Metabolic Profiling of Chestnut Shell (Castanea crenata) Cultivars Using UPLC-QTOF-MS and Their Antioxidant Capacity

The inner shell of the chestnut (Castanea crenata) has long been used in Asia as a medicinal herb for improving digestion and blood circulation, and treating diarrhea. However, most chestnut shells are now treated as waste materials in industrial peeling processes. In this study, we examined the metabolite variation among major cultivars of C. crenata shells using mass spectrometry. Among five representative cultivars, Okkwang, Porotan, and Ishizuuchi had higher levels of bioactive compounds, such as ellagic acid derivatives, ellagitannins, flavonoids, and gallic acid derivatives. Their antioxidant capacity was positively correlated with their chemical composition. The byproducts (whole shells) from the industrial peeling process were re-evaluated in comparison with the inner shell, a rich source of phenolic compounds. The phenolic acids and flavonoid glucoside derivatives were significantly higher in the whole shells, whereas the levels of flavonoids were higher in the inner shells. In addition, the whole shell extracts significantly reduced cellular reactive oxygen species production compared to the inner shell extracts. This study demonstrated the different biochemical benefits of different C. crenata cultivars through metabolic profiling and suggests that the whole shell could be used as a functional ingredient, as it has the highest levels of bioactive products and antioxidant effects.


Introduction
Chestnut (Castanea crenata Siebold & Zucc., Okkwang, Daebo, Riheiguri, Porotan, and Ishizuuchi) belongs to the Fagaceae family that grows wild throughout Korea and Japan. They are mainly used for food, especially processed foods, and in medicine. In 2019, the FAO (Food and Agriculture Organization of the United Nations) reported that 2,321,780 tons of chestnuts were produced globally [1]. Korea is the second-largest chestnut producer in Asia, producing 50,000 tons of chestnuts annually [2]. About 45% of Korean chestnuts are exported as peeled, canned, and raw chestnuts [3]. The integument of chestnuts consists of a hard hull (outer shell) and thin skin (inner shell), which make up 15-20% of the fresh fruit weight [1,4]. During industrial chestnut peeling processes, the shells are considered solid waste and thrown away. However, they remain a useful source of bioactive compounds [1,5]. The chestnut inner shell, in particular, has been used for many years in traditional medicine in Asian countries and is known for its antioxidant, antiallergic, antidiabetic, and anti-amnesic properties [6][7][8][9]. Several studies have demonstrated the biological activities of the Castanea genus, including antioxidant and anticancer activities [10][11][12]. These positive effects on human health and wellness

Chemicals
Methanol, water, and acetonitrile of LC-MS grade were purchased from Honeywell (St. Muskegon, MI, USA). Formic acid was obtained from Fisher Scientific (Pittsburgh, PA, USA). All authentic standards were bought from Sigma Chemical (St. Louis, MO, USA).

Plant Materials
Chestnuts from five cultivars (Okkwang, Daebo, Riheiguri, Porotan, and Ishizuuchi) of C. crenata were purchased fresh from local chestnut processors and local distributors between September and November 2020, from local markets in Republic of Korea. To ensure that variety-identified chestnuts were collected, the acquisition was made through local contacts. Figure S1 shows the samples photographed before being peeled and dried. The sample information of C. crenata are shown in Table S1. Fresh fruits were maintained under vacuum at −1 • C until extraction.
The whole shells (outer and inner shell) from chestnuts were automatically separated using a chestnut skin removal machine. The inner (which directly covers the kernel) shells were separated from the fruits by hand for the analysis. The samples were prepared using a freeze-drying process under 10 mTorr at −50 • C for 72 h using a FDCF-12003 freeze dryer (Operon, Gimpo-si, Republic of Korea). Dried shells were ground and stored at −80 • C until further use.
The chestnut shell samples were analyzed both in positive and negative ionization modes and the mass range was from 50 to m/z 1100. In positive and negative mode, the ion spray voltage was set at 5500 V and −4500 V, the curtain gas at 30 psi, nebulizing (GS1) at 50 psi; heating (GS2) at 60 psi, and the source temperature at 500 • C. The IDA function was performed, in which the five most intense mass peaks were fragmented. Sciex O.S. software V 1.5 (SCIEX) was employed for data acquisition and processing.
For targeted analysis, 14 selected compounds were quantified using LC-QTRAP/MS-MRM mode.
LC separation was implemented using a Syncronis C18 column (2.1 mm × 100 mm, 3 µm; Thermoscientific) at a flow rate of 0.3 mL/min. A linear gradient from 1 to 50% B in 6 min, from 50 to 99% B in 1 min, held at 99% B for 1 min, from 99 to 1% B in 1 min, and held at 1% B for 5 min was used. The injection volume was 5 µL. The MRM transitions and retention times of the metabolites are summarized in Table S2.

