Design and optimization of cyclized NK1 antagonists with controlled atropisomeric properties

Article abstract of DOI:10.1016/j.tet.2004.03.054

We have previously described a series of antagonists that showed high potency and selectivity for the NK₁ receptor. However, these compounds also had the undesirable property of existing as a mixture of four interconverting rotational isomers. Through biological and structural analysis of the atropisomers, a binding model was developed and used to guide the design of compounds, which were rigidified by installation of a cyclizing linkage. These compounds existed as a mixture of two atropisomers. Further elaboration of the ring system reinforced the desired conformation and eliminated atropisomeric properties. We found that the region distal to the 8-membered ring system could be modified while retaining NK₁ potency, and optimization led to further improvements in the in vivo activity.

Full text of DOI:10.1016/j.tet.2004.03.054

Tetrahedron 60 (2004) 4337–4347
Design and optimization of cyclized NK1 antagonists with
controlled atropisomeric properties
†
Jeffrey S. Albert, Cyrus Ohnmacht, Peter R. Bernstein, William L. Rumsey, David Aharony,
*
Brian B. Masek,‡ Bruce T. Dembofsky, Gerard M. Koether, William Potts and John L. Evenden
CNS Discovery Research, AstraZeneca Pharmaceuticals LP, 1800 Concord Pike, PO Box 15437, Wilmington, DE 19850-5437, USA
Received 10 October 2003; revised 12 March 2004; accepted 19 March 2004
Abstract—We have previously described a series of antagonists that showed high potency and selectivity for the NK1 receptor. However,
these compounds also had the undesirable property of existing as a mixture of four interconverting rotational isomers. Through biological and
structural analysis of the atropisomers, a binding model was developed and used to guide the design of compounds, which were rigidified by
installation of a cyclizing linkage. These compounds existed as a mixture of two atropisomers. Further elaboration of the ring system
reinforced the desired conformation and eliminated atropisomeric properties. We found that the region distal to the 8-membered ring system
could be modified while retaining NK1 potency, and optimization led to further improvements in the in vivo activity.
q 2004 Elsevier Ltd. All rights reserved.
1. Introduction
antagonist CP-96,345,7 the area has received intense
interest. This reflects the potential clinical importance that
tachykinin antagonists are viewed to have. Prompted by the
identification of the NK2 antagonist SR-48,968,8 our group
initiated a program to identify NK1 and NK2 antagonists.
These efforts led to the discovery of 1 (Fig. 1). As a result of
the napthyl 2-substituent, this compound was more than two
log units more potent as an NK1 antagonist than as an NK2
antagonist (pKB 9.6 vs. 7.3, respectively). This compound
was advanced for the treatment of depression.9
The tachykinins are a family of three mammalian
neuropeptides: substance P (SP), neurokinin A (NKA),
and neurokinin B (NKB). The preferred receptors for these
are termed NK1, NK2 and NK3, respectively. Substance P is
widely distributed in the CNS and peripheral tissue and acts
as a neurotransmitter or neuro-modulatory agent. The NK1
receptor may be involved in several pathophysiological
conditions including asthma, emesis, anxiety, depression,
and pain.1,2 The discovery of tachykinin antagonists has
been extensively reviewed.3–6
Compound 1 has two bonds with restricted rotation. The
amide bond can exist in the s-cis or s-trans forms, and the
carbonyl–aryl bond can exist in the S-axial or R-axial
Since the identification of the first non-peptidic tachykinin
Figure 1. Left: structure of 1, arrows indicate bonds with restricted rotation. Right: structure of 2, arrow indicates the naphthalene H8.
Keywords: Tachykinin; Neurokinin; Atropisomer.
*
†
Corresponding author. Tel.: þ1-302-886-4771; e-mail address: jeffrey.albert@astrazeneca.com
Present address: GlaxoSmithKline, 709 Swedeland Rd, UW2532, PO Box 1539, King of Prussia, PA 19406-0939, USA.
Present address: Optive Research, Inc., 7000 North Mopac Expressway, Austin, TX 78731, USA.
‡
0040–4020/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2004.03.054

4338
J. S. Albert et al. / Tetrahedron 60 (2004) 4337–4347
configurations. The barrier to rotation is sufficiently high
such that four atropisomers (interconverting rotational
isomers) can clearly be seen by HPLC (Fig. 2) and NMR
spectroscopy.
Figure 2. High pressure liquid chromatogram (HPLC) of 1 showing the
resolution of the atropisomeric forms (1a–1d). Conditions: column;
Phenomonex Luna C18(2) 3 mm, 4.6£75 mm; 60% methanol, 40% water
(0.1% TFA), 1.5 mL/min. UV detection at 220 nM.
Figure 3. Model for the interconversion path and structural assignments for
atropisomers 1a–1d.
antagonist potency. Combining the 8-membered ring with
the simplified piperidine led to the identification of 2;15 this
compound maintained high NK1 potency and selectivity
(Table 1). As expected, conformational properties are
simplified for compound 2 as a result of the cyclizing
tether; it exists as a mixture of only two atropisomers which
are presumably the R- and S-axial carbonyl-naphthalene
isomers.
For a candidate drug, equilibrating atropisomeric properties
would present possible safety concerns as well as very
substantial complications for manufacturing and analytical
development. To better understand the properties of 1, we
purified and independently studied each of the atropisomers
(designated 1a, 1b, 1c, and 1d) by preparative HPLC. We
observed that interconversion occurred primarily between
atropisomers 1a/1d and between 1b/1c; interconversion
between all other pairs (i.e. 1d/1b, 1b/1a, 1a/1c, and 1c/1d)
was not seen.10 Such behavior has been observed for related
naphthamide systems.11 In these cases, it is understood that
the fastest interconversions are due to the simultaneous,
paired rotation of the naphthamide–aryl and amide
bonds.11,12 This occurs because steric strain develops as
the naphthamide or amide undergoes rotation; as a result,
the rotations must occur together. This process is referred to
as ‘geared rotation.’13
Table 1. NK1 Receptor antagonist potency (pKB) for 1 and 2a
b
c
Compound
NK1
NK2
1
2
9.56^0.04
9.50^0.14
7.31^0.28
,7
a
pKB Determinations using rabbit pulmonary artery tissue (n¼2–6).
Agonist: Ac-(Arg6, Sar9, Met(O2)11) SP6–11 (ASMSP).
Agonist: BANK (b-ala8NKA(4-10)).
b
c
The long rotational half-life (approximately 1.8 days at
37 8C) allowed us to individually test each component of
1.10 We found that most of the in vivo NK1 activity was
associated with the single atropisomer, 1d. Using a
combination of kinetic and spectroscopic studies, we
assigned structures for each of the atropisomers as indicated
in Figure 3.10
After demonstrating that we could reduce the number of
isomers from four to two, our next goal became the
stabilization of a single conformation. In this way, all
potential concerns about atropisomerism would be elimi-
nated. This paper describes the modeling and structural
analysis of these 8-membered ring antagonists. In these
studies, we have developed a series of novel NK1
antagonists that exist in predominantly a single confor-
mation. Furthermore, by comparing these compounds, we
have constructed a detailed pharmacophore model for NK1
receptor binding.
