¹H-NMR-Based Biochemometric Strategy to Identify Transient Receptor Potential Vanilloid 1-Stimulating Compounds from Alpinia officinarum Rhizome

Abstract

This study established a ¹H-NMR-based biochemometric approach for the isolation of biologically active compounds from complex extracts. In both pharmacognosy and natural product chemistry, reliably isolating bioactive compounds typically necessitates repeating time-consuming and laborious isolation and purification steps, presenting a bottleneck in many studies. We applied biochemometric methods to accurately estimate active compounds, thus minimizing the number of assays and isolation steps. The rhizomes of Alpinia officinarum Hance (Zingiberaceae) have been continuously prescribed in traditional Japanese medicine as stomachics and analgesics, despite a limited understanding of the mechanisms underlying these effects. Additionally, transient receptor potential vanilloid subtype 1 (TRPV1) plays a role in modulating nociception, respiratory defense responses, and gastrointestinal protection. Accordingly, ¹H-NMR-based biochemometry was employed to search for TRPV1-active components in A. officinarum rhizome extracts by combining TRPV1 activity intensity with ¹H-NMR data. However, initially, the active component could not be identified because the principal component analysis loading plot primarily displayed only buckets of primary metabolites. Consequently, we applied orthogonal partial least squares to the ¹H-NMR spectra, which allowed us to identify specific spectral bins at 1.66 ppm (aliphatic) and 7.02, 6.98, 6.82, and 6.74–6.58 ppm (aromatic), correlating with TRPV1-stimulating activity. Based on this prediction, diarylheptanoids were swiftly identified, and their potential to activate TRPV1 was confirmed by administering the identified compounds to TRPV1-expressing cells. These findings highlight the potential of chemometric analysis using ¹H-NMR spectroscopy for identifying the chemical classes responsible for the bioactive properties of complex crude drug extracts.

Introduction

Metabolomic approaches have recently gained prominence for the identification of active compounds.^1–5) This high-throughput analytical strategy enables the comprehensive analysis of all metabolites within biological systems.⁶⁾ Metabolomics that incorporates biological activity is referred to as biochemometrics.^6,7) Bioassay-guided fractionation is an effective method for identifying active compounds in crude extracts, particularly when the active components consist of a small number of highly potent compounds that can be isolated by column fractionation.⁸⁾ However, it is difficult to effectively employ bioassay-guided fractionation when the activity of the mixture is low and attributed to the presence of multiple compounds. This process is both time-consuming and laborious, as it typically requires substantial quantities of crude drugs and solvents for extraction and fractionation.⁹⁾ In contrast, a biochemometric approach using machine learning facilitates the identification of compounds using minimal bioassays and overcomes these drawbacks. For metabolomics approaches that utilize analytical instruments, such as LC/MS or NMR, multivariate analysis is considered appropriate for comprehensively analyzing complex component layers containing compounds with various polarities.⁶⁾ As metabolomic datasets are high-dimensional and challenging to analyze, they should be transformed into lower dimensions and visually represented using supervised multivariate data analysis. This technique helps link chemicals to biological activity by integrating metabolite profiles obtained from analytical instruments with the corresponding biological data. In this study, supervised biochemometrics were used to identify the active ingredients of crude drugs.

Alpinia officinarum Hance (Zingiberaceae) is native to Taiwan, Southeast China, and the Indochina Peninsula. Its rhizome is used as a spice because of its distinct aromatic and pungent flavor.^10,11) Additionally, the A. officinarum rhizome has been used in traditional herbal medicine in Japan and China to warm the stomach, alleviate vomiting, disperse cold, and relieve pain.^12,13) The extract of A. officinarum rhizome contains monoterpenes, including 1,8-cineole, galangin, flavonoids, and diarylheptanoids.^14–19) Diarylheptanoids possess anti-inflammatory, antibacterial, antiviral, antioxidant, and anticancer properties.^20–23) Meanwhile, flavonoids prevent gastric ulcers by modulating the activation of toll-like receptor 4 and transient receptor potential vanilloid 1 (TRPV1) signaling pathways.²⁴⁾ However, although A. officinarum rhizome has a pungent taste and is associated with pain relief, no reports to date have detailed the direct activity of any of its constituents on TRPV1.

TRPV1, the capsaicin receptor, is a nonselective cation channel activated by noxious heat (>43°C), protons,^25,26) lipids,²⁷⁾ and various natural products.²⁸⁾ TRPV1 is expressed in peripheral sensory neurons and plays a crucial role in nociception, which is associated with pain perception.^29,30) Furthermore, in addition to inducing pain signaling, TRPV1 also regulates gastrointestinal protection.^31,32) Activation of TRPV1 expression in the epithelial cells of the gastrointestinal tract induces calcitonin gene-related peptide release, which increases gastric blood flow and protects against gastric injury.^33,34) Additionally, postoperative ileus, a transient disruption of coordinated intestinal peristalsis that occurs after abdominal surgery, can be treated by targeting TRPV1.³⁵⁾ The Kampo medicine Daikenchuto has been used clinically to activate TRPV1 to promote intestinal motility, enhance intestinal perfusion, and alleviate inflammatory responses.³⁶⁾ Accordingly, TRPV1-stimulating components are expected to have potential therapeutic value in treating gastrointestinal disorders, including inflammatory pain and oxidative stress.³¹⁾

This study aimed to identify TRPV1-stimulating components in the rhizomes of A. officinarum using a biochemometric approach. Although crude fractions of extracts are commonly employed for metabolite profiling,^37,38) this study minimized column fractionation and used 12 different A. officinarum rhizome MeOH extracts, each with a distinct chemical composition profile, for untargeted analysis across various compound classes. Although LC/MS is a widely used technique in biochemometric studies, it typically requires more time to optimize the measurement conditions than NMR because of the need for optimized ionization and chromatographic parameters. Additionally, data analysis can be time-consuming because of discrepancies in retention times. Consequently, we opted for ¹H-NMR because of its short analysis time, high robustness, and capability to detect compounds with a broad range of polarities.³⁹⁾ Although ¹H-NMR is less sensitive than LC/MS, it provides crucial information regarding the partial chemical structure. To identify the active compounds, ¹H-NMR data and TRPV1-stimulating activity intensities of the MeOH extracts were combined and analyzed using multivariate discriminant analysis. In this study, we developed biochemometrics to rapidly identify active compounds present in crude drugs. Furthermore, 3 TRPV1-stimulating compounds were identified in the MeOH extract of A. officinarum rhizomes by comparing their bioactivities.

