Synthesis and biological activity of new 1,4-benzodioxan-arylpiperazine derivatives. Further validation of a pharmacophore model for α1-adrenoceptor antagonists

Article abstract of DOI:10.1016/S0968-0896(01)00286-3

A series of WB4101 (1)-related benzodioxanes (2-17) have been synthesized by replacing the phenoxyethyl moiety of 1 with a N-alkyl piperazine bearing a cyclic substituent (a substituted or unsubstituted phenyl group, a pyridine or pyridazinone ring, a furoyl moiety) at the second nitrogen atom. The binding profile of these compounds has been assessed by radioligand receptor binding assay at α₁- and α₂-adrenoceptors, in comparison to prazosin and rauwolscine, respectively. Moreover, structure-activity relationships have been derived for compounds 2-17 based on their fitting to a pharmacophore model for α₁-adrenoceptor antagonists recently proposed by our research group. In a parallel way, the same compounds have been used to further test the predictive power and statistical significance of the model itself. The accuracy of the results obtained also in this case revealed the robustness of the calculated pharmacophore model and led to the identification of the molecular structural moieties which are thought to contribute to the biological activity. Copyright

Full text of DOI:10.1016/S0968-0896(01)00286-3

Bioorganic & Medicinal Chemistry 10 (2002) 361–369
Synthesis and Biological Activity of New 1,4-Benzodioxan-
arylpiperazine Derivatives. Further Validation of a
Pharmacophore Model for ꢀ -Adrenoceptor Antagonists
1
a
b
c,
c
b
Roberta Barbaro, Laura Betti, Maurizio Botta, * Federico Corelli, Gino Giannaccini,
a
c
c
a,
Laura Maccari, Fabrizio Manetti, Giovannella Strappaghetti * and Stefano Corsano
aIstituto di Chimica e Tecnologia del Farmaco, Universita` di Perugia, Via del Liceo 1, 06123 Perugia, Italy
bDipartimento di Psichiatria, Neurobiologia, Farmacologia e Biotecnologie, Universita` di Pisa, Via Bonanno 6, 56126 Pisa, Italy
cDipartimento Farmaco Chimico Tecnologico, Universita` degli Studi di Siena, Via A. Moro, 53100 Siena, Italy
Received 18 April 2001; accepted 10 August 2001
Abstract—Aseries of WB4101 ( 1)-related benzodioxanes (2–17) have been synthesized by replacing the phenoxyethyl moiety of 1
with a N-alkyl piperazine bearing a cyclic substituent (a substituted or unsubstituted phenyl group, a pyridine or pyridazinone ring,
a furoyl moiety) at the second nitrogen atom. The binding profile of these compounds has been assessed by radioligand receptor
binding assay at a - and a -adrenoceptors, in comparison to prazosin and rauwolscine, respectively. Moreover, structure–activity
1
2
1
relationships have been derived for compounds 2–17 based on their fitting to a pharmacophore model for a -adrenoceptor
antagonists recently proposed by our research group. In a parallel way, the same compounds have been used to further test the
predictive power and statistical significance of the model itself. The accuracy of the results obtained also in this case revealed the
robustness of the calculated pharmacophore model and led to the identification of the molecular structural moieties which are
thought to contribute to the biological activity. # 2001 Elsevier Science Ltd. All rights reserved.
Introduction
with affinity towards a -AR are expected to potentially
1
bind to both a and a receptors.
1
2
The a -adrenergic receptors (a -AR) are a family of G-
1
1
protein coupled seven-transmembrane helix receptors
which are mainly involved in the cardiovascular and
In this context, many literature reports have highlighted
that the arylpiperazinylalkyl moiety is a key element in
1
6
central nervous system. It is now clear that a -AR are
defining a -AR antagonist activity. Moreover, great
1
1
comprised of multiple subtypes. To date, they are class-
2
attention has been paid to compounds bearing the 2-
aminomethyl-1,4-benzodioxane fragment, such as 1
ified into a1A, a , and a , which possess high affinity
1B
1D
for prazosin, and the corresponding cloned counterparts
3
(Fig. 1), due to their potent and highly selective a -AR
antagonist activity.
1
7
a , a , a , respectively. In addition to these three
1
a
1b
1d
subtypes, the existence of an additional a -AR (called
1
4
a ) has been postulated.
Research efforts in the area of a -AR receptor antago-
1
L
1
nists have led to the discovery of some clinically useful
8
In a similar way, a -AR have been classified in four
antihypertensive drugs, acting by relaxation of the
vascular smooth muscle which contains high concentra-
2
2
c
subtypes, called a2A, a , a , a2D, respectively.
2B
2C
tions of a -AR. In a similar way, a -AR blockers have
1
1
5
Molecular cloning studies have shown that the a - and
been employed in the treatment of benign prostatic
hyperplasia, due to the significant improvements in
symptoms and flow rates in patients with bladder out-
1
a2-adrenoceptors have many common features which
could reflect their similar mechanisms of action. As a
consequence of such similarities, synthetic compounds
9À11
flow obstruction.
The goal of our research project was the synthesis of new
benzodioxane-arylpiperazine hybridized compounds
possibly characterized by high affinity and selectivity
towards a -AR with respect to a -AR. Thus, in an effort
*
0
5
Corresponding authors. M. Botta tel.: +39-0577-234306; fax: +39-
577-234333; G. Strappaghetti tel.: +39-075-585-5136; fax: +39-075-
85-5129. E-mail: botta@unisi.it (M. Botta); noemi@unipg.it (G.
Strappaghetti)
1
2
0
968-0896/02/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(01)00286-3

362
R. Barbaro et al. / Bioorg. Med. Chem. 10 (2002) 361–369
Chemistry
The synthesis of compounds 10–17 is highlighted in
Scheme 1. These compounds were prepared by reacting
the 1-(2-chloroethyl)- or 1-(3-chloropropyl)-4-arylpiper-
azines 19a–h, in turn synthesized following the method
Figure 1. 1 (WB 4101): X=O, R=R
athian): X=S, R=R =OMe, =H 21: X=O, R=R
=OMe.