Based on the structural model for 1d, we reasoned that we
could enforce the single, desired conformation by intro-
ducing a tether to connect the methyl group of the amide
with the naphthalene 2-position. Such a tether should
stabilize the NK1 binding conformation because the amide
would be locked in the trans configuration. A similar
approach has been demonstrated for a structurally unrelated
series of NK1 antagonists.14
2. Synthesis
Secondary and tertiary amine compounds could be con-
veniently prepared from aldehyde 315 or 415 by reductive
amination in the presence of sodium cyanoborohydride.
Alcohol and carboxylic acid derived compounds were
prepared by reduction or oxidation of 4 as indicated in the
Scheme 1.
We analyzed compounds with tethering ring systems
containing 6, 7, and 8-membered rings and found that the
8-membered ring system was preferred.15 Prior studies had
shown that the piperidine substituent in 1 could be
simplified15 without significant loss of NK1 receptor

J. S. Albert et al. / Tetrahedron 60 (2004) 4337–4347
4339
Scheme 1. For R2, R3, R5, R6 see Table 4.
3. Results
constrain the torsional angles in the 8-membered ring in a
manner similar to that for the aryl and amide bonds in 2.
Thus, COD serves as a convenient starting point because its
conformational properties have been extensively
studied,16–19 both theoretically and experimentally.
To better understand the factors responsible for the
atropisomeric properties, molecular modeling studies were
conducted for the cyclic analogs. With 2 as a starting point,
and with knowledge of the rotational properties of related
naphthamides, we sought to eliminate the atropisomeric
properties. A detailed analysis of the ground state energetics
was carried out for compounds in the cyclic series to explore
this possibility.
Previous studies indicate three relevant conformations for
COD, a twist boat chair (TBC) conformation, and the two
twist boats (TB1 and TB2) conformations (Fig. 5). We
began by verifying that the conformational energetics
calculated by our AESOP-Enigma program20 were quanti-
tatively consistent with empirical data and with energies
determined by other theoretical methods. Next, we
examined the energetics of these same conformations in
the more elaborate cyanobenzamide and cyano-
naphthamide-based ring systems (Fig. 5). The TBC
conformers remained the most stable for both the cyano-
benzamide and cyanonaphthamide adducts. Interestingly,
both TB1 and TB2 conformers are substantially destabilized
in these systems relative to COD. It is important to note that
3.1. Conformational analysis of the 8-membered ring
system
To more fully understand the conformational properties of 2
and related compounds, we began by considering simpler
model systems. Analysis of the possible conformations of
the 8-membered ring was carried out by first considering
1,3-cyclo-octadiene (COD) (Fig. 4). As a first approxi-
mation, the cis double bonds in COD are expected to

4340
J. S. Albert et al. / Tetrahedron 60 (2004) 4337–4347
Figure 4. Progressive simplification from the 8-membered ring system in 2 to the simplified naphthalene analog, to the simplified benzamide analog, to COD.
Figure 5. Relative conformational energies (kcal/mol) for COD, and cyanobenzamide and cyanonaphthamide model systems. The sign of the torsional angles
for each bond in the 8-membered ring are indicated with (2) and (þ).
all three conformations (TBC, TB1 and TB2) are chiral; a
corresponding enantiomeric conformation exists for each of
these conformations. (The enantiomeric conformation has
an inversion of the sign of each of the torsional angles.)
These pairs of enantiomeric conformations (denoted A and
B below) are energetically equivalent in an achiral
environment.
3.2. Conformational analysis including the pendant
phenethyl side chain
As the next step in progressively elaborating from our
simple starting model, a pendant phenethyl group (Fig. 6)
was added to model the dichloroaryl region of compound 2.
Because the phenethyl group contains a chiral atom of

J. S. Albert et al. / Tetrahedron 60 (2004) 4337–4347
4341
3.3. Structural assignments of the 8-membered ring
system
Experimentally, we observe two atropisomers for 2 with a
population distribution of about 1:2 according to HPLC and
1H NMR spectroscopy (Fig. 8). This implies that the
conformational energies must be similar. For 2, NMR
spectra show that each of the two atropisomers are resolved
for many of the protons. However, the separation of the
corresponding signals is particularly striking for
the naphthalene H8 proton. For the minor atropisomer, the
naphthalene H8 proton is in the expected region at about
7.4 ppm. (It is obscured by overlapping signals but can be
identified in two dimensional NMR experiments). For the
major atropisomer, the naphthalene H8 proton is shifted
upfield to about 6.4 ppm.
Figure 6. TBC conformations of the napthamide system with pendant
chiral phenyl containing side chain used for computational analysis. The
two structures differ by inversion of the torsion angles around each atom in
the 8-membered ring; the sign of the torsional angles for each bond in the 8-
membered ring are indicated with (2) and (þ).
known fixed absolute stereochemistry, the (previously
enantiomeric) conformers of the 8-membered ring become
diastereomeric and thus both forms must be considered
(Fig. 6). Prior studies had shown that NK1 and NK2 affinity
are sensitive to the stereochemistry of the aryl methine
carbon, with the S-enantiomer being preferred.21 Therefore,
all subsequent experimental and computational studies
focused on this form.
For each ring conformation (2 TBC, 2 TB1 and 2 TB2)
exhaustive conformational scanning of the pendant side
chain was performed using Enigma-ConfScan.20 The
phenyl side chain does not appear to perturb the confor-
mational preferences of the 8-membered ring; the confor-
mations containing TB1 or TB2 ring conformations
remained 3–4 kcal/mol higher in energy than those
containing TBC conformations for the 8-membered ring
system. Two conformers were found to be very close in
energy; they are designated as A and B (Fig. 7). These
conformers are based on TBC ring geometries of ‘opposite’
chirality as well as differing side chain geometries. The A
conformer was favored by 0.7 kcal/mol over the B
conformer. From this we conclude that the actual energies
should be close to each other; the small predicted energy
difference is within the uncertainty of the calculation.
Figure 8. 1H NMR spectrum of 2 in d6-DMSO.
In 1H NMR spectra of 2, an upfield shift might be expected
for the ring conformer A because the H8 proton is oriented
into the shielding region of the phenyl ring (Fig. 7). For the
B conformation, the H8 proton is distant from the phenyl
ring, and no shift would be expected. The phenethyl side
chain does not appear to perturb the conformational
preferences of the 8-membered ring; conformational scan-
ning indicated that for B, no stable conformation exists
(,3 kcal/mol) for the pendant phenyl ring which would
allow interaction of the naphthalene H8 proton. The
atropiosmeric distribution and NMR spectroscopic proper-
ties of 2 are fully consistent with the structures and energies
predicted by modeling. Based on this evidence, we assigned
conformations to the atropisomers as indicated in Figure 7.
