Based on the distribution of R. cordifolia and its origin, as described in the literature (Xue et al., 2009; Yu et al., 2017; Yan, 2021), Liaoning, Hebei, Henan, Shanxi, Shaanxi, Gansu, and Sichuan were selected for our sampling ( Figure 2 ). Samples were collected from July to August in 2021 and 2022, and 3–7 whole underground parts of R. cordifolia were collected from each sample point, while trying to ensure that the samples had the same plant age. Soil at 0–30 cm depth in the roots was collected for subsequent analysis. The coordinates of each sampling point were recorded for the subsequent acquisition of meteorological and topographical factors, as shown in Table 1 .

The R. cordifolia root powder (0.5 g) was precisely weighed and placed in a stoppered conical flask. Methanol (100 mL) was added to the flask and stored it overnight. Secondary metabolites were extracted using ultrasound (power 250W, frequency 40kHz, 30 minutes). After the ultrasound was completed, the sample was cooled to room temperature, shaken well, and filtered. The filtrate (50 mL) was measured precisely, and evaporate to dryness in a 60°C water bath. The residue was dissolved by adding 20 mL methanol:25% hydrochloric acid (4:1). The solution was heated and hydrolyzed in an 80°C water bath for 30 minutes. The solution obtained after hydrolysis was cooled immediately, 3 ml of triethylamine was added, mixed, transferred to a 25 ml measuring flask, and diluted to volume with methanol. The solution was filtered through a microporous membrane and used for detection (Chinese Pharmacopeia Commission, 2020).

Ultra-high-performance liquid chromatography (UHPLC) (Ultimate 3000 RSLC, Thermo Fisher Scientific, MA, USA) was used to determine the mollugin and purpurin contents in the extract. Octadecylsilane bonded silica gel was used as the filler of chromatographic column, and the 25:50:25 mixture of methanol, acetonitrile, and 0.2% phosphoric acid was used as the mobile phase The detection wavelength is 250 nm. According to the calculations of mollugin and purpurin peaks, the theoretical plate number should not be less than 4000.

Soil samples were taken back to the laboratory and sieved to remove stones, fine roots, and other debris. The soil was naturally dried and ground, and part of the ground samples were passed through a 0.25 mm sieve to determine the organic matter content of the soil. The remaining samples were sieved through a 1 mm sieve, and the soil pH, available nitrogen, available phosphorus, and available potassium were determined using the methods described by Hu and Wang (2020). Soil pH was determined by the potentiometric method, organic matter content was determined by the potassium dichromate volumetric method, available nitrogen content was determined by the alkaline hydrolysis diffusion method, available phosphorus content was determined by the NaHCO₃ extraction-colorimetric method, and available potassium content was determined by the ammonium acetate extraction-flame photometric method.

The annual mean temperature (TEM), annual precipitation (PRE), and annual sunshine hours (SSD) for 2010–2020 were obtained from the Resource and Environment Science and Data Center (https://www.resdc.cn/) (Xu, 2017). Elevation data were obtained from WorldClim (https://www.worldclim.org/). Finally, the climatic and geographic factors of each sampling site were extracted using the extraction tool in ArcGIS.

The data were pre-processed using Microsoft Excel 2019. The Spearman correlation analysis was conducted between environmental variables and secondary metabolites using the corrplot package in R 4.2.2. The stepwise regression was analyzed using the car package in R 4.4.2. Corresponding plots were constructed using ggplot2 package in R 4.2.2 and ArcGIS 10.7.

Sampling sites were covered from the east to the west and from the north to the south of China ( Figure 2 ). The longitude spanned 23.24°, and the latitude spanned 13.34°. The elevation in the study area varied from 5 to 3395 m, rising from east to west ( Table 1 ).

Soil properties varied considerably between the sampling sites ( Table 1 ). The average soil pH varied between 6.13 and 8.54, the average organic matter varied between 10.02 and 334.23 g kg^-1, the average available nitrogen content ranged from 41.18 to 1,168.00 mg kg^-1, the average available phosphorus content ranged from 5.87 to 100.23 mg kg^-1, and the average available potassium ranged from 71.76 to 1,038.73 mg kg^-1 ( Table 1 ).

