MCH-R1 antagonists based on an arginine scaffold: SAR studies on the amino-terminus

Article abstract of DOI:10.1016/j.bmcl.2006.10.056

We have identified a novel series of potent MCH-R1 antagonists based on l-arginine. As predicted by computational methods, there was an activity dependence on the π-electronic character of the aromatic systems corresponding to the amino-terminus of these molecules. These results have enhanced our understanding of the MCH-R1 receptor and the potential for a predictive homology model.

Full text of DOI:10.1016/j.bmcl.2006.10.056

Bioorganic & Medicinal Chemistry Letters 17 (2007) 832–835
MCH-R1 antagonists based on an arginine scaffold:
SAR studies on the amino-terminus
*
´
´
Jose Mendez-Andino, Anny-Odile Colson, Daniel Denton, Maria C. Mitchell,
Doreen Cross-Doersen and X. Eric Hu
Procter & Gamble Pharmaceuticals, 8700 Mason-Montgomery Road, Mason, OH 45039, USA
Received 11 September 2006; revised 13 October 2006; accepted 23 October 2006
Available online 25 October 2006
Abstract—We have identified a novel series of potent MCH-R1 antagonists based on L-arginine. As predicted by computational
methods, there was an activity dependence on the p-electronic character of the aromatic systems corresponding to the amino-
terminus of these molecules. These results have enhanced our understanding of the MCH-R1 receptor and the potential for a
predictive homology model.
ꢀ 2006 Elsevier Ltd. All rights reserved.
Estimates suggest that over 30% of the US adult popu-
lation is overweight and that obesity-related health care
cost is in the range of $100 billion per year.1 Obesity has
become a major health concern worldwide since many
countries are also experiencing increases in obese popu-
lation. These trends have led to an increased number of
drug discovery programs directed to find anti-obesity
therapies.2,3
focused on variations at the amino-terminus, carboxy
terminus, and the guanidine group were performed to
identify the key pharmacophore contact points leading
to MCH-R1 antagonist activity. The work presented
herein focuses on the optimization of the amino-termi-
nus of these arginine-based MCH-R1 antagonists
(Fig. 1).7
Optimization of the amino-terminus requires an under-
standing of its role in MCH-R1 binding and function.7
The G-protein-coupled melanin-concentrating hormone
receptor 1 (MCH-R1) has received significant attention
in recent years as a potential target for effective anti-
obesity therapies.3 It has been suggested that CNS-locat-
ed MCH-R1 is involved in biological processes related
to mammal feeding behaviors and energy expenditure.4
Small molecule MCH-R1 antagonists are being heavily
pursued by many laboratories trying to find an effective
drug-molecule for the treatment of obesity.5
H
H
N
NH2
NH
N
NH2
NH
CF3
CF3
O
O
H
N
H
N
N
N
H
H
O
O
O
1
2
Aided by molecular modeling studies of various truncat-
ed analogs of the natural ligand peptide for human
MCH-R1, we identified arginine-based peptidomimetics
1 and 2 as potent MCH-R1 antagonists in the early stag-
es of our research efforts.6 Subsequently, SAR studies
IC50 = 4 1 nM
IC50 = 5 1 nM
H
H
N
NH2
NH
N
NH2
NH
CF3
CF3
O
O
H
N
H
N
N
N
H
H
H
O
O
Keywords: Melanin-concentrating hormone; MCH-R1 antagonists;
Obesity; Cation–p interactions.
H
3
4
*
Corresponding author. Tel.: +1 513 622 0304; fax: +1 513 622
1433; e-mail: mendezandino.jl@pg.com
IC50 = 112 33 nM
IC50 = 763 85 nM
Present address: Genome Research Institute, University of Cincin-
nati, 2180 E. Galbraith Road, Cincinnati, OH 45237, USA.
Figure 1. MCH-R1 antagonists based on an arginine scaffold.
0960-894X/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.10.056

