Synthesis and quantitative structure-activity relationship of a novel series of small conductance Ca2+-activated K+ channel blockers related to dequalinium

Article abstract of DOI:10.1021/jm950520i

The synthesis, pharmacological testing, and quantitative structure- activity relationship studies of a novel series of bisquinolinium small conductance Ca²⁺-activated K⁺ channel blockers (23) related to dequalinium are described. In this series, two quinolinium rings are linked via the 4- position to an α,ω-diamino alkylene chain and the ring N atom is quaternized with a methyl or benzyl group. The exocyclic N atom can be replaced by 0, S, or CH₂ but with some loss of potency. The quinoline groups do not have to be quaternized for blocking activity, as long as they are basic enough to be protonated at the site of action. For the quaternary compounds, there is considerable steric tolerance for the group R attached to the ring N atom of the quinoline; a benzyl group gave the optimum potency in this series. Moreover, and in contrast to previously reported results for dequalinium analogues, there is no correlation of activity with N¹ charge or E(HOMO). On the other hand, a good correlation was obtained between the blocking potency of the compounds and E(LUMO) [pEMR = 1.16(±0.26)E(LUMO) + 5.33(±1.29) (n = 11, r = 0.83, s = 0.243)]. It has been possible to combine this equation with the previously reported E(LUMO) correlation for a series of dequalinium analogues to include all the compounds of both series [pEMR = 1.17(±0.15)E(LUMO) + 5.33(±0.76) (n = 24, r = 0.85, s = 0.249)]. A possible physical meaning for the E(LUMO) correlation based upon the principle of maximum hardness is discussed.

Full text of DOI:10.1021/jm950520i

J . Med. Chem. 1996, 39, 359-370
359
Articles
Syn th esis a n d Qu a n tita tive Str u ctu r e-Activity Rela tion sh ip of a Novel Ser ies
of Sm a ll Con d u cta n ce Ca 2+-Activa ted K+ Ch a n n el Block er s Rela ted to
Dequ a lin iu m
Dimitrios Galanakis,† Carole A. Davis,† C. Robin Ganellin,*,† and Philip M. Dunn‡
Departments of Chemistry and Pharmacology, University College London, Gower Street, London WC1E 6BT, U.K.
Received J uly 19, 1995X
The synthesis, pharmacological testing, and quantitative structure-activity relationship studies
of a novel series of bisquinolinium small conductance Ca2+-activated K+ channel blockers (23)
related to dequalinium are described. In this series, two quinolinium rings are linked via the
4-position to an R,ω-diamino alkylene chain and the ring N atom is quaternized with a methyl
or benzyl group. The exocyclic N atom can be replaced by O, S, or CH2 but with some loss of
potency. The quinoline groups do not have to be quaternized for blocking activity, as long as
they are basic enough to be protonated at the site of action. For the quaternary compounds,
there is considerable steric tolerance for the group R attached to the ring N atom of the
quinoline; a benzyl group gave the optimum potency in this series. Moreover, and in contrast
to previously reported results for dequalinium analogues, there is no correlation of activity
with N1 charge or EHOMO. On the other hand, a good correlation was obtained between the
blocking potency of the compounds and ELUMO [pEMR ) 1.16((0.26)ELUMO + 5.33((1.29) (n )
11, r ) 0.83, s ) 0.243)]. It has been possible to combine this equation with the previously
reported ELUMO correlation for a series of dequalinium analogues to include all the compounds
of both series [pEMR ) 1.17((0.15)ELUMO + 5.33((0.76) (n ) 24, r ) 0.85, s ) 0.249)]. A
possible physical meaning for the ELUMO correlation based upon the principle of maximum
hardness is discussed.
In tr od u ction
13 and 14, the charged guanidinium groups of which
are important for SKCa channel blockade, although,
alone, they cannot account for the potency of apamin.22
The bis-charged pharmacophore of this peptide has
prompted testing of other compounds bearing two
positively charged N atoms. Thus, the neuromuscular
blockers9,23,24 atracurium, tubocurarine, and pancuro-
nium (Chart 1) were found to be effective SKCa channel
blockers, having potencies in the low micromolar range.
The significantly lower potencies of these bis-charged
compounds compared with apamin add to the suggestion
that the interaction of apamin with the SKCa channel
involves not only Arg13 and Arg14 but also other amino
acids.
Furthermore, dequalinium (22, R2 ) CH3, R3 ) H,
R4 ) NH2, R7 ) H; Chart 1) has been shown to be a
potent and selective blocker of the SKCa channel25,26
(IC50 ) 1 µM). Studies have been undertaken to identify
the pharmacophore of dequalinium for SKCa channel
blockade.27 Replacements for the quinolinium groups
of this molecule have been investigated, and the pos-
sibilities that the molecules of the general structure 22
bind to an anionic or aromatic site on the channel have
been discussed.28 Moreover, it has been suggested that
the role of the NH2 group of dequalinium is electronic,
probably via delocalization of the positive charge.29
Quantitative structure-activity relationship (QSAR)
analysis on analogues of the general structure 22 has
revealed that the blocking potency correlates well with
the partial charge on the ring N atom (obtained from
AM1 semiempirical molecular orbital calculations) and
Small conductance Ca2+-activated K+ (SKCa) channels
comprise an important but relatively little studied
subtype.1,2 These channels are present in a variety of
cell types, and their physiological role is known in some
cases including intestinal smooth muscle,3-5 hepato-
cytes,6,7 and brown fat cells.8 In sympathetic neu-
rons,9,10 opening of SKCa channels mediates a long
hyperpolarization following the action potential (AHP)
which is important for spike frequency adaptation.11,12
Three natural peptidic toxins, apamin,13-15 leiuro-
toxin I16 (scyllatoxin), and PO5,17 are known to block
SKCa channels with potencies in the low nanomolar
range. In particular, apamin has been an invaluable
tool for the study of SKCa channels; using this as a probe,
a role of the SKCa channel in the genesis of myotonia
has been suggested, since the binding site for apamin
is expressed in muscles of myotonic muscular dystrophy
patients, while it is completely absent in normal human
muscle.18,19 In addition, injection of apamin into muscle
of myotonic patients significantly suppressed myotonic
discharges.20 There is also some evidence for the
involvement of SKCa channels in EtOH intoxication.21
Thus, there may be therapeutic possibilities for selective
SKCa channel blockers.
Apamin contains two arginine residues at positions
* Address for correspondence: University College London, Depart-
ment of Chemistry, Christopher Ingold Laboratories, 20 Gordon Str.,
London WC1H 0AJ , U.K.
† Department of Chemistry.
‡ Department of Pharmacology.
X Abstract published in Advance ACS Abstracts, December 1, 1995.
0022-2623/96/1839-0359$12.00/0 © 1996 American Chemical Society

360 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 2
Galanakis et al.
Ch a r t 1
Sch em e 1a
a
Methods: (i) 1, m ) 10, R2 ) CH3, n-pentanol, reflux, 30 h;
(ii) 2, m ) 12, R2 ) CH3, as for 1; (iii) 3, m ) 10, R2 ) H, n-butanol,
reflux, 96 h; (iv) 7a , m ) 8, R2 ) H, n-butanol, reflux, 78 h; (v) 7,
m ) 8, R2 ) H, R ) CH3, X ) I, MEK, reflux, 2 h; (vi) 8, m ) 10,
R2 ) H, R ) CH3, X ) I, as for 7; (vii) 13, m ) 10, R2 ) H, R ) Bn,
X ) Br, MEK, reflux, 24 h; (viii) 14, m ) 10, R2 ) CH3, R ) Bn,
X ) Br, PhNO2, 100-120 °C, 96 h f 120-140 °C, 12 h; (ix) 15, m
) 12, R2 ) CH3, R ) Bn, X ) Br, (iBu)2CO, reflux, 24 h; (x) 18, m
) 10, R2 ) CH3, R ) 3,5-diMeOC6H3CH2, X ) I, PhNO2, Ar, reflux,
100-120 °C, 144 h; (xi) 20, m ) 10, R2 ) CH3, R ) Ph2CHCH2CH2,
X ) I, 4-methylpentan-2-ol, reflux, Ar, 120 h.
Sch em e 2
with the energy of the lowest unoccupied molecular
orbital (ELUMO).30
Since the contribution of substituent R4 to activity
was found to be electronic in nature, it seemed possible
that analogues in which the quinolinium groups in
dequalinium are inverted might also be active, i.e.,
linking the rings via the exocyclic N atoms. This has
afforded another opportunity to investigate electronic
effects, and a second series of analogues having the
structure 23 has been synthesized in which the roles of
the heteroatom X and the group R1 attached to the ring
N atom have been investigated. Correlations have also
been sought between the blocking potency of the ana-
logues and electronic indices, and the results have been
compared with those from structures 22. Furthermore,
in this series it has also been possible to address the
question of the need for the quinoline groups to possess
a quaternized ring N atom. Such an examination was
not possible in the structures 22, which are necessarily
quaternary derivatives.
atom of these compounds is substantially hindered by
the Me group as well as the peri-H atom (at position 8).
Thus, 3 could be benzylated in MEK at reflux (80 °C)
in 24 h to give 13, while for the benzylation of 1, PhNO2
was used as the solvent and the reaction proceeded at
100-140 °C for 108 h. In all cases, alkylation of the
ring N atom of 1 and 2 yielded a mixture of mono- and
bisquaternary products which could be separated only
by reverse phase preparative HPLC. Despite consider-
able efforts to find reaction conditions to obtain exclu-
sively the bisquaternary products, this did not prove to
be possible. The monoquaternary compounds 16, 17,
and 19 (Table 2) were obtained through HPLC separa-
tion from their bisquaternary analogues 14, 15, and 18,
respectively. Of various solvents tested, PhNO2 was
found to give the best results. The iodides 18c and 20c,
required for the synthesis of 18 and 20, respectively,
were obtained as detailed in Scheme 2. Compound 21
can also be synthesized through Scheme 1, but we were
unable to obtain an analytically pure sample through
this route. Instead, the procedure shown in Scheme 3
was used to provide pure 21 but in low yield.
Ch em istr y
Scheme 1 shows the synthesis of compounds 1-3, 7a ,
7, 8, 13-15, 18, and 20. Either 4-chloroquinoline or
4-chloroquinaldine was treated with the necessary di-
amine to provide compounds 1-3 and 7a . These
analogues have been previously synthesized using PhOH
as the solvent31,32 which has been suggested to facilitate
the reaction,33 but we found it more convenient to use
an alcoholic solvent. Quaternization of the ring N atom
of 1-3 and 7a with the appropriate halide yielded
compounds 7, 8, 13-15, 18, and 20. Note that for 1
and 2, each bearing a Me group at position 2 of the
quinoline, high temperatures and longer reaction times
were required for the quaternization step. The ring N
The oxo analogues 4 and 9 were prepared via Scheme
4. Alkaline hydrolysis of 1,1′-decane-1,10-diylbis-
[isothiouronium] dihydrochloride45 yielded the dianion
of the corresponding dithiol which was not isolated but
reacted in situ with 4-chloroquinoline to afford 5 (Scheme
5). Methylation of the latter gave 10. Analogues 6 and

