Fluorescently labeled analogues of dofetilide as high-affinity fluorescence polarization ligands for the human ether-a-go-go-related gene (hERG) channel

Article abstract of DOI:10.1021/jm0700565

Novel fluorescent derivatives of dofetilide (1) have been synthesized. Analogues that feature a fluorescent probe attached through an aliphatic spacer to the central tertiary nitrogen of 1 have high affinity for the hERG channel, and affinity is dependent on both linker length and pendent dye. These variables have been optimized to generate Cy3B derivative 10e, which has hERG channel affinity equivalent to that of dofetilide. When bound to cell membranes expressing the hERG channel, 10e shows a robust increase in fluorescence polarization (FP) signal. In a FP binding assay using 10e as tracer ligand, K_i values for several known hERG channel blockers were measured and excellent agreement with the literature K_i values was observed over an affinity range of 2 nM to 3 μM. 10e blocks hERG channel current in electrophysiological patch clamp experiments, and computational docking experiments predict that the dofetilide core of 10e binds hERG channel in a conformation similar to that previously predicted for 1. These analogues enable high-throughput hERG channel binding assays that are rapid, economical, and predictive of test compounds' potential for prolonged QT liabilities.

Full text of DOI:10.1021/jm0700565

© Copyright 2007 by the American Chemical Society
Volume 50, Number 13
June 28, 2007
Articles
Fluorescently Labeled Analogues of Dofetilide as High-Affinity Fluorescence Polarization
Ligands for the Human Ether-a-go-go-Related Gene (hERG) Channel
David H. Singleton,† Helen Boyd,# Jill V. Steidl-Nichols,‡ Matt Deacon,§ Marcel J. de Groot,| David Price,|
David O. Nettleton,† Nora K. Wallace,† Matthew D. Troutman,∞ Christine Williams,# and James G. Boyd*,†
Exploratory Medicinal Sciences and ADME Technology Group, Pfizer Global Research and DeVelopment, Groton, Connecticut, HTS Center of
Emphasis, Primary Pharmacology Group, and Department of Medicinal Chemistry, Pfizer Global Research and DeVelopment, Sandwich, U.K.,
and Global Safety Pharmacology, Pfizer Global Research and DeVelopment, La Jolla, California 92037
ReceiVed January 15, 2007
Novel fluorescent derivatives of dofetilide (1) have been synthesized. Analogues that feature a fluorescent
probe attached through an aliphatic spacer to the central tertiary nitrogen of 1 have high affinity for the
hERG channel, and affinity is dependent on both linker length and pendent dye. These variables have been
optimized to generate Cy3B derivative 10e, which has hERG channel affinity equivalent to that of dofetilide.
When bound to cell membranes expressing the hERG channel, 10e shows a robust increase in fluorescence
polarization (FP) signal. In a FP binding assay using 10e as tracer ligand, Ki values for several known
hERG channel blockers were measured and excellent agreement with the literature Ki values was observed
over an affinity range of 2 nM to 3 µM. 10e blocks hERG channel current in electrophysiological patch
clamp experiments, and computational docking experiments predict that the dofetilide core of 10e binds
hERG channel in a conformation similar to that previously predicted for 1. These analogues enable high-
throughput hERG channel binding assays that are rapid, economical, and predictive of test compounds’
potential for prolonged QT liabilities.
Introduction
ventricular repolarization during a single cardiac cycle.2 LQT
can have adverse effects on cardiac rhythm and can lead to a
potentially lethal arrhythmia known as torsades de points. Since
the cloning and functional assignment of hERG, it has also been
The human ether-a-go-go-related gene (hERGa) encodes for
a potassium channel that is expressed in the myocardium and
whose activity is critical for repolarization of cardiac tissue
during the heart beat cycle. The hERG channel is the molecular
target of class III antiarrhythmic drugs,1 and functional mutations
in hERG are associated with a phenomenon known as QT
prolongation (LQT), a lengthening of the interval required for
a Abbreviations: ACN, acetonitrile; Boc, tert-butyloxycarbonyl; DIEA,
diisopropylethylamine; DMF, N,N-dimethylformamide; ESMS, electrospray
mass spectrometry; FP, fluorescence polarization; g, gravity; HATU,
N-[(dimethylamino)-1H-1,2,3-triazolo[4,5-b]pyridin-1-ylmethylene]hexaflu-
orophosphate N-oxide; HBTU, 2-(1H-benzotriazol-1-yl)-1,1,3,3-tetramethy-
luronium hexafluorophosphate; HEK293, human embryonic kidney (cells);
hERG, human ether-a-go-go-related gene; HILIC, hydrophilic interaction
chromatography; HOAt, 1-hydroxy-7-azabenzotriazole; HPLC, high-pres-
sure liquid chromatography; HRMS, high-resolution mass spectrometry;
HT, high throughput; HTS, high-throughput screening; LQT, prolonged QT
interval; MR121, 1-(3-carboxypropyl)-11-ethyl-1,2,3,4,8,9,10,11-octahy-
drodipyrido[3,2-b:2′,3′-i]phenoxazin-13-ium inner salt; MSA, methane-
sulfonanilide; NBD, 7-nitrobenzodiazole; NHS, N-hydroxysuccinimide;
NMP, N-methylpyrrolidinone; ; SAR, structure-activity relationship; SPA,
scintillation proximity assay; TFA, trifluoroacetic acid; TMR, tetrameth-
ylrhodamine.
* To whom correspondence should be addressed. Address: Exploratory
Medicinal Sciences, Pfizer Global Research and Development, Mail Stop
4014, Eastern Point Road, Groton, CT 06340. Phone: 860-441-1580. Fax:
860-441-3858. E-mail: james_g_boyd@pfizer.com.
† Exploratory Medicinal Sciences, Pfizer, CT.
# HTS Center of Emphasis, Pfizer, U.K.
‡ Global Safety Pharmacology, Pfizer, CA.
§ Primary Pharmacology Group, Pfizer, U.K.
| Department of Medicinal Chemistry, Pfizer, U.K.
∞ ADME Technology Group, Pfizer, CT.
10.1021/jm0700565 CCC: $37.00 © 2007 American Chemical Society
Published on Web 05/31/2007

2932 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 13
Singleton et al.
Scheme 1. Synthesis of Fluorescent Analogues of Dofetilide by
shown that a number of marketed therapeutic drugs have off-
target affinity for the hERG channel, and the observation of
LQT in patients using these drugs has been assigned to this
activity. Several launched drugs with LQT side effects have
been recalled.3
Direct Acylation or Arylation of the Central Core Nitrogena
Appropriately, regulatory agencies now look scrupulously for
evidence of QT prolongation during the drug application process.
In order to avoid bringing a compound with LQT side effects
to clinical trials, the pharmaceutical industry has devised an array
of in vitro assays that are predictive of this effect. The
electrophysiological patch clamp assay has emerged as an
industry standard for measuring functional hERG channel
blockage by small-molecule drug candidates or leads.4,5 How-
ever, in its current format, the patch clamp assay is not suitable
for rapid screening of large numbers of compounds.6 Higher
throughput patch clamp assays have been developed recently,7,8
but these likely lack the capacity needed to fully support high-
speed medicinal chemistry programs. Because affinity for the
hERG channel correlates well with channel current blockage,
drug discovery companies have developed radioligand-based
hERG channel competition binding assays that are predictive,
robust, and amenable to high throughput.9-12 It is important to
note that these assays reflect the binding mode of the particular
tracer ligand used. Compounds that do not compete with the
ligand may still block channel current but can be registered as
false negatives. Conversely, ligands that do compete with the
tracer but do not block the channel may be registered as false
positives. Thus, ligand-based methods are not ideal surrogates
for patch clamp experiments. They are used most effectively
for HTS hit assessment and in medicinal chemistry lead
development programs where speed and high throughput are
required.
aReagents and conditions: (a) NBD chloride (1.0 equiv), DMF, 2 h,
41%; (b) 5- and 6-carboxyfluorescein (mixed isomers) or 5-carboxy TMR
(1.0 equiv), HATU (1.0 equiv), HOAt (1.0 equiv), 10% DIEA in DMF, 4
h, 5-40%; (c) MR121 (0.6 equiv), HATU (1.0 equiv), HOAt (1.0 equiv),
DIEA in DMF, 4 h, 35%.
Table 1. SPA Derived hERG Channel Affinities of Fluorescent
Dofetilide Analoguesa
Radioligand-based high-throughput assays suffer from several
undesirable burdens. These include regulatory agency docu-
mentation, radioactive waste-disposal, special safety training,
and environmental monitoring. Fluorescence-based methods,13
particularly fluorescence polarization (FP) methods, have emerged
as versatile screening alternatives with greatly reduced envi-
ronmental and cost impacts compared with radioligand assays.
Additionally, FP assays are easily miniaturized, do not require
laborious filtration steps, and can be executed in a homogeneous
(mix-and-read) fashion, thus leading to assays that are truly high-
throughput. While there are clear incentives to move from
radioligand assays to FP assays, a major challenge is the need
to develop a fluorescent ligand that has high affinity for the
receptor of interest. Usually this is accomplished through
synthetic modification of a high affinity ligand so that it can
accommodate covalent coupling to a fluorescent dye without
negatively perturbing the binding affinity of the parent ligand
(for recent examples, see refs 14-17). This is an empirical
process, but the likelihood of success is increased by use of
existing SAR, binding models, and crystallographic data when
available. Herein, we describe the development of a family of
fluorescent ligands that bind the hERG channel with high affinity
and, when bound, give a robust FP signal.
compd
Ki (nM)b
standard deviation (nM)
n
1
3
4
5
6.4
0.39
980
>1000
3.6
7300
160
18
2.7
0.08
280
4
1c
2
2
2
2
2
2
4
2
3
4
4
1
6
0.4
10a
10b
10c
10d
10e
12a
12b
13
70
2.0
2.0
0.8
4.4
4.3
15
13
7.0
19
10
15
89
14
15
51
>1000
a Ki values determined from competition binding experiments in an SPA
format using HEK293 cell membranes expressing hERG protein and
[3H]dofetilide tracer.40 b Reported Ki values are the arithmetic mean of
multiple experiments performed in triplicate. c A second experiment showed
a similar degree of inhibition but did not allow for a calculated Ki.
diazole (NBD chloride) to produce fluorescent product 3
(Scheme 1). When tested in a standard scintillation proximity
assay (SPA) that uses 3H dofetilide as a radioligand, 3 exhibited
a 15-fold increase in affinity for hERG channel relative to 1
(Table 1). While 3 served as an excellent proof-of-concept
compound, it did not show a useful FP signal upon binding.
