Strategies to reduce hERG K+ channel blockade. Exploring heteroaromaticity and rigidity in novel pyridine analogues of dofetilide

Article abstract of DOI:10.1021/jm301564f

Drug-induced blockade of the human ether-a-go-go-related gene K⁺ channel (hERG) represents one of the major antitarget concerns in pharmaceutical industry. SAR studies of this ion channel have shed light on the structural requirements for hERG

Full text of DOI:10.1021/jm301564f

Article
pubs.acs.org/jmc
Strategies To Reduce hERG K+ Channel Blockade. Exploring
Heteroaromaticity and Rigidity in Novel Pyridine Analogues of
Dofetilide
Joao F. S. Carvalho, Julien Louvel, Maarten L. J. Doornbos, Elisabeth Klaasse, Zhiyi Yu, Johannes Brussee,
̃
and Adriaan P. IJzerman*
Leiden Academic Center for Drug Research, Leiden University, P.O. Box 9502, 2300 RA Leiden, The Netherlands
S
* Supporting Information
ABSTRACT: Drug-induced blockade of the human ether-a-go-go-related gene K+ channel
(hERG) represents one of the major antitarget concerns in pharmaceutical industry. SAR
studies of this ion channel have shed light on the structural requirements for hERG
interaction but most importantly may reveal drug design principles to reduce hERG affinity.
In the present study, a novel library of neutral and positively charged heteroaromatic
derivatives of the class III antiarrhythmic agent dofetilide was synthesized and assessed for
hERG affinity in radioligand binding and manual patch clamp assays. Structural
modifications of the pyridine moiety, side chain, and peripheral aromatic moieties were evaluated, thereby revealing approaches
for reducing hERG binding affinity. In particular, we found that the extra rigidity imposed close to the positively charged pyridine
moiety can be very efficient in decreasing hERG affinity.
We have taken that approach previously and were successful in
defining criteria for what a molecule should not have in order to
display low or negligible affinity for the hERG channel.4,5 In
this particular case, we focused on the linker between the two
aromatic moieties of dofetilide. Flexibility and a basic aliphatic
amine in the center of the linker have been suggested as
prominent characteristics in the hERG pharmacophore,6 which
we took as the starting point in the current study.
However, there are a number of hERG blockers which do
not contain a positively charged nitrogen center,7 and this
apparent contradiction was explored by substituting the
aliphatic amine with an aromatic pyridine moiety. The linker’s
flexibility was challenged by introducing rigidity in that part of
the molecule. This synthetic approach yielded over 70 final
products that were tested in a competition binding assay with
[3H]astemizole as the radioligand. A few of these were further
evaluated in a patch clamp assay. From these data we derived
chemical alerts and “do’s and don’ts” for reduced hERG affinity.
INTRODUCTION
■
Among the vast family of ion channels are the voltage-gated
potassium channels (Kv), including the human ether-a-go-go-
related gene (hERG) K+ channel. Class III antiarrhythmic drugs
interact with this channel. As a result, the QT interval of the
cardiac cycle is prolonged, as observed in the electrocardio-
gram. However, nonantiarrhythmic drugs and their metabolites
may also interact with the hERG channel, thereby increasing
the risk of ventricular arrhythmia and fibrillation and sometimes
culminating in Torsades de pointes and sudden death.1,2 Drugs
such as astemizole, cisapride, sertindole, and terfenadine have
been withdrawn from most markets because of these unwanted
cardiac side effects. Currently, hERG safety testing is an
indispensable check in the drug discovery process and is
required by the FDA.3
In the present study we synthesized many derivatives of
dofetilide (Figure 1), a potent class III antiarrhythmic agent.
We reasoned that such a repertoire of congeneric ligands could
provide chemical alerts for hERG potassium channel toxicity.
RESULTS AND DISCUSSION
■
Chemistry. Several pyridine derivatives of dofetilide were
synthesized as outlined in Schemes 1−3, with the aim of
studying their ability to interact with the hERG channel and
ultimately finding new chemical modifications that decrease
hERG affinity. As depicted in Figure 1, and as previously shown
by our group,4 the p-methylphenyl derivative of dofetilide (2)
displays about the same hERG binding affinity as dofetilide
itself (1). Therefore, pyridine analogues of compound 2 were
initially designed.
Figure 1. Schematic representation of the rationale behind
heteroaromatic compound design. A: triple bonds; B: oxygen; n:
chain length; R: CH3, H, OMe, Cl, phenyl; R1: CH3; R2: H or NH2.
Received: October 24, 2012
Published: March 11, 2013
© 2013 American Chemical Society
2828
dx.doi.org/10.1021/jm301564f | J. Med. Chem. 2013, 56, 2828−2840

Journal of Medicinal Chemistry
Article
a
Scheme 1. Synthesis of Asymmetric and Symmetric 4-Methylphenyl Pyridine Derivatives of Dofetilide
a
(a) POCl3, reflux, overnight; (b) 4-methylbenzylzinc, Pd(PPh3)4, THF, N2, 60 °C, overnight; (c) methyl p-toluenesulfonate, CH3CN, reflux,
overnight; (d) methyl trifluoromethanesulfonate, CH2Cl2, 0 °C → room temp, overnight; (e) p-cresol or phenol, K2CO3, CH3CN, reflux, N2,
overnight; (f) (4-methylphenyl) alcohol, NaH, THF, 0 °C → reflux, N2, overnight. For simplification purposes, positively charged pyridines 11a,
11b, and 13 are represented without the trifluoromethane sulfonate counterion, and 12a−c without the p-toluenesulfonate counterion.
Scheme 2. Synthesis of Pyridines Bearing Side Chains with Different Degrees of Rigidity and Different Peripheral Aromatic
a
Substituents
a
(a) Alkyne, PdCl2(PPh3)2, CuI, Et3N, or Et3N/CH3CN, room temp or reflux; (b) methyl trifluoromethanesulfonate, CH2Cl2, 0 °C → room temp,
overnight; (c) H2, Pd/C, THF/MeOH, room temp. Positively charged pyridines are represented without the trifluoromethane sulfonate counterion.
Compounds 5a and 5b were first designed (Scheme 1), with
the aim to understand the biological consequences of changing
the pKa of the basic amino compound 2 by introducing a
pyridine moiety. The synthetic route was started by preparing
the N-oxide 38 (see Supporting Information for full synthetic
details of all compounds). Chlorination of compound 3 in
POCl3 at reflux temperature afforded the ortho-chlorinated
product 4a and the para-chlorinated product 4b in moderate
yields (39% and 28%, respectively). Negishi cross-coupling of
the pyridine chlorides 4a and 4b with 4-methylbenzylzinc,
2829
dx.doi.org/10.1021/jm301564f | J. Med. Chem. 2013, 56, 2828−2840

Journal of Medicinal Chemistry
Article
Scheme 3. Synthesis of p-Aminopyridines, Asymmetric Substituted Pyridines, and Aromatic Derivatives of 2,6-
a
Diphenethylpyridine 17b
a
(a) 1-Methyl-4-(prop-2-yn-1-yloxy)benzene, PdCl2(PPh3)2, CuI, Et3N/CH3CN, reflux; (b) H2, Pd/C, THF/MeOH, room temp; (c) 1-methyl-4-
(prop-2-yn-1-yloxy)benzene, PdCl2(PPh3)2, CuI, Et3N, room temp; (d) H2, PtO2, EtOH, Et3N, room temp; (e) methyl trifluoromethanesulfonate,
CH2Cl2, 0 °C → room temp; (f) 4-ethynylbenzene, PdCl2(PPh3)2, CuI, Et3N, room temp. Positively charged pyridines are represented without the
counterion for simplification purposes. pKa values of pyridines 19 and 20 were estimated using Marvin software.9
catalyzed by Pd(PPh3)4, gave the desired ortho- and para-
substituted 4-methylbenzylpyridines 5a and 5b in good yields
(92% and 75%, respectively).
PdCl2(PPh3)2 and CuI and an excess of Et3N) at reflux
(CH3CN) or room temperature. The flexible and neutral
pyridines 17a,b,i−p were obtained in moderate to good yields
by hydrogenation of the rigid pyridines 15a,b,i−p. The rigid or
flexible positively charged pyridines 16a−p or 18a,b,i−p,
respectively, were prepared by methylation of the correspond-
ing neutral pyridines using methyl trifluoromethanesulfonate,
yielding moderate to high yields.
Next we explored how asymmetric rigidification and basicity
of the pyridine moiety affect hERG interactions (Scheme 3).
The p-aminopyridine compounds 19 and 20 were obtained
starting from Sonogashira coupling of the commercially
available dibromo-p-aminopyridine 14b, which gave the acidic
compound 19 in low yield, followed by efficient hydrogenation
allowing access to the basic pyridine 20 (pKa ∼ 9.17, estimated
using Marvin software),9 in high yield. The positively charged
hybrid pyridine 24, bearing both a flexible and a rigid side
chain, was prepared by hydrogenation of compound 21 using
PtO2 under basic conditions, subsequent Sonogashira coupling
with 1-methyl-4-(prop-2-yn-1-yloxy)benzene yielding com-
pound 23, and finally methylation of the latter compound
using methyl trifluoromethanesulfonate. Compound 27, lacking
the nitrogen on the central aromatic moiety, was prepared by
Sonogashira coupling of 1,3-diiodobenzene (25) with ethynyl
benzene and subsequent hydrogenation of the resulting
compound 26, as shown in Scheme 3.
Compounds 6a and 6b were prepared by methylation of the
pyridines 5a and 5b, in order to study the effect of the positive
charge on the dofetilide pyridine derivatives. Methylation was
achieved using two different methylating agents, specifically
methyl p-toluenesulfonate or methyl trifluoromethanesulfonate.
A new set of dofetilide derivatives with symmetric, neutral,
and positively charged pyridines (8−13), bearing extended side
chains, was prepared by alkylation of the 2,6-dibromomethyl
pyridine (7) with the corresponding alcohol using basic
conditions, and subsequent methylation of the pyridine using
one of the two methods described above. Compounds 8a and
8b were prepared very efficiently (yields >90%) by reaction of
the corresponding aromatic alcohol using an excess of K2CO3
in CH3CN at reflux temperature. Compounds 9a−c and 10
were obtained by deprotonation of the corresponding aliphatic
alcohol with NaH at 0 °C for 30 min and subsequent overnight
stirring in the presence of the dibromide 7 at reflux temperature
(THF), resulting in a very good yield (>75%).
The next step in our synthetic plan (Scheme 2) was
performed to evaluate the influence of rigidity and aromatic
modification on hERG affinity, for positively charged
methylated pyridines. The rigid and neutral pyridines 15a−p
were synthesized in good yields through cross-coupling
reactions of 2,6-dibromopyridine 14a with the corresponding
alkyne, using Sonogashira conditions (catalytic amounts of
Biology. Radioligand Binding Experiments. In the present
study, 74 aromatic derivatives of dofetilide were evaluated in
2830
dx.doi.org/10.1021/jm301564f | J. Med. Chem. 2013, 56, 2828−2840

