A screening library for peptide activated G-protein coupled receptors. 1. The test set

Article abstract of DOI:10.1021/jm040084c

One subset of the G-protein coupled receptor (GPCR) super-family is that which is activated by a peptide carrying an obligatory positively charged residue (GPCR-PA⁺). This subclass is exemplified by receptors for melanocortins, GnRH, galanin, MCH, orexin, and some chemokine receptors variously involved in eating disorders, reproductive disorders, pain, narcolepsy, obesity, and inflammation. Using the methods described in this study, a region of chemical property space enriched in GPCR ligands was identified. This information was used to design and synthesize a "test" library of 2025 single, pure compounds to sample portions of this property space associated with GPCR-PA⁺ ligands. The library was evaluated by high-throughput screening against three different receptors, rMCH, hMC4, and hGnRH, and found to be highly enriched in active ligands (4.5-61-fold) compared to a control set of 2024 randomly selected compounds. In addition, the analysis suggested that about 7000 compounds will be necessary to complete the sampling of this GPCR-PA⁺ ligand-rich region and to better define its borders.

Full text of DOI:10.1021/jm040084c

6864
J. Med. Chem. 2004, 47, 6864-6874
A Screening Library for Peptide Activated G-Protein Coupled Receptors.
1. The Test Set
Karine Lavrador,† Brian Murphy,‡ John Saunders,*,† Scott Struthers,‡ Xiaochuan Wang,† and John Williams†
Chemistry and Biology Departments, Neurocrine Biosciences, 12790 El Camino Real, San Diego, California 92130
Received April 16, 2004
One subset of the G-protein coupled receptor (GPCR) superfamily is that which is activated by
a peptide carrying an obligatory positively charged residue (GPCR-PA+). This subclass is
exemplified by receptors for melanocortins, GnRH, galanin, MCH, orexin, and some chemokine
receptors variously involved in eating disorders, reproductive disorders, pain, narcolepsy,
obesity, and inflammation. Using the methods described in this study, a region of chemical
property space enriched in GPCR ligands was identified. This information was used to design
and synthesize a “test” library of 2025 single, pure compounds to sample portions of this property
space associated with GPCR-PA+ ligands. The library was evaluated by high-throughput
screening against three different receptors, rMCH, hMC4, and hGnRH, and found to be highly
enriched in active ligands (4.5-61-fold) compared to a control set of 2024 randomly selected
compounds. In addition, the analysis suggested that about 7000 compounds will be necessary
to complete the sampling of this GPCR-PA+ ligand-rich region and to better define its borders.
Introduction
tors and also the belief that the design of non-peptide
ligands may now be a tractable problem, thereby
overcoming the well-documented limitations of peptides
themselves as drugs.5 In addition, many of the new
orphan GPCRs may turn out to be opportunities for
therapeutic intervention and those that are activated
by peptides would naturally fall into this subcategory.
Stimulation of receptors linked to G-protein activation
represents a primary mechanism by which cells sense
changes in their external environment and convey that
information to the cytosol through various effector
mechanisms.1,2 Historically, GPCR’s have represented
a “gold mine” for drug discovery, and over 30% of
currently approved prescription medicines act as either
agonists or antagonists at such sites3. With some
notable exceptions, most successes have been achieved
by drugs interacting at receptors for the simpler ligands,
particularly the monoamines; the purpose of this study
was to facilitate the design and synthesis of new drugs
that compete with the more complex peptide ligands.4
The GPCRs have been classified1 into five categories
(A, rhodopsin-like; B, secretin-like; C, metabotropic
glutamate-like; D, fungal pheromone; E, c-AMP) on the
basis of extensive phylogenetic studies, and these
categories have been further subdivided into subclasses
on the basis of the level of protein sequence homology
(e.g., A-I, melanocortin; B-II, parathyroid). Those
receptors that are activated by peptide ligands do not
neatly fall into one of these subcategories but are
distributed within subclasses A and B. The current
study was focused on receptors within these subclasses
that are activated by positively charged peptide ligands
exemplified by receptors for melanocortins (MC, A-I),
gonadotropin releasing hormone (GNRH, A-III), brady-
kinin (A-IV), melanin concentrating hormone (MCH,
A-V), calcitonin gene related peptide (CGRP, B-I), and
vasoactive intestinal peptide (VIP, B-III). The in-
creased interest in GPCRs activated by endogenous
peptides is a reflection of the growing number of gene
sequences that have now been assigned to such recep-
To date, strategies for the discovery of non-peptide
ligands4,6 have included high-throughput screening of
random, large compound collections and/or combinato-
rial libraries, progressive replacement of amide bonds
in fragments of peptides thought to be critical for
binding, and the synthesis of novel templates putatively
mimicking a presumed secondary structural feature,
such as a â-turn mimetic. While all of these approaches
are valid and have had some notable successes, they do
not offer a systematic approach for the family of recep-
tors under review. Central to the current study is the
concept that the “property space” associated for ligands
for peptide-activated GPCRs is definable and forms only
a small fragment of what is referred to as a “druglike”
space. Here we describe in detail how this space may
be characterized and hence associated with a designed
combinatorial “test” library consisting of around 2000
molecule that is highly enriched in ligands to these
receptors. Furthermore, we show that the designed
library performed significantly better than a random set
of equal size selected from our corporate compound
collection when tested against three relevant receptors.
It is essential to appreciate that the objective of this test
set, and eventually of the completed library, is to define
an area of drug space for GPCR-PA+. Once this has
been defined, specific areas can subsequently be highly
populated with novel molecules targeting a specific
receptor in an iterative approach. Therefore, it is not
anticipated that potent molecules for a given receptor
will be found directly in this library.
* Corresponding author. Tel: 858 617 7637. E-mail: jsaunders@
neurocrine.com.
† Chemistry Department.
‡ Biology Department.
10.1021/jm040084c CCC: $27.50 © 2004 American Chemical Society
Published on Web 12/01/2004

A Screening Library GPCR-PA+
Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27 6865
Computational Chemistry Methods
We have defined “drug space” computationally by an
analysis of drugs and druglike molecules known in the
literature and representative, biologically active “in-
house” compounds selected from over 20 historical
discovery programs at Neurocrine; collectively this set
totals 81 560 and the space it defines is termed “NBI-
drug space” (NBI-DS). To delineate GPCR-PA+ prop-
erty space as a portion of the property space occupied
by all drugs, a database composed of 630 molecules
active against over 40 distinct GPCRs was constructed
from data available in the literature and from Neuro-
crine’s proprietary compound database. So as not to
populate the set with the vast number of ligands
described for the “classical” monoaminergic receptors,
which were the subject of intense study during the
1970-1990 era, or for specific, well-studied GPCR-PA+
(e.g. somatostatin7), only representative structures for
each receptor were included.
Using the DiverseSolutions software package, BCUT
metrics8,9 were calculated for NBI-DS. Examination of
different combinations of metrics allowed us to select a
five-dimensional space (H-bond donor and acceptor, two
metrics of polarizability, and charge) in which the
compounds are widely distributed. To accelerate the
analysis of large virtual libraries, where spatial loca-
tions of millions of compounds must be calculated and
compared, we have typically employed cell-based meth-
ods. With this approach, each of the five axes is divided
into 10 bins resulting in the partitioning of the entire
space into 100 000 individual cells. Comparing the
location or index of the occupied cells can be used to
measure the diversity/similarity between compounds or
collections of compounds. Numbers of compounds within
a given cell or range of cells is also a convenient measure
of sampling density.
Importantly, we have introduced the concept of ana-
lyzing neighboring cells which, in the context of five-
dimensional space, requires that each occupied cell has
35 - 1 ) 242 neighbors. Empirically, this is important
for the following reasons: (1) The objective of the
screening library is to (almost) guarantee multiple hits
against any GPCR-PA+ target; with drug space for a
specific receptor so defined, subsequent, more dense
sampling focused on this space will lead to the nano-
molar ligands required of drug candidates. Thus, while
highly active molecules reside within a few cells or even
a single cell, we have shown that progressively less
active compounds (but still active!) are layered in
neighboring cells. (2) Different levels of receptor pro-
miscuity may determine the volume of drug space for a
given receptor; it is well-established, for example, that
the dopamine D4 receptor is highly promiscuous based
on the ease with which antagonists can be made. (3)
Compounds residing in one cell may lie at, or close to,
the boundaries of that cell so that compounds in the
nearest neighboring cell may actually be closer than if
they were located at the extremities of the occupied cell.
The locations of compounds known to be active
against GPCRs activated by positively charged peptide
ligands projected into three of the five dimensions of
this diversity space is shown graphically in Figure 1A.
Analysis of their cell occupancies together with other
categories of ligands in the full five-dimensional space
Figure 1. (A) Set of known non-peptide ligands for GPCR-
PA+ (filled circles) mapped into three dimensions of BCUT
space vs the WDI-NBI (gray dots; representative sampling
only). (B) Projection of the test set (crosses) into drug space
using the BCUT methodology each compared to GPCR-PA+
space (filled circles). (P ) polarizability axis).
Table 1. Composition of the GPCR Training, Test, and
Random Sets
no. of
cells
occupied +
neighbor
cells
% drug
space
(NBI-DS)
no. of
ligand set
ligands occupied
all drugs + NBI
(NBI-DS)
81560
8506
91080
91
all GPCR
630
201
397
165
34
97
79
1169
352
506
19225
8727
19
9
monoamines
monoacids
35
2853
3
GPCR-PA+
GPCR-PA-
random set
random basic set
test set
111
6437
6
106
4050
4
2024
1401
2025
31098
15174
10692
31
15
11
is presented in Table 1. The various GPCR ligand
classes occupy only small regions of this chemical space
as compared to the original 81 560 compounds used to
define its extent. Thus, GPCR ligands occupy 397 cells,
whereas the broader collection, the NBI drug space set,
occupies 8506 cells. If neighboring cells are included,
all GPCR ligands occupy only 19% of drug space; in

6866 Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27
Lavrador et al.
