Synthesis of aryl-1,2,4-triazine-3,5-diones as antagonists of the gonadotropin-releasing hormone receptor

Article abstract of DOI:10.1016/j.bmcl.2005.06.057

Several efficient synthetic routes for 2-, 4-, and 6-aryl-1,2,4-triazine-3, 5-diones were developed. Derivatives were synthesized and studied as gonadotropin-releasing hormone antagonists in an effort to understand structure-activity relationships of the monocyclic compounds.

Full text of DOI:10.1016/j.bmcl.2005.06.057

Bioorganic & Medicinal Chemistry Letters 15 (2005) 4363–4366
Synthesis of aryl-1,2,4-triazine-3,5-diones as antagonists of
the gonadotropin-releasing hormone receptor
Joseph Pontillo,a Zhiqiang Guo,a Dongpei Wu,a R. Scott Struthersb and Chen Chena,*
aDepartment of Medicinal Chemistry, Neurocrine Biosciences Inc., 12790 El Camino Real, San Diego, CA 92130, USA
bDepartment of Endocrinology, Neurocrine Biosciences Inc., 12790 El Camino Real, San Diego, CA 92130, USA
Received 15 April 2005; revised 8 June 2005; accepted 10 June 2005
Available online 19 July 2005
Abstract—Several efficient synthetic routes for 2-, 4-, and 6-aryl-1,2,4-triazine-3,5-diones were developed. Derivatives were synthe-
sized and studied as gonadotropin-releasing hormone antagonists in an effort to understand structure–activity relationships of the
monocyclic compounds.
Ó 2005 Elsevier Ltd. All rights reserved.
In our efforts to identify orally active antagonists of the
gonadotropin-releasing hormone receptor (GnRH-R)
for possible treatment of sex hormone related diseases,
such as endometriosis, uterine fibroids, and prostate
cancer,1 we have discovered a series of 6-methyluracils
(6-methylpyrimidin-2,4-diones) exemplified by 1d having
low-nanomolar binding affinity.2 We have also conclud-
ed that the 4-oxo moiety of the uracil ring is critical for
high-affinity binding based on a large difference in poten-
cy between two regioisomers.3 Based on an internal
receptor modeling study, the oxo functionality is specu-
lated to have a hydrogen-bonding interaction with the
GnRH receptor. Another important feature of 1 is the
6-methyl group, which is optimal at this position for
receptor binding. Thus, the non-methyl uracil 3a
(Ki = 5.3 nM; Fig. 1)4 is about 10-fold less potent than
1d (Ki = 0.56 nM), and its 6-ethyl analogue 3b
(Ki = 50 nM) is almost 100-fold less active than 1d.
Although the 5-(2-fluoro-3-methoxyphenyl) group in
both 1d and 3a has an orthogonal conformation relative
to the uracil core in their crystal structures, 1d can exist
as two atropisomers, as evidenced by two sets of signals
in NMR spectra, while 3a does not. This suggests that
the rotation about the carbon–carbon bond connecting
the 5-phenyl ring of 1d is slower than that in 3a. The dif-
ferent rotational barrier could contribute to the discrep-
ancy in potency between 1d and 3a, and the bulky ethyl
group could be the cause of the large loss of potency in
3b. To further understand the role of the 6-methyl group,
as well as the 2-oxo moiety of 1, we synthesized several
phenyl substituted 1,2,4-triazine-3,5-diones (6-azaura-
cils) and studied them in the GnRH binding assay. Here-
in, we report the development of synthetic routes and
structure–activity relationships of these compounds.
The synthesis of 2-aryl-6-methyl-1,2,4-triazine-3,5-
dione, described in a patent publication, involves cycli-
zation of 5,5-dimethyl-2-phenyl-[1,2,4]triazolidin-3-one
with pyruvic acid.5 To apply this reaction to the synthe-
sis of the 6-benzyl analogue (Scheme 1), 2,6-difluorophe-
nylpyruvic acid
corresponding benzaldehyde 8 according to a known
protocol.6 Cyclization of
with 2-aryl-5,5-
9
was synthesized from the
9
dimethyltriazolidinone under acidic conditions (concd
H2SO4/dioxane/reflux) afforded the 6-(2,6-difluoroben-
zyl)-2-aryl-1,2,4-triazine-3,5-dione 10 in 32% yield.