Results and Discussion

Measurement of TRPV1-Stimulating Activity to Identify Bioactive Compounds in Model Sets

First, 12 samples were evaluated for their in vitro TRPV1-stimulating activity based on a previous study,⁴⁰⁾ and the results are presented in Table 1. The crude MeOH extract was used without concentrating to minimize the number of steps and reduce the possibility of human error during sample preparation. Cells were treated with a buffer containing 2% of this MeOH extract. TRPV1-stimulating activity was determined by comparing the fluorescence intensity to that of 0.2 µM capsaicin as a ratio. A wide range of intensities was observed among the samples. TRPV1-stimulating activity intensity had a mean of 35.7%, a maximum of 44.1% (Lot. 90G0809), and a minimum of 26.6% (Lot. 23008001). 4-(3-Chloro-2-pyridinyl)-N-[4-(1,1-dimethylethyl)phenyl]-1-piperazinecarboxamide (BCTC), a TRPV1 antagonist, completely inhibited the TRPV1-stimulating activity of all samples (Supplementary Materials).

Table 1. Details of the Crude Drug Samples Used in This Study

No.	Lot.	Collection location	Collection year	TRPV1-stimulating activitya)
1	25033944	Guangdong	Unknown	0.427
2	75T0804	Guangdong	Unknown	0.347
3	21021611	Guangdong	2002	0.364
4	BC1341	Guangdong	2009	0.324
5	23008001	Guangdong	2009	0.441
6	91F0804	Guangdong	2009	0.247
7	B460804	Guangdong	2013	0.427
8	G2K0804	Guangdong	2017	0.334
9	23016001	Guangdong	2017	0.266
10	102809	Guangdong	Unknown	0.369
11	23009001	Guangdong	2017	0.313
12	90G0804	Guangdong	2023	0.429

^a) The intensity of TRPV1-stimulating activity was measured in cells treated with a buffer containing 2% MeOH extract of A. officinarum rhizomes. TRPV1-stimulating activity intensity was determined by comparing the ratio of fluorescence intensity to that of 0.2 µM capsaicin.

Principal Component Analysis of ¹H-NMR Spectra for Identifying Components with TRPV1-Stimulating Activity

Principal component analysis (PCA) was initially used to identify the TRPV1-stimulating activity of the components. PCA is a widely used unsupervised method for multivariate data analysis that aims to reduce the dimensionality of a dataset. Bucketed data derived from the ¹H-NMR spectra of A. officinarum rhizome samples were subjected to PCA. Spectral intensities within the region δ 10.0 to 0.00 ppm were grouped into integrated regions, termed buckets, of equal width (0.04 ppm). The ppm value at the center of each integration region was used as the bucket name. The pretreatment parameters were selected using Pareto scaling (Prt) with orthogonal signal correction (OSC) analysis. Prt is an intermediate between the non-scaling and unit-variance methods. We also used OSC to remove explanatory variables that did not correlate with the objective variables. PCA extracted 3 significant principal components, accounting for 79% of the total variance (PC1 = 50%, PC2 = 19%, and PC3 = 10%), illustrated in a 3D score plot (Fig. 1A), where each point represents an individual sample. The size of each point reflects the fluorescence intensity ratio of A. officinarum rhizome extract compared to that of 0.2 µM capsaicin (see the previous section). In the score plots, the clustering of dots was expected to correlate with the intensity of the stimulus activity; however, the distribution of the dots appeared scattered.

Fig. 1. (A) 3D Score Plot of Principal Component (PC1, PC2, and PC3) Scores

The sphere size reflects the activity intensity. (B) Score Plot of PC1 and PC2 Scores. The circle size reflects the activity intensity. (C) Loading Line Plot for PC1 and PC2 Red: sugars; blue: fatty acids; black: CD₃OD or water.

To further explore the components that contributed to the clustering of A. officinarum rhizome samples, we examined a score plot of the first 2 principal components generated from the 12 samples (Fig. 1B). The loading plot (PC1 and PC2, Fig. 1C) indicates the presence of buckets at 4.06, 4.02, 3.98, 3.82, 3.78, 3.74, 3.70, 3.66, 3.62, 3.46, 3.42, 3.38, and 3.34 on the positive PC1 axis, which are associated with sugar-derived ¹H-NMR signals.^41,42) Meanwhile, on the negative PC1 axis, buckets at 1.62, 1.58, 1.30, 1.26, and 1.22 were identified as derived from fatty acids. Additionally, buckets at 5.10, 5.06, 5.02, 4.58, and 4.54, associated with water, along with a bucket at 3.30, associated with MeOH, contributed to the PC2 axis.⁴³⁾ The PCA results indicated that the distribution of the score plots reflected the ¹H-NMR signals of the primary metabolites present in the MeOH extract of the plant. Notably, a TRPV1-active compound was not detected using PCA. Consequently, orthogonal partial least squares (OPLS) was conducted in the following section to identify the ¹H-NMR signals of the secondary metabolites with TRPV1 activity.

OPLS Model to Identify TRPV1-Activating Components in the A. officinarum Rhizome MeOH Extract

A biochemometric model was employed to investigate the TRPV1-stimulating activity of rhizome components. Specifically, the OPLS model, a supervised multivariate data analysis method, was utilized to analyze data obtained from ¹H-NMR, treating it as the X-variable, whereas the TRPV1-stimulating activity intensity of the A. officinarum rhizome MeOH extract served as the Y-variable. This approach enabled the inference of significant ¹H-NMR signals associated with TRPV1-stimulating activity. The fit parameters of this model were determined using 1 + 1 components, yielding an R²Y value of 0.86 for precision and a Q² value of 0.65 for confidence. Values closer to 1 indicate better validity and predictivity for the model; generally, an R²Y value greater than 0.65 and a Q² value greater than 0.5 are considered acceptable.^44,45) In the OPLS score plot, the dot sizes represent the fluorescence intensity ratio of the A. officinarum rhizome extract relative to 0.2 µM capsaicin, similar to PCA. Samples exhibiting stronger TRPV1-stimulating activity are plotted on the positive X-axis (Fig. 2). An S-plot, which displays the contribution on the X-axis and confidence on the Y-axis, was used to identify the chemical shifts associated with the TRPV1-active components (Fig. 3A). Buckets that were strongly correlated with TRPV1-active components appeared positive on the Y-axis, indicating larger p(corr) values. Buckets with statistically significant confidence levels of p(corr) > 0.8 are highlighted in orange. Buckets at 7.02, 6.98, 6.82, 6.74, 6.70, 6.66, 6.62, 6.58, and 1.66 strongly contributed to the observed activity. However, the bucket at 1.62 was excluded from consideration because it was believed to contain fatty acids. The ¹H-NMR spectrum of the A. officinarum rhizome MeOH extract was used to estimate the active compounds based on their coupling pattern (Fig. 3B). Signals derived from buckets at 7.02, 6.98, 6.82, 6.70, 6.66, 6.62, and 6.58 were identified as aromatic groups, while signals from the bucket at 1.66 were classed as methylene groups. These signals indicated that the bioactive metabolites of A. officinarum rhizomes were likely diarylheptanoids. Further analysis of the ¹H-NMR spectrum revealed coupling patterns in the aromatic region, suggesting tri- and para-substitution of an aromatic ring. The biochemometric system facilitated reliable and rapid analysis that encompassed the entire spectral range. Visualization and analysis of this spectrum were employed to identify the potential correlates of biological activity. This system can detect signals that may correlate with biological activity.

