(S)-4-Amino-5-phenoxypentanoate designed as a potential selective agonist of the bacterial transcription factor GabR

Article abstract of DOI:10.1002/pro.3905

Addressing molecular recognition in the context of evolution requires pursuing new molecular targets to enable the development of agonists or antagonists with new mechanisms of action. Disruption of transcriptional regulation through targeting transcription factors that regulate the expression of key enzymes in bacterial metabolism may provide a promising method for controlling the bacterial metabolic pathways. To this end, we have selectively targeted a bacterial transcription regulator through the design and synthesis of a series of γ-aminobutyric acid (GABA) derivatives, including (S)-4-amino-5-phenoxypentanoate (4-phenoxymethyl-GABA), which are based on docking insights gained from a previously-solved crystal structure of GabR from Bacillus subtilis. This target was selected because GabR strictly controls GABA metabolism by regulating the transcription of the gabT/D operon. These GabR transcription modulators are selective for the bacterial transcription factor GabR and are unable to bind to structural homologs of GabR due to distinct steric constraints. We have obtained a crystal structure of 4-phenoxymethyl-GABA bound as an external aldimine with PLP in the effector binding site of GabR, which suggests that this compound is capable of binding and reacting in the same manner as the native effector ligand. Inhibition assays demonstrate high selectivity of 4-phenoxymethyl-GABA for bacterial GabR versus several selected eukaryotic enzymes. Single-molecule fluorescence resonance energy transfer (smFRET) experiments reveal a ligand-induced DNA distortion that is very similar to that of the native effector GABA, suggesting that the compound functions as a potential selective agonist of GabR.

Full text of DOI:10.1002/pro.3905

Article
Protein Sci
DOI10.1002/pro.3905
(S)-4-Amino-5-phenoxypentanoate Designed as a Potential Selective Agonist
of the Bacterial Transcription Factor GabR
Running Title: Design a Potential Selective Agonist of GabR
Daniel S. Catlin1§, Cory T. Reidl1§, Thomas R. Trzupek2, Richard B. Silverman2,3, Brian
Cannon4, Daniel P. Becker1*, and Dali Liu1*
1Department of Chemistry and Biochemistry, Loyola University Chicago, 1032 West Sheridan
Road, Chicago IL 60660, USA; 2Department of Chemistry, Chemistry of Life Processes
Institute, Center for Molecular Innovation and Drug Discovery, 3Department of Molecular
Biosciences, Northwestern University, 2145 Sheridan Road, Evanston, IL 60208-3113, USA;
4Department of Physics, Loyola University Chicago, 1032 West Sheridan Road, Chicago IL
60660, USA.
§These authors contributed equally.
*To whom correspondence should be addressed: Dali Liu, Department of Chemistry and
Biochemistry, Loyola University Chicago, IL 60660, dliu@luc.edu, 773.508.3093
or Daniel P. Becker: Loyola University Chicago, Department of Chemistry and Biochemistry,
dbecke3@luc.edu, 773.508.3089
Manuscript Pages: 35
Supplemental Pages: 7
Tables: 1
Figures: 5
Schemes: 2
Supplemental Material: contains material and methods for organic synthesis of a series of
analogs of (S)-4-amino-5-phenoxypentanoate (4-phenoxymethyl-GABA), which is detailed in
main body, as well as figures to complement the discussion in the main text.
This article has been accepted for publication and undergone full peer review but has not been
through the copyediting, typesetting, pagination and proofreading process which may lead to
differences between this version and the Version of Record. Please cite this article as doi:
10.1002/pro.3905
© 2020 The Protein Society
Received: May 07, 2020;Revised: Jun 12, 2020;Accepted: Jun 16, 2020
This article is protected by copyright. All rights reserved.

Abstract
Addressing molecular recognition in the context of evolution requires pursuing new molecular
targets to enable the development of agonists or antagonists with new mechanisms of action.
Disruption of transcriptional regulation through targeting transcription factors that regulate the
expression of key enzymes in bacterial metabolism may provide a promising method for
controlling the bacterial metabolic pathways. To this end, we have selectively targeted a bacterial
transcription regulator through the design and synthesis of a series of γ-aminobutyric acid (GABA)
derivatives, including (S)-4-amino-5-phenoxypentanoate (4-phenoxymethyl-GABA), which are
based on docking insights gained from a previously-solved crystal structure of GabR from Bacillus
subtilis. This target was selected because GabR strictly controls GABA metabolism by regulating
the transcription of the gabT/D operon. These GabR transcription modulators are selective for the
bacterial transcription factor GabR and are unable to bind to structural homologs of GabR due to
distinct steric constraints. We have obtained a crystal structure of 4-phenoxymethyl-GABA bound
as an external aldimine with PLP in the effector binding site of GabR, which suggests that this
compound is capable of binding and reacting in the same manner as the native effector ligand.
Inhibition assays demonstrate high selectivity of 4-phenoxymethyl-GABA for bacterial GabR
versus several selected eukaryotic enzymes. Single-molecule fluorescence resonance energy
transfer (smFRET) experiments reveal a ligand-induced DNA distortion that is very similar to that
of the native effector GABA, suggesting that the compound functions as a potential selective
agonist of GabR.
This article is protected by copyright. All rights reserved.

Keywords: Bacterial transcriptional regulation, MocR, GabR, PLP, structure-based ligand design,
DNA distortion, X-ray crystallography, single-molecule FRET, organic synthesis.
This research demonstrates the design and synthesis of a tight-binding and potential
selective agonist of the transcriptional regulator GabR in non-pathogenic Bacillus subtilis, and
thus opens the door to the possibility of specifically targeting the modulation of bacterial
transcription. Analysis of the effects of a new and selective ligand is explored through the use of
biochemical techniques, including X-ray crystallography and enzymatic assays, in conjunction
with the biophysical technique single-molecule FRET. The manuscript proves the concept in
designing and synthesizing modulators of the transcriptional regulator GabR in key bacterial
metabolism pathways.
Abbreviations:
GABA: γ-aminobutyric acid
GABA-AT: GABA-aminotransferase
SSDH: succinate semialdehyde aminotransferase
wHTH: winged helix-turn-helix
eb/o: effector binding/oligomerization
Asp-AT: aspartate aminotransferase
This article is protected by copyright. All rights reserved.

