5-HT1A-versus D2-receptor selectivity of flesinoxan and analogous N4-substituted N1-arylpiperazines.

Article abstract of DOI:10.1021/jm960496o

We investigated the structural requirements for high 5-HT1A affinity of the agonist flesinoxan and its selectivity versus D2 receptors. For this purpose a series of arylpiperazine congeners of flesinoxan were synthesized and evaluated for their ability to

Full text of DOI:10.1021/jm960496o

300
J . Med. Chem. 1997, 40, 300-312
5-HT1A- ver su s D2-Recep tor Selectivity of F lesin oxa n a n d An a logou s
4
1
N -Su bstitu ted N -Ar ylp ip er a zin es
Wilma Kuipers,*,† Chris G. Kruse,† Ineke van Wijngaarden,† Piet J . Standaar,† Martin Th. M. Tulp,‡
Nora Veldman,§ Anthony L. Spek,§, and Adriaan P. IJ zerman∇
Departments of Medicinal Chemistry and CNS Pharmacology, Solvay Pharmaceuticals Research Laboratories, P.O. Box 900,
1380 DA Weesp, The Netherlands, Bijvoet Center for Biomolecular Research, Laboratory of Crystal and Structural Chemistry,
Utrecht University, Padualaan 8, 3584 CH Utrecht, The Netherlands, and Division of Medicinal Chemistry,
Leiden/ Amsterdam Center for Drug Research, P.O. Box 9502, 2300 RA Leiden, The Netherlands
Received J uly 8, 1996X
We investigated the structural requirements for high 5-HT1A affinity of the agonist flesinoxan
and its selectivity versus D2 receptors. For this purpose a series of arylpiperazine congeners
of flesinoxan were synthesized and evaluated for their ability to displace [3H]-8-OH-DPAT and
[3H]spiperone from their specific binding sites in rat frontal cortex homogenates and rat
4
striatum, respectively. Variations were made in the N -substituent and the arylpiperazine
4
region. Effects of N -substitution in the investigated compounds appeared to be quite similar
for 5-HT1A- and D2-receptor affinity. Lipophilicity at a distance of four carbon atoms from the
piperazine N4 atom seems to be the main contributing factor to affinity for both receptors.
4
Our data show that the amide group in the flesinoxan N -substituent is unlikely to interact
with the 5-HT1A receptor but, instead, acts as a spacer. In contrast to the structure-affinity
4
relationships (SARs) of the N -substituents, selectivity for 5-HT1A versus D2 receptors was
gained by the arylpiperazine substitution pattern of flesinoxan. Restriction of flexibility of
4
the N -(benzoylamino)ethyl substituent and its effect on 5-HT1A-receptor affinity and activ-
4
ity were also studied. Our data show that in the bioactive conformation, the N -[(p-
fluorobenzoyl)amino]ethyl substituent is probably directed anti-periplanar relative to the HN4
atom. These results were used to dock flesinoxan (1) and two of its congeners (27 and 33) into
a model of the 5-HT1A receptor that we previously reported. Amino acid residues surrounding
4
the N -[(p-fluorobenzoyl)amino]ethyl substituent of flesinoxan and its congeners are also present
in D2 receptors. In contrast, several residues that contact the benzodioxane moiety differ from
those in D2 receptors. These observations from the 3D model agree with the 5-HT1A SAR data
and probably account for the selectivity of flesinoxan versus D2 receptors.
In tr od u ction
In this study we investigated the structure-affinity
relationships (SARs) of 1 in order to gain insight in the
structural parameters determining its high affinity for
5-HT1A receptors and selectivity versus D2 receptors. For
this purpose, we synthesized three series of congeners
of 1 (sets A-C, structural variation is depicted in Chart
2). 5-HT1A- and D2-receptor affinities of these com-
pounds were determined in radioligand binding studies.
A substantial contribution of log P to 5-HT1A-receptor
The family of receptors that are activated by the
neurotransmitter serotonin (5-hydroxytryptamine, 5-HT)
consists of at least seven types, i.e., 5-HT1-7 (for recent
reviews, see Saudou and Hen1 and Peroutka2). Some
of these types have been further subdivided. The 5-HT1
class contains 5-HT1A, 5-HT1B/1Dâ, 5-HT1DR, 5-HT1E, and
5-HT1F, as well as the Drosophila 5-HTdro2A and 5-HTdro2B
,
receptors.1 The early discovery of 8-OH-DPAT, the
R-enantiomer of which being a highly selective and
potent 5-HT1A agonist,3 enabled the search for new
selective compounds for the 5-HT1A subtype. This
research has led to the discovery of several agonists
(flesinoxan, 1), partial agonists (buspirone, 2; ipsa-
pirone, 3; BMY 7378, 4), and, recently, antagonists ((S)-
UH-301, 5; (S)-WAY-100135, 6) at 5-HT1A receptors
(see Chart 1).4-7 Flesinoxan (1) is selective for
5-HT1A receptors. In contrast, considerable affinity for
dopamine D2 receptors was reported8 for many other
4
affinity has previously been reported for certain N -
substituents but not for the arylpiperazine part of the
molecule.9,10 Therefore, we experimentally determined
log P values of our compounds and investigated a
possible correlation with 5-HT1A- and D2-receptor af-
finities. As only the series A and B contain variations
4
in the N -substituent, the log P evaluation was limited
to these compounds (see also Chart 2).
We subsequently investigated the bioactive conforma-
tion of the 5-HT1A-receptor agonist 1 that was required
for docking in our model for the 5-HT1A receptor.11,12
Compound 1 is rather flexible, since it contains several
freely rotatable bonds. Therefore, the more rigid com-
pound 33 was chosen as a model compound to study the
bioactive conformation of this type of 5-HT1A-receptor
agonist. The receptor model has previously been shown
4
N -substituted 5-HT1A arylpiperazines.
* To whom correspondence should be addressed, at the following:
tel, 31-2944-79875; fax, 31-2944-80688; e-mail, kuipers-solvay@e-
mail.com (or W.Kuipers@duphar.nl).
† Department of Medicinal Chemistry.
‡ Department of CNS Pharmacology.
§ Utrecht University.
∇ Leiden/Amsterdam Center for Drug Research.
Address correspondence pertaining to crystallographic studies to
this author.
4
to rationalize 5-HT1A SARs of N -unsubstituted arylpip-
erazines,11 as well as affinity and selectivity of several
other agonist classes such as tryptamines and aminote-
X Abstract published in Advance ACS Abstracts, November 15, 1996.
S0022-2623(96)00496-7 CCC: $14.00 © 1997 American Chemical Society

5-HT1A- vs D2-Receptor Selectivity of Flesinoxan
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 3 301
Ch a r t 1. Structures of Several 5-HT1A-Receptor
Sch em e 1a
Ligands
a
(i) P2S5, Et3N (4 equiv), CH2Cl2, 20 °C; (ii) C6H11C(OEt)NH‚HCl,
EtOH, 20 °C.
Sch em e 2a
a
(i) MeSO2Cl, Et3N, CH2Cl2, 20 °C; (ii) N-(2-methoxyphen-
yl)piperazine, Et(i-Pr)2N, CH3CN, 60 °C; (iii) H2, Pd/C, EtOH, 20
°C.
Sch em e 3a
Ch a r t 2. Variations in the Flesinoxan Structure That
Were Investigated in This Study
a
(i) TfOCH(CH3)CO2Me, Et(i-Pr)2N, 90 °C, n-pentane; (ii) KOH
(3 equiv), H2O-MeOH followed by aqueous HCl (excess); (iii) CDI
(2 × 1.0 equiv), 5-aminobenzo-1,4-dioxane, THF, reflux; (iv) H2,
Pd/C, EtOAc/MeOH; (v) BH3‚Me2S (2.6 equiv), THF, reflux; then
6 N HCl (excess); (vi) N-(4-fluorobenzoyl)aziridine, 80 °C.
tralins, and (aryloxy)propanolamine antagonists.12 It
was also shown that agonists and antagonists may
occupy very different binding sites at the receptor. A
5-HT1A agonistic profile of the compounds 27 and 33
would support the hypothesis that they address the
same binding site at the receptor as the full agonist
flesinoxan (1). Therefore, we studied the 5-HT1A func-
tional behavior of the two flexinoxan congeners 27 and
33 prior to the docking studies. The conclusions from
the modeling experiments concerning the 3D structure
of 33 were supported by the crystal structure determi-
nation of this compound.
conveniently prepared by P2S5 treatment of the corre-
sponding amide 15 in dichloromethane in the presence
of an excess of triethylamine at room temperature. The
cyclohexylamidine 17 resulted from reaction of 35 with
the corresponding iminoethyl ether hydrochloride salt
in ethanol at room temperature.
The route to compounds 21 and 22 is depicted in
Scheme 2. trans-4-(4-Fluorophenyl)-3-butenol (36) was
treated with methanesulfonyl chloride and triethyl-
amine in dichloromethane at 20 °C overnight. The
resulting mesylate was subsequently reacted with N-(2-
methoxyphenyl)piperazine and diisopropylethylamine in
acetonitrile at 60 °C overnight. Finally, product 22
could be conveniently hydrogenated with Pd/C as the
catalyst in ethanol at 20 °C, giving 21.
The cis-dimethyl-substituted arylpiperazine 33 was
synthesized by a stepwise procedure as shown in
Scheme 3. N-Benzyl-D-alanine methyl ester (37) was
alkylated with the triflate of racemic methyl lactate13
to obtain the imino diester 38 as a diastereoisomeric
mixture of the meso and the d,l forms. After converting
this mixture into the corresponding imino diacid 39 by
saponification, the arylpiperazine moiety was con-
Ch em istr y
The preparation of a number of target compounds
(Tables 1-4) has already been described in the litera-
ture. The unknown N-2-(acylamino)ethyl-substituted
(2-methoxyphenyl)piperazines 7, 9, 10, and 12-14 were
prepared by acylation of the known N-2-aminoethyl
precursor 35,29 using either the anhydride (for 7), the
acyl chloride (for 9, 10, and 13), or the Mukayama ester
(for 12 and 14) of the corresponding acids.
The route of preparation for the amide isosters 16 and
17 is depicted in Scheme 1. The thioamide 16 was

