Isotope labeling studies on the formation of 5-(hydroxymethyl)-2- furaldehyde (HMF) from sucrose by pyrolysis-GC/MS

Article abstract of DOI:10.1021/jf8010245

Although it is generally assumed that the reactivity of sucrose, a nonreducing sugar, in the Maillard reaction is due to its hydrolysis into free glucose and fructose, however, no direct evidence has been provided for this pathway, especially in dry and high temperature systems. Using specifically Relabeled sucrose at C-1 of the fructose moiety, HMF formation was studied at different temperatures. Under dry pyrolytic conditions and at temperatures above 250°C, 90% of HMF originated from fructose moiety and only 10% originated from glucose. Alternatively, when sucrose was refluxed in acidic methanol at 65°C, 100% of HMF was generated from the glucose moiety. Moreover, the relative efficiency of the known HMF precursor 3-deoxyglucosone to generate HMF was compared to that of glucose, fructose and sucrose. Glucose exhibited a much lower conversion rate than 3-deoxyglucosone, however, both fructose and sucrose showed much higher conversion rates than 3-deoxyglucosone thus precluding it as a major precursor of HMF in fructose and sucrose solutions. Based on the data generated, a mechanism of HMF formation from sucrose is proposed. According to this proposal sucrose degrades into glucose and a very reactive fructofuranosyl cation. In dry systems this cation can be effectively converted directly into HMF.

Full text of DOI:10.1021/jf8010245

J. Agric. Food Chem. 2008, 56, 6717–6723 6717
Isotope Labeling Studies on the Formation of
5-(Hydroxymethyl)-2-furaldehyde (HMF) from Sucrose
by Pyrolysis-GC/MS
CAROLINA PEREZ LOCAS AND VAROUJAN A. YAYLAYAN*
Department of Food Science and Agricultural Chemistry, McGill University,
21111 Lakeshore, Ste. Anne de Bellevue, Quebec, Canada H9X 3V9
Although it is generally assumed that the reactivity of sucrose, a nonreducing sugar, in the Maillard
reaction is due to its hydrolysis into free glucose and fructose, however, no direct evidence has been
provided for this pathway, especially in dry and high temperature systems. Using specifically 13C-
labeled sucrose at C-1 of the fructose moiety, HMF formation was studied at different temperatures.
Under dry pyrolytic conditions and at temperatures above 250 °C, 90% of HMF originated from fructose
moiety and only 10% originated from glucose. Alternatively, when sucrose was refluxed in acidic
methanol at 65 °C, 100% of HMF was generated from the glucose moiety. Moreover, the relative
efficiency of the known HMF precursor 3-deoxyglucosone to generate HMF was compared to that of
glucose, fructose and sucrose. Glucose exhibited a much lower conversion rate than 3-deoxyglu-
cosone, however, both fructose and sucrose showed much higher conversion rates than 3-deoxy-
glucosone thus precluding it as a major precursor of HMF in fructose and sucrose solutions. Based
on the data generated, a mechanism of HMF formation from sucrose is proposed. According to this
proposal sucrose degrades into glucose and a very reactive fructofuranosyl cation. In dry systems
this cation can be effectively converted directly into HMF.
KEYWORDS: Isotope labeling studies; HMF formation mechanism from glucose; fructose and sucrose;
fructofuranosyl cation; levoglucosan formation; 3-deoxyglucosone
INTRODUCTION
Furthermore, rates of HMF formation from glucose and sucrose
showed slight enhancement in the presence of the amino acids,
whereas virtually no enhancement occurred when fructose was
the substrate. Without acid catalysis and at 250 °C the
conversion rate of glucose into HMF was 24% and for fructose
the rate was 36% (7). However, increasing the acid concentration
significantly improved the rate of HMF formation from fructose
relative to glucose. At 1 mM H2SO4, 42% of fructose was
converted into HMF versus 31% for glucose. Interestingly, when
sucrose was heated under identical conditions, the yield of HMF
per mole of fructose increased from 36% to 47% for the
uncatalyzed reaction and from 42% to 53% for the acid
catalyzed reaction (7). The enhanced HMF formation from
sucrose per mole of fructose moiety at high temperatures can
be justified by the fact that the glycosidic bond of sucrose can
be easily cleaved under mild acidic conditions to produce
fructofuranosyl cation, the direct precursor of HMF (see Figure
1), however, it is much more difficult for the free fructose to
generate the same cation (9) under identical conditions. Numer-
ous studies have also indicated the formation of fructofuranosyl
cation from fructose as the first step in the formation of
HMF (9–11). On the other hand, glucose cannot be converted
into HMF through dehydration from cyclic forms for obvious
reasons and is therefore recognized to generate HMF through
cyclization of 3-deoxyglucosone (3-DG) intermediate formed
Similar to the widespread occurrence of acrylamide in
thermally processed food, 5-(hydroxymethyl)-2-furaldehyde
(HMF) is also detected in variety of food products but in
relatively higher concentrations (exceeding 1 g/kg). HMF is one
of the major degradation products of carbohydrates that has been
studied extensively as an indicator of heat damage (1, 2). HMF
has been used successfully as a chemical index in ensuring
adequate heat processing or for monitoring storage conditions
for fruit juices, milk, honey, cereal products, cookies and
jams (3–5). Formation of HMF from carbohydrates has been
found to depend on many factors such as time, water activity,
temperature, amount and type of catalyst and sugar used (6).