Figure 2. Templates selected for virtual library enumeration (points of monomer attachment are via NH or COOH groups).
Numbers located near to the reacting functional group indicate an association with a specific monomer listed in Table 2. In some
instances, because of symmetry in the template, this nomenclature is redundant.
addition, GPCR-PA+ ligands form a distinct subspace
(6%). An important goal of this preliminary investiga-
tion was then to evaluate whether these smaller regions
of GPCR and GPCR-PA+ space reflect a true mapping
of properties common to these receptor classes and are
predictive for designing new ligands or are simply due
to the smaller numbers of compounds used to define the
regions, compared to the overall space.
The five BCUT metrics discussed above were calcu-
lated for all of the virtual compounds (Table 2) in order
to evaluate their location relative to the GPCR-PA+
space defined above. Virtual libraries selected for syn-
thesis by the design algorithm were first explored
synthetically by generation of a small (typically 20
compound) trial library, and synthetic failure resulted
in rejection of the template. A substitute template that
closely mirrored the drug space defined by the rejected
templates was then selected. Of the 19 templates
(Figure 2) that were subjected to computational analy-
sis, seven were utilized in library production (boxed in
Figure 2) to form the basis of the 2025 member test set.
Synthetic Chemistry. The design principle for the
test set required that a basic nitrogen atom, which will
be predominantly protonated at physiological pH, be
present in each molecule. As a result, the synthetic
chemistry was restricted to reactions that preserved at
least one basic nitrogen atom in the original core.
Combinatorial libraries were prepared around tem-
plates 3, 4, 6, 9, 13, 14, and 18 utilizing synthetic
procedures analogous to those described in Schemes 1
and 2. The template was orthogonally protected and
synthesized on a 20-g scale. Side chain reagents were
then reacted with the unprotected functional group and
the intermediates were purified via automated flash
chromatography.10 The final compounds were then
purified and characterized by automated preparative
high performance liquid chromatography coupled to a
mass spectrometer.11 Each was prepared on a 10-mg
scale and only compounds that were greater than 85%
pure were screened to avoid misinterpretation of bio-
logical data. Validated hits were purified to >95% purity
prior to Ki determination.
Virtual Library Enumeration. There were 19
templates in the Neurocrine template database with at
least one basic center in the central scaffold (Figure 2)
that were selected for computational analysis so that
this key feature would be present in all of the final
molecules. Electronic lists of commercially available
reagents (alkyl halides, aldehydes, carboxylic acids, and
sulfonyl halides) that serve as side chains in each of
these templates were prepared using the Available
Chemicals Directory (ACD). Reagents with reactive
functionality that could not be blocked with a protecting
group, potentially toxic moieties (polyaromatics, Michael
acceptors, etc.), or reagents that would give rise to
compounds in the virtual library with molecular weights
greater than 600 were discarded from the set. Clusters
within these lists were established so that a subset of
each reagent list could be selected, since the virtual
molecules were stored as structural data (sd) files and
electronic storage space was limited. Combination of
templates and reagents lists yielded virtual molecules
that spanned the NBI-drug space. The reagent lists
would be fully expanded in the event of a positive hit
from the screening process (i.e. as part of the optimiza-
tion process). The virtual libraries were assembled with
the aid of a commercially available computer program
(Afferent Structure 3.0) to yield a virtual collection
summarized in Table 2.

A Screening Library GPCR-PA+
Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27 6867
Table 2. Summary of the Virtual Library
available
monomer 1
NBI-VL
monomer 1
available
monomer 2
NBI-VL
monomer 2
available
monomer 3
NBI-VL
monomer 3
virtual
library
template
1
2
RCHO
RCl
666
459
459
634
459
282
459
719
666
666
666
80
100
100
100
100
634
459
634
158
634
634
666
RCl
459
666
634
459
666
459
666
459
282
282
719
80
NA
305694
305694
291006
291006
305694
129438
305694
330021
187812
187812
478854
640000
800000
800000
800000
640000
422244
291006
291006
72522
RCHO
RCO2H
RCl
NA
3a
3b
3c
4
RCl
NA
RCO2H
RCl
RCOCl
RCl
RNH2
RCHO
RCHO
RCHO
RNH2
RCO2H
RCl
RCO2H
RCl
RCO2H
RCl
RCO2H
RSO2Cl
RCO2H
RCO2H
RCHO
NA
RCHO
RCl
RCHO
RCl
RCOCl
RCOCl
RNH2
RNH2
RCl
NA
NA
5
NA
6
NA
7
NA
8
NA
9
NA
10a
11a
12aa
12ba
12ca
13
14
15
16
17
18
19
RCO2H
RCHO
RNH2
RNH2
RNH2
NA
100
80
80
80
80
100
100
80
RCO2H
RCHO
RCHO
RCHO
RCO2H
RCl
80
666
634
459
459
666
666
459
NA
NA
RCl
NA
RCHO
RCHO
RCl
NA
422244
422244
305694
9025685
NA
NA
total
a Template has three positions of diversity.
Scheme 1. Compounds Derived from Template 3 and Synthetic Routesa
a Reagents: (a) R1CHO, Na(OAc)3BH, DCE, rt, 4 h; or (2-bromoethyl)benzene or 1-bromo-3,5,5-trimethylhexane, DMF, Hunig’s base,
40°C, 12 h; (b) TFA, DCM, 1 h, rt; (c) R2CO2H, EDCI, HOBT, DMF/NMP, rt, 12 h; (d) 5-(4-chlorophenyl)-2-furaldehyde, Na(OAc)3BH,
DCE, rt, 12 h.
Scheme 2. Synthetic Chemistry Associated with Template 4a
a Reagents: (a) R1CO2H, EDCI, HOBT, DCM, rt, 4 h, 40-85%; (b) TFA, DCM, rt, 1 h, 100%; (c) R2Br, Hunig’s base, 40 °C, 16 h,
50-85%.
Synthetic procedures for two templates (3, 4), which
afforded several of the hits for the MCH1-R, are
discussed in more detail. Derivatives of template 3 were
prepared (Scheme 1) from the monoprotected diamine
3-Boc amino pyrrolidine. Orthogonal protection of the
primary amine was required in order to conduct a
selective N-alkylation of the secondary amine, and the
Boc protecting group was chosen due to the fact that it
is removed rapidly under acidic conditions at room
temperature. The intermediates 20a and 20b were
prepared by reductive amination of 3-Boc amino pyr-
rolidine with the corresponding aldehydes in the pres-
ence of diisopropylethylamine (3 equiv) and sodium
triacetoxyborohydride.12 The intermediates 20c and 20d
were respectively prepared by N-alkylation of 3-Boc
amino pyrrolidine with (2-bromoethyl)benzene and 1-bro-
mo-3,5,5-trimethylhexane in the presence of diisopro-
pylethylamine in DMF at 40 °C. Deprotection of 20a-d
was accomplished by treatment with trifluoroacetic acid
in dichloromethane, and then the intermediate primary
amines 21a-c were reacted with the corresponding
carboxylic acids in the presence of EDCI and HOBT to
yield the final compounds exemplified by 3a-3g. Di-
chloroethane solutions of amines 21b and 21d were
added to the corresponding aldehyde in the presence of
diisopropylethylamine (3 equiv) and sodium triacetoxy-

6868 Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27
Lavrador et al.
Table 3. Distribution of Confirmed Hits for HTS for the Test,
borohydride. The resultant mixtures were agitated at
room temperature for 12 h to yield the diamines 3h and
3i.
Random, and Random Basic Setsa
confirmed hits
random
hit rate (%)
random
The synthesis of the active compounds derived from
2-aminomethylpyrrolidine template 4 is outlined in
Scheme 2. Acylation of the monoprotected diamines (R)-
and (S)-2-aminomethyl-1-Boc-pyrrolidine with the cor-
responding carboxylic acids in DMF in the presence of
EDCI and HOBT at room temperature for 12 h yielded
the intermediate amides (R)- and (S)-22a,b. Deprotec-
tion of (R)- and (S)-22a,b was accomplished with TFA
in dichloromethane, and the product secondary amines
(R)- and (S)-23a,b were then reacted with either 3,4-
dichlorobenzyl bromide or phenethyl bromide in 3 equiv
of diisopropylethylamine and N-methylpyrrolidinone as
solvent to provide (R)- and (S)-4a,b.
test random
basic
set
test random
basic
set
receptor set
hMC4 61
rMCH1 123
GnRH
set
set
set
1
6
2
5
5
1
3.00
6.10
0.44
0.05
0.30
0.10
0.35
0.35
0.07
9
a Confirmed hits represent compounds that displaced 50% or
more binding in both the original HTS and the confirmation assay.
The hit rate was determined as (no. confirmed hits/total no. of
compounds tested) × 100. The test set, random set, and random
basic set contained 2025, 2024, and 1401 compounds, respectively.
a “hit” and their activity was confirmed by repeating
them in duplicate on a separate day. Compounds that
displaced 50% or more of the specifically bound radio-
ligand in the follow-up assay were defined as “confirmed
hits” (Table 3). Active compounds were then subjected
to 12-point dose-response curve analysis to determine
binding affinities (typical curves are exemplified for
MCH1-R in Figure 3) and the resulting Kis from
templates 3 and 4 are listed in Table 4. A second random
set of 1401 compounds, which all contain a basic
nitrogen, was selected from Neurocrine’s corporate
collection and used as a basic/random control library.
This random basic set was tested under identical
conditions to the test and random sets.
As an initial proof of concept for the design process,
the number of hits identified in the designed set was
significantly greater than those obtained from the
random sets for each of the three receptors (Table 3).
