Design of more potent squalene synthase inhibitors with multiple activities

Article abstract of DOI:10.1016/j.bmc.2010.09.008

With the increasing realization that modulating a multiplicity of targets can be an asset in the treatment of multifactorial disorders, we hereby report the synthesis and evaluation of the first compounds in which antioxidant, anti-inflammatory as well as squalene synthase (SQS) inhibitory activities are combined by design, in a series of simple molecules, extending their potential range of activities against the multifactorial disease of atherosclerosis. The activity of the initially synthesized antihyperlipidemic morpholine derivatives (1-6), in which we combined several pharmacophore moieties, was evaluated in vitro (antioxidant, inhibition of SQS and lipoxygenase) and in vivo (anti-dyslipidemic and anti-inflammatory effect). We further compared the in vitro SQS inhibitory action of these derivatives with theoretically derived molecular interactions by performing an in silico docking study using the X-ray crystal structure of human SQS. Based on low energy preferred binding modes, we designed potentially more potent SQS ligands. We proceeded with synthesizing and evaluating these new structures (7-12) in vitro and in vivo, to show that the new derivatives were significantly more active than formerly developed congeners, both as SQS inhibitors (20-70-fold increase in activity) and antioxidants (4-30-fold increase in activity). A significant correlation between experimental activity [Log(1/IC₅₀)] and the corresponding binding free energy (ΔG_b) of the docked compounds was shown. These results, taken together, show a promising alternative and novel approach for the design and development of multifunctional antiatherosclerosis agents.

Full text of DOI:10.1016/j.bmc.2010.09.008

Bioorganic & Medicinal Chemistry 18 (2010) 7402–7412
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Design of more potent squalene synthase inhibitors with multiple activities
⇑
Angeliki P. Kourounakis , Alexios N. Matralis, Anastasios Nikitakis
Department of Pharmaceutical Chemistry, School of Pharmacy, University of Athens, 15771 Athens, Greece
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 7 July 2010
Revised 27 August 2010
Accepted 2 September 2010
Available online 7 September 2010
With the increasing realization that modulating a multiplicity of targets can be an asset in the treatment
of multifactorial disorders, we hereby report the synthesis and evaluation of the first compounds in
which antioxidant, anti-inflammatory as well as squalene synthase (SQS) inhibitory activities are com-
bined by design, in a series of simple molecules, extending their potential range of activities against
the multifactorial disease of atherosclerosis. The activity of the initially synthesized antihyperlipidemic
morpholine derivatives (1–6), in which we combined several pharmacophore moieties, was evaluated
in vitro (antioxidant, inhibition of SQS and lipoxygenase) and in vivo (anti-dyslipidemic and anti-inflam-
matory effect). We further compared the in vitro SQS inhibitory action of these derivatives with theoret-
ically derived molecular interactions by performing an in silico docking study using the X-ray crystal
structure of human SQS. Based on low energy preferred binding modes, we designed potentially more
potent SQS ligands. We proceeded with synthesizing and evaluating these new structures (7–12)
in vitro and in vivo, to show that the new derivatives were significantly more active than formerly devel-
oped congeners, both as SQS inhibitors (20–70-fold increase in activity) and antioxidants (4–30-fold
increase in activity). A significant correlation between experimental activity [Log(1/IC50)] and the corre-
Keywords:
Aromatic tetrahydro-1,4-oxazines
Synthesis
Lipid peroxidation
Triglycerides
Total cholesterol
LDL-cholesterol
Squalene synthase inhibitors
Molecular modeling
Docking
sponding binding free energy (DGb) of the docked compounds was shown. These results, taken together,
show a promising alternative and novel approach for the design and development of multifunctional
antiatherosclerosis agents.
Drug design
Autodock
Squalene synthase
Lipoxygenase
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction
Current management of atherosclerosis focuses on reducing li-
pid blood levels mainly by interfering with cholesterol biosynthe-
Atherosclerosis, the underlying cause of coronary artery disease
that is the leading cause of death in most parts of the World,1 is a
multifactorial disease. Several mechanisms participate in its initial
manifestation and progression, mainly involving the triad: hyper-
lipidemia, oxidative stress, and inflammation.2 It is generally
accepted that elevated levels of cholesterol, that is, mainly LDL-
cholesterol, provide a higher substrate concentration for lipid
peroxidation by reactive oxygen species. The oxidatively altered
LDL is taken up in large amounts by macrophages which are trans-
formed into foam cells, which are deposited in the vasculature and
excrete several proinflammatory factors. Thus, lesion formation in
the vascular tissue progresses into atheromatic plaques and the
subsequent complications of the disease.
sis. To date, HMG-CoA reductase inhibitors (i.e., statins) are the
most effective class of therapeutics in this respect. However, by
inhibiting HMG-CoA, statins suppress the production of mevalo-
nate which is not only a precursor of cholesterol but also of non-
sterol molecules which have a vital role in diverse cellular func-
tions. Thus, the inhibition of their synthesis has been associated
with statins’ major side-effects, hepatotoxicity, and myotoxicity.
Squalene synthase, a key enzyme in the cholesterol biosynthetic
pathway, occupies the first and solely committed step towards the
biosynthesis of the sterol nucleus of cholesterol. Because of its un-
ique and strategic location in this pathway, it is an attractive target
for inhibition and the development of novel and improved antihy-
percholesterolemic agents. These agents are considered not to
interfere with the biosynthesis of other biologically important mol-
ecules and thus may offer a better side-effect profile than the
extensively used inhibitors of (upstream) HMG-CoA reductase.
Several squalene synthase inhibitors have been studied3,4
including zaragozic acids, quinuclidine as well as 4,1-benzoxaze-
pine-3-acetic acid derivatives, such as CP-320473 and TAK-475.
TAK-475 (or else ‘Lapaquistat’) reached clinical phase III trials but
was eventually discontinued based on safety concerns at higher
doses.5 Using a very different pharmacophore structure, we have
Abbreviations: SQS, squalene synthase; HMG-CoA, 3-hydroxy-3-methylglutaryl-
CoA; FPP, farnesyl pyrophosphate; BSA, bovine serum albumin; NADPH, nicotin-
amide adenine dinucleotide phosphate; IC50, inhibitory concentration (for 50% of
the reaction); TC, total cholesterol; LDL-C, low-density lipoprotein cholesterol; TG,
triglyceride.
⇑
Corresponding author. Tel.: +30 210 7274818; fax: +30 210 7274747.
E-mail address: angeliki@pharm.uoa.gr (A.P. Kourounakis).
0968-0896/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2010.09.008

A. P. Kourounakis et al. / Bioorg. Med. Chem. 18 (2010) 7402–7412
7403
O
O
HOOC
O
N
OH
OH
R1
R2
O
O
N
HOOC
N
H
N
Cl
R3
A
CP-320473
X
III
R1
OH
O
OH
R1
R2
O
R2
N
N
CH3
R3
O
R3
4-6
1-3
O
OH
OH
R1
O
R2
N
R3
CH3
O
R
7-12
Naproxen
Chart 1.
previously synthesized and investigated 2-biphenyl morpholine
derivatives (Chart 1, general structure A) that were shown to re-
duce cholesterol and triglyceride levels, acting as squalene syn-
thase inhibitors, as well as to combine antioxidant activity.6–8
The mode of action of two such derivatives was further investi-
gated both in vitro9,10 and in vivo.11
derivatives were evaluated and proven more potent inhibitors of
SQS.
2. Results and discussion
2.1. Chemistry
Oxidative stress and inflammation are major players in athero-
sclerosis development and subsequent cardiovascular pathology.12
Since the potential of natural antioxidants (i.e., vitamins E and C) in
cardiovascular risk reduction is controversial with several clinical
studies showing a beneficial13 and others no effect,14 new antiox-
idant strategies are needed to address the oxidative stress menace
in such conditions. Further, the same holds for anti-inflammatory
treatment, since the overall effect and safety of various known
NSAIDs (including naproxen) on cardiovascular risk outcome is
also not conclusive and may vary.15,16 However, there is an increas-
ing realization that modulating simultaneously a multiplicity of
targets17 may be an asset in the treatment of multifactorial disor-
ders such as atherosclerosis. We hereby designed and synthesised
new aromatic morpholine derivatives aiming to incorporate in one
structure not only increased antioxidant, SQS inhibition and anti-
dyslipidemic but also anti-inflammatory activity. Thus, we incor-
porated a condensed aromatic system (Chart 1, compounds 1–6)
in order (i) to increase the conjugation effect for radical stabiliza-
tion for antioxidant activity and (ii) to provide some structural
similarity with a known non-steroidal anti-inflammatory drug
namely naproxen. We investigated the in vitro and in vivo pharma-
cological activity of the synthesized derivatives and performed in
silico docking studies with SQS. The latter suggested limited lipo-
philic interactions with the inhibitor binding site of SQS and thus
new derivatives (compounds 7–12) with extended aromatic sub-
stituents were designed and synthesised. The newly synthesized
The target compounds (1–12, Chart 1) were synthesised in
mostly good yields and via the spontaneous cyclization of the cor-
responding hydroxyaminoketone intermediate to
a hemiketal
structure (Scheme 1). Structures and calculated18 lipophilicity val-
ues (c Log P) are presented in Table 1. As verified by our theoretical
studies hereby, as well as by previous experimental studies19,20
stable conformations of the final compounds include equatorial
position of the aromatic substituent/axial position of the OH, and
chair–chair (trans) conformation of the fused octahydro-oxazine
ring systems.
