Identification of the First Highly Subtype-Selective Inhibitor of Human GABA Transporter GAT3

Article abstract of DOI:10.1021/acschemneuro.5b00150

Screening a library of small-molecule compounds using a cell line expressing human GABA transporter 3 (hGAT3) in a [³H]GABA uptake assay identified isatin derivatives as a new class of hGAT3 inhibitors. A subsequent structure-activity relationship (SAR) study led to the identification of hGAT3-selective inhibitors (i.e., compounds 20 and 34) that were superior to the reference hGAT3 inhibitor, (S)-SNAP-5114, in terms of potency (low micromolar IC₅₀ values) and selectivity (>30-fold selective for hGAT3 over hGAT1/hGAT2/hBGT1). Further pharmacological characterization of compound 20 (5-(thiophen-2-yl)indoline-2,3-dione) revealed a noncompetitive mode of inhibition at hGAT3. This suggests that this compound class, which has no structural resemblance to GABA, has a binding site different from the substrate, GABA. This was supported by a molecular modeling study that suggested a unique binding site that matched the observed selectivity, inhibition kinetics, and SAR of the compound series. These compounds are the most potent GAT3 inhibitors reported to date that provide selectivity for GAT3 over other GABA transporter subtypes.

Full text of DOI:10.1021/acschemneuro.5b00150

Subscriber access provided by UNIV OF MISSISSIPPI
Article
Identification of the first highly subtype-selective
inhibitor of the human GABA transporter GAT3.
Maria Damgaard, Anas Al-Khawaja, Stine B Vogensen, Andreas Jurik, Maarten
Sijm, Maria E.K. Lie, Mathias I Bæk, Emil Rosentahl, Anders A Jensen, Gerhard
F. Ecker, Bente Frølund, Petrine Wellendorph, and Rasmus P. Clausen
ACS Chem. Neurosci., Just Accepted Manuscript • DOI: 10.1021/acschemneuro.5b00150 • Publication Date (Web): 08 Jul 2015
Downloaded from http://pubs.acs.org on July 10, 2015
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
ACS Chemical Neuroscience 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 22
ACS Chemical Neuroscience
1
2
3
4
Identification of the first highly subtype-selective inhibitor of the human GABA transporter GAT3.
5
6
7
8
Maria Damgaard, Anas Al-Khawaja, Stine B. Vogensen, Andreas Jurik,# Maarten Sijm, Maria E.K. Lie,
Mathias I. Bæk, Emil Rosenthal, Anders A. Jensen, Gerhard F. Ecker, # Bente Frølund, Petrine
Wellendorph* and Rasmus P. Clausen*
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
Department of Drug Design and Pharmacology, Faculty of Health and Medical Sciences, University of
Copenhagen, Universitetsparken 2, DK 2100 Copenhagen
#Department of Pharmaceutical Chemistry, University of Vienna, Althanstrasse 14, Aꢀ1090 Vienna,
Austria.
*Corresponding authors: Rasmus P. Clausen, Phone: +45 35 33 65 66, Eꢀmail: rac@sund.ku.dk,
Pharmacology: Petrine Wellendorph, Phone: +45 39 17 98 11, Eꢀmail: pw@sund.ku.dk.
Graphical table of contents
O
O
Cl
S
O
O
N
Bn
34
N
H
20
(S)ꢁSNAPꢁ5114
hGAT1:
hBGT1:
hGAT2:
hGAT3:
>600 ꢀM
405 ꢀM
158 ꢀM
45 ꢀM
>1000 ꢀM
749 ꢀM
203 ꢀM
6 ꢀM
>1000 ꢀM
>1000 ꢀM
>1000 ꢀM
29 ꢀM
Abstract
Screening of a library of smallꢀmolecule compounds at a cell line expressing the human GABA
transporter 3 (hGAT3) in a [3H]GABA uptake assay, identified isatin derivatives as a new class of
hGAT3 inhibitors. A subsequent structureꢀactivity relationship (SAR) study led to the identification of
hGAT3 selective inhibitors (i.e. compounds 20 and 34) that were superior to the reference hGAT3
inhibitor (S)ꢀSNAPꢀ5114 in terms of potency (low micromolar IC50 values) and selectivity (>30 fold
selective for hGAT3 over hGAT1/hGAT2/hBGT1). Further pharmacological characterization of
compound 20 (5ꢀ(thiophenꢀ2ꢀyl)indolineꢀ2,3ꢀdione) revealed a nonꢀcompetitive mode of inhibition at
hGAT3. This suggests that the compound class, which has no structural resemblance to GABA, has a
binding site different from the substrate GABA. This was supported by a molecular modeling study
suggesting a unique binding site matching the observed selectivity, inhibition kinetics and SAR of the
1
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 2 of 22
1
2
3
4
5
compound series. These compounds are the most potent GAT3 inhibitors reported yet that provides
selectivity for GAT3 over the other GABA transporter subtypes.
6
7
8
9
Keywords: GABA uptake, inhibitor, hGAT3 selective, nonꢀcompetitive, kinetics, isatin
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
The major inhibitory neurotransmitter in the central nervous system, γꢀaminobutyric acid (GABA) (1,
Figure 1), plays a central role in normal brain function, and it has been demonstrated that abnormalities in
GABAergic neurotransmission in the brain can lead to several disorders. Proteins involved in GABA
neurotransmission are therefore interesting therapeutic targets and compounds perturbing this system in a
selective manner are considered drug candidates for a number of diseases.1 For example, epileptic and
sleep disorders are treated with compounds acting at the GABAergic system, mainly by increasing
GABAergic neurotransmission. An important part of the GABAergic system is the uptake of GABA into
neurons and glial cells mediated by GABA transporters (GATs), and inhibition of this transport has been
shown to be effective in the treatment of seizures in epileptic disorders.2 The therapeutic relevance has
been demonstrated by the GABA uptake inhibitor tiagabine (GabitrilTM) (2) that is a clinically approved
drug used as addꢀon therapy for partial epilepsy.
Four different mammalian GABA transporters have been cloned, named GAT1, GAT2, GAT3 and BGT1
according to the IUPHAR nomenclature. GAT1 is the most studied transporter due to the existence of
highly potent and specific inhibitors, such as tiagabine (2).3 Development of highly selective compounds
towards the other GABA transporters has been limited. GAT1 and GAT3 are exclusively present in the
brain, but with a heterogenous distribution to primarily neurons (GAT1) and astrocytes (GAT3).4 By
contrast, BGT1 and GAT2 are also expressed in other organs, and although their expression levels in the
brain are debated,5,6 the role of BGT1 in relation to GABAergic neurotransmission is supported by the
anticonvulsant effect of the mixed GAT1/BGT1 inhibitor (RS)ꢀ4ꢀ[Nꢀ[1,1ꢀbis(3ꢀmethylꢀ2ꢀthienyl)butꢀ1ꢀenꢀ
4ꢀyl]ꢀNꢀmethylamino]ꢀ4,5,6,7ꢀtetrahydrobenzo[d]isoxazolꢀ3ꢀol (EF1502, 3),7 and a proposed role of
BGT1 in regulating extrasynaptic GABA neurotransmission.8 Much still remains to be learned about
these transporter subtypes, but this requires more selective compounds.