Ketoses generate more HMF than aldoses and the yield increases
with increase in the temperature and the concentration of the
acid catalyst although, it can also be formed in slightly lower
yields in the absence of a catalyst (7). Numerous studies have
indicated that fructose is the most reactive sugar relative to
sucrose and glucose, in the formation of HMF under acidic
conditions. According to Lee and Nagy (8) at 50 °C and pH of
3.5, fructose was 31.2 times faster than glucose, whereas sucrose
was 18.5 times faster than glucose in the rate of HMF formation.
* Corresponding author [telephone (514) 398-7918; fax (514) 398-
7977; e-mail varoujan.yaylayan@mcgill.ca].
10.1021/jf8010245 CCC: $40.75  2008 American Chemical Society
Published on Web 07/09/2008

6718 J. Agric. Food Chem., Vol. 56, No. 15, 2008
Perez Locas and Yaylayan
Figure 1. Proposed mechanism of HMF formation from glucose, fructose, 3-deoxyglucosone, and sucrose.
temperature was 180 °C. The ionization voltage was 70 eV, and the
electron multiplier was 2471 V. All injections were in splitless mode.
The mass range analyzed was 33-650 amu. The initial temperature of
the column was set at 37 °C for 2 min and was increased to 100 °C at
a rate of 30 °C/min, immediately the temperature was further increased
to 250 °C at a rate of 8 °C/min and kept at 250 °C for 5 min. The
identity and purity of the chromatographic peaks were determined using
NIST AMDIS version 2.1 (http://chemdata.nist.gov/mass-spc/amdis/).
The reported percent label incorporation values (corrected for natural
abundance and for % enrichment) are the average of duplicate analyses
and are rounded off to the nearest multiple of 5%.
from the open-ring form of glucose (see Figure 1). Due to the
low propensity of glucose to exists in open ring form and due
to many other side reactions of 3-DG, the rate of its conversion
into HMF is low compared to fructose especially at higher
temperatures where fructose can directly dehydrate from its
cyclic forms through the intermediacy of fructofuranosyl cation
without the need to undergo thermodynamically controlled ring
opening process.
Recent findings on the acute toxicity of HMF (12–14) and
lack of evidence from isotope labeling studies confirming the
above proposed mechanism of conversion of sucrose into HMF,
prompted us to investigate the mechanism of HMF formation
utilizing 13C-labeled precursors.
(1-13Cfru)Sucrose Reaction in Methanol. The (1-13Cfru)sucrose
(4 mg) was refluxed in methanol (200 µL) for 10 min in the presence
of p-toluenesulfonic acid monohydrate (3 mg). A 20 µL portion was
injected into the GC/MS and analyzed using the same method as
described above.
MATERIALS AND METHODS
Detection of Levoglucosan by Ion Chromatography. A Metrohm
MIC-8 modular IC system (Herisau, Switzerland) consisting of a pulsed
amperometric detector (E1 ) 0.15 v, t1 ) 400 ms, E2 ) 0.75v, t2 )
200 ms; E3 ) -0.15v, t3 ) 400 ms), pump and a sample injection unit
connected to Metrosep Carb1-150 anion exchange column thermo-
statted at 31 °C was used for the analysis of carbohydrate residue after
pyrolysis. The mobile phase was 0.1N NaOH and flow rate of 1 mL/
min. The retention time of the commercial levoglucosan was 2.62 min.
