Identification of anxiolytic/nonsedative agents among indol-3- ylglyoxylamides acting as functionally selective agonists at the γ-aminobutyric acid-A (GABAA) α2 benzodiazepine receptor

Article abstract of DOI:10.1021/jm9001154

Anxioselective agents may be identified among compounds binding selectively to the α₂β_xγ₂ subtype of the γ-aminobutyric acid-A (GABA_A)/central benzodiazepine receptor (BzR) complex and behaving as agonists or among compounds binding with comparable potency to various BzR subtypes but eliciting agonism only at the α₂β_xγ₂ receptor. Because of subtle steric differences among BzR subtypes, the latter approach has proved much more successful. A biological screening within the class of indol-3-ylglyoxylamides 1-3 allowed us to identify compounds 1c and 2b as potential anxiolytic/nonsedative agents showing α₂ selective efficacy in vitro and anxioselective effects in vivo. According to molecular modeling studies, and consistently with SARs accumulated in the past decade, 5-NO₂- and 5-H-indole derivatives would preferentially bind to BzR by placing the indole ring in the L_Di and the L₂ receptor binding sites, respectively.

Full text of DOI:10.1021/jm9001154

J. Med. Chem. 2009, 52, 3723–3734
3723
Identification of Anxiolytic/Nonsedative Agents among Indol-3-ylglyoxylamides Acting as
Functionally Selective Agonists at the γ-Aminobutyric Acid-A (GABAA) r2 Benzodiazepine
Receptor
Sabrina Taliani,†,* Barbara Cosimelli,‡,* Federico Da Settimo,† Anna Maria Marini,† Concettina La Motta,† Francesca Simorini,†
Silvia Salerno,† Ettore Novellino,‡ Giovanni Greco,‡ Sandro Cosconati,‡ Luciana Marinelli,‡ Francesca Salvetti,§
Gianluca L’Abbate,§ Silvia Trasciatti,§ Marina Montali, Barbara Costa,| and Claudia Martini
Dipartimento di Scienze Farmaceutiche, UniVersita` di Pisa, Via Bonanno 6, 56126 Pisa, Italy, Dipartimento di Chimica Farmaceutica e
Tossicologica, UniVersita` di Napoli “Federico II”, Via Domenico Montesano 49, 80131 Napoli, Italy, Abiogen Pharma SpA, Research Centre,
Via del Paradiso 6, 56019 Migliarino Pisano, Pisa, Italy, Dipartimento di Psichiatria, Neurobiologia, Farmacologia e Biotecnologie,
UniVersita` di Pisa, Via Bonanno 6, 56126 Pisa, Italy, Dipartimento di Morfologia Umana e Biologia Applicata, UniVersita` di Pisa,
Via Volta, 4, 56126 Pisa, Italy
ReceiVed January 29, 2009
Anxioselective agents may be identified among compounds binding selectively to the R2ꢀxγ2 subtype of the
γ-aminobutyric acid-A (GABAA)/central benzodiazepine receptor (BzR) complex and behaving as agonists
or among compounds binding with comparable potency to various BzR subtypes but eliciting agonism only
at the R2ꢀxγ2 receptor. Because of subtle steric differences among BzR subtypes, the latter approach has
proved much more successful. A biological screening within the class of indol-3-ylglyoxylamides 1-3 allowed
us to identify compounds 1c and 2b as potential anxiolytic/nonsedative agents showing R2 selective efficacy
in vitro and anxioselective effects in vivo. According to molecular modeling studies, and consistently with
SARs accumulated in the past decade, 5-NO2- and 5-H-indole derivatives would preferentially bind to
BzR by placing the indole ring in the LDi and the L2 receptor binding sites, respectively.
Introduction
subunits. The major benzodiazepine-sensitive GABAA receptor
subtypes in the brain are R1ꢀxγ2, R2ꢀxγ2, R3ꢀxγ2, and R5ꢀxγ2,
with the BzR being located between the R and γ subunits. The
R4ꢀxγ2 and R6ꢀxγ2 subtypes do not respond to benzodiazepines
and are therefore called benzodiazepine-insensitive receptors.7
The γ2 subunit is the major occurring γ subunit in the brain.
The ꢀ subunit does not seem to affect the pharmacology of
benzodiazepines, whereas it has been demonstrated that the R
subunit is the main determinant of BzR ligand selectivity
(therefore BzR subtypes take their names from the R subunit).7-9
In situ mRNA hybridization, subunit specific immunoprecipi-
tation, and immunoaffinity chromatography have allowed the
determination of the distribution of the GABAA subtypes in the
brain. The R1ꢀxγ2 subtypes are the dominant ones and are present
in both the cerebellum and the cortex. The R2ꢀxγ2 and R3ꢀxγ2
receptors are of medium abundance and are found mainly in
the cortex and hippocampus, whereas the R5ꢀxγ2 isoforms are
scarcely abundant as they are only largely expressed in the
hippocampus. This differential localization of the GABAA
receptor subtypes in brain areas has suggested that the different
subtypes may be associated with different physiological effects.
In particular, it was demonstrated that the R1 containing BzR
mediates sedative action, that the R2 subtype is involved in
anxiolytic and myorelaxation effects, and that the R5 subtype
is associated with cognition processes like learning and memo-
rizing. Finally, the role of the R3 subtype seems to be mainly
involved in mediating anxiety.7-9 The correlation between
pharmacological profile and specific action at the various BzR
subtypes provides a rationale for the search of molecules capable
of binding and/or eliciting allosteric modulation at a single or
a group of BzR subtype/s. Thus, an anxioselective BzR ligand
may be identified either among compounds binding selectively
to the R2 subtype and behaving as agonists (affinity-based
selective agents) or among compounds binding with comparable
Benzodiazepines are currently the drugs of first choice in the
treatment of anxiety. These compounds bind to an allosteric
site located at the γ-aminobutyric acid-A (GABAA ) receptor
a
complex, the so-called central benzodiazepine receptor (BzR).1
Actually, benzodiazepines represent just one of the many
chemically diverse classes of ligands of the GABAA/BzR
complex that display a pharmacological action ranging from
full agonism (anxiolytic, anticonvulsant, sedative-hypnotic, and
myorelaxant agents) to antagonism (agents to reverse sedation
caused by BzR agonists), and to inverse agonism (anxiogenic,
somnolytic, and proconvulsant agents). Agonists and inverse
agonists are positive and negative allosteric modulators of
GABA affinity, respectively, while antagonists do not modify
GABA binding.2,3
The GABAA/BzR complex contains a chloride channel and
is a membrane-bound heteropentameric protein that may be
assembled from at least 21 subunits belonging to eight different
classes (6R, 4ꢀ, 4γ, 1ε, 1δ, 3F, 1θ, and 1π).4-6 It has been
found that a fully functional GABAA/BzR must contain an R
subunit, a ꢀ subunit, and a γ subunit, and it is currently accepted
that the predominant native receptors comprise 2R, 2ꢀ, and 1γ
* To whom correspondence should be addressed. For S.T.: phone,
+390502219547; fax, +390502219605; E-mail, taliani@farm.unipi.it. For
B.C.: phone, +39081678614; fax, +39081678630; E-mail, barbara.cosimelli@
unina.it.
†
Dipartimento di Scienze Farmaceutiche, Universita` di Pisa.
Dipartimento di Chimica Farmaceutica e Tossicologica, Universita` di
‡
Napoli “Federico II”.
§
Abiogen Pharma SpA, Research Centre.
Dipartimento di Morfologia Umana e Biologia Applicata, Universita`
|
di Pisa.
Dipartimento di Psichiatria, Neurobiologia, Farmacologia e Biotec-
nologie, Universita` di Pisa.
a Abbreviations: GABA, γ-aminobutyric acid; BzR, central benzodiaz-
epine receptor; LD test, light-dark test; MD, molecular dynamics.
10.1021/jm9001154 CCC: $40.75
2009 American Chemical Society
Published on Web 05/26/2009

