Synthesis and structure-activity relationships of N-(3-phenylpropyl)-N′-benzylpiperazines: Potent ligands for σ1 and σ2 receptors

Article abstract of DOI:10.1016/j.bmc.2007.10.037

Ten N-(3-phenylpropyl)-N′-benzylpiperazines having different substituents on the benzyl moiety were synthesized and evaluated for σ₁ and σ₂ receptor binding. The σ₁ affinities were 0.37-2.80 nM, σ₂ affinities were 1.03-34.3 nM, and selectivities, as σ₂/σ₁ affinity ratios, ranged from 1.4 to 52. Three compounds tested in a phenytoin shift binding assay profiled as probable σ₁ antagonists. Quantitative structure-activity relationships depended on π_x, MR or E_s and Hammett σ values. The hydrophobicity term is negative for σ₁ binding but positive for σ₂ binding, indicating a major difference between the pharmacophores.

Full text of DOI:10.1016/j.bmc.2007.10.037

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry 16 (2008) 755–761
Synthesis and structure–activity relationships
of N-(3-phenylpropyl)-N0-benzylpiperazines:
Potent ligands for r1 and r2 receptors
Roger I. Nahas,a John R. Leverc,d,e and Susan Z. Levera,b,
*
aDepartment of Chemistry, University of Missouri-Columbia, Columbia, MO 65212, USA
bMU Research Reactor Center, University of Missouri-Columbia, Columbia, MO 65212, USA
cDepartment of Radiology, University of Missouri-Columbia, Columbia, MO 65212, USA
dDepartment of Medical Pharmacology and Physiology, University of Missouri-Columbia, Columbia, MO 65212, USA
eResearch Service, Harry S. Truman Veterans Administration Medical Center, Columbia, MO 65201, USA
Received 7 August 2007; revised 30 September 2007; accepted 10 October 2007
Available online 17 October 2007
Abstract—Ten N-(3-phenylpropyl)-N0-benzylpiperazines having different substituents on the benzyl moiety were synthesized and
evaluated for r1 and r2 receptor binding. The r1 affinities were 0.37–2.80 nM, r2 affinities were 1.03–34.3 nM, and selectivities,
as r2/r1 affinity ratios, ranged from 1.4 to 52. Three compounds tested in a phenytoin shift binding assay profiled as probable r1
antagonists. Quantitative structure–activity relationships depended on px, MR or Es and Hammett r values. The hydrophobicity
term is negative for r1 binding but positive for r2 binding, indicating a major difference between the pharmacophores.
ꢀ 2007 Elsevier Ltd. All rights reserved.
1. Introduction
N
R
N
Recent reports indicate that sigma (r) receptor antago-
nists attenuate the effects of (ꢀ)-cocaine in animal mod-
1
2
els.1 Anti-cocaine activity correlates with affinities for
both r1 and r2 receptors, but the r1 relationships appear
stronger and (ꢀ)-cocaine itself displays 10-fold higher
affinity for r1 over r2 receptors.1–3 Notable active series
(Fig. 1) include N-benzyl-N0-benzylpiperazines2 (1) and
N-phenylpropyl-N0-phenethylpiperazines3 (2) although
the structural features that confer antagonist activity are
not well defined. In the present study, we investigated
quantitative structure–activity relationships (QSARs)
for the r1 and r2 receptor binding of N-phenylpropyl-
N0-benzylpiperazines (3) that represent hybrids of 1 and
2, and might also be potent r receptor antagonists.
R
N
N
N
N
R
3 a, R = H; b, 2-Br; c, 2-NO2;
d, 3-I; e, 3-F; f, 3-OCH3; g, 3-NO2;
h, 4-OCH3; i, 4-NO2; j, 4-CH3
Figure 1. Prototype sigma ligands 1 and 2; ligands (3) synthesized in
this study.
Abbreviations: QSAR, Quantitative structure–activity relationship;
(+)PPP, (+)-3-(3-hydroxyphenyl)-N-(1-propyl)piperidine; SA4503, 1-(2-
(3,4-dimethoxy-phenyl)ethyl)-4-(3-phenylpropyl)piperazine; DAMGO,
[D-Ala2,N-MePhe4,Gly5-ol]enkephalin; U69,593, (+)-(5a,7a,8b)-N-
methyl-N-[7-(pyrrolidin-1-yl)-1-oxaspiro[4,5]dec-8-yl]benzeneacetamide.
The parent compound, 3a, was shown by Younes and
colleagues4 to exhibit an IC50 of 20 nM against the bind-
ing sites labeled by [3H](+)PPP, but the r1/r2 receptor
selectivity and functional activity were not defined.
