Identification of antioxidants from Taraxacum mongolicum by high-performance liquid chromatography–diode array detection–radical-scavenging detection–electrospray ionization mass spectrometry and nuclear magnetic resonance experiments

Journal of Chromatography A, 1209 (2008) 145–152

Contents lists available at ScienceDirect

Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma

Identification of antioxidants from Taraxacum mongolicum by high-performance liquid chromatography–diode array detection–radical-scavenging detection–electrospray ionization mass spectrometry and nuclear magnetic resonance experiments Shuyun Shi a,c,∗ , Yu Zhao b , Honghao Zhou c , Yuping Zhang a , Xinyu Jiang a , Kelong Huang a a

School of Chemistry and Chemical Engineering, Central South University, Changsha 410083, China College of Pharmaceutical Sciences, Zhejiang University, Hangzhou 310031, China c Pharmacogenetics Institute of Central South University, Changsha 410078, China b

a r t i c l e

i n f o

Article history: Received 19 June 2008 Received in revised form 6 August 2008 Accepted 2 September 2008 Available online 5 September 2008 Keywords: Taraxacum mongolicum Phenolic secondary metabolites Antioxidant activity HPLC–DAD–RSD–ESI-MS NMR

a b s t r a c t Taraxacum mongolicum was a traditional Chinese medicine for the treatment of inflammatory disorders and viral infectious diseases. Furthermore, fresh leaves of T. mongolicum have been used by local people as vegetable food in Northern China. An on-line rapid screening method, high-performance liquid chromatography–diode array detection–radical-scavenging detection–electrospray ionization mass spectrometry (HPLC–DAD–RSD–ESI-MS) system, has been developed for the separation and identification of radical scavengers from the methanolic extract of T. mongolicum. In addition, the detected antioxidants were isolated directly by preparative HPLC (PHPLC) and Sephadex LH-20, then the purified compounds were sampled to off-line nuclear magnetic resonance (NMR) spectrometer to acquire NMR spectra. The structure of the active compounds was elucidated by ultraviolet (UV), ESI-MS and NMR spectral data. Thirty-two radical-scavenging compounds including sixteen flavonoid derivatives, ten phenylpropanoid derivatives and six benzoic acid derivatives were screened, isolated and identified. Among them, seventeen compounds including three new compounds were first isolated from Taraxacum genus by our group. Caffeic acid (6), isoetin-7-O-␤-d-glucopyranosyl-2 -O-␣-l-arabinopyranoside (9), isoetin-7-O-␤-d-glucopyranosyl-2 -O-␣-d-glucopyranoside (10) and isoetin-7-O-␤-d-glucopyranosyl-2 O-␤-d-xyloypyranoside (12) were found to be the major metabolites in T. mongolicum based on their relative peaks in the HPLC chromatogram. Antioxidant activity of three new compounds was assessed for their scavenging capacity on 2,2-diphenyl-1-picrylhydrazyl (DPPH) radical, and all of them showed potent activity. © 2008 Elsevier B.V. All rights reserved.

1. Introduction The genus Taraxacum is a member of the family Asteraceae, which is widely distributed in the warmer temperature zones of the northern hemisphere, inhabiting fields, roadsides and ruderal sites possessing many biological activities. Taraxacum species have attracted the attention of researchers because of their antioxidant potential besides anti-inflammatory, anticarcinogenic, anti-allergic, anti-hyperglycemic, anti-coagulatory and analgesic activities [1]. In China, the genus Taraxacum includes

∗ Corresponding author at: School of Chemistry and Chemical Engineering, Central South University, Changsha 410083, China. Tel.: +86 731 8879850; fax: +86 731 8879850. E-mail address: [email protected] (S. Shi). 0021-9673/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2008.09.004

70 species and one variety, while Taraxacum mongolicum Hand.Mazz. is the famous traditional Chinese medicine, which was commonly used by Chinese local herbal physicians. The herb is frequently used to treat inflammatory disorders and viral infectious diseases in Pharmacopoeia Chinensis [2]. Furthermore, fresh leaves of T. mongolicum have been used by local people as vegetable food in Northern China. Additionally, extracts are used as flavour components in various food products, including alcoholic beverages and soft drink, frozen dairy desserts, candy, baked goods, gelatins and puddings and cheese for its reputed medicinal properties. Although T. mongolicum is a well-known traditional herb and food with a long history, only limited scientific information is available to justify the reputed use until recently. In fact, medical plant therapy is mainly based on the empirical findings during hundreds and thousands of years [3], which urgently demand for separation, identification of major bioactive secondary metabolites from T. mongolicum. In

146

S. Shi et al. / J. Chromatogr. A 1209 (2008) 145–152

our preliminary experiments, the methanolic extracts from T. mongolicum exhibited a high DPPH radical-scavenging activity. The free radical scavengers in food have received a great amount of attention as being primary preventive ingredients against various diseases such as cancer, inflammation and cardiovascular diseases [4–7]. Recent studies have described that the antioxidant properties of medicinal plants, foods and beverages are due to their phenolic compounds, such as flavonoids, phenolic acids, tannins and phenolic diterpenes [8,9]. Our previous phytochemical investigations of polar extract of T. mongolicum yielded several flavonoids, one lignan and one guaianolide [10–14], but the antioxidant activity has not been reported. The further isolation, identification and evaluation of the antioxidants are of particular interest not only for their beneficial physiological activity, but also because of the potential that they could replace synthetic antioxidants, which have side effects [15]. The standard procedure of searching for active metabolites was activity-guided fractionation followed with biological screening. However, bioassay-guided fractionation of plant extracts is a timeconsuming, labor intensive and expensive strategy, which also leads to loss of activity during the isolation and purification process due to dilution effects or decomposition [16]. Moreover, owing to the existence of hundreds of bioactive compounds with different types of structures in the natural products, the fast separation and identification of them were almost impossible. Several attempts have been made to accelerate the isolation, identification and bioactivity evaluation processes. Recent publications demonstrated the efficiency of the on-line coupling of HPLC separation and activity determination of radical-scavenging compounds [17–19]. Such techniques apply for a rapid and selective detection of radicalscavenging substance from complex extract. Thus, it is no longer necessary to purify every constituent for off-line assays, leading to very significant reductions of costs and time to obtain results. And the integration of biological and chemical screening into bioactive compound discovery programs reduces the chance of missing novel and unidentified compounds, prevents replication during separation. A combination of HPLC–RSD method with on-line MS for the analysis of radical-scavenging compounds would permit the rapid determination of antioxidant activity and provide the structural identification of the antioxidant compounds involved, and this technique was fast, sensitive and required only minor sample preparation [20]. On the other hand, the complete elucidation of complex plant secondary metabolites frequently requires the aid of more effective auxiliary techniques, such as nuclear magnetic resonance (NMR). The aim of the present study was to monitory the radicalscavenging compounds in T. mongolicum by applying the hyphenized method HPLC–DAD–RSD–ESI-MS and characterizes their chemical structure by NMR. Furthermore, the antioxidant activity of three new compounds was established by measuring their radical-scavenging potential.