PLP: pyridoxal-5’-phosphate
GAD: glutamate decarboxylase
OAT: ornithine aminotransferase
Ala-AT: alanine aminotransferase
smFRET: single-molecule Förster resonance energy transfer
FRET: Förster resonance energy transfer
INTRODUCTION
In B. subtilis, the gabT/D operon [Fig. 1(a)] encodes the genes responsible for γ-
aminobutyric acid (GABA) catabolism including GABA aminotransferase (GABA-AT), which is
the product of the gabT gene, and succinate semialdehyde dehydrogenase (SSDH), which is the
product of the gabD gene. Transcription of enzymes in this metabolic pathway is strictly regulated
by the transcription regulator GabR (UNIPROT: P94426), a member of the MocR/GabR subfamily
of the GntR family of transcription factors.1-3 GabR forms a head-to-tail homodimer in solution
and has high sequence homology to type-I aminotransferases. Each subunit of GabR consists of
an N-terminal winged helix-turn-helix (wHTH) DNA-binding domain and a C-terminal effector-
binding/oligomerization (eb/o) domain [Fig. 1(b)] homologous to highly conserved Type-I
aminotransferases, such as GABA-AT and aspartate aminotransferase (Asp-AT).2-4
In 2015, Wu et al. reported that GabR facilitates a “partial” aminotransferase-like reaction
involving GABA and the cofactor pyridoxal-5’-phosphate (PLP), forming an external aldimine.
However, GabR is incapable of completing the “Ping-Pong” transamination reaction due to
This article is protected by copyright. All rights reserved.

conformational restraints imposed by Tyr 281.4 As a result, GabR functions solely as a PLP-
dependent transcriptional regulator, responding to the concentration of GABA but without
catalytic capacity.4 The current GabR-DNA binding model suggests that GabR binds as a
homodimer to regulate transcription. GabR specifically recognizes GABA, which induces protein
conformational changes upon binding.5-7 The changes in protein conformation are believed to
result in DNA distortion, thereby activating transcription of the gabT/D operon.4–7 The specific
recognition of GABA by GabR as opposed to α-amino acids is critical to ensure activation of the
gabT/D operon only when needed, and is particularly important in the bacterial stress response and
survival. The importance of selectivity for GABA was revealed by gabR, gabT, and gabD gene
knockout studies in B. subtilis, which resulted in mutated organisms that were incapable of
utilizing GABA as their sole carbon/nitrogen source. Moreover, GabR has been shown to be
sensitive to increased intracellular GABA concentrations.1,2
To regulate gene expression, an organism may exercise control directly at the site of
transcription.8,9 Through transcription regulation, both E. coli and B. subtilis control metabolic
pathways responsible for acid elimination in response to low pH environments, albeit via different
mechanisms.10-12 Under acidic stress conditions, some bacteria, including E. coli, can maintain pH
homeostasis by consuming intracellular glutamate via the glutamate decarboxylase (GAD)
reaction. The GAD reaction eliminates one intracellular proton and in conjunction produces one
GABA molecule that is exported from the intracellular space via the antiporter GadT2. However,
This article is protected by copyright. All rights reserved.

this pathway requires the GABA shunt pathway to relinquish the key nitrogen source glutamate,
and glutamate depletion is detrimental to bacteria.
The role of the gabT/D operon in B. subtilis appears to function similarly to the GABA
shunt pathway in E. coli under acidic stress. This would place GabR in a key role affecting the
ability of B. subtilis to survive in acidic environments, thus making this protein a viable target to
manipulate the pH homeostasis stress response in bacteria. With the ultimate goal of developing a
new class of bacterial gabT/D transcription modulators, we have synthesized a series of aryl-
functionalized GABA derivatives, including (S)-4-amino-5-phenoxypentanoate (referred to
hereafter as 4-phenoxymethyl-GABA), that were designed to specifically target the binding site of
GabR based on a known crystal structure of GABA (PDB: 5T4J). We have obtained a 2.75 Å
resolution protein-ligand complex structure of 4-phenoxymethyl-GABA bound as an external
aldimine (4-Phenoxymethyl-GABA-PLP, hereafter referred to as PGP) in the eb/o domain of
GabR. Selectivity of this compounded was probed against other PLP-dependent enzymes
including GABA-AT, ornithine aminotransferase (OAT), alanine aminotransferase (Ala-AT), and
Asp-AT, finding little-to-no inhibition. To probe possible effects of 4-phenoxymethyl-GABA on
GabR-mediated transcription in bacteria, we conducted single-molecule Förster resonance energy
transfer (smFRET) experiments to examine DNA distortion induced by the binding of 4-
phenoxymethyl-GABA.
RESULTS
This article is protected by copyright. All rights reserved.

Chemistry. Tosylate 2 was prepared as outlined in Scheme 1 from commercially available
lactam 1 by reaction with p-toluenesulfonyl chloride in the presence of triethylamine and a
catalytic amount of N,N-dimethylaminopyridine (DMAP). Aryloxy- and arylthiolactams 8-12
were prepared through nucleophilic substitution by nucleophiles 3-7 with potassium carbonate in
refluxing acetonitrile over a 20-hour reaction period. Lactam intermediates 8-12 were then each
individually treated with 6 M HCl under reflux to produce the final aryloxy-GABA derivatives
(13-17) in quantitative or nearly quantitative yields. In most cases, final compounds were obtained
directly from the hydrolysis in purities greater than 95%. Further purification was achieved via
reverse phase flash chromatography where specified.
Crystallography. A crystal of truncated (positions 107-471) holo GabR was soaked
overnight with 4-phenoxymethyl-GABA 13. An X-ray crystal complex structure of 13 and PLP as
an external aldimine bound to the eb/o domain of GabR was solved by molecular replacement
using a dimer from a previously reported truncated structure of GabR (PDB code: 5T4J) after
deletion of all water molecules and ligand atoms. In space group P41212, one asymmetric unit
contained 4 monomers. Two monomers form one homodimer as a truncated biological assembly
(with the wHTH domain deleted), and there are two truncated biological assemblies in one
asymmetric unit. The final model was refined to a resolution of 2.75 Å with a Rfree/Rwork values of
20.18%/17.31%, respectively. Final refinement statistics are reported in Table 1. The solved
structure of GabR-PGP has been deposited in the PDB as (PDB code: 6UXZ).
This article is protected by copyright. All rights reserved.