Hit enrichment rates ranged from 4.5-fold (GnRH-R)
to 61-fold (MC4-R) when compared with the random
set. Although the random basic set picked up more hits
than the random set for MC4, the designed set was still
more impressive, even allowing for the smaller size of
the former. Thus, for MC4-R, the 2025 compound test
set gave a number of hits that would have required
screening more than 120000 compounds of a typical
corporate collection. The absolute number of hits varied
from 9 for the GnRH receptor to 123 for the MCH1
receptor.
On the basis of cell occupancy (Table 1), it is clear
that the designed set (506 and 10 692 cells occupied and
occupied plus neighboring cells, respectively) is less
diverse than the random set (1169 and 31098). This was
verified by a standard diversity/similarity analysis,
whereby MACCS keys were used as the fingerprint and
Tanimoto coefficients used as the similarity measure.
The method was used to calculate the similarity of every
possible pair of compounds in each set with the result
that the designed set was least diverse (mean similarity
0.56), the random set most diverse (0.33), and the
random basic set (0.48) somewhere between the two.
Structure-Activity Observations: MCH1-R. The
current library was not designed to establish structure-
activity predictions for each receptor; the intent was
more that the set should provide a starting point from
which such predictions can be made. Since many of the
highest affinity ligands that bound the MCH receptor
identified from this library originated from templates
3 and 4 (Table 4), some observations may be made.
SNAP-7941,16 a known selective and highly potent
Results
Screening Results. The 2025 compounds in the test
set were arrayed in 96-well plates and dissolved in
DMSO to a standard concentration of 15 µM. A “random
set” of 2024 compounds was selected from Neurocrine’s
corporate screening collection and used as a control
library. For practical convenience, it was possible to
retrieve and assay only plates of compounds rather than
discrete samples for the “random set”. Each of 23
random plates selected contained 88 compounds per
plate. We were concerned that the designed set may
generate more hits than the random set, simply because
the ubiquitous presence of a positive charge in the
former may bias the outcome. For that reason, we also
wished to compare the designed set’s performance
against a second set that was still random other than
having an obligatory basic center. Fortuitously, we were
in a position to make this comparison because all
compounds made at Neurocrine are simultaneously
assayed in all current screens, since we expect some
“cross-talk” between the GPCRs within the area of
interest. Thus, as before, compounds (not plates on this
occasion) were selected at random from a list of over
30 000 compounds bearing a basic center. The caveat
here is that compounds originally designed as ligands
for either of the three receptors employed in this study
were deselected from the pick for that receptor, which
resulted in a 1401 member “random basic” set. Both the
test and random sets of compounds were simultaneously
tested for activity in three high-throughput screens
(HTS) against receptors of high therapeutic potential.
The melanocortin-4 receptor (MC4-R, obesity13,14), mela-
nin concentrating hormone receptor (MCH1-R, obesity
and depression15,16), and gonadotropin-releasing hor-
mone receptor (GnRH-R, endometriosis, uterine fi-
broids, prostate cancer17,18) are members of the class A
subclass of GPCRs activated by peptides bearing a
positive charge. Compounds were screened in singlet at
15 µM in radioligand binding assays in standard HTS
mode in 96-well format. Each plate contained six “total”
wells containing no compounds and six “background”
wells containing an excess of a cognate competitor to
define specific binding for each plate (total cpm -
backgound cpm). The Z′ factor of each assay plate was
approximately 0.5, verifying the suitability of the assays
for HTS.19 Compounds that displaced 50% or more of
the specifically bound radioligand were designated as

A Screening Library GPCR-PA+
Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27 6869
Figure 3. Representative dose-response curves of MCH-R hits from the test set. The Ki values are derived from the IC50 values
of inhibition using the Cheng-Prusoff equation.
Table 4. Summary of Binding Data for MCH Hits from
Templates 3 and 4
philic than the corresponding region in the amino amide
derivatives. This suggests that the optimal relationship
between side chain structure and MCH1-R affinity of
the template 3 diamines will be distinct to that of
aforementioned amino amides 3a-g.
In general there were fewer and less active com-
pounds derived from template 4 versus template 3,
perhaps suggesting that the extra conformational free-
dom is unfavorable entropically. Where a direct com-
parison can be made, such as between 3g (Ki ) 0.36 µM)
and (S)-4a (Ki ) 6.0 µM), almost a 20-fold difference in
activity is observed. A slight improvement in MCH1-R
binding is realized when the R1 and R2 substituents are
phenethyl and phenpropyl, respectively [(R)-4b, Ki )
3.1 µM]. The difference in the electronic nature and
shape of the side chains as well as a difference in the
absolute stereochemistry of the template [(R)-4a and
(S)-4b are less active than the corresponding enanti-
omers] again suggests distinct SAR within this sub-
series of MCH1-R ligands.
Using these “hints” of SAR, the synthesis of more
focused libraries is underway; it is clear that both
templates 3 and 4 could be used as starting points for
further lead optimization. The designed library adds yet
another dimension; however, actives are spread across
more than one series of compounds (i.e. derived from
several templates). This facet strengthens the resulting
computational model, the next step being to derive 3-D
pharmacophores and further increases the chance of
being able to optimize activity below 10 nM following
further iterations of design, synthesis, and assay. In the
context of lead optimization, it is important to note that
the set contains 1036 and 864 mono- and diamines,
respectively (the remainder are triamines), and since
all the hits are within these two classes, these properties
should not present any difficulties.
MCH1-R
binding,
compd
Y
R1
R2
Ki (µM)a
SNAP-7941
0.002
((0.001)b
>10
>10
0.98
>10
3a
CO PhCH2
CO PhCH2CH2
CO 3,4-di-Cl-C6H3CH2 PhCH2
CO 3,4-di-Cl-C6H3CH2 PhCH2CH2
PhCH2
PhCH2
3b
3c
3d
3e
CO 3,4-di-Cl-C6H3CH2 4-PhO-C6H4CH2 0.51
3f
CO 3,4-di-Cl-C6H3CH2
CO 3,4-di-Cl-C6H3CH2
CH2 Ph2CHCH2
A
B
C
C
B
B
0.63
0.36
1.7
3g
3h
3i
CH2
D
5.0
(R)-4a
(S)-4a
(R)-4b
(S)-4b
CO 3,4-di-Cl-C6H3CH2
CO 3,4-di-Cl-C6H3CH2
CO PhCH2CH2
9.8
6.0
PhCH2CH2CH2 3.1
CO PhCH2CH2
PhCH2CH2CH2 >10
a Value is an average of three determinations unless reported
otherwise. b n ) 7.
MCH1-R ligand, was used as a positive reference in
the biological assays for Ki determination. Benzyl (3a)
and phenethyl (3b) substituents at the secondary ni-
trogen of the amino amide derivative of template 3 did
not result in a compound that bound MCH1-R, but the
addition of chlorine atoms at the 3 and 4 positions of
the benzyl group did result in a compound with moder-
ate affinity for the MCH1 receptor (3c, Ki ) 0.98 µM).
Increasing the distance between the core and the phenyl
group at R2 did not improve the biological activity, as
seen in compound 3d, but increased binding affinity was
observed when an additional phenoxy substituent was
appended to the R2 phenyl group (3e, Ki ) 0.51 µM).
Other related high-affinity MCH1-R compounds from
the library incorporate a bicyclic moiety, as illustrated
in compounds 3f and 3g, which possess Ki values of 0.63
and 0.36 µM, respectively.
Discussion
While the random set was shown to occupy 1169 cells
totaling 31098 when neighboring cells were included,
the test designed library (Figure 1B) physically occupied
506 and, with neighbors, 10692 cells, respectively. The
screening hits from each of the three receptors have
been projected into GPCR-PA+ space (Figure 4) noting,
of course, that we are unable to visualize simultaneously
the remaining two dimensions. Of the 193 total hits, it
was observed that MCH1-R, MC4-R, and GnRH-R
respectively occupied 73, 35, and 7 cells, and with the
exception of one GnRH-R hit, all are located in cells
occupied by GPCR-PA+ or their neighbors. This is very
suggestive that GPCR-PA+ space does indeed exist and
Reduction of the amide bond in template 3 also
resulted in diamines that are moderate affinity ligands
for MCH1-R. Two diamine derivatives, 3h (Ki ) 1.7
µM) and 3i (Ki ) 5 µM), share an arylfuranyl group that
is one atom closer to the core in the R2 region of the
template, and the R1 region is significantly more lipo-

6870 Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27
Lavrador et al.
Figure 4. Screening hits (squares) from the test and random sets projected into drug space against a background of GPCR-PA+
ligands (open circles). (A) MCH-R hits from the test set; (B) GnRH-R hits from the test set; (C) MC4-R hits from the test set;
(D) hits from the random set [(1) region of GnRH random hits; (2) MC4, and (3) MCH)].
can assist ligand design for this subclass of receptors.
Of the 73 cells occupied by the hits for MCH1-R, 15
and two overlapped with MC4-R and GnRH-R, re-
spectively. For the MC4 receptor, 35 occupied cells had
two cells in common with GnRH-R. Two cells had hits
for all three receptors, and this information may be
relevant for receptor selectivity. The GnRH hit referred
to above as an “exception” is a consequence of the
convenience of matrix synthesis, which results in sev-
eral compounds being prepared that fall outside of the
desired space. In contrast, the small number of hits from
the random set arises from a significant, albeit low,
occupation of GPCR-PA+ space.
