Mapping the peripheral benzodiazepine receptor binding site by conformationally restrained derivatives of 1-(2-chlorophenyl)-N-methyl-N-(1- methylpropyl)-3-isoquinolinecarboxamide (PK11195)

Article abstract of DOI:10.1021/jm960516m

A synthetic-computational approach to the study of the binding site of peripheral benzodiazepine receptor (PBR) ligands related to 1-(2- chlorophenyl)-N-methyl-N-(1-methylpropyl)-3-isoquinolinecarboxamide (PK11195, 1) within their receptor has been developed. A wide series of conformationally restrained derivatives of 1 has been designed with the aim of probing the PBR binding site systematically. The synthesis of these compounds involves palladium-catalyzed coupling and amidation as the key steps. Twenty-nine rigid and semirigid derivatives of 1 were tested in binding studies using [³H]-1, and most of these showed PBR affinities in the nanomolar range. The essential role of the carbonyl moiety as a primary pharmacophoric element in the recognition by and the binding to PBR has been confirmed, and the restricted range of the carbonyl orientations, which characterizes the moss potent ligands, points to a specific hydrogen-bonding interaction, mainly directed by the geometrical factors, when the electronic ones are fulfilled. Moreover, the fundamental importance of the short-range dispersive interactions in the modulation of the binding affinity and, hence, in the stabilization of the ligand-receptor complex, emerged from the QSAR models reported.

Full text of DOI:10.1021/jm960516m

2910
J . Med. Chem. 1997, 40, 2910-2921
Ma p p in g th e P er ip h er a l Ben zod ia zep in e Recep tor Bin d in g Site by
Con for m a tion a lly Restr a in ed Der iva tives of
1-(2-Ch lor op h en yl)-N-m eth yl-N-(1-m eth ylp r op yl)-3-isoqu in olin eca r boxa m id e
(P K11195)
Andrea Cappelli,*,† Maurizio Anzini,† Salvatore Vomero,† Pier G. De Benedetti,‡ Maria Cristina Menziani,‡
Gianluca Giorgi,§ and Cristina Manzoni|
Dipartimento Farmaco Chimico Tecnologico, Universita` di Siena, Via Banchi di Sotto 55, 53100 Siena, Italy,
Dipartimento di Chimica, Universita` degli Studi di Modena, Via Campi 183, 41100 Modena, Italy,
Centro Interdipartimentale di Analisi e Determinazioni Strutturali, Universita` di Siena, via P.A. Mattioli 10,
53100 Siena, Italy, and Istituto di Ricerche Farmacologiche “Mario Negri”, Via Eritrea, 62, 20157 Milano, Italy
Received J uly 16, 1996X
A synthetic-computational approach to the study of the binding site of peripheral benzodi-
azepine receptor (PBR) ligands related to 1-(2-chlorophenyl)-N-methyl-N-(1-methylpropyl)-3-
isoquinolinecarboxamide (PK11195, 1) within their receptor has been developed. A wide series
of conformationally restrained derivatives of 1 has been designed with the aim of probing the
PBR binding site systematically. The synthesis of these compounds involves palladium-
catalyzed coupling and amidation as the key steps. Twenty-nine rigid and semirigid derivatives
of 1 were tested in binding studies using [3H]-1, and most of these showed PBR affinities in
the nanomolar range. The essential role of the carbonyl moiety as a primary pharmacophoric
element in the recognition by and the binding to PBR has been confirmed, and the restricted
range of the carbonyl orientations, which characterizes the most potent ligands, points to a
specific hydrogen-bonding interaction, mainly directed by the geometrical factors, when the
electronic ones are fulfilled. Moreover, the fundamental importance of the short-range
dispersive interactions in the modulation of the binding affinity and, hence, in the stabilization
of the ligand-receptor complex, emerged from the QSAR models reported.
Ch a r t 1
In tr od u ction
The discovery of benzodiazepine binding sites in the
periphery (peripheral benzodiazepine receptor, PBR)
stimulated intensive research devoted to the character-
ization of this receptor.1 Despite these efforts, the
physiological role of PBR is still unclear. However, a
wide range of pharmacological activities, such as anti-
convulsant, anxiolytic, immunomodulating, and cardio-
vascular, has been related to the activation of this
receptor.2 PK11195 (1) is the first non-benzodiazepine
ligand which was found to bind the peripheral benzo-
diazepine receptors (PBR) with nanomolar affinity, and
it is now the most commonly used radioligand for this
receptor. The structure-affinity relationship data of
the analogs of 1 were reported in the patent by Du-
broeucq and co-workers and were later discussed by
Bourguignon.3 Georges and co-workers studied a lim-
ited set of 1-like compounds by means of molecular
orbital calculations in the attempt to define and map
the binding site of 1 in the PBR.4 However, a compre-
hensive and systematic investigation of binding site of
1 within PBR is still missing.
The primary structure of PBR is now available,5 and
the information concerning its three-dimensional struc-
ture is likely to be sufficient to undertake the determi-
nation of the active site of PBR by direct methods in
the near future.6 However, at the present time, this is
a very difficult task, and the development of potent
ligands must rely on receptor mapping studies by
indirect methods. In mapping studies one of the most
striking problem is the determination of the ligand
conformation bound to the receptor (bioactive conforma-
tion); that is why rigid, highly active ligands are the
best possible templates.7
† Dipartimento Farmaco Chimico Tecnologico, Universita` di Siena.
‡ Dipartimento di Chimica, Universita` degli Studi di Modena.
§ Centro Interdipartimentale di Analisi e Determinazioni Strut-
turali, Universita` di Siena.
| Istituto di Ricerche Farmacologiche “Mario Negri”.
X Abstract published in Advance ACS Abstracts, August 1, 1997.
S0022-2623(96)00516-X CCC: $14.00 © 1997 American Chemical Society

Mapping the Peripheral Benzodiazepine Receptor Binding Site
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 18 2911
Ta ble 1. Preparative and Analytical Data of Compounds 3-7
compd
bridge (B)
CH2
CH2
CH2
X
R1
s-Bu
s-Bu
Bn
Bn
Me
s-Bu
Bn
s-Bu
s-Bu
Bn
4-ClBn
4-ClBn
s-Bu
s-Bu
Bn
4-ClBn
4-ClBn
4-ClPh
4-MeOPh
s-Bu
Bn
R2
prep methoda
yield, %
81
mp, °C (recryst solv)b
formula
anal.c
CHN
3a
3b
3c
3d
3e
3f
3g
4a
4b
4c
4d
4e
4f
4g
4h
4i
4j
4k
4l
5a
5b
5c
5d
5e
6a
6b
6c
7a
7b
H
F
ref 10
1
197-198 (C-EA)
C21H19N2OF
H
H
H
H
H
H
F
H
H
F
H
F
H
H
F
ref 10
ref 10
ref 11
ref 11
ref 11
2 A
2 A
2 A
2 A
2 A
2 B
2 B
2 C
2 B
2 B
2 C
2 C
3
n-BuCH
CH2CH2CH2
CH2CH2CH2
CH2CH2CH2
H
H
H
H
66
75
81
76
87
64
87
84
66
81
72
64
90
95
64
65
98
86
84
86
59
82
157-158 (H-C)
161-162 (H-C)
127-128 (H-C)
123-124 (H-C)
131-133 (C-H)
117-118 (H)
oil
118-120
132-133 (H-C)
107-108
165-166 (C-EA)
133-134 (H-EA)
193-194 (EA-C)
243-244 (EA-CH)
153-154 (EA-H)
193-194 (EA)
179-180 (EA)
120-121 (H)
136-137
C
C
C
C
C
C
C
C
C
C
C
C
21H22N2O‚1/8H2O
21H21N2OF
CHN
CHN
CHN
CHN
CHN
CHN
CHN
CHN
CHN
CHN
CHN
CHN
CHN
CHN
CHN
CHN
CHN
CHN
CHN
CHN
CHN
CHN
24H20N2O‚1/8H2O
24H19N2OCl
H
24H18N2OClF
22H24N2O‚1/4H2O
22H23N2OF
Me
Me
Me
Me
Me
Me
Me
25H22N2O‚1/4H2O
25H21N2OCl
25H20N2OClF
24H19N2OCl
25H22N2O2
C21H20N2O
C24H18N2O
C23H24N2O
C23H22N2O
C26H20N2O
H
H
CH2
CH2
CH2CH2CH2
CHdCHCH2
CHdCHCH2
3
s-Bu
s-Bu
Bn
s-Bu
s-Bu
Bn
4 D
4 D
4 E
5 A
5 B
5 C
6 F
d
d
H
Me
Me
C
C
C
21H22N2O
22H24N2O
25H22N2O
oil
oil
154-155
CH2CH2
OCH2CH2
C26H22N2O‚1/4H2O
e
6 F
C26H22N2O2
a
In the preparation method the number indicates the scheme and the letter the method for the newly synthesized compounds, while
b
references are reported for the known compounds. Recrystallization solvents; EA, ethyl acetate; H, n-hexane; C, cyclohexane; CH,
chloroform. c Analyses for the elements indicated were within (0.4% of the theoretical values. See Scheme 4. e See Scheme 6.
d
Sch em e 1a
In the large amount of PBR ligands developed up to
now the most rigid high-affinity ligand is Ro5-4864 (2),
while 1 shows at least one important degree of confor-
mational freedom.4 These two ligands are supposed to
bind nonidentical binding domains, as they have been
reported to have different binding modalities.6a
In this paper we present a new systematic approach
to the study of the binding site of 1-like compounds
within PBR, based on the design of conformationally
constrained derivatives 3-7.
Moreover, electronic indexes8 as well as ad-hoc-
defined size and shape molecular descriptors9 have been
considered and employed in order to decipher, on the
basis of the complementarity principle, the binding site
features on quantitative grounds.
Ch em istr y
The synthesis of pyrrolo[3,4-b]quinolin-3-ones 3a ,c,d
has recently been published by some of us,10 and the
related compound 3b (Table 1) was prepared by using
a
the reported methodology starting from 2′-fluoroben-
zophenone 8 (Scheme 1). Friedla¨nder condensation
(base-catalyzed) of 8 with 2-oxobutyric acid gave 9,
Reagents: (i) EtCOCOOH, MeONa, MeOH; (ii) POCl3, EtOH;
(iii) NBS, (PhCOO)2, CCl4; (iv) s-BuNH2, K2CO3, EtOH.
1
which was characterized by H-NMR spectroscopy and
peroxide gave γ-bromo ester 11, which was converted
into γ-lactam 3b by using sec-butylamine in ethanol at
reflux in the presence of K2CO3.
transformed into the corresponding ethyl ester 10.
Bromination of 10 with NBS in the presence of dibenzoyl

