Photooxidation of tryptophan leading to 2-Aminoacetophenone - A possible reason for the untypical aging off-flavor in wine

Article abstract of DOI:10.1111/php.12321

2-Aminoacetophenone (AAP) was recognized as the key compound for the so-called untypical aging off-flavor (UTA) in Vitis vinifera wines. In this study, it was shown that AAP can be formed by photooxidation of free and protein-bound tryptophan (TRP) in combination with a subsequent storage in model wine. Solutions of TRP and lysozyme were exposed to artificial sunlight both in the presence and in the absence of the photosensitizer riboflavin. Aliquots of the irradiation batches were stored in model wine solutions containing tartaric acid, sulfite and ethanol in different combinations. AAP formation could be identified from both free and bound (lysozyme) TRP, while free TRP resulted in higher yields. The presence of riboflavin during irradiation generally favored the AAP formation. AAP formation increased with increasing irradiation times, but AAP was not detectable, if TRP was directly incubated in model wine. Not only the irradiation time but also the storage time of model wines favored the formation of AAP. Concerning the model wine composition, it became evident that the presence of tartaric acid resulted in the highest AAP formation during storage.

Full text of DOI:10.1111/php.12321

Photochemistry and Photobiology, 2014, 90: 1257–1263
Photooxidation of Tryptophan Leading to 2-Aminoacetophenone – A
Possible Reason for the Untypical Aging Off-flavor in Wine
Nora Horlacher and Wolfgang Schwack*
Institute of Food Chemistry, University of Hohenheim, Stuttgart, Germany
Received 27 June 2014, accepted 18 July 2014, DOI: 10.1111/php.12321
ABSTRACT
degradation products. Experiments with TRP under the same
conditions also yielded FAP and AAP, but to a lower extend,
whereas the formation of FAP and AAP could not be observed
from kynurenine (KYN) and N’-formylkynurenine (NFK). The
sulfite depending formation of AAP from IAA was later con-
firmed by Hoenicke et al. (5), while they also excluded KYN as
a precursor of AAP. During yeast fermentation experiments, IAA
spiked to model wines at levels far above the natural occurrences
revealed AAP in rather low yields (5–7). Therefore, this way of
formation was considered not being relevant. Concerning KYN,
the fermentative formation of AAP was reported both as not rele-
vant (5,6) and highly detectable (7). Different experimental con-
ditions, especially the type and amount of yeast, may explain the
discrepancies.
Since the influence of sunlight on the formation of AAP from
IAA was already shown (8), the photooxidation of TRP was of
major concern in literature. The published studies strongly dif-
fered in the experimental conditions, as for the form of TRP
(free or bound), source of light, kind of medium (buffer, pH),
presence of oxygen as well as photosensitizers (9,10). Thus, a
direct comparison of the results is difficult. Some of the major
reported photoproducts are NFK and KYN (Fig. 1B), which
were identified after irradiation of both free and protein-bound
TRP (11–15). Summaries of the pathways of the photooxidation
of TRP residues in proteins generally propose two major mecha-
nisms (9,10,16). Direct UV absorption generates excited state
species and radicals, resulting in C-3 peroxyl radicals after reac-
tion with oxygen (Fig. 1B). The singlet oxygen-mediated path-
way leads to a dioxetane or a hydroperoxide, which can
decompose to NFK yielding KYN by hydrolysis. As mentioned
above, the chemical relationship of NFK and KYN with FAP
and AAP led authors assume that they could be potential precur-
sors of AAP. The fact that NFK and KYN were identified as
photoproducts of TRP encouraged us to fill the gap between the
well-known TRP photochemistry and wine processing on the
way of AAP formation. Therefore, the aim of this study was to
show, if UV light irradiation of free or bound TRP in combina-
tion with a subsequent storage in model wine yields AAP to cor-
roborate the hypothesis that the exposure of extra- or
intracuticular wine berry proteins to sunlight results in potential
precursors of AAP. After extraction into the must and during
wine storage, the formation of AAP takes place. However, light
exposure could also occur at a later stage as during storage of
white wines in clear bottles. Thus, intrinsic wine peptides or resi-
dues of fining proteins like lysozyme, or even free TRP could
undergo photodegradation in the presence of riboflavin in bottled
2-Aminoacetophenone (AAP) was recognized as the key com-
pound for the so-called untypical aging off-flavor (UTA) in
Vitis vinifera wines. In this study, it was shown that AAP can
be formed by photooxidation of free and protein-bound tryp-
tophan (TRP) in combination with a subsequent storage in
model wine. Solutions of TRP and lysozyme were exposed to
artificial sunlight both in the presence and in the absence of
the photosensitizer riboflavin. Aliquots of the irradiation
batches were stored in model wine solutions containing tar-
taric acid, sulfite and ethanol in different combinations. AAP
formation could be identified from both free and bound (lyso-
zyme) TRP, while free TRP resulted in higher yields. The
presence of riboflavin during irradiation generally favored
the AAP formation. AAP formation increased with increasing
irradiation times, but AAP was not detectable, if TRP was
directly incubated in model wine. Not only the irradiation
time but also the storage time of model wines favored the
formation of AAP. Concerning the model wine composition,
it became evident that the presence of tartaric acid resulted
in the highest AAP formation during storage.