Journal of Medicinal Chemistry
Article
vitro for the ability to displace radiolabeled [3H]astemizole
from hERG K+ channels stably expressed on the plasma
membranes of HEK293 cells. In our biological approach,
compounds were initially screened at 10 μM for [3H]astemizole
displacement in membranes of HEK293 cells stably expressing
the hERG K+ channel. Those capable of displacing more than
50% of the [3H]astemizole at that concentration were then
selected for generating concentration−effect curves (using at
least 12 dilutions), and IC50 values were determined.
compounds 17b (Scheme 2, IC50 ∼ 1.6 μM) and 17m (IC50 ∼
9.9 μM), none of the other neutral pyridines evaluated in this
study displaced [3H]astemizole binding by more than 50% at
10 μM (IC50 > 10 μM).
This observation addressed one of our fundamental
questions regarding the possibility that compounds containing
a central and neutral pyridine moiety could be effective hERG
blockers. Indeed, it is known that a central basic amine is not
always a requirement for hERG blockade,7 and a recent study
suggests a perpendicular binding mode for hERG blockers.10
This new model is supported by the 4-fold symmetry of the
tetrameric channel, containing four flexible Tyr 652 residues, in
combination with the polarity of the selectivity filter.
Presumably, the affinity of pyridine analogues of dofetilide
could result from π−π stacking interactions between the
peripheral phenyl rings of the blocker and Tyr 652 of the
channel, in combination with hydrogen bonds between the
pyridine central moiety and the OH groups of Thr 623 and Ser
624. However, our biological results reveal that substituting a
basic nitrogen with a pyridine central moiety can be regarded as
a safe strategy to avoid potent hERG binding. The reduction in
hERG affinity may result from the drastic decrease in pKa and/
or from the increased rigidity and bulkiness imposed by this
modification. To better understand the reason for such a drastic
decrease, positively charged pyridines were synthesized.
Positively Charged Pyridines. Positively charged pyridine
compounds have been recently synthesized with a wide range
of biological effects.11−14 Nonetheless, scarce information is
available in the literature regarding the ability of such pyridines
to interact with the hERG channel.
As shown in Figure 2, the concentration−effect curves of
dofetilide (1) and the pyridine derivatives (12c and 18m)
Figure 2. Representative displacement curves of specific [3H]-
astemizole binding to HEK293 cell membranes stably expressing the
4
hERG K+ channel by dofetilide (1, , dashed bold line) and two
●
positively charged pyridine analogues: compound 12c (Δ, solid line)
○
and compound 18m ( , solid line). Curves are best described by a
two-site model, representing a high- and low-affinity binding site.
Changing the pKa, bulkiness, and rigidity of the potent hERG
blocker, by introducing a pyridine moiety between the p-
methylphenyl groups of dofetilide derivative 2 (Figure 1),
reveal two affinity sites when displacing [3H]astemizole with
increasing concentrations of the unlabeled (“cold”) ligand. This
biphasic behavior in [3H]astemizole binding studies is a more
general phenomenon, although not observed with all
compounds tested. It has been previously described by our
group,4,5 not only for dofetilide4 but also for the experimental
drug E-4031.5 Both studies showed that the IC50 value for the
high-affinity site is highly correlated with the IC50 value
determined by whole-cell voltage clamp measurements. Thus,
in the present study, the reported IC50 values represent the
high-affinity site found for compounds with biphasic behavior
(Tables 1−4).
For a few selected compounds, we eventually performed
manual patch clamp experiments using the same HEK cells and
compared the potencies in that assay with the affinities
determined in the radioligand binding experiments (Table 5).
Structure−Affinity Relationships. Our biological study
started by focusing on pyridine analogues of the p-
methylphenyl derivative of dofetilide 2 (Table 1: symmetric
and asymmetric analogues). Later we moved on to derivatives
containing different substituents in the peripheral aromatic
groups (Table 2: short-chained analogues, and Table 3:
extended side-chained pyridine derivatives). Table 4 outlines
hERG affinities of short-chained pyridines bearing phenyl
substituents in order to discuss hERG activity of neutral
pyridine 17b. Finally, in Table 5 a comparison between the
hERG affinity found for selected compounds by the radioligand
binding assay on cell membranes and the manual patch clamp
assay on intact cells is presented.
results in a large decrease in the hERG binding affinity (IC50
10 μM) as demonstrated by compounds 5a and 5b (Scheme 1,
9% 10 and −3%
5 [3H]astemizole displacement,
>
respectively, at 10 μM). Additionally, the position of the
nitrogen in the pyridine moiety does not affect the biological
outcome for these neutral pyridines. However, when positively
charged, distinct biological behavior was noticed. While the 2,6-
disubstituted pyridine 6a did not interact with the hERG
channel (Table 1, asymmetric pyridines, IC50 > 10 μM), the
2,4-disubstituted pyridine 6b bound effectively to the hERG
channel (IC50 = 346 125 nM, data not shown in the table).
These results demonstrate that introducing rigidity can affect
hERG affinity, yet the biological outcome is highly dependent
on the position of the positively charged nitrogen in the
aromatic moiety. Most probably, this difference in hERG
activity observed for the two analogues 6a and 6b resulted from
the combination of the extra rigidity imposed by the pyridine
moiety with the extra hindrance imposed on the positive charge
by the two relatively short side-chained ortho substituents.
These experiments demonstrate the importance of rigidity on
hERG affinity when a central positive charge is present in the
blocker. Additionally, the position of the positive charge on the
pyridine can influence activity, and ortho-disubstituted
methylated pyridines might be a solution to overcome the
hERG affinity.
Having established that a positively charged pyridine
analogue of dofetilide can effectively be accommodated by
the hERG channel, we discuss below the parameters affecting
hERG affinity of derivatives containing a central methylated
pyridine moiety.
Effect of Central Pyridine Moiety. Out of the 72
heteroaromatic molecules tested, 36 compounds contained a
central neutral pyridine moiety, and with the exceptions of
2831
dx.doi.org/10.1021/jm301564f | J. Med. Chem. 2013, 56, 2828−2840

