Introduction

Ophiocordyceps sinensis is a mixture of dead insects and fungi formed by Chinese fungus infecting Hepialus larvae1. In China, NO is mainly distributed near the Qinghai-Tibet Plateau, at an altitude of more than 3500 m in the alpine region2. The climate is characterized by low temperature, low oxygen content, high ultraviolet, and variable temperature3,4. These extreme climatic characteristics make NO have many special metabolites, which are widely used in the field of medicine5,6. Studies suggest that NO rich superoxide dismutase may inhibit the excessive production of oxygen-free radicals and treat cerebral ischemia by increasing the activity of superoxide dismutase (SOD), glutathione peroxidase (GSH-PX), and catalase (CAT) in brain tissue7. Modern pharmacology found that cordyceps polysaccharide can inhibit lipid peroxidation in hepatocytes and protect the liver8. Nucleosides may enhance the function of macrophages, showing anti-tumor immune regulation9. In addition, in daily life, NO, ginseng, and velvet antler are known as the three treasures of traditional Chinese medicine and have extremely high nutritional value10,11. Therefore, it is more and more popular among consumers and medical scientists and has attracted the interest of scientists. However, the production of NO is gradually unable to meet the needs of consumers12. Herdsmen obeyed the profit-seeking mentality and blindly expanded the collection of NO, resulting in grassland degradation, reduced ecosystem diversity, and gradual loss of a suitable living environment for O. sinensis. The natural reserves have also been reduced year by year13,14,15.

Therefore, the cultivation of CO has become the academic peak that scientists compete to occupy. Methodologically, the cultivation of CO is divided into three processes. The first is the domestication and cultivation of host insects. Scientists have completed the complete life history of mating, oviposition, hatching, pupation, and eclosion of adults captured in the field in the laboratory16,17. The host bat moth has been reared for multiple life histories to adapt to laboratory feeding conditions. When its survival rate exceeds 80%, it is considered that the domestication of the species is initially completed18. The second is the isolation of Chinese fungus with a high infection rate and high activity19. The most common method is to inoculate its stroma, mycelia, or sclerotia on the medium from NO in the laboratory, and obtain Chinese fungus by adjusting the medium scheme and culture conditions20. Finally, Chinese fungus was artificially inoculated into the host insects by feeding method, epidermal injection method, or spray method21,22. After some time, the larvae died to form a dead insect, the stroma germinated from the top of the larvae, and the CO culture was completed. Due to the lack of a variety of environmental factors, the composition and content of CO metabolites may be different from NO. This indirectly leads to the preference of NO and CO in medicinal value and health care function. At present, CO cultivation technology is gradually improving, and scientists hope to use CO as an alternative to NO in the field of medical and physical health care23,24. Therefore, the differences in metabolites and medicinal functions of CO and NO are one of the urgent problems to be solved.

When used as a Chinese herbal medicine, NO has two medicinal forms, namely naturally fresh O. sinensis (NFO) and naturally dired O. sinensis (NDO). In the existing research, NFO and CFO are more used for research, ignoring the medicinal value of NDO and CDO. Therefore, NFO, cultivated fresh O. sinensis (CFO), NDO, and cultivated dried O. sinensis (CDO) were included in this study. In this study, the differences in metabolites between NFO and CFO (NFO vs CFO), and between NDO and CDO (NDO vs CDO) were compared to provide valuable metabolite data for the development of their medicinal functions.

Results

Overview of metabolites

The metabolic profile showed that a total of 2,641 metabolites were identified and annotated, including 1,229 species that could be classified by KEGG, which were divided into 18 superclasses and 108 classes (Table S2). The top 6 superclasses were organic acids and derivatives (300 metabolites, 25.64%), lipids and lipid-like molecules (218 metabolites, 18.63%), organoheterocyclic compounds (168 metabolites, 14.36%), benzenoids (151 metabolites, 12.91%), phenylpropanoids and polyketides (140 metabolites, 11.97%), and organic oxygen compounds (139 metabolites, 11.88%) (Fig. 1).

Figure 1
figure 1

KEGG identified the composition and proportion of metabolites.