Substituted benzyl derivatives of this lead compound
*
Corresponding author. Tel.: +1 573 882 8395; fax: +1 573 882
2754; e-mail: levers@missouri.edu
0968-0896/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2007.10.037

756
R. I. Nahas et al. / Bioorg. Med. Chem. 16 (2008) 755–761
m-I
1.5
1.0
0.5
0.0
-0.5
1.5
1.0
0.5
0.0
-0.5
m-I
1.5
1.0
0.5
0.0
-0.5
p-OCH3
m-I
m, p-OCH3
o-Br
p-CH3
o-Br
m-NO2
o-Br
p-NO2
m-OCH3
m-F
o-NO2
p-CH3
p-CH3
m-F
m-OCH
o, m, p-NO2
m-F
p-OCH3
3m-NO2
p-NO2
o-NO2
-0.5
0.0
0.5
1.0
-0.5
0.0
0.5
1.0
1.5
-0.5
0.0
0.5
1.0
Hydrophobicity
Hammett Sigma Values
Hammett Sigma Values
Scheme 1. Craig plots for r, p, and MR values for substituents in calibration sets.
Table 1. In vitro affinity and selectivity of compounds 3a–3j for r1 and
r2 receptorsa
have not been reported, but might provide sharp corre-
lations in QSAR studies.5 Thus, we selected a set of
congeners (3a–3j, Fig. 1) with substituents chosen
according to the factorial design method6 for optimal
coverage of descriptor space for physicochemical
parameters representing hydrophobic contributions
(px), electronic characteristics (r, Hammett constant),
and Molar Refractivity (MR). We tested the data
according to classical statistics’ notions (average, vari-
ance, standard deviation, skewness, kurtosis, linear pair
correlation coefficient) in order to ensure that the values
are well dispersed, follow a normal distribution, and
that the chosen descriptors series are totally independent
of each other so that consequently the final results will
be statistically sound, and fairly representative. The re-
sults of the tests (data not shown) confirmed that the
compounds selected to build the correlation equation
are statistically sound, and therefore, proceeding in this
study is feasible. In Scheme 1, Craig plots illustrate the
dispersion in the dual parameter space of hydrophobic-
ity and Hammett sigma values, Molar Refractivity and
Hammett sigma values, and Molar Refractivity and
hydrophobicity, respectively. Determination of the r
receptor affinities for the series allowed fitting of robust
models for r1 and r2 binding, and complemented previ-
ous qualitative studies of related compounds7a,b and
quantitative8a,b studies of the r1 receptor binding of sev-
eral different structural classes. We also evaluated three
representative compounds (3a, d, h) for binding against
[3H](+)-pentazocine in the presence of the positive allo-
steric modulator phenytoin to determine if the series was
likely to exhibit r1 receptor antagonism.9a,b
Ki (nM)
r1
Ki (nM)
r2
Selectivity
Kir2/Kir1
3a (lead, R = H)
3b (R = 2-Br)
3c (R = 2-NO2)
3d (R = 3-I)
0.66 0.06
0.60 0.01
2.80 0.04
0.39 0.05
1.36 0.08
0.87 0.01
0.93 0.02
0.76 0.07
0.37 0.01
1.17 0.01
0.83 0.03
4.34 0.31
34.33 2.69
4.12 0.19
3.79 0.23
1.03 0.08
13.85 0.98
14.19 0.26
1.59 0.08
32.81 2.93
3.29 0.34
17.51 2.51
9.58 0.98
89.51 7.97
52.0
6.9
1.4
2.6
3e (R = 3-F)
10.2
16.3
1.7
3f (R = 3-OCH3)
3g (R = 3-NO2)
3h (R = 4-OCH3)
3i (R = 4-NO2)
3j (R = 4-CH3)
Haloperidol
43.2
8.9
15.0
11.5
20.6
SA4503
a Mean SEM; n = 3–6.
and 3h. Haloperidol and SA4503 (N-phenylpropyl-N0-
3,4-dimethoxyphenethyl piperazine) were included for
reference in these assays, and gave Ki values near those
reported.10 Previous studies4 showed excellent binding
selectivity for compound 3a to r receptors (20 nM) over
serotonin 5-HT1A and dopamine D2 receptors (IC50 val-
ues >100 lM). Affinities for additional recognition sites
(e.g., monoamine transporters) do not appear to have
been studied. However, some structurally related com-
pounds, such as SA4503,11 do exhibit high and preferen-
tial binding to r receptors over a number of other
receptors and ion channels. In the present work, congen-
ers 3a–3j were tested in binding assays for the classical
opioid receptors (l, d, and j) and uniformly exhibited
poor affinities (<5% displacement of radioligand binding
at 1–2 lM, data not shown).