The data presented above show that a test library of
2025 compounds synthesized on the basis of their
location within a GPCR-PA+ ligand-enriched region of
drug space is highly enriched (4.5-61-fold) in ligands
for three receptors of this subfamily; the absolute hit
rate ranged from 0.45% to 6%. If one postulates the
existence of a discrete region of chemistry space that is
enriched in ligands for this class of receptor, as is
suggested by our preliminary data, then this variability
in hit rate and enrichment ratio could be due to uneven
sampling of either compounds or receptors in these
experiments. As was discussed above, we observed that
individual receptors appear to prefer compounds that
reside in discrete subregions. It may be that the popula-
tion of compounds in the initial set was not evenly
distributed into regions of chemical space preferred by
each of the three test receptors. Alternatively, the three
receptors selected may have intrinsic characteristics
(such as the number, size, shape, or complexity of
potential ligand binding pockets) that result in greater
or lesser degrees of ligand selectivity. Further, by
querying the library with only three receptors, each of
which can detect ligands only from their cognate sub-
region, one would not expect to delineate the entire
range of the GPCR-PA+ ligand-enriched region.
The design and synthesis of the completed library are
now underway to attempt to fill, wherever possible, the
remaining unoccupied cells together with their neigh-
bors. What is already certain is that the task cannot be
completed by another round of matrix-based parallel
synthesis; rather a cherry-picking approach will be
adopted so as to avoid the synthesis of large numbers
of compounds falling outside the desired space and to
avoid overpopulating cells already covered. It is expected
that around 7000 compounds will be the minimum
required to cover the relevant space (GPCR-PA+ oc-
cupied plus neighboring cells ) 6437) although, with
some redundancy, this may be significantly larger. On
the other hand, initial analysis suggests that it may not
be possible to occupy some neighboring cells with real
molecules, and this will reduce the number of com-
pounds to be made.
Experimental Section
Synthetic Chemistry. General Methods. Proton NMR
spectra were obtained with a Bruker 500 MHz spectrometer
at Numega Labs. Inc., San Diego, CA, and chemical shifts are

A Screening Library GPCR-PA+
Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27 6871
reported in parts per million (δ) downfield from tetramethyl-
silane as internal standard. Matrix-assisted laser desorption/
ionization (MALDI) FTMS experiments were performed on an
IonSpec FTMS mass spectrometer at The Scripps Research
Institute, San Diego, CA. Samples were irradiated with a
nitrogen laser operated at 337 nm, and the laser beam was
controlled by a variable attenuator and focused on the sample
target. A Sciex API-150EX mass spectrometer operating in
electrospray positive ionization mode was used to obtain [MH+]
ion of the target molecules. Optical activity was measured at
589 nm, 25 °C, using a Perkin-Elmer 341 polarimeter. Key
intermediates were synthesized in normal laboratory glass-
ware and purified by automated flash chromatography (Jones
Chromatography, Lakewood, CO). Parallel synthesis experi-
ments were performed under a nitrogen atmosphere using a
liquid handler (Gilson 215, Middleton, WI) to deliver each
solution into a deep-well microtiter plate with 96 750-µL glass
inserts. The plates were agitated on a vortexing heating block
(J-Kem, St. Louis, MO). Evaporation of the volatiles in vacuo
was performed using a Savant Discovery SpeedVac. Purifica-
tion of final compounds was conducted on a prep-LCMS Dionex
system, using an Alpha C18 30 × 75 mm reverse phase column
at a flow rate of 45 mL/min. The mobile phase was a gradient
of 95/5 to 5/95 A/B, where A ) H2O-0.1% TFA, and B ) CH3-
CN-0.1% TFA. All compounds after purification were reana-
lyzed on a reverse phase HPLC-MS system (DIONEX with
ESI+ ionization mode) and shown to be at least 85% pure on
the basis of both UV wavelengths (220, 254 nm) and total ion
current (TIC) from the mass spectrum.
carbamic acid tert-butyl ester (1.3 g, 52.0%). MALDI-FTMS
calcd for C18H36N2O2: 312.5 [MH+]. Found: 313.2841. 1H NMR
(500 MHz, methanol-d4): δ 0.92 (s, 9H), 0.94 (m, 1H), 0.96 (d,
J ) 6.5 Hz, 3H), 1.12 (AB, J ) 5.8 and 14.0 Hz, 1H), 1.28 (AB,
J ) 2.4 and 14.0 Hz, 1H), 1.48 (s, 9H), 1.50 (m, 4H), 2.20 (m,
1H), 2.50 (m, 4H), 2.69 (m, 1H), 2.86 (m, 1H), 4.08 (m, 1H).
13C NMR (500 MHz, methanol-d4): δ 23.2, 28.7, 29.0, 30.4,
31.9, 32.3, 38.8, 50.6, 52.5, 54.0, 55.5, 55.7, 61.4, 80.2, 158.0.
1-(3,4-Dichlorobenzyl)pyrrolidin-3-ylamine, TFA Salt
(21a). To a stirred solution of [1-(3,4-dichlorobenzyl)pyrrolidin-
3-yl]carbamic acid tert-butyl ester 20a (350 mg, 1.0 mmol) in
CH2Cl2 (2 mL) was added trifluoroacetic acid (2 mL). The
resultant solution was stirred at room temperature for 2 h and
the CH2Cl2/TFA solvent was removed under reduced pressure.
The residue was coevaporated with CH2Cl2, dissolved in MeOH
(10 mL), and purified by prep-LCMS to yield the desired TFA
salt as a colorless liquid (211.6 mg, 86.4%). MALDI-FTMS
calcd for C11H14Cl2N2: 245.0607 [MH+]. Found: 245.0596. 1H
NMR (500 MHz, methanol-d4): δ 2.08 (m, 1H), 2.52 (m, 1H),
3.13 (m, 1H), 3.29 (m, 1H), 3.42 (m, 2H), 4.04 (m, 1H), 4.20
(m, 2H), 4.88 (s, 2H), 7.42 (dd, J ) 8.2 Hz, 1.9 Hz, 1H), 7.58
(d, J ) 8.2 Hz, 1H), 7.71 (d, J ) 1.9 Hz, 1H). 13C NMR (500
MHz, methanol-d4): δ 30.0, 50.0, 53.7, 57.4, 58.3, 131.0, 132.3,
133.2, 134.1, 144.5.
1-(2,2-Diphenylethyl)pyrrolidin-3-ylamine, TFA Salt
(21b). The compound was prepared according to the procedure
given immediately above using 20b in place of 20a (176.0 mg,
66.0.%). MALDI-FTMS calcd for C18H22N2: 267.1856 [MH+].
1
Found: 267.1851. H NMR (500 MHz, methanol-d4): δ 2.07
(m, 1H), 2.49 (m, 1H), 3.38 (m, 1H), 3.57 (m, 1H), 3.64 (m,
1H), 4.00 (m, 3H), 4.48 (t, J ) 7.8 Hz, 1H), 4.86 (s, 1H), 7.24-
7.31 (aromatic, 10H). 13C NMR (500 MHz, methanol-d4): δ
29.6, 49.6, 49.8, 55.2, 58.2, 60.9, 128.8, 129.0, 130.3, 130.4,
141.9, 142, 157.
[1-(3,4-Dichlorobenzyl)pyrrolidin-3-yl]carbamic Acid
tert-Butyl Ester (20a). To a solution of 3-(tert-butoxycarbo-
nylamino)pyrrolidine (1.5 g, 8.06 mmol, 1 equiv) and N,N-
diisopropylethylamine (2.80 mL, 2.0 equiv) in dichloroethane
(20 mL) was added 3,4-dichlorobenzaldehyde (1.455 g, 8.06
mmol). The solution was stirred at room temperature for 10
min and then NaBH(OAc)3 (2.570 g, 12.09 mmol) was added.
The resultant mixture was allowed to stir for an additional
10 h and then the reaction mixture was diluted with CH2Cl2
(10 mL) and washed with water (10 mL). The organic phase
was dried over sodium sulfate and concentrated under reduced
pressure. The crude product was purified by automated flash
chromatography (gradient elution: 30% EtOAc in hexane to
60% EtOAc in hexane) to afford [1-(3,4-dichlorobenzyl)pyrro-
lidin-3-yl]carbamic acid tert-butyl ester as an oil (1.80 g,
66.0%). MALDI-FTMS calcd for C16H22Cl2N2O2: 345.1131
[MH+]. Found: 345.1136. 1H NMR (500 MHz, CHCl3-d): δ 1.43
(s, 9H), 1.58 (m, 1H), 2.27 (m, 2H), 2.59 (m, 1H), 2.75 (m, 1H),
3.53 (m, 2H), 4.16 (s, 1H), 4.82 (s, 1H), 7.14 (d, J ) 8.0 Hz,
1H), 7.36 (d, J ) 8.0 Hz, 1H), 7.41 (s, 1H). 13C NMR (500 MHz,
CHCl3-d): δ 28.6, 32.8, 50.0, 52.5, 59.0, 61.0, 79.6, 128.1, 130.5,
130.7, 131.2, 132.6, 139.2, 155.5.
1-(3,5,5-Trimethylhexyl)pyrrolidin-3-ylamine, TFA Salt
(21c). The compound was prepared according to the procedure
given immediately above using 20c in place of 20a (255.0 mg,
81.5%). MALDI-FTMS calcd for C13H28N2: 213.2325 [MH+].
1
Found: 213.2320. H NMR (500 MHz, methanol-d4): δ 0.93
(s, 9H), 0.97 (m, 1H), 1.05 (d, 3H), 1.16 (AB, J ) 6.0 and 14.0
Hz, 1H), 1.27 (AB, J ) 3.3, 3.4, 1.3 Hz, 1H), 1.62 (m, 2H), 1.75
(m, 1H), 2.2 (m, 1H), 2.65 (m, 1H), 3.30 (m, 3H), 3.6 (m, 1H),
4.15 (m, 1H), 4.89 (m, 3H).
N-(1-Benzylpyrrolidin-3-yl)-2-phenylacetamide (3a).
To a solution of phenylacetic acid (40 mg, 0.3 mmol) in NMP
(1.5 mL) was added a solution of commercially available
1-benzyl-pyrrolidin-3-ylamine (44 mg, 0.25 mmol) and N,N-
diisopropylethylamine (130.0 µL, 0.75 mmol) in DMF (1.0 mL).
