Reductive decyanation of pyrazinecarbonitriles

Article abstract of DOI:10.1016/S0040-4039(02)01499-5

Pyrazinecarbonitriles can be decyanated by hydrogenation with platinum on carbon in the presence of activated carbon under acidic conditions. Pyrazine carbonitrile-N-oxides undergo a stepwise reduction to the deoxy-pyrazinecarbonitriles followed by decyan

Full text of DOI:10.1016/S0040-4039(02)01499-5

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 6747–6750
Reductive decyanation of pyrazinecarbonitriles
Jennifer Albaneze-Walker,* Matthew Zhao, Melinda D. Baker, Peter G. Dormer and
James McNamara
Department of Process Research, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065, USA
Received 3 June 2002; revised 15 July 2002; accepted 22 July 2002
Abstract—Pyrazinecarbonitriles can be decyanated by hydrogenation with platinum on carbon in the presence of activated carbon
under acidic conditions. Pyrazine carbonitrile-N-oxides undergo a stepwise reduction to the deoxy-pyrazinecarbonitriles followed
by decyanation to give pyrazines in good yields. © 2002 Published by Elsevier Science Ltd.
Pyrazine derivatives are naturally occurring and impor-
tant pharmaceutical intermediates, which have been
used in the synthesis of pterins, pteridines, quinoxa-
lines, nucleosides, polymers, antioxidants, pesticides,
antiviral and cytotoxic agents.1
tophenone in methanol occurred at room temperature
(Scheme 1).3 The product crystallized out of solution as
the reaction progressed. Isolation by filtration provided
the desired product 2 in 92% yield and >99% purity.
While there are many methods to reduce N-oxides,
decyanation reactions are relatively rare.4 We did not
pursue the transformation of the nitrile into the car-
boxylic acid, as it is well known that decarboxylation of
pyrazinecarboxylic acids can be very difficult and intol-
erant of many functional groups.5
We recently required kilogram quantities of 2-amino-5-
phenylpyrazine 1, and so we sought an efficient and
economical synthesis of this intermediate. Starting from
readily available, inexpensive starting materials we have
developed an operationally simple process that relies on
formation of the pyrazine by a Taylor condensation1a
followed by a double reduction (Scheme 1).
Several reductants were screened for their ability to
both reduce the N-oxide and remove the nitrile under
mild conditions. Initially sodium dithionite was used to
effect the double reduction (Scheme 2).6 Despite exten-
sive optimization, the reaction was not entirely reliable.
Also, the volume productivity was relatively low and
not amenable to scale-up. Titanium trichloride is
known to effect decyanation of 4-cyanopyridine,7 and
did indeed cause decyanation of 2 to give 1 in 73%
yield, but with several side products. Four equivalents
of samarium diiodide effected a much cleaner decyana-
tion reaction to produce 1 in 81% yield. While the price
of samarium has decreased in recent years, it is still too
expensive for scale-up.
Scheme 1.
The Taylor condensation is an attractive method to
generate 2-amino-pyrazine-N-oxides with varying func-
tionality in the five and six positions. The conditions
are very mild and the yields are typically very good. In
our hands, the condensation reaction of aminoma-
lononitrile methanesulfonate salt2 and isonitrosoace-
Na2S2O4 85%
N
Ph
NC
N
Ph
TiCl3 73%
H2N
N
H2N
N
O
2
1
SmI2 81%
* Corresponding author. Tel.: (732) 594-8468; fax: (732) 594-1499;
Scheme 2.
e-mail: jennifer albanezewalker@merck.com
–
0040-4039/02/$ - see front matter © 2002 Published by Elsevier Science Ltd.
PII: S0040-4039(02)01499-5