2.2. Antioxidant activity
The effect of the investigated derivatives on the non enzymatic
peroxidation of hepatic microsomal membrane lipids after 45 min
of incubation, expressed as IC50 values, is shown in Table 2. Under
the same experimental conditions, 2,6-di-tert-butyl-4-methylphe-
nol (BHT), a known potent antioxidant, and probucol exhibited
IC50 values of 25 lM and >1 mM, respectively. The time course of li-
pid peroxidation, as affected by several concentrations of one of the
most active compounds, 9, is depicted in Figure 1. The initially de-
signed and synthesised derivatives (1–6) did not demonstrate stron-
ger antioxidant activity than that of the reference compound III
(Chart 1 and Table 2). However, the performed structural modifica-
tions, that is, the extension of the conjugated system (compounds 7–

7404
A. P. Kourounakis et al. / Bioorg. Med. Chem. 18 (2010) 7402–7412
O
Br
O
OH
OH
(a)
+
B
H
3C
CH3
R
I
R
IIa R=OMe
II
13 R=OMe
14
b R=H
R=H
OH
O
OH
Ar
O
N
NH
Ar
CH3
CH3
13, 14, 20-21
HBr
(c)
CH3
OH
(b)
1, 4, 7, 10
H
OH
Ar
N
O
O
N
HBr
Ar
(c)
(c)
Br
15-18
2, 5, 8, 11
OH
O
Ar
HBr
OH
N
(d)
O
CH3
3, 6, 9, 12
NH
19
CH3
Scheme 1. General method for the synthesis of compounds 1–12. Reagents and conditions: (a) Pd(OAc)2, Bu4NBr, K2CO3, H2O, Ar, 70 °C, 1.5–2 h; (b) Br2, CHCl3, rt, 2.5 h for 15
or phenyltrimethylammonium tribromide, THF/H2O, rt, 1.5 h for 16–18; (c) (i) acetone, rt, 20–24 h; (ii) HBr/Et2O; (d) CH3NH2, rt, 19 h.
12) led consistently to a considerable (4–30-fold) improvement of
their antioxidant profile. As anticipated, and in accordance to previ-
ous observations,8 antioxidant activity seems to be favoured by con-
jugation or resonance effects of the further extended aromatic
substituent (compare compounds 1–6 with 7–12), with an addi-
tional contribution to this by the positive resonance (+R) effect of
the OMe substituent (i.e., compare 7–9 with 10–12). Further, lipo-
philicity, determined as c Log P values, in this series of derivatives
seems also to correlate (r2 = 0.69, P = 0.0009) with antioxidant activ-
ity, inagreementwiththenotionthatthelipophiliccharacterof anti-
oxidants contributes to the offered inhibition of lipid peroxidation,
as it may facilitate access to biological membranes.
biaryl substitution while a directional requirement for the rigid biar-
yl side chain was identified. It is hypothesized that the aromatic sub-
stituent may act as an isoster for the isoprenyl subunits of the
farnesyl chain of the endogenous enzyme substrate farnesyl pyro-
phosphate (FPP).
2.4. Anti-dyslipidemic activity
The in vivo anti-dyslipidemic activity of these derivatives is also
demonstrated in Table 2. Experimental hyperlipidemia was suc-
cessfully established 24 h after Triton WR 1339 administration,
with an increase in plasma total cholesterol, LDL-cholesterol, and
triglyceride levels of 93%, 77%, and 84%, respectively, compared
to normal values.8 All test compounds were found able to reduce
the examined parameters in the plasma of hyperlipidemic rats
(ca. 20–80% reduction), with activities somewhat comparable to
those of reference compound III, with the exception of derivative
8 which had no effect on triglycerides and 7 with no effect on
LDL-cholesterol. Under the same experimental conditions, probu-
col and simvastatin (used as anti-dyslipidemic drugs), at the same
dose, reduced plasma total cholesterol by 18% and 75%, LDL-C by
18% and 70%, and triglycerides by 11% and 0%, respectively.8
It seems that compounds 7–12 that incorporate an extended
lipophilic substituent retained the in vivo effect of lowering lipi-
demic parameters. As expected, other parameters such as in vivo
distribution and pharmacokinetics may affect the in vivo order of
activity of these compounds as compared to their in vitro activities.
2.3. Inhibition of rat microsomal squalene synthase
Inhibition of the activity of squalene synthase, from rat liver
microsomes, by the test compounds, expressed as IC50 values, is
shown in Table 2. Most compounds inhibited squalene synthase
activity significantly and dose-dependently as shown in Figure 2
which depicts the activity of the most active compound 12. It was
first observed that the activity of the initially designed derivatives
1–6 was relatively low with IC50 values ranging from 15–155 lM.
Results from our docking studies (as described below in Section
2.7) led to the development of derivatives 7–12 with ‘elongated’
hydrophobic substituents. These derivatives, with the exception of
8, exhibited ca. 20–70-fold higher inhibitory activity. The overall in-
crease in lipophilicity seems to favor a better interaction between
the molecule and enzyme, and a mild correlation between c Log
P and IC50 values was observed (r2 = 0.63, P = 0.004). This is in accor-
dance with our previous studies,8 as well as studies on a series of
3-hydroxy-3-biaryl quinuclidines21, that demonstrated an improved
squalene synthase inhibitory activity for compounds with planar
2.5. Lipoxygenase inhibition
Lipoxygenases are important inflammatory mediators impli-
cated in acute inflammatory reaction processes. Numerous studies

A. P. Kourounakis et al. / Bioorg. Med. Chem. 18 (2010) 7402–7412
7405
Table 1
100
Structures of substituted hydroxymorpholins, and c Log P values
OH
O
R1
R2
75
50
25
0
Ar
HBr
N
R3
Compd
R1
H
R2
H
R3
Ar
c Log P
1
CH3
2.17
OCH3
OCH3
OCH3
2
3
H
(CH2)4
3.09
3.60
-3
-2
-1
0
1
2
3
Log (µM)
(CH2)4
CH3
CH3
Figure 2. Representative graph showing the activity of squalene synthase as
affected by various concentrations of compound 12.
4
5
6
H
H
2.37
3.17
3.80
H
(CH2)4
manner to that of the rat mast cell lipoxygenase and is often used
as a reliable means for evaluating lipoxygenase inhibitors.24
Among all compounds, the methoxy-naphthyl derivatives 1–3
inhibited lipoxygenase activity in a concentration and time depen-
dent manner (Fig. 3, Table 2). Their IC50 values, after 6 min of incu-
bation, were 20, 53, and 84
of the other compounds was much lower (IC50 values above
300 V) (Table 2). The three active compounds 1–3 have incorpo-
(CH2)4
CH3
CH3
OCH3
OCH3
OCH3
7
8
H
H
4.06
4.97
5.49
4.25
5.05
5.69
lV, respectively, whereas the potency
H
(CH2)4
l
rated in their structure a large structural moiety of naproxen
(Chart 1), which is very likely to contribute to the above action. Na-
proxen’s activity compared to other anti-inflammatory agents is
particularly high, inhibiting soybean lipoxygenase with an IC50 va-
9
(CH2)4
H
CH3
CH3
10
11
12
H
lue of 24 l
V.24 In contrast, no significant effect against lipoxyge-
H
(CH2)4
nase activity was noted for the respective naphthyl derivatives
(4–6) or derivatives 7–12, confirming not only the positive contri-
bution of the methoxy group in the activity of 1–3, but also the lack
of the rest of the potential ‘isosteric’ moiety (Chart 1) found in the
‘naproxen resembling’ derivatives 1–3.
(CH2)4
CH3
2.6. In vivo anti-inflammatory activity
The anti-inflammatory activity of most derivatives (1–3, 5, 6, 8–
11) was evaluated by the method of carrageenan-induced paw
edema which is a non specific inflammation maintained by the re-
lease of histamine and serotonin and later by prostaglandins.25,26
The examined compounds were administrated ip at a dose of
1.5
1.0
0.5
0.0
Control
0.035 mM
0.050 mM
300 lmol/Kg right after the injection of carrageenan.
The effect of compounds 1–3, 5, 6, 8–11 and naproxen on paw
edema is depicted in Figure 4. All compounds demonstrated signif-
icant anti-inflammatory activity comparable to that of naproxen
(51%) at the same dose,16 with the most active being compounds
0.100 mM
0.250 mM
0
10
20
30
40
50
1.5
Time (min)
Control
Figure 1. Representative graph of the time course of lipid peroxidation as affected
by various concentrations of compound 9.
1.0
4 µ
M
µ
M
19
75
have suggested a role for this enzyme in atherosclerosis, since cho-
lesterol-loaded macrophages contain increased levels of 12-lipoxy-
genase, while both mRNA and protein for 15-lipoxygenase are
present in atherosclerotic lesions.22 Thus, lipoxygenase inhibitors
have been proposed as antiatherosclerotic agents.23 Because of
the built-in structural similarity of our compounds with naproxen,
an active lipoxygenase inhibitor, we investigated the effect of our
derivatives on soybean lipoxygenase. Soybean lipoxygenase is
inhibited by non-steroidal anti-inflammatory drugs in a similar
µ
M
0.5
0.0
µ
M
150
0
100
200
Time (sec)
300
400
Figure 3. Effect of various concentrations of compound 1 on lipoxygenase activity.

