High-Resolution Liquid Chromatography–Mass Spectrometry-Based

Jul 19, 2019 - Five Citrus-type crude drugs (40 samples) were classified using liquid chromatography–mass spectrometry (LC-MS)-based metabolomics...
0 downloads 0 Views 2MB Size
Article Cite This: J. Nat. Prod. XXXX, XXX, XXX−XXX

pubs.acs.org/jnp

High-Resolution Liquid Chromatography−Mass Spectrometry-Based Metabolomic Discrimination of Citrus-Type Crude Drugs and Comparison with Nuclear Magnetic Resonance Spectroscopy-Based Metabolomics Takashi Tsujimoto,† Taichi Yoshitomi,† Takuro Maruyama,† Yutaka Yamamoto,‡ Takashi Hakamatsuka,† and Nahoko Uchiyama*,† Downloaded via BUFFALO STATE on July 20, 2019 at 02:39:12 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.



Division of Pharmacognosy, Phytochemistry and Narcotics, National Institute of Health Sciences, 3-25-26, Tonomachi, Kawasaki-ku, Kawasaki, Kanagawa 210-9501, Japan ‡ Tochimoto Tenkaido Co., Ltd., Oniya Kaibara-cho, Tamba, Hyogo 669-3315, Japan S Supporting Information *

ABSTRACT: Five Citrus-type crude drugs (40 samples) were classified using liquid chromatography−mass spectrometry (LCMS)-based metabolomics. The following six flavonoid derivatives were identified as contributors from the loading plots of multivariate analysis: naringin (1), neohesperidin (2), neoeriocitrin (3), narirutin (9), hesperidin (10), and 3,5,6,7,8,3′,4′heptamethoxyflavone (12). Three coumarin derivatives, namely, meranzin (6), meranzin hydrate (7), and meranzin glucoside (8), were also identified as contributors. Furthermore, compared with our previous studies on proton (1H) and 13C NMR spectroscopy-based metabolomics, the present study revealed that the Citrus-type crude drugs were distinguished with the same pattern; however, the contributors differed between the 1H and 13C NMR spectroscopy-based metabolomics. The high dynamic range of NMR spectroscopy provided broad coverage of the metabolomes including the primary and secondary metabolites. However, LC-MS appeared to be superior in detecting secondary metabolites with high sensitivity, some of which occurred in quantities that were undetectable using NMR spectroscopy.

T

metabolomics is increasing as a strategy to ensure crude drugs have authenticity, reliability, stable quality, and acceptable pharmaceutical efficacy parameters. Metabolomics is a statistics-based method for the comprehensive analysis of metabolites in living organisms and is useful for the classification of natural products, including various organic compounds, and the identification of natural organic compounds contributing to the classification.2 There are many reports on metabolomics using data from liquid chromatography−mass spectrometry (LC-MS),3 gas chromatography (GC)-MS,4 and proton (1H) NMR spectroscopy5 analyses, and metabolomics studies have been widely reported in the field of crude drug research.6,7

he quality of crude drugs consisting of natural plant and animal products depends on many factors including source species, localities, growth conditions (wild or cultivated), and the processing method. Currently, determination of the quality of a crude drug involves evaluation of morphological features and fixed quantity of indicator constituent(s) and qualitative analysis using thin-layer chromatography. However, crude drugs comprise many constituents including multiple bioactive natural organic compounds, and it would be difficult to quantify each chemical substance present. In addition, these products need to be evaluated using alternative methods because existing methods do not also determine their bioactivity directly.1 Furthermore, the steady supply of crude drugs with potential high therapeutic use may encourage the wider use of traditional oriental medicinal preparations. Therefore, the importance of © XXXX American Chemical Society and American Society of Pharmacognosy

Received: November 20, 2018

A

DOI: 10.1021/acs.jnatprod.8b00977 J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products

Article

Table 1. Citrus-Type Crude Drugs Utilized in Japan no.