Alkylation of triazinedione 10 by the Mitsunobu reac-
tion7 with N-Boc protected phenylalaninol (11, PPh3/
DEAD/THF/rt), followed by N-Boc deprotection with
trifluoroacetic acid, gave the final products 4 (43% for
4a, 42% for 4b). The preparation of 4-aryl-1,2,4-tri-
azine-3,5-diones 5 started from the benzylpyruvic acid
9 (Scheme 2). Cyclization of 9 with N-aryl thiosemi-
carbazide8 (prepared from aryl isothiocyanate and
hydrazine) in refluxing hydrochloric acid afforded the
5-thioxo-1,2,4-triazin-3-one 12 in 93% yield, which was
converted to the corresponding triazinedione 13 under
basic conditions (KOH/MeOH/rt, then aq HCl, 86%).
Mitsunobu reaction of 13 with phenylalaninol 11
Keywords: Synthesis; Triazinedione; Azauracil; GnRH; Antagonist.
*
Corresponding author. Tel.: +1 858 617 7600; fax: +1 858 617
7967; e-mail: cchen@neurocrine.com
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.06.057

4364
J. Pontillo et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4363–4366
Figure 1. Some monocyclic GnRH antagonists.
Scheme 1. Reagents and conditions: (a) i. AcNHCH2COOH/NaOAc/Ac2O; ii. NaOH, then aq HCl, 59%; (b) H2SO4 (cat.)/dioxane/reflux, 16 h, 32%;
(c) i. 11/PPh3/DEAD/THF/rt, 11 h; ii. TFA/CH2Cl2/rt, 0.5 h.
Scheme 2. Reagents and conditions: (a) Ar2NHC(@S)NHNH2/aq HCl/reflux, 16 h, 93%; (b) aq KOH/MeOH/rt, 0.5 h, then aq HCl, 86%;
(c) i. 11/PPh3/DEAD/THF/rt, 16 h; ii. TFA/CH2Cl2/rt, 0.5 h.
successfully installed the side chain at the 2-position of
the core, and the final products 5 were obtained after
N-Boc deprotection with TFA (65% for 5a, 52% for 5b).
nol under Mitsunobu conditions to provide the bromo
compounds 19a,b. Suzuki reaction13 of 19a,b with a
variety of arylboronic acids [Pd(PPh3)4/Ba(OH)2/
DME/C6H6/EtOH/reflux] gave the final products 6 after
deprotection. Replacement of the 6-bromo group of 19a
with a nucleophile such as phenol or alkylamine14 (Py/
EtOH/reflux) provided compounds 7 after N-Boc
removal. All final compounds were purified and charac-
terized by an HPLC–MS system.15
The 6-phenyl-1,2,4-triazine-3,5-dione (5-phenyl-6-aza-
uracil) 15 was obtained from reaction of benzoylformic
acid, which can be readily prepared by oxidation of ace-
tophenone,9 with semicarbazide to form a hydrazone,
which was then cyclized under basic conditions at elevat-
ed temperature (NaOEt/HOCH2CH2OH/reflux, 75%).10
Alkylation of 15 at the 2-position was accomplished
using a known protocol with 2,6-difluorobenzyl bromide
promoted by N,O-bis(trimethylsilyl)acetamide (MeCN/
80 °C) to give 16 in 88% yield.11 The aminoalkyl side
chain was introduced via a Mitsunobu reaction (11/
PPh3/DEAD/THF/rt) to provide the products 6 after
TFA deprotection (Scheme 3).
All of the synthesized compounds were tested in the
GnRH binding assay as previously described.2 In com-
parison with the s-triazinetriones 2,3 and uracils 3,4 or
6-methyluracils 1,2 the 2-phenyl-1,2,4-triazine-3,5-dione
4a had a Ki of 2.3 lM, which was slightly less potent
than the triazine analogue 2a (Ki = 1.1 lM), but much
less active than the 6-methyluracil 1a (Ki = 15 nM).
