Differentiation of natural and synthetic phenylalanine and tyrosine through natural abundance 2H nuclear magnetic resonance

Article abstract of DOI:10.1021/jf0302759

The natural abundance deuterium NMR characterization of samples of the amino acids tyrosine (1) and phenylalanine (2), examined as the acetylated methyl esters 4 and 6, has been performed with the aim to identify by these means the contribution in animals of the hydroxylation of the diet L-phenylalanine (2) to the formation of L-tyrosine (1), a feature previously revealed on the same samples through the determination of the phenolic δ¹⁸O values. The study, which includes also the NMR examination of benzoic acid (5) from 2 and of tyrosol (7) from 1, substantially fails in providing the required information because the mode of deuterium labeling of tyrosine samples of different origins is quite similar but indicates a dramatic difference in the deuterium labeling pattern of the two amino acids 1 and 2. The most relevant variation is with regard to the deuterium enrichments at the CH₂ and CH positions, which are inverted in the two amino acids of natural derivation. Moreover, whereas the diastereotopic benzylic hydrogen atoms of L-tyrosine (1) appear to be equally deuterium enriched, in L-phenylalanine (2) the (D/H)_3R > (D/H)_3S. Similarly, benzoic acid (5) shows separate signals for the aromatic deuterium nuclei, which are quite indicative of the natural or synthetic derivation. The mode of deuterium labeling of the side chain of 1 and 2 is tentatively correlated to the different origins of the two amino acids, natural from animal sources for L-tyrosine and biotechnological probably from genetically modified microorganisms for L-phenylalanine.

Full text of DOI:10.1021/jf0302759

4866 J. Agric. Food Chem. 2003, 51, 4866−4872
Differentiation of Natural and Synthetic Phenylalanine and
Tyrosine through Natural Abundance 2H Nuclear Magnetic
Resonance
ELISABETTA BRENNA, GIOVANNI FRONZA,* CLAUDIO FUGANTI, AND
MATTEO PINCIROLI
Dipartimento di Chimica, Materiali e Ingegneria Chimica “Giulio Natta”,
Politecnico di Milano and CNR, Istituto di Chimica del Riconoscimento Molecolare,
Sezione “A. Quilico”, Via Mancinelli 7, 20131 Milano, Italy
The natural abundance deuterium NMR characterization of samples of the amino acids tyrosine (1)
and phenylalanine (2), examined as the acetylated methyl esters 4 and 6, has been performed with
the aim to identify by these means the contribution in animals of the hydroxylation of the diet
L-phenylalanine (2) to the formation of L-tyrosine (1), a feature previously revealed on the same
samples through the determination of the phenolic δ18O values. The study, which includes also the
NMR examination of benzoic acid (5) from 2 and of tyrosol (7) from 1, substantially fails in providing
the required information because the mode of deuterium labeling of tyrosine samples of different
origins is quite similar but indicates a dramatic difference in the deuterium labeling pattern of the two
amino acids 1 and 2. The most relevant variation is with regard to the deuterium enrichments at the
CH2 and CH positions, which are inverted in the two amino acids of natural derivation. Moreover,
whereas the diastereotopic benzylic hydrogen atoms of L-tyrosine (1) appear to be equally deuterium
enriched, in L-phenylalanine (2) the (D/H)3R > (D/H)3S. Similarly, benzoic acid (5) shows separate
signals for the aromatic deuterium nuclei, which are quite indicative of the natural or synthetic
derivation. The mode of deuterium labeling of the side chain of 1 and 2 is tentatively correlated to the
different origins of the two amino acids, natural from animal sources for L-tyrosine and biotechnological
probably from genetically modified microorganisms for L-phenylalanine.
KEYWORDS: Tyrosine; phenylalanine; tyrosol; animal vegetal origin; aspartame; natural abundance
deuterium NMR
INTRODUCTION
Adulteration of foodstuff is a matter of concern. Among the
analytical methods developed up to now to address this issue,
the determination of the stable isotope composition of relevant
food components seems to be particularly effective (1, 2). We
have recently reported a study designed to define the animal
rather than plant origin of the amino acid L-tyrosine (1) through
the determination of the δ18O values of the phenolic oxygen
atom (3). The premise of this method is the existence in nature
of two pathways regulating the formation of L-tyrosine (1),
characterized by different sources for the phenolic oxygen, that
is, leaf water versus molecular oxygen (4). The key advanced
intermediate in the sequence starting from shikimic acid and
phosphoenolpyruvate is arogenic acid (3), which provides both
L-tyrosine (1) and L-phenylalanine (2) (Figure 1). L-Tyrosine
(1) is obtained from 3 by loss of carbon dioxide and two
hydrogen atoms. Conversely, L-phenylalanine (2) is formed
when carbon dioxide is lost together with a molecule of water.
Figure 1. Biosynthetic pathways for the formation of tyrosine (1) and
phenylalanine (2) from arogenic acid (3).
However, animals, for which L-tyrosine is not essential, are able
to directly hydroxylate 2 to 1 (Figure 1), the extent of this
activity depending upon the content of 1 in the diet (5). The
δ18O values of the phenolic oxygen of tyrosine and derivatives
suggest that L-tyrosine from animals is a mixture of two
* Corresponding author (telephone +390223993027; fax +390223993080;
e-mail giovanni.fronza@polimi.it).
