Palladium-catalyzed regioselective arylation of imidazo[1,2-b][1,2,4] triazine: Synthesis of an α2/3-selective GABA agonist

Article abstract of DOI:10.1021/jo0507035

A convergent, practical, and efficient synthesis of 2′,6-difluoro- 5′-[3-(1-hydroxy-1-methylethy)-imidazo[1,2-b][1,2,4]triazin-7-yl] biphenyl-2-carbonitrile (1), an orally active GABAA dag-selective agonist, is described. The seven-step, chromatography-fr

Full text of DOI:10.1021/jo0507035

Palladium-Catalyzed Regioselective Arylation of
Imidazo[1,2-b][1,2,4]triazine: Synthesis of an r2/3-Selective GABA
Agonist
Donald R. Gauthier, Jr.,* John Limanto, Paul N. Devine, Richard A. Desmond,
Ronald H. Szumigala, Jr., Bruce S. Foster, and R. P. Volante
Department of Process Research, Merck Research Laboratories, Merck & Co., Inc., P.O. Box 2000,
Rahway, New Jersey 07065
donald_gauthier@merck.com
Received April 8, 2005
A convergent, practical, and efficient synthesis of 2′,6-difluoro-5′-[3-(1-hydroxy-1-methylethyl)-
imidazo[1,2-b][1,2,4]triazin-7-yl]biphenyl-2-carbonitrile (1), an orally active GABAA
R2/3-selective
agonist, is described. The seven-step, chromatography-free synthesis was demonstrated on a multi-
kilogram scale and utilized biaryl bromide 6 and imidazotriazine 22 as key intermediates. Biaryl
bromide 6 was prepared via a highly selective aromatic bromination. The regioselective condensation
of aminotriazine 15 with chloroacetaldehyde provided the desired imidazotriazine intermediate
22. A highly regioselective palladium-catalyzed arylation in the final step highlights the efficiency
of the route.
γ-Aminobutyric acid (GABA) is a major inhibitory
neurotransmitter of the central nervous system.1 Selec-
tive ligands for GABAA receptors,2 especially those having
affinity for the R2, R3, and R5 subunit, are potentially
useful in the treatment of adverse conditions affecting
the central nervous system such as anxiety, convulsions,
and cognitive disorders. The discovery and development
of therapeutic agents to treat deficiencies in this field
have garnered extensive research efforts.3 Compound 1,
containing a uniquely functionalized imidazo[1,2-b][1,2,4]-
triazine core,4 is an orally active R2/3-selective GABAA
agonist possessing promising half-life and suitable oral
bioavailability for clinical development.5 Herein, we
report our efforts to design an efficient and practical
synthesis to provide bulk quantities of 1 to support
extensive biological testing.
Our retrosynthetic strategy for compound 1 is outlined
in Scheme 1. Encouraged by recent reports of metal-
catalyzed arylations of imidazoles6 and other electron-
rich heteroaromatics,7,8 we envisioned a highly conver-
gent approach that utilizes a selective Pd-catalyzed
arylation of imidazotriazine 22 with biaryl bromide 6.
The imidazotriazine core is derived via annulation of
(5) Bettati, M.; Blurton, P.; Carling, W. R.; Chambers, M. S.; Hallett,
D. J.; Jennings, A.; Lewis, R. T.; Russell, M. G. N.; Street, L. J.;
Szekeres, H. J.; Van Niel, M. B. Preparation of 7-phenylimidazo[1,2-
b][1,2,4]triazines as ligands for GABA receptors. PCT Int. Appl. 2002,
156 pp; WO 2002038568.
(6) (a) Pivsa-Art, S.; Satoh, T.; Kawamura, Y.; Miura, M.; Nomura,
M. Bull. Chem. Soc. Jpn. 1998, 71, 467-473. (b) Kondo, Y.; Komine,
T.; Sakamoto, T. Org. Lett. 2000, 2, 3111-3113. (c) Aoyagi, Y.; Inoue,
A.; Koizumi, I.; Hashimoto, R.; Tokunaga, K.; Gohma, K.; Komatsu,
J.; Sekine, K.; Miyafuji, A.; Kunoh, J.; Honma, R.; Akita, Y.; Ohta, A.
Heterocycles 1992, 33, 257-272. (d) Sezen, B.; Sames, D. J. Am. Chem.
Soc. 2003, 125, 5274-5275. (e) Sezen, B.; Sames, D. J. Am. Chem.
Soc. 2003, 125, 10580-10585.
(1) Sivilotti, L.; Nistri, A. Prog. Neurobiol. 1991, 36, 35-92.
(2) Bormann, J. Trends Pharmacol. Sci. 2000, 21, 16-19.
(3) (a) Chambers, M. S.; Atack, J. R.; Broughton, H. B.; Collinson,
N.; Cook, S.; Dawson, G. R.; Hobbs, S. C.; Marshall, G.; Maubach, K.
A.; Pillai, G. V.; Reeve, A. J.; MacLeod, A. M. J. Med. Chem. 2003, 46,
2227-2240. (b) Chebib, M.; Johnston, G. A. R. Clin. Exp. Pharmacol.
Physiol. 1999, 26, 937-940. (c) Armer, R. Curr. Med. Chem. 2000, 7,
199.
(7) (a) Ohta, A.; Akita, Y.; Ohkuwa, T.; Chiba, M.; Fukunaga, R.;
Miyafuji, A.; Nakata, T.; Tani, N.; Aoyagi, Y. Heterocycles 1990, 31,
1951-1958. (b) McClure, M. S.; Glover, B.; McSorley, E.; Millar, A.;
Osterhout, M. H.; Roschangar, F. Org. Lett. 2001, 3, 1677-1680. (c)
Glover, B.; Harvey, K. A.; Liu, B.; Sharp, M. J.; Tymoschenko, M. F.
Org. Lett. 2003, 5, 301-304. (d) Okazawa, T.; Satoh, T.; Miura, M.;
Nomura, M. J. Am. Chem. Soc. 2002, 124, 5286-5287. (e) Lane, B. S.;
Sames, D. Org. Lett. 2004, 6, 2897-2900.
(4) Carling, W. R.; Hallett, D. J.; Russell, M. G. N.; Street, L. J.
Preparation of 7-phenylimidazo[1,2-b][1,2,4]triazines as ligands for
GABA receptors. PCT Int. Appl. 2003, 42 pp; WO 2003008418.
(8) Li, W.; Nelson, D. P.; Jensen, M. S.; Hoerrner, R. S.; Javadi, G.
J.; Cai, D.; Larsen, R. D. Org. Lett. 2003, 5, 4835-4837.
