Formation of peptide-bound Heyns compounds

Article abstract of DOI:10.1021/jf073256y

The reaction of the N^α-hippuryllysine (BzGK) with fructose was investigated in two model systems to obtain an insight in fructose-induced modification of lysine in bakery products. After BzGK and fructose had been heated in a buffered low-moisture model system (80°C, 60 min, a_W = 0.86, pH 7.4), formation of epimeric Heyns compounds N^α- hippuryl-N^ε-glucosyl-lysine (BzGGlcK) and N^α- hippuryl-N^ε-mannosyl-lysine (BzGManK) was clearly demonstrated using RP-HPLC with UV as well as MS detection. The Amadori compound N ^α-hippuryl-N^ε-fructosyl-lysine (BzGFruK) was formed in glucose-containing samples. When BzGK was added to the dough of fructose-containing biscuits, the Heyns compounds were detectable after baking at 175°C for 7 min. The yields of the Heyns compounds in the fructose-containing biscuits and the yield of the Amadori compound in the glucose-containing biscuits were determined to 33 and 63%, pointing to the fact that formation of substantial amounts of Heyns products is very likely in fructose-containing bakery products.

Full text of DOI:10.1021/jf073256y

2522 J. Agric. Food Chem. 2008, 56, 2522–2527
Formation of Peptide-Bound Heyns Compounds
RENÉ KRAUSE, KATHRIN SCHLEGEL, EVELYN SCHWARZER, AND THOMAS HENLE*
Institute of Food Chemistry, Technische Universität Dresden, D-01062 Dresden, Germany
The reaction of the NR-hippuryllysine (BzGK) with fructose was investigated in two model systems to
obtain an insight in fructose-induced modification of lysine in bakery products. After BzGK and fructose
had been heated in a buffered low-moisture model system (80 °C, 60 min, aW ) 0.86, pH 7.4),
formation of epimeric Heyns compounds NR-hippuryl-Nꢀ-glucosyl-lysine (BzGGlcK) and NR-hippuryl-
Nꢀ-mannosyl-lysine (BzGManK) was clearly demonstrated using RP-HPLC with UV as well as MS
detection. The Amadori compound NR-hippuryl-Nꢀ-fructosyl-lysine (BzGFruK) was formed in glucose-
containing samples. When BzGK was added to the dough of fructose-containing biscuits, the Heyns
compounds were detectable after baking at 175 °C for 7 min. The yields of the Heyns compounds in
the fructose-containing biscuits and the yield of the Amadori compound in the glucose-containing
biscuits were determined to 33 and 63%, pointing to the fact that formation of substantial amounts of
Heyns products is very likely in fructose-containing bakery products.
KEYWORDS: Maillard reaction; glycation; Heyns compound; Amadori compound; fructose; lysine
INTRODUCTION
of fructose with the ꢀ-amino group of protein-bound lysine and
the formation of Heyns compounds during thermal processing.
To examine this hypothesis, the reactions of the lysine-
containing dipeptide NR-hippuryllysine (BzGK) with fructose
and glucose were investigated in two model systems in detail
with regard to bakery products.
Thermal processing of food containing both reducing sugars
and proteins leads to the formation of pleasant flavors and a
desirable color intermediate between yellow and dark brown
in the course of the Maillard reaction. On the other hand, this
reaction can adversely affect the nutritional value of food by
the blocking of the essential amino acid lysine. The derivati-
zation of protein-bound lysine by aldoses and the formation of
Amadori compounds are well documented and widely used to
evaluate the extent of the early stage of the Maillard reaction
in food (1). In comparison to this, the derivatization of lysine
by ketoses is generally recognized (2, 3), but very little is known
about the occurrence of the so-called Heyns compounds as the
first stable products of this reaction in food. Heyns compounds
in particular were described within synthetic studies (4), and
low amounts of Heyns compounds of free amino acids were
found in long-stored lyophilized apricots (5), raw licorice (6),
tomato powder (7), and tobacco (8) in earlier work. Thin layer
chromatography (6, 8), paper chromatography (5, 8), and cation
exchange HPLC and refractive index detection (7) were used
for detection. In addition, protein-bound Heyns compounds were
identified in lens protein in vivo using a RP-HPLC method (9).
To the best of our knowledge, however, direct analysis of
protein-bound Heyns compounds in food samples has not been
described yet. As Heyns compounds could not be detected in
aqueous ketose-casein systems heated for up to 40 min at 120
°C, the occurrence of this type of Maillard products in food
was questioned generally (3).