After molecular replacement, the experimental map (Fo-Fc) showed electron density in the
binding site that could not be attributed to PLP alone, demonstrating the presence of the ligand 4-
phenoxymethyl-GABA, bound as an external aldimine formed between PLP and 13, and termed
PGP. In the refined model [Fig. 2(a)], hydrogen bonds were observed between Arg 207 (3.4 Å),
His 114 (2.7 Å), Arg 430 (2.5 Å), and the carboxylate of PGP [Fig. 2(a)]. A superimposition of
external aldimine PLP-GABA and external aldimine PGP highlights the unique space occupied by
the phenoxyl substituent [Fig. 2(b)]. A polder map (Fo-Fc) of the external aldimine was generated
by removing PGP, as shown in Figure 3.
Off-Target Screening. The eb/o domain of GabR type bacterial transcription regulator are
homologous and structurally similar to PLP-dependent aminotransferases including those in
eukaryotes. To screen for potential off-target activity and to ascertain the potential of our 4-
phenoxymethyl-GABA as a selective GabR agonist in bacteria without interaction with
mammalian enzymes, four important PLP-dependent human aminotransferase enzymes, known to
have off-target inhibition potential when targeting a PLP-dependent protein, were selected,
including GABA-AT, Asp-AT, Ala-AT and OAT. These enzymes were selected because all four
utilize the same co-factor (PLP), follow similar enzymatic reaction mechanisms, and represent the
general molecular recognition employed by PLP-dependent aminotransferases. They are
commonly used to test the selectivity of synthesized covalent inactivators against PLP-dependent
aminotransfeases.13-15 At the three concentrations tested, modest inhibition was only observed at
the highest concentration of 5.0 mM (Fig. 4). The conducted off-target screen also proves a
This article is protected by copyright. All rights reserved.

concept that designed ligand displays potential to treat bacterial pathogens that rely on GabR-
mediated transcription regulation for virulence or survival in the future.
Effects of 13 on oligomerization and DNA distortion. To test the effect of 4-
phenoxymethyl-GABA (13) as a potential agonist of GabR that may promote transcription, the
transcription activation signal in the DNA caused by GABA needed first to be elucidated. To this
end, smFRET experiments were performed using a 53 base-pair (bp) fragment of the GabR-DNA
binding region. A GabR-DNA binding model was generated based on SAXS data in an early
publication and was generously provided by Dr. Till Boecking at New South Wales University,
Australia via personal communication.5 Using this GabR-DNA complex model as a guide (Fig.
S2), fluorescent dyes (Cy3-donor and Cy5-acceptor) were positioned 19 bp apart (Fig. S2), so that
changes in the DNA structure could be detected via FRET signals correlating to the distance
between the two fluorophores.
For immobilized DNA constructs in the absence of GabR, smFRET time traces for
individual molecules exhibited an exclusively low FRET signal (Fig. S3). Histograms generated
from the accumulated behavior of the observed molecules were well fit by a single-peak, having a
Gaussian distribution with an average FRET (<FRET>) value of 0.23 [Fig. 5(a)]. This FRET value
is consistent with the expected distance between Cy3 and Cy5 fluorophores positioned 19 bp apart
on opposing strands (Table S2). For DNA preincubated with GabR prior to immobilization, the
time traces displayed two distinct FRET states: a low FRET state consistent with free DNA, and a
This article is protected by copyright. All rights reserved.

distinctly higher FRET state. The smFRET histogram for the GabR-DNA complexes have a
bimodal distribution [Fig. 5(b)]. Analysis by Gaussian fitting yielded two peaks with <FRET>
values of 0.26 and 0.48 (Table S1). Extended observation did not detect any conversion between
the two FRET distributions, indicating that these FRET values represent two distinct populations.
The low FRET population was thus assigned to free DNA, and the high FRET population
represents the GabR-DNA complex. The increased FRET level for the GabR-DNA complex
suggests that GabR binding to DNA induces a bent conformation, perhaps through local melting
within the linker region. This observation provides direct evidence of GabR modulating DNA
conformation and supports structural studies that modeled a bent DNA conformation for the GabR-
bound state. The time traces for the GabR-DNA complexes did not exhibit any detectable
transitions, indicating that the complex is stably formed with the DNA held in a static, nonlinear
orientation with no switching between conformations.
The high FRET state associated with the GabR-DNA complex provides a signal for the
inactivated complex and serves as a reference for testing whether addition of activating ligands,
such as GABA, will induce further conformational change in the GabR-DNA complex for the
activation state. Indeed, addition of GABA to GabR-DNA shifted the FRET distribution to a lower
value, <FRET> = 0.32 as shown in the smFRET histograms (Figure 5c) and the time traces (Figure
S3). The lower FRET state suggests that GABA binding relaxes the bent conformation or induces
an out-of-plane distortion. Similarly, addition of 13 to GabR-DNA lowered the <FRET> value to
This article is protected by copyright. All rights reserved.

0.41 [Fig. 5(d)], consistent with the model that 13 acts as an agonist by occupying the GABA-
binding pocket and inducing some local distortion in GabR-bound DNA.
Although distance is only one of many factors that can cause changes in FRET, interdye
distances calculated from average FRET values can provide useful insight into how GabR and its
effectors alter the conformation of the bound DNA. Based on prior SAXS results indicating that
bound DNA wraps around GabR, the interdye distances were interpreted via a multi-segmented
geometric model for the DNA conformation (Fig. S4). From this model, estimates of the bend
angles induced by GabR binding were calculated (Table S2). The unbound DNA had an interdye
distance of 6.6 nm, which yields a bend angle of 0° and is consistent with a 19-bp dye separation.
GabR binding to DNA reduced the interdye distance to 5.5 nm, which corresponds to a DNA bend
angle of 47°. The structural information for the bound DNA gleaned from the smFRET
experiments complements the SAXS-derived, global structure of GabR to provide a more complete
picture of the GabR-DNA complex.5 In the presence of GABA, the interdye distance increased to
6.1 nm, suggesting either relaxation of the DNA bending or out-of-plane twisting to further
separate the dyes.
DISCUSSION
In their findings, Wu et al. concluded that GabR is incapable of catalyzing a transamination
reaction and that GabR is only able to facilitate the formation of the external aldimine.4 In addition,
a synthesized GABA derivative (S)-4-amino-5-fluoropentanoic acid (AFPA) was proven to be an
This article is protected by copyright. All rights reserved.

