Synthesis of 3-Aminopyrazinone Mediated by 2-Pyridylthioimidate-ZnCl 2 Complexes. Development of an Efficient Route to a Thrombin Inhibitor

Article abstract of DOI:10.1021/jo034835e

A six-step preparation of thrombin inhibitor drug candidate 1 from pyrazinone 7 in 47% overall yield is described. The problem of low reactivity between weak amine nucleophile 4 and poor electrophile 3-bromopyrazinone 17 was overcome with the use of pyrid

Full text of DOI:10.1021/jo034835e

Syn th esis of 3-Am in op yr a zin on e Med ia ted by
2-P yr id ylth ioim id a te-Zn Cl2 Com p lexes. Develop m en t of a n
Efficien t Rou te to a Th r om bin In h ibitor
J ohn Y. L. Chung,* Raymond J . Cvetovich,* Fuh-Rong Tsay, Peter G. Dormer,
Lisa DiMichele, David J . Mathre, J ennifer R. Chilenski, Bing Mao,† and Robert Wenslow†
Departments of Process Research and Department of Analytical Research, Merck Research Laboratories,
P.O. Box 2000, Rahway, New J ersey 07065
john_chung@merck.com
Received J une 16, 2003
A six-step preparation of thrombin inhibitor drug candidate 1 from pyrazinone 7 in 47% overall
yield is described. The problem of low reactivity between weak amine nucleophile 4 and poor
electrophile 3-bromopyrazinone 17 was overcome with the use of pyridinylthioimidate 27 in the
presence of ZnCl2 to afford adduct 3 in high yield. Several zinc complexes were characterized by
solution and solid-state NMR and X-ray crystallographic analyses, and provided insight into the
reaction mechanism. Preparation of pyridine N-oxide amine 4 was accomplished via a selective
oxidation of the corresponding pyridinylamine 6. Pyridinylthioimidate 27 was prepared from
pyrazinone 7 via a two-step one-pot process in near quantitative yield. Chlorination of the pyrazinone
ring in 3 followed by hydrolysis and amide coupling completed the synthesis of 1. This chroma-
tography-free synthesis was used successfully to prepare multikilogram quantities of the drug with
reproducibility and high purity.
In tr od u ction
free pyridine and a different arylmethylamine (Scheme
1). Our initial synthetic efforts were predicated on the
expectation that a late intermediate common to com-
pound 2a could be easily oxidized to a pyridine N-oxide.
However, 2a -d rapidly degraded in the presence of
oxidants via fragmentation of the pyrazinone ring based
on LC/MS analyses. Therefore, the pyridine N-oxide
functionality required construction prior to appending to
the pyrazinone. Scheme 1 illustrates the three necessary
building blocks: 2,2-difluoro-2-(1-oxido-2-pyridinyl)ethy-
lamine (4), pyrazinone 7, and 2-fluorobenzylamine. Com-
pound 4 could be derived from a selective oxidation of
pyridine amine 6, which in turn would be made from
2-bromopyridine. The preparation of pyrazinone 7 has
been described previously from oxalate, glycine ester, and
glycinal acetal,1 and 2-fluorobenzylamine is commercially
available. We expected to join 4 and 7 via halogenated
imidate 5.1 However, this coupling reaction suffered due
to the weak nucleophilicity of amine 4 and the poor
reactivities of halo-imidates 5. This paper reports our
solution to this coupling problem via activation by
2-pyridylthioimidate and ZnCl2, which significantly im-
proved the coupling efficiency. In addition, we detail the
characterization of several ZnCl2 complexes in an effort
to gain further insight into the mechanism of the reac-
tion. We also describe a selective oxidation of pyridine
amine 4, an efficient preparation of 27, and the final
construction to 1.
Compound 1 is a potent, orally active thrombin inhibi-
tor potentially useful for the prevention and treatment
of venous and cardiogenic thromboembolism.1 The com-
pound is a rapid-binding, reversible competitive inhibitor
of thrombin with high selectivity versus other protease
targets. The compound has good oral bioavailability and
shows efficacy against thrombosis in the rat ferric
chloride and the African Green Monkey models. In
addition to a clean profile in the in vitro assays, the
compound exhibited minimum side effects in a Rhesus
model. To further study this molecule and its biological
activity, an efficient synthesis was required.
Compound 2a , a related thrombin inhibitor analogue
reported previously,1 differs from 1 by only a N-oxide-
† Department of Analytical Research.
(1) (a) Burgey, C. S.; Robinson, K. A.; Lyle, T. A.; Nantermet, P. G.;
Selnick, H. G.; Isaacs, R. C. A.; Lewis, S. D.; Lucas, B. J .; Krueger, J .
A.; Singh, R.; Miller-Stein, C.; White, R. B.; Wong, B.; Lyle, E. A.;
Stranieri, M. T.; Cook, J . J .; McMasters, D. R.; Pellicore, J . M.; Pal,
S.; Wallace, A. A.; Clayton, F. C.; Bohn, D.; Welsh, D. C.; Lynch, J . J .,
J r.; Yan, Y.; Chen, Z.; Kuo, L.; Gardell, S. J .; Shafer, J . A.; Vacca, J .
P. Bioorg. Med. Chem. Lett. 2003, 13, 1353-1357. (b) Burgey, C. S.;
Robinson, K. A.; Lyle, T. A.; Sanderson, P. E. J .; Lewis, S. D.; Lucas,
B. J .; Krueger, J . A.; Singh, R.; Miller-Stein, C.; White, R. B.; Wong,
B.; Lyle, E. A.; Williams, P. D.; Coburn, C. A.; Dorsey, B. D.; Barrow,
J . C.; Stranieri, M. T.; Holahan, M. A.; Sitko, G. R.; Cook, J . J .;
McMasters, D. R.; McDonough, C. M.; Sanders, W. M.; Wallace, A. A.;
Clayton, F. C.; Bohn, D.; Leonard, Y. M.; Detwiler, T. J ., J r.; Lynch, J .
J ., J r.; Yan, Y.; Chen, Z.; Kuo, L.; Gardell, S. J .; Shafer, J . A.; Vacca,
J . P. J . Med. Chem. 2003, 46, 461-473. (c) Dolling, U. H.; Alabaster,
R. J .; Ashwood, M. S.; Cottrell, I. F.; Cowden, C. J .; Davies, A. J . PCT
Int. Appl. WO 0246160, 2002. (d) Cowden, C.; Cooper, V. B.; Rush, D.
J .; Frech, P. PCT Int. Appl. WO 0192234, 2001. (e) Burgey, C.; Isaacs,
R. C.; Dorsey, B. D.; Robinson, K. A.; Staas, D.; Sanderson, P. E.;
Barrow, J . PCT Int. Appl. WO 0138323, 2001. (f) Burgey, C. S.;
Robinson, K. A.; Williams, P. D.; Coburn, C. A.; Lyle, T. A.; Sanderson,
P. E. PCT Int. Appl. WO 0075134, 2000.
Resu lts a n d Discu ssion
P yr id in e N-Oxid e Am in e 4. A seven-step route1 to
4 (Scheme 2, route A) was previously reported from these
laboratories involving condensation of 2-lithiopyridine
10.1021/jo034835e CCC: $25.00 © 2003 American Chemical Society
Published on Web 10/11/2003
8838
J . Org. Chem. 2003, 68, 8838-8846

Synthesis of 3-Aminopyrazinone
SCHEME 1. Retr osyn th etic An a lysis of 1
SCHEME 3. P yr id in e N-Oxid a tion of 6
SCHEME 4. P r ep a r a tion of Br om op yr a zin on es 16
a n d 19
SCHEME 2. Syn th esis of 4
temperatures. In addition, the conversion of alcohol 10
to amine 6 was accomplished by directly treating the
corresponding triflate with ammonium hydroxide. Pure
amine 6 was isolated as the benzenesulfonate salt (BSA).1
Two methods were developed for the conversion of
pyridine amine 6 to pyridine N-oxide 4 (Scheme 3) based
on Caron et al.’s trifluoroacetic anhydride (TFAA)/urea
hydrogen peroxide (UHP) procedure:4 (1) protection of the
primary amine as a trifluoroacetamide 14, pyridine
N-oxidation with TFAA/UHP to give crystalline deriva-
tive 15, solvolysis with ammonium hydroxide, and isola-
tion of 4 as the trifluoroacetic acid salt; or (2) direct
oxidation of the trifluoroacetic or the benzenesulfonic salt
of 6. We used the later route for the large-scale prepara-
tion due to its simplicity.
