UPLC-QTOF-MS-Based Metabolomics and Antioxidant Capacity of Codonopsis lanceolata from Different Geographical Origins

Codonopsis lanceolata (C. lanceolata) has been commonly utilized as a therapeutic plant in traditional medicine. In this study, we examined variations in metabolites in C. lanceolata roots grown in different regions using ultra-high performance liquid chromatography quadrupole time-of-flight mass spectrometry (UPLC-QTOF-MS). Multivariate analysis showed that the metabolite profiles of plants grown in Hoengseong and Jeongseon were more similar to each other than to that of C. lanceolata grown in Jeju. Most primary metabolites were present at higher levels in C. lanceolata grown in Jeju. In contrast, C. lanceolata grown in Hoengseong and Jeongseon had high levels of secondary metabolites such as phenylpropanoids and triterpenoid saponins, respectively. In addition, the bioactive compound content and antioxidant capacity of in C. lanceolata grown in Hoengseong and Jeongseon were observed to be higher than those of C. lanceolata grown in Jeju. This study suggests that metabolomics is an effective approach to investigate the difference of metabolite profiling in C. lanceolata from different geographical origins, and is useful for evaluating its pharmacological potential.


Introduction
Codonopsis lanceolata (C. lanceolata) is a perennial twining vine that is classified in the Campanulaceae family and codonopsis genus. It is widely distributed or cultivated in East Asia, such as South Korea, China, and Japan [1]. C. lanceolata is attracting interest due to its unique flavor and taste, and its roots and leaves are edible raw or cooked. C. lanceolata is composed of various bioactive components, including alkaloids, phenylpropanoids, polyphenols, saponins, and steroids [2]. C. lanceolata extract has been reported to have potential therapeutic effects against obesity, oxidative stress, inflammation, diabetes, and cancer and is associated with cognitive improvement [3][4][5][6]. Additionally, in oriental medicine, C. lanceolata has been used as a medicinal herb to treat various lung inflammationrelated symptoms such as cough, sputum, and bronchitis by increasing lung energy [7]. C. lanceolata has been used as a substitute for ginseng because it has a comparable effect, and its demand is gradually increasing because it is less expensive than ginseng [7][8][9]. Although C. lanceolata consumption is increasing due to its various pharmacological effects on human health, knowledge regarding its major components and biological function is insufficient compared to that of other similar plants.
Because plant metabolite compositions are influenced by environmental conditions such as temperature, soil conditions, and precipitation [10], metabolite composition is a good criterion for determining food quality and geographical origin. Plant metabolomics can be used to simultaneously determine primary and secondary plant metabolites using various analytical techniques such as mass spectrometry (MS) combined with gas chromatography or liquid chromatography (LC), and nuclear magnetic resonance [10,11]. Thus, metabolomics has the advantage of efficiently revealing changes in metabolites within a short time. Ultra-high performance liquid chromatography (UPLC)-MS techniques are rapid and have high sensitivity and chromatographic resolution, facilitating the full analysis of complex samples with a broad range of metabolites [12]. Recently, metabolomics studies of C. lanceolata have been conducted, and a UPLC-quadrupole time-of-flight-MS (UPLC-QTOF-MS)-based metabolomics study evaluating C. lanceolata from various regions of China was reported [13]. Although a previous metabolomic study demonstrated the discrimination of C. lanceolata grown in regions of South Korea, the selected areas are not major producing regions of C. lanceolata in South Korea [14].
In this study, we conducted an untargeted metabolomics using UPLC-QTOF-MS to determine the metabolite profile of C. lanceolata grown in three major production regions in South Korea: Jeju (JJ), Hoengseong (HS), and Jeongseon (JS). Additionally, we analyzed and compared the bioactive compound content and antioxidant capacity of C. lanceolata grown in each region to determine its pharmacological characteristics by region. Our study provided information on the metabolite characteristics influenced by the cultivation environment of C. lanceolata, and may contribute to understanding whether a positive correlation exists in the relationship between the antioxidant capacity and metabolite profile.