INTRODUCTION
The untypical aging (UTA) in wine is an off-flavor, which is
described as floor polish, wet wool, fusel alcohol, naphthalene or
acacia blossom (1,2). 2-Aminoacetophenone (AAP) was identi-
fied by Rapp et al. (3) as the key compound mainly being
responsible for UTA. Already AAP concentrations of 0.5–
1.5 lg LÀ1 and >1.5 lg LÀ1 in white and red wines, respec-
tively, can result in UTA (1). Masking effects by other wine fla-
vors and wine compounds could be responsible for higher odor
thresholds in strong white wines and red wines (1). Reasons for
the formation of AAP in wine were extensively studied. Because
of structural relations, tryptophan (TRP) and its metabolites were
regarded as the precursors of AAP, when the performed studies
differentiated in enzymatic and nonenzymatic pathways. Starting
from indole-3-acetic acid (IAA), reaction pathways of oxidative
degradation by radical cooxidation of sulfite were discussed by
Christoph et al. (4), resulting in AAP in wine (Fig. 1A). The
mechanisms were supported by the additional detection of
skatole (SKA) and N-formyl-2-aminoacetophenone (FAP) as
*Corresponding author email: wolfgang.schwack@uni-hohenheim.de (Wolfgang
Schwack)
© 2014 The American Society of Photobiology
1257

1258 Nora Horlacher and Wolfgang Schwack
Figure 1. (A) Postulated pathways of AAP formation from IAA (4); indole-3-acetic acid (IAA), skatole (SKA), N-formyl-2-aminoacetophenone (FAP),
2-aminoacetophenone (AAP). (B) Mechanism of sensitized photooxidation of tryptophan (TRP) (10); N’-formylkynurenine (NFK), kynurenine (KYN).
nitrogen, and the residue taken up in 10 mL ethanol. The solution was
diluted 1:100 with ethanol, to be used as internal standard solution. The
residual nonconverted AAP, determined by GC-MS/MS, was below the
limit of detection (LOD). For NMR analysis, AAP (5 mg, 0.037 mmol)
was treated as described for synthesis. After stirring for 4 h, d4-methanol
was slowly removed under a stream of nitrogen. Carbon tetrachloride
(500 lL) was added and removed under a stream of nitrogen, which was
repeated three times. The residue was taken up in 500 lL of CDCl3,
transferred through a syringe filter into an NMR tube and filled up to
with 600 lL CDCl3. MS (EI, 70 eV): m/z (%) 138 (92), 120 (100), 92
(56), 65 (29), 46 (10). 1H-NMR (300 MHz, CD3OD): d = 7.74 (dd,
J = 8.1 Hz, J = 1.3 Hz, 1 H, 6-H), 7.23 (ddd, J = 8.2 Hz, J = 7.0 Hz,
J = 1.4 Hz, 1 H, 4-H), 6.72 (br d, J = 8.2 Hz, 1 H, 3-H), 6.58 (ddd,
J = 8.1 Hz, J = 7.0 Hz, J = 1.0 Hz, 1 H, 5-H). 1H-NMR (300 MHz,
CDCl3): d = 7.71 (dd, J = 8.2 Hz, J = 1.4 Hz, 1 H, 6-H), 7.26 (ddd, 4-
H, overlaid by CDCl3), 6.61-6.68 (m-like, 2 H, 3-H and 5-H), 6.25 (br s,
2 H, -NH2).
N-Formyl-2-aminoacetophenone (FAP). Following a method described
in literature (5,19), AAP (0.9 g, 6.7 mmol) was added to a ice-cooled
mixture of acetic anhydride (1.4 mL) and formic acid (2.2 mL). After stir-
ring on ice for 30 min and further for 3.5 h at ambient temperature,
diethyl ether (30 mL) and saturated sodium hydrogencarbonate solution
(15 mL) were added. Extraction was performed by shaking in a separatory
funnel and repeated once with 30 mL diethyl ether. The combined organic
extracts were dried over sodium sulfate and rotatory evaporated. The resi-
due was recrystallized from ethanol, dried in a desiccator and yielded
FAP as colorless crystals (0.8 g, 4.9 mmol, 74%). GC-MS analysis con-
firmed the absence of AAP. MS (EI): m/z (%) 163 (26), 148 (10), 135
(61), 120 (100), 92 (35), 65 (23), 43 (28). 1H-NMR (300 MHz, CDCl3):
d = 11.60 (br s, 1 H, –CHO), 8.75 (br d, J = 8.3 Hz, 1 H, 6-H), 8.50 (s,
1 H, -NH-), 7.92 (br d, J = 7.9 Hz, 1 H, 3-H), 7.57 (br t, J = 7.9 Hz, 1
H, 4-H), 7.17 (br t, 1 H, J = 7.5 Hz, 5-H), 2,65 (s, 3 H, -CO-CH3).
wines, which also is associated with the oxygen management
during the vinery process (17).
MATERIALS AND METHODS
Materials. 2-Aminoacetophenone (AAP) (98%), L-tryptophan (≥99.5%),
ethyl acetate (≥99.8%), and ethanol (≥99.8%), d4-methanol (99.96 atom%
D), deuterochloroform (CDCl3, 99.8 atom% D) and formic acid (98–
100%) were purchased from Sigma-Aldrich (Steinheim, Germany).
Diethyl ether (≥98%) was purchased from AppliChem (Darmstadt, Ger-
many) and freshly distilled prior use. Potassium pyrosulfite was obtained
from Hefereinzucht Schlag (Aalen, Germany), and L-(+)-tartaric acid
(pure) from AppliChem (Darmstadt, Germany). Riboflavin (98%), sodium
tert-butylate (98%), sodium hydroxide pellets (≥99%), acetic anhydride
(99%), sodium hydrogen carbonate (≥99%), sodium sulfate (>99%) and
carbon tetrachloride (>99.8%) were from Merck (Darmstadt, Germany).
Carbon tetrachloride was distilled twice. Water was purified by a Synergy
ultrapure water system (Merck Millipore, Schwalbach, Germany). Lyso-
zyme from chicken egg white (81989 U mgÀ1) was obtained from
Sigma-Aldrich; according to (18), the TRP content was determined to
7.3 g/100 g (n = 2; %RSD = 0.5).