Trifluoroperacetic acid (TFPAA) was generated by the
addition of TFAA to a slurry of excess UHP at -20 °C.5
Pyridine amine 6‚BSA was then added to the peracid at
<0 °C and amine 4 was eventual isolated as the crystal-
line amine 4‚HCl salt. Formation of the free base
produced pyridine N-oxide 4 in isolated yields of 80% on
both the gram and the kilogram scale.
F ir st Gen er a tion Cou p lin g Rea ction s. Initially, we
examined the coupling of pyridine N-oxide amine 4 with
two bromopyrazinone derivatives, bromoester 16 and
bromoamide 19 (Schemes 4 and 5). Bromoester 16 was
prepared from pyrazinone 7 with use of POBr3 in aceto-
and diethyl oxalate followed by DAST difluorination,
NaBH4 reduction, triflate formation, and azide displace-
ment to afford 2,2-difluoro-2-(2-pyridyl)ethyl azide (12,
34% overall yield). Subsequent MCPBA N-oxidation,
followed by triphenylphosphine reduction of azide 13
afforded amine 4. A more efficient synthesis of 4 was later
devised based on Kobayashi et al.’s copper-mediated
coupling of 2-bromopyridine and iododifluoroacetate (route
B).2 Several modifications were made to make this
reaction more amenable for large-scale synthesis. The
more readily available ethyl bromodifluoroacetate re-
placed ethyl iododifluoroacetate.3 DMF replaced DMSO
due to safety concerns with Cu-DMSO exotherms at high
(4) Caron, S.; Do, N. M.; Sieser, J . E. Tetrahedron Lett. 2000, 41,
2299-2302.
(2) (a) Taguchi, T.; Kitagawa, O.; Morikawa, T.; Nishiwaki, T.;
Uehara, H.; Endo, H.; Kobayashi, Y. Tetrahedron Lett. 1986, 27, 6103-
6106. (b) Kitagawa, O.; Taguchi, T.; Kobayashi, Y. Chem. Lett. 1989,
389-392.
(3) Ashwood, M. S.; Cottrell, I. F.; Cowden, C. J .; Wallace, D. J .;
Davies, A. J .; Kennedy, D. J .; Dolling, U. H. Tetrahedron Lett. 2002,
43, 9271-9273.
(5) Early oxidation efforts with TFPAA, produced by the addition
of TFAA to UHP under acidic conditions in the presence of 1.2 equiv
of TFA or BSA to protect the primary amine, gave a selective oxidaton
of the pyridine ring in good yields, but the reaction mixture contained
about 30% of the trifluoroamide 15. Although an aqueous NaOH
workup cleaved the trifluoroacetamide, isolation via extraction of very
water soluble pyridine N-oxide 4 proved to be difficult.
J . Org. Chem, Vol. 68, No. 23, 2003 8839

Chung et al.
SCHEME 6. Su lfon yla tion of P yr a zin on e 7
SCHEME 5. Cou p lin g w ith Br om op yr a zin on es 16
a n d 19
addition of pyridine N-oxide oxygen to bromide 16 became
the major manifold. The resulting adduct 22 hydrolyzed
to 7 upon aqueous workup.
Solvents and bases were screened for the coupling
reaction of 4 and 16 and the best condition achieved was
toluene/EtOH (3:1; 1.5 M), 1.1 equiv of pyridine N-oxide
amine, 1.3 equiv of Hunig’s base (or 2.3 equiv with TFA
salts) at ∼100 °C. The reaction remained slow, requiring
5 days to reach 90% conversion, to give assay yields of
75-80% and isolated yields of 70-75%. Due to safety
concerns,6 the reaction temperature in the kilogram scale
coupling reactions was limited to 85 °C, which then
required 7 days to achieve 85% conversions and 78%
reaction yield.
nitrile at 50-65 °C in 92-97% assay yield and 82-89%
isolated yields as a crystalline solid.1,6 Bromoamide 19
was prepared in two steps by heating pyrazinone 7 with
2-fluorobenzylamine in propionitrile to give hydroxy
amide 18, which was precipitated from the reaction by
the addition of water, followed by reaction with POBr3
in acetonitrile.
The use of the HCl salt of amine 4 in the coupling
reaction was deleterious, resulting in the conversion of
bromopyrazinone 16 to chloropyrazinone 17,1 which
reacted significantly slower than bromide 16 and led to
poor coupling yields (50% vs 75%). In the presence of 2
equiv of NaI, the reaction proceeded through the inter-
mediary iodoimidate, but gave only 46% conversion after
4 days. Other approaches, such as Pd-catalyzed amina-
tion conditions9 (CsCO3, Pd2(dba)3, BINAP, toluene), gave
only ∼25% yields of product. Attempted coupling between
trifluoroamide 15 and bromide 16 with NaH, KOtBu, or
Mitsunobu conditions between 15 and 7 failed to give the
desired product. Several attempts to prepare the fluoro
analogue of 5 also failed.
Cou p lin g of P yr id in e N-Oxid e Am in e a n d Th io-
im id a te. Despite the ability of the above coupling
reaction to produce kilogram quantities of 3, the long
reaction time required was not practical. Although there
are few known approaches to enhance the nucleophilicity
of primary amines,10 we searched for alternatives for
bromopyrazinone 16 due to its instability and potential
shock sensitivity safety problems.6 We briefly examined
activation of 7 via O-sulfonyl imidates 23. However,
reaction of 7 with mesyl chloride,11 tosyl chloride,12 and
triflic anhydride13 provided predominantly N-sulfonylated
products 24, which did not react with 4 (Scheme 6).
We were attracted to 2-mercaptopyridine as an activa-
tion group based on its successful application in the areas
The coupling reactions of pyridine N-oxide 4 with
bromoester 16 to form 3 only reached 50% conversion
after refluxing in acetonitrile or toluene for 80 h. The
coupling reaction of 4 with bromo-amide 19 was further
complicated by the poor solubility in these solvents. Polar
solvents, such as DMSO or DMPU at 120 °C, gave
significant decomposition. In contrast, the coupling of
pyridine amine 6 with 16 to make 2d proceeded three
times faster under the same conditions.1,7 We speculated
that the pyridine N-oxide functionality in 4 further
contributes to the electron deficiency of the primary
amine relative to 6, and accounts for the decreased
nucleophilicity (calculated pKa values for 6 and 4 are 6.3
and 5.9, respectively). The poor nucleophilicity of amine
4 is also supported by the fact that the formation of bis-
adduct 21b is negligible relative to bis-adduct 21a formed
as a significant byproduct from the reaction with amine
6.
In addition to decreased basicity, the pyridine N-oxide
oxygen may be acting as a competing nucleophile. Pyri-
dine N-oxides are known to add to imidoyl chlorides at
or below room temperatures.8 To amplify this side reac-
tion, we performed the coupling reaction between pyri-
dine N-oxide amine 4 and bromide 16 in refluxing
anhydrous ethanol without base. The reaction reached
97% consumption of bromide 16 after 72 h, but gave only
50% assay yield of product, and upon workup significant
amounts of pyrazinone 7 (7:3 ) 42:57 relative LCAP at
330 nm) were found. The reaction profile of this coupling
showed that after production of 3 plateaus at about 50%,
the formation of pyrazinone 7 predominated. It appeared
that at 50% conversion, the remaining unreacted amine
4 was completely protonated as the HBr salt, and the
(9) (a) Wolfe, J . P.; Tomori, H.; Sadighi, J . P.; Yin, J .; Buchwald, S.
L. J . Org. Chem. 2000, 65, 1158-1174. (b) Wolfe, J . P.; Buchwald, S.
L. J . Org. Chem. 2000, 65, 1144-57. (c) Chakraborty, M.; McConville,
D. B.; Niu, Y.; Tessier, C. A.; Youngs, W. J . J . Org. Chem. 1998, 63,
7563-7567.
(10) Ashton, R. P.; Calcagno, P.; Spencer, N.; Harris, K. D. M.; Philp,
D. Org. Lett. 2000, 2, 1365-1368 and references therein.
(11) Holan, G.; Virgona, C. T.; Watson, K. G. Bioorg. Med. Chem.