Fig. 2.  OPLS Score Plot of PC1 and PC2 Scores

Fig. 3. (A) S-Plot of the OPLS-DA Model

(B) ¹H-NMR (400 MHz, methanol-d₄) Spectrum of MeOH Extract (No. 1)

Spectroscopic regions (bins) marked in orange correspond to those considered important for bioactivity, based on the S-plot of the OPLS model.

Isolation of Compounds and Comparison of Their Spectra Using S-Plot

Biochemometric-guided purification was conducted to identify the metabolites responsible for the TRPV1-stimulating activity of A. officinarum rhizome, starting with the MeOH extract. The extract underwent various column chromatographic separations based on characteristic signals associated with TRPV1-stimulating activity, including peaks at 7.02, 6.98, 6.82, 6.74–6.58, and 1.66, resulting in the isolation of 3 compounds (1,¹⁵⁾ 2,⁴⁶⁾ and 3¹³⁾; Fig. 4). In addition, 2 analogs were isolated concurrently (4¹⁵⁾ and 5¹⁵⁾). The ¹H-NMR spectra of the isolated compounds and their chemical shifts are presented in Table 2. Furthermore, it was confirmed that the chemical shift of the sp³-secondary carbon in diarylheptanoids appeared at lower magnetic fields owing to the influence of adjacent hydroxy, carbonyl, and aryl groups, which helped differentiate them from the secondary carbons of the fatty acids. The buckets derived from the ¹H-NMR signals of compounds 1–3 were plotted at p(corr) > 0.8 on the S-plot. In contrast, the buckets from the ¹H-NMR signals of compounds 4 and 5 did not meet the criterion of p(corr) > 0.8 (Fig. 5).

Fig. 4. Compounds Isolated from the MeOH Extract of A. officinarum Rhizome

Table 2. Chemical Shift of Isolated Compounds

Compound	Atom	δH (CD3OD)	Bucketa)
1	1, 2, 4, 7	2.87–2.49 (m, 8H)	2.86, 2.82, 2.78, 2.74, 2.70, 2.66, 2.62, 2.58, 2.54, 2.50
	5	4.01 (m, 1H)	4.02
	6	1.71–1.64 (m, 2H)	1.70, 1.66
	2′–6′	7.26–7.10 (m, 5H)	7.26, 7.22, 7.18, 7.14
	2″	6.60 (dd, J = 2.0, 8.1 Hz, 1H)	6.62, 6.58
	3″	6.68 (d, J = 8.1 Hz, 1H)	6.70, 6.66
	6″	6,75 (d, J = 2.0 Hz, 1H)	6.74
	5″–OCH3	3.82 (s, 3H)	3.82
2	1, 2, 4, 7	2.89–2.49 (m, 8H)	2.86, 2.82, 2.78, 2.74, 2.70, 2.66, 2.62, 2.58, 2.54, 2.50
	5	4.02 (m, 1H)	4.06, 4.02, 3.98
	6	1.75–1.65 (m, 2H)	1.74, 1.70, 1.66
	2′–6′, 2″–6″	7.29–7.10 (m, 10H)	7.30, 7.26, 7.22, 7.18, 7.14, 7.10
3	1, 2, 4, 7	2.86–2.47 (m, 8H)	2.86, 2.82, 2.78, 2.74, 2.66, 2.62, 2.58, 2.54, 2.50
	5	4.01 (m, 1H)	4.06, 4.02, 3.98
	6	1.66 (m, 2H)	1.70, 1.66
	2′–6′, 4″	7.26–7.09 (m, 6H)	7.26, 7.22, 7.18, 7.14, 7.10
	2″, 6″	6.98 (d, J = 8.8 Hz, 2H)	6.98
	3″, 5″	6.68 (d, J = 8.8 Hz, 2H)	6.70, 6.66
4	1, 2	2.85 (s, 4H)	2.86
	4	6.07 (dt, J = 16.0, 1.5 Hz, 1H)	6.06, 6.02
	5	6.89 (dt, J = 16.0, 6.8 Hz, 1H)	6.94, 6.90, 6.86
	6	2.51 (m, 2H)	2.54, 2.50
	7	2.75 (t, J = 8.0 Hz, 2H)	2.74, 2.70
	2′–6′, 2″–6″	7.27–7.12 (m, 10H)	7.26, 7.22, 7.18, 7.14
5	1, 2, 4, 7	2.87–2.49 (m, 8H)	2.86, 2.82, 2.78, 2.74, 2.70, 2.66, 2.62, 2.58, 2.54, 2.50
	5	3.65 (m, 1H)	3.70, 3.66
	6	1.83–1.66 (m, 2H)	1.82, 1.78; 1.74, 1.70
	2′-6′, 2″-6″	7.27–7.10 (m, 10H)	7.26, 7.22, 7.18, 7.14, 7.10
	5-OCH3	3.27 (s, 3H)	3.26

^a) Signals with p(corr) > 0.8 or greater on the S-plot of OPLS are indicated in bold.

Fig. 5. S-Plot of the OPLS-DA Model

Signals derived from isolated compounds are highlighted in green.

Evaluation of the TRPV1-Stimulating Activity of the Isolated Compounds

The isolated compounds were evaluated for their TRPV1-stimulating activity, and the results are shown in Fig. 6. Compounds 1, 2, and 3 stimulated TRPV1 activity in a concentration-dependent manner (0.16–100 µM). The fluorescence intensity of these compounds was completely inhibited by BCTC (Supplementary Fig. S2). Among the isolated compounds, compound 1 exhibited the highest TRPV1-stimulating activity, with a 50% effective concentration (EC₅₀) value of 1.11 µM, followed by compounds 2 and 3, with EC₅₀ values of 47.21 and 31.38 µM, respectively. In contrast, compounds 4 and 5 did not exhibit any TRPV1-stimulating activity. This indicates that only compounds derived from buckets with p(corr) > 0.8 on the S-plot exhibited TRPV1-stimulating activity. The classical stimulatory activity of these compounds can be categorized into 3 functional regions: the vanillin portion (A-region); the linking chain (B-region), which includes urea and amide groups; and the lipophilic portion (C-region). Only compound 1 satisfied these criteria and was highly active. It is believed that the activity of compounds 2 and 3 was diminished owing to the loss of the C-5 hydroxy group and the vanilloid structure, respectively. Furthermore, compound 1 demonstrated lower stimulatory ability than the positive control, capsaicin, but was comparable to that of [6]-gingerol and evodiamine. This suggests that it is one of the most active TRPV1-stimulating compounds found in raw materials of the Kampo formulations listed in the Japanese Pharmacopoeia, 18th edition (JP18).⁴⁷⁾

Fig. 6. The Ratio of Fluorescence Intensity Induced by Different Concentrations of Isolated Compounds Compared to That Induced by 0.2 µM Capsaicin

The error bar represents the standard error.

