LUMO energy of model compounds of bispyridinium compounds as an index for the inhibition of choline kinase

Article abstract of DOI:10.1016/S0223-5234(01)01219-3

Eleven derivatives of 1,1′-[1,2-ethylenebis(benzene-1,4-diylmethylene)]bis(4-pyridinium) dibromides bearing various groups at C-4 of the pyridinium moiety were synthesized and examined for their inhibition of choline kinase (ChoK) and antiproliferative activities. The C-4 substituents include electron-releasing, neutral or electron-withdrawing groups. A one-parameter regression equation has been derived which satisfactorily describes the ex vivo inhibitory potency of ChoK of the title compounds. The electronic effect plays a critical function in the ex vivo inhibition of ChoK although the role of electrostatic interactions could be altered due to a solvation process of both ChoK and ligands.

Full text of DOI:10.1016/S0223-5234(01)01219-3

Eur. J. Med. Chem. 36 (2001) 215–225
´
© 2001 Editions scientifiques et me´dicales Elsevier SAS. All rights reserved
PII: S0223-5234(01)01219-3/FLA
Original Article
LUMO energy of model compounds of bispyridinium compounds as an index
for the inhibition of choline kinase
Joaqu´ın Camposa, Mar´ıa del Carmen Nu´n˜eza, Vicente Rodr´ıgueza, Antonio Entrenaa,
Rube´n Herna´ndez-Alcocebab, Fe´lix Ferna´ndezb, Juan Carlos Lacalb, Miguel A. Galloa,
Antonio Espinosaa*
aDepartamento de Qu´ımica Orga´nica, Facultad de Farmacia, Campus de Cartuja s/n, E-18071 Granada, Spain
bInstituto de Investigaciones Biome´dicas, Consejo Superior de Investigaciones Cient´ıficas, c/ Arturo Duperier,
E-28029 Madrid, Spain
Received 2 October 2000; revised 2 December 2000; accepted 4 January 2001
Abstract – Eleven derivatives of 1,1%-[1,2-ethylenebis(benzene-1,4-diylmethylene)]bis(4-pyridinium) dibromides bearing various
groups at C-4 of the pyridinium moiety were synthesized and examined for their inhibition of choline kinase (ChoK) and
antiproliferative activities. The C-4 substituents include electron-releasing, neutral or electron-withdrawing groups. A one-parameter
regression equation has been derived which satisfactorily describes the ex vivo inhibitory potency of ChoK of the title compounds.
The electronic effect plays a critical function in the ex vivo inhibition of ChoK although the role of electrostatic interactions could
´
be altered due to a solvation process of both ChoK and ligands. © 2001 Editions scientifiques et me´dicales Elsevier SAS
antiproliferative agents / electrostatic effects / choline kinase / quantitative structure–activity relationship
1. Introduction
mal cells. In drug resistance, the use of combination
chemotherapy, which is the administration of several
drugs with different and complementary mechanisms
of action, is regarded as the more effective approach.
Therefore, the object of our current research is to
overcome the shortcomings of present cancer
chemotherapy with an antitumour drug with a new
During the past 20 years, a variety of approaches
has been made for cancer chemotherapy, and many
antitumour drugs have been developed for clinical
use. In the treatment of solid tumours, however, the
conventional approaches have met with only limited
success, and cancer still remains as one of the leading
mechanism of action, capable of discriminating tu-
causes of human mortality. Current chemotherapeutic
mour cells from normal proliferative cells and exhibit-
antitumour drugs suffer two major drawbacks: ad-
ing selective toxicity against cancer.
verse effects and drug resistance. Adverse effects asso-
ciated with conventional antitumour drugs are usually
As a new mechanism of action, we focused our
attention on choline kinase (ChoK, EC 2.7.1.32). This
caused by their indiscriminate cytotoxic effect on nor-
enzyme is the first one in the cytidine 5%-diphosphate-
choline pathway for the synthesis of phosphatidyl-
Abbreviations: ChoK, choline kinase; PCho, phosphorylcholine;
choline, and phosphorylates choline to phos-
phorylcholine (PCho) using adenosine 5%-triphos-
phate as the phosphate donor [1–3]. The ras oncoge-
nes are mutated in about 25% of human tumours [4].
These genes encode 21 kDa proteins called p21 or
Ras. In vitro studies of oncogenic Ras proteins, and
products and growth factors have shown that
MO, molecular orbital; AM1, Austin Model 1; QSAR, quantitative
structure–activity relationship; HOMO, highest occupied molecular
orbital; LUMO, lowest unoccupied molecular orbital; HF, Hartree–
Fock; IPCM, isodensity surface polarized continuum model; IC50
,
concentration resulting in 50% inhibition; PI3K, phosphatidylinositol
3%-kinase; PLD, phospholipase D; PI-PLC, phosphatidylinositol–
phospholipase C; MAPK, mitogen-activated protein kinase.
* Corresponding author
E-mail address: aespinos@ugr.es (A. Espinosa).

216
J. Campos et al. / European Journal of Medicinal Chemistry 36 (2001) 215–225
phosphocholine contributes to cellular growth regula-
tion and intracellular signal transduction. Ras
proteins play a pivotal role in cellular signal transduc-
tion, and help regulate cellular proliferation and ter-
minal differentiation [5–7]. NMR studies demon
strated that the levels of PCho were increased in a
large number of tumour cells from mice and human
beings when compared to their normal counterparts
[8]. These events together with our results on symmet-
rical bisquaternary derivatives with increased in-
hibitory activity towards ChoK [9], and very recently,
the in vivo antitumour activity of several ChoK in-
hibitors [10], provide strong support to the hypothesis
of the important role of phosphatidylcholine metabo-
lites in the regulation of normal cell growth and that
its alteration is important for the establishment of
cancer cells.
pyridines and bis-p-(bromomethyl)bibenzyl [12] in bu-
tanone at 100°C in a sealed tube (figure 1). After
filtration the products were recrystallized from EtOH/
Et2O, except
6
(MeOH/Et2O). 4-(N,N-Diallyl-
amino)pyridine [13] and 4-piperidinopyridine [14]
were prepared as reported previously. 4-Perhy-
droazepinopyridine (89%) is a new compound and
was obtained from 4-chloropyridine and the corre-
sponding secondary amine according to the method
used for 4-(N,N-diallylamino)pyridine [13]. All the
new compounds gave satisfactory 1H-NMR, 13C-
NMR, and FABHRMS.
