Design and Mechanism of Tetrahydrothiophene-Based γ-Aminobutyric Acid Aminotransferase Inactivators

Article abstract of DOI:10.1021/jacs.5b01155

Low levels of γ-aminobutyric acid (GABA), one of two major neurotransmitters that regulate brain neuronal activity, are associated with many neurological disorders, such as epilepsy, Parkinson's disease, Alzheimer's disease, Huntington's disease, and cocaine addiction. One of the main methods to raise the GABA level in human brain is to use small molecules that cross the blood-brain barrier and inhibit the activity of γ-aminobutyric acid aminotransferase (GABA-AT), the enzyme that degrades GABA. We have designed a series of conformationally restricted tetrahydrothiophene-based GABA analogues with a properly positioned leaving group that could facilitate a ring-opening mechanism, leading to inactivation of GABA-AT. One compound in the series is 8 times more efficient an inactivator of GABA-AT than vigabatrin, the only FDA-approved inactivator of GABA-AT. Our mechanistic studies show that the compound inactivates GABA-AT by a new mechanism. The metabolite resulting from inactivation does not covalently bind to amino acid residues of GABA-AT but stays in the active site via H-bonding interactions with Arg-192, a π-π interaction with Phe-189, and a weak nonbonded S......O=C interaction with Glu-270, thereby inactivating the enzyme. (Figure Presented).

Full text of DOI:10.1021/jacs.5b01155

Subscriber access provided by SUNY DOWNSTATE
Article
Design and Mechanism of Tetrahydrothiophene-
based GABA Aminotransferase Inactivators
Hoang V. Le, Dustin D Hawker, Rui Wu, Emma H. Doud, Julia Reed Widom,
Ruslan Sanishvili, Dali Liu, Neil L. Kelleher, and Richard B. Silverman
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.5b01155 • Publication Date (Web): 17 Mar 2015
Downloaded from http://pubs.acs.org on March 23, 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.
Journal of the American Chemical Society 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 12
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Design and Mechanism of Tetrahydrothiophene-based GABA
Aminotransferase Inactivators
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
Hoang V. Le,1 Dustin D. Hawker,1 Rui Wu,2 Emma Doud,3 Julia Widom,1 Ruslan Sanishvili,4 Dali
Liu,2 Neil L. Kelleher,3 Richard B. Silverman*,1
1 Departments of Chemistry and Molecular Biosciences, Chemistry of Life Processes Institute, and the Center for
Molecular Innovation and Drug Discovery, Northwestern University, Evanston, IL 60208
2 Department of Chemistry and Biochemistry, Loyola University Chicago, Chicago, IL 60660
3 Departments of Chemistry and Molecular Biosciences, and the Proteomics Center of Excellence, Northwestern
University, Evanston, IL 60208
4 X-ray Science Division, Advanced Photon Source, Argonne National Laboratory, Lemont, IL 60439
ABSTRACT: Low levels of γ-aminobutyric acid (GABA), one of two major neurotransmitters that regulate brain neuronal
activity, are associated with many neurological disorders, such as epilepsy, Parkinson’s disease, Alzheimer’s disease, Hun-
tington’s disease, and cocaine addiction. One of the main methods to raise the GABA level in human brain is to use small
molecules that cross the blood-brain barrier and inhibit the activity of γ-aminobutyric acid aminotransferase (GABA-AT),
the enzyme that degrades GABA. We have designed a series of conformationally-restricted, tetrahydrothiophene-based
GABA analogs with a properly-positioned leaving group that could facilitate a ring-opening mechanism, leading to inacti-
vation of GABA-AT. One compound in the series is eight times more efficient an inactivator of GABA-AT than vigabatrin,
the only FDA-approved inactivator of GABA-AT. Our mechanistic studies show that the compound inactivates GABA-AT
by a new mechanism. The metabolite resulting from inactivation does not covalently bind to amino acid residues of
GABA-AT but stays in the active site via H-bond interactions with Arg-192, a π-π interaction with Phe-189, and a weak
nonbonded S···O=C interaction with Glu-270, thereby inactivating the enzyme.
1. Introduction
from glial cells into the synapse and transported back to
GABAergic neurons to complete the metabolic cycle of L-
glutamate.
Epilepsy is a family of chronic neurological disorders
characterized by recurring convulsive seizures, which
result from abnormal, excessive neuronal activity in the
central nervous system.1 It is estimated that about 65 mil-
lion people worldwide have epilepsy.2 Epilepsy can arise
from an imbalance in two major neurotransmitters that
regulate brain neuronal activity, L-glutamate, an excitato-
ry neurotransmitter, and γ-aminobutyric acid (GABA), an
inhibitory neurotransmitter.3
Low levels of GABA are linked to not only epilepsy,6 but
also many other neurological disorders including Parkin-
son’s disease,7 Alzheimer’s disease,8 Huntington’s dis-
ease,9 and cocaine addiction.10 Raising GABA levels has
proven effective in stopping recurring convulsive seizures
in the treatment of epilepsy.11 However, GABA does not
cross the blood-brain barrier; therefore, an increase in
brain levels of GABA cannot be achieved by intravenous
administration.12 To increase brain levels of GABA, one
could speed up the activity of GAD, the enzyme that
makes GABA, inhibit the activity of GABA-AT, the en-
zyme that degrades GABA, or inhibit the reuptake of
GABA by blocking the action of the GABA transporters.
GABA is produced in GABAergic neurons from L-
glutamate by the enzyme glutamic acid decarboxylase
(GAD) (Figure 1).4,5 GABA is then released into the syn-
apse and transported to glial cells. The enzyme GABA
aminotransferase (GABA-AT) in glial cells degrades GABA
to succinic semialdehyde, which is further oxidized to
succinate and enters the Krebs cycle. In this first half of
the GABA-AT catalytic cycle, cofactor pyridoxal 5'-
phosphate (PLP) is converted to pyridoxamine 5'-
phosphate (PMP). GABA-AT also converts α-
ketoglutarate from the Krebs cycle to L-glutamate and
returns PMP to PLP in the second half of its catalytic cy-
cle. Because there is no GAD in glial cells, this newly
formed L-glutamate is not converted to GABA. It is in-
stead converted to L-glutamine, which is then released
Our laboratory has focused on the design of mechanism-
based inactivators of GABA-AT, unreactive compounds
that require GABA-AT catalysis to convert them into the
species that inactivates the enzyme. Because these mole-
cules are not initially reactive, but require the catalytic
activity of GABA-AT to become activated and form cova-
lent bonds, indiscriminate reactions with off-target pro-
teins, leading to undesired side effects, should be greatly
reduced. Even at lower dosages, these inactivators can
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 12
achieve the desired pharmacologic effects with enhanced
potency and selectivity than conventional inhibitors.13
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. Metabolic cycle of L-glutamate
Scheme 1. Mechanisms of inactivation of GABA-AT by vigabatrin
O
COO
8
Lys 329
NH
B
Lys 329
NH2
PMP
H2O
Lys 329
NH
B
Lys 329
b
H
H b
Lys 329
HN
NH3
3
a
70%
H
a
N
COO
NH
COO
H2N
Pyr
COO
NH
Pyr
COO
a
H
H3N
COO
Pyr
Pyr
H
B
4
Pyr
5
6
7
1
2
b
30%
Pyr = pyridine ring of PLP
Lys 329
HN
Lys 329
NH2
Lys 329
H2N
COO
Pyr
NH
Pyr
COO
H2N
COO
Pyr NH2
9
10
11
H2O
NH4
O
COO
12
Currently, the only FDA-approved inactivator of GABA-
AT is the drug vigabatrin (2) (Scheme 1), which was first
developed by Lippert et al.,14 and is used for the treatment
of epilepsy.15 However, a large dose of vigabatrin (1 – 3 g)
needs to be taken daily,16–18 and there are many serious
side effects that arise from its usage, including psychosis19
and permanent vision loss resulting from the damage of
the retinal nerve fiber layer.20 Therefore, the demand for
an alternative to vigabatrin in the treatment of epilepsy is
urgent.
ketimine 5. An active-site nucleophile then reacts with
Michael acceptor 5 to form 6, which is in equilibrium with
7. In the enamine mechanism, the Schiff base (4) is sub-
jected to γ-proton removal and tautomerization through
the vinyl bond, which leads to the release of enamine 10.