Hence, two other dyes commonly used in FP assays were
examined. Activation of 5- and 6-carboxyfluorescein and
5-carboxytetramethylrhodamine (TMR) with HATU (N-[(dim-
ethylamino)-1H-1,2,3-triazolo[4,5-b]pyridin-1-ylmethylene]-
hexafluorophosphate N-oxide) followed by reaction with 2
produced fluorescent derivatives 4 and 5, respectively. These
analogues showed poor affinity for the hERG channel in the
Ligand Design, Synthesis, and SAR
The antiarrhythmic agent dofetilide18 (1) was chosen as the
scaffold for fluorescent analogue development. Dofetilide is a
member of the methanesulfonanilide (MSA) class of antiar-
rhythmic drugs and is marketed under the brand name Tikoyan.
It binds to the hERG channel with high affinity19 and potently
blocks channel current in a patch clamp assay.4,19 N-Desmeth-
yldofetilide20 (2) was alkylated with 4-chloro-7-nitrobenzoxa-

Fluorescence Polarization Ligands for hERG Channel
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 13 2933
Scheme 2. Synthesis of Fluorescent Dofetilide Analogues with Variable-Length Spacersa
a Reagents and conditions: (a) Boc amino aldehyde (2.0 equiv), NaBH(OAc)3 (2.0 equiv), 4:1 DCE/THF, room temp, 16 h, 47%; (b) TFA, room temp,
5 min, >95%; (c) bromoalkylphthalimide (2.4 equiv), Na2CO3, ACN, reflux, 4 h, 53%; (d) hydrazine (6 equiv), methanol, reflux, 2 h, 58%; (e) Cy3B NHS
ester (1.0 equiv), 5% DIEA in DMF, room temp, 15 min, 86%; (f) Boc amino acid (0.75 equiv), HBTU (0.75 equiv), 10% DIEA in NMP, room temp, 30
min, 57%; (g) fluorescein NHS ester or TMR NHS ester (1.0 equiv), 5% DIEA in DMF, room temp, 15 min; (h) (i) TFA, 5 min, (ii) Cy3B NHS ester (1.1
equiv), 20% DIEA in DMF, room temp, 1 h, 29%.
SPA. Next, derivatives that feature spacer groups between the
dye and the central nitrogen of 2 were synthesized. The red-
shifted dye MR121 (1-(3-carboxypropyl)-11-ethyl-1,2,3,4,8,9,-
10,11-octahydrodipyrido[3,2-b:2′,3′-i]phenoxazin-13-ium inner
salt) includes a butyric acid spacer, which following activation
with HATU acylates the secondary nitrogen of 2. The resulting
compound 6 has an hERG channel affinity close to that of
dofetilide.
[1,2-a:1′,2′-a′]diindol-5-ium, inner salt) conjugates 10a-e were
prepared by mixing amines 9a-e each with an equivalent of
Cy3B NHS ester in DMF/DIEA, followed by purification with
reverse-phase HPLC. The N-ethylamide of Cy3B (15) was
prepared as a control compound in the same manner. Conjugates
12a,b were prepared similarly by first deblocking 11a,b with
TFA and reaction with Cy3B NHS ester. Fluorescein and TMR
conjugates 13 and 14 were prepared by reaction of 9e with
fluorescein and TMR NHS esters, respectively.
Many of the commercially available dyes useful for FP assays
are sold as N-hydroxysuccinimide (NHS) esters. In order to
make conjugates with these dyes, hydrocarbon spacers with
protected terminal amines were prepared. As shown in Scheme
2, 2 undergoes reductive alkylation with C2, C3, and C6 tert-
butyloxycarbonyl (Boc) protected amino aldehydes to give Boc
protected amines 7a, 7b, and 7e in good yield. Removal of the
Boc groups of 7a,b,e with trifluoroacetic acid (TFA) provided
compounds 9a,b,e as their TFA salts. The C4 and C5 Boc
protected amino aldehydes reacted poorly with 2, presumably
because of their tendency to form cyclic hemiaminals. The C4
and C5 protected amines were obtained via alkylation of 2 with
phthalimide protected aminoalkyl bromides to give phthalimides
8a,b. Deprotection with hydrazine provided amines 9c,d as free
bases. The synthesis of conjugates of 2 with longer amide-
containing linkers was accomplished by reaction of 9e with
either N-Boc-3-aminopropionic acid or N-Boc-6-aminocaproic
acid using HBTU (2-(1H-benzotriazol-1-yl)-1,1,3,3-tetramethy-
luronium hexafluorophosphate) activation to give Boc protected
conjugates 11a,b, respectively.
Compound purity of fluorescent derivatives of 2 was assayed
by two analytical HPLC methods as described in Supporting
Information. All compounds exceeded 95% purity except 4
(88% as a 4:3 mixture of regioisomers) and 10a (82%). Yields
were calculated on the basis of UV absorbance at λmax in neutral
aqueous buffers. Samples were then aliquoted, lyophilized as
TFA salts, and redissolved in water for testing.
Table 1 displays the SPA measured binding affinities of all
fluorescent derivatives of 2, and several trends are worthy of
note. First, for conjugates 3-5, there is a marked affinity
dependence on the structure of the attached dye. NBD derivative
3 experiences a 15-fold increase in affinity over dofetilide, while
fluorescein and TMR derivatives 4 and 5 suffer more than a
150-fold decrease in affinity. Second, for Cy3B conjugates
10a-e and 12a,b, a consistent affinity dependence on spacer
length was observed. Moving from a two-methylene spacer for
compound 10a to a four-methylene spacer for compound 10c
resulted in a 400-fold increase in affinity. Increasing the spacer
length from four to six methylenes gave a modest 3-fold increase
in affinity to the maximum observed for conjugate 10e, which
is equipotent to dofetilide. Additional linker length via amide-
Cy3B (2-(carboxymethyl)-6,7,9,10,16,18-hexahydro-16,16,-
18,18-tetramethyl-14-sulfopyrano[3′′,2′′:3,4;5′′,6′′:3′,4′]dipyrido-

2934 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 13
Singleton et al.
Table 2. SPA and FP-Derived Ki Values for Literature Standard hERG
Channel Inhibitorsa
compd
FP Ki (nM)b SPA Ki (nM)c SPA/FP literature (nM)d
astemizole
dofetilide
E-4031
bepridil
mibefradil
clozapine
2.9(0.9)
7.7(4.4)e
18.7(3.8)
184(25)
190(32)
4060(840)
12.8(3.0)
6.4(2.7)f
19.4(3.1)
291(63)
353(97)
3080(900)
4.4
2
10
30
0.91
1.04
1.6
1.85
0.77
170-450
660-890
1200-5800
a Comparison of Ki values determined from competition binding experi-
ments in an SPA or FP format. Reported Ki values are the arithmetic mean
of multiple experiments performed in triplicate. Standard deviations are in
parentheses. b FP assay using 10e as fluorescent tracer. c SPA as in Table
1. d From Diaz et al.19 Single Ki values from measurements at 10 mM K+.
Range reflects Ki values measured at 60 and 5 mM K+. e The Ki values
reported for dofetilide are the arithmetic mean of experiments run regularly
over many months (n ) 610). f From Table 1.
Figure 1. Change in FP upon ligand binding to hERG channel. Shown
are FP signals for Cy3B ligands 10e (3.0 nM), 12a (2.0 nM), and 12b
(2.0 nM) and MR121 ligand 6 (5.0 nM) absent membranes (white bars),
with hERG transfected HEK293 membranes (black bars) or with control
HEK293 membranes transfected with irrelevant control (serotonin
transporter, gray bars). Error bars represent standard deviations from
experiments run in triplicate.
Table 3. Patch Clamp IC50 and τ Values and Apparent Permeability
Dataa
patch
clamp SPA
IC50 Ki
e
e
Papp
absorptive
Papp
secretory
τ for onset
compd (nM)b (nM)c of block (s)d (×10-6 cm/s) (×10-6 cm/s)e
1
3
10e
9.2
<3
310
6.4
0.39
7.0
182(30)
575*(42)
316*(35)
15.7(2.6)
3.7(0.9)
5.4(0.2)
14.3(2.3)
18.0(2.6)
4.8(0.9)
a Comparison of patch clamp IC50 and τ values to apparent cell
permeability as determined in a Caco-2 assay. b IC50 values determined from
data shown in Figure 3D. c SPA data from Table 1. d Time constants (and
standard error values) derived from time course of onset of current block
at concentrations that produced 80% inhibition of the hERG channel current
(25 nM 1, 3.0 nM 3, and 1000 nM 10e). Asterisks denote a significant
difference from 1 (see Experimental Section for description of analysis).
e Caco-2 permeability data determined in the absorptive and secretory
transport directions in the presence of 5 µM cyclosporin A. Values represent
the mean (and standard deviations in parentheses) of two separate
experiments in triplicate.
Figure 2. FP competition binding experiment. The FP response for
Cy3B ligand 10e as a function of [1] is shown. Calculated Ki for 1 is
8.0 nM, and the Hill slope is 1.2. Error bars represent standard
deviations for experiments run in triplicate.
trations of dofetilide, 10e shows the expected sigmoidal FP
response curve with a Hill coefficient of 1.2 (Figure 2). The
22
containing spacers (12a,b) did not result in further increases in
affinity. As expected, no binding was detected for negative
control Cy3B compound 15. Third, even with optimal spacer
length, there appears to be an affinity dependence on conjugate
dye. Exchanging the Cy3B dye of 10e for fluorescein (13) did
not significantly alter binding, but the structurally similar dye
TMR (14) caused a 12-fold affinity reduction.
calculated dofetilide Ki from the experiment is 8.0 nM. The
arithmetic mean Ki for multiple experiments is 7.7 nM (n )
610, average Hill slope of 1.04), essentially equivalent to the
SPA derived Ki value of 6.4 nM. In order to judge the per-
formance of 10e as a tracer FP ligand, it was tested against
other known hERG channel binding molecules with affinities
in the low nanomolar to low micromolar range. Table 2 shows
the FP and SPA derived Ki values for astemizole, E-4031,
bepridil, mibefradil, and clozapine and compares them to values
reported in the literature. Four of these hERG channel inhibitors
showed excellent agreement between the two assay formats.