2. Results and discussion
2.1. Receptor binding and qualitative SAR
The r1 receptor binding proved sensitive to the position
of nitro substitution, with an 8-fold increase in affinity
observed over the ortho-, meta-, and para-isomers
(3i > 3g > 3c). By contrast, notable changes in r2 affinity
were not observed. Having electron-donating methoxy
and methyl groups in the para-position decreased both
r1 and r2 receptor affinity with respect to the electron-
withdrawing para-nitro substituent, with a greater detri-
mental effect on r2 binding than r1 binding. The halogen
series showed a qualitative relationship, regardless of
position of substitution, for higher r1 and r2 receptor
affinity with increasing size, hydrophobicity, and
polarizability (3d > 3b > 3e).
All of the N-phenylpropyl-N0-benzylpiperazines dis-
played high affinity for r1 sites (Table 1). Seven of the
10 compounds exhibited subnanomolar apparent affini-
ties (Ki values), with 3-iodo (3d) and 4-nitro (3i) substi-
tution exerting particularly strong effects. A wider range
of affinities was noted for r2 receptor binding, with the
most potent compound 3d (Ki 1.0 nM) having a 3-iodo
substituent, and the least potent compounds being the
parent 3a and the 4-methoxy analog 3h (Ki 33–34 nM).
Thus, r1/r2 selectivity is dominated by the relative r2
affinity, with 40- to 50-fold r1 selectivities noted for 3a

R. I. Nahas et al. / Bioorg. Med. Chem. 16 (2008) 755–761
757
r1 receptor antagonists that might have the functional
ability to block some actions of (ꢀ)-cocaine in vivo.
2.3. Quantitative SAR
We employed the Multiple Linear Regression (MLR)
approach, using the ordinary least squares method, that
is especially useful for analyzing data when the number
of physicochemical descriptors is smaller than the num-
ber of compounds.13 While a multitude of regression
equations were judged significant, the equations shown
below represent the highest correlation coefficients (r2)
and F statistics. The number of compounds (n) used to
establish the equation is listed, and values in parentheses
are the 95% confidence interval of the regression coeffi-
cients. The t parameter gave decisive information
regarding the importance of a single independent vari-
able in a model involving all other independent variables
(Table 2). The standard error of the multiple linear
regression, s, was also taken in consideration. The sym-
bol q denotes the predictive effectiveness of the model
and is the square root of the cross-validated coefficient
of determination, q2.
Figure 2. Phenytoin causes a 15-fold shift (arrow) to higher affinity in
GPB membranes for the r1 agonist dextromethorphan, but only a
small shift, to lower affinity, for the r1 antagonist rimcazole.
N-Phenylpropyl-N0-benzylpiperazine (3a) behaves like an antagonist
in this assay. Points are means SEM for 3–5 trials, each performed in
duplicate.
2.2. Agonist/antagonist profiling
Cobos et al.9a,b reported competition binding assays in
guinea pig brain (GPB) homogenates that use [3H]
(+)-pentazocine in the presence of the allosteric modula-
tor phenytoin to differentiate r1 receptor agonists from
antagonists. In brief, known agonists exhibit major
shifts to higher affinity, while known antagonists show
either little change or modest shifts to lower affinity.
As shown in Figure 2, binding parameter changes for
the lead compound (3a, R = H) under these conditions
were consistent with designation as an antagonist. The
IC50s were not significantly different (p > 0.05) in the
presence and absence of phenytoin, but the ratio
Eq. 1 shows that r1 binding affinity is dependent on the
hydrophobicity (px), Molar Refractivity (MR), and the
sigma Hammett constant values for the corresponding
meta and para-substituted compounds (rm,p).
2
Logð1=KiÞ ¼ ꢀ0:63ðꢁ0:47Þ ꢀ 1:54ðꢁ1:45ÞðpxÞ
þ 1:26ðꢁ1:01ÞðpxÞ
(1.47 0.29 nM/1.89 0.08 nM;
vehicle/phenytoin)
þ 0:96ðꢁ0:79ÞMR
was less than unity (0.8) as previously found for a series
of r1 receptor antagonists including haloperidol,
NE100, BD1063, BD1047, and progesterone.9a,b Similar
results were obtained for the 3-iodo (3d) and 4-methoxy
(3h) analogs (data not shown). In validation studies
(Fig. 2), we find the previously reported9a,b 15-fold shift
to higher affinity for the r1 agonist dextromethorphan;
viz., IC50 232 8.8 nM (Hill slope 0.95 0.01) in the ab-
sence of phenytoin and 15.2 0.5 nM (Hill slope
1.00 0.04) in the presence of phenytoin. We also tested
rimcazole, a low affinity r1 receptor antagonist and
dopamine transporter inhibitor that can block (ꢀ)-co-
caine effects in vivo.12 Binding parameter changes in
the presence (IC50 1087 241 nM; Hill slope 0.89
0.05) or absence (IC50 960 77 nM; Hill slope 0.90
0.05) of phenytoin were not significant (p > 0.05), and
the IC50 ratio was less than unity (0.9) as expected.
