Nitric oxide reacts with methoxide

Article abstract of DOI:10.1021/jo7020423

(Chemical Equation Presented) Despite over a century of reports to the contrary, sodium methoxide has been found to react with nitric oxide (NO). The reaction, whose final organic product is sodium formate, is postulated to occur via an intermediate O-bound diazeniumdiolate [CH₃O-N(O)=NO ^-] that decomposes to formaldehyde and nitrous oxide. Sodium formate forms from the aldehyde via a Cannizzaro reaction. Carboxylate salts have similarly been obtained by exposing sodium benzylate and sodium neopentoxide to NO in dioxane solution. Accordingly, sodium trimethylsilanoate should be considered as a substitute for sodium methoxide as the base used to accomplish the replacement of active hydrogens by the diazeniumdiolate functional group via the Traube reaction.

Full text of DOI:10.1021/jo7020423

original German).3 While carbanion equivalents such as enam-
ines4 and amidines5 react with NO as neutral compounds, the
addition of NO to carbanions generated in the presence of
strongly basic methoxide ion6-10 is still an attractive route to
carbon-bound diazeniumdiolates to be tested as potential NO
prodrugs. The continuing use of sodium methoxide in these
reactions implies the expectation that it is inert toward NO (a
belief that has remained unchallenged for 75 years). However,
in contrast to the prior reports, sodium methoxide has now been
found to react quite readily with NO to generate intermediates
which eventually produce sodium formate as the observable end
product. A series of tests using alternative alkoxide bases has
led to the recommendation that sodium trimethylsilanoate should
be considered the reagent-of-choice for the conduct of the
Traube reaction.
Nitric Oxide Reacts with Methoxide
Frank DeRosa,† Larry K. Keefer,† and Joseph A. Hrabie*,‡
Chemistry Section, Laboratory of ComparatiVe Carcinogenesis,
National Cancer Institute at Frederick, Frederick, Maryland
21702, and Basic Research Program, SAIC-Frederick, Inc.,
National Cancer Institute at Frederick, Frederick, Maryland
21702
hrabie@ncifcrf.goV
ReceiVed September 17, 2007
A report of the formation of bisdiazeniumdiolate 1 during
the treatment of sodium ethoxide solutions with NO11 prompted
Traube to initially attribute the result to the presence of acetone
impurity in the ethanol supply,12 but eventually led others to
the conclusion3 that the ethoxylate reacted directly with NO to
form acetaldehyde, which subsequently underwent the Traube
reaction (Scheme 1). In a detailed study of a wide variety of
alkoxides, methoxide was reported to be the sole alcoholate that
showed no signs of reacting with NO.3 That study included
methanol and benzyl alcohol as the only two representatives of
substrates which contain no hydrogens in the â position.
â-Hydrogens ultimately become the acidic protons that are
replaced by the diazeniumdiolate group after the initial formation
of the carbonyl compound. The absence of â-hydrogens in
benzyl alcohol prevents the occurrence of the Traube reaction
and results in the production of a benzaldehyde/benzoic acid
mixture upon treatment of sodium benzylate with NO. These
products most likely form in the three-step process illustrated
in Scheme 2. The sequence is initiated (eq 2) by the formation
Despite over a century of reports to the contrary, sodium
methoxide has been found to react with nitric oxide (NO).
The reaction, whose final organic product is sodium formate,
is postulated to occur via an intermediate O-bound diazeni-
umdiolate [CH3O-N(O)dNO-] that decomposes to form-
aldehyde and nitrous oxide. Sodium formate forms from the
aldehyde via a Cannizzaro reaction. Carboxylate salts have
similarly been obtained by exposing sodium benzylate and
sodium neopentoxide to NO in dioxane solution. Accord-
ingly, sodium trimethylsilanoate should be considered as a
substitute for sodium methoxide as the base used to ac-
complish the replacement of active hydrogens by the
diazeniumdiolate functional group via the Traube reaction.