6748
J. Albaneze-Walker et al. / Tetrahedron Letters 43 (2002) 6747–6750
Hydrogenation on Pt/C with 20% catalyst loading
returned a mixture of 1 plus 2 plus the deoxy-inter-
mediate 3 (Scheme 3). Increasing the catalyst load to
100% returned the doubly reduced 1 in 90% yield.
We reasoned that prevention of catalyst poisoning
would allow the reaction to proceed to completion
with a reasonable catalyst loading. Experimentation
with various additives such as strongly cyanide-chelat-
ing transition metal salts (e.g. ZnCl2) failed to prevent
poisoning of the catalyst, and in fact slowed the reac-
the C-3 position of the pyrazine ring, replacing the
cyano group, and in no other positions (Fig. 1).
D
N
N
Ph
H2N
Figure 1.
,
tion down appreciably. The addition of 3 A sieves
made little difference to the reaction, but the addition
of activated carbon Norit RO allowed the reaction to
progress to 70% completion with 25% catalyst load.
Screening of activated carbon products revealed that
Darco KB-B gave the highest and most consistent
yields for this reaction and allowed us to decrease the
platinum catalyst load to only 10%.
To explore the generality of this mild decyanation
reaction, several pyrazine substrates were synthesized
and subjected to the optimized conditions (Table 1).9
Compound 2 was N-acylated and subjected to the
decyanation conditions: acylation of the amine (entry
2) does not hinder the reaction. The amino group is
not necessary to direct the reaction: the parent com-
pound cyanopyrazine (entry 4) was transformed into
pyrazine in 78% yield. The camphor-derived pyrazine-
N-oxide (entry 6) did not eliminate cyanide under our
optimized conditions—or at elevated temperature and
pressure and 30% catalyst loading. Only the deoxy-
genated product was isolated. It may be that this
substrate is unable to adhere to the catalyst due to
the steric hindrance of the camphor methyl groups.
The decyanation is specific to pyrazines: 2-cyanopy-
rimidine and cyanopyridines are inert (Fig. 2). A
dicyanoimidazole was also subjected to these condi-
tions, and found to be inert as well.
Scheme 3.
The addition of acids was also examined. Addition of
sulfuric acid led to over-reduction of the pyrazine
ring. Addition of glacial acetic acid enhanced the rate
of reaction without over-reduction problems. Decya-
nation is much slower with Pd/C catalyst compared
to Pt/C, and deoxy-product 3 was isolated as the
major product. No reaction occurred in the absence
of hydrogen gas, even at elevated temperatures, which
tends to rule out an organometallic insertion process.8
Solvent, temperature and pressure were also opti-
mized (Scheme 4).
Figure 2.
In summary, we have developed a mild and robust
protocol for the deoxygenation/decyanation of pyrazi-
nes. We have demonstrated that the decyanation is a
general reaction for pyrazines, and a selective reaction
over other heterocycles.
Deuterium labeling studies were performed. In all
experiments, deuterium incorporation was found at
H2 40psi
N
Ph
NC
N
Ph
5%Pt/C 10% load
Darco KB-B 75wt%
AcOH 1.2 eq
Acknowledgements
H2N
N
H2N
N
O
2
1
MeOH 40ºC
90%
We thank Andrew Newell, Charles Bazaral and Tony
Houck for their expertise with high pressure equip-
ment. We thank Tony King, Cecile Savarin and
Spencer Dreher for helpful suggestions.
Scheme 4.

J. Albaneze-Walker et al. / Tetrahedron Letters 43 (2002) 6747–6750
6749
Table 1.
Entry
1
R-CN
Product
Yield
92%
NC
N
Ph
Ph
N
Ph
Ph
H2N
NC
N
O
H2N
N
N
N
N
2
3
87%
73%
AcN
H
N
O
AcN
H
NC
N
Ph
N
N
Ph
H2N
N
O
CH3
H2N
CH3
NC
N
N
N
N
4
5
78%
NC
N
N
N
OAc
OAc
99%
H2N
N
O
H2N
O
N
NH2
CN
90%a
N
6
7
NH2
N
N
N
CN
NC
H2N
OMe
N
N
OMe
89%
N
O
H2N
a Yield of deoxy product - no decyanation occurred.
2. For preparation of aminomalononitrile methanesulfonate
salt, see: Mettler, H. US 482015, our improved procedure
follows: (1) A solution of NaNO2 (2.17 kg in 5.0 L H2O)
was prepared. A slurry of malononitrile (1) (1.98 kg, 30.0
mol) in water (5.0 L) was cooled to 15°C, then NaNO2
solution (0.8 L, ꢀ15% of the total) was added followed by
H2SO4 (5N, 0.60 L, 10 mol%). The pH of the mixture
should be 3.0–3.5. The remaining NaNO2 solution was
added slowly (1–2 h) with ice-water bath to control the
batch at 15–20°C. The mixture was then aged at 15–20°C
for 1 h. Final pH should be 4.0–4.2. Sulfuric acid (5N, 5.70
L) was added to acidify the reaction mixture to pH ꢀ2.0
at 15–20°C. The oxime is unstable in acidic aqueous
solution and should be kept <20°C. The mixture was
extracted with EtOAc (15 L, then 5 L). The combined
batch was washed with brine (7.5 L) then submitted to
hydrogenation. (2) HOAc (7.5 L) was added to the crude
solution of oxime (ꢀ23 L). Pt/C (5%, 0.57 kg) was added
and the mixture was hydrogenated at 10 bar H2 for 20 h at
10–15°C. The mixture was filtered through a Solka-Floc
References
1. (a) Taylor, E. C.; Perlman, K. L.; Sword, I. P.; Sequin-
Frey, M.; Jacobi, P. A. J. Am. Chem. Soc. 1973, 95,
6407–6412; (b) Taylor, E. C.; Perlman, K. L.; Kim, Y.-H.;
Sword, I. P.; Jacobi, P. A. J. Am. Chem. Soc. 1973, 95,
6413–6418; (c) Liu, W.; Wise, D. S.; Townsend, L. B. J.
Org. Chem. 2001, 66, 4783–4786; (d) Zhang, C. Y.; Tour,
J. M. J. Am. Chem. Soc. 1999, 121, 8783–8790; (e) Singh,
S. B.; Tomassini, J. E. J. Org. Chem. 2001, 66, 5504–5516;
(f) Ramadas, K.; Janarthaana, N.; Rani, D. M. Org Proc.
Prep. Int. 1998, 4, 469–471; (g) Ple, N.; Turck, A.; Cou-
ture, K.; Queguiner, G. Synthesis 1996, 838–842; (h) Lee,
S.; Fuchs, P. L. Org. Lett. 2002, 4, 313–316; (i) Lee, S.;
Fuchs, P. L. Org. Lett. 2002, 4, 317–318; (j) Cavalier,
J.-F.; Burton, M.; De Tollenaere, C.; Dussart, F.;
Marchand, C.; Rees, J.-F.; Marchand-Brynaert, J. Synthe-
sis 2001, 768–772; (k) Brunner, H.; Rosenboem, S. Monat.
Chem. 2000, 131, 1371–1382.