2912 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 18
Cappelli et al.
Sch em e 2
Sch em e 3a
The preparation of azepino[3,4-b]quinolin-5-ones 3e,f,g
is the object of a recent publication by some of us.11
Secondary amides 4a -e (Scheme 2) were prepared
by reaction of the acid chloride generated from the
suitable carboxylic acids 9 or 1210 with an appropriate
primary amine (method A) and then converted into the
corresponding tertiary amides by alkylating the amide
anion (generated with NaH) using MeI (method B).
Tertiary amide 4h was prepared by reaction of the
chloride of acid 12 with N-methylbenzylamine (method
C).
a
Reagents: (i) Tf2O, Py; (ii) Me4Sn, LiCl, Pd(PPh3)2Cl2, DMF;
(iii) NBS, (PhCOO)2, CCl4; (iv) R1NH2, EtOH.
Sch em e 4a
The synthesis of all the compounds containing the
isoquinoline core was rationalized through the search
for a common intermediate, and aryl triflate 14 was
chosen as the starting material. This compound was
prepared from the corresponding enolate 1312 by reac-
tion with triflic anhydride in pyridine. Aryl triflate 14
underwent Stille coupling13 with tetramethyltin to give
the corresponding methyl derivative 15 in very good
yield. Compound 15 was brominated in its benzyl
position with NBS in CCl4 in the presence of a radical
initiator (dibenzoyl peroxide), and it was then reacted
with the suitable primary amine to give pyrroloiso-
quinolines 5a ,b (Scheme 3).
The synthesis of the compounds containing the new
azepino[3,4-c]isoquinoline ring system is shown in
Scheme 4. Aryl triflate 14 underwent palladium-
catalyzed coupling with propargyl alcohol14 in the pres-
ence of CuI and Et3N to give compound 17 in good yield.
Catalytic reduction of 17 in the presence of Pd/C gave
saturated alcohol 18. This was chlorinated with SOCl2,
allowed to react with sec-butylamine, and then cyclized
with the procedure previously described for the synthe-
sis of compound 3f11 (method D). Alternatively, com-
pound 17 was hydrogenated in the presence of PdS on
carbon to give allylic alcohol 19 the Z geometry of which
at the double bond was confirmed by 1H-NMR spectros-
copy (J ) 11.3 Hz for the vinylic hydrogens in DMSO).
Alcohol 19 was transformed into 5d by using method
D, or reacted first with SOCl2, and then with benzy-
lamine, and finally, cyclized in toluene in the presence
of 2-hydroxypyridine to give azepinoisoquinoline 5e
(method E).
a
Reagents: (i) HCtCCH2OH, CuI, Pd(PPH3)2Cl2, Et3N, DMF;
(ii) H2, Pd/C, EtOH; (iii) H2, PdS/C, EtOH.
Finally, the compounds with the locked phenyl group
7a ,b were synthesized by the palladium-catalyzed ami-
dation15 of the corresponding imidoyl chlorides 21a or
21b (Scheme 6).16
Hydrolysis of ester 15 gave carboxylic acid 20, which
was promptly converted into secondary amide 6a by
using method A or into tertiary amide 6c by using
method C (Scheme 5). The alkylation of secondary
amide 6a (method B) gave tertiary amide 6b.
Resu lts a n d Discu ssion
Deter m in a tion of th e Bioa ctive Con for m a tion .
The importance of the carbonyl group in the interaction

Mapping the Peripheral Benzodiazepine Receptor Binding Site
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 18 2913
Sch em e 5a
Sch em e 6
reduced to 60° < φ < 110° for compound 4g, and to 80°<
φ < 130° for compound 6b. A good agreement is
observed between the range of allowed conformations
(∆E e 3 kcal/mol) obtained for these compounds by
means of the AM1 semiempirical method and the ab
initio results at the 6-31G* level.20 The results are
summarized in Table 3.
If the same binding modalities are assumed for
compounds 4f-l, 6b,c, and 7a ,b, these should bind PBR
with the same carbonyl orientations. Following this
idea, the high affinity of compounds 4f-l, 6b-c, and 7a ,b
suggests that the conformational space shared by the
carbonyl moieties of compounds 4f-l, 6b,c, and 7a ,b
(corresponding to the carbonyl orientations comprised
in the range 80° < φ < 110°) defines the orientation in
which the amide carbonyl of the PBR ligands related
to 1 interacts with a suitable amino acid residue in the
receptor. In other words, the high-affinity PBR ligands
related to 1 interact with their receptor, showing nearly
perpendicular carbonyl orientations with respect to the
plane of the bicyclic aromatic system.4
a
Reagents: (i) NaOH, EtOH.
of PBR ligands with their receptors has been stressed
by several authors.3,4,17,18 This is one of the most
important groups in the structure of 1, which is, in
principle, capable of giving long-range (strong and
directional) electrostatic interactions with the receptor.
With the aim of systematically probing the PBR
bound conformation, a series of conformationally con-
strained derivatives of 1, in which the orientation of the
carbonyl dipole is suitably spaced around its degree of
conformational freedom, were designed. The 10-π aro-
matic system (quinoline or isoquinoline) with the pen-
dant phenyl ring, typical of the structure of 1, was
preserved in all the new compounds and provides the
reference frame for the rotation of the carbonyl dipole
around the bond linking to the aromatic system (φ )
dihedral angle between the plane of the bicyclic aro-
matic system and the plane of the carbonyl group (a-
b-c-d), see Chart 1).
All the compounds belonging to this series of confor-
mationally constrained derivatives of 1 were tested for
their potential ability to displace [3H]-1 from PBR in
rat cortex in comparison with 1, and the results of these
studies are summarized in Table 2. The values of the
dihedral angle φ found in the AM119 lowest minimum
energy conformers of compounds 3-7 are also tabulated
for comparative purposes.
From these results it appears that tertiary amides
4f-l, 6b,c, and 7a ,b inhibited [3H]-1 binding, showing
IC50 values in the low nanomolar range and very similar
to that shown by 1. In these compounds the steric
repulsion between the amide methyl group and the 10-π
aromatic system methyl substituent (or the bridge
methylene in the case of 7a ,b) makes a large portion of
conformational space energetically inaccessible. This
is shown in Figure 1, where the AM1 energy profiles of
the parent compounds 4g and 6b are compared with
the energy profile of compound 1.
The AM1 absolute minimum, which corresponds to φ
) 110° and 112° in compound 1 and 6b, respectively,
shifts to 85° in compound 4g. Moreover, the presence
of the methyl substituents on the 10-π aromatic system
significantly reduces the breadth of the minimum, as
expected. In fact, for compound 1 the conformations
found within 3 kcal/mol from the global minimum
correspond to dihedral angle values between the plane
of the bicyclic aromatic system and the carbonyl axis
in the range of 70° < φ < 140°, whereas the range is
The low affinities showed by the constrained deriva-
tives 3a -g and 5a -e strongly support this hypothesis
(see Table 2). In fact, compounds 3a -d , in which the
carbonyl dipole is locked on the same plane as the
bicyclic aromatic system and shows a trans orientation
with respect to the pendant phenyl ring (φ ≈ 180°), are
essentially inactive. In the case of compounds 3e-g,
PBR affinity significantly increases along with the
increasing size of the amide substituent, but it never
reaches the low nanomolar level. The AM1-optimized
structure of such compounds shows that the carbonyl
group tends to be oriented out of the plane of the bicylic
aromatic system with a dihedral angle φ ≈ 120°. Thus,
the carbonyl orientation of compounds 3e-g approaches
the assumed bioactive range (80° < φ < 110°), and it is
noteworthy that compound 3g is >59 times more active
than 3c (φ ≈ 180°) and 81 times less active than 4h (φ
) 85°). A similar consideration also applies to com-
pounds 5c-e for which X-ray diffraction studies per-
formed on 5d (see the Experimental Section) showed
that the dihedral angle φ was about 43° and relaxed to
51° after AM1 full optimization.
It is interesting to note that compound 5e is only
about 1 order of magnitude less potent than 1 and about
1 order of magnitude more active than compound 5b
where the carbonyl dipole is locked in the planar cis
orientation (φ ≈ 0°).
Secondary amides 4a -e and 6a show significantly
lower PBR affinities when compared to their N-methy-

2914 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 18
Cappelli et al.
Ta ble 2. PBR Binding Affinities and φ Dihedral Angle Values of the AM1 Lowest Minimum Energy Conformers of Compounds 3-7
compd
bridge (B)
CH2
CH2
CH2
X
R1
s-Bu
s-Bu
Bn
Bn
Me
s-Bu
Bn
s-Bu
s-Bu
Bn
4-ClBn
4-ClBn
s-Bu
s-Bu
Bn
4-ClBn
4-ClBn
4-ClPh
4-MeO-Ph
s-Bu
Bn
R2
IC50 (nM) ( SEMa
φ (deg)b
3a
3b
3c
3d
3e
3f
3g
4a
4b
4c
4d
4e
4f
4g
4h
4i
4j
4k
4l
5a
5b
5c
5d
5e
6a
6b
6c
7a
7b
1
H
F
>10000
175
175
175
175
122
122
122
64
64
64
64
64
85
85
85
85
85
85
85
2
2
51
51
51
132
112
112
85
85
110
>10000
H
H
H
H
H
H
F
H
H
F
H
F
H
H
F
>10000
n-BuCH
810 ( 110
>10000
CH2CH2CH2
CH2CH2CH2
CH2CH2CH2
1200 ( 200
170 ( 14
230 ( 70
13 ( 4.9
1200 ( 180
1700 ( 230
270 ( 66
2.1 ( 0.9
2.9 ( 0.5
2.1 ( 0.6
9.8 ( 2.0
3.4 ( 1.0
6.4 ( 1.1
8.8 ( 1.5
620 ( 100
480 ( 29
780 ( 20
490 ( 47
45 ( 15
H
H
H
H
H
Me
Me
Me
Me
Me
Me
Me
H
H
CH2
CH2
CH2CH2CH2
CHdCHCH2
CHdCHCH2
s-Bu
s-Bu
Bn
s-Bu
s-Bu
Bn
c
c
H
Me
Me
550 ( 93
11 ( 2.6
3.1 ( 0.8
8.9 ( 2.5
10 ( 0.7
2.2 ( 0.3
CH2CH2
OCH2CH2
d
a
b
Each value is the mean ( SEM of three determinations. φ is the dihedral angle between the plane of the bicyclic aromatic system
and the plane of the carbonyl group (a-b-c-d), see Chart 1. c See Scheme 4. See Scheme 6.
d
4b, and 80° < φ < 160° for compound 6a is observed.
On the other hand, the differences in the conformational
preferences between secondary and tertiary amides are
amplified by analyzing the range of allowed conforma-
tions at the ab initio level (Table 3), since this analysis
shows that the minimum comprises φ dihedral angle
values in the range 0-80° (compound 4b) and 90-180°
(compound 6a ). However, these data also appear to be
inadequate to completely account for the relevant dif-
ferences in PBR affinity between secondary and tertiary
amides, so further work is in progress in order to explain
this intriguing finding.
Along with the amide carbonyl, the pendant phenyl
is a pharmacophoric group3,4,18 in the structure of 1
which presents 1 degree of conformational freedom. To
determine the bioactive conformation of this phenyl
ring, 6-31G* calculations were performed on the most
active compound 4h . The strong steric repulsion be-
tween the hydrogen atoms in the 2,6-positions of the
phenyl group and that in 5-position of the bicyclic
system or those of the nuclear methyl group was found
to constrain the orientation of the phenyl moiety out of
the plane of the bicyclic system. Thus, for compound
4h the lowest energy minimum conformation shows a
dihedral angle between the bicyclic system and the
phenyl substituent (a′-b′-c′-d′) (see Chart 1) ψ ) 90°,
the energetically allowed phenyl orientations (within 3
kcal/mol from the minimum) are in the range of 60° <
F igu r e 1. Conformational energy ∆E (kcal/mol) vs dihedral
angle φ for compounds 4g and 6b. The energy profile of
compound 1 (PK11195) is reported for comparative purposes.
lated counterparts 4f-j and 6b. These differences in
PBR binding affinity appear to be very difficult to
rationalize when the conformational profile of these
compounds (Figure 2) is compared to those reported for
compounds 4g and 6b (Figure 1). In fact, only a
moderate shift and broadening of the minimum in the
region comprised between 40° < φ < 110° for compound

