Isoxazolo-[3,4-d]-pyridazin-7-(6H)-ones and their corresponding 4,5-disubstituted-3-(2H)-pyridazinone analogues as new substrates for α1-adrenoceptor selective antagonists: Synthesis, modeling, and binding studies

Article abstract of DOI:10.1016/S0968-0896(98)00056-X

A series of phenylpiperazinylalkyl moieties were attached to monocyclic or bicyclic substituted pyridazinones and the new compounds tested for their affinity towards α₁-adrenoceptor and its α(1a), α(1b) and α(1d) subtypes, as well as serotonin 5-HT(1A) receptor. Several analogues (5, 6, 9, and 10) showed remarkable potency and selectivity towards α(1a), and α(1d) with respect to α(1b) subtype. None of the test compounds exhibited significant affinity for 5-HT(1A) receptor. Finally, on the basis of the α₁-AR subtypes 3D models recently proposed, we have elaborated theoretical interaction models for the new compounds. The theoretical study allowed us to predict the affinity of the new compounds as well as to infer the structural/dynamics determinants of their interaction with the three α₁-AR subtypes. Copyright (C) 1998 Elsevier Science Ltd.

Full text of DOI:10.1016/S0968-0896(98)00056-X

BIOORGANIC &
MEDICINAL
CHEMISTRY
Bioorganic & Medicinal Chemistry 6 (1998) 925±935
Isoxazolo-[3,4-d]-pyridazin-7-(6H)-ones and their
Corresponding 4,5-Disubstituted-3-(2H)-pyridazinone
Analogues as New Substrates for ꢀ -Adrenoceptor Selective
1
Antagonists: Synthesis, Modeling, and Binding Studies
a
a
b
Federica Montesano, Daniela Barlocco, Vittorio Dal Piaz,
c
c
d
Amedeo Leonardi, Elena Poggesi, Francesca Fanelli
d,
and Piero G. De Benedetti *
aIst. Chimica Farmaceutica e Tossicologica, Viale Abruzzi 42, 20131 Milano, Italy
bDip. Scienze Farmaceutiche, Via G. Capponi, 9 - 50121 Firenze, Italy
cPharmaceutical R&D Division, Recordati s.p.a., Via M. Civitali 1, 20148 Milano, Italy
dDipartimento di Chimica, Via Campi 183, 41100 Modena, Italy
Received 22 April 1997; accepted 21 January 1998
AbstractÐA series of phenylpiperazinylalkyl moieties were attached to monocyclic or bicyclic substituted pyr-
idazinones and the new compounds tested for their anity towards a -adrenoceptor and its a1a, a1b and a1d subtypes,
as well as serotonin 5-HT1A receptor. Several analogues (5, 6, 9, and 10) showed remarkable potency and selectivity
1
towards a , and a with respect to a1b subtype. None of the test compounds exhibited signi®cant anity for 5-HT1A
1
a
1d
1
receptor. Finally, on the basis of the a -AR subtypes 3D models recently proposed, we have elaborated theoretical
interaction models for the new compounds. The theoretical study allowed us to predict the anity of the new com-
pounds as well as to infer the structural/dynamics determinants of their interaction with the three a -AR subtypes.
1
#
1998 Published by Elsevier Science Ltd. All rights reserved.
Introduction
Molecular biology techniques allowed the identi®cation
of cDNAs encoding three a
1
-ARs, namely the rat a1a/d
,
Originally, the
5±9
a
1
-Adrenergic receptors (a
superfamily of G protein coupled receptors (GPCR)
that transduce signal across the cell membrane. The a
1
-AR) are members of the
the hamster a1b and the bovine a1c
.
pharmacological characteristics of the cloned subtypes
appeared to be inconsistent with the pharmacological
properties of the native subtypes: the existence of four
subtypes was also suggested.9
1
-
ARs mediate the functional e€ects of catecholamines
like epinephrine and norepinephrine by coupling to the
Gq mediated activation of phospholipase C culminating
into the phosphoinositide (PI) hydrolysis. The initial
It is now clear that the three recombinant a
1
-ARs
subdivision of the a
1
-AR into subtypes was based pri-
correlate closely with the three a -AR subtypes that
1
marily on the selectivity of WB-4101 for the a A-sub-
were identi®ed in native tissues and which mediate
their functional responses.10±12 These receptors are now
designated as a1A (a1a), a1B (a1b), and a1D (a1d), with
lower case subscripts being used to designate the
recombinant receptors, and upper case subscripts to
denote the native receptors.13 The recombinant receptor
previously designated as a -AR has now been shown
1
type1 and the selective alkylation of the a1B-subtype
shown by chloroethylclonidine (CEC). Other antago-
2
nists selective for the a A-AR were identi®ed, such as 5-
1
3
4
methylurapidil and S-(+)-niguldipine.
1
c
to correspond to the native a1A-AR,1
0,14±16
and as
*
Corresponding author. Tel: 39 59 378463; Fax: 39 59 373543.
such, the terms a1C- or a1c-AR are no longer used.17
0968-0896/98/$19.00 # 1998 Published by Elsevier Science Ltd. All rights reserved
PII: S0968-0896(98)00056-X

9
26
F. Montesano et al./Bioorg. Med. Chem. 6 (1998) 925±935
Furthermore, the a1D-AR has now been characterized
in detail, and antagonists capable of discriminating
between this receptor and the a1A-AR are now avail-
able, con®rming that these two receptors represent dis-
investigated by some of us24 as a possible substituent
for piperazinyl derivatives. In the search of more potent
1
and selective a -AR compounds, we have now syn-
thesized and tested a new series in which a phenyl-
piperazinylalkyl moiety has been introduced at di€erent
positions of the previously reported25,26 3,4-disub-
stituted-isoxazolo-[3,4-d]-pyridazin-7-(6H)-one (1) and
of its open analogue 2.
1
8,19
tinct a -AR subtypes.
