HPLC-MS/MS Profiling of Tryptophan-Derived Alkaloids in Food: Identification of Tetrahydro-β-carbolinedicarboxylic Acids

Article abstract of DOI:10.1021/jf960952h

A method for selective detection of 1,2,3,4-tetrahydro-β-carbolinecarboxylic acids (THCCs) was developed based on electrospray ionization-tandem mass spectrometry coupled to liquid chromatography (HPLC-ESI-MS/MS). Low-energy collision-induced dissociation (CID) led to characteristic fragment ions due to neutral loss of 73 amu. Subsequently, constant neutral loss scanning was used for substructure specific screening of THCCs in food samples. Detection limits for HPLC-ESI-MS/MS analysis of THCCs applying neutral loss experiments were established at 100 ng mL^-1 (ca. 2.5 pmol on column). Application of this MS/MS method enabled us to detect THCC derivatives derived from Pictet-Spengler condensation of tryptophan with α-oxo acids. Subsequently, diastereomeric 1,2,3,4-tetrahydro-β-carboline-1,3-dicarboxylic acid 3a/b, 1-methyl-1,2,3,4-tetrahydro-β-carboline-1,3-dicarboxylic acid 4a/b, and 1-(2′-carboxyethyl)-1,2,3,4-tetrahydro-β-carboline-3-carboxylic acid 5a/b were identified in alcoholic beverages, seasoning sauces, yeast extract, and fruit products for the first time. Most food samples under study contained 3a/b and 4a/b in significant amounts. 5a/b was identified in soy sauce, worcestershire sauce, seasoning sauce, and yeast extract. Due to the excellent selectivity of tandem mass spectrometry coeluting tetrahydroβ-carboline derivatives could be identified unequivocally by HPLC-ESI-MS/MS.

Full text of DOI:10.1021/jf960952h

2458
J. Agric. Food Chem. 1997, 45, 2458−2462
HP LC-MS/MS P r ofilin g of Tr yp top h a n -Der ived Alk a loid s in F ood :
Id en tifica tion of Tetr a h yd r o-â-ca r bolin ed ica r boxylic Acid s
B. Gutsche and M. Herderich*
Lehrstuhl f u¨ r Lebensmittelchemie, Universit a¨ t W u¨ rzburg, Am Hubland, 97074 W u¨ rzburg, Germany
A method for selective detection of 1,2,3,4-tetrahydro-â-carbolinecarboxylic acids (THCCs) was
developed based on electrospray ionization-tandem mass spectrometry coupled to liquid chroma-
tography (HPLC-ESI-MS/MS). Low-energy collision-induced dissociation (CID) led to characteristic
fragment ions due to neutral loss of 73 amu. Subsequently, constant neutral loss scanning was
used for substructure specific screening of THCCs in food samples. Detection limits for HPLC-
-1
ESI-MS/MS analysis of THCCs applying neutral loss experiments were established at 100 ng mL
ca. 2.5 pmol on column). Application of this MS/MS method enabled us to detect THCC derivatives
(
derived from Pictet-Spengler condensation of tryptophan with R-oxo acids. Subsequently, diaster-
eomeric 1,2,3,4-tetrahydro-â-carboline-1,3-dicarboxylic acid 3a /b, 1-methyl-1,2,3,4-tetrahydro-â-
carboline-1,3-dicarboxylic acid 4a /b , and 1-(2′-carboxyethyl)-1,2,3,4-tetrahydro-â-carboline-3-
carboxylic acid 5a /b were identified in alcoholic beverages, seasoning sauces, yeast extract, and
fruit products for the first time. Most food samples under study contained 3a /b and 4a /b in
significant amounts. 5a /b was identified in soy sauce, worcestershire sauce, seasoning sauce, and
yeast extract. Due to the excellent selectivity of tandem mass spectrometry coeluting tetrahydro-
â-carboline derivatives could be identified unequivocally by HPLC-ESI-MS/MS.
Keyw or d s: Tetrahydro-â-carboline; electrospray ionization; tandem mass spectrometry; neutral loss
scanning
INTRODUCTION
,2,3,4-Tetrahydro-â-carboline-3-carboxylic acid de-
corresponding tetrahydro-â-carbolinedicarboxylic acids
which makes isolation and derivatization difficult. In
order to clarify the role of these polar tryptophan
derivatives in food, a reliable analytical method for
detection and determination of THCCs in complex
matrices was required.
