In vitro glucuronidation of five rhubarb anthraquinones by intestinal and liver microsomes from humans and rats

Article abstract of DOI:10.1016/j.cbi.2014.05.006

Anthraquinones naturally distribute in many plants including rhubarb and have widespread applications throughout industry and medicine. Recent studies provided new insights in potential applications of these traditional laxative constituents. Glucuronidation was the main metabolic pathway of rhubarb anthraquinones in vivo. This study examined the activity and regioselectivity of glucuronidation of rhubarb anthraquinones (aloe-emodin, emodin, chrysophanol, physcion, rhein) in liver and intestinal microsomes from rats and humans, by comparing with the core structure danthron. All anthraquinones formed mono-glucuronides and, except for rhein, the conjugation sites of the main metabolites were unambiguously identified. Two minor glucuronides of emodin were first reported together with the dominant emodin-3-O-β-d-glucuronide. The substitution on the anthraquinone ring was crucial to the activity and regioselectivity of glucuronidation. In general, the activity was decreased greatly with a β-COOH (rhein), while enhanced dramatically with a β-OH (emodin). Glucuronidation showed an absolute preference towards β-OH, followed by α-OH and β-alcoholic OH. The glucuronidation activity and regioselectivity also varied slightly with organs and species. All glucuronides of aloe-emodin, emodin, chrysophanol and physcion were formed by multiple human UGT isoforms with 1A9 being the most prominent in most cases. The UGT2B subfamily (2B7 and 2B15) only showed high activity towards a β-OH. In conclusion, the substitution at the anthraquinone ring was crucial to the rate and preference of glucuronidation. The high glucuronidation activity of UGT1A9 towards anthraquinones highlighted potential drug interactions.

Full text of DOI:10.1016/j.cbi.2014.05.006

Chemico-Biological Interactions 219 (2014) 18–27
Contents lists available at ScienceDirect
Chemico-Biological Interactions
journal homepage: www.elsevier.com/locate/chembioint
In vitro glucuronidation of five rhubarb anthraquinones by intestinal
and liver microsomes from humans and rats
a,b
Wenjin Wu a, Nan Hu a, Qingwen Zhang a, Yaping Li a, Peng Li a, Ru Yan a, , Yitao Wang
⇑
a State Key Laboratory of Quality Research in Chinese Medicine, Institute of Chinese Medical Sciences, University of Macau, Av. Padre Tomás Pereira Taipa, Macao, China
b School of Chinese Meteria Medica, Beijing University of Chinese Medicine, Beijing, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Anthraquinones naturally distribute in many plants including rhubarb and have widespread applications
throughout industry and medicine. Recent studies provided new insights in potential applications of
these traditional laxative constituents. Glucuronidation was the main metabolic pathway of rhubarb
anthraquinones in vivo. This study examined the activity and regioselectivity of glucuronidation of
rhubarb anthraquinones (aloe-emodin, emodin, chrysophanol, physcion, rhein) in liver and intestinal
microsomes from rats and humans, by comparing with the core structure danthron. All anthraquinones
formed mono-glucuronides and, except for rhein, the conjugation sites of the main metabolites were
unambiguously identified. Two minor glucuronides of emodin were first reported together with the
Received 18 February 2014
Received in revised form 5 May 2014
Accepted 11 May 2014
Available online 20 May 2014
Keywords:
Rhubarb
Anthraquinones
Glucuronidation
UGTs
dominant emodin-3-O-b-D-glucuronide. The substitution on the anthraquinone ring was crucial to the
activity and regioselectivity of glucuronidation. In general, the activity was decreased greatly with a
b-COOH (rhein), while enhanced dramatically with a b-OH (emodin). Glucuronidation showed an abso-
Regioselectivity
lute preference towards b-OH, followed by a-OH and b-alcoholic OH. The glucuronidation activity and
regioselectivity also varied slightly with organs and species. All glucuronides of aloe-emodin, emodin,
chrysophanol and physcion were formed by multiple human UGT isoforms with 1A9 being the most
prominent in most cases. The UGT2B subfamily (2B7 and 2B15) only showed high activity towards a
b-OH. In conclusion, the substitution at the anthraquinone ring was crucial to the rate and preference
of glucuronidation. The high glucuronidation activity of UGT1A9 towards anthraquinones highlighted
potential drug interactions.
Ó 2014 Published by Elsevier Ireland Ltd.
1. Introduction
new applications of these components and the herb, and thereby
have attracted increasing research interests.
Anthraquinones are widely distributed in many medicinal and
nutritional plants, including rhubarb. Rhubarb (Rhei Radix et
Rhizoma), an edible medicinal herb noted for its purgative activity,
originated from the dried roots and rhizomes of Rheum palmatum
L., Rheum officinale Baill., and Rheum tanguticum Maxim. ex Balf.
Although potentially toxic, anthraquinones, including the agly-
cones (aloe-emodin, rhein, emodin, chrysophanol, physcion) and
their glycosides, are considered as the main bioactive constituents
for the laxative effect of rhubarb [1]. Although not well studied in
humans, long-term use or overdose of rhubarb anthraquinones
may cause nephrotoxicity. In recent studies, these anthraquinone
constituents also showed antivirus [2], anticancer [3,4], anti-
oxidation [5] and anti-diabetic activities [6], supporting potential
Rhubarb anthraquinones share the core structure 1,8-dihy-
droxyanthraquinone (danthron) with varying substituents at
b-positions (C-3 and/or C-6), which may account for different
pharmacological activities of anthraquinones. For example, the
presence of an ortho-dihydroxy structure greatly enhanced the
radical scavenging activity, while glycosylation of the hydrox-
ylanthraquinones resulted in reduced activity [5]. The presence
of hydroxyl groups at the b- but not the
nones was essential for suppression of 2,3,7,8-tetrachlorodibenzo-
p-dioxin-induced DNA-binding activity of cytosolic aryl
a-positions of anthraqui-
a
hydrocarbon receptor [7]. The antimicrobial activity of anthraqui-
nones was also shown to be associated with substitutions [8].
Rhubarb anthraquinones exhibited low oral bioavailability in
rats and humans [9,10], which was mainly attributed to extensive
phase II conjugation reactions, in particular, glucuronidation
[10–12]. More interestingly, the conjugated metabolites (mainly
glucuronides) of rhubarb anthraquinones exhibited higher
⇑
Corresponding author. Tel.: +86 853 83974876; fax: +86 853 28841358.
E-mail address: ruyan@umac.mo (R. Yan).
http://dx.doi.org/10.1016/j.cbi.2014.05.006
0009-2797/Ó 2014 Published by Elsevier Ireland Ltd.

W. Wu et al. / Chemico-Biological Interactions 219 (2014) 18–27
19
free-radical scavenging activity than the parent compounds
2.2. Glucuronidation of rhubarb anthraquinones in human and rat
microsomes and recombinant human UGTs
[10,13]. These findings highlighted the importance of glucuronida-
tion to both the disposition of anthraquinones and their health-
promoting activities observed in vivo.