Lett. 1996, 6, 77-80.
(12) Sekine M. J . Org. Chem. 1989, 54, 2321-6. Kamimura, T.;
Masegi, T.; Sekine, M.; Hata, T. Tetrahedron Lett. 1984, 25, 4241-4.
(13) (a) Charette, A. B.; Grenon, M. Tetrahedron Lett. 2000, 41,
1677-1680. (b) Charett, A. B.; Chua, P. J . Org. Chem. 1998, 63, 908-
909. (c) Sforza, S.; Dossena, A.; Corradini, R.; Virgili, E.; Marchelli, R.
Tetrahedron Lett. 1998, 39, 7110714. (d) Thomas, E. W. Synthesis 1993,
767-768. (e) Sisti, N. J .; Zeller, E.; Grierson, D. S.; Fowler, F. W. J .
Org. Chem. 1997, 62, 2093-2097.
(6) CAUTION! Solid bromopyrazinone 16 is thermally unstable
with an exothermic initiation at ∼50 °C, potentially shock sensitive
(1 out of 6 positive drop weight tests), and gradually decomposes at
room temperature.
(7) Fleitz, F. J .; Lyle, T. A.; Zheng, N.; Armstrong, J . D., III; Volante,
R. P. Synth. Commun. 2000, 30, 7171-3180.
(8) (a) Manley, P. J .; Bilodeau, M. T. Org. Lett. 2002, 4, 3127-3129
and references therein. (b) Pyridine N-oxide addition to imidoyl
chloride: Tachikawa, R.; Wachi, K.; Sato, S.; Terada, A. Chem. Pharm.
Bull. 1981, 29, 3529-3535.
8840 J . Org. Chem., Vol. 68, No. 23, 2003

Synthesis of 3-Aminopyrazinone
SCHEME 7. P r ep a r a tion of P yr id ylth iop yr a zin on e
27
F IGURE 1. Solution NMR data for 31.
SCHEME 8. Cou p lin g w ith P yr id ylth iop yr a zin on e
27
proceeded slowly and resulted in only a 46% conversion
after 2 days but with a very clean purity profile.
The rate of the coupling of 27/4/ZnCl2 was slower at
lower temperatures, as noted by the reaction carried out
in refluxing THF, taking 4 days to reach completion at
67 °C compared to 1 day at 88 °C. Other higher boiling
solvents were screened, with acetonitrile, dioxane,
dimethoxyethane, and dichloroethane providing better
yields and better solubility, whereas aromatic solvents
such as toluene, chlorobenzene, and trifluorotoluene did
not offer enough solubility and the reaction mixtures
tended to form gelatinous mixtures. At a concentration
of 1 M in dioxane with 1 equiv of ZnCl2, the reaction
proceeded to 84% yield after 4.5 h at 88 °C, and 94% after
20 h. Reactions in acetonitrile (83 °C) or 1,2-dichloroet-
hane (83 °C) gave similar rates and yields.
Zn Cl2 Com p lexes. During the coupling reaction of 27/
4/ZnCl2 in acetonitrile a precipitate formed, which was
found to be a 1:2 ZnCl2/2-thiopyridone complex (28). The
structure was confirmed by single-crystal X-ray analysis
(recrystallized from EtOH). Interestingly, in the solid
state, the zinc was coordinated to the sulfurs, but in
DMSO-d6 solution, the NMR evidence (see the Support-
ing Information) was consistent with nitrogen coordina-
tion.18
of macrolide formation, peptide coupling, and glycoside
syntheses.14-16 Its ability to provide a template effect in
the presence of a metal should enhance reactivity. The
preparation of thioimidate 27 was accomplished by the
addition of 2-mercaptopyridine to chloroimidate 17, which
was generated in situ from pyrazinone 7 with oxalyl
chloride in acetonitrile/DMF (Scheme 7).17 Addition of
base (TEA) was unnecessary and undesirable due to a
marked slowing of the reaction of chloroimidate with
2-mercaptopyridine. Upon completion of reaction, addi-
tion of brine to the reaction mixture at pH 5-6 precipi-
tated thioimidate 27 in 95% yield and 98 wt % assay
purity. Control of the aqueous quench at pH <7 and <30
°C was important to avoid ester hydrolysis.
To gain further insight into the mechanism of the zinc-
(II)-mediated coupling, we studied the complexation of
ZnCl2 with each of the starting materials, pyridine
N-oxide 4 and thioimidate 27, by solution (1H, 13C, 15N)
and solid-state NMR and X-ray diffraction. In both cases,
solid complexes were formed upon mixing in acetonitrile
and were isolated by filtration. Solution NMR of com-
plexed thioimidate 27 provided evidence that the zinc was
coordinated to both N-1′ and N-5 as in 31 (Figure 1).
Specifically, N-1′ and N-5 were shifted upfield from ∼322
and ∼316 ppm, respectively, in 27 to ∼260 and 277 ppm,
respectively, for the complexed species 31. Also, there
were several diagnostic 13C shift changes: (1) both C-3
and C-4′ shifted downfield from 128.0 and 138.3 ppm,
respectively, in 27 to 132.4 and 142.7 ppm in 31; (2) both
C-2′ and C-6 shifted upfield from 153.8 and 158.5 ppm,
respectively, in 27 to 152.8 and 156.9 ppm in 31; and (3)
With thioimidate 27 in hand, we proceeded to study
its coupling reaction with pyridine N-oxide amine 4.
Initially, the reaction carried out in refluxing i-PrOAc
without any additive was very slow with generation of a
new impurity, which appeared to be the product derived
from attack on the pyridine carbon bearing sulfur. With
the use of ZnCl2, we observed significant rate enhance-
ment. The reaction of thioimidate 27 and amine 4 in the
presence of 1 equiv of ZnCl2 in THF (0.5 M) at 82 °C in
a sealed tube afforded a 92% assay yield of adduct 3 after
24 h with only 4% of 27 remaining (Scheme 8). Other
Lewis acids such as CuBr2 and MgCl2, and protic acid
(HCl) were inferior to ZnCl2. The coupling reaction
between amine 4 and chloroimidate 17 with catalytic
amounts of 2-mercaptopyridine (20%) and ZnCl2 (15%)
(14) (a) Schmidt, U.; Gleich, P. Angew. Chem. 1985, 97, 606-607.
(b) Lloyd, K.; Young, G. T. J . Chem. Soc. (C) 1971, 2890-6.
(15) For a 2-mercaptopyrdine review, see: Schmidt, B.; Ku¨hn, C.
J . Prakt. Chem. 1999, 341, 114-120.
(16) Zhou, X.-X.; Welch, C. J .; Chattopadhyaya, J . Acta Chem.
Scand. 1986, B40, 806-816.
(17) Early effort with POCl3 for the preparation of 16 gave a slower
reaction and lower yield, and coprecipitation of inorganic phosphate
salts with isolated product 27 after workup.
(18) (a) Moran, D.; Sukcharoenphon, K.; Puchta, R.; Schaefer, H.
F.; Schleyer, P. v. R.; Hoff, C. D. J . Org. Chem. 2002, 67, 9061-9069.
(b) Stoyanov, S.; Petkov, I.; Antonov, L.; Stoyanove, T. Can. J . Chem.
1990, 68, 1482-1489. (c) Shunmugam, R.; Sathyanarayana, D. N. J .
Coord. Chem. 1985, 14, 27-30. (d) Kennedy, B. P.; Lever, A. B. P. Can.
J . Chem. 1972, 50, 3488-3507. (e) Penfold, B. R. Acta Crystallogr.
1953, 6, 707-713.
J . Org. Chem, Vol. 68, No. 23, 2003 8841

Chung et al.
F IGURE 4. Chlorination reaction impurities.
information could be extracted. However, elemental
analysis data confirmed a 1:1 composition of 4 to ZnCl2.
A competitive complexation study with a 1:1:1 mixture
of 4, 27, and ZnCl2 was examined via NMR in CD3CN.
The NMR signals observed were intermediate values
between that of complexed and noncomplexed species,
indicating that there is rapid equilibrium between all the
species, and no one particular species predominates. This
is very likely the reason ZnCl2 is the optimum Lewis acid.