1
In addition to a1A, a1B and a1D-AR subtypes, which
share a high anity for prazosin, the existence of addi-
1
tional a -AR has been proposed. These are called a1L-
ARs and are characterized by a low functional anity
for prazosin, intermediate anity for WB 4101 and 5-
methylurapidil and resistance to CEC alkylation.2
1
Moreover, on the basis of the a -AR 3D models recently
proposed,23 we have elaborated theoretical interaction
models between the new compounds and the three a1-
AR subtypes. The theoretical study allowed us to pre-
dict the anity of the new compounds as well as to infer
the structural/dynamics determinants of their selective
0±22
1
The presence of these di€erent a -AR subtypes in blood
vessels and other smooth muscles points to the impor-
tance of developing selective drugs for receptor classi®-
cation and characterization as well as for therapeutic
e€ectiveness.
1
interaction with the three a -AR subtypes.
The understanding of the molecular determinants of
drug selectivity towards GPCRs is a very challenging
target. One obvious problem is our limited knowledge
of the 3D structure of these membrane proteins, because
of diculties linked to their non degenerative puri®ca-
tion and crystallization. Moreover, even if a highly
Chemistry
The 6-substituted isoxazolo-[3,4-d]-pyridazin-7-(6H)-
ones (5±7) were prepared by alkylation of compound 1
by the required dibromoalkane in dimethylformamide
and potassium carbonate at room temperature, followed
by condensation with the appropriate substituted phenyl-
resolved structure of
a GPCR were immediately
ꢀ
available, the description of the mechanism of ligand±
receptor selective interaction in terms of both structure
and dynamics would remain a daunting task. This
problem may be approached by combining 3D model
building of receptor structure and computational simu-
lation of receptor dynamics. In this context, some of us
recently presented a preliminary rationalization, at the
molecular level, of antagonist selectivity towards the
piperazine in acetone and potassium carbonate at 80 C.
According to the same scheme, the corresponding open
derivatives (9±11) were prepared, starting from 2 (see
Scheme 1).
Alternatively, compounds 5±7 and 9±11 could be pre-
pared by treatment with the required substituted-4-
phenylpiperazin-1-yl-alkyl chloride, in turn obtained by
condensation in acetone and potassium carbonate of the
appropriate phenylpiperazine and the dihaloalkane27
(see Scheme 1). No signi®cantly di€erent yields were
obtained from the two pathways. The lower homologs 4
and 8 were obtained in good yields by Mannich reac-
tion, starting from 1 and 2, respectively (see Scheme 2).
2
3
three cloned a -AR subtypes. Molecular dynamics
1
simulations allowed a structural/dynamics analysis of
the three receptor structures in their free forms. This
analysis was paralleled by docking simulations, using 16
selective and nonselective antagonists that provided
theoretical quantitative structure±anity relationship
models.23 This theoretical study led to the hypothesis
that the transmembrane domains of the a -AR subtypes
Finally, the 3-substituted 12 was prepared by nucleo-
philic substitution on the 3-bromomethylisoxazolo-
pyridazinone 328 by ortho-methoxyphenylpiperazine in
1
have di€erent dynamic behaviors and di€erent topo-
graphies of the binding sites, which are mainly con-
stituted by conserved residues. In particular, the a -AR
1
a
binding site is more ¯exible and topographically di€er-
ent with respect to the other two subtypes. In other
words, the nonconserved residues seemed to exert
intramolecular rather than intermolecular e€ects, con-
ferring di€erent structural/dynamics features to the
conserved ligand binding sites.
Amongst the compounds showing high anity towards
1
a -AR relevant attention has recently been devoted to a
series of phenylpiperazines of general structure I, where
Ar represents several aromatic residues nonstructurally
related. Amongst others, the pyridazine system was
Chart 1.

F. Montesano et al./Bioorg. Med. Chem. 6 (1998) 925±935
927
Scheme 1. Reagents and conditions: (a) 1. Br(CH
2
)nBr, DMF, K
2
CO , rt, 2.
3
, K
2
CO
3
, Á, acetone; (b)
2 2
2.
, Á acetone; (c) ortho-methoxyphenylpiperazine, HCHO, EtOH, Á; (d) NH -NH H O, 10% Pd-C,
,
K
2
CO
3
EtOH, Á.
ꢀ
chloroform at 60 C (Scheme 2). Treatment of 12 with
hydrazine hydrate in re¯uxing ethanol in the presence of
1
the same reductive opening of the isoxazole ring to the
above reported 4 always brought about a parallel loss of
the alkylpiperazinyl moiety at position 6. See Experi-
mental and Table 1 for the chemical and physical data
of the newly synthesized compounds.
0% Pd-C led to the open analogue 13 in 70% yield
Scheme 2). It should be noted that attempts to apply
(
Binding studies
All compounds were tested for their anity towards
3
native a -receptor ( H-prazosin as speci®c ligand on rat
1
3
cerebral cortex), 5-HT1A receptor ( H-8-OH-DPAT, rat
hippocampus) and the cloned receptor subtypes expressed
3
on CHO cells ( H-prazosin). In case of good anity
for the a1d subtype, additional testing on human
recombinant a1a and a1b AR subtypes was performed.
Prazosin, 8-OH-DPAT and BMY 7378 were used as
reference standards. Data are shown in Table 2.
Modeling
Scheme 2. Reagents and conditions: (a) ortho-methoxy-
ꢀ
2
2.
