Comprehensive Analysis of Secondary Metabolites in the Extracts from Different Lily Bulbs and Their Antioxidant Ability

The genus Lilium contains more than 100 wild species and numerous hybrid varieties. Some species of them have been used as medicine and food since ancient times. However, the research on the active components and the medical properties of lilies has only focused on a few species. In this study, the total phenolic acid content (TPC), total flavonoid content (TFC), and antioxidant capacity of 22 representative lilies were systematically investigated. The results showed that the TPC, TFC and antioxidant activity were highly variable among different lilies, but they were significantly positively correlated. Hierarchical cluster analysis indicated that L. henryi and L. regale were arranged in one group characterized by the highest TPC, TFC and antioxidant capacity, followed by Oriental hybrids and Trumpet and Oriental hybrids. The traditional edible and medicinal lilies were clustered in low TPC, TFC and antioxidant capacity group. A total of 577 secondary metabolites, including 201 flavonoids, 153 phenolic acids, were identified in the five species with great differences in antioxidant capacity by extensive targeted metabonomics. Differentially accumulated metabolites (DAMs) analysis reviewed that the DAMs were mainly enriched in secondary metabolic pathways such as isoflavonoid, folate, flavonoid, flavone, flavonol, phenylpropanoid, isoquinoline alkaloid biosynthesis, nicotinate and nicotinamide metabolism and so on. Correlation analysis identified that 64 metabolites were significantly positively correlated with antioxidant capacity (r ≥ 0.9 and p < 0.0001). These results suggested that the genus Lilium has great biodiversity in bioactive components. The data obtained greatly expand our knowledge of the bioactive constituents of Lilium spp. Additionally, it also highlights the potential application of Lilium plants as antioxidants, functional ingredients, cosmetic products and nutraceuticals.


Introduction
The genus Lilium, belonging to the family Liliaceae, is mainly distributed in temperate regions of the Northern Hemisphere including Eastern Asia, Europe, and North America [1]. There are approximately 115 species in this genus and 55 of them are native to China [2,3]. Due to their outstanding value as ornamental, edible, and medicinal plants, lilies are widely cultivated worldwide. Secondary metabolites of plants are important sources of many nutrients and natural medicines [4][5][6][7][8]. Plants of the genus Lilium are rich in chemical diversity and have been the focus of natural product chemistry research for decades. Phytochemical studies have shown that, in addition to essential primary metabolites, the chemical constituents of Lilium plants also include many secondary metabolites, including phenolic acids, flavonoids, saponins, sterols and alkaloids, which are important

Preparation of Methanol Extracts
The bioactive compounds were extracted by an ultrasound-assisted method using chromatographic grade methanol. In detail, 100 mg of each sample was weighed in a 10 mL sterile centrifuge tube, and 5 mL methanol was added. After shaking on a vortex oscillator, the extraction was carried out under 30 • C and 400 W ultrasonic conditions for 30 min by a numerically controlled ultrasonic cleaning instrument (UP4000HE, Nanjing Leijunda Ultrasonic Electronic Equipment Co., Ltd., Nanjing, China). The extracts were centrifuged (318K Labourzentrifugen, Sigma, Osterode, Germany) at 8000 rpm for 5 min at 25 • C. The extraction was repeated twice, and the combined supernatants were filtered through a 0.22 µm organic filter for determination of composition and antioxidant capacity.

Total Phenolic Acids
The total phenolic acid content of each extract was determined according to the method of Dowom, with some modifications, by using Folin-Ciocalteu reagent, and gallic acid was used as a standard [26]. Briefly, 400 µL methanol extract, 0.6 mL double distilled water and 0.25 mL 50% Folin-Ciocalteu were added to a 2 mL test tube, and the mixture was placed at room temperature for 2 min after mixing. Then, 750 µL 15% Na 2 CO 3 were added and mixed well, and again, the mixture was placed at room temperature in the dark for 2 h. After incubation, 200 µL of reaction solution were added to a microplate for absorbance determination at 765 nm using a microplate reader (SpectraMax M2, Molecular Devices, Sunnyvale, CA, USA), and 400 µL of methanol were used as a reagent blank. Assays were performed after appropriate dilution for samples with high phenolic acid concentrations. The content of total phenolics in the samples was measured using a calibration equation (y = 0.0052 + 0.0318x, r 2 = 0.9996) and expressed as mg of gallic acid equivalent per g of dry weight (GAE mg/g DW).

