Global Profiling of Various Metabolites in Platycodon grandiflorum by UPLC-QTOF/MS

In this study, a method of metabolite profiling based on UPLC-QTOF/MS was developed to analyze Platycodon grandiflorum. In the optimal UPLC, various metabolites, including major platycosides, were separated well in 15 min. The metabolite extraction protocols were also optimized by selecting a solvent for use in the study, the ratio of solvent to sample and sonication time. This method was used to profile two different parts of P. grandiflorum, i.e., the roots of P. grandiflorum (PR) and the stems and leaves of P. grandiflorum (PS), in the positive and negative ion modes. As a result, PR and PS showed qualitatively and quantitatively different metabolite profiles. Furthermore, their metabolite compositions differed according to individual plant samples. These results indicate that the UPLC-QTOF/MS-based profiling method is a good tool to analyze various metabolites in P. grandiflorum. This metabolomics approach can also be applied to evaluate the overall quality of P. grandiflorum, as well as to discriminate the cultivars for the medicinal plant industry.


Introduction
Platycodon grandiflorum, a perennial herb grown widely in Northeast Asia, contains triterpenoid saponins, carbohydrates, fibers, etc. [1,2]. P. grandiflorum has been used as a food material and a traditional medicine. Platycodin radix, the root of P. grandiflorum, has various benefits for health, and its biological usefulness has been reviewed [3]. Interest in platycodin saponins, the main components of P. grandiflorum, has increased recently due to their novel pharmacological activities, including anti-inflammatory, anti-oxidant, anti-lipidemic, anti-obesity, anti-cancer activities and their ability to improve insulin resistance [4][5][6][7][8][9][10][11]. Furthermore, in a previous study, various components were isolated from the flower of P. grandiflorum, and their biological activities were monitored [12]. Thus, it is critical to study not only Platycodin radix, but also different parts of P. grandiflorum in order to screen it for useful components.
Recently, metabolomics approaches have been used to assess the metabolite contents of individual plant species [13]. For plant metabolomics, well-constructed analytical platforms are necessary [14,15]. Nuclear magnetic resonance spectroscopy and gas chromatography or liquid chromatography (LC) coupled with mass spectrometry (MS) have been widely used to analyze plant metabolites [16][17][18][19][20]. In particular, metabolite profiling by LC/MS is applicable to phenotype and can discriminate individual plant species [21][22][23]. In this study, we constructed a profiling method based on LC/MS to analyze the metabolites of P. grandiflorum. Metabolite extraction protocols and LC/MS conditions were optimized to profile two different parts ((1) P. grandiflorum roots (PR) and (2) P. grandiflorum stems and leaves (PS)).

Construction of LC-MS Conditions to Profile Platycodon grandiflorum Metabolites
For the high-throughput and sensitive analysis of various metabolites in P. grandiflorum, it is necessary to construct a robust method of profiling. The UPLC system with its small particle size column enables the fast and effective separation of various molecules. Furthermore, QTOF/MS is a good tool for a full mass scan with high resolution. Thus, in this study, UPLC-QTOF/MS was applied to profile P. grandiflorum metabolites. First, we used seven standards (i.e., platycoside E, platycodin D3, platycodin D2, platycodin D, polygalacin D, platycogenic acid A and platycodigenin) to optimize the LC-MS conditions. In the negative mode of electrospray ionization (ESI), seven compounds were mainly detected as [M´H]´and [M + COOH]´ions (Table 1). These standards were separated well and eluted for 15 min ( Figure 1) at a flow rate of 450 µL/min by using an ACQUITY BEH C18 column (2.1 mmˆ100 mm, 1.7 µm particle size). Second, two extracts of PR and PS were analyzed in both positive and negative ion modes. As a result, chromatographic data from the positive mode showed poor efficiency of ionization (data not shown). Thus, we carried out the profiling of PR and PS in the negative mode only. Various metabolites of PR and PS were also separated well in 15 min ( Figure 1). The gradient elution program consisted of a first linear gradient from steady (A/B: 90/10) to solvent (A/B: 80/20) for 3 min; followed by a second linear gradient to solvent (A/B: 77/23) for 8 min; a third linear gradient to solvent (A/B: 5/95) for 9 min; and a forth linear gradient to solvent (A/B: 90/10) for 1 min. The column was equilibrated at 10% Solvent B for 4 min before reuse. The total run time was 25 min for each analysis. Third, we validated the performance of the P. grandiflorum metabolites' profiling method. For this, we drew the standard curves of seven standards and calculated their linearity range and correlation. The limits of detection (LODs) of each isolated compound are also listed ( Table 2).