Although at first glance the NMR and mass spectra
of compounds 1, 2, 4 and 6–11 looked clean, there
were problems with their elemental analyses. Despite
numerous recrystallizations from ethanol/diethyl
ether, elemental analyses of 1 gave C and N values
which were about 3.4 and 0.7% too low, indicating
some water content. Nevertheless, H values were
+0.2% of the theoretical values. The C, N values did
not improve significantly when drying of the samples
was carried out in high vacuum at 40°C instead of
room temperature. Obviously, water can be trapped
in 1. These phenomena were also observed for com-
pounds 2, 4 and 6–11. By adding the necessary
amounts of water the elemental analyses concord
perfectly (see Section 6). The physicochemical proper-
ties of compounds 1–11 are shown in table I. Later
on, it will be discussed that water (in both aspects, as
a reactive and as a solvent) plays a pivotal role in the
biological behaviour of these bisquaternary
compounds.
2. Chemistry
In this paper we have synthesized eleven symmetri-
cal bispyridinium derivatives 1–11 (figure 1). The C-4
substituents of the pyridinium moiety include elec-
tron-withdrawing (strong ones like –CꢀN and
–COOH), neutral (–H, –CH3, –CH2OH) or electron-
releasing groups (the –NH2 is a strong one, the
–NMe2 group is the most electron-releasing neutral
substituent known [11], and the diallylamino, pyrro-
lidino, piperidino and perhydroazepino groups).
The compounds for this study were synthesized by
reaction between the corresponding 4-substituted
Figure 1. Synthesis of bispyridinium derivatives.

J. Campos et al. / European Journal of Medicinal Chemistry 36 (2001) 215–225
217
Table I. Physicochemical properties for the symmetrical bispyridinium compounds 1–11.
a
Compound
R4
Formula
Mp (°C)
291–293
Crystal solvent b
umax (nm)
‡
Yield c (%)
1
2
ꢁNMe2
ꢁNH2
C30H36N4Br2·2H2O
C26H28N4Br2·0.2H2O 295–296
C28H30N2O2Br2
C28H30N2Br2·H2O
C28H26N2O4Br2
C28H24N4Br2·0.3H2O 255–257
C38H44N4Br2·H2O
C34H40N4Br2·2H2O
a
a
a
a
b
a
a
a
a
a
a
293
272
254
254
263
278
293
293
302
302
257
35 100
5230
4940
8470
8340
8980
5030
2640
3580
2590
8850
48
83
63
75
59
84
37
80
70
94
84
3
ꢁCH2OH
ꢁCH3
\310
\310
\310
4
5
ꢁCOOH
ꢁCꢀN
6
d
7
ꢁN(Allyl)2
e
8
ꢁNC4H8
307–309
f
9
ꢁNC5H10
C36H44N4Br2·2.4H2O 236–238
C38H48N4Br2·1.8H2O \310
C26H26N2Br2·1.5H2O 109–110
g
10
11
ꢁNC6H12
ꢁH
a Satisfactory microanalyses obtained: C, H, N values are within 90.4% of the theoretical values.
b Crystal solvent: a, ethyl alcohol/diethyl ether; b, methanol/diethyl ether.
c Yield related to last step.
d Very hygroscopic to determine its mp.
e Pyrrolidino.
f Piperidino.
g Perhydroazepino.
3. Pharmacology
that could be treated by using Hammett-type sub-
stituent constants. The conventional electronic
parameters are available for only a relatively small
number of substituents (e.g. |P and |R values for the
diallylamino, pyrrolidino, piperidino and perhy-
droazepino groups are not available). Quantum-
chemical descriptors have long been used in QSAR
studies [15]. A quantum chemical treatment of elec-
tronic effects is potentially more powerful than the
Hammett-type approach since it allows greater flexi-
bility in the construction of the data set. Thus, MO
calculations were made. To simplify the computa-
tional problem, the calculations were made on model
compounds consisting of only one of the two substi-
tuted pyridinium moieties and in which the spacer
between the two charged nitrogen atoms was replaced
by a methyl group. There is inherent error associated
with the assumptions required to facilitate the calcula-
tions. In using quantum chemistry-based descriptors
with a series of related compounds, the computational
error is considered to be approximately constant
throughout the series; thus, relative values of calcu-
lated descriptors can be meaningful even though their
absolute values are not directly applicable [16]. The
low-energy conformations obtained served as initial
geometries for ab initio quantum chemical calcula-
Compounds 1–11 were tested in an ex vivo system
using purified ChoK from yeast (in a test tube with-
out cells) as a target. This assay allowed us to evalu-
ate the effect on 1–11 activities without considering
the possible effects on other properties such as perme-
ability into intact cells, specific cellular environments,
putative intracellular modifications of the synthesized
compounds or enzyme compartmentalization. The ef-
fects on cell proliferation by the ChoK inhibitors in
ras-transformed cells were next investigated. The Hill
equation was fitted to the data to obtain estimates of
the IC50.
4. Results and discussion
The electronic influence of R4 on 11 alters its
charge distribution (data not shown) and molecular
orbital (MO) energies. During the past three decades
most of the QSARs derived for biological systems
have relied on the use of Hammett-type | constants
to account for electronic variation associated with
changes in molecular structure. As a result, these
studies have often been limited to sets of congeners

218
J. Campos et al. / European Journal of Medicinal Chemistry 36 (2001) 215–225
As can be seen from Eq. (1) there is a good correla-
tion between p(IC50)ex vivo and the partial charge on
the ring N atom of the pyridinium fragment:
tions. The structures of the model compounds are
given in table II. Emphasis was placed on both accu-
racy and computational feasibility. While the highest
accuracy is always desirable, computational accessi-
bility is necessary. The ab initio (6-31G* level) calcu-
lations indicated deficiency of a negative charge on
hydrogens atoms and excess charge on carbon and
particularly on nitrogen atoms. The charge is, there-
fore, distributed according to the electronegativity of
atoms. These results do not conform with our tradi-
tional beliefs, particularly in the case of quarternized
species, which habitually assume that deficiency of a
charge occurs on the nitrogen atom to which the
chain is attached. According to the theoretical predic-
tions the carbon and nitrogen atoms form a nega-
tively charged core surrounded by a positively
charged layer holding hydrogen atoms. Similar fea-
tures have been predicted for several protonated ni-
trogen organic bases [17]. The partial charges on the
ring N atom of the pyridinium moiety are shown in
table II.
p(IC50)ex vivo
= −3.82 ( 1.74)−13.4 ( 2.81) (N1 charge)
(1)
n=10, r= −0.860, s=0.220, F1,8=22.8, P<0.005
where pIC50 = −log IC50, bearing in mind that the
higher the value of pIC50 the more potent is the
compound, n is the number of compounds (com-
pounds 1–5 and 7–11 are included), r is the correla-
tion coefficient, s is the standard deviation, the
numbers in brackets are standard errors of estimate
and F is the F ratio between the variances of observed
and calculated activities.