10.1021/jo0507035 CCC: $30.25 © 2005 American Chemical Society
Published on Web 06/24/2005
5938
J. Org. Chem. 2005, 70, 5938-5945

Synthesis of an R2/3-Selective GABA Agonist
SCHEME 1
SCHEME 2
SCHEME 3
SCHEME 4
aminotriazine 15. Biaryl bromide 6 was prepared by
means of a Suzuki cross-coupling followed by a regiose-
lective bromination. The synthetic route is a conver-
gent seven-step sequence that relies on a regioselective
aminotriazine synthesis to prepare the imidazotriazine
core (22) and a robust, scaleable synthesis of biaryl
bromide 6.
Results and Discussion
Several synthetic strategies were examined for prepa-
ration of biaryl bromide 6. Our initial approach was to
convert 3-fluorobenzonitrile 2 into the nucleophilic com-
ponent of a transition metal catalyzed cross-coupling
reaction.9 Selective orthometalation of 3-fluorobenzoni-
trile gave only nitrile addition products with strong bases
such as n-butyllithium, LDA, or LHMDS, while weaker
bases such as Grignard or alkyl zinc halide reagents
failed to react. A directed metalation with lithium dialkyl-
2,2,6,6-tetramethylpiperidinozincate (TMP-ZnR2Li), pro-
vided aryl-zincate intermediate 3.10,11 All attempts to
cross-couple aryl-zincate 312 with 2,4-dibromofluoroben-
zene or iodobenzene were unsuccessful, generally giving
no reaction. Varying the alkyl groups of zincate 3 (R )
Me, Et, Bu, or t-Bu), the amine (TMP or LDA), or the Pd
ligand failed to promote the cross-coupling reaction.
Boronic acid 4 was then prepared by adding LDA to a
mixture of 3-fluorobenzonitrile and triisopropylborate,13
followed by acid hydrolysis. Similar to the zincate species
3, boronic acid 4 failed to provide the desired biaryl
product when reacted with 1,3-dibromo-4-fluorobenzene
or 2-bromo-4-chlorofluorobenzene under standard Suzuki
conditions with a variety of Pd catalysts and ligands. We
believe the complete lack of reactivity of zincate 3 and
boronic acid 4 toward cross-coupling is due to the
electron-withdrawing effects of the aryl substituents.Our
next approach was to prepare 2-bromo-3-fluorobenzoni-
trile 7 and then cross-couple with an appropriate orga-
nometallic nucleophile. We previously reported the syn-
thesis of 2-bromo-3-fluorobenzonitrile 7 via treatment of
boronic acid 4 with dibromodimethylhydantoin (DBDMH)
and found it was essential to conduct the bromodebor-
onation under slightly basic conditions (Scheme 3).14
Several cross-coupling partners were explored with bro-
mide 7, including organometallic reagents 8a, 8b, and
8c which were generated from 2,4-dibromofluorobenzene
via a selective metal-halogen exchange reaction with
iPrMgCl.15 Unfortunately, Pd-catalyzed reactions of 8a,
8b, and 8c with 7 gave a low yield due to polymerization
(Scheme 4). With use of a similar strategy, polymeriza-
tion was minimized by using organozinc 9, which was
derived from chlorofluorobenzene via selective ortholithi-
(9) Deiterich, F. Metal catalyzed cross-coupling reactions; Wiley-
VCH: New York, 1998.
(10) Kondo, Y.; Shilai, M.; Uchiyama, M.; Sakamoto, T. J. Am. Chem.
Soc. 1999, 121, 3539-3540.
(11) Confirmation of organozincate formation was established via
conversion to bromide 7 with the addition of 5 equiv of Br2 (75% yield).
(12) For examples of Pd-catalyzed cross-coupling with zincates,
see: Gauthier, D. R., Jr.; Szumigala, R. H., Jr.; Dormer, P. G.;
Armstrong, J. D., III; Volante, R. P.; Reider, P. J. Org. Lett. 2002, 4,
375-378.
(13) (a) Kristensen, J.; Lyse´n, M.; Vedsø, P.; Begtrup, M. Org. Lett.
2001, 3, 1435-1437. (b) Caron, S.; Hawkins, J. M. J. Org. Chem. 1998,
63, 2054-2055. (c) Li, W.; Nelson, D.; Jensen, M.; Hoerrner, R. S.; Cai,
D.; Larsen, R.; Reider, P. J. Org. Chem. 2002, 67, 5394-5397.
(14) Szumigala, R. H., Jr.; Devine, P. N.; Gauthier, D. R., Jr.;
Volante, R. P. J. Org. Chem. 2004, 69, 566-569.
(15) A sample of Grignard reagent 8a was quenched with water and
a GC assay revealed 20:1 selectivity of 4-bromofluorobenzene versus
2-bromofluorobenzene.
J. Org. Chem, Vol. 70, No. 15, 2005 5939

Gauthier et al.
SCHEME 5
SCHEME 6
We envisioned accessing the 5-substituted imidazo[1,2-
c][1,2,4]triazine 22 by performing a regioselective annu-
lation of aminotriazine 15 with chloroacetaldehyde. The
aminotriazine 15 precursor could be prepared by a
regioselective condensation of aminoguanidine and keto
aldehyde 17 or its equivalent. Accordingly, dibromoke-
tone 1420 was reacted with aminoguanidine yet the
condensation was nonselective, providing an equal mix-
ture of aminotriazines 15 and 16 (Scheme 7). Since the
terminal hydrazine nitrogen of aminoguanidine is the
most nucleophilic,21 we anticipated achieving high regio-
control using keto aldehyde 17.22 The aqueous base
hydrolysis of dibromide 14 gave an intractable mixture
of keto aldehyde 17, present as a complex hydrate/
aldehyde mixture along with several byproducts, which
was treated with aminoguanidine bicarbonate providing
the desired aminotriazine regioisomer in a 19:1 ratio.23
However, aminotriazine 15 was isolated in only 28%
overall yield from the dibromoketone 14 after a difficult
extractive workup. The low yield is attributed to byprod-
ucts 18a, 19a, and 19b24 which were formed during the
hydrolysis of dibromide 14. Reports on the preparation
of keto aldehyde 17 from 2-methyl-3-butyn-2-ol or 3-meth-
yl-2-butenone have appeared, yet these oxidations re-
quired stoichiometric use of indelible toxic promoters
such as SeO2 or Hg(OAc)2.25 Ultimately, the keto alde-
hyde formation was circumvented and aminotriazine 15
was prepared in 65% isolated yield using our recently
reported conditions for the regioselective cyclization of
aminoguanidine with bis-morpholine ketoaminals 20
(Scheme 8).26 This strategy takes advantage of an in situ
generated N,O-acetal species (21) that selectively reacts
with aminoguanidine acetate to provide a substantially
improved yield of the desired product. Additionally, the
isolation of aminiotriazine 15 was now straightforward,
simply requiring partial concentration and aging at 0 °C
for 1 h. The precipitated product was analytically pure
and isolated in an overall 65% yield from dibromide 14.
ation (Scheme 5).16 Negishi coupling of 7 with organozinc
9 gave a 73% yield of 5; however, biaryl chloride 5 proved
to be a poor penultimate intermediate for the synthesis
of compound 1 because the subsequent arylation reaction
proceeded in poor yield (vide infra). Thus, biaryl-Br 6 was
the desired intermediate and was ultimately prepared
via a two-step sequence entailing a Suzuki cross-coupling
followed by a highly regioselective aromatic bromination
(Scheme 6). Suzuki coupling of boronic acid 10 with
bromide 7 gave biaryl 11 in 85% yield. Biaryl 11 was then
selectively brominated with Fe (0.5 equiv) and Br2 (4-6
equiv) in DCE to give biaryl-Br 6 in 75% yield after
workup and crystallization from IPA/water.