Synthesis and QSAR of SKCa Channel Blockers
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 2 361
Sch em e 3
Sch em e 6
Sch em e 4
Ta ble 1. Structures and Biological Results for the
Nonquaternary Compounds
compd
k
X
R2
IC50 ( SD (µM) EMRa ( SD nb
1
2
3
4
5
6
10 NH
12 NH
10 NH
CH3
CH3
H
H
H
2.5 ( 0.5
3.8 ( 0.3
4.4 ( 1.1
18 ( 7.0
3.6 ( 1.2
7.1 ( 2.8
3.5 ( 1.2
16 ( 21
4
6
3
3
4
4
10
10
O
S
>10c
.10d
>10c
.10d
10 CH2
H
a
Equieffective molar ratio: the ratio of the concentrations of
the test compound and that of dequalinium that cause 50%
inhibition of the AHP, as determined in the same experiment. The
IC50 value for dequalinium in these experiments was 0.87 ( 0.07
b
µM (n ) 8). Number of neurons tested. c Insufficient activity up
Sch em e 5
d
to this concentration to determine IC50
concentration.
.
No activity up to this
at least three cells. Between one and three compounds
were examined at a time, and in each such series of
experiments, dequalinium was also included as a refer-
ence compound. The Hill equation was fitted to the data
to obtain estimates of the IC50. However, because there
was some variation in the potency of dequalinium
during the course of the study, equieffective molar ratios
(EMR; relative to dequalinium) were also obtained by
simultaneous nonlinear least-squares fitting of the data
with the Hill equation. These are also listed in Tables
1 and 2, and it is these values which have been used
for the comparison between compounds. It should be
noted that the compounds were applied in a continu-
ously flowing solution to isolated cells, so that differ-
ences in depletion as a consequence of variation in
lipophilicity are unlikely to have been a complicating
factor.
12a were synthesized via alkylation of the anion of
lepidine with the appropriate dihalide (Scheme 6), and
methylation yielded 11 and 12, respectively.
Although relatively simple, this assay relies on Ca2+
influx during the action potential to activate the SKCa
channels, and the potency of any compound interfering
with this influx may be overestimated. Dequalinium
itself is a highly selective blocker of the SKCa channel,
with no detectable effect on Ca2+ current even at the
relatively high concentration of 10 µM.26 As most of the
compounds tested in the present work have a similar
bis-cationic structure to dequalinium, an action on Ca2+
P h a r m a cology
The SKCa-blocking action of the compounds was
assessed from their ability to inhibit the after hyper-
polarization (AHP) in cultured rat sympathetic neu-
rons26 (see the Experimental Section). Briefly, each
compound was tested at two to four concentrations on

362 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 2
Galanakis et al.
Ta ble 2. Structuresa and Biological Results for the Quaternary Compounds
compd
m
X
R
R1
R2
R7
IC50 ( SD (µM)
EMRb ( SD
nc
7
8
9
8
NH
NH
O
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
PhCH2
PhCH2
PhCH2
H
H
H
H
H
H
H
H
CH3
CH3
CH3
CH3
CH3
CH3
CH3
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
Cl
2.7 ( 0.25
2.1 ( 0.2
6.8 ( 1.7
3.7 ( 0.6
3.9 ( 0.5
40 ( 39
0.7 ( 0.1
0.4 ( 0.05
2.4 ( 0.08
1.3 ( 1.2
4.4 ( 0.6
1.7 ( 0.3
0.75 ( 0.05
4.3 ( 1.3
1.3 ( 0.18
3.1 ( 1.4
1.9 ( 1.0
7.2 ( 3.4
6.2 ( 1.7
5.9 ( 3.9
26 ( 9.8
1.0 ( 0.4
0.6 ( 0.3
4.0 ( 2.0
2.4 ( 0.5
8.0 ( 5.9
2.1 ( 0.9
1.2 ( 0.8
6.2 ( 3.2
1.5 ( 0.6
3
4
7
3
3
3
7
7
3
3
4
4
5
6
4
10
10
10
10
8
10
10
12
10
12
10
10
10
10
10
11
12
13
14
15
16
17
18
19
20
21
S
CH2
CH2
NH
NH
NH
NH
NH
NH
NH
NH
NH
PhCH2
PhCH2
PhCH2
PhCH2
PhCH2
3,5-diMeOC6H3CH3
3,5-diMeOC6H3CH3
Ph2CHCH2CH2
CH3
H
3,5-diMeOC6H3CH3
H
Ph2CHCH2CH2
CH3
a
Compounds 7-12 and 21 were analyzed and tested as the diiodides, 13 as the dibromide, and 14-20 as the bistrifluoroacetates.
For the definition of EMR, see footnote a of Table 1. c Number of neurons tested.
b
channels seems unlikely. Nevertheless, because of the
indirectness of the assay, test concentrations of more
than 10-30 µM were generally avoided. None of the
compounds tested produced obvious broadening of the
action potential, suggesting that, like dequalinium, they
block neither voltage-gated nor the large conductance
calcium-activated (BKCa) K+ channels at the concentra-
tions tested.
Resu lts a n d Discu ssion
From Tables 1 and 2 it can be seen that extension of
the aliphatic chain from 10 to 12 methylene groups
results in some loss of potency (cf. compounds 1 and 2,
14 and 15, 16 and 17), and the same is true for reduction
of the number of methylene groups from 10 to 8 (cf.
compounds 7 and 8, 11 and 12). Nevertheless, the
dependence of activity on the length of the chain is not
striking. A methyl group at position 2 of the quinoline
hardly makes any contribution [cf. R2 ) H or CH3 in
compounds 1-3 (Table 1) and 13 and 14 (Table 2)], and
this is consistent with results previously reported on
dequalinium analogues.29
It is also evident from Table 2 that NH, O, S, and
CH2 groups can be tolerated as heteroatoms X (com-
pounds 8-11, respectively). The most potent is the
nitrogen analogue 8, with the other three compounds
(9-11) having similar potencies to each other and being
3 times less potent than 8.
A charged heterocycle, preferably a quinolinium
group, appears to be an important feature of the
pharmacophore of the SKCa channel blockers for potent
blockade of this channel.28 There is evidence that the
heterocycle does not have to be quaternized, as long as
it is basic enough to be protonated, and hence charged,
at physiological pH.28 The results of this study are
consistent with this model. Thus, of the four nonqua-
ternary compounds (3-6; Table 1) tested, the NH (3)
and O (4) analogues were active (3 being more potent
than 4), and the S compound 5 showed some activity at
10 µM but insufficient to determine an IC50 value, while
the CH2 analogue 6 was inactive. The results for these
F igu r e 1. pKa values35 for 4-substituted quinolines and
deprotonation of 1-alkyl-4-(alkylamino)quinolinium compounds.
compounds could be accounted for by the pKa values of
their quinoline groups. The pKa values for four analo-
gously substituted quinolines are given in Figure 1.35
It can be hypothesized that the quinoline groups of 5
and 6 are not sufficiently basic to be protonated to a
significant extent at the site of action. Although the
pH at the site of action is not known, this is a reasonable
hypothesis to explain the inactivity of 5 and 6, particu-
larly since their quaternary analogues 10 and 11 (Table
2), respectively, are active and equipotent with the
quaternary O compound 9. Proton dissociation in
quaternary aminoquinolinium compounds having the
general formula 24a (Figure 1) can, of course, give the
neutral imino bases 24b. However, it does appear to
be most likely that the active species is the cation since
we have previously shown that the dimethylamino
compound 22, R2 ) R3 ) R7 ) H, R4 ) NMe2 (Chart 1),
which cannot undergo proton dissociation, is equipotent
with the corresponding amino analogue (22, R2 ) R3 )
R7 ) H, R4 ) NH2).29
Quaternization of the ring nitrogen of the quinoline
with a CH3 group increases activity by 2-fold, as can be
seen from compound pairs 8 and 3 and 9 and 4. A

Synthesis and QSAR of SKCa Channel Blockers
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 2 363
Ta ble 3. Partial Charges and Frontier Orbital Energies for Model Compounds, Used for Correlation Eqs 1-6
charge
compda
X
R
R2
R7
C4
N1
EHOMO
ELUMO
pEMRc
b
b
1a
3a
4a
5a
6a
8a
9a
10a
11a
13a
14a
18a
20a
21a
NH
NH
O
H
H
H
H
CH3
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
Cl
0.266
0.271
0.265
-0.116
0.096
0.268
0.257
-0.121
0.087
0.260
0.261
0.260
0.260
0.270
-0.333
-0.333
-0.302
-0.286
-0.271
-0.258
-0.225
-0.208
-0.195
-0.248
-0.249
-0.245
-0.244
-0.260
-12.74
-12.82
-13.31
-12.99
-13.50
-12.64
-13.15
-12.84
-13.36
-12.37
-12.29
-11.43
-11.43
-12.76
-4.840
-4.940
-5.329
-5.551
-5.543
-4.846
-5.207
-5.424
-5.406
-4.716
-4.638
-4.637
-4.600
-4.968
-0.56
-0.54
-1.20
<-1
S
CH2
NH
O
H
,-1
CH3
CH3
CH3
CH3
PhCH2
PhCH2
3,5-diMeOC6H3CH2
Ph2CHCH2CH2
CH3
-0.28
-0.86
-0.79
-0.77
0.00
S
CH2
NH
NH
NH
NH
NH
H
CH3
CH3
CH3
H
0.22
-0.32
-0.79
-0.18
a
b
Model compound numbers correspond to the compound numbers of Tables 1 and 2. In eV. c The -log(EMR) of the corresponding
bisquinolinium compound of Table 1 or 2 is given for simplicity. For the definition of EMR, see footenote a of Table 1.
further 2-fold increase in potency is achieved if the CH3
group on the ring N (8) is replaced by the larger and
more lipophilic benzyl group (13, 14). Thus, there seems
to be considerable steric tolerance in the direction of the
group R (and R1) attached to the ring N atom of the
quinoline. However, the increase in potency is probably
too small to be attributed to the lipophilicity of the
benzyl group. Introduction of two MeO groups in each
of the phenyl rings of 14 reduced activity (18) and so
did replacement of the benzyl group of 14 with the
considerably bulkier diphenylpropyl group (20). Fur-
thermore, nonsymmetrical analogues, having one qua-
ternary and one nonquaternary quinoline, are either
approximately equipotent (19 and 18) or slightly less
potent (16 and 14, 17 and 15) than their symmetrical
counterparts. Finally, introduction of a Cl atom at
position 7 of the quinoline groups of 8 to give compound
21 had no effect on the blocking potency of the analogue.
We reported that the SKCa channel-blocking potency
of dequalinium analogues of the general structure 22
(Chart 1) correlates well with three electronic descrip-
tors obtained from AM1 MO calculations.30 These are
the partial charge on the ring N atom of the quinolinium
group (N1 charge), the EHOMO, and the ELUMO. Respec-
tive correlations were sought for the present series of
blockers. Hence, semiempirical MO calculations were
performed on the compounds of this study. For consis-
tency, the calculations were done on only the compounds
of Table 2 which are symmetrical and have a 10-
methylene chain. Furthermore, as before,30 the calcula-
tions were not performed on the whole molecule but on
model compounds consisting of one substituted quino-
linium group in which the 10-methylene chain was
replaced by a methyl. These simplifications were neces-
sary to increase the computational speed of the calcula-
tions and should not invalidate the comparability of the
results, since any errors introduced are expected to be
similar for all the compounds. The MOPAC MO pack-
age and the AM1 Hamiltonian34 were used for the
calculations.
results of the calculations are given in Table 3. There
is no correlation of pEMR with either the partial charge
on the C atom at position 4 of the quinoline (eq 1) or
the partial charge on the ring N atom of the quinoline
(eq 2). The fact that the pEMR of the analogues of Table
2 do not correlate with the N1 charges is in stark
contrast to the good correlation of pEMR with N1 charge
previously found for the compounds represented by 22,
although we commented at the time that it was difficult
to assign a physical meaning to this correlation.30
Moreover, there appears to be no correlation with the
energy of the highest occupied molecular orbital (EHOMO
,
eq 3), and the correlation with the energy of the lowest
unoccupied molecular orbital (ELUMO) is only a moderate
one (eq 4).
pEMR ) 1.03((1.04)C4 charge - 0.73((0.26) (1)
n ) 12, r ) 0.30, s ) 0.408
pEMR ) 0.19((2.93)N1 charge - 0.46((0.77) (2)
n ) 12, r ) 0.02, s ) 0.428
pEMR ) 0.26((0.19)EHOMO + 2.71((2.35) (3)
n ) 12, r ) 0.40, s ) 0.392
pEMR ) 0.89((0.31)ELUMO + 3.93((1.55) (4)
n ) 12, r ) 0.67, s ) 0.316 (significance at R < 0.02)
With respect to the ELUMO correlation, there appears
to be a major outlier, and this is compound 20a (Figure
2). This compound is much less active than predicted
by the equation. The ring nitrogen of the quinoline has
been alkylated with the bulky diphenylpropyl group;
hence, it is possible that this group interferes sterically
in the interaction of the compound with the channel.
On this basis, 20a can be excluded from the correlation
which then improves substantially (eq 5). Equation 5
is of comparable quality to the reported30 ELUMO cor-
The structures of the model compounds and the