The lone exception was high-affinity inhibitor astemizole, which
registered a 4.4-fold higher Ki in the SPA format. The FP derived
Ki values are also in agreement with those reported by Diaz et
al.,19 and the rank order of compound affinities is the same in
all three assays.
Fluorescence Polarization
Because of their high affinity for hERG channel and their
desirable spectral properties, ligands 6, 10e, 12a, and 12b were
selected for further evaluation. NBD ligand 3 displayed low
fluorescence intensity when bound to receptor and was not
pursued further. In order to detect a robust change in polarization
upon receptor binding, a HEK293 cell line that highly expresses
hERG protein was developed.21 Figure 1 shows the change in
FP signal upon Cy3B ligands 10e, 12a, 12b, and MR121 ligand
6 binding to hERG channel membranes. Membranes lacking
the hERG protein were used as a specificity control. For all
four ligands a robust fluorescence polarization effect was
observed. Ligands 10e and 12b showed more than a 100%
increase in polarization over background, while the effect with
ligands 6 and 12a was slightly less pronounced. Interestingly,
12b did not suffer a loss in polarization despite its long and
flexible 14-atom linker.
Patch Clamp Electrophysiology and Cell Permeability
To confirm that the hERG channel binding activities of
ligands 3 and 10e translate to actual functional blockade of
potassium current, ligands were tested using patch clamp
electrophysiology.23 Both 3 and 10e blocked hERG channel
current but at different potencies and with slower kinetics when
compared to 1 (Figure 3A-C, Table 3). By observation of the
percent inhibition at apparent steady-state block as a function
of ligand concentration, IC50 values of 9.2, <3, and 310 nM
were determined for ligands 1, 3, and 10e, respectively (Figure
3D). The IC50 value determined for 1 was comparable to the
Because of its lower background signal and better assay
performance, 10e was chosen for full pharmacological charac-
terization. In competition experiments with increasing concen-

Fluorescence Polarization Ligands for hERG Channel
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 13 2935
Consistent with this hypothesis, 10e displayed low permeability
in both absorptive and secretory transport directions while 1
was moderately permeable (Table 3). NBD compound 3
showed asymmetric permeability, low in the absorptive direc-
tion and moderate in the secretory direction. The results suggest
that 10e and 3 may have reduced access to the intracellular
compartment of the HEK293 cells used in patch clamp
experiments.
Molecular Modeling
A homology model of hERG channel was created using
Modeler 6,27 with the crystal structure of the bacterial potassium
channel KcsA28 as the template structure. Twenty homology
models were generated. On the basis of energy, visual inspec-
tion, and analysis using Procheck,29 the best model was selected.
The conformations of critical ligand binding residues30 Tyr652
and Phe656 were adjusted to the most commonly found confor-
mations in the PDB using Quanta2000. Compounds 3 and 10e
were separately docked in the homology model cavity without
restraints using GOLD.31 The ligand dockings were scored using
the GoldScore function, with the top 20 binding modes retained.
Figure 4A shows the 20 energetically most favorable docked
conformations of compound 3. In 17 of these models, one of
the two nonequivalent methanesulfonanilide (MSA) groups of
3 rests between juxtaposed residues Phe656 from each of the
four protein subunits. Figure 4B shows the three docked confor-
mations of 3 with the highest GoldScore values. While there
are significant differences between these models, all feature a
MSA low in the channel pore. Interestingly three of the top 20
models predict an alternative conformation with the NBD group
interacting with Phe656 and both MSA groups closer to Tyr652
(not shown). In Figure 4C the docked conformation of 10e with
the highest GoldScore is shown along with a model of 3 with
the second highest GoldScore. The polycyclic Cy3B dye of 10e
is predicted to occupy the water-filled cavity28 within the pore.
Other conformations where the Cy3B dye of 10e rests in the
channel pore and interacts with Phe656 were also observed (not
shown).
Discussion
Figure 3. Functional blockade of hERG channel currents assessed by
whole-cell patch clamp. The peak of the current tail is plotted as a
function of time, and typical examples of current blockade by 3 nM 3
(A), 1000 nM 10e (B), and 25 nM 1 (C) are shown at concentrations
that induced approximately 80% hERG channel current inhibition.
Initiation of continuous ligand perfusion is indicated by an arrow, and
extrapolation of the rundown of the control current is depicted with a
solid line. Upon achieving an apparent steady state of ligand blockade,
cells were perfused with 1 µM dofetilide to demonstrate complete
inhibition of current. (D) Percent inhibition of the hERG channel current
at apparent steady state is plotted as a function of ligand concentration.
Results for 1 and 10e were fit with the Hill equation (solid line) to
estimate IC50 values. Data are mean ( SEM; n g 4 except for 10 nM
10e for which n ) 2 cells.
FP assays are finding increasing utility in drug discovery in
part because of their compatibility with miniaturized formats
for HTS. A general limitation when converting small-molecule
radioligand assays to FP assays is the need to develop high-
affinity small-molecule fluorescent conjugates. The major
challenge is to append covalently a fluorescent dye to a known
ligand without suffering a significant loss of binding affinity.
In contrast to radioligand binding experiments, FP requires that
most of the added tracer ligand be bound to give robust
polarization signal. To achieve this endpoint with a fluorescent
ligand of moderate affinity (0.1-1 µM), significant increases
in receptor concentration may be necessary, perhaps to the point
where HTS becomes impractical. In addition to preserving
affinity, it is preferable that the reporter ligand retain the
pharmacological function of the parent ligand as well.
Ligand development is empirical and iterative16 in nature,
but success rates can be enhanced by effective use of preexisting
small-molecule SAR, detailed ligand-receptor models, and
ligand-receptor cocrystal structures. For the development of
FP ligands from 1, we considered the three nonequivalent methyl
groups as synthetic handles for dye linkage. Unpublished results
from these labs show that even modest modifications to either
of the MSA methyl groups result in >10-fold loss in hERG
channel affinity, so synthetic efforts were focused on replace-
SPA Ki of 6.4 nM, while the IC50 of 10e was 40-fold less potent
than the SPA Ki of 7.0 nM. An exact IC50 value for 3 could not
be determined because of slow kinetics of onset of block at
concentrations below 3 nM (τ ) 575 ( 42 s at 3.0 nM).
Since hERG channel ligands are believed to block the channel
from the cytoplasmic face of the pore,4,24 we asked if slow
blocking kinetics of 3 and 10e and reduced potency of 10e
could be attributed to poor membrane permeability conferred
by the fluorescence tags. All three ligands were tested in a stan-
dard Caco-2 permeability assay25 in which 5 µM cyclosporin
A was added as a nonspecific inhibitor of cell transporters.26

2936 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 13
Singleton et al.
Figure 4. Homology model of hERG protein with models of 3 and 10e docked into the pore. (A) Tetrameric hERG protein homology model with
channel-forming helices and loops rendered in gray ribbon style and inner pore aromatic residues Tyr652 and Phe656 shown in dark purple. The
intracellular opening of the channel is at the bottom of the figure. The 20 docked conformations of 3 with the highest GoldScore values are
displayed in the channel pore. The MSA sulfur atoms in yellow show good overlap below the Phe656 residues. (B) Close-up view of the three
docked conformations of 3 with the highest GoldScore values in light-purple, green, and blue. Pore residues Tyr652 and Phe656 are shown in
dark-purple. (C) Close-up view of 10e in gold and 3 in green. Model predicts a high degree of overlap of the lower MSA groups. The Cy3B group
is shown edge-on (orthogonal to the plane of the figure) in the pore cavity above residue Tyr652.
ments of the methyl group at the central tertiary amine of 1.
This led to ligands with poor affinity (4, 5, 10a), moderate
affinity (10b, 14), high affinity (6, 10c-e, 12a, 12b, 13), and
very high affinity (3). Among the high-affinity conjugates
described, Cy3B conjugate 10e was judged to be superior by
virtue of its excellent spectral properties and its performance
in FP competition binding experiments. MR121 conjugate 6
performed well in FP assay development experiments, but 10e
had a higher statistical Z-factor32 value and is preferred
(unpublished results).
low permeability in a Caco-2 assay compared to the moderately
permeable dofetilide, the discrepancy in kinetics of block may
be due in part to the different rates of intracellular accumulation
of the compounds. However, the permeability data cannot fully
account for the observed differences in patch clamp IC50 values.
Ligand 3 is 18-fold more potent than 10e in the SPA and at
least 100-fold more potent in patch clamp, but 3 and 10e show
similar cellular permeability across Caco-2 cells in the presence
of cyclosporin A. Since FP ligand binding studies are performed
using homogenized cell membrane vesicles of variable orienta-
tion (normal and inside out), membrane permeation is not
required for 10e to access the ligand binding site. Therefore,
reduced membrane permeation should have little impact on the
effectiveness of 10e as an FP ligand. One would expect that
FP derived binding data would be as predictive of functional
hERG channel block as a radioligand derived binding data, and
this is indeed the case. The FP derived Ki values for 18 known
hERG channel binding compounds were measured and com-
pared to their patch clamp derived IC50 values.21 Strong
correlation was observed for compounds with potencies ranging
from 8 nM to >10 µM.
It is important to note some key differences in 10e vs ligand
3 permeability observed from the Caco-2 studies. For 10e,
Caco-2 permeability is low because passive permeability was
found to be low. From these studies, one would predict that
cellular permeability of 10e would be low in other cell types.
Conversely, ligand 3 Caco-2 permeability is low in some or all
cell types because of attenuation via efflux transport rather than
solely because of poor passive membrane permeability. It is
likely that the absorptive cellular permeability determined for
3 underestimates the true passive membrane permeability of this
compound because of the attenuating activity of efflux trans-
porters not inhibited by cyclosporin A (it is, however, impossible
to determine from these experimental results if the observed
asymmetry was due to attenuation and/or enhancement of 3
passive membrane permeability during absorptive vs secretory
transport). Therefore, from the Caco-2 studies, the cellular per-
meability of 3 would be expected to depend on both passive
permeability and expression of efflux transporters in the cell
type. There are differences in transporter expression between
Caco-2 and HEK cells (unpublished results), and it is possible
that the efflux transporter responsible for imparting the observed
asymmetry to ligand 3 in Caco-2 cells may not be active, or
could show less activity vs Caco-2 than in HEK cells.