3724 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 12
Chart 1. Structures of Indol-3-ylglyoxylamide Derivatives 1-3
Taliani et al.
Figure 1. Hypothetical binding modes A and B of indole derivatives
to BzR within the framework of Cook’s pharmacophore/topological
model.27,28
Scheme 1. Synthesis of Indol-3-ylglyoxylamide Derivatives
3c-j
potency to various BzR subtypes that elicit agonism at the R2
subtype and antagonism at other subtypes (efficacy-based
selective agents). Compared to classical nonspecific BzR
agonists, such as diazepam, either affinity or efficacy-based R2
selective agonists are expected to retain anxiolytic properties
without side effects like sedation, ataxia, tolerance, dependence,
and impairment of cognitive processes.10,11
Molecular modeling studies performed by Cook’s group12
on structurally different classes of BzR ligands led to the
development of a comprehensive pharmacophore/topological
model consisting of several BzR interaction (sub)sites: (i) an
H-bond acceptor (A2), (ii) an H-bond donor (H1), (iii) an H-bond
donor/acceptor (H2/A3), (iv) four lipophilic pockets (L1, L2, L3,
and LDi), and (v) three sterically forbidden sites (S1, S2, and S3)
as boundaries of the receptor binding cleft.
Further studies by the same researchers have suggested that
the shapes of the BzR subtypes are very similar, with the
exception of R1 and R5 subtypes that seem to be slightly larger
in size at two distinct lipophilic regions, called LDi and L2
regions, respectively.13 The above steric differences have been
exploited to obtain ligands that bind selectively to either R1 or
R5 subtypes14 but have also hampered the identification of
affinity-based R2 and/or R3 selective ligands.15 The search of
efficacy-based R2 and/or R3 selective ligands by screening and
lead optimization methods has yielded better results in a variety
of chemical classes.10,16-25
In the past several years, our research group has described
the synthesis and biological evaluation of several BzR ligands
featuring an indol-3-ylglyoxyl scaffold: N-benzylindol-3-ylg-
lyoxylamides (1),26 (R) and (S) enantiomers of N-(R-substituted-
benzyl)indol-3-ylglyoxylamides (2),27 and N-(indan-1-yl)indol-
3-ylglyoxylamides (3)27 (Chart 1). The structure-affinity
relationships of these ligands were rationalized by assuming that
they adopt two alternative binding modes, called A and B in
their interaction with the BzR (Figure 1). Specifically, the 5-Cl/
NO2 derivatives adopt preferentially the binding mode A,
whereas the 5-H derivatives preferentially adopt the binding
mode B.27,28
Prompted by the interesting potency displayed by some of
the indoles 1-3 at the wild type BzR, as well as their favorable
pharmacokinetic properties,28,29 we tested them as potential
anxioselective agents in three steps: (i) determine the affinity
for the rat recombinant R1ꢀ2γ2, R2ꢀ2γ2, and R5ꢀ3γ2 BzR
subtypes, (ii) determine the efficacy profile for compounds
displaying the highest potency at the R2 BzR subtype, and (iii)
evaluate the in vivo pharmacological properties of compounds
that turned out to be affinity- or efficacy-based R2 selective.
The same screening was then extended to newly synthesized
derivatives of the relatively rigid N-(indan-1-yl)indol-3-ylgly-
oxylamide scaffold27 bearing different substituents (X ) F, NO2,
CH3, OCH3) at the 5′ position of the indane nucleus (compounds
3c-j).
Herein, we describe the synthesis of the novel derivatives
3c-j and the biological/pharmacological studies on a number
of indoles 1-3 leading to the identification of functionally
selective agonists of the R2 BzR subtype characterized by in
vivo anxioselective properties. Finally, molecular modeling
studies were carried out to rationalize the structure-affinity
relationships in the class of indol-3-ylglyoxylamides at the
molecular level.
Chemistry. Acylation of the appropriate indoles 4 and 5 with
oxalyl chloride,30 followed by reaction of the indolylglyoxylyl
chlorides 6 and 7 in toluene with the appropriate amines 8-11
under mild conditions and in the presence of triethylamine
afforded the new compounds 3c-j (Scheme 1).
The 1-indanamines 8,3110,32 and 1133 were prepared follow-
ing a published synthetic procedure33 starting from the corre-
sponding 5-substituted-1-indanones, whereas the 5-nitro sub-
stituted derivative 9 was obtained in three steps according to
another synthetic route (Scheme 2). The 5-nitro-1-indanone 13
was synthesized in good yield (66%) by oxidation of the
5-amino-1-indanone 12 with m-chloroperoxybenzoic acid in
dichloromethane. Reaction of 13 with methoxyamine led to the
intermediate methyloxime 14 that was converted into the final
amine 9 by treatment with diborane in tetrahydrofuran. All
1-indanamines 8-11 were used as racemic mixtures.
Biological and Pharmacological Studies. The binding
affinity of the newly synthesized indole derivatives 3c-j for
the BzR in bovine brain membranes was determined by

Anxiolytic/NonsedatiVe Agents among Indol-3-ylglyoxylamides
Scheme 2. Synthesis of 5-Nitro-1-indanamine 9
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 12 3725
derivatives, the insertion of electron-withdrawing groups on the
side chain of the aromatic ring does not improve the affinity,
differently from the trend observed in the benzylamine series
(compare 3c and 3e vs 3a, with respect to 1c and 1e vs 1a).
The novel compounds endowed with significant potency at
the wild type BzR (3c,d,f,i,j), together with some previously
described indoles (1c-h, 2a,b, and 3a,b), were assayed for their
affinity to recombinant rat R1ꢀ2γ2, R2ꢀ2γ2, and R5ꢀ3γ2 BzRs
(Table 1).
An overview of the binding data shows that all tested
compounds display enhanced affinities for the R1 isoform
compared with R2 and R5 subtypes with various degrees of
selectivity. The only exception is 1c, which showed a moderate
R2 binding selectivity. However, a number of ligands (1e, 2b,
3b, 3j) showed a fairly good R2 binding affinity, even though
they were not selective. These data indicate that the insertion
of small groups with various steric and electronic properties on
the side chain of the benzene ring (both in the benzylamines
1,2 and in their indanyl closed-chain analogues 3) does not lead
to a selective binding to the R2 BzR.
competition experiments against the radiolabeled antagonist
[3H]flumazenil34 and expressed as Ki values only for compounds
inhibiting radioligand binding by more than 80% at a fixed
concentration of 10 µM (Table 1). The binding data of the
previously described indol-3-ylglyoxylamides 1a-h,262a-b,27
and 3a-b27 are also included in Table 1.
As mentioned above, it has been widely reported in the
literature how difficult it is to obtain affinity-based R2 selective
ligands due to subtle steric differences among BzR subtypes,15
whereas efficacy-based R2 selective agents have been more
easily identified.16,19,44 Thus, we assessed the efficacy profile
of compounds 1c, 2b, and 3b, displaying the highest potency
at the R2 BzR. Figure 2 shows the maximal efficacy on
coapplication with GABA of the tested compounds at the rat
recombinant R1 and R2 BzRs, using diazepam as the reference
standard. At the R1 subtype, all three compounds displayed an
efficacy lower than the one of diazepam (11%, 35%, 1%, and
71% for 1c, 2b, 3b, and diazepam, respectively). In particular,
the efficacy ranged from antagonism for 3b to partial agonism
for 2b and antagonism/partial agonism for 1c. At the R2 subtype,
1c and 2b showed an efficacy profile comparable to the one of
diazepam (85%, 79%, and 81% for compounds 1c, 2b, and
diazepam, respectively), whereas 3b behaved as a partial agonist
(48% and 81% for 3b and diazepam, respectively). These results
may predict anxiolytic nonsedative properties for the tested
compounds.
The affinity of compounds 3c-j for BzR subtypes was
evaluated by measuring their ability to displace [3H]flumazenil
in membranes from HEK293 cells expressing rat R1ꢀ2γ2, R2ꢀ2γ2,
and R5ꢀ3γ2 BzRs (Table 1).27 The same binding assays were
performed for compounds 1c-h,26 2a-b,27 and 3a-b27(Table
1).
The compounds displaying the highest potency at the R2 BzR
(1c, 2b, 3b) were tested for their functional efficacy by
measuring their modulatory effect on 36Cl- influx through the
ion channel pore at a GABA concentration, evoking 20% of
the maximum influx (EC20) in cloned HEK293 cells expressing
R1ꢀ2γ2 and R2ꢀ2γ2 BzRs, as previously described.29,35 A 100
times higher concentration than the Ki value of the tested
compound provided the maximal effect on the GABA-evoked
Cl- influx.36 Efficacy results are shown in Figure 2, in which
the nonselective full agonist diazepam was included as the
reference standard.
Furthermore, the anxiolytic activity of the selected compounds
1c, 2b, and 3b was first evaluated by a mouse light/dark (LD)
test, after ip treatment with the test compound (20 mg/kg), or
diazepam (1.25 mg/kg) used as control (Figure 3).
Molecular Modeling Studies. Docking simulations were
carried out using the AutoDock4 (AD4) software.37 Compounds
1c and 2b were docked into the models of the BzR constructed
by Cromer et al.38 and by Ernst et al.39 The binary complexes
between ligands and the receptor were further investigated by
molecular dynamics (MD) simulations using the AMBER 9.0
package software.40 Visual inspections of the calculated com-
plexes were attained using the MGL Tools41 and the UCSF
Chimera packages.42
Because of their promising efficacy profile, 1c, 2b, and 3b
were evaluated in vivo for their effects on mouse anxiety by
means of a light/dark (LD) apparatus, using diazepam as the
anxiolytic control.
The LD test is based on the mouse conflict between its innate
aversion to brightly illuminated areas and its spontaneous
exploratory behavior in response to mild stress, that is, novel
environment and light. When given a choice between a large
brightly lit compartment and a small dark compartment, a mouse
spontaneously prefers the dark one. The conflict rises between
the natural tendency to explore and the initial tendency to avoid
the unfamiliar (neophobia). In this view, drug-induced increase
in behaviors in the brightly lit open area is proposed to reflect
anxiolytic activity.45 The time spent in the lit area and the
number of transitions between the light and dark chambers have
been reported as the most reliable parameters to assess anxi-
olytic-like activity.46-49 Another important parameter involves
the evaluation of the total activity of the animals that could be
considered as an index of potential sedative action. Thus, it
seems obvious that animals that spend less time in one
compartment demonstrate few movements and vice versa. To
avoid this problem, results of movements/exploratory behavior
in a specific area were expressed as a function of the time spent
Results and Discussion
The binding data reported in Table 1 indicate that most of
the newly synthesized compounds 3c-j show appreciable
affinity for the BzR in bovine brain membranes, with the 5-NO2
substituted derivatives being more potent than the corresponding
5-H counterparts. Compound 3j, bearing 5-NO2 and 5′-OCH3
substituents, stands out as the most potent in the whole series
with a Ki value of 13 nM. The higher potency of 3j (Ki 13 nM)
compared with the parent compound 3b (Ki 28 nM) may be
due to an H-bond between the 5′-OMe group of 3j (in the
binding mode A) and an H-bond donor that we believe is present
on the surface of the S1 site.43 Within the 5-H indanyl