Thus, we report an approach based on HPLC-ESI-
MS/MS for substructure specific detection of novel
THCC derivatives in food samples. In addition, results
obtained by analysis of tetrahydro-â-carboline-1,3-di-
carboxylic acids in alcoholic beverages, seasoning sauces,
yeast extract, and fruit products are presented.
1
rivatives (THCCs) are readily formed by Pictet-Spen-
gler condensation of tryptophan with various aldehydes
or R-oxo acids (Harvey et al., 1941). THCCs have been
demonstrated to inhibit monoamine oxidase, to alter the
re-uptake of biogenic amines, and to interfere with
benzodiazepine receptors (McKenna and Towers, 1984;
Rommelspacher et al., 1994). Because of their neuro-
pharmacological effects these alkaloids have attracted
much concern; thus, their participation in pathogenesis
of alcoholism and psychiatric disorders is under inves-
tigation (Rommelspacher et al., 1984; Spies et al., 1995).
Structures of THCCs can vary widely depending on
both indole amine derivatives and carbonyl compounds
involved in Pictet-Spengler condensation reaction.
However, only 1,2,3,4-tetrahydro-â-carboline-3-carboxyl-
ic acid 1 and 1-methyl-1,2,3,4-tetrahydro-â-carboline-
EXPERIMENTAL PROCEDURES
Ap p a r a tu s. Chromatographic separation was performed
by an Applied Biosystems 140b pump. For sample injection
a SunChrom Triathlon autosampler (BAI, Bensheim, Ger-
many) was used. HPLC-ESI-MS/MS analysis was performed
utilizing a TSQ 7000 tandem mass spectrometer system
equipped with an ESI interface (Finnigan MAT, Bremen,
Germany). Data acquisition and evaluation were conducted
on a DEC 5000/33 (Digital Equipment, Unterf o¨ hring, Ger-
many) with ICIS 8.1 software (Finnigan MAT, Bremen,
Germany). NMR spectra were acquired with a Bruker WM
400 spectrometer calibrating the chemical shifts utilizing the
3
-carboxylic acid 2a /b have been demonstrated to occur
in considerable amounts in smoked and cured meat
products and food samples like wine, beer, soy sauce
(
(
Wakabayashi et al., 1983), cheese, yoghurt, and bread
Bosin et al., 1986; Adachi et al., 1991; Papavergou and
Clifford, 1992; Herraiz, 1996). Besides the widespread
occurrence of â-carbolines 1 and 2a /b, condensation
products of tryptophan with glycolaldehyde (Sen et al.,
1
solvent signal (DMSO: 2.58 ppm for H-NMR, 39.7 ppm for
1
3
C-NMR. MeOH: 3.5 ppm for 1H-NMR, 49.0 ppm for C-
13
NMR) as reference.
1
1
995), hexanal (Arai et al., 1971), pyridoxal (Argoudelis,
994), and isobutyraldehyde (Herraiz et al., 1993) have
Rea gen ts. Water, acetonitrile, both of HPLC gradient
grade, L-tryptophan, and trifluoroacetic acid (spectroscopic
grade) were from Merck (Darmstadt, Germany). Glyoxylic
acid, pyruvic acid, and R-oxoglutaric acid were purchased from
Sigma (Steinheim, Germany). Formaldehyde, acetaldehyde,
and succinic semialdehyde were obtained from Aldrich
(Deisenhofen, Germany). L-[Indole-D5]tryptophan was from
Cambridge Isotope Laboratories (Andover, MA). All chemicals
were of analytical purity. Membrane filters of pore size 0.2
µm were from Ziemer (Mannheim, Germany). C18 cartridges
(200 mg, 3 mL) were from Varian (Harbor City, MD). Food
samples were purchased at local markets.
been described, but until now knowledge of the presence
of THCCs in food samples has been rather limited. For
example, no information is available on products formed
by Pictet-Spengler condensation of tryptophan with
R-oxo acids commonly found in several food samples.
One likely explanation is the labile character of the
*
Corresponding author (fax +49-931-8885484; e-mail
mherderi@pzlc.uni-wuerzburg.de).