Quantitative Analysis of Compounds 1, 2, and 3 in A. officinarum Rhizome Extract

A quantitative analysis was conducted to assess the contribution of the isolated compounds to the activity of the A. officinarum rhizome MeOH extract. HPLC analysis revealed that the contents of compounds 1, 2, and 3 in 100 mg of A. officinarum rhizome were approximately 0.161–0.415%, 0.027–0.45%, and 0.014–0.029%, respectively (Table 3). The MeOH extract used for the assay contained approximately 1.0 µM of compounds 1, 2, and 3. The concentration of compound 1 in the buffer containing 2% MeOH extract, which was administered to the cells, was approximately 4.9–12.7 µM, suggesting that the diarylheptanoid derivative contributes to the TRPV1 activity of the extract. The activity of A. officinarum rhizome number 9 was weaker, despite having higher contents of diarylheptanoids 1–3 compared to sample numbers 2 and 4. It is believed that there are numerous TRPV1-stimulating active compounds in the rhizome of A. officinarum, in addition to the isolated compounds mentioned. Identifying the additional TRPV1-stimulating active compounds and conducting quantitative experiments should help resolve this discrepancy.

Table 3. Concentration of Compounds in 100 mg of A. officinarum Rhizome (%)

No.	Compound 1	Compound 2	Compound 3
1	0.358	0.0265	0.0202
2	0.200	0.0322	0.0184
3	0.262	0.0380	0.0257
4	0.188	0.0270	0.0140
5	0.415	0.0454	0.0267
6	0.161	0.0340	0.0293
7	0.322	0.0288	0.0248
8	0.292	0.0256	0.0230
9	0.266	0.0373	0.0156
10	0.310	0.0368	0.0246
11	0.215	0.0322	0.0172
12	0.304	0.0274	0.0217

The quantitative analyses were compared with those of another crude drug, “Processed Ginger,” which is anticipated to protect against gastric injury through the TRPV1-stimulating activity of [6]-shogaol.^48,49) According to JP18, Processed Ginger must contain at least 0.1% [6]-shogaol, based on the dried material. Essentially, clinically significant TRPV1 activity is anticipated when [6]-shogaol is present in the crude drug at concentrations exceeding 0.1%. Consequently, given that compound 1 was found in A. officinarum rhizomes at concentrations exceeding 0.1%, it is expected that diarylheptanoids provide protection against gastric injury in traditional Kampo formulations.

Conclusion

In this study, ¹H-NMR-based biochemometrics enabled the efficient identification of TRPV1-active compounds in MeOH extracts of A. officinarum rhizomes. This method successfully predicted the TRPV1-active components using only 1 assay and could identify them through a few isolation processes. We are confident that this tool can be used to discover various pharmacologically active components by modifying the indicator bioactivity and the type of crude extract.

A non-supervised approach using PCA of ¹H-NMR spectra of CD₃OD extracts failed to identify TRPV1-activating compounds in A. officinarum rhizomes owing to interference from signals associated with primary metabolites. Consequently, we analyzed A. officinarum rhizome samples using OPLS and detected ¹H-NMR signals related to TRPV1 activity in the S-plot. Subsequently, employing ¹H-NMR-based biochemometrics as an indicator, we isolated 3 compounds and 2 of their derivatives. The compounds identified using ¹H-NMR-based biochemometrics exhibited TRPV1-activating capacity, whereas the other 2 compounds did not. Although there were no significant structural differences among the 5 isolated compounds, the S-plot from the OPLS revealed subtle ¹H-NMR variations and highlighted only the signals associated with the activity. The series of procedures described above could contribute to the discovery of new pharmacologically active components and the evaluation of the quality of natural products.

This study is the first to report that MeOH extracts of A. officinarum rhizomes activate TRPV1 in vitro. These findings suggest that the intestinal regulatory effect of A. officinarum rhizome extracts may be mediated by TRPV1. The TRPV1-active compounds identified in this study may represent only a small portion of those present in A. officinarum rhizomes. We utilized ¹H-NMR for rapid identification because of its robustness and its few measurement variables. However, more sensitive analytical instrumentation can be employed for the comprehensive identification of TRPV1-active compounds. Additionally, crude drugs are typically administered as hot water extracts; therefore, to confirm the intestinal regulatory effect of TRPV1-active ingredients in A. officinarum rhizomes, it is essential to verify their content in hot water extracts. We believe that further application of biochemometrics can rapidly elucidate the biological activities of crude drugs that are not yet well understood.

Experimental

Plant Material

Details of the A. officinarum rhizomes used in this study are presented in Table 1. The samples were provided by the Kitasato Institute Hospital, Kitasato University (Tokyo, Japan). The A. officinarum rhizome (Lot. G2K0804) used for isolation was purchased from Uchida Wakanyaku, Ltd. (Tokyo, Japan).

Sample Extraction

The A. officinarum rhizome sample used in the TRPV1 assay was prepared by extracting 100 mg of the material with 1 mL of CH₃OH by ultrasonication at room temperature for 1 h. The mixture was then centrifuged at 3000 rpm (Kubota 3740, Tokyo, Japan) for 5 min. After filtering the supernatant through an Ekicrodisc^® 30 mm syringe membrane filter (pore size: 0.45 µm), 4 µL of the filtrate was transferred to a 2 mL test tube and diluted with HBSS) buffer. For ¹H-NMR, samples were prepared by extracting 100 mg of material with 1.0 mL of CD₃OD-d₄ (99.8% D), following the same procedure.