3726 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 12
Taliani et al.
Table 1. Inhibition of [3H]Flumazenil Specific Binding to Bovine Brain Membranes and to Rat R1ꢀ2γ2, R2ꢀ2γ2, and R5ꢀ3γ2 GABAA/Bz Receptor
Subtypesa of Indol-3-ylglyoxylylamide Derivatives 1-3
a The ability of the compounds to displace [3H]flumazenil was measured in membranes from HEK293 cells expressing the R1ꢀ2γ2, R2ꢀ2γ2, and R5ꢀ3γ2
subtypes, as described in the Experimental Section. b Ki values are means ( SEM of three determinations carried out in triplicate. c Percentage inhibition
values of specific [3H]flumazenil binding at 10 µM concentration are means ( SEM of three determinations carried out in triplicate. d Data taken from ref
26. e Data taken from ref 29. f Not determined. g Data taken from ref 27.
in the compartment under consideration.45 Anxiolytic agents
selectively increased exploration rather than general activity.
In the present investigation, the time spent in the white
compartment (TW), the number of transfers from the lit to the
dark area and vice versa (Tran) were measured after the mouse
was treated ip with the test compounds 1c, 2b, and 3b (20 mg/
kg) or diazepam (1.25 mg/kg) used as control (Figure 3a,b).
To evaluate the total activity parameter, the number of
exploratory rearings in the white section and the number of line
crossings (total activity, AW, data not shown) were also
measured and expressed as a function of TW (AW/TW, Figure
3c).
Administration of the reference diazepam at a dose of 1.25
mg/kg produces anxiolytic-like effects, as evidenced by the
significant increase of TW and Tran parameters with respect to
vehicle treated group (Figure 3a,b). These results are in
agreement with the literature.46 This anxiolytic effect is devoid

Anxiolytic/NonsedatiVe Agents among Indol-3-ylglyoxylamides
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 12 3727
Figure 2. Maximal efficacy of compounds 1c, 2b, and 3b, and
diazepam on rat recombinant R1ꢀ2γ2 and R2ꢀ2γ2 GABAA receptor
subtypes. Results are expressed as percentage of increase in the response
to GABAA at EC20 (mean ( SEM) from at least five independent
experiments.
of sedative component, as demonstrated by the AW/TW
parameter that is even greater than the ones of the control group
(Figure 3c).
Administration of compounds 1c and 2b in mice induced an
increase in the TW and Tran values compared to vehicle-treated
mice (Figure 3a,b); however, the AW/TW parameter for both
compounds and control are comparable (Figure 3c). On the basis
of these in vivo data, 1c and 2b may be novel anxiolytic agents
lacking sedative activity.
In spite of its in vitro efficacy at the R2 BzR, 3b displayed
no significant differences in vehicle-treated animals for all
parameters measured, evidencing no appreciable in vivo activity
(Figure 3). In this respect, the in vivo effects of pharmacologi-
cally active substances may diverge from those expected on
the basis of in vitro experiments due to unfavorable pharma-
cokinetic properties (for example, 3b may be metabolized to a
different extent than 1c and 2b).
Molecular Modeling. Recently, Cook et al.50 have oriented
the previously proposed pharmacophore model12,51 into the
binding site of the R1ꢀ2γ2 BzR by considering the most recent
advances in ligand structure-affinity relationships and structural
studies of the receptor binding site. Taking into account this
unified pharmacophore/receptor model, we undertook a molec-
ular modeling study on our compounds in order to explain their
structure-affinity relationships.
Among the available models of R1ꢀ2γ2 BzR in its closed-
channel conformation (Figure 4), we considered those kindly
provided by Cromer38 and Ernst.39 A detailed comparison of
the two reveals that they share a common general folding,
although, as expected, some differences can be seen in the
orientation of the loops. Both these BzR models, most precisely
the ligand binding domains of the R and γ subunits, were used
to dock compounds 1c and 2b that were the most interesting
from a pharmacological point of view. Results of docking
simulations were analyzed by taking into account the predicted
binding free energy (∆GAD4) associated to each docked solution,
together with the consistency with ligand structure-affinity
relationships data and results of receptor site-directed mutagen-
esis studies.
Figure 3. Effects of compounds 1c, 2b, and 3b (20 mg/kg ip) on
behavioral parameters in the LD test in mice (n ) 4-6) in comparison
with diazepam (1.25 mg/kg ip) or with vehicle (1% carboxymethyl
cellulose). Each value represents mean ( SEM. Statistical significance
was calculated for each substance versus the control using a two-tailed
unpaired t-test. Mice that failed to move from the white to the black
compartment were eliminated from the final analysis. * p < 0.05.
ligand was unable to efficiently contact both subunits at the same
time were discarded.
All the remaining solutions found for 2b using Cromer’s
model belong to the two lowest energy clusters (the predicted
binding free energies are reported in Table 4 in Supporting
Information) and differ for the relative positions of the ligand
aromatic rings. In the first cluster (mode A) 2b projects its indole
ring toward the γ subunit in the putative LDi site (Y160R, T142γ,
L140γ, M130γ, and R132γ)50 and the phenyl ring is closer to
the R subunit lying in the putative L2 site (interface between
loop C and A of the R subunit).50 In the second cluster (mode
B), the ligand aromatic rings are swapped (for any details of
ligand-receptor interactions see Figure 1 in the Supporting
Information). Interestingly, the existence of two possible binding
modes for the indole derivatives had already been postulated
in our previous studies based exclusively on their structure-
affinity relationships.28
In particular, because experimental data have outlined that
both the R and γ subunits are required for recognition and
binding of BzR ligands,52,53 all the docked poses in which the

3728 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 12
Taliani et al.
segments. However, such an artifactual plasticity did not affect
the conformation of the ligand binding site, as demonstrated
by the low B-factor values calculated for the residues in this
portion (see Figures 8 and 9 in Supporting Information). Thus,
we are confident that considering the sole extracellular ligand
binding domain does not significantly hamper the correctness
of molecular modeling studies aimed at the description of a
ligand binding mode in BzR.
For the 2b/BzR complex in the binding mode A, the
simulation demonstrated that the ligand is stabilized in a
conformation similar to the starting one (Figure 5a and
Supporting Information). In the LDi site, the ligand 5-NO2 group
forms an H-bond with the Y210R side chain (Figure 6a) that
was found to be in direct interaction with other BzR ligands
according to covalent labeling data.54 Furthermore, during the
MD simulations, an additional T-shaped interaction is estab-
lished by the indole ring with Y160R that is reinforced by the
electron-withdrawing NO2 group and by the formation of an
H-bond between the ligand indole NH and the Y160R side chain
OH.
Our 2b/BzR interaction model is consistent with the phar-
macophoric role of the indole NH as H-bond donor because its
methylation as well as its replacement by an oxygen or a sulfur
atom diminishes or abolishes the potency.28,55 Moreover, the
inspection of the energy-minimized average structure of the
complex reveals the presence of an H-bond between the carbonyl
oxygen attached to the indole ring and S206R. In such a
conformation, the ligand side chain methyl group takes favorable
hydrophobic contacts with V203R, V213R, and H102R. Ad-
ditionally, the aromatic ring of H102R is involved in a well-
oriented T-shaped interaction with the ligand phenyl ring.
According to structure-affinity relationships in the benzyl-3-
indolylglyoxylamides, affinity is enhanced when the pendant
phenyl ring is substituted at the 3′ and 4′ positions with hydroxy/
methoxy groups (for instance, compare 3j with 3b). In this
respect, a visual inspection of pose A calculated for 2b reveals
that S205R (loop C) could indeed donate an H-bond to the above
cited groups when they are present. The same is true for K256R.
Moreover, it has been demonstrated that only the (R)-enantiomer
of 2b is endowed with affinity for BzR.27 Indeed, the (S)-
enantiomer of this ligand would hardly adopt pose A due to
steric clashes between its phenyl ring and the R receptor subunit.
MD simulations for mode B have shown certain instability
in the binding conformation calculated by AD4. In fact, during
the production run, 2b is displaced from its initial position and
after 8 ns it adopts a stable orientation that differs substantially
from the initial one (Figures 5b and 6b). Interestingly, the frames
calculated in the last 3 ns of the simulations have the lowest
potential energy values recorded during the production run.
However, in this new binding pose, some inconsistencies arise
from the position of the ligand phenyl ring that is now distant
from the LDi and that makes weak van der Waals contacts with
M81γ and L140γ side chains in a rather shallow solvent exposed
receptor region. Indeed, structure-affinity relationships data
indicate that affinity is retained by replacing the phenyl ring
with alkyl chains of suitable size and length, such as i-propyl
or n-butyl but not longer or branched such as n-pentyl or
t-butyl.29 Likewise, replacement of the same phenyl moiety with
different 5- or 6-membered aromatic rings allows retention of
the affinity.43 These data would suggest that the cleft hosting
the benzyl moiety should be mainly hydrophobic and have a
limited room. As mentioned before, while in mode A, the
terminal phenyl of 2b is placed in the lipophilic L2 site, in mode
B, the same ring is hosted in a rather shallow and large pocket
Figure 4. Side view of the used BzR model. The putative ligand
binding cavity is depicted as a surface while the rest of the protein as
ribbons. The R and γ subunits are colored as orange and green,
respectively.
However, when the docking of 2b was performed in the
structural model of a R1 BzR by Ernst and co-workers,39 the
ligand was unable to take direct contact with both the R and γ
subunits at the same time. Particularly, in the great majority of
the calculated binding poses, the ligand is closer to the R subunit.
Such a behavior is mainly ascribable to the presence of Q204R
that extends its bulky side chain toward the inner part of the
binding site preventing the ligand to span from the L2 to the
LDi site. Similar considerations could also be done when Ernst’s
model was used to dock compound 1c. Thus, Cromer’s model
was selected for further ligand-receptor model refinements.
Similarly to 2b, when 1c was docked in Cromer’s model, a
double binding pose was still found with the indole ring pointing
toward the γ subunit (in the LDi site, mode A) or toward the R
one (in the L2 pocket, mode B) (Figure 2 in Supporting
Information).
The binary complexes calculated by the docking program for
1c and 2b, in both modes A and B, were then subjected to a 10
ns MD simulations to refine the predicted binding geometries.
Most importantly, given the dichotomy of binding modes
(A and B) predicted for these ligands, results of MD simulations
could suggest a preferential binding orientation for each ligand.
Because in the present study only the extracellular portion
of the BzR ligand binding domain was considered, a preliminary
analysis of MD results was attained to probe whether the
absence of the other BzR subunits (two ꢀ and one R) might
have influenced the overall binding site architecture. To identify
the relative internal motions of the protein residues, the isotropic
temperature (B) factor for each residue in the simulations was
calculated (see Figures 8 and 9 in Supporting Information). As
expected, the truncation of the model led to a high degree of
flexibility of the regions in direct contact with the missing

