Inhibition of protein kinase C by dequalinium analogues: Structure-activity studies on head group variations

Article abstract of DOI:10.1016/j.bmc.2006.07.067

New dequalinium analogues and related heteroaromatic systems were synthesized and evaluated for inhibition of protein kinase Cα. In vitro assays with recombinant human PKCα showed that the number of the aromatic ring head groups as well as their electron-

Full text of DOI:10.1016/j.bmc.2006.07.067

Bioorganic & Medicinal Chemistry 14 (2006) 7796–7803
Inhibition of protein kinase C by dequalinium analogues:
Structure–activity studies on head group variationsq
Chandima Abeywickrama, Susan A. Rotenberg and Arthur David Baker*
Department of Chemistry, The Graduate Center, The City University of New York, New York, NY 10016-4309, USA
Department of Chemistry and Biochemistry, Queens College of the City University of New York, Flushing, NY 11367-1597, USA
Received 22 June 2006; revised 30 July 2006; accepted 31 July 2006
Available online 7 September 2006
Abstract—New dequalinium analogues and related heteroaromatic systems were synthesized and evaluated for inhibition of protein
kinase Ca. In vitro assays with recombinant human PKCa showed that the number of the aromatic ring head groups as well as their
electron-richness, are critical factors that determine potency. The inhibitory strengths of the synthesized compounds are shown to
correlate well with Mulliken charges on the head group ring nitrogen atoms making it possible to design likely candidate molecules
having improved protein kinase Ca inhibitory activity.
ꢀ 2006 Elsevier Ltd. All rights reserved.
1. Introduction
and B.16 For example 1 has an IC50 = 7 lM in contrast
to that of A (IC50 = 117 lM) (Table 1). The significantly
large difference led to the proposal of a ‘‘two point’’
binding model in which optimal activity occurs when
the two head groups have coincident contact with two
target sites in the catalytic domain of PKCa.16 In addi-
tion, the inhibitory activity was shown to increase with
increasing linker length.16 The incorporation of a single
trans double bond in the center of the C10-linker yielded
an inhibitory activity almost equivalent to that of the
C10 saturated analogue, while the cis analogue was sig-
nificantly weaker as an inhibitor. This finding suggested
that when bound at its target site in PKCa, the dequalin-
ium ion is positioned in a trans-oid geometry.16
The role of protein kinase C (PKC) in signaling path-
ways that govern the proliferation, differentiation, and
metastasis of cells makes this enzyme an attractive target
for the design of chemotherapeutic agents.1–10 Substan-
tial effort has been directed over the years to inhibit
PKC.11–13 Among the most potent leads, bis-quinolini-
um compounds have received much attention.14–19 An
initial discovery15 in 1990 that dequalinium diiodide 1,
a bis-quinolinium compound, inhibits PKC activity sug-
gested that studies of structurally similar salts would be
useful. Besides PKC inhibition, these compounds also
exhibit an array of biological effects such as anti-micro-
bial16,20 and fungicidal activity,21,22 potent anti-tumor
activity,23–25 potent and selective blockade of small
conductance Ca2+-activated K+ channels,26–32 and
anti-malarial effects.33
This paper describes our continuing efforts to elucidate
structure–activity relationships (SARs) of dequalinium
and dequalinium-type analogues with respect to in vitro
inhibition of PKC. In this study, we have focused on the
effects of structural variations in the head groups while
keeping the linkage constant, namely [(CH2)10], with
the aim of gaining insights into the rational design of
new PKCa inhibitors.
We earlier reported that the PKC inhibitory action of
bis-quaternary dequalinium salts is considerably greater
than that of the corresponding mono-quaternary salts A
Keywords: PKC Inhibition; Head group role; Substituent effects;
Quinolinium analogues.
2. Chemistry
q
The authors gratefully dedicate this article to Professor Koji
Nakanishi on receipt of the Tetrahedron Prize.
Compounds 1–5 were prepared by reaction of the
appropriate quinoline derivative with 1,10-diiododecane
in 2-butanone or 4-methyl 2-pentanol (Scheme 1).
Similar procedures were used for the synthesis of
*
Corresponding author. Tel.: +1 718 997 4219; fax: +1 718 997
5531; e-mail: arthur.baker@qc.cuny.edu
Present address: Columbia University, Department of Chemistry,
3000 Broadway, New York, NY 10027, USA.
0968-0896/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2006.07.067

C. Abeywickrama et al. / Bioorg. Med. Chem. 14 (2006) 7796–7803
Table 1. IC50 values of compounds studied
7797
Compound
Structure
Substituents
IC50 (lM)
Aa
117
8
Ba
3590 510
C
D
>250
94 16
R1
R2
1
2
3
4
5
CH3
H
NH2
NH2
N(CH3)2
H
7
30
1
8
H
29 12
CH3
H
22
72
2
6
H
6
7
30 11
30
8
X
R3
8
O
CH3
CH3
NH2
36 11
9
10
S
NCH3
17
14
6
1
R4
NH2
11
12
112
231 26
4
N(CH3)2
13
16
7
14
48
4
The assay studies reported in above table were performed under reversible binding conditions as described in Section 6. IC50 represent the analogue
concentrations needed for 50% inhibition, and are the average of two or more independent experiments, where PKC activity was analyzed in
triplicate measurements.
a Previously reported.16

7798
C. Abeywickrama et al. / Bioorg. Med. Chem. 14 (2006) 7796–7803
4. Results and discussion
Most of the compounds analyzed in this study are struc-
tural analogues of the parent compound 1. Compounds
A and B are mono-quaternized quinoline derivatives and
were included to provide reference benchmarks for PKC
inhibition (Table 1).16 For purposes of comparison, two
aliphatic chain containing salts C and D lacking het-
eroaromatic head groups were also examined.
