Design and synthesis of specific probes for human 5-HT4 receptor dimerization studies

Article abstract of DOI:10.1021/jm050234z

Recently, human 5-HT₄ receptors have been demonstrated to form constitutive dimers in living cells. To evaluate the role of dimerization on the 5-HT₄ receptor function, we investigated the conception and the synthesis of bivalent molecules able to influence the dimerization process. Their conception is based on a model of the 5-HT₄ receptor dimer derived from protein/protein docking experiments. These bivalent ligands are constituted by two ML10302 units, a specific 5-HT₄ ligand, linked through a spacer of different sizes and natures. These synthesized bivalent ligands were evaluated in binding assays and cyclic AMP production on the 5-HT_4(e/g) receptor isoform stably transfected in C6 glial cells. Our data showed that bivalent ligands conserved a similar affinity compared to the basal ML10302 unit. Nevertheless, according to the nature and the size of the spacer, the pharmacological profile of ML10302 is more or less conserved. In view of the interest of bivalent ligands for investigating the GPCR dimerization process, these 5-HT₄ specific bivalent ligands constitute valuable pharmacological tools for the study of 5-HT₄ receptor dimerization.

Full text of DOI:10.1021/jm050234z

6220
J. Med. Chem. 2005, 48, 6220-6228
Design and Synthesis of Specific Probes for Human 5-HT4 Receptor
Dimerization Studies
Jean-Louis Soulier,†,| Olivier Russo,†,| Mireille Giner,† Lucie Rivail,† Magali Berthouze,‡ Sandrine Ongeri,†
Bernard Maigret,§ Rodolphe Fischmeister,‡ Frank Lezoualc’h,‡ Sames Sicsic,† and Isabelle Berque-Bestel†,*
UMR C8076 (CNRS), Laboratoire de Reconnaissance Mole´culaire et Synthe`ses, and INSERM U-446, Laboratoire de
Cardiologie Mole´culaire et Cellulaire, IFR-75 ISIT, Faculte´ de Pharmacie, Universite´ de Paris XI, Biocis, 5, rue Jean-Baptiste
Cle´ment, 92296 Chaˆtenay-Malabry 92296, France, and Equipe de Dynamique des Assemblages Membranaires, UMR-7565,
Universite´ Henri Poincare´, 54506 Vandœuvre les Nancy, France.
Received March 15, 2005
Recently, human 5-HT4 receptors have been demonstrated to form constitutive dimers in living
cells. To evaluate the role of dimerization on the 5-HT4 receptor function, we investigated the
conception and the synthesis of bivalent molecules able to influence the dimerization process.
Their conception is based on a model of the 5-HT4 receptor dimer derived from protein/protein
docking experiments. These bivalent ligands are constituted by two ML10302 units, a specific
5-HT4 ligand, linked through a spacer of different sizes and natures. These synthesized bivalent
ligands were evaluated in binding assays and cyclic AMP production on the 5-HT4(e/g) receptor
isoform stably transfected in C6 glial cells. Our data showed that bivalent ligands conserved
a similar affinity compared to the basal ML10302 unit. Nevertheless, according to the nature
and the size of the spacer, the pharmacological profile of ML10302 is more or less conserved.
In view of the interest of bivalent ligands for investigating the GPCR dimerization process,
these 5-HT4 specific bivalent ligands constitute valuable pharmacological tools for the study
of 5-HT4 receptor dimerization.
Introduction
5-HT4 receptor in the amyloid precursor protein (APP)
metabolism, a key gene involved in Alzheimer’s disease
(AD).8 Together with its role in learning and memory,
these results suggest that the 5-HT4 receptor may
constitute a new pharmacological target for the treat-
ment of AD.9
Although many 5-HT4 ligands were synthesized dur-
ing this past decade, only two drugs indicated for the
irritable bowel syndrome became commercially avail-
able.7 It now appears evident that a better knowledge
of GPCR signaling pathways as well as the character-
ization of the physiological role of GPCR dimerization
process could open new avenues for the design of
efficient drugs.
In this context, bivalent ligands, which were proved
to modulate the dimerization process by binding dimeric
receptors, may represent valuable pharmacological tools.
In a recent study, Bushan et al. demonstrated the
importance of bivalent ligands approach in the study
of δ-κ opioid receptor heterodimers. Their heterobiva-
lent ligands selective for heterodimeric opioid receptors
constituted probes for targeting different tissues.10,11
In this paper, we describe the design and the synthe-
sis of specific 5-HT4 bivalent ligands, based on molecular
modeling studies and more particularly on receptor-
receptor docking experiments. Those bivalent ligands
are constituted by two 5-HT4 pharmacophoric parts
linked with a spacer. Many spacers were used varying
in nature and in size. The pharmacological properties
of the synthesized bivalent ligands were evaluated.
Identification of Dimer Interfaces. To understand
the functional role of dimerization, and the structural
mechanism for cross-talk between receptors in a dimeric
complex, we planed to design 5-HT4 bivalent ligands as
The concept that most G protein coupled receptors
(GPCRs) can exist as dimeric entities is now largely
accepted. However the reason for this dimerization
remains elusive. Although dimerization or oligomeriza-
tion of GPCR has been proved to participate in receptor
function, including agonist affinity, potency and efficacy
and G protein specificity, it is difficult to conclude
whether dimerization or oligomerization is functionally
essential for all GPCRs.1-5
Elucidating the role of dimerization in the activation
process of GPCR became a new challenge which might
lead to develop novel pharmaceutical agents able to
promote activation or inhibition of GPCR signaling.