General Equipment and Chemicals

Samples were powdered using a high-speed vibration sample mill (CMT T1-100; CMT, Tokyo, Japan). A Mettler Toledo XS105 dual-range analytical balance was used to prepare the extracts for analysis. The samples were sonicated using an ultrasonic cleaner (US-109; SND, Tokyo, Japan). KUBOTA 3740 (Kubota) and CVE-3000 (EYELA Tokyo Rikakikai Co., Ltd., Tokyo, Japan) centrifuges were used for centrifugation. NMR spectra were obtained using an Agilent Technologies 400-MR spectrometer (Agilent Technologies, Santa Clara, CA, U.S.A.) with ¹H-NMR at 400 MHz in CD₃OD-d₄ (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan). The chemical shifts are expressed in ppm downfield from the internal solvent peaks for CD₃OD-d₄ (δ 3.31 for ¹H-NMR), and J values were measured in Hertz. Semi-preparative HPLC was performed using a Jasco UV 2075 Plus detector coupled with a Jasco LC-Net 11/ADC valve, Jasco PU-2080 pump, and YMC-Actus Triart C₁₈ reversed-phase (RP) analytical column (150 × 20 mm). High-resolution mass spectra were measured on a JEOL (Tokyo, Japan) JMS-T100 LP Mass Spectrometer. Optical rotations ([α]_D) were measured using a JASCO P-2200 polarimeter. Analytical HPLC measurements were performed on a Shimadzu (Kyoto, Japan) LC-20A Prominence instrument equipped with a CBM-20A communication bus module, 2 LC-20AD dual-piston pumps, and a PDA detector. A YMC-Actus Triart C₁₈ RP analytical column (150 × 20 mm) was used. The UV absorption of the compounds was measured using a PDA detector; thus, log ε was not calculated. All IR spectra were measured on a JASCO FT/IR-460 spectrometer.

Cell Culture ​⁴⁰⁾

Human TRPV1/Flp-In293 (hTRPV1) cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 2 mM GlutaMAX, 0.1 mM MEM nonessential amino acid solution (MEM NEAA), and 200 µg/mL hygromycin B at 37°C in 5% CO₂.

TRPV1-Stimulating Activity Measurement

Intracellular Ca²⁺ concentration was measured, and stable human TRPV1/Flp-In293 cells were prepared as previously described.^50,51) The TRPV1-stimulating activity was measured according to a previously described procedure. Human TRPV1/Flp-In293 (hTRPV1) cells (4 × 10⁴ cells/well) were cultured in 100 µL of DMEM supplemented with 10% FBS, 2 mM GlutaMAX, and 0.1 mM MEM NEAA in 96-well, poly-d-lysine black-walled, clear-bottomed plates (Greiner Bio-One, Frickenhausen, Germany) for 24 h. The medium was exchanged, and cells were incubated in HBSS buffer and 20 mM HEPES buffer (pH 7.4) containing FLIPRÒ calcium 6 assay reagent (Molecular Devices, Sunnyvale, CA, U.S.A.) for 1 h at 37°C. Fluorescence was immediately measured using a FlexStation 3 microplate reader (Molecular Devices) (excitation: 485 nm and emission: 525 nm, using a 515-nm cut-off) for 20 s. Subsequently, HBSS buffer containing either the samples or 0.2 µM capsaicin was added, and the fluorescence was immediately measured.

To examine the effect of BCTC, a TRPV1 antagonist, on the sample- or capsaicin-induced increase in intracellular Ca²⁺ concentration, HBSS buffer containing 0–10 nM BCTC was added after the initial fluorescence measurement. After 60 s, HBSS buffer containing either the samples or 0.2 µM capsaicin was added, and the fluorescence was immediately measured. Capsaicin and BCTC were dissolved in MeOH and diluted in HBSS. The final MeOH concentration was adjusted to within 0.1–0.2%.

The data were analyzed using the software, Soft Max Pro^® 5.4 (Molecular Devices). The activities of the purified or authentic compounds were expressed as the EC₅₀, calculated using the GraphPad Prism 5^® software program (GraphPad Software, La Jolla, CA, U.S.A.). All data are expressed as the mean ± standard error of the mean and analyzed by one-way ANOVA. Significant differences between the control and treatment groups were determined using Dunnett’s multiple comparison test or Student’s t-test. All statistical analyses were performed using Prism 5^® software (GraphPad Software, Inc., San Diego, CA, U.S.A.). Statistical significance was determined based on p < 0.05, 0.01, and 0.001.

¹H-NMR Acquisition and ¹H-NMR Multivariate Data Analysis⁴²⁾

The 1D ¹H-NMR spectra were acquired at 298 K using a 45° excitation pulse (Bruker Pulprog: zg). The probe was frequency tuned and impedance matched before each acquisition. For each sample, 64 scans (ns) and 4 dummy scans were recorded using the following parameters: a spectral width of 16 ppm, relaxation delay (D1) of 3.0 s, and receiver gain of 256. The total duration of each ¹H-NMR acquisition was 15 min. Off-line data processing was performed using MNova Lite^® (Mestrelab Research, Santiago de Compostela, Spain) and ALICE2 software (JEOL). The ¹H-NMR spectra were automatically Fourier-transformed using ALICE2 software. The spectra were referenced to CH₃OH at 3.30 ppm. Spectral intensities were reduced to integrated region buckets using ALICE2 for Metabolome version 6 software (JEOL). For CD₃OD, the region between δ 5.0 and 4.6 ppm, corresponding to residual water signals, was excluded. The total integral value of the spectrum was set to 100 to obtain the bucket tables. In the resulting bucket tables, all rows were scaled to the total intensity, and Prt was applied to the columns before performing PCA and OPLS using SIMCA software.^42,43,52,53)

Extraction and Isolation

A. officinarum rhizome (300 g) was extracted 3 times with 3 L of MeOH, yielding 45.2 g (11.2%) of the MeOH extract after solvent evaporation. A portion of the extract (9.0 g) was passed through a Diaion HP-20^® column using H₂O and 20, 40, 60, 80, and 100% MeOH as the eluents to yield the fractions (Fr. aq), 20, 40, 60, 80, and 100M, respectively. Each fraction was analyzed using ¹H-NMR spectroscopy, and Fr. 100M exhibited signals corresponding to the bucket of plots with p(corr) > 0.8. Fr. 100M (11.9 g) was further purified by silica gel chromatography and eluted with a hexane-AcOEt gradient (100 : 0, 95 : 5, 5 : 1, 3 : 1, 1 : 1, 0 : 100, v/v) to obtain 10 fractions (100M-1 to 100M-10). A portion of Fr. 100M-8 (20.3 mg) was subjected to C₁₈ RP column chromatography by semi-preparative HPLC (MeOH-H₂O, 60 : 40, v/v) to yield compounds 1 (4.7 mg) and 2 (2.7 mg). A portion of Fr. 100M-5 (149.6 mg) was also subjected to C₁₈ RP column chromatography by semi-preparative HPLC (MeOH-H₂O, 65 : 35, v/v) to yield compound 3 (62.3 mg). A portion of Fr. 100M-3 (113.2 mg) was processed under the same conditions with MeOH-H₂O 60 : 40 (v/v) to yield compounds 4 (23.11 mg) and 5 (4.4 mg).