Mapping the Peripheral Benzodiazepine Receptor Binding Site
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 18 2915
Ta ble 3. Comparison between AM1 and ab Initio (RHF/6-31G*) Conformational Preferences of Some Selected Compounds
AM1
6-31G*
φ
range of conf within
φ
range of conf within
3 kcal/molb (deg)
∆Ec
(kcal/mol)
compd
pIC50
mina (deg)
3 kcal/molb (deg)
mina (deg)
4g
6b
4b
6a
1
8.54
7.95
7.89
6.26
8.66
85
112
64
132
110
60-110
80-130
40-110
80-160
70-140
70
112
30
132
110
60-90
90-130
0-80
90-180
80-150
0.06
0.00
1.41
0.00
0.00
a
b
φ min is the dihedral angle value of the lowest minimum energy conformer. Range of φ values in the conformers within 3 kcal/mol
from the lowest energy minimum. c ∆E is the energy difference, computed at ab initio level, between the conformer corrresponding to the
6-31G* absolute minimum and the conformer corresponding to the AM1 absolute minimum.
absent in compounds with high affinity (compare 6b
with 6c, and 4f with 4h ). This particular behavior
seems to claim for a short-range attractive interaction
sustained by the amide side chain, and the homoge-
neous SAFIR trend suggests that a common mode of
interaction with the receptor exists.
A similar SAFIR trend was found when the effect of
the 2′-substitution in compounds 4 was investigated. In
fact, the positive influence of the 2′-fluoro substitution
is more evident in compounds with medium to low
affinity and is reduced or absent in compounds with
high affinity. It is noteworthy that the effects of the
2′-fluoro substitution on the PBR affinities of compounds
4a -j are very similar to those described for the 2′-chloro
substitution in analogous isoquinolines (see ref 3),
suggesting for the alogen in 2′-position a hydrophobic
role. Since for the pendant phenyl group a short-range
attractive interaction is feasible,3,4 a further character-
ization of the kind of interaction sustained by the
F igu r e 2. Conformational energy ∆E (kcal/mol) vs dihedral
angle φ for compounds 4b and 6a . The energy profile of
pendant phenyl group was obtained by the conforma-
tional constraining of the phenyl ring in different space
positions as it occurs in compounds 7a and 7b. The lack
of strict structural requirements for the binding of the
phenyl ring (see the previous section) suggests that the
ring is likely to perform no more than a hydrophobic
space-filling role.
Finally, the position and role of the nitrogen atoms
in the quinoline and isoquinoline nuclei seem to be
irrelevant.
Qu a n tita tive Str u ctu r e-Affin ity Rela tion sh ip s.
A quantitative rationalization of this SAFIR study was
attempted by using electronic indexes and size and
shape descriptors derived by molecular modeling. A
large set of molecular orbital derived electronic reactiv-
ity indexes8 was calculated in the AM1 framework for
the whole series of compounds reported in Table 2, and
at ab initio level (RHF/6-31G*) for a selected subset of
representative compounds. Table 4 lists some selected
electronic indexes computed with the semiempirical
AM1 hamiltonian and with 6-31G* basis set. It can be
noted that in average 6-31G* charges on the carbonylic
oxygen atom (qO), derived from Mulliken population
analysis, are greater in magnitude than the correspond-
ing AM1 charges, but the scaling factors vary very
slightly from one molecule to another; therefore, a good
correlation (r ) 0.974) is observed between the variation
of the AM1 and 6-31G* charges along the series of
compounds considered. The same holds for the energy
of the highest occupied molecular orbital (EHOMO) (r )
0.986) and for the dipole moment of the molecules (µ)
(r ) 0.973).
compound 1 (PK11195) is reported for comparative purposes.
ψ < 110°, and the conformation showing ψ ) 40° is 39
kcal/mol less stable than the minimum. On the other
hand, compound 7a shows a dihedral angle ψ ) 27° and
is slightly less active than 4h , while 7b shows both a
dihedral angle ψ ) 50° and a PBR affinity comparable
to that of 7a . Therefore, the (quasi systematic) con-
straining of the pendant phenyl group fails to reveal a
clear preference of the receptor in the interaction with
a particular conformation of the phenyl. Thus, the
orientation of the phenyl substituent does not appear
to be of crucial importance for the interaction of PBR
ligands related to 1 with their receptor, according to the
hypothesis by Georges et al.4
Qu a lita tive Str u ctu r e-Affin ity Rela tion sh ip s.
Although the remarkable structural similarity existing
in all the new compounds might be mirrored by a
similarity in the nature of the ligand-receptor interac-
tions for all the constrained derivatives, a structure-
affinity relationship (SAFIR) study has been undertaken
in order to test this assumption. Initially, the role of
the amide substituents was characterized by introduc-
ing several highly lipophilic side chains on compounds
4. The results showed that both benzyl and sec-butyl
groups are optimal for the interaction with the receptor
despite their structural differences. These two amide
substituents were then selected for the characterization
of similar interactions with the PBR binding site of the
remaining new compounds 3, 5, 6, and in most cases
the benzyl group showed higher affinity than the sec-
butyl. It is interesting to remark that the positive
influence of the benzyl group is more evident in those
compounds provided with medium to low affinity (com-
pare 3f with 3g, and 5d with 5e) and it is reduced or
An accurate correlation analysis between the elec-
tronic and reactivity indexes and PBR binding affinity

2916 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 18
Cappelli et al.
Ta ble 4. Comparison between ab Initio (RHF/6-31G*) and AM1 Molecular Orbital Indices Computed on a Representative Subset of
Compounds
6-31G*
AM1
compd
pIC50
qOa
EHOMOb (eV)
µc (Debye)
qOa
EHOMOb (eV)
µc (Debye)
3a
3g
4a
4b
4f
4g
4h
4i
5b
5d
6a
6b
7a
1
NAd
6.77
6.64
7.89
8.68
8.54
8.68
8.01
6.32
6.31
6.26
7.95
8.05
8.66
-0.6182
-0.6263
-0.6314
-0.6323
-0.6292
-0.6294
-0.6350
-0.6353
-0.5391
-0.5317
-0.6438
-0.6395
-0.6351
-0.6346
-8.3561
-8.2439
-8.2347
-8.2688
-8.2106
-8.2106
-8.2372
-8.2984
-8.2439
-7.6954
-7.9351
-7.9349
-7.9137
-8.1424
6.9662
6.0407
3.3310
3.9540
4.9731
4.9731
4.8556
4.9776
6.1256
4.2840
4.0953
4.3953
4.9281
4.0784
-0.3075
-0.3397
-0.3649
-0.3563
-0.3496
-0.3509
-0.3544
-0.3523
-0.2658
-0.2797
-0.3765
-0.3630
-0.3524
-0.3616
-9.1942
-9.1561
-9.1053
-9.1724
-9.1178
-9.1571
-9.1398
-9.1818
-9.1244
-8.7829
-8.9230
-8.9305
-8.8738
-9.0930
5.6900
5.0586
2.8126
3.2837
4.2415
3.8818
4.2700
4.2623
4.9706
3.5042
3.1811
3.7329
4.3871
3.3748
a
b
qO is the charge on the carbonylic oxygen atom. EHOMO is the energy of the highest occupied molecular orbital. c µ is the dipole
d
moment of the molecule. NA ) not active (see Table 2).
Ta ble 5. PBR Binding Affinities (pIC50) and Ad-Hoc-Defined
Size and Shape Descriptors (Vd, Vin, Vout, d, and R) of
Conformationally Restrained Derivatives of 1
compd
pIC50
Vd
Vin (Å3) Vout (Å3) d (Å)
R (deg)
3d
3f
3g
4a
4b
4c
4d
4e
4f
4g
4h
4i
6.09
5.92
6.77
6.64
7.89
5.92
5.77
6.57
8.68
8.54
8.68
8.01
8.47
8.19
8.06
6.21
6.32
6.10
6.31
7.34
6.26
7.95
8.51
8.05
8.00
8.66
0.43
0.54
0.52
0.49
0.62
0.46
0.43
0.50
0.70
0.69
0.74
0.66
0.67
0.59
0.57
0.44
0.45
0.53
0.53
0.55
0.51
0.65
0.74
0.60
0.63
0.62
274.00
270.38
277.63
250.88
282.50
253.25
252.75
269.38
304.00
304.75
322.63
308.88
314.88
287.75
288.00
234.88
246.50
270.88
268.25
280.50
255.63
294.50
321.50
291.88
304.50
285.00
86.38
37.50
50.75
38.12
11.63
53.00
64.25
52.00
0.00
8.14
6.93
7.75
6.50
6.99
7.86
7.86
7.84
7.07
7.07
8.09
8.20
8.12
7.58
7.56
5.96
7.36
6.75
6.45
6.64
7.00
7.18
8.00
7.86
7.85
7.19
-165.50
-171.40
-141.20
-156.40
-153.40
-138.00
-138.00
-145.10
-158.30
-158.30
-163.50
-155.10
-157.90
-167.50
-167.10
-154.20
-139.00
-168.30
-141.50
-159.60
-179.30
-165.70
-168.40
-163.60
-164.20
-161.50
4.88
0.00
23.75
23.75
29.50
40.63
42.87
50.13
38.62
35.25
42.75
31.50
10.13
0.00
4j
4k
4l
5a
5b
5c
5d
5e
6a
6b
6c
7a
7b
1
F igu r e 3. Correlation between the PBR binding affinities
data values (pIC50) and the data values of the van der Waals
volume included (Vin) in the reference volume of the super-
molecule (see text) for the ligands listed in Table 5. The linear
regression equation is pIC50 ) 0.04((0.001)Vin - 3.62((1.29),
n ) 26, r ) 0.872, s ) 0.551, F ) 71.92, where n is the number
of compounds, r the correlation coefficient, s the standard
deviation, and F the value of the Fisher ratio; the numbers in
parentheses are the 95% confidence intervals of the regression
coefficient and of the intercept.
31.62
28.88
13.75
did not reveal any organized trend (data not reported).
This result suggests that, while the electronic features
of the ligands considered, and, in particular, those
localized on the high polar carbonyl group, are essential
for the recognition of the PBR, they are unable to
describe quantitatively the variation of the ligands
affinity toward the PBR. However, as previously shown
for molecular series of protonated ligands which bind
G-protein-coupled receptor9,21 the modulation of the
ligand binding affinity is explained by the “ad hoc”
defined size and shape descriptors.
Table 5 reports some selected geometrical as well as
size and shape molecular descriptors together with the
cologarithmic form of the considered ligands’ binding
affinity values.
We assume that the volume obtained by superimpos-
ing the most structurally different ligands showing the
highest affinities reflects the overall shape and the
conformational freedom of the PBR binding site. Thus,
compounds 4f, 4h , and 6c were chosen to constitute the
supermolecule and were superimposed as described in
the computational procedure section. Molecular de-
scriptors Vin and Vout are, respectively, the inner and
the outer van der Waals molecular volumes of the
ligands considered in relation to the resultant reference
volume of the supermolecule, and from an interpretative
point of view, they mimic the short-range attractive and
repulsive interactions with the receptor, respectively.9
Vd is defined as the difference between Vin and Vout and
is normalized with respect to the volume of the super-
molecule. The most significant QSAR models obtained
are shown in Figures 3 and 4, where Vin and Vd are
plotted against the PBR binding affinity data values.
The higher the molecular volume shared by a generic
ligand and the supermolecule (Vin), the higher the
binding affinity. A positive slope is also obtained in the
correlation involving Vd. Recalling the mathematical
formulation of this molecular descriptor, the correlation