While the Fe-catalyzed bromination of biaryl-H 11 with
Br2 afforded exclusively the desired bromide in good yield,
the reaction was only successful in halogenated solvents
and required incremental Br2 charges during the course
of the reaction. Since these reaction requirements were
undesirable upon scale-up, alternative bromination con-
ditions were necessary. Treatment of biaryl 11 with
N-bromosuccinamide17 (NBS, 1 equiv) or N,N′-dibro-
modimethylhydantoin (DBDMH, 0.5 equiv) in the pres-
ence of sulfuric or triflic acid18 (1 equiv) in solvents such
as cyclohexane, MeCN, and dichloroethane selectively
afforded the desired bromide; however, over brominated
products, mainly dibromobiaryl 12 began to form when
reactions were maximized for conversion. The combina-
tion of DBDMH and H2SO4 in MeCN was chosen because
the product was conveniently crystallized from the reac-
tion mixture by simply adding H2O.19 The amounts of
DBDMH and H2SO4 were then optimized based upon
reaction time, conversion, and minimization of over-
brominated products. After treatment of 11 with 1 equiv
of H2SO4 and 0.6 equiv of DBDMH for 24 h, <5%
dibromination occurred and was efficiently removed upon
crystallization, providing bromide 6 in 88% yield.
Aminotriazine 15 was then treated with chloroacetal-
dehyde in IPA/H2O to give imidazotriazine 22 (Scheme
9). The product was extracted from the crude reaction
with i-BuOH, solvent switched to MeCN via a constant
volume distillation, and then conveniently isolated as the
HCl salt monohydrate in 73% yield.
Initially, we envisioned the end-game strategy involv-
ing a selective bromination of imidazotriazine 22 followed
by a Pd cross-coupling reaction with a boronic acid (23)
generated from biaryl bromide 6 (Scheme 10). While this
(20) Prepared via bromination of the corresponding methyl ketone,
see: (a) Nan’ya, S.; Ishida, H.; Moiji, E. J.; Butsugan, Y.; Bajji, A. C.
J. Heterocycl. Chem. 1994, 31, 401. (b) Boeykens, M.; De Kimpe, N.
Tetrahedron 1994, 50, 12349.
(21) (a) Ueda, T.; Furukawa, M. Chem. Pharm. Bull. 1964, 12, 100-
103. (b) Osman, A. N.; Botros, S.; Kandeel, M. M.; Khalil, N. A.; Abd
El-Latif, H. A. Indian J. Chem. 1996, 35B, 446-455.
(22) Pfuller, O. C.; Sauer, J. Tetrahedron Lett. 1998, 39, 8821-8824.
(23) The regioselectivity was found to be highly pH dependent and
gave optimal selectivity at pH 8-9.
(16) Mongin, F.; Schlosser, M. Tetrahedron Lett. 1996, 37, 6551.
(17) For bromination with NBS in H2O/H2SO4, see: Lambert, F. L.;
Ellis, W. D.; Parry, R. J. J. Org. Chem. 1965, 30, 304-305.
(18) Only strong protic acids affected the desired bromination.
(19) With use of DBDMH, the hydantoin byproduct precipitated
during the course of the reaction and subsequently dissolved with the
addition of H2O, whereas the succinamide generated by using NBS
inhibited crystallization of the product.
(24) Cyclic byproduct 19a was found to be extremely stable to the
hydrolysis conditions. Thus, once formed it does not convert back to
keto aldehyde 17.
(25) (a) Ballistreri, F. P.; Failla, S.; Tomaselli, G. A. J. Org. Chem.
1988, 53, 830-831. (b) Colonge, J.; Cumet, L. Bull. Soc. Chim. Fr. 1947,
841-843.
(26) Limanto, J.; Desmond, R. A.; Gauthier, D. R., Jr.; Devine, P.
N.; Reamer, R. A.; Volante, R. P. Org. Lett. 2003, 5, 2271-2274.
5940 J. Org. Chem., Vol. 70, No. 15, 2005

Synthesis of an R2/3-Selective GABA Agonist
SCHEME 7
ter.28 Electrophilic aromatic substitution appears to be
the operative mechanism.29
SCHEME 8
The reaction parameters including Pd source, base,
solvent, and additives of the arylation reaction were
explored (Table 1). Arylation reactions with biary chloride
5 gave poor conversions even with catalysts that have
shown good reactivity with aryl chlorides in metal-
catalyzed couplings.30 For example, Pd(t-Bu)3P31 and Pd/
IMes32 were the only catalysts that produced any product
(Table 1, entries 1 and 2). Biaryl bromide 6 was found to
be reactive under a wide range of Pd/ligand systems.
Although several Pd(0) and Pd(II) sources were effective,
Pd(OAc)2 was robust and commercially less expensive
compared to other Pd sources. Many phosphine ligands
were evaluated, but none were superior to triphenylphos-
phine in terms of cost and conversion. The Pd/PPh3 ratio
had a significant effect on the reaction and more favor-
able results were obtained with equal molar amounts of
ligand and Pd(OAc)2 (Table 1, entry 4 vs 5).31 In accord
with Miura’s initial report with imidazoles, organic amine
bases were completely ineffective, while either Cs2CO3
or K2CO3 afforded much better conversion than the less
soluble Li and Na carbonates. High boiling polar aprotic
solvents such as DMF, 1,4-dioxane, NMP, and DMAc
gave the best conversions probably due the ability to
partially dissolve the inorganic base at high temperature.
Accordingly, reactions conducted in less polar, high
boiling solvents, such as toluene, provided poor conver-
sion (<20%). When imidazotriazine HCl hydrate 22 was
extensively dried in a vacuum oven to afford the anhy-
drous HCl salt, the reaction rate and conversion suffered
dramatically. The arylation was optimal with only the
water provided by the hydrated intermediate 22, al-
though having up to 5% water by volume in the reaction
medium did not significantly impede the reaction. Ini-
tially, our best conditions used 5 mol % of Pd(OAc)2, 5
mol % of PPh3, and 2 equiv of K2CO3 in DMF at 100 °C,
where reactions were typically complete in 4 h (entry 5).
