Inhibition of the Vesicular Glutamate Transporter (VGLUT) with Congo Red Analogs: New Binding Insights

Article abstract of DOI:10.1007/s11064-020-03182-0

Abstract: The vesicular glutamate transporter (VGLUT) facilitates the uptake of glutamate (Glu) into neuronal vesicles. VGLUT has not yet been fully characterized pharmacologically but a body of work established that certain azo-dyes bearing two Glu isosteres via a linker were potent inhibitors. However, the distance between the isostere groups that convey potent inhibition has not been delineated. This report describes the synthesis and pharmacologic assessment of Congo Red analogs that contain one or two glutamate isostere or mimic groups; the latter varied in the interatomic distance and spacer properties to probe strategic binding interactions within VGLUT. The more potent inhibitors had two glutamate isosteres symmetrically linked to a central aromatic group and showed IC₅₀ values ~ 0.3–2.0?μM at VGLUT. These compounds contained phenyl, diphenyl ether (PhOPh) or 1,2-diphenylethane as the linker connecting 4-aminonaphthalene sulfonic acid groups. A homology model for VGLUT2 using d-galactonate transporter (DgoT) to dock and identify R88, H199 and F219 as key protein interactions with Trypan Blue, Congo Red and selected potent analogs prepared and tested in this report. Graphic Abstract: [Figure not available: see fulltext.]

Full text of DOI:10.1007/s11064-020-03182-0

Neurochemical Research (2021) 46:494–503
https://doi.org/10.1007/s11064-020-03182-0
ORIGINAL PAPER
Inhibition of the Vesicular Glutamate Transporter (VGLUT) with Congo
Red Analogs: New Binding Insights
David M. Hitt1,2 · Jefery D. Zwicker1,3 · Chih‑Kai Chao1 · Sarjubhai A. Patel1 · John M. Gerdes1
Richard J. Bridges1 · Charles M. Thompson1
·
Received: 11 June 2020 / Revised: 11 September 2020 / Accepted: 17 November 2020 / Published online: 4 January 2021
© Springer Science+Business Media, LLC, part of Springer Nature 2021
Abstract
The vesicular glutamate transporter (VGLUT) facilitates the uptake of glutamate (Glu) into neuronal vesicles. VGLUT has
not yet been fully characterized pharmacologically but a body of work established that certain azo-dyes bearing two Glu
isosteres via a linker were potent inhibitors. However, the distance between the isostere groups that convey potent inhibition
has not been delineated. This report describes the synthesis and pharmacologic assessment of Congo Red analogs that contain
one or two glutamate isostere or mimic groups; the latter varied in the interatomic distance and spacer properties to probe
strategic binding interactions within VGLUT. The more potent inhibitors had two glutamate isosteres symmetrically linked
to a central aromatic group and showed IC50 values~0.3–2.0 μM at VGLUT. These compounds contained phenyl, diphenyl
ether (PhOPh) or 1,2-diphenylethane as the linker connecting 4-aminonaphthalene sulfonic acid groups. A homology model
for VGLUT2 using d-galactonate transporter (DgoT) to dock and identify R88, H199 and F219 as key protein interactions
with Trypan Blue, Congo Red and selected potent analogs prepared and tested in this report.
Graphic Abstract
NH2
NH2
N
N
N
linker
N
SO3H
SO3H
Congo Red: linker = -Ph-Ph-
Compound 2a: linker = -Ph-
Compound 2d: linker = -Ph-O-Ph-
IC50 ~ 550 nM (left)
IC50 ~ 300 nM (right)
IC50 ~ 1500 nM
Compound 2g: linker = -PhCH2CH2Ph- IC50 ~ 2300 nM
Congo Red and compound 2a shown
docked in a homology model of VGLUT2
Keywords Vesicular glutamate transporter · Isostere · Inhibitor · Congo red analogs · Docking · VGLUT
Abbreviations
CR
Congo Red
Glu
Glutamate
4-ANS
4-Aminonaphthalene sulfonic acid
VGLUT Vesicular glutamate transporter
DgoT
TB
d-galactonate transporter
Trypan Blue
* Charles M. Thompson
chuck.thompson@mso.umt.edu
Extended author information available on the last page of the article
Vol:.(1234567890)
1 3

Neurochemical Research (2021) 46:494–503
495
phenomena. VGLUTs are believed to transport substrate
using an alternating-access (rocker) mechanism [18] with
a substrate binding pocket midway in the protein channel
[19–25]. Phosphate serves as an alternative substrate but
binds at a second, undefned binding domain [3, 26]. Using
a homology model for VGLUT1 based on GlpT, a possible
second substrate site for glutamate was proposed [19]; a
notion supported in part by computational and experimen-
tal studies at other transporters [27–31]. The presence of
a second substrate binding site on VGLUT also appears
plausible because Trypan Blue (TB) and related azo-dye
structures possessing two symmetric glutamate or amino
acid mimics (Fig. 1) are potent inhibitors, whereas a ligand
containing.