S0021-8561(96)00952-1 CCC: $14.00
© 1997 American Chemical Society

Identification of Tetrahydro-
â
-carbolinedicarboxylic Acids
J. Agric. Food Chem., Vol. 45, No. 7, 1997 2459
Ta ble 1. 1,2,3,4-Tetr a h yd r o-â-ca r bolin e-3-ca r boxylic
DMSO-d
6
), 10.78 (s, 1H, H-9), 7.48 (d, 2H, H-5 and H-8), 7.12
Acid s Id en tified in F ood Sa m p les
(dd, 1H, H-7), 7.03 (dd, 1H, H-6), 4.17 (m, 1H, H-3), 3.22 (dd,
H, H-4ax), 2.96 (dd, 1H, H-4eq), 1.81 (s, 3H, 1-CH
8 Hz, J 6,7 ) 7 Hz, J 3,4ax ) 4 Hz, J 3,4eq ) 12 Hz, J 4ax,4eq ) 15
1
3
); J 5,6 ) J 7,8
)
1
3
Hz; C-NMR (100 MHz, DMSO-d
COOH′), 136.30 (C-8a), 133.70 (C-9a), 125.92 (C-4b), 121.22
C-7), 118.62 (C-6), 117.73 (C-5), 111.74 (C-8), 104.96 (C-4a),
5.75 (C-1), 52.15 (C-3), 24.49 (CH ), 24.05 (C-4). Minor
diastereomer: H-NMR (400 MHz, DMSO-d ), 11.16 (s, 1H,
H-9), 7.49 (d, 1H, H-5), 7.41 (d, 1H, H-8), 7.13 (dd, 1H, H-7),
6
), 169.83 (COOH and
(
5
3
1
6
compd no.
R1
R2
7
2
.04 (dd, 1H, H-6), 4.15 (dd, 1H, H-3), 3.16 (dd, 1H, H-4ax),
.83 (dd, 1H, H-4eq), 1.80 (s, 3H, 1-CH
7 Hz, J 3,4ax ) 4 Hz, J 3,4eq ) 12 Hz; J 4ax,4eq ) 15 Hz; C-NMR
1
H
CH3
H
H
H
3
); J 5,6 ) J 7,8 ) 8 Hz, J 6,7
13
2
2
3
3
4
4
5
5
a a
)
a
b
a
b
a
b
a
CH3
COOH
H
COOH
CH3
H
(
6
100 MHz, DMSO-d ), 172.64, 172.52 (COOH and COOH′),
a
H
a
136.50 (C-8a), 133.86 (C-9a), 125.81 (C-4b), 121.06 (C-7), 118.52
(C-6), 117.82 (C-5), 111.28 (C-8), 105.90 (C-4a), 60.19 (C-1),
54.08 (C-3), 25.07 (CH ), 24.00 (C-4).
3
COOH
CH3
COOH
CH2-CH2-COOH
H
b
b
a
1-(2′-Carboxyethyl)-1,2,3,4-tetrahydro-â-carboline-3-carboxy-
lic Acid (5a /b). Compound 5a /b was synthesized by reaction
of L-tryptophan and succinic semialdehyde. The major dia-
stereomer cis-1-(2′-carboxyethyl)-1,2,3,4-tetrahydro-â-carbo-
line-3-carboxylic acid 5a precipitated; the filtrate, a mixture
of both diastereomers, was lyophilized prior to NMR analysis.
Relative configuration of the diastereomers was determined
by 13C-NMR (Cox and Cook, 1995). Signal assignment (Table
a
b
CH2-CH2-COOH
a
Assignment of absolute configuration was based on the
observation that racemization at C3 is negligible while working
at room temperature in aqueous media (Bailey et al., 1993).
b
Assignment of absolute configuration is exchangeable.
Refer en ce Com p ou n d s. Synthesis of 1,2,3,4-tetrahydro-
â-carboline-3-carboxylic acid 1 and 1-methyl-1,2,3,4-tetrahy-
dro-â-carboline-3-carboxylic acid 2a /b was based on the pro-
cedure of J acobs and Craig (1936). Absolute configurations
of the resulting diastereomers (1S,3S)-1-methyl-1,2,3,4-tet-
rahydro-â-carboline-3-carboxylic acid 2a and (1R,3S)-1-methyl-
1
+
2
) was confirmed by CH-COSY experiments. ESI-MS, [M
+
H] m/ z 289; ESI-MS/MS (15 eV, 2.0 mTorr Ar), m/z 272,
16, 188, 130. Data for cis-1-(2′-carboxyethyl)-1,2,3,4-tetrahy-
1
dro-â-carboline-3-carboxylic acid 5a : H-NMR (400 MHz, CD
OD), 7.69 (d, 1H, H-5), 7.57 (d, 1H, H-8), 7.36 (dd, 1H, H-7),
.27 (dd, 1H, H-6), 5.02 (m, 1H, H-1), 4.59 (dd, 1H, H-3), 3.67
dd, 1H, H-4ax), 3.34 (dd, 1H, H-4eq), 2.86 (m, 3H, H-2′ and
H-1′ax), 2.50 (m, 1H, H-1′eq); J 5,6 ) J 7,8 ) 8 Hz, J 6,7 ) 7 Hz,
3
-
7
(
1
,2,3,4-tetrahydro-â-carboline-3-carboxylic acid 2b were as-
signed according to published data (Yamada and Akimoto,
969). The identity of 2a /b was confirmed by ESI-MS, ESI-
MS/MS, and NMR spectroscopy.