Sample Materials and Chemicals
C. lanceolata samples were collected from three regions which are Jeju, Hoengseong, and Jeongseon in South Korea. Samples were obtained from local markets, and each sample was grown on a different farm. C. lanceolata samples were harvested between August and October 2021. Table 1 shows the basic and geographic information of C. lanceolata samples, and Figure S1 shows geographical locations and representative samples of C. lanceolata. After washing several times in running water, the skin was removed. It was freeze-dried for 72 h, at −50 • C condenser temperature and a pressure of 10 mTorr using a freeze dryer (FDCF-12003, Operon, Korea). After lyophilization, the samples were finely powdered and stored at −80 • C.

Preparation of Samples for UPLC-QTOF-MS Analysis
For metabolites extraction, 1 mL of water/ethanol (30:70, v/v) was added to 50 mg of dried C. lanceolata as extraction solvent. The sample was mixed for 2 min and ultrasonicated for 20 min at room temperature. The extracted solution was then centrifuged at 12,000× g rpm for 20 min, and the supernatant solution was filtered through a 0.2 µm PTFE syringe filter (Millipore, Billerica, MA, USA) and transferred to LC-MS vial.

Preparation of the Ethanolic Extract of C. lanceolata
20 g of dried C. lanceolata was extracted with 400 mL of water/ethanol (30:70, v/v) at 70 • C for 4 h using a reflux condenser and cooled. The undissolved residues were filtered through filter paper (No. 2, ADVANTEX, Tokyo, Japan). After filtration, the sample was concentrated by a vacuum evaporator. The residue was freeze-dried and dried extracts were stored at −80 • C.

Determination of Bioactive Compounds
The total phenolic content (TPC) was determined according to the modified Folin-Ciocalteu method [15]. The extracts diluted with water/ethanol (30:70, v/v), were reacted with 0.2 N Folin-Ciocalteu reagent. After 6 min, the 10% sodium carbonate solution was added. After 1 h, the absorbance was measured at 765 nm. The calibration curve of gallic acid was used to calculate the TPC in the extract of C. lanceolata and the calibration curve equation was y = 0.0102x − 0.2243 (r 2 = 0.9994). The results were expressed as mg of gallic acid equivalents (GAE)/100 g of dry extracts. Synergy H1 microplate reader (Bio Tek, Winooski, VT, USA) was used to measure the absorbance.
The total flavonoid content (TFC) was measured based on the aluminum chloride colorimetric method [16]. The extracts in water/ethanol (30:70, v/v) were successively reacted with 5% sodium nitrite and distilled water for 5 min, with 10% aluminum chloride for 6 min, and with 1 M sodium hydroxide for 11 min. The absorbance was measured at 420 nm. The calibration curve of quercetin was used to calculate TFC in the extract of C. lanceolata and the calibration curve equation was y = 0.0013x + 0.0142 (r 2 = 0.9955). The results were shown as mg of quercetin equivalents (QE)/100 g of dry extracts.
2.6. Biological Assays 2.6.1. Cell Culture HepG2 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 100 U/mL penicillin, 100 µg/mL streptomycin, and 10% fetal bovine serum (FBS) at 37 • C in a humidified atmosphere of 5% CO 2 . The cells were obtained from the Korea Cell Line bank (KCLB, Seoul, Korea).  [17]. To investigate intracellular ROS levels, hepG2 cells were seeded in a black 96-well plate at a density of 2 × 10 4 cells/well. After 24 h, the medium was replaced with 10 µM DCF-DA (dissolved in DMSO) for 10 min at 37 • C in the dark. After washing out the excess probe with PBS, the cells were treated with the DW for control, or extracts (500 µg/mL) of C. lanceolata in the presence of 500 µM H2O2. After 40 min, fluorescence was detected by microplate reader using the λ (ex./em.) = 480/530 nm.

DPPH Radical Scavenging Assay
The DPPH free radical scavenging activity of C. lanceolata extracts was evaluated by a method modified from a previous study [15]. 100 µL of different concentrations of sample solutions was added to 96-well plates, with the addition of DPPH solution (0.2 mmol/L, 100 µL). The mixture was shaken and left the plate at room temperature in the dark. After 5 min, the absorbance was measured at 517 nm.

