Conversion of dehydrodiferulic acids by human intestinal microbiota

Article abstract of DOI:10.1021/jf900159h

Plant cell wall associated dehydrodiferulic acids (DFA) are abundant components of cereal insoluble dietary fibers ingested by humans. The ability of human intestinal microbiota to convert DFA was studied in vitro by incubating 8-O-4- and 5-5-coupled DFA with fecal suspensions. 8-O-4-DFA was completely degraded by the intestinal microbiota of the majority of donors, yielding homovanillic acid, 3-(3,4-dihydroxyphenyl)propionic acid, and 3,4-dihydroxyphenylacetic acid as the main metabolites. The transient formation of ferulic acid and presumably 3-(3-hydroxy-4-methoxyphenyl)pyruvic acid suggests an initial cleavage of the ether bond. In contrast to 8-O-4-DFA, the 5-5-coupled DFA was not cleaved into monomers by any of the fecal suspensions. Only the side chains were hydrogenated and the methoxy groups were demethylated. The cleavage of DFA by human intestinal microbiota, which depended on their coupling type, may affect both the bioavailability of DFA and the degradability of DFA-coupled fiber in the gut.

Full text of DOI:10.1021/jf900159h

3356 J. Agric. Food Chem. 2009, 57, 3356–3362
Conversion of Dehydrodiferulic Acids by Human
Intestinal Microbiota
ANNETT BRAUNE,*,†,§ MIRKO BUNZEL,†,# REIKO YONEKURA,⊥ AND
§
MICHAEL BLAUT
Department of Gastrointestinal Microbiology, German Institute of Human Nutrition
Potsdam-Rehbruecke, Arthur-Scheunert-Allee 114-116, 14558 Nuthetal, Germany; Department of
Food Science and Nutrition, University of Minnesota, 1334 Eckles Avenue, St. Paul,
Minnesota 55108; and Institute of Biochemistry and Food Chemistry, University of Hamburg,
Grindelallee 117, 20146 Hamburg, Germany
Plant cell wall associated dehydrodiferulic acids (DFA) are abundant components of cereal insoluble
dietary fibers ingested by humans. The ability of human intestinal microbiota to convert DFA was
studied in vitro by incubating 8-O-4- and 5-5-coupled DFA with fecal suspensions. 8-O-4-DFA was
completely degraded by the intestinal microbiota of the majority of donors, yielding homovanillic acid,
3-(3,4-dihydroxyphenyl)propionic acid, and 3,4-dihydroxyphenylacetic acid as the main metabolites.
The transient formation of ferulic acid and presumably 3-(3-hydroxy-4-methoxyphenyl)pyruvic acid
suggests an initial cleavage of the ether bond. In contrast to 8-O-4-DFA, the 5-5-coupled DFA was
not cleaved into monomers by any of the fecal suspensions. Only the side chains were hydrogenated
and the methoxy groups were demethylated. The cleavage of DFA by human intestinal microbiota,
which depended on their coupling type, may affect both the bioavailability of DFA and the degradability
of DFA-coupled fiber in the gut.
KEYWORDS: Diferulic acids; diferulate; ferulic acid; 8-O-4-dehydrodiferulic acid; 5-5-dehydrodiferulic acid;
human intestinal microbiota; conversion
INTRODUCTION
risk (12). However, protection could also be attributed to other
nutrients contained in high-fiber diets or phenolic compounds
associated with or bound to dietary fiber (13). Thus, hydroxy-
cinnamic acids such as ferulic acid and DFA have antioxidant
and chemoprotective effects in vitro and in vivo (14-17). After
intake, both mucosal and bacterial esterases are supposed to
release fiber-bound hydroxycinnamates such as DFA not only
from low molecular weight (model) compounds but also from
cereal brans into the gut lumen, which then are potentially
absorbed (18-20). However, there are no reports on the
absorption of DFA in humans, indicating that the bulk of dimers
is further metabolized by intestinal microbiota or excreted in
feces.
Plant cell wall associated dehydrodiferulic acids (DFA) are
abundant components of insoluble cereal dietary fiber. For
example, an intake of 10 g of wheat bran provides approximately
10 mg of DFA (1). The highest amount of DFA within cereal
grains studied was found in the insoluble fiber fraction of maize
amounting to 12.6 mg of DFA/g (2). Next to their occurrence
in monocotyledons, DFA are also cell-wall components of
dicotyledons, such as carrot and sugar beet (3, 4). The ferulate
dehydrodimers are formed via radical coupling of two mono-
meric ferulates (5) that are ester-linked to polymers in the plant
cell wall (6). Thus, by cross-linking polysaccharides to each
other or to lignin-like polymers, ferulate dimerization provides
stability to the cell wall but may inhibit intestinal fiber
degradability by host and microbial enzymes (7, 8).
In contrast to microbial hydrolysis of the ester bonds,
knowledge of the potential breakdown and further metabolism
of DFA by intestinal bacteria is very limited. The rates of
conversion of differently coupled DFA model compounds by
rumen and human intestinal microbiota in vitro were found to
be low (21). So far, microbial metabolites resulting from DFA
conversion have not been described.
The consumption of high-fiber diets has been implicated in
the prevention of colorectal cancer (9, 10), but this issue remains
controversial (11). In particular, the intake of insoluble dietary
fiber was reported to be inversely associated with colon cancer
* Author to whom correspondence should be addressed (telephone
+49 33200 88402; fax +49 33200 88407; e-mail braune@dife.de).
† A.B. and M.B. contributed equally to this work.
§ German Institute of Human Nutrition Potsdam-Rehbruecke.
