Stable isotope characterization of the ortho-oxygenated phenylpropanoids: Coumarin and melilotol

Article abstract of DOI:10.1021/jf0518507

The natural abundance ²H NMR spectra of extractive coumarin 10 and of its dihydroderivative melilotol 11 produced by baker's yeast reduction has been compared with synthetic materials. Diagnostic for the differentiation of 10 are the (D/H)_β values, which are in the 128.1-133.6 ppm interval for the natural compounds but 258.5 and 189.8 ppm for the synthetic materials. Such a dramatic difference is also found for methyl cinnamate 12, which shows (D/H)_β values of 127.2 and 515.8 ppm, respectively. In extractive 10, the ratio (D/H)_4(para)/(D/H)_6(ortho) = 1.24 is similar to that observed in structurally related salicin and methyl salicylate. Coumarin 10 is transformed in salicyl alcohol 9, providing diacetate 14, showing in the natural series the trend (D/H)_3(meta) > (D/H)_4(para) > (D/H)_5(meta) ～ (D/H) _6(ortho). A similar trend is shown also by the synthetic 10. A clear distinction between extractive and synthetic 10 is obtained through δ¹⁸O determinations on 10 and on chroman 13. The bulk δ¹⁸O values in the extractive series of 10 are 20.3, 23.6, and 22.6‰, while those of the aromatic oxygen are 2.3, 0.5, and -0.5‰. In the synthetic sample, the values are 12.6 and 5.6‰, respectively. As a final product, the reduction of 10 leads to the dihydroderivative 11. Both the baker's yeast reduction and the catalytic hydrogenation lead to a marked decrease of the deuterium content of 11, which is stronger for the β-position than for the α-position.

Full text of DOI:10.1021/jf0518507

J. Agric. Food Chem. 2005, 53, 9383−9388 9383
Stable Isotope Characterization of the ortho-Oxygenated
Phenylpropanoids: Coumarin and Melilotol
ELISABETTA BRENNA,† GIOVANNI FRONZA,*,‡ CLAUDIO FUGANTI,†
FRANCESCO G. GATTI,† VALENTINA GRANDE,† STEFANO SERRA,‡
§
CLAUDE GUILLOU,§ FABIANO RENIERO,§ AND FRANCESCA SERRA
Dipartimento di Chimica, Materiali e Ingegneria Chimica “Giulio Natta”, Politecnico di Milano, and
Istituto di Chimica del Riconoscimento Molecolare, CNR, Via Mancinelli 7, 20131 Milano, and
Physical and Chemical Exposure Unit, Institute for Health and Consumer Protection, Joint Research
Centre, European Commission, BEVABS, TP281, Via Fermi 2, 21020 Ispra, Italy
The natural abundance 2H NMR spectra of extractive coumarin 10 and of its dihydroderivative melilotol
11 produced by baker’s yeast reduction has been compared with synthetic materials. Diagnostic for
the differentiation of 10 are the (D/H)â values, which are in the 128.1-133.6 ppm interval for the
natural compounds but 258.5 and 189.8 ppm for the synthetic materials. Such a dramatic difference
is also found for methyl cinnamate 12, which shows (D/H)â values of 127.2 and 515.8 ppm,
respectively. In extractive 10, the ratio (D/H)4(para)/(D/H)6(ortho) ) 1.24 is similar to that observed in
structurally related salicin and methyl salicylate. Coumarin 10 is transformed in salicyl alcohol 9,
providing diacetate 14, showing in the natural series the trend (D/H)3(meta) > (D/H)4(para) > (D/H)5(meta)
∼ (D/H)6(ortho). A similar trend is shown also by the synthetic 10. A clear distinction between extractive
and synthetic 10 is obtained through δ18O determinations on 10 and on chroman 13. The bulk δ18O
values in the extractive series of 10 are 20.3, 23.6, and 22.6‰, while those of the aromatic oxygen
are 2.3, 0.5, and -0.5‰. In the synthetic sample, the values are 12.6 and 5.6‰, respectively. As a
final product, the reduction of 10 leads to the dihydroderivative 11. Both the baker’s yeast reduction
and the catalytic hydrogenation lead to a marked decrease of the deuterium content of 11, which is
stronger for the â-position than for the R-position.
KEYWORDS: Coumarin; extractive; synthetic; deuterium NMR; chroman; positional δ18O values;
dihydrocoumarin; methyl cinnamate
INTRODUCTION
position 3 to ferulic acid, the closest precursor of 1 (Figure 1)
(8). The 2H NMR analyses and the determination of the
positional δ18O values of extractive 1-3 have revealed the
changes occurring in both the deuterium pattern of the aromatic
ring during the introduction of the oxygen functions (9) and
the extent of the isotopic fractionation of oxygen in the process
(10).
In this context, the pattern of natural abundance 2H of
extractive and synthetic samples of the glycoside salicin 8
through NMR analyses has been recently determined (11).