Optimization of Extraction Protocols for Platycodon grandiflorum Metabolites
Next, we optimized the protocols to extract metabolites, including various isolated compounds from two different parts (PR and PS). There are several factors involved in the optimization of extraction protocols. First, it is critical to select suitable solvents for the extraction of various metabolites with different polarities. A single solvent is insufficient to dissolve the wide range of compounds, and several solvents, such as water, ethanol (EtOH) and methanol (MeOH), have been widely used to extract plant metabolites. To find the most effective solvent for extraction, we tried to compare six different solvents (50% EtOH, 70% EtOH, 80% EtOH, 50% MeOH, 70% MeOH and 80% MeOH) while other factors, such as solvent amount (20 mL) and sonication time (30 min), were kept constant. As a result, 70% EtOH exhibited the largest number of peaks and the highest intensity of several compounds. Second, the ratio of solvent to sample is also a critical factor in the metabolites' extraction due to the limited solubility of samples. Thus, we compared four different amounts (10, 20, 30 and 40 mL) of solvent, 70% EtOH, to extract metabolites from the 50 mg of samples while the sonication time (30 min) was kept constant. Among them, 40 mL provided the largest amount of extracts. Third, different sonication times (15,30,45,60 and 75 min) were tested in the use of 40 mL of 70% EtOH, and as a result, it was found that at least 60 min of sonication were required to extract many metabolites. Finally, we optimized the solvent used (70% EtOH), the ratio of solvent to sample (40 mL:50 mg) and the sonication time (60 min) for the effective extraction of metabolites from P. grandiflorum.

Analysis of Various Metabolites in the Stem, Leaf and Roots of Platycodon grandiflorum
We applied the proposed method based on UPLC-QTOF/MS to profile various metabolites in PR and PS. Three species of P. grandiflorum were used to obtain three PRs and three PSs. We extracted the metabolites from the three PRs and three PSs, respectively, analyzed each extract three times (n = 3) and then processed each set of data using the UNIFI TM software (Version 1.7.1; Waters Corp.). As a result, various metabolites, including several platycosides, were identified based on the library of UNIFI software containing information for the molecule's name and formula [24]. From the processed data, compounds identified repeatedly (n = 3) having high mass accuracy (ppm <˘5) are listed in Tables 3 and 4 respectively. In the peak assignment, it is critical to estimate retention time (RT) (RT tolerance: 0.2 min) and mass accuracy. In the samples of PR and PS, four compounds, such as platycoside E, platycodin D, platycodin D2 and platycodin D3, were analyzed with small RT shifts compared to the standard analysis. We also describe the Venn diagram of metabolites analyzed in three PRs and three PSs to show how the metabolic compositions of each plant species (PR-1, -2 and -3 or PS-1, -2 and -3) were different ( Figure 2). Two parts of P. grandiflorum (PR and PS) showed qualitatively and quantitatively different metabolite profiles. Furthermore, the metabolite profiles differed according to individual samples. In particular, several platycosides were abundantly different in the three PRs and three PSs, respectively ( Figure 3). For example, comparing to other PR species, PR-1 has 1.5-times more abundant platycodin D3, and PR-2 has two-times less abundant platycodin A. Compared to other PS species, PS-3 also has 2.5-times more abundant platycodin A.

Optimization of Extraction Protocols for Platycodon grandiflorum Metabolites
Next, we optimized the protocols to extract metabolites, including various isolated compounds from two different parts (PR and PS). There are several factors involved in the optimization of extraction protocols. First, it is critical to select suitable solvents for the extraction of various metabolites with different polarities. A single solvent is insufficient to dissolve the wide range of compounds, and several solvents, such as water, ethanol (EtOH) and methanol (MeOH), have been widely used to extract plant metabolites. To find the most effective solvent for extraction, we tried to compare six different solvents (50% EtOH, 70% EtOH, 80% EtOH, 50% MeOH, 70% MeOH and 80% MeOH) while other factors, such as solvent amount (20 mL) and sonication time (30 min), were kept constant. As a result, 70% EtOH exhibited the largest number of peaks and the highest intensity of several compounds. Second, the ratio of solvent to sample is also a critical factor in the metabolites' extraction due to the limited solubility of samples. Thus, we compared four different amounts (10, 20, 30 and 40 mL) of solvent, 70% EtOH, to extract metabolites from the 50 mg of samples while the sonication time (30 min) was kept constant. Among them, 40 mL provided the largest amount of extracts. Third, different sonication times (15,30,45,60 and 75 min) were tested in the use of 40 mL of 70% EtOH, and as a result, it was found that at least 60 min of sonication were required to extract many metabolites. Finally, we optimized the solvent used (70% EtOH), the ratio of solvent to sample (40 mL:50 mg) and the sonication time (60 min) for the effective extraction of metabolites from P. grandiflorum.

