Stepwise modulation of neurokinin-3 and neurokinin-2 receptor affinity and selectivity in quinoline tachykinin receptor antagonists

Article abstract of DOI:10.1021/jm000501v

A stepwise chemical modification from human neurokinin-3 receptor (hNK-3R)-selective antagonists to potent and combined hNK-3R and hNK-2R antagonists using the same 2-phenylquinoline template is described. Docking studies with 3-D models of the hNK-3 and

Full text of DOI:10.1021/jm000501v

J . Med. Chem. 2001, 44, 1675-1689
1675
Step w ise Mod u la tion of Neu r ok in in -3 a n d Neu r ok in in -2 Recep tor Affin ity a n d
Selectivity in Qu in olin e Ta ch yk in in Recep tor An ta gon ists
Frank E. Blaney,*,† Luca F. Raveglia,*,‡ Marco Artico,‡ Stefano Cavagnera,‡ Catherine Dartois,§ Carlo Farina,‡
Mario Grugni,‡ Stefania Gagliardi,‡ Mark A. Luttmann,| Marisa Martinelli,‡ Guy M. M. G. Nadler,§
Carlo Parini,‡ Paola Petrillo, Henry M. Sarau,| Mark A. Scheideler, Douglas W. P. Hay,| and
Giuseppe A. M. Giardina*,‡
Department of Computational & Structural Sciences, SmithKline Beecham Pharmaceuticals, New Frontiers Science Park,
Third Avenue, Harlow, Essex, England CM19 5AW, Departments of Medicinal Chemistry and Neurobiology Research,
SmithKline Beecham S.p.A., Via Zambeletti, 20021 Baranzate, Milano, Italy, SmithKline Beecham Laboratoires
Pharmaceutiques, Unite´ de Recherche, 4, rue du ChesnaysBeauregard, Boite Postale 96205,
F-35762 St. Gregoire, France, and Department of Pulmonary Biology, SmithKline Beecham Pharmaceuticals,
709 Swedeland Road, King of Prussia, Pennsylvania 19406-0939
Received November 27, 2000
A stepwise chemical modification from human neurokinin-3 receptor (hNK-3R)-selective
antagonists to potent and combined hNK-3R and hNK-2R antagonists using the same
2-phenylquinoline template is described. Docking studies with 3-D models of the hNK-3 and
hNK-2 receptors were used to drive the chemical design and speed up the identification of
potent and combined antagonsits at both receptors. (S)-(+)-N-(1-Cyclohexylethyl)-3-[(4-
morpholin-4-yl)piperidin-1-yl]methyl-2-phenylquinoline-4-carboxamide (compound 25, SB-
400238: hNK-3R binding affinity, Ki ) 0.8 nM; hNK-2R binding affinity, Ki ) 0.8 nM) emerged
as the best example in this approach. Further studies led to the identification of (S)-(+)-N-
(1,2,2-trimethylpropyl)-3-[(4-piperidin-1-yl)piperidin-1-yl]methyl-2-phenylquinoline-4-carboxa-
mide (compound 28, SB-414240: hNK-3R binding affinity, Ki ) 193 nM; hNK-2R binding
affinity, Ki ) 1.0 nM) as the first hNK-2R-selective antagonist belonging to the 2-phenylquino-
line chemical class. Since some members of this chemical series showed a significant binding
affinity for the human µ-opioid receptor (hMOR), docking studies were also conducted on a
3-D model of the hMOR, resulting in the identification of a viable chemical strategy to avoid
any significant µ-opioid component. Compounds 25 and 28 are therefore suitable pharmacologi-
cal tools in the tachykinin area to elucidate further the pathophysiological role of NK-3 and
NK-2 receptors and the therapeutic potential of selective NK-2 (28) or combined NK-3 and
NK-2 (25) receptor antagonists.
In tr od u ction
ground on the main biological aspects, profiles of neu-
rokinin receptors and neurokinin receptor antagonists,
and an extensive description of the lead generation and
lead optimization processes in the discovery of the
neurokinin receptor antagonists are given in recently
published reviews.1-7
The mammalian tachykininssor neurokininssin-
cluding substance P (SP), neurokinin A (NKA), and
neurokinin B (NKB), are a family of small peptides
widely distributed in the central and peripheral nervous
systems.1 They produce an array of biological responses,
including smooth muscle contraction and relaxation,
vasodilation, activation of the immune system, regula-
tion of pain transmission and neurogenic inflammation,
and they have been proposed to play a pathophysiologi-
cal role in a range of CNS and peripheral disorders,
including migraine, depression, anxiety, emesis, asthma,
and irritable bowel disease (IBD). These effects are
mediated by specific 7 transmembrane G-protein coupled
receptors (7 TM-GPCRs). The three tachykinins are able
to activate all the receptors, although SP has the highest
affinity for the neurokinin-1 receptor (NK-1R), NKA for
the NK-2R, and NKB for the NK-3R. A broad back-
In recent years an outstanding number of publications
have appeared in the literature describing both diverse
chemical classes and specific compounds as selective
antagonists of the NK-1, NK-2, and NK-3 receptors.2,3,6
In contrast, there has been minimal information in the
primary literature with regard to combined tachykinin
receptor antagonists; however, in the patent literature,
many nonselective tachykinin antagonists are claimed.8
The rationale of synthesizing combined antagonists for
the neurokinin receptors is mainly based on the poten-
tial synergistic effect associated with a multiple receptor
interaction.4,6 In particular, with regard to the NK-2 and
NK-3 receptors, the potential for synergistic pharma-
cological effects is supported by the different locations
of the two receptors. By its strategic location at the
ganglia of parasympathetic nerves,9,10 the NK-3R has
the potential to influence the release, and thus the
effect, of several rather than one neurotransmitter.
Accordingly, an NK-3R antagonist would be expected
to have therapeutic activity in pathological conditions
* To whom correspondence should be addressed. Phone: +39 02
3806-2265. Fax: +39 02 3806 2606. E-mail: giuseppe_giardina@
sbphrd.com.
† Department of Computational & Structural Sciences, SmithKline
Beecham Pharmaceuticals.
‡ Department of Medicinal Chemistry, SmithKline Beecham.
§ SmithKline Beecham Laboratoires Pharmaceutiques.
| Department of Pulmonary Biology, SmithKline Beecham Phar-
maceuticals.
Department of Neurobiology Research, SmithKline Beecham.
10.1021/jm000501v CCC: $20.00 © 2001 American Chemical Society
Published on Web 04/25/2001

1676 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 11
Ch a r t 1. Chemical Structure of Compound 1
Blaney et al.
formulas 5a -j. Acidic hydrolysis in refluxing 6 N HCl,
and a subsequent coupling reaction with the appropriate
primary amines (R1-H) in THF/DCM in the presence
of O-benzotriazol-1-yl-N,N,N′,N′-tetramethyluronium-
hexafluorophosphate (HBTU) as a condensing agent and
TEA gave the desired amides 6-9, 11-20, 22, and
24-28 of general formula I in Table 1.
Compounds 10, 21, and 23 were obtained by fmoc
deprotection of the corresponding compounds 9, 20, and
22, respectively, in MeCN at room temperature in the
presence of piperidine.
that have a significant neuronal component. NK-2R is
found on end-organs, such as airway, bladder, and
gastrointestinal (GI) tract smooth muscles.9 Pulmonary,
bladder, and GI tract disorders may therefore be the
main potential therapeutic targets for combined NK-3
and NK-2 receptor antagonists.
In principle it should be possible to test the hypothesis
of pharmacological synergism by utilizing a combination
of two molecules, a selective NK-3R antagonist and a
selective NK-2R antagonist. The issue associated with
this approach lies primarily in the likely different
pharmacokinetic profiles of the two compounds. An
alternative, more attractive approach would be to test
a single compound possessing combined NK-2R/NK-3R
antagonist activity and to compare its effects with those
of selective receptor antagonists.
The aim of this publication is to describe how the
chemical structure of quinoline-based selective NK-3R
antagonists, exemplified by (S)-(-)-N-(1-phenylpropyl)-
3-hydroxy-2-phenylquinoline-4-carboxamide (1, SB-
223412, talnetant, Chart 1),11 was modified in a step-
wise manner to obtain combined NK-3 and NK-2
receptor antagonists or selective NK-2R antagonists
without major changes of the chemical template.
Chemical synthesis (Scheme 1), physicochemical prop-
erties (Table 1), and radioligand binding affinities (Table
2) of compounds 6-8, 10-19, 21, and 23-28 will be
detailed.
Moreover, novel 3D models of the human NK-3,
NK-2, and µ-opioid receptors and docking hypotheses
for the selective and combined ligands will be described.
Inputs from these models were key drivers in the
stepwise optimization process mentioned above.
P h a r m a cology
Ra d ioliga n d Bin d in g Assa ys. Neurokinin receptors
binding assays were performed with crude membranes
from Chinese hamster ovary (CHO) cells expressing the
hNK-3, hNK-2, or hNK-1 receptors (hNK-3-CHO, hNK-
2-CHO, and hNK-1-CHO, respectively) as detailed
previously.12-14 For NK-3R competition binding studies,
[
125I]-[MePhe7]-NKB binding to hNK-3-CHO mem-
branes was performed using the procedure of Sadowski
and co-workers15 as described previously.14 For NK-2R
competition binding studies, [125I]-NKA binding to hNK-
2-CHO membranes was performed essentially as de-
scribed by Aharony et al.16 Competition binding studies
for the NK-1R were performed on hNK-1-CHO mem-
branes essentially as described by Payan et al.17
Stable expressions of human δ- (hDOR) and human
µ-opioid receptors (hMOR) in CHO cell lines and of
κ-opioid receptors (hKOR) in human embrionic kidney
(HEK-293) cell lines have been performed in house,
using pCDN vectors.18 Receptor binding assays were
performed in membranes prepared by lysis in hypotonic
phosphate buffer19 using the method described by
Kotzer et al.18 with minor modifications. Briefly, the
following radioligands were utilized: [3H]-[D-Ala2,Me-
Phe4,Gly-ol5]enkephalin ([3H]-DAMGO, New England
Nuclear, Belgium), [3H]-[D-Ala2,D-Leu5] enkephalin
([3H]-DADLE, New England Nuclear, Belgium), and
[3H]-U-69593 (Amersham, Italia) to label µ-, δ-, and
κ-opioid receptors, respectively.
Concentration-response curves in the hNK-3-CHO,
hNK-2-CHO, and hMOR-CHO binding assays were run
using duplicate samples for compounds 1, 6-8, 10-19,
21, and 23-28; among them, compounds that gave an
IC50 binding affinity lower than 2.5 µM in the first
experiment were run in three to five independent
experiments (n ) 3-5).
The percent inhibition of specific binding in the NK-3
and NK-2 receptor binding assays was determined for
each concentration of the compounds and the IC50,
defined as the concentration required to inhibit 50% of
the specific binding, obtained from concentration-
response curves. Values reported in Table 2 are the
apparent inhibition constant (Ki), which was calculated
from the IC50 as described by Cheng and Prusoff.20
Compounds 1, 6, and 24 were studied in the hNK-1-
CHO binding assay, while compounds 1, 25, and 28
were evaluated in the hKOR and hDOR binding assays.
Ch em istr y
The synthesis of compounds 6-28 was straightfor-
ward because they were all derived from a common key
intermediate (3-bromomethylquinoline derivative 4,
Scheme 1), which could be easily produced in large
amounts by radical bromination of methyl 3-methyl-2-
phenylquinoline-4-carboxylate (3) with N-bromosuccin-
imide (NBS) in refluxing MeCN and in the presence of
a catalytic amount of dibenzoylperoxide. Compound 3
in Scheme 1 could in turn be prepared by simple
esterification of the corresponding, commercially avail-
able carboxylic acid (2, [CAS 43071-45-0]); this esteri-
fication was found to increase the yields of the bromi-
nation step with respect to the same reaction directly
performed on the carboxylic acid. Nucleophilic displace-
ment of the reactive benzyl bromide by a variety of
secondary amines (R-H in Scheme 1) in THF in the
presence of diisopropylethylamine (DIEA) at a temper-
ature ranging from 20 to 50 °C afforded compounds of
Resu lts a n d Discu ssion
1. Mod elin g th e Neu r ok in in NK-2/NK-3 a n d th e
µ-Op ioid Recep tor s. For some years now, we21 and
others22 have described models of G protein-coupled

