Enhancing Action of Positive Allosteric Modulators through the Design of Dimeric Compounds

Article abstract of DOI:10.1021/acs.jmedchem.8b00250

The present study describes the identification of highly potent dimeric 1,2,4-benzothiadiazine 1,1-dioxide (BTD)-type positive allosteric modulators of the AMPA receptors (AMPApams) obtained by linking two monomeric BTD scaffolds through their respective 6-positions. Using previous X-ray data from monomeric BTDs cocrystallized with the GluA2 ligand-binding domain (LBD), a molecular modeling approach was performed to predict the preferred dimeric combinations. Two 6,6-ethylene-linked dimeric BTD compounds (16 and 22) were prepared and evaluated as AMPApams on HEK293 cells expressing GluA2_o(Q) (calcium flux experiment). These compounds were found to be about 10,000 times more potent than their respective monomers, the most active dimeric compound being the bis-4-cyclopropyl-substituted compound 22 [6,6′-(ethane-1,2-diyl)bis(4-cyclopropyl-3,4-dihydro-2H-1,2,4-benzothiadiazine 1,1-dioxide], with an EC₅₀ value of 1.4 nM. As a proof of concept, the bis-4-methyl-substituted dimeric compound 16 (EC₅₀ = 13 nM) was successfully cocrystallized with the GluA2_o-LBD and was found to occupy the two BTD binding sites at the LBD dimer interface.

Full text of DOI:10.1021/acs.jmedchem.8b00250

Subscriber access provided by Kaohsiung Medical University
Article
Enhancing action of positive allosteric modulators
through the design of dimeric compounds
Thomas Drapier, Pierre Geubelle, Charlotte Bouckaert, Lise Nielsen, Saara Laulumaa, Eric Goffin, Sébastien
Dilly, Pierre Francotte, Julien Hanson, Lionel Pochet, Jette Sandholm Kastrup, and Bernard Pirotte
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.8b00250 • Publication Date (Web): 18 May 2018
Downloaded from http://pubs.acs.org on May 18, 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.
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 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Enhancing action of positive allosteric modulators
through the design of dimeric compounds
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
Thomas Drapier1, Pierre Geubelle1,2, Charlotte Bouckaert3, Lise Nielsen4, Saara Laulumaa4,
Eric Goffin1, Sébastien Dilly1, Pierre Francotte1, Julien Hanson1,2*, Lionel Pochet3*, Jette
Sandholm Kastrup4*, Bernard Pirotte1*
1 Laboratory of Medicinal Chemistry, Center for Interdisciplinary Research on Medicines (CIRM), ULiège, Quartier
Hôpital – Avenue Hippocrate, 15, B36, Bꢀ4000 Liège, Belgium
2 Laboratory of Molecular Pharmacology, GIGAꢀ Molecular Biology of Diseases, ULiège, B34, Quartier Hôpital –
Avenue de l'hôpital, 11, Bꢀ4000 Liège, Belgium
3 NAmur MEdicine & Drug Innovation Center (NAMEDIC), NARILIS, UNamur, rue de Bruxelles 61, Bꢀ5000
Namur.
4 Biostructural Research, Department of Drug Design and Pharmacology, Faculty of Health and Medical Sciences,
University of Copenhagen, Universitetsparken 2, DKꢀ2100 Copenhagen, Denmark
Keywords
AMPA receptor; positive allosteric modulator; dimeric compound; crystallography; molecular
modeling
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 49
1
2
3
4
Abstract
5
6
7
8
9
The present study describes the identification of highly potent dimeric 1,2,4ꢀbenzothiadiazine
1,1ꢀdioxide (BTD)ꢀtype positive allosteric modulators of the AMPA receptors (AMPApams)
obtained by linking two monomeric BTD scaffolds through their respective 6ꢀpositions. Using
previous Xꢀray data from monomeric BTDs coꢀcrystallized with the GluA2o ligandꢀbinding
domain (LBD), a molecular modeling approach was performed to predict the preferred dimeric
combinations. Two 6,6ꢀethyleneꢀlinked dimeric BTD compounds (16 and 22) were prepared and
evaluated as AMPApams on HEK293 cells expressing GluA2o(Q) (calcium flux experiment).
These compounds were found to be about 10,000 times more potent than their respective
monomers, the most active dimeric compound being the bisꢀ4ꢀcyclopropylꢀsubstituted compound
22 [6,6'ꢀ(ethaneꢀ1,2ꢀdiyl)bis(4ꢀcyclopropylꢀ3,4ꢀdihydroꢀ2Hꢀ1,2,4ꢀbenzothiadiazine 1,1ꢀdioxide],
with an EC50 value of 1.4 nM. As a proof of concept, the bisꢀ4ꢀmethylꢀsubstituted dimeric
compound 16 (EC50 = 13 nM) was successfully coꢀcrystallized with the GluA2oꢀLBD and was
found to occupy the two BTD binding sites at the LBD dimer interface.
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
Introduction
It is well established that Lꢀglutamate is the key excitatory neurotransmitter in the central
1
nervous system (CNS). Lꢀglutamate exerts its biological effect on the CNS through the
activation of either metabotropic (mGluRs, G proteinꢀcoupled) or ionotropic (iGluRs, ligandꢀ
1,2
gated ion channel) receptors.
iGluRs can be classified into three subcategories depending on
their affinity toward nonꢀendogenous ligands: NꢀmethylꢀDꢀaspartic acid (NMDA) receptors, αꢀ
aminoꢀ3ꢀhydroxyꢀ5ꢀmethylꢀ4ꢀisoxazolepropionic acid (AMPA) receptors and kainic acid (KA)
ACS Paragon Plus Environment
2

Page 3 of 49
Journal of Medicinal Chemistry
1
2
3
4
receptors. 1,2 Due to rapid response to Lꢀglutamate in the synapses, iGluRs play a sustainable role
5
6
3ꢀ5
in the fast transmission of excitatory potentials from one neuron to another.
Several studies
7
8
have shown the impact of a decrease of the AMPA signaling in brain disorders such as attention
9
6ꢀ
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
deficit hyperactivity disorder (ADHD), Alzheimer’s disease, mild depression or schizophrenia.
8
Based on this, the AMPA receptors (AMPARs) seem to be an interesting pharmacological
target for the development of either probing tools or cognitive enhancers.
It is conceivable that agonists could be used as cognitive enhancers but their development is
questioned due to the possible occurrence of unwanted and adverse phenomena related to
excitotoxicity. 9 Therefore, AMPA positive allosteric modulators (soꢀcalled “AMPApams”) have
been postulated to be a more promising option. Because they do not bind into the orthostericꢀ
binding site but into an allosteric binding site, they are expected to be devoid of an intrinsic
activity on the AMPARs and to show an effect only when glutamate is released in the synapse.
This mode of action is assumed to give better fineꢀtuning of the AMPA receptor signaling and to
avoid the excitotoxicity problems resulting from overdose of a direct agonist.
AMPARs consist of various combinations of four different subunits (GluA1ꢀ4) that can bind to
each other in several ways, either on a homoꢀ or a heteromeric scaffold with heteromeric
10
constructions preferred in wild type neurons. They first assemble in homomeric dimers and
11
then in functional dimers of dimers (homodimers or heterodimers). Each subunit constituting
12ꢀ14
the receptors can be divided into four semiautonomous domains.
Two of them are
extracellular: the aminoꢀterminal domain (ATD) and the ligandꢀbinding domain (LBD). One is
inside the membrane, which constitutes the channel itself, the transmembrane domain (TMD).
The last one is intracellular, the carboxyꢀterminal domain (CTD). 12ꢀ14 Interactions with ligands
occur at the level of the LBD, either with Lꢀglutamate or with different allosteric modulators,
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 49
1
2
3
4
respectively in the orthosteric (agonist) binding site and in the allosteric binding site, which lies
5
6
2
at the interface of two subunits. The interactions of AMPApams with the AMPAR can lead to
7
8
9
two main outcomes: they can act on slowing down the receptor deactivation process by
enhancing the affinity of Lꢀglutamate for its orthosteric binding site or they can increase the
affinity between subunits, thereby inhibiting the receptor desensitization process, where
conformational changes lead to the closure of the channel even if Lꢀglutamate is still bound. 15ꢀ17
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The wellꢀknown allosteric binding site has a U shape and comprises LBDs of both subunits. It is
18,19
divided in three smaller subsites called A, B/B’ and C/C’.
Allosteric modulators interact
with the different subsites depending on their chemical structure. For example, the benzamideꢀ
18
type family of AMPApams was shown to interact mostly with the A subsite. Among those
compounds, aniracetam (1, Figure 1) is referred to as the oldest molecule acting as an
20
AMPApam.
The second major chemical class of AMPApams is the ringꢀfused thiadiazine
family of compounds among which the benzothiadiazine dioxide (BTD) members are the most
prominent representatives. 20 They were studied following the discovery of the in vitro AMPAR
potentiation by cyclothiazide (2, Figure 1), a compound used in clinical practice for its diuretic
21
action. However, this drug did not show marked in vivo activity on the CNS resulting from
AMPAR potentiation, probably because of its inability to cross the bloodꢀbrain barrier (BBB).
18,20
Cyclothiazide interacts with the B and C subsites of this allosteric binding site in a
symmetrical way resulting in an interaction of one molecule per subunit, thus two molecules at
18
the dimer interface. Unlike cyclothiazide, IDRA 21 (4, Figure 1), a saturated analog of the
potassium channel opener diazoxide (3, Figure 1), was found to be active in vitro as an
22
AMPApam and was further identified in vivo as a cognitive enhancer in animal models of
23,24
cognitive impairments.
Accordingly, many examples of ringꢀfused thiadiazine dioxides
ACS Paragon Plus Environment
4