Analysis of Various Metabolites in the Stem, Leaf and Roots of Platycodon grandiflorum
We applied the proposed method based on UPLC-QTOF/MS to profile various metabolites in PR and PS. Three species of P. grandiflorum were used to obtain three PRs and three PSs. We extracted the metabolites from the three PRs and three PSs, respectively, analyzed each extract three times (n = 3) and then processed each set of data using the UNIFI TM software (Version 1.7.1; Waters Corp.). As a result, various metabolites, including several platycosides, were identified based on the library of UNIFI software containing information for the molecule's name and formula [24]. From the processed data, compounds identified repeatedly (n = 3) having high mass accuracy (ppm < ±5) are listed in Tables 3 and 4, respectively. In the peak assignment, it is critical to estimate retention time (RT) (RT tolerance: 0.2 min) and mass accuracy. In the samples of PR and PS, four compounds, such as platycoside E, platycodin D, platycodin D2 and platycodin D3, were analyzed with small RT shifts compared to the standard analysis. We also describe the Venn diagram of metabolites analyzed in three PRs and three PSs to show how the metabolic compositions of each plant species (PR-1, -2 and -3 or PS-1, -2 and -3) were different ( Figure 2). Two parts of P. grandiflorum (PR and PS) showed qualitatively and quantitatively different metabolite profiles. Furthermore, the metabolite profiles differed according to individual samples. In particular, several platycosides were abundantly different in the three PRs and three PSs, respectively (Figure 3). For example, comparing to other PR species, PR-1 has 1.5-times more abundant platycodin D3, and PR-2 has two-times less abundant platycodin A. Compared to other PS species, PS-3 also has 2.5-times more abundant platycodin A.

Platycodon grandiflorum Samples
The samples of Platycodon grandiflorum (roots, stems and leaves) were purchased from a Daegu Herbal Market in Daegu Gyeongbuk Province, Korea, in 2014. Voucher specimen (NIHHS150128) was deposited at the herbarium of the Department of Herbal Crop Research, National Institute of Horticultural and Herbal Science, Rural Development Administration, Eumseong, Korea.

Standard Constituents and Reagents
HPLC-grade acetonitrile, methanol and water were obtained from Merck (Darmstadt, Germany). Formic acid was purchased from Sigma-Aldrich (St. Louis, MO, USA). Standard compounds were isolated and purified from Platycodon grandiflorum roots by a series of chromatography procedures in our laboratory, and their structures were elucidated by a comparison of the spectroscopic data (MS, 1 H-NMR and 13 C-NMR) with the literature data: platycoside E [25], platycodin D3 [26], platycodin D [27], platycodin D2 [26], polygalacin D [26], platycogenic acid A [2] and platycodigenin [28]. The purity of the isolated compounds was determined to be more than 98% by normalization of the peak areas detected by HPLC analysis.

Sample Preparation
Each sample was dried at 40 °C in a forced-air convection-drying oven for 48 h after washing, and then weighed. The main and lateral roots were used in experiments after removing the rhizomes and fine roots. The roots were ground (<0.5 mm) using a mixer (Hanil, Seoul, Korea) and thoroughly mixed, after which the subsamples were homogenized further using a Retsch MM400 mixer mill (Retsch GmbH, Haan, Germany) for the analyses. Fine powder was weighed (50 mg), suspended in 40 mL of 70% (v/v) ethanol and ultrasonically extracted for 1 h at 50 °C. The extract was filtered and evaporated by using a solvent evaporator, Genevac Ez-2 (Genevac Ltd, Suffolk, UK), and the residue (5 mg) was dissolved in the 1 mL of 70% methanol. The solution was filtered through a syringe filter (0.22 µm) and injected directly into the UPLC system.