Ferric Reducing Antioxidant Power (FRAP) Assay
The FRAP assay was evaluated by using the MAK369-1KT kit (Sigma, St Louis, MO, USA). The procedure was performed according to the manufacturer's instructions. After the samples were added into each 96-well plates, the assay buffer and the working solution were added and incubated at 37 • C. After 1 h the absorbance was measured at 594 nm.

Antioxidant Assay
Total antioxidant capacity of the extracts was measured using the antioxidant assay kit (CS0790, sigma), according to the manufacturer's protocol. Trolox was used as a standard. After the samples were added into each of the 96-well plates, myoglobin working solution and ABTS (2,2 -azino-bis 3-ethylbenzthiazoline-6-sulfonic acid) substrate working solution were added. All samples were then incubated at room temperature for 5 min. The reaction was stopped by adding stop solution (100 µL) to each well. The endpoint absorbance was measured at 405 nm by a microplate reader.

Statistical Analysis
Multivariate statistical analyses were performed with a unit variance scale using SIMCA-P+ software, version 17.0 (Umetrics, Umeå, Sweden). The statistical significance of the data was analyzed by Kruskal-Wallis test, followed by Dunn's post hoc test using GraphPad Prism (San Diego, CA, USA). A hierarchical cluster analysis (HCA) and heat map were generated using MetaboAnalyst Ver. 5.0.

Metabolite Profiling of C. lanceolata Extracts Using UPLC-QTOF-MS
C. lanceolata samples collected from three different major C. lanceolata-producing areas in South Korea, JJ, JS, and HS, were subjected to extraction and analysis. The factors contributing to the metabolite composition of C. lanceolata from different origins can be complex and may be influenced by environmental factors, such as climate, soil type, and fertilization. The JS and HS areas are in the Gangwon Province, South Korea, located in the central part of the Korean peninsula, are geographically close, and have more similar environmental factors, such as climate and soil type, to each other than to those of JJ. JJ is an island located in the southernmost part of South Korea. JS and HS have a combined mountain and continental climate and have cooler weather conditions; the average annual temperatures in these areas are 10.7 • C and 11.3 • C, respectively. Compared to the other two regions, JJ has a milder environment, with an average annual temperature of 16.5 • C, and it does not have a large annual temperature range. JS and HS are representative limestone regions, and JJ is a volcanic region. These distinct climatic and soil type factors may significantly influence the metabolite profile of C. lanceolata. In addition, the C. lanceolata produced in the three regions are of different ages. C. lanceolata cultivated in JJ, a temperate region, is harvested within 2-3 years. However, 5-6 years are required to cultivate C. lanceolata in JS, where it is cultivated at a high altitude, similar to the environment where wild C. lanceolata grows, at an elevation of 300 m or higher. This difference in maturity may be a factor that affects the differentiation of metabolite profiles in C. lanceolata.
To investigate differences in the metabolite composition of C. lanceolata cultivated in different areas, untargeted metabolic profiling of C. lanceolata extracts was conducted. Root extracts were analyzed by UPLC-QTOF-MS. Representative total ion chromatograms of the extracts from positive and negative UPLC-QTOF-MS ionization modes are shown in Figure S2.
To investigate the discriminatory characteristics and visualize metabolic differences among extracts of C. lanceolata grown in different areas, unsupervised principal component analysis (PCA) was performed with the mass spectra of C. lanceolata, and the results are presented in Figure S3. The PCA model was established using the positive and negative modes (positive, R 2 = 0.611, Q 2 = 0.335; negative, R 2 = 0.578, Q 2 = 0.328). The use of QC samples to assess the precision of detection is a common practice in untargeted metabolic profiling. QC samples were closely clustered in the PCA score plots obtained from positive and negative mode spectra, indicating good stability of the UPLC-QTOF-MS analysis throughout the experiment. The PCA score plots showed that some HS and JS samples partially overlapped with JJ samples, but the three groups of samples could be distinguished under unsupervised conditions.
In the present study, 92 metabolites were putatively identified in C. lanceolata samples from different geographical origins based on their molecular formula. The major classes of identified metabolites were organic acids, amino acids and derivatives, sugars, glycosides, triterpenoids, polyacetylenes, phenylpropanoids, flavonoids, and others. These metabolites are detailed in Table S1. Metabolites were identified as amino acids, organic acids, sugars, and phenylpropanoids by comparing the mass accuracy of precursor ions and their similarity of MS/MS spectra patterns to those in freely accessible metabolite databases, including the Human Metabolome Database and METLIN database. Other metabolite classes, such as glycosides, triterpenoids, polyacetylenes, phenylpropanoids, and flavonoids, were tentatively identified by comparing the exact mass of precursors and fragment ions with previously published data [4,14,[18][19][20][21][22][23][24][25].
To remove the many irrelevant variables and further our understanding of the identified metabolite patterns in C. lanceolata roots, partial least squares-discriminant analysis (PLS-DA), which can maximize the separation between observation groups and improve classification and prediction capabilities, was performed with the identified metabolites. The PLS-DA score plot shows separation patterns among the three groups ( Figure 1A, R 2 X = 0.344, R 2 Y = 0.634, Q 2 = 0.554). The JJ group was mainly separated from the HS and JS groups along the PLS1 axis, and the HS and JS groups were separated by the PLS2 axis. Variation within each group can be attributed to the fact that each sample was grown on a different farm, as each farm had different fertilizers, storage conditions after harvest, age of C. lanceolata, and harvest time. HCA, an unsupervised chemometric method, was also performed with the metabolites identified in C. lanceolata samples to determine the degree of similarity between the groups. A small distance indicates a high degree of association. As shown in Figure 1B, although some samples were located between groups, the samples were separated into two large groups: a group including samples from HS and JS, and another group with samples only from JJ. The multivariate analysis showed differences in the metabolite profiles of C. lanceolata grown in the three regions, although there might be variation within each group due to other factors except geographical origin. The metabolite profiles of C. lanceolata grown in HS and JS were more similar to each other than to that of C. lanceolata grown in JJ.
were separated into two large groups: a group including samples from HS and JS, and another group with samples only from JJ. The multivariate analysis showed differences in the metabolite profiles of C. lanceolata grown in the three regions, although there might be variation within each group due to other factors except geographical origin. The metabolite profiles of C. lanceolata grown in HS and JS were more similar to each other than to that of C. lanceolata grown in JJ.