Page 5 of 49
Journal of Medicinal Chemistry
1
2
3
4
based on this structure were developed as AMPApams (see for example compounds 5ꢀ8; Figure
5
6
20,25
1).
These molecules interact in different ways with the binding site compared to
7
8
cyclothiazide: being smaller, they rotate in the cavity and only interact with the hydrophobic
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
C/C’ subsites (still with a ratio of one molecule per subunit, thus two molecules per dimer
18
interface).
Recently, the largeꢀsized BTD modulator 8 (BPAM538) was found to bind the
26
“BTD” allosteric site with a high affinity and with only one molecule per dimer interface.
Other families of AMPApams that have their own way to interact with the allosteric binding site
20
have been reported,
among which are Nꢀbiaryl(cyclo)alkylꢀ2ꢀpropanesulfonamides,
phenyliminothiazoles and 3ꢀtrifluoromethylpyrazoles.
Among the reported AMPApams, Nꢀbiarylpropylmethanesulfonamides such as 9 synthesized by
27
Kaae et al. should be highlighted as the first examples of dimeric molecules modulating the
activity of AMPARs. The authors showed that some dimeric compounds acted as AMPApams
with a drastic one thousand fold elevation of the potency compared to the constitutive
27
monomers. This gain of potency was tentatively explained by the presence of the twoꢀfold
pattern fitting in the two subsites, C and C’, on both side of the AMPAR dimer interface.
Based on Xꢀray data obtained from some examples of BTDꢀtype AMPApams synthesized by our
team and coꢀcrystallized with the GluA2oꢀLBD, 28ꢀ33 we clearly observed a close proximity of
two modulators in two binding sites at the dimer interface. Thus, we hypothesized that it should
be possible to design dimeric compounds linked together via the 6ꢀposition of the respective
BTD monomers (Figure 2), providing new ligands with a significant enhancement of potency on
AMPARs. This assumption was reinforced by our recent work on 7ꢀphenoxyꢀsubstituted BTDs.
26 Indeed, we demonstrated that these compounds occupied the C subsite with their BTD moiety,
spanning the A subsite and pointing the methoxy substituent toward the B’ subsite where waterꢀ
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 49
1
2
3
4
26
mediated interactions took place between receptor subunits.
Therefore, we reasoned that it
5
6
7
8
could be possible to design rather large molecules that would interact on both sites of the
GluA2oꢀLBD dimer interface. Starting from such hypothesis, we elaborated our work with an
initial molecular modeling approach (docking studies) to predict what would be the most potent
dimeric modulator candidates. These molecules were synthesized, pharmacologically evaluated
and then compared to their respective monomers in order to demonstrate a gain of activity on the
AMPAR by the design of dimeric compounds, and finally coꢀcrystallized with the GluA2oꢀLBD
as an ultimate proof of concept.
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
Results and Discussion
Dimeric compounds design based on docking studies
In order to prioritize the synthesis of novel compounds, we decided to dock potential dimeric
combinations and to compare the obtained results with the position of the two isolated monomers
(compound 6) in the crystal structure (pdbꢀcode 4N07). 29 According to the distance between the
two monomers in the crystal structure, we estimated that an ethylene or a propylene spacer
would be appropriate to generate dimeric structures. The anchoring points for the linker were set
to the 6,6ꢀ, 7,7ꢀ or 6,7ꢀpositions. We also considered amide or methylamine linkers instead of the
alkylꢀlinking group. The alkyl substituent at the 4ꢀposition was chosen in accordance with
previously established structureꢀactivity relationships (SAR) (a methyl radical as the smallest
alkyl group and a cyclopropyl radical known to confer the highest potency in the BTD series of
AMPApams). Finally, we evaluated the possibility to simplify the structure by replacing one of
the BTD monomers by a simpler benzenesulfonamide group. On this ground, a total of 10
ACS Paragon Plus Environment
6

Page 7 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
potential dimeric compounds were designed and docked using the automated GOLD 5.3.0
program while the binding energy and interactions were obtained using Discovery Studio 4.0.
The docking solutions were ranked and analyzed according to the occurrence of compounds in
each cluster, the ChemPLP score, the calculated binding energy and the superimposition with
compound 6 in the crystal structure (see Supporting Information for details on the interactions
between the docking pose and the target).
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
From the results of this docking study, it seems that an ethylene linker would be more favorable
compared to a propylene, an amide or a methylamine linker, and that the anchoring point should
preferably be the 6,6ꢀposition (Table 1). While a propylene linker associated to a Nꢀmethyl
moiety gives a promising calculated binding energy, its superimposition with the monomer in the
crystal structure was found to be less satisfactory (see dimeric compounds C and D in the
Supporting Information). The simplification of the dimeric structure through the replacement of
one BTD monomer with a benzenesulfonamide group led to a marked decrease of estimated
binding energy. Taken together, the results of this analysis highlighted the potential interest of
the dimeric compounds A (16) and B (22) with an ethylene bridge between the two 6ꢀpositions
of the respective BTD scaffolds for potent AMPAR positive allosteric modulation.
Synthesis of the target compounds
The synthetic pathways giving access to the two desired dimeric molecules 16 and 22 are shown
in Schemes 1 and 2. It should be noted that the change of the nature of the substituent on the
nitrogen atom at the 4ꢀposition led to a complete change in the synthetic pathway. For the 4ꢀ
methylꢀsubstituted dimeric compound 16, a classical convergent method was used and we firstly
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 49
1
2
3
4
realized the synthesis of the monomeric BTD intermediate 14 according to previously described
5
6
29
protocols
except the use of a Suzuki’s coupling reaction for vinylation at the 6ꢀposition.
7
8
9
Finally, the 6ꢀvinylꢀsubstituted monomer 14 was coupled through a Grubbs’ metathesis for
dimerization to obtain the dimeric intermediate 15, which gave access to the final compound 16
after hydrogenation (Scheme 1). For the 4ꢀcyclopropylꢀsubstituted dimeric molecule 22 the same
synthetic pathway was investigated, but the classical ring formation with triethyl orthoformate on
intermediate 18 led to uncharacterized byproducts. A second pathway was developed in which
the dimerization took place before the ring closure reaction, and then the benzothiadiazine
dioxide cycle was obtained on both sides of the dimeric molecule according to classical protocols
(Scheme 2).
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
Concerning the monomeric compounds selected for comparison purposes in this work, the BTDs
29,34
34 and 35 devoid of a substituent at the 6ꢀposition were synthesized as previously described.
The corresponding 6ꢀmethylꢀsubstituted BTDs 28 and 33 were obtained using a slightly modified
pathway compared to their corresponding unsubstituted analogs at the 6ꢀposition (Schemes 3 and
4).
Starting from mꢀtoluidine (23), the ring closure reaction was performed with chlorosulfonyl
35
isocyanate in the classical experimental conditions described by Girard et al.
The reaction
provided a mixture of two geometric isomers (24a and 24b) (Scheme 3). The mixture was
engaged in the next step of hydrolysis by means of concentrated sulfuric acid in water (50%
w/v). After 6 hours at refluxing temperature, compound 24b was predominantly converted into
the corresponding ringꢀopened aminobenzenesulfonamide 25, while compound 24a remained
largely insoluble and unaffected in the acidic medium. After filtration, compound 25 was
ACS Paragon Plus Environment
8