Japanese

English

Latin

botanical origin

1

Kijitsu

immature orange

Aurantii Fructus Immaturus

2

Touhi

bitter orange peel

Aurantii Pericarpium

3

Chimpi

Citrus unshiu peel

Citri Reticulatae Pericarpium

4

Kippi (type 1)

Citrus peel

Tachibana Pericarpium

Citrus aurantium L. var. daidai Makino Citrus natsudaidai Hayata Citrus aurantium L. Citrus hassaku Hort. ex Tanaka Citrus aurantium L. var. daidai Makino Citrus aurantium L. Citrus unshiu Marcow. Citrus reticulata Blanco Citrus tachibana Tanaka Citrus leiocarpa Tanaka Citrus grandis Osbeck Citrus unshiu Marcow.

5

Kippi (type 2)

Citrus peel

Tachibana Pericarpium

6

Seihi

immature Citrus unshiu peel

Citri Unshiu Pericarpium Immaturus

Citrus reticulata Blanco Citrus unshiu Marcow.

part

category

immature fruit

JP1710

mature peel

JP17

mature peel

JP17

mature peel

NonJPS201811

mature peel

Non-JPS2018

immature peel or fruit

Non-JPS2018

Citrus reticulata Blanco

Figure 1. Structures of 21 compounds assigned in Citrus-type crude drugs using liquid chromatography−mass spectrometry (LC-MS) analysis.

Standards for Non-Pharmacopoeial Crude Drugs (NonJPS2018),11 namely, Kippi (type 1, Citrus peel, Tachibana Pericarpium, the pericarp of the ripe fruit of Citrus tachibana Tanaka or Citrus leiocarpa Tanaka or Citrus grandis Osbeck), Kippi (type 2, Citrus peel, Tachibana Pericarpium, the pericarp of the ripe fruit of C. unshiu Marcow. or C. reticulata Blanco), and Seihi (immature C. unshiu peel, Citri Unshiu Pericarpium Immaturus, the immature or crosswise cut fruit or the pericarp of the immature fruit of C. unshiu Marcow. or C. reticulata Blanco). However, the distinction and quality evaluation of the Citrus-type crude drugs is challenging because of their similar morphological characteristics and variety of ingredients. Currently, the following metabolomic studies have been conducted on these crude drug samples. The quality control of Citri Reticulatae Pericarpium using high-performance LC (HPLC) metabolomics12 and discrimination of Pericarpium

A metabolomic study on the differentiation of Citrus-type crude drugs, which have been used widely as medicines and foods for many years,8,9 was carried out. In Japan, six types of Citrus-type crude drugs are regulated, and their attributes are summarized in Table 1. Three Citrus-type crude drugs have been officially recorded in the Japanese Pharmacopeia 17th edition (JP17),10 which are Chimpi (Citrus unshiu peel, Citri Unshiu Pericarpium, the pericarp of the ripe fruit of C. unshiu Marcow. or Citrus reticulata Blanco), Kijitsu (immature orange, Aurantii Fructus Immaturus, the immature or crosswise cut fruit of Citrus aurantium L. var. daidai Makino, C. aurantium L., C. hassaku Hort. ex Tanaka, or Citrus natsudaidai Hayata), and Tohi (bitter orange peel, Aurantii Pericarpium, the pericarp of the ripe fruit of C. aurantium L. or C. aurantium L. var. daidai Makino). Moreover, the following three additional Citrus-type crude drugs have been defined officially in the Japanese B

DOI: 10.1021/acs.jnatprod.8b00977 J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products

Article

Figure 2. (+)-LC-MS chromatograms of selected Citrus-type crude drugs.

