Bioactivation of food genotoxicants 5-hydroxymethylfurfural and furfuryl alcohol by sulfotransferases from human, mouse and rat: a comparative study

Article abstract of DOI:10.1007/s00204-014-1392-6

5-Hydroxymethylfurfural (HMF) and furfuryl alcohol (FFA) are moderately potent rodent carcinogens that are present in thermally processed foodstuffs. The carcinogenic effects were hypothesized to originate from sulfotransferase (SULT)-mediated bioactivation yielding DNA-reactive and mutagenic sulfate esters, a confirmed metabolic pathway of HMF and FFA in mice. It is known that orthologous SULT forms substantially differ in substrate specificity and tissue distribution. This could influence HMF- and FFA-induced carcinogenic effects. Here, we studied HMF and FFA sulfoconjugation by 30 individual SULT forms of humans, mice and rats. The catalytic efficiencies (k_cat/K_M) of HMF sulfoconjugation of human SULT1A1 (13.7?s^?1?M^?1), mouse Sult1a1 (15.8?s^?1?M^?1) and 1d1 (4.8?s^?1?M^?1) and rat Sult1a1 (5.3?s^?1?M^?1) were considerably higher than those of all other SULT forms investigated (≤0.73?s^?1M^?1). FFA sulfoconjugation was monitored using adenosine as a nucleophilic scavenger for the reactive 2-sulfoxymethylfuran (t_1/2?=?20?s at 37?°C). The resulting adduct N⁶-((furan-2-yl)methyl)-adenosine (N⁶-MF-A) was quantified by isotope-dilution UPLC-MS/MS. The rates of N⁶-MF-A formation showed that hSULT1A1 and its orthologues in mice and rats were also the most important contributors to FFA sulfoconjugation in each of the species. Taken together, the catalytic capacity of hSULT1A1 is comparable to that of mSult1a1 in mice, the species in which carcinogenic effects of HMF and FFA were detected. This is of primary concern due to the expression of hSULT1A1 in many different tissues.

Full text of DOI:10.1007/s00204-014-1392-6

Arch Toxicol
DOI 10.1007/s00204-014-1392-6
TOXICOKINETICS AND METABOLISM
Bioactivation of food genotoxicants 5‑hydroxymethylfurfural
and furfuryl alcohol by sulfotransferases from human, mouse
and rat: a comparative study
Benjamin Sachse · Walter Meinl · Yasmin Sommer ·
Hansruedi Glatt · Albrecht Seidel ·
Bernhard H. Monien
Received: 14 July 2014 / Accepted: 16 October 2014
© The Author(s) 2014. This article is published with open access at Springerlink.com
Abstract 5-Hydroxymethylfurfural (HMF) and furfuryl
alcohol (FFA) are moderately potent rodent carcinogens
that are present in thermally processed foodstuffs. The car-
cinogenic effects were hypothesized to originate from sul-
fotransferase (SULT)-mediated bioactivation yielding DNA-
reactive and mutagenic sulfate esters, a confirmed metabolic
pathway of HMF and FFA in mice. It is known that ortholo-
gous SULT forms substantially differ in substrate specific-
ity and tissue distribution. This could influence HMF- and
FFA-induced carcinogenic effects. Here, we studied HMF
and FFA sulfoconjugation by 30 individual SULT forms of
humans, mice and rats. The catalytic efficiencies (kcat/KM) of
HMF sulfoconjugation of human SULT1A1 (13.7 s−1 M−1),
mouse Sult1a1 (15.8 s−1 M−1) and 1d1 (4.8 s−1 M−1) and
rat Sult1a1 (5.3 s−1 M−1) were considerably higher than
those of all other SULT forms investigated (≤0.73 s−1 M−1).
FFA sulfoconjugation was monitored using adenosine as a
nucleophilic scavenger for the reactive 2-sulfoxymethylfuran
(t1/2 = 20 s at 37 °C). The resulting adduct N6-((furan-2-yl)
methyl)-adenosine (N6-MF-A) was quantified by isotope-
dilution UPLC-MS/MS. The rates of N6-MF-A formation
showed that hSULT1A1 and its orthologues in mice and rats
were also the most important contributors to FFA sulfocon-
jugation in each of the species. Taken together, the catalytic
capacity of hSULT1A1 is comparable to that of mSult1a1
in mice, the species in which carcinogenic effects of HMF
and FFA were detected. This is of primary concern due to the
expression of hSULT1A1 in many different tissues.
Keywords Food carcinogens · 5-Hydroxymethylfurfural ·
Furfuryl alcohol · Sulfotransferase · UPLC-MS/MS
Introduction
5-Hydroxymethylfurfural (HMF) and furfuryl alcohol
(FFA) are furan derivatives present in many foodstuffs.
HMF is formed from reducing sugars in the Maillard
Electronic supplementary material The online version of this
article (doi:10.1007/s00204-014-1392-6) contains supplementary
material, which is available to authorized users.
H. Glatt
B. Sachse · B. H. Monien (*)
e-mail: glatt@dife.de
Research Group Genotoxic Food Contaminants, German
Institute of Human Nutrition (DIfE) Potsdam-Rehbrücke,
Arthur-Scheunert-Allee 114-116, 14558 Nuthetal, Germany
e-mail: monien@dife.de
Present Address:
Y. Sommer
German Federal Institute for Risk Assessment (BfR),
Max-Dohrn-Strasse 8-10, 10589 Berlin, Germany
B. Sachse
e-mail: sachse@dife.de
A. Seidel
W. Meinl · Y. Sommer · H. Glatt
Biochemical Institute for Environmental Carcinogens, Prof. Dr.
Gernot Grimmer-Foundation, 22927 Grosshansdorf, Germany
e-mail: albrecht.seidel@biu-grimmer.de
Department of Nutritional Toxicology, German Institute
of Human Nutrition (DIfE) Potsdam-Rehbrücke, 14558 Nuthetal,
Germany
e-mail: meinl@dife.de
Y. Sommer
e-mail: yasmin.sommer@bfr.bund.de
1 3