302 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 3
Kuipers et al.
F igu r e 2. Definition of the torsion angles τ1 (C6′-C1′-N1-
C6), which determines the angle between the benzene and
piperazine rings, τ2 (HN4-N4-C1′′-C2′′), which primarily
4
determines the direction of the N -substituent, τ3 (N4-C1′′-
C2′′-N3′′), and τ4 (C1′′-C2′′-N3′′-C4′′).
F igu r e 1. Sequences of the putative transmembrane domains
of the rat 5-HT1A receptor, which were used for construction
of the receptor model, and the corresponding alignment with
the rat D2-receptor sequence.12,48 Residues marked boldface
were found to be part of the binding site (residues within a
sphere of 4 Å) of flesinoxan (1), 27, and 33 in the 5-HT1A
-
receptor model. Residues in the D2 receptor sequence that
differ from those in the modeled binding site are printed in
italics. Residues in the D2 sequence that correspond with those
in the modeled 5-HT1A-receptor binding site are underlined.
F igu r e 3. Model compounds 43 and 44 that were used for
the study of the torsion angle τ2 of compounds 27 and 33,
respectively, with MOPAC/AM1 calculations. The three stag-
gered conformations of the C2′′ atom with respect to the HN4
atom with theoretical τ2 values of -60°, 60°, and 180° are found
as energy minima for 43. The first two gauche conformations
are not accessible to compound 44, caused by steric hindrance
with the methyl substituents at the piperazine ring. Com-
pounds 43 and 44 share the anti-periplanar conformation at
τ2 ) 180° as an energy minimum. This is also the conformation
of τ2 in the crystal structure of compound 33, for which 44
serves as a model compound.
structed by N,N-carbonyldiimidazole-promoted coupling
with 5-aminobenzo-1,4-dioxane.14 The imide 40 was
converted into the dimethyl-substituted arylpiperazines
41 and 42 by two successive reduction steps (debenzy-
lation and deoxygenation).14 After these steps the
diastereomers 41 and 42 were separated, and the
achiral cis-diastereomer 41 was subsequently converted
into 33 by treatment with N-(4-fluorobenzoyl)aziridine.
Bioch em istr y
Affinity values of compounds 7-32 were measured
as their ability to displace [3H]-8-OH-DPAT from central
5-HT1A recognition sites in rat frontal cortex and [3H]-
spiperone from D2 recognition sites in rat striatum. In
addition, 5-HT1A agonist activity of compounds 27 and
33 was measured as their potency to inhibit forskolin-
stimulated accumulation of intracellular cAMP in cells
expressing human 5-HT1A receptors.
esized in earlier docking studies.11,12 SAR of these
agonists could be rationalized with the putative binding
sites in the aforementioned 5-HT1A-receptor model. We
assumed that the binding site of the arylpiperazine
4
moiety of 1 overlaps with that of 30 and 31: the N -
unsubstituted analogs of its congeners 26 and 27 (Table
3).11 Also the conformation of the arylpiperazine moiety
4
was chosen as previously reported for the N -unsub-
Mod elin g
stituted compounds 30 and 31. Similar arrangements
were used for the docking of 27 and 33. The validity of
this choice was investigated by comparison of 5-HT1A
F lesin oxa n in th e 5-HT1A-Recep tor Mod el. Flesi-
noxan (1) was docked into a 5-HT1A-receptor model12 in
order to rationalize its high 5-HT1A affinity and its
selectivity versus D2 receptors. Two of its congeners
(i.e., 27 and 33) were also incorporated in the docking
study. The 5-HT1A-receptor model, which was built on
the structure of bacteriorhodopsin as a template, was
4
SAR of N -substituted arylpiperazines with that of the
4
corresponding N -unsubstituted compounds.
Bioa ctive Con for m a tion of P h en ylp ip er a zin es:
4
4
N -Su bstitu en t. The conformation of the N -sub-
stituent is primarily determined by the torsion angles
τ2, τ3, and τ4 (see Figure 2). The benzamide part is
considered to be fairly rigid (the plane angles between
the benzene ring and the amide moiety of the AM1-
calculated conformations of compound 33 are all in the
range 25-30°, vide infra). The 2,6-dimethyl substitu-
ents at the piperazine ring in compound 33 restrict the
4
previously shown to explain affinities of N -unsubsti-
tuted arylpiperazines.11,12 It consists of seven putative
R-helical transmembrane domains. Sequences of the
putative transmembrane domains are shown in Figure
1.12 For docking procedures possible bioactive confor-
mations were generated, and a putative binding site of
1 in the model was identified. Conformational aspects
that were investigated in this study are depicted in
Figure 2. For clarity, we use the numbering scheme
from Figure 2 throughout the entire paper.
4
conformational freedom of the N -[(p-fluorobenzoyl)-
amino]ethyl substituent. The effect of the dimethyl
substitution in this compound on 5-HT1A-receptor af-
finity and activation was determined.
Bin d in g Site of P h en ylp ip er a zin e 5-HT1A-Recep -
tor Agon ists in th e Recep tor Mod el. Flesinoxan (1)
is a full 5-HT1A-receptor agonist. It is likely, therefore,
that its binding site at the 5-HT1A receptor overlaps that
of other 5-HT1A agonists. Binding sites of these other
agonists, such as 5-HT and 8-OH-DPAT as well as
several arylpiperazine compounds, have been hypoth-
Also the conformational effects were investigated. For
this purpose, firstly the rotational barrier of the N -
4
ethyl chain in the hypothetical compounds 43 and 44
(see Figure 3) was studied with semiempirical MOPAC/
AM1 calculations. In these compounds, the benzamide
moiety of 27 and 33 is replaced by a hydrogen atom,
allowing the study of the torsion angle τ2 independently

5-HT1A- vs D2-Receptor Selectivity of Flesinoxan
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 3 303
Ta ble 1. 5-HT1A- and D2-Receptor Affinities of Compounds
7-17 (A)
F igu r e 4. PLUTON49 representation of the conformation of
compound 33 in the crystal structure. The torsion angles τ1
(C6′-C1′-N1-C6), τ2 (HN4-N4-C1′′-C2′′), τ3 (N4-C1′′-C2′′-
N3′′), and τ4 (C1′′-C2′′-N3′′-C4′′) are -11.3° (3), 176.9° (3),
169.7° (2), and 80.0° (3), respectively (see also Figure 2). For
clarity the fumarate counterion, which is hydrogen bonded to
the basic nitrogen N4 atom and the amide NH, is not shown.
from τ3 and τ4. Thus 43 and 44 may serve as model
compounds for the investigation of τ2 of compounds 27
and 33, respectively. Results of this conformational
analysis were supported by the crystal structure of
compound 33 (Figure 4). Secondly, possible conforma-
4
tions of the complete N -[(p-fluorobenzoyl)amino]ethyl
chain were generated by varying τ3 and τ4 in the
RANDOMSEARCH option of the SYBYL program,
using the crystal structure of compound 33 as the
starting geometry (for details, see the Experimental
Section). This option locates energy minima by ran-
domly adjusting the chosen torsion angles (here τ3 and
τ4) and minimizing the energy of the resulting geometry
(see also the Experimental Section). The conformation
of the arylpiperazine moiety was constrained during this
procedure because the Molecular Mechanics (MM) meth-
ods, as used in the RANDOMSEARCH option, are less
reliable for calculation of this moiety.11,15 The unique
conformations of 33 that were found were subsequently
fully energy-minimized with the semiempirical method
MOPAC/AM1,16 which has been shown to be suitable
for arylpiperazine conformational calculations.15 The
resulting conformations (energy e 3 kcal/mol above the
absolute minimum energy) were checked for their ability
to fit the 5-HT1A-receptor model and used for further
docking studies.
a
Ki values for the displacement of [3H]-8-OH-DPAT from central
5-HT1A recognition sites in rat frontal cortex homogenates and of
[3H]spiperone from D2 binding sites in rat striatum. Ki values are
based on n ) 3 determinations, each using four to six concentra-
tions in triplicate.
Resu lts a n d Discu ssion
SARs a t 5-HT1A a n d D2 Recep tor s: Va r ia tion s
in A (Ta ble 1). Replacement of the phenyl ring of
compound 8 by a methyl group, as in compound 7, is
detrimental for the affinity for 5-HT1A and dopamine
D2 receptors. Apparently, the benzene ring in 8 pro-
vides an additional interacting group for both receptors.
Replacement of the benzene ring of 8 by 2-thiophene as
in 9 does not change either the 5-HT1A- or D2-receptor
affinity. In contrast, its replacement by more polar
aromatic rings like 2-furan (10) and 2-pyrrole (11) or
polar six-membered rings like 4-pyridyl (12) and 2,4-
pyrimidyl (13) reduces 5-HT1A-receptor affinity about
10-fold, while the m-pyridine in 14 reduces the receptor
affinity more than 2 log units. Similar effects of
replacement of the benzene ring of 8 with other aromatic
rings were observed for D2-receptor affinity, although
the polar five-membered rings in 10 and 11 lower
affinity less dramatically (approximately 5-fold) than
the polar six-membered rings in 12-14 (15-44-fold
reduction). Thus, apolar aromatic rings are preferred
over more polar aromatic rings in this position of the
Saturation of the benzene ring in 8 to cyclohexane in
15 has little effect on the affinity for both 5-HT1A and
D2 receptors. The high affinity of 15 for 5-HT1A and D2
receptors indicates that the lipophilic rather than the
aromatic character of the substituent contributes to
receptor affinity. Lipophilicity may also play a role in
the approximately 3-fold increase in both 5-HT1A- and
D2-receptor affinity by replacement of the benzamide
CdO by CdS (compare 16 with 15). In agreement with
this idea, both affinities are reduced approximately
6-fold by replacement of CdO in 15 by CdNH in 17. Of
this series of compounds (A), the cyclohexylthioamide
analog 16 is the most potent ligand at both 5-HT1A and
D2 receptors.
Va r ia tion s in B (Ta ble 2). In Table 2 variations
(B) of the amide moiety of the [(p-fluorobenzoyl)amino]-
ethyl group are shown. The p-fluoro substituent in 18
has no influence on 5-HT1A- and slightly increases D2-
receptor affinity (compare with 8). Replacement of the
4
N -substituent for both 5-HT1A and D2 receptors.
amide NH by CH2 (in 19) has little influence on 5-HT1A-

304 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 3
Kuipers et al.
Ta ble 3. 5-HT1A- and D2-Receptor Affinities of Compounds
Ta ble 2. 5-HT1A- and D2-Receptor Affinities of Compounds
18-24 (B)
25-32 (C)
a
Ki values for the displacement of [3H]-8-OH-DPAT from central
5-HT1A recognition sites in rat frontal cortex homogenates and of
[3H]spiperone from D2 binding sites in rat striatum. Ki values are
based on n ) 3 determinations, each using four to six concentra-
tions in triplicate.
a
Ki values for the displacement of [3H]-8-OH-DPAT from central
5-HT1A recognition sites in rat frontal cortex homogenates and of
[3H]spiperone from D2 binding sites in rat striatum. Ki values are
based on n ) 3 determinations, each using four to six concentra-
or D2-receptor affinity. Reduction of the keto group in
19 to hydroxy in 20, and even its complete removal as
in the (p-fluorophenyl)butane analogs 21 and 22, has
only minor effects on affinity for both receptors. The
double bond in 22 may be conjugated with the aromatic
π-system of the benzene ring in a similar manner as
the amide group in 18. Thus, neither the electrostatic
effect of the CdO group nor the conformational effect
of the amide function in 18 seem to influence its affinity
for 5-HT1A and D2 receptors, respectively. In contrast
introduction of an ether oxygen as in 23 reduces affinity
for both receptors. Apparently, an electronegative atom
at this position is unfavorable. Shifting this oxygen
atom by one position toward the benzene ring, like in
compound 24, restores both 5-HT1A- and D2-receptor
affinities to similar values as observed for compounds
18-22. None of the variations of the amide moiety in
this series (B) results in selective 5-HT1A ligands.
Va r ia tion s in C (Ta ble 3). The results in Table 3
show that compounds with variation in the aryl part
(C; see Chart 2) may display larger differences between
5-HT1A- and D2-receptor affinities than compounds in
the former two series (A and B, respectively). Omitting
the 2-methoxy group of compound 18, giving the un-
substituted phenylpiperazine 25, for instance, enhances
the selectivity for 5-HT1A receptors by a factor of 4.
Apparently, the presence of a 2-methoxy group is more
favorable for D2 than 5-HT1A receptors. On the other
hand, incorporation of the 2-methoxy group into a furan
ring as in 26 is highly favorable for 5-HT1A-receptor
affinity, whereas the affinity for D2 receptors is not
b
tions in triplicate. 5-HT1A-receptor affinity data taken from van
Steen et al.18
affected. The result is a significant increase in selectiv-
ity (ratio D2/5-HT1A is 35.3). Replacement of 2-meth-
oxyphenyl in 18 by benzodioxane yields compound 27
which is slightly less potent than the benzofuran analog
26 but even more selective for 5-HT1A when compared
to D2 receptors (ratio D2/5-HT1A is 46.7). This selectivity
is further increased to a factor of 82.4 by substitution
of the dioxane ring at the 2-position with a CH2OH
group as in 1, which is less favorable for D2- than
5-HT1A-receptor affinity (compare with 27). Thus,
selectivity of 1 for 5-HT1A versus D2 receptors seems to
be largely due to the arylpiperazine aromatic ring
system rather than the N -[(p-fluorobenzoyl)amino]-
ethyl substituent.
5-HT1A SARs for Mod elin g: An a lysis of Bin d in g
Site a n d Con for m a tion of N -Su bstitu ted Com -
p a r ed t o N -Un su b st it u t ed P h en ylp ip er a zin es
(Ta ble 3). Table 3 also shows 5-HT1A SARs of phen-
ylpiperazines carrying N -[(p-fluorobenzoyl)amino]ethyl
substituents and those of the corresponding N -unsub-
4
4
4
4
4
stituted compounds. Addition of a 2-methoxy group to
25, as in 18, increases 5-HT1A-receptor affinity by a
factor of 5, while in the corresponding N -unsubstituted
compounds 28 and 29, this increase amounts to a factor
of 3. Replacement of the benzene ring of 25 and 28 with
benzofuran, yielding 26 and 30, raises 5-HT1A-receptor
affinity by factors of 33 and 38, respectively. A similar
4

