Synthesis and evaluation of potent KCNQ2/3-specific channel activators

Article abstract of DOI:10.1124/mol.115.103200

KQT-like subfamily (KCNQ) channels are voltage-gated, non-inactivating potassium ion channels, and their down-regulation has been implicated in several hyperexcitability-related disorders, including epilepsy, neuropathic pain, and tinnitus. Activators of these channels reduce the excitability of central and peripheral neurons, and, as such, have therapeutic utility. Here, we synthetically modified several moieties of the KCNQ2-5 channel activator retigabine, an anticonvulsant approved by the U.S. Food and Drug Administration. By introducing a CF₃-group at the 4-position of the benzylamine moiety, combined with a fluorine atom at the 3-position of the aniline ring, we generated Ethyl (2-amino-3-fluoro-4-((4-(trifluoromethyl)benzyl)amino)phenyl)carbamate (RL648-81), a new KCNQ2/3-specific activator that is >15 times more potent and also more selective than retigabine. We suggest that RL648-81 is a promising clinical candidate for treating or preventing neurologic disorders associated with neuronal hyperexcitability.

Full text of DOI:10.1124/mol.115.103200

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
Title page
Synthesis and Evaluation of Potent KCNQ2/3-specific Channel Activators
Manoj Kumar, Nicholas Reed, Ruiting Liu, Elias Aizenman, Peter Wipf and Thanos
Tzounopoulos
Department of Otolaryngology, University of Pittsburgh School of Medicine, Pittsburgh,
Pennsylvania 15261 (M.K., T.T); Department of Chemistry, University of Pittsburgh, Pittsburgh,
Pennsylvania 15260 (N.R., R.L., PW); Department of Neurobiology, University of Pittsburgh,
Pittsburgh, Pennsylvania 15261 (E.A., T.T.) and Pittsburgh Institute for Neurodegenerative
Diseases, University of Pittsburgh School of Medicine (E.A).

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
Running title page
Novel KCNQ2/3 channel activators
Corresponding authors:
Thanos Tzounopoulos, Department of Otolaryngology and Department of Neurobiology,
University of Pittsburgh, Pittsburgh, Pennsylvania 15261
Email: thanos@pitt.edu
Peter Wipf, Department of Chemistry, University of Pittsburgh, Pittsburgh, Pennsylvania 15260
Email: pwipf@pitt.edu
Number of Text pages: 36
Number of Tables: 0
Number of Figures: 9
Number of References: 43
Number of words in Abstract: 133
Number of words in Introduction: 709
Number of words in Discussion: 732

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
Abbreviations: ADME/tox, Absorption, distribution, metabolism, excretion/ toxicity; Cbz,
benzyloxycarbonyl; CHO, chinese hamster ovary; DMSO, dimethyl sulfoxide; DIPEA, di-
isopropylethylamine; EGTA, ethylene glycol-bis(β-aminoethyl ether)-N,N,N’,N’-tetraacetic
acid; HEPES, N-(2-hydroxyethyl)-1-piperazineethanesulfonic acid; Kv, voltage-gated potassium
channel; PTSA, para-toluenesulfonic acid; SAR, structure-activity relationship; SNAr,
nucleophilic aromatic substitution.

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
Abstract
KCNQ channels are voltage-gated, non-inactivating potassium ion channels, and their down-
regulation has been implicated in several hyperexcitability-related disorders, including epilepsy,
neuropathic pain and tinnitus. Activators of these channels reduce the excitability of central and
peripheral neurons, and, as such, have therapeutic utility. Here, we synthetically modified several
moieties of the KCNQ2-5 channel activator retigabine, an FDA approved anti-convulsant. By
introducing a CF3–group at the 4-position of the benzylamine moiety, combined with a fluorine
atom at the 3-position of the aniline ring, we generated RL648_81, a new KCNQ2/3-specific
activator (EC50 190 nM) that is >15 times more potent and also more selective than retigabine
(EC50 3.3 µM). We suggest that RL648_81 is a promising clinical candidate for treating or
preventing neurological disorders associated with neuronal hyperexcitability.

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
Introduction
Epilepsy is the most common neuronal hyperexcitability disorder, affecting 1% of the world
population. This neurological condition is generally managed with sodium channel blockers or
GABA receptor agonists (Bialer et al., 2010; Bialer and White, 2010). Although there are several
drugs in clinical use with distinct mechanism of action, unfortunately approximately 30% of
patients do not respond to these agents (Brodie, 2010; Sharma et al., 2015). Thus, there is an
urgent need for drug development to broaden the treatment options.
KCNQ (or Kv7) channels play a critical role in maintaining neuronal excitability and have
recently emerged as a potential target for the prevention and treatment of epilepsy and other
hyperexcitability-related disorders, including neuropathic pain and tinnitus (Brown and
Passmore, 2009; Gribkoff, 2008; Grunnet et al., 2014; Miceli et al., 2008; Wickenden and
McNaughton-Smith, 2009; Wulff et al., 2009; Li et al., 2013). KCNQ channels are voltage-
gated, non-inactivating potassium ion channels (Brown and Adams, 1980). These channels are
open at resting membrane potentials and function as a ‘brake’ on the excitability of central and
peripheral neurons (Robbins, 2001). The KCNQ family comprises of five subunits (KCNQ1-5):
KCNQ2-5 are confined to the nervous system including brainstem and inner ear whereas
KCNQ1 is limited to the heart and peripheral epithelial and smooth muscle cells (Howard et al.,
2007). Genetic mutations in either KCNQ2 or KCNQ3 subunits are linked to benign familial
neonatal convulsions, whereas noise-induced reduction in KCNQ2/3 channel current leads to
development of tinnitus in mice (Biervert et al., 1998; Jentsch, 2000; Li et al., 2013; Li et al.,
2015). Moreover, pathological reduction in KCNQ2/3 channel activity is involved in different
classes of seizures, neuropathic pain, migraine, anxiety, attention deficient-hyperactivity
disorder, schizophrenia, mania and bipolar disease (Grunnet et al., 2014; Hansen et al., 2008;