Indeed, salicyl alcohol 9, the aglycone moiety of 8, shares with
1-3 the derivation from cinnamic acid 4. In this instance,
cinnamic acid 4 is hydroxylated in position 2 to o-coumaric
acid 7, which subsequently undergoes C-2 chain shortening to
the C-6-C-1 framework of 9 (Figure 1). The deuterium pattern
of 9, obtained by enzymic hydrolysis of extractive 8, was
determined through NMR analyses on the diacetyl derivative
14. The main result was that (D/H)4(para) > (D/H)3(meta) ∼ (D/
H)5(meta) > (D/H)6(ortho). Another result of this study (11) was
that in these circumstances the H-3 and the H-5 of 9, which are
The stable isotope distribution of a number of aroma
compounds has been recently determined with the intent of
differentiating materials of chemically identical molecules of
natural or synthetic origin (1-3). In this context, the analysis
2
of the natural abundance H NMR spectra together with the
definition of the positional δ18O values achieved through
selective chemical degradation appeared particularly useful.
Relevant examples in the field of aromas are vanillin 1 (4, 5)
and raspberry ketone 2 (6) and the antioxidant resveratrol 3 (7).
These phenolic compounds share in nature a common derivation
from C-6-C-3 L-phenylalanine via cinnamic acid 4 and the
oxygenated derivative p-coumaric acid 5. In the production of
C-6-C-1 vanillin 1, the latter compound undergoes a second
hydroxylation to caffeic acid 6 and, in turn, is O-methylated at
* To whom correspondence shold be addressed. Tel: +39 02 23993027.
Fax: +39 02 23993080. E-mail: giovanni.fronza@polimi.it.
† Politecnico di Milano.
‡ CNR.
§ European Commission.
10.1021/jf0518507 CCC: $30.25
© 2005 American Chemical Society
Published on Web 11/05/2005

9384 J. Agric. Food Chem., Vol. 53, No. 24, 2005
Brenna et al.
Figure 1. Some biologically active phenolic phenylpropanoids and their relevant biosynthetic intermediates (1−10). Compounds 10−14 are the subject
of this work.
(6 g) of commercial Tonka bean absolute by liquid chromathograpphy
using a column with 200 g of SiO2 60 (Merck, Milano) with a hexane/
ethyl acetate mixture 6/4. A homogeneous solid was recrystallized from
the extract with a yield of 87%. Sample 10.1 was from material from
V. Mane & Fils (Bar-sur-Loup), and samples 10.2 and 10.3 were from
two different lots of absolute from Charabot (Grasse). Samples 10.4
and 10.5 were synthetic coumarins acquired from Fluka (Milano) and
Carlo Erba (Milano).
ortho-para located with respect to the phenolic oxygen, undergo
rapid equilibration with the hydrogen atoms of the solvent water
under acid catalysis.
To achieve a better picture of the mode of formation of natural
phenylpropanoids via the o-coumarate pathway, we submitted
to stable isotope characterization coumarin 10, the lactone of
the Z-isomer of o-coumaric acid, the impact flavor of Tonka
bean absolute [Dipteryx odorata (Aubl.) Willd. (Fabaceae)], a
widely used ingredient in the flavor industry (12). This study,
which might be useful for the development of an analytical
procedure allowing the differentiation of extractive molecules
from the synthetic counterparts, is based on the isotopic
The samples of dihydrocoumarin 11 were obtained by chemical and
biochemical hydrogenation of extractive and synthetic coumarin. For
the first process, 2 g of coumarin in 50 mL of ethyl acetate was
hydrogenated at ambient conditions in the presence of 0.2 g of 10%
Pd/C. After completeness of the reaction, the catalyst was removed by
filtration, and the residue obtained by evaporation of the solvent was
distilled bulb-to-bulb at 0.2 mm/Hg (oven temperature, 130 °C). The
solid obtained on cooling was then used as such for the deuterium
measurements. Alternatively, 5 g of coumarin dissolved in the minimum
amount of ethanol was added dropwise to a stirred mixture of 2 kg of
baker’s yeast (DSM, Milano) and 0.2 kg of D-glucose mixed in 2.5 L
of tap water at 35-38 °C. After 2 days under these conditions, 1 L of
acetone was added, followed by 1 L of ethyl acetate. The mixture was
filtered under vacuum through a large Celite pad. The organic phase
was separated, and the lower layer was extracted twice by means of 1
L of ethyl acetate. The combined organic extract was washed with 5%
NaHCO3 and 10% NaCl solution, dried over Na2SO4, and evaporated.
The dihydrocoumarin 11 was obtained from the residue by liquid
chromathography (SiO2, ethyl acetate in hexane) and purified by bulb-
to-bulb vacuum distillation, with a yield of 50%. The dihydrocoumarin
produced from sample 10.1 of coumarin (extractive) by chemical
hydrogenation was labeled sample 11.1, while the identical material
produced by baker’s yeast reduction was labeled sample 11.2. Sample
2
characterization through natural abundance H NMR of the
extractive and synthetic samples of coumarin 10 and further of
a series of related compounds, including dihydrocoumarin
(melilotol) 11 and methyl cinnamate 12 (Figure 1). In a
subsequent experiment, coumarin 10 was degraded to chroman
13, which would retain the aromatic oxygen of 10. The δ18O
value of chroman 13 of natural and synthetic derivation was
determined in order to gain information on the oxygen frac-
tionation occurring when 4 undergoes the transformation in
o-coumaric acid 7.
