Journal of Pharmaceutical and Biomedical Analysis 83 (2013) 108–121

Contents lists available at SciVerse ScienceDirect

Journal of Pharmaceutical and Biomedical Analysis

jou rnal homepage: www.elsevier.com/locate/jpba

The proﬁling and identiﬁcation of the absorbed constituents and metabolites of

Paeoniae Radix Rubra decoction in rat plasma and urine by the

HPLC–DAD–ESI-IT-TOF-MS technique: A novel strategy for the systematic

screening and identiﬁcation of absorbed constituents and metabolites from traditional Chinese medicines

1 1

Jing Liang , Feng Xu , Ya-Zhou Zhang, Shuai Huang, Xin-Yu Zang, Xin Zhao, Lei Zhang,

∗

Ming-Ying Shang, Dong-Hui Yang, Xuan Wang, Shao-Qing Cai

State Key Laboratory of Natural and Biomimetic Drugs, School of Pharmaceutical Sciences, Peking University, No. 38 Xueyuan Road, Beijing 100191, China

a r t i c l e i n f o a b s t r a c t

Article history: Paeoniae Radix Rubra (PRR, the dried roots of Paeonia lactiﬂora) is a commonly used traditional Chinese

Received 23 January 2013

medicine (TCM). A clear understanding of the absorption and metabolism of TCMs is very important

Received in revised form 13 April 2013

in their rational clinical use and pharmacological research. To ﬁnd more of the absorbed constituents

Accepted 16 April 2013

and metabolites of TCMs, a novel strategy was proposed. This strategy was characterized by the fol-

Available online 9 May 2013

lowing: the establishment and utilization of the databases of parent compounds, known metabolites

and characteristic neutral losses; the comparison of base peak chromatograms and ClogPs; and the use

Keywords: n

of the HPLC–DAD–ESI-IT-TOF-MS technique. This strategy was ﬁrst applied to screen and identify the

Paeonia lactiﬂora

Paeoniﬂorin absorbed constituents and metabolites of PRR decoction and paeoniﬂorin in rats. In total, 13 new absorbed

(Epi)catechin constituents and 90 new metabolites of PRR decoction were detected. Among these metabolites, the struc-

Gallic acid tures of 70 metabolites were identiﬁed, and the conjugation types and structure skeletons of the other

Effective forms 20 metabolites were preliminarily determined. Moreover, 35 new metabolites of some constituents of

PRR, i.e., 22 new metabolites of paeoniﬂorin, 10 new metabolites of gallic acid-related compounds, 1

new metabolite of (epi)catechin-related compounds, and 2 new metabolites of other compounds, were

reported for the ﬁrst time. The results also indicated that (epi)catechin-related compounds, gallic acid-

related compounds and paeoniﬂorin were the main precursors of these metabolites. Phase I reactions

(dehydroxylation, decarboxylation, dehydrogenation) and phase II reactions (sulfation, glucuronidation

and methylation) were observed as the main metabolic pathways of PRR. According to the literature, the

11 absorbed constituents and 11 metabolites have various bioactivities. This study is the ﬁrst to explore

the absorption and metabolism of PRR decoction, and the result also is a notable improvement in the dis-

covery of paeoniﬂorin metabolites in vivo. These ﬁndings enhance our understanding of the metabolism

and Effective forms (the truly active structures) of PRR decoction and paeoniﬂorin.

1. Introduction Pall. or Paeonia veitchii Lynch. According to TCM theory, PRR can

remove heat from blood, can eliminate blood stasis, and can relieve

Paeoniae Radix Rubra (PRR), which is one of the commonly pain. Recent pharmacological studies indicate that PRR has various

used traditional Chinese medicines (TCMs) and is called Chishao bioactivities, such as anti-platelet agglutination, anti-thrombus,

in Chinese, is derived from the dried roots of Paeonia lactiﬂora anticoagulation, anti-atherosclerosis, heart and liver protection,

and anti-tumor properties [1]. With respect to the chemical con-

stituents of PRR, more than 150 compounds have been isolated

and identiﬁed to date, among which monoterpene glycosides and

∗

Corresponding author at: Department of Natural Medicines, School of Pharma- phenolic compounds are considered to be the major constituents

ceutical Sciences, Peking University Health Science Center, No. 38 Xueyuan Road, [2,3].

Haidian District, Beijing 100191, China. Tel.: +86 10 82801693;

Many investigations on PRR have been conducted in the ﬁelds of

fax: +86 10 82801693.

phytochemistry, pharmacology and clinical application. However,

E-mail address: [email protected] (S.-Q. Cai).

1 until now, little has been known about either how many and which

These authors contributed equally to this work.

J. Liang et al. / Journal of Pharmaceutical and Biomedical Analysis 83 (2013) 108–121 109

constituents are absorbed into the blood after the administration of (Shanghai, China). The purities of all standards were above 98%

PRR decoction or what the ﬁnal fate of the decoction is in vivo. This based on HPLC analysis (area normalization method). Acetonitrile

presents a signiﬁcant obstacle for deeper pharmacological mech- (Merck KGaA, Darmstadt, Germany) and formic acid (Mreda

anism studies and increasing the clinical applications of PRR. As Technology, Beijing, China) were of HPLC-grade. De-ionized water

a result, it is extremely urgent to thoroughly proﬁle the absorbed was prepared using a Milli-Q water puriﬁcation system (Millipore,

constituents and metabolites of PRR. Billerica, MA, USA). Analytical grade sodium carboxymethyl cel-

Paeoniﬂorin is one of the main active constituents in PRR lulose (CMC-Na) was purchased from Kermel Chemical Reagents

decoction and has been used as a phytochemical marker for the Development Centre (Kermel, Tianjin, China). Paeoniae Radix

quality control of PRR in Chinese Pharmacopoeia. Previous stud- Rubra (Sample No. 6902) was purchased from Baohua Chinese

ies also showed that paeoniﬂorin exhibits diverse pharmacological Herbal Medicine Procurement Department (Heihe City, Hei-

effects, such as anti-inﬂammatory, neuromuscular blocking, neuro- longjiang Province, China) in October 2011, and was authenticated

protective, antihyperglycemic and cognition-enhancing effects [4]. as the roots of P. lactiﬂora Pall. by Prof. Shao-Qing Cai (School of

However, studies have shown that paeoniﬂorin has a poor absorp- Pharmaceutical Sciences, Peking University). A voucher sample

tion rate and very low bioavailability (3–4%) when administrated (No. 6902) was deposited in the Herbarium of Pharmacognosy,

orally [5]. Previously, only 5 metabolites (7R-paeonimetabolin I, School of Pharmaceutical Sciences, Peking University.

7S-paeonimetabolin I, 7R-paeonimetabolin II, 7S-paeonimetabolin

II and paeoniflorgenin) of paeoniflorin had been found, and all of 2.2. Preparation of the PRR decoction and paeoniflorin solution

these metabolites were formed by intestinal bacteria [6–8]. To gain

a more comprehensive understanding of the metabolism of PRR and Two hundred grams of PRR was cut into small pieces (approxi-

the Effective forms (truly active structures) of paeoniﬂorin, further mately 0.5 cm length) and immersed in 2000 mL de-ionized water

research on the metabolism of paeoniﬂorin was deemed necessary. for 0.5 h, then heated to boiling and maintained for 0.5 h. The aque-

Proﬁling the absorbed constituents and metabolites of a TCM is ous extract was ﬁltered, and the residue was decocted twice with

difﬁcult, because (1) endogenous metabolites and proteins often 1600 mL and 1200 mL de-ionized water under the same conditions.

interfere. Additionally, the concentrations of the absorbed con- The ﬁltered supernatants were combined and evaporated to 300 mL

stituents and metabolites are quite low, increasing the complexity and then were freeze-dried, yielding 70 g dried powder by a SIM

of their determination in biological samples; (2) their structures FD8-10B freezer dryer (SIM International Group Co. Ltd., Newark,

are highly diverse, and many are unknown metabolites, whose DE, USA) for animal studies. The dried powder was characterized by

structural information has not been previously reported; hence, the LCMS technique, and the major constituents were identiﬁed

unequivocally identifying their structures is a signiﬁcantly chal- by comparison with reference substances or by the interpretation

lenging endeavor [9]. Therefore, a powerful and full-scale strategy of the LCMS data. The corresponding base peak chromatogram

is essential; such a strategy must be applicable for the systematic is shown in Supplementary data – Fig. A1. Paeoniﬂorin powder

screening and identiﬁcation of the absorbed constituents and the was suspended in 0.5% CMC-Na solution and ultrasonic vibrated

metabolites of TCMs. for 30 min (TP-150 ultrasonic cleaner, Tianpong Electricity New

The primary aim of studies on the absorption and metabolism of Technology Co., Beijing, China) before administration.

TCMs is to understand better their pharmacological mechanism and

clinical application. Absorbed constituents and circulating metabo- 2.3. Animals and drug administration

lites ultimately reach biological targets, so they should be the main

actors in explanations of the health effects of TCMs. Additionally, Eighteen male Sprague-Dawley rats (weighing in the range from

the circulating metabolites are likely to possess different biologi- 250 to 300 g) were obtained from the Experimental Animal Cen-

cal properties in comparison to those of their original compounds. ter of Peking University Health Science Center (Beijing, China).

Therefore, it is necessary to clarify the bioactivities possessed by The rats were randomly divided into 3 groups with 6 rats per

the absorbed constituents and metabolites. group (Group A, the PRR decoction group; Group B, the paeoni-

Accordingly, the aims of this study are (1) to map out a novel ﬂorin group; Group C, the blank group). The animals were housed

strategy, which will be a powerful tool for the systematic screening in metabolic cages (Type: DXL-DL, Suzhou Fengshi Laboratory Ani-

and identiﬁcation of the absorbed constituents and metabolites of mal Equipment Co. Ltd., Suzhou, China) and were maintained in

TCMs, (2) to identify the absorbed constituents in rat bio-samples an environmentally controlled breeding room for 7 days before the

and thoroughly proﬁle the metabolites of PRR in rat urine and experiments were performed. The freeze-dried PRR powder was

plasma, (3) to discover more of the in vivo metabolites of paeoni- dissolved in de-ionized water and administrated orally to the rats

ﬂorin, and (4) to summarize the bioactivities of the absorbed of group A at a dose of 3.486 g/kg body weight (equivalent to 9.96 g

constituents and metabolites of PRR decoction. crude drug per kg). The paeoniﬂorin solution was administrated by

oral gavage at a dose of 500 mg/kg body weight to group B. The

blank group was orally administrated with de-ionized water in the

2. Materials and methods same way. The rats of all the three groups were administrated once

per day (at 8:30 a.m.) for 3 days.

