(S)-2-Amino-3-(5-methyl-3-hydroxyisoxazol-4-yl)propanoic Acid (AMPA) and Kainate Receptor Ligands: Further Exploration of Bioisosteric Replacements and Structural and Biological Investigation

Article abstract of DOI:10.1021/acs.jmedchem.8b00099

Starting from 1-4 and 7 structural templates, analogues based on bioisosteric replacements (5a-c vs 1, 2 and 6 vs 7) were synthesized for completing the SAR analysis. Interesting binding properties at GluA2, GluK1, and GluK3 receptors were discovered. The requirements for GluK3 interaction were elucidated by determining the X-ray structures of the GluK3-LBD with 2 and 5c and by computational studies. Antinociceptive potential was demonstrated for GluK1 partial agonist 3 and antagonist 7 (2 mg/kg ip).

Full text of DOI:10.1021/acs.jmedchem.8b00099

Subscriber access provided by UNIVERSITY OF THE SUNSHINE COAST
Brief Article
(S)-2-Amino-3-(5-methyl-3-hydroxyisoxazol-4-yl)propanoic Acid
(AMPA) and Kainate Receptor Ligands: Further Exploration of
Bioisosteric Replacements, Structural and Biological Investigation
Simone Brogi, Margherita Brindisi, Stefania Butini, Giridhar Uttam Kshirsagar, Samuele
Maramai, Giulia Chemi, Sandra Gemma, Giuseppe Campiani, Ettore Novellino, Paolo
Fiorenzani, Jessica Pinassi, Anna Maria Aloisi, Mikko Gynther, Raminta Venskutonyte,
Liwei Han, Karla Frydenvang, Jette Sandholm Kastrup, and Darryl S. Pickering
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.8b00099 • Publication Date (Web): 16 Feb 2018
Downloaded from http://pubs.acs.org on February 16, 2018
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 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 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 7
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(S)-2-Amino-3-(5-methyl-3-hydroxyisoxazol-4-yl)propanoic
Acid (AMPA) and Kainate Receptor Ligands: Further
Exploration of Bioisosteric Replacements, Structural and
Biological Investigation
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
Simone Brogi,¶ Margherita Brindisi,¶ Stefania Butini,¶,* Giridhar U. Kshirsagar,¶ Samuele
Maramai,¶ Giulia Chemi,¶ Sandra Gemma,¶ Giuseppe Campiani,¶,* Ettore Novellino,∫ Paolo
Fiorenzani,# Jessica Pinassi,# Anna Maria Aloisi,# Mikko Gynther,∇ Raminta Venskutonytė,⊥ Liwei
⊥
⊥
⊥
⊥
,
,
Han, Karla Frydenvang, Jette Sandholm Kastrup, * Darryl S. Pickering *
¶Department of Biotechnology, Chemistry and Pharmacy, NatSynDrugs, University of Siena, via A. Moro 2, 53100
∫
Siena, Italy; Department of Pharmacy, University of Napoli Federico II, via D. Montesano 49, 80131, Napoli, Italy
#Department of Medicine, Surgery and Neuroscience, University of Siena, viale M. Bracci 16, 53100 Siena, Italy;
⊥
∇School of Pharmacy, Faculty of Health Sciences, University of Eastern Finland, 70211 Kuopio, Finland; Department
of Drug Design and Pharmacology, University of Copenhagen, Jagtvej 162, DK-2100 Copenhagen, Denmark.
KEYWORDS. Kainate receptors, binding pharmacology, crystallography, computational modeling, inflammatory pain
ABSTRACT: Starting from 1-4 and 7 structural templates, analogues based on bioisosteric replacements (5a-c vs 1, 2 and 6
vs 7) were synthesized for completing the SAR analysis. Interesting binding properties at GluA2, GluK1 and GluK3 receptors
were discovered. The requirements for GluK3 interaction were elucidated determining the X-ray structures of the GluK3-
LBD with 2 and 5c and by computational studies. Antinociceptive potential was demonstrated for GluK1 partial agonist 3
and antagonist 7 (2 mg/kg ip).
Introduction
signaling to the thalamus.8 GluK1 relevance as therapeutic
targets in this field was assessed by a number of animal
studies, mainly in models of chronic pain, where KARs
inhibition or their genetic ablation was effective in
reducing the pain behaviors. Small clinical trials were also
performed using several orthosteric antagonists.8
Chart 1. Reference compounds (1-4 and 7) and title
compounds 5a-c and 6.
L-Glutamate (L-Glu) mediates most of the fast excitatory
synaptic signaling in the central nervous system (CNS) by
means of heteromeric ligand-gated ion channels
(ionotropic L-Glu receptors, iGluRs).1 Kainate receptors
(KARs) are, together with (S)-2-amino-3-(5-methyl-3-
hydroxyisoxazol-4-yl)propanoic acid (AMPA) and N-
methyl-D-aspartate receptors, tetrameric ligand-gated ion
channels. All the identified KAR (GluK1-5) subunits can
combine in vitro to form heteromeric receptors,1,2 likely in
a dimer-of-dimers topology.3 KARs are distributed in
various areas of the CNS;4 they are localized on both pre-
and postsynaptic cell membranes contributing to fast
excitatory signaling and synaptic plasticity. Among iGluRs,
GluA2, GluK1 and GluK3 subtypes are implicated in various
neurological diseases such as depression, pain,
neurodegeneration, and epilepsy; though their normal
physiological function is not fully understood yet.5-7
Nociception (the sensorial input mainly travelling through
the spinothalamic tract) and pain (the conscious aspect)
are supported by ascending and descending pathways.
