Synthesis, biological evaluation, and quantitative receptor docking simulations of 2-[(acylamino)ethyl]-1,4-benzodiazepines as novel tifluadom- like ligands with high affinity and selectivity for κ-opioid receptors

Article abstract of DOI:10.1021/jm950423p

The synthesis and biological evaluation of a series of 2-substitued 5- phenyl-1,4-benzodiazepines, structurally related to tifluadom (5), the only benzodiazepine that acts simultaneously as a κ-opioid agonist and a cholecystokinin-A (CCK-A) antagonist, ar

Full text of DOI:10.1021/jm950423p

860
J . Med. Chem. 1996, 39, 860-872
Syn th esis, Biologica l Eva lu a tion , a n d Qu a n tita tive Recep tor Dock in g
Sim u la tion s of 2-[(Acyla m in o)eth yl]-1,4-ben zod ia zep in es a s Novel
Tiflu a d om -lik e Liga n d s w ith High Affin ity a n d Selectivity for K-Op ioid
Recep tor s1
Andrea Cappelli,‡ Maurizio Anzini,*,‡ Salvatore Vomero,‡ M. Cristina Menziani,† Pier G. De Benedetti,†
Massimo Sbacchi,§ Geoffrey D. Clarke,§ and Laura Mennuni|
Dipartimento Farmaco Chimico Tecnologico, Universita´ di Siena, Banchi di Sotto 55, 53100 Siena, Italy, Dipartimento di
Chimica, Universita´ degli Studi di Modena, Via Campi 183, 41100 Modena, Italy, SmithKline Beecham Farmaceutici S.p.A.,
Via Zambeletti, Baranzate di Bollate, 20021 Milano, Italy, and Rotta Research Laboratorium S.p.A., Via Valosa di Sopra 7,
20052 Monza, Italy
Received J une 8, 1995X
The synthesis and biological evaluation of a series of 2-substitued 5-phenyl-1,4-benzodiazepines,
structurally related to tifluadom (5), the only benzodiazepine that acts simultaneously as a
κ-opioid agonist and a cholecystokinin-A (CCK-A) antagonist, are reported. The radioligand
binding models used in these studies were [125I](BH)-CCK-8 in rat pancreas (CCK-A), [3H]-
(MeNLE28,31)-CCK-8 in guinea pig cerebral cortex (CCK-B), and [3H]U-69593 (κ1), [3H]DAMGO
(µ), and [3H]DADLE (δ) in guinea pig brain. All the title compounds were devoid of significant
affinity for both CCK-A and CCK-B receptors, while some of them bound with nanomolar affinity
and high selectivity for κ-opioid receptors. In particular, the 2-thienyl derivative 7a (X ) H)
with a Ki ) 0.50 nM represents a clear improvement with respect to tifluadom, showing a
comparable potency but higher selectivity. The application of computational simulations and
linear regression analysis techniques to the complexes between guinea pig κ (κ1)-receptor and
the title compounds allowed the identification of the structural determinants for recognition
and quantitative elucidation of the structure-affinity relationships in this class of receptors.
In tr od u ction
subtypes of the κ-receptor (κ1, κ2, and κ3) described in
the literature.15 However, of these receptor subtypes,
only one, with the characteristics of a κ1-receptor
[specific ligand, the arylacetamide U-69,593 (2b)16], has
been isolated, cloned, and expressed from a variety of
species17-19 including man20 and, like the other opioid
receptors belongs to the G-protein coupled receptors
(GPCRs) superfamily, appears to be a classical seven-
transmembrane receptor.21
Tifluadom (5)22 is a 2-[(acylamino)methyl]-1,4-benzo-
diazepine with negligible effects upon classical central
benzodiazepine receptors,23 which also shows high af-
finity for the κ-opioid receptor24 and causes antinocicep-
tion in several animal models.25 Additional activities
noted with tifluadom include locomotor disturbances,26
diuresis,27 anti-inflammatory activity,28 and effects on
food consumption:29 all of these effects are thought to
be associated with activity at κ-opioid receptor. Tiflua-
dom, however, also has significant affinity for the µ- and
δ-opioid receptors,30 and activities at these sites may
serve to modulate the κ-related effects in vivo.31 Fur-
thermore, it has been shown that tifluadom is also a
cholecystokinin (CCK) antagonist,32 and several studies
have indicated that agonists and antagonists at the
CCK receptors can interact with the actions of opioids.33
For more than a decade, the search has continued for
a highly selective κ-opioid agonist with clinical potential
as a safe and effective analgesic.2,3 Most of the κ-ago-
nists synthesized during this period were derived from
4
the prototype selective κ-agonist U-50,488 (1), which
is an arylacetamide compound. Successive modifica-
tions of this structure led to U-62,066 (spiradoline) (2a ),5
CI 977 (enadoline) (2c), 6 GR 89696 (3), 7 and BRL 52656
(4).8 All these compounds are selective for the κ-opioid
receptor and are potent in a range of antinociceptive
models in animals;9 where tested, these compounds have
also produced analgesia in man.10
Administration of these compounds, however, is also
associated with the appearance of a spectrum of side
effects including diuresis, sedation/locomotor incoordi-
nation, and an array of central nervous system (CNS)
effects.11,12 The side-effect profile of the above com-
pounds is almost certainly related to their selective
interaction with κ-receptors. All feature, however, the
same pharmacophore sequence [N-C-C-N (sp2)]13 which
may be important in this respect, since it has been
hypothesized that, even at a single receptor, the chemi-
cal structure of an agonist may determine which second-
messenger response pathways are activated and hence
which biological response is apparent (see review by
Kenakin).14 Another possibility is that the diverse
biological activities of existing κ-opioid agonists repre-
sent the functional correlates of activities at the various
There are several reports in the literature of medici-
nal chemistry effort to separate the various opioid and
CCK components in tifluadom analogues.34 Bock et al.35
and Meurisse et al.36 prepared a series of tifluadom
derivatives to try to clarify the relationship between
affinity for the CCK receptors and chemical structure.
In the former study the influence of the molecular
flexibility of the compounds on the κ-opioid and CCK
receptors was also examined. The present study was
† Universita´ degli Studi di Modena.
‡ Universita´ di Siena.
§ SmithKline Beecham Farmaceutici S.p.A.
| Rotta Research Laboratorium S.p.A.
X Abstract published in Advance ACS Abstracts, J anuary 15, 1996.
0022-2623/96/1839-0860$12.00/0 © 1996 American Chemical Society

[(Acylamino)ethyl]benzodiazepines as Tifluadom-Like Ligands
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 4 861
Ch a r t 1
Sch em e 1a
a
(i) SOCl2, CH2Cl2; (ii) potassium phthalimide, NaI, DMF; (iii)
b
NH2NH2, MeOH; (iv) RCOCl, NEt3, THF. a , X ) H; b, X ) Cl.
µ- and δ-opioid receptors was also evaluated. The
SAFIR in the very limited series of the title compounds
are briefly discussed. Furthermore, in this study the
heuristic direct approach to quantitative structure-
activity relationship (QSAR) analysis38,39 has been ap-
plied to the three-dimensional models of the complexes
between the κ1-receptors and the newly synthesized
ligands in order to rationalize, quantitatively, through
the combination of molecular modeling, computational
simulations, and correlation analysis, their structure-
affinity features.
Ch em istr y
Scheme 1 summarizes the synthetic procedures used
to obtain the 2-[(acylamino)ethyl]-1,4-benzodiazepines
considered in this study. The title compounds were
prepared starting from the appropriate 2-hydroxyethyl
derivative 9a ,b. By treatment with thionyl chloride in
methylene chloride at room temperature, compounds
9a ,b were converted into the respective chloroethyl
derivatives 10a ,b which, via the classical Gabriel syn-
thesis, were transformed into the 2-(aminoethyl)-1,4-
benzodiazepines 12a ,b.40 These compounds were, in
turn, acylated by the suitable acyl chloride to afford the
expected 1,4-benzodiazepines 7a -c and 8a -e. While
the synthesis of compound 9b has been already re-
ported,41 compound 9a was prepared starting from the
1,4-benzodiazepin-2-ylidene acetic acid ethyl ester 4-ox-
ide (13) which we previously described42 (Scheme 2).
Briefly, by reduction with phosphorus trichloride, com-
pound 13 gave 14, which was in turn transformed into
its N1-methyl derivative 15 using methyl p-toluene-
sulfonate in a two-phase system (benzene/50% aqueous
sodium hydroxide), in the presence of tetrabutylammo-
nium bromide as a phase-transfer catalyst. The Z-
initiated to ascertain whether, using tifluadom as a
starting point, it would be possible to prepare selective
κ1-agonists which do not feature the [N-C-C-N (sp2)]
pharmacophore of existing compounds: such compounds
might ultimately gain insight into the role of the
chemical structure of a selective agonist on its biological
effects in vitro and in vivo.
The interpretation of the structure-affinity relation-
ship (SAFIR) data generated (for details, see Results
and Discussion section) prompted the design of com-
pounds 7 and 8 as receptor probes able to discriminate
κ-opioid from CCK receptors. Although it is generally
accepted that the constraint of molecules of various
chemical structures having a nonselective binding pro-
file confers selectivity for a particular receptor,37 the
title compounds which possess further degrees of con-
formational freedom represented more flexible tifluadom
analogues showing higher selectivity with respect to the
parent compound.
1
geometry of 13 was assigned on the basis of its H-NMR
spectrum42 only, while the E/ Z-geometry at the exocy-
clic double bond of compounds 14 and 15 was assigned
on the basis of their 1H-NMR spectra and 2D-NOESY
experiments. The 2D-NOESY spectrum of 14 showed
a nuclear Overhauser effect (NOE) between the olefinic
In the present study, the influence of the increasing
degree of molecular flexibility on binding affinities for