Synthesis of reference substances. The synthesized reference sub-
stances were identified by 1H-NMR and GC-MS. NMR spectra were
recorded on a Varian (Darmstadt, Germany) Unity Inova-300 spectrome-
ter at 300 MHz. The samples were dissolved in CDCl3, or CD3OD. The
chemical shifts were referenced to the residual solvent signals at d 7.26
(CDCl3) and 3.31 (CD3OD).
d3-2-Aminoacetophenone (d3-AAP). AAP (3 mg, 0.022 mmol) was
dissolved in 500 lL of d4-methanol, and catalytic amounts of sodium
tert-butylate (2 mg) were added. After stirring for 4 h, 15 lL of formic
acid (2.6 M) was added. The solvent was removed under a stream of

Photochemistry and Photobiology, 2014, 90 1259
Irradiation experiments. L-Tryptophan (0.046 g, 0.23 mmol) or lyso-
zyme (0.64 g; 0.045 mmol) was dissolved in 50 mL water. For the cata-
lyzed batches, 1 mg (2.7 lmol) or 12.5 lg (0.033 lmol) riboflavin was
added. Irradiation was performed with 25 mL of the solutions in 50 mL
quartz beakers (diameter 38 mm, Geyer, Renningen, Germany) with tef-
lon caps, which were placed in a self-constructed cooling chamber made
from quartz glass (12 9 5 9 5 cm) operated by a chiller (thermostat
constant pressure 15 psi; ion source: 230°C, electron energy 70 eV;
multiple reaction monitoring (MRM) mode: collision gas N2; quench
gas He. Retention times, collision energy and MRM transitions are
summarized in Table 1. Agilent MassHunter qualitative analysis B.06.00
was used for data acquisition.
Mass spectra for identification of reference substances were recorded
on an Agilent (Waldbronn, Germany) 6890 gas chromatograph coupled
to an 5973 MSD (Agilent) equipped with an HP5-ms Ultra Inert column
(30 m 9 0.25 mm 9 0.25 lm) (Agilent). Injector settings: splitless; inlet
temperature: 60°C, injection volume 2 lL; oven temperature program:
60°C (0.2 min hold) to 150°C (1 min hold) (ramp of 10°C minÀ1) to
250°C (1 min hold) (ramp of 50°C minÀ1); carrier gas: He, constant flow
(0.8 mL minÀ1); ion source: 230°C, electron energy 70 eV; scan mode:
30–800 amu. Agilent MSD ChemStation E.02.00.493 was used for data
acquisition.
€
model WK 230, Lauda, Lauda-Konigshofen, Germany). Temperature of
the solutions was kept at 25 Æ 0.5°C under constant stirring with a mag-
netic stirrer, Variomag Micro (Thermo Scientific). The light source was a
€
€
sun simulator SOL 500 (Dr. Honle, Grafelfing, Germany) with an inte-
grated metal halogen lamp (UVA 9.7 mW cmÀ² and UVB 0.25
mW cmÀ²). The front filter glass was replaced by an aluminum plate
with two windows (4 9 4 cm² each) to hold WG 295 glass filters
(Schott, Mainz, Germany). The cooling chamber was directly positioned
in front of the windows to completely expose the solutions to UV light.
Storage experiments. Irradiated and nonirradiated solutions were trans-
formed into different model wines. Depending on the aimed composi-
tions, 4 mL of sample solution was spiked with 50 lL of an aqueous
L-(+)tartaric acid solution (270 g LÀ1), 50 lL of an aqueous potassium
pyrosulfite solution (8.66 g LÀ1) and 650 lL of ethanol. The pH of the
model wines was adjusted to 3.3 with sodium hydroxide solution (1 M),
which finally were filled up with water to 5 mL. Solutions of TRP,
which were transformed into model wine W4, were stored at their native
pH values, which were between 3.6 and 3.8. Thus, following model
wines were obtained: model wine W1 (according to (20): 2.7 g LÀ1
L-(+)-tarÀta1ric acid, 50 mg LÀ1 SO2, 13 Vol% ethanol), model wine W2
Statistics. Statistics were performed using IBM SPSS Statistics for
Windows, Version 19.0. Armonk, NY: IBM Corp. released 2010. The
significance level was chosen to be 0.05 and was tested with two sample
t-test.
RESULTS AND DISCUSSION
Determination of AAP and FAP by GC-MS/MS
The difficulty of AAP quantification in wine is due to strong
matrix effects, caused by numerous volatile wine flavors, and
due to the low concentrations in wines. In literature, several
methods of the analysis of AAP in wine are described, mainly
using gas chromatography with mass spectrometric detection
(GC-MS) (20–24), when one-dimensional as well as multidimen-
sional GC methods (20,21) were applied. Extraction of AAP was
performed by liquid–liquid extraction with trichlorofluorome-
thane (4,21) or n-pentane (22), LiChrolut EN supported solid–
liquid extraction (20), and by direct immersion (23) or headspace
solid-phase microextraction (24). Besides ethyl vanillin (4), 2,
4-dichloroaniline (21), d8 -acetophenone (23), d3-AAP (20,22),
or d5-AAP (24) were used as internal standards. As we per-
formed our experiments only in model wine solutions, strong
matrix effects were not to be expected. To develop a fast and
simple method for the quantitation of AAP and FAP, a simple
liquid–liquid extraction with ethyl acetate was chosen. In litera-
ture, however, an influence of pH and sulfurous acid on the
extraction yields of AAP was discussed (20,22,24), but the
obtained results were inconsistent. Therefore, we decided to pre-
pare calibration standards in model wines to take the extraction
process into account, which also may refer to a kind of matrix-
matched standards. Fan et al. (23) also used model wines for
calibration, to be extracted by solid-phase microextraction. Repli-
cates of extraction (n = 6) from a model wine W1 (199 lg LÀ1
AAP) resulted in a relative standard deviation of 1.9%. The
determination of AAP and FAP was performed by GC-MS/MS
in the highly selective multiple reaction monitoring (MRM)
mode (Table 1), when, according to (22), d3-AAP was used as
internal standard. The obtained calibration graphs with high coef-
ficients of correlation are exemplarily shown in Fig. 2. Following
the DIN 32645 method (25), LODs of 0.5 lg LÀ1 and
(2.7 g L
L-(+)-tartaric acid, 50 mg LÀ1 SO2), model wine W3
(2.7 g LÀ1 L-(+)-tartaric acid, 13 Vol% ethanol) and model wine W4
(50 mg LÀ1 SO2, 13 Vol% ethanol). The model wines were transferred
into 10 mL screw-capped vials and stored protected from light at 40 or
18°C. After defined periods of storage (0, 1, 3, 5 and 20 weeks), aliquots
were sampled and prepared for GC-MS/MS analysis.