6750
J. Albaneze-Walker et al. / Tetrahedron Letters 43 (2002) 6747–6750
pad and rinsed with 10 L of EtOAc. Methanesulfonic acid
dithionite 3 equiv. and sodium carbonate 1 equiv., the
reaction gave 70–75% isolated yield. Unfortunately, about
50 volumes are required.
(3.17 kg, 2.14 L) was added over 2 h with a dropping
funnel at 10–15°C. The mixture was aged for 1 h then
filtered to collect the product, which was dried under a
stream of nitrogen to constant weight to give the mesylate
salt as a white crystalline solid (4.78 kg, 90% yield).
3. To a solution of isonitrosoacetophenone (1.69 kg, 10 mol,
88% pure) in methanol (15 L) and water (0.4 L) was added
in portions the mesylate salt (1.95 kg) over 5 h at rt to give
a very thick slurry. The product crystallizes as the mesylate
salt is added and the reaction proceeds. The mixture was
aged overnight and then quenched with a solution of
Na2CO3 (0.87 kg in 8 L water). The pH of the quenched
mixture was 8–9. The product was collected by filtration
and washed with 1/1 MeOH/H2O (8 L) and dried by
passing air through the filter cake to constant weight, to
give the product as an fluffy bright yellow solid, 1.934 kg,
99 w%, 90% corrected yield.
7. Clerici, A.; Porta, O. Tetrahedron Lett. 1980, 21, 1675–
1676.
8. For a rare example of CꢀCN bond activation by a nickel
complex which promotes cross coupling, see: Miller, J. A.
Tetrahedron Lett. 2001, 42, 6991–6993.
9. Experimental for decyanation reaction: A slurry of pyra-
zine N-oxide 2 (2.00 g, 9.4 mmol), 5% Pt/C (0.30 g, 15%
catalyst loading), Darco KB-B (1.50 g, 75% loading) and
glacial acetic acid (0.650 mL, 11.3 mmol) in methanol (40
mL) was submitted for hydrogenation at 40 psi and 40°C
for 18 h in pressure apparatus. After hydrogen atmosphere
was flushed with nitrogen, a solution of sodium hydroxide
(0.450 g, 11.3 mmol) in 2 mL water was added. The
reaction mixture was filtered through Solka-Floc to
remove catalyst. Solka Floc/catalyst was rinsed with 1 L
methanol. Caution! the filtrate contains the majority of
cyanide as NaCN! The filtrate was concentrated to ꢀ5–6
mL by rotary evaporator in a fume hood. Water (10 mL)
was added. Aged for 1 h. The precipitate was collected by
filtration and air-dried overnight in hood to give pale
yellow solids 1.47 g, 91% yield.
4. Alshaar, A. H. M.; Chambers, R. K.; Gilmour, D. W.;
Lythgoe, D. J.; McClenaghan, I.; Ramsden, C. A. J.
Chem. Soc., Perkin Trans. 1 1992, 2789–2811.
5. Sarma, M. S. P.; Czarnick, A. W. Synthesis 1988, 72–73.
6. Sodium dithionite is unstable at elevated temperatures and
under acidic conditions. By using a mixture of sodium