agonist of GabR, activating transcription of the gabT/D operon. Compared to GABA, AFPA
resulted in greater activation in vitro, although AFPA yielded lower activation in vivo.4 GABA
permease (GabP), a transmembrane protein responsible for the translocation of external GABA to
within the cell, may have low affinity for AFPA and be the cause of the discrepancy. Given that
AFPA was able to react similarly to GABA and activate transcription, this molecule was chosen
as the lead compound in this research to be further optimized. Improving selectivity toward GabR
was an important first step in the molecular probe design process. To validate and further improve
ligand design, the GabR-PLP-GABA co-crystal structure was explored to identify structural
characteristics unique to the GabR binding site. It was observed that GabR possesses a
hydrophobic pocket that is not found in the active sites of other PLP-dependent enzymes, including
GABA-AT, Asp-AT, and OAT. On this basis, we used a known GabR agonist AFPA as a
molecular probe for GabR to examine the transcription factors binding site reactivity.4 It was
hypothesized that aryloxyl- and arylthio moieties could replace the fluorine of AFPA and result in
increased binding affinity to GabR by forming additional stabilizing contacts in this hydrophobic
pocket in the GABA binding site while sterically excluding off-targets. To this end, we have
synthesized several aryloxyl- and arylthio derivatives, as summarized in Scheme 1.
The distances in the GabR-PGP structure deviate only slightly from the GabR-PLP-GABA
co-crystal structure [Fig. 2(b)]. These residues facilitate recognition of monocarboxylate ligands,
such as the native effector GABA, as well as synthetic GABA derivatives, such as AFPA, and
facilitate Schiff base formation with PLP. Formation of the Schiff base has been reported as the
This article is protected by copyright. All rights reserved.

chemical reaction responsible for the GabR mediated transcription activation of the gabT/D
operon.4 In addition, the conserved interactions between the phosphate group of PLP and Tyr 181,
Thr 309, Ser 311, Arg 319, and Ser 321, more commonly referred to as the “phosphate-binding
cup” for PLP dependent enzymes, were also observed in the GabR-PGP crystal structure, and these
interactions prevent co-factor displacement.4
As seen in Figure 2(a) the O3 atom of PGP and the acidic proton of Tyr 281 are within
hydrogen bonding distance (3.1 Å). This interaction is absent in the GabR-PLP-GABA co-crystal
structure (PDB code: 5T4J) because a hydrogen bonding acceptor does not exist in the PLP-GABA
external aldimine at the same position. The additional hydrogen bond likely further increases
selectivity and affinity of PGP compared to GABA alone in formation of the external aldimine of
PLP-GABA. It will be of interest to investigate any difference in binding or affinity with the
replacement O3 in 13 to an atom incapable of hydrogen bonding, such as in the arylthio analogue
17. A polder (omit) map16 shown in Figure 3 (Fo-Fc at 3.0σ) superimposed with PGP from the
GabR binding site further supported the formation and position of the external aldimine between
PLP and 13 [Fig. 2(a)].
To elucidate why the 4-phenoxymethyl-GABA derivative 13 is selective against the off-
target enzymes, the active sites from the inhibited structures of these off-targets were used to
visualize how PLP-13 (PGP) compared to the positions of the inhibitor complex in the enzyme’s
active sites. The PLP moieties were manually structurally aligned in Chimera and the van der
Waals radii of PGP were represented as a surface feature (Fig. S1). It was observed that PGP would
This article is protected by copyright. All rights reserved.

encounter unfavorable interactions within the active sites of these off-target enzymes. Either the
side chain residues or the backbone of the active sites clashed with the surface feature of PGP.
These unfavorable interactions should preclude 13 from binding to these off-target enzymes and
is the reasoning for the observed limited inhibition. As proposed, the phenoxyl moiety of 13
includes additional steric bulk that limits off-target inhibition.
Proposed mechanisms for transcription regulation include direct interactions between
activators and RNA polymerase, or as is suspected for GabR, activators may induce DNA
distortion that reposition promoters in more favorable orientations to interact with RNA
polymerase.17-20 We hypothesize a mechanism for GabR wherein GabR limits the promoter-RNA
polymerase interactions necessary for transcription during DNA oligomerization. Next, binding of
GABA within the eb/o domain induces protein conformational changes resulting in DNA
distortion resulting in the necessary promoter-RNA polymerase interactions being restored, and
transcription of the gabT/D operon is activated.
Given that transcription is only activated in the presence of an effector ligand such as
GABA, the FRET distribution in the presence of a ligand was expected to be distinguishable from
both free DNA and GabR-DNA complex. Indeed, the FRET distribution in the presence of GABA
was unique and intermediate to the values of free and GabR-DNA complex. This observation
suggests that 1) DNA is initially distorted during oligomerization with GabR and that 2) DNA
undergoes further distortion during the protein conformational changes induced by GABA. A
lower <FRET> value suggests that DNA distortion caused by GABA is a relaxation in the
This article is protected by copyright. All rights reserved.

curvature of the DNA; however, twisting in the DNA strand could also promote activation. This
twisting-based activation could originate from the manner in which DNA wraps around GabR,
with positive charges following an “S” type of shape along the surface of GabR. Various groups
have reported similar findings regarding the GabR-DNA complex in the presence of GABA.5-7,21
The results in this report are consistent with the findings of Al-Zyoud, whose group was the first
to suggest that the DNA wraps around GabR during oligomerization and that the addition of
GABA results in a larger overall complex structure of GabR.5 It was proposed that the larger
complex structure would cause the bending in the natural curvature of the DNA to decrease, thus
activating transcription.5 Similar DNA distortion mechanisms had been previously observed in the
well-studied MerR family of transcription factors. These proteins bind between the -35 and -10
elements, activating suboptimal promoters by contracting DNA in response to external stimuli.22-
25 Even though the -35 element of the gabT promoter is positioned within the GabR binding site,
it does not play a role during oligomerization and is free to interact with RNA polymerase. A
similar effect on DNA distortion and <FRET> values would be expected if 4-phenoxymethyl-
GABA 13 functions as an agonist of GabR. Similar to GABA, 13 caused the <FRET> distribution
to decrease in between the values of free and oligomeric DNA, indicating that 13 elicits a
comparable effect on DNA distortion.
This article is protected by copyright. All rights reserved.