5-HT1A- vs D2-Receptor Selectivity of Flesinoxan
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 3 305
Ta ble 4. In Vitro 5-HT1A-Receptor Data of Flesinoxan
Congeners 27 and 33
Table 4 show that compound 27, like 1, is a potent
agonist at 5-HT1A receptors, also with full intrinsic
activity (data not shown). cis-Dimethyl substitution of
the piperazine ring yields the similarly potent 5-HT1A
-
receptor agonist 33 (intrinsic activity R ) 0.86 ( 0.14).
Apparently, the methyl groups do not affect the binding
mode of compound 33 at 5-HT1A receptors, although
these groups dramatically reduce the conformational
4
freedom of its N -substituent. Therefore 33 was used
X2, X6
H, H
no.
Ki ( SEM (nM)a
EC50 ( SEM (nm)b
as a model compound for the study of the bioactive
conformation of 1 and 27.
27
33
0.30 ( 0.04
1.0 ( 0.4
0.90 ( 0.05
0.78 ( 0.16
cis-dimethyl
4
In flu en ce of Lip op h ilicity of N -Su bstitu en ts on
a
Ki values for the displacement of [3H]-8-OH-DPAT from central
5-HT1A recognition sites in rat frontal cortex homogenates. Ki
values are based on n ) 3 determinations, each using four to six
5-HT1A- a n d D2-Recep tor Affin ity. Partition coef-
ficients (n-octanol/water) of compounds 7-11 and 13-
24 were determined by a high-performance liquid
chromatographic (HPLC) method. These compounds of
b
concentrations in triplicate. Potency at human 5-HT1A receptors
in inhibiting forskolin-stimulated accumulation of intracellular
cAMP. EC50 values based on three independent experiments (see
also the Experimental Section).
4
series A and B contain varying N -substituents. The
results (reported as log P) are collected in Table 5. The
lipophilicity of the substituent attached to the amide
moiety appears to greatly influence the affinity of
compounds 7-11 and 13-15 for 5-HT1A as well as D2
receptors. The log P values of these compounds cor-
relate well with 5-HT1A- and D2-receptor affinity (cor-
relation coefficients for pKi versus log P are 0.96 and
0.93, respectively). Also replacement of the polar CdO
group in 15 with the more lipophilic CdS in 16 increases
both receptor affinities. Compound 17 may be some-
what more hydrophilic than 15 or 16 as a result of
(partial) protonation of the imino moiety (see also the
1H NMR data in the Experimental Section). The given
log P value represents the uncharged form (see the
protocol in the Experimental Section). The more hy-
drophilic nature of 17 might account for the lower
5-HT1A- and D2-receptor affinities of this compound
when compared to 15 and 16.
Ta ble 5. Measured log P Values for Compounds from Tables 1
and 2
pKib
a
compd
measured log P
5-HT1A
D2
7
8
9
0.5
2.2
2.0
1.6
1.6
ND
1.5
1.2
2.8
3.7
3.4
2.5
3.6
3.4
4.8
4.7
3.6
4.1
6.1
8.9
8.8
7.7
7.8
7.7
7.8
6.8
9.1
9.5
8.3
9.0
9.0
8.6
9.0
8.7
7.7
8.5
6.0
7.9
7.8
7.1
7.2
6.6
6.6
6.3
7.9
8.2
6.9
8.3
8.6
8.5
8.5
8.4
8.0
8.5
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
In contrast to the A series, lipophilic dependence of
5-HT1A- or D2-receptor affinity is completely absent for
compounds of series B. Correlation coefficients of log
P of compounds 18-24 with 5-HT1A- and D2-receptor
affinity values from Table 5 are 0.03 and 0.28, respec-
tively. However, the log P values of these compounds
are in general higher than those of compounds from
series A. Therefore, it may also be argued that the
lipophilicity of the entire chain should be sufficiently
high for good binding at 5-HT1A and D2 receptors. The
observed effects of log P for compounds of the A and B
series are similar for both 5-HT1A- and D2-receptor
types.
a
n-Octanol/water coefficients were determined by an HPLC
method; ND ) not determined. log P values below 2.1 were
obtained by extrapolation; interpolated values (in the range
b
2.1-5.7) have a 95% confidence interval of (0.4. pKi (i.e., -log
Ki) values were calculated from the Ki values (in M) in Tables 2
and 3.
replacement with benzodioxane in 27 and 31 gives 16-
and 13-fold increases of 5-HT1A-receptor affinity when
compared to 25 and 28. Addition of a CH2OH substitu-
4
ent to 27 and the corresponding N -unsubstituted
compound 31 (eltoprazine) gives compounds 1 and 32,
Our results for the 5-HT1A receptor are in agreement
respectively. This CH2OH group slightly lowers 5-HT1A
-
with previous studies by Mokrosz et al.9 and van Steen
4
receptor affinity, although the effect for the N -unsub-
stituted compound 32 is less pronounced than for the
4
et al.10 of arylpiperazines carrying N -alkyl substitu-
4
ents. In these studies it was shown that 5-HT1A
-
N -substituted compound 1. Thus, effects of substitu-
receptor affinity increases with growing alkyl chain
length, mainly in the region C4-C6. These alkyl
substituents may partly overlap the -C(dO)-R moiety
(R ) aryl or cyclohexyl) of our compounds. We found a
considerable contribution of lipophilicity to affinity for
tion in the aromatic part of the arylpiperazine moiety
4
for the investigated N -substituted arylpiperazines and
4
the corresponding N -unsubstituted compounds appear
to be rather similar. Therefore it seems likely that the
arylpiperazine parts of these compounds address similar
binding sites at 5-HT1A receptors. This hypothesis is
4
5-HT1A receptors for this part of the N -substituent,
4
4
which resembles the effect of N -alkyl substituents. In
corroborated by the fact that both N -unsubstituted
4
addition, our compounds displayed a similar correlation
between log P and their affinity for D2 receptors.
compounds, 30 and eltoprazine (31), as well as N -
substituted arylpiperazines from this series, 27 and 1,
display agonist properties at 5-HT1A receptors (for
compound 27, see Table 4).4,11
In Vitr o 5-HT1A F u n ction a l Beh a vior of F lesi-
n oxa n Con gen er s 27 a n d 33 (Ta ble 4). Results from
van Steen et al. previously found that the presence
of an amide moiety is not a prerequisite for high 5-HT1A
4
affinity.17 Using CoMFA models, the affinity of N -
substituents containing an amide moiety could be fully

306 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 3
Kuipers et al.
explained by steric field descriptors. Thus, the hydrogen-
bonding capacity of the amide moiety seems to be of
minor importance for 5-HT1A-receptor affinity. Our
experimental data confirm these results and show that
the amide moiety merely acts as a spacer instead. In
addition we show that the same rule applies to D2-
receptor affinity as well.
Cr ysta l Str u ctu r e Ver ifica tion of Com p ou n d 33.
The X-ray structure of compound 33 is shown in Figure
4. The observed torsion angle τ1 is -11.3° (3). This is
in agreement with previous results concerning the
4
1
bioactive conformation of N -unsubstituted N -arylpip-
erazine agonists at 5-HT1A receptors.11 We found that
singly ortho-substituted agonists of this class probably
bind at 5-HT1A receptors in a semicoplanar conformation
with torsion angle τ1 ≈ 0°. In this conformation, the
plane angle between the piperazine and the benzene
ring is approximately 30°. Thus, a similar conformation
for the arylpiperazine moiety was found in the crystal
4
structure of the (N -substituted) potent 5-HT1A receptor
agonist 33.
Mod elin g: P h en ylp ip er a zin es in th e 5-HT1A
-
Receptor Model. Bioactive Con for m ation of P h en -
4
ylp ip er a zin es. N -Su b st it u en t Con for m a t ion s
Ca lcu la tion s. Figure 5 shows the results of MOPAC/
AM1 calculations of τ2, for the hypothetical compounds
43 (a) and 44 (b), respectively (structures in Figure 3).
For 43, three conformations are found as minima, with
the C2′′ atom staggered relative to the HN4 atom. The
anti-periplanar conformation at τ2 ) 180° is also found
as a minimum for the cis-dimethyl analog 44. The two
other minima of 43 (gauche conformers) are not acces-
sible to 44 because of severe steric hindrance between
the C2′′ methyl group and the two equatorially posi-
tioned methyl substituents at the piperazine ring.
These results are in agreement with the crystal struc-
ture of compound 33, which is in the anti-periplanar
conformation (τ2 ) 176.9° (3); see Figure 4). Although
it might be argued that τ2 ) 0° is also accessible to both
43 and 44 (not a minimum for 43, but the energy is
within 3 kcal/mol of the absolute minimum energy), it
seems less likely to be the bioactive conformation. In
F igu r e 5. Results of MOPAC/AM1 calculations of the torsion
angle τ2 of compounds 43 (a) and 44 (b), which serve as model
compounds for 27 and 33, respectively. Energy is relative to
the calculated absolute minimum. For compound 43 three
staggered minimum energy conformations are calculated with
τ2 values of -45°, 45°, and 180°, respectively. In conformations
-45° and 45°, the C2′′ atom is positioned gauche relative to
the HN4 atom; in the 180° minimum this orientation is anti-
periplanar. Only the latter conformation is energetically
favored by compound 44 also, since in the gauche conforma-
tions the C2′′ atom is severely sterically hindered by the two
methyl substituents at the piperazine ring. The second mini-
mum of 44 is the eclipsed conformation at τ2 ) 0°, in which
the C2′′ atom points in the same direction as the HN4 atom.
This is not a minimum energy conformation for compound 43,
although its energy is within 3 kcal/mol of the absolute
minimum.
4
this conformation, the N -substituent is eclipsed rela-
differences occur mainly in the angle between the amide
tive to the HN4 atom and directed toward the putative
4
moiety and the benzene ring of the N -substituent
NH+-interacting group at the receptor. Thus, in the τ2
(i.e., N3′′-C4′′-C5′′-C6′′). For C and D this torsion
angle amounts to 153.2° and -150.9°, respectively. For
E-G this angle is -152.7°, 154.3°, and 150.5°, respec-
tively. Thus, conformations F and G are practically
identical.
4
) 0° conformation, the N -substituent is likely to bump
into essential receptor volume. Therefore, the crystal
structure of 33, with τ2 ) 176.9° (3), was used to further
study the bioactive conformation.
The RANDOMSEARCH analysis of τ3 and τ4 yielded
20 conformations with energies (Tripos force field)
ranging from 116.7 to 119.2 kcal/mol. Subsequent full
energy minimization with MOPAC/AM1 further reduced
the number of favorable conformations to eight (A-H;
see Table 6). The conformations A and B have τ3 values
of approximately 180°; in conformation H τ3 amounts
to -63.8°. The torsion angle τ3 for the conformations
C-G is in the range 64.0-76.1°. Conformations C-G
can be divided in two clusters, separated by their τ4
values. The torsion angle τ4 determines the orientation
of the amide moiety (see Figure 6). Cluster 1 is formed
by C and D (τ4 is -104.2° and -105.9°, respectively).
In conformations E-G (cluster 2), the direction of the
amide moiety is shifted approximately 180°, with τ4
values in the range 65.6-75.5°. Within the clusters
Structure A closely resembles the crystal structure
of 33. This indicates that our approach indeed yields
accessible low-energy conformations. The eight poten-
tially bioactive conformations (A-H) were further in-
vestigated for their ability to fit the 5-HT1A-receptor
model12 in a docking study.
Com p ou n d s 1, 27, a n d 33 in th e 5-HT1A-Recep tor
Mod el. Conformations A, B, and H of compound 33 and
the close analogs 1 and 27 (Table 6) could not be
accommodated in the 5-HT1A-receptor model, as the
4
benzamide part of their N -substituents caused severe
steric hindrance with the backbones of either helix VI
(conformation H) or helix VII (conformations A and B;
see Figure 7, top). The complexes of the resulting
conformations C-G with the 5-HT1A-receptor model
were further minimized. These conformations (C-G)