The reaction of nitric oxide (NO) with carbanions is one of
the most often-used methods for the preparation of carbon-bound
diazeniumdiolates (compounds also called nitrosohydroxyl-
amines).1 Sodium methoxide has remained the preferred base
for effecting the deprotonation of a variety of substrates to be
subjected to this “Traube reaction” (e.g., that of acetone shown
in eq 1) since Wilhelm Traube first introduced the procedure
of an unstable O-bound diazeniumdiolated intermediate (2) that
decomposes to generate benzaldehyde, a compound known to
undergo the Cannizzaro reaction to create benzoic acid. While
several O-alkylated diazeniumdiolated alcohols have been
(3) Wieland, H.; Kerr, F. N. Ber. Dtsch. Chem. Ges. 1930, 63, 570-
579.
(4) Hrabie, J. A.; Arnold, E. V.; Citro, M. L.; George, C.; Keefer, L. K.
J. Org. Chem. 2000, 65, 5745-5751.
(5) Hrabie, J. A.; Citro, M. L.; Chmurny, G. N.; Keefer, L. K. J. Org.
Chem. 2005, 70, 7647-7653.
(6) Bohle, D. S.; Imonigie, J. A. J. Org. Chem. 2000, 65, 5685-5692.
(7) Arulsamy, N.; Bohle, D. S. J. Am. Chem. Soc. 2001, 123, 10860-
10869.
over 100 years ago.2 Although Weiland and Kerr showed that
a variety of other alkoxide anions react with NO, ultimately
leading to the production of C-diazeniumdiolated aldehydes,
ketones, and carboxylic acid salts, they reported that methoxide
does not react “to the slightest degree” (translated from the
(8) Arnold, E. V.; Keefer, L. K.; Hrabie, J. A. Tetrahedron Lett. 2000,
41, 8421-8424.
(9) Arnold, E. V.; Citro, M. L.; Keefer, L. K.; Hrabie, J. A. Org. Lett.
2002, 4, 1323-1325.
(10) Arulsamy, N.; Bohle, D. S.; Korich, A. L.; Mondanaro, K. R.
Tetrahedron Lett. 2003, 44, 4267-4269.
† National Cancer Institute.
‡ SAIC-Frederick, Inc.
(11) MacDonald, G. W.; Masson, O. J. Chem. Soc., Trans. 1894, 65,
944-949.
(1) Hrabie, J. A.; Keefer, L. K. Chem. ReV. 2002, 102, 1135-1154.
(2) Traube, W. Justus Liebigs Ann. Chem. 1898, 300, 81-128.
(12) Traube, W. Ber. Dtsch. Chem. Ges. 1894, 27, 3291-3293.
10.1021/jo7020423 CCC: $40.75 © 2008 American Chemical Society
Published on Web 01/10/2008
J. Org. Chem. 2008, 73, 1139-1142
1139

SCHEME 1. The Reaction of Sodium Ethoxide with NO
The intermediate formation of benzaldehyde in the conversion
of sodium benzylate to sodium benzoate was verified by
conducting the reaction using only the parent alcohol as the
solvent. The 2,4-dinitrophenylhydrazone of benzaldehyde could
be isolated in 20% yield (based on sodium benzylate) when
this reaction was interrupted after 4 h. Attempts to trap the
intermediate O-bound diazeniumdiolate in both the methoxide/
NO and benzylate/NO reactions via alkylation with either
dimethyl sulfate or methyl iodide proved unsuccessful. However,
support for the first step in Scheme 2 comes from the fact that,
under the anhydrous conditions employed in this study, the use
of alkoxide bases in these reactions would lead to the formation
of esters in the Cannizzaro reaction (since there would be no
hydroxide ion to attack the aldehydic carbonyl in the first step)
rather than the acid salts which are isolated were it not for the
generation of sodium hydroxide in this step. Benzaldehyde has
previously been shown to be a stable end product of the reaction
of Schiff bases with NO.15 No further oxidation to benzoic acid
was observed in those reactions. We were unable to verify the
formation of nitrous oxide gas as a byproduct due to the
construction of our NO reactor, which does not allow for
convenient testing of the gases produced. The pressure decrease
during the course of these reactions does correspond roughly
to a net uptake of 2 mol of gas as shown in eq 3. Elemental
analysis as well as traditional wet analytical testing (see the
Experimental Section) revealed the presence of small yields of
sodium nitrite (1.25%) and sodium nitrate (0.39%) in one sample
of sodium formate removed directly from the reaction conducted
in dioxane/methanol. Since further testing revealed considerable
variation in these numbers, it is assumed that these impurities
arise from the presence of traces of oxygen in the reactions upon
addition of NO. It should be noted that these inorganic salts
are not present in quantities that could be attributable to any
significant alternate mechanistic pathway for the reaction.
