Qualitative and quantitative studies on two commercial specifications of Polygonatum odoratum

The rhizoma of Polygonatum odoratum (PO) is used to treat yin injuries of the lung and stomach in traditional Chinese medicine. The chemical constituents of this herb are steroidal saponins, homoisoflavanones, and alkaloids. Xiangyuzhu (XPO) and Guanyuzhu (GPO) are available in the market as two specifications of the commodity. Nonetheless, systematic research on the identification and comparison of chemical constituents of these two commercial specifications is yet lacking. Herein, an integrated method combing ultra-high-performance liquid chromatography-quadruple time-of-flight mass spectrometry (UHPLC-Q-TOF/MS) with ultra-high-performance liquid chromatography-charged aerosol detection (UHPLC-CAD) was employed for the comprehensively qualitative and quantitative analyses of PO. A total of 62 compounds were identified by UHPLC-Q-TOF/MS, among which 13 potential chemical markers were screened out to distinguish two commercial specifications. Subsequently, the absolute determination method for polygodoraside G, polygonatumoside F, and timosaponin H1 was established and validated by UHPLC-CAD. The contents of the three compounds were 13.33–236.24 μg/g, 50.55–545.04 μg/g, and 13.34–407.83 μg/g, respectively. Furthermore, the ratio of timosaponin H1/polygodoraside G could be applied to differentiate the two specifications. Samples with a ratio <2 are considered XPO and >5 are considered GPO. Therefore, the above results provide a valuable means for the quality control of PO.


Introduction
Plants of genus Polygonatum have been extensively used as traditional Chinese medicine. The Chinese Pharmacopoeia records the rhizome of Polygonatum odoratum (PO) as a yinnourishing medicine "Yuzhu", which can be used for the treatment of diabetes and cough Li et al., 2021). PO is widely distributed in the temperate regions of Eurasia (Xia et al., 2022), growing in the forest or on the shady slopes of mountains (Feng et al., 2022). The major chemical constituents of PO are steroidal saponins, homoisoflavanones, alkaloids, and polysaccharides (Zhao et al., 2018;Zhao et al., 2019). In recent years, PO has gained increasing attention (Ning et al., 2018;Wang H J et al., 2018;Pang et al., 2021 PO. Xiangyuzhu (XPO) and Guanyuzhu (GPO) are available in the market as two main commercial specifications with different origins and prices. In the book "Standard compilation of authentic medicinal materials," the appearance of XPO is thick, long, translucent, and light yellow in color, and that from Hunan is considered an authentic medicinal material, which occupies the majority of the Chinese market. GPO is a wild product in Northeast China, Inner Mongolia, and Hebei, and its cost is lower than XPO. Nonetheless, XPO is more frequently used as the raw material of dietary supplements, while GPO is more inclined to medicinal use. Currently, the quality control of PO was the only focus on the polysaccharide content (Jing et al., 2022), whereas some studies have pointed out that the polysaccharide content of GPO is higher than that of XPO (Tian et al., 2014). However, no comparison has been made between the other small molecule chemicals of these two commercial specifications to date. Therefore, it is essential to develop a specific method to distinguish the two commercial specifications in terms of chemical ingredients.
Ultra-high-performance liquid chromatography-quadrupole time-of-flight mass spectrometry (UHPLC-Q-TOF/MS) is a capital analytical tool with good resolution, excellent sensitivity, and strong structural characterization capability and has also been used in the qualitative analysis of Polygonatum genus Sui et al., 2022). Off-line solid phase extraction (SPE) has been used in the pretreatment of complex samples to remove polysaccharides that cannot be retained on C18 columns and enrich targeted components for analysis (Li et al., 2019). The present study aimed to establish a qualitative method by coupling off-line SPE with UHPLC-Q-TOF/MS to characterize PO in positive and negative ion modes. The measured molecular weight plus the fragment ion information obtained by MS/MS could identify the structures of the chemical components (Zhao et al., 2013). Multivariate statistical analyses [for example, principal component analysis (PCA), orthogonal partial least squares discriminant analysis (OPLS-DA), and variable importance in the projection (VIP) plot] were used to identify the potential differential components among XPO and GPO (Pan et al., 2021).
As an aerosol-based universal detector, charged aerosol detector (CAD) can be characterized by high sensitivity, broad dynamic range, less interanalyte response variability, and improved reproducibility, which has gained increasing popularity for LC analysis of organic compounds with poor UV chromophores (Haidar Ahmad et al., 2019;Zheng et al., 2022). The fitting method of regression equation relies on the concentration range. In many cases, the linearity of CAD response is sufficient over the range of interest (Pawellek et al., 2021). Due to its high sensitivity, the limit of detection (LOD) and limit of quantification (LOQ) may be at the nanogram level . Saponins are a class of active components used as key quality control indicators for content determination in many studies Nan et al., 2021). The furostonal saponin polygonatumoside F showed the higher content in the PO samples and may be served as the chemical marker for quality control . In the present study, the other two components polygodoraside G and timosaponin H1 were screened out as major chemical markers in XPO and GPO. In order to assess the quality of PO, UHPLC-CAD method was used, and a content ratio of timosaponin H1/polygodoraside G was proposed to differentiate the two commercial specifications of PO.