The glucuronidation of tested anthraquinones was assessed
under initial rate conditions. Each anthraquinone (danthron,
aloe-emodin, rhein, emodin, chrysophanol and physcion) was
Regioselectivity of glucuronidation of phenolic compounds has
been extensively reported [14,15]. The regioselectivity could vary
with species, organs, age, gender, UGT isoforms and substrate con-
centrations and sometimes, may result in distinct pharmacological
activities [16]. Therefore, it’s of great importance to investigate the
glucuronidation of rhubarb anthraquinones and identify the
regioselectivity, which, to our best knowledge, has not yet been
reported.
dissolved in DMSO and tested at 0.5
levels were partly dictated by the solubility (5–10
sophanol and physcion in 50 mM Tris–HCl (pH = 7.4). The higher
lM and 5
lM. The substrate
lM) of chry-
concentration was also close to the reported Km (ꢀ5
lM) of emodin
in RLMs [17].
Each anthraquinone was incubated individually in a 200-
lL
Glucuronidation is catalyzed by UDP-glucuronosyltransferases
(UGTs) which are generally highly expressed in the liver and intes-
tine. Both liver and intestine have been demonstrated to be
responsible for the first-pass elimination of anthraquinones via
glucuronidation [17,18]. Information on the differences in glucu-
ronidation of anthraquinones between the two organs is scarce.
Moreover, species differences in glucuronidation of emodin and
rhein were also reported [17]. Given that many pharmacology
studies of rhubarb and the anthraquinones are carried out on rats,
any extrapolation of rat data to humans must consider potential
species differences. So far, the species differences in glucuronida-
tion of aloe-emodin, chrysophanol and physcion by human and
rat microsomes have not been reported.
Therefore, in the present study, the glucuronidation of five
rhubarb anthraquinones was simultaneously characterized in
liver and intestinal microsomes from rats and humans. The influ-
ences of the substituents on the activity and regioselectivity of
glucuronidation of anthraquinones were discerned by comparing
with the glucuronidation of the synthetic analog danthron.
Moreover, the activity and regioselectivity were further dis-
cussed with data obtained from 12 recombinant human UGT
isoforms.
incubation system containing 8 mM MgCl2, 25 g/mL alamethicin,
l
0.05 mg protein/mL liver microsomes or 0.1 mg protein/mL intesti-
nal microsomes in 50 mM Tris–HCl buffer (pH = 7.4) at 37 °C. Reac-
tions were initiated by adding 1 mM UDPGA and kept for 10 min.
In the case of rhein, reactions were carried out at 0.5 mg/mL micro-
somal proteins and 10 mM UDPGA, and kept for 30 min due to its
low glucuronidation activity.
The incubations with UGTs were carried out following the
guideline from the manufacturer: each anthraquinone (aloe-
emodin, emodin, chrysophanol and physcion) was incubated with
0.05 mg protein/mL individual UGT isoform in the presence of
25 lg/mL alamethicin for 10 min. All other conditions were the
same as described above for reactions in microsomal proteins.
DMSO in the reaction system was 1%. Zero-minute incubations
and reactions without UDPGA or with denatured HLMs (heated at
100 °C for 10 min) served as controls. Each experiment was
repeated three times.
2.3. Sample preparation
Reactions were terminated by adding 3 volumes of ice-cold
ethyl acetate and 0.5 volume of 0.1 M HCl which contained
0.5 lg/mL daidzein as the internal standard. The resultant mix-
tures were vortexed, and then centrifuged at 15,000g for 5 min.
2. Materials and methods
The organic layer was evaporated to dryness under a stream of
nitrogen and reconstituted with 100
lL of methanol. An aliquot
2.1. Materials
(10 L) was subjected to LC–MS/MS analysis.
l
Aloe-emodin, rhein, emodin, chrysophanol, physcion, daidzein,
danthron (>98%) were purchased from Yuanye Bio-Technology
2.4. Isolation of anthraquinone glucuronides
Co., Ltd. (Shanghai, China). Physcion-8-O-b-
>98%) was isolated from the Rhei Radix et Rhizoma (Rhubarb) in
our laboratory. Chrysophanol-8-O-b- -glucoside (C-8-G, >98%)
D
-glucoside (P-8-G,
The glucuronides of aloe-emodin (AEG1, AEG2), emodin (EG3),
chrysophanol (CG1, CG2) and physcion (PG1, PG2) were prepared
from scale-up reaction. The reaction system consisted of 0.2 mg/
mL aloe-emodin or emodin, or 0.05 mg/mL chrysophanol or phys-
cion, 0.2 mg protein/mL RLMs, 1 mM UDPGA, in a final volume of
100 mL (aloe-emodin or emodin) or 20 mL (chrysophanol or phys-
cion) and was incubated at 37 °C for 4 h. The reactions were
stopped and samples were prepared in the same manner as
described above before being injected to a semi-preparative
HPLC–UV instrument. Samples were eluted using an Agilent Zob-
D
was provided by BioBioPha Co., Ltd. (Kunming, China).
Alamethicin, uridine -50-diphosphoglucuronic acid (UDPGA)
and dimethyl sulfoxide (DMSO) were obtained from Sigma (St.
Louis, MO). Formic acid and acetonitrile of HPLC grade were
purchased from Merck (Darmstadt, Germany). Ultra-water was
purified by a Milli-Q purification system (Millipore, Bedford,
MA).
Pooled human liver microsomes (HLMs) and intestinal micro-
somes (HIMs), and 12 recombinant human UGTs (UGT1A1, 1A3,
1A4, 1A6, 1A7, 1A8, 1A9, 1A10, 2B4, 2B7, 2B15 and 2B17)
expressed in baculovirus-infected insect cells were purchased from
BD Gentest (Woburn, MA). HLMs were prepared from 50 donors
(27 males, 23 females; range 41.2–61.4 y), and HIMs from 6 donors
(3 males, 3 females; range 36.5–70.9 y). Pooled rat liver micro-
somes (RLMs) and intestinal microsomes (RIMs) were prepared
from 6 untreated male Sprague–Dawley (SD) rats (200–250 g)
using a standard procedure [19,20]. Total contents of microsomal
proteins were measured using the Lowry’s method [19]. All
experiments were carried out according to a protocol (No.: ICMS-
AEC-2013-05) approved by the Animal Ethics Committee, Institute
of Chinese Medical Sciences, University of Macau.
rax SB-C18 column (5
l
m, 250 mm ꢁ 9.4 mm). The mobile phases
consisted of 0.1% aqueous formic acid (A) and acetonitrile (B) and
eluted isocratically at 2.5 mL/min as follows: AEG1 and AEG2
45% B, EG3 55% B, CG1 and CG2 62% B, PG1 and PG2 68% B. The elu-
ent containing each metabolite was collected, evaporated under
nitrogen to remove acetonitrile followed by freeze drying.
The purities of all glucuronides prepared were >98% as deter-
mined using HPLC–UV and LC–MS/MS.