-NH
H
2
O, 10% Pd-C,
The building of the 3D model of the three a
1
-AR sub-
phenylpiperazine, CHCl , 60 C; (b) NH
3
EtOH, Á.
types was described in detail in our previous paper.23

9
28
F. Montesano et al./Bioorg. Med. Chem. 6 (1998) 925±935
Table 1. Physicochemical data of compounds 4±11
ꢀ
Compd
n
R
% Yield
mp ( C)
Formula
4
5
6
7
8
9
1
2
2
3
1
2
2
3
H
H
Cl
Cl
H
H
Cl
H
79
70a
72a
68b
51
63a
70b
65b
166±168
128±129
160±162
130±131
198±200
154±155
154±155
oil
C
C
24
H
H
25
N
N
5
O
O
3
25
27
5
3
C
25
H
H
26ClN
28ClN
5
O
3
C
26
5
O
3
C
24
H
27
N
O
5 3
C
25
H
29
N
5
1
1
0
1
C
25
H
28ClN
5
O
3
C
26
31 5 3
H N O
aMethod B.
bMethod A.
i 1 1
Table 2. Anity constants (K , nM) of compounds 4±13 towards native a and 5-HT1A receptors and cloned a adrenoceptor
subtypes
K
i
(nM), native receptors (rat brain)
i
K (nM), cloned receptors (human brain)
Compd
a
1
5-HT1A
Ratio 5-HT1A/a
1
a
1a
a
1b
a
1d
4
5
6
7
8
9
2308
147.1
15.57
7.86
480.3
1101
2203
1434
395.7
123.1
NT
0.21
7.48
NT
1.49
1.85
1.55
NT
NT
31.09
7.72
2.57
NT
315.8
1.53
141.5
182.4
0.16
0.26
1.32
2395
596.8
24.56
42.68
46.73
515.6
0.74
5.01
4.75
6.27
18.81
NT
22.06
17.82
142.9
NT
0.93
0.75
1
1
1
0
1
3
NT
61.37
NT
2868
NT
137.8
84.74
0.29
1.28
NT
prazosin
BMY7378
2357
0.37
2.33
3185
0.00
0.00
0.58
378.4
NT
0.28
70.9
NT
281.9
18002
8
-OH-DPAT
Equilibrium dissociation constants (K
3
i
) were derived from IC50 using the Cheng±Pruso€ equation.34 The anity estimated were
3
derived from displacement of [ H]prazosin binding for a
1
-adrenoceptors and [ H]8-hydroxy-2-(di-n-propylamino)tetraline for 5-HT1A
receptor. Each experiment was performed in triplicate. The lable NT means not tested.
The average minimized structures of the a
1
-AR subtypes
as a spacer (see Table 2). These compounds show rea-
sonable a -AR binding anity values resulting from the
previously obtained were used for providing ligand-
receptor interaction models for the new synthesized
compounds. For each ligand±receptor minimized com-
plex, an intermolecular interaction descriptor was com-
puted and the linear regression equations previously
presented were used to calculate the anities of the new
1
di€erent contribution of a1a and a1b-AR subtypes; more-
over, they show low anity for the 5-HT1A receptor. In
general, 2-methoxy-5-chloro-phenylpiperazines show
higher a /5-HT selectivity than the corresponding 2-
1
1A
methoxyphenylpiperazines.29 None of compounds 4±13
showed signi®cant potency towards 5-HT1A, as clearly
indicated by the results reported in Table 2. In particular,
1
compounds for the three a -AR subtypes.
for the most potent a
ratio between the two K
1
-antagonists in this series (7) the
s approaches a value close to 200.
i
Results and Discussion
The most interesting compounds presented in this work
(5±7, 9, and 10) carry an ethylenic or a propylenic chain
In order to better de®ne the above trends, the mole-
cular determinants of anity and selectivity of the new

F. Montesano et al./Bioorg. Med. Chem. 6 (1998) 925±935
929
compounds towards the three a
been investigated by means of a theoretical study of
ligand±receptor interactions.
1
-AR subtypes have
Receptor and ligand distortion energies provide a
measure of both ligand and receptor penalties to go
from the free to the bound state. In other words, the
higher the receptor and ligand distortion energies the
lower the ligand±receptor interaction propensity and,
hence, the lower the ligand±receptor dynamic com-
plementarity. From these studies it was concluded
that, the `theoretical a1a selectivity', that is the selectiv-
ity encoded by theoretical descriptors, is mainly modu-
lated by receptor and ligand distortion energies. In other
words, subtype selectivity seemed to be mainly guided
by dynamic complementarity (induced ®t) between the
ligand and the receptor.
A theoretical intermolecular interaction descriptor as a
predictor of the anity/selectivity towards the three
1
ꢀ -AR subtypes
Recently, some of us presented a theoretical investiga-
tion, at the molecular level, of antagonist selectivity
2
3
towards the three cloned a -AR subtypes. A quanti-
1
tative description of the theoretical results, which were
based on the structural/dynamics analysis of the free
receptor forms, was provided by docking each of the
antagonists into the average minimized structures of the
In the present work, the average minimized structures of
three a
1
-AR subtypes. For each selected ligand±receptor
1
the a -AR subtypes previously obtained have been used
minimized complex the theoretical binding energies (BE)
were computed and correlated with the experimental
for providing ligand±receptor interaction models for the
new synthesized compounds. Each compound has been
docked into the average minimized structures of the
i
binding anities (pK ) measured on the three animal
cloned subtypes; the resulting linear regression equa-
tions are reported as follows:
1
three a -AR subtypes following the docking criteria and
the results of the previous theoretical study.23 More-
over, for each ligand±receptor minimized complex, the
BE have been computed and the linear regression
equations previously presented (see above in the text)
have been used for calculating the anities of the new
compounds. Table 3 shows the theoretical binding a-
nities as well as the theoretical intermolecular interac-
tion descriptors computed on the ligand-receptor
minimized complexes. Interestingly, the theoretical
binding anities (Table 3) interpolated from the pre-
viously reported linear regression models, calibrated
(
(
(
a) pK1a=5.45(‹0.37)� 0.08(‹0.01)BE1a, n=16,
r=0.90, s=0.52;
b) pK1b=4.20(‹0.32)� 0.12(‹0.01)BE1b
r=0.94, s=0.37;
,
n=16,
n=16,
c) pK1d=4.43(‹0.36)� 0.09(‹0.01)BE1d
,
r=0.92, s=0.39;
where n is the number of compounds, r is the corre-
lation coecient, s is the standard deviation and the
number in parentheses give the 95% con®dence
intervals.