2. Experimental 2.1. Materials and chemicals The aerial part of T. mongolicum Hand.-Mazz. was purchased from Bozhou, Anhui province in January, 2004, and identified by Professor Liurong Chen. A voucher specimen (TM200401-02) was deposited in Department of Traditional Chinese Medicine and Natural Drug Research, College of Pharmaceutical Sciences, Zhejiang University. Methanol used for analytical HPLC was of chromatographic grade (Merck, Darmstadt, Germany). All aqueous solutions used

in the experiments were prepared with pure water produced by Milli-Q water system (Millipore, Bedford, MA, USA). Other analytical grade chemicals were purchased from Chemical Reagent Factory of Hunan Normal University (Changsha, Hunan, China). Deuterated dimethyl sulfoxide used for NMR was bought from Sigma–Aldrich Chemie GmbH (Steinheim, Germany). DPPH (Sigma–Aldrich Chemie GmbH, Steinheim, Germany), quercetin (National Institute of the Control of Pharmaceutical and Biological Products, Ministry of Health, Beijing, China), multiwell plates (Greiner Bio-One GmbH, Frickenhausen, Germany) and multi-well plates reader (Bio-Tek Instruments, Winooski, Vermont, USA) were used in the antioxidant activity experiments. 2.2. Sample preparation for analytical HPLC Fresh plants of T. mongolicum were dried at 50 ◦ C for 3 days before grinding. About 10 g of the powder was extracted by refluxing with methanol (100 ml) for three times (1.5 h for each time). All the filtrates were combined and concentrated to dryness under reduced pressure by rotary evaporation at 45 ◦ C. The residues were then diluted to 100 ml, stored at 4 ◦ C and then brought to room temperature before analysis. The solution was filtered through 0.45 ␮m membranes, and 20 ␮l of the sample solution was injected for HPLC analysis. 2.3. Instrumental 2.3.1. HPLC analyses Analytical HPLC was consisted of two LC-8A pumps, a Prominence SPD-M20A diode array detector performing the wavelength scanning from 190 to 950 nm, a manual injection valve with a 20 ␮l loop and an LC Solution workstation (Shimadzu, Kyoto, Japan). The target compounds were separated by using a reversed phase Symmetry C18 (250 mm × 4.6 mm I.D., 5 ␮m, Waters, Milford, MA, USA) column and a security guard C18 ODS (4.0 mm × 3.0 mm I.D.) from Phenomenex (Torrance, CA, USA). The mobile phase consisted of 0.1% aqueous acetic acid (solvent A) and methanol (solvent B) in a gradient elution mode. The flow rate was kept at 0.9 ml/min, while the ambient temperature was controlled at 20 ◦ C by air conditioner. Spectra were recorded from 200 to 500 nm (peak width 0.2 min and data rate 1.25 s−1 ) while the chromatogram was acquired at 254 nm. 2.3.2. HPLC–DAD–RSD–ESI-MS instrumentation An on-line HPLC–DAD–RSD–ESI-MS method has been described for a rapid screening and identification of radical-scavenging components by using a methanolic solution of DPPH stable free radicals [20,21]. The analytical HPLC analysis was conducted on a Waters Alliance 2690 liquid chromatographic system. The stationary phase and the elution gradient were the same as those in the HPLC analysis. The crude extract was dissolved in methanol and injected into the HPLC system. Before delivering into the system the solvent was filtered through 0.45 ␮m polytetrafluoroethylene (PTEE) filter and degassed using vacuum. The flow rate was 0.9 ml/min at 20 ◦ C. The column eluent was split at a ratio of 3:1 using an adjustable highpressure stream splitter (Supelco Port, Bellefonte, PA), the lower one was introduced into a Micromass ZQ 2000 ESI-MS while a Masslynx 4.0 data system were equipped with a XTerra MS. Negative ion mode for ESI-MS was selected. Mass detection was performed in full scan mode for m/z in the range 100–800. The following settings were applied to the instrument: capillary and cone voltage were 2500 and 40 V, respectively; nebulizer nitrogen gas flow rate was 500 l/h; the ionization sources were worked at 120 ◦ C. The desolvation temperature was 450 ◦ C. The larger flow was again split into two streams using an adjustable high-pressure stream splitter. One part (0.2 ml/min) was used for the radical-scavenging detection.

S. Shi et al. / J. Chromatogr. A 1209 (2008) 145–152

147

Fig. 1. The separation scheme of T. mongolicum.