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
Munro and Dalby-Brown, 2007). As a result, KCNQ channel activators, which lead to the
opening of these channels at more hyperpolarized potentials, have recently been employed to
treat or prevent epilepsy (Gribkoff, 2008; Miceli et al., 2008; Xiong et al., 2008).
Several Kv7 channel openers are under active development for the management of
hyperexcitability disorders (Dalby-Brown et al., 2013; Davoren et al., 2015; Grunnet et al., 2014;
Stott et al., 2014). Retigabine, which activates all KCNQ2-5 channels, is the only FDA approved
anti-convulsant KCNQ activator (Gunthorpe et al., 2012; Tatulian et al., 2001). However, recent
data showed severe side effects associated with retigabine, including urinary retention, blue skin
discoloration and retinal abnormalities (Jankovic and Ilickovic, 2013). As a result, the FDA
limited its use to patients who have not responded to alternative treatments. The undesirable side
effects are likely due to the poor selectivity of retigabine among KCNQ2-5 channels as well as
metabolic degradation products of its aniline ring. For example, retigabine activates KCNQ4 and
KCNQ5, which are not involved in the pathology of hyperexcitability-related disorders. KCNQ4
is the primary potassium channel in the smooth muscle of the bladder, where it regulates
contractility (Greenwood and Ohya, 2009; Jentsch, 2000). Activation of KCNQ4 leads to
membrane hyperpolarization and results in significantly reduced contractility, which may be the
cause for urinary retention associated with the use of retigabine. Moreover, a form of dominant
deafness arises from loss of function of KCNQ4 (Kharkovets et al., 2000), and therefore opening
of these channels may affect hearing. In addition to expression in the CNS, KCNQ4 and KCNQ5
are also found in skeletal muscle (Iannotti et al., 2013; Iannotti et al., 2010; Jentsch, 2000).
Accordingly, there is an urgent need for the development of potent and selective KCNQ2/3
channel activators, which, unlike retigabine, do not activate KCNQ4 and KCNQ5 channels.

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
To achieve this aim, we synthesized and evaluated several novel KCNQ2/3-specific channel
activators. To maximize potency, we manipulated the different chemical components or “zones”
of retigabine (Figure 1). Particularly, by introducing a CF3-group in zone 1 at the 4-position of
the benzylamine moiety, combined with a fluorine atom in zone 2 at the 3-position of the aniline
ring, we generated RL648_81, a new KCNQ2/3-specific activator that is 3 times more potent and
also more selective than SF0034, a recently described retigabine analog with increased potency
and selectivity for KCNQ2/3 channels (Kalappa et al., 2015). We propose that RL648_81 is a
promising clinical candidate for the treatment or prevention of hyperexcitability-related
neurological disorders.

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
Materials and Methods
Constructs and Chemicals. The KCNQ2, KCNQ3, KCNQ4, KCNQ5 and KCNQ2 (W236L)
constructs, as well as the PIPKI 90 plasmid used in this study have been described previously
(Soh and Tzingounis, 2010; Kalappa et al., 2015) and were generously provided by Dr.
Anastassios Tzingounis (University of Connecticut, Connecticut, USA). Buffers and salts were
purchased from Sigma-Aldrich (St. Louis, Missouri, USA). Compounds stock solutions (20 mM)
were made in DMSO. All stock solutions were stored at -20 °C. On the day of experiment, fresh
working drug concentrations were prepared from stock solutions by dissolving them in
physiological buffer solution.
Cell Culture and Transfection. Chinese hamster ovary (CHO) cells were platted on glass
coverslips in 35-mm culture dishes and were incubated and maintained at 37 °C in a humidified
incubator with 5% CO2. To get heterologous KCNQ2/3 configuration, CHO cells were
transfected with human KCNQ2 and KCNQ3 subunits cDNA in a 1:1 (0.5 g : 0.5 g) ratio with
0.5 µg of GFP plasmid using lipofectamine transfection reagent (Thermo Fisher Scientific, MA,
USA) and used for recording within 24-72 h of post transfection. For homomeric KCNQ4 and
KCNQ5 channels, 1 µg of plasmid cDNA was used. To record robust KCNQ5 currents, we
cotransfected KCNQ5 with pIRES-dsRed-PIPKI 90. PIPKI 90 increases KCNQ channel open
probability (Li et al., 2005).
Whole cell patch clamp electrophysiology. We used whole-cell patch clamp electrophysiology
used to assess effects of retigabine and synthesized compounds on KCNQ channel currents. We
conducted electrophysiology experiments at room temperature (22-25 °C). Patch pipettes of
borosilicate glass (BF150-110-10; Sutter Instrument Company, Novato, CA) were pulled to a tip
resistance of 4–6 M . Patch pipettes were filled with a solution consisting of (in mM): 132 K-