Neurokinin-3 and Neurokinin-2
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 11 1677
Sch em e 1. Synthesis of Compounds 6-28a
a
Reagents: (a) (COCl)2, DCM, room temperature, 1 h; MeOH, DCM, room temperature, 18 h; (b) NBS, dibenzoylperoxide, CH3CN,
reflux, 28 h; (c) RH (a secondary amine), DIEA, THF, room temperature to 50 °C, overnight; (d) HCl (6 N), reflux, 4 h; HBTU, R1H (a
primary amine), TEA, THF/DCM, 25 °C to reflux, 18 h; (e) (only for compounds 9, 20, and 22) piperidine, CH3CN, room temperature,
overnight.
receptors (GPCRs), which were initially based on the
transmembrane helix packing bundle of bacteriorho-
dopsin (BR). These were modified when the early 9 Å
cryoelectron diffraction map of bovine rhodopsin23 sug-
gested that the relative tilt of the helices differed
from that of BR. A major breakthrough was reported
in 1997 by Unger et al,24a when a 6 Å electron den-
sity map of rhodopsin was described as a series of slices
at 4 Å intervals through the membrane. The implica-
tions of this, particularly the unexpected tilt angle of
transmembrane helix 3 (TM3), led us to conclude that
a totally new template structure was necessary be-
fore further ligand design work could reliably be per-
formed.
It was assumed in the description below that the same
helix packing arrangement is present in all GPCR

1678 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 11
Blaney et al.
Ta ble 1. Physical Properties of Compounds 6-8, 10-19, 21, and 23-28 of General Formula I

Neurokinin-3 and Neurokinin-2
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 11 1679
Ta ble 1 (Continued)
a
b
d
c ) 1, MeOH. c ) 0.25, MeOH; nd ) not determined. c Cl: calcd, 17.26; found, 16.78. C: calcd, 63.21; found, 63.88. H: calcd, 7.46;
found, 7.03. Cl: calcd, 17.49; found,17.01. e C: calcd, 65.90; found, 65.44. f C: calcd, 75.25; found, 75.67.
Ta ble 2. Binding Affinities of Compounds 1, 6-8, 10-19, 21,
and 23-28 at Cloned hNK-3, hNK-2, and hMOR Receptors
adjusting these where necessary to take account of
irregularities in the helical axes.
Use was made at this stage of the program Helanal27
developed by us some years ago to define the interior
face and the hydrophobic exposure of a transmembrane
helix. This defines the conservation moment arising
from a multiple sequence alignment of closely related
proteins, as a Fourier power spectrum, using the
algorithm originally described by Donnelly et al.28 The
moment is then mapped onto a color-coded helical
wheel. A set of rules are used to define a hydrophobic
arc on the wheel, which gives an indication of the degree
to which the helix is exposed to the lipid bilayer. When
this is used together with site-directed mutagenesis
information from a variety of sources, the helical
domains were folded around the axes described above.
Density around TM6 suggests a strongly bent helical
structure that corresponds to a kink produced by the
highly conserved proline. This places two highly con-
served aromatic residues (phenylalanine F 261 and
tryptophan W265 of bovine rhodopsin) on the inside of
the helix bundle. In the aminergic receptors, the con-
served TM6 proline is followed by two phenylalanine
residues. The orientation in the model suggests that the
latter of these is inward-facing, whereas the former is
interhelical. Mutation studies of these phenylalanines
in the 5HT2a receptor29 have shown that it is the latter
residue that is mainly involved in ligand binding, in
keeping with the model.
binding affinities
hNK-2/hNK-3
(Ki, mean ( SEM (nM))
receptor
compd
hNK-3
hNK-2
hMOR
selectivity ratio
1a,b 1.4 ( 0.2
144 ( 22
172 ( 18
1860 ( 560
1680 ( 480
103
78
16
39
4.7
3.1
1.3
1.2
0.4
0.02
<0.1
0.015
0.003
3.1
0.7
1.1
0.3
1
6a
7
8
2.2 ( 0.3
3.3 ( 1.1
7.0 ( 1.1
4.7 ( 1.0
1.8 ( 0.4
1.2 ( 0.1
1.2 ( 0.1
1.9 ( 0.4
454 ( 129
>12000
51.6 ( 18.8 >2500
276 ( 99
22.2 ( 3.6
5.7 ( 0.5
1.6 ( 0.1
1.5 ( 0.2
0.7 ( 0.1
9.2 ( 0.8
811 ( 211
>2500
10
203 ( 12
98.6 ( 10.3
15.8 ( 0.6
120 ( 13
195 ( 59
d
11
12
13
14
15
16
17
18
19
21
23
d
d
1210 ( 180 18.5 ( 9.3
1250 ( 30
1.5 ( 0.4
3.4 ( 0.6
31.2 ( 2.4
4.2 ( 0.3
4.6 ( 0.4
2.4 ( 0.4
33.0 ( 8.9
0.6 ( 0.1
0.8 ( 0.1
6.1 ( 0.5
1.2 ( 0.1
1.0 ( 0.2
226 ( 15
919 ( 89
172 ( 6.8
1780 ( 450
1320 ( 170
>2500
24a,c 1.7 ( 0.1
25b
26
0.8 ( 0.1
4.6 ( 0.8
1.4 ( 0.3
193 ( 48
>2500
1.3
0.9
0.005
27
>2500
28b
g2000
a
Compounds 1, 6, and 24 were assessed for their ability to
displace [3H]-SP binding from hNK-1-CHO membranes with the
method described in ref 17; they were devoid of any significant
b
effects up to concentration of 100 µM. Compounds 1, 25, and 28
were assessed for their ability to displace [3H]-U69593 binding
from HEK membranes transfected with the hKOR and [3H]-
DADLE from CHO membranes transfected with the hDOR and
were devoid of any significant effects up to concentration of 2.5
µM. c A representative member of this chemical series, compound
24, was assessed for its cellular functional antagonist activity in
the NKB-induced CA2+ mobilization assay in HEK cells stably
expressing the hNK-3R (KB ) 2.6 nM, pA2 ) 8.7) following the
method described in ref 12, and in the NKA-induced Ca2+
mobilization assay in HEK cells stably expressing the hNK-2R
(KB ) 9.9 nM, pA2 ) 8.0); these data confirm the well-known
antagonist properties of quinoline-based NK-3 and NK-2 receptor
TM3 was placed with the cysteine at the extracellular
side oriented toward TM5, in anticipation of the forma-
tion of a disulfide bond with the second extracellular
loop between TM4 and TM5. The seventh residue below
this cysteine has been identified from mutagenesis
work30 as the key TM3 binding residue for many
receptors; e.g., it is an aspartate in the aminergic, opioid,
and several peptide receptors and binds to a protonated
amino group in the natural ligands. As a result of the
cysteine placement, this was also oriented toward the
TM5-TM6 region, which has again been identified from
mutagenesis studies29,31 to be crucial for the binding of
small molecule agonists such as the monoamines. This
orientation of TM3 is quite different from the model
described by Baldwin,24b where the above residues in
TM3 are oriented more toward TM2 and TM7.
ligands.11,14 Not determined.
d
family 1 receptors. Therefore, ligand binding and site-
directed mutagenesis data were collated from a number
of different receptors. As a starting point for the
construction of the new template, a computer program
was written that would convert the electron density slice
images into accurate three-dimensional representations
(see Experimental Section for details of the algorithm
used). These density contours were then used to define
the axes of the helices in a manner similar to that
described by Baldwin.24b The helix sequences derived
from hydropathy plots25 were then manually folded
around these axes, assuming the standard values of
-59° and -44° for φ and ψ in a hydrophobic helix,26 but
TM4 contains a highly conserved tryptophan found
in both family 1 and family 2 receptors, which was
predicted to be inward-facing by Helanal. The unusual
tilt of TM3 suggested that this trypyophan played a
structural role, acting as a “wedge” packing against TM3