7406
A. P. Kourounakis et al. / Bioorg. Med. Chem. 18 (2010) 7402–7412
100
90
80
70
60
50
40
30
20
10
0
Control
1
2
3
5
6
10
11
naproxen
8
9
Figure 4. Anti-inflammatory effect of compounds 1–3, 5, 6, 8–11 and naproxen (nap) on carrageenan-induced paw edema in rats. Results are expressed as percent inhibition
of edema compared to the control group. Statistical significance (Student’s t-test) from control: P <0.005.
1, 3, and 6 (58%, 53%, and 54% edema inhibition, respectively). The
incorporated structural similarity of compounds 1–3 with naprox-
en, may contribute to their increased anti-inflammatory activity in
vivo, while the naphthyl derivatives 5 and 6 also exhibited good
anti-inflammatory activity, as expected, since several other naph-
thyl derivatives with anti-inflammatory activity have been re-
ported in the literature.27 Surprisingly, the activity of the
‘elongated’ derivatives 8–11 was more or less retained as far as
in vivo anti-inflammatory action is concerned and comparable to
compounds 1–6.
overlaps with a similar one in the structural superimposition of
SQS and FPS (farnesyl pyrophosphate synthase) that belong to
the same class I of isoprenoid biosynthetic enzymes.29,31 This se-
quence is considered to be the binding site for the phosphate
groups, via Mg2+, of the substrate farnesyl pyrophosphate (FPP).
Visual inspection of the lowest binding free energies of SQS
complexes with our initially synthesized morpholine derivatives
(1–6) revealed a hydrogen bond between the 2-hydroxyl group
of the substituted morpholine ring of all compounds studied in this
series and the side chain carboxyl group of ASP80, possibly sug-
gesting that this hydrogen bond is significant for their inhibitory
action (Fig. 5a).
2.7. Molecular docking with squalene synthase (SQS)
Other interactions suggested by the docking results were the
hydrophobic interactions of the bulky hydrophobic groups of com-
pounds 1–6 as they were found oriented towards the hydrophobic
cavity under the ‘flap’ in close proximity to the tert-butyl group of
the co-crystallized ligand CP-320473 suggesting similar hydropho-
bic interactions. Hydrophobic residues such as VAL175, ALA176,
VAL 179, LEU 211, and LEU183, forming the inner cavity of the ac-
tive site, surround the hydrophobic side chains of the compounds.
Realizing that hydrophobic interactions of our morpholine
derivatives (1–6) were limited at least when compared to those
of the bound inhibitor (compare the position of the naphthyl
groups of the respective molecules, Fig. 5a) we sought to reach dee-
per in the hydrophobic cavity by elongating the hydrophobic moi-
ety of our morpholine derivative. This was accomplished by
extending this naphthyl group with the addition of an extra phenyl
ring, as seen in the designed compounds 7–12.
To gain new insights in the interactions between the initially
synthesized morpholine derivatives (1–6) and squalene synthase
and to design theoretically more potent SQS inhibitors, we per-
formed rigid docking studies using the available X-ray crystal
structure of SQS with a bound inhibitor. This theoretical study
was performed using the software Autodock 4.28
In order to test Autodock’s accuracy a docking run was per-
formed on the initial co-crystallized ligand CP-32047329 (structure
in Chart 1) under the same conditions and parameters. This pro-
duced a very good prediction of the actual original binding mode
of the ligand with a reference difference of 0.95 Å root mean square
deviation (RMSD).
The final binding modes of the morpholine derivatives, after the
respective docking runs, were selected on the basis of the highest
score (best fit), corresponding to the complex with the lowest free
energy of binding (
process, the lowest conformations of our morpholine derivatives
were used (derived as described in Section 3).
D
Gb) measured in kcal/mol. For the docking
It seemed from the docking studies of the designed new deriv-
atives that not only the previous lipophilic interactions (previously
exhibited by the naphthyl moiety of compounds 1–6) were main-
tained (via the additional phenyl ring of the new derivatives) but
now also further hydrophobic interactions deeper in the hydropho-
bic cavity could be achieved with the re-positioning of the naph-
thyl moiety ( Fig. 5b). So these derivatives seemed to have a
better binding capacity as also estimated by their relatively lower
binding free energy values.
In specific, the naphthyl and tert-butyl part of CP-320473 were
found in the same area as the naphthyl and phenyl groups, respec-
tively, of the best docking poses of compounds 7–12, which turned
out to be the most active compounds of the series (Table 2). This
can be exemplified by comparing the position of the naphthyl
group in compound 12 (the most active compound of this series)
The inhibitor binding site in SQS is a solvent exposed external
cavity formed at the end of a central channel running through
the protein. This cavity is partially covered from one side by a ‘flap’
formed by residues 50–54 with the side chain of Phe54 forming
one wall of a large hydrophobic cavity under the flap. The deep
hydrophobic cavity leads to an aspartate-rich opening. Both the
aspartate-rich sequences and the above mentioned residues form-
ing the ‘flap’ are among the most highly conserved in squalene syn-
thases from different species.30 The aminoacids belonging to the
aspartate-rich sequence of SQS (80DTLED84) are considered to
play an important role in the catalytic reaction mechanism. As spe-
cifically discussed in Pandit et al., this aspartate-rich sequence

A. P. Kourounakis et al. / Bioorg. Med. Chem. 18 (2010) 7402–7412
7407
Figure 5. Binding mode visualization of compound 6 (a) and 12 (b). The binding mode of CP-320473 (green) is also included for comparison.
Table 2
Antioxidant, SQS inhibitory, lipoxygenase inhibitory, and anti-dyslipidemic effects of the investigated compounds. Autodock’s binding free energy (
D
G) derived from the docking
studies on SQS
Compd
Inhibition of lipid
peroxidation IC50
Microsomal SQS
D
Gb (Kcal/mol) Lipoxygenase
% Decrease compared to hyperlipidemic controls (56 lmol/kg ip)
(
l
M) inhibition IC50
(l
V)
inhibition IC50 (lV)
TC
LDL-C
TG
1
2
3
4
5
6
7
8
9
10
11
12
900
390
220
640
440
600
30
65
48
71
155.0
18.7
40.0
135.0
130.0
15.1
6.9
18.0
1.4
1.9
À7.42
20
53
84
62b
45b
36b
51a
33b
43b
41a
19b
32a
37a
49a
43a
54a
18b
—
24c
46a
33b
36b
42b
39a
9d
51a
37b
47a
40a
41a
51a
30b
6d
À8.20
À8.75
À7.32
À7.95
À8.52
À9.69
À9.83
À10.58
À9.37
À10.39
À10.29
À8.82
—
>300
>300
>300
>100
>100
>100
>100
>100
>100
110
29b
54b
59a
33b
48a
51b
18d
—
82a
28b
34b
32b
49a
11d
—
117
118
450
n
n.d.
0.8
36.0
—
III
Probucol
Naproxen
—
n
—
—
24e
n.d. = not determined, n = above 1 mM.
a
P <0.005.
P <0.05.
P <0.1.
b
c
d
Not significant (P >0.1) (Student’s test).
e
Data from Taraporewala et al.24
and that of compound 6 (Fig. 5b). The naphthyl group of compound
6 is found to occupy approximately the same area (i.e., closer to the
surface of the cavity) as the phenyl group of compound 12
(Fig. 5a,b) and as the tert-butyl hydrophobic group of CP-320473.
In addition, however, compound 12 exhibits further hydrophobic
interactions deeper in the hydrophobic cavity with its naphthyl
moiety. Similar observations can be made when comparing bind-
ing modes of compounds 7–12 versus their ‘shorter’ counterparts
1–6. Thus, these extended hydrophobic interactions, may explain
the higher binding and inhibiting capacity of the former com-
pounds. The presence of an OCH3 group in compounds 1–3 and
7–9 did not particularly influence (i.e., increase) binding.
Comparison of the binding free energy (DGb as calculated by the
docking runs, Table 2) and the experimentally recorded inhibitory
activity [expressed as Log(1/IC50)] indicated that lower energy val-
ues (i.e., stronger theoretical binding to SQS) correspond to lower
IC50 values and thus more significant inhibition. Simple regression
calculations between Log(1/IC50) of all compounds and the corre-
sponding
P = 0.0001) (Fig. 6).
D
Gb revealed a significant correlation (R2 = 0.79, F = 37,

7408
A. P. Kourounakis et al. / Bioorg. Med. Chem. 18 (2010) 7402–7412
3.2.1. General procedure for the preparation of the final
compounds 1–12
1
0
r2 =0.79
The final products (Table 1) were obtained by the reaction of
2.2 mmol of 2-methylaminoethanol or 2-piperidinemethanol or
trans-2-methylamino-cyclohexanol with 1.0 mmol of either
2-bromoacetyl-6-methoxynaphthalene (compounds 1–3) or 2-
bromoacetylnaphthalene (compounds 4–6) or 2-bromo-1-[4-(6-
methoxy-2-naphthalene)-phenyl]-ethanone (compounds 7–9) or 2-
bromo-1-[4-(2-naphthalene)-phenyl]-ethanone (compounds 10–12),
in anhydrous acetone (40 mL) at room temperature with stirring for
20–24 h. Acetone was then distilled off, ether was added to the residue,
the mixture was washed with saturated sodium chloride solution,
dried (K2CO3), and the products were isolated as hydrobromide salts.
-1
-2
-3
-12
-11
-10
-9
-8
-7
-6
Log1/IC50
3.2.1.1. 2-(6-Methoxy-2-naphthalene)-4-methylmorpholin-2-ol
Figure 6. Correlation between the binding free energy (
compounds 1–12.