Full size image

The top 6 classes were carboxylic acids and derivatives (249 metabolites, 20.26%), organooxygen compounds (138 metabolites, 11.23%), fatty acyls (105 metabolites, 8.54%), benzene and substituted derivatives (97 metabolites, 7.89%), prenol lipids (67 metabolites, 5.45%), and flavonoids (56 metabolites, 4.56%).

Multivariate statistical analysis

Multivariate statistical analysis was used to determine the differences in metabolites between NO and CO. Principal component analysis (PCA) results showed that CDO, CFO, NDO, and NFO were significantly divided into 4 groups and the distance was far (Fig. 2A), indicating that the metabolites of NO and CO were quite different. The clustering and correlation analysis of the samples (Fig. 2B, C), all support that CDO and NDO are clustered into one branch, CFO and NFO are clustered into one branch, indicating that there is a significant difference in the metabolic spectrum between fresh and dry O. sinensis. It also shows that the inclusion of dry O. sinensis in the research scope of this study has important reference value for the discovery of their medicinal functions.

Figure 2
figure 2

Multivariate statistical analysis of 4 O. sinensis samples. (A) PCA analysis of 4 groups samples. (B) Cluster analysis of 4 groups samples. (C) Correlation analysis between 4 groups samples.

Full size image

In the OPLS-DA score plot, different species of O. sinensis were significantly distinguished. The OPLS-DA model had high interpretability (\({R}_{X(NFO VS CFO)}^{2}\)=0.716, \({R}_{Y(NFO VS CFO)}^{2}\)=1, \({R}_{X(NDO VS CDO)}^{2}\)=0.718, and \({R}_{Y(NDO VS CDO)}^{2}\)=1) and predictability (\({Q}_{NFO VS CFO}^{2}\)=0.997, and \({Q}_{NDO VS CDO}^{2}\)=0.993). It shows that the model has good fitting and strong predictability (Figure S1A, 1B). To prevent over-fitting of the model, this study used 200 permutation tests for verification. The results show that the verification intercepts of R2 and Q2 are \({R}_{NFO VS CFO}^{2}\)=0.77, \({Q}_{NFO VS CFO}^{2}\)=0.04, \({R}_{NDO VS CDO}^{2}\)=0.73, and \({Q}_{NDO VS CDO}^{2}\)=− 0.09 (Figure S1C, 1D), respectively, indicating that the model is not over-fitting.

Among the identified metabolites, NFO had more abundant lpc 18:2, glycerophosphocholine, acetyl coenzyme a, phosphorylcholine, oleamide, and fenpropidin than CFO. CFO is rich in bis (2-ethylhexyl) adipate, tyramine, spermine, l-arginine, n6, n6, n6-trimethyl-l-lysine, and paminoazobenzene (Figure S1E). Compared with CDO, NDO is rich in leu-pro, medermycin, traumatic acid, and l-carnitine. CDO is rich in 2-amino-1-phenylethanol, 2-thio-s-acetyl-sn-glyceryl-3-phosphorylcholine, acetyl coenzyme a, paminoazobenzene, glycerophosphocholine, lpc 18:2, fluvoxamine, chlormadinone acetate, N6, N6. N6-trimethyl-l-lysin, and l-o-hexadecyl-2-deoxy (Figure S1F). The results showed that N6, N6, N6-trimethyl-l-lysin, and paminoazobenzene were abundant in CO (CDO and CFO), but not in NO (NFO and NDO), which may be potential biomarkers of NO and CO.

Identification and abundance analysis of differentially accumulated metabolites (DAMs)

By comparing the metabolites in wild and artificial O. sinensis samples (NFO vs CFO, and NDO vs CDO), VIP ≥ 1 was used to screen differential metabolites (Fig. 3A). In NDO vs CDO, a total of 508 metabolites were screened, of which 200 were up-regulated and 308 were down-regulated (Fig. 3B), mostly belonging to organic acids and derivatives (88 metabolites, 17.32%) and lipids and lipid-like molecules (55 metabolites, 10.83%). In NFO vs CFO, a total of 492 metabolites were screened, of which 331 were up-regulated and 161 were down-regulated (Fig. 3C), most of which belonged to organic acids and derivatives (79 metabolites, 16.06%) and lipids and lipid-like molecules (49 metabolites, 11.42%).