Op tim iza tion of Zn Cl2 Stoich iom etr y. In theory,
the reaction requires only 0.5 equiv of ZnCl2 to form the
2:1 complex 28. We screened 0.25, 0.6, 0.75, and 1.0 equiv
of ZnCl2 in the coupling reaction, and the yields for all
these runs fell within the 80-94% range. The fact that
sub-stoichiometric 0.25 equiv of ZnCl2 still gives a
reasonable yield suggests that the 2-thiopyridone-ZnCl2
complexes must be an active catalyst or these complexes
could easily disproportionate to liberate free ZnCl2.
Indeed, when complex 28 was used in place of ZnCl2 in
the coupling reaction, a 75-80% assay yield was obtained
after 24 h.
F IGURE 2. Solid-state 15N NMR for 27 and 31.
F IGURE 3. Solution NMR data for 4 and 32.
no significant change in the 13C shift of C-1 was observed,
154.5 ppm in 27 versus 154.8 ppm in 31.
The optimized coupling condition uses 0.75 equiv of
ZnCl2, 1.1 equiv of amine 4, and 1.0 equiv of thioimidate
27 at a concentration of 1 M in acetonitrile at reflux (82-
84 °C) for 24 h. This was successfully demonstrated on
2-10 kg scales and achieved >98% conversion and 90-
95% reaction yield in 24 h. Product isolation involved
filtering the reaction mixture to remove the insoluble zinc
salts and crystallization from ethanol to afford amidine
3 in 87% isolated yield and g95 wt % purity.
Use of this material in the subsequent chlorination step
gave 10-15% lower yields than expected. It was deter-
mined that residual zinc salts present in 3 were the cause
for the lower yield and high impurity levels. This was
confirmed by spiking ZnCl2 or Zn salt 28 into a reaction
mixture starting with Zn-free starting materials. Some
of the impurities generated were identified to be dimers
33, 34, and thiopyridine adducts 35, presumably arising
from radical reactions (Figure 4). Indeed, ZnCl2 is known
to act as an NCS radical promoter.19 This problem was
resolved by EDTA treatment of amidine 3 in an aqueous
acetonitrile solution. Subsequent removal of acetonitrile
by vacuum distillation resulted in the crystallization of
amidine 3 from water. In this manner, the Zn content in
amidine 3 was consistently reduced to <10 ppm with
>95% recovery of product.
As a complement to the solution NMR work, solid-state
15N CP/MAS NMR was also performed to gain insight
into zinc coordination. Akin to the solution results, 31
displayed three nitrogen sites representing the three
unique nitrogen environments in the molecule (Figure
2). The 15N spectrum of the undried precipitate after
ZnCl2 reaction displayed an upfield shift of the N-1′and
N-5 sites in 27, as well as a minimal downfield shift for
the N-2 species. These 15N shift values are consistent with
the solution spectra and suggest zinc coordination to both
N-1′ and N-5 in 31. Additionally, it is clear that the
undried precipitate is an acetonitrile solvate as the peak
representing incorporated acetonitrile is present in the
spectrum at 241.4 ppm. Solvation is also confirmed by
13C solid-state NMR work (not displayed). Finally, the
structure of 31 as the acetonitrile solvate was confirmed
by single-crystal X-ray.
For pyridine N-oxide 4, the zinc complex appears to
be 32 by (1H, 13C, 15N) solution NMR (Figure 3). This
structure is based upon the upfield 15N shift for N-1 from
∼291 ppm in 4 to ∼281 ppm in 32 and the downfield shift
for NH N-9 from ∼7 ppm in 4 to ∼13 ppm in 32. In
addition, there were changes in the 13C and 1H shifts:
(1) C-4 and C-8 shifted downfield from 126.2 and 45.3
ppm, respectively, in 4 to 130.7 and 46.3 ppm in 32. (2)
H-4, H-6, and H-8 were all shifted downfield from 7.38,
8.18, and 3.63 ppm, respectively, in 1 to 7.66, 8.42, and
3.80 ppm in 4.
Oth er Th ioim id a te An a logu es. Three other thio-
imidate analogues were also prepared and evaluated in
the coupling reaction with amine 4. In all three cases,
Amine 4-ZnCl2 complex was shown by solid-state
NMR to be an amorphous solid, therefore not much
(19) Ko, M.; Sato, T.; Otsu, T. J . Macromol. Sci., Chem. 1975, A9,
199-210.
8842 J . Org. Chem., Vol. 68, No. 23, 2003

Synthesis of 3-Aminopyrazinone
F IGURE 5. Other thioimidates.
SCHEME 9. Ch lor in a tion of P yr a zin on e 3
F IGURE 6. Acyl urea impurity.
36, 37, and 38 (Figure 5), the yields of their preparation
and subsequent coupling reactions were significantly
inferior relative to those of pyridine thioimidate 27,
confirming that bidendate chelation as in 31 is important
for rate-enhanced, high-yielding reactions.
Ch lor in a tion of P yr a zin on e 3. With ester 3 in hand,
chlorination was attempted on amide 3a , which was
prepared from ester 3 via hydrolysis followed by EDC
coupling with 2-fluorobenzylamine (Scheme 9). Amide 3a
turned out to be poorly soluble in acetonitrile and gave
very little reaction with NCS even at 65 °C. Complete
dissolution was achieved at 80 °C, but the reaction
produced large amounts of bis-chlorination byproduct
40a . On the other hand, ester 3 was completely soluble
in acetonitrile at 55-65 °C, and slow addition of NCS at
these temperatures produced chloropyrazinone 39 in
>90% yield.
A slight excess of NCS was necessary to ensure
complete chlorination, but formation of bis-chloro pyrazi-
none 40 became competitive. Because ester 3 is easier to
reject by crystallization than dichloro 40, it is important
to limit the amount of excess NCS to achieve acceptable
levels of 40 in the final isolation of 1.
We found that small amounts of water in the reaction
mixture led to the formation of an unstable impurity
identified as hydroxy dione 41. This material rapidly
decomposes to amine 4 and an unknown byproduct. It
was therefore important to control the water content of
the reaction (<100 µg/mL) to maintain high yield. After
the chlorination was completed, addition of the acetoni-
trile reaction mixture to a 10% aq NaCl solution precipi-
tated chloropyrazinone 39 in 85-90% isolated yield.
Other chlorinating agents, such as dichloro dimethylhy-
dantoin, gave messy reactions, using various solvents and
temperatures. In most cases, the formation of hydantoin
adducts predominated.
F IGURE 7. Bis-HCl and mesylate salts of compound 1.
pH 11 coupling required 2 days at 67 °C and gave an
isolated yield of 74%. At pH 6, the coupling reaction
completed in less than 3 h at 22 °C and the isolated yield
of 1 was 90%. In initial experiments we observed forma-
tion of acyl urea 42 based on LCMS (515) (Figure 6),
which varied from a few percent to upward of 20% in
reactions. To control this impurity, we changed the order
of the reagent addition after the hydrolysis reaction to
adding 2-fluorobenzylamine first, followed by pH adjust-
ment, then adding HOBt and EDC-HCl. This modifica-
tion resulted in a reproducible and discernible yield
improvement (>92% isolated yield with <1 area % 42).
Compound 1 could also be isolated as the bis-HCl salt or
the mesylate salt (Figure 7).
In summary, we have developed a practical synthesis
of 1 in high yield and high purity, suitable for large-scale
preparation. The utilization of thioimidate/ZnCl2 leaving
group/activation in the key coupling step overcame the
problem of weak amine nucleophile coupling with a poor
electrophile, and provided the coupling rate acceleration
through a template effect. In addition, we have gained
some mechanistic understanding through the character-
ization of several ZnCl2 complexes by NMR and X-ray
diffraction studies.
Exp er im en ta l Section
2,2-Diflu or o-2-(1-oxid o-2-p yr id in yl)eth yla m in e (4). To
a slurry of urea hydroperoxide (UHP, 114.8 mmol, 10.8 g) in
acetonitrile (30 mL) cooled to -10 °C was added a solution of
trifluoroacetic anhydride (TFAA, 99.5 mmol, 20.9 g) in aceto-
nitrile (7 mL) over 75 min while maintaining the temperature
at <0 °C. The mixture was stirred 30 min at -5 °C to give a
near homogeneous solution, then cooled to -10 °C. To the
solution was added solid 6-benzenesulfonate salt1 (18.1 g, 90
wt %, 50.0 mmol assay) over 30 min, maintaining the tem-
perature at <0 °C. The reaction remained completely homo-
geneous during the addition of the first ∼20% of pyridine 6,
F or m a tion of Com p ou n d 1. The formation of 1 was
accomplished as a one-pot hydrolysis of ester 39 and
coupling with 2-fluorobenzylamine, using EDC and HOBt.