DGb) and Log1/IC50 of
hydrobromide (1). Yield 40%, mp 148.5–150 °C. IR: 3206 (
m JY),
2720 (
m
+NY). 1H NMR (DMSO-d6): d 2.72 (s, 3H, N–CH3), 3.11 (d,
J = 12.08 Hz, 1H, 3-Hax), 3.19 (d, J = 10.61 Hz, 1H, 5-Hax), 3.33 (m,
2H, 5-Heq and –OH), 3.49 (d, J = 12.07 Hz, 1H, 3-Heq), 3.80 (s, 3H,
O–CH3), 3.97 (dd, J1 = 3.29 Hz, J2 = 12.63 Hz, 1H, 6-Heq), 4.24 (dt,
J1 = 3.21 Hz, J2 = 12.53 Hz, 1H, 6-Hax), 7.13 (dd, J1 = 2.56 Hz,
J2 = 8.77 Hz, 1H, 7-H Ar), 7.28 (ds, J = 2.19 Hz, 1H, 5-H Ar), 7.56 (dd,
J1 = 1.56 Hz, J2 = 8.51 Hz, 1H, 3-H Ar), 7.80 (d, J = 8.60 Hz, 1H, 4-H
Ar), 7.83 (d, J = 8.97 Hz, 1H, 8-H, Ar), 7.95 (s, 1H, 1-H Ar), 9.87 (br s,
1H, NH). Anal. (C16H20BrNO3Á0.1H2O) C, H.
2.8. Conclusion
We designed and synthesized new morpholine derivatives (1–
6) aiming to incorporate in one structure squalene synthase inhib-
itory, antioxidant as well as anti-inflammatory activity. The new
multiple acting molecules that combined several pharmacophore
moieties showed in vivo anti-dyslipidemic and anti-inflammatory
effect as well in vitro antioxidant, lipoxygenase and SQS inhibitory
effect. In order to gain new insights in the interactions between
these molecules and squalene synthase, we performed in silico
docking studies using the X-ray crystal structure of human SQS
which is available with a co-crystallized bound inhibitor. To our
knowledge, this is the first time that an in silico docking study
has been performed on human SQS. By visualizing and evaluating
low energy docking poses, we designed and developed new deriv-
atives (7–12) with ‘elongated’ hydrophobic moieties which, as pre-
dicted, showed increased SQS inhibitory activity, as well as
antioxidant, while retaining in vivo effects. We further found a
good correlation between theoretical and experimental binding
of all compounds to SQS. Thus, application of findings from this
docking study led to the design and development of more potent
SQS inhibitors with multiple activities that are aimed against the
multifactorial disorder of atherosclerosis.
3.2.1.2. 3-(6-Methoxy-2-naphthalene)-octahydro-1,4-pyrido[2,1-c]
oxazin-3-ol hydrobromide (2). Yield 81%, mp 191.5–192.5 °C. 1H
NMR (CDCl3): d 1.62–1.89 (m, 4H, 2 Â 8-H and 2 Â 9-H), 2.23–2.43
(m, 4H, 6-Hax, –OH and 2 Â 7-H), 2.72–2.78 (m, 1H, 6-Heq), 2.94 (d,
J = 12.71 Hz, 1H, 4-Hax), 3.23–3.32 (m, 2H, 4-Heq and 9a-Hax), 3.35
(dd, J1 = 3.62 Hz, J2 = 12.12 Hz, 1Y, 1-Heq), 3.76 (s, 3H, O–CH3), 4.32
(t, J = 12.46 Hz, 1H, 1-Hax), 6.96–7.01 (m, 2H, 5,7-H Ar), 7.52 (dd,
J1 = 1.71 Hz, J2 = 8.62 Hz, 1H, 3-H Ar), 7.55–7.60 (m, 2H, 4,8-H Ar),
7.93 (s, 1H, 1-H Ar), 10.72 (br s, 1H, NH). Anal. (C19H24BrNO3) C, H.
3.2.1.3. 2-(6-Methoxy-2-naphthalene)-4-methyl-octahydro-1,4-
benzoxazin-2-ol hydrobromide (3). Yield 70%, mp 158–160 °C.
1H NMR (CDCl3): d 1.26–1.54 (m, 4H, 5-Hax, 2 Â 7-H and 8-Hax),
1.81–1.99 (m, 4H, 2 Â 6-H, 8-Heq and –OH), 2.18–2.30 (m, 1H, 5-
Heq), 2.73 (d, J = 4.70 Hz, 3H, N–CH3), 2.91–3.00 (m, 2H, 3-Hax and
4a-Hax), 3.61 (dd, J1 = 2.31 Hz, J2 = 12.55 Hz, 1H, 3-Heq), 3.86 (s, 3H,
O–CH3), 4.49 (m, 1H, 8a-Hax), 7.06–7.14 (m, 2H, 5,7-H Ar), 7.61 (dd,
J1 = 1.71 Hz, J2 = 8.62 Hz, 1H, 3-H Ar), 7.64–7.71 (m, 2H, 4,8-H Ar),
8.05 (s, 1H, 1-H Ar), 11.78 (br s, 1H, NH). Anal. (C20H26BrNO3) C, H.
3. Materials and methods
3.1. Materials
3.2.1.4. 2-(2-Naphthalene)-4-methylmorpholin-2-ol hydrobro-
mide (4). Yield 40%, mp 80–84 °C. 1H NMR (CDCl3 + DMSO-d6): d
2.27 (d, J = 4.16 Hz, 3H, N–CH3), 2.40 (d, J = 11.25 Hz, 1H, 3-Hax),
2.49–2.56 (m, 2H, 5-Hax and –OH), 2.65 (d, J = 13.20 Hz, 1H, 5-Heq),
2.98 (d, J = 11.86 Hz, 1H, 3-Heq), 3.40 (dd, J1 = 3.55 Hz,
J2 = 12.83 Hz, 1H, 6-Heq), 4.16 (dt, J1 = 3.28 Hz, J2 = 12.84 Hz, 1H, 6-
All commercially available chemicals are of the appropriate pur-
ity and purchased from standard sources. [3H] FPP (21.5 Ci/mmol),
NADPH, FPP, and BSA were purchased from Sigma–Aldrich (Ger-
many). Soybean Lipoxygenase (250 U/mL) was purchased from Sig-
ma Chemical Co. (St Louis). For the in vivo experiments, Wistar male
rats (200–250 g) were used. Animals were kept in a controlled
temperature room (22 2 °C), having free access to laboratory chow
and tap water, under a 12 h light/dark cycle.
Hax), 6.80–6.87 (m, 2H, 6,7-H Ar), 7.04 (dd, J1 = 1.71 Hz,
J2 = 8.56 Hz, 1H, 3-H Ar), 7.15–7.24 (m, 3H, 4,5,8-H Ar), 7.48 (ds,
J = 0.98 Hz, 1H, 1-H Ar), 10.22 (br s, 1H, NY). 13C NMR (DMSO-d6): d
43.53 (N–CH3), 52.11 (5-C), 57.35 (6-C), 59.70 (3-C), 94.29 (2-C),
124.03 (3-C Ar), 125.30 (1-C Ar), 126.98 (6-C Ar), 127.19 (7-C Ar),
3.2. Synthesis
128.00 (8-C Ar), 128.32 (5-C Ar), 128.77 (4-C Ar), 132.76 (4
a-C Ar),
133.39 (8
a
-C Ar), 139.03 (2-C Ar). Anal. (C15H18BrNO2Á0.2H2O) C, H.
Melting points (mp) were determined with a digital Electro-
thermal IA 9000 series apparatus and are uncorrected. Infared
(IR) spectra were recorded on a Perkin–Elmer FT-IR spectrometer
(Spectrum RXI). 1H NMR spectra and 2D-NMR (COSY and HMQC)
were obtained with a Brücker Avance DRX 400 (400 MHz) spec-
trometer and 13C NMR spectra with a Brücker Avance DPX 200
(200 MHz) spectrometer. Elemental analyses were performed by
the Service Central de Microanalyse, France and are reported in
the Supplementary data.