Subsequent nucleophilic addition of 10 to the lysine-
bound PLP on GABA-AT gives rise to 11.
The Michael addition mechanism and the enamine mech-
anism happen concurrently in a 70/30 ratio, respectively.
It was discovered that ketimine 5 in the Michael addition
mechanism, and enamine 10 in the enamine mechanism,
underwent partial hydrolysis to form the α,β-unsaturated
ketone (8) and the saturated ketone (12), respectively.
While 8 is a reactive electrophile, possibly responsible for
some side effects, 12 is not a reactive metabolite. There-
fore, it is imperative to find vigabatrin analogs that either
follow the enamine mechanism exclusively to avoid the
We have found that vigabatrin inactivates GABA-AT via
two pathways: a Michael addition mechanism and an
enamine mechanism, as shown in Scheme 1.21 In the Mi-
chael addition mechanism, the resulting Schiff base (4)
from the reaction of vigabatrin and the lysine-bound PLP
(1) on GABA-AT is subjected to γ-proton (boldfaced hy-
drogen in 2) removal and tautomerization that leads to
ACS Paragon Plus Environment

Page 3 of 12
Journal of the American Chemical Society
F
formation of 8 or speed up the Michael addition pathway
F
1
so that 5 would have much lower probability to undergo
2
3
4
H3N
COO
H3N
H3N
COO
H3N
COO
hydrolysis.
COO
14
16
CPP-115
13
15
5
6
7
8
Figure 2. Vigabatrin analogs that follow one GABA-AT
inactivation mechanism exclusively
Scheme 3. Syntheses of GABA Analogs 17 – 19
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
A.
1. Boc2O
Et3N
CH2Cl2
S
O
O
COOMe
OMe
TMSOTf
Et3N
HCl
SH
S
S
O
O
LiTMP
BocHN
OTMS
O
O
O
THF, -70 oC
CH2Cl2, 0 o
C
BocHN
H2N
BocN
O
O
26
O
O
TMS O
2.
Br
36% (2 steps)
23
24
74%
25
HCl
S
4N HCl
AcOH
S
COOH
COOMe
COOMe
70 oC, 5h
H2N
BocHN
BocHN
17
89%
S
29
25%
COOMe
1. HF/MeOH
TBAF, rt
2. NaBH4, 0 oC
MsCl
Et3N
S
Mg (s)
NH4Cl
MeOH
COOMe
OH
+
BocHN
CH2Cl2
4N HCl
AcOH
HCl
BocHN
S
S
COOH
28
64% (2 steps)
27
70 oC, 5h
H2N
18
87%
30
32%
B.
O O
O O
S
4N HCl
AcOH
MnSO4.H2O
H2O2 (aq)
S
S
COOMe
COOMe
HCl
H2N
COOH
70 oC, 5h
CH3CN
BocHN
BocHN
31
81%
19
29 or 30
86%
An energy minimized molecular model of vigabatrin
bound to PLP in GABA-AT revealed that after tautomeri-
zation, the vinyl bond in 5 needs to rotate toward Lys-329
for the Michael addition to occur.22 Therefore, conforma-
tionally-restricted analogs such as 13 and 14 (Figure 2)
would prevent the rotation of the vinyl bond, thereby
blocking the Michael addition mechanism. Experiments
showed that 13 inactivated GABA-AT following the
enamine mechanism exclusively. However, its potency
remained low. In the alternative approach, conformation-
ally-restricted analogs 15 and 16 have the vinyl bond
readily pointed toward Lys-329 for rapid Michael addition
to occur, thereby minimizing the hydrolysis of the
ketimine intermediate.23 Experiments showed that 16 was
186 times more efficient in inactivating GABA-AT than
vigabatrin. Furthermore, unlike vigabatrin,24 16 did not
inactivate or inhibit off-target enzymes, such as alanine
aminotransferase and aspartate aminotransferase,25 and
therefore is less likely to produce side effects. Indeed, we
tested 16 in a multiple-hit rat model of infantile spasms,26
and the results showed that 16 suppressed spasms at dos-
es of 0.1-1 mg/kg/day, which were >100-fold lower than
those for vigabatrin. The spasms suppression by 16 stayed
effective longer (3 days vs. 1 day for vigabatrin), and 16
also had a much larger margin of safety than vigabatrin.
With those results, 16, now known as CPP-115, was grant-
ed Orphan Drug Designation by the FDA for the treat-
ment of infantile spasms. It recently finished a Phase I
clinical trial.
O
O
S
S
S
H3N
COO
H3N
COO
H3N
COO
18
19
17
O
S
O
S
S
H3N
COO
H3N
COO
H3N
COO
22
20
21
Figure 3. Tetrahydrothiophene-based GABA analogs
In our ongoing effort to develop new classes of antiepilep-
tic drugs, we are interested in new GABA analogs that
inactivate GABA-AT by new mechanisms. Compounds
with a leaving group adjacent to the carbanion formed
after the γ-proton removal seem to inactivate GABA-AT
by an enamine mechanism. Therefore, we synthesized a
series of conformationally-restricted, tetrahydrothio-
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 12
phene-based analogs (Figure 3), which have a properly- inactivation rate constant almost 20 times that of vigaba-
1
2
3
4
5
6
positioned leaving group that could facilitate a ring-
opening mechanism in the inactivation of GABA-AT
(Scheme 2). Here we report the synthesis, biological eval-
uation, mechanistic studies, including mass spectral and
X-ray crystallographic results, of these analogs, which
reveal an unexpected inactivation mechanism.
trin), and 18 is half as efficient as vigabatrin.
Table 1. Kinetic constants for the inhibition and inactiva-
tion of GABA-AT by 17 – 19
7
8
9
Scheme 2. Michael addition (pathway a) and enamine
addition (pathway b) mechanisms for 17
Lys 329
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
Lys 329
NH2
S
HN
H
S
COO
H3N
COO
NH
Pyr
Lys 329
Pyr
17
1
The X-ray crystal structure GABA-AT inactivated by 17 (at
1.66 Å) showed that the inactivating metabolite contained
a buckled 5-membered ring covalently bound to PLP
(Figure 4A), and Lys-329 was not covalently modified. The
final ligand interpretation is strongly supported by the
electron density of the simulated annealing omit map (fo-
fc). Furthermore, the omit map density at a higher con-
tour level also revealed that the refined atom positions of
the dihydrothiophene ring were accurate, resulting in a
nonplanar ring (Figures 4B, 4C). (S)-4-Amino-4,5-
dihydro-2-thiophenecarboxylic acid, the corresponding
dihydrothiophene analog of 17, is a known inactivator of
GABA-AT that inactivates by an aromatization mecha-
nism resulting in a thiophene ring;28 it also is an inactiva-
tor of aspartate aminotransferase.30 Here, the crystal
structure of 17-inactivated GABA-AT suggests that the
inactivation by 17 is likely to follow the mechanism shown
in Scheme 4. The resulting Schiff base (32) from the reac-
tion of 17 and the lysine-bound PLP on GABA-AT under-
goes γ-proton removal, leading to enamine 34. The crystal
structure revealed that 34 was stabilized in the active site
by an interaction between its carboxylate group and the
guanidinum group of Arg-192, by the interaction between
the enamine alkene and the phenyl ring of Phe-189, and
the sulfur atom in 34 is in close proximity (3.3 Å) to a car-
boxyl oxygen atom of Glu-270 (Figure 5A). This distance
suggests a weak nonbonded interaction between the diva-
lent sulfur and the carboxyl carbonyl oxygen. Weak non-
bonded S···O and S···N interactions have been reported
and are mainly characterized as stabilizing forces of pro-
tein structures31 and of some organic sulfur
compounds.32,33 These interactions have recently attracted
growing attention because they could play important
roles in the structure and the biological activity of some
sulfur compounds and might also regulate enzymatic
functions.34 Until now, all reported weak nonbonded in-
teractions have been intramolecular. To the best of our
knowledge, no intermolecular weak nonbonded S···O and
S···N interactions have been reported. However, the dis-
tance and directionality of intermolecular nonbonded
S···O interactions has been suggested in theoretical stud-
ies.34 For intermolecular nonbonded S···O interactions,
the nucleophilic O atom approaches the S atom from the
backside of the S - Y and S - Z bonds (the σs* direction),
and the S atom lies in the direction of the O lone pairs
(the nO direction) (Figure 5B). The stabilization of this
Pyr = pyridine ring of PLP
Lys 329
HN
a
:
H2N
HS
HS
COO
a
COO
HN
HN
Pyr
H
Pyr
b
B
Lys 329
H2N
Lys 329
HN
COO
SH
HS
NH2
COO
..
