Identification of a cytotoxic molecule in heat-modified citrus pectin

Article abstract of DOI:10.1016/j.carbpol.2015.10.055

Modified forms of citrus pectin possess anticancer properties. However, their mechanism of action and the structural features involved remain unclear. Here, we showed that citrus pectin modified by heat treatment displayed cytotoxic effects in cancer cells. A fractionation approach was used aiming to identify active molecules. Dialysis and ethanol precipitation followed by HPLC analysis evidenced that most of the activity was related to molecules with molecular weight corresponding to low degree of polymerization oligogalacturonic acid. Heat-treatment of galacturonic acid also generated cytotoxic molecules. Furthermore, heat-modified galacturonic acid and heat-fragmented pectin contained the same molecule that induced cell death when isolated by HPLC separation. Mass spectrometry analyses revealed that 4,5-dihydroxy-2-cyclopenten-1-one was one cytotoxic molecule present in heat-treated pectin. Finally, we synthesized the enantiopure (4R,5R)-4,5-dihydroxy-2-cyclopenten-1-one and demonstrated that this molecule was cytotoxic and induced a similar pattern of apoptotic-like features than heat-modified pectin.

Full text of DOI:10.1016/j.carbpol.2015.10.055

Carbohydrate Polymers 137 (2016) 39–51
Contents lists available at ScienceDirect
Carbohydrate Polymers
journal homepage: www.elsevier.com/locate/carbpol
Identification of a cytotoxic molecule in heat-modified citrus pectin
Lionel Leclerea, Maude Fransoleta, Pierre Cambierb, Sandy El Bkassinyc, Abdellatif Tikadc,
Marc Dieua, Stéphane P. Vincentc, Pierre Van Cutsemb, Carine Michielsa,∗
a Laboratory of Biochemistry and Cellular Biology-URBC, NARILIS, University of Namur, 61 rue de Bruxelles, 5000 Namur, Belgium
b Laboratory of Plant Cellular Biology-URBV, University of Namur, 61 rue de Bruxelles, 5000 Namur, Belgium
c Organic Chemistry Research Unit (UCO), University of Namur, 61 rue de Bruxelles, 5000 Namur, Belgium
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 18 June 2015
Received in revised form 13 October 2015
Accepted 14 October 2015
Available online 19 October 2015
Modified forms of citrus pectin possess anticancer properties. However, their mechanism of action and
the structural features involved remain unclear. Here, we showed that citrus pectin modified by heat
treatment displayed cytotoxic effects in cancer cells. A fractionation approach was used aiming to iden-
tify active molecules. Dialysis and ethanol precipitation followed by HPLC analysis evidenced that most
of the activity was related to molecules with molecular weight corresponding to low degree of polymer-
ization oligogalacturonic acid. Heat-treatment of galacturonic acid also generated cytotoxic molecules.
Furthermore, heat-modified galacturonic acid and heat-fragmented pectin contained the same molecule
that induced cell death when isolated by HPLC separation. Mass spectrometry analyses revealed that
4,5-dihydroxy-2-cyclopenten-1-one was one cytotoxic molecule present in heat-treated pectin. Finally,
we synthesized the enantiopure (4R,5R)-4,5-dihydroxy-2-cyclopenten-1-one and demonstrated that this
molecule was cytotoxic and induced a similar pattern of apoptotic-like features than heat-modified
pectin.
Keywords:
Cancer
Cytotoxic activity
Fragmented pectin
4,5-Dihydroxy-2-cyclopenten-1-one
© 2015 Elsevier Ltd. All rights reserved.
1. Introduction
ities against some aggressive and recurrent cancers (Leclere, Van
Cutsem, & Michiels, 2013).
Despite the development of “smart” or targeted drugs as new
therapeutic approach for cancer patients, refractory disease still
represents a high hurdle hampering successful chemotherapy.
More than 60% of anti-cancer drugs in use nowadays are derived
from natural products (Cragg & Newman, 2005, 2013). Vinca-
alkaloids, camptothecin, etoposide or taxanes are such examples.
Successful remission can be achieved with chemotherapy in most
cases but refractory diseases and relapses remain a major obsta-
cle. Indeed, resistance to chemotherapy develops and is still a
high hurdle to get over for efficient cancer patient treatment.
New molecules are thus needed that could be used alone or in
combination with currently used drugs to enhance therapeutic suc-
cess. Among promising avenues, pectin and pH- or heat-modified
pectin have demonstrated chemopreventive and antitumoral activ-
Pectins are abundant and complex polysaccharides present in
the primary cell wall of plants. It is composed of homogalacturo-
nan (HG), substituted galacturonans, rhamnogalacturonan-I (RG-I)
and rhamnogalacturonan-II (RG-II). HG is a polymer of ␣-1,4-
linked-galacturonic acid. HG residues can be methyl-esterified at
the C-6 carboxyl or acetylated at the O-2 or O-3 depending from
the pectin source. Side chains of RG-I and RG-II, whose structure
can be very complex, are attached to a backbone of HG (Harholt,
Suttangkakul, & Vibe Scheller, 2010; Mohnen, 2008). Citrus pectin
consists mainly of HG, lower amount of RG-I and very few RG-II
structural domains (Yapo, Lerouge, Thibault, & Ralet, 2007). Acidic
pH has been used to modify pectin and generate lower molecular
weight fragments able to compete with galectin-3 (Gao et al., 2012).
Galectin-3 is a lectin that possesses a conserved carbohydrate-
binding domain that specifically recognizes ß-galactoside moieties.
It has been described to be overexpressed in metastatic cancers
(de Oliveira et al., 2010) and promotes cell migration and survival
(Fortuna-Costa, Gomes, Kozlowski, Stelling, & Pavao, 2014). Several
works have demonstrated its anti-cancer activities both in vitro and
in vivo (Chauhan et al., 2005; Inohara & Raz, 1994; Nangia-Makker
et al., 2002).
∗
Corresponding author. Tel.: +32 81724131; fax: +32 81724135.
E-mail addresses: lionel.leclere@slbo.be (L. Leclere),
maude.fransolet@unamur.be (M. Fransolet), pierre.cambiuer@unamur.be
(P. Cambier), sandy.elbkassiny@unamur.be (S. El Bkassiny),
abdellatif.tikad@unamur.be (A. Tikad), marc.dieu@unamur.be (M. Dieu),
stephane.vincent@unamur.be (S.P. Vincent), pierre.vancutsem@unamur.be
(P. Van Cutsem), carine.michiels@unamur.be (C. Michiels).
In another study, pectin can also be modified by heat treat-
ment. Different polysaccharide fractions isolated from ginseng rich
http://dx.doi.org/10.1016/j.carbpol.2015.10.055
0144-8617/© 2015 Elsevier Ltd. All rights reserved.

