Combined Acid Hydrolysis and Fermentation Improves Bioactivity of Citrus Flavonoids in Vitro and in Vivo

ARTICLE
https://doi.org/10.1038/s42003-023-05424-7 OPEN
Combined acid hydrolysis and fermentation

improves bioactivity of citrus flavonoids in vitro and
in vivo
Alice König 1,2,4, Nadiia Sadova 1,4, Marion Dornmayr1,2, Bettina Schwarzinger1,2, Cathrina Neuhauser1,
Verena Stadlbauer1,2, Melanie Wallner 1,2, Jakob Woischitzschläger1, Andreas Müller3, Rolf Tona3,
Daniel Kofel3 & Julian Weghuber 1,2 ✉
Many bioactive plant compounds, known as phytochemicals, have the potential to improve
1234567890():,;
health. Unfortunately, the bioavailability and bioactivity of phytochemicals such as poly-

phenolic flavonoids are reduced due to conjugation with sugar moieties. Here, we combine
acid hydrolysis and tailored fermentation by lactic acid bacteria (Lactiplantibacillus plantarum)
to convert the biologically less active flavonoid glycosides hesperidin and naringin into the
more active aglycones hesperetin and naringenin. Using a comprehensive approach, we
identify the most effective hydrolysis and fermentation conditions to increase the con-
centration of the aglycones in citrus extracts. The higher cellular transport and bioactivity of
the biotransformed citrus extract are also demonstrated in vitro and in vivo. Superior anti-
oxidant, anti-inflammatory and cell migration activities in vitro, as well as intestinal barrier
protecting and antioxidant activities in Drosophila melanogaster are identified. In conclusion,
the presented biotransformation approach improves the bioactivity of flavonoids, clearly
traced back to the increase in aglycone content.
1 Center of Excellence Food Technology and Nutrition, University of Applied Sciences Upper Austria, Stelzhamerstraße 23, Wels 4600, Austria. 2 FFoQSI
GmbH-Austrian Competence Centre for Feed and Food Quality, Safety and Innovation, Technopark 1D, Tulln 3430, Austria. 3 TriPlant AG, Industriestrasse 17,
Buetzberg 4922, Switzerland. 4These authors contributed equally: Alice König, Nadiia Sadova. ✉email: [email protected]
COMMUNICATIONS BIOLOGY | (2023)6:1083 | https://doi.org/10.1038/s42003-023-05424-7 | www.nature.com/commsbio 1

ARTICLE COMMUNICATIONS BIOLOGY | https://doi.org/10.1038/s42003-023-05424-7
T
he health-promoting effects of secondary plant com- in which molecules are degraded by the enzymatic activities of
pounds, known as phytochemicals, have been a major microorganisms. Lactic acid bacteria (LAB), especially Lactoba-
focus of plant research in recent years1,2. For instance, cillus species, are of particular interest here, as they catabolize
citrus flavonoids, polyphenolic phytochemicals found in plants of flavonoids in the human colon8. Many Lactobacillus species, such
the Rutaceae family such as sweet oranges (Citrus sinensis), bitter as Lactiplantibacillus plantarum (L. plantarum) are generally
oranges (Citrus aurantium) or grapefruits (Citrus paradisi) have recognized as safe by the United States Food and Drug
been of special interest3–5 due to their anti-inflammatory5–7 and Administration14 and included to the list of biological agents with
antioxidant4,5 properties. One subclass of flavonoids are flava- qualified presumption of safety by the European Food Safety
nones with naringin and hesperidin as the most abundant Authorities15, justifying their use in food and feed. Previous
representatives8. These flavanones naturally occur in their gly- studies have demonstrated that various Lactobacillus species can
cosidic form, in which the respective aglycones naringenin and cleave flavanones9,16,17, however, the reported conversion took
hesperetin are bound to a sugar moiety that is either a neohe- 10 days and the aglycone yield was moderate9.
speridose (2-O-L-rhamnosyl-D-glucose) or a rutinose (6-O-L- In the present study, we aimed to accelerate the conversion of
rhamnosyl-D-glucose)5,8,9. These sugar moieties prevent an effi- biologically less active flavonoid glycosides into more active
cient absorption in the small intestine8,9 and therefore reduce the aglycones using a combination of acid hydrolysis and LAB fer-
bioavailability and bioactivity of flavanones5,10,11. Intact flavo- mentation of citrus extracts. Our purpose was not to develop a
noids enter the colon, where they undergo extensive metabolism highly specific hydrolysis method, but to establish a bio-
by the intestinal microbiota8. transformation method that can be used in the food and feed
One strategy to improve bioavailability and bioactivity of fla- industry as an alternative to commercially available enzymes. In
vonoids is to increase aglycone levels by deglycosylation12,13. A addition, we wanted to determine whether the biotransformation
randomized, double-blind crossover trial with human subjects of flavonoids changes the bioactivity of the citrus extracts. For this
could show that enzymatic removal of the sugar moiety from purpose, we implemented suitable in vitro and in vivo models to
hesperidin in orange juice changes the site of absorption from study flavonoid transport and uptake, antioxidant and anti-
colon to the small intestine, thereby improving flavanone inflammatory properties, as well as intestinal barrier integrity.
bioavailability12. The classical approach for conversion of citrus A requirement for bioactivity is that the substances are available
flavanones requires the hydrolases α-L-rhamnosidase (EC to the cells. Thus, we first tested the uptake and transport of
3.2.1.40) and β-D-glucosidase (EC 3.2.1.21)9. As shown in Fig. 1, flavonoids from our extracts in the human intestinal epithelial cell
α-L-rhamnosidase first cleaves terminal L-rhamnose from narir- line Caco-2. Differentiated Caco-2 cells express tissue-typical cell
utin and naringin (1a) or hesperidin and neohesperidin (1b) and membrane and transport proteins and have been established as
converts them to naringenin-7-O-glucoside or hesperetin-7-O- in vitro model to study drug absorption18,19. Since oxidative
glucoside, respectively. Next, β-D-glucosidase splits off D-glucose stress is a major cause of barrier disruption and inflammation in
and releases the aglycones naringenin and hesperetin. A sus- the small intestine20, experiments on oxidative stress and cell
tainable and cost-efficient alternative to enzymes is fermentation, migration were performed in the intestinal cell lines, Caco-2 and
Fig. 1 Degradation of flavanones by α-L-rhamnosidase and β-D-glucosidase. a Narirutin (naringenin-7-O-rutinoside) or naringin (naringenin-7-O-
neohesperidoside) are converted to naringenin-7-O-glucoside by α-L-rhamnosidase. Next, β-D-glucosidase catalyzes the conversion of naringenin-7-O-
glucoside to naringenin. b Hesperidin (hesperetin-7-O-rutinoside) or neohesperidin (hesperetin-7-O-neohesperidoside) are hydrolyzed to hesperetin-7-O-
glucoside by α-L-rhamnosidase. Hesperetin-7-O-glucoside is converted to hesperetin by β-D-glucosidase.
2 COMMUNICATIONS BIOLOGY | (2023)6:1083 | https://doi.org/10.1038/s42003-023-05424-7 | www.nature.com/commsbio

COMMUNICATIONS BIOLOGY | https://doi.org/10.1038/s42003-023-05424-7 ARTICLE
IPEC-J2. The porcine IPEC-J2 intestinal epithelial cell line is a the flavanones naringin, neohesperidin, narirutin and hesperidin
valuable tool to study the effects of phytochemicals on intracel- were hardly metabolized during incubation with bacteria for
lular oxidative stress and wound healing20,21. Cell migration 144 h (6 d) (Supplementary Figure 3a). These findings led us to
assays are simplified in vitro models of wound healing and the assumption that all strains have β-glucosidase activity but no
require cells with high migration rates, such as IPEC-J2, to dif- noticeable α-rhamnosidase activity. Addition of 2% of glucose to
ferentiate between cell migration and cell division22. The anti- AQE improved bacterial growth (Supplementary Figure 3b) but
inflammatory activities of flavanones were determined in human reduced the β-glucosidase activity of the bacteria resulting in
monocytic THP-1 cells, a suitable model to study inflammatory significantly lower aglycone levels (Supplementary Figure 3c).
immune responses because differentiated THP-1 cells behave Thus, medium without glucose was used for fermentation of
similarly to native monocyte-derived macrophages22,23. To verify extracts.
whether deglycosylation of flavanones in biotransformed citrus Next, we enhanced the efficiency of glycolytic cleavage by
extracts protects the intestinal barrier in vivo, we used the fruit fly adding acid hydrolysis. First, the L-rhamnose group was split off
Drosophila melanogaster (D. melanogaster) model. The midgut of from naringin, neohesperidin, hesperidin and narirutin using
D. melanogaster is well comparable with the mammalian small citric acid, resulting in citric acid hydrolyzed extract (CAE). In a
intestine24–27 making the fruit fly a valuable model organism for next step, the resulting 7-O-glucosides, namely naringenin-7-O-
studying the effects of bioactive compounds on the intestinal glucoside and hesperetin-7-O-glucosides, were converted to
barrier. We also investigated the antioxidant properties of gly- aglycones by the bacterial β-glucosidase activity. A fermentation
cosylated flavanones and their aglycones in the fruit fly, as D. period of 24 h was chosen because a longer incubation time did
melanogaster is a well-established model in oxidative stress not show any improvement in our previous experiments. As
research28–30. presented in Fig. 2b, fermentation of CAE with bacteria resulted
Taken together, we demonstrated an efficient method of fla- in significantly higher aglycone levels compared to the control
vanone biotransformation via a combination of acid hydrolysis treatment incubated without bacteria. The maximum increases in
and 24 h-fermentation and provided a comprehensive approach naringenin (463 µM) and hesperetin (144 µM) above the initial
to study the bioactivity of these biotransformed flavanones. concentrations were achieved by fermenting CAE with L.
Superior antioxidant and anti-inflammatory activities were plantarum. Thus, this strain was used for preparation of the
observed for biotransformed citrus extracts with increased final extracts (Fig. 3a). As shown in Supplementary Figure 4, co-
hesperetin and naringenin content. cultivation of L. plantarum with different LAB strains did not
improve the aglycone yield significantly.
Next, we performed acid hydrolysis with different citric acid
Results concentrations to optimize the deglycosylation process. We first
Combination of acid hydrolysis and fermentation increases aimed to increase the content of 7-O-glucosides by acid
aglycone content in citrus extracts. First, we confirmed the hydrolysis so that they could be converted to aglycones by
degradation process described in Fig. 1 by incubating an aqueous bacterial β-glucosidase activity. As shown in Fig. 3b, citric acid
extract (AQE) made of dried citrus fruits containing all eight hydrolysis significantly enhanced the levels of 7-O-glucosides in a
flavanones of interest (naringin, neohesperidin, hesperidin, nar- concentration-dependent manner. The higher the acid concen-
irutin, naringenin-7-O-glucoside, hesperetin-7-O-glucoside, nar- tration, the more efficient the hydrolysis of naringin to
ingenin and hesperetin) with the enzymes α-rhamnosidase or β- naringenin-7-O-glucoside as well as neohesperidin to hesper-
glucosidase. Our results indicated that α-L-rhamnosidase from etin-7-O-glucoside was. Subsequent fermentation by L. plan-
prokaryotic source was able to hydrolyze the rutinosides narirutin tarum significantly reduced these 7-O-glucosides to aglycones
and hesperidin, in which L-rhamnose is α-1,6-linked to β-D- (Fig. 3c). The increase of naringenin and hesperetin was
glucoside, but did not cleave the neohesperidosides naringin and significantly higher in fermented extract that was previously
neohesperidin, in which L-rhamnose is α-1,2-linked to β-D- treated with 0.25 M of citric acid compared to fermented aqueous
glucoside (Supplementary Figure 1a). In contrast, β-D- extract (FermAQE; 0 M citric acid). Contrary to expectations,
glucosidase increased naringenin and hesperetin by degradation higher initial 7-O-glucoside levels due to higher citric acid
of naringenin-7-O-glucoside and hesperetin-7-O-glucoside, concentrations (0.50 M or 1.00 M) did not improve the final
respectively (Supplementary Figure 1b). aglycone content, indicating limited β-D-glucosidase activity.
Next, we screened four different LAB strains, namely L. plantarum Thus, we used the lowest citric acid concentration (0.25 M) to
DSM 20205, Lacticaseibacillus rhamnosus NCTC 10302 prepare the final CAE and fermented citric acid hydrolyzed
(L. rhamnosus), Levilactobacillus brevis DSM 6235 (L. brevis) and extracts (FermCAE). Flavanone profiles of the four final extracts
Lacticaseibacillus paracasei DSM 20312 (L. paracasei), to determine (AQE, FermAQE, CAE and FermCAE) that were used for in vitro
whether they produce the enzymes required for increasing the and in vivo experiments are provided in Fig. 3d and
aglycone levels in AQE. The specific LAB strains were selected based representative chromatograms in Supplementary Figure 5. Citric
on literature research16,17,31 and their safety for use in food and feed. acid hydrolysis proved to be an effective way to increase the
Bacterial growth was not inhibited by the concentration of citrus overall yield of 7-O-glucosides since CAE contained the highest
extract used (5% of citrus extract; Supplementary Figure 2). naringenin-7-O-glucoside (638 ± 201 µM) and hesperetin-7-O-
Although flavonoids have a low water solubility, water was used as glucoside (180 ± 40 µM) levels. Nevertheless, the maximum
solvent for citrus extract preparation because other organic solvents content of aglycones was achieved after fermentation in
such as ethanol would have greatly inhibited bacterial growth during FermCAE (812 ± 132 µM of naringenin and 291 ± 30 µM of
fermentation. After fermentation, the citrus extracts were centrifuged hesperetin), which was 4.5-fold (naringenin) and 3.1-fold
to remove the bacteria from the extract. Thus, the stated (hesperetin) higher than the aglycone content in AQE.
concentrations refer to flavonoids dissolved in the supernatant. As mentioned above, only dissolved flavonoids in the super-
All tested bacteria were able to convert naringenin-7-O- natant after centrifugation were considered for our final extracts.
glucoside and hesperetin-7-O-glucoside into the aglycones However, the entire fermentation product also includes pre-
naringenin and hesperetin, respectively. The highest aglycone cipitated flavonoids in the residue. We could show that the
increase with 168 µM of naringenin and 105 µM of hesperetin was percentage of flavonoids in the pellet was low, with a maximum of
observed after 24 h of fermentation with L. rhamnosus. However, 2.20% of the total flavonoid content (FermCAE). Of these,

