Ma

ZJa

b

a

ARR1AA

KTFRMAE

1

atoPoeid2fo1

0d

Journal of Ethnopharmacology 140 (2012) 83– 90

Contents lists available at SciVerse ScienceDirect

Journal of Ethnopharmacology

journa l h o me page: www.elsev ier .com/ locate / je thpharm

etabolomic profiling of the flower bud and rachis of Tussilago farfara withntitussive and expectorant effects on mice

hen-Yu Lia, Hai-Juan Zhia,b, Shui-Yu Xuea,b, Hai-Feng Suna,b, Fu-Sheng Zhanga,in-Ping Jiaa, Jie Xinga, Li-Zeng Zhanga, Xue-Mei Qina,∗

Modern Research Center for Traditional Chinese Medicine of Shanxi University, No. 92, Wucheng Road, Taiyuan 030006, Shanxi, People’s Republic of ChinaCollege of Chemistry and Chemical Engineering of Shanxi University, No. 92, Wucheng Road, Taiyuan 030006, Shanxi, People’s Republic of China

r t i c l e i n f o

rticle history:eceived 30 June 2011eceived in revised form6 December 2011ccepted 16 December 2011vailable online 24 December 2011

eywords:ussilago farfara L.lower budachisetabolomics

ntitussivexpectorant

a b s t r a c t

Ethnopharmacological relevance: Flower bud of Tussilago farfara L. is widely used for the treatment ofcough, bronchitis and asthmatic disorders in the Traditional Chinese Medicine. However, due to theincreasing demands, adulteration with rachis is frequently encountered in the marketplace. No reportdemonstrated the chemical and pharmacological differences between flower bud and rachis before.Materials and methods: The water extracts were orally administrated to mice. Ammonia induced micecoughing model was used to evaluate the antitussive activity. The expectorant activity was evaluatedby volume of phenol red in mice’s tracheas. Metabolites were identified directly from the crude extractsthrough 1D- and 2D-NMR spectra. A metabolic profiling carried out by 1H NMR spectroscopy and multi-variate data analysis was applied to crude extracts from flower bud and rachis.Results: Flower bud significantly lengthened the latent period of cough, decreased cough frequency causedby ammonia and enhanced tracheal phenol red output in expectorant evaluation. Principal componentanalysis (PCA) yielded good separation between flower bud and rachis, and corresponding loading plot

showed that the phenolic compounds, organic acid, sugar, amino acid, terpene and sterol contributed tothe discrimination.Conclusions: These findings provide pharmacological and chemical evidence that only flower bud can beused as the antitussive and expectorant herbal drug. The high concentration of chlorogenic acid, 3,5-dicaffeoylquinic acid, rutin in flower buds may be related with the antitussive and expectorant effects ofFlos Farfara. To guarantee the clinical effect, rachis should be picked out before use.

© 2011 Elsevier Ireland Ltd. All rights reserved.

. Introduction

The flower buds of Tussilago farfara L. (Asteraceae), Flos Farfara,lso locally known as “Kuandonghua” in China, are an impor-ant traditional Chinese medicine, widely used for the treatmentf cough, bronchitis and asthmatic disorders (Committee for theharmacopoeia of People’s Republic of China, 2010). The extractsf Flos Farfara also exhibit various activities, such as antioxidantffect, antimicrobial activities, �-glucosidase inhibitory effect, andnhibitory effects on NO synthesis in LPS-activated macrophage andiacylglycerol acyltransferase activity (Gao et al., 2008; Park et al.,008). Extensive phytochemical studies have revealed that Flos Far-

ara contains a number of diverse components including essentialils (Liu et al., 2006), sesquiterpenes (Ryu et al., 1999; Yaoita et al.,999, 2001), flavonoids (Kim et al., 2006; Liu et al., 2007a; Wu

∗ Corresponding author. Tel.: +86 351 7018379; fax: +86 351 7011202.E-mail address: qinxm@sxu.edu.cn (X.-M. Qin).

378-8741/$ – see front matter © 2011 Elsevier Ireland Ltd. All rights reserved.oi:10.1016/j.jep.2011.12.027

et al., 2010), phenylpropanoids (Liu et al., 2007a; Gao et al., 2008;Wu et al., 2010), chromones (Wu et al., 2008), and pyrrolizidinealkaloids (Jiang et al., 2009).

Flos Farfara has been used in many Traditional Chinese Medicine(TCM) prescriptions, such as Juhongwan, Chuanbeixueligao, Kuan-dongdingchuangtang. According to the theory of TCM, only theflower bud can be used as herbal drug. Due to the increasingdemands of Flos Farfara, adulteration with rachis is frequentlyencountered in the marketplace (Fig. 1). However, to our knowl-edge, there is no report on the pharmacological and chemicaldifferences between flower bud and rachis.

In Chinese pharmacopoeia, the quality of Flos Farfara was eval-uated by the content of tussilagone. However, synergistic effectsof different components are vital for the therapeutic efficacy ofphytomedicine (Williamson, 2001; Liu et al., 2007b). To confine

the analysis to a small range of metabolites may be misleadingfor quality evaluations of TCM. As a result, a non-targeted chemi-cal fingerprinting, which can cover a wide range of metabolites, isnecessary.