Nociception propagates to the dorsal horn of the spinal
cord, through the cell soma residing in the dorsal root
ganglion (where GluK1 receptors are widely expressed)
and is sent to the cortex for higher-order processing
(pain). From the brain, descending pathways travel
towards the spinal cord. These pathways may produce
analgesic effects by stimulation of opioid receptors and by
modulation of the GABAergic tone on the nociceptive
dorsal horn neurons.8 Further, the presynaptic GluK1
receptors are also relevant to the supraspinal pain
pathways as it has been demonstrated that GluK1
inhibition can attenuate, among others, the spinothalamic
With the specific aim to extend our previous work and
explore novel bicyclic pyrimidinedione-based compounds
typified by 19 and its analogues 2,10 310 and 411 (Chart 1),
and to investigate GluK3 interaction with ligands, the five
membered fused system was modified in compounds 5a-c
for completing the structure-activity relationship (SAR)
analysis of bioisosteric-replacements at the five membered
system. Further, our GluK1 antagonist 7 inspired the
development of derivative
6
(Chart 1). The 2-
hydroxyoxadiazole system was used as
a potential
bioisosteric alternative12 to the thiophene-2-carboxylate
system of 711 (Fig. S1).
RESULTS AND DISCUSSION
Chemistry. The chemistry for 5a-c and 6 is reported in
Schemes 1-3. For 5a-c (Schemes 1, 2) β-keto esters 913 and
1514,15 were needed for the synthesis of the key
pyrimidinediones 12, 20 and 21. After N1 alkylation of
these latter compounds and Boc deprotection of the amino
acidic lateral chain 5a-c were obtained. For the synthesis
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
of 6 (Scheme 3) cyclization of the ureido derivative 24
Page 2 of 7
In vitro Activity. Compounds 5a-c displayed micromolar
affinities at GluA2, GluK1 and GluK3 subunits (Table 1).
They also displayed from good to moderate selectivity for
GluK1 over GluK2, GluK3 and GluA2 with the exception of
5a which displayed a higher affinity for GluA2 (Ki GluA2 =
0.331 μM) than 5b (Ki GluA2 = 2.37 μM) and 5c (Ki GluA2 =
1.99 μM). It is noteworthy that while analogues 5b and 5c
displayed comparable affinity at GluK1 their affinity
towards GluK3 differed by at least one order of magnitude,
while 5a was found to have Ki > 100 µM at GluK3. The
order of affinity at AMPA receptors and KARs follows the
1
2
3
4
5
6
7
8
gave the pyrimidinedione 25 which underwent N1
alkylation (for a more detailed description please refers to
the Supporting Information, SI).
Scheme 1. Synthesis of 5aa
aReagents and conditions: (a) NaH, Et2O, 1 h, then methyl
acrylate, DMSO, 0 °C, 1 h, then rt, 5 h, 81%; (b) NH4OAc, EtOH,
reflux, 12 h, 32%; (c) phosgene 20% in toluene, pyridine, 1,2-
dichloroethane, 0 °C, 3 h, then NH4OH, reflux, 13 h, 19%; (d)
MeONa, MeOH, rt, 12 h, 52%; (e) urea, conc. HCl, MeOH,
reflux, 2 h, then 2 N NaOH, H2O, reflux 1.5 h, 45%; (f) NaH,
DMF, rt, 2 h, then -65 °C (S)-N-Boc-3-amino-2-oxetanone, rt,
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
trend
GluK1>GluA2>GluK3
(for
5b,c)
and
GluA2>GluK1>GluK3 (for 5a). The pharmacological profile
of 5a substantially differs from that of its furan-based
analogue (4). 4, compared to the other compounds, has a
lower affinity at GluA2 and GluK111 and is inactive at
inactive at GluK2 and GluK3 receptors. This observation
prompted an in-depth computational analysis, as discussed
below. By this series of ligands we observed that the
introduction of nitrogen in the pyrimidinedione-fused
system was the best choice to maintain GluK3 affinity
while lowering GluK1 and GluA2 affinity compared to the
other ligands. Accordingly, 5c was identified as the most
promising ligand for the study of the GluK3 receptor
subtype. Compound 6, with a structure inspired by our
antagonist 7,11 resulted a weak GluK1 ligand. We assume
that the observed low affinity is due to the short length of
the (mono-methylene) side-chain spacing the 1,3,4-
oxadiazole subunit from the pyrimidinedione system. This
may hamper the acidic proton from perfectly overlapping
the position occupied by the carboxylate moiety of our lead
compound 7. Accordingly, the hydroxyl functionality may
fail to establish suitable contacts within the ligand-binding
domain (LBD) of the target receptors as observed by
molecular modeling studies (not shown).
14 h, 8%; (g) TFA, DCM, rt, 4 h, 35%
.
Scheme 2. Synthesis of 5b,ca
aReagents and conditions: (a) NaH, ethyl acrylate, toluene, 0
°C, 0.5 h then, 60 °C, 4 h, 50%; (b) NH4OAc, EtOH, reflux, 12 h,
90%
;
(c) phosgene 20% in toluene, pyridine, 1,2-
dichloroethane, 0 °C, 3 h, then NH4OH, reflux, 13 h, 80%; (d)
EtONa, MeOH, rt, 12 h, 77%; (e) NaH, DMF for 19, DMF/DMSO
for 22, rt, 2 h, then -65 °C (S)-N-Boc-3-amino-2-oxetanone, rt,
14 h, 9% for 19 and 29% for 22; (f) TFA, DCM, rt, 4 h, 80%
for 5b and 50% for 5c; (g) HCl, MeOH, 0 °C, 0.5 h then, rt, 3 h,
91%; (h) formaldehyde, NaBH(OAc)3, EtOH, 13 h, 45%
.