,2,3,4-Tetrahydro-â-carboline-1,3-dicarboxylic Acid (3a /b).
.5 mM L-tryptophan was dissolved in 7.5 mL of water. After
1
13
J
3,4ax ) 5 Hz, J 3,4eq ) 12 Hz, J 4ax,4eq ) 16 Hz; C-NMR (100
MHz, DMSO-d + deuterated TFA), 174.07, 170.53 (COOH
6
1
and COOH′), 136.77 (C-8a), 129.64 (C-9a), 125.88 (C-4b),
122.35 (C-7), 119.54 (C-6), 118.47 (C-5), 111.77 (C-8), 105.67
(C-4a), 55.53, 53.42 (C-1 and C-3), 29.45 (CH -2′), 26.08 (CH -
2
addition of 1 mL of 0.5 M sulfuric acid (pH 2) and 5 mM
glyoxylic acid, the reaction mixture was allowed to stand at
room temperature for 2 days (Harvey et al., 1941). The
precipitate consisting of the major diastereomer cis-1,2,3,4-
tetrahydro-â-carboline-1,3-dicarboxylic acid 3b was collected
by filtration, washed with water, and dried. Both lyophiliza-
tion and preparative HPLC of the trans-configurated isomer
failed as the compound rapidly decomposed on concentration.
Thus, the remaining filtrate containing a mixture of both
diastereomers was concentrated by C18 solid phase extraction
2
2
1′), 22.40 (C-4). Data for trans-1-(2′-carboxyethyl)-1,2,3,4-
tetrahydro-â-carboline-3-carboxylic acid 5b:
MHz, DMSO-d + deuterated TFA), 51.92, 51.34 (C-1 and C-3).
1
3
C-NMR (100
6
An a lytica l P r oced u r es. Sample Preparation. Prior to
any preconcentration step food samples were spiked with
L-tryptophan-d (10 µg mL-1) for monitoring artifact formation
5
during sample handling (Gutsche et al., 1996). Seasoning
sauce, soy sauce, and worcestershire sauce were filtered
through membrane filters of pore size 0.2 µm, and the resulting
solutions were directly subjected to HPLC-ESI-MS/MS analy-
sis. All other samples were concentrated by C18 solid phase
extraction as follows: Aliquots of alcoholic beverages and fruit
products (normally 10 mL) were diluted to 20 mL end-volume
with distilled water. In the case of yeast extract and caramel
color, the sample (1-10 g) was diluted to 20 mL with distilled
water and centrifuged, and the supernatant was used for
analysis. Each sample was adjusted to pH 1 by addition of 2
M hydrochloric acid and subjected to solid phase extraction
(SPE) (Visiprep SPE vacuum manifold system; Supelco, Belle-
fonte, MD, U.S.A.). SPE was performed with C18 cartridges
(200 mg) conditioned with 3 mL of methanol and 6 mL of
water. After application of the acidified sample and washing
with 3 mL of water adjusted to pH 1 with HCl, each column
was eluted (0.2 mL/min) with 1 mL of methanol containing
1% TFA (v/v). The eluate was evaporated under a gentle
stream of nitrogen and redissolved in 100 µL of aqueous (90
% (v/v)) acetonitrile.
