Novel syntheses of the amino-1,2,4-triazine GW356194: Identification of a synthesis amenable to scale up

Article abstract of DOI:10.1016/S0040-4039(03)01359-5

New syntheses of the amino-1,2,4-triazine, GW356194 have been developed to improve on existing methodology. The use of different amidrazones in cyclisations with α-keto and α-amido carbonyl compounds as the key step for the synthesis of the 1,2,4-triazine core was evaluated and the results are presented here.

Full text of DOI:10.1016/S0040-4039(03)01359-5

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 44 (2003) 5657–5659
Novel syntheses of the amino-1,2,4-triazine GW356194:
identification of a synthesis amenable to scale up
Fiona M. Adam, Andrew J. Burton,* Kevin S. Cardwell, Richard A. Cox, Richard A. Henson,
Keith Mills, Jeremy C. Prodger, Mark B. Schilling and Daniel T. Tape
GlaxoSmithKline R&D, Chemical Development, Medicines Research Centre, Stevenage, Hertfordshire SG1 2NY, UK
Received 23 April 2003; accepted 30 May 2003
Abstract—New syntheses of the amino-1,2,4-triazine, GW356194 have been developed to improve on existing methodology. The
use of different amidrazones in cyclisations with a-keto and a-amido carbonyl compounds as the key step for the synthesis of the
1,2,4-triazine core was evaluated and the results are presented here.
© 2003 Elsevier Ltd. All rights reserved.
Lamotrigine 3, a sodium channel blocker (Fig. 1), is the
active component of Lamictal® and is in clinical use as
an anticonvulsant therapy. As part of a programme to
identify other potential sodium channel blockers for the
treatment of disorders of the central nervous system,
chloride gave 8. Subsequent photolytic isomerisation
and concurrent thermally mediated cyclisation gave
amino-1,2,4-triazine 9. The SMe group was then
removed by oxidation to the sulfone and sodium boro-
hydride mediated reduction.1
the
5-amino-1,2,4-triazine,
GW356194
4
was
discovered.1
It was decided that the initial synthesis would not be
viable in the long term and therefore we initiated a
programme to identify alternative routes for scale up
and manufacture of GW356194. The synthesis of 1,2,4-
triazines has been extensively reviewed.2 These hetero-
cycles are often derived from 1,2,4-triazinones,
accessible via condensation of amidrazones3 with a-
keto acids, amides and esters and so we envisaged that
these would be key intermediates in the synthesis of
GW356194.
The synthesis of Lamotrigine involves the condensa-
tion/cyclisation of aminoguanidine 2 with the acylnitrile
1. Unfortunately, condensation of the corresponding
trichlorobenzoylnitrile with amidrazones results in poor
1,2,4-triazine formation with acylation being the main
reaction pathway. Therefore, during the lead optimisa-
tion phase it was necessary to use a stepwise synthesis
of GW356194 4 (Scheme 1). Thus, condensation of the
methylthiosemicarbazide HI salt 6 with the primary
ketoamide 5, followed by dehydration using phosphoryl
Initially we found that the low yield in the condensa-
tion of 5 with 6 was due to the formation of the
1,2,4-triazinone 10 as a side product.4 Further investi-
gation of this reaction demonstrated that reaction of 5
and 6 under isomerising conditions gave smooth con-
version to 10 in 91% yield. The 1,2,4-triazinone 10
could also be converted to amino-1,2,4-triazine 9 using
standard chlorination/amination methodology (Scheme
1). This result encouraged us to consider the use of
other 1,2,4-triazinones as key intermediates for the
synthesis of GW356194 4.
Figure 1.
The most direct approach to GW356194 4, via the
1,2,4-triazinone 13 (Scheme 2) was investigated initially.
We anticipated that the 1,2,4-triazine core of this
molecule could be produced by reacting the readily
available amidrazones 12, 15 and 21 with the a-keto
Keywords: 1,2,4-triazine; 1,2,4-triazinone; amidrazone; thiosemicar-
bazide; formamidrazone; ethyl oxamidrazonate.
* Corresponding author. Tel.: +44 1438 763667; fax: +44 1438
764869; e-mail: andrew.j.burton@gsk.com
0040-4039/$ - see front matter © 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0040-4039(03)01359-5