The 2-fluoro analogue 4b (Ki = 580 nM), however, was
much less potent than the corresponding triazine 2b
(Ki = 89 nM) and 6-methyluracil 1b (Ki = 5.6 nM), sug-
gesting the core structure of 4 possesses an unfavored
Alternatively, benzylation (ArCH2Br/BSA/MeCN/
80 °C) of the 6-bromo-1,2,4-triazine-3,5-dione 1712
afforded 18a,b, which were alkylated with phenylalani-

J. Pontillo et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4363–4366
4365
Scheme 3. Reagents and conditions: (a) i. NH2C(@O)NHNH2/EtOH/H2O/rt, 16 h, 88%; ii. NaOEt/(CH2OH)2/reflux, 15 h, 85%; (b) BSA/MeCN/
80 °C, 3 h; then (2,6-F)C6H3CH2Br/80 °C, 15 h, 74–88%; (c) 11/PPh3/DEAD/THF/rt, 11 h; (d) TFA/CH2Cl2/rt, 0.5 h; (e) Ar3B(OH)2/Pd(PPh3)4/aq
Ba(OH)2/DME/C6H6/EtOH/90 °C, 5 h, 52% for 6s; (f) RH/Py/EtOH/reflux, 15 h.
Table 2. SAR of 6-substituted 1,2,4-triazine-3,5-diones 6 and 7 at
GnRH-R
conformation for receptor binding. The binding
affinity of the 4-phenyl-1,2,4-triazine-3,5-dione 5a
(Ki = 450 nM) was slightly better than that of the tri-
Compound
X
Ar3 or R
Ki (nM)
azine 2a and 2-phenyltriazinedione 4a. Similarly, the
4-(3-methoxylphenyl)triazinedione 5b (Ki = 110 nM)
was only about 3-fold less potent than the triazine 2c
(Ki = 37 nM), although it was still much less active than
the 6-methyluracil 1c. The data summarized in Table 1
suggest that the 6-oxo moiety is more important for
receptor binding than the 2-oxo moiety of the triazine-
trione 2. Apparently, both 2- and 4-aryl-1,2,4-triazine-
3,5-diones are less desirable templates than either the
6a
6b
6c
6d
6e
6f
F
Ph–
2-ClC6H4–
1600
850
F
F
2,4-Cl2C6H3–
2-MeC6H4–
2600
1200
F
F
2,3-Me2C6H3–
2,5-Me2C6H3–
2,4,6-Me3C6H2–
2-EtC6H4–
1000
4700
F
6g
6h
6i
F
11,000
8400
F
F
2-(Me2NCH2)C6H4–
2-CF3C6H4–
>10,000
1300
8700
6j
F
triazinetriones
antagonists.
2 or 6-methyluracils 1 as GnRH
6k
6l
F
2-MeOC6H4–
2,4-MeO2C6H3–
2-MeO-5-FC6H3–
2-MeO-5-ClC6H3–
2-MeO-5-MeC6H3–
2,5-MeO2C6H3–
2-MeO-5-iPrC6H3–
2-EtOC6H4–
F
5200
6m
6n
6o
6p
6q
6r
6s
6t
F
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
92
Next, we conducted a more detailed study on a series of
1,2,4-triazine-3,5-diones with a 6-aryl ring (6, Table 2),
since this template (5-aryl-6-azauracils) is structurally
closest to the 5-aryluracils 3. Further, the described syn-
thesis allowed a rapid survey of substitution effects
at the 6-position. The 6-phenyl triazinedione 6a
(Ki = 1.6 lM) possessed binding affinity similar to the
2-phenyl analogue 4a. As an orthogonal relation be-
tween the 6-phenyl ring and the 1,2,4-triazine-3,5-dione
core is expected to be critical for high affinity, we sur-
veyed a series of ortho-substituted phenyl and bicyclic
aromatic analogues. 2-Chloro (6b) or 2,4-dichloro (6c)
substitution had minimal effect on potency, despite the
fact that incorporation of a 2-chloro in the 5-phenyl of
uracil 3 increases potency 6-fold.4 The 2-methyl substi-
F
F
F
F
F
F
2-F-3-MeOC6H3–
8-Quinolinyl–
1-Naphthyl–
F
>10,000
4300
6u
6v
6w
6x
7a
7b
7c
F
F
3-Benzothienyl–
2-F-3-MeOC6H3–
2-Cl-3-MeOC6H3–
PhO–
>10,000
13
CF3
CF3
F
2.3
>10,000
>10,000
>10,000
F
PhS–
1-Morpholinyl–
F
tution (6d), including the 2,3- or 2,5-dimethyl analogues
(6e,f), also had little impact on the binding. The more
bulky 2,4,6-trimethyl (6g, Ki = 11 lM), ethyl (6h,
Ki = 8.4 lM), and methoxy derivatives (6k, Ki = 8.7 lM)
also displayed reduced potency.