Recently, constitutive human 5-HT4 receptor dimers
were observed in living cells and membrane prepara-
tions of CHO and HEK293 cells as detected by co-
immunoprecipitation and BRET experiments. 5-HT4
receptor dimerization was not influenced by the binding
of classical 5-HT4 ligands: agonists such as serotonin,
prucalopride, ML10302 or antagonists such as
GR113808.6 Considering the ubiquitous involvement of
the 5-HT4 receptor in major diseases such as memory
disorders, gastrointestinal disorders, hypertension, atri-
al arrhythmia and dysfunction of the urinary tract, the
research of new 5-HT4-related drugs is fundamental.7
Moreover, recent data have underlined a role of the
* To whom correspondence should be addressed: Faculte´ de Phar-
macie, 5, rue J. B. Cle´ment, 92296 Chaˆtenay-Malabry Cedex, France.
Phone: 33(0)146835743, fax: 33(0)146835740, isabelle.bestel@
cep.u-psud.fr.
† UMR C8076 (CNRS).
‡ INSERM U-446.
§ Universite´ Henri Poincare´.
| Contributed equally to this work.
10.1021/jm050234z CCC: $30.25 © 2005 American Chemical Society
Published on Web 09/08/2005

Probes for Human 5-HT4 Receptor Dimerization Studies
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 20 6221
chemical tools. For that we focused first on the identi-
fication of the dimer interfaces.
Molecular modeling studies were performed and more
particularly protein-protein docking experiments for
investigating likely dimer arrangements. Docking was
realized from the three-dimensional model of 5-HT4
receptor recently published, based on the crystallized
bovine rhodopsin structure,12 and the specialized soft-
ware GRAMM (global range molecular matching), which
considers the proteins as rigid entities, was used.13 The
procedure implements an exhaustive grid search for the
receptor-receptor structure matches. For quantifying
possible interactions between two receptors, GRAMM
places each monomer in a three-dimensional grid. Then,
an exhaustive research in the three-dimensional spaces
is realized using translations and rotations. Finally, this
software evaluates the surface complementarities,
penalizes covering and orders the obtained results.
In this study we have applied the low-resolution
docking with a 6.8 Å grid step, using the special mode
for helix in privileging the hydrophobic interactions. One
of the two monomers was rotated with 10°-angle inter-
vals. For each complex, the 100 lowest-energy matches
were sorted from low to high energy and then analyzed
visually. The values of the different parameters were
determined as quasi optimal after many experiments
of validation and based to the results of Vasker et al.14,15
From the 100 structures generated by GRAMM with
the optimized parameters, two main structures of
dimers emerged with 40% of II-IV/II-IV complexes
including helices II and IV in their dimer interfaces and
40% of IV-VI/IV-VI complexes including helices IV and
VI (Figure 1). So, the two most favorable complexes
obtained by this docking procedure both involve helix
IV. This result agrees well with results of literature
which established that in GPCR A family, helix IV is
likely to be involved in receptor-receptor contacts.
Different biological experiments have identified this
transmembrane domain as potential interfaces for
dopamine D2 receptors dimerization,16 γ-opioid receptors
or C5a receptors.17 The most convincing argument
derives from the application of atomic force microscopy
to the organization of rhodopsin in its native mem-
brane.18 Models based on these pictures were inferred
to indicate that receptor-receptor links might be pro-
vided by contacts between helices IV and V. Moreover,
there is some evidence in the literature that similar
dimerization motifs in transmembrane domains may be
involved in dimerization interfaces. Thus in several
mammalian GPCRs, (AGSTP)XXX(GAS) sequence types
in transmembrane domains are known to play a key role
in receptor dimerization and can be used as a consensus
motif to identify potential dimerization sequences.19 In
the 5-HT4 receptor, analysis of the receptor sequence
revealed the presence of a A4.43XXXG4.47 motif in helix
IV and visualization of our molecular model indicated
a correct orientation of this motif to allow receptor-
receptor interactions.
Figure 1. 5-HT4 receptor dimer models obtained by protein-
protein docking experiments using the GRAMM software. A:
dimer where dimer interfaces included helices II and IV. B:
dimer where dimer interfaces included helices IV and VI.
Scheme 1
the methoxy group affording an ether bivalent ligand,
or on the amino function giving rise to bivalent ligands
linked through the anilic position of ML10302. Finally,
a link from the 4 position of the piperidine ring of
ML10302 can also be afforded.
The two complexes suggested by the GRAMM docking
procedure can be used to orient the design of bivalent
ligands. For example, in the hypothesis of receptor
dimers interfaced by helices II and IV, a convenient
positioning of bivalent ligands suggests to link ML10302
parts through the 4 position of the piperidine ring of
ML10302 (Scheme 2). In contrary, the hypothesis of
complexes IV-VI/IV-VI should require bivalent ligands
linked through the methoxy or the amino function of
the aromatic ring of ML10302.
Design of Bivalent Ligands. Bivalent ligands should
be constituted by two 5-HT4 ligands linked through a
spacer. ML10302,20 a 5-HT4 receptor specific ligand
synthesized in our laboratory was selected (Scheme 1).
This molecule allows different anchor points for the
spacer. For example, a spacer could be introduced on
We first focused on II-IV/II-IV complexes and on
bivalent ligands linked through the 4 position of the
piperidine ring of ML10302. Previous studies on
ML10302 molecules had underlined the possibility of
substitution of the basic amino part with a voluminous

6222 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 20
Soulier et al.
Scheme 2
Scheme 3a
a Reagents and conditions: (i) 4-amino-5-chloro-2-methoxybenzoic acid 2-bromoethyl ester,20 DIEA, DMF, 40-50 °C, 24 h; (ii) CF3COOH,
CH2Cl2, rt, 24 h; (iii) EDC, HOBt, NEt3, CH2Cl2, Fmoc-2-(2-aminoethoxy)ethylamine hydrochloride, or Fmoc-1-amino-3,6-dioxa-8-octanamine
hydrochloride, or Fmoc-1-amino-4,7,10-trioxa-13-tridecamine hydrochloride, 0 °C to room temperature, 24 h; (iv) 20% piperidine, DMF,
10-15 min, 4 N MeOH/HCl; (v) 3, EDC, HOBt, NEt3, CH2Cl2, 0 °C to room temperature, 24 h.
group such as fluorescent probes without penalizing the
affinity of molecules.21 These results were supported by
the site-directed mutagenesis and molecular modeling
results which showed a large pocket between helices III
and VI extended to helix VII.12 Altogether, these obser-
vations indicated that the substitution of the spacer on
this part of ML10302 should not affect the affinity of
bivalent ligands (Scheme 2).
