Network pharmacology-based analysis on three amicoumacintype isocoumarin compounds from an endophytic bacterium in Houttuynia cordata

Yiyao Lu,Yinjie Xiong,Mengyue Zhang,Yingying Wu,Jingjing Xu,Linyu Zhang,Xin Wu

School of life science and biopharmaceutics,Shenyang Pharmaceutical University,Shenyang 110016,China

Abstract Amicoumacin isocoumarin-type compounds are mainly derived from the secondary metabolites of Bacillus bacteria,and have potential biological activities such as antibacterial,anti-tumor,anti-inflammatory,and anti-ulcer.In our previous research,three amicoumacin-type isocoumarin compounds,Paenicoumacin A,Paenicoumacin B,and Paenicoumacin C (B and C are novel),were isolated from an endophytic bacterium in Houttuynia cordata.In this study,network pharmacology and bioinformatics techniques were used to analyze the potential activities of amicoumacin compounds in the treatment of liver cancer.These compounds were found to mainly regulate cellular oxidative stress,cancer process,nucleic acid metabolism,nutritional metabolism and other pathways.The findings provide useful information for subsequent experimental verification.

Keywords:Houttuynia cordata;Paenibacillus;isocoumarin;liver cancer;network pharmacology

1 Introduction

Amicoumacin-type isocoumarin compounds are 3,4-dihydroisocoumarin derivatives with unique dihydroxy-amino acid side chains.They are mainly derived from Bacillus bacteria,such as

Bacillus pumilus

and

Bacillus subtilis

,and also isolated from secondary metabolites of

Xanthomonas

,

Streptomyces

and marine fungi [1].Studies have found that amicoumacin-type isocoumarin compounds have a wide range of biological activities,such as antibacterial [2],anti-tumor [3],antiinflammatory [4],anti-ulcer [5],and anti-insect [6].Therefore,this type of compound is potential drug,and can be used as the lead compounds for new antibiotics or innovative anti-ulcer drugs.Liver cancer is a common disease,usually caused by long-term accumulation of chronic hepatitis,such as hepatitis B viral (HBV),hepatitis C viral (HCV) infection,alcoholic or non-alcoholic liver disease [7,8].Studies have shown that amicoumacin compounds induce cancer cell death by targeting eukaryotic ribosomes.Experiments have shown that,compared with non-cancer cell lines,these compounds have higher sensitivity to human cancer cells [9].Therefore,amicoumacin isocoumarin-type compounds obtained from the secondary metabolites of endophytic bacterium

Paenibacillus

can be used as potential anti-cancer molecules.

The concept of network pharmacology was first proposed by Andrew L Hopkins in 2008.Based on the theory of systems biology,network pharmacology uses bioinformatics technology to predict multiple targets and multiple pathways of active ingredients of drugs [10].Since cancer is characterized by complex pathological mechanisms and the active components of drugs acting on a single target or a single pathway are no longer suitable for establishing new anti-cancer drug models,examples of cancer drug development based on network pharmacology have been documented in the literature [11].Network pharmacology analyzes the network relationship between active molecules and potential targets based on biological functions,and multi-functional and multi-channel characteristics of the potential drug leads indicate the pharmacodynamic value [12,13].Compared with the drug-single target model,this model considers drug activity and toxicity more comprehensively [10].

In our previous research,the active components,Paenicoumacin A,Paenicoumacin B,and Paenicoumacin C have been isolated from the secondary metabolites of

Paenibacillus

sp.Xy-2,an endophytic bacterium in

Houttuynia cordata

.In this study,network pharmacology and bioinformatics techniques were used to analyze the potential activities of these three amicoumacin compounds in the treatment of liver cancer,aiming to provide useful information of the obtained target,gene function,and signal pathway for subsequent experimental verification.

2 Materials and methods

2.1 Strain identification and separation and extraction of secondary metabolites

The plant endophytic bacteria

Paenibacillus

sp.Xy-2 was isolated from the root soil of

Houttuynia cordata

in QiPan Mountains of Shenyang [14].The strain Xy-2 was separated and purified on LB solid medium.The colony was white and the surface was sticky and shiny.The functional genes of

Paenibacillus

sp.