Fig. 2 Enzymatic activity of various LAB strains for the formation of aglycones from 7-O-glucosides present in citric acid hydrolyzed citrus extract.
a Representative images of L. plantarum, L. rhamnosus, L. brevis and L. paracasei showing gram-positive rods. Scale bar: 10 µm. b Concentrations of flavanones
in citric acid hydrolyzed extract (CAE) after 24 h incubation with each strain at 37 °C. Control refers to the sample without bacteria. Data are mean ± SD of
n = 6 samples/strain. Differences to control are analyzed by Kruskal-Wallis’s test with Dunn’s multiple comparisons test, with exception of data set for
hesperetin which is analyzed by ordinary one-way ANOVA with Dunnett’s multiple comparison test.
aglycones accounted for the largest proportion with 0.63% and 0.57 nmol · mg-1 protein for naringenin and hesperetin)
(FermCAE). Supplementary Figure 6 summarizes the proportions compared to AQE (0.96 and 0.25 nmol · mg−1 protein,
of flavonoids in supernatant and pellet for all extracts. respectively).
Taken together, the combination of citric acid hydrolysis and For transport tests, Caco-2 monolayers were incubated with
fermentation with L. plantarum was proven to be the most AQE or FermCAE (2.5% in HBSS) for 4 h and the transport of
effective method to transform glycosides into aglycones in the flavanones from the apical to the basolateral compartment was
citrus extract. determined (Fig. 4a, 2). The monolayer integrity was controlled
by transepithelial electrical resistance (TEER) measurements. As
shown in Fig. 4e, TEER values remained above 500 Ω throughout
Intestinal epithelial cells take up and transport more bioactive the experiment indicating intact monolayers. We could confirm
flavonoids from biotransformed citrus extracts. To demon- that the aglycones have higher transport rates (above 32%) than
strate, whether the amount of aglycones present in bio- their respective glycosylated forms (below 2%). More impor-
transformed citrus extracts is bioavailable to intestinal epithelial tantly, we showed that this remains true when flavonoids were
cells, we investigated the cellular uptake and transport of flava- applied as a mixture in form of the citrus extracts AQE and
nones from AQE and FermCAE in differentiated Caco-2 cells. FermCAE (Fig. 4g, h). When epithelial cells received higher
For uptake tests, the cells were incubated with AQE or FermCAE concentrations of flavonoid aglycones in FermCAE, they could
(2.5% in Hank’s balanced salt solution; HBSS) for 4 h (Fig. 4a, 1). transfer significantly higher amounts of hesperetin and narin-
First, we could prove that cellular uptake of the aglycones nar- genin compared to AQE (p < 0.0001; Fig. 4f). Neither narirutin
ingenin and hesperetin was significantly better than uptake of nor the 7-O-glucosides could be detected in the basolateral
glycosylated flavonoids in both AQE and FermCAE (Fig. 4c, d). compartment after 4 h.
Moreover, when cells received higher concentrations of hesper- Our findings showed that citric acid hydrolysis and fermenta-
etin and naringenin in form of FermCAE (Fig. 4b), they absorbed tion of citrus extract did not change the dynamics of cellular
significantly higher absolute amounts of these aglycones (3.26 uptake or hinder cellular transport. On the contrary, we

demonstrated that epithelial cells could take up and transport Biotransformed citrus extracts reduce oxidative stress and
higher absolute amounts of flavonoids without reaching satura- improve cell migration of intestinal cells under challenging
tion within the concentrations in FermCAE. Taken together, conditions. The influence of individual flavanones (naringin,
higher flavonoid contents, provided by biotransformed citrus naringenin, hesperidin and hesperetin) and the four citrus
extract, resulted in a significantly higher uptake and transport extracts of different biotransformation level (AQE, FermAQE,
creating conditions for superior bioactivity of FermCAE. CAE and FermCAE) on cell viability and intracellular reactive

Fig. 3 Citric acid-based hydrolysis and fermentation affect flavanone profiles in citrus extracts. a Schematic diagram of extract preparation. For citric
acid hydrolyzed extract (CAE), bioflavonoid complex was treated with citric acid solution at 90 °C for 4 h. Then, sodium citrate solution was added to reach
a final concentration of 10% (w/w). Aqueous extract (AQE) was prepared with water instead of citric acid. For fermentation, AQE and CAE were incubated
with L. plantarum at 37 °C for 24 h (referred to as FermAQE and FermCAE). Control extracts were diluted in modified MRS and incubated without bacteria
(referred to as AQE and CAE). b Influence of hydrolysis with increasing citric acid (CA) concentrations (0.25 M CA, 0.50 M CA, or 1.00 M CA) on
flavanone concentrations in citrus extracts before incubation with bacteria. Data are mean ± SD of n = 4 samples/treatment. Differences to 0 M CA are
analyzed by ordinary one-way ANOVAs with Dunnett’s multiple comparison test, with exception of data set for hesperidin which is analyzed with Kruskal’s
test with Dunn’s multiple comparisons test. c Increase in aglycones after 24 h-incubation with L. plantarum. For calculation of the aglycone increase, the
initial aglycone concentration (at 0 h) was subtracted from the total concentration at 24 h. Data are mean ± SD of n = 5 samples/treatment. Differences to
0 M CA are analyzed by Kruskal’s test with Dunn’s multiple comparisons test. d Flavonoid concentrations in the final extracts prepared with water (AQE
and FermAQE) or 0.25 M CA (CAE and FermCAE). Data are mean ± SD of n = 9 samples/extract. Differences between treatments are analyzed by
ordinary one-way ANOVAs with Tukey’s multiple comparison test (narirutin, neohesperidin and hesperetin) or Kruskal’s test with Dunn’s multiple
comparisons test (naringin, naringenin-7-O-glucoside, naringenin, hesperidin and hesperetin-7-O-glucoside).
Fig. 4 Transport and uptake of flavonoids from AQE or FermCAE in Caco-2 cells. a Schematic overview of uptake test (1) or transport test (2) in
differentiated Caco-2 cells. b Comparison of aglycone uptake in Caco-2 cells treated with AQE or FermCAE for 4 h, normalized to cell protein content.
Uptake of detectable flavonoids from (c) aqueous citrus extract (AQE) and (d) fermented citric acid hydrolyzed citrus extract (FermCAE) in Caco-2 cells,
normalized to cell protein content. e TEER values of Caco-2 monolayers before and during transport studies. f Comparison of absolute aglycone
concentrations detected on the basolateral side after treatment with AQE or FermCAE on the apical side for 4 h. Transport rate of detectable flavonoids
from AQE (g) and FermCAE (h) across a Caco-2 monolayer. Data are mean ± SD of n = 6 samples/treatment. Differences are analyzed by ordinary one-
way ANOVAs with Šídák’s multiple comparison test.

Fig. 5 Impact of citrus extracts of different biotransformation level on ROS production and cell migration of intestinal epithelial cells under challenging
conditions. Percentage of intracellular reactive oxygen species (ROS) accumulation in human Caco-2 cells after treatment with pure flavanones (a), or
citrus extracts (b) and subsequent stress induction by 2,2’-azobis (2-amidinopropane) dihydrochloride (AAPH), normalized to stressor treatment. AQE
stands for aqueous citrus extract, CAE for citric acid hydrolyzed citrus extract, FermAQE for fermented aqueous citrus extract, FermCAE for fermented
citric acid hydrolyzed citrus extract. Percentage of intracellular ROS accumulation in swine epithelial IPEC-J2 cells after treatment with pure flavanones (c)
or citrus extracts (d) and subsequent stress induction by tert-butylhydroperoxide (tBHP), normalized to stressor treatment. Control refers to untreated
cells; quercetin was used as antioxidant positive control. Data are mean ± SD of n = 9 samples/treatment. Differences are analyzed by ordinary one-way
ANOVAs with Šídák’s multiple comparison test. e Cell front velocity [µm · h−1] of IPEC-J2 cells pre-treated with citrus extracts for 6 h and stressed with
tBHP after scratching. Data are mean ± SD of n = 6 samples/treatment. Differences are analyzed by ordinary one-way ANOVAs with Šídák’s multiple
comparison test. f Representative brightfield images of the scratch-wounded area at start (15 min after scratching) and end point (420 min). The wounded
area is highlighted in yellow ocher.
oxygen species (ROS) formation was investigated in Caco-2 and glycosylated flavanones (naringin and hesperidin) with their
IPEC-J2 cells. For ROS measurements, Caco-2 and IPEC-J2 cells respective aglycones (naringenin and hesperetin) when applied at
were first treated with the phytochemicals and then with suitable the same concentration (flavonoid stocks prepared in dimethyl
stressors, respectively. The applied concentrations of pure flava- sulfoxide, DMSO). Both aglycones, naringenin and hesperetin,
nones (150 µM) or extracts (1.25% and 2.5%) and subsequent exhibited a significantly stronger reduction of ROS formation in
treatment with the respective stressor did not affect cell viability stressed Caco-2 cells than their glycosylated forms naringin and
(Supplementary Table 1). We compared antioxidant effects of hesperidin (Fig. 5a). Along with this, the biotransformed extract

