Introduction

Hawk tea, an herbal tea, is mainly produced from buds or leaves of Litsea coreana Lévl. var. lanuginose and is traditionally and widely consumed in China1,2. Several hawk tea products have been developed using different tea processing methods. Generally, following the collection of buds, leaves, or tender shoots, they need to be withered, fixated, rolled (or not), fermented (or not), and dried to obtain hawk tea products3. Marketed hawk tea usually includes green tea-type hawk tea (GHT), traditional Chinese hawk tea (a white tea, TCHT), oolong tea-type hawk tea (OHT), and black tea-type hawk tea (BHT) made from green tea, white tea, oolong tea, and black tea processing techniques, respectively. Where GHT is unfermented, TCHT is slightly fermented, OHT is semi-fermented, and BHT is fully fermented4. Another important hawk tea is the insect-type hawk tea (IHT), the feces of specialized insects such as Aglossa dimidiata, Martyringa xulaula, and Pyralis farinalis that are fed with the leaves of L. coreana3. The infusion of hawk tea possesses a mellow and sweet taste, a slight odor of camphor and aromaticity, and an absence of bitterness5,6. These beverages are highly appreciated by consumers (more than 30 million) for their unique flavor and quality7,8.

Hawk tea is not only a beverage but also provides potential health benefits and pharmacological properties6. Phytochemical analyses showed that hawk tea contained characteristic polysaccharides, polyphenols, volatile oils, and various flavonoids, contributing to significant biological activities5. Antioxidant activity is one of the most significant bioactivities of hawk tea. The antioxidant capacity of infusions from hawk teas of different raw material maturity (bud, primary leaf, and mature leaf) varied, in which the primary leaf tea infusion possessed higher contents of total polyphenols and showed better DPPH radical scavenging activity and ferric-reducing activity power7. Individual flavonoids from mature leaves picked in different months did not differ significantly in composition, but there were differences in the proportion of content and differences in antioxidant activity6. Different processing methods, pan-frying (PFHT), sun-drying (SDHT), and fermentation (FHT), altered the chemical composition of hawk tea, where PFHT showed the highest content of phenolics and flavonoids, and the in vitro antioxidant activity of PFHT and SDHT was higher than that of FHT8. In addition, storage time also affects the antioxidant activity. The antioxidant capacity of new hawk tea was the strongest and decreased with the extension of storage time9.

The antioxidant activity of tea is related to its antioxidant compounds10,11. Air-dried and powdered hawk tea leaves contained six major flavonoids (hyperin, isoquercitrin, quercitrin, astragalin, quercetin, and kaempferol) and volatile oils, which showed promising antioxidant activity in vitro12,13. High-performance liquid chromatography (HPLC) analysis showed that the contents of hyperin, isoquercitrin, trifolin, astragalin/quercitrin, quercetin, and kaempferol in mature leaf hawk tea were 28.59–49.11, 80.28 –137.35, 14.82 –45.43, 232.63–461.90, 5.38 –11.01, and 4.33–26.57 µg/g, respectively6. Further antioxidant experiments suggested that the antioxidant effects were correlated significantly with the contents of total flavonoids6. The liquid chromatography-tandem mass spectrometry (LC-MS/MS) was used to determine 15 phenolic compounds in hawk tea, combined with an antioxidant activity test and partial least squares regression (PLSR) analysis14. The results showed that isoquercitrin, catechin, astragalin, chlorogenic acid, hyperoside, p-coumaric acid, and kaempferol were the main phenolic compounds that exert their antioxidant efficiency14.

Although a large number of previous studies from different perspectives have shown that hawk tea has good antioxidant activity, and several compounds with antioxidant activity have also been detected, the metabolic characteristics, potential antioxidant compounds, and possible mechanism of action in different processed types of hawk tea products are not clear. Therefore, this study aimed to collect buds and leaves of L. coreana, prepare various types of hawk tea products, reveal the metabolic profiles, analyze potential antioxidant compounds, and explore possible antioxidant mechanisms through untargeted metabolomics, network pharmacology, and molecular docking.