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
gluconate, 10 KCl, 4 Mg•ATP, 20 HEPES, and 1 EGTA•KOH, pH 7.2–7.3. Coverslips
containing cultured cells were placed in the recording chamber on the stage of an inverted light
microscope and superfused continuously with an external solution consisting of (in mM): 144
NaCl, 2.5 KCl, 2.25 CaCl2, 1.2 MgCl2, 10 HEPES, and 22 D-glucose, pH 7.2–7.3. Osmolarity
was adjusted to 300–305 mOsm and pH to 7.2–7.3 with NaOH. Cells were clamped at -85 mV
and currents were elicited by 1 s depolarization potentials, in 10 mV increments, from -105 to +
15 mV followed by a return step to -70 mV. Currents elicited by each voltage step were
measured and used to generate the conductance-voltage (G-V) curves as described in the figure
legends. These values are adjusted for the calculated junction potential, which was -15 mV.
Series resistance was compensated by 75%. To quantify the potency of the tested compounds, we
measured the shift in V1/2 at increasing concentrations and used a Hill equation fit to calculate
their EC50, which is the concentration of the compound that produces a half-maximal shift in
V1/2. Data were acquired through an Axopatch 200 amplifier (Molecular Devices, Sunnyvale,
CA), low-pass filtered at 2 kHz, sampled at 10 kHz using pClamp 10 data acquisition system.
Data analysis and statistics. We used the Boltzmann function to fit the conductance–voltage
curves and determine the maximal conductance (Gmax) and half-maximal activation voltage
(V1/2) of KCNQ currents, where G = Gmax/[1 + e^{-(V - V1/2)/k}] and k is the slope factor. To
calculate the dependence of the shift in V1/2 with the concentration of different KCNQ channel
activators, we measured the V1/2 of KCNQ2/3 currents in presence of various concentrations of
compounds (100 nM-30 µM). We then fitted the agonist dependence of the shift of the V1/2
obtained by different concentrations with a Hill equation, where V1/2 = V1/2 max * [activator]
n
n/([activator]n + EC50 . V1/2 is the change in V1/2 caused by the activator, EC50 is the
concentration of the KCNQ channel activator that causes 50% of the maximal effect in V1/2, n is

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
the Hill coefficient, and [activator] is the concentration of KCNQ activator. All data are
presented as mean values ± S.E.M. Statistical significance between control and test conditions
was determined using Student’s t-test (paired or unpaired) and one-way analysis of variance.
Tukey-Kramer post hoc test for multiple comparisons was performed as needed. Data analysis
and statistical tests were performed using GraphPad Prism version 6 and OriginLab 2015.

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
Results
Retigabine and its derivative SF0034 shift the voltage dependent opening of KCNQ channels
towards more hyperpolarized potentials, leading to increased K+ currents at resting membrane
potential (Kalappa et al., 2015; Tatulian et al., 2001). To generate KCNQ2/3 channel activators
with increased potency and selectivity, we partitioned retigabine’s chemical structure into three
distinct zones (Figure 1). During the first round of chemical synthesis we mainly modified zones
1 and 3, whereas in the second iteration of our structure-activity relationship (SAR) analysis, we
combined the beneficial modifications found for zones 1 and 3 with previously known beneficial
modifications on zone 2, which led to the development of SF0034 (Figure 1, Kalappa et al.,
2015). Based on the site of modification, we also categorized 1st generation compounds into
three classes. Class I compounds had substitutions at the phenyl ring in zone 1, class II
compounds had substitutions at the methylene group in zone 1, and class III compounds featured
substitutions in zone 3 (Figure 2). Our guiding principle for this round of SAR studies was to
modulate both steric and, in particular, electronic features in zone 1 by introducing fluoride,
trifluoromethyl, and, importantly, a novel pentafluorosulfanyl group, which is considered a
“super-trifluoromethyl” substituent (Alverez et al., 2015; Mo et al., 2010; Wipf et al., 2009).
These modifications led to our class I analogs NR561_29, NR561_40, NR561_45, and
NR561_50. As part of our class II analog design, we introduced substituents at the benzylic
methylene group, designed to slow down metabolic degradation (trifluoromethyl: NR579_04;
deuterium: NR561_87). Moreover, we introduced a preliminary series of heterocycle analogs of
the phenyl group in zone 1, i.e. a more electron-rich thiophene (NR579_46) and a more electron-
deficient thiazole (NR579_38). Finally, our class III design modified the steric size of the
ethylcarbamate in zone 3 by introducing an iso-propyl group (NR561_62) and the solubilizing

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
oxetane derivatives NR579_45 and NR579_36 (Skoda et al., 2014; Sprachman and Wipf, 2012;
for synthesis details and spectroscopic information on these compounds, see supplemental data).
Incorporation of highly fluorinated substituents at the phenyl ring of retigabine increases
the potency of KCNQ2/3 channel activation
To set the control conditions for assessing the potency and selectivity of the newly synthesized
compounds, we first evaluated the ability of two standard activators, retigabine and SF0034, in
potentiating KCNQ2/3 channel-mediated K+-currents under our assay conditions (Fig. 3). We
transiently expressed heterologous KCNQ2/3 channels in CHO cells and tested the effect of
increasing concentrations (100 nM to 30 µM) of retigabine and SF0034 on KCNQ2/3 currents.
We employed whole cell patch clamp electrophysiological techniques (Kalappa et al., 2015). 100
nM SF0034 increased KCNQ2/3 currents at hyperpolarized potentials, whereas 100 nM
retigabine failed to show any effect (Fig. 3A1, B1). Consistent with our previous studies (Kalappa
et al., 2015), SF0034 was approximately five times more potent than retigabine in shifting the
V1/2 of KCNQ2/3 currents (Fig. 3A-D; retigabine: EC50 3.3 ± 0.8 µM, n=4-11; SF0034: EC50
0.60 ± 0.06 µM, n=5-21, p <0.01).
Next, we tested the ability of the newly synthesized compounds to shift the V1/2 of KCNQ2/3
currents. We used different concentrations of these analogs to evaluate whether the new
substituents resulted in gain or loss of potency in activating KCNQ2/3 currents compared to
retigabine and SF0034. A concentration of 100 nM of NR561_40 (CF3-group at the 4-position of
the phenyl ring) increased the KCNQ2/3 currents at hyperpolarized potentials (Fig. 4A1).