with the highest aglycone content, FermCAE at a concentration these cytokines than the non-biotransformed extract AQE. Three
of 2.5%, reduced ROS levels in stressed Caco-2 cells by 34%, being cytokines, epithelial neutrophil-activating protein 78 (ENA-78 or
significantly different to AQE (p < 0.0001) at the same con- CXCL5), granulocyte macrophage-colony stimulating factor
centration (Fig. 5b). Similar findings were obtained for stressed (GM-CSF) and vascular endothelial-derived growth factor
IPEC-J2 cells. Here, the difference between hesperidin (ROS (VEGF), were upregulated by treatment with both extracts but
reduction by 12%) and its aglycone hesperetin (ROS reduction by not with LPS alone. Additional data of cytokine array analysis
25%) was significant with p < 0.0001 (Fig. 5c). In addition, with the aglycone naringenin are provided in Supplementary
FermCAE at a concentration of 2.5% had the best antioxidant Figure 8, showing a similar cytokine profile to the one with the
effect, reducing ROS by 37% (Fig. 5d). This was significantly extracts.
different from the non-biotransformed extract AQE with ROS Based on our findings, we selected five cytokines for further
reduction by 16% at a concentration of 2.5% (p < 0.0001). Sam- investigation via bead-based multiplex immunoassay (Fig. 7a),
ples containing only bacteria (FermBlank) or citric acid solution namely IL-6, CXCL11, CCL7, CXCL9 and TNF-α. As expected,
(CA Blank) without flavanones did not reduce ROS formation in LPS significantly upregulated the secretion of these cytokines
stressed intestinal epithelial cells. This finding suggests that the compared to unstimulated control. Concurrent treatment with
superior effect of FermCAE over other citrus extracts was due to pure flavanones significantly (p < 0.0001) reduced IL-6 (Fig. 7b),
its higher aglycone content. CXCL11 (Fig. 7c) and CCL7 (Fig. 7d) and showed a slight
Since high ROS levels are associated with impaired wound decrease in CXCL9 (Fig. 7e) and TNF-α (Fig. 7f). Distinctively,
healing32, we next investigated whether the extracts could also the anti-inflammatory effects of the aglycones naringenin and
improve cell migration and thereby potentially wound healing hesperetin were more pronounced than of their glycosylated
capacity of IPEC-J2 cells, using a scratch assay. For this purpose, forms. Significant differences between naringin and its aglycone
cells were pre-treated with 1.25% of citrus extracts for 6 h. After naringenin were found for IL-6 (p = 0.0042), CCL7 (p = 0.0003)
scratching, cells were incubated with the stressor in presence of and TNF-α (p = 0.0076), while the aglycone hesperetin showed a
the citrus extracts and cell migration was recorded by automated significantly stronger reduction of CCL7 than hesperidin
time lapse microscopy for 7 h (Supplementary Movie). To (p = 0.0497). Consistent with these data, flavonoid-rich extracts
compare cell migration between treatments, the cell front velocity, also strongly decreased IL-6, CXCL11, CCL7, CXCL9 and TNF-α
defined as the speed at which cells move toward each other, was (Fig. 7g–k) in a concentration-dependent manner. Furthermore,
calculated. the extract with the highest aglycone content, FermCAE, reduced
As shown in Fig. 5e, f, stressor treatment strongly impaired cell pro-inflammatory cytokines more efficiently than the other
migration of IPEC-J2 cells. In contrast, all citrus extracts restored extracts.
cell migration ability, as evident by a significant rise in cell front In conclusion, the results of our in vitro experiments revealed
velocities compared to stressor treatment. The best effect was that the aglycones naringenin and hesperetin have a higher
observed with FermCAE, which increased cell front velocity by bioactivity than respective glycosylated flavonoids found in citrus
2.38 µm · h−1 (p < 0.0001). In comparison, the effect of AQE was extracts. Accordingly, the increase of the aglycone concentration
lower, showing an increase in cell front velocity by 1.35 µm · h−1 in citrus extracts improved their anti-inflammatory properties.
with p = 0.0069, but not significantly different from FermCAE.
To summarize, these data indicate that the increase of
naringenin and hesperetin in biotransformed citrus extracts leads Biotransformed citrus extracts reduce intestinal barrier
to an enhanced reduction of oxidative stress and improved cell damage and mortality in female D. melanogaster. Based on the
migration compared to the respective non-biotransformed in vitro experimental results, we further investigated antioxidant
extracts. and intestinal barrier protective properties of AQE and its cor-
responding biotransformed extract FermCAE in female D. mel-
anogaster. Considering the metabolic complexity of the whole
Biotransformed citrus extracts attenuate LPS-induced inflam- organism compared to the cell culture, the effect of food matrix
mation in macrophages. To study the influence of investigated on bioavailability33,34 and the involvement of gut microbiota of
extracts on inflammatory processes, we first performed a semi- D. melanogaster in intestinal function35,36, we doubled the
quantitative cytokine array analysis. Therefore, differentiated treatment extract concentrations compared to the in vitro
THP-1 macrophages were challenged with lipopolysaccharides experiments. The dextran sulphate sodium (DSS) induced
(LPS) for 24 h to trigger an inflammatory response and simul- intestinal barrier challenge was performed simultaneously with
taneously treated with 1.25% FermCAE or AQE. The con- AQE or FermCAE treatment (Fig. 8a, 1). DSS-induced rupture of
centrations were selected based on cytotoxicity testing intestinal barrier led to leakage of the blue dye through the entire
(Supplementary Figure 7). Membranes pre-loaded with anti- fruit fly, turning them into so called Smurfs (Fig. 8b, 1). The
bodies specific for 105 human cytokines were incubated with the choice of the fruit fly sex was based on the technical aspect of the
supernatants of treated THP-1 cells. Captured analytes were then experiment, such as clear visual identification of Smurfs, and was
visualized as spots (Fig. 6a). Spot intensities of important cyto- not due to the expectation of sex-specific effects. We scored and
kines that were upregulated upon LPS stimulation compared to compared the mortality and visible midgut leakage (Smurf phe-
control are presented in Fig. 6b, the overall expression profiles are notype) separately, because some DSS-fed flies were not
summarized in a heat map in Fig. 6c. Both extracts, AQE and demonstrating Smurf phenotype (Fig. 8b, 2). Our findings con-
FermCAE, reduced the concentrations of insulin-like growth firmed, that DSS significantly increases mortality and induces
factor-binding protein 3 (IGFBP-3), interleukin 6 (IL-6), inter- intestinal barrier rupture compared to control fruit flies. Impor-
leukin 17A (IL-17A), C-X-C motif chemokine ligand 11 tantly, we observed a significant decrease of mortality in D.
(CXCL11), CC-chemokine ligand 2 (CCL2), CC-chemokine melanogaster supplemented with higher concentrations of both
ligand 7 (CCL7), C-X-C motif chemokine ligand 9 (CXCL9), citrus extracts (Fig. 8c), 5% AQE (p = 0.0069) or 5% FermCAE
CC-chemokine ligand 20 (CCL20), osteopontin, regulated upon (p < 0.0001), respectively, with FermCAE demonstrating better
activation, normal T-cell expressed and presumably secreted protective effect compared to AQE (14.43% vs. 8.61% of mortality
(RANTES) and tumor necrosis factor α (TNF-α). The previously reduction compared to untreated DSS-challenged group). Visible
most promising extract FermCAE caused stronger reduction of damage of intestinal barrier was reduced for fruit flies treated

Fig. 6 Change in cytokine expression profiles in LPS-stimulated THP-1 macrophages treated with citrus extracts of different biotransformation level.
a Representative images of cytokine arrays incubated with supernatants of differentiated THP-1 macrophages treated with aqueous citrus extract (AQE) or
fermented citric acid hydrolyzed citrus extract (FermCAE) under lipopolysaccharide (LPS) challenge. Cytokines expressed are presented as duplicate spots.
Important cytokines are outlined and marked with numbers. The corresponding spot intensities (stated as arbitrary units, A.U.) of these analytes are
presented in b. Data of technical duplicates are shown. c Heat map of 105 cytokines with mean spot intensities in a color range from white (low level) to
blue (high level).
with both extracts in a concentration-dependent manner physiological level37,38 (Fig. 8a, 2). ROS formation and intracel-
(Fig. 8d). Moreover, Smurf phenotype reduction was comparable lular metabolic activity were quantitated after 70 h of induced
in 5% AQE and 5% FermCAE treatments (both p < 0.0001) with oxidative stress. In consistency with our in vitro results, citrus
19.87% and 21.06% respectively, compared to the untreated DSS- extracts significantly reduced ROS in fruit flies in a
challenged group. concentration-dependent manner (Fig. 8e). Particularly, the
Taken together, the aglycone-enriched biotransformed citrus extract with the highest aglycone content FermCAE at a con-
extract demonstrated a significant protective effect in female D. centration of 5% showed the best antioxidant effect reducing ROS
melanogaster under induced intestinal barrier challenge and by 64.27% compared to the group with stressor only. Both con-
reduced mortality in stressed fruit flies. centrations of FermCAE also showed a trend towards better
antioxidant activity compared to respective concentrations of
AQE, with a significant difference between the 2.5% extract
Biotransformed citrus extracts reduce ROS level and improve groups (p = 0.0014). No significant improvement of suppressed
climbing performance in stressed female D. melanogaster. To metabolic activity was observed in any of the stressed and extract-
test whether the observed antioxidant effect of biotransformed treated flies (Fig. 8f), compared to the untreated control group
extracts in vitro could be transferred to an in vivo model, we with significantly higher metabolic activity (p < 0.0001). All
induced oxidative stress in female D. melanogaster using sup- groups with ferrous iron diet demonstrated nearly equal levels of
plementation of ferrous iron (Fe2+) several times above the metabolic activity reduction.

Fig. 7 Biotransformed citrus extracts reduce pro-inflammatory cytokine levels in LPS-stimulated THP-1 macrophages. a Schematic graph of
lipopolysaccharide (LPS) challenge of THP-1 cells (1) with subsequent multiplex bead-based cytokine immunoassay (2). AQE stands for aqueous citrus
extract, CAE for citric acid hydrolyzed citrus extract, FermAQE for fermented aqueous citrus extract, FermCAE for fermented citric acid hydrolyzed citrus
extract. Effect of pure flavonoids (b−f) and citrus extracts (g−k) on expression of selected pro-inflammatory cytokines by differentiated THP-1
macrophages. Data are mean ± SD of n = 4 samples/treatment, measured in technical duplicates. Differences between treatments are analyzed by
ordinary one-way ANOVAs with Šídák’s multiple comparison test with exception of CXCL9 in pure flavonoid treatments (e), which is analyzed with
Kruskal-Wallis’s test with Dunn’s multiple comparisons test due to not normally distributed data.
In parallel to biochemical parameters, we investigated climbing making this plant material well applicable in nutraceuticals, food
activity and mortality of D. melanogaster due to oxidative stress. supplements or functional food and feed.
Our findings revealed that FermCAE significantly reduced Providing mechanistic detail and quantification, we demon-
mortality in flies upon oxidative stress (Fig. 8g) in a strated the superior antioxidant activity of biotransformed citrus
concentration-dependent manner; p = 0.0252 for 2.5% and extract in two mammalian intestinal epithelial cell lines, IPEC-J2
p < 0.0001 for 5% FermCAE. Similar to in vitro experiments, a and Caco-2, compared to non-biotransformed citrus extracts.
positive but much weaker effect was also observed for AQE, Moreover, we were able to show that improvement of bioactivity by
which was significant only for 5% AQE with p = 0.0002, biotransformation was attributed to the increase in naringenin and
indicating superior protective properties of biotransformed hesperetin levels, which were better taken up by cells than glyco-
extract FermCAE. This weak positive effect of AQE completely sylated flavonoids. Our findings are consistent with literature on
faded when the fitness level of D. melanogaster was concerned. In flavonoid uptake in Caco-2 cells, where uptake of naringin nor-
the climbing assay, where negative geotaxis of fruit flies was tested malized to cell protein was previously found to be low39. In con-
against time, only the group on a diet with the highest trast, uptake of its aglycone naringenin was comparatively higher, as
concentration of FermCAE was significantly better (p = 0.0019) confirmed in other studies40,41. Similarly, the transport rate of the
than the stressor-only group (Fig. 8h). aglycone hesperetin was also found to be superior than that of the
In conclusion, biotransformed extract with highest aglycone glycoside hesperidin42. However, none of the aforementioned stu-
content showed a significant antioxidant effect upon induced dies investigated the cellular uptake and transport of target flavo-
oxidative stress in female D. melanogaster, while the protective noids in a mixture within a matrix of a plant extract, in which
effect of non-biotransformed extract was weak to absent. compounds other than naringin, hesperidin and the respective
aglycones were present. Noteworthy, the extracts were dissolved in
HBSS for uptake and transport studies, although HBSS does not
Discussion mimic the properties and composition of gastrointestinal fluids.
In this study, we demonstrated that combination of acid hydro- Due to technical limitations in HPLC analysis, it was not possible to
lysis and fermentation with the probiotic bacterium L. plantarum use a solvent that better simulated intestinal fluids. The necessary
improves the cellular bioavailability and bioactivity of food grade purification steps for removing the interfering compounds would
citrus extract due to deglycosylation of its prototypic flavonoids, have resulted in analyte concentrations below the detection limit.