Conclusion
The novel synthetic GABA derivative, (S)-4-amino-5-phenoxypentanoate (13) and related
analogs were designed to selectively bind to GabR and thereby to potentially modulate
transcription induced by GabR, and these studies demonstrate that 13 can potentially functions as
a GabR agonist. A crystal structure of truncated GabR with 13 revealed the formation of an
external aldimine between 13 and the cofactor (PLP) in the binding site. smFRET studies support
an agonistic response elicited by 13 that is similar to the native effector ligand, GABA, via DNA
distortion. The distortion of DNA is necessary to activate transcription of the gabT/D operon.
Furthermore, in vitro studies indicate that 13 is selective toward GabR with very limited inhibition
of off-target human aminotransferases.
MATERIALS AND METHODS
Anhydrous solvents (THF, CH2Cl2, MeOH, Et3N, and DMF) were distilled prior to use.
All other solvents, reactants, and reagents were purchased from commercial vendors and were
used without further purification. Melting points were determined in capillary tubes using a Büchi
melting point B-540 apparatus and are uncorrected. 1H-NMR spectra were recorded at 500 MHz,
13
using a Bruker Avance III 500 (direct cryoprobe), and C-NMR spectra were obtained at 126
MHz using the same instrument. High-resolution mass spectral data were obtained at the Integrated
Molecular Structure Education and Research Center (IMSERC, Northwestern University) on an
Agilent 6210A TOF mass spectrometer in positive ion mode coupled to an Agilent 1200 series
This article is protected by copyright. All rights reserved.

HPLC system. Data were processed using MassHunter software version B.04.00. Flash column
chromatography was performed using an Agilent 971-FP automated flash purification system with
a Varian column station and SiliCycle cartridges (12-80 g, both normal and High Performance).
Analytical HPLC was performed using an Agilent Infinity 1260 HPLC system and injection
volumes of 5-10 µL. A Phenomenex Luna 5 µm C-8(2) 100 Å column, 50 x 4.60 mm, was used
for all HPLC experiments, using a 10-minute gradient of 95% H2O/5% acetonitrile + 0.05% TFA
to 95% acetonitrile/5% H2O + 0.05% TFA, at 1.5 mL/min. The purity of all final target compounds
was found to be ≥95% by HPLC. Analytical thin-layer chromatography was performed on
Silicycle extra-hard 250 µm TLC plates. Compounds were visualized with short-wavelength UV
light, and with ninhydrin or with phosphomolybdic acid, where appropriate.
(S)-(5-Oxopyrrolidin-2-yl)methyl 4-Methylbenzenesulfonate (2) Synthesis: To a stirred
solution of (S)-5-(hydroxymethyl)-2-pyrrolidinone 7 (1.10 g, 9.56 mmol) and p-toluenesulfonyl
chloride (1.83 g, 9.59 mmol) in CH2Cl2 (22 mL) at 0°C was added DMAP (117 mg, 0.96 mmol)
and Et3N (2.68 mL, 19.2 mmol). The resulting mixture was allowed to warm to 23°C and stirred
for 20 h. The reaction was then quenched with water (40 mL), and the aqueous layer was extracted
with CH2Cl2. The combined organic extracts were washed with 1 M HCl and dried over anhydrous
Na2SO4. Removal of solvent under reduced pressure followed by dry load flash chromatography
purification (1/99 MeOH-EA as the eluent) yielded 1 as a white solid (2.37 g, 92%): mp 128.5-
133.5°C, 1H NMR (500 MHz, CDCl3) δ 1.75-1.80 (m, 1H), 2.19-2.35 (m, 3H), 2.44(s, 3H), 3.85-
This article is protected by copyright. All rights reserved.

3.92 (m, 2H), 4.00-4.03 (m, 1H), 6.49 (s, 1H), 7.35 (d, 2H, J = 8.0 Hz), 7.77 (d, 2H, J = 8.1 Hz);
13C NMR (100 MHz, CDCl3) δ 21.6, 22.7, 29.2, 52.5, 71.9, 121.9, 130.0, 132.3, 145.3, 178.2.
HRMS (ESI) calculated for C11H15NO3 [M+H]+ 210.1125, found 210.1125. HPLC purity = 99%
pure.
This article is protected by copyright. All rights reserved.

General Procedure 1: Nucleophilic displacement of tosylate 2 by phenol and thiophenol
derivatives. To a 20 mL dram vial charged with 2 and K2CO3 (5.0 eq.) was added dry acetonitrile
(0.3M) followed by an excess amount of phenol or thiophenol derivative (3.0 eq.). The reaction
vessel was sealed under argon and sonicated for 2 min before stirring at 110°C overnight under an
inert atmosphere. HPLC and TLC analysis with PMA staining were used to monitor consumption
of starting material. Once the starting material disappeared, the reaction mixture was cooled and
partitioned between EtOAc (10 mL) and H2O (10 mL). The layers were separated, and the aqueous
phase was extracted with EtOAc (3 x 5 mL). The organic fractions were combined, dried over
anhydrous sodium sulfate, and concentrated. Purification, if necessary, is described below under
subheadings for individual compounds.
(S)-5-(Phenoxymethyl)pyrrolidin-2-one (8).
The title compound was obtained according to the general procedure described above
starting from tosylate 2 (0.420 g, 1.57 mmol) and phenol (0.249 g, 2.65 mmol). Purification by
flash column chromatography (silica gel; gradient from 5% EtOAc/hexane to 100% EtOAc over
10 min, then 100% EtOAc for 10 min) gave lactam 8 (0.227 g, 75% yield) as a white fluffy solid:
mp 73-77°C; 1H NMR (500 MHz, chloroform-d) δ 7.34 – 7.23 (m, 2H), 6.98 (td, J = 7.4, 1.1 Hz,
1H), 6.92 – 6.83 (m, 2H), 6.18 (s, 1H), 4.17 – 4.02 (m, 1H), 3.99 (dd, J = 9.2, 3.9 Hz, 1H), 3.82
(dd, J = 9.1, 7.8 Hz, 1H), 2.52 – 2.25 (m, 3H), 1.99 – 1.83 (m, 1H); 13C NMR (126 MHz, CDCl3)
This article is protected by copyright. All rights reserved.