364 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 2
Galanakis et al.
F igu r e 2. Plot of pEMR vs ELUMO for the model compounds
of Table 3, showing the least-squares-fitted regression line.
F igu r e 4. HOMO contour plots for representative model
compounds of Table 3. The HOMO is not localized on the
quinolinium part of the molecule in all compounds as exempli-
fied using 8a , 14a , 18a , and 20a . The molecules are drawn to
match the general structure of Table 3. Contours were drawn
in SYBYL 6.03 at the PSI2 mode, at the 0.0009 e/Å3 level.
F igu r e 3. Plot of pEMR vs EHOMO for the model compounds
of Table 3, showing the least-squares-fitted regression line.
relation for structures 22.
pEMR ) 1.16((0.26)ELUMO + 5.33((1.29) (5)
n ) 11, r ) 0.83, s ) 0.243
(significance at R < 0.005)
F igu r e 5. LUMO contour plots for representative model
compounds 8a , 14a , 18a , and 20a of Table 3. The LUMO is
localized on the quinolinium part of the molecule in all
analogues of Table 3. The molecules are drawn to match the
general structure of Table 3. Contours were drawn in SYBYL
6.03 at the PSI2 mode, at the 0.0009 e/Å3 level.
On the other hand, two compounds, 18a and 20a , are
outliers in the EHOMO correlation (Figure 3). Although
20a can be excluded on the same basis as above, there
appears to be no well-justified reason for excluding 18a .
We have already noted30 that the EHOMO correlation for
the series represented by structure 22 is qualitatively
inconsistent, since the EHOMO does not refer to the same
orbital in all analogues. This is also the case for the
model compounds of Table 3. Whereas in analogues 1a ,
3a -6a , and 8a -11a the HOMO is a π orbital localized
on the quinolinium ring, in analogues 13a , 14a , 18a ,
and 20a there are contributions to the HOMO only from
the benzene ring(s) (Figure 4). Thus, it is also incon-
sistent to use the EHOMO for the present series of
blockers. On the other hand, the LUMO of all the model
compounds of Table 2 is localized on the quinolinium
part of the molecule (Figure 5). In addition to consis-
tency problems, the physical meaning of a correlation
with EHOMO is not clear since such a correlation suggests
that the higher the energy of the HOMO the more
potent the compound is. If the contribution of the
HOMO is assumed to be at the level of the interaction
of the compound with the channel, then a charge
transfer interaction with the latter is implied in which
the compound has the role of the electron donor. This
is in disagreement with fundamental chemical concepts
since it is hard to see how these compounds could act
as electron donors when they are electron deficient.
F igu r e 6. Plot of pEMR vs ELUMO for the model compounds
of Tables 3 and 4, showing the least-squares-fitted regression
line.
It therefore seems that ELUMO is the best electronic
parameter to use both for the series represented by
structure 22 and for the compounds of Tables 1 and 2.
The ELUMO correlations for both these series can be
combined to a single correlation of pEMR with ELUMO
(eq 6, Figure 6) of similar quality to eq 5. This includes
the model compounds of Table 2 (except the inactive
nonquaternary 5a and 6a and the bulky 20a ) and the

Synthesis and QSAR of SKCa Channel Blockers
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 2 365
Ta ble 4. Partial Charges and ELUMO Values for Model Compounds Corresponding to Structure 22, Used for Correlation Eq 6a
R2
R3
R4
R7
ELUMO
pEMRc
b
compd
1b
2b
3b
4b
5b
6b
7b
8b
H
H
H
H
PhCH2NH
PhCH2NH
NH2
H
NH2
H
H
H
H
H
H
H
H
H
H
H
-4.693
-4.341
-4.743
-4.847
-4.347
-4.952
-4.810
-4.652
-4.735
-5.148
-5.058
-5.564
-5.456
0.15
0.05
0.05
0.00
0.00
-0.11
-0.15
-0.26
-0.53
-0.74
-0.81
-1.18
-1.41
-(CH2)4-
CH3
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
NH2
2,4,6-triMeOC6H2CH2NH
NH2
(CH3)2N
Ph(CH3)N
PhNH
CH3CONH
PhO
9b
10b
11b
12b
13b
H
CH3
H
a
b
Data obtained from ref 30. In eV. c The -log(EMR) of the corresponding bisquinolinium compound of ref 30 is given for simplicity.
For the definition of EMR, see footnote a of Table 1.
model compounds of Table 4 (from ref 29).
pEMR ) 1.17((0.15)ELUMO + 5.33((0.76) (6)
n ) 24, r ) 0.85, s ) 0.249
(significance at R < 0.001)
Previously we suggested30 that the ELUMO correlation
is inconsistent with an electron donor-acceptor (EDA)
interaction of the channel with the compounds, since
activity is increased with increasing ELUMO. We pro-
posed that the ELUMO correlation may be related to
processes other than the drug-ion channel itself, such
F igu r e 7. Frontier orbital energy changes on formation of
an ion pair.
in the formation of an ion pair is shown in Figure 7.37
As the cation (compound) approaches the anion (binding
site), the potential of the anion raises the ELUMO of the
compound and the potential of the cation lowers the
EHOMO of the binding site. Since in the quinolinium
compounds the ELUMO is progressively increased, the
ELUMO-EHOMO gap of the ion pair (drug-K+ channel
complex) is also increased, and the latter becomes more
stable. To the best of our knowledge this hypothesis
represents the first attempt to apply the principle of
maximum hardness to structure-activity relationships
in the field of pharmacology.
as the solvation of the compounds; the higher the ELUMO
,
the weaker is the solvation of the blocker and the
stronger becomes the interaction with the channel. An
alternative explanation for the physical meaning of the
ELUMO correlation may be provided by the principle of
maximum hardness initially suggested by Pearson36,37
and described mathematically by Parr using statistical
mechanics and density functional theory.38 Briefly, the
hardness of a chemical system (a chemical system being
an atom, an ion, a radical, a molecule, or several
molecules in a state of interaction37) can be defined as
one-half the energy gap between the LUMO and the
HOMO. The larger the gap, the harder the system and
the more stable it is. There appears to be a general
trend in nature for chemical systems to adjust them-
selves so as to maximize the HOMO-LUMO gap and
thereby minimize their energy and become more
stable.36-38
In order to apply this principle to the interaction of
the quinolinium compounds with the channel, one must
assume that the compounds bind to an anionic site. This
assumption is reasonable since, as has been discussed
above, the positive charge on the quinoline rings is an
essential feature for blocking activity and since all K+
channel subtypes that have been cloned so far have
negatively charged amino acids in the sequence of the
putative pore-forming region of the protein.39-43 These
negatively charged amino acids are believed to facilitate
energetically the conduction of the positively charged
K+ ions.44 The alteration of the frontier orbital energies
Con clu sion
The synthesis and pharmacological testing of a novel
series of SKCa channel blockers have been described. It
has been shown that linking of the two quinolinium
groups from the exocyclic N atom rather than the ring
N atom (as in dequalinium) is tolerated. The exocyclic
N atom can be replaced with O, S. or the CH2 group
but with some loss of potency. The O, S, and CH2
analogues have similar potencies to each other. Fur-
thermore, it appears that the quinoline groups of the
blockers do not have to be quaternized, as long as they
are basic enough to be protonated at the site of action.
For quaternary compounds, there is considerable steric
tolerance for the group R attached to the ring N atom
of the quinoline; a benzyl group gave the optimum
potency in this series.
Moreover, and in contrast to previously reported
results for dequalinium analogues,30 there is no cor-