Results and discussion

Metabolic profiles of five types of hawk tea

Different processing significantly influences the biotransformation of phytochemicals, such as phenolic compounds, in tea, leading to variations in flavor and bioactivity15,16. In this study, five different types, BHT, GHT, IHT, OHT, and TCHT, were prepared for analyzing the metabolite profiles (Fig. 1A). As shown in Fig. 1B, the numbers of BHT, GHT, IHT, OHT, and TCHT metabolites were 345, 348, 341, 341, and 349, respectively. Following the classification of metabolites at Class I, hawk tea consisted of ten primary categories, in which, the relatively high content of phenylpropanoids and polyketides (27.73%), lipids and lipid-like molecules (14.29%), organic acids and derivatives (11.2%), benzenoids (8.12%), organic oxygen compounds (6.72%), organoheterocyclic compounds (6.16%), and nucleosides, nucleotides, and analogues (4.48%) (Fig. 1C). Polyphenols exhibit antioxidant, anti-inflammatory, and antimicrobial properties due to their multi-phenolic rings and phenolic OH groups and are subclassified into flavonoids and non-flavonoids17, synthesized through the shikimic acid-derived phenylpropanoid and/or polyketide pathways in plants18. The content of phenylpropanoids and polyketides in hawk tea was over a quarter, which could confer multiple bioactivities to these beverages. Sample clustering results showed that GHT (unfermented) and TCHT (slightly fermented) clustered in a single clade, BHT (fully fermented) and OHT (semi-fermented) clustered in another clade, while IHT (insect tea) clustered alone, which differed significantly from the other four types (Fig. 1D). It is associated with the fact that the biochemical composition varies considerably among different degrees of fermented teas19. PCA analysis was done on all the samples to observe the differences in metabolites among the five different types, which showed that these samples fell into five distinct groups, with PC1 accounting for 39.26% and PC2 accounting for 33.73% of the total variance (Supplementary Fig. S1 online). The PCA score plot suggested that the metabolic differences among the five different types of hawk tea resemble the clustering effect observed in the hierarchical cluster analysis in Fig. 1D. The OPLS-DA score plot also displayed a notable separation among the five groups based on the principle of p-value (< 0.05) and VIP (≥ 1), indicating that the five types of hawk tea shared different metabolic profiles (Fig. 1E).

Fig. 1
figure 1

Metabolic profiles of five different types of hawk tea. (A) Five different types of hawk tea. (B) Number of metabolites in five different types of hawk tea. (C) Percentage of Class I classification of hawk tea metabolites, metabolites in small quantities or without chemical classification attribution were considered “Others”. (D) Hierarchical cluster heatmap of five different types of hawk tea metabolites. (E) Orthogonal projections to latent structure-discriminant analysis (OPLS-DA) of hawk tea metabolites.

Full size image

For the classification of class II metabolites, the five different types of hawk tea mainly consisted of flavonoids, fatty acyls, carboxylic acids and derivatives, organooxygen compounds, and prenol lipids (Fig. 2A). Of these, flavonoids accounted for approximately 50% of the total metabolites, with IHT being the highest and OHT the lowest (Fig. 2A). A comprehensive study of the chemical constituents of essential oils, flavonoids, polyphenols, and polysaccharides has demonstrated that flavonoids are the most dominant component of hawk tea5. The antioxidant capacity of insect tea was higher than other types of hawk tea20, which may be attributable to the highest content of flavonoids. In addition, the flavonoid content of the IHT was reported to be higher than those of Huaxiang and Sanye insect tea21. Metabolites with an average relative content of ≥ 1% in five different types of hawk tea are shown in Table 1. Among them, astragalin is the most abundant, with tribuloside ranking highest only in GHT. The major metabolic compounds of different types of hawk tea were presented in Fig. 2B, with adenosine, astragalin, isoquercetrin, kaempferol-3-glucuronide, tribuloside, 2’,4’,6’-trihydroxydihydrochalcone, and others. Herein, astragalin, kaempferol-3-glucuronide, and tribuloside belonged to flavonoids. Sheng et al. applied widely targeted metabolomics to identify the nutritional components of hawk tea20. The results showed that metabolites such as adenosine, astragalin, tribuloside, (-)-epicatechin, kaempferol, etc. are also the main components of TCHT and IHT20. This is basically consistent with our results (Table 1). However, there are few reports on the characteristics of other types of hawk tea metabolites.

