Integrated UPLC/Q-TOF-MS/MS Analysis and Network Pharmacology to Reveal the Neuroprotective Mechanisms and Potential Pharmacological Ingredients of Aurantii Fructus Immaturus and Aurantii Fructus

Aurantii Fructus (AF) and Aurantii Fructus Immaturus (AFI) have been used for thousands of years as traditional Chinese medicine (TCM) with sedative effects. Modern studies have shown that Citrus plants also have protective effects on the nervous system. However, the effective substances and mechanisms of action in Citrus TCMs still remain unclear. In order to explore the pharmacodynamic profiles of identified substances and the action mechanism of these herbs, a comprehensive approach combining ultra-high-performance liquid chromatography with quadrupole time-of-flight mass spectrometry (UPLC/Q-TOF-MS/MS) analysis and network pharmacology was employed. Firstly, UNIFI 2.1.1 software was used to identify the chemical characteristics of AF and AFI. Secondly, the SwissTargetPrediction database was used to predict the targets of chemical components in AF and AFI. Targets for neuroprotection were also collected from GeneCards: The Human Gene Database (GeneCards-Human Genes|Gene Database|Gene Search). The networks between targets and compounds or diseases were then constructed using Cytoscape 3.9.1. Finally, the Annotation, Visualization and Integrated Discovery Database (DAVID) (DAVID Functional Annotation Bioinformatics Microarray Analysis) was used for GO and pathway enrichment analysis. The results showed that 50 of 188 compounds in AF and AFI may have neuroprotective biological activities. These activities are associated with the regulatory effects of related components on 146 important signaling pathways, derived from the KEGG (KEGG: Kyoto Encyclopedia of Genes and Genomes), such as neurodegeneration (hsa05022), the Alzheimer’s disease pathway (hsa05010), the NF-kappa B signaling pathway (hsa04064), the hypoxia-inducible factor (HIF)-1 signaling pathway (hsa04066), apoptosis (hsa04210), the epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor resistance signaling pathway (hsa01521), and others, by targeting 108 proteins, including xanthine dehydrogenase (XDH), glutamate ionotropic receptor NMDA type subunit 2B (GRIN2B), and glucose-6-phosphate dehydrogenase (G6PD), among others. These targets are thought to be related to inflammation, neural function and cell growth.


Introduction
Aurantii Fructus (AF) and Aurantii Fructus Immaturus (AFI) have been used in traditional Chinese medicine (TCM) for thousands of years [1].AF and AFI are the fruits of Citrus aurantium L. (CA) (bitter orange) and their cultivated varieties [2].And Citrus aurantium L. Cv. daidai (CAD) is the most commonly used cultivated variety of Citrus aurantium L. and is widely grown as a medicinal plant [3].AF and AFI are collected at different stages of fruit growth with diverse clinical efficacy; the effect of AFI on promoting qi is obviously better than that of AF, and they are thus are recorded in the Chinese Pharmacopoeia as two distinct medicinal materials [4].According to TCM theory, AF and AFI each have their own unique clinical applications [5].Although AF and AFI have common effects of regulating visceral functions [6], AF is always used to alleviate chest pain and improve gastrointestinal functions, such as alleviating dyspepsia in a gentle yet efficient manner [7].AFI, compared to AF, expresses a more rapid and robust method of action and is often employed to disperse severe abdominal distention and to eliminate phlegm [8].We found that Citrus plants, including Citrus aurantium L., have beneficial effects on those with neurodegenerative diseases [9], suggesting AF and AFI to have potential protective effects on the nervous system.Therefore, it is reasonable to explore the protective effects of AF and AFI on nervous system.Currently, excitotoxicity and oxidative stress are recognized as two important aspects of nervous system damage [10].Hence, we believe that it is meaningful to study the chemical components related to excitotoxicity and oxidative stress in AF and AFI.At present, chemical analysis methods, including chromatography [11], nuclear magnetic resonance (NMR) spectroscopy [12], and mass spectrometry (MS) [13], are usually used to study the chemical constituents of plant drugs.Among them, ultra-highperformance liquid chromatography (UPLC) alongside high-resolution mass spectrometry (HR-MS) can simultaneously detect a variety of chemical components in plant drugs [14]; however, to obtain accurate identification results, the UPLC-HR-MS detection results must be compared with the standard chromatogram of chemical components or the mass spectrometry database [15].As an auxiliary mass spectrum analysis software, UNIFI supports multi-user, server-based workgroups to complete liquid chromatography (LC), LC/MS, and LC/MS/MS data collection, storage, management, mining, and sharing, which can greatly improve collaboration efficiency [16,17].
In this study, the chemical compositions of AF and AFI derived from Citrus aurantium L. and Citrus aurantium L. Cv. daidai were systematically evaluated with UNIFI software with UPLC/quadrupole time-of-flight (Q-TOF)-MS/MS.The chemical similarities and differences between AF and AFI were summarized.Furthermore, the target of compounds and the target of neuroprotection were predicted using the method of network pharmacology [18].Finally, identifying bioactive compounds, potential targets, and signaling pathways relevant to the neuroprotection with AF and AFI was realized using an integrative network analysis [19].