This followed by a solution of EDCI (86 mg, 0.45 mmol) and
HOBT (59 mg, 0.45 mmol) in NMP (1.0 mL). The reaction
mixture was stirred at room temperature for 16 h and then
purified by prep-LCMS to give the desired product as the TFA
salt (62.0 mg, 85.0%). MALDI-FTMS calcd for C19H22N2O:
295.1805 [MH+]. Found: 295.1808. 1H NMR (500 MHz, CHCl3-
d): δ 1.77 (m, 1H), 2.12 (m, 1H), 2.41 (m, 1H), 2.89 (m, 1H),
3.02 (m, 1H), 3.36 m, 1H), 3.46 (s, 2H), 3.70 (m, 1H), 4.08 (m,
1H), 4.26 (m, 1H), 4.79 (m, 1H), 7.25 (m, 5H), 7.43 (m, 5H).
13C NMR (500 MHz, CHCl3-d): δ 30.6, 43.5, 47.2, 52.6, 58.8,
59.0, 115.7, 118.0, 127.1, 128.8, 128.9, 129.2, 129.4, 129.8,
130.4, 130.6, 135.2, 171.7.
[1-(2,2-Diphenylethyl)pyrrolidin-3-yl]carbamic Acid
tert-Butyl Ester (20b). This compound was prepared accord-
ing to the procedure described above using diphenylacetalde-
hyde in place of 3,4-dichlorobenzaldehyde (2.5 g, 84.5%).
MALDI-FTMS calcd for C23H30N2O2: 367.2380 [MH+].
1
Found: 367.2362. H NMR (500 MHz, methanol-d4): δ 1.40
(s, 9H), 1.50 (m, 1H), 2.10 (m, 1H), 2.40 (m, 1H), 2.48 (m, 1H),
2.68 (m, 1H), 2.77 (m, 1H), 3.12 (d, J ) 7.6 Hz, 2H), 3.31 (m,
1H), 3.98 (m, 1H), 4.16 (t, J ) 7.6 Hz, 1H), 7.14-7.30
(aromatic, 10H). 13C NMR (500 MHz, methanol-d4): δ 28.8,
32.6, 50.8, 52.0, 54.4, 62.1, 62.6, 80.2, 127.5, 129.2, 129.6, 145.3,
158.0.
N-(1-Phenethylpyrrolidin-3-yl)-2-phenylacetamide (3b).
The compound was prepared according to the procedure
described for compound 3a using 1-phenethylpyrrolidin-3-
ylamine20 in place of 1-benzylpyrrolidin-3-ylamine (54.9 mg,
70%). MALDI-FTMS calcd for C20H24N2O: 309.1961 [MH+].
[1-(3,5,5-Trimethylhexyl)pyrrolidin-3-yl]carbamic Acid
tert-Butyl Ester (20c). To a solution of 3-(tert-butoxycarbo-
nylamino)pyrrolidine (1.5 g, 8.06 mmol, 1 equiv) and N,N-
diisopropylethylamine (2.80 mL, 2.0 equiv) in anhydrous DMF
(20 mL) was added 1-bromo-3,5,5-trimethylhexane (1.75 g, 8.46
mmol, 1.05 equiv). The reaction mixture was stirred at 40 °C
for 12 h, and after cooling to room temperature, the solvent
was removed under reduced pressure. The crude product was
purified by automated flash chromatography (gradient elu-
tion: 30% EtOAc in hexane to 60% EtOAc in hexane with 5%
methanol) to afford [1-(3,5,5-trimethylhexyl)pyrrolidin-3-yl]-
1
Found: 309.1968. H NMR (500 MHz, CHCl3-d): δ 2.13 (m,
1H), 2.43 (m, 1H), 2.81 (m, 1H), 2.96 (m, 1H), 3.07 (m, 2H),
3.24 (m, 2H), 3.49 (s, 2H), 3.63 (m, 1H), 3.83 (m, 1H), 4.22 (m,
2H), 7.15-7.52 (aromatics, 10H). 13C NMR (500 MHz, CHCl3-
d): δ 24.0, 30.6, 43.5, 47.5, 53.5, 57.2, 60.2, 127.2-135.8, 171.8
N-[1-(3,4-Dichlorobenzyl)pyrrolidin-3-yl]-2-phenyl-
acetamide (3c). To a solution of phenylacetic acid (40 mg,
0.3 mmol) in NMP (1.5 mL) was added a solution of 1-(3,4-
dichlorobenzyl)pyrrolidin-3-ylamine 21a (73 mg, 0.25 mmol)

6872 Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27
Lavrador et al.
and N,N-diisopropylethylamine (130.0 µL, 0.75 mmol) in DMF
(1.0 mL). This was followed by a solution of EDCI (86 mg, 0.45
mmol) and HOBT (59 mg, 0.45 mmol) in NMP (1.0 mL). The
reaction mixture was stirred at room temperature for 16 h and
then purified by prep-LCMS to give the desired product as the
TFA salt (63.5 mg, 70.0%). MALDI-FTMS calcd for C19H20-
Cl2N2O: 363.1025 [MH+]. Found: 363.1026. 1H NMR (500
MHz, CHCl3-d): δ 2.12 (m, 1H), 2.44 (m, 1H), 2.87 (m, 1H),
3.03 (m, 1H), 3.34 d, J ) 12.5 Hz, 1H), 4.17 (d, J ) 12.5 Hz,
1H), 4.88 (m, 1H), 7.26 (m, 5H), 7.34 (dd, J ) 8.2 Hz, 1.7 Hz,
1H), 7.50 (d, J ) 8.2 Hz, 1H), 7.54 (s, 1H). 13C NMR (500 MHz,
CHCl3-d): δ 30.4, 43.4, 47.2, 52.9, 57.5, 59.3, 127.2, 128.8,
129.4, 129.7, 131.8, 132.4, 134.0, 135.1, 135.2, 171.8.
N-[1-(3,4-Dichlorobenzyl)pyrrolidin-3-yl]-2-(4-phen-
oxyphenyl)acetamide (3e). The compound was prepared
according to the procedure described for compound 3c using
4-phenoxyphenylacetic acid in place of phenylacetic acid (70.3
mg, 62%). MALDI-FTMS calcd for C25H24Cl2N2O2: 455.1288
[MH+]. Found: 455.1283. 1H NMR (500 MHz, CHCl3-d): δ 2.08
(m, 1H), 2.45 (m, 1H), 2.87 (m, 1H), 3.03 (m, 1H), 3.29 (m,
1H), 3.44 (s, 2H), 3.60 (m, 1H), 3.99 (d, J ) 12.9 Hz, 1H), 4.12
(d, J ) 12.9 Hz, 1H), 4.74 (m, 1H), 6.91-7.55 (aromatics, 12H).
13C NMR (500 MHz, CHCl3-d): δ 30.6, 42.6, 47.6, 52.9, 57.7,
59.2, 118.0, 119.1, 123.5, 129.5, 129.9, 130.7, 131.7, 132.3,
133.9, 156.5, 157.3, 171.7.
2-Benzo[1,3]dioxol-5-yl-N-[1-(3,4-dichlorobenzyl)pyr-
rolidin-3-yl]acetamide (3f). The compound was prepared
according to the procedure described for compound 3c, using
3,4-methylenedioxyphenylacetic acid in place of phenylacetic
acid (86.0 mg, 85.0%). MALDI-FTMS calcd for C20H20-
Cl2N2O3: 407.0924 [MH+]. Found: 407.0915. 1H NMR (500
MHz, CHCl3-d): δ 2.12 (m, 1H), 2.45 (m, 1H), 2.89 (m, 1H),
3.06 (m, 1H), 3.35 (m, 1H), 3.37 (s, 2H), 3.68 (m, 1H), 4.03 (m,
1H), 4.19 (d, J ) 13.0 Hz, 1H), 4.80 (m, 1H), 5.90 (s, 2H), 7.12
(dd, J ) 8.2 Hz, 1.8 Hz, 1H), 7.32-7.54 (aromatics, 5H). 13C
NMR (500 MHz, CHCl3-d): δ 30.4, 43.0, 47.3, 53.0, 57.5, 59.2,
101.2, 108.6, 109.7, 122.5, 129.6, 131.8, 132.4, 133.9, 135.2,
146.9, 148.0, 172.0.
2-Benzo[b]thiophen-3-yl-N-[1-(3,4-dichlorobenzyl)pyr-
rolidin-3-yl]acetamide (3g). This compound was prepared
according to the procedure described for compound 3c, using
benzo[b]thiophene-3-acetic acid in place of phenylacetic acid
(52.3 mg, 50.0%). MALDI-FTMS calcd for C21H20Cl2N2OS:
419.0746 [MH+]. Found: 419.0756. 1H NMR (500 MHz, CHCl3-
d): δ 2.04 (m, 1H), 2.38 (m, 1H), 2.77 (m, 1H), 2.97 (m, 1H),
3.24 (m, 1H), 3.47 (s, 1H), 3.55 (m, 1H), 3.72 (m, 1H), 3.88 (d,
J ) 12.8 Hz, 1H), 4.7 (d, J ) 12.8 Hz, 1H), 4.74 (m, 1H), 7.26-
7.82 (11H aromatic). 13C NMR (500 MHz, CHCl3-d): δ 30.4,
36.5, 48.0, 52.9, 57.5, 59.1, 122.1, 123.0, 124.4, 124.6, 125.0,
129.4, 129.6, 129.9, 131.7, 132.3, 133.9, 135.0, 138.8, 140.5,
170.6.
pound 3h, using 1-(3,5,5-trimethylhexyl)pyrrolidin-3-ylamine,
TFA salt 21c in place of 1-(2,2-diphenylethyl)pyrrolidin-3-
ylamine, TFA salt 21b (100.0 mg, 83.7%). MALDI-FTMS calcd
for C24H35ClN2O: 403.2511 [MH+]. Found: 403.2508. 1H NMR
(500 MHz, methanol-d4): 0.92 (s, 9H), 0.97 (d, J ) 6.5 Hz, 3H),
1.14 (AB, J ) 6.0 Hz and 14.0 Hz, 1H), 1.23 (AB, 1H), 1.6 (m,
4H), 2.25 (m, 1H), 2.6 (m, 1H), 3.25 (m, 2H), 3.64 (m, 3H), 4.12
(m, 1H), 4.38 (m, 2H), 6.72 (d, J ) 3.4 Hz, 1H), 6.87 (d, J )
3.4 Hz, 1H), 7.40 (m, 2H), 7.74 (m, 2H). 13C NMR (500 MHz,
methanol-d4): δ 22.7, 28.7, 29.2, 30.4, 31.9, 35.9, 44.0, 52.0,
54.0, 55.2, 55.9, 56.5, 108.1, 115.6, 126.6, 130.2, 135.0, 148.0,
156.0.