Introduction
As the primary excitatory neurotransmitter in the mam-
malian CNS, l-glutamate (Glu) is produced in the cytosol
of neurons from glutamine or α-ketoglutarate [1]. Synaptic
vesicles are then loaded with Glu via ATP/H+-dependent
vesicular glutamate transporters (VGLUTs 1–3; SLC17A6,
SLC17A8 and SLC17A7) [2] via Cl− stimulation that can
occur from either side of the vesicle membrane [3–5]
(Scheme 1). Vesicles that fuse with the plasma membrane
during synaptic transmission release Glu into the synaptic
cleft where it can activate ionotropic (iGluR) and metabo-
tropic receptors (mGluR). To terminate the transmission,
high-afnity, Na+-dependent protein transporters (EAATs
1–5) remove Glu from the synapse by sequestering it in
glia and neurons (Scheme 1) [6–11].
only one amino acid mimic is much less active. In this
study, we frst evaluated the possibility of dual binding
sites by conducting preliminary docking of VGLUT2 to
the known inhibitors Trypan Blue and Congo Red (CR) to
determine the interatomic distance between glutamate isos-
tere groups. Using docking information, analogs of CR were
designed and prepared with glutamate isosteres attached to
linker groups that varied in interatomic length to aford the
optimal binding interactions to efectively inhibit VGLUT.
High afnity, selective ligands have been essential to
the pharmacological characterization of the glutamate
receptors and transporters [2, 14], but by comparison far
fewer have been identifed that act at VGLUTs, where
there is an unexplored potential to control vesicle flling,
regulate pre-synaptic quantal size (Fig. 1) [15, 16] and
cause a corresponding change in Glu-mediated signaling
[17]. Although numerous advances have been made toward
understanding VGLUT function, selective ligands for rou-
tine pharmacologic and/or physiologic manipulation of the
transporter have been limited. Considering VGLUTs difer
in function, transport criteria, substrate specifcity and cel-
lular location from EAATs and GluRs suggest they can be
selectively targeted and pharmacologically-manipulated
by small molecules without disrupting other transport
Materials and Methods
Solutions were concentrated by rotary evaporation (10 to
160 torr) or on a high-vacuum manifold (ca. 0.1 to 0.25 torr)
at 25–60 °C. Flash column chromatography was performed
using silica gel (60 Å, particle size 0.043–0.060 mm, EMD
Chemicals). Protocols calling for the cooling of reaction
Scheme 1 Proteins involved in
glutamate neurotransmission
[12, 13] with a focus on the
Glu
Glu
EAAT 2,5
EAAT 3,4
glutamate vesicular transporter
ADP+ Pi
Glu
iGluR
(VGLUT). VGLUT transports
2H+
-
Cl
anionic substrates using the
Glu
membrane potential as the driv-
ATP
Glu
Glu
ing force
-
2H+
Cl
Glu
mGluR
Glu
Glu
Glu
Glu
ꢀꢁ
=
ꢀ
pH + ꢀꢂ
H+
Glu
Glu
Glu
Glu
Glu
Glu
Glu
Glu
Glu
iGluR
-
Cl
Glu
H+
-
Cl
VGLUT
Vesicle fusion
Glu
Glu
mGluR
mGluR2
Glu
Glu
Glu
EAAT 1,2
Glu
Glu
Glu
astrocyte
1 3

496
Neurochemical Research (2021) 46:494–503
H
HO
N
O
OH NH2
O
O
N
CO2H
SO3H
N
N
H
H
N
R
O
HN
Br
H
R
N
H
CO2H
SO3H
2
R = Me; Chicago Sky Blue (3000 nM)
R = OMe; Evans Blue (90 nM)
Ki = 300-500 nM
Bromocryptine (BCT)
25 nM
quinoline dicarboxylates (QDC)
20 -50 ꢀM
HO3S
SO3H
azobiphenyl
Me
N
N
HO2C
H2N
HO3S
SO3H
CO2H
OH NH2
NH2 OH
Me
CO2H
N
N
O
OH NH2
HO3S
SO3H
1-amino-8-naphthol
3,6-disulfonic acid
Ki ~ 940 ꢀM
trans-ACPD
Ki ~ 700 ꢀM
NH2
HO
naphthalene disulfonic acid
Trypan Blue (Ki = 50 nM)
Fig. 1 Structures and potency of select VGLUT inhibitors
mixtures to 0 °C were done so in a brine/ice bath. All rea-
gents and solvents were used as received from commercial
suppliers. NMR spectra were recorded on a Bruker Avance
Software and Methods Used for the Homology
Modeling and Docking Studies
III (1H, 400 MHz; 13C 100 MHz) spectrometer. 1H and 13
C
Homology model and molecular docking were conducted
using YASARA (Version 18.10.11) [35]. A homology
model for VGLUT2 was built using d-galactonate trans-
porter (DgoT) [36] and the model subsequently employed
to conduct docking experiments. VGLUT2 secondary struc-
ture was predicted using PSI-Pred algorithm [37, 38] and
produced 21 homology models that were further refned to
a model with a quality Z-core of -1.14. Molecular dock-
ing in the VGLUT2 homology model was performed using
YASARA using customized versions of AutoDock and
VINA [39, 40]. Inhibitors were built using standard bond
lengths and angles and minimized at pH 7.4 in the YAM-
BER force feld. Clusters of poses were obtained that were
scored in binding energy inclusive of typical binding inter-
actions (hydrogen bonding, hydrophobic, polar, etc.) and
repulsive forces. Substrate binding domains were initially
identifed by comparison to reported models [3, 5, 19, 26,
41, 42]. The simulation cell coordinates spanned two-thirds
the length of VGLUT2 (Y-axis~50 Å) with X-axis=30 Å
and Z-axis = 40 Å inclusive of each of the following ten
residues: R88, H128, R184, E191, H199, R211, R322,
Y327, H356 and T454. The docking results represent the
top energy binding pose some of which were exported into
2D protein ligand map programs to identify interactions,
and visualized in PyMol ver 2.1 (Schrodinger) or YASARA.