84 Z.-Y. Li et al. / Journal of Ethnopharmacology 140 (2012) 83– 90

ara flo

c(fiCNrsciDnobse

brtflmfif

2

2

WtTaqtTiV(eU

2

C

the mice were exposed to a 1000 ml special glass chamber withammonium hydroxide (1 ml) cotton ball after 1 h of the last admin-

Fig. 1. Tussilago farf

The term metabolome has been used to describe the observablehemical profile or fingerprint of the metabolites in whole tissuesKim et al., 2005a). To obtain the most complete metabolomic pro-les, it is necessary to use a wide spectrum of analytical techniques.ompared with the GC–MS or LC–MS based metabolomic studies,MR has some unique advantages, such as rapid, non-selectiveness,

eproducible, and stable (Kooy et al., 2009). In addition, detailedtructural information of metabolites, including chemical shifts andoupling constants, can be directly obtained. This makes NMR andeal choice for the profiling of the broad range of plant metabolites.ifferent multivariate data analysis tools, such as principal compo-ent analysis (PCA), are important tools for the analysis of the databtained by NMR. Recently, NMR based metabolomic studies haveeen used in the quality evaluation of several kinds of herbal drugs,uch as ginseng (Yang et al., 2006), Angelica acutiloba (Tarachiwint al., 2008), Artemisia afra (Liu et al., 2010).

In this study, the antitussive and expectorant activities of flowerud and rachis were evaluated. Then a metabolic profiling was car-ied out, using 1H NMR spectroscopy and multivariate data analysisechniques. Diverse kinds of compounds, including sesquiterpenes,avonoids, phenylpropanoids and fatty acids, were detected, andetabolic differences between flower bud and rachis were identi-

ed. This procedure suggested the possible metabolites responsibleor the antitussive and expectorant properties of Flos Farfara.

. Materials and methods

.1. Plant materials

The flower buds of Tussilago farfara were purchased fromanshengHuangQi Co., Ltd. in Hunyuan, Shanxi, China and authen-

icated by Prof. Xue-Mei Qin from Modern Research Center forraditional Chinese Medicine of Shanxi University. The rachisesdulterated in the flower buds were picked out for the subse-uent study. For metabolic profiling, ten flower bud samples anden rachis samples having different lot numbers were evaluated.hese samples were freeze-dried and grinded to fine powdersn a pestle and mortar, and then stored at −80 ◦C until analysis.oucher specimens coded FB (flower bud) 201001-201010 and RA

rachis) 201001-201010 were deposited in herbarium of the Mod-rn Research Center for Traditional Chinese Medicine of Shanxiniversity.

.2. Solvents and chemicals

Analytical grade chloroform was purchased from Fengchuanhemical Co. Ltd. (Tianjin, China) and analytical grade methanol

wer bud and rachis.

was from Beijing Chemical Works (Beijing, China). Deuterated chlo-roform (CDCl3, 99.8%D) containing tetramethylsilane (TMS, 0.03%,m/v) and methnol-d4 (99.8% D) were obtained from Merck (Darm-stadt, Germany). D2O was bought from Norell (Landisville, USA).Sodium 3-trimethlysilyl [2,2,3,3-d4] propionate (TSP) was fromCambridge Isotope Laboratories Inc. (Andover, MA), and NaOD waspurchased from Armar (Dottingen, Switzerland). Pentoxyverinewas obtained from CSPC Ouyi Pharmaceutical Co., Ltd. (Shiji-azhuang, China). Ammonium chloride was from Beijing ChemicalWorks (Beijing, China) and phenol red from Tianjin Chemical Works(Tianjin, China).

2.3. Samples for antitussive and expectorant analysis

According to the traditional use, the flower buds and rachis wereextracted with water by refluxing for three times (2 h each). Thecombined solution was filtered and concentrated under reducedpressure to afford the water extract.

2.4. Animals

ICR mice of either sex (19–24 g) were purchased from Bei-jing Vital Laboratory Animal Technology Company (license numberSCKX-2006-0008). All animals were housed at room temperature(20–25 ◦C) and constant humidity (40–70%) under a 12 h light–darkcycle in SPF (Specific Pathogen Free) grade laboratory. The animalstudy was performed according to the international rules consid-ering animal experiments and the internationally accepted ethicalprinciples for laboratory animal use and care.

2.5. Antitussive effects against ammonia induced coughing

The antitussive activity was performed as previously describedwith minor modifications (Xu et al., 2005). After 3 days of adap-tation, 60 mice were divided into 4 groups (n = 15) randomly andorally administered, including control group (distilled water), pos-itive group (pentoxyverine 50 mg/kg), flower bud group (2.8 g/kg),rachis group (3.5 g/kg). The administration dose of extract wascalculated according to the clinical dose of this plant and yieldof extract. The administration had been lasted for 5 days and

istration and the cough incubation period was recorded. After1 min, the mice were taken out from the chamber and placed in abeaker and the frequency of cough within 2 min was observed andrecorded.

ophar

2

gcrpSft0cfciwwt

2

2

aTucK01N

2

cofcaiuCm6oa1

2

(ea(istdtttpfTr

Z.-Y. Li et al. / Journal of Ethn

.6. Expectorant test

After 1 week of adaptation, 48 ICR mice were divided into 4roups (12/group) randomly and orally administered, includingontrol group (distilled water), positive group (ammonium chlo-ide 1 g/kg), flower bud group (2.8 g/kg), rachis group (3.5 g/kg). Therocedures were performed as described previously (Engler andzelenyi, 1984). All groups were treated with a single dose dailyor 5 days and the last dose were given 30 min before intraperi-oneal injection of phenol red solution (5% in saline solution, w/v,.1 ml/10 g body weight). Mice were sacrificed by cervical dislo-ation 45 min after application of phenol red. After dissected freerom adjacent organs, the trachea was removed from the thyroidartilage to the main stem bronchi and put into 2 ml normal salinemmediately. After ultrasonic for 15 min, 2 ml NaHCO3 solution (5%,

/v) was added to the saline and optical density of the mixtureere measured at 558 nm using UV-7501 UV–vis spectrophotome-

er (Wuxi Keda Instrument Co., Ltd., China).