2.5. NMR analysis of AEG1, AEG2 and EG3
AEG1, AEG2, and EG3 were dissolved in DMSO-d6 at 10, 6,
10 mg/mL, respectively. Their nuclear magnetic resonance (NMR)
spectra (1H, 13C, heteronuclear single quantum coherence (HSQC),

20
W. Wu et al. / Chemico-Biological Interactions 219 (2014) 18–27
heteronuclear multiple bond correlation (HMBC), rotating-frame
overhauser effect spectroscopy (ROESY) and nuclear overhauser
effect spectroscopy (NOESY)) were recorded on a Bruker AV-500
(Bruker, Newark, Germany) using tetramethylsilane as the internal
standard. Chemical shifts were expressed in d (ppm) and coupling
constants (J) were reported in Hertz (Hz).
Afterwards, the stock solutions of the glucuronides at known
concentrations were diluted with DMSO to obtain working solu-
tions at serial concentrations. Two microliters of each working
solution were mixed with the in vitro reaction system, respectively,
and processed in the same manner as described in Section 2.3
before subjected to LC–MS/MS analysis. The calibration curves
were constructed by plotting the peak area ratios of the analyte
to the internal standard versus nominal concentrations of the
analyte. All glucuronides exhibited good linearity (R2 > 0.99)
within the concentration ranges tested (DG, 1.3–652 nM; AEG1,
1.9–972 nM; AEG2, 1.2–597 nM; EG3, 3.9–4020 nM; CG1, 1.4–
367 nM; CG2, 0.7–371 nM; PG1, 0.6–154 nM; PG2, 1.0–251 nM)
except for EG3, of which the calibration curve fitted a quadratic
equation with a regression coefficient higher than 0.99 within
3.9–4020 nM (data not shown).
2.6. Identification of the conjugation sites of the glucuronides of
physcion and chrysophanol
To determine the conjugation sites of the glucuronides of phys-
cion (PG1 and PG2), physcion-8-glucoside (P-8-G, 50 lM) was incu-
bated with 0.5 mg protein/mL RLMs in the presence of 1 mM
UDPGA for 2 h. The glucuronide of P-8-G (PGG) formed was isolated
using an analytical column (Alltech Alltima™ C18 column, 5 lm,
250 mm ꢁ 4.6 mm), and then hydrolyzed by incubating with HCl
(0.1 M) and ascorbic acid (5 mg/mL) at 60 °C for 15 min. The hydro-
lytic mixture containing P-8-G and physcion-1-O-glucuronide was
analyzed before and after spiking with the separated PG1 or PG2
using LC–MS/MS on the analytical column (Alltech Alltima™ C18
2.8. HPLC–MS/MS analysis
Mass spectrometry analysis was performed on an ABI 4000 Q-
Trap™ hybrid triple quadrupole linear ion trap mass spectrometer
(Applied Biosystems, Foster City, CA) interfaced online with an HP
1200 series binary pump SL system. Data acquisition and processing
were conducted using Analyst (version 1.5.1, Applied Biosystems,
Foster City, CA). Analytes were eluted on an Agilent XDB C18 column
column, 5
l
m, 250 mm ꢁ 4.6 mm). The identities of PG1 and PG2
were assigned by comparing their retention times and mass spectra
with those of physcion-1-O-glucuronide in the mixture.
Chrysophanol-8-glucoside (C-8-G, 50 lM) was processed in the
(3.5
l
m, 100 mm ꢁ 2.1 mm). The mobile phase consisted of 0.1%
same manner as described above. Chrysophanol-1-O-glucuronide
formed in the mixture was used for determination of the conjuga-
tion sites of CG1 and CG2.
aqueous formic acid (A) and acetonitrile (B) and eluted in a gradient
mode: 0–5 min,10–30% B, 0.3 mL/min; 5–10 min, 30–60% B, 0.3 mL/
min; 10–12 min, 60–100% B, 0.4 mL/min, 12–13 min, 100–10%,
0.4 mL/min, re-equilibration 10% B, 0.3 mL/min, 13–18 min. Elec-
trospray ionization mass spectrometry analysis was operated in
negative ion mode under the following conditions: dry gas N2 8 L/
min, dry temperature 350 °C, nebulizer pressure 40 psi, capillary
voltage ꢂ3500 V. Structural identification was performed at a mass
range of m/z 100–700 for the glucuronides of C-8-G (CGG) and P-8-G
(PGG), and m/z 100–600 for all aglycones and their glucuronides.
Quantitative analysis was performed in multiple reaction
monitoring (MRM) mode. The aglycones and the respective
mono-glucuronides were monitored at the selected precursor/frag-
ment ion pairs: m/z 238.9/211.1 (danthron), 414.9/238.9 (DG),
268.9/239.1 (aloe-emodin), 444.9/268.9 (AEG1, AEG2), 283.1/
238.9 (rhein), 459.1/282.9 (RG1, RG2, RG3), 268.9/224.9 (emodin),
444.8/268.9 (EG1, EG2, EG3), 252.9/224.9 (chrysophanol), 428.9/
252.8 (CG1, CG2), 282.9/239.8 (physcion), 458.9/282.8 (PG1,
PG2). For all analytes, enhanced potential was ꢂ10 V with excep-
tion of ꢂ4.5 V for AEG1 and AEG2, collision cell potential was ꢂ5
to ꢂ10 V, declustering potential was ꢂ40 to ꢂ80 V and collision
energy was ꢂ21 to ꢂ45 V.
2.7. Construction of calibration curves of rhubarb anthraquinones and
their glucuronides
Calibration curves of anthraquinones (aglycones) were con-
structed using HPLC–UV analysis. Five microliters of stock solutions
of each anthraquinone at serial concentrations were added to 95
lL
of methanol, respectively, and 10 L of the solution were then sub-
l
jected to HPLC–UV analysis. The calibration curve of each aglycone
was obtained by plotting peak area measured at 268 nm versus
nominal concentrations of the analyte. Good linearity (r2 P 0.999)
was observed for each aglycone over tested concentration range
(emodin 250–1000 lM, other anthraquinones 50–500 lM).
Anthraquinone glucuronides generated in the in vitro incuba-
tion system were quantitatively determined with the metabolites
isolated (DG, AEG1, AEG2, EG3, CG1, CG2, PG1 and PG2), except
for rhein glucuronides which were semi-quantitated with mass
responses (RG1, RG2, RG3), using HPLC–MS/MS analysis. Each
anthraquinone glucuronide was dissolved in DMSO to make a stock
solution, respectively. Since the amounts of anthraquinone glucu-
ronides isolated from the in vitro study were too low to be weighed
accurately, the concentrations of the stock solutions were deter-
mined from the calibration curve of respective aglycone with a
UV conversion factor (K) between each aglycone and the respective
glucuronide at equal molar concentration. The K value was deter-
2.9. Data analysis
The glucuronidation activity of each anthraquinone was
calculated from the formations of all its mono-glucuronides with
exceptions of rhein and emodin. Glucuronidation of emodin was
approximately measured from the formation of the major glucuro-
nide EG3, which accounted for >95% of all three glucuronides
generated. The elimination of rhein was roughly estimated from
substrate loss due to minor formations and the unavailability of
the pure glucuronides. Data were presented as means standard
deviation (S.D) from triplicate reactions.
mined as follows: (1) 5
lL of stock solution of each glucuronide
was added to 95 L of the aqueous ascorbic acid (5 mg/mL) alone
l
or aqueous ascorbic acid (5 mg/mL) containing b-glucuronidase
(100 U/mL), respectively. The mixed solutions were incubated at
37 °C for 1 h before 100 lL of methanol was added to terminate
the reactions. After centrifugation, the supernatants were analyzed
by HPLC–UV at 268 nm. (2) The differences between the peak area
of the glucuronide (
cone ( Peak area A) in samples obtained before and after the
hydrolysis treatment were calculated. K value was the ratio of
Peak area A to Peak area G (K = Peak area A/ Peak area G).