and eluted with CD
3
OD containing 0.1% deuterated trifluo-
roacetic acid (TFA) (v/v) prior to NMR analysis of the trans-
isomer. Relative configurations were determined by 13C-NMR
spectroscopy (Cox and Cook, 1995). Assignment of NMR
signals (Table 1) was confirmed by DEPT experiments. ESI-
+
MS, [M + H] m/ z 261; ESI-MS/MS (15 eV, 2.0 mTorr Ar),
m/ z 217, 188, 145. Data for cis-1,2,3,4-tetrahydro-â-carboline-
1
1
,3-dicarboxylic acid 3b: H-NMR (400 MHz, DMSO-d
6
), 10.79
(s, 1H, H-9), 7.57 (d, 1H, H-5), 7.52 (d, 1H, H-8), 7.15 (dd, 1H,
H-7), 7.05 (dd, 1H, H-6), 4.99 (s, 1H, H-1), 4.29 (dd, 1H, H-3),
3
J
.28 (dd, 1H, H-4ax), 3.09 (m, 1H, H-4eq); J 5,6 ) J 7,8 ) 8 Hz,
6,7 ) 7 Hz, J 3,4ax ) 5 Hz, J 3,4eq ) 12 Hz, J 4ax,4eq ) 16 Hz; 13C-
NMR (100 MHz, CD
1
1
3
OD + deuterated TFA), 170.70 and
67.73 (COOH and COOH′), 138.75 (C-8a), 127.09 (C-9a),
24.04 (C-4b), 120.96 (C-7), 119.08 (C-6), 118.35 (C-5), 112.92
(C-8), 107.80 (C-4a), 56.40, 56.10 (C-1 and C-3), 23.28 (C-4).
Data for trans-1,2,3,4-tetrahydro-â-carboline-1,3-dicarboxylic
1
3
acid 3a : C-NMR (100 MHz, CD
4.70, 54.25 (C-1 and C-3), 22.79 (C-4).
-Methyl-1,2,3,4-tetrahydro-â-carboline-1,3-dicarboxylic Acid
4a /b). 4a /b was synthesized as described above using 5 mM
3
OD + deuterated TFA),
5
Mass Spectrometric Analysis of Tetrahydro-â-carbolinecar-
boxylic Acids. Chromatographic separation for HPLC-ESI-
MS/MS was performed on an Eurospher 100 C18 column (100
× 2.0 m i.d., 5 µm) (Knauer, Berlin, Germany) using a binary
gradient. Solvent A was 0.05% TFA in water (v/v), solvent B
was 0.05% TFA in acetonitrile (v/v). HPLC was programmed
as follows: pressurizing with 50% B, equilibration time 5 min
at 10% solvent B and linear gradient elution (0 min, 10% B;
20 min, 30% B; 30 min, 50% B). The flow rate was 200 µL
1
(
pyruvic acid as carbonyl compound. The major diastereomer
precipitated. The minor diastereomer was isolated by pre-
parative HPLC (Eurospher 100 C18 column 16 × 250 mm,
5
µm; water-acetonitrile 8-2 containing 0.05% (v/v) TFA).
NMR signal assignment (Table 1) was confirmed by DEPT,
+
CH-COSY, and HMBC experiments. ESI-MS, [M + H] m/ z
2
1
-
1
75; ESI-MS/MS (15 eV, 2.0 mTorr Ar), m/z 258, 231, 202,
min and the injection volume was 5 µL, respectively. For
pneumatically assisted electrospray ionization, the spray
84, 159, 146, 130. Major diastereomer: 1H-NMR (400 MHz,

2460 J. Agric. Food Chem., Vol. 45, No. 7, 1997
Gutsche and Herderich
F igu r e 1. HPLC-MS/MS analysis of soy sauce, constant neutral loss scanning (-73 amu; 19 eV; 2.0 mTorr Ar). (A - E)
+
Reconstructed ion chromatograms showing precursor ions [M + H] of compounds 1-5. (F) Reconstructed ion chromatogram
(RIC).
+
F igu r e 2. Product ion spectrum of 1,2,3,4-tetrahydro-â-carboline-1,3-dicarboxylic acid 3a /b, precursor ion m/ z 261 [M + H] (15
eV, 2.0 mTorr Ar).
capillary voltage was set to 3.5 kV and the temperature of the
heated inlet capillary was 220 °C. Nitrogen served both as
sheath (50 psi) and auxiliary gas (10 units). Positive ions were
detected with a total scan duration of 1.0 s for a single full
spectrum. MS/MS experiments, such as product ion scanning
of the iminoacetic acid moiety C H NO (-73 amu) due
2
3
2
to retro-Diels-Alder fragmentation (Gutsche et al.,
1
996).
In order to extend our knowledge to the occurrence
and relevance of tetrahydro-â-carboline derivatives
other than 1 and 2a /b we analyzed numerous food
samples by means of HPLC coupled to tandem mass
spectrometry. The application of constant neutral loss
scanning in combination with subsequent product ion
experiments directly led to the substructure specific
identification (“profiling”) of various Pictet-Spengler
condensation products in complex matrices such as
fermented beverages, seasoning sauces, yeast extract,
and fruit products.