The residue after pyrolysis of sucrose was dissolved in distilled water
and diluted before injection. One of the major peaks had a retention
time of 2.62 min identical to the levoglucosan standard.
All reagents, chemicals were purchased from Aldrich Chemical
company (Milwaukee, WI) and used without further purification.
3-Deoxyglucosone was purchased from Toronto Research Chemicals
(Ontario, Canada). The [13C]glucoses were purchased from Cambridge
Isotope Laboratories (Andover, MA). [1-13C]fructose and [1-13Cfru]su-
crose were purchased from Omicron Biochemicals Inc. (IN).
Pyrolysis-GC/MS Analysis of HMF. A Helwett-Packard (Palo Alto,
California) GC with Mass selective detector (5890 GC/5971B MSD)
interfaced to a CDS pyroprobe 2000 unit (CDS Analytical, Oxford,
PA) was used for the Py-GC/MS analysis. One mg samples of reactants
were mixed either with silica gel or introduced into the quartz tube
(0.3 mm thickness) as is, plugged with quartz wool, and inserted inside
the coil probe and pyrolyzed at indicated temperatures with a total
heating time of 20s. The column was a fused silica DB-5 column (50
m length × 0.2 mm i.d. × 0.33 µm film thickness; J&W Scientific,
Folsom, CA). The pyroprobe interface temperature was set at 250 °C.
Capillary direct MS interface temperature was 280 °C; ion source
RESULTS AND DISCUSSION
Sucrose Degradation and Origin of HMF. Although there
are some reports in the literature (6) to indicate that fructose is
the main moiety in sucrose that contributes to the generation

Formation of HMF from Sucrose
J. Agric. Food Chem., Vol. 56, No. 15, 2008 6719
Figure 2. Mass spectrum of HMF and the structure of the major ions.
of HMF, however, there is no quantitative data. According to
Antal et al. (7) at 250 °C and under 34.5 MPa, 30% of the
uncatalyzed sucrose solution (0.05M) can be converted into
HMF in 32 s. This conversion rate increases to 50% under acid
catalysis. Using the data provided by the Antal et al. (7)
regarding the amount of free glucose and fructose remaining
after sucrose hydrolysis and the rates of HMF formation from
glucose and fructose under identical conditions in addition to
the amount of other products formed, we were able to estimate
the amount of HMF formed from fructose moiety of sucrose to
be ∼84% during both uncatalyzed and acid catalyzed reactions.
In order to verify this number, the ability of sucrose labeled
only at C-1 of fructose moiety to generate HMF was analyzed
by Py-GC/MS at various temperatures. If HMF can be produced
only from glucose, no incorporation of 13C-label will be
observed in the parent ion of HMF at m/z 126, whereas if HMF
was generated only from fructose moiety 100% 13C-label
incorporation should be observed in HMF and finally if both
sugar moieties are responsible for HMF formation less than
100% incorporation will be observed. Moreover, percent label
incorporation will indicate percent contribution of fructose to
total HMF production from sucrose.
In order to extract mechanistic information from such labeling
studies, knowledge of the elemental composition and the
structures of the important mass spectral fragments are essential.
Consequently, to gain insight into mass spectral fragmentation
patterns of HMF singly labeled glucoses were pyrolyzed to
generate HMF and label incorporation was analyzed (see
Figures 2 and 3 and Tables 1 and 2). These figures show the
structures of the relevant fragment ions identified based on the
label incorporation pattern listed in Tables 1 and 2. It is
important to note that similar to any aldehyde, HMF exhibits
an M-1 peak due to the loss of aldehydic hydrogen. During
labeling studies the percent of M-1 peak will be used to indicate
unlabeled HMF arising from the glucose moiety, therefore it is
important to confirm this number accurately. Table 1 indicates
that based on six labeled glucoses and one labeled fructose this
value is 15-16% of the intensity of the parent ion at m/z 126.
When (1-13Cfru)sucrose was pyrolyzed to generate labeled HMF
the corresponding M-1 peak for the labeled HMF was 26%
indicating that 10% was due to the unlabeled HMF arising from
glucose and 16% due to loss of aldehydic proton. Therefore,
90% of HMF generated from sucrose arises from fructose
moiety during pyrolysis at high temperatures (see Table 1).