Fig. 8 Biotransformed citrus extracts improve intestinal barrier integrity and reduce oxidative stress in female w1118 D. melanogaster under challenge
conditions. a Schematic overview of in vivo challenges: (1) dextran sulphate sodium (DSS) challenge to reduce intestinal barrier integrity; (2) ferrous iron
induced oxidative stress in female w1118 D. melanogaster. AQE stands for aqueous citrus extract, FermCAE stands for fermented citric acid hydrolyzed citrus
extract, DCF(DA) stands for 2’,7’-dichlorodihydrofluorescein (diacetate). b D. melanogaster fed with blue dye (1) with DSS-challenged intestinal barrier
(Smurf phenotype) and (2) in control group. Scale bar: 500 µm. c, d Mortality and Smurf phenotype observed in D. melanogaster challenged by DSS and
treated with citrus extracts for 7 d. Data are mean ± SD of n = 300 flies/treatment. e, f Reactive oxygen species (ROS) level and metabolic activity of D.
melanogaster stressed with ferrous iron and treated with citrus extracts for 70 h, normalized to protein content. RFU stands for relative fluorescence units.
Data are mean ± SD of n = 12 samples (25 flies each)/treatment. g, h Mortality and climbing performance of D. melanogaster stressed with ferrous iron and
treated with citrus extracts for 7 d. Data are mean ± SD of n = 225 flies/treatment. Differences are analyzed by ordinary one-way ANOVAs with Šídák’s
multiple comparison test.
Considering the connection between oxidative stress and contrast, the aglycones naringenin and hesperetin or plant
inhibition of wound healing, we could also confirm enhanced cell extracts containing them have not been investigated in D. mela-
migration and thus wound healing capacity of IPEC-J2 cells nogaster so far. Here, we demonstrated an enhanced antioxidant
incubated with biotransformed citrus extracts. No studies were capacity of biotransformed citrus extract with increased con-
found on naringin, hesperidin and their respective aglycones in centration of the respective aglycones fed to D. melanogaster.
the context of our IPEC-J2 experiments. Fermented citrus extract not only reduced ROS level in chemi-
Moreover, the superior bioactivity of fermented citrus extract cally stressed fruit flies significantly stronger compared to non-
over non-fermented was confirmed in D. melanogaster. Previous biotransformed extract, but also led to superior reduction of
studies indicated antioxidant activities of naringin43,44 and stress-induced mortality.
hesperidin37,45 in fruit flies. In particular, the survival as well as In consensus with our in vitro cell migration experiments in
climbing performance of D. melanogaster stressed by iron excess intestinal epithelial cells, intestinal barrier challenge induced by
and treated with purified hesperidin was improved37,45, which is DSS confirmed protective properties of biotransformed citrus
also confirmed with our results despite the slight differences in extracts in vivo. DSS was proven to induce intestinal tissue
treatment duration and the way of hesperidin supplementation. damage in D. melanogaster46, but remains an underrated
The simultaneous treatment with a different oxidative stres- instrument to study the protective capacity of phytochemical
sor – trichloroethylene – and hesperidin also significantly reduced compounds. It was previously shown that DSS-induced intestinal
the ROS level and restored acetylcholinesterase activity in the inflammation could be reduced by anthocyanins47. There, a
fruit flies compared to the stressor only group37,45. The anti- reduction of intestinal barrier damage of ~20% was achieved in
oxidant effects of the glycoside naringin were as well examined the simultaneous treatment of experimental flies with DSS and
with a different stressor, such as paraquat, and resulted in bilberry anthocyanins extracts compared to DSS only. Another
improved survival rate, as well as locomotor activity43,44. In study showed that the aglycone hesperetin ameliorated DSS-

induced damage in Caco-2 cells, as well as reduces the damage in with high costs, since enzyme production is energy- and labor-
mice48. In the present study, we demonstrated that DSS-induced intensive56, and free enzymes cannot be conveniently reused. In
intestinal barrier damage and related mortality in D. melanogaster this regard, immobilized enzymes offer the advantage that they
were as well reduced due to supplementation of flavonoids in can be reused for multiple runs. For example, immobilized nar-
form of citrus extracts (hesperetin and naringenin included) and inginase was recently shown to retain 80% of its original activity
this reduction was even greater when aglycone-enriched extracts after 10 runs57. While highly efficient, enzymatic cleavage, as well
were applied. It is important to mention that gut microbiota of as application of bioengineered microorganisms also imply
living organisms can have an effect on flavonoid bioavailability. application limitations58 and heavier environmental footprint.
Therefore, in the conducted experiments, the intestinal bacteria of In our study, we presented a combination of citric acid
D. melanogaster might have influenced the intestinal deglycosy- hydrolysis with tailored LAB fermentation for conversion of
lation or further metabolization of naringin and hesperidin in the glycosylated flavanones to aglycones in food grade citrus extracts.
fed extracts as well. When material costs are considered, our suggested method is a
In the respective in vitro THP-1 model of inflammation, a cost-efficient alternative to the enzymatic deglycosylation and a
significantly stronger reduction of pro-inflammatory cytokines sustainable technology that assists deglycosylation of flavanones.
could also be traced back to aglycones with naringenin expressing For instance, costs of citric acid, LAB strains and nutrient media
the maximal anti-inflammatory effect. Our findings on reduction for initial LAB cultivation, required for biotransformation of one
of IL-6 and TNF-α are in line with previously investigated anti- unit of food grade citrus extract, are lower compared to the costs
inflammatory properties of naringenin49 and hesperetin50,51 and for pure naringinase, required for deglycosylation of the same
expands these data with the anti-inflammatory properties of the weight unit, even if the enzyme will be reused multiple times.
biotransformed citrus extracts that contain both aglycones. To Moreover, food-grade components allow the biotransformed
our knowledge, we first demonstrated flavanone-related decreases plant products to be used as dietary supplements, functional food
of the chemokines CXCL11, CCL7, and CXCL9 in the LPS- and beverages, as well as additives for animal feed. Therefore, it
stimulated macrophage model. In particular, we showed the might be of interest to use the entire fermentation product
superior anti-inflammatory properties of naringenin and including dissolved and undissolved flavonoids as well as bacteria
hesperetin compared to their glycosylated forms naringin and for large-scale applications. In this study, however, we only used
hesperidin, respectively. Reduction of CCL7 upon naringenin the supernatant with dissolved flavonoids as final extracts for
treatment was once demonstrated in high-fat diet-induced obese in vitro and in vivo experiments. Nevertheless, we consider our
mice52, while CXCL9 and CXCL11 were not investigated in biotransformation approach a promising method to improve the
context of flavanones at all. bioavailability and bioactivity of plant extracts rich in flavanones.
The topic whether aglycones are more beneficial in the human
body than glycosylated citrus flavonoids has been discussed Methods
extensively in literature, and to our knowledge, there is still no Extract preparation and biotransformation of flavonoids.
clear answer. For example, it has been reported that O-glycosy- Citrus bioflavonoids complex 60 T0110890 (named complex 60)
lation of naringenin and hesperetin generally reduces the biolo- was purchased from Evesa Extractos Vegetales S.A (Cadiz, Spain).
gical activity, such as anti-inflammatory and antioxidant Complex 60, a light brown powder, consists of whole dried fruit
properties of the compounds but can enhance other biological of Citrus aurantium var. amara L., Citrus sinensis, and Citrus
benefits such as anti-HIV activity, tyrosinase inhibition or anti- paradisi collected from Seville, Spain and has a total flavonoid
rotavirus activity10. At this point it has to be emphasized that content of > 60% with the main flavonoids identified as naringin,
findings from in vitro and in vivo studies are not directly neohesperidin and hesperidin.
applicable to humans without further investigations. Therefore, it For preparation of the non-biotransformed aqueous extract
can neither be generalized that aglycones always have a better (referred to as AQE), a 10% (weight per weight) solution of
effect on human health nor that conversion to aglycones is complex 60 in deionized water was prepared, incubated at 90 °C
required to achieve a beneficial effect on health. for 4 h and centrifuged at 4,000× g for 10 min. After extraction,
Given the demonstrated bioactivity of the aglycones naringenin approximately 60% of the flavonoids of complex 60 were
and hesperetin, the question arises whether existing methods of dissolved in the supernatant. Only the supernatant was used for
flavonoid deglycosylation, such as enzymatic cleavage, acid further experiments.
treatment or application of genetically modified organisms, are
optimal in terms of application field, sustainability and costs. It
was recently demonstrated that incubation with 0.5 M hydro- Citric acid hydrolysis. For citric acid hydrolysis, 10 g of complex
chloric acid in 80% ethanol effectively hydrolyzed the glycosy- 60 were incubated with 50 ml of citric acid solution (0.25 M,
lated flavonoid rutin to its aglycone form quercetin within 3 h53. 0.50 M or 1.00 M) in a water bath under constant stirring at 90 °C
However, application of hydrogen chloride is limited in food and for 4 h. After incubation, sodium citrate solution (0.25 M, 0.50 M,
feed technology, unlike citric acid54 applied in this study. 1.00 M, respectively) was added to reach a final weight of 100 g.
Moreover, citric acid is considered more environmentally friendly The final citric acid hydrolyzed extract (referred to as CAE) was
than hydrogen chloride because it is metabolized to carbon prepared with 0.25 M of citric acid and sodium citrate solution.
dioxide and water. In our approach, a 4.5-fold (naringenin) and Extracts were centrifuged at 4,000× g for 10 min to obtain the
3.1-fold (hesperetin) higher aglycone content was obtained by 4 h supernatants which were used for fermentation.
of citric acid hydrolysis and 24 h fermentation with L. plantarum
compared to the citrus extract prepared with water. There have Cultivation of bacteria. L. plantarum (DSM 20205), L. brevis
been previous attempts to use citric acid for deglycosylation of (DSM 6235) and L. paracasei (DSM 20312) were obtained from
flavonoids55. The researchers increased the extraction yield of the the German Collection of Microorganisms and Cell Cultures
aglycones apigenin and luteolin in celery extract by citric acid GmbH (Braunschweig, Germany), while L. rhamnosus (NCTC
hydrolysis and treatment with the purified enzyme β-glucosidase. 10302) was purchased from the National Collection of Type
They could even increase the content of the aglycones luteolin Cultures (NCTC, Salisbury, UK). All strains were cultivated
and apigenin by a factor of 14.5 and 83.6, respectively. However, microanaerobically in de Man, Rogosa and Sharpe (MRS) broth
the biotechnological application of purified enzymes is associated (Carl Roth GmbH, Karlsruhe, Germany) under agitation at 37 °C.