The length of the capillary used for the post-column reaction was adjusted to achieve a reaction time of 0.6 min. The antioxidants reacted post-column with the DPPH radical at a concentration of 50 mg/l in methanol. The flow of the DPPH radical solution was set to 0.2 ml/min. The DPPH radical-scavenging detection chromatogram is detected as a negative peak at 517 nm with a variable wavelength PC300 detector and the chromatogram was accordingly recorded on a model SCJS-3000 workstation (Tianjin Scientific Instrument Ltd., Tianjin, China). The other part (0.4 ml/min) was continuously monitored by a 2487 dual wavelength UV detector in the range of 200–500 nm, and the chromatograms were recorded at 254 nm. 2.3.3. PHPLC experiment The PHPLC experiments were performed on a self-assembled instrument, which was composed of a P3000 delivery pump, a 2PB00C sample injection pump, UV3000 variable wavelength detector with detection monitored at 254 nm and an SCJS-3000 ChemStation. A preparative column (500 mm × 80 mm I.D.) packed

with 5 ␮m ODS C18 (Fuji, Japan) was used for the preparative separation. The flow rate was 25 ml/min at 20 ◦ C. 2.4. Preparative isolation of radical-scavenging compounds The pulverized material of T. mongolicum (5.0 kg) was exhaustively extracted with methanol under reflux for 8 h and concentrated under reduced pressure to give brown syrup (637 g). A mass of 400 g of syrup was suspended in water and was carried out by open column chromatography over C18 (2.0 kg, 40 ␮m, Elite Co., Dalian, China) using H2 O–MeOH (1:0–0:1) mixtures as gradient elution solvent. Four main fractions were collected (F1 , 10% methanol elution; F2 , 30% methanol elution; F3 , 50% methanol elution; F4 , 70% methanol elution) according to the TLC. Every fraction was further purified by preparative HPLC. In the preparative HPLC, a linear gradient elution of solvent A (0.1% aqueous acetic acid) and solvent B (methanol) was used according to polarity of every fraction. Elution of F1 run from 20% to 25% B for 35 min, and then with 25% B for 60 min to afford pure compounds 1 (12 mg), 2 (9 mg),

Fig. 2. HPLC–UV (254 nm) (A) and DPPH radical-scavenging detection profile (B) of T. mongolicum methanolic extract. Peaks 1–32, see Fig. 3.

148

Table 1 Spectral data of the main antioxidants acquired by HPLC–DAD–RSD–ESI–MS and NMR from methanolic extract of T. mongolicum 1

