New insights into the acid-promoted reaction of caffeic acid and its esters with nitrite: Decarboxylation drives chain nitrosation pathways toward novel oxime derivatives and oxidation/fragmentation products thereof

Article abstract of DOI:10.1021/jo015965v

In 0.05 M acetate buffer, pH 4, containing 1% methanol, caffeic acid (1a) (2 × 10^-3 M) reacted smoothly with nitrite (NO₂^-) (4 × 10^-3 M) to afford as main products the novel 2-hydroxy- and 2-methoxyaldoximes 7a,b, the 2-oxoaldoxime 9a, 3,4-dihydroxybenzoic acid, 3,4-dihydroxybenzaldehyde, and the known furoxan 3c and benzoxazinone 4b in smaller amounts. At lower 1a concentration (e.g., 1 × 10^-4 M), 7a was the main product, whereas with 0.1 M 1a and 0.5 M NO₂^- 3c and 9a were prevailing. At pH 2, 7a was still the most abundant product, together with 3,4-dihydroxybenzaldehyde and some 9a, whereas at pH 1 9a and 3,4-dihydroxybenzaldehyde were formed in higher yields. No evidence for ring nitration products, including the previously reported 4,5-dihydroxy-2-nitrobenzaldehyde, was obtained. At 2 × 10^-3 M concentration and at pH 4, caffeic acid methyl ester (1b) reacted with NO₂^- chiefly via ring nitration and/or dimerization to give 5a, the novel nitrated neolignan derivative 10, and the parent 6. Chlorogenic acid (1c) afforded only the ring nitrated derivative 5b. A unifying mechanism for the reaction of 1a and its esters with NO₂^- is proposed involving reversible formation of nitroso intermediates via chain nitrosation at the 2-position of the (E)-3-(3,4-dihydroxyphenyl)propenoic system. In the case of 1a, decarboxylation would drive the nitroso intermediates toward the formation of oximes 7a,b and 3c, reflecting nucleophilic addition of water, methanol, and NO₂^-, and their oxidation or breakdown products, viz. 9a, 3,4-dihydroxybenzaldehyde, 3,4-dihydroxybenzoic acid, and the benzoxazinone 4b. In the case of esters 1b,c, to which decarboxylation is precluded, ring nitration or dimerization become the favored routes, triggered by preliminary oxidation at the catechol moiety.

Full text of DOI:10.1021/jo015965v

J . Org. Chem. 2002, 67, 803-810
803
New In sigh ts in to th e Acid -P r om oted Rea ction of Ca ffeic Acid a n d
Its Ester s w ith Nitr ite: Deca r boxyla tion Dr ives Ch a in Nitr osa tion
P a th w a ys tow a r d Novel Oxim e Der iva tives a n d Oxid a tion /
F r a gm en ta tion P r od u cts Th er eof
Alessandra Napolitano and Marco d’Ischia*
Department of Organic Chemistry and Biochemistry, University of Naples “Federico II”,
Via Cinthia 4, I-80126 Naples, Italy
dischia@cds.unina.it
Received J uly 26, 2001
In 0.05 M acetate buffer, pH 4, containing 1% methanol, caffeic acid (1a ) (2 × 10-3 M) reacted
smoothly with nitrite (NO2-) (4 × 10-3 M) to afford as main products the novel 2-hydroxy- and
2-methoxyaldoximes 7a ,b, the 2-oxoaldoxime 9a , 3,4-dihydroxybenzoic acid, 3,4-dihydroxybenz-
aldehyde, and the known furoxan 3c and benzoxazinone 4b in smaller amounts. At lower 1a
concentration (e.g., 1 × 10-4 M), 7a was the main product, whereas with 0.1 M 1a and 0.5 M NO2
-
3c and 9a were prevailing. At pH 2, 7a was still the most abundant product, together with 3,4-
dihydroxybenzaldehyde and some 9a , whereas at pH 1 9a and 3,4-dihydroxybenzaldehyde were
formed in higher yields. No evidence for ring nitration products, including the previously reported
4,5-dihydroxy-2-nitrobenzaldehyde, was obtained. At 2 × 10-3 M concentration and at pH 4, caffeic
acid methyl ester (1b) reacted with NO2- chiefly via ring nitration and/or dimerization to give 5a ,
the novel nitrated neolignan derivative 10, and the parent 6. Chlorogenic acid (1c) afforded only
the ring nitrated derivative 5b. A unifying mechanism for the reaction of 1a and its esters with
-
NO2 is proposed involving reversible formation of nitroso intermediates via chain nitrosation at
the 2-position of the (E)-3-(3,4-dihydroxyphenyl)propenoic system. In the case of 1a , decarboxylation
would drive the nitroso intermediates toward the formation of oximes 7a ,b and 3c, reflecting
nucleophilic addition of water, methanol, and NO2-, and their oxidation or breakdown products,
viz. 9a , 3,4-dihydroxybenzaldehyde, 3,4-dihydroxybenzoic acid, and the benzoxazinone 4b. In the
case of esters 1b,c, to which decarboxylation is precluded, ring nitration or dimerization become
the favored routes, triggered by preliminary oxidation at the catechol moiety.
In tr od u ction
in part from their potent nitrite (NO2-) scavenging
properties, suggesting an inhibitory role in acid-depend-
ent mutagenic N-nitrosamine formation in the stomach.
In simulated gastric juice, 1a was reported to react
rapidly and completely with an equimolar quantity of
NO2-, blocking the elevation of serum N-nitrosodimethyl-
amine in rats receiving aminopyrine and NO2-.5 Com-
parative studies indicated that 1a was much more
effective than its esters, e.g., 1c,6 and other related
compounds, such as ferulic acid and p-coumaric acid7 as
Epidemiological1 and biomedical data2,3 that accrued
over the past decades consistently supported a chemo-
preventive role of caffeic acid ((E)-3-(3,4-dihydroxyphen-
yl)propenoic acid, 1a ) and other related polyphenolic
dietary constituents (e.g., chlorogenic acid, 1c) widely
found in vegetables, fruits, cereals, coffee, and honeybee
propolis,4 in the development of gastric and colorectal
tumors.
-
scavenger of NO2 and inhibitor of nitrosamine forma-
tion, yet the causes for such differences were not ad-
dressed at the chemical level.
Although the emphasis in all those studies was on the
beneficial consequences of the interaction of 1a with
(3) (a) Mahmoud, N. N.; Carothers, A. M.; Grunberger, D.; Bilinski,
R. T.; Churchill, M. R.; Martucci, C.; Newmark, H. L.; Bertagnolli, M.
M. Carcinogenesis 2000, 21, 921-927. (b) Michaluart, P.; Masferrer,
J . L.; Carothers, A. M.; Subbaramaiah, K.; Zweifel, B. S., Koboldt, C.;
Mestre, J . R. Grunberger, D.; Sacks, P. G., Tanabe, T.; Dannenberg,
A. J . Cancer Res. 1999, 59, 2347-2352.
(4) Clifford, M. N. J . Sci. Food Agric. 1999, 79, 362-372.