Due to its exclusive expression in the brain and primary localization to astrocytes, GAT3 is a very
interesting subtype and of potential clinical relevance in epilepsy and stroke,4, 9 but a role in neurons has
also been demonstrated in some studies.10,11 Although (S)ꢀSNAPꢀ5114 (4) is often applied as a “selective”
2
ACS Paragon Plus Environment

Page 3 of 22
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
GAT3 inhibitor, the selectivity of the compound, and of several synthesized analogues12 is not
sufficiently pronounced to exclude effects on other GABA transporter subtypes. In addition, (S)ꢀSNAPꢀ
5114 has a limited brain uptake and a modest chemical stability, thus limiting its use in vivo.13
Consequently, there is a need to develop more selective GAT3 inhibitors that can be used to probe the
physiological roles and therapeutic potential of this transporter.
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 GABA transporters belong to the superfamily 6 of solute carriers (SLC6) that are Na+/Clꢀ dependent
neurotransmitter transporters and share a common topology of 12 transmembrane (TM) spanning
segments. Biochemical and Xꢀray crystallographic studies of members of this transporter family have
provided knowledge on the transport mechanism. Currently, GABA is believed to be transported
according to an alternate access model of substrate translocation, where GABA and coꢀtransported ions
are assumed to enter a central binding site from the extracellular space in an outwardꢀopen state of the
transporter, followed by transition to a state, where GABA is occluded. Subsequently, an inwardꢀfacing
conformation of the protein arises, with lower substrate affinity that triggers the release of the substrate
into the cell.14 Although still debated, it has been proposed that a secondary substrate binding site (S2)
exists at the extracellular part of the occluded state that needs to be occupied to induce transport.15,16
Currently, most of the known inhibitors of the GABA reuptake system display competitive inhibition at
the central substrate binding site, but compounds with nonꢀcompetitive profile have also been described.17
Biostructural studies has enabled comparison of aligned GAT homology models, based on the crystal
structures of the homologous bacterial leucine transporter LeuT and the Drosophila
melanogaster dopamine transporter (DAT),18,19 which offers the possibility of investigating potential
subtypeꢀselective binding cavities. Combined knowledge about steric demands of the inhibitor molecules
and sequence differences facilitate docking studies for the generation of binding hypotheses.
In the current study we report the development of a new structural class of highly subtypeꢀselective
hGAT3 inhibitors that work in a nonꢀcompetitive manner, possibly by acting at a new binding pocket for
inhibition of the GABA transporters.
Results and Discussion
A commercial compound library consisting of 4,000 structurally diverse compounds (Enamine,
Monmouth, NJ, USA) was screened for their ability to inhibit the uptake of 30 nM [3H]GABA into
Chinese Hamster Ovary (CHO) cells stably expressing the human GAT3 (hGAT3) subtype.20 The bromoꢀ
/fluoroꢀsubstituted isatin 5 (Table 1) was identified as an hGAT3 inhibitor, thus representing a new
structural scaffold among GABA uptake inhibitors bearing no structural resemblance to GABA. An
extended characterization of commercially available analogues of 5 was performed as an initial structureꢀ
3
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 4 of 22
1
2
3
4
5
6
7
8
9
activity relationship (SAR) study. 48 commercially available isatin analogues were tested in the
[3H]GABA uptake assay at the hGAT3ꢀCHO cell line in a single concentration. The structures of the 48
analogues and the results from this followꢀup screening are presented in sFig. 1 in the Supporting
Information. From this screening, nine compounds (6ꢀ14, Table 1) showed inhibition to a greater degree
than the original hit (5) at hGAT3 and were consequently selected for further characterization at CHO cell
lines expressing each of the four human GAT subtypes: hGAT1, hBGT1, hGAT2 and hGAT3.
Concentrationꢀinhibition curves were generated for each of the compounds, and the estimated IC50 values
are presented in Table 1. The initial screening hit (5) displayed an inhibitory potency for hGAT3 of 108
ꢀM and an hGAT3/1 selectivity ratio of 5, however it did not discriminate between hBGT1, hGAT2 and
hGAT3. Compounds with a small lipophilic substituent in the R3 position (6, 10 and 13) displayed both
higher hGAT3 potency and hGAT3/1 selectivity, whereas compounds lacking a substituent in the R3
position were weaker inhibitors at hGAT3. Moving the halogen to position R2, R4 and R5 completely
abolished hGAT3 inhibition. The Nꢀmethylated compound (11) did not display more potent inhibition
than the original hit, while the Nꢀchloro compound (12) both improved hGAT3 potency and hGAT3/1
selectivity. Thus, activities of this scaffold responded well to the changes in substituents, and there
seemed to be room for improvement of both hGAT3 potency and hGAT3/1 selectivity. Based on these
findings, a series of compounds with different R1 and R3 substituents were designed to investigate critical
areas for GAT3 inhibition and selectivity further. In particular, we wanted to explore how larger
substituents in the R3 position and on the ring nitrogen would affect the GAT3 selectivity observed in the
initial SAR.
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 procedures for the synthesis of compounds 15ꢀ39 are depicted in scheme 1. The 5ꢀalkyl substituted
analogues 15 and 16 were synthesized via the Sandmeyer isonitrosoacetanilide isatin synthesis as
previously described.21 The reaction was not applicable for the 5ꢀaryl substituted analogues. Instead, these
analogues were synthesized via the Suzuki crossꢀcoupling reaction.22 Conditions previously described for
Suzuki reactions on iodoisatins were tested.23,24 However, low yields were obtained when coupling the 5ꢀ
iodo or 5ꢀbromoisatin under the described conditions. This has previously been described in cross
coupling reactions of isatins.23 Ketal protection of the free carbonyl group was tried to optimize the
yields.23 5ꢀIodoisatin (13) and 5ꢀbromoisatin (40) were refluxed with ethylene glycol and catalytic
amounts of pꢀtoluenesulfonic acid in toluene to yield the ketalꢀprotected compounds 41 and 42. The ketalꢀ
protected isatins 41 and 42 were crossꢀcoupled with phenylboronic acid and (4ꢀmethoxyphenyl)boronic
acid using the previously reported conditions25 affording 43 and 44. Deprotection of the ketal moiety in
concentrated aqueous hydrochloric acid gave the desired compounds 17 and 23, however in low yields.
The overall yield of the Suzuki reactions did not differ substantially between the protected and
4
ACS Paragon Plus Environment

Page 5 of 22
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
unprotected isatins, hence the unprotected reaction was preferred. A number of 5ꢀaryl substituted
analogues were synthesized using 5ꢀiodoisatin directly in Suzuki crossꢀcoupling reactions giving
compounds 19ꢀ28. Nꢀalkylated analogues 29ꢀ39 were obtained using the alkyl halide and potassium
carbonate in DMF under microwaveꢀheating in fair yields.