These results are consistent with the above estimation of 84%
based on literature data generated at 250 °C.
H-Rearrangement and Scrambling of Labels in the Mo-
lecular Ion of HMF Generated under Electron Impact (EI)
Conditions. During studies on the fragmentation patterns of
HMF, inspection of ion at m/z 97 (see Figures 2 and 3) has
indicated an unexpected label incorporation pattern as shown
in Table 2. The ion at m/z 97 arises from ion at M - 1 by the
loss of aldehydic CO as shown in Figure 3. According to Figure
3, when the precursor of HMF is either glucose-1-13C or
fructose-1-13C, the ion at m/z 97 is expected to completely lose
the label as 13CO, however, both labeled sugars retained 20%
of the label as shown in Table 2. In addition, the remaining
80% was incorporated into the C-6 as indicated from the data
on pyrolysis of glucose-6-13C (see Table 2). These observations
can be explained by the formation of two molecular ions as
shown in Figure 3, one by the loss of electron from carbonyl
oxygen (60%) and the other by the loss of electron from
hydroxyl oxygen (40%). The latter can initiate a series of two
hydrogen rearrangement reactions to generate two isotopomers
of m/z 126 in equimolar amounts (20% each) due to the
symmetrical nature of the intermediate formed after the first
rearrangement. Consequently, 80% of the HMF in the mass
detector will incorporate C-1 as the aldehydic carbon and 20%
will incorporate C-6 as the aldehydic carbon.
Proposed Mechanism of Thermal Generation of HMF
from Sucrose. In order to confirm the literature data (7)
generated at 250 °C and under a pressure of 34.5 MPa regarding
the relative conversion efficiency of sucrose into HMF; fructose,
glucose and sucrose were pyrolyzed at 250, 300, and 350 °C

6720 J. Agric. Food Chem., Vol. 56, No. 15, 2008
Perez Locas and Yaylayan
Figure 3. Mass spectral fragmentation pattern of HMF and formation of two molecular ions at m/z 126. rH ) hydrogen rearrangement.
Table 1. Percent Labela Incorporation in M + 1, M, and M - 1 Ionsb of
Table 2. Percent labela incorporation in fragment at m/z 97b of HMF
Generated from Different Precursorsc
HMF Generated from Various Precursorsc
compound
M + 1, m/z 127
M, m/z 126
M - 1, m/z 125
compound
m/z 97
m/z 98
3-deoxyglucosone
D-glucose
0
0
100
100
16 (% of M)
15 (% of M)
3-deoxyglucosone
D-glucose
100
100
20
0
0
0
0
0
80
100
100
100
100
20
D-glucose-6-13C
D-glucose-5-13C
D-glucose-4-13C
D-glucose-3-13C
D-glucose-2-13C
D-glucose-1-13C
D-fructose-1-13C
(1-13Cfru)sucrosed
(1-13Cfru)sucrosee
(1-13Cfru)sucrosef
100
100
100
100
100
100
100
90
16 (% of M + 1)
16 (% of M + 1)
15 (% of M + 1)
16 (% of M + 1)
15 (% of M + 1)
16 (% of M + 1)
16 (% of M + 1)
26 (% of M + 1)
26 (% of M + 1)
16 (% of M + 1)
D-glucose-6-13C
D-glucose-5-13C
D-glucose-4-13C
D-glucose-3-13C
D-glucose-2-13C
D-glucose-1-13C
D-fructose-1-13C
(1-13Cfru)sucrosed
(1-13Cfru)sucrosee
0
80
80
81
81
20
19
19
90
0
a All singly labeled and corrected for 13C natural abundance. b See Figures 3
and 4. c Values represent average of two replicates with standard deviation not
more than 5%. d Pyrolyzed at 250 °C. e Pyrolyzed at 350 °C.
a All singly labeled and corrected for 13C natural abundance. b See Figures 3
and 4. c Values represent average of two replicates with standard deviation of not
more than 5%. d Pyrolyzed at 250 °C and corrected for loss of aldehydic hydrogen.
e Pyrolyzed at 350 °C and corrected for loss of aldehydic hydrogen. f (1-
13Cfru)sucrose (4 mg) was refluxed in methanol for 10 min in the presence of
p-toluenesulfonic acid monohydrate (3 mg).
glycosidically linked terminal fructose (as in sucrose) to generate
more HMF relative to free fructose, other oligosaccharides such
as raffinose and stacchiose having similar fructose linkages were
also analyzed and the results are shown in Table 5. According
to this table, both raffinose and stacchiose exhibited higher
efficiency of HMF formation compared to lactose a disaccharide
lacking a terminal fructose moiety as in sucrose.