and Seihi samples, rutinose-containing flavanone glycosides such as narirutin (9), hesperidin (10), and didymin (11) and the polymethoxyflavones 3,5,6,7,8,3′,4′-heptamethoxyflavone (12) and sinensetin (13) were detected characteristically (Figure 2c−e). The polymethoxyflavone compounds tangeretin (14) and nobiletin (15) were detected in all crude drugs. Differences in the component content of the abovementioned botanical constituents were also reflected in principal component analysis (PCA) plots using (+)-LC-MS. When the sample clusters were in the two-dimensional space of two vectors, principal component 1 (PC1 = 54.2%) and PC2 (17.5%), the Citrus-type crude drugs were classified into four groups with high statistical values of Rx2 (0.981) and Q2 (0.942, Figure 3): (A) Kijitsu, (B) Tohi, (C) Chimpi and Kippi (type 2), and (D) Seihi. The PCA plot positioned group (C) Chimpi and Kippi (type 2) and (D) Seihi in the positive direction (in the first PC [PC1]) with respect to the X axis, and (A) Kijitsu and (B) Tohi were located in the negative direction. All three crude drugs in the positive direction, namely, (C) Chimpi and Kippi (type 2) and (D) Seihi, were prepared from C. unshiu or C. reticulata, and the two negatively positioned crude drugs, i.e., (A) Kijitsu and (B) Tohi, were derived from sources such as C. aurantium and C. natsudaidai. Therefore, the possibility that the species of the botanical source was evaluated in the PC1 was considered. In the loading plot, the flavanone rutinosides narirutin (9) and hesperidin (10) showed a positive contribution to the PC1, corresponding to (C) Chimpi and Kippi (type 2) and (D) Seihi (Figure S1, Supporting Information). In contrast, the flavanone neohesperidosides naringin (1) and neohesperidin (2) and coumarin derivatives meranzin (6) and meranzin hydrate (7) contributed negatively to PC1, corresponding to (A) Kijitsu and (B) Tohi (Figure S1, Supporting Information). Furthermore, naringin (1), neohesperidin (2), and narirutin (9) showed a positive contribution to the PC2, which corresponded to (A) Kijitsu and (D) Seihi. In addition, meranzin (6), meranzin hydrate (7), hesperidin (10), and 3,5,6,7,8,3′,4′-heptamethoxyflavone (12) showed a negative

Citri Reticulatae and Pericarpium Citri Reticulatae Viride13 were reported. HPLC metabolomics were also used for the quality control of Fructus Aurantii Immaturus,14 discrimination of three medicinal materials (Pericarpium Citri Reticulatae, Pericarpium Citri Reticulatae Viride, and Fructus Aurantii),15 and classification of seven Chinese Citrus herbs.16 In addition, there have been reports of the use of GC-MS to discriminate Pericarpium Citri Reticulatae and Pericarpium Citri Reticulatae Viride17,18 and C. reticulata Blanco and C. reticulata “Chachi”.19 Recently, five classifications for Citrustype crude drugs were proposed based on metabolomics using 1 H and 13C NMR spectroscopy.20 In the present study, ingredient analysis of five Citrus-type crude drugs using LC-MS and metabolomics was performed with a provided chromatogram. Moreover, this study reports a comparison of results and output analysis using MS and NMR spectroscopic techniques to determine the differences between these two techniques.



RESULTS AND DISCUSSION (+)-LC-MS and Chemometrics. Forty Citrus-type crude drug samples were analyzed in the present study (Table S1, Supporting Information). Initially, 21 compounds were assigned using LC-MS analysis and identified using direct comparison with standards or assignments of their highresolution MS (HRMS) and tandem MS (MS/MS) spectra. Their structures are presented in Figure 1. In the (+)-LC-MS chromatograms, differences in botanical origin were clearly observed (Figure 2). Characteristic neohesperidose-containing flavonoid glycosides such as naringin (1), neohesperidin (2), neoeriocitrin (3), and rhoifolin (4) were detected in the chromatograms of Kijitsu and Tohi samples derived from C. aurantium and C. natsudaidai (Figure 2a and b). Poncirin (5) was found characteristically in the chromatogram of Kijitsu samples (Figure 2a). Coumarin derivatives such as meranzin (6), meranzin hydrate (7), and meranzin glucoside (8)21,22 were also apparent in chromatograms of Kijitsu and Tohi samples (Figure 2a and b). In contrast, in the chromatograms of C. unshiu and C. reticulata-derived Chimpi, Kippi (type 2), C

DOI: 10.1021/acs.jnatprod.8b00977 J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products