1
=OMe, R
2
=H 20 (Benox-
1
R
2
1
=H,
1
4
R
2
described by Bourdais,
1
with 2-aminomethyl-1,4-
benzodioxane (18), in dry ethanol and in the presence
5
to improve the a -AR binding affinity, we decided to
of sodium carbonate.
1
modify the lead structure of 1 to propose a new class of
compounds that would allow exploration of the role
played by the arylpiperazinylalkyl moiety in the activity
Details on the synthesis of compounds 2–9 have been
1
2
reported elsewhere.
towards a -AR. In the course of our studies in the field
1
of new and potentially selective a -AR antagonists, we
1
have recently synthesized several compounds (2–9,
Results and Discussion
1
2
Table 1) structurally related to 1 and characterized by
the arylpiperazine ring attached through an ethyl or
propyl spacer to the 2-aminomethyl-1,4-benzodioxane
moiety. Since some of them showed an interesting
This work is a part of a project aimed at synthesizing
arylpiperazine derivatives displaying high affinity and
selectivity for a -AR with respect to a -AR, to be
1
2
activity (i.e., compounds 2 and 3 towards a -AR, and
potentially applied in the treatment of hypertension or
benign prostatic hyperplasia.
1
compound 9 toward a -AR, respectively), in order to
2
better define the optimum structural properties required
to interact with a-AR, new compounds (10–17, Scheme
1 and Table 1) with diverse substituents on the piper-
azine ring have been synthesized.
Being aware that both the a - and a -AR are hetero-
1
2
genic species, but taking into account that our major
interest is the synthesis of selective a -AR antagonists
1
(
with respect to a -AR antagonists), the following com-
2
We report here the synthesis of the new benzodioxane-
ments on the structural features of compounds 2–17 are
only referred to the native a - and a -adrenoceptors,
and not to their relative subtypes.
arylpiperazines 10–17 and the a -, a -adrenoceptor
1
2
1
2
blocking properties of compounds 2–17. Moreover, the
relationships between chemical features of these com-
pounds and their binding affinity data have been derived
on the basis of a five-feature pharmacophore model for a1-
AR antagonists previously built by our research group,
and consisting of a positive ionizable, three hydrophobic,
By taking as a starting point an hybridized structure
containing both the arylpiperazinylalkyl and the
benzodioxane moieties, the role of either the aryl or the
alkyl group in defining affinity for a - and a -AR was
1
2
1
3
and a hydrogen bond acceptor pharmacophore features.
investigated.
Table 1.
1 2
a - and a -Adrenergic receptors binding affinities for benzodioxane-arylpiperazine derivatives 2–17
a
Compd
n
R
i
K (nM)
a
1
-ARb
a
2
-AR
2 1
a /a
c
2
2
3
2
3
2
3
2
3
2
3
2
3
2
3
2
3
o-Methoxyphenyl
o-Methoxyphenyl
o-Chlorophenyl
o-Chlorophenyl
Phenyl
37.0Æ5.3 (64)
16.5Æ4.0 (3.9)
67.0Æ9.2
134.0Æ15.0
221.1Æ37.0
104.9Æ9.0
186.7Æ35.0
380.2Æ50.0
238.8Æ17.0
25.6Æ3.0
1.81
8.12
1.95
0.82
0.92
2.58
0.57
0.08
0.16
0.13
0.08
4.21
1.09
0.48
0.89
0.18
c
3
c
4
113.7Æ15.2 (71)
127.0Æ13.9 (67)
201.3Æ34.0 (290)
147.7Æ20.0 (110)
415.8Æ39.0 (710)
346.7Æ30.0 (280)
986.0Æ87.0 (1040)
718.2Æ74.0 (790)
359.4Æ45.0 (290)
203.0Æ14.0 (63)
269.0Æ40.0 (100)
292.5Æ35.0 (120)
177.2Æ23.0 (120)
331.0Æ74.0 (250)
0.24Æ0.05
c
5
c
6
c
7
Phenyl
2-Pyridinyl
2-Pyridinyl
c
8
c
9
10
11
12
13
14
15
16
17
p-Methoxyphenyl
p-Methoxyphenyl
o-Fluorophenyl
163.7Æ32.0
96.8Æ7.0
30.0Æ4.0
o-Fluorophenyl
855.4Æ79.0
293.1Æ25.0
140.8Æ15.0
158.5Æ16.0
61.0Æ5.0
2-Methyl-4-chloropyridazin-3(2H)-one-5-yl
2-Methyl-4-chloropyridazin-3(2H)-one-5-yl
2-Furoyl
2-Furoyl
prazosin
rauwolscine
4.0Æ0.3
aThe K
triplicate. Inhibition constants (K
concentration and K
binding data were calculated as described in the Experimental. The K
values are meansÆSD of series separate assays, each performed in
i
i
27
i
) were calculated according to the equation of Cheng and Prusoff:
3
K
i
=IC50/1+(L/K
d
) and K of [ H]-rauwolscine
d
) when [L] is the ligand
3
d
its dissociation constant. K
d
of [ H]-prazosin binding to rat cortex membranes was 0.24 nM (a
1
binding to rat cortex membranes was 4 nM (a
2
).
b
In parentheses, calculated affinity values.
c
The synthesis of compounds 2–9 has been reported elsewhere.12

R. Barbaro et al. / Bioorg. Med. Chem. 10 (2002) 361–369
363
In a parallel way, the benzodioxan-arylpiperazine deri-
vatives reported in this paper have been used as a test
set to further assess the statistical significance and the
predictive power of the pharmacophore model. In fact,
benzodioxan-arylpiperazines 2–17 belong to a class of
a1-AR antagonists never included either in the training
set or in the test set considered in our previous work to
generate the pharmacophore model. As a consequence, the
aim of the current study was the investigation of the
structural features of compounds 2–17 responsible for a1-
AR affinity and compare them to the features previously
identified with a view to contribute to the understanding of
the interaction of all these compounds with the corre-
Scheme 1. Compounds: 10 and 19a, n=2, R=p-methoxy phenyl; 11
and 19b, n=3, R=p-methoxyphenyl; 12 and 19c, n=2, R=o-fluoro-
phenyl; 13 and 19d, n=3, R=o-fluorophenyl; 14 and 19e, n=2, R=2-
methyl-4-chloropyridazin-3(2H)-one-5-yl; 15 and 19f, n=3, R=2-
methyl-4-chloropyridazin-3(2H)-one-5-yl; 16 and 19g, n=2, R=2-
sponding receptor (a -AR). In this context, we have
1
assumed that they all bind in the same way to the receptor.