3.2.1.5. 3-(2-Naphthalene)-octahydro-1,4-pyrido[2,1-c]oxazin-
3-ol hydrobromide (5). Yield 82%, mp 193.5–194.5 °C. IR: 3276
(m JY), 2654 (m
+NY). 1H NMR (DMSO-d6): d 1.36–1.47 (m, 2H, 8-
Hax and 9-Hax), 1.65–1.87 (m, 4H, 2 Â 7-H, 8-Heq and 9-Heq), 2.00
(s, 1H, -OH), 2.96 (q, J = 22.11 Hz, 1H, 6-Hax), 3.18–3.28 (m, 2H,
6-Heq and 4-Hax), 3.40–3.48 (m, 2H, 4-Heq and 9
J1 = 3.52 Hz, J2 = 12.52 Hz, 1H, 1-Heq), 4.01 (t, J = 12.52 Hz, 1H, 1-
a-Hax), 3.91 (dd,

A. P. Kourounakis et al. / Bioorg. Med. Chem. 18 (2010) 7402–7412
7409
Hax), 7.48–7.51 (m, 2H, 6,7-H Ar), 7.63 (dd, J1 = 1.57 Hz,
J2 = 8.61 Hz, 1H, 3-H Ar), 7.87–7.94 (m, 3H, 4,5,8-H Ar), 8.05 (s,
1H, 1-H Ar), 9.80 (br s, 1H, NY). 13C NMR (DMSO-d6): d 21.70 (8-
1.88–1.91 (m, 1H, 4
3.02 (d, J = 9.98 Hz, 1H, 3-Hax), 3.19 (t, J = 10.56 Hz, 1H, 3-Heq),
3.83 (s, 3H, O–CH3), 4.00–4.06 (m, 1H, -Yax), 7.15 (d,
a
-Yax), 2.20 (s, 1H, –OH), 2.72 (s, 3H, N–CH3),
8
a
C), 22.07 (7-C), 23.90 (9-C), 53.82 (6-C), 59.69 (4-C), 61.26 (9
C), 61.87 (1-C), 94.65 (3-C), 124.11 (3-C Ar), 125.31 (1-C Ar),
126.99 (6-C Ar), 127.20 (7-C Ar), 128.00 (8-C Ar), 128.31 (5-C Ar),
a
-
J = 8.81 Hz, 1H, 7-H naphthalene), 7.31 (s, 1H, 5-H naphthalene),
7.46 (s, 1H, 8-H naphthalene), 7.62 (d, J = 8.22 Hz, 2H, 2,6-H phe-
nyl), 7.76–7.88 (m, 4H, 3,5-H phenyl and 3,4-H naphthalene),
8.13 (s, 1H, 1-H naphthalene), 9.92 (br s, 1H, NH). 13C NMR
(DMSO-d6): d 23.83 (6,7-C), 24.39 (5-C), 24.72 (8-C), 31.05 (N–
128.78 (4-C Ar), 132.74 (4a-C Ar), 133.38 (8a-C Ar), 138.94 (2-C
Ar). Anal. (C18H22BrNO2) C, H.
CH3), 55.73 (O–CH3), 61.33 (3-C), 66.21 (4
94.11 (2-C), 106.18 (5-C naphthalene), 119.50 (7-C naphthalene),
125.62 (3-C naphthalene), 125.85 (8 -C naphthalene), 126.91
a-C), 70.31 (8a-C),
3.2.1.6. 2-(2-Naphthalene)-4-methyl-octahydro-1,4-benzoxa-
zin-2-ol hydrobromide (6). Yield 95%, mp 158–160 °C. IR:
a
3166 (
m
JY), 2656 (
m
+NY). 1H NMR (DMSO-d6): d 1.12–1.54 (m,
(1,4-C naphthalene), 127.91 (3,5-C phenyl), 129.17 (8-C naphtha-
4H, 8-Hax, 5-Hax and 2 Â 7-H), 1.73 (m, 2H, 2 Â 6-H), 1.89 (d,
lene), 130.19 (2,6-C phenyl), 134.17 (2-C naphthalene), 134.93
(4a-C naphthalene), 140.23 (4-C phenyl), 141.09 (1-C phenyl),
J = 10.06 Hz, 1H, 8-Heq), 2.18 (d, J = 10.43 Hz, 1H, 5-Heq), 2.42 (s,
1H, -OH), 2.71 (s, 3H, N–CH3), 3.07 (m, 1H, 4
3-Hax), 3.59 (d, J = 12.26 Hz, 1H, 3-Heq), 4.06 (dt, J1 = 3.90 Hz,
J2 = 10.98 Hz, 1H, 8 -Yax), 7.47–7.50 (m, 2H, 6,7-H Ar), 7.63 (dd,
J1 = 1.47 Hz, J2 = 8.60 Hz, 1H, 3-H Ar), 7.87–7.94 (m, 3H, 4,5,8-H
Ar), 8.05 (s, 1H, 1-H Ar), 9.95 (br s, 1H, NY). 13C NMR (DMSO-
d6): d 23.84 (6,7-C), 24.39 (5-C), 24.70 (8-C), 31.08 (N–CH3),
a
-Yax), 3.25 (m, 1H,
158.05 (6-C naphthalene). Anal. (C26H30BrNO3) C, H.
a
3.2.1.10. 2-[4-(2-Naphthalen)-phenyl]-4-methylmorpholin-2-ol
hydrobromide (10). Yield 40%, mp 116–118 °C. 1H NMR DMSO-d6
(DMSO-d6): d 2.73 (s, 3H, N–CH3), 2.91 (s, 1H, –OH), 3.09 (d,
J = 11.93 Hz, 1H, 3-Hax), 3.15–3.21 (m, 1H, 5-Hax), 3.48 (d,
J = 11.93 Hz, 1H, 3-Heq), 3.76 (m, 1H,5-Heq), 3.96 (dd, J1 = 2.93 Hz,
J2 = 12.52 Hz, 1H, 6-Heq), 4.24 (dt, J1 = 2.51 Hz, J2 = 12.33 Hz, 1H, 6-
Hax), 7.44–7.51 (m, 2H, 6,7-H naphthalene), 7.62 (d, J = 8.22 Hz, 2H,
2,6-H phenyl), 7.80–8.04 (m, 6H, 3,5-H phenyl and 3,4,5,8-H naph-
thalene), 8.20 (s, 1H, 1-H naphthalene), 9.80 (br s, 1H, NH). 13C
NMR (DMSO-d6): d 43.48 (N–CH3), 51.99 (5-C), 57.27 (6-C), 59.77
(3-C), 94.12 (2-C), 125.45 (1-C naphthalene), 125.83 (6,7-C naphtha-
lene), 126.75 (4-C naphthalene), 126.96 (3,5-C phenyl), 127.18 (2,6-
C phenyl), 127.98 (3-C naphthalene), 128.67 (5-C naphthalene),
61.26 (3-C), 66.19 (4
Ar), 125.32 (1-C Ar), 127.02 (6-C Ar), 127.21 (7-C Ar), 128.03 (8-
C Ar), 128.34 (5-C Ar), 128.78 (4-C Ar), 132.73 (4 -C Ar), 133.40
(8
-C Ar), 138.94 (2-C Ar). Anal. (C19H24BrNO2Á0.2H2O) C, H.
a-C), 70.40 (8a-C), 94.34 (2-C), 124.14 (3-C
a
a
3.2.1.7.
2-[4-(6-Methoxy-2-naphthalen)-phenyl]-4-methylmor
pholin-2-ol hydrobromide (7). Yield 57%, mp
157–160 °C. 1H NMR (CDCl3 + DMSO-d6): d 2.82 (ds, J = 3.91 Hz, 3H,
–NCH3), 3.02–3.10 (m, 1H, 5-Hax), 3.21–3.29 (m, 1H, 5-Heq), 3.37 (d,
J = 23.48 Hz, 2H, 2 Â 3-H), 3.38 (s, 1H, –OH), 3.85 (s, 3H, –OCH3),
3.96 (dd, J1 = 3.53 Hz, J2 = 12.52 Hz, 1H, 6-Heq), 4.47 (dt, J1 = 3.05 Hz,
J2 = 12.52 Hz, 1H, 6-Hax), 7.07–7.10 (m, 2H, 5,7-H naphthalene),
7.62–7.69 (m, 5H, 3,4,8-H naphthalene and 3,5-H phenyl), 7.72 (d,
J = 9.98 Hz, 2H, 2,6-H phenyl), 7.91 (s, 1H, 1-H naphthalene) 10.25
(br s, 1H, NH). 13C NMR (DMSO-d6): d 43.50 (N–CH3), 52.11 (5-C),
55.78 (O–CH3), 57.33 (6-C), 59.77 (3-C), 94.15 (2-C), 106.21 (5-C
naphthalene), 119.56 (7-C naphthalene), 125.65 (3-C naphthalene),
129.01 (8-C naphthalene), 132.82 (4a-C naphthalene), 133.76 (8a-
C naphthalene), 137.21 (2-C naphthalene), 140.74 (4-C phenyl),
140.90 (1-C phenyl). Anal. (C21H22BrNO2Á0.75H2O) C, H.
3.2.1.11. 3-[4-(2-Naphthalen)-phenyl]-octahydro-1,4-pyrido[2,1-c]
oxazin-3-ol hydrobromide (11). Yield 75%, mp 215–217 °C. 1H
NMR (DMSO-d6): d 1.35–1.45 (m, 2H, 8-Hax and 9-Hax), 1.66–1.85
(m, 4H, 2 Â 7-H, 8-Heq and 9-Heq), 2.93 (q, J = 20.64 Hz, 1H, 6-Hax),
3.15 (t, J = 10.57 Hz, 1H, 4-Hax), 3.23 (d, J = 11.54 Hz, 1H, 6-Heq),
125.84 (8
(3,5-C phenyl), 129.19 (8-C naphthalene), 130.24 (2,6-C phenyl),
134.20 (2-C naphthalene), 134.87 (4 -C naphthalene), 140.36 (4-C
a-C naphthalene), 126.87 (1,4-C naphthalene), 127.92
3.35–3.37 (m, 1H, 9a-Hax), 3.43 (d, J = 11.93 Hz, 1H, 4-Heq), 3.89
a
(dd, J1 = 3.33 Hz, J2 = 12.53 Hz, 1H, 1-Heq), 3.98 (t, J = 11.64 Hz, 1H,
1-Hax), 7.44–7.49 (m, 2H, 6,7-H naphthalene), 7.63 (d, J = 8.41 Hz,
2H, 2,6-H phenyl), 7.80–7.89 (m, 4H, 3,5-H phenyl and 3,4-H naph-
thalene), 7.93–7.97 (m, 2H, 5,8-H naphthalene), 8.20 (s, 1H, 1-H
naphthalene), 9.73 (br s, 1H, NH). 13C NMR (DMSO-d6): d 21.72 (8-
phenyl), 141.04 (1-C phenyl), 158.03 (6-C naphthalene).Anal.