Total Flavonoids
The total flavonoid content in each sample was measured according to a previously described spectrophotometric method with some modifications, and quercetin was used as a standard [27]. Briefly, 300 µL of the extract were mixed with 40 µL of a 5% NaNO 2 solution in a test tube. After 6 min, 100 µL of a 10% AlCl 3 ·6H 2 0 solution were added and allowed to stand for another 5 min before 400 µL of 1 M NaOH were added. The mixture was mixed well for each step. Two hundred microliters of the final mixture were added to the microplate for absorbance determination at 510 nm using a microplate reader (SpectraMax M2, Molecular Devices, Sunnyvale, CA, USA), and 300 µL methanol were used as a blank. The total flavonoid content in the samples was calculated by the standard curve equation (y = 0.0012x + 0.0023, r 2 = 0.9990), and the results were expressed as mg quercetin equivalent per g of dry weight (QE mg/g DW).

DPPH Radical Scavenging
The DPPH free radical scavenging ability was assessed according to the method of Jao, with some modifications [28,29], and trolox was used as standard. Briefly, 300 µL of the extract and 950 µL of 0.1 mM DPPH solution were added to a test tube and incubated for 30 min. Two hundred microliters of the final mixture were added to a microplate for absorbance determination at 517 nm using a microplate reader (SpectraMax M2, Molecular Devices, Sunnyvale, CA, USA). Assays were performed after appropriate dilution for samples with high antioxidant activity. The DPPH free radical scavenging ability of the samples was calculated by the standard curve equation (y = 0.0035x − 0.0151, r 2 = 0.9962), and the results were expressed as µg trolox equivalent per g of dry weight (TE µg/g DW).

Ferric Reducing Antioxidant Power (FRAP)
The ferric reducing ability was measured following previous methods with slight modifications [30], and trolox was used as the standard. Briefly, a 300 µL sample solution was mixed with 900 µL Fe 3+ -TPTZ working solution (10 mM solution of 2,4,6-tri(2-pyridyl)-S-triazine (TPTZ) in 40 mM HCl, 20 mM FeCl 3 ·6H 2 O and volumes of acetate buffer pH 3.6). After 5 min of incubation, 200 µL of reaction solution were added to the microplate for absorbance determination at 593 nm using a microplate reader (SpectraMax M2, Molecular Devices, Sunnyvale, CA, USA), and 300 µL of methanol were used as a blank. Assays were performed after appropriate dilution for samples with high antioxidant ability. The FRAP in the samples was calculated by the standard curve equation (y = 0.0261x + 0.0248, r 2 = 0.9991), and the results were expressed as µg trolox equivalent per g of dry weight (TE µg/g DW).

Cupricion Reducing Capacity (CUPRAC)
The cupric ion reducing capacity was determined as described previously with slight changes [31]. Briefly, 100 µL extract or different concentrations of trolox, 0.5 mL of 10 mM CuSO 4 , and 0.5 mL of 7.5 mM neocuproine were added into a 2 mL test tube sequentially. After mixing well, the mixture was incubated for 30 min at room temperature. Then, 200 µL of the final mixture were added to the microplate for absorbance determination at 450 nm using a microplate reader (SpectraMax M2, Molecular Devices, Sunnyvale, CA, USA), and 100 µL methanol were used as a blank. Assays for samples with high antioxidant ability were performed after appropriate dilution. The CUPRAC of the samples was calculated by the standard curve equation (y = 0.003x + 0.1391, r 2 = 0.9983), and the results were expressed as µg trolox equivalent per g of dry weight (TE µg/g DW).

UPLC Conditions
The sample extracts were analyzed using an UPLC-ESI-MS/MS system (UPLC, SHIMADZU Nexera X2, www.shimadzu.com.cn/ (assessed on 21 April 2021); MS, Ap-plied Biosystems 4500 Q TRAP, www.appliedbiosystems.com.cn/ (assessed on 21 April 2021)). The analytical conditions were as follows, UPLC: column, Agilent SB-C18 (1.8 µm, 2.1 mm × 100 mm); the mobile phase consisted of solvent A, pure water with 0.1% formic acid, and solvent B, acetonitrile with 0.1% formic acid. Sample measurements were performed with a gradient program that employed the starting conditions of 95% A, 5% B. Within 9 min, a linear gradient to 5% A, 95% B was programmed, and a composition of 5% A, 95% B was kept for 1 min. Subsequently, a composition of 95% A, 5.0% B was adjusted within 1.1 min and kept for 2.9 min. The flow velocity was set as 0.35 mL per minute; the column oven was set to 40 • C; the injection volume was 4 µL. The effluent was alternatively connected to an ESI-triple quadrupole-linear ion trap (QTRAP)-MS.