UPLC-QTOF-MS Analysis
UPLC was performed using a Waters ACQUITY H-Class UPLC (Waters Corp, Milford, MA, USA). Chromatographic separations were performed on an ACQUITY BEH C18 column (2.1 mm × 100 mm, 1.7 µm). The column oven was maintained at 40 °C, and the mobile phases consisted of Solvent A (5% acetonitrile + 0.1% formic acid (v/v)) and Solvent B (95% acetonitrile + 0.1% formic acid (v/v)). The flow rate was 450 µL/min, and the injection volume was 2 µL for each run. Next, MS analysis was performed using a Waters Xevo G2-S QTOF MS (Waters Corp.) operating in positive and negative ion mode. The mass spectrometers performed alternative high-and low-energy scans, known as the MS E acquisition mode. The operating parameters were set as follows: cone voltage,

Platycodon grandiflorum Samples
The samples of Platycodon grandiflorum (roots, stems and leaves) were purchased from a Daegu Herbal Market in Daegu Gyeongbuk Province, Korea, in 2014. Voucher specimen (NIHHS150128) was deposited at the herbarium of the Department of Herbal Crop Research, National Institute of Horticultural and Herbal Science, Rural Development Administration, Eumseong, Korea.

Standard Constituents and Reagents
HPLC-grade acetonitrile, methanol and water were obtained from Merck (Darmstadt, Germany). Formic acid was purchased from Sigma-Aldrich (St. Louis, MO, USA). Standard compounds were isolated and purified from Platycodon grandiflorum roots by a series of chromatography procedures in our laboratory, and their structures were elucidated by a comparison of the spectroscopic data (MS, 1 H-NMR and 13 C-NMR) with the literature data: platycoside E [25], platycodin D3 [26], platycodin D [27], platycodin D2 [26], polygalacin D [26], platycogenic acid A [2] and platycodigenin [28]. The purity of the isolated compounds was determined to be more than 98% by normalization of the peak areas detected by HPLC analysis.

Sample Preparation
Each sample was dried at 40˝C in a forced-air convection-drying oven for 48 h after washing, and then weighed. The main and lateral roots were used in experiments after removing the rhizomes and fine roots. The roots were ground (<0.5 mm) using a mixer (Hanil, Seoul, Korea) and thoroughly mixed, after which the subsamples were homogenized further using a Retsch MM400 mixer mill (Retsch GmbH, Haan, Germany) for the analyses. Fine powder was weighed (50 mg), suspended in 40 mL of 70% (v/v) ethanol and ultrasonically extracted for 1 h at 50˝C. The extract was filtered and evaporated by using a solvent evaporator, Genevac Ez-2 (Genevac Ltd, Suffolk, UK), and the residue (5 mg) was dissolved in the 1 mL of 70% methanol. The solution was filtered through a syringe filter (0.22 µm) and injected directly into the UPLC system.

UPLC-QTOF-MS Analysis
UPLC was performed using a Waters ACQUITY H-Class UPLC (Waters Corp, Milford, MA, USA). Chromatographic separations were performed on an ACQUITY BEH C18 column (2.1 mm1 00 mm, 1.7 µm). The column oven was maintained at 40˝C, and the mobile phases consisted of Solvent A (5% acetonitrile + 0.1% formic acid (v/v)) and Solvent B (95% acetonitrile + 0.1% formic acid (v/v)). The flow rate was 450 µL/min, and the injection volume was 2 µL for each run. Next, MS analysis was performed using a Waters Xevo G2-S QTOF MS (Waters Corp.) operating in positive and negative ion mode. The mass spectrometers performed alternative high-and low-energy scans, known as the MS E acquisition mode. The operating parameters were set as follows: cone voltage, 40 V; capillary, 3.0 kV; source temperature, 120˝C; desolvation temperature, 300˝C; cone gas flow, 30 L/h; and desolvation gas flow, 600 L/h. Accurate mass measurements were obtained by means of an automated calibration delivery system, which contains the internal reference (Leucine, m/z 556.276 (ESI+), m/z 554.262(ESI-)). Data were collected between 100 and 2000 m/z. In the quantitative analysis of each metabolite, the height of peaks was used to measure the intensity.

Conclusions
A profiling method based on UPLC-QTOF/MS was developed to analyze various metabolites contained in P. grandiflorum. UPLC separation conditions were optimized by using seven isolated compounds. The protocols for extracting P. grandiflorum metabolites were optimized as follows: solvent used, 70% EtOH; the ratio of solvent to sample, 40 mL:50 mg; and sonication time, 60 min. We applied this method to profile two different parts of P. grandiflorum (PR and PS), in what was a first attempt to characterize various metabolites in PR and PS, respectively. In the negative ion modes, PR and PS showed qualitatively and quantitatively different metabolite profiles. The metabolite compositions also differed according to the each species. These results indicated that the UPLC-QTOF/MS-based profiling method has potential as a tool to analyze various metabolites in P. grandiflorum. Hence, this study is of great significance regarding evaluations of the overall quality of P. grandiflorum in pharmacological and clinical investigations of drug products. Furthermore, this metabolomics approach could be applied to discriminate cultivars for use in the agricultural and pharmacological industries.