(R)-2-[(2-Benzo[b]thiophen-3-yl-acetylamino)methyl]-
pyrrolidine-1-carboxylic Acid tert-Butyl Ester ((R)-22a).
To a solution of benzo[b]thiophene-3-acetic acid (1.15 g, 6
mmol) and N,N-diisopropylethylamine (3.1 mL, 15 mmol) in
CH2Cl2 (20 mL) was added (R)-1-tert-butoxycarbonyl-2-amino-
methylpyrrolidine (1.0 g, 5 mmol), followed by EDCI (1.432 g,
7.5 mmol) and HOBT (1.01 g, 7.5 mmol). The reaction mixture
was stirred at room temperature for 16 h, diluted with CH2-
Cl2 (25 mL), and washed with aqueous 2.0 M NaOH solution
(20 mL) and then brine (20 mL). The organic extract was dried
over sodium sulfate and the solvent was removed in vacuo.
The residue was purified by automated flash chromatography
(gradient elution: 10% EtOAc in hexane to 35% EtOAc in
hexane) to yield the desired compound as a white oil (1.3 g,
71.0%). MALDI-FTMS calcd for C20H26N2O3S: 397.1556
[MNa+]. Found: 397.1564. 1H NMR (500 MHz, methanol-d4):
1.43 (d, 9H), 1.72 (m, 3H), 1.84 (m, 1H), 3.17 (m, 1H), 3.26 (m,
2H), 3.43 (m, 1H), 3.78 (s, 2H), 3.85 (m, 1H), 7.36 (m, 3H),
7.85 (m, 2H). 13C NMR (500 MHz, methanol-d4): δ 23.8, 28.9,
29.7, 37.2, 42.7, 47.6, 81.2, 122.9, 123.9, 125.3, 125.6, 125.7,
125.9, 140.1, 141.9, 173.3.
(S)-2-[(2-Benzo[b]thiophen-3-yl-acetylamino)methyl]-
pyrrolidine-1-carboxylic Acid tert-Butyl Ester ((S)-22a).
This compound was prepared according to the procedure
described for compound (R)-22a using (S)-1-tert-butoxycarbo-
nyl-2-aminomethylpyrrolidine in place of (R)-1-tert-butoxycar-
bonyl-2-aminomethylpyrrolidine (1.0 g, 53.4%). MALDI-
FTMS calcd for C20H26N2O3S: 397.1556 [MNa+]. Found:
397.1564. 1H NMR (500 MHz, methanol-d4): 1.44 (d, 9H), 1.72
(m, 3H), 1.84 (m, 1H), 3.18 (m, 1H), 3.27 (m, 2H), 3.45 (m,
1H), 3.78 (s, 2H), 3.85 (m, 1H), 7.36 (m, 3H), 7.86 (m, 2H). 13
C
NMR (500 MHz, methanol-d4): δ 23.8, 28.4, 29.5, 37.2, 42.7,
48.1, 81.2, 122.9, 124.0, 125.3, 125.6, 125.7, 125.9, 140.1, 141.9,
173.3.
(R)-2-[(4-Phenylbutyrylamino)methyl]pyrrolidine-1-
carboxylic Acid tert-Butyl Ester ((R)-22b). This compound
was prepared according to the procedure described for com-
pound (S)-22a using 4-phenylbutyric acid in place of benzo-
[b]thiophene-3-acetic acid (1.6 g, 94.6%). MALDI-FTMS calcd
for C20H30N2O3: 369.2149 [MNa+]. Found: 369.2138. 1H NMR
(500 MHz, CHCl3-d): δ 1.44 (s, 9H), 1.66 (m, 1H), 1.87 (m,
3H), 1.94 (m, 2H), 2.19 (m, 2H), 2.63 (m, 2H), 3.19 (m, 1H),
3.32 (m, 2H), 4.02 (m, 1H), 7.14-7.27 (m, 5H). 13C NMR (500
MHz, CHCl3-d): δ 24.1, 27.4, 28.6, 29.7, 35.6, 36.4, 46.4, 47.3,
56.4, 80.3, 126.0, 128.5, 128.7, 141.9, 157.0, 173.4.
(S)-2-[(4-Phenyl-butyrylamino)methyl]pyrrolidine-1-
carboxylic Acid tert-Butyl Ester ((S)-22b). This compound
was prepared according to the procedure described for com-
pound (S)-22a using 4-phenylbutyric acid in place of benzo-
[b]thiophene-3-acetic acid (1.3 g, 80%). MALDI-FTMS calcd
for C20H30N2O3: 369.2149 [MNa+]. Found: 369.2138. 1H NMR
(500 MHz, methanol-d4): δ 1.46 (d, 9H), 1.87 (m, 4H), 1.91
(m, 2H), 2.20 (t, J ) 7.5 Hz, 2H), 2.62 (t, J ) 7.6 Hz, 2H), 3.38
(m, 4H), 3.9 (m, 1H), 7.17-7.25 (m, 5H). 13C NMR (500 MHz,
methanol-d4): δ 23.8, 28.9, 36.7, 42.5, 47.6, 58.4, 81.0, 127.1,
129.5, 129.6, 129.7, 143.0, 176.2.
[5-(4-Chlorophenyl)furan-2-ylmethyl][1-(2,2-diphenyl-
ethyl)pyrrolidin-3-yl]amine (3h). To a stirred solution of
1-(2,2-diphenylethyl)pyrrolidin-3-ylamine, TFA salt 21b (114
mg, 0.3 mmol) in dichloroethane (750 µL) in the presence of
N,N-diisopropylethylamine (110 µL, 0.6 mmol) was added 5-(4-
chlorophenyl)-2-furaldehyde (57.0 mg, 0.28 mmol). The reac-
tion mixture was stirred at room temperature for 10 min and
then NaBH(OAc)3 (63 mg, 0.3 mmol) was added and the
mixture was further stirred for 2 h. The reaction mixture was
then diluted with water (100 µL) and the crude mixture was
directly purified by prep-LCMS without any workup to give
the desired diamine as a bis-TFA salt (25.6 mg, 20.0%).
MALDI-FTMS calcd for C29H29ClN2O: 457.2041 [MH+].
1
Found: 457.2039. H NMR (500 MHz, CHCl3-d): δ 1.97 (m,
1H), 2.22 (m, 1H), 2.91 (m, 1H), 3.15 (m, 2H), 3.46 (s, 2H),
3.60 (m, 2H), 3.84 (m, 1H), 3.89 (s, 1H), 4.87 (t, J ) 7.10 Hz,
1H), 6.37 (d, J ) 2.8 Hz, 1H), 6.50 (d, J ) 2.8 Hz, 1H), 7.20-
7.53 (aromatics, 14H). 13C NMR (500 MHz, CHCl3-d): δ 28.4,
42.6, 48.8, 53.7, 54.9, 57.0, 60.2, 106.5, 113.7, 125.3, 127.7,
127.8, 129.2, 134.0, 140.6, 140.7, 154.3, 162.7.
(R)-2-Benzo[b]thiophen-3-yl-N-1-pyrrolidin-2-ylmeth-
ylacetamide ((R)-23a). To a stirred solution of (R)-2-[(2-
benzo[b]thiophen-3-yl-acetylamino)methyl]pyrrolidine-1-car-
boxylic acid tert-butyl ester (R)-22a (800 mg, 2.1 mmol) in
CH2Cl2 (5 mL) was added trifluoroacetic acid (5 mL). The
resultant solution was stirred at room temperature for 2 h and
[5-(4-Chlorophenyl)-furan-2-ylmethyl][1-(3,5,5-tri-
methylhexyl)pyrrolidin-3-yl]amine (3i). This compound
was prepared according to the procedure described for com-

A Screening Library GPCR-PA+
Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27 6873
the CH2Cl2/TFA solvent was removed under reduced pressure.
The residue was coevaporated with CH2Cl2, dissolved in MeOH
(10 mL), and purified by prep-LCMS to yield the desired
compound as TFA salt (500.0 mg, 84.8%). MALDI-FTMS calcd
for C15H18N2OS: 275.1213 [MH+]. Found: 275.1205. 1H NMR
(500 MHz, methanol-d4): δ 1.74 (m, 1H), 1.98 (m, 2H), 2.08
(m, 1H), 3.22 (t, J ) 7.6 Hz, 2H), 3.49 (d, J ) 5.8 Hz, 2H),
3.66 (m, 2H), 3.86 (s, 2H), 7.6 (m, 2H), 7.47 (s, 1H), 7.83 (dd,
1H, J ) 1.0 and 10 Hz), 7.89 (dd, J ) 1.0 and 10 Hz). 13C NMR
(500 MHz, methanol-d4): δ 24.5, 28.5, 36.6, 41.9, 46.7, 62.2,
122.9, 123.9, 125.4, 125.7, 126.2, 130.5, 140.1, 141.9, 175.1.
din-2-ylmethyl-acetamide (S)-23a in place of (R)-2-Benzo[b]-
thiophen-3-yl-N-1-pyrrolidin-2-ylmethyl-acetamide (R)-23a
(126.0 mg, 50.0%). MALDI-FTMS calcd for C22H22Cl2N2OS:
433.0903 [MH+]. Found: 433.0912. [R]D +1.769 (c 1, MeOH).