Preparation of the Extracts
Approximately 50 mg of dried shell and inner shell were extracted with 1 mL of ethanol/water (70:30, v/v). The sample was vortexed for 2 min and sonicated for 1 h at room temperature. After centrifugation (12,000 rpm, 20 min, room temperature), the supernatants were filtered through a 0.45 µm PTFE syringe filter (Millipore). The filtered extracts were prepared using a freeze-drying process under 10 mTorr at −50 • C for 72 h using a FDCF-12003 freeze dryer (Operon, Republic of Korea). Dried extracts were stored at −80 • C until further use.

Total Phenolic Content (TPC)
The total phenolic content (TPC) was determined by the modified Folin-Ciocalteu method [22]. The extracts (0.5 mg/mL), diluted with ethanol/water (70:30, v/v), were mixed with 0.2 N Folin-Ciocalteu reagent for 6 min and then reacted with 10% sodium carbonate solution for 1 h. Then, the absorbance was read at 765 nm using a Synergy H1 microplate reader (Bio Tek, Winooski, VT, USA). The phenolic content in the extract of whole shells was shown as milligram of gallic acid equivalents (GAE) per 100 g of dry extract of each sample (mg GAE/100 g dry weight).

2,2-Diphenylpicrylhydrazyl (DPPH) Radical Scavenging Assay
The DPPH free-radical scavenging activity of whole shell extracts (0.1 mg/mL) was estimated as described in [22], with some modifications. DPPH (0.2 mmol/L, 100 µL) solution was added to a 96 well plate and 100 µL of different concentrations of sample solution was added.
The mixture was quenched and left at room temperature for 5 min in a darkroom. The absorbance was then measured at 517 nm with a microplate reader. The DPPH radicalscavenging activity (RSA) was calculated using the following equation: % RSA = [(A control − A sample )/A control ] × 100, where A control and A sample are the absorbance. Results are shown as half-minimal inhibitory concentrations (IC 50 ) calculated by comparing RSA graphs against extract concentrations.

Ferric Reducing Antioxidant Power (FRAP) Assay
A FRAP assay was performed using a MAK369-1KT kit (Sigma, USA). The procedure was carried out in accordance with the manufacturer's instructions. Samples (0.2 mg/mL) were added to a 96 well plate, then assay buffer and working solution were added and incubated at 37 • C for 1 h. Absorbance was measured at a wavelength of 594 nm using a microplate reader.
2.5.5. Determination of Intracellular ROS Scavenging Activity GES1 cells were purchased from the Korea Cell Line bank (KCLB, Seoul, Republic of Korea). The cells were grown at 37 • C in humidified 5% CO 2 , 95% air mixture in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 100 U/mL of penicillin, and 100 µg/mL of streptomycin.
Intracellular ROS levels were measured using a DCF-DA assay [23]. Briefly, GES1 cells were seeded in a black 96 well plate (2 × 10 4 cells/well) and pre-incubated with 10 µM DCF-DA (dissolved in DMSO) for 45 min at 37 • C in the dark. After washing the excess probe using phosphate-buffered saline (PBS), the cells were treated with DW for the control group, 6 mM N-acetylcysteine amide (NAC) (positive control), or extracts (100 µg/mL) of shells in the presence of 2 mM H 2 O 2 for 4 h, and then washed with PBS. Fluorescence was detected using a microplate reader using a λ (ex./em.) = 485/535 nm.

Statistical Methods
To verify differences between more than two groups, a Kruskal-Wallis test followed by Dunn's post hoc test was performed using SPSS 12.0 software (SPSS Inc., Chicago, IL, USA). A heat map and hierarchical cluster analysis (HCA) were generated using MetaboAnalyst Ver. 5.0.

Results and Discussion
In this study, representative chestnut cultivars (Okkwang, Daebo, Riheiguri, Porotan, and Ishizuuchi), originating from Buyeo-gun or Gongju-si, South Chungcheong Province, the main chestnut-producing regions in Korea, were used to determine the metabolite differences among C. crenata cultivars. Their origin and crossbreeding information are shown in Table S1.