10.1021/jf0302759 CCC: $25.00 © 2003 American Chemical Society
Published on Web 07/17/2003

Natural Abundance 2H NMR of Phenylalanine and Tyrosine
J. Agric. Food Chem., Vol. 51, No. 17, 2003 4867
pyridine (15 g; 0.189 mol). After 24 h at room temperature, the reaction
mixture was poured onto crushed ice containing NaHCO3 (5 g; 0.059
mol). The organic phase was separated, and the aqueous phase was
extracted with CH2Cl2 (100 mL × 2). The combined organic extract,
washed twice with cold 5% HCl (100 mL), saturated aqueous NaHCO3,
and water (100 mL each), was dried (Na2SO4) and evaporated. The
residue was crystallized from hexane-ethyl acetate to give the desired
N-acetylphenylalanine methyl ester 4 (3 g; 75%): 1H NMR (CH2Cl2)
δ 7.30-7.14 (5H, m, phenyl), 6.63 (1H, s br, NH), 4.79 (1H, m, CH-
2), 3.68 (3H, s, OCH3), 3.13 (1H, dd, J ) 5.9 and 13.6 Hz, H3R), 3.02
(1H, dd, J ) 7.1 and 13.6 Hz, H3S), 1.92 (3H, s, COCH3). L-
Phenylalanine samples from aspartame (R-L-aspartyl-L-phenylalanine
methyl ester) were obtained as follows: aspartame (9.8 g; 0.033 mol)
was refluxed for 6 h with 150 mL of 20% aqueous NaOH. The cooled
solution was brought to pH 5-6 with 20% HCl. The precipitated
L-phenylalanine was collected by suction and repeatedly washed on
the filter with water (4.5 g; 83%).
Benzoic Acid (5) from Phenylalanine (2). The oxidation of
phenylalanine samples was performed in alkaline aqueous solution with
3% potassium permanganate as previously reported for phenylacetic
acid (9) (yeld ∼90%): 1H NMR of 5 (DMSO) δ 7.92 (2H, m, Hortho),
7.53 (1H, m, Hpara), 7.42 (2H, m, Hmeta).
N,O-Diacetyltyrosine Methyl Ester (6). The transformation of
tyrosine 1 into 6 was performed as reported for the preparation of 4
from 2 (yield ) 84%): 1H NMR of 6 (CH3CN) δ 7.20 (2H, d, J ) 8.9
Hz, Hortho), 7.00 (2H, d, J ) 8.9 Hz, Hmeta), 6.74 (1H, s br, NH), 4.62
(1H, td, J ) 5.9 and 7.7 Hz, CH-2), 3.64 (3H, s, OCH3), 3.09 (1H, dd,
J ) 5.9 and 13.6 Hz, H3R), 2.95 (1H, dd, J ) 7.7 and 13.6 Hz, H3S),
2.21 (3H, s, COCH3), 1.84 (3H, s, COCH3).
Baker’s Yeast Conversion of Tyrosine (1) into Tyrosol (7). To a
mixture made up with Baker’s yeast (Distillerie Italiane, San Pancrazio,
Italy) (2 kg) and D-glucose (200 g), in tap water (3 L), under stirring
at ∼30 °C was added within 3 h a solution of tyrosine 1 (3 g; 0.016
mol) (6 g in the case of the D,L compound], in tap water (600 mL),
treated with enough 10% aqueous NaOH to obtain complete dissolution.
After 48 h of stirring, acetone (2 L) was added, and the mixture was
filtered by suction through a large Bu¨chner funnel using a Celite pad.
The cake was repeatedly washed with acetone. The filtrate was
concentrated under vacuum to ∼1 L volume. The aqueous phase was
extracted with ethyl acetate (300 mL × 3). The residue obtained upon
evaporation of the solvent was chromatographed on a silica column
with increasing amounts of ethyl acetate in hexane, tyrosol (7) (45%)
eluting with a ∼1:1 mixture, which was crystallized from hexane-
ethyl acetate (3): 1H NMR of 7 (acetone) δ 7.98 (1H, s br, OH), 6.95
(2H, d, J ) 8.4 Hz, Hortho), 6.64 (2H, d, J ) 8.4 Hz, Hmeta), 3.60 (2H,
t, J ) 7.1 Hz, CH2OH), 2.62 (2H, t, J ) 7.1 Hz, CH2).
Figure 2. Products submitted to natural abundance 2H NMR measurement.
materials of different origins, having, as a characteristic
fingerprint, quite distinct δ18O values. One of these is that of
plant origin, acquired through the diet, directly formed from
arogenic acid (3) and retaining as its phenolic oxygen atom that
originally present in position 4 of shikimic acid; the other is
that produced in the body from L-phenylalanine (2) of the diet
by hydroxylation, thus incorporating atmospheric oxygen.
The diagnostic potential of site-specific deuterium distribution
data obtained by natural abundance 2H NMR measurements in
the definition of the origin of food components is well-known
(2). This technique was applied in previous studies to the
characterization of several amino acids, showing that wide
variations of deuterium content occur in samples of different
2
origins (6, 7). We now present a H NMR study of L-tyrosine
(1) and L-phenylalanine (2) performed with the aim to identify
the relative proportions of the material directly derived from
arogenic acid and that produced via phenylalanine by hydroxyl-
ation in tyrosine samples extracted from animal sources.