# University of Minnesota.
In this study, we demonstrate the ability of human gut
microbiota to transform two selected DFA regioisomers, 5-5-
DFA and 8-O-4-DFA, and identify their microbial meta-
bolites.
⊥ University of Hamburg.
10.1021/jf900159h CCC: $40.75  2009 American Chemical Society
Published on Web 03/10/2009

Microbial Conversion of Dehydrodiferulic Acids
J. Agric. Food Chem., Vol. 57, No. 8, 2009 3357
MATERIALS AND METHODS
thermostated column compartment TCC-100 with eluent preconditioner,
a photodiode array detector UVD 340U, and a reversed-phase C18
column (LiChroCART 250-4 LiChrospher 100 RP-18, 5 µm; 250 × 4
mm i.d.; Merck, Darmstadt, Germany). The column temperature was
maintained at 50 °C. Aqueous 0.1% trifluoroacetic acid (v/v, solvent
system A) and methanol (solvent system B) served as the mobile phase
in a gradient mode with a flow rate of 1 mL/min. Detection was at 280
nm. For analysis of 8-O-4-DFA, 3-(4-hydroxy-3-methoxyphenyl)pyru-
vic acid, and their metabolites, the following gradient was applied: B
from 5 to 30% in 20 min, from 30 to 50% in 5 min, from 50 to 80%
in 10 min, and from 80 to 100% in 4 min. For analysis of 5-5-DFA
and its metabolites, the following gradient was used: B from 5 to 52.5%
in 70 min and from 52.5 to 100% in 5 min. If applicable, compounds
were identified by their retention times and UV spectra (λ ) 200-400
nm) in comparison to reference substances. External calibration curves
were used for quantification.
Sample Preparation for GC-MS Analysis. To gather additional
structural information, aliquots of some samples from the conversion
experiments were also investigated by GC-MS. For this purpose, sample
aliquots were centrifuged (12000g, 5 min), and the supernatant (200
µL) was lyophilized (Alpha 2-4, Christ, Osterode, Germany). The
residues were redissolved in water, acidified to pH 2.0, and extracted
twice with ethyl acetate. The combined organic layers were dried in a
Chemicals. 8-O-4-DFA and 5-5-DFA were preparatively isolated
from maize bran by gel permeation chromatography following saponi-
fication (22). Dihydroferulic acid-5-5-dihydroferulic acid and dihydro-
ferulic acid were synthesized by catalytic hydrogenation as briefly
described: Ferulic acid or 5-5-DFA was dissolved in methanol.
Palladium on activated carbon was carefully added. A balloon filled
with hydrogen was cautiously capped on the flask, and the mixture
was stirred at room temperature overnight. The catalyst was carefully
filtered off, and the product solution was dried. Hydrogenation was
verified by 1H and 13C NMR (360 MHz Bruker DRX-360 instrument).
3-(4-Hydroxy-3-methoxyphenyl)pyruvic acid was obtained from TCI
Europe (Zwijndrecht, Belgium). Ferulic acid, caffeic acid, homovanillic
acid, 3-(3,4-dihydroxyphenyl)propionic acid, and 3,4-dihydroxyphe-
nylacetic acid were purchased from Fluka (Deisenhofen, Germany),
and 3-(3-hydroxyphenyl)propionic acid was obtained from Alfa Aesar
(Karlsruhe, Germany).
Conversion Experiments with Human Fecal Slurries. Fecal
samples were collected from seven healthy human volunteers (three
males and four females) with a mean age of 34 years (range of 25-52
years), who consumed an unspecified Western diet and had not taken
antibiotics in the 6 months prior to the study. Freshly voided feces
were kept under anoxic conditions using AnaeroGen sachets (Oxoid,
Hampshire, U.K.) and stored at 4 °C for a maximum of 1 h before
processing. Fecal suspensions were prepared at room temperature in
an anaerobic glovebox (MACS Anaerobic Workstation, Don Whitley
Scientific Ltd., Shipley, U.K.) with a gas atmosphere of N2/CO2/H2
(80:10:10%, v/v/v) using a defined bicarbonate-buffered medium
(medium B) (23) supplemented with 10 mM glucose. An aliquot of
1 g of feces was transferred into a 16 mL tube fitted with a butyl rubber
stopper and a screw cap (Hungate tube) containing 10 mL of medium
B and subsequently homogenized by vigorous shaking.
For conversion experiments, aliquots of 500 µL of the fecal
suspension were used to inoculate 16 mL Hungate tubes gassed with
N2/CO2 (80:20%, v/v) and containing each 10 mL of medium B
supplemented with 8-O-4-DFA, 5-5-DFA, or 3-(4-hydroxy-3-methox-
yphenyl)pyruvic acid at a final concentration of 190 µM. DFA and
3-(4-hydroxy-3-methoxyphenyl)pyruvic acid were dissolved in dimethyl
sulfoxide (DMSO), and 50 µL of the respective stock solution was
added to the medium with a syringe. For control, the compounds were
incubated without fecal inoculum (addition of 500 µL of medium B).
In addition, the fecal slurries were incubated in the absence of the
compounds (addition of 50 µL of DMSO). The tubes were incubated
at 37 °C in a culture tube rotator at 6 rpm (SC1, Stuart Scientific,
Redhill, U.K.). At the times indicated, 500 µL samples were withdrawn
by using a syringe and stored at -20 °C. The assigned microbial
metabolites were exclusively detected following the incubation of 8-O-
4-DFA, 5-5-DFA, and 3-(3-hydroxy-4-methoxyphenyl)pyruvic acid,
respectively, with the fecal slurries. None of these metabolites was
observed in the absence of either the compounds or microbiota.