.7. Extraction of plant samples for metabolomics

.7.1. Extracted with hot waterPlant material (10 g) was transferred into round-bottomed flask

nd extracted with water by refluxing for three times (2 h each).he combined solution was filtered and dried with a rotary vac-um evaporator. Then 50 mg dried extract was weighed into 2 mlentrifuge tube and added 800 �l mixture (1:1) of CD3OD andH2PO4 buffer in D2O (adjusted to pH 6.0 by 1 N NaOD) containing.05% TSP and sonicated for 20 min. After centrifuging for 15 min at3,000 rpm, the supernatant (600 �l) was transferred into a 5 mmMR tube for NMR analysis.

.7.2. Extracted with two phase systemGround material (200 mg) was transferred into 10-ml glass

entrifuge tube. 3 ml of 50% water–methanol mixture and 3 mlf chloroform were added to the tube followed by vortexingor 1 min and ultrasonicating for 20 min. The material was thenentrifuged at 3500 rpm for 25 min. The chloroform (lower) andqueous methanol (upper) fractions were transferred separatelynto a 25 ml round-bottomed flask and dried with a rotary vac-um evaporator. Chloroform fractions were dissolved in 800 �lDCl3, and aqueous methanol fractions were dissolved in 800 �lixture (1:1) of CD3OD and KH2PO4 buffer in D2O (adjusted to pH

.0 by 1 N NaOD) containing 0.05% TSP. The supernatants (600 �l)f all the samples were transferred into 5-mm NMR tube for NMRnalysis after 10 min of sonication and centrifuging for 15 min at3,000 rpm.

.8. NMR measurements

1H NMR, 2D J-resolved and 1H–1H-correlated spectroscopyCOSY), heteronuclear multiple bonds coherence (HMBC), het-ronuclear single quantum coherence (HSQC) were recordedt 25 ◦C on a Bruker 600-MHz AVANCE III NMR spectrometer600.13 MHz proton frequency). CD3OD and CDCl3 were used fornternal lock purposes. Each 1H NMR spectrum consisted of 64cans requiring 5 min acquisition time with the following parame-ers: 0.18 Hz/point, pulse width (PW) = 30◦ (12.7 �s), and relaxationelay (RD) = 5.0 s. A presaturation sequence was used to suppresshe residual H2O signal with low power selective irradiation athe H2O frequency during the recycle delay. FIDs were Fourierransformed with LB = 0.3 Hz. The resulting spectra were manually

hased and baseline-corrected, and calibrated to TSP at 0.00 ppmor water fractions and TMS at 0.00 ppm for organic fractions.wo-dimensional J-resolved NMR spectra were acquired with 2.0 selaxation delay using 16 scans and spectral widths of 10,000 Hz

macology 140 (2012) 83– 90 85

in F2 and 78 Hz in F1. J-resolved spectra were tilted by 45◦ andsymmetrized about F1. The COSY spectra were acquired with 2.0 srelaxation delay, 5411 Hz spectra width in both dimensions. Thewindow function for COSY spectra was sine-bell (SSB = 0). TheHMBC spectra were obtained with 1.5 s relaxation delay, using8012 Hz spectral width in F2 and 133,200 Hz in F1. For HSQC spec-tra, the 1.5 s relaxation delay was used, 6009 Hz spectral width inF2 and 108,000 Hz in F1. All 2D spectra were calibrated at 0.00 ppmto TSP or TMS.

2.9. Data analysis

The 1H NMR spectra were processed using MestReNova (ver-sion 5.2.5, Mestrelab Research, Santiago de Compostella, Spain). Foraqueous methanol fraction and aqueous extract, spectral intensi-ties were scaled to TSP and reduced to integrated regions of equalwidth (0.04 ppm) corresponding to the region of � 0.50–10.02. Theregions of � 4.70–5.02 and � 3.30–3.38 were excluded from theanalysis because of the residual signal of HDO and CD3OD, respec-tively. For the chloroform fraction, spectral intensities were scaledto TMS and reduced to integrated regions 0.04 ppm correspond-ing to the region of � 10.02–0.50. The region between � 7.22 and �7.30 was removed from the analysis because of the residual signalof CHCl3. Principal component analysis (PCA) with pareto scalingwas performed with SIMCA-P11.0 (Umetrics, umeå, Sweden). Rela-tive amount of metabolites was evaluated based on the integratedregions (bucket) of the NMR spectra using Origin (8.0, OriginLab,USA).

The antitussive and expectorant results are expressed asmean ± standard error of mean (S.E.M.). Significance was evaluatedusing the Student’s t-test (Woodson, 1987). Values of p < 0.05 implysignificance of the pharmacological effects in the experiments.