(5) Kuenzig, W.; Chau, J .; Norkus, E.; Holowaschenko, H.; New-
mark, H.; Mergens, W.; Conney, A. H. Carcinogenesis 1984, 5, 309-
313.
(6) Kono, Y.; Shibata, H.; Kodama, Y.; Sawa, Y. Biochem. J . 1995,
312, 947-953.
(7) (a) Stich, H. F. Prev. Med. 1992, 21, 377-384. (b) Rundloef, T.;
Olsson, E.; Wiernik, A.; Back, S.; Aune, M.; J ohansson, L.; Wahlberg,
I. J . Agric. Food Chem. 2000, 48, 4381-4388.
There is general consensus that the antimutagenic and
anticancer effects of 1a and its congeners stem at least
* To whom correspondence should be addressed. Phone: +39-081-
674132. Fax: +39-081-674393.
(1) (a) Hudson, E. A.; Dinh, P. A.; Kokubun, T.; Simmonds, M. S.
J .; Gescher, A. Cancer Epidemiol., Biomarkers Prev. 2000, 9, 1163-
1170. (b) Owen, R. W.; Giacosa, A.; Hull, W. E.; Haubner, R.;
Spiegelhalder, B.; Bartsch, H. Eur. J . Cancer 2000, 36, 1235-1247.
(2) (a) Shureiqi, I.; Chen, D.; Lee, J . J .; Yang, P.; Newman, R.;
Brenner, D. E.; Lotan, R.; Fischer, S. M.; Lippman, S. M. J . Natl.
Cancer Inst. 2000, 92, 1136-1142. (b) Karekar, V.; J oshi, S.; Shinde,
S. L. Mutat. Res. 2000, 468, 183-194.
10.1021/jo015965v CCC: $22.00 © 2002 American Chemical Society
Published on Web 01/10/2002

804 J . Org. Chem., Vol. 67, No. 3, 2002
Napolitano and d’Ischia
NO2-, scattered reports in the literature indicated that
1a and its derivatives may cause genotoxicity upon
nitrosation in the Ames tester strain TA1008 and may
enhance nitrosation and nitrosamine formation in the
gastric juice.9,10 In the latter studies, the balance of
stimulating versus suppressive effects on nitrosation
processes was found to be critically dependent on the
concentration of 1a and individual features of the gastric
juice.
Despite the unabated interest raised by the interaction
of 1a with NO2-, elucidation of the underlying chemistry
has been hurdled by the intrinsic complexity of the
reaction pathways and the elusive character of the
products. Comparative investigations11 of the reaction of
-
1a and 1c with NO2 in acetate buffer at pH 4 revealed
markedly different spectrophotometric courses, although
no explanation was offered. Indirect insights derived from
two related studies on the reactions of p-coumaric acid
and ferulic acid with NO2.12-14 These indicated substan-
tial conversion to decarboxylation and chain breakdown
products, such as benzaldehydes, 2-(4-hydroxyphenyl)-
oximinoacetaldehyde (2), furoxans (3a ,b), 2-methoxy-4,6-
dinitrophenol, and the unusual 7-hydroxy-6-methoxy-1,2-
(4H)benzoxazin-4-one (4a ).
With a view to filling this longlasting gap in the
chemistry of caffeic acid derivatives, we undertook a
detailed investigation into the main products formed by
reaction of 1a with NO2-, in comparison with the methyl
ester 1b and 1c, under mildly acidic conditions mimick-
ing those occurring in the stomach. When the study was
-
NO2 (2 molar equiv) in acetate buffer containing 1%
methanol at pH 4. Under such conditions, reversed-phase
HPLC analysis indicated a complex mixture of products,
most of which eluted faster than 1a ,while only two
displayed higher retention times. These latter proved to
be the furoxan 3c and the benzoxazinone 4b by straight-
forward 1H and 13C NMR analysis in comparison with
literature data.15 Of the faster moving products, two were
readily identified as 3,4-dihydroxybenzaldehyde and 3,4-
dihydroxybenzoic acid. The other two products, one of
which was the main component of the mixture, were
isolated by PTLC fractionation of the ethyl acetate
extractable fraction and were subjected to extensive
spectral analysis. The 1H NMR spectrum of the most
abundant product featured the resonances of a 3,4-
dihydroxyphenyl moiety and two additional doublets for
coupled protons at δ 7.36 and 5.06, exhibiting one-bond
connectivities in the HMQC spectrum with 13C NMR
signals at δ 153.9 and 73.2, respectively. A similar
pattern of resonances was displayed by the companion
product, with the additional feature of a 3H singlet at δ
3.29 for a methoxyl group.
being completed, a paper appeared reporting preliminary
- 15
data on the reactions of 1a -c with acidic NO2
.
At pH
2, 1a was shown to react at the propenoic acid chain to
give the furoxan 3c, the 1,2-(4H)-benzoxazin-4-one 4b,
and 4,5-dihydroxy-2-nitrobenzaldehyde, the latter denot-
ing ring nitration, whereas 1b and 1c were converted into
their nitro derivatives 5a and 5b, respectively, as the sole
products. Under more acidic conditions (pH 1), both 1a
and 1b were reported to give dark polymeric precipitates;
in the case of 1b, ethyl acetate extraction of the reaction
mixture led to the isolation of the dihydrobenzofuran
lignan 6, an antimitotic agent inhibiting tubulin polym-
erization.16
Because of certain inconsistencies and discrepancies
between these and our results, we were then prompted
to extend the study to a reexamination of the same
reactions under the conditions described by the previous
authors.15
On this basis, the products were formulated as the
novel oximes 7a ,b. In product 7b, the methoxyl group
derived evidently from the methanol added as cosolvent,
as confirmed in separate experiments carried out with
acetone in the place of methanol. The E configuration of
the aldoxime functionalities was deduced on the basis of
the chemical shifts of the relevant protons and carbons.17
Attempts to obtain satisfactory mass spectra were de-
feated under a variety of conditions. Accordingly, 7a ,b
were acetylated with Ac2O/pyridine to afford similar sets
Resu lts a n d Discu ssion
Acid -P r om oted Rea ction of 1a w ith NO2-. In a first
series of experiments 1a (2 × 10-3 M) was reacted with
(8) Duarte, M. P.; Laires, A.; Gaspar, J .; Oliveira, J . S.; Rueff, J .
Teratog., Carcinog. Mutagen. 2000, 20, 241-249.
(9) Ermilov, V. B.; Shendrikova, I. A.; Dikun, P. P. Vopr. Onkol.
1989, 35, 38-44.
1
of two related products each. Both H NMR analysis and
(10) Dikun, P. P.; Ermilov, V. B.; Shendrikova, I. A. IARC Sci. Publ.
1991, 105, 552-557.
EI-MS spectra consistently indicated for the products the
structures of the peracetylated derivatives 7c,d and of
the nitriles 8a ,b, the latter derived by the Ac2O-promoted
dehydration of the aldoxime functionalities.
Another product, yellow in color, could be obtained
along with 3c in much higher yields by similar reaction
(11) Kono, Y.; Shibata, H.; Kodama, Y.; Sawa, Y. Biochem. J . 1995,
312, 947-953.
(12) Kikugawa, K.; Hakamada, T.; Hasunuma, M.; Kurechi, T. J .