2 3
furoyl; 17 and 19h, n=3, R=2-furoyl. Reagents: ethanol, K CO , Á.
Moreover, because no experimental data on the biol-
ogically relevant conformations of the selected com-
pounds are available, we resorted to a molecular
mechanics approach (the 2D–3D sketcher of Catalyst)
to build the conformational models to be used in the
fitting procedure to the pharmacophore.
The synthesis and pharmacological profile on isolated
rat vas deferens of compounds with an o-methoxy-
phenyl (2,3), o-chlorophenyl (4,5), phenyl (6,7), and 2-
pyridinyl (8,9) substituent have been previously reported
1
6
1
2
by some of us. Moreover, with the aim of obtaining
the maximum coverage in the kinds (i.e., substituents)
and relative positions (i.e., substitution pattern) of the
molecular chemical features, the benzodioxane-arylpi-
perazine class has been enlarged by synthesizing addi-
tional compounds (10–17) with a different substitution
on the piperazine ring and the alkyl chain characterized
by two or three methylene units.
In addition, due to the fact that the biological evalua-
tion of all the chiral compounds was carried out using
racemic mixtures, while the biological activity of chiral
compounds is usually due to one of the enantiomers [as
an example, (S)-1 has been reported to be much more
active than the corresponding R enantiomer], it was
arbitrarily decided to model the chiral compounds with
undefined chirality, thus allowing the program to
choose which configuration of the asymmetric carbon
atom (i.e., C2 on the benzodioxane ring) common to all
compounds 2–17 was most appropriate.
The a - and a -adrenoceptor binding affinities of
1
2
benzodioxane-arylpiperazine derivatives 2–17, expressed
as K values, have been assessed by radioligand receptor
binding assay and reported in Table 1.
i
Evaluation of how well derivatives 2–17 are able to fit
the pharmacophore highlighted that the chemical func-
tionalities of the model are all matched by the chemical
groups of 3 (Fig. 2), the most active compound of the
whole set, taken as a representative example of all ben-
zodioxane derivatives. Particularly, the arylpiperazinyl
moiety maps the region where a cluster of features,
known as crucial elements to interact with the putative
receptor, lies. While the ortho-substituted phenyl ring
(corresponding to the aromatic feature of the De Marinis’
Considering the binding data, it is possible to observe
that affinity for a -AR is markedly affected by replace-
1
ment of the o-methoxyphenyl group with all other sub-
stituents, with the p-methoxy derivatives 10 and 11
showing the lowest affinity. In the o-substituted series,
the chloro and fluoro derivatives are all less active than
the corresponding methoxy counterparts 2 and 3, with
the chloro substituent associated with a modest
enhancement in affinity with respect to the fluoro deri-
vatives (see compounds 4 and 5 vs 12 and 13).
1
7
pharmacophore model) occupies both HY1 and HY2,
the piperazine N4 nitrogen atom is located inside the
positive ionizable feature PI, also identified by all the
To further analyze, at a quantitative level, the relation-
ships between the structural properties of com-
pounds 2–17 and their relative affinity data, we have
previously proposed pharmacophore models for a -AR
1
antagonists. Among all the orientations of 3 into the
pharmacophore, it was found that in most cases the
remaining two features (HBAand HY3, respectively) of
the model are interacting with the ligand in such a way as
the secondary amine nitrogen overlaps HBA(correspond-
ing to the polar region in the De Marinis’ pharmaco-
phore), while the phenyl ring of the benzodioxane moiety
only partially matches HY3 of the pharmacophore. Affi-
nity of 3 in this orientation within the model was predicted
to be 3.9 nM versus an experimental value of 16.5 nM.
applied a pharmacophore model for a -AR antagonists
1
1
3
recently reported by our research group. Starting
from some arylpiperazine-pyridazinone derivatives and
various other molecules collected from the literature,
the model has been developed resorting to a ligand-
based drug design method, with the aim of gaining an
insight into the structural factors responsible for a1
affinity.
Aquantitative S AR has been derived on the basis of
how well each of compounds 2–17 was able to fit the
pharmacophore features of the model proposed for a1-
AR antagonists.
Afurther analysis of mapping modes of compound 3
and other benzodioxanes onto the pharmacophore
showed two additional interaction pathways. In parti-

364
R. Barbaro et al. / Bioorg. Med. Chem. 10 (2002) 361–369
cular, 3 is able to locate itself in such a manner that the
benzodioxane ring maps both HY1 and HY2, while the
o-methoxyphenyl substituent lies into HY3 (Fig. 3).
Consequently, the secondary amino nitrogen and the
piperazine nitrogen bearing the propyl chain inter-
changed their positions, becoming the first one the PI
group and the second one the HBAfeature, respec-
tively. This arrangement, derived by simply reversing
the orientation of the ligand, is not as favoured, mainly
because of the conformational strain associated to the
piperazine ring.
different substituents usually led to derivatives less
active than the parent compound. In this context, we
have performed a superposition of several benzodioxane
derivatives (taken from the literature) into the model. In
particular, the S enantiomer of 1, a potent and selective
a1-AR antagonist, lies into the pharmacophore in such
a manner that O1 interacts with HBA(Fig. 4). The
amine nitrogen and the o-methoxyphenyl substituent
match PI and HY1–HY2, respectively. Spatial location
of the condensed phenyl ring allows only a partial fit
into HY3, not in accordance with the very high affinity
of (S)-1 for a -AR reported to be 0.16, 2.5, and 0.25 nM
1
1
9
Athird different orientation was found characterized by
a conformational rearrangement mainly involving the
alkylamino spacer of the ligand. In fact, while the aryl-
piperazine group maintains the usual location within the
pharmacophore (HY1–HY2–PI), the alkylamino chain
causes the shortening of the distance between the term-
inal rings of the ligand, assuming a C-shaped con-
formation. As a consequence, the benzodioxane system
undergoes a translation resulting in a perfect match of
the oxygen atom at the 1-position (O1) to HBA.
for a1A, a , a1D, respectively. The difference in the
1B
predicted vs actual affinity of (S)-1 (predicted value
15 nM) led us to re-consider the possible size of HY3.