Article

hydroxy-3-methylglutarate substituent, namely, melitidin (16) and brutieridin (17), were observed in the Kijitsu and Tohi chromatograms. Similar to the (+)-LC-MS, poncirin (5) was found characteristically in the Kijitsu chromatogram (Figure 5a). In contrast, the rutinose-containing flavanone glycosides narirutin (9), hesperidin (10), and didymin (11) were detected characteristically in the Chimpi, Kippi (type 2), and Seihi chromatograms (Figure 5e). Some coumarin derivatives such as meranzin glucoside (8) were detected in the Tohi chromatogram (Figure 5b). In contrast, in the Kijitsu chromatogram, bergaptol (18) was detected. Few compound groups without acidic hydrogen, such as polymethoxyflavone, were detected. However, the limonoids nomilinic acid (19) in the Kijitsu chromatogram and limonin (20) and isoobacunoic acid (21) in the Seihi chromatogram, which are characteristic for citrus fruits, were detected (Figure 5a and e). Differences in the constituents of the above-mentioned botanical sources were also reflected in the PCA plot of the (−)-LC-MS. When the sample clusters were in the twodimensional space of two vectors, PC1 (64.6%) and PC2 (13.8%), the Citrus-type crude drugs were separated into four distinct groups with high statistical values of Rx2 (0.996) and Q2 (0.968, Figure 6): (A) Kijitsu, (B) Tohi, (C) Chimpi and Kippi (type 2), and (D) Seihi. From this PCA plot, groups (A) Kijitsu and (B) Tohi were in the positive direction in the PC2 with respect to the X axis, and groups (C) Chimpi and Kippi (type 2) and (D) Seihi were located in the negative direction in the PC1. Both herbal products in the positive direction with respect to the X axis, (A) Kijitsu and (B) Tohi, were from C. aurantium or C. natsudaidai, and the three crude drugs located in the negative direction, i.e., (C) Chimpi and Kippi (type 2) and (D) Seihi, were prepared from C. unshiu or C. reticulata. Therefore, (−)-LC-MS could also be considered suitable for evaluating the species of the botanical source in the PC1 as well as (+)-LC-MS. In the loading plot (Figure S2, Supporting Information), the neohesperidose glucosides naringin (1) and neohesperidin (2) positively contributed to the PC1, corresponding to (A) Kijitsu. In contrast, the rutinose glycosides, narirutin (9) and hesperidin (10), contributed negatively, which corresponded to (C) Chimpi and Kippi (type 2) and (D) Seihi. Flavanone glucosides with a high content in the second main component, namely, naringin (1), neohesperidin (2), narirutin (9), and hesperidin (10), contributed positively, which corresponded to (A) Kijitsu and (C) Chimpi and Kippi (type 2). Moreover, neoeriocitrin (3), meranzin glucoside (8), melitidin (16), and brutieridin (17) negatively contributed in an equivalent manner to (B) Tohi. The results showed there was a difference in the sugar chain of the flavanone glycoside as the first main component [naringin (1) and neohesperidin (2) vs narirutin (9) and hesperidin (10)]. Therefore, similar to (+)-LC-MS, discrimination was considered to have been achieved based on the original plant species present. The LC-MS conditions used in this study provided information on the secondary metabolites present (Figure 7). The flavanone glycosides in the Citrus species contained biased sugar chains that depended on the actual plant species present. Therefore, the results of differentiation using secondary metabolites also reflected the differences in botanical origin. Comparison of Metabolomics between LC-MS and NMR Spectroscopy. The results of the PCA plot and contributors in this study and our previous study using 1H and

Figure 3. Principal component analysis (PCA) score plot of (+)-LCMS chromatograms of Citrus samples.

contribution equivalent to (B) Tohi and (C) Chimpi and Kippi (type 2). The contributors assigned from the loading plot are illustrated in Figure 4.

Figure 4. Contributors assigned from the loading plot of (+)-LC-MS chromatograms.