5-Hydroxy-7-(4″-hydroxy-3″methoxyphenyl)-1-phenyl-3-heptanone (1)¹⁵⁾

Colorless oil; HRESIMS m/z: 351.1566 [M + Na]⁺ (Calcd for C₂₀H₂₄NaO₄: 351.1572); IR ν_max cm^–1: 2936, 1705, 1515, 1271, 1033, 818, 750, 701; UV λ_max (AcCN) nm: 278, 231, 281; ¹H-NMR (400 MHz, CDCl₃) δ: 7.36–7.13 (5H, m, H2′–H6′), 6.82 (1H, d, J = 7.8 Hz, H-2″), 6.80 (1H, d, J = 1.8 Hz, H-5″), 6.67 (1H, dd, J = 8.0, 1.8 Hz, H-6″), 5.49 (1H, s, 4″-OH), 4.06 (1H, m, H-5), 3.86 (3H, s, 3″-OCH₃), 2.84–2.80 (1H, m, H-13), 3.05 (1H, s, 5-OH), 2.90 (2H, t, J = 7.7 Hz, H-4), 2.79–2.47 (6H, m, H-1, 2, 7), 1.92–1.86 (1H, m, H-12a), 1.64–1.88 (2H, m, H-6); ¹³C-NMR (100 MHz, CDCl₃) δ: 211.2 (C-3), 146.3 (C-3″), 143.7 (C-4″), 140.6 (C-1′), 133.7 (C-1″), 128.5 (C-3′, 5′), 128.3 (C-2′, 6′), 126.2 (C-4′), 120.9 (C-6″), 114.2 (C-5″), 111.0 (C-2″), 66.8 (C-5), 55.8 (3″-OCH₃), 49.3 (C-4), 45.0 (C-2), 38.3 (C-6), 31.4 (C-7), 29.4 (C-1).

5-Methoxy-1,7-diphenylheptan-3-one (2)⁴⁶⁾

Yellow amorphous solid; HRESIMS m/z: 305.1511 [M + Na]⁺ (Calcd for C₁₉H₂₂NaO₂: 305.1518); IR ν_max cm^–1: 2912, 1705, 1454, 1078, 750, 700; UV λ_max (AcCN) nm: 230, 280; ¹H-NMR (400 MHz, CDCl₃) δ: 7.31-7.13 (10H, m, H2′–6′, 2″–6″), 4.05 (1H, m, H-5), 2.90 (2H, t, J = 7.7 Hz, H-4), 2.84–2.48 (6H, m, H-1, 2, 7), 1.64–1.88 (2H, m, H-6); ¹³C-NMR (100 MHz, CDCl₃) δ: 211.2 (C-3), 141.8 (C-1′), 140.6 (C-1″), 128.5 (C-3′, 5′), 128.4 (C-3″, 5″), 128.4 (C-2′, 6′), 128.3 (C-2″, 6″), 126.2 (C-4′), 125.9 (C-4″), 66.8 (C-5), 49.2 (C-4), 45.0 (C-2), 38.0 (C-6), 31.7 (C-7), 29.5 (C-1).

5-Hydroxy-7-(4-hydroxyphenyl)-1-phenyl-3-heptanone (3)¹³⁾

Colorless oil; HRESIMS m/z 321.1455 [M + Na]⁺ (Calcd for C₁₉H₂₂NaO₃: 321.1466); IR ν_max cm^–1 3025, 2828, 1703, 1514, 1230, 838, 701; UV λ_max (AcCN) nm: 216, 230, 276; ¹H-NMR (400 MHz, CDCl₃) δ: 7.13–7.36 (5H, m, H2′–H6′), 7.29 (2H, d, J = 7.0 Hz, H-3′, 5′), 7.20 (1H, m), 7.16 (2H, m), 7.04 (2H, d, J = 8.6 Hz, H-2″, 6″), 6.74 (2H, dd, J = 8.6 Hz, H-3″, 5″), 4.83 (1H, s, 4″-OH), 4.03 (1H, m, H-5), 3.05 (1H, s, 1H, 5-OH), 2.97–2.82 (6H, m, H-1, 2, 7), 2.77 (2H, t, J = 8.2 Hz, H-4), 1.57–1.50 (2H, m, H-6); ¹³C-NMR (100 MHz, CDCl₃) δ: 211.2 (C-3), 153.7 (C-3″), 140.6 (C-1′), 133.7 (C-1″), 129.5 (C-2″, 6″), 128.6 (C-3′, 5′), 128.4 (C-2′, 6′), 126.2 (C-4′), 115.2 (C-3″, 5″), 66.8 (C-5), 49.2 (C-4), 45.0 (C-2), 38.2 (C-6), 30.7 (C-7), 29.4 (C-1).

1,7-Diphenylhept-4-en-3-one (4)¹⁵⁾

Colorless oil; HRESIMS m/z: 287.1413 [M + Na]⁺ (Calcd for C₁₉H₂₀NaO₁: 287.1412); IR ν_max cm^–1: 3026, 2827, 1628, 1498, 1454, 976, 747, 699; UV λ_max (AcCN) nm: 216, 279; ¹H-NMR (400 MHz, CDCl₃) δ: 7.33–7.15 (10H, m, H2′–6′, 2″–6″), 6.85 (1H, dt, J = 16.0, 7.4 Hz, H-5), 6.11 (1H, dt, J = 16.0, 1.5 Hz, H-4), 2.90 (2H, t, J = 7.7 Hz, H-4), 2.84–2.48 (6H, m, H-1, 2, 7), 1.92–1.86 (1H, m, H-12a), 1.64–1.88 (2H, m, H-6); ¹³C-NMR (100 MHz, CDCl₃) δ: 199.5 (C-3), 146.1 (C-5), 141.2 (C-1′), 140.5 (C-1″), 130.5 (C-4), 128.5 (C-3′, 3″, 5′, 5″), 128.3 (C-2′, 2′, 6′, 6″), 126.2 (C-4′), 126.0 (C-4″), 120.9 (C-6″), 41.7 (C-2), 34.4 (C-6), 34.1 (C-7), 30.1 (C-1).