366 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 2
Galanakis et al.
1
relation of pEMR with N1 charge. In addition, there is
no correlation with EHOMO either, due to the presence
of two major outliers, compounds 18a and 20a . The
EHOMO does not refer to the same orbital in all the
compounds, and this is in line with previous results for
dequalinium analogues.30 On the other hand, the ELUMO
does refer to the same orbital in all the analogues, and
a good correlation has been obtained between the
blocking potency of the compounds and ELUMO, provided
that 20a is excluded. The absence of any correlation of
pEMR with the N1 charge of the analogues of this study
and the fact that the ELUMO correlates with pEMR both
in the present series as well as in a previously reported30
series of dequalinium analogues suggest that ELUMO is
the best electronic parameter to use. Assuming an
interaction of the blockers with an anionic site on the
channel, molecules with a higher ELUMO would form a
harder and, hence, more stable ion pair. This hypoth-
esis, which is based on the principle of maximum
hardness, provides a possible physical meaning for the
ELUMO correlation.
°C; H NMR (400 MHz, CDCl3) δ 1.31, 1.47 (m, 16 H, -CH2-),
1.75, (quint, 4 H, -CH2-), 2.61 (s, 6 H, -CH3), 3.29 (q, 4 H, -CH2-
), 4.90 (sbr, 2 H, N-H), 6.32 (s, 2 H, quinoline-H3), 7.36 (t, 2 H,
quinoline-H6 or -H7), 7.59 (t, 2 H, quinoline-H7 or -H6), 7.67
(d, 2 H, quinoline-H4), 7.91 (d, 2 H, quinoline-H8); MS (FAB,
matrix MNOBA + NaI) [M + H]+ 483, fragments at m/z 325,
311, 297, 283, 269. Anal. (C32H42N4‚1.5MeOH) N; C: calcd,
76.39; found, 76.84. H: calcd, 9.19; found, 9.66.
1,10-Bis(N-qu in olin -4-yla m in o)d eca n e (3). A solution of
4-chloroquinoline (0.5 g, 3.1 mmol) and 1,10-diaminodecane
(0.26 g, 1.5 mmol) in n-butanol (30 mL) was heated under
reflux with stirring for 96 h. The solvent was removed in
vacuo, and the resulting dark solid was treated with NH4OH.
After the latter had been removed in vacuo, the resulting
material was purified by column chromatography using 30%
MeOH in EtOAc. This gave a solid material which was
recrystallized twice from MeOH to give white needles (0.283
g): mp 170-170.5 °C; 1H NMR (400 MHz, CDCl3) δ 1.4 (m, 14
H, -CH2-), 1.75 (m, 4 H, -CH2-), 3.30 (q, 4 H, N-CH2), 5.15 (tbr
,
2 H, NH), 6.42 (d, 2 H, quinoline-H3), 7.40 (t, 2 H, quinoline-
H6 or -H7), 7.62 (t, 2 H, quinoline-H7 or -H6), 7.76 (d, 2 H,
quinoline-H5 or -H8), 8.00 (d, 2 H, quinoline-H8 or -H5), 8.55
(d, 2 H, quinoline-H2); MS (EI) [M]+ 426, fragments at m/z 283,
269, 255, 241, 227, 213, 199, 185, 171, 157, 144; HPLC (column
L, C:D ) 40:60) major peak at 5.26 min representing 99.8% of
the absorption at 220 nm. Anal. (C28H36N4) H, N; C: calcd,
78.83; found, 78.26.
Exp er im en ta l Section
(a ) Ch em istr y. Melting points (mp) were obtained on an
Electrothermal melting point apparatus and are uncorrected.
Infrared (IR) spectra were run on a Perkin-Elmer 983 spec-
trophotometer. Nuclear magnetic resonance (NMR) spectra
were recorded on a Varian XL-200 (200 MHz) or VXR-400 (400
MHz) spectrometer, and chemical shifts (ppm) are reported
relative to the solvent peak (CHCl3 in CDCl3 at 7.24 ppm and
DMSO in DMSO-d6 at 2.49 ppm) or TMS. Signals are
designated as follows: s, singlet; sbr, broad singlet; d, doublet;
dd, doublet of doublets; t, triplet; q, quadruplet; quint, quintet;
m, multiplet. Mass spectra were run on a ZAB SE or VG
7070H spectrometer. Analytical reverse phase high-perfor-
mance liquid chromatography (HPLC) was performed on either
a Gilson or Shimadzu HPLC apparatus with a UV detector at
215 or 254 nm and a Kromasil C18 7 µm (K) or Lichrosorb RP
SELECT B 7 µm (L) column. Isocratic elutions using solvent
mixtures of A ) water + 0.1% TFA and B ) MeOH + 0.1%
TFA or C ) water + 0.5% sodium salt of hexanesulfonic acid
+ 0.5% orthophosphoric acid and D ) MeOH + 0.5% sodium
salt of hexanesulfonic acid + 0.5% orthophosphoric acid were
performed unless otherwise stated. The ratio of A:B or C:D is
indicated for each individual compound. The flow rate was 1
mL/min. For preparative reverse phase HPLC, a Kromasil
C18 7 µm column was used, the solvent mixture being the
same as the one used for analytical HPLC and the flow rate
being 15 mL/min.
4,4′-[Deca n e-1,10-d iylbis(oxy)]bis[qu in olin e] (4). A so-
lution of 4-hydroxyquinoline (1 g, 5 mmol) in DMF (50 mL)
was treated with NaH (0.15 g, 6.25 mmol) with stirring at room
temperature for 15 min. 1,10-Diiododecane (0.98 g, 2.5 mmol)
in DMF was then added and the solution heated to 100 °C for
20 h. After cooling to room temperature, the DMF was
removed in vacuo, and the resulting material was diluted with
water and extracted three times with EtOAc. The combined
extracts were dried, and the solvent was removed in vacuo to
give a colorless oil (1.57 g). This was purified by column
chromatography using 7% MeOH and 1% NH4OH in EtOAc.
This gave 0.302 g of product which was recrystallized from
EtOAc and petroleum ether (small white needles): mp 122-
1
124 °C; H NMR (400 MHz, CDCl3) δ 1.49 (m, 12 H, -CH2-),
1.93 (m, 4 H, -CH2-), 4.17 (t, J ) 6.4 Hz, 4 H, O-CH2), 6.69 (d,
2 H, quinoline-H3), 7.49 (t, 2 H, quinoline-H6 or -H7), 7.67 (t,
2 H, quinoline-H7 or -H6), 8.03 (d, 2 H, quinoline-H5 or -H8),
8.20 (d, 2 H, quinoline-H8 or -H5), 8.72 (d, 2 H, quinoline-H2);
13C NMR (100 MHz, CDCl3) δ 26.1, 28.8, 29.3, 29.4, 68.4, 100.6,
121.5, 121.8, 125.4, 128.9, 129.6, 149.2, 151.4, 161.7; MS (EI)
[M - H]+ 427, fragments at m/z 256, 242, 228, 214, 200, 186,
172, 158, 145, 128; HPLC (column L, H2O + 0.1% TEA:MeOH
+ 0.1% TEA ) 40:60) major peak at 4.42 min representing
95% of the absorption at 240 nm. Anal. (C28H32N2O2) H, N;
C: calcd, 78.40; found, 77.86.
Gen er a l P r oced u r e for th e P r ep a r a tion of 1,10-Bis[N-
(2-m eth ylqu in olin -4-yl)a m in o]d eca n e31,32 (1) a n d 1,12-
Bis[N-(2-m et h ylq u in olin -4-yl)a m in o]d od eca n e31,32 (2).
4-Chloroquinaldine (2 g, 11.26 mmol) and the corresponding
diamine (1,8-diaminooctane, 1,10-diaminodecane, or 1,12-
diaminododecane; 5.68 mmol) were disolved with heating in
n-pentanol (30 mL). The solution was heated under reflux for
30 h. On cooling a creamy precipitate formed, which was
collected by vacuum filtration and dried under vacuum at 50
°C. This was dissolved in methanol, the solution was made
alkaline by addition of NaOH (10% in water), excess of water
was added, and the mixture was kept at 4 °C overnight. The
precipitate formed was collected by vacuum filtration, washed
well with water, and dried under high vacuum over P2O5.
1,10-Bis[(2-m eth ylqu in olin -4-yl)a m in o]d eca n e31,32 (1):
84%; mp 188-190 °C; 1H NMR (400 MHz, CDCl3) δ 1.35, 1.48
(m, 12 H, -CH2-), 1.75 (quint, 4 H, -CH2-), 2.61 (s, 6 H, -CH3),
3.29 (q, 4 H, -CH2-), 4.93 (sbr, 2 H, N-H), 6.32 (s, 2 H, quinoline-
H3), 7.36 (t, 2 H, quinoline-H6 or -H7), 7.59 (t, 2 H, quinoline-
H7 or -H6), 7.67 (d, 2 H, quinoline-H5), 7.90 (d, 2 H, quinoline-
H8); MS (EI) [M]+ 454, fragments at m/z 297, 283, 269, 255,
241. Anal. (C30H38N4‚0.55MeOH) C, H, N.
4,4′-[Deca n e-1,10-d iylbis(th io)]bis[qu in olin e] (5). 1,1′-
Decane-1,10-diylbis[isothiouronium] dihydrochloride45 (0.788
g, 2.17 mmol), 4-chloroquinoline (0.71 g, 4.34 mmol), and KOH
(0.789 g, 14.06 mmol) were dissolved with heating in absolute
EtOH (30 mL), and the solution was heated under reflux for
3 h. The white precipitate formed was removed by vacuum
filtration while the solution was still warm. The filtrate was
concentrated to dryness in vacuo, dispersed in water, and
extracted with CH2Cl2 (5 × 50 mL). The extracts were
combined and dried with K2CO3, and the solvent was removed
in vacuo to yield a yellow oil. This was dissolved in the
minimum amount of EtOAc, excess of petroleum ether was
added, and the white solid formed was collected by filtration
and washed with EtOAc. This was recrystallized from EtOAc
to yield a white hygroscopic powder (0.518 g, 52%): mp 96-
97 °C; 1H NMR (400 MHz, CDCl3, TMS) δ 1.34 (m, 8 H, -CH2-
), 1.52 (quint, 4 H, -CH2-), 1.82 (quint, 4 H, -CH2-), 3.10 (t, 4
H, S-CH2), 7.17 (d, 2 H, quinoline-H3), 7.54 (t, 2 H, quinoline-
H6 or -H7), 7.71 (t, 2 H, quinoline-H7 or -H6), 8.06 (d, 2 H,
quinoline-H5), 8.13 (d, 2 H, quinoline-H8), 8.71 (d, 2 H,
quinoline-H2); MS (FAB, matrix MNOBA) [M + H]+ 461,
fragments at m/z 332, 300, 286, 272, 258, 244, 230, 216, 188,
175, 162, 129; HPLC (column K, A:B ) 30:70) major peak at
1,12-Bis[(2-m eth ylqu in olin -4-yl)am in o]dodecan e31,32 (2):
79%; for experimental details see compound 1; mp 153-155