˚
Thus, we have increased by 0.5 A the radius of the
sphere representing this hydrophobic feature. As a
result, an improved fit (from 9 to 11) and a predicted
affinity value of 0.17 nM, in good agreement with the
experimental data, were found. Although this finding
has to be considered as a preliminary result, it suggested
that a hydrophobic region larger than HY3 might lead
to an improvement of the ligand affinity to the receptor.
Evaluation of the influence of the HY3 size on the
The last finding is worthy of further consideration on
1
8
the basis of a literature report describing that, while
O4 does not interact with the receptor by hydrogen
bond, O1 is a crucial key in defining the role of benzo-
dioxane ring in receptor binding. In fact, with the
exception of a carbonyl group, replacement of O1 with
affinity towards a -AR is currently under investigation.
1
Substitution of the oxygen atom at 4-position of
benzodioxane moiety led to benoxathian (20, Fig. 1)
with a slightly reduced activity (0.2, 4.0, and 0.4 nM for
1
9
a1A, a , a1D, respectively) with respect to (S)-1. Since
1B
both O4 of (S)-1 and S4 of 20 lie in an ‘empty’ region of
space (where no pharmacophore features are present),
the program well predicts (0.59 nM) the affinity of such a
thioderivative. In a similar way, the model also accounts
for the marked drop in affinity found for the p-methoxy
2
0
derivative (21, Fig. 1) of (S)-1, the calculated affinity
value being 340 nM. The main reason for this decreased
activity was the complete inability of 21 to match one of
the HY1–HY2 features of the model, according to the
trend found for diverse p-substituted phenyl derivatives.
Finally, with the purpose of evaluating the ability of
the pharmacophore model to discriminate between
enantiomeric pairs, the difference between predicted and
experimental values has been calculated for S and R
enantiomers of some compounds used in this study. As
Figure 2. Compound 3, the most active of the whole class of benzo-
dioxane-arylpiperazine derivatives, mapped to the pharmacophore
model for a -adrenoceptor antagonists. Pharmacophore features are
1
color coded: cyan for hydrophobics (HY1, HY2, and HY3), green for
a hydrogen bond acceptor (HBA), and red for a positive ionizable
feature (PI).
Figure 3. Alternative orientation of compound 3 within the pharma-
cophore model. This arrangement is not as favoured due to the
strained conformation of piperazine. Pharmacophore features are
color coded as in Figure 2.
Figure 4. The S enantiomer of WB4101 (1) superposed to the phar-
macophore model. The oxygen atom at the 1-position of the benzo-
dioxane ring corresponds to the HBAfeature of the model.
Pharmacophore features are color coded as in Figure 2.

R. Barbaro et al. / Bioorg. Med. Chem. 10 (2002) 361–369
365
an example, the predicted affinity value for (R)-1 was
affinity of the unsubstituted phenyl compounds versus
the corresponding o-fluoro derivatives. For this pur-
pose, it is important to note that Catalyst is unable to
take into account electrostatic effects possibly involved
in defining the activity of these phenyl and pyridine
derivatives.
4
3
.0 nM versus an experimentally determined affinity of
2
0
9.8 nM. While no substantial differences in orienta-
tion of (S)-1 and (R)-1 with respect to the pharmaco-
phore features was found, the enhanced predicted value
for affinity of (S)-1 toward a -AR was due to a slightly
1
more productive interaction with HBA.
The distance between the arylpiperazine and benzo-
dioxane moieties is influenced by the length of the
alkylamino chain acting as a spacer between the two
rings of compounds 2–17. Both the intrinsically high
In a similar way, S enantiomers of the benzodioxane-
arylpiperazine derivatives have been selected by the
program as the optimum structures to interact with the
model, although predicted values for S enantiomers are
in all cases almost identical to the affinity values calcu-
lated for R enantiomers.
3
3
conformational flexibility of the Csp –Csp poly-
methylene sequence and the length of the whole chain
are crucial parameters influencing the goodness-of-fit of
each compound to the chemical features of the phar-
macophore model. In fact, while the PI, HY1, and HY2
features are mapped by all ligands, HBAand HY3 can
be mapped only depending on the length of the poly-
methylene chain. Particularly, compounds characterized
by n=2 are estimated to be weakly active, due to diffi-
cult in mapping the HY3 feature. As an example, affi-
nity of compound 2 has been predicted 64 nM versus an
experimental value of 37 nM, while the corresponding
propyl derivative 3 has been predicted to have an affi-
nity of 3.9 nM versus an experimental value of 16.5 nM).
This finding is in agreement with the experimental data
of phenyl derivatives. In fact, with the exception of 4
and 5, among compounds bearing the phenylpiperazine
moiety, the ethyl spacer-bearing derivatives are all less
active than the corresponding propyl analogues. Alter-
native orientations of compounds bearing the ethylene
chain are characterized by an enhanced mapping to
HY3, with the piperazinyl ring undergoing a deep con-
formational rearrangement leading to a very strained
structure of the ligand.