Page 9 of 49
Journal of Medicinal Chemistry
1
2
3
4
recovered in the filtrate, and then engaged in the next usual steps 34 up to the target compound 28
5
6
(Scheme 3).
7
8
9
For the synthesis of 33, starting compound 2ꢀfluoroꢀ4ꢀmethylaniline (29) was converted into the
corresponding 2ꢀfluoroꢀ4ꢀmethylbenzenesulfonamide 30 after diazotization according to the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
36
Meerwein conditions. The oꢀfluorobenzenesulfonamide 30 was then converted to the target
compound 33 according to a multistep sequence previously described for 4ꢀcyclopropylꢀ
substituted 3,4ꢀdihydroꢀ2Hꢀ1,2,4ꢀbenzothiadiazine 1,1ꢀdioxides (scheme 4). 29
Dramatic increase of the potency of AMPApams by the dimeric design
The evaluation of the two dimeric molecules 16 and 22 and their constitutive monomers 28, 33,
34 and 35 as AMPApams was achieved with a fluorescenceꢀbased calcium assay as previously
described. 26 HEK293 cells stably expressing GluA2o(Q) were loaded with the calciumꢀsensitive
fluorescent probe Fluoꢀ4ꢀAM. The cells were incubated with the modulators at different
37
concentrations and stimulated with Lꢀglutamate (1 mM).
The intensity of fluorescence
emission is positively correlated to the cytoplasmic concentration of calcium ions and hence can
be used as a surrogate measure to estimate the calcium influx from extracellular medium through
the open AMPAR channel. Therefore, this assay is suitable to determine the potentiating activity
of the compounds on the Lꢀglutamateꢀinduced AMPAR opening. From the data obtained, the
pEC50 values (negative logarithm of the modulator concentration responsible for 50% of the
maximum effect) were calculated for the six selected molecules and compared with the pEC50
values obtained in the same conditions for the reference compounds 6 and 8 (Table 2; Figure 3).
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 49
1
2
3
4
20,25
As expected from previous SAR studies on BTDꢀtype AMPApams,
the cyclopropyl
5
6
7
8
substituent at the 4ꢀposition was always found to be the optimal (cyclo)alkyl chain with a gain of
potency compared to the corresponding 4ꢀmethylꢀsubstituted compounds of at least one log unit
(compare 35 versus 34; 33 versus 28; 22 versus 16). Introduction of a methyl group at the 6ꢀ
position in both series giving compounds 28 and 33 was responsible for a marked decrease of
potency on AMPARs compared to the unsubstituted compounds 34 and 35 (compare 28 versus
34 and 33 versus 35) supporting the view that the absence of a substituent at this position should
be preferred. On the contrary, with the dimeric compounds bearing an alkyl link at this key
position, a substantial improvement of activity on AMPARs, up to more than 10,000 times (4 log
units) compared to their corresponding monomers, was observed (compare 16 versus 34 and 22
versus 35). This major enhancement led to the identification of one of the most potent
AMPApams (compound 22) ever described with an EC50 value of 1.4 nM (pEC50 value of ꢀ8.90).
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
We recently demonstrated that the Hill coefficient (HC), which is sometimes improperly referred
to as the "slope" of the sigmoidal curve, 38 of the concentrationꢀresponse curves obtained in our
26
functional assay was a good surrogate to estimate the binding stoichiometry of AMPApams.
Therefore, calculation of the HC has been included in Table 2. It was noticed that compounds 16
and 22 were characterized by a HC closer to unity, a profile that we previously observed for
26
compound 8. Such observations strongly suggest that the dimeric compounds have a different
binding stoichiometry compared to their monomeric counterparts 28 and 33. These data are
consistent with the nature of these dimeric compounds establishing a bridge between two
AMPAR subunits, forming a single binding pocket at this interface.
The observation that the interaction of 7ꢀphenoxyꢀsubstituted BTDs with two distant AMPAR
26
subunits led to a drastic enhancement of the activity
seems to be confirmed here with the
ACS Paragon Plus Environment
10

Page 11 of 49
Journal of Medicinal Chemistry
1
2
3
4
dimeric compounds. To our knowledge, the BTDꢀtype AMPApam family shows an effect either
5
6
2
on the deactivation or the desensitization process. With the dimeric modulators we observe a
7
8
9
“staple effect”, or at least a bridge effect, in which the dimeric compound acts by “sticking” the
two LBD subunits together, leading to a locked position of the 2ꢀfold symmetric GluA2 dimer
and thus most probably inhibiting completely the desensitization phenomenon. This hypothesis is
strengthened by studies on iGluRs involving rigid bridging through cysteineꢀcysteine interaction
where the linking of the two different parts of the LBD with a similar bridge significantly
affected desensitization. 39ꢀ41
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
Co-crystallization of dimeric compound 16 with GluA2o-LBD shows 1:2 stoichiometry
To establish the binding mode of 16 at AMPARs, we crystallized 16 in complex with GluA2oꢀ
LBD and Lꢀglutamate. Diffraction data were collected to a high resolution of 1.4 Å (Table 3).
The complex crystallized with two GluA2oꢀLBD molecules in the crystal asymmetric unit,
forming a dimer (Figure 4A). Unambiguous electron density was observed for Lꢀglutamate in
both orthosteric binding sites and for one molecule of 16 at the dimer interface (Figure 4B). In
the presence of 16, Lꢀglutamate induces a domain closure of 20.7° (chain A) and 21.7° (chain B)
relative to the apo structure of GluA2oꢀLBD (pdbꢀcode 1FTJ, chain A). These domain closures
are similar to those observed with Lꢀglutamate with no modulator bound. 42
Compound 16 binds at GluA2 in a low energy conformation (2.8 kJ/mol above the global energy
minimum). Therefore, introducing an ethylene bridge in 16 only gives rise to a minor
conformational penalty. For the dimeric modulator 16 the potency at GluA2 was increased
~30,000 and ~13,000 times, respectively, compared to its monomeric parent compound 28 (6ꢀ
ACS Paragon Plus Environment
11

Journal of Medicinal Chemistry
Page 12 of 49
1
2
3
4
5
6
7
8
9
methyl) and the monomeric compound containing a 6ꢀhydrogen (34) (Table 1). Similarly,
compound 22 was superior to its monomeric parent compound 33 (~34,000) and 35 containing
6ꢀhydrogen (~7,000). The increase in potencies corresponds to a more favorable energy of 22ꢀ26
kJ/mol. It has previously been shown that the potency of a bivalent modulator was greater than
the sum of its fragments.27 This could be due to a decreased loss of rotational and translational
entropy (from freezing of the overall molecular motion) in the dimeric compound resulting from
joining the fragments compared to the binding of two individual components of the monomer.
However, other aspects related to enthalpy might also be involved such as decreased rate of
dissociation of the dimer. The 6ꢀhydrogen containing compounds 34 and 35 have a greater
potency than 6ꢀmethyl compounds 28 and 33 (~2 and 5 times, respectively). This is in agreement
with that 34 and 35 can bind in a similar manner in GluA2 as 16, whereas a different binding
mode is required for 28 and 33 to avoid clash between the two methyl groups.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Compound 16 binds to the allosteric binding site in the same region as other BTD modulators.
The binding mode of the dimeric modulator 16 is similar to that of the monomeric modulator 6
(Figure 4C). However, a slight displacement of the BTD ring systems of ~0.5 Å was observed.
Previously, it was shown that introduction of an cyclopropyl group at the 4ꢀN nitrogen in the
monomeric modulator (6) led to 10 times improved binding affinity compared to an ethyl group
29
(5).
This is in agreement with what we observed in the present study, namely a 10 times
greater potency of 22 (with cyclopropyl) compared to 16 (with methyl) (Table 2). Compound 16
binds symmetrically in both the B/B’ and C/C’ subsites (Fig. 4C).
Similarly to 16 and 22, compound 8 can be considered a dimeric modulator, but it is asymmetric.
However, the 3ꢀmethoxyphenoxy substituent of 8 occupies a different space than the
ACS Paragon Plus Environment
12

Page 13 of 49
Journal of Medicinal Chemistry
1
2
3
4
benzothiadiazine dioxide of 16, with the benzene ring positioned perpendicular to the backbone
5
6
of Met517 and Ser518. 26
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
The following residues are within 4 Å of 16: Lys514, Pro515, Phe516, Met517, Ser518, Ser750,
Lys751, Gly752, Leu772 and Asn775 of both subunits comprising the GluA2oꢀLBD dimer
(numbering with signal peptide). Of these residues, Lys514, Pro515, Ser518, Ser750, Leu772
and Asn775 are protruding into the modulator binding site (Figure 4B). Notably, Ser518 and
Ser750 are present in two sideꢀchain conformations in both subunits of which one conformation
of Ser518 in both subunits are lining the ethylene bridge of 16. Both conformations of Ser518
form waterꢀmediated contacts to both conformations of Ser750 (Figure 4B). Compound 16 only
forms two hydrogen bonds to the receptor subunits: a hydrogen bond is seen from both
sulfonamide nitrogen atoms of 16 to the backbone oxygen atom of Pro515 in each subunit
(Figure 4B). In addition, a waterꢀmediated hydrogen bond is formed from an oxygen atom of
each sulfonamide in 16 to the side chain of Asn775. Thus, the monomeric parts of 16 make polar
contacts to the receptor subunits in a symmetrical manner.
Conclusions
The aim of the present study was to develop dimeric compounds able to fit the allosteric binding
sites located on both sides of the AMPAR dimer interface. The design of such molecules was
motivated by the putative gain of activity resulting from a double occupancy of the binding site.
Previous Xꢀray data obtained with “monomeric” BTDꢀtype AMPApams clearly highlighted the
close proximity of two molecules of such modulator on two close receptor subunits and that it
ACS Paragon Plus Environment
13