Sample preparation. Samples of the stored model wines (2 mL) were
pipetted into 5 mL screw-capped centrifuge tubes, spiked with 100 lL of
the internal standard solution, followed by the addition of 1 mL ethyl
acetate. After extraction for 1 min, the emulsions were centrifuged at
10 000 U minÀ1 for 20 min. The organic layer was drawn into a syringe,
transferred through a 0.45-lm filter tip (VWR International, Darmstadt,
Germany) into a GC vial. The extraction was performed in duplicates
and the mean was calculated. Samples of irradiated free TRP, which were
stored in model wine W1 at 40°C, were diluted 1:10 in model wine
before extraction.
Calibration. For calibration, stock solutions of AAP and FAP (ca
40 mg LÀ1) were prepared in water and diluted 1:20 and 1:2000 with
water. The obtained standard solutions with concentrations of ca
0.02 mg LÀ1 (standard 1) and 2 mg LÀ1 (standard 2) were transformed
into model wines. Therefore, 500 lL of the standard 1 solution and 25,
60, 125, 250 and 1250 lL of the standard 2 solution were transformed
into the respective model wine, as described in Storage Experiments,
resulting in AAP and FAP concentrations of 2–500 lg LÀ1. Aliquots
(2 mL) were treated as described for sample preparation. Thus, calibra-
tion points were expressed as model wine concentrations. The extraction
was performed in duplicates and the mean was calculated.
Gas chromatography–mass spectrometry (GC-MS/MS and GC-MS).
An Agilent (Waldbronn, Germany) 7890A gas chromatograph equipped
with an ALS 7693 autosampler and coupled to
Quad 7000 (Agilent) was used for GC-MS/MS analysis. Separation
was performed on an HP5-ms Ultra Inert column (30
a GC-MS Triple
m
9
0.25 mm 9 0.25 lm) (Agilent). Injector settings: pulsed splitless; inlet
temperature program: 60°C (1 min hold) to 250°C (10 min hold) (ramp
of 900°C minÀ1), injection volume 2 lL; oven temperature program:
60°C (0.2 min hold) to 150°C (1 min hold) (ramp of 10°C minÀ1
)
to 250°C (1 min hold) (ramp of 50°C minÀ1); carrier gas: He,
Table 1. Retention times, collision energy, MRM transitions for quantifiers and qualifiers and sensitivity of the GC-MS/MS method; LODs and LOQs
were calculated from the calibration data according to DIN 32645 (25).
Compound
Retention time (min)
Collision energy (eV)
Quantifier MRM
Qualifier MRM
LOD (lg LÀ1
)
LOQ (lg LÀ1
)
d3-AAP
AAP
FAP
9.62
9.66
12.37
20
20
20
138 > 120
135 > 120
135 > 120
138 > 92
135 > 92
163 > 120
–
0.5
0.8
–
1.5
2.0

1260 Nora Horlacher and Wolfgang Schwack
that is discussed to be a direct precursor of AAP (4,5) could also
be identified in the stored model wines prepared from irradiated
TRP and lysozyme solutions, but was below the LOQ and LOD,
if nonirradiated free and lysozyme-bound TRP was applied. As
compared to AAP, the FAP concentrations were clearly lower
both from free and protein-bound TRP (Table 2). Calculating the
sum of AAP and FAP, ca 0.09 and 0.009 mol% of free and lyso-
zyme-bound TRP, respectively, were productively transformed,
resulting in final AAP concentrations of 432 and 45 lg LÀ1 in
the studied model wines.
Influence of storage time and storage temperature
The influence of storage time and storage temperature on the for-
mation of AAP was studied with nonirradiated and irradiated
samples of lysozyme, transformed into model wine W1 that was
extracted immediately and after storage for 1–20 weeks at 40°C
and 18°C. Meanwhile, it did not come as a surprise that AAP
formation increased with both storage time and storage tempera-
ture. As compared to 18°C, storage at 40°C significantly
enhanced the formation of AAP (Fig. 3). Only after 1 week at
40°C, AAP strongly exceeded the value obtained after 5 weeks
at 18°C. It is known from the literature that a higher temperature
and increasing storage time enforces the formation of AAP (4,8).
Studies on the existence of AAP at different stages of vinifica-
tion showed that in the wine berry, in the must and in the wine
direct after fermentation, AAP concentrations were mostly below
the odor threshold (1,7,35). Higher concentrations were only
found after fermentation in the course of storage, which is in
accordance with our results. In model wines immediately
extracted after preparation, AAP could well be determined in
samples from irradiated free TRP, but were below LOQ in sam-
ples from irradiated lysozyme (Table 2). Thereafter, AAP con-
centration further increased with increasing storage time.
Concerning FAP, the highest concentrations were found immedi-
ately after the irradiation process (Table 2), which decreased
with storage time, especially at the higher temperature of 40°C
(Table 2, Fig. 3), when already after 3 weeks it was below the
LOQ. At the lower temperature of 18°C, FAP only decreased by
ca 60% within the longer storage time of 5 weeks (Fig. 3).
These findings are well in agreement with the results published
by Christoph et al. (4). They also reported an enormous thermal
instability of FAP. Under the influence of heat—also in the
absence of sulfite—FAP degraded rapidly. In solutions of FAP
stored in model wine for 4 days at 45°C and further 4 weeks at
20°C, 90 mol% of the added FAP was transformed to AAP.
Figure 2. Linear calibration graphs for AAP and FAP (duplicates of
extraction).