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
All synthesized analogues were assayed at all four hGAT subtypes in the [3H]GABA uptake assay, and
the results are presented in Table 2. Substitution of the isatin scaffold in the R3 position with small
aliphatic substituents (15 and 16) yielded IC50 values of 49 and 15 ꢀM at hGAT3 and an hGAT3/1
selectivity of 11 and 16, respectively. Increasing the size of the substituents in this position to larger
aromatic substituents (17, 19ꢀ23) increased the inhibitory potency at hGAT3 further, as these analogues
exhibited IC50 values in the low micromolar range and higher hGAT3/1 selectivity ratios (>29 or higher).
In contrast, substitution of the distal aryl with groups larger than a methyl (18, 24ꢀ28) led to complete loss
of activity. For the Nꢀsubstituted compounds, small substituents in the R1 position (29ꢀ30) impaired
potency for hGAT3, however potency was regained for larger substituents (31ꢀ33) and the R1 benzylꢀ
substituted compound 34 displayed an hGAT3 IC50 value of 29 µM and an hGAT3/1 ratio greater than
100. Substitution on the R1 benzyl group was detrimental to the activity of the compounds as the
combination of an aryl group in the R3 position and a benzyl in the R1 position (39) completely abolished
activity.
As can be seen from Tables 1 and 2, a number of the isatin analogues displayed partial inhibition of
[3H]GABA uptake at their saturating concentrations. This phenomenon was seen in the concentrationꢀ
inhibition curves as a bottom plateau less than the control value of 100% inhibition with 3 mM GABA.
An example of such a curve is shown in figure 2A for 34 at hGAT3. For comparison, the concentrationꢀ
inhibition curves for the isatin analogues 17 and 20 as well as the reference compound (S)ꢀSNAPꢀ5114
(4), which all displayed full uptake inhibition, are also included in Figure 2. Several of the isatin
analogues (e.g. 34 at hGAT3, Figure 2A) displayed partial inhibition of [3H]GABA uptake with apparent
plateaus ranging from 36% to 90% of control, a phenomenon reported previously for the dopamine
transporter.26 Whether the observed partial behavior of some isatin analogues is related merely to intrinsic
pharmacological/structural properties of the compounds or to physicochemical properties, such as
aggregation or reaching the solubility limits of the compounds at higher concentrations, is currently
unclear. Clearly, solubility limits could mask either a full inhibition or a potential biphasic concentrationꢀ
inhibition curve. The latter could exist if a compound binds with distinct affinities to two sites within the
same transporter monomer, or with different affinities to two monomers of a transporter dimer. An
example of biphasic uptake kinetics has been reported for the dopamine transporter,27 and dimerization of
GATs has been reported for the rat GAT1.28
5
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 6 of 22
1
2
3
4
5
6
7
8
9
Figure 2B highlights the pronounced hGAT3 over hGAT1 subtypeꢀselectivity of compound 20 (more
than 167 times), and that 20 is more potent than (S)ꢀSNAPꢀ5114 (4) (data summarized in Tables 1 and 2).
In addition, compound 20 showed an hGAT3 selectivity of at least 30ꢀfold over the other GAT subtypes
while this is approximately 3 for (S)ꢀSNAPꢀ5114 (4). As described earlier, (S)ꢀSNAPꢀ5114 (4) has a
limited use in vivo due to a limited brain uptake and a modest chemical stability.13
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 addition to investigating the inhibitory activity of the synthesized compounds at the four hGATs, the
mechanism underlying the inhibitions of GAT3 transport exerted by 20 was further investigated by
measuring the transport rate of [3H]GABA in the absence and presence of different fixed concentrations
of compound 20. The results are shown in Figure 2C. Increasing the concentration of 20 resulted in a
statistically significant decrease in Vmax, while the pKmꢀvalue was not significantly altered. This suggested
that compound 20 inhibits the uptake of GABA through hGAT3 in a nonꢀcompetitive manner. Kinetics
studies were also performed for the structurally related compound 6, showing the same kinetics profile
and thereby confirming a nonꢀcompetitive inhibition mode for the isatin analogues (results not shown). A
nonꢀcompetitive inhibition mode for these compounds is not surprising, since they show no structural
resemblance with GABA.
A molecular modeling study was employed to suggest potential inhibitory binding sites for the
compounds in the transporter that would be in accordance with the observed SAR and subtypeꢀselectivity
of the compounds. Cavity analysis of comparative models for all subtypes revealed a potential site in an
outwardꢀopen state of GAT3 between TM domains 10 and 11, containing residues on positions 10.49
(S468) and 11.60 (F531) (generic numbering according to Beuming et al. 200629 see Figure 3 and
alignment in sFig. 2 in the Supporting Information) that are not conserved in the three other human
GATs. As it can be seen in Figure 3, substrates could still enter the central binding cavity, and also the
disputed secondary substrate binding site S2 is accessible. Thus, a nonꢀcompetitive mode of inhibition
could root from impaired TM rearrangement needed for substrate translocation to the intracellular space.
Placement of the isatin derivatives into this site using a flexible docking workflow indicated a possible
binding mode, which is in line with the SAR. For selected compounds, short molecular dynamics
simulations in a membrane environment were performed for exploring solventꢀmediated interactions with
surrounding residues.
The hydrophobic side of the compounds was mainly accommodated in the hydrophobic part of the
pocket. An aromatic residue in position 10.64, which is present in all subtypes except hGAT1, was
observed to support hydrophobic interactions with aromatic substituents in the 1ꢀ or 5ꢀposition and
supporting the selectivity trends of the dataset, as exemplified by compounds 17-23 and 34. The position
6
ACS Paragon Plus Environment

Page 7 of 22
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
of the polar oxo groups seem to be stabilized by a complex waterꢀmediated hydrogen bond network,
occasionally involving polar residues of TM1, 6 and the extracellular loop EL4. The fragile character of
this network is underlined by inactivity of compounds 45 and 46, being substituted at the position of the
3ꢀoxo group.
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
As exemplified by hGAT1 and 3 in Figure 4, similar initial ligand placements could be observed for all
four hGAT subtypes, but nonꢀhGAT3 complexes were significantly less stable and displayed pronounced
drifting of the ligands in the short molecular dynamics simulations. This underlines the importance of
hydrophobic properties, as residues 10.49 (Y/Y/Y/S) and 11.60 (S/S/S/F) are, albeit switched, nearly
identical for all of the GAT subtypes. Hence, the small difference of a phenolic hydroxyl group could
cause the particular pharmacological properties of the data set.