The Metabolic Composition of Whole C. crenata Shells
To evaluate the differences in chemical composition between C. crenata cultivars, nontarget metabolite profiling of chestnut extracts was performed using UPLC-QTOF/MS. The metabolites were tentatively identified with the exact mass values of the precursor and the MS/MS spectral data of authentic standards, previously published data, or the MassBank and METLIN databases [24]. The total ion chromatograms (TIC) of the chestnut extracts are given in Figure S2. The characteristic ions corresponding to each compound are showed in Table  The primary compounds identified from the C. crenata shell extracts were ellagitannins. Ellagitannins are hydrolyzable tannins containing polyphenolic compounds and a sugar core [25]. These compounds protect against oxidative stress-related diseases and prevent degenerative diseases, such as cardiovascular and inflammatory diseases and cancers [4,11,13]. Each compound was tentatively characterized through loss of a hexahydroxydiphenyl (HHDP), a sugar unit, and gallic acid moieties. The predominant ion at m/z 301 was formed from the ellagic acid anion produced by the rearrangement of the HHDP moiety [5]. Ellagitannins were also reported in C. sativa Mill. shells in a previous study, and they possessed a wide range of biological activities [5].
In agreement with a previous report, various proanthocyanidins were detected in chestnuts [29]. The highly enriched proanthocyanidins showed significant anti-inflammatory activity. Proanthocyanidins are condensed tannins found in the nuts, bark, fruits, and seeds of various plants [30] and are anti-inflammatory and have beneficial effects on metabolic syndrome, atherosclerosis, and cancer [29]. Proanthocyanidins are formed by catechins and epicatechins [31]. These compounds produced the same fragment ions at m/z 289, 303, and 305 in the MS/MS experiment, suggesting that these compounds were composed of either catechin (C) or gallocatechin (GC) aglycones [32]. The loss of 170 Da or 188 Da (170 + 18) was caused by the release of a gallic acid and a water molecule.
Flavonoids are one of the major classes of phenylpropanoids abundantly found in various foods and beverages [33]. Ellagic acid is a phenolic acid found in various fruits and vegetables, with antioxidant and antiviral properties [35,36]. Ellagic acid derivatives, such as ellagic acid hexose,pentose, and -deoxyhexose, yielded the predominant fragment ion at m/z 301 in negative mode from an ellagic acid moiety. Methylated ellagic acids were generated as methyl, dimethyl, and trimethylellagic acid at m/z 315, 329, and 243, respectively, by the addition of methyl groups to ellagic acid.
Gallic acid and derivatives showed high antioxidant activity and may play protective roles, including anticancer, antiviral, antifungal, and antibacterial activities [37]. Gallic acid displayed [M−H] − ions at m/z 169 and an abundant fragment ion at m/z 125, due to the loss of carboxylic acid. Digalloyl-, trigalloyl-, and tetragalloyl glucose were characterized by a product ion at m/z 465, which could be attributed to the loss of H 2 O and gallic acid moieties from the [M−H] − ion.

Differences in Metabolite Levels Associated with Whole C. crenata Shells
To assess the differences in the chemical compositions of the C. crenata cultivars, principal component analysis, specifically partial least squares discriminant analysis (PLS-DA), was performed on the mass spectra. The PLS-DA score plots derived from the positive ( Figure S3A) and negative ( Figure S3B) modes showed a significant separation in C. crenata cultivars (positive, R 2 = 0.547, Q 2 = 0.371; negative, R 2 = 0.770, Q 2 = 0.711). The five C. crenata cultivars in this study were clustered in both polarity modes. Okkwang, Porotan, and Ishizuuchi were separated from the other cultivars, Daebo and Riheiguri, which formed a separate group. These results showed the close relationship due to their origins. Daebo originated by crossbreeding with "Sangmyeon," a Korean native cultivar, and Riheiguri (Table S1).
In order to visualize the relationship of the differential metabolites identified in the C. crenata cultivars, a heatmap visualization was employed (Figure 1). In Figure 1, the chestnut shell samples were divided into three branches according to their cultivars: Daebo and Riheiguri formed one group, Okkwang formed another group, and Porotan and Ishizuuchi formed a final group. The differential compounds were separated into four groups, which indicated that the identified metabolites could be used to identify the cultivars of chestnuts from their shells. The identified metabolites included in each group are shown in Table S3. To obtain an insight into the behavior of these metabolites in chestnut shells from different cultivars, the relative contents of each group of metabolites are displayed as boxplots in Figure 2. The Group 1 compounds are most abundant in Porotan and Ishizuuchi, which contain high levels of flavonols and methylellagic acids. The Group 2 compounds, which were higher in Okkwang, are mainly ellagitannins and ellagic acid glucosides. Most ellagic acid is present as hydrolyzable ellagitannin in the vacuoles of plant cells, and ellagitannins produce ellagic acid during hydrolysis of the HHDP group [38]. In this study, the level of ellagic acid was positively correlated with ellagitannins. The compounds in Group 3, which are mainly flavonoids and proanthocyanidins, had a high degree of variation within the samples in each group. Amino, organic, and phenolic acids were the dominant compounds in Group 4 and were abundant in Daebo and Riheiguri. To determine the relationship between the metabolite contents and the antioxidant effects, the total phenolic content and the antioxidant capacity of the C. crenata shell extracts were measured. The results of the total phenolic content, DPPH radical scavenging assay, and FRAP assay are shown in Table 2. The antioxidant effects of the C. crenata shell