Known hERG channel blockers dofetilide, E-4031, bepridil,
mibefradil, and clozapine were shown to compete equally well
with [3H]dofetilide or 10e in SPA or FP formats, respectively
(Table 2). The compounds tested feature hERG channel affinities
from low nanomolar to low micromolar concentrations, dem-
onstrating consistent results over a dynamic range of at least 3
orders of magnitude. Reassuringly, the values reported here
correspond well with those from another laboratory.19 The high-
affinity blocker astemizole was the only compound to show
greater than a 2-fold difference between the two formats. The
Ki for astemizole binding to hERG channel membranes is
reported to vary more than 10-fold with changes in potassium
ion concentration, but this probably does not explain the
discrepancy reported here because all experiments were run at
10 mM K+. Nonetheless, the rank order of all six compounds
is the same between the two assays. A comprehensive assay
validation study using a panel of well-known hERG channel
binding compounds has been conducted and is reported in a
separate manuscript.21 The data highlight that the Ki values
determined for a range of known hERG channel blockers
(ranging from low nanomolar to tens of micromolar concentra-
tion) are equivalent to those obtained with radiometric methods.
To be predictive of functional hERG channel blockade, an
ideal FP ligand should not only have a high affinity for the
hERG protein but also bind in a manner that inhibits the channel
current. In patch clamp electrophysiology experiments, 10e
reversibly inhibited hERG channel currents21 in a use-dependent
manner similar to 1. However, 3 and 10e displayed slower onset
of channel block (Figure 3A-C, Table 3). At concentrations
that induced 80% current inhibition, time constants (τ) for onset
of block were significantly larger for 3 and 10e than for 1.
Dofetilide and similar molecules are believed to block current
by crossing the cell membrane and entering the channel via the
cytoplasmic face of the protein.4,5,24 Because 3 and 10e displayed

Fluorescence Polarization Ligands for hERG Channel
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 13 2937
Conversely, if the passive membrane permeability of ligands 3
and 10e is comparable, it is possible that the HEK cells used in
patch clamp are able to selectively efflux 10e but not ligand 3.
Each scenario would give rise to differences in permeability
between the compounds and subsequent differences in intrac-
ellular accumulation over time and observed potency between
compounds 10e and 3.
model where a long flexible spacer threads down through the
inner pore and conjugate dye rests in the cytoplasm. However,
with this orientation one would predict that channel affinity
would be relatively independent of conjugate dye. The decrease
in affinity observed when the Cy3B of 10e is exchanged for
TMR in 14 indicates that the dyes are interacting with hERG
protein residues. The cavity-filled model in Figure 4C seems
more consistent with this result. Also, analogue 12b has a
flexible 14-atom spacer between the dofetilide core and the
Cy3B dye and yet retains the magnitude of polarization observed
with 10e when bound (Figure 1). A model with the dye in the
cytoplasm would allow for free dye rotation and perhaps a loss
of polarization. A cavity-packed model would not allow for free
rotation, and polarization would be preserved.
In summary, novel fluorescent analogues of dofetilide (1)
have been discovered. Several analogues have affinity for hERG
channel that equals or exceeds that measured for 1. Most
notably, the Cy3B ligand 10e and the MR121 ligand 6 show
robust increases in polarization when bound to membranes
prepared from HEK293 cells that overexpress hERG protein.
The excellent spectral properties of these ligands have enabled
the development of an economical HT FP binding assay that is
predictive of test compounds’ ability to block hERG channel
current.21 We anticipate that this or similar assays will become
standard medicinal chemistry methods for early assessment of
potential QT prolongation liabilities.
To rationalize our SAR, we referred to models of methane-
sulfonanilide (MSA) compounds docked into the channel of a
hERG protein homology model, which was built upon the crystal
structure of the bacterial potassium channel KcsA.33,34 The KcsA
and subsequent potassium channel crystal structures demonstrate
that pores have conserved features, including an intracellular
membrane proximal inner pore about 18 Å in length, a central
water filled cavity about 10 Å in diameter, and an extracellular
membrane proximal selectivity filter.28 Interpretation of the SAR
reported here is greatly simplified by assuming that the binding
modes of fluorescent dofetilide derivatives are not substantially
perturbed from the consensus MSA binding mode. Our models
and others35,36 predict that MSA inhibitors bind in extended
conformations with one of the MSA groups sandwiched between
critical binding Phe656 residues30 from each protein subunit.
Computational experiments in which high-affinity analogues 3
and 10e were docked into a hERG channel homology model
were performed, and low-energy bound conformations are
shown in Figure 4. The MSA-Phe656 interactions appear to
anchor the dofetilide core within the inner pore (Figure 4A).
Low-energy docked conformations of 3 (Figure 4B) lack overtly
unfavorable nonbonded interactions with the channel but do not
reveal any obvious favorable interactions that would explain
the 15-fold increase in affinity relative to 1. Most likely the
jump in affinity is due at least in part to the significant increase
in hydrophobicity that the NBD confers upon 3 relative to 1.
Once bound, the NBD group is available to interact with Tyr652
or other hydrophobic residues. Docked conformations of 3 where
the NBD group interacts with Phe656 and both MSA groups
interact with Tyr652 were also observed (data not shown). These
do not match well to consensus MSA binding models but should
not be dismissed entirely.
Experimental Section
Chemistry. Reactions were performed using commercially
available reagents from Aldrich (Milwaukee, WI) and solvents from
JT Baker or Applied Biosystems Inc. (Foster City, CA) that were
used as purchased. 5-Carboxyfluorescein NHS ester, 5-carboxy
TMR NHS ester, 5- and 6-carboxyfluorescein (mixed isomers),
5-carboxy TMR, and NBD chloride were from Molecular Probes
(Eugene, OR). Cy3B NHS ester was from GE Healthcare (Amer-
sham Biosciences Piscataway, NJ). MR121 was purchased from
Roche Biosciences (NJ). A synthesis of N-desmethyldofetilide20
(2) has been reported. Boc amino aldehydes were prepared from
commercially available Boc amino alcohols as described.39 HATU
and aminomethylpolystyrene resin were purchased from EMD
Biosciences (San Diego, CA). Astemizole, bepridil hydrochloride,
clozapine, mibefradil dihydrochloride hydrate, and E-4031 (N-[4-
[[1-[2-(6-methyl-2-pyridinyl)ethyl]-4-piperidinyl]carbonyl]phe-
nyl] methanesulfonamide dihydrochloride) are sold by Sigma (St.
Louis, MO). Silica gel chromatography was performed on a Biotage
flash column gradient pump system using 15 cm long columns.
NMR was performed on a 400 MHz field strength spectrometer
from Varian Inc. (Palo Alto, CA) in CD3OD. Chemical shifts are
reported in parts per million (ppm). Methanol solvent peaks were
used as references (1H ) 3.32 ppm, 13C ) 49.0 ppm). High-
resolution mass spectrometry was peformed in electrospray mode.
Analytical samples for liquid chromatography electrospray mass
spectrometry (LC ESMS) were dissolved at 4 µM in 1:1 ACN/
H2O. An amount of 5 µL each was injected onto an Aquity 2.1
mm × 100 mm UPLC 1.7 Å BEH C18 column and eluted using a
linear gradient from 100% 0.1% formic acid/water to 100% ACN
over 10 min with a flow rate of 0.4 mL/min. Mass detection was
performed using a Micromass LCT spectrometer employing external
calibration with sodium iodide/cesium iodide clusters to determine
experimental masses. Observed peaks were evaluated by summing
those scans at one-half peak height. HRMS experiments were
performed using electrospray ionization on a Micromass MUX-
LCTP instrument that had been calibrated with leucine enkaphalin
as the lock mass reference. Compound purity was assessed by
complementary HPLC methods: C18 reverse-phase chromatogra-
phy and hydrophilic interaction chromatography using diode array
UV detection as described in Supporting Information. All com-
pounds met or exceeded 95% purity except 4 (roughly a 4:3 mixture
For high-affinity conjugates 6, 10c-e, and 12a,b (SPA Ki <
25nM) to bind in a mode that is consistent with models proposed
for MSA inhibitors, the polycyclic conjugate dyes need to
occupy space outside the inner pore. An obvious compartment
is the cavity just above the inner pore. hERG protein homology
models predict that hydrophobic residues line the cavity, creating
a suitable environment for aromatic dyes.28 The docked model
of 10e shown in Figure 4C orients the Cy3B moiety in the
central cavity of the pore with the planar fused ring system
orthogonal to the pore axis. This model requires a flexible spacer
between the dye and dofetilide core and therefore is consistent
with the spacer length dependence observed for analogues 10a-
e. Alternative conformations where the dye resides in the lower
channel or outside the pore cannot be ruled out. Testing for the
molecular determinants of blockade using mutagenesis of the
channel would lead to greater certainty that the binding does
occur at similar residues as does dofetilide.
The KcsA structure and hence the homology model shown
here exist in closed conformation. Other workers have proposed
hERG channel homology models based on open potassium
channel structures.37,38 Those models have larger pore openings
that allow docked compounds to reside lower in the pore (closer
to the cytoplasmic opening), but the predicted interactions with
binding residue Phe656 are qualitatively similar. Open channel
homology models might also accommodate a ligand binding

2938 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 13
Singleton et al.
of 5- and 6-carboxyfluorescein isomers) and 10a (82% pure).
Preparative HPLC purifications were performed with either a Vydac
(Hesperia, CA) C18 column (part number 218TP1022) or a Vydac
semipreparative diphenyl column (Vydac part number 219TP510)
using a 0.1% TFA/water and neat ACN elution gradient. Yields
for fluorescent products were calculated from UV absorbance
measurements performed on HPLC purified products. Briefly, an
aliquot of pooled product in HPLC eluent was diluted into pH 7
aqueous buffer and the absorbance measured at the following
wavelengths: 466 nm for compound 3 (ꢀ ) 22 000 L mol-1 cm-1);
492 nm for 4 and 13 (ꢀ ) 75 000 L mol-1 cm-1); 550 nm for 5
and 14 (ꢀ ) 80 000 L mol-1 cm-1); 660 nm for 6 (ꢀ ) 100 000 L
mol-1 cm-1); 562 nm for 10a-e and 12a,b (ꢀ ) 130 000 L mol-1
cm-1).