Mapping the Peripheral Benzodiazepine Receptor Binding Site
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 18 2917
restricted range of carbonyl orientations (comprised
between 80° and 110°) which characterizes the ligands
with the greatest affinity implies a specific hydrogen-
bonding interaction. A second important pharmacoph-
oric element has emerged from the quantitative struc-
ture-activity relationship (QSAR) models obtained and
concerns the amidic substituents. In fact, ad-hoc-
defined size and shape descriptors have proved to be
able to rationalize the variation in PBR binding affinity
quantitatively. They are assumed to interpret the
variation of the short-range intermolecular interaction
in the ligand-receptor complex. The pharmacophoric
function of the pendant phenyl ring is essential, al-
though its constraining fails to reveal a clear preference
of the receptor in interacting with a particular confor-
mation. Thus, the lack of strict structural requirements
for the binding of the phenyl group suggests a dispersive
nature of its interaction with the receptor. For the
fluorine atom in the 2′-position a hydrophobic role may
be suggested. Finally, the position and role of the
nitrogen atoms in the quinoline and isoquinoline nuclei
seem to be irrelevant.
F igu r e 4. Correlation between the PBR binding affinities
data values (pIC50) and the data values of the Vd ) (Vin - Vout)/
Vsup for the ligands listed in Table 5. The linear regression
equation is pIC50 ) 10.25((0.97)Vd + 1.43((0.56), n ) 26, r )
0.912, s ) 0.462, F ) 111.09.
Exp er im en ta l Section
Melting points were determined in open capillaries on a
Gallenkamp apparatus and are uncorrected. Microanalyses
were carried out using a Perkin-Elmer 240C elemental ana-
lyzer. Merck silica gel 60 (70-230 mesh or 230-400 mesh)
was used for column chromatography, and Riedel-de Haen DC-
Mikrokarten SI F 37341 were used for TLC. 1H-NMR spectra
were recorded with a Bruker AC 200 spectrometer in the
indicated solvents (TMS as internal standard); the values of
the chemical shifts are expressed in ppm and coupling
constants (J ) in hertz (Hz). Mass spectra (EI, 70 eV) were
recorded on a VG 70-250S spectrometer. NMR spectra and
elemental analyses were performed by the Dipartimento
Farmaco Chimico Tecnologico, Universita` di Siena. Mass
spectra were performed by Centro di Analisi e Determinazioni
Strutturali, Universita` di Siena.
suggests that, in the molecular series of the compounds
considered, the binding affinities are modulated by the
molecular shape of the amide substituents through the
optimization of dispersive and steric interactions.
The need for severe steric binding requirements is
also supported by a qualitative inspection of the last two
columns of Table 5, which shows (a) the distance (d)
between the centroid of the 10-π aromatic system and
the centroid of the amidic substituents and (b) the
dihedral angle (R) which defines the spatial orientation
of the main plane on which the amidic substituents lie
with respect to the bicyclic aromatic system.
Eth yl 4-(2-F lu or op h en yl)-3-m eth ylqu in olin e-2-ca r box-
yla te (10). To a solution of 2-amino-2′-fluorobenzophenone
(8) (6.5 g, 30 mmol) in dry methanol (150 mL) with 2-oxobu-
tyric acid (3.4 g, 33 mmol) was added a 30% solution of sodium
methoxide (10 mL, 54 mmol), and the resulting mixture was
heated at reflux for 6 days. After removal of the solvent in
vacuo, the residue was partitioned between ether-THF and
water; the organic layer was discarded, while the aqueous
solution was acidified with acetic acid and extracted with ethyl
acetate. The extracts were dried over sodium sulfate and
evaporated to dryness in vacuo to obtain a thick oil which
crystallized on standing. By washing with ether, pure 9 was
obtained (2.4 g, 8.5 mmol) and dissolved in dry ethanol (100
mL), and phosphorus oxycloride (5 mL, 546 mmol) was added.
The resulting mixture was heated at reflux overnight and the
solvent evaporated in vacuo; the residue was diluted with
water, basified by addition of solid Na2CO3, and finally
extracted with dichloromethane. The extracts were washed
with water, dried over sodium sulfate, and evaporated in vacuo
to give 10 (2.6 g, yield 28%) as colorless crystals. An analytical
sample recrystallized from n-hexane melted at 107-108 °C:
1H-NMR (CDCl3) 1.49 (t, J ) 7.0, 3H), 2.35 (s, 3H), 4.56 (q, J
) 7.0, 2H), 7.19-7.58 (m, 6H), 7.66-7.74 (m, 1H), 8.21 (d, J
) 8.5, 1H). Anal. (C19H16NO2F) C, H, N.
The data values of these two geometrical descriptors
seem to indicate that in order to obtain high binding
affinities in this molecular series of ligands the two
conditions d > 7 Å and -168° < R < -153° have to be
satisfied simultaneously.
Any attempt to improve the linear QSAR models
presented by multiregression analysis of the above
mentioned size and shape descriptors with a large
variety of electronic and reactivity indexes8 was unsuc-
cessful. These elements taken together seem to support
the hypothesis that specific highly directional interac-
tion with a suitable hydrogen-bonding donor amino acid
residue of the receptor can easily be achieved from the
point of view of the electronic requirements (being the
electronic character of the CO group quite similar in
the ligands considered), once the conformational re-
quirements are satisfied.
Con clu sion s
A combined synthetic-computational approach to the
receptor mapping has been applied and a large series
of conformationally constrained derivatives of 1 has
been investigated in order to probe the PBR binding site
in a systematic way. On this basis, structural, confor-
mational, and electronic requirements for the binding
are proposed. The essential role of the carbonyl function
as a primary pharmacophoric element in the recognition
by and the binding to PBR has been confirmed and the
E t h yl 3-(Br om om et h yl)-4-(2-flu or op h en yl)q u in olin e-
2-ca r boxyla te (11). A mixture of 10 (0.8 g, 2.6 mmol) in CCl4
(50 mL) with N-bromosuccinimide (0.45 g, 2.5 mmol) and
dibenzoyl peroxide (0.1 g, 0.4 mmol) was heated at reflux
overnight, the solvent was evaporated in vacuo, and the
residue was diluted with a small portion of the same solvent
and filtered. The filtrate was concentrated in vacuo, and the
residue was purified by column chromatography by eluting
with dichloromethane. Thus pure 11 was obtained as colorless