1
on the animal cloned a -AR subtypes, are consistent
with the experimental binding anity data, measured
towards the three human cloned a -AR subtypes
(Table 2).
1
The BE of the ligand±receptor minimized complexes
were computed according to the following formula:
BE=IE+ER+EL, where IE is the ligand±receptor
interaction energy and ER and EL are the distortion
energies of the receptor and of the ligand, respectively,
calculated as di€erences between the energies of the
bound and of the free optimized molecular forms.
1
It can be seen that the generic a -AR anities of com-
pounds 5, 6, 9, and 10 are quite increased if the anity
values towards the cloned a1a- and a1d-ARs are con-
sidered. This is mainly due to the fact that these com-
pounds show a high anity as well as a remarkable
Table 3. Computed binding anities (K) and binding (BE, kcal/mol), interaction (IE, kcal/mol), receptor distortion (ER, kcal/mol),
and ligand distortion (EL, kcal/mol) energies of the ligand-a -adrenoceptor subtype complexes
1
Compd
Kla K1b K1d
BE1a
IE1a ER1a EL1a BE1b
IE1b ER1b EL1b BE1d
IE1d ER1d EL1d
4
5
6
7
8
9
426.58
� 21.62 � 77.26 47.58 8.06
2.75 46.77
0.45 34.67
0.55 0.78
2.00 � 38.88 � 87.17 43.83 4.46 � 26.11 � 79.13 48.24 4.78 � 47.78 � 84.13 30.33 6.32
0.52 � 48.80 � 90.21 35.27 6.14 � 27.21 � 80.70 48.95 4.54 � 53.85 � 90.21 25.70 10.66
4.68 � 47.58 � 92.60 36.97 8.05 � 40.91 � 81.76 36.26 4.59 � 43.43 � 89.92 38.85 7.64
524.81
� 20.56 � 76.52 42.58 13.38
0.14 � 39.55 � 89.28 42.71 7.02 � 31.60 � 83.37 47.15 4.62 � 60.27 � 93.99 24.74 8.98
0.15 � 42.28 � 91.53 40.73 8.52 � 35.65 � 85.25 45.28 4.32 � 59.87 � 98.03 25.34 12.82
2.45 10.47
1.48 3.31
10
11
13
16.89 32.36 75.86 � 28.95 � 81.73 47.45 5.33 � 27.44 � 81.20 45.96 7.80 � 21.91 � 93.88 52.70 11.27
20.89
� 36.12 � 80.38 34.22 10.04

9
30
F. Montesano et al./Bioorg. Med. Chem. 6 (1998) 925±935
selectivity for these two a
1
-AR subtypes than for the
values are associated with low interaction energies and
low receptor distortion energies.
a -AR subtype. Interestingly, the high anity com-
1b
pounds 5, 6, 9, and 10 show slight a1d/a1a and high a1d
/
1b selectivities, suggesting that they can constitute good
templates for designing high anity and selective
antagonists towards the a1d-AR subtype. The impor-
tance of these compounds is remarkable considering the
low number of available high anity and selective
a
Structural features of ligand±receptor selective
interaction
The drawings shown in Figure 1 may give a general idea
about how the considered antagonists are oriented into
antagonists for the a1d or the a1b
.
1
the binding sites of the three a -AR subtypes. In gen-
eral, the protonated nitrogen atom of all the antagonists
makes charge reinforced H-bonding interaction with the
carboxylate of D3:07 (the ®rst digit indicates the helix
and the last two digits indicate the position of the resi-
due in the helix), whereas the other molecular moieties
®t two binding pockets formed by TM1, TM2, and TM7
on one side, and by TM4, TM5, and TM6 on the other
one, lying in an almost symmetrical topography with
By analysing the theoretical descriptors reported in
Table 3 it can be inferred that low a1d binding energies
(
high anity) are reached by means of both low inter-
action energies (low negative values) and low receptor
distortion energies (low positive values). In general, the
a1d/a1b selective compounds show higher interaction
energies and higher distortion energies when they inter-
act with the a -AR than with the a -AR. The highest
respect to D3:07. The a -AR antagonists usually are
1
b
1d
1
anity towards the a1b-AR is shown by compound 7
and is associated with the lowest a1b-AR distortion
energy. As for the a -AR subtype, the highest anity
constituted by two aromatic moieties on the two sides of
the protonated nitrogen atom. In the proposed interaction
models, the aromatic moiety closer to the protonated
1
a
1
Figure 1. Details of the interaction between compound 9 and the three a -AR subtypes. The antagonist±receptor complexes are
viewed from the intracellular side of the membrane in a direction parallel to the helix main axes. Residues are labeled according to an
arbitrary numbering: the ®rst digit indicates the helix and the last two digits indicate the position of the residue in the helix. Helices 1,
2, 3, 4, 5, 6 and 7 are colored in blue, orange, green, pink, yellow, light blue and violet, respectively.

F. Montesano et al./Bioorg. Med. Chem. 6 (1998) 925±935
931
nitrogen atom docks into the hydrophobic pocket
formed by residues of TM4, TM5, and TM6.
only a change in the interaction pattern involving the
4,5,6-trisubstituted-3-(2H)-pyridazinone moiety and the
aromatic cluster formed by W3:03, F7:07 and Y7:11,
but also the geometry of the charge reinforced H-bond-
ing interaction between the protonated nitrogen atom of
the ligand and the carboxylate side chain of D3:07 and,
hence, the orientation of the phenyl-piperazine. More-
over, the shortening of the spacer of compounds 5 and 9
into a methylenic chain induces the 3,4-disubstituted-
isoxazolo-[3,4-d]-pyridazin-7-(6H)-one and the 4,5,6-tri-
substituted-3-(2H)-pyridazinone moieties of compounds
4 and 8, respectively, to interfere with the interaction
between the protonated nitrogen atom of the ligands
and D3:07 of the receptor (Figure 2). As a consequence,
the 4- and 8-a1d receptor complexes show higher inter-
action energies and higher receptor distortion energies
than the corresponding homo-derivatives 5 and 9, which
is consistent with their having anity pro®les. As shown
by Figure 2, the two compounds 4 and 8 give a very
similar interaction pattern, which is consistent with their
having binding anity pro®les.