Prior studies on 1 demonstrated that it was atropisomer 1d
which was preferred for NK1 antagonism10 (Fig. 3). The
conformation of 1d overlays well with conformer A of the
8-membered system. In particular, they share the same
orientation of the amide and naphthalene–carbonyl bonds,
and they both have a potential aryl–aryl stacking interaction
between the dichloroaryl ring and the naphthalene ring. For
these reasons, we assign conformer A as the NK1 preferring
analogue.
3.4. Rigidification of the 8-membered ring by additional
substitution
Figure 7. Relaxed-eye stereoview representation of the two low energy
conformers of the model compounds from Figure 6. Conformation A shows
an edge-to-face aryl–aryl stacking interaction (green line) which is
hypothesized in the NK1-active atropisomer. Conformation B cannot
accommodate the stacking interaction and is hypothesized to be the less
active atropisomer. Hydrogen atoms have been omitted for clarity.
Despite the potent in vitro activity observed for 2 (Table 1)
its existence as a mixture of two atropisomers limited
further consideration of this compound for drug develop-
ment. Kinetic analysis showed that the interconversion

4342
J. S. Albert et al. / Tetrahedron 60 (2004) 4337–4347
half-life between the two atropisomers was 9.7 h (37 8C,
50% acetonitrile/water, 25 mM phosphate, pH 7.0; data not
shown). This rate is fast enough that it would be impractical
to isolate and administer the compound as a single
atropisomer. We sought to eliminate the atropisomeric
properties while further improving potency.
Table 2. Inhibition of ASMSP-induced foot tapping response in gerbil and
receptor antagonist potency (pKB) in rabbit pulmonary artery tissue
a
pKB
% Inhibition
of GFT
responseb
Isomer
distribution;
shifted/unshiftedc
5
6
7
8
9
9.10^0.03
8.63^0.13
8.64^0.09
8.29^0.15
8.34^0.24
38^11
3^1
8^3
23^13
63^23
68:32
.98:2
10:90
31:68
.98:2
We speculated that addition of a methyl substituent in the
8-membered ring might perturb the energies enough to
stabilize one conformer with respect to the other con-
former.14 Building upon the computational analysis
described above, we analyzed the energies of each of the
methyl substituted analogs in Figure 9. These modeling
studies were used to aid conformational analysis; we have
not attempted to extend the analysis for the interpretation of
in vivo results which would require the addition of
parameters to account for distribution, metabolism and
other factors. In each of the examples, we found two
predicted low energy conformers of the 8-membered ring
system, which were analogous to the A and B conformations
of the unsubstituted case. However, the predicted energy
differences between the A and B conformers were small.
Therefore, we could not reliably predict which position for
substitution would most optimally stabilize the desired
A-type conformation.
a
pKB Determinations using RPA tissue (n¼2–6), agonist: ASMSP.
Determined 4 h after oral dosing of antagonist at 5 mmol/kg and initiated
by CNS administration of ASMSP (100 pmol).
Isomer distribution determined by HPLC. Assignments (‘shifted’ and
‘unshifted’) refer to the position of the naphthalene H8 signal in NMR
spectra.
b
c
For 5–9, the distribution and assignment (H8-shifted and
non-shifted) of the conformational forms are indicated in
Table 2. It was evident that 7 and 8 existed as a mixture of
atropisomers, and the conformational equilibrium favored
the H8-nonshifted (and hence NK1 non-preferred) confor-
mation. In comparison, 6 and 9 existed predominantly as a
single atropisomer in the H8 shifted conformation.
Detailed characterization of the four isomers will be
presented separately, but key observations are summarized
here. Central administration of ASMSP to gerbils induces a
foot tapping (GFT) response which can be attenuated by
prior dosing with an orally available CNS-penetrant, NK1
receptor antagonist. This can provide a convenient way to
assess potency of NK1 receptor antagonists.22 Gerbils were
orally treated with antagonist at 5 mmol/kg at 4 or 6 h prior
to administration of agonist. The CNS potency was then
quantified by comparing the foot tapping response with
control animals with vehicle only. Among the four isomers,
compound 9 was the most potent in the GFT model
(Table 2). Furthermore, it was more potent than the
unsubstituted 5. We were fortunate that this was one of
the two compounds to exist as a single atropisomer; this
could not have been predicted a priori.
Pharmacological activity was further assessed in vitro using
pulmonary artery isolated from rabbit (RPA). Rabbit
neurokinin receptors are homologous to human. It is
noteworthy that while 9 is the most potent isomer in the
in vivo model, it does not show the highest potency (pKB) in
the in vitro functional model. While each of the methyl-
substituted isomers have similar RPA potencies (pKB 8.3–
8.6), the potency is clearly greater for 5 (pKB 9.1).
Obviously, receptor binding results from multiple inter-
actions, and it is likely that addition of the methyl groups in
6–9 introduces some unfavorable interactions in addition to
altering the atropisomeric distribution. The apparent
discrepancy between the in vivo and in vitro potency for 5
and 9 cannot be attributed to overall drug exposure since
gerbil dose-normalized exposure levels in brain are similar
between 5 (1478 ng/g following 100 mmol/kg oral dose) and
9 (224 ng/g following 30 mmol/kg oral dose).23 Possible
explanations include differences between the rabbit and
gerbil receptors, or differences in exposure levels at specific
brain regions. Pharmacokinetic parameters for 9 are shown
in Table 3.
Figure 9. Structures of 5–9.
Each of the four methyl substituted analogs 6–9 were
prepared and the atropisomeric distributions were analyzed
by HPLC. As described above, the NK1-preferential
conformers of 1 and 2 showed a characteristic upfield
shift in NMR spectra for the naphthalene H8 proton. We
interpret this shift as resulting from an aryl–aryl interaction
when the compound adopts a conformation like that shown
in Figure 7(A). Therefore, we expected that the NK1
preferred conformers in 6–9 would show a similar upfield
shift for the naphthalene H8 proton. We determined which
peaks in the HPLC chromatograms corresponded to the
H8-shifted and non-shifted conformations by analysis of
NMR integrations and comparison with HPLC integrations
(Table 2). By extension of this, we assigned the one with the
upfield shifted H8 resonance as the NK1 preferred
conformer. This assumption allowed us to expedite
compound evaluation by avoiding the need to separate
and test each atropisomer for each compound.

J. S. Albert et al. / Tetrahedron 60 (2004) 4337–4347
4343
Table 3. Pharmacokinetic analysis of 9 in rat (n¼2)
3.6. Structural analysis and pharmacophore model for
13
Pharmacokinetic parameter
Result
Unfortunately, we were unable to crystallize 13, however;
we succeeded for the analog 17 (Fig. 10). Crystallographic
analysis showed excellent correspondence with the assigned
structure for the NK1-preferential atropisomer 1d (Fig. 3)
and modeling of the 8-membered ring system (Fig. 7). In
particular, the crystal structure shows the 8-membered ring
adopting the TBC configuration with the naphthalene
Oral dose (mmol/kg)
Formulation
Cmax (nM)
Tmax (h)
AUC-PO(0-i) (ng h/mL)
Bioavailability (%)
20
75% PEG/Saline
265
7
2370
25
˚
oriented such that the H8 proton is 3.1 A from the
dichloroaryl ring centroid. This close proximity is consistent
with the upfield shift observed for this proton in NMR
spectroscopy. Based on the potent in vivo activity and the
structural information, we propose an NK1 pharmacophore
model involving an edge-to-face aryl–aryl stacking inter-
action and a polar group served by the amide carbonyl in the
8-membered ring system. The exocyclic methyl group may
serve to stabilize the desired conformational form, but may
also introduce penalizing interactions with the NK1
receptor. Finally, we note that there is substantial tolerance
to substitution at the left-side region in this series which
could therefore be used to further optimize distribution and
metabolic properties.