Sample Preparation for HPLC Analysis. Aliquots of samples from
the conversion experiments with 8-O-4-DFA and 3-(4-hydroxy-3-
methoxyphenyl)pyruvic acid were centrifuged (12000g, 5 min), and
112.5 µL of the supernatant was acidified by adding 2.5 µL of aqueous
4 M HCl. After centrifugation (12000g, 5 min), 102 µL of the resulting
supernatant was applied to the HPLC system. Whereas the recoveries
of 8-O-4-DFA, 3-(3-hydroxy-4-methoxyphenyl)pyruvic acid, and their
metabolites were maximal when the acidified supernatants were
analyzed directly, 5-5-DFA and its metabolites were recovered most
efficiently by extraction with ethyl acetate. For analysis of samples
from the conversion experiments with 5-5-DFA, aliquots of 200 µL
were extracted twice with 400 µL each of ethyl acetate after the addition
of 5 µL of aqueous 2 M HCl. The combined organic layers were dried
by vacuum centrifugation (RC 10.22; Jouan, Saint-Nazaire, France),
and the residue was dissolved in 40 µL of 70% aqueous methanol
(v/v). After centrifugation (12000g, 5 min), 20 µL of the resulting
supernatant was applied to the HPLC system.
reacti-vial under
derivatization.
a gentle stream of nitrogen and used for
Isolation of Metabolites for GC-MS Analysis. Some metabolites
could not be identified by HPLC analysis and comparison with standard
reference compounds. To identify these metabolites, additional GC-
MS experiments were necessary. Unidentified metabolites were isolated
from samples of the conversion experiments. Samples of 102 µL each
prepared as described above were run on the HPLC system as described
above. Fractions containing the metabolites of interest were manually
collected, pooled, and dried by vacuum centrifugation. Samples were
transferred into reacti-vials by sequentially using ethyl acetate and
acetone. After drying, the samples were derivatized as described
below.
GC-MS Analysis. Samples were silylated by using bis(trimethyl-
silyl)trifluoroacetamide (BSTFA) (40 µL) and pyridine (10 µL);
however, in some cases the amounts of reagents were equally reduced.
Silylation was carried out for 30 min at 60 °C in sealed reacti-vials.
Over the time of the study, two different GC-MS systems were used.
System 1 used a Trace 2000 ThermoQuest GC coupled to a PolarisQ
ion trap (Thermo Scientific, Dreieich, Germany). Samples were injected
in the splitless mode (injector temperature ) 300 °C). Separation was
carried out on a 30 m × 0.32 mm i.d., 0.25 µm film, HP-5-MS fused
silica capillary column (Hewlett-Packard, Waldbronn, Germany).
Compounds were characterized by their electron impact mass data (70
eV, scan range m/z 50-800). He (1 mL/min) was used as a carrier
gas. GC conditions were as follows: initial column temperature, 150 °C,
ramped at 3 °C/min to 250 °C, ramped at 30 °C/min or at 4 °C/min to
300 °C, held for 30 min.
System 2 used a 2010 GC coupled to a QP2010 quadrupole mass
spectrometer (Shimadzu, Columbia, MD). Samples were injected using
a split ratio of 10 (injector temperature ) 250 °C). Separation was
carried out on a 30 m × 0.25 mm i.d., 0.25 µm film, SHR5XLB
capillary column (Shimadzu). Compounds were characterized by their
electron impact mass data (70 eV, scan range m/z 45-900). He (0.85
mL/min) was used as carrier gas. GC conditions were as follows: initial
column temperature, 150 °C, held for 1 min, ramped at 10 °C/min to
310 °C, held for 25 min.
RESULTS
Conversion of 8-O-4-DFA by Human Intestinal Micro-
biota. Fecal suspensions of seven human donors were tested
for their ability to transform 8-O-4-DFA. Whereas the compound
was stable during incubation in medium without an inoculum,
8-O-4-DFA was completely converted by six of seven fecal
slurries within 15-40 h of incubation (Table 1). After 35 h,
no major changes in the metabolite patterns were observed for
most of the fecal suspensions. The complete conversion of 8-O-
HPLC Analysis. DFA and aromatic metabolites were determined
using an HPLC system with a photodiode array detector (Summit HPLC
Analytical System, Dionex, Idstein, Germany). The HPLC system was
equipped with a pump P680A LPG-4, an autosampler ASI-100T, a

3358 J. Agric. Food Chem., Vol. 57, No. 8, 2009
Braune et al.
Table 1. Conversion of 8-O-4-Dehydrodiferulic Acid (8-O-4-DFA),
5-5-Dehydrodiferulic Acid (5-5-DFA), and
3-(4-Hydroxy-3-methoxyphenyl)pyruvic Acid (HMPP) by Human Fecal
Suspensions (FS) within 164 h of Anoxic Incubation
period of complete conversion (h)
FS
8-O-4-DFA
5-5-DFA
HMPP
1
2
3
4
5
6
7
15
35
25
40
25
nc
25
nca
nc
nc
nc
40
nc
nc
ndb
48
48
nd
nd
nd
nd
a Not completed. b Not determined.