(C22H24BrNO3Á2.0H2O) C, H.
3.2.1.8. 3-[4-(6-Methoxy-2-naphthalen)-phenyl]-octahydro-1,4-
pyrido[2,1-c]oxazin-3-ol hydrobromide (8). Yield 67%, mp 238–
239 °C. 1H NMR (CDCl3 + DMSO-d6): d 1.63–2.09 (m, 6H, 2 Â 7-H,
2 Â 8-H and 2 Â 9-H), 2.20 (q, J = 22.22 Hz, 1H, 6-Hax), 3.01 (q,
J = 22.01 Hz, 1H, 6-Heq), 3.14 (s, 1H, –OH), 3.28 (d, J = 12.03 Hz, 1H,
C), 22.08 (7-C), 23.92 (9-C), 53.88 (6-C), 59.76 (4-C), 61.32 (9a-C),
61.84 (1-C), 94.48 (3-C), 125.47 (1-C naphthalene), 125.83 (6,7-C
naphthalene), 126.76 (4-C naphthalene), 126.92 (3,5-C phenyl),
127.18 (2,6-C phenyl), 127.98 (3-C naphthalene) 128.67 (5-C naph-
4-Hax), 3.35 (d, J = 12.33 Hz, 1H, 4-Heq), 3.42–3.45 (m, 1H, 9a-Yax),
thalene), 129.03 (8-C naphthalene), 132.82 (4a-C naphthalene),
3.84 (dd, J1 = 3.04 Hz, J2 = 13.01 Hz, 1H, 1-Heq), 3.87 (s, 3H, –OCH3),
4.30 (t, J = 12.52 Hz, 1H, 1-Hax), 7.09–7.10 (m, 2H, 5,7-H naphtha-
lene), 7.61–7.76 (m, 7H, 3,4,8-H naphthalene and 2,3,5,6-H phenyl),
7.90 (s, 1H, 1-H naphthalene), 10.57 (br s, 1H, NH). 13C NMR (DMSO-
d6): d 21.73 (8-C), 22.10 (7-C), 23.93 (9-C), 53.87 (6-C), 55.72 (O–
133.75 (8 -C naphthalene), 137.20 (2-C naphthalene), 140.66 (4-C
a
phenyl), 140.90 (1-C phenyl). Anal. (C24H26BrNO2) C, H.
3.2.1.12. 2-[4-(2-Naphthalen)-phenyl]-4-methyl-octahydro-1,4-
benzoxazin-2-ol hydrobromide (12). Yield 79%, mp 169–170 °C.
1H NMR (DMSO-d6): d 0.98–1.43 (m, 5H, 8-Hax, 5-Hax, 2 Â 7-H and
–OH), 1.73 (m, 2H, 2 Â 6-H), 1.87 (d, J = 9.39 Hz, 1H, 8-Heq), 2.18 (d,
CH3), 59.81 (4-C), 61.33 (9
C naphthalene), 119.55 (7-C naphthalene), 125.65 (3-C naphtha-
lene), 125.84 (8 -C naphthalene), 126.89 (1,4-C naphthalene),
127.93 (3,5-C phenyl), 129.19 (8-C naphthalene), 130.23 (2,6-C phe-
nyl), 134.21 (2-C naphthalene), 134.88 (4 -C naphthalene), 140.29
a-C), 61.83 (1-C), 94.48 (3-C), 106.23 (5-
a
J = 11.54 Hz, 1H, 5-Heq), 2.71 (s, 3H, N–CH3), 3.04 (m, 1H, 4a-Yax),
3.21 (m, 1H, 3-Hax), 3.56 (d, J = 12.33 Hz, 1H, 3-Heq), 4.03 (dt,
a
J1 = 4.11 Hz, J2 = 10.57 Hz, 1H, 8 -Yax), 7.44–7.50 (m, 2H, 6,7-H
a
(4-C phenyl), 141.05 (1-C phenyl), 158.01 (6-C naphthalene). Anal.
(C25H28BrNO3Á0.25H2O) C, H.
naphthalene), 7.63 (d, J = 8.41 Hz, 2H, 2,6-H phenyl), 7.79–7.89
(m, 4H, 3,5-H phenyl and 3,4-H naphthalene), 7.93–7.97 (m, 2H,
5,8-H naphthalene), 8.20 (s, 1H, 1-H naphthalene), 9.93 (br s, 1H,
NH). 13C NMR (DMSO-d6): d 23.82 (6,7-C), 24.38 (5-C), 24.70 (8-
C), 31.07 (N–CH3), 61.32 (3-C), 66.22 (4a-C), 70.33 (8a-C), 94.15
(2-C), 125.50 (1-C naphthalene), 125.83 (6,7-H naphthalene),
3.2.1.9. 2-[4-(6-Methoxy-2-naphthalen)-phenyl]-4-methyl-octa-
hydro-1,4-benzoxazin-2-ol hydrobromide (9). Yield 60%, mp
106–107 °C. 1H NMR (DMSO-d6): d 1.14–1.45 (m, 6H, 8-Hax
,
5-Hax, 2 Â 7-H and 2 Â 6-H), 1.58–1.79 (m, 2H, 8-Heq and 5-Heq),
126.73 (4-C naphthalene), 126.97 (3,5-C phenyl), 127.18 (2,6-C

7410
A. P. Kourounakis et al. / Bioorg. Med. Chem. 18 (2010) 7402–7412
phenyl), 127.98 (3-C naphthalene), 128.66 (5-C naphthalene),
129.02 (8-C naphthalene), 132.83 (4 -C naphthalene), 133.75
(8 -C naphthalene), 137.25 (2-C naphthalene), 140.63 (4-C phe-
CH2Cl2–petroleum ether 1:4, yield 67%, mp 170.5–171.5 °C. 1H
NMR (CDCl3): d 3.89 (s, 3H, –OCH3), 4.43 (s, 2H, –CH2–), 7.11–
7.15 (m, 2H, 5,7-H naphthalene), 7.75–7.79 (m, 5H, 3,4,8-H naph-
thalene and 3,5-H phenyl), 7.97 (s, 1H, 1-H naphthalene), 8.03 (d,
J = 8.13 Hz, 2H, 2,6-H phenyl). Anal. (C19H15BrO2Á0.5H2O) C, H.
a
a
nyl), 140.92 (1-C phenyl). Anal. (C25H28BrNO2) C, H.
3.2.2. Preparation of intermediate and starting materials
3.2.2.1. General procedure for the cross-coupling reactions (13–
14). To a three-necked flask were added 2 mmol of arylhalide,
2.2 mmol of 1-(4-acetyl-phenyl)-boronic acid, 0.2 mol % of
Pd(OAc)2, 5 mmol of powdered K2CO3, and 2 mmol of Bu4NBr.
The flask was flushed with argon and equipped with a rubber sep-
tum. Water (2.2 mL) was added with a syringe, and the resulting
suspension was energetically stirred and degassed to remove O2.
The mixture was stirred and heated for 1.5–2 h at 70 °C under ar-
gon. It was then cooled to room temperature, diluted with water,
and extracted with EtOAc. The solution was dried (Na2SO4), the sol-
vent was removed in vacuum, and the crude product purified by
silica gel column chromatography to yield the desirable product.
3.2.2.3.3.
2-Bromo-1-[4-(2-naphthalene)-phenyl]-ethanone
Purified by flash chromatography eluting with CH2Cl2–
(18)35
.
petroleum ether 1:6, yield 76%, mp 154–156 °C. 1H NMR (CDCl3):
d 4.43 (ds, J = 1.23 Hz, 2H, –CH2–), 7.45–7.50 (m, 2H, 6,7-H naph-
thalene), 7.70 (dd, J1 = 1.59 Hz, J2 = 8.44 Hz, 1H, 1-H naphthalene),
7.77–7.90 (m, 5H, 3,5-H phenyl and 3,4,5-H naphthalene), 8.03–
8.05 (m, 3H, 8-H naphthalene and 2,6-H phenyl).
3.2.2.4. trans-2-Methylamino-cyclohexanol (19). To a solution
of cyclohexene oxide (55.7 mmol) in 20 mL methanol and 15 mL
aqueous solution (40%) of methylamine, was stirred for 19 h at
rt. The product was isolated by vacuum distillation (bp4 66–
69 °C). Yield 70%.
3.2.2.1.1.
1-[4-(6-Methoxy-2-naphthalen)-phenyl]-ethanone
(13). Purified by flash chromatography eluting with EtOAc–
petroleum ether 1:3, yield 90%, mp 186.5–188.5. 1H NMR (CDCl3):
d 2.66 (s, 3H, –CH3), 3.96 (s, 3H, –OCH3), 7.18–7.22 (m, 2H, 5,7-H
naphthalene), 7.73–7.85 (m, 5H, 3,4,8-H naphthalene and 3,5-H
phenyl), 8.04 (s, 1H, 1-H naphthalene), 8.07 (d, J = 8.21 Hz, 2H,
2,6-H phenyl). Anal. (C19H16O2Á0.2H2O) C, H.
3.3. In vitro lipid peroxidation
Heat-inactivated hepatic microsomes from untreated rats were
prepared as described.36 The incubation mixture contained heat-
inactivated (90 °C for 90 s) microsomal fraction (corresponding to
2.5 mg of hepatic protein per milliliter or 4 mM fatty acid resi-
dues37), ascorbic acid (0.2 mM) in Tris–HCl/KCl buffer (50 mM/
3.2.2.1.2. 1-[4-(2-Naphthalen)-phenyl]-ethanone (14)32
.