N-(4-{2-[[2-(4-Methanesulfonylaminophenoxy)ethyl](7-
nitrobenzo[1,2,5]oxadiazol-4-yl)amino]ethyl}phenyl)-
methanesulfonamide (3). To a solution of 25.5 mg (60 µmol) of
N-desmethyldofetilide (2)20 in 0.3 mL of DMF was added 12 mg
(60 µmol) of 4-chloro-7-nitrobenz-2-oxa-1,3-diazole (NBD chlo-
ride). After 2 h the reaction was judged by LCMS to be >90%
consumed. The reaction mixture was directly purified by preparative
C18 HPLC. Collected fractions were analyzed by analytical LCMS,
and those judged as having adequate purity were pooled. Product
yield was calculated as 14.5 mg (41%), and the pooled fractions
were lyophilized to dryness. 1H NMR (CD3OD): δ 8.49 (d, 1H, J
) 10 Hz), 7.26 (d, 2H, J ) 8.3 Hz), 7.15 (d, 2H, J ) 8.3 Hz), 7.14
(d, 2H, J ) 8.8 Hz), 6.87 (d, 2H, J ) 8.8 Hz), 6.53 (d, 1H, J ) 10
Hz), 4.16 (m 4H), 3.26 (m, 2H), 3.05 (t, 2H, J ) 7.0 Hz), 2.87 (s,
3H), 2.82 (s, 3H). ESMS: calculated m/z for (MH)+ ) 591.13,
found ) 591.10. HRMS: m/z calculated for C24H27N6O8S2 (MH)+
) 591.1326, found ) 591.1288.
Synthesis of 5 (5-Carboxy TMR Conjugate of 2), TFA Salt.
To a solution of 4.3 mg (10 µmol) of 5-carboxy TMR and 0.10
mL of DMF were added 3.8 mg (10 µmol) of HATU and 20 µL of
a 0.5 M solution of HOAt (10 µmol) in DMF. Then 20 µL (120
µmol) DIEA was added, and after 1 min the resulting deep-red
solution was added to a solution of 4.3 mg (12 µmol) of 2 dissolved
in 0.20 mL of DMF. After 20 min, the product was purified by
direct injection of the reaction mixture onto a semipreparative
diphenyl column. Collected fractions were analyzed by HPLC, and
appropriate fractions were pooled and lyophilized. Calculated
product yield ) 3.32 mg (40%). ESMS: calculated m/z for (MH)+
) 840.27, found ) 840.11. HRMS: m/z calculated for C43H46N5O9S2
(MH)+ ) 840.2726, found ) 840.2761.
Synthesis of 4 (5- and 6-Carboxyfluorescein Conjugates of
2). The procedure is similar to the procedure for 5. Calculated
product yield ) 5%. ESMS: calculated m/z for (MH)+ ) 786.18,
found ) 786.12. HRMS: m/z calculated for C39H36N3O11S2 (MH)+
) 786.1782, found ) 786.1851
Synthesis of 6 (MR121 Conjugate of 2), TFA Salt. The
procedure is similar to the procedure for 5. Calculated product yield
) 35%. ESMS: calculated m/z for (M)+ ) 815.33, found ) 815.24.
HRMS: m/z calculated for C42H51N6O7S2 (M)+ ) 815.3249, found
) 815.3273.
tert-Butyl 3-((4-(Methylsulfonamido)phenethyl)(2-(4-(meth-
ylsulfonamido)phenoxy)ethyl)amino)propylcarbamate (7b).
3-[(tert-Butoxycarbonyl)-amino]propanal39 (81 mg, 468 µmol) in
8 mL of 1,2-dichloroethane was mixed with 100 mg (234 µmol)
of 2 followed by 100 mg (468 µmol) of sodium triacetoxyboro-
hydride. To enhance solubility, 2 mL of tetrahydrofuran was added
to the mixture which was stirred at room temperature. After 40
min, 50 mg (234 µmol) of sodium triacetoxyborohydride was added.
After a total of 5 h, 100 mg (468 µmol) of sodium triacetoxyboro-
hydride was added. After 16 h, excess residual aldehyde was
absorbed by the addition of 100 mg of aminomethylpolystyrene
resin followed by stirring for 30 min. The resin was removed by
filtration, and the remaining filtrate was diluted with 50 mL of water
and extracted with 100 mL of dichloromethane. The organic layer
was washed with 1 × 50 mL saturated NaCl, dried over MgSO4,
filtered, and concentrated to an oily solid. Purification was
performed by silica gel chromatography, eluting with 90-100%
EtOAc/hexanes. Fractions judged as containing product were pooled
and solvent was evaporated, affording 64.2 mg (47%) of product.
1H NMR (CD3OD): δ 7.20 (m, 6H), 6.82 (d, 2H, J ) 7 Hz), 4.81
(s, 4H), 3.96 (m, 2H), 3.99 (dd, 2H, J ) 6.3, 6.5), 2.80 (s, 3H),
2.78 (s, 3H), 2.70 (m, 2H), 2.60 (t, 2H, J ) 7.2), 1.57 (m, 2H),
1.34 (m, 9H). 13C NMR (CD3OD): δ 156.9, 129.9, 129.6, 124.1,
121.1, 121.0, 115.2, 80.6, 66.6, 59.6, 59.3, 56.4, 52.8, 52.12, 37.8,
37.6, 37.1, 32.4, 27.6, 25.2. ESMS: calculated m/z for (MH)+
)
585.24, found ) 585.15. HRMS: m/z calculated for C26H41N4O7S2
(MH)+ ) 585.2407, found ) 585.2413.
tert-Butyl 2-((4-(Methylsulfonamido)phenethyl)(2-(4-(meth-
ylsulfonamido)phenoxy)ethyl)amino)ethylcarbamate (7a). The
1
procedure is similar to the procedure for 7b. H NMR (CD3OD):
δ 7.29 (d, 2H, J ) 8 Hz), 7.22 (m, 4H), 7.01 (d, 2H, J ) 10 Hz),
4.37 (m, 2H), 3.77 (m, 2H), 3.52 (m, 6H), 3.09 (m, 2H), 2.92 (s,
3H), 2.87 (s, 3H), 1.48 (d, 9H). ESMS: calculated m/z for (MH)+
) 571.23, found ) 571.15. HRMS: m/z calculated for C25H39N4O7S2
(MH)+ ) 571.2251, found ) 571.2292.
tert-Butyl 6-((4-(Methylsulfonamido)phenethyl)(2-(4-(meth-
ylsulfonamido)phenoxy)ethyl)amino)hexylcarbamate (7e). The
1
procedure is similar to the procedure for 7b. H NMR (CD3OD):
δ 7.25 (d, 2H, J ) 9hz), 7.23 (2d, 4H), 6.99 (d, 2H, J ) 9 Hz),
4.36 (m, 2H), 3.71 (m, 2H), 3.46 (m, 2H), 3.30 (m, 2H), 3.07 (t,
2H), 3.01 (t, 2H), 2.92 (s, 3H), 2.88 (s, 3H), 1.78 (m, 2H), 1.47
(m, 6H), 1.40 (s, 9H). 13C NMR (CD3OD): δ 156.8, 139.0, 133.7,
133.4, 131.1, 125.2, 122.3, 116.6, 63.9, 56.2, 55.4, 53.6, 41.2, 39.4,
39.1, 30.9, 30.5, 28.9, 27.4, 27.3, 24.9. ESMS: calculated m/z for
(MH)+ ) 627.29, found ) 627.38. HRMS: m/z calculated for
C29H47N4O7S2 (MH)+ ) 627.2875, found ) 627.2844.
N-[4-(2-{(3-Aminopropyl)[2-(4-methanesulfonylaminophenoxy)-
ethyl]amino}ethyl)phenyl]methanesulfonamide, TFA Salt (9b).
An amount of 5.4 mg (9.2 µmol) of 7b was dissolved in 2 mL of
TFA. After 5 min, the solvent was evaporated. The resulting oil
was rubbed with 25 mL of diethyl ether three times and then
pumped on vacuum. The oil was analyzed by LC-ESMS, judged
to be at least 95% pure, and then was used directly without further
purification. ESMS: calculated m/z for (MH)+ ) 485.19, found )
485.14.
N-[4-(2-{(3-Aminoethyl)[2-(4-methanesulfonylaminophenoxy)-
ethyl]amino}ethyl)phenyl]methanesulfonamide (9a), TFA Salt.
The procedure is similar to the procedure for 9b. ESMS: calculated
m/z for (MH)+ ) 471.17, found ) 471.01.
N-[4-(2-{(3-Aminohexyl)[2-(4-methanesulfonylaminophenoxy)-
ethyl]amino}ethyl)phenyl]methanesulfonamide (9e), TFA Salt.
The procedure is similar to the procedure for 9b. ESMS: calculated
m/z for (MH)+ ) 527.24, found ) 527.51.
N-[4-(2-{[4-(1,3-Dioxo-1,3-dihydroisoindol-2-yl)butyl][2-(4-
methanesulfonylaminophenoxy)ethyl]amino}ethyl)phenyl]-
methanesulfonamide (8a), TFA Salt. To a solution of 21 mg (50
µmol) of 2 in 2.0 mL of acetonitrile was added 5.0 mg (50 µmol)
of Na2CO3 followed by 19 mg of N-(4-bromobutyl)phthalimide.