MATERIALS AND METHODS
Samples. A total of 31 samples were examined. These included eight
phenylalanine samples [sample 1, commercial L-phenylalanine certified
as natural; sample 2, D,L-phenylalanine from Fluka (Milano, Italy);
sample 3, L-phenylalanine of undefined origin; sample 4, D,L-phenyl-
alanine from Aldrich (Milano, Italy); sample 5, L-phenylalanine from
aspartame from HSC; sample 6, l-phenylalanine from aspartame from
Miwon; sample 7, L-phenylalanine from aspartame from Ajinomoto;
sample 8, L-phenylalanine from aspartame from Caremoli (all aspartame
samples were gifts from Perfetti, Lainate, Italy)], eight benzoic acid
samples (samples 1-8) produced by oxidation of phenylalanine samples
1-8, nine tyrosine samples [sample 1, D,L-tyrosine from Merck (VWR
International, Milano, Italy); sample 2, L-tyrosine from hen feathers;
sample 3, D,L-tyrosine from Fluka (Milano, Italy); sample 4, L-tyrosine
from Fluka (Milano, Italy); sample 5, L-tyrosine from Fluka (different
production batch, Milano, Italy); samples 6-8, L-tyrosine from human
hair (three different batches); sample 9, L-tyrosine of unspecified animal
origin], and six tyrosol samples [sample 1, tyrosol from Ligustrum
oValifolium; samples 2-4 and 6, tyrosol obtained from tyrosine samples
1, 5, 2, and 6, respectively; sample 5, tyrosol from L-tyrosine from
Aldrich (Milano, Italy)].
NMR Experiments. The 2H NMR experiments were performed on
a Bruker Avance 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 298 K for
phenylalanine and tyrosine derivatives 4 and 6 and at 328 K for the
acids 5 and 7. 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. The spectra
were recorded by dissolving 0.4-0.7 g of material in 2.5-3.0 mL of
solvent, adding 70 µL of C6F6 for the lock and 100-130 mg of tert-
butyl disulfide as internal standard [(D/H) 129.2 ppm]. The solvents
used were CH2Cl2 for 4, CH3CN for 6, DMSO for 5, and acetone for
7.
At least five spectra were run for each sample, collecting 4000-
8000 scans to reach an S/N > 100 (methyl signals) and using the
following parameters: 6.8 s acquisition time, 1200 Hz spectral width,
and 16K memory size. Each FID was Fourier transformed with a line
broadening of 1.5-2 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.
Internal isotopic ratios Rij are defined as
Solubility problems prevented the direct acquisition of the NMR
spectra of the two amino acids as such. Accordingly, 1 and 2 were
converted through reported procedures into the corresponding highly
soluble acetylated methyl esters 6 and 4, respectively (Figure 2).
N-Acetylphenylalanine Methyl Ester 4. In a typical experiment
(8), phenylalanine (3 g; 0.018 mol) was added at 0 °C in portions under
stirring to methanol (100 mL) to which had previously been added
SOCl2 (8 mL). The mixture was kept at room temperature for 24 h
and then evaporated to dryness under vacuum. The residue was taken
up in methanol and evaporated, repeating the operation three times.
The solid residue was suspended in CH2Cl2 (100 mL) and treated under
stirring at 0 °C with acetic anhydride (10 g; 0.098 mol), followed by
Rij ) njSi/Sj
(1)
where Si is the area of the ith site, Sj is the area of a reference peak,

4868 J. Agric. Food Chem., Vol. 51, No. 17, 2003
Brenna et al.
and nj is the number of isochronous hydrogens at the j site. A statistical
distribution of deuterium among the n molecular sites would originate
Rij factors equal to the number of hydrogens of the corresponding ith
sites.
The absolute values of the site-specific (D/H) ratios were calculated
according to the formula
(D/H)i ) nWSgWS(MW)LSi(D/H)WS/(nigL(MWWS)SWS
)
(2)
where WS stands for the working standard with a known isotope ratio
(D/H)WS and L for the product under examination; nWS and ni are the
number of equivalent deuterium atoms of the standard and of the ith
peak; gWS and gL are the weights of the standard and the sample; MWL
and MWWS are the corresponding molecular weights; and Si and SWS
are the areas of the ith peak and of the standard, respectively.
13C Isotopic Mass Determination. The mass spectrometry deter-
mination of the 13C/12C isotope ratios (IRMS) was determined using a
Sira II-VG Fisons (Rodano, Italy) mass spectrometer interfaced with a
Carlo Erba (Rodano, Italy) NA 1500 elemental analyzer for sample
combustion. The measurements are expressed in per mil (‰) versus
an international standard (the carbonate Pee Dee Belemnite, PDB)
according to the equation
Figure 3. Natural abundance 2H NMR spectra of phenylalanine derivative
4: (a) commercial sample 2; (b) natural sample 1; (/) signal of calibrated
tert-butyl disulfide used as internal standard; (O) CH Cl2 solvent signal.
2
δ13C (‰) ) (Rsample/Rstandard - 1) × 103
(3)
which is now mainly based on microbial production with
genetically modified organisms using carbohydrates as the
carbon source.
where R is the 13C/12C ratio.
Benzoic Acids 5. The 2H NMR spectrum of the phenylalanine
derivatives 4 provides an unresolved signal for the aromatic
nuclei (Figure 3). Accordingly, the phenylalanine samples were
converted into the corresponding benzoic acids 5, which show
much more dispersed signals allowing the differences occurring
in the aromatic part of the molecule to be revealed.
RESULTS AND DISCUSSION
Origin of the Samples Examined. Concerning the origin of
the amino acids, it has to be pointed out that tyrosine (1) is
rather difficult to obtain by chemical synthesis. Accordingly,
its world demand (100 tons/year) is apparently completely
satisfied by the product obtained by acid hydrolysis of keratin-
rich materials of animal origin. Currently, human hairs and hen
feathers constitute the sources of tyrosine-rich proteins, as a
substitute for those previously obtained from other animals. The
commercially available D,L-form is manufactured by racemiza-
tion of the L-enantiomer. In the study on the origin of tyrosine
(1) based on the determination of the δ18O values of the phenolic
oxygen, a sample of tyrosol (7) (Figure 2) extracted from
Ligustrum oValifolium was also included. This material is a
direct metabolite of 1, and in the above context it formally
represents the equivalent of an authentic sample of “plant”
tyrosine. To compare the 2H NMR properties of the samples of
tyrosine (1) here examined with those of plant origin 7, selected
samples of 1 were converted into tyrosol (7) by fermentation
with baker’s yeast. The biotransformation, which implies
decarboxylation and deamination of 1 with the introduction of
one or two new hydrogen atoms in position 1 of 7, depending
on the mechanism of the reaction (10, 11), leaves the atoms in
position 2 untouched. However, the deuterium enrichments at
position 1 of the samples of tyrosol (7) generated from 1 were
not considered in the present study as an element of differentia-
tion.
