Synthesis and structure-activity relationships of 1-arylmethyl-3-(1-methyl-2-amino)ethyl-5-aryl-6-methyluracils as antagonists of the human GnRH receptor

Article abstract of DOI:10.1016/S0960-894X(03)00619-X

A new class of small molecule GnRH antagonists, the 1-arylmethyl-3-(1-methyl-2-amino)ethyl-5-aryl-6-methyluracils, was designed and a novel stereoselective synthesis for these compounds was developed. The stereochemical integrities of key intermediates (S

Full text of DOI:10.1016/S0960-894X(03)00619-X

Bioorganic & Medicinal Chemistry Letters 13 (2003) 3317–3322
Synthesis and Structure–Activity Relationships of 1-Arylmethyl-3-
(1-methyl-2-amino)ethyl-5-aryl-6-methyluracils as Antagonists of
the Human GnRH Receptor
Fabio C. Tucci,a,* Yun-Fei Zhu,a Zhiqiang Guo,a Timothy D. Gross,a
Patrick J. Connors, Jr.,a R. Scott Struthers,b Greg J. Reinhart,b John Saundersa
and Chen Chena,*
aDepartment of Medicinal Chemistry, Neurocrine Biosciences, Inc., 10555 Science Center Drive, San Diego, CA92121, USA
bDepartment of Exploratory Discovery, Neurocrine Biosciences, Inc., 10555 Science Center Drive, San Diego, CA92121, USA
Received 4 March 2003; accepted 19 May 2003
Abstract—A new class of small molecule GnRH antagonists, the 1-arylmethyl-3-(1-methyl-2-amino)ethyl-5-aryl-6-methyluracils,
was designed and a novel stereoselective synthesis for these compounds was developed. The stereochemical integrities of key inter-
mediates (S)-6 and (R)-6 were confirmed by a combination of X-ray crystallography and chiral HPLC determinations. SAR studies
were performed, which allowed the identification of derivatives (R)-9f, (R)-9h and (R)-12 as potent hGnRH antagonists (Ki=20
nM).
# 2003 Elsevier Ltd. All rights reserved.
In the previous letter,1 we described the initial SAR of
the 1-arylmethyl-5-aryl-6-methyluracils containing a
substituted 2-aminopropyl side-chain at the N-3 posi-
tion as gonadotropin-releasing hormone (GnRH)
antagonists. In that study, compound 1 was identified as
a potent antagonist of the human GnRH receptor
(hGnRH) (Ki=5 nM), and it represented an improve-
ment over our previous lead compound 22 (Fig. 1).
Compound 1 differs from 2 by a methyl group of (R)-
configuration at the b-carbon of the N-3 side-chain,
which we speculate may contribute to rotational
restriction to the molecule as well as orienting the basic
amine towards the respective acidic residue in the
receptor. Based on this premise, we decided to investigate
the effects of the introduction of a methyl group at the
a-position of the N-3 side-chain, and the results are
presented herein.
derivates and used in the Mitsunobu alkylation of uracil
3 (Scheme 1). The Mitsunobu reactions on the second-
ary alcohols (R)-5 and (S)-5 occurred with complete
inversion of configuration, leading to the corresponding
alkylated uracils (S)-6 and (R)-6, with (S)- and (R)-
configurations, respectively. The stereochemical integri-
ties of (S)-6 and (R)-6 were confirmed by chiral HPLC3
analysis, and the absolute (R)- stereochemistry of (R)-6
was determined by X-ray crystallography (Fig. 2),
which was possible due to the anomalous contributions
of the heavy atom (bromine).4 The 5-(3-methoxy)-
phenyl group was introduced through Suzuki couplings
5-Bromo-1-(2,6-difluorobenzyl)- 6-methyluracil (3) was
utilized as our starting point in this study, and was pre-
pared according to the procedure developed by our
group.1 (R)- and (S)-1-amino-2-propanol [(R)-4, (S)-4]
were separately protected as their tert-butyl carbamate
*Corresponding author. tel.: +1-858-658-7677; fax: +1-858-658-7601;
e-mail: ftucci@neurocrine.com; http://www.neurocrine.com
Figure 1. First-generation uracil-based GnRH antagonists.
0960-894X/03/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00619-X