Anxiolytic/NonsedatiVe Agents among Indol-3-ylglyoxylamides
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 12 3729
Figure 5. Binding conformations of 2b (a and b) and 1c (c and d) in pose A (a and c) and B (b and d) into the BzR cleft. The receptor is
represented as green (γ subunit) and orange (R subunit) ribbons. Ligands in their docked conformations are represented in purple sticks, while
ligands in conformations calculated through MD simulations are represented as white sticks.
Figure 6. Energy-minimized average structures of the 2b/BzR complex
in binding modes A (a) and B (b). The ligand is depicted as white
sticks, the R subunit as orange sticks and ribbons, and the γ one is
represented as green sticks and ribbons. For clarity, only interacting
residues are shown.
Figure 7. Energy-minimized average structures of the 1c/BzR complex
in binding modes A (a) and B (b). The ligand is depicted as white
sticks, the R subunit as orange sticks and ribbons, and the γ one is
represented as green sticks and ribbons. For clarity, only interacting
residues are shown.
that can possibly host bulky substituents. From these consid-
erations, mode A for 2b, rather than B, seems to be more
consistent with the above outlined experimental data.
Analysis of the MD trajectories for the 1c/BzR complex in
the binding mode A revealed that the ligand fluctuates among
different orientations reaching after 6 ns a stable pose that is
significantly different from the initial one (Figures 5c and 7a).
Particularly, the ligand phenyl ring is displaced from the L2 site
in the R subunit to make weak contacts with loop F of the γ
subunit (Figure 7a and Supporting Information).
During the 10 ns MD simulation of 1c in the binding mode
B, the ligand remains stably adapted in the receptor cavity (rmsd
) 1.5 Å), contacting both R and γ subunits and having strong
interactions (Figure 5d and Supporting Information). Specifi-
cally, the indole moiety is engaged in a π-π interaction with
the Y160R aromatic ring, the p-fluorophenyl ring favorably
contacts M132γ, Y141γ, P127R, and A161R residues and engages
a T-shaped interaction with the Y160R side chain, and the amide
NH donates an H-bond to the T142γ backbone CO (Figure 7b).
Interestingly, for compounds of the 5-H indoles series, the
replacement of the terminal benzyl moiety with various alkyl
chains always results in a loss of potency.29 This suggests the
involvement of the pendant phenyl ring of these ligands in strong
interactions with the receptor. Thus, pose B rather than A would
be more consistent with structure-activity relationships in the
5-H indole series.
Interestingly, rescoring the minimized average complexes
found in the MD simulations for 2b/BzR and 1c/BzR in the
two binding solutions (A and B, respectively) with a new
empirical scoring function implemented in AD4,37 we have

3730 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 12
Taliani et al.
General Procedure for the Synthesis of N-(Indan-1-yl)-(5-
substituted indol-3-yl)glyoxylamide Derivatives 3c-j. Triethy-
lamine (3.0 mmol) was added dropwise to a stirred suspension,
cooled at 0 °C, of indolglyoxylyl chlorides 6 and 7 (2.5 mmol)
and the appropriate amine 8-11 (2.75 mmol) in 50 mL of dry
toluene. The reaction mixture was left to warm to room temperature,
stirred for 24-48 h (TLC analysis), and then filtered. The
precipitated collected was triturated with a saturated NaHCO3
aqueous solution, washed with water, and collected again to give
a first portion of crude product. The toluene solution was evaporated
to dryness, and the residue was treated with saturated NaHCO3
aqueous solution, washed with water, and collected to yield an
additional amount of crude product. The quantities of amide
derivatives obtained from the initial insoluble precipitate or from
the toluene solution were variable depending upon the solubility
of the various compounds. All products 3c-j were purified by
chromatography (chloroform:methanol ) 9:1 v/v as eluant) and
recrystallization from the appropriate solvent whenever a solid was
obtained. Yields, recrystallization solvents, melting points, and
spectral data are reported in the Supporting Information.
found that for 2b pose A is energetically favored (∆G ) -7.0
kcal/mol) over pose B (∆G ) -4.6 kcal/mol) (Table 5 in
Supporting Information). On the other hand, for compound 1c,
the AD4 rescoring predicted a ∆G value more favorable for
pose B than for pose A (-8.5 and -5.1 kcal/mol, respectively).
Analysis of the energy components of the four ∆G values
reveals that the differences are mainly related to van der Waals
contacts (corresponding to hydrophobic interactions) that are
maximized in pose A for 2b and in pose B for 1c (Table 4 in
Supporting Information).
Altogether, these results suggest that compound 2b binds to
the R/γ interface of the BzR adopting pose A, while 1c
preferentially adopts pose B in agreement with our previously
reported pharmacophore/topological model.27
As reported in the present paper, 2b and 1c display a different
efficacy profile at the BzR, with the first being a partial agonist
and the second acting mostly as an antagonist. Nevertheless,
we have not attempted to relate the different efficacies of these
two compounds with their different putative binding modes to
the BzR. So far, the molecular underpinnings responsible for
the efficacy behaviors of indol-3-ylglyoxylamides, as well as
of other classes of ligands of the BzR subtypes, still remain to
be determined.
5-Nitro-1-indanone 13. A solution of 5-amino-1-indanone 12
(1.5 g, 10.2 mmol) in 10 mL of dichloromethane was treated at 0
°C with 1.75 g (10.2 mmol) of 51% m-CPBA (meta-chloroper-
oxybenzoic acid), and the mixture was stirred at the same
temperature for 48 h. The organic solution was washed with a
saturated NaHCO3 aqueous solution, dried over Na2SO4, and
evaporated to dryness. The obtained residue was purified by
chromatography (dichloromethane as eluant) to obtain 13 as a
yellow solid: yield 66%; mp 132-133 °C (lit.56 131-132.5 °C
Conclusions
Anxioselective agents may be identified either among com-
pounds binding selectively to the R2 BzR subtype that behave
as agonists or among compounds binding with comparable
potency to various BzR subtypes that elicit agonism at the R2
subtype and antagonism at other subtypes. Prompted by the
success of the second approach, a number of indol-3-ylglyoxy-
lamides belonging to our in-house collection were screened for
their binding affinity to the rat recombinant R1, R2, and R5 BzRs.
Among the compounds with highest potency at the R2 subtype,
compounds 1c and 2b exhibited interesting properties either in
vitro (full R2 agonism and R1 partial agonism/antagonism) or
in vivo (anxiolytic/nonsedative activity in mice). Molecular
docking calculations using a model of the BzR ligand binding
domain in the closed state indicated the presence of two possible
binding positions for 2b and 1c in which the ligand aromatic
rings (indole and phenyl) can be alternatively lodged in
the receptor L2 and LDI sites. Analysis of the stability of the
predicted complex through MD simulations suggested that the
presence of the 5-nitro group in 2b would allow the formation
of more productive interactions when the indole ring is
embedded in the LDi receptor region. Conversely, compounds
devoid of such a substituent, as in 1c, should produce more
effective interactions when the indole ring is placed in the L2
site.
1
dec.). H NMR (400 MHz, CDCl3) δ 2.65-2.80 (m, 2H, 3-CH2),
3.12-3.22 (m, 2H, 2-CH2), 7.27-7.33 (m, 1H, H-7), 7.42-7.48
(m, 1H, H-4), 7.61-7.75 (m, 1H, H-6). Anal. (C9H7NO3) C, H, N.
5-Nitroindan-1-methoxyimine 14. 5-Nitro-1-indanone 13 (1.3
g, 7.3 mmol) and methoxyamine hydrochloride (1.5 g, 18.2 mmol)
were dissolved in a mixture of ethanol (15 mL) and pyridine (15
mL) and heated on a steam bath for 30 min. The solution was
allowed to return to room temperature and then was diluted with
water (30 mL) to precipitate practically pure 14, which was collected
1
by filtration: yield 84%; mp 174-176 °C. H NMR (400 MHz,
CDCl3) δ 2.88-3.21 (m, 4H, 3-CH2 and 2-CH2), 4.02 (s, 3H,
OCH3), 7.55-7.70 (m, 1H, H-7), 8.13-8.33 (m, 2H, H-4 and H-6).
Anal. (C10H10N2O3) C, H, N.
5-Nitro-1-indanamine 9. A solution of 5-nitroindan-1-meth-
oxyimine 14 (1.3 g, 6.1 mmol) in 8 mL of THF was treated with
1 M diborane in THF (38.5 mL) and refluxed for 4 h in an argon
atmosphere. Methanol (10 mL) was added and the solvent
evaporated. The residue was treated with 10% HCl and heated on
a steam bath for 30 min. The mixture was made alkaline with
saturated NaHCO3 aqueous solution and extracted with ethyl ether;
the organic phase was dried (Na2SO4) and evaporated. The obtained
residue was purified by chromatography (chloroform:methanol )
8:2 v/v as eluant), yielding the product 9 as a yellow solid: yield
35%; mp 279-281 °C. 1H NMR (400 MHz, CDCl3) δ 1.73-1.92
(m, 1H, H-2), 2.52-2.74 (m, 1H, H-2), 2.78-3.18 (m, 2H, 3-CH2),
4.33-4.48 (m, 1H, H-1), 7.43-7.51 (m, 1H, H-7), 8.04-8.18 (m,
2H, H-4 and H-6). Anal. (C9H10N2O2) C, H, N.
Experimental Section
Chemistry. Melting points were determined using a Bu¨chi
apparatus B 540 and are uncorrected. Routine nuclear magnetic
resonance spectra were recorded on a Varian Mercury 400
spectrometer operating at 400 MHz. EI-HRMS mass spectra were
obtained on a Finnigan MAT95XP spectrometer using a direct
injection probe and an electron beam energy of 70 eV. Evaporation
was performed in vacuo (rotary evaporator). Analytical TLC was
carried out on Merck 0.2 mm precoated silica gel aluminum sheets
(60 F-254). Silica gel 60 (230-400 mesh) was used for column
flash chromatography. Combustion analyses on target compounds
were performed by our Analytical Laboratory in Pisa. All com-
pounds showed g95% purity.
Biological Methods. Materials. [3H]Flumazenil (specific activity
70.8 Ci/mmol) was obtained from NEN Life Sciences Products.
All other chemicals were of reagent grade and were obtained from
commercial suppliers. HEK293 cells stably expressing rat GABAA
receptor subtypes (R1ꢀ2γ2, R2ꢀ2γ2, R5ꢀ3γ2) were kindly supplied
by Dr. Franc¸ois Besnard, Department of Molecular and Functional
Genomics, Synthe´labo, France.57
Radioligand Binding Studies. Bovine cerebral cortex mem-
branes were prepared in accordance with ref 58. The membrane
preparations were subjected to a freeze-thaw cycle, washed by
suspension and centrifugation in 50 mM tris-citrate buffer pH 7.4
(T1), and then used in the binding assay. Protein concentration was
assayed by the method of Lowry et al.59 [3H]Flumazenil binding
studies were performed as previously reported.27
Besides the commercially available starting materials, the
5-substituted indanyl-1-amines 8, 10, and 11 were prepared in
accordance with a reported method.33