Quite different from that of tyrosine (1) is the manufacturing
source of L-phenylalanine (2). Before the advent of aspartame
(R-L-aspartyl-L-phenylalanine methyl ester) as an artificial
sweetener, 2 was required by the market in amounts similar to
those of tyrosine 1. The L-material was obtained by extraction
from protein sources or by enzymic hydrolysis of N-acyl
derivatives of the racemic form produced by chemical synthesis
using benzaldehyde as a source of the C-6sC-1 part of the
molecule. The sudden requirement for L-phenylalanine (2)
(∼3000 tons in 1995) for the manufacture of aspartame
necessarily required a modification of the manufacturing process,
The properties of the deuterium isotopic distribution for the
three different positions (i.e., ortho, meta, and para) of the phenyl
ring present in natural compounds are well-known from previous
studies carried out on important food flavors such as phenyl-
acetic acid (9) and natural benzaldehyde (12-14). It was shown
that for such natural material the isotopic enrichment follows
the trend para > ortho > meta. In contrast, samples of synthetic
origin show very similar deuterium content in all positions. Such
different behavior can be shown also by using appropriate
indicators such as the ratio between the molar fraction at the
ortho position and the sum of the molar fractions at the meta
and para positions [fortho/(fmeta + fpara)] (12) or the internal
isotopic ratios Rij (2). In Table 1 are reported the (D/H) values,
the ratio of the molar fractions and the internal Rmeta/para ratio
for the examined samples 1-8 of benzoic acids obtained by
degradation of the corresponding phenylalanines. As expected,
the synthetic samples 2 and 4 show quite similar deuterium
contents in the three positions; the indicators fortho/(fmeta + fpara
)
and Rmeta/para display values near 0.667 and 2, respectively,
characteristic of a statistical distribution of deuterium nuclei
among the five positions of the phenyl ring. Very similar results
were obtained previously for synthetic benzoic acids and
benzaldehydes (12, 13).
All other samples in Table 1 show deuterium contents
following the trend para > ortho > meta, except for sample 1
in which the ortho and meta positions are nearly equally
populated. The indicators fortho/(fmeta + fpara) and Rmeta/para are
very different with respect to those of synthetic samples. Very
marked changes are shown by the ratio Rmeta/para ranging from
1.2 to 1.58. The ratio fortho/(fmeta + fpara) display values in the
range of 0.60-0.65 (less than the statistical value) for samples
1, 3, 5, and 7, or ∼0.72 (greater than the statistical value) for
samples 6 and 8. Interestingly, similar values were found also

Natural Abundance 2H NMR of Phenylalanine and Tyrosine
J. Agric. Food Chem., Vol. 51, No. 17, 2003 4869
Table 1. Origin, (D/H) Isotopic Ratios (ppm), Deuterium Molar Fraction Ratios, Deuterium Internal Isotopic Ratios, and δ13C (‰) Global Isotope
i
Ratios of Benzoic Acids 5a from Phenylalanine (2)
sample
origin
(D/H)
(D/H)ortho
(D/H)para
(D/H)meta
fo/(fm + fp)
Rm/p
δ13C (‰)
total
1
2
3
4
5
6
7
8
natural (certified)
D,L Fluka
anonymous
141.8
137.4
143.0
137.8
128.3
134.1
150.1
129.1
132.6
139.2
141.0
138.9
125.4
141.5
144.4
134.3
169.1
130.6
162.8
128.8
153.7
175.4
180.2
152.3
134.3
139.1
134.0
141.2
119.1
106.1
140.9
112.3
0.61
0.68
0.65
0.67
0.64
0.73
0.62
0.71
1.59
2.13
1.65
2.19
1.55
1.21
1.56
1.47
−25.4
−27.3
−26.3
−29.4
−26.7
−13.8
−25.8
−12.4
D,L Aldrich
aspartame HSC
aspartame Miwon
aspartame Ajinomoto
aspartame Caremoli
a Mean calculated standard deviation of (D/H) values is 2.7.
i
Table 2. Origin and (D/H) Isotopic Ratios (ppm) of Phenylalanine Derivatives 4a
i
b
sample
origin
(D/H)Ph
(D/H)2
(D/H)3
(D/H)H-3S
(D/H)H-3R
(D/H)
total
1
2
3
4
5
6
7
8
natural (certified)
D,L Fluka
anonymous
144.0
135.0
145.0
130.7
134.8
141.1
148.1
136.0
128.0
98.5
75.0
102.5
127.5
118.4
110.6
122.1
283.0
118.0
323.0
90.6
352.7
141.2
118.0
131.1
167.4
114.7
254.4
88.7
326.3
119.7
112.8
105.0
374.8
113.9
393.5
92.6
391.3
159.3
127.4
142.4
185.0
117.2
181.0
107.9
205.0
133.6
125.6
129.7
D,L Aldrich
aspartame HSC
aspartame Miwon
aspartame Ajinomoto
aspartame Caremoli
a Mean calculated standard deviation of (D/H) values is 2.5. b (D/H) values are calculated from the individual positional deuterium contents.
i
total
for natural benzaldehyde samples obtained from bitter almond
oil (0.58) and from cassia oil (0.75) (12).
found (7). Among them in particular several samples of alanine
and proline displayed global enrichments in the range 165-
180 ppm, not too far from the values found here for phenyl-
alanine samples 1, 3, and 5.