MATERIALS AND METHODS
Chemicals and Reagents. NR-Hippuryllysine (BzGK, NR-benzoyl-
glycyl-L-lysine) was obtained from Bachem (Weil, Germany). HPLC
gradient grade methanol was from Sigma-Aldrich (Taufkirchen,
Germany). Carboxypeptidase B (EC 3.4.17.2) (DFP-treated, 133 units/
mg of protein, 5 mg of protein/mL), and all other chemicals used were
of analytical grade and purchased from Fluka (Taufkirchen, Germany).
The water used for the preparation of buffers and solutions was obtained
using a Purelab plus purification system (USFilter, Ransbach-Baumbach,
Germany).
High-Pressure Liquid Chromatography (HPLC). All analytical
HPLC analyses were performed with a gradient pump system from
Knauer (Berlin, Germany) with an online degasser, a K1500 solvent
organizer, a K1001 pump, a K2501 Knauer variable-wavelength
detector, a dynamic mixing chamber, and a column oven.
HPLC Analysis of Synthesis Mixtures and Samples of the Low-
Moisture Model System. The analytical gradient pump system de-
scribed above was used. Analytical separation was achieved using a
250 × 4.6 mm i.d., column filled with Eurosphere 100 RP 18, 5 µm,
with an integrated guard column, 5 × 4 mm, filled with the same
material (Knauer). The injection volume was 50 µL, the column
temperature was set to 20 °C, and ultraviolet detection was performed
at 230 nm. The mobile phase consisted of 0.01 M sodium phosphate
buffer, pH 7.0 (solvent A), and methanol (solvent B). A linear gradient
from 5 - 20 % B in 45 min at a flow rate of 0.5 mL/min was used.
HPLC Analysis of the Samples of Baked Biscuits. The analytical
gradient pump system described above was used. Analytical separation
was achieved using a stainless steel column, 250 × 3 mm, filled with
However, as fructose is often used as sweetening agent, in
particular, in food for diabetics, we decided to study the reaction
* Corresponding author (telephone ++49-351-463-34647; fax ++49-
351-463-34138; e-mail Thomas.Henle@chemie.tu-dresden.de).
10.1021/jf073256y CCC: $40.75  2008 American Chemical Society
Published on Web 03/05/2008

Peptide-Bound Heyns Products
J. Agric. Food Chem., Vol. 56, No. 7, 2008 2523
Knauer Eurospher 100, RP18-material of 5 µm particle size, with an
integrated guard column, 5 × 3 mm, filled with the same material
(Knauer). The injection volume was 50 µL, the column temperature
was set to 20 °C, and ultraviolet detection was performed at 230 nm.
The mobile phase consisted of 0.1% (v/v) formic acid (solvent A) and
0.1% (v/v) formic acid in methanol (solvent B). A linear gradient from
10 to 25% B in 45 min at a flow rate of 0.2 mL/min was used.
Semipreparative HPLC. Semipreparative HPLC was performed
with a gradient pump system from Knauer consisting of two K1001
pumps with 50 mL pump heads, an online degasser, a K1500 solvent
organizer, a dynamic mixing chamber, a column oven, and a K2501
Knauer variable-wavelength detector. All semipreparative separations
were performed using a stainless steel column, 250 × 16 mm, with a
guard column, 30 × 16 mm, both filled with Knauer Eurospher 100
RP18-material of 15-25 µm particle size (Knauer). The flow rate was
6 mL/min, the temperature was set to 20 °C, and ultraviolet detection
was at 280 nm. The first chromatographic stage was isocratic with a
mixture of 0.01 M sodium phosphate buffer, pH 7.0, and methanol
(94:6, v/v). The second stage (desalting) was realized using 0.05 M
acetic acid (solvent A) and methanol (solvent B) as well as a linear
gradient from 5 to 15% B in 45 min.
Mass Spectrometry. Mass spectrometric analysis was performed
with a PerSeptive Biosystems Mariner time-of-flight mass spectrometry
(TOF-MS) instrument equipped with an electrospray ionization source
(ESI) working in the positive mode (Applied Biosystems, Stafford, TX).
Calibration of the mass scale was established using a mixture of
bradykinin, angiotensin I, and neurotensin.
After appropriate dilution with 1% acetic acid in 50% methanol,
the sample was injected at a flow rate of 5 µL/min into the ESI source,
using a syringe pump for direct ESI-TOF-MS analysis. In LC-ESI-MS
experiments the outlet of the HPLC system was coupled to the ESI
interface.
°C under reduced pressure and the solid was dissolved in 9 mL of
0.01 M sodium phosphate buffer, pH 7.0. Isolation of BzGFruK was
achieved by semipreparative HPLC as described above. After desalting,
the BzGFruK-containing fractions were lyophilized and stored at –20
°C. Purity and identity were checked using analytical RP-HPLC, 1H
NMR, 13C NMR, and mass spectrometry as well as elemental analysis.