Thus, differences in the sugar chains of the flavanone glycoside in the PC1 were observed [naringin (1) and neohesperidin (2) vs narirutin (9) and hesperidin (10)]. Therefore, for this study, the discrimination could be regarded as based on botanical origin. (−)-LC-MS and Chemometrics. In the (−)-LC-MS chromatogram (Figure 5), a flavanone glycoside was observed as the main component, which confirmed clearly the difference in the botanical origin of the components. In the chromatograms of the Kijitsu and Tohi samples, flavanone glycosides containing a neohesperidose moiety such as naringin (1), neohesperidin (2), neoeriocitrin (3), and rhoifolin (4) were characteristically detected similar to those in the (+)-LC-MS (Figure 5a and b). Furthermore, flavanone glycosides with a 3D

DOI: 10.1021/acs.jnatprod.8b00977 J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products

Article

Figure 5. (−)-LC-MS chromatograms of Citrus-type crude drugs.

Figure 6. PCA score plot of (−)-LC-MS chromatograms of Citrustype crude drugs. Figure 7. Contributors assigned from the loading plot of (−)-LC-MS chromatograms.

C NMR spectroscopy-based metabolomics20 are summarized in Table S2, Supporting Information. When the statistical values of PCA score plots were compared, the Rx2 value and Q2 value of the LC-MS based metabolomics showed higher values than those of the NMR spectroscopy-based metabolomics (Table 2). This was dependent on the method used to create the data matrix, and since it was performed using bucket integration of a fixed range in NMR spectroscopy, the baseline was included as part of the data. Therefore, since the part that did not contain the component was included in the statistics, the value was 13

Table 2. Comparison of Statistical Values statistic value Rx (+)-LC-MS (−)-LC-MS 1 H NMR20 13 C NMR20

E

2

0.981 0.996 0.969 0.849

Q2 0.942 0.968 0.934 0.647

DOI: 10.1021/acs.jnatprod.8b00977 J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products

Article

Figure 8. Structures of contributors derived from LC-MS and NMR metabolomics of Citrus-type crude drugs.

Next, the differences in the contributing components were confirmed according to the analytical method. In the loading plot using 1H NMR spectroscopy, sucrose (22) and glucose (23) showed a positive contribution to the PC1, whereas naringin (1) showed a negative contribution. In the PC2, naringin (1), sucrose (22), and glucose (23) showed a positive contribution. This result indicated that the location of the Tohi and Chimpi and Kippi (type 2) samples, which contained higher amounts of sugar, corresponded to sucrose (22) and glucose (23), whereas that of Kijitsu corresponded to naringin (1) and that of Seihi was not related to these compounds. This mapping also supported the notion that the sugar or flavanone glucoside content in Seihi was not as much as those in the other four drugs, which was also confirmed from their spectra. Moreover, in the loading plot of the 13C NMR spectroscopic procedure, sucrose (22) and glucose (23) contributed positively to the PC1, and naringin (1) and neohesperidin (2) negatively contributed. In the second main component, naringin (1), sucrose (22), glucose (23), and limonene (24) showed a positive contribution, and narirutin (9) and synephrine (25) showed a negative contribution. As a result, the locations of the Tohi, Chimpi, and Kippi (type 2) samples, which contained sugars in higher amounts, corresponded to sucrose (22) and glucose (23). Furthermore, the location of the Kijitsu sample corresponded to naringin (1) and neohesperidin (2), whereas that of Seihi corresponded to narirutin (9) and synephrine (25). The contributing components identified in the (+)-LC-MS chromatograms and the 1H and 13C NMR spectroscopic procedures are presented in Figure 8. Three flavanone glycosides, naringin (1), neohesperidin (2), and narirutin (9), were observed in both the NMR spectroscopic and LCMS approaches. Compounds 1 and 2 contributed greatly to