Some binding orientations, docked molecules and structure
overlays for VGLUT2-ligand displays had occluding protein
NMR chemical shifts (δ) are reported in parts per million
(ppm) and referenced to the solvent (13C) or the residual pro-
ton resonance of the solvent (1H) as reported in literature.23
UV–Vis absorption and fuorescent spectra were recorded
on a Synergy Mx microplate reader. High-resolution mass
spectra were obtained by the Mass Spectrometry Facility of
the Core Laboratory for Neuromolecular Production at the
University of Montana. Melting / decomposition points are
uncorrected and were recorded on a Melt Temp II apparatus
(Laboratory Devices Inc.).
Synthesis of the Compounds
The synthetic procedures and full characterizations are pro-
vided in the Online Resources.
Glutamate Vesicular Uptake Assay
Whole rat brains were purchased from Innovative Research
(Novi, MI USA), the synaptic vesicles isolated, and the
glutamate uptake assay conducted as reported previously
[32–34] in which the control studies were done in the
absence of ATP. Intact rat brain vesicles contain more than
one VGLUT isoform. All eforts were made to minimize
animal sufering and to reduce the number of animals used
in this study.
1 3

Neurochemical Research (2021) 46:494–503
497
domains made transparent to assist with clarity. Additional
details on the procedures and methods used for the compu-
tational studies are provided in the Online Resources.
bearing dual Glu mimics would also be likely to block Pi
transport.
Docking of TB led to a high binding energy pose
(~16 kcal/mol) with the inward pointing naphthalene disul-
fonate (glutamate mimic) interacting with R88 and T454,
and a second naphthalene disulfonate glutamate mimic
forming interactions near the cytosolic entry or vestibule
binding site with R211 and other residues (Fig. 2) [41]. Like-
wise, the high energy binding pose for CR reveals inter-
actions between the sulfonate and R88 and also a second
ionic interaction between the lower sulfonate with R211.
The initial docking experiments support aryl disulfonates
as glutamate mimics binding these residues. Whereas TB
contains four substituents on each naphthalene moiety, CR
has only two, and the biphenyl group is unsubstituted in CR.
Symmetric structures capable of engaging a second bind-
ing site may be advantageous to enhance the selectivity at
VGLUTs. If two distinct binding domains were present,
ligands constructed with a matched inter-domain distance
may display improved binding over other known inhibitor
types. In this study, a panel of new structures were designed
containing dual carboxylate mimics separated by a linker
moiety to vary the interatomic distance and potentially capi-
talize on simultaneously binding two domains in VGLUT.
Results and Discussion
VGLUT2 Homology Model
To investigate the possibility of multiple binding domains
for glutamate, a homology model for VGLUT2 based on the
d-galactonate transporter (DgoT) [36] was constructed and
docking performed with TB and Congo Red (CR) (Fig. 2)
[41]. The VGLUT2-DgoT homology structure is similar
to that previously reported from GlpT [19, 26] and the
outward-facing, open, substrate-bound conformer (6E9O)
reveals a long channel lined with ionizable residues. In the
computational studies, residues considered important to
docking were derived from prior models and reports that
VGLUTs contain two anion binding sites and one cation
binding site that permit vesicle flling to adjust to changing
ionic conditions [5]. It should be noted that during the sub-
mission of this work, a structure for rat VGLUT2 obtained
by cryo-electron microscopy was reported at 3.8-angstrom
resolution [43]. Future studies will utilize the cryo-EM data
and compare with the DgoT model. It was also understood
that the modeling takes into account that VGLUTs have
been shown to transport glutamate and inorganic phosphate
(Pi) using a single, substrate binding site although coupling
to cation gradients difer [44]. Although the possibility of
a second binding domain would not alter the Glu and Pi
transport mechanisms, VGLUT occupancy by an inhibitor
Congo Red Analog Design and Syntheses
Due to structural simplicity and known VGLUT inhibition
(Ki ~550 μM), CR was selected as a scafold to develop the
new analogs (Fig. 3). The two 4-aminonaphthylsulfonic acid
(ANS) groups present in CR were considered as glutamate
isosteres to probe dual binding sites, and the intramolecular
Fig. 2 VGLUT2 homology
model with TB (cyan) and CR
spanning the part of the channel
while interacting with diferent
ionizable residues (gray); bot-
tom faces cytoplasm (cutaway
front domains)
1 3

498
Neurochemical Research (2021) 46:494–503
Fig. 3 Structure of Congo Red
(2a) and proposed analogs
NH2
SO3H
NH2
NH2
N
N
N
linker
N
N
N
NH2
SO3H
SO3H
n
N
N
Congo Red: linker = biphenyl (PhPh); n = 1
New analogs: linker = Ph, PhOPh, PhSPh,
SO3H
Congo Red: 550 nM
PhSO2Ph, PhCH2Ph, PhCH2CH2Ph,
PhCH=CHPh;Ph
C C Ph where n = 0, 1.
distance between these isosteres could be manipulated
by replacement of the biphenyl linker with spacer groups
including phenyl, diphenyl methyl, diphenyl ether, diphenyl
thioether, diphenyl sulfone, diphenyldisulfde, 1,2-diphe-
nylethane, 1,2-diphenylethylene and 1,2-diphenylethyne
(Fig. 3). The diazo moiety was retained for structure–activity
comparisons to CR and due to the known reliable chemistry
connecting the linkers to substituted naphthalenes.
of solid NaCl. For some products, silica gel column chro-
matography using CH2Cl2/CH3OH (9:1 to 4:1) was used to
purify the analogs. The diphenylether analog 2d was isolated
cleanly without additional purifcation. The side product
from the coupling reactions was the mono-addition product
3. In the preparation of 2c and 2g, sufcient quantity of the
mono-addition adduct was isolated for characterization and
VGLUT screening.