Analytical data were consistent with those reported previously (10).
Synthesis and Isolation of NR-Hippuryl-Nꢀ-(2-deoxy-D-glucosyl)lysine
(BzGGlcK) and NR-Hippuryl-Nꢀ-(2-deoxy-D-mannosyl)lysine (BzG-
ManK). A mixture of the Heyns compounds BzGGlcK and BzGManK
was synthesized according to the method of Heyns et al. (4, 11), using
NR-hippuryl-lysine as starting material. The used HPLC procedure for
purification of BzGFruK (10) was adapted for isolation of the Heyns
compounds. NR-Hippuryl-lysine (0.62 g, 2.0 mmol) and 3.61 g (20.0
mmol) of anhydrous fructose were heated at 75 °C in 25 mL of dry
dimethyl sulfoxide for 4 h. The reaction mixture was evaporated to
dryness at 35 °C in vacuo (0.25 mbar), and the residue was dissolved
in 9.0 mL of 0.2 M N-ethylmorpholine/acetic acid buffer, pH 8.0. The
pH value was adjusted to 8.0 with N-ethylmorpholine. Twenty-seven
microliters of a solution of carboxypeptidase B (665 units/mL) was
added, to a final activity of 2 units/mL. The solution was incubated in
a screw-cap culture tube at 25 °C. Hydrolysis of unreacted NR-hippuryl-
lysine was followed by analytical HPLC and was complete after 24 h.
After that, the solvent was evaporated at 40 °C under reduced pressure
and the solid was dissolved in 9 mL of 0.01 M sodium phosphate buffer,
pH 7.0, and subjected to semipreparative HPLC as described above.
The desalted mixed fractions containing BzGGlcK and BzGManK were
lyophilized and stored at –20 °C.
BzGGlcK: ESI-MS, positive mode, [M + H]+ m/z 470.2; 1H NMR
(500 MHz, D2O, R-pyranose), δ 1.29 (2H, m, K-H4), 1.61 (1H, m,
K-H3A), 1.61 (2H, m, K-H5), 1.75 (1H, m, K-H3B), 3.02 (2H, m,
K-H6), 3.13 (1H, dd, H2′), 3.34 (1H, dd, H4′), 3.65 (1H, dd, H6A′),
3.72 (1H, dd, H5′), 3.72 (1H, d, H6B′), 3.80 (1H, dd, H3′), 3.97 (1H,
d, G-H2A), 4.03 (1H, d, G-H2B), 4.15 (1H, dd, K-H2), 5.41 (1H, d,
H1′), 7.42 (1H, t, Bz-Hm), 7.52 (2H, t, Bz-Hp), 7.70 (2H, d, Bz-Ho);
(ꢀ-pyranose), δ 4.91 (1H, d, H1′), 2.90 (1H, dd, H2′): 13C NMR (125
MHz, D2O, R-pyranose), δ 21.93 (t, K-C4), 24.81 (t, K-C5), 30.77
(t, K-C3), 43.00 (t, G-C2), 45.36 (t, K-C6), 54.28 (d, K-C2), 59.75
(d, C2′), 60.17 (t, C6′), 69.40 (d, C3′), 69.52 (d, C4′), 71.14 (d, C5′),
87.66 (d, C1′), 127.11 (d, Bz-Co), 128.74 (d, Bz-Cm), 132.40 (d,
Bz-Cp), 132.68 (s, Bz-Ci), 170.80 (s, G-C1), 171.08 (s, Bz-C1),
178.27 (s, K-C1); (ꢀ-pyranose), δ 91.9 (d, C1′). Elemental analysis:
C21H31N3O9 × 0.6 CH3COOH × 1.6 H2O (MW ) 534.34); calcd, C
49.90%, H 6.90%, N 7.86%; found, C 49.94%, H 6.92%, N 7.81%.
Yield ) 222 mg (molar yield ) 20.8%).
The monoisotopic molecular masses were determined using the peak
with the lowest m/z ratio (monoisotopic peak) from prominent multiple-
charged ions and the equation Mr ) z × Mz - 1.0078z, where Mr is
the monoisotopic molecular mass, Mz is the m/z ratio, z is the number
of charges, and 1.0078 is the mass of a proton.
LC-ESI-MS Analysis of the Samples of the Low-Moisture Model
System and Baked Biscuits. LC-ESI-MS analysis was performed on
an Agilent 1100 series HPLC system (Agilent Technologies, Palo Alto,
CA) consisting of a high-pressure gradient pump system, column oven,
and diode array detector, which was coupled to a Mariner ESI-TOF-
MS instrument (see above). A sample volume of 100 µL was applied.