The study area spans multiple climate zones in China, including temperate, warm temperate, subtropical, and alpine. The annual mean temperature of each sampling site varied between 3.12 and 17.31°C, annual precipitation varied between 446.45 and 1,206.15 mm, and annual sunshine hours varied between 1,037.23 and 2,690.20 h ( Table 1 ).

The contents of purpurin and mollugin at each sampling point were calculated according to the UHPLC chromatogram ( Table 2 ). The chromatograms of standard and sample solutions are shown in Figure 3 . The purpurin and mollugin contents in R. cordifolia roots varied greatly among sampling sites in the study area. Purpurin content ranged from 0.00 to 3.03 mg g^-1, the minimum and maximum values were found at sampling points in Sichuan. The content of mollugin varied from 0.03 to 10.09 mg g^-1. The minimum value was found in Sichuan sampling site, while the maximum value was found in Liaoning sampling site. The contents of the two secondary metabolites in R. cordifolia roots varied greatly in different provinces. For example, the content of purpurin in more than half of the samples in Gansu was less than 1.00 mg g^-1, and the highest was 1.34 mg g^-1. The content of purpurin in in the samples in Hebei was higher than 1.00 mg g^-1, and the highest content was 2.05 mg g^-1. However, the content of mollugin in the samples in Hebei was less than 4.00 mg g^-1, and the highest content was 5.10 mg g^-1. The content of purpurin in the roots of R. cordifolia in nearly half of the sampling sites in Sichuan was less than 1.00 mg g^-1, and the highest was 3.03 mg g^-1. The content of mollugin in the roots in almost all sampling sites was not up to standard, indicating that the quality of R. cordifolia in each sampling site in Sichuan was uneven. The content of secondary metabolites in the roots of R. cordifolia in Shanxi, Shaanxi and Henan provinces can basically reach the standard. In addition, the content of secondary metabolites in the roots of R. cordifolia in a small number of sampling points in Liaoning was not up to standard, mainly distributed in the western part of Liaoning, while the quality of R. cordifolia in the southeastern part of Liaoning is higher ( Figure 4 ).

Purpurin and mollugin contents showed different relationships with soil, climate, and topographic factors ( Figure 5 ). Longitude showed a significant positive correlation with purpurin content (p < 0.05) ( Figure 5A ), while elevation (p < 0.001) showed a significant negative correlation ( Figure 5C ). Annual mean temperature (p < 0.001) and annual precipitation (p < 0.001) showed highly significant positive correlations with purpurin content ( Figures 5D, E ), whereas annual sunshine hours showed a significant negative correlation with purpurin content (p < 0.05) ( Figure 5F ). Organic matter content (p < 0.05), available phosphorus content (p < 0.05), and available potassium content (p < 0.01) were significantly negatively correlated with purpurin content ( Figures 5H, J, K ). For mollugin, longitude (p < 0.05) and latitude (p < 0.05) were significantly positively correlated with mollugin content ( Figures 5A, B ). Organic matter content (p < 0.05) showed a significant and negative correlation with mollugin content (p < 0.05) ( Figure 5H ). There was no significant correlation between mollugin content and climatic factors ( Figures 5D–F ), however, there was a negative correlation between annual precipitation and mollugin content (p = 0.052) ( Figure 5E ).

In order to further explore the relationship between the two secondary metabolites and environmental factors, a stepwise regression model between purpurin, mollugin and environmental factors was established ( Table 3 ). Model results showed that purpurin and mollugin were affected by different factors. Purpurin in R. cordifolia roots was affected by latitude, annual mean temperature and annual precipitation, and the three factors had a positive impact on purpurin. Otherwise, the model results showed that root mollugin content was only negatively affected by annual precipitation.

R. cordifolia is widely distributed in China; therefore, there are several places of origin recorded in the literature, including Shanxi, Sichuan, Henan, Shaanxi, Jiangsu, Hubei, Shandong, Anhui, and Hebei (Xue et al., 2009; Yan, 2021). Several studies have investigated the medicinal components in the roots of R. cordifolia from different origins and found that R. cordifolia from Shanxi, Shaanxi, Henan, Hebei, Jiangsu, Shandong, and Anhui met the Chinese Pharmacopoeia standards (Xue et al., 2009; Yu et al., 2017). Henan and Shaanxi are recognized R. cordifolia producing areas (Wen et al., 2022). The purpurin and mollugin contents of R. cordifolia roots from eastern and central regions in this study investigated met the requirements of the Chinese Pharmacopoeia, especially in central regions. Therefore, the study confirmed that R. cordifolia produced in Henan, Shaanxi, and Shanxi contained higher levels of purpurin and mollugin. However, R. cordifolia resources in Northeastern China have not yet been investigated. In this study, R. cordifolia from Liaoning contained higher purpurin and mollugin contents, which met the requirements of the Chinese Pharmacopoeia, suggesting that Liaoning could be a choice for R. cordifolia cultivation.