SCHEME 2. The Reaction of Sodium Benzylate with NO
reported,1 these were not prepared directly from the parent
alcohols and no alkylating agents are present in this reaction.
Among the O-bound diazeniumdiolates, the only frequently
studied example remains the inorganic compound known as
Angeli’s salt [Na+O--N(O)dNO-Na+]. Interestingly, Angeli’s
salt has been shown to convert 1-naphthaldehyde to 1-naphthoic
acid under aqueous conditions,13 so the Cannizzaro reaction may
not be the only source of the benzoic acid observed above. While
2 must undergo an isomerization from the cis diazeniumdiolate
shown here to the trans isomer to enable its fragmentation to
the aldehyde via a cyclic transition state, formation of these
types of ring systems is well-known.1 The previous studies of
NO/alkoxide reactions seem to indicate that, although the final
products vary, some type of reaction occurs in every instance.
They also suggest no apparent reason for the reported and
heavily relied upon stability of methanol under these reaction
conditions.
We have on many occasions been troubled by the appearance
of a seemingly unexplainable singlet at 8.46 ppm in the proton
NMR spectra of both carbon- and nitrogen-bound diazenium-
diolate preparations that employed the assistance of sodium
methoxide. After realizing that this impurity was sodium
formate, the product of the Cannizzaro reaction of formalde-
hyde14 produced in a side reaction of the methoxide, we sought
reaction conditions that would finally prove the existence of
this reaction.
Other mechanisms for the oxidation of the alcoholates in the
first step of this reaction are possible and the data at hand do
not allow exclusion of these pathways. In these pressurized
systems NO may be reacting either as the monomer or dimer.
Alkoxides are capable of acting as single electron donors16 so
the formation of the aldehydes via the corresponding alkyl
nitrites and/or alkoxide radicals cannot be excluded.17 It is likely
that different pathways prevail for each alkoxide and the present
study was not designed to resolve all these issues.
Although a solution of methoxide in methanol forms sodium
formate crystals slowly on stirring under a blanket of NO at 4
atm (15% yield after 48 h), the use of a dioxane/methanol
solution (5:1) affords a nearly complete reaction in that same
amount of time. Assuming that the processes described above
are the only ones occurring in this solution (the present work
does not prove this!) then the overall reaction of methanol with
NO under the influence of strong base may be written as shown
in eq 3. Dioxane has not been used previously as a solvent for
This newly observed reaction of NO with methoxide repre-
sents a potential complication to be overcome in the preparation
of any diazeniumdiolate requiring the use of added base. In
addition to the carbon-bound materials described above, sodium
methoxide has been enlisted to assist in the formation of some
of the less readily prepared nitrogen-bound diazeniumdiolates.
It has been used in the preparation of diazeniumdiolates of amino
acids such as proline18 and a variety of other obstinate
nucleophiles,19 and in the treatment of polymeric materials with
NO reactions but seems to provide the ideal combination of
NO solubility and product insolubility often required to drive
these reactions to completion. Sodium benzylate also reacts
rapidly with NO in dioxane to afford sodium benzoate in
reasonable yield. Neopentyl alcohol, a third example containing
no â-hydrogens, is converted to sodium pivalate by this process.