δ 177.81, 158.27, 129.59, 121.41, 114.48, 71.28, 53.28, 29.53, 23.15; HRMS (ESI) calcd for
C11H14NO2 [M+H]+ 192.1021, found 192.1019; HPLC purity = 98% pure.
This article is protected by copyright. All rights reserved.

General Procedure 2: Hydrolytic Pyrrolidine Ring Opening: A measured amount of the
purified pyrrolidine derivative was added to a small dram vial alone with a stir bar. Then a 1-2
mL amount of 6 M HCl(aq.) was added to the reaction vessel followed by degassing and sealing
with parafilm. The reaction vial was then heated at 115°C for 2-3 h. The reaction was monitored
using TLC and HPLC (method A). Once the starting material was consumed, the reaction vial was
heated at 80°C under reduced pressure to remove solvents and crystallize the HCl salt product
which was dried overnight in vacuo.
(S)-4-Amino-5-phenoxypentanoic acid hydrochloride (13)
The title compound was obtained according to the general procedure described above
starting from lactam 3 (88 mg, 0.458 mmol). Purification by flash chromatography (C18-silica gel;
isocratic 100% H2O elution) gave 4-phenoxymethyl-GABA 13 (49 mg, 43% yield) as a white
solid: mp 155.6-160°C; 1H NMR (500 MHz, d6-DMSO) δ 12.12 (bs, 1H), 8.40 (bs, 3H), 7.32 (t,
J=8.7Hz, 2H), 7.00 (d, J=8.4Hz, 3H), 4.15 (dd, J=36.3, 10.2Hz, 1H), 4.12 (dd, J=36.3, 10.2Hz,
1H), 3.52 (m, 1H), 2.47 (m, 1H), 1.94 (m, 2H); 13C NMR (500 MHz, d6-DMSO) δ 173.68, 157.84,
129.51, 121.23, 114.74, 67.01, 49.41, 29.49, 24.70; HRMS (ESI) calcd for C11H15NO3 [M+H]+
210.1125, found 210.1125; HPLC purity = > 99% pure.
Protein Purification: For protein expression and purification, E. coli BL21(DE3) competent cells
were transformed with pET24-GabR-eb/o and stored as a glycerol stock. A starter culture in Luria-
This article is protected by copyright. All rights reserved.

Betrani (LB) medium (50 mL) containing kanamycin (50 μg/mL) grown overnight at 37°C and
shaken at 250 rpm was used to inoculate a 2 L (50-fold dilution) of LB medium supplemented with
kanamycin (50 μg/mL). The expression culture was grown at 37°C, 250 rpm until OD600 ≈ 0.6,
then cooled to 4°C for 1 h before induction with isopropyl β-D-1-thiogalactopyranoside (0.5 mM).
After the culture was incubated for an additional 16-18 h at 25°C and shaken at 200 rpm, cells
were harvested with 3 rounds of centrifugation at 28048 × g for 15 min. The cell pellet was flash
frozen in liquid nitrogen for subsequent storage at -80°C.
Cell pellets were resuspended in Buffer A (1X PBS Buffer, 15 mM imidazole, 100 uM
PLP, pH 8.5) and lysed by sonication (5 min) alternating 10 s bursts and 20 s rests (Thermo Fisher
Sonic Dimembrator 500, 30% amplitude, microtip). Cell debris was removed by centrifugation at
31,000 × g for 1 h. The remaining supernatant was subjected to Ni2+ affinity chromatography
using a HisTrap FF column (GE Healthcare) pre-equilibrated with Buffer A. Protein was eluted
using a linear gradient with Buffer B (from 5 mM to 250 mM imidazole, supplemented with 2X
PBS Buffer, 100 uM PLP, pH 8.5). Fractions containing protein, identified by observing
absorbance at 280 nm, were pooled together and subjected to size exclusion chromatography
(HiLoad 16/60 Superdex 200 pg; GE Healthcare) using gel filtration Buffer (Buffer B with 300
mM imidazole). A single eluting peak was observed during this purification step and fractions
containing protein were pooled together. Both an aliquot from the Ni2+ affinity column eluate and
from the final purified truncated GabR protein appeared homogenous when characterized by a
Coomassie-stained 12% SDS-PAGE gel. Protein concentrations were measured using the
This article is protected by copyright. All rights reserved.

Bradford assay (BioRad) with bovine serum albumin standards.
Protein Crystallization: For crystallization experiments, protein solution was buffer exchanged
with Buffer A. Crystals of truncated GabR were obtained via hanging drop vapor diffusion by
mixing 2.5 uL of (10 mg/mL) protein solution with 2.5 uL of well solution (10% PEG 3350, 100
mM imidazole pH 8.0, and 200 mM Li2SO4). Crystals were allowed to grow at room temperature
and reached optimal size and morphology for data collection after 7-10 days. Truncated GabR
(eb/o domain) crystals were then soaked overnight in well solution supplemented with compound
13 (25 mM). Soaked crystals were transferred to the well solution supplemented with 30% PEG400
as cryo protectant and flash frozen in liquid N2.
Data Collection and Processing: Monochromatic data was collected at Beam Line 19-BM at the
Structural Biology Center (SBC), Advanced Photon Source (APS), Argonne National Laboratory
(ANL). Diffraction data was collected at a wavelength of 0.98 Å at 100 K using a ADSC 315r
Charged Couple Device (CCD) detector. Data sets were indexed and integrated using iMosflm and
scaled with SCALA both in the CCP4 software suite.26 Data statistics are summarized in Data
Table 1.
Model Building and Refinement: The GabR-PGP structure was solved by molecular replacement
using PHASER in the Phenix software suite.27 The first search model was based on a previously
This article is protected by copyright. All rights reserved.