The theoretical II-IV/II-IV complexes obtained were
used as templates in order to evaluate the optimal
length of the spacer. To identify compounds specifically
interacting with receptor dimers, we synthesized biva-
lent ligands possessing different spacer lengths: short
length spacers, insufficient for allowing the two phar-
macophoric parts to bind to receptor dimers, longer
spacers, allowing this binding via the lipid bilayer and
excessively long spacer to permit bridging of receptor
dimers. The influence of the chemical nature of the
spacers was also examined with polar oxygenated
chains, hydrophobic alkyl chains and alkyl chains
including a central cyclic core. The spacers were con-
nected through an amidic function to the 4 position of
the piperidine ring of ML10302. To introduce some
flexibility at this level, the amidic function was first
introduced directly on the piperidine ring and then
through one and two methylene groups. Taken as a
whole, three different series of bivalent ligands were
synthesized according to the nature of the spacer: (i) a
polar oxygenated spacer series (compounds 6a-c); (ii)
a
hydrophobic spacer series (compounds 10a-d,
11a-d, 12a-d, 13a-d, 16a-d and 18a-b); (iii) a
central cyclic core alkyl spacer series (compounds 23,
24, 27 and 32). In hydrophobic spacer series, two ligands
possessing only one pharmacophore head linked to
capped spacers were also synthesized, for comparison
with bivalent ligands.
Synthesis of Bivalent Ligands. Molecules 6a-c
were prepared as described in Scheme 3. In the first
step, condensation of piperidine-4-carboxylic acid tert-
butyl ester with 4-amino-5-chloro-2-methoxy-benzoic
acid 2-bromoethyl ester20 in the presence of diisoprop-
ylethylamine at 40-50 °C in dry DMF afforded 2. The
acidic hydrolysis of 2 with excess of CF3COOH in dry
CH2Cl2 gave the acid 3 with a good yield. Coupling of 3
with Fmoc-2-(2-aminoethoxy)-ethylamine hydrochloride
or Fmoc-1-amino-3,6-dioxa-8-octanamine or Fmoc-1-

Probes for Human 5-HT4 Receptor Dimerization Studies
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 20 6223
Scheme 4a
a Reagents and conditions: (i) HOOC(CH2)nNHBoc, EDC, HOBt, NEt3, CH2Cl2, 0 °C to room temperature, 24 h; (ii) 4 N MeOH/HCl,
rt, 3-5 h; (iii) 3, EDC, HOBt, NEt3, CH2Cl2, 0 °C to room temperature, 24 h; (iv) HOOC(CH2)nCOOH, EDC, HOBt, NEt3, CH2Cl2, 0 °C
to room temperature, 24 h; (v) NH2(CH2)nNH2, EDC, HOBt, NEt3, CH2Cl2, 0 °C to room temperature, 24 h; (vi) NH2(CH2)5NHBoc, EDC,
HOBt, NEt3, CH2Cl2, 0 °C to room temperature, 24 h.
Scheme 5a
amino-4,7,10-trioxa-13-tridecamine hydrochloride using
a general procedure (EDC, HOBt, and NEt3 in dry CH2-
Cl2) provided respectively 4a-c with moderate yields.
Removal of the Fmoc protecting group using piperidine
20% in DMF and subsequent treatment with MeOH/
HCl gave compounds 5a-c. Compounds 5a-c were
then coupled to 3 by the general coupling reaction as
previously described to yield dimers 6a-c.
Synthesis of molecules of hydrophobic spacer series
10a-d, 11a-d, 12a-d, 13a-d and 16a-d, is described
in Scheme 4. The synthesized molecules differ in the
length of the alkyl spacer n, in the direction of the
amidic function (CONH or NHCO) and in the distance
between the ML10302 part and the amidic function (m),
constituting five groups of hydrophobic bivalent ligands.
For the group which possesses the NHCO- - - -NHCO
motif, directly fixed on the ML10302 pharmacophoric
part, the bivalent molecules 10a-d were synthesized
from compound 7a21 by a first coupling reaction with
appropriate N-Boc protected amino acids activated with
EDC, HOBt as previously described, and, after depro-
tection of the obtained compounds 8a-d, a second
coupling reaction between the resulted compounds
9a-d and molecule 3.
For the group which possesses the NHCO- - - -OCNH
motif fixed on the ML10302 pharmacophoric part di-
rectly or through a chain of one and two carbons
respectively, bivalent molecules 11a-d, 12a-d, 13a-d
were yielded by direct coupling of 7a-c21 with suitable
acids.
a Reagents and conditions: (i) HOOC(CH)nCOOH, EDC, HOBt,
NEt3,CH2Cl2, 0 °C to room temperature, 24 h.
coupling of 3 with tert-butyl N-(6-aminohexyl)carbamate
using the general procedure afforded compound 14,
followed by a classical deprotection with MeOH/HCl and
coupling of the resulted compound 15 with 3 to afford
the bivalent ligand 16c.
The two bivalent ligands 18a,b, possessing a pipera-
zine ring instead of a piperidine ring were synthesized
according to Scheme 5, using appropriate diacides by
the same way.
For comparison of activities of the bivalents ligands,
compound 20, an analogue of compound 11d, swapping
one ML10302 unit for a benzylpiperidine group was
synthesized (Scheme 6). Coupling of the commercially
available 1-benzyl-piperidine-4-ylamine with dode-
canedioic acid monoethyl ester22 using the general
coupling conditions gave 19 in a moderate yield. The
ester 19 was saponified with LiOH in H2O/dioxane and
used without purification in a second coupling step with
7a21 using the general coupling protocol except that
CH2Cl2 was replaced by anhydrous DMF as solvent.