reduced. This notion was confirmed by the fact that the statistics were lower using 13C NMR spectroscopy, where statistics were performed with a wider range of bucket integrals. In contrast, in LC-MS-based metabolomics, since the data matrix was created by extracting the peaks, all components used in the statistical processing contained some component. Therefore, the statistical value obtained using LCMS analysis was higher than that of the NMR spectroscopic analysis. First, the pattern of discrimination in the score plots was similar regardless of the instruments, charge (LC-MS), and nuclei (NMR spectroscopy). In the PCA score plot of the 1H NMR spectroscopic analysis, the crude drugs were divided into four groups: (A) Kijistu, (B) Tohi, (C) Chimpi and Kippi (type 2), and (D) Seihi. Chimpi and Kippi (type 2) are chemically regarded as belonging to the same group because they are both derived from the mature peel of C. reticulata and C. unshiu. In this PCA plot, groups (B) Tohi and (C) Chimpi and Kippi (type 2) were positioned in the positive direction along the PC1 axis, and (A) Kijitsu and (D) Seihi were positioned in the negative direction. Therefore, the degree of maturity of the botanical origin could be evaluated in the PC1. The PCA score plot using 13C NMR spectroscopy also afforded four groups: (A) Kijitsu, (B) Tohi, (C) Chimpi and Kippi (type 2), and (D) Seihi, similar to that with 1H NMR spectroscopy. In this PCA plot, groups (B) Tohi and (C) Chimpi and Kippi (type 2) were in the positive direction along the PC1 axis and (A) Kijitsu and (D) Seihi were also located in the negative direction. Therefore, the PC1 axis of the 13C NMR spectroscopic procedure also likely reflected the degree of maturity of the botanical source. F

DOI: 10.1021/acs.jnatprod.8b00977 J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products the discrimination of the Kijitsu sample, whereas narirutin (9) contributed to the discrimination of the Seihi sample. These were effective constituents for discriminating these crude drugs derived from immature fruits and peels. Characteristically, from the 1H and 13C NMR spectroscopic methods, sucrose (22), glucose (23), limonene (24), and synephrine (25) were identified as contributing components. These components were contained in the Tohi, Chimpi, and Kippi (type 2) samples and contributed to the characterization of these drugs. Therefore, NMR spectroscopy-based confirmation of these compounds was considered an effective technique for discriminating crude drugs derived from mature peels. In the (±)-LC-MS methods, hesperidin (10) was observed as a component that strongly characterized the Chimpi and Kippi (type 2) samples. The (−)-LC-MS identified neoeriocitrin (3), a flavanone glycoside, as a contributing component of the Tohi sample. The (+)-LC-MS identified 3,5,6,7,8,3′,4′-heptamethoxyflavone (11) as a contributing component of the Chimpi and Kippi (type 2) samples. Furthermore, coumarin derivatives, i.e., meranzin (6), meranzin hydrate (7), and meranzin glycoside (8), were also identified as contributing components of the Kijitsu sample and, especially, the Tohi sample. The features of each analytical method used for chemometrics are shown in Table S3, Supporting Information. LCMS showed a superior sensitivity to that of 1H and 13C NMR spectroscopy because of its ability to detect trace amounts of components. In terms of the required time for a series of operations, 1H NMR spectroscopy (∼5 min) and LC-MS (∼20 min) were superior to 13C NMR spectroscopy (∼2 h). In contrast, LC-MS separated each component the best because the corresponding signals appeared as a single peak. NMR spectroscopy was inferior to LC-MS because there were multiple signals corresponding to the components. However, since 13C NMR spectroscopy with decoupling showed individual signals as a singlet, it would be superior to 1H NMR spectroscopy, where signal overlap is likely to occur. Finally, considering the range of the detectable components, many primary metabolites could be identified using NMR spectroscopy. Moreover, the extent of the variety of secondary metabolites detectable using NMR was superior to that using LC-MS, which largely depended on the measurement conditions. Particularly with the 13C NMR spectroscopic method, signal separation was superior to 1 H NMR spectroscopic method from the viewpoint of the detection of quaternary carbons, reduced overlapping of signals, and being able to ignore the presence of water. Therefore, 13C NMR spectroscopy would likely provide more structural information about the contributing component than 1H NMR spectroscopy would, so it could be considered superior to 1H NMR spectroscopy based on these observations. Each analytical method had its unique and superior points as summarized in Table S3, Supporting Information. 1H NMR spectroscopy was useful for collecting structural information on components in a short time. 13C NMR spectroscopy would be useful for the collection of detailed component information. LC-MS would be better where secondary metabolite information including minor constituents is required. Moreover, by combining the results of metabolomics using multiple instrument analyses, numerous compounds are handled as index components, so the chemical quality evaluation of crude drugs could be performed in greater detail.