3. Results and discussion

3.1. Antitussive and expectorant effects

In this experiment, we used chemical stimulus (ammonium)method due to its simple procedure that omitted of anestheti-zation and usually used in new drug development for traditionalChinese medicines. Therefore, the antitussive activity of the flowerbud and rachis extracts were demonstrated in vivo experiment byprolonging the incubation time, and reducing coughing times in2 min. The longer cough incubation period showed stronger effectof the drug on relieving cough and the less cough times exhib-ited its stronger antitussive effect. The antitussive effects of flowerbuds and rachis extracts on the ammonia liquor induced cough inmice were shown in Fig. 2. Positive and flower bud groups sig-nificantly prolonged latent period of cough compared with thecontrol group. The data of cough frequency in 2 min were shownin Fig. 2B. It was observed that the cough frequency of flower budgroup could be significantly reduced, comparable with that of theeffect produced by pentoxyverine, a prototype antitussive agent.However, rachis group had no significant effect on the antitussiveactivity.

In the expectorant experiment, phenolsulphonphthalein wasapplied and discharged with sputum some minutes after i.p. injec-tions into mice. Excretion of phenolsulphonphthalein would takeplace of the excretion of sputum because the agent was elimi-nated partly from trachea. The stronger eliminating phlegm activity

and expectorant effect showed greater UV absorption. As shown inFig. 3, ammonium chloride and flower bud groups could enhancetracheal red output, indicating the strong expectorant effect. Forthe rachis, no obvious effect was observed.

86 Z.-Y. Li et al. / Journal of Ethnopharmacology 140 (2012) 83– 90

F ed co1 chis; 4d

3w

jstd0TrtrigwmdwertJ(

Fewm

ig. 2. Antitussive effect of flower bud and rachis on the ammonia hydroxide-induc: Control mice without any treatment; 2: the mice were given pentoxyverine; 3: raifferent from control.

.2. Chemical differences between flower bud and rachis in hotater extracts

All the hot water extracts of flower bud and rachis were sub-ected to 1H NMR analysis. The assignment of the 1H spectrum,ee Fig. 4, was based on careful analysis of 1D and 2D NMR spec-ra, and further confirmed by comparing with the reported NMRata in literature. In the 1H NMR spectrum, the area between �.50 and � 4.00 corresponds to amino acids and organic acids.he region of � 4.00–� 5.50 is considered to be the carbohydrateegion and the remaining part, i.e. � 5.50–� 10.02, is known ashe aromatic region. The high signal intensities in the amino acidegion were helpful to elucidate a number of amino acid signals,ncluding isoleucine, leucine, valine, threonine, alanine, proline,lutamic acid, and aspartic acid. In addition, a number of signalsere assigned to the organic acids, such as acetic acid, succinic acid,alic acid, citric acid. Other compounds, including 1-O-ethyl-�-

-glucoside, ethanolamine, choline, phosphatidylcholine, creatine,ere also identified (Kim et al., 2010a; Liu et al., 2010; Lopez-Gresa

t al., 2010; Lubbe et al., 2011). The signals in the carbohydrateegions were highly clustered and overlapped. This region showed

he signals of the anomeric protons of �-glucose at � 4.60 (d,

= 7.9 Hz), �-glucose at � 5.20 (d, J = 3.7 Hz), and sucrose at � 5.42d, J = 3.8 Hz) and 4.18 (d, J = 8.7 Hz) (Kim et al., 2010a; Ali et al.,

ig. 3. Effect of flower bud and rachis extracts on the phenolsulphonphthaleinxcretion in mice (n = 12). 1: Control mice without any treatment; 2: the miceere given ammonium chloride; 3: rachis; 4: flower bud. Values are represented asean ± SE. *p < 0.05 significantly different from control.

ughing incubation periods (A) and coughing times within 2 min (B) in mice (n = 15).: flower bud. Values are represented as mean ± SE. *p < 0.05, **p < 0.01 significantly

2011). The aromatic part of the 1H NMR spectra showed a seriesof signals of phenylpropanoids, including caffeic acid, chlorogenicacid, sinapic acid, 4,5-dicaffeoylquinic acid, 3,5-dicaffeoylquinicacid, and 3,4-dicaffeoylquinic acid (S. Table 1). The characteristicdoublets of 16.0 Hz in the range of � 6.39–� 6.50 and � 7.59–� 7.70(H-8′ and H-7′) and double triplets at � 5.43, � 5.49 and � 5.67 werealso evidence of these compounds (Verpoorte et al., 2007; Kim et al.,2010b; Liu et al., 2010). In the aromatic region, the flavonoids rutinand kampferol analogues were evident (Fig. 4). Maleic acid, fumaricacid, formic acid, p-hydroxybenzoic acid, adenine and adenosinewere also identified in the aromatic region (Liu et al., 2010; Lopez-Gresa et al., 2010) (S. Table 1; S. Fig. 1).

Visual inspection of the NMR spectra indicated that flower budhad more phenylproponoids (S. Fig. 2). However, it was not possibleto get a consistent idea about the differences in the metabolic pro-file with simple visual inspection due to the intra-group variations.Therefore, multivariate data analysis tools were used to providecomparative interpretations and visualization of chemical differ-ences between the flower bud and rachis.