Agric. Food Chem. 1983, 31, 780-785.
(13) Rousseau, B.; Rosazza, J . P. N. J . Agric. Food Chem. 1998, 46,
3314-3317.
(14) Torres y Torres, J . L.; Rosazza, J . P. N. J . Agric. Food Chem.
2001, 49, 1486-1492
(15) Cotelle, P.; Vezin, H. Tetrahedron Lett. 2001, 42, 3303-3305.
(16) Pieters, L.; Van Dyck, S.; Gao, M.; Bai, R.; Hamel, E.; Vlietinck,
A.; Lemiere, G. J . Med. Chem. 1999, 42, 5475-5481.
(17) (a) Rykowski, A.; Guzik, E., Makosza, M.; Holzer, W. J .
Heterocyclic Chem. 1993, 30, 413. (b) Karabatsos, G. J .; Taller, R. A.
Tetrahedron 1968, 24, 3347-3360.

Reaction of Caffeic Acid and Its Esters with Nitrite
J . Org. Chem., Vol. 67, No. 3, 2002 805
of 1a at 0.1 M concentration with 0.5 M NO2- in acetate
buffer, pH 4/methanol 1:1 v/v. The compound shared with
7a ,b the proton spin system for a 4-substituted catechol
ring, although the H-2 and H-6 protons resonated rela-
tively downfield. It also exhibited a singlet at δ 7.91
correlating with a carbon signal at δ 147.8 and displaying
cross-peaks in the HMBC spectrum with carbon signals
at δ 187.1 (denoting a carbonyl group) and 129.2. On this
basis, the product was formulated as the novel 2-oxo-
aldoxime 9a . This conclusion was corroborated by the EI-
MS spectrum of the trimethyl derivative 9b, obtained by
reaction with diazomethane, which displayed a discern-
ible molecular ion peak at m/z 223 and a well-defined
pattern of fragmentation peaks at m/z 165 (M - CH3O
- HCN) and 137 (M - CH3O - HCN - CO) indicating
sequential losses of the elements of the O-methyl 2-ke-
toaldoxime chain.
rapidly to give mainly 7a , all other products being below
detection limits. At 0.1 M concentration, on the other
hand, 1a afforded 3c as the chief product, with some 9a .
When NO2- concentration was varied in the range 1-10
× 10-3 M keeping 1a at 2 × 10-3 M concentration, HPLC
analysis revealed consistent changes in substrate con-
sumption but only barely appreciable variations in
product distribution. Moreover, in all cases examined,
product distribution from 1a remained virtually un-
changed during the reaction course, for at least 4 h.
Changing the nature of the buffer (e.g., phosphate in
the place of acetate)11 or excluding oxygen from the
medium (purging with Ar) reduced only slightly the
extent of substrate consumption without, however, af-
fecting product distribution.
The reaction of 1a (2 × 10-3 M) with NO2 (4 × 10-3
-
M) was also carried out at pH 1 and 2, as reported.15
Under both conditions, 3,4-dihydroxybenzaldehyde and
9a were among the main products. At pH 2, oximes 7a ,b
were also formed. Neither 4,5-dihydroxy-2-nitrobenz-
aldehyde nor oximinoaldehyde products related to those
previously reported from ferulic acid12 were detected in
To assess whether the material balance could be
improved by detection of putative aromatic nitration
products of 1a escaping HPLC and TLC analysis, the
synthesis of 6-nitrocaffeic acid18 was pursued by nitration
of 1a with tetranitromethane in aqueous bicarbonate at
pH 8.0. Careful scrutiny of the reaction mixtures failed
the reaction mixtures from 1a .
Acid -P r om oted Rea ction of 1b a n d 1c w ith NO2
however to reveal detectable formation of this product
-
.
-
Reaction of 2 mM 1b with NO2- (2 molar equiv) in acetate
buffer at pH 4 resulted in the formation of two main
products, the most abundant of which (43% yield) proved
to be 5a . The other product (10% yield) was apparently
a dimer sharing several spectral features (1H,13C HMQC,
HMBC) in common with those of the dihydrobenzofuran
neolignan 6,16 along with some telltale differences. These
included the presence of four 1H singlets in the aromatic
by reaction of 1a with acidic NO2
.
Under the typical conditions used for 1a , both 3,4-
dihydroxybenzoic acid and 3,4-dihydroxybenzaldehyde
-
reacted slowly with NO2 to afford complex patterns of
products which did not figure among those obtained from
1a . Notably, no detectable formation of 4,5-dihydroxy-2-
nitrobenzaldehyde19 from 3,4-dihydroxybenzaldehyde was
observed under such conditions.
To provide a more complete inventory of the products
formed from 2 × 10-3 M 1a and 4 × 10-3 M NO2- at pH
4, and to rule out the possibility that the isolated products
could be artifactually generated during work up or
chromatographic separation, a typical reaction was con-
tinued until most of the starting material was consumed
(about 4 h), then the mixture was carefully neutralized
1
region of the H NMR spectrum, one of which appearing
relatively downfield (δ 7.67), and a pronounced downfield
shift of the H-2 proton on the dihydrobenzofuran moiety
(δ 6.65 vs. 5.97 for 6).16 Both H-2 and H-3 proton
resonances correlated in the HMBC spectrum with
quaternary carbon signals at δ 151.5 and 132.9, the latter
giving a cross-peak also with the proton signal at δ 7.67.
The typical pH-dependent nitrocatechol chromophore,
and the conversion to a triacetyl derivative by acetylation,
provided definitive evidence for the structure of the
nitrated dihydrobenzofuran lignan 10. Although two
different stereoisomers were possible, spectral analysis
suggested a single isomer, which was regarded as having
a 2,3-trans configuration, in accord with literature data16
and with preliminary molecular mechanics calculations
(MM+). Morever, no detectable nOe was observed be-
tween the protons on the 2- and 3-positions of the
benzofuran ring. Although the value for the vicinal
coupling constant for the relevant protons (J ) 4.4 Hz)
was smaller than that of the parent dihydrobenzofuran
neolignan 6 (J ) 8.3 Hz),16 it is possible that the nitro
group on the adjacent aromatic ring causes a significant
distortion of the dihydrofuran moiety. This influence of
the nitro group is apparent from the dramatic downfield
shift of the H-2 proton of the dihydrobenzofuran moiety.
Despite careful analysis, no apparent chain nitrosation/
breakdown product of 1b could be detected.
1
and extracted with ethyl acetate. The H NMR spectrum
of the whole extractable fraction displayed the charac-
teristic resonances of 7a as the main constituent, as well
as those of 7b (ca. 50% of 7a ), 3c (ca. 15% of 7a ), 3,4-
dihydroxybenzaldehyde (ca. 10% of 7a ) and of the ben-
zoxazinone 4b (ca. 10% of 7a ). Apparently, no dimeriza-
tion product of 1a related, e.g., to the neolignan 6, was
detected.
When 1b (2 × 10-3 M) was allowed to react with NO2
-
(2 molar equiv) at pH 1, formation of 5a was decreased
(ca. 21% yield) but not suppressed, at variance with the
In a subsequent series of experiments, the concentra-
tion of 1a was varied systematically maintaining NO2
at 2 molar equiv. At 100 µM concentration, 1a reacted
(18) Grenier, J .-L.; Cotelle, N.; Cotelle, P.; Catteau, J .-P. Bioorg.