5-HT1A- vs D2-Receptor Selectivity of Flesinoxan
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 3 307
formations C and D) or hydrophilic (conformations E-G)
residues. This observation from the model may account
for the fact that 5-HT1A-receptor affinity is hardly
affected by the replacement of the polar amide moiety
with more lipophilic groups in the compounds of series
B.
4
In the model, the residues that surround the N -
substituent in the 5-HT1A-receptor model (i.e., Asp116,
Cys120, Ser123, Ile124, and Leu127, helix III; Phe354,
Cys357, Trp358, and Phe361, helix VI; Ser393, helix
VII) are also present in D2 receptors (see Figure 8,
bottom, and the alignment in Figure 1). This might
4
explain the similar SARs of the N -substituent at
5-HT1A and D2 receptors. In contrast, a number of
residues in the model that surround the heteroaryl
moiety (i.e., Thr160, Trp161, and Gly164, helix IV;
Ser199, Thr200, Ala203, and Phe204, helix V; Phe362,
Ala365, and Leu366, helix VI) differ from those in D2
receptors. Observed differences between 5-HT1A and D2
receptors in this region are Thr160 T Val, Gly164 T
Ser, Thr200 T Ser, Ala203 T Ser, Ala365 T His, and
Leu366 T Ile, respectively (see Figure 8, bottom, and
the alignment in Figure 1). This is in good agreement
with the SAR data of our compounds, which show that
the arylpiperazine part itself is not very well accom-
modated by the D2 receptor. In contrast, the 5-HT1A
receptor can accommodate rather bulky arylpiperazine
substituents like the dioxane ring in 27 and 33 and even
the additional CH2OH group of 1. In the model, the
dioxane ring of these compounds has a close contact with
the small and lipophilic Ala365. This residue is replaced
with a bulky and polar histidine in D2 receptors (see
Figure 1), which may account for the low tolerance of
bulky arylpiperazine substituents at D2 receptors. Thus,
selectivity for 5-HT1A versus D2 receptors is mainly
caused by the aryl substitution pattern of the arylpip-
erazine moiety.
F igu r e 6. Potentially bioactive conformations C and D
(cluster 1) and E-G (cluster 2) of compound 33, resulting from
RANDOMSEARCH analysis and subsequent minimization
with MOPAC/AM1. The compounds differ in the values for τ4,
which determines the direction of the amide group. This
torsion angle is approximately -105° for C and D and varies
between 65.6° and 75.5° for E-G. As a result, the CdO group
(red) of C and D is pointing to the ‘right’, while that of
conformations E-G is pointing to the ‘left’ side. Still the
p-fluorophenyl substituents of both clusters overlap in space.
Within the clusters differences occur mainly in the angle
4
between the amide moiety and the benzene ring of the N -
substituent (i.e., N3′′-C4′′-C5′′-C6′′). For C and D this
torsion angle amounts to 153.2° and -150.9°, respectively. For
E-G this angle is -152.7°, 154.3°, and 150.5°, respectively.
all have τ3 values of approximately 70° and consequently
occupied a similar region in the 5-HT1A-receptor model
(see also Figure 6). For 1, as well as for 27 and 33, the
conformations C and D converged to stable receptor-
ligand complexes. For example, Figure 7, bottom, shows
1 in the receptor model in conformation C. The confor-
mations E-G fitted the model slightly worse, due to
steric hindrance with the side chain of Trp358 in helix
VI. This steric hindrance, or even that of conformations
A, B, and H with the helical backbones, does not
completely rule out these conformations as potentially
bioactive. It is of interest, however, that the surround-
ings of compounds 1, 27, and 33 in conformations C and
D after minimization are in good agreement with SAR
data. Thus, despite the uncertainty in the exact atomic
coordinate positions, which is considered to be the major
imperfection of all bacteriorhodopsin-based models of G
protein-coupled receptors (GPCRs), our ligand-receptor
model accounts for 5-HT1A affinity and selectivity versus
D2 receptors (vide infra).
In our model, the CH2OH group has no additional
interaction with the receptor. This is in agreement with
SAR data that show no contribution of this group to
5-HT1A-receptor affinity.
Con clu sion s
5-HT1A-receptor affinity and selectivity versus D2
receptors of the potent 5-HT1A receptor agonist flesi-
4
noxan (1) and a series of analogous N -substituted
arylpiperazines were studied. The selectivity of 1 was
1
shown to emerge mainly from the N -aryl substitution
pattern. For both 5-HT1A and D2 receptors, SARs at the
For example, a lipophilic pocket is formed by Cys120,
Ile124, Leu127, Phe354, Cys357, Trp358, and Phe361
and might account for the considerable contribution of
log P to both 5-HT1A- and D2-receptor affinity. This
pocket is located in the direction of the CdO group of
4
N -substituent appeared to be dominated by lipophi-
licity. Therefore, variation in this part of the structure
did not yield selective compounds.
5-HT1A-receptor affinities could be rationalized with
a computer modeling study. The bioactive conformation
was investigated. The results were used to study the
interactions of the ligands in a 5-HT1A-receptor model
that we have previously reported. The coplanar con-
4
the N -substituent and its benzene ring (see Figure 8,
top). The conformations E-G essentially occupy a
similar binding region at the 5-HT1A-receptor model as
conformations C and D (see also Figure 6). Therefore,
the residues surrounding the p-fluorophenyl moiety of
compounds 1, 27, and 33 in conformations E-G are
similar to those in conformations C and D. However,
the CdO group in conformations E-G is directed away
from the lipophilic pocket and toward Ser393 at helix
VII. Thus, two possible orientations of the amide moiety
in the model are found, in which the CdO group of the
amide moiety points in the direction of lipophilic (con-
formation of the arylpiperazine moiety, which was
4
shown to be preferred by N -unsubstituted 5-HT1A
-
4
receptor agonists, can also be adopted by the N -sub-
stituted agonists 1, 27, and 33. In this conformation,
the plane angle between the benzene and piperazine
ring is approximately 30°. In the bioactive conformation
4
of the agonists 1, 27, and 33, the N -substituent is
probably directed anti-periplanar with respect to the

308 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 3
Kuipers et al.
Ta ble 6. Crystal Structurea and Conformations (A-H) Resulting from RANDOMSEARCH Analysis of τ3 and τ4 of Compound 33 and
Subsequent Minimization with MOPAC/AM1
AM1 energy
(kcal/mol)
τ1 (deg),
C6′-C1′-N1-C6
τ2 (deg),
HN4-N4-C1′′-C2′′
τ3 (deg),
N4-C1′′-C2′′-N3′′
τ4 (deg),
C1′′-C2′′-N3′′-C4′′
conformation
crystal structure
ND
-11.3
-12.8
-16.1
-16.7
-14.6
-14.1
-13.1
-18.3
-16.4
176.9
-178.9
-177.5
167.9
168.2
166.1
170.2
163.8
-163.5
169.7
-178.4
174.0
76.1
80.0
66.2
-64.3
-104.2
-105.9
65.6
75.5
66.9
-63.9
A
B
C
D
E
F
G
H
55.3
55.2
54.2
54.2
55.2
56.3
55.8
55.2
75.2
67.2
68.1
64.0
-63.8
a
Standard deviations amount 3°, 3°, 2°, and 3° for τ1, τ2, τ3, and τ4, respectively.
F igu r e 7. (Top) Conformations A (red), B (yellow), and H
(blue) of compound 27 in the 5-HT1A-receptor model. The
overlapping parts of the conformations are white. The N -
substituents of A and B cause severe steric hindrance with
the backbone of helix VII, while the p-fluorophenyl substituent
of conformation H sticks right through the backbone of helix
F igu r e 8. More detailed view of the environment of flesinoxan
(1) docked into the 5-HT1A-receptor model in conformation C
and subsequently minimized. (Top) The hydrophobic residues
Cys120, Ile124, Leu127, and Cys357 (green) and the aromatic
residues Phe354, Trp358, and Phe361 (blue) form a lipophilic
4
4
pocket at one side of the N -substituent. At the other side of
VI. (Bottom) Position of flesinoxan (1) docked into the 5-HT1A
-
the pocket two hydrophilic (pink) serine residues are observed
(at positions 123 and 393). (Bottom) Residues in the binding
pocket that are similar to (yellow) or different from (pink) those
receptor model in conformation C and subsequently minimized.
The model is viewed from the extracellular side. In contrast
4
4
to the conformations A, B, and H, the N -substituent in
in D2 receptors. All residues that surround the N -substituent
conformations C-G is located in the available space between
the helices III, VI, and VII.
in the 5-H1A-receptor model are identical with those at similar
positions in the D2 rat receptor, according to the alignment of
Figure 1 (i.e., yellow residues Asp116, Cys120, Ser123, Ile124,
Leu127, Phe354, Cys357, Trp358, Phe361, and Ser393). Sev-
1
eral of the residues that surround the N -aryl part are
HN4 atom. Eight of such anti-periplanar conformations
(A-H), among which is the crystal structure of 33, could
be identified as potentially bioactive at 5-HT1A receptors.
Five of these conformations fitted the 5-HT1A receptor
model well. The predominantly hydrophobic residues
different from those in D2 receptors (i.e., pink residues Thr160,
Gly164, Thr200, Ala203, Ala365, and Leu366).
with 5-HT1A SARs and account for the selectivity of
flesinoxan (1) versus D2 receptors.
4
of helices III, VI, and VII, which surround the N -
substituent, are identical for 5-HT1A and D2 receptors.
Six of the residues that surround the benzodioxane ring
of flesinoxan (1) in the 5-HT1A-receptor model are
different from those in D2 receptors. Thus, the positions
of the 5-HT1A agonists 1, 27, and 33 in the receptor
model, in the potentially bioactive conformations, agree
Exp er im en ta l Section
Ch em istr y. Melting points are uncorrected. 1H NMR
spectra were recorded on
a Bruker WP-200 or AM-400
spectrometer. Chemical shifts (δ) are expressed in parts per
million (ppm) relative to internal tetramethylsilane (TMS);