H2N
Pyr
Pyr
2. Results and Discussion
The syntheses of analogs 17 – 19 are shown in Scheme 3,
starting from commercially available D-cysteine methyl
ester hydrochloride (23). The route up to the generation
of dihydrothiophene 28 was achieved by following a mod-
ified procedure from Adam et al.27 Reduction of 28 by
magnesium in methanol resulted in diastereomers 29 and
30, which were separable by flash column chromatog-
raphy. Deprotection of the amino group and hydrolysis of
the ester in 29 and 30 using aqueous HCl provided the
desired analogs 17 and 18, respectively. Oxidation of 29 or
30 by MnSO4 and H2O2 resulted in a 1:1 mixture of the
corresponding sulfones (31) as a result of epimerization at
the C-2 position. Subsequent deprotection of the amino
group and hydrolysis of the ester in 31 using aqueous HCl
gave desired analog 19. Synthesis of compounds 20 – 22
followed an identical route starting from L-cysteine me-
thyl ester hydrochloride. The purity of compounds 17 – 22
was confirmed by HPLC and HRMS, which showed that
there was none of the corresponding dihydrothiophene
analog of 17, a known inactivator of GABA-AT.28
Preliminary in vitro results showed that 19 – 22 were weak
reversible inhibitors, while 17 and 18 were potent inactiva-
tors of GABA-AT (Table 1). The kinetic constants for inac-
tivation of GABA-AT by 17 and 18 could not be deter-
mined accurately under optimal conditions (pH 8.5, 25
oC),23 where the enzyme exhibited maximum activity, be-
cause inactivation occurred too rapidly. The inhibition
constant (KI) and the rate constant of enzyme inactiva-
tion (kinact) for inactivation of GABA-AT by 17 and 18 were
then measured under non-optimal conditions (pH 6.5, 25
oC) using a Kitz and Wilson replot.29 From kinact/KI values
(Table 1), we can conclude that 17 is eight times more effi-
cient an inactivator of GABA-AT than vigabatrin (with an
ACS Paragon Plus Environment

Page 5 of 12
Journal of the American Chemical Society
S···O=C interaction is described by an nO→ σs* orbital
ed S···O interactions. Therefore, the interaction between
1
2
3
4
interaction, and 3.3 Å is well within the predicted dis-
tance. As shown in Figure 5A, the directionality of the
interaction between the sulfur atom in metabolite 34 and
the carboxyl group of Glu-270 matches the description of
the directionality of theoretical intermolecular nonbond-
the sulfur atom in metabolite 34 and the carboxyl group
of Glu-270 might be the first reported example of an in-
termolecular nonbonded S···O interaction, which con-
tributes to the stabilization of metabolite 34 in the active
site of GABA-AT.
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. The PLP-dihydrothiophene complex, shown in ball-and-stick form, is the final adduct, which contains a non-
planar dihydrothiophene ring. Carbons are in beige, nitrogens are in blue, oxygens are in red, phosphorus is in orange,
and sulfur is in yellow. (A) The electron density of the simulated annealing omit map (fo-fc) is shown as a grey mesh at 3.5
σ around the PLP-dihydrothiophone adduct. (B) and (C) are two different orientations of the same image: the electron
density of the simulated omit map (fo-fc) is shown in green mesh at 4.1 σ around atoms in the dihydrothiophene ring (at a
0.6 Å radius). (B) This orientation displays the refined atom positions; (C) this orientation displays a buckled ring plane.
tionality of the intermolecular weak nonbonded S···O
interaction in theoretical studies,34 representing an nO→
σs* orbital interaction
Scheme 4. Proposed mechanism for the inactivation of
GABA-AT by 17
A
B
Figure 5. (A) Interactions between the PLP-
dihydrothiophene adduct and nearby residues. (B) Direc-
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 12
investigated the activity of 39 from (1S,4R)-2-
1
2
3
4
5
6
7
8
azabicyclo[2.2.1]hept-5-en-3-one (37, Scheme 6). Results
showed that 39 is not an inactivator but is a good compet-
itive inhibitor of GABA-AT with a Ki of 0.87 mM. A com-
puter model of the energy-minimized hypothetical ad-
duct of 39 bound to PLP after tautomerization and depro-
tonation (i.e., the cyclopentene analog of 34) docked into
GABA-AT using GOLD35 gave the pose with the highest
fitness score that was almost identical to that shown in
Figure 5 (Supporting Information Figure S25). These inhi-
bition and modeling results further support the crucial
role of the sulfur atom in retaining the product in the
active site of GABA-AT, thereby inactivating the enzyme.
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 6. Radioactive-labeling experiment for the inacti-
vation of GABA-AT by 17: [7-3H]PLP-GABA-AT was pre-
pared from apoGABA-AT and [7-3H]PLP then inactivated
by 17, followed by denaturation and submission to HPLC.
Fractions were collected each minute and counted for
radioactivity. A solution of 1 mM PMP and 1 mM PLP was
treated identically as a control.
3. Conclusion
We have developed two new GABA-AT inactivators. Pre-
liminary in vitro results show that 17 is eight times more
efficient an inactivator of GABA-AT than vigabatrin, an
FDA-approved antiepilepsy drug, and 18 is half as efficient
as vigabatrin. Mechanistic studies of the inactivation of
GABA-AT by 17 showed that the sulfur atom in 17 plays a
crucial role in keeping the resulting adduct bound to the
active site of GABA-AT, thereby inactivating the enzyme.
An intermolecular nonbonded interaction between the
carboxyl oxygen of Glu-270 and the sulfur atom in 17, the
first observed example of this kind, is important for stabi-
lizing the adduct in the active site.
Mass spectrometric analysis (via ESI-mass spectrometry)
was run on a sample of GABA-AT, inactivated by 17, but
the presence of 34 was not observed. When a small
amount of formic acid was added to another sample of
GABA-AT inactivated by 17 to disrupt H-bonding before
running the spectrum, metabolite 36 (m/z 144.9954, Sup-
porting Information Figure S1) was detected instead of 34.
Fragmentation data for m/z 144.9954 confirmed the struc-
ture of 36 (Supporting Information Figure S2), the likely
result of hydrolysis of 34 (Scheme 5).
4. Experimental Section
General Procedures
Scheme 5. Hydrolysis of Metabolite 34
S
S
COOH
COOH
S
N
Chemicals were obtained from TCI America, Sigma-
Aldrich, Alfa Aesar, and American Radiolabeled Chemi-
cals, and used as received unless specified. All syntheses
were conducted under anhydrous conditions in an at-
mosphere of argon, using flame-dried apparatus and em-
ploying standard techniques in handling air-sensitive ma-
terials, unless otherwise noted. All solvents were distilled
and stored under an argon or nitrogen atmosphere before
use. 1H NMR and 13C NMR spectra were taken on a Bruker
AVANCE III 500 spectrometer using CDCl3, MeOD,
(CD3)2CO, or D2O as solvents, recorded in δ (ppm) and
referenced to CDCl3 (7.26 ppm for 1H NMR and 77.16 ppm
HN
O
P
COOH
O
P
O
OH
O
OH
O
36
O
O
HO
H2O
tautomerization
HO
+
N
N
PMP
35
34
Treatment of [7-3H]PLP-reconstituted GABA-AT with 17
was performed to determine the fate of the coenzyme
upon inactivation. A solution of 1 mM PMP and 1 mM PLP
was treated identically as controls. The results showed
that the denaturation of GABA-AT, inactivated by 17, re-
leased PMP exclusively (Figure 6).