3318
F. C. Tucci et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3317–3322
Scheme 1. (a) (Boc)2O, Et3N, CH2Cl2, 0 ꢀC to rt, 2 h; (b) (R)-5 or (S)-5, DEAD, Ph3P, THF, rt, 17 h; (c) 3-methoxyphenyl boronic acid, Pd[Ph3P]4,
K2CO3, toluene/EtOH/H2O, 110 ꢀC, 18 h; (d) TFA/CH2Cl2 1:1 v/v, rt, 1 h; (e) carbonyl compound, MeOH, NaBH4, rt, 1 h; (f) CH2O aq, MeOH,
NaBH4, rt, 15 min.
of the appropriate boronic acid with bromides (S)-6 and
(R)-6. The products (S)-7 and (R)-7 were treated with
trifluoroacetic acid in CH2Cl2 to furnish the corre-
sponding amines (S)-8 and (R)-8 in excellent yields.
Each compound was then reacted with a variety of
aldehydes and ketones, under reductive conditions, to
give the corresponding secondary amines (S)-9 and (R)-
9. Some of the secondary amines were further alkylated
with formaldehyde and NaBH4 in MeOH to give the
tertiary amines (S)-10 and (R)-10.
Scheme 2. Both amines (S)-8 and (R)-8 were reacted
with 2-vinylpyridine in EtOH at reflux to give the sec-
ondary amines (S)-11 and (R)-11. These compounds
were then methylated with aqueous formaldehyde and
NaBH4 in MeOH to provide (S)-12 and (R)-12.
All compounds were then evaluated for their competi-
tive binding to the cloned human GnRH receptor
expressed in HEK293 cells, using a 96-well filtration
setup5,6 and des-Gly10[125I-Tyr5, DLeu6, NMeLeu7,
Pro9-NEt]GnRH as the radiolabel. The results of the
binding assay are presented in Tables 1 and 2. The
functional antagonism of the compounds herein syn-
thesized was confirmed by their ability to inhibit GnRH
stimulated Ca2+ flux in the transfected cells, in a dose-
dependent manner.7 For instance, compounds (R)-9f
and (R)-9h were found to be functional antagonists with
IC50’s=12 and 26 nM, respectively. Analogously to our
previous findings,1,8a the compounds described in this
letter were specific for the human variant of the recep-
tor. Compared to the hGnRH receptor, compound (R)-
9f is 8-fold less potent (Ki=160 nM) in the monkey
GnRH receptor and 600-fold less potent (Ki=12,700
nM) in the rat receptor.
The derivatives containing the 2-pyridylethyl side-chain
on the basic nitrogen were synthesized according to
Previous knowledge from past SAR studies of our
bicyclic and monocyclic series1,2,8 indicated that the best
binding affinity towards the h-GnRH receptor was
achieved when the basic nitrogen was alkylated with
both methyl and 2-pyridylethyl side-chains. We postu-
lated that the ‘bidentated’ basic nature of this motif, in
Figure 2. X-ray structure of (R)-6.

F. C. Tucci et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3317–3322
3319
Scheme 2. (a) 2-Vinylpyridine, Et3N, EtOH, reflux, 3 days; (b) CH2O aq, MeOH, NaBH4, rt, 15 min.
Table 1. Binding affinities of 1-(2,6-difluorobenzyl)-3-(1-methyl-2-
amino)ethyl-5-(3-methoxyphenyl)-6-methyluracils [(S)- or (R)-9a–d,
(S)- or (R)-10a–b, (S)-12 and (R)-12] towards the hGnRH receptor
Table 2. Binding affinities of 1-(2,6-difluorobenzyl)-3-(1-methyl-2-
amino)ethyl-5-(3-methoxyphenyl)-6-methyluracils [(S)- or (R)-9e–j]
towards the hGnRH receptor
Compd
R
R0
Ki (nM)
Compd
R
R0
Ki (nM)
(S)-12
CH3
1800
(S)-9e
H
400
(R)-12
(S)-9a
(R)-9a
(S)-10a
(R)-10a
(S)-9b
(R)-9b
(S)-10b
(R)-10b
(S)-9c
CH3
H
20
370
(R)-9e
(S)-9f
(R)-9f
(S)-9g
(R)-9g
(R)-9 h
(R)-9i
(R)-9j
(R)-9k
H
H
H
H
H
H
H
H
H
70
320
20
H
75
CH3
CH3
H
2000
150
550
90
1600
780
H
20
CH3
CH3
H
7000
590
720
85
950
(R)-9c
(S)-9d
(R)-9d
H
430
250
H
1200
190
methyl group of (R)-configuration on the a-carbon of
the N-3 side-chain contributes to enhance the binding
affinity, albeit only marginally, over the unsubstituted 2.
However, the gain in energy is not as large as that real-
ized by the introduction of a methyl group on the b-
carbon, as in 1.
H
addition to the p interaction provided by the pyridine
ring, could account for the enhanced affinity towards
the receptor.8f Compound (R)-12, which represents a
direct comparison to both 1 and 2 bound with Ki=20
nM, whereas its (S)-counterpart (S)-12 was 90 times less
potent (Ki=1800 nM). Therefore, the introduction of a
The (R)-enantiomeric preference for the receptor was
further confirmed by alkylation of amines (S)-8 and (R)-
8 with various carbonyl compounds. For instance, the
N-benzylated compound (S)-9a (Ki=370 nM) was