DPeak area G) and that of the respective agly-
D
3. Results
D
D
D
D
3.1. Metabolism of anthraquinones in rat and human microsomes
The conversion factors (K) of DG, AEG1, AEG2, EG3, CG1, CG2,
PG1 and PG2 from their respective parent compounds were within
0.82–0.94.
When incubated with liver or intestinal microsomes from rats
or humans, all tested anthraquinones were metabolized to their

W. Wu et al. / Chemico-Biological Interactions 219 (2014) 18–27
21
glucuronides in the presence of UDPGA. The representative MRM
3.2. HPLC–MS/MS identification of anthraquinone metabolites
chromatograms of the incubates of each anthraquinone in RLMs
are shown in Fig. 1. The metabolites of each anthraquinone were
the same between liver and intestine within the two species at
The mass spectra of each anthraquinone and its most abundant
glucuronide are shown in Supplementary Fig. S1. The metabolites
of danthron and rhubarb anthraquinones showed their pseudo-
molecular ions 176 mass units higher than the respective agly-
cones. Moreover, the MS2 of the characteristic [M-Glu]ꢂ ion of each
anthraquinone glucuronide exhibited similar fragment ions to
those of the corresponding aglycone (data not shown), supporting
DG1, AEG1 and AEG2, RG1–RG3, EG1–EG3, CG1 and CG2, PG1 and
two substrate levels tested (0.5 and 5 lM) as judged from the
retention times and mass spectra data (Supplemental Fig. S1). Dan-
thron produced only one metabolite (DG), aloe-emodin, chrysoph-
anol and physcion formed two metabolites (AEG1 and AEG2; CG1
and CG2; PG1 and PG2), while both rhein and emodin generated
three glucuronides (RG1–RG3; EG1–EG3).
Fig. 1. Chemical structures and MRM chromatograms of incubated samples of danthron (A), aloe-emodin (B), rhein (C), emodin (D), chrysophanol (E) and physcion (F) in rat
liver microsomes. Each anthraquinone (5 M) was incubated with 0.05 mg protein/mL of rat liver microsomes in the presence of 1 mM UDPGA in 50 mM Tris buffer (pH 7.4)
l
at 37 °C for 10 min, except for rhein which was incubated with 0.5 mg protein/mL of rat liver microsomes in the presence of 10 mM UDPGA for 30 min. Inserts in (C and D)
were the enlarged MRM chromatograms of the incubates of rhein and emodin, respectively.

22
W. Wu et al. / Chemico-Biological Interactions 219 (2014) 18–27
Table 1
NMR data of anthraquinone metabolites AEG1, AEG2 and EG3 in DMSO-d6.
No.
AEG1
AEG2
EG3
dH
dC
dH
dH
dC
1
2
3
4
5
6
7
8
–
161.6
120.7
152.2
116.0
120.7
135.9
122.2
157.8
187.5
182.0
132.2
116.0
115.4
134.9
62.0
–
–
164.5
108.0
165.6
108.8
120.6
148.4
124.2
161.5
189.8
181.5
135.2
113.5
109.1
132.9
–
21.5
99.4
72.8
75.4
71.2
75.4
169.8
7.28 (1H, s)
–
7.66 (1H, s)
7.88 (1H, m)
7.87 (1H, m)
7.70 (1H, d, J = 7.2 Hz)
–
–
–
–
–
–
7.62 (1H, s)
–
7.88 (1H, s)
7.68 (1H, d, J = 7.1 Hz)
7.77 (1H, m)
7.36 (1H, d, J = 7.0 Hz)
–
–
–
–
–
–
6.98 (1H, d, J = 2.5 Hz)
–
7.25 (1H, d, J = 2.5 Hz)
7.50 (1H, s)
–
7.16 (1H, s)
–
–
–
–
–
–
–
9
10
4a
8a
9a
10a
CH2OH
CH3
–
–
4.62 (2H, s)
–
4.64 (2H, s)
–
–
–
2.41 (3H, s)
5.32 (1H, d, J = 7.3 Hz)
3.33 (1H, m)
3.35 (1H, m)
3.34 (1H, m)
4.07 (1H, d, J = 9.5 Hz)
–
GluA-1
5.32 (1H, d, J = 7.5 Hz)
3.49 (1H, m)
3.17 (1H, m)
3.42 (1H, m)
3.93 (1H, d, J = 9.6 Hz)
–
100.0
73.0
75.9
71.3
75.2
5.27 (1H, d, J = 7.2 Hz)
3.49 (1H, m)
3.36 (1H, m)
3.38 (1H, m)
3.82 (1H, d like)
2
3
4
5
6
170.1
PG2 as mono-glucuronidated metabolites of danthron, aloe-
emodin, rhein, emodin, chrysophanol and physcion, respectively.
was mixed with PG1, the MRM chromatogram of selected transi-
tions for physcion mono-glucuronides only showed one peak,
while there were two peaks when PG2 was spiked into the
hydrolytic mixture (Fig. 3C). Thus, PG1 and PG2 were assigned as
physcion-1-O-glucuronide and physcion-8-O-glucuronide, respec-
tively. Using the same strategy, CG1 and CG2 were identified as
chrysophanol-1-O-glucuronide and chrysophanol-8-O-glucuro-
nide, respectively.
3.3. NMR analysis of AEG1, AEG2 and EG3
The 1H NMR and 13C NMR spectra of AEG1 showed the signals of
an aglycone (aloe-emodin) and a sugar moiety (glucuronic acid,
GluA) (Table 1). The 1H and 13C NMR signals assigned to the
aglycone were very similar to those of aloe-emodin [21] except
for the chemical shift of C-8 (ꢂ3.6 ppm) and C-7 (ꢂ2.3 ppm) in
3.5. Activity and regioselectivity of glucuronidation of anthraquinones
in microsomes from rats and humans
AEG1, suggesting that the D-glucuronyl moiety is attached to C-8.
The linkage site was further confirmed by the HMBC correlation
(Fig. 2) between the anomeric proton (H-GluA-1) and C-8, as well
as the ROESY correlation between H-7 and the anomeric proton
(H-GluA-1). Thus, AEG1 was identified as aloe-emodin-8-O-b-D-
glucuronide.