Differences in C. lanceolata Metabolites According to Geographical Origin
Metabolites that significantly contributed to the differentiation among C. lanceolata samples from different origins were determined using univariate statistical analysis. A total of 35 metabolites with significant p-values using the Kruskal-Wallis test (p < 0.001) were visualized as a heat map, and these metabolites were divided into three groups that showed different concentration patterns (Figure 2). Group 1 metabolites were most abundant in samples from the JJ region, while group 2 and group 3 compound levels were highest in the HS and JS region samples, respectively. In addition, group 2 and group 3 metabolites could be merged into one of the two main clusters, which again showed that a small metabolite profile gap existed between JS and HS samples and that JJ samples had the most distinct metabolite profile.
samples obtained from three different geographical origins.

Differences in C. lanceolata Metabolites According to Geographical Origin
Metabolites that significantly contributed to the differentiation among C. lanceolata samples from different origins were determined using univariate statistical analysis. A total of 35 metabolites with significant p-values using the Kruskal-Wallis test (p < 0.001) were visualized as a heat map, and these metabolites were divided into three groups that showed different concentration patterns (Figure 2). Group 1 metabolites were most abundant in samples from the JJ region, while group 2 and group 3 compound levels were highest in the HS and JS region samples, respectively. In addition, group 2 and group 3 metabolites could be merged into one of the two main clusters, which again showed that a small metabolite profile gap existed between JS and HS samples and that JJ samples had the most distinct metabolite profile. The metabolites in group 1 were mainly primary metabolites ( Figure 3A), such as amino acids and their derivatives (arginine, proline, histidine, lysine, aspartate, leucine, serine, and glutathione), sugars (sucrose, raffinose, and glucose), organic acids (dihydrojasmonic acid and azelaic acid), and vitamins (pantothenic acid). Plant primary metabolites help to carry out the basic functions of living cells, thus substantially contributing to growth and agricultural yields [26]. The high content of primary metabolites in C. lanceolata from JJ allows it to grow in 2-3 years to a size similar to that of C. lanceolata grown in JS for 5-6 years. The metabolites in group 1 were mainly primary metabolites ( Figure 3A), such as amino acids and their derivatives (arginine, proline, histidine, lysine, aspartate, leucine, serine, and glutathione), sugars (sucrose, raffinose, and glucose), organic acids (dihydrojasmonic acid and azelaic acid), and vitamins (pantothenic acid). Plant primary metabolites help to carry out the basic functions of living cells, thus substantially contributing to growth and agricultural yields [26]. The high content of primary metabolites in C. lanceolata from JJ allows it to grow in 2-3 years to a size similar to that of C. lanceolata grown in JS for 5-6 years.
The concentrations of free amino acids in plant tissues are related to the nitrogen supply, and an abundant nitrogen supply increases the total free amino acid concentration in many plant species [27]. A previous study reported that amino acid synthesis and concentrations in soybean leaves grown in the JJ area were increased compared with those grown in Yeongwol, because the soils of JJ have greater levels of organic matter and nitrogen than those in the Yeongwol area near HS and JS, which belong to Gangwon Province [28]. The abundant nitrogen supply in the soil of the JJ area would have caused an increase in amino acid levels in C. lanceolata samples from JJ.  