5658
F. M. Adam et al. / Tetrahedron Letters 44 (2003) 5657–5659
Scheme 1. Reagents and conditions: (a) EtOH, reflux, 63%; (b)
(POCl2)2O, 1,4-dioxane, rt, 61%; (c) propan-1-ol, hw, 77%; (d)
i. mCPBA, DMA, EtOAc, ii. NaBH4, EtOH, 63%; (e) EtOH,
reflux, hw, 91%; (f) i. oxalyl chloride, 1,4-dioxane, ii. NH3,
propan-2-ol, 91%.
Scheme 2. Reagents and conditions: (a) EtOH, rt; (b) i.
NH2NH2, AcOH; ii. HC(OMe)3, MeSO3H; (c) NH3, EtOH,
hw, reflux, yields up to 40% (two stages); (d) EtOH, hw, reflux,
54%; (e) NaOH, reflux, 81%.
ester 11, available by acid catalysed ethanolysis of 5, or
one of the a-keto amides 17, 18, 19 and 20, the synthe-
ses of which are described in the preceding paper.5
isomerisation of the intermediate hydrazone was medi-
ated by base rather than photolysis. The thio-1,2,4-tri-
azinone 22 could then be treated with aqueous
hydrogen peroxide, according to an optimised proce-
dure,9 to give mixtures of the corresponding sulfinic
and sulfonic acids which, on treatment with concen-
trated HCl, eliminated to give 13 in good yield.
Formamidrazone 12, formed in situ by the treatment of
formamidinium acetate with hydrazine in ethanol,6 was
the amidrazone of choice for the synthesis of 13, due to
potential formation of 13 without the need for further
modification. However, despite some success in reacting
formamidrazone with methyl benzoylformate, for-
mamidrazone failed to react with the a-keto ester 11.
We assume this is due to increased steric hindrance of
the ketone and the instability of formamidrazone.
Therefore, we resorted to a stepwise build up of the
1,2,4-triazinone 13 as shown in Scheme 2. Thus, hydra-
zone formation, followed by formation of iminoether
14 and treatment with ammonia, without isolation,
under photolysis conditions gave 13 in moderate yield.
Having developed several syntheses of 13 we needed to
effect the final conversion to GW356194 4. In stark
contrast to the conversion of 10 to 9, this proved more
difficult than first imagined and we found that the
chloro-1,2,4-triazine 23 was particularly unstable. How-
ever, after screening a large number of chlorinating
agents, we were able to form chloro-1,2,4-triazine 23,
using known conditions by treatment of 13 with >2
equiv. PCl5 in toluene at 90°C.10 Treatment of the
toluene solution of 23 with ammonia in THF gave
GW356194 4 in 40% yield. Significant quantities of
polymeric material were formed both in this reaction
and also on deviation from the successful chlorinating
conditions or on attempted isolation of 23. This poly-
merisation is probably due to the increased elec-
trophilicity of the 3 position of the 1,2,4-chlorotriazine
and ring opening of the heterocycle after attack in this
position.
Reaction of ethyl oxamidrazonate 15,7 with the a-keto
ester 11, under photolysis conditions gave the 1,2,4-tri-
azinone 16 in 54% yield and subsequent decarboxyla-
tion in refluxing sodium hydroxide gave 13 in good
yield (Scheme 2). Reaction of 15 with a-keto amides
was unsuccessful and degradation occurred, we assume
that this is because of the larger energy barrier for
cyclisation with an amide compared to an ester.
Thiosemicarbazide 21 is a commercially available
amidrazone and we envisaged that if the thio-1,2,4-tri-
azinone 22 could be formed, the unwanted thiol could
be removed by oxidative elimination to give 13 (Scheme
3).8 We found that the best conditions for effecting the
cyclisation were to treat the a-keto amides 17, 18, 19
and 20 with thiosemicarbazide in refluxing sodium
hydroxide and the resulting thio-1,2,4-triazinone 22
could be isolated by neutralisation of the reaction mix-
ture. This procedure was a significant improvement on
the other methods for 1,2,4-triazinone formation, as the
Despite the success of using 1,2,4-triazinone 13 as an
intermediate for the synthesis of GW356194 4, we felt
that the final chlorination–amination was not robust
enough to be practicable on a large scale. Therefore, we
decided to see if intermediate 22 could be converted
into the SMe-1,2,4-triazinone 10. Indeed, treatment of
22 with methyl iodide gave 10 in excellent yield and it
was decided that this was the approach to GW356194 4
most amenable to the scale up and manufacture of this
compound.

F. M. Adam et al. / Tetrahedron Letters 44 (2003) 5657–5659
5659
Scheme 3. Reagents and conditions: (a) NaOH, reflux; (b) i. NaOH, EtOH, H2O2, ii. cHCl, 75%; (c) PCl5, toluene 90°C, >95%
solution yield; (d) NH3, THF, toluene, 40%; (e) MeI, NaOH, EtOH, 91%.
In conclusion, we have explored several new
approaches to GW356194 4 and identified an efficient
approach, based on the use of thiosemicarbazide as a
key building block for the synthesis of the core hetero-
cycle. The robustness of this key reaction has been
demonstrated by the scale up of this process for a-keto
amides 18 and 19 to produce >100 g of 22.
3. Neilson, D. G.; Roger, R. J.; Heatlie, J. W. M.; New-
lands, L. R. Chem. Rev. 1970, 70, 151.
4. All new compounds gave concordant 1H NMR, 13C
NMR and mass spectral analyses.
5. Burton, A. J.; Cardwell, K. S.; Fuchter, M. J.; Lindvall,
M. K.; Patel, R.; Packham, T. W.; Prodger, J. C.;
Schilling, M. B.; Walker, M. D. Tetrahedron Lett. 2003,
44, 5653.
6. Neunhoeffer, H.; Weischedel, F. Liebigs Ann. Chem.
1971, 749, 16.
References
7. Schmidt, P.; Druey, J. Helv. Chim. Acta 1955, 38, 1560.
8. Gianturco, M. Gazz. Chim. Ital. 1952, 82, 595.
9. Kamiyama, T.; Enomoto, S.; Inoue, M. Chem. Pharm.
Bull. 1988, 36, 2652.
1. PCT Int. Appl. (1999), WO 9932462 A1.
2. See for example H. Neunhoeffer in Methods of Organic
Chemistry (Houben-Weyl); Ernst Schaumann, Ed.;
George Thieme Verlag: Stuttgart, New York, 1998; Vol.
E9 c, pp. 582–666 and references cited therein.
10. Neunhoeffer, H.; Reichel, D.; Cullmann, B.; Rehn, I.
Liebigs Ann. Chem. 1990, 631.