2918 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 18
Cappelli et al.
crystals (0.84 g, yield 83%). Recrystallization from n-hexane
gave an analytical sample melting at 91-92 °C: 1H-NMR
(CDCl3) 1.52 (t, J ) 7.2, 3H), 4.51-4.68 (m, 3H), 5.00 (d, J )
10.0, 1H), 7.30-7.42 (m, 4H), 7.49-7.64 (m, 2H), 7.73-7.81
(m, 1H), 8.25 (d, J ) 8.3, 1H). Anal. (C19H15NO2BrF) C, H,
N.
(()-1,2-Dih yd r o-9-(2-flu or op h en yl)-2-(1-m eth ylp r op yl)-
3H-p yr r olo[3,4-b]qu in olin -3-on e (3b). A mixture of 11 (0.23
g, 0.59 mmol) in ethanol (20 mL) and sec-butylamine (0.16 mL,
1.6 mmol) with K2CO3 (0.5 g, 3.6 mmol) was refluxed for 1 h
and the solvent then removed in vacuo. The residue was
partitioned between CHCl3 and water, and the organic layer
then was washed with water, dried over sodium sulfate, and
concentrated in vacuo. Purification of the residue by column
chromatography, eluting with CH2Cl2-ethyl acetate (8:2), gave
3b as a white powder: 1H-NMR (CDCl3) 0.86-0.95 (m, 3H),
1.24-1.30 (m, 3H), 1.55-1.71 (m, 2H), 4.10-4.34 (m, 2H),
4.53-4.71 (m, 1H), 7.28-7.43 (m, 3H), 7.50-7.82 (m, 4H), 8.46
(d, J ) 8.4, 1H).
Gen er a l P r oced u r e for th e Syn th esis of Com p ou n d s
4 a n d 6. Meth od A. To a solution of the suitable carboxylic
acid (9, 1210 or 20) (1.0 mmol) in CH2Cl2 (20 mL) cooled at
0-5 °C was added SOCl2 (1.0 mL), and the resulting solution
was stirred at room temperature for 30 min. The solvent was
distilled under reduced pressure, and excess SOCl2 was
removed by distillation with toluene. The residue was diluted
with CH2Cl2 (20 mL) and cooled at 0-5 °C, and the suitable
primary amine (30 mmol) was added (in the case of compounds
4c-e 1.1 mmol of the suitable primary amine and 20 mmol of
triethylamine were used). The resulting mixture was stirred
overnight at room temperature, washed with water, dried over
sodium sulfate, and concentrated under reduced pressure. The
residue was recrystallized from the suitable solvent, and the
corresponding secondary amide (4a -e or 6a ) was obtained in
analytically pure form.
ence of two different rotamers in equilibrium. For the sake
of simplification, the integral intensities have not been given.)
(()-N,3-Dim eth yl-N-(1-m eth ylp r op yl)-4-p h en ylqu in o-
lin e-2-ca r boxa m id e (4f):
1.03 (t, J ) 7.4), 1.22-1.69 (m), 2.20 (s), 2.23 (s), 2.72 (s), 3.04
(s), 3.40-3.58 (m), 4.82-4.92 (m), 7.26-7.68 (m), 8.07-8.14
(m).
(()-N,3-Dim et h yl-4-(2-flu or op h en yl)-N-(1-m et h ylp r o-
p yl)qu in olin e-2-ca r boxa m id e (4g): 1H-NMR (CDCl3) 0.79-
0.92 (m), 0.98-1.09 (m), 1.19-1.71 (m), 2.22-2.25 (m), 2.71
(s), 3.04 (s), 3.41-3.49 (m), 4.85-4.95 (m), 7.22-7.70 (m),
8.09-8.15 (m); MS m/ z calcd for C22H23N2OF (M+) 350.1794,
found 350.1796.
N-[(4-Ch lor op h en yl)m et h yl]-N,3-d im et h yl-4-p h en yl-
qu in olin e-2-ca r boxa m id e (4i): 1H-NMR (CDCl3) 2.19 (s),
2.20 (s), 2.83 (s), 3.13 (s), 4.43 (s), 4.83 (s), 7.20-7.69 (m), 8.07-
8.14 (m).
N-[(4-Ch lor op h en yl)m et h yl]-N,3-d im et h yl-4-(2-flu o-
r op h en yl)qu in olin e-2-ca r boxa m id e (4j): 1H-NMR (CDCl3)
2.20 (s), 2.22 (s), 2.81 (s), 3.13 (s), 4.41 (s), 4.82 (AB q, J )
14.6), 7.22-7.71 (m), 8.08-8.15 (m).
(()-N,4-Dim eth yl-N-(1-m eth ylp r op yl)-1-p h en ylisoqu in -
olin e-3-ca r boxa m id e (6b): 1H-NMR (CDCl3) 0.80 (t, J ) 7.3),
1.01 (t, J ) 7.4), 1.19-1.73 (m), 2.65 (s), 2.67 (s), 2.69 (s), 3.00
(s), 3.38-3.53 (m), 4.79-4.98 (m), 7.46-7.78 (m), 8.04-8.11
(m).
1H-NMR (CDCl3) 0.86 (t, J ) 7.4),
Meth od C. A solution of the appropriate carboxylic acid
(1210 or 20) (1.0 mmol) in CH2Cl2 (20 mL) with SOCl2 (1.0 mL,
13.7 mmol) was stirred at room temperature for 30 min, and
the solvent was evaporated under reduced pressure. The
excess of SOCl2 was removed by azeotropic distillation with
toluene and the residue diluted with CH2Cl2 (20 mL) and then
cooled to 0-5 °C. To the resulting mixture the suitable
secondary amine (1.1 mmol) (in the case of compound 4k ,
N-methyl-4-chloroaniline hydrochloride was used) and triethy-
lamine (1.4 mL, 10 mmol) were added in sequence. The
reaction mixture was stirred overnight at room temperature,
washed with water, dried over sodium sulfate, and concen-
trated under reduced pressure. Purification of the residue by
column chromatography, eluting with CH2Cl2-ethyl acetate
(8:2), gave the corresponding tertiary amide (4h ,k ,l or 6c).
Compounds 4k ,l were recrystallized from the suitable solvent.
(Since the amide nitrogen of these compounds bears two
different substituents, their 1H-NMR spectra show the pres-
ence of two different rotamers in equilibrium. For the sake
of simplification the integral intensities have not been given).
N,3-Dim et h yl-4-p h en yl-N-(p h en ym et h yl)qu in olin e-2-
(()-3-Meth yl-N-(1-m eth ylp r op yl)-4-p h en ylqu in olin e-2-
ca r boxa m id e (4a ): 1H-NMR (CDCl3) 1.03 (t, J ) 7.5, 3H),
1.32 (d, J ) 6.9, 3H), 1.57-1.79 (m, 2H), 2.56 (s, 3H), 4.07-
4.21 (m, 1H), 7.19-7.69 (m, 8H), 7.88 (d, J ) 8.1, 1H), 8.09 (d,
J ) 8.3, 1H).
(()-4-(2-F lu or op h en yl)-3-m et h yl-N-(1-m et h ylp r op yl)-
1
qu in olin e-2-ca r boxa m id e (4b): H-NMR (CDCl3) 1.03 (t, J
) 7.4, 3H), 1.30-1.33 (m, 3H), 1.60-1.76 (m, 2H), 2.59 (s, 3H),
4.05-4.20 (m, 1H), 7.16-7.71 (m, 7H), 7.88 (br s, 1H), 8.10 (d,
J ) 8.4, 1H).
3-Meth yl-4-p h en yl-N-(p h en ylm eth yl)qu in olin e-2-ca r -
boxa m id e (4c): 1H-NMR (CDCl3) 2.60 (s, 3H), 4.73 (d, J )
6.2, 2H), 7.20-7.69 (m, 13H), 8.05 (d, J ) 8.4, 1H) 8.47 (t, J )
6.2, 1H).
1
ca r boxa m id e (4h ): H-NMR (CDCl3) 2.19 (s), 2.23 (s), 2.83
(s), 3.15 (s), 4.48 (s), 4.87 (s), 7.21-7.70 (m), 8.12 (d, J ) 8.6).
N-(4-Ch lor op h en yl)-N,3-d im eth yl-4-p h en ylqu in olin e-
2-ca r boxa m id e (4k ): 1H-NMR (CDCl3) 2.14 (s), 2.31 (s), 3.30
(s), 3.56 (s), 7.07-7.63 (m), 7.98 (d, J ) 8.4), 8.15 (d, J ) 8.4).
N,3-Dim eth yl-N-(4-m eth oxyph en yl)-4-ph en ylqu in olin e-
N-[(4-Ch lor op h en yl)m eth yl]-3-m eth yl-4-p h en ylqu in o-
lin e-2-ca r boxa m id e (4d ): 1H-NMR (CDCl3) 2.59 (s, 3H), 4.69
(d, J ) 6.3, 2H), 7.21-7.69 (m, 12H), 8.06 (d, J ) 8.4, 1H)
8.53 (t, J ) 6.3, 1H).
1
N-[(4-Ch lor op h en yl)m eth yl]-4-(2-flu or op h en yl)-3-m e-
2-ca r boxa m id e (4l): H-NMR (CDCl3) 2.14 (s), 2.32 (s), 3.27
1
th ylqu in olin e-2-ca r boxa m id e (4e): H-NMR (CDCl3) 2.63
(s), 3.54 (s), 3.66 (s), 3.84 (s), 6.61-6.65 (m), 6.97-7.61 (m),
8.00 (d, J ) 8.4), 8.16 (d, J ) 8.4).
(s, 3H), 4.69 (d, J ) 6.4, 2H), 7.17-7.71 (m, 11H), 8.07 (d, J )
8.3, 1H) 8.54 (t, J ) 6.4, 1H).
N,4-Dim eth yl-1-p h en yl-N-(p h en ylm eth yl)isoqu in olin e-
3-ca r boxa m id e (6c): 1H-NMR (CDCl3) 2.69 (s), 2.70 (s), 2.80
(s), 3.10 (s), 4.44 (s), 4.84 (s), 7.25-7.79 (m), 8.03-8.11 (m).
Eth yl 1-P h en yl-4-[[(tr iflu or om eth yl)su lfon yl]oxy]iso-
qu in olin e-3-ca r boxyla te (14). To a solution of 13 (0.96 g,
3.3 mmol) in anhydrous pyridine (15 mL) cooled at 0 °C was
added dropwise triflic anhydride (0.66 mL, 3.9 mmol). The
resulting mixture was stirred for 5 min at 0 °C and overnight
at room temperature. The reaction mixture was poured into
ice-water and the precipitate collected by filtration, washed
with water, and dried to give pure 14 (1.3 g, yield 93%). An
analytical sample was recrystallized from methanol (mp 129-
130 °C):
(()-4-Meth yl-N-(1-m eth ylpr opyl)-1-ph en ylisoqu in olin e-
1
3-ca r boxa m id e (6a ): H-NMR (CDCl3) 0.97 (t, J ) 7.4, 3H),
1.24 (d, J ) 6.7, 3H), 1.52-1.74 (m, 2H), 3.20 (s, 3H), 4.05-
4.20 (m, 1H), 7.53-7.81 (m, 7H), 8.11 (d, J ) 8.4, 1H), 8.28 (d,
J ) 8.6, 1H), 8.34 (br s, 1H).
Meth od B. To a solution of the appropriate secondary
amide (4a ,b,d ,e or 6a ) (0.5 mmol) in anhydrous DMF (5 mL)
with CH3I (0.5 mL, 8.0 mmol) cooled at 0 °C under argon
atmosphere was added NaH (0.05 g, 2.1 mmol). The resulting
mixture was stirred at 0 °C for 30 min. The reaction mixture
was poured into ice-water, neutralized with 1 N HCl, and
extracted with CHCl3 (3 × 25 mL). The combined organic
extracts were washed with water, dried over sodium sulfate,
and concentrated under reduced pressure to give the corre-
sponding tertiary amide, which was purified by recrystalliza-
tion from the suitable solvent (4f,i) or by column chromatog-
raphy by eluting with CHCl3-ethyl acetate (8:2) (4g,j or 6b).
(Since the amide nitrogen of these compounds bears two
different substituents, their 1H-NMR spectra show the pres-
1H-NMR (CDCl3) 1.47 (t, J ) 6.9, 3H), 4.55 (q, J )
6.9, 2H), 7.53-7.56 (m, 3H), 7.70-7.79 (m, 3H), 7.89-7.97 (m,
1H), 8.18-8.27 (m, 2H). Anal. (C19H14NO5SF3) C, H, N.
Eth yl 4-Meth yl-1-ph en ylisoqu in olin e-3-car boxylate (15).
A solution of 14 (0.64 g, 1.5 mmol) in anhydrous DMF with
LiCl (0.35 g, 8.3 mmol) and tetramethyltin (1.0 mL, 7.2 mmol)
was stirred under an argon stream for 10 min, and Pd(PPh3)-
Cl2 (0.11 g, 0.16 mmol) was added. The mixture was heated