862 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 4
Cappelli et al.
Ta ble 1. Physicochemical Data of Compounds 7-16
Sch em e 2a
MS,b
molecular
ion
yield,a
%
compd
formula
mp, °C
anal.c
7a
7b
7c
8a
8b
8c
8d
8e
9a
10a
10b
11a
11b
12a
12b
14
C23H23N3OS
52
40
42
65
74
78
66
62
80
83
76
74
83
70
85
75
68
84
d
d
d
389
401
401
423
435
435
452
456
280
298
333
409
443
279
313
306
320
322
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
25H24FN3O
25H24FN3O
23H22ClN3OS
25H23ClFN3O
25H23ClFN3O
25H23Cl2N3O
27H25ClN4O
18H20N2O
18H19ClN2
18H18Cl2N2
26H23N3O2
26H22ClN3O2
18H21N3
85-86
62-63
79-80
82-83
138-139
d
d
d
d
65-66
d
d
122-123
d
d
18H20ClN3
19H18N2O2
20H20N2O2
20H22N2O2
15
16
a
b
Yields were not optimized. Determined by low-resolution
mass spectrometry; fragmentation patterns were consistent with
the expected structure. c Combustion analyses are within (0.4%
d
of the theoretical value. Oil.
of the 3-thienyl group of tifluadom with a 2-indolyl or
4-chlorophenyl moiety led to subnanomolar κ-ligands
provided with a CCK-A affinity (expressed as IC50)
ranging from 29 to 800 nM. The critical inspection of
the literature data about the influence of the 7-substitu-
tion of the benzodiazepine ring revealed that the pres-
ence of a chlorine atom depresses CCK-A affinity, while
up to now, no data concerning its importance on κ-opioid
receptors affinity were available. So, the designed
7-chloro substitution in the compounds of the present
series could lead to a better κ/CCK-A selectivity. On
the other hand, in a series of devazepide-like com-
pounds,47 2′-fluoro substitution appears in some in-
stances to have a positive effect on CCK-A affinity.
Thus, all these considerations allowed us to predict high
κ-selectivity versus CCK-A receptors for compounds 7
and 8. The results of receptor binding studies per-
formed on compounds 7 and 8, and summarized in Table
2, confirmed that our working hypothesis proved to be
correct. All the title compounds were completely devoid
of affinity for CCK receptors, while most of them bound
with high affinity and selectivity to the κ-opioid site. In
particular, 2-thienyl derivative 7a represents a clear
improvement with respect to tifluadom, showing a
comparable potency but higher κ/CCK and κ/µ selectiv-
ity.
a
(i) PCl3, CHCl3; (ii) TsOCH3, (t-Bu)4N+Br-, NaOH (50%), C6H6;
(iii) Et3SiH, TFA; (iv) LAH, ether.
proton (δ 4.89) and the C-3 methylene protons (δ 4.30),
supporting Z-geometry at the exocyclic olefin. In the
case of compound 15, a detectable NOE, between the
olefinic proton and the protons of the N1-methyl group,
allowed to assign the E-geometry to this compound.
Compounds 14 and 15 were isolated as single isomers.
Selective reduction of 15 with triethylsilane in trifluo-
roacetic acid yielded the saturated ester 16, which was
reduced with lithium aluminum hydride in dry ether
at room temperature to afford the expected 2-(hydroxy-
ethyl)-1,4-benzodiazepine 9a .
Biology
In vitro receptor binding assays were used to measure
affinities at κ-, µ-, and δ-opioid and CCK receptors.The
radioligand binding models used in these studies were
[
125I](BH)-CCK-8 in rat pancreas (CCK-A),43 [3H](MeN-
LE28,31)-CCK-8 in guinea pig cerebral cortex (CCK-
B),44,45 and [3H]U-69,593 (κ1),16 [3H]DAMGO (µ), and
[3H]DADLE (δ) in guinea pig brain tissues,46 respec-
tively. The results are reported in Table 2.
It should be noted that all the compounds in this
series are racemic, and therefore no conclusions can be
drawn concerning the relative contributions of the
corresponding enantiomers to the inhibition of [3H]U-
69,593 guinea pig brain binding. From the data re-
ported in Table 2, some SAFIR can be obtained about
the influence of the acyl substituent. In the series of
the acyl groups examined, negligible difference in the
κ-opioid affinities could be found, suggesting that, most
probably, the amidic functionality of the acylamino
moiety gives the main contribution to the binding
process. However, additional steric hindrance in the
zone corresponding to the para-position of the phenyl
moiety of compounds 8 is not well tolerated, and both
p-chlorophenyl and 2-indolyl derivatives 8d ,e (Chart 1)
resulted less active at κ-opioid receptors. The designed
Resu lts a n d Discu ssion
A careful study of the SAFIR performed on tiflua-
dom,35 devazepide (MK-329, formerly L-364,718) (6),
47
and related 1,4-benzodiazepines48 revealed that the
insertion of a methylene bridge between the 1,4-benzo-
diazepine nucleus and the exocyclic amidic nitrogen in
devazepide structure led to an increase of κ-opioid
affinity while dramatically decreasing the affinity for
CCK-A receptors. Further speculation about the dif-
ferent steric tolerance in the receptor areas correspond-
ing to position 2 or 3 of the 1,4-benzodiazepine core of
tifluadom-like compounds is possible. The replacement

[(Acylamino)ethyl]benzodiazepines as Tifluadom-Like Ligands
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 4 863
Ta ble 2. Receptor Binding Affinities for the 2-[(Acylamino)ethyl]-1,4-benzodiazepines 7 and 8
Ki, nMa
κ1,
µ,
δ,
CCK-A,
CCK-B,
compdb
[3H]U-65,693
[3H]DAMGO
[3H]DADLE
[
125I](BH)-CCK-8
[3H](MeNLE28,31)-CCK-8
5, tifluadom
7a
7b
7c
8a
8b
8c
8d
0.78 ( 0.19
0.50 ( 0.10
1.03 ( 0.42
0.56 ( 0.10
8.84 ( 1.47
17.9 ( 6.44
4.74 ( 1.33
35.0 ( 5.1
1.93 ( 0.08
6.42 ( 0.96
4.69 ( 0.46
2.26 ( 0.03
69.4 ( 15
153 ( 1.71
105 ( 48
>300
275 ( 43
>1000 (2)
>1000 (2)
>1000 (2)
>1000 (2)
>1000 (2)
47 ( 16c
NAd
>100 000c
13 000
NAe
NAd
NAe
NAe
NAd
17 000
62.1 ( 16
NAf
NAd
53.2 ( 1.33
g100 (2)
>100 (2)
NAf
NAd
NAf
NAd
8e
296 ( 37
NAf
NAd
lorglumide
6, devazepide
129 ( 19
3.69 ( 0.84
a
b
Ki values followed by SEM derived from three independent experiments (except where otherwise indicated). All compounds were
tested as racemates. c Data drawn from ref 32. d-f NA: compound which did not reach the half-maximal displacement at the maximal
concentration tested, d30 µM, 3 µM, and 10 µM.
e
f
Ta ble 3. Energies and Geometric Description of the
Hydrogen-Bonding Network Established in the Minimized
Average Structure of the κ1-Receptor
7-chloro substitution has proved to be inappropriate
because it depresses κ -opioid affinity to a higher extent
than CCK-A affinity. Although, as pointed out above,
the high κ/CCK-A selectivity could be easily predicted
on the basis of available SAFIR, it was rather difficult
to rationalize them in terms of ligand-receptor interac-
tion in absence of three-dimensional models of the
receptors.
Com p u ta tion a l Sim u la tion s. 1. 3D Mod el of th e
K1-Recep tor . A comparison of the primary sequences
of the κ-, δ-, and µ-opioid receptors and of their hydro-
phobic plots50 was carried out, in order to identify the
seven putative transmembrane (TM) regions. The
alignment was facilitated by the occurrence of several
residues of the TM domains which are highly conserved
across the GPCR superfamily. Figure 1 shows the zones
selected as the TM portions of guinea pig κ (κ1)-, mouse
κ-, rat κ-, mouse δ-, rat δ-, and rat µ-receptors.19
Residues conserved in most of the GCPR superfamily
are in bold, while residues mainly conserved in the
peptide receptor subclass are boxed. The putative TM
regions of the CCK-A receptor sequence50 are also
reported in Figure 1 for comparison.
atom
atom
IE
IE(HB) D-H-A d(H-A)
residue type residue type (kcal/mol) (kcal/mol) (deg)
(Å)
Thr104 HG1 P r o219 O
Ser108 Thr216 HG1
Asn 118 HD22 Phe205 O
-7.06
-6.78
-6.61
-6.53
-7.15
-3.82
-3.98
-3.22
-3.97
-3.37
-2.37
-4.54
-2.64
-4.15
-3.59
-4.11
-3.92
-3.32
-4.16
-3.91
-3.88
-3.15
-3.26
-4.10
-4.12
-3.26
-4.06
169.64
171.50
171.29
167.19
162.00
155.60
165.34
175.70
176.92
164.78
176.06
174.21
173.47
153.56
173.24
172.34
167.63
169.29
174.89
168.15
169.29
178.23
2.00
1.93
1.97
1.92
1.99
2.06
2.04
2.49
1.91
1.97
1.93
2.03
2.05
2.19
1.97
1.97
2.03
1.93
1.92
1.93
1.94
1.98
O
Tyr107 HH
Ala709
O
Leu 207 O
Thr314 HG1
Asp 211 OD1 Asn311 HD22 -4.72
Asp 211 OD2 Ser716 HG
Cys301 HT1 Ile421
-7.59
-5.06
-7.35
-5.32
-7.05
-8.12
-6.34
O
Tyr309 OH
Tyr309 HH
Ser414 HG
Ser418 OG
Thr320 HG1 Asn405 OD1
Asp 308 OD1 Tyr 712 HH
Asn311 HD21 Ser 715 OG
Ser 315
Leu319
Ser 412 HG
O
O
Asn 718 HD22 -5.80
Tyr722 HH
Phe506 O
-7.14
-7.75
-7.83
-5.62
-9.37
-6.09
Phe504
Ala507
Tyr 519 HH
O
O
His617 HE2
Tr p 613 HE1
Leu603 O
Val 622 HCT Ser703 OG
P h e609 O
Asn 714 HD22 -7.22
Ser702 HG -7.82
Phe619
O
a
The residues which are conserved in most of the GPCRs are
in bold.
The seven TM domains were constructed as right-
handed R-helices with φ ) -59° and ψ ) -44°. The
backbone coordinates of bacteriorhodopsin (Brookhaven
Protein Data Bank, code 1BRD)51 were used as initial
templates for the packing of the bundle. The helices
were rotated around their principal axes in order to
orient the maximum number of charged, polar, and
conserved residues toward the inside of the receptor,
according to the helical wheel projection model proposed
by Baldwin.52 Uncertainty in the relative vertical
position of the helices was partially overcome by taking
into account data from site-directed mutagenesis ex-
periments on bovine rhodopsine53 and cationic neu-
rotransmitter or peptide receptors54,55 together with the
need for an efficient side-chain packing.56,57
The structure of the receptor model obtained was
energy minimized to relieve steric conflicts of the amino
acid side chains. The minimized coordinates of the
receptor structure were then used as starting point for
a simulation which combines a sequence of quenched
molecular dynamics (T ) 0->300 K) and a 80 ps
molecular dynamics (MD) run at 300 K. In this way
the search for conformational space can be expanded
in a limited fashion and the conformational behavior of
the conserved amino acids, which are invoked to play a
role in the receptor folding and function, can be inves-
tigated during the simulation period of the dynamics
run. Table 3 lists the amino acids that participate in
the complex network of interhelix interactions estab-
lished during dynamics. The energy contributions of the
interacting amino acids (IE), the hydrogen-bonding
components of the IE [IE(HB)], the hydrogen-acceptor
(H-A) distances, and the donor-acceptor angles (D-
H-A) of each hydrogen bond are computed over the
minimized average structure of κ1-receptor.
Most of the residues present in virtually all the
GPCRs are involved in this network, as expected. They
are reported in bold in Table 3. Among the other
residues, the dynamics behavior of the ones specifically
conserved in the subclass of opioid receptors or in the
κ-receptor subtypes is particularly important as they
might play a crucial role in ligand recognition and
binding, in a direct or indirect manner. Thr104, Thr320,
and Ser716 (the first digit corresponds to the helix, the
next two digits indicate the position of the residue in
the helix), which have the function of hydrogen bond
donors with respect to Pro219, Asn405, and Asp211,
respectively, are conserved only in the κ - and δ -receptor
subtypes. The µ-receptor subtype presents Met, Cys,

864 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 4
Cappelli et al.
F igu r e 2. General view of the κ1-receptor model. The amino
acids which constitute the binding site are highlighted.
among a number of peptide receptors (e.g., cannabinol,
adenosine, bombesin, endothelin, and tachykinin). Their
role in the ligand binding and receptor function of the
δ- and µ-receptor subtypes has been investigated by site-
directed mutagenesis studies.58,59 In agreement with
mutagenesis studies on other GPCRs,54,55 the data show
the fundamental importance of a negatively charged
residue at position 211 for efficient G-protein coupling,
the implication of Asp308 in direct ligand recognition,
and the contribution of His617 to the binding of both
agonist and antagonist ligands, although its implication
in the intrinsic activity of the receptor is not clear.
Chemical modifications of the κ-opioid receptor ex-
pressed on the mouse lymphoma cell line R1.1 indicate
that a sulfhydryl group or groups, but not a disulfide
bond, are present at or near the binding site.60 More-
over, the existence of a Cys residue in the ligand-binding
site is emphasised by the finding that a sulfhydryl
affinity labeling ligand for opioid receptor is able to
modify the receptor covalently.61 In our 3D-minimized
model of κ1-receptor subtype, this sulfhydryl group may
be identified as the Cys707 which points toward the
inside of the receptor pore at a minimal side-chain
distance to Asp308 of 3.6 Å. The same residue has been
proposed to participate in the binding of S-activated (-)-
6â-sulfhydryldihydromorphine with the µ-receptor in a
very recently published model building study of the
opioid receptor.63
F igu r e 1. Comparison of the transmembrane amino acid
sequences of opioid and CCK-A receptors. Residues conserved
in most of the GPCRs are in bold, and residues conserved in
the peptide receptor subclass are boxed.
2. K1-Ten t a t ive Bin d in g Dom a in a n d Dock in g
Exp er im en ts. On the basis of this information, a
tentative binding domain can be identified for the
nonpeptide κ1-ligands considered in this study. It
consists of the Asp308, His617, and Cys707 amino acid
residues and a large hydrophobic pocket containing
widely conserved residues: Trp409, Phe508, Pro511,
Phe609, and Trp613. The distance between the center
of the cavity and the Asp308 side chain is 5.5 Å. On
the contrary, several polar residues are clustered in a
pocket delimited by H1, H2, H3, and H7. Among these,
Tyr712 and Ser715 are conserved in the GPCR super-
family, and Asn311, Thr215, and Tyr107 are present
only in the opioid receptor subtypes. A general view of
the κ1-receptor model obtained is given in Figure 2,
where the amino acids which constitute the binding site
are highlighted.
and Cys in these positions. A conservative mutation is
found for the hydrogen bond donor Thr216 which is
exchanged with a Ser in the δ- and µ-receptor subtypes.
Ser414 and Ser702 act as hydrogen bond donors with
respect to Tyr309 and Phe619, while Ser418 and Ser703
accept hydrogen bonds from Tyr309 and Val622. These
serines are the most variable residues among the opioid
receptor subtypes. In fact, Ser414 is exchanged with a
Gly in the δ-receptor subtype and with an Ala in the
µ-receptor subtype; Ser418 substitutes a Pro residue,
which is widely conserved among the GPCR superfam-
ily. Ala is found in positions 702 and 703 in the
δ-subtype, while Ser702 is exchanged with a Val in the
µ-receptor subtype.
Three charged amino acid residues are predicted to
lie within the TM region of the opioid receptors. They
are Asp211 and Asp308, highly conserved within the
GCPR superfamily, and His617, which is conserved
Compounds 7a -c and 8a -c were used for docking
purposes together with tifluadom (5) and the higly