Our initial studies on the head groups were designed to
explore the relative importance of the 2-methyl (R1) and
4-amino (R2) groups as contributors to inhibitory prow-
ess in bis-quinolinium compounds. Dequalinium ana-
logues 1–5 were synthesized in which these groups
were replaced by hydrogen or other groups. The IC50
values of analogues 2 and 4 were both in the range
22–30 lM, indicating a weaker inhibition than that of
1. For analogue 5, in which both the 2-methyl and the
4-amino groups were replaced with H, a 10-fold lower
potency than 1 was observed. These results suggest that
both 2-methyl and 4-amino groups contribute to inhib-
itory activity.
Scheme 1. Synthesis of dequalinium analogues. Reagents and condi-
tions: (a) 2-butanone/4-methyl 2-pentanol, reflux, 48–72 h.
The role of the 4-amino group was further investigated.
In principle its contribution toward PKC inhibition
could be related to its hydrogen bonding ability and/or
to its ability to donate electron density into the aromatic
head group via resonance. Analogue 3 was prepared to
investigate these possibilities. This compound contains
a 4-N,N-dimethylamino (–NMe2) group rather than 4-
amino. While the –NMe2 group is approximately equiv-
alent to –NH2 group in terms of its ability to donate
electron density into the aromatic ring, it cannot func-
tion as a hydrogen bond donor unlike the –NH2 group.
Revealingly (Table 1) the inhibitory activity of 3 was
found to be essentially the same as that of 2, implying
that the electron-donating ability (by delocalization) of
the 4-substituent (R2) is more significant than its ability
to function as a hydrogen bond donor. The 2-methyl
group is an electron-donating group, albeit inferior to
the 4-amino group. However, removal of the methyl
group from parent compound 1 gives analogue 2, that
has almost the same IC50 as 4, which is obtained when
the 4-amino group is removed from 1. This seems to re-
flect the closer proximity of methyl to the ring nitrogen.
In summary, these results suggested to us that the degree
of electron-richness in the heteroaromatic head group
correlates with PKC inhibitory ability.
Scheme 2. Synthesis of exocyclic dequalinium analogue 6. Reagents
and conditions: (a) phenol, 180 ꢁC; (b) acetone, reflux, 3d.
compounds 7–12. Corresponding starting head groups
for compounds 2 and 3 (4-aminoquinoline and 4-N,N-
dimethylaminoquinoline) were synthesized by nucleo-
philic displacement of the chlorine atom from 4-chloro-
quinoline by ammonia and N,N-dimethylamine
respectively, by following a modified literature proce-
dure.34 Scheme 2 shows the preparation of exocyclic
dequalinium analogue 6 from 4-chloroquinaldine by
reaction with 1,6 diaminohexane. This reaction pro-
duced bis(quinaldinylamino)hexane, which was then
subjected to quaternization with methyl iodide to give 6.
3. Biological activity
Recombinant human PKCa (95% pure) was used for
testing the inhibitory action of dequalinium analogues
1–12, and salts 13, 14, C, and D in vitro (Table 1).
The total catalytic activity of PKCa was analyzed in
triplicate at increasing concentrations of a selected
dequalinium analogue. When compared to the level
of activity measured in the absence of an analogue,
the extent to which a given concentration of an ana-
logue decreased the formation of phosphorylated pep-
tide indicated the extent of PKCa inhibition. The
error for these assays was typically less than 10%.
IC50 values were calculated from dose–response curves
as that concentration producing 50% inhibition. This
analysis was carried out at least twice for each ana-
logue and the IC50 values were averaged and are given
in Table 1.