EXPERIMENTAL SECTION



ASSOCIATED CONTENT

Article

General Experimental Procedures. Naringin (1), neohesperidin (2), methanol (HPLC grade), 0.1% formic acid, and acetonitrile (containing 0.1% formic acid) were purchased from Kanto Chemical Co., Ltd. (Tokyo, Japan). Neoeriocitrin (3), rhoifolin (4), poncirin (5), sinensetin (12), and nobiletin (14) were purchased from Extrasynthase SAS (Lyon, France). Meranzin (6) and meranzin hydrate (7) were purchased from Ark Pharm Inc. (Arlington Heights, IL, USA). Narirutin (9) and 3,5,6,7,8,3′,4′-heptamethoxyflavone (11) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Hesperidin (10) was purchased from ACROS (Geel, Belgium). Tangeretin (13) was purchased from Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan). Plant Material. All the crude drugs used in the present studies were obtained as JP17 or Non-JPS2018 grade, and the appropriate details are summarized in Table S1, Supporting Information. Commercially available products for 11 Kijitsu species (JP17 product), 8 Tohi species (JP17 product), and 12 Chimpi species (JP17 product), as well as 3 Kippi species (type 2, Non-JPS2018 product), and 6 Seihi species [Non-JPS2018 products provided by Tochimoto Tenkaido Co., Ltd. (Osaka, Japan)] were used. Each sample was ground into powder using an MM-300 ball mill (Qiagen) at 30 Hz for 2 min. The powder obtained in each case was used for sample preparation. Sample Preparation for LC-MS Analysis. Briefly, 100 mg of each crude drug powder was suspended in 1 mL of methanol, sonicated for 10 min, and centrifuged at 2500g for 10 min. The supernatant was filtered through a membrane filter (0.45 μm; Merck, Kenilworth, NJ, USA) to obtain a sample stock solution of 100 mg/ mL. This stock solution was diluted 20-fold and used as a sample solution at 5 mg/mL for LC-MS measurements. Ultra-HPLC (UHPLC)-MS Analysis. The ultraperformance LC (UPLC) system was interfaced with a Q Exactive hybrid quadrupole− orbitrap mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA). A 2 μL portion of each sample was introduced using full-loop injection into an UltiMate 3000 RS LC system with a photodiode array detector (Thermo Fisher Scientific). Separation was performed using an Acquity UPLC HSS T3 column (100 × 2.1 mm, particle size 1.8 μm, Waters), maintained at 40 °C. The mobile phase consisted of a 0.1% aqueous solution of formic acid (phase A) and acetonitrile containing 0.1% formic acid (phase B) run at a flow rate of 0.4 mL/ min. Gradient elution was performed on the following parameters: 5% B to 40% B in the initial 5 min, 95% B in successive 9 min increments, holding for 3 min, and then returned to the initial ratio in 0.2 min. MS was measured in the positive- and negative-ion electrospray modes. Nitrogen was used as the desolvation gas at 300 °C. The capillary and cone voltages were set to 4000 and 35 V, respectively. Data were collected over the range of m/z 150 to 2000 and were centroided during the acquisition. Multivariate Analysis. All data obtained from the four assays in the two systems in both the positive- and negative-ion modes were processed using Progenesis QI data analysis software (Nonlinear Dynamics, Newcastle upon Tyne, UK). This was used for peak picking, alignment, and normalization to produce peak intensities for retention time and m/z data pairs. The ranges of the automatic peak picking assays were between 1.5 and 14 min. The resultant data matrices were then imported into SIMCA version 14.0 (Umetrics, Umeå, Sweden) for further multivariate statistical analysis with the Pareto scaling.