Interestingly, as clearly shown by Figure 2, the 4,5,6-
trisubstituted-3-(2H)-pyridazinones and the corres-
ponding 3,4-disubstituted-isoxazolo-[3,4-d]-pyridazin-7-
(
6H)-ones give quite similar interaction patterns with
each of the three a -AR subtypes, which is consistent
1
with their having similar binding anity pro®les. Simi-
larly, the addition of a chlorine atom in the 5-position of
the 2-methoxyphenylpiperazine moieties of compounds
5 and 9, giving the derivatives 6 and 10, respectively,
does not add any change in their interaction patterns.
On the contrary, the lengthening (i.e., 11) or the
shortening (i.e., 4 and 8) of the spacer even if by a single
methylenic unit, causes dramatic di€erences in the
interaction mechanisms in comparison with the corre-
sponding ethylenic derivatives (i.e., 5, 6, 9, and 10). In
particular, as for compound 11, the lengthening of the
ethylenic spacer into a propylenic chain induces not
Figure 2. Details of the interaction between compounds 9 and 5 and the three a
1
-AR subtypes. The complexes between compound 9
-AR subtype.
and each of the three a -AR subtypes have been superimposed to those between compound 5 and the corresponding a
1
1
Moreover, in the bottom right side of this ®gure, the superimposed complexes between compounds 4 and 8 and the a1d-AR subtype
are also represented.

9
32
F. Montesano et al./Bioorg. Med. Chem. 6 (1998) 925±935
In summary, among the new synthesized compounds,
the most interesting derivatives carry an ethylenic spacer
between the protonated 2-methoxyphenylpiperazine or
the 2-methoxy-5-chloro-phenyl-piperazine and the
Taken together, these results suggest that the new N-
arylalkylpiperazines can be considered good templates
for the development of novel a1d selective ligands, with
1
respect to both of the other two a -AR subtypes and the
4
,5,6-trisubstituted-3-(2H)-pyridazinone or the 3,4-disub-
stituted-isoxazolo-[3,4-d]-pyridazin-7-(6H)-one moieties.
In fact, these antagonists show slight a1d/a1a, high a1d
a , and very high a /5-HT and a /5-HT selec-
5-HT1A serotonergic receptor. The importance of such
compounds is remarkable, considering that the develop-
ment of high anity and selective antagonists for the
a - or the a -subtypes is quite dicult. In fact, BMY
/
1b
1a
1A
1d
1A
1d
1b
tivities. The ethylenic spacer serves to place the 4,5,6-
trisubstituted-3-(2H)-pyridazinone or the 3,4-disub-
stituted-isoxazolo-[3,4-d]-pyridazin-7-(6H)-one moieties
into a hydrophobic pocket mainly formed by residues of
TM1, TM2, and TM7 in such a way that, as for the a1a
subtype, the pyridazinone moiety and the substituent
phenyl ring form a p±p sandwich interaction with F7:07
and Y7:11 and an orthogonal s±p interaction with
W3:03, respectively, while as for the a1d subtype, the
pyridazinonic carbonylic oxygen atom forms an H-
bonding interaction with N7:13 and the substituent
phenyl ring forms orthogonal s±p interactions with
both W3:03 and F7:07 (Figure 2). These interactions
together with the van der Waals attractive interactions
with L2:13 and V2:20 might contribute to the stronger
interaction energies and the lower distortion energy
shown by 5, 6, 9, and 10 when they interact with both
the a1a and the a1d-AR subtypes rather than with the
7378, one of the few known a1d-selective antagonists,
with respect to the two other AR-subtypes, binds also
the 5-HT1A receptor with nanomolar anity.
Finally, the computational simulated interaction
between the new synthesized compounds and the three
a1-AR subtypes con®rms our previous hypothesis that
selective and nonselective antagonists give di€erent
1
interaction patterns when they bind the three a -AR
subtypes.23 This is mainly due to the fact that the ligand
binding sites of these subtypes, even if constituted by
conserved residues, have di€erent structural/dynamics
features as a consequence of the intramolecular in¯u-
ence exerted by the nonconserved residues.
Experimental
Chemistry
a1b subtype. Interestingly, the protonated nitrogen atom
of these compounds gives the strongest charge rein-
forced H-bonding interactions with the carboxylate ion
È
Melting points were determined on a Buchi 510 capil-
lary melting points apparatus and are uncorrected. Ele-
of D3:07 in the a
two subtypes.
1
d subtypes with respect to the other
mental analyses for the test compounds were within
1
‹0.4% of the theoretical values. H NMR spectra were
È
recorded on a Bruker AC200 spectrometer; chemical
Either selective or nonselective compounds give dif-
ferent interaction patterns when they bind to the three
a1-AR subtypes. This is mainly due to the fact that
the ligand binding site of these subtypes, even if
constituted by conserved residues, have di€erent struc-
tural/dynamics features as a consequence of the intra-
molecular in¯uence exerted by the non-conserved
residues.
shifts are reported as d (ppm) relative to tetra-
methylsilane as internal standard. CDCl was used as
3
the solvent, unless otherwise noted. TLC on silica gel
plates was used to check product purity. Silica gel 60
(Merck; 70±230 mesh) was used for column chromato-
graphy and silica gel 60 (Merck, 230±400 mesh) for
¯ash chromatography. The structures of all compounds
were consistent with their analytical and spectroscopic
data.