3320
F. C. Tucci et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3317–3322
about 5 times less potent than its counterpart (R)-9a
(Ki=75 nM). Interestingly, when both compounds were
methylated, the corresponding tertiary amines (S)-10a
and (R)-10a displayed much lower affinity for the
receptor [(S)-10a Ki=2000 nM and (R)-10a Ki=150
nM]. This observation contrasts with the 2-(pyr-
idylethyl) side-chain compound series, where the N-
methylated tertiary amines are the more active spe-
cies.1,2 According to the data presented in Table 1, it is
clear that benzylic side-chains were preferred over ali-
phatic ones. For example, when compared to (R)-9a,
compound (R)-9b, which is the cyclopropylmethyl deri-
vative, is about 10-fold less potent. In addition, the iso-
lipophilic isobutyl derivative (R)-9d and the cyclopentyl
analogue (R)-9c are 3- and 6-fold less active, respec-
tively. Further N-methylation of these aliphatic secondary
amines did not translate into a significant improvement
in potency. For example, compound (R)-9b had similar
binding to (R)-10b.
to a dramatic decrease in affinity. The same trend was
observed for the fluoro-substituted series. Both the
ortho-[(R)-9g] and the meta-[(R)-9j] fluorinated ana-
logues are very similar in potency (Ki’s=90 and 85 nM,
respectively), whereas the para-compound (R)-9k has a
Ki value of 250 nM.
In addition to the compounds described in Tables 1 and
2, we were interested in further restraining the rota-
tional freedom of the N-3 side-chain. Thus, by formally
connecting the methyl groups at both the basic nitrogen
and the a- position of compound (R)-12, we arrived at
3-pyrrolidinyl substituted derivatives (Scheme 3).
The synthesis of these compounds started with N-boc
protection of the racemic 3-pyrrolidinol in quantitative
yield. The resulting protected alcohol was coupled to
5-bromouracil 3 under Mitsunobu conditions. The
product 13 was obtained in 77% yield and was subse-
quently subjected to Suzuki coupling with 3-methoxy-
phenyl boronic acid, in the presence of catalytic amounts
of Pd[Ph3P]4, to give 14 in excellent yield. Deprotection,
followed by alkylation with either benzaldehyde or
2-vinylpyridine gave 16a and 16b, respectively.
The substitution pattern of the benzylic side-chain was
investigated next (Table 2). Introduction of an ortho-
methyl group [(R)-9e] caused little change in affinity
when compared to (R)-9a. When an ortho- or meta-
methoxy groups were introduced, however, the corre-
sponding compounds (R)-9f and (R)-9h were about 4-
fold more potent than (R)-9a, and matched the potency
of the pyridyl derivative (R)-12 (Ki=20 nM). Substitu-
tion at the para-position [(R)-9i], on the other hand, led
Disappointingly, neither compound showed improved
binding over their acyclic counterparts (Table 3).
In fact, the binding affinity of 16b was dramatically
inferior to the one of (R)-12 (Ki=8,300 nM vs 20 nM,
Scheme 3. (a) (Boc)2O, Et3N, CH2Cl2, 0 ꢀC to rt, 4 h; (b) DEAD, Ph3P, THF, rt, 17 h; (c) 3-methoxyphenyl boronic acid, Pd[Ph3P]4, K2CO3,
toluene/EtOH/H2O, 110 ꢀC, 15 h; (d) TFA/CH2Cl2 1:1 v/v, rt, 1 h; (e) benzaldehyde, MeOH, NaBH4, rt, 1 h; (f) 2-vinylpyridine, EtOH, Et3N, reflux,
48 h.