Fig. 2
figure 2

Major metabolites (Class II and Class III) of five different types of hawk tea. (A) Percentage of Class II classification of hawk tea metabolites, metabolites in small quantities or without chemical classification attribution were considered “Others”. (B) Major metabolic compounds of five different types of hawk tea.

Full size image
Table 1 Metabolites and average relative contents (≥ 1%) in five types of hawk tea.
Full size table

Differential metabolites of five types of hawk tea

In this study, 357 metabolites were detected in five types of hawk tea, with 322 identical metabolites present in each type (Fig. 3A). Thirty-five metabolites, excluding the 322 identical ones, were selected for further study to screen the individualized and specific metabolites of different types of hawk tea. The hierarchical clustering heatmap of the content of 35 metabolites in the five types of hawk tea suggested that these metabolites were separated into two main clades, one with only IHT and the other with another four types (Fig. 3B). Metabolites with significantly higher levels than the other types were present in each type of hawk tea, with the characteristic metabolites of BHT being less pronounced. K-means analysis was used to centralize and standardize the metabolites based on the correlation coefficients among the metabolite concentrations to reveal the changing trend of these 35 metabolites in five types of hawk tea. These metabolites were clustered into four sub-classes based on the variation patterns, as shown in Fig. 3C–F, and these trends were consistent with the hierarchical cluster analysis (Fig. 3B).

Fig. 3
figure 3

Overview of differential metabolites in five types of hawk tea. (A) Overlapping relationships of the differential metabolites among BHT, GHT, IHT, OHT, and TCHT are plotted in a Venn diagram. (B) Clustered heat map showing the contents of 35 metabolites in the five types of hawk tea. Not included were the 322 metabolites that were present in all five types. (C)–(F) K-means clustering analysis of 35 metabolites.

Full size image

A comparison of significant differences in metabolites in each sub-class was presented in Supplementary Fig. S2 online. Excluding metabolic compounds that were not significant (marked as “ns”), subclasses 1 to 4 consisted of 12, 9, 4, and 2 metabolites, respectively, that were significantly different (p < 0.05). As shown in Supplementary Table S1 online, Metabolites with p < 0.05, FC ≥ 2 or ≤ 0.5, and VIP > 1 were recognized as differential metabolites. As shown in Supplementary Table S1 online, 15 differential metabolites were screened out, of which Sub-class 1 contained kaempferol 3,7,4’-trimethyl ether, 3,4-dimethoxyphenylethylamine, velutin, deferrioxamine E, nevadensin, sulfanilic acid, artemisinin, and bergenin, which were mainly from IHT; Sub-class 2 contained UDP-L-rhamnose, NAD+, isoquercitrin, and haemoventosine, which were primarily from GHT; Sub-class 3 contained (E)-1-[3-[(2,3-dihydroxyphenyl)methyl]-2,4-dihydroxy-6-methoxyphenyl]-3-phenylprop-2-en-1-one and coclaurine, which were present in high levels in TCHT; Sub-class 4 contained tomasin with the highest levels in OHT (Fig. 3B). Where, kaempferol 3,7,4’-trimethyl ether, velutin, nevadensin, and isoquercitrin belonged to flavonoids. Several identified differential metabolites have been reported to possess significant antioxidant activity, such as bergenin22, isoquercitrin23, and velutin24.