Although agonist and antagonist properties are best
determined in functional assays, this series of N-phenyl-
propyl-N0-benzylpiperazines profiles as a set of probable
ð1Þ
þ 0:62ðꢁ0:37Þrm;p
;
n ¼ 9; r2 ¼ 0:89; F 4;8 ¼ 7:93; s ¼ 0:13;
q2 ¼ 0:43; 0:01 < P < 0:05
The inclusion of compound 3a (R = H) gave an equa-
tion having less statistical significance. Interestingly,
the negative dependence of p and binding affinity is
similar to the relationship described by Mascarella and
co-workers for the r1 binding of a series of normetazo-
cines.8a Utilizing the Molar Refractivity term gave a
better correlation than the Taft Steric Effect (Es) (data
not shown). This implies that increasing the substituent
size enhances the binding potency, and that the polariz-
ability component of Molar Refractivity also plays a
role in binding. The dependence on sigma Hammett val-
ues (rm,p) suggests that electron withdrawing groups in
the meta- and para-positions improve r1 receptor
affinity. The use of rp instead of rꢀ for compound 3i
Table 2. Values for t parameters
Variable
Eq. 1
Eq. 2
Eq. 3
Number (significance)
t
Number (significance)
t
Number (significance)
t
Constant
p
ꢀ0.630 (0.020)
1.255 (0.026)
ꢀ1.535 (0.042)
0.965 (0.027)
0.617 (0.011)
ꢀ3.728
3.453
ꢀ1.495 (0.000)
ꢀ0.968 (0.037)
1.502 (0.019)
ꢀ0.405 (0.084)
0.505 (056)
ꢀ11.227
ꢀ2.830
3.391
ꢀ1.495 (0.114)
ꢀ1.001 (0.279)
1.493 (0.374)
ꢀ0.370 (0.162)
0.548 (0.173)
ꢀ13.071
ꢀ3.586
3.990
p2
MR/Es
ꢀ2.937
3.399
4.486
ꢀ2.155
2.477
ꢀ2.281
3.174
r

758
R. I. Nahas et al. / Bioorg. Med. Chem. 16 (2008) 755–761
(R = 4-NO2) decreases the correlation coefficient and
the statistical significance of the model. This suggests
that a nitro group in the para-position increases the
binding affinity dramatically due to geometric require-
ments, or favorable charge interactions with an un-
known binding region of the protein. These results
agree with those from Ruoho’s group on the r1 receptor
binding of (ꢀ)-cocaine derivatives, where reduced p elec-
tron density caused by a para-nitro substituent on the
phenyl ring promotes high-affinity binding through
favorable interactions with electron-rich residues of the
steroid binding domain-like regions.14 Further, differen-
tial binding data across the present nitro series indicate
that the precise orientation of the dipole moment is an
important contributor to the r1 receptor–ligand binding
interaction. In general, the statistical parameters of Eq.
1 may have been more robust had the Ki values been dis-
persed over a broader range.
Observed Ki values for all ten compounds, ranging from
1.0 to 34.3 nM, yielded a quality correlation equation
for r2 binding affinity where substituent hydrophobicity
proved a major contributor (Eq. 2).
2
Logð1=KiÞ ¼ ꢀ1:50ðꢁ0:34Þ þ 1:50ðꢁ1:14ÞðpxÞ
Figure 3. Actual pKi versus calculated pKi for r1 receptor binding
affinity using Eq. 1 (upper panel). Actual pKi versus calculated pKi for
r2 receptor binding affinity using Eq. 3 (lower panel).
ꢀ 0:96ðꢁ0:88ÞðpxÞ
ꢀ 0:41ðꢁ0:48ÞEs
ð2Þ
þ 0:50ðꢁ0:53Þrm;p
;
between the actual pKi values obtained and the values
derived from Eqs. 1 and 3 for the r1 and r2 receptor sub-
types, respectively (Fig. 3).
n ¼ 10; r2 ¼ 0:95; F 4;9 ¼ 22:22; s ¼ 0:17;
q2 ¼ 0:79; P < 0:01
Introducing Hammett sigma values for the two ortho-
substituted compounds gave a poorer performance in
the equation derived for r1 binding, but a better perfor-
mance for r2 binding as shown below Eq. 3.
3. Conclusion
A foundation has been established to correlate structure
with r1 and r2 receptor binding affinity for a series of
N-(3-phenylpropyl)-N0-benzylpiperazines. The results
revealed a number of potent ligands, in particular for
the r1 subtype. The predictive ability of these equations
is of good statistical quality and showed interesting re-
sults regarding the quantified dependence of the binding
affinity upon the studied physicochemical parameters.
The hydrophobicity terms in the equations exhibited
opposite signs for r1 and r2, which can be relied upon
to design selective ligands, and a better understanding
of the binding pharmacophore difference between the
two subtypes.
