Synthesis of benzoylpyrimidines as antagonists of the corticotropin- releasing factor-1 receptor

Article abstract of DOI:10.1016/j.bmcl.2004.05.072

A series of benzoylpyrimidines derived from the anilinepyrimidine CRF ₁ antagonists were synthesized. Several synthetic routes were developed to explore the SAR of this series of compounds. Compounds such as 8d (K_i=15nM) exhibited high binding affinities at the human CRF ₁ receptor.

Full text of DOI:10.1016/j.bmcl.2004.05.072

Bioorganic & Medicinal Chemistry Letters 14 (2004) 3869–3873
Synthesis of benzoylpyrimidines as antagonists
of the corticotropin-releasing factor-1 receptor
Thomas R. Webb,*,ꢀ Terry Moran, Charles Q. Huang, James R. McCarthy,
Dimitri E. Grigoriadis and Chen Chen*
Department of Medicinal Chemistry and Department of Pharmacology, Neurocrine Bioscience, Inc.,
10555 Science Center Drive, San Diego, CA 92121, USA
Received 1 April 2004; revised 25 May 2004; accepted 27 May 2004
Available online 19 June 2004
Abstract—A series of benzoylpyrimidines derived from the anilinepyrimidine CRF1 antagonists were synthesized. Several synthetic
routes were developed to explore the SAR of this series of compounds. Compounds such as 8d (Ki ¼ 15 nM) exhibited high binding
affinities at the human CRF1 receptor.
Ó 2004 Elsevier Ltd. All rights reserved.
1. Introduction
N
N
N
2
1
5
Cl
R
4
R
N
N
N
Corticotropin-releasing factor (CRF) is a 41 amino acid
peptide C-terminal amide, which stimulates adrenocor-
ticotropic hormone (ACTH) release from the pituitary1
and is a neuromodulator or neurotransmitter, which is
involved in neural stress induced responses in the brain.2
CRF exerts its action by binding to and activating the
CRF1 and CRF2 receptors, both belong to the class B
G-protein-coupled receptor superfamily, which have
recently been cloned and expressed.3 The pharmacology,
physiology and role for CRF in CNS disorders have
been extensively reviewed.4
R
6
N
HN
N
Cl
N 3
Cl
O
N
1
2
Ar
Cl
NBI 27914 (2)
7-10
CP-154,526 (1)
Figure 1. Small molecule CRF1 antagonists.
proposed that the dihedral angle between the phenyl
group and the pyrimidine ring of 2 is very important for
receptor binding.7 To further explore and understand
the structure–activity relationships of these CRF1
antagonists, we replaced the nitrogen linker in 2 with a
carbon moiety, which may alter the relation of the two
aromatic rings. In addition, this change eliminates the
potential liability of the aniline functionality. Here we
report the synthesis and SAR of this class of compounds
as CRF1 receptor antagonists.
Clinical evidence suggests that over-stimulation of CRF
may result in several diverse neuropsychiatric diseases
including depression, anxiety and stress related disor-
ders, and therefore CRF antagonists have the potential
to treat these diseases. Since the first nonpeptide CRF
antagonist CP-154,526 (1) disclosed in 1996,5 many
potent small molecule CRF antagonists from different
chemical classes have been reported.6 A series of analino-
pyrimidines exemplified by NBI 27914 (2) has been
reported as potent CRF1 antagonists (Fig. 1).7 We have
2. Chemistry
Several different synthetic routes were investigated in
order to obtain the benzoylpyrimidine derivatives with
various substitutions at the 4-, 5- and 6-positions of the
pyrimidine ring. The initial synthesis of 7 is showed
in Scheme 1. 4-Chloro-2,6-dimethylpyrimidine was
Keywords: Benzoylpyrimidine; CRF; Antagonist.
* Corresponding authors. Tel.: +1-858-658-7600; fax: +1-858-658-76-
19; e-mail: cchen@neurocrine.com
ꢀ
Present address: ChemBridge Corporation, 16981 Via Tazon, San
Diego, CA 92127, USA.
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.05.072

3870
T. R. Webb et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3869–3873
N
N
N
c
N
N
Ar
N
COOMe
N
Cl
O
5
b
6
a
N
N
+
N
N
N
N
N
N
4
3
c
N
N
Ar
MeOOC
O
14a
7
Scheme 1. Reagents and conditions: (a) c-PrCH2NHPr/heat, 93%; (b) i. SeO2, ii. SOCl2/MeOH, then separation, 20%; (c) ArMgX/THF, 40–60%.
allowed to react with N-propyl-N-cyclopropanemethyl-
amine to give the aminopyrimidine 4 in 93% yield.