Identification of Compounds in AF and AFI
The total ion chromatograms of AF and AFI in both positive and negative ion modes are presented in Figure 1A-F.The process of identifying compounds via UNIFI software is shown in Figure 1G.The retention times and the MS data of the characterized compounds are summarized in Table 1.A total of 188 compounds were identified by UNIFI software based on the self-built database.Among these compounds, compounds (46, 47, 92, 106, 119, 130, 154) were unambiguously identified by comparison with reference compounds.
software based on the self-built database.Among these compounds, compounds (46, 47, 92, 106, 119, 130, 154) were unambiguously identified by comparison with reference compounds.Using the SwissTargetPrediction databases, we obtained the 9021 target proteins of the 188 compounds in AFI and AF.The entire list of targets of each compound is provided in Supplementary Table S2.After removing redundancy, we identified 1052 AFI-or AFassociated targets (Supplementary Table S3).Compound-target networks were constructed on the basis of compounds 1 (7-Hydroxycoumarin), 6 (Limonin), 46 ((+/−)-Naringenin), 61 (Helenalin), and 63 (Kaempferol) and their corresponding targets, as shown in Figure 2. The round, yellow nodes and round, blue nodes represent the compounds and targets, respectively, and the edges represent the interactions between compounds and targets.

Identification of the Neuroprotective Targets and Analysis of the "Disease-Target" Network
By means of the available resource, namely, the GeneCards: The Human Gene Database.we obtained 151 excitotoxicity-associated targets (relevance > 1.0) and 187 antioxidant-associated targets (relevance > 1.0).And detailed information on the collected targets is provided in Supplementary Table S4 (excitotoxicity-associated targets) and Supplementary Table S5 (antioxidant-associated).Disease-target networks were constructed, as shown in Figure 3.The network consisted of two parts (A: an excitotoxicity-associated target network with 151 nodes; B: an antioxidation target network with 187 nodes).The round, blue nodes and round, yellow nodes represent the targets and diseases, respectively, and the edges represent the interactions between diseases and targets.
Naringenin), 61 (Helenalin), and 63 (Kaempferol) and their corresponding targ shown in Figure 2. The round, yellow nodes and round, blue nodes represent th pounds and targets, respectively, and the edges represent the interactions betwee pounds and targets.

Identification of the Neuroprotective Targets and Analysis of the "Disease-Target" Ne
By means of the available resource, namely, the GeneCards: The Human Gen base.we obtained 151 excitotoxicity-associated targets (relevance > 1.0) and 187 a dant-associated targets (relevance > 1.0).And detailed information on the collected is provided in Supplementary Table S4 (excitotoxicity-associated targets) and Supp tary Table S5 (antioxidant-associated).Disease-target networks were construc shown in Figure 3.The network consisted of two parts (A: an excitotoxicity-asso target network with 151 nodes; B: an antioxidation target network with 187 node round, blue nodes and round, yellow nodes represent the targets and diseases, tively, and the edges represent the interactions between diseases and targets.

Identification of the Neuroprotective Targets and Analysis of the "Disease-Target" Network
By means of the available resource, namely, the GeneCards: The Human Gene Database.we obtained 151 excitotoxicity-associated targets (relevance > 1.0) and 187 antioxidant-associated targets (relevance > 1.0).And detailed information on the collected targets is provided in Supplementary Table S4 (excitotoxicity-associated targets) and Supplementary Table S5 (antioxidant-associated).Disease-target networks were constructed, as shown in Figure 3.The network consisted of two parts (A: an excitotoxicity-associated target network with 151 nodes; B: an antioxidation target network with 187 nodes).The round, blue nodes and round, yellow nodes represent the targets and diseases, respectively, and the edges represent the interactions between diseases and targets.