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
Consistent with this finding, a Boltzmann fit of the G-V relationship revealed a significant shift
in the V1/2 towards hyperpolarized potentials in the presence of 100 nM or 10 µM NR561_40
without altering Gmax (Gmax:nor: control: 0.95 ± 0.015; n=9, 100 nM: 0.99 ± 0.043; n=5, 10 µM:
0.97 ± 0.108; n=4) (Fig. 4A2, A3). Similarly, we measured the shift in V1/2 in the presence of
increasing concentrations of NR561_40 (100 nM - 30 µM) and calculated the EC50 values (Fig.
4A4). NR561_40 showed a similar potency as SF0034 (Fig. 5A). Because retigabine has only
one fluorine atom at the 4-position of the phenyl ring, these results suggested that increasing the
steric bulk and electron-withdrawing properties at this position improves the potency at
KCNQ2/3 channels. NR561_29, which has an even larger and more electronegative SF5-group at
this position, demonstrated a 4-5 fold higher potency compared to retigabine, but showed a
significantly lower maximal V1/2 compared to NR561_40 and SF0034, maximal V1/2 values
were calculated from the observed shift in V1/2 at 10 M concentrations of the compound (Fig.
5B). This result suggested that a large steric size in zone 1 might limit the potency/efficacy at
KCNQ2/3 channels (Fig. 5). Furthermore, when either the smaller CF3-group or the larger SF5-
group were introduced at the 3-position of the phenyl ring, the resulting NR561_50 (Fig. 4B) and
NR561_45, both showed similar potency but a lower maximal V1/2 compared to NR561_40 and
SF0034. From this result we concluded that the exact position of the fluorinated groups at the
phenyl ring plays a critical role in determining the potency/efficacy of KCNQ activators at
KCNQ2/3 channels.
Likewise, we tested class II and class III compounds for their potency and efficacy at shifting the
V1/2 of KCNQ2/3 currents. Class II compounds NR579_38, NR579_46 and NR561_87 shifted
the V1/2 of KCNQ2/3 channels, but did not show any improvement in EC50 values relative to
retigabine. Interestingly, NR579_04 failed to show any effect on KCNQ2/3 channel currents,

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
even at concentrations as high as 10 µM, demonstrating the need for a small linker between the
phenyl ring and the benzylic amine (Figs. 4C, 5). Among class III compounds, only NR561_62
demonstrated potency that was 2-3 fold better than retigabine; the more hydrophilic substituents
were not tolerated (Fig. 4D). Taken together, our SAR results demonstrate that the position and
the steric size of fluorinated groups in zone 1 of retigabine are critical determinants of compound
potency and efficacy. Addition of CF3- and SF5-groups in positions 3 and 4 of the phenyl ring of
retigabine generates KCNQ2/3 activators with increased potency over retigabine. Incorporation
of a trifluoromethyl substituent specifically at the para-position of the benzylamine moiety
resulted in the maximal improvement in potency and efficacy at KCNQ2/3 channels, eclipsing
that of SF0034. Manipulation of the retigabine structure as probed in class II and III compounds
did not provide further improvements in KCNQ2/3 channel potency, with the exception of
NR561_62, where the ethyl group was replaced with a more lipophilic iso-propyl chain in the
carbamate moiety.
KCNQ2/3 selectivity of CF3- and SF5-containing analogs
Next, we determined the selectivity profile of class I compounds that showed increased potency
for KCNQ2/3 channels, i.e. NR561_40, NR561_50 and NR561_29. To assess the selectivity of
these analogs, we quantified their effect on homomeric KCNQ4 and KCNQ5 channels. First, we
tested NR561_50 at homomeric KCNQ4 channels. A concentration of 100 nM of NR561_50
failed to alter the V1/2 of KCNQ4 channels (Fig. 6A1). A Boltzmann fit of the G-V relationship in
presence of 100 nM or 1 µM compound did not shift the V1/2 of KCNQ4 channels (Fig. 6A2, A3).
Similarly, NR561_50 did not shift the V1/2 of KCNQ5 channels (Fig. 6B). Although 100 nM
NR561_40 and NR561_29 did not change the V1/2, at 1 µM concentration, both NR561_40 and