Anxiolytic/NonsedatiVe Agents among Indol-3-ylglyoxylamides
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 12 3731
HEK293 cells stably expressing rat GABAA receptor subtypes
(R1ꢀ2γ2, R2ꢀ2γ2, R5ꢀ3γ2) were maintained, as previously described,57
in DMEM/Nut Mix F-12 with Glut-I (GIBCO), supplemented with
10% fetal bovine serum, L-glutamine (2 mM), penicillin (100 units/
mL), and streptomycin (100 µg/mL) in a humidified atmosphere
of 5% CO2/95% air at 37 °C. Cells were harvested and then
centrifuged at 500g. The crude membranes were prepared after
homogenization in 10 mM potassium phosphate, pH 7.4, and
differential centrifugation at 48000g for 30 min at 4 °C. The pellets
were washed twice in this manner before final resuspension in 10
mM potassium phosphate, pH 7.4, containing 100 mM potassium
chloride.57 [3H]Flumazenil binding assays to transfected cells
membranes were carried out as previously described.57 In brief,
the cell lines membranes were incubated in a volume of 500 µL,
which contained [3H]flumazenil at a concentration of 1-2 nM and
the test compound in the range 10-9-10-5 M. Nonspecific binding
was defined by 10-5 M diazepam. Assays were incubated to
equilibrium for 1 h at 4 °C.
fifths painted black illuminated under a red light (four red bulbs,
each 15 W, 10 lx total) and the remainder of the box painted white
and brightly illuminated with a 60 W (400 lx) light source. The
compartments were separated by a divider and connected by an
opening 7.5 cm × 7.5 cm located at the floor level in the center of
the partition. Each mouse was tested by placing it in the center of
the lit chamber, facing away from the dark one. The total duration
of the test was 5 min. Each test was recorded on videotape, and
the behavioral analysis was performed after recording. The time
spent in the white compartment (TW) and the number of transfers
from lit to dark areas and vice versa were measured. To obtain the
total activity parameter, the number of exploratory rearings in the
white section and the number of line crossings were measured.
Results are given as the mean ( SEM. Statistical analysis was
performed using single-factor analysis of variance (ANOVA) and
were appropriately followed by the procedure (Fisher’s PLSD) for
comparing the treatments to each other. Data were analyzed using
the software StatView 5.0 Abacus Concept.
Functional Efficacy Studies. 36Cl- uptake was measured in
transfected HEK 293 cells stably expressing rat GABAA receptor
subtypes (R1ꢀ2γ2, R2ꢀ2γ2), essentially as previously described.29,35
In brief, coverlips were placed 8-10/cell culture plate (100 mm)
and were incubated overnight under germicidal light in a solution
containing 0.01 mg/mL polylysine in 0.1 M boric acid (pH 8.4).
The following day, coverlips were washed twice with phosphate
buffered saline and were placed into individual 6-well cell culture
plates. DMEM/Nut Mix F-12 with Glut-I (GIBCO), supplemented
with 10% fetal bovine serum, penicillin (100 units/mL), and
streptomycin (100 µg/mL) were added to each well. Cells were
then plated into each well at a density of 50 × 105 cells/well and
were grown on the coverlips for 3 days at 37 °C with 5% CO2.
Cells were washed twice (10 s/wash) in wash buffer 136 mM NaCl,
5.4 mM KCl, 1.4 mM MgCl2, 1.2 mM CaCl2, 1 mM NaH2PO4,
and 20 mM 4-(2-hydroxyethyl)-1-piperazin ethane sulfonic acid,
adjusted to a pH of 7.4 with Tris base. All steps were conducted at
room temperature. Cells were then dipped into 2 mL of a 36Cl-
solution (2 µCi/mL) containing the drug to be studied, with or
without GABA. Influx were terminated after 8 s by transfer
coverlips to 500 mL ice-cold assay buffer (with 0.1 mM picrotoxin).
Coverlips were drained rapidly and placed into 1.6 mL of 0.2 N
NaOH in 20 mL scintillation vials and were left overnight. A 0.1
mL aliquot was removed and assayed for protein determination.
The remaining 1.5 mL were neutralized with 0.3 mL of 1 N acetic
acid and 20 mL of BioSafe II were added for counting by liquid
scintillation spectrometry. Values for 36Cl- influx were expressed
as nanomol/mg protein.
Pharmacology. Mouse Light/Dark Box Test. The experiments
were carried out in accordance with the Animal Protection Law of
the Republic of Italy, DL no. 116/1992, on the basis of the European
Communities council Directive of 24 November 1986 (86/609/
EEC). All efforts were made to minimize animal suffering and to
reduce the number of animals involved (n ) 4-6). Male CD-1
albino mice (35-40 g) (Harlan, Italy) were used. Ten mice were
housed in cages under 12 h reversed lighting conditions (lights off
at 2:00 and on at 14:00) in rooms held at constant temperature
(22 ( 2 °C) and fed ad libitum on standard laboratory chow and
tap water. After transfer from the breeding unit to the holding rooms,
mice were allowed to acclimatize during 4 weeks to the reversed
light cycle before the beginning of the experiments. Mice were only
used once.
Computational Methods. Docking Simulations. The new
version of the docking program AutoDock (version 4, AD4),37 as
implemented through the graphical user interface called AutoDock-
Tools (ADT), was used to dock both 1c and 2b derivatives. This
new release implements a new force field (FF) that, using an
improved thermodynamic model of the binding process, allows
inclusion of intramolecular terms in the estimated free energy.
Moreover, the new FF includes a full desolvation model that
contains terms for all atom types, including the favorable energetics
of desolvating carbon atoms as well as the unfavorable energetics
of desolvating polar and charged atoms. This FF also incorporates
an improved model of directionality in hydrogen bonds, now
predicting the proper alignment of groups with multiple hydrogen
bonds. The structures of compounds 1c and 2b were built by using
the Builder module in Insight2000.1 (Accelrys, San Diego, CA).60
Atomic potentials and charges were assigned using the cvff force
field.61 Built conformations were geometrically optimized (Discover
module, Insight2000.1) using a distance-dependent dielectric con-
stant mimicking an aqueous environment (ε ) 80r). Energy
minimizations were performed using the conjugate gradient62 as
the minimization algorithm until the maximum rms derivative was
less than 0.001 kcal mol-1 Å-1.
BzR structures that were kindly provided Cromer et al.38 and
by Ernst et al.39 were energy minimized using Amber 9.0 software
package.40 Then the constructed compounds and BzR structure were
converted to AD4 format files using ADT generating automatically
all other atom values. The docking area was assigned visually
around the ligands binding site. A grid of 60 Å × 60 Å × 60 Å
with 0.375 Å spacing was calculated around the docking area for
the ligand atom types using AutoGrid4. For each ligand, 100
separate docking calculations were performed. Each docking
calculation consisted of 10 million energy evaluations using the
Lamarckian genetic algorithm local search (GALS) method. The
GALS method evaluates a population of possible docking solutions
and propagates the most successful individuals from each generation
into the subsequent generation of possible solutions. A low-
frequency local search according to the method of Solis and Wets
is applied to docking trials to ensure that the final solution represents
a local minimum. All dockings described in this paper were
performed with a population size of 250, and 300 rounds of Solis
and Wets local search were applied with a probability of 0.06. A
mutation rate of 0.02 and a crossover rate of 0.8 were used to
generate new docking trials for subsequent generations, and the
best individual from each generation was propagated over the next
generation. The docking results from each of the 100 calculations
were clustered on the basis of root-mean-square deviation (rmsd)
(solutions differing by less than 2.0 Å) between the Cartesian
coordinates of the atoms and were ranked on the basis of free energy
of binding (∆GAD4). The top-ranked compounds were visually
inspected for good chemical geometry.
This test exploited the conflict between the animal’s tendency
to explore a new environment and its fear of bright light. Test for
a change in anxiety were conducted essentially as described by
Costall B. et al.47 between 10.00-13.00 in a quiet darkened room
illuminated by red light. The mice were taken from the dark holding
rooms in a dark container to the experimental room. After 30 min,
mice received ip either test compounds 20 mg/kg, or diazepam
(Sigma) 1.25 mg/kg, or a vehicle (1% carboxymethyl cellulose).
After a 1 h period of adaptation to the new environment, mice were
placed into the test box. The open-topped test box (45cm × 27 cm
× 27 cm high) had the floor area lined into 9 cm squares: two-
Molecular Dynamics and Energy Minimization Simula-
tions. The binary complexes between BzR and 2b and1c in both
binding mode A and B were investigated by means of MD