1680 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 11
Blaney et al.
and forcing it through the middle of the bundle. TM4
was thus aligned with its axis in such a way that the
tryptophan was stacked against a tryptophan (W126)
of bovine rhodopsin in the intracellular half of TM3. The
structural rather than binding role of the TM4 tryp-
tophan has been confirmed in the muscarinic receptor
where mutation to phenylalanine had no effect on ligand
binding.32
quickly became clear that these ligands would have to
extend into the extracellular domains. The extracellular
loops were therefore added using a new semiautomated
procedure (see Experimental Section for details of this
method). Initially the long second loop between TM4 and
TM5 was built with its disulfide bond between the
cysteine at the top of TM3 and the cysteine in the loop.
The shorter TM2-TM3 and TM6-TM7 loops were then
added afterward.
As with TM6, a similar though less pronounced kink
was observed in the helix axis of TM5. This was again
aligned with the conserved proline in this region.
Mutation studies in the adrenergic and dopamine recep-
tors have suggested that two serine residues in TM5
are involved in agonist binding.31 The helix was rotated
to ensure that these residues were oriented toward the
interior of the bundle. The density slice information has
suggested that TM5 has relatively little interhelical
interaction with TM4 and TM6 at the extracellular side.
This is supported by SCAM mutations carried out on
the dopamine D2 receptor,33b which show that all
residues at the extracellular side of TM5 are relatively
accessible. The main implication of this is that the
assumed binding pocket between TM3, TM5, and TM6
is larger than in previous models.
2. Str u ctu r e-Activity Relation sh ips (SARs, Table
2). Structure-activity considerations on a large set of
selective NK-3R antagonists reported in a previous
paper in this journal11 identified position 3 of the
quinoline nucleus as the best point for chemical varia-
tion. This observation prompted a broad investigation
of substitution at this position with the aim of improving
the affinity for the NK-2R. In addition, a particular focus
was given to tertiary amines with the objective of
improving water solubility and brain penetration (com-
pared to the same parameters obtained for compound
1).12
A conformational analysis study on the rotatable
bonds of 2,3,4-trisubstituted quinolines revealed, not
suprisingly, that these compounds are very rigid. The
ortho substitution in the quinoline means that the
2-phenyl substituent of the series sits well out of plane
with the bicyclic ring. For the same reason, the 4-amido
group adopts one of two narrow ranges of conformations
that are essentially orthogonal to the quinoline ring and
180° apart from each other. This rigidity limited the
number of rotational conformers that needed to be
examined during docking studies in the receptor models.
The remaining three helices proved to be much harder
to align with their axes. The main guidance here was
that there is ample evidence suggesting that three
highly conserved residues, an asparagine in TM1, an
aspartate in TM2, and an asparagine in TM7, together
form a hydrogen-bonded network that is involved in the
functional response of family 1 GPCRs.34 The placement
of TM2 was carried out on the basis of an unusual
double bend in the helix axis, which was attributed to
the double glycine (G89/G90) in TM2 of rhodopsin. The
conserved aspartate (D83) was oriented toward TM7.
This then allowed placement of the TM7 asparagine
(N302) within hydrogen-bonding distance of the aspar-
tate. Much adjustment of the φ-ψ angles of TM7 was
necessary to allow the lysine (K296) of rhodopsin to
orient itself in a suitable rotamer into the bundle so that
it could form the Schiff base with retinal. Finally, TM1
was placed by aligning the kink in the axis with proline
(P 53) in rhodopsin and rotating the helix so that the
TM1 asparagine (N55) was oriented toward asparagine
(N302) in TM7.
Compound 1 was manually placed in the NK-3R
model in the cavity between TM3, TM4, TM5, TM6,
and TM7. A variety of different starting orientations
were used. Care was taken with this and other ligands
to ensure that only low-energy rotational conformers
were considered during the initial placements. Each
starting orientation was then minimized fully using the
CHARMm program. Where likely hydrogen bonding or
charged interactions were identified, during the initial
placements, no distance constraints were applied to the
first stages of the minimization procedure. It became
clear that one particular binding orientation explained
much of the previously observed SAR of these com-
pounds.11,14 In Figure 1, it can be seen that the essential
quinoline nitrogen forms a hydrogen bond with a
tyrosine in TM6 (Y315). The quinoline ring itself is
surrounded by a number of other aromatic residues,
including a tyrosine (Y256) in TM5 and a phenylalanine
(F 319) and histidine (H316) in TM6. This histidine can
also form a hydrogen bond with the amide hydrogen on
the 4-substituent of the quinoline. The 2-phenyl sub-
stituent is forced to sit orthogonal to the quinoline ring,
in keeping with the conformational analysis results, and
fits into a very tight pocket bounded by an asparagine
(N162) and proline (P 165) in TM3, a phenylalanine
(F 234) in the second extracellular loop, and a tyrosine
(Y338) and phenylalanine (F 342) in TM7. The tight
dimensions of this pocket explain previous observations
that only very limited substitution (a 2-fluoro group) is
tolerated on this ring.14
Models of the TM domains of numerous GPCRs have
been built on the basis of this template, using the auto-
mated program GPCR_Builder.35 The resulting struc-
tures were all subjected to several rounds of side chain
conformational refinement using the Karplus rotamer
library36 option in Quanta,37 prior to their final mini-
mization with CHARMm.38 Comparison of the models
with experimental data is generally very good. For ex-
ample the dopamine D2 model is in agreement with the
SCAM data published by J avitch33 and the 5HT2 models
explain most of the published mutagenesis experi-
ments.29,39 What was particularly encouraging was that
the model of the double histidine mutant of the NK-1R,
constructed by Elling et al.40 in their work on GPCR
zinc binding sites, exhibited a near-perfect zinc-histi-
dine triad arrangement between TM5 and TM6.
The NK-2, NK-3, and µ-opioid receptor models were
built from the template as described above and used in
initial docking studies of the quinoline antagonists. It
With the exception of phenylalanine (F 234), all these
residues are conserved in the NK-2R. However, docking

Neurokinin-3 and Neurokinin-2
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 11 1681
F igu r e 1. Compound 1 bound in the NK-3 receptor model. Most residues have been removed for clarity.
F igu r e 2. Sequence alignment of the NK-3 and NK-2 receptors showing the various regions used in the modeling (TM )
transmembrane helix; EC ) extracellular loop; IC ) intracellular loop). The residues involved in binding are in bold below the
1 symbol. Conserved residues are indicated by the / symbol below the alignment. The cysteines of the disulfide bond are below
the 9 symbol.
of 1 into the NK-2R model revealed some interesting
differences between the two receptors; these differences
are highlighted by the alignment of the two sequences
shown in Figure 2. In both models, the 3-hydroxy
substituent forms a hydrogen bond with the TM3
asparagine (N162). In the NK-3R, however, there is an
additional strong electrostatic interaction with a lysine
(K223) at the top (extracellular side) of TM4. Indeed it
is likely that the acidic nature of the phenolic group
would imply that it is ionized in the NK-3R, which
would increase the strength of the interaction. The
corresponding residue in the NK-2R, threonine (T171),
is too small to make any interactions with the ligand.
This would suggest that increasing the size of the
3-substituent, together with the introduction of suitable
functionality in this substituent, should increase the
affinity for the NK-2R. Thus, the selectivity of compound
1 for the NK-3R can be explained by the presence of
this salt bridge alone.
Further differences between the two models can be
seen in the binding pocket for the (S)-1-phenylpropyl
amide substituent. In the NK-3R, this binds in a
hydrophobic pocket bounded by isoleucine (I166), valine
(V169), and phenylalanine (F 170) in TM3 and by the
valines (V251 and V255) and proline (P 259) in TM5.
In the NK-2R, valine (V169) is replaced by a methionine

1682 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 11
Blaney et al.
and valine (V255) is replaced by an isoleucine. Again,
this would suggest that subtle changes in the bulk of
the amide substituent could affect the selectivity for the
two receptors.
benzylic position was predicted to be more favored for
protonation over the secondary piperazine nitrogen.
N-alkylation of the piperazine might be expected to shift
this equilibrium in favor of the nonbenzylic position.
However, calculations showed that the tertiary pipera-
zines still favored protonation at the benzylic position
(albeit with a smaller energy difference), thus maintain-
ing the interaction with the TM3 asparagine. While
N-alkylation of the piperazine would not change the
protonation site, examination of the models suggested
that extension of an alkyl substituent into the cavity
would pick up further hydrophobic interactions with the
TM4 threonine (T171) and the loop histidine (H198) in
the NK-2 receptor. N-alkylation at the piperazine ring
with the i-Pr group (compound 11) produced an ap-
proximate 3-fold increase in both hNK-3R (Ki ) 1.8 nM)
and hNK-2R (Ki ) 5.7 nM) binding affinities over the
unsubstituted piperazine 10, resulting in a slightly
increased combined profile (hNK-2R/hNK-3R binding
affinity ratio of 3.1). As predicted, an even more
pronounced effect was observed with compound 12
featuring the more bulky cyclohexyl group (hNK-2R/
hNK-3R binding affinity ratio of 1.3).
The proposed introduction of tertiary amino substit-
uents at the 3-position was chemically facilitated by the
possibility of utilizing a common key intermediate (the
3-bromomethylquinoline derivative 4) that could be
easily produced in large amounts and coupled with a
set of diverse secondary amines in high yields.
Whereas the smallest tertiary amine (the dimethy-
lamine) produced the moderately NK-3R selective com-
pound 6 (hNK-2R/hNK-3R binding affinity ratio ) 78),
introduction of the piperidino group (compound 8)
decreased the selectivity by a factor of 2, thus resulting
in a hNK-2R/hNK-3R receptor binding affinity ratio of
39. Docking of these compounds into the NK-2R model
suggested that the protonated nitrogen of the 3-sub-
stituent could now form a hydrogen bond with the amide
carbonyl in the side chain of asparagine (N110) in TM3.
In the NK-3R, the same interaction can be seen, but in
addition, it is likely that proton exchange occurs with
the primary amino group of the TM4 lysine (K223). As
predicted earlier, the increased bulk of the piperidine
ring allows it to make favorable interactions with
cysteine (hNK-2R C167) in TM4 and valine (hNK-2
V201) in TM5, thus increasing the affinity for the
NK-2R.
Because compound 11 (and 12) showed potent binding
affinities at both receptors, further attention was fo-
cused on the role of the amidic portion while maintain-
ing the i-Pr-piperazine group in position 3 of the
quinoline nucleus. As has already been mentioned, there
are subtle differences between the receptor models in
this binding pocket. Accordingly, compounds 13-18 in
Table 2 feature several amide variations. The first
modification consisted of the simple replacement of the
Et at the benzylic position with an i-Pr group, giving
compound 13, which showed a 3-fold increase in hNK-
2R binding affinity. Thus, this compound had equivalent
affinity at both receptors (Ki about 1 nM). Docking of
this compound into the two receptor models showed that
the increase in NK-2R affinity could readily be explained
by the increased hydrophobic interaction of the bulkier
isopropyl group with the unique TM3 methionine (M117).
In the NK-3R model, the phenyl group of the 4-amido
substituent is sandwiched between the two valines
(V169 and V255) that are specific to this sequence.
Replacement of these in the NK-2R by the larger
methionine and isoleucine residues, respectively, sug-
gested that increase in the hydrophobic bulk of the
phenyl substituent would give an increased affinity for
the NK-2R. Indeed a further increase in the hNK-2R
binding affinity was obtained by replacing the phenyl
ring with the cyclohexyl and, concomitantly, the Et with
the Me group to give compound 14 (hNK-3R binding
affinity, Ki ) 1.9 nM; hNK-2R binding affinity, Ki ) 0.7
nM). The larger hydrophobic cavity in the NK-3R would
also suggest that reduction in size of the amide sub-
stituent would have a greater detrimental effect on
binding than in the NK-2R. Replacement of the cyclo-
hexyl group with i-Pr (15) produced a compound that,
although less potent at both receptors, was clearly much
more hNK-2R selective (hNK-2R/hNK-3R receptor bind-
ing affinity ratio of 0.02). Therefore, in agreement with
the prediction from the docking studies, the data
obtained for compounds 14 and 15 suggest that a
decrease in the steric hindrance of substituents at the
amidic nitrogen should be associated with a strong
One significant difference between the NK-2 and
NK-3 receptor sequences (Figure 2) was the presence
of a lysine (K180) located in the former receptor in the
second extracellular loop, adjacent to the disulfide bond.
In the NK-3R sequence, this amino acid is leucine. In
our model of the NK-2R, this side chain could extend
into the cavity toward the 3-substituent. In addition,
the different lengths of the second extracellular loop in
the two receptors resulted in a different orientation of
the conserved histidine (hNK-2R H198) at the C-
terminal end of this loop. In the NK-2R, this histidine
is also pointing into the binding cavity. These observa-
tions suggested that a second polar substituent in the
piperidine ring could improve the NK-2R affinity.
Significantly, a considerable increase in the combined
NK-2R/NK-3R receptor binding profile was obtained by
introducing an extra basic nitrogen in the piperidine
ring, producing the piperazine derivative 10. Docking
of this compound into the NK-2R model showed that
the loop lysine (K180) would now form an interaction
with the piperazine nitrogen. This compound showed a
marked combined profile (hNK-3R binding affinity, Ki
) 4.7 nM; hNK-2 binding affinity, Ki ) 22.2 nM) with
a hNK-2R/hNK-3R receptor binding affinity ratio of 4.7.
The introduction of the second amino group meant that
two possible protonated states would need to be con-
sidered in the modeling. Estimations of the relative
basicities were made on the basis of the relative heats
of formation of the isomers, as calculated by ab initio
Hartree-Fock methods.41 A second approach to this was
to calculate the maximum positive electrostatic potential
of the protonated isomer at the surface of this proton.41
This follows from the fact that the more basic the
nitrogen, the more polarized the NH bond of the cation
will be, and hence, it will develop a more positive charge
at the NH center. With compound 10, the tertiary