F. C. Tucci et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3317–3322
3321
Table 3. Binding affinities of 1-(2,6-difluorobenzyl)-3-(3-pyrrolidinyl)-
5-(3-methoxyphenyl)-6-methyluracils 15, 16a and 16b towards the
hGnRH receptor
to Mrs. Mila Lagman (Neurocrine Biosciences) and Dr.
John Huffman (Indiana University Molecular Structure
Center) for assistance in obtaining the chiral HPLC and
the X-ray crystallographic data, respectively.
References and Notes
1. Guo, Z.; Zhu, Y.-F.; Tucci, F. C.; Gao, Y.-H.; Struthers,
R. S.; Saunders, J.; Gross, T, D.; Xie, Q.; Reinhart, G. J.;
Chen, C. Bioorg. Med. Chem. Lett. See preceding paper.
doi:10.1016/50960-894X(03)00612-7.
2. Zhu, Y.-F.; Gross, T. D.; Guo, Z.; Connors, P. J., Jr.; Gao,
Y.-H.; Tucci, F. C.; Struthers, R. S.; Reinhart, G. J.; Saun-
ders, J.; Chen, C. J. Med. Chem. 2003, 46, 2023.
Compd
R
H
Ki (nM)
15
15,000
3. The enantiomeric excesses (% e.e.) of compounds (S)-6 and
(R)-6 were determined by HPLC, using a long Pirkle Covalent
(R,R) WHELK column (25 cmꢁ4.6 mm), manufactured by
REGIS. The samples were eluted with an isocratic 80:20 (v/v)
mixture of hexanes and ethanol, respectively. For (S)-6,
rt=11.10 min, e.e.=>99%. For (R)-6, rt=9.93 min, e.e.=
>99%.
16a
16b
280
8300
4. (R)-6 was crystallized from slow diffusion of diethyl ether
into
a dichloromethane solution. Crystallographic data
(excluding structure factors) for the structure in this paper
have been deposited with the Cambridge Crystallographic
Data Centre as supplementary publication numbers CCDC
204509. Copies of the data can be obtained, free of charge, on
application to CCDC, 12 Union Road, Cambridge CB2 1EZ,
UK (Fax: +44-1223-336033; e-mail: deposit@ccdc.cam.ac.uk).
5. Perrin, M. H.; Haas, Y.; Rivier, J. E.; Vale, W. W. Mol.
Pharmacol. 1983, 23, 44.
6. On each assay plate, a standard antagonist of comparable
affinity to those being tested was included as a control for
plate-to-plate variability. Overall, Ki values were highly
reproducible with an average standard deviation of 45% for
replicate Ki determinations. Key compounds were assayed in
3–8 independent experiments.
7. Inhibition of GnRH stimulatd Ca2+ flux: functional activ-
ity of compounds for the hGnRH receptor was determined by
inhibition of GnRH stimulated Ca2+ flux. RBL cells stably
expressing the full-length human GnRH receptor were seeded
into 96-well, black wall clear bottom plates (Corning) at a
density of 50,000 cells/well and the cells allowed to attach
overnight. Cells were then loaded with the Ca2+ sensitive dye,
Fluo-4 (Molecular Probes), by incubation in loading medium
[(DMEM with 20 mM Hepes, 10% FBS, 2 mM Fluo-4, 0.02%
pluronic acid (Molecular Probes) and 2.5 mM probenecid
(Sigma)] for 1 h at 37 ꢀC. Cells were then washed three times
with assay buffer (Hanks balanced salt, 20 mM Hepes, 2.5
mM probenecid). Compounds at varying concentrations in
assay buffer were pre-incubated with cells for 1 min prior to
stimulation with GnRH (5 nM). Measurement of fluorescence
due to GnRH stimulated Ca2+ flux was performed according
to manufacturer’s instructions on a FLIPR system (Molecular
Devices, FLIPR384 system). IC50 values for the inhibition of