Med. Chem. Lett. 1996, 6, 431-434.
(19) Murphy, B. P. J . Org. Chem. 1985, 50, 5873-5875.
-

806 J . Org. Chem., Vol. 67, No. 3, 2002
Napolitano and d’Ischia
Ta ble 1. Tota l Ch a r ge Den sities (TCD) a n d HOMO
Coefficien ts of 1a /b
HOMO
TCD
compd
carbon
AM1
PM3
AM1
PM3
1a
2
3
2′
5′
6′
2
0.328
0.139
0.019
0.169
0.362
0.331
0.146
0.013
0.169
0.363
0.315
0.137
0.018
0.170
0.362
0.318
0.144
0.012
0.171
0.363
-0.203
-0.021
-0.161
-0.195
-0.092
-0.199
-0.026
-0.162
-0.194
-0.093
-0.200
0.012
-0.140
-0.171
-0.074
-0.202
-0.005
-0.142
-0.170
-0.075
previous report.15 Compound 10 was likewise present in
detectable amounts (5%), along with 6 (8%). In fact, 6
was also detected by HPLC as a minor component in the
early stages of the reaction of 1b with NO2- at pH 4, but
its levels gradually decreased as 10 began to accumulate.
1b
3
2′
5′
6′
-
Reaction of 2 × 10-3 M 1c with NO2 (2 molar equiv)
greater stability compared to the Z isomers,21 as deduced
from brief molecular mechanics calculations (MM+) on
structures 7a and 9a , but may also stand in witness of a
concerted decarboxylation mechanism.15
proceeded virtually to completion to give mainly 5b, in
accord with the previous report.15
Mech a n istic Issu es. The present survey confirms the
radically divergent reaction pathways of 1a and its esters
1b/c with acidic NO2-. The structures of main products
from 1a support previous hypotheses15 suggesting regio-
selective nitrosation at the 2-position of the propenoic
acid moiety as the initial reaction step. This would be
brought about by NO+ or a related nitrosating species
In Scheme 1, oxime 9a is shown to arise from oxidation
of 7a . This was confirmed in separate experiments in
which 7a was reacted with NO2- at pH 4 to give 9a along
with 3,4-dihydroxybenzaldehyde. Formation of the latter
may follow from O-nitrosation of the oxime OH group in
7a , which would set the scene for sequential or concerted
losses of HNO2 and HCN by a process reminiscent of the
Wohl degradation of aldoses.22 The importance of the
unsubstituted OH group R to the oxime group in provid-
ing the driving force for chain breakdown was apparent
from the resilience of 7b to give rise to 3,4-dihydroxy-
benzaldehyde under the reaction conditions.
-
produced by acid promoted decomposition of NO2 via
HNO2.
Brief semiempirical (AM1/PM3) calculations predicted
for both 1a and 1b the 2-position of the propenoic acid
sector as the most reactive site, as apparent from the
relatively high HOMO coefficient and the highest total
charge density (Table 1).
The formation of a benzoic acid derivative is apparently
unprecedented in the chemistry of caffeic acid analogues
and derivatives with acidic NO2-. In principle, this acid
could arise by oxidation of 3,4-dihydroxybenzaldehyde;
however, this possibility was ruled out by the lack of
appreciable conversion of the latter compound to 3,4-
dihydroxybenzoic acid under the reaction conditions.
Thus, the only plausible route to 3,4-dihydroxybenzoic
acid was oxidative breakdown of the oxime 9a or of a
related species by a mechanism akin to that operative
in the case of 3,4-dihydroxybenzaldehyde formation.
Unfortunately, this mechanism could not be convincingly
validated by direct experimental proof, since neither 9a
nor its precursor 7a afforded significant amounts of 3,4-
This result conforms to expectations based on mere
inspection of canonical resonance forms for 1a and 1b
(not shown) which underscore the primary role of the
p-hydroxyl group on the catechol ring in pushing π
electron density toward the 2-position of the propenoic
acid chain. In accord with the chain activating effect of
the aromatic p-hydroxyl group are the similar patterns
-
of reactivity with acidic NO2 displayed by p-coumaric
and ferulic acids12-14 and the complete lack of reactivity
of cinnamic acid (as observed in separate experiments).
On this basis, it can be safely concluded that the
different behaviors exhibited by 1a and its esters toward
-
acidic NO2 do not reflect an intrinsically different π
electron makeup, but are determined by the availability
of a decarboxylation route, paving the way to otherwise
hardly accessible evolution pathways of the first formed
nitroso intermediate(s) (Scheme 1).
-
dihydroxybenzoic acid upon reacting with NO2 under
various acidic conditions. A possible explanation is that
3,4-dihydroxybenzoic acid is formed mainly via a 2-oxo-
nitrile intermediate produced specifically from Z oximes,
because of the favorable anti arrangement of the breaking
bonds, and that E-Z isomerization does not occur to a
sufficient extent with isolated 7a or 9a under the
conditions of the mechanistic experiments. Alternatively,
Z oximes may be directly generated by decarboxylation
of the nitroso intermediates, which cannot be ruled out
on the basis of available evidence. Indirect proof of the
actual formation in the pathway of Z oximes would be
provided by the generation of benzoxazinone 4b, which
purportedly involves intramolecular cyclization of an
o-quinone Z oxime intermediate.
The facile decarboxylation of the putative nitroso
intermediates from 1a would be warranted by concomi-
tant isomerization of the nitroso group to a stable
aldoxime functionality. This step would be preceded by
competing attacks of the various nucleophiles to the
resonance-stabilized nitroso cation, or a related species,
wherebyproduct distribution would mirror the relative
nucleophilicities and concentrations of all species in the
medium. Consistent with this view, under conditions of
-
low methanol and NO2 concentration, H2O is the most
effective nucleophile causing 7a to prevail over 7b and
-
3c, whereas with relatively high NO2 concentrations,
Many of the mechanistic scenarios proposed in this
study show noticeable similarities with those reported
formation of a 1,2-nitronitroso compound, precursor to
3c,20 would become the main route. The preferential
formation of E oximes would be in accord with their
(21) Saunders: W. H.; Cockerill, A. F. Mechanism of Elimination
Reactions; J . Wiley and Sons: New York, 1973.
(20) (a) Boulton, A. J .; Hadjimihalakis, P.; Katrisky, A. R.; Majid
Hamid, A. J . Chem. Soc. C 1969, 1901-1903. (b) Gasco, A.; Boulton,
A. J . Adv. Heterocycl. Chem. 1981, 29, 281.
(22) Barton, D.; Ollis, D. N. Comprehensive Organic Chemistry: The
Synthesis and Reaction of Organic Compounds; Hoslam, E., Ed.;
Pergamon Press: New York, 1979; Vol. V, p 696.