Potential antioxidants and possible action mechanisms in hawk tea

Network pharmacology-based antioxidant targets screening

In this study, six identical major metabolites and 15 differential metabolites of hawk tea were selected to explore the antioxidant molecular mechanism. Identical metabolites included adenosine, astragalin, isoquercetrin, kaempferol-3-glucuronide, tribuloside, and 2’,4’,6’-trihydro- xydihydrochalcone. Differential metabolites were kaempferol 3,7,4’-trimethyl ether, 3,4-dimethoxyphenylethylamine, velutin, deferrioxamine E, nevadensin, sulfanilic acid, artemisinin, bergenin, UDP-L-rhamnose, NAD+, isoquercitrin, haemoventosine, coclaurine, (E)-1-[3-[(2,3-dihydroxyphenyl)methyl]-2,4-dihydroxy-6-methoxyphenyl]-3-phenylprop-2-en-1-one, and tomasin.

Network pharmacology is a multi-effect drug analysis method that combines systems biology and network analysis, which reveals the regulatory network of a drug in the organism at a systems level, enabling analysis of drug-target interactions and prediction of potential action mechanisms25. In this study, 21 active compounds of hawk tea were annotated in the SwissTargetPrediction database. Among these, 2’,4’,6’-trihydroxydihydrochalcone without a corresponding target, deferrioxamine E, NAD+, and haemoventosine with targets of “probability < 0.1”, finally 344 targets were obtained after removing duplicates from the other 17 compounds. In addition, 2,690 antioxidant-related targets (score > 5) were filtered by the GeneCards database. Following, 209 potential targets associated with antioxidant compounds were identified in hawk tea based on their intersection (Fig. 4A). The PPI network of 209 common targets was constructed using the STRING database with Cytoscape software to screen the core antioxidant targets. Then, 26 core targets were found based on the “highest confidence > 0.9” and BC, CC, and DC above the mean values (average values of BC, CC, and DC are 473.3690, 0.2715, and 12.3571, respectively) (Fig. 4A). The PPI network consisted of 168 nodes and 1038 edges, suggesting that several potential antioxidant compounds in hawk tea may play antioxidant roles through multiple targets and pathways.

Fig. 4
figure 4

Schematic diagram of the network pharmacological analysis of antioxidant in hawk tea. (A) Compound targets in hawk tea and antioxidant-related targets, left: intersection targets, right: protein-protein interaction (PPI) network, the outermost circle represents the core targets, arranged in descending order of degree centrality (DC) value. (B) Hawk tea-component-target-pathway-antioxidation network, arranged in descending order of DC value, K374T*: Kaempferol 3,7,4’-trimethyl ether, K3G*: Kaempferol-3-glucuronide.

Full size image

Hawk tea-component-target-pathway-antioxidation network construction

Twenty-six core targets were used for KEGG annotation in Metascape, resulting in the identification of 158 pathways. The 20 pathways with the highest number of entries, along with 17 active compounds and 26 core targets, were selected to construct a hawk tea-component-target-pathway-antioxidant network using Cytoscape software. Six compounds not associated with the targets and pathways: coclaurine, 3,4-dimethoxyphenylethylamine, sulfanilic acid, tribuloside, nevadensin, and (E)-1-[3-[(2,3-dihydroxyphenyl)methyl]-2,4-dihydroxy-6-methoxyphenyl]-3-phenylprop-2-en-1-one were excluded, as was the pathway-unrelated target, PARP1. The remaining 11 compounds and the 25 core targets constructed a network as shown in Fig. 4B. The results suggested that 11 compounds, including adenosine, artemisinin, astragalin, bergenin, isoquercetrin, isoquercitrin, kaempferol-3-glucuronide, kaempferol 3,7,4’-trimethyl ether, tomasin, UDP-L-rhamnose, and velutin may be critical antioxidant metabolites of hawk tea, which mainly involved 25 targets and exerted their antioxidant activities through 20 pathways (Fig. 4). Adenosine is a naturally occurring nucleoside comprising an adenine nitrogenous base linked to a D-ribose sugar via a β-glycosidic bond26. It exhibits free radical scavenging activity, which confers cardioprotective effects in vivo26. Artemisinin is not only an antimalarial drug but also protects neuronal HT-22 cells from glutamate-induced oxidative injury by activating the AKT signaling pathway27. Astragalin is a flavonoid known to have antioxidant effects and also has cardioprotective effects28,29. Bergenin exhibits antioxidant potential by significantly increasing Nrf-2 levels in the hippocampus30. Compared with ascorbic acid standard, isoquercetin showed higher antioxidant activity with an IC50 value of 3.30 µg/ml31. Isoquercitrin could remove reactive oxygen/nitrogen species and exert antioxidant activity23. Velutin is one of the flavonoids contained in natural plants and has many beneficial activities such as antioxidants32. Tomasin and UDP-L-rhamnose belong to coumarins and its derivatives and organooxygen compounds respectively, while kaempferol-3-glucuronide and kaempferol 3,7,4’-trimethyl ether are classified as flavonoids, the antioxidant properties of these four compounds need to be strengthened.