X-ray Crystallography. X-ray crystallography was used to
investigate structural requirements for GluK3 ligands. The
two ligands with highest affinity at GluK3, 2 and 5c, were
crystallized in complex with GluK3-LBD. The selection
criteria were based on in vitro potency (5c) while 2 was
selected on the basis of structural similarity to the
previously developed compounds. The structure of GluK3-
LBD with 2, determined at 2.6 Å resolution, contains one
molecule of GluK3-LBD in the asymmetric unit of the
crystal (Table S1). The structure of GluK3-LBD with 5c,
determined at 2.4 Å resolution, contains two molecules of
GluK3-LBD in the asymmetric unit of the crystal. Both
structures correspond to GluK3-LBD in its monomeric
form (Fig. 1A). 2 and 5c interact in a similar way with the
LBD residues (Fig. 1B-C). The α-carboxylate of the ligands
forms H-bonds with the side-chain guanidino group of
Arg525 (GluK3 numbering according to UNP GRIK3_RAT
P42264) and the backbone nitrogen atoms of Thr520 and
Ala691. The α-amino group of the ligands interacts with
the side-chain of Glu739, the backbone oxygen atom of
Pro518 and the side-chain of Thr520. The two
pyrimidinedione oxygen atoms form contacts to the
backbone nitrogen atom of Thr692 and Glu739,
respectively, whereas the pyrimidinedione nitrogen atom
forms a contact to the side-chain hydroxyl group of
Thr692. The sulfur atom in 2 and N-methyl part in 5c are
located in a predominantly hydrophobic region within 5 Å
Scheme 3. Synthesis of 6a
aReagents and conditions: (a) methyl 4-aminothiophene-3-
carboxylate, triphosgene, toluene, reflux, 3 h, then 24, rt, 13 h,
34%; (b) MeONa, MeOH, rt, 12 h, 93%; (c) NaH, DMF, rt, 2 h,
then -65 °C, (S)-N-Boc-3-amino-2-oxetanone, rt, 14 h, 3%; (d)
TFA, DCM, rt, 4 h, 42%
.
Table 1. Binding affinities at recombinant homomeric rat
AMPA receptors and KARs expressed in Sf9 cells.
Ki (µM)
Selectivity
GluK3/GluK3/
GluK1 GluA2
Cmpds
GluA2
GluK1
GluK2
GluK3
19 0.223± 0.012 0.044± 0.004 >1000 5.09± 1.19 116
210 0.080± 0.013 0.0052±0.0003 >1000 3.25± 0.30 625
310 0.465± 0.002 0.0039± 0.0004 >1000 1.02± 0.25 262
23
41
21
411
44 ± 4
2.98 ± 0.52
>1000 > 1000 > 336 23
5a 0.331± 0.058 0.586 ± 0.116 >1000 111 ± 26 190
5b 2.37 ± 0.21 0.869 ± 0.063 >1000 18.1 ± 1.7 21
336
8
1
5c
1.99 ± 0.57 0.356 ± 0.023 >1000 2.15± 0.10
6
6
711
> 100
73 ± 15
36 ± 3
>1000
≈ 100
> 3
-
0.157 ± 0.022 99 ± 22 7.45± 0.30 47
0.1
Mean ± SEM of at least three competition experiments at 12-
16 drug concentrations performed in triplicate.
ACS Paragon Plus Environment

Page 3 of 7
Journal of Medicinal Chemistry
of the following residues: Glu443, Phe446, Tyr491, Pro518,
Asn722, Glu739, Thr742 and Tyr765 (not shown). The
sulfur atom of 2 forms a contact with the side-chain of
Tyr765 (Fig. 1B), whereas the N-methyl in 5c is in close
contact to the side-chain carboxylate of Glu443 (molA; Fig.
1C) or the side-chain amide of Asn722 (molB, Fig. 1D).
A similar, although slightly smaller, domain closure is seen
in molA of the GluK3-LBD structure with 5c (33°), whereas
the domain closure in molB of the GluK3-LBD structure
with 5c is only 28° (Fig. 1D). Such a difference in domain
closure between two GluK3-LBD molecules in the same
crystal has also been observed when crystallized with
kainate in presence of Zn ions (molA: 32°, molB 29°; PDB-
code 3U9219). When crystallized without Zn ions, kainate
induces a domain closure of 32° in GluK3-LBD (PBD-code
4E0W20). So, the introduction of a 5-membered ring fused
to the pyrimidinedione system seems to allow a similar
domain closure as L-Glu (Fig. 1E), but can also be stabilized
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
in
a
more opened conformation (Fig. 1F). The
pyrimidinedione system of 2 and 5c acts as a bioisoster of
the distal carboxylate of L-Glu and also displaces one water
molecule (Fig. 1E). In addition, the 5-membered ring in 2
and 5c displaces two extra water molecules. The sulfur
atom of 2 and the N-methyl of 5c seem to affect the
important D1-D2 interlobe interaction between the side-
chains of Glu443 and Asn722 (Fig. 1D,E). Whereas the
interaction between Glu443 and Asn722 is slightly
changed in the structure of GluK3-LBD with 2 compared to
the structure with L-Glu, Glu443 and Asn722 are poorly
defined in the structure of GluK3-LBD with 5c and need to
adopt a different conformation to avoid steric clash with
the N-methyl of 5c. So, the presence of the bulkier Asn772
in GluK3 compared to Ser721 in GluK1 might explain the
more favorable affinity at GluK1 compared to GluK3.