595

10

22.23

257, 356

625

12

27.90

258, 357

595

15

31.61

253, 268, 345

447

17

35.86

253, 360

625

18

36.94

253, 268, 345

447

19

37.41

258, 347

625

20 21

39.21 41.90

256, 360 262, 370

301 301

22

44.59

254, 349

609

23

46.78

254, 362

269

25

51.26

250, 344

285

26

53.34

264, 323

591

28

58.06

256, 351

283

31

64.70

256, 356

343

32

67.76

253, 349

387

13.11 (1H, br s, 5-OH), 7.31 (1H, s, H-6 ), 7.16 (1H, s, H-3), 6.77 (1H, s, H-3 ), 6.72 (1H, s, H-8), 6.44 (1H, s, H-6), 5.08 (1H, d, J = 8.0 Hz, H-1 ), 4.87 (1H, d, J = 6.0 Hz, H-1 ), 3.20–3.60 (6H, m, H-2 to H-6 ), 3.20–3.60 (5H, m, H-2 to H-5 ) 13.07 (1H, s, 5-OH), 7.31 (1H, s, H-6 ), 7.11 (1H, s, H-3), 6.80 (1H, s, H-3 ), 6.73 (1H, s, H-8), 6.45 (1H, s, H-6), 5.08 (1H, d, J = 8.0 Hz, H-1 ), 4.91 (1H, br s, H-1 ), 3.20–3.60 (6H, m, H-2 to H-6 ), 3.20–3.60 (6H, m, H-2 to H-6 ) 7.30 (1H, s, H-6 ), 7.08 (1H, s, H-3), 6.73 (2H, br s, H-3 and H-8), 6.44 (1H, s, H-6), 5.07 (1H, d, J = 7.2 Hz, H-1 ), 4.89 (1H, d, J = 6.4 Hz, H-1 ), 3.20–3.60 (6H, m, H-2 to H-6 ), 3.20–3.60 (5H, m, H-2 to H-5 ) 7.43 (1H, dd, J = 8.4, 2.0 Hz, H-6 ), 7.42 (1H, d, J = 2.0 Hz, H-2 ), 6.90 (1H, d, J = 8.0 Hz, H-5 ), 6.81 (1H, d, J = 2.0 Hz, H-8), 6.70 (1H, s, H-3), 6.51 (1H, d, J = 2.0 Hz, H-6), 5.08 (1H, d, J = 6.4 Hz, H-1 ), 3.20–3.60 (6H, m, H-2 to H-6 ) 7.91 (1H, dd, J = 8.4, 2.0 Hz, H-6 ), 7.83 (1H, d, J = 2.0 Hz, H-2 ), 7.58 (1H, J = 8.4 Hz, H-5 ), 6.73 (1H, d, J = 2.0 Hz, H-6), 6.68 (1H, d, J = 2.0 Hz, H-8), 5.67 (1H, d, J = 7.6 Hz, H-1 ), 5.58 (1H, d, J = 6.4 Hz,H-1 ), 4.20–4.60 (6H, m, H-2 to H-6 ), 4.20–4.60 (6H, m, H-2 to H-6 ) 12.96 (1H, br s, 5-OH), 7.44 (1H, dd, J = 8.4, 2.0 Hz, H-6 ), 7.42 (1H, d, J = 2.0 Hz, H-2 ), 6.91 (1H, d, J = 8.0 Hz, H-5 ), 6.77 (1H, d, J = 2.0 Hz, H-8), 6.72 (1H, s, H-3), 6.45 (1H, d, J = 2.0 Hz, H-6), 5.07 (1H, d, J = 7.2 Hz, H-1 ), 3.73 (1H, dd, J = 10.0, 2.8 Hz, H-6 a), 3.51 (1H, t, J = 8.5 Hz, H-3 ), 3.42 (1H, dd, J = 10.0, 2.8 Hz, H-6 b), 3.39 (1H, t, J = 8.5 Hz, H-2 ), 3.28 (1H, t, J = 8.5 Hz, H-4 ), 3.19 (1H, m, H-5 ) 7.91 (1H, dd, J = 8.4, 2.0 Hz, H-6 ), 7.83 (1H, d, J = 2.0 Hz, H-2 ), 7.58 (1H, J = 8.4 Hz, H-5 ), 6.72 (1H, d, J = 2.0 Hz, H-6), 6.66 (1H, d, J = 2.0 Hz, H-8), 5.83 (1H, d, J = 7.6 Hz, H-1 ), 5.58 (1H, d, J = 6.4 Hz, H-1 ), 4.20–4.60 (6H, m, H-2 to H-6 ), 4.20–4.60 (6H, m, H-2 to H-6 ) 13.05 (1H, s, 5-OH), 7.26 (1H, s, H-6 ), 7.01 (1H, s, H-3), 6.50 (1H, s, H-3 ), 6.38 (1H, d, J = 1.2 Hz, H-8), 6.16 (1H, d, J = 1.2 Hz, H-6) 12.45 (1H, br s, 5-OH), 7.68 (1H, d, J = 2.0 Hz, H-2 ), 7.54 (1H, dd, J = 8.4, 2.0 Hz, H-6 ), 6.89 (1H, J = 8.4 Hz, H-5 ), 6.41 (1H, d, J = 2.0 Hz, H-6), 6.20 (1H, d, J = 2.0 Hz, H-8) 12.01 (1H, br s, 5-OH), 6.97 (1H, d, J = 2.0 Hz, H-2 ), 6.88 (1H, d, J = 8.0 Hz, H-5 ), 6.83 (1H, dd, J = 8.0, 2.0 Hz, H-6 ), 6.14 (1H, d, J = 2.0 Hz, H-8), 6.13 (1H, d, J = 2.0 Hz, H-6), 5.50 (1H, dd, J = 11.0, 5.0 Hz, H-2), 4.97 (1H, d, J = 7.2 Hz, H-1 ), 4.54 (1H, br s, H-1 ), 3.78 (3H, s, 4-OCH3 ), 3.20–3.60 (6H, m, H-2 to H-6 ), 3.20–3.60 (3H, m, H-2 to H-6 ), 3.11 (1H, dd, J = 17.0, 11.0 Hz, H-3a), 2.78 (1H, dd, J = 17.0, 5.0 Hz, H-3b), 2.51 (1H, d, J = 6.0 Hz, H-5 ), 1.09 (3H, d, J = 6.0 Hz, H-6 ) 12.01 (1H, br s, 5-OH), 6.94 (1H, d, J = 2.0 Hz, H-2 ), 6.94 (1H, dd, J = 8.6, 2.0 Hz, H-6 ), 6.88 (1H, d, J = 8.6 Hz, H-5 ), 5.91 (1H, d, J = 1.2 Hz, H-6), 5.89 (1H, d, J = 1.2 Hz, H-8), 5.43 (1H, dd, J = 11.0, 5.0 Hz, H-2), 3.78 (3H, s, 4 -OCH3 ), 3.20 (1H, dd, J = 17.0, 11.0 Hz, H-3a), 2.72 (1H, dd, J = 17.0, 5.0 Hz, H-3b) 12.94 (1H, br s, 5-OH), 7.41 (1H, dd, J = 8.4, 2.0 Hz, H-6 ), 7.39 (1H, d, J = 2.0 Hz, H-2 ), 6.89 (1H, d, J = 8.4 Hz, H-5 ), 6.64 (1H, s, H-3), 6.44 (1H, d, J = 2.0 Hz, H-8), 6.19 (1H, d, J = 2.0 Hz, H-6) 12.89 (1H, br s, 5-OH), 8.04 (2H, dd, J = 9.0, 2.0 Hz, H-6 and H-2 ), 7.15 (2H, dd, J = 9.0, 2.0 Hz, H-3 and H-5 ), 6.92 (1H, s, H-3), 6.80 (1H, d, J = 2.0 Hz, H-8), 6.47 (1H, d, J = 2.0 Hz, H-6), 5.07 (1H, d, J = 7.2 Hz, H-1 ), 4.58 (1H, br s, H-1 ), 3.88 (3H, s, 7-OCH3 ), 3.20–3.60 (6H, m, H-2 to H-6 ), 3.20–3.60 (3H, m, H-2 to H-6 ), 2.51 (1H, d, J = 6.0 Hz, H-5 ), 1.10 (3H, d, J = 6.0 Hz, H-6 ) 12.89 (1H, br s, 5-OH), 8.00 (2H, dd, J = 9.0, 2.0 Hz, H-6 and H-2 ), 7.08 (2H, dd, J = 9.0, 2.0 Hz, H-3 and H-5 ), 6.84 (1H, s, H-3), 6.48 (1H, d, J = 2.0 Hz, H-8), 6.18 (1H, d, J = 2.0 Hz, H-6), 3.84 (3H, s, 7-OCH3 ) 12.66 (1H, br s, 5-OH), 9.89 (1H, br s, 3-OH), 7.68 (1H, d, J = 2.0 Hz, H-2 ), 7.63 (1H, dd, J = 8.4, 2.4 Hz, H-6 ), 6.97 (1H, J = 8.4 Hz, H-5 ), 6.77 (1H, d, J = 2.0 Hz, H-6), 6.38 (1H, d, J = 2.0 Hz, H-8), 3.97 (6H, s, 2↑OCH3 ), 3.81 (3H, s, OCH3 ). 12.58 (1H, br s, 5-OH), 7.73 (1H, dd, J = 8.8, 2.4 Hz, H-6 ), 7.67 (1H, d, J = 2.4 Hz, H-2 ), 7.17 (1H, J = 8.8 Hz, H-5 ), 6.93 (1H, s, H-8), 3.93 (4 -OMe), 3.87 (3 -OMe), 3.86 (7-OMe), 3.83 (6-OMe), 3.74 (3-OMe)

tR (min)

UV (max , nm)

Phenylpropanoid derivatives 4 8.99 258, 301 sh, 352 6 14.98 243, 300 sh, 332

177 179

8

16.85

244, 300 sh, 330

353

11

26.78

223, 302 sh, 310

163

13

29.19

233, 300 sh, 321

193

16

33.71

244, 299 sh, 329

207

H NMR ([2 H6 ]DMSO, 400 MHz), ı (ppm)

