Rapid Characterization and Action Mechanism of the Antidiabetic Effect of Diospyros lotus L Using UHPLC-Q-Exactive Orbitrap MS and Network Pharmacology

Diospyros lotus L, F. Ebenaceae, is an edible fruit that is widely distributed in China and other Asian countries. Presently, Diospyros lotus L can be used to treat patients with diabetes; however, its chemical composition and pharmacological profiles remain to be elucidated. This study investigated the potential bioactive compounds of Diospyros lotus L and their mechanisms of action using LC-MS and network pharmacology analysis. First, the components of Diospyros lotus L were identify using a reliable strategy for UHPLC-Q-Exactive Orbitrap mass spectrometry combined with parallel reaction monitoring (PRM) in the negative ion mode. Second, a network pharmacology study, including target gene prediction and functional enrichment, was applied to screen the main quality markers of Diospyros lotus L and explore its potential mechanism for the treatment of diabetes. The results showed that a total of 159 compounds were identified from Diospyros lotus L, among which, 140 were reported for the first time. Furthermore, 40 active components, such as quercetin, luteolin, and kaempferol, were proposed as active components of Diospyros lotus L for the treatment of diabetes based on network pharmacology analysis. In addition, 92 relevant antidiabetic targets were mainly related to positive regulation of transcription from the RNA polymerase II promoter, extracellular space, and protein binding, suggesting the involvement of TNF, PI3K-Akt, and HIF-1 signaling pathways in the antidiabetic effect of Diospyros lotus L. Our results may provide a useful approach to identify potential active components and molecular mechanisms of Diospyros lotus L for the treatment of diabetes.


Introduction
Diospyros lotus L, a genus of the family Ebenaceae, is an edible fruit that is widely distributed in China and other Asian countries. Diospyros lotus L fruit extract has antidiabetic, antitumor, antinociceptive, and antiinfammatory efects [1][2][3][4] and is used for treating various diseases, such as hypertension, diarrhea, and dry cough. However, to date, few studies have investigated the chemical composition and mechanism of the antidiabetic efect of Diospyros lotus L. remain unknown. Network pharmacology is based on the chemical components of traditional Chinese medicine in the existing database to explore its mechanism from multiple perspectives, such as target gene identifcation and function prediction [11,12].
In this study, a UHPLC-Orbitrap-MS combined with PRM was developed for component identifcation of Diospyros lotus L. Te bioactive ingredients and mechanism of action of Diospyros lotus L on the targets of diabetes were investigated by network pharmacology, which is of great signifcance for further research on Diospyros lotus L.

Standard and Sample
Preparation. Diospyros lotus L (1 g) was extracted with 70% methanol (20 mL) by sonication (1 h). Ten, the extract was centrifuged (15 min, 10°C, 12000 rpm) to obtain the supernatant. Finally, 2 μL of the supernatant was injected into the LC-MS system for analysis.
All the reference standards were accurately weighed using an electronic analytical balance (1 mg) and dissolved in methanol (1 mL). Ten, 10 μL of each standard solution was added to a 1-mL volumetric fask to prepare a mixed standard solution. Te obtained standard solution was stored below 4°C before analysis.
Mass detection was performed on a Q-Exactive Orbitrap MS equipped with an electrospray ionization source operating in negative mode with the following operating parameters: spray voltage at −3.0 kV; sheath gas fow rate at 30 arbs; auxiliary gas fow rate at 10 arbs; capillary temperature at 320°C; heater temperature at 350°C; S-lens RF level at 50; and normalized collision energies at 30%. Te MS spectra were recorded over an m/z range of 80-1000. All data were acquired and processed using Xcalibur software version 4.2.

Candidate Ingredient Screening.
To select the components that have better biological availability in vivo, the components were fltered using the principle of "drug-like soft" in FAFDrugs4. Screening parameters included restriction to molecular weight, logP, and hydrogen bond acceptors (HBA). Details of the physicochemical property flters are listed in Table 1.

Targets of Diospyros lotus L and Diseases.
Te targets of the fltered components were obtained from the traditional Chinese medicine systems pharmacology database and analysis Platform (TCMSP) and predicted using Swiss TargetPrediction (STP). Setting the organism "Homo sapiens" and targets with a probability value greater than 0.1 were considered as potential efective targets for these compounds in the STP database.
Diabetes-related targets were searched in the Online Mendelian Inheritance in Man and GeneCards platform with "diabetes" as the keyword. Te collected targets were amalgamated and duplicated. Potential target genes of Diospyros lotus L therapy for diabetes were obtained through the jvenn intersection.

Construction of Protein-Protein Interaction (PPI)
Network. Te PPI network between target proteins of the related ingredients in Diospyros lotus L and diabetes was obtained by STRING and then imported into Cytoscape software version 3.8.2 to construct and validate a visual network. Te species was set as "Homo sapiens," and the protein interaction was obtained with a medium confdence score of 0.4 to ensure the reliability of our analysis. In the PPI network, topology parameters were calculated to obtain promising candidate targets that were visually characterized by the colors of nodes and to screen remarkable targets.