Environmental factors, including soil, climate, and topography, showed different degrees of correlation with the two secondary metabolites in the roots. The effects of geographic factors on plant secondary metabolite content have been extensively reported. The content of proanthocyanidins in Ribes nigrum L. was positively correlated with latitude (Yang et al., 2019b), whereas harpagoside content in S. ningpoensis was negatively correlated with latitude (Yang et al., 2011). Different essential oil contents in H. italicum showed different correlations with elevation (Melito et al., 2016). In this study, purpurin and mollugin contents were significantly and positively correlated with longitude and latitude ( Figure 5 ; Table 2 ), suggesting that when the environment was closer to northern and eastern China, the purpurin and mollugin contents in R. cordifolia roots were higher, which is similar to the findings of Weston et al. (2013). The two secondary metabolites can cause oxidative stress in cells and therefore act as natural biocides (Robertson et al., 2009; Hook et al., 2018). The higher purpurin and mollugin content found in the eastern and central regions of this study might be due to the fact that the environmental conditions in these regions cause some kind of biotic stress, which in turn leads to the synthesis of more purpurin and mollugin. Previous studies have shown that anthraquinone content in medicinal plants Rheum officinale Baill. (Wang et al., 2013; Yan et al., 2017; Ge et al., 2018) and Rumex nepalensis Spreng (Farooq et al., 2013) increased with elevation, but in the present study elevation negatively affected purpurin content in R. cordifolia roots, and a similar result was found in Rheum australe D. Don (Pandith et al., 2014). The negative correlation between secondary metabolite content and elevation may be related to factors such as ultraviolet radiation or temperature (Spitaler et al., 2006). The positive correlation between purpurin content and annual mean temperature and the negative correlation between elevation and temperature just proved this point.

Soil nutrients play a crucial role in supporting plant growth. In the case of medicinal plants, the presence of appropriate soil nutrients not only enhances their yield but also improves their overall quality. Changes in soil nutrition will inevitably lead to changes in secondary metabolites, understanding the relationship between the content of secondary metabolites and soil nutrients can guide the cultivation of medicinal plants to improve the effectiveness of medicinal plants. However, specific soil factors that cause an increase in secondary metabolite levels in medicinal plants are not the same. Soil pH may be an important factor determining the secondary metabolite content of Echinacea purpurea (L.) Moench (Xu et al., 2022). The available phosphorus in the soil is thought to be a key factor influencing the secondary metabolite content in Salvia miltiorrhiza Bunge (Liang et al., 2021). For Rhodiola sachalinensis A. Bor, available phosphorus and available potassium are the main factors limiting the synthesis of salidroside (Yan et al., 2004). In this study, the findings found that higher soil nutrients were not necessarily beneficial to the synthesis of purpurin and mollugin, which is similar to the study by Shen et al. (2017), where the content of anthraquinone compounds in the roots of Rheum tanguticum Maxim. ex Balf was higher in nutrient-poor soil. This phenomenon might be due to the fact that under nutrient-sufficient conditions R. cordifolia mainly carries out primary metabolism for growth, while under poor soil conditions the plants are stressed and then carry out secondary metabolism and produce secondary metabolites to alleviate the stress.