Neurokinin-3 and Neurokinin-2
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 11 1683
decrease in the hNK-3R binding affinity while poten-
tially maintaining an adequate hNK-2R component. To
further test this hypothesis, we decided to introduce
small alkyl symmetric groups, which led to the achiral
compounds 16-18. While replacement of the i-Pr group
in compound 15 with the simple Me afforded a poorly
active compound at both receptors (16), incorporation
of the isopentyl group on the amidic nitrogen was clearly
beneficial and provided compound 17 with a selectivity
ratio of 0.015. The most hNK-2R selective compound
was obtained by introducing a cyclohexyl group directly
attached to the amidic nitrogen (18). This had an
outstanding hNK-2R/hNK-3R receptor binding affinity
ratio of 0.003 with a hNK-3R binding affinity of 1245
nM and hNK-2R binding affinity of 4.2 nM. Docking
studies showed that the cyclohexyl group could interact
with the TM3 methionine in the NK-2R but could not
interact with the corresponding smaller valine in the
NK-3R.
scribed. On the other hand, only weak, nonspecific polar
interactions were observed with the other protonated
isomer. As with other opioid receptors, the overriding
driving force for binding to human µ opioid is the
interaction of the protonated nitrogen with a conserved
aspartate on TM3 (human µ D149). The µ receptor does
not have the TM3 asparagine or the TM6 tyrosine,
which are essential residues for NK receptor binding,
and it has a bulky tyrosine instead of the TM3 isoleucine
in the phenylpropyl pocket. Thus, the µ receptor is
believed to bind the compounds in a totally different
manner. Docking studies of 12 in the human µ-opioid
receptor model (Figure 3) showed that the protonated
nitrogen at the nonbenzylic position could interact with
the aspartate on TM3. The NH of the 4-amido substitu-
ent could still hydrogen-bond to the histidine at the top
of TM6 (human µ H299), which is conserved in the three
receptors. The molecule is essentially binding “upside
down” compared to the NK receptor binding mode, with
no major contributions other than a shape complemen-
tarity (see Figure 4). The opioid binding mode showed
that the position and directionality of the protonated
nitrogen were crucial. Further docking studies sug-
gested that if the nitrogen of 12 was shifted into the
cyclohexyl ring to give the 4-(piperidin-1-yl)piperidin-
1-yl derivative 19, then interaction with the TM3
aspartate in the µ-opioid receptor would be abolished
without affecting its NK receptor binding.
To test this hypothesis compound 19 was synthesized.
In perfect agreement with the docking study prediction,
19 displayed a human µ-opioid receptor binding affinity
(Ki) of 919 nM while maintaining low nM binding
affinities for both the hNK-3R (Ki ) 1.5 nM) and hNK-
2R (Ki ) 4.6 nM).
This trend in SARs was also verified in another amide
series, featuring the (S)-1-cyclohexylethylamine. As with
10, compound 21 (incorporating the piperazinylmethyl
group at position 3 of the quinoline ring) showed a high
binding affinity for both the hNK-3R (Ki ) 3.4 nM) and
hNK-2R (Ki ) 2.4 nM) and similar human µ-opioid
receptor binding affinity (Ki ) 172 nM).
It is worth noting that this achievement of such good
hNK-2R selectivity was made with a compound devoid
of stereogenic centers, whereas previous literature
evidence suggested that the presence of the stereogenic
center in the amidic portion of the molecule was
necessary for strong hNK-3R binding affinity in quino-
line-based hNK-3R selective antagonists.13,14
In view of the potential in vivo use of some of the
compounds described above, it was decided to profile
them in a battery of receptor, ion channel, and enzyme
assays, paying particular attention to the opioid receptor
family. The reason for a specific opioid profile came
from the observation that another NK-3R antagonist,
(R)-(+)-N-{{3-[1-benzoyl-3-(3,4-dichlorophenyl)piperidin-
3-yl]prop-1-yl}-4-phenylpiperidin-4-yl}-N-methylaceta-
mide (SR 142,801, osanetant), albeit chemically unre-
lated, has been described as interacting with opioid
receptors with IC50s in the 0.1-1 uM range (human
µ-opioid receptor binding affinity, Ki ) 480 nM, in-house
data).42 As a result, compounds 10-14 and 18 showed
human µ-opioid receptor binding affinities in the range
16-226 nM, whereas compounds 1 and 6-8 were devoid
of any significant human µ-opioid receptor binding
affinity.
Again, when the nitrogen position is shifted as in
compounds 23-25, the human µ-opioid binding affinity
was significantly reduced or even essentially abolished
(it ranged from 1322 to >2500 nM).
From these results, it could be speculated that the
µ-opioid receptor binding component was related to the
presence of the second piperazine nitrogen (compare
compound 10, human µ-opioid binding affinity, Ki ) 203
nM with compound 8, human µ-opioid binding affinity,
Ki > 2.5 µM). To study this further, compound 12, which
showed the greatest affinity for the human µ-opioid
receptor (human µ-opioid Ki ) 15.8 nM), was examined
in detail in all three receptor models. Ab initio quantum
calculations, carried out as described earlier, demon-
strated that the preferred site of protonation was at the
nonbenzylic nitrogen, although the difference between
the two sites was calculated to be only 4.3 kcal (in the
gas phase). An estimation of the difference in pKa
between the two protonated states using electrostatic
potential calculations showed that they were quite close
(ca. 0.4 kcal), which would suggest that they are in
equilibrium. Docking studies with compound 12 in both
the NK-2 and NK-3 receptors showed that with benzylic
protonation, the protonated site formed a strong hydro-
gen bond to the TM3 asparagine, as previously de-
In particular, compound 25, featuring the 4-(morpho-
lin-4-yl)piperidin-1-yl derivative, showed an impressive
binding profile and can be considered as a viable
pharmacological tool to assess the therapeutic potential
of combined NK-3 and NK-2 receptor antagonists. In
fact, in addition to the exquisite selectivity over the
µ-opioid receptor (human µ-opioid receptor binding
affinity, Ki > 2.5 µM), 25 diplayed very high hNK-3R
(Ki ) 0.8 nM) and hNK-2R (Ki ) 0.8 nM) binding
affinities. Docking studies (Figure 5) showed that the
oxygen of the morpholine ring was able to pick up a
hydrogen bond with a conserved serine at the top of
TM4 (S170 in NK-2R/S222 in NK-3R) and a conserved
histidine (H198 in NK-2R/H248 in NK-3R) at the top
of TM5. A lack of any significant binding affinity for the
κ- and δ-opioid receptors completed the in vitro biologi-
cal profile.
A further validation of the SAR hypothesis explaining
the µ-opioid component on the basis of the nitrogen

1684 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 11
Blaney et al.
F igu r e 3. Docking of compound 12 in the human µ-opioid receptor model. The protonated nitrogen is interacting with aspartate
(D149) in TM helix 3. Note that this orientation is essentially inverted with respect to the neurokinin docking mode.
F igu r e 4. Side view of compound 12 in the NK-2 (right) and human µ-opioid (left) receptor models showing the different binding
orientations. The extracellular side is on the top of the picture.
position and its basic nature was given by the inactivity
at the human µ-opioid receptor of compounds 26 (lack
of the second piperazine basic nitrogen) and 27 (lack of
the basic nature of the second piperazine nitrogen). Both
26 and 27 showed binding affinities for the hNK-3 and
hNK-2 receptors lower than 6.1 nM.
combine the 4-(piperidin-1-yl)piperidin-1-yl basic moiety
with a variety of amides possessing those features
believed to be crucial in minimizing the hNK-3R com-
ponent while maintaining a strong hNK-2R binding
affinity. Compound 28, featuring the (S)-3,3-dimethyl-
but-2-yl residue, was found to be the most selective
antagonist among all the compounds synthesized, show-
ing a hNK-2R binding affinity of 1.0 nM (Ki), a hNK-2/
Finally, to obtain also a selective hNK-2R compound
with no opioid receptor binding affinity, we decided to