The overall glucuronidation activities of each anthraquinone in
human and rat microsomal proteins are shown in Fig. 4 and Sup-
plementary Fig. S2. Compared with danthron, rhubarb anthraqui-
nones, except for emodin, showed lower glucuronidation activity
in RIMs and comparable activities in RLMs, and lower activities
in HIMs and HLMs. The glucuronidation activity of emodin was
higher than danthron and was the highest in livers of both species
and HIMs. In RIMs, the glucuronidation activity of emodin was
lower than danthron and comparable with other rhubarb anthra-
quinones. The glucuronidation activity of rhein over the incubation
period was the lowest in both human and rat microsomes. The
average rate of loss of rhein (0.033 0.014 nmol/min/mg protein)
determined at a much higher protein level, a longer reaction time
and a higher UDPGA level were still much lower than those of
other anthraquinones (Fig. 4). When comparing between the two
species and the two organs tested, RLMs always showed higher
activities than HLMs, and liver exhibited higher activities than
intestine with exception that the activities of HIMs towards chry-
sophanol and physcion were comparable to those of HLMs.
The 1H NMR data, including the signals of an aglycone (aloe-
emodin) and a sugar moiety (glucuronic acid), of AEG2 were very
similar to those of AEG1 (Table 1). In the ROESY spectrum
(Fig. 2), the anomeric proton (H-GluA-1) showed the correlation
to H-2 and no correlation between the methylene protons at C-3
and the anomeric proton (H-GluA-1) was found, indicating a link-
age of the
D-glucuronyl moiety to C1. Thus, AEG2 was determined
as aloe-emodin-1-O-b-
D
-glucuronide.
The 1H NMR and 13C NMR spectra of EG3 exhibited the signals
of emodin and a sugar moiety (glucuronic acid) (Table 1). The
HMBC correlation between the anomeric proton (H-GluA-1) and
C-3 suggested the D-glucuronyl moiety is attached to C-3, corre-
sponding well with the ROESY correlations between H-2 and the
anomeric proton (H-GluA-1), and between H-4 and the anomeric
proton. Thus, EG3 was identified as emodin-3-O-b-D-glucuronide.
The amount of each glucuronide and the percent of each to the
total glucuronides formed from respective anthraquinone in liver
and intestinal microsomes are shown in Table 2, Supplementary
Tables S1 and S2. In RLMs, the major glucuronides of aloe-emodin,
rhein, emodin, chrysophanol and physcion were 8-O-glucuronide
(AEG1, 76.2%), RG1 (73.2%), 3-O-glucuronide (EG3, 96.5%), 8-O-glu-
curonide (CG2, 64.3%) and 8-O-glucuronide (PG2, 68.9%), respec-
tively. The 3-O-glucuronide (EG3, 98.4%), 8-O-glucuronide (CG2,
59.0%) and 8-O-glucuronide (PG2, 64.8%) were also the prominent
metabolites in HLMs, while the glucuronidation of aloe-emodin
3.4. Elucidation of the conjugation sites of the glucuronides of physcion
and chrysophanol
When incubated with RLMs, physcion-8-glucoside (P-8-G) pro-
duced one metabolite (PGG) (Fig. 3A and 3B), which exhibited the
pseudo-molecular ion at m/z 621 and characteristic ions at m/z 459
and 283 (Fig. 3B), corresponding to a mono-glucuronide of P-8-G.
After treatment with HCl, PGG was hydrolyzed into physcion-1-
O-glucuronide and P-8-G. When the resultant hydrolytic mixture

W. Wu et al. / Chemico-Biological Interactions 219 (2014) 18–27
23
Fig. 2. Key HMBC and ROESY correlations of AEG1 and EG3, and key ROESY correlation of AEG2. AEG1, AEG2 and EG3 were dissolved in DMSO-d6 at a concentration of 10, 6
and 10 mg/mL, respectively, and TMS as the internal standard. Chemical shifts were expressed in d (ppm) and coupling constants (J) were reported in Hertz (Hz).
and rhein showed different positional preference in HLMs. The
ratio of AEG1 (8-glucuronide) to AEG2 (1-glucuronide) was about
3:1 in RLMs, but it was about 2:3 in HLMs. As for rhein, RG1 was
decreased to about 10% and RG3 appeared to be the main metabo-
lite in HLMs (55.5% versus 8.0% in RLMs) (Table 2). The regioselec-
tivity in intestinal microsomes was similar to liver microsomes
from the same species with the exception of rhein, of which RG2
appeared to be the main metabolite in HIMs compared to RG3 in
human liver (Table 2). The regioselectivity in glucuronidation of
anthraquinones was similar in liver and intestinal microsomes
from humans or rats (Table 2, Supplementary Tables S1 and S2).
All tested anthraquinones underwent glucuronidation and
formed mono-glucuronides in microsomes from rats and humans.
In general, the number of the glucuronides generated matched the
number of the free hydroxyl and/or carboxylic groups of the
substrate with exception of danthron and aloe-emodin, which
produced one due to the symmetry of the structure and two
metabolites in the presence of a b-CH2OH, respectively. Two
aloe-emodin glucuronides and the major metabolite of emodin
(emodin-3-O-glucuronide, EG3) were unambiguously identified
using NMR analysis. The exact conjugation sites of chrysophanol
and physcion were firstly identified adopting a new strategy com-
bining biosynthesis with acidic hydrolysis. In addition to EG3
which has been reported previously [17], two minor metabolites
of emodin (EG1 and EG2) were first reported. Three rhein glucu-
ronides (RG1, RG2, RG3) were found in our study in comparison
to two previously reported [23]. The conjugation sites of EG1,
EG2 and the three glucuronides of rhein could not be definitely
assigned due to minor formations in the present study.
3.6. Activity and regioselectivity of glucuronidation of anthraquinones
in recombinant human UGTs
Formation rates of glucuronides of aloe-emodin, emodin, chry-
sophanol and physcion by 12 recombinant human UGTs at two
substrate levels are shown in Fig. 5. Most isoforms from UGT1A
subfamily showed glucuronidation activity towards rhubarb
anthraquinones. UGT1A1, 1A3, 1A7, 1A8 (except for aloe-emodin),
1A9 and 1A10 catalyzed glucuronidation of all four anthraquinones
at varied rates and positional preferences. The top UGT isoforms
conjugating each anthraquinone were summarized in site-specific
manner (Supplementary Table S3). UGT1A9 exhibited the highest
activities in most cases with the exception that UGT1A3 showed
higher/comparable activities towards formations of 8-glucuronides
The findings obtained from the present study revealed that the
activity and positional preference of glucuronidation of anthraqui-
nones varied with organs, species, substrate concentrations, UGT
isoforms, and the substitution at b-positions was the dominant
factor. Many carboxylic acid-containing substances were actively
glucuronidated as acylglucuronides [24]. However, the presence
of a b-COOH in rhein significantly decreased its glucuronidation
activity even at much higher protein (Fig. 4, Supplementary
Fig. 2) and UDPGA levels. This low in vitro glucuronidation activity
of rhein agreed well with our findings that compared to other
anthraquinones, rhein mainly exists as the free form in rhubarb
and in plasma samples collected from rats receiving one oral dose
of the rhubarb extract [12]. The glucuronidation activity of rhein,
although low, was evidenced by the observation of three mono-
glucuronides formed in microsomes. In contrast, with a b-OH, emo-
din acquired much higher glucuronidation activities in RLMs, HLMs
and HIMs. With two b-OH groups at the anthraquinone ring,
anthraflavic acid showed an even higher glucuronidation activity
than emodin [25,26]. These findings indicated that the presence
of b-OH markedly enhanced the glucuronidation activity of anthra-
quinones in most cases. When substituted with b-alcoholic OH,
b-CH3 or b-OCH3, the glucuronidation activity was affected to a less
extent (Fig. 4, Supplementary Fig. 2). Taken together, these findings
support the impact of substituent at b-positions on the overall
glucuronidation activity of rhubarb anthraquinones: b-OH >
b-CH3 ꢃ b-OCH3 ꢃ b-alcoholic OH > b-carboxyl group.
of chrysophanol and physcion at the higher substrate level (5 lM).