Results from the radioactive-labeling experiment and
mass spectrometric analysis suggested that metabolite 34
was not stable outside of the active site and would under-
go hydrolysis to produce PMP and 36, supporting the
proposed mechanism for the inactivation of GABA-AT by
17 shown in Scheme 4.
1
for 13C NMR) or MeOD (3.31 ppm for H NMR and 49.00
ppm for 13C NMR) or (CD3)2CO (2.05 ppm for 1H NMR and
1
29.84 ppm for 13C NMR) or D2O (4.79 ppm for H NMR).
Nuclear Overhauser Effect (NOE) correlation experiments
were performed using an Agilent DDR2 400 MHz spec-
trometer. High resolution mass spectra (HRMS) were
measured with an Agilent 6210 LC-TOF (ESI, APCI, APPI)
mass spectrometer. The purity of the synthesized final
compounds was determined by HPLC analysis to be
>95%. The column used was a Chiralcel OD-H 5 μm, 4.6 x
250 mm. After thorough column equilibration, com-
pounds were eluted with a mobile phase of 2% EtOH in
hexanes at 0.6 mL/min. Biochemical assays were per-
formed using a Biotek Synergy H1 microplate reader. Prior
to their evaluation, initial experiments were performed to
confirm the synthesized analogues do not inhibit the
Scheme 6. Synthesis of Cyclopentane Analog 39
If the interaction between the sulfur atom in 34 and the
O=C of Glu-270 is an intermolecular nonbonded S···O
interaction, then the corresponding cyclopentane analog
(39) (Scheme 6) should form a less stable metabolite in
the active site of GABA-AT than 34. We have made and
ACS Paragon Plus Environment

Page 7 of 12
Journal of the American Chemical Society
(10 mL), dried (MgSO4), and concentrated. The crude
coupling enzymes utilized in the substrate and inhibition
1
2
3
4
5
6
7
8
assays. Mass spectrometric analysis: LC gradient was em-
ployed at a flow rate of 200 μL/min on an Agilent 1150 LC
system (Agilent, Santa Clara, CA, USA); mass spectrome-
try was performed on a Q-Exactive mass spectrometer
(Thermo Fisher Scientific, Waltham, MA, USA). Crystal-
lographic data were collected on beamlines 23ID-B and
23ID-D of GM/CA@APS of the Advanced Photon Source
(APS) using X-rays of 0.99 Å wavelength and Rayonix
(formerly MAR-USA) 4×4 tiled CCD detector with a 300
mm2 sensitive area.
product was purified by chromatography (ethyl ace-
tate/hexanes, 3:17) to yield the major diastereomer (26) as
1
a white solid (2.91 g, 36%). H NMR matched literature
value.27 1H NMR (500 MHz, CDCl3) δ 6.76 (d, J = 9.5 Hz,
1H), 4.22 (ddd, J = 9.6, 5.1, 1.8 Hz, 1H), 4.06 (s, 1H), 3.70 (s,
3H), 3.31 (s, 3H), 3.23 (dd, J = 10.9, 5.1 Hz, 1H), 2.80 (dd, J =
10.9, 1.9 Hz, 1H), 1.40 (s, 9H), 0.11 (s, 9H). 13C NMR (126
MHz, CDCl3) δ 173.09, 155.37, 110.31, 79.39, 59.44, 52.83,
50.78, 49.65, 36.83, 28.50, 0.89. HRMS (LC-TOF): Calcu-
lated for C15H29NO6SSi [M+Na]+ 402.1377; found 402.1388.
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
(S)-Methyl
methoxy-2-oxoethyl)thio)propanoate (24)
2-((tert-butoxycarbonyl)amino)-3-((2-
(S)-Methyl
dihydrothiophene-2-carboxylate (28)
4-((tert-butoxycarbonyl)amino)-4,5-
To a stirred light suspension of D-cysteine methyl ester
hydrochloride (23, 5 g, 29.1 mmol) and Boc2O (7 mL, 30.6
mmol) in anhydrous CH2Cl2 (250 mL) at 0 C was added
To a stirred solution of 26 (979 mg, 2.58 mmol) in 1 M HF
solution (13 mL, prepared by diluting 48% aqueous HF in
dry methanol) at rt was added TBAF (2.84 mL, 2.84 mmol;
1 M solution in THF). The reaction was stirred at rt for 2 h
before being cooled in an ice/brine bath. Once cooled,
NaBH4 (199 mg, 5.16 mmol) was added in small portions
o
Et3N (15.4 mL, 111 mmol) over a 10 min period. After addi-
tion, the cooling bath was removed, and the reaction so-
lution was stirred at rt overnight. After being cooled to 0
oC, methyl bromoacetate (3.3 mL, 35 mmol) was added to
the reaction solution and was stirred for 30 min before
removal of the cooling bath. Stirring was continued for 2h
at rt, followed by removal of the bulk of the solvent under
reduced pressure. The resulting crude mixture was dilut-
ed with ether (60 mL), washed with water (3 x 30 mL) and
brine (5 mL), dried (MgSO4), and concentrated. Chroma-
tography (ethyl acetate/hexanes, 3:7) afforded the desired
o
while maintaining a reaction temp of 0 C. Following ad-
dition, the reaction was stirred for 1 h at 0 oC before being
quenched with acetone (1.3 mL) and allowed to continue
stirring at rt. After 1 h, acetic acid (161 μL) was added, fol-
lowed by removal most of the solvent under reduced
pressure. The resulting crude mixture was diluted with
ethyl acetate (30 mL) and washed with 1:1 saturated
brine:water (15 mL), water (2 x 15 mL), and brine (3 mL),
dried (MgSO4), and concentrated to yield the crude alco-
hol, which was used in the next step without purification.
1
product as a clear oil (6.79 g, 74%). H NMR matched lit-
erature value.27 1H NMR (500 MHz, CDCl3) δ 5.40 (d, J =
8.2 Hz, 1H), 4.55 (m, 1H), 3.75 (s, 3H), 3.73 (s, 3H), 3.26 (q,
J = 15.2 Hz, 2H), 3.07 (m, 2H), 1.43 (s, 9H). 13C NMR (126
MHz, CDCl3) δ 171.50, 170.61, 155.27, 80.35, 53.16, 52.78,
52.66, 35.02, 33.90, 28.38. HRMS (LC-TOF): Calculated for
C12H21NO6S [M+Na]+ 330.0982; found 330.1001.
To a stirred solution of the crude alcohol in CH2Cl2 (13
o
mL) at 0 C was added Et3N (1.44 mL, 10.3 mmol), fol-
lowed by mesyl chloride (400 μL, 5.17 mmol) dropwise.
The reaction was stirred for 30 min at 0 C and overnight
o
at rt. After removal of most of the solvent under reduced
pressure, the resulting crude mixture was diluted with
ethyl ether (30 mL), washed with water (2x 15 mL), 0.5 M
HCl (15 mL), and brine (3 mL), dried (MgSO4), and con-
centrated. Chromatography (ethyl acetate/hexanes, 3:7)
afforded the desired product as a white solid (430 mg,
64%). 1H NMR (500 MHz, CDCl3) δ 6.51 (d, J = 3.3 Hz, 1H),
5.15 (m, 1H), 4.90 (d, J = 9.2 Hz, 1H), 3.79 (s, 3H), 3.61 (dd,
J = 12.3, 8.3 Hz, 1H), 3.13 (dd, J = 12.3, 4.2 Hz, 1H), 1.43 (s,
9H). 13C NMR (126 MHz, CDCl3) δ 162.68, 154.74, 138.37,
131.93, 80.35, 58.29, 52.76, 39.40, 28.44. HRMS (LC-TOF):
Calculated for C11H17NO4S [M+Na]+ 282.0770; found
282.0773.