In the third series, four molecules were synthesized.
The bivalent ligand 23 containing a benzyl core in the
Bivalent ligands 16a-b, d, which possess the
CONH- - - -HNCO motif directly fixed on ML10302
pharmacophore, were obtained by direct coupling of 3
according to the general procedure with the suitable
diamine. This procedure applied to the synthesis of
compound 16c gave no product, for this reason this
compound was obtained after two steps. In a first step

6224 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 20
Soulier et al.
Scheme 6a
a Reagents and conditions: (i) HOOC(CH2)10COOEt,22 EDC, HOBt, NEt3, CH2Cl2, rt, 5 h; (ii) LiOH, H2O/dioxane 20:80, rt, 7 h; (iii)
7a,21 EDC, HOBt, NEt3, DMF, rt, 8 h.
Scheme 7a
a Reagents and conditions: (i) 3, EDC, HOBt, NEt3, CH2Cl2, 0 °C to room temperature, 24 h; (ii) 4 N MeOH/HCl,2 h; (iii) 3, EDC,
HOBt, NEt3, CH2Cl2/DMF, 0 °C to room temperature, 24 h.
Scheme 8a
Compound 27 was obtained via the route depicted in
Scheme 9. Reaction of benzyl-(4,6-dichloro-[1,3,5]triazin-
2-yl)-amine24 with a slight excess of 11-amino-unde-
canoic acid methyl ester25 in refluxing DMF in the
presence of DIEA afforded diester 25 with a good yield.
Saponification of the resulting compound with NaOH
in H2O/dioxane and acidification with concentrated HCl
gave the desired diacid 26. Coupling of compound 26
with 7a using the general procedure provided the
bivalent ligand 27 with a good yield.
Finally the bivalent ligand 32 was synthesized as
shown in Scheme 10 from the 1-Boc-4-aminomethyl-
piperidine hydrochloride. The general coupling reaction
with Z-Gly-OH yielded compound 28 which was further
hydrogenated affording 29. The alkylation by methyl 11-
bromo-undecanoic acid methyl ester in the presence of
Cs2CO3 in DMF gave 30, with a moderate yield. The
methyl ester was then hydrolyzed with lithium hydrox-
ide, and the resulting acid was coupled with 7a21
according to the general procedure to yield the bivalent
a Reagents and conditions: (i) 7a,21 EDC, HOBt, NEt3, CH2Cl2/
DMF, 0 °C to room temperature, 24 h.
middle of the spacer was synthesized from the 3-(Boc-
aminomethyl)-benzylamine hydrochloride as described
in Scheme 7. A first coupling step with 3 using the
general procedure afforded compound 21. Classical
removal of the Boc protecting group with HCl/MeOH
gave the hydrochloride amine salt 22 which was further
coupled with 3 according to the general procedure.
The preparation of the bivalent ligand 24, outlined
in Scheme 8, resulted from the coupling reaction (gen-
eral procedure) of 3-[6-(2-carboxy-ethyl)-dibenzofuran-
4-yl]-propionic acid23 with 7a.21
Scheme 9a
a Reagents and conditions: (i) NH2(CH2)10COOMe,25 DIEA, CH3CN, reflux, 24 h; (ii) NaOH, H2O/dioxane 8/2, rt, 3 h, then reflux 1 h;
(iii) 7a,21 EDC, HOBt, NEt3, DMF, 0 °C to room temperature, 24 h.
Scheme 10a
a Reagents and conditions: (i) Z-Gly-OH, EDC, HOBt, NEt3, CH2Cl2, 0 °C to room temperature, 24 h; (ii) Pd/C, H2, MeOH, rt, 3 h; (iii)
Br-(CH2)10-COOMe, Cs2CO3, DMF, rt, 48 h; (iv) LiOH, THF/H2O, rt, 7 h; (v) EDC, HOBt, NEt3, 7a,21 DMF, 0 °C to room temperature, 24
h; (vi) HCl/MeOH 4N, rt, 3 h.

Probes for Human 5-HT4 Receptor Dimerization Studies
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 20 6225
Table 1
no.
X1
Y
X2
spacer size (atom number)
Ki (nM)
cAMP (%)
ML10302
6a
5 ( 2.5
19 ( 9.4
3 ( 2.6
45
0
CONH
CONH
CONH
NHCO
NHCO
NHCO
NHCO
NHCO
NHCO
NHCO
NHCO
CONH
CONH
CONH
CONH
(CH2)2O(CH2)2
(CH2)2O(CH2)2O(CH2)2
CH2(CH2CH2O)3CH2
(CH2)2
NHCO
NHCO
NHCO
NHCO
NHCO
NHCO
NHCO
CONH
CONH
CONH
CONH
NHCO
NHCO
NHCO
NHCO
9
6b
12
15
6
3
6c
6 ( 1.8
0
10a
10b
10c
10d
11a
11b
11c
11d
16a
16b
16c
16d
24 ( 11.6
31 ( 6.9
7 ( 3.2
10
42
46
5
(CH2)3
7
9
14
6
7
9
14
6
(CH2)5
(CH2)10
20 ( 12.7
21 ( 1,4
5 ( 0,8
(CH2)2
47
32
26
20
3
13
5
8
(CH2)3
(CH2)5
9 ( 3,1
(CH2)10
16 ( 0,5
37 ( 12.6
13 ( 4.1
5 ( 3.5
(CH2)2
(CH2)3
7
9
14
(CH2)5
(CH2)10
8 ( 1.7
ligand 31 which was classically deprotected to afford the
bivalent ligand 32.