5-HT1A- vs D2-Receptor Selectivity of Flesinoxan
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 3 309
coupling constants (J ) are in hertz; for clarity the piperazine
nitrogen atoms were numbered as in Figure 2. Elemental
analyses were performed at the Mikroanalytisches Labor
Pascher, Remagen-Bandorf, Germany. Thin-layer chroma-
tography (TLC) was run on Merck silica gel 60 F-254 plates.
For normal pressure and flash chromatography, Merck silica
gel type 60 (size 70-230 and 230-400 mesh, respectively) was
used. Unless stated otherwise, all starting materials were
high-grade commercial products. All reactions were performed
under a nitrogen atmosphere.
Compounds 28 and 29 are commercially available. Com-
pounds 1, 26, and 27 were prepared according to Patent
Application EP 0138280,18 and compounds 8 and 18 were made
according to patent application EP 0104614.19 Compounds 11,
15, 20, and 23-25 were synthesized as described in Patent
Applications ES 2027898,20 J P 60169742,21 DE 2053759,22 WO
9109594,23 FR 1349636,24 and FR 1537901,25 respectively. The
syntheses of compounds 19 and 30 have been described in refs
26 and 27, respectively. Compounds 31 and 32 were synthe-
sized according to Patent Application J P 61152655.28
The N-2-(acylamino)ethyl-substituted (2-methoxyphenyl)-
piperazines 7, 9, 10, and 12-14 were prepared by acylation
of the known N-2-aminoethyl precursor 35,29 using either the
anhydride, the acyl chloride, or the Mukayama ester of the
corresponding acids. Compounds 16 and 17 were obtained by
the method in Scheme 1 and compounds 21 and 22 by the
method in Scheme 2, and compound 33 was obtained according
to Scheme 3.
2.35 g (10 mmol) of 35 in 40 mL of CH2Cl2. The reaction
mixture was heated at reflux for 2 h, stirred at room temper-
ature for 12 h, acidified with 50 mL of 2 N HCl, and twice
extracted with 50 mL of CH2Cl2. The organic layers were
discarded, and the residue was brought at pH > 13 with Na2-
CO3 and 2 N NaOH followed by three times extraction with
70 mL of CH2Cl2. The combined organic layers were dried on
anhydrous Na2SO4, filtered once, concentrated in vacuo, and
purified by silica gel (300 g) column chromatography using
CH2Cl2-MeOH, 90:10: yield 2.2 g. The white HCl salt was
obtained by addition of 3 equiv of HCl in 50 mL of EtOH.
Recrystallization from EtOH yielded 1.60 g (39%): mp 228-
1
231 °C; H NMR (D2O) δ 3.5-3.8 (cluster, 10H, CH2 pip, N4-
CH2), 3.93 (s, 3H, OCH3), 3.98 (t, 2H, CONHCH2, J ) 6), 7.05-
7.4 (cluster, 4H, PhOCH3), 8.43 (m, 2H, pyridine H-3, H-5),
9.01 (m, 2H, pyridine H-2, H-6). Anal. (C19H24N4O2‚3HCl‚
0.25H2O) C, H, N.
N-[2-[4-(2-Met h oxyp h en yl)-1-p ip er a zin yl]et h yl]-4-p y-
r im id in eca r boxa m id e (13). A portion of 1.42 g of 4-pyrim-
idinecarboxylic acid chloride (10 mmol) was prepared from
4-pyrimidinecarboxylic acid according to a literature proce-
dure.30 It was suspended in 15 mL of CH2Cl2 and added to a
solution of 2.35 g (10 mmol) of 35 in 30 mL of CH2Cl2, resulting
in a exothermic reaction. The reaction mixture was stirred
for 1 h at room temperature and then filtered over Hyflo. The
solvent was evaporated, and the remainder was treated with
active coal in an EtOAc-EtOH mixture, recrystallized from
EtOH-i-PrOH, isolated by filtration, and washed with iso-
propyl alcohol and diethyl ether: yield 1.61 g (43%); mp 249.5-
250.5 °C; Rf 0.5 (CH2Cl2-MeOH, 90:10); 1H NMR (DMSO-
CDCl3, 4:1) δ 3.0-3.9 (cluster, 15H, N-CH2’s, OCH3), 6.9
(cluster, 4H, Ph), 8.76 (dd, 1H, pyrimidine H-6, J ) 2 and 3),
8.89 (d, 1H, pyrimidine H-5, J ) 3), 9.24 (d, 1H, pyrimidine
H-3, J ) 2), 9.31 (t, 1H, CONH, J ) 6), 10.6 (broad peak, 1H,
NH+). Anal. (C18H23N5O2‚HCl) C, H, N.
N-[2-[4-(2-Met h oxyp h en yl)-1-p ip er a zin yl]et h yl]-3-p y-
r id in eca r boxa m id e (14). Compound 14 was synthesized
analogous to compound 12, using 2.35 g (10 mmol) of 35 and
a Mukayama reagent of 1.25 g (10 mmol) of 3-pyridinecar-
boxylic acid, 3.4 mL (25 mmol) of triethylamine, and 3.2 g (12.5
mmol) of 2-chloro-1-methylpyridinium iodide, prepared ac-
cording to a literature procedure:31 yield 1.45 g (35%); mp 216-
220.5 °C; 1H NMR (DMSO-CDCl3, 4:1) δ 3.2-3.4 (cluster, 4H,
CHax pip), 3.45 (t, 2H, N4-CH2), 3.53 (d, 2H, N4-CHeq pip), 3.73
(br d, 2H, N1-CHeq pip), 3.83 (s, 3H, OCH3), 3.83 (m, 2H,
CONHCH2), 6.86-7.08 (m, 4H, Ph), 8.09 (dd, 1H, pyridine H-5,
J ) 6 and 8), 9.00 (m, 1H, pyridine H-4), 9.03 (m, 1H, pyridine
H-6), 9.46 (br s, 1H, pyridine H-2), 9.70 (t, 1H, CONH, J ) 6).
Anal. (C19H24N4O2‚3HCl) C, H, N.
N-[2-[4-(2-Meth oxyp h en yl)-1-p ip er a zin yl]eth yl]cyclo-
h exa n eca r both ioa m id e (16) (Sch em e 1, i). A total of 3.13
g (10 mmol) of 15 and 2.22 g (10 mmol) of P2S5 were suspended
in 30 mL of CH2Cl2. A 5.12-mL volume (40 mmol) of Et3N
was added while the reaction mixture was stirred and cooled.
After 24-h storage at room temperature, the whole was
extracted with water. After drying (Na2SO4), the organic layer
was loaded on a dry silica gel (170 g) column and eluted with
EtOAc, yielding 1.9 g of a yellow viscous oil (free base). The
HCl salt was obtained by addition of 1 equiv of HCl in absolute
EtOH and subsequent recrystallization from EtOH-petroleum
ether (40-60 °C): yield 1.25 g (35%); mp 151-151.5 °C; 1H
NMR (DMSO-CDCl3, 4:1) δ 1.1-1.9 (cluster, 10H, cyclohexyl),
2.68 (broad peak, 1H, cyclohexyl), 3.0-3.8 (cluster, 10H,
N-CH2’s), 3.80 (s, 3H, OCH3), 4.0 (m, 2H, CONHCH2), 6.9
(cluster, 4H, Ph), 10.24 (t, 1H, CSNH, J ) 6), 11.1 (broad peak,
1H, NH+). Anal. (C20H31N3OS‚HCl‚0.25H2O) C, H, N.
N-[2-[4-(2-Meth oxyp h en yl)-1-p ip er a zin yl]eth yl]cyclo-
h exa n eca r boxim id a m id e (17) (Sch em e 1, ii). A 5.7-g
portion (30 mmol) of cyclohexanecarboximidic acid ethyl ester
was prepared from cyanocyclohexane via the method given in
Patent Application EP 528633.33 It was added to a solution
of 7.0 g (30 mmol) of 35 in 30 mL of EtOH and kept at room
temperature for 5 days. The solvent was removed in vacuo,
taken up in absolute EtOH, and filtered. The HCl salt was
N-[2-[4-(2-Met h oxyp h en yl)-1-p ip er a zin yl]et h yl]a cet -
a m id e (7). A 3.03-g (12.8 mmol) sample of 35 was dissolved
in 20 mL of toluene, and a 1.8-mL portion (19 mmol) of acetic
anhydride was slowly added during which the temperature
increased from 20 to 40 °C. After stirring for 2 h, the solvent
was removed in vacuo, and the white product was recrystal-
lized from diethyl ether: yield 2.04 g (57%); Rf 0.25 (CH2Cl2-
1
MeOH, 95:5); mp 96-98 °C; H NMR (CDCl3) δ 1.98 (s, 3H,
COCH3), 2.4-2.8 (cluster, 6H, CH2 pip and N4-CH2), 3.09 (m,
4H, CH2 pip), 3.38 (q, 2H, CONHCH2, J ) 6), 3.84 (s, 3H,
OCH3), 6,14 (broad peak, 1H, CONH), 6.94 (cluster, 4H, Ph).
Anal. (C15H23N3O2) C, H, N.
N -[2-[4-(2-Me t h oxyp h e n yl)-1-p ip e r a zin yl]e t h yl]-2-
th iop h en eca r boxa m id e (9). A 1.50-g (16 mmol) portion of
2-thiophenecarboxylic acid chloride was prepared from 2-thio-
phenecarboxylic acid according to a literature procedure.30 It
was dissolved in 10 mL of CH2Cl2 and slowly added to a
solution of 2.34 g (18 mmol) of 35 in 10 mL of CH2Cl2, resulting
in an exothermic reaction. After cooling, the white product
crystallized from the solvent, was filtered off, and was washed
with cold CH2Cl2: yield 1.50 g (39.5%); mp 207-208 °C
(dissolution); 1H NMR (DMSO-CDCl3, 4:1) δ 3.0-3.8 (cluster,
12H, N-CH2’s), 3.81 (s, 3H, OCH3), 6.9 (cluster, 4H, Ph), 7.14
(dd, 1H, thiophene H-3, J ) 4 and 5), 7.71 (dd, 1H, thiophene
H-2, J ) 1 and 5), 7.94 (dd, 1H, thiophene H-4, J ) 1 and 4),
9.03 (t, 1H, CONH, J ) 6), 10.6 (broad peak, 1H, NH+). Anal.
(C18H23N3O2S‚HCl) C, H, N.
N-[2-[4-(2-Meth oxyph en yl)-1-piper azin yl]eth yl]-2-fu r an -
ca r boxa m id e (10). A 1.93-g portion of 2-furoic acid chloride
(15 mmol) was prepared from 2-furoic acid according to a
literature procedure.30 It was dissolved in 10 mL of CH2Cl2
and slowly added to a solution of 3.5 g (15 mmol) of 35 in 10
mL of CH2Cl2, during which a precipitate was formed. The
reaction mixture was stirred for 12 h at room temperature;
the precipitate was filtered off and recrystallized from EtOH:
yield 1.95 g (36%); mp 210-211 °C; 1H NMR (DMSO-CDCl3,
4:1) δ 2.9-3.8 (cluster, 12H, N-CH2’s), 3.80 (s, 3H, OCH3), 6.61
(dd, 1H, furan H-3, J ) 2 and 3), 6.9 (cluster, 4H, Ph), 7.18
(dd, 1H, furan H-4, J ) 1 and 3), 7.80 (dd, 1H, furan H-2, J )
2 and 1), 8.76 (t, 1H, CONH, J ) 6), 10.4 (broad peak, 1H,
NH+). Anal. (C18H23N3O3‚HCl‚0.10H2O) C, H, N.
N-[2-[4-(2-Met h oxyp h en yl)-1-p ip er a zin yl]et h yl]-4-p y-
r id in eca r boxa m id e (12). The intermediate Mukayama ester
of 4-pyridinecarboxylic acid was prepared in situ from 1.23 g
(10 mmol) of the acid, 3.4 mL (25 mmol) of triethylamine, and
3.2 g (12.5 mmol) of 2-chloro-1-methylpyridinium iodide in 80
mL of CH2Cl2, according to a literature procedure.31 The
Mukayama reagent was subsequently added to a solution of