In the present study, elevated levels of sugars such as glucose, sucrose, and raffinose were found in C. lanceolata roots grown in JJ compared with those grown in HS and JS. Factors such as temperature and year of cultivation can affect sugar levels in plants. Root respiration rates decrease at low temperatures, which is negatively correlated with reducing sugar concentration [29]. In a previous study, sugars such as fructose, glucose, and sucrose in ginseng roots were abundant in samples collected at 3 and 4 years compared with those collected at 6 years [30]. Although the temperature in JJ is higher than that in HS or JS, the high sugar concentration in C. lanceolata from JJ may have been influenced by the short cultivation year.
Most group 2 metabolites exhibited the highest content in the HS group and were secondary plant metabolites, including phenylpropanoids ( Figure 3B). As one of the most critical metabolic pathways in plants, the phenylpropanoid pathway contributes to plant development and environmental interactions [31].
Syringin, a bioactive compound, has pharmacological significance because of its antioxidant and anti-inflammatory effects [32]. Syringin were detected at m/z 390 from [M + NH4] + and fragment ions at m/z 161 [18]. Tangshenoside, which is considered a syringin molecule bound to meglutol glucoside, is the major constituent of the genus codonopsis. Tangshenosides I  Tangshenosides I, IV, and VIII have been found in the roots of C. lanceolata [13,20,24]. In vitro and in vivo studies in mice have shown that tangshenoside I contributes to ameliorating skeletal muscle atrophy via activating the PI3K/Akt/mTORC1 pathway and upregulating the SIRT1/PGC-1α pathway [33].
Myristicin, 6-methylcoumarin, and cinnamaldehyde are also types of phenylpropanoids. Myristicin is effective in treating diarrhea, abdominal pain, and anxiety and is known to possess anti-inflammatory activity in the inflammatory response [34]. 6-Methylcoumarin inhibits inflammation by downregulating the MAPK and NF-κB pathways as a potential treatment for inflammatory diseases [35]. Cinnamaldehyde is a potential antidiabetic compound with both hypoglycemic and hypolipidemic effects, that affect glucose concentration and decreases glucolipid levels in streptozotocin-induced diabetic rats [36].
Most group 3 metabolites were triterpenoid saponins (codonolaside I, echinocystic acid, lancemaside D, lancemaside E, aster saponin Hb, and foetidissimoside A) and had the highest abundance in C. lanceolata roots grown in JS ( Figure 3C). Lancemaside D, aster saponin Hb, and foetidissimoside A showed [M − H] − ion at m/z 1087, 925, and 1057, respectively, and a product ion m/z 647 were generated by neutral loss of a tetrasaccharide unit (two xyloses, arabinose, and rhamnose) [19]. Triterpenoid saponins are plant secondary metabolites with a structure derived from the precursor oxidosqualene, with one or more sugar residues added [37]. More than 10 triterpenoid saponins have been identified from C. lanceolata [4]. The saponins found in C. lanceolata are considered mainly responsible for their anti-inflammatory properties [6,25], and their stress regulating and radical scavenging actions [38].
Overall, the JJ group had a high primary metabolite content, and the HS and JS groups had a high secondary metabolite content, although differences in metabolite classes were observed. Environmental stresses such as high salinity, drought, ultraviolet irradiation, nutrient deficiencies, and extreme temperature increases the accumulation of secondary metabolites, which overcomes the stressful condition and allows the plant to adapt to the environment [39,40]. Environmental factors such as the low temperature and low nitrogen supply in the soil compared to that in JJ may have acted as environmental stresses, resulting in secondary metabolites accumulation in C. lanceolata grown in HS and JS. These differences in secondary metabolite content may also cause differences in the amount of bioactive compounds (flavonoids and phenolics) that contribute significantly to antioxidant capacity.