Although Sakasai et al.8 have shown that it is possible to
regioselectively oxidize 2,4-dimethylpyrimidine to give
pyrimidine-4-carboxylic acid, applying these conditions
(SeO2/pyridine, reflux, followed by esterification) to 4
gave a mixture of isomers 5 and 14a, along with a diester
by-product. Compounds 5 and 14a were readily sepa-
rated by flash column chromatography in about 20%
yield each. Reaction of 5 or 14a with an aryl Grignard
reagent afforded the ketones 6 and 7, respectively, in
moderate yields (40–60%). The regio-isomeric structures
of 6a and 7a were assigned based on observed NOE
between the 5-proton and the 6-methyl group of 6a.
into the desired 6-benzoylpyrimidines 7–10 in good
yields (40–80%) with various aryl Grignard reagents
(Scheme 2). Alternately, the 4-chloropyrimidine 13b was
treated with 2,4,6-trimethylphenylmagnesium bromide
to give the 4-chloro-6-(2,4,6-trimethylbenzoyl)pyrim-
idine 15b, in 90% yield, which was subjected to a nucle-
ophilic replacement with various alkylamines to afford
the final products 8g–p.
Since the cyclization of 2-oxosuccinate with acetamidine
provided the pyrimidines in relative poor yield (10% for
12a),9 we explored other alternatives to synthesize the
pyrimidine-esters 12a–d. Thus, reaction of diethyl
fumarate (16a, containing a trace of DMF) with chlo-
rine gas gave a chlorinated intermediate,10 which up on
treatment with sodium ethoxide in ethanol gave pre-
sumably the chlorofumarate 17a. Cyclization of 17a
with acetamidine hydrochloride in the presence of a base
such as sodium ethoxide afforded the desired pyrimidine
12a in 56% yield. Similarly, 12b was synthesized from
16b.
The 5-alkyl pyrimidinones 12b–d were obtained in
moderate yields (10–40%) by the reaction of 3-alkyl-2-
oxosuccinates 11b–d with acetamidine hydrochloride in
the presence of NaOEt, similar to the reported proce-
dure.9 Treatment of 12 with POCl3 gave the corre-
sponding 4-chlorpyrimidines 13, in almost quantitative
yields, which were converted to the 4-aminopyridines 14
with N-propyl-N-cyclopropanemethylamine in about
90% after purification. Finally, 14a–d were transformed
The benzoylpyrimidines 8b–e were further modified as
showed in Scheme 3. Palladium-catalyzed hydrogena-
2
1
R
R
N
N
Cl
N
OH
R
R
EtOOC
O
R
R
c
b
a
N
N
N
EtOOC
EtOOC
COOEt
EtOOC
N
14a-d: R1NR2 = c-PrHCH2NPr
12a-d
11b-d
13a-d
e
d
2
1
R
R
N
N
Cl
OEt
O
R
R
O
R
f
EtOOC
R
g
N
N
Ar
Ar
N
EtOOC Cl
COOEt
O
16a-b
7: R = H
8: R = Me
9: R = Et
17a-b
15b: R = Me;
Ar = 2,4,6-Me3Ph
a: R = H; b: R = Me; c: R = Et; d: CH2CH=CH2
10: R = CH2CH=CH2
Scheme 2. Reagents and conditions: (a) acetamidineÆHCl/NaOEt/EtOH, 10–40%; (b) POCl3/reflux, ꢀ100%; (c) c-PrCH2NHPr/heat, 90%; (d)
ArMgX/THF, 40–80%; (e) 2,4,6-Me3PhMgBr/THF, 93%; (f) R1R2NH/heat, 30–90%; (g) i. Cl2/CH2Cl2, ii. NaOEt/EtOH, iii. acetamidineÆHCl,
ꢀ60%.

T. R. Webb et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3869–3873
3871
N
N
N
N
N
N
N
N
a
d
e
O
N
R'
HO
O
N
O
N
O
N
O
R'
O
O
R'
O
O
O
8b, 8d or 8e
22
21
18
b
c
a
N
N
N
N
N
N
N
N
f
HO
R'
O
N
O
N
R'
N
O
N
O
O
O
R'
O
19
20a: R' = MeO
20b: R = Me
20c: R = Et
24
23
Scheme 3. Reagents and conditions: (a) H2/Pd/EtOH; (b) Me3Al/Me3SiOTf/CH2Cl2; (c) NaBH4/MeOH; (d) MeMgBr/THF; (e) MsCl/Et3N; (f)
MsCl/Et3N, then NaOMe/MeOH.
tion of 8b afforded the methylene 18 in 45% yield. The
gem-dimethyl compound 19 was isolated from the
treatment of 8b with trimethylaluminium–trimethyl-
silyltriflate in dichloromethane.11 Reactions of 8b with
methyl Grignard reagent provided the tertiary alcohol
21, which was further modified. Treatment of 21 with
methanesulfonyl chloride in the presence of triethyl-
amine afforded the olefin 22, which was reduced under
hydrogenation conditions to give the ethylidene 23.