Xy-2 were screened by Polymerase Chain Reaction (PCR)amplification technology,and found to have the potential to produce amicoumacin-type isocoumarin compounds.Since this strain was the first Bacillus producing amicoumacin-type isocoumarin compounds,it was selected for research.According to the previous research,the Blast homologous sequence was aligned based on the ITSrDNA sequence of the strain [15] and the homology with

Paenibacillus

was found to be 99%.Therefore,the plant endophytic bacteria Xy-2 was identified as

Paenibacillus

sp.The fermentation and cultivation of

Paebibacillus

sp.Xy-2 and the separation and purification of secondary metabolites were performed in accordance with previous literature reports [15].

2.2 Bioactivity target prediction

To predict the active targets of Paenicoumacin A,Paenicoumacin B,Paenicoumacin C,we used Pubchem website (https://pubchem.ncbi.nlm.nih.gov/) and Chemdraw 19.0.0.22 to obtain 2D and 3D structures of the compounds,and imported target prediction database PharmMapper 2017 (http://www.lilab-ecust.cn/pharmmapper/) in sdf format [16-18].PharmMapper was used to screen the potential targets based on pharmacophore features,and then the target prediction information was obtained on March 15,2021.The keyword“liver cancer”was used to obtain liver disease targets from the disease database GeneCard 5.5 (https://www.genecards.org/).After sing UniProKB search function in the Uniprot database (https://www.uniprot.org/),entering the name of the target protein,and limiting the species to“Homo sapiens”,the target protein name was obtained and converted into a gene symbol.

2.3 Network construction

2.3.1 Construction of strain-compound-target network

According to ranking of the correction score in the potential target prediction information of the PharmMapper database,the first 10% targets were selected for subsequent analysis and straincompound-target network construction.Network relationship files were prepared and attributed separately.These files were then imported into Cytoscape 3.7.2 in csv format,the format between nodes was adjusted,and the network diagrams were finally made.

2.3.2 Construction of target interaction network

The intersection targets of the three new compounds,Paenicoumacin A,Paenicoumacin B,and Paenicoumacin C,and liver cancer disease were obtained through the Venn diagram 2.1.0,(https://bioinfogp.cnb.csic.es/tools/venny/).The String website 11.5 (https://string-db.org/) was used for PPI analysis,and the results were imported into Cytoscape 3.7.2 to construct the interaction relationship network between the targets and core role target network diagram.The size and color of the nodes in the network reflect the degree value,and the thickness of the edge reflects the combined score.

2.4 Analysis of the core target GO and KEGG enrichment pathway

The Metascape website (http://metascape.org/gp/index.html#/main/step1) was used to perform Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis on the core targets of the liver cancer and the three amicoumacin-type isocoumarin compounds isolated from the secondary metabolites of strain Xy-2 pathway.Enrichment analysis was carried out on the ontology sources,namely GO molecular functions,GO biological processes,GO cell composition and KEGG pathway.All genes in the genome were used as the enrichment background.Terms with

P

<0.01,minimum count of 3,and enrichment factor >1.5 (enrichment factor is the ratio of the observed count to the count expected by chance) were collected and grouped into clusters based on their membership similarities.

P

-value was calculated based on the accumulative hypergeometric distribution [19],and

q

-value was calculated using the Banjamini-Hochberg procedure to account for multiple testings [20].Kappa score [21] was used as the similarity metric when hierachical clustering was performed on the enriched terms,and sub-trees with similarity >0.3 formed a cluster.The most statistically significant term within a cluster was chosen to represent the cluster.

3 Results and discussion

3.1 Paebibacillus sp.Xy-2 fermentation product acquisition

In the early stage,the research group has fermented the strain Xy-2 using the original fermentation conditions and separated two amicoumacin-type isocoumarin compounds from its secondary metabolites,namely PM-94128 and paenicoumacin A.

In order to isolate more novel amicoumacintype compounds,the OSMAC strategy was used in this experiment to activate the silent genes in strain Xy-2.It was found that changes in culture conditions could affect the strain’s ability to produce amicoumacin-type isocoumarins.After repeated explorations,the new medium TSA was finally selected as the fermentation conditions of strain Xy-2 in this experiment.

After changing the culture conditions,TSA was finally used as the liquid fermentation medium of strain Xy-2.With silica gel column chromatography,gel column chromatography,ODS open column chromatography,high performance liquid chromatography and other separation methods,and directional tracking separation by thin layer chromatography and HPLC-DAD,5 were finally isolated from the strain

Paebibacillus

sp.Xy-2

.