4-DFA [recovery at 0 h (( SD), 159 ( 2 µM] as performed by
the six fecal slurries resulted in cleavage of the dimer and
formation of a similar pattern of phenolic acids, which were
identified by comparison to standard reference compounds using
HPLC-DAD and when necessary by GC-MS. Ferulic acid,
dihydroferulic acid, 3-(3,4-dihydroxyphenyl)propionic acid,
homovanillic acid, and 3,4-dihydroxyphenylacetic acid were
detected in all of the incubations, whereas 3-(4-hydroxy-3-
methoxyphenyl)pyruvic acid, caffeic acid, and 3-(3-hydrox-
yphenyl)propionic acid were each observed in incubations with
only three to four of the fecal slurries (for chemical structures
see Figure 2). The main metabolites occurring at maximal
concentrations of >80 µM in the course of 8-O-4-DFA
conversion by the majority of fecal slurries were 3-(3,4-
dihydroxyphenyl)propionic acid, homovanillic acid, and 3,4-
dihydroxyphenylacetic acid. These metabolites, together with
3-(3-hydroxyphenyl)propionic acid, also represented the final
products of 8-O-4-DFA conversion under the conditions used.
Major differences between incubations with fecal slurries of
individual donors were observed with regard to metabolite
kinetics. The time course of 8-O-4-DFA conversion by a
representative fecal slurry forming the complete spectrum of
metabolites (FS 5) is depicted in Figure 1A.
The fecal microbiota of one of seven human subjects only
partially transformed 8-O-4-DFA. After 58 h of incubation with
this fecal suspension (FS 6), approximately half of the 8-O-4-
DFA was converted and no further transformation was observed
(Figure 1B). From 8-O-4-DFA was formed one main product
showing a spectrum [λmax 218, 238, 290(S), 327 nm] nearly
identical to that of the parental compound [λmax 218, 235, 290(S),
326 nm] but eluting earlier using RP-HPLC. Further charac-
terization of the metabolite by GC-MS indicated the loss of a
methyl group. The characteristic ions of the parental, silylated
8-O-4-DFA with m/z 602 (molecular ion), 484 (M - TMSOH
- CO), and 394 (M - 2 TMSOH - CO) were shifted toward
the ions with m/z 660 (molecular ion), 542, and 452. The mass
difference of 58 is explained by the loss of the methyl group,
creating another hydroxyl group, which is silylated. In the
parental 8-O-4-DFA two methoxy groups are available to be
theoretically demethylated. We hypothesize that due to sterical
hindrance, demethylation did not preferentially occur at the C-4-
coupled ferulate monomer but occurred at the methoxy group
of the C-8-coupled ferulate monomer, yielding caffeic acid-8-
O-4-ferulic acid. However, this question needs to be addressed
in more detail in the future, for example, by semipreparative
isolation and NMR characterization of this compound.
Figure 1. (A) Time course of degradation of 8-O-4-dehydrodiferulic acid
(8-O-4-DFA) by a human fecal suspension (FS 5). From 8-O-4-DFA (b)
were formed 3-(4-hydroxy-3-methoxyphenyl)pyruvic acid (· · ·), ferulic acid
(4), homovanillic acid (1), dihydroferulic acid (O), caffeic acid (]), 3-(3,4-
dihydroxyphenyl)propionic acid (9), 3-(3-hydroxyphenyl)propionic acid ([),
and 3,4-dihydroxyphenylacetic acid (2). For the 8-O-4-DFA control without
fecal inoculum see panel B. (B) Time course of partial conversion of
8-O-4-DFA by a human fecal suspension (FS 6). From 8-O-4-DFA (b)
was formed the proposed demethylated metabolite, caffeic acid-8-O-4-
ferulic acid (2) (y axis on the right). 8-O-4-DFA without fecal inoculum
(O) served as a control.
cleaved, resulting in the formation of ferulic acid and likely
3-(4-hydroxy-3-methoxyphenyl)pyruvic acid. Both metabolites
were further converted by decarboxylation, reduction, O-
demethylation, and dehydroxylation, respectively. From 3-(4-
hydroxy-3-methoxyphenyl)pyruvic acid was formed 3,4-dihy-
droxyphenylacetic acid via homovanillic acid. Ferulic acid was
converted mainly to dihydroferulic acid but also to caffeic acid.
From the latter two compounds was formed 3-(3,4-dihydrox-
yphenyl)propionic acid, which was further dehydroxylated to
3-(3-hydroxyphenyl)propionic acid. The incomplete conversion
of homovanillic acid and dihydroferulic acid to 3,4-dihydrox-
yphenylacetic acid and 3-(3,4-dihydroxyphenyl)propionic acid,
respectively, by some of the individual fecal slurries indicates
that the O-demethylating activity of bacteria was limited.
However, besides the monomeric phenolic acids, also the
dimeric 8-O-4-DFA was a substrate for microbial O-demethy-
lation, resulting in the formation of an 8-O-4-coupled caffeic
acid/ferulic acid dimer. As observed for demethylation, dehy-
droxylation of 3-(3,4-dihydroxyphenyl)propionic acid was cata-
lyzed by only four of the fecal suspensions.
On the basis of the metabolites detected, the following
pathways of 8-O-4-DFA conversion by human fecal microbiota
are proposed (Figure 2). The ether bond of 8-O-4-DFA was

Microbial Conversion of Dehydrodiferulic Acids
J. Agric. Food Chem., Vol. 57, No. 8, 2009 3359
Figure 2. Proposed pathways of 8-O-4-dehydrodiferulic acid (8-O-4-DFA) conversion by human intestinal microbiota. Pathways indicated by dashed
arrows could not be verified in this study.