7.66 (1H, d, J = 9.2 Hz, H-4), 7.30 (1H, s, H-5), 7.17 (1H, s, H-8), 6.29 (1H, d, J = 9.2 Hz, H-3) 7.14 (1H, d, J = 16.0 Hz, H-7), 7.02 (1H, d, J = 2.0 Hz, H-2), 6.95 (1H, dd, J = 8.0, 2.0 Hz, H-6), 6.76 (1H, d, J = 8.0 Hz, H-5), 6.15 (1H, d, J = 16.0 Hz, H-8) 7.43 (1H, d, J = 16.0 Hz, H-7 ), 7.03 (1H, d, J = 2.0 Hz, H-2 ), 6.98 (1H, dd, J = 8.4, 2.0 Hz, H-6 ), 6.76 (2H, d, J = 8.2 Hz, H-5 ), 6.15 (1H, d, J = 16.0 Hz, H-8 ), 5.10 (1H, ddd, J = 10.1, 9.8, 4.6 Hz, H-3), 3.95 (1H, m, H-5), 3.57 (1H, m, H-4), 2.04 (1H, m, H-6a), 2.00 (2H, m, H-2), 1.81 (1H, m, H-6b) 12.07 (1H, br s, 1-COOH), 9.91 (1H, br s, 4-OH), 7.51 (1H, d, J = 16 Hz, H-7), 7.49 (2H, d, J = 7.6 Hz, H-2 and H-6), 6.80 (2H, d, J = 7.6 Hz, H-3 and H-5), 6.28 (1H, d, J = 16 Hz, H-8) 7.49 (1H, d, J = 16.0 Hz, H-7), 7.27 (1H, d, J = 1.6 Hz, H-2), 7.09 (1H, dd, J = 8.4, 2.0 Hz, H-6), 6.80 (1H, d, J = 8.0 Hz, H-5), 6.35 (1H, d, J = 16.0 Hz, H-8), 3.82 (3H, s, 3-OCH3 ) 7.46 (1H, d, J = 16.0 Hz, H-7), 7.03 (1H, d, J = 2.0 Hz, H-2), 6.98 (1H, dd, J = 8.0, 2.0 Hz, H-6), 6.75 (1H, d, J = 8.0 Hz, H-5), 6.24 (1H, d, J = 16.0 Hz, H-8), 4.14 (2H, q, J = 7.2 Hz, CH2 ), 1.22 (3H, t, J = 7.2 Hz, CH3 )

Structure assignment Isoetin-7-O-␤-d-glucopyranosyl-2 -O-␣-larabinopyranosidec Isoetin-7-O-␤-d-glucopyranosyl-2 -O-␣-dglucopyranosidea Isoetin-7-O-␤-d-glucopyranosyl-2 -O-␤-dxyloypyranosidea Luteolin-7-O-␤-d-galactopyranosideb Quercetin-7-O-[␤-dglucopyranosyl(1 → 6)-␤-dglucopyranoside]b Luteolin-7-O-␤-d-glucopyranosided

Quercetin-3,7-di-O-␤-d-diglucopyranosideb Isoetinb Quercetind Hesperidinb

4 ,5,7-Trihydroxy-3 -methoxyflava-noneb

Luteolind Genkwanin-4 -O-␤-d-lutinosideb

Genkwaninb Quercetin-3 ,4 ,7-trimethyl etherb Artemitind

Esculetind Caffeic acidd 3-O-Caffeoylquinic acidd

p-Coumaric acidc Ferulic acidc Caffeic acid ethyl esterb

S. Shi et al. / J. Chromatogr. A 1209 (2008) 145–152

ESI [M−H]−

Flavonoid derivatives 9 18.87 258, 357

149

Table 2 The scavenging effects of compounds 10, 12 and 30 isolated from T. mongolicum on DPPH radical Compounds 10 12 30 Quercetinb a b

DPPH radical scavenging, IC50 (␮mol/l)a 21.57 19.76 34.19 5.53

± ± ± ±

2.53 2.83 2.10 0.76

The values presented mean ± SD from triple experiments. Positive control, which was only accessed in off-line experiment.

3 (19 mg), 4 (13 mg), 5 (15 mg), 6 (64 mg), 7 (11 mg), 8 (5 mg), 9 (41 mg), 10 (32 mg) and 12 (29 mg). Elution of F2 started with 25% B for 25 min, then from 25% to 35% B for 20 min, and then 35% B for 35 min, to yield pure compounds 11 (21 mg), 13 (18 mg), 14 (15 mg), 15 (13 mg), 17 (9 mg), 18 (11 mg) and 19 (23 mg). For F3 , the elution run with 40% B for 40 min, reaching 55% B after 35 min. Pure compounds 16 (12 mg), 20 (21 mg), 22 (8 mg), 24 (34 mg), 26 (9 mg), 27 (11 mg) and 29 (18 mg) were obtained. F4 was submitted to isocratic conditions of 50% B for 30 min, linear gradient from 50% B to 70% B in 30 min, which remained at 70% B for 20 min, and then left to reach 90% B for 25 min, affording pure compounds 21 (13 mg), 23 (11 mg), 25 (15 mg), 28 (7 mg), 30 (8 mg), 31 (9 mg) and 32 (6 mg). All the fractions eluted were monitored by UV light at 254 nm. The collected fractions were evaporated to dryness in vacuum, while small portions of water were added occasionally to avoid acid glycoside hydrolysis, and the residues were lyophilized. The samples were redissolved in methanol and further purified by Sephadex LH-20 column (40 cm × 2.5 cm) using methanol as the mobile phase. Thirty-two compounds were separated and their chromatographic purities were determined by HPLC as higher than 97% for all individual compounds. The scheme in Fig. 1 summarizes the partitioning and the entire separation process from T. mongolicum. 2.5. Characterization of the target compounds Characterization of the target compounds was accomplished by their spectroscopic spectra, mass data and NMR spectra. The NMR experiments were performed on a VARIAN INOVA-400 NMR spectrometer (Varian Corporation, Palo Alto, CA, USA). The reference compound TMS was used as internal standard for the determination of chemical shifts.

c

d

a

Substances first reported by our group. Known substances first isolated from Taraxacum genus. Known substances first found in T. mongolim. Substances already identified elsewhere.