2.1. Chemicals and materials The animal experiments were carried out in accordance with

the Guide for the Care and Use of Laboratory Animals of the US

Paeoniﬂorin, epicatechin and ferulic acid were isolated and puri- National Institutes of Health; the animal experiment protocols

ﬁed in our laboratory. Catechin (Lot No. BCBB 18143 10898283) was were approved by the Biomedical Ethical Committee of Peking Uni-

purchased from Sigma Chemical Co. (St. Louis, MO, USA). Albiﬂorin versity (Approval No. LA2011-058).

(Lot No. 20100722), benzoylpaeoniﬂorin (Lot No. 20100715), oxy-

paeoniﬂorin (Lot No. 20100923), gallic acid (Lot No. 20110905) and 2.4. Urine and blood sample collection and pretreatment

benzoic acid (Lot No. 20110209) were purchased from Shanghai

Zhanshu Chemical Technology Co., Ltd. (Shanghai, China). 3-(3,4- After the ﬁrst and second administrations, urine samples were

Dihydroxyphenyl)propionic acid (Lot No. 10166081), m-cinnamic collected over 0–6 h, 6–12 h, 12–24 h periods respectively; then, all

acid (Lot No. LA10M21), p-cinnamic acid (Lot No. 380L06) and urine samples from the same group of rats were combined into

vanillic acid (Lot No. LT90M21) were purchased from J&K Scientiﬁc one sample. The blood samples (approximately 4 mL from each of

110 J. Liang et al. / Journal of Pharmaceutical and Biomedical Analysis 83 (2013) 108–121

2 rats at each collection time) were collected in heparinized tubes LCMSsolution Version 3.60, Formula Predictor Version 1.2, and

under anesthesia at 0.5 h, 1 h, 1.5 h after the third administration Accurate Mass Calculator (Shimadzu).

[10,11] and then were centrifuged for 15 min at 3000 rpm (Type: The chromatography separations were performed on a Phen-

3K15, Sigma Laborzentrifugen GmbH, Osterode am Harz, Germany) omenex Gemini C18 column (250 mm × 4.6 mm, 5 ␮m) protected

to obtain the plasma samples. All plasma samples from one group with a Phenomenex Security Guard column (4 mm × 3.0 mm,

were combined into one sample. Finally, all urine and plasma sam- 5.0 ␮m) (Phenomenex, Torrance, CA, USA), and the temperature

◦ ◦

ples were stored at −80 C in a Revco Value PLUS freezer (Thermo of the column oven was maintained at 30 C. The mobile phase

Fisher Scientiﬁc Inc., Asheville, NC, USA) until additional pretreat- consisted of water–formic acid (100:0.1, v/v) (A) and acetonitrile (B)

ment was performed. at the ﬂow rate of 1 mL/min. A gradient program was adopted, spe-

The urine samples were concentrated to dryness under reduced ciﬁc as 0% B at 0–12 min, 0–8% B at 12–33 min, 8% B at 33–37.5 min,

◦

pressure at 50 C using a Heidolph Laborota 4001 rotatory evapo- 8–12% B at 37.5–52.5 min, 12–25% at 52.5–82.5 min, 25–60% B at

rator (Heidolph Instruments GmbH & Co., Schwabach, Germany). 82.5–105 min, and 60–100% B at 105–120 min. At the end of each

Next, 1.00 g of dried sample was extracted with 4 mL methanol in run, 100% B was allowed to ﬂush the column for 10 min. For mass

an ultrasonic bath for 30 min and then was centrifuged at 5000 rpm detection, the mass spectrometer was programmed to carry out a

1 2 3

for 15 min. Afterward, the supernatant was transferred to a clean full scan over m/z 100–1000 (MS ) and m/z 50–1000 (MS and MS )

polypropylene tube. For the plasma samples, 5 mL was suspended in both positive ion (PI) and negative ion (NI) detection mode; the

in 20 mL methanol and then was ultrasonically vibrated for 20 min ﬂow rate was 0.2000 mL/min; the heat block and curved desolva-

◦

and centrifuged at 5000 rpm for 15 min to remove the protein. tion line temperature was 200 C; the nebulizing nitrogen gas ﬂow

Next, the supernatant was concentrated to dryness, and the residue was 1.5 L/min; the interface voltage was (+), 4.5 kV; (−), −3.5 kV;

was dissolved in 5 mL methanol and ﬁltered through a 0.45-␮m the detector voltage was 1.70 kV; the relative collision-induced

nylon ﬁlter (Tianjin Jinteng Experiment Equipment Co. Ltd., Tianjin, dissociation energy was 50%; and a triﬂuoroacetic acid sodium solu-

China). Afterward, the ﬁltrate was evaporated to dryness at room tion (2.5 mM) was used to calibrate the mass range from 50 to

temperature by nitrogen, followed by ultrasonic vibration with 1000 Da.

␮

200 L methanol for 30 min and then centrifugation at 5000 rpm

for 15 min. Finally, the supernatant was collected.

◦ 2.6. The strategy for systematic analysis of the absorbed

All pretreated samples were stored at −80 C and were ﬁltered

constituents and metabolites

through a 0.45-␮m nylon membrane once more before analysis. For

each sample, an aliquot of 5 ␮L was injected into the HPLC column.

To ﬁnd more absorbed constituents and metabolites, in this

work, we proposed a novel strategy for systematic metabolite

2.5. Instruments and conditions screening and identiﬁcation. The general procedures are shown in

Fig. 1.

HPLC–DAD–ESI-IT-TOF-MS analysis was performed on a Our strategy consisted of three steps, including (1) the construc-

Shimadzu HPLC instrument (two LC-20AD pumps, an SIL-20AC tion of a digital library of PRR, which included the parent compound

autosampler, a CTO-20A column oven, an SPD-M20A PDA detec- database, the known metabolites database, the characteristic neu-

tor, a CBM-20A system controller) coupled to an IT-TOF mass tral losses database, (2) searching the absorbed constituents and

spectrometer (Shimadzu, Kyoto, Japan) through an ESI interface. metabolites by comparing the extracted ion chromatograms (EICs)

All data were processed by Shimadzu software, speciﬁcally, and base peak chromatograms (BPCs) of the test group with those

Fig. 1. The strategy for the systematic screening and identiﬁcation of the absorbed constituents and metabolites of Paeoniae Radix Rubra.

J. Liang et al. / Journal of Pharmaceutical and Biomedical Analysis 83 (2013) 108–121 111

of the control group and collected the related information, and (3) (−1.58134) in comparison to the latter (ClogP = −1.38153). Gener-

the identiﬁcation of absorbed constituents and metabolites. The ally, the compound with the larger ClogP has the longer retention

absorbed constituents and known metabolites were identiﬁed with time in reverse phase HPLC, which can give valuable indications in

the digital library described above. The unknown metabolites were the discrimination of metabolite isomers [12].

elucidated in several steps: ﬁrst, the conjugation type was identi- According to previous literature, the metabolic reactions

ﬁed according to the characteristic neutral losses in MS; second, of some parent compounds exhibited position selectivity. For

the skeleton structure was elucidated by querying the CAS Registry example, epicatechin-3 -glucuronide was the major glucuronida-

and other bioinformatic tools using the molecular formula; third, tion metabolite of epicatechin in human urine; meanwhile,

the proposed structure was further conﬁrmed by MS fragment 3 -O-methyl-epicatechin and 4 -O-methyl-epicatechin were the

ions and maximum UV absorption; and ﬁnally, the exact conjuga- major methylated metabolites of epicatechin with 4 -O-methyl-

tion site was determined for some metabolites based on the ClogP epicatechin identiﬁed as the most abundant [13]. In rat urine, the

and predominant conjugation site. situation was found to be quite different: the major methylated

site of (epi)catechin was at position C-3 -OH, and (epi)catechin-5-

O-glucuronide was the major glucuronide of (epi)catechin [14,15].

3. Results and discussion

Comprehension of this information will be helpful in determining

the exact conjugation site.

3.1. Preparation for the systematic analysis of the absorbed

constituents and metabolites

3.1.3. The online chemical database and bioinformatic tools

Many of the unknown metabolites were quite difﬁcult to iden-

3.1.1. Construction of the digital library

tify, and the details recorded in our digital library were insufﬁcient

In many cases, some absorbed constituents and metabolites

to determine the structures of the unknown metabolites. Hence, in

already had been reported. Finding the possible absorbed con-

this case, we chose to obtain evidence and inspiration from interna-

stituents and metabolites in bio-samples is signiﬁcantly more

tional chemical databases and bioinformatic databases as follows.

effective when one is familiar with the detailed information of these

The Chemical Abstracts Service (CAS) database is a particularly

molecules. Hence, in this work, we constructed a digital library of

useful tool, because the quantity of substances that it contains

PRR (containing the roots of both P. lactiﬂora and P. veitchii) based on

is greater than 70 million. In our work, to elucidate some of the

the overall documents retrieved from the CAS Registry. The library

unknown metabolites, we ﬁrst listed the possible candidates by

included the parent compounds database, the known metabo-

querying the molecular formula in the CAS database. Then, the

lites database, and the characteristic neutral losses database. The

initial candidate list was narrowed by using orthogonal ﬁlters,

database was created by ChemFinder Ultra 12.0 (Cambridge Soft,

including the number and type of associated references. Finally,

Cambridge, MA, USA).

the candidate structures were further substantiated with additional

The database consisted of 21 searchable ﬁelds, including struc- n

information, such as the MS fragment ions, nuclear magnetic

ture, English name, Chinese name, molecular ID, molecular formula,

resonance data, relative retention times, and UV–vis diode array

molecular weight, accurate weight, measured weight, original plant

spectra. Meanwhile, for mutual veriﬁcation, we also queried other

or compound, physical properties, Chemical Abstracts Service (CAS)

bioinformatic tools, such as ChemSpider (www.chemspider.com)

registry number, ClogP, UV data, mass spectrum data, infrared data,

1 13 and METLIN (http://metlin.scripps.edu).