3732 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 12
Taliani et al.
simulations carried out with the AMBER 9.0 package software.40
Ligand force-field parameters were derived using the Antecham-
ber program,63 and partial charges for the substrates were derived
using the AM1-BCC procedure in Antechamber. Eleven ns MD
simulations on the aforementioned four complexes were carried
out in explicit solvent and periodic boundary conditions. One
Cl- counterion was added to the solvent bulk of the protein-water
complexes to maintain neutrality in the systems. First, water
shells and counterion were minimized using steepest descent and
conjugate gradient algorithms. Then a minimization of the entire
ensemble was performed setting a convergence criterion on the
gradient of 0.001 kcal mol-1 Å-1. Equilibration runs were carried
out by heating the system to 300 K in 1 ns. This was followed
by 10 ns MD simulations in the NPT ensemble (constant
temperature and pressure). The parm99 version64 of the all-atom
Amber force field65 was used for the protein and the counterion,
whereas the TIP3P model66 was employed to explicitly represent
water molecules. van der Waals and short-range electrostatic
interactions were estimated within a 10 Å cutoff, whereas the
long-range electrostatic interactions were assessed by using the
particle mesh Ewald (PME) method,67 with a 1 Å charge grid
spacing interpolated by a fourth-order B-spline, and by setting
the direct sum tolerance to 10-5. Bonds involving hydrogen
atoms were constrained by using the SHAKE algorithm68 with
a relative geometric tolerance for coordinate resetting of 0.00001
Å. Berendsen’s coupling algorithms69 were employed to maintain
constant temperature and pressure with the same scaling factor
for both solvent and solutes and with the time constant for heat
bath coupling maintained at 1.5 ps. The pressure for the
isothermal-isobaric ensemble was regulated by using a pressure
relaxation time of 1 ps in Berendsen’s algorithm. The simulations
of the solvated protein models were performed using a constant
pressure of 1 atm and a constant temperature of 300 K. Analysis
of MD trajectories was attained using the ptraj software.70 This
software was also used to calculate the average structures of
the four complexes that were energy minimized by employing
the same geometrical optimization protocol mentioned above.
All calculations were performed on a Linux machine employing
a Fedora Core 7 architecture.
(9) Whiting, P. J. GABA-A Receptor Subtypes in the Brain: A Paradigm
for CNS Drug Discovery. Drug DiscoVery Today 2003, 8, 445–450.
(10) Da Settimo, F.; Taliani, S.; Trincavelli, M. L.; Montali, M.; Martini,
C. GABAA/Bz Receptor Subtypes as Targets for Selective Drugs. Curr.
Med. Chem. 2007, 14, 2680–2701.
(11) Atack, J. R. The Benzodiazepine Binding Site of GABAA Receptors
as a Target for the Development of Novel Anxiolytics. Expert. Opin.
InVest. Drugs 2005, 14, 601–618.
(12) Zhang, W.; Koehler, K. F.; Zhang, P.; Cook, J. M. Development of a
Comprehensive Pharmacophore Model for the Benzodiazepine Recep-
tor. Drug Des. DiscoVery 1995, 12, 193–248.
(13) Huang, Q.; He, X.; Ma, C.; Liu, R.; Yu, S.; Dayer, C. A.; Wenger,
G. R.; McKernan, R.; Cook, J. M. Pharmacophore/Receptor Models
for GABAA/BzR Subtypes (R1ꢀ3γ2, R5ꢀ3γ2, and R6ꢀ3γ2) via a
Comprehensive Ligand-Mapping Approach. J. Med. Chem. 2000, 43,
71–95.
(14) He, X.; Huang, Q.; Ma, C.; Yu, S.; McKernan, R.; Cook, J. M.
Pharmacophore/Receptor Models for GABAA/BzR R2ꢀ3γ2, R3ꢀ3γ2, and
R4ꢀ3γ2 Recombinant Subtypes. Included Volume Analysis and Com-
parison to R1ꢀ3γ2, R5ꢀ3γ2, and R6ꢀ3γ2. Drug Des. DiscoVery 2000,
17, 131–171.
(15) Yu, S.; He, X.; Ma, C.; McKernan, R.; Cook, J. M. Studies in Search
of R2 Selective Ligands for GABAA/BzR Receptor Subtypes. Part I.
Evidence for the Conservation of Pharmacophoric Descriptors for DS
Subtypes. Med. Chem. Res 1999, 9, 186–202.
(16) Collins, I.; Moyes, C.; Davey, W. B.; Rowley, M.; Bromidge, F. A.;
Quirk, K.; Atack, J. R.; McKernan, R. M.; Thompson, S. A.; Wafford,
K.; Dawson, G. R.; Pike, A.; Sohal, B.; Tsou, N. N.; Ball, R. G.;
Castro, J. L. 3-Heteroaryl-2-pyridones: Benzodiazepine Site Ligands
with Functional Selectivity for R2/R3-Subtypes of Human GABAA
Receptor-Ion Channels. J. Med. Chem. 2002, 45, 1887–1900.
(17) Loughhead, D. G.; Novakovic, S.; O’Yang, C.; Putman, D. G.; Soth,
M. Preparation of Pyrazolopyridines and Other Heterocycle-fused
Pyrazoles as GABAA R2 Subtype Selective Receptor Modulators. PCT
Int. Appl. WO 2005077363 A1, 2005.
(18) Atack, J. R.; Wafford, K. A.; Tye, S. J.; Cook, S. M.; Sohal, B.; Pike,
A.; Sur, C.; Melillo, D.; Bristow, L.; Bromidge, F.; Ragan, I.; Kerby,
J.; Street, L.; Carling, R.; Castro, J. L.; Whiting, P.; Dawson, G. R.;
McKernan, R. M. TPA023 [7-(1,1-Dimethylethyl)-6-(2-ethyl-2H-1,2,4-
triazol-3-ylmethoxy)-3-(2-fluorophenyl)-1,2,4-triazolo[4,3-b]py-
ridazine], an Agonist Selective for R2- and R3-containing GABAA
Receptors, Is a Nonsedating Anxiolytic in Rodents and Primates.
J. Pharmacol. Exp. Ther. 2006, 316, 410–422.
(19) Griebel, G.; Perrault, G.; Simiand, J.; Cohen, C.; Granger, P.; Decobert,
M.; Franc¸on, D.; Avenet, P.; Depoortere, H.; Tan, S.; Oblin, A.;
Schoemaker, H.; Evanno, Y.; Sevrin, M.; George, P.; Scatton, B.
SL651498: An Anxioselective Compound with Functional Selectivity
for R2- and R3-containing γ-aminobutyric acidA (GABAA) Receptors.
J. Pharmacol. Exp. Ther. 2001, 298, 753–768.
(20) Licata, S. C.; Platt, D. M.; Cook, J. M.; Sarma, P. V.; Griebel, G.;
Rowlett, J. K. Contribution of GABAA Receptor Subtypes to the
Anxiolytic-Like, Motor, and Discriminative Stimulus Effects of
Benzodiazepines: Studies with the Functionally Selective Ligand
SL651498 [6-Fluoro-9-methyl-2-phenyl-4-(pyrrolidin-1-yl-carbonyl)-
2,9-dihydro-1H-pyridol[3,4-b]indol-1-one]. J. Pharmacol. Exp. Ther.
2005, 313, 1118–1125.
Acknowledgment. We are grateful to Prof. Cromer and Prof.
Ernst for kindly providing the homology models of the R1ꢀ2γ2
GABAA receptor. Furthermore, we thank Dr. Franc¸ois Besnard
for his generous gift of HEK293 cells stably expressing rat
GABAA receptor subtypes (R1ꢀ2γ2, R2ꢀ2γ2, R5ꢀ3γ2).
Supporting Information Available: Tables including physical
properties, spectral and analytical data of compounds
3c-j,Tables reporting the ∆GAD4 calculated for the ligand/BzR
complexes, together with additional molecular modeling pictures.
This material is available free of charge via the Internet at http://
pubs.acs.org.
(21) Crawforth, J.; Atack, J. R.; Cook, S. M.; Gibson, K. R.; Nadin, A.;
Owens, A. P.; Pike, A.; Rowley, M.; Smith, A. J.; Sohal, B.; Sternfeld,
F.; Wafford, K.; Street, L. J. Tricyclic Pyridones as Functionally
Selective Human GABAAR2/3 Receptor-Ion Channel Ligands. Bioorg.
Med. Chem. Lett. 2004, 14, 1679–1682.
(22) Albaugh, P. A.; Marshall, L.; Gregory, J.; White, G.; Hutchinson, A.;
Ross, P. C.; Gallagher, D. W.; Tallman, J. F.; Crago, M.; Cassella,
J. V. Synthesis and Biological Evaluation of 7,8,9,10-Tetrahydroimi-
dazo[1,2-c]pyrido[3,4-e]pyrimidin-5(6H)-ones as Functionally Selec-
tive Ligands of the Benzodiazepine Receptor Site on the GABAA
Receptor. J. Med. Chem. 2002, 45, 5043–5051.
(23) Russell, M. G. N.; Carling, R. W.; Atack, J. R.; Bromidge, F. A.;
Cook, S. M.; Hunt, P.; Isted, C.; Lucas, M.; McKernan, R. M.;
Mitchinson, A.; Moore, K. W.; Narquizian, R.; Macaulay, A. J.;
Thomas, D.; Thompson, S. A.; Wafford, K. A.; Castro, J. L. Discovery
of Functionally Selective 7,8,9,10-Tetrahydro-7,10-ethano-1,2,4-tria-
zolo[3,4-a]phthalazines as GABAA Receptor Agonists at the R3
Subunit. J. Med. Chem. 2005, 48, 1367–1383.
(24) Blackaby, W. P.; Atack, J. R.; Bromidge, F.; Lewis, R.; Russel,
M. G. N.; Smith, A.; Wafford, K.; McKernan, R. M.; Street, L. J.;
Castro, J. L. Pyrazolopyridinones as Functionally Selective GABAA
Ligands. Bioorg. Med. Chem. Lett. 2005, 15, 4998–5002.
(25) Goodacre, S. C.; Street, L. J.; Hallett, D. J.; Crawforth, J. M.; Kelly,
S.; Owens, A. P.; Blackaby, W. P.; Lewis, R. T.; Stanley, J.; Smith,
A. J.; Ferris, P.; Sohal, B.; Cook, S. M.; Pike, A.; Brown, N.; Wafford,
K. A.; Marshall, G.; Castro, J. L.; Atack, J. R. Imidazo[1,2-a]-
References
(1) Sigel, E. Mapping of the Benzodiazepine Recognition Site on GABAA
Receptors. Curr. Top. Med. Chem. 2002, 2, 833–839.
(2) Teuber, L.; Watjens, F.; Jensen, L. H. Ligands for the Benzodiazepine
Binding Site-A Survey. Curr. Pharm. Des. 1999, 5, 317–343.
(3) Gardner, C. R. Interpretation of the Behavioral Effects of Benzodi-
azepine Receptor Ligands. Drugs Future 1989, 14, 51–67.
(4) Sieghart, W. Structure and Pharmacology of γ-Aminobutyric AcidA
Receptor Subtypes. Pharmacol. ReV. 1995, 47, 181–234.
(5) Chebib, M.; Johnston, G. A. R. GABA-Activated Ligand Gated Ion
Channels: Medicinal Chemistry and Molecular Biology. J. Med. Chem.
2000, 43, 1427–1447.
(6) Mehta, A. K.; Ticku, M. K. An Update on GABA-A Receptors. Brain
Res. ReV. 1999, 29, 196–217.
(7) Sieghart, W.; Sperk, G. Subunit Composition, Distribution and
Function of GABA(A) Receptor Subtypes. Curr. Top. Med. Chem.
2002, 2, 795–816.
(8) Mohler, H.; Fritschy, J. M.; Rudolph, U. A New Benzodiazepine
Pharmacology. J. Pharmacol. Exp. Ther. 2002, 300, 2–8.