Synthesis and QSAR of SKCa Channel Blockers
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 2 367
7.06 min representing 98.2% of the absorption at 215 nm. Anal.
(C28H32N2S2‚0.7H2O) C, H, N.
and dried (73 mg, 88%): mp 158-161 °C; 1H NMR (400 MHz,
CD3OD) δ 1.49-1.63 (m, 12 H, -CH2-), 2.06 (quint, J ) 6.5
Hz, 4 H, -CH2-), 4.47 (s, 6 H, CH3), 4.60 (t, J ) 6.5 Hz, 4 H,
O-CH2), 7.50 (d, 2 H, quinoline-H3), 7.90 (t, 2 H, quinoline-H6
or -H7), 8.22 (t, 2 H, quinoline-H7 or -H6), 8.32 (d, 2 H,
quinoline-H5 or -H8), 8.55 (d, 2 H, quinoline-H8 or -H5), 9.12
(d, 2 H, quinoline-H2); MS (FAB, matrix MNOBA) [M + I +
H]+ 586, [M]+ 458, fragments at m/z 444, 427, 298, 284, 270,
256, 242, 229, 214, 200. Anal. (C30H38N2O2I2) C, H, N, I.
4,4′-[Deca n e-1,10-d iylbis(th io)]bis[1-m eth ylqu in olin i-
u m ] Diiod id e Hyd r a te (10). Compound 5 (0.124 g, 0.27
mmol) and MeI (2 mL, 32.13 mmol) were dissolved in MEK
(20 mL), and the solution was heated under reflux for 4 h
under Ar. A yellow precipitate formed which, after cooling to
room temperature, was collected by vacuum filtration, washed
with MEK, and dried in vacuo (0.172 g, 86%): mp 196-198
°C; 1H NMR (400 MHz, DMSO-d6, TMS) δ 1.32 (m, 8 H, -CH2-
), 1.51 (quint, 4 H, -CH2-), 1.79 (quint, 4 H, -CH2-), 3.46 (t, 4
H, S-CH2), 4.48 (s, 6 H, N+-CH3), 7.96-8.03 (m, 4 H, quinoline),
8.25 (t, 2 H, quinoline-H7 or -H6), 8.43 (dd, 4 H, quinoline-H5
+ -H8), 9.14 (d, 2 H, quinoline-H2); MS (FAB, matrix MNOBA)
[M - H]+ 489, fragments at m/z 475, 315, 300, 286, 272, 258,
245, 230, 216, 202, 189, 175, 143; HPLC (column K, A:B )
40:60) major peak at 14.13 min representing 100% of the
absorption at 215 nm. Anal. (C30H38N2S2I2‚H2O) C, H, N.
1,1′-Dim e t h yl-4,4′-d od e ca n e -1,12-d iylb is[q u in olin i-
u m ] Diiod id e (11). Compound 6 (0.2 g, 0.47 mmol) and MeI
(2 mL, 32.13 mmol) were dissolved in MEK (40 mL), and the
solution was heated under reflux for 4 h under argon. The
reaction mixture was cooled to room temperature and the
yellow solid collected by filtration, washed with the solvent,
4,4′-Dod eca n e-1,12-d iylbis[qu in olin e] (6). Na (0.482 g,
20.96 mmol) was dispersed in liquid NH3 containing a catalytic
amount of Fe(NO3)3 under argon. When the dark blue color
of the suspension turned gray, lepidine (3 g, 20.95 mmol) was
added and the reaction mixture was stirred for 1 h. 1,10-
Dibromodecane (2.075 g, 6.91 mmol) was then gradualy added,
and the NH3 was allowed to evaporate overnight. Water was
added to the residue and the aqueous phase extracted with
Et2O. The extracts were combined, dried (Na2SO4), and
concentrated to dryness in vacuo to yield an oil which
contained mainly lepidine and the title compound. The latter
was purified by column chromatography and isolated as a
yellow oil. This was dissolved in the minimum amount of hot
MeOH and placed at -20 °C. White crystals separated which
were collected by filtration, washed with cold MeOH, and dried
under vacuum (1.275 g, 43.4%): mp 50-51 °C; 1H NMR (400
MHz, CDCl3) δ 1.27-1.36 (m, 12 H, -CH2-), 1.43 (quint, 4 H,
-CH2-), 1.76 (quint, 4 H, -CH2-), 3.08 (t, 4 H, Ar-CH2-), 7.27 (d,
2 H, quinoline-H3), 7.58 (t, 2 H, quinoline-H6 or -H7), 7.73 (t,
2 H, quinoline-H7 or -H6), 8.06 (d, 2 H, quinoline-H5), 8.18 (d,
2 H, quinoline-H8), 8.80 (d, 2 H, quinoline-H2); MS (FAB,
matrix MNOBA) [M + H]+ 425; HPLC (column K, A:B ) 35:
65) major peak at 10.16 min representing 97.3% of the
absorption at 215 nm. Anal. (C30H36N2) C, H, N.
1,8-Bis(N-qu in olin -4-yld ia m in o)octa n e (7a ).32 A solu-
tion of 4-chloroquinoline (0.3 g, 2 mmol) in n-butanol (30 mL)
was treated with 1,8-diaminooctane (0.18 g, 1 mmol) under
reflux with stirring for 78 h. After removal of the solvent in
vacuo, the residue was treated with NH4OH and extracted
with CHCl3. Drying (Mg2SO4) of the extracts and removal of
the CHCl3 afforded a white solid which was purified by column
chromatography using 10% MeOH in EtOAc. This gave 0.28
g of product which was recrystallized from a mixture of iPrOH
and MeOH to yield pale yellow platelets: MS (FAB, matrix
MNOBA) [M + H]+ 399, fragments at m/z 270, 255, 241, 213,
199, 185, 171, 157, 144, 129.
1,8-Bis[N-(1-m eth ylqu in olin iu m -4-yl)a m in o]octa n e Di-
iod id e Hyd r a te (7). Compound 7a (89 mg, 0.22 mmol) was
dissolved in MEK (20 mL) and treated with MeI (2 mL, 32.15
mmol) under reflux for 2 h. The creamy white precipitate was
collected and recrystallized from absolute EtOH to give cream
microcrystalline needles (94 mg, 63%): mp 279-281 °C; 1H
NMR (400 MHz, CD3OD) δ 1.51 (m, 8 H, -CH2-), 1.86 (quint,
4 H, -CH2-), 3.64 (t, J ) 7.3 Hz, 4 H, NH-CH2), 4.21 (s, 6 H,
CH3), 6.90 (d, J ) 7.4 Hz, 2 H, quinoline-H3), 7.81 (t, J ) 6.4
Hz, 2 H, quinoline-H6 or -H7), 8.09 (m, 4 H, quinoline), 8.50
(m, 4 H, quinoline); MS (FAB, matrix MNOBA) [M + I - H]+
554, [M - H]+ 427, fragments at m/z 413, 269, 255, 241, 227,
213, 199, 185, 171, 157, 143; HPLC (column L, C:D ) 40:60)
major peak at 5.61 min representing 97% of the absorption at
215 nm. Anal. (C28H36N4I2‚1.5H2O) C, H, N, I.
1,10-Bis[N-(1-m eth ylqu in olin iu m -4-yl)am in o]decan e Di-
iod id e (8). 3 (0.2 g, 0.5 mmol) was dissolved in MEK (30 mL)
and treated with MeI (1 mL, 16.07 mmol) under reflux for 2
h. After cooling, the solvent was removed in vacuo and the
resulting solid was dissolved in hot EtOH, treated with
charcoal, and filtered and the filtrate allowed to cool. This
gave a creamy powdery solid which was collected, washed with
a small amount of iPrOH, and dried (0.162 g, 71%): mp 233-
236 °C; 1H NMR (400 MHz, DMSO-d6) δ 1.30 (m, 12 H, -CH2-
), 1.69 (quint, 4 H, -CH2-), 3.54 (t, J ) 7.2 Hz, 4 H, NH-CH2),
4.14 (s, 6 H, N+-CH3), 6.90 (d, J ) 6.5 Hz, 2 H, quinoline-H3),
7.80 (t, J ) 6.5 Hz, 2 H, quinoline-H6 or -H7), 8.08 (m, 4 H,
quinoline), 8.59 (m, 4 H, quinoline), 9.23 (sbr, 2 H, NH); MS
(FAB, matrix MNOBA) [M + I]+ 583, [M - H]+ 455, fragments
at m/z 441, 313, 297, 283, 269, 255, 242, 227, 213, 199, 185,
171, 143; HPLC (column L, C:D ) 20:80) major peak at 3.95
min representing 99.8% of the absorption at 240 nm. Anal.
(C30H40N4I2‚H2O) C, H, N.
1
and dried under vacuum (0.284 g, 85%): mp 156-158 °C; H
NMR (400 MHz, DMSO-d6) δ 1.25 (m, 8 H, -CH2-), 1.32 (m, 4
H, -CH2-), 1.41 (m, 4 H, -CH2-), 1.73 (m, 4 H, -CH2-), 3.34-
3.38 (m, Ar-CH2-, water), 4.57 (s, 6 H, N+-CH3), 8.03-8.07 (m,
4 H, quinoline-H3, -H6, or -H7), 8.27 (t, 2 H, quinoline-H7 or
-H6), 8.49 (d, 2 H, quinoline-H5), 8.60 (d, 2 H, quinoline-H8),
9.36 (d, 2 H, quinoline-H2); MS (FAB, matrix MNOBA + Li +
Na) [M - H]+ 453, fragments at m/z 438, 412, 170; HPLC
(column K, A:B ) 35:65) major peak at 5.23 min representing
99.2% of the absorption at 215 nm. Anal. (C32H42N2I2) C, H,
N.
4,4′-Decan e-1,10-diylbis[qu in olin e] (12a). Lepidine (2.762
g, 19.28 mmol), Na (0.52 g, 22.6 mmol), and 1,8-diiodooctane
(3.53 g, 9.64 mmol) were reacted in a manner similar to that
described under 6. The product was isolated as a yellow oil
which solidified after drying under vacuum to yield a creamy
solid (1.28 g, 33.5%): mp 61-62 °C; 1H NMR (400 MHz, CDCl3,
TMS) δ 1.31-1.45 (m, 12 H, -CH2-), 1.76 (quint, 4 H, -CH2-),
3.06 (t, 4 H, Ar-CH3), 7.23 (d, 2 H, quinoline-H3), 7.55 (t, 2 H,
quinoline-H6 or -H7), 7.70 (t, 2 H, quinoline-H7 or -H6), 8.04
(d, 2 H, quinoline-H5), 8.12 (d, 2 H, quinoline-H8), 8.80 (d, 2
H, quinoline-H2). Anal. (C28H32N2‚0.3H2O) C, H, N.
1,1′-Dim eth yl-4,4′-d eca n e-1,10-diylbis[qu in olin iu m ] Di-
iod id e (12). Compound 12a was methylated as described in
the preparation of 11 in 88.5% yield: mp 198-200 °C; 1H NMR
(400 MHz, DMSO-d6 + 10 drops of TFA, TMS) δ 1.27-1.44
(m, 12 H, -CH2-), 1.75 (quint, 4 H, -CH2-), 3.38 (t, 4 H, Ar-
CH2-), 4.59 (s, 6 H, N+-CH3), 8.03-8.08 (m, 4 H, quinoline-H3,
-H6, or -H7), 8.27 (t, 2 H, quinoline-H7 or -H6), 8.51 (d, 2 H,
quinoline-H5), 8.60 (d, 2 H, quinoline-H8), 9.38 (d, 2H, quino-
line-H2); MS (FAB, matrix MNOBA + Li + Na) [(M + I) -
H]+ 552, [M]+ 426, fragments at m/z 412, 282, 268, 254; HPLC
(column K, A:B ) 50:50) major peak at 13.30 min representing
100% of the absorption at 215 nm. Anal. (C30H38N2I2‚H2O)
C, H, N.
1,10-Bis[N-(1-ben zylqu in olin iu m -4-yl)am in o]decan e Di-
br om id e (13). Compound 3 (0.5 g, 1.17 mmol) and benzyl
bromide (0.4 g, 2.34 mmol) were dissolved in MEK, and the
solution was heated under reflux for 24 h. The white precipi-
tate formed was collected and washed with MeOH (0.46 g,
64%): mp 270-272 °C; 1H NMR (400 MHz, CD3OD) δ 1.29
(m, 12 H, CH2), 1.73 (quint, 4 H, CH2), 3.53 (t, 4 H, NH-CH2),
5.72 (s, 4 H, N+-CH2), 6.87 (d, 2 H, quinoline-H3), 7.14 (d, 4 H,
Ph), 7.24 (m, 6 H, Ph), 7.60 (ddd, 2 H, quinoline-H6 or -H7),
7.78 (ddd, 2 H, quinoline-H7 or -H6), 7.87 (d, 2 H, quinoline-
4,4′-[Deca n e-1,10-d iylb is(oxy)]b is[1-m et h ylq u in olin i-
u m ] Diiod id e (9). Compound 4 (50 mg, 0.12 mmol) was
dissolved in MEK (10 mL) and treated with MeI (1 mL, 16.07
mmol) under reflux for 3 h. After cooling to room temperature,
the pale yellow precipitate was collected, washed with MEK,