3.5. Optimization of compounds based on the ring
structure in 9
Compound 9 had excellent potency in the gerbil in vivo
model and existed in predominantly a single conformational
form. Based on this, we sought to further increase potency in
compounds containing the same 8-membered ring system.
We surveyed analogs of 9 in which the ‘left side’ dimethyl
amine group was replaced with groups of varying properties
(Table 4). We observed that antagonist potency (pKB) was
not greatly affected over the range of substituent groups,
which included basic, acidic, hydrophobic, and polar
functionality. Additionally, we found that antagonist in
vitro potency (pKB) did not directly correlate with in vivo
potency in the GFT model; this most likely reflects
substantial differences in distribution and metabolism. For
example, 9 had much greater potency than 11 in the in vivo
GFT model (63% vs 3% inhibition at 4 h); whereas 9 had
weaker potency than 11 in the in vitro model (pKB 8.2 vs
9.0). The most active compound identified was 13. It
showed excellent potency with a very long duration of
action (81% inhibition of response after 6 h).
Figure 10. Crystal structure of 17. The dichloroaryl ring exists both in the
orientation shown, and with rotation such that 3-chloro group is on the
‘alternate’ 3-carbon. The structure is shown as a relaxed-eye stereoview
with hydrogen atoms omitted for clarity.
Table 4. In vitro potency (pKB) in RPA and inhibition of ASMSP-induced
foot tapping response in gerbil
4. Summary and conclusions
In earlier work, we developed a series of potent and
selective NK1 antagonists. However, this series had the
undesirable property of existing as a mixture of four
atropisomers. Structural analysis and comparisons of the
activities of these atropisomers allowed us to construct an
NK1 pharmacophore model. This binding model was then
used to design a new series in which the naphthalene amide
system was constrained through a tether, thus forming an
8-membered ring. As a result of this constraint, the
conformational features were simplified, and resulting
compounds existed as a mixture of only two atropisomers.
Through detailed computational analysis we constructed a
model for both atropisomers and assigned a structure for the
active conformer. In the conformational mixture, we found
the NK1 preferred component had a characteristic shift in
the H8 naphthalene proton in NMR spectroscopy. The shift
was suggestive of the orientation of the naphthalene, and it
provided a convenient way to assign provisional confor-
mations and population distributions to the mixture of
atropisomers.
a
R
pKB
GFT, 4 hb
GFT, 6 hb
9
Me2NCH2–
MeNH–
(N-Piperidinyl)–CH2–
Me3CNCH2–
MeONHCH2–
HOCH2–
8.3^0.1
9.3^0.1
9.0^0.1
7.8^0.1
8.2^0.1
8.6^0.3
7.9^0.1
8.2^0.2
8.8^0.1
9.0^0.3
8.2^0.3
63^23
5^8
3^1
1^0.6
73^6
15^7
35^5
1^0.3
4^2
52^24
9^8
3^2
38^31
81^19
6^3
63^13
1^0.8
3^1
4^2
38^31
10
11
12
13
14
15
16
17
18
19
ClCH2–
HOC(O)–
H2NC(O)–
MeNHC(O)–
Me2NC(O)–
4^2
17^12
a
pKB Determinations using rabbit pulmonary artery (RPA) tissue
(n¼2–6), agonist: ASMSP.
b
% Inhibition of gerbil foot tap (GFT) response determined after oral
dosing of antagonist at 5 mmol/kg at 4 or 6 h (as indicated) prior to
initiation by CNS administration of ASMSP (100 pmol).
Our key goal was to eliminate atropisomeric properties
completely. Thus, we modified the 8-membered ring system

4344
J. S. Albert et al. / Tetrahedron 60 (2004) 4337–4347
of 5 by placing a methyl substituent on each of the four
positions of the ethyl segment of the ring (6–9) and
compared the atropisomeric distributions. As expected, the
addition of the methyl group did influence the atropisomeric
distribution, and two of the four isomers existed in
predominantly a single conformational form. Among these
two, 9 was the most potent in our in vivo model for CNS
activity.
otherwise noted. Mass spectral data were obtained on a
Micromass QTOF mass spectrometer. Silica gel chroma-
˚
tography was performed with ICN silica 32–63, 60 A. Thin-
layer chromatography was done on silica gel 60 F-254
(0.25 mm thickness) plates, and visualization was accom-
plished with UV light. Unless otherwise noted, all materials
were obtained commercially and used without further
purification. For compounds which exist as a mixture of
1
atropisomers, H NMR spectra and HPLC chromatograms
1
are complex. In these cases, H NMR integrations are not
Starting from the ring system in 9, we made alterations to
the left side region in attempt to further improve activity. In
this manner, we identified 13, which showed excellent
potency in the in vivo model and long duration of action.
Crystallographic analysis of 17 (a close analogue of 13)
confirmed the structural features which had been predicted
from structural analysis and modeling. Based on these
results, we propose a binding conformation and NK1
pharmacophore model which involves an edge-to-face
aryl–aryl stacking interaction and a hydrogen bonding
acceptor group.
given. Mass spectra were acquired using atmospheric
pressure chemical ionization. High resolution mass spectra
were acquired using electrospray (ES). The synthesis of
compounds 3, 4, 6, 7, 8, and 9 are described separately.15
5.5.1. 2-[(2S)-2-(3,4-Dichlorophenyl)-4-(diethylamino)-
butyl]-1-oxo-1,3,4,6-tetrahydro-2H-naphtho[1,2-f][1,4]-
oxazocine-7-carbonitrile (2). Aldehyde 326 (100 mg,
0.21 mmol) was dissolved in methanol (2 mL) under a
nitrogen atmosphere. To this was added diethylamine
hydrochloride (35 mg, 0.32 mmol) and triethylamine
(34 mL, 0.25 mmol). Acetic acid was added dropwise until
the pH was between 4 and 5. The mixture was stirred for
10 min, then 22 mg (0.36 mmol) sodium cyanoborohydride
was added as a solution in methanol (approx. 1 mL). After
stirring for 1.5 h the pH was adjusted to approximately 5.5
by addition of triethylamine and stirring was continued
overnight. The mixture was then concentrated, diluted with
ethyl acetate (20 mL), washed with water then brine (20 mL
each), dried over magnesium sulfate, filtered and concen-
trated. The remaining residue was then purified via reverse
phase HPLC (using a gradient of acetonitrile in water) to give
the title compound (80 mg, 60%) as a gum, dH (300 MHz, d6-
DMSO) 9.09 (s), 8.68 (s), 8.62 (s), 8.10 (d, J¼7.7 Hz), 8.02 (d,
J¼8.2 Hz), 7.90 (d, J¼8.3 Hz), 7.70 (m), 7.42 (m), 6.44 (d,
J¼8.5 Hz), 4.81 (m), 4.34 (m), 4.00 (t, J¼11.3 Hz), 3.45 (m),
2.71 (m), 2.08 (s), 1.13 (t, J¼7.1 Hz); m/z (APCI) 524 MHþ.