Reaction of Caffeic Acid and Its Esters with Nitrite
J . Org. Chem., Vol. 67, No. 3, 2002 807
Sch em e 1
in a recent paper dealing with the reaction of p-coumaric
acid with nitrite.14 In that paper the authors provided a
detailed spectral characterization of hydroxy- and oxo-
aldoxime derivatives, as well as of benzaldehyde, ben-
zoxazinone and furoxan products, in excellent agreement
with our data. By elegant experiments using 18O -labeled
water, they conclusively demonstrated that the hydroxy
and oxo group in the aldoxime products were derived
from the water of the medium. However, in contrast with
the mechanistic pathways in Scheme 1, in the previous
paper14 decarboxylation was suggested to precede the
addition of water, but not of nitrite, whereby chain
hydroxylation would be the result of the conjugated
addition of water to a â-nitrosostyrene derivative. Al-
though the proposed routes are not mutually exclusive,
and may well concur to product distribution, we would
favor the one indicated in Scheme 1, in which water or
methanol would add to a highly electrophilic nitroso
cation or a related species (e.g. a quinone methide).
In the case of esters 1b,c, the unfavorable equilibrium
with the putative nitroso intermediate(s) cannot be
driven by decarboxylation or other irreversible routes,
whereby aromatic nitration and dimerization become the
only viable pathways (Scheme 2). Three possible options
for catechol ring nitration leading to 5a and 5b could be
envisaged, namely NO2- induced oxidation of the catechol
to semiquinone followed by radical coupling with NO2;
Sch em e 2
NO2-, indicating that the free o-dihydroxy functionality
is an essential requisite for nitration. This point was
definitively confirmed in separate experiments in which
the O,O-dimethyl derivative of 1b was prepared and
-
nucleophilic attack of NO2 to the o-quinone; or electro-
philic nitrosation of the catechol ring followed by oxida-
tion. The latter route seems untenable in view of previous
experience gained in catechol nitration23 with acidic
shown to remain virtually unaffected upon prolonged
-
exposure to acidic NO2
.
Whether nitration occurs via the semiquinone or the
quinone route cannot be assessed on the basis of available
data. Two points, however, are worthy of note. First,
nitration of 1b is significant at pH 4 and is drastically
(23) (a) Napolitano, A.; d’Ischia, M.; Costantini, C.; Prota, G.
Tetrahedron 1992, 48, 8515-8522. (b) d’Ischia, M.; Costantini, C.
Bioorg. Med. Chem. 1995, 3, 923-927.

808 J . Org. Chem., Vol. 67, No. 3, 2002
Napolitano and d’Ischia
decreased at pH 1, as noted by the previous authors,15
at variance with simple catechols, which are conveniently
converted to 6-nitrocatechols with NO2 under strongly
concentrations, viz. with high NO2-/1a ratios, it may be
converted to products with potential nitrosative proper-
ties, e.g. the furoxan 3b.
-
acidic conditions (e.g., 1% sulfuric acid).23,24 Second,
formation of 6 from 1b at acidic pH is convincing evidence
for the generation of 1b quinone which would suffer
nucleophilic attack by 1b through the 2-position (a
mechanistic point that might be construed as confirma-
tory proof of the high reactivity of this site predicted also
for 1b).
Exp er im en ta l Section
Gen er a l Meth od s. Caffeic acid (1a ), chlorogenic acid (1c),
cinnamic acid, 3,4-dihydroxybenzaldehyde, 3,4-dihydroxyben-
zoic acid, and tetranitromethane were used as obtained. Caffeic
acid methyl ester (1b) was prepared from caffeic acid by
Fischer esterification (HCl (g)/dry methanol). Diazomethane
was prepared from N-methyl-N-nitroso-p-toluensulfonamide
in ethanolic KOH. CAUTION! Tetranitromethane is highly
toxic and is no more commercially available. Diazomethane is
explosive and must be collected in peroxide-free ether in a dry
ice/acetone bath and kept at -20 °C.
On these bases, and considering that NO2- concentra-
tion may be still high at pH 4 but very low at pH 1 (pKa
HNO2 ) 3.25),25 the NO2--quinone route seems a plau-
sible, though not exclusive option.
Formation of 10 from 1b raised a relevant issue
concerning the temporal sequence of the catechol nitra-
tion and dimerization steps. The finding that 6 was
UV spectra were performed with a diode array spectropho-
tometer. 1H (13C) NMR spectra were recorded at 400.1 (100.6)
1
MHz. 1H-1H COSY, 1H-13C HMQC, and H-13C HMBC NMR
-
smoothly converted to 10 by reaction with NO2 at pH
experiments were run at 400.1 MHz using standard pulse
programs from the Bruker library. For EI/MS spectra samples
were ionized with a 70 eV beam, and the source was taken at
230 °C. Main fragmentation peaks are reported with their
relative intensities (percent values are in parentheses).
Analytical and preparative TLC analyses were performed
on F254 0.25 and 0.5 mm silica gel plates or high performance
TLC (HPTLC) using chloroform-methanol 95:5 (eluant A), 90:
10 (eluant B), 85:15 (eluant C), chloroform-methanol 90:10
containing 1% or 0.2% acetic acid (eluant D or E), dichlo-
romethane-ethyl acetate 90:10 (eluant F) or benzene-ethyl
acetate 70:30 (eluant G). Griess reagent (1% sulfanilamide and
0.1% naphthylethylenediamine in 5% phosphoric acid),26 10%
ferric chloride in ethanol and 2 M sodium hydroxide were used
for product detection on TLC plates. Column chromatography
was performed on silica gel 60, 0.040-0.063 mm, 230-400
mesh.
4, and that 1b and 5a did not exhibit mutual reactivity
at the same pH, was taken as definitive evidence of the
sequence depicted in Scheme 2.
In product 10, the selective nitration of the unconju-
gated catechol ring argues further in favor of an oxidative
mechanism, which would be hindered in the hydroxydi-
hydrobenzofuran moiety.
Con clu sion s
The study of the reactions of 1a and related antioxidant
and antinitrosating agents with NO2- is an area of great
promise for elucidating mechanisms of colorectal car-
cinogenesis and for the design of novel antioxidants and
anticancer agents for rational chemopreventive strate-
gies. In this perspective, the results of the present
investigation (a) substantially modify and expand the
current knowledge of the reaction behavior of 1a and
related compounds with acidic NO2-; (b) throw light on
a range of reaction products and mechanisms that
apparently eluded the attention of previous workers; (c)
provide an improved background to understand at chemi-
cal level the biological properties of these important
dietary constituents.
Analytical and preparative HPLC was performed with an
instrument equipped with a UV detector set at 280 nm.
Octadecylsilane coated columns, 4.6 × 250 mm or 22 × 250
mm, 5 µm particle size were used for analytical or preparative
runs, respectively. Flow rates of 1 or 12 mL/min were used.
Different isocratic and gradient elution conditions were used
as follows: 0.1 M formic acid-methanol 97:3 (solvent A), 0.1
M formic acid-methanol 10:90 (solvent B), 0-5 min solvent A,
5-35 min from 0 to 40% solvent B, 35-45 40% solvent B
(eluant system I); 0.1 M formic acid /methanol 70:30 v/v (eluant
II) or 80:20 (eluant III); 0.1 M formic acid/acetonitrile 65:35
v/v (eluant IV).