Recognition of the Candidate Compounds and Potential Targets and Analysis of the "Compound-Disease-Target" Network
A total of 125 overlapping protein targets were recognized, and 50 candidate compounds were obtained, as described in Supplementary Table S6. Figure 4 shows the compound-disease-target network, which was composed of one hundred and seventyseven nodes (one hundred and twenty-five targets, fifty compounds, and two diseases) and two hundred and fifty edges.The round, yellow nodes, round, red nodes, and green nodes represent the compounds, targets, and diseases, respectively, and each node size is proportional to its degree.The edges represent the interactions between any two types of nodes.The results showed that the 50 compounds and 125 targets may be the candidate bio-active substances and the potential pharmacological targets for neuroprotection of AF and AFI.In particular, the neuroprotective candidate compounds are shown in Table 2 and Figure 5, and the potential pharmacological targets are shown in Table 3.There are significant differences in the chemical composition of AF and AFI [2], and we found that the neuroprotective effects of the compounds of AF and AFI are less different, as shown in Figure 5. Limonin in Table 2 is present in four samples, and studies have shown that it has a neuroprotective effect [20].1), targets and diseases, respectively, and a node's size is proportional to its degree.The edges represent the interactions between any two nodes.1), targets and diseases, respectively, and a node's size is proportional to its degree.The edges represent the interactions between any two nodes.

No. Compound Name AFI-CAD AF-CAD AFI-CA AF-CA No. Compound Name AFI-CAD AF-CAD AFI-CA AF-CA
Compound-disease-target network.The yellow, red, and green nodes represent the compounds (the numbers represent the serial numbers of the compounds in Table 1), targets and diseases, respectively, and a node's size is proportional to its degree.The edges represent the interactions between any two nodes.

GO and Pathway Enrichment Analyses of Potential Targets
One of the functions of GO processes is to predict genes related to a disease [21].GO and pathway enrichment analyses of the 108 potential targets for neuroprotection in AF and AFI were performed using the DAVID database to understand the relationships between functional units and their underlying significance in the biological system networks [22].All of the biological processes and pathways were extracted (p ≤ 0.05).Figure 6 lists the top 30 most significantly enriched GOBP terms.Supplementary Tables S7 and S8 provide detailed information about the biological processes and signaling pathways.In total, 146 related pathways were identified, including pathways of neurodegeneration (hsa05022), the Alzheimer's disease pathway (hsa05010), the NF-kappa B signaling pathway (hsa04064), the HIF-1 signaling pathway (hsa04066), apoptosis (hsa04210), and the EGFR tyrosine kinase inhibitor resistance signaling pathway (hsa01521).And numerous targets were involved in the memory process, gene expression, the rhythmic process, the neuron apoptotic process, and the apoptotic process.range, 100 m/z to 1200 m/z; positive adducts including H + , Na + , and K + ; and negative adducts containing H − , HCOO − , and Cl − .Finally, using the MassLynx workstation, the above identification results were reviewed in combination with the precise mass of excimer ions, retention time, fragment ion information, and the literature [17].According to our study (Section 3.1), all of the compounds in AFI and AF were chosen to predict the biological targets.The canonical SMILES [24] of the compounds were uploaded into the SwissTargetPrediction database (http://www.swisstargetprediction.ch/ accessed on 31 January 2024) to obtain the UniProt IDs for predicting targets [25].

Screening Candidate Compounds and Potential Targets
We selected the overlapping targets of AF and AFI for neuroprotection and used the compounds corresponding to these targets as candidate compounds.

Gene Ontology (GO) and Pathway Enrichment of Potential Targets
The Gene Ontology (GO) biological process (BP) was analyzed to further validate whether the potential targets were indeed matched for neuroprotection [29].GO and Kyoto Encyclopedia of Genes and Genomes (KEGG) [30] signaling pathway analyses were carried out using the Database for Annotation, Visualization and Integrated Discovery (DAVID) (https://david.ncifcrf.gov/,accessed on 31 January 2024, version v2023q1).A p-value ≤ 0.05 was considered significant.