MATERIALS AND METHODS
Origin of Samples and Preparation of Derivatives. A total of 14
samples, including five samples of coumarin 10, three samples of
dihydrocoumarin 11, four samples of diacetyl salicyl alcohol 14, and
two samples of methyl cinnamate 12 were analyzed for the deuterium
content (Table 1). The extractive coumarin was isolated from an aliquot

Isotopic Characterization of Coumarin and Melilotol
J. Agric. Food Chem., Vol. 53, No. 24, 2005 9385
Table 1. Origin and (D/H)i Isotopic Ratios (Parts per Million)a of Coumarin 10, Dihydrocoumarin 11, and Diacetyl Salicyl Alcohol 14
Coumarin 10
b
c
sample (origin)
(D/H)3,5
(D/H)4
(D/H)6
(D/H)Ar
(D/H)
(D/H)
R4
/6
â
R
10.1 (natural from F. tonka)
10.2 (natural from Charabot)d
10.3 (natural from Charabot)d
10.4 (synthetic from Fluka)
10.5 (synthetic from Carlo Erba)
157.7 (1.7)
157.0 (2.6)
155.2 (3.6)
167.5 (3.3)
130.6 (1.0)
166.3 (4.7)
161.4 (4.3)
159.7 (6.3)
158.6 (2.8)
138.1 (4.6)
128.1 (2.1)
133.0 (3.2)
131.0 (5.4)
146.9 (3.2)
125.5 (3.5)
152.4
152.1
150.3
160.1
131.2
131.9 (1.9)
133.6 (2.9)
128.1 (5.2)
258.5 (4.3)
189.8 (4,3)
111.2 (3.7)
104.6 (4.6)
116.0 (2.5)
146.9 (3.2)
122.9 (2.7)
1.30
1.21
1.22
1.08
1.10
Dihydrocoumarin 11
sample (origin)
(D/H)3
(D/H)4,6
(D/H)5
(D/H)Ar
(D/H)
(D/H)
â
R
11.1 (from 10.1 by H2/Pd reduction)
11.2 (from 10.1 by b.y. reduction)
11.3 (from 10.4 by H2/Pd reduction)
171.4 (3.4)
168.4 (4.4)
179.0 (3.5)
150.4 (2.8)
144.7 (3.7)
160.7 (3.3)
135.0 (3.8)
132.1 (3.7)
158.2 (2.5)
151.8
147.4
164.6
81.1 (3.3)
85.5 (0.2)
121.9 (2.1)
95.7 (3.0)
93.7 (1.1)
122.0 (1.9)
Diacetyl salicyl alcohol 14
sample (origin)
(D/H)3
(D/H)4
(D/H)5
(D/H)6
(D/H)Ar
(D/H)7
R4
/
6
14.1 (natural from 10.1)
14.2 (natural from 10.2)
14.3 (natural from 10.3)
14.4 (synthetic from 10.4)
175.4 (5.7)
168.7 (3.7)
170.8 (5.1)
167.2 (2.5)
164.8 (3.2)
160.6 (5.0)
156.1 (2.3)
161.4 (1.3)
133.5 (3.2)
133.5 (4.9)
136.6 (2.8)
151.0 (1.3)
133.9 (4.5)
130.5 (3.2)
128.8 (3.2)
146.3 (6.6)
151.9
148.3
148.2
156.5
85.0 (2.0)
102.4 (3.3)
112.3 (4.0)
137.7 (2.5)
1.23
1.23
1.21
1.10
a The (D/H)i standard deviations are reported within parentheses. b (D/H)Ar is the mean deuterium content of the aromatic nuclei calculated from the individual positional
(D/H)i. c Ratio (D/H)4/(D/H)6. d Coumarin different lots.
11.3 was a dihydrocoumarin produced by chemical hydrogenation of
a synthetic coumarin (sample 10.4).
temperature, 90 °C). Chroman samples 13.1-13.3 were produced from
extractive coumarin samples 10.1-10.3, while sample 13.4 was
obtained from the synthetic sample 10.4. The studied samples of methyl
cinnamate 12 were a synthetic compound from Fluka at Milano (sample
12.1) and a commercially available “natural” product from H&R
(sample 12.2).
The diacetate of salicyl alcohol 14 was obtained from coumarin by
ozonolysis and NaBH4 reduction, followed by a chemical acetylation.
Three grams of coumarin dissolved in 100 mL of a 1:1 mixture of
CH2Cl2-MeOH was submitted to the action of ozonized oxygen at
-78 °C until the blue color persisted. Nitrogen was then flushed
through, followed by the dropwise addition, at the same temperature,
of an ethanolic solution of 0.9 g of NaBH4. The mixture was kept at
room temperature overnight and then diluted with 100 mL of icy water.
The organic phase was subsequently separated, and the reaction mixture
was extracted twice with CH2Cl2 (100 mL). The organic extract was
washed with 10% NaCl solution and dried to give, upon evaporation
of the solvent, salicyl alcohol in the solid phase. The crude material
was treated with 15 mL of pyridine with 10 mL of acetic anhydride.