Molecular Modeling. We focused our efforts to rationalize
the observed differences in affinity for GluK1 and GluK3
receptors of the compounds 1-7. A comparative analysis of
GluK1 and GluK3 LBDs was performed using their X-ray
structures in complex with kainate (4E0X20 for GluK1 and
4E0W20 for GluK3). The two LBDs are highly similar in
amino-acid composition, although with crucial differences:
Ser741 and Ser721 in GluK1 are replaced by more
hindering Thr742 and Asn722 in GluK3. Another
difference is the replacement of Ser689 (GluK1) with
Ala691 (GluK3). These variations result in a different size
of the binding pockets (GluK1 volume 660.7 Å3, area
641Å2; GluK3 volume 567 Å3, area 520 Å2; Fig. S2). These
variations, influencing the accessible volume, can affect
ligand interaction.18,21 We performed molecular docking
(the protocol was validated as detailed in SI, Figs S3,S4) in
GluK1 and GluK3 LBDs to study the binding modes of our
compounds. Our calculations revealed that 2 establishes
similar interactions into both LBDs and targets the same
residues (Fig. 2A,B). It forms a salt bridge with Arg523 in
GluK1-LBD and Arg525 in GluK3-LBD as seen in its X-ray
structure (Fig. 1B), while the pyrimidinedione system
forms H-bonds with Thr residues (Thr518 and Thr690 in
GluK1; Thr520 and Thr692 in GluK3), with the backbone of
Pro516 in GluK1 and Pro518 in GluK3, Gly688 in GluK1
and Gly690 in GluK3 and with Met737 in GluK1 and
Met738 in GluK3 as well as with the side-chain carboxylate
of Glu738 in GluK1 and Glu739 in GluK3. Moreover, we
observed H-bonds with the backbone of Ala691 in GluK3.
Due to the presence of Ser689 in GluK1 in place of Ala691
in GluK3 we noted a strong network of H-bonds between 2
and GluK1. Compound 2 can form H-bonds with Ser689
(backbone and side-chain) as well as an additional polar
contact with Thr690. Due to the shrinkage of GluK3-LBD
Figure 1. X-ray crystal structures of GluK3-LBD with 2 and 5c
(PDB-codes 6F29 and 6F28, respectively). (A) Cartoon
representation of GluK3-LBD in complex with 2 (sticks, with C
in orange, N in blue, O in red and S in yellow), lobe D1 in light
blue and lobe D2 in dark green. The arrow indicates D1-D2
domain closure. (B) Binding site interactions with 2. Polar
contacts (up to 3.2 Å (except from the ligand to Thr692: 3.4
Å); black stippled lines) between binding-site residues and 2
are shown. A simple PHENIX omit 2Fo-Fc map carved around
2 and contoured at 1 sigma is shown in light grey. (C) Binding
site interactions with 5c (molA, sticks with yellow carbon
atoms. (D) Superimposition (on D1 residues) of molA and
molB of GluK3-LBD with 5c to illustrate the differences in
domain closure and conformation of Glu443 and Asn722. 5c
in sticks with yellow (molA) and grey (molB) carbon atoms.
(E) Comparison of binding mode of 2, 5c (molA), L-Glu (green;
PDB-code 4MH5) and kainate (cyan; PDB-code 3U92, molA).
Proteins were superimposed on D1 residues. GluK3 residues
and D1-D2 interlobe contacts are colored according to the
ligands. The three water molecules belonging to the structure
with L-Glu are shown as black spheres and the two water
molecules belonging to the structure with kainate as cyan
spheres. (F) Comparison of binding mode of 5c (molB) and
kainate (molB, cyan).
The closures of lobe D1 and D2 (named domain closure;
see Fig. 1A) introduced by binding of the ligands were
calculated using DynDom16 and are relative to an
antagonist-bound structure of GluK1-LBD (PDB-code 3S2V,
molB) as described.17 Compound 2 induced a domain
closure of 35° (Fig. 1D), similar to that induced by L-Glu in
GluK3-LBD (without zinc ions present: 36°; PDB-code
4MH5;18 with zinc ions present: 35-36°; PDB-code 3U9319).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
we noted a reduction in H-bonds network precluding
Page 4 of 7
for the volume estimation (see SI). After analyzing the
results (Table S2) we noted a correlation between the
molecules’ volume and their binding affinities. The only
exception is 5c which contains an additional carbon atom
on the pyrrolidine fused system (the Me group). As
expected, 4 has the lowest volume, among the tested
ligands. This data allowed us to speculate that the small
size of 4 may lead to two different scenarios as captured by
docking calculation: (i) the compound could flip when it is
already inside the receptor or (ii) during the entrance in
the LBD, the compound is accommodated in a non-classical
and less effective way, due to its different size, with
consequent reduction in binding affinity.