Based on the above observations it can be proposed that the
major pathway of sucrose decomposition is the direct formation
of fructofuranosyl cation in addition to glucose and 1,6-anhydro-
glucose (levoglucosan), a known degradation product of glucose
and cellulose (see Figure 4). To confirm the formation of
levoglucosan from different sugars, the sugars were pyrolyzed
at 250, 300, and 350 °C and the data are reported in Table 6.
According to this table, glucose is the most efficient precursor
of levoglucosan followed by sucrose. Fructose however, did not
generate any levoglucosan. Pyrolysis of levoglucosan itself
indicated that it is volatile enough to be detected at high
temperatures (see Table 6) and that it did not produce any HMF
(see Table 3). Furthermore, the formation of levoglucosan from
sucrose was also confirmed by ion chromatography, when the
and the areas of HMF peaks produced are reported as area per
mmol of the starting sugar (see Table 3). The data indicated
that at all temperatures studied sucrose indeed generated more
HMF per mol relative to both fructose and glucose. For
mechanistic considerations the efficiency of HMF formation
from these sugars at 300 °C relative to 3-deoxyglucosone (3-
DG) was also studied (see Table 4). According to this table
and relative to 3-DG, sucrose generated 4.5 fold more HMF
and fructose generated 2.4 fold more HMF, on the other hand,
glucose generated only 0.16 fold relative to 3-DG. These results
clearly show that 3-DG is not the main precursor of HMF in
the case of fructose and sucrose otherwise it would have
generated more HMF as is the case relative to glucose. These
conclusions are consistent with the above assertion (shown in
Figure 1) that sucrose and fructose generate HMF through
fructopyranosyl cation pathway and glucose generates HMF
through 3-DG pathway. Furthermore, to confirm the ability of

Formation of HMF from Sucrose
J. Agric. Food Chem., Vol. 56, No. 15, 2008 6721
Figure 4. Proposed mechanism of thermal generation of glucose, levoglucosan, and fructofuranosyl cation, showing percent contribution of glucose and
fructose moieties to HMF formation at 350 and 65 °C.
Table 3. Efficiencya (× 109) of HMF Formation at Different Temperatures
Table 5. Comparison of Relative Efficiencya of HMF Formation from
Different Oligosaccharides Containing Terminal Fructose (Except Lactose)
at 250 °C
from Selected Sugars
temperature
sugarb
relative efficiency
sugarb
sucrose
fructose
glucose
levoglucosan
250 °C
300 °C
350 °C
sucrosec
raffinosed
stacchyosee
lactosef
1
1.79
1.07
0.78
0.0
5.94
3.13
1.51
0.0
7.49
3.78
1.78
0.0
1.3
0.8
0.2
a Based on chromatographic peak area of HMF/ mmol of the sugar. Values
represent average of two replicates with percent standard deviation <5%. b Sugars
were homogenized with silica gel (60%) to maximize reproducibility. c A disaccharide
(Glu-Fru). d A trisaccharide (Gal-Glu-Fru). e A tetrasaccharide (Gal-Gal-Glu-Fru).
Decomposes at 250 °C. f As control (Gal-Glu).
a Expressed as chromatographic peak area of HMF/mmol of the sugar. Values
represent average of two replicates with percent standard deviation <5%. b Sugars
were homogenized with silica gel (60%) to maximize reproducibility.
Table 4. Comparison of Relative Efficiencya of 3-DG in HMF Formation
Relative to Glucose and Fructose at 300 °C
Table 6. Efficiencya (× 108) of Levoglucosan Formation at Different
Temperatures from Selected Sugars
sugar
relative efficiency
glucose
3-DG
sucrose
fructose
0.16
1
4.5
2.4
temperature
sugarb
sucrose
fructose
glucose
levoglucosan
250 °C
300 °C
350 °C
0.0
0.0
2.74
0.0
0.57
0.0
3.58
9.66
2.87
0.0
8.35
19.7
a Based on chromatographic peak area of HMF/mmol of the sugar in the absence
of silica. Values represent average of two replicates with percent standard deviation
<5%.
a Expressed as chromatographic peak area of levoglucosan/mmol of sugar.