Reduction of ketone 8b, 8d and 8e gave the corre-
sponding alcohols 20a–c in good yields. Conversion of
20c to the corresponding methyl ether 24 was accom-
plished with methanesulfonyl chloride and triethyl-
amine, followed by sodium methoxide in methanol.
affinity to the CRF1 receptor. Since we speculated a
replacement of the nitrogen linker of 2 with a sp2-carbon
of a carbonyl group may alter the geometry of the two
aromatic rings, this change may provide compounds
with different side-chain decorations. In addition, this
modification should also change the physicochemical
property of this series, and remove the potentially
undesirable aniline moiety.
The initial 5-unsubstituted pyridines 7 had modest
binding affinities at the CRF1 receptor. For example, 7b
with a 2,4,6-trimethylphenyl group had a Ki value of
200 nM. Introduction of a methyl group at the 5-posi-
tion of the pyrimidine increased the binding affinity over
13-fold (8d, Ki ¼ 15 nM). In comparison to 8d, the 2-
methylphenyl analog 8c exhibited a Ki value of 810 nM,
a 54-fold reduction in binding affinity, the 2,4,6-tri-
methoxyphenyl analog (8b, Ki ¼ 38 nM), however, was
only slightly less active, and the 2,4,6-triethylphenyl
compound displayed a similar Ki value (8e, Ki ¼ 13 nM).
Both the 5-position of the pyrimidine and the ortho-
position of the 6-benzoyl group have direct impact on
the relative conformation of these two aromatic rings.
Interestingly, 5-ethyl and the bigger 5-allyl group had
minimal effect on the receptor binding (Ki ¼ 55 and 31
for 9 and 10, respectively).
3. Results and discussion
The synthesized compounds were tested for their bind-
ing affinities at the cloned human CRF1 receptor in a
competition binding assay as previously described, and
the Ki values were determined from concentration–
response curves using concentrations ranging from 1 nM
to 10 lM.7;12a The structure–activity relationships of
these compounds are summarized in Table 1. Selected
compounds were also measured for their abilities to
inhibit CRF-stimulated cAMP production in cells
expressing CRF1 receptor to assess their functional
antagonism.12
The SAR study from the anilinopyrimidine series has
previously established for the alkylamine at the 4-posi-
tion of the pyrimidine, and a small aliphatic side-chain
such as dipropylamine and N-propyl-N-cyclopropane-
methylamine is optimal for the receptor binding.7 For
this benzoylpyrimidine series, it also seems to be true.
Thus, N-propyl-N-cyclopropylmethylamine proved to
be the most effective side-chain at the 4-position of the
pyrimidine. Other small hydrophobic amine substituents
were tolerated (8g and 8h). Cyclic amines yielded
derivatives that were much less active (8m and 8o).
These results seem to indicate a preference for the size
Previous studies have led to a proposed pharmacophore
for potent CRF receptor antagonists.7;13 A major fea-
ture of this pharmacophore is the orthogonal relation
between an aromatic group and a nitrogen-containing
heterocyclic ring. In the anilinopyrimidine series (i.e., 2),
a 2,6-substituted analine at the 6-position and a small
group such as methyl and chloro moiety at the 5-posi-
tion of the pyrimidine are crucial for high binding

3872
T. R. Webb et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3869–3873
Table 1. Structure–activity relationships of benzoylpyrimidines 6–10
2
1
R
R
N
N
N
R
O
N
O
N
O
N
Ar
O
O
6a
R1NR2
7-10
Ar
Compound
R
Ki (nM)
2
2.7
630
6a
7a
7b
7c
8a
8b
8c
8d
8e
8f
H
n-PrNCH2Pr-c
n-PrNCH2Pr-c
n-PrNCH2Pr-c
n-PrNCH2 Pr-c
n-PrNCH2Pr-c
n-PrNCH2Pr-c
n-PrNCH2Pr-c
n-PrNCH2Pr-c
n-PrNCH2Pr-c
n-Pr2N
2,4,6-MeOPh
2,4,6-MePh
2,3,4,5,6-MePh
2,4-MeOPh
2,4,6-MeOPh
2-MePh
1100
200
H
H
Me
>10,000
810
38
Me
Me
810
15
Me
Me
2,4,6-MePh
2,4,6-EtPh
13
Me
Me
2,3,4,5,6-MePh
2,4,6-MePh
2,4,6-MePh
2,4,6-MePh
2,4,6-MePh
2,4,6-MePh
2,4,6-MePh
2,4,6-MePh
2,4,6-MePh
2,4,6-MePh
2,4,6-MePh
2,4,6-MePh
2,4,6-MePh
5100
58
8g
8h
8i
Me
Me
(CH2@CHCH2)2N
EtNHBn-n
PrNBn-n
38
470
8j
Me
Me
180
540
8k
8l
n-PrNCH2CH2OH
(MeOCH2CH2)2N
Pyrrolidin-1-yl
Me
Me
1200
>10,000
380
8m
8n
8o
8p
9
Me
Me
(2S)-MeOCH2-pyrrolidin-1-yl
Morpholin-4-yl
Et2CHNH
>10,000
220
55
Me
Et
n-PrNCH2Pr-c
n-PrNCH2Pr-c
10
CH2@CHCH2
31
Table 2. Structure–activity relationships of benzylpyrimidines 18–24
and/or orientation of the hydrophobic substituents at
this position.