The remarkable facility to decarboxylation may be the
Molecular mechanics (MM+) and semiempirical AM1/PM3
calculations were carried out with the Hyperchem 5.0 package
produced by Hypercube Inc. (Waterloo, Ontario, Canada) 1997.
Meth yl (E)-3-(3,4-Dim eth oxyp h en yl)p r op en oa te. Com-
pound 1a (200 mg, 1.1 mmol) in DMF (2 mL) was treated with
methyl iodide (1 mL) and sodium carbonate (0.5 g), and the
mixture was stirred at 60 °C for 48 h. The mixture was then
diluted with water (10 mL) and extracted with ethyl acetate
(3 × 5 mL). The combined organic layers were dried over
sodium sulfate and taken to dryness to give a residue which,
after purification by silica gel chromatography (30 × 1.5 cm
column) using chloroform as the eluant, afforded the title
-
underlying cause of the greater efficiency of 1a as NO2
scavenger compared to its esters. Loss of CO2 would
provide the driving force for an irreversible sequence of
reaction steps that outcompete otherwise facile nitrosa-
tion/nitration at the catechol moiety. This latter moiety,
in fact, directs reactivity toward the side chain of 1a
through resonance effects and comes into play as target
site only when decarboxylation is precluded, as in 1b,c.
Elucidation of the reaction sequences of 1a delineated
a mechanism by which the nature and concentration of
nucleophilic constituents in the medium gain importance
in the early stages, governing the balance of competing
routes. This variability may have several biological
implications, currently under assessment in our labora-
tory. As a relevant example, it may offer a plausible
explanation to the reported concentration-dependent
change of effect, from stimulating to suppressive, of 1a
on gastric nitrosamine formation: whereas at high
concentration 1a can efficiently scavenge NO2-, at low
1
compound (40 mg, 16% yield). H NMR (acetone-d6) δ (ppm):
3.73 (s, 3H), 3.84 (s, 3H), 3.87 (s, 3H), 6.43 (d, J ) 15.8 Hz,
1H), 6.97 (d, J ) 8.4 Hz, 1H), 7.19 (dd, J ) 8.4, 1.8 Hz, 1H),
7.30 (d, J ) 1.8 Hz, 1H), 7.60 (d, J ) 15.8 Hz, 1H). Anal. Calcd
for C12H14O4: C, 64.85; H, 6.35. Found: C, 64.60; H, 6.37.
(E)-3-(4,5-Dih yd r oxy-2-n itr op h en yl)p r op en oic Acid . To
1a (500 mg, 2.8 mmol) dissolved in methanol (3 mL) 0.1 M
sodium carbonate (20 mL) was added under vigorous stirring
followed by a solution of tetranitromethane (335 µL, 2.8 mmol)
in ethyl acetate (6 mL) in three aliquots over 30 min. The
(24) (a) de la Breteche, M. L.; Servy, C.; Lenfant, M.; Ducrocq, C.
Tetrahedron Lett. 1994, 35, 7231-7232. (b) Daveu, C.; Servy, C.;
Dendane, M.; Marin, P.; Ducrocq, C. Nitric Oxide 1997, 1, 234-243.
(25) Lide, D. R., Ed. CRC Handbook of Chemistry and Physics, 75th
ed.; CRC Press: Boca Raton, FL, 1995.
(26) Klebanoff, S. J . Free Rad Biol. Med. 1993, 14, 351-360.

Reaction of Caffeic Acid and Its Esters with Nitrite
J . Org. Chem., Vol. 67, No. 3, 2002 809
mixture was extracted with ethyl acetate (3 × 30 mL) and the
combined organic layers were dried over sodium sulfate and
taken to dryness to give the title compound19 as an oily residue
(410 mg, 65% yield).
(E)-N-Acetoxy-2-acetoxy-2-(3,4-diacetoxyph en yl)eth an -
a loxim e (7c). H NMR (CD3OD) δ (ppm): 2.15 (s, 6H), 2.26
1
(s, 6H), 6.42 (d, 1H, J ) 6.0 Hz), 7.27 (d, 1H, J ) 8.0 Hz), 7.34
(d, 1H, J ) 2.0 Hz), 7.37 (dd, 1H, J ) 8.0, 2.0 Hz), 8.00 (d, 1H,
J ) 6.0 Hz). EI/MS m/z 351 (M+, 1), 309 (7), 291 (3), 267 (12),
249 (13), 207 (50), 165. (95), 147 (50), 138 (40), 137 (100).
HRMS for C16H17NO8: calcd 351.0954, found 351.0932.
(E)-N-Acetoxy-2-(3,4-diacetoxyph en yl)-2-m eth oxyeth an -
Gen er a l P r oced u r e for Rea ction of 1a /b a n d Oth er
Ca tech olic Com p ou n d s w ith Nitr ite. To 1a /b (1 mmol)
dissolved in methanol (5 mL) were added sequentially 0.05 M
acetate buffer pH 4 (500 mL) and sodium nitrite (2 mmol),
and the mixture was kept under stirring at room temperature.
The reaction course was followed by HPLC (eluant system I
for 1a and eluant III for 1b). After 4 h or after complete
consumption of the substrate, the pH of the reaction mixture
was raised to 7.0 by addition of sodium hydrogencarbonate.
The mixture was extracted with ethyl acetate (3 × 150 mL),
and the combined organic layers were dried over sodium
sulfate and taken to dryness.
1
a loxim e (7d ). H NMR (CD3OD) δ (ppm): 2.12 (s, 3H), 2.26
(s, 6H), 3.41 (s, 3H), 4.98 (d, 1H, J ) 7.6 Hz), 7.25 (d, 1H, J )
8.4 Hz), 7.29 (d, 1H, J ) 2.0 Hz), 7.32 (dd, 1H, J ) 8.4, 2.0
Hz), 7.78 (d, 1H, J ) 7.6 Hz). EI/MS m/z 323 (M+, 1), 281 (11),
239 (13), 221 (29), 179 (100), 148 (100). HRMS for C15H17NO7:
calcd 323.1005, found 323.1011.
2-Acet oxy-2-(3,4-d ia cet oxyp h en yl)a cet on it r ile (8a ).
1
UV: λmax (CH3OH) 300 nm. H NMR (CD3OD) δ (ppm): 2.15
In other experiments, the reaction of 1a was run as
follows: (i) as above but without addition of methanol; (ii)
purging with argon the solution of 1a in the acetate buffer
prior to addition of sodium nitrite; (iii) using substrate
concentrations in the range 0.1-100 mM and 2 molar equiv
of nitrite; (iv) with nitrite varying in the range 0.5-5 molar
equiv with respect to the substrate at 2 mM concentration;
(v) as in the general procedure but in 0.1 M phosphoric acid,
pH 1 or 2.
(s, 3H), 2.27 (s, 6H), 6.55 (s, 1H), 7.34 (d, 1H, J ) 8.4 Hz),
7.45 (d, 1H, J ) 2.0 Hz), 7.49 (dd, 1H, J ) 8.4, 2.0 Hz). EI/MS
m/z 291 (M+, 6), 249 (84), 207 (100), 165 (100), 147 (100).