H NMR data, C NMR data, biological activities, biological activ-

ities references, references, and notes. The accurate weight and

3.1.4. The approach for tracing the metabolite origin

measured weight were calculated by Formula Predictor Version 1.2

PRR contains more than 150 constituents, each with speciﬁc

and Accurate Mass Calculator (Shimadzu). ClogP was calculated by

structure. When the PRR decoction was administrated orally to the

ChemBioDraw Ultra 12.0 (Cambridge Soft).

rats, some constituents were absorbed into circulation and ﬁnally

The records of 158 compounds were input into the parent

were converted into a mixture of metabolites of unknown origin. To

compounds database, and 59 known metabolites (their original

clearly understand the metabolic fate of the parent drug, we tried

compounds exist in PRR) were stored in the known metabolites

to trace the origin of the metabolites. In the study, we took two

database; Twenty-four characteristic neutral losses of metabolites

approaches. (1) We used a reference assistant: in previous reports,

were recorded in the characteristic neutral losses database.

the metabolic pathways of some known metabolites had already

been illustrated, by which we were able to infer the probable ori-

3.1.2. A new approach for the determination of the exact gin of these metabolites. (2) Next, we studied the metabolic fate of a

conjugation site of metabolite isomers single compound. Paeoniﬂorin was an important active constituent

Many isomers were commonly identiﬁed for a single compound in PRR. Until now, researchers only have found ﬁve metabolites of

in the corresponding biological matrices. The isomers had the same paeoniﬂorin. The related literature gives insufﬁcient information

molecular formula, and little differences presented in their MS for the identiﬁcation of more metabolites and the determination

spectra. Although the details of mass spectra can be used to narrow of the origin of potential paeoniﬂorin-related metabolites. In this

the choices of conceivable structures, it was still quite challeng- case, we administered paeoniﬂorin to the rats to facilitate the iden-

ing to discriminate the exact conjugation site of these isomers. In tiﬁcation of paeoniﬂorin-related metabolites.

this study, we made progress in locating the conjugation site by

reference to the predominant conjugation site and ClogP. 3.2. Characterization of the absorbed constituents

ClogP is the LogP (n-octanol/water partition coefﬁcient) value

calculated by the program ClogP (BioBype Co., Claremont, CA, USA). The peaks which appeared at the same position in the

This is an important parameter in estimating the hydrophobicity of LC–MS chromatograms of both the dosed rat bio-samples and

a compound; it describes the tendency of a compound to distribute the PRR decoction but not in the chromatograms of the con-

within aqueous and hydrophobic phases and is more accurate trolled rat bio-samples were regarded as absorbed constituents.

than other kinds of calculated LogPs. For example, catechin-4 - To increase the detection sensitivity, we used EICs and BPCs

O-glucuronide and catechin-7-O-glucuronide had the same LogP simultaneously. Once the peaks were determined to be absorbed

values (−0.38, as calculated by ChemBiodraw Ultra 12.0), but their constituents, they were further conﬁrmed by carefully compar-

ClogP values were different: the former had a smaller ClogP value ing their MS and MS data with those recorded in the parent

112 J. Liang et al. / Journal of Pharmaceutical and Biomedical Analysis 83 (2013) 108–121 isomer isomer

ellagic ellagic -

d - ␤ - F O

result

3,8-Dimethyl 3,7-Dimethyl

or or

- O glucopyranoside Identification Catechin Glucopyranosyl- paeonisuffrone Mudanpioside Desbenzoylpaeoniflorin Desbenzoylpaeoniflorin Desbenzoylpaeoniflorin Paeoniflorin Oxypaeoniflorin Benzoylpaeoniflorin 3,7- 3,8- Dihydroapigenin Galloylpaeoniflorin 1,8-Cineole-2- 4- Methyldesbenzoylpaeoniflorin acid acid I II

Plasma − + + + + + + + − + + − − + +

b Urine + − − + + + + + + + + + + −

(ppm)

error

0.00 0.16 2.20 3.45 1.38 2.47 5.6 5.33 3.95 1.52 4.06 3.01 3.02 NI − − − − − − − − − −

(Da)

pred.

NI 289.0718 405.1402 375.1297 375.1297 479.1559 495.1508 629.1876 329.0303 329.0303 631.1668 331.1762 435.1508

(Da)

meas.

289.0714 271.0601 271.0612 375.1276 495.1518 631.165 629.1875 389.1463 389.1453 2.57 435.1523 375.1277 NI 479.1559 329.0289 329.0298 331.1752 375.1302 375.1297 1.33 + decoction.

Rubra

(NI)

Radix

mode

ion 165.0573, 183.0691, 165.0571 291.0752, 183.0674, 159.0346, 165.0580, 367.0878, 313.0534, 187.0367, 119.0543,

Paeoniae

negative

195.0670, 327.1019, 195.0729, 327.1055, 165.0673 431.1343, 195.0660, 181.0963, 281.0645, 491.1207, 179.0814 405.1392 270.9874 161.0638, 270.9871 151.0060,

detected administration 345.1177, 345.1189, 345.1215, 449.1436, 273.1483, 227.0936, 182.0843, 299.0801, 553.1697, 203.0721, 197.0877, 509.1242, 298.9825 298.9825, 165.0206, 298.9813,

oral ions

after

343.1408, 375.1265, 375.1276, 375.1277, 479.1533, 465.1402, 583.1807, 331.1752, 389.1447, 245.0814, 123.0508 165.0612 314.0037, 314.0060, 314.0037, 613.1520, 168.0110 125.0273 359.1322, 177.0198, 65.0007

fragment substances.

urine

and

Major 289.0715, 405.1450, 389.1481, 421.1326, 421.1346, 421.1352, 525.1619, 495.1518, 629.1875, 329.0273, 329.0298, 271.0601, 631.1650, 377.1798, 435.1523, 239.0575, 271.0426, 109.0266, 151.0810 149.0640 137.0671, 151.0823 329.0289, 139.0921 107.0148, reference

plasma 6 9 8 10 10 10 11 12 5 7 10 12 8 8 15

O O O O O O O O O O O O O O O with

rat 14 24 24 24 24 24 28 28 10 10 12 28 26 32 32

H H H H H H H H H H H H H H H in

15 16 16 16 16 16 23 23 30 16 16 15 30 16 17 Formula C C C C C C C C C C

undetected.

comparison

−

constituents

(min) R t 44.147 23.21034.918 C 12.342 C 9.967 C 20.928 60.495 45.102 92.917 91.61390.480 C 94.27572.235 C 54.962 28.077 1 detected;

a a a a absorbed Conﬁrmed +,

a No. A1 A15 A2 A3 A5 A8 A9 A10 A13 A14 A11 A12 A6 A7 A4 b Table The

J. Liang et al. / Journal of Pharmaceutical and Biomedical Analysis 83 (2013) 108–121 113

compounds database. Finally, 15 compounds (Table 1) were (ClogP = −1.38153), it was supposed that (epi)catechin 4 -O-

identiﬁed as the absorbed constituents, including catechin (A1), glucuronide would have a smaller retention time. Accordingly,

glucopyranosyl-paeonisuffrone (A2), mudanpioside F (A3), des- M2 was identiﬁed as (epi)catechin-4 -O-glucuronide, and M3 was

benzoylpaeoniﬂorin (A4), desbenzoylpaeoniﬂorin isomer I (A5), inferred to be (epi)catechin-7-O-glucuronide.

desbenzoylpaeoniﬂorin isomer II (A6), paeoniﬂorin (A7), oxy- Based on the HRMS data for M4 and M5, the molecular for-

paeoniﬂorin (A8), benzoylpaeoniﬂorin (A9), 3,7-dimethyl ellagic mula of these metabolites was determined to be C22H24O12,

acid or 3,8-dimethyl ellagic acid (A10), 3,8-dimethyl ellagic acid with an additional CH2 unit in comparison with (epi)catechin

or 3,7-dimethyl ellagic acid (A11), dihydroapigenin (A12), galloyl- glucuronide. According to the previously conﬁrmed major meth-

paeoniﬂorin (A13), 1,8-cineole-2-O-␤-d-glucopyranoside (A14), ylation site of (epi)catechin [14,15], M4 (peak area: 901,792,917)

and 4-O-methyldesbenzoylpaeoniﬂorin (A15). was unequivocally identiﬁed as 3 -O-methyl (epi)catechin-5-O-

Among these 15 absorbed constituents, 11 were monoterpene glucuronide because of its massive quantity, and M5 (peak area:

glycosides with a cage-like pinane skeleton, and A13 had a galloyl 40,431,605) was determined to be 3 -O-methyl (epi)catechin 7- or

group; the others (A1, A10, A11, A12) were phenolic compounds. 4 -O-glucuronide.

The results indicated with high probability that monoterpene Metabolites M6 and M7 both had the molecular formula

glycosides and phenolic compounds are the major bioactive con- C16H16O9S, as calculated from their HRMS data. In their NI MS

−

stituents of PRR. Except for paeoniﬂorin (A7) and oxypaeoniﬂorin spectra, the [aglycon−H] at m/z 303.08 was observed, which was

(A8), the other 13 compounds were identiﬁed as the absorbed con- formed by a neutral loss of 79.95 Da (elemental composition: SO3)

− −

−

stituents of PRR decoction for the ﬁrst time. from [M H] . The fragmentation pathway of [aglycon−H] was

identical to that of methyl-O-(epi)catechin, as previously reported

3.3. Characterization of the metabolites in rat plasma and urine [16,17]. Additionally, with reference to the major methylation site

previously reported [14,15], we inferred that methylation occurred

Metabolite identiﬁcation was carried out according to the strat- at position C-3 -OH. In their NI MS spectra, characteristic frag-

egy described above (illustrated in Fig. 1). Accordingly, we searched ment ions at m/z 216.98 (elemental composition: C7H5O6S) and m/z

for all probable metabolites, and a total of 90 peaks were assigned 137.02 (elemental composition: C7H5O3) were observed, which

as PRR metabolites with their corresponding EICs shown in Fig. 4 were the typical A-ring fragment ions of (epi)catechin, as generated

and Supplementary data – Figs. A2–A6. The results showed that by Retro-Diels-Alder (RDA) cleavage [18]. Therefore, the sulfate

the metabolites varied greatly and were highly associated with group should be conjugated at the B-ring, although the exact site

their likely original compounds; hence, in the following discussion could not be determined. Finally, M6 and M7 were identiﬁed to be

of structure elucidation, the metabolites were categorized accord- 3 -O-methyl (epi)catechin 5-O- or 7-O-sulfate.

ing to their probable original compounds. In total, we found 27

(epi)catechin-related metabolites, 25 gallic acid-related metabo- 3.3.1.2. Aromatic acid metabolites. Phenylvalerolactones and phenyl-

lites, 6 (epi)catechin-related and gallic acid-related metabolites, 25 valeric acids. M8–M13 (their structures are shown in Fig. 2). In

n −

paeoniﬂorin metabolites, and 7 other metabolites. All 90 metabo- the MS spectra of M8 and M9, the same [aglycon−H] at m/z

lites were new metabolites of PRR, including 22 new metabolites of 207.06 was observed, formed by the neutral losses of 176.02 Da

paeoniﬂorin, 10 new metabolites of gallic acid-related compounds, (elemental composition: C6H8O6) and 79.95 Da (elemental com-