Modified MRS (pH 6.5 ± 0.2) was prepared in-house by dissol- (all Vanquish™) equipped with the Chromeleon™ software (Version
ving 10 g of peptone from casein, 4 g of yeast extract, 6 g dipo- 7.3.1, Thermo Scientific™ Dionex™, Waltham, MA, USA). Analyte
tassium hydrogen phosphate, 5 g of diammonium hydrogen separation was performed on an Accucore™ C18 column (150 mm
citrate and 0.04 g of manganese sulphate per liter (all Carl Roth × 3.0 mm inner diameter, 2.6 µm particle size; Thermo Scientific).
GmbH). The medium was heat-treated by autoclaving at 121 °C The column temperature was set to 40 °C, and the injection volume
for 12 min and stored at 4 °C before use. was 10 µl. UV-detection was performed at wavelength 260 nm.
Gram staining was performed with a Gram staining kit (Merck, Analytes were separated by gradient elution, with mobile phase A
Darmstadt, Germany) according to the manufacturer’s instruc- containing 0.1% formic acid (FA) in water, and mobile phase B
tions. Cells were observed via light microscopy (Nikon Eclipse containing 0.1% FA in acetonitrile, at a rate of 0.5 ml · min−1
80i, objective Plan Fluor 100 × 1.30; Nikon Instruments, (FA and acetonitrile of analytical grade). The elution gradient
Amsterdam, Netherlands) using differential interference contrast starting conditions were 95% A and 5% B. After 5 min of equili-
(DIC). Photos were taken with the Nikon DS U1 camera (5 MPx) bration time, the starting conditions were maintained for 2 min
using the NIS-Elements software (Version 5.02.01, Nikon before the proportion of B was increased to 20% at 5 min, to 30%
Instruments). at 15 min, to 40% at 20 min, to 60% at 23 min, and to 80% at
25 min. The last condition was held for 3 min, followed by
Fermentation. Fresh overnight cultures of each strain were pre- reducing the proportion of B to 5% within 2 min. Finally, the
pared in MRS by incubating under agitation at 37 °C for 24 h. On condition with 95% A and 5% B was maintained for 10 min.
the next day, bacteria were sedimented by centrifugation (4,000× The internal standards narirutin, naringin, naringenin-7-O-gluco-
g, 10 min) and resuspended in 12 ml of modified MRS with pH side, hesperidin, neohesperidin, hesperetin-7-O-glucoside, nar-
6.5 to adjust the optical density at 600 nm (OD600) to 0.8 (cor- ingenin (all Extrasynthese, Genay, France) and hesperetin
responds approximately to 5 × 108 CFU · ml−1). Next, 3 ml of (Sigma-Aldrich, Schnelldorf, Germany) were used for calibration in
either AQE or CAE were added. The pH value of the mixture of a linear range. Detailed information on calibration curves, linearity,
medium and extract was approximately 5. For non-fermented limit of detection (LOD) and limit of quantification (LOQ) are
extracts, AQE and CAE were diluted in modified MRS without provided in Supplementary Table 2 and Supplementary Figure 9.
bacteria (abbreviations remain AQE and CAE, respectively).
Culture vials were incubated under agitation at 37 °C for up to Cell culture maintenance. All chemicals and reagents were
144 h. The number of viable cells was determined using the plate obtained from Sigma-Aldrich, if not stated otherwise. Porcine
spread method. Extracts were incubated at 95 °C for 10 min to IPEC-J2 (ACC 701) and human Caco-2 (ACC 169) cells were
inhibit bacterial enzymes. Final extracts were fermented with L. purchased from German Collection of Microorganisms and Cell
plantarum for 24 h, and clarified by centrifugation (4,000× g, Cultures GmbH, while human THP-1 (TIB-202™) monocytes
10 min) before usage as 100% stocks (referred to as FermAQE were obtained from American Type Culture Collection (Mana-
and FermCAE) for in vitro and in vivo experiments. ssas, VA, USA). IPEC-J2 cells were maintained in Dulbecco’s
Changes in the flavonoid composition between start point Modified Eagle’s Medium (DMEM), Caco-2 cells in MEM Eagle
(t = 0 h) and end of fermentation (t = 24 h) were determined by with non-essential amino acids and THP-1 cells in RPMI-1640
high-performance liquid chromatography with ultraviolet detec- medium, all supplemented with 10% fetal bovine serum (FBS),
tion (HPLC-UV) as described in Section “Analysis of flavonoids 100 µg · ml−1 penicillin and 100 µg · ml−1 streptomycin (all PAN-
by HPLC-UV”. For analysis, 200 µl of sample were dried at 30 °C Biotech, Aidenbach, Germany). Growth medium for THP-1 cells
for 2 h using a vacuum concentrator (Labconco™ CentriVap™; was additionally enriched with 0.05 mM of 2‐mercaptoethanol
Fisher Scientific GmbH, Schwerte, Germany) and redissolved (PAN-Biotech), as recommended by ATCC. Cells were grown at
with 200 µl of DMSO and 1800 µl of HPLC solvent (50% of 37 °C in a humidified atmosphere (≥95%) with 5% CO2.
acetonitrile in deionized water). Redissolved samples were
centrifuged at 17,000× g for 10 min and diluted 1:10 in Cell viability assay. Toxic effects of test extracts and flavanones
deionized water. were evaluated using a resazurin-based in vitro toxicology assay
To determine precipitated flavonoids after centrifugation, the kit according to the manufacturer’s instructions. Briefly, cells
pellet (from 250 µl sample) was redissolved in 65 µl of DMSO and were seeded into black 96‐well plates at a density of 2 × 104
185 µl of 0.1% formic acid in methanol/50% acetone, centrifuged (IPEC-J2), 1.5 × 105 (Caco-2) or 105 (THP-1) cells per well,
at 17,000× g at 4 °C for 10 min and analyzed by HPLC. grown to 90% confluency, and incubated with test substances in
growth medium for 20 min (IPEC-J2 and Caco-2) or 24 h (THP-
Enzymatic hydrolysis by α-L-rhamnosidase and β-D- 1) at 37 °C. Then, medium was aspirated, and cells were incu-
glucosidase. For enzymatic hydrolysis of flavonoids, AQE was bated with 10% resazurin in growth medium for 1.5 h. The
diluted 200-fold in 100 mM sodium phosphate buffer with pH 6.5 amount of resorufin was measured using a microplate reader in
and incubated with either a recombinant α-L-rhamnosidase from fluorescence mode (544 nm excitation, 590 nm emission;
prokaryotic source (1 U · ml−1; Megazyme, Wicklow, Ireland) or POLARstar Omega, BMG LABTECH, Ortenberg, Germany).
β-D-glucosidase from almonds (1 U · ml−1; Merck) at 37 °C for Data were analyzed using the Omega MARS Data analysis soft-
up to 120 min. Negative controls were prepared without enzymes. ware package (BMG LABTECH). Cell viability was normalized to
Samples were taken every 30 min, immediately incubated at 95 °C untreated cells (control) grown under the same conditions.
for 5 min to inhibit enzymatic activity and clarified using cen-
trifugation at 17,000× g and 4 °C for 15 min. Supernatants were Transport and uptake tests in vitro. For uptake studies, Caco-2
analyzed using HPLC-UV as described in Section “Analysis of cells were seeded in 6 cm dishes at 4.4 × 106 cells per dish and
flavonoids by HPLC-UV”. grown overnight. Cells were differentiated on day 2 using DMEM
supplemented with 100 µg · ml−1 penicillin and 100 µg · ml−1
Analysis of flavonoids by HPLC-UV. Reversed phase chroma- streptomycin, 0.1% MITO+ Serum Extender (Corning, FAL-
tography was conducted using a Vanquish™ Core HPLC System 355006) and 5 mM butyric acid, which was refreshed on the third
comprised of a quaternary pump, a split sampler, a diode day. Cell layers acquired a differentiated phenotype, which was
array detector, and a temperature-controlled column compartment observed as the formation of domes under the microscope59.

Differentiated Caco-2 cells were washed with HBSS and incu- angle using a 100 μl plastic pipette tip. Excess debris was removed
bated with 5 ml of AQE or FermCAE (2.5% in HBSS) for 4 h at by washing with DPBS. Cells were stressed with 100 µM of tBHP
37 °C. Then, the supernatant was removed, cell layers were while incubated with the extracts. Extract treatments were com-
washed twice with Dulbecco’s phosphate buffered saline (DPBS; pared to cells stressed with tBHP only. Untreated cells were used
PAN-Biotech) and detached using a cell scraper. After cen- as control. For scanning of the scratched area, automated
trifugation and removal of DPBS, cell pellets were resuspended in brightfield imaging was performed every 15 min for 420 min on
50% acetonitrile and placed into the ultrasonic bath for 15 min to an epi-fluorescent microscope (Nikon Eclipse Ti2, Tokyo, Japan)
break up the cells. For precipitation of cellular proteins, the using a 2× CFI Plan Apochromat objective (NA = 0.1) with
suspensions were frozen at −20 °C for 30 min. Samples were additional 1.5× magnification. For temperature and CO2 control,
centrifuged and the obtained supernatants were analyzed by the microscope was equipped with a cage incubator (Okolab,
HPLC to quantitate flavonoids taken up by the cells. The Shanghai, China) allowing imaging of cell culture at 37 °C and 5%
remaining pellets were dissolved in 1 M NaOH, and the protein CO2. NIS-Elements software package (Version 5.02.01, Nikon,
content was determined by Quick Start™ Bradford 1× Dye Tokyo, Japan) enabled definition of acquisition positions with
Reagent (Bio-Rad Laboratories, CA, USA) according to the respect to the scratch in each well. For scanning of multiple stage
manufacturer’s protocol. positions in one experiment, the microscope was further equip-
For transport studies Caco-2 cells were seeded at 1.65 × 105 ped with a x-y-stage (CMR-STG-MHIX2-motorized table,
cells per trans-well insert (Thin-Cert, 0.336 cm2, 0.4 µm pore size; Märzhäuser, Germany) and a sCMOS camera (Zyla 4.2, Andor,
Greiner-Bio One, GRE-662640) and grown overnight. Cell layers Northern Ireland). Image analysis was carried out using ImageJ
were differentiated as described before and cell layer integrity was and the Wound_healing_size_tool.ijm plugin61. Cell front velocity
assessed by TEER measurement with a Millicell-ERS-2 Voltohm- (μm · h−1) was calculated as indicated in Eq. 1:
meter (Merck, Darmstadt, Germany). For the experiment,
differentiated cells were washed twice with HBSS and the inserts wound closure speed
cell front velocity ¼ ð1Þ
were placed in a fresh microtiter plate prefilled with 1040 µl HBSS length of cell front ´ 2
per well. Cell layers were provided with 300 µl of AQE or
FermCAE (2.5% in HBSS) in the apical compartment and with wound closure speed calculated from the slope of the linear
incubated for 4 h at 37 °C. To ensure cell layer integrity is trend line (area vs. time; standardized to 15−420 min) and the
maintained throughout the experiment, TEER measurements length of cell front × 2 corresponded to the length of the gap area
were performed directly after the treatment (t = 0 h) and at the considering two cell fronts that migrate towards each other (see
end of the transport test (t = 4 h). Flavonoid concentrations were also ibidi-cells in focus62).
determined by HPLC analysis as described in Section “Analysis of
flavonoids by HPLC-UV”. LPS challenge assay in THP-1 cells. THP-1 cells were seeded in
6-well plates at a density of 2.75 × 106 cells per well and stimu-
lated with 50 ng · ml−1 of phorbol-12-myristate-13-acetate
ROS quantification assay in vitro. Cells were seeded into black (PMA) for differentiation into macrophages. After incubation for
96‐well plates at a density of 2 × 104 (IPEC-J2) or 1.5 × 105 48 h, depleted medium was aspirated, and adherent macrophages
(Caco-2) cells per well and grown overnight. Next, cells were were incubated simultaneously with 250 ng · ml−1 of LPS and
washed with DPBS and incubated simultaneously with 50 µM of extracts (1.25% or 2.5%) in growth medium with 1% of FBS (in
the ROS probe 2’,7’-dichlorodihydrofluorescein diacetate total 2.5 ml per well). Cell culture supernatants containing
(H2DCFDA)60 and extracts (1.25% and 2.5% in HBSS) or flavo- secreted cytokines were harvested after 24 h and stored at −80 °C
noids (150 µM of naringin, naringenin, hesperidin or hesperetin prior to analysis. Supernatants of extract treated cells were
in HBSS). Quercetin (20 µM) was used as an antioxidant positive compared to the ones of cells challenged with LPS only.
control. Flavonoid stocks were prepared in DMSO, but con- Untreated cells were used as control.
centrations applied to cells did not exceed 0.1% of DMSO. After
incubation for 20 min, cells were washed twice with DPBS and Cytokine array analysis. The expression profile of cytokines,
stressed with 100 µM of tert-butyl hydroperoxide (tBHP; for including chemokines, and growth factors was assessed using
IPEC-J2) or 500 µM of 2,2’-azobis(2-methylpropionamidine) Human XL Cytokine Array Kit (Proteome Profiler™ Array, Bio-
dihydrochloride (AAPH; for Caco-2). Two different stressors Techne Ltd., Abingdon, UK) following the manufacturer’s
were used because not every cell line responds to each stressor. instructions. Briefly, the membranes were blocked at room tem-
Cells treated with HBSS were used as control. The amount of perature (RT) for 60 min and incubated with samples at 4 °C
produced dichlorodihydrofluorescein (DCF) was determined with overnight. After washing, the membranes were treated with bio-
a microplate reader in fluorescence mode (485 nm excitation, tinylated detection antibodies at RT for 90 min. Finally, they were
520 nm emission; POLARstar Omega, BMG LABTECH). Data washed and incubated with streptavidin-horseradish peroxide at
were analyzed using the Omega MARS Data analysis software RT for 30 min. Captured proteins were visualized by incubation
package (BMG LABTECH). Measurements started immediately with chemiluminescent detection reagents for 1 min, followed by
after stressor addition at 0, 15, 30, 60 and 90 min. This oxidative measuring the luminescence using a ChemiDoc MP Imaging
status over time was summarized by calculation of the area under System (Bio-Rad Laboratories). Semi-quantitative comparison
the curve normalized to the stressor treatment. was performed using the Image Lab Software (Bio-Rad Labora-
tories). The pixel intensity in each spot of the array was analyzed
Cell migration assay in IPEC-J2 cells. The effect of bio- and the background signal was subtracted from each spot.
transformed citrus extracts on migration of stressed IPEC-J2 cells
was assessed using a scratch assay20. IPEC-J2 cells were seeded in Cytokine multiplex immunoassay. Based on cytokine membrane
24-well imaging plates with cover glass (MoBiTec, Goettingen, array results, the selected cytokines CCL7, CXCL11, CXCL9, IL-6,
Germany) at a density of 2.5 × 105 cells per well and grown and TNF-α were quantified with Luminex xMAP® custom mul-
overnight. When confluency was achieved, cells were pretreated tiplex assay (Bio-Techne Ltd.). For this purpose, cell culture
with the extracts at a concentration of 1.25% for 6 h. Then, a supernatants of LPS-stimulated THP-1 cells were applied undi-
straight scratch was generated with constant pressure, speed and luted for CCL7, CXCL11, CXCL9 and diluted 1:100 for IL-6 and