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
NR561_29 shifted the V1/2 of KCNQ4 currents towards more hyperpolarized potentials (Fig. 6C,
E). Furthermore, NR561_40 and NR561_29 increased KCNQ5 currents and shifted the V1/2 to
more hyperpolarized potentials at 100 nM and 1 µM (Fig. 6D, F). Taken together, NR561_50
proved to be selective for KCNQ2/3 channels whereas NR561_40 and NR561_29 were found to
be less selective. This profile suggests that incorporation of a trifluoromethyl group at the meta-
position of the benzylamine generates activators with enhanced KCNQ2/3 selectivity.
Incorporation of a fluorine substituent in zone 2 in analogs with CF3- and SF5- functions in
zone 1 generates the most potent KCNQ2/3 activator yet described
Because incorporation of a fluorine atom at the 3-position of the aniline ring of retigabine
improved the potency and selectivity at KCNQ2/3 channels (Kalappa et al., 2015), we
hypothesized that addition of fluorinated groups in zone 2 might also further improve the
potency and selectivity of class I compounds. To test our hypothesis, we selected the most potent
and efficacious compounds based on the 1st generation SAR (Fig. 5) and introduced a fluorine
atom at the 3-position of the aniline. In this fashion, RL648_81 was obtained as the fluorinated
analogue of NR561_40, RL648_73 as the fluorinated analogue of NR648_50, RL648_86 as the
fluorinated analogue of NR561_62, and RL673_02 as the fluorinated analogue of NR561_45
(Fig. 7; for synthetic and spectroscopic details, see supplemental data).
We first evaluated the effect of different concentrations of RL648_81 at KCNQ2/3-mediated
currents (Fig. 8A1). Application of 100 nM RL648_81 robustly shifted the V1/2 of KCNQ2/3
channels towards hyperpolarized potentials. Also, increased concentrations of 1 µM or 10 µM
compound showed a more robust shift in V1/2 compared to the shift caused by 1 µM or 10 µM

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
SF0034 (Fig. 8A2-A4). Evaluation of the EC50 revealed that RL648_81 is 3 times more potent
than SF0034 at shifting the voltage dependence of KCNQ2/3 channels to more hyperpolarized
potentials (Fig. 8D; EC50 0.19 ± 0.02 µM, n= 5). This shift in V1/2 was not associated by changes
in Gmax (Fig. 8A1-3; Gmax:nor: control: 0.96 ± 0.012; n=9, 100 nM: 1.01 ± 0.108; n=5, 01 µM: 1.01
± 0.012; n=4, 10 µM: 0.95 ± 0.070; n=4). To assess the selectivity of RL648_81, we tested 100
nM, 1 M and 10 µM of RL648_81 on homomeric KCNQ4 and KCNQ5 channel currents (Fig.
8B, C). RL648_81 did not shift the V1/2 of either KCNQ4 or KCNQ5, suggesting that RL648_81
is a KCNQ2/3-specific activator (Fig. 8B, C).
Next, we evaluated RL648_76, RL648_86 and RL673_02 for their potency and selectivity at
KCNQ2/3 channels. RL648_73 and RL648_86 showed 2-fold improved potency compared to
SF0034 at KCNQ2/3 channels, whereas RL673_02 did not show a significant difference from
SF0034, in agreement with the previously noted attenuation of potency with the sterically
demanding SF5 substituent (Fig. 8D left). These compounds, with the exception of the effect of
10 µM RL673_02 on KCNQ4 channels, also did not show any significant effect on either
KCNQ4 or KCNQ5 channel currents (Fig. 8E, F).
The conserved residue, tryptophan (W) in the intracellular end of the S5 helix, W236 for
KCNQ2 numbering and W265 for KCNQ3 numbering, is necessary for the enhancing effect of
retigabine (Schenzer et al., 2005). To determine whether the gating effect of RL648_81 also
requires the same W residue, we examined the influence of RL648_81 in KCNQ2 channels that
lack W236. Indeed, although RL648_81 had a strong gating effect on KCNQ2 channels, this
efficacy was abolished upon substitution of W236 to L (Fig. 9 A, B). These results suggest that
RL648_81, like retigabine and SF0034, requires W236 to exert its enhanced gating properties.

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
Taken together, our SAR investigations led to the discovery of several highly selective and
potent channel agonists, including the most potent KCNQ2/3-specific channel activator reported
to date, RL648_81, which demonstrated three times higher potency than SF0034. Moreover,
RL648_73 and RL648_86, which are two times more potent than SF0034, represent useful
secondary lead compounds. ADME/tox data for SF0034 are not available yet, but in view of the
undesired side effects associated with retigabine, we suggest that these more potent, selective
and electronically deactivated KCNQ2/3 activators may be attractive candidates for treating or
preventing hyperexcitability disorders at lower dosage and with reduced side effects.

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
Discussion
Epilepsy is a hyperexcitability disorder that affects approximately 65 million people worldwide
(Bialer and White, 2010). Despite the availability of more than 20 antiepileptic drugs, 25-40% of
epileptics are refractory to the treatment (Brodie, 2010). Moreover, even when currently
available drugs are helpful, they are not without adverse effects (Bialer et al., 2010; Thurman et
al., 2011). Therefore, there is an urgent need for novel drugs with improved therapeutic index
characterized by increased potency and reduced toxicity.
Retigabine, which stabilizes the open state of KCNQ2-5 channels and shifts the voltage-
dependence towards more hyperpolarized potentials (Wuttke et al., 2005), is an FDA approved
KCNQ channel activator that serves as an add-on for the treatment of resistant partial-onset
seizures (Gunthorpe et al., 2012). However, recently identified side effects of retigabine,
including retinal abnormalities, skin discoloration and urinary retention, are significantly limiting
its clinical use.
Here, we disclose the preparation and biological analysis of RL648_81 and related analogs, and
report that this compound is the most potent KCNQ2/3-specific channel activator known to date.
Due to its higher, yet selective agonism on KCNQ2/3 channels, RL648_81 is expected to be
more effective at a lower dose than retigabine and its analog, SF0034, in preventing seizures. In
addition, RL648_81 does not activate either KCNQ4 or KCNQ5. Because of its enhanced
specificity over retigabine, RL648_81 is expected to have fewer side effects.
Although the mechanism by which retigabine-mediated toxicity influences skin and retina
remains poorly understood, one hypothesis is that UV radiation may cause photodegradation and
oxidation of retigabine’s aniline ring, which may lead to the formation of colored deposits in skin
and eyes. The incorporation of electron-withdrawing highly fluorinated substituents significantly