Four of them are amicoumacintype isocoumarin compounds,including two previously isolated compounds,except PM-94128 (

1

) and Paenicoumacin A (

2

),and two newly isolated compounds,Paenicoumacin B (

3

)and Paenicoumacin C (

4

);One is another type of isocoumarin compound,

N

-acetyltyramine (

5

) (Fig.1).Based on the physical and chemical properties and spectral data,the structures of the above five compounds were identified [15].This study focused on the three amicoumacin-type isocoumarin compounds,Paenicoumacin A,Paenicoumacin B and Paenicoumacin C.

Fig.1 Isocoumarin compounds obtained from the TSA fermentation broth of Paebibacillus sp.Xy-2

3.2 Compound potential target prediction

The potential targets of the three compounds,Paenicoumacin A,Paenicoumacin B,and Paenicoumacin C,were obtained from the online target prediction website PharmMapper 2017.The first 10% of potential targets were selected and Cytoscape 7.3.2 was used to construct

Paebibacillus

sp.Xy-2-compound-target network(Fig.2).Diamonds and hexagons represent target strains and target compounds,respectively,while circles represent their respective potential target genes.Triangles in Fig.2 represent the intersection potential targets of Paenicoumacin A,B,and C.In addition,16,730 related targets of liver cancer disease were obtained using the keyword“liver cancer”in the disease data website GeneCard 5.5.

Fig.2 Paebibacillus sp.Xy-2-compound-target network

3.3 PPI analysis of the common target of compound and liver cancer

Protein target interaction and related functional analysis were the conditions for treating cells as a whole system to regulate physiological mechanisms [22].PPI analysis provided the basis for subsequent research on biological function and biological process.The functions of many proteins were manifested in the interaction and regulation with other proteins [22].The core targets obtained from PPI analysis of the common targets of the compound and liver cancer are shown in Table 1.

Table 1 The core target information

Continued table 1

There are 105 intersecting targets of Paenicoumacin A and liver cancer.PPI results were obtained using the string website,and PPI network diagram was constructed using Cytoscape 3.7.2.Topological analysis was performed according to the core parameter degree value (degree) that evaluates the importance of network nodes in the research of network pharmacology.Network indicates that many components act on the target.The size of the displayed graphics reflects the core degree of the corresponding target,and the same color represents the same number of targets in the intersection of the PPI analysis.PPI analysis results of Paenicoumacin B and C presented similar forms.Among them,the target cellular tumor antigen p53 (TP53) possesses the most relations to other targets,with a total of 25 action edges,followed by the targets catalase(CAT),cytochrome c (CYCS),notch receptor 1 (NOTCH1),superoxide dismutase 2 (SOD2),isocitrate dehydrogenase 1 (IDH1),and annexin A5 (ANXA5),and the higher scores belong to the targets CAT-SOD2,PARP1-POLB,PARP1-TP53,and CAT-HSD17B4.The PPI network diagram of the core target is shown in Fig.3.

Fig.3 PPI analysis of the intersection target between Paenicoumacin A and liver cancer

TP53

is an important tumor suppressor gene,which encodes P53 protein and exists on chromosome 17p13.1 [23],It is a common mutant gene in Asian and African races [24].p53 protein is involved in inhibiting tumor cell division,inducing apoptosis and repairing DNA damage [25].Chronic inflammation and intracellular oxygen or nitrogen metabolites may stimulate DNA damage and P53 gene mutations in liver cells [26].According to the literature,about 60% of TP30 loses functional activity in liver cancer diseases [27].A small heme protein encoded by

CYCS

plays a role in the mitochondrial electron transport chain,binds to the inner mitochondrial membrane,accepts electrons from cytochrome b and transfers it to the cytochrome oxidase complex.In addition,this protein is also involved in regulating cell apoptosis.

NOTCH1

encodes one of the NOTCH protein family,and the conservative Notch signaling pathway regulates the interaction between adjacent cells through the binding of this protein family receptor and its corresponding ligand.The

SOD2

gene,as a member of the iron/manganese superoxide dismutase family,encodes a mitochondrial protein.Studies have shown that,compared with normal liver tissues adjacent to hepatocellular carcinoma,circular RNA from the

SOD2

gene is expressed higher in tumor tissues [28],and this may be related to the activation of the JAK2/STAT3 signaling pathway.A new type of signaling pathway axis SOD2/miR-5025p/DNMT3a/JAK2/STAT3/circSOD2 provides reference for revealing the mechanism of hepatocellular carcinoma [29].