Arch Toxicol
reaction and by acid-catalyzed dehydration. It occurs in
carbohydrate-rich foods (Husoy et al. 2008; Murkovic and
Swasti 2013), coffee, fruit juices and other beverage prod-
ucts (Murkovic and Pichler 2006; Teixido et al. 2006) and
cigarette smoke (Crump and Gardner 1989). Estimates of
the mean daily dietary intake are in the range of 4–30 mg
HMF per person (Federal Institute of Risk Assessment
(BfR) 2011). The structural congener FFA originates
from cyclization and aromatization of 2-deoxypentose, a
key intermediate of heat-induced degradation of glucose
or fructose (Brands and van Boekel 2001). It is found in
thermally processed foods such as cocoa, tea, coffee, dehy-
drated orange products, cooked meat and milk products
(Maga 1979; Murkovic and Swasti 2013). High concen-
trations of FFA in coffee (267–1,490 µg/g in three differ-
ent brands) result in intake amounts of several mg FFA
per serving (Murkovic and Swasti 2013). Humans are also
exposed by inhalation, particularly in occupational environ-
ments due to FFA use in the manufacturing of furan resins
(Ahman et al. 1991), but also in residencies, where high
levels of FFA (0.4–500 µg/g dust) were found in particulate
matter (Nilsson et al. 2005).
The dose of 188 mg HMF/kg body weight significantly
increased the incidence of liver adenoma in female mice
in a two-year bioassay. This is merely 440- to 3,300-fold
higher compared with the human HMF uptake {calcu-
lated with 70 kg human body weight and the HMF uptake
range of 4–30 mg per person given by the German Federal
Institute of Risk Assessment (BfR)] (Federal Institute of
Risk Assessment (BfR) 2011; National Toxicology Pro-
gram 2008). The European Food Safety Authority (EFSA)
stated that a margin of exposure below 10,000 indicates a
high priority for risk assessment, if the carcinogenicity was
based on a genotoxic mechanism (European Food Safety
Authority 2005). Moreover, due to considerable differences
in SULT expression and substrate specificities of ortholo-
gous SULT forms between species (Glatt 2002), human
sensitivity to HMF- and FFA-induced effects may differ
from that of rodents. Here, we present a comprehensive
overview about sulfoconjugation of HMF and FFA by indi-
vidual SULTs of human and rodent origin using cytosolic
preparations from SULT-expressing Salmonella typhimu-
rium TA1538 strains.
HMF and FFA are rodent carcinogens. It was reported
that HMF caused the formation of aberrant crypt foci in the
colon of rats (Zhang et al. 1993) and neonatal Min/+ mice
(Svendsen et al. 2009) and initiated papillomas in mouse
skin (Surh et al. 1994). Oral administration of 188 mg
HMF/kg for 104 weeks led to increased incidences of hepa-
tocellular adenomas in female B6C3F1 mice compared
with the control animals (National Toxicology Program
2008). The exposure of male B6C3F1 mice to 32 ppm FFA
in the air (equivalent to 60 mg FFA/kg bw/day) induced
renal tubule neoplasms, which were also observed in some
of the female rats (National Toxicology Program 1999).
The mutagenic potential of HMF (Glatt et al. 2005; Jan-
zowski et al. 2000; National Toxicology Program 2008)
and FFA (Aeschbacher et al. 1989; National Toxicology
Program 1999) was small in standard in vitro tests. How-
ever, the mutagenic and genotoxic activity observed in
sulfotransferase (SULT)-proficient bacteria (Glatt et al.
2011) and mammalian cell lines suggests that the tumor-
initiating potential of HMF and FFA may depend, in part or
completely, on sulfoconjugation. This hypothesis was lend
further credibility by a toxicokinetic study demonstrating
that HMF is converted into mutagenic 5-sulfoxymethylfur-
fural (SMF) in FVB/N mice (Monien et al. 2009) and that
2-methylfuran adducts such as N2-((furan-2-yl)methyl)-2′-
deoxyguanosine were detected in DNA from liver, kidney
and lung of FVB/N mice that received FFA in the drinking
water over 4 weeks (Monien et al. 2011).
Materials and methods
Chemicals
FFA was purchased from Merck Schuchardt (Hohenb-
runn, Germany). HMF, adenosine, ammonium bicarbo-
nate and all other reagents (analytical grade) were from
Sigma-Aldrich (Taufkirchen, Germany). HPLC-grade
methanol, 2-propanol and formic acid were from Carl
Roth GmbH (Karlsruhe, Germany), and HPLC-grade
water was prepared using a Milli-Q Integral Water Purifi-
cation System (Millipore Merck, Darmstadt, Germany).
Stable isotope-labeled [15N5]adenosine was from Silantes
GmbH (Munich, Germany). Sulfation reactions of HMF
and FFA yielding SMF (Monien et al. 2009) and 2-sulfoxy-
methylfuran (Monien et al. 2011), respectively, were con-
ducted as described. The cofactor 3′-phosphoadenosine-5′-
phosphosulfate (PAPS) was prepared using human PAPS
synthetase 2 expressed in Escherichia coli and was purified
using anion-exchange chromatography (purity > 99.0 %)
(Muckel et al. 2001).
Synthesis of N6-((furan-2-yl)methyl)-adenosine
(N6-MF-A) and [15N5]N6-((furan-2-yl)methyl)-adenosine
([15N5]N6-MF-A)
The ribonucleoside adduct of FFA was prepared by dis-
solving 100 mg adenosine (374 µmol) in 50 ml sodium
phosphate buffer (pH 8.0) together with 80 mg (400 µmol)
of freshly prepared 2-sulfoxymethylfuran (Monien et al.
It is disturbing that the margins between the carcino-
genic doses found in animal experiments and the estimated
mean intake values for HMF and FFA are relatively small.
1 3

Arch Toxicol
2011). The solution was stirred for 15 min at 37 °C. The
product mixture was subjected to HPLC purification using
a Prep LC 150 (Waters, Eschborn, Germany) coupled to a
PDA detector 996 (Waters) and a preparative column Sun-
Fire C18 OBD (5 µm, Waters). The product was eluted with
5 % methanol (solvent A) and 75 % methanol (solvent B)
using a linear gradient from 100 % solvent A to 10 % sol-
vent A in 20 min at a flow rate of 20 ml/min. The prod-
uct containing fractions were pooled and freeze-dried. The
purity of the product was >99 % as determined by UPLC-
UV-MS/MS (MS (ESI+) m/z = 348.1 [M + H]+). 1H-NMR
of N6-MF-A (300 MHz, dimethyl sulfoxide-d6) δ [ppm]:
8,40 (s, 1H, H8), 8,26 (bs, 1H, N-H), 8,15 (s, 1H, H2); 7,55
(s, 1H, H5″), 6,37–6,38 (d, 1H, H1′), 6,24–6,25 (d, 1H,
H4″), 5,90–5,91 (d, 1H, H3″), 5,49–5,50 (m, 1H, O-H2′),
5,42–5,44 (m, 1H, O-H5′), 5,24 (m, 1H, O-H3′), 4,62–4,63
(m, 2H, N-CH2); 4,60–4,61 (m, 1H, H3′), 4,15–4,16 (m,
1H, H2′), 3,97–3,98 (m, 1H, H4′), 3,74–3,75 (m, 1H, H5′),
3,67–3,69 (m, 1H, H5′2). The stable isotope-labeled sub-
stance [15N5]N6-MF-A was prepared in the same manner
downscaling the reaction by factor 50.
HMF sulfoconjugation assay
The reactions were conducted in mixtures of 100 µl con-
taining 25–200 µg protein, 50 mM potassium phosphate
buffer (pH 7.4), 5 mM MgCl2, 50 µM PAPS and differ-
ent concentrations of HMF at 37 °C. Preliminary experi-
ments ensured that measurements were recorded under
‘linear’ conditions, i.e., doubling of protein concentration
or incubation time was doubling the amount of sulfate ester
formed. The reactions were stopped after 30 min by add-
ing 300 µl ice-cold 2-propanol, mixed and centrifuged at
15,000 g for 10 min.
SMF concentrations of the clear supernatants were ana-
lyzed by UPLC-MS/MS. The protocol was adapted from
the quantification of SMF in plasma samples of HMF-
treated mice (Monien et al. 2009). In short, the superna-
tants were analyzed using a UPLC system (Waters) con-
nected to a Quattro Premier XE tandem quadrupole mass
spectrometer (Waters) using the negative electrospray
ionization mode. Samples of 4 µl were injected into a HSS
T3 column (1.8 µm, 2.1 × 100 mm; Waters) and eluted
with an isocratic flow of 0.35 ml/min 10 mM ammonium
acetate/methanol (95:5). The fragmentation reactions of
SMF yielding two principal ions, the sulfate ion radical
(m/z = 204.9 → 96.0) and the protonated sulfonate ion
(m/z = 204.9 → 81.0) were monitored. The tune param-
eters were as follows: temperature of the electrospray
source 120 °C; desolvation temperature 475 °C; desolva-
tion gas nitrogen (800 l/h); cone gas nitrogen (100 l/h); col-
lision gas argon (indicated cell pressure ~5 × 10−3 mbar).
For the fragmentation of SMF, collision energies were 20
and 23 eV for the transitions m/z = 204.9 → 81.0 and
m/z = 204.9 → 96.0, respectively. The dwell time was set
to 100 ms, and capillary voltage was set to 0.35 kV. The
cone and RF1 lens voltages were 32 and 0.2 V, respectively.
SMF was quantified using an external calibration line that
was linear in the concentration range of 1.0 nM (4 fmol/
injection) to 2,500 (10 pmol/injection) SMF (r2 ≥ 0.99).
Data acquisition and handling were performed with Mass-
Lynx 4.1 software (Waters). The limit of detection (LOD)
for this technique of 2.5 nM SMF has been reported (Mon-
ien et al. 2009). In the current study, this corresponds to
10 fmol SMF per injection or a reaction rate of 0.5 pmol
SMF/mg/min.
Cytosolic preparations of SULT-expressing S. typhimurium
TA1538
Cytosolic fractions containing individual SULT forms were
prepared from transgenic S. typhimurium TA1538. Table S1
in the Supplementary Material provides a summary of the
thirty SULTs used in the current study together with refer-
ences describing the cloning of the respective S. typhimurium
TA1538 strains. The expressed SULT forms are denoted
with the prefixes h, m and r for human, mouse and rat
enzymes, respectively. Bacteria were grown in Luria Broth
medium containing ampicillin (100 µg/ml) by stirring for 8 h
at 200 revolutions/min at 37 °C. The bacterial suspensions
were centrifuged at 5,000 g for 10 min. The sediments were
washed twice with 25 ml of 10 mM potassium phosphate
buffer (pH 7.4)/50 mM potassium chloride (KCP buffer)
and centrifuged again. The pellets were taken up in 800 µl
KCP buffer and sonicated thrice for 15 s with 30 s pauses
in between. The soluble fraction was attained by centrifug-
ing for 1 h at 100,000 g. The supernatants (in the following
termed cytosolic preparations) were frozen at −80 °C. The
protein concentrations were determined according to a bicin-
choninic acid (BCA) assay protocol. The SULT concentra-
tions in the cytosolic preparations were estimated after elec-
trophoretic separation and immunodetection using inclusion
bodies of different SULT forms as standards (Teubner et al.
2007). The preparation of SULT inclusion bodies and the
immunodetection of SULTs is described in the Supplemen-
tary Material, which also contains a representative immuno-
blot of different amounts of rSult1a1standard and a cytosolic
preparation of S. typhimurium expressing rSult1a1 (Fig. S1).
FFA sulfoconjugation assay
The reaction mixtures were prepared as described for HMF.
In addition, they contained 10 mM adenosine as a nucleo-
philic scavenger. Following incubation for 20 min at 37 °C,
sulfoconjugation reactions were stopped by adding 100 µl
cold methanol containing 175 fmol of [15N5]N6-MF-A as
the internal standard, mixed and centrifuged at 15,000 g
1 3