Altering the linker properties could also help character-
ize favorable interactions within the interdomain region. As
controls, select structures were also prepared with only one
aminonaphthalene sulfonic acid attached to a linker. The
syntheses, in vitro inhibition of VGLUT, preliminary struc-
ture–activity relationships, and molecular modeling studies
were conducted to support these binding hypotheses. Con-
sidering these analogs may possess emission spectra that
could serve as reporters of VGLUT activity and ligand bind-
ing, we briefy investigated the UV–Vis spectral properties.
The syntheses of the mono- and bis-containing ami-
nonaphthyl sulfonic acids were conducted as shown in
Scheme 2. Diamine 1 was diazotized with NaNO2, the unre-
acted HNO2 quenched with urea, and the mixture added to a
basic solution of ANS. The crude bis-ANS products 2 pre-
cipitated for most reactions, and could be collected by fltra-
tion. In other cases, precipitation was induced by addition
In general, isolated yields for bis-linked analogs 2 were
low (<25%) due to incomplete reaction, competing reduc-
tion, diazotization decomposition, and chromatographic loss.
However, these reactions were amenable to scale-up and the
desired compounds were easily isolated. Compounds 2a and
2c–j representing the bis-linked analogs were character-
ized by 1H and 13C-NMR, mass spectrometry, and UV–Vis
spectroscopy (see: Online Resources). As noted, compounds
containing one ANS group, i.e., analogs 3a–d and 3i repre-
sent important controls lacking the second isotere. In order
to expand this analog set, three additional unsymmetric ana-
logs were synthesized using the same reaction conditions
(Scheme 2) with the monoamine coupling partners aniline,
4-phenylaniline, and 4-phenoxyaniline as diazotization sub-
strates. The resultant mono-substituted analogs bearing azo-
phenyl (3a), azobiphenyl (3b) and azo diphenyl ether (3d)
were isolated in 30–40% yield. Compound 3b represents
Scheme 2 Synthesis of Congo
1. NaNO2, HCl,
Red analogs
NH2
NH2
NH2
N
N
N
0 C, urea
N
X
N
X
N
H2N X NH2
+
2. 4-ANS,
Na2CO3, 0 C
1
SO3H
;
SO3H
3
SO3H
;
2
H2
;
O
d
C
a
b
c
X =
S
;
;
S
e
f
g
1 3

Neurochemical Research (2021) 46:494–503
499
the truncated mono-substituted analog of Congo Red. The
synthetic analogs gave UV–Vis absorption spectra with
extinction coefcients~104 M−1 cm−1, and most exhibited
emission bands between 415 and 440 nm when excited at
the maximum UV absorbance wavelength (c.f., Congo Red
λmax = 340, 498 nm with an λex at 417 nm [45]), however,
none were suitable for spectrometric assay development.
while 4-ANS proved to be only a moderate to weak inhibitor
itself (≈ 50% inhibition at 100 µM vs 250 µM 3H-l-Glu),
the presence of two ANS groups linked via a biphenyl moi-
ety (i.e., CR, 2b) resulted in a marked increase in inhibi-
3
tory activity (IC50 = 0.6 0.2 µM with 250 µM H-l-Glu)
in agreement with more recent reports [49]. All of the com-
pounds tested showed some inhibition of VGLUT with some
preliminary diferences observed between the compounds
containing two ANS moieties 2a–2j (symmetric) or those
with one 3a–d, 3i (asymmetric). Compounds 2a, 2c–2j
reduced VGLUT uptake from 63–100% inhibition. Based
on the VGLUT2 model, the bis-aryl or longer linkers were
anticipated to be more potent owing to similarity to azo dyes
such as Trypan Blue, however, compound 2a bearing a sin-
gle phenyl linker was the most efective inhibitor exhibiting
an IC50 ≈ 300 nM. To this point, compounds 2d (–PhOPh–),
2g (–PhCH2CH2Ph–) and 2j (–PhC≡CPh–) with slightly
longer linkers were also among the more potent inhibitors
in this new series. While compound 2c with a diphenylmeth-
ylene linker and compounds 2ef (thioether and sulfone) were
only tested at a single concentration, the resulting levels of
inhibiton demonstrate all are less potent than 2d, indicating
the importance of a possible interaction between VGLUT
residues and the linker oxygen atom. Among ligands with
two atom spacers between phenyl groups, compound 2g
(–CH2CH2–), which exhibited an IC50 ≈ 2.3 µM was a more
efective inhibitor than the trans-stibene 2h, disulfde 2i or
alkyne 2j. Although 2g would be expected to have greater
conformational fexibility owing to its ethane bridge, than
the other three, this fexibility alone cannot account for the
favorable interaction with VGLUT.
Inhibition of VGLUT by the Congo Red Analogs
CR analogs were screened as VGLUT inhibitors anticipat-
ing that analogs with two ANS isosteres would show greater
potency if the proper interatomic positions were obtained.