Journal of Medicinal Chemistry
Article
a
Table 1. hERG Channel Affinities of Symmetric and Asymmetric Pyridines
a
High-affinity IC50 SEM or displacement (%) at 10 μM (n = 3) of [3H]astemizole binding to membranes of HEK293 cells stably expressing the
hERG K+ channel. Positively charged pyridines are represented without the counterion.
Side-Chain Length (Table 1, symmetric pyridines). If
flexibility is a limiting parameter for hERG interaction, then
increasing the side-chain length of the nonactive positively
charged pyridine 6a results in a more flexible molecule which
can hypothetically be more prone to hERG interaction. Indeed,
the biological results of extended positively charged pyridines
(11a, 12a−c) revealed that increasing flexibility can enhance
hERG binding. The positively charged pyridines, 11a, 12a−c,
were all active in the low nanomolar range (24 to 135 nM).
Interestingly, side-chain length is not directly proportional to
hERG affinity. Remarkably, compound 12c, with three carbons
between the oxygen and aromatic group, displayed the most
potent hERG affinity (24 nM, and is only about 6-fold less
active than dofetilide, 1).
into hERG affinity, allowing the molecule to assume optimal
position on the hERG binding site. Increasing side-chain
rigidity may, therefore, decrease hERG affinity.
To test the influence of rigidity in hERG affinity, triple bonds
were introduced at different positions of the side chain.
Introducing a triple bond close to the peripheral aromatic
groups, such as in the symmetric pyridine 13, only slightly
affected hERG binding affinity (1.7-fold loss, compared to
compound 12c). In contrast, when the triple bond was
introduced close to the pyridine moiety, a considerable
decrease in hERG affinity was observed (compare 16a, 16i,
and 16j with 18a, 18i, and 18j, respectively). The decrease in
hERG affinity is related to the length of the side chain, where
introducing triple bonds was less effective in reducing hERG
affinity of shorter side-chained compounds (two carbon side
chains, 12-fold decrease, compare compound 16a with 18a).
Quite remarkably, introducing triple bonds close to the
Side-Chain Rigidity (Table 1). Several hERG blockers
described in the literature contain flexible linkers joining the
different fragments of the molecule. This flexibility translates
2832
dx.doi.org/10.1021/jm301564f | J. Med. Chem. 2013, 56, 2828−2840

Journal of Medicinal Chemistry
Article
pyridine moiety in four-bonded side chains led to a strong
decrease in hERG affinity (>116-fold, compare compound 16i
with 18i). Increasing further flexibility by adding two more
carbons to the side chain allowed some recovery of hERG
affinity (∼44-fold decrease, compare compound 16j with 18j).
Generally, the maximum free plasma drug concentration
allowed for clinical use of a hERG blocker is at least 30-fold less
than its IC50,15−18 as this is considered the safety margin of a
hERG blocker drug. We therefore considered a 30-fold
decrease in hERG affinity as our threshold for desirability.
The introduction of triple bonds close to the pyridine moiety is
therefore a very efficient strategy to reduce hERG affinity for
compounds possessing side chains with more than two bonds.
Having established that introducing triple bonds in each side
chain can completely remove hERG affinity on molecules of a
particular size, our attention was driven toward the biological
relevance of a triple bond in only one of the side chains, as in
the asymmetric pyridine 24. This modification was still very
effective in hampering hERG affinity, as it displayed a more
than 30-fold decrease in hERG affinity (IC50 ∼ 2.9 μM), when
affinities for hERG observed in this study (same hERG affinity
as compound 18m, Table 3, discussed later). These experi-
ments show that aromaticity is not essential for potent binding
affinity, suggesting that van der Waals interactions can be as
effective as π−π stacking interactions. In fact, these results are
consistent with the evidence for a lipophilic binding site in the
hERG cavity.24 van der Waals interactions with the side chain
of Phe656 were better correlated with the potency of hERG
blockage than π−π stacking interactions for several structurally
diverse drugs. Moreover, our results are also in agreement with
the large body of literature showing a relationship between
lipophilicity and potent hERG affinity.19 Additionally, compar-
ing the hERG affinity of compounds 16p and 18p reveals once
more the ability to reduce hERG affinity by triple bonds close
to the positively charged pyridine (hERG affinity decreases
>116-fold for aromatic side chains and about 133-fold for
aliphatic side chains).
p-Amino Substituent on the Pyridine Moiety (Table 1, p-
aminopyridines). Quaternary pyridine compounds, because of
their permanently positive charge, are pharmacokinetically
problematic, particularly when it comes to drug absorption.
Alternatively, p-aminopyridine compounds can be used.
Resonance stabilization of the protonated nitrogen on the
pyridine moiety renders p-aminopyridines basic and thus
positively charged at physiological pH. Because the p-
aminopyridine moiety is also present in its nonionized form,
it can cross membranes25,26 and thus potentially induce hERG
blockade. Indeed, p-aminopyridine itself is known to block
voltage-activated K+ channels,25,26 and quite recently, p-
aminopyridine-containing compounds designed to target the
neuropeptide Y Y1 receptor27 or the δ-opioid receptor28 were
found to display potent hERG affinity.
By comparing the hERG affinity of the basic p-aminopyridine
derivative 20 (estimated pKa ∼ 9.2) with the quaternary
pyridine 18i, we found that the p-amino substituent is 2-fold
less active. This difference in hERG affinity may result from the
increased hydrophilicity and/or bulkiness introduced on the
para-position of the pyridine moiety.
Introduction of triple bonds close to the p-aminopyridine
moiety increases rigidity and substantially decreases the pKa
(pKa < 7.4), resulting in predictable loss of hERG affinity, as
confirmed by compound 19 (IC50> 10 μM).
Taking into consideration the knowledge gained with the
asymmetric pyridine 24, where introduction of a single triple
bond reduced hERG activity (>30-fold), one can speculate that
introducing a triple bond in one of the side chains of
compound 20 would also reduce hERG affinity efficiently,
without interfering massively with the positive charge of the p-
aminopyridine moiety. The estimated pKa value of such a
compound is 7.9 using Marvin software. Hypothetically, when
designing a molecule for which a positive charge is crucial for
desired activity, introducing a p-aminopyridine moiety con-
jugated with a single triple bond might allow a transiently
positively charged compound to reach the inside of the cell,
without displaying considerable hERG affinity or bioavailability
problems.
Peripheral Aromatic Substituents. Introducing triple bonds
close to the positive charge of pyridines leads to a very different
hERG outcome, which varies according to the length of the side
chain. Consequently, different chemical modifications on the
peripheral aromatic groups were introduced to study the
influence on hERG activity of the resulting molecules.
compared to the saturated side-chained analogue 18i (IC50
∼
86 nM).
Several strategies to circumvent hERG blockade have been
reported in the literature using pharmacologically active
compounds that target different receptors.19 Based on the
classical model for hERG binding,20−22 modulation of pKa
recurs frequently as a satisfactory approach to reduce hERG
affinity.19 However, disruption of the nitrogen’s basicity can
also lead to decreased binding affinity for the desired target.
This feature is particularly relevant for compounds where
cation−π interactions are important for binding, such as
aminergic GPCR ligands.23 Here, we found a novel chemical
modification, comprising a methylated pyridine conjugated with
triple bonds, which allows substantial reduction of hERG
affinity without interfering with the positive charge of the
molecule.
Presence of Oxygens in the Side Chain (Table 1,
symmetric pyridines). Comparing the hERG affinity of
compound 18a with the oxy analogue compound 11a reveals
that the presence of oxygens in the side chain plays a role in
hERG binding. Specifically, approximately a 2-fold decrease in
hERG activity is observed by introducing oxygen atoms in the
side chains. Additionally, if we compare the activity of the long
side-chained pyridine 12b with analogue 18i, a decrease in
binding affinity was noticed when oxygens were moved toward
the positively charged pyridine moiety, indicating that the
location of the oxygen in the side chain also influences hERG
activity.
Furthermore, by comparing the hERG activity of compounds
11a, 16a, and 18a, rigidification of the side chains of short-
chained pyridines results in a 5-fold increase in efficiency
toward reducing hERG affinity compared with the introduction
of oxygens.
Peripheral Aromaticity (Table 1, aliphatic pyridines). The
literature indicates that π−π stacking interactions between the
blocker and the hERG binding pocket are important for potent
hERG affinity.10,20−22 To evaluate the importance of peripheral
aromaticity, compounds 16p and 18p, aliphatic analogues of
compounds 16i and 18i, were evaluated for hERG affinity.
Interestingly, positively charged pyridines bearing aliphatic side
chains were found to be about 5-fold more efficient in binding
the hERG channel than the corresponding aromatic derivatives.
Remarkably, compound 18p displayed one of the highest
2833
dx.doi.org/10.1021/jm301564f | J. Med. Chem. 2013, 56, 2828−2840