We therefore decided to seek a parameter that would
quantitatively measure the electron-richness of the head
groups represented in compounds 1–5. We expected and
found that the magnitude of the charge on N1 (Mulliken
charges) to be a suitable parameter.26,28,29 The needed
Mulliken charges were obtained from HF/6-31G MO
calculations using Gaussian 03. To simplify the compu-
tations, the calculations were not performed on the
whole bis-quinolinium compounds but rather on model
compounds consisting of only one of the quinolinium
head groups in which a simple methyl group was the
quaternizing group in place of the C10 methylene linker
present in the actual analogues (Table 2). Since there is

C. Abeywickrama et al. / Bioorg. Med. Chem. 14 (2006) 7796–7803
7799
Table 2. Mulliken Charges on N1 of the model compounds used for correlation in Figure 1
Compounda
Structure
Substituents
N1 Charge
Ca
Da
ꢀ0.555
1.135b
R1
CH3
H
R2
1a
2a
3a
4a
5a
NH2
NH2
N(CH3)2
H
ꢀ0. 790
ꢀ0. 743
ꢀ0. 720
ꢀ0. 764
ꢀ0. 683
H
CH3
H
H
7a
ꢀ0.723
X
O
S
R3
8a
9a
CH3
CH3
NH2
ꢀ0.713
ꢀ0.633
ꢀ0.776
10a
NCH3
R4
NH2
11a
12a
ꢀ0.624
ꢀ0.624
N(CH3)2
13a
14a
ꢀ0.771
ꢀ0.814
HF/6-31G molecular orbital (MO) calculations were performed on model methylated quaternized derivatives as shown in this table using Gaussian
03.
a Model compound numbers correspond to the compound numbers of Table 1.
b Charge on P.
no electronic interaction between the two quinolinium
head groups that are well separated by a C10 saturated
linker, this is a valid approximation. Any errors intro-
duced (due to different inductive and hyperconjugation
effects) are expected to be common for all the com-
pounds and the results should be comparable.26–28,35
The calculated Mulliken charges on the ring nitrogen
atoms for the model compounds are summarized in Ta-
ble 2. A good correlation between IC50 values and the
Mulliken charges on N1 was obtained for compounds1,
2, 4, and 5 as shown in Figure 1. Here it can be seen that
the analogues with lower IC50 values showed a larger
negative Mulliken charge on N1 and those with higher
IC50 values showed a less negative charge on N1.36
To determine whether this correlation could be used to
predict the potency of novel analogues, we designed,
synthesized, and studied a variety of electron-rich ana-
logues (compounds 6–10) in which the head groups con-
sisted of rings other than quinoline, or in which the
heteroaromatic rings are connected via the 4-amino
Figure 1. Plot of IC50 values of compounds studied (Table 1) vs
Mulliken charges on N1 of corresponding model compounds in Table
2. All compounds except 6 (exocyclic dequalinium analogue) were
included into the calculation of R2.
electron density (compounds 11 and 12) that therefore
were expected to produce weaker inhibition. In all cases,
the IC50 values for all compounds exhibited good corre-
lation with the magnitude of the N1 Mulliken charge
(Fig. 1). The weaker PKCa inhibitors produced with
groups rather than the ring nitrogen atoms. As expected,
the calculated Mulliken charges for most of these newly
designed model compounds were in agreement with a
more negative charge density (Table 2). Preparation
was carried out of two pyridine analogues having lower

7800
C. Abeywickrama et al. / Bioorg. Med. Chem. 14 (2006) 7796–7803
pyridine head groups (compounds 11 and 12) are consis-
tent with the diminished electron-richness of these rings
as compared to quinoline rings. In contrast, compounds
7–10 with isoquinoline, benzoxazole, benzthiazole, and
benzimidazole head groups, respectively, have similar
elevated PKC inhibition to those of quinoline deriva-
tives 1–5. Compound 6 deviates a little from the gener-
ally linear correlation of Figure 1, but this is hardly
surprising as its structure is significantly different from
those of analogues linked through their ring nitrogen
atoms.
the nature of the linkage of head groups is tolerated
since two head groups linked by atoms other than the
ring N atoms, as in 6, do not appreciably alter inhibitory
strength.
6. Experimental
6.1. Assay of inhibition of catalytic activity
Recombinant human PKCa (95% pure) (Pan Vera
Corp., Madison, WI) was used for testing dequalinium
analogues in vitro. Because of the purity and high basal
activity of the enzyme preparation, it was possible to
conduct assays in the absence of phosphatidylserine
which is customarily added to assist in the activation
of the enzyme.16 It was excluded from these assays in or-
der to avoid non-specific hydrophobic interactions that
could pose a variable due to the different head groups
represented in this collection of analogues.
We had hoped to round out our studies by examining
other dequalinium analogues of type 1–5 with a range
of R2 groups (electron-donating and electron-withdraw-
ing) in the 4-position of the quinoline ring. However, at-
tempts to prepare an analogue with R2 = –OCH3 were
unsuccessful because reaction of 4-methoxyquinoline
with 1,10 diiododecane results in the formation of 1,10-
(decane-1,10-diyl) bis(4-(1H)-quinolone) instead of the
intended dequalinium analogue.27 In addition, attempts
failed to synthesize dequalinium analogues containing
electron-withdrawing groups (such as R2 = –CO2Et or
–Cl), presumably because such groups cause the quino-
line ring nitrogen atom to be so weakly nucleophilic to-
ward alkylating agents that no evidence of reaction
could be detected even after many weeks of reflux.27,28
Therefore, all compounds successfully prepared in these
studies possessed R2 substituents that were either elec-
tron-donating or neutral.