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
reduces the highest occupied molecular orbital energy of RL648_81 (-8.33 eV) vs. retigabine (-
8.06 eV), and thus should render the former compound less prone to formation of reactive
metabolites (Kawai et al., 2007). The trifluoromethylated RL648_81 is also expected to be more
resistant to photodegradation and therefore less toxic to the eye and to the skin (Dow et al.;
2006). In particular, the electronic properties of RL648_81 (Fig. S2C) compared to retigabine
(Fig. S2A) suggest that its aniline -system as well as all three attached nitrogen atoms are
considerably more electron-deficient. Furthermore, the CF3-substituent is a stronger deactivator
of the benzylamine -system, and also presents a greater steric barrier to potential cytochrome
P450-induced arene hydroxylation. The electrostatic potential map for RL673_02 (Fig. S2B) and
RL648_73 (Fig. S2D) is closely related to that of RL648_81, with the SF5-substituent in
RL673_02 providing even greater steric and electronic deactivation of the compound.
The conserved residue W236 is necessary for the enhancing effect of retigabine, SF0034, and
RL648_81 on the gating properties of KCNQ2 channels (Schenzer et al., 2005). Whereas it was
thought that the critical property of W was its hydrophobicity (Wuttke et al., 2005), recent
studies revealed that the ability of W to form H-bonds with the carbonyl/carbamate oxygen atom
present in retigabine makes this contact critical (Kim et al., 2015). These studies suggest that the
strength of the H-bond between retigabine and W236 determines the potency of retigabine. In
part, the improved potency of SF0034 and RL648_81 is probably due to their ability to form
stronger H-bonds with W236 than retigabine. However, besides W265, other residues are
important for the gating effects of retigabine, such as L272, L314, and Leu-338 (KCNQ3
numbering) and G301/G340 (KCNQ2/KCNQ3 numbering) (Schenzer et al., 2005; Wuttke et al.,
2005; Lange et al., 2009). Thus based on our SAR studies, we also suggest that the activity of
analogs can be further modulated by hydrophobic interactions at the benzylamine moiety, which

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
the carbamate-W236 interaction positions into the vicinity of L272, L314, and L338 (Lange et
al., 2008), and that the more lipophilic RL648_81 (logP=3.4) is a better fit than retigabine
(logP=2.5) for this hydrophobic pocket at the pore-forming S5 inner loop and S6 helix domains,
located near the intracellular voltage-operated gate of KCNQ2–5 channels.
In conclusion, our investigation of fluorinated substituents on retigabine, in particular the effect
of highly fluorinated polar and lipophilic CF3- and SF5-groups in the benzylamine portion of the
KCNQ2-5 channel activator, yielded a series of submicromolar affinity activators with exquisite
selectivity for KCNQ2/3. We propose that the combination of increased potency and selectivity
of RL648_81, as well as its structural features and modified electrostatic properties, may provide
a solution to the problems associated with the undesirable side effects associated with retigabine.

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
Acknowledgements
We thank Dr. Anastassios Tzingounis for providing KCNQ2/3 mutants and Karen Hartnet for
help with cell culture and plasmid preparation.

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
Authorship Contributions
Participated in research design: MK, PW, EA and TT
Conducted experiments: MK, NR, PW and RL
Contributed new reagents or analytic tools: RL, MK, and EA
Performed data analysis: MK, TT, RL, NR and PW
Wrote or contributed to the writing of the manuscript: MK, EA, PW, RL and TT

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
References
Alverez C, Arkin MR, Bulfer SL, Colombo R, Kovaliov M, LaPorte MG, Lim C, Liang M,
Moore WJ, Neitz RJ, Yan Y, Yue Z, Huryn DM and Wipf P (2015) Structure–Activity
Study of Bioisosteric Trifluoromethyl and Pentafluorosulfanyl Indole Inhibitors of the
AAA ATPase p97. ACS Medicinal Chemistry Letters 6(12): 1225-1230.
Bialer M, Johannessen SI, Levy RH, Perucca E, Tomson T and White HS (2010) Progress report
on new antiepileptic drugs: a summary of the Tenth Eilat Conference (EILAT X).
Epilepsy research 92(2-3): 89-124.
Bialer M and White HS (2010) Key factors in the discovery and development of new
antiepileptic drugs. Nature reviews Drug discovery 9(1): 68-82.
Biervert C, Schroeder BC, Kubisch C, Berkovic SF, Propping P, Jentsch TJ and Steinlein OK
(1998) A potassium channel mutation in neonatal human epilepsy. Science (New York,
NY) 279(5349): 403-406.
Brodie MJ (2010) Antiepileptic drug therapy the story so far. Seizure 19(10): 650-655.
Brown DA and Adams PR (1980) Muscarinic suppression of a novel voltage-sensitive K+
current in a vertebrate neurone. Nature 283(5748): 673-676.
Brown DA and Passmore GM (2009) Neural KCNQ (Kv7) channels. British journal of
pharmacology 156(8): 1185-1195.
Dalby-Brown W, Jessen C, Hougaard C, Jensen ML, Jacobsen TA, Nielsen KS, Erichsen HK,
Grunnet M, Ahring PK, Christophersen P, Strobaek D and Jorgensen S (2013)