In most cases, UGT1A9 and UGT1A7 favored the formations of aloe-
emodin 1-O-glucuronide (AEG1), chrysophanol 1-O-glucuronide
(CG1) and physcion 8-O-glucuronide (PG2), while the opposite
was true with UGT1A1 and UGT1A3. All positive UGT1A isoforms
showed glucuronidation preference on b-OH.
Most UGT2B isoforms showed low to no activities towards
anthraquinone glucuronidation, except that UGT2B7 and UGT2B15
catalyzed the b-OH glucuronidation of emodin at activities compa-
rable to UGT1A3 (Fig. 5, Supplementary Table S3).
4. Discussion
UGTs-mediated glucuronidation is the most important conjuga-
tion reaction in eliminating both exogenous and endogenous com-
pounds for humans and many other species [22]. Previous studies
revealed glucuronidation as the most efficient metabolic pathway
of rhubarb anthraquinones [9,10]. In the present study, the glucu-
ronidation of rhubarb anthraquinones in microsomes from humans
and rats were characterized in vitro.
In most cases, emodin showed the highest glucuronidation activ-
ity with an absolute positional preference at the b-OH, leading to the
b-glucuronide (EG3) accounting for >95% of three glucuronides

24
W. Wu et al. / Chemico-Biological Interactions 219 (2014) 18–27
Fig. 3. Structural identification of PG1 and PG2. (A) Biosynthesis and acidic hydrolysis of PGG; (B) MRM chromatograms and MS1 spectra of PGG and P-8-G. (C) MRM
chromatograms of physcion-1-O-glucuronide in the hydrolytic mixture of PGG (C-1), PG1 spiked with the hydrolytic mixture (C-2), PG2 spiked with the hydrolytic mixture
(C-3). Physcion-8-O-glucoside (P-8-G) (50
lM) was incubated in the presence of 1 mM UDPGA for 2 h to form the glucuronide (PGG). PGG was isolated, and then hydrolyzed
by incubation with HCl (0.1 M) in the presence of ascorbic acid (5 mg/mL) at 60 °C for 15 min to give physcion-1-O-glucuronide.
formed in microsomes from both species and human UGTs. In pres-
glucuronidation of aloe-emodin and rhein were observed and
ence of a b-alcoholic OH (aloe-emodin), glucuronidation only
should be attributed to varied hepatic and intestinal UGTs contents
between rats and humans [27,28]. The preferential formations of
emodin 3-O-glucuronide and the 8-O-glucuronides of chrysopha-
nol and physcion were the same in liver and intestinal microsomes
from humans and rats. It is worthy of note that, in addition to the
microsomal enzyme activity, other factors, such as organ blood
flow, total organ mass, also differ between species and contribute
to species difference in vivo. Thus, an in vivo study is warranted
to evaluate whether the species difference in glucuronidation of
anthraquinones observed from the microsomes-based assay
resembles the in vivo situation.
occurred at the
a-OH. Slight regioselectivity in glucuronidation at
a-OH, in particular, the C-8 OH, of anthraquinones was observed
in presence of b-CH3 (chrysophanol), b-OCH3 and b-CH3 (physcion)
substituents. It’s difficult to address the influence of the b-COOH
(rhein) on the regioselectivity of glucuronidation of anthraquinones
without knowing the structures of the glucuronides formed. Hence,
the positional preference of glucuronidation of anthraquinones fol-
lows the descending order: b-OH > a-OH > b-alcoholic OH.
This study revealed the differences in overall glucuronidation
activity of rhubarb anthraquinones between rat and human
microsomes, which are firstly reported for aloe-emodin, chrysoph-
anol and physcion. Species differences in regioselectivity of
Many isoforms (UGT1A1, 1A3, 1A7, 1A9 and 1A10) from UGT1A
family catalyzed glucuronidation of all tested anthraquinones with

W. Wu et al. / Chemico-Biological Interactions 219 (2014) 18–27
25
Fig. 4. Overall glucuronidation activities of danthron, aloe-emodin, rhein, emodin, chrysophanol and physcion at 5
lM by RLMs and RIMs (A), HLMs and HIMs (B). Rhein was
incubated with 0.5 mg/mL of microsomal proteins from rat or human in the presence of 10 mM UDPGA for 30 min at 37 °C. Other anthraquinones were incubated individually
with 0.05 mg/mL (liver) or 0.1 mg/mL (intestine) microsomal proteins in the presence of 1 mM UDPGA for 10 min at 37 °C. *: substrate loss < 0.05 nmol/min/mg protein. Data
represented mean S.D from triplicate reactions.
Table 2
The amounts of anthraquinone glucuronides formed (nmol/min/mg protein) and the proportion of each to the total glucuronides in the incubates of liver microsomes with the
substrate at 5
lM.
Substrate
Species
Metabolite
G1
G2
G3
Aloe emodin
Rheina
Rat
Human
1.06 0.05 (76.2%)
0.10 0.02 (38.4%)
0.33 0.02 (23.8%)
0.17 0.06 (61.6%)
–
–
Rat
–
–
–
(73.2%)
–
(18.8%)
–
(8.0%)
–
Human
(10.9%)
(33.6%)
(55.5%)
Emodina
Rat
–
–
9.03 0.18
(96.5%)
6.99 0.21
(98.4%)
(1.5%)
–
(0.8%)
(1.9%)
–
(0.9%)
Human
Chrysophanol
Physcion
Rat
Human
0.63 0.23 (35.7%)
0.15 0.01 (41.0%)
1.14 0.32 (64.3%)
0.21 0.02 (59.0%)
–
–
Rat
Human
0.35 0.18 (31.1%)
0.08 0.01 (35.2%)
0.76 0.26 (68.9%)
0.14 0.01 (64.8%)
–
–
–: Not applicable.
a
Proportion calculated based on peak area ratio data.