[(Acylamino)ethyl]benzodiazepines as Tifluadom-Like Ligands
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 4 865
Ta ble 4. Experimental Binding Constants and Computed
Energy Descriptors (kcal/mol) for the Ligand-κ1-Receptor
Interactions
d
d
compd
pKi
BE
IEk-L
EL
Ek
5a
1
9.10
9.04
8.75
9.30
8.99
9.25
8.05
7.75
8.32
-49.62
-49.63
-44.82
-51.72
-51.25
-50.57
-31.24
-21.73
-35.56
-73.06
-71.59
-74.94
-73.27
-74.94
-76.16
-70.86
-75.74
-80.95
6.42
5.46
8.93
4.84
5.71
6.53
5.66
8.62
6.90
20.09
16.50
21.19
16.71
17.98
19.02
33.96
45.43
38.49
2b
7a
7b
7c
8a
8b
8c
selective κ1-agonists U-50,488 (1) and U-69,593 (2b)
whose experimental binding affinity constants were
determined in our laboratories. The amine nitrogen
atom in compounds 1 and 2b and the imine nitrogen
atom in compounds 5, 7a -c, and 8a -c were considered
as protonated for docking purposes.
F igu r e 3. Correlation between the binding affinity values and
the calculated binding energies for the minimized ligand-
receptor complexes. The linear regression is pKi ) 6.56((0.16)
- 0.05((0.003)BE, n ) 9, r ) 0.98, s ) 0.11, where n is the
number of compounds, r is the correlation coefficient, s is the
standard deviation, and the values in parentheses give the 95%
confidence intervals.
In order to test the ability of the receptor model thus
obtained to reproduce the stereoselectivity observed for
the (-)-2S-enantiomer of tifluadom [(-)-(2S)-tifluadom
(5a ), IC50 ) 1.1 nM; (+)-(2R)-tifluadom (5b), IC50 ) 50
nM],36 the docking of the two enantiomers was carried
out, whereas 2S was assumed to be the most active
configuration of ligands 7a -c and 8a -c, as in the case
of tifluadom. The ligands were manually docked in the
minimized average structure of the κ1-receptor subtype
using as a main criterium the formation of a charge-
reinforced hydrogen bond between the protonated ni-
trogen N4 atom of the ligands 5, 7a -c, and 8a -c (Chart
1) and the Asp308. This can only be achieved at the
expense of disrupting the charge-reinforced hydrogen-
bonding interaction with the Tyr712 residue which
occurs in the free receptor. The rest of the ligands were
accommodated in the adjacent hydrophobic pocket in
such a way that the carbonylic oxygen of the substituent
and the HD1 atom of the His617 residue were found at
hydrogen-bonding distance, subject to the removal of the
interhelix hydrogen bond which stabilized the isolated
receptor.
This is mainly achieved by means of Pro15 in helix
6, positioned just below the putative active site, which
allows the kinking of the C-terminal part of the helix
away from the core of the receptor. The extent of this
structural change depends on the structural features
of the ligands and influences the overall packing of the
bundle. This is also shown by the data of Table 5, where
the amino acid residues of the κ1-receptor mainly
involved in the interactions with the ligands are listed
and the energy contributions of each residues are given.
A few important observations arise from a detailed
analysis of the output structures of the complexes. (a)
(-)-(2S)-Tifluadom performs the best interaction with
Asp308. It is the only ligand able to achieve a good
geometry for hydrogen-bonding interaction with both
the protonated nitrogen and amidic nitrogen atoms
simultaneously. In fact, the hydrogen bond contributors
to the Asp308 total IE is 6.8 kcal/mol, while it is only
3.0 and 3.89 kcal/mol for compounds 1 and 7a , respec-
tively. (b) In order to interact with the ligands, residues
Asn311, Trp613, His617, Tyr712, Asn714, and Ser715
induce slight modifications in the interhelix hydrogen-
bonding network observed in the isolated receptor.
Tyr712 gives dispersion interactions with the ligand
while it establishes new interhelix hydrogen-bonding
interactions with the Thr215 and Asn311 or Asp211
residues. The protonated state of His617 is able to
maintain the hydrogen bond with the backbone of H5
present in the isolated receptor and establishes at the
same time a hydrogen bond with the carbonyl oxygen
of the substituent side chain of the ligands 7a -c and
8a -c. Only dispersive forces are involved in the
interaction of this residue with the other ligands. (c)
The stereoselectivity of the κ1-receptor for the (-)-(2S)-
tifluadom (5a ) can be partially explained, in this model,
by the inability of the carbonylic oxygen of the 2-sub-
stituent in the R-enantiomer to achieve a correct spatial
orientation for the formation of a good hydrogen bond
with His617. Moreover, the lack of this specific interac-
tion cannot be compensated by dispersion interaction
3. En er gy Descr ip tion of th e K1-Bin d in g Site
Mod el a n d QSAR An a lysis. The energies for the
selected κ1-ligand complexes are reported in Table 4,
together with their experimental binding affinities (pKi).
The variation in the measured binding constants is very
well explained by the computed binding energies (BE)
which take into account the ligand-receptor interaction
energy (IEk-L) and the mutual conformational adjust-
ment of the ligand (EdL) and of the receptor (Edk) upon
binding. The quantitative relationship obtained be-
tween the binding affinity data values and the computed
binding energies for the ligands considered is shown in
Figure 3.
From the analysis of the data reported in Table 4
clearly emerges that the changes in the potential energy
of the receptor upon formation of the complexes are
mainly responsible for the correlation obtained, being
the changes in the interaction energies disorganized and
ranging in a smaller interval. This implies that repul-
sion forces play the most important role in discriminat-
ing between the compounds considered. Thus, from a
dynamic point of view, a reorganization of the receptor
architecture is needed in order to accommodate the
ligands.

866 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 4
Cappelli et al.
Ta ble 5. Energy Contribution (kcal/mol) to the Total IE of κ1-Receptor Residues Interacting with the Ligand Considered
compd Ile307 Asp308 Asn311 Met312 Ile417 Ile420 Trp613 Ile616 His617 Ile620 Cys707 Tyr712 Asn714 Ser715
5a
5b
1
2b
7a
7b
7c
8a
8b
8c
-2.34 -20.25
-2.29 -16.07
-1.67 -19.18
-16.18
-2.77 -14.80
-2.75 -15.21
-2.75 -15.56
-2.88 -15.07
-2.31 -14.10
-2.53 -15.12
-2.16
-0.56
-2.82
-3.59
-1.83
-2.00
-2.05
-1.52
-1.56
-1.93
-3.11
-1.77
-3.31
-3.05
-4.12
-4.53
-4.69
-5.33
-4.38
-4.54
-1.66 -1.59
-2.79 -1.29
-3.57 -2.87
-3.65 -2.05
-2.41 -2.06
-3.08 -2.63
-2.81 -2.35
-2.12 -1.89
-1.03 -3.61
-2.52 -4.04
-2.33
-1.15
-1.60
-6.58
-4.77
-2.22
-2.63
-2.95
-5.70
-5.75
-2.58 -6.34 -3.13
-3.15 -1.24 -1.64
-2.49 -6.11 -2.03
-2.81 -7.65 -2.92
-3.09 -7.07 -2.32
-2.68 -7.35 -1.97
-2.74 -7.54 -2.07
-3.84 -4.70 -2.57
-4.47 -5.56
-2.07
-1.50
-2.67
-2.37
-2.08
-1.84
-1.82
-1.66
-1.53
-1.97
-3.64
2.52
-3.28
-2.55
-3.94
-1.19
-2.62
-3.15
-3.37
-3.02
-2.82
-3.75
-3.74
-3.90
-2.60
-2.09
-3.87
-3.49
-3.61
-3.09
-3.76
-3.14
-1.82
-3.01
-3.29
-2.01
-2.01
-1.13
-2.10 -7.27 -2.02
F igu r e 4. Details of the interaction preferences between the κ1-binding site residues and the most relevant ligands: tifluadom,
7a , and U-50,488 (clockwise from top left).
between this residue and the thienyl ring. However,
the substantial difference in the total interaction ener-
gies of the two enantiomers (∆IE )15 kcal/mol) indicates
that a large reorganization of the receptor architecture
is needed in order to accommodate the R-enantiomer.
(d) Ile417, Ile420, and Ile620 are only present in the
κ1-subtypes. Although highly conservative substitutions
occur in these positions among the opioid receptor
subtypes (see Figure 1), they might contribute to recep-
tor selectivity affecting the local conformation of the
binding pocket. In fact, variability in the aliphatic side
chain of a single amino acid residue in the H6 domain
of the cholecystokinin receptor (from Val to Leu) was
proved to be entirely responsible for ligand selectivity.63
(e) The amino acids showed to be involved in the
interaction with these ligands are not conserved in the
CCK-A receptor (Figure 1). On this basis, different loci
for ligand-receptor interaction can be invoked in order
to explain the observed selectivity. Figure 4 shows a
detailed picture of the interaction preferences of the 1,
5, and 7a ligands with the κ1-binding site residues.
Common κ-recognition requirements for nonpeptide
and peptide ligands can be identified by comparing the
three-dimensional structure of the compounds consid-
ered with a low-energy conformation of the endogenous
peptide agonist. Better clues are provided when the
interacting conformations are considered.
Structural analysis of the endogenous opioid peptides
indicates that the N-terminal tetrapeptide sequence,
shared by dynorphins and enkephalins, is required for
agonist activity and binds to the message subsite of the
receptor which is located in the TM region. The
divergent C-terminal residues define the subtype speci-
ficity by binding to the address domain located on the
extracellular loops of the receptors.64,65 This view has,
very recently, been supported by the results obtained
from studies on chimeric µ/κ receptor.66 On this basis,
the N-terminal “message” fragment of dynorphin
[(COOCH3)-Leu-Phe-Gly-Gly-Tyr-NH3+] was modeled in
the κ1-receptor, and the conformation assumed in the
minimized complex is compared in Figure 5 with those
assumed by the most relevant compounds studied,
tifluadom (1), U-50,488 (2), and compound 7a . The
overlay of these ligands (Figure 6) helps to formulate a
peptidomimetic hypothesis which may be used to guide
future synthesis of 1,4-benzodiazepine derivatives with
improved characteristics.
The results obtained show the ability of the heuristic
direct QSAR approach to handle the available hetero-
geneous experimental information on receptors and

