Aconitum pendulum subsp. based, N. Busch, N. Busch

Wang, Jun-Jie, Lou, Hua-Yong, Liu, Ying, Han, Hong-Ping, Ma, Feng-Wei, Pan, Wei-Dong & Chen, Zhi, 2022, Profiling alkaloids in Aconitum pendulum N. Busch collected from different elevations of Qinghai province using widely targeted metabolomics, Phytochemistry (113047) 195, pp. 1-10 : 3-7

2.2. Metabolite profiling of A. pendulum based on UPLC-MS/MS

Using ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS), we profiled the metabolites and selected distinct markers for A. pendulum extracts from different locations. A total of 80 chemical compounds were identified ( Table S1 ), including 58 diterpenoid alkaloids ( Hu et al., 2019; Li et al., 2020), 5 apomorphline alkaloids ( Liu et al., 2016), 2 pyrrole alkaloids ( Zou et al., 2000), 1 imidazole alkaloid ( Liu et al., 2014), 1 steroid alkaloid ( Yang and Duan, 2012), 1 guanidine alkaloid ( Bouaicha et al., 1994), 3 matrine alkaloids ( Tan et al., 2015), 3 promorphinane alkaloids ( Gan et al., 2009), 3 isoquinoline alkaloids ( Shi et al., 2017), 3 aromatic alkylamine alkaloids, and 1 amino acid and its derivatives. Some alkaloid chemical compounds, including azitine, napelline, aconine, aconitine, spicatine A, and polyschistine A, were reported in our previous study ( Wang et al., 2021).

We divided the 80 chemical compounds into 11 types of alkaloids and compared their peak areas of the different A. pendulum samples ( Fig. 1 ). We detected diterpenoid alkaloids, apomorphline alkaloids, guanidine alkaloids, promorphinane alkaloids, acids and their derivatives, isoquinoline alkaloids, and aromatic alkylamine alkaloids in the samples from every sampling site. The peak area of the diterpenoid alkaloids was considerably larger than other types of alkaloids. However, the pyrrole alkaloids, imidazole alkaloids, steroid alkaloids, and matrine alkaloids were associated with specific sampling areas, the contents thereof also differed. Results revealed that these 80 nitrogen-containing compounds were the main alkaloid metabolites and there were differences in the types and contents of alkaloids in the A. pendulum samples from different locations.

2.3. Multivariate statistical analysis

2.3.1. PCA and HCA

Multivariate statistical analysis was used to evaluate the metabolites in A. pendulum . Prior to the differential analysis, a principal components analysis (PCA) was conducted on the grouped samples to observe the degree of variation between different groups and samples within groups.

First, PCA was used to identify patterns in the data and the separation of each group was investigated to evaluate the interpretation and prediction ability of the established model. The PCA score scatter is shown in Fig. 2A . In the PCA analysis, the triplicate data points are closely grouped or overlapping demonstrating good reproducibility. The first 2 principal components, PC1 and PC2, explained 34.95% and 24.67% of the variability in the dataset, respectively, and were associated with geographical differences. In the PCA plot, the biological replicates of GLM and YSZ were concentrated on the left side of the plot, HZX and MYG were distributed on the right, and GNG and ZKW were distributed in the middle. Samples from different locations grouped into 6 distinct groups based on their locations ( Fig. 2A ).

A hierarchical cluster analysis (HCA) using the Z -score normalized metabolite content was performed to evaluate the relationships of the 80 nitrogenous metabolites from the 6 locations. Metabolites with the same characteristics were identified using Euclidean distance and were grouped according to complete linkage, following which the intergroup variation of the metabolite characteristics was assessed. Fig. 2B shows that there were 3 main groups among the different samples along the horizontal direction. The first group included GNG and ZKW, the second group included YSZ and GLM, and the third group included HZX and MYG. Moreover, metabolites with the same characteristics were classified in a heatmap, and the inter-group variation of the metabolites was assessed along the vertical direction. The red areas indicate specific substances between samples in Fig. 2B . Thus, the PCA and HCA results suggest that environmental differences may be responsible for the variation between sample groupings.

2.3.2. OPLS-DA

The PCA and HCA results provided an overview of metabolite differences between populations. OPLS-DA was further used to evaluate the differences observed among samples from different geographical origins. Using HZX as a reference, pairwise sample comparisons were conducted for 5 groups as follows: HZX vs. MYG, HZX vs. ZKW, HZX vs. GNG, HZX vs. YSZ, and HZX vs. GLM. The results of permutation test (p <0.05) indicated the models are reliable ( Fig. S3 ). High predictability

(Q 2) and strong goodness of fit (R 2 X, R 2 Y) of the OPLS-DA models were observed for the comparisons between HZX and MYG (Q 2 = 0.996, R 2 X = 0.844, R 2 Y = 0.996), HZX and ZKW (Q 2 = 0.998, R 2 X = 0.917, R 2 Y = 0.999), HZX and GNG (Q 2 = 0.998, R 2 X = 0.873, R 2 Y = 0.999), HZX and YSZ (Q 2 = 0.998, R 2 X = 0.908, R 2 Y = 0.999), and HZX and GLM (Q 2 = 0.999, R 2 X = 0.915, R 2 Y = 0.999). The OPLS-DA scores indicated that there was large variability between HZX and the other groups with respect to the X-axis ( Fig. 3 ). In addition, the extracts of samples from different regions were found dispersed along the Y-axis, illustrating their chemical variability. Collectively, these results indicate significant differences among the alkaloids in the extracts. The R 2 and Q 2 values of the OPLS-DA model were high in each group, confirming that the models had good prediction ability and reliability, which could thus be used to further identify differentially accumulated metabolites.