Herein, we have reported the design, synthesis and pharmacological evaluation of a new type of hGAT3
selective inhibitors, based on the isatin scaffold. Compound 20 is a more potent analogue than the
reference GAT3 inhibitor (S)ꢀSNAPꢀ5114 (4), and also the selectivity for hGAT3 compared to the other
subtypes is significantly improved. Thus, selectivity for hGAT3 over hGAT2 and hBGT1 is increased
from 3.5ꢀ and 9ꢀ fold to 34ꢀ and 124ꢀfold, respectively. Some of the compounds display partial inhibition
of the GABA transport. Interestingly, these compounds were shown to have a nonꢀcompetitive inhibition
mode at hGAT3, which together with the lack of structural resemblance to GABA suggests a different
binding site to previous inhibitors. Many of the previous GABA uptake inhibitors bear an amino and acid
group, suggesting a binding site partially overlapping with GABA. A molecular modeling study evaluated
potential binding sites in a full range of GABA transporter homology models and only one binding site
had the size and physiochemical characteristics that were in accordance with observed potency and
selectivity trends, which awaits further experimental validation. This is part of future studies that will also
be directed at determining the bloodꢀbrain barrier permeability of selected analogues. Also, studies on
the effect of compound 20 at neurons and astrocytes under normal and diseased conditions, and how it
could influence synaptic vesicle recycling, transporter exo/endocytosis are of high importance. In
conclusion, we have presented a new class of hGAT3 selective compounds that can be used as tool
compounds to understand the pharmacology of hGAT3 and its involvement in disease.
Methods
Chemistry
General procedures. All anhydrous reactions were carried out in ovenꢀdried glassware, under
nitrogen. Solvents were of chromatography grade and dried using an SG Water solvent purification
7
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 8 of 22
1
2
3
4
5
6
7
8
9
system (DCM, DMF, and THF) or with 3 Å molecular sieves. Commercially obtained chemicals were
used without further purification. TLC was carried out on Merck precoated silica gel 60 F254 plates and
visualized using UV (254 nm). Column purifications were carried out manually using silica gel: 40–63
µm. NMR spectra were recorded on a 400 MHz Bruker Avance or a 600 MHz Bruker Avance instrument.
Data reported as follows: chemical shift, multiplicity (s = singlet, d = doublet, dd = double doublet, t =
triplet, bs = broad singlet, m = multiplet), integration and coupling constants (Hz). Chemical shifts are
reported in ppm with the solvent resonance resulting from incomplete deuteration as the internal reference
(CHCl3 in CDCl3: δ 7.26, (CH3)2SO in (CD3)2SO: δ 2.50), exception being samples in D2O, to which 40
µl of dioxane per 100 ml D2O was added as reference (dioxane in D2O: δ 3.75). A few NMR spectra have
been attached in appendix T. Analytical HPLC was performed on a MerckꢀHitachi HPLC system
consisting of an Lꢀ7100 pump, an Lꢀ7200 auto sampler, and an Lꢀ7400 UV detector (254 nm), using an
Chromolith SpeedROD RPꢀ18 column (4.6 × 50 mm) eluted at a flow rate of 4.0 mL/min. A linear
gradient elution was performed with eluent A (H2O/TFA 100:0.1) containing 0% of solvent B
(MeCN/H2O/TFA, 90:10:0.01) rising to 100% of B during 5 min. Data were acquired and processed
using the EZChrom Elite Software version 3.1.7 by Hitachi. HPLC purity is ≥ 95%, unless otherwise
stated. LCMS method A: LCꢀMS was performed using a Agilent 1100 HPLC systems with a XBridge
3.5 ꢁm Cꢀ18 column (100 x 4.60 mm) using gradient elution from buffer A (H2O:CH3CN:HCOOH,
95:5:0.1) to buffer B (H2O:CH3CN: HCOOH, 5:95:0.05) over 10 min flow rate: 1.0 mL/min, coupled to
an Hewlett Packard 1100 series mass spectrometer with an electrospray ionization source.. LCMS method
B: HPLCꢀMS was recorded on an Agilent 1200 series system using an Xbridge RP C18 column (3.5 µm,
100 × 4.6 mm) with UV detection at 210, 254 and 280 nm. Mobile phase (MP) A: 0.2% HCOOH, 99.8%
H2O (v/v). Mobile phase B: 0.2% HCOOH, 99.8% MeCN (v/v). Flow rate: 1 ml/min. Gradient: 0ꢀ7 min:
0ꢀ90% MP B, 7ꢀ9 min: 90% MP B, 9ꢀ16 min: 10% MP B.
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
5-(thiophen-2-yl)indoline-2,3-dione (20). 5ꢀiodoindolineꢀ2,3ꢀdione (13) (0.1 g, 0.36 mmol),
thiopheneꢀ2ꢀboronic acid pinacol ester (0.090 g, 0.43 mmol) and NaHCO3 (0.06 g, 0.72 mmol) were
dissolved in a mixture of DME:H2O (5:1, 4.8 mL). The mixture was degassed with N2 gas, after which
Pd(dppf)Cl2 (15 mg, 0.018 mmol) was added and the reaction vial was sealed and heated at 90oC for 18
hour. The crude reaction mixture was filtered over Celite and diluted with EtOAc, after which the organic
layer was washed with a saturated NaHCO3 solution. Organic layer was dried over MgSO4, after which
volatiles were evaporated under reduced pressure. The resulting crude solids were purified over SiO2
using a gradient of 4:1 nꢀheptane:EtOAc towards 2:1 nꢀheptane:EtOAc yielding 10 mg of title compound
1
(0.042 mmol, 12%). H NMR (600 MHz, DMSOꢀd6) δ 11.14 (s, 1H), 7.86 (dd, J = 8.2, 2.0 Hz, 1H), 7.75
13
(d, J = 2.0 Hz, 1H), 7.55 – 7.51 (m, 1H), 7.15 – 7.11 (m, 1H), 6.96 (d, J = 8.2 Hz, 1H). C NMR (101
MHz, DMSOꢀd6) δ 184.63, 159.91, 150.11, 142.43, 135.43, 129.19, 129.05, 125.88, 124.11, 121.46,
8
ACS Paragon Plus Environment

Page 9 of 22
ACS Chemical Neuroscience
1
2
3
4
5
118.92, 113.25. LCꢀMS (ESI) m/z calculated for C12H8NO2S+: 230.26 found: 230.1 [M+H]+, retention
time: 6.054 minutes.
6
7
8
9
Materials, [3H]GABA uptake assay
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
Ham’s F12 Nutrient Mix, fetal bovine serum (FBS), penicillin–streptomycin, hygromycin B, trypsin and
Hank’s Balanced Salt Solution (HBSS, catalogue number 14175ꢀ095) were purchased from Life
Technologies (Paisley, UK). Plasmocin was obtained from InvivoGen (San Diego, CA, USA). PolyꢀDꢀ
lysine (PDL) and HEPES (4ꢀ(2ꢀhydroxyethyl) piperazineꢀ1ꢀethanesulfonic acid) were purchased from
SigmaꢀAldrich. [2,3ꢀ3H(N)] GABA (specific radioactivity 35.0 Ci/mmol) and MicroScintꢀ20 were
purchased from PerkinElmer (Boston, MA, USA). GABA was purchased from SigmaꢀAldrich (St. Louis,
MO, USA). The isatin analogues 5ꢀ14 (see Table 1) were generously provided by 7TM Pharma,
Denmark.