0.8 lg LÀ1 and limits of quantitation (LOQs) of 1.5 and
2.0 lg LÀ1 were achieved for AAP and FAP, respectively
(Table 1).
AAP formation from free and bound tryptophan
The formation of AAP by photooxidation of TRP in combination
with a subsequent storage under model wine conditions was first
studied with free TRP in the presence of the photosensitizer ribo-
flavin (2.7 lmol/50 mL), when the irradiation time was 15 h.
After storage for 1 week at 40°C in model wine W1, AAP clearly
could be detected (Table 2). Under the same conditions, equimo-
lar amounts of protein-bound TRP (lysozyme) also resulted in the
formation of AAP, although to a nearly ten-fold lower extent
(Table 2). This is to be explained by a lower degree of photooxi-
dation of the lysozyme-bound than of free TRP. Edwards and
Silva (26) already showed for lysozyme and a-lactalbumine that
the accessibility of the TRP residues to the solvent, thus exposure
to UV light, influences their photosensitivity. Increasing accessi-
bility led to a higher degree of photooxidation. They reported that
only two of the six TRP residues of lysozyme are exposed (26).
Besides the accessibility of TRP residues, the yield of AAP from
photooxidized TRP residues is another variable, which altogether
will explain our results. However, in model wines prepared from
nonirradiated TRP solutions and stored under the same condi-
tions, AAP formation was below the LOD for free and lysozyme-
bound TRP. From these first results, it clearly could be concluded
that a former photooxidation is a prerequisite for the formation of
AAP during the incubation under model wine conditions. FAP
Table 2. AAP and FAP formation from free and lysozyme-bound TRP, irradiated for 15 h in the presence of riboflavin (2.7 lmol/50 mL) and
transformed into model wines stored at 40°C.
Free TRP
Lysozyme
Storage time
(weeks)
Model wine
AAP Æ SD (lg LÀ)
FAP Æ SD (lg LÀ)
AAP Æ SD (lg LÀ)
FAP Æ SD (lg LÀ)
W1
W1
0
1
3
5
1
3
5
24.9 Æ 3.5 (n = 3)
432 Æ 4.5 (n = 2)
444 Æ 3.3 (n = 2)
824 Æ 10.2 (n = 2)
232 Æ 7.6 (n = 2)
377 Æ 2.2 (n = 2)
382 Æ 9.1 (n = 2)
581 Æ 88.3 (n = 3)
142 Æ 10.6 (n = 2)
<LOQ (n = 3)
27.9 Æ 0.9 (n = 3)
14.5 Æ 0.8 (n = 3)
<LOQ (n = 5)
45.0 Æ 1.3 (n = 3)
71.9 Æ 7.5 (n = 5)
92.7 Æ 5.0 (n = 6)
28.5 Æ 0.7 (n = 2)
50.1 Æ 0.8 (n = 2)
54.5 Æ 1.6 (n = 2)
<LOQ (n = 2)
<LOQ (n = 2)
<LOQ (n = 6)
W4
572 Æ 47.8 (n = 2)
204 Æ 10.8 (n = 2)
115 Æ 16.3 (n = 2)
9.8 Æ 2.3 (n = 2)
3.7 Æ 0.4 (n = 2)
<LOQ (n = 2)

Photochemistry and Photobiology, 2014, 90 1261
used the model protein lysozyme for our experiments, an appro-
priate TRP source has to be identified to transform the results to
real conditions. Both intrinsic wine proteins and proteins present
on the surface of wine berries as well as free TRP may be
involved. Thus, exudates of micro- or macroorganisms exempl-
arily come into question, or the biological insecticide Bacillus
thuringiensis (Bt), which is used for vine protection. The
Cry1Ab toxin, for example, which is produced during sporula-
tion of Bt (30), owns 17 TRP residues (31,32). The sunlight-
mediated destruction of up to 65% of TRP residues in endotoxin
crystals from Bt was already shown in literature (33,34).
Influence of model wine composition
The so far presented results were obtained by the transformation
of irradiated solutions of free and bound TRP into the well-estab-
lished model wine W1 and stored. To study the effects of acid,
SO2 and ethanol during storage on the formation of AAP, sam-
ples of lysozyme, irradiated in the presence of riboflavin
(2.7 lmol/50 mL), were transformed into three additional model
wines. As compared to model wine W1, model wine W2 was
free of ethanol, model wine W3 free of SO2 and model wine W4
free of acid, while the other respective components remained
constant. The thus obtained model wines were generally stored
at the optimal conditions of 5 weeks and 40°C. The highest
AAP amounts were found in model wines containing all three
components, tartaric acid, SO2 and ethanol, i.e. model wine W1
(Fig. 5). The absence of ethanol (model wine W2) or SO2
(model wine W3) generally had the consequence of significantly
lower AAP formation, but the presence of only SO2 (W2) or eth-
anol (W3) under acidic conditions did not turn out a significant
difference concerning the formation of AAP. Although the high-
est AAP amounts could be determined, if tartaric acid was added
to the model wines (W1, W2 and W3), the presence of tartaric
acid was not essential for AAP formation. Samples transformed
into model wine W4 (without tartaric acid) also revealed AAP,
but the found AAP amounts were significantly lower than in the
presence of acid (model wine W1, W2 and W3) (Fig. 5). As to
Figure 3. Influence of storage time and temperature on AAP and FAP
formation. Solutions of lysozyme were irradiated for 15 h in the presence
of riboflavin (2.7 lmol/50 mL) and transformed into model wine W1.
Samples were analyzed without storage and after storage at 40°C or
18°C for 1, 3, 5 and 20 weeks; means Æ SD, n = 3 for the nonstored
samples, n = 3, 5, 6 and 2 for the samples stored at 40°C for 1, 3, 5 and
20 weeks, respectively, n = 4 (AAP) and n = 2 (FAP) for the samples
stored at 18°C; * <LOQ; **<LOD.