1H NMR (methanol-d4): δ 1.83 (m, 2H), 2.03 (m, 2H), 2.21 (m,
1H), 3.13 (m, 1H), 3.24 (m, 1H), 3.47 (m, 1H), 3.64 (m, 2H),
3.89 (s, 2H), 4.04 (d, J ) 12.5 Hz, 1H), 4.51 (d, J ) 12.5 Hz,
1H), 7.33 (m, 3H), 7.51 (s, 1H), 7.55 (d, 1H), 7.66 (d, 1H), 7.86
(m, 2H). 13C NMR (500 MHz, methanol-d4): δ 23.2, 28.0, 36.6,
41.4, 55.3, 58.0, 70.4, 122.9, 124.0, 125.4, 125.7, 126.4, 130.4,
131.5, 132.5, 133.8, 134.3, 139.9, 141.9, 176.3.
(R)-N-(1-Phenethylpyrrolidin-2-ylmethyl)-4-phenylbu-
tyramide ((R)-4b). This compound was prepared according
to the procedure described for compound (R)-4a using (2-
chloroethyl)benzene in place of 3,4-dichlorobenzyl bromide (126
mg, 50.0%). MALDI-FTMS calcd for C23H30N2O: 351.2431
[MH+]. Found: 351.2423. [R]D +11.382 (c 1, MeOH). 1H NMR
(methanol-d4): δ 1.93 (m, 5H), 2.2 (m, 4H), 2.64 (t, J ) 7.5
Hz, 2H), 3.0.6 (m, 2H), 3.30 (m, 2H), 3.45 (m, 1H), 3.66 (m,
3H), 7.26 (m, 10H). 13C NMR (500 MHz, methanol-d4): δ 23.5,
27.9, 28.5, 33.2, 36.1, 36.4, 40.7, 55.6, 57.3, 70.1, 127.2, 128.5,
129.5, 129.6, 129.9, 130.1, 137.6, 142.8, 178.7.
(S)-N-(1-Phenethylpyrrolidin-2-ylmethyl)-4-phenylbu-
tyramide ((S)-4b). This compound was prepared according
to the procedure described for compound (S)-4a using (2-
chloroethyl)benzene in place of 3,4-dichlorobenzyl bromide
(189.0 mg, 75.0%). MALDI-FTMS calcd for C23H30N2O:
351.2431 [MH+]. Found: 351.2426. [R]D -15.314 (c 1, MeOH).
1H NMR (500 MHz, methanol-d4): δ 1.36 (m, 2H), 1.93 (m,
3H), 2.13 (m, 1H), 2.24 (m, 3H), 2.64 (t, J ) 7.5 Hz, 2H), 3.09
(m, 2H), 3.30 (m, 1H), 3.44 (m, 1H), 3.67 (m, 5H), 7.22 (m,
10H). 13C NMR (500 MHz, methanol-d4): δ 23.54, 28.0, 28.5,
33.2, 36.1, 36.4, 40.7, 55.6, 57.3, 70.1, 127.2, 128.5, 129.6, 137.6,
142.8, 178.7.
Computational Chemistry. Using an input SD file of the
81 560 compounds making up NBI-DS, standard 3D BCUT
metrics were calculated for each member of the collection. A
ø2 approach provided by Diverse Solutions automatically chose
and optimized which metric combinations would comprise the
optimal chemistry space to represent maximum diversity of
this set. The result was a chemical space defined by five BCUT
metrics: charge, H-bond donor, H-bond acceptor, polarizability
I (low eigenvalue), and polarizability II (high eigenvalue).
Arbitrarily dividing each axis into 10 bins afforded 100 000
cells, so cell occupancy can be used as a measure of the
diversity/similarity between the molecules based on the coor-
dinates of each cell.
A similar process was used to describe the location of the
known GPCR ligands having reported affinity for specific a
GPCR (630 molecules), the training set (111 ligands for
peptide-activated GPCRs), and the virtual libraries from which
the test set (2025 compounds) was selected. In the case of the
latter, virtual libraries for each of the 19 templates were
enumerated separately using the Afferent and Project Library
(MDL), and the resulting SD files contained over 9 million
compounds. For reasons already described, the location of the
training set (97 occupied cells) was expanded to include all
neighboring cells (6437 cells). For every molecule in each
virtual library, the same five BCUT metrics were calculated
to give 19 cell-index files, which were then compared with the
locations of molecules within the expanded training set. The
cell-index files also provided information on how many cells
were occupied by molecules from each virtual library, and the
degree of overlap with the training set was assessed using an
in house algorithm, “Cell Compare”. Those virtual libraries
having the highest percentage of occupied cells overlapping
with the GPCR-PA+ subspace were selected for synthetic
trials; in all, 10 templates were thereby selected. As additional
selection criteria, so as to achieve a balance between relevant
space coverage and density of cell population, no more than
25 compounds per cell were selected for the test set, selection
was biased to maximize the number of templates used, and
finally, matrixes for parallel synthesis were chosen so as to
maximize cell occupancy in GPCR-PA+ subspace.
(S)-2-Benzo[b]thiophen-3-yl-N-1-pyrrolidin-2-ylmeth-
ylacetamide ((S)-23a). This compound was prepared accord-
ing to the procedure described for compound (R)-23a using
(S)-2-[(2-benzo[b]thiophen-3-ylacetylamino)methyl]pyrrolidine-
1-carboxylic acid tert-butyl ester (S)-22a in place of (R)-2-[(2-
benzo[b]thiophen-3-ylacetylamino)methyl]pyrrolidine-1-car-
boxylic acid tert-butyl ester (R)-22a (527.0 mg, 90.0%). MALDI-
FTMS calcd for C15H18N2OS: 275.1213 [MH+]. Found: 275.1209.
1H NMR (500 MHz, methanol-d4): δ 1.93 (m, 1H), 1.98 (m,
2H), 2.08 (m, 1H), 3.22 (t, J ) 7.3 Hz, 2H), 3.49 (d, J ) 5.8
Hz, 2H), 3.66 (m, 2H), 3.86 (s, 2H), 7.6 (m, 2H), 7.47 (s, 1H),
7.83 (d, 1H, J ) 7.4 Hz), 7.89 (d, J ) 7.4 Hz). 13C NMR (500
MHz, methanol-d4): δ 24.5, 28.5, 36.6, 41.9, 46.7, 62.2, 122.9,
123.9, 125.4, 125.7, 126.2, 130.5, 140.1, 141.9, 175.1
(R)-4-Phenyl-N-1-pyrrolidin-2-ylmethylbutyramide ((R)-
23b). This compound was prepared according to the procedure
described for compound (R)-23a using (R)-2-[(4-phenylbutyry-
lamino)methyl]pyrrolidine-1-carboxylic acid tert-butyl ester
(R)-22b in place of (R)-2-[(2-benzo[b]thiophen-3-ylacetylami-
no)methyl]pyrrolidine-1-carboxylic acid tert-butyl ester (R)-
22a (484.0 mg, 92.0%). MALDI-FTMS calcd for C15H22N2O:
247.1805 [MH+]. Found: 247.1802. 1H NMR (500 MHz,
methanol-d4): δ 1.74 (m, 1H), 1.94 (m, 2H), 2.02 (m, 1H), 2.03
(m, 1H), 2.07 (m, 1H), 2.26 (t, J ) 7.5 Hz, 2H), 2.64 (t, J ) 7.5
Hz, 2H), 3.30 (m, 2H), 3.45 (m, 2H), 3.67 (m, 1H), 7.17-7.25
(m, 5H). 13C NMR (500 MHz, methanol-d4): δ 24.5, 28.6, 36.2,
36.4, 41.7, 46.7, 62.2, 127.1, 129.5, 129.6, 143.0, 177.8.
(S)-4-Phenyl-N-1-pyrrolidin-2-ylmethylbutyramide ((S)-
23b). This compound was prepared according to the procedure
described for compound (R)-23a using (S)-2-[(4-phenylbutyry-
lamino)methyl]pyrrolidine-1-carboxylic acid tert-butyl ester
(S)-22b in place of (R)-2-[(2-benzo[b]thiophen-3-ylacetylami-
no)methyl]pyrrolidine-1-carboxylic acid tert-butyl ester (R)-
22a (447.0 mg, 85.0%). MALDI-FTMS calcd for C15H22N2O:
247.1805 [MH+]. Found: 247.1800. 1H NMR (500 MHz,
methanol-d4): δ 1.74 (m, 1H), 2.01 (m, 5H), 2.27 (t, J ) 7.3
Hz, 2H), 2.64 (t, J ) 7.5 Hz, 2H), 3.27 (m, 1H), 3.44 (d, J )
5.8 Hz, 2H), 3.66 (m, 1H), 7.2 (m, 5H). 13C NMR (500 MHz,
methanol-d4): δ 24.5, 28.5, 36.2, 36.4, 41.7, 46.7, 62.3, 127.1,
129.5, 129.6, 129.7, 142.9, 177.9.