5-Methoxy-7-(4″-hydroxy-3″-methoxyphenyl)-1-phenyl-3-heptanone (5)¹⁵⁾

Colorless oil; HRESIMS m/z: 319.1667 [M + Na]⁺ (Calcd for C₂₀H₂₄NaO₂: 319.1674); IR ν_max cm^–1: 2828, 1715, 1098, 748, 700; UV λ_max (AcCN) nm: 214, 225, 277; ¹H-NMR (400 MHz, CDCl₃) δ: 7.29–7.12 (10H, m, H2′–6′, 2″–6″), 3.69 (1H, m, H-5), 3.29 (3H, s, 5-OCH₃), 2.88 (2H, t, J = 8.3 Hz, H-4), 2.77–2.56 (6H, m, H-1, 2, 7), 1.88–1.64 (2H, m, H-6); ¹³C-NMR (100 MHz, CDCl₃) δ: 208.5 (C-3), 141.7 (C-1′), 141.0 (C-1″), 128.4 (C-3′, 5′), 128.4 (C-3″, 5″), 128.3 (C-2′, 2″, 6′, 6″), 126.1 (C-4′), 125.8 (C-4″), 76.6 (C-5), 57.0 (5-OCH₃), 47.4 (C-4), 45.4 (C-2), 35.6 (C-6), 31.3 (C-7), 29.5 (C-1).

Determination of Isolated Compounds by HPLC

To determine the concentrations of the isolated compounds, analytical HPLC was performed using a Shimadzu LC-20A Prominence instrument equipped with a CBM-20A communication bus module, 2 LC-20AD dual-piston pumps, and a PDA detector. A YMC-Pack Pro C₁₈ RP analytical column (150 × 4.6 mm) was used. The column temperature was set at 40°C, and the detection wavelength was set at 210 nm. The mobile phase consisted of solvents A (Milli-Q^®) and B (HPLC-grade acetonitrile) at a flow rate of 1.0 mL/min. The gradient elution was programmed as follows: 0–100% B from 0 to 35 min, 100% B maintained until 45 min, and returned to 0% B after 46 min, with the gradient elution re-equilibration time set at 10 min. The flow rate was set at 1.0 mL/min, and the injection volume was 10 µL. The UV spectra were recorded in the wavelength range 190–800 nm. Ten milligrams of the sample (n = 3) was extracted by ultrasound with 1 mL of MeOH for 60 min, followed by centrifugation and filtration. A 10 µL sample solution was injected into the HPLC system.

Acknowledgments

The authors thank Dr. Tadashi Adachi (Mitsubishi Chemical Corp.) for providing Diaion HP-20^® and for helpful discussions.

Author Contributions

Y. Kobayashi, H. J., and T. T.-K. initiated and directed the project. Y. Kobayashi, S. N., and T. S. designed the study. Y. Koizumi. and T.H. collected and stored the plant samples. S. O. provided hTRPV1/Flp-In293 cells. T. S. and T. K. designed the experiments. T. K., R. S., and A. H. conducted the experiments and analyzed and interpreted the results. K. N. provided technical support for the NMR experiments. T. K., T. S., S. N., and Y. Kobayashi wrote the manuscript. All authors have read and approved the final version of the manuscript.

Conflict of Interest

The authors declare no conflict of interest.

Supplementary Materials

This article contains supplementary materials.