To test the inhibitory potency of these compounds, the
total catalytic activity of PKCa was analyzed in tripli-
cate with increasing concentrations of a selected dequa-
linium analogue. PKCa activity was measured in a
reaction volume of 0.12 mL consisting of 20 mM Tris,
pH 7.4, 10 mM Mg2+, 0.5 mM Ca2+, 26.6 lM peptide
substrate (RFARKGSLRQKNV), PKCa (28 ng per as-
say), 66 lM [c-32P]ATP, and 5 lL dequalinium analogue
(or 5 lL DMSO as control) added to the specified con-
centration. Stock solutions for dequalinium analogues
were prepared in DMSO and standardized spectropho-
tometrically on a Perkin-Elmer Lambda II spectropho-
tometer. The extinction coefficient for each compound
is given in the Supplemental information.
Based on our finding of a correlation between electron
density of the head groups and the corresponding PKCa
inhibitory potency, it should be possible to make predic-
tions about the PKC activity of other alkylated nitrogen
heterocycles having no direct structural relationship to
dequalinium salts. To test this correlation, we examined
propidiumiodide, 13, and safranine O, 14,37 which
would be predicted to have appreciable PKC inhibitory
activity due to the presence of both multiple electron-do-
nating (NH2 and CH3) groups and several aromatic
rings to augment electron-richness. Indeed, both these
compounds were found to be much more potent as
PKC inhibitors than previously investigated compounds
such as A and B which also contain a single heteroaro-
matic head group. In fact, compound 13 proved to be al-
most as active as dequalinium diiodide itself (see Table
1). This finding suggests that preparation of an analogue
based on 13 with two heteroaromatic head groups rather
than one would be a worthwhile objective for future
studies.
Each phosphotransferase reaction was initiated by the
addition of [c-32P]ATP to consecutive tubes in staggered
15-s intervals. After a 10 min reaction period, each tube
was quenched in the same consecutive style by applying
100 lL of the reaction medium to a 2 · 2 cm square of
phosphocellulose paper. The squares retained only the
peptide following five wash steps with 1 L tap water,
and the radioactive content of each square was analyzed
by b-scintillation counting. When compared to the level
of activity measured in the absence of an analogue, the
extent to which a given concentration of an analogue de-
creased the formation of phosphorylated peptide indi-
cated the extent of PKCa inhibition. The error for
triplicate assays was typically less than 10%. IC50 values
were calculated from dose–response curves as that con-
centration producing 50% inhibition. This analysis was
carried out at least twice for each analogue and the
IC50 values were averaged.
5. Conclusion
By SAR studies we demonstrated that there is a good
correlation between the PKC potency and the N1 charge
for the compounds studied. The greater the electron
density in the head group, the more negative is the N1
charge and consequently the higher potency of PKCa
inhibition. The influence of R2 was found to be electron-
ic via delocalization of positive charge thus making the
head groups more electron-rich. In addition, varying
6.2. General experimental methods
1H NMR and 13C NMR spectra were recorded at
400 MHz and 100 MHz on a Bruker spectrometer,
respectively. Chemical shifts were expressed in ppm rel-
1
ative to TMS (0.00 ppm), MeOD (3.31 ppm for H and
49.01 ppm for 13C), DMSO (2.50 ppm for 1H and
39.51 ppm for 13C) or CDCl3 (7.26 ppm for 1H and

C. Abeywickrama et al. / Bioorg. Med. Chem. 14 (2006) 7796–7803
7801
77.22 ppm for 13C). Elemental analyses were performed
by Desert Analytics Laboratory, Arizona. Low and high
resolution mass spectra were recorded by use of electro-
spray ionization at core facilities at Hunter College of
the City University of New York and at University of
Illinois Urbana-Champaign. Reaction solvents were dis-
tilled under N2 as follows: hexane and Et2O from sodi-
um and benzophenone immediately before use,
methanol from magnesium, CH2Cl2, and CHCl3 from
calcium hydride (CaH2). 2 M ammonia in methanol
was obtained from Aldrich in SureSealTM bottles. TLC
was carried out with Merck 60F254 (0.25 mm thick)
sheets. All air- and water-sensitive reactions were per-
formed under N2 in oven- or flame-dried glassware.
J = 8.4 Hz), 8.08 (d, H, J = 8.4 Hz), 8.64 (d, H,
J = 4.8 Hz); 13C NMR (DMSO) d 43.6, 107.3, 122.2,
124.2, 124.6, 128.8, 129.6, 149.5, 150.5, 156.8; LRMS
(MH+) calcd for C11H12N2 m/z 173.1, found 173.1.
6.2.4. 1,10-(1,10-Decanediyl)bis[4-aminoquinolinium] di-
iodide (2). The pale yellow precipitate isolated was 67%.