SCHEME 9
SCHEME 10
route was plausible,27 a more efficient strategy based
upon recent reports of Pd-catalyzed arylation reactions
with electron-rich heterocycles, such as oxazoles, imida-
zoles, thiazoles, furans, and imidazopyrazines, led us to
explore the direct coupling of imidazotriazine 22 with
biaryl bromide 6. Similar to the imidazopyrazines,8
arylation is selective and parallels aromatic bromination,
reacting at the sight of the highest nucleophilic charac-
(28) Imidazotriazine 22 was selectively brominated at the 7-position
to give 24 in 75% yield by treatment with NaOAc and Br2 in HOAc;
see ref 5.
(29) Park, C.-H.; Ryabova, V.; Seregin, I. V.; Sromek, A. W.;
Gevorgyan, V. Org. Lett. 2004, 6, 1159-1162.
(30) For an excellent review, see: Littke, A. F.; Fu, G. C. Angew.
Chem., Int. Ed. 2002, 41, 4176-4211.
(31) Littke, A. F.; Dai, C.; Fu, G. C. J. Am. Chem. Soc. 2000, 122,
4020-4028.
(32) (a) Arduengo, A. J., III Acc. Chem. Res. 1999, 32, 913-921. (b)
Zhang, C.; Huang, J.; Trudell, M. L.; Nolan, S. P. J. Org. Chem. 1999,
64, 3804-3805.
(27) Bettati, M.; Chambers, M. S.; Goodacre, S. C.; Hallett, D. J.;
Russell, M. G. N.; Street, L. J. Preparation of imidazotriazines as
ligands for GABA receptors. PCT Int. Appl. 2003, 67 pp; WO
2003093272.
J. Org. Chem, Vol. 70, No. 15, 2005 5941

Gauthier et al.
TABLE 1. Arylation of Imidazotriazine 22
reaction conditions
entry
X
solvent
base (equiv)
catalyst/ligand (ratio)
temp, °C
time, h
yield, %
1
2
3
4
5
6
7
8
9
Cl
Cl
Br
Br
Br
Br
Br
Br
Br
dioxane
dioxane
DMF
Cy2NMe
5% Pd2dba3/t-Bu3P (1:1)
5% Pd/IMes
100
100
100
100
100
100
100
130
130
48
48
4
14
20
70
70
82
45
88
85
86
Cs2CO3
Cs2CO3 (2)
K2CO3 (2)
K2CO3 (2)
K2CO3 (2)
KOAc (4)
KOAc (4)
KOAc (2.5)
5% Pd(OAc)2/PPh3 (1:2)
5% Pd(OAc)2/PPh3 (1:2)
5% Pd(OAc)2/PPh3 (1:1)
2% Pd(OAc)2/PPh3 (1:1)
5% Pd(OAc)2/PPh3 (1:1)
1% Pd(OAc)2/PPh3 (1:1)
1% Pd(OAc)2/PPh3 (1:1)
DMF
4
DMF
DMF
DMAc
DMAc
DMAc
4
24
24
36
4
SCHEME 11
Although these conditions were robust, a lower catalyst
loading could not be tolerated (entry 6, reaction stalled
at 50% conversion, 45% yield). Switching the base to
KOAc (4 equiv) gave a slower reaction with 5 mol % of
catalyst (entry 7), but catalyst loading could be reduced
to 2 mol % when run at 130 °C for 36 h in DMAc (entry
8). Temperatures over 100 °C in DMF could not be
tolerated due to debromination of 6 generating 11 as the
major product. Significant rate effects were observed by
varying the amount of KOAc, the only base that was
completely soluble under the reaction conditions. At least
2 mol of base are necessary to arylate the imidazotriazine
HCl salt (22), but excess KOAc apparently retards the
rate while still maintaining an active catalyst. For
example, the optimal arylation conditions were found by
reducing the amount of KOAc from 4 to 2.5 equiv,
yielding an 86% yield in only 4 h vs 36 h (Table 1,
compare entries 8 and 9).
that compound 1 has a particularly high affinity toward
Pd since standard workup and crystallization procedures
gave nominal rejection (e.g., 2500 ppm of Pd with 1 mol
% of Pd(OAc)2). The use of immobilized palladium
catalysts such as Pd/C35 resulted in extensive leaching
since the isolated product contained >1000 ppm of Pd.
Chemical additives such as n-Bu3P36 and trimercaptot-
riazine37 or even high-throughput screening processes38
with adsorbents such as activated carbon, diatomaceous
earth, and immobilized phosphine ligands on polymer
supports did not adequately reduce the Pd levels. For-
tuitously, crude 1 containing up to 8000 ppm of Pd was
recrystallized in EtOH and the resulting 1‚EtOH solvate
contained <20 ppm of Pd. Apparently the affinity for Pd
greatly diminishes when EtOH enters the crystalline
The arylation reaction produced only two significant
side products, biaryl 11 and imidazotriazine dimer 25,
both formed in <3%. We presumed that the dimerization
occurs via bis arylation of a Pd(II) species followed by
reductive elimination to give 25. When imidazotriazine
22 was heated with catalyst in the absence of the aryl
bromide 6, only a trace of dimer 25 formed after 24 h at
130 °C. When the same reaction was exposed to atmo-
spheric oxygen, 22 was quantitatively converted to dimer
25 after 12 h (Scheme 11).
The high convergence achieved by using the arylation
reaction in the final synthetic step came with the burden
of removing the palladium, a necessity required for the
preparation of an active pharmaceutical ingredient (API).33
Although Pd-catalyzed processes are ubiquitous in the
literature, only few reports have addressed this problem
of removing the residual metal because palladium-
catalyzed reactions are typically avoided in the final step
toward an API.34 Indeed, we were disappointed to find
(35) (a) LeBlond, C. R.; Andrews, A. T.; Sun, Y.; Sowa, J. R., Jr.
Org. Lett. 2001, 3, 1555-1557. (b) Sakurai, H.; Tsukuda, T.; Hirao, T.
J. Org. Chem. 2002, 67, 2721-2722.
(36) Larsen, R. D.; King, A. O.; Chen, C. Y.; Corley, E. G.; Foster,
B. S.; Roberts, F. E.; Yang, C.; Lieberman, D. R.; Reamer, R. A.;
Tschaen, D. M.; Verhoeven, T. R.; Reider, P. J.; Lo, Y. S.; Rossano, L.
T.; Brookes, A. S.; Meloni, D.; Moore, J. R.; Arnett, J. F. J. Org. Chem.
1994, 59, 6391-6394.
(37) Rosso, V. W.; Lust, D. A.; Bernot, P. J.; Grosso, J. A.; Modi, S.