1
2
3
4
5
6
7
8
optimal accommodation of 2 into the LBD. These little
differences of accommodation observed in the docking
poses may help to rationalize the differences in binding
affinities of 2 at GluK1 and GluK3. Regarding 3 (Fig. S5) we
obtained comparable docking results as for 2 with a
reduction of the number of contacts in the GluK3-LBD. The
replacement of the sulfur atom of 2 with NH and N-Me
groups in 5b and 5c causes a decrease in binding affinity at
GluK1 (Table 1) while 5c shows a slightly better binding
affinity at GluK3 compared to 2. The docking of 5b into
both receptors (Fig. S6) gives similar outputs with minor
differences in terms of contacts, highlighting the
interaction of the NH group with Glu441 and Ser741 in
GluK1 and Glu443 and Thr742 in GluK3. Docking outputs
of 5c (Fig. S7A,B) showed similar interactions of the
amino-acid moiety as those described for 2. Contrarily, the
pyrimidinedione moiety of 5c is differently accommodated
into GluK1-LBD, interacting with the backbones of Thr690
and Met737, lacking the H-bond with the side-chain of
Thr690. The N atom of the N-Me group establishes an H-
bond with the side-chain of Ser741 (Fig. S7A). The binding
mode of 5c in GluK3-LBD (Fig. S7B) revealed interactions
with Thr692 and Met738. 5a docked in a similar manner in
GluK1-LBD and GluK3-LBD established a similar pattern of
interaction (Fig. S8). Hence, the docking calculations are
not able to explain the favorable binding affinity of the
compound against GluK1 over GluK3. Interestingly,
docking of 4 into GluK1 and GluK3-LBDs revealed different
binding modes. We observed that when adopting an
orientation into the LBD similar to that seen for 2 and 5c in
their X-ray structures, the pyrimidinedione system of 4 is
slightly twisted in an orientation that precludes key
contacts in the LBD (Fig. S9). Besides these results we
found, for 4, another accommodation where the ligand is
completely flipped, missing the “correct” orientation and
position of H-bonding atoms for the optimal interaction
within the receptor (Fig. S9). Remarkably, we got this
result only for 4. For the other analogues we found a
unique and reliable cluster of docked solutions (in both
receptors) where the pyrimidinedione system is, as usual,
projected toward Glu441, Ser741 and Tyr764 in GluK1 and
Glu443, Thr742 and Tyr765 in GluK3. Then, the undefined
binding mode of 4 may provide a reason for its reduced
binding affinity at GluK1 and GluK3 receptors.22 We also
estimated ligand-binding energies (ΔGbind) at GluK1 and
GluK3 using Prime MM/GBSA (Table S2). At GluK1 the best
predicted ΔGbind values were obtained for 2 and 3, followed
by other ligands that exhibited nM affinities (1 and 5a-c).
In agreement with experimental data for 4, the calculation
also revealed that it possesses the poorest affinity at the
GluK1 receptor. Interestingly, we noted that the flipped
orientation showed the highest ΔGbind, denoting that this
orientation may contribute to a dramatic reduction in
GluK1 binding affinity. A similar trend was also observed
at GluK3. We noted a significant increase in ΔGbind for the
less active compounds such as 5a with respect to the most
active ligands. Also in this case 4 showed the lowest
predicted affinity. For further substantiating the above
observations, we evaluated the molecules’ volume for
correlating it with binding affinities.23 Calculations were
carried out using different software with diverse methods
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. Representative docked poses of 2 (bright orange
sticks) into GluK1-LBD (PDB-code 4E0X; green cartoon) (A)
and GluK3-LBD (PDB-code 4NWC; light blue cartoon) (B). LBD
residues are represented as sticks. H-bonds are represented
by solid blue lines and the salt bridges as a grey dotted line
between two spheres.
In vivo Studies. The class of pyrimidinediones typified by
CPW399 (1), and represented also by the analogues
herein, was never investigated in animal models of pain.
Thus, in order to characterize this class of GluK1 ligands,
the most potent and selective analogues of the series were
selected to enter into in vivo studies. For in vivo tests the
GluK1 partial agonist 3 and the antagonist 711 were
selected and enrolled, while the new, less potent, analogs
5a-c and 6 were not investigated.
The formalin test is a well-known method to study the
behavioral response to an acute, long lasting painful
stimulation.24 The possibility to record different behavioral
responses to the painful stimulation allows a deeper
insight into pain mechanisms. Indeed, the time spent
licking, flexing or jerking the injected paw depends on the
activation of different circuits at supraspinal and/or spinal
levels. Moreover, the formalin test leaves the animal free to
move in a large cage, thus allowing the recording of other
spontaneous, not painful responses, important for
a
comprehensive analysis of the drug-induced effect on the
animal behavior. The formalin concentration used (5%)
induced a first phase of pain behavior (10-15 min, first
phase) followed by a longer lasting phase (45-50 min). The
statistical analysis was carried out in the two phases
separately to better determine the effects of the drugs on
the first or second phase.24 For evaluating the therapeutic
potential of our GluK1 ligands in the formalin test we
resynthesized compounds 3 and 7.11 Collected data show
how 3 and 7 efficiently reduce the pain sensation already
at 2 mg/kg dose i.p. (Fig. 3). The formalin-treated mice
showed typical behaviors: licking, flexing, and paw jerk.
These responses showed a classical biphasic time course
characterized by an initial rush of pain behavior (0-15
ACS Paragon Plus Environment

Page 5 of 7
Journal of Medicinal Chemistry
min) followed by a second phase characterized by longer-
compartment) of 7 can be achieved despite the low brain
permeability.25
1
lasting inflammatory events. The administration of 3 and 7
decreased the pain sensation as expressed by a statistically
significant reduction of the pain behaviors. We flanked the
classical pain-elicited behavioral observation with the
locomotor activity evaluation for the same animals (Fig.