HRMS for C14H13NO6: calcd 291.0743, found 291.0735.
2-(3,4-Dia cet oxyp h en yl)-2-m et h oxya cet on it r ile (8b ).
1
UV: λmax (CH3OH) 300 nm. H NMR (CD3OD) δ (ppm): 2.27
(s, 6H), 3.52 (s, 3H), 5.45 (s, 1H), 7.31 (d, 1H, J ) 8.4 Hz),
7.37 (d, 1H, J ) 2.0 Hz), 7.42 (dd, 1H, J ) 8.4, 2.0 Hz). EI/MS
m/z 263 (M+, 12), 221 (100), 179 (100), 148 (100). HRMS C13H13
-
Reactions of 3,4-dihydroxybenzoic acid, 3,4-dihydroxyben-
zaldehyde, cinnamic acid, or methyl (E)-3-(3,4-dimethoxyphen-
yl)propenoate (1 mmol) dissolved in methanol (5 mL) with
nitrite (2 mmol) was carried out under the general conditions.
The reaction course was followed by HPLC (eluant system IV
in the case of 3,4-dihydroxybenzoic acid or I in the case of 3,4-
dihydroxybenzaldehyde) or by TLC (eluant B for cinnamic acid
or eluant F for methyl (E)-3-(3,4-dimethoxyphenyl)propenoate).
Isola t ion of 3,4-Dih yd r oxyb en za ld eh yd e, 3,4-Dih y-
d r oxyben zoic a cid , 4b, 7a -d , a n d 8a /b. For preparative
purposes, the reaction of 1a with nitrite was carried out in
0.05 M acetate buffer, pH 4.0 using 200 mg of the starting
material. After workup of the reaction mixture, the residue
(60 mg) was fractionated by PTLC (eluant D) to give four main
bands. The less polar (Rf 0.33, 13 mg, 9% yield) was identified
as 3,4-dihydroxybenzaldehyde by comparison of the spectral
NO5: calcd 263.0794, found 263.0772.
Isola tion of 3c a n d 9a . The reaction of 1a (500 mg, 2.78
mmol) dissolved in methanol (12 mL) with nitrite (1.0 g) was
carried out as above in 0.05 M, acetate buffer, pH 4.0 (12 mL).
After workup of the reaction mixture, the residue (376 mg)
was fractionated by chromatography on silica gel (30 × 1.5
cm column) using chloroform/ethyl acetate (9:1 to 1:1 gradient
mixtures) to afford two main fractions. The fraction collected
with chloroform/ethyl acetate from 9:1 to 7:3 proved to consist
15
of 3c
(63 mg, 12% yield), while the fraction eluted with
chloroform/ethyl acetate 1:1 (67 mg) was further purified by
PTLC (eluant C) to give pure 9a (Rf 0.51, 38 mg, 8% yield).
The latter compound was subjected to methylation by treat-
ment with diazomethane to afford, after PTLC fractionation
(eluant A), 9b.
2-(3,4-Dih yd r oxyp h en yl)-2-oxoeth a n a loxim e (9a ). UV:
1
(UV, H NMR) and chromatographic properties with those of
λ
max (CH3OH) 234 nm, 273 nm, 322 nm (CH3OH, 0.1 M NaOH)
an authentic sample. Another fraction (Rf 0.30, 4 mg, 2% yield)
consisted of 4b identified by comparison of the spectral
properties (1H NMR) with those reported.15 The other two
bands were found to consist of pure 7b (Rf 0.19, 10 mg, 5%
yield) and 7a (Rf 0.07, 20 mg, 10% yield). Purification of the
aqueous layer by preparative HPLC (eluant II) afforded 3,4-
dihydroxybenzoic acid (tR 16 min, 10 mg, 6% yield) identified
by spectral analysis (1H and 13C NMR) and comparison with
an authentic sample.
Compound 7a or 7b (15 mg) was treated with acetic
anhydride (1 mL) and pyridine (40 µL) overnight at room
temperature. After removal of the volatile components, the
residues were fractionated by HPTLC (eluant G) to afford in
each case two main fractions. The less polar bands were found
to consist of 8a (Rf 0.64, 11 mg, 46% yield) or 8b (Rf 0.64, 9
mg, 45% yield), while the others contained pure 7c (Rf 0.49,
12 mg, 42% yield) or 7d (Rf 0.50, 7 mg, 28% yield).
1
267 nm, 379 nm. H NMR (acetone-d6) δ (ppm): 6.91 (d, 1 H,
J ) 8.8 Hz), 7.61 (dd, 1H, J ) 8.8, 2.0 Hz), 7.62 (d, 1H, J )
2.0 Hz), 7.91 (s, 1H). 13C NMR (acetone-d6) δ (ppm): 115.2
(CH), 117.1 (CH), 124.5 (CH), 129.2 (C), 141.7 (C), 145.3 (C),
148.7 (CH), 187.1 (C). Anal. Calcd for C8H7NO4: C, 53.04; H,
3.89; N, 7.73. Found: C, 53.22; H, 3.92; N, 7.80.
N -Me t h oxy-2-(3,4-d im e t h oxyp h e n yl)-2-oxoe t h a n a l-
oxim e (9b). EI/MS: m/z 223 (M+, 32), 165 (100), 137 (13), 122
(11), 107 (15). HRMS for C11H13NO4: calcd 223.0844, found
223.0865.
2-Oxy-3-(3,4-d ih yd r oxyp h en yl)-1,2,5-oxa d ia zole (3c).
1
UV: λmax (CH3OH) 247, 316. H NMR (CD3OD) δ (ppm): 6.89
(d, J ) 8.4 Hz, 1H), 7.32 (dd, J ) 8.4, 2.0 Hz), 7.54 (d, J ) 2.0
Hz, 1H), 8.92 (s, 1H). 13C NMR δ (ppm): 113.9 (CH), 115.4
(C), 116.2 (C), 117.3 (CH), 120.1 (CH), 146.4 (CH), 147.5 (C),
149.7 (C). EI/MS: m/z 194 (M+, 33), 134 (100).
(E)-2-Hyd r oxy-2-(3,4-d ih yd r oxyp h en yl)et h a n a loxim e
(7a ). UV: λmax (CH3OH) 280 nm. H NMR (CD3OD) δ (ppm):
Isola tion of 5a , 6, a n d 10. For preparative purposes,
reaction of 1b with nitrite was carried out in 0.05 M acetate
buffer pH 4.0 as described above, using 200 mg (1.0 mmol) of
the starting material. After workup of the reaction mixture,
the residue (192 mg) was fractionated by PTLC (eluant E) to
afford two main fractions consisting of 5a (Rf 0.44, 107 mg,
45% yield) identified by comparison of the spectral properties
(1H NMR) with those previously reported 15 and of 10 (Rf 0.35,
46 mg, 10% yield). Compound 10 was acetylated with acetic
anhydride/pyridine as described above for 7a /b to afford a main
product purified by PTLC (eluant F). The reaction of 1b (200
mg) with nitrite (2 molar equiv) was also carried out in 0.1 M
phosphate buffer, pH 1.0. Fractionation of the residue (296
mg) obtained from work up of the reaction mixture on PTLC
(eluant D) afforded 5a (Rf 0.38, 51 mg, 21% yield), 6 (Rf 0.35,
30 mg, 8% yield) identified by comparison of the spectral
1
5.06 (d, J ) 7.0 Hz, 1H), 6.71 (d, J ) 2.0 Hz, 1H), 6.72 (d, J )
7.6 Hz, 1H), 6.80 (dd, J ) 7.6, 2.0 Hz, 1H), 7.36 (d, J ) 7.0 Hz,
1H). 13C NMR (CD3OD) δ (ppm): 73.2 (CH), 115.4 (CH), 117.2
(CH), 119.7 (CH), 134.7 (C), 147.0 (C), 147.3 (C), 153.9 (CH).