368 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 2
Galanakis et al.
H5 or -H8), 8.38 (dd, 2 H, quinoline-H8 or -H5), 8.57 (d, 2 H,
quinoline-H2); MS (FAB, matrix MNOBA) [M]+ 608, fragment
at m/z 517; HPLC (column K, A:B ) 33:67) major peak at 7.24
min representing 95.6% of the absorption at 215 nm. Anal.
(C42H48N4Br2) C, H, N.
solidified after thorough drying in vacuo over P2O5 (greenish,
hygroscopic crystals): mp 77-79 °C melts, 104-106 °C liqui-
1
fies; H NMR (400 MHz, DMSO-d6) δ 1.27 (m, 16 H, -CH2-),
1.70 (m, 4 H, -CH2-), 2.64 (s, 3 H, -CH3′), 2.74 (s, 3 H, -CH3),
3.48 (q, 2 H, -CH2-), 3.52 (q, 2 H, -CH2-), 5.85 (s, 2 H, Ph-CH2-
N+), 6.78 (s, 1 H, quinoline-H3′), 7.06 (d, 3 H, Ph-Ho, quinoline-
H3), 7.32 (t, 1 H, Ph-Hp), 7.36 (t, 2 H, Ph-Hm), 7,67 (ddd, 2 H,
quinoline-H6 or -H7, quinoline-H6′ or -H7′), 7.82-7.95 (m, 4 H,
quinoline-H5, -H5′, quinoline-H7 or -H6, quinoline-H7′ or -H6′),
8.46 (d, 1 H, quinoline-H8′), 8.56 (d, 1 H, quinoline-H8), 9.03
(tbr, 1 H, N-H′), 9.30 (tbr, 1 H, N-H); MS (FAB, matrix MNOBA
+ NaI) [M + H]+ 574, fragment at m/z 482; HPLC (column K,
A:B ) 35:65) major peak at 13.48 min representing 97.5% of
the absorption at 215 nm. Anal. (C43H50F6N4O4‚0.6 CF3-
COOH) C, H, N.
1,10-Bis[N-(1-ben zyl-2-m eth ylqu in olin iu m -4-yl)am in o]-
d eca n e Bis(tr iflu or oa ceta te) (14). Compound 1 (0.3 g, 0.66
mmol) was dissolved with heating in previously dried (molec-
ular sieves, 4 Å) nitrobenzene (10 mL), and benzyl bromide
(0.288 g, 1.68 mmol) was added. The solution was heated in
an oil bath under argon at 100-120 °C for 96 h and then at
120-140 °C for 12 h. More benzyl bromide was added after
48 h (0.144 g, 0.84 mmol) and 72 h (0.144 g, 0.84 mmol). The
reaction mixture was cooled to room temperature, excess of
ether was added, and a gum came out of solution. The
supernatant solution was decanted and the gum washed
several times with ether. This consisted mainly of the mono-
and bisquaternary products which were separated by reverse
phase preparative HPLC. The product was isolated as a green
oil which solidified after drying under vacuum: mp 58-60 °C;
1H NMR (400 MHz, DMSO-d6) δ 1.31-1.40 (m, 12 H, -CH2-),
1.71 (quint, 4 H, -CH2-), 2.73 (s, 6 H, -CH3), 3.54 (q, 4 H,
N-CH2-), 5.84 (s, 4 H, -CH2-N+), 7.05 (d, 6 H, C3-H, Ph-Ho),
7.29 (t, 2 H, Ph-Hp), 7.35 (t, 4 H, Ph-Hm), 7.69 (t, 4 H, quinoline-
H6 or -H7), 7.88 (t, 2 H, quinoline-H7 or -H6), 7.93 (d, 2 H,
quinoline-H5), 8.56 (d, 2 H, quinoline-H8), 9.31 (t, 2 H, N-H);
MS (FAB, matrix MNOBA) [M - H]+ 635; HPLC (column K,
A:B ) 35:65) major peak at 19.42 min representing 100% of
the absorption at 215 nm. Anal. (C48H52N4F6O8‚1.2CF3CO2H)
C, H, N.
3,5-Dim eth oxyben zyl iod id e (18c). 3,5-Dimethoxybenzyl
chloride (0.5 g, 2.68 mmol) and dry NaI (0.44 g, 2.94 mmol)
were dissolved in dry (K2CO3) acetone (10 mL), and the
solution was heated to 50 °C for 3 h. The solvent was removed
in vacuo and the residue treated with CH2Cl2 and filtered in
order to remove the NaCl formed and the excess of NaI. The
filtrate was concentrated to dryness, and the last traces of
solvents were removed under vacuum to yield a yellowish
1
solid: mp 84-85 °C; H NMR (200 MHz, CDCl3) δ 3.77 (s, 6
H, -CH3), 4.37 (s, 2 H, -CH2-), 6.31 (m, 1 H, Ph-H4), 6.51 (d, J
) 2.25 Hz, 2 H, Ph-H2, Ph-H6).
1,10-Bis[N-[1-(3,5-d im et h oxyb en zyl)-2-m et h ylq u in o-
lin iu m -4-yl]am in o]decan e Bis(tr iflu or oacetate) (18). Com-
pounds 1 (0.3 g, 0.66 mmol) and 18c (0.367 g, 1.32 mmol) were
dissolved with heating in previously dried (molecular sieves,
4 Å) nitrobenzene (10 mL), and the solution was heated under
argon in an oil bath at 100-120 °C for 144 h. The reaction
mixture was cooled to room temperature, and dry Et2O was
added until the solution became cloudy. The creamy precipi-
tate formed was collected by filtration, washed three times
with EtOH, and dried under vacuum at 60 °C over P2O5. The
later was a mixture of the mono- and bisquaternary com-
pounds, which were separated by reverse phase preparative
HPLC. The product was isolated as a creamy oil which
solidified after drying in vacuo: mp 96-98 °C melts, 115-
119 °C solidifies, 213-215 °C melts; 1H NMR (400 MHz,
DMSO-d6) δ 1.30-1.45 (m, 12 H, -CH2-), 1.71 (quint, 4 H, -CH2-
), 2.72 (s, 6 H, -CH3), 3.54 (q, overlaps with water signal,
N-CH2-), 3.67 (s, 12 H, O-CH3), 5.74 (s, 4 H, -CH2-N+), 6.14 (d,
4 H, Ph-Ho), 6.44 (t, 2 H, Ph-Hp), 7.04 (s, 2 H, quinoline-H3),
7.69 (m, 2 H, quinoline-H6 or -H7), 7.90 (d, 4 H, quinoline-H5,
-H7 or quinoline-H5, -H6), 8.58 (d, 2 H, quinoline-H8), 9.30 (t,
2 H, N-H); MS (FAB, matrix glycerol + thioglycerol + TFA)
[M]+ 756, fragments at m/z 605, 335, 185, 151; HPLC (column
K, A:B ) 35:65) major peak at 34.83 min representing 99.6%
of the absorption at 215 nm. Anal. (C52H60N4F6O8‚0.5CF3-
CO2H) C, H, N.
1-[N-[1-(3,5-Dim et h oxyb en zyl)-2-m et h ylqu in olin iu m -
4-yl]a m in o]-10-[N′-(2-m eth ylqu in olin iu m -4-yl)a m in o]d e-
ca n e Bis(tr iflu or oa ceta te) (19). The preparation is de-
scribed under 18. The product was isolated as a creamy oil
which solidified after drying at high vacuum: mp 77-78 °C;
1H NMR (400 MHz, DMSO-d6) δ 1.34-1.38 (m, 12 H, -CH2-),
1.72 (m, 4 H, -CH2-), 2.62 (s, 3 H, -CH3′), 2.72 (s, 3 H, -CH3′),
3.53 (m, 4 H, N-CH2-), 3.67 (s, 6 H, O-CH3), 5.74 (s, 2 H, -CH2-
N+), 6.15 (s, 2 H, Ph-Ho), 6.44 (s, 1 H, Ph-Hp), 6.78 (s, 1 H,
quinoline-H3′), 7.04 (s, 1 H, quinoline-H3), 7.64-7.70 (m, 2 H,
quinoline-H6, -H6′ or quinoline-H7, -H7′), 7.81 (d, 1 H, quinoline-
H5′), 7.90 (m, 3 H, quinoline-H5, -H7, -H7′ or quinoline-H5, -H6,
-H6′), 8.45 (d, 1 H, quinoline-H8′), 8.55 (d, 1 H, quinoline-H8),
9.04 (t, 1 H, N-H′), 9.30 (t, 1 H, N-H); MS (FAB, matrix glycerol
+ thioglycerol + TFA) [M]+ 605, fragments at m/z 455, 185,
171, 151; HPLC (column K, A:B ) 35:65) major peak at 14.71
min representing 99.5% of the absorption at 215 nm. Anal.
(C41H50N4F6O8‚0.5CF3CO2H) C, H, N.
1,12-Bis[N-(1-ben zyl-2-m eth ylqu in olin iu m -4-yl)am in o]-
d od eca n e Bis(tr iflu or oa ceta te) (15). Compound 2 (0.3g,
0.621 mmol) and benzyl bromide (0.319 g, 1.865 mmol) were
dissolved with heating in distilled diisobutyl ketone (10 mL),
and the solution was heated under reflux for 24 h. On cooling
a
dark gum came out of solution which crystallized on
standing. This was collected by filtration and dried overnight
in vacuo over P2O5. It was shown (MS, HPLC) that this was
a mixture of three compounds, namely, 36 and the monoben-
zylated (37) and dibenzylated (38) products. These were
separated by reverse phase preparative HPLC. The compound
was isolated as a greenish oil which solidified after thorough
drying in vacuo over P2O5 (greenish, hygroscopic crystals): mp
1
133-134 °C; H NMR (400 MHz, DMSO-d6) δ 1.26-1.41 (m,
16 H, -CH2-), 1.72 (quint, 4 H, -CH2-), 2.74 (s, 6 H, -CH3), 3.55
(q, 4 H, -CH2-), 5.85 (s, 4 H, Ph-CH2-N+), 7.06 (d, s, 6 H, Ph-
Ho, quinoline-H3), 7.30 (t, 2 H, Ph-Hp), 7.36 (t, 4 H, Ph-Hm),
7.69 (t, 2 H, quinoline-H6 or -H7), 7.89 (t, 2 H, quinoline-H7 or
-H6), 7.94 (d, 2 H, quinoline-H5), 8.57 (d, 2 H, quinoline-H8),
9.30 (tbr, 2 H, N-H); MS (FAB, matrix MNOBA + NaI) [M]+
664, fragments at m/z 547, 482, 431, 415, 401, 387; HPLC
(column K, A:B ) 35:65) major peak at 27.28 min representing
98.1% of the absorption at 215 nm. Anal. (C50H56F6N4O4‚2CF3-
COOH‚4H2O) C, H, N [agreed after allowing for the presence
of 10.14% of inorganic matter (determined by an ash test)].
1-[N-(1-Ben zyl-2-m et h ylq u in olin iu m -4-yl)a m in o]-10-
[N′-(2-m eth ylqu in olin iu m -4-yl)a m in o]d eca n e Bis(tr iflu o-
r oa ceta te) (16). The preparation is described under 14. The
compound was isolated as a greenish oil which solidified after
drying under vacuum: mp 72-74 °C; 1H NMR (400 MHz,
DMSO-d6, TMS) δ 1.30 (m, 12 H, -CH2-), 1.70 (quint, 4 H,
-CH2-), 2.64 (s, 3 H, -CH3′), 2.74 (s, 3 H, -CH3), 2.48 (q, 2 H,
N-CH2-′), 2.55 (q, 2 H, N-CH2-), 5.85 (s, 2 H, -CH2-N+), 6.79
(s, 1 H, quinoline-H3′), 7.06 (d, 3 H, quinoline-H3, Ph-Ho), 7.30
(t, 1 H, Ph-Hp), 7.36 (t, 2 H, Ph-Hm), 7.67 (m, 2 H, quinoline-
H6, -H6′ or quinoline-H7, -H7′), 7.83 (d, 1 H, quinoline-H5′), 7.89
(t, 2 H, quinoline-H7, -H7′ or quinoline-H6, -H6′), 7.94 (d, 1 H,
quinoline-H5), 8.46 (d, 1 H, quinoline-H8′), 8.56 (d, 1 H,
quinoline-H8), 9.06 (t, 1 H, N-H′), 9.32 (t, 1 H, N-H); MS (FAB,
matrix MNOBA) [M - H]+ 544; HPLC (column K, A:B ) 35:
65) major peak at 9.52 min representing 96.8% of the absorp-
tion at 215 nm. Anal. (C41H46N4F6O8‚0.5CF3CO2H) C, H, N.
1-[N-(1-Ben zyl-2-m et h ylq u in olin iu m -4-yl)a m in o]-12-
(N′-(2-m eth ylqu in olin iu m -4-yl)a m in o]d od eca n e Bis(tr i-
flu or oa ceta te) (17). The experimental procedure is described
under 15. The compound was isolated as a greenish oil which
1,1′-(3-Iod op r op ylid en e)bis[ben zen e] (20c). To a stirred
and cooled to 0 °C solution of 3,3-diphenyl-1-propanol (2.5 g,
11.78 mmol) in dry CH2Cl2 were successively added Ph3P
(3.243 g, 12.36 mmol) and iodine (3.048 g, 12.01 mmol). The
work was carried out under an atmosphere of argon. The
orange solution was stirred for 15 min, and an orange