(
mass range 20-350 amu) and neutral loss experiments
(scanning from 200 to 350 amu), were performed at a collision
gas pressure of 2.0 mTorr Ar with a total scan duration of 3.0
s for a single spectrum. Quantitative evaluations were carried
out using standard solutions (0, 100, 1000, and 10000 ng/mL)
of reference compound 3b in H
2
O for external calibration.
Values were calculated from signal intensities of the respective
+
[
M + H] ions as obtained by HPLC-ESI-MS analysis.
RESULTS AND DISCUSSION
Detection limits for HPLC-ESI-MS/MS analysis of
Initial experiments had revealed that the electrospray
THCCs by means of neutral loss scanning were estab-
-
1
process could effectively ionize tetrahydro-â-carboline-
lished at 100 ng mL (ca. 2.5 pmol on column) using
standard solutions of reference compounds. As a rep-
resentative example the profile of THCCs from soy sauce
analysis is outlined in Figure 1. In addition to proto-
3
-carboxylic acids such as 1 and 2a /b; in the positive
+
mode, exclusively protonated molecules [M + H] were
obtained. In addition, low-energy CID of protonated
molecules had been demonstrated to yield characteristic
product ion spectra for tetrahydro-â-carbolines, i.e., the
most abundant product ions were formed by neutral loss
+
nated molecular ions [M + H] of well-documented
1,2,3,4-tetrahydro-â-carboline-3-carboxylic acid 1 (m/ z
217) and 1-methyl-1,2,3,4-tetrahydro-â-carboline-3-car-

Identification of Tetrahydro-
â
-carbolinedicarboxylic Acids
J. Agric. Food Chem., Vol. 45, No. 7, 1997 2461
F igu r e 3. Product ion spectrum of 1-methyl-1,2,3,4-tetrahydro-â-carboline-1,3-dicarboxylic acid 4a /b, precursor ion m/ z 275 [M
+
+
H] (15 eV, 2.0 mTorr Ar).
F igu r e 4. Product ion spectrum of 1-(2′-carboxyethyl)-1,2,3,4-tetrahydro-â-carboline-3-carboxylic acid 5a /b, precursor ion m/ z
+
2
89 [M + H] (15 eV, 2.0 mTorr Ar).
boxylic acid 2a /b (m/ z 231) (traces C and E, Figure 1),
precursor ions m/ z 261, m/ z 275, and m/ z 289 revealed
for the first time the presence of additional THCC
derivatives (traces A, B, and D, Figure 1). Product ion
spectra of these putative tetrahydro-â-carbolines as
obtained by low-energy CID showed characteristic frag-
mentation patterns dominated by the loss of 73 amu
Therefore, only application of tandem mass spectrom-
etry provided an approach for the reliable assignment
of numerous structurally related THCC derivatives in
complex food samples as coeluting tetrahydro-â-carbo-
line derivatives were identified unequivocally by HPLC-
ESI-MS/MS.
On the basis of structural information as implied by
tandem mass spectrometry of putative THCC deriva-
tives, authentic reference compounds were synthesized
and characterized spectroscopically. Upon reaction with
glyoxylic acid or pyruvic acid, L-tryptophan yielded
diastereomers of 1,2,3,4-tetrahydro-â-carboline-1,3-di-
carboxylic acid 3a /b and 1-methyl-1,2,3,4-tetrahydro-
â-carboline-1,3-dicarboxylic acid 4a/b, respectively (Table
1). In contrast, reaction of L-tryptophan and R-oxoglu-
taric acid led to a complex mixture of several unstable
conjugates. 1-(2′-Carboxyethyl)-1,2,3,4-tetrahydro-â-
carboline-3-carboxylic acid 5a /b was identified as the
major product by HPLC-ESI-MS/MS analysis. Un-
fortunately, 5a /b rapidly decomposed during the isola-
tion procedure. Thus, 5a /b was synthesized directly by
reaction of L-tryptophan and succinic semialdehyde.
Identity of purified 1,2,3,4-tetrahydro-â-carbolines
3a /b, 4a /b, and 5a /b was established by DEPT-,
(
Figures 2-4). In addition, product ions m/ z 217
(Figure 2; 3a /b) and m/ z 231 (Figure 3; 4a /b) apparently
resulted from the loss of CO2 (-44 amu) characteristic
for carboxylic acids; product ions m/ z 145 ([M + H -
+
+
1
16] ) of 3a /b (Figure 2) and m/ z 159 ([M + H - 116] )
of 4a /b (Figure 3) were indicative of the loss of the indole
moiety.