Separation was achieved using the same column, temperature, gradient,
and solvents as described for HPLC analysis of the baked biscuits.
Nuclear Magnetic Resonance Spectroscopy. NMR spectra were
recorded on a Bruker DRX 500 instrument (Reinstetten, Germany) with
Studies Related to the Formation of Heyns Compounds in a Low-
Moisture Model System. Phosphate buffers (0.1 M), pH 6.0 or 7.4,
and 1.78 g of disodium hydrogen phosphate were dissolved in 90 mL
of water, the pH value was adjusted using hydrochloric acid, and the
volume was adjusted to 100 mL.
1
500 MHz for H and 125 MHz for 13C experiments. Proton chemical
shifts are given relative to internal HOD signal (4.70 ppm) for deuterium
oxide solutions. Carbon chemical shifts are given relative to the signal
of the external standard tetramethylsilane for deuterium oxide solutions.
Assignments of 1H and 13C signals were based on 1H-1H COSY
(correlation spectroscopy), HSQC (heteronuclear single quantum coher-
ence), HMBC (heteronuclear multiple bond correlation), and DEPT
(distortionless enhancement by polarization transfer) experiments.
Elemental Analysis. Elemental analysis data were obtained on a
Euro EA 3000 elemental analyzer (Eurovector, Milano, Italy).
Synthesis and Isolation of NR-Hippuryl-Nꢀ-(1-deoxy-D-fructosyl)lysine
(BzGFruK). The Amadori compound NR-hippuryl-Nꢀ-(1-deoxy-D-
fructosyl)lysine (BzGFruK) was prepared as described previously, using
NR-hippuryl-lysine as starting material (10). NR-Hippuryl-lysine (0.62
g, 2.0 mmol) and 2.16 g (12.0 mmol) of anhydrous glucose were
refluxed in 84 mL of methanol for 4 h. The reaction mixture was
evaporated to dryness at 20 °C under reduced pressure, and the residue
was dissolved in 9.0 mL of 0.2 M N-ethylmorpholine/acetic acid buffer,
pH 8.0. The pH value was adjusted to 8.0 with N-ethylmorpholine.
Twenty-seven microliters of a solution of carboxypeptidase B (665
units/mL) was added, to a final activity of 2 units/mL. The solution
was incubated in a screw-cap culture tube at 25 °C. Hydrolysis of
unreacted NR-hippuryl-lysine was followed by analytical HPLC and
was complete after 24 h. After that, the solvent was evaporated at 40
The model mixtures were prepared with a lysine-to-sugar ratio of
1:6 as follows: 30.7 mg of BzGK and 108.0 mg of fructose (or glucose)
were dissolved in 20 mL of 0.1 M phosphate buffer, pH 6.0 or 7.4.
After the addition of 3.861 g of microcristalline cellulose (Avicel), the
mixture were intensively mixed using an Ultra Turrax (Janke & Kunkel
GmbH, Staufen, Germany). The samples were then deep frozen,
lyophilized, and homogenized. The water activity of the powdered
model mixtures was adjusted to an aW value of 0.52 or 0.86 by storing
them in Petri dishes placed in desiccators over saturated solutions of
magnesium nitrate hexahydrate (aW ) 0.52) or potassium chloride (aW
) 0.86) at ambient temperature (22 °C) for 6 days (12).
Aliquots of 100 mg of the equilibrated model mixtures were heated
in sealed screw-cap vials for up to 4 h at 80 or 100 °C in a sand bath
in a drying oven. After heating, the samples were extracted with 1.0
mL of 0.2 M N-ethylmorpholine/acetic acid buffer, pH 8.0, and
centrifuged. Seven hundred microliters of the supernatant was mixed
with 2.1 µL of a solution of carboxypeptidase B (665 units/mL), to a
final activity of 2 units/mL. After incubation at 25 °C for 24 h, an
aliquot of 600 µL was diluted with 600 µL of solvent A (10 mM sodium
phosphate buffer, pH 7.0) and subjected to HPLC analysis.

2524 J. Agric. Food Chem., Vol. 56, No. 7, 2008
Krause et al.
The same procedure of analysis was applied for an additional aliquot
of 100 mg of the equilibrated model mixture but without heating and
treatment with carboxypeptidase B to determine the initial content of
BzGK. Each experiment, beginning with or without heating, was made
in triplicate.
The identities of the two product peaks in the chromatogram of the
heated BzGK-fructose mixtures were elucidated using LC-ESI-MS.
Therefore, the fractions of the two peaks were collected manually using
HPLC with the phosphate-containing eluent system (six runs), con-
centrated, and subsequently analyzed using LC-ESI-MS with the formic
acid containing eluent system (see above).