ESI-Q TRAP-MS/MS
LIT and triple quadrupole (QQQ) scans were acquired on a triple quadrupole-linear ion trap mass spectrometer (Q TRAP), AB4500 Q TRAP UPLC/MS/MS System, equipped with an ESI Turbo Ion-Spray interface, operating in positive and negative ion mode and controlled by Analyst 1.6.3 software (AB Sciex). The ESI source operation parameters were as follows: ion source, turbo spray; source temperature 550 • C; ion spray voltage (IS) 5500 V (positive ion mode)/−4500 V (negative ion mode); ion source gas I (GSI), gas II (GSII), curtain gas (CUR) were set at 50, 60, and 25.0 psi, respectively; the collisionactivated dissociation (CAD) was high. Instrument tuning and mass calibration were performed with 10 and 100 µmol/L polypropylene glycol solutions in QQQ and LIT modes, respectively. QQQ scans were acquired as MRM experiments with collision gas (nitrogen) set to medium. DP and CE for individual MRM transitions were performed with further DP and CE optimization. A specific set of MRM transitions was monitored for each period according to the metabolites eluted within this period.

Statistical Analysis
The TPC, TFC and antioxidant ability data of different lilies were sorted in Excel 2020 for Windows and analyzed in triplicate, and the results were expressed as the mean ± standard deviation (SD) using SPSS 16.0 for Windows. One-way analysis of variance (ANOVA) and Duncan's multiple range tests were used to determine the significance of the differences among samples, with a significance level of 0.05. A two-tailed Pearson's correlation test was performed to determine the correlations among mean values of different indexes, and hierarchical cluster analysis was used to group Lilium species.
For secondary metabolome data, significantly regulated metabolites between groups were determined by VIP ≥ 1 and absolute Log2FC (fold change) ≥ 1. VIP values were extracted from the OPLS-DA results, which also contained score plots and permutation plots and were generated using the R package MetaboAnalystR. The data were log transformed (log2) and mean centered before OPLS-DA. To avoid overfitting, a permutation test (200 permutations) was performed. The functions of DAMs were annotated based on the KEGG compound database to determine the metabolic pathways.

Total Phenolic Acid and Flavonoid Content
Due to their excellent antioxidant, bacteriostatic and anticancer functions, phenolics and flavonoids are among the most prevalent active components [32,33]. As secondary metabolites, phenolic acids and flavonoids can be extensively found in plant-derived foods, composing significant parts of our daily diet. The levels of TPC and TFC of 22 lily bulbs are shown in Table 2. The TPC was presented by GAE mg/g DW, and the TFC was presented by QE mg/g DW. The TPC of different lilies varies from L. 'Jinghe' (0.6 ± 0.07 GAE mg/g DW) to L. regale (13.73 ± 0.35 GAE mg/g DW), among which wild species L. regale and L. henryi stood out from all the tested lilies, followed by the O and OT types (2.83 ± 0.01~4.21 ± 0.07 GAE mg/g DW). The traditional edible and medicinal lily L. lancifolium, had a higher TPC of 2.25 ± 0.07 GAE mg/g DW than the others, and followed by the edible L. davidii var. willmottiae and Lilium longiflorum hybrid L. 'White Heaven'. The remaining lilies had the lowest, having similar TPCs. The results showed that the TPCs of different genotype lilies varied greatly, which coincides with results from potato in which cultivar was the main variable affecting the biosynthesis of the studied phenolics [34]. The TFCs of the different lilies tested showed a similar trend with the TPCs, but with some differences. The TFCs of L. regale and L. henryi were 3-10 times higher than those of the other tested materials. Although both L. lancifolium and L. brownii var. viridulum are medicinal lilies, our results showed that there were significant differences in the contents of phenolic acids and flavonoids between them ( Table 2). For traditional edible lilies, L. davidii, L. leichtlinii var. maximowiczii, L. davidii var. willmottiae, L. lancifolium and L. brownii var. viridulum, L. lancifolium had the highest levels of both the TPC and TFC, and L. brownii var. viridulum had the lowest levels of TPC and TFC. There was no significant difference in the TFCs of L. davidii, L. leichtlinii var. maximowiczii, and L. davidii var. willmottiae (Table 2). This means that the nutritional and health effects of different lilies may be different. In a previous study, L. regale had higher contents of total phenolics, flavonoids and flavanols than the other five species (L. concolor, L. pumilum, L. leucanthum, L. davidii var. unicolor and L. lancifolium), consequently exhibiting the highest antioxidant ability [31]. We confirmed this result by the study of a broader selection of material, while our study also demonstrated that L. henryi is richer in TPC and TFC than other lilies except L. regale.