Journal of Medicinal Chemistry
Page 14 of 49
1
2
3
4
5
6
could be possible to design dimeric molecules interacting with both sides of the LBD dimer
interface.
7
8
9
Molecular modeling was used to compute the virtual docking of a series of large dimeric
molecules in the doubleꢀsided allosteric pocket. The best fits were obtained with dimeric
compounds linked together by an ethylene bridge between their respective 6ꢀpositions. It should
be noticed that there was a little difference in the predicted binding affinities between
compounds differing by the nature of the substituent (methyl or cyclopropyl) introduced at the 4ꢀ
position of the two BTD scaffolds.
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
According to these data, two dimeric molecules, compounds 16 and 22, were selected for
chemical synthesis. This was done following two different synthetic pathways. For the 4ꢀmethylꢀ
substituted compound 16, a convergent pathway was developed in which the 1,2,4ꢀ
benzothiadiazine 1,1ꢀdioxide ring was first built and for which the cross linkage took place in the
last steps. For the 4ꢀcyclopropylꢀsubstituted compound 22, the cross linking step was performed
much earlier in the synthesis route and the ring closure of the heterocycle was built in the final
steps.
The pharmacological evaluation of the target compounds and their corresponding constitutive
monomers was performed on HEK293 cells expressing the GluA2o(Q) receptor and the
variations of cytoplasmic calcium content were followed with a fluorescent probe. The
AMPApam activity of dimeric compound 22 was found to be approximately 10 times higher
than dimeric compound 16, an observation that differ only slightly from the docking results. For
the whole series of tested compounds, the cyclopropyl group was always found to be the best
choice of alkyl substituent at the 4ꢀposition of the BTD ring. Interestingly, both dimeric
ACS Paragon Plus Environment
14

Page 15 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
molecules 16 and 22 showed about 10,000 times lower EC50 values than their corresponding
unsubstituted BTD monomers (compounds 34 and 35). To the best of our knowledge, compound
22, with an EC50 value of 1.4 nM, may be considered as one of the most potent AMPApam ever
described in the literature, expressing a potency in the same order of magnitude as the recently
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
26
described compound 8. Ultimately, the coꢀcrystallization of dimeric compound 16 with the
GluA2oꢀLBD clearly demonstrated that the two allosteric binding sites located at the LBD dimer
interface were able to host the dimeric compound occupying the position of two separated
monomers.
Experimental section
General procedures. Melting points were determined on a Büchi Tottoli capillary apparatus and
are uncorrected. The 1H and 13C NMR spectra were recorded on a Bruker Avance (500 MHz for
1H; 125 MHz for 13C) instrument using deuterated dimethyl sulfoxide (DMSOꢀd6) as the solvent
with tetramethylsilane (TMS) as an internal standard; chemical shifts are reported in δ values
(ppm) relative to that of internal TMS. The abbreviations s = singlet, d = doublet, t = triplet, q =
quadruplet, m = multiplet, dd = doublet of doublet, qd = quadruplet of doublet, dt = doublet of
triplet, tt = triplet of triplet and bs = broad singlet are used throughout. Elemental analyses (C, H,
N, S) were realized on a Thermo Scientific Flash EA 1112 elemental analyzer and were within ±
0.4% of the theoretical values for carbon, hydrogen and nitrogen. This analytical method
certified a purity of ≥ 95% for each tested compound. All reactions were routinely checked by
TLC on silica gel Merck 60 F254.
4ꢀBromoꢀ2ꢀ(methylamino)benzenesulfonamide (11). 4ꢀBromoꢀ2ꢀfluorobenzenesulfonamide
ACS Paragon Plus Environment
15

Journal of Medicinal Chemistry
Page 16 of 49
1
2
3
4
5
6
7
8
9
(10) (4 g, 15.7 mmol) was dissolved in a 1:3 mixture of 1,4ꢀdioxane and methylamine solution
40% in water (20 mL). The solution was heated in a microwave oven at 135°C for 25 min. After
cooling, the solvent was removed by distillation under reduced pressure. The pink solid was
suspended in water, collected by filtration and dried in a vacuum desiccator (yield: 96%); m.p.:
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
171ꢀ173°C; H NMR (500 MHz, DMSOꢀd6) δ 7.51 (d, J=8.4 Hz, 1H, 6ꢀH), 7.38 (s, 2H,
SO2NH2), 6.87 (d, J = 1.8 Hz, 1H, 3ꢀH), 6.82 (dd, J = 8.4 Hz/1.9 Hz, 1H, 5ꢀH), 6.01 (q, J = 4.5
Hz, 1H, NH), 2.83 (d, J = 4.9 Hz, 3H, NHCH3) ; 13C NMR (125 MHz, DMSOꢀd6) δ 146.7 (Cꢀ1),
130.0 (Cꢀ6), 127.2 (Cꢀ4), 124.1 (Cꢀ2), 117.0 (Cꢀ5), 113.3 (Cꢀ3), 30.0 (NCH3).
2ꢀ(Methylamino)ꢀ4ꢀvinylbenzenesulfonamide (12). To
a
solution of 4ꢀbromoꢀ2ꢀ
(methylamino)benzenesulfonamide (11) (1.9 g, 7.17 mmol) in 1,4ꢀdioxane (14 mL), potassium
vinyltrifluoroborate (1.15 g, 8.6 mmol, 1.2 eq.) was added. Once the salt was completely
solubilized, an aqueous solution of NaOH 10% (6.3 mL) was added, followed by Pd(OAc)2 (4.3
mg, 0.27 mol%). This mixture was heated to reflux for 24 h. Once returned to room temperature,
the crude black suspension was filtered. The aqueous filtrate was then extracted with ethyl
acetate (3 x 25 mL). The organic layers were combined, washed with brine (15 mL) and dried
over MgSO4. The solvent was removed under reduced pressure and the oily residue was
1
recrystallized in a mixture of methanol/water 1:2 (yield: 33%); m.p.: 128ꢀ131°C; H NMR (500
MHz, DMSOꢀd6) δ 7.57 (d, J = 8.2 Hz, 1H, 6ꢀH), 7.26 (s, 2H, SO2NH2), 6.80 (d, J = 8.3 Hz, 1H,
5ꢀH), 6.76 (s, 1H, 3ꢀH), 6.72 (dd, J = 17.7, 10.9 Hz, 1H, CH2CH), 5.92 (d, J = 17.6 Hz, 1H, Zꢀ
CH2CH), 5.87 (q, J = 4.9 Hz, 1H, NHCH3), 5.35 (d, J = 10.9 Hz, 1H, EꢀCH2CH), 2.86 (d, J = 4.9
Hz, 3H, NCH3); 13C NMR (125 MHz, DMSOꢀd6) δ 145.9 (Cꢀ1), 141.8 (Cꢀ4), 136.4 (CH), 128.5
(Cꢀ6), 124.0 (Cꢀ2), 116.4 (CH2), 111.9 (Cꢀ5), 109.0 (Cꢀ3), 29.6 (NCH3).
ACS Paragon Plus Environment
16