Enrichment Analysis.
Gene Ontology (GO) and KEGG pathway enrichment analyses were performed using DAVID software. Subsequently, correlated "histograms" and "bubble graphs" were established.

Construction of Active Component-KeyGene-Pathway
Interaction Network. To further explore the mechanism of the antidiabetic efect of Diospyros lotus L, an active component-keygene-pathway interaction network was constructed using Cytoscape 3.9.0 software. In the network, nodes with diferent shapes represented the active compounds, key genes, and related pathways, and an "edge" was an association between the nodes.

Establishment of Qualitative Analysis Strategy.
In this study, an analytical method of UHPLC-Q-Exactive Orbitrap MS combined with the acquisition mode of the PRM mode was used to identify the chemical components of Diospyros lotus L. First, the extraction method and UHPLC-MS conditions of Diospyros lotus L were optimized. Second, the sample was injected into the UHPLC-Q-Exactive Orbitrap MS to obtain high-resolution mass data, including MS and MS 2 . Tird, the compounds were predicted using the Compound Discover version 3.0 workstation with the aid of the metabolism workfow template by adjusting relevant parameters. Finally, the compounds were characterized based on full-scan MS and MS 2 , retention times, standards, and literature.

Optimization of UHPLC-MS Conditions.
To achieve good chromatographic separation, UHPLC parameters were optimized, including the mobile phase (methanol/ water and acetonitrile/water); type and content of acid (acetic acid and formic acid, 0.05%, 0.1%, and 0.2%); column (Waters ACQUITY BEH C18 column, 100 mm × 2.1 mm, 1.7 μm, and HYPERSIL GOLD C18 column, 100 mm × 2.1 mm, 1.9 µm); column temperature (30, 35, and 40°C); fow rate of the mobile phase (0.2, 0.3, and 0.4 mL/min); and the diference gradient of mobile phase.. Te MS parameters, including the fow rate of the sheath gas and auxiliary, temperature of the capillary and auxiliary, heater temperature, spray voltage, and collision energies were examined. In the optimized conditions of UHPLC-Q-Exactive Orbitrap MS, most of the components in the Diospyros lotus L showed efcient separation and parent/daughter ion pairs with high responses.
Compounds 50 and 65 were eluted at 6.07 and 6.53 min, respectively, and possessed the same quasi-molecular ion [M-H] − at m/z 319.0459 and fragment ions at m/z 125.0232, 193.0134, and 151.0026. Tey were tentatively assigned as dihydromyricetin isomers by referring to the literature [22].

Identifcation of Phenylpropanoids in Diospyros lotus L.
Compound 61 possessed the same quasi-molecular ions, and the characteristic fragment ion of compound 42 was characterized as a cafeic acid isomer. Similarly, compounds 95, 104, and 129 were ferulic acid isomers.
Compounds (56, t R 6.22 min, and 63, t R 6.52 min) had the same quasi-molecular ions [M-H] − at m/z 355.1035 and the fragment ion at m/z 193.0500, corresponding to the neutral loss of the glucose group (162 Da) and further generation of the fragment ions of compound 87. Terefore, they were tentatively assigned as ferulic acid acyl-β-Dglucoside isomers [17].  [20], compounds 11 and 22 were 3-methylglutaric acid isomers, compound 27 was syringic acid glucoside [29], compounds 17 and 18 were pantothenic acid isomers, compound 55 was dihydrophaseic acid, compound 120 was azelaic acid, and compounds 130 and 145 were abscisic acid isomers [16,27]. Compound [16]. Compound 49 appeared at a t R of 5.95 min, possessing quasimolecular ions at m/z 329.0878 and the main fragment ion at m/z 167.0341 owing to the loss of a glucose residue (162 Da), which was characterized as vanillic acid glucoside [32]. Similarly, compounds 34 and 54 were confrmed as vanillic acid-O-rutinosides.
Compound 94 at m/z 137.0244 with the molecular formula C 7 H 6 O 3 and appearing at a t R of 7.58 min was suggested to be p-hydroxybenzoic acid based on the MS 2 data [21]. Compounds 23,24,25,26,31, and 32 (t R 2.65, 2.85, 2.89, 3.00, 3.53, and 3.75 min, respectively) had the same quasi-molecular ion [M-H] − at m/z 299.0772 and the characteristic fragment ion at m/z 137.0233 based on the neutral loss of a glucose residue (162 Da). Tey were tentatively characterized as p-hydroxybenzoic acid-O-glucoside isomers [33].

Active Components and Related Targets.
Te active compounds were selected by using the "drug-like soft" in FAFDrugs4 with the criteria of 100 ≤ MW ≤ 600, −3 ≤ logP ≤ 6, and HBA ≤ 12. Eventually, a total of 40 compounds were screened (Supplementary Table 1). Combined with TCMSP and STP database search and prediction, 445 component targets were obtained after removing duplicate targets. Furthermore, 521 diabetes-related targets were identifed by screening the disease-target database. Finally, 92 overlapping genes of compound targets and diabetes-related targets were regarded as potential targets of Diospyros lotus L for the treatment of diabetes ( Figure 2).