Determination of the Bioactive Compound Content in C. lanceolata
To investigate the relationship between the secondary metabolite contents and the bioactive compound, we measured the TPC and TFC of ethanolic extracts of C. lanceolata related to its antioxidant capacity (Table 2). Phenolic and flavonoid compounds in plants significantly contribute to their antioxidant capacities [41] and are influenced by many factors, including origin, maturation, and climate conditions [42]. The TPC content was greatest in ethanolic extracts of the JS group (426.98 ± 70.05 mg GAE/100 g dry weight), followed by that in the HS group (409.62 ± 42.04 mg GAE/100 g dry weight) and the JJ group (349.98 ± 71.70 mg GAE/100 g dry weight). The TFC content in ethanolic extracts of C. lanceolata was also highest in the JS group (281.00 ± 42.50 mg QE/100 g dry weight), followed by the HS group (256.94 ± 33.30 mg QE/100 g dry weight) and the JJ group (185.49 ± 32.61 mg QE/100 g dry weight). Regarding the metabolomics results, C. lanceolata in the JS and HS groups had a high amount of secondary metabolites, and similarly, the TPC and TFC contents were higher in C. lanceolata grown in JS and HS than that grown in JJ. The results are presented as mean ± SD. The results are presented as mean ± SD. p-values were calculated from Kruskal-Wallis test followed by Dunn's post hoc test. HS group comparing to JJ group: * p < 0.05, and **** p <0.0001; JS group comparing to JJ group: ## p < 0.01, and #### p <0.0001.

Determination of Antioxidant Capacity
To determine the antioxidant capacity of ethanolic extracts of C. lanceolata, DPPH radical scavenging, FRAP assays, and antioxidant assays were performed. Table 3 shows the antioxidant activity of C. lanceolata grown in the three different areas. When the radical scavenging activity was determined using the DPPH assay, the highest DPPH radical scavenging effect was found in HS groups (IC 50 = 13.89 ± 2.03), and the lowest DPPH radical scavenging effect was found in JJ groups (IC 50 = 20.04 ± 5.72). The FRAP results ranged from 32.06 to 22.53 µmol Ferrous equivalents (FE)/g dry weight, and it is the highest in HS groups and lowest in JJ groups, respectively. The antioxidant assay results ranged from 10.14 to 8.19 µmol Trolox equivalents (TE)/g dry weight. The highest were found in the JS and the lowest values were found in JJ groups. This result may have occurred because of the high content of the bioactive compounds TPC and TFC in the HS and JS groups, and the secondary metabolite levels were also high in these two groups.