1 new metabolite of (epi)catechin-related compounds, and 2 new position: SO3), respectively, and the fragmentation behaviors

−

metabolites of other compounds. of [aglycon−H] were in high accordance with that of 5-(3,4-

␥

dihydroxyphenyl)- -valerolactone [19]. Thus, M8 and M9 must be

3.3.1. Identiﬁcation of the (epi)catechin-related metabolites 5-(3,4-dihydroxyphenyl)-␥-valerolactone glucuronide and 5-(3,4-

␥

A total of 33 compounds were assigned as metabolites orig- dihydroxyphenyl)- -valerolactone sulfate, respectively.

inating from (epi)catechin or polyphenols that can be degraded Based on the HRMS data, the molecular formulae of M10

into (epi)catechin in vivo. Seven were conjugates of (epi)catechin, and M11 were determined to be C12H14O7S and C18H22O10, with

while the other 26 compounds were identiﬁed as aromatic acid an additional CH2 unit in comparison with the formula of M9

metabolites. Their LC–MS data are summarized in Table 2 and and M8. This suggested that M10 and M11 were the methyl-

Supplemental Table A1, and the possible metabolic pathways are ation products of M9 and M8, respectively. A previous study

presented in Fig. 2. conﬁrmed that the skeleton structure of 5-(3,4-dihydroxyphenyl)-

␥-valerolactone was the metabolite of (epi)catechin, and the

3.3.1.1. Conjugates of (epi)catechin. M1–M7 (their structures are predominant methylation site of (epi)catechin was the C-3 -

shown in Fig. 2). The molecular formulae of metabolites M1, M2, OH. Because the C-3 position in (epi)catechin corresponded

␥

and M3 were determined to be C21H22O12 based on their HRMS to the C-3 position in 5-(3,4-dihydroxyphenyl)- -valerolactone,

data (Table 2 and Supplemental Table A1). In their NI MS spectra, we deduced that the metabolites of (epi)catechin should share

−

the [aglycon−H] at m/z 289.07 was observed, formed by the neu- the same conjugation rule. Therefore, M10 and M11 were

␥

tral loss of 176.03 Da (elemental composition: C6H8O6). Further, identiﬁed to be 5-(3-methoxyl,4-hydroxyphenyl)- -valerolactone

␥

the aglycon yielded characteristic fragment ions in high accordance sulfate and 5-(3-methoxyl,4-hydroxyphenyl)- -valerolactone glu-

with that of the reference catechin and epicatechin, so they were curonide, respectively [20].

preliminary identiﬁed as glucuronic acid conjugates of catechin or The molecular formula of M12 was C11H14O8S calculated from

−

epicatechin. To determine their exact conjugation sites, we adopted its HRMS data. It showed [aglycon H] at m/z 225.0762 formed

the approach described in Section 3.1.2. In rat urine, a previ- by a neutral loss of 79.95 Da (elemental composition: SO3) in the

2 −

−

ous study showed that glucuronidation of (epi)catechin preferably NI MS spectrum, and the [aglycon H] yielded fragment ions

occurred at the C-5-OH of the A-ring [14]. In our study, the quantity at m/z 207.0673 and m/z 163.0784, which were the character-

of M1 (peak area: 971,949,559) was signiﬁcantly larger than that istic fragment ions of 5-(3,4-dihydroxyphenyl)-␥-valerolactone.

of M2 (peak area: 27,784,282) and M3 (peak area: 157,685,397), This indicated that M12 was the ring-opening product of M9 [5-

so M1 was identiﬁed as (epi)catechin 5-O-glucuronide, and M2, (3, 4-dihydroxyphenyl)-␥-valerolactone sulfate, with molecular

M3 were inferred to be (epi)catechin 7- or 4 -O-glucuronide. formula of C11H12O7S]. As a result, M12 was identiﬁed as 5-(3,4-

Further, because (epi)catechin-4 -O-glucuronide had a smaller dihydroxyphenyl)-4-hydroxypentanoic acid sulfate. M13 showed

−

ClogP value (−1.58134) than that of (epi)catechin-7-O-glucuronide [M−H] at m/z 271.0280 in the NI MS spectrum, which indicated

114 J. Liang et al. / Journal of Pharmaceutical and Biomedical Analysis 83 (2013) 108–121 sulfate

sulfate glucuronide

acid

sulfate sulfate

glucuronide sulfate

acid sulfate acid

sulfate

sulfate sulfate

-glucuronide acid

-sulfate -sulfate sulfate O

O O - sulfate sulfate sulfate acid

acid

4 7- 7-

-glucuronide sulfate sulfate sulfate or or or

acid acid acid

-O sulfate -valerolactone -valerolactone I II

5 7- 5- 5- -sulfate -sulfate ␥ ␥

sulfate

acid O O acid acid

-sulfate

-sulfate -sulfate -valerolactone acid O O

␥ glucuronide disulfate sulfate glucuronide sulfate glucuronide sulfate

-O

glucoside isomer isomer

acid -glucuronide

sulfate acid acid acid acid -glucuronide -glucuronide gallic

sulfate benzoic

-diglucuronide

O O - acid-3- acid-4 acid phenylpropionic phenylacetic

O -glucuronide -glucuronide -sulfate -sulfate

result 5- 4 7-

O O O O acid (epi)catechin (epi)catechin (epi)catechin (epi)catechin sulfate -glucuronide

acid

sulfate

phenylpropionic phenylpropionic phenylacetic phenylacetic O sulfate glucuronide

disulfate disulfate

-methyl glucuronide acid

acid O

acid acid

4-Hydroxy

-Methyl -Methyl -Methyl -Methyl -Methylgallic -Methylgallic -Methylgallic -Methylgallic -Methylpyrogallol -Methylpyrogallol -Methylpyrogallol -Methylpyrogallol -Methylpyrogallol-2- -Methylpyrogallol-3- O O O O or O O O O O O O O O O

-Coumaric - - - - -Coumaric rats. Identiﬁcation (Epi)catechin (Epi)catechin (Epi)catechin 3 3 3 3 5-(3,4-Dihydroxyphenyl)- 5-(3,4-Dihydroxyphenyl)- 5-(3-Methoxyl-4-hydroxyphenyl)-valerolactone 5-(3-Methoxyl-4-hydroxyphenyl)-valerolactone 5-(3,4-Dihydroxyphenyl)-4-hydroxypentanoic 5-(3-Hydroxyphenyl)- 3-Hydroxy 4-Hydroxy 3,4-Dihydroxy 3-Hydroxy-4-methoxy-phenylpropionic 3-Methoxy-4-hydroxy-phenylpropionic m p Ferulic 3-Hydroxy 4-Hydroxy 3-Hydroxy-4-methoxyphenylacetic 3,4-Dihydroxy Protocatechuic Protocatechuic Protocatechuic 3-Hydroxy-4-methoxyl-benzoic Vanillic 3- Benzoyl 3-Hydroxyhipuric Gallic Gallic 4- 4- 4- 3- 3,4-Di- 2-Deoxy-pyrogallol-1- Pyrogallol-1- Pyrogallol-2- Pyrogallol-2- Pyrogallol-1- Pyrogallol-1,3- Pyrogallol Pyrogallol Pyrogallol- 2- 2- 2- 2- 1- 1-

plasma

and

acid-related acid-related acid-related acid-related acid-related acid-related

urine

in gallic gallic gallic gallic gallic gallic

and and and and and and

(PRR)

compound

related related related related related related Rubra

original

Radix

acid-related acid-related acid-related acid-related acid-related acid-related acid-related acid-related acid-related acid-related acid-related acid-related acid-related acid-related acid-related acid-related acid-related acid-related acid-related acid-related acid-related acid-related

Possible (Epi)catechin-related (Epi)catechin-related (Epi)catechin-related (Epi)catechin-related (Epi)catechin-related (Epi)catechin-related (Epi)catechin-related (Epi)catechin-related (Epi)catechin-related (Epi)catechin-related (Epi)catechin-related (Epi)catechin-related (Epi)catechin-related (Epi)catechin-related (Epi)catechin-related (Epi)catechin-related (Epi)catechin-related (Epi)catechin-related (Epi)catechin-related (Epi)catechin-related (Epi)catechin-related (Epi)catechin-related (Epi)catechin-related (Epi)catechin-related (Epi)catechin-related (Epi)catechin (Epi)catechin (Epi)catechin (Epi)catechin (Epi)catechin (Epi)catechin (Epi)catechin-related (Epi)catechin-related Gallic Gallic Gallic Gallic Gallic Gallic Gallic Gallic Gallic Gallic Gallic Gallic Gallic Gallic Gallic Gallic Gallic Gallic Gallic Gallic Gallic Gallic Paeoniae

− − Plasma − − − + + + + − − − − − − − + + + + + + − − − − − + + + + − − − − − + − + + − − − − − − − − − + − − − + − metabolites

b new

Urine + + + + − + + + + + + + + − + + + + + + + + + − + + + + + + + + + + + + + + + + + + + + + + + + + 90

(ppm)

0.86 0.33 0.74 0.67 0.21 0.42 0.00 0.00 0.00 0.58 0.00 0.70 0.00 0.00 1.22 2.30 + 4.88 + 3.24 4.00 + 1.72 1.82 3.83 1.29 1.62 6.70 3.82 3.34 2.12 1.46 2.93 1.08 2.35 2.35 2.33 2.27 3.67 1.83 1.30 4.76 4.86 1.72 2.02 1.38 2.28 3.99 2.66 2.68 2.74 4.56 2.28 2.74 Error − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − identiﬁcation

for

(NI) (Da)

301.0387 305.0337 230.9969 230.9969 301.0565 301.0565 204.9812 204.9812 Pred. 465.1038 465.1038 465.1038 479.1195 383.0442 383.0984 397.1140 271.0282 245.0125 275.0231 242.9969 273.0074 261.0074 246.9918 232.9761 315.0722 246.9918 246.9918 216.9812 297.0616 194.0459 248.9711 345.0463 183.0299 359.0620 262.9867 277.0024 188.9863 477.0886 284.9380 381.0133 315.0722 298.9537 218.9969 395.0290 218.9969 mode