Neurokinin-3 and Neurokinin-2
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 11 1685
affinity, Ki ) 0.8 nM) emerged as the best example of
combined NK-3 and NK-2 receptor antagonist, (S)-(+)-
N-(1,2,2-trimethylpropyl)-3-[(4-piperidin-1-yl)piperidin-
1-yl]methyl-2-phenylquinoline-4-carboxamide (compound
28, SB-414240, hNK-3R binding affinity, Ki ) 193 nM;
hNK-2R binding affinity, Ki ) 1.0 nM) was found to be
the most NK-2R-selective. Both these compounds are
devoid of any significant affinity for the µ, δ, and κ-opioid
receptors.
On the basis of the above considerations we have
demonstrated the possibility of utilizing a common
chemical template to modulate the NK-2 and NK-3
receptor affinity and selectivity by subtle chemical
modifications.
Compounds 25 and 28 join compound 1 as viable
pharmacological tools that will be used to better eluci-
date the pathophysiological role of the NK-3 and NK-2
receptors and to clarify the therapeutic potential of
NK-3R-selective, NK-2R-selective, or the combined an-
tagonists at both these receptors.
F igu r e 5. Compound 25 docked in the NK-2 receptor model.
Note the additional interactions of the morpholine oxygen with
serine (S170) and histidine (H198).
F ootn ote
hNK-3 receptor binding affinity ratio of 0.005, and
human µ-opioid receptor binding affinity greater than
2000 nM (Ki).
While the design work described above was carried
out on theoretical models of the receptors, during the
time of writing, the first X-ray crystal structure of a
GPCR, viz., bovine rhodopsin, was published.43 Com-
parison of our theoretical models with homology models
of the receptors showed that they were in fact very close
in structure. TM helices 3, 6, and most of 7 aligned
almost perfectly with the crystal-based models. Helices
1 and 2 were correctly oriented toward the inner face
of the bundle but were approximately displaced by one
helical turn in the transmembrane axis. Helices 4 and
5 were correct in their placement in the membrane axis,
but the inward faces were out by approximately 60°. The
overall result was that the binding pocket described
above was largely unchanged. For NK-3R, tyrosine
(Y315), phenylalanine (F 319), and histidine (H316) still
form the main binding pocket for the quinoline ring. The
3-phenyl substituent remains in a pocket bounded by
asparagine (N162), proline (P 165), tyrosine (Y338), and
phenylalanine (F 342). The asparagine can still form a
hydrogen bond with the protonated nitrogen. The only
major difference between our theoretical model and the
crystal-based homology model is in the binding pocket
for the hydrophobic substituent on the 4-amido group.
On one side, the TM3 residues isoleucine (I166), valine
(V169), and phenylalanine (F 170) remain in the same
orientation. However, the differing orientation of TM5
means that the role of valine (V255) in NK-3R is
replaced by tyrosine (Y256). However, the differing
selectivities in this pocket can still be explained by the
valine (V169 in NK3) to methionine (NK-2R) change as
described earlier. While the TM helices of GPCR models
can be placed with much greater confidence as a result
of the crystal structure, the extracellular loop regions
would be expected to have much greater flexibility. The
deeply buried second extracellular loop of rhodopsin is
unlikely to be mimicked by other antagonist bound
receptors because such ligands would have no access to
the binding pocket. Loop modeling, as described earlier,
is still therefore necessary in the consideration of GPCR
homology structures and their ligand docking.
Con clu sion s
In this paper, optimization of the combined hNK-3
and hNK-2 receptor profile of compounds 1, 6-8, 10-
19, 21, and 23-28 by means of stepwise chemical
modifications was presented. This optimization was
carried out in conjunction with docking studies, using
novel receptor models that were based on de novo
folding of the TM helices around published electron
density information. Another key feature of the models
was the inclusion of the extracellular loops that were
built using a novel combined distance geometry-locally
enhanced sampling (LES) simulation method. The
optimization process was conducted by starting from
compounds 1 and 6, which displayed a hNK-3R-selective
profile. Two portions of the molecule, where modeling
has suggested that differences in the receptors could be
utilized (namely, the basic group at position 3 and the
amide at position 4 of the quinoline ring), were broadly
investigated.
The introduction of a basic, sterically hindered and
cyclized group at position 3 increased the hNK-2R
component. However, incorporation of the piperazine
group (compounds 10-18 and 21) resulted in a signifi-
cant µ-opioid binding interaction that was attributed to
the nonbenzylic piperazine nitrogen. As suggested by
docking studies, this issue was overcome by a simple
shift of the position of this nitrogen (compounds 19, 23-
25, and 28) or by removal of its basic nature (compound
27).
Substituents on the amidic portion proved to be
crucial with regard to affinity for the hNK-3R, thus
allowing the modulation of the neurokinin selectivity
profile. In particular, compounds lacking the benzyl or
cyclohexylmethyl groups had significantly lower affinity
for the hNK-3R. Whereas (S)-(+)-N-(1-cyclohexylethyl)-
3-[(4-morpholin-4-yl)piperidin-1-yl]methyl-2-phenylquin-
oline-4-carboxamide (compound 25, SB-400238, hNK-
3R binding affinity, Ki ) 0.8 nM; hNK-2R binding

1686 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 11
Blaney et al.
at the high speed. Pellets were resuspended in buffer A
containing 0.32 M sucrose and 5 mM EDTA (buffer B) to wash
and concentrate the membranes. The final pellets were
resuspended in buffer B at a final concentration of 1-2 mg
protein/mL (ca. 40 × 106 cells/mL) and stored at -80 °C.
Protein was determined with the protein assay kit from Sigma
Chemical Co. (Milan, Italy). MgCl2 (3 mM) was added to buffer
A in the case of cells expressing hMOR and hKOR.
Binding experiments were performed in triplicate at a final
protein concentration of 10 µg/mL in buffer A and a radiola-
beled ligand concentration of 0.4-0.5 nM. The nonspecific
binding was determined in the presence of naloxone 10 µM
and under these assay conditions was less than 1% of the
radioligand added. Samples (final volume of 2 mL) were
incubated for 60 min at 37, 30, and 25 °C for δ-, µ-, and
κ-receptor binding assays, respectively. The reaction was
terminated by rapid filtration through Whatman GF/B filters
and two washes with cold assay buffer A (4 mL) using a M48
Brandel cell harvester (Biomedical Research and Development
Laboratories, Inc., Gaithersburg, MD). Filters used for [3H]-
U69593 binding were presoaked in buffer A containing 0.05%
polyethylenimine. Radioactivity on the filters was measured
by liquid scintillation counting with a Camberra Packard
2500TR â counter (Milan, Italy).
3. Com p u ta tion a l Ch em istr y. 3.1. Gen er a l. Visualization
of the density maps, the various receptor models, and their
ligands, as well as the generation of the initial rhodopsin model
and the manual docking of the ligands, was carried out with
the program Quanta.37 Molecular mechanics energy refine-
ments, minimizations, and molecular dynamics simulations
were performed with the program CHARMm.38 Ligand mol-
ecules were built using the builder option of the program
Spartan.41 They were further optimized at the semiempirical
level with the Vamp program,44 using the AM1 Hamiltonian
and suitable symmetry constraints to maintain planar amide
bonds. The atom-centered charges for the ligands, used in the
docking studies, were natural atomic orbital charges calculated
at the Hartree-Fock level within Spartan, using a 3-21G*
basis set. The electrostatic potential calculations used in the
estimation of relative basicities were also carried out with
Spartan at the same level. Following manual docking of the
ligands into the receptor models, they were minimized with
CHARMm using a distance-dependent dielectric and 1000
steps of steepest descent, followed by 5000 steps of ABNR.
Generation of the transmembrane helical bundles of the
NK-2, NK-3, and µ-opioid receptor models from the rhodopsin
template was done with the GPCR_Builder program.35
3.2. Gen er a tion of th e Extr a cellu la r Loop s. A semiau-
tomated procedure for the generation of the extracellular loops
has been developed. Initial conformational sampling was
performed using the distance geometry program DGEOM95.45
Distance constraints were automatically derived from the N-
and C-terminal helical five residues, together with the disulfide
bond information in the case of the second extracellular loop
and appropriate steric information from previously generated
loops. Typically 500 random conformations were generated.
A modification to the DGEOM95 program allowed the applica-
tion of the Karplus rotamer library36 prior to export of the files.
Each loop structure was minimized with CHARMm using 1000
steps of steepest descent and 5000 steps of ABNR. A constant
dielectric of 7 was used in this case to decrease the influence
of electrostatic terms. The conformations were analyzed by
goodness of backbone φ-ψ angles using the criteria described
by Wilmot et al.46 They were clustered by backbone dihedral
“families” and further ranked by energy.The best conforma-
tions were chosen for subsequent molecular dynamics analysis
using the locally enhanced sampling (LES) method of Elber,47
as utilized in CHARMm. Following heating and equilibration
at 400 K, the loops were run for 250 ps, and the trajectories
were again analyzed by backbone φ-ψ angles. The best
resulting conformations were further minimized as described
above.
Exp er im en ta l Section
1. Ra d ioliga n d Bin d in g Assa y. Neu r ok in in Recep tor s.
The CHO cells expressing the hNK-1, hNK-2, and hNK-3
receptors were cultured at 37 °C in a humidifier incubator
under 5% CO2/95% air in 1017 SO3 (in-house formulation)
media containing nucleosides plus Geneticin (400 mg/L). The
cells were harvested by centrifugation at 600 g for 10 min.
The cell pellet was resuspended in hypotonic buffer (10 mM
Tris, pH 7.4, 1.0 mM EDTA, 10 µg/mL soybean trypsin
inhibitor, 100 µg/mL bacitracin, 100 µM benzamidine, and 10
µM phenylmethylsulfonyl fluoride), and the mixture was then
rapidly frozen and thawed (3 times) followed by Dounce
homogenization for preparation of crude membranes. For NK-
3R competition binding studies,15 membranes (∼15 µg protein)
were incubated with 0.15 nM [125I]-[MePhe7]-NKB in a total
of 150 µL of 50 mM Tris, pH 7.4, 4 mM MnCl2, 1 µM
phosphoramidon, and 0.1% ovalbumin, with or without various
concentrations of antagonist, for 90 min at 25 °C. Incubations
were stopped by rapid filtration with
a Brandell tissue
harvestor (Gaithersburg, MD) through Whatman GF/C filters
that were presoaked for 60 min in 0.5% bovine serum albumin
(BSA). Membranes were washed with 10 mL of ice cold 20 mM
Tris, pH 7.4 containing 0.1% BSA, then placed in vials with
10 mL of Beckman “Ready Safe” and counted in a Beckman
LS 6000 (Fullerton, CA) liquid scintillation counter. Specific
binding was determined by subtracting total binding from
nonspecific binding, which was assessed as the binding in the
presence of 0.5 µM cold [MePhe7]-NKB.
For NK-2R competition binding studies,16 cells were grown
and membranes were prepared as described above for the
hNK-3R binding assay. The assay buffer was the same as that
utilized for the hNK-3R binding assay with a total volume of
150 µL, ∼10 µg of membrane protein and 0.15 nM [125I]-NKA,
which was incubated for 90 min at 25 °C, with various
concentrations of antagonist. Filtration was through Whatman
GF/C filters soaked for 30 min in 0.1% polyethylenimine (PEI),
and membranes were washed as described above for the hNK-
3R binding assay. Nonspecific binding was determined in the
presence of 0.5 µM cold NKA. For NK-1R competition binding
studies,17 the assay volume was 300 µL and the assay buffer
(25 mM Tris, pH 7.4, 2.0 mM CaCl2, 2.0 mM MgCl2, 1 µM
phosphoramidon, and 0.1% ovalbumin) contained various
concentrations of antagonist and 1.0 nM [3H]-substance P.
Membranes were incubated for 45 min at 25 °C; Whatman
filters were presoaked with BSA and membranes washed as
described for the hNK-3R binding assay. Nonspecific binding
was determined in the presence of 1 µM cold substance P and
Ki determined as for the NK-3 and NK-2 receptor binding
assays.
2. Op ioid Recep tor s. CHO cells expressing the hDOR and
hMOR were grown in suspension culture in serum-free media
(1017 SO3) in the presence of 0.05% pluronic acid (F18) and
maintained at 37 °C and gassed with 5% CO2. Selection for
CHO transfectants was performed by growth in the absence
of nucleotides. The maximum cell density for these cell lines
was 4 × 106 cells/mL. From binding studies it was determined
that transfected CHO cell lines express hDOR and hMOR at
a density of 11.5 and 6.9 pmol/mg protein, respectively.
HEK-293 cells expressing hKOR were grown attached to
T-150 flasks in Earles’s minimum essential medium (E-MEM)
supplemented with 10% FBS and 2 mM L-glutamine; G418
was included for selection. The stable cell line expresses hKOR
at a density of 3.6 pmol/mg protein from binding studies.
Cells were harvested in phosphate-buffered saline (1 × 106
cells/ mL) and collected by centrifugation (800 g for 5 min).
The pellets were resuspended in the same volume of ice-cold
10 mM potassium phosphate buffer pH 7.2 (buffer A) and
centrifuged at 40 000 g for 10 min. The cells were hypo-
osmotically lysed by resuspension in the same volume of buffer
A for 20 min in ice and centrifuged at 800 g (5 min), saving
the supernatant. The resulting pellets were resuspended in
buffer A, and the last step was repeated two more times,
saving supernatants each time. Supernatants derived from the
three low-speed centrifugations were pooled and centrifuged
3.3. Der iva tion of th e Th r ee-Dim en sion a l Den sity
Slices. The images from the paper of Unger et al.24a were