N
N
4
We intended to break the possible conjugation between
the carbonyl group and the aromatic rings by modifying
the C@O moiety. The methylene analog (18,
Ki ¼ 26 nM) of 8b exhibited the similar binding affinity
as the parent, however, the ethylidene (23, Ki ¼ 480 nM)
and the hydroxymethylene (20a, Ki ¼ 1:9 lM) were
much less active. This decrease in binding affinity was
most likely caused by conformational change since both
the CHMe and CHOH groups are not much larger in
size than the C@O moiety. The bulkier gem-dimethyl
group abolished the binding affinity of 19. The low
activity of the alcohols 20 caused at least partially by the
polarity of the hydroxyl group, since the corresponding
methyl ether (24, Ki ¼ 90 nM) recovered some activity
from its hydroxyl analog (20c, Ki ¼ 360 nM). An at-
tempt to mimic the C@O group with a C@CH2 moiety
was also unsuccessful and the resultant compound
exhibited almost 20-fold reduction in binding (22,
Ki ¼ 560 nM). These results imply the importance of the
correct dihedral angle between the phenyl group and the
pyrimidine ring for high binding affinity to the CRF1
receptor (Table 2).
R
3
N
R
Ar
18-24
CR3R4
Compound
Ar
Ki (nM)
18
23
2,4,6-MeOPh
2,4,6-MeOPh
2,4,6-MeOPh
2,4,6-MePh
2,4,6-EtPh
CH2
26
480
CHMe
CHOH
CHOH
CHOH
CHOMe
C(Me)2
C(OH)Me
C@CH2
20a
20b
20c
24
1900
79
360
90
2,4,6-EtPh
19
21
2,4,6-MeOPh
2,4,6-MeOPh
2,4,6-MeOPh
>10,000
1900
560
22
1.0 lM (average of four independent measurements),
and 8d had an IC50 of 1.2 lM (average of three mea-
surements). In addition, 8d was also tested and found to
inhibit CRF-stimulated ACTH release from primary rat
anterior pituitary cell cultures with an IC50 value of
0.98 lM.14 Neither compounds 8b and 8d nor any other
analogs examined had any functional intrinsic activity
for cAMP release suggesting that these molecules are
indeed functional CRF1 receptor antagonists.
Selected compounds were tested for functional antago-
nism, and these compounds were able to inhibit CRF-
stimulated release in CHO cells expressing the CRF1
receptor. For example, compound 8b dose-dependently
inhibited cAMP production with an IC50 value of
Conformational analyses of compounds 2, 8b, 18 and 19
using MedChem Explorer15 revealed that there was a

T. R. Webb et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3869–3873
3873
References and notes
1. Vale, W.; Spiess, J.; Rivier, C.; Rivier, J. Science 1981, 213,
1394.
2. Koob, G. F. Perspect. Behav. Med. 1985, 2, 39.
3. (a) Chang, C. P.; Pearse, R. I.; O’Connell, S.; Rosenfeld,
M. G. Neuron 1993, 11, 1187; (b) Chen, R.; Lewis, K. A.;
Perrin, M. H.; Vale, W. W. Proc. Natl. Acad. Sci. U.S.A.
1993, 90, 8967; (c) Vita, N.; Laurent, P.; Lefort, S.;
Chalon, P.; Lelias, J. M.; Kaghad, M.; Le, F. G.; Caput,
D.; Ferrara, P. FEBS Lett. 1993, 335, 1; (d) Perrin, M. H.;
Donaldson, C. J.; Chen, R.; Lewis, K. A.; Vale, W. W.