Influence of irradiation time
To demonstrate the influence of light exposure on the formation
of AAP, the irradiation time was varied from 0 to 45 h for solu-
tions of lysozyme in the presence of riboflavin (2.7 lmol/
50 mL). After the irradiations, W1 model wines were prepared
and stored for 3 weeks at 40°C. As to be expected, time of irra-
diation highly correlated with the formation of AAP in model
wines (Fig. 4). Already an irradiation time of 3 h led to an AAP
formation of 12.9 lg LÀ1, while it reached 99.8 lg LÀ1 after
45 h. Nonirradiated samples, however, did not result in detect-
able AAP amounts. The formation of FAP was below the LOD
up to 3 h and remained below the LOQ above 6 h of irradiation.
FAP was nearly completely degraded after a storage time of
3 weeks, which was only 1 week during the first experiments.
The influence of UV light on UTA is obvious, as UTA often
occurred in years with high insolation and low water supply
(27–29). Accordingly, the longer lysozyme was being exposed to
UV light the more AAP was formed in model wines. Since we
Figure 5. Influence of model wine composition on AAP formation dur-
ing storage. Model wines were prepared from lysozyme solutions irradi-
ated for 15 h in the presence of riboflavin (2.7 lmol/50 mL) and stored
for 5 weeks at 40°C. In addition, the influence of the photosensitizer
riboflavin during irradiation is exemplarily shown. Means Æ SD, n = 3
except the samples irradiated in the presence of riboflavin and stored in
W1 (n = 6) or W4 (n = 2); *Not studied; #Significant (P < 0.05), x not
significant.
Figure 4. Influence of irradiation time on AAP formation. Solutions of
lysozyme were irradiated in the presence of riboflavin (2.7 lmol/50 mL)
for 0–45 h and stored in model wine W1 for 3 weeks at 40°C;
means Æ SD, n = 2 except 0 h and 15 h (n = 5).

1262 Nora Horlacher and Wolfgang Schwack
be expected from the former results, the formation of FAP could
not be detected during these experiments (high temperature, long
storage time). To show the influence of acid on the formation of
FAP and its transformation to AAP, further experiments with the
following modifications were undertaken: (1) irradiated samples
were prepared without storage (t = 0) and storage time was addi-
tionally shortened to 1 week and 3 weeks; (2) besides lysozyme,
irradiated solution of free TRP was additionally applied; (3) stor-
age was performed in the acid and nonacid model wines W1 and
W4. As mentioned above, AAP formation increased with
extended storage times. This also held true for the present experi-
ments with model wines W1 and W4 prepared from irradiated
solutions of both free and lysozyme-bound TRP (Table 2). In
model wines prepared from irradiated solutions of free TRP, sig-
nificantly higher AAP concentrations could be determined in the
presence (W1) than in the absence of acid (W4) (Table 2). After
storage of 5 weeks, more than twofold of AAP was determined
in W1 compared to W4. The same trend could be shown at each
stage of storage of model wines W1 made from irradiated lyso-
zyme, whereas the AAP values were approximately half of them
in model wines W4 (Table 2). FAP was generally well detect-
able in model wines W1 and W4 freshly prepared (without stor-
age) from samples of both irradiated free and lysozyme-bound
TRP, but FAP was less stable during storage under the acidic
conditions of W1 (Table 2). In model wines prepared from irra-
diated solutions of free TRP, FAP decreased more than fourfold
in W1 only after 1 week, whereas it was still detectable after 3
and 5 weeks in W4 (Table 2). Concerning irradiated lysozyme,
FAP also more rapidly disappeared under the acidic conditions
of W1 than in W4 (Table 2).
Figure 6. Proposed pathways of AAP formation through free and pro-
tein-bound NFK and KYN.
at higher temperatures (40°C), FAP degradation occurred, while
the AAP concentration increased. However, FAP obviously was
not exclusively transformed into AAP. In model wines prepared
from TRP, steady losses of FAP occurred, while AAP did not
correspondingly increase (Table 2). Therefore, the dissipation of
FAP must follow at least one other route than hydrolysis, which,
however, could not be identified. In addition, AAP still increased
in model wine W1, when FAP was already disappeared. Thus,
there obviously was another AAP precursor different from FAP.
With regard to the known TRP photochemistry (Fig. 1B) (13–
17), NFK should provide the first access to AAP, followed by a
possible hydrolysis to KYN (Fig. 6). The ongoing formation of
AAP or FAP from KYN or NFK requires an elimination of the
amino acid head group. Thereby, the hydrolysis of NFK presum-
ably proceeds more rapidly than the elimination, still providing
KYN as AAP precursor, when NFK and FAP already disap-
peared (<LOQ). This could explain why AAP increased during
the storage of model wines, although FAP was not detectable
anymore. In the case of lysozyme, FAP was detectable after irra-
diation, but AAP also steadily increased, when FAP was below
the LOQ, and the finally determined AAP strongly exceeded the
original FAP concentrations (Table 2). Here, protein-bound KYN
and NFK have additionally to be taken into account as precur-
sors of FAP and AAP. Exploring the mechanisms in more detail
is surely highly interesting, but was out of scope of this study.
Influence of the photosensitizer riboflavin during irradiation
It is known that the presence of a photosensitizer during irradia-
tions increases the photochemical degradation of TRP. Due to
the formation of TRP-riboflavin adducts, especially for lysozyme,
a high efficiency of riboflavin as photosensitizer was described
(36,37). Riboflavin is present in most living systems, in the free
form or as flavin mononucleotide (FMN) and flavin adenine
dinucleotide (FAD) (41). Thus, its presence on the surface of
wine berries is not unlikely. To guarantee optimal conditions for
the formation of AAP, the so far chosen riboflavin concentrations
(2.7 lmol/50 mL) were rather high as compared to those, which
are typically present in grapes (38) and wines (39,40). Therefore,
additional irradiations of lysozyme were exemplarily performed
in the presence of more relevant riboflavin concentrations
(0.033 lmol/50 mL) and without riboflavin, when the irradiation
time was 15 h, and the storage conditions were 5 weeks at
40°C. AAP could successfully be detected also in model wine
W1 prepared from nonsensitized irradiation batches, but the
AAP concentration was about two-fold and six-fold, if
0.033 lmol and 2.7 lmol riboflavin, respectively, was present
during the irradiation (Fig. 5). Concerning the presence and the
absence of riboflavin during lysozyme irradiations, the model
wines W2 and W3 provided nearly the same results.