2
Logð1=KiÞ ¼ ꢀ1:45ðꢁ0:30Þ þ 1:48ðꢁ0:97ÞðpxÞ
ꢀ 0:99ðꢁ0:72ÞðpxÞ
ꢀ 0:38ðꢁ0:42ÞEs
ð3Þ
þ 0:55ðꢁ0:45Þro;m;p
n ¼ 10; r2 ¼ 0:96; F 4;9 ¼ 29:95;
s ¼ 0:15; q2 ¼ 0:86; P < 0:001
;
The positive parabolic form of the hydrophobicity term
of Eqs. 2 and 3 as opposed to a negative sign in the case
of Eq. 1 constitutes a major binding pharmacophore dif-
ference between the two r receptors. The second appar-
ent difference stems from utilizing Es in Eqs. 2 and 3
instead of MR, suggesting the polarizability of the sub-
stituent has a pronounced effect on r2, but not r1, recep-
tor binding. Further, both equations describe a binding
affinity increase for electron-withdrawing groups in the
meta- and para-positions. However, including ro values
for the ortho substituents improves the quality of the
correlation equation for r2 binding. This third differen-
tiation indicates that ortho substituents interact favor-
ably with the r2 receptor, leading to moderately
improved ligand binding. A good correlation is observed
4. Experimental
4.1. General procedures
1H NMR spectra were determined on Bruker 250 or
300 MHz spectrometers. Chemical shifts are reported
as parts per million (d) relative to internal Me4Si in
CDCl3, with coupling constants (J) given in hertz
(Hz). Elemental analyses were performed by Atlantic
Microlab, Inc. (Norcross, GA), and were in agreement
with calculated values (C, H, N: 0.4%). Short-path sil-
ica gel (Merck 7729, <230 mesh) chromatography was

R. I. Nahas et al. / Bioorg. Med. Chem. 16 (2008) 755–761
759
conducted under N2 pressure. Analytical TLC was per-
formed with Macherey-Nagel silica gel 60 UV-254 plates
(250 lm). N-Phenylpropylpiperazine was prepared as
previously described.15 OptiPhaseꢁ HiSafe 2 scintilla-
tion cocktail, [3H](+)-pentazocine ([3H](+)-PTZ, 34 Ci/
mmol), [3H]1,3-di(2-tolyl)guanidine ([3H]DTG, 58 Ci/
mmol), and [3H]naltrindole ([3H]NTI, 20 Ci/mmol) were
from Perkin-Elmer Life Sciences, Inc. (Boston, MA),
while [3H]DAMGO (50 Ci/mmol) and [3H]U69,593
(50 Ci/mmol) were from GE Healthcare Bio-Sciences
Corp. (Piscataway, NJ). Other chemicals and solvents
were of the best grades available, and were used as re-
ceived from commercial sources.
Ph), 2.52 (4H, m, Pip), 2.6 (4H, m, Pip), 2.67 (2H, t,
J = 7.5 Hz, –N–CH2–CH2–), 3.64 (2H, s, Ph–CH2–N–),
7.09–7.55 (8H, m, Ph), 7.57 (1H, d, ortho to Br,
J = 9 Hz). Yield: 70%. Anal (C20H25N2Br). Theory: C,
64.34; H, 6.75; N, 7.50. Found: C, 64.05; H, 6.77; N,
7.49. HPLC: tR = 9.3, k0 = 4.8.
5.3. 1-(2-Nitrobenzyl)-4-(3-phenylpropyl)piperazine (3c)
1H NMR d 1.75 (2H, apparent pentet, J = 7.5 Hz,
–CH2–CH2–CH2–), 2.35 (2H, t, J = 7.5 Hz, –CH2
–CH2–Ph), 2.47 (8H, m, Pip), 2.65 (2H, t, J = 7.5 Hz,
–N–CH2–CH2–), 3.8 (2H, s, Ph–CH2–N–), 7.15–7.79
(8H, m, Ph), 7.82 (1H, d, ortho to NO2, J = 12.5 Hz).
Yield: 83%. Anal (C20H25N3O2). Theory: C, 70.77; H,
7.42; N, 12.38. Found: C, 70.50; H, 7.48; N, 12.26.
HPLC: tR = 6.1, k0 = 2.6.
4.2. General method for the preparation of compounds
3a–3j
The appropriate benzyl bromide (1.0 mmol), N-phenyl-
propylpiperazine (1.0 mmol), NaI (150 mg, 1.0 mmol),
and K2CO3 (0.41 g, 3.0 mmol) were heated in DMF
(10 mL) for 2 h at 60 ꢂC. The mixture was filtered, and
then concentrated under reduced pressure at 60 ꢂC. The
residue was partitioned between water and EtOAc, and
the organic layer washed with brine, dried (MgSO4),
and concentrated under reduced pressure. Column chro-
matography using a gradient of n-hexane/EtOAc (5:1–
1:4) gave the target compounds in 70–88% yields as color-
less to pale yellow oils that were stored in free base form.