The rate of the coupling reaction was pH dependent. At
J . Org. Chem, Vol. 68, No. 23, 2003 8843

Chung et al.
then benzenesulfonic acid (BSA) precipitated. The slurry was
maintained at -5 °C, and stirred ∼30 min until the reaction
reached completion (<0.1 LC area % @ 210 nm of 6 vs 4). The
slurry was cooled to -10 °C. The mixture was filtered, and
the solids washed with cold (-10 °C) acetonitrile (20 mL). The
filtrate was maintained at -5 °C and concentrated aq HCl (6.6
mL) was added over 20 min to produce a thick white precipi-
tate. The slurry was stirred at -5 °C for 1 h and filtered, and
the solids were washed with cold (-5 °C) acetonitrile (20 mL)
and dried with vacuum/N2 at 20 °C until a constant weight
(13.5 g). The assay of isolated solid was 69.2 wt % as 4-HCl,
83.4 LC area % with 16.6 LC area % BSA. The solid was added
to a mixture of ethyl acetate (40 mL) and triethylamine (13
mL) at 20 °C. The resulting slurry was stirred for 19 h, diluted
with isopropyl acetate (6 L), and filtered. The solids were
washed with isopropyl acetate (35 mL). The solution was
concentrated in vacuo at <35 °C to an oil. The oil was diluted
with acetonitrile (2 L) and re-concentrated. The dilution and
concentration was repeated twice to obtain an oil (7.7 g, 91
wt % as 4) for an overall isolated yield of 80%. After this
preparation, safety studies restricted the concentration of 4
to <17 wt % in either ethyl acetate, isopropyl acetate, or
acetonitrile. HPLC assay: column, Luna C18, 4.6 × 250 mm;
mobile phase, acetonitrile/0.1% aqueous H3PO4 1% to 50% over
20 min; flow rate, 1.0 mL/min; wavelength, 205 nm; tR for 4,
2.7 min, for 14, 4.7 min, for BSA, 11.6 min. Free base 4: mp
38-40 °C; 1H NMR (CDCl3) δ 1.40 (br s, NH2), 3.74 (t, J )
15.2 Hz, 2H), 7.34 (m, 2H), 7.67 (m, 1H), 8.24 (m, 1H). 13C
NMR (CDCl3) δ 143.6, 141.1, 127.2, 125.4, 125.3, 119.1, 44.8.
1H NMR (499.87 MHz, CD3CN) δ 8.18 (br d, J ) 6.4 Hz, 1H),
7.66 (dd, J ) 8.0, 2.0 Hz, 1H), 7.44 (ddd, J ) 8.0, 6.4, 2.0 Hz,
1H), 7.38 (td, J ) 8.0, 1.2 Hz, 1H), 3.63 (t, J HF ) 15.1 Hz, 2H);
13C NMR (125.69 MHz, CD3CN) δ 144.1 (t, J CF ) 30.1 Hz, C-2),
142.1 (C-6), 128.9 (C-5), 126.6 (t, J CF ) 7.4 Hz, C-3), 126.2 (C-
4), 120.9 (t, J CF ) 242.7 Hz, C-7), 45.3 (t, J CF ) 27.4 Hz, C-8);
15N NMR (50.66 MHz, CD3CN, reference to external NH3,
burgundy solution (pH 0) was adjusted to pH 5-6 with solid
NaHCO3 (∼42 g), during which time two clear layers of liquid
formed. The mixture was seeded with several milligrams of
product and concentrated under reduced pressure (30-50
mmHg, 24 °C) to about 850 mL, during which time the product
crystallized. The resulting mixture (pH 6.8) was adjusted to
pH 6 with 1.4 mL of 1 N HCl and then stirred overnight at 21
°C. The mixture was stirred at 0-5 °C for 1-2 h, filtered, and
washed with 0-5 °C water (3 × 100 mL). The wet cake was
vacuum dried at 22 °C with a nitrogen sweep overnight. Yield
) 66.6 g of 27 as a light beige solid. LCAP @ 215 nm ) 99.5
area %; LCWP ) 99.8% (66.5 g, 95.1% yield); no impurity
>0.11 area %. Product loss to combined ML/washes was 2.5%
(2.2 g). HPLC conditions: column, Luna C18, 5 µm; 4.6 × 250
mm; mobile phase, CH3CN/0.1% aqueous H3PO4, 20% to 80%
over 20 min; flow rate, 1.0 mL min-1; UV detection, 215 nm;
column temperature, 25 °C; tR for DMF, 3.10 min, for dione 7,
4.00 min, for 2-mercaptopyridine, 3.61 min, for 2-pyridylsul-
fide, 8.50 min, for chloropyrazinone 17, 9.23 min, for thio-
pyrazinone 27, 9.43 min. 27: mp 97-99 °C; 1H NMR (CDCl3)
δ 8.64 (m, 1H), 7.72 (m, 2H), 7.29 (m, 1H), 7.10 (d, J ) 4.4 Hz,
1H), 6.87 (d, J ) 4.4 Hz, 1H), 4.62 (s, 2H), 4.25 (q, J ) 7.2 Hz,
2H), 1.29 (t, J ) 7.2 Hz, 3H). 13C NMR (CDCl3) δ 166.4, 159.1,
153.5, 152.7, 150.6, 137.1, 130.4, 125.2, 123.3, 123.2, 62.3, 50.0,
14.1. 1H NMR (499.87 MHz, CD3CN) δ 8.58 (ddd, J ) 4.8, 2.0,
1.0 Hz, 1H), 7.79 (td, J ) 7.8, 2.0 Hz, 1H), 7.73 (dt, J ) 8.0,
1.2 Hz, 1H), 7.35 (ddd, J ) 7.6, 4.8, 1.2 Hz, 1H), 7.13 (d, J )
4.4 Hz, 1H), 7.06 (d, J ) 4.4 Hz, 1H), 4.61 (s, 2H), 4.21 (q, J )
7.2 Hz, 2H), 1.25 (t, J ) 7.2, 3H); 13C NMR (125.69 MHz, CD3-
CN) δ 168.0 (C-8), 158.5 (C-6), 154.5 (C-1), 153.8 (C-2′), 151.4
(C-6′), 138.3 (C-4′), 131.3 (C-3′), 128.0 (C-3), 124.5 (C-5′), 123.6
(C-4), 62.9 (C-9), 51.3 (C-7), 14.5 (C-10); 15N NMR (50.66 MHz,
CD3CN, reference to external NH3, shifts were measured from
2D 1H-15N HMBC) δ ∼322 (N-1′), ∼316 (N-5), ∼161 (N-2).
Anal. Calcd for C13H13N3O3S: C, 53.60; H, 4.50; N, 14.42; S,
11.01. Found: C, 53.49; H, 4.44; N, 14.33; S, 11.09.
1
shifts were measured from 2D H-15N HMBC) δ ∼291 (N-1),
{3-[2,2-Diflu or o-2-(1-oxy-p yr id in -2-yl)-et h yla m in o]-2-
oxo-2H-p yr a zin -1-yl}-a cetic Acid Eth yl Ester (3). Meth od
A. Cou p lin g via Br om oim id a te 16. To a round-bottom flask
equipped with a condenser, a thermometer, and nitrogen inlet/
outlet was charged 3:1 toluene/ethanol (4.8 L), amine 4 (1104
g, 6.34 mol), bromopyrazinone ester 16 (1530 g, 5.76 mol), and
N,N-diisopropylethylamine (1.305 L, 7.36 mol). The mixture
was stirred and heated to reflux at 85-88 °C under a nitrogen
atmosphere for 7 days. The reaction mixture (85% conversion
by LC assay) was cooled to 22 °C, diluted with 1/4 of a solution
of H2O (21.7 L) and NaCl (2.17 kg), and concentrated under
reduced pressure at 25-30 °C. After most of the organic
solvents were removed, the rest of the brine solution was added
and the mixture was stirred at ambient for 2-4 h. The slurry
was filtered and washed with NaCl solution (10.9 L of H2O +
1.09 kg of NaCl) and H2O (5.4 L). The wet cake was vacuum
dried at 22 °C under nitrogen overnight. Yield ) 1.63 kg of 3
as a beige solid; 82% yield. LCAP @ 330 nm ) 96.8 area %;
LCWP ) 96.9%; 2.5 area % bis-adduct 21b; 0.7 area % (0.53
wt %) bromo pyrazinone 16; 0.03 wt % ionic Br-. HPLC
conditions: column, Discovery RP Amide C16, 4.6 × 250 mm;
mobile phase, CH3CN/0.1% aqueous H3PO4 15/85; flow rate,
1.0 mL min-1; UV detection, 203 and 330 nm; column tem-
perature, 35 °C; tR for 2,2-difluoro-2-(1-oxy-pyridin-2-yl)-ethy-
lamine 4, 2.43 min, for ionic bromide, 3.85 min, for bis-adduct
21b, 5.21 min, for coupled product 3, 9.55 min, for bromo
pyrazinone 16, 11.9 min.