Puri-
fied by flash chromatography eluting with EtOAc–petroleum ether
1:8, yield 84%, mp 138.5–139.5 °C. 1H NMR (CDCl3): d 2.61 (s, 3H, –
CH3), 7.44–7.50 (m, 2H, 6,7-H naphthalene), 7.71 (dd, J1 = 1.65 Hz,
J2 = 8.60 Hz, 1H, 3-H naphthalene), 7.76 (d, J = 8.05 Hz, 2H, 3,5-H
phenyl), 7.81–7.90 (m, 3H, 4,5,8-H naphthalene), 8.01 (s, 1H, 1-H
naphthalene), 8.03 (d, J = 7.98 Hz, 2H, 2,6-H phenyl).
150 mM, pH 7.4), and the studied compounds (10
solved in dimethyl sulfoxide. The reaction was initiated by the
addition of a freshly prepared FeSO4 solution (10 M), and the mix-
lV–1 mM) dis-
l
ture was incubated at 37 °C for 45 min. Aliquots were taken at var-
ious time intervals and lipid peroxidation was assessed by
spectrophotometric (535 against 600 nm) determination of the 2-
thiobarbituric acid reactive material.8 All compounds and solvents
were tested and found not to interfere with the assay. Each assay
was performed in duplicate and IC50 values represent the mean
concentration of compounds that inhibits the peroxidation of con-
trol microsomes by 50% after 45 min of incubation. All standard er-
rors are within ten percent of the respective reported values.
3.2.2.2. 2-Bromoacetylnaphthalene (15)33
. Bromine (5.5 mmol)
in dry chloroform (20 mL) was added dropwise to a stirred solution
of the corresponding acetyl derivative 4-cyclohexyl-a-acetophe-
none (19) or 2-acetylnaphthalene (20) (5 mmol) in dry chloroform
(10 mL). After the addition of bromine, the solution was stirred at
room temperature for 2.5 h. The solution was washed with H2O, 5%
NaHCO3, and H2O, dried (Na2SO4) and concentrated. The residue
was subjected to flash column chromatography on silica gel eluting
with CH2Cl2–petroleum ether 1:2, to give the desirable product at a
yield of 82%. mp 78–79.5 °C. 1H NMR (CDCl3): d 4.61 (s, 2H, –CH2–),
7.58–7.67 (m, 2H, 6,7-Y Ar), 7.90–7.96 (m, 2H, 3,4-Y Ar), 7.99–8.06
(m, 2H, 5,8-Y Ar), 8.53 (s, 1H, 1-Y Ar).
3.4. In vitro squalene synthase activity assay
SQS activity was evaluated by determining the amount of [3H]
FPP converted to squalene as previously described.8,38 The assay
was performed in 1 mL of 50 mM phosphate buffer, pH 7.4, con-
taining 10 mM MgCl2, 0.5 mM NADPH, rat liver microsomes
(18
l
g protein/mL), various concentrations of the test compounds
3.2.2.3. General procedure for the preparation of bromo-acetyl
derivatives 16–18. To a solutionofthe acetyl derivative 2-acetyl-6-
methoxynaphthalene (21) or 1-[4-(6-methoxy-2-naphthalen)-phe-
nyl]-ethanone (13) or 1-[4-(2-naphthalen)-phenyl]-ethanone (14)
(5 mmol) in 10 mL of anhydrous tetrahydrofuran contained in a
100 mL flask was added phenyltrimethylammonium tribromide
(5 mmol) in small portions over a 30 min period. A white precipitate
forms immediately. The mixture was stirred at room temperature for
1.5–2 h and then cold water was added. The crystalline precipitate
wasfiltered,washedwithwater, dried, andwassubjectedtoflashcol-
umn chromatography on silica gel to give the desirable product.
or simvastatin dissolved in ethanol, and [3H] FPP (0.5
l
M,
0.27 Ci/mmol) in a 10x100 mm glass screw-cap tube. All compo-
nents except [3H] FPP were preincubated for 10 min at 37 °C. The
reaction was initiated by the addition of [3H] FPP. After 10 min at
37 °C, the reaction was terminated by the addition of 1 mL 15%
KOH in ethanol. The tubes were incubated at 65 °C for 30 min to
solubilize proteins. The mixture was extracted with 5 mL petro-
leum ether for 10 min. After freezing the lower aqueous phase,
the organic phase was transferred to glass tubes containing 2 mL
distilled water. After washing, 1.5 mL of the upper organic phase
was removed and counted with 3 mL scintillation liquid using a
Beckman scintillation counter. Each assay was performed in tripli-
cate and IC50 values represent the mean concentration of com-
pounds that inhibits the activity of the enzyme by 50%. All
standard errors are within 10% of the respective reported values.
3.2.2.3.1. 2-Bromoacetyl-6-methoxynaphthalene (16)34
.
Purified
by flash chromatography eluting with CH2Cl2–petroleum ether 1:1,
yield 73%, mp 108–110 °C. 1H NMR (CDCl3): d 3.97 (s, 3H, O–CH3),
4.56 (s, 2H, –CH2–), 7.17 (ds, J = 2.27 Hz, 1H, 5-H Ar), 7.23 (dd,
J1 = 2.45 Hz, J2 = 8.95 Hz, 1H, 3-H Ar), 7.80 (d, J = 8.70 Hz, 1H, 7-H
Ar), 7.88 (d, J = 8.98 Hz, 1H, 4-H Ar), 8.01 (dd, J1 = 1.65 Hz,
J2 = 8.66 Hz, 1H, 8-H Ar), 8.45 (s, 1H, 1-H Ar).
3.5. In vitro lipoxygenase activity assay
3.2.2.3.2. 2-Bromo-1-[4-(6-methoxy-2-naphthalene)-phenyl]-eth-
anone (17). Purified by flash chromatography eluting with
Lipoxygenase activity was determined using soybean lipoxyge-
nase (250 U/mL) and sodium linoleate (100 lV) as substrate, in

A. P. Kourounakis et al. / Bioorg. Med. Chem. 18 (2010) 7402–7412
7411
Tris–HCl buffer pH 9.0. The test compounds in 60% ethanol were
3.10.3. Docking studies
added and the reaction was monitored for 6 min at 28 °C, recording
absorbance at 234 nm.24 The assay for each concentration was car-
ried out twice. All standard errors are within 10% of the respective
reported values.
Autodock Tools42 was used in order to set up the docking runs.
All files (protein structure and ligands) were transformed in the
appropriate pdbqt formats with proper gasteiger charges assigned.
The active site of the protein was easily identified as the cavity
containing the co-crystallized ligand CP-320473. The docking
search space was defined as the lattice containing the co-crystal-
lized ligand, covering the entire volume of the cavity. Autogrid 4
was then used in order to generate the appropriate grid lattice files
for every candidate ligand to be used in the subsequent docking
runs.
3.6. In vivo evaluation of antihypercholesterolemic and antihy
perlipidemic activity
An aqueous solution of Triton WR 1339 was given ip to rats
(200 mg/kg) and one hour later the test compounds (56 lmol/
kg), dissolved in saline, or saline only were administered ip.39 After
24 h, blood was taken from the aorta and used for the determina-
tion of plasma total cholesterol (TC), LDL-cholesterol (LDL-C) and
triglyceride (TG) levels, using commercially available kits.8 Levels
of plasma lipids were determined in duplicate while values pre-
sented are the mean from 8–10 rats (per compound). All standard
errors are within 12% of the respective reported values.
Docking was performed using Autodock 4’s Lamarckian Genetic
Algorithm search method.28 The protein structure for this proce-
dure was kept rigid, while the ligand (compound) structures were
assigned (by Autodock Tools) the appropriate degrees of freedom
in terms of their flexibility. The genetic algorithm search parame-
ters included: 300 binding modes as the initial algorithm popula-
tion and 250,000 as the maximum number of energy evaluations.
These values are in accordance to the ones suggested for the num-
ber of torsions assigned to the ligands (two to three torsions in our
case). Default docking parameters (e.g., random number generator,
energy outside the grid, number of retries, output format, cluster
analysis) were used.
3.7. In vivo inhibition of carrageenan-induced rat-paw edema25
For the in vivo anti-inflammatory activity, Wistar male rats
(150–180 g) were injected with 0.1 mL carrageenan (2% w/v solu-
tion in physiological saline) id into the right hind paw, the left
Acknowledgments
paw serving as control. The test compounds (300 lmol/kg) were
given ip after the carrageenan injection, and 3.5 h later the
produced edema was estimated as paw weight increase. Values
are the mean from 8–10 rats (per compound). All standard errors
are within 12% of the respective reported values.
The authors wish to acknowledge the Special Account for Re-
search Grants of the University of Athens (Grant 70/4/7842) and
the Hellenic Society of Atherosclerosis, for financial support.
Supplementary data
3.8. Protein determination
Supplementary data (elemental analysis data of all novel com-
pounds (i.e., all final and some intermediates)) associated with this
article can be found, in the online version, at doi:10.1016/
j.bmc.2010.09.008.
The protein content of microsomal fractions, used in lipid per-
oxidation and squalene synthase inhibitory assay, was determined
according to Lowry’s method.40
3.9. Statistical analysis
References and notes
Data are expressed as mean standard deviation. Where indi-
cated, statistical comparisons were made using Student’s t-test
and a statistically significant difference was inferred if P <0.05.