Metabolite Quantification in Inner and Whole Shells of C. crenata
Chestnut inner shells are a rich source of total phenols and hydrolyzable tannins and flavonoids [10,42], but they are separated from the kernel with the outer shell during industrial peeling processes and are considered waste materials. To separate only the inner shell of the chestnuts is a difficult, time-consuming, and labor-intensive process. To reevaluate the effective use of whole shells, the metabolic composition and antioxidant capacities of the inner and whole shells of chestnuts were compared.
To investigate the contribution of the different parts of the shells to the phytochemical composition and biological activity, chestnut extracts from the inner parts and whole shells were prepared separately, as described in Section 2.2. The metabolic profiling data To determine the relationship between the metabolite contents and the antioxidant effects, the total phenolic content and the antioxidant capacity of the C. crenata shell extracts were measured. The results of the total phenolic content, DPPH radical scavenging assay, and FRAP assay are shown in Table 2. The antioxidant effects of the C. crenata shell extracts determined using the DPPH and FRAP assays were correlated with the total phenolic content. The Okkwang, Porotan, and Riheiguri extracts presented higher antioxidant activities than the Daebo and Riheiguri extracts, similarly to the results obtained from the metabolite profiling of C. crenata. In previous reports, differences in the chemical composition and antioxidant capacities among chestnut (C. sativa Mill. and C. mollissima) cultivars were observed [5,[39][40][41].

Metabolite Quantification in Inner and Whole Shells of C. crenata
Chestnut inner shells are a rich source of total phenols and hydrolyzable tannins and flavonoids [10,42], but they are separated from the kernel with the outer shell during industrial peeling processes and are considered waste materials. To separate only the inner shell of the chestnuts is a difficult, time-consuming, and labor-intensive process. To re-evaluate the effective use of whole shells, the metabolic composition and antioxidant capacities of the inner and whole shells of chestnuts were compared.
To investigate the contribution of the different parts of the shells to the phytochemical composition and biological activity, chestnut extracts from the inner parts and whole shells were prepared separately, as described in Section 2.2. The metabolic profiling data showed that bioactive compounds, such as phenolic acid derivatives, flavonoids, tannins, and proanthocyanidins, had the highest intensities in the Okkwang, Porotan, and Ishizuuchi cultivars ( Figure 1). Thus, these three cultivars were selected to determine the differences between the inner and whole shells of C. crenata.
A targeted analysis was performed on an LC-QTRAP/MS, to determine the quantities of differential metabolites in the inner and whole shells. Targeted analyses of C. crenata shell samples were performed for the phenolic acids, namely caffeic acid, ferulic acid, gallic acid, shikimic acid, chlorogenic acid, and coumaric acid; the flavonoids, namely catechin, quercetin, quercetin glucoside, rutin, apigenin, luteolin, and naringenin; and the amino acids, namely tryptophan. To quantify the metabolites, the phenylalanine-13 C 6 were used as internal standards, improving the quantification's precision and accuracy.
The calibration standards were generated using a solution of combined standards for the appropriate range of each compound. The equations for the calibration curves and the linear regression coefficients for the targeted metabolites are given in Table S4. The calibration curves were constructed using a linear least squares regression analysis of the analyte and internal standard. The calibration curves of the compounds showed correlation coefficients (R 2 ) greater than 0.99 within the given concentration ranges. Figure 3 shows the difference in the levels of significant metabolites in the different parts and cultivars of chestnut shells.
The inner shell extracts contained higher levels of flavonoids, including apigenin, luteolin, naringenin, and quercetin, than the whole shell extracts. In comparison, the whole shell extracts showed predominant concentrations of flavonoid glucosides, such as quercetin glucoside and rutin. The phenolic acids, including caffeic acid, chlorogenic acid, coumaric acid, ferulic acid, and shikimic acid, were more abundant in the whole shell extracts. However, the inner shell extracts contained more gallic acid. The distribution of catechin in the inner and whole shell extracts differed for each cultivar; in particular, tryptophan was barely detected in the inner shell extracts.
In a previous study of the potential anti-gastritis properties of chestnut (C. sativa Mill.), chestnut flour was devoid of any anti-inflammatory activity, while the inner and outer shells retained the ability to inhibit IL-8 secretion. In addition, it has reported that the inhibitory activity on IL-8 secretion was highly similar between the inner and outer shells [29]. We utilized a GES1 cell-based assay to evaluate the ability of the inner and whole shell extracts of C. crenata to reduce the cellular reactive oxygen species (ROS) production stimulated by H 2 O. The intracellular levels of ROS generated after stimulation of the GES1 cells with H 2 O 2 were significantly reduced, by 100 µg/mL, for both the inner and whole shell extracts ( Figure 4). The whole shell extracts more significantly reduced the intracellular levels of ROS than the inner shell extracts. In summary, the whole shells produced by mechanical peeling had as much antioxidant and biological activity as the inner shells.