All the obtained NMR data of flower bud and rachis were ana-lyzed by principal component analysis (PCA). In the score plot, aclear separation was found between the flower bud and rachis,with the first two components explaining 71.6% of total variation(Fig. 5). The corresponding loading plot revealed that the rachiswas characterized by high signals for the compounds of succinicacid, malic acid, aspartic acid, sucrose. The flower bud showedhigher aromatic compounds such as caffeic acid, chlorogenic acid,3,5-dicaffeoylquinic acid and rutin. Some amino acids, includingthreonine, alanine, proline, and glutamic acid, as well as glucose,choline, phosphatidylcholine, acetic acid, formic acid, were alsofound to be higher in flower bud. Some unidentified signals werealso found to largely contribute to the discrimination betweenflower bud and rachis. These signals mostly belong to sugars andare difficult to identify.

To confirm their biased distribution, the relative amounts (NMRsignal buckets) are shown in Fig. 6. Independent student’s t-testindicated that chlorogenic acid, 3,5-dicaffeoylquinic acid, rutinwere significantly higher in the flower buds. However, the rachiscontained more malic acid, sucrose.

3.3. Metabolomic profiling using two-phase extraction method

Extraction with hot water is the traditional use in TCM (Hanet al., 2010; Li et al., 2011). However, only the abundant polar

metabolites, such as amino acids, organic acids, can be detectedby NMR in the hot water extracts. Non-polar metabolites suchas terpenoids are also believed to play an important role in theherbal materials. To evaluate the overall chemical differences

Z.-Y. Li et al. / Journal of Ethnopharmacology 140 (2012) 83– 90 87

Fig. 4. 600 MHz 1H NMR spectra of the hot water extracts from Tussilago farfara. 1: isoleucine, 2: leucine, 3: valine, 4: 1-O-ethyl-�-d-glucoside, 5: threonine, 6: alanine,7: acetic acid, 8: proline, 9: glutamic acid, 10: succinic acid, 11: malic acid, 12: citric acid, 13: aspartic acid, 14: ethanolamine, 15: choline, 16: phosphatidyl-choline, 17:creatine, 18: �-glucose, 19: �-glucose, 20: sucrose, 21: maleic acid, 22: fumaric acid, 23: 4,5-dicaffeoylquinic acid, 24: caffeic acid, 25: chlorogenic acid, 26: sinapic acid, 27:3,5-dicaffeoylquinic acid, 28: 3,4-dicaffeoylquinic acid, 29: rutin, 30: kampferol analogues, 31: p-hydroxy benzoic acid, 32: adenine, 33: adenosine, 34: formic acid.

Fig. 5. Score (PC 1 vs. PC 2, A) and loading (PC 1, B) plot of PCA results obtained from1H NMR spectra of the hot water extracts: flower bud (�) and rachis ( ).

between flower bud and rachis, a different extraction method,the biphase extraction, CHCl3–MeOH–H2O (2:1:1), was used. Afterextraction, two fractions, chloroform and aqueous methanol parts,were obtained and subjected to 1H NMR analysis and multivariateanalysis, respectively (S. Fig. 3). The non-polar metabolites wereconcentrated in the chloroform fraction.

For the aqueous methanol fractions, the 1H NMR spectra weresimilar with those of hot water extracts (S. Fig. 4; S. Fig. 5),except that some minor differences in the aromatic and amino acidregion. The differential metabolites between flower bud and rachis,revealed by the corresponding loading plot (S. Fig. 6), were also sim-ilar with those of hot water extracts (Fig. 5), except that two aminoacids, isoleucine, leucine, were found to be higher in rachis.

For the chloroform fractions, six metabolites (Fig. 7) have beenidentified, including two sesquiterpenes, tussilagone and 7�-(3-Ethyl-ciscrotonoyloxy)-1�-(2-methylbutyryloxy)-3(14)-dehydro-Z-notonipetranone (EMDNT), a triterpene, bauer-7-ene-3�,16�-diol, two sterols, �-sitosterol and sitosterone, and fattyacids analogues. � 0.78 (d, J = 6.6 Hz, CH3-13), � 0.98 (d, J = 6.6 Hz,CH3-12), � 1.08 (t, J = 7.3 Hz, CH3-5′), � 1.22 (d, J = 6.7 Hz, CH3-15), �2.10 (s, OAC), � 2.15 (s, CH3-6′), � 4.78 (s, H-10a), � 5.14 (s, H-10b),� 5.57 (br t, J = 2.3 Hz, H-7), and � 5.63 (s, H-2′) were assigned tobe tussilagone, a major sesquiterpene in Flos Farfara (Park et al.,2008). EMDNT, another structurally similar sesquiterpene, was at� 0.88 (t, J = 7.4 Hz, CH3-4′′), � 0.90 (d, J = 7.0 Hz, CH3-13), � 0.97 (d,J = 6.4 Hz, CH3-12), � 1.07 (t, J = 7.5 Hz, CH3-5′), � 1.12 (d, J = 7.1 Hz,CH3-5′′), � 2.15 (s, CH3-6′), � 2.16 (d, J = 7.2 Hz, CH3-15), � 4.81(s, H-10a), � 5.17 (s, H-10b), � 5.51 (d, J = 3.4 Hz, H-7), � 5.53 (d,

J = 4.0 Hz, H-1), � 5.61 (s, H-2′), � 6.38 (q, J = 6.4 Hz, H-14) (Parket al., 2008). In addition, resonances at � 0.73 (s, CH3-25) and� 1.15 (d, J = 6.8 Hz, CH3-29) were identified as bauer-7-ene-3�,16�-diol, a triterpene in the plant (Santer and Stevenson, 1962;

88 Z.-Y. Li et al. / Journal of Ethnopharmacology 140 (2012) 83– 90

Fig. 6. Student t-test for the relative amount of compounds based on mean peak areas of associated signals. ( ): chemical shifts of signals used for the student t-test. 25:chlorogenic acid (6.36), 27: 3,5-dicaffeoylquinic (6.47), 29: rutin (6.98), 11: malic acid (2.47), 20: sucrose (5.42).