Analytical TLC Rf values ranged from 0.2 to 0.4 (1:1,
n-hexane/EtOAc). 1H NMR and elemental analysis data
agreed with the assigned structures. Analytical reversed-
5.4. 1-(3-Iodobenzyl)-4-(3-phenylpropyl)piperazine (3d)
1H NMR d 1.85 (2H, apparent pentet, J = 7.6 Hz,
–CH2–CH2–CH2–), 2.4 (2H, t, J = 7.6 Hz, –CH2–CH2
–Ph), 2.49 (8H, m, Pip), 2.65 (2H, t, J = 7.6 Hz, –N
–CH2–CH2–), 3.46 (2H, s, Ph–CH2–N–), 7.02–7.31
(7H, m, Ph), 7.60 (1H, d, meta to I, J = 7.8 Hz), 7.71
(1H, s, ortho to I and CH2–). Yield: 76%. Anal
(C20H25N2I). Theory: C, 57.15; H, 5.99; N, 6.66. Found:
C, 57.03; H, 5.95; N, 6.61. HPLC: tR = 13.1, k0 = 6.7.
5.5. 1-(3-Fluorobenzyl)-4-(3-phenylpropyl)piperazine (3e)
phase HPLC, using
a
Symmetry C18 column
1H NMR CDCl3
d 1.86 (2H, apparent pentet,
(4.6 · 150 mm, 5 lm; Waters Corp., Milford, MA) and
a ternary mobile phase of MeOH (25%), CH3CN (25%),
and water (50%) containing Et3N (1.5%) and HOAc
(2%) at a flow rate of 1 mL/min with detection at
254 nm, showed P98% purity for each compound
and provided retention times (tR) and capacity factors
(k0).
J = 7.5 Hz, –CH2–CH2–CH2–), 2.42 (2H, t, J = 7.5
Hz, –CH2–CH2–Ph), 2.52 (8H, m, Pip), 2.69 (2H, t,
J = 7.5 Hz, –N–CH2–CH2–), 3.59 (2H, s, Ph–CH2–N–),
6.95–7.36 (9H, m, Ph). Yield: 88%. Anal
(C20H25N2F). Theory: C, 76.84; H, 8.07; N, 8.97.
Found: C, 76.91; H, 8.08; N, 9.10. HPLC: tR = 5.9,
k0 = 2.5.
5.6. 1-(3-Methoxybenzyl)-4-(3-phenylpropyl)piperazine
(3f)
5. Chemistry
A common precursor, N-phenylpropylpiperazine, was
prepared according to the literature procedure.15 Alkyl-
ation by the corresponding benzyl bromide, in the pres-
ence of potassium carbonate and KI, provided 3a–3j as
oils in 70–88% yields.
1H NMR d 1.85 (2H, apparent pentet, J = 7.6 Hz,
–CH2–CH2–CH2–), 2.4 (2H, t, J = 7.6 Hz, –CH2–CH2
–Ph), 2.51 (8H, m, Pip), 2.63 (2H, t, J = 7.6 Hz, –N
–CH2–CH2–), 3.51 (2H, s, Ph–CH2–N–), 3.83 (3H, s,
Ph–OCH3), 6.78–7.34 (9H, m, Ph). Yield: 84%. Anal
(C21H28N2O). Theory: C, 76.67; H, 8.73; N, 8.52.
Found: C, 76.82; H, 8.73; N, 8.53. HPLC: tR = 5.2,
k0 = 2.1.
5.1. 1-Benzyl-4-(3-phenylpropyl)piperazine (3a)
1H NMR d 1.85 (2H, apparent pentet, J = 7.5 Hz,
–CH2–CH2–CH2–), 2.42 (2H, t, J = 7.5 Hz, –CH2–
CH2–Ph), 2.53 (8H, m, Pip), 2.67 (2H, t, J = 7.5 Hz,
–N–CH2–CH2–), 3.56 (2H, s, Ph–CH2–N), 7.19–7.38
(10H, m, Ph). Yield: 78%. Anal (C20H26N2). Theory:
C, 81.59; H, 8.90; N, 9.51. Found: C, 81.22; H, 8.95;
N, 9.59. HPLC: tR = 5.1, k0 = 2.0.
5.7. 1-(3-Nitrobenzyl)-4-(3-phenylpropyl)piperazine (3g)
1H NMR d 1.81 (2H, apparent pentet, J = 7.5 Hz,
–CH2–CH2–CH2–), 2.39 (2H, t, J = 7.5 Hz, –CH2
–CH2–Ph), 2.50 (8H, m, Pip), 2.64 (2H, t, J = 7.5 Hz,
–N–CH2–CH2–), 3.6 (2H, s, Ph–CH2–N–), 7.16–7.68
(7H, m, Ph), 8.12 (1H, d, meta to NO2, J = 9.6 Hz),
8.21 (1H, s, ortho to NO2 and CH2–). Yield: 70%. Anal
(C20H25N3O2). Theory: C, 70.77; H, 7.42; N, 12.38.