40
L. Leclere et al. / Carbohydrate Polymers 137 (2016) 39–51
in HG stopped cell cycle in G2/M phase on colon cancer HT-29
cells. On the other hand, fractions rich in HG and modified by heat
treatment exerted a much higher anti-proliferative activity, which
was accompanied by caspase-3 activation and apoptosis induction
(Cheng et al., 2011). Similarly, rhamnogalacturonan I domain-rich
pectin isolated from potato inhibited in vitro HT-29 cell prolifera-
tion and provoked a cell cycle arrest in G2/M phase. This inhibition
was due to a decrease in cyclin B1 expression and in CDK-1 activity
(Cheng et al., 2013). Kang et al. (2006) also produced a citrus pectin-
derived oligosaccharide, which was biologically active, but they
used irradiation instead of heat treatment. Pectin irradiated with
20 kGy and then dialyzed (molecular weight <10,000) inhibited
cancer cell growth.
Jackson et al. (2007) showed that different treatment proto-
cols of pectin can lead to differences in pectin apoptosis-inducing
activity and that heat-fragmented pectin has a cytotoxic effect in
galectin-3 expressing but also in non-expressing prostate cancer
cells. This indicates that the active molecules in pH-modified and
heat-treated pectin are not the same. Treatment of heat-modified
pectin with pectinmethylesterase to cleave galacturonosyl car-
boxymethylesters and/or with endopolygalacturonase to remove
non-methylesterified HG did not change the pro-apoptotic activity
of the modified pectin. On the other hand, mild base treatment that
eliminated the ester bounds destroyed the pro-apoptotic activity.
Cytotoxic activity thus requires a base-sensitive linkage in oligosac-
charides other than a carboxymethylester bound. Size analyses of
the active fragments suggested low mass (10–20 kDa) oligosaccha-
rides (Jackson et al., 2007). Recently, we showed that citrus pectin
modified by heat treatment induced cell death in HepG2 and A549
cells. The induced cell death differs from classical apoptosis since
a different pattern of caspase 3 cleavage was detected. Autophagy
was also induced (Leclere et al., 2015).
10270), and the cells were kept at 37 ◦C in an atmosphere of
95% air and 5% CO2. MCF10A cells were cultured in DMEM/F12
medium (Gibco 11320-074) containing 5% horse serum (Gibco
16050-122), 20 ␮g/mL EGF, 0.5 ␮g/mL hydrocortisone, 10 ␮g/mL
insulin and 100 ng/mL cholera toxin. They were maintained in cul-
ture in 75 cm2 polystyrene flasks (Corning) with Basal Medium
Eagle (BME, Invitrogen), supplemented with L-glutamine (2 mM,
supplied by Sigma-Aldrich). For routine culture, both media were
supplemented with 10% fetal bovine serum (Gibco 10270) and cells
were kept at 37 ◦C in an atmosphere of 95% air and 5% CO2. For treat-
ment, cells were allowed to adhere for 24 h after seeding. Media
were discarded, cells washed twice with PBS (Lonza BE17-516F)
and placed in their corresponding medium, without serum, con-
taining the following: 3 mg/mL of filter-sterilized HFCP, 3 mg/mL of
filter-sterilized citrus pectin or 50 ␮M etoposide, used as a positive
control. Negative controls were cells incubated in media alone.
2.3. Cell viability assay
HepG2 cells were seeded at 50 000 cells/well in 24-well
plates before treatments for 24 h. MTT (3-[4,5-dimethylthiazol-2-
yl]-2,5-diphenyltetrazolium bromide, Sigma M2128) solution was
prepared at a concentration of 2.5 mg/mL in phosphate-buffered
saline and 500 ␮L were added per well. After 2 h at 37◦C and 5% CO2
atmosphere, media and MTT solution were removed before adding
lysis buffer. Optical density was measured 1 h later at 570 nm using
a microplate spectrophotometer (Biorad × Mark Microplate spec-
trophotometer) and Microplate manager 6 software. The MTT test
measures the number of metabolically active (viable) cells.
2.4. Cell cytotoxicity assay
However, besides these different studies, the nature of the active
molecules in both products is not known. In this work, we used
a fractionation approach of heat-modified citrus pectin aiming to
identify molecules exerting a cytotoxic activity toward cancer cells.
We report herein the isolation, synthesis and structural analysis of
one such molecule, 4,5-dihydroxy-2-cyclopenten-1-one.
HepG2 cells were seeded at 50 000 cells/well in 24-well plates
before incubation for 24 or 48 h. For each well, lactate dehydroge-
nase activity was measured in the supernatant, in the detached cells
and in adherent cells after lysis in PBS containing 10% Triton X100
(Merck 9036-19-5). Lactate dehydrogenase activity was detected
by colorimetric assay using a cytotoxicity kit (Roche 11644 793
001) and a microplate spectrophotometer. Cytotoxicity percent-
ages were calculated by the ratio of the quantity of LDH present
in the supernatant and in detached cells on the total quantity of
LDH as in the formula: 100 × (a + b)/(a + b + c) where a = supernatant
LDH; b = detached cells LDH; c = adherent cells LDH. 0% LDH release
means that all cells are viable while 100% LDH release indicates that
all the cells are dead.
2. Material and methods
2.1. Fractionation of citrus pectin by heat treatment
Heat fragmented citrus pectin (HFCP) was obtained according
to the method described by Jackson et al. (2007). A solution of 0.1%
citrus pectin (Sigma P9135) composed (on a dry weight basis) of
74% homopolygalacturonic acid, 8.7% neutral sugars, 6.7% methoxy
groups and 2.5% Na, was heated at pH 4.2 for 60 min at 121 ◦C under
a pressure of 1 bar. The solution was then frozen at −80 ◦C and
lyophilized. The dry material was stored at 4◦C. Galactose, galac-
turonic acid, glucose and glucuronic acid monomers (1 g/l, Sigma)
were heat-treated at pH 2.9 for 7.5 h.
2.5. Western blot analysis
Cell lysates were prepared in lysis buffer (40 mM Tris; pH 7.5,
150 mM KCl, 1 mM EDTA, 1% Triton X-100) containing a protease
inhibitor cocktail (Complete from Roche Molecular Biochemicals;
1 tablet in 2 mL of H2O, added at a 1:25 dilution) and phosphatase
inhibitor buffer (25 mM NaVO3, 250 mM PNPP, 250 mM ␣-
glycerophosphate and 125 mM NaF, at a 1:25 dilution). The medium
was centrifuged, and pelleted cells were added to cell lysates.
The lysates were then centrifuged at 12,000 × g for 5 min, and the
supernatants were collected. The proteins (15 ␮g) were denatured
with the addition of LDS sample buffer (Invitrogen NP0007) and
heated to 70 ◦C for 10 min. The proteins were resolved on a 4–12%
NuPAGE (Invitrogen) gel and transferred to a low-fluorescence
membrane (Millipore IPFL00010). The membranes were kept for
1 h in LiCor blocking solution and probed overnight with either
an anti-caspase-3 rabbit antibody (Cell Signaling #9662S) that
recognizes the full-length and cleaved forms of caspase-3 at a
dilution of 1/500, an anti-PARP mouse antibody (BD Biosciences
#551025) at a dilution of 1/1000 or a mouse anti-ubiquitin antibody
2.2. Cell culture and pectin incubation
HepG2, A549, A431, HeLa, MDA-MB-231 and MCF10A cells were
obtained from the American Type Culture Collection HepG2 cells
were cultured in DMEM medium (Gibco 31825-023), A549 cells in
MEM medium (Gibco 41090-028), A431 cells in DMEM high glucose
and pyruvate (Gibco 41966-029), Hela cells in MEM, HEPES, Gluta-
MAX supplement (Gibco 42360-024) supplemented with sodium
pyruvate (100 mM, Gibco 11360-039) and MEM non-essential
amino acids solution (100x, Gibco 11140-035) and MDA-MB-231
cells in RPMI 1640 medium (Gibco 21875-034). For routine cul-
ture, media were supplemented with 10% fetal bovine serum (Gibco

L. Leclere et al. / Carbohydrate Polymers 137 (2016) 39–51
41
(LifeSensor VU101) at a dilution of 1/1000. All antibodies were
diluted in LiCor solution containing 0.1% Tween 20 (Sigma P1379).
The membranes were rinsed with PBS + 0.1% Tween 20 and incu-
bated with fluorochrome-conjugated anti-mouse or anti-rabbit
antibodies at a dilution of 1/10,000 for 1 h in the dark at room tem-
perature. The membranes were rinsed with PBS + 0.1% Tween 20
and dried before being scanned and analyzed using Odyssey soft-
ware. For anti-ubiquitin labeling, the membranes were washed in
PBS after transfer, fixed in PBS containing 0.5% glutaraldehyde (pH
7) and rinsed three times in PBS before blocking. The membranes
were rinsed once more with PBS + 0.1% Tween 20 and dried before
being scanned and analyzed using Odyssey software.
(4 × 250 mm). NaOAc gradient and constant 0.1 M NaOH at a flow
rate of 1 mL/min were used under constant helium pressure as the
mobile phase.
2.8.2. Reverse phase chromatography
Separation of heat-modified galacturonic acid and HFCP and
collection of peak at 3.9–4.4 min was performed with High
Performance Reverse Phase Chromatography Dionex ICS 3000
(Sunnyvale, USA) equipped with UV detection at 210 nm. Chro-
matographic separation of 50 ␮L samples was performed on
Alltima HP C18 column (4.6 mm × 250 mm) (Grace Davison Discov-
ery Sciences). Constant elution of aqueous 0.1% formic acid at a flow
rate of 1 mL/min was used under constant helium pressure as the
mobile phase.
2.6. Dialysis
Data analysis was performed with Chroméléon (Dionex-Thermo
Fisher, USA) version 6.70 SP3.
Regenerated cellulose dialysis membranes Spectra/Por (Spec-
trumLab, USA) with cut off of 3500 or 6–8000 Da were used. Before
being used, membranes were bathed in distilled H2O at 100 ◦C for
10 min then rinsed with cold distilled H2O. Membranes were then
filled with HFCP, clipped and bathed in distilled H2O (volume 10
fold larger than the sample volume) at 4 ◦C, under light agitation.
The dialysis solution was changed every 2 h for 6 h. Both in and
out solutions were then frozen at −80 ◦C and lyophilized. The dry
material was stored at 4 ◦C.
2.9. Mass spectrometry analysis
Samples were analyzed by mass spectrometry, using an ESI-
QTOF maXis 4G UHR-TOF (Bruker). Direct injection was performed
by a syringe pump at a flow rate of 3 ␮L/min. MS spectra were
acquired for 1 s in the m/z range between 50 and 400 m/z. Raw
data were processed using DataAnalysis 4.0 and smart formula 3D
(Bruker). Smart Formula 3D, a formula generation software based
on the combination of MS and MSMS information, was used to gen-
erate the best possible chemical structure from one m/z ratio and
its MS/MS spectrum.
2.7. Precipitation with ethanol
500 mL aqueous pectin solution at 1 mg/mL were heat treated,
then concentrated by evaporation at 4 ◦C up to a volume of about
200 mL. Ethanol was then added to reach a final concentration of
70% or 85%. Precipitation took place overnight at 4 ◦C. The samples
were then centrifuged at 1000 rpm for 10 min at 4 ◦C. The super-
natant was recovered and evaporated up to half the volume. Then
distilled H2O was added to eliminate the remaining ethanol and
the solution was lyophilized.
2.10. 4,5-Dihydroxy-2-cyclopenten-1-one synthesis
The synthesis of (4R,5R)-4,5-dihydroxy-2-cyclopentenone 8 has
been achieved through a 7 step synthetic sequence starting from
the commercially available d-ribose 1 (Fig. 1). The protection of 1
by an isopropylidene afforded the furanoside 2 which was trans-
formed into 3 via iodination of the primary alcohol in 2. The
reductive elimination of 3, using Zinc under acidic conditions gen-
erated the acyclic intermediate 4. The addition of vinyl Grignard
reagent on the volatile aldehyde 4 provided a second olefin nec-
essary for the formation of the cycle. Ring closing metathesis of
the dienes afforded the cyclopentenol 6 which was oxidized by
the Dess–Martin reagent to form the cyclopentenone 7. The ace-
tonide in 7 was then cleaved under acidic conditions affording
target molecule 8 in 20% yield over 7 steps.
2.8. HPLC analysis
2.8.1. Ion exchange chromatography
High Performance Anion Exchange Chromatography with
Pulsed Amperometric Detection (HPAEC–PAD) was performed
using a Dionex ICS 3000 (Sunnyvale, USA) in order to identify
and compare the degree of polymerization of polygalacturonic
acid and HFCP. Chromatographic separation of 50 ␮L polysaccha-
ride samples was performed on a Dionex CarboPac PA-100 column
Fig. 1. Synthesis of molecule 8.