(3R,4S)-Methyl
methoxy-3-
lyl)oxy)tetrahydrothiophene-2-carboxylate (26)
4-((tert-butoxycarbonyl)amino)-3-
((trimethylsi-
To a solution of 24 (6.59 g, 21.4 mmol) in anhydrous di-
chloromethane (100 mL) at 0 °C was added Et3N (3.3 mL,
23.6 mmol) followed by dropwise addition of TMSOTf
(4.25 mL, 23.5 mmol) over 20 min. The mixture was
stirred for 10 min at 0 °C then allowed to warm to rt. After
being quenched with saturated sodium bicarbonate (50
mL), the organic layer was separated and washed with
saturated NaHCO3 (2 x 50 mL), dried (MgSO4) and con-
centrated to obtain crude intermediate 25.
(2S,4S)-
and
(2R,4S)-Methyl
4-((tert-
In a separate flask, a solution of lithium tetramethylpiper-
idide was prepared by the dropwise addition of n-BuLi
(14.0 mL, 22.4 mmol; 1.6 M solution in hexanes) to a solu-
tion of 2,2,6,6-tetramethylpiperidide (4.17 mL, 24.5 mmol)
in THF (100 mL) at -78 oC. After a brief warm-up to rt, the
solution was cooled to -78 C, and crude intermediate 25
in THF (50 mL) was added dropwise over 30 min. After
the addition, the reaction was stirred at -78 C for 30 min
butoxycarbonyl)amino)tetrahydrothiophene-2-
carboxylate (29 and 30):
Magnesium turnings (484 mg, 19.9 mmol) were added to
a mixture of (S)-methyl 4-((tert-butoxycarbonyl)amino)-
4,5-dihydrothiophene-2-carboxylate (28, 430 mg, 1.66
mmol) and NH4Cl (5.33 g, 99.6 mmol) in MeOH (15 mL),
and the resulting mixture was vigorously stirred overnight
at rt. After removal of most of the solvent under reduced
pressure, the resulting crude mixture was diluted with
water (20 mL) and extracted with ethyl ether (3 x 40 mL).
The combined organics were washed with brine (2 mL),
o
o
and at -40 oC for another 30 min. The reaction was cooled
again to -78 oC and quenched with acetic acid (3 mL). The
reaction mixture was diluted in ether (100 mL), washed
with water (3 x 100 mL), 0.5 N HCl (3 x 30 mL), and brine
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 12
dried (MgSO4), and concentrated. Chromatography (ethyl (dd, J = 9.1, 6.1 Hz), 1H], [3.88 (s), 3.85 (s), 3H], [3.47 (dd, J
1
2
3
4
5
6
7
8
ether/toluene, 2:8) afforded 29 (109 mg, 25%) and 30 (141
mg, 32%) as white solids. (29): 1H NMR (500 MHz, CDCl3)
δ 5.86 (d, J = 7.6 Hz, 1H), 4.55-4.40 (m, 1H), 3.97 (dd, J =
8.6, 3.1 Hz, 1H), 3.72 (s, 3H), 3.11 (dd, J = 11.0, 5.0 Hz, 1H),
2.92 (dd, J = 11.3, 3.6 Hz, 1H), 2.37-2.26 (m, 1H), 2.17 (ddd, J
= 14.0, 8.5, 6.0 Hz, 1H), 1.41 (s, 9H); 13C NMR (126 MHz,
CDCl3) δ 175.10, 155.31, 79.48, 55.54, 52.95, 45.68, 40.16,
37.28, 28.48; HRMS (LC-TOF): Calculated for C11H19NO4S
= 13.5, 6.9 Hz), 3.42 (dd, J = 13.7, 7.0 Hz), 1H], [3.18 (t, J =
4.9 Hz), 3.15 (t, J = 5.4 Hz), 1H], [2.83 (dt, J = 13.9, 6.8 Hz),
2.69 (ddd, J = 14.1, 8.9, 6.9 Hz), 1H], 2.50 (m, 1H), 1.44 (s,
9H). 13C NMR (126 MHz, CDCl3) δ [166.49, 165.77], [155.00,
154.78], [80.76, 80.59], [65.14, 64.02], [57.07, 56.05], [53.98,
53.78], [45.54, 45.37], [33.30, 32.23], [28.43, 28.41]. HRMS
(LC-TOF): Calculated for C11H19NO6S [M+Na]+ 316.0825;
found 316.0833.
1
[M+Na]+ 284.0927, found 284.0931. (30): H NMR (500
9
(4S)-4-Aminotetrahydrothiophene-2-carboxylic acid
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
MHz, (CD3)2CO) δ 6.25 (d, J = 3.4 Hz, 1H), 4.49 (tt, J = 7.4,
3.8 Hz, 1H), 4.06 (dd, J = 7.5, 5.1 Hz, 1H), 3.69 (s, 3H), 3.14
(dd, J = 10.5, 5.6 Hz, 1H), 2.78 (dd, J = 10.5, 6.2 Hz, 1H),
2.42 (dt, J = 13.0, 5.3 Hz, 1H), 2.15-2.05 (m, 1H), 1.42 (s, 9H);
13C NMR (126 MHz, (CD3)2CO) δ 174.00, 155.92, 78.96,
55.98, 52.52, 44.49, 37.69, 37.53, 28.53; HRMS (LC-TOF):
Calculated for C11H19NO4S [M+Na]+ 284.0927, found
284.0931.
1,1-dioxide hydrochloride (19):
Compound 19 was synthesized from 31 as an inseparable
1:1 mixture of diastereomers using a procedure similar to
1
that for 17 from 29 (86%). H NMR (500 MHz, MeOD) δ
[4.41 (dd, J = 8.6, 4.9 Hz), 4.36 (dd, J = 9.5, 7.9 Hz), 1H],
[4.27 (p, J = 7.4 Hz), 4.12 (p, J = 8.1 Hz), 1H], [3.73 (dd, J =
13.8, 8.1 Hz), 3.68 (dd, J = 13.8, 8.0 Hz), 1H], [3.35 (dd, J =
13.9, 7.1 Hz), 3.29 (dd, J = 13.5, 8.1 Hz), 1H], [2.96 (ddd, J =
14.2, 7.0, 5.0 Hz), 2.84 (dt, J = 14.2, 7.2 Hz), 1H], [2.55 (dt, J
= 13.9, 9.3 Hz), 2.46 (dt, J = 14.3, 8.1 Hz), 1H]. 13C NMR (126
MHz, MeOD) δ [167.13, 167.04], [66.52, 65.56], [54.61,
54.44], [46.38, 45.40], [31.40, 31.37]. HRMS (LC-TOF): Cal-
culated for C5H9NO4S [M+Na]+ 202.0144; found 202.014.
(2S,4S)-4-Aminotetrahydrothiophene-2-carboxylic
acid hydrochloride (17):
Boc-protected amino acid ester 29 (100 mg, 0.38 mmol)
was dissolved in 4 N HCl (5 mL) and acetic acid (5 mL).
o
The resulting solution was heated to 70 C and stirred for
5 h before being concentrated in vacuo to afford a solid.
The solid was purified by ion-exchange chromatography
(AG 50W-X8), eluting with a gradient from 0.4 N to 2.0 N
HCl, giving the desired amino acid hydrochloride product
as a white solid (63 mg, 89%). 1H NMR (500 MHz, MeOD)
δ 4.10 (m, 2H), 3.32 (dd, J = 11.8, 5.6 Hz, 1H), 3.09 (dd, J =
11.8, 5.0 Hz, 1H), 2.52 - 2.43 (m, 1H). 13C NMR (126 MHz,
MeOD) δ 177.11, 55.99, 46.14, 37.28, 36.77. HRMS (LC-
TOF): Calculated for C5H9NO2S [M-H]- 146.0281; found
146.0278.
(1R,4S)-2-azabicyclo[2.2.1]heptan-3-one (38):
Compound
38
was
prepared
from
(1S,4R)-2-
azabicyclo[2.2.1]hept-5-en-3-one (37) by following a pub-
lished procedure (94%).36 1H NMR matched literature
value.37 1H NMR (500 MHz, CDCl3) δ 5.53 (br s, 1H), 3.90
(m, 1H), 2.76 (m, 1H), 1.95-1.40 (m, 6H). 13C NMR (126
MHz, CDCl3) δ 181.05, 55.54, 45.04, 41.37, 30.34, 23.79.
HRMS (LC-TOF): Calculated for C6H9NO [M+Na]+
134.0576; found 134.0578.