Biological Results and Discussion
All compounds were evaluated on C6 glial cells stably
transfected with the human 5-HT4(e/g) receptor isoform.
For some ligands, the pharmacological evaluation was
completed by studies on CHO cell lines transfected with
the same 5-HT4(e/g) receptor isoform. The affinities on
the 5-HT4 bivalent ligands were determined in binding
studies using the 5-HT4 receptor antagonist [3H]-
GR113808, and their functional properties were inves-
tigated by measuring their ability to regulate cAMP
production.
Figure 2. Concentration-effect curves for 5-HT on adenylyl
cyclase activity in C6 glial cells stably transfected with the
5-HT4(e/g) receptor isoform: 9, 5-HT alone; 2, 5-HT + 10d (200
nM).
The interest of mono-oxygenated spacer (compound
6a), or poly-oxygenated spacer (compounds 6b,c) was
to introduce a hydrophilic character at the spacer chain
level. These compounds showed good affinities (Table
1), more particularly compounds 6b,c with nanomolar
Ki close to that of ML10302 (Ki ) 5 nM). However in
contrast to ML10302, compounds 6b-c shared a clear
antagonist profile. Compared to analogous more hydro-
phobic compounds 16c,d which possess equivalent
spacer length chains (Table 1), no difference appeared
on both the affinity and the activity of bivalent ligands.
Compounds 10a-d, 11a-d and 16a-d have their
spacer chain attached to the pharmacophoric ML10302
part, directly through NHCO or CONH functions. In
this set of compounds the spacer was a polyalkyl chain
of variable size from 2 to 10 methylene units. All these
compounds have good nanomolar affinities (Table 1),
with a slight superiority for a spacer size of 9 atoms.
However we observed differences in their activity pro-
files. Thus, the partial agonist tendency of ML10302 is
more or less conserved when the two functional groups
through which the 4 position of piperidine ring is
attached to the spacer are NHCO (compounds 11a-d),
since an antagonist profile is favored when these two
functional groups are CONH (compounds 16a-d). A
mixed situation occurred when the attachment groups
are NHCO and CONH (compounds 10a-d). In this case
compounds 10b and 10c behave as partial agonists since
10d is an antagonist with Kb ) 7 nM (Figures 2 and 3).
This phenomenon is confirmed by the observation that
compounds 9d (Ki ) 13 nM, cAMP ) 50%) and 20 (Ki )
14 nM, and cAMP ) 35%) which both contain only one
pharmacophoric part with the 4 position of the piperi-
dine ring substituted with a chain through a NHCO
function, conserve the partial agonist profile of ML10302
(Figure 3A). Taken into account that a partial agonist
can behave as an antagonist in a different cellular
system, the pharmacological profiles were also evalu-
ated on CHO cells and similar results were obtained
(Figure 3B). It was also interesting to examine if the
flexibility at the attachment point level between the
spacer chain and the piperidine ring plays a sensitive
role on the biological properties of the bivalent ligands.
For that, we introduced one (compounds 12a-d) or two
(compounds 13a-d) methylene units between the at-
tachment function NHCO and the piperidine ring. All
those compounds showed nanomolar affinities, and we
did not find any differences in their biological activity
(Table 2). These results indicate that for up to 18 atoms
the size of a polyalkyl spacer could not clearly dif-
ferentiate the biological properties of bivalent ligands.
Previous results have shown that replacement of pip-
eridine ring with piperazine led to 5-HT4 ligands with
antagonist profile.26 Piperazine ring is interesting since
it allows easy derivatization of the ring in the 4 position.
However, in contrast to the piperidine-based ligands,
the synthesized bivalents compounds 18a,b showed

6226 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 20
Soulier et al.
function to the pharmacophoric parts behave like
ML10302 as partial agonists, since their analogues
attached symmetrically by a CONH function (com-
pounds 16a-d), and the oxygenated spacer-containing
compounds 6a-c behave clearly as antagonists.
In conclusion, since none of the bivalent ligands
displayed a higher affinity than ML10302,10 these
results do not provide clear information about the
possibility for any bivalent ligand to bind its two
pharmacophoric parts to the 5-HT4 receptor in its
dimeric form. It would now be very interesting to
evaluate these ligands on more pertinent biological
models including notably BRET experiments. Taken
into account that classical monovalent ligands of the
5-HT4 receptor do not influence the receptor dimeriza-
tion process,6 these bivalent ligands could provide new
useful information concerning the structure of 5-HT4
receptor dimers. Co-immunoprecipitation and BRET
experiments have revealed the presence of constitutive
dimers but the monomer-dimer proportion is not clear.6
The binding of bivalent ligands to dimeric receptors
could enhance the dimer proportion which would be
revealed by BRET signal modification.
Directed mutagenesis studies are also in progress in
order to identity the dimer interfaces. The results of
these studies and of BRET evaluation of our ligands will
orient the conception and the synthesis of other bivalent
ligands for example by linking the ML10302 parts
through its methoxy or amino group. Moreover, the
study of heterodimers has revealed the possible dimer-
ization process between adrenergic receptors and 5-HT4
receptors. In this context, the synthesis of bivalent
ligands including adrenergic and 5-HT4 pharmacophoric
parts is also envisaged.
Figure 3. Concentration-effect curves for ligands on adenylyl
cyclase activity. A: In C6 glial cells stably transfected with
the h5-HT4(e/g) receptor isoform: 9, 5-HT; 3, ML10302; [, 9d;
2, 20; 1, 10d. B: In CHO cells stably transfected with the
h5-HT4(e/g) receptor isoform: 9, 5-HT; [, 9d; 1, 10d.