42
L. Leclere et al. / Carbohydrate Polymers 137 (2016) 39–51
2.10.1. Materials and methods
and concentrated. The residue was purified through silica gel using
Cyclohexane: EtOAc 9:1 as eluent to afford the product 3 (1 g, 68%)
as an oil. The analytical data were identical to literature data.
1H NMR (400 MHz, CDCl3): ı = 5.05 (s, 1H), 4.77 (d, J = 5.8 Hz, 1H),
4.63 (d, J = 5.8 Hz, 1H), 4.44 (dd, J = 10.1, J = 5.9 Hz, 1H), 3.37 (s, 3H),
3.28 (dd, J = 5.9, J = 9.9 Hz, 1H), 3.16 (t, J = 10.1 Hz, 1H), 1.48 (s, 3H),
1.33 (s, 3H).
All reactions were carried out under an argon atmosphere. Yields
refer to chromatographically and spectroscopically homogeneous
materials. Reagents and chemicals were purchased from Sigma-
Aldrich or Acros at ACS grade and were used without purification.
All reactions were performed using purified and dried solvents:
tetrahydrofuran (THF) was refluxed over sodium-benzophenone,
dichloromethane (CH2Cl2) was refluxed over calcium hydride
(CaH2). All reactions were monitored by thin-layer chromatogra-
phy (TLC) carried out on Merck aluminum roll silica gel 60-F254
using UV light and a molybdate-sufuric acid solution as revelator.
Merck silica gel (60, particle size 0.040–0.063 mm) was employed
for flash column chromatography and preparative thin layer chro-
matography using technically solvent distilled prior to use as
eluting solvents. NMR spectra were recorded on a JEOL ECX 400
with solvent peaks as reference. All compounds were characterized
by 1H and 13C NMR as well as by 1H–1H and 1H–13C correla-
tion experiment when necessary. The following abbreviations are
used to describe the multiplicities: s = singlet, d = doublet, t = triplet,
q = quartet, m = multiplet. Chemical shifts (ı) are reported in ppm
and referenced indirectly to TMS via the solvent (or residual sol-
vent) signals.
13C NMR (100 MHz, CDCl3): ı = 112.8, 109.8, 87.6, 85.5, 83.1,
55.4, 26.5, 25.1, 6.8.
2.10.4.
4,5-Dideoxy-2,3-O-isopropylidene-D-erythro-4-pentenose 4
(Chen et al., 2012; Palmer & Jager, 2001)
To a stirred solution of 3 (500 mg, 1.6 mmol, 1.0 equiv.) in
MeOH (5 mL) was added Zn powder (115 mg, 1.76 mmol, 1.1 equiv.,
Aldrich, dust, <10 ␮m) in one batch. After addition of AcOH (11 ␮L,
0.19 mmol, 0.12 equiv.), the resulting mixture was heated to reflux
for 4 h. Then, the reaction was cooled down to RT and filtered
through a plug of celite followed by washing the cake with a mix-
ture of THF/hexane (1:1, 3 mL) and MeOH (100 mL). The filtrate was
then concentrated in vacuo. The aldehyde (liquid oil) was volatile,
so the pressure was kept above 20 mbar. The residue was used in
the next step without further purification.
2.10.2. 1-Methoxy-2,3-O-isopropylidene-˛/ˇ-D-ribofuranoside
(2˛/2ˇ)
1H NMR (400 MHz, CDCl3): ı = 9.56 (d, J = 3.2 Hz, 1H), 5.76
(ddd, J = 6.8, J = 10.3, J = 17.2 Hz, 1H), 5.47 (d, J = 17.2 Hz, 1H), 5.33
(d, J = 10.5 Hz, 1H), 4.86 (t, J = 6.8, J = 7.5 Hz, 1H), 4.42 (dd, J = 3.2,
J = 7.5 Hz, 1H), 1.63 (s, 3H), 1.45 (s, 3H).
The anomeric mixture 2␣/2␤ (1.09 g, 82%) was prepared from
d-ribose 1 (1 g, 6.66 mmol) according to the procedure of Klepper,
Jahn, Hickmann, and Carell (2007).
13C NMR (100 MHz, CDCl3): ı = 200.7, 131.3, 119.7, 111.3, 82.2,
79.1, 27.4, 25.3.
1H NMR (400 MHz, CDCl3): ı = 4.98 (s, 1H␤), 4.85 (d, J = 5.9 Hz,
1H␤), 4.63 (d, J = 5.5 Hz, 1H␣), 4.58 (d, J = 5.9 Hz, 1H␤), 4.49–4.47 (m,
1H␣), 4.43 (m, 1H␤), 4.40–4.37 (m, 1H␣), 3.84–3.6 (m, 5H, H␤␣), 3.46
(s, 3H␣), 3.44 (s, 3H␤), 3.25–3.22 (dd, J = 10.7, J = 2.6 Hz, 1H␤), 1.55
(s, 3H␣), 1.48 (s, 3H␤), 1.37 (s, 3H␣), 1.32 (s, 3H␤).
2.10.5. (4S,5R)-1-(2,2-Dimethyl-5-vinyl-[1,3]dioxolan-4-yl)-
prop-2-en-1-ol 5 (Smith, Han, Breslin, & Beauchamp, 2005; Yang,
Ye, & Schneller, 2004)
2.10.3. Methyl
5-Deoxy-5-iodo-2,3-O-isopropylidene-ˇ-d-ribofuranoside (Chen,
Ye, lLiu, & Schnelle, 2012; Palmer & Jager, 2001)
To a stirred solution of 4 (100 mg, 0.64 mmol) in dry THF
(0.7 mL) was added vinylmagnesium bromide (0.75 mL, 0.75 mmol,
1.2 equiv., 1 M in THF) at −78 ◦C. After addition, the reaction
was stirred for 10 min at the same temperature then for 1 h at
0 ◦C, before saturated solution of ammonium chloride (10 mL) and
brine (10 mL) were added. The aqueous layer was separated and
extracted with ethyl acetate (3 × 20 mL). The combined organic
solutions were dried over MgSO4 and concentrated under reduced
pressure to give a mixture of diastereoisomers 5a/5b (d.r. = 65:35),
which was used in the next step without further purification. The
analytical data were identical to literature data (Yang et al., 2004).
To a solution of 2 (1 g, 4.9 mmol, 1.0 equiv.) in a mixture
of MeCN/toluene 1:1 (16 mL) was added imidazole (500 mg,
7.35 mmol, 1.5 equiv.) and triphenylphosphine (1.4 g, 5.4 mmol,
1.1 equiv.). After the addition of I2 (1.4 g, 5.4 mmol, 1.1 equiv.) in
portions, the reaction was stirred at RT for 2 h. TLC (Cyclohex-
ane/EtOAc 8:2) indicated the formation of the major product and
the presence of starting material. An additional addition of imid-
azole (500 mg, 7.35 mmol, 1.5 equiv.), triphenylphosphine (1.4 g,
5.4 mmol, 1.1 equiv.) and I2 (1.4 g, 5.4 mmol, 1.1 equiv.) was per-
formed and the reaction was stirred for 1 h at RT. Water (12 mL)
and sodium thiosulfate (728 mg) were then added until the solution
became clear. The organic layer was separated, dried over MgSO4
2.10.6. (4R,5S)-4,5-O-Isopropylidenecyclopenten-1-ol 6 (Chen
et al., 2012; Smith et al., 2005; Yang et al., 2004)