Page 17 of 49
Journal of Medicinal Chemistry
1
2
3
4
4ꢀMethylꢀ6ꢀvinylꢀ4Hꢀ1,2,4ꢀbenzothiadiazine
1,1ꢀdioxide
(13)
.
2ꢀ(Methylamino)ꢀ4ꢀ
5
6
7
8
vinylbenzenesulfonamide (12) (1.13 g, 5.3 mmol) was heated under stirring in triethyl
othoformate (5 mL) at 135°C for 4 hours. The resulting suspension was cooled on an ice bath
and the solid product was collected by filtration, washed with ether and dried (yield: 81%); m.p.:
>300°C; 1H NMR (500 MHz, DMSOꢀd6) δ 8.08 (s, 1H, 3ꢀH), 7.85 (d, J = 8.2 Hz, 1H, 8ꢀH), 7.70
(dd, J = 8.3, 1.4 Hz, 1H, 7ꢀH), 7.52 (d, J = 1.4 Hz, 1H, 5ꢀH), 6.88 (dd, J = 17.7, 11.0 Hz, 1H
CH2CH), 6.14 (d, J = 17.6 Hz, 1H, ZꢀCH2CH), 5.53 (d, J = 10.9 Hz, 1H, EꢀCH2CH), 3.65 (s, 3H,
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
13
NCH3); C NMR (125 MHz, DMSOꢀd6) δ 151.3 (Cꢀ3), 141.7 (Cꢀ6), 136.2 (Cꢀ4a), 135.1 (CH),
124.4 (Cꢀ8), 123.9 (Cꢀ7), 121.6 (Cꢀ8a) 118.7 (CH2), 114.3 (Cꢀ5), 38.2 (NCH3).
4ꢀMethylꢀ6ꢀvinylꢀ3,4ꢀdihydroꢀ2Hꢀ1,2,4ꢀbenzothiadiazine 1,1ꢀdioxide (14). 4ꢀMethylꢀ6ꢀvinylꢀ
4Hꢀ1,2,4ꢀbenzothiadiazine 1,1ꢀdioxide (13) (0.4 g, 1.83 mmol) was stirred in suspension in
isopropanol (10 mL) at 60°C. NaBH4 (0.2 g, 3 eq.) was added and the mixture was stirred for 15
min. The solvent was removed by distillation under reduced pressure and the crude product was
taken in water and the suspension was adjusted to acidic pH by means of 6N HCl. The aqueous
suspension was then extracted with methylene chloride (3 x 30 mL). The combined organic
layers were washed with brine and dried over MgSO4. The filtrate was evaporated to dryness and
the resulting solid residue was suspended in water, collected by filtration, washed with water and
1
dried (yield: 61%); m.p: > 300°C; H NMR (500 MHz, DMSOꢀd6) δ 8.03 (t, J = 7.9 Hz, 1H,
NH), 7.49 (d, J = 8.1 Hz, 1H, 8ꢀH), 6.95 (d, J = 8.1 Hz, 1H, 7ꢀH), 6.88 (d, J = 1.4 Hz, 1H, 5ꢀH),
6.72 (dd, J = 17.6, 10.9 Hz, 1H, CH2CH), 5.95 (dd, J = 17.6, 0.8 Hz, 1H, ZꢀCH2CH), 5.38 (d, J =
13
11.0 Hz, 1H, EꢀCH2CH), 4.65 (d, J = 8.1 Hz, 2H, CH2), 2.97 (s, 3H, NCH3); C NMR (125
MHz, DMSOꢀd6) δ 144.4 (Cꢀ6), 141.6 (Cꢀ4a), 136.2 (CH2CH), 124.4 (Cꢀ8), 122.0 (Cꢀ8a), 116.5
(CH2CH), 113.6 (Cꢀ7), 111.7 (Cꢀ5), 62.2 (Cꢀ3), 35.9 (NCH3).
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 49
1
2
3
4
5
6
7
8
9
(E)ꢀ6,6'ꢀ(Etheneꢀ1,2ꢀdiyl)bis(4ꢀmethylꢀ3,4ꢀdihydroꢀ2Hꢀ1,2,4ꢀbenzothiadiazine 1,1ꢀdioxide
(15). 4ꢀMethylꢀ6ꢀvinylꢀ3,4ꢀdihydroꢀ2Hꢀ1,2,4ꢀbenzothiadiazine 1,1ꢀdioxide (14) (0.2 g, 0.89
mmol) was solubilized in methylene chloride (3 mL) and supplemented with HoveydaꢀGrubbs II
catalyst (0.01 mol%) under nitrogen, and the mixture was heated to reflux. Every 30 min, the
nitrogen was fluxed and replaced with fresh volume. After 4 hours, the reaction mixture was
cooled to room temperature and the solid was collected by filtration, washed with ether and dried
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
(yield: 65%); m.p.: 331°C (decomposition); H NMR (500 MHz, DMSOꢀd6) δ 8.06 (t, J = 7.9
Hz, 1H, NH), 7.54 (d, J = 8.1 Hz, 1H, 8ꢀH), 7.37 (s, 1H, CH), 7.10 (dd, J = 8.2, 1.6 Hz, 1H, 7ꢀ
H), 7.05 (d, J = 1.5 Hz, 1H, 5ꢀH), 4.68 (d, J = 7.6 Hz, 2H, CH2), 3.01 (s, 3H, NCH3); 13C NMR
(125 MHz, DMSOꢀd6) δ 144.3 (Cꢀ6), 141.2 (Cꢀ4a), 130.0 (CH linker), 124.5 (Cꢀ8), 121.8 (Cꢀ8a),
114.3 (Cꢀ7), 112.2 (Cꢀ5), 62.3 (Cꢀ3), 36.2 (NCH3).
6,6'ꢀ(Ethaneꢀ1,2ꢀdiyl)bis(4ꢀmethylꢀ3,4ꢀdihydroꢀ2Hꢀ1,2,4ꢀbenzothiadiazine 1,1ꢀdioxide) (16).
(E)ꢀ6,6'ꢀ(Etheneꢀ1,2ꢀdiyl)bis(4ꢀmethylꢀ3,4ꢀdihydroꢀ2Hꢀ1,2,4ꢀbenzothiadiazine 1,1ꢀdioxide (15)
(0.13 g, 0.31 mmol) was dissolved in DMF (5 mL) and was supplemented with 10% palladium
on charcoal (13 mg). The resulting suspension was placed under 12 bars of H2 for one hour. The
solvent was removed by distillation under reduced pressure and the residue was recrystallized in
1
methanol, collected by filtration and dried (yield: 25%); m.p.: 313°C (decomposition); H NMR
(500 MHz, DMSOꢀd6) δ 7.91 (s, 1H, NH), 7.42 (d, J = 7.4 Hz, 1H, 8ꢀH), 6.71 (s, 1H, 5ꢀH), 6.70
(dd, J = 7.1 Hz / 1.2 Hz, 1H, 7ꢀH), 4.62 (s, 2H, CH2), 2.91 (s, 3H, NCH3), 2.86 (s, 2H, CH2
13
(linker)); C NMR (125 MHz, DMSOꢀd6) δ 146.9 (Cꢀ6), 144.0 (Cꢀ4a), 124.0 (Cꢀ8), 120.6 (Cꢀ
8a), 116.8 (Cꢀ7), 113.4 (Cꢀ5), 62.4 (Cꢀ3), 36.7 (CH2 linker), 36.2 (NCH3); Anal. (C18H22N4O4S2):
calculated: N 13.26%, C 51.17%, H 5.25%, S 15.18%; found: N 13.22%, C 51.29%, H 5.29%, S
14.71%.
ACS Paragon Plus Environment
18