Arch Toxicol
for 10 min. The clear supernatants were subjected to
UPLC-MS/MS analyses. Incubations were performed as
quadruplicates. At higher concentrations, FFA reacts spon-
taneously with adenosine at low, but detectable, levels.
To correct for the SULT-independent adduct formation,
mixtures with SULT-deficient cytosols were carried along
simultaneously in duplicate. Controls were subtracted from
the experimental samples.
area of the daily background signal from 10 to 12 blank
incubations by more than twofold was considered to be
above the LOD. The resulting LOD values for incubations
with 0.1, 1 and 10 mM FFA were 0.18, 0.20 and 0.44 nM
N6-MF-A, respectively, corresponding to reaction rates of
18, 20 and 44 fmol N6-MF-A/mg/min.
The ribonucleoside adduct N6-((furan-2-yl)methyl)-
adenosine (N6-MF-A) was analyzed by isotope-dilution
UPLC-MS/MS with the same system described above
using a HSS T3 column (1.8 µm, 2.1 × 100 mm, Waters).
Samples of 8 µl were eluted with 10 mM ammonium
bicarbonate buffer, pH 8.0 (A) and methanol (B) using an
6-min gradient from 98 % solvent A to 15 % solvent A at
0.35 ml/min flow rate (Yin et al. 2013). The electrospray
interface of the mass spectrometer operated in the posi-
tive ion mode. The fragmentations of N6-MF-A into the
aglycone cation [N6-((furan-2-yl)methyl)-adenine-H]+
(m/z = 348.1 → 216.1), the cleavage of the 2-methylfuranyl
cation (m/z = 348.1 → 81.0) and the release of [N6-methyl-
adenine-H]+ (m/z = 348.1 → 148.0) were monitored
together with the corresponding fragmentations of the iso-
tope-labeled internal standard [15N5]N6-MF-A for the for-
mation of the aglycone (m/z = 353.1 → 221.1), the 2-meth-
ylfuran cation (m/z = 353.1 → 81.0) and 15N5-labeled
N6-methyl-adenine (m/z = 353.1 → 153.0). The N6-MF-A
content of the incubation mixture was calculated from the
ratio of peak areas resulting from the total ion current of
all three MRM signals of the analyte. For the fragmenta-
tion of N6-MF-A ([15N5]N6-MF-A), collision energies were
20, 25 and 35 eV for the transitions m/z = 348.1 → 216.1
(m/z = 353.1 → 221.1), m/z = 348.1 → 148.0
(m/z = 353.1 → 153.0) and m/z = 348.1 → 81.0
(m/z = 353.1 → 81.0), respectively. The tune parameters
were as follows: temperature of the electrospray source
110 °C; desolvation temperature 450 °C; desolvation gas
nitrogen (850 l/h); cone gas nitrogen (50 l/h); collision gas
argon (indicated cell pressure ~5 × 10−3 mbar). The dwell
time was set to 100 ms. The cone gas voltage and capillary
voltage were set to 25 V and 0.7 kV, respectively. The RF1
lens voltage was 0.1 V. Data acquisition and handling were
performed with MassLynx 4.1 software (Waters).
Results
Quantification of SMF by UPLC-MS/MS
The analytical technique used in this study was adapted
from the SMF quantification in plasma samples of HMF-
treated mice (Monien et al. 2009). This is beset with two
potential sources of inaccuracy. First, SMF is a reactive ana-
lyte with half-life times of 114, 126 and 108 min in water,
urine and plasma, respectively, at 37 °C (Monien et al.
2009). Here, we chose a reaction time of 30 min. Assuming
a continuous linear formation of SMF during the incubation
and a half-life of 120 min, about 6 % of newly formed SMF
should degrade before stopping the reaction by adding cold
2-propanol. Notably, SMF was more stable in solutions con-
taining 75 % 2-propanol at 4 °C without any detectable deg-
radation within four weeks. SMF was analyzed by UPLC-
MS/MS in the MRM mode using the specific fragmentation
reactions of the sulfate ion radical (m/z = 204.9 → 96.0)
and the protonated sulfonate ion (m/z = 204.9 → 81.0)
(Monien et al. 2009). Representative chromatograms are
shown in Fig. S2 of the Supplementary Material. The SMF
levels resulting from HMF sulfoconjugation were quantified
using SMF solutions at various concentrations as external
calibration line prepared in water/2-propanol (1:3), which
represents the solvent composition in the incubation mixture
following instant stopping and precipitating of proteins. The
second source of imprecision of mass spectrometric SMF
quantification with an external calibration is the possible
ion suppression by residues from bacterial cytosol or buffer
salts in the samples. This was tested in a separate experi-
ment measuring the peak areas of six series of solutions
with five different SMF concentrations, either prepared with
water or with the contents of the incubation mixture includ-
ing 1 mg/ml protein from a cytosolic preparation of native
S. typhimurium TA1538. The results, summarized in Table
S2 of the Supplementary Material, showed that ion suppres-
sion did not interfere with SMF detection.
The trace amounts of N6-MF-A also formed in the
absence of SULTs increased with increasing concentrations
of FFA in the incubation mixture. Due to this background,
the LOD values for the quantification of N6-MF-A were
determined individually for each of the FFA concentrations
from the background integral of the total ion current from
all three transitions of N6-MF-A monitored at the precise
retention time in incubations containing all components
and SULT-deficient cytosol from S. typhimurium. Only a
concentration of N6-MF-A exceeding the arithmetic mean
Sulfoconjugation of HMF by individual human, mouse
and rat SULTs
HMF sulfoconjugation was assessed in cytosolic prepa-
rations from various strains of S. typhimurium TA1538
expressing individual SULT forms of human, mouse and
1 3