Mapping the Peripheral Benzodiazepine Receptor Binding Site
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 18 2919
at 140 °C for 40 min, poured into ice-water, and extracted
with CH2Cl2 (5 × 15 mL). The combined organic extracts were
washed with water, dried over sodium sulfate, and concen-
trated under reduced pressure. The residue was purified by
column chromatography by eluting with CH2Cl2-ethyl acetate
(9:1), and analytically pure 15 (0.42 g, yield 96%) was obtained
as a pale yellow oil: 1H-NMR (CDCl3) 1.45 (t, J ) 6.9, 3H),
2.85 (s, 3H), 4.50 (q, J ) 6.9, 2H), 7.48-7.83 (m, 7H), 8.12 (d,
J ) 8.1, 1H), 8.18 (d, J ) 8.5, 1H). Anal. (C19H17NO2) C, H,
N.
The catalyst was removed by filtration and washed with hot
ethanol and the filtrate concentrated under reduced pressure.
Purification of the residue by chromatography with CH2Cl2-
ethyl acetate (8:2) as eluent gave 19 (0.14 g, yield 74%) as a
colorless oil which crystallized by treatment with diethyl
ether-n-hexane. An analytical sample was recrystallized from
diethyl ether-n-hexane (mp 101-102 °C): 1H-NMR (DMSO-
d6) 1.30 (t, J ) 7.0, 3H), 3.65-3.72 (m, 2H), 4.31 (q, J ) 7.2,
2H), 4.65 (t, J ) 5.0, 1H), 6.09-6.21 (m, 1H), 6.84 (d, J ) 11.3,
1H), 7.56-7.80 (m, 6H), 7.86-7.93 (m, 1H), 8.05 (d, J ) 8.3,
1H), 8.11 (d, J ) 8.3, 1H). Anal. (C21H19NO3) C, H, N.
E t h yl 4-(Br om om et h yl)-1-p h en ylisoq u in olin e-3-ca r -
boxyla te (16). A mixture of 15 (0.18 g, 0.62 mmol) in CCl4
(20 mL) with N-bromosuccinimide (0.11 g, 0.62 mmol) and
dibenzoyl peroxide (0.03 g, 0.12 mmol) was heated at reflux
for 2 h and 15 min, the solvent was evaporated in vacuo, and
the residue was diluted with a small portion of the same
solvent and filtered. The filtrate was concentrated in vacuo
and the residue purified by column chromatography by eluting
with dichloromethane. Thus, pure 16 was obtained as color-
less crystals (0.22 g, yield 96%); recrystallization from n-
hexane gave an analytical sample melting at 109-110 °C: 1H-
NMR (CDCl3) 1.48 (t, J ) 7.3, 3H), 4.54 (q, J ) 7.2, 2H), 5.30
(s, 2H), 7.51-7.73 (m, 6H), 7.85-7.93 (m, 1H), 8.18 (d, J )
8.4, 1H), 8.32 (d, J ) 8.6, 1H). Anal. (C19H16NO2Br) C, H, N.
(()-1,2-Dih yd r o-2-(1-m et h ylp r op yl)-5-p h en yl-3H -p yr -
r olo[3,4-c]isoqu in olin -3-on e (5a ). A mixture of 16 (0.11 g,
0.30 mmol) in ethanol (20 mL) with sec-butylamine (0.18 mL,
1.8 mmol) was heated at reflux for 2 h, and then the volatiles
were removed under reduced pressure. Purification by column
chromatography of the residue, using CH2Cl2-ethyl acetate
(8:2) as eluent, gave pure 5a as white crystals (0.086 g, yield
90%): 1H-NMR (CDCl3) 0.96 (t, J ) 7.4, 3H), 1.37 (d, J ) 6.8,
3H), 1.67-1.82 (m, 2H) 4.56-4.76 (m, 3H), 7.49-7.53 (m, 3H),
7.62-7.74 (m, 3H), 7.80-7.88 (m, 1H), 7.96 (d, J ) 8.0, 1H),
8.23 (d, J ) 8.5, 1H).
Meth od D. To a solution of the suitable alcohol (18 or 19)
(0.4 mmol) in CH2Cl2 (20 mL) cooled at 0-5 °C was added,
SOCl2 (1.0 mL, 13.7 mmol) and the mixture was stirred at
room temperature for 2 h. The solvent was evaporated under
reduced pressure, and the excess of SOCl2 was removed by
azeotropic distillation with benzene. The residue was dis-
solved with ethanol (20 mL), sec-butylamine (3.0 mL, 29.7
mmol) was added, and the resulting mixture was heated to
reflux overnight. The volatiles were removed under reduced
pressure, the residue was diluted with absolute ethanol (30
mL), and to the resulting mixture was added a 30% MeONa/
MeOH solution (1.0 mL, 5.4 mmol). After 2 h of heating at
reflux, the reaction mixture was cooled in an ice-water bath
and neutralized with 1 N ethanolic HCl. The solvent was
removed under reduced pressure, the residue was diluted with
CH2Cl2 (30 mL), and SOCl2 (1.0 mL, 13.7 mmol) was added.
After 30 min of stirring at room temperature, the volatiles were
distilled under reduced pressure, benzene (20 mL) was added,
and the solvent was evaporated. The residue was partitioned
between CH2Cl2 (30 mL) and a saturated solution of NaHCO3
(20 mL). The organic layer was washed with water, dried over
sodium sulfate, and concentrated under reduced pressure.
Purification of the residue by column chromatography using
CHCl3-ethyl acetate (8:2) as eluent gave the corresponding
lactam (5c or 5d ) which was recrystallized from the suitable
solvent.
1,2-Dih yd r o-5-p h en yl-2-(p h en ylm et h yl)-3H -p yr r olo-
[3,4-c]isoqu in olin -3-on e (5b). Starting from 16 (0.11 g, 0.30
mmol), 5b was prepared as described for 5a , by using benzy-
lamine (0.20 mL, 1.8 mmol) instead of sec-butylamine: 1H-
NMR (CDCl3) 4.61 (s, 2H), 4.97 (s, 2H), 7.26-7.87 (m, 13H),
8.23 (d, J ) 8.6, 1H).
(()-4-(1-Meth ylp r op yl)-7-p h en yl-1,2,3,4-tetr a h yd r o-5H-
a zep in o[3,4-c]isoqu in olin -5-on e (5c): 1H-NMR (CDCl3) 1.01
(t, J ) 7.3, 3H), 1.27 (d, J ) 6.8, 3H), 1.50-1.70 (m, 2H), 2.04-
2.18 (m, 2H), 3.19-3.33 (m, 4H), 4.87-4.98 (m, 1H), 7.46-
7.61 (m, 4H), 7.70-7.80 (m, 3H), 8.13-8.22 (m, 2H).
Eth yl 4-(3-Hyd r oxyp r op yn -1-yl)-1-p h en ylisoqu in olin e-
3-ca r boxyla te (17). A mixture of 14 (0.43 g, 1.0 mmol),
propargyl alcohol (0.60 mL, 10.3 mmol), triethylamine (1.0 ml,
7.2 mmol), cuprous iodide (0.016 g, 0.084 mmol), and anhy-
drous DMF (10 mL) was degassed with argon for 10 min. Pd-
(PPh3)Cl2 (0.071 g, 0.1 mmol) was added, and the resulting
mixture was heated to 60-65 °C for 45 min, poured into ice-
water, and extracted with CHCl3 (3 × 20 mL). The combined
organic extracts were washed with water, dried over sodium
sulfate, and concentrated under reduced pressure. The residue
was purified by column chromatography, eluting with CH2-
Cl2-ethyl acetate (8:2) to give 17 (0.21 g, yield 63%). An
analytical sample was recrystallized from diethyl ether-n-
hexane (mp 86-87 °C): 1H-NMR (DMSO-d6) 1.36 (t, J ) 7.3,
3H), 4.40 (q, J ) 7.2, 2H), 4.51 (d, J ) 6.0, 2H), 5.54 (t, J )
6.2, 1H), 7.56-7.71 (m, 5H), 7.78-7.86 (m, 1H), 7.97-8.11 (m,
2H), 8.48 (d, J ) 8.2, 1H). Anal. (C21H17NO3) C, H, N.
E t h yl 4-(3-H yd r oxyp r op yl)-1-p h en ylisoq u in olin e-3-
ca r boxyla t e (18). A solution of 17 (0.14 g, 0.42 mmol) in
ethanol (20 mL) with 10% Pd/C (0.07 g) was hydrogenated in
a Parr apparatus at 25 psi at room temperature for 4 h. The
catalyst was removed by filtration and washed with hot
ethanol and the filtrate concentrated under reduced pressure.
Purification of the residue by chromatography with CHCl3-
ethyl acetate (8:2) as eluent gave 18 (0.11 g, yield 78%) as a
colorless oil which crystallized on standing. An analytical
sample was obtained from diethyl ether-n-hexane (mp 62-
63 °C): 1H-NMR (DMSO-d6) 1.34 (t, J ) 7.1, 3H), 1.76-1.90
(m, 2H), 3.15-3.23 (m, 2H), 3.56 (q, J ) 5.8, 2H), 4.39 (q, J )
7.1, 2H), 4.64 (t, J ) 5.3, 1H), 7.54-7.77 (m, 6H), 7.88-8.04
(m, 2H), 8.34 (d, J ) 8.5, 1H). Anal. (C21H21NO3) C, H, N.
(z)-Eth yl 4-(3-Hyd r oxyp r op en -1-yl)-1-p h en ylisoqu in o-
lin e-3-ca r boxyla te (19). A solution of 17 (0.19 g, 0.57 mmol)
in ethanol (30 mL) with 5% PdS/C (0.1 g) was hydrogenated
in a Parr apparatus at 20 psi at room temperature for 2 h.
(()-3,4-Dih yd r o-4-(1-m et h ylp r op yl)-7-p h en yl-5H -a ze-
p in o[3,4-c]isoqu in olin -5-on e (5d ): 1H-NMR (CDCl3) 0.97 (t,
J ) 7.4, 3H), 1.24 (d, J ) 6.8, 3H), 1.52-1.67 (m, 2H), 3.67 (d,
J ) 7.1, 2H), 4.73-4.91 (m, 1H), 6.52-6.64 (m, 1H), 7.43-
7.65 (m, 5H), 7.74-7.82 (m, 3H), 8.13-8.20 (m, 2H).
3,4-Dih yd r o-7-p h en yl-4-(p h en ylm et h yl)-5H -a zep in o-
[3,4-c]isoqu in olin -5-on e (5e) (Meth od E). To a solution of
19 (0.08 g, 0.24 mmol) in CH2Cl2 (20 mL) cooled at 0-5 °C
was added, SOCl2 (1.0 mL, 13.7 mmol) and the mixture was
stirred at room temperature for 2 h. The solvent was
evaporated under reduced pressure, and the excess of SOCl2
was removed by azeotropic distillation with benzene. The
residue was dissolved with ethanol (20 mL), benzylamine (1.0
mL, 9.1 mmol) was added, and the resulting mixture was
heated at reflux overnight. The volatiles were removed under
reduced pressure, the residue was diluted with toluene,
2-hydroxypyridine (0.05 g, 0.52 mmol) was added, and the
resulting mixture was heated at reflux for an additional 24 h.
The solvent was removed under reduced pressure and the
residue purified by chromatography. Elution with CHCl3-
ethyl acetate (8:2) gave pure 5e: 1H-NMR (CDCl3) 3.76 (d, J
) 7.0, 2H), 4.89 (s, 2H), 6.29-6.40 (m, 1H), 7.26-7.66 (m,
10H), 7.75-7.82 (m, 3H), 8.13-8.20 (m, 2H).
4-Meth yl-1-p h en ylisoqu in olin e-3-ca r boxylic Acid (20).
A solution of 15 (0.41 g, 1.4 mmol) in ethanol (30 mL) with 3
N NaOH (10 mL, 30 mmol) was heated at reflux for 20 min
and the solvent evaporated under reduced pressure. The
residue was diluted with cold water (30 mL) and acidified with
3 N HCl (pH 5-6). The precipitate was extracted with CHCl3
(3 × 10 mL), and the combined organic extracts were washed
with water, dried over sodium sulfate, and concentrated under
reduced pressure. Thus, 20 (0.33 g, yield 89%) was obtained
as a white solid and used in the next steps without further