L. Leclere et al. / Carbohydrate Polymers 137 (2016) 39–51
43
manner after 24 h of incubation, as measured by a LDH release assay
(Fig. 2A). On the other hand, unmodified pectin did not affect cell
viability. At 3 mg/mL, HFCP induced cell death to a higher extent
than etoposide at 50 ␮M, which was used as a positive control. It
must be noted that etoposide is still much more potent in elici-
ting apoptosis in comparison with HFCP when analyzed mass-wise
(etoposide 50 ␮M = approximately 30 ␮g/mL in comparison to the
3.0 mg/mL of HFCP in Fig. 2A). Its effect was also time-dependent
and observed in two other cancer cell lines (Leclere et al., 2015). The
concentration of 3 mg/mL was chosen as the reference for the fol-
lowing experiments. These effects were confirmed in a MTT assay
assessing the number of viable cells (Fig. 2B). A kinetic analysis
has also been performed (Fig. 2C). The results showed that a strong
cytotoxic effect was observed for HFCP, but not for untreated pectin
already after 24 h of incubation, similarly to what was observed for
etoposide.
A solution of compound 5 (120 mg, 0.65 mmol) in fresh dis-
tilled dichloromethane (1.56 mL) was bubbled with N2 for 15 min to
remove O2. To this solution was added under argon Grubbs’s sec-
ond generation catalyst (5.5 mg, 0.0065 mmol, 0.01 equiv.). After
overnight at RT, the solvent was concentrated under reduced pres-
sure to give acetonide 6, which was used in the next step without
further purification. The analytical data were identical to literature
data (Yang et al., 2004).
In order to identify the fraction and/or the molecules in HFCP
that are responsible for its cytotoxic activity, we used a stepwise
approach with different separation techniques and the subsequent
“active” fractions were tested for their cytotoxic activity on HepG2
cells. When pectin is heat-treated, two degradation mechanisms
can occur: ß-elimination and acidic hydrolysis (Kral & McFeeters,
1998). Both lead to pectin fragmentation at the level of the osidic
bounds, generating fragments of different degrees of polymeriza-
tion (DP). We thus attempted to separate the active fragments
according to their size. HFCP was first dialyzed using a membrane
with a cut off of 6–8000 Da. As shown in Fig. 2D, dialyzed HFCP
completely lost its cytotoxic activity. Anion exchange HPLC anal-
ysis of dialyzed HFCP in comparison to HFCP and to a standard
composed of oligogalacturonic acid with DPs lower than 6 (Fig. 2E)
evidenced that dialyzed HFCP contained a mixture of fragments
with a large variation of degree of polymerization (from monomer
to DP ≥ 15). The analysis was performed in the presence of NaOH,
which demethylated the fragments: a preliminary HPLC analy-
sis had shown that all fragments were methylesterified (data not
shown). Comparison between HFCP and dialyzed HFCP indicated
that it was mainly molecules with a DP of 1, i.e. monomers, or
neutral sugars that were lost upon dialysis. In order to confirm
this observation and to investigate whether these low molecular
weight fragments would display the cytotoxic activity, dialysis of
HFCP with a cut off of 3500 Da was performed and the permeat (out
fraction) was recovered and tested for potential activity. Fig. 3A
shows that the permeat contained active molecules that exerted a
concentration-dependent cytotoxic activity in HepG2 cells.
A second approach using ethanol precipitation was developed
to confirm that the active fragments were of low molecular weight.
Increasing the concentration of ethanol added to a solution induces
the precipitation of molecules with lower and lower molecular
weight, leaving the very low molecular weight molecules dissolved
in the supernatant. We showed that the molecules from HFCP
that were still soluble in 70 and 85% ethanol were the active ones
(Fig. 3A), confirming that they were of very low molecular weight.
Indeed, less than 20% of galacturonic acid monomers and less than
5% galactose precipitate in the presence of 85% ethanol (data not
shown).
2.10.7. (4R,5R)-4,5-O-Isopropylidene-2-cyclopentenone 7 (Chen
et al., 2012; Smith et al., 2005; Yang et al., 2004)
To a solution of 6 (101 mg, 0.65 mmol) in dichloromethane
(5.4 mL) at 0 ◦C was added Dess–Martin periodinane (413 mg,
0.975 mmol, 1.5 equiv.). After addition, the reaction was warmed
to room temperature and stirred for 2 h, before saturated NaHCO3
solution (5 mL) and Na2S2O3 solution (2 mL) were added. The aque-
ous layer was separated and extracted with CH2Cl2 (3 × 10 mL). The
combined organic solutions were dried (MgSO4), filtered and con-
centrated under reduced pressure to give the compound 7, which
was used in the next step without further purification. The ana-
lytical data were identical to literature data (Yang et al., 2004).
2.10.8. (4R,5R)-4,5-dihydroxy-2-cyclopentenone 8 (Sugahara,
Fukada, & Iwabuchi, 2004)
To a solution of 7 (100 mg, 0.65 mmol) in dichloromethane
(60 mL) was added TFA/H2O 1:1 (2 mL). After addition, the reaction
was stirred 16 h at RT. Then, the solvent was concentrated in vacuo
and the residue was purified through preparative TLC by using
CHCl3/Acetone 6/4 as eluant to give the final compound (25 mg,
34% over 4 steps) as a liquid oil.
1H NMR (400 MHz, CDCl3): ı = 6.38 (dd, J = 6.2, J = 2.9 Hz, 1H),
5.91–5.77 (m, 2H), 5.00 (d, J = 6.2 Hz, 1H), 2.81 (d, J = 5.5 Hz, 1H).
13C NMR (100 MHz, CDCl3): ı = 170.8, 161.5, 132.9, 71.22, 68.99.
HRMS (ESI+): meas. 137.014 calc. 137.021 for C5H6NaO3.
Since pectin is mainly composed of polygalacturonic acid, we
hypothesized that using oligogalacturonic acid with a DP lower
than 6 would reproduce the effect observed with the super-
natant obtained after ethanol precipitation and of the permeat
obtained after dialysis. However, this was not the case (Fig. 3B):
polygalacturonic acid with a DP lower than 6 did not induce
HepG2 cell death. Of note, polygalacturonic acid with a DP higher
than 7 had no effect either. Two hypotheses can be proposed
to explain this result: either, the cytotoxic activity did not orig-
inate from (poly)galacturonic acid whatever its DP but from
neutral sugars derived from the rhamnogalacuronans of pectin; or
3. Results
3.1. Low molecular fraction of HFCP induces HepG2 cell death
In order to study the cell death-inducing effects of HFCP,
HepG2 cells were exposed to different concentration of HFCP.
HFCP induced cytotoxic effect in
a concentration-dependent