Arch Toxicol
Table 1 Rates of sulfoconjugation assayed at four HMF concentrations in cytosolic preparations of S. typhimurium TA1538 containing different
SULT forms from human, mouse and rat
Species
SULT
Rate (pmol/mg/min)
0.1 mM HMF
1 mM HMF
10 mM HMF
100 mM HMF
Human
1A1
1A2
1A3
1B1
1C1
1C2
1C3
1E1
2A1
2B1a
2B1b
4A1
1a1
20.4 2.8
<0.5
171 18
15.2 4.5
12.1 1.5
22.2 5.2
<0.5
641 100
130 53
117 12
133 28
4.7 0.3
115 14
7.2 0.9
5.7 0.8
31.1 5.4
<0.5
699 170
628 233
406 39
272 49
11.0 1.4
285 71
3.6 0.2
7.7 1.2
201 39
<0.5
1.2 0.1
2.0 0.5
<0.5
1.1 0.3
<0.5
14.0 1.8
1.0 0.1
1.0 0.2
2.4 0.6
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
Mouse
76.6 14.4
0.9 0.1
<0.5
722 93
8.2 0.1
2.0 0.2
1,160 260
35.5 2.6
1.3 0.2
4.4 0.8
<0.5
2,010 210
66.2 6.8
16.2 1.1
6,750 1,190
1,970 70
64.4 5.0
92.4 6.7
22,400 2,400
534 52
13.1 1.7
77.5 11.7
<0.5
1b1
1c2
1d1
1e1
124 33
3.7 0.3
<0.5
226
9
2a1
9.0 0.9
32.4 3.2
<0.5
2a2
<0.5
2a3
<0.5
2b1b
5a1
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
Rat
1a1
<0.5
6.5 2.1
3.4 0.2
29.1 9.0
<0.5
43.6 13.7
25.2 0.4
120 39
2.3 0.2
17.2 0.6
62.8 4.6
<0.5
36.1 16.8
30.8 1.2
47.5 27.3
15.0 1.1
58.9 4.6
76.5 8.8
<0.5
1b1
1c1
<0.5
2.1 1.3
<0.5
1c2
2a1
<0.5
1.8 0.1
6.2 0.2
<0.5
2a3
0.6 0.0
<0.5
2a4
2b1
<0.5
<0.5
<0.5
<0.5
Values are mean SE of three or four measurements. The limit of detection (signal-to-noise ratio = 4) was 2.5 nM SMF or 10 fmol SMF on
column corresponding to a rate of 0.5 pmol/mg/min under the assay conditions used (Monien et al. 2009)
rat. This allowed including 30 SULT forms in this study
without time-consuming, multi-step chromatographic pro-
cedures for the purification of individual enzymes. The
approach offers two additional advantages. First, we found
that SULTs in cytosolic preparations are more stable com-
pared with purified enzymes (Teubner et al. 2007), and, sec-
ond, consumed cofactor 3′-phosphoadenosine-5′-phosphate
(PAP) acts as a SULT inhibitor in assays with purified
enzymes, but is rapidly decomposed in cytosolic bacte-
rial preparations. Sulfoconjugation of all 30 SULT forms
was assessed using 0.1, 1, 10 and 100 mM HMF (Table 1).
Among twelve human SULTs included in the study, nine
enzymes were found to catalyze HMF conjugation with
highest rates observed for SULT1A1, 1A2, 1A3, 1B1 and
1C2. From seven mouse Sult forms catalyzing HMF sul-
foconjugation mSult1a1, 1d1 and 1e1 stood out due to the
highest turnover rates. Generally, rat Sults showed low
activity toward HMF, the highest activity being found for
rSult1c1.
Eleven of the most active SULTs, five human, three
mouse and three rat enzymes, were subjected to a com-
prehensive kinetic investigation (compare Table 2). The
selection was based on the HMF sulfoconjugation rate
(Table 1), but also on the quantitative importance of indi-
vidual SULTs, determined by the tissue distribution in a
given adult species and the expression levels reported in
1 3

Arch Toxicol
Table 2 Kinetic parameters of HMF sulfoconjugation catalyzed by most prominent SULT forms from human, mouse and rata
a,b
Species
Human
SULT
KM (mM)a
V
(pmol/mg/min)
SULT concentration (%)
VcMAX (pmol/mg/min)
kcat/KM (s−1 M−1
)
MAX,cyt
1A1
1A2
1A3
1B1
1C2
1a1
3.2 0.4
56.5 5.4
16.7 1.8
13.3 0.3
18.8 5.4
2.0 0.3
9.5 1.1
31.9 5.7
4.9 0.7
15.7 0.9
18.4 4.7
778 24
998 47
1.0
1.4
1.4f
77,800 2,400
71,300 3,400
7,570 360
13.7
0.71
0.26
0.31
0.73
15.8
0.36
4.8
106
310
5
2
4.2
7,390 40
242 24
2,350 80
121
1.0f
4.2
24,200 2,400
55,900 1,800
6,030 250
Mouse
Rat
1b1
1d1
1a1
5
2.0
30,000 2,200
59.5 2.7
61.7 1.4
391 76
11d
0.13
2.5f
10d
273,000 20,000
45,800 2,100
2,470 60
5.3
1b1
1c1e
0.09
0.12
3,910 760
a
Values of KM and VMAX,cyt
( SE) were determined by fitting the HMF sulfoconjugation data averaged from three to five measurements con-
ducted on different occasions with the Michaelis–Menten equation (compare Fig. 1)
b
Values of VMAX,cyt were calculated from whole protein concentrations in the cytosolic preparations
c
VMAX was calculated from VMAX,cyt by correction with the concentrations of individual SULT forms in the cytosolic preparations. These were
determined by immunoblotting and comparison with various amounts of inclusion bodies containing >99 % SULT as described by Meinl et al.
(2006), unless specified otherwise
d
The enzyme levels were estimated from the intensity of the SULT band in Coomassie blue-stained polyacrylamide gels after electrophoresis
(Meinl et al. 2006)
e
Due to the inhibition of sulfoconjugation at concentrations ≥ 50 mM HMF, the data were fitted to the equation V = VMAX/(1 + KM/
[HMF] + [HMF]/KI) yielding KI = 13.2 3.4 mM. The fitting curve is shown in the Supplementary Material Figure S3
f
Meinl et al. (2013)
relevant organ tissues, especially liver, kidney and colon
(Glatt 2002; Riches et al. 2009; Teubner et al. 2002). Cur-
rent knowledge of these parameters will be discussed. Sul-
foconjugation rates of selected enzymes were monitored
at ten HMF concentrations. The resulting data were fitted
to the Michaelis–Menten equation yielding the Michaelis–
Menten constant (KM) and the maximum rate (VMAX,cyt). It
is of note that VMAX,cyt was based on the total protein con-
tent in the cytosolic preparations. Figure 1 shows represent-
ative kinetics of HMF sulfoconjugation with hSULT1A1,
immunoblotting using inclusion bodies of homogenous
SULTs as reference proteins. Usually, cytosolic SULT con-
centrations were in the range of those reported previously
(Meinl et al. 2006, 2013). The actual reaction rates VMAX
calculated from VMAX,cyt and the SULT content of individ-
ual cytosolic preparations are included in Table 2.
The KM values of HMF turnover observed for SULT1A1
forms of all three species and rSult1c1 were markedly
lower compared with those of all other SULT forms stud-
ied. The reaction rates VMAX were relatively high with
values of 77.8, 55.9 and 45.8 nmol/mg/min for SULT1A1
from human, mouse and rat, respectively. Similar or higher
rates were only found for hSULT1A2 (71.3 nmol/mg/min)
and mSult1d1 (273 nmol/mg/min). However, the high
activities of these enzymes were associated with high KM
values. The calculation of kcat/KM as a measure for the cata-
lytic efficiency of HMF sulfoconjugation confirmed the
predominant role of SULT1A1 in all three species with cat-
alytic efficiencies of 13.7, 15.8 and 5.3 s−1 M−1 for human,
mouse and rat SULT1A1, respectively.
mSult1a1 and rSult1a1 with KM values of 3.2
0.4,
2.0 0.3 and 4.9 0.7 mM, respectively. The values of the
catalytic parameters KM and VMAX,cyt of all selected SULT
forms are summarized in Table 2. The sulfoconjugation by
rSult1c1 was notable due to a decrease in the rate observed
at concentrations higher than 50 mM HMF (Supplemen-
tary Material Fig. S3). Assuming a substrate inhibitory
effect, KM and VMAX,cyt were derived from the extended
Michaelis–Menten equation V = VMAX,cyt/(1 + KM/
[HMF] + [HMF]/KI)that also allowed estimating the dis-
sociation constant (KI) of the complex formed by rSult1c1-
HMF and a second HMF molecule (KI = 13.2 3.4 mM).
Fitting of HMF sulfoconjugation rates yielded values of
apparent VMAX,cyt related to the overall protein content of
Quantification of N6-((furan-2-yl)methyl)-adenosine
(N6-MF-A) by UPLC-MS/MS
the cytosolic preparations. In order to correct the VMAX,cyt
we determined the actual SULT levels of the cytosolic
preparations containing the most effective enzymes by
,
The protocol of HMF sulfoconjugation by individual SULTs
was modified for FFA, because the reactivity of 2-sulfoxym-
ethylfuran (t1/2 = 20 s in water at 37 °C, (Glatt et al. 2012))
1 3