The mixture was warmed to reflux. After 2 h, an additional 15 mg
(53 µmol) of N-(4-bromobutyl)phthalimide was added. After an
additional 2 h, the mixture was cooled to room temperature, 0.5
mL of ACN and 0.5 mL of water were added, and the product was
isolated by purification via preparative C18 HPLC (10 mL/ min
flow rate, 100% (0.1% TFA) water to 80% ACN 30 min). Collected
fractions were analyzed by HPLC and appropriate fractions pooled
1
and lyophilized. Yield ) 16.6 mg (53%). H NMR (CD3OD): δ
7.81 (m, 4H), 7.27 (d, 2H, J ) 9 Hz), 7.21 (m, 4H), 6.97 (d, 2H,
J ) 9 Hz), 4.36 (t, 2H, J ) 5 Hz), 3.73 (m, 4H), 3.42 (m, 4H),
3.07 (m, 2H), 2.91 (s, 3H), 2.87 (s, 3H), 1.80 (m, 4H). 13C NMR
(CD3OD): δ 168.8, 155.4, 137.6, 134.3, 132.4, 132.1, 129.7, 123.8,
123.0, 121.0, 115.4, 112.5, 62.6, 55.0, 53.6, 52.5, 38.1, 37.7, 36.5,
29.1, 25.5, 20.9. ESMS: calculated m/z for (MH)+ ) 629.21, found
) 629.24. HRMS: m/z calculated for C30H37N4O7S2 (MH)+
629.2095, found ) 629.2125.
)
N-[4-(2-{[4-(1,3-Dioxo-1,3-dihydroisoindol-2-yl)pentyl][2-(4-
methanesulfonylaminophenoxy)ethyl]amino}ethyl)phenyl]-
methanesulfonamide (8b), TFA Salt. The procedure is similar to
1
the procedure for 8a. H NMR (CD3OD): δ 7.82 (m, 4H,), 7.28

Fluorescence Polarization Ligands for hERG Channel
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 13 2939
(d, 2H, J ) 9 Hz), 7.22 (m, 4H), 6.98 (d, 2H, J ) 9 Hz), 4.37 (t,
2H, J ) 5 Hz), 3.70 (m, 4H), 3.47 (m, 2H), 3.29 (m, 2H), 3.08 (m,
2H), 2.92 (s, 3H), 2.87 (s, 3H), 1.86 (m, 2H), 1.75 (m, 2H), 1.41
(m, 2H). 13C NMR (CD3OD): δ 168.8, 155.4, 137.6, 134.3, 132.4,
132.1, 129.7, 123.8, 122.9, 121.0, 115.3, 62.5, 58.7, 54.9, 53.8,
52.3, 38.1, 37.7, 36.8, 29.1, 27.7, 26.2, 23.4. ESMS: calculated
m/z for (MH)+ ) 643.23, found ) 643.27. HRMS: m/z calculated
for C31H39N4O7S2 (MH)+ ) 643.2251, found ) 643.2421.
N-[4-(2-{(3-Aminopropyl)[2-(4-methanesulfonylaminophenoxy)-
ethyl]amino}ethyl)phenyl]methanesulfonamide (9c), TFA Salt.
To a solution containing 6.3 mg (10 µmol) of 10c dissolved in 500
µL of methanol was added 10 µL (310 µmol) of hydrazine. After
5 min at room temperature, the mixture was warmed to reflux for
2 h. The reaction was judged to be complete by HPLC, and the
mixture was cooled to ambient temperature. The solution was
purified by direct injection onto a preparative C18 HPLC column
(as for 8a). Collected fractions were analyzed by HPLC and
appropriate fractions pooled and lyophilized. Yield ) 2.9 mg (58%).
ESMS: calculated m/z for (MH)+ ) 499.20, found ) 498.97.
Synthesis of tert-Butyl 6-(6-((4-(Methylsulfonamido)phen-
ethyl)(2-(4-(methylsulfonamido)phenoxy)ethyl)amino)hexylamino)-
6-oxohexylcarbamate (11b). The procedure is similar to the
procedure for 11a. ESMS: calculated m/z for (MH)+ ) 740.37,
found ) 740.78.
Synthesis of 12a (Cy3B Conjugate of Boc Deprotected 11a),
TFA Salt. An amount of 3.5 mg (5.0 µmol) of compound 11a was
dissolved in 0.5 mL of TFA. After 2 min, the TFA was removed
under a stream of nitrogen and placed under high vacuum for 3 h.
The residue was dissolved in 400 µL of DMF and 100 µL of DIEA,
and 5.0 mg (4.3 µmol) of Cy3B NHS ester was added. After 1 h
at room temperature, the reaction mixture was directly purified by
preparative C18 HPLC as for compound 8a. Collected fractions
were analyzed by HPLC and appropriate fractions pooled. Calcu-
lated product yield ) 1.86 mg (23%). Pooled fractions were
lyophilized to dryness. ESMS: calculated m/z for (MH2)+2
570.73, found ) 570.67.
)
Synthesis of 12b (Cy3B Conjugate of Boc Deprotected 11b),
TFA Salt. The procedure is similar to the procedure for 12a.
Calculated product yield ) 1.67 mg (20%). ESMS (MH2)+2 m/z
calculated ) 591.76, found ) 591.71.
Synthesis of 13 (the 5-Carboxyfluorescein Conjugate of 9e),
TFA Salt. An amount of 7.1 mg (15 µmol) of 5-carboxyfluorescein
NHS ester was mixed with 9.8 mg (16 µmol) of 9e using the same
conditions and methods as for 10e. Yield ) 4.32 mg (32%).
ESMS: m/z calculated for (MH)+ ) 885.2, found ) 885.33; m/z
calculated for (MH2)+2 ) 443.15, found ) 443.09.
Synthesis of 14 (5-Carboxy TMR Conjugate of 9e), TFA Salt.
An amount of 7.5 mg (14 µmol) of of 5-carboxy TMR NHS ester
was mixed with 9.8 mg (16 µmol) of 9e using the same conditions
and methods as for 10e. Yield ) 2.2 mg (18%). ESMS: calculated
m/z for MH+ ) 939.37, found ) 939.46; calculated m/z for (MH2)+2
) 470.19, found ) 470.14. HRMS: m/z calculated for C49H60N6O9S2
(MH2)+2 ) 470.1925, found ) 470.2032.
Synthesis of 15 (Cy3B Ethylamide), TFA Salt. To 0.50 mg
(0.15 µmol) of Cy3B NHS ester was added 100 µL of DMF
followed by 0.50 mL of 2 M (1.0 mmol) ethylamine in methanol.
After 10 min, the reaction was judged to be complete by LC-
ESMS. The reaction solution was directly purified by passage over
a preparative C18 HPLC column as for 8a. Collected fractions were
analyzed by LC-ESMS, and appropriate fractions were pooled and
lyophilized. Calculated product yield ) 0.22 mg (47%). ESMS:
m/z calculated for (MH)+ ) 588.25, observed ) 588.04.
N-[4-(2-{(4-Aminobutyl)[2-(4-methanesulfonylaminophenoxy)-
ethyl]amino}ethyl)phenyl]methanesulfonamide (9d), TFA Salt.
The procedure is similar to the procedure for 9c. Yield ) 58%.
ESMS: calculated m/z for (MH)+ ) 513.22, found ) 512.95.
Synthesis of 10e (Cy3B Conjugate of 9e), TFA Salt. To 5.0
mg (15.8 µmol) of Cy3B-NHS ester was added 1.0 mL of DMF
followed by 50 µL of DIEA. The solution was added to 9.8 mg
(15.8 µmol) of 9e. After 5 min, the reaction was judged to be 85%
complete by LC-ESMS. The reaction solution was directly purified
by passage over a C18 preparative HPLC column as for 8a.
Collected fractions were analyzed by LC-ESMS, and appropriate
fractions were pooled. Calculated product yield ) 7.0 mg (86%).
The solution was aliquoted and lyophilized to a powder. 1H NMR
(CD3OD): δ 8.17 (s, 1H), 7.90 (s, 1H), 7.87 (d, 1H, J ) 1hz),
7.43 (s, 1H), 7.36 (dd, 1H, J ) 1 Hz, 8 Hz), 7.28 (m, 4H), 7.23
(m, 4H), 6.98 (d, 2H, J ) 9 Hz), 4.67 (m, 2H), 4.37 (t, 2H), 4.34
(m, 1H), 3.91 (m, 1H), 3.71 (m, 1H), 3.33 (s,1H), 3.48 (m, 2H),
3.24 (m, 4H), 3.21 (m, 4H), 3.13 (m, 2H), 3.07 (m, 2H), 2.92 (s,
3H), 2.87 (s, 3H), 2.59 (m, 2H), 2.00 (m, 2H), 1.75 (m, 12H), 1.52
(m, 2H), 1.37 (m, 4H). ESMS: calculated m/z for (MH)+ ) 1069.4,
found ) 1068.9; calculated m/z for (MH2)+2 ) 535.21, found )
535.14. HRMS: m/z calculated for C55H70N6O10S3 (MH2)+2
535.2150, found ) 535.2267.
)
Synthesis of 10a (Cy3B Conjugate of 9a), TFA Salt. The
procedure is similar to the procedure for 10e. ESMS: calculated
m/z for (MH)+ ) 1013.36, found ) 1012.95; calculated m/z for
(MH2)+2 ) 507.19, found ) 507.13.
Synthesis of 10b (Cy3B Conjugate of 9b), TFA Salt. The
procedure is similar to the procedure for 10e. ESMS: calculated
m/z for (MH2)+2 ) 514.19, found ) 514.11. HRMS: m/z calculated
for C52H64N6O10S3 (MH2)+2) 514.1916, found ) 514.2021.
Synthesis of 10c (Cy3B Conjugate of 9c), TFA Salt. The
procedure is similar to the procedure for 10e. ESMS: calculated
m/z for (MH2)+2 ) 521.20, found ) 521.14.
Synthesis of 10d (Cy3B Conjugate of 9d), TFA Salt. The
procedure is similar to the procedure for 10e. ESMS: calculated
m/z for (MH2)+2 ) 528.21, found ) 528.14.
Synthesis of [2-(6-{[2-(4-Methanesulfonylaminophenoxy)-
ethyl][2-(4-methanesulfonylaminophenyl)ethyl]amino}-
hexylcarbamoyl)ethyl]carbamic Acid tert-Butyl Ester (11a), TFA
Salt. A solution of 4.9 mg (9.3 µmol) of 9e in 100 µL of 10%
DIEA in NMP was prepared. An amount of 10 µL of a 0.5 M
solution containing 3-tert-butoxycarbonylaminopropionic acid (Boc-
â-Ala-OH) was activated by HBTU and DIEA in stoichiometric
amounts. After 5 min, an additional 5 µL was added. The reaction
was judged by LC-ESMS to be approximately 75% complete. The
reaction solution was directly purified by passage over a C18
preparative HPLC column as for 8a. Collected fractions were
analyzed by HPLC and appropriate fractions pooled and evaporated.