published structure of the truncated GabR eb/o domain (PDB Code: 5T4J). The model was rebuilt
using COOT,28 refined using CCP4 and analyzed in COOT and USCF Chimera.29 All ligands were
built and regularized using JLigand in the CCP4 program suite. Final refinement statistics are
reported in Table 1. Structural figures were made in USCF Chimera.
Human Aminotransferase Assay Panel: All reagents were purchased from Sigma-Aldrich (St.
Louis, MO, USA). Recombinant human pyrroline-5-carboxylate reductase 1 was purchased from
Creative Biomart. GABA-AT was isolated and purified from freshly harvested pig brain according
to a published procedure.30 OAT was cloned, expressed, and purified according to a published
procedure.31 Literature procedures were used to determine the inhibition of GABA-AT,32 human
OAT,33 L-aspartate aminotransferase,34 and L-alanine aminotransferase34 by the GABA analogues.
Assays were recorded on a Synergy H1 hybrid multi-mode microplate reader (Biotek, USA), with
transparent 96 well plates. All assays were performed in duplicate.
Single-molecule Fluorescence Resonance Energy Transfer (smFRET) Imaging: Single-
stranded oligonucleotides (Integrated DNA Technology Inc., Coralville, IA) containing internal
Cy3 and Cy5 fluorescence probes inserted as base substitutions were designed based on the GabR
binding region.
Forward sequence:
5’-
TCTGATACCATCAAAAAGATTATAATTGGTA/iCy5/CTTTTCATCATACCAAAGATTATAATTT
TCA-3’-Biotin
This article is protected by copyright. All rights reserved.

Reverse sequence:
5’-
TGAAAATTATAATCTTTGGTATGATGAAAAGATACCAATTATAA/iCy3/CTTTTTGATGGTATC
AGA-3’
The 3’ of the forward sequence was biotinylated to serve as the tether site for
immobilization of DNA to the slide surface by biotin-streptavidin conjugation. The oligos were
checked for appropriate emission signals by excitation of the Cy5 dye at 610 nm and by excitation
of the Cy3 dye at 530 nm. Sets of the oligos were mixed in 20 mM Tris-HCl buffer to a final
concentration of 20 uM and annealed using a personal Mastercycler (Eppendorf). DNA was stored
at -20°C. DNA samples were serially diluted with buffer A (50 mM Tris, 50 mM NaCl, pH 8.0)
to a final concentration of 10 pM. For DNA+GabR samples, GabR (5000 nM) was incubated with
DNA at 25°C for 30 min in DNA binding buffer (20 mM Tris HCl, pH 8.0 50 mM KCl, 2 mM
MgCl2, 5% (vol/vol) glycerol, 1 mM EDTA, and 0.05% Nonidet P-40) then serially diluted with
DNA binding buffer to a final DNA concentration of 10 pM. Flow chambers constructed from
cleaned quartz slides (Technical Glass Products, Painesville, OH) were prepared for DNA
immobilization by flowing 0.5 mg/mL BSA-biotin and streptavidin (0.1 mg/mL) with washes by
buffer A. Solutions containing DNA or pre-incubated GabR-DNA complexes were then flowed
through the chamber and allowed to incubate for 4 min. The sample was washed with DNA binding
buffer and OSS buffer (25 mM Tris pH 8.0, 50 mM 8 mM Trolox, 5% glycerol, 1 mM EDTA, 1
mM DTT, 0.05% Nonident P-40, 50 mM KCl, 2 mM MgCl2, 8% (m/v) glucose, 0.2 (mg/mL)
glucose oxidase, 0.1 mg/mL catalase) was applied immediately before imaging to reduce
photobleaching. The samples were imaged by prism-type total internal reflection with a 60X water
This article is protected by copyright. All rights reserved.

objective on an Olympus IX-83 with an EMCCD (Andor iXon DU897; Belfast, UK) cooled to -
60°C. The Cy3 and Cy5 emission signals were spatially separated for simultaneous imaging by a
series of dichroic mirrors (640 nm cutoff; Chroma Technologies, Bellow Falls, VT). Each pair of
emission signals corresponding to the same molecule was identified using an affine transformation
calibrated from a control slide coated with fluorescent fiducial markers. Data were acquired at 10
Hz. Data analysis was performed using custom Python software using established methodologies.
FRET histograms were fit by Gaussian distributions using the Levenberg−Marquardt optimization
with OriginPro (OriginLab, Northampton, MA).
Supplementary Material
The supplementary materials section provides a list of the synthesis protocols for all of the GABA
analogs described in the series of compounds synthesized. The SI contains figures that were
discussed in the main body of text, but not included because these figures are provide serve a
supporting role to the main findings. The SI also includes alternate views to some figures to
enhance display of experimental design and results.
This article is protected by copyright. All rights reserved.

Funding
The work was supported by NIH NIGMS R15GM113229 to DL and DB and by R01 DA030604
to RBS. This work made use of the IMSERC at Northwestern University, which has received
support from the Soft and Hybrid Nanotechnology Experimental (SHyNE) Resource (NSF ECCS-
1542205), the State of Illinois, and the International Institute for Nanotechnology (IIN).
This article is protected by copyright. All rights reserved.

REFERENCES
1.
Belitsky BR, Sonenshein AL (2002) GabR, a member of a novel protein family, regulates
the utilization of gamma-aminobutyrate in Bacillus subtilis. Mol Microbiol 45:569-583.
2.
Belitsky BR (2004) Bacillus subtilis GabR, a protein with DNA-binding and
aminotransferase domains, is a PLP-dependent transcriptional regulator. J Mol Biol 340:655-
664.
3.
Edayathumangalam R, Wu R, Garcia R, Wang Y, Wang W, Kreinbring CA, Bach A,
Liao J, Stone TA, Terwilliger TC, Hoang QQ, Belitsky BR, Petsko GA, Ringe D, Liu D (2013)
Crystal structure of Bacillus subtilis GabR, an autorepressor and transcriptional activator of
gabT. Proc Natl Acad Sci USA 110:17820-17825.
4.
Wu R, Sanishvili R, Belitsky BR, Juncosa JI, Le HV, Lehrer HJ, Farley M, Silverman
RB, Petsko GA, Ringe D, Liu D (2017) PLP and GABA trigger GabR-mediated transcription
regulation in Bacillus subtilis via external aldimine formation. Proc Natl Acad Sci USA
114:3891-3896.
5.
Al-Zyoud WA, Hynson RM, Ganuelas LA, Coster AC, Duff AP, Baker MA, Stewart AG,
Giannoulatou E, Ho JW, Gaus K, Liu D, Lee LK, Bocking T (2016) Binding of transcription
factor GabR to DNA requires recognition of DNA shape at a location distinct from its cognate
binding site. Nucleic Acids Res 44:1411-1420.
6.
Amidani D, Tramonti A, Canosa AV, Campanini B, Maggi S, Milano T, di Salvo ML,
Pascarella S, Contestabile R, Bettati S, Rivetti C (2017) Study of DNA binding and bending by
Bacillus subtilis GabR, a PLP-dependent transcription factor. Biochim Biophys Acta Gen Subj
1861:3474-3489.
7.
Park SA, Park YS, Lee KS (2017) Crystal structure of the C-terminal domain of Bacillus
subtilis GabR reveals a closed conformation by gamma-aminobutyric acid binding, inducing
transcriptional activation. Biochem Biophys Res Commun 487:287-291.
8.
Slattery M, Zhou T, Yang L, Dantas Machado AC, Gordan R, Rohs R (2014) Absence of
a simple code: how transcription factors read the genome. Trends Biochem Sci 39:381-399.
9.
Rohs R, Jin X, West SM, Joshi R, Honig B, Mann RS (2010) Origins of specificity in
protein-DNA recognition. Annu Rev Biochem 79:233-269.
This article is protected by copyright. All rights reserved.