CONCLUSIONS
AAP, the character impact compound of the untypical aging off-
flavor (UTA) in wine, was clearly formed by UV light irradiation
of TRP (free or bound) in combination with a subsequent storage
in model wine. Extending the irradiation time increased the AAP
formation, but nonirradiated samples did not yield AAP during
storage in model wines. A clear conclusion is that AAP forma-
tion was photochemically initiated, and TRP can be regarded as
its primary precursor. However, lower AAP amounts were
formed from lysozyme-bound TRP than from the free amino
acid. Furthermore, we could show that the AAP formation mech-
anism proceeded via FAP, however, not exclusively, and is
affected by the set storage conditions. A faster transformation
was observed in the presence of acid and at the higher tempera-
ture of 40°C. Further research will be focused on appropriate
TRP sources and to prove its formation—according to this
model—under real conditions.
Pathways of AAP formation
As reported in literature (4,5), FAP was also identified in our
experiments as a precursor of AAP. During storage in model
wines, reinforced in the presence of acid (model wine W1) and

Photochemistry and Photobiology, 2014, 90 1263
20. Schmarr, H.-G., S. Ganß, W. Sang and T. Potouridis (2007) Analysis
of 2-aminoacetophenone in wine using a stable isotope dilution assay
and multidimensional gas chromatography-mass spectrometry.
J. Chromatogr. A 1150, 78–84.
21. Rapp, A., G. Versini and L. Engel (1995) Determination of
2-aminoacetophenone in fermented model wine solutions. Vitis 34,
193–194.
Acknowledgements—We express many thanks to Analytisches Institut
Bostel (Stuttgart, Germany) for providing the GC-MS/MS system and the
€
Landesanstalt f€ur Landwirtschaftliche Chemie, Universitat Hohenheim
(Stuttgart, Germany) for tryptophan analyses.
22. Dollmann, B., D. Wichmann, A. Schmitt, H. Koehler and P. Schreier
(1996) Quantitative analysis of 2-aminoacetophenone in off-flavored
wines by stable isotope dilution assay. J. AOAC Int. 79, 583–586.
23. Fan, W., I.-M. Tsai and M. C. Qian (2007) Analysis of 2-aminoace-
tophenone by direct-immersion solid-phase microextraction and gas
chromatography-mass spectrometry and its sensory impact in Char-
donnay and Pinot gris wines. Food Chem. 105, 1144–1150.
24. Gulan, S. and U. Arzberger (2008) Quantification of 2-aminoace-
tophenone in white wine by headspace solid phase micro extraction
and gas chromatography-mass spectrometry. Dtsch. Lebensm.-Run-
dsch. 104, 41–46.
REFERENCES
€
1. Christoph, N., C. Bauer-Christoph, M. Geßner and H. J. Kohler
(1995) Die “Untypische Alterungsnote” im Wein Teil I: Untersuch-
ungen zum Auftreten und zur sensorischen Charakterisierung der
“Untypischen Alterungsnote”. Rebe & Wein 48, 350–356.
2. Fischer, U. and R. Sponholz (2000) Die sensorische Beschreibung
der Untypischen Alterungsnote. Dtsch. Weinbau 3, 16–21.
3. Rapp, A., G. Versini and H. Ullemeyer (1993) 2-Aminoacetophe-
none: Causal component of “untypical aging flavour” (“naphthalene
note”, “hybrid note”) of wine. Vitis 32, 61–62.
25. Deutsches Institut f€ur Normung e.V. (2008) DIN 32645 Chemical
analysis – Decision limit, detection limit and determination limit
under repeatability conditions - Terms, methods, evaluation. Beuth,
Berlin.
€
4. Christoph, N., C. Bauer-Christoph, M. Geßner, H. J. Kohler, T. J. Si-
mat and K. Hoenicke (1998) Formation of 2-aminoacetophenone and
formylaminoacetophenone in wine by reaction of sulfurous acid with
indole-3-acetic acid. Vitic. Enol. Sci. 53, 79–86.
26. Edwards, A. M. and E. Silva (1985) Photochemical reactivity of the
homologous proteins a-lactalbumin and lysozyme. Radiat. Environ.
Biophys. 24, 141–148.
5. Hoenicke, K., O. Borchert, K. Gr€uning and T. J. Simat (2002)
“Untypical aging off-flavor” in wine: Synthesis of potential degrada-
tion compounds of indole-3-acetic acid and kynurenine and their
evaluation as precursors of 2-aminoacetophenone. J. Agric. Food
Chem. 50, 4303–4309.
€
27. Schwab, A., M. Peternel, J. Kohler and K.-P. Heigel (1996) Die Un-
typische Alterungsnote im Wein Teil IV: Beeinflussung durch we-
inbauliche Maßnahmen. Rebe & Wein 49, 181–187.
€
6. Geßner, M., H. J. Kohler, N. Christoph and C. Bauer-Christoph
€
28. Sponholz, W. R., T. Huhn, A. Engelmann and A. Siben (1997) Pos-
(1996) Die “Untypische Alterungsnote im Wein” Teil VII Untersuch-
ungen zur Bildung von o-Aminoacetophenon aus Produkten des
Tryptophan-Stoffwechsels bei der alkoholischen Garung. Rebe &
sible influence of some viticultural parameters on the incidence of
“untypical” off-flavours in Riesling wines. Vitic. Enol. Sci. 52, 41–
50.
€
Wein 49, 251–255.