Fn

ig. 7. 600 MHz 1H NMR spectra of the chloroform fractions of Tussilago farfarotonipetranone, 36: tussilagone, 37: bauer-7-ene-3�, 16�-diol, 38: �-sitosterol, 39: sito

a. 35: 7�-(3-Ethyl-ciscrotonoyloxy)-1�-(2-methylbutyryloxy)-3(14)-dehydro-Z-sterone, 40: fatty acid.

Z.-Y. Li et al. / Journal of Ethnophar

F1

Y2apeN

c(li(

4

rmdsi

stSdwlas

ig. 8. Score (PC 1 vs. PC 2, A) and loading (PC 1, B) plot of PCA results obtained fromH NMR spectra of the chloroform fractions: flower bud ( ) and rachis (�).

aoita and Kikuchi, 1998). The terminal methyl (� 0.90), �-CH2 (�.3), �-CH2 (� 1.6), allylic CH2 (� 2.05), and bis-allylic CH2 (� 2.77),ll the other protons of hydrocarbon chain (� 1.2–1.3), and olefinicrotons (� 5.35) were detected as fatty acids analogues (Chatterjeet al., 2010). The metabolites were further confirmed by the 2DMR spectra and purified standards.

PCA analysis of the chloroform fraction indicated that the rachisould also be clearly separated from the flower bud along PC1Fig. 8A). The loading plot suggests that in the flower bud, tussi-agone, EMDNT, bauer-7-ene-3�, 16�-diol, are predominant; whilen rachis, �-sitosterol, sitosterone, fatty acids are predominantFig. 8B).

. Conclusions

For thousands of years, the flower bud was decocted for treatingespiratory diseases. However, the adulteration with rachis is com-only encountered in the drug market for a long time. This study

emonstrated that the rachis extract showed no obvious antitus-ive and expectorant activities, implying that it should not be usednstead of flower bud.

Corresponding metabolomic approach based on 1H NMRpectroscopy coupled with multivariate data analysis revealedhe chemical differences between flower bud and rachis.ome secondary metabolites, including chlorogenic acid, 3,5-icaffeoylquinic acid, rutin were highly accumulated in flower bud,

hile the rachis showed higher concentration of primary metabo-

ites, like malic acid, sucrose, fatty acids analogues. Chlorogeniccid, which is the active compound in many medicinal plants, hasignificant antibacterial and antiviral activities (Peluso et al., 1995;

macology 140 (2012) 83– 90 89

Zhang et al., 2001). Rutin, one of the major flavonoids found ina variety of plants, is also the active compound for many medic-inal plants. Rutin has significant antioxidant, radic scavenging(Song et al., 2010), anti-inflammatory (Guardia et al., 2001; Kimet al., 2005b), anti-viral, and anti-bacterial (Pereira et al., 2008;Orhan et al., 2010) activities. No direct antitussive and expecto-rant effects of these compounds have been reported, however,synergistic effect may be existed for relieving the syndrome ofcough and asthma. The high concentration of chlorogenic acid,3,5-dicaffeoylquinic acid, rutin, tussilagone in flower buds may berelated with the antitussive and expectorant effects of Flos Farfara.

The overall chemical differences between flower bud and rachisalso revealed that the flower bud contained more tussilagone,EMDNT, bauer-7-ene-3�, 16�-diol. Tussilagone has been deter-mined as the main active compound of Flos Farfara, and was usedas marker compound for the quality evaluation of Flos Farfarain the Chinese Pharmacopoeia. Previous pharmacological studieshave revealed that tussilagone exerts anti-inflammatory activitiesin murine macrophages by inducing heme oxygenase-1 (HO-1)expression (Hwangbo et al., 2009). The relationship of non-polarmetabolites and the antitussive and expectorant activities remainsunknown. And further studies should be conducted on systematicresearch for the active components of Flos Farfara for the antitus-sive and expectorant effects.

Previous studies showed that Tussilago farfara contained thetoxic pyrrolizidine alkaloids (PAs), mainly senkirkine and traces ofsenecionine (Pu et al., 2004). The PAs can be hepatoxic, causingdamage to the liver and may even cause liver cancer when minutequantities are consumed (Lin et al., 2007). However, no signals ofPAs were found in the 1H NMR spectra of hot water extracts, aque-ous methanol fractions or chloroform fractions (S. Fig. 7; S. Fig. 8),due to the low amount of PAs and also the low sensitivity of NMRcompared with MS. Further study should be conducted to evaluatethe PAs levels between flower bud and rachis by UPLC–MS.

This is the first time that a metabolic profiling analysis, usingNMR and multivariate data analysis, was applied to Flos Farfara.The information obtained in this work is valuable for further phy-tochemical and activity research, and may be used to establish theanalytical requirements for quality control purpose in the future.

Acknowledgments

This study was financially supported by National Natural Sci-ence Foundation of China (No. 30900118), Program of InternationalScience and Technology Cooperation of Shanxi Province (No.2010081070).

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.jep.2011.12.027.

References

Ali, K., Maltese, F., Toepfer, R., Choi, Y.H., Verpoorte, R., 2011. Metabolic characteriza-tion of Palatinate German white wines according to sensory attributes, varieties,and vintages using NMR spectroscopy and multivariate data analyses. Journalof Biomolecular NMR 49, 255–266.