Cell culture and GAT expression
Generation and culture of the Chinese hamster ovary (CHO) Flpꢀin cell line stably expressing the human
GATs have been previously described.17,20
[3H]GABA uptake assay and compound library screening
The [3H]GABA competition uptake assay and kinetics studies were performed exactly as previously
described using a 96ꢀwell format setup.20
Data analysis
Data and statistical analysis were performed in GraphPad Prism 6 (GraphPad Software, San Diego, CA,
USA). Concentrationꢀinhibition curves generated in the competition [3H]GABA uptake assay were fitted
by nonlinear regression using the equation for sigmoidal concentrationꢀresponse with variable slope: Y =
Bottom + (Top ꢀ Bottom)/(1 + 10^((logIC50 ꢀ X) × HillSlope)), where Y is the response, X is the
logarithm of the concentration, Top and Bottom are the plateaus in same units as Y, and logIC50 is the
concentration giving a response half way between Bottom and Top. The HillSlope is the steepness of the
curve. Concentrationꢀinhibition curves were fitted to 100% inhibition by 3 mM GABA if the highest
9
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 10 of 22
1
2
3
4
5
6
7
8
9
concentration of the inhibitor displayed less than 90% but more than 50% inhibition of [3H]GABA
uptake. The IC50 values of compounds displaying less than 50% but more than 10% inhibition were
estimated as being greater than the highest tested concentration. Compounds showing less than 10%
[3H]GABA uptake inhibition at the highest tested concentration were considered not active.
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
Data from kinetics experiments were analyzed using the MichaelisꢀMenten equation: Y=Vmax*X/(Km+X),
where Y is the uptake velocity, X is substrate concentration, Vmax is the maximum uptake velocity in the
same units as Y, and Km is the MichaelisꢀMenten constant in the same unit as X.
Molecular modeling
Models of hGAT1ꢀ3 and hBGT1 were constructed using Modeller 9.14, using inwardꢀfacing, outwardꢀ
occludedꢀand outwardꢀopen template structure of LeuT (PDB codes 3TT3, 2A65, 3F3A). Initial sequence
alignment according to Beuming et al.29 was further optimized in loop regions (see Supporting
Information sFig.2). Crystal structures of MhsT and DATcryst were consulted for resolving secondary
structural ambiguities. 200 models for each transporter in all three conformations were generated and
sorted according to DOPE Score.
Site Finder module of MOE 2014.09 was used for cavity analysis. Docking studies were carried out using
the Schrödinger 2014.3 Suite. Structure preparation was performed with LigPrep, docking with the
Induced Fit Docking protocol. Isatin derivatives were drawn within MOE and prepared with LigPrep,
sampled for a physiological pH range of 7.2 +/ꢀ 0.2.
MD simulations were performed for 10 ns using Desmond 3.6 and the OPLS 2005 force field, n=3 for
hGAT1 and hGAT3. Protein models were placed in a POPC membrane manually adjusted after being set
to TM residues, and solvated using an SPC water model and 0.15 M NaCl. Trajectory analysis was
carried out using VMD and Desmond's Simulation Interaction Diagram.
Supporting information
Synthetic procedures for the synthesis of for compound 15ꢀ39 can be found in the Supporting Information
along with results and structures of the followꢀup screening of 48 isatin analogues and molecular
modeling data. This material is available free of charge via the Internet at http://pubs.acs.org.
Abbreviations.
10
ACS Paragon Plus Environment

Page 11 of 22
ACS Chemical Neuroscience
1
2
3
4
5
BGT1, betaineꢀGABA transporter; CHO, chinese hamster ovary; CNS, central nervous system; GABA, γꢀ
aminobutyric acid; GAT, GABA transporter; SAR, structureꢀactivity relationship; (S)ꢀSNAP 5114, 1ꢀ[2ꢀ
[tris(4ꢀmethoxyphenyl)methoxy]ethyl]ꢀ(S)ꢀ3ꢀpiperidinecarboxylic acid; TM, transmembrane.
6
7
8
Author Information.
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
Current address: Department of Drug Design and Pharmacology, Faculty of Health and Medical Sciences,
University of Copenhagen, Universitetsparken 2, DK 2100 Copenhagen
Funding: This work was supported by the Lundbeck Foundation, the Novo Nordisk Foundation, the A.P.
Møller Foundation for the Advancement of Medical Sciences, and the Drug Research Academy.
The authors declare no competing financial interest.
11
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 12 of 22
1
2
3
4
References.
5
6
7
8
9
1. Rowley, N. M., Madsen, K. K., Schousboe, A., and Steve White, H. (2012) Glutamate and GABA
synthesis, release, transport and metabolism as targets for seizure control, Neurochem. Int. 61, 546ꢀ558.
2. Meldrum, B. S., and Chapman, A. G. (1999) Basic Mechanisms of Gabitril (Tiagabine) and Future
Potential Developments, Epilepsia 40, S2ꢀS6.
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
3. Nielsen, E. B., Suzdak, P. D., Andersen, K. E., Knutsen, L. J., Sonnewald, U., and Braestrup, C. (1991)
Characterization of tiagabine (NOꢀ328), a new potent and selective GABA uptake inhibitor, Eur.J.
Pharmacol. 196, 257ꢀ266.
4. Madsen, K. K., White, H. S., and Schousboe, A. (2010) Neuronal and nonꢀneuronal GABA
transporters as targets for antiepileptic drugs, Pharmacol. Therapeut 125, 394ꢀ401.
5. Borden, L. A. (1996) GABA transporter heterogeneity: pharmacology and cellular localization,
Neurochem. Int.29, 335ꢀ356.
6. Lehre, A. C., Rowley, N. M., Zhou, Y., Holmseth, S., Guo, C., Holen, T., Hua, R., Laake, P., Olofsson,
A. M., PobleteꢀNaredo, I., Rusakov, D. A., Madsen, K. K., Clausen, R. P., Schousboe, A., White, H. S.,
and Danbolt, N. C. (2011) Deletion of the betaineꢀGABA transporter (BGT1; slc6a12) gene does not
affect seizure thresholds of adult mice, Epilepsy Res. 95, 70ꢀ81.
7. White, H. S., Watson, W. P., Hansen, S. L., Slough, S., Perregaard, J., Sarup, A., Bolvig, T., Petersen,
G., Larsson, O. M., Clausen, R. P., Frolund, B., Falch, E., KrogsgaardꢀLarsen, P., and Schousboe, A.
(2005) First demonstration of a functional role for central nervous system betaine/{gamma}ꢀaminobutyric
acid transporter (mGAT2) based on synergistic anticonvulsant action among inhibitors of mGAT1 and
mGAT2, J. Pharmacol. Exp. Therapeut.312, 866ꢀ874.