5. Experimental
5.1. Molecular modeling
Molecular mechanics computations were performed in
vacuo, using the AESOP20 force field with full geometry
optimization of all conformations examined and results
were visualized using the in-house molecular graphics
program ENIGMA.24
5.2. Biological studies
Isolated tissue response (pKB), studies were carried out as
previouslydescribed.25 Inhibition ofgerbilfoottapping (GFT)
response studies were carried out as previously described.22
5.5.2. 2-[(2S)-2-(3,4-Dichlorophenyl)-4-(dimethyl-
amino)butyl]-1-oxo-1,3,4,6-tetrahydro-2H-naphtho[1,2-
f][1,4]oxazocine-7-carbonitrile (5). Aldehyde 326
(150 mg, 0.32 mmol) was dissolved in methanol (3 mL)
under a nitrogen atmosphere. To this was added dimethyl-
amine hydrochloride (40 mg, 0.48 mmol) and triethylamine
(50 mL, 0.35 mmol). Acetic acid was added dropwise until
the pH was between 4 and 5. The mixture was stirred for
0.5 h, then 34 mg (0.54 mmol) sodium cyanoborohydride
was added as a solution in methanol (approx. 1 mL). After
stirring for 2 h the mixture was then concentrated, diluted
with ethyl acetate (30 mL), washed with water then brine
(30 mL each), dried over magnesium sulfate, filtered and
concentrated. The remaining residue was then purified silica
gel chromatrography (6–10% MeOH/CH2Cl2) to afford the
title product (120 mg, 80%) a white solid, dH (300 MHz, d6-
DMSO) 8.68 (s), 8.62 (s), 8.09 (d, J¼7.8 Hz), 8.01 (d,
J¼8.1 Hz), 7.90 (d, J¼8.3 Hz), 7.82–7.59 (m), 7.40 (m),
6.43 (d, J¼8.5 Hz), 4.87 (m), 4.71 (t, J¼12.0 Hz), 4.32 (m),
3.99 (t, J¼11.1 Hz), 3.89–3.64 (m), 3.42–2.94 (m), 2.83–
2.55 (m), 2.10 (s); m/z (APCI) 496 MHþ.
5.3. Drug exposure analysis
Studies were conducted in fed mongolian gerbils. The
animals received 100 mmol/kg drug by oral administration
in a dosing volume of 6 mL/kg in a 75% PEG400/saline
solution. The gerbils were euthanized at 4 h after dosing by
CO2 inhalation and brain tissue samples were collected.
Samples were stored at 220 8C or below until analysis.
Brain homogenates were prepared in a tissue homogenizer
using four milliliters of saline per gram of tissue. Samples
were processed for analysis by the addition of acidified
acetonitrile. The resulting extracts were analyzed by LC/MS.
5.4. Bioavailability analysis
Compounds were orally administered to rat (n¼2) as a
solution in 75% PEG400/saline solution. Blood samples
were taken via surgically implanted cannula or by
venipuncture over a 24 h period and plasma was analyzed
for unchanged compound by LC/MS.
5.5. General chemistry methods
5.5.3. (4R)-2-[(2S)-2-(3,4-Dichlorophenyl)-4-(methyl-
amino)butyl]-4-methyl-1-oxo-1,3,4,6-tetrahydro-2H-
naphtho[1,2-f][1,4]oxazocine-7-carbonitrile (10). Aldehyde
1H NMR spectra were obtained at 300 MHz using a Bruker
DPX 300 spectrometer and were referenced to TMS unless

J. S. Albert et al. / Tetrahedron 60 (2004) 4337–4347
4345
415 (70 mg, 0.145 mmol) was dissolved in 1 mL of THF and
methylamine in THF (2 M, 87 mL, 0.174 mmol) was added.
Themixturewasdilutedwithmethanol(2 mL), andthenacetic
acid (5 mL) was added and the mixture stirred for 10 min. A
solution of sodium cyanoborohydride (1 M in THF, 0.5 mL,
0.500 mmol) was added and the reaction was stirred overnight
at ambient temperature. The reaction mixture was evaporated
and the residue purified by preparative HPLC using a
Phenomenex LUNA C-18(2), 250£21.2 mm (10 m) column
eluting with acetonitrile–water gradient containing0.1% TFA
(40–70% acetonitrile over 20 min). Product containing
fractions were pooled and partially concentrated to remove
acetonitirile. The remaining aqueous solution was made basic
by addition of 10% aqueous sodium carbonate, and the
solution extracted with ethyl acetate (3 times, 20 mL). The
organic extracts were dried (Na2SO4), filtered and evaporated
to afford the title compound (40 mg, 55%) as a white foamy
solid, dH (300 MHz, d6-DMSO) 8.66 (s, 1H), 8.30 (1H, br),
8.05 (1H, d, J¼8.3 Hz), 7.67 (3H, m), 7.42 (2H, m), 6.54 (1H,
d,J¼8.3 Hz),4.95(1H, d,J¼13.8 Hz),4.71(1H, m),4.51(1H,
d, J¼13.8 Hz), 4.03(1H, m), 3.24(1H, m), 2.99–2.53(6H, m),
2.37 (2H, m), 2.01 (2H, m), 1.10 (3H, d, J¼6.1 Hz); m/z
(APCI) 496 MHþ.
m/z (APCI) 538 MHþ; HRMS (ES) Mþ, found 538.2051.
C30H33Cl2N3O2 requires 538.2028.
5.5.6. (4R)-2-[(2S)-2-(3,4-Dichlorophenyl)-4-(methoxy-
amino)butyl]-4-methyl-1-oxo-1,3,4,6-tetrahydro-2H-
naphtho[1,2-f][1,4]oxazocine-7-carbonitrile (13). Alde-
hyde 415 (70 mg, 0.145 mmol) was dissolved THF (1 mL)
and O-methylhydroxylamine hydrochloride (14 mg,
0.167 mmol) was added. The mixture was diluted with
methanol (2 mL), then acetic acid (5 mL) was added and the
mixture stirred for 10 min. A solution of sodium cyano-
borohydride (1 M in THF, 0.5 mL, 0.500 mmol) was added
and the reaction was stirred overnight at ambient tempera-
ture. The reaction mixture was evaporated and the residue
purified by preparative HPLC according to the procedure
described for 10 to afford the title compound as a white
foamy solid (21 mg, 28%), dH (300 MHz, CDCl3) 8.25 (1H,
s), 8.81 (1H, d, J¼7.4 Hz), 7.60–7.46 (3H, m), 7.39 (1H, d,
J¼1.7 Hz), 7.31–7.25 (1H, m), 6.68 (1H, d, J¼8.3 Hz),
5.53 (1H, br), 5.14 (1H, d, J¼13.6 Hz), 4.02 (1H, t,
J¼13.7 Hz), 4.56 (1H, d, J¼13.7 Hz), 3.96 (1H, m), 3.52
(3H, s), 3.24–3.02 (4H. m), 2.80 (2H, m), 1.96 (2H, m), 1.17
(3H, d, J¼6.6 Hz); m/z (APCI) 512 MHþ; HRMS (ES) Mþ,
found 512.1343. C27H27Cl2N3O3 requires 512.1508.