Antioxidant Activity
Due to the convenient and fast features of in vitro antioxidant activity assays, a variety of in vitro research methods have been widely used to evaluate the antioxidant activity of plant extracts [35]. Here, DPPH radical-scavenging activity, ferric reducing antioxidant power (FRAP), and cupric ion reducing capacity (CUPRAC) were used to determine the antioxidant ability of the extracts of different lilies [36]. Different concentrations of trolox were used to establish the standard curve for each test. The radical scavenging capacity (DPPH) and reducing power (FRAP and CUPRAC) of various extracts were expressed by the trolox equivalent antioxidant capacity (TE µg/g DW) in Table 2.
As shown in Table 2, the radical scavenging capacity of different lily extracts was significantly different. The best DPPH scavenging property was exhibited by the L. regale extract (16,707.07 ± 847.51 TE µg/g DW), which was 38 times as much as that of the lowest value observed for L. brownii var. viridulum (439.53 ± 171.03 TE µg/g DW). The second was L. henryi (13,249.75 ± 390.47 TE µg/g DW), being 30 times higher than the lowest value. The O and OT hybrids had moderate DPPH scavenging abilities from 2738.94 ± 265.84 to 4971.99 ± 402.33 TE µg/g DW, following a similar pattern as the TPC and TFC. For edible lily, L. lancifolium has the highest antioxidant capacity, although this value was not significantly different than that of L. davidii var. willmottiae. For the reducing power of FRAP and CUPRAC, L. regale (11,622.55 ± 344.81 and 32,304.44 ± 1878.93 TE µg/g DW, respectively) and L. henryi (8736.93 ± 645.45 and 26,593.33 ± 1205.54 TE µg/g DW, respectively) continued to show much higher levels than the other materials. In Rhamnus petiolaris, the contents of flavonols and anthocyanins were highly correlated with DPPH, and phenolic compounds were the main contributors to the antioxidant abilities of the members of the genus Rhamnus [37]. Our study confirmed that there were similar trends among the DPPH, FRAP and CUPRAC values and the TPC and TFC of the different lily bulb extracts, which means that the phenolic acids and flavonoids in lily bulbs might play vital roles in their antioxidant capacity. There were significant differences in TPC and TFC between L. davidii var. willmottiae and L. lancifolium, but there were no significant differences in antioxidant capacity in terms of DPPH, FRAP and CUPRAC values, which might be caused by the differences in antioxidant substances between them [38].

Correlation and Hierarchical Cluster Analysis of TPC, TFC and Antioxidant Capacity
The correlations among the TPC, TFC and antioxidant ability of DPPH, FRAP and CUPRAC were comprehensively analyzed based on the means of different tested indexes of each material. The results showed that all the tested factors were significantly highly correlated at the 0.01 level, and all the coefficients were above 0.977 (Table 3). The correlation coefficients between TPC and antioxidant indexes were 0.997 for DPPH, 0.992 for FRAP and 0.997 for CUPRAC, and the coefficients of TFC were 0.983, 0.986 and 0.977, respectively. This result indicated that although both TPC and TFC exhibited significant positive correlations with antioxidant capacity, TPC was more highly correlated with this capacity. The same pattern was also found in Actinidia extracts [34]. Our results revealed that the TPC was significantly related to the TFC, with a correlation coefficient of 0.977, in lily bulb extracts. This is reasonable since the biosynthesis of many flavonoids occurs downstream of phenolic acids in plants [39,40]. Previous studies suggested that the substances extracted by 80% methanol with 1 M HCl from lilies had no significant correlation with DPPH [31], but our results indicated that the TPC extracted by methanol was highly correlated with DPPH, which coincided with other reported results [38]. This means that the differences in substances extracted by different extraction methods from the same sample may cause divergences in antioxidant capacity [41]. Again, the significant correlation between methods was confirmed with three methods (DPPH, FRAP and CUPRAC). Above all, our results suggested that all the tested bulb extracts exhibited strong antioxidant activities, both in radical scavenging capacity Hierarchical cluster analysis is the most widely used method to classify the objects according to their characteristics [42]. Data on the TPC, TFC and antioxidant capacity (DPPH, FRAP and CUPRAC) were used to carry out hierarchical cluster analysis of the tested lilies, and 22 lilies were separated into three clusters, as shown in Figure 1. As the results showed, L. regale and L. henryi were clustered together with high TPC, TFC and antioxidant capacity, O and OT hybrid sections were arranged in one group characterized by moderate TPC, TFC and antioxidant properties, and the others, including traditional edible and medical lilies, belonged to a group with low TPC and TFC and weak antioxidant activity (Figure 1). The hierarchical clustering results showed the same change trend as the TPC, TFC and antioxidant capacity, which proved the reliability of our results above. These results suggested that, in addition to the traditional edible and medicinal lily, a large number of other species and cultivated varieties also have high antioxidant capacity, which can be used as important sources of natural antioxidants.