Table 2
Conclusion
To understand the functional importance of the 5-HT4
receptor dimerization process, we synthesized a family
of bivalent ligands possessing two 5-HT4 pharmacoph-
oric parts linked through a spacer to the amino basic
function. Their conception was based on the molecular
modeling of the 5-HT4 receptor dimer from our pub-
lished model of the monomer receptor using protein/
protein docking experiments. In this context, the basal
monomer was ML10302, and several spacers differing
in their nature and in their size were chosen and linked
by an amidic function to the fourth position of the
piperidine ring of ML10302. Three series of bivalent
ligands were then defined, synthesized and pharmaco-
logically evaluated. However, no synthesized bivalent
ligand showed biological properties which could reflect
a specific interaction with a receptor dimer, which would
mean binding of each ML10302 part in a distinct
monomer of receptor dimers. The evaluation of the
synthesized bivalent ligands in BRET experiments is
now envisaged.
spacer size
cAMP
(%)
no.
m
X1
Y
X2
(atom number) Ki (nM)
11a
11b
11c
11d
12a
12b
12c
12d
13a
13b
13c
13d
0
0
0
0
1
1
1
1
2
2
2
2
NHCO (CH2)2 CONH
NHCO (CH2)3 CONH
NHCO (CH2)5 CONH
NHCO (CH2)10 CONH
NHCO (CH2)2 CONH
NHCO (CH2)3 CONH
NHCO (CH2)5 CONH
NHCO (CH2)10 CONH
NHCO (CH2)2 CONH
NHCO (CH2)3 CONH
NHCO (CH2)5 CONH
NHCO (CH2)10 CONH
6
21 ( 1.4
5 ( 2.8
9 ( 3.1
16 ( 9.3
7 ( 2.3
13 ( 6.5
26 ( 15.7
22 ( 13.1
5 ( 2.4
3 ( 3.2
11 ( 3.4
9 ( 4.0
47
32
26
20
16
20
20
21
15
24
31
6
7
9
14
8
9
11
16
10
11
13
18
weak affinities (Table 3). Finally, to introduce more
important modulations of the spacer size up to 29 atoms,
molecules 23, 24, 27 and 32 were designed and synthe-
sized. Independently of the chemical structure of the
central core of these molecules, results of Table 3
indicate that increasing spacer size from 9 atoms
(compound 23) to 29 atoms (compound 27) is detrimen-
tal to affinity, since no correlation of their activities can
be established. Taken together, these data show that,
compared to ML10302, the bivalent piperidine-based
ligands have generally good nanomolar affinities for the
5-HT4 receptor except for those with large spacer sizes
(15 or more atoms). However, these compounds have
different activity profiles. So, compounds 11a-d having
a polyalkyl chain attached symmetrically by a NHCO
Experimental Section
Chemistry. Melting points were determined on a Kofler
melting point apparatus. NMR spectra were performed on a
Bruker AMX 200 (1H, 200 MHz; 13C, 50 MHz) or Bruker
AVANCE 400 (1H, 400 MHz; 13C, 100 MHz). Unless otherwise
stated, CDCl3 was used as solvent. Chemical shifts δ are in
ppm, and the following abbreviations are used: singlet (s),
doublet (d), triplet (t), multiplet (m), quintet (q), broad doublet

Probes for Human 5-HT4 Receptor Dimerization Studies
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 20 6227
Table 3
(bd), broad multiplet (bm), broad triplet (bt) and broad singlet
(bs). Elemental analyses (C, H, N) were performed at the
Microanalyses Service of the Faculty of Pharmacy at Chaˆt-
enay-Malabry (France) and were within 0.4% of the theorical
values otherwise stated. Mass spectra were obtained using a
Bruker Esquire electrospray ionization apparatus.
chromatography (CH2Cl2/MeOH 90:10) to afford 0.31 g (55%)
of 4a as a resinous amber solid. Rf (CH2Cl2/MeOH 90:10) 0.30;
1H NMR (200 MHz): δ 7.80 (s, 1H), 7.75 (m, 2H), 7.58 (m,
2H), 7.40 (m, 2H), 7.32 (m, 2H), 6.25 (s, 1H), 5.88 (bs, 1H),
5.11 (bs, 1H), 4.41 (m, 4H), 4.33 (t, J ) 5.9 Hz, 2H), 4.22 (t, J
) 6.07 Hz, 1H), 3.82 (s, 3H), 3.60-3.25 (m, 8H), 3.00 (m, 2H),
2.70 (t, J ) 5.9 Hz, 2H), 2.19-1.99 (m, 3H), 1.88-1.58 (m, 4H).
Anal. (C35H41ClN4O7‚0.75H2O), C, H, N.
Biology. Membrane Preparation and Radioligand
Binding Assays. Briefly, C6 glial cells stably transfected with
h5-HT4(e) receptors, grown to confluence, were incubated with
serum-free medium for 4 h, washed twice with phosphate-
buffered-saline (PBS) and centrifuged at 300g for 5 min. The
pellet was used immediately or stored at -80 °C. The pellet
was resuspended in 10 volumes of ice-cold HEPES buffer (50
mM, pH 7.4) and centrifuged at 40 000g for 20 min at 4°C.
The resulting pellet was resuspended in 15 volumes of HEPES
buffer (50 mM, pH 7.4). The protein concentration was
determined by the method of Bradford using bovine serum
albumin as the standard.
Radioligand binding studies were performed in 500 µL of
HEPES buffer (50 mM, pH 7.4), 20 µL of the studied ligand (7
concentrations), 20 µL of [3H]GR113808 at a concentration of
0.2 nM and 50 µL of membranes preparation (100-200 µg of
protein). Nonspecific binding was determined with 10 µM
GR113808. Tubes were incubated at 25 °C for 30 min, and
the reaction was terminated by filtration through Whatman
GF/B Filter paper using the Brandel 48R cell harvester. Filters
were presoaked in a 0.1% solution of polyethylenimine. Filters
were subsequently washed with ice-cold buffer (50 mM Tris-
HCl, pH 7.4) and placed overnight in 4 mL of ready-safe
scintillation cocktail. Radioactivity was measured using a
Beckman model LS 6500C liquid scintillation counter. Binding
data (Ki) were analyzed by computer-assisted nonlinear re-
gression analysis (Prism, Graphpad Software, San Diego, CA).