5.5.4. (4R)-2-[(2S)-2-(3,4-Dichlorophenyl)-4-piperidin-1-
ylbutyl]-4-methyl-1-oxo-1,3,4,6-tetrahydro-2H-
naphtho[1,2-f][1,4]oxazocine-7-carbonitrile (11). Alde-
hyde 415 (70 mg, 0.145 mmol) was dissolved in THF
(1 mL) and piperidine (15 mg, 0.176 mmol) was added.
The mixture was diluted with methanol (2 mL), then acetic
acid (5 mL) was added and the mixture stirred for 10 min. A
solution of sodium cyanoborohydride (1 M in THF, 0.5 mL,
0.500 mmol) was added and the reaction was stirred
overnight at ambient temperature. The reaction mixture
was evaporated and the residue purified by preparative
HPLC according to the procedure described for 10 to afford
of the title compound as a white foamy solid (50 mg, 62%),
dH (300 MHz, d6-DMSO) 8.68 (1H, s), 8.39 (1H, br), 8.04
(1H, d, J¼7.9 Hz), 7.68 (3H, m), 7.43 (2H, m), 6.55 (1H, d,
J¼8.3 Hz), 4.96 (1H, d, J¼13.8 Hz), 4.72 (1H, m), 4.51
(1H, d, J¼13.8 Hz), 4.04 (1H, m), 3.02–2.52 (8H, m), 2.05
(2H, m), 1.09 (3H, d, J¼6.6 Hz), 0.96 (1H, m), 0.55 (2H.
m), 0.30 (2H, m); m/z (APCI) 550 MHþ; HRMS (ES) Mþ,
found 550.2007. C31H33Cl2N3O2 requires 550.2028.
5.5.7. (4R)-2-[(2S)-2-(3,4-Dichlorophenyl)-4-hydroxy-
butyl]-4-methyl-1-oxo-1,3,4,6-tetrahydro-2H-
naphtho[1,2-f][1,4]oxazocine-7-carbonitrile (14). Alde-
hyde 415 (250 mg, 0.52 mmol) was dissolved in methanol
(5 mL) under a nitrogen atmosphere and cooled to 5 8C.
Sodium borohydride (22 mg, 0.57 mmol) was then added,
the cooling bath removed, and the reaction allowed to stir
for 15 min. It was then concentrated, diluted with ethyl
acetate, washed with water then brine, dried over mag-
nesium sulfate, filtered and concentrated. The residue was
then purified via silica gel chromatography (60–80% ethyl
acetate/hexanes) to give the title compound as a foamy solid
(200 mg, 80%), dH (300 MHz, CDCl3) 8.25 (1H, s), 7.82
(1H, d, J¼7.7 Hz), 7.54 (3H, m), 7.41 (1H, d, J¼2.0 Hz),
7.29 (1H, m), 6.69 (1H, d, J¼8.3 Hz), 5.15 (1H, d,
J¼13.8 Hz), 4.94 (1H, t, J¼12.4 Hz), 4.57 (1H, d,
J¼13.8 Hz), 3.96 (1H, m), 3.70 (1H, m), 3.47 (1H, m),
3.29 (2H, m), 3.14 (3H, m), 1.98 (2H, m), 1.17 (3H, d,
J¼6.5 Hz); m/z (APCI) 438 MHþ; HRMS (ES) Mþ, found
483.1223. C26H24Cl2N2O3 requires 483.1242.
5.5.5. (4R)-2-[(2S)-4-(tert-Butylamino)-2-(3,4-dichloro-
phenyl)butyl]-4-methyl-1-oxo-1,3,4,6-tetrahydro-2H-
naphtho[1,2-f][1,4]oxazocine-7-carbonitrile (12). Alde-
hyde 415 (70 mg, 0.145 mmol) was dissolved THF (1 mL)
and tert-butylamine (13 mg, 0.177 mmol) was added. The
mixture was diluted with methanol (2 mL), then acetic acid
(5 mL) was added and the mixture stirred for 10 min. A
solution of sodium cyanoborohydride (1 M in THF, 0.5 mL,
0.500 mmol) was added and the reaction was stirred
overnight at ambient temperature. The reaction mixture
was evaporated and the residue purified by preparative
HPLC according to the procedure described for 10 to afford
the title compound as a white foamy solid (50 mg, 64%), dH
(300 MHz, d6-DMSO) 8.67 (1H, s), 8.27 (1H, br), 8.05 (1H,
d, J¼8.3 Hz), 7.72–7.64 (3H, m), 7.43 (2H, m), 6.60 (1H, d,
J¼8.7 Hz), 4.96 (1H, d, J¼14.0 Hz), 4.73 (1H, m), 4.52
(1H, d, J¼14.0 Hz), 4.05 (1H, m), 2.94 (3H, m), 2.75–2.53
(3H, m), 2.02 (2H, m), 1.21 (9H, s), 1.10 (3H, d, J¼6.5 Hz);
5.5.8. (4R)-2-[(2S)-4-Chloro-2-(3,4-dichlorophenyl)-
butyl]-4-methyl-1-oxo-1,3,4,6-tetrahydro-2H-
naphtho[1,2-f][1,4]oxazocine-7-carbonitrile (15). Alcohol
14 (80 mg, 0.17 mmol) was dissolved in toluene (3.5 mL)
under a nitrogen atmosphere and to this was added
triphenylphosphine (48 mg, 0.18 mmol), then hexachloro-
acetone (264 mg, 0.42 mmol) over one min. as a solution in
toluene. The reaction was allowed to stir for 1 h, heated briefly
to 60 8C, cooled and allowed to stir overnight at room
temperature. A second portion of triphenylphosphine (10 mg)
and hexachloroacetone (2 drops) were added and reaction
again heated briefly to 60 8C, cooled, then concentrated. The
remaining residue was purified via silica gel chromatography
(30–40% ethyl acetate/hexanes) to give the title compound as
a foam solid (80 mg, 96%), dH (300 MHz, CDCl3) 8.26 (1H,
s), 7.83 (1H, d, J¼7.8 Hz), 7.54 (3H, m), 7.43 (1H, d,
J¼2.0 Hz), 7.28 (1H, m), 6.73 (1H, d, J¼8.4 Hz), 5.17 (1H, d,

4346
J. S. Albert et al. / Tetrahedron 60 (2004) 4337–4347
J¼13.8 Hz), 4.95 (1H, t, J¼12.2 Hz), 4.58 (1H, d,
J¼13.8 Hz), 3.99 (1H, m), 3.60 (1H, m), 3.23 (5H, m), 2.18
(2H, m), 1.19 (3H, d, J¼6.5 Hz); m/z (APCI) 501 MHþ;
HRMS (ES) Mþ, found 501.0905. C26H23Cl3N2O2 requires
501.0903.