Preparation
of
(1S,3R)-3-aminocyclopentane-1-
(2R,4S)-4-Aminotetrahydrothiophene-2-carboxylic
acid hydrochloride (18):
carboxylic acid (39):
Compound 39 was prepared from 38 by following a pub-
lished procedure (80%).37 1H NMR matched literature
value.37 1H NMR (500 MHz, D2O) δ 3.76 (m, 1H), 3.01 (m,
Compound 18 (61 mg, 87%) was synthesized from 30 (100
mg, 0.38 mmol) using a similar procedure to that for 17
1
from 29. H NMR (500 MHz, MeOD) δ 4.15 (p, J = 5.9 Hz,
13
1H), 2.44-1.79 (m, 6H). C NMR (126 MHz, D2O) δ 180.01,
1H), 4.05 (dd, J = 7.6, 4.5 Hz, 1H), 3.28 (dd, J = 11.6, 5.7 Hz,
1H), 2.95 (dd, J = 11.5, 5.7 Hz, 1H), 2.65 (dt, J = 13.6, 5.0 Hz,
1H), 2.18 (dt, J = 13.9, 7.2 Hz, 1H). 13C NMR (126 MHz,
MeOD) δ 175.62, 55.75, 45.19, 37.22, 35.81. HRMS (LC-
TOF): Calculated for C5H9NO2S [M-H]- 146.0281; found
146.0278.
51.47, 42.40, 33.65, 29.81, 27.68. HRMS (LC-TOF): Calcu-
lated for C6H11NO2 [M+H]+ 130.0863; found 130.0864.
Purification of GABA Aminotransferase (GABA-AT)
from Pig Brain
GABA-AT was isolated and purified from pig brain by a
published procedure.38 The purified GABA-AT used in
these experiments was found to have a concentration of
6.41 mg/mL with a specific activity of 1.84 units/mg.
(4S)-Methyl
4-((tert-
butoxycarbonyl)amino)tetrahydrothiophene-2-
carboxylate 1,1-dioxide (31):
Evaluation of Compounds as Time-Dependent Inhib-
itors of GABA-AT
To a stirred solution of 29 (60 mg, 0.23 mmol) and
MnSO4•H2O (1 mg) in CH3CN (5 mL) was added at room
temperature a mixture of 30% H2O2 (1.15 mmol, 118 μL)
GABA-AT (17.5 µL) was incubated in the presence of vary-
ing concentrations of each compound (70 µL final vol-
o
and 0.2 M NaHCO3 (3.4 mL), previously prepared at 0 C.
o
After 15 min the reaction was quenched with brine, ex-
tracted with ethyl acetate (3 x 10 mL), dried (Na2SO4), and
concentrated. Chromatography (ethyl acetate/hexanes;
1:9) provided the desired product as a 1:1 mixture of the
ume) at 25 C in 50 mM potassium pyrophosphate buffer
solution, pH 6.5, containing 5 mM α-ketoglutarate and 1
mM β-mercaptoethanol. Aliquots (10 µL) were withdrawn
at timed intervals and were added immediately to the
assay solution (137 µL, see below) followed by the addition
of SSDH (3 µL). The relative enzyme activity was deter-
mined by normalizing the rate of increasing absorbance
1
two diastereomers (55 mg, 81%). H NMR (500 MHz,
CDCl3) δ [5.57 (d, J = 7.1 Hz); 5.21 (d, J = 6.2 Hz), 1H], [4.65
(br s), 4.54 (sex, J = 6.6 Hz), 1H], [4.14 (t, J = 7.9 Hz), 4.11
ACS Paragon Plus Environment

Page 9 of 12
Journal of the American Chemical Society
column (4.6 mm × 250 mm, 5 µ, 110 A). Water (with 0.1%
at 340 nm to a control. A Kitz and Wilson replot was used
to determine the kinetic constants KI and kinact
29
1
.
TFA) and acetonitrile (with 0.1% TFA) were used as the
mobile phase (5% acetonitrile, 0.5 mL/min, 25 min; then
gradient to 95% acetonitrile, 0.5 mL/min, 20 min). Under
these conditions, PLP eluted at 38 min. Fractions running
with the PLP peak were collected, counted for radioactivi-
ty using liquid scintillation counting, and then lyophi-
lized, affording [7-3H]PLP.
2
3
4
5
6
7
8
9
Evaluation of Compounds as Inhibitors of GABA-AT
Inhibition constants were determined by monitoring
GABA-AT activity in the presence of 0-50 mM concentra-
tions of synthesized analogues using a coupled assay with
the enzyme succinic semialdehyde dehydrogenase
(SSDH). The assay solution consisted of 10 mM GABA, 5
Preparation of [7-3H]PLP-Reconstituted GABA-AT
mM α-ketoglutarate,
1
mM NADP+,
5
mM β-
mercaptoethanol, and excess SSDH in 50 mM potassium
pyrophosphate buffer, pH 8.5. Enzyme activity was de-
termined by observing the change in absorbance at 340
nm at 25 °C. IC50 values were obtained using non-linear
regression in GraphPad Prism5 software. Subsequent Ki
values were determined using the Cheng-Prusoff relation-
ship.39
To potassium phosphate buffer (0.5 mL, 100 mM, pH 7.4)
containing β-mercaptoethanol (0.25 mM) (buffer A) and
GABA-AT (170 µg, 1.55 nmol) protected from light was
added GABA (10 mg, 0.097 mmol). The resulting solution
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
o
was stirred at rt for 1 h and then was dialyzed at 4 C
against potassium phosphate buffer (200 mL, 500 mM,
pH 5.5) containing β-mercaptoethanol (0.25 mM) (buffer
B) and GABA (4.0 g, 39 mmol) for 3 h. The solution was
Evaluation of Compounds as Substrates for GABA-AT
o
then dialyzed at 4 C against 1800 mL of buffer B over-
Compounds were tested using an experiment in which
the conversion of α-ketoglutarate to L-glutamic acid was
monitored as an indication of the rate of PLP reduction to
PMP, which in turn corresponds to amine oxidation to
the corresponding aldehyde. Enzyme reactions were pre-
pared at 5 mM concentrations of compounds in 100 µL
pyrophosphate buffer (50 mM, pH 8.5) containing 5 mM
α-ketoglutarate and 0.13 mg/mL purified GABA-AT and
allowed to incubate at room temperature for 24 h. The L-
glutamic acid content was determined by combining 50
µL of each incubation mixture with 50 µL of Tris-HCl
buffer (100 mM, pH 7.5) containing 100 µM Ampliflu™
Red (Sigma-Aldrich), 0.25 units/mL horseradish peroxi-
dase and 0.08 units/mL L-glutamate oxidase in a 96-well
black walled plate. After incubation at 37 °C for 30 min
fluorescence was recorded with the aid of a microplate
reader (BioTek Synergy H1) with 530 nm excitation and
590 nm emission wavelengths, where fluorescence is pro-
portional to the L-glutamate concentration.
o
night, followed by dialysis at 4 C against 4 x 500 mL of
potassium phosphate buffer (100 mM, pH 8.0) containing
β-mercaptoethanol (0.25 mM) (buffer C) at 1.5 h interval.
An aliquot (1 µL) of the dialyzed solution was assayed to
confirm that there was no enzyme activity remaining. To
the dialyzed solution of enzyme (apo-GABA-AT) was
added a solution of PLP (40 µL, 20 mM) and [3H]PLP (1/5
the amount prepared above) and stirred at rt for 5 h until
the enzyme activity returned and the reactivation was
complete. The excess [3H]PLP was removed from the re-
constituted GABA-AT solution by centrifugation with 5 x
400 µL of buffer C using a 10K molecular weight cutoff
o
filter. The resulting enzyme solution was dialyzed at 4 C
against 2 L of buffer A overnight, affording the [3H]PLP-
reconstituted GABA-AT solution.