The data in Table 1 allow the synthetic or natural origin of
the examined benzoic acids and consequently also the origin
of the starting phenylalanines to be assigned with confidence.
In particular, the values of the benzoic acids 5-8 suggest that
the phenylalanines used for the synthesis of the corresponding
aspartame samples are of natural derivation.
In addition the methylene nuclei of the phenylalanine
framework are diastereotopic and give rise to two separate NMR
signals, which show strong relative intensity variations (Table
2 and Figure 3). The resonances can be assigned on the basis
of the vicinal coupling constants with the proton in position 2.
It was shown for derivative 4 dissolved in acetonitrile that the
signal at low field (3.08 ppm, Jvic ) 5.9 Hz) belongs to the
pro-S hydrogen, whereas the signal at high field (2.93 ppm,
Jvic ) 8.3 Hz) is due to the pro-R hydrogen (16). The assignment
can be easily transferred to the spectra taken in other solvents.
The (D/H)H-3R and (D/H)H-3S values, reported in Table 2, are
substantially identical for the synthetic samples 2 and 4, whereas
for the other samples (D/H)H-3S is always smaller than (D/
H)H-3R, such a difference being dramatic for the natural sample
1 (167.4 vs 374.8 ppm, 69% of the 3R-isotopomer). Imbalances
of the deuterium enrichment at the diastereotopic positions of
the methylene group, although not as strong as in the present
case, were observed previously for aspartic and glutamic acids
(6, 7). Because such differences were detected only for natural
samples the method was proposed as a tool for recognizing, in
an absolute and simple way, the biological origin of a compound
(6). By applying this criterion of naturalness to phenylalanine
samples 1-8, clearly samples 2 and 4 are confirmed to be
synthetic and all other samples are natural, in agreement with
the conclusion based on the isotopic analysis of the benzoic
acids.
Phenylalanine. The (D/H)i values of the eight phenylalanine
samples examined as derivatives 4 are reported in Table 2. It
is noteworthy that the (D/H)Ph ratios are very similar to those
of the derived benzoic acids, definitely proving that the chemical
degradation did not alter substantially the deuterium enrichment
of the target molecule. Inspection of Table 2 highlights as the
main element of differentiation the deuterium content at position
2
3 of the CH2CH moiety, well exemplified in the H NMR
spectra of samples 1 and 2 (Figure 3) obtained from the natural
and racemic commercial 1. Samples 1, 3, and 5 form a group
characterized by an extremely high deuterium enrichment at C-3
with (D/H)3 values in the range of 283-351 ppm. The other
samples show more modest (D/H)3 ratios going from 90.6 ppm
of sample 4 to 141.2 ppm of sample 6. It may be of interest in
this context to examine the global (D/H) values for the reported
samples, which span from 107.9 to 205 ppm (Table 2). For
natural aromatic amino acids derived from C-3 plants the global
(D/H) ratios should occur at ∼140 ppm (15). In our case, apart
the racemic synthetic samples 2 and 4, which show quite low
global enrichments of 117.2 and 107.9 ppm, respectively, all
other samples are of natural derivation as established on the
basis of isotopic analysis of the derived benzoic acids discussed
above. However, only samples 6-8 show global (D/H) in the
range of 125-133 ppm near the values characteristic of natural
samples. Samples 1, 3, and 5 are peculiar, showing very high
(D/H)total (181-205 ppm) as a direct consequence of the
abnormal deuterium content at position 3. Nevertheless, other
examples are reported to show particularly high (D/H)total. In a
study carried out on the amino acids alanine, aspartic acid,
glutamic acid, proline, and lysine a very wide distribution of
the global (D/H) ratios determined by the IRMS technique was
Considering together the data relative to derivatives 4 and 5,
obtained from the eight samples of phenylalanine (Tables 1
and 2), it is possible to make a representation of the main
differences of phenylalanine samples combined with those
present in the derived benzoic acids by a multivariate data
analysis (17). Starting from the (D/H)i values of benzoic acid
and phenylalanine, it is observed that two principal components
C1 and C2 explain 84.1% of the overall variance. Thus, Figure
4, relative to the principal component analysis (PCA), indicates

4870 J. Agric. Food Chem., Vol. 51, No. 17, 2003
Brenna et al.
to the fact that racemic tyrosine is prepared from the L-
enantiomer via azlactone, which is hydrolyzed with aqueous
acid causing the introduction at position 2 of a hydrogen atom
from water in substitution of that originally present in the parent
L-amino acid (18). Conversely, the (D/H)3 values of 1 do not
show the extremely strong variations observed for 2. In addition,
all tyrosine samples display practically equal deuterium contents
at the diastereotopic positions 3R and 3S, a fact that is quite
puzzling because 1 presumably is obtained from natural sources,
that is, keratin-rich animal material. This implies that the amino
acid has been synthesized directly by plants or partially by
animal hydroxylation of plant-synthesized phenylalanine (3),
both following the shikimate pathway. Due to the probable
natural derivation of the samples a more or less evident
imbalance of (D/H)3R and (D/H)3S values should be expected,
a fact that does not occur. Moreover, the global (D/H)total values
of C-3 plant products from the shikimate pathway should range
from 125 to 145 ppm (15). The global values of 1 reported in
Table 3 are well within the predicted range for samples 1, 3, 5,
and 9 (∼130 ppm), whereas for the other samples they are
somewhat lower than expected (102.6-119.7 ppm). In any case
they cannot be used as an element of differentiation among the
samples.