Found: C, 70.31; H, 7.50; N, 12.12. HPLC: tR = 5.6,
k0 = 2.3.
5.2. 1-(2-Bromobenzyl)-4-(3-phenylpropyl)piperazine (3b)
1H NMR d 1.85 (2H, apparent pentet, J = 7.5 Hz,
–CH2–CH2–CH2–), 2.4 (2H, t, J = 7.5 Hz, –CH2–CH2–

760
R. I. Nahas et al. / Bioorg. Med. Chem. 16 (2008) 755–761
5.8. 1-(4-Methoxybenzyl)-4-(3-phenylpropyl)piperazine
(3h)
7. QSAR analysis
The QSAR of compounds 3a–3j were analyzed by the
Hansch–Fujita method,18 using physicochemical descrip-
tors that represent hydrophobic, electronic, and steric
effects. Physicochemical parameters were taken from
Hansch et al.19 and we employed published r values20a,b
for 2-Br and 2-NO2 in Eq. 3. We utilized Multiple Linear
Regression analysis with SAS statistical software (v. 9.0)
to generate regression equations that were evaluated by
their correlation coefficients (r), F statistics, t parameters,
and standard errors, s.
1H NMR d 1.85 (2H, apparent pentet, J = 7.6 Hz,
–CH2–CH2–CH2–), 2.39 (2H, t, J = 7.6 Hz, –CH2
–CH2–Ph), 2.48 (8H, m, Pip), 2.64 (2H, t, J = 7.6 Hz,
–N–CH2–CH2–), 3.46 (2H, s, Ph–CH2–N–), 3.81 (3H,
s, Ph–OCH3), 6.78–7.31 (9H, m, Ph). Yield: 84%. Anal
(C21H28N2OÆ0.5H2O). Theory: C, 75.68; H, 8.55; N,
8.40. Found: C, 76.06; H, 8.60; N, 8.25. HPLC:
tR = 5.0, k0 = 1.9.
5.9. 1-(4-Nitrobenzyl)-4-(3-phenylpropyl)piperazine (3i)
1H NMR d 1.84 (2H, apparent pentet, J = 7.5 Hz,
–CH2–CH2–CH2–), 2.37 (2H, t, J = 7.5 Hz –CH2
–CH2–Ph), 2.5 (8H, m, Pip), 2.67 (2H, t, J = 7.5 Hz,
–N–CH2–CH2–), 3.6 (2H, s, Ph–CH2–N–), 7.2–7.31
(5H, m, Ph), 7.52 (2H, d, ortho to –CH2, J = 8.7 Hz),
8.18 (2H, d, ortho to NO2, J = 8.7 Hz). Yield: 70%. Anal
(C20H25N3O2). Theory: C, 70.77; H, 7.42; N, 12.38.
Found: C, 70.65; H, 7.51; N, 12.20. HPLC: tR = 6.5,
k0 = 2.8.
Acknowledgments
We thank NSF CHE-95-31247 for NMR instrumenta-
tion support, and Ms. Sarah A. Violand for assistance
with the phenytoin binding assays.
References and notes
1. Matsumoto, R. R.; Liu, Y.; Lerner, M.; Howard, E. W.;
Brackett, D. J. Eur. J. Pharmacol. 2003, 469, 1–12.
2. Foster, A.; Wu, H.; Chen, W.; Williams, W.; Bowen, W.
D.; Matsumoto, R. R. Coop, A 2003, 13, 749–751.
3. Matsumoto, R. R.; Hourcade-Potelleret, F.; Mack, A.;
Pouw, B.; Zhang, Y.; Bowen, W. D. Pharmacol. Biochem.
Behav. 2004, 77, 775–781.
4. Younes, S.; Labssita, Y.; Baziard-Mouysset, G.; Payard,
M.; Rettori, M.; Renard, P.; Pfeiffer, B.; Caignard, D. Eur.
J. Med. Chem. 2000, 35, 107–121.
5.10. 1-(4-Methylbenzyl)-4-(3-phenylpropyl)piperazine
(3j)
1H NMR d 1.85 (2H, apparent pentet, J = 7.5 Hz,
–CH2–CH2–CH2–), 2.35 (2H, s, Ph–CH3), 2.39 (2H, t,
J = 7.5 Hz, –CH2–CH2–Ph), 2.65 (2H, t, J = 7.5 Hz,
N–CH2–CH2–), 3.49 (2H, s, Ph–CH2–N), 7.12–7.29
(9H, m, Ph). Yield: 82%. Anal (C21H28N2). Theory: C,
81.77; H, 9.15; N, 9.08. Found: C, 81.91; H, 9.30; N,
9.19. HPLC: tR = 5.6, k0 = 2.3.
5. Leo, A. L.; Hansch, C. Perpect. Drug Discov. Design 1999,
17, 1–25.