∼7 (N-9). Anal. Calcd for C7H8F2N2O‚1/5H2O: C, 47.30; H, 4.76;
N, 15.76; F, 21.38. Found: C, 47.33; H, 4.70; N, 15.87; F, 21.22.
1
4‚HCl salt: mp 230-240 °C dec. H NMR (400.25 MHz, D2O)
δ 8.39 (d, J ) 6.2 Hz, 1H), 7.92 (m, 1H), 7.79 (br t, J ) 7.9 Hz,
1H), 7.70 (M, 1H), 4.68 (OH, 3H), 4.12 (br t, J ) 15.1 Hz, 2H).
13C NMR (100.65 MHz, D2O) δ 141.3, 140.7 (t, J ) 28.6 Hz),
131.7, 129.7, 125.8 (t, J ) 7.2 Hz), 115.4 (t, J ) 246.1 Hz),
42.2 (t, J ) 26.2 Hz). 19F NMR (376.61 MHz, D2O) δ -104.79.
Anal. Calcd for C7H9ClF2N2O: C, 39.92; H, 4.31; N, 13.30; F,
18.04; Cl, 16.83. Found: C, 39.84; H, 4.10; N, 13.18; F, 18.32;
Cl, 16.67.
Br om oim id a te 16. This material was prepared by using a
published procedure except 1,2-dichloroethane was replaced
by acetonitrile.1f,6
Eth yl [2-Oxo-3-(p yr id in -2-ylth io)p yr a zin -1(2H)-yl]a c-
eta te (27). To a round-bottom flask equipped with a con-
denser, HCl trap, and nitrogen inlet was charged acetonitrile
(350 mL), DMF (6.5 mL), and pyrazinone 7 (48.0 g, 99.5%, 0.24
mol). Oxalyl chloride (23.0 mL, 0.26 mol) was added to the
slurry mixture over 2.5 h at 20-25 °C. The resulting yellow
solution was heated at 40 °C for 1-2 h, or until pyrazinone 7
was <0.5% relative to chloropyrazinone 171 by HPLC assay
(relative area % at 215 nm). An analytical sample of 17 was
obtained by aqueous workup and extraction with IPAC:
1HNMR (CDCl3) δ 7.14 (d, J ) 4.4 Hz, 1H), 7.12 (d, J ) 4.4
Hz, 1H), 4.64 (s, 2H), 4.24 (q, J ) 7.1 Hz, 2H), 1.28 (t, J ) 7.1
Hz, 3H); 13CNMR (CDCl3) δ 161.1, 152.7, 148.7, 139.8, 129.1,
121.6, 62.6, 50.9, 14.1. The resulting brown solution was cooled
to 20 °C, and 2-mercaptopyridine (29.4 g, 0.26 mol) was added
in three equal portions at 20-min intervals. The mixture was
stirred at 22 °C until chloropyrazinone 17 was e0.5 LC area
% relative to 27 (3-20 h). More 2-mercaptopyridine (3-5 mol
%) was added to complete the reaction if required. The
resulting dark green solution was quenched into a solution of
water (800 mL) and NaCl (80 g) at <20 °C. The resulting
Meth od B. Cou p lin g via Th ioim id a te 27. To a round-
bottom flask equipped with a condenser, a thermometer, and
nitrogen inlet was charged amine 4 oil (1510 g, 91.0%, 7.89
mol), acetonitrile (7 L), and thiopyrazinone ester 27 (2410 g,
86.7%, 7.17 mol) at 20 °C. The mixing was endothermic, which
brought down the temperature to 11 °C. The mixture was
warmed to 20 °C, then zinc chloride (748 g, 98%, 5.38 mol)
was added. The mixture exothermed to ∼40 °C and was heated
to reflux at 82-84 °C under a N2 atmosphere for 24 h. The
8844 J . Org. Chem., Vol. 68, No. 23, 2003

Synthesis of 3-Aminopyrazinone
reaction mixture was cooled to 22 °C, diluted with acetonitrile
(65.6 L), and stirred for 1 h at 22 °C. The mixture was filtered
through a silica plug (6.6 kg, wetted with acetonitrile with 3.3
kg of sand on top), and the filter cake was washed with 72 L
of acetonitrile. The filtrate was concentrated to 3.5-4 L,
flushed with ethanol (4 × 4 L) until e0.1% acetonitrile.
Ethanol was added to give a final concentration of ∼10 mL of
ethanol/g of product (∼24 L total). The slurry was heated at
reflux (for 1 h) to give a clear yellow solution, and then cooled
slowly to crystallize the product. The mixture was concentrated
to half of the original volume (∼12 L) and then stirred at 0 °C
for 1 h. The slurry was filtered and washed with ice cold
ethanol (7.2 L) and the wet cake was vacuum dried at +22 °C
under nitrogen overnight. Yield ) 2.32 kg of 3 as a light beige
solid. LCAP @ 215 nm ) 99.2 area %; LCWP ) 94.4% (2.19
kg, 86% yield); 0.8 area % 2-thiolpyridine; 7700 ppm Zn.
Product loss to ML and washes was 1.6% (41 g). HPLC
conditions: column, Luna C18, 5 um; 4.6 × 250 mm; mobile
phase, CH3CN/0.1% aqueous H3PO4 20% to 80% over 20 min;
flow rate, 1.0 mL min-1; UV detection, 215 nm; column
temperature, 25 °C; tR for 2,2-difluoro-2-(1-oxy-pyridin-2-yl)-
ethylamine 4, 1.90 min, for 2-thiolpyridine, 3.61 min, for bis-
adduct 21b, 4.24 min, for coupled product 3, 7.25 min, for
thiopyrazinone 27, 9.43 min.
min, for dimer 33, 8.84 min, for thio adduct 35, 9.86 min, for
Cl-dimer 34, 10.17 min, for di-Cl pyrazinone 40, 11.18 min,
for des-oxide 2c, 12.92 min. 39: mp 162-164 °C; 1H NMR
(DMSO-d6) δ 8.34 (d, J ) 6.4 Hz, 1H), 7.65-7.50 (m, 2H +
NH), 7.38 (t, J ) 7.7 Hz, 1H), 6.82 (s, 1H), 4.85 (s, 2H), 4.45
(dt, J ) 13.4, 6.7 Hz), 4.16 (q, J ) 7.1 Hz, 2H), 1.19 (t, J ) 7.1
Hz, 3H); 13C NMR (DMSO-d6) δ 167.4, 151.4, 149.7, 142.4 (t,
J ) 28.5 Hz), 141.3, 128.7, 125.7, 125.3 (t, J ) 6.7 Hz), 119.9,
117.9 (t, J ) 245.7 Hz), 117.6, 62.0, 46.7, 42.0 (t, J ) 28.0
Hz), 14.4. Anal. Calcd for C15H15ClF2N4O4: C, 46.34; H, 3.89;
Cl, 9.12; F, 9.77; N, 14.41. Found: C, 46.34; H, 4.04; Cl, 9.14;
F, 9.77; N, 14.38.