2920 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 18
Cappelli et al.
purification: 1H-NMR (CDCl3) 3.25 (s, 3H), 7.55-7.75 (m, 6H),
7.85-7.93 (m, 1H), 8.20 (d, J ) 8.6, 1H), 8.36 (d, J ) 8.4, 1H).
Meth od F . To a solution of the suitable imidoyl chloride
(21a or 21b) (0.41 mmol) in anhydrous DMF (5 mL) with PPh3
(0.026 g, 0.1 mmol) under a CO atmosphere were added PdAc2
(0.011 g, 0.049 mmol) and N-methylbenzylamine (0.27 mL, 2.1
mmol) in sequence. The resulting mixture was heated at 80-
90 °C for 7 h, poured into ice-water, and extracted with CHCl3
(4 × 20 mL). The organic layer was washed with water, dried
over sodium sulfate, and evaporated under reduced pressure
to give an oil residue which was purified by column chroma-
tography. Elution with n-hexane-ethyl acetate (65:35) gave
the corresponding amide (7a or 7b). (Since the amide nitrogen
5d : C23H22N2O (mol wt 342.4), a colorless single crystal,
dimensions 0.2 × 0.3 × 0.2 mm was used for data collection;
monoclinic; Pc space group; a ) 8.825(2) Å, b ) 9.261(2) Å, c
) 11.719(2) Å, â ) 108.40(3)°, V ) 908.8(3) Å3, Z ) 2, Dc )
1.25 g/cm3. A total of 3195 unique reflections (Rint ) 0.011)
were collected at 22 °C, of which 2762 were observed with Fo
> 4σ(Fo). No absorption correction was applied. The structure
was solved by direct methods of the SHELXTL PC package.23
As 5d crystallizes in
a polar space group, polar axis
restraints were applied by Flack and Schwarzenbach method.24
Atomic scattering factors including f ′ and f ′′ were taken from
ref 25. The refinement was carried out by full-matrix aniso-
tropic least-squares on F2 for all reflections for non-H atoms
by using SHELXL-93 program.26 Function ∑w(Fo2 - Fc2)2 was
minimized with weighing scheme w ) 1/[σ2(Fo)2 + aP2 + bP],
1
of these compounds bears two different substituents, their H-
NMR spectra show the presence of two different rotamers in
equilibrium. For the sake of simplification the integral
intensities have not been given.)
2
where P ) (Fo + 2Fc2)/3, a ) 0.0570 and b ) 0.1891.
The final refinement converged to R1 ) 0.054 and wR2 )
0.1103, goodness of fit ) 1.037 for all data. R1 refers to the
conventional R factor based on F, while wR2 and goodness of
fit are based on F2. R factors based on F2 are statistically
about twice as large as those based on F.26
7,8-Dih yd r o -N -m e t h yl-N -(p h e n ylm e t h yl)b e n zo[k ]-
p h en a n th r id in e-6-ca r boxa m id e (7a ): 1H-NMR (CDCl3) 2.82
(s), 2.86-2.93 (m), 3.15 (s), 4.50 (s), 4.87 (s), 7.25-7.43 (m),
7.52-7.75 (m), 7.90-7.98 (m), 8.13-8.19 (m), 8.43-8.49 (m).
6,7-Dih yd r o-N-m et h yl-N-(p h en ylm et h yl)-[1]b en zoxe-
p in o[4,5-c]qu in olin e-8-ca r boxa m id e (7b): 1H-NMR (CDCl3)
2.72-3.11 (m), 3.15 (s), 4.42-4.70 (m), 4.87 (AB q, J ) 14.7),
7.25-7.59 (m), 7.67-7.76 (m), 8.03-8.18 (m).
Com p u ta tion a l P r oced u r e. The structures of all com-
pounds studied were fully optimized by means of molecular
orbital calculations (AM1),19 using the MOPAC 6.0 (QCPE 455)
program, by applying the molecular mechanics correction to
the amide bond. The three-dimensional structure of 14 and
of compound 5d , solved by X-ray crystallography in our
laboratory (see text), were taken as the input for the AM1
optimization. The starting geometry of the other derivatives
considered were constructed within the Chemnote module of
Quanta 4.1 program. Enantiomer R was arbitrarily chosen
for those compounds which show an asymmetric carbon in
their structures. Both R and S enantiomers of 1 have been
found to be nearly equipotent toward PBR ((R)-1, IC50 ) 9 nM;
(S)-1, IC50 ) 19 nM; (R,S)-1, IC50 ) 9 nM).3
Conformational analysis was performed for the parent
compounds 4b,g and 6a , and for 1. The most important
conformational freedom is the rotation around the torsional
angle φ (see Chart 1), which modulates the space orientation
of the amide function. Uniform scanning of this dihedral angle
was performed, in a rigid manner, in steps of 10° on the fully
optimized structures. The relaxation of the other degrees of
freedom does not affect the qualitative nature of the potential
energy trends, although it usually reduces the barrier to the
rotation between conformers.
Min and max height in last ∆F map were -0.19 and 0.236
e Å -3, respectively.
During the refinement a statistical disorder was shown by
the CH and by one CH3 of the pendant substituent. For these
atoms two different positions were refined with the sum of the
site occupation factors (sof) fixed at 1. Refined values of sof
equal to 0.849(5) and 0.151(5) were found for each of the two
positions.
Full crystallographic details will be given elsewhere.
In Vitr o Bin d in g Assa ys. Male CRL: CD(SD)BR (Charles
River Italia, Calco, CO, Italy) rats were killed by decapitation,
their brains were rapidly dissected into the various areas and
stored at -80°C until the day of assay. The binding assays
were performed as described by Cantoni et al.27 The frozen
cortices were homogenized in 50 vols of ice-cold phosphate-
buffered saline (PBS) 50 mM, pH 7.4 containing 100 mM NaCl,
using an Ultra Turrax TP-1810 (2 × 20 sec). The homogenate
was centrifuged for 10 min at 1000g at 4° C (Beckman model
J -21B refrigerated centrifuge). The pellet was washed three
more times by resuspension in the same volume of fresh buffer
and centrifuged again at 50000g for 10 min. The pellet
obtained was finally resuspended in the incubation buffer
(PBS) just before the binding assay.
[3H]-1 (s.a. 86.4 Ci/mmol, NEN) binding was assayed in a
final incubation volume of 1.0 mL, consisting of 0.5 mL
membrane suspension, 0.5 mL of [3H]ligand and 20 µl of
displacing agent or solvent. Tissue concentration and [3H]-1
final concentration were 2.3 mg tissue/sample and 1 nM,
respectively.
Incubations (120 min at 4° C) were stopped by rapid
filtration under vacuum through GF/B fiber filters which were
then washed with 12 mL (3 × 4 mL) of ice-cold PBS using a
Brandel M-48RP cell harvester.
Conformational preferences in the region of the absolute
AM1 minimum were studied at the ab initio level (RHF/6-
31G*); the Gaussian94 program was used.20
Molecu la r Su p er im p osition . Compounds 4f,h , and 6c,
the most structurally different ligands which show high
affinity for PBR, were chosen for the construction of the
reference supermolecule and were superimposed, by a rigid
fit procedure, on the basis of their common quinoline or
isoquinoline nucleus. All of the other compounds were satis-
factorily matched on the supermolecule by using their global
minimum conformer or conformers whose energy is only 1 kcal/
mol above the global minimum.
Nonspecific binding was assayed in the presence of 1 (1 µM).
Dried filters were immersed in vials containing 4 mL of
Filter Count (Packard) liquid scintillation cocktail for the
measurement of trapped radioactivity with a Packard LKB
1214 RACKBETA liquid scintillation spectrometer at a count-
ing efficiency of about 60%. The concentration of the test
compounds that inhibited [3H]ligand binding by 50% (IC50) was
28
determined using the Allfit program with 6 concentrations
The ad-hoc-defined size and shape descriptor, Vd, is com-
puted according to the following formulas: Vd ) (Vin-Vout)/
Vsup where Vsup is the resultant van der Waals volume of the
reference supermolecule. Vin and Vout are, respectively, the
intersection and the outer van der Waals volume of the ligands
considered in relation to the supermolecule.
The QUANTA 4.122 molecular modeling software imple-
mented on a HP 720 personal workstation was used for
molecular comparisons, matching, and computation of van der
Waals volumes.
X-Ra y Cr ysta llogr a p h y. A single crystal of 5d was
submitted to X-ray data collection on a Siemens P4 four-circle
diffractometer with graphite monochromated Mo KR radiation.
The ω/2θ scan technique was used.
of the displacers, each performed in triplicate.
Ack n ow led gm en t. The authors are grateful to Dr.
J ean J acques Bourguignon for his collaboration. This
work was financially supported by Italian MURST (60%
and 40% funds).
Refer en ces
(1) (a) Verma, A.; Snyder, S. H. Peripheral Type Benzodiazepine
Receptors. Annu. Rev. Pharmacol. Toxicol. 1989, 29, 307-322.
(b) Saano, V.; Ra¨go, L.; Ra¨ty, M. Peripheral Benzodiazepine
Binding Sites. Pharmacol. Ther. 1989, 41, 503-514. (c) Gavish,
M.; Katz, Y.; Bar-Ami, S.; Weizman, R. Biochemical, Physiologi-
cal, and Pathological Aspects of the Peripheral Benzodiazepine
Receptor. J . Neurochem. 1992, 58, 1589-1601.