Adenosine, astragalin, isoquercetrin, and kaempferol-3-glucuronide belong to the same metabolites in five types of hawk tea (Fig. 2B). Of which, high levels of astragalin (relative content 10–28%) and kaempferol-3-glucuronide (7–10.5%) were found in hawk tea, which target TNF and act through twelve pathways, such as hsa05417. Adenosine, another important metabolite in hawk tea, had eight targets of action, namely, MAPK1, EGFR, SRC, CCND1, PRKACA, GAPDH, HSPA8, and EZH2, which involves 20 pathways in Fig. 4B. Artemisinin, bergenin, kaempferol 3,7,4’-trimethyl ether, and velutin were mainly generated from IHT, significantly different from BHT, GHT, OHT, and TCHT. Of these, velutin had a higher BC value and nine targets of action, such as AKT1, PIK3R1, and EGFR, exerting antioxidant activity through these 20 pathways. In addition, isoquercitrin and UDP-L-rhamnose were obtained primarily from GHT, whereas tomasin was found in OHT (Supplementary Fig. S2 online). The action target of isoquercitrin was TNF, whereas the target of UDP-L-rhamnose was HSPA8, which involved the hsa05417, hsa04010, and hsa04915 pathways. As shown in Fig. 4B, tomasin had the highest BC value with 12 action targets, such as MAPK1, PIK3R1, and EGFR, covering these 20 pathways.

According to the DC values in descending order, MAPK1, AKT1, PIK3R1, EGFR, NFKB1, and RELA, etc. were the core targets that were highly correlated with the antioxidant of hawk tea (Fig. 4B). Dietary compounds can activate several cellular kinases, such as MAPK and PI3K, associated with the up-regulation of several antioxidant/detoxifying enzyme activities33. The MAPK and PI3K/AKT pathways are involved in the activation mechanism of the transcription factor Nrf2, which initiates gene transcription via antioxidant response elements34. As another important core target in this study, EGFR involved five key antioxidant compounds (tomasin, velutin, adenosine, kaempferol 3,7,4’-trimethyl ether, and artemisinin) and was associated with 13 pathways. EGFR and EGFR-mediated signaling pathways are essential for a cell’s ability to defend against oxidative stress, and enhanced expression of EGFR increased reactive oxygen species (ROS) scavenging capacity and attenuated ROS-mediated cell death35.

The above results indicated that the antioxidant activity of hawk tea was a synergistic effect of multi-compounds, multi-targets, and multi-pathways. Network pharmacological analyses revealed synergistic multi-component and multi-target properties, such as Chinese Cabernet Sauvignon red wines functioned mainly through eight antioxidant compounds involved in multiple metabolic pathways in vivo36. Generally, consuming whole fruits or vegetables provides more potent effects than individual dietary supplements, owing to the interactions of phytochemicals that co-exist in whole foods37. Several phytochemicals, such as phenolic acids and β-carotene, curcumin and (-)-epicatechin, exhibit remarkable synergistic effects on antioxidant activity38,39.