Profiling of Secondary Metabolites
The latest review summarized the information of 183 substances found in the genus Lilium thus far, including 95 saponins, 9 sterols, 23 phenylpropenoid glycerides, 10 alkaloids and 31 flavonoids, which is far from meeting the exploration on the medicinal function of lily. Therefore, it is necessary to identify the compounds contained in the genus Lilium more extensively [1]. Here, considering the results of TPC, TFC and antioxidant capacity, widely targeted metabolomics based on multiple reaction monitoring was used for qualitative and relative quantitative analysis of secondary metabolites in five representatively species (three traditional edible and medicinal lilies L. davidii var. willmottiae (LDW), L. lancifolium (LL), L. brownii var. viridulum (LBV), and two high TPC, TFC and antioxidant capacity species L. henryi (LH) and L. regale (LR)).
Data validation and the results of total substance analysis are shown in Figure 2 or Figure S1. Unsupervised PCA (principal component analysis) was performed by statistics function prcomp within R (www.r-project.org) (Assessed on April 21, 2021). The correlation heat map between samples shows the high correlation within biological repeats and the differences between different species ( Figure S2A). The differences among the five tested lilies were distinguished and verified by 2D PCA (Figure 2A) and 3D PCA ( Figure  S1B). The results showed that the first two principal components (PC1 and PC2) explained