44
L. Leclere et al. / Carbohydrate Polymers 137 (2016) 39–51
Fig. 2. Low molecular fraction of HFCP induces HepG2 cell death. (A) HepG2 cells were incubated with medium alone (CTL), 50 ␮M etoposide (Eto), different concentrations of
heat fragmented citrus pectin (HFCP) or 3 mg/mL citrus pectin (Pectin) for 24 h. Cytotoxicity was assayed using LDH cytotoxicity detection kit and results are expressed in
percentage of total LDH activity. (B) HepG2 cells were incubated with medium alone (CTL), 50 ␮M etoposide (Eto), 3 mg/mL heat fragmented citrus pectin (HFCP) or 3 mg/mL
citrus pectin (Pectin) for 24 h. Cell viability was assayed with MTT assay and results are expressed in number of living cells (% reported to control). (C) HepG2 cells were
incubated with medium alone (CTL), 50 ␮M etoposide (Eto), 3 mg/mL heat fragmented citrus pectin (HFCP) or 3 mg/mL citrus pectin (Pectin) for different incubation time.
Cell viability was assayed with MTT assay. Data are means 1 SD (n = 3). (D) HepG2 cells were incubated with medium alone (CTL), 50 ␮M etoposide (Eto), 3 mg/mL heat
fragmented citrus pectin (HFCP) or 3 mg/mL dialyzed HFCP (diHFCP, cut off 6–8000 Da) for 24 h. Cytotoxicity was assayed using LDH cytotoxicity detection kit. Data are
means 1 SD (n = 3). Statistical analyses performed were ANOVA I with Bonferroni post-hoc tests. p Values in comparison to the control are **: p < 0.01; ***: p < 0.001. (E)
Elution profile from HPLC separation (column PA100 with 0.2 to 0.5 M NaAc gradient and constant 0.1 M NaOH) of HFCP, dialyzed HFCP (diHFCP, cut off 6–8000 Da) or of
oligogalacturonic acid with a degree of polymerization lower than 6 (DP).
(poly)galacturonic acid was chemically modified by heat treatment
to generate the active molecule(s). The second hypothesis seems
to be the correct one since heat-treatment of polygalacturonic acid
with a DP lower than 6 generated molecules with cytotoxic activity
toward HepG2 cells (Fig. 3C).
results showed that there was a peak at 3.9–4.4 min of elution that
was detected in HFCP and heat-treated galacturonic acid monomer
(see arrow in Fig. 5A), which was not observed in untreated pectin
or in untreated galacturonic acid monomer. Moreover, the height
of this peak was increased when pectin was heat-treated for 7.5 h
in comparison to the usual 60 min treatment, suggesting that
increasing the time of treatment generated more of the active
molecule(s). This peak was collected, lyophilized and tested for
its cytotoxic activity in HepG2 cells. The molecule(s) contained
in this sample was(were) more active that heat-modified galac-
turonic acid monomer: at 0.25 mg/mL, these molecules induced a
decrease of 93% in cell viability while heat-modified galacturonic
acid monomer diminished cell viability by 46% (Fig. 5B).
Since heat treatment seemed to be necessary to generate the
cytotoxic activity and since HPLC analyses revealed that molecules
with a molecular weight corresponding to a DP of 1 were con-
tained in the active fraction of HFCP, we investigated whether
heat treatment of galacturonic acid monomer would also gener-
ate cytotoxic molecule(s). Fig. 3D shows that, while galacturonic
acid monomer at 3 mg/mL did not induce HepG2 cell death, heat-
modified galacturonic acid monomer was toxic for HepG2 cells at
1 mg/mL. This was confirmed by measuring the number of viable
cells using a MTT assay: heat-modified galacturonic acid monomer
decreased the number of viable cells in a concentration-dependent
manner after 24 h of incubation (Fig. 4A). It is important to note
that the cytotoxic activity was not generated by heat treatment of
neutral sugars like galactose or glucose. On the other hand, heat-
treated glucuronic acid induced a similar cytotoxic activity than
heat-treated galacturonic acid (Fig. 4B).
3.2. 4,5-Dihydroxy-2-cyclopenten-1-one is one of the active
molecules of HFCP
In order to investigate the structural features of the molecule(s)
contained in heat-treated galacturonic acid monomer, mass spec-
trometry analyses were performed. Mass spectrometry spectra of
heat-treated galacturonic acid monomer in comparison to unmo-
dified galacturonic acid showed the presence of new peaks which
might have derived from the active molecule(s): M/Z = 115, 137,
191, 200 and 281 (see circles in Fig. 5C). The peak with m/z = 137
(Fig. 5D) seemed to be the most interesting one because this
peak was also present in the mass spectrometry analysis of
In order to investigate whether HFCP contained the same
molecule(s) than heat-modified galacturonic acid monomer, both
products were analyzed by HPLC using a C18 column and formic
acid 0.1% in water as the eluent, which allowed a separation
according to the hydrophobic properties of the molecules. The

L. Leclere et al. / Carbohydrate Polymers 137 (2016) 39–51
45
Fig. 3. Heat-modified galacturonic acid induces HepG2 cell death. (A) HepG2 cells were incubated with medium alone (CTL), 50 ␮M etoposide (Eto), 3 mg/mL heat fragmented
citrus pectin (HFCP), 3 mg/mL citrus pectin (Pectin), different concentrations of the permeat obtained after HFCP dialysis (cut off 3500 Da) or different concentration of the
supernatant of HFCP from ethanol precipitation (70 or 85% ethanol) for 24 h. Cytotoxicity was assayed using LDH cytotoxicity detection kit and results are expressed in
percentage of total LDH activity. (B) HepG2 cells were incubated with medium alone (CTL), 50 ␮M etoposide (Eto), 3 mg/mL HFCP, 3 mg/mL polygalacturonic acid with a
degree of polymerization lower than 6 (GalUA DP < 6), 3 mg/mL polygalacturonic acid with a degree of polymerization higher than 7 (GalUA DP > 7) or 3 mg/mL citrus pectin
(Pectin) for 24 h. Cytotoxicity was assayed using LDH cytotoxicity detection kit. (C) HepG2 cells were incubated with medium alone (CTL), 50 ␮M etoposide (Eto), 3 mg/mL
heat-modified polygalacturonic acid with a degree of polymerization lower than 6 (GalUA DP < 6*) or 3 mg/mL citrus pectin (Pectin) for 24 h. Cytotoxicity was assayed using
LDH cytotoxicity detection kit. (D) HepG2 cells were incubated with medium alone (CTL), 50 ␮M etoposide (Eto), 3 mg/mL HFCP, 3 mg/mL pectin, 1 mg/mL heat-modified
galacturonic acid (GalUA*) or 1 and 3 mg/mL galacturonic acid (GalUA) for 24 h. Cytotoxicity was assayed using LDH cytotoxicity detection kit. Data are means 1 SD (n = 3
or 4). Statistical analyses performed were ANOVA I with Bonferroni post-hoc tests. P values in comparison to the control are **: p < 0.01; ***: p < 0.001.
Fig. 4. Heat-modified galacturonic acid heat-modified glucuronic acid induced HepG2 cell death in a concentration-dependent manner. (A) HepG2 cells were incubated
with medium alone (CTL), 1 mg/mL galacturonic acid (GalUA) or different concentration of heat-modified galacturonic acid (GalUA*) for 24 h. Cell viability was assayed with
MTT assay and results are expressed in number of living cells (% reported to control). Data are means 1 SD (n = 3 or 4). Statistical analyses performed were ANOVA I with
Bonferroni post-hoc tests. p Values in comparison to the control are ***: p < 0.001. (B) HepG2 cells were incubated with medium alone (CTL), 50 ␮M etoposide (Eto), 3 mg/mL
citrus pectin (Pectin) or with galactose (Gal), heat-modified galactose (Gal*), glucose (Glu), heat-modified glucose (Glu), galacturonic acid (GalUA), heat-modified galacturonic
acid (GalUA*), glucuronic acid (GluA) or heat-modified glucuronic acid (GluA*), all at 3 mg/mL for 24 h. Cell viability was assayed with MTT assay. Data are means 1 SD (n = 3
or 4). Statistical analyses performed were ANOVA I with Bonferroni post-hoc tests. p Values in comparison to the control are *: p < 0.05; **: p < 0.01.
4,5-dihydroxy-2-cyclopenten-1-one (Fig. 6A), a molecule derived
from heat treatment of uronic acid, uronic acid derivatives or
a substance containing a saccharide compound which contains
uronic acid, such as pectin. This molecule was claimed to possess
anti-cancer activity (Koyama et al., 2000). According to these
authors, 4,5-dihydroxy-2-cyclopenten-1-one presents a m/z value
of 105 when ionized with one H+ and of 137 when ionized with one
Na+. MS/MS analysis of this peak from heat-treated galacturonic

46
L. Leclere et al. / Carbohydrate Polymers 137 (2016) 39–51
Fig. 5. Heat-modified galacturonic acid and HFCP contain the same molecule that induces HepG2 cell death. (A) Elution profile from HPLC separation (C18 reverse phase with 0.1%
formic acid in water as the eluent) of pectin modified during 7.5 h by heat treatment, pectin, HFCP, heat-modified galacturonic acid (GalUA*), galacturonic acid (GalUA) and of
the eluent. The arrow points at the 3.9–4.4 min peak. (B) HepG2 cells were incubated with medium alone (CTL), 50 ␮M etoposide (Eto), 3 mg/mL heat fragmented citrus pectin
(HFCP), 3 mg/mL citrus pectin (Pectin), different concentrations of the molecules contained in the 3.9–4.4 min peak from C18 separation of heat-modified galacturonic acid,
the corresponding volume of the HPLC eluent (formic acid) or different concentration of heat-modified galacturonic acid (GalUA*) or 0.5 mg/mL galacturonic acid (GalUA)
for 24 h. Cell viability was assayed with MTT assay and results are expressed in number of living cells (% reported to control). Statistical analyses performed were ANOVA I
with Bonferroni post-hoc tests. Data are means 1 SD (n = 3 or 4). p Values in comparison to the control are ***: p < 0.001. (C) MS spectra of the 3.9–4.4 min peak from C18
separation of galacturonic acid (GalUA) or heat-modified galacturonic acid (GalUA*) in comparison to the eluent. (D) Zoom of the spectra from (C) on the peak with m/z = 137.
Fig. 6. 4,5-Dihydroxy-2-cyclopenten-1-one is one cytotoxic molecule present in HFCP. (A) Chemical structure of 4,5-dihydroxy-2-cyclopenten-1-one (molecule 8). (B) MS/MS
spectra of molecule 8 in comparison to the solvent (0.1% formic acid).
acid (Fig. 7A) and smart Formula 3D analysis suggested the formula
C5H6NaO3, which corresponds to 4,5-dihydroxy-2-cyclopenten-
1-one ionized with one Na+. Theoretical fragmentation of
4,5-dihydroxy-2-cyclopenten-1-one gives the exact same peaks as
the one observed in the MS/MS spectrum (data not shown). Spec-
tra of MS analysis of HFCP in comparison to untreated pectin are
very difficult to interpret because numerous peaks were detected
(Fig. 7B). However, peaks with m/z values of 115 and 191, as

L. Leclere et al. / Carbohydrate Polymers 137 (2016) 39–51
47
Fig. 7. Structural analysis of HFCP and HFCP-derived molecules. (A) MS/MS analysis of heat-modified galacturonic acid, peak with a m/z = 137. (B) MS/MS analysis of pectin and
of heat-fragmented citrus pectin (HFCP).