Materials and reagents
Acetonitrile (HPLC grade) was purchased from Fisher Scientific Co. (Loughborough, United Kingdom). Distilled water was purchased from Watsons. Formic acid (HPLC grade) was purchased from Acros Co. Ltd. (NJ, United States). The other reagents of analytical grade were also obtained commercially (Beijing, China).

Preparation of samples and standard solutions
An equivalent of 1.0 g accurately weighed fine powder (<40 mesh) of each sample was mixed with 20 mL of 70% ethanol was added. After ultrasonication for 60 min, the solutions were cooled to room temperature, and the weight loss was replaced with 70% aqueous EtOH. Then, the solutions were filtered through a 0.45 µm membrane before quantitative analysis. A volume of 10 mL of the supernatant was concentrated under reduced pressure and diluted to 4 mL with deionized water. A solid-phase extraction cartridge (C18-SPE, 6 mL) was activated with 10 mL of methanol, rinsed with 10 mL of water, and then reconstituted in 4 mL solution to load the sample; first, 10 mL water was used for elution, followed by 4 mL of 95% ethanol. Finally, the ethanol eluate was collected and filtered through a 0.22-μm membrane for qualitative analysis.
The data acquisition mode was MS E . The data were obtained at 50-1500 Da. The source temperature was 100°C, the desolvation temperature was 450°C with desolvation gas flow 850 L/h, leucine enkephaline was used as lock mass, and the capillary voltage was 3 kV. At low CE scan, the cone voltage was 30 V for ESI, and the collision energy was 6 eV (trap) and 4 eV (transfer), while it was 40-60 eV ramp (trap) and 12 eV (transfer) for ESI − and 15-25 eV ramp (trap) and 12 eV (transfer) for ESI + . The instrument was controlled by MassLynx 4.1 software (Waters Corp.).

Validation of UHPLC-CAD analysis
The linearity of regression curves was tested by diluting the mixed stock solution to a series of concentrations of working solutions, and each was subjected to UHPLC analysis. The lower LODs and LOQs were determined by analyzing the serially diluted reference solutions of each compound until the signal-to-noise (S/ N) proportion was about 3 and 10, respectively. The samples were analyzed in six replicates. The stability was tested by analyzing the sample solutions at different time points (0, 2, 4, 6, 8, 12, and 24 h).
The RSD values of peak areas were calculated to examine the precision, repeatability, and stability of the quantitative method. Its accuracy was evaluated by recovery experiments. A specific amount of individual reference standards was spiked into PO sample. Five fortified samples were extracted and analyzed as described above. The recovery value (%) was calculated by the following equation: recovery (%) = 100 × (detected amount − original amount)/spiked amount.

Multivariate analysis
The ESI-MS E centroid data were processed by MarkerLynx version 4.1 (Waters Corp. Manchester, United Kingdom). The analysis included deconvolution, alignment, and data reduction to obtain a list of mass and retention time pairs with the corresponding areas for all the detected peaks from each file in the dataset. The processed data list was then imported by the PCA and OPLS-DA. All the test groups were discriminated in the PCA to investigate whether different groups could be separated. The method parameters were as follows: retention time range, 1-25 min; mass range, 100-1500 Da; mass tolerance, 0.02 Da; 6.00 for noise elimination level, 5% of the base peak intensity of minimum intensity; 0.20 min for RT tolerance. Moreover, isotopic peaks were excluded from analysis. Then, OPLS-DA was carried out to discriminate the ions contributing to the classification of the samples. The results were visualized in a score plot to show group clusters, and a VIP plot showed variables contributing to the classification.