Table 1. Binding affinity of 2- and 4-aryl-1,2,4-triazine-3,5-diones 4
and 5
Compound
Ar1 or Ar2
Ki (nM)
4a
4b
5a
5b
Ph–
2300
580
450
110
The basic dimethylaminomethyl 6i was inactive, and the
strongly electron-withdrawing trifluoromethyl com-
pound (6j, Ki = 1.3 lM) possessed an almost identical
Ki value to the 2-methyl analogue 6d, suggesting no clear
2-FC6H4–
Ph–
3-MeOC6H4–

4366
J. Pontillo et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4363–4366
electronic preference for this phenyl ring. Any 2-meth-
oxyphenyl compound with an additional substituent
on the phenyl ring (6m–6g) showed no competitive bind-
ing, except the 2,4-dimethoxy analogue 6l (Ki = 5.2 lM).
The ethoxy 6r, as well as the 8-quinolinyl 6t and the
3-benzothienyl compound 6v, also displayed no binding
activity, while the 1-naphthyl derivative exhibited poor
affinity (6u, Ki = 4.3 lM). As predicted, the 2-fluoro-3-
methoxyphenyl compound 6s (Ki = 92 nM) displayed
an almost 20-fold increase in binding affinity over the
phenyl derivative 6a. Although these results are parallel
to those of the 6-methyluracils 1a (Ki = 15 nM) and 1d
(Ki = 0.56 nM), the 6-azauracil 6s was about 17-fold less
potent than its uracil analogue 3a. Thus, replacement
of the CH moiety of 3a with the smaller N atom at
the 6-position reduced binding affinity.
interactions with the receptor. 6x (Ki = 2.3 nM) was
identified as a potent GnRH antagonist from this series.
The improved potency of this compound was possibly
due to an increase in binding interactions of the 6-(2-
chloro-3-methoxyphenyl) group with amino acid resi-
dues such as Asp-122 on the human GnRH receptor.
References and notes
1. Zhu, Y.-F.; Chen, C.; Struthers, R. S. Annu. Rep. Med.
Chem. 2004, 39, 99.
2. Tucci, F. C.; Zhu, Y.-F.; Struthers, R. S.; Guo, Z.; Gross,
T. D.; Rowbottom, M. W.; Acevedo, O.; Gao, Y.;
Saunders, J.; Xie, Q.; Reinhart, G. J.; Liu, X.-J.; Ling,
N.; Bonneville, A. K. L.; Chen, T.; Bozigian, H.; Chen, C.
J. Med. Chem. 2005, 48, 1169.
3. Guo, Z.; Wu, D.; Zhu, Y.-F.; Tucci, F. C.; Pontillo, J.;
Saunders, J.; Xie, Q.; Struthers, R. S.; Chen, C. Bioorg.
Med. Chem. Lett. 2005, 15, 693.
Previously, we have shown in the non-methyl uracils
that a 2-fluoro-6-trifluoromethylbenzyl group at the
1-position (3c, Ki = 0.64 nM) possesses about 9-fold in-
creased affinity over the corresponding 2,6-difluoroben-
zyl analogue (3a, Ki = 5.3 nM).4 In the azauracil series,
the 2-fluoro-6-trifluoromethylbenzyl analogue 6w
(Ki = 13 nM) was about 7-fold better than the 2,6-diflu-
orobenzyl analogue 6s. Interestingly, the 2-chloro-3-
methoxyphenyl compound 6x (Ki = 2.3 nM) showed
another 6-fold binding affinity increase over the corre-
sponding 2-fluoro analogue 6w. These results may sug-
gest that the 3-methoxy group of the 6-phenyl ring in
6 plays a much more important role in receptor interac-
tion than that in 3a or 3c. On the basis of our receptor
modeling, the counterpart for this interaction could be
Asn-212 on transmembrane domain 5 of the human
GnRH receptor. This residue is known to be important
for hydrogen-bonding interaction with the pyro-gluta-
mine at the 1-position of the GnRH peptide.16,17
4. Guo, Z.; Chen, Y.; Huang, C. Q.; Gross, T. D.; Pontillo,
J.; Rowbottom, M. W.; Saunders, J.; Struthers, R. S.;
Tucci, F. C.; Xie, Q.; Wade, W.; Zhu, Y.-F.; Wu, D.;
Chen, C. Bioorg. Med. Chem. Lett. 2005, 15, 2519.