Chatterjee, S., Srivastava, S., Khalid, A., Singh, N., Sangwan, R.S., Sidhu, O.P., Roy,R., Khetrapal, C.L., Tuli, R., 2010. Comprehensive metabolic fingerprinting ofWithania somnifera leaf and root extracts. Phytochemistry 71, 1085–1094.

Committee for the Pharmacopoeia of People’s Republic of China, 2010. Pharma-

copoeia of People’s Republic of China, Part 1. Chemical Industry PublishingHouse, Beijing, China, 312 pp.

Engler, H., Szelenyi, I., 1984. Tracheal phenol red secretion, a new method forscreening mucosecretolytic compounds. Journal of Pharmacological Methods11, 151–157.

9 ophar

G

G

H

H

J

K

K

K

K

K

K

L

L

L

L

L

L

L

0 Z.-Y. Li et al. / Journal of Ethn

ao, H., Huang, Y.N., Gao, B., Xu, P.Y., Inagaki, C., Kawabata, J., 2008. �-Glucosidaseinhibitory effect by the flower buds of Tussilago farfara L. Food Chemistry 106,1195–1201.

uardia, T., Rotelli, A.E., Juarez, A.O., Pelzer, L.E., 2001. Anti-inflammatory proper-ties of plant flavonoids. Effects of rutin, quercetin and hesperidin on adjuvantarthritis in rat. Il Farmaco 56, 683–687.

an, N., Chang, C.L., Wang, Y.C., Huang, T., Liu, Z.H., Yin, J., 2010. The in vivo expec-torant and antitussive activity of extract and fractions from Reineckia carnea.Journal of Ethnopharmacology 131, 220–223.

wangbo, C., Lee, H.S., Park, J., Choe, J., Lee, J.H., 2009. The anti-inflammatory effectof tussilagone, from Tussilago farfara, is mediated by the induction of hemeoxygenase-1 in murine macrophages. International Immunopharmacology 9,1578–1584.

iang, Z.J., Liu, F., Goh, J.J.L., Yu, L.J., Li, S.F.Y., Ong, E.S., Ong, C.N., 2009. Determina-tion of senkirkine and senecionine in Tussilago farfara using microwave-assistedextraction and pressurized hot water extraction with liquid chromatographytandem mass spectrometry. Talanta 79, 539–546.

im, H.K., Choi, Y.H., Erkelens, C., Lefeber, A.W.M., Verpoorte, R., 2005a. Metabolicfingerprinting of Ephedra species using 1H NMR spectroscopy and principal com-ponent analysis. Chemical and Pharmaceutical Bulletin 53, 105–109.

im, H.K., Choi, Y.H., Verpoorte, R., 2010a. NMR-based metabolomic analysis ofplants. Nature Protocols 5, 536–549.

im, H.K., Saifullah, Khan, S., Wilson, E.G., Kricun, S.D.P., Meissner, A., Goraler,S., Deelder, A.M., Choi, Y.H., Verpoorte, R., 2010b. Metabolic classification ofSouth American IIex species by NMR-based metabolimics. Phytochemistry 71,773–784.

im, K.H., Lee, K.W., Kim, D.Y., Park, H.H., Kwon, I.B., Lee, H.J., 2005b. Optimal recov-ery of high-purity rutin crystals from the whole plant of Fagopyrum esculentumMoench (buckwheat) by extraction, fractionation, and recrystallization. Biore-source Technology 96, 1709–1712.

im, M.R., Lee, J.Y., Lee, H.H., Aryal, D.K., Kim, Y.G., Kim, S.K., Woo, E.R., Kang, K.W.,2006. Antioxidative effects of quercetin-glycosides isolated from the flower budsof Tussilago farfara L. Food and Chemical Toxicology 44, 1299–1307.

ooy, F.V.D., Maltese, F., Choi, Y.H., Kim, H.K., Verpoorte, R., 2009. Quality control ofherbal material and phytopharmaceuticals with MS and NMR based metabolicfingerprinting. Planta Medica 75, 763–775.

i, D.Y., He, Y.G., Liu, L.P., Cui, X., 2011. Anti-fatigue effects and changes of seru-mal ANP levels of Radix Astragali in mice. Journal of Jilin Institute of PhysicalEducation 27, 77–78.

in, G., Tang, J., Liu, X.Q., Jiang, Y., Zheng, J., 2007. Deacetylclivorine: a gender-selective metabolite of clivorine formed in female Sprague-Dawley rat livermicrosomes. The American Society for Pharmacology and Experimental Thera-peutics 35, 607–613.

iu, N.Q., Cao, M., Frederich, M., Choi, Y.H., Verpoorte, R., Kooy, F.V.D., 2010.Metabolomic investigation of the ethnopharmacological use of Artemisia afrawith NMR spectroscopy and multivariate data analysis. Journal of Ethnophar-macology 128, 230–235.

iu, Y.F., Yang, X.W., Wu, B., 2006. GC-M S analysis of essential oil constituents frombuds of Tussilago farfara L. Journal of Chinese Pharmaceutical Science 15, 10–14.

iu, Y.F., Yang, X.W., Wu, B., 2007a. Studies on chemical constituents in the buds ofTussilago farfara. China Journal of Chinese Materia Medica 32, 2378–2381.

iu, Y.H., Kong, J.M., Chia, L.S., Goh, N.K., 2007b. Synergistic effect of traditionalChinese medicine. Asian Journal of Chemistry 19, 867–882.

opez-Gresa, M.P., Maltese, F., Belles, J.M., Conejero, V., Kim, H.K., Choi, Y.H.,Verpoorte, R., 2010. Metabolic response of tomato leaves upon differentplant–pathogen interactions. Phytochemical Analysis 21, 89–94.

macology 140 (2012) 83– 90

Lubbe, A., Choi, Y.H., Vreeburg, P., Verpoorte, R., 2011. Effect of fertilizers on galan-thamine and metabolite profiles in narcissus bulbs by 1H NMR. Journal ofAgricultural and Food Chemistry 59, 3155–3161.