(Da)

detection

ion

301.0380 305.0336 230.9966 230.9958 301.0533 301.0557 204.9815 204.9818 Meas. 465.1046 465.1034 465.1033 479.1197 479.1209383.0433 479.1195383.0975 2.92 397.1131 271.0280 245.0116 275.0236 242.9969 273.0069 261.0085 246.9906 232.9764 315.0722 246.9922 246.9913 216.9809 297.0614 194.0472 248.9711 345.0464 183.0306 359.0631 262.9861 277.0024 188.9867 477.0887 284.9382 284.9384381.0133 315.0722 284.9380298.9529 218.9963 1.40395.0272 218.9964 +

S S

S SS 383.0446S 287.0217 383.0442S S 287.0231 1.04 + S S 275.0220 275.0231 S negative

2 2 2

SS S 245.0122 261.0068 245.0125 261.0074 S 12 12 12 12 12 9 9 7 7 10 8 6 9 8 11 11 9 9 15 12 9 12 10 7 7 7

4 S SS 242.9970S S 242.9969SS 0.41 232.9757S S S 232.9761 + S SS 262.9858S 262.9867 S S S S S S S S 218.9962 218.9969 6 6 7 7 5 O O O O O O O O O O O O O O O O O O O O O O O O O O in 6 6 6 6 7 7 7 7 7 6 8 5 8 8 5 6 6 9 9 9 6 6 6

O O O O O O O O O O O O O O O NO O O O O O O O O O O O O O 22 22 22 24 24 16 16 20 12 14 22 14 12 16 14 14 16 14 14 22 14 16 16 12 12 10 10 10 10 8 8 8 8 10 8 6 6 8 8 6 9 6 8 8 8 10 6 6 6 6 6 8 8 8 8 H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H 21 21 21 22 22 16 16 17 11 12 18 11 11 9 9 9 10 10 9 9 10 8 8 9 8 7 7 13 8 8 7 13 9 7 13 8 8 14 8 9 6 12 12 6 6 18 6 6 12 13 7 7 13 7 7 Formula C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C obtained

data

770

(min) R 60.212 C 50.357 60.488 50.073 t 39.282 34.992 46.613 55.470 86.828 83.72787.618 C 86.488 55.867 C 67.448 91.217 79.67074.267 73.368 C 68.707 75.060 C 75.902 83.387 C 78.440 99.557 C 59.255 63.263 89.250 62.52752.510 35.105 C 59.368 53.980 51.150 58.40536.513 29.902 C 36.117 52.39728.938 57.045 C 77.577 61.232 38.318 22.733 32.13554.350 33.810 C 53.398 66.26732.628 31.738 C 76.240 36.570 74. 69.202 C a a a a a a a a a a a 2

No. M1 M2 M4 M3 M34 M22 M42 M43 M44 M45 M46 M55 M47 M41 M27 M29 M31 M21 M33 M32 M54 M23 M24 M5 M20 M37 M38 M39 M40 M35 M36 M15 M16 M7 M8 M26 M6 M17 M13 M25 M51 M52 M53 M30 M48 M49 M19 M50 M9 M10 M14 M28 M18 M11 M12 Table LC-ESI-IT-TOF-MS

J. Liang et al. / Journal of Pharmaceutical and Biomedical Analysis 83 (2013) 108–121 115 I II

I isomer isomer

sulfate

glucuronide

acid

isomer

acid acid

ethanol

I I

glucuronide glucuronide

glucuronide II I I

benzoic

ellagic ellagic

phenyl

result

sulfate sulfate glucuronide glucuronide glucuronide glucuronide glucuronide glucuronide glucuronide glucuronide glucuronide glucuronide glucuronide glucuronide glucuronide glucuronide glucuronide glucuronide acid

3 3 2 2 2 2 2 2 2 2 2 2 4 4 3 2 2 4 glucuronide glucuronide salicylate glucuronide sulfate

O O O O O O O O O O O O O O O O O O 3 3 4-Hydroxy O O 14 14 18 18 18 18 18 18 18 18 18 18 18 18 20 14 18 18

8 8 -Methyldebenzoylpaeoniﬂorin H H H H H H H H H H H H H H H H H H or -Paeonimetabolin H H -Paeonimetabolin O

10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 8 8 10 10 S R Identiﬁcation 3- Phenol Phenol Paeonimetabolin 7 7 Desbenzoylpaeoniﬂorin 4- Paeonimetabolin Paeonimetabolin C C C C C C C C C C C C C C C C C C 3,7,8-Trimethyl 3,7,8-Trimethyl Methyl 3,4-Dihydroxy Hippuric C C

compound

original

acid acid

acid-related acid-related acid-related

Possible Gallic Gallic Gallic Paeoniflorin Paeoniflorin Paeoniflorin Paeoniflorin Paeoniflorin Paeoniflorin Paeoniflorin Paeoniflorin Paeoniflorin Paeoniflorin Paeoniflorin Paeoniflorin Paeoniflorin Paeoniflorin Paeoniflorin Paeoniflorin Paeoniflorin Paeoniflorin Paeoniflorin Paeoniflorin Paeoniflorin Paeoniflorin Paeoniflorin Paeoniflorin Paeoniflorin Ellagic Ellagic Uncertain Uncertain Uncertain Uncertain Uncertain

Plasma + + + − + − + + − − + + + + + − + + + − + + + + − − + − + − + + + − −

Urine − + + + + + + − + + + + + + + + + + + + + + + + + + + + − + + + + + +

(ppm)

5.11 2.97 5.60 4.29 1.61 1.53 4.21 4.06 2.03 2.32 3.77 1.16 2.61 2.32 1.16 4.35 1.45 4.77 4.24 4.75 3.92 2.45 4.59 7.58 2.45 2.61 2.32 3.47 1.51 1.01 4.06 3.45 4.62 1.72 4.49 Error − − − − − − − − − − − − − − − − − − − − − − − − − − − compounds

(Da)

Pred. 137.0303 269.0667 172.9914 199.0976 197.0819 197.0819 375.1297 435.1508 373.1140 373.1140 261.0438 261.0438 345.1555 345.1555 345.1555 345.1555 345.1555 345.1555 345.1555 345.1555 345.1555 345.1555 377.1453 377.1453 379.1610 357.1191 327.0722 327.0722 343.0459 519.0780 327.0722 233.0125 178.0510 345.1555 345.1555 acid-related

gallic

(Da)

9920

Meas. 137.0296 269.0659 172. 199.0979 197.0821 197.0827 375.1276 435.1523 373.1124 373.1134 261.0434 261.0427 345.1541 345.1548 345.1547 345.1542 345.1551 345.1546 345.1547 345.1551 345.1540 345.1550 377.1435 377.1437 379.1592 357.1177 327.0714 327.0707 343.0433 519.0804 327.0714 233.0129 178.0518 345.1546 345.1547 paeoniﬂorin,

, e.g.

S S PRR,

S 7 10 10 10 10 8 8 8 8 8 8 8 8 8 8 10 10 10 9 9 9 8 9 8 8 4 4 4 6 6 14

3 S 6 O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O 3 4 of O

O O NO 14 24 26 22 22 26 26 26 26 26 26 26 26 26 26 26 26 28 22 16 16 12 20 16 26 26 16 14 14 14 14 6 6 10 9 H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H 6 12 6 10 10 10 16 17 16 16 10 10 16 16 16 16 16 16 16 16 16 16 16 16 16 16 14 14 17 23 14 8 9 16 16 Formula C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C constituents

some

(min) of

9.9665 R 60.778 70.950 80.060 30.015 t 76.718 38.942 71.565 82.707 37.525 42.245 28.077 34.822 24.998 67.222 74.663 67.562 68.298 73.765 75.513 76.637 77.817 81.573 26.787 28.995 36.627 49.012 32.523 96.648 88.457 49.012 44.818 41.070 72.13 85.762 undetected. )

, −

metabolites

Continued (A5) (A15) (

new a a a a a a a a a a a a a a a a a a a a a a a a

detected; 2

35 +, a No. M56 M64 M89 M87 M61 M60 M81 M67 M68 M69 M70 M71 M72 M73 M74 M75 M76 M77 M78 M79 M80 M82 M90 M62 M83 M84 M58 M59 M85 M88 M86 M65 M66 M57 M63 b Table

116 J. Liang et al. / Journal of Pharmaceutical and Biomedical Analysis 83 (2013) 108–121

Fig. 2. The possible metabolic pathways of (epi)catechin and its related compounds in rats orally administered with Paeoniae Radix Rubra decoction.

that its molecular formula was C11H12O6S. In its NI MS spectrum be proposed that the C-3-OH was the preferred methylation site,

−

[aglycon−H] at m/z 191.0726 was observed, formed by the neu- based on the argument described above in the structure elucidation

tral loss of 79.95 Da (elemental composition: SO3). Further, the of M10 and M11. Accordingly, M18 (peak area: 290,399,813) was

−

fragmentation behavior of [aglycon−H] was the same as that of present in a massive quantity, assigned as 3-methoxyl-4-hydroxy-

␥

5-(3-hydroxyphenyl)- -valerolactone. Accordingly, M13 was iden- phenylpropionic acid sulfate, and M17 (peak area: 16,774,823) was

tiﬁed as 5-(3-hydroxyphenyl)-␥-valerolactone sulfate. identiﬁed to be 3-hydroxy-4-methoxyl-phenylpropionic acid sul-

Phenylpropionic acids. M14–M18 (their structures are shown fate.

in Fig. 2). The molecular formulae of M14 and M15 were deter- Cinnamic acids. M19–M21 (their structures are shown in Fig. 2).

mined to be C9H10O6S based on their HRMS data (Table 2). In Based on the HRMS data of M19 and M20, the molecular formula

2 −

their NI MS spectra, the [aglycon−H] at m/z 165.0569 was of these metabolites was determined to be C9H8O6S. In their NI

2 −

−

observed, implying the neutral loss of 79.95 Da (elemental com- MS spectra, the [aglycon H] at m/z 163.04 was observed, imply-

−

position: SO3) from [M−H] . The fragmentation pathways of ing the neutral loss of 79.95 Da (elemental composition: SO3). The

− −

[aglycon H] were identical to those of hydroxy phenylpropionic fragmentation behaviors of [aglycon H] were the same as that

acid, as previously reported [19]. Thus, we preliminarily identi- of coumaric acid. In a previous study, it was conﬁrmed that m-

ﬁed M14 and M15 as hydroxy phenylpropionic acid sulfate. It coumaric acid is a major metabolite in rat in comparison with

has been reported that 3-hydroxy phenylpropionic acid is a pre- p-coumaric acid [16]. Therefore, M19 (peak area: 839,949,768) with

ferred metabolite of (epi)catechin [21]; meanwhile, we found a massive quantity was identiﬁed as m-coumaric acid sulfate, and

M14 (peak area: 453,044,549) to be more abundant than M15 M20 (peak area: 45,034,536) was inferred to be p-coumaric acid

(peak area:77,107,498) in the rat urine. Accordingly, we assigned sulfate.