[(Acylamino)ethyl]benzodiazepines as Tifluadom-Like Ligands
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 4 867
plexes between the κ1-receptors and the newly synthe-
sized ligands allowed us to rationalize, quantitatively,
through the combination of molecular modeling, com-
putational simulations, and correlation analysis, their
structure-affinity features. Notwithstanding the lim-
ited set of compounds studied, the detailed picture of
the ligand-receptor interactions emerging from this
approach constitutes an important working hypothesis
to direct the design and synthesis of new congeneric
ligands with improved characteristics, which, in turn,
will be used to evaluate the capability and the limita-
tions of the model, enlarging, in this way, its local
reliability.
Exp er im en ta l Section
Melting points were determined in open capillaries on a
Bu¨chi 510 apparatus and are uncorrected. Microanalyses were
carried out using a Perkin-Elmer 240C elemental analyzer.
Merck silica gel 60, 70-230 mesh, was used for column
chromatography, and Riedel-de Haen DC-Mikrokarten SI F
37341 were used as TLC. Infrared spectra (IR) were recorded
in Nujol mulls with a Perkin-Elmer model 397 spectropho-
tometer. 1H-NMR spectra were recorded with a Bruker AC
200 spectrometer in the indicated solvents (TMS as internal
standard); the values of chemical shifts are expressed in ppm
and coupling constants (J ) in Hz. Mass spectra (EI, 70 eV)
were recorded on a VG 70-250S spectrometer. IR, NMR
spectra, and elemental analyses were performed in the Di-
partimento Farmaco Chimico Tecnologico, Universita` di Siena.
Mass spectra were performed by Centro di Analisi e Deter-
minazioni Strutturali, Universita` di Siena. Commercial chemi-
cals were used as obtained without further purification, except
for solvents, which were purified and dried, where appropriate,
before use by standard methods.
2,3-Dih yd r o-2-(2-ch lor oet h yl)-1-m et h yl-5-p h en yl-1H -
1,4-ben zod ia zep in e (10a ). To an ice-cooled solution of 2,3-
dihydro-2-(2-hydroxyethyl)-1-methyl-5-phenyl-1H-1,4-benzo-
diazepine, 9a (1.8 g, 6.2 mmol), in dry CH2Cl2 (40 mL) was
added thionyl chloride (2.2 mL, 30 mmol) dropwise. The ice
bath was removed, and the resulting mixture was stirred at
room temperature for 4 h. The solvent was evaporated under
reduced pressure, and thionyl chloride excess was removed by
azeotropic distillation with benzene. The residue was parti-
tioned between methylene chloride and saturated NaHCO3
solution. The organic phase was washed with a saturated
NaHCO3 solution (2 × 20 mL) and brine, dried (Na2SO4), and
concentrated under reduced pressure. The crude product was
purified by column chromatography (CH2Cl2-EtOAc, 95:5, v/v)
to give 1.54 g (83%) of 10a as a light yellow oil: 1H NMR
(CDCl3) δ 2.12-2.22 (m, 2 H), 2.88 (s, 3 H), 3.64 (t, 2 H, J )
6.9), 3.69-3.76 (m, 2 H), 3.88-3.99 (m, 1 H), 6.81-7.10 (m, 3
H), 7.30-7.48 (m, 4 H), 7.54-7.58 (m, 2 H). Anal. Calcd for
C18H19ClN2: C, H, N.
F igu r e 5. Conformation assumed in the minimized complexes
by the N-terminal pentapeptide sequence of dynorphin,
U-50,488, tifluadom, and 7a (clockwise from top left).
F igu r e 6. Superposition of the N-terminal pentapeptide
sequence of dynorphin (green), tifluadom (violet), U-50,488
(pink), and 7a (blue).
ligands and to summarize and translate them into
quantitative models. Moreover, this procedure consti-
tutes a powerful tool for the design of new leads based
on intermolecular interactions and for the suggestion
of site-directed mutagenesis studies, in order to give,
interactively, further support and improvement to the
predictive and interpretive aspects of the model.
Su m m a r y a n d Con clu sion s
The present study describes the synthesis of a series
of tifluadom analogues which feature the introduction
of a methylenic bridge between the 1,4-benzodiazepine
nucleus and the (acylamino)methyl group of the parent
compound.
7-Ch lor o-2,3-dih ydr o-2-(2-ch lor oeth yl)-1-m eth yl-5-ph en -
yl-1H-1,4-ben zod ia zep in e (10b). This compound was pre-
pared by the identical procedure described for 10a , starting
from 9b,41 and 1.57 g (76%) of 10b was isolated as an oil. An
analytical sample was obtained by column chromatography
(CH2Cl2-EtOAc, 95:5, v/v) as a light yellow oil: 1H NMR
(CDCl3) δ 2.11-2.22 (m, 2 H), 2.85 (s, 3 H), 3.63 (t, 2 H, J )
7.0), 3.69-3.74 (m, 2 H), 3.86-3.94 (m, 1 H), 6.93-7.0 (m, 2
H), 7.26-7.44 (m, 4 H), 7.52-7.57 (m, 2 H). Anal. Calcd for
C18H18Cl2N2: C, H, N.
2,3-Dih yd r o-2-(2-p h th a lim id oeth yl)-1-m eth yl-5-p h en -
yl-1H-1,4-ben zod ia zep in e (11a ). To a solution of 10a (1.19
g, 4 mmol) in dry DMF (30 mL) was added sodium iodide (0.6
g, 4 mmol), and the resulting mixture was stirred at room
temperature for 0.5 h. Potassium phthalimide (0.83 g, 4.5
mmol) was then added, and the reaction mixture was heated
at 70 °C for 14 h. After cooling the mixture was poured onto
crushed ice and extracted with CH2Cl2. The organic phase was
thoroughly washed with water to remove the DMF excess,
dried (Na2SO4) and concentrated under reduced pressure to
Affinity for the κ1-receptor was maintained with close
analogues of tifluadom which had the 2-thienyl, 2-fluo-
rophenyl, 4-fluorophenyl, and 2-chlorophenyl substitu-
ents. Strict derivatives of these compounds which had
a chlorine atom in position 7 of the benzodiazepine ring
showed, however, a marked loss of activity at the κ1-
receptor. Compared to tifluadom, however, all the
compounds synthesized in this study had lower affinity
for the CCK receptors; thus, they exhibited a greater
CCK-A/κ selectivity ratio. This was particularly evident
in compound 7a , the 2-[(2-thienylcarbonyl)amino]ethyl
derivative, which had a ratio of 60 000 compared to 60
for tifluadom [each value is the ratio of Ki value for the
CCK-A site over that for κ1-receptor (from data reported
in Table 2)].
The heuristic direct approach to QSAR analysis
applied to the three-dimensional models of the com-