2-F lu or oben zyl 3-(2,2-Diflu or o-2-(2-p yr id yl-N-oxid e)-
eth yla m in o)-6-ch lor op yr a zin -2-on e-1-a ceta m id e (1). To a
100-L vessel, equipped with mechanical stirrer, thermocouple,
addition funnel, and nitrogen inlet was charged sequentially
THF (23 L), chloropyrazinone 39 (1965 g, 90.8 wt %, 4.59 mol),
and 1 M KOH (11.5 L). The slurry was stirred at 20 °C and it
became a clear solution. The progress of the reaction was
monitored by HPLC. When the starting material was <1 area
% relative to product (about 1 h), 2-fluorobenzylamine (776 g,
6.2 mol) was added. The solution was adjusted to pH 5-6 with
2 N HCl (∼5 L). Then, HOBt hydrate (250 g, 1.85 mol) and
finally EDC-HCl (1364 g, 7.12 mol) were added. The reaction
mixture was stirred at 18-22 °C. HPLC conditions: column,
Waters Symmetry Shield RP18, 4.6 × 150 mm, 3.5 µm; mobile
phase, CH3CN/0.1% aqueous H3PO4 20% to 28% over 3 min,
hold 2 min, to 70% over 8 min, hold 2 min; flow rate, 1.0 mL/
min; UV detection, 210 and 330 nm; temperature, 35 °C; tR
for hydroxy dione 41, 3.1 min, for amidine 3, 4.27 min, for
dimer 33, 8.84 min, for thio adduct 35, 9.86 min, for Cl-dimer
34, 10.17 min, for di-Cl pyrazinone 40, 11.18 min, for des-oxide
2c, 12.92 min. When the starting material was <1 area %
relative to product (2-3 h), 7 wt % aqueous NaHCO3 (46 L)
was added and the solution was stirred for 1 h and filtered.
The wet cake was washed with water (38.5 L) and cold EtOH
(8 L) and vacuum-dried under nitrogen to afford 2164 g of
compound 1. The purity was 95.4 area % and 93.7 wt % by
HPLC and the corrected yield was 94%. HPLC conditions:
column, Waters Symmetry Shield RP18, 4.6 × 150 mm, 3.5
µm; mobile phase, CH3CN/0.1% aqueous H3PO4 20% to 28%
over 3 min, hold 8 min, to 70% over 15 min; flow rate, 1.0 mL/
min; UV detection, 210 and 330 nm; temperature, 35 °C; tR
for des-Cl 3a , 4.27 min, for des-F of 1, 12.6 min, for 1, 14.15
min, for; di-Cl 40a , 22.04 min, for mono-Cl dimmer 34, 22.49
min, for des-oxide 2b, 25.1 min.
To a round-bottom flask equipped with a thermometer and
nitrogen inlet/outlet was charged water (240 mL), EDTA‚4Na‚
3.6H2O (2.4 g, 5.39 mmol) and NaCl (9.7 g). The resulting pH
10.3 solution was adjusted to pH 7.4 with 12 N HCl (0.48 mL).
To this solution was added crude 3 (20 g, 94.8 wt %, 53.5 mmol,
containing 7800 ppm Zn) and acetonitrile (300 mL) to give a
homogeneous solution. The solution was concentrated under
reduced pressure (50-70 Torr, 27 °C water bath) to about 230
mL with acetonitrile content at e5 vol %. The resulting slurry
was filtered and washed with water (3 × 40 mL). The wet cake
was vacuum dried at 22 °C under nitrogen to afford 18.2 g of
3 as an off-white solid. The purity was 99.9 area % and 99.3
wt % (18.1 g, 95% yield) by HPLC. The zinc content was 4
ppm by elemental analysis. 2-Thiolpyridine was not detected.
1
3: mp 160-162 °C; H NMR (CDCl3) δ 8.25 (d, J ) 6.4 Hz,
1H), 7.60 (dd, J ) 7.6, 2.1 Hz, 1H), 7.33 (m, 1H), 7.26 (m, 1H),
6.76, (d, J ) 4.7 Hz, 1H), 6.41 (br t, J ) 6.9 Hz, NH), 6.37 (d,
J ) 4.7 Hz, 1H), 4.65 (dt, J ) 13.8, 6.9 Hz, 2H), 4.52 (s, 2H),
4.22 (q, J ) 7.2 Hz, 3H), 1.27 (t, J ) 7.1 Hz, 3H). 13C NMR
(CDCl3) δ 166.8, 151.4, 151.1, 143.3, 141.0, 127.4, 125.3, 125.2,
122.0, 117.5, 117.4, 62.0, 49.6, 42.9, 14.0. Anal. Calcd for
C
15H16F2N4O4: C, 50.85; H, 4.55; N, 15.81; F, 10.72. Found:
C, 50.69; H, 4.31; N, 15.81; F, 10.81.
Compound 1 was further purified by crystallization as the
di-HCl salt: Compound 1 (2084 kg, 93.7 wt %, 4.17 mol) in
glacial acetic acid (9.2 L) was warmed to 60-70 °C under
nitrogen to give a homogeneous solution. To this was added 1
N HCl in acetic acid (9.0 L, 9.0 mol) over 15 min, which
produced a crystalline precipitate. The slurry was cooled to
20 °C and filtered, washed with glacial acetic acid (2 × 4 L),
washed with isopropyl acetate (4 × 4 L), then dried with a
flow of nitrogen through the filter cake at 20 °C overnight to
give 2340 g of 1‚di-HCl salt. The dried cake (2310 g) was added
to water (50 L) and stirred for 12-20 h, during which time
the pH dropped to 1, and the slurry became thicker. The slurry
was filtered and washed with water (3 × 20 L) to give a filtrate
with pH >4. The solids were dried via nitrogen drawn through
the cake with vacuum to a constant weight at 20 °C to give
compound 1 free base (1960 g, 97 wt %, 97.4% recovery).
Compound 1 (1930 g) was combined with dry DMF (9.5 L) and
heated to 75 °C to dissolve. The solution was filtered (5 µm)
and cooled slowly. Methanol (21 L) was added after crystals
appeared at about 30 °C. After further cooling to 23 °C, water
(9.5 L) was added. The slurry was stirred at 18-23 °C
overnight, then filtered. The wet cake was washed with water
(40 L) and dried by drawing nitrogen through the cake with a
vacuum. The recovery yield of compound 1 free base was 96.5%
Eth yl 3-(2,2-Diflu or o-2-(2-pyr idyl-N-oxide)eth ylam in o)-
6-ch lor op yr a zin -2-on e-1-a ceta te (39). To a 50-L vessel,
equipped with mechanical stirrer, thermocouple, and con-
denser with nitrogen inlet, was charged acetonitrile (6.5 L, KF
e 100 µg) and amidine 3 (1973 g, 5.57 mol). The slurry was
heated to 55 °C and N-chlorosuccinimide (751.3 g, 5.63 mol)
in acetonitrile (4.3 L, KF e 100 µg) was added over 5 min.
The slurry was heated to 60 °C to dissolve, then cooled to 50
°C. The clear solution was stirred at 50 °C for 20 min, then
cooled to 35-40 °C and stirred at this temperature for 1 h.
When dichloropyrazinone 40 was between <0.4 and 0.25
HPLC area % relative to product (about 0.5-1 h), the reaction
mixture was cooled to 27 °C, and cold 10% aqueous NaCl
solution (24 L) was added. The mixture was cooled to 0 °C,
and more cold 10% aqueous NaCl solution (52.8 L) was added.
After the mixture was stirred for 2 h at 0 °C, the slurry was
filtered. The wet cake was washed with cold water (2 × 13 L)
and dried by drawing nitrogen through the cake with a vacuum
to afford 2070 g of chloropyrazinone 39. The purity was 90.8
wt % and the corrected yield was 86.8%. HPLC conditions:
column, Waters Symmetry Shield RP18, 4.6 × 150 mm, 3.5
µm; mobile phase, CH3CN/0.1% aqueous H3PO4 20% to 28%
over 3 min, hold 2 min, to 70% over 8 min, hold 2 min; flow
rate, 1.0 mL/min; UV detection, 210 and 330 nm; temperature,
35 °C; tR for hydroxy dione 41, 3.1 min, for amidine 3, 4.27
1
(1840 g, 98.5 wt %). 1: mp 224-225 °C; H NMR (DMSO-d6)
δ 8.75 (t, J ) 5.6 Hz, NH), 8.34 (d, J ) 6.4 Hz, 1H), 7.60 (dd,
J . Org. Chem, Vol. 68, No. 23, 2003 8845

Chung et al.