Cellular Antioxidant Capacity in H 2 O 2 -Stimulated HepG2 Cells
We evaluated the cellular antioxidant activities of C. lanceolata with different geographic origins using DCF-DA in H 2 O 2 -treated HepG2 cells. The changes in intracellular ROS fluorescence are shown in Figure 4. When the cells were treated with 500 µM H 2 O 2 , a more than 2.1-fold increase in ROS production compared with that in the control sample was observed. Treatment with C. lanceolata from JJ, HS, and JS effectively reduced the H 2 O 2 -mediated production of ROS. This result indicates that C. lanceolata grown in all three regions has excellent antioxidant capacity. Among them, C. lanceolata from JS removed ROS most effectively, and C. lanceolata from JJ had the weakest antioxidant ability. These results support the correlations between the abundance of secondary metabolites and bioactive compounds, and potent antioxidant capacity.

Cellular Antioxidant Capacity in H2O2-Stimulated HepG2 Cells
We evaluated the cellular antioxidant activities of C. lanceolata with different geographic origins using DCF-DA in H2O2-treated HepG2 cells. The changes in intracellular ROS fluorescence are shown in Figure 4. When the cells were treated with 500 μM H2O2, a more than 2.1-fold increase in ROS production compared with that in the control sample was observed. Treatment with C. lanceolata from JJ, HS, and JS effectively reduced the H2O2-mediated production of ROS. This result indicates that C. lanceolata grown in all three regions has excellent antioxidant capacity. Among them, C. lanceolata from JS removed ROS most effectively, and C. lanceolata from JJ had the weakest antioxidant ability. These results support the correlations between the abundance of secondary metabolites and bioactive compounds, and potent antioxidant capacity.

Conclusions
In the present study, ethanolic extracts from C. lanceolata grown in different regions were characterized according to UPLC-QTOF-MS-based metabolomics data and their antioxidant capacity. C. lanceolata grown in different regions yielded significant differences in metabolic profiles, bioactive compound content, and antioxidant capacity. We performed metabolite profiling using UPLC-QTOF-MS for the first time in C. lanceolata grown in the main production area in South Korea, and a total of 92 metabolites were identified in C. lanceolata roots. We found that levels of most primary metabolites were higher in C. lanceolata grown in JJ, and that C. lanceolata grown in HS and JS was rich in secondary metabolites, including phenylpropanoids and triterpenoid saponins, respectively. In addition, a high bioactive compound content and strong antioxidant capacity were observed in C. lanceolata grown in HS and JS. These results indicate that health-beneficial secondary metabolites content in C. lanceolata is proportional to the antioxidant capacity of C. lanceolata. A limitation of our study is that the age of C. lanceolata grown in each region was not consistent. The age of commercially distributed C. lanceolata from various regions differs

Conclusions
In the present study, ethanolic extracts from C. lanceolata grown in different regions were characterized according to UPLC-QTOF-MS-based metabolomics data and their antioxidant capacity. C. lanceolata grown in different regions yielded significant differences in metabolic profiles, bioactive compound content, and antioxidant capacity. We performed metabolite profiling using UPLC-QTOF-MS for the first time in C. lanceolata grown in the main production area in South Korea, and a total of 92 metabolites were identified in C. lanceolata roots. We found that levels of most primary metabolites were higher in C. lanceolata grown in JJ, and that C. lanceolata grown in HS and JS was rich in secondary metabolites, including phenylpropanoids and triterpenoid saponins, respectively. In addition, a high bioactive compound content and strong antioxidant capacity were observed in C. lanceolata grown in HS and JS. These results indicate that health-beneficial secondary metabolites content in C. lanceolata is proportional to the antioxidant capacity of C. lanceolata. A limitation of our study is that the age of C. lanceolata grown in each region was not consistent. The age of commercially distributed C. lanceolata from various regions differs due to its different growth rates depending on the environment, which may have had a critical impact on the metabolite composition. Further studies are needed to discriminate the differences in metabolite profiles caused by environment and maturity. Nevertheless, we determined the metabolite characteristics of C. lanceolata grown in each region, thereby providing information about the production environment conducive to C. lanceolata quality and pharmacological value.