Eq. (1) shows that the more negative the partial
charge on N1, the more potent is the compound. The
influence of the variation of N1 charge on the in-
hibitory potency of ChoK of the analogues is in
agreement with the 13C chemical shifts (in CD3OD) of
the methylene group bearing the endocyclic pyri-
Table II. Partial charges, frontier orbital and calculated solvation-free energies for model compounds used for correlation Eqs.
(1)–(6)
a
b
b
c
d
d
Compound
R4
N1 charge
EHOMO
ELUMO
DGsolv (kcal/mol)
p(IC50)ex vivo
p(IC50)antiprol
1a
2a
ꢁNMe2
ꢁNH2
−0.640
−0.639
−0.596
−0.598
−0.583
−0.588
−0.639
−0.642
−0.642
−0.641
−0.586
−296.7
−315.4
−338.4
−341.0
−346.4
−357.7
−285.1
−293.2
−290.0
−290.8
−344.6
−30.62
−36.77
−55.03
−56.41
−78.69
−88.26
−27.92
−28.61
−26.17
−27.36
−63.50
−45.01
−54.84
−54.72
−49.80
−64.99
−62.90
−46.21
−43.05
−44.73
−43.61
−53.01
4.77
4.64
4.00
4.00
5.70
5.40
e
3a
ꢁCH2OH
ꢁCH3
4a
4.70
f
5a
ꢁCOOH
ꢁCꢀN
3.86
f
6a
3.70
6.26
6.00
6.40
6.40
4.22
7a
ꢁN(Allyl)2
4.77
4.70
5.02
4.82
4.51
g
8a
ꢁNC4H8
h
9a
ꢁNC5H10
ꢁNC6H12
ꢁH
i
10a
11a
a Model compound numbers correspond to the compound numbers of table I.
b In kcal/mol.
c Solvation-free energies were calculated with the HF/6-31G* method using the IPCM model implemented in GAUSSIAN94 [24]. The
solvent used in the ab initio calculations was water at 298 K, with dielectric constant: 78.36.
d All values are the means of two independent determinations performed in duplicate; pIC50=−log IC50; the pIC50 of the
corresponding bispyridinium compound of table I is given for simplicity.
e It cannot be accurately calculated because IC50\100 mM.
f It cannot be accurately calculated because IC50\1000 mM.
g Pyrrolidino.
h Piperidino.
i Perhydroazepino.

J. Campos et al. / European Journal of Medicinal Chemistry 36 (2001) 215–225
219
dinium nitrogen atom of the symmetrical bisquater-
nary molecules 1–5 and 7–10 [18]. The more elec-
tron-releasing R4 is, the more electron density is fed
into the ring and the more positive charge is with-
drawn from it; accordingly, the electron-donor groups
shift the methylene resonance upfield. As a result, the
negative charge on N1 is increased, and so is potency.
Obviously, the opposite holds for electron-withdraw-
ing groups. Eq. (1) suggests that substituents that
tend to donate electrons to the pyridinium ring make
more potent inhibitors.
In any thorough investigation of substituent effects,
it is essential to prove that one’s conclusions are both
statistically valid and make chemical sense. Neither
approach is sufficient alone [19]. It is appropriate to
attempt to obtain insight into the physical meaning of
the above-mentioned correlations 1, 2 and 3. This
may suggest which of the three electronic parameters
(N1 charge, ELUMO and EHOMO, which are clearly
themselves colinear, as shown explicity by Eq. (5) is
more consistent to use.
The physical meaning of the correlation with N1
charge was not clear at this stage. Nevertheless, the
involvement of electronic factors suggests the occur-
rence of either charge transfer or dipolar interactions.
The transfer of a pair of electrons from the HOMO to
the LUMO is, by definition, a reaction between a
Lewis acid and a Lewis base. The compound that
reacts via its HOMO is functioning as an electron-
pair donor, that is, a Lewis base or a nucleophile. A
nucleophile, or Lewis base, participates in reactions
via its HOMO. The compound that reacts via its
LUMO is functioning as an electron-pair acceptor,
that is, a Lewis acid or an electrophile. An elec-
trophile, or Lewis acid, participates in reactions via
the LUMO.
In principle, the previous reasoning seems contra-
dictory: a simple electrostatic interaction between the
quaternized nitrogen atom and an anionic group in
the active site of ChoK should be increased with
electron-withdrawing groups at position 4 of the sym-
metrical bispyridinium compounds. Such atomic
charges are suitable for characterizing interactions
according to classical point-charge electrostatic
model. The electrostatic interaction may involve
matching parts of the electric fields of the two inter-
acting species, which have opposite signs, rather than
matching point charges. For a better understanding
of the experimental results, the energies of the
HOMO (highest occupied molecular orbital, a mea-
sure of the ease of electron loss) and LUMO (lowest
unoccupied molecular orbital, a measure of the ease
of electron gain) of the compounds were calculated.
The energies of the frontier orbitals for the model
compounds are also shown in table II. Normally, the
spectroscopic transition energies are expressed in elec-
tron volts (eV) while the relative energies are ex-
pressed in kcal mol−1. However, the frontier orbital
energies (HOMO and LUMO) will be expressed in
kcal mol−1 as Eq. (4) (see later) correlates ELUMO and
solvation-free energy (Gsolv). The inhibitory potency
of ChoK of the analogues correlates well with the
energy of the lowest unoccupied molecular orbital
(ELUMO) (Eq. (2)) and with the energy of the highest
occupied molecular orbital (EHOMO) (Eq. (3)):
With the exception of the hydrogen cation, H+,
which has no electrons and cannot have a HOMO, all
molecules and ions possess both a HOMO and a
LUMO. Therefore there are always, in principle, two
possible HOMO–LUMO combinations for reactions
in which H+ is not a participant. Fortunately, one of
these combinations is usually obviously right and the
other is usually obviously wrong.
It seems that in the EHOMO and ELUMO correlations,
the higher the energies of the HOMO or the LUMO
are (i.e. get closer to zero), the more potent the
compound. If HOMO is considered to contribute at
the level of interaction of the compound with ChoK,
it follows that the compound acts as an electron
donor, with the enzyme acting as an ‘electron sink’.