TNF-α quantitation, respectively. Samples or standards were solution of 5% sucrose and 1% Brilliant Blue FCF only. Each
incubated for 2 h in presence of antibody-coated magnetic experimental condition was represented by 4 vials (≈100 fruit
microbeads, followed by washing step (Fig. 7a). This and all the flies) per experiment. These experimental flies were maintained in
next incubation periods were performed in a sealed, light- standard climate conditions at 25 °C and 60% RH for 7 d and
protected black well plate agitated with thermo-shaker at RT, were transferred to a fresh agar vial with a soaked filter paper 4
800 rpm. For the next 1 h of incubation, biotinylated detection times per experiment. The number of dead flies as well as of flies
antibodies were added. Microbeads with analyte sandwiches were with Smurf phenotype was scored daily. Female D. melanogaster
then washed and incubated with streptavidin-phycoerythrin were scored as Smurfs as described by Rera et al.64 when they
solution for 30 min. Afterwards, unbound reagent was removed. demonstrated blue coloration observed outside intestinal area, e.g.
Prepared samples were resuspended with washing buffer, in thorax and limbs (Fig. 8b, 1).
acquired by Luminex® 200™ and analyzed with xPonent® acqui-
sition software, Version 4.3 (both Luminex Corp., Austin, TX,
Induction of oxidative stress in D. melanogaster. To evaluate
USA) in technical duplicates. Calibration and interpolation of
possible antioxidant effects of the biotransformed extract
unknown concentrations was performed with weighted logistic
FermCAE in vivo, series of twelve experiments were conducted
5P curve with intra-assay goodness-of-fit R2 = 0.995. Intra-assay
using ferrous iron as oxidative stressor. For that 5 d. o. w1118
coefficient of variation equaled 5.26%, intra-assay recovery of
female D. melanogaster were sorted to groups of 25 ± 2 adult flies
standards equaled 100.53%. Detailed inter-assay characteristics
per vial and placed on solid experimental media that was
per analyte are listed in Supplementary Table 3.
exchanged every 2−3 d. Experimental media were based on
sugar-yeast medium described above (also applied as control
Husbandry of D. melanogaster and preparation of experi- group) with addition of 30 mM iron (II) sulphate heptahydrate
mental fruit flies. Parental fruit flies D. melanogaster, strain as oxidation-inducing stressor37,65 and 2.5% or 5% of AQE
w1118 (University of Kiel, Kiel, Germany; strain no. 5905 Bloo- or FermCAE in one of the 4 treatment groups, respectively
mington Drosophila Stock Center, Bloomington, IN, USA), were (Fig. 8a, 2). Each experimental group was represented with ≈ 100
reared in our laboratory for over 60 generations in standardized flies (4 vials) per experiment. With respect to physiological
conditions at 25 °C, 60% relative humidity (RH), 12:12 h light- changes related to circadian rhythm66–68 fruit flies were placed
dark cycle. Controlled climate conditions were insured by cli- on experimental media at the exact same day time (11:30 am) in
mate chamber HPP750 eco (Memmert GmbH + Co. KG, Ger- each of 3 experiments for biochemical assays or for survivorship
many). If not stated otherwise, all reagents in the D. melanogaster and climbing performance.
experiments were purchased from Carl Roth GmbH. To prepare
aged-synchronized fruit flies, parental flies were transferred to
ROS quantification and metabolic activity assays in D. mela-
the embryo collection cages with Nutri-Fly® grape agar petri
nogaster. For both biochemical assays, female D. melanogaster
dishes (both Genesee Scientific, San Diego, CA, USA) prepared
were sacrificed exactly after 70 h. Homogenizer tubes with
according to the manufacturer’s instructions. To induce egg
stainless steel beads (Precellys Lysing Kit for Hard Tissue
laying behavior, parental flies were provided with fresh yeast
Grinding, Bertin Technologies, Main, Germany) were filled with
paste consisting of active dry yeast “Red star” (Genesee Scien-
300 µl of ice-cold DPBS containing 1 µl · ml−1 protease inhibitor
tific) and deionized water. After 24 h, the petri dish was replaced
cocktail (PIC; Sigma-Aldrich). Flies were anesthetized with CO2
with a fresh one. In ca. 20 h, eggs were collected from grape agar,
for 10 s, transferred with a funnel to the homogenizer tubes and
washed with sterile DPBS 3 times, and dispensed on larval
homogenized in Precellys Evolution tissue homogenizer (Bertin
growth medium (35 µl per stock bottle). In 9 d freshly eclosed
Technologies) at 6000 rpm for 40 s. Afterwards, 700 µl of DPBS-
synchronized adults were transferred to the stock bottles with
PIC mix were added to homogenate on ice. 800 µl of sample
sugar-yeast medium and allowed to mate for 2 d before the
were transferred to the fresh tubes without steel beads and
sorting to sexes. The medium for larval growth was prepared as
centrifuged for 10 min at 4 °C and 14,000× g. Clear supernatant
described22,63 from 1% agar, 5.5% glucose, 3% sucrose, 2.5%
was kept on ice and assayed immediately in ROS quantification
inactive dry yeast, 6% yellow cornmeal (both Genesee Scientific),
and metabolic activity assays, as well as in Bradford assay for
all percentage is weight per volume. Sugar-yeast medium con-
protein normalization. For ROS quantification, 90 µl of DPBS,
sisted of 1.5% agar, 5% sucrose, 10% inactive dry yeast (weight
100 µl of samples and 10 µl of 100 µM H2DCFDA were incu-
per volume). Both media were preserved with 1% of methyl
bated in a black 96‐well plate light-protected for 2 h. Addi-
4-hydroxybenzoate solution (10% solution in absolute ethanol,
tionally blanks of homogenization buffer with H2DCFDA and
weight per volume), and 0.48% propionic acid (volume per
sample supernatants without H2DCFDA were tested to exclude
volume).
autofluorescence interference. End point measurement of
fluorescent DCF product was performed as described in Section
Intestinal barrier challenge in D. melanogaster. To investigate “ROS quantification assay in vitro”. The metabolic activity assay
the protective effect of biotransformed citrus extracts in vivo, a was based on reduction of resazurin to fluorescent resorufin by
series of three experiments with DSS challenge were conducted. NADH, NADPH, or FADH. In brief, 50 µl of sample and 50 µl of
The experiments were based on the method of Amcheslavsky 25 µM resazurin solution (Thermo Fisher Scientific) were incu-
et al.46 with our modifications22 (Fig. 8a, 1). 5 days old (d. o.) bated light-protected for 2 h. Analogue to ROS assay, blanks
w1118 female D. melanogaster were sorted to groups of 25 ± 2 were tested to exclude variating autofluorescence of samples or
adult flies per vial and transferred to non-nutritional medium homogenization buffer. End point measurement of fluorescent
(1.5% agar in water) in order to prevent drying out of the resorufin was performed as described in Section “Cell viability
experimental media in the empty vials. Experimental liquid media assay”. To normalize obtained values of biochemical assays,
consisted of aqueous solution of 5% sucrose, 1% Brilliant Blue proteins in sample supernatants were quantified with Bradford
FCF (C.I. 42090), 5% DSS with average MW 40,000 g · mol−1, assay69 with modifications. Obtained supernatants were diluted
and either 2.5% or 5% of AQE or FermCAE, and were introduced 1:20 in 20 mM sodium chloride solution and assayed with Quick
on a piece of 1.5 × 3 cm, 1.5 mm thick gel blotting paper Start™ Bradford 1× Dye Reagent (Bio-Rad Laboratories) in 1:5
(Whatman, Cytiva, UK). The control group was fed with aqueous proportion. Calibration was performed with 2-fold dilution of