2.6. Off-line DPPH free radical-scavenging activity

b

Gallic acidc 3,5-Dihydroxybenzoic acidc Gallicinc p-Hydroxybenzoic acidc Syringic acidc 1-Hydroxymethyl-5-hydroxy-phenyl-2-O␤-d-glucopyranosideb 12.12 (1H, br s, 1-COOH), 9.12 (2H, br s, 3-OH and 5-OH), 8.75 (br s, 4-OH), 6.91 (2H, d, J = 3.2 Hz, H-2 and H-6) 12.60 (1H, br s, 1-COOH), 9.49 (2H, br s, 3-OH and 5-OH), 6.80 (2H, t, J = 2.0 Hz, H-2 and H-6), 6.42 (1H, t, J = 2.0 Hz, H-4) 9.17 (3H, br s, 3-OH, 4-OH and 5-OH), 6.94 (2H, d, J = 3.2 Hz, H-2 and H-6), 3.74 (3H, s, OCH3 ) 12.36 (1H, br s, 1-COOH), 10.17 (1H, br s, 4-OH), 7.78 (2H, d, J = 7.6 Hz, H-2 and H-6), 6.84 (2H, d, J = 7.6 Hz, H-3 and H-5) 12.48 (1H, br s, COOH), 9.13 (1H, br s, 4-OH), 7.23 (2H, s, H-2 and H-6), 3.81 (6H, s, 3-OCH3 and 5-OCH3 ) 7.07 (1H, m, H-3), 6.81 (1H, d, J = 2.6 Hz, H-6), 6.68 (1H, dd, J = 8.5, 2.6 Hz, H-4), 4.76 (1H, d, J = 6.9 Hz, H-1 ), 4.54 (2H, dd, J = 13.3, 4.6 Hz, H-7), 3.30–3.90 (5H, m, H-2 to H-6 ) Hydroxybenzoic acid derivatives 1 4.02 273 2 5.83 271 3 7.69 276 5 11.25 275 7 15.37 272 14 30.58 259

169 153 183 137 197 301

Mongolicumin Aa 353 62.99 30

226, 276, 317, 373

3,4-di-O-Caffeoylquinic acidb 515 59.41 29

247, 300 sh, 330

4,5-di-O-Caffeoylquinic acidb 515 56.67 27

247, 300 sh, 326

49.00 24

248, 300 sh, 329

515

3,5-di-O-Caffeoylquinic acidb

7.59 (1H, d, J = 16.0 Hz, H-7 ), 7.55 (1H, d, J = 16.0 Hz, H-7 ), 7.04 (2H, br s, H-2 and H-2 ), 6.94 (2H, br d, J = 8.4 Hz, H-6 and H-6 ), 6.75 (2H, d, J = 8.4 Hz, H-5 and H-5 ), 6.33 (1H, d, J = 16.0 Hz, H-8 ), 6.24 (1H, d, J = 16.0 Hz, H-8 ), 5.39 (1H, m, H-5), 5.36 (1H, m, H-3), 3.94 (1H, d, J = 4.4 Hz, H-4), 2.16 (2H, m, H-2), 2.29 (1H, br d, J = 13.2 Hz, H-6a), 2.16 (1H, m, H-6b) 7.56 (1H, d, J = 5.6 Hz, H-7 ), 7.48 (1H, d, J = 15.6 Hz, H-7 ), 6.99 (1H, br s, H-2 ), 6.96 (1H, br s, H-2 ), 6.88 (1H, d, J = 8.0 Hz, H-6 ), 6.85 (1H, d, J = 8.0 Hz, H-6 ), 6.71 (2H, m, H-5 and H-5 ), 6.25 (1H, d, J = 15.6 Hz, H-8 ), 6.15 (1H, d, J = 15.6 Hz, H-8 ), 5.74 (1H, br s, H-5), 5.09 (1H, d, J = 7.2 Hz, H-4), 4.35 (1H, br s, H-3), 2.26 (2H, m, H-2), 2.26 (1H, m, H-6), 2.10 (1H, m, H-6) 7.61 (1H, d, J = 16.0 Hz, H-7 ), 7.57 (1H, d, J = 16.0 Hz, H-7 ), 7.08 (1H, br s, H-2 ), 7.07 (1H, br s, H-2 ), 6.95 (2H, br d, J = 8.4 Hz, H-6 and H-6 ), 6.78 (2H, d, J = 8.4 Hz, H-5 and H-5 ), 6.32 (1H, d, J = 16.0 Hz, H-8 ), 6.30 (1H, d, J = 16.0 Hz, H-8 ), 5.68 (1H, br s, H-3), 5.16 (1H, br s, H-4), 4.27 (1H, br s, H-5), 2.16 (2H, m, H-6), 2.03 (2H, m, H-2) 8.05 (1H, s, H-7), 7.46 (1H, d, J = 8.8 Hz, H-6), 7.40 (1H, s, H-6 ), 7.25 (1H, d, J = 8.8 Hz, H-5), 6.62 (1H, s, H-3 )

S. Shi et al. / J. Chromatogr. A 1209 (2008) 145–152

The quenching of free radicals by T. mongolicum extract and the target compounds were assayed spectrophotometrically at 517 nm against the absorbance of the stable free DPPH radical. The free radical-scavenging efficiency of the compounds was determined by decoloration of the DPPH radical. In brief, reaction mixtures contained various concentrations of the test compounds which were dissolved in dimethyl sulfoxide (DMSO) and DPPH (0.4 mg/ml) dissolved in methanol. Negative and positive controls of the experiment were established by a methanolic solution of DPPH and quercetin (free radical scavenger), respectively. The absorbance was measured at 517 nm after incubating the mixture at 37 ◦ C for 30 min. The antiradical activity was expressed as IC50 (antiradical dose required to cause a 50% inhibition), which was calculated by the following formula: [(Ablank − Asample )/Asample ] × 100, where Ablank is the absorbance of the DPPH radical solution and Asample is the absorbance of the DPPH radical solution after the addition of the sample [22].