10.
Wilks JC, Kitko RD, Cleeton SH, Lee GE, Ugwu CS, Jones BD, BonDurant SS,
Slonczewski JL (2009) Acid and base stress and transcriptomic responses in Bacillus subtilis.
Appl Environ Microbiol 75:981-990.
11.
Feehily C, Karatzas KA (2013) Role of glutamate metabolism in bacterial responses
towards acid and other stresses. J Appl Microbiol 114:11-24.
12.
De Biase D, Pennacchietti E (2012) Glutamate decarboxylase-dependent acid resistance
in orally acquired bacteria: function, distribution and biomedical implications of the gadBC
operon. Mol Microbiol 86:770-786.
13.
Lee H, Doud EH, Wu R, Sanishvili R, Juncosa JI, Liu D, Kelleher NL, Silverman RB
(2015) Mechanism of inactivation of gamma-aminobutyric acid aminotransferase by (1S,3S)-3-
amino-4-difluoromethylene-1-cyclopentanoic acid (CPP-115). J Am Chem Soc 137:2628-2640.
14.
Lee H, Le HV, Wu R, Doud E, Sanishvili R, Kellie JF, Compton PD, Pachaiyappan B,
Liu D, Kelleher NL, Silverman RB (2015) Mechanism of inactivation of GABA
aminotransferase by (E)- and (Z)-(1S,3S)-3-amino-4-fluoromethylenyl-1-cyclopentanoic acid.
ACS Chem Biol 10:2087-2098.
15.
Sarhan S, Knodgen B, Grauffel C, Seiler N (1993) Effects of inhibition of ornithine
aminotransferase on thioacetamide-induced hepatogenic encephalopathy. Neurochem Res
18:539-549.
16.
Liebschner D, Afonine PV, Moriarty NW, Poon BK, Sobolev OV, Terwilliger TC,
Adams PD (2017) Polder maps: improving OMIT maps by excluding bulk solvent. Acta Cryst
D73:148-157.
17.
van Hijum SA, Medema MH, Kuipers OP (2009) Mechanisms and evolution of control
logic in prokaryotic transcriptional regulation. Microbiol Mol Biol Rev 73:481-509.
18.
Lee DJ, Minchin SD, Busby SJ (2012) Activating transcription in bacteria. Annu Rev
Microbiol 66:125-152.
19.
Browning DF, Busby SJ (2004) The regulation of bacterial transcription initiation. Nat
Rev Microbiol 2:57-65.
20.
Busby SJW, Savery NJ (2007) Transcription Activation at Bacterial Promoters. In
Encyclopedia of Life Sciences, John Wiley & Sons: Chichester.
This article is protected by copyright. All rights reserved.

21.
Okuda K, Kato S, Ito T, Shiraki S, Kawase Y, Goto M, Kawashima S, Hemmi H, Fukada
H, Yoshimura T (2015) Role of the aminotransferase domain in Bacillus subtilis GabR, a
pyridoxal 5'-phosphate-dependent transcriptional regulator. Mol Microbiol 95:245-257.
22.
Brown NL, Stoyanov JV, Kidd SP, Hobman JL (2003) The MerR family of
transcriptional regulators. FEMS Microbiol Rev 27:145-163.
23.
Molina-Henares AJ, Godoy P, Duque E, Ramos JL (2009) A general profile for the MerR
family of transcriptional regulators constructed using the semi-automated Provalidator tool.
Environ Microbiol Rep 1:518-523.
24.
Hobman JL (2007) MerR family transcription activators: similar designs, different
specificities. Mol Microbiol 63:1275-1278.
25.
Heldwein EE, Brennan RG (2001) Crystal structure of the transcription activator BmrR
bound to DNA and a drug. Nature 409:378-382.
26.
Winn MD (2003) An overview of the CCP4 project in protein crystallography: an
example of a collaborative project. J Synchrotron Radiat 10:23-25.
27.
Adams PD, Grosse-Kunstleve RW, Hung LW, Ioerger TR, McCoy AJ, Moriarty NW,
Read RJ, Sacchettini JC, Sauter NK, Terwilliger TC (2002) PHENIX: building new software for
automated crystallographic structure determination. Acta Cryst D58:1948-1954.
28.
Emsley P, Lohkamp B, Scott WG, Cowtan K (2010) Features and development of Coot.
Acta Cryst D66:486-501.
29.
Paettersen E, Goddard T, Huang C, Couch G, Greenblatt D, Meng E, Ferrin T (2004)
UCSF Chimera--a visualization system for exploratory research and analysis. J Comput Chem
25:1605-1612.
30.
Koo YK, Nandi D, Silverman RB (2000) The multiple active enzyme species of gamma-
aminobutyric acid aminotransferase are not isozymes. Arch Biochem Biophys 374:248-254.
31.
Mascarenhas R, Le HV, Clevenger KD, Lehrer HJ, Ringe D, Kelleher NL, Silverman
RB, Liu D (2017) Selective targeting by a mechanism-based inactivator against pyridoxal 5'-
phosphate-dependent enzymes: Mechanisms of inactivation and alternative turnover.
Biochemistry 56:4951-4961.
This article is protected by copyright. All rights reserved.