The climate is crucial for the growth and quality of medicinal plants. Studies have shown that different medicinal plants have different requirements for climatic conditions, and secondary metabolite production can be affected by climatic conditions (Liu et al., 2020). Temperature has multiple effects on the synthesis of secondary metabolites in medicinal plants. On the one hand, temperature stress may cause medicinal plants to produce secondary metabolites for defense (Dong et al., 2011); on the other hand, high or low temperatures may increase the expression of genes involved in the synthesis of secondary metabolites (Tao et al., 2017; Yang et al., 2019b). The harpagoside content in Scrophularia ningpoensis Hemsl (Yang et al., 2011), and the nerolidol content in Helichrysum italicum (Roth) G. Don (Melito et al., 2016) are correlated positively with temperature, while there is a negative correlation between high temperature and total proanthocyanidins in R. nigrum fruits (Yang et al., 2019b). Temperature promoted the synthesis of purpurin in the roots of R. cordifolia, which may be due to high temperature stress or increased expression of related genes. Precipitation also has an impact on the synthesis of secondary metabolites in plants. Typically, drought conditions stimulate the synthesis and accumulation of secondary metabolites in plants (Humbal and Pathak, 2023). However, chlorogenic acid, choline, 3,5-o-dicaffeoylquinic acid, and coumaric acid in E. purpurea are positively correlated with annual precipitation (Xu et al., 2022). Two secondary metabolites in the roots of R. cordifolia were oppositely affected by precipitation, which promoted the synthesis of purpurin but limited the synthesis of mollugin. The impact of precipitation on the secondary metabolism of medicinal plants is complex. This complexity arises due to the potential changes in soil properties, microbial fungi, and root activity that can occur as a result of precipitation (Wang et al., 2021b; Xu et al., 2022). Sunshine hours also affect the production of plant secondary metabolites. For instance, a longer duration of sunshine can increase the content of geniposidic acid in Eucommia ulmoides Oliv. (Dong et al., 2011) and tanshinones in S. miltiorrhiza (Zhang et al., 2018). However, in the present study, correlation analysis suggested that sunshine duration might have an inhibitory effect on purpurin synthesis in R. cordifolia root.

The cultivation of medicinal plants needs to ensure both yield and quality, so Daodi medicinal materials are recognized as high-quality Chinese medicinal products. However, as wild medicinal plants have been over-exploited, they can no longer meet market demand. Therefore, field cultivation has become the main way to ensure the market supply of medicinal plants (Kling, 2016). In the case of field planting, many factors need to be considered comprehensively, including geographical location, soil, climate, etc. First of all, it is necessary to select the appropriate planting area, which needs to consider geographical and climatic factors. Therefore, eastern and northern China can be used as a choice when planting R. cordifolia. Higher temperatures facilitate the synthesis of purpurin, so the temperature in the suitable area needs to be considered. Precipitation has opposite effects on the two secondary metabolites in the roots of R. cordifolia. When planting R. cordifolia, the precipitation situation needs to be weighed, and the amount of watering in the field must also be considered. In order to ensure the yield of medicinal plants, fertilization is a necessary measure. However, the synthesis of two major secondary metabolites in R. cordifolia roots is inhibited by soil nutrients, therefore necessary fertilization experiments need to be performed to determine the appropriate fertilization ratio.

The two main secondary metabolites in the root of R. cordifolia belong to quinone compounds, among which purpurin belongs to anthraquinone compounds (Singh et al., 2021), and mollugin belongs to naphthoquinone compounds (Morita et al., 2015). The Chinese Pharmacopoeia requires that the contents of the two secondary metabolites in the roots of R. cordifolia should be up to standard before they can be used as medicines (Chinese Pharmacopeia Commission, 2020), so the products of R. cordifolia on the market are uneven. How to improve the quality when planting R. cordifolia has become a problem. At present, some researchers have carried out transcriptome studies on R. cordifolia, and found that the synthesis of anthraquinones in the roots involves the synthesis of A, B, and C rings. Among them, A and B rings are completed by the shikimic acid pathway, while the synthesis of C ring comes from isopentenyl diphosphate (Peng et al., 2018). Previous studies have confirmed that isopentenyl diphosphate is synthesized through the 2-C-methyl-D-erythritol 4-phosphate pathway in the Rubiaceae. Isochorismate synthase gene, o-succinylbenzoic acid gene, 1-deoxy-D-xylulose 5-phosphate reductoisomerase gene, and 1-deoxy-D-xylulose 5-phosphate synthase gene are all key genes for the synthesis of anthraquinones (Han et al., 2001). However, the synthetic pathway of naphthoquinone compounds in the roots of R. cordifolia has not been reported. Therefore, in the future, on the one hand, the biosynthesis pathway of anthraquinones will be explored by molecular biology methods, and key genes will be screened. On the other hand, the relationship between the key genes in the synthesis pathway of anthraquinones and naphthoquinones and the ecological environment factors should be analyzed to find the key factors, and then measures can be taken to promote the synthesis of two secondary metabolites.