Mapping the Peripheral Benzodiazepine Receptor Binding Site
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 18 2921
(2) Peripheral Benzodiazepine Receptors; Giesen-Crouse, E., Ed.;
Academic Press: London, 1993.
(16) Anzini, M.; Cappelli, A.; Vomero, S.; Giorgi, G.; Langer, T.;
Hamon, M.; Nacera, M.; Emerit, B. M.; Cagnotto, A.; Skorupska,
M.; Mennini, T.; Pinto, J . C. Novel Potent and Selective 5-HT3
Receptor Antagonists Based on Arylpiperazine Skeleton: Syn-
thesis, Structure, Biological Activity and CoMFA Studies. J .
Med. Chem. 1995, 38, 2692-2704.
(17) Kozikowski, A. P.; Ma, D.; Brewer, J .; Sun, S.; Costa, E.; Romeo,
E.; Guidotti, A. Chemistry, Binding Affinities, and Behavioral
Properties of a New Class of “Antineophobic” Mitochondrial DBI
Receptor Complex (mDRC) Ligands. J . Med. Chem. 1993, 36,
2908-2920.
(18) Dalpiaz, A.; Bertolasi, V.; Borea, P. A.; Nacci, V.; Fiorini, I.;
Campiani, G.; Mennini, T.; Manzoni, C.; Novellino, E.; Greco,
G. A Concerted Study Using Binding Measurements, X-ray
Structural Data, and Molecular Modeling on the Stereochemical
Features Responsible for the Affinity of 6-Arylpyrrolo[2,1-d][1,5]-
benzothiazepines toward Mitochondrial Benzodiazepine Recep-
tors. J . Med. Chem. 1995, 38, 4730-4738.
(19) Dewar, M. J . S.; Zoebisch, E. G.; Healey, E. F.; Stewart, J . J . P.
AM1: A New General Purpose Quantum Mechanical Molecular
Model. J . Am. Chem. Soc. 1985, 107, 3902-3909.
(20) Frisch, M. J .; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
J ohnson, B. G.; Robb, M. A.; Cheeseman, J . R.; Keith, T. A.;
Petersson, G. A.; Montgomery, J . A.; Raghavachari, K.; Al-
Laham, M. A.; Zakrzewski, V. G.; Ortiz, J . V.; Foresman, J . B.;
Cioslowski, J .; Stefanov, B. B.; Nanayakkara, A.; Challacombe,
M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.; Wong, M. W.; Andres,
J . L.; Replogle, E. S.; Gomperts, S.; Martin, R. L.; Fox, D. J .;
Binkley, J . S.; Defrees, D. J .; Baker, J .; Stewart, J . J . P.; Head-
Gordon, M.; Gonzalez, C.; People, J . A. GAUSSIAN94: Gaussian
Inc., Pittsbourgh, PA, 1995.
(21) (a) De Benedetti, P. G.; Cocchi, M.; Menziani, M. C.; Fanelli, F.
Theoretical Quantitative Structure-Activity Analysis and Phar-
macophore Modelling of Selective Non Congeneric R1A-Adren-
ergic Antagonists. J . Mol. Struct. (THEOCHEM) 1993, 280,
283-290. (b) Cocchi, M.; Menziani, M. C.; Fanelli, F.; De
Benedetti, P. G. Theoretical Quantitative Structure-Activity
Relationship Analysis of Congeneric and Non Concgeneric R1-
Adrenoceptor Antagonists. A Chemometric Study. J . Mol. Struct.
(THEOCHEM), 1995, 331, 79-93.
(3) Bourguignon, J . J . Endogenous and Synthetic Ligands of Mito-
chondrial Benzodiazepine Receptors: Structure-Affinity Rela-
tionships. In Peripheral Benzodiazepine Receptors; Giesen-
Crouse, E., Ed.; Academic Press: London, 1993; pp 59-85.
(4) Georges, G.; Vercauteren, D. P.; Vanderveken, D. J .; Horion, R.;
Evrard, G.; Fripiat, J . G.; Andre, J . M.; Durant, F. Structural
and Electronic Analysis of Peripheral Benzodiazepine Ligands:
Description of the Pharmacophoric Elements for Their Receptor.
Int. J . Quant. Chem. Quant. Biol. Symp. 1990, 17, 1-25.
(5) Sprengel, R.; Werner, P.; Seeburg, P. H.; Mukhin, A. G.; Santi,
M. R.; Grayson, D. R.; Guidotti, A.; Krueger, K. E. Molecular
Cloning and Expression of cDNA Encoding a Peripheral-Type
Benzodiazepine Receptor. J . Biol. Chem. 1989, 264, 20415-
20421.
(6) (a) Farges, R.; J oseph-Liauzun, E.; Shire, D.; Caput, D.; Le Fur,
G.; Ferrara, P. Site-Directed Mutagenesis of the Peripheral
Benzodiazepine Receptor: Identification of Amino Acids Impli-
cated in the Binding Site of Ro5-4864. Mol. Pharmacol. 1994,
46, 1160-1167. (b) Bernassau, J . M.; Reversat, J . L.; Ferrara,
P.; Caput, D.; Le Fur, G. A 3D Model of the Peripheral
Benzodiazepine Receptor and its Implication in Intra Mitochon-
drial Cholesterol Transport. J . Mol. Graphics 1993, 11, 236-
244.
(7) Marshall, G. R. In Medicinal Chemistry VI, Proceedings of 6th
International Symposium on Medicinal Chemistry; Brighton,
U.K., September 4-7, 1978; Simkins, M. A., Ed.; Costwold
Press: Oxford, 1979; pp 225-235.
(8) Karelson, M.; Lobanov, V. S.; Katritzky, A. R. Quantum Chemi-
cal Descriptors in QSAR/QSPR Studies. Chem. Rev. 1996, 96,
1027-1043
(9) De Benedetti, P. G.; Cocchi, M.; Menziani, M. C.; Fanelli, F.
Theoretical Quantitative Size and Shape Activity Selectivity
Analyses of 5-HT1A Serotonin and R1-Adrenergic Receptor
Ligands. J . Mol. Struct. (THEOCHEM) 1994, 305, 101-110.
(10) Anzini, M.; Cappelli, A.; Vomero, S. Synthesis of 2-Substituted-
2,3-dihydro-9-phenyl-1H-pyrrolo[3,4-b]quinolin-3-ones as Poten-
tial Peripheral Benzodiazepine-receptor Ligands. Heterocycles
1994, 38, 103-111.
(11) Cappelli, A.; Anzini, M.; Vomero, S. Synthesis of 4-Substituted-
1,2,3,4-tetrahydro-11-phenyl-5H-azepino[3,4-b]quinolin-5-one as
Potential Peripheral Benzodiazepine-receptor Ligands. Hetero-
cycles 1994, 38, 1265-1272
(12) Marsili, A.; Ricci, P. Trasformazione di Indoni in Derivati della
Chinolina e della Isochinolina. Nota II. Un Caso di Reazione di
Schmidt Anomala. (Transformation of Indenones into Quinoline
and Isoquinoline Derivatives. II. A Case of Anomalous Schmidt
Reaction.) Ann. Chim. 1962, 52, 112-120.
(13) Echavarren, A. M.; Stille, J . K. Palladium-Catalyzed Coupling
of Aryl Triflates with Organostannanes. J . Am. Chem. Soc. 1987,
109, 5478-5486.
(14) (a) Chen, Q.-Y.; Yang, Z.-Y. Palladium-Catalyzed Reaction of
Phenyl Fluoroalkanesulfonates with Alkynes and Alkenes.
Tetrahedron Lett. 1986, 27, 1171-1174. (b) Tilley, J . W.;
Zawoiski, S. A Convenient Palladium-Catalyzed Coupling Ap-
proach to 2,5-Disubstituted Pyridines. J . Org. Chem. 1988, 53,
386-390. (c) Walser, A.; Flynn, T.; Mason, C.; Crowley, H.;
Maresca, C.; O’Donnell, M. Thienotriazolodiazepines as Platelet-
Activating Factor Antagonists. Steric Limitations for the Sub-
stituent in Position 2. J . Med. Chem. 1991, 34, 1440-1446.
(15) Schoenberg, A.; Heck, R. F. Palladium-Catalyzed Amidation of
Aryl, Heterocyclic, and Vinylic Halides. J . Org. Chem. 1974, 39,
3327-3331.
(22) QUANTA/CHARMm, 1995 Molecular Simulations Inc., 16 New
England Executive Park, Burlington, MA 01803-5297.
(23) SHELXTL PC, Siemens Analytical X-Ray Instruments, Inc.,
Madison, WI; Rel. 4.1, 1990.
(24) Flack, H. D.; Schwarzenbach, D. On the Use of Least-Squares
Restraints for Origin Fixing in Polar Space Groups. Acta
Crystallogr. 1988, A44, 499-506.
(25) International Tables for Crystallography; Dordrecht: Kluwer
Academic Publishers: Dordrecht, 1992; Vol. C.
(26) Sheldrick, G. M. SHELXL-93, University of Goettingen, 1993.
(27) Cantoni, L.; Rizzardini, M.; Skorupska, M.; Cagnotto, A.; Cat-
egoni, A.; Pecora, N.; Frigo, L.; Ferrarese, C.; Mennini, T. Hepatic
Protoporphyria is Associated with a Decrease in Ligand Binding
for Mitochondrial Benzodiazepine Receptors in the Liver. Bio-
chem. Pharmacol. 1992, 44, 1159-1164.
(28) Munson, P. G.; Roadbard, D. Computers in Endocrinology; Raven
Press: New York, 1984; pp 117-145.
J M960516M