−

M14 as 3-hydroxy phenylpropionic acid sulfate and M15 as 4- M21 showed [M H] at m/z 273.0069 (calculated to be

− − n

−

hydroxy phenylpropionic acid sulfate. M16 showed [M−H] at m/z C10H9O7S) and [aglycon H] at m/z 193.0513 in its MS spectrum,

− −

−

261.0068 (calculated to be C9H9O7S), and yielded [aglycon−H] at and the [aglycon H] produced characteristic fragment ions at m/z

−

m/z 181.0534. The fragmentation behavior of [aglycon−H] was in 134.0408, m/z 178.0278, m/z 149.0644, which were in high accor-

high accordance with that of 3,4-dihydroxy phenylpropionic acid dance with those of the reference standard ferulic acid. Accordingly,

[19]. Hence, M16 was unequivocally identiﬁed as 3,4-dihydroxy M21 was identiﬁed to be ferulic acid sulfate.

phenylpropionic acid sulfate. Phenylacetic acids. M22–M25 (their structures are shown in

The molecular formulae of M17 and M18 were determined to be Fig. 2). M22 and M23 had the same molecular formula (C8H8O6S),

2 2

C10H12O7S based on their HRMS data. In their NI MS spectra, the as calculated from their HRMS data. In their NI MS spectra, the

− −

[aglycon H] at m/z 195.06 (C10H11O4) formed by the neutral loss [aglycon H] at m/z 151.04 was observed, indicating a neutral loss

−

of 79.95 Da (elemental composition: SO3) was observed, which had of 79.95 Da (elemental composition: SO3), and the [aglycon H]

−

an additional CH2 unit in comparison with [aglycon−H] of M16. yielded the same characteristic ions as those of hydroxy phenyl-

This indicated that the aglycons of M17 and M18 were methyla- acetic acid. In a previous study, it was conﬁrmed that 3-hydroxy

tion products of 3,4-dihydroxy phenylpropionic acid. Further, it can phenylacetic acid forms in a larger quantity in comparison with

J. Liang et al. / Journal of Pharmaceutical and Biomedical Analysis 83 (2013) 108–121 117

−

4-hydroxy phenylacetic acid in rat urine [16]. Hence, M22 (peak [aglycon−H] at m/z 167.03 formed by the neutral loss of 79.95 Da

−

area: 2,121,993,281) with the larger amount, was identiﬁed (elemental composition: SO3) from [M−H] was observed. Com-

as 3-hydroxy phenylacetic acid sulfate, and M23 (peak area: pared with M26 and M27, a CH2 unit was added to M29 and M30;

779,174,736) was assigned as 4-hydroxy phenylacetic acid sulfate. thus, it can be ascertained that a methyl group was linked to the

The molecular formula of M24 was determined to be C9H10O7S protocatechuic acid in M29 and M30. It has been reported that

−

based on its HRMS data. The [aglycon−H] at m/z 181.0528 was vanillic acid is a major metabolite of catechins in rat urine [21].

detected as the base peak in its NI MS spectrum, formed by As a result, M30 (peak area: 427,840,748) with a signiﬁcant quan-

the neutral loss of 79.95 Da (elemental composition: SO3). Addi- tity, was assigned as vanillic acid sulfate, and M29 (peak area:

−

tionally, [aglycon H] yielded fragment ions at m/z 137.0658, 374,654,520) was inferred to be 3-hydroxy-4-methoxyl-benzoic

m/z 122.042, which are the typical fragment ions of 3-hydroxy- acid sulfate.

4-methoxyphenylacetic acid, as recorded in the METLIN database M31 has the molecular formula of C7H6O6S, calculated from its

n −

(http://metlin.scripps.edu/). Accordingly, M24 was identiﬁed as HRMS data. In the NI MS spectrum [aglycon−H] at m/z 137.0280

3-hydroxy-4-methoxyphenylacetic acid sulfate. M25 showed and fragment ion at m/z 93.0377 were observed, which are the

− −

[M H] at m/z 246.9906 (calculated to be C8H7O7S) in the NI MS typical fragment ions of 3- or 4-hydroxybenzoic acid [21]. There-

−

spectrum, showing [aglycon−H] at m/z 167.0364 (calculated to be fore, M31 was assigned as 3- or 4-hydroxybenzoic acid sulfate.

C8H7O4) in the NI MS spectrum, which was formed by the loss of According to the HRMS data of M32, its molecular formula was

− 2 −

−

79.95 Da (elemental composition: SO3) from [M H] . The fragmen- determined to be C13H14O8. In its NI MS spectrum [aglycon−H] at

− −

−

tation pathways of [aglycon H] were in high accordance with that m/z 121.0355 and [glucuronyl−H] at m/z 175.0264 were observed.

of 3,4-dihydroxy phenylacetic acid, as previously reported. There- Additionally, the aglycon was veriﬁed as benzoic acid by compar-

fore, M25 was assigned as 3,4-dihydroxy phenylacetic acid sulfate ing with a reference substance. Therefore, M32 was identiﬁed to be

[19]. benzoyl glucuronide. M33 was identiﬁed as 3-hydroxyhipuric acid,

Benzoic acids. M26–M33 (their structures are shown in Fig. 2). In which is one of the major metabolites of catechin.

−

their NI MS spectra, M26–M28 showed [M−H] at m/z 232.9757,

m/z 232.9764 and m/z 315.0722, which indicated that their molec- 3.3.2. Identiﬁcation of the gallic acid-related metabolites

ular formulae were C7H6O7S, C7H6O7S and C13H16O9, respectively. In this work, a total of 31 compounds were identiﬁed as the

− 2

−

The same [aglycon H] at m/z 153.02 was observed in their MS metabolites of gallic acid or as the metabolites of related com-

spectra, suggesting that M26 and M27 were sulfate conjugates and pounds that can be degraded into gallic acid in vivo. Among them,

M28 was a glucoside conjugate. Further, the fragmentation behav- 7 compounds were conjugates of gallic acid, 15 compounds were

−

iors of their [aglycon H] were the same as that of protocatechuic pyrogallol-related metabolites, and 9 compounds were benzoic

acid [19,22]. Therefore, M28 was determined to be protocatechuic acids and others. Their LC–MS data are summarized in Table 2 and

acid glucoside, and M26 (peak area: 106,695,266) which exhibited Supplemental Table A1, and the possible metabolic pathways of

a signiﬁcant quantity, was assigned as protocatechuic acid-3- gallic acid and the related compounds are presented in Fig. 3.

O-sulfate with reference to previous reports, while M27 (peak

area: 25,350,969) was identiﬁed to be protocatechuic acid-4-O- 3.3.2.1. The conjugates of gallic acid. M34–M40 (their structures are

sulfate. shown in Fig. 3). The molecular formulae of M34 and M35 were

The molecular formulae of M29 and M30 were determined determined to be C7H6O8S and C13H14O11 according to their HRMS

2 2 −

to be C8H8O7S according to their HRMS data. In their NI MS data. Their MS spectra showed [aglycon−H] at m/z 169.0167,

Fig. 3. The possible metabolic pathways of gallic acid and its related compounds in rats orally administered with Paeoniae Radix Rubra decoction.

118 J. Liang et al. / Journal of Pharmaceutical and Biomedical Analysis 83 (2013) 108–121

with fragmentation behaviors that were the same as that of the molecular formula of their aglycon was C6H6O. Therefore, M57

reference standard gallic acid. Accordingly, M34 and M35 were and M58 were assigned as phenol glucuronide and phenol sulfate,

assigned as gallic acid sulfate and gallic acid glucuronide, respec- respectively. Additionally, M26–M31, which were identiﬁed in Sec-

tively. tion 3.3.1 as metabolites of (epi)catechin-related components, also

M36 showed characteristic ions at m/z 183.0306, m/z 168.0098, were regarded as metabolites of gallic acid and related ingredients,

and m/z 124.0188; hence, it was identiﬁed as 4-O-methylgallic acid. according to previous literature.

M37–M39 had the molecular formulae of C8H8O8S, C14H16O11, and

2 −

C8H8O8S, respectively. In their NI MS spectra [aglycon−H] at 3.3.3. Identiﬁcation of the paeoniﬂorin metabolites

m/z 183.03 was observed, formed by the neutral loss of 79.95 Da, In our study, we discovered 25 metabolites with the likely origin

176.03 Da, and 79.95 Da, respectively. The fragmentation pathways of paeoniﬂorin, found by comparing the rat bio-samples of the PRR

−

of [aglycon H] were in high accordance with that of methylgallic decoction group with that of the blank group. Then, to clarify their

acid. In a previous study, it was conﬁrmed that 4-O-methylgallic origin, we administrated rats with paeoniﬂorin, and searched the

acid is a major metabolite in comparison with 3-O-methylgallic metabolites in the rat bio-samples of the paeoniﬂorin group, ﬁnd-

acid [23]. Therefore, M37 (peak area: 1,210,546,832) with the sig- ing that all 25 metabolites existed in the paeoniﬂorin bio-samples

niﬁcant quantity, was identiﬁed as 4-O-methylgallic acid sulfate, (shown in Fig. 4 and Fig. A2). Except for M59 (paeonimetabolin II),

while M38 and M39 (peak area: 343,791,941) were assigned to be M60 (7S-paeonimetabolin I) and M61 (7R-paeonimetabolin I), the

4-O-methylgallic acid glucuronide and 3-O-methylgallic acid sul- other 22 metabolites were discovered as metabolites of paeoni-

fate, respectively. Based on its HRMS data, the molecular formula ﬂorin for the ﬁrst time. The process of structural elucidation was

of M40 was determined to be C9H10O5, which had an additional performed as follows. The LC–MS data are summarized in Table 2

CH2 unit in comparison with 4-O-methylgallic acid. Thus, M40 was and Supplemental Table A1, and the EICs of these metabolites are

identiﬁed as 3,4-di-O-methyl gallic acid sulfate. shown in Fig. 4 and Supplemental Fig. A2.