310 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 3
Kuipers et al.
obtained by addition of 1 equiv of HCl in absolute EtOH and
subsequent recrystallization from absolute EtOH: yield 6.7 g
(59%); mp >258 °C (dissolution); 1H NMR (DMSO-CDCl3, 4:1)
δ 1.1-1.9 (cluster, 10H, cyclohexyl), 2.6 (broad peak, 1H,
cyclohexyl), 3.1-3.9 (cluster, 12H, N-CH2’s), 3.82 (s, 3H,
OCH3), 6.9 (cluster, 4H, Ph), 9.20 (broad peak, 1H, imino
protonated), 9.31 (broad peak, 1H, imino protonated), 9.82 (t,
1H, CONH, J ) 6), 11.4 (broad peak, 1H, NH+). Anal.
(C20H32N4O‚2HCl‚0.50H2O) C, H, N.
1-[4-(4-F lu o r o p h e n y l)-1-b u t -3-e n y l]-4-(2-m e t h o x y -
p h en yl)p ip er a zin e (22) (Sch em e 2, i, ii). The intermediate
compound 4-(4-fluorophenyl)-3-buten-1-ol (36) was synthesized
according to a literature procedure.34 To a solution of 2 g (12
mmol) of 36 and 5 mL (34 mmol) of Et3N in 100 mL of CH2Cl2
at -20 °C (CO2/acetone) was added 5 mL of methanesulfonyl
chloride (26 mmol) while stirring. The reaction mixture was
stirred at 0 °C for 30 min and then for 12 h at room
temperature, poured into 100 mL of 1 N HCl, and extracted
twice with 100 mL of CH2Cl2. The combined organic layers
were dried (Na2SO4), and the solvent was removed in vacuo,
yielding 3.17 g of a light-brown oil.
could not be stirred. The mixture was heated at 90 °C
overnight. After cooling, 1.5 L of n-pentane was added, and
the precipitate was removed using suction filtration and
washed three times with large volumes of n-pentane. The
solvent of the filtrate was evaporated in vacuo at 50 °C yielding
235.1 g of a mixture of the meso and d,l forms of 38 (99.5%)
as light-brown crystals.
N-(1-Ca r b oxyet h yl)-N-(p h en ylm et h yl)a la n in e (39)
(Sch em e 3, ii). A 139.65-g portion (0.5 mol) of 38 was
dissolved in 585 mL of MeOH, and a solution of 96.2 g of 87.5%
KOH (1.5 mol) in 300 mL of water was added. This mixture
was stirred and heated at reflux for 4.5 h. After cooling the
reaction mixture was concentrated to ca. 250 mL using rotary
evaporation and filtered over Hyflo. The filtrate was heated
on a steam bath to prevent crystallization of the dipotassium
salt, and 125 mL of 12 N HCl was added in portions. At the
last portion the desired product started crystallizing, and the
gray solid was collected at room temperature. The crude
product was recrystallized by concentrating a solution in 4.5
L of absolute EtOH to ca. 1.75 L on a steam bath. After cooling
the product was collected, washed with diethyl ether, and dried
in vacuo yielding 97.9 g of 39 (77.9%).
The oil was dissolved in 100 mL of acetonitrile, and 2.26
mL of diisopropylethylamine (13 mmol) was added followed
by a 2.28-mL portion of N-(2-methoxyphenyl)piperazine (13
mmol). The reaction mixture was stirred for 12 h at 60 °C.
The solvent was removed, and the residue was taken up in
EtOAc and extracted with 200 mL of 2 N NaOH and 100 mL
of saturated NaCl solution, respectively. The organic layer
was dried (Na2SO4), the solvent was removed in vacuo, and
the residue was purified over 300 mL of silica gel using EtOAc
as eluent. Solvent evaporation of the combined fractions
yielded 2.6 g of light-yellow oil (free base of compound 22).
A 1.25-g portion (3.7 mmol) of this oil was dissolved in 15
mL of EtOAc, and a solution of 0.43 g (3.7 mmol) of fumaric
acid in 2.5 mL of absolute EtOH was added. The solvent was
removed, and the residue was stirred in diisopropyl ether,
collected by filtration, and washed with diisopropyl ether and
petroleum ether. The obtained white solid was dried in vacuo
at 38 °C above KOH: yield 1.38 g (82%); mp 165-167 °C
(dissolution); 1H NMR (DMSO-CDCl3, 4:1) δ 2.44 (m, 2H,
HdCHCH2), 2.63 (m, 2H, N4-CH2’s), 2.70 (m, 4H, N4-CH2 pip),
3.04 (m, 2H, N1-CH2 pip), 3.80 (s, 3H, OCH3), 6.21 (dt, 1H,
HdCHCH2, J ) 7, 13 and 16), 6.44 (broad peak, 1H, Ph-
CHdCH, J ) 16), 6.62 (s, 2H, fum), 6.85-7.0 (cluster, 4H,
PhOCH3), 7.06 (m, 2H, p-F-Ph meta H’s), 7.38 (m, 2H, p-F-Ph
ortho H’s). Anal. (C21H25FN2O‚C4H4O4) C, H, N.
1-[4-(4-F lu or op h en yl)-1-b u t yl]-4-(2-m et h oxyp h en yl)-
p ip er a zin e (21) (Sch em e 2, iii). A 1.37-g portion (4 mmol)
of 22 was dissolved in 100 mL of EtOH with 0.5 g of 10% Pd
on C catalyst and hydrogenated at room temperature for 2 h.
The catalyst was removed by filtration over Hyflo followed by
rinsing with EtOH. The solvent of the filtrate was removed
in vacuo yielding 1.2 g of a light-yellow oil (free base of
compound 21). This was taken up in 25 mL of EtOAc and a
solution of 0.41 g of fumaric acid (3.5 mmol) in 4 mL of absolute
EtOH was added during which a white precipitate was formed.
This was stirred at room temperature for 1 h and then collected
and washed with EtOH, diisopropyl ether, and petroleum
ether. Drying in vacuo at 38 °C above KOH yielded 1.36 g of
21 (74%): mp 170-171 °C (dissolution); 1H NMR (DMSO-
CDCl3, 4:1) δ 1.48-1.66 (cluster, 4H, NCH2CH2CH2), 2.51 (t,
2H, PhCH2), 2.60 (t, 2H, N4-CH2), 2.67 (m, 4H, N4-CH2 pip),
3.02 (m, 4H, N1-CH2 pip), 3.79 (s, 3H, OCH3), 6.61(s, 2H, fum),
6.83-6.98 (cluster, 4H, PhOCH3), 7.02 (m, 2H, p-F-Ph meta
H’s), 7.20 (m, 2H, p-F-Ph ortho H’s). Anal. (C21H27FN2O‚
C4H4O4) C, H, N.
1-(2,3-Dih yd r o-1,4-b en zod ioxin -5-yl)-3,5,-d im et h yl-4-
(p h en ylm eth yl)-2,6-p ip er a zin ed ion e (40) (Sch em e 3, iii).
Compound 40 was synthesized from 7.8 g (31 mmol) of 39, 4.7
g (31 mmol) of 5-aminobenzo-1,4-dioxane, and 5 g (31 mmol)
of N,N ′-carbonyldiimidazole in THF according to a literature
procedure:14 yield 6.7 g (59%).
1-(2,3-Dih yd r o-1,4-ben zod ioxin -5-yl)-3,5-d im eth ylp ip -
er a zin e (41, 42) (Sch em e 3, iv, v). A mixture of compounds
41 and 42 was obtained via hydrogenation of 18.3 mmol of 40
followed by reaction with 2.8 equiv of borane-methyl sulfide
complex in THF and 6 N HCl, similar to a procedure described
before.14 The cis- and trans-isomers 41 and 42 were separated
using flash chromatography (CH2Cl2-MeOH-ammonia, 97:
2.5:0.5), giving 1.7 g of 41, 2.6 g of 42, and 1.4 g of a cis/trans-
mixture: total yield 5.7 g (63%).
N-[2-[4-(2,3-Dih yd r o-1,4-b en zod ioxin -5-yl)-cis-2,6-d i-
m e t h yl-1-p ip e r a zin yl]e t h yl]-4-flu or ob e n ze n e ca r b ox-
a m id e (33) (Sch em e 3, vi). A 1.7-g sample (6.85 mmol) of
the cis-isomer 41 was dissolved in 50 mL of acetone, and 1.29
g (7.7 mmol) of N-(4-fluorobenzoyl)aziridine (obtained via a
literature procedure)36 was added. The solvent was removed
in vacuo from the reaction mixture using rotary evaporation,
yielding a light-brown solid. The fumarate was obtained by
dissolving the solid in 20 mL of absolute EtOH and addition
of 1 equiv of fumaric acid in 10 mL of absolute EtOH followed
by concentration to ca. 10 mL (rotary evaporation). A 75-mL
volume of diethyl ether was added, and the desired compound
started crystallizing after scratching the laboratory glass. The
product was collected as white crystals, washed with diethyl
ether, and air-dried: yield 1.0 g (28%); mp 213-217 °C; 1H
NMR (DMSO-CDCl3, 4:1) δ 1.14 (d, 6H, -CH3, J ) 6), 2.32 (t,
2H, N1-CHax pip, J ) 10), 2.8 (cluster, 4H, N4-CH2’s), 3.25
(broad peak, 2H, N1-CHeq pip, J ) 10), 3.34 (m, 2H, CON-
HCH2), 4.22 (m, 4H, Bzd CH2’s), 6.42 (dd, 1H, Bzd ortho H, J
) 2 and 8), 6.48 (dd, 1H, Bzd para H, J ) 2 and 8), 6.62 (s,
2H, fum), 6.69 (t, 1H, Bzd meta H, J ) 8), 7.25 (m, 2H, p-F-Ph
meta H’s), 7.9 (m, 2H, p-F-Ph ortho H’s), 8.55 (t, 1H, CONH,
J ) 6). Anal. (C23H28FN3O3‚0.5C4H4O4‚0.5H2O) C, H, N.
log P Deter m in a tion . Partition coefficients (n-octanol/
water) were measured by a high-performance liquid chromato-
graphic (HPLC) method based on an OECD (Organisation for
Economic Co-operation and Development) method,37 using a
mobile phase buffered at pH > 11 and an aluminum-based
octadecyl-modified stationary phase. The measured retention
factor k of a compound was correlated with its partition
coefficient using a calibration graph based on 10 reference
compounds with well-known log Pow values in the range 2.1-
5.7.
N-[2-(1-Met h oxy-1-oxop r op yl)]-N-(p h en ylm et h yl)a la -
n in e Meth yl Ester (38) (Sch em e 3, i). N-(Phenylmethyl)-
D-alanine methyl ester (37)35 and 3-[[(trifluoromethyl)sulfonyl]-
oxy]propanoic acid methyl ester13 were synthesized according
to published procedures. A 229.7-g portion (0.97 mol) of
racemic 3-[[(trifluoromethyl)sulfonyl]oxy]propanoic acid meth-
yl ester was mixed with 177 mL of diisopropylethylamine.
While cooling, a 163.4-g (0.845 mol) portion of 37 was slowly
added, resulting in a very exothermic reaction and producing
a viscous precipitate. Consequently, the reaction mixture
Cr ysta l Str u ctu r e Deter m in a tion a n d Refin em en t of
33. Determination was carried out (monoclinic, space group
P21/c (No. 14) with a ) 10.4183(6), b ) 10.7524(5), and c )
21.656(2) Å; â ) 107.218(5)°, V ) 2317.2(2) Å3; Z ) 4, R1
)
0.0595, ωR2 ) 0.1438, S ) 1.04). A colorless, plate-shaped
crystal was cut to size and glued to the tip of a Lindemann