(15) Hrabie, J. A.; Srinivasan, A.; George, C.; Keefer, L. K. Tetrahedron
Lett. 1998, 39, 5933-5936.
(16) Ashby, E. C.; Coleman, D.; Gamasa, M. J. Org. Chem. 1987, 52,
4079-4085.
(17) He, Y.; Sanders, W. A.; Lin, M. C. J. Phys. Chem. 1988, 92, 5474-
5481.
(18) Saavedra, J. E.; Southan, G. J.; Davies, K. M.; Lundell, A.; Markou,
C.; Hanson, S. R.; Adrie, C.; Hurford, W. E.; Zapol, W. M.; Keefer, L. K.
J. Med. Chem. 1996, 39, 4361-4365.
(19) Saavedra, J. E.; Booth, M. N.; Hrabie, J. A.; Davies, K. M.; Keefer,
L. K. J. Org. Chem. 1999, 64, 5124-5131.
(13) Torun, L.; Mohammad, T.; Morrison, H. Tetrahedron Lett. 1999,
40, 5279-5282.
(14) Martin, R. J. L. Aust. J. Chem. 1954, 7, 335-347.
1140 J. Org. Chem., Vol. 73, No. 3, 2008

NO.20 Generation of sodium formate might be expected to
compete with the formation of the desired products in any or
all of these reactions, where diazeniumdiolate formation is either
slow or readily reversible. Thus, in addition to avoiding the
presence of â-hydrogens to prevent the Traube reaction, it is
the avoidance of R-hydrogens that should be paramount in the
selection of alcohols as solvents for these reactions.
One possible solution to this problem is the use of sodium
trimethylsilanoate as a replacement for sodium methoxide in
sluggish diazeniumdiolation reactions. This material offers a
higher solubility in organic solvents than sodium tert-butoxide
and it is somewhat less basic than the standard alkoxides.21 The
lower basicity is advantageous in those cases requiring the
presence of methanol cosolvent because it prevents the depro-
tonation (and subsequent involvement) of the solvent in the
reaction. In fact, we have found that a solution of sodium
trimethylsilanoate in methanol only very slowly develops a
cloudy appearance due to the possible precipitation of sodium
formate after being stirred under NO for several days, and no
isolable quantity is produced.
The presence of an extra salt such as sodium formate may
significantly influence the release of NO from diazeniumdiolated
materials, and we have previously noted the detrimental
consequences of the presence of even carbonate salts.22 Since
many diazeniumdiolates are not sufficiently stable to be purified
by recrystallization,22 the present results favor the use of sodium
trimethylsilanoate as an alternative to sodium methoxide in the
Traube reaction and in the diazeniumdiolation of other nucleo-
philes.
to maintain a pressure of 50-80 psig. After 47 h, copious amounts
of precipitate had formed. The pressure was released and the
solution was purged with argon. The precipitate was filtered, rinsed
with dioxane followed by diethyl ether, and dried in vacuo resulting
1
in a white powder. Yield: 3.10 g (98%). H NMR (D2O) δ 8.46
(s); 13C NMR (D2O) δ 173.9. Addition of a small quantity of a
standard sodium formate sample to either of these NMR solutions
did not result in the appearance of any new signals. Anal. Calcd
for C1H1O2Na1: C, 17.66; H, 1.48. Found: C, 17.29; H, 1.63; N,
3.08.
A sample of this product was analyzed for nitrite using the
standard Griess assay, and then for total nitrite/nitrate via the
chemiluminescence signal generated on complete reduction with
vanadium(III) chloride followed by reaction of the NO produced
with ozone in a Sievers 280i NO Analyzer. In an average of two
determinations, it was found to contain 1.79% total nitrite/nitrate
and 1.29% nitrite. These represent yields (based on NaOMe) of
1.25% for NaNO2 and 0.39% for NaNO3.