Arch Toxicol
Fig. 2 Molecular structure and fragmentation pattern of N6-MF-
A observed by positive ESI MS/MS collision-induced dissocia-
tion. Principal fragmentation ions of N6-MF-A were as follows:
m/z = 216.1 (aglycone of N6-MF-A), m/z = 81.0 (the cation of meth-
ylfuran) and m/z = 148.0 (N6-methyladenosine)
Fig. 1 HMF sulfoconjugation by hSULT1A1 (a), mSult1a1 (b) and
rSult1a1 (c). The rates at single HMF concentrations are mean SE
of three or four measurements. Fitting of the data with the Michae-
lis–Menten model yielded values for the catalytic parameters KM and
apparent VMAX,cyt. These were subsequently corrected for the actual
concentrations of particular SULT forms in the cytosolic preparations
(Table 2)
ribonucleoside with a single 2-methylfuran moiety attached
to the exocyclic nitrogen of the adenine, a structural scaffold
that has been observed previously for many 2′-deoxyadeno-
sine adducts of SULT-activated hydrocarbons (Herrmann et al.
2012; Monien et al. 2008, 2011, 2012). For the quantification
of N6-MF-A by UPLC-MS/MS MRM in incubation mix-
tures containing SULTs and FFA, we synthesized the isotope-
labeled reference compound [15N5]N6-MF-A. Figure 3 shows
the MRM chromatograms resulting from the neutral loss of
the ribose of N6-MF-A (m/z = 348.1 → 216.1) and the cleav-
age of the methylfuran moiety (m/z = 348.1 → 81.0). As an
additional qualifier signal, the release of the N6-methylated
adenine (m/z = 348.1 → 148.0) was monitored, which further
increased the specificity of the mass spectrometric detection
(not shown). The limit of detection (signal-to-noise ratio = 4)
of the technique was 0.8 fmol N6-MF-A per injection.
prohibited its direct quantification. We tested various nucleo-
philes as trapping agents for 2-sulfoxymethylfuran, including
adenosine, guanosine, 2′-deoxyadenosine, 2′-deoxyguano-
sine, Boc-lysine, glutathione and N-acetylcysteine. Adenosine
was selected because it was well soluble in the incubation
medium and formed a single reaction product with 2-sul-
foxymethylfuran. The presumed product N6-MF-A was syn-
thesized via a biomimetic approach. The mass spectrometric
collision-induced dissociation showed the expected fragmen-
tation patterns (Fig. 2): the cleavage of the 2-methylfuran
cation (m/z = 348.1 → 81.0), the neutral loss of the ribose
unit (m/z = 348.1 → 216.1) and the release of the positively
charged N6-methylated adenine (m/z = 348.1 → 148.0). The
1H-NMR spectrum (Supplementary Material Fig. S4) cor-
roborated the expected structure of N6-MF-A to consist of the
Sulfoconjugation of FFA by individual SULT forms
from human, mouse and rat
The sulfoconjugation assay was performed under simi-
lar conditions as used for HMF, with three different
1 3

Arch Toxicol
about 1.5 % of the 2-sulfoxymethylfuran was converted
into N6-MF-A (data not shown). However, the estimation
was beset with different inaccuracies and, therefore, was
not used to determine FFA sulfoconjugation quantitatively.
First, the chemical syntheses of 2-sulfoxymethylfuran pro-
vide mixtures of different salts, e.g., sodium sulfate, which
contained only 18.3 % 2-sulfoxymethylfuran in a previous
study (Glatt et al. 2012) and only 13.2 % 2-sulfoxymethyl-
furan in the current study. Consequently, incubation mix-
tures of chemically synthesized 2-sulfoxymethylfuran with
adenosine did not exactly match the chemical composition
of the sulfoconjugation assay. Furthermore, the adduct for-
mation in mixtures of chemically synthesized 2-sulfoxym-
ethylfuran with adenosine may differ from that in incuba-
tions in which 2-sulfoxymethylfuran is formed slowly from
FFA and SULTs.
The N6-MF-A formation rates of the most important
SULT forms observed at 1 mM FFA were corrected by the
actual SULT level of the cytosolic preparations (Table 2).
SULT1A1 was the most efficacious catalyst of FFA sul-
foconjugation in human, mouse and rat, leading to for-
mation of 57,500, 61,700 and 24,400 fmol N6-MF-A/mg/
min, respectively (Table 3). Among the human enzymes,
SULT1C2 was also very efficacious (45,900 fmol N6-MF-
A/mg/min). In contrast, the FFA turnover by other SULT
forms was remarkably less effective. The rate of N6-MF-
A appearance observed in incubation mixtures containing
hSULT1A1 was about 11- to 30-fold greater compared
with hSULT1A2, 1A3, 1B1, 1C3 and 2A1. Mouse Sult1a1
catalyzed FFA bioactivation about 14 times more effec-
tively compared with mSult1d1, whereas N6-MF-A was not
detectable at 1 mM FFA in the presence of mSult1b1. In
rats, the formation of N6-MF-A was about 6- and 71-fold
greater in cytosolic preparations containing rSult1a1 com-
pared with those of rSult1c1 and 1b1, respectively. How-
ever, the data indicated that hydroxysteroid rSult2a3, which
is predominantly expressed in female liver (Dunn and
Klaassen 1998), may be the most effective catalyst of FFA
sulfoconjugation in rats.
Fig. 3 UPLC-MS/MS MRM analysis of N6-MF-A in a cytosolic
preparation containing hSULT1A1, FFA and adenosine. The chroma-
tograms originate from the fragmentations m/z = 348.1 → 216.1 (a)
and m/z = 348.1 → 81.0 (b) of N6-MF-A and from the transitions
m/z = 353.1 → 221.1 (c) and m/z = 353.1 → 81.0 (d) of the internal
isotope-labeled standard [15N5]N6-MF-A (7.0 fmol/injection)
concentrations of FFA (0.1, 1 and 10 mM). Table 3 sum-
marizes the formation rates of N6-MF-A, which were used
as surrogate parameter for the SULT-specific capacities of
FFA turnover. In cytosolic preparations containing human
SULTs and 1 mM FFA, N6-MF-A was most rapidly formed
in the presence of SULT1A1, 1C2 and 1B1. Other human
enzymes leading to detectable levels of N6-MF-A were
SULT1A2, 1A3, 1C3, 1E1 and 2A1. Among all mouse Sult
forms tested, FFA sulfoconjugation by Sult1a1 and 1d1 led
to relatively high levels of N6-MF-A, which was less effi-
ciently formed in incubations containing mSult1b1, 1e1,
2a1 and 2a2. In rats, Sult2a3 was an efficient catalyst of
FFA sulfoconjugation, whereas rSult1a1, 1b1, 1c1 and 2a1
showed lower, but measurable activity.
Discussion
Considerable species-dependent differences of SULT
expression and substrate specificities have been found
(Glatt 2002). For example, tamoxifen is a potent carcino-
gen in rat but not in human liver. A major phase I metab-
olite, α-hydroxytamoxifen is activated by hepatic rat
SULT2A3 producing a DNA-reactive sulfate ester (Glatt
et al. 1998). Further, the benzylic alcohol 2-hydroxy-
3-methylcholanthrene, a primary metabolite of the carcino-
gen 3-methylcholanthrene, was efficiently bioactivated by
hSULT1A1 expressed in S. typhimurium, but not by the
We determined the reaction efficiency of chemically syn-
thesized 2-sulfoxymethylfuran with adenosine in the pres-
ence of cytosolic protein of native S. typhimurium TA1538
under normal incubation conditions. In this experiment,
1 3