Journal of Medicinal Chemistry
Article
Short-Chain and Rigid Pyridines (Table 2). The short-
chained and rigid pyridine 16a, with 4-methylphenyl
substituents, has a remarkably high affinity for the hERG
channel (IC50 ∼ 285 nM). Removing the p-methyl substituent
leads to a 2-fold decrease in activity (compare compound 16a
with 16b). Most probably, this loss in hERG activity is due to
the resulting decrease in lipophilicity.
from the electron-withdrawing properties of the chloro
substituent, destabilizing the required π−π stacking inter-
actions. Alternatively, the less impressive decrease in hERG
activity observed for the p-methoxy analogue 16e can be
explained by a decrease in lipophilicity and/or increased
bulkiness of this substituent when compared to the p-methyl
group.
Additionally, introducing m- or o-chloro substituents on the
peripheral phenyl moieties causes further loss of hERG activity
(compounds 16g and 16h, about 14-fold decrease when
compared to compound 16a). Compounds 16g and 16h
display hERG activity similar to that of the cyclohexyl derivative
16c, which confirms the role of π−π stacking destabilization
mediated by the chloro substituent on the phenyl moieties.
These observations also highlight a positive effect of the p-
chloro substituent which counteracts its π−π stacking
destabilization features.
Table 2. hERG Channel Affinities of Rigid and Short-
ab
,
Chained Pyridines Bearing Different Substituents
Long-Chained Pyridines (Table 3). Contrary to the rigid and
short-chained compound 16a with 4-methylphenyl substitu-
ents, the long-chained analogue 16i was unable to target the
hERG channel (IC50> 10 μM). To explore how triple bonds in
long-chained and positively charged pyridines impede inter-
actions with hERG, other rigid analogues containing different
aromatic substituents, such as phenyl 16k, p-methoxy 16l, and
p-chloro 16m analogues and their corresponding flexible
derivatives, were evaluated. Results showed that rigid and
extended pyridines, in contrast with shorter pyridines (Table
2), were unable to significantly interact with the hERG channel
(IC50 > 10 μM). Once again, introducing triple bonds close to
the positively charged pyridine proved to be very efficient in
decreasing hERG affinity (>30-fold decrease), as depicted by
the chemical index parameter presented in Table 3. Indeed,
rigidity can decrease hERG activity from 67-fold (for phenyl
analogues) to 2173-fold (p-chloro analogues), clearly indicating
how effective this chemical transformation can be in reducing
hERG affinity.
Considering hERG affinity of the flexible and positively
charged pyridines, the p-chloro derivative 18m was the most
potent ligand (6-fold more potent than the p-methyl analogue
18i). In fact, it is the strongest hERG blocker found in this
study, with an IC50 value comparable to the extended and
aliphatic side-chained 18p (Table 1). In contrast with the short
and rigid analogues discussed above, where introduction of a p-
chloro substituent disrupted hERG activity, the introduction of
a p-chloro substituent on long-chained and flexible positively
charged pyridines enhances hERG binding. As discussed
previously, the electron-withdrawing properties of the chloro
substituent destabilize π−π stacking interactions and thus van
der Waals forces may be the prevalent interactions driving the
potent hERG affinity observed for long-chained compounds.
The 11-fold decrease in hERG activity of the p-methoxy
analogue 18l, when compared to the p-chloro derivative 18m,
may result from increased bulkiness and decreased lipophilicity
of the methoxy substituent.
a
Positively charged pyridines are represented without the counterion.
The corresponding rigid (15a−h) and flexible (17a,b) neutral
b
pyridines are not shown due to their negligible affinity for the hERG
channel at 10 μM, with the exception of compound 17b (see Table 4).
c
High-affinity IC50 SEM of [3H]astemizole binding to membranes of
d
HEK293 cells stably expressing the hERG K+ channel; Synthesis is
schematically presented in Scheme 2.
We have shown that peripheral aromaticity is not required
for potent hERG affinity of long side-chained and positively
charged pyridines (see IC50 of aliphatic pyridine 18p, Table 1).
The loss of hERG affinity of the aliphatic derivative 16c (8-fold
decrease when compared to the aromatic analogue 16b) shows
that, in contrast to long-chained derivatives, the peripheral
aromaticity of short-chained and rigid pyridines is crucial for
hERG affinity. The role of aromaticity on the hERG affinity of
short-chained and rigid pyridines is stronger than the
lipophilicity gained by substituting a phenyl ring with a
cyclohexyl group. Conversely, additional aromaticity introduced
by a biphenyl substituent, as in compound 16d, only slightly
improved hERG affinity (compare compounds 16b and 16d).
The fact that biphenyl compound 16d is as active as the 4-
methylphenyl analogue 16a suggests a negative effect from the
bulkiness of the second aromatic moiety.
The effect of electron-donating and -withdrawing groups on
hERG affinity was studied by preparing aromatic analogues
containing p-methoxy and p-chloro substituents (compounds
16e and 16f, respectively). Interestingly, both compounds were
less active than the p-methyl analogue 16a. Notably, the p-
chloro analogue 16f was about 4-fold less active than the p-
methyl analogue 16a. This decrease in hERG activity may result
The m- (16n and 18n) and o-choride analogues (16o and
18o) were also evaluated for hERG affinity. While the rigid
analogues (16n and 16o) were not active, as expected, the
activity of the flexible derivatives proved to be dependent on
the position of the chloro substituent. The hERG affinity was
decreased when the chloro atom moved away from the para-
position, being 5-fold less active on the ortho-position
(compound 18o) when compared to the p-chloro analogue
18m. Interestingly, the same trend was noticed for short-
2834
dx.doi.org/10.1021/jm301564f | J. Med. Chem. 2013, 56, 2828−2840

Journal of Medicinal Chemistry
Article
Table 3. hERG Channel Affinities of Positively Charged Pyridines Bearing Extended Side Chains with Different Aromatic
a b
,
Substituents
a
b
Positively charged pyridines are represented without the counterion. The corresponding neutral pyridines are not shown due to their nonaffinity
c
for the hERG channel at 10 μM; with the exception of compound 17m (IC50= 9857 239 nM). High-affinity IC50 SEM or displacement (%) at
d
10 μM (n = 3) of [3H]astemizole binding to membranes of HEK293 cells stably expressing the hERG K+ channel. Synthesis is schematically
e
f
presented in Scheme 2. Chemical index = IC50 of triple bond compound/IC50 of single bond compound. Compound 18n was insufficiently pure
(<80%).
chained and rigid derivatives (16f and 16h, Table 2), whereas
moving the chloro substituent from the para- to the ortho-
position led to an approximately 3-fold decrease in activity.
Interestingly, and corroborating how chloro substituents on
the phenyl moieties of extended pyridines impact hERG
affinity, we found that out of the 12 neutral derivatives of the
compounds outlined in Table 3, only the p-chloro and flexible
side-chained analogue 17m (Scheme 2) displayed some hERG
affinity (IC50 ∼ 9.9 μM). Quite remarkably, this neutral and
flexible pyridine 17m is 3-fold more active on hERG than the
positively charged and rigid analogue 16m. This example shows
that a neutral pyridine can be more active than a positively
charged pyridine and reveals how effective introducing triple
bonds can be in reducing hERG activity, because the resulting
positively charged pyridine is considerably less active than the
neutral analogue.
hERG Affinity of Neutral Pyridines (Table 4). In this study
we found a neutral and flexible pyridine capable of inducing
significant hERG blockade (Table 4, compound 17b, IC50 ∼ 1.6
μM). Interestingly, adding a 4-methyl substituent to the phenyl
groups, as in compound 17a (Scheme 2), hampered hERG
affinity (IC50 > 10 μM, data not shown). This exceptional
neutral pyridine 17b displays only 3.5-fold less activity than the
positively charged analogue 18b or 2-fold less activity than the
positively charged oxy analogue 11b. As far as we know, this is
the first report of a neutral compound with a central pyridine
moiety capable of displaying a considerable binding affinity
toward the hERG channel. Generally, this study shows that
replacing a basic nitrogen with a pyridine moiety results in
reduced hERG affinity. However, the significative hERG activity
of 17b reminds us that every general rule has an exception. This
result should keep medicinal chemists alert about the
introduction of pyridine moieties in look-a-like compounds.
The loss of activity (more than 6-fold) observed for
compound 27, the aromatic analogue of 17b, reveals the
nitrogen requirement for hERG interactions. Introducing triple
bonds in the side chain of both heteroaromatic and aromatic
rigid compounds 15b and 26, respectively, resulted in
nondisplacement of [3H]astemizole from the hERG channel
at 10 μM. Introducing oxygens in the side chain, such as in the
oxy analogue 8b, also proved to be as capable of reducing the
hERG affinity as rigidification (IC50> 10 μM). In this respect,
the introduction of oxygens in the side chains was much more
effective in neutral pyridines than in positively charged pyridine
derivatives. Although hERG affinity of neutral and positively
charged pyridines appear to be ruled by different interactions,
the same chemical modifications are effective for avoiding
hERG activity in positively charged and neutral pyridines.
Patch Clamp Studies (Table 5). Previous work performed
by our group showed that quaternary and tertiary amine
compounds can be accommodated by the hERG binding
pocket equally well.5 Results obtained with the radioligand
binding assay and the patch clamp assay were found to be quite
comparable,4,5 with the exception of quaternary amine
derivatives.5 In this regard, a considerable decrease in hERG
potency (up to 75-fold) was observed for permanently
positively charged compounds, as these compounds have
difficulty in passing through the cell membrane. Generally,
hERG blockers access the binding site inside the cell,29 so the
cell membrane becomes a barrier for hERG activity. This
feature can be explored to abrogate off-target hERG binding of
pharmacologically active compounds. In fact, chemical
modifications that limit membrane permeability, such as the
2835
dx.doi.org/10.1021/jm301564f | J. Med. Chem. 2013, 56, 2828−2840