dichlorophenyl)-N-methylbutanamide (18). Carboxylic
acid 16 (80 mg, 0.16 mmol) was dissolved in dichloro-
methane (2 mL) under a nitrogen atmosphere; to this was
added oxalyl chloride (28 mL, 0.32 mmol) and one drop of
dimethylformamide. After stirring for 0.5 h, the mixture
was concentrated, redissolved in dichloromethane (1 mL)
and concentrated again. The residue was dissolved in
dichloromethane (2 mL) and to this was added methylamine
(2 M solution in THF, 120 mL, 0.24 mmol), and triethyl-
amine (32 mL, 0.24 mmol). The mixture was stirred for 1 h,
concentrated and purified using reverse phase HPLC to give
the title compound as a white solid (60 mg, 60%), dH
(300 MHz, CDCl3) 8.26 (1H, s), 7.83 (1H, d, J¼7.8 Hz),
7.53 (4H, m), 7.26 (1H, m), 6.67 (1H, d, J¼8.4 Hz), 5.74
(1H, d, J¼4.1 Hz), 5.16 (1H, m), 4.93 (1H, m), 4.56 (1H,
m), 3.98 (1H, m), 3.65 (1H, m), 3.27 (1H, m), 3.12 (2H, m),
2.78 (d, J¼4.8 Hz), 2.62 (2H, m), 1.18 (3H, d, J¼6.5 Hz;
m/z (APCI) 510 MHþ; HRMS (ES) Mþ, found 510.1356.
C27H25Cl2N3O3 requires 510.1351.
5.5.9. (3S)-4-[(4R)-7-Cyano-4-methyl-1-oxo-1,3,4,6-
tetrahydro-2H-naphtho[1,2-f][1,4]oxazocin-2-yl]-3-(3,4-
dichlorophenyl)butanoic acid (16). Aldehyde 415
(850 mg, 1.77 mmol) was dissolved in acetone (10 mL)
under a nitrogen atmosphere and cooled to 0 8C. Jones
reagent (2 M, 1.3 mL, 2.65 mmol) was added as a solution
in acetone (4 mL) over 5 min. The reaction was stirred for
5 min, then 2-propanol (3 mL) was added and stirred for
5 min. The mixture was concentrated, diluted with ethyl
acetate (15 mL), washed with water then brine (15 mL
each), dried over magnesium sulfate, filtered and concen-
trated to give the title compound as a foamy solid (850 mg,
97%), dH (300 MHz, CDCl3) 8.27 (1H, s), 7.84 (1H, d,
J¼7.8 Hz), 7.55 (3H, m), 7.42 (1H, d, J¼2.0 Hz), 7.29
(1H, m), 6.67 (1H, d, J¼8.4 Hz), 5.17 (1H, d, J¼13.9 Hz),
4.94 (H, t, J¼12.2 Hz), 4.56 (1H, d, J¼13.9 Hz), 3.9
(1H, m), 3.51 (1H, m), 3.19 (3H, m), 2.83 (2H, d,
J¼7.2 Hz), 1.20 (3H, d, J¼6.5 Hz); m/z (APCI) 497
MHþ; HRMS (ES) Mþ, found 497.1027. C26H22Cl2N2O4
requires 497.1035.
5.5.12. (3S)-4-[(4R)-7-Cyano-4-methyl-1-oxo-1,3,4,6-
tetrahydro-2H-naphtho[1,2-f][1,4]oxazocin-2-yl]-3-(3,4-
dichlorophenyl)-N,N-dimethylbutanamide (19). Car-
boxylic acid 16 (80 mg, 0.16 mmol) was dissolved in
dichloromethane (2 mL) under a nitrogen atmosphere; to
this was added oxalyl chloride (28 mL, 0.32 mmol) and one
small drop of dimethylforamide. After stirring for 0.5 h, the
mixture was concentrated, redissolved in dichloromethane
(1 mL) and concentrated again. The residue was dissolved
in dichloromethane (2 mL) and to this was added dimethyl-
amine (120 mL of 2 M solution in tetrahydrofuran,
0.24 mmol), and triethylamine (32 mL, 0.24 mmol). The
mixture was stirred for 15 min, concentrated and purified
using silica gel chromatography (2–3% methanol/dichloro-
methane) to give the title compound as a foamy solid
(80 mg, 95%), dH (300 MHz, CDCl3) 8.24 (1H, s), 7.81 (1H,
d, J¼7.8 Hz), 7.51 (4H, m), 7.30 (1H, m), 6.62 (1H, d,
J¼8.3 Hz), 5.15 (1H, d, J¼13.8 Hz), 4.92 (1H, t,
J¼12.6 Hz), 4.56 (1H, d, J¼13.8 Hz), 4.00 (1H, m), 3.64
(1H, m), 3.31 (2H, m), 3.11 (1H, m), 3.00 (3H, s), 2.95 (3H,
s), 2.70 (2H, d, J¼6.7 Hz), 1.22 (3H, d, J¼6.5 Hz); m/z
(APCI) 524 MHþ; HRMS (ES) Mþ, found 524.1501.
C28H27Cl2N3O3 requires 524.1508.
5.5.10. (3S)-4-[(4R)-7-Cyano-4-methyl-1-oxo-1,3,4,6-
tetrahydro-2H-naphtho[1,2-f][1,4]oxazocin-2-yl]-3-(3,4-
dichlorophenyl)butanamide (17). Carboxylic acid 16
(80 mg, 0.16 mmol) was dissolved in dichloromethane
(3 mL) under a nitrogen atmosphere; to this was added
diisopropylethylamine (55 mL, 0.32 mmol), then fluoro-
N,N,N0-tetramethylforamidinium
(51 mg, 0.19 mmol). After 1 h HOBt–NH3
hexafluorophosphate
27
(43 mg,
0.32 mmol) was added and the mixture stirred for 2 h.
Saturated aqueous sodium bicarbonate and ethyl acetate
were added (15 mL each) and the mixture was washed with
aqueous sodium bicarbonate, then brine (15 mL each), dried
over magnesium sulfate and purified via silica gel
chromatography (5–7% methanol/methylene chloride) to
give the title compound as a foamy solid (60 mg, 76%), dH
(300 MHz, CDCl3) 8.25 (1H, s), 7.82 (1H, d, J¼7.8 Hz),
7.52 (4H, m), 7.28 (1H, m), 6.68 (1H, d, J¼8.3 Hz), 5.43
(2H, m), 5.16 (1H, d, J¼13.9 Hz), 4.94 (1H, t, J¼12.6 Hz),
4.56 (1H, d, J¼13.9 Hz), 3.97 (1H, m), 3.60 (1H, m), 3.26
(1H, m), 3.15 (2H, d, J¼7.1 Hz), 2.64 (2H, m), 1.19 (3H, d,
J¼6.5 Hz); m/z (APCI) 496 MHþ; HRMS (ES) Mþ, found
496.1194. C26H23Cl2N3O3 requires 496.1194. Crystals
suitable for diffraction were prepared by dissolving the
compound in ethyl acetate/methylene chloride (10:1) and
allowing diethyl ether to diffuse into the solution in a closed
vial, mp 168–170 8C, [a]2D0¼2130 (c 0.5, methanol).