Figure 3
figure 3

Differentially expressed metabolites screening. (A) Differential VIP plots the top 20 DAMs in CDO, CFO, NDO, and NFO. (B) Volcano plot of DAMs in NDO vs CDO. (C) Volcano plot of DAMs in NFO vs CFO.

Full size image

It was found that NO is rich in organic acids and derivatives, and CO is rich in lipids and lipid-like molecules. Quantitative analysis of metabolites showed that NO was rich in linoleic acid and oleic acid, and CO was rich in thymol-β-d-glucoside, gly-his-lys, and hydroquinidine (Fig. 4). The discovery of these DAMs not only effectively identifies NO and CO, but also provides a reference for the medicinal research of NO and CO.

Figure 4
figure 4

Comparative analysis of the abundance of DAMs in the top 20 species of 4 species of O. sinensis. At the level of P < 0.05, the representation of different capital letters was significantly different.

Full size image

Hierarchical cluster analysis

In NDO vs CDO, DAMs were clustered into 2 groups, and both cluster 1 and cluster 2 contained organooxygen compounds and cinchona alkaloids. In organooxygen compounds, CDO is rich in sucrose and perseitol, and NDO is rich in l-iditol. In cinchona alkaloid, CDO is rich in hydroquinidine, and NDO is rich in malate (Fig. 5A). This indicated that in NDO and CDO, sucrose and perseitol showed opposite accumulation with l-iditol, and hydroquinidine and malate also showed opposite accumulation. In NFO vs CFO, DAMs were clustered into two groups, and both cluster 1 and cluster 2 included fatty acyls. Oleic acid was more abundant in NFO, while nonanoic acid and 1-hydroxy-2-naphthoic acid were more abundant in CFO (Fig. 5B). This indicated that Oleic acid and 1-hydroxy-2-naphthoic acid accumulated oppositely in NFO and CFO.

Figure 5
figure 5

HCA is based on the top 20 DAMs. (A) NDO vs CDO, (B) NFO vs CFO.

Full size image

NO has more abundant l-iditol, malate, and oleic acid, while CO is rich in sucrose, perseitol, hydroquinidine, nonanoic acid, and 1-hydroxy-2-naphthoic acid. The differential accumulation of these compounds may be one of the main reasons for the differences in the metabolic profiles of NO and CO, and the similar DAMs often have functional similarities or complementarities in biology, which will promote the mining of different medicinal functions of NO and CO.

KEGG pathway enrichment analysis of DAMs

DAMs were annotated by KEGG, and the results showed that DAMs were also enriched in a total of 94 metabolic pathways such as phenylalanine metabolism, glycerolipid metabolism, lysine biosynthesis, and arginine biosynthesis in NFO vs CFO (Table S2). In NDO vs CDO, DAMs were enriched in 86 metabolic pathways such as glutathione metabolism, cysteine and methionine metabolism, and biosynthesis of amino acids (Table S3). The top 20 enrichment pathways were selected for analysis. It was found that ascorbate and aldarate metabolism, carbon metabolism, pyrimidine metabolism, fatty acid biosynthesis were the main enrichment pathways of NO and CO (Figure 6A, and B ).

Figure 6
figure 6

KEGG enrichment circle diagram of DAMs. (A) NDO vs CDO; (B) NFO vs CFO. Among them, the first circle from outside to inside is the pathway of the first 20 enrichment, and the outside circle is the coordinate scale of the number of differential metabolites. Different colors represent different superclasses. The second circle is the number of differential metabolites in the pathway and the −log10Q value. The more the number of differential metabolites, the longer the strip, the smaller the −log10Q value, and the redder the color. The third circle is the bar chart of the proportion of up-regulated and down-regulated differential metabolites. Dark purple represents the proportion of up-regulated differential metabolites, and light purple represents the proportion of down-regulated differential metabolites. The specific values are shown below. The fourth circle is the Rich Factor value of each pathway, the background grid line, each grid represents 0.1.