Yield ) 4.2 mg (65%). ESMS: calculated m/z for (MH)+ ) 698.33,
found ) 698.73.
Binding Assays. [3H]Dofetilide scintillation proximity assays
(SPA) were performed as previously described.40 Fluorescence
polarization (FP) binding assay methods are detailed in ref 21 and
described briefly here. Membrane homogenates from HEK-293S
(cell line no. 15-08) cells expressing the hERG product were
prepared as follows. Cell pellets were thawed at room temperature
and kept on ice. Buffer (50 mM Tris-HCl, 1 mM MgCl2, 10 mM
KCl, pH 7.4, at 4 °C) was added to each cell pellet (10 mL of
buffer per 10 g of packed cell pellet), and the mixture was
homogenized using an Omni LabTek homogenizer (20 000 rpm
for 30 s). The homogenate was centrifuged at 48000g for 20 min
at 3-5 °C in a Sorvall Evolution RC centrifuge and the supernatant
discarded. The pellet was resuspended, homogenized (20 000 rpm
for 10 s), and centrifuged as before. The resultant supernatant was
discarded, and the final pellet was resuspended (100 mL of the
above buffer per 10 g of packed cell pellet), homogenized (20 000
rpm for 10 s), dispensed in to tubes in 1.2 and 5 mL aliquots, and
stored between -75 and -85 °C until use. Protein concentration
was determined using a Coomassie Blue kit as per manufacturer’s
instructions (Sigma 610A and 610-11). Specific hERG protein levels
of g10 pmol/mg total protein were required for FP assay.
Membranes with <5 pmol/mg protein did not work in FP mode.
The Cy3B ligands 10e, 12a, and 12b and MR121 ligand 6 were
stored in 100% DMSO and diluted to 6 nM in assay buffer (50
mM Tris-HCl, 1 mM MgCl, 10 mM KCl, 0.05% Pluronic F127
(BASF, Mount Olive, NJ), pH 7.4, at 4 °C) on the day of the
experiment. Test samples and controls were diluted in 6% DMSO
and 0.05% Pluronic F127. Cell membranes were removed from

2940 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 13
Singleton et al.
the -80 °C freezer and placed on ice after defrosting. When
required, the defrosted membranes were homogenized using a
polytronic device for no more than 10 s; they were then diluted in
the above assay buffer to produce a working solution of 0.3 mg/
mL. The assay was compiled by adding 10 µL of test compound
or control solution, 10 µL of the Cy3B ligand, and 10 µL of cell
membranes to a black 384-well plate (Matrix, catalog no. 4318).
The plates were mixed and then incubated for a minimum of 2 h
prior to reading on a Tecan Ultra (for Cy3B, excitation at 530 nm
and emission at 590 nm; for MR121, excitation at 595 nm and
emission at 680 nm). Ki values22 were generated using Pfizer
proprietary software.
70% ACN, and saved for analysis. Permeability assays were
performed in triplicate. Caco-2 cell monolayer integrity was
assessed immediately following transport experiments by determin-
ing the flux of 0.1 mg/mL Lucifer yellow over 1 h in the absorptive
transport direction. Compounds were detected using quantitative
HPLC-MS-MS.
The apparent permeability (Papp) was calculated using the
equation
dMr
dt
1
Papp
)
(area)(CD(0)
)
Patch Clamp Electrophysiology. Testing was carried out in
HEK293 cells transfected with the hERG gene41 maintained at
37 °C in Minimum Essential Medium with Earle’s salts and
l-glutamine, 10% fetal bovine serum, 1 mM sodium pyruvate, 0.1
mM nonessential amino acids, and 0.4 mg/mL Geneticin. Membrane
currents of cells continuously perfused with 35 °C extracellular
recording saline (137 mM NaCl, 4 mM KCl, 1.8 mM CaCl2, 1
mM MgCl2, 10 mM HEPES, 10 mM D-glucose, pH 7.4, 335 mM
mOs) were measured using whole-cell patch clamp23 with a
MultiClamp 700A amplifier (Molecular Devices) and 3-5 MΩ
glass pipettes filled with intracellular recording saline (130 mM
KCl, 1 mM MgCl2, 10 mM HEPES, 5 mM Mg-ATP, 5 mM EGTA,
pH 7.2, 320 mM mOs). The liquid junction potential was 4 mV,
and access resistance was compensated by at least 80%. For
characterization of ligand potency, the hERG channel current was
activated with a voltage step to +20 mV for 1 s followed by a
ramp to -80 mV at 0.5 V/s delivered at 0.25 Hz. Following 5 min
of stable control recording in which rundown of the tail current
peak was <2% per minute, ligand solution was continuously
perfused into the recording chamber for evaluation of current
blockade (10-25 min). Upon reaching an apparent steady state of
current inhibition, cells were exposed to 1 µM 1 (dofetilide) to
block any remaining channel current. The peak of the hERG channel
tail current produced by the voltage ramp was plotted as a function
of time. A linear regression of 5 min of control data prior to drug
application was performed, and the linear fit of the control data
was extrapolated for calculation of percent inhibition. Data were
fit with the Hill equation in order to estimate IC50 values. To derive
a time constant (τ) for onset of block, currents in the presence of
compound were normalized to the rundown-corrected control
current amplitude and plotted as a function of time and the decay
phase was fit with a single-exponential equation. Time constants
were compared using an ANOVA and Bonferroni procedure to
control for multiple comparisons (a Bonferroni-corrected p value
of <0.05 was considered significant).
Permeability Assays. Caco-2 cells were obtained from American
Type Culture Collection (Rockville, MD). Cells were cultured at
37 °C with cell culture medium in an atmosphere of 10% CO2 and
90% relative humidity. The cells were passaged upon reaching
approximately 75-85% confluence from T-flasks using 0.05%
trypsin-EDTA (Invitrogen, Gibco Laboratories, Grand Island, NY).
Caco-2 cells were seeded in Transwells (Corning Costar, Cam-
bridge, MA) at a density of 60 000 cells/cm2. The cell culture
medium was changed biweekly (every 72 h) and was changed 24
h prior to experimentation. Caco-2 cell monolayers were used for
experimentation between 17 and 25 days after seeding. Caco-2 cell
monolayers were preincubated for 30 min at 37 °C (temperature
maintained throughout the experiment) with transport buffer
containing 5 µM cyclosporin A. To begin transport experiments,
donor solutions of test compound with 5 µM cyclosporin A were
added to the donor compartment. For absorptive (apical (AP) to
basolateral (BL)) transport, the AP compartment is the donor and
the BL compartment is receiver. For secretory (from BL to AP)
transport, the BL compartment is the donor and AL compartment
is receiver. The appearance of test compound into receiver
compartments was measured at 1 and 2 h. At the conclusion of the
experiments, samples were taken from the donor compartments
along with samples from the initial donor solutions and were saved
for analysis. Cell monolayers on filters were exised, lysed with
where “area” is the surface area of the cell monolayer on the filter
support (1.0 cm2), CD(0) is the initial concentration of test compound
applied to the donor chamber, t is time, Mr is the mass of compound
appearing in the receiver compartment as a function of time, and
dMr/dt is the flux of the compound across the cell monolayer. The
mass recovery of test compounds was assessed and found to be
86-117%.
Acknowledgment. The authors are grateful to Leonard
Contillo for model compound synthesis, Navin Varshney for
HRMS, Dr. John Soglia for expert mass spectral analyses,
Rodger Pasieczny for patch clamp analysis, and Dr. C. Preston
Hensley for helpful comments.
Supporting Information Available: Details of HPLC purity
assessments for new compounds, table listing purity data, and HPLC
chromatograms of 3, 6, and 10e. This material is available free of
charge via the Internet at http://pubs.acs.org.
References
(1) Spector, P. S.; Curran, M. E.; Keating, M. T.; Sanguinetti, M. C.
Class III antiarrhythmic drugs block hERG, a human cardiac delayed
rectifier K+ channel: open-channel block by methanesulfonanilides.
Circ. Res. 1996, 78, 499-503.
(2) Curran, M. E.; Splawski, I.; Timothy, K. W.; Vincent, G. M.; Green,
E. D.; Keating, M. T. A molecular basis for cardiac arrhythmia:
HERG mutations cause long QT syndrome. Cell (Cambridge, Mass.)
1995, 80, 795-803.
(3) Meadows, M. Why drugs get pulled off the market. FDA Consum.
2002, 36 (1), 11-17.
(4) Snyders, D. J.; Chaudhary, A. High affinity open channel block by
dofetilide of HERG expressed in a human cell line. Mol. Pharmacol.
1996, 49, 949-955.
(5) Kiehn, J.; Lacerda, A. E.; Wible, B.; Brown, A. M. Molecular
physiology and pharmacology of HERG. Single-channel currents and
block by dofetilide. Circulation 1996, 94, 2572-2579.
(6) Brown, A. M. hERG Assay, QT Liability, and sUdden Cardiac Death.
In Cardiac Safety of Noncardiac Drugs; Morganroth, J., Gussak, I.,
Eds.; Humana Press: Totowa, NJ, 2005; pp 67-81.
(7) Dubin, A. E.; Nasser, N.; Rohrbacher, J.; Hermans, A. N.; Marrannes,
R.; Grantham, C.; Van Rossem, K.; Cik, M.; Chaplan, S. R.;
Gallacher, D.; Xu, J.; Guia, A.; Byrne, N. G.; Mathes, C. Identifying
modulators of hERG channel activity using the PatchXpress planar
patch clamp. J. Biomol. Screening 2005, 10, 168-181.
(8) Bridgland-Taylor, M. H.; Hargreaves, A. C.; Easter, A.; Orme, A.;
Henthorn, D. C.; Ding, M.; Davis, A. M.; Small, B. G.; Heapy, C.
G.; Abi-Gerges, N.; Persson, F.; Jacobson, I.; Sullivan, M.; Albertson,
N.; Hammond, T. G.; Sullivan, E.; Valentin, J. P.; Pollard, C. E.