1H NMR (DMSO) d 1.17 (m, 12H), 1.74 (m, 4H), 4.5 (t,
4H, J = 7.1 Hz), 6.78 (d, 2H, J = 7.1 Hz), 7.73 (t, 2H,
J = 7.7 Hz) 8.02 (t, 2H, J = 7.7 Hz), 8.15 (d, 2H,
J = 8.6 Hz), 8.47 (d, 2H, J = 8.6 Hz), 8.52 (d, 2H,
J = 7.1 Hz), 9.02 (br s, 4H); 13C NMR (DMSO) d
25.5, 28.2, 28.4, 28.5, 53.5, 101.7, 116.8, 118.0, 124.2,
126.1, 134.2, 137.7, 145.9, 157.5; Anal. Calcd for
C28H36N4I2: C, 49.28; H, 5.32; N, 8.21; I, 37.19. Found:
C, 49.15; H, 5.46; N, 8.09; I, 37.55. LRMS (M-2I)/2
calcd for C28H36N4I2 m/z 214.2, found 214.2.
6.2.1. 1,10-(1,10-Decanediyl)bis[4-amino-2-methyl quino-
linium] diiodide (1). 2.4 equivalents of 4-aminoquinaldine
(0.5 g, 3.16 mmol) was refluxed with 1,10-diiododecane
(0.518 g, 1.32 mmol) in 2-butanone for 48 h. The reac-
tion was monitored by TLC using 2 M ammonia in
MeOH/CH2Cl2 0.25:0.75 (Rf 0.45) and the presence of
1,10-diiododecane was observed by hexane. The precip-
itate obtained was filtered, washed thoroughly with 2-
butanone, acetone, and anhyd Et2O to yield 0.75 g
(80%) of 1 as a pale yellow precipitate. 1H NMR
(DMSO) d 1.31 (m, 10H), 1.72 (m, 6H), 2.74 (s, 6H),
4.47 (t, 4H, J = 7.9 Hz), 6.65 (s, 2H), 7.73 (t, 2H,
J = 7.8 Hz), 8.03 (t, 2H, J = 7.8 Hz), 8.17 (d, 2H,
J = 8.4 Hz), 8.45 (d, 2H, J = 8.4 Hz), 8.84 (br s, 4H);
13C NMR (DMSO) d 9.2, 21.6, 25.9, 29.8, 32.8, 47.9,
103.9, 116.6, 118.4, 124.3, 125.9, 134.4, 139.0, 155.0,
156.7; LRMS (Mꢀ2I/2) calcd for C30H40N4I2 m/z
228.2, found 228.2. Compounds 2–12 were synthesized
following the procedure described above.
6.2.5.
1,10-(1,10-Decanediyl)bis[4-N,N,dimethylamino-
quinolinium] diiodide (3). The pale yellow precipitate iso-
lated was 73%. 1H NMR (DMSO) d 1.24 (m, 12H), 1.79
(m, 4H), 3.46 (s, 12H), 4.56 (t, 4H, J = 7.1 Hz), 7.01 (d,
2H, J = 7.5 Hz), 7.70 (t, 2H, J = 7.7 Hz), 8.01 (t, 2H,
J = 7.7 Hz), 8.15 (d, 2H, J = 8.6 Hz), 8.47 (d, 2H,
J = 8.6 Hz), 8.61 (d, 2H, J = 7.5 Hz); 13C NMR (CDCl3)
d 26.1, 28.4, 28.5, 28.9, 45.0, 55.0, 104.4, 117.9, 119.6,
125.7, 127.9, 134.0, 139.1, 146.3, 160.2; HRMS (MꢀI)
calcd for C32H44N4I2 m/z 611.2611, found 611.2640.
6.2.6. 1,10-(1,10-Decanediyl)bis[2-methylquinolinium] di-
iodide (4). The pale yellow precipitate isolated was 70%.
1H NMR (DMSO) d 1.35 (m, 8H), 1.57 (m, 4H), 1.90
(m, 4H), 3.12 (s, 6H), 4.92 (m, 4H), 8.00 (t, 2H,
J = 7.6 Hz), 8.13 (d, 2H, J = 8.4 Hz), 8.24 (t, 2H,
J = 7.6 Hz), 8.42 (d, 2H, J = 8.0 Hz), 8.59 (d, 2H,
J = 8.0 Hz), 9.11 (d, 2H, J = 8.4 Hz); 13C NMR
(DMSO) d 22.5, 25.9, 28.1, 28.6, 28.9, 51.4, 118.9,
125.6, 128.2, 129.0, 130.6, 135.3, 138.2, 145.7, 160.5;
Anal. Calcd for C30H38N2I2: C, 52.95; H, 5.63; N,
4.12. Found: C, 52.87; H, 5.86; N, 3.89. HRMS (MꢀI)
calcd for C30H38N2I2 m/z 553.2080, found 553.2100.
6.2.2. 4-Aminoquinoline. 4-Chloroquinoline (3 g,
0.018 mol) was heated to 180 ꢁC with approximately
9 g of phenol. Ammonia dried over quicklime was
passed through the solution for 3 h. The hydrochloride
of the amine was separated and the excess of phenol
was removed by steam-distillation.34 The clear solution
was concentrated by evaporation. The resulting solution
was cooled, made alkaline with NaOH, and the crude
obtained was recrystallized with MeOH to afford
2.26 g (87%) of 4-aminoquinoline as a pale yellow
solid. 1H NMR (DMSO) d 6.55 (d, H, J = 4.8 Hz),
6.77 (br s, 2H), 7.38 (t, H, J = 7.6 Hz), 7.59 (t, H,
J = 7.6 Hz ), 7.76 (d, H, J = 8.0 Hz), 8.15 (d, H,
J = 8.0 Hz), 8.32 (br s, H); 13C NMR (DMSO) d
102.3, 118.6, 122.3, 123.4, 128.6, 128.8, 148.8, 150.3,
151.4; LRMS (MH+) calcd for C9H8N2 m/z 145.1, found
145.1. 4-N,N-dimethylaminoquinoline and bis-(qui-
naldinylamino)hexane were synthesized following the
procedure described above.