3.3.2.2. The pyrogallol-related metabolites. M41–M55 (their struc- 3.3.3.1. The known metabolites of paeoniﬂorin. M59, M60 and M61

tures are shown in Fig. 3). M41 was identiﬁed to be 2-deoxy- were identiﬁed as paeonimetabolin II, paeonimetabolin I isomer I,

pyrogallol-1-O-sulfate, according to a previous report. According to paeonimetabolin I isomer II, respectively, according to their HRMS

the NI HRMS data, the molecular formulae of M42–M49 were deter- data (Table 2 and Supplemental Table A1). According to a previ-

mined to be C12H14O9, C12H14O9, C6H6O6S, C6H6O6S, C18H22O15, ous report, 7S-paeonimetabolin I is considered to be the ﬁrst major

C6H6O9S2, C6H6O9S2, and C12H14O12S, respectively. These metabo- metabolite of paeoniﬂorin; hence, M60 (peak area: 65,721,144)

−

lites yielded the same [aglycon−H] (elemental composition: with the larger quantity was identiﬁed as 7S-paeonimetabolin

C6H5O3) by neutral losses of SO3 or C6H8O6, and the aglycon of I, while M61 (peak area: 19,724,250) was inferred to be 7R-

each was ascertained as pyrogallol based on reference [23]. Among paeonimetabolin I [6–8].

these metabolites, two isomer groups were identiﬁed. M42 and

M43 were determined to be pyrogallol glucuronides. Moreover, the 3.3.3.2. New metabolites of paeoniﬂorin. The molecular formulae of

retention time of M43 was smaller than that of M42, so we deduced M62 (A5) and M63 (A15) were determined to be C16H24O10 and

that M43 should have the smaller ClogP value. Considering the C17H26O10, respectively. In their NI MS spectra, the fragmentation

−

fact that pyrogallol-2-O-glucuronide had the smaller ClogP value behaviors of [M+HCOOH−H] were identical to that of desben-

−

( 2.10004) in comparison to that of pyrogallol-1-O-glucuronide zoylpaeoniﬂorin isomer I and 4-O-methyldebenzoylpaeoniﬂorin,

(ClogP = −1.78004), we determined that M42 was pyrogallol-1-O- respectively [24], which were recorded in our digital library,

glucuronide and M43 was pyrogallol-2-O-glucuronide. Similarly, created from the CAS database. In Section 3.1, both metabo-

M44 (tR = 32.135 min) was assigned as pyrogallol-2-O-sulfate lites already were identiﬁed as absorbed constituents. However,

(ClogP = −1.18474), and M45 (tR = 54.35 min) was conﬁrmed to be they also were found to exist in the rat bio-samples of the

pyrogallol-1-O-sulfate (ClogP = −1.11474). According to the agly- paeoniﬂorin group. Therefore, it can be concluded that some pro-

con structure and the above-mentioned metabolic pathways, we totype constituents, i.e., desbenzoylpaeoniﬂorin isomer I (M62)

concluded that M46–M49 were pyrogallol-1,3-O-diglucuronide, and 4-O-methyldebenzoylpaeoniﬂorin (M63), also can be pro-

pyrogallol disulfate isomer I, pyrogallol disulfate isomer II, and duced through the metabolism of paeoniﬂorin in vivo. M64 and

pyrogallol-O-glucuronide sulfate, respectively. M65 were found to have the molecular formula C16H22O10, as

Metabolites M50–M55 had product ions at m/z 139.04 and calculated from their HRMS data. In their NI MS spectra, they

− −

m/z 124.02, which indicated that the aglycone of these metabo- showed [aglycon−H] and [glucuronyl−H] at m/z 197.0482 and

lites was methyl pyrogallol. Based on previous reports, we m/z 175.03, respectively; hence, we tentatively assigned these

inferred that 2-O-methylpyrogallol was the major metabolite metabolites as paeonimetabolin I glucuronides.

in comparison with 1-O-methylpyrogallol. Finally, it was ascer- For most of the new paeoniﬂorin metabolites, we found that

tained that M50–M53 were 2-O-methylpyrogallol glucuronide, their aglycon ions were difﬁcult to fragment in either the ESI ion

2-O-methylpyrogallol disulfate, 2-O-methylpyrogallol sulfate, and source or APCI ion source. As a result, we can only preliminarily

2-O-methylpyrogallol glucuronide sulfate, respectively [23]. The determine their conjugation types and structural skeletons. M66

− −

isomers of M54 and M55 were discriminated by their ClogP values. and M67 showed [M−H] at m/z 261.04 and [aglycon−H] at m/z

M54 (tR = 74.77 min) was identiﬁed as 1-O-methylpyrogallol-2-O- 181.08 in their NI MS spectra, which indicated that they were

sulfate (ClogP = −0.53873); M55 (tR = 69.202 min) was inferred to sulfates and that their aglycons had the elemental composition

be 1-O-methylpyrogallol-3-O-sulfate (ClogP = −0.7873). of C10H14O3. Based on the HRMS data, the molecular formulae

of M68–M77 were calculated to be C16H26O8. In their NI MS

−

3.3.2.3. Benzoic acids and others. M56–M58, M26–M31 (their spectra [glucuronyl−H] at m/z 175.02 was observed, so these

structures are shown in Fig. 3). Based on the HRMS data, the isomers were identiﬁed as glucuronides of the aglycons which

molecular formula of M56 was determined to be C6H6O3, and had the molecular formula of C10H18O2. Isomers M78 and M79

−

the characteristic fragment ions at m/z 137.0274 and m/z 93.0218 with the molecular formulae C16H26O10 showed [glucuronyl−H]

indicated that M56 was likely 3- or 4-hydroxybenzoic acid. Accord- at m/z 175.02 in their NI MS spectra. Thus, they were inferred

ing to the HRMS data, the molecular formulae of M57 and M58 to be glucuronide, whose aglycons had the molecular formula of

were determined to be C12H14O7 and C6H6O4S, meanwhile the C10H18O4. By the same method, M80 (C16H28O10), M81 (C16H22O9),

J. Liang et al. / Journal of Pharmaceutical and Biomedical Analysis 83 (2013) 108–121 119

Fig. 4. The negative extracted ion chromatograms (EICs) at m/z 199.09, 197.08, 421.13, 373.11, 261.04, 345.15, 377.14, 379.15, 357.11, 327.07 of rat urine (A, the PRR group;

B, the paeoniflorin group; C, the blank group) for the 24 identified metabolites of paeoniflorin.

M82 (C14H16O9), M83 (C14H16O9) were identiﬁed as glucuronic sulfation, glucuronidation and methylation were detected as the

acid conjugates, and the molecular formulae of their aglycons were main phase II reactions of (epi)catechins, gallic acids, paeoni-

C10H20O4, C10H14O3, C8H8O3, and C8H8O3, respectively. ﬂorin and their related compounds; phase I reactions, such

as dehydroxylation, dehydrogenation, were observed in the

metabolism of (epi)catechins and gallic acids. Further, decar-

3.3.4. Identiﬁcation of the other metabolites

boxylation is the typical reaction of gallic acids. According

M84–M90 (their LC–MS data are summarized in Table 2). M84

to the metabolites identiﬁed here, the metabolic pathways of

and M85 were assigned as 3,7,8-trimethyl ellagic acid and 3,7,8-

the major compounds were proposed and are illustrated in

trimethyl ellagic acid glucuronide based on their fragmentation

Figs. 2 and 3.

pathways. They were metabolites of dimethyl ellagic acid (A10 or

A11). M86–M90 were identiﬁed (shown in Table 2), but their ori-

gins were undetermined because they could be formed via several

3.5. Bioactivities of the absorbed constituents and the metabolites

pathways.

of PRR decoction

3.4. Analysis of the metabolite proﬁle of PRR decoction As mentioned before, a total of 15 absorbed constituents

and 90 metabolites of PRR decoction were detected and iden-

The absorbed constituents and metabolites identified in our tified in vivo. Among these 90 metabolites, 25 were confirmed

work provided a global view of the absorption and metabolism as the metabolites of paeoniﬂorin, which also might be the

of PRR decoction in rats. Monoterpene glycosides and phenolic metabolites of other paeoniﬂorin-related compounds (such as

compounds were found to be the main absorbed constituents. oxypaeoniﬂorin, benzoylpaeoniﬂorin); 27 metabolites might orig-

In total, 25 (epi)catechin-related metabolites, 24 gallic acid- inate from (epi)catechin or (epi)catechin-related compounds;

related metabolites, 5 (epi)catechin-related and gallic acid-related 25 metabolites probably originated from gallic acid or gallic

metabolites, 24 paeoniﬂorin metabolites and 6 other metabolites acid-related compounds; 6 metabolites can be produced from

were identiﬁed in rat urine. Additionally, 10 (epi)catechin-related both (epi)catechin-related and gallic acid-related compounds.

metabolites, 8 gallic acid-related metabolites, 4 (epi)catechin- Obviously, many metabolites could be formed from the constituent

related and gallic acid-related metabolites, 16 paeoniﬂorin of PRR decoction in vivo.

metabolites and 4 other metabolites were found in rat plasma. However, whether the original constituents are the Effec-

The fact that almost all of the compounds found in the plasma tive forms or the metabolites are the Effective forms or the

sample also were present in the urine sample indicates that both are the Effective forms of the PRR decoction became

urine was more suitable than plasma for metabolite proﬁling questions. Furthermore, whether only some substances among

research. the original constituents and the metabolites are the Effec-

The constituents absorbed into rat plasma after PRR admin- tive forms or the aggregate of both are the Effective forms

istration, such as (epi)catechin and paeoniﬂorin, were further are questions. Is it possible that the summation of these orig-

metabolized by various metabolic enzymes. In our work, inal constituents and metabolites play the core role in the

120 J. Liang et al. / Journal of Pharmaceutical and Biomedical Analysis 83 (2013) 108–121

pharmacological effects of the PRR decotion? All these ques- Acknowledgments

tions need further studies and clarifying these questions may

make an important breakthrough for understanding of the real This study was ﬁnancially supported by Key Program of National

substances and mechanisms of the pharmacological effects of Natural Science Foundation of China (Grant No. 30830120), Special-

TCMs. ized Research Fund for the Doctoral Program of Higher Education

To provide a preliminary answer to these questions, we made (Grant No. 20110001120036), and China Postdoctoral Science

a summary of the bioactivities of the absorbed constituents and Foundation (Grant No. 20080430293). We are grateful to Dr. Jun

metabolites, which was carried out based on the overall litera- Li for her routine management and careful maintenance of the

ture retrieval. It was found that 11 of 15 absorbed constituents LCMS-IT-TOF instrument.

and 11 of 90 metabolites had been reported to possess various

bioactivities (shown in Supplemental data Table A2). Among these Appendix A. Supplementary data

compounds, 11 absorbed constituents (A1, A2, A4–11, A13) and

6 metabolites (M1, M2, M4, M28, M31, M60) showed bioactivities Supplementary data associated with this article can be found, in

related to the pharmacological actions of PRR decoction. The related the online version, at http://dx.doi.org/10.1016/j.jpba.2013.04.029.

pharmacological effects were recorded as follows: A1, A7, A8, A9

and A13 can inhibit blood platelet aggregation; A4–6, A7, A10–11 References

and M1 showed noticeable neuroprotection effects; A7, M1 and

M2 brought cardiovascular beneﬁts; A4–6 and M2 exhibited anti- [1] J. Ruan, Z. Zhao, Q. Zeng, Z. Qian, Recent advances in study of components and

pharmacological roles of Radix Paeoniae Rubra, Zhongguo Yaolixue Tongbao

hyperglycemic effects; A7 can ameliorate hepatic ﬁbrosis; A1 and

19 (2003) 965–970.