Constructing the Network of Compounds, Diseases, and Targets
To comprehensively understand the neuroprotection of AF and AFI, the compoundtarget and disease-target networks were constructed using Cytoscape 3.9.1 (Bethesda, MD, USA) [31].In these networks, the nodes represented the compounds, diseases, targets, or signaling pathways, and the edges represented their interactions [32].

Conclusions
In this study, a comprehensive method combining UPLC/Q-TOF-MS/MS analysis and network pharmacology was used to reveal the differences in the chemical components of AF and AFI that applied to their neuroprotective effects.The results indicated that 50 of the 188 compounds in AF and AFI may be bioactive, which may be related to their targeting of 108 targets such as XDH, GRIN2B, AKT1, PRKCG, CAPN1, CSNK2A1, G6PD.One hundred and forty-six important signaling pathways were implicated, including neurodegeneration (hsa05022), the Alzheimer's disease pathway (hsa05010), the NF-kappa B signaling pathway (hsa04064), the HIF-1 signaling pathway (hsa04066), apoptosis (hsa04210), and the EGFR tyrosine kinase inhibitor resistance signaling pathway (hsa01521).These findings fully reflect the multi-component, multi-target, and multi-approach characteristics of TCM in disease treatment.This study shows that AF and AFI have great potential in neuroprotection, and their neuroprotective effects deserve further study.
In some network pharmacological studies, compounds are collected indiscriminately from databases; however, this can produce false-positive results.The method we applied in this research was built on the basis of experimentally identified components and corre-sponding targets, which will greatly reduce the prediction range and improve the accuracy of the prediction results.However, further pharmacological experiments are needed to verify its main biological components and related targets, so as to deeply understand the neuroprotective mechanism of AF and AFI, which will be the direction of our further research.

Figure 1 .
Figure 1.Identification of compounds in AF and AFI.(A) AF-CA; (B) AFI-CA; (C) AF-CAD; (D) AFI-CAD; (E) total ion chromatography of samples in positive ion mode; (F) total ion chromatography of samples in negative ion mode; (G) the identification process of compounds in UNIFI software: No. 142 compound identified as Gardenin A.

Figure 1 .
Figure 1.Identification of compounds in AF and AFI.(A) AF-CA; (B) AFI-CA; (C) AF-CAD; (D) AFI-CAD; (E) total ion chromatography of samples in positive ion mode; (F) total ion chromatography of samples in negative ion mode; (G) the identification process of compounds in UNIFI software: No. 142 compound identified as Gardenin A.

Figure 4 .
Figure 4. Compound-disease-target network.The yellow, red, and green nodes represent the compounds (the numbers represent the serial numbers of the compounds in Table1), targets and diseases, respectively, and a node's size is proportional to its degree.The edges represent the interactions between any two nodes.

Figure 4 .
Figure 4. Compound-disease-target network.The yellow, red, and green nodes represent the compounds (the numbers represent the serial numbers of the compounds in Table1), targets and diseases, respectively, and a node's size is proportional to its degree.The edges represent the interactions between any two nodes.

Figure 6 .
Figure 6.The top 30 enriched gene ontology terms for the biological processes of potential targets.

Figure 6 .
Figure 6.The top 30 enriched gene ontology terms for the biological processes of potential targets.

3.
Materials and Methods 3.1.Experimental Compounds Discovery 3.1.1.Chemicals and Materials AF-CA and AFI-CA (batch number: S202108-0932, S202101-0929) were collected from Xinyu County, Jiangxi Province, China.AF-CAD and AFI-CAD (batch number: S202108-0933, S202106-0930) were collected from Jinhua County, Zhejiang Province, China.And all samples were stored at room temperature until experimentation.All collected samples have accompanying voucher specimens held in the National Engineering Research Center of TCM Standardization Technology, Shanghai Institute of Materia Medica (Shanghai Institute

3. 2 .
Target Prediction of the Compounds in AFI and AF and Neuroprotective Target Collection 3.2.1.Predicting Targets of Compounds in AFI and AF

Table 1 .
Compounds identified in AF and AFI by UNIFI software.

Table 1 .
Compounds identified in AF and AFI by UNIFI software.
2.2.Identification of the AFI-and AF-Associated Targets and Analysis of the "Compound-Target" Network

Table 2 .
Neuroprotective candidate compounds in AF and AFI. No.

Table 3 .
The potential neuroprotective pharmacological targets of AF and AFI.