Fundamental chemical concepts are in discordance
with this because these compounds are electron defi-
cient and could not act as electron donors. It is
therefore difficult to attribute a chemical meaning to
the EHOMO correlation.
p(IC50)ex vivo =5.33 ( 0.17)+0.02 ( 0.00) ELUMO
(2)
n=10, r=0.878, s=0.207, F1,8 =26.83, P<0.001
p(IC50)ex vivo =8.83 ( 0.85)+0.01 ( 0.00) EHOMO
4.1. Solvation of the model compounds
(3)
However, if LUMO is assumed to contribute at the
level of interaction of the compound with ChoK, it is
n=10, r=0.874, s=0.210, F1,8 =25.81, P<0.001

220
J. Campos et al. / European Journal of Medicinal Chemistry 36 (2001) 215–225
n=10, r=0.874, s=9.66, F1,8=26.02, P<0.001
LUMO= −20.7 ( 2.30)−30.3 ( 2.71) (N1 charge)
necessary to consider the possibility of the formation of
a charge transfer interaction with the latter, with the
consequent interaction of the HOMO of ChoK
(HOMOChoK) with the LUMO of the compound
(LUMOcompd). With the solvent (H2O) molecules acting
as electron donors, the bispyridinium compounds could
act as electron acceptors. The increase of the ELUMO of
the molecule would result in its weaker solvation, and
thus, a stronger interaction with ChoK. A possible
chemical meaning for the ELUMO correlation is there-
fore offered by the hypothesis that a higher value for
ELUMO indicates a weaker solvation of the compound.
A fact that strongly supports the solvation hypothesis
has already been used by Galanakis et al. to explain
the K+ channel blocking activity of dequalinium ana-
logues [20]. Thus net ion–ion interactions are greatly
reduced on hydration and only the consequent libera-
tion of solvent molecules may create a significant en-
tropy effect and favourable interaction [21]. Very
recently, Miller et al. [22] have suggested that water
can significantly enhance the electronic coupling be-
tween a donor and an acceptor in aqueous electron
transfer processes.
E
(5)
n=10, r=0.970, s=0.211, F1,8=125.26, P<0.001
4.2. Relationship between antiproliferative activity and
ELUMO
Finally, as the main goal of this research is to search
for new anticancer agents, the correlation between the
antiproliferative activity and the ELUMO has been calcu-
lated and the result is excellent (Eq. (6)):
p(IC50)antiprol
=7.32 ( 0.21)+4.42×10−2 ( 4.39×10−3) ELUMO
(6)
n=9, r=0.967, s=0.270, F1,7=101.85, P<0.001
4.3. Relationship between ChoK inhibition and
The Hartree–Fock (HF) and the isodensity surface
polarized continuum model (IPCM) [23] implemented
in the GAUSSIAN94 [24] package were used to calcu-
late absolute aqueous solvation-free energies (kcal
mol−1) because it has been shown that HF/IPCM gave
the best balance between computational accessibility
and accuracy between experimental and calculated val-
ues for pyridinium ions [25]. Both the active site of
ChoK and the compounds would solvate but the dif-
ferent interactions between both species would depend
on the nature of the R4 group of the pyridinium model
compounds, since the solvation of ChoK is always the
same. As can be seen from table II, the model com-
pounds with higher ELUMO (7a, 8a, 9a and 10a) show a
decreased free solvation energy, whereas those with
lower ELUMO have a higher Gsolv (5a and 6a). The
unsubstituted compound (R4=H) represents the inter-
mediate case with an ELUMO of −63.50 kcal mol−1 and
an Gsolv of −53.01 kcal mol−1. There is a fairly good
correlation between the ELUMO and the free solvation
energies of the model compounds (Eq. (4)), whereas
there is a much better correlation between ELUMO and
the partial charge on N1 of the pyridinium moiety (Eq.
(5)):
antiproliferative activity
All compounds (with the exception of 11a) are more
potent against the HT-29 cell line (in vitro assay) than
against ChoK (ex vivo assay). The level of the enzyme
inhibition needed to perform a proliferation inhibition
is not known. In other words, there is not a direct
mathematical relationship between p(IC50)antiprol and
p(IC50)ex vivo, the reason for this being that the inhibi-
tion level for the PCho production necessary to induce
the proliferation inhibition is not known. What is
known is that all the ChoK inhibitors produce this
inhibitor effect on proliferation. Nevertheless, taken to-
gether, the data presented here could be interpreted by
means of a different unknown mechanism. With regard
to the specificity, the parameters so far analyzed are
negative except for ChoK. The title compounds are
specific for ChoK since no inhibitory effect was ob-
served against the phosphorylation of endogenous sub-
strates such as phosphatidylinositol 3%-kinase (PI3K),
two of the most relevant enzymes involved in the regu-
lation of phospholipid metabolism such as phospholi-
pase D (PLD) and phosphatidylinositol–phospholipase
C (PI–PLC) or the Raf/mitogen-activated protein ki-
nase (MAPK) pathway [26]. Moreover, compounds 1–
11 showed no intercalating activity on DNA.
ELUMO=74.9 ( 23.4)+2.36 ( 0.46) Gsolv
(4)

J. Campos et al. / European Journal of Medicinal Chemistry 36 (2001) 215–225
221
5. Conclusions
Bruker AM-300 spectrometers and chemical shifts are
1
reported relative to the residual H and 13C signals of
The quantum-chemically derived descriptors have
definite chemical meaning, and they have thus proved
to be especially useful in the clarification of the inter-
molecular interaction mechanism determining the
molecular property or chemical process under study.
The fitting of the three-dimensional structure and the
complementary surface properties of a drug to its
binding site are conditions for its biological activity.
Even in simple ex vivo systems, the surrounding water
molecules compete to solvate the binding site and to
the functional groups of the ligand, and this balance
can modify affinity. LUMOs seem to be essential for
both inhibition of ChoK and for antiproliferative
activity. In spite of being considered rather small, the
variation of substituents presented here have been
proved to be of general applicability for further exten-
sive structure–activity relationship studies within the
bisquaternary heterocycles as ChoK inhibitors. The
regressor ELUMO under study is proving to be useful
for the design principles of even more potent in-
hibitors of ChoK, the results of which will be pre-
sented elsewhere.
the deuterated solvent (CHCl3, l 7.24 and 77.1 ppm;
CD2HOD, l 3.31 and 49.9 ppm). Chemical shifts (l)
quoted in the case of multiplets were measured from the
approximate centre. Signals are designated as follows: s,
singlet; d, doublet; dd, doublet of doublets; ddt, double
double triplet; pst, pseudotriplet; psq, pseudoquartet; m,
multiplet. Coupling constants are expressed in hertz.
High resolution liquid secondary ion mass spectra (HR
LSIMS) were carried out on a VG AutoSpec Q high
resolution mass spectrometer (Fisons Instruments). The
compounds gave accurate mass spectra, having no ex-
traneous peaks. Analyses indicated by the symbols of
the elements or functions were within 0.4% of the
theoretical values. All compounds were dried at 40–
50°C and 0.1 mmHg for 16 h, but many tenaciously held
on to water which appears to be a solvate.
6.1.1. General procedure for compounds 1–11
4-Substituted pyridine (1.58 mmol), bis-p-(bro-
momethyl)bibenzyl [12] (0.79 mmol), and butanone (50
mL) were heated at 100°C in a sealed tube. After
filtration and thorough washing with butanone and
CHCl3, the solid product was purified by crystallization
from ethanol (or methanol in the case of 6) by adding
diethyl ether to turbidity. The m.p.s, crystalization sol-
vent, UV spectral data (the umax and the extinction
coefficient, o) and yields of compounds 1–11 are pre-
sented in table I.