Inactivation of [7-3H]PLP-Reconstituted GABA-AT by
17 and Cofactor Analysis
A 60 µL portion of buffer A containing [7-3H]PLP-
reconstituted GABA-AT (1/5 the amount prepared above),
α-ketoglutarate (3 mM), and 17 (4 mM) was protected
from light and incubated at rt overnight until the activity
of GABA-AT was less than 1%. To the inactivated enzyme
solution was added KOH (1 drop, 1 M) to adjust the pH to
11. The mixture was allowed to stand at rt for 1 h, then
trifluoroacetic acid (TFA) (6.7 µL) was added, and it was
allowed to stand for another 5 min. The resulting dena-
tured enzyme solution was centrifuged at 13,400 rpm for 5
min. The supernatant was collected, and the pellet was
rinsed with 2 x 50 µL of buffer A containing 10% TFA and
centrifuged. The supernatant and rinses were combined
and lyophilized. The resulting solid was dissolved in 100
µL of a solution containing 1 mM PLP and 1 mM PMP
standards and injected into an HPLC with an Econosil C18
column (4.6 mm × 150 mm, 10 µ). Water (with 0.1% TFA)
and acetonitrile (with 0.1% TFA) were used as the mobile
phase (0% acetonitrile, 0.5 mL/min, 25 min; then gradient
to 90% acetonitrile, 1.0 mL/min, 30 min). Under these
conditions, PMP eluted at 8 min and PLP at 13 min. Frac-
tions were collected every minute, and the radioactivity
was measured by liquid scintillation counting.
Preparation of [7-3H]-Pyridoxal 5'-Phosphate ([7-
3H]PLP)
To a solution of pyridoxal 5'-phosphate (PLP) in water (1.8
mL, 0.28 M) was added thirty drops of 1 M NaOH. The
mixture was then cooled to 0 °C in an ice bath, and a solu-
tion of NaBH4 (5.86 mg, 0.15 mmol) and NaB[3H]4 (100
mCi) in 450 µL of 0.1 M NaOH was added in small por-
tions and stirred for 1 h at 0 °C. Concentrated HCl (120 µL)
was then added to the solution very slowly (the pH of the
resulting solution was 4), and the reaction was stirred for
5 min at 0 oC. Ground MnO2 (200 mg, 2.3 mmol) was add-
ed, and the resulting mixture was stirred for 2 h at rt. A
solution of 1 M NaOH was added dropwise to bring the
pH to 8, and the resulting solution was centrifuged. The
supernatants were collected and loaded onto a gel filtra-
tion column packed with Bio-Rad AG1-X8 resin (hydrox-
ide form). Water and 5 M acetic acid was used as the mo-
bile phase (gradient from 90% water to 0% water, 1.5
mL/min, 300 min). Fractions (10 mL each) were collected
and tested for UV absorption and radioactivity. The frac-
tions with the desired product were lyophilized and then
loaded onto an HPLC with a Phenomenex Gemini C18
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 12
Mass spectrometric analysis of the Inactivated GABA- show a well-defined electron density for both PLP and the
1
2
3
4
5
6
7
8
AT
inactivator tetrahydrothiophene (Figure 4A). The struc-
ture of inactivator tetrahydrothiophene was built in
Chemdraw (a “Mol” file); the molecule was regularized
and then the chemical restrains were generated in pro-
gram JLigand.43 The PLP and the inactivator were fitted
into the residual electron density in COOT after the rest
of the structure, including most of the solvent molecules,
had been refined. The Rfree and Rfactor for inactivated
GABA-AT were 21.5% and 19.8%, respectively (Supporting
Information Table S1). All structural figures were made in
UCSF Chimera.44
A 50 µL portion of ammonium bicarbonate buffer (50
mM, pH 7.4) containing GABA-AT (23 µg, 0.21 nmol), α-
ketoglutarate (5 mM), and 17 (43 mM) was protected from
light and incubated at rt overnight until the activity of
GABA-AT was less than 1%. Another 50 µL of ammonium
bicarbonate buffer (50 mM, pH 7.4) containing GABA-AT
(23 µg, 0.21 nmol) and α-ketoglutarate (5 mM) was sub-
jected to the same condition as a control. After incuba-
tion, formic acid (10 µL) was added to each sample, and
20 µL of each resulting solution was loaded onto a 5-μm
Luna C18 column (2 mm i.d.; 150 mm) (Phenomenex, Tor-
rance, CA, USA). A 30-min LC gradient was employed at a
flow rate of 200 μl/min on an Agilent 1150 LC system (Ag-
ilent, Santa Clara, CA, USA). Mass spectrometry was per-
formed on a Q-Exactive mass spectrometer (Thermo
Fisher Scientific, Waltham, MA, USA). Intact MS spectra
were acquired at a resolution of 35,000. The top-five most
intense ions were selected for fragmentation in a data-
dependent acquisition mode. Mass spectra were acquired
at a resolution of 17,500.
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
ASSOCIATED CONTENT
Supporting Information
NMR spectroscopy, mass spectroscopy, and crystallograph-
ic data collection and processing statistics. This material is
available free of charge via the Internet at
http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
* Agman@chem.northwestern.edu
Crystallization and Data Collection
Potassium pyrophosphate buffer (500 µL, 50 mM, pH 8.5)
containing GABA-AT (200 µg, 1.82 nmol), α-ketoglutarate
(5 mM), β-mercaptoethanol (5 mM), and 17 (37 mM) was
protected from light and incubated at rt overnight until
the activity of GABA-AT was less than 3%. The inactivated
GABA-AT was then buffer-exchanged into a sodium ace-
tate buffer (40 mM, pH 5.5) by centrifugation before the
initial crystallization screening and optimization. The
crystals were obtained in hanging drops comprising 1 µL
of 10 mg/mL inactivated GABA-AT and 1 μL reservoir so-
lution, containing 0.1 M ammonium acetate, 0.1 M Bis-
Tris (pH 5.5), and 17% w/v PEG 10,000. Diffraction quality
crystals grew within 4-5 days at ambient temperature.
For X-ray data collection, these crystals were briefly
soaked in the reservoir solution with additional 20% (v/v)
glycerol as cryo-protectant before flash freezing in liquid
nitrogen. Crystallographic data were collected on beam-
lines 23ID-B and 23ID-D of GM/CA@APS of the Advanced
Photon Source (APS) using X-rays of 0.99 Å wavelength
and Rayonix (formerly MAR-USA) 4×4 tiled CCD detector
with a 300 mm2 sensitive area. All data were indexed,
integrated, and scaled with HKL2000. Data collection
and processing statistics are given in Table S1.
ACKNOWLEDGMENT
We are grateful to the National Institutes of Health for finan-
cial support (Grants GM066132 and DA030604 to RBS;
GM067725 to NLK). GM/CA@APS has been funded in whole
or in part with Federal funds from the National Cancer Insti-
tute (ACB-12002) and the National Institute of General Medi-
cal Sciences (AGM-12006). This research used resources of
the Advanced Photon Source, a U.S. Department of Energy
(DOE) Office of Science User Facility operated for the DOE
Office of Science by Argonne National Laboratory under
Contract No. DE-AC02-06CH11357. The authors wish to
thank Drs. Jose Juncosa and Hyunbeom Lee for helpful dis-
cussions and Dr. Boobalan Pachaiyappan for constructing the
computer model of 39 docked into the enzyme using GOLD
(Supporting Information Figure S25). The authors would also
like to thank Park Packing Co. (Chicago, IL) for their gener-
osity in providing fresh pig brains for this study. Support for
the spectrometer funding has been provided by the Interna-
tional Institute of Nanotechnology.
REFERENCES
(1) Chapter 1: Basic Mechanisms Underlying Seizures and Epilep-
sy. In An Introduction to Epilepsy [Internet]; Bromfield, E. B.;
Cavazos, J. E.; Sirven, J. I., Eds.; American Epilepsy Society: West
Hartford, CT, 2006.
Phasing, Model Building, and Refinement
(2) Epilepsy Foundation. About Epilepsy: The Basics.
http://www.epilepsy.com/learn/about-epilepsy-basics (accessed
Jul 29, 2014).
(3) Karlsson, A.; Fonnum, F.; Malthe-Sørenssen, D.; Storm-
Mathisen, J. Biochem. Pharmacol. 1974, 23, 3053–3061.