Among the three orientations of 3 into the pharmaco-
phore, the program underscores the first one to best
rationalize the effect deriving from the substitution on
the phenyl ring. The substitution pattern on the phenyl
ring of the arylpiperazine moiety is a crucial element to
be accounted for with the purpose to rationalize the
relationships between structural properties and affinities
of compounds 2–17. Pharmacological tests highlighted
that the methoxy group at the ortho position is the
substituent associated with the highest a -AR antago-
1
nist activity within the whole set of benzodioxanes
under study. Moreover, the presence of a chloride,
fluoride, or hydrogen atom instead of the methoxy
group led in all cases to compounds about one order of
magnitude (or more) less active than 3 (Table 1). The
effect of replacement of the o-methoxy group with a
chlorine atom was difficult to rationalize on the basis of
previous work pointing out that both the two sub-
stituents are able to interact, equally weighted by the
program, with the same hydrophobic portion of the
1
3
pharmacophore model. Decreased activity associated
with 4 and 5 with respect to 2 and 3 can be justified
taking into account that, during estimation of affinity
values based on the orientation of the ligand, the pro-
gram runs with the aim of finding the highest number of
The lengthening of the ethyl chain to a propyl spacer led
to several compounds characterized by a higher con-
formational flexibility. This structural variation is
responsible for the enhanced fit to the pharmacophore
with a consequent improvement of the calculated affi-
nity values. On the contrary, in the pyridazinone and
furoyl series, compounds bearing an ethyl spacer (14
and 16) where slightly more active than the corre-
sponding propyl derivatives (15 and 17). Both pyr-
idazinone and furoyl derivatives showed the same
orientation within the model, with the benzodioxane
moiety matching both HY1–HY2. In addition, com-
pounds 14 and 15 are characterized by a PI group cor-
responding to the piperazine nitrogen bearing the alkyl
chain, while the methyl group of the pyridazinone ring was
the HY3 feature. In this orientation, both compounds
were unable to fit HBA, while alternative mappings,
characterized by the carbonyl moiety located into HBA
and the methyl group in a very partial fit to HY3, led to
predicted affinity values of 100 and 120 nM for 14 and
15, respectively. On the other hand, the amino nitrogen
of compounds 16 and 17 was the PI feature, while the
piperazine nitrogen bearing the alkyl chain and the
furane ring of 17 matched HBAand HY3, respectively.
Adifferent situation was found for 16 where a higher
affinity was predicted as the consequence of fitting
between the carbonyl group and HBA.
2
1
contacts between pharmacophore and ligand. A s a
consequence of forcing the benzodioxane ring into
HY3, the ortho substituent at the opposite terminal
portion of the molecule, with the exception of the
methoxy group, is only able to partially map one of the
HY1–HY2 features. In summary, the pharmacophore
model accounts for this trend in the biological data with
the ability of each of the above substituents to fit the
HY2 feature. Particularly, while the o-methoxyphenyl
moiety possesses optimum structural features to perfectly
map the HY1–HY2 system of the model, the o-chloro-
and o-fluorophenyl groups are only able to partially
match into HY1–HY2, with a consequent reduction in
the predicted activity (i.e., 67 and 63 nM are the calcu-
lated affinity values for compounds 5 and 13, respec-
tively). In a similar way, derivatives where the aromatic
group is unsubstituted (6 and 7) or corresponding to a
pyridine ring (8 and 9), are unable to map one of the
HY1-HY2 features, with consequent decreased activity
(
i.e., calculated values for compounds 7 and 9 are 110
and 280 nM, respectively). Based on these considera-
tions, we were unable to justify by means of the phar-
macophore model the slight improvement in a -AR
1

366
R. Barbaro et al. / Bioorg. Med. Chem. 10 (2002) 361–369
The preferential orientation of the pyridazinone and
furoyl moieties of compounds 14, 15 and 16, 17 into the
pharmacophore was similar to that previously found for
pyridazinone derivatives used to build the pharmaco-
phore itself and for prazosin, respectively. Moreover,
the ethyl spacer represents the optimum structural
requirement (with respect to the propyl chain) to allow for
a fit into both PI and HY3, according to experimental
data showing for compounds 14 and 16 a slightly
increased affinity with respect to 15 and 17, respectively.
Conclusions
Several compounds belonging to a hybridized class of
benzodioxane-arylpiperazine derivatives have been pre-
pared by variation of the alkyl spacer linking the term-
inal benzodioxane and arylpiperazine rings and by
modification of the aryl moiety bound to the piperazine.
All these compounds have been tested for their a - and
a2-AR antagonist properties.
1
7
1
Some structural features have been demonstrated to
markedly affect affinity for both a - and a -AR. In par-
Finally, in accordance with experimental data, p-meth-
oxy substituted compounds 6 and 7 have been predicted
to have the highest activity values within the whole set,
in agreement with some recent findings reporting para
substituents as unfavourable to ligand-receptor bind-
1
2
ticular, the o-methoxyphenyl group bound to the piper-
azine ring led to the best a - affinity profile. In fact,
derivatives 2 and 3 showed an interesting affinity
1
towards a -AR. On the contrary, the remaining com-
1
6
b,22
ing. As an example, the predicted affinity of 11 was
790 nM, versus an experimental value of 718 nM.
pounds, characterized by a different substitution on the
piperazine ring, are all less active than the o-methoxy-
phenyl derivatives, with p-methoxyphenyl compounds
showing the highest values in affinity (i.e., they are the
less active compounds).
From the above considerations, a summary of the
interaction pathways between benzodioxane derivatives
and the pharmacophore model for a -AR antagonists
1
can be drawn as follows. (i) Phenyl or pyridine deriva-
tives show, as a preferential orientation, the arylpiper-
azine moiety located within the HY1–HY2–HBA
system, in agreement with other arylpiperazines used to
build the model itself. On the contrary, furoyl- and
pyridazinonepiperazine undergo a reversion of the
orientation, having the benzodioxane moiety located
within HY1–HY2. As a consequence, the basic, positive
ionizable nitrogen, accessible to the receptor and easily
protonable at physiological pH, can be alternately
represented by the piperazine N4 nitrogen or the sec-
ondary amine nitrogen. (ii) The o-methoxy substituted
phenyl ring of compounds 2 and 3 is associated with the
The alkyl spacer is also important in defining affinity,
the propyl spacer being the optimum linker between the
terminal heterocycles.