P.; Rusowicz, A.; Sedergran, T. C.; Simpson, J. H.; Srivastava, S. K.;
Humora, M. J.; Anderson, N. G. Org. Proc. Res. Dev. 1997, 1, 311-
314.
(38) Welch, C. J.; Shaimi, M.; Biba, M.; Chilenski, J. R.; Szumigala,
R. H., Jr.; Dolling, U.; Mathre, D. J.; Reider, P. J. J. Sep. Sci. 2002,
25, 847-850.
(33) Only a few parts per million of a heavy metal such as palladium
can be tolerated in an API according to the guidelines of the
International Conference on Harmonization (ICH).
(34) Chen, C.; Dagneau, P.; Grabowski, E. J. J.; Oballa, R.; O’Shea,
P.; Prasit, P.; Robichaud, J.; Tillyer, R.; Wang, X. J. Org. Chem. 2003,
68, 2633-2638.
5942 J. Org. Chem., Vol. 70, No. 15, 2005

Synthesis of an R2/3-Selective GABA Agonist
ing slurry of DBDMH was added over 1-2 h with the
temperature maintained between 15 and 25 °C. The reaction
was aged for 1 h after all DBDMH was added. The reaction
was cooled to 5-10 °C, quenched with 20% Na2SO3 (18 L), and
diluted with MTBE (18 L). Upon agitating and settling, the
aqueous layer was cut and the organic layer was washed with
1 N NaOH (36 L). The MTBE solution of 7 was assayed for
yield by HPLC (3.80 kg, 88% yield).
lattice. A final recrystallization from THF/heptane pro-
vided the unsolvated GABAA agonist 1 in high purity.
Conclusion
We have developed a convergent, practical, and ef-
ficient synthesis of a GABA agonist using the crystalline
intermediates biaryl bromide 6 and imidazotriazine 22.
Biaryl bromide 6 was prepared via a highly selective
aromatic bromination with DBDMH and H2SO4. The
regioselective condensation of aminotriazine 15 with
chloroacetaldehyde provided the desired imidazotriazine
intermediate 22 as the HCl hydrate salt. In the final step,
imidazotriazine 22 was selectively arylated under pal-
ladium-catalyzed conditions, and a simple recrystalliza-
tion of 1 as the EtOH solvate effectively removed the
homogeneous palladium allowing the API to be accept-
able for human clinical trials. The seven-step, chroma-
tography-free synthesis was demonstrated on a multi-
kilogram scale.
5′-Chloro-2′,6-difluorobiphenyl-2-carbonitrile (5). To a
-40 °C solution of 4-chlorofluorobenzene (10.0 mmol) in dry
THF (13.0 mL) was added hexyllithium (2.5 M in hexane; 11.0
mmol) with the temperature maintained at <-20 °C. The
reaction was aged for 1 h at -20 to -25 °C. Anhydrous zinc
chloride (∼1.5 M in THF; 11.0 mmol) was slowly added to the
aryllithium with the temperature maintained at <0 °C. Upon
aging for 30 min at 0 °C, the mixture was degassed (five
vacuum/N2 purges) and Pd2(dba)3 (0.125 mmol), t-Bu3P (0.25
mmol), and 7 (10.0 mmol) were added sequentially to the
reaction mixture. The reaction warmed to 23 °C and was aged
for 1 h. After aqueous workup (10% brine; 2 × 10 mL), 5 was
isolated as a white solid in 73% yield after silica gel chroma-
tography. Mp 107-110 °C; IR (KBr pellet) 3080, 2923, 2854,
2237, 1496, 1456, 1260, 1217, 962, 819, 799 cm-1 1H NMR
;
(CDCl3, 400 MHz) δ 7.61 (d, J ) 7.7 Hz, 1H), 7.53 (dt, J ) 8.0,
5.2 Hz, 1H), 7.48-7.39 (m, 3H), 7.19 (t, J ) 8.9 Hz, 1H); 13C
NMR (CDCl3, 100 MHz) δ 160.2 (JCF ) 136.3 Hz), 157.7 (JCF
) 134.5 Hz), 131.4 (JCF ) 8.6 Hz), 131.3, 130.9 (JCF ) 8.9 Hz),
129.4 (JCF ) 3.3 Hz), 129.1 (JCF ) 3.9 Hz), 125.7 (JCF ) 19.5
Hz), 120.7 (JCF ) 17.7 Hz), 120.6 (JCF ) 22.5 Hz), 117.4 (JCF
) 23.8 Hz), 116.3 (JCF ) 4.0 Hz), 114.9 (JCF ) 4.2 Hz); 19F
NMR (CDCl3, 376.4 MHz) -110.6 (JFF ) 10.3 Hz), -116.3 (JFF
) 8.9 Hz). C13H6ClF2N required: C, 62.54; H, 2.42; N, 5.61.
Found: C, 62.82; H, 2.29; N, 5.38.
Experimental Section
General. Solvents and common reagents were purchased
from a commercial source and used without purification. All
reactions were conducted under N2 atmosphere with standard
air-free manipulation techniques. 1H and 13C NMR spectra
were recorded on a 400 MHz spectrometer and chemical shifts
are reported in ppm downfield from tetramethylsilane (TMS).
High Performance Liquid Chromatography (HPLC) analysis
was performed with a YMC-Pack Pro C18 (250 × 4.6 mm i.d.)
column under the following conditions: 80% MeCN:20% 0.1%
vol H3PO4 in H2O mobile phase, 1 mL/min flow rate, 210 nm
wavelength detection, 35 °C column temperature. Melting
points are uncorrected. All new compounds are fully character-
ized below.
(2-Cyano-6-fluorophenyl)boronic Acid (4). To a 100-L
round-bottom flask equipped with an overhead stirrer, ther-
mocouple, and N2 inlet was added 2 (4.00 kg, 33.03 mol),
B(OiPr)3 (6.50 kg, 34.7 mol), and THF (20 L). The solution was
cooled to -20 °C and a 1.8 M LDA solution in heptane/THF/
ethylbenzene (18.7 L, 33.7 mol) was added to the solution over
∼1 h with the internal temperature maintained at <5 °C. The
reaction was aged for 15 min at 0 °C and assayed for
completion by HPLC. The reaction was quenched into a 0-5
°C mixture of 5 N HCl (26 L) and MTBE (20 L) over 30 min.
The resulting mixture was agitated for 30 min at 23 °C. Upon
settling, the aqueous layer was cut and back-extracted with
MTBE (20 L). The organic layers were combined and washed
with a 20 wt % brine/2.5 wt % KHCO3 wash (17 L). The organic
solution was concentrated under vacuum to approximately 10
L and a constant volume solvent switch was performed: MeCN
(∼20 L) was continuously added to the MTBE/THF solution
of boronic acid with the volume maintained at approximately
10 L (temperature maintained <30 °C, 20 to 28 in. Hg). After
solvent switch to MeCN, the water concentration was 2.5-
5.0% as determined by Karl Fisher titration. The solution of
4 (4.08 kg, 75% yield) in MeCN was taken directly into the
next reaction without further manipulation.