S10, comparing the effect of 3 and 7 with formalin-treated
and naïve groups). The general locomotor activity (walking
and explorative behavior) was assessed by recording the
time spent in movement for 60 min after treatment. The
data revealed that in all the groups the early explorative
phase usually undergoes a time-dependent decrease up to
extinction. In the groups treated with 3 and 7 the initial
locomotor activity paralleled that of the formalin treated
animals. Interestingly, starting from 30 min of observation,
the registered increase of locomotor activity nicely
correlates with the analgesic profile of the tested
compounds as depicted in Fig. 3. In particular, with 3
starting from 45 min of observation, the animals, almost
devoid of pain, seem to recover the exploration phase.
Collectively, in vivo studies highlighted that the analgesic
effect is present in all the observed behaviors suggesting a
central effect localized not only at the spinal level but also
at supraspinal levels. Since the effect is present in both
formalin phases, we hypothesize that analgesia blocks both
the nociceptive input as well as the longer term effect
induced by formalin injection. These data further support
GluK1 partial agonists (such as 3)11 and antagonists (such
as 7)11 as valuable option for the effective treatment of
pain although with different behavioral outcomes.
2
3
4
5
6
7
8
9
CONCLUSION
We further extended the structural and biological
investigation of the class of pyrimidinediones structurally
related to 1 as AMPA and KA receptors ligands. We
described herein the synthesis, X-ray structural studies
and biological evaluation of a new set of analogues 5a-c
and 6. Specifically, we sought to explore focused
bioisosteric replacements of the condensed heterocyclic
moiety. Computational studies provided a rationalization
of their binding properties. The new analogues 5a-c
preserved GluK1 affinity over GluA2 (for 5b,c) with 5c
showing increased affinity for GluK3 (GluK3/GluK1 affinity
ratio 6.0 for 5c; and 625 for 2). Brain permeability studies
proved the capability of 7 to cross the BBB and, as GluK1
subtypes are involved in sensory transmission, the
biological characterization of the series was continued
assessing the analgesic potential in the formalin test of 3
and 7. Both the antagonist 7 and the partial agonist 3
proved efficacy at 2 mg/kg, i.p. The present work provides
further insights into the structural requirements for the
design of AMPA receptor and KAR subtype selective
ligands as well as additional evidences about the analgesic
potential of GluK1/GluK3 ligands.
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
Experimental Section
Final compounds were analyzed by combustion analysis
(CHN) to confirm purity >95%.
(S)-1-[2'-Amino-2'-carboxyethyl]-5,7-dihydrofuro[3,4-
d]pyrimidine-2,4(1H,3H)-dione (5a). To a suspension of
13 (121 mg, 0.355 mmol) in dry DCM (10 mL), TFA (0.55
mL) was added dropwise at 0 °C and the mixture was
stirred 4 h at rt. After evaporation, the crude was purified
by ion-exchange resin (Dowex 50WX 8-400), using a 1:1
mixture of water/EtOH, then a 1% to 10% gradient of
diluted NH3 in water, to recover 5a as a white solid (30 mg,
1
0.124 mmol, 35% yield). H NMR (400 MHz, D2O + DCl) δ
4.94 (t, J = 3.5 Hz, 2H), 4.79 (t, J = 3.5 Hz, 2H), 4.37-4.30 (m,
1H) 4.14-4.02 (m, 2H); 13C NMR (100 MHz, D2O + DCl) δ
168.9, 161.1, 154.8, 153.7, 109.1, 71.7, 71.0, 51.8, 46.0; ESI-
MS m/z 242 [M+H]+. Anal. (C9H11N3O5) C, H, N.
(S)-1-[2'-Amino-2'-carboxyethyl]-5,7-
dihydropyrrolo[3,4-d]pyrimidin-2,4(1H,3H)-dione
(5b). Starting from 19 (48 mg, 0.109 mmol) and as
described for 5a, title compound was obtained as white
1
solid (21 mg, 0.087 mmol, 80% yield). H NMR (300 MHz,
D2O) δ 4.27-4.08 (m, 1H), 4.02 (d, J = 6.0 Hz, 2H), 3.95-3.84
(m, 4H). 13C NMR (75 MHz, D2O) δ 172.4, 162.5, 156.7,
154.3, 110.2, 54.2, 50.7, 48.5, 47.6. ESI-MS m/z 241
[M+H]+. Anal. (C9H12N4O4) C, H, N.
(S)-1-[2'-Amino-2'-carboxyethyl]-6-methyl-5,7-
dihydropyrrolo[3,4-d]pyrimidin-2,4(1H,3H)-dione
(5c). Starting from 22 (31 mg, 0.087 mmol) and as
described for 5a, title compound was obtained as white
Figure 3. Behavioral responses to formalin injection (5%, 50
µL, dorsal paw), first (0-15’) and second (15’-60’) phases;
animal experiments were conducted in compliance with
institutional guidelines. Licking (A) and flexing (B) duration;
(C) paw jerk frequency. #p<0.01, ##p<0.001 vs control.
Brain permeability of 7 was also evaluated using in situ rat
brain perfusion technique. The obtained data showed that
1
solid (11 mg, 0.043 mmol, 50% yield). H NMR (300 MHz,
7 can cross the rat BBB with a permeability of 5.9 × 10-7
±
D2O) δ 4.09-3.95 (m, 4H), 3.89 (d, J = 5.1 Hz, 1H), 3.69 (s,
2H), 2.43 (s, 3H). 13C NMR (75 MHz, D2O) δ 171.4, 162.1,
155.5, 154.0, 109.4, 58.4, 56.2, 54.1, 47.1, 41.7. ESI-MS m/z
253 [M-H]‾. Anal. (C10H14N4O4) C, H, N.