Molecular docking of the potential antioxidant compounds and core antioxidant targets

Molecular docking is widely used to predict interactions and mechanisms between small molecules and proteins and has been applied to structure-based drug design40. In this study, molecular docking of 11 potential antioxidant compounds and 10 core targets with top-ranked DC values was performed in the CB-Dock2 online service, and their binding energies are shown in Table 2. The results suggested that the 11 potential antioxidant compounds exhibited higher affinity for AKT1, RELA, and MTOR than the other core antioxidant targets. Of these, artemisinin, astragalin, isoquercetrin, and kaempferol-3-glucuronide ranked high in binding energies to AKT1, with − 10.0, -10.0, -11.5, and − 10.9 kcal mol−1, respectively. Those with higher binding energies to RELA were astragalin, isoquercetrin, isoquercitrin, and kaempferol-3-glucuronide, which were − 10.3, -10.2, -10.5, and − 10.4 kcal mol−1, respectively, while the higher binding energies with MTOR were those of isoquercetrin, isoquercitrin, kaempferol-3-glucuronide, and UDP-L-rhamnose with − 11.1, -10.9, -10.0, and − 10.9 kcal mol−1, respectively.

Table 2 Binding energy between critical antioxidant compounds and core antioxidant targets in hawk tea.
Full size table

Lower energy values indicate more stable structures, whereas energies below − 5 kcal mol−1 are usually considered thermodynamically stable41. In particular, binding energies below − 10 kcal mol−1 indicate strong interactions between small molecules and amino acid residues42. Multiple interactions between drug molecules and target binding sites were analyzed using LigPlot + software to evaluate the binding capacity of critical antioxidant compounds to their corresponding targets43. The results revealed that the potential antioxidant metabolites artemisinin, astragalin, isoquercetrin, isoquercitrin, kaempferol-3-glucuronide, and UDP-L-rhamnose interacted strongly with the corresponding core targets (binding energies ≤ -10 kcal mol−1). These antioxidant compounds were all embedded in the active pocket of the target proteins and interacted with several amino acid residues of the core targets in various forms, including hydrophobic interactions and/or the formation of hydrogen bonds (Supplementary Table S2 online). Such as kaempferol-3-glucuronide occupied the hydrophobically active cavity formed by amino acid residues of Lys268, Val270, Trp80, Leu210, Ala212, Leu264, Asp292, Thr82, Gln79 in AKT1 (PDB ID: 3O96) protein, and interacted with Tyr272 (2.93 Å), Val271 (forms two hydrogen bonds, 2.93 Å and 2.75 Å), Thr81 (3.29 Å and 3.15 Å), Thr211 (2.98 Å), and Ser205 (3.07 Å) that bound by forming hydrogen bonds (Fig. 5). Similarly, kaempferol-3-glucuronide formed hydrophobic interactions with amino acid residues of Glu154, Tyr297, Leu153, Arg152, Phe299, Ala73, Val72, Tyr223, Ile249, and Gly74 in RELA (PDB ID: 3QXY) and bound with Tyr285 (2.94 Å and 2.84 Å), Ala222 (2.73 Å), Asp248 (3.11 Å), Leu250 (2.97 Å), Tyr75 (2.85 Å), Asn251 (2.91 Å), and His252 (2.82 Å) that formed hydrogen bonding interactions (Fig. 5). The above results also further validated the reliability of predicting the antioxidant targets of hawk tea based on network pharmacology.

Fig. 5
figure 5

Schematic representation of the molecular docking of kaempferol-3-glucuronide with AKT1 (PDB ID: 3O96) and RELA (PDB ID: 3QXY).