(4) Baxter, C. F.; Roberts, E. J. Biol. Chem. 1958, 233, 1135–1139.
(5) Durkin, M. M.; Smith, K. E.; Borden, L. A.; Weinshank, R. L.;
Branchek, T. A.; Gustafson, E. L. Mol. Brain. Res. 1995, 33, 7–21.
(6) Yogeeswari, P.; Sriram, D.; Vaigundaragavendran, J. Curr.
Drug Metab. 2005, 6, 127–139.
Molecular replacement for the inactivated GABA-AT was
carried out using the program Phaser from CCP4 software
suite.40 The tetrameric structure of native GABA-AT from
pig brain was used as the starting search model, in which
all solvent and PLP molecules were deleted. Initial Rfree
and R factor of the correct solution were 29.59% and
28.91%, respectively. The rigid body refinement was fol-
lowed by restrained refinement with Refmac541 and fur-
ther manual model inspection and adjustments with
Coot.42 When refinement converged, the Fo-Fc difference
maps, before incorporation of ligands in the structures,
(7) Nishino, N.; Fujiwara, H.; Noguchi-Kuno, S. A.; Tanaka, C.
Jpn. J. Pharmacol. 1988, 48, 331–339.
ACS Paragon Plus Environment

Page 11 of 12
Journal of the American Chemical Society
(8) Aoyagi, T.; Wada, T.; Nagai, M.; Kojima, F.; Harada, S.;
Osborne, H.; Craft, C. M.; Brodie, J. D.; Schiffer, W. K.; Dewey, S.
1
2
3
4
5
6
7
8
Takeuchi, T.; Takahashi, H.; Hirokawa, K.; Tsumita, T. Chem.
Pharm. Bull. 1990, 38, 1748–1749.
(9) Iversen, L. L.; Bird, E. D.; Mackay, A. V; Rayner, C. N. J. Psy-
chiatr. Res. 1974, 11, 255–256.
(10) Dewey, S. L.; Morgan, A. E.; Ashby, C. R.; Horan, B.; Kush-
ner, S. A.; Logan, J.; Volkow, N. D.; Fowler, J. S.; Gardner, E. L.;
Brodie, J. D. Synapse 1998, 30, 119–129.
(11) Gale, K. Epilepsia 1989, 30 Suppl 3, S1–11.
(12) Van Gelder, N. M.; Elliott, K. A. C. J. Neurochem. 1958, 3,
139–143.
(13) Singh, J.; Petter, R. C.; Baillie, T. A.; Whitty, A. Nat. Rev.
Drug Discovery 2011, 10, 307–317.
(14) Lippert, B.; Metcalf, B. W.; Jung, M. J.; Casara, P. Eur. J. Bio-
chem. 1977, 74, 441–445.
(15) Waterhouse, E. J.; Mims, K. N.; Gowda, S. N. Neuropsychiatr
Dis Treat. 2009, 5, 505–515.
(16) Tassinari, C. A.; Michelucci, R.; Ambrosetto, G.; Salvi, F.
Arch. Neurol. 1987, 44, 907–910.
(17) Browne, T. R.; Mattson, R. H.; Penry, J. K.; Smith, D. B.;
Treiman, D. M.; Wilder, B. J.; Ben-Menachem, E.; Miketta, R. M.;
Sherry, K. M.; Szabo, G. K. Br. J. Clin. Pharmacol. 1989, 27 Suppl 1,
95S–100S.
(18) Sivenius, M. R.; Ylinen, A.; Murros, K.; Matilainen, R.; Riek-
kinen, P. Epilepsia 1987, 28, 688–692.
(19) Sander, J. W.; Hart, Y. M.; Trimble, M. R.; Shorvon, S. D. J.
Neurol. Neurosurg. Psychiatry. 1991, 54, 435–439.
(20) Wild, J. M.; Chiron, C.; Ahn, H.; Baulac, M.; Bursztyn, J.;
Gandolfo, E.; Goldberg, I.; Goñi, F. J.; Mercier, F.; Nordmann, J.-
P.; Safran, A. B.; Schiefer, U.; Perucca, E. CNS Drugs 2009, 23,
965–982.
(21) Nanavati, S. M.; Silverman, R. B. J. Am. Chem. Soc. 1991, 113,
9341–9349.
(22) Choi, S.; Storici, P.; Schirmer, T.; Silverman, R. B. J. Am.
Chem. Soc. 2002, 124, 1620–1624.
L.; Miller, S. R.; Silverman, R. B. J. Med. Chem. 2012, 55, 357–366.
(26) Silverman, R. B. J. Med. Chem. 2012, 55, 567–575.
(27) Adams, J. L.; Chen, T. M.; Metcalf, B. W. J. Org. Chem. 1985,
50, 2730–2736.
(28) Fu, M.; Nikolic, D.; Van Breemen, R. B.; Silverman, R. B. J.
Am. Chem. Soc. 1999, 121, 7751–7759.
(29) Kitz, R.; Wilson, I. B. J. Biol. Chem. 1962, 237, 3245–3249.
(30) Liu, D.; Pozharski, E.; Lepore, B. W.; Fu, M.; Silverman, R.
B.; Petsko, G. A.; Ringe, D. Biochemistry 2007, 46, 10517–10527.
(31) Iwaoka, M.; Takemoto, S.; Okada, M.; Tomoda, S. Chem.
Lett. 2001, 132–133.
(32) Nagao, Y.; Hirata, T.; Goto, S.; Sano, S.; Kakehi, A.; Iizuka,
K.; Shiro, M. J. Am. Chem. Soc. 1998, 120, 3104–3110.
(33) Ioannidis, S.; Lamb, M. L.; Almeida, L.; Guan, H.; Peng, B.;
Bebernitz, G.; Bell, K.; Alimzhanov, M.; Zinda, M. Bioorg. Med.
Chem. Lett. 2010, 20, 1669–1673.
(34) Iwaoka, M.; Takemoto, S.; Tomoda, S. J. Am. Chem. Soc.
2002, 124, 10613–10620.
(35) Jones, G.; Willett, P.; Glen, R. C. J. Mol. Biol. 1995, 245, 43–53.
(36) Evans, C.; McCague, R.; Roberts, S. M.; Sutherland, A. G. J.
Chem. Soc., Perkin Trans. 1 1991, 656–657.
(37) Forró, E.; Fülöp, F. Eur. J. Org. Chem. 2008, 2008, 5263–5268.
(38) Koo, Y. K.; Nandi, D.; Silverman, R. B. Arch. Biochem. Bio-
phys. 2000, 374, 248–254.
(39) Yung-Chi, C.; Prusoff, W. H. Biochem. Pharmacol. 1973, 22,
3099–3108.
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
(40) Collaborative Computational Project Number 4 Acta Crys-
tallogr. D Biol. Crystallogr. 1994, 50, 760–763.
(41) Emsley, P.; Cowtan, K. Acta Crystallogr. D Biol. Crystallogr.
2004, 60, 2126–2132.
(42) Murshudov, G. N.; Vagin, A. A.; Dodson, E. J. Acta Crystal-
logr. D Biol. Crystallogr. 1997, 53, 240–255.
(43) Lebedev, A. A.; Young, P.; Isupov, M. N.; Moroz, O. V;
Vagin, A. A.; Murshudov, G. N. Acta Crystallogr. D Biol. Crystal-
logr. 2012, 68, 431–440.
(23) Pan, Y.; Qiu, J.; Silverman, R. B. J. Med. Chem. 2003, 46,
5292–5293.
(24) Okumura, H.; Omote, M.; Takeshita, S. Arzneimittel-
forschung. 1996, 46, 459–462.
(25) Pan, Y.; Gerasimov, M. R.; Kvist, T.; Wellendorph, P.; Mad-
sen, K. K.; Pera, E.; Lee, H.; Schousboe, A.; Chebib, M.; Bräuner-
(44) Pettersen, E. F.; Goddard, T. D.; Huang, C. C.; Couch, G. S.;
Greenblatt, D. M.; Meng, E. C.; Ferrin, T. E. J. Comput. Chem.
2004, 25, 1605–1612.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 12
Insert Table of Contents artwork here
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
12
ACS Paragon Plus Environment