After 24 h, the reaction mixture was evaporated under vacuum to give
a residue, which was taken up with water and CH2Cl2. The organic
extract was washed with cold diluted HCl and a saturated solution of
NaHCO3. The oily residue obtained upon evaporation was separated
by liquid chromathography on SiO2 with 15% ethyl acetate in hexane
providing the desired diacetate 14 in 60% yield. The latter was purified
by bulb-to-bulb vacuum distillation, as above. Samples 14.1, 14.2, and
14.3 were derived in this way from coumarin samples 10.1, 10.2, and
10.3, respectively, while sample 14.4 was from the synthetic coumarin
sample 10.4. The conversion of coumarin into chroman 13 (3,4-dihydro-
2H-1-benzopyran) was achieved by the following procedure. Coumarin
was chemically hydrogenated to dihydrocoumarin, as described above.
The latter material (3 g) in diethyl ether (50 mL) was added dropwise
under stirring to a boiling suspension of 2 g of LiAlH4 in 50 mL of
diethyl ether. After 5 h under these conditions, ethyl acetate was added
to destroy the excess of hydride, followed by cold diluted HCl. The
residue, obtained upon evaporation of the washed and dried extract,
contained only oily 3-(2-hydroxyphenyl)-1-propanol. The crude material
in 15 mL of pyridine was treated overnight at 0 °C with p-
toluenesulfonyl chloride (4.7 g, 1.2 mol equiv). The mixture was poured
onto crushed ice containing 10 mL of concentrated HCl, and the mixture
was subsequently extracted with diethyl ether (2 × 150 mL). The
washed and dried organic phase, after evaporation, gave an oily residue
that was dissolved in 50 mL of 2-propanol and added dropwise under
stirring conditions to 200 mL of 20% NaOH at 70-80 °C. After 5 h,
the cooled reaction mixture was diluted with water and extracted with
diethyl ether (2 × 150 mL). The organic residue was chromatographed
on SiO2 with 10% ethyl acetate in hexane to provide chroman 13 and
further purified by bulb-to-bulb vacuum distillation (0.5 mm/Hg; oven
The purity of the samples, checked by determination of melting point,
thin-layer chromatography, and proton NMR spectroscopy, was 100%.
To minimize eventual isotopic fractionation during the purification
process, all compounds were treated in the same way. They were
purified through column chromatography, pooling all pure fractions,
and the crystallizations were carried out using the minimum amount
of solvent to lower the loss of material in the mother liquors.
NMR Experiments. The 2H NMR experiments were performed on
a Bruker Advance 500 spectrometer equipped with a 10 mm probehead
and a 19F lock (C6F6) channel, under CPD (Waltz 16 sequence) proton
decoupling conditions. The spectra were recorded at 300-310 K. The
reference used for (D/H)i calculations was tert-butyl disulfide calibrated
against the official standard TMU (Community Bureau of References,
BCR) with a certified (D/H) ratio (123.38 ppm). The spectra were
recorded dissolving about 0.8 g of material in ca. 3.0 mL of solvent,
adding 50 µL of C6F6 for the lock and ca. 120 mg of tert-butyl disulfide
as internal standard [(D/H) 129.2 ppm]. The solvents used were acetone
for 10, 11, and 14 and CH2Cl2 for 12.
At least three spectra were run for each sample collecting ca. 8000
scans using the following parameters: 6.8 s acquisition time, 1200 Hz
width, 16 K time domain, 8 K FT data points, 3 s recycle delay. The
longest relaxation time (T1) estimated previously (11) belongs to the
methyl nuclei of the reference material (ca. 0.8 s). Thus, the repetition
time of 9.8 s is more than enough to ensure a complete nuclear
relaxation between two consecutive scans. Each free induction decay
was Fourier transformed with a line broadening of 0.5-1.5 Hz,
manually phased, and integrated after an accurate correction of the
spectrum baseline. For partially overlapped signals, the peak areas were
determined through the deconvolution routine of the Bruker NMR
software using a Lorentzian line shape.
The absolute values of the site specific (D/H) ratios were calculated
according to the formula:
(D/H)i ) nWS gWS MWL Si (D/H)WS/ni gL MWWS SWS
(1)
where WS stands for the working standard with a known isotope ratio
(D/H)WS and L stands for the product under examination; nWS and ni
are the number of equivalent deuterium atoms of the standard and of

9386 J. Agric. Food Chem., Vol. 53, No. 24, 2005
Brenna et al.
the i-th peak; gWS and gL are the weights of the standard and the sample;
MWL and MWWS are the corresponding molecular weights; Si and SWS
are the areas of the i-th peak and of the standard, respectively.
The signal assignment for compounds 10-12 and 14 was obtained
from the analysis of the proton spectrum taken in the same solvent
and at the same concentration of the deuterium spectrum. The signals
were assigned through chemical shift correlation experiments (COSY)
and determination of the nuclear Overhauser enhancements (NOESY).