7. Dollmann, B., A. Schmitt, H. Kohler and P. Schreier (1996) Forma-
29. Schwab, A. and M. Peternel (2001) Ungen€ugende Traubenreife: Urs-
ache der “Untypischen Alterungsnote (UTA)” im Wein. Rebe &
Wein 49, 14–17.
30. Schnepf, E., N. Crickmore, J. Van Rie, D. Lereclus, J. Baum, J. Fei-
telson, D. R. Zeigler and D. H. Dean (1998) Bacillus thuringiensis
and its pesticidal crystal proteins. Microbiol. Mol. Biol. Rev. 62,
775–806.
€
tion of the “untypical aging off-flavour” in wine: Generation of 2-
aminoacetophenone in model studies with Saccharomyces cerevisiae.
Vitic. Enol. Sci. 51, 122–125.
€
8. Christoph, N., C. Bauer-Christoph, M. Geßner and H. J. Kohler
(1996) Die “Untypische Alterungsnote” im Wein Teil VI Untersuch-
ungen zur Bildung von o-Aminoacetophenon aus Produkten des
31. The Universal Protein Resource (UniProt) UniProt Knowledgebase
(UniProtKB) P0A372 (CR1AB_BACTA). Available at: http://www.
uniprot.org/uniprot/P0A372. Accessed on 21 January 2014.
32. The Universal Protein Resource (UniProt) UniProt Knowledgebase
(UniProtKB) P0A370 (CR1AB_BACTK). Available at: http://www.
uniprot.org/uniprot/P0A370. Accessed on 21 January 2014.
33. Pozsgay, M., P. Fast, H. Kaplan and P. R. Carey (1987) The effect
of sunlight on the protein crystals from Bacillus thuringiensis var.
kurstaki HD1 and NRD12: A raman spectroscopic study. J. Inverte-
br. Pathol. 50, 246–253.
€
Tryptophan-Stoffwechsels vor der alkoholischen Garung. Rebe &
Wein 49, 246–250.
9. Creed, D. (1984) The photophysics and photochemistry of the near-
UV absorbing amino acids - I. Tryptophan and its simple derivatives.
Photochem. Photobiol. 39, 537–562.
10. Pattison, D. I., A. S. Rahmanto and M. J. Davies (2012) Photo-oxi-
dation of proteins. Photochem. Photobiol. Sci. 11, 38–53.
11. Bolte, M. and J. Lemaire (1989) Monohydrogen chromate photosen-
sitized oxidation of tryptophan exposed to near-UV-visible light.
J. Photochem. Photobiol., A 46, 285–294.
34. Pusztai, M., P. Fast, L. Gringorten, H. Kaplan, T. Lessard and P. R.
Carey (1991) The mechanism of sunlight-mediated inactivation of
Bacillus thuringiensis crystals. Biochem. J. 273, 43–47.
12. Benassi, C. A., E. Scoffone, G. Galiazzo and G. Iori (1967) Profla-
vine-sensitized photooxidation of tryptophan and related peptides.
Photochem. Photobiol. 6, 857–866.
€
35. Kohler, H., N. Christoph, C. Bauer-Christoph, M. Geßner and K.
13. Pirie, A. (1971) Formation of N’-formylkynurenine in proteins from
lens and other sources by exposure to sunlight. Biochem. J. 125,
203–208.
Curschmann (1996) Die “Untypische Alterungsnote” im Wein Teil
€
V: Einfluß kellerwirtschaftlicher Maßnahmen auf die Auspragung der
UTA. Rebe & Wein 49, 213–218.
14. Gomyo, T. and M. Fujimaki (1970) Studies on changes of protein
by dye sensitized photooxidation Part III. On the photodecomposi-
tion products of lysozyme. Agr. Biol. Chem. 34, 302–309.
15. Asquith, R. S. and D. E. Rivett (1971) Studies on the photooxidation
of tryptophan. Biochim. Biophys. Acta 252, 111–116.
16. Kerwin, B. A. and R. L. Remmele (2007) Protect from light: Photo-
degradation and protein biologics. J. Pharm. Sci. 96, 1468–1479.
17. Morozova, K., O. Schmidt and W. Schwack (2014) Impact of head-
space oxygen and copper and iron addition on oxygen consumption
rate, sulphur dioxide loss, colour and sensory properties of Riesling
wine. Eur. Food Res. Technol. 238, 653–663.
36. Silva, E. and J. Gaule (1977) Light-induced binding of riboflavin to
lysozyme. Radiat. Environ. Biophys. 14, 303–310.
37. Silva, E. (1979) Rate constants studies of the dye-sensitized photoin-
activation of lysozyme. Radiat. Environ. Biophys. 16, 71–79.
38. Souci, S. W., W. Fachmann and H. Kraut (2000) Food Composition
and Nutrition Tables (6th revised and completed ed.), p. 930. Medp-
harm, Stuttgart.
ꢀ
39. Andres-Lacueva, C., F. Mattivi and D. Tonon (1998) Determination
of riboflavin, flavin mononucleotide and flavin-adenine dinucleotide
in wine and other beverages by high-performance liquid chromatog-
raphy with fluorescence detection. J. Chromatogr. A 823, 355–363.
18. European Commision (2009) Commission Regulation (EC) No 152/
2009 of 27 January 2009 laying down the methods of sampling and
analysis for the official control of feed. Off. J. Eur. Commun. 54,
32–37.
ꢀ
40. Mattivi, F., A. Monetti, U. Vrhovsek, D. Tonon and C. Andres-La-
cueva (2000) High-performance liquid chromatographic determina-
tion of the riboflavin concentration in white wines for predicting
their resistance to light. J. Chromatogr. A 888, 121–127.
19. F€urstner, A. and D. N. Jumbam (1992) Titanium-induced syntheses
of furans, benzofurans and indoles. Tetrahedron 48, 5991–6010.
41. Choe, E., R. Huang and D. B. Min (2005) Chemical reactions and
stability of riboflavin in foods. J. Food Sci. 70, R28–R36.