2.4. Screening and identification of differential metabolites of A.

pendulum from different locations

To further our understanding of the metabolite differences between HZX vs. MYG, HZX vs. ZKW, HZX vs. GNG, HZX vs. YSZ, and HZX vs. GLM, differential metabolite screening was performed using all 80 chemical compounds identified with a fold-change score of ≥ 2 or ≤ 0.5 and VIP score ≥ 1 ( Ali et al., 2021). The volcano plots further showed the results of the OPLS-DA. Volcano plots of the different comparisons are shown in Fig. 4. A total of 51 compounds were identified as discriminatory metabolites (26 upregulated, 25 downregulated) between HZX vs. MYG, 52 compounds (36 upregulated, 16 downregulated) between HZX vs. ZKW, 57 compounds (33 upregulated, 24 downregulated) between HZX vs. GNG, 58 compounds (38 upregulated, 20 downregulated) between HZX vs. YSZ, and 60 compounds (41 upregulated, 19 downregulated) between HZX vs. GLM ( Fig. 4 , Table S2 ).

We compared the ion intensity of each significantly abundant metabolite between samples from different locations. According to the Venn diagram, there were 19 significant differentially abundant alkaloid metabolites shared by samples from the 6 locations ( Fig. 5 ), including hordenine, pallidine, corydine, argemonine, 12-epi-dehydronapelline, lepenine, polyschistine A, vilmoridine, vilmorrianine G, karakanine, turupellin, neostemonine, 11-acetyllepenine, lasiandroline, 14-acetylsachaconitine, condelphine, 14- O -acetylneoline, spicatine A, and N - deethyl- N -19-didehydrosachaconitine (Table S3). These were categorized into 9 C 19 diterpenoid alkaloids, 5 C 20 diterpenoid alkaloids, 1 apomorphline alkaloid, 1 pyrrole alkaloid, 1 promorphinane alkaloid, 1 isoquinoline alkaloid, and 1 aromatic alkylamine alkaloid. Terpenoid alkaloids varied greatly in quantity and relative content, which were the main contributors to the metabolite diversity of A. pendulum samples ( Fig. 6 ).

2.5. Anti-inflammatory and analgesic activity analysis of the extracts

Aconitum alkaloids have a wide range of anti-inflammatory and analgesic properties and have thus been used to treat inflammatory and neuropathic pain, especially diterpenoid alkaloids ( Huang et al., 2017). In present study, A. pendulum samples exhibited analgesic activity (Table S7). Meanwhile, we evaluated the anti-inflammatory activities of the A. pendulum samples from different locations using macrophage cells (RAW264.7) ( Table 1 , Table S8). Results showed that the anti-inflammatory activities differed among samples from different locations. The HZX samples (low altitude) demonstrated the best inhibition rate (23.1 ± 3.54%) at 50 μg/mL The content of polyschistine A in HZX was higher than the samples from other regions. Thus, we speculated that this content contributed to the high inhibitory activity of the low-altitude HZX samples.

2.6. Correlation analysis among environmental parameters, differential alkaloid abundance, and anti-inflammatory activity

The correlation analysis of environmental factors and differentially abundant metabolites showed that latitude, longitude, altitude, and aspect were significantly correlated (Table S4), where latitude and longitude were negatively correlated with lepenine, vilmoridine, and

14- O -acetylneoline. Aspect was significantly negatively correlated with argemonine, condelphine, spicatine A, and polyschistine A (Table S4), while altitude was positively correlated with hordenine ( Fig. 7a ). Moreover, in the correlation analysis between the differentially abundant metabolites and anti-inflammatory data, argemonine, polyschistine A, and spicatine A were significantly positively correlated (Table S5), while N -deethyl- N -19-didehydrosachaconitine was significantly negatively correlated (Table S5) ( Fig. 7b ). These results indicated that the chemical compositions and contents of the samples from different regions were affected by environmental parameters. Furthermore, argemonine, polyschistine A, and spicatine A were significantly positively correlated with activation inhibition rate (Table S5), indicating their possible anti-inflammatory activities.