A7 showed dramatic analgesic effect; A1, A2, A4–8, A10–11 and

[2] H. Ding, H. Lin, C. Teng, Y. Wu, Phytochemical and pharmacological studies on

A13 were reported to possess anti-inﬂammatory activity; A1, A7, Chinese Paeonia species, J. Chin. Chem. Soc. (Taipei) 47 (2000) 381–388.

[3] S. Li, J. Song, F.F.K. Choi, C. Qiao, Y. Zhou, Q. Han, H. Xu, Chemical proﬁling of

M4 and M28 exhibited radical scavenging effects; A1, M1, M2 and

Radix Paeoniae evaluated by ultra-performance liquid chromatography/photo-

M31 showed antioxidant activity; and M60 demonstrated anticon-

diode-array/quadrupole time-of-ﬂight mass spectrometry, J. Pharm. Biomed.

vulsant activity. Anal. 49 (2009) 253–266.

[4] H. Wu, Z. Zhu, G. Zhang, L. Zhao, H. Zhang, D. Zhu, Y. Chai, Comparative pharma-

As a result, we consider A1, A2, A4–11 and A13 to be the

cokinetic study of paeoniﬂorin after oral administration of pure paeoniﬂorin,

active constituents of PRR. Because M1, M2, M4, M28, M31 and

extract of Cortex Moutan and Shuang-Dan prescription to rats, J. Ethnophar-

M60 have related activities, their original compounds (Supple- macol. 125 (2009) 444–449.

mental data Table A2) also might be the active constituents [5] L.Z. Qiu, Z. Hua, L. Liang, J.Z. Hong, W.Y. Fan, X. Ying, C. Xiong, X.H. Xi, C. Kelvin,

Inﬂuence of co-administrated sinomenine on pharmacokinetic fate of paeoni-

of PRR, and they might be Effective forms of their original

ﬂorin in unrestrained conscious rats, J. Ethnopharmacol. 99 (2005) 61–67.

compounds.

[6] M. Hattori, Y.Z. Shu, M. Shimizu, T. Hayashi, N. Morita, K. Kobashi, G. Xu, T.

Namba, Metabolism of paeoniﬂorin and related compounds by human intesti-

nal bacteria, Chem. Pharm. Bull. 33 (1985) 3838–3846.

[7] Y.Z. Shu, M. Hattori, T. Akao, K. Kobashi, K. Kagei, K. Fukuyama, T. Tsukihara, T.

Namba, Metabolism of paeoniﬂorin and related compounds by human intesti-

4. Conclusions nal bacteria. II: structures of 7S- and 7R-paeonimetabolines I and II formed

by Bacteroides fragilis and Lactobacillus brevis, Chem. Pharm. Bull. 35 (1987)

3726–3733.

In this work, a novel strategy was mapped out. In this strategy,

[8] S. Hsiu, Y. Lin, K. Wen, Y. Hou, P.L. Chao, A deglucosylated metabolite of paeoni-

BPCs with a higher signal-to-noise ratio than TICs were used to

ﬂorin of the root of Paeonia lactiﬂora and its pharmacokinetics in rats, Planta

screen the absorbed constituents and metabolites in complex bio- Med. 69 (2003) 1113–1118.

[9] M. Baker, Metabolomics: from small molecules to big ideas, Nat. Methods 8

samples. Additionally, the more accurate ClogP value was selected

(2011) 117–121.

to estimate the hydrophobicity of a compound, other than the

[10] C. Feng, M. Liu, S. Shi, W. Yang, D. Kong, K. Duan, Q. Wang, Pharmacokinetic

LogP value. Further, the HPLC–DAD–ESI-IT-TOF-MS technique was properties of paeoniflorin, albiflorin and oxypaeoniflorin after oral gavage of

extracts of Radix Paeoniae Rubra and Radix Paeoniae Alba in rats, J. Ethnophar-

applied to analyze the absorbed constituents and metabolites in

macol. 130 (2010) 407–413.

biological samples. This strategy was a powerful tool for the sys-

[11] S. Baba, N. Osakabe, M. Natsume, Y. Muto, T. Takizawa, J. Terao, In vivo compar-

tematic screening and identiﬁcation of the absorbed constituents ison of the bioavailability of (+)-catechin, (−)-epicatechin and their mixture in

orally administered rats, J. Nutr. 131 (2001) 2885–2891.

and metabolites of TCMs.

[12] Y. Zhang, F. Xu, J. Dong, J. Liang, Y. Hashi, M. Shang, D. Yang, X. Wang, S. Cai,

This is the ﬁrst time that the absorption and metabolism of PRR

Proﬁling and identiﬁcation of the metabolites of calycosin in rat hepatic super-

decoction has been explored. As a result, 13 new absorbed con- natant incubation system and the metabolites of calycosin-7-O-␤-d-glucoside

in rat urine, J. Pharm. Biomed. Anal. 70 (2012) 425–439.

stituents and 90 new metabolites of PRR decoction were found in

[13] S. Sang, M. Lee, I. Yang, B. Buckley, C.S. Yang, Human urinary metabolite proﬁle

rat urine and plasma samples; in the meantime, 2 known absorbed

of tea polyphenols analyzed by liquid chromatography/electrospray ionization

constituents also were detected. Based on previous reports, 11 of tandem mass spectrometry with data-dependent acquisition, Rapid Commun.

15 absorbed constituents and 11 of 90 metabolites were found to Mass Spectrom. 22 (2008) 1567–1578.

[14] N.A. Shali, C.G. Curtis, G.M. Powell, A.B. Roy, Sulfation of the ﬂavonoids

have various bioactivities, and all 11 absorbed constituents and

quercetin and catechin by rat liver, Xenobiotica 21 (1991) 881–893.

6 metabolites demonstrated bioactivities related to the pharma-

[15] K. Okushio, M. Suzuki, N. Matsumoto, F. Nanjo, Y. Hara, Identiﬁcation of (−)-

cological actions of PRR decoction. These 90 metabolites were epicatechin metabolites and their metabolic fate in the rat, Drug Metab. Dispos.

27 (1999) 309–316.

categorized into ﬁve groups, i.e., 27 (epi)catechin-related metabo-

[16] S. Tourino, J. Perez-Jimenez, M.L. Mateos-Martin, E. Fuguet, M.P. Vinardell, M.

lites, 25 gallic acid-related metabolites, 6 (epi)catechin-related and

Cascante, J.L. Torres, Metabolites in contact with the rat digestive tract after

gallic acid-related metabolites, 25 paeoniﬂorin metabolites, and ingestion of a phenolic-rich dietary ﬁber matrix, J. Agric. Food Chem. 59 (2011)

5955–5963.

7 other metabolites. In this work, 22 new metabolites of paeoni-

[17] A. Fardet, R. Llorach, J. Martin, C. Besson, B. Lyan, E. Pujos-Guillot, A. Scal-

ﬂorin, 10 new metabolites of gallic acid-related compounds, 1

bert, A liquid chromatography–quadrupole time-of-ﬂight (LC–QTOF)-based

new metabolite of (epi)catechin-related compounds, and 2 new metabolomic approach reveals new metabolic effects of catechin in rats fed

high-fat diets, J. Proteome Res. 7 (2008) 2388–2398.

metabolites of other compounds were reported for the ﬁrst time.

[18] G. Susana, G. Ana, S. Celestino, D. Montserrat, Preparation and characterization

As for the 22 new metabolites of paeoniﬂorin, we were only able to

of catechin sulfates, glucuronides, and methylethers with metabolic interest, J,

determine their conjugation types and structural skeletons using Agric. Food Chem. 57 (2009) 1231–1238.

[19] U. Mireia, G. Ignacio, M. Maria, G. Carmen, M. Alexander, A. Cristina, B. Begona,

the latest LC–MS technique. To identify their exact structures, the

Proﬁle of plasma and urine metabolites after the intake of almond [Prunus

isolation and accumulation of these metabolites from rat urine is

dulcis (Mill.) D.A. Webb] polyphenols in humans, J. Agric. Food Chem. 57 (2009)

ongoing work in our lab. 10134–10142.

J. Liang et al. / Journal of Pharmaceutical and Biomedical Analysis 83 (2013) 108–121 121

[20] N.P. Das, L.A. Griffiths, Studies on flavonoid metabolism. Metabolism of [23] Z. Yan, Y. Chen, T. Li, J. Zhang, X. Yang, Identification of metabolites of Si–Ni–San,

(+)-[14C] catechin in the rat and guinea pig, Biochem. J. 115 (1969) a traditional Chinese medicine formula, in rat plasma and urine using liq-

831–836. uid chromatography/diode array detection/triple-quadrupole spectrometry, J.

[21] M. Gonthier, J.L. Donovan, O. Texier, C. Felgines, C. Remesy, A. Scalbert, Chromatogr. B: Anal. Technol. Biomed. Life Sci. 885/886 (2012) 73–82.

Metabolism of dietary procyanidins in rats, Free Radic. Biol. Med. 35 (2003) [24] W. Duan, J. Yang, L. Chen, L. Zhang, Z. Jiang, X. Cai, X. Zhang, F. Qiu, Monoter-

837–844. penes from Paeonia albiﬂora and their inhibitory activity on nitric oxide

[22] N. Vasudevan, A. Mahadevan, Utilization of catechin by micro organisms, Curr. production by lipopolysaccharide-activated microglia, J. Nat. Prod. 72 (2009)

Sci. 59 (1990) 1323–1325. 1579–1584.