6. Experimental
6.1. Chemistry
All solvents were used dried and freshly distilled. All
evaporations were carried out in vacuo with a rotary
evaporator. Solutions were dried over MgSO4 before
concentration under reduced pressure. Analytical thin-
layer chromatography (TLC) was done on Merck sil-
icagel F-254 plates with detection with iodine, an UV
lamp or by charring with dilute sulfuric acid, using
mixtures of CH2Cl2:MeOH (10:0.1) as developing sol-
vent. All analytical samples were TLC homogeneous.
For normal column chromatography Merck silica gel 60
was used with a particle size 0.063–0.200 mm (70–230
mesh ASTM). For flash chromatography Merck silica
gel 60 was used with a particle size 0.040–0.063 mm
(230–400 mesh ASTM). Melting points (mp) were ob-
tained on an Electrothermal melting point apparatus
and are uncorrected. Ultraviolet–visible (UV–vis) ab-
sorption spectra were recorded on a Spectronic Genesys
5 spectrophotometer. NMR spectra were recorded on a
400.1 MHz 1H- and 100.03 MHz 13C-NMR Bruker
6.1.1.1. 1,1%-[1,2-Ethylenebis(benzene-1,4-
diylmethylene)]bis[(4-dimethylamino)pyridinium]
dibromide (1)
1H-NMR (300.13 MHz, CD3OD): l 8.22 (d, J2,3
=
7.9, 4H, H-2), 7.29 (d, J=8.3, 4H, H-3benz and H-5benz),
7.24 (d, J=8.3, 4H, H-2benz and H-6benz), 7.00 (d,
J
2,3=7.9, 4H, H-3), 5.33 (s, 4H, CH2N+), 3.25 (s, 12H,
CH3), 2.91 (s, 4H, CH2). 13C-NMR (75.57 MHz,
CD3OD): l 158.00 (C-4), 144.13 (C-1benz), 143.06 (C-2),
133.78 (C-4benz), 130.56 (C-2benz and C-6benz), 129.46
(C-3benz and C-5benz), 109.11 (C-3), 61.54 (CH2N+),
40.39 (CH3), 38.30 (CH2). HR LSIMS (thioglycerol)
Calc. m/z for C30H36N4Br [M−Br]+ 531.2123. Found
m/z: 531.2130. Anal. C30H36N4Br2·2H2O (C, H, N).
6.1.1.2. 1,1%-[1,2-Ethylenebis(benzene-1,4-
diylmethylene)]bis(4-aminopyridinium) dibromide (2)
1H-NMR (300.13 MHz, CD3OD): l 8.15 (d, J2,3
=
1
ARX 400, a 300.13 MHz H- and 75.78 MHz 13C-NMR
7.6, 4H, H-2), 7.28 (d, J=8.4, 4H, H-3benz and H-5benz),

222
J. Campos et al. / European Journal of Medicinal Chemistry 36 (2001) 215–225
7.23 (d, J=8.4, 4H, H-2benz and H-6benz), 6.86 (d,
m/z for C28H25N2O4 [M−BrH−Br]+ 453.1814. Found
m/z: 453.1817. Anal. C28H26N2O4Br2 (C, H, N).
J
2,3=7.9, 4H, H-3), 5.29 (s, 4H, CH2N+), 2.91 (s, 4H,
CH2). 13C-NMR (75.57 MHz, CD3OD): l 160.82 (C-4),
144.12 (C-1benz), 144.00 (C-2), 133.74 (C-4benz), 130.59
(C-2benz and C-6benz), 129.41 (C-3benz and C-5benz),
110.95 (C-3), 61.80 (CH2N+), 38.29 (CH2). HR LSIMS
(thioglycerol) Calc. m/z for C26H27N4 [M−BrH−Br]+
6.1.1.6. 1,1%-[1,2-Ethylenebis(benzene-1,4-
diylmethylene)]bis(4-cyanopyridinium) dibromide (6)
1H-NMR (400.13 MHz, CD3OD): l 9.32 (d, J2,3
=
6.7, 4H, H-2), 8.52 (d, J2,3=6.7, 4H, H-3), 7.46 (d,
J=8.1, H-3benz and H-5benz), 7.33 (d, J=8.1, 4H, H-
2benz and H-6benz), 5.91 (s, 4H, CH2N+), 2.45 (s, 4H,
CH2). 13C-NMR (100.62 MHz, CD3OD): l 147.33 (C-
2), 145.40 (C-1benz), 132.49 (C-3), 131.30 (C-4benz),
130.99 (C-2benz and C-6benz), 130.73 (C-3benz and C-
5benz), 115.16 (CN), 66.49 (CH2N+), 38.15 (CH2). HR
LSIMS (thioglycerol+H+) Calc. m/z for C28H25N4 [M−
2Br+H]+ 417.2079. Found m/z: 417.2075. Anal.
C28H24N4Br2·0.3H2O (C, H, N).
395.2236.
Found
m/z:
395.2234.
Anal.
C26H28N4Br2·0.2H2O (C, H, N).
6.1.1.3. 1,1%-[1,2-Ethylenebis(benzene-1,4-
diylmethylene)]bis[(4-hydroxymethyl)pyridinium]
dibromide (3)
1H-NMR (300.13 MHz, CD3OD): l 8.96 (d, J2,3
=
6.9, 4H, H-2), 8.07 (d, J2,3=6.9, 4H, H-3), 7.41 (d,
J=8.2, H-3benz and H-5benz), 7.27 (d, J=8.2, 4H, H-
2benz and H-6benz), 5.78 (s, 4H, CH2N+), 4.92 (s, 4H,
CH2OH), 2.93 (s, 4H, CH2). 13C-NMR (75.57 MHz,
CD3OD): l 165.21 (C-4), 145.16 (C-2), 144.79 (C-1benz),
132.45 (C-4benz), 130.82 (C-2benz and C-6benz), 130.15
(C-3benz and C-5benz), 126.05 (C-3), 64.79 (CH2N+),
62.86 (CH2OH), 38.17 (CH2). HR LSIMS (thioglycerol)
Calc. m/z for C28H30N2O2Br [M-Br]+ 505.1491. Found
m/z: 505.1493. Anal. C28H30N2O2Br2 (C, H, N).