Profiling of Secondary Metabolites
The latest review summarized the information of 183 substances found in the genus Lilium thus far, including 95 saponins, 9 sterols, 23 phenylpropenoid glycerides, 10 alkaloids and 31 flavonoids, which is far from meeting the exploration on the medicinal function of lily. Therefore, it is necessary to identify the compounds contained in the genus Lilium more extensively [1]. Here, considering the results of TPC, TFC and antioxidant capacity, widely targeted metabolomics based on multiple reaction monitoring was used for qualitative and relative quantitative analysis of secondary metabolites in five representatively species (three traditional edible and medicinal lilies L. davidii var. willmottiae (LDW), L. lancifolium (LL), L. brownii var. viridulum (LBV), and two high TPC, TFC and antioxidant capacity species L. henryi (LH) and L. regale (LR)).
Data validation and the results of total substance analysis are shown in Figure 2 or Figure S1. Unsupervised PCA (principal component analysis) was performed by statistics function prcomp within R (www.r-project.org) (Assessed on 21 April 2021). The correlation heat map between samples shows the high correlation within biological repeats and the differences between different species ( Figure S2A). The differences among the five tested lilies were distinguished and verified by 2D PCA (Figure 2A) and 3D PCA ( Figure S1B). The results showed that the first two principal components (PC1 and PC2) explained 72.84% of the overall variance (Figure 2A). Chinese traditional edible lilies LDW, LL and LBV were clearly separated from LH and LR in PC1 (58.32%), and LH and LR, LDW, and LBV and LL were clearly separated in PC2 (14.52%), respectively. LDW and LBV could be separated by PCA2 but the distinction was not obvious. The PCA highlighted a clear metabolic difference among the different Lilium species. In summary, the metabolic variation between LH, LR and LDW, LBV, LL was more obvious, whereas LH and LR, LDW and LL showed more similar metabolite profiling separately.
( Figure 2C). Detailed information on these substances is shown in Supplementary Table  S1. Additionally, 541, 549 374, 305 and 329 compounds were detected in LR, LH, LL, LDW and LBV, respectively, and all the profiled metabolites were identified as nine main types ( Figure 2C or Figure S1C). For instance, flavonoids and phenolic acids are the main secondary metabolites in lily, accounting for more than 50% of the total detected substances ( Figure 2C or Figure S1D). Venn diagram analysis of the substances contained in the different species showed that 219 compounds were common to all five species, with few or no specific metabolites to a single species, indicating some similarity in metabolic composition among five tested species ( Figure 2D). However, compared to the traditional edible lilies LDW, LBV and LL, LR and LH showed more numerous and higher contents of many substances ( Figure 2B,C); LR and LH had 135 and 130 special compounds, respectively ( Figure 2D). The phenylpropane metabolic pathway involves the synthesis of many secondary metabolites, and most flavonoids are synthesized downstream of phenolic acids [39,40]. Since flavonoids and phenolic acids are the major antioxidants in plants [43] and account for the largest proportion of lily constituents ( Figure 2C or Figure S2D), we performed correlation analysis for all flavonoids and phenolic acids detected in the five materials. The results indicated that most of the flavonoid have broad correlations with phenolic The cluster heatmap and class heatmap of the metabolites clearly showed the similarity of components among biological repeats and the difference of components amongst different species ( Figure 2B or Figure S1C). A total of 577 secondary metabolites were target-identified, including 201 flavonoids, 153 phenolic acids, 88 alkaloids, 40 steroids, 30 lignans and coumarins, 23 terpenoids, 9 quinones, 6 tannins and 20 other compounds ( Figure 2C). Detailed information on these substances is shown in Supplementary Table S1. Additionally, 541, 549 374, 305 and 329 compounds were detected in LR, LH, LL, LDW and LBV, respectively, and all the profiled metabolites were identified as nine main types ( Figure 2C or Figure S1C). For instance, flavonoids and phenolic acids are the main secondary metabolites in lily, accounting for more than 50% of the total detected substances ( Figure 2C or Figure S1D). Venn diagram analysis of the substances contained in the different species showed that 219 compounds were common to all five species, with few or no specific metabolites to a single species, indicating some similarity in metabolic composition among five tested species ( Figure 2D). However, compared to the traditional edible lilies LDW, LBV and LL, LR and LH showed more numerous and higher contents of many substances ( Figure 2B,C); LR and LH had 135 and 130 special compounds, respectively ( Figure 2D).
The phenylpropane metabolic pathway involves the synthesis of many secondary metabolites, and most flavonoids are synthesized downstream of phenolic acids [39,40]. Since flavonoids and phenolic acids are the major antioxidants in plants [43] and account for the largest proportion of lily constituents ( Figure 2C or Figure S2D), we performed correlation analysis for all flavonoids and phenolic acids detected in the five materials. The results indicated that most of the flavonoid have broad correlations with phenolic acid, and vice versa ( Figure S1E). This result supports our findings of a significant positive correlation between the TPC and TFC (with a correlation coefficient of 0.977, Table 3) and provides new insights into the synergistic regulation of the biosynthesis and accumulation of phenolic acids and flavonoids in Lilium plants.
Difference analysis was performed for the substances detected from all samples, and compounds with fold change (FC) ≥ 2 or FC ≤ 0.5 and OPLS-DA VIP value ≥ 1 were defined as differentially accumulated metabolites (DAMs). The number of up/downregulated compounds resulting from pairwise comparison of tested species is shown in Figure 3A. The smallest number of differential metabolites was found in the LR vs. LH group with 177 (95 were upregulated and 82 were downregulated). Moreover, LH and LR had more differential metabolites than LDW, LBV, and LL, and among them, the substances whose contents were upregulated accounted for the majority ( Figure 3A). A total of 112 differential metabolites were shared for three traditional edible species (LDW, LBV and LL), and LL was predominant compared with LDW and LBV not only in the total number of metabolites but also in the upregulated DAMs ( Figure 3A,B). The metabolic comparative signatures among three traditional edible and medicinal lilies and the top 20 KEGG enrichment pathways of DAMs are presented in Supplementary Figure S2. Interestingly, LBV was dominant over LDW in the quantitative abundance of substances, but LDW contained more differentially upregulated metabolites than LBV ( Figure 2C or Figure 3A). To explore the metabolite information in LH and LR, which had higher antioxidant activities and abundance and numbers of active compounds, we performed KEGG enrichment analysis of differential metabolites comparing them with those from Chinese traditional medicinal and edible lily LL. The top 20 most enriched pathways for the differential metabolites are shown in Figure 3E,F. Multiple amino acid metabolic pathways and secondary metabolic pathways (isoflavonoid biosynthesis, folate biosynthesis, flavonoid biosynthesis, flavone and flavonol biosynthesis, phenylpropanoid biosynthesis) were among them. These results may reveal the reason why the TPC and TFC were higher in LH and LR as well as their stronger antioxidant activities.
To gain further insight into the types of antioxidants present in lily, we examined the correlation (spearman correlation coefficient) between the different metabolites detected and antioxidant capacity (DPPH, FRAP, CUPRAC), and identified 64 metabolites that were significantly positively correlated (r ≥ 0.9 and p < 0.0001) with antioxidant capacity. In which, 22 phenolic acid (vanillic acid-4-O-glucuronide, α-hydroxycinnamic acid, salicylic acid, 3,6diferuloylsucrose, 1-O-caffeoyl-β-D-glucose, vanillic acid, 2,5-dihydroxybenzoic acid, 3,4dihydroxybenzoic acid, vanillin, 4-hydroxybenzaldehyde, syringaldehyde, ethyl caffeate, caffeic acid, tyrosol, benzamide, trans-5-O-(p-coumaroyl)shikimate, 3,4-dimethoxycinnamic acid, 5-O-p-coumaroylshikimic acid O-glucoside, 2-(formylamino)benzoic acid, 1-O-pcumaroylglycerol, 1-feruloyl-sn-glycerol, 1-O-p-hydroxycinnamoyl-3-O-caffeoylglycerol, 6-O-feruloyl-D-glucose), 13 flavonoids (luteolin-7-O-gentiobioside, kaempferol-3-O-(6 -malonyl)galactoside, kaempferol-3-O-(6 -acetyl)glucoside, dihydroquercetin, hesperetin, pinobanksin, homoeriodictyol, dihydrokaempferol, naringenin chalcone, butin, 2,4,2 ,4tetrahydroxy-3 -prenylchalcone, safflor yellow A), 13 alkaloids (dihydrocaffeoylputrescine, caffeoylspermine, bis(caffeoyl)spermidine, zarzissine, tryptamine, phenethylamine, Noleoylethanolamine, sophoridine, choline, N-glucosyl-p-coumaroylputrescine, N-feruloylagmatine, 9α-hydroxysophoramine, ebeinone) and 5 quinones (4,8-dihydroxy naphthol-1-O-glucoside, emodin, chrysophanol-8-O-glucoside, 6-methylaloe emodin, isoemodin), and 3 terpenoids (isopimaric acid, kaurenoic acid, 3 -O-D-glucosylgentiopicroside) were significantly positively correlated with at least one antioxidant capacity, suggesting that in addition to phenolic acids and flavonoids, some alkaloids and other substances in lily may also be important antioxidants (Figure 4, Table S2). Phenylpropanoid glycerol glucosides, also named as regalosides because they were first isolated from Lilium regale, are typical compounds in genus Lilium [44]. Studies have shown that regalosides have anti-oxidation, bacteriostasis, anti-inflammatory and other A had a higher proportion in all five tested lilies. These findings are of great value for the production of specific regalosides.  Phenylpropanoid glycerol glucosides, also named as regalosides because they were first isolated from Lilium regale, are typical compounds in genus Lilium [44]. Studies have shown that regalosides have anti-oxidation, bacteriostasis, anti-inflammatory and other functions. Murray et al. [45] proved that phenylpropanoid glycerol glucosides can be used as an inhibitor of hepatic glucose production and may have potential to regulate metabolic syndrome and delay type II diabetes. The total glycosides of lily, including six regalosides, can protect cardiomyocytes by protecting mitochondria [46]. Regaloside I could be the main constituent in protecting human dermal fibroblasts from UVA-induced morphological changes [47]. In this study, 13 regalosides were detected in the five tested materials ( Figure 5, Table S3). The results showed that the proportions of different regalosides varied in five lilies. In LR and LH, regaloside A, regaloside E and regaloside F accounted for a relatively high proportion; while, regaloside, regaloside A, regaloside B, regaloside D, regaloside F and regaloside I accounted for a relatively high proportion in LL; regaloside A had a higher proportion in all five tested lilies. These findings are of great value for the production of specific regalosides. Antioxidants 2021, 10, x FOR PEER REVIEW 14 of 17 LDW, LL and LBV are Chinese traditional edible lilies, meanwhile, LL and LBV are used as TCM. At present, there is no specific description of the composition of the substances available for edible and medicinal lilies. The information on the composition of substances in lily that we obtained in the current study can provide a useful reference for the development of standards related to edible and medicinal lilies. The present results suggested that, besides traditional edible lilies, bulbs of the other lilies were also rich in phenolic acids, flavonoids and other bioactive ingredients, and had extensive antioxidant activity. Moreover, these characteristics of LR, LH, and O and OT hybrids were superior to those of the traditional edible lily, indicating the potential applications of a large number of other lily species and hybrids. Although some lilies have high TPC and TFC and strong antioxidant capacity, they also contain some substances that are not found in traditional edible and medicinal lilies. Considering the possible potential toxicity of different ingredients, even though there is almost no obvious acute toxic reaction to the human body, further studies on the constituents contained in different lilies are needed [1].