Studies Related to the Formation of Heyns Compounds in
Biscuits. A biscuit dough was prepared from 10.00 g of wheat flour,
5.00 g of butter, 5.00 g of fructose (or another sugar component:
glucose, sucrose, or sorbitol), 0.30 g of commercial baking powder
(sodium bicarbonate, sodium diphosphate), 25 mg of sodium chloride,
and 60 mg of BzGK (dissolved in 1.0 mL of water). The dough was
rolled out to a thickness of about 5 mm, and biscuits were cut out with
a diameter of about 5 cm. The biscuits were baked in a laboratory drying
oven at 175 °C for 7 min. After cooling, the biscuits were ground using
a mill. A sample of 6.5 g was defatted using 20 mL of petroleum ether
(boiling point range ) 30-50 °C) in a screw-cap centrifuge tube. The
sample were centrifuged, and the petroleum ether was discarded. After
this procedure had been repeated, residual petroleum ether was removed
with a nitrogen stream. The residue was transferred in a 100 mL
volumetric flask and extracted with 70 mL of 40% ethanol (v/v) under
stirring (magnetic stirrer) for 60 min and using ultrasonic treatment
for 5 min. After that, the volume was adjusted to 100 mL. The extract
was filtered and centrifuged (15000g, 10 min). Two milliliters of the
purified extract was evaporated to dryness at 40 °C under reduced
pressure, and the residue was dissolved in 2.00 mL of 0.01 M sodium
phosphate buffer, pH 7.0. An aliquot of 500 µL was mixed with 500
µL of 0.4 M N-ethylmorpholine/acetic acid buffer, pH 8.0 () solution
E). An aliquot of 500 µL of solution E was mixed with 1.5 µL of a
solution of carboxypeptidase B (665 units/mL), to a final activity of 2
units/mL () solution F). After incubation at 25 °C for 24 h, an aliquot
of 68 µL was diluted with 932 µL of solvent A (0.1% formic acid)
and subjected to HPLC analysis for determination of the content of
BzGGlcK and BzGManK. An aliquot of 68 µL of solution E was diluted
with 932 µL of solvent A (0.1% formic acid) and subjected to HPLC
analysis for determination of the total content of BzGK, BzGGlcK,
and BzGManK. Each experiment, starting with the preparation of the
dough, was made in triplicate.
Figure 1. Structures of BzGK (1), BzGGlcK (2), BzGManK (3), and
BzGFruKe (4).
formation of both compounds, it was concluded that the isolate
was sufficient as reference material. The Amadori compound
BzGFruK was prepared using BzGK and glucose as starting
material in chromatographically pure form. The structural
identity of the desired compounds was confirmed by mass
spectrometry and NMR spectroscopy. Elemental analysis and
NMR spectroscopy indicated residual acetic acid from the HPLC
eluent and some water after lyophilization. Because chroma-
tography and NMR analysis revealed no byproduct, it was
deduced that the purity of the isolates was adequate for the
purposes of this study.
To compare the formation of Heyns and Amadori products
under reduced water activity, first the reaction of BzGK with
fructose or glucose was investigated in a buffered low-moisture
model system with microcrystalline cellulose as support.
Mixtures consisting of BzGK and the sugar component (molar
ratio 1:6) in phosphate buffer (pH 6.0 or 7.4) as well as
microcrystalline cellulose were prepared and lyophilized. After
that, the water activity was adjusted to aW ) 0.52 or 0.86 at 20
°C. Water activities and pH values were chosen with respect to
bakery products. It is interesting that commercial biscuits have
a relatively high pH value of 7.4, if a 10% suspension is
measured. The reason may be the bicarbonate salt in the baking
powder. The molar ratio between BzGK and fructose, namely,
1:6, was chosen to keep the Maillard reaction in an “early stage”
and to avoid caramelization and intense browning reactions.
The model mixtures were heated and the degradation of
BzGK and formation of peptide bound rearrangement products
were monitored using RP-HPLC with UV detection. Chromato-
graphic separation of the modified lysine derivatives from the
starting peptide was optimized by treatment with carboxypep-
tidase B (EC 3.4.17.2) before chromatography. This enzyme is
used in sequence analysis of peptides and hydrolyses the basic
amino acids arginine and lysine from the C terminus of peptides
(13). Thus, unreacted BzGK was hydrolyzed to hippuric acid
(BzG) and lysine, whereas the modified lysine peptides remained
unchanged. On this basis, separation of the rearrangement
products was possible (Figure 2). The epimeric Heyns com-
pounds BzGGlcK and BzGManK were detected in the fructose-
containing samples and the Amadori compound BzGFruK was
detected in that with glucose. Identity of the products was
examined and confirmed by co-injection of sample and reference
material in the HPLC system as well as by LC-ESI-MS. Figure
3 shows the time dependency of the product concentrations in
the samples at pH 7.4 and aW ) 0.86. As can be seen from
Figure 3, the formation of Heyns or Amadori compounds,
respectively, is significant in both model systems. Lysine
An aliquot of 200 µL of carboxypeptidase B-treated solution F was
diluted with 800 µL of solvent A (0.1% formic acid) and subjected to
LC-ESI-MS analysis.