The data are the results of two or three determinations in
triplicate.
Measurement of cAMP. C6 glial cells stably transfected
with human 5-HT4(e) receptors were grown to confluence and
incubated with serum-free medium for 4 h before the beginning
of the assay. Then the cells were preincubated for 15 min with
serum-free medium supplemented with 5 mM theophylline and
10 µM pargyline. 5-HT (1 µM) and/or compounds were added
and incubated for an additional 15 min at 37 °C in 5% CO2.
The reaction was stopped by aspiration of the medium and
addition of 50 µL of ice-cold perchloric acid (20%). After a 30
min period, neutralization buffer was added (HEPES 25 mM,
KOH 2 N), supernatant was extracted after centrifugation at
2000g for 5 min and cAMP was quantified using radioimmu-
Materials. DMF distilled from BaO, CH2Cl2 distilled from
calcium hydride, and usual solvents were purchased from SDS
(Paris, France). Liquid chromatography was performed on
Merck silica gel 60 (70/30 mesh), and TLC was performed on
silica gel, 60F-254 (0.26 mm thickness) plates. Visualization
was achieved with UV light and Dragendorff reagent unless
otherwise stated. Boc-â-Ala-OH, Boc-4-aminobutyric acid, Boc-
6-aminocaproic acid, Boc-11-aminoundecanoic acid, 3-(Boc-
aminomethyl)-benzylamine hydrochloride, 1-Boc-4-aminomethyl-
piperidine hydrochloride, succinic acid, glutaric acid, pimelic
acid, dodecanedio¨ıc acid, Fmoc-2-(2-aminoethoxy)-ethylamine
hydrochloride, Fmoc-1-amino-3,6-dioxa-3-octanamine, Fmoc-
1-amino-4,7,10-trioxa-13-tridecamine hydrochloride, ethylene-
diamine, 1,3-diaminopropane, tert-butyl N-(6-aminohexyl)-
carbamate, 4-amino-1-benzylpiperidine dihydrochloride and 1,
10-diaminodecane and methyl 11-bromodecane were pur-
chased from commercial sources. tert-Butyl 4-piperidine car-
boxylate,27 dodecanedioic acid monoethyl ester,22 N-benzyl-4,6-
dichloro-1,3,5-triazin-2-amine,24 methyl 11-aminododecanoate25
and 3-[6- (3-hydroxy-3-oxopropyl)dibenzo[b,d]furan-4-yl]pro-
panoic acid23 were prepared according to methods reported in
the literature. 4-Amino-5-chloro-2-methoxybenzoic acid 2-bro-
moethyl ester20 and compounds 7a-c21 and 1726 were synthe-
sized as previously described.
Representative Procedure for the Synthesis of Biva-
lent Compounds. Preparation of 2-{4-[11-(9H-Fluoren-
9-yl)-9-oxo-5,10-dioxa-2,8-diazaundec-1-anoyl]ethylpip-
eridino-4-amino-5-chloro-2-methoxybenzoate (4a). To a
solution of the trifluoroacetate salt of 3 (0.4 g, 0.85 mmol, 1
equiv) in dry CH2Cl2 (60-70 mL) and cooled at 0 °C were
added HOBt (0.22 g, 1.66 mmol, 1.95 equiv), EDC (0.21 g, 85
mmol, 1.3 equiv) and NEt3 (0.30 g, 2.97 mmol, 3.5 equiv). The
reaction mixture was stirred at this temperature for 5-10 min,
and then Fmoc-2-(2-aminoethoxy)-ethylamine hydrochloride
(0.31 g, 0.11 mmol, 1 equiv) was added. After maintaining the
mixture for 24 h at room temperature, the organic layer was
washed with 10% aqueous KHSO4 (25 mL), saturated NaHCO3
(25 mL) and brine (25 mL) and dried (MgSO4). A gum which
separated from the CH2Cl2 layer when washing with KHSO4
was immediately taken up in MeOH, and the methanolic
solution was dried (MgSO4). This methanolic solution was
added to the dried CH2Cl2 organic phase. After concentration
of the combined organic layers, the residue was purified by

6228 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 20
Soulier et al.
(14) Vakser, I. A. Evaluation of GRAMM low-resolution docking
methodology on the hemagglutinin-antibody complex. Proteins
1997, Suppl 1, 226-230.
(15) Vakser, I. A.; Jiang, S. Strategies for modeling the interactions
of transmembrane helices of G protein-coupled receptors by
geometric complementarity using the GRAMM computer algo-
rithm. Methods Enzymol. 2002, 343, 313-328.
(16) Guo, W.; Shi, L.; Javitch, J. A. The fourth transmembrane
segments forms the interface oh the dopamine D2 receptor
homodimer. J. Biol. Chem. 2003, 278, 4385-4388.
(17) Klco, J. M.; Lassere, T. B.; Baranski, T. J. C5a receptor
oligomerization. I. Disulfide trapping reveals oligomers and
potential contact surfaces in a G protein-coupled receptor. J.
Biol. Chem. 2003, 278, 35345-35353.
noassay kit (cAMP competitive radioimmunoassay, Beckman,
France). The 5-HT concentration-effect curve was calculated
using seven concentrations (10-10-10-5) alone or in the pres-
ence of compounds. The ligand concentration-effect curves were
calculated using seven concentrations (10-10-10-5).
Supporting Information Available: Chemistry experi-
mental, spectroscopic data and results from elemental analyses
of all the listed compounds. This material is available free of
charge via the Internet at http://pubs.acs.org.
References
(18) Fotiadis, D.; Liang, Y.; Filipek, S.; Sapertstein, D. A.; Engel, A.;
Palczewski, K. The G protein-coupled receptor rhodopsin in the
native membrane. FEBS Lett. 2004, 564, 281-288.