8. Madsen, K. K., Ebert, B., Clausen, R. P., KrogsgaardꢀLarsen, P., Schousboe, A., and White, H. S.
(2011) Selective GABA transporter inhibitors tiagabine and EF1502 exhibit mechanistic differences in
their ability to modulate the ataxia and anticonvulsant action of the extrasynaptic GABA(A) receptor
agonist gaboxadol, J. Pharmacol. Exp. Therapeut.338, 214ꢀ219.
9. Clarkson, A. N., Huang, B. S., Macisaac, S. E., Mody, I., and Carmichael, S. T. (2010) Reducing
excessive GABAꢀmediated tonic inhibition promotes functional recovery after stroke, Nature 468, 305ꢀ
309.
10. Melone, M., Cozzi, A., PellegriniꢀGiampietro, D.E. and Conti F. (2003) Transient focal ischemia
triggers neuronal expression of GATꢀ3 in the rat perilesional cortex. Neurobiol Dis. 14, 120ꢀ132.
11. Pozdnyakova, N., Dudarenko, M., Yatsenko, L., Himmelreich, N., Krupko, O., and Borisova, T.
(2014). Perinatal hypoxia: different effects of the inhibitors of GABA transporters GAT1 and GAT3 on
the initial velocity of [3H]GABA uptake by cortical, hippocampal, and thalamic nerve terminals. Croat.
Med. J. 55, 250–258.
12. Pabel, J., Faust, M., Prehn, C., Worlein, B., Allmendinger, L., Hofner, G., and Wanner, K. T. (2012)
Development of an (S)ꢀ1ꢀ{2ꢀ[tris(4ꢀmethoxyphenyl)methoxy]ethyl}piperidineꢀ3ꢀcarboxylic acid [(S)ꢀ
SNAPꢀ5114] carba analogue inhibitor for murine gammaꢀaminobutyric acid transporter type 4,
ChemMedChem 7, 1245ꢀ1255.
13. Dalby, N. O. (2000) GABAꢀlevel increasing and anticonvulsant effects of three different GABA
uptake inhibitors, Neuropharmacol. 39, 2399ꢀ2407.
12
ACS Paragon Plus Environment

Page 13 of 22
ACS Chemical Neuroscience
1
2
3
4
5
6
14. Kristensen, A. S., Andersen, J., Jorgensen, T. N., Sorensen, L., Eriksen, J., Loland, C. J., Stromgaard,
K., and Gether, U. (2011) SLC6 neurotransmitter transporters: structure, function, and regulation,
Pharmacol. Rev. 63, 585ꢀ640.
15. Reyes, N., and Tavoulari, S. (2011) To be, or not to be two sites: that is the question about LeuT
substrate binding, J. Gen. Physiol. 138, 467ꢀ471.
7
8
9
16. Quick, M., Shi, L., Zehnpfennig, B., Weinstein, H., and Javitch, J. A. (2012) Experimental conditions
can obscure the second highꢀaffinity site in LeuT, Nat. Struct. Mol. Biol. 19, 207ꢀ211.
17. Kragholm, B., Kvist, T., Madsen, K. K., Jorgensen, L., Vogensen, S. B., Schousboe, A., Clausen, R.
P., Jensen, A. A., and BraunerꢀOsborne, H. (2013) Discovery of a subtype selective inhibitor of the
human betaine/GABA transporter 1 (BGTꢀ1) with a nonꢀcompetitive pharmacological profile, Biochem.
Pharmacol. 86, 521ꢀ528.
18. Yamashita, A., Singh, S. K., Kawate, T., Jin, Y., and Gouaux, E. (2005) Crystal structure of a
bacterial homologue of Na+/Clꢀꢀdependent neurotransmitter transporters, Nature 437, 215ꢀ223.
19. Penmatsa, A., Wang, K. H., and Gouaux, E. (2013) Xꢀray structure of dopamine transporter elucidates
antidepressant mechanism, Nature 503, 85ꢀ90.
20. AlꢀKhawaja, A., Petersen, J. G., Damgaard, M., Jensen, M. H., Vogensen, S. B., Lie, M. E.,
Kragholm, B., BraunerꢀOsborne, H., Clausen, R. P., Frolund, B., and Wellendorph, P. (2014)
Pharmacological identification of a guanidineꢀcontaining betaꢀalanine analogue with low micromolar
potency and selectivity for the betaine/GABA transporter 1 (BGT1), Neurochem. Res. 39, 1988ꢀ1996.
21. Sandmeyer, T. (1919) Über Isonitrosoacetanilide und deren Kondensation zu Isatinen, Helv. Chim.
Acta 2, 234ꢀ242.
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
22. Miyaura, N., and Suzuki, A. (1995) PalladiumꢀCatalyzed CrossꢀCoupling Reactions of Organoboron
Compounds, Chem. Rev. 95, 2457ꢀ2483.
23. Gérard, A. L., Lisowski, V., and Rault, S. (2005) Direct synthesis of new arylanthranilic acids via a
Suzuki crossꢀcoupling reaction from iodoisatins, Tetrahedron 61, 6082ꢀ6087.
24. Liu, Y. C., Ye, C. J., Chen, Q., and Yang, G. F. (2013) Efficient synthesis of bulky 4ꢀsubstitutedꢀ
isatins via microwaveꢀpromoted Suzuki crossꢀcoupling reaction, Tetrahedron Lett. 54, 949ꢀ955.
25. Wang, C. H., White, A. R., Schwartz, S. N., Alluri, S., Cattabiani, T. M., Zhang, L. K., Chan, T. M.,
Buevich, A. V., and Ganguly, A. K. (2012) Novel synthesis and functionalization of ortho–ortho
disubstituted biphenyls and a highly condensed novel heterocycle using radical cyclization reaction,
Tetrahedron 68, 9750ꢀ9762.
26. Ananthan, S., Saini, S. K., Khare, R., Clayton, S. D., Dersch, C. M., and Rothman, R. B. (2002)
Identification of a novel partial inhibitor of dopamine transporter among 4ꢀsubstituted 2ꢀ
phenylquinazolines, Bioorg. Med. Chem. Lett. 12, 2225ꢀ2228.
27. Lee, F. J., Pristupa, Z. B., Ciliax, B. J., Levey, A. I., and Niznik, H. B. (1996) The dopamine
transporter carboxylꢀterminal tail. Truncation/substitution mutants selectively confer high affinity
dopamine uptake while attenuating recognition of the ligand binding domain, J. Biol. Chem. 271, 20885ꢀ
20894.
28. Scholze, P., Freissmuth, M., and Sitte, H. H. (2002) Mutations within an intramembrane leucine
heptad repeat disrupt oligomer formation of the rat GABA transporter 1, J. Biol. Chem. 277, 43682ꢀ
43690.
29. Beuming, T., Shi, L., Javitch, J. A., and Weinstein, H. (2006) A comprehensive structureꢀbased
alignment of prokaryotic and eukaryotic neurotransmitter/Na+ symporters (NSS) aids in the use of the
LeuT structure to probe NSS structure and function, Mol. Pharmacol. 70, 1630ꢀ1642.