RESULTS AND DISCUSSION
The reaction of the ꢀ-amino group of lysine with fructose or
glucose and the formation of Heyns and Amadori compounds
was studied in two model systems to investigate a possible
formation of these lysine derivatives in bakery products. The
peptide derivative BzGK was used as model for protein-bound
lysine because the NR-hippuryl moiety allowed a convenient
direct analysis of the samples by RP-HPLC and UV detection.
The putative derivatives resulting from the reaction with fructose
are the epimeric Heyns compounds BzGGlcK and BzGManK,
whereas the reaction with glucose leads to the Amadori
compound BzGFruK (Figure 1). These compounds were
synthesized as reference materials for chromatographic analysis.
Isolation from the synthesis mixtures was achieved using a two-
step semipreparative RP-HPLC method after hydrolysis of
unreacted BzGK with carboxypeptidase B (EC 3.4.17.2) as
described previously (10). Because the reaction of BzGK and
fructose leads to the formation of both Heyns compounds, which
could not be separated by HPLC, a mixture of the epimeric
Heyns compounds was obtained. As heating experiments with
the fructose-containing model systems resulted also in the

Peptide-Bound Heyns Products
J. Agric. Food Chem., Vol. 56, No. 7, 2008 2525
than at aW ) 0.86 for the same product. This might be explained
by a higher stability and a certain accumulation of the rear-
rangement products at aW ) 0.52, as is known for Amadori
compounds (14). Furthermore, the samples at pH 7.4 show after
heating at 80 °C a higher content than the samples at pH 6.0.
Consequently, heating of the samples at 80 °C for 60 min at
pH 7.4 versus pH 6.0 favors more the formation of the
rearrangement products. The results have reversed after heating
at 100 °C for 60 min, samples with pH 6.0 show a higher content
than the samples with pH 7.4 under these conditions. At the
increased temperature of 100 °C it may by of importance that
the higher pH value accelerates not only the formation of the
rearrangement products but also its degradation.
In conclusion, the studies clearly revealed the formation of
Heyns or Amadori compounds for the low-moisture model
systems used. According these results, the formation of Heyns
compounds could be expected in fructose-containing bakery
products. To examine this hypothesis, the formation of the
rearrangement products was investigated under conditions more
closely resembling baked food. Biscuits were baked for this
purpose. A usual recipe for rolled biscuits was used with the
ingredients of wheat flour, butter, fructose or an other sugar
component, baking powder, and salt. Additionally, BzGK
dissolved in a small amount of water was added to the dough
as a model for protein-bound lysine. Therewith, the detection
of rearrangement products was possible using RP-HPLC with
UV detection as in the model systems used before. The molar
ratio of added BzGK to lysine from the ingredients was 1:1.
The temperature of the baking oven was set to 175 °C, and a
baking time of 7 min was chosen to obtain sensorially pleasant,
slightly browned biscuits.
Figure 2. HPLC analysis of heated mixtures of BzGK (peak 1) with
fructose (molar ratio 1:6, pH 7.4, aw ) 0.52, 60 min, 100 °C) (A) before
and (B) after hydrolysis with carboxypeptidase B. Peaks: 2, BzGGlcK; 3,
BzGManK; 5, hippuric acid.
Figure 5 shows the HPLC analysis of the biscuits containing
different sugar components using 0.1% formic acid and 0.1%
formic acid in methanol as eluents. In comparison to the
phosphate buffer containing eluent system used for analysis of
cellulose-containing model mixtures, this eluent system im-
proved the peak form of the BzGFruK peak in the biscuit
samples containing glucose. HPLC analysis revealed the forma-
tion of the epimeric Heyns compounds BzGGlcK and BzG-
ManK in the fructose-containing biscuits and of the Amadori
compound BzGFruK in the glucose-containing biscuits. Com-
pared with this, no rearrangement product was detected in the
biscuits with sorbitol or sucrose as ingredient (Figure 5). This
was expected, because the sugar alcohol sorbitol does not
contain a reactive aldehyde or keto group. Sucrose is a
nonreductive disaccharide and apparently was not hydrolyzed
under the baking conditions used. As no Heyns compounds were
detectable in the fructose-containing biscuit samples before
heating (data not shown), it was clearly proven that its formation
is exclusively due to heat-induced reactions between peptide-
bound lysine and fructose.