868 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 4
Cappelli et al.
give 1.21 g (74%) of nearly pure 11a as a yellow oil which was
used as such in the next step. An analytical sample was
obtained by column chromatography (CH2Cl2-EtOAc, 90:10,
v/v) as a light yellow oil: 1H NMR (CDCl3) δ 1.90-2.18 (m, 2
H), 2.88 (s, 3 H), 3.65-3.98 (m, 5 H), 6.88-7.12 (m, 3 H), 7.25-
7.45 (m, 4 H), 7.52-7.63 (m, 2 H), 7.68-7.75 (m, 2 H), 7.78-
7.89 (m, 2 H). Anal. Calcd for C26H23N3O2: C, H, N.
7-Ch lor o-2,3-d ih yd r o-2-(2-p h th a lim id oeth yl)-1-m eth yl-
5-p h en yl-1H-1,4-ben zod ia zep in e (11b). This compound
was prepared with the same procedure used for 11a , starting
from 10b, and allowed to react for 20 h. When the reaction
mixture was poured onto crushed ice, nearly pure 11b pre-
cipitated as a yellow solid (1.47 g, 83%) which was filtered,
dried, and used with no further purification in the next step.
An analytical sample melting at 65-66 °C was obtained after
recrystallization from EtOAc-n-hexane as light yellow
needles: 1H NMR (CDCl3) δ 1.97-2.12 (m, 2 H), 2.88 (s, 3 H),
3.76-3.94 (m, 5 H), 6.97-7.02 (m, 2 H), 7.28-7.45 (m, 4 H),
7.53-7.58 (m, 2 H), 7.69 -7.75 (m, 2 H), 7.81-7.86 (m, 2 H).
Anal. Calcd for C26H22ClN3O2: C, H, N.
2,3-Dih yd r o-2-(2-a m in oet h yl)-1-m et h yl-5-p h en yl-1H -
1,4-ben zod ia zep in e (12a ). A suspension of the phthalimido
derivative 11a (1.0 g, 2.5 mmol) in MeOH (30 mL) treated with
98% hydrazine (1.21 mL, 25 mmol) was stirred at room
temperature for 4 h, diluted with MeOH (20 mL), and filtered.
The precipitate was discarded, and the filtrate was concen-
trated under reduced pressure. The residue was partitioned
between CH2Cl2 (50 mL) and ice-cold aqueous dilute hydro-
chloric acid (10 mL). The organic layer was discarded, and
the acidic aqueous solution was made alkaline with ice-cold 3
M sodium hydroxide solution. The reaction product was then
extracted with CH2Cl2 and the organic phase washed to
neutrality with brine and dried over MgSO4. Evaporation
under reduced pressure gave pure 12a (0.56 g, 80%) as a
colorless oil, which was used as such in the next reaction. An
analytical sample was obtained by column chromatography
(EtOAc-MeOH-Et3N, 90:10:10, v/v): 1H NMR (CDCl3) δ
1.69-2.10 (m, 4 H, two of which disappeared with D2O), 2.57-
3.0 (m, 5 H), 3.52-3.85 (m, 3 H) 6.84-7.0 (m, 3 H), 7.26-7.41
(m, 4 H), 7.52-7.57 (m, 2 H). Anal. Calcd for C18H21N3: C,
H, N.
7-Ch lor o-2,3-dih ydr o-2-(2-am in oeth yl)-1-m eth yl-5-ph en -
yl-1H-1,4-ben zod ia zep in e (12b). Starting from compound
11b and following the same procedure as described for 12a ,
amine 12b (0.66 g, 85%) was obtained as a yellowish oil and
used with no further purification in the next step. An
analytical sample was obtained by column chromatography
(EtOAc-MeOH-Et3N, 90:10:10, v/v) as a light yellow oil: 1H
NMR (CDCl3) δ 1.75-2.10 (m, 4 H, two of which disappeared
with D2O), 2.72-2.95 (m, 5 H), 3.54-3.85 (m, 3 H) 6.91-7.0
(m, 2 H), 7.28-7.56 (m, 6 H). Anal. Calcd for C18H20ClN3: C,
H, N.
Gen er a l P r oced u r e for t h e P r ep a r a t ion of 2-[(Acyl-
a m in o)eth yl]-1,4-ben zod ia zep in es 7 a n d 8. The 2-(ami-
noethyl)-1,4-benzodiazepine derivatives 12a ,b were acylated
as previously described.35 Briefly, a magnetically stirred
solution of the suitable amine (1.76 mmol) in dry THF (12 mL)
cooled at 0 °C was treated in succession with Et3N (245 µL)
and the appropriate acyl chloride (1.76 mmol) dissolved in dry
THF (3 mL). The ice bath was removed, and the reaction
mixture was warmed at 23 °C on overnight stirring and then
diluted with ethyl acetate (100 mL). The organic layer was
washed with a saturated NaHCO3 solution (2 × 30 mL) and
brine, dried (MgSO4), and concentrated under reduced pres-
sure to give the crude product, which after purification by
column chromatography (EtOAc-n-hexane, 70:30, v/v) af-
forded the corresponding acylamino derivative.
1H NMR (CDCl3) δ 2.02-2.14 (m, 2 H), 2.90 (s, 3 H), 3.35-
3.54 (m, 2 H), 3.57-3.91 (m, 3 H), 6.82 (d, 1 H, J ) 7.5), 6.95-
7.12 (m, 3 H), 7.19-7.70 (m, 6 H), 7.73-7.78 (m, 2 H), 7.97 (t,
1 H, J ) 5.4, disappeared with D2O). Anal. Calcd for
C23H23N3OS: C, H, N.
Compound 8a was obtained as a solid in 65% yield. Re-
crystallization from cyclohexane-EtOAc afforded an analytical
sample as light orange fluffy flakes melting at 85-86 °C: 1H
NMR (DMSO-d6) δ 1.72-2.07 (m, 2 H), 2.84 (s, 3 H), 3.30-
3.40 (m, 2 H), 3.61-3.82 (m, 3 H), 6.93 (d, 1 H, J ) 2.7), 7.11-
7.20 (m, 2 H), 7.45-7.54 (m, 6 H), 7.73-7.78 (m, 2 H), 8.54 (t,
1 H, J ) 5.4, disappeared with D2O). Anal. Calcd for C23H22
ClN3OS: C, H, N.
-
2,3-Dih yd r o-2-[[(2-flu or oben zoyl)a m in o]eth yl]-1-m eth -
yl-5-p h en yl-1H-1,4-ben zod ia zep in e (7b) a n d 7-Ch lor o-2,3-
d ih yd r o-2-[[(2-flu or ob en zoyl)a m in o]et h yl]-1-m et h yl-5-
p h en yl-1H -1,4-b en zod ia zep in e (8b ). Compounds 12a ,b
were acylated with 2-fluorobenzoyl chloride as reported above
to give 7b (yield 40%) and 8b (yield 74%), respectively. An
analytical sample of compound 7b was obtained by column
chromatography (EtOAc-n-hexane, 70:30, v/v) as an orange
oil: 1H NMR (CDCl3) δ 1.96-2.15 (m, 2 H), 2.86 (s, 3 H), 3.44-
3.58 (m, 2 H), 3.62-3.85 (m, 3 H), 6.84 (d, 1 H, J ) 3.6), 6.90-
7.65 (m, 12 H), 7.91 (t, 1H, J ) 6.2, disappeared with D2O).
Anal. Calcd for C25H24FN3O: C, H, N.
Compound 8b was obtained as a solid and recrystallization
from cyclohexane-CH2Cl2 afforded an analytical sample melt-
ing at 62-63 °C as yellowish flakes: 1H NMR (DMSO-d6) δ
1.71-2.12 (m, 2 H), 2.83 (s, 3 H), 3.29-3.42 (m, 2 H), 3.64-
3.83 (m, 3 H), 6.93 (d, 1 H, J ) 2.2), 7.14 (d, 1 H), 7.25-7.66
(m, 10 H), 8.39 (t, 1 H, J ) 6.1, disappeared with D2O). Anal.
Calcd for C25H23ClFN3O: C, H, N.
2,3-Dih yd r o-2-[[(4-flu or oben zoyl)a m in o]eth yl]-1-m eth -
yl-5-p h en yl-1H-1,4-ben zod ia zep in e (7c) a n d 7-Ch lor o-2,3-
d ih yd r o-2-[[(4-flu or ob en zoyl)a m in o]et h yl]-1-m et h yl-5-
p h en yl-1H -1,4-b en zod ia zep in e (8c). Compounds 12a ,b
were acylated with 4-fluorobenzoyl chloride as reported above
to give 7c (yield 42%) and 8c (yield 78%), respectively. An
analytical sample of compound 7c was obtained by column
chromatography (EtOAc-n-hexane, 70:30, v/v) as an orange
oil: 1H NMR (CDCl3) δ 1.78-2.13 (m, 2 H), 2.86 (s, 3 H), 3.05-
3.49 (m, 2 H), 3.52-3.81 (m, 2 H), 399-4.07 (m, 1 H), 6.81 (t,
1 H, J ) 7.4), 7.0-7.11 (m, 4 H), 7.31-7.58 (m, 6 H), 7.75 -7.88
(m, 2 H), 8.15 (t, 1H, J ) 5.4, disappeared with D2O). Anal.
Calcd for C25H24FN3O: C, H, N.
Compound 8c was obtained as a solid, and recrystallization
from cyclohexane-CH2Cl2 afforded an analytical sample melt-
ing at 79-80 °C as yellow flakes: 1H NMR (DMSO-d6) δ 1.74-
2.10 (m, 2 H), 2.83 (s, 3H), 3.32-3.42 (m, 2H), 3.63-3.80 (m,
3H), 6.93 (d, 1 H, J ) 2.1), 7.14 (d, 1 H, J ) 8.8), 7.32 (t, 2 H,
J ) 8.8), 7.40-7.54 (m, 6 H), 7.95 (q, 2 H), 8.56 (t, 1H, J )
5.3, disappeared with D2O). Anal. Calcd for C25H23ClFN3O:
C, H, N.
7-Ch lor o-2,3-dih ydr o-2-[[(4-ch lor oben zoyl)am in o]eth yl]-
1-m eth yl-5-p h en yl-1H-1,4-ben zod ia zep in e (8d ). This com-
pound, prepared by acylation of 12b with 4-chlorobenzoyl
chloride according to the reported procedure was obtained as
a solid in 66% yield. Recrystallization from cyclohexane-CH2-
Cl2 gave an analytical sample melting at 82-83 °C as yellow
flakes: 1H NMR (DMSO-d6) δ 1.74-2.06 (m, 2 H), 2.83 (s, 3
H), 3.32-3.46 (m, 2 H), 3.63-3.80 (m, 3 H), 6.93 (d, 1 H, J )
2.5), 7.13 (d, 1 H, J ) 8.7), 7.39-7.58 (m, 8 H), 7.89 (d, 2 H, J
) 8.7), 8.62 (t, 1 H, J ) 5.2, disappeared with D2O). Anal.
Calcd for C25H23Cl2N3O: C, H, N.
7-Ch lor o-2,3-d ih yd r o-2-[(1H-in d ol-2-yla m in o)eth yl]-1-
m eth yl-5-p h en yl-1H-1,4-ben zod ia zep in e (8e). By acyla-
tion of compound 12b with 2-indolylcarbonyl chloride,67 com-
pound 8e was obtained as a solid in 62% yield. Recrystallization
from EtOAc-n-hexane gave an analytical sample melting at
138-139 °C as light yellow needles: 1H NMR (DMSO-d6) δ
1.76-2.12 (m, 2 H), 2.84 (s, 3 H), 3.35-3.42 (m, 2 H), 3.70-
3.83 (m, 3 H), 6.93 (d, 1 H, J ) 2.8), 7.05-7.24 (m, 4 H), 7.39-
7.65 (m, 8 H), 8.53 (t, 1 H, J ) 5.3, disappeared with D2O),
2,3-Dih yd r o-2-[[(2-t h ie n ylca r b on yl)a m in o]e t h yl]-1-
m et h yl-5-p h en yl-1H -1,4-b en zod ia zep in e (7a ) a n d 7-
Ch lor o-2,3-d ih yd r o-2-[[(2-th ien ylca r bon yl)a m in o]eth yl]-
1-m eth yl-5-ph en yl-1H-1,4-ben zodiazepin e (8a). These com-
pounds were prepared according to the general procedure
starting from 12a ,b and using 2-thiophenecarbonyl chloride
as acylating agent. Compound 7a was obtained in 52% yield.
The purification by column chromatography (EtOAc-n-hex-
ane, 70:30, v/v) gave an analytical sample as an orange oil:
11.56 (s, 1H, disappeared with D2O). Anal. Calcd for C27H25
ClN4O: C, H, N.
-