Neurokinin-3 and Neurokinin-2
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 11 1687
scanned into the computer and were then converted to pure
black and white 24 bit (8 bit each of red, green, and blue)
formatted bitmap postscript files. These have the white
represented as “ffffff” in hexadecimal notation and the black
as “000000”. The computer program written to convert these
files into three-dimensional information searched each line of
the image, noting the coordinates of the midpoint of any
transition from white to black to white. These were interpreted
as points on the density contour lines. Adjacent points were
then connected using a simple distance search algorithm, and
the coordinates of the points were converted to angstroms
using the scale factor in the original paper. The density
information for each slice was then written as a formatted
“QM coordinate” file suitable for input into the Quanta
program.
N-bromosuccinimide (13 g, 72 mmol) was added, and the
reaction mixture was heated to reflux. After the addition of
dibenzoylperoxide (1 g, 4.1 mmol), the reaction was refluxed
for 24 h, then additional N-bromosuccinimide (4 g, 22.5 mmol)
and dibenzoylperoxide (0.5 g, 2.0 mmol) were added, and the
reaction was refluxed for 4 h. The solvent was evaporated to
dryness, and the residue was taken up with EtOAc and washed
with H2O, 5% NaHCO3, and H2O. The organic layer was dried
over anhydrous Na2SO4, filtered, evaporated to dryness, and
triturated with Et2O to yield 4 (9.0 g, 25.3 mmol, 70%) as a
white solid; mp 118-120 °C. IR (KBr): 3453, 3062, 3037, 2952,
1
1735, 1575, 1556 cm-1. H NMR (DMSO): δ 8.09 (d, J ) 8.0
Hz, 1H), 7.89 (d, J ) 8.0 Hz, 2H), 7.88 (dd, J ) 8.0, 8.0 Hz,
1H), 7.73 (dd, J ) 8.0, 8.0 Hz, 1H), 7.67 (m, 2H), 7.55 (m, 3H),
4.69 (s, 2H), 4.11 (s, 3H). EI-MS (source 180 °C, 70 eV, 200
mA): m/z 355 (M+•), 276, 232, 216.
4. Syn th etic Ch em istr y. Melting points were determined
with a Bu¨chi 530 hot stage apparatus and are uncorrected.
Proton NMR spectra were recorded on a Bruker ARX 300
spectrometer at 303 K unless otherwise indicated. Chemical
shifts were recorded in parts per million (δ) downfield from
tetramethylsilane (TMS); NMR spectral data are reported as
a list. IR spectra were recorded in Nujol mull or neat on sodium
chloride disks or in KBr with a Perkin-Elmer 1420 spectro-
photometer; mass spectra were obtained on a Finnigan MAT
TSQ-700 (or TSQ 70) spectrometer. Optical rotations were
determined in MeOH solution at the indicated concentration
with a Perkin-Elmer 341 polarimeter at the sodium D-line.
Silica gel used for flash column chromatography was Kiesegel
60 (230-400 mesh) (E. Merck AG, Darmstadt, Germany).
Evaporations were performed at reduced pressure, and all oily
products were dried at 0.1 mbar for 16 h. Combustion
elemental analyses were performed by Redox s.n.c., Milan,
Italy, and values found were within 0.4% of the theoretical
values (unless otherwise indicated).
Tetrahydrofuran (THF) was dried by distillation over LiAlH4
and stored over 4 Å molecular sieves under nitrogen atmo-
sphere; DCM was dried over CaCl2; MeCN and DMF were
stored over 4 Å molecular sieves. Triethylamine (TEA) was
dried by distillation and stored over KOH. Amines RH, where
R has the meanings reported in Scheme 1 for compounds 5a -g
and 5i, are commercially available compounds; 4-piperidin-4-
ylmorpholine was prepared according to Freifelder et al.,48 and
2-methyl-1-piperazin-1-ylpropan-1-one was prepared according
to Meyer et al.49 Amines R1H, where R1 has the meanings
reported in Scheme 1 for compounds 6, 8-12, and 14-28, are
commercially available compounds; (S)-2-methyl-1-phenylpro-
pylamine was prepared according to Pallavicini et al.50 Com-
pound 2 is a commercially available compound or can be
prepared according to the synthetic procedure reported by
Giardina et al.11
5.3. Gen er a l P r oced u r e for th e Syn th esis of Com -
p ou n d s 5a -j in Sch em e 1. Methyl 3-bromomethyl-2-phen-
ylquinoline-4-carboxylate (4, 5 g, 14 mmol), amine RH (where
R has the meanings reported in Scheme 1 for compounds 5a -
j) (15.4 mmol), and ethyldiisopropylamine (2.7 mL, 15.4 mmol)
were dissolved in dry THF (100 mL), and the mixture was
stirred for 1 night at 50 °C. The solvent was concentrated, the
residue was dissolved in CH2Cl2 and washed with water, and
the organic phase was dried over MgSO4. After concentration
of the solvent, the residue was purified by flash column
chromatography over silica gel, affording compounds 5a -j in
50-80% yield.
As an example, spectroscopic data of compound 5f are
reported. Mp 118-120 °C. IR (KBr): 3572, 3054, 2968, 2935,
1
2784, 1715, 1573, 1556 cm-1. H NMR (DMSO): δ 8.06 (d, J
) 8.8 Hz, 1H), 7.86 (d, J ) 8.8 Hz; 1H), 7.81 (dd, J ) 8.8, 8.8
Hz, 1H), 7.68 (dd, J ) 8.8, 8.8 Hz, 1H), 7.51 (m, 5H), 3.98 (s,
3H), 3.61 (s, 2H), 2.67-2.58 (m, 2H), 2.60-2.35 (m, 4H), 2.11
(m, 1H), 1.78 (m, 2H), 1.62-1.28 (m, 10H). ESI-MS (positive,
solvent methanol, spray 4.5 keV, skimmer 60 eV, capillary 220
°C): m/z 444 (MH+).
6.0. Syn th esis of F in a l Com p ou n d s. 6.1. Gen er a l P r o-
ced u r e for th e Syn th esis of Com p ou n d s 6-9, 11-20, 22,
a n d 24-28. Compounds 5a -j (7.9 mmol) and 6 N HCl (50
mL) were refluxed for 4 h and then concentrated to dryness.
The residue was triturated with diethyl ether to obtain the
corresponding carboxylic acids, which were used without
further purification in the coupling step. These carboxylic acids
(7.9 mmol), amines R1H (where R1 has the meanings reported
in Scheme 1 for compounds 6-28) (12.6 mmol), HBTU (4.8 g,
12.6 mmol), and TEA (5.1 mL, 37 mmol) were dissolved in THF
(100 mL) and DCM (25 mL), and the mixture was refluxed
for 18 h. The solvent was evaporated, and the residue was
dissolved in DCM and washed with H2O, 0.5 N NaOH, and
H2O. The organic phase was dried over MgSO4, filtered, and
evaporated to dryness. The residue was purified by flash
column chromatography, affording compounds 6-9, 11-20,
22, and 24-28 in 60-85% yield. In some cases the free base
was dissolved in acetone and acidified with Et2O saturated
with HCl to obtain the corresponding dihydrochloride or
trihydrochloride.
5. Syn th esis of In ter m ed ia tes. 5.1. Meth yl 3-Meth yl-
2-ph en ylqu in olin e-4-car boxylate (3). 3-Methyl-2-phenylquin-
oline-4-carboxylic acid (2, 30 g, 114 mmol, CAS [43071-45-0])
were suspended in dry DCM (250 mL); oxalyl chloride (20 mL,
230 mmol) dissolved in DCM (120 mL) was added dropwise,
and the reaction mixture was stirred at room temperature for
30 min. Two drops of DMF were added, and the reaction was
stirred for an additional 30 min. The solvent was evaporated
to dryness, the residue was taken up with DCM (100 mL), and
MeOH (100 mL) dissolved in DCM (400 mL) was added
dropwise. After the mixture was stirred for 18 h, the solvent
was evaporated to dryness and the residue was taken up with
DCM and washed with 1% NaHCO3. The organic layer was
dried over anhydrous Na2SO4, filtered, and evaporated to
dryness to yield 3 (31.4 g, 113 mmol, 99%) as a solid, which
was used in the following reaction without further purification;
mp 73-75 °C. IR (KBr): 3441, 3051, 2954, 1731, 1582, 1556
As an example, spectroscopic data of compound 24 are
20
reported. Mp 180-184 °C. [R]D +6.9° (c 0.5, MeOH). IR
(KBr): 3422, 2928, 2852, 2659, 1647, 1546 cm-1 1H NMR
.
(DMSO; free base): δ 8.27 (d br, J ) 8.5 Hz; 1H), 8.01 (d, J )
8.2 Hz, 1H), 7.86 (d, J ) 8.2 Hz, 1H), 7.75 (dd, J ) 8.2, 8.2 Hz,
1H), 7.62 (dd, J ) 8.2, 8.2 Hz, 1H), 7.55 (m, 2H), 7.50-7.42
(m, 3H), 4.09-3.98 (m, 1H), 3.56 (s, 2H), 2.52 (m, 2H), 2.34
(m, 4H), 2.03-1.93 (m, 1H), 1.88-1.60 (m, 6H), 1.56-1.05 (m,
17H), 1.20 (d, J ) 6.9 Hz, 3H). EI-MS (source 180 °C, 70 eV,
200 mA): m/z 538 (M+•), 372, 261, 167. Anal. (C35H46N4O‚3HCl)
C, H, N, Cl.
1
cm-1. H NMR (CDCl3): δ 8.15 (d, J ) 8.0 Hz, 1H), 7.73 (m,
2H), 7.62-7.40 (m, 7H), 4.10 (s, 3H), 2.40 (s, 3H). EI-MS
(source 180 °C, 70 eV, 200 mA): m/z 277 (M+•), 276, 262, 246,
217, 166.
5.2. Meth yl 3-Br om om eth yl-2-p h en ylqu in olin e-4-ca r -
boxyla te (4). Methyl 3-methyl-2-phenyl-quinoline-4-carboxy-
late (3, 10 g, 36 mmol) was dissolved in CH3CN (500 mL);
6.2. Gen er a l P r oced u r e for th e Syn th esis of Com -
p ou n d s 10, 21, a n d 23. Fmoc deprotection of compounds 9,
20, and 22 was obtained by dissolving these compounds in
CH3CN, adding piperidine (1.5 equiv), and stirring them at
room temperature overnight. Evaporation to dryness and
purification by flash column chromatography afforded the