Figure 4. Graphical representation of the distribution of samples 1−8 of
phenylalanine derivative 4 and benzoic acid (5) by PCA. Principal
components C1 and C2 explain >80% of the overall variance. Samples
can be divided into three groups: natural 1 (samples 1, 3, and 5), natural
2 (samples 6−8), and synthetic (samples 2 and 4).
From the data in Table 3 it appears that the nine samples of
tyrosine (1) are poorly differentiable also on the basis of their
site-specific deuterium content. The (D/H) values of the aromatic
hydrogen atoms of samples 2 (hen feathers), 3 (D,L Fluka), 5 (L
Fluka), and 9 (animal origin) are quite similar and, in general,
higher than those of the remaining samples. The three batches
of tyrosine (1) from human hairs (samples 6-8) are character-
ized by similar low values of the aromatic moiety and by side-
chain values consistent with those of the other materials. In
addition, it should be noted that the majority of tyrosine samples
follow the trend (D/H)ortho > (D/H)meta (where meta and ortho
are defined with respect to the side chain), although generally
the reverse occurs for para-hydroxylated natural compounds
(15). Possibly this behavior may be due to the fact that tyrosine
is produced from protein hydrolysates in strong acidic condi-
tions. It is reported that in these conditions the tyrosine meta
hydrogen atoms exchange with those of water (19). We have
verified that when a sample of the tyrosine is left in acidic D2O/
H2O medium, the deuterium spectrum shows a strong enhance-
ment of the meta signal (ortho to the OH group), whereas all
other signals remain unchanged. Thus, the deuterium content
Figure 5. Natural abundance 2H NMR spectra of tyrosine derivative 6:
(a) commercial sample 1; (b) natural sample 2 (hen feathers); (o) signal
of nuclei ortho to the side chain; (m) signal of nuclei meta to the side
chain; (/) signal of calibrated tert-butyl disulfide used as internal standard;
(O) CH CN solvent signal.
3
the presence of a group of three samples (5, 1, and 3, named
“natural 1”), quite distinct from a second group (samples 6-8,
named “natural 2”), and, finally, a third set (samples 2 and 4).
The first two groups are certainly of natural origin due to the
fingerprint of naturalness shown by the labeling pattern of
benzoic acid. Benzoic acid produced by oxidation of the
components of the third group shows the typical statistical
labeling pattern of products of petrochemical origin. An
additional element of differentiation of the latter material from
the products of natural origin is the equal deuterium content at
positions 3R and 3S of the phenylalanine framework.
Tyrosine and Tyrosols. Figure 5 shows the NMR spectra
of derivative 6 of samples 1 and 2 (racemic material from Merck
and the L-isomer from hen feathers) of tyrosine (1), whereas in
Table 3 are reported the (D/H)i (ppm) values relative to the
nine examined samples. A direct comparison of the data of
Tables 3 and 2, relative to the derivatives of tyrosine (1) and
phenylalanine (2), shows that the values greatly differentiating
the two amino acids are those of the CHCH2 moiety, that is,
the (D/H)3 and (D/H)2 ratios. The (D/H)2 values for tyrosine
(1) are, in general, higher than those of phenylalanine (2)
(119.0-195.6 vs 75.0-128.8). Significantly high is the value
195.6 ppm of sample 1 (D,L-tyrosine from Merck), possibly due
2
of the tyrosine meta position reflects the H abundances of
hydrolysis water medium rather than the result of the metabolic
pathway. For this reason the tyrosine samples, although certainly
of natural origin, cannot be compared easily with other natural
substances sharing the common derivation from 3.
In Table 4 are included the (D/H)i values relative to tyrosol
(7) of plant extraction (sample 1) and to the five samples
produced in baker’s yeast from tyrosine (1) (samples 2-4 and
6 from samples 1, 5, 2, and 6 of tyrosine; sample 5 from
L-tyrosine from Aldrich). Because during the bioconversion new
hydrogen atoms are introduced in position 1 of 7 (10), the (D/
H) values of the biogenerated CH2OH moiety and consequently
the global (D/H) ratios were not taken into account as an element
of comparison with respect to the parent tyrosine samples. On
the contrary, the aromatic and the benzylic portions of tyrosine
do not undergo variations during the biotransformation.
As expected the deuterium distribution of tyrosol samples
2-6 is in agreement with that of the tyrosine from which they
derive. Consequently, tyrosols such as tyrosines are poorly
differentiable. Apart from the deuterium content at position 1

Natural Abundance 2H NMR of Phenylalanine and Tyrosine
J. Agric. Food Chem., Vol. 51, No. 17, 2003 4871
Table 3. Origin and (D/H) Isotopic Ratios (ppm) of Tyrosine Derivative Samples 6a
i
b
b
c
sample
origin
D,L Merck
hen feathers
D,L Fluka
Fluka (old)
L Fluka
hair (primary sample)
hair (primary sample)
hair (primary sample)
unknown animal sample
(D/H)ortho
(D/H)meta
(D/H)2
(D/H)3
(D/H)
total
1
2
3
4
5
6
7
8
9
104.4
115.2
128.4
108.1
134.4
97.9
103.4
96.7
134.4
106.3
106.4
124.1
93.9
126.0
92.6
96.9
79.1
126.8
195.6
118.9
151.1
135.3
150.8
134.2
159.0
125.7
136.9
108.7
108.1
119.6
111.4
109.2
115.9
119.8
108.9
116.9
128.7
112.1
130.8
112.2
130.1
110.1
119.8
102.6
128.7
a Mean calculated standard deviation of (D/H) values is 2.5. b Positions ortho and meta refer to the side chain. c (D/H) values are calculated from the individual
i
total
positional deuterium contents.