The chemical shifts are expressed in ppm from internal TMS (δ), and
the coupling constants are expressed in hertz. Coumarin 10 (δ, acetone-
d6, Figure 1): 7.83 (Hâ), 7.54 (H6), 7.48 (H4), 7.22 (H3 + H5), 6.34
(HR); J3,4 ) J4,5 ) 7.9, J4,6 ) 1.8, J5,6 ) 7.7, JR,â ) 9.7. Dihydrocoumarin
11 (δ, acetone-d6, Figure 1): 7.21 (H4 + H6), 7.06 (H5), 6.94 (H3),
2.96 (CH2-â), 2.70 (CH2-R); J3,4 ) 8.3, J3,5 ) 1.4, J4,5 ) J5,6 ) 7.5,
JR,â ∼ 7.6). Methyl cinnamate 12 (δ, CD2Cl2, Figure 1): 7.71 (Hâ),
7.54 (H2,6), 7.38 (H3,4,5), 6.48 (HR), 3.79 (OCH3); JR,â ) 16.0. Diacetyl
salicyl alcohol 14 (δ, acetone-d6, Figure 1): 7.47 (H6), 7.39 (H4), 7.26
(H5), 7.14 (H3), 5.07 (CH2-7), 2.29, and 2.02 (two COCH3 groups);
J3,4 ) 8.2, J3,5 ) 1.3, J3,6 ) 0.5, J4,5 ) 7.6, J4,6 ) 1.8, J5,6 ) 7.6.
Isotopic Oxygen Determinations. The determination of δ18O of
aromatic compounds by IRMS is fast and does not generally require
any manipulation of the sample. The only constraint is posed by
unintentional absorption of water that must be avoided; this problem
can be easily prevented by keeping the samples in a dry chamber. For
the analytical determination of 18O, the samples were pyrolyzed in a
TC/EA high-temperature conversion elemental analyzer (Thermo-
Finnigan) and measured with a DeltaPlus XP. The method of measure-
ment of δ18O, due to the absence of nitrogen, which is the only element
that could cause an isobaric interference during the measurement of
CO, followed the standard settings suggested by ThermoQuest, pyro-
lyzing the samples in a ceramic tube filled with glassy carbon chips at
1450 °C.
The results of oxygen isotope ratio analyses are reported in per mil
(‰) relative to V-SMOW, defined as 0‰ point in the δ18O scale. The
precision (standard deviation) for analysis of the laboratory standards
(beet and cane sugar previously intercalibrated in collaboration with
15 different laboratories taking part in the European Project Sugar 18O-
SMT4-CT98-2219) is (0.18‰ (n ) 10). To evaluate the precision of
the analyses of the unknown samples, several replicates (n ) 5) were
performed giving as a result a standard deviation of (0.20‰, similar
to the values obtained for the internal standards. The rest of the samples
were analyzed in duplicate or triplicate and calculated against reference
gases calibrated with an international reference material (benzoic acid
IAEA-601, δ18O ) +23.3‰).
Figure 2. Natural abundance 2H NMR spectra of coumarin 10 (acetone):
(A) synthetic sample 10.4 and (B) extractive sample 10.1. Natural
abundance 2H NMR spectra of methyl cinnamate 12 (CH2Cl2): (C)
synthetic sample 12.1 and (D) natural sample 12.2.
the (D/H)6(ortho) (positions referred to the alkyl side chain) range
from 166.3 to 159.7 and from 133 to 128.1 ppm, respectively.
These values are higher than those previously determined for
the corresponding positions of salicin 8. Indeed, in the latter
instance, values around 139 and 107 ppm, respectively, were
measured (11). However, if we consider the ratio R4/6 between
(D/H)4(para) and (D/H)6(ortho) for the two biosynthetically related
products 8 and 10, mean values of 1.28 and 1.24 are obtained
(Table 1). For the synthetic samples, the same ratio is 1.03 and
1.09, respectively. Very recently, Le Grand et. al. (16) in a study
on the origin of methyl salicylate, a natural product sharing with
salicin and coumarin many steps of the biosynthetic pathway,
proposed the ratio R4/6 as a simple and significant parameter to
distinguish between samples of natural and synthetic derivation.
This parameter in the case of methyl salicylate showed values
around 1.0 for the synthetic samples and around 1.25 for the
natural ones, in good agreement with those determined by us
for salicin 8 (∆R4/6 ) 0.25). For coumarin, the difference
All of the results were calculated according to the following equation:
δ ‰ ) [(sample ratio - reference ratio)/reference ratio] × 1000
(2)
RESULTS AND DISCUSSION
Coumarin. Records concerning the natural occurrence of
coumarin 10 and of its dihydroderivative melilotol 11 date back
to the beginning of the chemical literature (13, 14). Both
products are present in a large number of plant extracts and
FaVa tonka absolute has entered food flavors as a source of
coumarin. This compound was one of the first synthetic
chemicals available, being produced by condensation of salicylic
aldehyde with acetic anhydride in the presence of sodium acetate
(15). The coumarin samples of the present study were natural
materials isolated by chromatography from commercially avail-
able F. tonka absolute and commercial synthetic samples.
Examples of the natural abundance 2H NMR spectra of
extractive and synthetic coumarin (samples 10.1 and 10.4) are
reported in Figure 2, while in Table 1 are included the (D/H)i
values relative to the five examined samples.
between natural and synthetic samples is less marked (∆R4/6
0.15) but still significant.
)
The signals of H-3 and H-5 are unresolved in coumarin 10.