Figure 3. Yield of Heyns compounds in heated mixtures of BzGK and
fructose (gray bars) and of Amadori compound in heated mixtures of BzGK
and glucose (black bars) (molar ratio of BzGK to sugar 1:6, pH 7.4, aW
) 0.86, 80 °C). Mean value ( standard deviation of three separate
determinations is shown.
derivatization was calculated by comparing peak areas of BzGK
before and after heating. For example, with respect to the initial
lysine content, the yields were measured with 8% for the Heyns
compounds and with 17% for the Amadori compound (pH 7.4,
aW ) 0.86, 100 °C, 60 min). Lysine derivatization was
calculated by comparing peak areas of BzGK before and after
heating. The total derivatization of lysine, however, was
considerably higher: 48 and 70% were measured in the fructose-
containing and in the glucose-containing samples, respectively.
According to these results, besides the formation of the
rearrangement products, degradation of the Maillard products
takes place at the same time and. consequently, the measured
contents are the result of these two processes.
Figure 4 shows the yields of rearrangement products at
different aW and pH values after heating at 80 or 100 °C for 60
min. When compared with the Amadori compound, lower
amounts of Heyns compounds are formed at the same reaction
conditions. This may be attributed to the specific properties of
Heyns compounds, their lower rate of formation, the reversibility
of the formation reaction, and significantly faster conversion to
other products. To interpret the different yields of the rear-
rangement products at the chosen aW and pH values, both the
formation reactions and degradation reactions must be consid-
ered as mentioned above. At aW ) 0.52, the yields are higher
The detection of the rearrangement products in the biscuits
with fructose or glucose was confirmed by co-injection of
sample and reference material in the HPLC system as well as
by mass spectrometric measurement via LC-ESI-MS. Figure 6
shows the mass spectrometric verification of BzGGlcK detection
in a the sample of the fructose-containing biscuits. As can be
seen, the mass spectrum of the product peak of a sample of the
fructose-containing biscuits was dominated by a [M + H]+ ion
with m/z 470.42, which led to a monoisotopic mass Mr of 469.41
(Figure 6). This result was consistent with the expected value
for BzGGlcK (Mr of 469.49). Accordingly, BzGFruK was
identified by LC-ESI-MS (Figure 7).

2526 J. Agric. Food Chem., Vol. 56, No. 7, 2008
Krause et al.
Figure 4. Yield of Heyns compounds in heated mixtures of BzGK and fructose (gray bars) and of Amadori compound in heated mixtures of BzGK and
glucose (black bars) (molar ratio of BzGK to sugar 1:6, heating time ) 60 min). Mean value ( standard deviation of three separate determinations is
shown.
Figure 7. Mass spectrometric analysis of the product peak (BzGFruK) in
the fructose-containing biscuits using LC-ESI-MS in positive mode.
conditions used for the biscuits, for example, oven temperature
(175 °C) and pH value (pH 7.4, if a 10% suspension is
measured), favor the formation of the rearrangement products.
Due to the short baking time, degradation reactions apparently
are not of substantial importance.
Figure 5. HPLC analysis of baked biscuits containing BzGK as well as
fructose (A), glucose (B), sorbitol (C), or sucrose (D) after hydrolysis with
carboxypeptidase B. Peaks: 2, BzGGlcK; 3, BzGManK; 4, BzGFruK.
Whereas the formation of Amadori compounds in the model
systems is well-documented for baked food, very little is known
about the occurrence of Heyns compounds in food. Because
the model studies unequivocally showed the formation of Heyns
compounds, it can be concluded that they are also formed in
substantial amounts in fructose-containing bakery products. In
particular, the formation of Heyns compounds may be of
importance in bakery products for diabetic patients, in which
saccharose as sweetener is substituted by fructose. Furthermore,
products containing honey, high-fructose corn syrup, or corre-
sponding sweeteners may contain Heyns compounds depending
on the heating conditions applied during food processing. With
that, a fructose-induced lysine derivatization could have nutri-
tional importance. It is well-known that Amadori compounds
of lysine are not available as lysine source during digestion (1).
Furthermore, a possible antinutritional role for Maillard com-
pounds is under debate (15). With that in mind, further studies
Figure 6. Mass spectrometric analysis of the product peak (BzGGlcK) in
the fructose-containing biscuits using LC-ESI-MS in positive mode.
must deal with analytical approaches for the identification and
quantification of protein-bound Heyns compounds in food to
clarify a possible impact of these lysine derivatives on the
nutritional quality of relevant food items.