References

• 1)  Stewart A. K., Ravindra R., Van Wagoner R. M., Wright J. L. C., J. Nat. Prod., 81, 349–355 (2018).
• 2)  Tawfike A., Abbott G., Young L., Edrada-Ebel R., Planta Med., 84, 182–190 (2018).
• 3)  Inui T., Wang Y., Pro S. M., Franzblau S. G., Pauli G. F., Fitoterapia, 83, 1218–1225 (2012).
• 4)  Li P., Anandhi Senthilkumar H., Figueroa M., Wu S. -B., Fata J. E., Kennelly E. J., Long C., J. Nat. Prod., 79, 1619–1627 (2016).
• 5)  Fuchino H., Anjiki N., Murase S., Matsuo H., Hayashi S., Kawahara N., Yoshimatsu K., Chem. Pharm. Bull., 70, 848–858 (2022).
• 6)  Britton E. R., Kellogg J. J., Kvalheim O. M., Cech N. B., J. Nat. Prod., 81, 484–493 (2018).
• 7)  Caesar L. K., Kellogg J. J., Kvalheim O. M., Cech N. B., J. Nat. Prod., 82, 469–484 (2019).
• 8)  Junio H. A., Sy-Cordero A. A., Ettefagh K. A., Burns J. T., Micko K. T., Graf T. N., Richter S. J., Cannon R. E., Oberlies N. H., Cech N. B., J. Nat. Prod., 74, 1621–1629 (2011).
• 9)  Chan K. -M., Yue G. G. -L., Li P., Wong E. C. -W., Lee J. K. -M., Kennelly E. J., Lau C. B. -S., J. Chromatogr. A, 1487, 162–167 (2017).
• 10)  Ivanović M., Makoter K., Islamčević Razboršek M., Plants, 10, 501 (2021).
• 11)  Wongphan P., Nampanya P., Chakpha W., Promhuad K., Laorenza Y., Leelaphiwat P., Bumbudsanpharoke N., Sodsai J., Lorenzo J. M., Harnkarnsujarit N., Food Packag. Shelf Life, 37, 101077 (2023).
• 12)  Liu H., Wang X., Shi Q., Li L., Zhang Q., Wu Z. -L., Huang X. -J., Zhang Q. -W., Ye W. -C., Wang Y., Shi L., ACS Omega, 5, 10167–10175 (2020).
• 13)  Shin D., Kinoshita K., Koyama K., Takahashi K., J. Nat. Prod., 65, 1315–1318 (2002).
• 14)  Sun Y., Tabata K., Matsubara H., Kitanaka S., Suzuki T., Yasukawa K., Planta Med., 74, 427–431 (2008).
• 15)  Liu Z., Sang S., Hartman T. G., Ho C. T., Rosen R. T., Phytochem. Anal., 16, 252–256 (2005).
• 16)  Ly T. N., Shimoyamada M., Kato K., Yamauchi R., J. Agric. Food Chem., 51, 4924–4929 (2003).
• 17)  Ly T. N., Yamauchi R., Shimoyamada M., Kato K., J. Agric. Food Chem., 50, 4919–4924 (2002).
• 18)  Itokawa H., Morita H., Midorikawa I., Aiyama R., Morita M., Chem. Pharm. Bull., 33, 4889–4893 (1985).
• 19)  Matsuda H., Ando S., Kato T., Morikawa T., Yoshikawa M., Bioorg. Med. Chem., 14, 138–142 (2006).
• 20)  Subramanian K., Selvakkumar C., Vinaykumar K. S., Goswami N., Meenakshisundaram S., Balakrishnan A., Lakshmi B. S., Int. J. Antimicrob. Agents, 33, 244–250 (2009).
• 21)  Honmore V. S., Kandhare A. D., Kadam P. P., Khedkar V. M., Sarkar D., Bodhankar S. L., Zanwar A. A., Rojatkar S. R., Natu A. D., Int. Immunopharmacol., 33, 8–17 (2016).
• 22)  Wu D., Ge D., Dai Y., Chen Y., Fu Q., Jin Y., J. Sep. Sci., 46, e2300156 (2023).
• 23)  Bitari A., Oualdi I., Touzani R., Elachouri M., Legssyer A., Mater. Today Proc., 72, 3869–3874 (2023).
• 24)  Lin K., Deng T., Qu H., Ou H., Huang Q., Gao B., Li X., Wei N., Pharm. Biol., 61, 50–60 (2023).
• 25)  Caterina M. J., Schumacher M. A., Tominaga M., Rosen T. A., Levine J. D., Julius D., Nature (London), 389, 816–824 (1997).
• 26)  Tominaga M., Caterina M. J., Malmberg A. B., Rosen T. A., Gilbert H., Skinner K., Raumann B. E., Basbaum A. I., Julius D., Neuron, 21, 531–543 (1998).
• 27)  Huang S. M., Bisogno T., Trevisani M., Al-Hayani A., De Petrocellis L., Fezza F., Tognetto M., Petros T. J., Krey J. F., Chu C. J., Miller J. D., Davies S. N., Geppetti P., Walker J. M., Di Marzo V., Proc. Natl. Acad. Sci. U.S.A., 99, 8400–8405 (2002).
• 28)  Liang Q., Wang J. W., Bai Y. R., Li R. L., Wu C. J., Peng W., “Targeting TRPV1 and TRPA1: A feasible strategy for natural herbal medicines to combat postoperative ileus,” Academic Press, 2023.
• 29)  Cao E., Liao M., Cheng Y., Julius D., Nature (London), 504, 113–118 (2013).
• 30)  Ye W., Lui S. T., Zhao Q., Wong Y. M., Cheng A., Sung H. H. Y., Williams I. D., Qian P. Y., Huang P., Int. J. Biol. Macromol., 253, 127136 (2023).
• 31)  Peng J., Li Y. -J., Eur. J. Pharmacol., 627, 1–7 (2010).
• 32)  Horie S., Yamamoto H., Michael G. J., Uchida M., Belai A., Watanabe K., Priestley J. V., Murayama T., Scand. J. Gastroenterol., 39, 303–312 (2004).
• 33)  Nishihara K., Nozawa Y., Nakano M., Ajioka H., Matsuura N., Br. J. Pharmacol., 135, 1487–1494 (2002).
• 34)  Mózsik G., Szolcsányi J., Dömötör A., Inflammopharmacology, 15, 232–245 (2007).
• 35)  Bragg D., El-Sharkawy A. M., Psaltis E., Maxwell-Armstrong C. A., Lobo D. N., “Postoperative ileus: Recent developments in pathophysiology and management,“ Churchill Livingstone, New York, 2015.
• 36)  Tokita Y., Yamamoto M., Satoh K., Nishiyama M., Iizuka S., Imamura S., Kase Y., J. Pharmacol. Sci., 115, 75–83 (2011).
• 37)  Casanova L. M., Rodrigues L. M., de Aguiar P. F., Tinoco L. W., J. Nat. Prod., 83, 243–254 (2020).
• 38)  Cui W. -L., Guo D. -X., Wang N., Wang Z. -F., Ji J. -B., Wang X., Yang C. -G., Lin Y. -Q., Wang S. -Q., Front. Chem., 10, 1069591 (2023).
• 39)  Ueno Y., Suzuki R., Kitamura M., Chem. Pharm. Bull., 70, 589–862 (2022).
• 40)  Nakamori S., Takahashi J., Hyuga S., Tanaka-Kagawa T., Jinno H., Hyuga M., Hakamatsuka T., Odaguchi H., Goda Y., Hanawa T., Kobayashi Y., J. Nat. Med., 71, 105–113 (2017).
• 41)  Yoshitomi T., Wakana D., Uchiyama N., Tsujimoto T., Kawano N., Yokokura T., Yamamoto Y., Fuchino H., Hakamatsuka T., Komatsu K., Kawahara N., Maruyama T., J. Nat. Med., 74, 65–75 (2020).
• 42)  Shirahata T., Ishikawa H., Kudo T., Takada Y., Hoshino A., Taga Y., Minakuchi Y., Hasegawa T., Horiguchi R., Hirayama T., Konishi T., Takemoto H., Sato N., Aragane M., Oikawa T., Odaguchi H., Hanawa T., Kodaira E., Fukuda T., Kobayashi Y., J. Nat. Med., 75, 475–488 (2021).
• 43)  Kanai T., Shirahata T., Nakamori S., Koizumi Y., Kodaira E., Sato N., Fuchino H., Kawano N., Kawahara N., Hoshino T., Yoshimatsu K., Kobayashi Y., J. Nat. Med., 78, 952–969 (2024).
• 44)  Triba M. N., Le Moyec L., Amathieu R., Goossens C., Bouchemal N., Nahon P., Rutledge D. N., Savarin P., Mol. Biosyst., 11, 13–19 (2015).
• 45)  Szymańska E., Saccenti E., Smilde A. K., Westerhuis J. A., Metabolomics, 8(S1), 3–16 (2012).
• 46)  Pulkkinen J. T., Honkakoski P., Peräkylä M., Berczi I., Laatikainen R., J. Med. Chem., 51, 3562–3571 (2008).
• 47)  Japanese Pharmacopoeia Commission, Ministry of Health, Labour and Welfare, (2021), “The Japanese Pharmacopoeia 18th Edition, English version,” Japanese Government.
• 48)  Yoshikubo M., Tashima K., Raimura M., Watanabe Y., Noma Y., Horie S., Planta Med., 9, e72–e79 (2022).
• 49)  Satoh K., Kase Y., Hayakawa T., Murata P., Ishige A., Sasaki H., Biol. Pharm. Bull., 24, 1122–1126 (2001).
• 50)  Ohkawara S., Tanaka-Kagawa T., Furukawa Y., Nishimura T., Jinno H., Biol. Pharm. Bull., 33, 1434–1437 (2010).
• 51)  Ohkawara S., Tanaka-Kagawa T., Furukawa Y., Jinno H., J. Toxicol. Sci., 37, 831–835 (2012).
• 52)  Shirahata T., Kanazawa A., Uematsu M., Fuchino H., Kawano N., Kawahara N., Yoshimatsu K., Hanawa T., Odaguchi H., Kobayashi Y., Chem. Pharm. Bull., 70, c21–c01102 (2022).
• 53)  Kitazoe T., Usui C., Kodaira E., Maruyama T., Kawano N., Fuchino H., Yamamoto K., Kitano Y., Kawahara N., Yoshimatsu K., Shirahata T., Kobayashi Y., J. Nat. Med., 78, 296–311 (2024).