6.2.7. 1,10-(1,10-Decanediyl)bis[quinolinium] diiodide (5).
The pale yellow precipitate isolated was 73%. H NMR
1
(DMSO) d 1.29 (m, 12H), 1.93 (m, 4H), 5.04 (t, 4H,
J = 7.6 Hz), 8.06 (t, 2H, J = 7.6 Hz), 8.20 (t, 2H,
J = 7.6 Hz) 8.30 (t, 2H, J = 8.1 Hz), 8.49 (d, 2H,
J = 7.8 Hz), 8.65 (d, 2H, J = 8.6 Hz), 9.29 (d, 2H,
J = 8.6 Hz), 9.54 (d, 2H, J = 5.8 Hz); 13C NMR
(DMSO) d 25.8, 28.5, 28.8, 29.5, 57.3, 118.9, 122.1,
129.7, 129.9, 130.8, 135.6, 137.4, 147.4, 149.6; HRMS
(MꢀI) calcd for C28H34N2I2 m/z 525.1767, found
525.1783.
6.2.8. 1,6-Bis[N-(1-methylquinolinium-2-methyl)amino]-
hexane diiodide (6). 1,6-Diaminohexane (0.61 g,
5.25 mmol) was added dropwise to the reaction mixture
containing 4-chloroquinaldine (2.06 g, 11.6 mmol) for
1 h. The crude product was recrystallized with ethanol
to afford 1.46 g (70%) of bis-(quinaldinylamino)hexane
as pale yellow crystals. H NMR (DMSO) d 1.45 (m,
4H), 1.70 (m, 4H), 2.98 (s, 6H), 3.52 (m, 4H), 7.3 (s,
2H), 7.90 (t, 2H, J = 7.6 Hz), 8.00 (t, 2H, J = 8.4 Hz),
6.2.3. 4-N,N-Dimethylaminoquinoline. This compound
was synthesized by nucleophilic displacement of chloride
ion in 4-chloroquinoline with N,N-dimethylamine. The
alkaline solution was extracted to CH2Cl2 and dried to
afford 67% of 4-N,N-dimethylaminoquinoline as a
brown liquid. H NMR (DMSO) d 2.97 (s, 6H), 6.85
(d, H, J = 4.8 Hz), 7.52 (dt, H, J = 1.2 Hz, J = 8.4 Hz),
7.69 (dt, H, J = 1.2 Hz, J = 8.4 Hz), 7.98 (d, H,
1
1

7802
C. Abeywickrama et al. / Bioorg. Med. Chem. 14 (2006) 7796–7803
8.48 (d, 2H, J = 6.8 Hz), 8.73 (d, 2H, J = 8.4 Hz), 9.56
(br s, 2H); 13C NMR (DMSO) d 26.0, 27.5, 42.7, 97.9,
116.7, 120.3, 123.5, 126.2, 133.0, 138.1, 142.3, 153.7,
155.1; HRMS (MH+) calcd for C26H30N4 m/z
399.2549, found 399.2568. Bis-(quinaldinylamino) hex-
ane (0.246 g, 0.618 mmol) was refluxed with acetone
(10 mL) and excess of MeI (0.25 g, 1.76 mmol) for 3d
in order to methylate the ring nitrogen. The crude ob-
tained was recrystallized with methanol to afford
0.34 g (80%) of 6 as a pale yellow precipitate. 1H
NMR (DMSO) d 1.47 (m, 4H), 1.72 (m, 4H), 2.78 (s,
6H), 3.52 (m, 4H), 3.99 (s, 6H), 6.96 (s, 2H), 7.73 (t,
2H, J = 7.6 Hz), 8.01 (t, 2H, J = 7.6 Hz), 8.16 (d, 2H,
J = 8.4 Hz), 8.51 (d, 2H, J = 8.4 Hz), 9.01 (br s, 2H);
13C NMR (DMSO) d 22.3, 26.1, 27.7, 36.4, 42.9,
100.2, 116.9, 118.5, 123.4, 126.0, 133.7, 139.4, 153.9,
156.5 Anal. Calcd for C28H36N4I20.3HCl: C, 48.49; H,
5.24; N, 8.08. Found: C, 48.20; H, 4.99; N, 7.95. HRMS
(MꢀI) calcd for C28H36N4I2 m/z 555.1985 found
555.2000.
for C26H38N6I2: C, 45.36; H, 5.56; N 12.21; I, 36.87.