Quantitative studies demonstrated high yields of the rear-
rangement products under the baking conditions used in the
baked biscuits containing fructose or glucose. The yields of
the Heyns compounds in the fructose-containing biscuits and
the yield of the Amadori compounds in the biscuits with glucose
were 33 and 63%, respectively. Consequently, the baking
ABBREVIATIONS USED
BzGFruK, N-R-hippuryl-Nꢀ-(1-deoxy-D-fructosyl)lysine; BzG-
GlcK, N-R-hippuryl-Nꢀ-(2-deoxy-D-glucosyl)lysine; BzGK, N-R-

Peptide-Bound Heyns Products
J. Agric. Food Chem., Vol. 56, No. 7, 2008 2527
hippuryllysine; BzGManK, N-R-hippuryl-Nꢀ-(2-deoxy-D-man-
nosyl)lysine; HPLC, high-pressure liquid chromatography;
NMR, nuclear magnetic resonance; RP, reversed phase.
(6) Vondenhof, T.; Glombitza, K. W.; Steiner, M. Lösliche Kohlen-
hydrate in Rohlakritzen (Succus Liquiritiae). Z. Lebensm. Unters.
Forsch. 1973, 152, 345–347.
(7) Van den Ouweland, G. A. M.; Peer, H. G.; Tjan, S. B. Occurrence
of Amadori and Heyns rearrangement products in processed foods
and their role in flavor formation. In FlaVor of Foods and
BeVerages; Charalambous, G., Inglett, G. E., Eds.; Academic
Press: New York, 1978; Vol. 1, pp 131–143.
(8) Wahl, R. Aminozucker in Tabak. Z. Lebensm. Unters. Forsch.
1972, 148, 94–96.
(9) McPherson, J. D.; Shilton, B. H.; Walton, D. J. Role of fructose
in glycation and cross-linking of proteins. Biochemistry 1988, 27,
1901–1907.
ACKNOWLEDGMENT
We thank Dr. Uwe Schwarzenbolz, Institute of Food Chem-
istry, for acquisition of the mass spectrometric data. We are
grateful to the members of the Institute of Organic Chemistry,
Technische Universität Dresden, namely, Dr. Margit Gruner and
Annett Rudolf, for recording the NMR spectra and Anke Peritz
for performing the elemental analysis.
(10) Krause, R.; Knoll, K.; Henle, T. Studies on the formation of
furosine and pyridosine during acid hydrolysis of different
Amadori products of lysine. Eur. Food Res. Technol. 2003, 216,
277–283.
(11) Heyns, K.; Noack, H. Die Umsetzung von D-Fructose mit L-Lysin
und Arginin und deren Beziehung zu nichtenzymatischen Bräu-
nungsreaktionen. Chem. Ber. 1962, 95, 720–727.
(12) Rockland, L. B. Saturated salt solutions for static control of relative
humidity between 5 and 40 °C. Anal. Chem. 1960, 32, 1375–
1376.
(13) Ambler, R. P. Enzymatic hydrolysis with carboxypeptidases.
Methods Enzymol. 1972, 25, 143–154.
(14) Ciner-Doruk, M.; Eichner, K. Bildung und Stabilität von Amadori-
Verbindungen in wasserarmen Lebensmitteln. Z. Lebensm. Unters.
Forsch. 1979, 168, 9–20.
LITERATURE CITED
(1) Henle, T. Protein-bound advanced glycation endproducts (AGEs)
as bioactive amino acid derivatives in foods. Amino Acids 2005,
29, 313–322.
(2) Matsuda, T.; Kato, Y.; Nakamura, R. Lysine loss and polymer-
ization of bovine ꢀ-lactoglobulin by amino carbonyl reaction with
lactulose (4-O-ꢀ-D-galactopyranosyl-D-fructose). J. Agric. Food
Chem. 1991, 39, 1201–1204.
(3) Brands, C. M. J.; Van Boekel, M. A. J. S. Reactions of
monosaccharides during heating of sugar-casein systems: building
of a reaction network model. J. Agric. Food Chem. 2001, 49,
4667–4675.
(4) Heyns, K.; Breuer, H.; Paulsen, H. Darstellung und Verhalten der
2-N-Aminosäure-2-desoxy-glucosen (“Glucose-Aminosäuren”) aus
Glycin, Alanin, Leucin und Fructose. Chem. Ber. 1957, 90, 1374–
1386.
(15) Henle, T. Dietary advanced glycation end productssa risk to
human health. Mol. Nutr. Food Res. 2007, 51, 1075–1078.
Received for review November 7, 2007. Revised manuscript received
January 24, 2008. Accepted January 29, 2008.
(5) Anet, E. F. J. L.; Reynolds, T. M. Reactions between amino acids,
organic acids, and sugars in freeze-dried apricots and peaches.
Aust. J. Chem. 1957, 10, 193–197.
JF073256Y