150

S. Shi et al. / J. Chromatogr. A 1209 (2008) 145–152

3. Results and discussion

DPPH is a free radical compound and has been used widely to test the free radical-scavenging ability of various samples. Some plant extracts of the genus Taraxacum were found to possess significant DPPH radical-scavenging activity [1,23,24]. In the present study, the antioxidant activity of the T. mongolicum methanolic extract was assessed by the ability to scavenge DPPH radical, and the IC50 value of it was 70.1 ␮g/ml. The result implies that there are abundant antioxidative compounds present in the methanolic extract of T. mongolicum.

gradient program (gradient time, gradient shape and initial composition of the mobile phase), column temperature and detection wavelength (relatively higher absorption) were investigated. The final results showed that best resolution and shortest analysis time were achieved when a gradient elution mode composed of solvent A 0.1% acetic acid and solvent B methanol was used in the following linear gradient combination: at 0 min, 78% A; at 20 min, 65% A; at 35 min, 50% A; at 45 min, 50% A, at 55 min, 30% A; at 70 min, 0% A. The flow rate was 0.9 ml/min, the column temperature was set at 20 ◦ C, and the 254 nm was selected as the detection wavelength. Under the optimum gradient elution, the compounds in the methanolic extract of T. mongolicum reached base-line separation (Fig. 2A).

3.2. Selection of suitable chromatographic conditions

3.3. On-line HPLC–DAD–RSD–ESI-MS method

The separation of all compounds in complex extract is one of the challenging tasks in analytical HPLC. According to the present knowledge, the Taraxacum genus comprised a mixture of different bioactive compounds belonging to different chemical types, such as flavonoids, sesquiterpenes, triterpenes, phenolic acids, sterols, etc. Among them, flavonoids and phenolic acids were the more abundant metabolites, which always had similar structure skeleton. For HPLC-MS, the compositions of the mobile phase were limited, where phosphate buffer solutions and extreme pH conditions should be avoided for preventing hydrolysis during the sample process. In the course of optimizing the conditions of separation, the system conditions including the mobile phase (methanol–acetic acid, acetonitrile–water and different concentrations of acetic acid in water were compared to get the most suitable mobile phase),

The on-line coupling of separation and activity determination techniques (HPLC–RSD) can be used for a rapid and selective assessment of radical-scavenging substances in complex mixtures, particularly plant extract with a minimum of sample preparation. The more rapidly the absorbance decreases, the stronger will be the ability of a specific compound as hydrogen-donor, and so its potency as an antioxidant agent [21]. The methanolic extract of T. mongolicum has been assessed for the free radical-scavenging activity in on-line HPLC–DAD–RSD–ESIMS assay. As shown in Fig. 2, at least 32 eluted constituents were detected and gave positive peaks on the UV detector (254 nm) and negative peaks on the DPPH quenching chromatogram (517 nm). The retention times, UV and MS data of the antioxidants obtained by HPLC–DAD–RSD–ESI-MS were presented in Table 1.

3.1. DPPH radical-scavenging activity of methanolic extract

Fig. 3. Structures of the identified antioxidants.

S. Shi et al. / J. Chromatogr. A 1209 (2008) 145–152

151

Fig. 3. (Continued ).

Compounds identification relied first on UV spectra and reasonable molecular formulae calculated from accurate mass measurements, both obtained from HPLC–DAD–ESI-MS analyses, and comparison of these data with the metabolites previously reported from the Taraxacum genus [1]. Combination of the UV and ESI data allowed the characterization of 32 phenolic structures, which could be classified into three groups, flavonoid derivatives, phenylpropanoid derivatives and hydroxybenzoic acid derivatives. All three types of compounds are known to exist in natural products and to possess a variety of biological activities, particular to their antioxidant activity. However, owing to the unavailability of authentic compounds, the peaks could only be tentatively assigned. For unambiguous identification, further studies are required by using authentic compounds or their off-line NMR spectra, which can determine the type of the substituent groups such as glycosyl and the position for isomers.

3.4. Structure elucidation of the antioxidants The structure of three types of phenolic compounds from T. mongolicum could be elucidated by their UV spectra, molecular weight in negative ion mode and the chemical shifts and spin–spin coupling pattern in 1 H NMR spectra. Compounds 9, 10, 12, 15, 17–23, 25, 26, 28, 30 and 31 had two maximum absorption bands at 250–270 and 340–370 nm in UV spectra, which were the typical spectra of flavonoid derivatives, whereas the presence of higher molecular weight could be interpreted as attached saccharide structures. Compounds 6, 8, 11, 13, 16, 24, 27 and 29 had a similar type of UV spectrum among them, with maximum absorbance at about 254 and 326 nm, presumably corresponding to cinnamic acid derivatives [25]. Compounds 1–3, 5, 7 and 14, with a maximal absorbance at about 275 nm and the hydrogen signals in aromatic region, were identified as