BSA solutions (Sigma-Aldrich) in concentrations from 0 to Received: 3 March 2023; Accepted: 5 October 2023;
100 µg · ml−1. After 1 h incubation, absorbance was measured
with POLARstar® Omega Microplate Reader at 590 nm and
normalized at 450 nm. The second wavelength was used for
linearization of calibration graph as suggested by Ernst&Zor70.
All fluorescence and absorbance measurements were processed References
with Omega Software, Version 2.10 and Mars Data Analysis 1. Guan, R. et al. A review of dietary phytochemicals and their relation to
Software, Version 2.30 (both BMG LABTECH GmbH). All oxidative stress and human diseases. Chemosphere 271, 129499 (2021).
measurements were performed in technical duplicates except 2. Liu, S. et al. Review of phytochemical and nutritional characteristics and food
applications of Citrus L. fruits. Front. Nutr. 9, 968604 (2022).
Bradford assays, which were run in technical triplicates.
3. Zhang, M., Zhu, S., Yang, W., Huang, Q. & Ho, C.-T. The biological fate and
bioefficacy of citrus flavonoids: bioavailability, biotransformation, and delivery
Mortality and climbing performance experiments in D. mela- systems. Food Funct. 12, 3307–3323 (2021).
nogaster. To further investigate physiological effects of citrus 4. Wang, Y. et al. Citrus flavonoids and their antioxidant evaluation. Crit. Rev.
Food Sci. Nutr. 62, 3833–3854 (2021).
bioflavonoids on fruit flies under the oxidative challenge, 5 d. o.
5. Barreca, D. et al. Flavanones: citrus phytochemical with health-promoting
w1118 female D. melanogaster flies were reared on experimental properties. BioFactors 43, 495–507 (2017).
media with ferrous iron for 7 d. During that period D. melano- 6. Ren, H. et al. Hesperetin suppresses inflammatory responses in
gaster were transferred to fresh media 4 times, dead flies were lipopolysaccharide-induced RAW 264.7 cells via the Inhibition of NF-κB and
scored during every transfer. Results were expressed as percent of activation of Nrf2/HO-1 pathways. Inflammation 39, 964–973 (2016).
7. Pinho-Ribeiro, F. A. et al. The citrus flavonone naringenin reduces
dead insects from all fruit flies in each vial. Climbing activity of D.
lipopolysaccharide-induced inflammatory pain and leukocyte recruitment by
melanogaster was assessed as previously described71,72 with inhibiting NF-κB activation. J. Nutr. Biochem. 33, 8–14 (2016).
minor modifications: on day 7 percent of subjects, climbed to the 8. Stevens, Y. et al. The intestinal fate of citrus flavanones and their effects on
top of the vial (5 cm distance) in 20 s after being tapped to the gastrointestinal health. Nutrients 11, 1464 (2019).
bottom, was scored. For each vial, the measurement was per- 9. Park, C.-M., Kim, G.-M. & Cha, G.-S. Biotransformation of flavonoids by
newly isolated and characterized Lactobacillus pentosus NGI01 strain from
formed twice, and the mean value was obtained.
kimchi. Microorganisms 9, 1075 (2021).
10. Xiao, J. Dietary flavonoid aglycones and their glycosides: Which show better
biological significance? Crit. Rev. Food Sci. Nutr. 57, 1874–1905 (2017).
Statistics, reproducibility and visual content. All statistical tests 11. Di Majo, D. et al. Flavanones in citrus fruit: structure–antioxidant activity
were performed in GraphPad Prism, Version 9.4.1 (GraphPad relationships. Food Res. Int. 38, 1161–1166 (2005).
Software, San Diego, CA, USA) as recommended in GraphPad 12. Nielsen, I. L. F. et al. Bioavailability is improved by enzymatic modification of
Statistics Guide. Prior to significance tests each data set was tested the citrus flavonoid hesperidin in humans. A randomized, double-blind,
crossover trial. J. Nutr. 136, 404–8 (2006).
to normality of data distribution with Shapiro-Wilk test. Nor- 13. Da Silva, C. M. G. et al. Enhancement of the antioxidant activity of orange and
mally distributed data were analyzed with two-tailed t-test (for 2 lime juices by flavonoid enzymatic de-glycosylation. Food Res. Int. 52,
data sets) or ordinary one-way ANOVA (for > 2 data sets). 308–314 (2013).
Dependent on the research question, ANOVA was extended with 14. FDA. GRAS Notices. Available at https://www.cfsanappsexternal.fda.gov/
one of the following multiple comparison tests: 1) comparison of scripts/fdcc/index.cfm?set=GRASNotices&sort=GRN_No&order=
DESC&startrow=1&type=basic&search=plantarum (2022).
all data sets with each other—Tukey’s test, 2) comparison of 15. Koutsoumanis, K. et al. Update of the list of QPS-recommended biological
treatment group vs. control or stressor group—Dunnett’s test, 3) agents intentionally added to food or feed as notified to EFSA 14: suitability of
comparison of selected groups with control, stressor, or each taxonomic units notified to EFSA until March 2021. EFSA J. 19, e06689
other — Šídák’s test. Data that did not pass normality distribution (2021).
test were analyzed with Kruskal-Wallis’s test with Dunn’s mul- 16. Mueller, M. et al. Rhamnosidase activity of selected probiotics and their ability
to hydrolyse flavonoid rhamnoglucosides. Bioprocess Biosyst. Eng. 41, 221–228
tiple comparisons test (for > 2 data sets). All tests were conducted (2018).
with 95% confidence interval (CI). Detailed summaries of sta- 17. Pereira-Caro, G. et al. Catabolism of citrus flavanones by the probiotics
tistical tests including number of subjects in each group n, F-, H- Bifidobacterium longum and Lactobacillus rhamnosus. Eur. J. Nutr. 57,
or t-values, as well as degree of freedom are provided in Sup- 231–242 (2018).
plementary Tables 4−6. p values ≤ 0.05 were considered sig- 18. Müller, U. et al. In vitro and in vivo inhibition of intestinal glucose transport
by guava (Psidium guajava) extracts. Mol. Nutr. Food Res. 62, e1701012
nificant and indicated as * (p ≤ 0.05), ** (p ≤ 0.01), *** (2018).
(p ≤ 0.001) or **** (p ≤ 0.0001). The uncertainty of the visualized 19. Blank-Landeshammer, B. et al. Improved bioavailability and bioaccessibility of
data was expressed via standard deviations (SD) shown as error lutein and isoflavones in cultured cells in vitro through interaction with
bars. All graphs were constructed in GraphPad Prism, Version ginger, curcuma and black pepper extracts. Antioxidants 11, 1917 (2022).
9.4.1. Schematic overviews of respective methods were created by 20. Vergauwen, H. et al. Trolox and ascorbic acid reduce direct and indirect
oxidative stress in the IPEC-J2 cells, an in vitro model for the porcine
authors in BioRender application (Science Suite Inc., Toronto, gastrointestinal tract. PLoS one 10, e0120485 (2015).
ON, Canada). Images and photos were processed in Corel- 21. Cao, S. et al. AMPK-PINK1/Parkin mediated mitophagy is necessary for
DRAW® Graphics Suite, Version 21.0.0.593 (Corel Corporation, alleviating oxidative stress-induced intestinal epithelial barrier damage and
Ottawa, ON, Canada). mitochondrial energy metabolism dysfunction in IPEC-J2. Antioxidants 10,
2010 (2021).
22. Heckmann, M. et al. Extracts prepared from feed supplements containing
Reporting summary. Further information on research design is wood lignans improve intestinal health by strengthening barrier integrity and
available in the Nature Portfolio Reporting Summary linked to reducing inflammation. Molecules 27, 6327 (2022).
this article. 23. Chanput, W., Mes, J. J. & Wichers, H. J. THP-1 cell line: an in vitro cell
model for immune modulation approach. Int. Immunopharmacol. 23, 37–45
(2014).
24. Staats, S., Lüersen, K., Wagner, A. E. & Rimbach, G. Drosophila melanogaster
Data availability as a versatile model organism in food and nutrition research. J. Agric. Food
Numerical source data for all charts and graphs can be found in Supplementary Data. Chem. 66, 3737–3753 (2018).
Further data that support the findings of this study are not openly available due to 25. Wong, A. C. N., Vanhove, A. S. & Watnick, P. I. The interplay between
reasons of sensitivity and are available from the corresponding author upon reasonable intestinal bacteria and host metabolism in health and disease: lessons from
request. Drosophila melanogaster. Dis. Model. Mech. 9, 271–281 (2016).