Page 19 of 49
Journal of Medicinal Chemistry
1
2
3
4
4ꢀBromoꢀ2ꢀ(cyclopropylamino)benzenesulfonamide
(17).
4ꢀBromoꢀ2ꢀ
5
6
7
8
fluorobenzenesulfonamide (10) (4.0 g, 15.7 mmol) was dissolved in a 1:1 mixture of 1,4ꢀdioxane
and cyclopropylamine (20 mL). The solution was heated in a microwave oven to 135°C for 25
min. After cooling, the solvent was removed by distillation under reduced pressure. The solid
residue was recrystallized from methanol/water 1:2, collected by filtration, washed with water
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
1
and dried in a vacuum desiccator (yield: 75%); m.p.: 201ꢀ204°C; H NMR (500 MHz, DMSOꢀ
d6) δ 7.52 (d, J = 8.4 Hz, 1H, 6ꢀH), 7.43 (bs, 2H, SO2NH2), 7.25 (d, J = 1.9 Hz, 1H, 3ꢀH), 6.90
(dd, J = 8.4, 1.9 Hz, 1H, 5ꢀH), 6.21 (d, J = 1.8 Hz, 1H, NH), 2.50 (m, 1H, CH(CH2)2), 0.87 –
13
0.76 (m, 2H, CH(CH2)2), 0.56 – 0.49 (m, 2H, CH(CH2)2). C NMR (125 MHz, DMSOꢀd6) δ
146.6 (Cꢀ1), 129.9 (Cꢀ6), 126.8 (Cꢀ4), 124.4 (Cꢀ2), 118.0 (Cꢀ5), 114.9 (Cꢀ3), 24.2 (NCH3), 7.1
(CH(CH2)2).
2ꢀ(Cyclopropylamino)ꢀ4ꢀvinylbenzenesulfonamide (18). The title compound was obtained
according to the procedure described for compound 12 starting from compound 17 (yield: 86%);
1
m.p.: 155ꢀ161°C; H NMR (500 MHz, DMSOꢀd6) δ 7.61 – 7.51 (m, 1H, 6ꢀH), 7.31 (s, 2H,
SO2NH2), 7.15 (d, J = 1.6 Hz, 1H, 3ꢀH), 6.87 (dd, J = 8.2, 1.6 Hz, 1H, 5ꢀH), 6.74 (dd, J = 17.6,
10.9 Hz, 1H, CHCH2), 6.11 (s, 1H, NH), , 5.90 (d, J = 1.7 Hz, 1H, ZꢀCHCH2), 5.37 (d, J = 17.5
Hz, 1H, EꢀCHCH2), 2.50 (m, , 1H, CH(CH2)2), 0.88 – 0.72 (m, 2H, CH(CH2)2), 0.57 – 0.43 (m,
13
2H, CH(CH2)2); C NMR (125 MHz, DMSOꢀd6) δ 146.0 (Cꢀ4), 141.0 (Cꢀ1), 136.3 (CHCH2),
128.0 (Cꢀ6), 124.0 (Cꢀ2), 116.4 (CH2), 113.0 (Cꢀ5), 110.7 (Cꢀ3), 24.5 (CH(CH2)2), 7.2
(CH(CH2)2).
(E)ꢀ4,4'ꢀ(Etheneꢀ1,2ꢀdiyl)bis(2ꢀ(cyclopropylamino)benzenesulfonamide) (19). The title
compound was obtained according to the procedure described for compound 15 starting from
1
compound 18 (yield: 89%); m.p.: 247°C (decomposition); H NMR (500 MHz, DMSOꢀd6) δ
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 49
1
2
3
4
5
6
7
8
9
7.61 (d, J = 8.3 Hz, 1H, 6ꢀH), 7.37 – 7.28 (m, 3H, SO2NH2 and 3ꢀH), 7.09 (dd, J = 8.4, 1.5 Hz,
1H, 5ꢀH), 6.16 – 6.11 (m, 1H, NH), 2.56 (tt, J = 6.4, 3.3 Hz, 1H, CH(CH2)2), 0.86 (m, J = 6.6,
3.3 Hz, 2H, CH(CH2)2), 0.55 (m, J = 4.4 Hz, 2H, CH(CH2)2); 13C NMR (125 MHz, DMSOꢀd6) δ
145.6 (Cꢀ4), 141.1 (Cꢀ1) 130.0 (Cꢀ6), 128.5 (CH linker), 124.1 (Cꢀ2), 113.3 (Cꢀ5), 111.6 (Cꢀ3),
24.6 (NCH), 7.4 (CH(CH2)2).
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
4,4'ꢀ(Ethaneꢀ1,2ꢀdiyl)bis(2ꢀ(cyclopropylamino)benzenesulfonamide)
(20).
The
title
compound was obtained according to the procedure described for compound 16 starting from
compound 19 (yield: 90%); m.p.: 220ꢀ222°C; 1H NMR (500 MHz, DMSOꢀd6) δ 7.50 (d, J = 8.1
Hz, 1H, 6ꢀH), 7.23 (s, 2H, SO2NH2), 6.94 (d, J = 1.6 Hz, 1H, 3ꢀH), 6.62 (dd, J = 8.1, 1.7 Hz, 1H,
5ꢀH), 6.05 (d, J = 1.6 Hz, 1H, NH), 2.90 (s, 2H, CH2), 2.43 – 2.37 (m, 1H, CH(CH2)2), 0.76 (dt, J
= 6.6, 3.3 Hz, 2H, CH(CH2)2), 0.50 – 0.41 (m, 2H, CH(CH2)2); 13C NMR (125 MHz, DMSOꢀd6)
δ 146.8 (Cꢀ4), 145.5 (Cꢀ1), 128.0 (Cꢀ6) 123.0 (Cꢀ2), 115.9 (Cꢀ5), 112.6 (Cꢀ3), 36.4 (CH2 linker),
24.5 (NCH), 7.3 (CH(CH2)2).
6,6'ꢀ(Ethaneꢀ1,2ꢀdiyl)bis(4ꢀcyclopropylꢀ4Hꢀ1,2,4ꢀbenzothiadiazine 1,1ꢀdioxide) (21). The
title compound was obtained according to the procedure described for compound 13 starting
from compound 20 (yield: 87%); m.p.: 298°C (decomposition); 1H NMR (500 MHz, DMSOꢀd6)
δ 8.09 (s, 1H, 3ꢀH), 7.81 (d, J = 8.1 Hz, 1H, 8ꢀH), 7.53 (s, 1H, 5ꢀH), 7.47 (d, J = 8.1 Hz, 1H, 7ꢀ
H), 3.26 (m, J = 7.2, 3.9 Hz, 1H, CH(CH2)2), 3.15 (s, 2H, CH2 linker), 1.09 (t, J = 6.6 Hz, 2H
CH(CH2)2), 0.93 – 0.84 (m, 2H, CH(CH2)2); 13C NMR (125 MHz, DMSOꢀd6) δ 151.4 (Cꢀ3),
146.6 (Cꢀ6), 136.3 (Cꢀ4a), 127.2 (Cꢀ8), 124.1 (Cꢀ7), 120.3 (Cꢀ8a), 116.3 (Cꢀ5), 36.3 (CH2
linker), 32.1 (NCH), 7.3 (CH(CH2)2).
ACS Paragon Plus Environment
20

Page 21 of 49
Journal of Medicinal Chemistry
1
2
3
4
6,6'ꢀ(Ethaneꢀ1,2ꢀdiyl)bis(4ꢀcyclopropylꢀ3,4ꢀdihydroꢀ2Hꢀ1,2,4ꢀbenzothiadiazine
1,1ꢀ
5
6
7
8
dioxide) (22). The title compound was obtained according to the procedure described for
compound 14 starting from compound 21 suspended in isopropanol and using DMF drop by drop
until the reactant was in solution before the addition of NaBH4 (yield: 82%); m.p.: 290°C
(decomposition); 1H NMR (500 MHz, DMSOꢀd6) δ 7.78 (t, J = 7.9 Hz, 1H, NH), 7.43 (d, J = 8.0
Hz, 1H, 8ꢀH), 7.02 (d, J = 1.5 Hz, 1H, 5ꢀH), 6.75 (dd, J = 8.1, 1.5 Hz, 1H, 7ꢀH), 4.61 (d, J = 7.7
Hz, 2H, 3ꢀH), 2.92 (s, 2H, CH2 linker), 2.41 (m, 1H, CH(CH2)2), 0.90 – 0.84 (m, 2H, CH(CH2)2),
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
13
0.58 (m, 2H, CH(CH2)2); C NMR (125 MHz, DMSOꢀd6) δ 146.1 (Cꢀ6), 143.9 (Cꢀ4a), 124.2
(Cꢀ8), 121.2 (Cꢀ8a), 117.8 (Cꢀ7), 114.2 (Cꢀ5), 61.0 (Cꢀ3), 36.5 (CH2 linker), 29.6 (NCH), 8.4
(CH(CH2)2). Anal. (C22H26N4O4S2): calculated: N 11.80% C 55.68% H 5.52% S 13.51%; found:
N 11.77% C 55.88% H 5.67% S 13.05%.
2ꢀAminoꢀ4ꢀmethylbenzenesulfonamide (25). A solution of mꢀtoluidine (23) (10 g, 93.3 mmol)
in nitromethane (40 mL) was added to a stirred, iceꢀcooled solution of chlorosulfonyl isocyanate
(9 mL) in nitromethane (25 mL) during 5 minutes. This addition resulted in a suspension.
Aluminium chloride (14.0 g, 105 mmol) was added at once, resulting in a clear solution. The
reaction mixture was then refluxed for 30 minutes, cooled, and poured into iceꢀwater. The
precipitated solid was collected by filtration and washed with water. The precipitate was
suspended in water and then extracted three times with ethyl acetate. The combined organic
layers were dried over MgSO4 and filtered. The filtrate was concentrated to dryness under
reduced pressure. The residue was taken in water (100 mL) and the mixture was adjusted to pH
12 with NaOH 10% w/w in water, treated with charcoal and filtered. The filtrate was then
acidified by addition of 12N HCl, and the solid that precipitated was collected by filtration. This
solid consisted in the mixture of the two isomers (24a and 24b). This solid was dispersed in 50%
ACS Paragon Plus Environment
21