[(Acylamino)ethyl]benzodiazepines as Tifluadom-Like Ligands
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 4 869
(Z)-2,3-Dih ydr o-2-[(eth oxycar bon yl)m eth ylen e]-5-ph en -
yl-1H-1,4-ben zod ia zep in e (14). A solution of 2,3-dihydro-
2-[(ethoxycarbonyl)methylene]-5-phenyl-1H-1,4-benzodiaz-
epine 4-oxide, 1342 (2.25 g, 7.0 mmol), in CHCl3 (70 mL)
containing phosphorus trichloride (6.72 g, 49.0 mmol) was
refluxed for 45 min while stirring. The solution was cooled,
the reaction quenched with ice, and the mixture basified with
a saturated Na2CO3 solution and then extracted with CH2Cl2.
The organic phase was washed to neutrality with brine, dried
(Na2SO4), and evaporated under reduced pressure to give a
gummy residue. The crude product was purified by column
chromatography (CHCl3) affording an oil (75%) which crystal-
lized on standing. After recrystallization from cyclohexane an
analytical sample melting at 122-123 °C was obtained as
extensively minced, and dispersed into 30 vol of Krebs-
Henseleit buffer (KHB) (118 mM NaCI, 25 mM NaHCO3, 4.7
mM KCI, 1.2 mM NaH2PO4, 0.1 mM CaCI2, 14 mM glucose,
0.1 mg/mL SBTI, 1% BSA) adjusted to pH 7.4. To this
medium, continuously shaken and gassed with 95% O2-5%
CO2, was added 0.5 mg/mL crude collagenase. The resulting
suspension was filtered with a nylon mesh (320 µm), layered
over KHB containing 4% BSA, 0.5 mM CaCI2, and 0.1 mg/mL
SBTI, and centrifuged for 5 min at 13200g. The final resulting
pellet was suspended in 10 vol of Hepes-Ringer buffer, pH 7.4
(118 mM NaCl, 10 mM Hepes, 1.13 mM MgCI2, 1.28 mM
CaCI2, 1% BSA, 0.2 mg/mL SBTI). This preparation gave a
final concentration of (1-3) × 107 cells/mL (assessed by light
microscope with the aid of a Neubauer chamber), which was
diluted in Hepes-Ringer buffer to about 5 × 106 cells/mL. Acini
(400 µL), [l25I](BH)-CCK-8 (25 pM), and displacing agents were
incubated in 0.5 mL total volume in polypropylene tubes for
30 min at 37 °C. At the end of incubation 1 mL of ice-cold
assay buffer was added, and the tubes were centrifuged for 5
min at 12500g. The supernatant was aspirated and the
radioactivity associated to the pellet measured in a Canberra
Packard B-5002 γ-counter (80% efficiency). Specific binding
was estimated as the difference between the [l25I](BH)-CCK-8
binding values found in the absence of 1 µM CCK-8 and those
in the presence.
CCK-B receptor binding assays were carried out on mem-
branes obtained from male albino guinea pig (Bettinardi,
Momo, Italy) cerebral cortex as previously described44 with
modifications. Tissue (about 1.5 mg) was homogenized in ice-
cold 10 mM Hepes (pH 7.4) and centrifuged twice at 4 °C for
15 min at 48000g. The final membrane pellet was suspended
in tissue buffer (10 mM Hepes, 118 mM NaCI, 4.7 mM KCI,
5.0 mM MgCl2, 1.0 mM EGTA, pH 7.4) to a concentration of
100 mg/mL (original wet weight). Protein concentration was
determined according to the method of Bradford.45 Cortical
membranes (0.35-0.4 mg of protein/tube), [3H](N-Me,NLe28,31)-
CCK-8 (0.5 nM), and displacing agents were incubated for 150
min at 25 °C. All the components, other than cortical
membrane suspension, were prepared in an assay buffer
consisting of 1 mg/mL BSA, 50 µM bestatin, and 0.1 mg/mL
bacitracin dissolved in tissue buffer. Specific binding was
estimated as the difference between the [3H](N-Me,NLe28,31)-
CCK-8 binding values in absence of 1 µM CCK-8 and those in
the presence. Bound radioligand was separated by rapid
filtration on glass fiber filters (GFB; Whatman), pretreated
for at least 1 h with 0.1% BSA solution, and washed three
times with 4.0 mL of ice-cold normal saline. Filters were
counted in a liquid scintillation spectrophotometer (55% ef-
ficiency), after adding 8 mL of Ultima Gold fluid (Canberra
Packard).
colorless prisms: IR 3200 (NH), 1635 (CO) cm-1 1H NMR
;
(CDCl3) δ 1.27 (t, 3 H, J ) 7.3), 4.15 (q, 2 H), 4.30 (bs, 2 H),
4.89 (s, 1 H), 6.98-7.51 (m, 9 H), 10.5 (s, 1 H, disappeared
with D2O). Anal. Calcd for C19H18N2O2: C, H, N.
(E)-2,3-Dih ydr o-2-[(eth oxycar bon yl)m eth ylen e]-1-m eth -
yl-5-p h en yl-1H-1,4-ben zod ia zep in e (15). To a well-stirred
solution of compound 14 (1.78 g, 5.8 mmol) in benzene (30 mL
50%) were added aqueous sodium hydroxide (30 mL), tetrabu-
tylammoniun bromide (0.386 g, 1.2 mmol), and methyl p-
toluenesulfonate (2.45 g,13.2 mmol) in sequence. The two-
phase system was refluxed under vigorous stirring for 90 min,
cooled, diluted with water (100 mL), and extracted with CH2-
Cl2 (3 × 50 mL). The organic layer, washed to neutrality with
brine, was dried over Na2SO4 and evaporated to dryness in
vacuo. The residue, purified by column chromatography (CH2-
Cl2-EtOAc, 95:5, v/v), gave an orange oil in 68% yield: IR 1640
-1
1
cm
(CO); H NMR (CDCl3) δ 1.27 (t, 3 H, J ) 7.1), 4.15 (q,
2 H), 4.30 (bs, 2 H), 4.89 (s, 1 H), 6.98-7.51 (m, 9 H), 10.5 (s,
1 H, disappeared with D2O). Anal. Calcd for C20H20N2O2: C,
H, N.
2,3-Dih ydr o-1-m eth yl-5-ph en yl-1H-1,4-ben zodiazepin e-
2-a cetic Acid Eth yl Ester (16). A solution of compound 15
(1.0 g, 3.1 mmol) in TFA (15 mL) was stirred under N2
atmosphere and cooled in an ice bath. The solution was
treated with triethylsilane (0.72 g, 6.2 mmol) added in a period
of 5 min. The ice bath was removed, and the stirring was
continued for 30 min. The mixture was then poured onto
crushed ice, made alkaline by addition of ammonia, and
extracted with CH2Cl2. The extracts were combined, dried
(Na2SO4), and evaporated under reduced pressure. The resi-
due was purified by column chromatography (CH2Cl2-EtOAc,
95:5, v/v), and 16 was obtained in 84% yield as a thick yellow
oil. Further purification by chromatography afforded an
analytical sample as a colorless oil: 1H NMR (CDCl3) δ 1.25
(t, 3 H, J ) 7.4), 2.46 (dd, 1 H, J ) 14.8, 7.7), 2.70 (dd, 1 H, J
) 15, 4.3), 2.86 (s, 3 H), 3.51 (dd, 1 H, J ) 11, 9.4), 3.87 (dd,
1 H, J ) 11, 4.2), 4.11 (q, 2 H), 4.29 (m, 1 H), 6.92-7.05 (m, 3
H), 7.26-7.40 (m, 4 H), 7.53-7.57 (m, 2 H). Anal. Calcd for
C20H22N2O2: C, H, N.
2,3-Dih yd r o-2-(2-h yd r oxyeth yl)-1-m eth yl-5-p h en yl-1H-
1,4-ben zod ia zep in e (9a ). A solution of compound 16 (2.7
mmol) in dry ether (25 mL) was added dropwise to a stirred
suspension of lithium aluminum hydride (5.7 mmol) in dry
ether (25 mL). After stirring under N2 atmosphere for 15 min,
the excess of reducing agent was decomposed by careful
addition of H2O (1 mL). The inorganic material was filtered
off and washed with ether. The filtrate was dried (Na2SO4)
and evaporated under reduced pressure. A light yellow oil was
obtained in 80% yield and used in the next step without further
purification. An analytical sample was obtained by column
chromatography using EtOAc as eluant: 1H NMR (DMSO-
d6) δ 1.54-1.68 (m, 1 H), 1.80-1.91 (m, 1 H), 2.76 (s, 3H),
3.44-3.62 (m, 3 H), 3.64-3.77 (m, 2 H), 4.48 (bs, 1 H,
disappeared with D2O), 6.91 (d, 2 H, J ) 4.5), 7.05 (d, 1 H, J
) 8.1), 7.32-7.50 (m, 6 H). Anal. Calcd for C18H20N2O: C,
H, N.
BSA (fraction V), soybean trypsin inhibitor (SBTI), and
crude collagenase A (from Clostridium hystoliticum) were
purchased from Boehringer Mannheim GmbH (West Ger-
many). [125I](BH)-CCK-8 (2000 Ci/mmol) was purchased from
Amersham Int. plc (Buchinghamshire, U.K.). [3H](MeNle28,31
)
CCK-8 (2000 Ci/mmol) was purchased from New England
Nuclear (Boston, MA); bacitracin was purchased from Sigma
(St. Louis, MO).
2. K1-, µ-, a n d δ-Op ioid Recep tor Bin d in g Stu d ies.
Male guinea pigs (Dunkin-Hartley; 300-350 g) were killed by
decapitation and the brains (without cerebella) rapidly re-
moved. The tissues were homogenized in 10 vol of TRIS-HCl
(50 mM, pH 7.4), using a PBI polytron (setting 5 for 15 s),
and centrifuged at 49000g for 10 min at 4 °C. The pellets
obtained were resuspended in 10 vol of TRIS buffer, incubated
at 37 °C for 45 min, and centrifuged at 49000g for 10 min.
The resultant pellets were finally resuspended in 100 vol of
TRIS buffer (0.5-0.6 mg of protein/mL), and 1.9 mL aliquots
were used for the assay. The tubes containing the homogenate
and tritiated and unlabeled compounds were incubated at 25
°C at the final volume of 2 mL.
In Vitr o Bin d in g Assa ys. 1. CCK Recep tor Bin d in g
Stu d ies. CCK-A receptor binding assays were carried out on
rat isolated pancreatic acini, prepared by enzymatic digestion
of pancreas obtained from male Sprague-Dawley (SD) rats
(Charles River, Calco, Italy) weighing 250-300 g and fasted
24 h before sacrifice.43 Briefly, tissue was quickly removed,
The following radioligands were used, [3H]U-69,593 (65 Ci/
mmol; Amersham, Italy) to bind κ1-sites,16 [3H]-[DAla2,MePhe4,
Gly-ol5]-enkephalin ([3H]DAMGO) (50.5 Ci/mmol; New
England Nuclear, Italy) to bind µ-sites, and [3H]-[DAla2, DLeu5]-
enkephalin ([3H]DADLE) (33.5 Ci/mmol; New England Nuclear,

870 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 4
Cappelli et al.
Italy) to label δ-sites, in the presence of 40 nM of unlabeled
DAMGO to prevent cross-reactivity with µ-receptors.46 The
nonspecific binding was determined in the presence of nalox-
one, 10 µM. The time necessary to reach equilibrium condi-
tions, previously determined with specific kinetic studies, was
40 min for DAMGO and DADLE and 50 min for U-69,593. The
bound ligand was separated from the free ligand by filtration
through Whatman GF/B filters using a Brandell cell harvester;
the radioactivity on the disc was measured by liquid scintil-
lation counting using a Camberra Packard 2500TR â-counter.
3. An a lysis of Bin d in g Da ta . The binding parameters
deriving from competition experiments (IC50 values) were
calculated by nonlinear regression analysis using the RS/1
software.68 The inhibition constants (Ki) were calculated from
IC50 values using the Cheng and Prusoff equation.69
receptor separately, giving as geometrical input the same
geometry they assume in each minimized complex.
Ack n ow led gm en t. We are grateful to Dr. F. Mak-
ovec (Rotta Research Laboratorium, Monza, Italy) for
his comments and support throughout the course of this
work. We also acknowledge the contributions of Dr. G.
Delle Fave and Dr. C. Anania (Dept. Gastroenterology,
Dept. of Internal Medicine 2, University “La Sapienza”,
Rome, Italy) for CCK preliminary testing and Dr. G.
Giorgi (CIADS, University of Siena, Italy) for the
recording of mass spectra. Financial support from CNR
and Italian MURST (40% and 60% funds) is acknowl-
edged.
Com p u ta tion a l Meth od ology. Model building, docking
experiments, energy minimization, and molecular dynamics
studies were carried out using the software package QUANTA,
version 3.3 (Molecular Simulations, Inc., 200 Fifth Ave.,
Waltham, MA 02154), implemented on a Hewlett-Packard 720
workstation.
The minimization procedure consisted of 50 steps of steepest
descent followed by a conjugate gradient minimization until
the rms gradient of the potential energy was <0.001 kcal/mol/
Å. The united atom force field parameters, a distance-
dependent dielectric term (ꢀ ) 4r), and a 12 Å nonbonded cutoff
were employed. During dynamics the lengths of the bonds
involving hydrogen atoms were constrained according to the
SHAKE algorithm, allowing an integration time step of 0.001
ps. Moreover, weak harmonic constraints (30 kJ /mol/Å) were
applied between the backbone oxygen atoms of residue i and
backbone nitrogen atom of residue i+4 by using the NOE
facility in the CHARMM program70 in order to maintain the
helical structure.
Crystallographic coordinates are available for tifluadom,71
U-69,593,72 and U-50,488.73 The other ligands were obtained
by modification of the atomic coordinates of (+)-(2S)-tifluadom.
Standard geometric parameters were used. The fully opti-
mized geometries of the ligands were obtained in the AM1
framework,74 by applying the molecular mechanics correction
to the amidic bonds. The N-terminal pentapeptide portion of
dynorphin was constructed in the sequence builder module of
QUANTA, and the minimized average conformation obtained
from a 50 ps dynamics run was used for docking purposes.
The ligands were manually docked in the minimized average
structure of the κ1-receptor subtype using as a main criterium
the formation of a charge-reinforced hydrogen bond between
the protonated nitrogen atom of the ligand and the Asp308.
In order to obtain optimal complementarity between each
ligand and the receptor cavity, several AM174 energetically
permitted conformations of the protonated ligands were tested
and, as a consequence, rotation of the side-chain torsional
angles of different residues forming the binding site were
necessary to remove bad steric contacts from each input
structure of the complexes. Several complexes were, therefore,
obtained for each ligand which were energy minimized by
means of molecular mechanics calculation. The choice of the
“best complex” among those obtained was performed on the
basis of the following criteria: (a) the most favorable energy
state for the complex achieved by favorable energy states of
the individual components (CHARMm energy70), (b) a reason-
able distorsion of the ligand upon binding, judged by compar-
ing the AM1 energy of the bound conformation of the ligand
to that of the global mimimun (a range of 6 kcal/mol was
allowed), and (c) consistency in the binding preferences among
the ligands considered.
Refer en ces
(1) Portions of this work were presented previously at the 8th
Camerino Noordwijkerhout Symposium Trends in Receptor
Research, Camerino, Italy, September 8-12, 1991, and at the
11th Convegno Nazionale Divisione di Chimica Farmaceutica
della Societa' Chimica Italiana, Bari, Italy, October 2-5, 1994.
(2) For review, see: Millan, M. J . κ -Opioid Receptor and Analgesia.
Trends Pharmacol. Sci. 1990, 11, 70-76.
(3) J affe, H. I.; Martin, W. R. Opioid Analgesics and Antagonists.
In The Pharmacologic Basis of Therapeutics, 8th ed.; Goodman
Gilman, A., Rall, T. W., Nies, A. S., Taylor, P., Eds.; Pergamon
Press: New York, 1990; pp 485-521.
(4) Szmuszkovicz, J .; Von Voightlander, P. F. Benzeneacetamide
Amines: Structurally Novel Non-µ-Opioids J . Med. Chem. 1982,
25, 1125-1126.
(5) Szmuskovicz, J . U.S. Pat. No. 443 130, 1985; German Pat. No.
3241933, 1985; Chem. Abstr. 1985, 103, 184969b.
(6) Hunter, J . C.; Leighton, G. E.; Meecham, K. G.; Boyle, S. J .;
Horwell, D. C.; Rees, D. C.; Hughes, J . Cl-977, a Novel and
Selective Agonist for the κ -Opioid Receptor. Br. J . Pharmacol.
1990, 101, 183-189. Halfpenny, P. R.; Horwell, D. C.; Hughes,
J .; Hunter, J . C.; Rees, D. C. J . Med. Chem. 1990, 33, 286-291.
(7) Naylor, A.; J udd, D. B.; Lloyd, J . E.; Scopes, D. I. C.; Hayes, A.
G.; Birch, P. J . A Potent New Class of κ-Receptor Agonist:
4-Substituted 1-(Arylacetyl)-2-[(dialkylamino)methyl]pipera-
zines. J . Med. Chem. 1993, 36, 2075-2083.
(8) Vecchietti, V.; Giordani, A.; Giardina, G.; Colle, R.; Clarke, G.
D. (2S)-1-(Arylacetyl)-2-(aminomethyl)piperidine Derivatives:
Novel Highly Selective κ-Opioid Analgesics. J . Med. Chem. 1991,
34, 397-403.
(9) Rees, D. C. Non-Peptide Ligands for Membrane-Bound Peptide
Receptors. Curr. Med. Chem. 1994, 1, 145-158. Hughes, J .;
Woodruff, G. N. Neuropeptides. Function and Clinical Applica-
tions Arzneim.-Forsch./ Drug Res. 1992, 42, 250-255.
(10) Dionne, R. A.; Dobbins, K. R.; Hargreaves, K. M. Evaluation of
a Kappa Opioid Agonist, Spiradoline, in Comparison to Morphine
and Placebo in the Local Surgery Model. Clin. Pharmacol. Ther.
1991, 49, 183 (PTS).
(11) Mucha, R. F.; Herz, A. Motivational Properties of Kappa and
Mu Opioid Receptor Agonists Studied with Place and Taste
Preference Conditioning. Psychopharmacology 1985, 86, 274-
280. Zaratin, P.; Petrone, G.; Pizzi, A.; Sbacchi, M.; Clarke, G.
D. The Antinociceptive and Behavioural Properties of BRL
53001A: a New Kappa Opioid Agonist. Pharmacol. Res. 1992,
23 (Suppl. 2), 269-270.
(12) Peters, G.; Gaylor, S. Human Central Nervous System (CNS)
Effects of a Selective Kappa Opioid Agonist. Clin. Pharmacol.
Ther. 1989, 45, 130 (PPF-5). Peters, G. R.; Ward, N. J .; Antal,
E. G.; Lai, P. Y.; DeMaar, E. W. Diuretic Actions in Man of a
Selective Kappa Opioid Agonist: U-62,066E. J . Pharmacol. Exp.
Ther. 1987, 240, 128-131. Reece, P. A.; Sedman, A. J .; Rose, S.;
Scott Wright, D.; Dawkins, R.; Rajagopalan, R. D. Diuretic
Effects, Pharmacokinetics, and Safety of a New Centrally Acting
Kappa-Opioid Agonist (CI-977) in Humans. J . Clin. Pharmacol.
1994, 34, 1126-1132.
(13) Giardina, G.; Clarke, G. D.; Dondio, G.; Petrone, G.; Sbacchi,
M.; Vecchietti, V. Selective κ-Opioid Agonists: Synthesis and
Structure-Activity Relationships of Piperidines Incorporating an
Oxo-Containing Group. J . Med. Chem. 1994, 37, 3482-3491.
(14) Kenakin, T. Agonist-Receptor Efficacy: Mechanisms of Efficacy
and Receptor Promiscuity. Trends Pharmacol. Sci. 1995, 16,
188-192.
(15) For review, see: Clark, J . A.; Liu, L.; Price, M.; Hersh, B.;
Edelson, M.; Pasternak, G. W. Kappa Opiate Receptor Multiplic-
ity: Evidence for Two U50,488-Sensitive κ1 Subtypes and a
Novel κ3 Subtype. J . Pharmacol. Exp. Ther. 1989, 251, 461-
468.
(16) Raynor, K.; Kong, H.; Chen, Y.; Yasuda, K.; Yu, L.; Bell, G. B.;
Reisine T. Pharmacological Characterization of the Cloned κ-,
δ- and µ-Opioid Receptors. Mol. Pharmacol. 1994, 45, 330-334.
Tiberi, M.; Magnan, J . Pharmacological Characterization of the
The binding energies (BE) of the ligand-receptor-minimized
complexes were computed according to the following equa-
tion:75 BE ) IEk-L + Ed + Edk, where IEk-L is the total
L
interaction energy between the receptor and the ligand, Ed
L
is the distorsion energy of the ligand calculated with respect
to the optimized energy of the free molecule, and Ed is a
k
measure of the conformational energy change in the receptor
induced by ligand binding. Ligand and receptor distorsion
energies were obtained by minimizing the ligand and the