Benzodioxane-arylpiperazine derivatives 2–17 were also
used to further test the predictive power of a
pharmacophore model for a -AR antagonists. In par-
1
ticular, the model was able to predict accurately the
receptor affinity of such compounds not used in the
construction of the model itself. Moreover, this model
justifies the importance of the main pharmacophore
groups (particularly, the o-methoxyphenyl moiety and
the basic nitrogen atom) as well as of their relative dis-
tance. In fact, while the substitution pattern on the
phenyl ring considerably affect the affinity properties
(with the o-methoxy substituent associated with the best
activity), spatial considerations are particularly impor-
tant for the phenyl ring on the piperazine moiety and
the basic nitrogen atom, since their spatial locations are
critical for the biological activity, as generally accepted
best activity toward a -AR. A marked decrease in
1
activity was found in the o-chloro, o-fluoro, unsub-
stituted or pyridine derivatives. As expected, the p-
methoxy substituent led to a dramatic drop in affinity.
(
iii) Apolar group corresponding to the carbonyl moi-
ety of 14–17 and the secondary amino nitrogen in the
remaining compounds, is the preferential hydrogen
bond acceptor substituent. (iv) The polymethylene chain
linking the arylpiperazine to the benzodioxane ring is an
important structural feature in determining affinity of
compounds 2–17. In fact, as demonstrated by the
pharmacophore fitting studies reported here, the alkyl
moiety determines the distance at which both the
benzodioxane and arylpiperazine rings are located, and,
accordingly, the goodness-of-fit to the pockets defined
by the HY1–HY2–PI and HBA–HY3 features, respec-
tively. These results are in good agreement with the
for a -AR antagonists.
1
Considering that the model was able to rationalize the
major structure–activity relationships for these com-
pounds, and its statistical significance has been pre-
viously highlighted, we can conclude that this
pharmacophore model may represent a powerful tool
for subsequent three dimensional quantitative struc-
ture–activity relationships (3D-QSAR) studies on the
antagonists of a -adrenoceptors.
1
generally accepted statement that an a -AR antagonist
1
interacts with the corresponding receptor by a three-
pocket binding pathway: the most important structural
Experimental
Chemistry
element for blocking a -AR is a protonated nitrogen
1
interacting with the carboxylate of an aspartic residue.
The remaining molecular portions of the ligand fit two
binding pockets almost symmetrically located with
respect to the aspartate. (v) Finally, an additional
pharmacophore element (HY3) accommodates (por-
tions of) the benzodioxane, furoyl, and pyridazinone
moieties of compounds 2–17.
Melting points were determined using a Kofler hot-stage
1
apparatus and are uncorrected. H NMR spectra were
recorded on a Bruker AC 200 MHz instrument. Chemi-
cal shift values (ppm) are relative to tetramethylsilane
used as an internal standard. Elemental analyses are

R. Barbaro et al. / Bioorg. Med. Chem. 10 (2002) 361–369
367
within Æ0.4% of theoretical values. Precoated Kiesegel
0 F plates (Merck) were used for TLC.
solvent scintillation fluid was added. The radioactivity
was counted with an Packard 1600 TR scintillation
counter. Specific binding was obtained by subtracting
nonspecific binding from total binding and was
approximated to 85–90% of the total binding.
6
2
54
Method for the preparation of compounds 10–17. An ex-
ample. N-{2-[4-(4-Methoxyphenyl)piperazin-1-yl]ethyl}-
2
-aminomethyl-1,4-benzodioxane (10). Amixture of 1-
(
(
(
0
6
2-chloroethyl)-4-(4-methoxyphenyl)piperazine
(19a)
0.25 g, 0.001 mol), 2-aminomethyl-1,4-benzodioxane
18) (0.49 g, 0.003 mol) and sodium carbonate (0.10 g,
ꢀ -Receptor binding. Cerebral cortex was dissected from
2
rat brain and the tissue was homogenized in 20 volumes
of ice-cold 50 mM Tris–HCl buffer at pH 7.7 containing
5 mM EDTA, as reported above (buffer T ). The
.001 mol) in dry ethanol was refluxed under stirring for
h. The mixture was filtered and the organic phase was
1
homogenate was centrifuged at 48,000g for 15 min at
ꢀ
evaporated under reduced pressure. The residue was
purified by chromatography on a silica gel column elut-
ing with EtOH/CH Cl (1:9) to give 1 (55%) as an oil;
4 C. The resulting pellet was diluted in 20 volumes of
50 mM Tris–HCl buffer at pH 7.7 and used in the bind-
ing assay.
2
2
1
H NMR (CDCl ) d 1.90 (s, 1H, NH), 2.45–2.60 (m,
3
6
H, 4H-pip, CH ), 2.70–2.90 (m, 4H, 2CH ), 3.00–3.10
Binding assay was perfomed in triplicate, by incubating
aliquots of the membrane fraction (0.2–0.3 mg protein)
in Tris–HCl buffer at pH 7.7 with approximately 2 nM
2
2
(m, 4H, H-pip), 3.70 (s, 3H, OCH ), 3.90–4.00 (m, 1H,
3
H-benzodiox), 4.20–4.30 (m, 2H, H-benzodiox), 6.70–
3
6
chloride had mp 140–145 C. Anal. calcd for
.90 (m, 8H, H-arom). The corresponding hydro-
[ H]-rauwolscine in a final volume of 1 mL. Incubation
ꢀ
ꢀ
C H N O 3HCl: C, 53.62; H, 6.50; N, 8.50. Found:
C, 53.20; H, 6.90; N, 8.15.
was carried out at 25 C for 60 min. Non-specific
.
binding was defined in the presence of 10 mM
rauwolscine. The binding reaction was concluded by
filtration through Whatman GF/C glass fiber filters
under reduced pressure. Filters were washed four times
with 5 mL aliquots of ice-cold buffer and placed in
scintillation vials. Specific binding was obtained by
subtracting non specific binding from total binding
and approximated to 85–90% of total binding. The
receptor-bound radioactivity was measured as described
above.
2
2
29
3
3
Method for the preparation of compounds 19a–h. An ex-
ample. 1-(2-Chloroethyl)4-(4-methoxyphenyl)piperazine
(
19a). This compound was prepared using the method
1
4
of Bourdais. Amixture of 1-(4-methoxyphenyl)piper-
azine (2 g, 0.01 mol), 1-bromo-2-chloroethane (1.79 g,
0
0
2
.012 mol) and dry potassium carbonate (1.72 g,
.012 mol) in DMF (10 mL) was stirred for 24 h at
ꢀ
5 C. The mixture was diluted with water and extracted
with CH Cl . The organic phase was dried with sodium
2
Compounds were dissolved in buffer or DMSO (2%
buffer concentration) and added to the assay mixture. A
blank experiment was carried out to determine the effect
of the solvent on binding.