Journal of Medicinal Chemistry
Article
whether a particular compound interacts with the hERG
channel.30
Table 4. hERG Channel Affinities of Short-Chained
Compounds Bearing Phenyl Substituents
a b
,
Using the radioligand binding assay, out of the four
compounds selected, the p-aminopyridine derivative 20 was
the least potent on the hERG channel. Conversely, in the patch
clamp experiments, compound 20 was the most active (IC50
=
146 nM). In fact, this compound showed comparable hERG
affinities in both methods. These results support our hypothesis
that a transiently positive charge allows compound 20 to cross
membranes easily. On the other hand, all the permanently
positive charged pyridines (compounds 12c, 13, and 18i)
showed significant loss of hERG activity (>5-fold) in the patch
clamp assay when compared to the respective radioligand
binding experiments. Specifically, the flexible pyridines 12c and
18i are more than 8-fold less active on hERG when compared
to the radioligand binding experiments, while rigidification of
the side chain distant from the positive charge, such as in
pyridine 13, enhances hERG affinity, suggesting that it may
facilitate membrane crossing.
In summary, radioligand binding experiments showed that
permanently positively charged pyridines interact very
efficiently with the hERG channel (low nanomolar range).
Patch clamp results, however, which better reflect the
consequences of in vivo hERG liability, reveal that these
compounds can be up to 11-fold less active when compared to
the values obtained with the radioligand binding assay.
Previously, we had set a threshold of 30-fold decrease in
activity for a chemical modification to be considered desirable.
By comparing the hERG activity obtained with radioligand
binding and patch clamp assays, we found that although a
positively charged pyridine can suppress affinity, it is not
enough to be considered a safe strategy by itself. Conversely,
considering that quaternary amine compounds showed up to
75-fold decreased hERG potency with the patch clamp assay5
while quaternary pyridines displayed only up to 11-fold
impairment suggests that positively charged pyridines are less
pharmacokinetically problematic (about 7-fold more cell
membrane permeable) than positively charged amines.
a
High-affinity IC50 SEM or displacement (%) at 10 μM (n = 3) of
[3H]astemizole binding to membranes of HEK293 cells stably
b
expressing the hERG K+ channel. Positively charged pyridines are
represented without the counterion.
formation of zwitterions, have already been explored as a
strategy to mitigate hERG liability.19 Our current results, using
a radiolabeled binding assay, reveal that flexible positively
charged pyridines can effectively fit in the hERG channel. To
investigate if permanently positively charged pyridine deriva-
tives can be a strategy to overcome hERG liability, we selected
three such compounds (12c, 13, and 18i) and a p-amino-
pyridine derivative (20) for manual patch clamp studies on
intact cells (Table 5). Manual patch clamping is considered the
most accurate, albeit laborious, functional method to unravel
Table 5. Comparison of hERG Affinity of Selected Compounds As Determined by the Radioligand Binding Assay on Cell
a
Membranes and the Manual Patch Clamp Assay on Intact Cells
a
Positively charged pyridines are represented without the counterion. Compounds with p-toluene sulfonate counterion: 12c. Compounds with
b
trifluoromethane sulfonate counterion: 13 and 18i. IC50 SEM determined by radioligand binding assay (RBA): displacement of [3H]astemizole in
membranes of HEK293 cells stably overexpressing the hERG channel. IC50 values, determined by patch clamp assay (PCA, n = 1). Assay index =
c
d
IC50 of PCA/IC50 of RBA.
2836
dx.doi.org/10.1021/jm301564f | J. Med. Chem. 2013, 56, 2828−2840

Journal of Medicinal Chemistry
Article
Figure 3. Graphical depiction of the proposed strategies to reduce the hERG K+ affinity based on SAR found with the radioligand binding assay.
Generally, introducing a central pyridine moiety reduces hERG affinity. When a positive charge is critical for the desired activity, methylation or p-
amino substitution of the pyridine moiety renders the molecule positive. Permanently positively charged pyridines showed substantial loss of hERG
activity (up to 11-fold) when evaluated with the patch clamp assay, in contrast to p-aminopyridines. Illustrated strategies for reduced hERG K+
affinity were found on positively charged pyridines. Introducing rigidity and oxygens in the side chain of a neutral pyridine also reduced hERG
affinity.
with longer side chains, which are less dependent on
aromaticity for interacting with the hERG channel. Rigid-
ification of short side-chained pyridines was less effective in
reducing hERG affinity than in extended side-chained pyridines.
In contrast, introduction of the electron-withdrawing chloro
substituent on the aromatic groups of short side-chained
pyridines was very efficient in mitigating hERG activity while
for more extended pyridines the presence of the chloro
substituent in peripheral aromatic moieties promoted stronger
hERG affinity. Nonetheless, both classes showed decreased
hERG affinity when the chlorine substituent was moved away
from the para-position.
Rigidifying a molecule reduces its conformational flexibility
and is often regarded in medicinal chemistry as an approach to
increase affinity and selectivity.36 Here we show that rigid-
ification of an aliphatic amine by introduction of a neutral
pyridine moiety promotes loss of hERG affinity. If positive
charge is essential for primary activity, the pyridine moiety can
be methylated without inducing hERG affinity for very short
compounds. For lengthy side-chained compounds, introducing
a positively charged pyridine moiety requires extra rigidity
(imposed by adjacent triple bonds) for reduction of hERG
interactions. Rigidification of only one of the side chains
resulted in a lowering of hERG affinity, and the conjugation of
this chemical modification with a p-aminopyridine moiety may
lead to a transiently charged compound with reduced hERG
affinity, which is able to cross cell membranes and thus be less
prone to bioavailability problems. Finally, the chemical features
found to reduce the affinity of potent positively charged
pyridine hERG blockers could also be applied to reduce hERG
affinity of neutral pyridines.
CONCLUSIONS
■
The pyridine moiety is found in a wide variety of biologically
active molecules,31−33 and, hence, in this study 72 aromatic
analogues of dofetilide containing a central pyridine moiety
were synthesized and evaluated for hERG affinity in a
radioligand binding assay using [3H]astemizole. Sixty nine of
these pyridine analogues of dofetilide are novel chemical
entities. This large collection of novel compounds allowed us to
demonstrate that substituting a basic nitrogen on an effective
hERG blocker with a pyridine moiety can be regarded as a
reliable measure to reduce potent hERG affinity. However, such
chemical modifications can also disrupt the interaction with the
desired pharmacological target, particularly for compounds in
which a positive charge determines the desired receptor
interaction. In this regard, we found that the potent hERG
activity of the basic dofetilide derivative 2 was efficiently
suppressed by introducing a positively charged pyridine, as
confirmed by compound 6a. However, we also demonstrate
that longer side-chained positively charged pyridines (perma-
nently: methylated pyridines, or transiently: p-aminopyridines)
can effectively bind to the hERG channel. Actually, in this study
we identified four positively charged pyridine compounds (12c,
18a, 18p, and 18m) with IC50 values ranging from 14 to 24
nM, as determined in radioligand binding assays, which are
almost equivalent to the most potent hERG blockers known to
date.34
In patch clamp assays, permanently positively charged
pyridines were less effective (up to 11-fold), indicating this
chemical modification as a strategy to reduce in vivo hERG
affinity. Alone, though, such modification may not be enough to
be considered a reliable approach. Thus, we searched for other
chemical modifications that can decrease the hERG affinity of
lengthy positively charged pyridines. We found that rigid-
ification and oxygenation of both side chains can affect hERG
binding affinity. In particular, the introduction of triple bonds
and oxygens close to the positive charge lead to a stronger
decrease in hERG affinity than distant modifications. Rigid-
ification proved to be much more effective in decreasing hERG
activity than oxygenation, in particular for more extended
pyridine analogues. Successful mitigation of hERG activity by
rigidification evidenced in this study is consistent with a recent
report on the synthesis of histone deacetylase inhibitors.35
Upon closer inspection we identified two classes of positively
charged pyridines, those with a short side chain in which hERG
interaction is dependent on peripheral aromaticity, and those
We believe these findings might have wide applications in the
development of new drugs where a low off-target hERG profile
is desired, and Figure 3 presents a schematic overview of the
major strategies to modulate hERG affinity based on SAR
found with the radioligand binding assay.
EXPERIMENTAL SECTION
■
Chemistry. General Methods. All commercially available chem-
icals were used without purification. Reactions were monitored by
TLC using silica gel 60 F254 aluminum sheets. Reaction yields were
determined by compound isolation through flash column chromatog-
raphy using silica gel 60 (230−400 mesh ASTM). H and 13C NMR
1
spectra were recorded on a Bruker AV400 spectrometer (1H NMR,
400 MHz; 13C NMR, 100 MHz) at ambient temperature and analyzed
with MestReNova software. Sample solutions were prepared in CDCl3,
MeOD, or mixtures of both deuterated solvents. Chemical shifts (δ)
2837
dx.doi.org/10.1021/jm301564f | J. Med. Chem. 2013, 56, 2828−2840