6.1.1.7. 1,1%-[1,2-Ethylenebis(benzene-1,4-
diylmethylene)]bis[(4-diallylamino)pyridinium] dibromide
(7)
1H-NMR (400.13 MHz, CD3OD): l 8.26 (d, J2,3
=
7.9, 4H, H-2), 7.31 (d, J=8.2, 4H, H-3benz and H-5benz),
7.25 (d, J=8.2, 4H, H-2benz and H-6benz), 7.03 (d,
J2,3=7.9, 4H, H-3), 5.90 (ddt, J=17.2, 10.4 and 4.9,
4H, H-2allyl), 5.35 (s, 4H, CH2N+), 5.28 (ddt, J=10.4,
1.2 and 1.1, 4H, H-3allyl), 5.21 (ddt, J=17.2, 1.2 and
1.1, 4H, H-3%allyl), 4.24 (dt, J=4.9 and 1.2, 8H, H-1allyl),
2.91 (s, 4H, CH2). 13C-NMR (100.62 MHz, CD3OD): l
157.89 (C-4), 144.25 (C-1benz), 143.50 (C-2), 133.56 (C-
4benz), 131.49 (C-2allyl), 130.59 (C-2benz and C-6benz),
129.57 (C-3benz and C-5benz),118.38 (C-3allyl), 109.92 (C-
3), 61.78 (CH2N+), 54.24 (C-1allyl), 38.28 (CH2). HR
LSIMS (thioglycerol) Calc. m/z for C38H44N4Br [M−
Br]+ 635.2749. Found m/z: 635.2747. Anal.
C38H44N4Br2·H2O (C, H, N).
6.1.1.4. 1,1%-[1,2-Ethylenebis(benzene-1,4-
diylmethylene)]bis(4-methylpyridinium) dibromide (4)
1H-NMR (300.13 MHz, CD3OD): l 8.88 (d, J2,3
=
6.6, 4H, H-2), 7.94 (d, J2,3=6.6, 4H, H-3), 7.41 (d,
J=8.2, H-3benz and H-5benz), 7.27 (d, J=8.2, 4H, H-
2benz and H-6benz), 5.75 (s, 4H, CH2N+), 2.93 (s, 4H,
CH2), 2.67 (s, 6H, CH3). 13C-NMR (75.57 MHz,
CD3OD): l 161.75 (C-4), 144.79 (C-2), 132.50 (C-4benz),
130.81 (C-2benz and C-6benz), 130.15 (C-3benz and C-
5benz), 130.10 (C-3), 64.63 (CH2N+), 38.21 (CH2), 22.03
(CH3). HR LSIMS (thioglycerol) Calc. m/z for
C28H30N2Br [M−Br]+ 473.1592. Found m/z: 473.1578.
Anal. C28H30N2Br2·H2O (C, H, N).
6.1.1.8. 1,1%-[1,2-Ethylenebis(benzene-1,4-
diylmethylene)]bis(4-pyrrolidino)pyridinium) dibromide
(8)
1H-NMR (400.13 MHz, CD3OD): l 8.20 (d, J2,3
=
6.1.1.5. 1,1%-[1,2-Ethylenebis(benzene-1,4-
7.6, 4H, H-2), 7.29 (d, J=8.1, 4H, H-3benz and H-5benz),
7.23 (d, J=8.1, 4H, H-2benz and H-6benz), 6.86 (d,
diylmethylene)]bis(4-carboxypyridinium) dibromide (5)
1H-NMR (400.13 MHz, CD3OD): l 9.24 (d, J2,3
=
J
2,3=7.6, 4H, H-3), 5.32 (s, 4H, CH2N+), 3.55 (pst,
6.6, 4H, H-2), 8.54 (d, J2,3=6.6, 4H, H-3), 7.47 (d,
J=8.1, H-3benz and H-5benz), 7.31 (d, J=8.1, 4H, H-
2benz and H-6benz), 5.91 (s, 4H, CH2N+), 2.94 (s, 4H,
CH2). 13C-NMR (100.62 MHz, CD3OD): l 147.11 (C-
2), 145.09 (C-1benz), 131.85 (C-4benz), 130.93 (C-2benz and
C-6benz), 130.52 (C-3benz and C-5benz), 129.10 (C-3), 65.83
(CH2N+), 38.15 (CH2). HR LSIMS (thioglycerol) Calc.
J=6.8, 8H, H-2pyrrolid and H-5pyrrolid), 2.90 (s, 4H,
CH2), 2.11 (psq, J=6.8, 8H, H-3pyrrolid and H-4pyrrolid).
13C-NMR (100.62 MHz, CD3OD): l 155.14 (C-4),
144.08 (C-1benz), 142.99 (C-2), 133.92 (C-4benz), 130.55
(C-2benz and C-6benz), 129.43 (C-3benz and C-5benz),
109.70 (C-3), 61.55 (CH2N+), 49.74 (C-2pyrrolid and C-
5pyrrolid), 38.33 (CH2), 26.14 (C-3pyrrolid and C-4pyrrolid).

J. Campos et al. / European Journal of Medicinal Chemistry 36 (2001) 215–225
223
HR LSIMS (thioglycerol) Calc. m/z for C34H40N4Br
[M−Br]+ 583.2436. Found m/z: 583.2438. Anal.
C34H40N4Br2·2H2O (C, H, N).
C-5benz), 129.72 (C-3), 65.52 (CH2N+), 38.19 (CH2). HR
LSIMS (thioglycerol) Calc. m/z for C26H26N2Br [M−
Br]+ 445.1279. Found m/z: 445.1275. Anal.
C26H26N2Br2·1.5H2O (C, H, N).
6.1.1.9. 1,1%-[1,2-Ethylenebis(benzene-1,4-
6.1.2. 4-Chloropyridine
diylmethylene)]bis(4-piperidino)pyridinium) dibromide (9)
1H-NMR (300.13 MHz, CD3OD): l 8.17 (d, J2,3
=
4-Chloropyridine hydrochloride (10 g) was dissolved
in water (20 mL), and excess 6 N NaOH was added to
the solution to reach a pH of 12. The 4-chloropyridine
formed an organic layer above the aqueous phase. The
mixture was extracted with diethyl ether (4×50 mL).
The combined ether extracts were dried (MgSO4),
filtered and concentrated in vacuo, and the remaining
dark brown liquid was kept at high vacuum for 20 min
(6.42 g, 86%).
7.9, 4H, H-2), 7.32 (d, J=8.2, 4H, H-3benz and H-5benz),
7.26 (d, J=8.2, 4H, H-2benz and H-6benz), 7.12 (d,
J
2,3=7.9, 4H, H-3), 5.30 (s, 4H, CH2N+), 3.70 (t, J=
6.8, 8H, H-2piperid and H-6piperid), 2.91 (s, 4H, CH2), 1.75
(m, 12H, H-3piperid, H-4piperid and H-5piperid). 13C-NMR
(75.57 MHz, CD3OD): 156.85 (C-4), 144.13 (C-1benz),
143.49 (C-2), 133.76 (C-4benz), 130.56 (C-2benz and C-
6benz), 129.47 (C-3benz and C-5benz), 109.27 (C-3), 61.38
(CH2N+), 49.10 (C-2piperid and C-6piperid), 38.32 (CH2),
26.67 (C-3piperid and C-5piperid), 24.93 (C-4piperid). HR
LSIMS (thioglycerol) Calc. m/z for C36H44N4Br [M−
Br]+ 611.2749. Found m/z: 611.2773. Anal.