48
L. Leclere et al. / Carbohydrate Polymers 137 (2016) 39–51
Fig. 8. 4,5-Dihydroxy-2-cyclopenten-1-one is cytotoxic. (A) HepG2 cells were incubated with medium alone (CTL) or with different concentrations of molecule 8 for 24 h. Cell
viability was assayed with MTT assay and results are expressed in number of living cells (% reported to control). Data are means 1 SD (n = 3). (B) HepG2 cells were incubated
with medium alone (CTL), with 3 mg/mL heat-fragmented citrus pectin (HFCP) or with 80 ␮M of molecule 8 for different incubation time. Cell viability was assayed with
MTT assay. Data are means 1 SD (n = 3). (C) MCF10A cells were incubated with medium alone (CTL) or with different concentrations of molecule 8 for 24 h. Cell viability
was assayed with MTT assay. Data are means 1 SD (n = 6). The IC50 has been determined using Prism software. (D) HepG2 were incubated with medium alone (Ctl-), 50 ␮M
etoposide (Etop), 3 mg/mL heat-fragmented citrus pectin (HFCP), 3 mg/mL citrus pectin (Pectin) or with 80 ␮M of molecule 8 for 24 h. Cleaved caspase-3, cleaved PARP and
protein ubiquitination were detected by western blot analysis.
observed in the profile of heat-treated galacturonic acid, were
revealed in HFCP, which were not present in the profile of untreated
pectin. A peak with m/z of 137 was also detected but with a very
low abundance.
It is to mention that MS analysis of heat-modified galacturonic
acid monomer also revealed a peak with m/z values of 191 in addi-
tion to the 137 one. MS/MS analysis of these two peaks generated
fragments with a m/z value of 134 and 97, which suggested that heat
treatment of galacturonic acid may lead to the formation of dimers
or trimers of 4,5-dihydroxy-2-cyclopenten-1-one. However, this
still needs to be confirmed.
In order to confirm that 4,5-dihydroxy-2-cyclopenten-1-one
could lead to such a fragmentation pattern in MS/MS analysis and to
study its possible cytotoxic activity, this molecule was synthetized
in our laboratory. 1H and 13C NMR analyses confirmed the struc-
ture of the synthetized molecule 8 while MS analysis showed that
the fragmentation of 4,5-dihydroxy-2-cyclopenten-1-one indeed
led to a peak with a m/z value of 137 (Fig. 6B). These results show
that 4,5-dihydroxy-2-cyclopenten-1-one was one of the molecule
generated by heat treatment of pectin and of galacturonic acid.
We then evaluated the cytotoxic activity of 4,5-dihydroxy-
2-cyclopenten-1-one on HepG2 cells. Fig. 8A shows that this
molecule induced a strong decrease in the number of living cells,
in a concentration-dependent manner: a decrease in cell viabil-
ity of about 50% was observed at 80 ␮M, indicating a very potent
molecule. Indeed, in comparison, etoposide provoked a decrease in
cell viability of 27% at 50 ␮M and heat-treated galacturonic acid, a
decrease of 47% at 0.25 mg/mL (Fig. 5A). 0.25 mg/mL of galacturonic
acid corresponds to 1.3 mM.
Then, we wanted to investigate whether these molecules would
display a specific cytotoxic activity toward cancer cells. The cyto-
toxic activity of 4,5-dihydroxy-2-cyclopenten-1-one was assessed
in MCF10A cells in comparison to HepG2 cells. MCF10A cells are
an immortalized, non-transformed epithelial cell line derived from
human fibrocystic mammary tissue. They are often considered as a
normal control for cancer cells. As shown in Fig. 8B, HFCP induced a
toxic effect in MCF10A similar to the one observed for HepG2 cells
(compared with Fig. 2B). A kinetic analysis was performed (Fig. 2C).
The results showed that a strong cytotoxic effect was observed for
molecule 8, similarly to what was observed for HFCP. (4R,5R)-4,5-
dihydroxy-2-cyclopenten-1-one also led to MCF10A cell death to a
similar extent than the one observed in HepG2 cells: a decrease in
cell viability of 80% was observed at 80 ␮M (Fig. 8C).
A more quantitative approach to assess HFCP activity was per-
formed by determining its IC50 using 5 different cancer cell lines
as well as MCF10A. Results showed that the IC50 varied from 0.8
to 3 mg/mL, with no difference between the cancer cell lines and
MCF10A cells (Fig. 9). These results indicate that HFCP is probably
not specific to cancer cells. It has to be noted that etoposide was
also as toxic for MCF10A cells as for HepG2 cells (data not shown)
while etoposide is presently used in clinics to treat several types
of cancer. Similar effects were obtained for another non-cancerous
cell type, BJ-1 dermal fibroblasts (data not shown).
Finally, since HFCP was shown to induce specific features of
apoptotic-cell death, i.e. classical PARP cleavage but atypical cleav-
age pattern of caspase 3 (Leclere et al., 2015), we wanted to
investigated whether would induce cell death by a similar path-
way. For that, PARP and caspase 3 cleavage was studied by western

L. Leclere et al. / Carbohydrate Polymers 137 (2016) 39–51
49
Fig. 9. HFPC IC50 on different cell lines. HepG2 (A), A549 (B), HeLa (C), A431 (D), MDA-MB-231 (E) and MCF10A (F) cells were incubated with medium alone or different
concentrations of heat fragmented citrus pectin (HFCP) for 24h. Cell viability was assayed with MTT assay and results are expressed in number of living cells (% reported to
control). Data are means 1 SD (n = 3). The IC50 has been determined using Prism software.
blot analysis in HepG2 cells incubated in the presence of (4R,5R)-
4,5-dihydroxy-2-cyclopenten-1-one at 80 ␮M in comparison to
3 mg/mL HFCP. (4R,5R)-4,5-dihydroxy-2-cyclopenten-1-one, sim-
ilarly to etoposide and to HFCP induced the cleavage of PARP.
On the other hand, HFCP induced an atypical caspase 3 modifica-
tion since a band of about 60 kDa was observed, which was not
observed in the extract of cells incubated in the presence of etopo-
side. (4R,5R)-4,5-dihydroxy-2-cyclopenten-1-one led to the same
caspase 3 migration pattern than the one induced by HFCP (Fig. 8D).
Because HFCP was shown to trigger caspase 3 activation via pro-
tein ubiquitination and caspase 8 activation (Leclere et al., 2015),
protein ubiquitination was assessed by a western blot analysis.
An early increase in ubiquitination was observed when the cells
were incubated in the presence of HFCP or (4R,5R)-4,5-dihydroxy-
2-cyclopenten-1-one for 6 h but not in the presence of medium
alone, etoposide or unmodified pectin (Fig. 8D).
4,5-dihydroxy-2-cyclopenten-1-one as one molecule present
in the active fraction and chemipure (4R,5R)-4,5-dihydroxy-2-
cyclopenten-1-one synthetized in our laboratory recapitulated the
activity. It has to be noted that this molecule was probably not the
only active molecule present in HFCP and that synergistic effects
might have resulted from their association in this mixture.
The known mechanisms of pectin degradation upon heat
treatment are ß-elimination and acidic hydrolysis (Diez-Dacal &
Perez-Sala, 2012; Kral & McFeeters, 1998). However, these two
types of reactions cannot explain the generation of 4,5-dihydroxy-
2-cyclopenten-1-one 8. Since we noted that HFCP and heat-treated
galacturonic acid momoners were brown and smelling like caramel,
it reminded us of the non-enzymatic browning of fruits upon
storage or heating (Bharate & Bharate, 2012). The presence of galac-
turonic acid favors this process (Ibartz, Garza, & Pagan, 2008). This
phenomenon is due to the Maillard reaction initiated by the con-
densation of a carbonyl group from a reductive sugar by an amine
group from an amino acid or a protein. This reaction involves
several complex and interconnected processes leading to the for-
mation of glucosamines, ketosamines via Amadori rearrangements,
of diketosamines and to the degradation of these compounds,
resulting in the production of hydroxymethylfurfural. Some of
these hydroxymethylfurfurals have toxic effects (Bauer-Marinovic,
Taugner, Florian, & Glatt, 2012). Similarly, Products from the Mail-
lard reaction are able to induce DNA single and double strand
breaks (Hiramoto, Ishihara, Sakui, Daishima, & Kikugawa, 1998), to
inhibit cell proliferation, probably by interfering with microtubules
(Marko et al., 2002) and to display anti-tumoral activities (Marko
et al., 2003).
4. Discussion
Pectin fragmented and modified by pH- or heat-treatment
was shown to have potent anti-cancer effects (Leclere et al.,
2013). A few hints have been unraveled: in pH-modified pectin,
a functional motif of the ␤-1 → 4 galactan fragment, lying in the
terminal residues, is required for its inhibitory activity on galectin-
3 (Gao et al., 2012). On the other hand, low molecular weight
pectin fragments bearing a base-sensitive linkage other than a
carboxymethylester-bound seems to be responsible for their cyto-
toxic activity toward prostate cancer cells (Jackson et al., 2007).
However, the exact structure of the compounds generated in mod-
ified pectin that are responsible for these activities is currently
unknown.
To address this issue, we used several separation methods
and tested the different fractions for their cytotoxic activity on
HepG2 cells. It was determined that acidic sugars with low
degree of polymerization (down to 1, i.e. monomers), mainly
methyl esters, cleaved from homopolygalacturonans and chem-
ically modified by the heat treatment displayed the cytotoxic
activity. Mass spectrometry analysis of these molecules identified
How do these molecules exert their cytotoxic effects? 4,5-
Dihydroxy-2-cyclopenten-1-one possesses an ␣, ␤ unsaturated
carbonyl group. This kind of chemical bound may form cova-
lent adducts with thiols through a Michael conjugate addition
(Nakayachi et al., 2004), which leads to the denaturation of pro-
teins. This has been described for another molecule possessing
an ␣, ␤ unsaturated carbonyl group, 3-hydroxy-4-[(E)-(2-fyryl)
methylidene]methyl-3-cyclopentene-1,2-dione, which reacts with
cysteines in tubulin, thus impeding microtubule formation and