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
Characterization of a novel high-potency positive modulator of K(v)7 channels.
European journal of pharmacology 709(1-3): 52-63.
Davoren JE, Claffey MM, Snow SL, Reese MR, Arora G, Butler CR, Boscoe BP, Chenard L,
DeNinno SL, Drozda SE, Duplantier AJ, Moine L, Rogers BN, Rong S, Schuyten K,
Wright AS, Zhang L, Serpa KA, Weber ML, Stolyar P, Whisman TL, Baker K, Tse K,
Clark AJ, Rong H, Mather RJ and Lowe JA, 3rd (2015) Discovery of a novel Kv7
channel opener as a treatment for epilepsy. Bioorganic & medicinal chemistry letters
25(21): 4941-4944.
Dow GS, Heady TN, Bhattacharjee AK, Caridha D, Gerena L, Gettayacamin M, Lanteri CA,
Obaldia N, Roncal N, Shearer T, Smith PL, Tungtaeng A, Wolf L, Cabezas M, Yourick
D, Smith,KS (2006) "Utility of alkylaminoquinolinyl methanols as new antimalarial
drugs." Antimicrob. Agents Chemother 50, 4132-4143
Greenwood IA and Ohya S (2009) New tricks for old dogs: KCNQ expression and role in
smooth muscle. British journal of pharmacology 156(8): 1196-1203.
Gribkoff VK (2008) The therapeutic potential of neuronal K V 7 (KCNQ) channel modulators:
an update. Expert opinion on therapeutic targets 12(5): 565-581.
Grunnet M, Strobaek D, Hougaard C and Christophersen P (2014) Kv7 channels as targets for
anti-epileptic and psychiatric drug-development. European journal of pharmacology 726:
133-137.
Gunthorpe MJ, Large CH and Sankar R (2012) The mechanism of action of retigabine
(ezogabine), a first-in-class K+ channel opener for the treatment of epilepsy. Epilepsia
53(3): 412-424.

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
Hansen HH, Waroux O, Seutin V, Jentsch TJ, Aznar S and Mikkelsen JD (2008) Kv7 channels:
interaction with dopaminergic and serotonergic neurotransmission in the CNS. The
Journal of physiology 586(7): 1823-1832.
Howard RJ, Clark KA, Holton JM and Minor DL, Jr. (2007) Structural insight into KCNQ (Kv7)
channel assembly and channelopathy. Neuron 53(5): 663-675.
Iannotti FA, Barrese V, Formisano L, Miceli F and Taglialatela M (2013) Specification of
skeletal muscle differentiation by repressor element-1 silencing transcription factor
(REST)-regulated Kv7.4 potassium channels. Molecular biology of the cell 24(3): 274-
284.
Iannotti FA, Panza E, Barrese V, Viggiano D, Soldovieri MV and Taglialatela M (2010)
Expression, localization, and pharmacological role of Kv7 potassium channels in skeletal
muscle proliferation, differentiation, and survival after myotoxic insults. The Journal of
pharmacology and experimental therapeutics 332(3): 811-820.
Jankovic S and Ilickovic I (2013) The preclinical discovery and development of ezogabine for
the treatment of epilepsy. Expert opinion on drug discovery 8(11): 1429-1437.
Jentsch TJ (2000) Neuronal KCNQ potassium channels: physiology and role in disease. Nature
reviews Neuroscience 1(1): 21-30.
Kalappa BI, Soh H, Duignan KM, Furuya T, Edwards S, Tzingounis AV and Tzounopoulos T
(2015) Potent KCNQ2/3-specific channel activator suppresses in vivo epileptic activity
and prevents the development of tinnitus. The Journal of neuroscience : the official
journal of the Society for Neuroscience 35(23): 8829-8842.

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
Kawai, M, Sakurada I, Morita A, Iwamuro Y, Ando K, Omura H, Sakakibara S, Masuda T,
Koike H, Honma T, Hattori K, Takashima T, Mizuno K, Mizutani M, Kawamura M
(2007) "Structure-activity relationship study of novel NR2B-selective antagonists with
arylamides to avoid reactive metabolites formation." Bioorg. Med. Chem. Lett. 17, 5537-
5542.
Kharkovets T, Hardelin JP, Safieddine S, Schweizer M, El-Amraoui A, Petit C and Jentsch TJ
(2000) KCNQ4, a K+ channel mutated in a form of dominant deafness, is expressed in
the inner ear and the central auditory pathway. Proceedings of the National Academy of
Sciences of the United States of America 97(8): 4333-4338.
Kim RY, Yau MC, Galpin JD, Seebohm G, Ahern CA, Pless SA and Kurata HT (2015) Atomic
basis for therapeutic activation of neuronal potassium channels. Nature communications
6: 8116.
Lange W, Geißendorfer J, Schenzer A, Grotzinger J, Seebohm G,Friedrich T, and Michael
Schwake M (2008) Refinement of the Binding Site and Mode of Action of the
Anticonvulsant Retigabine on KCNQ K_ Channels. Molecular pharmacology 75(2): 272-
80.
Li S, Choi V and Tzounopoulos T (2013) Pathogenic plasticity of Kv7.2/3 channel activity is
essential for the induction of tinnitus. Proceedings of the National Academy of Sciences
of the United States of America 110(24): 9980-9985.
Li S, Kalappa BI and Tzounopoulos T (2015) Noise-induced plasticity of KCNQ2/3 and HCN
channels underlies vulnerability and resilience to tinnitus. eLife 4.