´
J. Mendez-Andino et al. / Bioorg. Med. Chem. Lett. 17 (2007) 832–835
833
Simple structural changes around this region, as exem-
plified in compounds 3 and 4, led to a significant loss
in biological activity. We initially hypothesized that
the interaction was hydrophobic in nature, but our re-
sults did not support that assumption and demanded
an alternative explanation.
to hypothesize that the delocalized p-electron-rich xan-
thene and fluorene groups in 1 and 2 were key for
MCH-R1 antagonist activity and that they were partic-
ipating in cation–p interaction with arginine residues of
the receptor. Based on our model, the cation–p interac-
tion only occurs over one of the phenyl rings of the sym-
metric xanthene unit while the additional phenyl ring is
only necessary to achieve the required spatial arrange-
ment and electronic properties of the aromatic ring
system.
To help generate new hypotheses, further docking
experiments were performed in our homology model
of the MCH-R1 receptor. The model of the receptor
was built using the known human MCH-R1 receptor
amino acid sequence (SWISS_PROT Primary Accession
No. Q96S47) mapped onto the crystal structure of rho-
dopsin.8 Details of the construction are presented else-
where.6b,9 The model was subsequently refined using
inhouse and published site-directed mutagenesis stud-
ies,10 as well as SAR data generated in our laboratories.
Individual ligands were then docked manually and ener-
gy minimized (SYBYL Version 6.7. Tripos: St. Louis,
MO, USA) within the putative binding site of the recep-
tor located between transmembrane helices (TMH) 3, 5,
6, and 7. At the bottom of this site lies Asp 123 (TMH3)
which has been shown to be essential for cationic ago-
nist and antagonist binding at homologous GPCRs.11
For the antagonist–receptor complex, the charge-rein-
forced H-bond between the protonated nitrogen atom
of the ligand and the carboxyl of Asp 123 was a required
interaction for docked poses to be considered further.
Upon docking of the arginine-based compound 1
(Fig. 2), our model suggests that two arginine residues,
Arg197 (extra-cellular loop 2) and Arg284 (TMH6),
are key components of the binding site where the aro-
matic groups putatively interact.
In order to test this hypothesis, we focused our efforts on
finding replacements for the xanthene and fluorene
groups based on cation–p interaction theory. Further-
more, we aimed at producing effective MCH-R1 antag-
onists with lower molecular weight and/or solubility
properties suitable for in vivo dosing. We therefore envi-
sioned that a heterocyclic electron-rich naphthalene
analog could mimic the electronic properties of the xan-
thene and fluorene rings while also reducing the molec-
ular weight and potentially improving aqueous
solubility.
The magnitude of the cation–p interaction depends on
distance, orientation, and is directly proportional to
the delocalization of p electrons (aromaticity).13 The or-
der of aromaticity for heteroatomic naphthalene analogs
on the basis of the electronegativity of heteroatoms is as
follows: benzothiophene > indole > benzofuran.14 How-
ever, it is well known that indoles display great cation–
p-stacking capabilities in combination with the ability of
the NH bond to engage in additional H-bonding
interactions.13c,d
It has been suggested in the literature that the guanidine
unit of an arginine can engage in strong cation–p inter-
actions with electron-rich aromatic functionalities in
aqueous environments.12 It is worth pointing out that
our MCH-R1 model predicts that this interaction occurs
on the extra-cellular side of the receptor. This informa-
tion, along with careful analysis of our SAR data, led us
In order to understand, the activity dependence on the
p-electronic character of the aromatic systems, we iden-
tified a series of p-excessive and p-deficient naphthalene
analogs. Isoquinolines and quinolines are a common
example of p-deficient naphthalene analogs, while ben-
zofurans, benzothiophenes, and indoles represent p-aex-
cessive naphthalene analogs.14 Spatial arrangements
based on substitution patterns were part of our survey
as well. Based on these criteria, a series of compounds
containing naphthalene-analog units were prepared
(Table 1).
The compounds needed for our study were easily acces-
sible by a four-step synthetic route from readily avail-
able starting materials (Scheme 1).15 The synthesis
initiated with the coupling of the L-arginine derivative
5 with 4-trifluoromethylbenzylamine followed by cleav-
age of the t-butylcarbamate protecting group. The
resulting primary amine was coupled to the correspond-
ing naphthoic acid analog followed by the deprotection
of the guanidine functionality to yield the desired com-
pounds 7.
As predicted, the p electron-excessive series produced the
more potent compounds, whereas, the p-deficient series
primarily led to less potent antagonists (Table 1).16 The
indole analog 7a was one of the most active compounds.
It was suggested by our model (Fig. 3) to form p–cation
interactions with both arginines in addition to forming a
Figure 2. Bound conformation of compound 1 in the h-MCH-R1
model.