3
The VGLUT assay [32, 46, 47] quantifes H-l-glutamate
transported into synaptic vesicles, with the amount of radio-
activity correspondingly reduced by inhibitors. All values
have been corrected for non-specifc uptake as determined
3
by the accumulation of H-l-glutamate in the absence
of ATP. The activities of the analogs were frst screened
at a concentration of 10 µM against 250 µM l-glutamate
and reported as percent of control (i.e., 100% glutamate
uptake=no inhibition and 0% uptake=complete blockade of
glutamate uptake into vesicles). A more detailed IC50 evalu-
ation was conducted on select analogs that reduced gluta-
mate uptake to less than 10% of control (Table 1). Previously
reported values for Congo Red (2b) and ANS were included
for comparison [48, 49].
Prior work characterized the activity of 4-ANS and CR
(2b) as inhibitors of VGLUT-mediated uptake [38]. Thus,
Table 1 Inhibition of VGLUT by Congo Red Analogs
In general, the asymmetric mono-ANS analogs exhibited
less than or similar levels of VGLUT inhibition when com-
pared with the corresponding bis-ANS analogs, although
generalization such as these should be interpreted cautiously,
as comparisons of % uptake (or % inhibition) are based on
activity determinations at single concentrations. The diphe-
nylmethylene and diphenyldisulfide were equivalent in
activity indicating, at least with these derivatives, that no
additional favorable interactions occurred with the inclu-
sion of a second ANS group. ANS, which lacked any diazo
linkers, showed modest inhibition of VGLUT (~50%), while
the mono-substituted structures with phenyl 3a and biphenyl
3b groups did not appear to enhance binding. Extending
the linker length in compounds 3cdi improved inhibition
likely through additional hydrophobic interactions. Still, the
fnding that the shortest symmetric analog 2a was the most
potent blocker of VGLUT was highly unexpected and may
help advance the development of future inhibitor structures.
Further investigation was conducted to identify the activ-
ity of compounds 2a, 2d and 2g that appear comparable if
not greater than Congo Red as VGLUT inhibitors. Detailed
dose/response analyses for 2a (Fig. 4; IC50 plot shown),
Cmpd
Symmetric, dual ANS
groups
VGLUT VGLUT IC50 (μM)
% uptakea
2a
2c
2d
2e
2f
2 g
2 h
2i
–Ph–
–PhCH2Ph–
–PhOPh–
–PhSPh–
–PhSO2Ph–
–PhCH2CH2Ph–
–PhCH=CHPh– (trans) 24
–PhSSPh–
–PhC≡CPh–
0
35
6
0
0.3 0.05
–
1.5 1.15
3
3
27
37
6
5
6
–
–
2
2.3 0.3
–
–
3
6
3
28
18
2j
Asymmetric, single ANS group
3a
3b
3c
3d
3i
ANS [48]
Ph
58
59
39
28.5
29
4
2
2
–
–
–
–
–
–
–Ph–Ph
–PhCH2Ph
–PhOPh
–PhSSPh–
7
6
54
aControl uptake=100%. VGLUT assay concentrations: l-gluta-
mate=250 μM, inhibitors tested at 10 μM. All data n≥3 except 3d;
n=2)
1 3

500
Neurochemical Research (2021) 46:494–503
ANS moiety, retained some ability to block VGLUT activity.
To investigate this, fve additional ANS-diazophenyl analogs
were prepared and tested based on the structure of 3a. When
assayed at VGLUT, the p-methyl (51% 11), p-hydroxy
(42% 4), p-methoxy (63% 6), p-bromo (65% 5), and
p-iodo (56% 3) 3a analogs reduced uptake levels similarly.
Docking did not reveal any deeper insights for these ana-
logs but the p-hydroxy analog of 3a can form an additional
hydrogen bonding interaction with VGLUT.
The role of the sulfonate group in CR as a glutamate
mimic was also briefy investigated. Analogs in which the
sulfonic acid group of ANS and compound 3a were sub-
stituted for a sulfonamide were prepared and tested. The
reasoning for this exchange was to alter the pKa and degree
of ionization, and consequently to attempt to change the
interaction with R88 from ionic to hydrogen bonding and/or
assess the role of Tyr195 and Tyr327. However, neither sul-
fonamide showed any activity as an inhibitor at the concen-
trations tested possibly suggesting the importance of charge
interactions. Also, no analogs were prepared with substitu-
ents on the linker because benzopurpurin 4B, the 2,2′-bis-
methyl diphenyl analog of CR is less active IC50 ~680 μM
[51] potentially due to reduced pi-stacking interactions.
Fig. 4 Concentration dependent inhibition of VGLUT by compounds
2a, 2g and 2d
2d and 2g aforded IC50 values of 300 nM, 1500 nM and
2300 nM, respectively, with plots consistent with competi-
tive inhibition. Only 2a was more potent than Congo Red
(Ki~490 nM; IC50 =550 nM) [48] (Table 1). Attempts were
made to assess compound 2j, but poor solubility limited this
experiment.
Conclusions
Acher et al. [52] reported a similar approach with modif-
cation of the linker but using the naphthalene scafold of
Trypan Blue rather than Congo Red. They were able to show
that bi-aryl derivatives and 8-hydroxylated substituents are
key to inhibitory activity—a useful and crucial fnding as it
aligns with our prior work using lipophilic groups improve
the activity quinoline-2,4-dicarboxylates as inhibitors [53,
54].
Docking of Congo Red Analog 2a (Table 2)
Since compound 2a proved to be a more potent inhibitor
than CR, despite being shorter, docking experiments
in YASARA [35] were conducted to assess possible
interactions. The docking revealed a relatively high
energy binding orientation for 2a (12.6 kcal/mol) with
ionic interactions identifed between sulfonates and R88
and H199 (Table 2; Entries 1 and 2) but not R211, which
was found to be important for CR binding. A 2D protein–
ligand interaction profle (PLIP) [50] revealed additional
interactions and distances for ionic interactions His199
(3.8 Å) and R88 (4.3 Å) and hydrophobic, hydrogen
bonding and a key pi-stacking interaction with Phe219
(Table 2). Overall, the docking uncovered a number of
interactions between 2a and VGLUT that support the
unexpected inhibitory potency of this analog; although,
additional analogs will need to be prepared to fully assess
the validity of these interactions.