6. Methods and Principles in Medicinal Chemistry 2: Chemo-
metric Methods in Molecular Design; Mannhold, R.,
Krogsgaard-Larsen, P., Timmerman, H., Van de Water-
beemd, H., Eds.; Springer: Germany, 1995 (2).
7. (a) Glennon, R. A. Mini-Rev. Med. Chem. 2005, 5, 927–
940; (b) Glennon, R. A. Brazil J. Pharm. Sci. 2005, 41,
1–12.
8. (a) Mascarella, S. W.; Bai, X.; Sine, B.; Bowen, W. D.;
Carroll, F. I. J. Med. Chem. 1995, 38, 565–569; (b) Gund,
T. M.; Floyd, J.; Jung, D. J. Mol. Graph. Model. 2004, 22,
221–230.
9. (a) Cobos, E. J.; Baeyens, J. M.; Del Pozo, E. Synapse
2005, 55, 192–195; (b) Cobos, E. J.; Lucena, G.; Baeyens,
J. M.; Del Pozo, E. Synapse 2006, 59, 152–161.
10. Lever, J. R.; Gustafson, J. L.; Xu, R.; Allmon, R. L.;
Lever, S. Z. Synapse 2006, 59, 350–358.
11. Matsuno, K.; Nakazawa, M.; Okamoto, K.; Kawashima,
Y.; Mita, S. Eur. J. Pharmacol. 1996, 306, 271–279.
12. Matsumoto, R. R.; Hewett, K. L.; Pouw, B.; Bowen, W.
D.; Husbands, S. M.; Cao, J. J.; Hauck Newman, A.
Neuropharmacology 2001, 41, 878–886.
6. Receptor binding assays
Competition binding assays for ligands at r1 and r2
receptors were performed using [3H](+)-PTZ (r1),
[3H]DTG/ 200 nM (+)-PTZ (r2), and membranes from
fresh-frozen, male English Hartley guinea pig brains
(Rockland Immunochemicals, Inc.; Gilbertsville, PA)
as previously described.10 Experiments were performed
in duplicate and repeated 3–6 times. Ki values were cal-
culated from the inhibition data using the Cheng-Prus-
off equation,16 and a r1 Kd of 2.3 nM for [3H](+)-PTZ
and a r2 Kd of 23.9 nM for [3H]DTG.10 Phenytoin
modulation of ligand binding to r1 receptors was inves-
tigated using minor modifications of reported proce-
dures.9 Assays were conducted as described above for
[3H](+)-PTZ, except DPH (50 lL, 20 mm) in NaOH
vehicle (0.15 M) was added to every tube. Control
experiments, where only NaOH (50 lL, 0.15 M) was
added, also were conducted. The incubation medium
for these assays had pH 7.44 at 37 ꢂC, while the med-
ium had pH 7.06 at 37 ꢂC for [3H](+)-PTZ assays done
in the absence of NaOH or DPH/NaOH. Opioid recep-
tor binding assays were conducted in membranes from
guinea pig (l,j) or mouse (d) brains using [3H]NTI (d),
[3H]DAMGO (l), and [3H]U69,593 (j) as previously
reported.17
13. Hadjipavlou-Litina, D.; Garg, R.; Hansch, C. Chem. Rev.
2004, 104, 3751–3793.
14. Chen, Y.; Hajipour, A. R.; Sievert, M. K.; Arbabian, M.;
Ruoho, A. E. Biochemistry 2007, 46, 3532–3542.
15. Kawamura, K.; Elsinga, P. H.; Kobayashi, T.; Ishii, S.-I.;
Wang, W.-F.; Matsuno, K.; Vaalburg, W.; Ishiwata, K.
Nucl. Med. Biol. 2003, 30, 273–284.
16. Cheng, Y. C.; Prusoff, W. H. Biochem. Pharmacol. 1973,
22, 3099–3108.
17. Duval, R. A.; Allmon, R. L.; Lever, J. R. J. Med. Chem.
2007, 50, 2144–2156.

R. I. Nahas et al. / Bioorg. Med. Chem. 16 (2008) 755–761
761
18. Hansch, C.; Fujita, T. J. Am. Chem. Soc. 1964, 86, 1616–
1626.
ogy; American Chemical Society: Washington, DC,
1995.
19. Hansch, C.; Leo, A.; Hoekman, D. Hydrophobic,
Electronic, and Steric Constants in Exploring QSAR.
Fundamentals and Applications in Chemistry and Biol-
20. (a) Smith, G. G.; Lum, K. K. Chem. Commun. (London)
1968, 20, 1208; (b) Pressman, D.; Brown, D. H. J. Am.
Chem. Soc. 1943, 65, 540–543.