Sodium tert-Pentanoate (Sodium Pivalate) from Reaction of
Sodium Neopentoxide with NO. A stirring solution of neopentyl
alcohol (7.0 g, 79 mmol) in anhydrous dioxane (50 mL) was treated
with sodium hydride (0.80 g, 33 mmol) in small portions under an
inert atmosphere. As the solution became foamy, aliquots of
anhydrous dioxane were added bringing the total volume to
approximately 300 mL. Upon completion, the solution was gently
heated for 15 min. The resulting light yellow solution was filtered,
poured into a 500-mL glass Parr hydrogenation bottle, and purged
with argon. The solution was then placed under approximately 80
psi of NO gas and stirred at ambient temperature for 24 h. After
approximately 3 h, a white precipitate formed. After 24 h, the
pressure was released and the solution was purged with argon. The
precipitate was filtered, rinsed with dioxane followed by diethyl
ether, and dried in vacuo resulting in a white powder. Yield: 1.08
g (26%). 1H NMR (D2O) δ 1.11(s); 13C NMR (D2O) δ 30.3 (3C),
42.6, 191.7. A sample of sodium pivalate was prepared from the
free acid using sodium hydroxide; a portion was added to each of
these NMR solutions, and no additional signals were observed.
Experimental Section
General Information. NO was purchased in UHP grade and
allowed to stand in a ballast tank at about 5 atm pressure over
potassium hydroxide pellets for several hours before use. The NO
pressure apparatus employed has been described previously.22 The
glass bottles employed should be pressure tested and used behind
a shield with caution. NMR spectra were recorded at 400 MHz for
proton and 100 MHz for carbon.
Sodium Formate from Reaction of Sodium Methoxide with
NO (in Methanol). A 25% NaOMe/MeOH solution (60 mL, 0.26
mol) was placed in a 250-mL glass Parr hydrogenation bottle and
purged with argon. The solution was then placed under ap-
proximately 80 psi of NO gas and stirred at ambient temperature
for 48 h. After approximately 24 h, a white precipitate formed.
After 48 h, the pressure was released and the solution was purged
with argon. The precipitate was filtered, rinsed with methanol
followed by diethyl ether, and dried in vacuo resulting in a white
powder. Yield: 1.37 g (7.7%). 1H NMR (D2O) δ 8.46 (s); 13C NMR
(D2O) δ 173.9.
Sodium Formate from Reaction of Sodium Methoxide with
NO (in Dioxane/Methanol). A mixture (containing some suspended
solid NaOMe as evidenced by its cloudy appearance) of dioxane
(250 mL) and 25% NaOMe/MeOH solution (10.0 g, 46.3 mmol)
was prepared in a 500-mL glass Parr hydrogenation bottle and
treated in an analogous manner as above. The formation of
additional white solid became apparent within approximately 1 h.
During the course of 2 d, the pressure dropped substantially and
additional NO was added to the apparatus occasionally as required
Sodium Benzoate from Reaction of Sodium Benzylate with
NO. A stirring solution of benzyl alcohol (4.0 g, 37.0 mmol) in
anhydrous dioxane (50 mL) was treated with sodium hydride (0.445
g, 18.5 mmol) in small portions under an inert atmosphere. As the
solution became foamy, aliquots of anhydrous dioxane were added
bringing the total volume to approximately 150 mL. Upon comple-
tion, the solution was gently heated for 30 min. The resulting light
yellow solution was filtered, poured into a 250-mL glass Parr
hydrogenation bottle, and purged with argon. The solution was then
placed under approximately 80 psi of NO gas and stirred at ambient
temperature for 24 h. After approximately 10 min, a white
precipitate formed. After 24 h, the pressure was released and the
solution was purged with argon. The precipitate was filtered, rinsed
with dioxane followed by diethyl ether, and dried in vacuo resulting
1
in a white powder. Yield: 1.15 g (43%). H NMR (CD3OD) δ
7.33-7.39 (m, 3H), 7.86-7.89 (m, 2H); 13C NMR (CD3OD) δ
128.3 (2C), 129.8 (2C), 130.4, 139.5, 172.0. These NMR signals
were identical with those observed using a solution of commercial
sodium benzoate.