Arch Toxicol
Table 3 Rates of N6-
Species
Human
SULT
Rate (fmol N6-MF-A/mg/min)
MF-A formation after FFA
sulfoconjugation in cytosolic
preparations of S. typhimurium
TA1538 containing different
SULT forms and adenosine as
nucleophilic scavenger
0.1 mM FFA
1 mM FFA
10 mM FFA
1 mM FFA
(corr.)a
1A1
1A2
1A3
1B1
1C1
1C2
1C3
1E1
2A1
2B1a
2B1b
4A1
1a1
73.5 1.1
<18
575
2
931
150
226
5
4
3
57,500 200
2,480 30
1,950 30
5,010 50
34.8 0.5
27.3 0.5
<18
23.3 0.3
<18
211
<20
459
2
1,070 20
<44
51.3 0.8
<18
2
2,230 10
45,900 200
3,090 10b
61.8 0.3
<20
248
3
<18
47.0 0.9
<18
37.0 0.6
<20
386
<44
<44
<44
6
2,110 40b
<18
<18
<20
<18
<20
Mouse
361
<18
<18
4
2,590 20
<20
5,140 60
74.3 1.6
<44
61,700 400
4,260 60
1b1
1c2
<20
1d1
1e1
52.5 1.7
<18
468
6
2,610 30
49.3 1.2
<20
425
154
216
<44
<44
<44
<44
314
3
2
3
2a1
<18
Values are mean SE of
four measurements from
2a2
<18
<20
independent incubations. Only
N6-MF-A levels exceeding the
background signal by more than
twofold were considered to be
above the detection limit
2a3
<18
<20
2b1b
5a1
<18
<20
<18
<20
Rat
1a1
<18
31.8 0.8
34.3 1.0
97.8 2.3
<20
24,400 600
343 10
a
The rate of N6-MF-A
1b1
1c1
<18
4
formation at 1 mM FFA
<18
492 13
<44
3,910 90
was corrected by the actual
concentrations of individual
SULT forms in the cytosolic
preparations (compare Table 2)
1c2
<18
2a1
<18
24.8 1.4
228
7
2a3
48.3 1.3
<18
461
<20
<20
8
3,420 30
<44
b
If not listed in Table 2, the
2a4
SULT concentrations were
reported in (Meinl et al. 2006)
2b1
<18
<44
orthologues of mouse, rat and dog (Meinl et al. 2013). Par-
allel species-dependent differences may introduce incalcu-
lable uncertainties in human risk assessment of SULT-acti-
vated compounds from the results of two-year bioassays.
The substituted furans HMF and FFA are rodent carcino-
gens present at high concentrations in the human diet. The
carcinogenicity was proposed to originate from sulfocon-
jugation yielding reactive sulfate esters (Glatt and Sommer
2006; Surh 1998; Surh et al. 1990) that form specific DNA
adducts in mammalian cell culture (Monien et al. 2012)
and mice (Monien et al. 2011). It is of utmost importance
to know whether humans may be more (or less) suscepti-
ble to the carcinogenic effects of HMF and FFA intake than
rodents.
mouse and eight rat SULT forms (Tables 1, 3). Eleven
enzymes were selected for detailed kinetic analyses, based
on the results of the initial screening and the expression
levels in liver (target of HMF-induced carcinogenicity in
the mouse), kidney (target of FFA-induced carcinogenic-
ity in the mouse) and intestine (induction of preneoplasia
by HMF in rats and mice). In humans, SULT1A1 has a
dominant role in the activation of many benzylic alcohols,
a broad substrate tolerance, and is highly expressed in
numerous tissues (Glatt et al. 2001; Glatt and Meinl 2005).
SULT1C2 was of special interest because HMF and FFA
were particularly mutagenic in S. typhimurium TA100-
hSULT1C2 (Glatt et al. 2012), and hSULT1B1 is present
at high levels in colon, rectum and in other compartments
of the gastrointestinal tract and in the liver (Teubner et al.
2007; Wang et al. 1998), all of which are potential target
Here, we studied the sulfoconjugation of the furan
derivatives in vitro taking into account twelve human, ten
1 3

Arch Toxicol
organs for HMF-induced carcinogenicity (National Toxi-
cology Program 2008; Svendsen et al. 2009, 524; Zhang
et al. 1993, 83). SULT1A3 is an abundant enzyme in many
tissues (except liver) with especially high levels in the gas-
trointestinal tract (Teubner et al. 2007).
The relative importance of SULT1A1 for the turnover
of many other substrates has been reported. For example,
a survey of eight human SULTs identified SULT1A1 as
the predominant enzyme for the sulfoconjugation of vari-
ous endogenous compounds, e.g., epinephrine, 2-hydrox-
yestradiol, T2 (3,5-diiodo-l-thyronine) and cholesterol, as
well as xenobiotica, e.g., minoxidil, 4-methylphenol and
α-zearalenol (Allali-Hassani et al. 2007). It is of note that
the KM values of the SULT1A1 forms for the turnover of
HMF were high compared with those of other substrates.
Honma reported KM values of 3.0, 18 and 130 µM for the
sulfoconjugation of p-nitrophenol, 6-hydroxymelatonin
and dopamine by hSULT1A1, respectively (Honma et al.
2001). Similar to HMF, a KM of 2.4 mM was determined
for the turnover of paracetamol by hSULT1A1 (Adjei et al.
2008). The high KM values of all SULT1A1 forms observed
in the current study indicate a relatively weak binding affin-
ity of HMF. This is a possible reason why only 500 ppm of
100 mg HMF/kg body weight administered to mice were
converted to SMF (Monien et al. 2009). To our knowledge,
KM values for the conversion of HMF by alcohol and alde-
hyde dehydrogenases have not been reported.
Mouse Sults were chosen in consideration of the major
target tissues of HMF- and FFA-mediated carcinogenic
effects in liver (National Toxicology Program 2008) and in
kidneys (National Toxicology Program 1999), respectively.
Alnouti and Klaassen reported that mSult1a1 is the predom-
inant hepatic form as judged from mRNA levels. Sult1d1 is
highly expressed in kidney and at moderate levels through-
out the gastrointestinal tract of mice, which is also rich in
mSult1b1 (Alnouti and Klaassen 2006). Further, HMF and
FFA produced strong mutagenic effects in mSult1a1- and
mSult1d1-expressing TA100 strains (Glatt et al. 2012). Rat
Sults were selected as follows. In male rats, Sult1a1 and
1c1 are expressed at substantial levels in the liver but also in
the kidneys (Dunn and Klaassen 1998; Honma et al. 2001).
Rat Sult1b1 is expressed in liver and kidneys of males and
females but also in the intestine, a reported target organ for
the neoplastic effect of HMF (Zhang et al. 1992).
The catalytic efficiencies of HMF turnover by the
selected SULT forms were calculated from Michaelis–
Menten constants KM and turnover numbers kcat (Table 2).
Human SULT1A1 and the orthologous enzymes in mice
and rats were the predominant catalysts of HMF sulfo-
conjugation. The data showed that among the other SULT
forms only mSult1d1 was similarly effective, but required
high substrate concentrations (KM = 31.9 5.7 mM). The
formation rates of N6-MF-A, used as a surrogate marker
of FFA sulfoconjugation (Table 3), largely paralleled the
ranking of HMF sulfoconjugation capacity. There were
two notable exceptions. First, N6-MF-A was formed almost
equally well by hSULT1C2 and hSULT1A1 in the pres-
ence of 1 mM FFA, whereas HMF sulfoconjugation was
catalyzed about 19 times more effective by hSULT1A1
compared with hSULT1C2. We did not further investi-
gate this difference, because hSULT1C2 may be of minor
importance for the metabolism due to its low expression.
The mRNA was primarily found in fetal tissues (Sakak-
ibara et al. 1998), while the hSULT1C2 protein was hith-
erto detected only in Caco-2 cells (Meinl et al. 2008). The
second deviation between HMF and FFA sulfoconjuga-
tion was observed for mSul1d1. The catalytic efficiency of
mSult1d1 regarding HMF sulfoconjugation was three times
less than that observed for mSult1a1, whereas the N6-MF-
A formation rate in the presence of mSul1d1 at 1 mM FFA
was about 14-fold lower compared with that observed for
mSult1a1. Apart from these differences, the corrected rates
of N6-MF-A formation indicated that, in agreement to HMF
sulfoconjugation, hSULT1A1 and its rodent orthologues
were the most effective catalysts of FFA turnover.
Due to the lability of the reactive sulfate esters, the bioac-
tivation of benzylic alcohols by individual SULT forms was
frequently assessed by the mutagenic effect of the substrates
in SULT-expressing S. typhimurium TA100 or TA1538
strains (Glatt and Meinl 2004). It was found that hSULT1A1
was most important for the sulfoconjugation of various pro-
mutagens, e.g., ( )-1′-hydroxymethyleugenol (Herrmann
et al. 2012), nitrofen (Glatt and Meinl 2004), 1-hydroxym-
ethylpyrene (Meinl et al. 2002) and 2-hydroxy-3-methyl-
cholanthrene (Meinl et al. 2013), whereas other SULT forms
of human origin or from other species contributed little or
nothing at all to the mutagenic activity of these compounds.
The results presented in this study were in agreement with
a previous report about the mutagenicity of HMF and FFA
in S. typhimurium TA100 strains expressing different SULT
forms. Mutagenic effects of HMF, albeit small, were detect-
able in those strains that expressed hSULT1A1, 1A3 and
1C2 as well as mSult1a1 and 1d1. The highest mutagenic-
ity of FFA was observed in bacteria expressing hSULT1A1,
1C2 and mSult1a1 (Glatt et al. 2012).
Besides the species-dependent efficiency of SULT1A1-
mediated bioactivation, also the expression levels of the
enzyme may influence the carcinogenic risk of HMF and
FFA in each of the organisms. Despite the principal role
of SULT1A1 in the metabolism of many endogenous and
exogenous compounds, the knowledge about its expres-
sion in different organisms is limited. In humans, SULT1A1
is found in many tissues at high levels, e.g., in liver, lung,
brain, throughout the gastrointestinal tract and in kidneys
(Glatt 2002; Meinl et al. 2006; Riches et al. 2009; Teub-
ner et al. 2007). The SULT1A1 levels found in human liver
1 3