However, the (D/H)H-5+H-3(meta) remains in the narrow range
of 157.7-155.2 ppm for all three extractive samples. Strangely,
the two synthetic samples, 10.4 and 10.5, showed different
aromatic deuterium patterns: (D/H)6(ortho) ) 146.9 and 125.5
ppm and (D/H)4(para) ) 158.6 and 138.1 and (D/H)H-5+H-3(meta)
) 167.5 and 130.6 ppm, respectively. The deuterium signals
At first, we consider the deuterium values of positions 4 and
6 of 10, i.e., those not activated toward acid-catalyzed proton
exchange. In natural samples 10.1-10.3, the (D/H)4(para) and

Isotopic Characterization of Coumarin and Melilotol
J. Agric. Food Chem., Vol. 53, No. 24, 2005 9387
Table 2. Origin and (D/H)i Isotopic Ratios (Parts per Million)a of Methyl Cinnamate 12
b
sample (origin)
(D/H)2,6
(D/H)3,4,5
(D/H)Ar
(D/H)
(D/H)
(D/H)OMe
â
R
12.1 (synthetic from Fluka)
12.2 (natural commercial from H&R)
150.5 (1.7)
126.3 (0.7)
144.0 (2.4)
150.4 (1.7)
146.6
140.7
515.8 (1.9)
127.2 (1.2)
121.6 (3.7)
131.2 (5.1)
143.8 (2.6)
121.1 (2.3)
a The (D/H)i standard deviations are repoted within parentheses. b (D/H)Ar is the mean deuterium content of the aromatic nuclei calculated from the individual positional
(D/H)i.
of the double bond of the extractive samples 10.1-10.3 differed
were 126.3 and 150.4 ppm, respectively. This result confirms
the hypothesis that the aromatic moiety of the synthetic product
is labeled in a uniform way, whereas the sugar-derived aromatic
framework of 12.2 is characterized by a deuterium enrichment
for the positions 2 and 6 much smaller that the one of the three
remaining hydrogen atoms in positions 3, 4, and 5. Interestingly,
the (D/H)â of 12.1 was 515.8 ppm, dramatically higher than
the 127.2 ppm value of the natural material 12.2. On the other
hand, the (D/H)R values of the two samples were 121.6 and
131.2 ppm, values consistent with the ones observed in the
corresponding position in samples 10.1-10.5 (Table 1). The
significant deuterium enrichment for positions â in synthetic
10 and 12, as compared with the natural counterparts, might be
the consequence of a kinetic isotopic effect in the condensation,
which drives the conversion of salicyl aldehyde and benzalde-
hyde into 10 and 12, respectively. It is worth noting that the
(D/H)R values of extractive coumarin 10 and of natural methyl
cinnamate 12 are all in the range of 131.2-104.6 ppm. Samples
of L-phenylalanine of natural origin display in the corresponding
position (D/H) values of the same magnitude (9).
Melilotol. Melilotol 11 is present in small amounts in many
essential oils, and it is the component responsible for the
therapeutic effect of the extract of Melilotus officinalis (L.) Lam.
(Leguminosae). The preparation of aliquots of 11 by enzymic
reduction of easily accessible extractive coumarin 10 might be
economically interesting, because the material produced in this
way is considered “natural”. In this case, the determination of
the profile of the deuterium natural abundance of 11 should be
a sound method to demonstrate the origin of the product
obtained by different methods. Accordingly, 11 was prepared
by chemical hydrogenation and baker’s yeast reduction of the
coumarin sample 10.1 (samples 11.1 and 11.2, respectively).
Chemical hydrogenation of the synthetic coumarin 10.4 provided
the third material included in the study (sample 11.3).
The (D/H)i values of samples 11.1-11.3 are reported in Table
1. As expected, the pattern of the aromatic moiety of 11
corresponded to that of the parent unsaturated precursors. The
part of the spectrum that might be of some interest in the
differentiation of the various samples is the one concerning the
CH2-R and CH2-â. The data indicate that samples 11.1 and 11.2
obtained by chemical and biochemical hydrogenation, respec-
tively, have almost identical values, and therefore, they are
indistinguishable one from the other. It should be noted that
the (D/H)â and (D/H)R values of 11.1-11.3 are smaller than
those of their precursors 10.1 and 10.4. The lower deuterium
content exhibited by the side chain methylene groups of samples
11.1 and 11.3, obtained by catalytic reduction, is due to the
combined effect of the isotopic composition of the hydrogen
employed, a gas produced from methane of petrochemical
origin, depleted in deuterium and of a deuterium kinetic isotope
effect. In the biological reduction of the carbonyl-activated
double bond (sample 11.2), it is expected that the R-hydrogen
derives from water, while the one in the â-position is delivered
by the intermediacy of the reduced nicotine cofactor(s) (18). A
deuterium kinetic isotope effect is expected to be present for
both processes (19).
substantially from 10.4 and 10.5, obtained by synthesis. Specif-
ically, samples 10.1-10.3 showed a (D/H)CH-â around 131 ppm.
The corespective values in samples 10.4 and 10.5 were 258.5
and 189.8 ppm, respectively. When the R-position is considered,
the two sets appear less differentiated.