2′,6-Difluoro-1,1′-biphenyl-2-carbonitrile (11). 2-Fluo-
rophenylboronic acid 10 (3.0 kg, 21.4 mol) and K3PO4 (4.0 kg,
18.7 mol) were added to a solution of THF (25.5 L) and water
(8.5 L). The resulting mixture was thoroughly degassed via
vacuum/N2 purges (5 times). A solution of 10% tri-tert-
butylphosphine in hexanes (1.3 L, 0.42 mol) was added to the
reaction, followed by allylpalladium chloride dimer (77.7 g, 0.21
mol). A solution of bromobenzonitrile 7 (3.4 kg, 17.0 mol) in
MeCN (18 L) was slowly added to the reaction over 1 h to
control the exotherm. The reaction was aged for 2 h at 45-50
°C, cooled to 5 °C, and quenched with 1 N NaOH (17.0 L). The
biphasic mixture was vigorously agitated for 15 min and
transferred to an extraction vessel. The aqueous layer was cut
upon settling. The organic layer was sequentially washed with
1 N NaOH (17.0 L) and then saturated brine (17.0 L). The
crude organic solution of 11 was assayed at 90% yield by
HPLC, using an analytically pure standard as reference. The
solution of 11 in THF was converted over to a solution in IPA
via a constant volume distillation (17.0 L, 5 mL/g vs assay yield
of 11). The resulting slurry was heated to 50 °C for 30 min to
dissolve all solids. The solution was allowed to slowly cool to
23 °C. To the resulting slurry was added water (17.0 L) over
1 h and the mixture was cooled to 0-5 °C. The slurry was
filtered and the filter-cake was washed with a 5 °C 1:1 IPA/
water solution (10.2 L). The solids were dried under a N2
stream for 12 h and then in a vacuum oven at 40 °C to remove
any remaining water. Biaryl 11 was isolated as a white solid
(3.53 kg, 88.0 wt %, 85% yield) with 96.3% purity by HPLC.
An analytically pure sample of 11 was obtained by recrystal-
lization from 1:1 IPA/H2O then dried in a vacuum oven. Mp
75-75.5 °C; IR (thin film) 3085, 2231, 1615, 1603, 1583, 1569,
2-Bromo-3-fluorobenzonitrile (7). To a 100-L round-
bottom flask equipped with an overhead stirrer, thermocouple,
and N2 inlet was added a solution of 4 (3.56 kg, 21.6 mol) in
MeCN. The solution was cooled to 15 °C and 25% NaOMe in
MeOH (247 mL, 1.10 mol) was added. In a separate vessel, a
slurry of DBDMH (1,3-dibromo-5,5-dimethylhydantoin, 6.79
kg, 23.7 mol) in MeCN (18 L) was prepared. Approximately
10-15% of the DBDMH slurry was transferred to the boronic
acid solution with the temperature maintained between 15 and
25 °C. The reaction was aged for 15 min during which an
exotherm occurred indicating reaction initiation. The remain-
1
1080, 1039, 1008, 983, 919 cm-1; H NMR (CDCl3, 400 MHz)
δ 7.60 (d, J ) 7.6 Hz, 1H), 7.53-7.41 (m, 4H), 7.32-7.22 (m,
2H); 13C NMR (CDCl3, 100 MHz) δ 160.9 (JCF ) 7.3 Hz), 158.4
(JCF ) 5.9 Hz), 131.5 (JCF ) 8.2 Hz), 130.3 (JCF ) 8.8 Hz), 129.0
(JCF ) 3.7 Hz), 127.2 (JCF ) 20.1 Hz), 124.3 (JCF ) 3.5 Hz),
120.4 (JCF ) 22.8 Hz), 119.2 (JCF ) 15.7 Hz), 116.7 (JCF ) 4.0
Hz), 115.9 (JCF ) 21.7 Hz), 114.9 (JCF ) 4.6 Hz); 19F NMR
(CDCl3, 376.4 MHz) -110.8 (JFF ) 10.8 Hz), -113.8 (JFF
)
J. Org. Chem, Vol. 70, No. 15, 2005 5943

Gauthier et al.
10.5 Hz). C13H7F2N required: C, 72.56; H, 3.28; N, 6.51.
Found: C, 72.40; H, 3.25; N, 6.41.
begins to precipitate. The resulting suspension was then cooled
to 3-5 °C and aged for 6-12 h. The suspension was filtered
and the wet cake was washed with cold H2O (10-14 L). The
yellowish solid was then dried overnight in vacuo at 25-35
°C under a stream of N2 to afford 3.74 kg (65% yield) of 15 in
g99.5% regioisomeric ratio.
5′-Bromo-2′,6-difluorobiphenyl-2-carbonitrile (6). To a
round-bottom flask equipped with a condenser under N2 was
added solid biaryl 11 (3.31 kg, 15.39 mol) and acetonitrile (24.8
L). The solution of 11 in MeCN should have <5000 ppm water
by Karl Fisher titration to achieve an efficient transformation.