Synthesis and QSAR of SKCa Channel Blockers
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 2 369
precipitate formed. Imidazole (1.042 g, 15.31 mmol) was
added, the color disappeared, and a white precipitate formed.
The reaction mixture was stirred at 0 °C for 30 min and then
heated under reflux for 2 h. After cooling to room temperature,
the reaction mixture was washed with water (20 mL) and the
water layer extracted with CH2Cl2 (20 mL). The CH2Cl2
extracts were combined, dried with Na2SO4, and concentrated
to dryness in vacuo to yield a white solid. The iodide was
purified by column chromatography (silica gel, petroleum ether
30-40 °C:CH2Cl2 ) 5:1) and isolated as a white solid (2.910
the solution contained 5 mM HCl. Superior cervical ganglia
from 17 day old rats were treated with collagenase and trypsin
and then dissociated using a fire-polished pipette. The result-
ant cell suspension was plated onto laminin-coated plastic
dishes and maintained in tissue culture for 3-10 days.
Culture dishes were placed on the stage of an inverted
microscope and perfused with physiological salt solution of the
following composition (in mM): NaCl, 118; KCl, 4.8; CaCl2,
2.5; MgSO4, 1.19; NaHCO3, 25; KH2PO4, 1.18; and glucose, 11;
warmed to 30 °C and equilibrated with 95% O2:5% CO2.
Intracellular recordings were made with convetional micro-
electrodes filled with 1 M KCl (resistance 80-120 MΩ),
connected to the headstage of a Neurolog NL 102 amplifier.
Action potentials were evoked by injection of 30 ms depolar-
izing current pulses every 5 s. Data aquisition and analysis
were performed on a microcomputer using the pClamp suite
of programs (Axon Instruments). Inhibition of the AHP was
measured by digitally subtracting records obtained in the
presence of blocker from controls and expressing the peak
difference as a percentage of the control AHP at the same time
point.
1
g, 76.7%): mp 54-55 °C; H NMR (200 MHz, CDCl3) δ 2.49
(q, 2 H, -CH2-), 3.02 (t, 2 H, -CH2-I), 4.04 (t, 1 H, Ph2-CH),
7.09-7.28 (m, 10 H, Ph); MS (EI) [M]+ 322, fragments at m/z
245, 168, 167, 155, 141, 127.
1,10-Bis[N-[1-(3,3-d ip h en ylp r op -1-yl)-2-m et h ylq u in o-
lin iu m -4-yl]am in o]decan e Bis(tr iflu or oacetate) (20). Com-
pounds 1 (0.3 g, 0.66 mmol) and 20c (0.5 g, 1.552 mmol) were
dissolved with heating in 4-methylpentan-2-ol (10 mL), and
the solution was heated under reflux for 120 h under argon,
more iodide being added after 36 h (0.2 g, 0.621 mmol) and 60
h (0.2 g, 0.621 mmol). On cooling, a red solid came out of
solution, which was collected by filtration, washed with the
solvent, and dried. This was shown (MS, HPLC) to contain
mainly the desired product together with some of the mono-
quaternary compound. Attempts to purify it by crystallization
failed; therefore, it was purified by reverse phase preparative
HPLC. The product was isolated as a pink oil which was
disolved in the minimum amount of cold iPrOH and filtered
and the solvent removed in vacuo to yield a pink oil. This
solidified after drying in vacuo over P2O5 (0.22 g, 30.3%): mp
77-78 °C; 1H NMR (400 MHz, DMSO-d6) δ 1.28 (m, 12 H,
-CH2-), 1.64 (q, 4 H, -CH2-), 2.56 (m, 10 H, -CH2-, -CH3), 3.46
(q, 4 H, NH-CH2-), 4.29 (t, 6 H, Ph2CH, -CH2-N+), 6.87 (s, 2 H,
quinoline-H3), 7.20 (t, 4 H, Ph-Hp), 7.30 (t, 8 H, Ph-Hm), 7.39
(d, 8 H, Ph-Ho), 7.70 (t, 2 H, quinoline-H6 or -H7), 7.93 (t, 2 H,
quinoline-H7 or -H6), 7.98 (d, 2 H, quinoline-H5), 8.48 (d, 2 H,
quinoline-H8), 9.10 (t, 2 H, N-H); MS (FAB, matrix glycerol +
thioglycerol + TFA) [M]+ 844, fragments at m/z 649, 422, 380,
365, 200, 185; HPLC (column K, A:B ) 25:75) major peak at
18.35 min representing 99.6% of the absorption at 215 nm.
Anal. (C64H68F6N4O4‚0.4HI‚0.4CF3COOH‚H2O) C, H, N, I.
4,7-Dich lor o-1-m eth ylqu in olin iu m Iod id e (21c). A so-
lution of 4,7-dichloroquinoline (2.53 g, 12.7 mmol) and MeI (1.5
mL, 24.11 mmol) in MEK (20 mL) was stirred and heated
under reflux for 5 h. After cooling, the precipitate was
collected and purified by column chromatography using MeOH
as the eluant. The solid thus obtained was suspended in
boiling EtOH, filtered, and washed twice with hot EtOH to
yield a yellow powder (1.26 g, 24%): mp 240-242 °C; 1H NMR
(400 MHz, DMSO-d6) δ 3.94 (s, 3 H, CH3), 6.36 (d, J ) 7.5 Hz,
1 H, quinoline-H3), 7.56 (dd, J 1 ) 8.7 Hz, J 2 ) 1.8 Hz, 1 H,
quinoline-H6), 7.94 (d, J ) 1.7 Hz, 1 H, quinoline-H8), 8.22 (d,
J ) 8.7 Hz, 1 H, quinoline-H5), 8.24 (d, J ) 7.5 Hz, 1 H,
quinoline-H2); MS (FAB, matrix MNOBA) [M + I + H - Cl]+
303, [M + H]+ 211, fragment at m/z 177.
(c) MO Ca lcu la tion s. These were performed on the model
compounds of Table 2. The structures were initially built in
SYBYL 5.546 running on a Silicon Graphics IRIS 4D worksta-
tion and then submitted for semiempirical MO calculations
using the MOPAC 5.0 MO package and the AM134 Hamilto-
nian. The normal self-consistent field (SCF) convergence
procedure (default setting) was used, and full geometry
optimization of the structures was carried out. Note that
reliable energy minimization of the molecules cannot be
performed using the Tripos force field47 since it lacks an sp2-
charged nitrogen atom type which is present in these mol-
ecules. The charge on the system was set to 1. A Mulliken
population analysis was also performed, and the charges
obtained were used in the correlation studies. The MO
calculations were performed on the protonated forms of the
nonquaternary analogues 1a -6a .
Ack n ow led gm en t. This work was supported by the
Wellcome Trust (including Fellowships to C.A.D. and
P.M.D.). We thank Mr. C. S. Owen and Dr. B. Del Rey
Herrero for the synthesis of compounds 13 and 21,
respectively. We are grateful to Drs. J . G. Vinter and J .
A. D. Calder for helpful discussions on the MO calcula-
tions and to Professor D. H. J enkinson for helpful
discussions on the pharmacology.
Refer en ces
(1) Haylett, D. G.; J enkinson, D. H. In Potassium Channels; Cook,
N. S., Ed.; Ellis Horwood Ltd.: Chichester, 1990; pp 70-95.
(2) Lazdunski, M.; Romey, G.; Schmid-Antomarchi, H.; Renaud, J .
F.; Mourre, C.; Hugues, M.; Fosset, M. The Apamin-Sensitive
Ca2+-Dependent K+ Channel: Molecular Properties, Differentia-
tion, Involvement in Muscle Disease, and Endogenous Ligands
in Mammalian Brain. Handb. Exp. Pharmacol. 1988, 83, 135-
145.
(3) Ferrero, J . D.; Cocks, T.; Burnstock, G. A comparison between
ATP and bradykinin as possible mediators of the responces of
smooth muscle to non-adrenergic non-cholinergic nerves. Eur.
J . Pharmacol. 1980, 63, 295-302.
(4) Brown, C. M.; Burnstock, G. Evidence in support of the P1/P2
purinoceptor hypothesis in the guinea-pig taenia coli. Br. J .
Pharmacol. 1981, 73, 617-624.
1,10-Bis[N-(7-ch lor o-1-m eth ylqu in olin iu m -4-yl)a m in o]-
d eca n e Diiod id e Dih yd r a te (21). Compound 21c (1 g, 4.7
mmol), 1,10-diaminodecane (0.38 g, 2 mmol), and PhOH (2.3
g) were heated at 135 °C with stirring for 3.5 h. After cooling,
the reaction mixture was powdered, poured into EtOAc,
filtered, and washed twice with EtOAc. The resulting solid
was suspended in boiling EtOH, filtered, and washed with hot
EtOH. This procedure was repeated twice to yield a brown-
yellow powder (0.175 g, 18%): mp 272-274 °C; 1H NMR (400
MHz, DMSO-d6) δ 1.27-1.36 (m, 12 H, -CH2-), 1.67 (m, 4 H,
-CH2-), 3.44-3.51 (m, 4 H, NH-CH2), 4.10 (s, 6 H, CH3), 6.93
(5) Gater, P. R.; Haylett, D. G.; J enkinson, D. H. Neuromuscular
blocking agents inhibit receptor-mediated increases in the
potassium permeability of intestinal smooth muscle. Br. J .
Pharmacol. 1985, 86, 861-868.
(d, J ) 7.6 Hz, 2 H, quinoline-H3), 7.88 (dd, J 1 ) 9.0 Hz, J 2
)
(6) Coats, R. A. Effects of apamin on R-adrenoceptor-mediated
changes in plasma potassium in guinea-pigs. Br. J . Pharmacol.
1983, 80, 573-580.
1.9 Hz, 2 H, quinoline-H6), 8.18 (d, J ) 1.9 Hz, 2 H, quinoline-
H8), 8.58 (d, J ) 9.0 Hz, 2 H, quinoline-H5), 8.59 (d, J ) 7.4
Hz, 2 H, quinoline-H2), 9.34 (t, J ) 5.1 Hz, 2 H, NH); MS (FAB,
matrix glycerol + thioglycerol + TFA) [M + I]+ 651, [M - H]+
523; HPLC (column K, A:B ) 40:60) major peak at 10.9 min
representing 96.6% of the absorption at 254 nm. Anal.
(C30H38Cl2N4I2‚2.5H2O) C, N; H: calcd, 5.26; found, 4.79.
(b) Biologica l Testin g. Stock solutions (2 mM) of all
compounds were prepared in MeOH. For the free bases 1-6,
(7) Coats, R. A. The effects of adrenoceptor agonists and antagonists
on plasma potassium concentration in anaesthetized guinea-pigs,
rabbits and rats. Br. J . Pharmacol. 1985, 86, 827-836.
(8) Na˚nberg, E.; Connolly, E.; Nedergaard, J . Presence of Ca2+
-
dependent K+ channel in brown adipocytes: possible role in
maintenance of R1-adrenergic stimulation. Biochim. Biophys.
Acta 1985, 844, 42-49.
(9) Nohmi, M.; Kuba, K. (+)-Tubocurarine blocks Ca2+-dependent