Full size image

Conclusions

In this study, the metabolic profiles, potential antioxidant metabolites, and possible molecular action mechanisms of five types of hawk tea were investigated by integrating untargeted metabolomics, network pharmacology, and molecular docking. Untargeted metabolomics analyses showed that flavonoids were the main components of hawk tea, and differential metabolic profiles among the five types existed. Subsequently, network pharmacology and KEGG enrichment analyses suggested that 11 potential antioxidant metabolites in hawk tea were associated with 25 core targets and may be involved in several metabolic pathways in vivo, including the MAPK and PI3K/AKT signaling pathways, which are closely related to antioxidant activity. Moreover, molecular docking revealed that the potential antioxidant metabolites formed strong interactions with core antioxidant targets through hydrogen bonding and hydrophobic interactions with stable conformations. This work provides a theoretical foundation for further research on the physiological activities of hawk tea. In the future, it is essential to verify the definite action mechanisms of the antioxidant metabolites in hawk tea in vitro and in vivo models.

Materials and methods

Chemicals and reagents

Methanol, acetonitrile, formic acid, and isopropyl alcohol were purchased from ANPEL Laboratory Technologies (Shanghai) Inc., China. All solvents were liquid chromatograph-mass spectrometer (LC-MS) grade. Ultrapure water was prepared in-house using a Milli-Q water purification system (Millipore, Bedford, MA, USA).

Hawk tea materials

Litsea coreana is cultivated on a tea plantation in Meitan County, Zunyi City, Guizhou Province, China (27°39′ N, 107°38′ E), with the same management practices. One bud and two leaves of L. coreana were collected in April 2024, when the tender shoots were green. More than thirty tea trees were used for sampling, and different types of hawk tea (BHT, GHT, IHT, OHT, and TCHT) were prepared following green tea, white tea, oolong tea, black tea, and insect tea processing techniques. Thirty samples, each weighing approximately 100 g, were prepared for each type of hawk tea processing, resulting in 150 samples. Three samples were randomly selected from each processing type, immediately fixed with liquid nitrogen, and stored at -80℃ for further study. All sample preparation was done at Hengyuan Agricultural and Animal Husbandry Co Ltd, Meitan County, Guizhou Province, China. Quality control (QC, n = 3) samples were prepared by mixing equal amounts of each analyzed hawk tea sample.

Non-targeted metabolomics analysis

Sample preparation and extraction

Sample preparation and extraction were performed as described previously with slight modifications44,45. The hawk tea samples were thawed slowly at 4℃. Then, weighed accurately 50 − 100 mg of the sample in a centrifuge tube, and added 1.0 mL of pre-cooled extraction solution (water-acetonitrile-isopropyl alcohol mixed solution, 1∶1∶1, v/v/v). The mixture was homogenized for 60 s and followed by low-temperature ultrasonic extraction for 30 min. After that, it was centrifuged at 12,000 rpm for 10 min at 4℃, followed by one hour of still setting to precipitate the protein at -20℃. The next step involved another centrifugation at 12,000 rpm for 10 min at 4℃. The supernatant solution was then dried in a vacuum, and 0.2 mL of acetonitrile-water solution (3∶7, v/v) was added for reconstitution. After homogenization, the mixture was centrifuged at 14,000 rpm for 15 min at 4℃, and the resulting supernatant was used for detection.

UPLC conditions

The UPLC running parameters and conditions were slightly modified from Li et al.46. The extracted samples were analyzed using a UPLC-Orbitrap-MS system (UPLC, Vanquish; MS, HFX). UPLC conditions: column, Waters HSS T3 (100 × 2.1 mm, 1.8 μm); column temperature, 40℃; flow rate, 0.3 mL/min; injection volume, 2.0 µL; solvent system, phase A: Milli-Q water (0.1% formic acid), phase B: acetonitrile (0.1% formic acid); gradient program, 0 min: phase A/phase B (100∶0, v/v), 1 min: phase A/phase B (100∶0, v/v), 12 min: phase A/phase B (5∶95, v/v), 13 min: phase A/phase B (5∶95, v/v), 13.1 min: phase A/phase B (100∶0, v/v), 17 min: phase A/phase B (100∶0, v/v).