Arch Toxicol
Conflict of interest The authors declare that they have no conflicts
of interest.
(420–4,900 ng/mg cytosolic protein, n = 28) (Riches et al.
2009) and the ileum (990–1,600 ng/mg cytosolic protein,
n = 4) (Teubner et al. 2007) were considerably higher com-
pared with those in kidneys (30–270 ng/mg cytosolic pro-
tein, n = 10) (Meinl et al. 2006; Riches et al. 2009). Little
data of Sult1a1 expression on the protein level are available
for mice and rats. Honma et al. reported hepatic Sult1a1
Open Access This article is distributed under the terms of the Crea-
tive Commons Attribution License which permits any use, distribu-
tion, and reproduction in any medium, provided the original author(s)
and the source are credited.
amounts of 2,300
100 ng/mg cytosolic protein (n = 3)
and 4,400
200 ng/mg cytosolic protein (n = 3) in male
References
and female mice, respectively, and 9,100 200 ng/mg cyto-
solic protein (n = 3) and 6,300 500 ng/mg cytosolic pro-
tein (n = 3) in liver of male and female rats, respectively.
Judged from concentrations of mRNA, the highest amounts
of mSult1a1 were expressed in liver, lung and large intestine,
whereas the Sult1a1 expression in the small intestine of mice
was negligible as judged from the mRNA levels (Alnouti and
Klaassen 2006). In rats, mRNA of Sult1a1 is expressed in
many tissues, albeit at levels that are minute in comparison
with the liver (Dunn and Klaassen 1998). Taken together,
the current knowledge about SULT1A1 expression in dif-
ferent species does not allow assuming a species-dependent
carcinogenic risk of SULT1A1-activated compounds. How-
ever, HMF and FFA were more effectively bioactivated
by mSult1a1 than by rSult1a1. This may explain why car-
cinogenic effects of HMF and FFA were observed in mice
and not in rats (National Toxicology Program 1999, 2008).
Our data further indicate that hSULT1A1 conjugates HMF
and FFA as efficient as mSult1a1. However, looking only
at SULTs would present a simplification. The concentration
of the reactive sulfate esters in particular tissues and species
may also be influenced by various other parameters, such as
detoxification of the precursors and the sulfate esters, tis-
sue-specific directed export and import of the sulfate esters,
transport via the blood stream and excretion.
In summary, it becomes increasingly clear that HMF
and FFA are sulfoconjugated not only in mice (Monien
et al. 2009, 2011) but also in the human body. The char-
acterization of HMF and FFA sulfoconjugation capacity
of individual human, mouse and rat SULT forms showed
that hSULT1A1 and the orthologous enzymes in mice and
rats catalyzed the bioactivation of the substituted furans
most efficiently compared with all other SULTs. Human
SULT1A1 was about as effective as Sult1a1 of mice, the
principal species in which carcinogenic effects of HMF
(liver) and FFA (kidney) were observed. This raises concern
because hSULT1A1 is ubiquitously expressed at high levels
in several tissues, e.g., liver, lung, gastrointestinal tract and
kidney (Glatt 2002; Meinl et al. 2006; Teubner et al. 2007).
Adjei AA, Gaedigk A, Simon SD, Weinshilboum RM, Leeder JS
(2008) Interindividual variability in acetaminophen sulfation by
human fetal liver: implications for pharmacogenetic investiga-
tions of drug-induced birth defects. Birth Defects Res A Clin Mol
Teratol 82:155–165
Aeschbacher HU, Wolleb U, Loliger J, Spadone JC, Liardon R (1989)
Contribution of coffee aroma constituents to the mutagenicity of
coffee. Food Chem Toxicol 27:227–232
Ahman M, Alexandersson R, Ekholm U, Bergstrom B, Dahlqvist M,
Ulfvarson U (1991) Impeded lung function in moulders and core-
makers handling furan resin sand. Int Arch Occup Environ Health
63:175–180
Allali-Hassani A, Pan PW, Dombrovski L et al (2007) Structural
and chemical profiling of the human cytosolic sulfotransferases.
PLoS Biol 5:e97
Alnouti Y, Klaassen CD (2006) Tissue distribution and ontogeny of
sulfotransferase enzymes in mice. Toxicol Sci 93:242–255
Brands CM, van Boekel MA (2001) Reactions of monosaccharides
during heating of sugar-casein systems: building of a reaction
network model. J Agric Food Chem 49:4667–4675
Crump DR, Gardner D (1989) Sources and concentrations of alde-
hydes and ketones in indoor environments in the UK. Environ Int
15:455–462
Dunn RT 2nd, Klaassen CD (1998) Tissue-specific expression of rat
sulfotransferase messenger RNAs. Drug Metab Dispos 26:598–604
European Food Safety Authority (2005) Opinion of the Scientific
Committee on a request from EFSA related to a harmonised
approach for risk assessment of substances which are both geno-
toxic and carcinogenic. EFSA J 282:1–31
Federal Institute of Risk Assessment (BfR) (2011) Statement No.
030/2011, 5-HMF-Gehalte in Lebensmitteln sind nach derzeiti-
gem wissenschaftlichen Kenntnisstand gesundheitlich unprob-
lematisch. vol 30
Glatt HR (2002) Sulphotransferases. In: Ioannides C (ed) Handbook
of enzyme systems that metabolise drugs and other xenobiotics.
Wiley, Sussex, pp 353–439
Glatt HR, Meinl W (2004) Use of genetically manipulated Salmo-
nella typhimurium strains to evaluate the role of sulfotransferases
and acetyltransferases in nitrofen mutagenicity. Carcinogenesis
25:779–786
Glatt HR, Meinl W (2005) Sulfotransferases and acetyltransferases in
mutagenicity testing: technical aspects. Meth Enzymol 400:230–249
Glatt HR, Sommer Y (2006) Health risks by 5-hydroxymethylfur-
fural (HMF) and related compounds. In: Skog K, Alexander J
(eds) Acrylamide and other health hazardous compounds in heat-
treated foods. Woodhead Publishing, Cambridge, pp 328–357
Glatt HR, Davis W, Meinl W, Hermersdörfer H, Venitt S, Phillips DH
(1998) Rat, but not human, sulfotransferase activates a tamoxifen
metabolite to produce DNA adducts and gene mutations in bacte-
ria and mammalian cells in culture. Carcinogenesis 19:1709–1713
Glatt HR, Boeing H, Engelke CEH et al (2001) Human cytosolic sul-
photransferases: genetics, characteristics, toxicological aspects.
Mutat Res 482:27–40
Acknowledgments We thank Martina Scholtyssek for her excel-
lent technical assistance. This work was supported by a grant from
the German Research Foundation (MO 2520/1-1) and by the German
Institute of Human Nutrition (DIfE), Potsdam-Rehbrücke, Germany.
1 3