[(Acylamino)ethyl]benzodiazepines as Tifluadom-Like Ligands
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 4 871
Binding of [3H]Bremazocine in Guinea-Pig Brain: Evidence for
Multiplicity of the κ-Opioid Receptors. Can. J . Physiol. Phar-
macol. 1989, 67, 1336-1344. Lahti, R. A.; Mickelson, M. M.;
McCall, J . M.; Von Voightlander, P. F. [3H]U-69653 a Highly
Selective Ligand for the Opioid κ-receptor. Eur. J . Pharmacol.
1985, 109, 281-284.
(35) Bock, M. G.; DiPardo, R. M.; Evans, B. E.; Rittle, K. E.; Whitter,
W. L.; Veber, D. F.; Freidinger, R. M.; Chang, R. S. L.; Chen, T.
B.; Lotti, V. J . Cholecystokinin-A Receptor Ligands Based on
the κ -Opioid Agonist Tifluadom. J . Med. Chem. 1990, 33, 450-
455.
(36) Meurisse, R. L.; Hevlen, L.; Leysen, J . E.; Meunier, J . Cl.; De
Ranter, C. Radioligand Binding Profile of Enantiomers of
Tifluadom Analogues for Opioid, CCK, and Central Benzodiaz-
epine Receptors: Versatility of the Benzodiazepine Skeleton.
Regul. Pept. 1994 (Suppl. 1), S247-S248.
(37) Schiller, P. W. Rational Design of Peptidomimetics: Structural
and Pharmacological Aspects. In Medicinal Chemistry for the
21th Century, Wermuth, C. G., Ed.; Blackwell Scientific Publi-
cations: Oxford, 1992; pp 215-232.
(17) Minami, M.; Toya, T.; Katao, Y.; Maekawa, K.; Nakamura, S.;
Onogi, T.; Kaneko, S.; Satoh, M. Cloning and Expression of a
cDNA for the Rat κ-Opioid Receptor. FEBS Lett. 1993, 329, 291-
295.
(18) Yasuda, K.; Raynor, K.; Kong, H.; Breder, C. D.; Takeda, J .;
Reisine, T.; Bell, G. l. Cloning and Functional Comparison of κ-
and δ-Opioid Receptors from Mouse Brain. Proc. Natl. Acad. Sci.
U.S.A. 1993, 90, 6736-6739.
(19) Xie, G. X.; Meng, F.; Mansour, A.; Thompson, R. C.; Hoversten,
M. T.; Goldstein, A.; Watson, S. J .; Akil, H. Primary Structure
and Functional Expression of a Guinea Pig κ -Opioid (dynorphin)
Receptor. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 3779-3783.
(20) Zhu, J .; Chen, C.; Xue, J . C.; Kunapuli, S.; DeRiel, J . K.; Liu-
Chen, L. Y. Cloning of a Human κ-Opioid Receptor from the
Brain. Life Sci. 1995, 56, 201-207. Mansson, E.; Hare, L.; Yang,
D. Isolation of a Human κ-Opioid Receptor cDNA from Placenta.
Biochem. Biophys. Res. Commun. 1994, 202, 1431-1437.
(21) Evans, C. J .; Keith, D. E., J r.; Morrison, H.; Magendzo, K.;
Edwards, R. H. Cloning of a Delta-Opioid Receptor by Functional
Expression. Science (Washington, D.C.) 1992, 258, 1952-1955.
Ueda, H.; Nozaki, M.; Satoh, M. Multiple Opioid Receptors and
GTP-Binding Protein. Comp. Biochem. Physiol. 1991, 98C, 157-
169.
(38) De Benedetti, P. G.; Menziani, M. C.; Fanelli, F.; Cocchi, M. The
Heuristic-Direct Approach to Quantitative Structure-Activity
Relationship Analysis. THEOCHEM 1993, 285, 147-153.
(39) Menziani, M. C.; Fanelli, F.; Cocchi, M.; De Benedetti, P. G. The
Heuristic-Direct Approach to Quantitative Structure-Activity
Relationship Analysis of Ligand G Protein-Coupled Receptor
Complexes. The Endothelin Receptors as a Case Study. In
Membrane Protein Models: Experiment, Theory and Speculation;
Findley, J ., Ed.; Bios Science Pub., in press.
(40) The aminoethyl derivatives, obtained as thick oily products, were
stable at
4 °C for several weeks, while the corresponding
hydrochloric salts resulted hygroscopic and deliquescent com-
pounds.
(41) Anzini, M.; Cappelli, A.; Vomero, S. Synthesis and Receptor
Binding Studies of 2-Functionalized 1,4-Benzodiazepine Deriva-
tives as Potential Metaclazepam-like Antianxiety Agents. Far-
maco 1993, 48, 897-905.
(42) Anzini, M.; Garofalo, A.; Vomero, S. Syntheses of Functionalized
Derivatives of Quinazolines and 1,4-Benzodiazepines. Hetero-
cycles 1989, 29, 1477-1487.
(43) Makovec, F.; Bani, M.; Chiste, R.; Revel, L.; Rovati, L. C.; Rovati,
L. A. Differentiation of Central and Peripheral Cholecystokinin
Receptors by New Glutaramic Acid Derivatives with Cholecys-
tokinin Antagonistic Activity. Arzneim. Forsch./ Drug Res. 1986,
36, 98-102.
(44) Knapp, R. J .; Vaughn, L. K.; Fang, S. N.; Bogert, C. L.;
Yamamura, M. S.; Hruby, V. J .; Yamamura, H. l. A New, Highly
Selective CCK-B Receptor Radioligand [3H][N-Methyl-NLe28,31]-
CCK26-33: Evidence for CCK-B Receptor Heterogeneity. J .
Pharmacol. Exp. Ther. 1990, 255, 1278-1286.
(45) Bradford, M. M. A Rapid and Sensitive Method for the Quan-
titation of Microgram Quantities of Protein Utilizing the Prin-
ciple of Protein-Dye Binding. Anal. Biochem. 1976, 72, 248-
254.
(46) Wang, Y. X.; Clarke, G. D.; Sbacchi, M.; Petrone, G.; Brooks, D.
P. Contribution of Alpha-2 Adrenoceptors to Kappa Opioid
Agonist-Induced Water Diuresis in the Rat. J . Pharmacol. Exp.
Ther. 1994, 270, 244-249.
(47) Evans, B. E.; Rittle, K. E.; Bock, M. G.; Di Pardo, R. M.;
Freidinger, R. M.; Whitter, W. L.; Gould, N. P.; Lundell, G. F.;
Homnick, C. F.; Veber, D. F.; Anderson, P. S.; Chang, R. S. L.;
Lotti, V. J .; Cerino, D. J .; Chen, T. B.; King, P. J .; Kunkel, K.
A.; Springer, J . P.; Hirshfield, J . Design of Nonpeptidal Ligands
for a Peptide Receptor: Cholecystokinin Antagonists. J . Med.
Chem. 1987, 30, 1229-1239.
(48) Bock, M. G.; Di Pardo, R. M.; Evans, B. E.; Rittle, K. E.; Veber,
D. F.; Freidinger, R. M.; Chang, R. S. L.; Lotti, V. J . Cholecys-
tokinin Antagonists: Synthesis and Biological Evaluation of
4-Substituted 4H-[1,2,4]Triazolo[4,3-a][1,4]benzodiazepines. J .
Med. Chem. 1988, 31, 176-181.
(49) Kyte, J .; Doolittle R. F. A Simple Method for Displaying the
Hydropathic Character of a Protein. J . Mol. Biol. 1982, 157,
105-132.
(50) de Weerth, A.; Pisegna, J . R.; Huppi, K.; Wank, S. A. Molecular
Cloning, Functional Expression and Chromosomal Localization
of the Human Cholecystokinin Type A Receptor. Biochem.
Biophys. Res. Commun. 1993, 194, 811-818.
(51) Henderson, R.; Baldwin, J .; Ceska, T. H.; Zemlin, F.; Beckmamn,
E.; Downing, K. Model of the Structure of Bacteriorhodopsin
Based on High Resolution Electron Cryomicroscopy. J . Mol. Biol.
1990, 213, 899-919.
(52) Baldwin, J . M. The Probable Arrangement of the Helices in G
Protein-Coupled Receptors. EMBO J . 1993, 12, 1693-1703.
(53) Tomoko, A.; Gorbind Korana, H. Mapping of the Amino Acids
in Membrane Embedded Helices that Interact with the Retinal
Chromophore in Bovine Rhodopsin. J . Biol. Chem. 1991, 266,
4269-4275.
(22) Romer, D.; Buscher, H. H.; Hill, R. C.; Maurer, R.; Petcher, T.
J .; Zeugner, H.; Benson, W.; Finner, E.; Milkowsky, W.; Thies,
P. W. An Opioid Benzodiazepine. Nature (London) 1982, 298,
759-760.
(23) Kley, H.; Scheidemantel, U.; Bering, B.; Muller, W. E. Reverse
Stereoselectivity of Opiate and Benzodiazepine Receptors for the
Opioid Benzodiazepine Tifluadom. Eur. J . Pharmacol. 1983, 87,
503-504. Romer, D.; Buscher, H. H.; Hill, R. C.; Maurer, R.;
Petcher, T. J .; Zeugner, H.; Benson, W.; Finner, E.; Milkowsky,
W.; Thies, P. W. Unexpected Opioid Activity in a Known Class
of Drug. Life Sci. 1982, 31, 1217-1220.
(24) Burkard, W. P. [3H]Tifluadom Binding in Guinea-Pig Brain
Membranes Eur. J . Pharmacol. 1984, 97, 337-338.
(25) Genovese, R. F.; Dykstra, L. A. Tifluadom's Effects Under
Electric Shock Titration and Tail-Immersion Procedures in
Squirrel Monkeys. Life Sci. 1986, 39, 1713-1719.
(26) Dykstra, L. A.; Gmerek, D. E.; Winger, G.; Woods, J . H. Kappa-
Opioids in Rhesus Monkeys. I. Diuresis, Sedation, Analgesia and
Discriminative Stimulus Effects. J . Pharmacol. Exp. Ther. 1987,
242, 413-420.
(27) Shearman, G. T.; Tolcsvai, L. Tifluadom-Induced Diuresis in
Rats. Evidence for an Opioid Receptor-Mediated Central Action.
Neuropharmacology 1986, 25, 853-856. Blackburn, T. P.;
Borkowski, K. R.; Friend, J .; Rance, M. J . On the Mechanism of
κ-Opioid-Induced Diuresis. Br. J . Pharmacol. 1986, 89, 593-
598.
(28) Carey, P.; Haworth, M. A.; Stone, J . R. M.; Heron, J .; Wayne,
A.; J amisson, A.; Russel, N. J . W. Involvement of Corticosterone
in the Anti-Inflammatory Activity of the Opioid κ-Agonists
Tifluadom and U50488. Br. J . Pharmacol. 1987, 92 (Proc. Suppl.
629P).
(29) Morley, J . E.; Levine, A. S.; Grace, M.; Kneip, J .; Zeugner, H.
The Effect of Opioid-Benzodiazepine Tifluadom on Ingestive
Behaviors. Eur. J . Pharmacol. 1983, 93, 265-269.
(30) Petrillo, P.; Amato, M.; Tavani, A. The interaction of the Two
Isomers of the Opioid Benzodiazepine Tifluadom with µ-, δ-, and
κ-Binding Sites and Their Analgesic and Intestinal Effects in
Rats. Neuropeptides 1985, 5, 403-406.
(31) Kamei, J .; Tanihara, H.; Kasuya, Y. Modulation of κ-Mediated
Antitussive Activity in Rats by a δ-Agonist. Res. Commun. Chem.
Pathol. Pharmacol. 1992, 76, 375-378. Vonvoigtlander, P. F.;
Lewis, R. A. Analgesic and Mechanistic Evaluation of Spirado-
line, a Potent Kappa Opioid. J . Pharmacol. Exp. Ther. 1988, 246,
459-462.
(32) Chang, R. S. L.; Lotti, V. J .; Chen, T. B.; Keegan, M. E.
Tifluadom, a κ -Opiate Agonist, Acts as a Peripheral Cholecys-
tokinin Receptor Antagonist. Neurosci. Lett. 1986, 72, 211-214.
(33) Dourish, C. T.; Hawley, D.; Iversen, S. D. Enhancement of
Morphine Analgesia and Prevention of Morphine Tolerance in
the Rat by the Cholecystokinin Antagonist L-364,718. Eur. J .
Pharmacol. 1988, 147, 469-472. Noble, F.; Derrien, M.; Roques,
B. P. Modulation of Opioid Antinociception by CCK at the
Supraspinal Level: Evidence for
a Regulatory Mechanism
(54) Strader, D. C.; Fong, T. M.; Tota, M. R.; Underwood, D.; Dixon,
R. A. F. Structure and Function of G Protein-Coupled Receptors.
Annu. Rev. Biochem. 1994, 63, 101-132.
(55) Wess, J . Mutational Analysis of Muscarinic Acetylcholine Recep-
tors: Structural Basis of Ligand/Receptor/G Protein Interactions.
Life Sci. 1993, 53, 1447-1463.
Between CCK and Enkephalin System in the Control of Pain.
Br. J . Pharmacol. 1993, 109, 1064-1070.
(34) Freidinger, R. M. Non-Peptide Ligands for Peptide Receptors.
In Medicinal Chemistry for the 21th
Century; Wermuth, C. G.,
Ed.; Blackwell Scientific Publications: Oxford, 1992; pp 233-
245.