´
J. Mendez-Andino et al. / Bioorg. Med. Chem. Lett. 17 (2007) 832–835
834
Table 1. SAR studies:16 examples of naphthalene analogs and electronic properties
p Electron-excessive systems p Electron-deficient systems
Structure IC50 (nM) Structure IC50 (nM)
20 10
Substitution patterns
IC50 (nM)
Structure
96
8
212 76
O
O
O
NH
N
7m
7n
7g
7a
O
O
25 17
311 75
443 118
904 52
243 30
512 104
952 80
NH
F
O
O
F
7h
7b
7c
O
O
45
78
4
6
S
N
H
O
7o
7i
O
O
S
O
N
NH
O
7j
7p
7d
O
O
O
O
313 184
622 344
H
1034 330
1008 189
N
7k
S
7e
7f
7q
O
O
O
O
O
1814 387
1797 594
N
7l
7r
H
HN NH2
NH
HN
N
HN
O
N
NO2
NO2
NH
NH
CF3
O
c, d
a, b
81-87%
CF3
H
N
O
H
N
OH
Ar
N
H
O
N
51-65%
H2N
O
H
O
7
6
5
Scheme 1. Reagents and conditions: (a) 4-trifluoromethylbenzylamine, EDCI, HOBT, NMM, DMF; (b) TFA/CH2Cl2 (1:1); (c) ArCOOH, EDCI,
HOBT, NMM, DMF; (d) 5% Pd/BaSO4, H2, AcOH/MeOH (1:4).
hydrophobic interaction with F289. It is worth noting
that arginines are known to participate in p–cation inter-
action in two limiting geometries: T-shaped and paral-
lel.13c Both of these geometries are illustrated with the
docked poses of compounds 1 and 7a (Fig. 3). The indole
compound 7a not only offers a good biological activity,
but it also has a lower molecular weight than the bench-
marks 2 and 3. Furthermore, its significantly improved
aqueous solubility (433 lg/mL) over both prior analogs
(7–14 lg/mL) made it readily formulable for subsequent
in vivo studies.
to mimic compound 1. Synthesis and subsequent testing
of these compounds showed a 10-fold increase in func-
tional activity for 7n compared to 7r. Although there
was no improvement over compound 1, the correct pre-
diction of the relative functional activity of these two
compounds infers a certain degree of reliability of our
model.
We report here the design, modeling, and synthesis of
novel, effective MCH-R1 antagonists based on an argi-
nine scaffold. In these systems, the indole unit (7a)
proved to be an effective replacement for the xanthene
and fluorene groups displaying similar biological activi-
ty, but improved aqueous solubility over 1 and 2. This
work, which was led by p–cation interaction theory aid-
ed by docking experiments within a putative receptor
binding model, suggests that rational design employing
in silico approaches contributes to efficient optimization
It is interesting to note that upon docking of compound
7r, the naphthyl ring positioned itself off-the-plane of
the limiting geometries to form optimal p–cation inter-
actions with the arginines, hence suggesting poor mimics
of either system 1 or 7a. The addition of a methylene
linker (as in 7n) satisfactorily positioned the ring system