Under chromatographic conditions selected in the
reconstructed ion chromatogram RIC (Figure 1, trace
F) as well as in the fluorescence chromatogram (excita-
tion 290 nm, emission 360 nm, data not shown) dia-
stereomers of 1-methyl-1,2,3,4-tetrahydro-â-carboline-
3
-carboxylic acid 2a /b were separated neither from the
second eluting diastereomer of 4 nor from the first
eluting diastereomer of 5. Additionally, the first eluting
diastereomer of 4 and the second eluting diastereomer
of 3 could not be differentiated by chromatography.

2462 J. Agric. Food Chem., Vol. 45, No. 7, 1997
Gutsche and Herderich
Ta ble 2. Novel Tetr a h yd r o-â-ca r bolin es in F ood a
Bailey, P. D.; Hollinshead, S. P.; McLay, N. R.; Morgan, K.;
Palmer, S. J .; Prince, S. N.; Reynolds, C. D.; Wood, S. D.
Diastereo- and enantioselectivity in the Pictet-Spengler
reaction. J . Chem. Soc., Perkin Trans. 1 1993, 431-439.
Bosin, T. R.; Krogh, S.; Mais, D. Identification and quantitation
of 1,2,3,4-tetrahydro-â-carboline-3-carboxylic acid and 1-meth-
yl-1,2,3,4-tetrahydro-â-carboline-3-carboxylic acid in beer
and wine. J . Agric. Food Chem. 1986, 34, 843-847.
Cox, E. D.; Cook, J . M. The Pictet-Spengler condensation: A
new direction for an old reaction. Chem. Rev. 1995, 95,
1797-1842.
Gutsche, B.; Herderich, M. High performance liquid chroma-
tography-electrospray ionization-tandem mass spectrom-
etry (HPLC-ESI-MS/MS) for the analysis of 1,2,3,4-
tetrahydro-â-carboline derivatives. J . Chromatogr. A 1997,
767, 101-106.
food sample
seasoning sauce
3a /b
4a /b
5a /b
+++
+++
+
+++
+++
+
+++
+++
+
+
nd
nd
nd
nd
nd
nd
nd
nd
nd
soy sauce
worcestershire sauce
yeast extract
white wine
red wine
vinegar 1
vinegar 2
sherry
beer
++
+
++
+
+
+
+
+
+
+
+
+
+
+
juice prepared from dried plums
fruit syrup
caramel color
+
nd
+
+
nd
nd
a
Symbols: +++, >10 µg/mL; ++ >500 ng/mL; +, >10 ng/mL;
nd; not detectable.
Harvey, D. G., Miller, E. J ., Robsen, W. The Adamkiewicz,
Hopkins and Cole, and Rosenheim tests for Tryptophan. J .
Chem. Soc. 1941, 153-158.
Herraiz, T. Occurrence of Tetrahydro-â-carboline-3-carboxylic
acids in commercial foodstuffs. J . Agric. Food Chem. 1996,
CH-COSY-, and HMBC-NMR experiments. Subse-
quently, analytical data such as retention times, pro-
tonated molecular ions, and characteristic product ion
spectra of synthesized reference compounds confirmed
identification of novel THCC derivatives in food samples
under study. Any artifactual formation of THCCs
during sample preparation could be excluded because
no d4-labeled THCC derivatives were detectable after
addition of L-tryptophan-d5 prior to preconcentration
4
4, 3057-3065.
Herraiz, T.; Huang, Z.; Ough, C. S. 1,2,3,4-Tetrahydro-â-
carboline-3-carboxylic acid and 1-methyl-1,2,3,4-tetrahydro-
â-carboline-3-carboxylic acid in wines. J . Agric. Food Chem.
1
993, 41, 455-459.
J acobs, W. A.; Craig, L. C. The ergot alkaloids: The synthesis
of 4-carboline carbonic acids. J . Biol. Chem. 1936, 113,
(Gutsche et al., 1996).
7
59-765.
In order to clarify the relevance of the newly identified
McKenna, D. J .; Towers, G. H. N. Biochemistry and pharma-
cology of tryptamines and â-carbolines: A Minireview. J .
Psychoact. Drugs 1984, 16, 347-358.