The ability of human gut bacteria to convert 3-(4-hydroxy-
3-methoxyphenyl)pyruvic acid, the proposed initial product of
8-O-4-DFA cleavage besides ferulic acid, was tested. 3-(4-
Hydroxy-3-methoxyphenyl)pyruvic acid was incubated with
fecal suspensions of two donors (FS 2, FS 3; Table 1), which
in previous experiments showed a complete conversion of 8-O-
4-DFA. 3-(4-Hydroxy-3-methoxyphenyl)pyruvic acid (190 µM)
was completely converted within 5 h of incubation (data not
shown). The end product was identified as 3,4-dihydroxyphe-
nylacetic acid. Two intermediates, M1 (λmax 227, 280 nm) and
M2 [λmax 229(S), 281 nm], were formed consecutively. Both
metabolites could not be identified by comparison to standard
reference compounds. Therefore, the compounds were isolated
from supernatants of incubation experiments and subjected to
GC-MS analysis. On the basis of the mass spectra of the
silylated metabolites, M1 and M2 were tentatively identified as
3-(4-hydroxy-3-methoxyphenyl)lactic acid and 3-(3,4-dihydrox-
yphenyl)lactic acid, respectively. Metabolite M1 showed a
molecular ion with m/z 428 (6% of the base peak) and
characteristic fragments of m/z 338 (M - TMSOH) and 209
(base peak). A similar fragmentation pattern was described for
3-(4-hydroxyphenyl)lactic acid in the literature. However, due
to the lack of an aromatic methoxy group the molecular ion is
shifted to m/z 398, and the characteristic fragments are shifted
to m/z 308 and 179 (base peak) (24). Metabolite M2 had a
molecular ion with m/z 486 (8% of the base peak), and
characteristic fragments were noted at m/z 396 (M - TMSOH)
and 267 (base peak). From this mass spectrum it can be deduced
that the aromatic methoxy group was demethylated, creating a
new hydroxyl group to be silylated during the derivatization
process. Following their assignment as metabolites of the
conversion of 3-(4-hydroxy-3-methoxyphenyl)pyruvic acid
(Figure 2), 3-(4-hydroxy-3-methoxyphenyl)lactic acid and
3-(3,4-dihydroxyphenyl)lactic acid were also detected in several
samples taken during initial 8-O-4-DFA transformation by some
of the fecal slurries. However, homovanillic acid was not
observed in the course of the conversion of 3-(4-hydroxy-3-
methoxyphenyl)pyruvic acid. Homovanillic acid had been
identified as an intermediate of 8-O-4-DFA transformation and
proposed to be a metabolite resulting from 3-(4-hydroxy-3-
methoxyphenyl)pyruvic acid (Figure 2).

3360 J. Agric. Food Chem., Vol. 57, No. 8, 2009
Braune et al.
transformed. The conversion of 5-5-DFA essentially occurred
within 40 h of incubation. Two metabolites were formed from
5-5-DFA at various concentrations depending on the fecal
suspension. One of these compounds, ferulic acid-5-5-dihydro-
ferulic acid, had already been observed as an intermediate of
5-5-DFA conversion by the fecal suspension FS 5. The second
metabolite was identified as dihydroferulic acid-5-5-dihydrof-
erulic acid by comparison to a standard reference compound.
The transformation of 5-5-DFA by the fecal suspension forming
the highest amounts of both metabolites (FS 4) is shown in
Figure 3B. On the basis of the metabolites detected following
incubation with the individual human fecal slurries, Figure 4
summarizes the pathways of microbial 5-5-DFA conversion. The
ferulate dimer was not cleaved by human intestinal microbiota.
Only the side-chain double bonds were reduced, and methoxy
groups were partially demethylated.
DISCUSSION
The fermentation of dietary fiber by gut bacteria is proposed
to have beneficial effects on human health (25). The fiber
degradability may be affected by cross-linking polysaccharide
chains by ferulate dimerization. In vitro analyses of carbohydrate
fermentation products show that cross-linking via ferulate
derivatives reduces the rate of arabinoxylan degradation (8).
On the other hand, low to moderate levels of diferulates
(0.8-2.6 mg/g) do not alter the release of carbohydrates or the
overall degradation of nonlignified cell walls by human gut
microbiota (26). The ester bonds between the polysaccharides
and the (di)ferulates may be cleaved by both intestinal and
microbial enzymes (19, 20, 27). Microbial feruloyl esterases
have been characterized from several species, including the gut-
relevant Lactobacillus acidophilus (28). In contrast, data on both
the cleavage of fiber-bound diferulates and the conversion of
the free dimers as released by esterases in the gut are still
limited. In the present study, the ability of the human intestinal
microbiota to convert two regioisomeric DFA, 8-O-4-DFA and
5-5-DFA, was elucidated. These dimers are among the dominat-
ing diferulates of cereal grains. For example, in insoluble dietary
fiber from maize (total DFA content ) 12.6 mg/g) 8-O-4-DFA
and 5-5-DFA contribute 21 and 25% to the total diferulates (2).
The 8-O-4-coupled DFA was completely transformed by
human fecal suspensions within 15-40 h. This time period
corresponds to the average colonic transit time in healthy
humans of 36 h (29). The observed metabolites indicate that
the first step of 8-O-4-DFA conversion was the cleavage of the
ether bond (Figure 2). Ether linkages in both biological and
xenobiotic compounds show a high degree of resistance against
enzymatic degradation. Although only a few ether-cleaving
enzymes have been described to date, these enzymes exhibit a
variety of mechanisms (30). We assume that 8-O-4-DFA was
cleaved analogously to the hydrolysis of vinyl ethers catalyzed
by the isochorismate pyruvate hydrolase (isochorismatase, EC
3.3.2.1), which is found in several bacteria including Escherichia
coli (31). According to the reaction mechanism described for
the isochorismatase (30, 32), the cleavage of 8-O-4-DFA may
take place by hydration of the double bond to yield the
hemiacetal, cleavage via ꢀ-elimination releasing ferulic acid and
3-(4-hydroxy-3-methoxyphenyl)pyruvic acid, the latter formed
by tautomerization. The transformation of ferulic acid by
intestinal microbiota has been described in the past (21, 33).