Benzaldehyde 2,4-Dinitrophenylhydrazone. A solution of 2,4-
dinitrophenylhydrazine (0.37 g, 1.9 mmol) in methanol (40 mL)
was treated with 6 M HCl (1.0 mL) and warmed to ensure
dissolution. Separately, the reaction of sodium benzylate with nitric
oxide (described above) was performed with slight variation. Benzyl
alcohol (1.0 g, 9.25 mmol) was directly treated in a vial with sodium
hydride (47 mg, 1.88 mmol), placed into a 250-mL glass Parr
hydrogenation bottle, and purged with argon. The solution was then
placed under approximately 70 psi of NO gas and stirred at ambient
temperature. The reaction was quenched after 4 h via addition of
6 M HCl (0.5 mL). At this time, the previously prepared
2,4-dinitrophenylhydrazine solution was brought to a gentle boil.
(20) Zhang, H.; Annich, G. M.; Miskulin, J.; Stankiewicz, K.; Oster-
holzer, K.; Merz, S. I.; Bartlett, R. H.; Meyerhoff, M. E. J. Am. Chem.
Soc. 2003, 125, 5015-5024.
(21) West, R.; Baney, R. H. J. Am. Chem. Soc. 1959, 81, 6145-6148.
(22) Hrabie, J. A.; Klose, J. R.; Wink, D. A.; Keefer, L. K. J. Org. Chem.
1993, 58, 1472-1476.
J. Org. Chem, Vol. 73, No. 3, 2008 1141

The biphasic mixture that resulted upon quenching of the reaction
was then added to the boiling solution. A bright orange precipitate
formed immediately and the hot solution was stirred for ap-
proximately 1 min. Upon cooling to room temperature, the mixture
was filtered, rinsed with the minimum amount of cold water, and
dried in vacuo resulting in a bright orange fluffy precipitate.
Yield: 0.11 g (20%). 1H NMR (DMSO-d6) δ 7.48-7.50 (m, 3H),
7.77-7.81 (m, 2H), 8.10 (d, J ) 9.6 Hz, 1H), 8.37 (dd, J ) 9.6
Hz, J ) 2.7 Hz, 1H), 8.70 (s, 1H), 8.86 (d, J ) 2.7 Hz, 1H), 11.67
(s, 1H); 13C NMR (DMSO-d6) δ 116.8, 123.0, 127.3 (2C), 128.9
(2C), 129.5, 129.7, 130.5, 133.8, 137.0, 144.5, 149.4; UV (CH3-
CN): λmax (ꢀ) 380 nm (27.0 mM-1 cm-1) [lit.23 UV (95% EtOH):
λmax (ꢀ) 378 nm (29.2 mM-1 cm-1)]. Mp 247-248 °C (lit.23 mp
238-239 °C).
Acknowledgment. We thank Mr. Michael L. Citro for
conducting the nitrite/nitrate and Griess assays. This project has
been funded in part with federal funds from the National Cancer
Institute, National Institutes of Health, under Contract No. N01-
CO-12400, and by the Intramural Research Program of the NIH,
National Cancer Institute, Center for Cancer Research.
Supporting Information Available: Copies of the NMR spectra
for each compound. This material is available free of charge via
the Internet at http://pubs.acs.org.
(23) Roberts, J. D.; Green, C. J. Am. Chem. Soc. 1946, 68, 214-216.
JO7020423
1142 J. Org. Chem., Vol. 73, No. 3, 2008