50
L. Leclere et al. / Carbohydrate Polymers 137 (2016) 39–51
Fig. 10. Schematic representation of HFCP-induced cell death. Heat treatment triggers Maillard reactions that generate 4,5-dihydroxy-2-cyclopenten-1-one. This compound is
able to form covalent adducts on cysteines, which leads to protein denaturation and ubiquitination. Tubulin adducts induce the degradation of this protein, hence inhibiting
mitosis. On the other hand, ubiquitinylated protein aggregates inhibit proteasome, leading to caspase 8 activation and apoptosis.
cell cycle (Marko et al., 2002). It is important to note that
HepG2 cells incubated in the presence of HFCP displayed a
decrease in the abundance of tubulin (data not shown). Another
family of molecules displaying an ␣, ␤ unsaturated carbonyl
group are cyclopentenone-containing prostaglandins. Some of
them have anti-cancer properties (Diez-Dacal & Perez-Sala, 2012;
etoposide, one of the most used anti-cancer agents nowadays. In
conclusion, this work has unraveled a new avenue to develop new
drugs with a new mechanism of action that could overcome resis-
tance of cancer cells to conventional chemotherapies.
Acknowledgements
Sanchez-Gomez, Cernuda-Morollon, Stamatakis,
& Perez-Sala,
2004). Moreover, prostaglandins J inhibit the activity of ubiqui-
tin isopeptidase, leading to polyubiquitinylated misfolded protein
accumulation (Mullally, Moos, Edes, & Fitzpatrick, 2001). Similarly,
cyclopentenone-containing prostaglandins provoke the denatur-
ation and aggregation of the UCH-L1 protein (ubiquitin C-terminal
hydrolase-L1), which is a deubiquitination enzyme, via a reac-
tion with the thiol group of cysteines (Koharudin et al., 2010).
Finally, an increase in cyclopentenone-containing prostaglandin
level was detected in ischemic brain, that led to ubiquitination
and aggregation of proteins in neurons (Liu et al., 2013). We
actually observed an accumulation of polyubiquitinylated pro-
teins in HepG2 cells exposed to HFCP (Leclere et al., 2015) as
well as to (4R,5R)-4,5-dihydroxy-2-cyclopenten-1-one (Fig. 8C).
We further demonstrated that this accumulation of polyubiqui-
tinylated proteins initiated autophagy, as a protective mechanism,
but that this process was overwhelmed, leading to caspase 8 acti-
vation and atypical apoptosis cell death (Leclere et al., 2015).
It must be noted that cells incubated in the presence of 4,5-
dihydroxy-2-cyclopenten-1-one displayed similar apoptotic-like
features, suggesting that it is one of the active molecules in HFCP.
The whole process is summarized in Fig. 10.
To the best of our knowledge, it is the first time that results
regarding the effects of modified pectin on normal cells are
reported. A similar cytotoxic effect was observed in several cancer
cell lines and two normal cell types both for HFCP and 4,5-
dihydroxy-2-cyclopenten-1-one. This observation urges everyone
to look for the possible toxicity of new putative anti-cancer agent
toward normal cells. However, future chemical modification of 4,5-
dihydroxy-2-cyclopenten-1-one may decrease its toxicity toward
normal cells, while keeping it for cancer cells, hence enhancing
its specificity, as was performed for podophyllotoxin that became
L. Leclère was recipient of a FRIA fellowship (Fonds pour la for-
mation à la Recherche dans l’Industrie et dans l’Agriculture). We
are grateful to ARC for a PhD grant to S. El Bkassiny. The authors
would like to thank the “MaSUN” technological platform from the
University of Namur.
References
Bauer-Marinovic, M., Taugner, F., Florian, S., & Glatt, H. (2012). Toxicity studies
with 5-hydroxymethylfurfural and its metabolite 5-sulphooxymethylfurfural
in wild-type mice and transgenic mice expressing human sulphotransferases
1A1 and 1A2. Archives of Toxicology, 86, 701–711. http://dx.doi.org/10.1007/
s00204-012-0807-5
Bharate, S. S., & Bharate, S. B. (2012). Non-enzymatic browning in citrus juice:
Chemical markers, their detection and ways to improve product quality.
Journal of Food Science and Technology, http://dx.doi.org/10.1007/s13197-012-
0718-8. 10.3389/fphar.2013.00128
Chauhan, D., Li, G., Podar, K., Hideshima, T., Neri, P., He, D., et al. (2005). A novel
carbohydrate-based therapeutic GCS-100 overcomes bortezomib resistance
and enhances dexamethasone-induced apoptosis in multiple myeloma cells.
Cancer Research, 65, 8350–8358. http://dx.doi.org/10.1158/0008-5472.CAN-05-
0163
Chen, C., Ye, W., lLiu, C., & Schnelle, S. W. (2012). 2ꢀ-Fluoro-3-deazaaristeromycin.
Tetrahedron, 68, 3908–3914.
Cheng, H., Li, S., Fan, Y., Gao, X., Hao, M., Wang, J., et al. (2011). Comparative studies
of the antiproliferative effects of ginseng polysaccharides on HT-29 human
colon cancer cells. Medical Oncology, 28, 175–181. http://dx.doi.org/10.1007/
s12032-010-9449-8
Cheng, H., Zhang, Z., Leng, J., Liu, D., Hao, M., Gao, X., et al. (2013). The inhibitory
effects and mechanisms of rhamnogalacturonan I pectin from potato on HT-29
colon cancer cell proliferation and cell cycle progression. International Journal
of Food Sciences and Nutrition, 64, 36–43. http://dx.doi.org/10.3109/09637486.
2012.694853
Cragg, G. M., & Newman, D. J. (2005). Plants as a source of anti-cancer agents.
Journal of Ethnopharmacology, 100, 72–79. http://dx.doi.org/10.1016/j.jep.2005.
05.011