LC-MS/MS analysis

The high-resolution mass spectrometer (HRMS) data were recorded using a Q Exactive HFX Hybrid Quadrupole Orbitrap mass spectrometer equipped with a heated ESI source from Thermo Fisher Scientific. The Full-ms-ddMS2 acquisition method was utilized. ESI source parameters were set as described previously47: sheath gas pressure at 40 Arb, auxiliary gas pressure at 10 Arb, spray voltage at + 3000 V/-2800 V, temperature at 350℃, and ion transport tube temperature at 320℃. Primary scan range (scan m/z range): 70–1050 Da, with a primary resolution of 70,000 and a secondary resolution of 17,500. Data processing was carried out as described by Mu et al.48. The relative content of each compound was determined using peak area normalization.

Hawk tea metabolites analysis

Data analysis and graphing, such as unsupervised principal component analysis (PCA), hierarchical cluster analysis, orthogonal projections to latent structure-discriminant analysis (OPLS-DA), Venn diagrams, and k-means clustering analysis, were conducted using the free platform Metware Cloud (https://cloud.metware.cn). Cluster heat maps were performed using TBtools-II (V1.120) software.

Network pharmacology analysis

Common targets screening for metabolites and antioxidant activity

Several shared metabolites that ranked high in content in five different types of hawk tea, as well as specific and differential metabolites for each kind of tea, were selected for network pharmacology and antioxidant target screening. These compounds were downloaded in Smile format from the Pubchem website (https://pubchem.ncbi.nlm.nih.gov/) and then screened for targets in the SwissTargetPrediction database (http://swisstargetprediction.ch/) using the species “Homo sapiens” and “probability > 0.1” as criteria49. The GeneCards database (https://www.genecards.org/) was used to search for relevant targets by entering the keywords “oxidative” and “antioxidant” respectively. Antioxidant targets were obtained by removing duplicates and selecting genes with a “score > 5”36. The obtained metabolite and antioxidant targets were used to obtain common targets by making a Venn diagram in Venny 2.1.0 (https://bioinfogp.cnb.csic.es/tools/venny/ index.html).

Core targets screening

Common targets were uploaded to STRING 12.0 online server (https://string-db.org/cgi/input.pl), “Organisms” was set to “Homo sapiens” to obtain the protein-protein interaction (PPI) networks, the minimum interaction threshold was set to “highest confidence > 0.9”, and the rest of the settings were set as defaults50. The information on PPI networks was obtained by Cytoscape 3.10.2. Core targets were identified using a topological approach and by the degree centrality (DC), betweenness centrality (BC), and closeness centrality (CC) above their mean values36.

Hawk tea-component-target-pathway-antioxidation network construction

The KEGG enrichment analysis of intersection targets was performed using the Metascape (https://metascape.org/), p-value < 0.01. Hawk tea-component-target- pathway-antioxidation network was constructed using Cytoscape software, selecting the top 20 pathways with the highest number of enriched genes51,52.

Molecular docking

The molecular docking was used to analyze the binding capacity of the active components screened from the hawk tea-component-target-pathway-antioxidation network to the potential antioxidant targets. Active components were downloaded in SDF format files from the PubChem database (http://pubchem.ncbi.nlm.nih.gov). For the target proteins, PDB format files were obtained from the RCSB PDB database (https://www.rcsb.org/). Molecular docking was performed using the CB-Dock2 online service (https://cadd.labshare.cn/cb-dock2/). Docking results were displayed as binding energy (kcal mol−1) and visualized in 3D and 2D using CB-Dock2 and Ligplot 2.2.9, respectively.

Statistical analysis

All assays were performed in triplicate, and the results were presented as mean ± standard deviation. Statistical differences were analyzed using one-way ANOVA and LSD test (p < 0.05) in the Data Processing System (DPS v9.50). Differential metabolites were identified based on their p-value, fold change (FC), and variable importance in projection (VIP). Metabolites with p < 0.05, FC ≥ 2 or ≤ 0.5, and VIP > 1 were considered significant53.