Optimization and validation of a medium-throughput electrophysi-
ology-based hERG assay using IonWorks HT. J. Pharmacol. Toxicol.
Methods 2006, 54, 189-199.
(9) Finlayson, K.; Sharkey, J. A High-Throughput Binding Assay for
HERG. In Optimization in Drug DiscoVery. In Vitro Methods;
Zhengyin, Y., Caldwell, G. W., Eds.; Humana Press: Totowa, NJ,
2004; pp 353-368.
(10) Cheng, C. S.; Alderman, D.; Kwash, J.; Dessaint, J.; Patel, R.; Lescoe,
M. K.; Kinrade, M. B.; Yu, W. A high-throughput HERG potassium
channel function assay: an old assay with a new look. Drug DeV.
Ind. Pharm. 2002, 28, 177-191.
(11) Chiu, P. J. S.; Marcoe, K. F.; Bounds, S. E.; Lin, C.-H.; Feng, J.-J.;
Lin, A.; Cheng, F.-C.; Crumb, W. J.; Mitchell, R. Validation of a
[3H]astemizole binding assay in HEK293 cells expressing HERG K+
channels. J. Pharmacol. Sci. (Tokyo, Jpn) 2004, 95, 311-319.

Fluorescence Polarization Ligands for hERG Channel
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 13 2941
(12) Raab, C. E.; Butcher, J. W.; Connolly, T. M.; Karczewski, J.; Yu,
N. X.; Staskiewicz, S. J.; Liverton, N.; Dean, D. C.; Melillo, D. G.
Synthesis of the first sulfur-35-labeled hERG radioligand. Bioorg.
Med. Chem. Lett. 2006, 16, 1692-1695.
(13) Dorn, A.; Hermann, F.; Ebneth, A.; Bothmann, H.; Trube, G.;
Christensen, K.; Apfel, C. Evaluation of a high-throughput fluores-
cence assay method for hERG potassium channel inhibition. J.
Biomol. Screening 2005, 10, 339-347.
(14) Berque-Bestel, I.; Soulier, J.-L.; Giner, M.; Rivail, L.; Langlois, M.;
Sicsic, S. Synthesis and characterization of the first fluorescent
antagonists for human 5-HT4 receptors. J. Med. Chem. 2003, 46,
2606-2620.
(15) Cha, J. H.; Zou, M.-F.; Adkins, E. M.; Rasmussen, S. G. F.; Loland,
C. J.; Schoenenberger, B.; Gether, U.; Newman, A. H. Rhodamine-
labeled 2b-carbomethoxy-3b-(3,4-dichlorophenyl)tropane analogues
as high-affinity fluorescent probes for the dopamine transporter. J.
Med. Chem. 2005, 48, 7513-7516.
(16) Tahtaoui, C.; Parrot, I.; Klotz, P.; Guillier, F.; Galzi, J.-L.; Hibert,
M.; Ilien, B. Fluorescent pirenzepine derivatives as potential bitopic
ligands of the human M1 muscarinic receptor. J. Med. Chem. 2004,
47, 4300-4315.
(17) Zaheer, A.; Wheat, T. E.; Frangioni, J. V. IRDye78 conjugates for
near-infrared fluorescence imaging. Mol. Imaging 2002, 1, 354-
364.
(18) Cross, P. E.; Arrowsmith, J. E.; Thomas, G. N.; Gwilt, M.; Burges,
R. A.; Higgins, A. J. Selective class III antiarrhythmic agents. 1.
Bis(arylalkyl)amines. J. Med. Chem. 1990, 33, 1151-1165.
(19) Diaz, G. J.; Daniell, K.; Leitza, S. T.; Martin, R. L.; Su, Z.;
McDermott, J. S.; Cox, B. F.; Gintant, G. A. The [3H]dofetilide
binding assay is a predictive screening tool for hERG blockade and
proarrhythmia: comparison of intact cell and membrane preparations
and effects of altering [K+]. J. Pharmacol. Toxicol. Methods 2004,
50, 187-199.
(20) Liu, H.; Ji, M.; Luo, X.; Shen, J.; Huang, X.; Hua, W.; Jiang, H.;
Chen, K. New p-Methylsulfonamido phenylethylamine analogues as
class III antiarrhythmic agents: design, synthesis, biological assay,
and 3D-QSAR analysis. J. Med. Chem. 2002, 45, 2953-2969.
(21) Deacon, M.; Singleton, D.; Szalkai, N.; Pasieczny, R.; Peacock, C.;
Price, D.; Boyd, J.; Boyd, H.; Steidl-Nichols, J.; Williams, C. Early
evaluation of compound QT prolongation effects: a predictive 384-
well fluorescence polarisation binding assay for measuring hERG
blockade. J. Pharmacol. Toxicol. Methods 2007, 55, 238-247.
(22) Cheng, Y.-C.; Prusoff, W. H. Relation between the inhibition constant
K1) and the concentration of inhibitor which causes fifty per cent
inhibition (I50) of an enzymic reaction. Biochem. Pharmacol. 1973,
22, 3099-3108.
(23) Hamill, O. P.; Marty, A.; Neher, E.; Sakmann, B.; Sigworth, F. J.
Improved patch-clamp techniques for high-resolution current record-
ing from cells and cell-free membrane patches. Eur. J. Physiol. 1981,
391, 85-100.
(24) Veldkamp, M. W.; Van Ginneken, A. C. G.; Bouman, L. N. Single
delayed rectifier channels in the membrane of rabbit ventricular
myocytes. Circ. Res. 1993, 72, 865-878.
(25) Troutman, M. D.; Thakker, D. R. Novel experimental parameters to
quantify the modulation of absorptive and secretory transport of
compounds by P-glycoprotein in cell culture models of intestinal
epithelium. Pharm. Res. 2003, 20, 1210-1224.
(26) Hamilton, G.; Cosentini, E. P.; Teleky, B.; Koperna, T.; Zacheri, J.;
Riegler, M.; Feil, W.; Schiessel, R.; Wenzi, E. The multidrug-
resistance modifiers verapamil, cyclosporine A and tamoxifen induce
an intracellular acidification in colon carcinoma cell lines in vitro.
Anticancer Res. 1993, 13, 2059-2063.
(27) Fiser, A.; Sali, A. MODELLER: generation and refinement of
homology-based protein structure models. Methods Enzymol. 2003,
374, , 461-491 (Macromolecular Crystallography, Part D).
(28) Doyle, D. A.; Cabral, J. M.; Pfuetzner, R. A.; Kuo, A.; Gulbis, J.
M.; Cohen, S. L.; Chait, B. T.; MacKinnon, R. The structure of the
potassium channel: molecular basis of K+ conduction and selectivity.
Science (Washington, D.C.) 1998, 280, 69-77.
(29) Laskowski, R. A.; McArthur, M. W.; Moss, D. S.; Thornton, J. M.
PROCHECK: A program to check the stereochemical quality of
protein structures. J. Appl. Crystallogr. 1993, 26, 283-291.
(30) Lees-Miller, J. P.; Duan, Y.; Teng, G. Q.; Duff, H. J. Molecular
determinant of high-affinity dofetilide binding to HERG1 expressed
in Xenopus oocytes: involvement of S6 sites. Mol. Pharmacol. 2000,
57, 367-374.
(31) Jones, G.; Willett, P.; Glen, R. C.; Leach, A. R.; Taylor, R.
Development and validation of a genetic algorithm for flexible
docking. J. Mol. Biol. 1997, 267, 727-748.
(32) Zhang, J.-H.; Chung, T. D. Y.; Oldenburg, K. R. A Simple statistical
parameter for use in evaluation and validation of high throughput
screening assays. J. Biomol. Screening 1999, 4, 67-73.
(33) Perry, M.; De Groot, M. J.; Helliwell, R.; Leishman, D.; Tristani-
Firouzi, M.; Sanguinetti, M. C.; Mitcheson, J. Structural determinants
of HERG channel block by clofilium and ibutilide. Mol. Pharmacol.
2004, 66, 240-249.
(34) Perry, M.; Stansfeld, P. J.; Leaney, J.; Wood, C.; de Groot, M. J.;
Leishman, D.; Sutcliffe, M. J.; Mitcheson, J. S. Drug binding
interactions in the inner cavity of hERG channels: molecular insights
from structure-activity relationships of clofilium and ibutilide
analogs. Mol. Pharmacol. 2006, 69, 509-519.
(35) Pearlstein, R. A. V.; Roy, J.; Kang, J.; Chen, X.-L.; Preobrazhenskaya,
M.; Shchekotikhin, A. E.; Korolev, A. M.; Lysenkova, L. N. M.;
Olga, V.; Hendrixa, J.; Rampea, D. Characterization of HERG
potassium channel inhibition using CoMSiA 3D QSAR and homology
modeling approaches. Bioorg. Med. Chem. Lett. 2003, 13, 1829-
1835.
(36) Rajamani, R.; Tounge, B. A.; Li, J.; Reynolds, C. H. A two-state
homology model of the hERG K+ channel: application to ligand
binding. Bioorg. Med. Chem. Lett. 2005, 15, 1737-1741.
(37) Oesterberg, F.; Aqvist, J. Exploring blocker binding to a homology
model of the open hERG K+ channel using docking and molecular
dynamics methods. FEBS Lett. 2005, 579, 2939-2944.
(38) Rajamani, R. T.; Brett, A.; Li, J.; Reynolds, C. H. A two-state
homology model of the hERG K+ channel: application to ligand
binding. Bioorg. Med. Chem. Lett. 2005, 15, 1737-1741.
(39) More, J. D.; Finney, N. S. A simple and advantageous protocol for
the oxidation of alcohols with o-iodoxybenzoic acid (IBX). Org. Lett.
2002, 4, 3001-3003.
(40) Greengrass, P. M.; Stewart, M.; Wood, C. M. Assay to establish the
affinity of compounds at the ERG potassium channel, and use in
identification of compounds with undesirable cardiac side effects.
WO2003021271 A2, 2003.
(41) Zhou, Z.; Gong, Q.; Ye, B.; Fan, Z.; Makielski, J. C.; Robertson, G.
A.; January, C. T. Properties of HERG channels stably expressed in
HEK 293 cells studied at physiological temperature. Biophys. J. 1998,
74, 230-241.
JM0700565