Journal of Medicinal Chemistry
Article
are given in ppm and coupling constants (J) in hertz. Trimethylsilane
132.50, 132.01, 130.27 (2C), 129.92 (2C), 129.09 (2C), 127.33,
126.77, 114.70 (2C), 65.38, 44.82, 41.03, 20.99, 20.44. HRMS (ESI)
m/z calcd for C33H21N+: 318.18518, found: 318.18524. HPLC purity:
99.6% (tR: 13.66 min).
was used as internal standard for calibrating chemical shift for 1H, 13
C
NMR spectroscopy. HRMS was recorded on a Thermo Finnigan LTQ
Orbitrap mass spectrometer. HPLC analysis, carried out on a C18
reversed-phase column (125 × 4.6 mm, particle size 5 μm, pore size
180 Å) at 25 °C coupled with a UV detector at 254 nm, was used to
determine the purity of final compounds. Final compounds were
dissolved in mixtures of CH3CN/H2O and eluted from the column at
a flow rate of 0.6 mL min−1, with a two-component system of
CH3CN/H2O. Purity was confirmed to be ≥95% for all compounds,
with the exception of compounds 10 (91%), 16i (93%), and 23
(94%). Compounds 17n and 18n could not be purified to a desired
standard (purity <80%) and the biological data obtained from these
compounds was only used qualitatively (see Table 4). With the
exception of compounds 8a,37 15b,38 17b,39 26,38 and 27,40 all
compounds synthesized and evaluated for hERG activity were
considered novel chemical entities using CrossFire Beilstein Database.
pKa values of pyridines 19 and 20 were estimated using Marvin
software (http://www.chemaxon.com/marvin/sketch/index.jsp).9
General Procedures. General Procedure for the Negishi Coupling
Reaction of Chloropyridines. A 0.5 M solution of 4-methylbenzylzinc
chloride in THF (20.7 mL, 10.356 mmol) was added dropwise to a
solution of 6-chloro-2-((p-tolyloxy)methyl)pyridine (4a, 605 mg,
2.589 mmol) in dry THF (50 mL) at 60 °C and under inert
atmosphere. The mixture was stirred overnight, and the reaction was
stopped by evaporation under vacuum. FCC (petroleum ether/ethyl
acetate, 60:1 to 20:1) afforded the desired product 2-(4-methyl-
benzyl)-6-((p-tolyloxy)methyl)pyridine (5a, 720 mg, 92%). Product
was further purified by hydrochloride salt formation using ethanolic
General Procedure for Alkylation with Aromatic Alcohols. To a
solution of p-cresol (1714 mg, 15.85 mmol) in acetonitrile (113 mL)
under inert atmosphere were added sequentially K2CO3 (4172 mg,
30.19 mmol) and 2,6-bis(bromomethyl)pyridine (7, 2000 mg, 7.55
mmol) at room temperature. The resulting mixture was then stirred at
reflux temperature overnight, and the reaction was stopped by
filtration and evaporation under vacuum. FCC (petroleum ether/ethyl
acetate, 20:1 to 10:1) afforded the desired 2,6-bis((p-tolyloxy)methyl)-
pyridine (8a, 2210 mg, 92%). 1H NMR (400 MHz, CDCl3) δ 7.70 (t, J
= 7.8 Hz, 1H), 7.43 (d, J = 7.8 Hz, 2H), 7.08 (d, J = 8.4 Hz, 4H), 6.88
(d, J = 8.3 Hz, 4H), 5.18 (s, 4H), 2.28 (s, 6H). 13C NMR (101 MHz,
CDCl3) δ 156.99 (2C), 156.22 (2C), 137.52, 130.34 (2C), 129.94
(4C), 119.98 (2C), 114.58 (4C), 70.57 (2C), 20.45 (2C). HRMS
(ESI) m/z calcd for C21H21NO2H+: 320.16451, found: 320.16452.
HPLC purity: 98.6% (tR: 15. 80 min).
General Procedure for Alkylation with Aliphatic Alcohols. p-
Tolylmethanol (338.6 mg, 2.77 mmol) was added to a suspension of
NaH (60%, 147.8 mg, 3,69 mmol) in dry THF (30 mL) at 0 °C under
inert atmosphere and the resulting mixture stirred for 30 min. Next, a
solution of 2,6-bis(bromomethyl)pyridine (7, 350 mg, 1.32 mmol) in
THF (30 mL) was added dropwise at 0 °C and then refluxed and
stirred overnight. The reaction was stopped by filtration through Celite
and evaporation under vacuum. The resulting crude product was
purified by FCC (petroleum ether/ethyl acetate, 10:1 to 4:1) and
afforded the pure 2,6-bis(((4-methylbenzyl)oxy)methyl)pyridine (9a,
1
1
435.6 mg, 95%). H NMR (400 MHz, CDCl3) δ 7.66 (t, J = 7.7 Hz,
HCl and recrystallized from a mixture of EtOH and Et2O. H NMR
1H), 7.37 (d, J = 7.7 Hz, 2H), 7.26 (d, J = 7.9 Hz, 4H), 7.14 (d, J = 7.9
Hz, 4H), 4.64 (s, 4H), 4.58 (s, 4H), 2.32 (s, 6H). 13C NMR (101
MHz, CDCl3) δ 157.84 (2C), 137.25, 137.06 (2C), 134.78 (2C),
128.95 (4C), 127.80 (4C), 119.76 (2C), 72.75 (2C), 72.65 (2C),
21.04 (2C). HRMS (ESI) m/z calcd for C23H25NO2H+: 348.19581,
found: 348.19578. HPLC purity: 98.6% (tR: 15. 59 min).Method A
General Procedure for the Sonogashira Coupling of Bromopyr-
idines PdCl2(PPh3)2 (148 mg, 0.21 mmol) and CuI (120.6 mg, 0.63
mmol) were added to a solution of 2,6-dibromopyridine (1000 mg,
4.22 mmol) in Et3N (10 mL) under N2, and the resulting mixture was
stirred for 30 min. Then 4-ethynyltoluene (1177 mg, 10.13 mmol) was
added dropwise and the reaction mixture stirred overnight at room
temperature. The reaction was stopped by addition of silica gel and
evaporated under vacuum. FCC (petroleum ether/ethyl acetate, 40:1
to 1:1) afforded the desired 2,6-bis(3-(p-tolyloxy)prop-1-yn-1-yl)-
pyridine product (15a, 1200 mg, 92%). 1H NMR (400 MHz, CDCl3)
δ 7.64 (t, J = 7.6 Hz, 1H), 7.49 (d, J = 7.2 Hz, 4H), 7.43 (d, J = 7.6 Hz,
2H), 7.16 (d, J = 7.2 Hz, 4H), 2.37 (s, 6H). 13C NMR (101 MHz,
CDCl3) δ 143.91 (2C), 139.30 (2C), 136.28, 131.97 (4C), 129.12
(4C), 125.92 (2C), 119.03 (2C), 89.87 (2C), 87.82 (2C), 21.54 (2C).
HRMS (ESI) m/z calcd for C23H17NH+: 308.14338; found:
308.14322. HPLC purity: 97.7% (tR: 17.64 min).Method B
PdCl2(PPh3)2 (140.4 mg, 0.2 mmol), CuI (114.3 mg, 0.6 mmol),
ethynylbenzene (2451 mg, 24 mmol), and Et3N (3036 mg, 30 mmol)
were added to a solution of 2,6-dibromopyridine (2369 mg, 10 mmol)
in CH3CN (20 mL) under N2. The reaction mixture was stirred
overnight at reflux temperature. The reaction was stopped by addition
of silica gel and evaporated under vacuum. FCC (petroleum ether/
ethyl acetate, 20:1 to 3:1) afforded the desired 2,6-bis(phenylethynyl)-
pyridine product (15b, 2486 mg, 89%).1H NMR (400 MHz, CDCl3) δ
7.66−7.60 (m, 5H), 7.45 (d, J = 7.7 Hz, 2H), 7.36 (s, 6H). 13C NMR
(101 MHz, CDCl3) δ 143.68 (2C), 136.36, 132.00 (4C), 129.01 (2C),
128.30 (4C), 126.14 (2C), 121.99 (2C), 89.56 (2C), 88.22 (2C).
HRMS (ESI) m/z calcd for C21H13NH+: 280.11208, found:
280.11188. HPLC purity: 100.0% (tR: 15.44 min).
(400 MHz, CDCl3) δ 8.20 (t, J = 7.8 Hz, 1H), 7.93 (d, J = 7.8 Hz,
1H), 7.44 (d, J = 7.8 Hz, 1H), 7.31 (d, J = 7.8 Hz, 2H), 7.17 (d, J = 7.8
Hz, 2H), 7.11 (d, J = 7.8 Hz, 2H), 6.95 (d, J = 8.1 Hz, 2H), 5.78 (s,
2H), 4.67 (s, 2H), 2.34 (s, 3H), 2.30 (s, 2H). 13C NMR (101 MHz,
CDCl3) δ 157.10, 154.74, 153.25, 145.03, 137.61, 131.58 (2C), 130.18
(2C), 129.93 (2C), 129.29 (2C), 125.15, 121.93, 114.57 (2C), 64.62,
38.10, 21.02, 20.42. HRMS (ESI) m/z calcd for C21H21NOH+:
304.16959, found: 304.16926. HPLC purity: 100.0% (tR: 15.70 min).
General Procedure for the Methylation of Pyridines Using Methyl
p-Toluenesulfonate. Methyl p-toluenesulfonate (906 μL, 6 mmol)
was added to a solution of 2-(4-methylbenzyl)-6-((p-tolyloxy)methyl)-
pyridine (5a, 303.4 mg, 1 mmol) in acetonitrile (15 mL) under inert
atmosphere and the resulting mixture stirred at reflux temperature
during overnight. The reaction was stopped by evaporation under
vacuum and crystallization (petroleum ether/ethyl acetate) gave the
desired product 1-methyl-2-(4-methylbenzyl)-6-((p-tolyloxy)methyl)-
pyridin-1-ium 4-methylbenzenesulfonate (6a, 333 mg, 67%). 1H NMR
(400 MHz, CDCl3) δ 8.10 (t, J = 7.8 Hz, 1H), 7.89 (d, J = 7.8 Hz,
1H), 7.57 (d, J = 7.8 Hz, 2H), 7.37 (d, J = 7.8 Hz, 1H), 7.11 (d, J = 7.8
Hz, 2H), 7.12−6.88 (m, 6H), 6.89 (d, J = 7.8 Hz, 2H), 5.39 (s, 2H),
4.42 (s, 2H), 4.23 (s, 3H), 2.32 (s, 3H), 2.25 (s, 6H). 13C NMR (101
MHz, CDCl3) δ 159.21, 154.74, 153.97, 144.06, 143.42, 139.00,
137.52, 131.54, 130.39 (2C), 130.13 (2C), 129.96, 129.35 (2C),
128.44 (2C), 127.92, 125.70 (2C), 125.40, 114.76 (2C), 66.28, 40.43,
39.30, 21.12, 21.01, 20.39. HRMS (ESI) m/z calcd for C22H24NO+:
318.18524, found: 318.18488. HPLC purity: 95.6% (tR: 13.59 min).
General Procedure for the Methylation of Pyridines Using Methyl
Trifluoromethanesulfonate. Methyl trifluoromethanesulfonate (50
μL, 0.456 mmol) was added to a solution of 4-(4-methylbenzyl)-2-((p-
tolyloxy)methyl)pyridine (5b, 50 mg, 0.165 mmol) in dry DCM (5
mL) at 0 °C and under inert atmosphere. The mixture was stirred
overnight, and the reaction was stopped by pouring the mixture into
silica. FCC (dichloromethane/methanol, 60:1 to 30:1) afforded the
desired 1-methyl-4-(4-methylbenzyl)-2-((p-tolyloxy)methyl)pyridin-1-
1
ium trifluoromethanesulfonate product (6b, 70.2 mg, 91%). H NMR
(400 MHz, CDCl3) δ 8.76 (d, J = 6.4 Hz, 1H), 7.82 (d, J = 2.3 Hz,
1H), 7.59 (dd, J = 6.4, 2.3 Hz, 1H), 7.12 (d, J = 7.8 Hz, 2H), 7.08(d, J
= 8.3 Hz, 2H), 7.02 (d, J = 7.8 Hz, 2H), 6.87 (d, J = 8.3 Hz, 2H), 5.29
(s, 2H), 4.29 (s, 3H), 4.12 (s, 2H), 2.31 (s, 3H), 2.27 (s, 3H). 13C
NMR (101 MHz, CDCl3) δ 162.08, 154.68, 152.24, 146.80, 137.39,
General Procedure for Reduction of Pyridine Alkynes. To a
solution of 2,6-bis(3-(p-tolyloxy)prop-1-yn-1-yl)pyridine (15a, 400
mg, 1.3 mmol) in a mixture of THF (12 mL) and MeOH (12 mL) was
added 10% Pd on carbon (80 mg), and the atmosphere was changed
to H2 (2.5 bar). The resulting mixture was stirred in a Parr
2838
dx.doi.org/10.1021/jm301564f | J. Med. Chem. 2013, 56, 2828−2840