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jnatprod.8b00977. Additional information (PDF) G

DOI: 10.1021/acs.jnatprod.8b00977 J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products



Article

AUTHOR INFORMATION

Corresponding Author

*Tel: +81-44-270-6521. Fax: +81-44-270-6523. E-mail (N. Uchiyama): [email protected]. ORCID

Takashi Tsujimoto: 0000-0002-7897-0221 Taichi Yoshitomi: 0000-0002-2296-0868 Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS This research was supported by AMED under Grant Number JP18mk0101102j0101. REFERENCES

(1) Länger, R.; Stöger, E.; Kubelka, W.; Helliwell, K. Planta Med. 2018, 84, 350−360. (2) Zhao, Q.; Zhang, J. L.; Li, F. Nat. Prod. Bioprospect. 2018, 8, 321−334. (3) Allwood, J. W.; Goodacre, R. Phytochem. Anal. 2010, 21, 33−47. (4) Almstetter, M. F.; Oefner, P. J.; Dettmer, K. Anal. Bioanal. Chem. 2012, 402, 1993−2013. (5) Kumar, D. Crit. Rev. Anal. Chem. 2016, 46, 400−412. (6) Liu, S.; Liang, Y. Z.; Liu, H. T. J. Chromatogr. B: Anal. Technol. Biomed. Life Sci. 2016, 1015−1016, 82−91. (7) Allard, P. M.; Bisson, J.; Azzollini, A.; Pauli, G. F.; Cordell, G. A.; Wolfender, J. L. Curr. Opin. Biotechnol. 2018, 54, 57−64. (8) Lv, X.; Zhao, S.; Ning, Z.; Zeng, H.; Shu, Y.; Tao, O.; Xiao, C.; Lu, C.; Liu, Y. Chem. Cent. J. 2015, 9, 68. (9) Yu, X.; Sun, S.; Guo, Y.; Liu, Y.; Yang, D.; Li, G.; Lü, S. J. Ethnopharmacol. 2018, 220, 265−282. (10) Japanese Pharmacopoeia, 17th ed.; Ministry of Health, Labour and Welfare: Tokyo, Japan, 2016. (11) Japanese Standards for Non-pharmacopoeial Crude Drugs; Ministry of Health, Labour and Welfare, Tokyo, Japan, 2018. (12) Tistaert, C.; Thierry, L.; Szandrach, A.; Dejaegher, B.; Fan, G.; Frédérich, M.; Heyden, Y. V. Anal. Chim. Acta 2011, 705, 111−122. (13) Yi, L. Z.; Yuan, D. L.; Liang, Y. Z.; Xie, P. S.; Zhao, Y. Anal. Chim. Acta 2007, 588, 207−215. (14) Xu, X.; Jiang, Y.; Liang, Y.; Yi, L.; Cheng, J. Anal. Methods 2010, 2, 2002−2010. (15) Qin, X. M.; Dai, Y. T.; Zhang, L. Z.; Guo, X. Q.; Shao, H. X. Phytochem. Anal. 2009, 20, 307−313. (16) Duan, L.; Guo, L.; Liu, K.; Liu, E. H.; Li, P. J. Chromatogr. A 2014, 1339, 118−127. (17) Wang, Y.; Yi, L.; Liang, Y.; Li, H.; Yuan, D.; Gao, H.; Zeng, M. J. Pharm. Biomed. Anal. 2008, 46, 66−74. (18) Yi, L.; Dong, N.; Liu, S.; Yi, Z.; Zhang, Y. Food Chem. 2015, 186, 192−199. (19) Duan, L.; Guo, L.; Dou, L. L.; Zhou, C. L.; Xu, F. G.; Zheng, G. D.; Li, P.; Liu, E. H. Food Chem. 2016, 212, 123−127. (20) Tsujimoto, T.; Yoshitomi, T.; Maruyama, T.; Yamamoto, Y.; Hakamatsuka, T.; Uchiyama, N. J. Pharm. Biomed. Anal. 2018, 161, 305−312. (21) Wu, M.; Zhang, H.; Zhou, C.; Jia, H.; Ma, Z.; Zou, Z. Molecules 2015, 20, 6925−6940. (22) Mencherini, T.; Campone, L.; Piccinelli, A. L.; Mesa, M. G.; Sánchez, D. M.; Aquino, R. P.; Rastrelli, L. J. Agric. Food Chem. 2013, 61, 1686−1693.

H

DOI: 10.1021/acs.jnatprod.8b00977 J. Nat. Prod. XXXX, XXX, XXX−XXX