Papavergou, E.; Clifford, M. N. Tetrahydro-â-carbolinecar-
boxylic acids in smoked foods. Food Addit. Contam. 1992,
THCCs numerous food samples including seasoning
sauce, alcoholic beverages, vinegar, and yeast extract
were analyzed. Results are summarized in Table 2.
Notably, most samples under study contained dicar-
boxylic acids 3a /b and 4a /b together with established
THCCs 1 and 2a /b. 5a /b was identified in soy sauce,
worcestershire sauce, seasoning sauce, and yeast ex-
tract.
9
, 83-95.
Rommelspacher, H.; Damm, H.; Strauss, S.; Schmidt, G.
Ethanol induces an increase of harman in the brain and
urine of rats. Naunyn-Schmiedeberg’s Arch. Pharmacol.
1
984, 327, 107-113.
CONCLUSION
Rommelspacher, H.; May, T.; Salewski, B. Harman (1-methyl-
â-carboline) is a natural inhibitor of monoamine oxidase type
A in rats. Eur. J . Pharmacol. 1994, 252, 51-59.
Sen, N. P.; Seaman, S. W.; Lau, P. P.-Y.; Weber, D. Determi-
nation and occurrence of various tetrahydro-â-carboline-3-
carboxylic acids and the corresponding N-nitroso compounds
in foods and alcoholic beverages. Food Chem. 1995, 54,
327-337.
Spies, C. D.; Rommelspacher, H.; Schnapper, C.; Muller, C.;
Marks, C.; Berger, G.; Conrad, C.; Blum, S.; Specht, M.;
Hannemann, L.; Striebel, H. W.; Schaffartzik, W. â-Carbo-
lines in chronic alcoholics undergoing elective tumor resec-
tion. AlcoholismsClin. Exp. Res. 1995, 19, 969-976.
Wakabayashi, K.; Ochiai, M.; Saito, H.; Tsuda, M.; Suwa, Y.;
Nagao, M.; Sugimura, T. Presence of 1-methyl-1,2,3,4-
tetrahydro-â-carboline-3-carboxylic acid, a precursor of a
mutagenic nitroso compound, in soy sauce. Proc. Natl.
Acad. Sci. U.S.A. 1983, 80, 2912-2916.
Tetrahydro-â-carboline-1,3-dicarboxylic acids 3a/b and
a /b as well as 1-(2′-carboxyethyl) derivatives 5a /b were
4
identified for the first time in numerous food samples
and underlined the relevance of Pictet-Spengler con-
densation for the formation of tryptophan-derived deg-
radation products from R-oxo acids. In addition, HPLC-
MS/MS was demonstrated to efficiently facilitate studies
examining the structural variety of tryptophan-derived
alkaloids. In combination with research on nutritional
biochemistry and toxicology of THCC derivatives 3a /b,
4
a /b, and 5a /b, HPLC-MS/MS will provide a reliable
analytical basis for elucidating the relevance of indi-
vidual â-carbolines in food.
ACKNOWLEDGMENT
Yamada, S.; Akimoto, H. Reductive decyanization of R-amino
We gratefully acknowledge the assistance of M.
Schmitt and B. Pink with NMR spectroscopy.
nitriles with NaBH
4
: A new synthetic approach to isoquino-
line- and indole-alkaloids. Tetrahedron Lett. 1969, 36,
105-3108.
3
LITERATURE CITED
Adachi, J .; Mizoi, Y.; Naito, T.; Ogawa, Y.; Uetani, Y.;
Ninomiya, I. Identification of tetrahydro-â-carboline-3-car-
boxylic acid in foodstuffs, human urine and human milk.
J . Nutr. 1991, 121, 646-652.
Arai, S.; Abe, M.; Yamashita, M.; Kato, H.; Fujimaki, M.
Applying proteolytic enzymes on soybean. Agric. Biol.
Chem. 1971, 35, 552-559.
Received for review December 13, 1996. Revised manuscript
received April 9, 1997. Accepted April 10, 1997. This study
was supported by a generous grant (He 2599/1-2) of the
Deutsche Forschungsgemeinschaft DFG, Bonn, Germany.
X
J F960952H
Argoudelis, C. J . In vitro reaction of L-tryptophan and vitamin
X
B
6
. Synthesis of the corresponding â-tetrahydrocarbolines.
Abstract published in Advance ACS Abstracts, J une
J . Agric. Food Chem. 1994, 42, 2372-2375.
15, 1997.