Found: C, 45.50; H, 5.59; N 11.86; I, 36.40. HRMS
(MꢀI) calcd for m/z C26H38N6I2 561.2203, found
561.2216.
Acknowledgments
We thank PSC-CUNY for financial support (Grant
# 66380) and Dr. Sergei V. Dzyuba for helpful
discussions.
Supplementary data
NMR spectral data for compounds 11, 12, C, and D and
wavelengths (kmax), and molar extinction coefficients (e)
of compounds studied. This material is available free of
charge via the Internet at http://www.sciencedirect.com.
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.bmc.
2006.07.067.
6.2.9. 1,10-(1,10-Decanediyl)bis[1-amino isoquinolinium]
diiodide (7). The pale yellow precipitate isolated was
79%. 1H NMR (MeOD) d 1.41 (m, 12H), 1.88 (m,
4H), 4.27 (t, 4H, J = 7.7 Hz), 7.25 (d, 2H, J = 7.3 Hz),
7.70 (d, 2H, J = 7.7 Hz), 7.79 (m, 2H), 7.94 (m, 4H),
8.48 (d, 2H, J = 9.0 Hz) (NH2 signals were not observed
due to exchange with the solvent); 13C NMR (MeOD) d
27.4, 28.7, 30.4, 30.5, 54.9, 113.9, 120.1, 126.1, 128.9,
130.5, 133.6, 135.7, 137.5, 155.2; HRMS (MꢀI) calcd
for C28H36N4I2 m/z 555.1987, found 555.2007.
References and notes
1. Rotenberg, S. A.; Weinstein, I. B.. In Biochemical and
Molecular Aspects of Selected Cancers; Pretlow, T. G., II,
Pretlow, T. P., Eds.; Academic Press: Orlando, 1991; Vol.
1, pp 25–73.
2. Rotenberg, S. A.; Zhu, J.; Hansen, H.; Li, X.; Sun,
X.; Michels, C. A.; Riedel, H. J. Biochem. 1998, 124,
756.
3. Krauss, R. S.; Housey, G. M.; Hsiao, W. L. W.; Johnson,
M. K.; Rotenberg, S. A.; Borner, C. M. B.; Weinstein, I.
B. Prog. Clin. Biol. Res. 1990, 340, 175.
4. Sun, X.; Rotenberg, S. A. Cell Growth Differ. 1999, 10,
343.
5. Sanz-Navares, E.; Fernandez, N.; Kazanietz, M. G.;
Rotenberg, S. A. Cell Growth Differ. 2001, 12, 517.
6. Hanauske, A. R.; Sundell, K.; Lahn, M. Curr. Pharm.
Des. 2004, 10, 1923.
6.2.10. 1,10-(1,10-Decanediyl)bis[2-methylbenzoxazolium]
diiodide (8). The orange precipitate isolated was 80%. 1H
NMR (MeOD) d 1.41 (m, 12H), 2.00 (m, 4H), 3.14 (s,
6H), 4.61 (t, 4H, J = 7.8 Hz), 7.80 (m, 4H), 8.00 (m,
2H), 8.08 (m, 2H); 13C NMR (DMSO), d 13.4, 27.2,
28.9, 29.7, 30.0, 48.1, 113.7, 115.2, 128.9, 129.9, 130.7,
149.2, 169.5; Anal. Calcd for C26H34N2O2I2: C, 47.29;
H, 5.19; N, 4.24; I, 38.43. Found: C, 47.05; H, 5.35; N
3.86; I, 38.65. LRMS (Mꢀ2I)/2 calcd for C26H34N2O2I2
m/z 203.2, found 203.2.
7. Silva, D. Curr. Cancer Drug Targets 2004, 4, 327.
8. Chung, S. H.; Polgar, J.; Reed, G. L. J. Biol. Chem. 2000,
275, 25286.
9. House, C.; Kemp, B. E. Science 1987, 238, 1726.
10. Parker, P. J.; Coussens, L.; Totty, N.; Rhee, L.; Young, S.;
Chen, E.; Stabel, S.; Waterfield, M. D.; Ullrich, A. Science
1986, 233, 853.
11. Rotenberg, S. A.; Calogeropoulou, T.; Jaworski, J. S.;
Weinstein, I. B.; Rideout, D. Proc. Natl. Acad. Sci. U.S.A.
1991, 88, 2490.
12. Rotenberg, S. A.; Huang, M. H.; Zhu, J.; Su, L.; Riedel,
H. Mol. Carcinog. 1995, 12, 42.
13. Tammela, P.; Ekokoski, E.; Garcia-Horsman, A.; Talman,
V.; Finel, M.; Tuominen, R.; Vuorela, P. Drug Dev. Res.
2004, 63, 76.
14. Rotenberg, S. A.; Baker, A. D. U.S. Patent 2002114769,
2002.
15. Rotenberg, S. A.; Smiley, S.; Ueffing, M.; Krauss, R. S.;
Chen, L. B.; Weinstein, B. Cancer Res. 1990, 50, 677.
16. Qin, D.; Sullivan, R.; Berkowitz, W. F.; Bittman, R.;
Rotenberg, S. A. J. Med. Chem. 2000, 43, 1413.