(S)-1-[2'-Amino-2'-carboxyethyl]-3-((5-oxo-4,5-
dihydro-1,3,4-oxadiazol-2-yl)methyl)thieno[3,4-
2.1 × 10-7 cm/s (mean ± sem, n = 3). The low permeability
was expected for a polar molecule like 7. As high polarity
of compounds is correlated with high unbound fraction
and low distribution volume in the brain, effective
concentration at the target site (brain extracellular
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 7
8. Bhangoo, S. K.; Swanson, G. T. Kainate receptor signaling in pain
pathways. Mol. Pharmacol. 2013, 83, 307-315.
d]pyrimidine-2,4(1H,3H)-dione (6). Starting from 26
(31 mg, 0.067 mmol) and as described for 5a, title
compound was obtained as white solid (10 mg, 0.34 mmol,
1
2
3
4
5
6
7
8
9. Campiani, G.; Morelli, E.; Nacci, V.; Fattorusso, C.; Ramunno, A.;
Novellino, E.; Greenwood, J.; Liljefors, T.; Griffiths, R.; Sinclair, C.;
Reavy, H.; Kristensen, A. S.; Pickering, D. S.; Schousboe, A.; Cagnotto,
A.; Fumagalli, E.; Mennini, T. Characterization of the 1H-
cyclopentapyrimidine-2,4(1H,3H)-dione derivative (S)-CPW399 as a
novel, potent, and subtype-selective AMPA receptor full agonist with
partial desensitization properties. J. Med. Chem. 2001, 44, 4501-4504.
10. Butini, S.; Pickering, D. S.; Morelli, E.; Coccone, S. S.; Trotta, F.; De
Angelis, M.; Guarino, E.; Fiorini, I.; Campiani, G.; Novellino, E.;
Schousboe, A.; Christensen, J. K.; Gemma, S. 1H-cyclopentapyrimidine-
2,4(1H,3H)-dione-related ionotropic glutamate receptors ligands.
structure-activity relationships and identification of potent and
Selective iGluR5 modulators. J. Med. Chem. 2008, 51, 6614-6618.
11. Venskutonyte, R.; Butini, S.; Coccone, S. S.; Gemma, S.; Brindisi, M.;
Kumar, V.; Guarino, E.; Maramai, S.; Valenti, S.; Amir, A.; Valades, E. A.;
Frydenvang, K.; Kastrup, J. S.; Novellino, E.; Campiani, G.; Pickering, D.
S. Selective kainate receptor (GluK1) ligands structurally based upon
1H-cyclopentapyrimidin-2,4(1H,3H)-dione: synthesis, molecular
modeling, and pharmacological and biostructural characterization. J.
Med. Chem. 2011, 54, 4793-4805.
1
42% yield). H NMR (300 MHz, D2O) δ 8.18 (d, J = 3.2 Hz,
1H), 6.77 (d, J = 3.2 Hz, 1H), 4.59 (d, J = 2.3 Hz, 2H), 4.13-
4.07 (m, 1H), 3.83-3.77 (m, 1H), 3.52-342 (m, 1H); 13C NMR
(300 MHz, D2O) δ 171.8, 169.9, 162.3, 160.0, 151.8, 135.1,
131.6, 120.2, 104.5, 52.2, 50.7, 41.8. ESI-MS m/z 352 [M-H]-
. Anal. (C12H11N5O6S) C, H, N.
9
PDB ACCESSION CODES
6F29 (2) and 6F28 (5c). Authors will release the atomic
coordinates and experimental data upon article publication.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ASSOCIATED CONTENT
Supporting Information. Experimental procedures, Figures
and Tables for computational and X-ray studies and molecular
formula strings. This material is available free of charge via
the Internet at http://pubs.acs.org
12. Morelli, E.; Gemma, S.; Budriesi, R.; Campiani, G.; Novellino, E.;
Fattorusso, C.; Catalanotti, B.; Coccone, S. S.; Ros, S.; Borrelli, G.;
Persico, M.; Fiorini, I.; Nacci, V.; Ioan, P.; Chiarini, A.; Hamon, M.;
Cagnotto, A.; Mennini, T.; Fracasso, C.; Colovic, M.; Caccia, S.; Butini, S.
Specific targeting of peripheral serotonin 5-HT(3) receptors.
AUTHOR INFORMATION
Corresponding Author
*Stefania Butini email: butini3@unisi.it; Giuseppe Campiani
email: campiani@unisi.it
*For pharmacology, email: picker@sund.ku.dk
*For crystallography, email: jsk@sund.ku.dk
Author Contributions
The manuscript was written with contributions of all authors.
All authors approved the final version of the manuscript.
Funding Sources
The Lundbeck Foundation (R.V., J.S.K.), GluTarget (R.V., L.H.,
K.F., J.S.K., D.S.P.), Danscatt (R.V., K.F., J.S.K).
Synthesis,
biological
investigation,
and
structure-activity
relationships. J Med Chem 2009, 52, 3548-3562.
13. Schönbrunn, E.; Lawrence, N. J.; Lawrence, H. R. Potent Dual Brd4-
kinase Inhibitors as Cancer Therapeutics. WO 2016022460 A1, Feb
11, 2016.
14. Chavan, S. P.; Dhawane, A. N.; Kalkote, U. R. Tandem Aza-Michael-
Condensation-Aldol Cyclization Reaction: Approach to the
Construction of DE Synthon of (+/-)-Camptothecin. Synlett. 2008, 18,
2781-2784.