5-HT1A- vs D2-Receptor Selectivity of Flesinoxan
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 3 311
glass capillary and transferred to an Enraf-Nonius CAD4-T
diffractometer equipped with a rotating anode. Unit-cell
parameters and an orientation matrix were determined by
least-squares refinement of 25 well-centered reflections (SET4)38
in the range 10.0° < Θ < 14.0°. Reduced-cell calculations did
not indicate higher lattice symmetry.39 All data were collected
in ω scan mode. Data were corrected for Lp effects and the
observed linear decay of 2% of the three periodically measured
reference reflections. The structure was solved by automated
direct methods (SHELXS86).40 Refinement on F 2 was carried
out by full-matrix least-squares techniques (SHELXL-93);41 no
observance criterion was applied during refinement. All non-
hydrogen atoms were calculated and refined with anisotropic
thermal parameters. The hydrogen atoms were refined with
a fixed isotropic thermal parameter amounting to 1.5 or 1.2
times the value of the equivalent isotropic thermal parameter
of their carrier atoms, for the hydrogen atoms on the methyl
groups and the other hydrogen atoms, respectively. The
hydrogens on N were refined without constraints. Weights
were introduced in the final refinement cycles. Crystal data
and details on data collection and refinement, as well as final
positional parameters, can be obtained as Supporting Informa-
tion. Neutral atom scattering factors and anomalous disper-
sion corrections were taken from International Tables for
Crystallography.42 Geometrical calculations and illustrations
were performed with PLATON.43 All calculations were per-
formed on a DEC station 5000/125.
were kept fixed. The Tripos force field tends to underestimate
the conjugation of the nitrogen lone pair with the aromatic
ring system. Therefore, the torsion angle τ1 was constrained
at the current values as calculated by MOPAC/AM1 (force
constant was 100 kcal/(mol deg2)). The distance between the
Asp116 oxygen atoms and the HN4 atom was constrained at
2.0 Å (force constant was 100 kcal/(mol Å)2).
Bioch em istr y. 1. Recep tor Bin d in g Assa y. Radioli-
gand binding studies were carried out using [3H]-2-(di-n-
propylamino)-8-hydroxytetralin (8-OH-DPAT)
44 and [3H]-8-[3-
(p-fluorobenzoyl)propyl]-1-phenyl-1,3,8-triazaspiro[4.5]decan-
4-one (spiperone)45 as radioligands on rat frontal cortex and
rat corpus striatum, respectively.
2. In Vitr o F u n ction a l Assa y. CHO cells, stably express-
ing human 5-HT1A receptors, were obtained from Allelix
Biopharmaceuticals Inc., Ontario, Canada.
Cells were cultured in R-DMEM (Dulbecco’s Modified Eagle’s
Medium) containing 10% fetal calf serum. Three days prior
to experimentation, cells were seeded in 24-well plates with a
density of 0.5 × 105 cells/well. Assessment of adenylate cyclase
activity was done essentially according to Weiss et al.46 Cells
were incubated with [3H]adenine (NEN; 2 µCi/mL) in R-DMEM
for 2 h. Subsequently, cells were rinsed once with PBS
containing 10 mM IBMX, after which cells were stimulated
for 10 min in PBS/IBMX with forskolin (10-7 M) in the
presence of compounds at various concentrations, determined
in quadruplicate. Cells were extracted with 5% TCA, contain-
ing excess nonradioactive ATP and cAMP. [3H]cAMP and [3H]-
ATP were separated by sequential chromatography over
Dowex 50-WX4 and aluminum oxide according to Salomon et
al.,47 and eluates were analyzed for radioactivity by liquid
scintillation counting.
Mod elin g. 1. Softw a r e a n d Ha r d w a r e. Small-ligand
building, docking procedures, and all computations (MAXI-
MIN, MOPAC) were performed with the SYBYL package,
version 6.1a (Tripos Associates, Inc., St. Louis, MO), running
on a Silicon Graphics Iris Indigo Elan 4000. For MAXIMIN
calculations (Tripos force field), the Powell method was chosen
(default values).
5-HT1A-receptor-mediated inhibition of forskolin-stimulated
accumulation of cAMP was analyzed by a four-parameter
logistic regression function, using the nonlinear curve-fitting
program INPLOT (ISI Software, Philadelphia, PA). EC50
values of three independent experiments were obtained and
are presented as mean value ( SEM.
2. Liga n d Bu ild in g a n d Min im iza tion s. The geometry
of the crystal structure of compound 33 was used as a basis
for modeling the other ligands. All phenylpiperazines were
built by modification of this structure with the SKECTCH
option in SYBYL. For the study of the first torsion angle τ2
with MOPAC (AM1, CHARGE)1, PRECISE),16 the structures
of compounds 43 and 44 (Figure 3) were minimized over all
bonds and angles, except the torsion angle τ2, which was
constrained at values between 0° and 360° (10° intervals). The
crystal structure coordinates of 33 were also used in the
RANDOMSEARCH option of the SYBYL program for the
study of τ3 and τ4 (default values were used; a chirality check
was applied; energy cutoff was at 120 kcal/mol; the arylpip-
erazine moiety was constrained by defining it as an aggregate).
It was previously shown that MOPAC/AM1 is more suitable
for full energy minimization of arylpiperazine structures than
molecular mechanics (MM) calculations (e.g., Tripos force
field).15 Therefore, structures that emerged from the RAN-
DOMSEARCH analysis were further minimized with MOPAC/
AM1 (default values; additional keywords MMOK, PRECISE,
and CHARGE)1). Only low-energy conformations (energy less
than 3 kcal/mol above the absolute minimum energy) were
investigated further.
3. Mod el Bu ild in g a n d Dock in g. A model of the 5-HT1A
receptor was built according to Kuipers et al.11,12 The position
of eltoprazine (31) in the receptor model, which is described
in ref 11, was taken as a starting point. The compounds 1,
27, and 33 were docked into the receptor model in conforma-
tions A-H, by fitting them on (docked) compound 31. The
relative orientation with respect to 31 was derived by a
straightforward fit of the arylpiperazine parts. Two atoms in
the phenyl ring (C3′ and C5′), the basic nitrogen atom, and
its proton were used as fitting points. Subsequently, eltopra-
zine (31) was removed from the model. Conformations C-G
could fit the model without causing severe hindrance with
backbone atoms. The complexes of each compound 1, 27, and
33 in conformations C-G with the receptor were energy-
minimized using molecular mechanics calculations. For this
purpose, an “active site” was created, containing all side chains
of residues within a distance of 4 Å from the ligand (see Figure
8). These side chains, and the ligand itself, were allowed to
optimize their position and conformation; the backbone atoms
Ack n ow led gm en t. We wish to thank Erik Ronken
and Annemarije Menkveld of the Pharmacology Depart-
ment, CNS Cell Biology, for determination of the
functional properties of compounds 27 and 33. We
thank Piet Hoogkamer and Gemma Feenstra-Bielders
for the experimental determination of log P values. We
also thank Rolf Feenstra, Paul van Vliet (IRI, Delft, The
Netherlands), and Koos Tipker for helpful discussions
and Petra Lelieveld and Netty Rengelink for their
substantial secretarial support.
Su p p or tin g In for m a tion Ava ila ble: Crystal data and
details on data collection and crystal structure determination,
including final positional parameters, atomic coordinates for
the hydrogen atoms, bond lengths and angles, thermal pa-
rameters, and a thermal motion ellipsoid plot for 33 (8 pages).
Listings of observed and calculated structure factor amplitudes
for 33 (13 pages). Ordering information is given on any
current masthead page.
Refer en ces
(1) Saudou, F.; Hen, R. 5-Hydroxytryptamine Receptor Subtypes in
Vertebrates and Invertebrates. Neurochem. Int. 1994, 25, 503-
532.
(2) Peroutka, S. J . Serotonin Receptor Subtypes. Their Evolution
and Clinical Relevance. CNS Drugs 1995, 4 (Suppl. 1), 18-28.
(3) Middlemiss, D. N.; Fozard, J . R. 8-Hydroxy-2-(di-n-propylamino)-
tetralin Discriminates between Subtypes of the 5-HT1 Recogni-
tion Site. Eur. J . Pharmacol. 1983, 90, 151-153.
(4) van Wijngaarden, I.; Tulp, M. Th. M.; Soudijn, W. The Concept
of Selectivity in 5-HT Receptor Research. Eur. J . Pharmacol.-
Mol. Pharmacol. Sect. 1990, 188, 301-312.
(5) Bjo¨rk, L.; Cornfield, L. J .; Nelson, D. L.; Hillver, S. E.; Anden,
N. E.; Lewander, T. Pharmacology of the Novel 5-Hydroxy-
tryptamine Receptor Antagonist (S)-5-Fluoro-8-Hydroxy-2-
(Dipropylamino) Tetralin: Inhibition of (R)-8-Hydroxy-2-(Dipro-
pylamino) Tetralin-Induced Effects. J . Pharmacol. Exp. Ther.
1991, 258, 58-65.