´
J. Mendez-Andino et al. / Bioorg. Med. Chem. Lett. 17 (2007) 832–835
835
Ulven, T.; Little, P. B.; Norregaard, P. K.; Hogberg,
T. Bioorg. Med. Chem. Lett. 2004, 14, 5075.
5. (a) Kanuma, K.; Omodera, K.; Nishiguchi, M.; Funako-
shi, T.; Chaki, S.; Semple, G.; Tran, T.-A.; Kramer, B.;
Hsu, D.; Casper, M.; Thomsen, B.; Sekiguchi, Y. Bioorg.
Med. Chem. Lett. 2005, 15, 3853; (b) Ulven, T.; Frimurer,
T. M.; Receveur, J.-M.; Little, P. B.; Rist, Ø.; Nørregaard,
P. K.; Ho¨gberg, T. J. Med. Chem. 2005, 48, 5684.
6. (a) Chen, Z.; Kende, A.; Colson, A.-O.; Mendezandino, J.;
Ebetino, F.; Bush, R.; Hu, X. Synth. Commun. 2006, 36,
473; (b) Mieling, G. E.; Mieling, K. K.; Bush, R. D.;
Colson, A.-O. US 2005170433 A1, 2005.
7. For examples of p-rings in molecular recognition, see:
Sulpizi, M.; Carloni, P. J. Phys. Chem. B 2000, 104, 10087,
and references therein.
8. Palczewski, K.; Kumasaka, T.; Hori, T.; Behnke, C. A.;
Motoshima, H.; Fox, B. A.; Le Trong, I.; Teller, D. C.;
Okada, T.; Stenkamp, R. E.; Yamamoto, M.; Miyano, M.
Science 2000, 289, 739.
9. Colson, A.-O.; Perlman, J. H.; Smolyar, A.; Gershengorn,
M. C.; Osman, R. Biophys. J. 1998, 74, 1087.
10. Macdonald, D.; Murgolo, N.; Zhang, R.; Durkin, J. P.;
Yao, X.; Strader, C. D.; Graziano, M. P. Mol. Pharmacol.
2000, 58, 217.
Figure 3. Bound conformation of 1 (magenta) and 7a (green) in the
MCH-R1 model.
11. (a) Gether, U. Endocr. Rev. 2000, 21, 90; (b) Cavalli, A.;
Fanelli, F.; Taddei, C.; De Benedetti, P. G.; Cotecchia, S.
FEBS Lett. 1996, 399, 9.
12. (a) Tsou, L. K.; Tatko, C. D.; Waters, M. L. J. Am. Chem.
Soc. 2002, 124, 14917; (b) Flocco, M. M.; Mowbray, Sl.
J. Mol. Biol. 1994, 235, 709.
of in vitro active series. Further optimization will be
pursued, including the addition of new contact residues
within the pharmacophore, the modeling of explicit sol-
vent on the extra-cellular side of the receptor and the
exploration of alternative binding orientation.
13. (a) Gallivan, J. P.; Dougherty, D. A. J. Am. Chem. Soc.
2000, 122, 870; (b) Minoux, H.; Chipot, C. J. Am. Chem.
Soc. 1999, 121, 10366; (c) Gallivan, J. P.; Dougherty, D.
A. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 9459; (d)
Dougherty, D. A. Science 1996, 271, 163.
Acknowledgments
The authors extend their gratitude to Dr. Frank H. Ebe-
tino and Dr. John A. Wos for the invaluable technical
discussions and assistance.
14. Gupta, R. R.; Kumar, M.; Gupta, V. Heterocyclic
Chemistry I; Spinger-Verlag: Germany, 1998.
15. (a) For compound 6 see Ryan, J. W.; Chung, A. Eur. Pat.
Appl., 78703, 11 May 1983; (b) 7a–r: a mixture of the
corresponding nitro-guanidine compound (0.25 mmol) and
5% Pd/BaSO4 (15 mg) in 20% AcOH/MeOH (3 mL) was
stirred for 8–12 h under H2 atmosphere. The mixture
diluted with MeOH (5 mL) and filtered through a syringe-
filter (0.45 lm). The filtrates purified by preparative HPLC
(Polaris C18-A 10l, 250 · 500 R, 1% TFA-water/acetoni-
trile as eluent) to yield the desired products as TFA
salts.
References and notes
1. The Surgeon General’s Call to Action to Prevent and
Decrease Overweight and Obesity 2001, Public Health
Service: Office of the Surgeon General, US Department of
Health and Human Sevices: Rockville, MD, 2001;
www.surgeongeneral.gov/topics/obesity.
2. (a) Powell, D. R. Obesity Rev. 2006, 7, 89; (b) Fong, T. M.
Expert Opin. Invest. Drugs 2005, 14, 243.
3. Dyke, H. J.; Ray, N. C. Expert Opin. Ther. Patents 2005,
15, 1303.
4. (a) Kowalski, T. J.; McBriar, M. D. Expert Opin. Invest.
Drugs 2004, 13, 1113; (b) Receveur, J. M.; Bjurling, E.;
16. MCH functional antagonist was detected using HEK 293
cells that expressed the MCH-R1 (SLC-1) receptor in a
firefly luciferase reporter assay. Cells were incubated for
4 h in the presence of 25 nM MCH and varying concen-
trations of drug of interest. Receptor activation was
measured by luminescence.