Conclusions
In the present study, we determined the metabolite profile of chestnut shells different cultivars of C. crenata using UPLC-QTOF/MS. The identification of metabo was confirmed using high-resolution MS/MS analysis, complemented by authentic st ards and the calculation of the corresponding molecular formulas. We found that b tive compounds, such as ellagic acid derivatives, ellagitannins, flavonoids, proantho nidins, and gallic acid derivatives, were more abundant in the Okkwang, Porotan, Ishizuuchi cultivars. Corresponding higher antioxidant activities were observed in t cultivars using the DPPH and FRAP assays. Our study suggests the biochemical ben of C. crenata shells, which can help growers and consumers intentionally choose ches cultivars.
To investigate the contribution of the inner and whole shells to the biological acti several phenolic compounds were quantified using MRM transitions. The concentra of the flavonoids were higher in the inner shells; however, the levels of phenolic acids flavonoid glucoside derivatives were higher in the whole shells. The whole shell ext significantly reduced the intracellular levels of ROS compared to the inner shell extr indicating that the whole shell has as much biological activity as the inner shell and resents a valuable new source of bioactive phytochemicals.  Table S1: Sample list of C. cr analyzed in the present study; Table S2: Retention times and MRM transitions of metabolites q tified from C. crenata shells by LC-QTRAP/MS; Table S3: The list of compounds in heatmap vi zation; Table S4: LC-QTRAP/MS-MRM calibration curve equations and linear correlation c cients (R 2 ) for metabolites quantified in C. crenata shells.

Conclusions
In the present study, we determined the metabolite profile of chestnut shells from different cultivars of C. crenata using UPLC-QTOF/MS. The identification of metabolites was confirmed using high-resolution MS/MS analysis, complemented by authentic standards and the calculation of the corresponding molecular formulas. We found that bioactive compounds, such as ellagic acid derivatives, ellagitannins, flavonoids, proanthocyanidins, and gallic acid derivatives, were more abundant in the Okkwang, Porotan, and Ishizuuchi cultivars. Corresponding higher antioxidant activities were observed in these cultivars using the DPPH and FRAP assays. Our study suggests the biochemical benefits of C. crenata shells, which can help growers and consumers intentionally choose chestnut cultivars.
To investigate the contribution of the inner and whole shells to the biological activity, several phenolic compounds were quantified using MRM transitions. The concentrations of the flavonoids were higher in the inner shells; however, the levels of phenolic acids and flavonoid glucoside derivatives were higher in the whole shells. The whole shell extracts significantly reduced the intracellular levels of ROS compared to the inner shell extracts, indicating that the whole shell has as much biological activity as the inner shell and represents a valuable new source of bioactive phytochemicals.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/biom12121797/s1, Figure S1: Morphological features of five C. crenata cultivars; Figure S2: Total ion chromatograms in (A) ESI-positive and (B) ESI-negative modes from UPLC-QTOF/MS of whole C. crenata shells; Figure S3: PLS-DA score plot derived from UPLC-QTOF/MS spectra of whole C. crenata shells extracts. (A) positive, (B) negative mode; Table S1: Sample list of C. crenata analyzed in the present study; Table S2: Retention times and MRM transitions of metabolites quantified from C. crenata shells by LC-QTRAP/MS; Table S3: The list of compounds in heatmap visualization; Table S4: LC-QTRAP/MS-MRM calibration curve equations and linear correlation coefficients (R 2 ) for metabolites quantified in C. crenata shells.