As reported previously, hydrogenation of the side chain
proceeded more quickly than O-demethylation and dehydroxy-
lation. On the basis of kinetic data, homovanillic acid and 3,4-
dihydroxyphenylacetic acid were most likely formed from 3-(4-
Figure 3. (A) Time course of conversion of 5-5-dehydrodiferulic acid (5-
5-DFA) by a human fecal suspension (FS 5). From 5-5-DFA (b) were
formed ferulic acid-5-5-dihydroferulic acid (· · ·), ferulic acid-5-5-caffeic
acid ([), caffeic acid-5-5-caffeic acid (9), and caffeic acid-5-5-dihydro-
caffeic acid (2). Broken lines refer to the y axis on the right. For the
5-5-DFA control without fecal inoculum see panel B. (B) Time course of
partial conversion of 5-5-DFA by a human fecal suspension (FS 4). From
5-5-DFA (b) were formed ferulic acid-5-5-dihydroferulic acid (2) (y axis
on the right) and dihydroferulic acid-5-5-dihydroferulic acid (9). 5-5-DFA
without fecal inoculum (O) served as a control.
Conversion of 5-5-DFA by Human Intestinal Microbiota.
The fecal suspensions used for conversion experiments with
8-O-4-DFA were also tested for their ability to transform 5-5-
DFA (Table 1). The compound was stable during the total
incubation period of 164 h in medium without an inoculum.
Only one of seven different fecal suspensions catalyzed the
complete conversion of 5-5-DFA [recovery at 0 h ((SD), 210
( 10 µM] within 40 h of incubation (Figure 3A). In the course
of 5-5-DFA conversion by this fecal suspension (FS 5), four
main metabolites were detected (Figure 3A). Following isola-
tion, these metabolites were characterized by GC-MS analysis
as ferulic acid-5-5-dihydroferulic acid (molecular ion and base
peak m/z 676), ferulic acid-5-5-caffeic acid (molecular ion and
base peak m/z 732), caffeic acid-5-5-caffeic acid (molecular ion
and base peak m/z 790), and caffeic acid-5-5-dihydrocaffeic acid
(molecular ion and base peak m/z 792). Whereas ferulic acid-
5-5-dihydroferulic acid and ferulic acid-5-5-caffeic acid were
formed transiently, caffeic acid-5-5-caffeic acid and caffeic acid-
5-5-dihydrocaffeic acid represented the end products of micro-
bial 5-5-DFA conversion. After 70 h of incubation, no major
changes in the concentration of the two end products were
observed.
In the course of incubation for 164 h with the other six fecal
suspensions, 10-60% of the initially available 5-5-DFA was

Microbial Conversion of Dehydrodiferulic Acids
J. Agric. Food Chem., Vol. 57, No. 8, 2009 3361
Figure 4. Proposed pathways of 5-5-dehydrodiferulic acid (5-5-DFA) conversion by human intestinal microbiota.
hydroxy-3-methoxyphenyl)pyruvic acid. However, when incubated
with fecal suspensions, 3-(4-hydroxy-3-methoxyphenyl)pyruvic
acid was converted to 3,4-dihydroxyphenylacetic acid via 3-(4-
hydroxy-3-methoxyphenyl)lactic acid and 3-(3,4-dihydroxyphe-
nyl)lactic acid (Figure 2), but we were not able to detect
homovanillic acid. Therefore, we cannot exclude a concurrent
mechanism in which homovanillic acid is formed from the
parental 8-O-4-DFA or from one of the primary cleavage
products described here.
the depolymerization of dietary fiber in the gut by cleaving the
existing 8-O-4-DFA cross-links prior to ester hydrolysis, thereby
increasing the fiber fermentability.
ABBREVIATIONS USED
BSTFA, bis(trimethylsilyl)trifluoroacetamide; DFA, dehy-
drodiferulic acid(s); 5-5-DFA, 5-5-dehydrodiferulic acid; 8-O-
4-DFA, 8-O-4-dehydrodiferulic acid; FS, fecal suspension; TMS,
trimethylsilyl.
In contrast to 8-O-4-DFA, the 5-5-coupled DFA was not
cleaved into monomers by any of the human fecal suspensions
tested. Only the side chains of 5-5-DFA were finally hydroge-
nated, yielding dihydroferulic acid-5-5-dihydroferulic acid (Fig-
ure 4). Demethylation of the methoxy groups was less
frequently observed. However, methoxy-substituted phenolic
acids formed in the course of 8-O-4-DFA conversion were
readily demethylated by the same microbiota. The conversion
of 5-5-DFA and other synthesized diphenolic model compounds
by rumen and human intestinal microbiota was tested in a
previous study (21). Following incubation of the differently
coupled diphenolic compounds with rumen or fecal slurries in
buffer, some of the investigated compounds including the 5-5-
DFA were mostly recovered intact. The use of a culture medium
containing glucose, vitamins, and trace elements may have led
to a more efficient transformation of 5-5-DFA as observed in
our study.