5. Lyga, J. W. PCT patent application WO8600072, 1986.
6. Audia, J. E.; Evrard, D. A.; Murdoch, G. R.; Froste, J. J.;
Nissen, J. S.; Schenck, K. W.; Fludzinski, P.; Lucaites, V.
L.; Nelson, D. L.; Cohen, M. L. J. Med. Chem. 1996, 39,
2773.
7. For a review, see: Hughes, D. L. Org. Prep. Proced. Int.
1996, 28, 127.
8. Mansourm, A. K.; Eid, M. M.; Hassa, R. A.; Haemers,
A.; Pattyn, S. R.; Vanden Berghe, D. A.; Van Hoof, L.
J. Heterocycl. Chem. 1988, 25, 279.
9. Beebe, X.; Nilius, A. M.; Merta, P. J.; Soni, N. B.; Bui, M.
H.; Wagner, R.; Beutel, B. A. Bioorg. Med. Chem. Lett.
2003, 13, 3133.
10. Sharma, B. P.; Lakha, R. J. Napal Chem. Soc. 1999, 17–
18, 8.
11. Jin, Y. H.; Liu, P.; Wang, J.; Baker, R.; Huggins, J.; Chu,
C. K. J. Org. Chem. 2003, 68, 9012.
12. Dudfield, P. J.; Le, V.-D.; Lindell, S. D.; Rees, C. W. J.
Chem. Soc., Perkin Trans. 1 1999, 2929.
13. For reviews, see: (a) Bellina, F.; Carpita, A.; Rossi, R.
Synthesis 2004, 15, 2419; (b) Kotha, S.; Lahiri, K.;
Kashinath, D. Tetrahedron 2002, 58, 9633; (c) Negishi,
E. In Handbook of Organopalladium Chemistry for Organic
Synthesis; Wiley-Interscience: New York, 2002, pp 767–
789; (d) Stanforth, S. P. Tetrahedron 1998, 54, 263.
14. Tzeng, C. C.; Motola, N. C.; Panzica, R. P. J. Org. Chem.
1983, 48, 1271.
15. For experimental details, see: Pontillo, J.; Guo, Z.; Wu, D.
PTC patent application WO03011841, 2003.
16. Hoffmann, S. H.; ter Laak, T.; Kuhne, R.; Reilander, H.;
Beckers, T. Mol. Endocrinol. 2000, 14, 1099.
17. Millar, R. P.; Lu, Z.-L.; Pawson, A. J.; Flanagan, C. A.;
Morgan, K.; Maudsley, S. R. Endocr. Rev. 2004, 25,
235.
Finally, 1,2,4-triazine-3,5-diones with 6-phenoxy (7a),
phenylsulfinyl (7b), and 1-morpholinyl group (7c) were
tested but these compounds showed no affinity in the
competitive binding, suggesting that the 6-aryl group
of 6 is important for receptor interaction.
In conclusion, several synthetic routes for 2-, 4-, and 6-
aryl-1,2,4-triazine-3,5-diones (1-, 3-, and 5-aryl-6-aza-
uracils) were developed for structure–activity relation-
ship study of the monocyclic GnRH antagonists. The
successful Mitsunobu reactions enabled us to install a
favored (2R)-aminophenethyl side chain at both the 2-
and 4-positions of the triazinediones. The results from
these SAR studies provide further evidence for the role
of the 6-methyl group in the uracil 1, which is important
for invoking a conformation that positions two key ele-
ments, the 5-phenyl and 1-benzyl groups, for favored