Values represent average of two replicates with percent standard deviation <5%.
b Sugars were homogenized with silica gel (60%) to maximize reproducibility.
sucrose residue generated after pyrolysis was dissolved in water
and analyzed. In addition to glucose, comparable amounts of
levoglucosan were also detected using commercially available
levoglucosan as a standard.
Finally, to confirm that HMF mainly arises from fructofura-
nosyl cation, (1-13Cfru)sucrose was heated in refluxing methanol
in the presence of p-toluenesulfonic acid as catalyst. According
to Moody and Richards (15) when methanol is used as solvent
under acidic conditions, the fructofuranosyl cation, if formed,
will immediately react with the solvent to produce methyl
fructofuranoside, thus preventing the formation of HMF from
the fructose moiety through the 3-DG pathway and the low
temperature of refluxing methanol (65 °C) will prevent formation
of HMF through fructofuranosyl cation pathway. Alternatively,
if sucrose was being hydrolyzed into glucose and fructose,
without the formation of fructofuranosyl cation as an intermedi-
ate, then both fructose and glucose moieties can generate HMF

6722 J. Agric. Food Chem., Vol. 56, No. 15, 2008
Perez Locas and Yaylayan
Figure 5. Fate of fructofuranosyl cation under various conditions.
through the less efficient 3-DG pathway as shown in Figure 1.
In effect, generation of HMF exclusively from the glucose
moiety in the refluxing methanol solution of (1-13Cfru)sucrose
can be considered as evidence for the fructofuranosyl cation
formation and its generation from both glucose and fructose
moieties can be considered as evidence against the fructofura-
nosyl cation formation.
was pyrolyzed in the presence of asparagine for example, more
acrylamide was generated compared to glucose and fructose
combined (17). However, if the fructofuranosyl cation is
generated under catalysis by dilute acid and at lower temper-
atures it will mainly be converted into fructose due to its fast
reaction with water. As the temperature increases, it is more
likely for the fructofuranosyl cation to be converted into HMF
or react with nucleophiles especially in low moisture systems.
The amount of HMF formed from sucrose is expected therefore
to be higher than that of fructose or glucose at higher
temperatures due to the more efficient conversion pathway of
fructofuranosyl cation into HMF relative to less efficient 3-DG
pathway that glucose and fructose follow at lower temperatures
under dilute aqueous conditions.
Consequently, when (1-13Cfru)sucrose was refluxed in metha-
nol for 10 min in the presence of p-toluenesulfonic acid
monohydrate and the solution was analyzed by GC/MS, the data
indicated that the low concentration of HMF detected was
exclusively produced from glucose alone. This was illustrated
by the lack of any label incorporation in the HMF (see Table
1), thus confirming that hydrolysis of sucrose proceeds exclu-
sively through the formation of fructofuranosyl cation. Based
on the above observations, a mechanism of HMF formation from
sucrose is proposed as shown in Figure 4. According to this
figure, sucrose under thermal treatment and/or acid catalysis
can easily cleave the glycosidic bond of the fructose moiety
with the assistance of the lone pair electrons of the fructofura-
nosyl ring oxygen to release a free glucose and a fructofuranosyl
cation as a reactive intermediate. At high temperatures and in
dry systems this cation can quickly be converted into HMF,
whereas in methanol and at low temperatures it can be trapped
as methyl fructofuranoside and therefore only free glucose
moiety can be converted into HMF through 3-deoxyglucosone
pathway following ring opening and enolization steps (see
Figure 1). This pathway is less efficient than direct dehydrations
from cyclic forms (see Table 4).
The facile formation of fructofuranosyl cation from sucrose
either under acid catalysis or just thermally generated, can also
explain the unusual reactivity of sucrose observed in the Maillard
reaction (16). The fate of this cation depends on the condition
of its generation. When it is generated at high temperatures under
dry conditions it can directly dehydrate into HMF or react with
nucleophiles such as amino acids if present. The fructofuranosyl
amine formed as a result of this interaction, can rearrange into
the well-known Heyns product (see Figure 5). When sucrose
LITERATURE CITED
(1) Association of the Industry of Juices and Nectars from Fruits
and Vegetables (AIJN), Association of the industry of juices
and nectars of the european economic community code of
practice for evaluation of fruit and vegetable juices. Brussels:
AIJN, 1996.