872 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 4
Cappelli et al.
(56) Ballesteros, J . A.; Weinstein, H. Integrated Methods for the
Construction of Three Dimensional Models and Computational
Probing of Structure-Function Relations in G-Protein Coupled
Receptors. Methods Neurosci. 1994, in press.
(57) Fanelli, F.; Menziani, M. C.; Cocchi, M.; De Benedetti, P. G.
Comparative Molecular Dynamics Study on the Seven-Helix
Bundle Arrangement of G-Protein Coupled Receptors. THEO-
CHEM 1995, 333, 49-69.
(58) Kong, H.; Raynor, K.; Yasuda, K.; Moe, S. T.; Portoghese, P. S.;
Bell, G. l.; Reisine, T. A Single Residue, Aspartic Acid 95, in the
δ-Opioid Receptor Specifies Selective High Affinity Agonist
Binding. J . Biol. Chem. 1993, 268, 23055-23058.
(59) Surratt, C. K.; J ohnson, P. S.; Moriwaki, A.; Seidleck, B. K.;
Blaschak, C. J .; Wang, J . B.; Uhl, G. R. µ-Opiate Receptor.
Charged Transmembrane Domain Amino Acids are Critical for
Agonist Recognition and Intrinsic Activity. J . Biol. Chem. 1994,
269, 20548-20553.
Non-Peptide Ligands in the Cloned Rat κ -Opioid Receptor. J .
Biol. Chem. 1994, 48, 30195-30199.
(67) Prepared by chlorination of the commercially available corre-
sponding acid as reported by J ohnson, J . R.; Hasbrouck, R. B.;
Dutcher, J . D.; Bruce, W. F. The Structure of Certain Indole
Derivatives Related to Gliotoxin. J . Am. Chem. Soc. 1945, 67,
423-430.
(68) Baron, B. M.; Siegel, B. W.; Slone, A. L.; Harrison, B. L.;
Palfreyman, M. G.; Hurt, S. D. [3H]5,7-Dichlorokynurenic Acid,
a Novel Radioligand Labels NMDA Receptor-Associated Glycine
Binding Sites. Eur. J . Pharmacol. 1991, 206, 149-154.
(69) Cheng, Y. C.; Prusoff, W. H. Relationship Between the Inhibition
Constant (Ki) and the Concentration of Inhibitor Which Causes
50 Per Cent Inhibition (I50) of an Enzymatic Reaction. Biochem.
Pharmacol. 1973, 22, 3099-3108.
(70) Brooks, B. R.; Bruccoleri, R. E.; Olafson, B. D.; States, D. J .;
Swaminathan, S.; Karplus, M. CHARMM:
a Program for
(60) J oseph, D. B.; Bidlack, J . M. The κ-Opioid Receptor Expressed
on the Mouse Lymphoma Cell Line R1.1 Contains a Sulfhydryl
Group at the Binding Site. Eur. J . Pharmacol. 1994, 267, 1-6.
(61) Kanematsu, K.; Naito, R.; Shimohigashi, Y.; Ohno, M.; Ogasawara,
T.; Kurono, M.; Yagi, K. Design and Synthesis of an Opioid
Receptor Probe: Mode of Binding of S-Activated(-)-6â-Sulfhy-
drylhydromorphine with the SH Group in the µ-Opioid Receptor.
Chem. Pharm. Bull. 1990, 38, 1438-1440.
Macromolecular Energy Minimization and Dynamics Calcula-
tions. J . Comput. Chem. 1983, 4, 187-217.
(71) Petcher, T. J .; Widmer, A.; Maetzel, U.; Zeugner, H. Structures
of Tifluadom [5-(2-fluorophenyl)-1-methyl-2-(3-thenoyl-aminom-
ethyl)-2,3-dihydro-1H-1,4-benzodiazepine] hydrochloride, C22H21
-
FN3OS‚CI-, and of (+)-Tifluadom p-toluenesulphonate C22H21
FN3OS‚C7H703S-, and the Absolute Configuration of the Latter.
Acta Crystallogr. 1985, C41, 909-912.
(62) Fujii, I.; Nakamura, H.; Sagara, T.; Kanematsu, K. Molecular
Model Building of Multiple Opioid Receptor Subtypes. Med.
Chem. Res. 1994, 4, 424-431.
(72) Doi, M.; Ishida, T.; Inoue, M. Conformational Characteristics of
Opioid κ -Receptor Agonist: Crystal Structure of (5S,7S,8S)-
(-)-N-methyl-N-[7-(1-pyrrolidinyl)-1-oxaspiro[4.5]dec-8-yl]benze-
neacetamide (U69,593), and Conformational Comparison with
Some κ -Agonists. Chem. Pharm. Bull. 1990, 38, 1815-1818.
(73) Doi, M.; Ishida, T.; Inoue, M. Structure of κ -Agonist, U-50488.
Acta Crystallogr. 1990, C46, 676-678.
(63) Beinborn, M.; Lee, Y. M.; McBride, E. W.; Quinn, S. M.; Kopin,
A. S. A Single Amino Acid of the Cholecystokinin-B/Gastrin
Receptor Determines Specificity for Non-Peptide Antagonists.
Nature 1993, 362, 348-350.
(64) Portoghese, P. S. Bivalent Ligands and the Message-Address
Concept in the Design of Opioid Receptor Antagonists. TIPS
1989, 10, 230-235.
(65) Lin, C. E.; Takemori, A. E.; Portoghese, P. S. Synthesis and κ
-Opioid Antagonist Selectivity of a Norbinaltorphimine Conge-
ner. Identification of the Address Moiety Required for k -An-
tagonist Activity. J . Med. Chem. 1993, 36, 2412-2415.
(66) Xue, J . C.; Chen, C.; Zhu, J .; Kunapuli, S.; DeRiel, J . K.; Yu, L.;
Liu-Chen, L. Y. Differential Binding Domains of Peptide and
(74) Dewar, M. J . S.; Zoebisch, E. G.; Healeyand, E. F.; Stewart, J .
J . P. AM1:
a New General Purpose Quantum Mechanical
Molecula Model. J . Am. Chem. Soc. 1985, 107, 3902-3909.
(75) Menziani, M. C.; De Benedetti, P. G.; Gago, F.; Richards, W. G.
The Binding of Benzensulphonamides to Carbonic Anhidrase
Enzyme. A Molecular Mechanics Study and Quantitative Struc-
ture-Activity Relationships. J . Med. Chem. 1989, 32, 951-956.
J M950423P