Crystallographic data (excluding structure factors) for the
structure in this paper was deposited with the Cambridge
Crystallographic Data Center as supplementary publication
numbers CCDC 230503. Copies of the data can be obtained,
free of charge, on application to CCDC, 12 Union Road,
Cambridge CB2 1EZ, UK [fax: 144-(0)1223-336033 or
e-mail: deposit@ccdc.cam.ac.uk].
References and notes
1. Stout, S. C.; Owens, M. J.; Nemeroff, C. B. Annu. Rev.
Pharmacol. Toxicol. 2001, 41, 877–906.
2. Chahl, L. A.; Urban, L. A. Pharmacol. Rev. Commun. 1999,
10, 197–203.
3. Swain, C.; Rupniak, N. M. J. Annu. Rep. Med. Chem. 1999, 34,
51–60.
4. Gerspacher, M.; Von Sprecher, A. Drugs Future 1999, 24,
883–892.
5. Gao, Z.; Peet, N. P. Curr. Med. Chem. 1999, 6, 375–388.
6. Sakurada, T.; Sakurada, C.; Tan-No, K.; Kisara, K. CNS Drugs
1997, 8, 436–447.
7. Snider, R. M.; Constantine, J. W.; Lowe, J. A., III; Longo,
K. P.; Lebel, W. S.; Woody, H. A.; Drozda, S. E.; Desai, M. C.;
Vinick, F. J.; Science, F. J. Science 1991, 251, 435–437.
8. Emonds-Alt, X.; Proietto, V.; Van Broeck, D.; Vilain, P.;
5.5.11. (3S)-4-[(4R)-7-Cyano-4-methyl-1-oxo-1,3,4,6-
tetrahydro-2H-naphtho[1,2-f][1,4]oxazocin-2-yl]-3-(3,4-

J. S. Albert et al. / Tetrahedron 60 (2004) 4337–4347
4347
Advenier, C.; Neliat, G.; Le Fur, G.; Brellere, J. C. Bioorg.
Med. Chem. Lett. 1993, 3, 925–930.
19. Sandstrom, J. Dynamic NMR spectroscopy; Academic: New
York, 1982.
9. Albert, J. S.; Aharony, D.; Andisik, D.; Barthlow, H.;
Bernstein, P. R.; Bialecki, R. A.; Dedinas, R.; Dembofsky,
B. T.; Hill, D.; Kirkland, K.; Koether, G. M.; Kosmider, B. J.;
Ohnmacht, C.; Palmer, W.; Potts, W.; Rumsey, W.; Shen, L.;
Shenvi, A.; Sherwood, S.; Warwick, P. J.; Russell, K. J. Med.
Chem. 2002, 45, 3972–3983.
20. AESOP is an in-house molecular mechanics program (Masek,
B.) derived from BIGSTRN-3 (QCPE-514) Nachbar, R., Jr.;
Mislow, K. QCPE Bull. 1986, 6, 96.
21. Bernstein, P. R.; Aharony, D.; Albert, J. S.; Andisik, D.;
Barthlow, H. G.; Bialecki, R.; Davenport, T.; Dedinas, R. F.;
Dembofsky, B. T.; Koether, G.; Kosmider, B. J.; Kirkland, K.;
Ohnmacht, C. J.; Potts, W.; Rumsey, W. L.; Shen, L.; Shenvi,
A.; Sherwood, S.; Stollman, D.; Russell, K. Bioorg. Med.
Chem. Lett. 2001, 11, 2769–2773.
10. Albert, J. S.; Ohnmacht, C.; Bernstein, P. R.; Rumsey, W. L.;
Aharony, D.; Alelyunas, Y.; Russell, D. J.; Potts, W.;
Sherwood, S. A.; Shen, L.; Dedinas, R. F.; Palmer, W. E.;
Abbott, B. M.; Russell, K. J. Org. Chem. 2004, 47, 519–529.
11. Clayden, J.; Pink, J. H. Angew. Chem., Int. Ed. 1998, 37,
1937–1939.
22. Rupniak, N. M. J.; Tattersall, F. D.; Williams, A. R.; Rycroft,
W.; Carlson, E. J.; Cascieri, M. A.; Sadowski, S.; Ber, E.;
Hale, J. J.; Mills, S. G.; MacCoss, M.; Seward, E.; Huscroft, I.;
Owen, S.; Swain, C. J.; Hill, R. G.; Hargreaves, R. J. Eur.
J. Pharmacol. 1997, 326, 201–209.
12. Ahmed, A.; Bragg, R. A.; Clayden, J.; Lai, L. W.; McCarthy,
C.; Pink, J. H.; Westlund, N.; Yasin, S. A. Tetrahedron 1998,
54, 13277–13294.
23. Drug exposure is expressed in terms of amount of drug (ng)
per unit of brain tissue (g) 4 h following oral dosing at the
indicated quantity; see Section 5.
13. Iwamura, H.; Mislow, K. Acc. Chem. Res. 1988, 21, 175–182.
14. Natsugari, H.; Ikeura, Y.; Kamo, I.; Ishimaru, T.; Ishichi, Y.;
Fujishima, A.; Tanaka, T.; Kasahara, F.; Kawada, M.; Doi, T.
J. Med. Chem. 1999, 42, 3982–3993.
24. ENIGMA is an in-house molecular graphics program.
25. Rumsey, W. L.; Aharony, D.; Bialecki, R. A.; Abbott, B. M.;
Barthlow, H. G.; Caccese, R.; Ghanekar, S.; Lengel, D.;
McCarthy, M.; Wenrich, B.; Undem, B.; Ohnmacht, C.;
Shenvi, A.; Albert, J. S.; Brown, F.; Bernstein, P. R.; Russell,
K. J. Pharmacol. Exp. Ther. 2001, 298, 307–315.
26. Manuscript in preparation.
15. Ohnmacht, C. J.; Albert, J. S.; Bernstein, P. R.; Rumsey, W. L.;
Masek, B. B.; Dembofsky, B. T.; Koether, G. M.; Andisik,
D. W.; Aharony, D. Bioorg. Med. Chem. 2004, 12, see doi:
10.1016/j.bmc.2004.03.015.
16. Anet, F. A. L.; Yavari, I. Tetrahedron 1978, 34, 2879–2886.
17. Anet, F. A. L.; Yavari, I. J. Am. Chem. Soc. 1978, 100,
7814–7819.
27. Bajusz, S.; Ronai, A. Z.; Szekely, J. I.; Graf, L.; Dunai-
Kovacs, Z.; Berzetei, I. FEBS Lett. 1977, 76, 91–92.
18. Burkert, U.; Allinger, N. L. Molecular mechanics; American
Chemical Society: Washington, DC, 1982.