Rev. Neurosci. 6 (2005) 215e229.
[47] N. Collinson, F.M. Kuenzi, W. Jarolimek, K.A. Maubach, R. Cothliff, C. Sur,
A. Smith, F.M. Otu, O. Howell, J.R. Atack, R.M. McKernan, G.R. Seabrook,
G.R. Dawson, P.J. Whiting, T.W. Rosahl, Enhanced learning and memory and
altered GABAergic synaptic transmission in mice lacking the alpha 5 subunit
of the GABAA receptor, J. Neurosci. 22 (2002) 5572e5580.
[28] C. Koradin, W. Dohle, A.L. Rodriguez, B. Schmid, P. Knochel, Synthesis of
polyfunctional indoles and related heterocycles mediated by cesium and
potassium bases, Tetrahedron 59 (2003) 1571e1587.
[29] P.G. Dunbar, G.J. Durant, T. Rho, B. Ojo, J.J. Huzl III, D.A. Smith, A.A. El-Assadi,
S. Sbeih, D.O. Ngur, Design, synthesis, and neurochemical evaluation of 2-
amino-5-(alkoxycarbonyl)-3,4,5,6-tetrahydropyridines
and
2-Amino-5-
(alkoxycarbonyl)-1,4,5,6-tetrahydropyrimidines as M1 muscarinic receptor
agonists, J. Med. Chem. 37 (1994) 2774e2782.
[48] D.S. Pedersen, C. Rosenbohm, Dry column vacuum chromatography, Synthesis
(2001) 2431e2434.
[30] D. Cartwright, J.R. Ferguson, T. Giannopoulos, G. Varvounis, B.J. Wakerfield,
[49] J. Comer, Ionization constants and ionization profiles, in: D.J. Triggle,
J.B. Taylor (Eds.), Comprehensive Medicinal Chemistry II, Elsevier, Oxford,
2006, pp. 357e397.
Synthesis of some b-trichloromethyl-azines and -diazines, J. Chem. Soc. Perkin
Trans. 1 (1995) 2595e2597.
[31] B. Ebert, M. Mortensen, S.A. Thompson, J. Kehler, K.A. Wafford, P. Krogsgaard-
Larsen, Bioisosteric determinants for subtype selectivity of ligands for het-
eromeric GABAA receptors, Bioorg. Med. Chem. Lett. 11 (2001) 1573e1577.
[32] P. Krogsgaard-Larsen, H. Hjeds, E. Falch, F.J. Jørgensen, L. Nielsen, Recent ad-
vances in GABA agonists, antagonists and uptake inhibitors: structure-activity
relationships and therapeutic potential, Adv. Drug Res. (1988) 382e456.
[33] P. Krogsgaard-Larsen, H. Hjeds, D.R. Curtis, J.D. Leah, M.J. Peet, Glycine an-
tagonists structurally related to muscimol, THIP, or isoguvacine, J. Neurochem.
39 (1982) 1319e1324.
[50] R.W. Ransom, N.L. Stec, Cooperative modulation of [3H]MK-801 binding to the
N-methyl-d-aspartate receptor-ion channel complex by l-glutamate, glycine,
and polyamines, J. Neurochem. 51 (1988) 830e836.
[51] P. Leff, I.G. Dougall, Further concerns over Cheng-Prusoff analysis, Trends
Pharmacol. Sci. 14 (1993) 110e112.
[52] C. Madsen, A.A. Jensen, T. Liljefors, U. Kristiansen, B. Nielsen, C.P. Hansen,
M. Larsen, B. Ebert, B. Bang-Andersen, P. Krogsgaard-Larsen, B. Frolund, 5-
Substituted imidazole-4-acetic acid analogues: synthesis, modeling, and
pharmacological characterization of a series of novel gamma-aminobutyric
acid(C) receptor agonists, J. Med. Chem. 50 (2007) 4147e4161.
[34] A. Neu, H. Neuhoff, G. Trube, S. Fehr, K. Ullrich, J. Roeper, D. Isbrandt, Acti-
vation of GABAA receptors by guanidinoacetate: a novel pathophysiological
mechanism, Neurobiol. Dis. 11 (2002) 298e307.
€
[53] 2013-2, I.F.D. p. Glide Version 5.9, Prime Version 3.2, Schrodinger, LLC, New
York, NY, 2013.
[35] M. Chebib, N. Gavande, K. Wong, A. Park, I. Premoli, K. Mewett, R. Allan,
[54] P.J. Goodford, A computational procedure for determining energetically
R. Duke, G. Johnston, J. Hanrahan, Guanidino acids act as
antagonists, Neurochem. Res. 34 (2009) 1704e1711.
r1 GABAC receptor
favorable binding sites on biologically important macromolecules, J. Med.
Chem. 28 (1985) 849e857.
[36] M.M. Campbell, S.J. Mickel, G. Singh, ( )-2-Amino-2-thiazoline-4-ethanoic
acid; a novel, specific GABAA receptor agonist, Bioorg. Med. Chem. Lett. 1
(1991) 247e248.
[55] Molecular Discovery L, G., Pinner, Middelsex, UK, 2005.
€
[56] The PYMOL Molecular Graphics System, Version 1.6.0.0., L. Schrodinger, LLC.
€
[57] MacroModel v. 10.1, L. Schrodinger, N.Y., NY, 2013.
[37] C.G. Wermuth, J.J. Bourguignon, G. Schlewer, J.P. Gies, A. Schoenfelder,
A. Melikian, M.J. Bouchet, D. Chantreux, J.C. Molimard, Synthesis and
structure-activity relationships of a series of aminopyridazine derivatives of
[58] P. Krogsgaard-Larsen, L. Nielsen, E. Falch, D.R. Curtis, GABA agonists. Resolu-
tion, absolute stereochemistry and enantioselectivity of (S)-(þ)- and (R)-(-)
dihydromuscimol, J. Med. Chem. 28 (1985) 1612e1617.