Very recently, experimental evidences have emerged
1
which support our hypothesis that the three a -AR
subtypes have di€erent structural/dynamics behavior
and that, in particular, the a1a-AR subtype is more
3-Methyl-4-phenyl-6-[(4-substitutedphenylpiperazin-1-yl)
alkyl]-isoxazolo-[3,4-d]-pyridazin-7-(6H)-ones
(5±7):
¯exible and more sensitive to the environmental condi-
tions with respect to the other two subtypes.
General methods. Method A. (a) A mixture of 1
(1 mmol), the required dibromoalkane (10 mmol) and
K CO (1.5 mmol) in DMF (8 mL) was stirred at rt for
3
0
2
3
5
h. The mixture was then poured into water (15 mL)
and extracted with dicloromethane (3Â10 mL). After
drying and evaporation of the solvent, the residue was
normally used as such for the next step. When neces-
sary, it could be puri®ed by silica gel chromatography,
eluting with cyclohexane:EtAc (8:2). (b) A mixture of
the required bromoalkylpyridazinone (1 mmol), the
Conclusion
The new compounds presented in this work show, in
general, a very good selectivity between the a -adreno-
ceptor and the 5-HT1A serotonergic receptor. Moreover,
four of them (5, 6, 9, and 10) are provided with
remarkable a1d/a1b and a slight a1d/a1a-selectivity.
1
2 3
appropriate phenylpiperazine (1.5 mmol) and K CO

F. Montesano et al./Bioorg. Med. Chem. 6 (1998) 925±935
933
(
1.5 mmol) in acetone (20 mL) was re¯uxed overnight.
Radioligand binding assays at native receptors. Binding
studies on native a adrenergic and 5-HT serotonergic
receptors were carried out in membranes of rat cerebral
After cooling, the inorganic salts were ®ltered o€ and
the residue puri®ed by silica gel chromatography, elut-
1
1A
ing with CH
for data.
2
Cl
2
/MeOH 98/2. See Scheme 1 and Table 1
cortex (a ) and hippocampus (5-HT1A). Male Sprague±
1
Dawley rats (200±300 g, Charles River, Italy) were killed
by cervical dislocation and di€erent tissues were excised
ꢀ
Method B. A mixture of 1 (1 mmol), K CO
2
3
(1.5 mmol)
and immediately frozen and stored at � 70 C until use.
and the required chloroalkylphenylpiperazine (1.5 mmol)
in acetone (20 mL) was re¯uxed overnight. After cool-
ing, the inorganic salts were ®ltered o€, the solvent
evaporated under vacuum and the residue puri®ed by
silica gel chromatography eluting with CH Cl :EtAc
Tissues were homogenized (2Â20 s) in 50 vols of cold
Tris±HCl bu€er pH 7.4, using a Politron homogenizer
(speed 7). Homogenates were centrifuged at 49000g for
10 min, resuspended in 50 vols of the same bu€er, incu-
ꢀ
bated at 37 C for 15 min and centrifuged and resus-
pended twice more. The ®nal pellets were suspended in
2
2
(
4:6). See Scheme 2 and Table 1 for data.
1
pargiline and 0.1% ascorbic acid. Membranes were
00 vols of Tris±HCl bu€er pH 7.4, containing 10 mM
4
-Amino-5-acetyl-6-phenyl-2-[(4-substitutedphenylpiper-
ꢀ
azin-1-yl)alkyl]-3(2H)-pyridazinones (9±11): General
method. Compounds were prepared according to the
above reported methods A and B, starting from 2. See
Table 1 for data.
incubated in a ®nal volume of 1 mL for 30 min at 25 C
3
3
1
with 0.1±0.5 nM [ H]prazosin (a ) or 0.5±1.5 nM [ H]8-
OH-DPAT (5-HT1A), in absence or presence of com-
peting drugs; nonspeci®c binding was determined in the
1
presence of 10 mM phentolamine (a ), or 10 mM 5-HT
3
-Methyl-4-phenyl-6-[(4-ortho-methoxyphenylpiperazin-
-yl)methyl]-isoxazolo-[3,4-d]-pyridazin-7-(6H)-one (4). A
(5-HT1A).
1
mixture of 1 (0.1 g, 0.44 mmol), 37% formaldehyde
1.7 mL, 0.44 mmol) and ortho-methoxyphenylpiper-
The incubation was stopped by addition of ice-cold
Tris±HCl bu€er and rapid ®ltration through 0.2%
polyethyleneimine pretreated Whatman GF/B or
Schleicher & Schuell GF52 ®lters. The ®lters were then
washed with ice-cold bu€er and the radioactivity
retained on the ®lters was counted by liquid scintillation
spectrometry.
(
azine (0.13 g; 0.67 mmol) in EtOH (6 mL) was re¯uxed
for 2 h. After cooling and concentration of the solvent,
the so formed precipitate was ®ltered to give 0.15 g
ꢀ
(
79%) of 4, mp 166±168 C (see also Table 1)
4
-Amino-5-acetyl-6-phenyl-2-[(4-ortho-methoxyphenyl-
piperazin-1-yl)methyl]-3(2H)-pyridazinone (8). Compound
was prepared as above reported for 4, starting from 2.
Yield 83%, mp 198±200 C (see also Table 1).
Radioligand binding assay at cloned ꢀ -adrenoceptors.
1
ꢀ
1
Binding to cloned human a -adrenoceptor subtypes
was performed in membranes from CHO cells (Chinese
hamster ovary cells) transfected by electroporation with
4
1
-Phenyl-6-methyl-3-[(4-ortho-methoxyphenylpiperazin-
-yl)methyl]-isoxazolo-[3,4-d]-pyridazin-7-(6H)-one (12).