152

S. Shi et al. / J. Chromatogr. A 1209 (2008) 145–152

hydroxybenzoic acid derivatives [26]. The structure of all the antioxidants were finally established by comparison of their spectral data with literature: gallic acid (1), 3,5-dihydroxybenzoic acid (2), gallicin (3), esculetin (4), p-hydroxybenzoic acid (5), caffeic acid (6), syringic acid (7), 3-O-caffeoylquinic acid (8), isoetin-7-O-␤-d-glucopyranosyl-2 -O-␣-l-arabinopyranoside (9), isoetin-7-O-␤-d-glucopyranosyl-2 -O-␣-d-glucopyranoside (10), p-coumaric acid (11), isoetin-7-O-␤-d-glucopyranosyl2 -O-␤-d-xyloypyranoside (12), ferulic acid (13), 1-hydroxymethyl-5-hydroxy-phenyl-2-O-␤-d-glucopyranoside (14), luteolin-7-O-␤-d-galactopyranoside (15), caffeic acid ethyl ester (16), quercetin-7-O-[␤-d-glucopyranosyl(1 → 6)␤-d-glucopyranoside] (17), luteolin-7-O-␤-d-glucopyranoside (18), quercetin-3,7-di-O-␤-d-diglucopyranoside (19), isoetin (20), quercetin (21), hesperidin (22), 4 ,5,7-trihydroxy-3 methoxyflavanone (23), 3,5-di-O-caffeoylquinic acid (24), luteolin (25), genkwanin-4 -O-␤-d-lutinoside (26), 4,5-di-O-caffeoylquinic acid (27), genkwanin (28), 3,4-di-O-caffeoylquinic acid (29), mongolicumin A (30), quercetin-3 ,4 ,7-trimethyl ether (31) and artemitin (32) (Fig. 3). Among them, four compounds 6, 9, 10 and 12 were found to be the major metabolites. Although phenolic compounds are common in Taraxacum genus, seventeen compounds 10, 12, 14–17, 19, 20, 22–24, 26–29, 30 and 31 were isolated from this genus for the first time, while separation of 10, 12 and 30 were first reported by our group, as well as were the separation of some lignans from this genus [11,13,14]. 3.5. Antioxidant activity of the new compounds Phenolic compounds are well known to occur in plant extracts and to possess many different biological activities besides antioxidant activity. In this work, the off-line DPPH radical-scavenging activity of three new compounds was evaluated and the results were incorporated into Table 2. The results showed that two flavonoids (compound 10 and 12) had higher activity than the lignan (compound 30). The potency of a molecule for scavenging the radical is due to the number of hydrogens available for donation by the hydroxyl group. Flavonoids bearing free hydroxyl groups are known to be good radical scavengers, an assumption that is particularly true for the flavonoid B-ring [27]. On the other hand, glycosyl groups do not contribute effectively to radical scavenging [28]. Both of these reasons may support the quite similar antioxidant activity for 10 and 12. 4. Conclusion An on-line HPLC–DAD–RSD–ESI-MS coupled with off-line NMR experiment has been developed for the rapid screening and

identification of radical-scavenging compounds from methanolic compounds of T. mongolicum. The hyphenated techniques used herein provided a rapid and selective detection of natural antioxidants in complex matrices with simple operation. Acknowledgements This work was supported by the National Scientific Foundation of China (no. 20775092) and the Postdoctoral Science Foundation of Central South University (no. 743410000). References [1] K. Schütz, R. Carle, A. Schieber, J. Ethnopharmacol. 107 (2006) 313. [2] Chinese National Pharmacopoeia Committee, Pharmacopoeia Chinensis (Zhong Hua Ren Min Gong He Guo Yao Dian), Chemistry Engineering Publishing, Beijing, 2000, p. 289. [3] A. Gurib-Fakim, Mol. Asp. Med. 27 (2006) 1. [4] C. Borek, J. Nutr. 131 (2001) 1010. [5] E. Middleton, Int. J. Pharm. 34 (1996) 344. [6] A.K. Verma, J.A. Johnson, M.N. Gould, M.A. Tanner, Cancer Res. 48 (1988) 5754. [7] H.C. Wei, L. Tye, E. Bresnick, D.F. Birt, Cancer Res. 50 (1990) 499. [8] U. Krings, R.G. Berger, Food Chem. 72 (2001) 223. [9] H. Wang, X.D. Gao, G.C. Zhou, L. Cai, W.B. Yao, Food Chem. 106 (2008) 888. [10] S.Y. Shi, Y.P. Zhang, K.L. Huang, S.Q. Liu, Y. Zhao, Food Chem. 108 (2008) 402. [11] S.Y. Shi, Y.P. Zhang, Y. Zhao, K.L. Huang, J. Sep. Sci. 31 (2008) 683. [12] S.Y. Shi, K.L. Huang, Y.P. Zhang, S.Q. Liu, Sep. Purif. Technol. 60 (2008) 81. [13] S.Y. Shi, Y.P. Zhang, K.L. Huang, Y. Zhao, S.Q. Liu, Biochem. Syst. Ecol. 36 (2008) 437. [14] S.Y. Shi, Q. Zhou, H. Peng, C.X. Zhou, M.H. Hu, Q.F. Tao, X.J. Hao, J. Stöekigt, Y. Zhao, Chin. Chem. Lett. 18 (2007) 1367. [15] H.S. Choi, H.S. Song, H. Ukeda, M. Sawamura, J. Agric. Food Chem. 48 (2000) 4156. [16] K. Hostettmann, J.L. Wolfender, C. Terreaux, Pharm. Biol. 39 (2001) 18. [17] J.H. Wu, C.Y. Huang, Y.T. Tung, S.T. Chang, J. Agric. Food Chem. 56 (2008) 328. [18] M. Perez-Bonilla, S. Salido, T.A. van Beek, P.J. Linares-Palomino, J. Altarejos, M. Nogueras, A. Sanchez, J. Chromatogr. A 1112 (2006) 311. [19] D. Chandrasekar, K. Madhusudhana, S. Ramakrishna, P.V. Diwan, J. Pharm. Biomed. Anal. 40 (2006) 460. [20] N. Nuengchamnong, C.F. de Jong, B. Bruyneel, W.M.A. Niessen, H. Irth, K. Ingkaninan, Phytochem. Anal. 16 (2005) 422. [21] I.I. Koleva, H.A.G. Niederländer, T.A. van Beek, Anal. Chem. 72 (2000) 2323. [22] A. Tapia, J. Rodriguez, C. Theoduloz, S. Lopez, G.E. Feresin, G.J. SchmedaHirschmann, J. Ethnopharmacol. 95 (2004) 155. [23] H.J. Jeon, H.J. Kang, H.J. Jung, Y.S. Kang, C.J. Lim, Y.M. Kim, E.H. Park, J. Ethnopharmacol. 115 (2008) 82. [24] K.M. Yoo, C.H. Lee, H. Lee, B.K. Moon, C.Y. Lee, Food Chem. 106 (2008) 929. [25] P. Arapitsas, Food Chem. 110 (2008) 1041. [26] J.P. Sun, F. Liang, B. Yuan, P. Li, C.Q. Duan, Molecules 12 (2007) 679. [27] S.V. Jovanovic, S. Steenken, M. Tosic, B. Marjanovie, M.G. Simic, J. Am. Chem. Soc. 116 (1994) 4846. [28] Q.Z. Du, Y.J. Xu, L. Li, Y. Zhao, G. Jerz, P. Winterhalter, J. Agric. Food Chem. 54 (2006) 4186.