GnRH-stimulated Ca2+ flux were calculated using the Prism
software package (GraphPad Software) with a ‘sigmoidal
dose–response (variable slope)’ option for curve fitting.
8. (a) Zhu, Y.-F.; Struthers, R. S.; Gao, Y.-H.; Connors, P. J.;
Gross, T. D.; Saunders, J.; Wilcoxen, K.; Reinhart, G. J.;
Ling, N.; Chen, C. Bioorg. Med. Chem. Lett. 2002, 12, 399. (b)
Zhu, Y.-F.; Wilcoxen, K.; Saunders, J.; Guo, Z.; Gao, Y.-H.;
Connors, P. J.; Gross, T. D.; Tucci, F. C.; Struthers, R. S.;
Reinhart, G. J.; Chen, C. Bioorg. Med. Chem. Lett. 2002, 12,
403. (c) Wilcoxen, K.; Zhu, Y.-F.; Connors, P. J., Jr.; Saun-
ders, J.; Gross, T. D.; Gao, Y.; Reinhart, G. J.; Struthers,
respectively), which may reflect the extent to which the
contacts between the pyridine and the receptor residues
were disrupted. On the other hand, the difference
between 16a and (R)-10a is much smaller (Ki=280 nM
vs 150 nM, respectively), especially if one considers that
16a is a racemate. The main difference between these
two pairs of compounds is the group attached to the
basic nitrogen (pyridylethyl vs benzyl), and this clearly
points to two distinct binding modes. The constraint
imposed by the pyrrolidine ring in 16b, may disrupt the
interaction between the pyridylethyl-substituent and,
possibly, a polar residue in the receptor, whereas in the
case of 16a, the predominantly aromatic/lipophilic
interaction is maintained. Based on the results obtained
thus far, we speculate that the benzyl group most probably
interacts with Phe 309 on TM7 of the hGnRH receptor.
In summary, we have developed a novel synthesis of a
new class of small-molecule antagonists to the human
GnRH receptor, the 1-arylmethyl-3-(1-methyl-2-ami-
no)ethyl-5-aryl-6-methyluracils. We have also demon-
strated that the introduction of a methyl group of (R)-
configuration at the a-carbon of the N-3 side-chain gave
a modest improvement in activity when compared to the
unsubstituted ethylene analogues.2 The SAR of this
class of compounds was explored and three potent
antagonists were identified, (R)-9f, (R)-9h and (R)-12
(Ki=20 nM). In addition, we have demonstrated for the
first time in the uracil series1,2 that compounds lacking
the 2-pyridylethyl motif on the basic nitrogen can still
display excellent potency against the hGnRH receptor
[e.g., (R)-9f, (R)-9h], as long as the a-methyl group of
(R)-configuration is present.
Acknowledgements
This work was partly supported by NIH grants 1-R43-
HD38625-01 and 2-R44-HD38625-02. We are indebted

3322
F. C. Tucci et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3317–3322
R. S.; Chen, C. Bioorg. Med. Chem. Lett. 2002, 12, 2179. (d)
Gross, T. D.; Zhu, Y.-F.; Saunders, J.; Gao, Y.; Connors,
P. J., Jr.; Guo, Z.; Struthers, R. S.; Reinhart, G. J.; Chen, C.
Bioorg. Med. Chem. Lett. 2002, 12, 2185. (e) Tucci, F. C.; Zhu,
Y.-F.; Guo, Z.; Gross, T. D.; Connors, P. J., Jr.; Struthers,
R. S.; Reinhart, G. J.; Wang, X.; Saunders, J.; Chen, C.
Bioorg. Med. Chem. Lett. 2002, 12, 3491. (f) Zhu, Y. F.; Guo,
Z.; Gross, T. D.; Gao, Y.; Connors, P. J.; Struthers, R. S.; Xie,
Q.; Tucci, F. C.; Reinhart, G. J.; Wu, D.; Saunders, J.; Chen,
C. J. Med. Chem. 2003, 46, 1769.