2
sulfate and evaporated under reduced pressure. The
residue was purified by chromatography on a silica gel
column eluting with ethyl acetate to afford the title
1
compound as an oil. Yield: 50%. H NMR (CDCl ) d
3
2
3
3
.60–2.70 (m, 4H, H-pip), 2.80 (t, J=7 Hz, 2H, CH2),
.10–3.20 (m, 4H, H-pip), 3.65 (t, J=7 Hz, 2H, CH2),
.80 (s, 3H, OCH ), 6.80–7.00 (m, 4H, H-arom).
Protein estimation was based on a reported method,26
after solubilization with 0.75 N sodium hydroxide, using
bovine serum albumin as the standard.
3
The concentration of tested compound that produces
3
Biology
3
50% inhibition of specific [ H]-prazosin or [ H]-rau-
wolscine binding (IC ) was determined by log-probit
ꢀ1-Receptor binding. Rat cerebral cortex was homo-
genized in 20 volumes ice-cold 50 mM Tris–HCl buffer
at pH 7.7 containing 5 mM EDTA(buffer T ) in an
1
50
analysis with seven concentrations of the displacer, each
performed in triplicate. Inhibition constants (K ) were
i
2
7
ultra-turrax homogenizer. The homogenate was cen-
trifuged at 48,000g for 15 min at 4 C. The pellet (P )
calculated according the equation: K =IC /1+([L]/
i 50
ꢀ
was suspended in 20 volumes of ice-cold buffer T . It
K ) where [L] is the ligand concentration and K its dis-
d d
1
3
sociation constant. K of [ H]-prazosin binding to cortex
d
1
3
was then homogenized and centrifuged at 48,000g for
ꢀ
membranes was 0.24 nM (a ) and K of [ H]-rauwolscine
1
d
1
5 min at 4 C. The resulting pellet (P ) was frozen at
2
binding to cortex membranes was 4.0 nM (a2).
ꢀ
À80 C until the time of assay.
Computational methods
The pellet P was suspended in 20 volumes of ice-cold
2
5
binding assay was performed in triplicate by incubating
at 25 C for 60 min in 1 mL T buffer containing aliquots
2
0 mM Tris–HCl buffer at pH 7.7 (T buffer) and a1
All calculations and graphic manipulations were per-
formed on a Silicon Graphics O2 workstation by means
of the Catalyst 4.6 software package.
2
ꢀ
of the membrane fraction (0.2–0.3 mg protein) and
3
0
.1 nM [ H]-prazosin in the absence or presence of
All the compounds used in this study were built using
the 2D–3D sketcher of the program. Arepresentative
family of conformations were generated for each mole-
cule using the poling algorithm and the ‘best quality
unlabelled 1 mM prazosin. The binding reaction was
terminated by filtering through Whatman GF/C glass
fiber filters under suction and washing twice with 5 mL
ice-cold Tris–buffer. The filters were placed in scintilla-
tion vials and 4 mL Ultima Gold MN Cocktail-Packard
2
3
conformational analysis’ method. The parameter set
employed to perform all the conformational calculations

368
R. Barbaro et al. / Bioorg. Med. Chem. 10 (2002) 361–369
2
4
derives from the CHARMm force field, opportunely
2
Pharmacol. Rev. 1995, 47, 266. (c) Hieble, J. P.; Ruffolo, R. R.,
Jr. Prog. Drug Res. 1996, 47, 81.
5
modified and corrected.
3
. Price, D. T.; Schwinn, D. A.; Lomasney, J. W.; Allen, L. F.;
Caron, M. G.; Lefkowitz, R. J. J. Urol 1993, 250, 546.
. (a) Chang, D. J.; Chang, T. K.; Yamanishi, S. S.; Salazar,
The best quality conformational analysis approach has
been selected because it provides the best possible con-
formational coverage within Catalyst.
4
F. H. R.; Kosaka, A. H.; Khare, R.; Bhakta, S.; Jasper, J. R.;
Shieh, I.-S.; Lesnick, J. D.; Ford, A. P. D. W.; Daniels, D. V.;
Eglen, R. M.; Clarke, D. E.; Bach, C.; Chan, H. W. FEBS
Lett. 1998, 422, 279. (b) Daniels, D. V.; Gevel, J. R.; Jasper,
J. R.; Kava, M. S.; Lesnick, J. D.; Meloy, T. D.; Stepan, G.;
Williams, T. J.; Clarke, D. E.; Chang, D. J.; Ford, A. P. D. W.
Eur. J. Pharmacol. 1999, 370, 337.
Accordingly to some literature reports, conformational
behavior of the arylpiperazine moiety shows the het-
eroring characterized by a N1–N4 di-equatorial chair.
The structures of the conformers generated for
compounds 14–17 showed a quasi-coplanar orienta-
tion of the piperazine ring with respect to the pyr-
idazinone or furoyl moiety directly linked to it, in
agreement with previous findings on piperazine
5
1
. (a) Strader, C. D.; Sigal, I. S.; Dixon, R. A. FASEB J.
989, 3, 1825. (b) Strader, C. D.; Sigal, I. S.; Dixon, R. A.
Trends Pharmacol. Sci. 1989, (Suppl.), 26.
6. (a) Mokrosz, M. J.; Paluchowska, M. H.; Char-
akchieva-Minol, S.; Bien, S. Arch. Pharm. Pharm. Med.
Chem. 1997, 330, 177. (b) Lopez-Rodriguez, M. L.; Rosado,
M. L.; Benhamu, B.; Morcillo, M. J.; Fernandez, E.; Schaper,
K.-J. J. Med. Chem. 1997, 40, 1648. (c) Kuo, G.-H.; Prouty,
C.; Murray, W. V.; Pulito, V.; Jolliffe, L.; Cheung, P.; Varga,
S.; Evangelisto, M.; Shaw, C. Bioorg. Med. Chem. 2000, 8,
2
2b
derivatives.