Anal. Calcd for C8H9NO4: C, 52.46; H, 4.95; N, 7.65. Found:
C, 52.50; H, 4.84; N, 7.47.
(E )-2-(3,4-D i h y d r o x y p h e n y l)-2-m e t h o x y e t h a n a l-
1
oxim e (7b). UV: λmax (CH3OH) 280 nm; H NMR (CD3OD) δ
(ppm): 3.29 (s, 3H), 4.61 (d, 1H, J ) 7.4 Hz), 6.65 (dd, 1H, J
) 8.0, 2.0 Hz), 6.73 (d, 1H, J ) 8.0 Hz), 6.77 (d, 1H, J ) 2.0
Hz), 7.30 (d, 1H, J ) 7.4 Hz). 13C NMR (CD3OD) δ (ppm): 57.3
(CH3), 82.6 (CH), 115.7 (CH), 117.2 (CH), 120.4 (CH), 132.1
(C), 147.3 (C), 152.4 (CH). Anal. Calcd for C9H11NO4: C, 54.82;
H, 5.62; N, 7.10. Found: C, 54.93; H, 5.45, N, 7.01.

810 J . Org. Chem., Vol. 67, No. 3, 2002
Napolitano and d’Ischia
properties with those of an authentic sample,16 and 10 (Rf 0.30,
22 mg, 5% yield).
sequentially 0.05 M acetate buffer pH 4 (3 mL) and sodium
nitrite (204 mg) under vigorous stirring at room temperature.
The reaction course was followed by HPLC (eluant III). After
1 h or after complete consumption of the substrate, the mixture
was taken to a small volume and fractionated by preparative
HPLC (eluant II) to give 5b (tR 23 min, 90 mg, 40% yield)
identified by comparison of the spectral properties (1H NMR)
with those reported.15 UV: λmax (MeOH) 276, 339, 430 nm; λmax
Meth yl (E)-3-(4,5-Dih ydr oxy-2-n itr oph en yl)pr open oate
(5a ). UV: λmax (CH3OH) 274 nm, (CH3OH, 0.1M NaOH) 258,
306, 322, 432 nm. 1H NMR (CD3OD) δ (ppm): 3.78 (s, 3H),
6.26 (d, J ) 16.0 Hz, 1H), 7.03 (s, 1H), 7.52 (s, 1H), 8.08 (d, J
) 16 Hz, 1H). 13C NMR (CD3OD) δ (ppm): 113.8 (CH), 115.4
(CH), 121.6 (CH), 125.6 (C), 142.7 (C), 144.3 (CH), 149.3 (C),
153.2 (C), 169.7 (C).
1
(CH3OH, NaOH 0.01 M) 262, 341, 439 nm. H NMR (CD3OD)
Meth yl (E)-3-[7-Hyd r oxy-2-(4,5-d ih yd r oxy-2-n itr op h en -
yl)-3-m eth oxycar bon yl-2,3-dih ydr o-1-ben zofu r an -5-yl]pr o-
p en oa te (10). UV: λmax (CH3OH) 303, 325 nm (CH3OH, NaOH
0.01 M) 275, 319, 367, 483 nm;1H NMR (CD3OD) δ (ppm): 3.75
(s, 3H), 3.81 (s, 3H), 4.13 (d, 1H, J ) 4.4 Hz), 6.30 (d, 1H, J )
16.0 Hz), 6.65 (d, 1H, J ) 4.4 Hz), 6.96 (s, 1H), 7.03 (s, 1H),
7.06 (s, 1H), 7.54 (d, 1H, J ) 16.0 Hz), 7.67 (s, 1H). 13C NMR
(CD3OD) δ (ppm): 53.0 (CH3), 54.2 (CH3), 58.7 (CH), 85.9 (CH),
114.3 (CH), 114.6 (CH), 117.1 (CH), 118.5 (CH), 118.5 (CH),
128.1 (C), 131.4 (C), 132.9 (C), 139.5 (C), 144.2 (C), 147.1 (CH),
147.4 (C), 151.5 (C), 154.3 (C), 170.3 (C), 174.5 (C). Anal. Calcd
for C21H17NO10: C, 55.69; H, 3.97; N, 3.25. Found: C, 55.47;
H, 3.85; N, 3.32. Acetyl Der iva tive. 1H NMR (CDCl3) δ
(ppm): 2.30 (s, 3H), 2.32 (s, 3H), 2.34 (s, 3H), 3.79 (s, 3H),
3.86 (s, 3H), 4.26 (d, 1H, J ) 4.4 Hz), 6.29 (d, 1H, J ) 16.0
Hz), 6.78 (d, 1H, J ) 4.4 Hz), 7.19 (d, 1H, J ) 1.6 Hz), 7.35 (d,
1H, J ) 1.6 Hz), 7.57 (d, 1H, J ) 16.0 Hz), 7.67 (s, 1H), 8.10
(s, 1H). EI/MS: m/z 470 (97), 428 (24), 386 (66), 326 (47), 267
(29), 194 (100), 134 (21).
δ (ppm): 2.12 (m, 4H), 3.74 (dd, J ) 10.0, 3.2 Hz, 1H), 4.20 (m,
1H), 5.42 (m, 1H), 6.31 (d, J ) 15.6 Hz, 1H), 7.05 (s, 1H), 7.57
(s, 1H), 8.15 (d, J ) 15.6 Hz, 1H). 13C NMR (CD3OD) δ (ppm):
39.5 (CH2), 41.3 (CH2), 73.2 (CH), 73.3 (CH), 75.3 (C), 78.2
(CH), 113.2 (CH), 116.0 (CH), 121.2 (CH), 125.6 (C), 140.7 (C),
144.0 (CH), 150.2 (C), 156.2 (C), 169.1 (C), 181.8 (C).
Ack n ow led gm en t. This work was carried out in the
frame of a MURST (PRIN 2001) project. We thank the
“Centro Interdipartimentale di Metodologie Chimico-
Fisiche of Naples University” for NMR facilities and
mass spectra.
13
Su p p or tin g In for m a tion Ava ila ble: 1H and
C NMR
spectra of compounds 7a /7b and 9a . H NMR, H--H COSY,
1H-13 C HMQC, and HMBC spectra of compound 10. This
material is available free of charge via the Internet at
http://pubs.acs.org.
1
1
Rea ction of 1c w ith Nitr ite. Isola tion of 5b. To 1c (200
mg, 0.56 mmol) dissolved in methanol (3 mL) were added
J O015965V