Conclusions
The TPC, TFC and antioxidant capacity (DPPH, FRAP and CUPRAC) in 22 representative lilies, including wild species and different hybrid sections, were systematically studied for the first time. L. regale and L. henryi have the highest TPC and TFC, leading to the highest antioxidant activity, followed by the O and OT hybrids, and the L, LA, A hybrids and the remaining wild species have the lowest TPC, TFC and antioxidant capacity. A total of 577 secondary metabolites, including 201 flavonoids and 153 phenolic acids, were profiled by extensively targeted metabolomics in five wild species (L. davidii var. willmottiae, L. lancifolium, L. brownii var. viridulum, L. henryi and L. regale). L. regale and L. henryi have great advantages in terms of both the quantity and content of active substances, leading to stronger antioxidant ability. A total of 64 compounds (including 22 phenolic acids, 13 flavonoids and 13 alkaloids) were identified to be significantly related to antioxidant capacity. Different species were all rich in phenylpropanoid glycerol glucosides, but the composition and content are quite different. These data greatly enriched our understanding of the active components of the genus Lilium. The present study reviewed the meaningful bioactive potential of traditional edible and medicinal lilies, as LDW, LL and LBV are Chinese traditional edible lilies, meanwhile, LL and LBV are used as TCM. At present, there is no specific description of the composition of the substances available for edible and medicinal lilies. The information on the composition of substances in lily that we obtained in the current study can provide a useful reference for the development of standards related to edible and medicinal lilies. The present results suggested that, besides traditional edible lilies, bulbs of the other lilies were also rich in phenolic acids, flavonoids and other bioactive ingredients, and had extensive antioxidant activity. Moreover, these characteristics of LR, LH, and O and OT hybrids were superior to those of the traditional edible lily, indicating the potential applications of a large number of other lily species and hybrids. Although some lilies have high TPC and TFC and strong antioxidant capacity, they also contain some substances that are not found in traditional edible and medicinal lilies. Considering the possible potential toxicity of different ingredients, even though there is almost no obvious acute toxic reaction to the human body, further studies on the constituents contained in different lilies are needed [1].