Journal of Medicinal Chemistry
Article
hydrogenation apparatus overnight at room temperature, followed by
filtration over Celite and evaporation under vacuum. FCC (petroleum
ether/ethyl acetate, 20:1) afforded the desired 2,6-bis(4-
or Induce Torsade de Pointes: An Overview. Drug Safety 2002, 25,
263−286.
(7) Aronov, A. M. Common Pharmacophores for Uncharged Human
Ether-a-go-go-Related Gene (hERG) Blockers. J. Med. Chem. 2006, 49,
6917−6921.
1
methylphenethyl)pyridine product (17a, 221.4 mg, 54%). H NMR
(400 MHz, CDCl3) δ 7.41 (t, J= 7.7 Hz, 1H), 7.09−7.04 (m, 8H),
6.86 (d, J = 7.7 Hz, 2H), 3.07 (ddd, J = 9.1, 6.1, 2.1 Hz, 4H), 3.00
(ddd, J = 9.1, 6.1, 2.1 Hz, 4H), 2.29 (s, 6H). 13C NMR (101 MHz,
CDCl3) δ 160.70 (2C), 138.48 (2C), 136.38, 135.12 (2C), 128.87
(4C), 128.29 (4C), 120.19 (2C), 40.07 (2C), 35.64 (2C), 20.93 (2C).
HRMS (ESI) m/z calcd for C23H25NH+: 316.20598, found:
316.20585. HPLC purity: 99.6% (tR: 13.99 min).
Biology. Radioligand Binding Experiments. Radioligand binding
studies were performed as recently described by our group.4,5
Patch Clamp Experiments. These experiments are detailed in the
Supporting Information.
́
(8) Carvalho, J. F. S.; Silva, M. M. C.; Sa e Melo, M. L. Highly
Efficient Epoxidation of Unsaturated Steroids Using Magnesium
Bis(monoperoxyphthalate) Hexahydrate. Tetrahedron 2009, 65, 2773−
2781.
(9) Manchester, J.; Walkup, G.; Rivin, O.; You, Z. Evaluation of pKa
Estimation Methods on 211 Druglike Compounds. J. Chem. Inf. Model.
2010, 50, 565−571.
(10) Zachariae, U.; Giordanetto, F.; Leach, A. G. Side Chain
Flexibilities in the Human Ether-a-go-go Related Gene Potassium
Channel (hERG) Together with Matched-Pair Binding Studies
Suggest a New Binding Mode for Channel Blockers. J. Med. Chem.
2009, 52, 4266−4276.
(11) Lavrado, J.; Gani, K.; Nobre, P. A.; Santos, S. A.; Figueiredo, P.;
́
Lopes, D.; Rosario, V.; Gut, J.; Rosenthal, P. J.; Moreira, R.; Paulo, A.
Bis-alkylamine Quindolone Derivatives as New Antimalarial Leads.
ASSOCIATED CONTENT
■
S
* Supporting Information
Structural characterization of starting materials, intermediates,
and final products synthesized and details of the patch-clamp
assay. This material is available free of charge via the Internet at
http://pubs.acs.org.
Bioorg. Med. Chem. Lett. 2010, 20, 5634−5637.
(12) Alptuzun, V.; Parlar, S.; Tasļ ı, H.; Erciyas, E. Synthesis and
̈
̈
Antimicrobial Activity of Some Pyridinium Salts. Molecules 2009, 14,
5203−5215.
AUTHOR INFORMATION
■
(13) Robertson, L.; Hartley, R. C. Synthesis of N-Arylpyridinium
Salts Bearing a Nitrone Spin Trap as Potential Mitochondria-Targeted
Antioxidants. Tetrahedron 2009, 65, 5284−5292.
Corresponding Author
*Tel: +31 (0)71 527 4651. E-mail: ijzerman@lacdr.leidenuniv.
(14) Musilek, K.; Kucera, J.; Jun, D.; Dohnal, V.; Opletalova, V.;
Kuca, K. Monoquaternary Pyridinium Salts with Modified Side Chain-
Synthesis and Evaluation on Model of Tabun- and Paraoxon-Inhibited
Acetylcholinesterase. Bioorg. Med. Chem. 2008, 16, 8218−8223.
(15) Kang, J.; Wang, L.; Chen, X.-L.; Triggle, D. J.; Rampe, D.
Interactions of a Series of Fluoroquinolone Antibacterial Drugs with
the Human Cardiac K+ Channel HERG. Mol. Pharmacol. 2001, 59,
122−126.
nl.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This study was financially supported by the Dutch Top
Institute Pharma, project number D2-201.
(16) Redfern, W. S.; Carlsson, L.; Davis, A. S.; Lynch, W. G.;
MacKenzie, I.; Palethorpe, S.; Siegl, P. K. S.; Strang, I.; Sullivan, A. T.;
Wallis, R.; Camm, A. J.; Hammond, T. G. Relationships between
Preclinical Cardiac Electrophysiology, Clinical QT Interval Prolonga-
tion and Torsade de Pointes for a Broad Range of Drugs: Evidence for
a Provisional Safety Margin in Drug Development. Cardiovasc. Res.
2003, 58, 32−45.
(17) De Bruin, M. L.; Pettersson, M.; Meyboom, R. H. B.; Hoes, A.
W.; Leufkens, H. G. M. Anti-HERG Activity and the Risk of Drug-
Induced Arrhythmias and Sudden Death. Eur. Heart J. 2005, 26, 590−
597.
ABBREVIATIONS USED
■
AI, assay index; approx, approximately; CI, chemical index;
FCC, flash column chromatography; GPCR, G-protein-coupled
receptors; hERG, human ether-a-go-go-related gene; IC50,
concentration of unlabeled ligand which displaces 50% of
[3H]astemizole binding to membranes of HEK293 cells stably
expressing the hERG K+ channel; PCA, patch clamp assay;
RBA, radioligand binding assay; SAR, structure−activity
relationships; TdP, Torsades de pointes; temp, temperature
(18) Webster, R.; Leishman, D.; Walker, D. Towards a Drug
Concentration Effect Relationship for QT Prolongation and Torsades
de Pointes. Curr. Opin. Drug Discovery Dev. 2002, 5, 116−126.
(19) Jamieson, C.; Moir, E. M.; Rankovic, Z.; Wishart, G. Medicinal
Chemistry of hERG Optimizations: Highlights and Hang-Ups. J. Med.
Chem. 2006, 49, 5029−5046.
(20) Cavalli, A.; Poluzzi, E.; De Ponti, F.; Recanatini, M. Toward a
Pharmacophore for Drugs Inducing the Long QT Syndrome: Insights
from a CoMFA Study of HERG K+ Channel Blockers. J. Med. Chem.
2002, 45, 3844−3853.
(21) Aronov, A. M.; Goldman, B. B. A Model for Identifying HERG
K+ Channel Blockers. Bioorg. Med. Chem. 2004, 12, 2307−2315.
(22) Du, L.-P.; Tsai, K.-C.; Li, M.-Y.; You, Q.-D.; Xia, L. The
Pharmacophore Hypotheses of IKr Potassium Channel Blockers:
Novel Class III Antiarrhythmic Agents. Bioorg. Med. Chem. Lett. 2004,
14, 4771−4777.
(23) Zacharias, N.; Dougherty, D. A. Cation-π Interactions in Ligand
Recognition and Catalysis. Trends Pharmacol. Sci. 2002, 23, 281−287.
(24) Fernandez, D.; Ghanta, A.; Kauffman, G. W.; Sanguinetti, M. C.
Physicochemical Features of the hERG Channel Drug Binding Site. J.
Biol. Chem. 2004, 279, 10120−10127.
REFERENCES
■
(1) Ruta, V.; Chen, J.; MacKinnon, R. Calibrated Measurement of
Gating-Charge Arginine Displacement in the KvAP Voltage-Depend-
ent K+ Channel. Cell 2005, 123, 463−475.
(2) Sanguinetti, M. C.; Tristani-Firouzi, M. hERG Potassium
Channels and Cardiac Arrhythmia. Nature 2006, 440, 463−469.
(3) Valentin, J.-P.; Hammond, T. Safety and Secondary Pharmacol-
ogy: Successes, Threats, Challenges and Opportunities. J. Pharmacol.
Toxicol. Methods 2008, 58, 77−87.
(4) Shagufta; Guo, D.; Klaasse, E.; de Vries, H.; Brussee, J.; Nalos, L.;
Rook, M. B.; Vos, M. A.; van der Heyden, M. A. G.; IJzerman, A. P.
Exploring Chemical Substructures Essential for hERG K+ Channel
Blockade by Synthesis and Biological Evaluation of Dofetilide
Analogues. ChemMedChem 2009, 4, 1722−1732.
(5) Vilums, M.; Overman, J.; Klaasse, E.; Scheel, O.; Brussee, J.;
IJzerman, A. P. Understanding of Molecular Substructures that
Contribute to hERG K+ Channel Blockade: Synthesis and Biological
Evaluation of E-4031 Analogues. ChemMedChem 2012, 7, 107−113.
(6) Ponti, F. D.; Poluzzi, E.; Cavalli, A.; Recanatini, M.; Montanaro,
N. Safety of Non-Antiarrhythmic Drugs that Prolong the QT Interval
2839
dx.doi.org/10.1021/jm301564f | J. Med. Chem. 2013, 56, 2828−2840