L. Leclere et al. / Carbohydrate Polymers 137 (2016) 39–51
51
Cragg, G. M., & Newman, D. J. (2013). Natural products: A continuing source of
Leclere, L., Van Cutsem, P., & Michiels, C. (2013). Anti-cancer activities of pH- or
heat-modified pectin. Frontiers in Pharmacology, 4, 128. http://dx.doi.org/10.
3389/fphar.2013.00128
novel drug leads. Biochimica et Biophysica Acta, 1830, 3670–3695. http://dx.doi.
org/10.1016/j.bbagen.2013.02.008
de Oliveira, J. T., de Matos, A. J., Gomes, J., Vilanova, M., Hespanhol, V., Manninen,
A., et al. (2010). Coordinated expression of galectin-3 and galectin-3-binding
sites in malignant mammary tumors: Implications for tumor metastasis.
Glycobiology, 20, 1341–1352. http://dx.doi.org/10.1093/glycob/cwq103
Diez-Dacal, B., & Perez-Sala, D. (2012). A-class prostaglandins: Early findings and
new perspectives for overcoming tumor chemoresistance. Cancer Letters, 320,
150–157. http://dx.doi.org/10.1016/j.canlet.2012.03.003
Fortuna-Costa, A., Gomes, A. M., Kozlowski, E. O., Stelling, M. P., & Pavao, M. S.
(2014). Extracellular galectin-3 in tumor progression and metastasis. Frontiers
in Oncology, 4, 138. http://dx.doi.org/10.3389/fonc.2014.00138
Gao, X., Zhi, Y., Zhang, T., Xue, H., Wang, X., Foday, A. D., et al. (2012). Analysis of
the neutral polysaccharide fraction of MCP and its inhibitory activity on
galectin-3. Glycoconjugate Journal, 29, 159–165. http://dx.doi.org/10.1007/
s10719-012-9382-5
Harholt, J., Suttangkakul, A., & Vibe Scheller, H. (2010). Biosynthesis of pectin. Plant
Physiology, 153, 384–395. http://dx.doi.org/10.1104/pp.110.156588
Hiramoto, K., Ishihara, A., Sakui, N., Daishima, S., & Kikugawa, K. (1998). Formation
of hydroxyfuranone and hydroxypyranone derivatives with DNA-breaking
activity in the Maillard reaction of glucose and albumin under physiological
conditions. Biological & Pharmaceutical Bulletin, 21, 102–104.
Ibartz, A., Garza, D., & Pagan, J. (2008). Nonenzymatic browning of selected fruit
juices affected by d-galacturonic acid. International Journal of Food Science &
Technology, 43, 908–914. http://dx.doi.org/10.3389/fphar.2013.00128
Inohara, H., & Raz, A. (1994). Effects of natural complex carbohydrate (citrus
pectin) on murine melanoma cell properties related to galectin-3 functions.
Glycoconjugate Journal, 11, 527–532.
Liu, H., Li, W., Ahmad, M., Rose, M. E., Miller, T. M., Yu, M., et al. (2013). Increased
generation of cyclopentenone prostaglandins after brain ischemia and their
role in aggregation of ubiquitinated proteins in neurons. Neurotoxicity
Research, 24, 191–204. http://dx.doi.org/10.1007/s12640-013-9377-4
Marko, D., Habermeyer, M., Kemeny, M., Weyand, U., Niederberger, E., Frank, O.,
et al. (2003). Maillard reaction products modulating the growth of human
tumor cells in vitro. Chemical Research in Toxicology, 16, 48–55. http://dx.doi.
org/10.1021/tx025531a
Marko, D., Kemeny, M., Bernady, E., Habermeyer, M., Weyand, U., Meiers, S., et al.
(2002). Studies on the inhibition of tumor cell growth and microtubule
assembly by
3-hydroxy-4-[(E)-(2-furyl)methylidene]methyl-3-cyclopentene-1,2-dione, an
intensively coloured Maillard reaction product. Food and Chemical Toxicology,
40, 9–18 (an international journal published for the British Industrial Biological
Research Association).
Mohnen, D. (2008). Pectin structure and biosynthesis. Current Opinion in Plant
Biology, 11, 266–277. http://dx.doi.org/10.1016/j.pbi.2008.03.006
Mullally, J. E., Moos, P. J., Edes, K., & Fitzpatrick, F. A. (2001). Cyclopentenone
prostaglandins of the J series inhibit the ubiquitin isopeptidase activity of the
proteasome pathway. The Journal of Biological Chemistry, 276, 30366–30373.
http://dx.doi.org/10.1074/jbc.M102198200
Nakayachi, T., Yasumoto, E., Nakano, K., Morshed, S. R., Hashimoto, K., Kikuchi, H.,
et al. (2004). Structure-activity relationships of alpha, beta-unsaturated
ketones as assessed by their cytotoxicity against oral tumor cells. Anticancer
Research, 24, 737–742.
Nangia-Makker, P., Hogan, V., Honjo, Y., Baccarini, S., Tait, L., Bresalier, R., et al.
(2002). Inhibition of human cancer cell growth and metastasis in nude mice by
oral intake of modified citrus pectin. Journal of the National Cancer Institute, 94,
1854–1862.
Palmer, A. M., & Jager, V. (2001). Pyrrolidine N-oxides by stereoselective addition
of grignard and lithium compounds to
Jackson, C. L., Dreaden, T. M., Theobald, L. K., Tran, N. M., Beal, T. L., Eid, M., et al.
(2007). Pectin induces apoptosis in human prostate cancer cells: Correlation of
apoptotic function with pectin structure. Glycobiology, 17, 805–819. http://dx.
doi.org/10.1093/glycob/cwm054
Kang, H. J., Jo, C., Kwon, J. H., Son, J. H., An, B. J., & Byun, M. W. (2006). Antioxidant
and cancer cell proliferation inhibition effect of citrus pectin-oligosaccharide
prepared by irradiation. Journal of Medicinal Food, 9, 313–320. http://dx.doi.
org/10.1089/jmf.2006.9.313
4,5-dideoxy-2,3-O-isopropylidene-d-erythro-4-pentenose N-benzyl nitrone
and subsequent cope–house cyclization. European Journal of Organic Chemistry,
2001, 1293–1308.
Klepper, F., Jahn, E.-M., Hickmann, V., & Carell, T. (2007). Synthesis of the
transfer-RNA nucleoside queuosine by using a chiral allyl azide intermediate.
Angewandte Chemie International Edition, 46, 23–27.
Koharudin, L. M., Liu, H., Di Maio, R., Kodali, R. B., Graham, S. H., & Gronenborn, A.
M. (2010). Cyclopentenone prostaglandin-induced unfolding and aggregation
of the Parkinson disease-associated UCH-L1. Proceedings of the National
Academy of Sciences of the United States of America, 107, 6835–6840. http://dx.
doi.org/10.1073/pnas.1002295107
Sanchez-Gomez, F. J., Cernuda-Morollon, E., Stamatakis, K., & Perez-Sala, D. (2004).
Protein thiol modification by 15-deoxy-delta12,14-prostaglandin J2 addition
in mesangial cells: Role in the inhibition of pro-inflammatory genes. Molecular
Pharmacology, 66, 1349–1358. http://dx.doi.org/10.1124/mol.104.002824
Smith, A. B., Han, Q., Breslin, P. A. S., & Beauchamp, G. K. (2005). Synthesis and
assignment of absolute conformation of (–)-oleocanthal: A potent naturally
occurring non-steroidal anti-inflammatory and anti-oxidant agent derived
from extra virgin olive oils. Organic Letters, 7, 23–27.
Koyama, N., Sagawa, H., Kobayashi, E., Enoki, T., Wu, H.-K., Nishiyama, E. et al.
(2000). Cyclopentenones, process for preparing the seame, and the use thereof.
US patent, 6,087, 401.
Kral, S. M., & McFeeters, R. F. (1998). Pectin hydrolysis: Effect of temperature,
degree of methylation, pH, and calcium on hydrolysis rates. Journal of
Agricultural and Food Chemistry, 46, 1311–1315. http://dx.doi.org/10.3389/
fphar.2013.00128
Leclere, L., Fransolet, M., Cote, F., Cambier, P., Arnould, T., Van Cutsem, P., et al.
(2015). Heat-modified citrus pectin induces apoptosis-like cell death and
autophagy in HepG2 and A549 cancer cells. PLoS ONE, 10, e0115831. http://dx.
doi.org/10.1371/journal.pone.0115831
Sugahara, T., Fukada, H., & Iwabuchi, Y. (2004). Total synthesis and absolute
stereochemistry of pentenocin B, a novel interleukin-1 beta converting
enzyme inhibitor. Journal of Organic Chemistry, 69, 1744–1747.
Yang, M., Ye, W., & Schneller, S. (2004). Preparation of carbocyclic
S-adenosylazamethionine accompanied by a practical synthesis of
(−)-aristeromycin. Journal of Organic Chemistry, 69, 3993–3996.
Yapo, B. M., Lerouge, P., Thibault, J. F., & Ralet, M. C. (2007). Pectins from citrus peel
cell walls contain homogalacturonans homogenous with respect to molar
mass, rhamnogalacturon I and rhamnogalacturonan II. Carbohydrate Polymers,
69, 426–435. http://dx.doi.org/10.3389/fphar.2013.00128