Anxiolytic/NonsedatiVe Agents among Indol-3-ylglyoxylamides
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 12 3733
pyrimidines as Functionally Selective and Orally Bioavailable GABAA
R2/R3 Binding Site Agonists for the Treatment of Anxiety Disorders.
J. Med. Chem. 2006, 49, 35–38.
Visualization System for Exploratory Research and Analysis. J. Com-
put. Chem. 2004, 25, 1605–1612.
(43) Primofiore, G.; Da Settimo, F.; Marini, A. M.; Taliani, S.; La Motta,
C.; Simorini, F.; Novellino, E.; Greco, G.; Cosimelli, B.; Ehlardo,
M.; Sala, A.; Besnard, F.; Montali, M.; Martini, C. Refinement of the
Benzodiazepine Receptor Site Topology by Structure-Activity Re-
lationships of New N-(Heteroarylmethyl)indol-3-ylglyoxylamides.
J. Med. Chem. 2006, 49, 2489–2495.
(44) McKernan, R. M.; Rosahl, T. W.; Reynolds, D. S.; Sur, C.; Wafford,
K. A.; Atack, J. R.; Farrar, S.; Myers, J.; Cook, G.; Ferris, P.; Garrett,
L.; Bristow, L.; Marshall, G.; Macaulay, A.; Brown, N.; Howell, O.;
Moore, K. W.; Carling, R. W.; Street, L. J.; Castro, J. L.; Ragan, C. I.;
Dawson, G. R.; Whi, P. J. Sedative but not Anxiolytic Properties of
Benzodiazepines are Mediated by the GABAA Receptor R1 Subtype.
Nat. Neurosci. 2000, 3, 587–592.
(26) Da Settimo, A.; Primofiore, G.; Da Settimo, F.; Marini, A. M.;
Novellino, E.; Greco, G.; Martini, C.; Giannaccini, G.; Lucacchini,
A. Synthesis, Structure-Activity Relationships, and Molecular Model-
ing Studies of N-(Indol-3-ylglyoxylyl)benzylamine Derivatives Acting
at the Benzodiazepine Receptor. J. Med. Chem. 1996, 39, 5083–5091.
(27) Primofiore, G.; Da Settimo, F.; Taliani, S.; Marini, A. M.; Novellino,
E.; Greco, G.; Lavecchia, A.; Besnard, F.; Trincavelli, L.; Costa, B.;
Martini, C. Novel N-(Arylalkyl)indol-3-ylglyoxylylamides Targeted
as Ligands of the Benzodiazepine Receptor: Synthesis, Biological
Evaluation, and Molecular Modeling Analysis of the Structure-
Activity Relationships. J. Med. Chem. 2001, 44, 2286–2297.
(28) Da Settimo, A.; Primofiore, G.; Da Settimo, F.; Marini, A. M.;
Novellino, E.; Greco, G.; Gesi, M.; Martini, C.; Giannaccini, G.;
Lucacchini, A. N′-Phenylindol-3-ylglyoxylohydrazide Derivatives:
Synthesis, Structure-Activity Relationships, Molecular Modeling
Studies, and Pharmacological Action on Brain Benzodiazepine Recep-
tors. J. Med. Chem. 1998, 41, 3821–3830.
(45) Bourin, M.; Hascoe¨t, M. The Mouse Light/Dark Box Test. Eur.
J. Pharmacol. 2003, 463, 55–65.
(46) Hascoe¨t, M.; Bourin, M. A New Approach to the Light/Dark Test
Procedure in Mice. Pharmacol., Biochem. BehaV. 1998, 60, 645–
653.
(47) Costall, B.; Jones, B. J.; Kelly, M. E.; Naylor, R. J.; Tomlins, D. M.
Exploration of Mice in a Black and White Test Box: Validation as
a Model of Anxiety. Pharmacol., Biochem. BehaV. 1989, 32, 777–
785.
(48) Kilfoil, T.; Michel, A.; Montgomery, D.; Whiting, R. L. Effects of
Anxiolytic and Anxiogenic Drugs on Exploratory Activity in a
Simple Model of Anxiety in Mice. Neuropharmacology 1989, 28,
901–905.
(29) Primofiore, G.; Taliani, S.; Da Settimo, F.; Marini, A. M.; La Motta,
C.; Simorini, F.; Patrizi, M. P.; Sergianni, V.; Novellino, E.; Greco,
G.; Cosimelli, B.; Calderone, V.; Montali, M.; Besnard, F.; Martini,
C. Novel N-Substituted Indol-3-ylglyoxylamides Probing the LDi and
L1/L2 Lipophilic Regions of the Benzodiazepine Receptor Site in
Search for Subtype-Selective Ligands. J. Med. Chem. 2007, 50, 1627–
1634.
(30) Da Settimo, A.; Primofiore, G.; Marini, A. M.; Ferrarini, P. L.;
Franzone, J. S.; Cirillo, L.; Reboani, M. C. N-(Indol-3ylglyoxylyl) -
methionine Derivatives: Preparation and Gastric Antisecretory Activity.
Eur. J. Med. Chem. 1988, 23, 21–24.
(49) Young, R.; Johnson, D. N. A Fully Automated Light/Dark Apparatus
Useful for Comparing Anxiolytic Agents. Pharmacol., Biochem.
BehaV. 1991, 40, 739–743.
(31) Soerensen, U. S.; Teuber, L.; Peters, D.; Stroebaek, D.; Johansen, T. H.;
Nielsen, K. S.; Christophersen, P. Preparation of Novel 2-Aminoben-
zimidazole Derivatives as Modulators of Small-Conductance Calcium-
Activated Potassium Channels. PCT Int. Appl. WO 2006013210 A2,
2006.
(32) Borne, R. F.; Forrester, M. L.; Waters, I. W. Conformational Analogues
of Antihypertensive Agents Related to Guanethidine. J. Med. Chem.
1977, 20, 771–776.
(33) Gomtsyan, A.; Bayburt, E. K.; Lee, C.-H.; Koenig, J. R.; Schmidt,
R.; Lukin, K.; Chambournier, G.; Hsu, M.; Leanna, M. R.; Cink, R. D.
Preparation of Fused Heterocyclic Compounds as Vanilloid Receptor
Subtype 1 (VR1) Inhibitors. PCT Int. Appl. WO 2004111009 A1,
2004.
(34) Giannaccini, G.; Giacomelli, M.; Martini, C.; Lucacchini, A.; Piccolino,
M. Binding of the Benzodiazepine Ligand [3H]Ro 15-1788 to
Membrane Preparations of the Rabbit and Turtle Retina. Comp.
Biochem. Physiol., Part C: Toxicol. Pharmacol. 1992, 101, 337–342.
(35) Harris, R. A.; Proctor, W. R.; McQuilkin, S. J.; Klein, R. L.; Mascia,
M. P.; Whatley, V.; Whiting, P. J.; Dunwiddie, T. V. Ethanol Increases
GABAA Responses in Cells Stably Transfected with Receptor Subunits.
Alcohol.: Clin. Exp. Res. 1995, 19, 226–232.
(36) Street, L. J.; Sternfeld, F.; Jelley, R. A.; Reeve, A. J.; Carling, R. W.;
Moore, K. W.; McKernan, R. M.; Sohal, B.; Cook, S.; Pike, A.;
Dawson, G. R.; Bromidge, F. A.; Wafford, K. A.; Seabrook, G. R.;
Thompson, S. A.; Marshall, G.; Pillai, G. V.; Castro, J. L.; Atack,
J. R.; MacLeod, A. M. Synthesis and Biological Evaluation of
3-Heterocyclyl-7,8,9,10-tetrahydro-(7,10-ethano)-1,2,4-triazolo[3,4-
a]phthalazines and Analogues as Subtype-Selective Inverse Agonists
for the GABAAR5 Benzodiazepine Binding Site. J. Med. Chem. 2004,
47, 3642–3657.
(50) Clayton, T.; Chen, J. L.; Ernst, M.; Richter, L.; Cromer, B. A.;
Morton, C. J.; Ng, H.; Kaczorowski, C. C.; Helmstetter, F. J.;
Furtmu¨ller, R.; Ecker, G.; Parker, M. W.; Sieghart, W.; Cook, J. M.
An Updated Unified Pharmacophore Model of the Benzodiazepine
Binding Site on γ-Aminobutyric Acida Receptors: Correlation with
Comparative Models. Curr. Med. Chem. 2007, 14, 2755–2775.
(51) Zhang, W.; Diaz-Arauzo, H.; Allen, M. S.; Koehler, K. F.; Cook, J. M.
Chemical and Computer Assisted Development of the Inclusive
Pharmacophore of Benzodiazepine Receptors. In Studies in Medicinal
Chemistry; ChoudharyM. I., Ed.; Harwood Academic Publishers:
Chichester, U.K., 1996; pp 303.
(52) Buhr, A.; Baur, R.; Malherbe, P.; Sigel, E. Point Mutations of the
R1ꢀ2γ2 γ-aminobutyric acidA Receptor Affecting Modulation of the
Channel by Ligands of the Benzodiazepine Binding Site. Mol.
Pharmacol. 1996, 49, 1080–1084.
(53) Sigel, E.; Baur, R.; Trube, G.; Mo¨hler, H.; Malherbe, P. The Effect
of Subunit Composition of Rat Brain GABAA Receptors on Channel
Function. Neuron 1990, 5, 703–711.
(54) Sawyer, G. W.; Chiara, D. C.; Olsen, R. W.; Cohen, J. B. Identification
of the Bovine γ-Aminobutyric Acid Type A Receptor R Subunit
Residues Photolabeled by the Imidazobenzodiazepine [3H]Ro15-4513.
J. Biol. Chem. 2002, 277, 50036–50045.
(55) Da Settimo, A.; Primofiore, G.; Da Settimo, F.; Marini, A. M.;
Novellino, E.; Greco, G.; Martini, C.; Senatore, G.; Lucacchini,
A. Synthesis of 3-(2′-Furoyl)indole Derivatives as Potential New
Ligands at the Benzodiazepine Receptor, Structurally More Re-
strained Analogues of Indoleglyoxylylamides. Farmaco 1995, 50,
311–320.
(56) Miller, R. B.; Frincke, J. M. Synthesis of Isoquinolines from Indenes.
J. Org. Chem. 1980, 45, 5312–5315.
(37) Huey, R.; Morris, G. M.; Olson, A. J.; Goodsell, D. S. A Semiempirical
Free Energy Force Field with Charge-Based Desolvation. J. Comput.
Chem. 2007, 28, 1145–1152.
(38) Cromer, B. A.; Morton, C. J.; Parker, M. W. Anxiety over GABA(A)
Receptor Structure Relieved by AChBP. Trends Biochem. Sci. 2002,
27, 280–287.
(39) Ernst, M.; Bruckner, S.; Boresch, S.; Sieghart, W. Comparative models
of GABAA Receptor Extracellular and Transmembrane Domains:
Important Insights in Pharmacology and Function. Mol. Pharmacol.
2005, 68, 1291–1300.
(40) Case, D. A.; Darden, T. E.; Cheatham, T. E. I.; Simmerling, C. L.;
Wang, J.; Duke, R. E.; Luo, R.; Merz, K. M., Wang, B.; Pearlman,
D. A.; Crowley, M.; Brozell, S.; Tsui, V.; Gohlke, H.; Mongan, J.;
Hornak, V.; Cui, G.; Beroza, P.; Schafmeister, C.; Caldwell, J. W.;
Ross, W. S.; Kollman, P. A. , AMBER 9; University of California:
San Francisco.
(57) Besnard, F.; Even, Y.; Itier, V.; Granger, P.; Partiseti, M.; Avenet,
P.; Depoortere, H.; Graham, D. Development of Stable Cell Lines
Expressing Different Subtypes of GABAA Receptors. J. Recept. Signal
Transduction Res. 1997, 17, 99–113.
(58) Martini, C.; Lucacchini, A.; Ronca, G.; Hrelia, S.; Rossi, C. A.
Isolation of Putative Benzodiazepine Receptors From Rat Brain
Membranes by Affinity Chromatography. J. Neurochem. 1982, 38,
15–19.
(59) Lowry, O. H.; Rosebrough, N. J.; Farr, A. L.; Randall, R. J. Protein
Measurement With the Folin Phenol Reagent. J. Biol. Chem. 1951,
193, 265–275.
(60) Accelrys, Inc., San Diego, CA.
(61) Dauber-Osguthorpe, P.; Roberts, V. A.; Osguthorpe, D. J.; Wolff,
J.; Genest, M.; Hagler, A. T. Structure and energetics of ligand
binding to proteins. Proteins: Struct., Funct., Genet. 1988, 4, 31-
47.
(41) Sanner, M. F.; Python, A. Programming Language for Software
Integration and Development. J. Mol. Graphics Modell. 1999, 17, 57–
61.
(42) Pettersen, E. F.; Goddard, T. D.; Huang, C. C.; Couch, G. S.;
Greenblatt, D. M.; Meng, E. C.; Ferrin, T. E. UCSF ChimerasA
(62) Fletcher, R. Unconstrained Optimization. In Practical Methods of
Optimization; John Wiley & Sons: New York, 1980; Vol. 1.
(63) Wang, J.; Wolf, R. M.; Caldwell, J. W.; Kollman, P. A.; Case, D. A.
Development and Testing of a General Amber Force Field. J. Comput.
Chem. 2004, 25, 1157–1174.