Journal of Medicinal Chemistry
Article
(25) Howe, J. R.; Ritchie, J. M. On the Active Form of 4-
Aminopyridine: Block of K+ Currents in Rabbit Schwann Cells. J.
Physiol. 1991, 433, 183−205.
(26) Choquet, D.; Korn, H. Mechanism of 4-Aminopyridine Action
on Voltage-Gated Potassium Channels in Lymphocytes. J. Gen. Physiol.
1992, 99, 217−240.
(27) Kameda, M.; Kobayashi, K.; Ito, H.; Miyazoe, H.; Tsujino, T.;
Nakama, C.; Kawamoto, H.; Ando, M.; Ito, S.; Suzuki, T.; Kanno, T.;
Tanaka, T.; Tahara, Y.; Tani, T.; Tanaka, S.; Tokita, S.; Sato, N.
Optimization of a Series of 2,4-Diaminopyridines as Neuropeptide Y
Y1 Receptor Antagonists with Reduced hERG Activity. Bioorg. Med.
Chem. Lett. 2009, 19, 4325−4329.
(28) Owen, D. R.; Rodriguez-Lens, M.; Corless, M. D.; Gaulier, S.
M.; Horne, V. A.; Kinloch, R. A.; Maw, G. N.; Pearce, D. W.; Rees, H.;
Ringer, T. J.; Ryckmans, T.; Stammen, B. L. C. 2,4-Diaminopyridine δ-
Opioid Receptor Agonists and Their Associated hERG Pharmacology.
Bioorg. Med. Chem. Lett. 2009, 19, 1702−1706.
(29) 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.
(30) Scheel, O.; Himmel, H.; Rascher-Eggstein, G.; Knott, T.
Introduction of a Modular Automated Voltage-Clamp Platform and Its
̀
Correlation with Manual Human Ether-a-go-go Related Gene Voltage-
Clamp Data. Assay Drug Dev. Technol. 2011, 9, 600−607.
́
(31) Lukevits, E. Pyridine Derivatives in the Drug Arsenal (150 years
of pyridine chemistry). Chem. Heterocycl. Compd. 1995, 31, 639−650.
(32) Chaubey, A.; Pandeya, S. N. Pyridine: a Versatile Nucleuse in
Pharmaceutical Field. Asian J. Pharm. Clin. Res. 2011, 4, 5−8.
(33) Kumar, S.; Sharma, P. K.; Dudhe, R.; Kumar, N. Pyridine:
Potential for Biological Activities. J. Chronother. Drug Delivery 2011, 2,
71−78.
(34) Cavalli, A.; Buonfiglio, R.; Ianni, C.; Masetti, M.; Ceccarini, L.;
Caves, R.; Chang, M. W. Y.; Mitcheson, J. S.; Roberti, M.; Recanatini,
M. Computational Design and Discovery of “Minimally Structured”
hERG Blockers. J. Med. Chem. 2012, 55, 4010−4014.
(35) Cho, Y. S.; Whitehead, L.; Li, J.; Chen, C. H. T.; Jiang, L.;
Vogtle, M.; Francotte, E.; Richert, P.; Wagner, T.; Traebert, M.; Lu,
̈
Q.; Cao, X.; Dumotier, B.; Fejzo, J.; Rajan, S.; Wang, P.; Yan-Neale, Y.;
Shao, W.; Atadja, P.; Shultz, M. Conformational Refinement of
Hydroxamate-Based Histone Deacetylase Inhibitors and Exploration
of 3-Piperidin-3-ylindole Analogues of Dacinostat (LAQ824). J. Med.
Chem. 2010, 53, 2952−2963.
(36) Patrick, G. L. An Introduction to Medicinal Chemistry; Oxford
University Press: New York, 2001.
(37) Lamsa, M.; Kiviniemi, S.; Kettukangas, E.-R.; Nissinen, M.;
̈
̈
Pursiainen, J.; Rissanen, K. Synthesis and Characterization of New
Aromatic Tweezers and Complex Formation with Tropylium Ion in
1,2-Dichloroethane. J. Phys. Org. Chem. 2001, 14, 551−558.
(38) Yamaguchi, Y.; Kobayashi, S.; Wakamiya, T.; Matsubara, Y.;
Yoshida, Z.-i. Banana-Shaped Oligo(aryleneethynylene)s: Synthesis
and Light-Emitting Characteristics. Angew. Chem., Int. Ed. 2005, 44,
7040−7044.
(39) Zheng, G.; Dwoskin, L. P.; Deaciuc, A. G.; Norrholm, S. D.;
Crooks, P. A. Defunctionalized Lobeline Analogues: Structure−
Activity of Novel Ligands for the Vesicular Monoamine Transporter.
J. Med. Chem. 2005, 48, 5551−5560.
(40) Sisido, K.; Yosikawa, Y. The Organic Reactions with Aluminum
Chloride. XX. The Action of Aluminum Chloride upon Ethylene
Chloride. J. Am. Chem. Soc. 1948, 70, 1608−1609.
2840
dx.doi.org/10.1021/jm301564f | J. Med. Chem. 2013, 56, 2828−2840