derivation through a comparison of the deuterium labeling
pattern of samples of 1 and 2 of different origins. This primary
goal has not been achieved because among the nine examined
tyrosine samples no differences in the natural abundance
deuterium pattern that might be indicative of different origins
were observed. However, the mode of deuterium labeling of
the tyrosine and phenylalanine samples shows interesting
features that require some comments. The main differences
between L-tyrosine (1) and L-phenylalanine (2) regard the
deuterium enrichment at the CH2 and CH groups of the side
chain. First of all, the (D/H)2 values are smaller than the (D/
H)3 ratios for 1, whereas they are greater for 2. Second, the
deuterium enrichments at the two diastereotopic positions of
the benzylic methylene group are identical for L-tyrosine (1)
samples but different for L-phenylalanine (2) samples, which
show (D/H)3R > (D/H)3S. Tentatively, an explanation for this
phenomenon can be proposed by considering the biosyntheses
of 1 and 2 in light of the recent observations on the mode of
deuterium labeling pattern of D-glucose and D-fructose in
different plants (20). Indeed, Martin et al. (20) have elegantly
demonstrated that hexoses produced through the three photo-
synthetic pathways C3, C4, and CAM differ in their deuterium
distributions. An important observable difference regards the
deuterium enrichment at positions 6R and 6S of D-glucose. (D/
H)6R and (D/H)6S are practically identical for D-glucose from
C3 plants, whereas in the case of C4 plants (D/H)6R is greater
than (D/H)6S with differences in the range of 10-15 ppm. The
reverse is true for CAM plants.
Table 4. Origin and (D/H) Isotopic Ratios (ppm) of Tyrosol Samples
i
7a
b
b
sample
origin
(D/H)ortho
(D/H)meta
(D/H)2
(D/H)1
1
2
3
4
5
6
ligustrum
Merck
Fluka
hen feathers
Aldrich
hair
100.0
98.8
106.7
116.4
111.9
97.8
103.0
98.3
102.1
109.8
103.7
106.6
94.1
107.6
102.0
114.6
100.8
110.0
95.1
71.0
54.7
62.2
61.7
47.5
a Mean calculated standard deviation of (D/H) values is 2.9. Tyrosol samples
i
2−4 and 6 were obtained from tyrosine samples 1, 5, 2, and 6, respectively, and
sample 5 was from L-tyrosine from Aldrich. b Positions ortho and meta refer to the
side chain.
Figure 6. Discrimination of tyrosols through PCA. Principal components
C1 and C2 explain >84% of the overall variance.
the deuterium pattern of plant tyrosol 7 (sample 1) is quite
similar to that of the other samples. This observation cor-
roborates the hypothesis that the commercial tyrosine is
produced entirely from natural sources (animal keratin). Con-
sidering as variables the site-specific deuterium enrichments of
all tyrosol samples including that of the CH2 in position 1, the
graphical representation of Figure 6 can be drawn, where the
two principal components C1 and C2 account for 84.1% of the
overall variance. The graph shows two well-separated regions,
the first containing the isolated sample of “plant” tyrosol from
Ligustrum and the second the other samples grouped together.
Thus, the deuterium isotope analysis is not able to differentiate
between tyrosine and tyrosol samples of different origins. On
the contrary, for the phenolic oxygen, the same samples show
quite distinct δ18O values: +26.0 and +31.5‰, respectively,
for samples 1 and 5, obtained from the parent amino acid of
plant origin [tyrosine directly derived from arogenic acid (3)]
and +18.9 and +17.8‰, respectively, for samples 4 and 6 of
animal origin, indicative of the presence in the samples, close
to plant tyrosine (1) taken up by the animal with the diet, of ca.
46 and 51%, respectively, of the same chemical species, formed
in the body by hydroxylation of phenylalanine (2) (3).
It is well-known (4) that in the biosynthesis of 1 and 2 the
benzylic CH2 group incorporates both of the methylene hydro-
gens at positions 1 and 6 of fructose-1,6-diphosphate, which is
formed from glucose at the beginning of the glycolysis process,
with conservation of the stereochemical information associated
with the biosynthetic mechanism. Thus, the substantial differ-
ences here observed in the deuterium labeling of benzylic
position in L-tyrosine (1) and L-phenylalanine (2) might possibly
be consequences of the mode of labeling of the parent sugars.
That is, the examined L-tyrosine (1), in which (D/H)3R ) (D/
H)3S, can seemingly derive from C3 plants, the glucose of which
is equally labeled at position 6. Conversely, L-phenylalanine
(2) isolated from commercial aspartame and obtained on the
market can possibly be produced by fermentation processes
supplemented with cane sugar molasses and/or corn syrup,
which represent the cheapest carbon source for microorganisms.
Both sugar cane and maize are C4 plants, shown to provide
D-glucose (20) in which the deuterium signals (D/H)3R > (D/
H)3S.
The interpretation above may be reasonable for L-tyrosine
samples that are obtained from animal keratin derived originally
with high probability from C3 plants. In contrast, the theory is
The basic aim of this study was the search for a fingerprint
of the formation pathway of tyrosine samples of animal

4872 J. Agric. Food Chem., Vol. 51, No. 17, 2003
Brenna et al.
(8) Li, W.-R.; Yo, Y.-C.; Lin, Y.-S. Efficient one-pot formation of
amides from benzyl carbamates: Application to solid-phase
synthesis. Tetrahedron 2000, 56, 8867-8875.