Characterization of chemical compounds in PO
To obtain satisfactory separation and high analytical efficiency, a series of preliminary experiment conditions, including chromatographic column particle size, mobile phase composition, and column temperature, were optimized. Various columns, such as ACQUITY UHPLC BEH C18 (2.1 mm × 100 mm, 1.7 μm) and CORTECS T3 (2.1 mm × 100 mm, 1.6 μm) were compared. Finally, the result was obtained on UHPLC HSS T3 column (100 × 2.1 mm, 1.8 µm). The acetonitrile-water system showed better separation and a more satisfactory peak shape than the methanol-water system. Moreover, 0.1% formic acid was added to the ACN-water system to enhance the peak capacity and improve the peak shape of saponins and flavonoids. The column showed excellent separation performance at 40°C.
Simultaneously, for the identification of the chemical components in PO by UHPLC-Q-TOF/MS, the MS parameters, such as ionization mode, capillary voltage, and different collision energy ranges, were optimized. The herbal extract samples were analyzed in the positive and negative ion modes with the same LC conditions. As shown in Supplementary Table S1, 62 compounds were identified from the PO extract tentatively. The characterization was validated by the data of high-resolution mass spectrometry (Figure 1). The combination of the measured molecular weight with the fragment ion information obtained by collision-induced dissociation (CID) identified the structures and deduced that alkaloids, flavonoids, and steroidal saponins were the major constituents of the components.
Frontiers in Chemistry frontiersin.org

Alkaloid derivatives
Alkaloids are nitrogen-containing basic organic compounds existing in nature. A total of four pairs of alkaloid components were identified in Supplementary Table S1; their ion chromatographic signals were well-displayed in positive and negative ion modes. Each group of alkaloids had cis and trans structures, and the isomers could be determined by referring to the retention time of reference materials. Peak

Flavonoids
The homoisoflavanones of PO are critical bioactive compounds (Wang Y et al., 2018;Xia G. H. et al., 2021;Xia G. H. et al., 2021), and the potential biosynthetic pathway for isoflavonoid formation has been defined in the PO . It can generate characteristic fragment ions by RAD cleavage in the C ring at high energy levels (Ying et al., 2020), and the bond connecting the two sides of the methylene group between the B and C rings is broken to form other characteristic fragment ions.
Peak 32 produced a deprotonated molecular ion at m/z 301.0716 [M-H]in the MS 1 spectrum. In the CID spectrum, the peak produced characteristic fragment ions m/z 179.0370 [M-H-C 7 H 6 O 2 ]and m/z 121.0269 [M-H-C 9 H 8 O 4 ]due to α-cleavage at the f bond (Ren et al., 2021) and the characteristic fragment ion m/z 191.0311 [M-H-C 6 H 6 O 2 ]due to α-cleavage at the g bond. Finally, peak 32 was determined to be disporopsin, and the mass spectrogram in the negative mode and the potential cracking pathway are shown in Figure 3.

Steroidal saponins
Diosgenin and yamogenin are the main parent nucleus types in PO, and the C-3 position is often substituted by 3-4 sugar groups. The sugar group mainly consists of arabinose, glucose, xylose, and rhamnose. The negative ion mode displays deglycosylation fragments, while the positive ion mode presents patent ion characteristic fragments. Thus, these compounds could be identified by the combination of the positive and negative ion modes.

Acetylated steroidal saponins
Acetylated steroidal saponins are primary metabolites that have been reported in the same genus of plants (Ahn et al., 2006). The position of the acetyl group is not fixed and could appear in the C-1 position or the sugar group inside the C-3 position. Such compounds mainly exist in GPO, and it is feasible to classify such compounds by the presence of a neutral loss of 42 Da in the mass spectrum.

Identification of novel compounds
The summary of the chromatographic rules of reference substances could be used to deduce the structure of unknown   Frontiers in Chemistry frontiersin.org 06 compounds to identify novel compounds. For example, the retention time of compounds with the glucose terminal group is less than that of xylose, and peak 7 can be inferred from this rule. The retention time of peak 7 was close to polygodoraside these fragments were the same as those of polygodoraside F, and the retention time is relatively close. Therefore, it is inferred that there is difference between peak 7 and polygodoraside F in only the terminal sugar group; the peak 7 is finally determined to be polygodoraside F-Xyl + Glc. The mass spectrogram in the negative mode of polygodoraside F-Xyl + Glc and polygodoraside F are shown in Supplementary Figure S2.
As a result, 62 compounds were identified from PO tentatively, including 43 steroidal saponins, 11 flavonoids, 8 alkaloids, and 10 possible new components.

Discrimination of PO samples with two commercial specifications by PCA and OPLS-DA analysis
Multivariate statistical analysis was carried out on the metabolite data to discriminate the two commercial specifications of PO. First, the obtained multivariate dataset of 27 batches of samples was analyzed by PCA. The results showed that the samples from different specifications were classified into two categories ( Figure 5A), indicating a great variation in the chemical profile between XPO and GPO.
Frontiers in Chemistry frontiersin.org (peak 50) showed high contents in XPO samples. These chemical markers make it possible to distinguish the two groups of PO samples; then, two conspicuously characteristic markers, polygodoraside G and timosaponin H1, were selected for further quantitative analysis to evaluate the quality of PO from two specifications.