Conclusions
The TPC, TFC and antioxidant capacity (DPPH, FRAP and CUPRAC) in 22 representative lilies, including wild species and different hybrid sections, were systematically studied for the first time. L. regale and L. henryi have the highest TPC and TFC, leading to the highest antioxidant activity, followed by the O and OT hybrids, and the L, LA, A hybrids and the remaining wild species have the lowest TPC, TFC and antioxidant capacity. A total of 577 secondary metabolites, including 201 flavonoids and 153 phenolic acids, were profiled by extensively targeted metabolomics in five wild species (L. davidii var. willmottiae, L. lancifolium, L. brownii var. viridulum, L. henryi and L. regale). L. regale and L. henryi have great advantages in terms of both the quantity and content of active substances, leading to stronger antioxidant ability. A total of 64 compounds (including 22 phenolic acids, 13 flavonoids and 13 alkaloids) were identified to be significantly related to antioxidant capacity. Different species were all rich in phenylpropanoid glycerol glucosides, but the composition and content are quite different. These data greatly enriched our understanding of the active components of the genus Lilium. The present study reviewed the meaningful bioactive potential of traditional edible and medicinal lilies, as well as other wild and hybrid lilies, indicating that many wild and commercialized lilies can be also applied to obtain high added-value products, such as antioxidants, functional ingredients, cosmetic products and nutraceuticals.