312 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 3
Kuipers et al.
(6) Cliffe, I. A.; Brightwell, C. I.; Fletcher, A.; Forster, E. A.; Mansell,
H. L.; Mansell, Y. R.; Routledge, C.; White, A. C. (S)-N-tert-Butyl-
3-(4-(2-methoxyphenyl)piperazin-1-yl)-2-phenylpropanamide [(S)-
(26) Soudijn, W.; van Wijngaarden, I.; Allewijn, F. T. N. Distribution,
Excretion and Metabolism of Neuroleptics of the Butyrophenone
Type. I. Excretion and Metabolism of Haloperidol and Nine
Related Butyrophenone Derivatives in Wistar Rats. Eur. J .
Pharmacol. 1967, 1, 47-57.
(27) Van Wijngaarden, I.; Kruse, C. G.; van der Heyden, J . A. M.;
Tulp, M. Th. M. 2-Phenylpyrroles as Conformationally Restricted
Benzamide Analogues. A New Class of Potential Antipsychotics.
2. J . Med. Chem. 1988, 31, 1934-1940.
(28) Preparation of Piperazines as Psychotropics. Patent Application
J P61152655; Chem. Abstr. 1986, 106, 5080.
WAY-100135]:
A Selective Antagonist at Presynaptic and
Postsynaptic 5-HT1A Receptors. J . Med. Chem. 1993, 36, 1509-
1510.
(7) Glennon, R. A.; Naiman, N. A.; Lyon, R. A.; Titeler, M. Arylpip-
erazine Derivatives as High-Affinity 5-HT1A Serotonin Ligands.
J . Med. Chem. 1988, 31, 1968-1971.
(8) For examples, see: (a) Abou-Gharbia, M.; Patel, U. R.; Webb,
M. B.; Moyer, J . A.; Andree, T. H.; Muth, E. A. Polycyclic Aryl
and Heteroarylpiperazinyl Imides as 5-HT1A Receptor Ligands
and Potential Anxiolytic Agents: Synthesis and Structure-
Activity Relationship Studies. J . Med. Chem. 1988, 31, 1382-
1392. (b) Scott, M. K.; Martin, G. E.; DiStefano, D. L.; Fedde, C.
L.; Kubkla, M. J .; Barrett, D. L.; Baldy, W. J .; Elgin, R. J ., J r.;
Kesslick, J . M.; Mathiasen, J . R.; Shank, R. P.; Vaught, J . R.
Pyrrole Mannich Bases as Potential Antipsychotic Agents. J .
Med. Chem. 1992, 35, 552-558. (c) Peglion, J .; Canton, H.;
Bervoets, K.; Audinot, V.; Brocco, M.; Gobert, A.; Le Marouille-
Girardon, S.; Millan, M. J . Characterization of Potent and
Selective Antagonists at Postsynaptic 5-HT1A Receptors in a
Series of N4-Substituted Arylpiperazines. J . Med. Chem. 1995,
38, 4044-4055.
(29) Mull, R. P.; Mizzoni, R. H.; Dapero, M. R.; Egbert, M. E.
Guanidines with Antihypertensive Activity. II. J . Med. Pharm.
Chem. 1962, 5, 944.
(30) Barcock, R.; Chadwick, D.; Storr, R.; Fuller, L.; Young, J .
Synthesis of Some Novel Imidate Derivatives of Thiophene and
Furan: Investigations of Their Methylation Properties and Some
Synthetic Applications. Tetrahedron 1994, 50, 4149.
(31) Degani, Y.; Heller, A. Direct Electrical Communication Between
Chemically Modified Enzymes and Metal Electrodes. 2. Methods
for Bonding Electron-Transfer Relays to Glucose Oxidase and
D-Amino-Acid Oxidase. J . Am. Chem. Soc. 1988, 110, 2615.
(32) Feng, Z.; Gollamudi, R.; Purcell, W.; Hill, R.; Korfmacher, W.
Molecular Determinants of the Platelet Aggregation Inhibitory
Activity of Carbamoylpiperidines. J . Med. Chem. 1992, 35, 2952.
(33) Preparation of N-(Oxoalkyl)-5-(Acylamino)-6-Oxopyrimidin-1-yl-
acetamides as Elastase Inhibitors. Patent Application EP0528633;
Chem. Abstr. 1993, 119, 72617.
(34) Cheik-Rouhou, F.; Le Bigot, Y.; El Gharbi, R.; Delmas, M.; Gaset,
A. A Simplified Wittig Synthesis Using a Solid-Liquid Transfer
Process: VI. Synthesis of γ-Unsaturated Alcohols from Aromatic
and Heteraoaromatic Aldehydes. Synth. Commun. 1986, 16,
1617.
(35) Kubota, H.; Kubo, A.; Takahashi, H.; Shimizu, R.; Date, T.;
Okamura, K.; Nunami, K. Stereospecific Amination by Dynamic
Kinetic Resolution Utilizing 2-Oxoimidazolidine-4-carboxylate
as a Novel Chiral Auxiliary. J . Org. Chem. 1995, 60, 6776.
(36) Pettit, G.; Gupta, S.; Whitehouse, P. Antineoplastic Agents
XVIII. N-(2-Haloethyl)benzylamines. J . Med. Chem. 1967, 10,
692-696.
(9) Mokrosz, J . L.; Pietrasiewicz, M.; Duszynska, B.; Cegla, M. T.
Structure Activity Relationship Studies of Central Nervous
System Agents. 5. Effect of the Hydrocarbon Chain on the
Affinity of 4-Substituted 1-(3-chlororphenyl)piperazines for 5-HT1A
Receptor Site. J . Med. Chem. 1992, 35, 2369-2374.
(10) van Steen, B. J .; van Wijngaarden, I.; Tulp, M. Th. M.; Soudijn,
W. Structure-Affinity Relationship Studies on 5-HT1A Receptor
Ligands. 1. Heterobicyclic Phenylpiperazines with N4-Alkyl
Substituents. J . Med. Chem. 1993, 36, 2751-2760.
(11) Kuipers, W.; van Wijngaarden, I.; Kruse, C. G.; ter Horst-van
Amstel, M.; Tulp, M. Th. M.; IJ zerman, A. P. N4-unsubstituted
N
1-Arylpiperazines as High-Affinity 5-HT1A Receptor Ligands.
J . Med. Chem. 1995, 38, 1942-1954.
(12) Kuipers, W.; van Wijngaarden, I.; IJ zerman, A. P. A Model of
the Serotonin 5-HT1A Receptor: Agonist and Antagonist Binding
Sites. Drug Des. Discovery 1994, 11, 231-249.
(13) Effenberger, F.; Burkard, U.; Willfahrt, J . Trifluoromethane-
sulfonates of R-Hydroxycarboxylates - Educts for the Racemiza-
tion - Free Synthesis of N-substituted R-Amino Acids. Angew.
Chem., Int. Ed. Engl. 1983, 22, 65.
(37) OECD Guideline for Testing of Chemicals No. 117, adopted at
Mar. 30, 1989.
(38) de Boer, J . L.; Duisenberg, A. J . M. Diffractometer Software
(CAD4F). Acta Crystallogr. 1984, A40, C410.
(14) Kruse, C. G.; Troost, J . J .; Cohen-Fernandes, P.; van der Linden,
H.; van Loon, J . D. Single-Step Conversion of Aliphatic, Aro-
matic, and Heteroaromatic Primary Amines into Piperazine-2,6,-
diones. Recl. Trav. Chim. Pays-Bas 1988, 107, 303-309.
(15) Dijkstra, G. D. H. Conformational Analysis of 1-Arylpiperazines
and 4- Arylpiperidines. Recl. Trav. Chim. Pays-Bas 1993, 112,
151-160.
(16) Stewart, J . J . P. MOPAC: A Semiempirical Molecular Orbital
Program. J . Comput.-Aided Mol. Des. 1990, 4, 1-105.
(17) van Steen, B. J .; van Wijngaarden, I.; Tulp, M. Th. M.; Soudijn,
W. Structure-Affinity Relationship Studies on 5-HT1A Receptor
Ligands. 2. Heterobicyclic Phenylpiperazines with N4-Aralkyl
Substituents. J . Med. Chem. 1994, 37, 2761-2773.
(18) Blood-Pressure Lowering Piperazine Derivatives. Patent Ap-
plication EP0138280; Chem. Abstr. 1985, 103, 123520.
(19) Phenylpiperazine Derivatives. Patent Application EP0104614;
Chem. Abstr. 1984, 101, 110952.
(20) Process for Preparation of New (2-Methoxyphenyl)piperazine
Derivatives with 5-HT1A Receptor Activity. Patent Application
ES2027898; Chem. Abstr. 1992, 118, 124560.
(21) Phenylpiperazine Derivatives. Patent Application J P60169472;
Chem. Abstr. 1985, 104, 109689.
(22) 1-Hydroxy- and 1-Oxo-1-phenyl- (or Heteroaryl)-4-(1-piperazi-
nyl)butanes. Patent Application DE2053759; Chem. Abstr. 1971,
75, 49136.
(23) Preparation of Substituted Phenylisopropylamines and Analogs
as Sigma Receptor Ligands for Treatment of Schizophrenia and
Psychoses. Patent Application WO9109594; Chem. Abstr. 1991,
115, 182801.
(39) Spek, A. L. LEPAGE - An MS-DOS Program for the Determi-
nation of the Metrical Symmetry of a Translation Lattice. J .
Appl. Crystallogr. 1988, 21, 578-579.
(40) Sheldrick, G. M. SHELXS86 Program for Crystal Structure
Determination; University of Go¨ttingen: Go¨ttingen, Germany,
1986.
(41) Sheldrick, G. M. SHELXL-93 Program for Crystal Structure
Refinement; University of Go¨ttingen: Go¨ttingen, Germany, 1993.
(42) Wilson, A. J . C., Ed. International Tables for Crystallography;
Kluwer Academic Publishers: Dordrecht, The Netherlands,
1992; Vol. C.
(43) Spek, A. L. PLATON, An Integrated Tool for the Analysis of the
Results of Single Crystal Structure Determination. Acta Crys-
tallogr. 1990, A46, C34.
(44) Gozlan, H.; El Mestikawy, S.; Pichet, L.; Glowinsky, J .; Hamon,
M. Identification of Presynaptic Serotonin Autoreceptors Using
a New Ligand: [3H]-PAT. Nature 1983, 305, 140-142.
(45) Creese, I.; Schneider, R.; Snyder, S. H. [3H]Spiroperidol Labels
Dopamine Receptors in Pituitary and Brain. Eur. J . Pharmacol.
1977, 46, 377-381.
(46) Weiss, S.; Sebben, M.; Bockaert, J . Corticotropin-Peptide Regu-
lation of Intracellular Cyclic AMP Production in Cortical Neu-
rons in Primary Culture. J . Neurochem. 1985, 45, 869-874.
(47) Salomon, Y.; Londos, C.; Rodbell, M. A Highly Sensitive Ade-
nylate Cyclase Assay. Anal. Biochem. 1974, 58, 541-548.
(48) Sequences were obtained from the SWISSPROT database, A.
Bairoch, Dept. Biochimie Medicinale, Centre Medicinal Univer-
sitaire, 1211 Geneva 4, Switzerland, and SWISS-PROT Protein
sequence database, EMBL Data Library, D-69117 Heidelberg,
Germany.
(24) Nouveaux Derives de la Piperazine, Utilisables en Therapeu-
tique. Patent Application FR1349636; World Patent Index 1966,
10820F.
(25) N-(4-Phenyl-1-Piperazinyl)alkyl Halo (and Nitro) Benzamides.
Patent Application FR1537901; Chem. Abstr. 1969, 71, 70637.
(49) Spek, A. L. PLUTON Molecular Graphics Program; Utrecht Uni-
versity: Utrecht, The Netherlands, 1995.
J M960496O