26. Capo, F., Wilson, A. & Di Cara, F. The intestine of Drosophila melanogaster: 52. Tsuhako, R., Yoshida, H., Sugita, C. & Kurokawa, M. Naringenin suppresses
an emerging versatile model system to study intestinal epithelial neutrophil infiltration into adipose tissue in high-fat diet-induced obese mice.
homeostasis and host-microbial interactions in humans. Microorganisms 7, J. Nat. Med. 74, 229–237 (2020).
336 (2019). 53. Yang, J. et al. Conversion of rutin to quercetin by acid treatment in relation to
27. Sandner, G., König, A., Wallner, M. & Weghuber, J. Alternative model biological activities. Prev. Nutr. Food Sci. 24, 313–320 (2019).
organisms for toxicological fingerprinting of relevant parameters in food and 54. Scientific opinion on the safety and efficacy of citric acid when used as a
nutrition. Crit. Rev. Food Sci. Nutr. 62, 5965–5982 (2022). technological additive (preservative) for all animal species. EFSA J. 13, (2015).
28. Le Bourg, E. Oxidative stress, aging and longevity in Drosophila melanogaster. 55. Che, D. N. et al. Citric acid and enzyme-assisted modification of flavonoids
FEBS Lett. 498, 183–186 (2001). from celery (Apium graveolens) extract and their anti-inflammatory activity in
29. Seto, N. O., Hayashi, S. & Tener, G. M. Overexpression of Cu-Zn superoxide HMC-1.2 cells. J. Food Biochem. 45, e13774 (2021).
dismutase in Drosophila does not affect life-span. Proc. Natl Acad. Sci. USA 56. Puri, M. & Banerjee, U. C. Production, purification, and characterization of
87, 4270–4274 (1990). the debittering enzyme naringinase. Biotechnol. Adv. 18, 207–217 (2000).
30. Orr, W. C., Arnold, L. A. & Sohal, R. S. Relationship between catalase activity, 57. Bodakowska-Boczniewicz, J. & Garncarek, Z. Immobilization of naringinase
life span and some parameters associated with antioxidant defenses in from Aspergillus niger on a magnetic polysaccharide carrier. Molecules 25,
Drosophila melanogaster. Mech. Ageing Dev. 63, 287–296 (1992). 2731 (2020).
31. Beekwilder, J. et al. Characterization of rhamnosidases from Lactobacillus 58. Leonard, W., Zhang, P., Ying, D., Adhikari, B. & Fang, Z. Fermentation
plantarum and Lactobacillus acidophilus. Appl. Environ. Microbiol. 75, transforms the phenolic profiles and bioactivities of plant-based foods.
3447–3454 (2009). Biotechnol. Adv. 49, 107763 (2021).
32. Kurahashi, T. & Fujii, J. Roles of antioxidative enzymes in wound healing. J. 59. Marziano, M. et al. Monitoring Caco-2 to enterocyte-like cells differentiation
Dev. Biol. 3, 57–70 (2015). by means of electric impedance analysis on printed sensors. Biochim. Biophys.
33. Kamiloglu, S., Tomas, M., Ozdal, T. & Capanoglu, E. Effect of food matrix on Acta Gen. Subj. 1863, 893–902 (2019).
the content and bioavailability of flavonoids. Trends Food Sci. Technol. 117, 60. Wang, H. & Joseph, J. A. Quantifying cellular oxidative stress by
15–33 (2021). dichlorofluorescein assay using microplate reader. Free Radic. Biol. Med. 27,
34. Dima, C., Assadpour, E., Dima, S. & Jafari, S. M. Bioavailability of 612–616 (1999).
nutraceuticals: role of the food matrix, processing conditions, the 61. Suarez-Arnedo, A. et al. An Image J plugin for the high throughput image
gastrointestinal tract, and nanodelivery systems. Compr. Rev. Food Sci. Food analysis of in vitro scratch wound healing assays. PloS one 15, e0232565
Saf. 19, 954–994 (2020). (2020).
35. Miguel-Aliaga, I., Jasper, H. & Lemaitre, B. Anatomy and physiology of 62. ibidi cells in focus. Experimental Setup Optimization and Data Analysis of
the digestive tract of Drosophila melanogaster. Genetics 210, 357–396 Wound Healing Assays. Application Note 30, Version 1.3. Available at https://
(2018). ibidi.com/img/cms/support/AN/AN30_Wound_Healing_Data_Analysis.pdf
36. Lesperance, D. N. A. & Broderick, N. A. Gut bacteria mediate nutrient (2019).
availability in Drosophila diets. Appl. Environ. Microbiol. 87, e01401–20 63. Piegholdt, S., Rimbach, G. & Wagner, A. E. Effects of the isoflavone prunetin
(2020). on gut health and stress response in male Drosophila melanogaster. Redox Biol.
37. Poetini, M. R. et al. Hesperidin attenuates iron-induced oxidative damage and 8, 119–126 (2016).
dopamine depletion in Drosophila melanogaster model of Parkinson’s disease. 64. Rera, M., Clark, R. I. & Walker, D. W. Intestinal barrier dysfunction links
Chem. Biol. Interact. 279, 177–186 (2018). metabolic and inflammatory markers of aging to death in Drosophila. Proc.
38. Jimenez-Del-Rio, M., Guzman-Martinez, C. & Velez-Pardo, C. The effects of Natl Acad. Sci. USA 109, 21528–21533 (2012).
polyphenols on survival and locomotor activity in Drosophila melanogaster 65. Jomova, K. & Valko, M. Advances in metal-induced oxidative stress and
exposed to iron and paraquat. Neurochem. Res. 35, 227–238 (2010). human disease. Toxicology 283, 65–87 (2011).
39. Turner, M. D., Nedjai, B., Hurst, T. & Pennington, D. J. Cytokines and 66. Xu, K., Zheng, X. & Sehgal, A. Regulation of feeding and metabolism by
chemokines: At the crossroads of cell signalling and inflammatory disease. neuronal and peripheral clocks in Drosophila. Cell Metab. 8, 289–300 (2008).
Biochim. Biophys. Acta 1843, 2563–2582 (2014). 67. Dubowy, C. & Sehgal, A. Circadian rhythms and sleep in Drosophila
40. Zhang, Z.-D. et al. Uptake and transport of naringenin and its antioxidant melanogaster. Genetics 205, 1373–1397 (2017).
effects in human intestinal epithelial Caco-2 cells. Front. Nutr. 9, 894117 68. Zhang, Y. & Emery, P. Molecular and neural control of insect circadian
(2022). rhythms. Insect Biochem. Mol. Biol. (Elsevier), 513–551 (2012).
41. Nait Chabane, M., Al Ahmad, A., Peluso, J., Muller, C. D. & Ubeaud, G. 69. Bradford, M. M. A rapid and sensitive method for the quantitation of
Quercetin and naringenin transport across human intestinal Caco-2 cells. J. microgram quantities of protein utilizing the principle of protein-dye binding.
Pharm. Pharmacol. 61, 1473–1483 (2009). Anal. Biochem. 72, 248–254 (1976).
42. Kobayashi, S., Tanabe, S., Sugiyama, M. & Konishi, Y. Transepithelial 70. Ernst, O. & Zor, T. Linearization of the Bradford protein assay. J. Vis. Exp. 38,
transport of hesperetin and hesperidin in intestinal Caco-2 cell monolayers. 1918 (2010).
Biochim. Biophys. Acta 1778, 33–41 (2008). 71. Chambers, R. P. et al. Nicotine increases lifespan and rescues olfactory and
43. Lashmanova, E. et al. The Evaluation of geroprotective effects of selected motor deficits in a Drosophila model of Parkinson’s disease. Behav. Brain Res.
flavonoids in Drosophila melanogaster and Caenorhabditis elegans. Front. 253, 95–102 (2013).
Pharmacol. 8, 884 (2017). 72. Bartholomew, N. R., Burdett, J. M., VandenBrooks, J. M., Quinlan, M. C. &
44. Anh, N. T. T., Nishitani, M., Harada, S., Yamaguchi, M. & Kamei, K. A Call, G. B. Impaired climbing and flight behaviour in Drosophila melanogaster
Drosophila model for the screening of bioavailable NADPH oxidase inhibitors following carbon dioxide anaesthesia. Sci. Rep. 5, 15298 (2015).
and antioxidants. Mol. Cell. Biochem. 352, 91–98 (2011).
45. Abolaji, A. O., Babalola, O. V., Adegoke, A. K. & Farombi, E. O. Hesperidin, a
citrus bioflavonoid, alleviates trichloroethylene-induced oxidative stress in
Drosophila melanogaster. Environ. Toxicol. Pharmacol. 55, 202–207 (2017).
Acknowledgements
This work was created within a research project of the Austrian Competence Centre for
46. Amcheslavsky, A., Jiang, J. & Ip, Y. T. Tissue damage-induced intestinal stem
Feed and Food Quality, Safety and Innovation (FFoQSI, grant number 850681). The
cell division in Drosophila. Cell Stem Cell 4, 49–61 (2009).
COMET-K1 Competence Centre FFoQSI is funded by the Austrian ministries BMVIT,
47. Zhang, G., Gu, Y. & Dai, X. Protective effect of bilberry anthocyanin extracts
BMDW and the Austrian provinces Lower Austria, Upper Austria and Vienna within the
on dextran sulfate sodium-induced intestinal damage in Drosophila
scope of COMET—Competence Centers for Excellent Technologies. The program
melanogaster. Nutrients 14, 2875 (2022).
COMET is handled by the Austrian Research Promotion Agency FFG. This research was
48. Zhang, J., Lei, H., Hu, X. & Dong, W. Hesperetin ameliorates DSS-induced
also supported by a PhD grant of the University of Applied Sciences Upper Austria
colitis by maintaining the epithelial barrier via blocking RIPK3/MLKL
which is funded by the Austrian province Upper Austria and the Austrian Research
necroptosis signaling. Eur. J. Pharmacol. 873, 172992 (2020).
Promotion Agency FFG (project number 881301), and by the Christian Doppler
49. Bodet, C., La, V. D., Epifano, F. & Grenier, D. Naringenin has anti-
Gesellschaft (Josef Ressel Center for Phytogenic Drug Research).
inflammatory properties in macrophage and ex vivo human whole-blood
models. J. Periodontal Res. 43, 400–407 (2008).
50. Xagorari, A. et al. Luteolin inhibits an endotoxin-stimulated phosphorylation
cascade and proinflammatory cytokine production in macrophages. J. Author contributions
Pharmacol. Exp. Ther. 296, 181–187 (2001). Conceptualization: A.K., N.S., M.D., A.M., R.T., D.K., and Ju.W.; investigation: A.K.,
51. Comalada, M. et al. Inhibition of pro-inflammatory markers in primary bone N.S., M.D., B.S., C.N., M.W., Ja.W., V.S.; data curation: A.K., N.S., Ju.W.; wri-
marrow-derived mouse macrophages by naturally occurring flavonoids: ting―original draft preparation: A.K., N.S., and Ju.W.; writing―review and editing:
analysis of the structure-activity relationship. Biochem. Pharmacol. 72, Ju.W., A.K., N.S.; supervision: Ju.W.; funding acquisition: Ju.W. All authors have read
1010–1021 (2006). and agreed to the published version of the manuscript.

Competing interests Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
The authors declare the following competing interests: TriPlant AG provided support in published maps and institutional affiliations.
the form of salaries for the authors A.M., D.K. and R.T. but did not have any additional
role in the study design. TriPlant AG had no role in the design of the study; in the
collection, analysis, or interpretation of data; in the writing of the manuscript, or in the Open Access This article is licensed under a Creative Commons
decision to publish the results.
Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as long as you give
Additional information appropriate credit to the original author(s) and the source, provide a link to the Creative
Supplementary information The online version contains supplementary material Commons licence, and indicate if changes were made. The images or other third party
available at https://doi.org/10.1038/s42003-023-05424-7. material in this article are included in the article’s Creative Commons licence, unless
indicated otherwise in a credit line to the material. If material is not included in the
Correspondence and requests for materials should be addressed to Julian Weghuber. article’s Creative Commons licence and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from
Peer review information Communications Biology thanks Young Sung Jung, Ewa the copyright holder. To view a copy of this licence, visit http://creativecommons.org/
Huszcza and the other, anonymous, reviewer(s) for their contribution to the peer review licenses/by/4.0/.
of this work. Primary Handling Editors: Cesar de la Fuente and Joao Valente.
Reprints and permission information is available at http://www.nature.com/reprints © The Author(s) 2023

Combined Acid Hydrolysis and Fermentation Improves Bioactivity of Citrus Flavonoids in Vitro and in Vivo

Uploaded by

Copyright:

Available Formats

Combined Acid Hydrolysis and Fermentation Improves Bioactivity of Citrus Flavonoids in Vitro and in Vivo

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Combined Acid Hydrolysis and Fermentation Improves Bioactivity of Citrus Flavonoids in Vitro and in Vivo

Uploaded by

Copyright:

Available Formats

ARTICLE

Combined acid hydrolysis and fermentation

health. Unfortunately, the bioavailability and bioactivity of phytochemicals such as poly-

COMMUNICATIONS BIOLOGY | (2023)6:1083 | https://doi.org/10.1038/s42003-023-05424-7 | www.nature.com/commsbio 1

2 COMMUNICATIONS BIOLOGY | (2023)6:1083 | https://doi.org/10.1038/s42003-023-05424-7 | www.nature.com/commsbio

COMMUNICATIONS BIOLOGY | (2023)6:1083 | https://doi.org/10.1038/s42003-023-05424-7 | www.nature.com/commsbio 3

4 COMMUNICATIONS BIOLOGY | (2023)6:1083 | https://doi.org/10.1038/s42003-023-05424-7 | www.nature.com/commsbio

COMMUNICATIONS BIOLOGY | (2023)6:1083 | https://doi.org/10.1038/s42003-023-05424-7 | www.nature.com/commsbio 5

6 COMMUNICATIONS BIOLOGY | (2023)6:1083 | https://doi.org/10.1038/s42003-023-05424-7 | www.nature.com/commsbio

COMMUNICATIONS BIOLOGY | (2023)6:1083 | https://doi.org/10.1038/s42003-023-05424-7 | www.nature.com/commsbio 7

8 COMMUNICATIONS BIOLOGY | (2023)6:1083 | https://doi.org/10.1038/s42003-023-05424-7 | www.nature.com/commsbio

COMMUNICATIONS BIOLOGY | (2023)6:1083 | https://doi.org/10.1038/s42003-023-05424-7 | www.nature.com/commsbio 9

10 COMMUNICATIONS BIOLOGY | (2023)6:1083 | https://doi.org/10.1038/s42003-023-05424-7 | www.nature.com/commsbio

COMMUNICATIONS BIOLOGY | (2023)6:1083 | https://doi.org/10.1038/s42003-023-05424-7 | www.nature.com/commsbio 11

12 COMMUNICATIONS BIOLOGY | (2023)6:1083 | https://doi.org/10.1038/s42003-023-05424-7 | www.nature.com/commsbio

COMMUNICATIONS BIOLOGY | (2023)6:1083 | https://doi.org/10.1038/s42003-023-05424-7 | www.nature.com/commsbio 13

14 COMMUNICATIONS BIOLOGY | (2023)6:1083 | https://doi.org/10.1038/s42003-023-05424-7 | www.nature.com/commsbio

COMMUNICATIONS BIOLOGY | (2023)6:1083 | https://doi.org/10.1038/s42003-023-05424-7 | www.nature.com/commsbio 15

16 COMMUNICATIONS BIOLOGY | (2023)6:1083 | https://doi.org/10.1038/s42003-023-05424-7 | www.nature.com/commsbio

COMMUNICATIONS BIOLOGY | (2023)6:1083 | https://doi.org/10.1038/s42003-023-05424-7 | www.nature.com/commsbio 17

Reprints and permission information is available at http://www.nature.com/reprints © The Author(s) 2023

18 COMMUNICATIONS BIOLOGY | (2023)6:1083 | https://doi.org/10.1038/s42003-023-05424-7 | www.nature.com/commsbio

You might also like