varied activities and preferences (Supplemental Table S3). The
most abundant isoforms from UGT1A family in human liver were
1A1, 1A4, 1A6 and 1A9 [29], while 1A1, 1A6, 1A8 and 1A10 showed
high relative mRNA levels in human small intestine [29]. In the
present study, 1A1 showed considerable activities towards glucu-
ronidation of anthraquinones, thus, it likely plays an important role
in both intestinal and hepatic first-pass elimination of rhubarb
anthraquinones in human body. UGT1A4 showed no glucuronida-
tion activity on tested anthraquinones. This result agreed with
the previous report that UGT1A4 was active in the N-glucuronida-
tion of tertiary amine, yet showed no activity towards carboxylic
acid or phenolic-hydroxy [24]. Both UGT 1A6 and 1A9 were classi-
fied as ‘‘phenol UGT isoforms’’ due to their high turnover rates
towards phenols [30]. In this study, 1A9 showed the highest glucu-
ronidation activities towards rhubarb anthraquinones in most
cases. Yet UGT1A6 showed very low or no activity to anthraqui-
nones tested in contrast to a high activity towards the probe sub-
strate 1-naphthol (data not shown). This finding agreed well with
Ebner’s report that UGT1A6 preferentially catalyzed the glucuron-
idation of small planar phenols, while UGT1A9 exhibited high
activity towards bulky, larger nonplanar phenols [30]. A much
higher activity of UGT1A8 than UGT1A10 on emodin glucuronida-
tion was reported [31], which was in contrast with our observa-
tion. The reason for this discrepancy warrants further study. The
activity and preference of each UGT isoform towards glucuronida-
tion of anthraquinones at two substrate levels were almost the
same, except for UGT1A3 which exhibited a higher activity than
1A9 in the formations of 8-O-glucuronides of chrysophanol and
physcion at 5 lM of the substrate but vice versa at 0.5 lM. This
might be due to the different affinities to UGTs at different sub-
strate levels [14,32,33]. Although 1A3 and 1A7 also showed high
glucuronidation activity, they are not likely to play important roles
in clearance of rhubarb anthraquinones in human body due to their
low expression in human liver and small intestine. Taken together,
it is reasonable to speculate a main contribution of UGT1A9 and
UGT1A10 in hepatic and intestinal first-pass elimination of these
anthraquinones, respectively. UGT1A1 might be another important
isoform which is responsible for the glucuronidation of anthraqui-
nones in both organs.
It was found that the top isoforms that catalyzed a-OH glucuron-
idation (aloe-emodin, chrysophanol, physcion) were highly similar,
but very much different from those involved in b-OH glucuronida-
tion (emodin) (Fig. 5, Supplemental Tables S1 and S2). The present
study revealed the involvement of UGT2B7, 2B15 and 2B4 in glucu-
ronidation of rhubarb anthraquinones. Considering that UGT2B7
and UGT2B15 are abundantly expressed in human small intestine
and liver [29], the data imply important contributions of these
two isoforms in glucuronidation of b-OH of anthraquinones in both

26
W. Wu et al. / Chemico-Biological Interactions 219 (2014) 18–27
Fig. 5. Formations of glucuronides of aloe-emodin, emodin, chrysophanol and physcion by 12 recombinant human UGT isoforms. Each anthraquinone at low (0.5
(5 M) concentrations were incubated with 0.05 mg protein/mL of each UGT isoform in the presence of 1 mM UDPGA for 10 min at 37 °C. Data represented mean S.D. from
triplicate reactions.
lM) or high
l

W. Wu et al. / Chemico-Biological Interactions 219 (2014) 18–27
27
[9] J.H. Lee, J.M. Kim, C. Kim, Pharmacokinetic analysis of rhein in Rheum
undulatum L, J. Ethnopharmacol. 84 (2003) 5–9.
[10] C.S. Shia, S.H. Juang, S.Y. Tsai, P.H. Chang, S.C. Kuo, Y.C. Hou, et al., Metabolism
and pharmacokinetics of anthraquinones in Rheum palmatum in rats and
ex vivo antioxidant activity, Planta Med. 75 (2009) 1386–1392.
organs in human body. In addition to alizarin, emodin and anthra-
flavic acid [34], UGT2B15 also showed glucuronidation activity
towards chrysophanol in this study (Fig. 5).
[11] R. Song, L. Xu, F. Xu, Z. Li, H. Dong, Y. Tian, et al., In vivo metabolism study of
rhubarb decoction in rat using high-performance liquid chromatography with
5. Conclusion
UV photodiode-array and mass-spectrometric detection:
a strategy for
systematic analysis of metabolites from traditional Chinese medicines in
biological samples, J. Chromatogr. A 1217 (2010) 7144–7152.
[12] W. Wu, R. Yan, M. Yao, Y. Zhan, Y. Wang, Pharmacokinetics of anthraquinones
This study reported the glucuronidation of rhubarb anthraqui-
nones in intestinal and liver microsomal proteins from humans
and rats. In general, the activity was decreased greatly in the
presence of b-COOH, and enhanced dramatically with b-OH. Glucu-
ronidation of anthraquinones showed absolute preference towards
in rat plasma after oral administration of
Chromatogr. 28 (2014) 564–572.
a rhubarb extract, Biomed.
[13] C.S. Shia, Y.C. Hou, S.Y. Tsai, P.H. Huieh, Y.L. Leu, P.D. Chao, Differences in
pharmacokinetics and ex vivo antioxidant activity following intravenous and
oral administrations of emodin to rats, J. Pharm. Sci. 99 (2010) 2185–2195.
[14] L. Tang, L. Ye, R. Singh, B. Wu, C. Lv, J. Zhao, et al., Use of glucuronidation
fingerprinting to describe and predict mono- and dihydroxyflavone
metabolism by recombinant UGT isoforms and human intestinal and liver
microsomes, Mol. Pharm. 7 (2010) 664–679.
[15] B. Wu, S. Basu, S. Meng, X. Wang, M. Hu, Regioselective sulfation and
glucuronidation of phenolics: insights into the structural basis, Curr. Drug
Metab. 12 (2011) 900–916.
[16] K. Shimomura, O. Kamata, S. Ueki, S. Ida, K. Oguri, H. Yoshimura, et al.,
Analgesic effect of morphine glucuronides, Tohoku J. Exp. Med. 105 (1971)
45–52.
[17] W. Liu, L. Tang, L. Ye, Z. Cai, B. Xia, J. Zhang, et al., Species and gender
differences affect the metabolism of emodin via glucuronidation, AAPS J. 12
(2010) 424–436.
[18] Z.H. Teng, S.Y. Zhou, R.T. Yang, X.Y. Liu, R.W. Liu, X. Yang, et al., Quantitation
assay for absorption and first-pass metabolism of emodin in isolated rat small
intestine using liquid chromatography–tandem mass spectrometry, Biol.
Pharm. Bull. 30 (2007) 1628–1633.
[19] O.H. Lowry, N.J. Rosebrough, A.L. Farr, R.J. Randall, Protein measurement with
the Folin phenol reagent, J. Biol. Chem. 193 (1951) 265–275.