Journal of Medicinal Chemistry
Page 22 of 49
1
2
3
4
5
6
7
8
9
w/w sulfuric acid in water and refluxed during 6 hours. After being cooled, the mixture was
filtered and afforded a solid (mainly unchanged 24a) and a filtrate containing compound 25.
After neutralizing the filtrate with NaOH 40% w/w in water, the solvent was evaporated under
reduced pressure. The residue was suspended in hot ethanol (200 mL) and filtered. The resulting
filtrate was concentrated to dryness under reduced pressure and the residue of the title compound
was recrystallized in methanol:water 1:2 (yield: 19%); m.p.: 121ꢀ122°C (lit. 124ꢀ126°C 43).
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
6ꢀMethylꢀ4Hꢀ1,2,4ꢀbenzothiadiazine 1,1ꢀdioxide (26). The title compound was obtained
according to the procedure described for compound 13 starting from compound 25 (yield: 76%);
1
m.p.: 186ꢀ195°C; H NMR (DMSOꢀd6, 500 MHz) δ 12.17 (s, 1H, NH), 7.94 (s, 1H, 3ꢀH), 7.69
13
(d, 1H, 8ꢀH), 7.58 (s, 1H, 5ꢀH), 7.25 (d, 1H, 7ꢀH), 2.39 (s, 3H, 6ꢀCH3); C NMR (125 MHz,
DMSOꢀd6) δ 147.5 (Cꢀ3), 143.5 (Cꢀ6), 134.6 (C4a), 127.7 (Cꢀ8a), 123.6 (Cꢀ8), 120.1 (Cꢀ7),
117.0 (Cꢀ5), 21.1 (CH3).
4,6ꢀDimethylꢀ4Hꢀ1,2,4ꢀbenzothiadiazine 1,1ꢀdioxide (27). The mixture of 6ꢀmethylꢀ4Hꢀ1,2,4ꢀ
benzothiadiazine 1,1ꢀdioxide (26) (1 g, 4.62 mmol), potassium carbonate (2 g), and methyl
iodide (1.7 mL) in acetonitrile (30 mL) was heated at 60°C for 3 h. The solvent was removed by
distillation under reduced pressure and the residue was suspended in water (40 mL). The
resulting insoluble material was collected by filtration, washed with water, dried and
1
recrystallized in ethyl acetate (yield: 75%); m.p.: 269ꢀ272°C; H NMR (DMSOꢀd6, 500 MHz)
8.04 (s, 1H, 3ꢀH), 7.76 (d, 1H, 8ꢀH), 7.37 (d, 1H, 7ꢀH), 7.32 (s, 1H, 5ꢀH), 3.59 (s, 3H, NCH3),
13
2.44 (s, 3H, 6ꢀCH3); C NMR (125 MHz, DMSOꢀd6) δ 151.1 (Cꢀ3), 143.8 (Cꢀ6), 135.8 (Cꢀ4a),
127.7 (Cꢀ7), 124.0 (Cꢀ8), 120.2 (Cꢀ8a), 116.4 (Cꢀ5), 38.0 (NCH3), 21.4 (6ꢀCH3).
ACS Paragon Plus Environment
22

Page 23 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
4,6ꢀDimethylꢀ3,4ꢀdihydroꢀ2Hꢀ1,2,4ꢀbenzothiadiazine 1,1ꢀdioxide (28). The title compound
was obtained according to the procedure described for compound 14 starting from compound 27
7
8
1
(yield: 75%); m.p.: 141ꢀ142°C; H NMR (DMSOꢀd6, 500 MHz) δ 7.95 (t, 1H, NH) 7.40 (d, 1H,
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
8ꢀH), 6.66 (s, 1H, 5ꢀH), 6.62 (d, 1H, 7ꢀH), 4.63 (d, 2H, 3ꢀH2), 2.92 (s, 3H, NCH3), 2.29 (s, 3H, 6ꢀ
13
CH3); C NMR (125 MHz, DMSOꢀ d6) δ 144.0 (Cꢀ6), 143.3 (Cꢀ4a), 124.0 (Cꢀ8), 120.3 (Cꢀ8a),
117.3 (Cꢀ7), 113.8 (Cꢀ5), 62.3 (Cꢀ3), 36.1 (NCH3), 21.6 (6ꢀCH3). Anal. (C9H12N2O2S):
calculated: N 13.20%, C 50.92%, H 5.70% S 15.10%; found: N 13.14%, C 50.61%, H 5.50% S
15.53%.
2ꢀFluoroꢀ4ꢀmethylbenzenesulfonamide (30). A portion of glacial acetic acid (160 mL) was
saturated for 30 min with gaseous sulfur dioxide. To the resulting solution, cooled on an ice bath,
was added under stirring an aqueous solution of CuCl2 (7 g in 20 mL) (suspension A). 2ꢀFluoroꢀ
4ꢀmethylaniline (29) (15 g, 87 mmol) was dissolved in a mixture of glacial acetic acid (160 mL)
and 16N HCl (40 mL). To the resulting solution, cooled in an ice−salt bath (−5 °C), was added
dropwise under stirring an aqueous solution of NaNO2 (8 g in 20 mL, 116 mmol). At the end of
the addition, the resulting solution was slowly mixed with suspension A. After being stirred for
15 min, the suspension was poured onto ice (400 g). The resulting precipitate was collected by
filtration, washed with water, and immediately dissolved in 1,4ꢀdioxane (150 mL). The solution
obtained was added gradually, under stirring, to a concentrated ammonium hydroxide solution
(300 mL) previously cooled on an ice bath. After the resulting solution was stirred for 30 min,
the organic solvent and part of the ammonia were removed by evaporation under reduced
pressure. The aqueous solution/suspension obtained was adjusted to neutral pH by adding 6N
HCl. The precipitate formed was collected by filtration, washed with water and dried (yield:
37%); m.p.: 136ꢀ137°C; 1H NMR (500 MHz, DMSOꢀd6) δ 7.66 (t, J = 7.9 Hz, 1H, 8ꢀH), 7.59 (d,
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 49
1
2
3
4
J = 4.7 Hz, 2H, SO2NH2), 7.26 (d, J = 11.3 Hz, 1H, 7ꢀH), 7.17 (t, J = 6.4 Hz, 1H, 3ꢀH), 2.38 (d, J
5
6
13
= 4.0 Hz, 3H, CH3); C NMR (125 MHz, DMSOꢀd6) δ 158.9 ꢀ 156.9 (d, J = 252.3 Hz, Cꢀ2),
7
8
9
145.5 (Cꢀ1), 128.8 ꢀ 128.7 (d, J = 14.4 Hz, Cꢀ6), 128.1 (Cꢀ4), 124.9 (Cꢀ5), 117.3 ꢀ 117.1 (d, 20.8
Hz, Cꢀ3), 20.8 (CH3).
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
2ꢀ(Cyclopropylamino)ꢀ4ꢀmethylbenzenesulfonamide
(31).
2ꢀFluoroꢀ4ꢀ
methylbenzenesulfonamide (30) (5 g, 21 mmol) was introduced into a hermetically closed vessel
containing a mixture of 1,4ꢀdioxane (70 mL) and cyclopropylamine (3.5 mL, 50 mmol). The
closed vessel was placed in an oven at 100 °C for 240 h. After this time period, the solvent and
the reagent were removed by distillation under reduced pressure. The residue was taken up in
methanol (20 mL), and the insoluble material, which contained the title product, was collected by
filtration, washed with methanol, and dried. This compound was used in the next step without
1
further purification (yield: 35%) ; H NMR (500 MHz, DMSOꢀd6) δ 7.48 (d, J = 7.9 Hz, 1H, 6ꢀ
H), 7.23 (s, 2H, SO2NH2), 6.94 (s, 1H, 3ꢀH), 6.58 – 6.49 (m, 1H, 5ꢀH), 6.06 (s, 1H, NH), 2.45
(dp, J = 9.7, 3.7, 3.2 Hz, 1H, CH(CH2)2), 2.29 (s, 3H, CH3), 0.85 – 0.71 (m, 2H, CH(CH2)2), 0.50
(p, J = 4.3 Hz, 2H, CH(CH2)2); 13C NMR (125 MHz, DMSOꢀd6) δ 145.4 (Cꢀ4), 143.2 (Cꢀ1),
128.1 (Cꢀ6), 122.6 (Cꢀ2), 116.4 (Cꢀ5), 113.0 (Cꢀ3), 24.5 (CH(CH2)2), 21.5 (CH3), 7.3
(CH(CH2)2).
4ꢀCyclopropylꢀ6ꢀmethylꢀ4Hꢀ1,2,4ꢀbenzothiadiazine 1,1ꢀdioxide (32). The title compound was
obtained according to the procedure described for compound 13 starting form compound 31
(yield: 35%); m.p.: 224ꢀ226°C; 1H NMR (500 MHz, DMSOꢀd6) δ 8.12 (s, 1H, 3ꢀH), 7.77 (d, J =
7.9 Hz, 1H, 8ꢀH), 7.64 (s, 1H, 5ꢀH), 7.39 (d, J = 8.4 Hz, 1H, 7ꢀH), 3,34 (bs with water, 1H,
NCH), 2.48 (s, 3H, CH3), 1.30 – 1.08 (m, 2H, CH(CH2)2), 1.00 (dt, J = 6.9, 4.9 Hz, 2H,
ACS Paragon Plus Environment
24