13
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 14 of 22
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. Structures of γꢀaminobutyric acid (GABA, 1), the highly GAT1 selective inhibitor tiagabine
(2), and EF 1502 (3), a mixed BGT1/GAT1 inhibitor and the moderately potent GAT3 inhibitor (S)ꢀ
SNAPꢀ5114 (4).
14
ACS Paragon Plus Environment

Page 15 of 22
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Pharmacological characterization of novel isatin analogues at hGATs, stably expressed in CHO
cells. (A) Concentrationꢀinhibition curves for 17, 20, 34 and (S)ꢀSNAPꢀ5114 (4) at hGAT3. 34 displayed
partial inhibition of [3H]GABA uptake. The responses are normalized mean values of at least three
independent experiments ± S.E.M. (see tables 1 and 2 for mean IC50 values). (B) Concentrationꢀinhibition
curves for compound 20 at hGAT1 and hGAT3. The responses are normalized mean values of three
independent experiments ± S.E.M. Estimated IC50 value at hGAT1: >1000 µM; and at hGAT3: 6 µM
(table 2). (C) Effect of compound 20 on the GABA transport velocity through hGAT3. Specific transport
velocity of increasing concentrations of GABA was measured in the [3H]GABA uptake assay. The
maximum concentration of [3H]GABA was 100 nM. Data are given as nmol GABA transported per
minute and normalized to protein amount per well. Results are mean values of three independent
experiments ± S.E.M. The Km values expressed in µM (pKm ± S.E.M.) were estimated to be: 0 µM 20,
172 (3.77 ± 0.03); 10 µM 20, 201 (3.82 ± 0.24); 50 µM 20, 246 (3.65 ± 0.13). The Vmax values expressed
as nmol GABA/(min×mg protein) ± S.E.M. were estimated to be: 0 µM 20, 0.39 ± 0.06; 10 µM 20, 0.28
± 0.04; 50 µM 20, 0.13 ± 0.01 (the Vmax value was constrained to the average of the maximal level). The
15
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 16 of 22
1
2
3
4
5
Vmax value in the presence of 50 µM 20 was significantly decreased compared to control, but no
significant different in pKm was present (Oneꢀway ANOVA followed by Dunnett’s test, P<0.01).
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. Localization of GAT3 selective residues (orange). Amino acids 10.49 (S468) and 11.60 (F531)
located in the putative pocket depicted in Van der Waals surface.
16
ACS Paragon Plus Environment

Page 17 of 22
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. Molecular dynamics behavior of 34 at hGAT1 and hGAT3. A: proteinꢀligand contacts over the
trajectory. Hydrophobic contacts depicted in purple, hydrogen bonds in green, ionic in pink, water bridges
in blue. B: ligand position relative to key residues at the beginning and the end of the simulation. C:
protein (blue) and ligand (purple) root mean square of deviation (RMSD).
17
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 18 of 22
1
2
3
4
5
6
7
8
9
Scheme 1. Synthesis of 17ꢀ46a
41, R=I
O
42, R=Br
O
O
Tol
R
c
43
, R=Ph
O
O
N
H
44, R=4ꢁMeOPh
N
Bn
39
d
e
b
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
17 Ph
18 4ꢁOMeꢁPh
19: 3ꢁthienyl
20: 2ꢁthienyl
21: pꢁtoluyl
22: mꢁtoluyl
23: oꢁtoluyl
O
O
a
X
R
O
O
N
N
H
H
g
6, X=Cl
13, X=I
40, X=Br
24: 2,6ꢁ(MeO)2ꢁPh
: 2,3ꢁdihydrobenzo(
R
N
25
b
)
Cl
(1,4)ꢁdioxinꢁ6ꢁyl
26: Eꢁstyrenyl
27: naphtalenꢁ1ꢁyl
O
N
H
28: 4ꢁmorpholinoꢁphenyl
f
45, R= Me
46, R= NH2
O
29: Me
30: Et
31: Pr
32: iꢁPr
33: n-Bu
34: Benzyl
Cl
35: 4ꢁmethylbenzyl
36: 3ꢁmethylbenzyl
37: 3ꢁphenoxyꢁbenzyl
38: naphtalenꢁ2ꢁylmethyl
O
N
R
aReagents and conditions: (a) RꢀB(OH)2, Pd(dppf)Cl2, NaHCO3, DME, H2O, heating, 2ꢀ12%, (b)
Ethylene glycol, PTSA (pꢀtoluenesulfonic acid), Toluene, reflux, 68%, (c) PhꢀB(OH)2 or pꢀOMeꢀPhꢀ
B(OH)2, Pd(PPh3)4 or Pd(dppf)Cl2, NaHCO3, DME, H2O, heating, 41ꢀ94%, (d) conc. HCl, MeOH, 80°C,
(e) BnꢀBr, K2CO3, DMF, microwave heating, 85%, (f) RꢀI or RꢀBr, K2CO3, DMF, microwave heating,
20ꢀ65%. (g) Hydrazin or MeNH2.
18
ACS Paragon Plus Environment

Page 19 of 22
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
Table 1. Pharmacological characterization of GABA (1), (S)ꢀSNAPꢀ5114 (4) and a number of
commercially available isatin analogues at the four human GAT subtypes. Compounds were tested for
their ability to inhibit uptake of 30 nM [3H]GABA at the four human GAT subtypes stably expressed in
CHO cells. IC50 values were determined by curveꢀfitting analysis. Compounds were selected based on
their ability to inhibit hGAT3 at a 50 µM concentration (Supporting Information sFig. 1). All data points
were determined in triplicates and repeated in at least three independent experiments.
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
IC50 (ꢀM) (pIC50 ± S.E.M)
Cmpd
R1 R2
R3
R4
R5
hGAT1
hBGT1
hGAT2
hGAT3
GAT3/1
selectivity
ratio
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
2.5
5
2.2
1.4
1.8
GABA (1)
(5.60±0.04) (5.29±0.02) (5.71±0.14) (5.87±0.05)
(S)-SNAP-
5114 (4)
ꢀ
ꢀ
ꢀ
Br
ꢀ
ꢀ
H
>600c
405
158
45
>13.3
5.0
11
(3.40±0.05) (3.81±0.06) (4.37±0.10)
H
H
H
H
H
H
H
Cl
H
H
H
H
H
H
H
H
H
H
H
Cl
F
541a
55b 65b 108a
(3.27±0.05) (4.26±0.02) (4.19±0.05) (4.00±0.07)
5
Cl
H
H
H
H
H
H
199 146 223 18
6
(3.72±0.08) (3.83±0.03) (3.66±0.06) (4.78±0.09)
114 55 76 17
Cl
Cl
Me
H
6.7
60
7
(3.98±0.10) (4.26±0.02) (4.14±0.08) (4.80±0.08)
Cl
1080a 27b 105b 18b
(3.13±0.28) (4.56±0.01) (4.04±0.16) (4.78±0.09)
181 47 51 16
8
OCF3
iPr
11
9
(3.75±0.04) (4.33±0.03) (4.30±0.03) (4.79±0.05)
H
218 155 174 25
8.7
4.2
17
10
11
12
13
14
(3.67±0.04) (3.81±0.04) (3.76±0.03) (4.72±0.18)
229 250 266 55
NO2 Me Me
(3.64±0.01) (3.60±0.03) (3.58±0.04) (4.38±0.17)
H
I
H
H
H
H
H
H
363 130 200 22
(3.46±0.09) (3.90±0.08) (3.70±0.05) (4.67±0.08)
46b 21b 28b
9
5.1
8.5
(4.34±0.06) (4.70±0.07) (4.58±0.08) (5.06±0.02)
Cl
85 54 109 10
(4.07±0.03) (4.27±0.04) (3.97±0.05) (5.01±0.11)
a The compounds displayed less than 90% but more than 50% inhibition, and the concentrationꢀinhibition
curves were accordingly fitted to the value of 100% inhibition (3 mM GABA).