Endocrinology 1993, 133, 3058; (e) Liaw, C. W.; Loven-
berg, T. W.; Barry, G.; Oltersdorf, T.; Grigoriadis, D. E.;
De Souza, E. B. Endocrinology 1996, 137, 72; (f) Loven-
berg, T. W.; Liaw, C. W.; Grigoriadis, D. E.; Clevenger,
W.; Chalmers, D. T.; De Souza, E. B.; Oltersdorf, T. Proc.
Natl. Acad. Sci. U.S.A. 1995, 92, 836.
Figure 2. Overlay of low-energy conformers of compounds 2 (light
blue), 8b (yellow), 18 (green) and 19 (red).
4. (a) Owens, M. J.; Nemeroff, C. B. Pharmacol. Rev. 1991,
43, 425; (b) Holsboer, F. Curr. Opin. Invest. Drugs 2003, 4,
46.
5. Schulz, D. W.; Mansbach, R. S.; Sprouse, J.; Braselton, J.
P.; Collins, J.; Corman, M.; Dunaiskis, A.; Faraci, S.;
Schmidt, A. W.; Seeger, T.; Seymour, P.; Tingley, F. D.,
III; Winston, E. N.; Chen, Y. L.; Heym, J. Proc. Natl.
Acad. Sci. U.S.A. 1996, 93, 10477.
6. For recent reviews, see: (a) Grigoriadis, D. E.; Haddach,
M.; Ling, N.; Saunders, J. Curr. Med. Chem. 2001, 1, 63;
(b) Saunders, J.; Williams, J. Prog. Med. Chem. 2003, 41,
195.
good match between the low-energy conformers of 2
and 8b. The phenyl ring of 18 or 19 was not aligned well
with that of 2 (Fig. 2). Since the methylene moiety of 18
is much more flexible than the gem-dimethyl substituted
carbon of 19, the correct conformer of 18 for receptor
binding will cost much less energy than that of 19. That
may explain why both 8b and 18 had similar binding
affinity while 19 was inactive.
7. Chen, C.; De Souza, E. B.; Grigoriadis, D. E.; Huang, C.
Q.; Kim, K. I.; Lui, Z.; Moran, T.; Webb, T. R.; Whitten,
J. P.; Xie, M.; McCarthy, J. R. J. Med. Chem. 1996, 39,
4358.
4. Conclusion
8. Sakasai, T.; Sakamoto, T.; Yamanaka, H. Heterocycles
1979, 13, 235.
9. Redd, J. T.; Bradshaw, J. S.; Huszthy, P.; Izatt, R. M. J.
Heterocycl. Chem. 1994, 31, 1047.
10. Yasuda, N.; Yamatani, T.; Ohnuki, T.; Okutsu, M. J.
Heterocycl. Chem. 1984, 21, 1845.
11. Kim, C. U.; Misco, P. F.; Buh, B. Y.; Mansuri, M. M.
Tetrahedron Lett. 1994, 35, 3019.
12. (a) Grigoriadis, D. E.; Liu, X. J.; Vaughn, J.; Palmer, S.
F.; True, C. D.; Vale, W. W.; Ling, N.; De Souza, E. B.
Mol. Pharmacol. 1996, 50, 679; (b) De Souza, E. B. J.
Neurosci. 1987, 7, 88.
The structure–activity relationships of the novel benzo-
ylpyrimidines as antagonists of the CRF1 receptor were
explored. Synthetic approaches were developed that
allowed for the facile divergent modification of the
6-aryl or the 4-amino substituents. The carbonyl func-
tionality of the benzoyl group was found to be the most
desirable among the carbon linkers explored. This is
probably related to a proper dihedral angle between the
two aromatic rings. Several potent compounds were
discovered during this study (e.g., 8d and 8e).
13. Hodge, C. N.; Aldrich, P. E.; Wasserman, Z. R.; Fernan-
dez, C. H.; Nemeth, G. A.; Arvanitis, A.; Cheeseman, R.
S.; Chorvat, R. J.; Ciganek, E.; Christos, T. E.; Gilligan,
P. J.; Krenitsky, P.; Scholfield, E.; Strucely, P. J. Med.
Chem. 1999, 42, 819.
14. Battaglia, G.; Webster, E. L.; De Souza, E. B. Synapse
1987, 1, 572–581.
15. MedChem Explorer Version 2.1.1., Accelrys, Inc., http://
www.accelrys.com.
Acknowledgements
This work was supported in part by a grant funded
through the Small Business Innovative Research (SBIR)
program at NIH, identification number 1R43
NS334879.