DNA expressing the gene encoding each a -adreno-
1
ceptor subtype. Cloning and stable expression of the
A solution of 3* (0.2 g, 0.6 mmol) and ortho-methoxy-
phenylpiperazine (0.32 g, 1.6 mmol) in CHCl3 (6 mL)
was re¯uxed for 2 h. After cooling and evaporation of
the solvent, the residue was puri®ed by ¯ash chromato-
graphy eluting with cyclohexane:EtAc (8:2). Yield 79%,
1
human a -adrenoceptor gene was performed as pre-
viously described.9 CHO cell membranes (30 mg pro-
teins) were incubated in 50 mM Tris±HCl, pH 7.4, with
3
0
.1±0.4 nM [ H]prazosin, in a ®nal volume of 1 mL for
ꢀ
ꢀ
1
30 min at 25 C, in absence or presence of competing
drugs (1 pM±10 mM). Nonspeci®c binding was deter-
mined in the presence of 10 mM phentolamine. The
incubation was stopped by addition of ice-cold Tris±
HCl bu€er and rapid ®ltration through 0.2% poly-
ethyleneimine pretreated Whatman GF/B or Schleicher
mp 177±178 C. H NMR: d 2.4±2.5 (m, 4H), 2.9±3.0
m, 4H), 3.7 (s, 2H), 3.8 (2s, 6H), 6.9±7.0 (m, 4H), 7.5±
(
7
.7 (m, 5H). C H N O (C,H,N)
23 25 5 3
2
-Methyl-4-amino-6-phenyl-5-[(4-ortho-methoxyphenyl-
piperazin-1-yl)acetyl]-3(2H)-pyridazinone (13). A mixture
of 12 (0.08 g, 0.00018 mol), hydrazine hydrate (0.04 mL,
&
Schuell GF52 ®lters.
0
.0007 mol) and 10% Pd-C (0.02 g) in EtOH (5 mL) was
re¯uxed for 1 h. The still-hot mixture was ®ltered, the
catalyst thoroughly washed with CH Cl , the solvent
Data analysis
2
2
evaporated and the residue puri®ed by silica gel chro-
matography, eluting with CH Cl :MeOH (98:2). Yield
0%. H NMR: d 2.7±2.8 (m, 2H), 3.0±3.2 (m, 4H), 3.4±
.5 (m, 2H), 3.7±3.8 (app. d., 6H), 6.9±7.0 (m, 6H), 7.5±
.7 (m, 5H). C23 (C,H,N).
The inhibition of speci®c binding of the radioligands by
the tested drugs was analyzed to estimate the IC50 value
by using the nonlinear curve-®tting program All®t.31
2
2
1
3
3
7
i
The IC50 value is converted to an anity constant (K )
by the equation of Cheng and Pruso€.32
H
27
5
N O
3

9
34
F. Montesano et al./Bioorg. Med. Chem. 6 (1998) 925±935
Modeling
D Model building of the ꢀ
simulations. The building of the transmembrane (TM)
8. Perez, D. M.; Piascik, M. T.; Graham, R. M. Mol. Phar-
macol. 1991, 40, 876.
3
1
-AR subtypes and docking
9. Schwinn, D. A.; Lomasney, J. W. Eur. J. Pharmacol. 1992,
227, 433.
model of the three-a -AR subtypes was described in
10. Laz, T. M.; Forray, C.; Smith, K. E.; Bard, J. A.; Vaysse,
P. J. J.; Branchek, T. A.; Weinshank, R. L. Mol. Pharmacol.
1
2
3,33
detail in our previous papers.
The input structures
1
994, 46, 414.
of the ligand±receptor complexes were obtained by
docking each ligand into the target receptor structure
averaged over the last 100 picoseconds of the Molecular
Dynamics simulation. For each complex several mini-
mizations were performed in order to probe di€erent
orientations and conformations of the ligands and to
optimize several fundamental interactions (i.e., the
charge reinforced H-bond between the protonated
nitrogen atom of the ligands and the carboxylate of
D3:07), which is the putative key residue for the cationic
ligand-receptor recognition.34
11. Forray, C.; Bard, J. A.; Wetzel, J. M.; Chiu, G.; Shapiro,
E.; Tang, R.; Lepor, H.; Hartig P. R.; Weinshank, R. L.;
Branchek, T. A.; Gluchowski, C. Mol. Pharmacol. 1994, 45,
7
1
03.
2. Forray, C.; Bard, J. A.; Laz, T. M.; Smith, K. E.; Vaysse,
P. J.; Weinshank, R. L.; Gluchowski, C.; Branchek, T. A.
Faseb J. 1994, 8, A353.
13. Hieble, J. P.; Bylund, D. B.; Clarke, A. E.; Eikenberg, D.
C.; Langer, S. Z.; Lefkowitz, R. J.; Minneman, K. P.; Ru€olo,
R. R. Pharmacol. Rev. 1995, 47, 267.
1
4. Ford, A. P. D. W.; Williams, T. J.; Blue, D. R.; Clarke, D.
E. Trends Pharmacol. Sci. 1994, 15, 167.
5. Testa, R.; Taddei, C.; Poggesi, E.; De Stefani, C.; Cotec-
Modeling studies were performed with the molecular
graphics package QUANTA (version 4.0).35 Energy
minimizations and MD simulations were obtained by
means of the program CHARMm (version 22).36 Mini-
mizations were carried out by using 200 steps of steepest
descent followed by a conjugate gradient minimization,
1
chia, S.; Hieble, J. P.; Sulpizio, A. C.; Naselsky, D. P.;
Bergsma, D. J.; Ellis, C.; Swift, A.; Ganguly, S.; Ru€olo, R. R.
Jr.; Leonardi, A. Pharmacol. Comm. 1995, 6, 79.
1
6. Pimoule, C.; Langer, S.; Graham, D. Eur. J. Pharmacol-
Mol. Pharmacol. 1995, 290, 49.