(9) Aleu, J.; Fronza, G.; Fuganti, C.; Serra, S.; Fauhl, C.; Guillou,
C.; Reniero, F. Differentiation of natural and synthetic phenyl-
acetic acids by 2H NMR of the derived benzoic acids. Eur. Food
Res. Technol. 2002, 214, 63-66.
(10) Cheetam, P. S. J. Novel specific pathways for flavor production.
In Bioformation of FlaVors; Patterson, R. L. S., Charlwood, B.
V., MacLeod, G., Williams, A. A., Eds.; The Royal Society of
Chemistry: Cambridge, U.K., 1992; p 103.
(11) Fuganti, C.; Ghiringhelli, D.; Grasselli, P.; Santopietro-Amisano,
A. On the stereochemistry of the conversion of tyramine and
hordenine into tyrosol by Willia anomala Hansen. J. Chem. Soc.,
Chem. Commun. 1973, 862-863.
(12) Hagedorn, M. L. Differentiation of natural and synthetic benz-
aldehydes by 2H nuclear magnetic resonance. J. Agric. Food
Chem. 1992, 40, 634-637.
(13) Remaud, G.; Debon, A. A.; Martin, Y.; Martin, G. G.; Martin,
G. J. Authentification of bitter almond oil and cinnamon oil:
Application of the SNIF-NMR method to benzaldehyde. J. Agric.
Food Chem. 1997, 45, 4042-4048.
rather weak as an explanation for the behavior of L-phenyl-
alanine samples because the observed differences of the
deuterium population at the methylene group are much stronger
than that reported for the hypothesized starting sugar material.
To gain more information about this point, the values of δ13C
of the benzoic acids 1-8 obtained by degradation of the
corresponding phenylalanine samples have been measured
(Table 1). The material deriving from C3 plants is expected to
have δ13C of ∼26-30‰, whereas that from C4 plants should
have δ13C of 10-13‰ (21). The 13C isotopic ratios in Table 1
show that only the benzoic acid samples 6 and 8 from the
aspartame Miwon (-13.8‰) and aspartame Caremoli (-12.4‰)
originate from C4 plants, whereas all remaining samples must
derive with certainty from C3 plant material (values from -25.4
to -29.4‰). Thus, the correlation of the asymmetry of the
deuterium population at the methylene group of L-phenylalanine
(2) with that of the starting sugar material fails. Most probably
such a strong difference is due to drastic biotechnological effects
connected to the kind of microorganism used in the fermentation
processes.
(14) Schmidt, H.-L.; Eisenreich, W. Systematic and regularities in
the origin of 2H patterns in natural compounds. Isotopes EnViron.
Health Stud. 2001, 37, 253-254.
(15) Schmidt, H.-L.; Werner, R. A.; Eisenreich, W. Systematics of
2H patterns in natural compounds and its importance for the
elucidation of biosynthetic pathways. Phytochem. ReV. 2003, in
press.
ACKNOWLEDGMENT
We thank Prof. H.-L. Schmidt for his interest in and suggestions
for this work.
LITERATURE CITED
(16) Newmark, R. A.; Miller, M. A. Nuclear magnetic resonance study
of the conformations of valine and phenylalanine derivatives. J.
Phys. Chem. 1971, 75, 505-508.
(17) Podani, J. MultiVariate Data Analysis in Ecology and Systemat-
ics; Academic Publishing: The Hague, The Netherlands, 1994.
(18) Greenstein, J. P.; Winitz, M. The Chemistry of Amino Acids;
Krieger Publishing: Malabar, FL, 1984; p 2364.
(19) Rittenberg, D.; Keston, A. S.; Schoenheimer, R.; Foster, G. L.
Deuterium as an indicator in the study of intermediary metabo-
lism. XIII. The stability of hydrogen in amino acids. J. Biol.
Chem. 1938, 125, 1-12.
(1) Schmidt, H.-L.; Werner, R. A.; Rossman, A. 18O pattern and
biosynthesis of natural plant products. Phytochemistry 2001, 58,
9-32.
(2) Martin, G. J.; Martin, M. L. Stable isotope analysis of food and
beverages by NMR spectroscopy. Annu. Rep. NMR Spectrosc.
1995, 31, 91-104.
(3) Fronza, G.; Fuganti, C.; Schmidt, H.-L.; Werner, R. A. The δ18O-
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.
(4) Haslam, E. Shikimic Acid, Metabolism and Metabolites; Wiley:
Chichester, U.K., 1993.
(5) Thorpe, J. M.; Roberts, S. A.; Ball, R. O.; Pencharz, P. B. Effect
of tyrosine intake on the rate of phenylalanine hydroxylation in
adult males. Metabolism 2000, 49, 444-449.
(6) Martin, M. L.; Zhang, B.; Martin, G. J. Natural chirality of
methylene sites applied to the recognition of origin and to the
study of biochemical mechanisms. FEBS Lett. 1983, 158, 131-
133.
(20) Zhang, B.-L.; Billault, I.; Li, X.; Mabon, F.; Remaud, G.; Martin,
M. L. Hydrogen isotopic profile in the characterization of sugars.
Influence of the metabolic pathway. J. Agric. Food Chem. 2002,
50, 1574-1580.
(21) Bender, M. M. Variations in 13C/12C ratios of plants in relation
to pathway of photosynthetic carbon dioxide fixation. Phyto-
chemistry 1971, 10, 1239-1244.
(7) Vallet, C.; Arendt, M.; Mabon, F.; Naulet, N.; Martin, G. J.
Combination of mass spectrometry and site-specific NMR isotope
analyses in the characterization of amino acids. J. Sci. Food
Agric. 1991, 56, 167-185.
Received for review April 10, 2003. Revised manuscript received May
29, 2003. Accepted May 29, 2003.
JF0302759