Quantitative analysis of the PO samples by UHPLC-CAD with two commercial specifications
The furostonal saponin polygonatumoside F showed the higher content in all PO samples and thus could serve as the chemical marker for quality control, while the other components polygodoraside G and timosaponin H1 were screened out as chemical markers in XPO and GPO. To further understand the variation in the contents of the main steroidal saponins, a UHPLC-CAD approach was developed for the quantitative analysis of three furostanol saponins (polygodoraside G, polygonatumoside F, and timosaponin H1) in 27 samples with two commercial specifications; the chemical structures showed in Supplementary Figure S3.
The performance of Waters ACQUITY BEH C18 and ACQUITY CORTECS T3 was compared. Polygodoraside G and polygonatumoside F could be separated by Waters ACQUITY ™ HSS T3. Therefore, the ACQUITY ™ HSS T3 column was used for further separation. The column temperature and injection volume were also considered, and finally, the gradient was optimized under the condition of isocratic elution for the separation of the compounds. Subsequently, the established method was validated. The regression equations of the three analytes were calculated in the form of Y = aX 2 + bX + c; X and Y indicated the concentrations of the compound and the corresponding peak area, respectively. A good

FIGURE 6
Contents of the three compounds in 27 batches of PO.

FIGURE 7
Ratio of timosaponin H1/polygodoraside G in 27 batches of PO.
Frontiers in Chemistry frontiersin.org 09 linear correlation of the three analytes was gained (r 2 > 0.999) with a specific concentration range. Due to the high sensitivity of CAD, the LOD and LOQ were in the nanogram level. The results of the method validation are summarized in Table 2. The data indicated that the UHPLC-CAD method possessed good accuracy with recoveries from 97.38% to 106.76%, and all the RSDs of precision, stability, and repeatability were <3%.
Then, the developed UHPLC-CAD method was applied to determine the three targeted compounds in different batches of PO samples (Supplementary Figure S4). Each sample was determined in triplicate, and the contents of targeted analytes are presented in Table 3. The contents of polygodoraside G, polygonatumoside F, and timosaponin H1 ranged from 13.33 to 236.24 μg/g, 50.55-545.04 μg/g, and 13.34-407.83 μg/g, respectively.
Considering the total content of the three compounds (Figure 6), the quality of GPO was stable in the range of 349.19-859.64 μg/g, while the quality of different batches of XPO varied greatly, and the total content was 137.70-1005.55 μg/g. Moreover, polygonatumoside F was the main characteristic compound in the samples of PO. The average content of polygonatumoside F in XPO was 195.49 μg/g and was lower in GPO with 329.35 μg/g, indicating that GPO should be focused upon from the perspective of saponin content. In addition, polygodoraside G was present in XPO with an average content of 113.04 μg/g, and that in GPO was 19.82 μg/g, while timosaponin H1 was mainly present in XPO with an average content of 78.85 μg/g and that in GPO was 291.22 μg/g; the content ratio of the two compounds in different specifications differed markedly. Therefore, it can be concluded that the samples with a ratio <2 of timosaponin H1/polygodoraside G were considered as XPO, and >5 were considered as GPO (Figure 7). In conclusion, the absolute content of the main characteristic components in PO has been investigated. We found that XPO and GPO have significant differences in the quality of medicinal materials. Thus, timosaponin H1/polygodoraside G seems to be a promising indicator to distinguish the two commercial specifications.

Conclusion
In order to understand the chemical composition of PO comprehensively, an effective and sensitive UHPLC-Q-TOF/MS method was developed for the characterization of PO with two commercial specifications. A total of 62 components were identified, of which 13 robust known chemical markers (including 6 flavonoids and 7 steroidal saponins) were screened out to differentiate Xiangyuzhu (XPO) and Guanyuzhu (GPO). A simultaneous determination method for polygodoraside G, polygonatumoside F, and timosaponin H1 was established and validated by UHPLC-CAD before applying it to determine the differences in the contents of the three components in XPO and GPO samples. The ratio of timosaponin H1/polygodoraside G can be used as an indicator to distinguish the two commercial specifications of PO. The samples with a ratio <2 are considered XPO and >5 are considered GPO.
In this study, the qualitative analysis and quantitative analysis of PO were carried out to elucidate the composition. The findings provided a method for distinguishing the two commercial specifications of PO in the market and laid a theoretical foundation for the appropriate utilization of the resources.

Data availability statement
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.