C36H44N4Br2·2.4H2O (C, H, N).
6.1.3. General procedure for the synthesis of
4-(diallylamino)- and 4-(cycloamino)pyridines
4-Chloropyridine (38 mmol) was refluxed with excess
of the amine (95 mmol) under an argon atmosphere for
3 days. The mixture was neutralized with 1 N NaOH,
extracted with diethyl ether (4×100 mL), dried
(MgSO4), filtered and rotaevaporated off. This residue
was purified by flash chromatography using the mixture
Cl3CH:MeOH, 9:1 as eluant, giving the title compound.
6.1.1.10. 1,1%-[1,2-Ethylenebis(benzene-1,4-
diylmethylene)]bis(4-perhydroazepino)pyridinium)
dibromide (10)
1H-NMR (300.13 MHz, CD3OD): l 8.20 (d, J2,3
=
7.9, 4H, H-2), 7.31 (d, J=8.2, 4H, H-3benz and H-5benz),
7.25 (d, J=8.2, 4H, H-2benz and H-6benz), 7.05 (d,
J
6.1.3.1. 4-(Diallylamino)pyridine
2,3=7.9, 4H, H-3), 5.32 (s, 4H, CH2N+), 3.72 (t, J=
Following the general procedure (Section 6.1.3) and
using diallylamine, 4-(diallylamino)pyridine was pre-
pared in 34.5% yield as a dark brown viscous liquid;
TLC (Cl3CH:MeOH, 8.5:1.5) Rf: 0.43. 1H-NMR (300.13
MHz, CD3OD): l 8.09 (d, J2,3=7.3, 2H, H-2pyr), 6.84
6.8, 8H, H-2perhhydoazep and H-7perhydroazep), 2.91 (s, 4H,
CH2), 1.85 (m, 8H, H-3perhydroazep and H-6perhydroazep),
1.60 (m, 8H, H-4perhydroazep and H-5perhydroazep). 13C-
NMR (75.57 MHz, CD3OD): l 157.19 (C-4), 144.17
(C-1benz), 143.33 (C-2), 133.75 (C-4benz), 130.57 (C-2benz
and C-6benz), 129.51 (C-3benz and C-5benz), 109.01 (C-3),
61.48 (CH2N+), 51.33 (C-2perhydroazep and C-7perhydroazep),
38.33 (CH2), 27.34 (C-3perhydroazep, C-4perhydroazep, C-
5perhydroazep and C-6perhydroazep). HR LSIMS (thioglyc-
erol) Calc. m/z for C38H48N4Br [M−Br]+ 639.3062.
Found m/z: 639.3050. Anal. C38H48N4Br2·1.8H2O (C,
H, N).
(d, J2,3=7.3, 2H, H-3pyr), 5.89 (ddt, J2,3%=17.2, J2,3
=
10.1, J1,2=4.7, 2H, H-2allyl), 5.24 (ddt, J2,3=10.1,
J3,3%=1.5, J1,3=1.7, 2H, H-3allyl), 5.18 (ddt, J2,3%=17.2,
J3,3%=1.5, J1,3%=1.7, 2H, H-3%allyl), 4.13 (dt, J1,2=4.7,
J
1,3=1.7, J1,3%=1.7, 4H, H-1allyl). 13C-NMR (75.57
MHz, CD3OD): l 157.32 (C-4pyr), 144.56 (C-2pyr),
132.49 (C-2allyl), 117.53 (C-3allyl), 108.69 (C-3pyr), 53.57
(C-1allyl). Anal. C11H14N2 (C, H, N).
6.1.1.11. 1,1%-[1,2-Ethylenebis(benzene-1,4-
diylmethylene)]bispyridiniumdibromide (11)
6.1.3.2. 4-Piperidinopyridine
Yield: 95%; m.p. 78–79°C (lit.14 m.p.: 80°C); TLC
1
1H-NMR (300.13 MHz, CD3OD): l 9.08 (d, J2,3
=
(Cl3CH:MeOH, 1:1) Rf: 0.41. H-NMR (300.13 MHz,
6.6, 4H, H-2), 8.62 (t, 2H, H-4), 8.14 (pst, 4H, H-3),
7.43 (d, J=8.2, 4H, H-3benz and H-5benz), 7.29 (d,
J=8.2, 4H, H-2benz and H-6benz), 5.83 (s, 4H, CH2N+),
2.94 (s, 4H, CH2). 13C-NMR (75.57 MHz, CD3OD): l
147.27 (C-2), 145.90 (C-4), 144.94 (C-1benz), 132.27 (C-
4benz), 130.87 (C-2benz and C-6benz), 130.28 (C-3benz and
CDCl3): l 8.17 (d, J2,3=6.6, 2H, H-2pyr), 6.63 (d, J2,3=
6.6, 2H, H-3pyr), 3.33 (t, 4H, H-2piperid and H-6piperid),
1.63 (m, 6H, H-3piperid and 4piperid). 13C-NMR (75.57
MHz, CDCl3): l 155.11 (C-4pyr), 148.88 (C-2pyr), 108.10
(C-3pyr), 47.25 (C-2piperid and C-6piperid), 25.13 (C-3piperid
and C-5piperid), 24.28 (C-4piperid).

224
J. Campos et al. / European Journal of Medicinal Chemistry 36 (2001) 215–225
6.2. Pharmacology
6.1.3.3. 4-(Perhydroazepino)pyridine
1
Yield: 89%; TLC (Cl3CH:MeOH, 1:1) Rf: 0.43. H-
NMR (300.13 MHz, CDCl3): l 8.12 (d, J2,3=6,7, 2H,
H-2pyr), 6.44 (d, J2,3=6.7, 2H, H-3pyr), 3.41 (t, 4H,
H-2perhydroazep and H-7perhydroazep), 1.73 (m, 4H, H-
3perhydroazep and H-6perhydroazep), 1.49 (m, 4H, H-
4perhydroazep and H-5perhydroazep). 13C-NMR (75.57 MHz,
CDCl3): l 153.23 (C-4pyr), 149.39 (C-2pyr), 106.16 (C-
3pyr), 48.61 (C-2perhydroazep and C-7perhydroazep), 27.03 (C-
3perhydroazep and C-6perhydroazep), 26.76 (C-4perhydroazep and
C-5perhydroazep). Anal. C11H16N2 (C, H, N).