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
Li, Y., Gamper, N., Hilgemann, D.W. & Shapiro, M.S. (2005) Regulation of Kv7 (KCNQ) K+
channel open probability by phosphatidylinositol 4,5-bisphosphate. The Journal of
neuroscience : the official journal of the Society for Neuroscience, 25, 9825-9835.
Miceli F, Soldovieri MV, Martire M and Taglialatela M (2008) Molecular pharmacology and
therapeutic potential of neuronal Kv7-modulating drugs. Current opinion in
pharmacology 8(1): 65-74.
Mo T, Mi X, Milner EE, Dow GS and Wipf P (2010) Synthesis of an 8-pentafluorosulfanyl
analog of the antimalarial agent mefloquine. Tetrahedron Letters 51(39): 5137-5140.
Munro G and Dalby-Brown W (2007) Kv7 (KCNQ) channel modulators and neuropathic pain.
Journal of medicinal chemistry 50(11): 2576-2582.
Robbins J (2001) KCNQ potassium channels: physiology, pathophysiology, and pharmacology.
Pharmacology & therapeutics 90(1): 1-19.
Schenzer A, Friedrich T, Pusch M, Saftig P, Jentsch TJ, Grotzinger J and Schwake M (2005)
Molecular determinants of KCNQ (Kv7) K+ channel sensitivity to the anticonvulsant
retigabine. The Journal of neuroscience : the official journal of the Society for
Neuroscience 25(20): 5051-5060.
Sharma AK, Rani E, Waheed A and Rajput SK (2015) Pharmacoresistant Epilepsy: A Current
Update on Non-Conventional Pharmacological and Non-Pharmacological Interventions.
Journal of epilepsy research 5(1): 1-8.

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
Skoda EM, Sacher JR, Kazancioglu MZ, Saha J and Wipf P (2014) An uncharged oxetanyl
sulfoxide as a covalent modifier for improving aqueous solubility. ACS Med Chem Lett
5(8): 900-904.
Soh H and Tzingounis AV (2010) The specific slow afterhyperpolarization inhibitor UCL2077 is
a subtype-selective blocker of the epilepsy associated KCNQ channels. Molecular
pharmacology 78(6): 1088-1095.
Sprachman MM and Wipf P (2012) A bifunctional dimethylsulfoxide substitute enhances the
aqueous solubility of small organic molecules. Assay and drug development technologies
10(3): 269-277.
Stott JB, Jepps TA and Greenwood IA (2014) K(V)7 potassium channels: a new therapeutic
target in smooth muscle disorders. Drug discovery today 19(4): 413-424.
Tatulian L, Delmas P, Abogadie FC and Brown DA (2001) Activation of expressed KCNQ
potassium currents and native neuronal M-type potassium currents by the anti-convulsant
drug retigabine. The Journal of neuroscience : the official journal of the Society for
Neuroscience 21(15): 5535-5545.
Thurman DJ, Beghi E, Begley CE, Berg AT, Buchhalter JR, Ding D, Hesdorffer DC, Hauser
WA, Kazis L, Kobau R, Kroner B, Labiner D, Liow K, Logroscino G, Medina MT,
Newton CR, Parko K, Paschal A, Preux PM, Sander JW, Selassie A, Theodore W,
Tomson T and Wiebe S (2011) Standards for epidemiologic studies and surveillance of
epilepsy. Epilepsia 52 Suppl 7: 2-26.
Wickenden AD and McNaughton-Smith G (2009) Kv7 channels as targets for the treatment of
pain. Current pharmaceutical design 15(15): 1773-1798.

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
Wipf P, Mo T, Geib SJ, Caridha D, Dow GS, Gerena L, Roncal N and Milner EE (2009)
Synthesis and biological evaluation of the first pentafluorosulfanyl analogs of
mefloquine. Organic & Biomolecular Chemistry 7(20): 4163-4165.
Wulff H, Castle NA and Pardo LA (2009) Voltage-gated potassium channels as therapeutic
targets. Nature reviews Drug discovery 8(12): 982-1001.
Wuttke TV, Seebohm G, Bail S, Maljevic S and Lerche H (2005) The new anticonvulsant
retigabine favors voltage-dependent opening of the Kv7.2 (KCNQ2) channel by binding
to its activation gate. Molecular pharmacology 67(4): 1009-1017.
Xiong Q, Gao Z, Wang W and Li M (2008) Activation of Kv7 (KCNQ) voltage-gated potassium
channels by synthetic compounds. Trends in pharmacological sciences 29(2): 99-107.

Molecular Pharmacology Fast Forward. Published on March 22, 2016 as DOI: 10.1124/mol.115.103200
This article has not been copyedited and formatted. The final version may differ from this version.
Footnotes
This work was supported by the Department of Defense, the Joint Warfighter Medical Research
Program [W81XWH-14-1-0117] to T.T. and P.W..