7. Schwinn, D. A.; Johnston, G. I.; Page, S. O.; Mosley, M.
Ê
until the rms gradient was less than 0.001 kcal/mol A. A
Ê
1
distance dependent dielectric term ("=4r) and a 12 A
J.; Wilson, K. H.; Worman, N. P.; Campbell, S.; Fidock, M.
D.; Furness, L. M.; Parry-Smith, D. J.; Peter, B.; Bailey, D. S.
J. Pharmacol. Exp. Ther. 1995, 272, 134.
nonbonded cuto€ distance were chosen. The `united atom
approximation' was used for computational eciency.36
The charge distributions of the ligands in their proto-
nated form were obtained in the AM1 framework.37
1
8. Goetz, A. S.; King, H. K.; Ward, S. D. C.; True, T. A.;
Rimele, T. J.; Saussy, D. L. Jr. Eur. J. Pharmacol. 1995, 272, R5.
9. Hieble, J. P.; Ru€olo, R. R. Jr.; Sulpizio, A. C.; Naselsky,
1
D. P.; Conway, T. M.; Ellis, C.; Swift, A. M.; Ganguly, S.;
Bergsma, D. J. Pharmacol. Comm. 1995, 6, 91.
Acknowledgements
Financial support from CNR and Ministero dell' Uni-
versita e della Ricerca Scienti®ca (funds 40%) is
Á
acknowledged. The CICAIA (University of Modena) is
acknowledged for the technical support and for the
computer facilities.
2
0. Flavahan, N. A.; Vanhoutte, P. M. Trends Pharmacol. Sci.
1986, 7, 347.
1. Muramatsu, I.; Ohmura, T.; Kigoshi, S.; Hashimoto, S.;
Oshita, M. Br. J. Pharmacol. 1990, 99, 197.
2. Muramatsu, I.; Ohmura, T.; Hashimoto, S.; Oshita, M.
Pharmacol. Comm. 1995, 6, 23.
3. De Benedetti, P. G.; Fanelli, F.; Menziani, M. C.; Cocchi,
M.; Testa, R; Leonardi, A. Bioorg. Med. Chem. 1997, 5, 809.
4. Montesano, F.; Barlocco, D.; Cignarella, G.; De Benedetti,
2
2
2
References and Notes
1
. Morrow, A. L.; Creese, I. Mol. Pharmacol. 1986, 29, 321.
2
2
. Han, C.; Abel, P. V.; Minneman, K. P. Nature (London)
987, 329, 333.
P. G.; Fanelli, F.; Leonardi, A.; Taddei, C. 5th International
Symposium on the Chemistry and Pharmacology of Pyridazines,
Sopron (Hungary), 1996, P13.
1
3
4
. Hanft, G.; Gross, G. Br. J. Pharmacol. 1989, 97, 691.
25. Sprio, V.; Aiello, E.; Mazza, A. Annali di Chimica, 1967,
. Boer R.; Grassegger, A.; Schudt, C. H.; Glossman, H. Eur.
5
7, 836.
J. Pharmacol. 1989, 172, 131.
. Cotecchia, S.; Schwinn, D. A.; Randall, L. L.; Lefkowitz, R.
J.; Caron, M. G.; Kobilka, B. K. Proc. Natl. Acad. Sci. U.S.A.
26. Romeo, G.; Russo, F.; Guccione, S.; Barbarulo, D.; De
Blasi, A. Il Farmaco 1995, 50, 471.
5
1
988, 85, 7159.
27. Amschler, H.; Krastinat, W. Ger. O€en. 2,258,561; Chem.
Abstr. 1973, 79, 78841a.
6. Schwinn, D. A.; Lomasney, J. W.; Lorenz, W.; Szklut, P. J.;
Fremeau, R. T.; Yang-Feng, T. L.; Caron, M. G.; Lefkowitz,
R. J.; Cotecchia, S. J. Biol. Chem. 1990, 265, 8183.
28. Dal Piaz, V. Personal communication.
29. Leonardi, A.; Testa, R.; Motta, G.; De Benedetti, P. G.;
7
. Lomasney, J. W.; Cotecchia, S.; Lorenz, S.; Leung, W.;
Schwinn, D. A; Yang-Feng, T. L.; Brownstein, M.; Lefkowitz,
Hieble, P.; Giardina', D. In Perspective in Receptor Research;
Giardina', D.; Piergentili, A.; Pigini, M., Eds; Elsevier Science
B.V.: Amsterdam, 1995; p 135.
R. J.; Caron, M. G. J. Biol. Chem. 1991, 266, 6365.

F. Montesano et al./Bioorg. Med. Chem. 6 (1998) 925±935
935
30. Ford, A. P. D. W; Daniels, D. V.; Chang, D. J.; Diaz, M.
R.; Gever, J. R.; Jasper, J. R.; Lesnick, J. D.; Clarke, D. E.
Brit. J. Pharmacol. 1996, 118, 29P.
34. Cavalli, A.; Fanelli, F.; Taddei, C.; De Benedetti, P. G.;
Cotecchia, S. FEBS Lett. 1996, 399, 9.
3
5. Quanta/CHARMm, 1990 Molecular Simulations, 200 Fifth
Avenue.
6. Brooks, B. R.; Bruccoleri, R. E.; Olafson, B. D.; States, D.
3
1. DeLean, A.; Munson, P. J.; Rodbard, D. Am. J. Physiol.
978, 235, E97.
1
3
3
2. Cheng, Y. C.; Pruso€, W. H. Biochem. Pharmacol. 1973,
2, 3099.
J.; Swaminathan, S.; Karplus, M. J. Comput. Chem. 1983, 4,
187.
2
33. Fanelli, F.; Menziani, M. C.; Cocchi, M.; De Benedetti, P.
G. J. Mol. Struct. (Theochem) 1995, 333, 49.
37. Dewar, M. J. S.; Zoebisch, E. G.; Healey, E. F.; Stewart, J.
J. P. J. Am. Chem. Soc. 1985, 107, 3902.