Brilliant Yellow (BY) is among the most potent VGLUT
inhibitors yet identifed and possesses sulfonic acid groups
on the linker portion. Ueda et al. used a clever substitution
of sulfonate to carboxylate and showed these BY analogs
retained excellent inhibitory activity [55]. Molecular mode-
ling indicated conserved distances between the azenyl nitro-
gen and sulfonate oxygens for the more potent inhibitors.
As speculated by the authors [55], this arrangement might
engage Tyr195, which was an insightful prediction that is
supported by our docking study showing clear involvement
of Tyr195 at both azo and sulfonates (Table 2). We are cur-
rently investigating new chemical entities that will utilize
these collective fndings to identify increasingly potent and
selective inhibtors of VGLUT.
The pi-stacking interaction between 2a and Phe219
(Entry 5) may explain, in part, why 3a, which lacks a second
1 3

Neurochemical Research (2021) 46:494–503
501
Table 2 Compound 2a docked in VGLUT2 (DgoT homology model) showing sulfonate groups spanning R88 and H199
Entry
1
Residue (type)
Distance in Å (FG)
Arg88
(SB/I)
4.3
(SO3H)
2
3
4
5
6
7
His199
(SB/I)
3.8
(SO3H)
Tyr195
3.2, 3.6 and 3.9 (SO3H, N=N)
(HB)
Tyr327
(HB)
2.4–2.6
(SO3H)
Phe219
4.0
(PS,HP)
(Ph)
Ile427
3.0
(HB)
(N=N)
Phe326
(HP)
3.9
(naphthyl)
Residues, interactions and distances obtained from PLIP [50]
SB salt bridge, I ionic, HB hydrogen bond, PS pi-stacking, HP hydrophobic, FG functional group
Acknowledgements The authors wish to thank Dr David Holley for
assistance with the molecular modeling and the University of Mon-
tana’s Center for Structural & Functional Neuroscience. Research in
the authors laboratories is funded by a grant to The Center for Biomo-
lecular and Structural Dynamics from the National Institutes of Health,
P20GM103546.
Compliance with Ethical Standards
Conflict of interest The authors declare no competing fnancial inter-
est.
References
Author Contributions The manuscript was written by C.M.T. with
contributions by all authors. DH, JZ, SP and C-KC conducted the
investigation, methodology and formal analyses. Conceptualization
was conducted by JG, RB and CT. All authors have given approval to
the fnal version of the manuscript.
1. Takeda K, Ishida A, Takahashi K, Ueda T (2012) Synaptic vesi-
cles are capable of synthesizing the VGLUT substrate glutamate
from alpha-ketoglutarate for vesicular loading. J Neurochem
121:184–196
2. Bridges RJ, Patel SA (2009) Pharmacology of Glutamate Trans-
port in the CNS: substrates and inhibitors of excitatoryamino acid
transporters (EAATs) and the glutamate/cystine exchanger system
xc−. In: Napier S, Bingham M (eds) Transporters as targets for
drugs. Springer, Berlin , pp 187–222
Funding Financial support for University of Montana Molecular Com-
putational and Magnetic Resonance Core Facilities utilized in this
study was provided by NIH NIGMS P20GM103546 (S. Sprang, PI)
and gratefully acknowledged.
3. Juge N, Gray JA, Omote H, Miyaji T, Inoue T, Hara C, Uney-
ama H, Edwards RH, Nicoll RA, Moriyama Y (2010) Metabolic
1 3

502
Neurochemical Research (2021) 46:494–503
24. Patching SG, Psakis G, Baldwin SA, Baldwin J, Henderson PJ,
Middleton DA (2008) Relative substrate afnities of wild-type
and mutant forms of the Escherichia coli sugar transporter GalP
determined by solid-state NMR. Mol Membr Biol 25:474–484
25. Yang Q, Wang X, Ye L, Mentrikoski M, Mohammadi E, Kim YM,
Maloney PC (2005) Experimental tests of a homology model for
OxlT, the oxalate transporter of Oxalobacter formigenes. Proc
Natl Acad Sci USA 102:8513–8518
control of vesicular glutamate transport and release. Neuron
68:99–112
4. Wolosker H, de Souza DO, de Meis L (1996) Regulation of gluta-
mate transport into synaptic vesicles by chloride and proton gradi-
ent. J Biol Chem 271:11726–11731
5. Preobraschenski J, Zander JF, Suzuki T, Ahnert-Hilger G, Jahn
R (2014) Vesicular glutamate transporters use fexible anion and
cation binding sites for efcient accumulation of neurotransmitter.