1688 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 11
Blaney et al.
Nonpeptide tachykinin receptor antagonists. I. Pharmacological
and pharmacokinetic characterization of SB 223412, a novel,
potent and selective NK-3 receptor antagonist. J . Pharmacol.
Exp. Ther. 1997, 281, 1303-1311.
corresponding deprotected compounds 10, 21, and 23 in nearly
quantitative yields.
As an example, spectroscopic data of compound 21 are
20
reported. Mp 112-114 °C. [R]D +20.5 (c 1, MeOH). IR
(13) Giardina, G. A. M.; Sarau, H. M.; Farina, C.; Medhurst, A. D.;
Grugni, M.; Foley, J . J .; Raveglia, L. F.; Schmidt, D. B.; Rigolio,
R.; Vassallo, M.; Vecchietti, V.; Hay, D. W. P. 2-Phenyl-4-
quinolinecarboxamides: A Novel Class of Potent and Selective
Non-Peptide Competitive Antagonists for the Human Neuroki-
nin-3 Receptor. J . Med. Chem. 1996, 39, 2281-2284.
(14) Giardina, G. A. M.; Sarau, H. M.; Farina, C.; Medhurst, A. D.;
Grugni, M.; Raveglia, L. F.; Schmidt, D. B.; Rigolio, R.; Lutt-
mann, M.; Vecchietti, V.; Hay, D. W. P. Discovery of a novel class
of selective non-peptide antagonists for the human neurokinin-3
receptor. 1. Identification of the 4-quinolinecarboxamide frame-
work. J . Med. Chem. 1997, 40, 1794-1807.
(15) Sadowski, S.; Huang, R.-R. C.; Fong, T. M.; Marko, O.; Cascieri,
M. A. Characterization of the Binding of [125I-Iodo-histidyl,
methyl-Phe7] Neurokinin B to the Neurokinin-3 Receptor. Neu-
ropeptides 1993, 24, 317-319.
(16) Aharony, D.; Conner, G. E.; Woodhouse, D. P. Pharmacologic
Characterization of the Novel Ligand [4,5-3H-Leu9]-Neurokinin-
A Binding to NK-2 Receptors on Hamster Urinary Bladder
Membranes. Neuropeptides 1992, 23, 121-130.
(17) Payan, D. G.; McGillis, J . P.; Organist, M. L. Binding Charac-
teristics and Affinity Labeling of Protein Constituents of the
Human IM-9 Lymphoblast Receptor for Substance P. J . Biol.
Chem. 1986, 261, 14321-14329.
(18) Kotzer, C. J .; Hay, D. W. P.; Dondio, M.; Giardina, G. A. M.;
Petrillo, P.; Underwood, D. The antitussive activity of delta
opioid receptor stimulation in guinea pigs. J . Pharmacol. Exp.
Ther. 2000, 292, 803-809.
(Nujol): 3250, 3100, 1650, 1550, 1470. 1H NMR (DMSO): δ
8.51 (d br, J ) 8.5 Hz, 1H), 8.01 (d, J ) 8.2 Hz, 1H), 7.84 (d,
J ) 8.2 Hz, 1H), 7.77 (dd, J ) 8.2, 8.2 Hz, 1H), 7.64 (dd, J )
8.2, 8.2 Hz, 1H), 7.57 (m, 2H), 7.50-7.42 (m, 3H), 4.07-3.95
(m, 1H), 3.51 (s, 2H), 2.41 (m, 4H), 2.02 (m, 4H), 1.86-1.60
(m, 5H), 1.54-1.42 (m, 1H), 1.32-0.99 (m, 6H), 1.17 (d, J )
6.9 Hz, 3H). EI-MS (source 180 °C, 70 eV, 200 mA): m/z 456
(M+•), 372, 302, 261, 218, 217. Anal. (C29H36N4O) C, H, N.
Ack n ow led gm en t . We are grateful to Roberto
Rigolio, Davide Graziani, Lorenzo Perugini, Alberto
Cerri, and Renzo Mena (SmithKline Beecham SpA,
Italy) for synthetic assistance and mass and NMR
spectroscopic determinations. We also thank Isabelle
Leger, Isabelle Delimoge, and Marcel Morvan (Smith-
Kline Beecham Laboratoires Pharmaceutiques, France)
for NMR and synthetic assistance.
Su p p or tin g In for m a tion Ava ila ble: Detailed spectro-
scopic data (IR, MS, and 1H NMR) for compounds 6-8, 10-
19, 21, and 23-28. This material is available free of charge
via the Internet at http://pubs.acs.org.
(19) Scheideler, M. A.; Zukin, R. S. Reconstitution of solubilized delta-
opiate receptor binding sites in lipidic vesicles. J . Biol. Chem.
1990, 265, 15176-15182.
(20) Cheng, Y.-C.; Prusoff, W. H. Relationship Between the Inhibition
Constant (Ki) and the Concentration of Inhibitor Which Cause
50 Percent Inhibition (IC50) of an Enzymatic Reaction. Biochem.
Pharmacol. 1973, 22, 3099-3108.
Refer en ces
(1) Regoli, D.; Boudon, A.; Fauchere, J .-L. Receptors and antagonists
for substance P and related peptides. Pharmacol. Rev. 1994, 46,
551-599.
(2) Giardina, G. A. M.; Raveglia, L. F.; Grugni, M. Lead generation
and lead optimization processes in the discovery of selective
nonpeptide neurokinin receptor antagonists. Drugs Future 1997,
22, 1235-1257.
(3) Gao, Z.; Peet, N. P. Recent advances in neurokinin receptor
antagonists. Curr. Med. Chem. 1999, 6, 375-388.
(4) Von Sprecher, A.; Gerspacher, M.; Anderson, G. P. Neurokinin
antagonists as potential therapies for inflammation and rheu-
matoid arthritis. IDrugs 1998, 1, 73-91.
(5) Swain, C. J . Neurokinin receptor antagonists. In Progress in
Medicinal Chemistry; Ellis, G. P., Luscombe, D. K., Oxford, A.
W., Eds.; Elsevier Science B.V.: Amsterdam, 1998; Vol. 35, pp
57-81.
(6) Seward, E. M.; Swain, C. J .; Neurokinin receptor antagonists.
Expert Opin. Ther. Pat. 1999, 9, 571-582.
(21) (a) Bromidge, S. M.; Dabbs, S.; Davies, D. T.; Davies, S.;
Duckworth, D. M.; Forbes, I. T.; Gaster, L. M.; Ham, P.; J ones,
G. E.; King, F. D.; Mulholland, K. R.; Saunders, D. V.; Wyman,
P. A.; Blaney, F. E.; Clarke, S. E.; Blackburn, T. P.; Holland, V.;
Kennett, G. A.; Lightowler, S.; Middlemiss, D. N.; Trail, B.; Riley,
G. J .; Wood, M. D. Biarylcarbamoylindolines Are Novel and
Selective 5-HT2C Receptor Inverse Agonists: Identification of
5-Methyl-1-[[2-[(2-methyl-3-pyridyl)oxy]- 5-pyridyl]carbamoyl]-
6-trifluoromethylindoline (SB-243213) as a Potential Antidepres-
sant/Anxiolytic Agent. J . Med. Chem. 2000, 43, 1123-1134. (b)
Gaster, L. M.; Blaney, F. E.; Duckworth, D. M.; Ham, P.;
J enkins, S.; J ennings, A. J .; J oiner, G. F.; King, F. D.; Mulhol-
land, K. R.; Starr, S.; Wyman, P. A.; Hagan, J . J .; Hatcher, J .;
J ones, B. J .; Middlemiss, D. N.; Price, G. W.; Riley, G.; Roberts,
C.; Routledge, C.; Selkirk, J .; Slade, P. D. The Selective 5-HT1B
Receptor Inverse Agonist 1′-Methyl-5-[[2′-methyl-4′- (5-methyl-
1,2,4-oxadiazol-3-yl)biphenyl-4-yl]carbonyl]-2,3,6,7-te trahydro-
spiro[furo[2,3-f]indole-3,4′-piperidine] (SB-224289) Potently Blocks
Terminal 5-HT Autoreceptor Function Both in Vitro and in Vivo.
J . Med. Chem. 1998, 41, 1218-1235.
(22) Bikker, J . A.; Trumpp-Kallmeyer, S.; Humblet, C. G-Protein
Coupled Receptors: Models, Mutagenesis, and Drug Design. J .
Med. Chem. 1998, 41, 2911-2927.
(23) Schertler, G. F. X.; Villa, C.; Henderson, R. Projection structure
of rhodopsin. Nature (London) 1993, 362, 770-772.
(24) (a) Unger, V. M.; Hargrave, P. A.; Baldwin, J . M.; Schertler, G.
F. X. Arrangement of rhodopsin transmembrane R-helixes.
Nature (London) 1997, 389, 203-206. (b) Baldwin, J . M.;
Schertler, G. F. X.; Unger, V. M. An alpha-carbon template for
the transmembrane helixes in the rhodopsin family of G-protein-
coupled receptors. J . Mol. Biol. 1997, 272, 144-164.
(25) Kyte, J .; Dolittle, R. F. A simple method for displaying the
hydropathic character of a protein. J . Mol. Biol. 1982, 157, 105-
132.
(7) Giardina, G. A. M.; Grugni, M.; Raveglia, L. F. Recent advances
in neurokinin-3 receptor antagonists. Expert Opin. Ther. Pat.
2000, 10, 939-960.
(8) (a) Burkholder, T. P.; Maynard, G. D.; Kudlacz, E. M. (Assignor
to Hoechst Marion Roussel, Inc.). Novel carboxy substituted
cyclic carboxamide derivatives. International Patent WO/
9827085, 1998; Chem. Abstr. 1998, 129, 95498. (b) Reichard, G.
A.; Alaimo, C. A.; Shih, N.-Y.; Ting, P. C.; Carruthers, N. I.;
Lavey, B. J . (Assignor to Schering Corp.). Substituted oximes
as neurokinin antagonists. U.S. Patent 6,063,926, 2000; Chem.
Abstr. 2000, 132, 334363. (c) Esser, F.; Schnorrenberg, G.;
Dollinger, H.; Lu¨ttke, S.; J ung, B.; Speck, G. (Assignor to
Boehringer Ingelheim Pharma KG). Novel neurokinin antago-
nists, method for the production thereof, and pharmaceutical
composition containing said compounds. International Patent
DE/19824470, 1999; Chem. Abstr. 1999, 132, 12330. (d) Nishi,
T.; Yamaguchi, T.; Iio, Y.; Takemoto, T.; Nakajima, K. (Assignor
to Sankyo Co. Ltd.). Spiropiperidine derivatives. International
Patent WO/9937642, 1999; Chem. Abstr. 1999, 131, 116241.
(9) Myers, A. C.; Undem, B. J . Electrophysiological effects of
tachykinins and capsaicin on guinea-pig bronchial parasympa-
thetic neurones. J . Physiol. 1993, 470, 665-679.
(26) Blundell, T.; Barlow, D.; Borkakoti, N.; Thornton, J . Solvent
induced distortions and the curvature of R-helices. Nature
(London) 1983, 306, 281-283.
(27) Blaney, F. E.; Tennant, M. HelanalsA Novel Algorithm for the
Construction of Transmembrane Helical Bundles. Poster pre-
sented at the International Conference on Protein Folding and
Design, Bethesda, MD, 1996. See also ref 35.
(28) Donnelly, D.; Overington, J . P.; Blundell, T. L. The prediction
and orientation of R-helixes from sequence alignments: the
combined use of environment-dependent substitution tables.
Fourier transform methods and helix capping rules. Protein Eng.
1994, 7, 645-653.
(10) Maggi, C. A. Tachykinins in the autonomic nervous system.
Pharmacol. Res. 1996, 33, 161-170.
(11) Giardina, G. A. M.; Raveglia, L. F.; Grugni, M.; Sarau, H. M.;
Farina, C.; Medhurts, A. D.; Graziani, D.; Schmidt, D. B.; Rigolio,
R.; Luttmann, M.; Cavagnera, S.; Foley, J . J .; Vecchietti, V.; Hay,
D. W. P. Discovery of a novel class of selective non-peptide
antagonists for the human neurokinin-3 receptor. 2. Identifica-
tion of (S)-N-(1-phenylpropyl)-3-hydroxy-2-phenylquinoline-4-
carboxamide (SB 223412). J . Med. Chem. 1999, 42, 1053-1065.
(12) Sarau, H. M.; Griswold, D. E.; Potts, W.; Foley, J . J .; Schmidt,
D. B.; Webb, E. F.; Martin, L. D.; Brawner, M. E.; Elshourbagy,
N. A.; Medhurst, A. D.; Giardina, G. A. M.; Hay, D. W. P.