(R)-2-Benzo[b]thiophen-3-yl-N-[1-(3,4-dichlorobenzyl)-
pyrrolidin-2-ylmethyl]acetamide ((R)-4a). To a solution of
(R)-2-benzo[b]thiophen-3-yl-N-1-pyrrolidin-2-ylmethylaceta-
mide (R)-23a (200.0 mg, 0.72 mmol) in the presence of N,N-
diisopropylethylamine (375 µL, 2.16 mmol) in anhydrous DMF
(3.6 mL) was added 3,4-dichlorobenzyl bromide (209.0 mg, 0.87
mmol). The reaction mixture was stirred at 40 °C for 12 h,
cooled to room temperature, and purified by prep-LCMS to give
the desired compound as the TFA salt (243.0 mg, 78.0%).
MALDI-FTMS calcd for C22H22Cl2N2OS: 433.0903 [MH+].
Found: 433.0899. [R]D -10.457 (c 1, MeOH). NMR (methanol-
d4): δ 1.84 (m, 2H), 2.04 (m, 2H), 2.23 (m, 1H), 3.13 (m, 1H),
3.26 (m, 1H), 3.46 (m, 1H), 3.68 (m, 2H), 3.89 (s, 2H), 4.05 (d,
J ) 12.5 Hz, 1H), 4.48 (d, J ) 12.5 Hz, 1H), 7.34 (m, 3H), 7.5
(s, 1H), 7.55 (d, 1H), 7.66 (d, 1H), 7.85 (m, 2H). 13C NMR (500
MHz, methanol-d4): δ 23.1, 28.0, 36.6, 41.3, 55.3, 57.9, 70.3,
122.9, 124.0, 125.4, 125.7, 126.4, 130.4, 131.5, 132.4, 132.5,
133.8, 134.3, 135.3, 139.9, 141.9, 176.2.
(S)-2-Benzo[b]thiophen-3-yl-N-[1-(3,4-dichloro-benzyl)-
pyrrolidin-2-ylmethyl] -acetamide ((S)-4a). This compound
was prepared according to the procedure described for com-
pound (R)-4a using (S)-2-Benzo[b]thiophen-3-yl-N-1-pyrroli-

6874 Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27
Lavrador et al.
(3) For a list of the Top 200 Prescription Medicines see: http://
www.rxlist.com/top200.htm.
(4) Marshall, G. R. Peptide Interactions with G-Protein Coupled
Receptors. Biopolymers (Pept. Sci.) 2001, 60, 246-277.
(5) Rees, D. C. Non-Peptide Ligands for Neuropeptide Receptors.
Annu. Rep. Med. Chem. 1993, 28, 59-68.
(6) Sugg, E. E. Non-peptide Agonists for Peptide Receptors: Lessons
from Ligands. Annu. Rep. Med Chem. 1997, 32, 277.
(7) Yang, L. Non-Peptide Somatostatin Receptor Ligands. Annu.
Rep. Med. Chem. 1999, 34, 209-218.
Biological Assays. MC4-R and MCH1-R binding assays
were carried out with membrane preparations derived from
mammalian cell lines stably transfected with human MC4
receptor (2) or rat MCH1 receptor (3). Crude membrane
preparations were generated by nitrogen cavitation of cells
followed by differential centrifugation. Specifically, cells were
suspended in “binding buffer” and subjected to nitrogen
cavitation at 900 psi on ice for 30 min. Lysed cells were
centrifuged at 800g for 10 min, and the resultant supernatant
was centrifuged at 38 000g for 20 min. Pellets were resus-
pended in the appropriate binding buffer and stored at -80
°C until use. MC4-R binding assays were carried out in 25
mM HEPES (pH 7.0), 1.5 mM CaCl2, 1.0 mM MgSO4, and 100
mM NaCl. MCH1-R binding assays were carried out in 50
mM HEPES (pH 7.0), 10 mM MgCl2, 2 mM EGTA. Binding
reactions were carried out at equilibrium in 96-well format
and were terminated by rapid vacuum filtration using a
Packard Unifilter-96 plate harvester, over glass-fiber (GF-
C) filter plates pretreated for 15 min with 0.5% polyethylen-
imine. Filters were washed four times with 0.2 mL of binding
buffer.
GnRH receptor binding assays were carried out in whole-
cell format using mammalian cells stably transfected with
human GnRH receptor (4). Cells were harvested on the day of
assay and assayed at 0.5 × 106 cells per well in 10 mM Hepes
(pH 7.5), 150 mM NaCl, and 0.1% BSA. Binding reactions were
carried out at equilibrium in 96-well format and terminated
by rapid vacuum filtration over Millipore Multiscreen (GF-
C) filter plates pretreated for 120 min with 0.1% polyethyl-
enimine. Filters were washed two times with 0.2 mL phosphate-
buffered saline (pH 7.4). In all binding assays, filters were
dried, 0.05 mL Microscint-20 was added, and radioactivity was
monitored by liquid scintillation counting.
(8) Pearlman, R. S.; Smith, K. M. Novel Software Tools for Chemical
Diversity. Perspect. Drug Discovery 1998, 9, 339-353.
(9) Pearlman, R. S.; Smith, K. M. Metric Validation and the
Receptor-Relevant Subspace Concept. J. Chem. Inf. Comput. Sci.
1999, 39, 28-35.
(10) Williams, J. P.; Lavrador, K. A Solution-Phase Combinatorial
Synthesis of Selective Dopamine D4
2000, 3, 43-50.
Ligands. Comb. Chem. HTS
(11) Zeng, L.; Burton, L.; Yung, K.; Shushan, B.; Kassel, D. B.
Automated Analytical/Preparative High-Performance Liquid
Chromatography-Mass Spectrometry System for the Rapid
Characterization and Purification of Compound Libraries. J.
Chromatogr. A 1998, 794, 3-13.
(12) Abdel-Magid, A. F.; Carson, K. G.; Harris, B. D.; Maryanoff, C.
A.; Shah, R. D. Reductive Amination of Aldehydes and Ketones
with Sodium Triacetoxyborohydride. Studies on Direct and
Indirect Reductive Amination Procedures. J. Org. Chem. 1996,
61, 3849-3862.
(13) Cone, R. D.; Lu, D.; Koppula, S.; Vage, D. I.; Klungland, H.;
Boston, B.; Chen, W.; Orth, D. N.; Pouton, C.; Kesterson, R. A.
The Melanocortin Receptors: Agonists, Antagonists, and the
Hormonal Control of Pigmentation. Rec. Prog. Hor. Res. 1996,
51, 287-318.
(14) Goodfellow, V. S.; Saunders, J. The Melanocortin System and
Its Role in Obesity and Cachexia. Curr. Top. Med. Chem. 2003,
3, 855-883.
(15) Qu, D.; Ludwig, D. S.; Gammeltoft, S.; Piper, M.; Pelleymounter,
M. A.; Cullen, M. J.; Mathes, W. F.; Przypek, J.; Kanarek, R.;
Maratos-Flier, E. A Role For Melanin-Concentrating Hormone
in the Central Regulation of Feeding Behavior. Nature 1996,
380, 243-7.
(16) Borowsky, B.; Durkin, M. M.; Ogozalek, K.; Marzabadi, M. R.;
DeLeon, J.; Lagu, B.; Heurich, R.; Lichtblau, H.; Shaposhnik,
Z.; Daniewska, I.; Blackburn, T. P.; Branchek, T. A.; Gerald, C.;
Vaysse, P. J.; Forray, C. Antidepressant, Anxiolytic and Anorec-
tic Effects of a Melanin-Concentrating Hormone-1 Receptor
Antagonist. Nat. Med. 2002, 8, 825-30.
(17) Sealfon, S. C.; Weinstein, H.; Millar, R. P. Molecular Mechanisms
of Ligand Interaction with the Gonadotropin-Releasing Hormone
Receptor. Endocr. Rev. 1997, 18, 180-205.
(18) Emons, G.; Grundker, C.; Gunthert, A. R.; Westphalen, S.;
Kavanagh, J.; Versschraege, C. GnRH Antagonists in the
Treatment of Gynaecological and Breast Cancers. Endocr. Relat.
Cancer 2003, 10, 291-299.
(19) Zhang, J.-H.; Chung, T. D. Y.; Oldenburg, K. R. A Simple
Statistical Parameter for Use in Evaluation and Validation of
High Throughput Screening Assays. J. Biomol. Screening 1999,
4, 67-73.
In all binding experiments, specific binding is defined as
that displaced by a known, pharmacologically relevant unla-
beled competitor. The concentration of competitor was defined
from competition curves for each assay where the competitor
inhibited greater than 95% of tracer binding. Percent inhibi-
tion of binding for test compounds is defined as [1 - (specific
binding in the presence of test compound)/(control specific
binding)] × 100. Binding data for Ki determinations were
analyzed by the nonlinear least squares curve-fitting program
GraphPad Prism.
Acknowledgment. This work was supported in part
by a grant from the National Institutes of Health (Grant
Number 5R44 GM61491-03). In addition, we thank Dr.
Christian Baber for help in analyzing the final data set.
Supporting Information Available: HPLC purity data
and MALDI-FTMS data. Color versions of Figures 1 and 4
are available form the authors on request. This material is
available free of charge via the Internet at htpp://pubs.acs.org.
(20) Ohmori, J.; Maeno, K.; Hidaka, K.; Nakato, K.; Matsumoto, M.;
Tada, S.; Hattori, H.; Sakamoto, S.; Tsukamoto, S. Dopamine
D3 and D4 receptors Antagonists: Synthesis and Structure-
Activity Relationships of (S)-(+)-N-(1-Benzyl-3-pyrrolidinyl)-5-
chloro-4-[(cyclopropylcarbonyl)amino]-2-methoxybenzamide (YM-
43611) and related compounds. J. Med. Chem. 1996, 39,
2764-2772.
References
(1) The most up to date listing can be viewed and searched on the
GPCRDB home page at http://www.darmstadt.gmd.de/∼gpcrdb.
(2) Watson, S.; Arkinstall, S. The G-Protein Linked Receptor Facts-
Book; Academic Press: San Diego, CA, 1996.
JM040084C