[20] G. Lin, Y.Y. Cui, E.M. Hawes, Characterization of rat liver microsomal
metabolites of clivorine, an hepatotoxic otonecine-type pyrrolizidine
alkaloid, Drug Metab. Dispos. 28 (2000) 1475–1483.
b-OH followed by a-OH and b-alcoholic OH. The overall glucuron-
idation activity and regioselectivity of anthraquinones tested
subjected to slight changes with organs and species. Human
recombinant UGT1A9 showed the highest activity and similar reg-
ioselectivity to that of HLMs in most cases, while UGT2B isoforms
(UGT2B7 and 2B15) only showed high activity towards b-OH.
These findings lay a foundation for further pharmacological evalu-
ation of the main glucuronides of these anthraquinones as well as
prediction of drug interactions.
Transparency Document
The Transparency document associated with this article can be
found in the online version.
Acknowledgements
This work is supported by the Science and Technology
Development Fund of Macao SAR (Ref. No. 043/2011/A2), the
National Basic Research Program of China (973 program, Grant
No. 2009CB522707), and the Research Committee of University
of Macau (Ref. No. RG086/10-11S/YR/ICMS). We also thank Bio
Bio Pha Co., Ltd. (Yunnan, China) for offering C-8-G.
[21] K. Danielsen, D.W. Aksnes, G.W. Francis, NMR study of some anthraquinones
from rhubarb, Magn. Reson. Chem. 30 (1992) 359–360.
[22] J.O. Miners, P.I. Mackenzie, Drug glucuronidation in humans, Pharmacol. Ther.
51 (1991) 347–369.
[23] M. Dahms, R. Lotz, W. Lang, U. Renner, E. Bayer, H. Spahn-Langguth,
Elucidation of phase I and phase II metabolic pathways of rhein: species
differences and their potential relevance, Drug Metab. Dispos. 25 (1997)
442–452.
[24] R.H. Tukey, C.P. Strassburg, Human UDP-glucuronosyltransferases:
metabolism, expression, and disease, Annu. Rev. Pharmacol. Toxicol. 40
(2000) 581–616.
Appendix A. Supplementary data
[25] J. Terao, K. Murota, Y. Kawai, Conjugated quercetin glucuronides as bioactive
metabolites and precursors of aglycone in vivo, Food Funct. 2 (2011) 11–17.
[26] S.B. Senafi, D.J. Clarke, B. Burchell, Investigation of the substrate specificity of a
cloned expressed human bilirubin UDP-glucuronosyltransferase: UDP-sugar
specificity and involvement in steroid and xenobiotic glucuronidation,
Biochem. J. 303 (Pt 1) (1994) 233–240.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.cbi.2014.05.006.
References
[27] S. Ohno, S. Nakajin, Determination of mRNA expression of human UDP-
glucuronosyltransferases and application for localization in various human
tissues by real-time reverse transcriptase-polymerase chain reaction, Drug
Metab. Dispos. 37 (2009) 32–40.
[28] M.K. Shelby, N.J. Cherrington, N.R. Vansell, C.D. Klaassen, Tissue mRNA
expression of the rat UDP-glucuronosyltransferase gene family, Drug Metab.
Dispos. 31 (2003) 326–333.
[29] B. Wu, K. Kulkarni, S. Basu, S. Zhang, M. Hu, First-pass metabolism via UDP-
glucuronosyltransferase: a barrier to oral bioavailability of phenolics, J. Pharm.
Sci. 100 (2011) 3655–3681.
[30] T. Ebner, B. Burchell, Substrate specificities of two stably expressed human
liver UDP-glucuronosyltransferases of the UGT1 gene family, Drug Metab.
Dispos. 21 (1993) 50–55.
[31] Z. Cheng, A. Radominska-Pandya, T.R. Tephly, Studies on the substrate
specificity of human intestinal UDP-lucuronosyltransferases 1A8 and 1A10,
Drug Metab. Dispos. 27 (1999) 1165–1170.
[32] Y. Fan, L. Tang, J. Zhou, Q. Feng, B. Xia, Z. Liu, Simultaneous determination of
sulfation and glucuronidation of flavones in FVB mouse intestine in vitro and
in vivo, J. Appl. Toxicol. 33 (2013) 273–280.
[33] L. Tang, J. Zhou, C.H. Yang, B.J. Xia, M. Hu, Z.Q. Liu, Systematic studies of
sulfation and glucuronidation of 12 flavonoids in the mouse liver S9 fraction
reveal both unique and shared positional preferences, J. Agric. Food Chem. 60
(2012) 3223–3233.
[34] D. Turgeon, J.S. Carrier, S. Chouinard, A. Belanger, Glucuronidation activity
of the UGT2B17 enzyme toward xenobiotics, Drug Metab. Dispos. 31
(2003) 670–676.
[1] S.Y. Wei, W.X. Yao, W.Y. Ji, J.Q. Wei, S.Q. Peng, Qualitative and quantitative
analysis of anthraquinones in rhubarbs by high performance liquid
chromatography with diode array detector and mass spectrometry, Food
Chem. 141 (2013) 1710–1715.
[2] Z. Li, L.J. Li, Y. Sun, J. Li, Identification of natural compounds with anti-hepatitis
B virus activity from Rheum palmatum L. ethanol extract, Chemotherapy 53
(2007) 320–326.
[3] J.C. Ko, M.S. Tsai, Y.H. Kuo, Y.F. Chiu, S.H. Weng, Y.C. Su, et al., Modulation of
Rad51, ERCC1, and thymidine phosphorylase by emodin result in synergistic
cytotoxic effect in combination with capecitabine, Biochem. Pharmacol. 81
(2011) 680–690.
[4] Q. Huang, G. Lu, H.M. Shen, M.C. Chung, C.N. Ong, Anti-cancer properties of
anthraquinones from rhubarb, Med. Res. Rev. 27 (2007) 609–630.
[5] Y. Cai, M. Sun, J. Xing, H. Corke, Antioxidant phenolic constituents in roots of
Rheum officinale and Rubia cordifolia: structure-radical scavenging activity
relationships, J. Agric. Food Chem. 52 (2004) 7884–7890.
[6] J.M. Zheng, J.M. Zhu, L.S. Li, Z.H. Liu, Rhein reverses the diabetic phenotype of
mesangial cells over-expressing the glucose transporter (GLUT1) by inhibiting
the hexosamine pathway, Br. J. Pharmacol. 153 (2008) 1456–1464.
[7] I. Fukuda, A. Kaneko, S. Nishiumi, M. Kawase, R. Nishikiori, N. Fujitake, et al.,
Structure–activity relationships of anthraquinones on the suppression of
DNA-binding activity of the aryl hydrocarbon receptor induced by 2,3,7,8-
tetrachlorodibenzo-p-dioxin, J. Biosci. Bioeng. 107 (2009) 296–300.
[8] J. Wang, H. Zhao, W. Kong, C. Jin, Y. Zhao, Y. Qu, et al., Microcalorimetric assay
on the antimicrobial property of five hydroxyanthraquinone derivatives in
rhubarb (Rheum palmatum L.) to Bifidobacterium adolescentis, Phytomedicine
17 (2010) 684–689.