Neurokinin-3 and Neurokinin-2
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 11 1689
(29) Choudhary, M. S.; Craigo, S.; Roth, B. L. A single point mutation
(Phe340fLeu340) of a conserved phenylalanine abolishes 4-[125I]-
iodo-(2,5-dimethoxy)phenylisopropylamine and [3H]mesulergine
but not [3H]ketanserin binding to 5-hydroxytryptamine2 recep-
tors. Mol. Pharmacol. 1993, 43, 755-761.
(30) (a) Strader, C. D.; Sigal, I. S.; Register, R. B.; Candelore, M. R.;
Rands, E.; Dixon, R. A. F. Identification of residues required for
ligand binding to the â-adrenergic receptor. Proc. Natl. Acad.
Sci. U.S.A. 1987, 84, 4384-4388. (b) Mansour, A.; Meng, F.;
Meador-Woodruff, J . H.; Taylor, L. P.; Civelli, O.; Akil, H. Site-
directed mutagenesis of the human dopamine D2 receptor. Eur.
J . Pharmacol., Mol. Pharmacol. Sect. 1992, 227, 205-214.
(31) Strader, C. D.; Candelore, M. R.; Hill, W. S.; Sigal, I. S.; Dixon,
R. A. F. Identification of two serine residues involved in agonist
activation of the â-adrenergic receptor. J . Biol. Chem. 1989, 264,
13572-13578.
(37) Quanta98, version 98.1111; Molecular Simulations, Inc.: San
Diego, 1999.
(38) Brooks, B. R.; Bruccoleri, R. E.; Olafson, B. D.; States, D. J .;
Swaminathan, S.; Karplus, M. CHARMM: a program for mac-
romolecular energy, minimization, and dynamics calculations.
J . Comput. Chem. 1983, 4, 187-217. CHARMm, version 25.2,
revision 98.0731; Molecular Simulations, Inc.: San Diego, 1999.
(39) Kao, H. T.; Adham, N.; Olsen, M. A.; Weinshank, R. L.;
Branchek, T. A.; Hartig, P. R. Site-directed mutagenesis of a
single residue changes the binding properties of the serotonin
5-HT2 receptor from a human to a rat pharmacology. FEBS Lett.
1992, 307, 324-328.
(40) Elling, C. E.; Moeller, N.; Nielsen, S. M.; Schwartz, T. W.
Conversion of antagonist-binding site to metal-ion site in the
tachykinin NK-1 receptor. Nature (London) 1995, 374, 74-77.
(41) SPARTAN SGI, version 5.1.3; Wavefunction, Inc.: Irvine, CA,
1999.
(42) Edmonds-Alt, X.; Bichon, D.; Ducoux, J . P.; Heaulme, M.; Miloux,
B.; Poncelet, M.; Proietto, V.; Van Broeck, D.; Vilain, P.; Neliat,
G.; Soubrie´, P.; Le Fur, G.; Breliere, J . C. SR 142801, the First
Potent Non-Peptide Antagonist of the Tachykinin NK-3 Recep-
tor. Life Sci. 1995, 56, PL27-PL32.
(43) Polczewski, K.; Kumasaka, T.; Hori, T.; Behnke, C. A.; Mo-
toshima, H.; Fox, B. A.; Le Trong, I.; Teller, D. C.; Okada, T.;
Stenkamp, R. E.; Yamamoto, M.; Miyano, M. Crystal structure
of rhodopsin: A G protein-coupled receptor. Science (Washington,
D.C.) 2000, 289, 739-745.
(32) Wess, J .; Nanavati, S.; Vogel, Z.; Maggio, R. Functional role of
proline and tryptophan residues highly conserved among G
protein-coupled receptors studied by mutational analysis of the
m3 muscarinic receptor. EMBO J . 1993, 12, 331-338.
(33) (a) J avitch, J . A.; Li, X.; Kaback, J .; Karlin, A. A cysteine residue
in the third membrane-spanning segment of the human D2
dopamine receptor is exposed in the binding-site crevice. Proc.
Natl. Acad. Sci. U.S.A. 1994, 91, 10355-10359. (b) J avitch, J .
A.; Fu, D.; Chen, J . Residues in the Fifth Membrane-Spanning
Segment of the Dopamine D2 Receptor Exposed in the Binding-
Site Crevice. Biochemistry 1995, 34, 16433-16439. (c) Fu, D.;
Ballesteros, J . A.; Weinstein, H.; Chen, J .; J avitch, J . A. Residues
in the Seventh Membrane-Spanning Segment of the Dopamine
D2 Receptor Accessible in the Binding-Site Crevice. Biochemistry
1996, 35, 11278-11285. (d) J avitch, J . A.; Ballesteros, J . A.;
Weinstein, H.; Chen, J . A Cluster of Aromatic Residues in the
Sixth Membrane-Spanning Segment of the Dopamine D2 Recep-
tor Is Accessible in the Binding-Site Crevice. Biochemistry 1998,
37, 998-1006. (e) J avitch, J . A.; Ballesteros, J . A.; Chen, J .;
Chiappa, V.; Simpson, M. M. Electrostatic and Aromatic Micro-
domains within the Binding-Site Crevice of the D2 Receptor:
Contributions of the Second Membrane-Spanning Segment.
Biochemistry 1999, 38, 7961-7968.
(34) (a) Oliveira, L.; Paiva, A. C. M.; Sander, C.; Vriend, G. A common
step for signal transduction in G protein-coupled receptors.
Trends Pharmacol. Sci. 1994, 15, 170-172. (b) Sealfon, S. C.;
Chi, L.; Ebersole, B. J .; Rodic, V.; Zhang, D.; Ballesteros, J . A.;
Weinstein, H. Related contribution of specific helix 2 and 7
residues to conformational activation of the serotonin 5-HT2A
receptor. J . Biol. Chem. 1995, 270, 16683-16688. (c) Perlman,
J . H.; Colson, A.-O.; Wang, W.; Bence, K.; Osman, R.; Gershen-
gorn, M. C. Interactions between conserved residues in trans-
membrane helixes 1, 2, and 7 of the thyrotropin-releasing
hormone receptor. J . Biol. Chem. 1997, 272, 11937-11942.
(35) Blaney, F. E.; Tennant, M. Computational tools and results in
(44) Clark, T.; Alex, A.; Beck, B.; Chandrasekhar, J .; Gedeck, P.;
Horn, A.; Hutter, M.; Rauhut, G.; Sauer, W.; Steinke, T. Vamp,
version 7.0; Oxford Molecular, Ltd.: Oxford, 1998.
(45) Blaney, J . M.; Crippen, G. M.; Dearing, A.; Dixon, J . S.;
Spellmeyer, D. C. DGEOM95 is available from the Quantum
Chemistry Program Exchange, Bloomington, Indiana, 1995.
(46) Wilmot, C. M.; Thornton, J . M. â-Turns and their distortions: a
proposed new nomenclature. Protein Eng. 1990, 3, 479-493.
(47) (a) Elber, R.; Karplus, M. Enhanced sampling in molecular
dynamics: use of the time-dependent Hartree approximation for
a simulation of carbon monoxide diffusion through myoglobin.
J . Am. Chem. Soc. 1990, 112, 9161-9175. (b) Roitberg, A.; Elber,
R. Modeling side chains in peptides and proteins: application
of the locally enhanced sampling and the simulated annealing
methods to find minimum energy conformations. J . Chem. Phys.
1991, 95, 9277-9287.
(48) Freifelder, M.; Stone, G. R. Reductions with ruthenium. II. Its
use in the hydrogenation of pyridines. J . Org. Chem. 1961, 26,
3805-3808.
(49) Meyer, W. E.; Tomcufcik, A. S.; Chan, P. S.; Haug, M. 5-(1-
Piperazinyl)-1H-1,2,3,4-triazol-3-amines as antihypertensive
agents. J . Med. Chem. 1989, 32, 593-597.
the construction of
G protein-coupled receptor models. In
Membrane Protein Models; Findlay, J . B. C., Ed.; Bios Scientific
Publishers, Ltd.: Oxford, 1996; pp 161-176.
(36) Dunbrack, R. L., J r.; Karplus, M. Backbone-dependent rotamer
library for proteins. Application to side-chain prediction. J . Mol.
Biol. 1993, 230, 543-574.
(50) Pallavicini, M.; Valoti, E.; Villa, L.; Piccolo, O. Resolution of
amines with isopropylidene glycerol hydrogen phthlate. Tetra-
hedron: Asymmetry 1997, 8, 1069-1073.
J M000501V