19
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 20 of 22
1
2
3
b
The bottom plateaus of the fitted curves were less than the GABA maximal inhibition plateau (set to
4
5
6
7
100%) and consequently gave the following relative maximum inhibition levels (% inhibition ± S.D.):
hGAT1: 13, 53 ± 4; hBGT1: 5, 60 ± 5; 8, 69 ± 2; 13, 79 ± 4; hGAT2: 5, 60 ± 8; 8, 61 ± 7; 13, 72 ± 8;
hGAT3: 8, 87 ± 5.
c The compound displayed less than 50% inhibition, and the IC50 value was accordingly estimated as
being greater than the highest tested concentration.
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
20
ACS Paragon Plus Environment

Page 21 of 22
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
Table 2. Structureꢀactivity relationships of isatin analogues at the four human GAT subtypes.
Compounds were tested for their ability to inhibit uptake of 30 nM [3H]GABA at the four human GAT
subtypes stably expressed in CHO cells. IC50 values were determined by curveꢀfitting analysis and are
based on at least three independent experiments, whereas compounds with no activity are based on at
least two independent experiments. All data points were carried out in triplicates.
9
IC50 (ꢀM) (pIC50 ± S.E.M)
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
hGAT1
hBGT1
hGAT2
hGAT3
GAT3/1
selectivity
ratio
cmpd
15
R1
R2
H
Et
532a
(3.27±0.01)
680a
643a
49
11
16
(3.17±0.03) (3.20±0.06
)
(4.34±0.12)
H
H
SecꢀBu
Ph
233a
(3.67±0.13)
100
171
15
16
17
(4.01±0.08) (3.77±0.03
)
(4.84±0.10)
876b
(3.07±0.07)
314c
63c
8
110
(3.64±0.27) (4.24±0.13
)
(5.09±0.09)
H
H
4ꢀOMeꢀPh
3ꢀthienyl
>1000
>1000
>1000b
N.A.
>1000b
ꢀ
18
19
33c
(4.48±0.003)
50c
(4.32±0.02
)
17c
(4.79±0.11)
>59
H
2ꢀthienyl
>1000
749c
203c
6c
>167
20
(3.21±0.19) (3.74±0.14
)
(5.20±0.05)
H
H
pꢀtoluyl
mꢀtoluyl
>1000
>1000
>1000b
N.A.
22c
(4.66±0.05)
>45
>29
21
22
55c
63c
35c
(4.47±0.08)
(4.20±0.002
(4.26±0.003)
)
H
oꢀtoluyl
>1000
50c
(4.30±0.02)
52c
(4.29±0.02)
24c
(4.63±0.09)
>42
23
H
H
2,6ꢀOMeꢀPh
N.T.
N.T.
N.T.
N.T.
N.T.
N.T.
>300b
>300b
ꢀ
ꢀ
24
25
(2,3ꢀdiꢀhydrobenzo
[b][1,4]ꢀdioxinꢀ6ꢀyl
H
H
(E)ꢀ styrenyl
N.T.
N.T.
N.T.
N.T.
N.T.
N.T.
>300b
>300b
ꢀ
ꢀ
26
27
naphtalenꢀ1ꢀyl
21
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 22 of 22
1
2
3
4
5
6
7
8
9
H
4ꢀmorpholinoꢀphenyl
Cl
N.T.
N.T.
533a
N.T.
421a
>300b
ꢀ
28
29
Me
906a
(3.04±0.03)
232a
(3.67±0.11)
3.9
(3.28±0.04) (3.39±0.07
)
Et
nPr
iPr
Cl
Cl
Cl
Cl
Cl
563a
(3.25±0.01)
453a
453a
124
(3.96±0.15)
4.5
4.8
10
30
31
32
33
34
(3.35±0.05) (3.35±0.05
)
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
248a
(3.61±0.03)
198
198
52
(3.71±0.04) (3.71±0.04
)
(4.31±0.12)
464a
(3.34±0.04)
235a
235a
46
(3.63±0.02) (3.63±0.02
)
(4.40±0.16)
nBu
Bn
184
(3.74±0.04)
140
140
53
3.5
>34
(3.86±0.05) (3.86±0.05
)
(4.28±0.05)
>1000
>1000
>1000
29c
(4.65±0.18)
4ꢀMeBn
3ꢀMeBn
Cl
Cl
Cl
N.T.
N.T.
N.T.
N.T.
N.T.
N.T.
N.T.
N.T.
N.T.
>1000b
>300b
>300b
ꢀ
ꢀ
ꢀ
35
36
37
Naphtalenꢀ2ꢀ
ylmethyl
3ꢀphenoxyꢀbenzyl
Bn
Cl
N.T.
N.T.
N.T.
N.T.
N.T.
N.T.
N.T.
N.T.
N.T.
N.T.
N.T.
N.T.
>300b
>1000b
>1000b
>1000b
ꢀ
ꢀ
ꢀ
ꢀ
38
39
45
46
pꢀtoluyl
R=Me
R=NH2
N.T. not tested
a The compounds displayed less than 90% but more than 50% inhibition, and the concentrationꢀinhibition
curves were accordingly fitted to the value of 100% inhibition (3 mM GABA).
b
The compounds displayed less than 50% inhibition, and the IC50 values were accordingly estimated as
being greater than the highest tested concentration.
c
The bottom plateaus of these fitted curves were less than the GABA maximal inhibition plateau (set to
100%) and consequently gave the following relative maximum inhibition levels (% inhibition ± std. dev.):
hBGT1: compound 17, 62 ± 9; compound 19, compound 20, 50 ± 11; 36 ± 14; compound 22, 63 ± 9;
compound 23, 61 ± 6; hGAT2: compound 17, 78 ± 7; compound 19; 47 ± 21; compound 20, 66 ± 7;
compound 22, 58 ± 17; compound 23, 52 ± 14; hGAT3: compound 19, 80 ± 10; compound 20, 90 ± 2;
compound 21, 73 ± 1; compound 22, 81 ± 8; compound 23, 83 ± 8; compound 34, 43 ± 11.
22
ACS Paragon Plus Environment