Neuron 84:1287–1301
26. Juge N, Yoshida Y, Yatsushiro S, Omote H, Moriyama Y (2006)
Vesicular glutamate transporter contains two independent trans-
port machineries. J Biol Chem 281:39499–39506
6. Gonzalez MI, Robinson MB (2004) Neurotransmitter transporters:
why dance with so many partners? Curr Opin Pharmacol 4:30–35
7. Hinoi E, Takarada T, Tsuchihashi Y, Yoneda Y (2005) Glutamate
transporters as drug targets. Curr Drug Targets CNS Neurol Dis-
ord 4:211–220
27. Adler J, Bibi E (2004) Determinants of substrate recognition by
the Escherichia coli multidrug transporter MdfA identifed on
both sides of the membrane. J Biol Chem 279:8957–8965
28. Adler J, Lewinson O, Bibi E (2004) Role of a conserved mem-
brane-embedded acidic residue in the multidrug transporter
MdfA. Biochemistry 43:518–525
8. Moriyama Y, Yamamoto A (2004) Glutamatergic chemical trans-
mission: look! Here, there, and anywhere. J Biochem 135:155–163
9. Reimer RJ, Edwards RH (2004) Organic anion transport is the
primary function of the SLC17/type I phosphate transporter fam-
ily. Pfugers Arch 447:629–635
29. Adler J, Bibi E (2005) Promiscuity in the geometry of electro-
static interactions between the Escherichia coli multidrug resist-
ance transporter MdfA and cationic substrates. J Biol Chem
280:2721–2729
10. Reimer RJ, Fremeau RT Jr, Bellocchio EE, Edwards RH (2001)
The essence of excitation. Curr Opin Cell Biol 13:417–421
11. Takamori S (2006) VGLUTs: “exciting” times for glutamatergic
research? Neurosci Res 55:343–351
30. Tsigelny IF, Greenberg J, Kouznetsova V, Nigam SK (2008)
Modeling of glycerol-3-phosphate transporter suggests a poten-
tial “tilt” mechanism involved in its function. J Bioinform Comput
Biol 6:885–904
12. Shigeri Y, Seal RP, Shimamoto K (2004) Molecular pharmacol-
ogy of glutamate transporters, EAATs and VGLUTs. Brain Res
Brain Res Rev 45:250–265
31. Varoqui H, Erickson JD (1997) Vesicular neurotransmitter trans-
porters. Potential sites for the regulation of synaptic function. Mol
Neurobiol 15:165–191
13. Zhou Y, Danbolt NC (2013) GABA and glutamate transporters in
brain. Front Endocrinol 4:165
32. Kish PE, Ueda T (1989) Glutamate accumulation into synaptic
vesicles. Methods Enzymol 174:9–25
14. Kew JN, Kemp JA (2005) Ionotropic and metabotropic gluta-
mate receptor structure and pharmacology. Psychopharmacology
179:4–29
33. Carrigan CN, Esslinger CS, Bartlett RD, Bridges RJ, Thompson
CM (1999) Quinoline-2,4-dicarboxylic acids: synthesis and evalu-
ation as inhibitors of the glutamate vesicular transport system.
Bioorg Med Chem Lett 9:2607–2612
15. Goh GY, Huang H, Ullman J, Borre L, Hnasko TS, Trussell LO,
Edwards RH (2011) Presynaptic regulation of quantal size: K+/
H+ exchange stimulates vesicular glutamate transport. Nat Neu-
rosci 14:1285–1292
34. Carrigan CN, Patel SA, Cox HD, Bolstad ES, Gerdes JM, Smith
WE, Bridges RJ, Thompson CM (2014) The development of
benzo- and naphtho-fused quinoline-2,4-dicarboxylic acids as
vesicular glutamate transporter (VGLUT) inhibitors reveals a
possible role for neuroactive steroids. Bioorg Med Chem Lett
24:850–854
16. Wilson NR, Kang J, Hueske EV, Leung T, Varoqui H, Murnick
JG, Erickson JD, Liu G (2005) Presynaptic regulation of quantal
size by the vesicular glutamate transporter VGLUT1. J Neurosci
25:6221–6234
17. Chaudhry FA, Edwards RH, Fonnum F (2008) Vesicular neuro-
transmitter transporters as targets for endogenous and exogenous
toxic substances. Annu Rev Pharmacol Toxicol 48:277–301
18. Huang Y, Lemieux MJ, Song J, Auer M, Wang DN (2003) Struc-
ture and mechanism of the glycerol-3-phosphate transporter from
Escherichia coli. Science 301:616–620
35. Krieger E, Vriend G (2014) YASARA View—molecular graphics
for all devices—from smartphones to workstations. Bioinformat-
ics 30:2981–2982
36. Leano JB, Batarni S, Eriksen J, Juge N, Pak JE, Kimura-Someya
T, Robles-Colmenares Y, Moriyama Y, Stroud RM, Edwards RH
(2019) Structures suggest a mechanism for energy coupling by a
family of organic anion transporters. PLoS Biol 17:e3000260
37. Jones DT (1999) Protein secondary structure prediction based on
position-specifc scoring matrices. J Mol Biol 292:195–202
38. Jones DT, Taylor WR, Thornton JM (1994) A model recognition
approach to the prediction of all-helical membrane protein struc-
ture and topology. Biochemistry 33:3038–3049
19. Almqvist J, Huang Y, Laaksonen A, Wang DN, Hovmoller
S (2007) Docking and homology modeling explain inhibi-
tion of the human vesicular glutamate transporters. Protein Sci
16:1819–1829
20. Eudes A, Kunji ER, Noiriel A, Klaus SM, Vickers TJ, Bever-
ley SM, Gregory JF 3rd, Hanson AD (2010) Identifcation of
transport-critical residues in a folate transporter from the folate-