15. Simpson, G. L.; Gordon, A. H.; Lindsay, D. M.; Promsawan, N.;
Crump, M. P.; Mulholland, K.; Hayter, B. R.; Gallagher, T. Glycosylated
foldamers to probe the carbohydrate-carbohydrate interaction. J. Am.
Chem. Soc. 2006, 128, 10638-10639.
16. Hayward, S.; Berendsen, H. J. Systematic analysis of domain
motions in proteins from conformational change: new results on
citrate synthase and T4 lysozyme. Proteins 1998, 30, 144-154.
17. Mollerud, S.; Frydenvang, K.; Pickering, D. S.; Kastrup, J. S. Lessons
from crystal structures of kainate receptors. Neuropharmacology
2017, 112, 16-28.
18. Venskutonyte, R.; Frydenvang, K.; Gajhede, M.; Bunch, L.;
Pickering, D. S.; Kastrup, J. S. Binding site and interlobe interactions of
the ionotropic glutamate receptor GluK3 ligand binding domain
revealed by high resolution crystal structure in complex with (S)-
glutamate. J. Struct. Biol. 2011, 176, 307-314.
19. Veran, J.; Kumar, J.; Pinheiro, P. S.; Athane, A.; Mayer, M. L.; Perrais,
D.; Mulle, C. Zinc potentiates GluK3 glutamate receptor function by
stabilizing the ligand binding domain dimer interface. Neuron 2012,
76, 565-578.
20. Venskutonyte, R.; Frydenvang, K.; Hald, H.; Rabassa, A. C.; Gajhede,
M.; Ahring, P. K.; Kastrup, J. S. Kainate induces various domain
closures in AMPA and kainate receptors. Neurochem. Int. 2012, 61,
536-545.
21. Mayer, M. L. Crystal structures of the GluR5 and GluR6 ligand
binding cores: molecular mechanisms underlying kainate receptor
selectivity. Neuron 2005, 45, 539-552.
ACKNOWLEDGMENT
We thank MAX-lab, Lund, Sweden for providing beamtime and
beamline scientists for their help (R.V., K.F., J.S.K.). We thank
Dr Andrew Orry (Molsoft LLC) for the trial license of ICM-Pro.
ABBREVIATIONS
AMPA,
yl)propanoic
(S)-2-amino-3-(5-methyl-3-hydroxyisoxazol-4-
acid; BBB, blood-brain-barrier; DCM,
dichloromethane; DMF, N,N-dimethylformamide; DMSO,
dimethylsulfoxide; LBD, ligand-binding domain; iGluRs,
ionotropic glutamate receptors; KARs, kainate receptors; L-
Glu, L-glutamate; TFA, trifluoroacetic acid.
REFERENCES
1. Traynelis, S. F.; Wollmuth, L. P.; McBain, C. J.; Menniti, F. S.; Vance, K.
M.; Ogden, K. K.; Hansen, K. B.; Yuan, H.; Myers, S. J.; Dingledine, R.
Glutamate receptor ion channels: structure, regulation, and function.
Pharmacol. Rev. 2010, 62, 405-496.
2. Pinheiro, P. S.; Perrais, D.; Coussen, F.; Barhanin, J.; Bettler, B.;
Mann, J. R.; Malva, J. O.; Heinemann, S. F.; Mulle, C. GluR7 is an
essential subunit of presynaptic kainate autoreceptors at
hippocampal mossy fiber synapses. Proc. Natl. Acad. Sci. U. S. A. 2007,
104, 12181-12186.
3. Sobolevsky, A. I.; Rosconi, M. P.; Gouaux, E. X-ray structure,
symmetry and mechanism of an AMPA-subtype glutamate receptor.
Nature 2009, 462, 745-756.
4. Huettner, J. E. Kainate receptors and synaptic transmission. Prog.
Neurobiol. 2003, 70, 387-407.
22. Vistoli, G.; Mazzolari, A.; Testa, B.; Pedretti, A. Binding space
concept: a new approach to enhance the reliability of docking scores
and its application to predicting butyrylcholinesterase hydrolytic
activity. J. Chem. Inf. Model. 2017, 57, 1691-1702.
23. Maicheen, C.; Phosrithong, N.; Ungwitayatorn, J. Biological activity
evaluation and molecular docking study of chromone derivatives as
cyclooxygenase-2 inhibitors. Med. Chem. Res. 2017, 26, 662-671.
24. Capone, F.; Aloisi, A. M. Refinement of pain evaluation techniques.
The formalin test. Ann. Ist. Super. Sanita 2004, 40, 223-229.
25. Gynther, M.; Petsalo, A.; Hansen, S. H.; Bunch, L.; Pickering, D. S.
Blood-brain barrier permeability and brain uptake mechanism of
kainic acid and dihydrokainic acid. Neurochem Res 2015, 40, 542-549.
5. Jane, D. E.; Lodge, D.; Collingridge, G. L. Kainate receptors:
pharmacology,
function
and
therapeutic
potential.
Neuropharmacology 2009, 56, 90-113.
6. Lerma, J. Kainate receptor physiology. Curr. Opin. Pharmacol. 2006,
6, 89-97.
7. Matute, C. Therapeutic potential of kainate receptors. CNS Neurosci.
Ther. 2011, 17, 661-669.
ACS Paragon Plus Environment

Page 7 of 7
Journal of Medicinal Chemistry
SYNOPSIS TOC
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
7
ACS Paragon Plus Environment