370 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 2
Galanakis et al.
K+ channel of the bullfrog sympathetic ganglion cell. Brain Res.
1984, 301, 146-148.
(28) Galanakis, D.; Davis, C. A.; Del Rey Herrero, B.; Ganellin, C.
R.; Dunn, P. M.; J enkinson, D. H. Synthesis and structure-
activity relationships of deqalinium analogues as K+ channel
blockers. Investigations on the role of the charged heterocycle.
J . Med. Chem. 1995, 38, 595-606.
(10) Pennefather, P.; Lancaster, B.; Adams, P. R.; Nicoll, R. A. Two
distinct Ca-activated K currents in bullfrog sympathetic ganglion
cells. Proc. Natl. Acad. Sci. U.S.A. 1985, 82, 3040-3044.
(11) Bourque, C. W.; Brown, D. A. Apamin and d-tubocurarine block
the afterhyperpolarizationof rat supraoptic neurosecretory neu-
rons. Neurosci. Lett. 1987, 82, 185-190.
(29) Galanakis, D.; Davis, C. A.; Del Rey Herrero, B.; Ganellin, C.
R.; Dunn, P. M.; J enkinson, D. H. Synthesis and QSAR of
dequalinium analogues as K+
channel blockers. Investigations
(12) Zhang, L.; Kmjevic, K. Apamin depresses selectively the after-
hyperpolarization of cat spinal motoneurons. Neurosci. Lett.
1987, 74, 58-62.
(13) Banks, B. E. C.; Brown, C.; Burgess, G. M.; Burnstock, G.; Claret,
M.; Cocks, T. M.; J enkinson, D. H. Apamin blocks certain
neurotransmitter-induced increases in potassium permeability.
Nature 1979, 282, 415-417.
(14) Burgess, G. M.; Claret, M.; J enkinson, D. H. Effects of quinine
and apamin on the calcium-dependent potassium permeability
of mammalian hepatocytes and red cells. J . Physiol. (London)
1981, 317, 67-90.
on the role of 4-NH2 group. Bioorg. Med. Chem. Lett. 1995, 5,
559-562.
(30) Galanakis, D.; Calder, J . A. D.; Ganellin, C. R.; Owen, C. S.;
Dunn, P. M. Synthesis and quantitative structure-activity
relationships of dequalinium analogues as K+ channel block-
ers: Investigation into the role of the substituent at position 4
of the quinoline ring. J . Med. Chem. 1995, 38, 3536-3546.
(31) Deshpande, S. M.; Singh, A. K. Preparation of some N,N′-bis-
(4-quinaldino)/(4-quinazolino)-R,ω-diaminialkanes(HCl) as po-
tential antibacterial and antitubercular compounds. Agric. Biol.
Chem. 1971, 35, 119-121.
(15) Lazdunski, M. Apamin, a neurotoxin specific for one class of
Ca2+-dependent K+ channels. Cell Calcium 1983, 4, 421-428.
(16) Auguste, P.; Hugues, M.; Grave, B.; Gesquiere, J .-C.; Maes, P.;
Tartar, A.; Romey, G.; Schweitz, H.; Lazdunski, M. Leiurotoxin
I (Scyllatoxin), a peptide ligand for Ca2+-activated K+ channels.
J . Biol. Chem. 1990, 265, 4753-4759.
(17) Zerrouk, H.; Mansuelle, P.; Benslimane, A.; Rochat, H.; Martin-
Eauclaire, M. F. Characterization of a new leiurotoxin I-like
scorpion toxin PO5 from Androctonus mauretanicus mauretani-
cus. FEBS Lett. 1993, 320, 189-192.
(32) Singh, A. K. Antibacterial agents. III. Synthesis of N,N′-bis-(4-
quinolinyl-, 4-quinaldinyl-, 4-quinazolinyl or 9-acridinyl)polym-
ethylenediamines and their sulfonamides. Chem. Pharm. Bull.
1975, 23, 1869-1873.
(33) Surrey, A. R.; Cutler, R. K. The role of phenol in the reaction of
4,7-dichloroquinoline with novol diamine. J . Am. Chem. Soc.
1961, 73, 2623-2626.
(34) Dewar, M. J . S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J . J . P.
AM1: A New General Purpose Quantum Mechanical Molecular
Model. J . Am. Chem. Soc. 1985, 107, 3902-3909.
(35) Perrin, D. D. Dissociation constants of organic bases in aqueous
solution; Butterworths: London, 1965.
(36) Pearson, R. G. Recent advances in the concept of hard and soft
acids and bases. J . Chem. Ed. 1987, 64, 561-567.
(37) Pearson, R. G. The principle of maximum hardness. Acc. Chem.
Res. 1993, 26, 250-255.
(38) Parr, R. G.; Chattaraj, P. K. Principle of maximum hardness. J .
Am. Chem. Soc. 1991, 113, 1854-1856.
(39) Kubo, Y.; Baldwin, T. J .; J an, Y. N.; J an, L. Y. Primary structure
and functional expression of a mouse inward rectifier potassium
channel. Nature 1993, 362, 127-133.
(40) Ho, K.; Nichols, C. G.; Lederer, W. J .; Lytton, J .; Vassilev, P.
M.; Kanazirska, M. V.; Hebert, S. C. Cloning and expression of
an inwardly rectifying ATP-regulated potassium channel. Nature
1993, 362, 31-38.
(41) Butler, A.; Tsunoda, S.; McCobb, D. P.; Wei, A.; Salkoff, L. mSlo,
a Complex Mouse Gene Encoding “Maxi” Calcium-Activated
Potassium Channels. Science 1993, 261, 221-224.
(42) Pongs, O. Structural basis of voltage-gated K+ channel phar-
macology. Trends Pharmacol. Sci. 1992, 13, 359-365.
(43) Kubo, Y.; Reuveny, E.; Slesinger, P. A.; J an, Y. N.; J an, L. Y.
Primary structure and functional expression of a rat G protein-
coupled muscarinic potassium channel. Nature 1993, 364, 802-
806.
(18) Renaud, J . F.; Desnuelle, C.; Schmid-Antomarchi, H.; Hugues,
M.; Serratrice, G.; Lazdunski, M. Expression of apamin receptor
in muscles of patients with myotonic muscular dystrophy. Nature
1986, 319, 678-680.
(19) Behrens, M. I.; Vergara, C. Increase of apamin receptors in
skeletal muscle induced by colchicine: possible role in myotonia.
Am. J . Physiol. 1992, 263, C974-C802.
(20) Behrens, M. I.; J alil, P.; Serani, A.; Vergara, F.; Alvarez, O.
Possible role of apamin-sensitive K+ channels in myotonic
dystrophy. Muscle Nerve 1994, 17, 1264-1270.
(21) Yamamoto, H.-A.; Harris, R. A. Calcium-dependent 86Rb efflux
and ethanol intoxication: Studies of human red blood cells and
rodent brain synaptosomes. Eur. J . Pharmacol. 1983, 88, 357-
363.
(22) Demonchaux, P.; Ganellin, C. R.; Dunn, P. M.; Haylett, D. G.;
J enkinson, D. H. Search for the pharmacophore of the K+
channel blocker, apamin. Eur. J . Med. Chem. 1991, 26, 915-
920.
(23) Cook, N. S.; Haylett, D. G. Effects of apamin, quinine and
neuromuscular blockers on calcium-activated potassium chan-
nels in guinea pig hepatocytes. J . Physiol. 1985, 358, 373-394.
(24) J enkinson, D. H.; Haylett, D. G.; Cook, N. S. Calcium-activated
Potassium channels in liver cells. Cell Calcium 1983, 4, 429-
437.
(25) Castle, N. A.; Haylett, D. G.; Morgan, J . M.; J enkinson, D. H.
Dequalinium; a potent inhibitor of apamin-sensitive K+ channels
in hepatocytes and of nicotinic responses in skeletal muscle. Eur.
J . Pharmacol. 1993, 236, 201-207.
(44) Hille, B. Ionic channels of excitable membranes; 2nd ed.; Sinauer
Associaties Inc.: Sunderland, MA, 1992.
(45) Grogan, C. H.; Rice, L. M.; Sullivan, M. X. Diisothiuronium
dihydrohalide salts. J . Org. Chem. 1953, 18, 728-735.
(46) Sybyl Molecular Modelling; Tripos Associates, Inc.: St. Louis,
MO.
(47) Clark, M.; Cramer, R. D., III; Vann Opdenbosch, N. Validation
of the general-purpose TRIPOS 5.2 force-field. J . Comput. Chem.
1989, 10, 982-1012.
(26) Dunn, P. M. Dequalinium, a selective blocker of the slow
afterhyperpolarization in rat sympathetic neurones in cultures.
Eur. J . Pharmacol. 1994, 252, 189-194.
(27) Dunn, P. M.; Davis, C. A.; Ganellin, C. R.; Haylett, D. G.;
Morgan, J . M.; J enkinson, D. H. Potassium channel blocking
activity of dequalinium analogues in guinea-pig hepatocytes and
rat sympathetic neurones. Br. J . Pharmacol. 1991, 104, 165P.
J M950520I