(2) Council Directive 2001/110/EC of 20 December 2001 relating to
honey. Off. J. Eur. Commun. 2002, 10, L 58-66.
(3) Cortes, C.; Esteve, M. J.; Frigola, A. Color of orange juice treated
by high intensity pulsed electric fields during refrigerated storage
and comparison with pasteurized juice. Food Control 2007, 19,
151–158.
(4) Berg, H. E.; van Boekel, M. A. J. S. Degradation of lactose during
heating of milk. 1. Reaction pathways. Neth. Milk Dairy J. 1994,
48, 157–175.
(5) Morales, F. J.; Romero, C.; Jimenez-Perez, S. Chromatographic
determination of bound hydroxymethylfurfural as an index of milk
protein glycosylation. J. Agric. Food Chem. 1997, 45, 1570–
1573.
(6) Kuster, B. F. M. Manufacture of 5-hydroxymethylfurfural. Starch/
Staerke 1990, 42, 314–321.
(7) Antal, M. J.; Mok, W. S. L.; Richards, G. N. Mechanism of
formation of 5-hydroxymethy)-2-furaldehyde from D-fructose and
sucrose. Carbohydr. Res. 1990, 199, 91–109.

Formation of HMF from Sucrose
J. Agric. Food Chem., Vol. 56, No. 15, 2008 6723
(8) Lee, H. S.; Nagy, S. Relative reactivities of sugars in the formation
of 5-hydroxymethyl furfural in sugar-catalyst model systems. J.
Food Process. PreserV. 1990, 14, 171–178.
(9) Queneau, Y.; Jarosz, S.; Lewandowski, B.; Fitremann, J. Sucrose
chemistry and applications of sucrochemicals. AdV. Carbohydr.
Chem. Biochem. 2008, 61, 217–292.
(10) Manley-Harris, M.; Richards, G. N. A novel fructoglucan from
the thermal polymerization of sucrose. Carbohydr. Res. 1993, 240,
183–196.
(11) Roman-Leshkov, Y.; Chheda, J. N.; Dumesic, J. A. Phase
modifiers promote efficient production of hydroxymethylfurfural
from fructose. Science 2006, 312, 1933–1937.
(12) Sommer, Y.; Hollnagel, H.; Schneider, H.; Glatt, H. R. Metabolism
of 5-hydroxymethyl-2-furfural (HMF) to the mutagen, 5-sul-
foxymethyl-2-furfural (SMF) by individual human sulfotrans-
ferases. Naunyn-Schmiedeberg’s Arch. Pharmacol. 2003, 367, 166.
(13) Svendsen, C.; Husoy, T.; Glatt, H.; Haugen, M.; Alexander, J.
5-Sulfooxymethylfurfural (SMF), the metabolite of 5-hydroxym-
ethylfurfural (HMF), increases the numbers of adenoma and
aberrant crypt foci in the intestine of min-mice. Toxicol. Lett. 2007,
172, S202–S202.
(14) Glatt, H.; Schneider, H.; Liu, Y. G. V79-hCYP2E1-hSULT1A1,
a cell line for the sensitive detection of genotoxic effects induced
by carbohydrate pyrolysis products and other food-borne chemi-
cals. Mutat. Res., Genet. Toxicol. EnViron. Mutagen. 2005, 580,
41–52.
(15) Moody, W.; Richards, G. N. Formation and equilibration of
D-fructosides and 2-thio-D-fructosides in acidified dimethyl sul-
foxide: synthetic and mechanistic aspects. Carbohydr. Chem. 1983,
124, 201–213.
(16) Karel, M.; Labuza, T. P. Nonenzymatic browning in model
systems containing sucrose. J. Agric. Food Chem. 1968, 16, 717–
719.
(17) Yaylayan, V. A.; Wnorowski, A.; Perez-Locas, C. Why asparagine
needs carbohydrates to generate acrylamide. J. Agric. Food Chem.
2003, 51, 1753–1757.
Received for review April 1, 2008. Revised manuscript received May
27, 2008. Accepted June 1, 2008. We acknowledge funding for this
research by the Natural Sciences and Engineering Research Council
of Canada (NSERC).
JF8010245