biopterin transporter (FBT) family. J Biol Chem 285:2867–2875
21. Jeon J, Yang JS, Kim S (2009) Integration of evolutionary fea-
tures for the identifcation of functionally important residues in
major facilitator superfamily transporters. PLoS Comput Biol
5:e1000522
39. Trott O, Olson AJ (2010) AutoDock Vina: improving the speed
and accuracy of docking with a new scoring function, efcient
optimization, and multithreading. J Comput Chem 31:455–461
40. Morris GM, Huey R, Lindstrom W, Sanner MF, Belew RK, Good-
sell DS, Olson AJ (2009) AutoDock4 and AutoDockTools4: auto-
mated docking with selective receptor fexibility. J Comput Chem
30:2785–2791
22. Khare P, Mulakaluri A, Parsons SM (2010) Search for the ace-
tylcholine and vesamicol binding sites in vesicular acetylcholine
transporter: the region around the lumenal end of the transport
channel. J Neurochem 115:984–993
41. Thompson CM, Chao CK (2020) VGLUT substrates and
inhibitors: a computational viewpoint. Biochim Biophys Acta
Biomembr 183175
23. Law CJ, Enkavi G, Wang DN, Tajkhorshid E (2009) Structural
basis of substrate selectivity in the glycerol-3-phosphate: phos-
phate antiporter GlpT. Biophys J 97:1346–1353
42. Zhang FX, Ge SN, Dong YL, Shi J, Feng YP, Li Y, Li YQ, Li
JL (2018) Vesicular glutamate transporter isoforms: the essential
players in the somatosensory systems. Prog Neurobiol 171:72–89
1 3

Neurochemical Research (2021) 46:494–503
503
43. Li F, Eriksen J, Finer-Moore J, Chang R, Nguyen P, Bowen A,
Myasnikov A, Yu Z, Bulkley D, Cheng Y, Edwards RH, Stroud
RM (2020) Ion transport and regulation in a synaptic vesicle glu-
tamate transporter. Science 368:893–897
biphenyl and amino- and sulphonic acid-substituted naphthyl
groups. Biochem Pharmacol 56:1243–1249
52. Favre-Besse FC, Poirel O, Bersot T, Kim-Grellier E, Daumas S, El
Mestikawy S, Acher FC, Pietrancosta N (2014) Design, synthesis
and biological evaluation of small-azo-dyes as potent Vesicular
Glutamate Transporters inhibitors. Eur J Med Chem 78:236–247
53. Carrigan CN, Bartlett RD, Esslinger CS, Cybulski KA, Tongchar-
oensirikul P, Bridges RJ, Thompson CM (2002) Synthesis and
in vitro pharmacology of substituted quinoline-2,4-dicarboxylic
acids as inhibitors of vesicular glutamate transport. J Med Chem
45:2260–2276
44. Preobraschenski J, Cheret C, Ganzella M, Zander JF, Richter K,
Schenck S, Jahn R, Ahnert-Hilger G (2018) Dual and direction-
selective mechanisms of phosphate transport by the vesicular
glutamate transporter. Cell Rep 23:535–545
45. Iwunze MO (2010) Aqueous photophysical parameters of Congo
Red. Spectrosc Lett 43:16–21
46. Carlson MD, Kish PE, Ueda T (1989) Glutamate uptake into syn-
aptic vesicles: competitive inhibition by bromocriptine. J Neuro-
chem 53:1889–1894
54. Thompson CM, Davis E, Carrigan CN, Cox HD, Bridges RJ, Ger-
des JM (2005) Inhibitor of the glutamate vesicular transporter
(VGLUT). Curr Med Chem 12:2041–2056
47. Naito S, Ueda T (1985) Characterization of glutamate uptake into
synaptic vesicles. J Neurochem 44:99–109
55. Tamura Y, Ogita K, Ueda T (2014) A new VGLUT-specifc potent
inhibitor: pharmacophore of Brilliant Yellow. Neurochem Res
39:117–128
48. Bartlett RD (1999) Identifcation and characterization of inhibi-
tors of L-glutamate transport into rat brain synaptic vesicles. PhD
Thesis; The University of Montana, Missoula
49. Neale SA, Copeland CS, Salt TE (2014) Efect of VGLUT inhib-
itors on glutamatergic synaptic transmission in the rodent hip-
pocampus and prefrontal cortex. Neurochem Int 73:159–165
50. Salentin S, Schreiber S, Haupt VJ, Adasme MF, Schroeder M
(2015) PLIP: fully automated protein-ligand interaction profler.
Nucleic Acids Res 43:W443-447
Publisher’s Note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional afliations.
51. Roseth S, Fykse EM, Fonnum F (1998) Uptake of L-glutamate
into synaptic vesicles: competitive inhibition by dyes with
Afliations
David M. Hitt1,2 · Jefery D. Zwicker1,3 · Chih‑Kai Chao1 · Sarjubhai A. Patel1 · John M. Gerdes1
Richard J. Bridges1 · Charles M. Thompson1
·
David M. Hitt
Richard J. Bridges
davidmhitt@gmail.com
richard.bridges@mso.umt.edu
Jefery D. Zwicker
1
Department of Biomedical and Pharmaceutical Sciences,
Center for Structural and Functional Neuroscience, College
of Health Professions and Biomedical Sciences, University
of Montana, Missoula, MT 59812, USA
zwickjd@gmail.com
Chih-Kai Chao
chihkai.chao@mso.umt.edu
2
3
Sarjubhai A. Patel
Present Address: Department of Chemistry, Carroll College,
1601 N Benton Ave., Helena, MT 59625, USA
sarjubhai.patel@umontana.edu
John M. Gerdes
Present Address: Deciphera Pharmaceuticals, 643
Massachusetts St, Lawrence, KS 66044, USA
jgerdesum@gmail.com
1 3