3734 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 12
Taliani et al.
(64) Wang, J.; Cieplak, P.; Kollman, P. A. How Well Does a Restrained
Electrostatic Potential (RESP) Model Perform in Calculating Confor-
mational Energies of Organic and Biological Molecules? J. Comput.
Chem. 2000, 21, 1049–1074.
(65) Cornell, W. D.; Cieplak, P.; Bayly, C. I.; Gould, I. R.; Merz, K. M.;
Ferguson, D. M.; Spellmeyer, D. C.; Fox, T.; Caldwell, J. W.;
Kollman, P. A. A Second Generation Force Field for the Simulation
of Proteins, Nucleic Acids, and Organic Molecules. J. Am. Chem.
Soc. 1995, 117, 5179–5197.
(66) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.;
Klein, L. M. Comparison of Simple Potential Functions for
Simulating Liquid Water. J. Chem. Phys. 1983, 79, 926–935.
(67) Essmann, U.; Perera, L.; Berkowitz, M. L.; Darden, T.; Lee, H.;
Pedersen, L. G. A Smooth Particle Mesh Ewald Method. J. Chem.
Phys. 1995, 103, 8577–8593.
(68) Ryckaert, J. P.; Ciccotti, G.; Berendsen, H. J. C. Numerical Integration
of the Cartesian Equations of Motion of a System with Constraints:
Molecular Dynamics of n-Alkanes. J. Comput. Phys. 1977, 23, 327–
341.
(69) Berendsen, H. J. C.; Postma, J. P. M.; Van Gunsteren, W. F.; Di Nola,
A.; Haak, J. R. Molecular Dynamics with Coupling to an External
Bath. J. Chem. Phys. 1984, 81, 3684–3690.
(70) Shao, J.; Tanner, S. W.; Thompson, N.; Cheatham, T. E. I. Clustering
Molecular Dynamics Trajectories: I. Characterizing the Performance
of Different Clustering Algorithms. J. Chem. Theory Comput. 2007,
3, 2312–2334.
JM9001154