The results obtained in this study indicate that human
intestinal microbiota may structure-dependently transform DFA,
which are released from fiber carbohydrates by esterases into
the gut lumen. This in turn may affect the absorption of DFA,
which has been demonstrated for several of these dimers to occur
in the rat intestine (19). However, no absorption of unaltered
DFA was observed in humans after high-bran cereal consump-
tion, which indicates that the bulk of free or still ester-linked
dimers remains in the intestinal content and is available for
metabolization by gut bacteria (34). For both the DFA and the
phenolic acids, which were formed from 8-O-4-DFA, such as
ferulic acid, caffeic acid, and 3,4-dihydroxyphenylacetic acid,
antioxidant and chemoprotective effects have been reported
(14-16, 35, 36). Because the phenolic acids are presumably
better absorbed, for example, by the monocarboxylic acid
transporter (37), the microbial conversion of DFA might result
in bioactivation of these compounds. Besides cleavage of the
free 8-O-4-DFA, intestinal microbiota could also contribute to
ACKNOWLEDGMENT
We thank Sabine Zimmermann and Hoon Kim for technical
assistance.
LITERATURE CITED
(1) Garcia-Conesa, M. T.; Plumb, G. W.; Waldron, K. W.; Ralph, J.;
Williamson, G. Ferulic acid dehydrodimers from wheat bran:
isolation, purification and antioxidant properties of 8-O-4-diferulic
acid. Redox Rep. 1997, 3, 319–323.
(2) Bunzel, M.; Ralph, J.; Marita, J. M.; Hatfield, R. D.; Steinhart,
H. Diferulates as structural components in soluble and insoluble
cereal dietary fibre. J. Sci. Food Agric. 2001, 81, 653–660.
(3) Parr, A. J.; Ng, A.; Waldron, K. W. Ester-linked phenolic
components of carrot cell walls. J. Agric. Food Chem. 1997, 45,
2468–2471.
(4) Ralet, M. C.; Andre-Leroux, G.; Quemener, B.; Thibault, J. F.
Sugar beet (Beta Vulgaris) pectins are covalently cross-linked
through diferulic bridges in the cell wall. Phytochemistry 2005,
66, 2800–2814.
(5) Ralph, J.; Quideau, S.; Grabber, J. H.; Hatfield, R. D. Identification
and synthesis of new ferulic acid dehydrodimers present in grass
cell walls. J. Chem. Soc., Perkin Trans. 1 1994, 3485–3498.
(6) Ishii, T. Structure and functions of feruloylated polysaccharides.
Plant Sci. 1997, 127, 111–127.
(7) Grabber, J. H.; Hatfield, R. D.; Ralph, J. Diferulate cross-links
impede the enzymatic degradation of non-lignified maize walls.
J. Sci. Food Agric. 1998, 77, 193–200.
(8) Hopkins, M. J.; Englyst, H. N.; Macfarlane, S.; Furrie, E.;
Macfarlane, G. T.; McBain, A. J. Degradation of cross-linked and
non-cross-linked arabinoxylans by the intestinal microbiota in
children. Appl. EnViron. Microbiol. 2003, 69, 6354–6360.
(9) Bingham, S. A.; Day, N. E.; Luben, R.; Ferrari, P.; Slimani, N.;
Norat, T.; Clavel-Chapelon, F.; Kesse, E.; Nieters, A.; Boeing,
H.; et al. Dietary fibre in food and protection against colorectal
cancer in the European Prospective Investigation into Cancer and

3362 J. Agric. Food Chem., Vol. 57, No. 8, 2009
Braune et al.
Nutrition (EPIC): an observational study. Lancet 2003, 361, 1496–
1501.
(25) Rose, D. J.; DeMeo, M. T.; Keshavarzian, A.; Hamaker, B. R.
Influence of dietary fiber on inflammatory bowel disease and colon
cancer: importance of fermentation pattern. Nutr. ReV. 2007, 65,
51–62.
(26) Funk, C.; Braune, A.; Grabber, J. H.; Steinhart, H.; Bunzel, M.
Moderate ferulate and diferulate levels do not impede maize cell
wall degradation by human intestinal microbiota. J. Agric. Food
Chem. 2007, 55, 2418–2423.
(27) Wende, G.; Buchanan, C. J.; Fry, S. C. Hydrolysis and fermenta-
tion by rat gut microorganisms of 2-O-ꢀ-D-xylopyranosyl(5-O-
feruloyl)-L-arabinose derived from grass cell wall arabinoxylan.
J. Sci. Food Agric. 1997, 73, 296–300.
(28) Wang, X.; Geng, X.; Egashira, Y.; Sanada, H. Purification and
characterization of a feruloyl esterase from the intestinal bacterium
Lactobacillus acidophilus. Appl. EnViron. Microbiol. 2004, 70,
2367–2372.
(10) Peters, U.; Sinha, R.; Chatterjee, N.; Subar, A. F.; Ziegler, R. G.;
Kulldorff, M.; Bresalier, R.; Weissfeld, J. L.; Flood, A.; Schatzkin,
A.; Hayes, R. B. Dietary fibre and colorectal adenoma in a
colorectal cancer early detection programme. Lancet 2003, 361,
1491–1495.
(11) Tan, K. Y.; Seow-Choen, F. Fiber and colorectal diseases:
separating fact from fiction. World J. Gastroenterol 2007, 13,
4161–4167.
(12) Wakai, K.; Hirose, K.; Matsuo, K.; Ito, H.; Kuriki, K.; Suzuki,
T.; Kato, T.; Hirai, T.; Kanemitsu, Y.; Tajima, K. Dietary risk
factors for colon and rectal cancers: a comparative case-control
study. J. Epidemiol. 2006, 16, 125–135.
(13) Ferguson, L. R.; Harris, P. J. The dietary fibre debate: more food
for thought. Lancet 2003, 361, 1487–1488.
(14) Kikuzaki, H.; Hisamoto, M.; Hirose, K.; Akiyama, K.; Taniguchi,
H. Antioxidant properties of ferulic acid and its related com-
pounds. J. Agric. Food Chem. 2002, 50, 2161–2168.