PPI Network of Overlapping
Genes. Te PPI network graph was obtained by importing 92 overlapping targets into STRING and removing one disconnected point. Tere were 91 nodes and 1488 edges; the average number of nodes was 32.8, and the average local clustering coefcient was 0.703. TSV data were downloaded and imported into Cytoscape 3.9.0 software to show the protein interaction network.
Te results are shown in Figure 3, where the node size is positively correlated with the degree value and the lines represent interactions. As betweenness centrality increases, the color of the node changes from yellow to turquoise. Degree and betweenness centrality indicate the importance of the targets. Te target whose degree value was greater than the average value was considered the key target.

Enrichment Analysis.
Te key targets were further analyzed by functional association clustering to integrate functional genomics annotations of the most important cluster of targets and pathways, which facilitates further understanding of the mechanism of the antidiabetic efect of Diospyros lotus L.
As shown in Figure 4, the most representative GO-BP terms were "positive regulation of transcription from RNA polymerase II promoter" and "infammatory response," whereas the most representative GO-CC terms were "extracellular space," "extracellular region," and "plasma membrane." Te most representative GO-MF terms were "protein binding" and "enzyme binding." Te two most representative KEGG pathways ( Figure 5) were the "MAPK signaling pathway" and "AGE-RAGE signaling pathway." After exclusion of broad pathways, 47 core common target genes were mainly related to the TNF, PI3K-Akt, HIF-1, NAFLD, toll-like receptor, and other multiple signaling pathways. Tis suggests that the efect of Diospyros lotus L on diabetes may involve multiple pathways as well as complex interactions among these pathways.

Active Component-KeyGene-Pathway
Interaction Network Analysis. As shown in Figure 6, the active componentkeygene-pathway interaction network contained 104 nodes (47 key genes, 37 active components, and 20 KEGG pathways (top 20)) and 410 edges. In the network, the diamond, oval, and elliptical nodes correspond to diferent active compounds, pathways, and targets, respectively. Te degrees of quercetin, luteolin, kaempferol, TNF signaling pathway, PI3K-Akt signaling pathway, HIF-1 signaling pathway, PTGS2, AKT1, IL6, and TNF were 34,20,18,12,11,10,25,21,20, and 17, respectively. Te average degrees of the diamond and elliptical nodes were 5.93°and 8.72°, respectively. In addition, at least nine genes were potentially involved in each diabetes-related pathway, suggesting that one active component can potentially target multiple genes and have the action characteristics of multiple active compounds, targets, and pathways of Diospyros lotus L in the treatment of diabetes. Quercetin has many antihyperglycemic efects, such as enhancing insulin sensitivity, promoting glycogen synthesis, inhibiting α-glucosidase activity, and improving insulin resistance [34]. Luteolin can play an antioxidant role by enhancing the activity of superoxide dismutase in microvascular lesions in diabetes [35]. Kaempferol is a favonoid compound that plays an active role in the prevention and treatment of diabetes and has antiinfammatory and antioxidant properties. It can reduce oxidative stress and infammation through the MAPK pathway to alleviate myocardial ischemia-reperfusion injury in diabetic rats [36]. Myricetin can enhance the antioxidant defense system in mice, increase insulin secretion, substantially reduce blood glucose levels, and efectively protect the liver and kidney from oxidative damage in diabetic mice [37,38]. IL-6 interferes with the insulin signaling pathway and promotes apoptosis of pancreatic β-cells, which promotes insulin resistance in multiple organs through a variety of infammatory signaling pathways [39]. TNF is one of the cytokines constituting the acute infammatory response, which can trigger the MAPK and NF-κB pathways, leading to insulin resistance [40].

Conclusion
In this study, an integrated approach combining UHPLC-Q-Exactive Orbitrap MS and network pharmacology analysis was adopted to explore the potential active ingredients and ameliorative mechanisms of Diospyros lotus L against hyperglycemia. Eventually, 159 compounds were identifed in Diospyros lotus L (140 of which were reported for the frst time). According to the results of the active components and key gene-pathway interaction network, the antihyperglycemic efect of Diospyros lotus L is attributed to quercetin, luteolin, kaempferol, myricetin, and dihydromyricetin, which act on PTGS2, AKT1, IL6, TNF, and MMP9 and participate in the TNF, PI3K-Akt, and HIF-1 signaling pathways, as well as NAFLD. In conclusion, the integrated approach combining UHPLC-Q-Exactive Orbitrap MS and network pharmacology analysis provided insights into the potential active ingredients and ameliorative mechanism of Diospyros lotus L on hyperglycemia.

Data Availability
Te data used to support the fnding of this study are available from the corresponding author upon request.

Disclosure
Te funder has no role/infuence in this study.

Conflicts of Interest
Te authors declare that there are no conficts of interest.