Page 25 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
CH(CH2)2); 13C NMR (125 MHz, DMSOꢀd6) δ 151.4 (Cꢀ3), 143.7 (Cꢀ6), 136.5 (Cꢀ4a), 127.6 (Cꢀ
7), 124.0 (Cꢀ8), 119.8 (Cꢀ8a), 116.5 (Cꢀ5), 32.1 (CH3), 21.6 (NCH), 7.3 (CH(CH2)2).
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
4ꢀCyclopropylꢀ6ꢀmethylꢀ3,4ꢀdihydroꢀ2Hꢀ1,2,4ꢀbenzothiadiazine 1,1ꢀdioxide (33). The title
compound was obtained according to the procedure described for compound 14 starting form
1
compound 32 (yield: 75%); m.p.: 157ꢀ158°C; H NMR (500 MHz, DMSOꢀd6) δ 7.81 (s, 1H,
NH), 7.45 – 7.38 (d, J = 8.0 Hz, 1H, 8ꢀH), 7.08 (s, 1H, 5ꢀH), 6.70 (d, J = 8.0 Hz, 1H, 7ꢀH), 4.63
(d, J = 4.9 Hz, 2H, CH2), 2.32 (d, J = 5.1 Hz, 3H, CH3), 0.92 (qd, J = 6.5, 4.2 Hz, 2H,
13
CH(CH2)2), 0.65 (q, J = 4.1, 2.8 Hz, 2H, CH(CH2)2); C NMR (125 MHz, DMSOꢀd6) δ 144.1
(Cꢀ6), 142.9 (Cꢀ4a), 124.2 (Cꢀ8), 120.8 (Cꢀ8a), 118.3 (Cꢀ7), 114.5 (Cꢀ5), 61.1 (Cꢀ3), 29.6 (CH3),
21.7 (NCH), 8.4(CH(CH2)2). Anal. (C11H14N2O2S) calculated: N 11.76%, C 55.44%, H 5.92%, S
13.45%; found: N 12.10%, C 55.78%, H 5.99%, S 13.60%.
Fluorescentꢀbased calcium assay on GluA2o(Q) cells. The pharmacological evaluation of the
26
target molecules was performed as described previously.
Briefly, HEK293 cells stably
expressing AMPA channel GluA2 (Q) were routinely grown in Dulbecco’s Modified Eagle’s
o
Medium (DMEM) containing 10% fetal bovine serum (FBS), 1% penicillin and streptomycin
(5000 U/mL), 1% Lꢀglutamine (200 mM) and Geneticin (G418 – 600 mg/mL). The measurement
of the fluorescence intensity was carried out using fluorescent microplate reader Fluoroskan
Ascent FT equipped with two dispensers (Thermo Electron Corporation, Finland). After
trypsinization, the cells were washed twice using Hanks balanced salt solution (HBS, 120 mM
NaCl, 2 mM KCl, 2 mM CaCl , 2 mM MgCl , 10 mM HEPES, pH 7.4) and incubated for 1
2
2
h with fluorescent dye Fluoꢀ4/AM (5 ꢁg/mL; Molecular Probes, Invitrogen, Merelbeke
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 49
1
2
3
4
5
6
7
8
9
Belgium). The cells were rinsed twice with HBS, and 100 ꢁL of the resulting cell suspension in
HBS buffer was introduced into each well of a 96ꢀwell plate (density of 150 000 cells/well). The
evaluated compounds were applied at various concentrations (from 10−11 M to 3x10−3 M, with
maximal DMSO concentration = 1%). After shaking the plate 30 s at 1200 rpm, the emission
was read at 538 nm after an excitation at 485 nm during 500 ms per measures. Measurement of a
vehicle control was followed by injection of a 10 mM Lꢀglutamate solution (10 ꢁL), a
fluorescence measurement (2 min), 15 s of shaking at 1200 rpm and a final fluorescence
measurement (5 min). The results are expressed as the EC50 value, which is the concentration
required to reach half of the maximal intensity, and are calculated by nonlinear regression
analysis (GraphPad Prism software) from at least three independent concentrationꢀresponse
curves made in triplicate.
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
Molecular modeling. The coꢀcrystal structure of monomer 6 with the GluA2oꢀLBDꢀL504Yꢀ
N775S (pdbꢀcode 4N07) was used in the docking study. The target protein was prepared with
Discovery Studio 4.0 while the docking procedure was carried out with the automated GOLD
5.3.0 program. 44 The binding site was defined as a 10 Å sphere starting from the two monomers
6 in the crystal structure, allowing the incorporation of both monomers as well as the amino
acids interacting with them. Potential dimeric compounds were sketched with ChemDraw
(Perkin Elmer Informatics) and prepared for docking with Discovery Studio 4.0 (Accelrys Inc¸
San Diego, California, USA). For each ligand, the number of genetic algorithm (GA) run was set
at 100. The default ChemPLP score function was used and the search efficiency was fixed at
200%. For the output, we asked GOLD to keep the twenty best solutions for each ligand. From
these solutions, clusters based on the orientation adopted by the docking poses were identified.
From the cluster having the highest occurrence, we kept the best representative (higher PLP
ACS Paragon Plus Environment
26

Page 27 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
fitness score). This representative was then used for the minimization in situ and free binding
energy calculation. The in situ minimization, free binding energy calculations and interactions
visualization were performed with Discovery Studio (DS) 4.0 (Accelrys Inc¸ San Diego,
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
California, USA). The in situ minimization and free binding energy calculation were launched in
45,46
the same DS protocol. In this protocol, the ligand conformational entropy
was also
considered (conformers generated with the BEST algorithm) and the Generalized Born with
Molecular Volume (GBMV) was used as implicit solvent model.
Crystallography
GluA2oꢀLBD was expressed and purified as previously described. 47 A suspension of 16 in buffer
with glutamate and 10% DMSO (5 mM Lꢀglutamate, 20 mM 16, 10 mM HEPES (pH 7.0), 20
mM NaCl, 1 mM EDTA, 10% DMSO) was mixed using shaking in the cold room overnight.
Then to the GluA2oꢀLBD solution (in 10 mM HEPES (pH 7.0), 20 mM NaCl, 1 mM EDTA) was
added 16 suspension to a resulting concentration of 5 mg/mL protein and 7 mM 16 suspension.
The proteinꢀligand suspension was left at 6°C for 4 days with periodically shaking. The hanging
drop vapor diffusion method was used for crystallization of the complex at 6 °C. The drops
consisted of 1 ꢁL proteinꢀligand solution and 1 ꢁL reservoir solution. The reservoir volume was
0.5 mL. The crystal used for data collection was obtained using reservoir solution consisting of
15% PEG4000, 0.3 M ammonium sulfate, 0.1 M phosphateꢀcitrate buffer pH 4.5. The crystal
was cryoꢀprotected using reservoir solution containing 20% glycerol before flashꢀcooling in
liquid nitrogen.
Xꢀray diffraction data were collected at BioMAX, MAX IV, Lund, Sweden to a resolution of 1.4
48
49
50
Å. The data were processed with XDS and scaled using SCALA in CCP4 . The structure
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 49
1
2
3
4
51
was solved with molecular replacement using PHASER
in CCP4, with the structure of
5
6
7
8
GluA2oꢀL504YꢀN775SꢀLBD in complex with glutamate and BPAM344 as search model (pdbꢀ
code 4N07; chain A). The structure was built with AUTOBUILD 52 in PHENIX 53. The program
MAESTRO [version 10.1.013, Schrödinger, LLC, New York, NY, 2015] was used to generate
coordinate file for compound 16. The energy minimization and conformational search were
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
calculated in water. The parameter file of 16 was obtained using eLBOW 54, keeping the
55
geometry obtained from MAESTRO. The structure was manually adjusted in COOT
and
refined in PHENIX with individual isotropic Bꢀfactors in the initial refinements and anisotropic
Bꢀfactors at the final refinements, as well as riding H atoms. The structure validation was done
using tools in COOT, MolProbity 56 in PHENIX and the wwPDB Validation Service. The LBD
57
domain closure induced by 16 was calculated using the DynDom server relative to the apo
structure of GluA2oꢀLBD (pdbꢀcode 1FTO, chain A). Figure 4 was made in PyMOL [version
2.0.3, The PyMOL Molecular Graphics Systems, V.S., LLC].
ACS Paragon Plus Environment
28

Page 29 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figures
Figure 1: Known AMPA receptor modulators from different families. Benzamides: (1)
aniracetam; 1,2,4ꢀbenzothiadiazine 1,1ꢀdioxides: (2) cyclothiazide, (3) diazoxide, (4) IDRAꢀ21,
58
(5) BPAM97, (6) BPAM344, (7) 5ꢀ(3ꢀfuryl)ꢀIDRAꢀ21,
methanesulfonamides: dimeric compound (9).
(8) BPAM538; Nꢀbiarylpropylꢀ
ACS Paragon Plus Environment
29