Full size image

Discussion

O. sinensis has a long history as a medicinal material25. In recent years, scientists have successfully cultured CO in the laboratory26,27. Scientists have carried out a lot of research on whether CO can be a substitute for NO28,29. However, there is a lack of sufficient metabolite evidence in the scientific literature to effectively distinguish NO from CO, especially the difference between NDO and CDO. In this study, the metabolite characteristics of NFO vs CFO, and NDO vs CDO were compared and analyzed, and the differences in metabolic profiles of NO and CO were found, which not only provided data basis for the development of the medicinal value of O. sinensis, but also provided guidance for consumers to choose NO and CO.

In NO, the abundance of organic acids and derivatives is higher, specifically organooxygen compounds, and cinchona alkaloids. Studies have shown that O. sinensis is rich in alkaloids (3'-deoxyadenosine, pyrimidines, adenosine, etc.)30,31. We observed that they used NFO, which can effectively inhibit tumor cell growth in a concentration-dependent manner32,33. This study found that although the overall content of organic acids and derivatives in NO was higher, its nucleoside abundance was lower than that of CO, and more leu-pro, medermycin, traumatic acid, and l-carnitine were detected in NDO. This proves that in NO, the preservation process promotes the accumulation of nucleosides, which may lead to different medicinal functions of NFO and NDO, especially NDO may play a better role in antitumor function. This study complements the missing data on the metabolites of post-storage O. sinensis (NDO and CDO), and more animal model experiments are needed to verify their detailed medicinal functions. In addition, adenosine was used for NO quality control standards at the same time34,35, this study found that it can also effectively distinguish CO. Adenosine has been reported to trigger and mediate ischemic preconditioning to reduce myocardial ischemia-reperfusion injury and has a good protective effect on the myocardium36,37,38. Therefore, NO may have better myocardial protection. In addition, this study suggests that glycine, histidine, l-lysine, linoleic acid, nonanoic acid, oleic acid, sucrose, l-iditol, and perseitol are also potential biomarkers of NO and CO.

Lipids and lipid-like molecules were more abundant in CO. Lipids such as unsaturated FFAs and eicosanoids have been reported to have multiple biological activities, such as improving the anti-inflammatory ability of organisms, improving memory, and improving cognitive deficits39,40,41. Modern pharmacology believes that O. sinensis contains at least eight essential amino acids42,43, which play an important role in the treatment of nervous system diseases, inhibition of bacteria, and enhancement of immune function44,45,46,47. Glutamate, tryptophan, and tyrosine have been reported to have immune-enhancing physiological activities48,49. This study found that CO is rich in glycine, histidine, and l-lysine. Therefore, CO may have a better effect in improving the nervous system and regulating the immune function of organisms. In NO and CO, there was no significant difference in the content and composition of antibacterial active substances, cordyceps polysaccharides, and sterols. This indicates that CO can be an alternative to NO when these compounds need to be used.

Methods

Sample collection

CFO (6 repetitions) was purchased from Shenzhen Dongyang Industrial Development Co., Ltd., China, and NFO (6 repetitions) was purchased from Baohuitang Co., Ltd., Qinghai Province, China, from Zaduo County (95° 33′ 29′′E, 33° 12′ 31′′N, Altitude: 4568 m), Yushu City, Qinghai Province, China. On July 17, 2023, CFO and NFO were placed in a petri dish covered with filter paper in the dark, dried at natural temperature, weighed once every 3 h, until the weight did not change significantly twice in a row. It was considered that the drying was completed, and CDO (6 repetitions) and NDO (6 repetitions) were obtained respectively. Four samples were divided into two groups, the NO group included NDO and NFO, and the CO group included CDO and CFO.