Orhan, D.D., Ozcelik, B., Ozgen, S., Ergun, F., 2010. Antibacterial, antifungal, andantiviral activities of some flavonoids. Research in Microbiology 165, 496–504.

Park, H.R., Yoo, M.Y., Seo, J.H., Kim, I.S., Kim, N.Y., Kang, J.Y., Cui, L., Lee, C.S., Lee,C.H., Lee, H.S., 2008. Sesquiterpenoids isolated from the flower buds of Tussilagofarfara L. inhibit diacylglycerol acyltransferase. Journal of Agricultural and FoodChemistry 56, 10493–10497.

Peluso, G., Feo, V.D., Simone, F.D., Bresciano, E., Vuotto, M.L., 1995. Studies on theinhibitory effects of caffeoylquinic acids on monocyte migration and superoxideion production. Journal of Natural Products 58, 639–646.

Pereira, A.C., Oliveira, D.F., Silva, G.H., Figueiredo, H.C.P., Cavalheiro, A.J., Carvalho,D.A., Souza, L.P., Chalfoun, S.M., 2008. Identification of the antimicrobial sub-stances produced by Solanum palinacanthum (Solanaceae). Anais da AcademiaBrasileira de Ciências 80, 427–432.

Pu, S.B., Xu, D.R., Zhang, M., Zhou, H.H., Wang, Z.T., Yu, G.D., 2004. The detection ofhepatotoxic pyrrolizidine alkaloids in Flos Farfarae by LC/MSn . Chinese Journalof Natural Medicine 2, 293–297.

Ryu, J.H., Jeong, Y.S., Sohn, D.H., 1999. A new bisabolene epoxide from Tussilagofarfara, and inhibition of nitric oxide synthesis in LPS-activated macrophages.Journal of Natural Products 62, 1437–1438.

Santer, J.O., Stevenson, R., 1962. Arnidiol and faradiol. The Journal of Organic Chem-istry 27, 3204–3208.

Song, F.L., Gan, R.Y., Zhang, Y., Xiao, Q., Kuang, L., Li, H.B., 2010. Total phenoliccontents and antioxidant capacities of selected Chinese medicinal plants. Inter-national Journal of Molecular Sciences 11, 2362–2372.

Tarachiwin, L., Katoh, A., Ute, K., Fukusaki, E., 2008. Quality evaluation of Angelicaacutiloba Kitagawa roots by 1H NMR-based metabolic fingerprinting. Journal ofPharmaceutical and Biomedical Analysis 48, 42–48.

Verpoorte, R., Choi, Y.H., Kim, H.K., 2007. NMR-based metabolomics artwork in phy-tochemistry. Phytochemistry Reviews 6, 3–14.

Williamson, E.M., 2001. Synergy and other interactions in phytomedicine. Phy-tomedicine 8, 401–409.

Woodson, R.F., 1987. Statistical methods for the analysis of biomedical data. WileySeries in Probability and Mathematical Statistics, Wiley, New York, 315–316.

Wu, D., Zhang, M., Zhang, C.F., Wang, Z.T., 2008. Chromones from the flower buds ofTussilago farfara. Biochemical Systematics and Ecology 36, 219–222.

Wu, D., Zhang, M., Zhang, C.F., Wang, Z.T., 2010. Flavonoids and phenolic acid deriva-tives from Flos Farfarae. China Journal of Chinese Materia Medica 35, 1142–1144.

Xu, S.Y., Bian, R.L., Chen, X., 2005. Pharmacology Experimental Method, third ed.People’s Medical Publishing House, Beijing, China.

Yang, S.Y., Kim, H.K., Lefeber, A.W.M., Erkelens, C., Angelova, N., Choi, Y.H., Ver-poorte, R., 2006. Application of two-dimensional nuclear magnetic resonancespectroscopy to quality control of Ginseng commercial products. Planta Medica72, 364–369.

Yaoita, Y., Kamazawa, H., Kikuchi, M., 1999. Structures of new oplopane-typesesquiterpenoids from the flower buds of Tussilago farfara L. Chemical and Phar-maceutical Bulletin 47, 705–707.

Yaoita, Y., Kikuchi, M., 1998. Triterpenoids from flower buds of Tussilago farfara L.Journal of Natural Medicines 52, 273–275.

Yaoita, Y., Suzuki, N., Kikuchi, M., 2001. Structures of new sesquiterpenoids fromFarfarae Flos. Chemical and Pharmaceutical Bulletin 49, 645–648.

Zhang, A.L., Ma, Q., Gao, J.M., Zhang, K.J., Wang, L., 2001. Studies on bioactivities ofchlorogenic acid and its analogues. Chinese Traditional and Herbal Drugs 32,173–176.