(15) Andreasen, M. F.; Landbo, A. K.; Christensen, L. P.; Hansen,
A.; Meyer, A. S. Antioxidant effects of phenolic rye (Secale
cereale L.) extracts, monomeric hydroxycinnamates, and ferulic
acid dehydrodimers on human low-density lipoproteins. J. Agric.
Food Chem. 2001, 49, 4090–4096.
(16) Kawabata, K.; Yamamoto, T.; Hara, A.; Shimizu, M.; Yamada,
Y.; Matsunaga, K.; Tanaka, T.; Mori, H. Modifying effects of
ferulic acid on azoxymethane-induced colon carcinogenesis in
F344 rats. Cancer Lett. 2000, 157, 15–21.
(17) Drankhan, K.; Carter, J.; Madl, R.; Klopfenstein, C.; Padula, F.;
Lu, Y.; Warren, T.; Schmitz, N.; Takemoto, D. J. Antitumor
activity of wheats with high orthophenolic content. Nutr. Cancer
2003, 47, 188–194.
(18) Poquet, L.; Clifford, M. N.; Williamson, G. Transport and
metabolism of ferulic acid through the colonic epithelium. Drug
Metab. Dispos. 2008, 36, 190–197.
(29) Meier, R.; Beglinger, C.; Dederding, J. P.; Meyer-Wyss, B.;
Fumagalli, M.; Rowedder, A.; Turberg, Y.; Brignoli, R. Influence
of age, gender, hormonal status and smoking habits on colonic
transit time. Neurogastroenterol. Motil. 1995, 7, 235–238.
(30) White, G. F.; Russell, N. J.; Tidswell, E. C. Bacterial scission of
ether bonds. Microbiol. ReV. 1996, 60, 216–232.
(31) Rusnak, F.; Liu, J.; Quinn, N.; Berchtold, G. A.; Walsh, C. T.
Subcloning of the enterobactin biosynthetic gene entB: expression,
purification, characterization, and substrate specificity of isocho-
rismatase. Biochemistry 1990, 29, 1425–1435.
(32) Parsons, J. F.; Calabrese, K.; Eisenstein, E.; Ladner, J. E. Structure
and mechanism of Pseudomonas aeruginosa PhzD, an isocho-
rismatase from the phenazine biosynthetic pathway. Biochemistry
2003, 42, 5684–5693.
(33) Rosazza, J. P. N.; Huang, Z.; Dostal, L.; Volm, T.; Rousseau, B.
Review: Biocatalytic transformations of ferulic acid: an abundant
aromatic natural product. J. Ind. Microbiol. 1995, 15, 457–471.
(34) Kern, S. M.; Bennett, R. N.; Mellon, F. A.; Kroon, P. A.; Garcia-
Conesa, M. T. Absorption of hydroxycinnamates in humans after
high-bran cereal consumption. J. Agric. Food Chem. 2003, 51,
6050–6055.
(19) Andreasen, M. F.; Kroon, P. A.; Williamson, G.; Garcia-Conesa,
M. T. Intestinal release and uptake of phenolic antioxidant diferulic
acids. Free Radical Biol. Med. 2001, 31, 304–314.
(35) Raneva, V.; Shimasaki, H.; Ishida, Y.; Ueta, N.; Niki, E.
Antioxidative activity of 3,4-dihydroxyphenylacetic acid and
caffeic acid in rat plasma. Lipids 2001, 36, 1111–1116.
(36) Gao, K.; Xu, A.; Krul, C.; Venema, K.; Liu, Y.; Niu, Y.; Lu, J.;
Bensoussan, L.; Seeram, N. P.; Heber, D.; Henning, S. M. Of the
major phenolic acids formed during human microbial fermentation
of tea, citrus, and soy flavonoid supplements, only 3,4-dihydrox-
yphenylacetic acid has antiproliferative activity. J. Nutr. 2006,
136, 52–57.
(37) Konishi, Y.; Kobayashi, S. Microbial metabolites of ingested
caffeic acid are absorbed by the monocarboxylic acid transporter
(MCT) in intestinal Caco-2 cell monolayers. J. Agric. Food Chem.
2004, 52, 6418–6424.
(20) Andreasen, M. F.; Kroon, P. A.; Williamson, G.; Garcia-Conesa,
M. T. Esterase activity able to hydrolyze dietary antioxidant
hydroxycinnamates is distributed along the intestine of mammals.
J. Agric. Food Chem. 2001, 49, 5679–5684.
(21) Chesson, A.; Provan, G. J.; Russell, W. R.; Scobbie, L.; Rich-
ardson, A. J.; Stewart, C. Hydroxycinnamic acids in the digestive
tract of livestock and humans. J. Sci. Food Agric. 1999, 79, 373–
378.
(22) Bunzel, M.; Funk, C.; Steinhart, H. Semipreparative isolation of
dehydrodiferulic and dehydrotriferulic acids as standard substances
from maize bran. J. Sep. Sci. 2004, 27, 1080–1086.
(23) Schneider, H.; Schwiertz, A.; Collins, M. D.; Blaut, M. Anaerobic
transformation of quercetin-3-glucoside by bacteria from the
human intestinal tract. Arch. Microbiol. 1999, 171, 81–91.
(24) Packert, M. Analytik und Bedeutung gebundener aromatischer
Carbonsa¨uren der Nahrungsfaser aus Getreide und anderen
Nutzpflanzen. Ph.D. Thesis, University of Hamburg, Germany,
1993.
Received for review January 15, 2009. Revised manuscript received
February 19, 2009. Accepted February 23, 2009.
JF900159H