The ex vivo ChoK inhibition and antiproliferative
assays against HT-29 cells were followed in accordance
with the protocols reported previously [9, 10].
Acknowledgements
We thank the Spanish CICYT (project SAF98-0112-
C02-01) for financial support. The award of a grant
from the Junta de Andaluc´ıa to M.C.N. and Dr. Ali
1
Haidour for recording the H- and 13C-NMR spectra
are gratefully acknowledged. We also want to thank one
of the referees for the helpful comments given.
6.1.4. Computational methods
Molecular modelling and preliminary geometry opti-
mizations were performed using SYBYL software pack-
age. Then the structures were reoptimized using
semiempirical quantum chemical calculations with the
semiempirical AM1 (Austin Model 1) method [27], pre-
vious to their treatment by the ab initio methods. The
AM1 method [27] was used because indicated that the
ring N atom carries a negative partial charge (data not
shown). Gas-phase geometry optimizations were per-
formed via ab initio quantum mechanical methods. Ab
initio calculations were performed using GAUSSIAN-94
[24] with the 6-31G* basis set. Geometry optimizations
were performed at the Hartree–Fock (HF) level of
theory.
Free energies of solvation in aqueous solution were
calculated using the isodensity surface polarized contin-
uum model (IPCM) [23]. The IPCM is implemented in
the GAUSSIAN-94 [24] package. All solvation calcula-
tions were performed on the gas-phase optimized struc-
tures. Due to the pyridinium ring, the presence of
solvent is not expected to have a significant influence on
the optimized geometries, as was verified in several tests
(not described). For ab initio reaction-field calculations
the dielectric of water was set to 78.36 and the isoden-
sity cutoff values were 0.001 a.u. Solvation-free energies
in aqueous solution were obtained as the difference
between the HF/6-31G* gas-phase energies and the cor-
responding energies of the same level of theory with the
IPCM solvation model.
References
[1] Kennedy E.P., Fed. Proc. 20 (1961) 934–940.
[2] Kent C., Prog. Lipid Res. 29 (1990) 87–105.
[3] Vance D.E., Biochem. Cell Biol. 68 (1990) 1151–1165.
[4] Barbacid M., Ann. Rev. Biochem. 56 (1987) 779–827.
[5] Abdellatif M., MacLellan W.R., Schneider M.D., J. Biol.
Chem. 269 (1994) 15423–15426.
[6] Barbacid M., Eur. J. Clin. Invest. 20 (1990) 225–235.
[7] Wiesmu¨ller L., Wittinghofer F., Cell. Signal. 6 (1994) 247–267.
[8] Bhakoo K.K., Williams S.R., Florian C.L., Land H., Noble
M.D., Cancer Res. 56 (1996) 4630–4635.
[9] Herna´ndez-Alcoceba R., Saniger L., Campos J., Nu´n˜ez M.C.,
Khaless F., Gallo M.A., Espinosa A., Lacal J.C., Oncogene 15
(1997) 2289–2301.
[10] Herna´ndez-Alcoceba R., Ferna´ndez F., Lacal J.C., Cancer Res.
59 (1999) 3112–3118.
[11] Hansch C., Leo A., Substituent Constants for Correlation
Analysis in Chemistry and Biology, Wiley, New York, 1979,
pp. 1–8.
[12] Cram D.J., Steinberg H., J. Am. Chem. Soc. 73 (1951) 5691–
5704.
[13] Vaidya R.A., Mathias L., J. Am. Chem. Soc. 108 (1986)
5514–5520.
[14] Jerchel D., Jakob L., Chem. Ber. 91 (1958) 1266–1273.
[15] Karelson M., Lobanov V.S., Katritzky A.R., Chem. Rev. 96
(1996) 1027–1043.
[16] Redl G., Cramer R.D. III, Berkoff C.E., Chem. Soc. Rev. 3
(1974) 273–292.
[17] Lubkowski J., Blazejoswski J., J. Phys. Chem. 95 (1991) 2311–
2316.
Statistical analysis of the relationships between ChoK
inhibitory or antiproliferative activities and substituent
properties of interest was performed by partial least-
squares algorithm using the QSAR module of SYBYL
software [28].
[18] Campos J., Nu´n˜ez M.C., Rodr´ıguez V., Gallo M.A., Espinosa
A., Bioorg. Med. Chem. Lett. 10 (2000) 767–770.
[19] Reynolds W.F., Gomes A., Maron A., MacIntyre D.W., Tanin
A., Can. J. Chem. 61 (1983) 2376–2384.
[20] Galanakis D., Calder J.A.D., Ganellin C.R., Owen C.S., Dunn
P.M., J. Med. Chem. 38 (1995) 3536–3546.

J. Campos et al. / European Journal of Medicinal Chemistry 36 (2001) 215–225
225
[21] Davies R.H., Timms D., in: Smith H.J. (Ed.), Introduction to
the Principles of Drug Design and Action, Harwood Academic
Publishers, Amsterdam, 1998, pp. 51–96.
Wong, J.L. Andres, E.S. Replogle, R. Gomperts, R.L. Martin,
D.J. Fox, J.S. Binkley, D.J. Defrees, J. Baker, J.P. Stewart, M.
Head-Gordon, C. Gonzalez, J.A. Pople, Gaussian Inc., Pitts-
burgh, PA, 1995.
[22] Miller N.E., Wander M.C., Cave R.J., J. Phys. Chem. A 103
(1999) 1084–1093.
[25] Chen I.-J., MacKerell A.D. Jr, Theor. Chem. Acc. 103 (1999)
483–494.
[23] Foresman J.B., Keith T.A., Wiberg K.B., Snoonian J., Frisch
M.J., J. Phys. Chem. 100 (1996) 16098–16104.
[26] To know the molecular mechanisms of signal transduction, see:
C.-H. Heldin, M. Purton (Eds.), Signal Transduction, Chap-
man and Hall, London, 1996.
[24] GAUSSIAN-94; Revision C.2, M.J. Frisch, G.W. Trucks, H.B.
Schlegel, P.M.W. Gill, B.G. Johnson, M.A. Robb, J.R. Cheese-
man, T. Keith, G.A. Petersson, J.A. Montgomery, K.
Raghavachari, M.A. Al-Laham, V.G. Zakrzewski, J.V. Ortiz,
J.B. Foresman, J. Cioslowski, B.B. Stefanov, A. Nanayakkara,
M. Challacombe, C.Y. Peng, P.Y. Ayala, W. Chen, M.W.
[27] Dewar M.J.S., Zoebish E.G., Healy E.F., Stewart J.J.P., J. Am.
Chem. Soc. 107 (1985) 3902–3909.
[28] SYBYL, version 6.5, Tripos Associates, St. Louis, MO, 1998.
.