J ) 7.8, 1.8 Hz, 1H), 7.57-7.50 (m, 1H + NH), 7.39 (t, J ) 7.8
Hz, 1H), 7.31 (m, 2H), 7.16 (m, 2H), 6.79 (s, 1H), 4.73 (s, 2H),
4.46 (dt, J ) 13.4, 6.7 Hz, 2H), 4.33 (d, J ) 5.6 Hz, 2H); 13C
NMR (DMSO-d6) δ 166.0, 160.5 (d, J ) 244.1 Hz), 151.7, 149.9,
142.5 (t, J ) 28.5 Hz), 141.3, 129.9 (d, J ) 4.0 Hz), 129.5 (d,
J ) 8.0 Hz), 128.6, 125.9 (d, J ) 14.5 Hz), 125.8, 125.3 (t, J )
6.7 Hz), 124.8 (d, J ) 3.6 Hz), 119.6, 118.3, 118.0 (t, J ) 244.9
Hz), 115.5 (d, J ) 20.9 Hz), 47.7, 42.1 (t, J ) 28.0 Hz), 36.6 (d,
J ) 4.5 Hz). Anal. Calcd for C20H17ClF3N5O3: C, 51.35; H, 3.66;
Cl, 7.58; F, 12.18; N, 14.97. Found: C, 51.25; H, 3.40; Cl, 7.57;
F, 12.24; N, 14.87.
Com p ou n d 1 Meth a n esu lfon a te Sa lt. To a slurry of
compound 1 (475 g, 98.5 wt %, 1.0 mol) in glacial acetic acid
(3.61 L) was added methanesulfonic acid (78 mL, 1.2 mol).
After being stirred at 20 °C for 5 min, the mixture became a
homogeneous solution. The solution was heated to 80-85 °C,
then isopropyl acetate (7.22 L) was added over 1 h, while
keeping the internal temperature >70 °C. After cooling to 67
°C, the solution was seeded with 1‚MSA salt (0.6 g) and cooled
slowly to 20 °C over 6 h. After stirring overnight, isopropyl
acetate (7.22 L) was added to the slurry, stirred for additional
4 h, and then filtered. The wet cake was washed with isopropyl
acetate (7.22 L) and vacuum dried under nitrogen to afford
pure 1‚MSA salt (549 g, 96.5% yield): mp 192-194 °C; both
1H NMR and 13C NMR are essentially identical with those of
the free base in DMSO, except for the methanesulfonic acid
nitrogen to afford 1.49 g (87%) of 31 as a crystalline solid
suitable for single crystal X-ray. The acetonitrile solvate was
gradually replaced by water upon exposure to air. Mp 180-
1
182 °C; H NMR (499.87 MHz, CD3CN) δ 8.79 (ddd, J ) 5.2,
2.0, 0.8 Hz, 1H), 8.10 (td, J ) 8.0, 2.0 Hz, 1H), 7.89 (dt, J )
8.0, 1.2 Hz, 1H), 7.70 (ddd, J ) 7.6, 5.2, 1.2 Hz, 1H), 7.52 (d,
J ) 4.8 Hz, 1H), 7.38 (d, J ) 4.8 Hz, 1H), 4.68 (s, 2H), 4.21 (q,
J ) 7.2 Hz, 2H), 1.96 (s, CH3CN), 1.25 (t, J ) 7.2, 3H); 13C
NMR (125.69 MHz, CD3CN) δ 167.3 (C-8), 156.9 (C-6), 154.8
(C-1), 152.8 (C-2′), 151.5 (C-6′), 142.7 (C-4′), 132.4 (C-3), 130.8
(C-3′), 126.7 (C-5′), 122.2 (C-4), 63.2 (C-9), 51.9 (C-7), 14.4 (C-
10); 15N NMR (50.66 MHz, CD3CN, reference to external NH3,
1
shifts were measured from 2D H-15N HMBC) δ ∼277 (N-5),
∼260 (N-1′), ∼166 (N-2). Anal. Calcd for C13H13Cl2N3O3SZn‚
KCH3CN‚J H2O: C, 37.35; H, 3.43; Cl, 15.38; N, 11.14; S, 6.96;
Zn, 14.18. Found: C, 37.14; H, 3.42; Cl, 15.19; N, 11.03; S,
6.81; Zn, 14.42.
P r ep a r a tion of Zin c Com p lex 32. To a solution of 2,2-
difluoro-2-(1-oxido-2-pyridinyl)ethylamine (4) (7.8 g, 44.5 mmol)
and dry CH3CN (81 mL) was added ZnCl2 (5.5 g, 30.4 mmol)
in one portion without cooling. The temperature rose from 22
to 37 °C. After the mixture was stirred overnight at 22 °C,
the slurry was filtered and washed with CH3CN. The solid was
1
vacuum dried under nitrogen: mp 223-225 °C dec; H NMR
(499.87 MHz, CD3CN) δ 8.42 (br d, J ) 6.4 Hz, 1H), 7.81 (dd,
J ) 8.0, 2.0 Hz, 1H), 7.66 (td, J ) 6.4, 1.2 Hz, 1H), 7.61 (ddd,
J ) 8.0, 6.4, 2.0 Hz, 1H), 3.80 (t, J HF ) 14.5 Hz, 2H); 13C NMR
(125.69 MHz, CD3CN) δ 143.5 (t, J CF ) 29.5 Hz, C-2), 142.8
(C-6), 130.7 (C-4), 129.9 (C-5), 126.8 (t, J CF ) 7.4 Hz, C-3), 118.6
(t, J CF ) 244.9 Hz, C-7), 46.3 (t, J CF ) 28.9 Hz, C-8); 15N NMR
(50.66 MHz, CD3CN, reference to external NH3, shifts mea-
signals at δ 2.46 (s, 3H) and 39.7. Anal. Calcd for C21H21
-
ClF3N5O6S: C, 44.73; H, 3.75; Cl, 6.29; F, 10.11; N, 12.42; S,
5.69. Found: C, 44.59; H, 3.78; Cl, 6.37; F, 10.15; N, 12.42; S,
5.77.
Bis-2(1H)-P yr id in eth ion e Zin c Dich lor id e Com p lex
(28 or 30). Filtration of the reaction slurry of the coupling
reaction by method B, followed by washing with acetonitrile
and vacuum drying afforded 28 as a beige solid. Complex 28
was also prepared by combining 2-mercaptopyridine with
ZnCl2 (2:1) in ethanol. Mp 241-243 °C; 1H NMR (DMSO-d6) δ
13.50, (br s, 1H), 7.70 (m, 1H), 7.45 (m, 1H), 7.35 (m, 1H), 6.80
(m, 1H); 13NMR (DMSO-d6) δ 176.7, 138.6, 138.5, 133.2, 113.9;
see text for 15N NMR. Anal. Calcd for C10H10Cl2N2S2Zn: C,
33.49; H, 2.81; Cl, 19.77; N, 7.81; S, 17.88; Zn, 18.23. Found:
C, 33.51; H, 2.75; Cl, 19.51; N, 7.80; S, 17.66; Zn, 18.02.
Eth yl [2-Oxo-3-(p yr id in -2-ylth io)p yr a zin -1(2H)-yl]a c-
eta te Zin c Dich lor id e Com p lex (31). Combined thioimidate
27 (1.2 g, 4.0 mmol), ZnCl2 (0.6 g, 4.3 mmol), and acetonitrile
(7.2 m) were heated at 35 °C for 17 h. After the mixture was
cooled to 22 °C over 3 h, the solid was filtered and washed
with acetonitrile (5 mL). The wet cake was vacuum dried under
1
sured were from 2D H-15N HMBC) δ ∼281 (N-1), ∼13 (N-9).
Anal. Calcd for C7H8Cl2F2N2OZn: C, 27.08; H, 2.60; Cl, 22.84;
F, 12.24; N, 9.02; Zn, 21.06. Found: C, 27.44; H, 2.35; Cl, 22.61;
F, 12.26; N, 9.16; Zn, 21.15.
Ack n ow led gm en t. We thank Dr. F. Xu for supply-
ing bromopyrazinone 16.
Su p p or tin g In for m a tion Ava ila ble: Complete details
supporting the X-ray crystallographic determination for 28 and
31 including the CIF files, NMR data for 2-mercaptopyridine
and 28. This material is available free of charge via the
Internet at http://pubs.acs.org.
J O034835E
8846 J . Org. Chem., Vol. 68, No. 23, 2003