17. Rotenberg, S. A.; Sun, X.-g. J. Biol. Chem. 1998, 273,
2390.
6.2.11. 1,10-(1,10-Decanediyl)bis[2-methylbenzothiazoli-
um] diiodide (9). The purple precipitate isolated was
87%. H NMR (DMSO) d 1.28 (m, 8H), 1.31 (m, 4H),
1.84 (m, 4H), 3.22 (s, 6H), 4.71 (t, 4H, J = 7.8 Hz),
7.81 (m, 2H), 7.90 (m, 2H), 8.35 (d, 2H, J = 8.2 Hz),
8.46 (d, 2H, J = 8.2 Hz); 13C NMR (DMSO) d 16.9,
25.9, 27.8, 28.6, 28.7, 49.2, 116.9, 124.6, 128.1, 129.1,
129.4, 140.8, 177.0; Anal. Calcd for C26H34N2S2I2: C,
45.09; H, 4.95; N 4.05; I, 36.65. Found: C, 45.09; H,
4.84; N 3.91; I, 36.34. HRMS (MꢀI) calcd for
C26H34N2S2I2 m/z 565.1242, found 565.1226.
1
6.2.12. 1,10-(1,10-Decanediyl)bis[2-amino-1-methylbenz-
imidazolium] diiodide (10). The pale yellow precipitate
1
isolated was 77%. H NMR (MeOD) d 1.30 (m, 12H),
1.78 (m, 4H), 3.71 (s, 6H), 4.17 (t, 4H, J = 7.3 Hz),
7.38 (m, 4H), 7.48 (m, 4H) (NH2 signals were not ob-
served due to exchange with the solvent); 13C NMR
(MeOD) d 28.0, 29.5, 30.3, 30.8, 30.9, 44.8, 111.7,
111.9, 125.5, 125.6, 132.5, 132.2, 151.6; Anal. Calcd

C. Abeywickrama et al. / Bioorg. Med. Chem. 14 (2006) 7796–7803
7803
18. Rotenberg, S. A.; Zhu, J.; Hansen, H.; Li, X.-d.; Sun, X.-
g.; Michels, C. A.; Riedel, H. J. Biochem. 1998, 124, 756.
19. Sullivan, R. M.; Stone, M.; Marshall, J. F.; Uberall, F.;
Rotenberg, S. A. Mol. Pharmacol. 2000, 58, 729.
20. Babbs, M.; Collier, H. O. J.; Austin, W. C.; Potter, M. D.;
Taylor, E. P. J. Pharm. Pharmacol. 1956, 8, 110.
21. Toshio, A.; Shozo, K.; Sachiko, K. Yakuzaigaku 1966, 26,
131.
22. Toshio, A.; Shozo, K.; Sachiko, K. Yakuzaigaku 1966, 26,
189.
23. Weiss, M. J.; Wong, J. R.; Ha, C. S.; Bleday, R.; Salem, R.
R.; Steele, G. D.; Chen, L. B. Proc. Natl. Acad. Sci. U.S.A.
1987, 84, 5444.
28. Galanakis, D.; Calder, J. A. D.; Ganellin, C. R.; Owen, C.
S.; Dunn, P. M. J. Med. Chem. 1995, 38, 3536.
29. Galanakis, D.; Davis, C. A.; Ganellin, C. R.; Dunn, P. M.
J. Med. Chem. 1996, 39, 359.
30. Galanakis, D.; Ganellin, C. R.; Malik, S.; Dunn, P. M.
J. Med. Chem. 1996, 39, 3592.
31. Cook, N. S. Trends Pharmacol. Sci. 1988, 9, 21.
32. Robertson, D. W.; Steinberg, M. I. J. Med. Chem. 1990,
33, 1529.
33. Makler, M. T. PCT Int. Appl. PIXXD2 WO 9004644 A1
19900503, 1990.
34. The procedure was a modification of that described by
Backeberg, O. G.; Marais, J. L. C. J. Chem. Soc.
(London) 1942, 381.
35. Use of a methyl group to quaternize the head groups is a
common practice in computational studies for this type of
dimers.
24. Zhuo, S.; Allison, W. S. Biochem. Biophys. Res. Commun.
1988, 152, 968.
25. Bodden, W. L.; Palayoor, S. T.; Hait, W. N. Biochem.
Biophys. Res. Commun. 1986, 135, 574.
26. Galanakis, D.; Davis, C. A.; Herrero, B. D. R.; Ganellin,
C. R.; Dunn, P. M.; Jenkinson, D. H. Bioorg. Med. Chem.
Lett. 1995, 5, 559.
27. Galanakis, D.; Davis, C. A.; Herrero, B. D. R.; Ganellin,
C. R.; Dunn, P. M. J. Med. Chem. 1995, 38, 595.
36. A feature of MO computations of Mulliken charges on N1
is that they come out negative even for N atoms with a
formal plus charge. Electron-donating groups then make
the Mulliken charges even more negative.
37. Compounds 13 and 14 were purchased from Aldrich.