1. Cassar, A.; Holmes, D. R., Jr.; Rihal, C. S.; Gersh, B. J. Mayo Clin. Proc. 2009, 84,
1130.
2. Libby, P.; Ridker, P. M. Circulation 2002, 105, 1135.
3. Biller, S. A.; Neuenschwander, K.; Ponpipom, M. M.; Poulter, C. Curr. Pharm. Des.
1996, 2, 1.
4. Elsayed, R. K.; Evans, J. D. Expert. Opin. Emerg. Drugs 2008, 13, 309.
5. Seiki, S.; Frishman, W. H. Cardiol. Rev. 2009, 17, 70.
3.10. Molecular modeling
6. Chrysselis, M. C.; Rekka, E. A.; Kourounakis, P. N. J. Med. Chem. 2002, 43, 609.
7. Chrysselis, M. C.; Rekka, E. A.; Siskou, I. C.; Kourounakis, P. N. J. Med. Chem.
2002, 45, 5406.
8. Kourounakis, A. P.; Charitos, C.; Rekka, E. A.; Kourounakis, P. N. J. Med. Chem.
2008, 51, 5861.
9. Tavridou, A.; Kaklamanis, L.; Megaritis, G.; Kourounakis, A. P.; Papalois, A.;
Roukounas, D.; Rekka, E. A.; Kourounakis, P. N.; Charalambous, A.;
Manolopoulos, V. G. Eur. J. Pharmacol. 2006, 535, 34.
10. Tavridou, A.; Megaritis, G.; Kourounakis, A. P.; Charalambous, A.;
Manolopoulos, V. G. Endothelium 2007, 14, 239.
3.10.1. Preperation of ligands
Molecular modeling of the compounds was performed using
SPARTAN’04, (Wavefunction, Inc. Irvine, CA. 2004). All initial struc-
tures underwent conformational analysis (by use of molecular
mechanics) and the lowest energy conformers were subsequently
optimized using the semiempirical AM1 method. This procedure
produced lowest energy conformations that were used in the dock-
ing process.
11. Tavridou, A.; Kaklamanis, L.; Papalois, A.; Kourounakis, A. P.; Rekka, E. A.;
Kourounakis, P. N.; Charalambous, A.; Manolopoulos, V. G. J. Pharmacol. Exp.
Ther. 2007, 323, 794.
12. Stocker, R.; Keaney, J. F. Physiol. Rev. 2004, 84, 1381.
13. (a) Willcox, B. J.; Curb, J. D.; Rodriguez, B. L. Am. J. Cardiol. 2008, 101, 75D; (b)
Jaxa-Chamiec, T.; Bednarz, B.; Herbaczynska-Cedro, K.; Maciejewski, P.;
Ceremuzynski, L. Cardiology 2009, 112, 219.
14. (a) Briasoulis, A.; Tousoulis, D.; Antoniades, C.; Stefanadis, C. Curr. Pharm. Des.
2009, 15, 3078; (b) Sesso, H. D.; Buring, J. E.; Christen, W. G.; Kurth, T.;
Belanger, C.; MacFadyen, J.; Bubes, V.; Manson, J. E.; Glynn, R. J.; Gaziano, J. M.
JAMA 2008, 300, 2123.
15. (a) Ray, W. A.; Varas-Lorenzo, C.; Chung, C. P.; Castellsague, J.; Murray, K. T.;
Stein, C. M.; Daugherty, J. R.; Arbogast, P. G.; García-Rodríguez, L. A. Circ.
Cardiovasc. Qual. Outcomes 2009, 2, 155; (b) Hermann, M. Curr. Rheumatol. Rep.
2009, 11, 31.
16. Doulgeris, C. M.; Galanakis, D.; Kourounakis, A. P.; Tsiakitzis, K. C.; Gavalas, D.;
Eleftheriou, P. T.; Victoratos, P.; Rekka, E. A.; Kourounakis, P. N. Bioorg. Med.
Chem. 2006, 16, 825.
3.10.2. Selection and preparation of protein crystal structure
The protein crystal structure of human squalene synthase 1EZF
was obtained from the Brookhaven Protein Data Bank. The enzyme
structure, identified by X-ray diffraction at 2.15 Å resolution,
contained a co-crystallized inhibitor codenamed CP-320473 (N-{2-
[trans-7-chloro-1-(2,2-dimethyl-propyl)-5-naphthalen-1-yl-
2-oxo-1,2,3,5-tetrahydrobenzo[e] [1,4]oxazepin-3-yl]-acetyl}-as
partic acid) and several water molecules. For the docking studies,
all water molecules were removed and the single protein chain, that
is, the monomer with the respective bound inhibitor, was used. This
preparation procedure was performed using Chimera.41

7412
A. P. Kourounakis et al. / Bioorg. Med. Chem. 18 (2010) 7402–7412
17. Cavalli, A.; Bolognesi, M. L.; Minarini, A.; Rosini, M.; Tumiatti, V.; Recanatini,
M.; Melchiorre, C. J. Med. Chem. 2008, 51, 347.
18. Hansch, C.; Leo, A. Substituent Constants for Correlation Analysis in Chemistry and
Biology; J. Wiley & Sons: New York, 1991.
19. Rankin, G. O.; Riley, T. N.; Murphy, J. C. J. Med. Chem. 1978, 21, 460.
20. Bettoni, G.; Franchini, C.; Perrone, R.; Tortorella, V. Tetrahedron 1980, 36,
409.
29. Pandit, J.; Danley, D. E.; Schulte, G. K.; Mazzalupo, S.; Pauly, T. A.; Hayward, C.
M.; Hamanaka, E. S.; Thompson, J. F.; Harwood, H. J. J. Biol. Chem. 2000, 275,
30610.
30. Robinson, G. W.; Tsay, Y. H.; Kienzle, B. K.; Smith-Monroy, C. A.; Bishop, R. W.
Mol. Cell. Biol. 1993, 13, 2706.
31. Tarshis, L. C.; Proteau, P. J.; Kellogg, B. A.; Sacchettini, J. C.; Poulter, C. D. Proc.
Natl. Acad. Sci. U.S.A. 1996, 93, 15018.
21. Brown, G. R.; Clarke, D. S.; Foubister, A. J.; Freeman, S.; Harrison, P. J.; Johnson,
M. C.; Mallion, K. B.; McCormick, J.; McTaggart, F.; Reid, A. C.; Smith, G. J.;
Taylor, M. J. J. Med. Chem. 1996, 39, 2971.
22. (a) Mathur, S. N.; Field, F. J.; Spector, A. A.; Armstrong, M. A. Biochim. Biophys.
Acta 1985, 837, 13; (b) Yla-Herttuala, S.; Rosenfeld, M. E.; Parthasarathy, S.;
Glass, C. K.; Sigal, E.; Witztum, J. L.; Steinberg, D. Proc. Natl. Acad. Sci. U.S.A.
1990, 87, 6959.
32. Duan, J. P.; Cheng, C. H. Tetrahedron Lett. 1993, 34, 4019.
33. Immediata, T.; Day, A. R. J. Org. Chem. 1940, 38, 512.
34. Marquet, A.; Jacques, J. Tetrahedron Lett. 1959, 9, 24.
35. Black, C.; Mancini, J. A.; Lau, C. K.; Prasit, P.; Vickers, P. J. GB 2283745, 1995;
Chem. Abstr. 1995, 123, 227989.
36. Rekka, E.; Kolstee, J.; Timmerman, H.; Bast, A. Eur. J. Med. Chem. 1989, 24, 43.
37. Eichenberger, K.; Bohni, P.; Winterhalter, K. H.; Kawato, S.; Richter, C. FEBS Lett.
1982, 142, 59.
23. Cornicelli, J. A. IDrugs 1998, 1, 206.
24. Taraporewala, I.; Kauffman, J. M. J. Pharm. Sci. 1990, 79, 173.
25. Hadjipetrou-Kourounakis, L.; Rekka, E.; Kourounakis, A. Ann. N.Y. Acad. Sci.
1992, 650, 19.
26. Di Rosa, M.; Giroud, J. P.; Willoughby, D. A. J. Pathol. 1971, 104, 15.
27. Huang, M. H.; Wu, S. N.; Wang, J. P.; Lin, C. H.; Lu, S. I.; Liao, L. F.; Shen, A. Y.
Drug Dev. Res. 2003, 60, 261.
38. Amin, D.; Cornell, S. A.; Gustafson, S. K.; Needle, S. J.; Ullrich, J. W.; Bilder, G. E.;
Perrone, M. H. J. Lipid Res. 1992, 33, 1657.
39. Kuroda, M.; Tanzawa, K.; Tsujita, Y.; Endo, A. Biochim. Biophys. Acta 1977, 489, 119.
40. Lowry, O. H.; Rosebrough, N. J.; Farr, A. L.; Randall, R. J. J. Biol. Chem. 1951, 193,
265.
41. Pettersen, E. F.; Goddard, T. D.; Huang, C. C.; Couch, G. S.; Greenblatt, D. M.;
Meng, E. C.; Ferrin, T. E. J. Comput. Chem. 2004, 25, 1605.
42. Sanner, M. F. J. Mol. Graphics Modell. 1999, 17, 57.
28. Morris, G. M.; Goodsell, D. S.; Halliday, R. W.; Huey, R.; Hart, W. E.; Belew, R. K.;
Olson, A. J. J. Comput. Chem. 1998, 19, 1639.