To this solution was added solid DBDMH (4.4 kg, 15.39 mol)
in one portion at room temperature to give a yellow-orange
suspension. The addition of DBDMH is endothermic, which
brings the batch temperature to about 10-13 °C. Concentrated
sulfuric acid (1.28 L, 23.08 mol) was then added slowly over
15-20 min to give a turbid solution, which was then heated
to 45 °C and aged for 12 h. The reaction was monitored by
HPLC and a 98% conversion was typically obtained after 12
h. At the end of the reaction, water (24.8 L) was slowly added
over 1.5 h to the cloudy reaction mixture while maintaining
the internal temperature at 45-50 °C. Addition of the first
5% of H2O results in a 10-15 °C exotherm. During addition
of the first half of H2O, the reaction mixture turns from orange
cloudy to a clear red solution, after which the desired product
begins to precipitate. Vigorous stirring is usually required to
ensure homogeneous crystallization. At the end of the H2O
addition, the resulting suspension was allowed to cool to room
temperature (20-23 °C), aged for 3 h, and then filtered. The
filter cake was then washed with 1:1 MeCN:H2O (24.8 L) and
dried in vacuo under a stream of N2. Biaryl bromide 6 was
isolated as a white crystalline solid in 88% yield (3.98 kg). Mp
130-131 °C; IR (thin film) 3048, 2238, 1603, 1563, 1290, 1162,
1121, 1020, 980, 838, 760 cm-1; 1H NMR (CDCl3, 400 MHz) δ
7.62-7.50 (m, 4H), 7.44 (dt, J ) 9.2, 1.2 Hz, 1H), 7.14 (t, J )
8.9 Hz, 1H); 13C NMR (CDCl3, 100 MHz) δ 160.4 (JCF ) 80.4
Hz), 157.9 (JCF ) 79.1 Hz), 134.4 (JCF ) 8.2 Hz), 134.1, 130.9
(JCF ) 8.8 Hz), 129.1 (JCF ) 3.9 Hz), 125.6 (JCF ) 19.7 Hz),
121.2 (JCF ) 17.2 Hz), 120.6 (JCF ) 22.5 Hz), 117.8 (JCF ) 23.4
2-Imidazo[1,2-b][1,2,4]triazin-3-ylpropan-2-ol Hydro-
chloride Hydrate (22). To a solution of aminotriazine 15
(2.00 kg, 12.97 mol) in isobutanol (16 L) was added 50 wt %
aqueous chloroacetaldehyde (5 L, 39.00 mol). The reaction was
warmed to 85 °C and the resulting dark mixture was aged at
85 °C for 22 h. The reaction was cooled to room temperature
and water (16 L) was added. The layers were mixed and
separated. The acidic water solution containing the imidazo-
triazine was made basic (pH 8-9) with 50% NaOH and
extracted with EtOAc (3 × 16 L). The EtOAc solution of the
imidazotriazine was distilled and solvent switched to aceto-
nitrile. To the acetonitrile solution of 22 was added 1.1 equiv
of conc HCl over 1 h and the resulting slurry was aged for 4 h
at 0-3 °C. The solid was isolated by filtration and the cake
was washed with cold acetonitrile. The solid was air-dried to
give 1.76 of kg imidazotriazine-HCl salt (60% yield). The solid
typically contained ∼5% water. A sample was further dried
in a vacuum oven to give anhydrous 22 as an off-white solid.
Mp 225-230 °C dec; IR (thin film) 3245, 3124, 3068, 2982,
2849, 2779, 1730, 1561, 1523, 1503, 1456, 1400, 1139, 1275,
1003, 800 cm-1; 1H NMR (d6-DMSO, 400 MHz) δ 9.34 (s, 1H),
8.63 (s, 1H), 8.35 (s, 1H), 7.90-7.00 (br s, 2H), 1.55 (s, 6H);
13C NMR (d6-DMSO, 100 MHz) δ 167.4, 141.1, 139.0, 126.2,
116.5, 72.5, 29.5. HRMS for C8H11N4O (179.0933) found
179.0975. C8H11ClNO required: C, 44.76; H, 5.17; N, 26.10.
Found: C, 44.48; H, 4.98; N, 25.78.
2′,6-Difluoro-5′-[3-(1-hydroxy-1-methylethyl)imidazo-
[1,2-b][1,2,4]triazin-7-yl]biphenyl-2-carbonitrile (1). To a
72-L round-bottom flask under N2 is added imidazotriazine
22 (2.21 kg, 10.29 mol), biaryl bromide 6 (2.75 kg, 9.35 mol),
KOAc (2.29 kg, 23.4 mol), Pd(OAc)2 (31.5 g, 0.140 mol), PPh3
(36.8 g, 0.140 mol), and DMAc (22.1 L). The slurry was warmed
to 130 °C and the near homogeneous dark solution was aged
at 130 °C until >90% conversion. After 4 h, the solution was
cooled to 80 °C and H2O (4.42 L) and Darco G-60 (1.1 kg) were
added. The resulting slurry was aged at 60 °C for 1 h then
filtered through solka floc at 60 °C. The filter-cake was washed
with 5:1 DMAc/H2O (13.3 L). The combined filtrate and wash
was warmed to 100 °C and water (17.7 L) was added while
maintaining the temperature at >90 °C. The resulting solution
is seeded at 90-95 °C with a sample of crystalline 1 (0.5%,
168.0 g). To the resulting thin slurry was added H2O (14.6 L)
over 1 h at 90-95 °C. The resulting slurry was aged 1 h at
90-95 °C, then cooled to 20-25 °C over 2 h before filtering.
The batch was filtered and the filter-cake was washed with
1:1 DMAc/H2O (13.8 L) and then H2O (13.8 L) and dried in a
vacuum oven set at 80 °C, 130 Torr. 1 was isolated (3.15 kg,
86% yield) as a mustard yellow solid contaminated with 4500
ppm Pd.
Hz), 116.6 (JCF ) 3.8 Hz), 116.3 (JCF ) 4.0 Hz), 114.9 (JCF
)
4.2 Hz); 19F NMR (CDCl3, 376.4 MHz) -110.6 (JFF ) 10.2 Hz),
-115.7 (JFF ) 10.3 Hz). C13H6BrF2N required: C, 53.09; H,
2.06; N, 4.76. Found: C, 53.01; H, 1.77; N, 4.53.
2-(3-Amino-1,2,4-triazin-5-yl)propan-2-ol (15). Forma-
tion of aminal and N,O-acetal intermediates: To a
solution of 1,1-dibromo-3-hydroxy-3-methyl-2-butanone (9.71
kg, 37.35 mol) in THF (48 L) under N2 at 45 °C was slowly
added neat morpholine (13.4 L, 53.15 mol) over a period of 4
h. The resulting suspension was then heated to 66-68 °C and
aged for 18 h. The reaction was monitored by 1H NMR
spectroscopy and judged complete when g97% conversion was
obtained. The yellow suspension was then cooled to room
temperature and filtered. The wet filter-cake obtained (i.e.,
morpholine hydrobromide salt) was washed with THF (38 L)
and the combined filtrate was then concentrated under a
partial vacuum (5-10 Torr, pot temperature )15-20 °C) and
solvent switched to MeOH with a final volume of 25 L.
Preparation of aminoguanidine acetate solution: To
a suspension of aminoguanidine bicarbonate (5.08 kg, 37.35
mol) in MeOH (25 L) under N2 at 25 °C was slowly added neat
AcOH (6.4 L, 112.06 mol) over a period of 1 h. The resulting
white suspension was then aged at room temperature for 4-12
h, after which the CO2 evolution would cease and a thin
suspension was typically obtained.