In addition, twisted or even orth-
ogonal conformations of the piperazine C2–C3–C5–
C6 system with respect to the phenyl ring of the arylpi-
perazinyl moiety of the ligands 2–13, demonstrated to
be highly profitable for a -AR antagonism, were also
1
6
a
found.
2
7
263.
. (a) Bolognesi, M. L.; Budriesi, R.; Cavalli, A.; Chiarini, A.;
Conformational diversity was emphasized by selection
of the conformers that fell within 20 kcal/mol range
above the lowest energy conformation found.
Gotti, R.; Leonardi, A.; Minarini, A.; Poggesi, E.; Recanatini,
M.; Rosini, M.; Tumiatti, V.; Melchiorre, C. J. Med. Chem.
1
999, 42, 4214. (b) Quaglia, W.; Pigini, M.; Piergentili, A.;
Giannella, M.; Marucci, G.; Poggesi, E.; Leonardi, A.; Mel-
chiorre, C. J. Med. Chem. 1999, 42, 2961.
8. Graham, R. M.; Pettinger, W. A. N. Engl. J. Med. 1979,
300, 232.
The Compare/Fit command within Catalyst has been
used to predict affinity values of the studied compounds.
Particularly, the Best Fit option has been selected which
manipulates the conformers of each compound to find,
when possible, different mapping modes of the ligand
within the model. As a consequence, a value of the bio-
logical activity will be associated to each mapping mode
satisfying the constraints imposed by the location of the
pharmacophore features.
9. Lepor, H. J. Androl 1991, 12, 389.
10. Caine, M. Urol. Clin. North Amer 1990, 17, 641.
11. Kenny, B.; Ballard, S.; Blagg, J.; Fox, D. J. Med. Chem.
1997, 40, 1293.
1
2. Corsano, S.; Strappaghetti, G.; Scapicchi, R.; Marucci, G.
Arch. Pharm. Pharm. Med. Chem. 1996, 39, 468.
3. Barbaro, R.; Betti, L.; Botta, M.; Corelli, F.; Giannaccini,
1
G.; Maccari, L.; Manetti, F.; Strappaghetti, G.; Corsano, S. J.
Med. Chem. 2001, 44, 2118.
For each Compare/Fit operation, the program provides
the measure (indicated as a fit value) of how closely the
pharmacophore features correspond to the molecular
groups of the ligand.
1
1
1
1
4. Bourdais, J. Bull. Soc. Chim. Fr. 1968, 8, 3246.
5. Green, P. N.; Shapero, M.; Wilson, C. J. Med. Chem.
969, 12, 326.
6. Catalyst 4.6; Molecular Simulation, Inc. (MSI): San
Diego, CA92121, USA.
7. De Marinis, R. M.; Wise, M.; Hieble, J. P.; Ruffolo R. R.,
Jr. In The Alpha-1 Adrenergic Receptor; Ruffolo, R. R., Jr.,
Acknowledgements
1
Finantial support provided by the Italian ‘Ministero
dell’Universita e della Ricerca Scientifica e Tecnologica’
`
MURST) and Italian Research National Council of
Italy (CNR Target Project on ‘Biotechnology’ is
acknowledged.
Ed.; Humana: Clifton, NJ, USA, 1987; p 211.
1
8. Pigini, M.; Brasili, L.; Giannella, M.; Giardina
Gulini, U.; Quaglia, W.; Melchiorre, C. J. Med. Chem. 1988,
1, 2300.
`
, D.;
(
3
19. Bremner, J. B.; Coban, B.; Griffith, R.; Groenewoud,
K. M.; Yates, B. F. Bioorg. Med. Chem. 2000, 8, 201.
2
0. Bremner, J. B.; Coban, B.; Griffith, R. J. Comp.-Aid. Mol.
Des 1996, 10, 545.
1. Help on Line provided by MSI along with the program
Catalyst 4.6.
2. (a) Russo, F.; Romeo, G.; Guccione, S.; De Blasi, A. J.
2
References and Notes
2
1
. Bylund, D. B.; Eikenberg, D. C.; Hieble, J. P.; Langer,
S. Z.; Lefkowitz, R. J.; Minneman, K. P.; Molinoff, P. B.;
Med. Chem. 1991, 34, 1850. (b) Cinone, N.; Carrieri, A.;
Strappaghetti, G.; Corsano, S.; Barbaro, R.; Carotti, ?.
Bioorg. Med. Chem. 1999, 7, 2615.
23. (a) Smellie, A.; Teig, S. L.; Towbin, P. J. Comp. Chem.
1995, 16, 171. (b) Smellie, A.; Kahn, S. D.; Teig, S. L. J.
Chem. Inf. Comp. Sci. 1995, 35, 285. (c) Smellie, A.; Kahn,
S. D.; Teig, S. L. J. Chem. Inf. Comp. Sci. 1995, 35, 295.
24. Brooks, B. R.; Bruccoleri, R. E.; Olafson, B. D.; States,
Ruffolo, R. R., Jr.; Trendelenburg, U. Pharmacol. Rev. 1994,
4
2
D. E. Trends Pharmacol. Sci. 1994, 15, 167. (b) Hieble, J. P.;
Bylund, D. B.; Clarke, A. E.; Eikenberg, D. C.; Langer, S. Z.;
Lefkowitz, R. J.; Minneman, K. P.; Ruffolo, R. R., Jr.
6, 121.
. (a) Ford, A. P. D. W.; Williams, T. J.; Blue, D. R.; Clarke,

R. Barbaro et al. / Bioorg. Med. Chem. 10 (2002) 361–369
369
D. J.; Swaminathan, S.; Karplus, M. J. Comp. Chem. 1983, 4,
26. Lowry, O. H.; Rosenbrough, N. J.; Farr, A. L. J. Biol.
Chem. 1951, 193, 265.
27. Cheng, Y.; Prusoff, W. H. Biochem. Pharmacol. 1973, 22,
3099.
1
2
87.
5. Catalyst force field and potential energy functions are
described in the file REFenergy.doc.html provided by MSI
along with the program Catalyst 4.6.