More detailed information on the aromatic deuterium pattern
of 10 comes from the examination of the spectra of 14, the
diacetyl derivative of salicyl alcohol 9, the material used in the
characterization of salicin 8, showing separated signals for the
four aromatic hydrogen atoms (11). The (D/H)i values of
samples 14.1-14.4, prepared from the coumarin samples 10.1-
10.4, are reported in Table 1. The total (D/H)aromatic is in
satisfactory agreement with the data of the precursors. In detail,
the (D/H)6(ortho) and the (D/H)4(para) in 14.1-14.4 were quite
similar to those observed in 10.1-10.4, while the (D/H)5(meta)
and the (D/H)3(meta), for the extractive samples, fell around 134
and 172 ppm, respectively. Similarly, in the synthetic sample
14.4, the corresponding values were 151.0 and 167.2 ppm,
respectively. Thus, in all of these samples, position 3 of the
salicyl moiety, which is ortho to the phenolic oxygen, is more
deuterium enriched than position 5, para to the activating group.
These positions for the two samples of 14 derived from
extractive salicin 8 showed a much more similar degree of
deuterium labeling, since the (D/H)3(meta) and the (D/H)5(meta)
were 127 and 119 ppm and 134 and 115 ppm, respectively.
The (D/H)7 values relative to the benzylic position are without
significance in the considered situation because one of the two
hydrogen atoms comes from the NaBH4 used in the reduction
of the intermediate ozonide.
So far, it seems quite difficult to attribute a biosynthetic
relevance to the fact that in natural coumarin position 3, ortho
to the oxygen atom introduced in the framework of cinnamic
acid 4 when it is converted into o-coumaric acid 7, is more
deuterium-enriched than position 5, symmetrical to position 3
in precursor 4, because the same is not true in salicin 8. It is
worth noting that in our exchange experiments, in which salicyl
alcohol was incubated with deuterated water in the presence of
mineral acid, position 3 was more rapidly labeled than position
5 (11).
In these experiments, the observation made on salicin and
on methyl salicylate (16) was, however, confirmed as follows:
H-4 (para) is more deuterium-enriched than H-6 (ortho). As a
general trend, it can be deduced that the ratio R4/6 of the natural
o-hydroxylated phenylpropanoids (range 1.19-1.30) is greater
than that of the synthetic samples (range 1.0-1.11) (Table 1).
This behavior is partially similar to that of the natural phenyl-
propanoids, for which as a consequence of the modality of
labeling of the sugar fragments, providing shikimic acid (17),
the deuterium enrichment follows the trend para > ortho > meta.
2
Support of this view derives from the H NMR analyses
carried out on methyl cinnamate 12. The spectra of the synthetic
and natural samples 12.1 and 12.2 are reported in Figure 2,
and the (D/H)i values are collected in Table 2. The synthetic
sample 12.1 shows a (D/H)2,6(ortho) ) 150.5 and (D/H)3,4,5(meta+para)
) 144.0 ppm, while for natural 12.2 the corresponding values

9388 J. Agric. Food Chem., Vol. 53, No. 24, 2005
Brenna et al.
LITERATURE CITED
(1) Schmidt, H.-L.; Werner, R. A.; Eisenreich, W. Systematics of
2H patterns in natural compounds and its importance fort he
elucidation of biosynthetic pathways. Phytochem. ReV. 2003, 2,
61-85.
(2) Schmidt, H.-L.; Werner, R. A.; Rossman, A. 18O Pattern and
biosynthesis of natural plant products. Phytochemistry 2001, 58,
9-32.
(3) Martin, G. J.; Martin, M. L. Stable isotope analysis of food and
beverages by NMR spectroscopy. Annu. Rep. NMR Spectrosc.
1995, 31, 91-104.
Figure 3. Conversion of 10 into 13 by hydrogenation, reduction, and
cyclization.
(4) Remaud, G. S.; Martin, Y. L.; Martin, G. G.; Martin, G. J.
Detection of sophisticated adulterations of natural vanilla flavors
and extracts: Application of the SNIF-NMR method to vanillin
and p-hydroxybenzaldehyde. J. Agric. Food Chem. 1997, 45,
859-866.
(5) Fronza, G.; Fuganti, C.; Serra, S.; Burke, A.; Guillou, C.; Reniero,
F. The positional δ(18O) values of extracted and synthetic
vanillin. HelV. Chim. Acta 2001, 84, 351-358.
Table 3. Positional and Global
δ
18O Values (‰) for Coumarin 10
sample (origin)
Cd
O
ArO
total
10.1 (natural from F. tonka)
10.2 (natural from Charabot)a
10.3 (natural from Charabot)a
10.4 (synthetic from Fluka)
38.3
46.7
45.7
19.6
2.3
0.5
20.3
23.6
22.6
12.6
−0.5
5.6
(6) Fronza, G.; Fuganti, C.; Pedrocchi-Fantoni, G.; Serra, S.; Zucchi,
G.; Fauhl, C.; Guillou, C.; Reniero, F. Stable isotope character-
ization of raspberry ketone extracted from Taxus baccata and
obtained by oxidation of the accompanying alcohol (betuligenol).
J. Agric. Food Chem. 1999, 47, 1150-1155.
(7) Fronza, G.; Fuganti, C.; Serra, S.; Cisero, M.; Koziet, J. Stable
isotope labeling of resveratrol and related natural stilbenes. J.
Agric. Food Chem. 2002, 50, 2748-2754.
a Different lots.