Cyclization-aminotriazine formation: To the MeOH
solution of aminal/N,O-acetal intermediates obtained above
was slowly added the pre-made solution of aminoguanidine
acetate at room temperature over a period of 1 h. The resulting
solution was then heated to 67 °C and aged for 14 h. At the
end of the reaction, the dark brown solution was cooled to room
temperature and concentrated to about half its volume (25-
28 L). An equal volume of H2O (28 L) was added, followed by
heptane (15 L). The resulting biphasic layers were stirred
vigorously for 0.5 h and the aqueous layer was separated. The
MeOH/H2O solution was then concentrated to half its volume
(28 L), after which the desired aminotriazine regioisomer
2′,6-Difluoro-5′-[3-(1-hydroxy-1-methylethyl)imidazo-
[1,2-b][1,2,4]triazin-7-yl]biphenyl-2-carbonitrile-Etha-
nol (1:1) (1‚EtOH solvate). Crude 1 (2.50 kg, 6.38 mol) was
slurried in absolute EtOH (82.5 L) and warmed to reflux. After
all solids were dissolved, the solution was slowly cool to 20-
25 °C. The EtOH solvate precipitates during the cool-down
period. The slurry was concentrated to 25 L under vacuum
with a temperature <25 °C. The slurry was aged for 1 h at 23
°C and then filtered. The filter-cake was washed with EtOH
(14 L) and then dried in a vacuum oven set at 60 °C with a
nitrogen purge (23 “Hg). 1‚EtOH solvate was isolated as a
bright yellow solid in 93% yield (2.60 kg) and contained <20
ppm Pd. Mp 120-122 °C; IR (thin film) 3201, 2974, 2236, 1592,
1571, 1545, 1513, 1497, 1411, 1313, 987, 917, 768 cm-1 1H
;
NMR (CDCl3, 400 MHz) δ 8.85 (s, 1H), 8.19 (s, 1H), 8.14-
8.09 (m, 2H), 7.64 (dd, J ) 8.3, 0.7 Hz, 1H), 7.54 (dt, J ) 8.1,
5.0 Hz, 1H), 7.46 (dt, J ) 8.6, 1.1 Hz, 1H), 7.36 (t, J ) 8.8 Hz,
5944 J. Org. Chem., Vol. 70, No. 15, 2005

Synthesis of an R2/3-Selective GABA Agonist
1H), 4.04 (s, 1H), 3.71 (dt, J ) 11.7, 7.0 Hz, 2H), 1.69 (s, 6H),
1.23 (t, J ) 6.9 Hz, 3H); 13C NMR (CDCl3, 100 MHz) δ 160.8
(JCF ) 39.5 Hz), 159.3, 158.3 (JCF ) 40.7 Hz), 141.6, 136.2,
133.7, 130.7 (JCF ) 7.2 Hz), 129.6, 129.5, 129.2 (JCF ) 3.7 Hz),
126.5 (JCF ) 19.7 Hz), 125.1, 124.3 (JCF ) 3.9 Hz), 120.6 (JCF
1H), 7.82-7.72 (m, 2H), 7.62 (t, J ) 9.3 Hz, 1H), 5.78 (s, 1H),
1.56 (s, 6H); 13C NMR (d6-DMSO, 100 MHz) δ 160.7, 160.3
(JCF ) 61.0 Hz), 157.9 (JCF ) 62.1 Hz), 142.0, 137.5, 134.4,
132.3 (JCF ) 9.0 Hz), 130.2 (JCF ) 3.4 Hz), 129.9 (JCF ) 8.5
Hz), 129.7, 126.0 (JCF ) 20.0 Hz), 125.2 (JCF ) 3.4 Hz), 124.2,
121.7 (JCF ) 22.3 Hz), 119.9 (JCF ) 16.6 Hz), 117.0 (JCF ) 11.0
) 22.3 Hz), 119.9 (JCF ) 16.6 Hz), 116.8, 116.5, 114.9 (JCF
)
4.3 Hz), 72.6, 58.2, 29.6, 18.3; 19F NMR (CDCl3, 376.4 MHz)
-111.4 (JFF ) 12.1 Hz), -112.9 (JFF ) 10.9 Hz). HRMS for
C21H16F2N5O (392.1323) found 392.1329. C23H21F2N5O2 re-
quired: C, 63.15; H, 4.84; N, 16.01. Found: C, 62.82; H, 4.52;
N, 16.02.
Hz), 116.9 (JCF ) 7.3 Hz), 114.4 (JCF ) 4.2 Hz), 72.3, 29.7; 19
F
NMR (d6-DMSO, 376.4 MHz) -111.9 (JFF ) 8.7 Hz), -115.0
(JFF ) 9.2 Hz). HRMS for C21H16F2N5O (392.1323) found
392.1329. C21H15F2N5O required: C, 64.45; H, 3.86; N, 17.89.
Found: C, 64.20; H, 3.66; N, 17.82.
2′,6-Difluoro-5′-[3-(1-hydroxy-1-methylethyl)imidazo-
[1,2-b][1,2,4]triazin-7-yl]biphenyl-2-carbonitrile (1). 1‚
EtOH solvate (2.00 kg, 4.57 mol) was slurried in THF (60 L)
and warmed to 30 °C. The cloudy solution was transferred to
a separate vessel via a 1 µm line filter to remove the
precipitated imidazotriazine dimer 25. The resulting solution
was concentrated under vacuum at <25 °C to 20 L. A slurry
forms during the concentration, and the resulting slurry was
aged at 20-22 °C for 0.5-1 h. Heptane (20 L) was added over
1-2 h and then the slurry was aged for 1 h at 20-22 °C. The
slurry was filtered, washed with 1:1 THF:heptane (4 L), and
then dried under a stream of nitrogen for 15 h. Pure 1 was
isolated as a bright yellow solid (1.9 kg, 95% yield). Mp 223-
224 °C; IR (thin film) 3302, 2238, 1577, 1509, 1392, 1273, 1001,
2,2′-(7,7′-Biimidazo[1,2-b][1,2,4]triazine-3,3′-diyl)dipro-
pan-2-ol (25). Yellow solid isolated from filtration of 1 in THF.
Mp 360-362 °C; IR (thin film) 3218, 3129, 2980, 2020, 1528,
1315, 1188, 1147, 1097, 960, 929, 849, 782 cm-1; 1H NMR (d6-
DMSO, 400 MHz) δ 9.12 (s, 2H), 8.74 (s, 2H), 5.82 (br s, 2H),
1.58 (s, 12H); 13C NMR (d6-DMSO, 100 MHz) δ 161.4, 141.6,
138.1, 133.5, 115.8, 72.4, 29.7. HRMS for C16H19N8O2 (355.1631)
found 355.1637. C16H18N8O2 required: C, 54.23; H, 5.12; N,
31.62. Found: C, 54.64; H, 4.81; N, 31.58.
Acknowledgment. The authors thank Mr. Robert
Reamer for assistance with NMR, Ms. Mirlinda Biba
for HRMS analysis, and Dr. Thomas Loughlin for HPLC
development.
1
990, 747, 717 cm-1; H NMR (d6-DMSO, 400 MHz) δ 9.03 (s,
1H), 8.49 (s, 1H), 8.41-8.34 (m, 2H), 7.91 (dd, J ) 7.4, 1.2 Hz,
JO0507035
J. Org. Chem, Vol. 70, No. 15, 2005 5945