Sample pretreatment and UPLC-MS/MS

Analysis Based on the existing research results, this study improved the determination method in the reference50,51. Four groups were washed three times with sterile distilled water, and then the stroma and sclerotia mixed liquid nitrogen was ground. (1) Six samples were taken for each component, and 50 mg of each sample was weighed and placed in a 2 mL EP tube. The medium-sized grinding beads were added to assist grinding, and 250uL 4 °C pre-cooled liquid extractant (methanol: water = 4:1) was added for the extraction of metabolites. (2) Homogenate in the tissue disruptor, add 1 mL 4 °C pre-cooled extract, ice bath ultrasonic extraction for 20 min, and then stand at − 20 °C for 1 h. (3) Using a low-temperature high-speed centrifuge, the parameters were set to 15,000 g, 4 °C, and extracted for 20 min. The supernatant was taken for UPLC-MS/MS analysis. Methods are as follows, column: agilent 1290 infinity LC (100 nm × 2.1 mm, 1.7 um), flow rate: 0.4 mL/min, column temperature: 40 °C, injection volume: 2 μL. The mobile phase was ultrapure water (containing 0.04% acetic acid) and acetonitrile (containing 0.04% acetic acid). Elution gradient: 0.0–0.5 min water: acetonitrile = 95:5 (V/V); 0.5–7.0 min, 5:95; 7.0–8.0 min, 5:95; 8.1–12.0 min, 95:5. Mass spectrometry conditions were as follows: electrospray ionization (ESI) temperature was 500 °C; the mass spectrometry voltage was 5500 V; the helium pressure was 25psi; the collision activation dissociation (CAD) parameter is set to high. In the triple quadrupole (QQQ) system, each ion pair is scanned based on the optimized clustering potential (DP) and collision energy (CE). In addition, 10 uL was taken from each sample and mixed well to make QC samples, and the same analysis was performed.

Data processing and metabolite identification

The original mass spectrometry data was converted into MzXML format using Analyst 1.6.3 software and imported into XCMS format for processing. The XCMS parameters are as follows: for peak picking, centWave m/z = 10 ppm, peakwidth = c (10,60), prefilter = c (10,100). For peak grouping, bw = 5, mzwid = 0.025, minfrac = 0.5. Firstly, baseline filtering, peak recognition, integration, peak alignment and retention time correction were performed. The characteristic peaks with relative standard deviation (RSD) > 30% in QC samples were filtered to obtain the data matrix of retention time (RT), mass-to-charge ratio (m/z) and peak intensity. The MS and MSMS data were matched and annotated with commonly used metabolic databases (KEGG, HMDB, Metlin, MoNA and self-built databases) to obtain metabolite information. The total peak area normalization and log10 logarithm of the response intensity of the sample mass spectrum peak were performed to reduce the error caused by the sample preparation process and the instrument, and the data matrix for subsequent analysis was obtained. The content of metabolites was imported into OriginPro 2018 software for principal component analysis, orthogonal partial least squares discriminant analysis (OPLS-DA) and S-Plot analysis difference multiple analysis. The screening criteria for differential metabolites also met the following conditions: P ≤ 0.05, VIP > 1. The relative quantitative hierarchical clustering of metabolites and differential metabolites was performed by heatmap package in Rv3.3.2. KEGG pathway enrichment (http://www.kegg.jp/kegg/kegg1.html)52,53,54 analysis of metabolites was performed through the OmicShare cloud platform (https://www.omicshare.com).

Conclusions

In this study, non-targeted metabolomics was used to compare the metabolic profiles of NO (NDO and NFO) and CO (CDO and CFO). The differences in their metabolic profiles were mainly derived from Organic acids and derivatives (NO>CO) and lipids and lipid-like molecules (NO<CO). Through comprehensive analysis, it was found that the abundances of organooxygen compounds, cinchona alkaloids, and fatty acyls were different in NO and CO. Specifically, NO is rich in l-iditol, malate, linoleic acid, and oleic acid; CO is rich in sucrose, perseitol, hydroquinidine, nonanoic acid, 1-hydroxy-2-naphthoic acid, hymol-beta-d-glucoside, and gly-his-lys. These compounds have the potential to be biomarkers of NO and CO. KEGG enrichment analysis showed that their differences mainly came from ascorbate and aldarate metabolism, carbon metabolism, pyrimidine metabolism, and fatty acid biosynthesis.