Determination of the Positional δ18O Values of Coumarin.
To complete the isotopic characterization of coumarin 10, the
positional δ18O values of extractive and synthetic samples were
determined. For this purpose, 10 was converted to chroman 13,
which retains the aromatic oxygen atom of 10. The transforma-
tion was performed chemically, hydrogenating 10 to dihydro-
coumarin 11. The latter compound, on LiAlH4 reduction,
provided 3-(2-hydroxyphenyl)propan-1-ol. This material, treated
with 4-toluenesulfonyl chloride in pyridine, gave the primary
tosylate, whose ring closed to chroman 13 under basic conditions
(Figure 3). Samples 10.1-10.4 and samples 13.1-13.4 were
subsequently submitted to the isotopic oxygen determinations.
The bulk and positional δ18O values of coumarin samples 10.1-
10.4 are reported in Table 3. The numerical data representing
the δ18O values of the carbonyl carbon were calculated from
the δ18O values of 10 and 13.
(8) Haslam, E. Shikimic Acid. Metabolism and Metabolites; Wiley:
Chicester, United Kingdom, 1993.
(9) Brenna, E.; Fronza, G.; Fuganti, C.; Pinciroli, M. Differentiation
of natural and synthetic phenylalanine and tyrosine through
natural abundance 2H nuclear magnetic resonance. J. Agric. Food
Chem. 2003, 51, 4866-4872.
(10) Fronza, G.; Fuganti, C.; Schmidt, H.-L.; Werner, R. A. The δ18
O-value of the p-OH group of L-tyrosine permits the assignment
of its origin to plant or animal sources. Eur. Food Res. Technol.
2002, 215, 55-58.
(11) Brenna, E.; Fronza, G.; Fuganti, C.; Gatti, F. G.; Pinciroli, M.;
Serra, S. Differentiation of extractive and synthetic salicin: The
2H aromatic pattern of natural 2-hydroxybenzyl alcohol. J. Agric.
Food Chem. 2004, 52, 7747-7751.
A closer look at the whole set of values of 10 allows the
identification of two distinct groups, represented by samples
10.1-10.3 of extractive origin (mean value 22.1‰) and by the
sample 10.4 produced by chemical synthesis (12.6‰). Taking
into consideration also the positional values, the first set is
characterized by an isotope ratio mean value of 0.6‰ of the
aromatic oxygen atom, while a value of 5.6‰ is observed for
the synthetic material 10.4. Moreover, the carbonyl oxygen in
samples 10.1-10.4 displays a mean value of 43.5‰, consider-
ably higher than the value of 19.6‰ of sample 10.4. Therefore,
the positional δ18O values are relevant for the determination of
the origin of coumarin. Additionally, the present data provide
information on the possible mechanism of introduction of the
oxygen function in the aromatic ring of cinnamic acid 4 during
the conversion into o-coumaric acid 7. The mean δ18O value
of 0.6‰ here observed for the aromatic oxygen atom of
extractive 10 is consistent with the values of phenolic oxygen
of p-coumarate-derived products such as raspberry ketone 2,
-0.8 and +0.6‰ (6), and vanillin 1, 5.3 and 6.3‰ (5),
respectively, thus suggesting a similar mechanism for the oxygen
activation, i.e., the direct introduction of the O-atom into the
aromatic ring from O2 through the action of a specific enzyme.
Seen together, the results of the multiple isotope characterization
of coumarin 10 and related materials so far reported are not
only useful for differentiating extractive natural products from
the synthetic ones but also provide insights into the means of
formation of natural products.
(12) Bauer, K.; Garbe, D.; Surburg, H. Common Fragrance and
FlaVor Materials; Wiley-VCH: Weinheim, Germany, 2001; p
221.
(13) Bleitreu, H. Untersuchungen u¨ber Cumarin. Liebig’s Ann. 1846,
59, 177-198.
(14) Phipson, T. L. U¨ ber das melilotol. Chemisches Central-Blatt.
1875, 646.
(15) Perkin, W. H. Bildung von Cumarin und zimmtsa¨ure, und
anderen analogen sa¨ure aus aromatischen aldehyden. Ber. 1877,
10, 299-300.
(16) Le Grand, F.; George, G.; Akoka, S. Natural abundance 2H-
ERETIC NMR authentication of the origin of methyl salicylate.
J. Agric. Food. Chem. 2005, 53, 5125-5129.
(17) Walsh, C. Enzymic Reactions Mechanism; Freeman: San
Francisco, CA, 1979; pp 923-929.
(18) Sedgwick, B.; Morris, C. Stereohemical course of hydrogen
transfer catalysed by enoyl reductase enzymes of the yeast fatty
acid synthetase. J. Chem. Soc., Chem. Commun. 1980, 96-97.
(19) Fronza, G.; Fuganti, C.; Guillou, C.; Reniero, F.; Joulain, D.
2
Natural abundance H nuclear magnetic resonance study of the
origin of raspberry ketone. J. Agric. Food Chem. 1998, 46, 248-
256.
Received for review July 28, 2005. Revised manuscript received October
5, 2005. Accepted October 5, 2005. COFIN “Aromi e Fragranze” is
acknowledged for partial financial support.
JF0518507