3. Discussion

Our assessment of the alkaloid compounds present in A. pendulum indicates that this herb possesses considerable potential as a source of anti-inflammatory and analgesic agents. This is the first study to evaluate the differences in alkaloid constituents and their anti-inflammatory activities in A. pendulum samples collected from 6 different habitats in the Qinghai region of the Qinghai-Tibet Plateau. Our widely targeted metabolomics analysis identified 80 nitrogen-containing chemical compounds. Among them, pingbeimine C, neostemonine, argemonine, pallidine, norrisocorydine, armepavine, isosophocarpine, and 7,11- dehydromatrine were detected for the first time in the Aconitum genus, while szechenyianine F, pseudaconine, aldohypaconitine, flavaconitine, and 58 other alkaloids were detected in A. pendulum for the first time.

The Qinghai-Tibet Plateau is referred to as a ‘natural laboratory of plant diversity’, and its unique ecological environment produces rich medicinal plant resources. Alkaloids are a major active ingredient in A. pendulum and therefore their study in this species has attracted great attention. A metabolomics approach was applied to explore the metabolic changes in samples from 6 regions with different elevation levels and environmental parameters. Using chemometrics, 19 compounds were identified as potential metabolic markers of the samples from different areas. Results revealed that samples from low altitudes contained more diverse alkaloids than samples from higher altitudes. However, some alkaloids were more abundant in the samples from high altitudes than from low altitudes, such as 14- O -acetylneoline, vilmoridine, and 11-acetyllepenine. At large spatial scales, plants produce specific chemical compounds under environmental stress, resulting in regional differences in metabolites ( He et al., 2020; Wang et al., 2019a). Thus, we hypothesized that high altitude samples were more stressed and thereby produced greater amounts of these compounds to adapt to the harsh high-altitude environment.

Previous studies have reported that plants produce different metabolites in different ecological environments. Altitude, temperature, and other environmental factors affect the quality, composition, and efficacy of metabolites ( Sheng et al., 2018). For example, samples of Zanthoxylum species were collected from different locations in the Tequendama region and their inhibitory activities against acetylcholinesterase and butyrylcholinesterase differed between locations ( Plazas et al., 2019). Similarly, the chemical constituents and biological activities of Juniperus przewalskii in the Qinghai-Tibet Plateau were significantly affected by altitude ( Liu, 2019). Many excellent traditional medicinal plants exist in Qinghai (referred to as “Dao-di” herbs). In this study, the anti-inflammatory activities of samples from different regions were evaluated. We found discrepancies in the anti-inflammatory activities among different regions, with the high-altitude samples containing fewer anti-inflammatory compounds.

Correlation analysis elucidated the relationships among the environmental parameters, differentially abundant alkaloids, and anti-inflammatory activities, suggesting that the environment was responsible for the differences in chemical compositions, ultimately influencing the anti-inflammatory activities of the samples. More importantly, the correlation analysis provided a basis for elucidating the diversity in metabolites of A. pendulum samples from different locations.

In conclusion, using a widely targeted metabolomics approach, a total of 80 putative chemical compounds were detected, 19 of which were identified as potential metabolic markers of A. pendulum samples from 6 regions. The anti-inflammatory activities of these samples were compared. C 19 diterpenoid alkaloids showed higher inhibition rates than C 20 diterpenoid alkaloids (Table S9). The high-altitude samples contained fewer anti-inflammatory compounds than samples from other regions. Moreover, the correlation analysis determined the factors responsible for the observed differences in metabolites of the A. pendulum samples based on their site of origin. These findings enhance our understanding of the chemical compositions of different A. pendulum ecotypes.

Figures

Fig. 1. Comparison of the total peak areas of various classes of metabolites among HZX, MYG, ZKW, GLM, YSZ, and GNG samples from different regions. Bars represent the sum of the peak areas for all metabolites belonging to each class.

Fig. 2. Overview of the global metabolic profiles of the rhizomes collected from 6 growing areas. (A) Principal component analysis (PCA), and (B) heat map hierarchical clustering analysis, The content of each metabolite was normalized to the complete linkage hierarchical clustering. Each example is visualized in a single column and each metabolite is represented by a single row. Red indicates high abundance, whereas metabolites with low relative abundance are shown in green. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 3. Differential metabolite analysis OPLS-DA plots of MYG, ZKW, GLM, YSZ, and GNG compared to HZX.

Fig. 4. Volcano plots for (a) MYG vs. HZX, (b) ZKW vs. HZX, (c) GLM vs. HZX, (d) YSZ vs. HZX, and (e) GNG vs. HZX. The green dots indicate differential metabolites that were significantly downregulated, red dots indicate differential metabolites that were significantly upregulated, and black dots indicate metabolites that were detected in the samples but were not significant. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 5. Venn diagram illustrating shared or unique metabolite contents that differed significantly among the different comparison groups.

Fig. 6. Comparison of the peak areas of various classes of significantly different abundant metabolites among HZX, MYG, ZKW, GLM, YSZ, and GNG. Bars represent the peak areas of the significantly differentially abundant metabolites.

Fig. 7. Correlation analysis between the environmental parameters and differentially abundant alkaloids (a), and. between differentially abundant alkaloids and anti-inflammatory activity (b). The red block indicates a positive correlation; the green block indicates a negative correlation; * indicates significant correlation; ** indicates extremely significant correlation.. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)