Arch Toxicol
Glatt HR, Schneider H, Liu Y-G (2005) 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 580:41–52
Glatt HR, Baasanjav-Gerber C, Schumacher F et al (2011) 1-Meth-
oxy-3-indolylmethyl glucosinolate; a potent genotoxicant in bac-
terial and mammalian cells: mechanisms of bioactivation. Chem
Biol Interact 192:81–86
Glatt HR, Schneider H, Murkovic M, Monien BH, Meinl W (2012)
Hydroxymethyl-substituted furans: mutagenicity in Salmonella
typhimurium strains engineered for expression of various human
and rodent sulphotransferases. Mutagenesis 27:41–48
high performance liquid chromatography. Naunyn-Schmiede-
berg’s Arch Pharmacol 363:R137
Murkovic M, Pichler N (2006) Analysis of 5-hydroxymethylfurfural
in coffee, dried fruits and urine. Mol Nutr Food Res 50:842–846
Murkovic M, Swasti YR (2013) 5-Hydroxymethyl-furfural and fur-
furyl alcohol: occurrence, exposure and detection. In: Ciesarova
Z, Pedreschi Plasencia F (eds) Chemical food safety and health.
Food science and technology. Nova, New York, pp 43–55
National Toxicology Program (1999) Toxicology and carcinogenesis
studies of furfuryl alcohol (CAS No.98-00-0) in F344/N rats and
B6C3F1 mice (inhalation studies). National Toxicology Program
technical report series 482
Herrmann K, Engst W, Appel KE, Monien BH, Glatt HR (2012)
Identification of human and murine sulfotransferases able to
activate hydroxylated metabolites of methyleugenol to mutagens
in Salmonella typhimurium and detection of associated DNA
adducts using UPLC-MS/MS methods. Mutagenesis 27:453–462
Honma W, Kamiyama Y, Yoshinari K et al (2001) Enzymatic charac-
terization and interspecies difference of phenol sulfotransferases,
ST1A forms. Drug Metab Dispos 29:274–281
National Toxicology Program (2008) Toxicology and carcinogenesis
studies of 5-(Hydroxymethyl)-2-furfural (CAS No. 67-47-0) in
F344/N rats and B6C3F1 mice (gavage studies). National Toxi-
cology Program technical report series554
Nilsson A, Lagesson V, Bornehag CG, Sundell J, Tagesson C (2005)
Quantitative determination of volatile organic compounds in
indoor dust using gas chromatography-UV spectrometry. Environ
Int 31:1141–1148
Husoy T, Haugen M, Murkovic M et al (2008) Dietary exposure to
5-hydroxymethylfurfural from Norwegian food and correlations
with urine metabolites of short-term exposure. Food Chem Toxi-
col 46:3697–3702
Riches Z, Stanley EL, Bloomer JC, Coughtrie MW (2009) Quantita-
tive evaluation of the expression and activity of five major sul-
fotransferases (SULTs) in human tissues: the SULT “pie”. Drug
Metab Dispos 37:2255–2261
Janzowski C, Glaab V, Samimi E, Schlatter J, Eisenbrand G (2000)
5-Hydroxymethylfurfural: assessment of mutagenicity, DNA-
damaging potential and reactivity towards cellular glutathione.
Food Chem Toxicol 38:801–809
Maga JA (1979) Furans in foods. CRC Crit Rev Food Sci Nutr
11:355–400
Sakakibara Y, Yanagisawa K, Katafuchi J et al (1998) Molecular clon-
ing, expression, and characterization of novel human SULT1C
sulfotransferases that catalyze the sulfonation of N-hydroxy-
2-acetylaminofluorene. J Biol Chem 273:33929–33935
Surh YJ (1998) Bioactivation of benzylic and allylic alcohols via
sulfo-conjugation. Chem Biol Interact 109:221–235
Meinl W, Meerman JH, Glatt HR (2002) Differential activation of pro-
mutagens by alloenzymes of human sulfotransferase 1A2 expressed
in Salmonella typhimurium. Pharmacogenetics 12:677–689
Meinl W, Pabel U, Osterloh-Quiroz M, Hengstler JG, Glatt HR
(2006) Human sulfotransferases are involved in the activation of
aristolochic acids and are expressed in renal target tissue. Int J
Cancer 118:1090–1097
Meinl W, Ebert B, Glatt H, Lampen A (2008) Sulfotransferase forms
expressed in human intestinal Caco-2 and TC7 cells at varying
stages of differentiation and role in benzo[a]pyrene metabolism.
Drug Metab Dispos 36:276–283
Meinl W, Tsoi C, Swedmark S, Tibbs ZE, Falany CN, Glatt H (2013)
Highly selective bioactivation of 1- and 2-hydroxy-3-methylcho-
lanthrene to mutagens by individual human and other mamma-
lian sulphotransferases expressed in Salmonella typhimurium.
Mutagenesis 28:609–619
SurhYJ, Blomquist JC, Liem A, Miller JA (1990) Metabolic activation of
9-hydroxymethyl-10-methylanthracene and 1-hydroxymethylpyrene
to electrophilic, mutagenic and tumorigenic sulfuric acid esters by
rat hepatic sulfotransferase activity. Carcinogenesis 11:1451–1460
Surh YJ, Liem A, Miller JA, Tannenbaum SR (1994) 5-Sulfooxym-
ethylfurfural as a possible ultimate mutagenic and carcinogenic
metabolite of the Maillard reaction product, 5-hydroxymethylfur-
fural. Carcinogenesis 15:2375–2377
Svendsen C, Husoy T, Glatt HR, Paulsen JE, Alexander J (2009)
5-Hydroxymethylfurfural and 5-sulfooxymethylfurfural increase
adenoma and flat ACF number in the intestine of Min/+ mice.
Anticancer Res 29:1921–1926
Teixido E, Santos FJ, Puignou L, Galceran MT (2006) Analysis of
5-hydroxymethylfurfural in foods by gas chromatography-mass
spectrometry. J Chromatogr A 1135:85–90
Teubner W, Meinl W, Glatt HR (2002) Stable expression of rat sul-
fotransferase 1B1 in V79 cells: activation of benzylic alcohols to
mutagens. Carcinogenesis 23:1877–1884
Monien BH, Müller C, Engst W, Frank H, Seidel A, Glatt HR (2008)
Time course of hepatic 1-methylpyrene DNA adducts in rats
determined by isotope dilution LC-MS/MS and 32
Teubner W, Meinl W, Florian S, Kretzschmar M, Glatt H (2007)
Identification and localization of soluble sulfotransferases in the
human gastrointestinal tract. Biochem J 404:207–215
P-postlabeling.
Chem Res Toxicol 21:2017–2025
Monien BH, Frank H, Seidel A, Glatt HR (2009) Conversion of the
common food constituent 5-hydroxymethylfurfural into a muta-
genic and carcinogenic sulfuric acid ester in the mouse in vivo.
Chem Res Toxicol 22:1123–1128
Wang J, Falany JL, Falany CN (1998) Expression and characteriza-
tion of a novel thyroid hormone-sulfating form of cytosolic sul-
fotransferase from human liver. Mol Pharmacol 53:274–282
Yin R, Liu S, Zhao C, Lu M, Tang MS, Wang H (2013) An ammo-
nium bicarbonate-enhanced stable isotope dilution UHPLC-MS/
MS method for sensitive and accurate quantification of acrolein-
DNA adducts in human leukocytes. Anal Chem 85:3190–3197
Zhang XM, Stamp D, Minkin S et al (1992) Promotion of aberrant
crypt foci and cancer in rat colon by thermolyzed protein. J Natl
Cancer Inst 84:1026–1030
Monien BH, Herrmann K, Florian S, Glatt HR (2011) Metabolic activa-
tion of furfuryl alcohol: formation of 2-methylfuranyl DNA adducts
in Salmonella typhimurium strains expressing human sulfotrans-
ferase 1A1 and in FVB/N mice. Carcinogenesis 32:1533–1539
Monien BH, Engst W, Barknowitz G, Seidel A, Glatt HR (2012)
Mutagenicity of 5-hydroxymethylfurfural in V79 cells expressing
human SULT1A1: identification and mass spectrometric quantifi-
cation of DNA adducts formed. Chem Res Toxicol 25:1484–1492
Muckel E, Landsiedel R, Glatt HR (2001) Preparation of highly pure
3´-phosphoadenosine-5´-phosphosulfate (PAPS) using recombi-
nant human PAPS synthetase and purification by ion exchange
Zhang XM, Chan CC, Stamp D, Minkin S, Archer MC, Bruce WR
(1993) Initiation and promotion of colonic aberrant crypt foci in
rats by 5-hydroxymethyl-2-furaldehyde in thermolyzed sucrose.
Carcinogenesis 14:773–775
1 3