(19) Overton, M. C.; Chinault, S. L.; Blumer, K. J. Oligomerization,
biogenesis, and signaling is promoted by a glycophorin A-like
dimerization motif in transmembrane domain 1 of a yeast G
protein-coupled receptor. J. Biol. Chem. 2003, 278, 49369-
49377.
(20) Yang, D.; Soulier, J. L.; Sicsic, S.; Mathe-Allainmat, M.; Bremont,
B. et al. New esters of 4-amino-5-chloro-2-methoxybenzoic acid
as potent agonists and antagonists for 5-HT4 receptors. J. Med.
Chem. 1997, 40, 608-621.
(21) Berque-Bestel, I.; Soulier, J. L.; Giner, M.; Rivail, L.; Langlois,
M. et al. Synthesis and characterization of the first fluorescent
antagonists for human 5-HT4 receptors. J. Med. Chem. 2003,
46, 2606-2620.
(22) Roy, B. C.; Hormas, D.; Mallik, S. Synthesis of new, pyrene-
containing metal, chelating lipids and sensing of cupric ions. Org.
Lett. 2003, 5, 11-14.
(23) Diaz, H. K., J. W. The synthesis of dibenzofuran based diacics
and amino acids designed to nucleate parallel and antiparallel
beta-sheat formation. Tetrahedron Lett. 1991, 32, 5725-5728.
(24) Rathod, K. T.; Patel, P. M.; Patel, S. K.; Patel, K. C. Synthesis
and physico chemical properties based on s-triazine. Ultra Sci.
1999, 11, 36-41.
(25) Leydet, A.; Barragan, V.; Boyer, B.; Montero, J. L.; Roque, J. P.
et al. Polyanion inhibitors of human immunodeficiency virus and
other viruses. 5. Telomerized anionic surfactants derived from
amino acids. J. Med. Chem. 1997, 40, 342-349.
(26) Curtet, S.; Soulier, J. L.; Zahradnik, I.; Giner, M.; Berque-Bestel,
I.; Mialet, J.; Lezoualc’h, F.; Donzeau-Gouge P.; Sicsic, S.;
Fischmeister, R.; Langlois, M. New arylpiperazine derivates as
antagonits of the human cloned 5-HT4 receptor isoforms. J. Med.
Chem. 2000, 43, 3761-3769.
(27) Kawanishi, Y.; Ishihara, S.; Takahashi, K.; Tsushima, T.;
Hagishita, S.; Ishikawa, M.; Ishihara, Y. Synthesis and biological
evaluation of a new reversely linked type of dual histamine H2
and gastrin receptor antagonist. Chem. Pharm. Bull. 1997, 45
(1), 116-124.
(1) Bulenger, S. M., S.; Bouvier, M. Emerging role of homo- and
heterodimerization in G-protein-coupled receptor biosynthesis
and maturation. Trends Pharmacol. Sci. 2005, in press.
(2) Breitwieser, G. G protein-coupled receptor oligomerization:
implications for G protein activation and cell signaling. Circ. Res.
2004, 94, 17-27.
(3) Bai, M. Dimerization of G-protein-coupled receptors: roles in
signal transduction. Cell. Signal. 2004, 16, 175-186.
(4) Milligan, G. Oligomerization of G-protein-coupled receptors. J.
Cell. Sci. 2001, 114, 1265-1271.
(5) Hansen, J. L. Sheikh, S. P. Functional consequences of 7TM
receptor dimerization. Eur. J. Pharm. Sci. 2004, 23, 301-317.
(6) Berthouze, M. Ayoub, M.; Russo, O.; Rivail, L.; Sicsic, S.;
Fischmeister, R.; Berque-Bestel, I.; Jockers, R.; Lezoualc’h, F.
Constitutive dimerization of human serotonin 5-HT4 receptors
in living cells. FEBS Lett. 2005, in press.
(7) Langlois, M.; Fischmeister, R. 5-HT4 Receptor Ligands: Ap-
plications and New Prospects. J. Med. Chem. 2003, 46, 319-
344.
(8) Robert, S. J.; Zugaza, J. L.; Fischmeister, R.; Gardier, A. M.;
Lezoualc’h, F. The human serotonin 5-HT4 receptor regulates
secretion of non-amyloidogenic precursor protein. J. Biol. Chem.
2001, 276, 44881-44888.
(9) Maillet, M.; Robert, S. J.; Lezoualc’h, F. New insights into
serotonin 5-HT4 receptors: a novel therapeutic target for Alzhe-
imer’s disease? Curr. Alzheimer Res. 2004, 1, 79-86.
(10) Portoghese, P. S. 2000 Alfred Burger Award Address in Medici-
nal Chemistry. From Models to Molecules: Opioid Receptor
Dimers, Bivalent Ligands, and Selective Opioid Receptor Probes.
J. Med. Chem. 2001, 44, 3758-3758.
(11) Bhushan, R. G.; Sharma, S. K.; Xie, Z.; Daniels, D. J.; Portoghese,
P. S. A bivalent ligand (KDN-21) reveals spinal delta and kappa
opioid receptors are organized as heterodimers that give rise to
delta(1) and kappa(2) phenotypes. Selective targeting of delta-
kappa heterodimers. J. Med. Chem. 2004, 47, 2969-2972.
(12) Rivail, L.; Giner, M.; Gastineau, M.; Berthouze, M.; Soulier, J.
L.; Fischmeister, R.; Lezoualc’h, F.; Maigret, B.; Sicsic, S. and
Berque-Bestel, I. First exploration of the hydrophobic pocket
essential for the binding of bulky ligands on human 5-HT4
receptors. Br. J. Pharmacol. 2004, 143, 361-370.
(13) Vakser, I. A. Protein docking for low-resolution structures.
Protein Eng. 1995, 8, 371-377.
JM050234Z