There are 67 intersecting targets between Paenicoumacin B and liver cancer.The target albumin (ALB) has the most relations to other targets,with a total of 11 action edges,followed by the CAT,proliferating cell nuclear antigen (PCNA),aminolevulinate dehydratase (ALAD) and so on(Fig.4).There are 68 intersecting targets between Paenicoumacin C and liver cancer.The target CAT has the most relations to other targets,with a total of 9 active edges,followed by the targets aldehyde dehydrogenase 18 family member A1 (ALDH18A1),poly (ADP-ribose),polymerase 1 (PARP1) and so on.The PPI network diagram of its core target point is shown in Fig.5.

Fig.4 PPI analysis of the intersection target between Paenicoumacin B and liver cancer

Fig.5 PPI analysis of the intersection target between Paenicoumacin C and liver cancer

ALB

gene is located on chromosome 4q13.3,which encodes albumin.It is the most abundant protein in human blood,and it is expressed restrictedly only in the liver.Studies have shown that albumin-bilirubin score (ALBI score) can be used to assess liver function [30].

CAT

is located on chromosome 11p13 and is commonly expressed in the liver.The encoded protein is catalase.This enzyme plays an important role in the body’s defense against oxidative stress.Many chronic or late-onset diseases are characterized by oxidative stress,such as diabetes,asthma,systemic lupus erythematosus,rheumatoid arthritis,and cancer.Studies have shown that HCV infection is also related to increased oxidative stress in liver tissue [31,32].HCV infection is a serious infectious disease clinically manifested as chronic hepatitis,which gradually leads to liver cirrhosis,and eventually to hepatocellular carcinoma (HCC) [32].Based on the situation of resisting oxidative stress,the liver has its own protective system,among which the key enzymes are CAT,SOD,glutathione reductase (GSR),and glutathione transferase(GST) [32].

3.4 Core target GO and KEGG enrichment analysis

The gene function and signal pathway enrichment analysis results of the compound Paenicoumacin A,B,C and the core target of liver cancer disease were obtained from the Metascape website.

The intersection targets of Paenicoumacin A and liver cancer involve 236 entries in GO enrichment,184 entries in biological process,31 entries in GO molecular functions,and 21 entries in GO cell composition.Items with the highest enrichment level were selected to draw a histogram(Fig.6 B (abc)).Biological process (BP) was mainly manifested in response to steroid hormone (GO:0048545),mRNA metabolic process (GO:0016071),growth factor (GO:0070848),oxygen levels(GO:0070482),and oxidative stress (GO:0006979).GO molecular function was mainly manifested in oxidoreductase activity (GO:0016491),hydrolase activity,and the action on ester bonds (GO:0016788).GO cell composition was mainly manifested in secretory granule lumen (GO:0034774),focal adhesion (GO:0005925),and mitochondrial matrix(GO:0005759).KEGG pathway enriched 38 entries and a bubble chart for the highly enriched entries were drawn (Fig.6B (d)).The main pathways involved were as follows:herpes simplex infection(hsa05168),pathways in cancer (hsa05200),focal adhesion (hsa04510),peroxisome (hsa04146) and so on.

Fig.6 GO,KEGG analysis of common core target of Paenicoumacin A and liver cancer

Paenicoumacin B and hepatocellular carcinoma targeted 112 entries in GO enrichment (Fig.7).There were 71 entries on BP,and the main pathways involved were response to nutrient (GO:0007584),carbohydrate derivative biosynthetic process(GO:1901137),and proteolysis involved in cellular protein catabolic process (GO:0051603) (Fig.8B(a)).19 and 22 entries were enriched on GO MF and GO CC,mainly in oxidoreductase activity(GO:0016702),hydrolase activity (GO:0016811),serine-type peptidase activity (GO:0008236),protein homodimerization activity (GO:0042803),secretory granule lumen (GO:0034774),primary lysosome (GO:0005766),and focal adhesion(GO:0005925).10 signal pathways were enriched on KEGG pathway (Fig.7B (d)),including arachidonic acid metabolism (hsa00590),protein processing in endoplasmic reticulum (ko04141),purine metabolism (hsa00230),focal adhesion (hsa04510),and estrogen signaling pathway (ko04915).

Fig.7 GO,KEGG analysis of common core target of Paenicoumacin B and liver cancer

Paenicoumacin C and hepatocellular carcinoma targeted 153 entries in GO enrichment(Fig.8).The enrichment pathways in BP mainly included compound metabolic process(GO:0072521),esponse to metal ion (GO:0010038),and small molecule biosynthetic process(GO:0044283).The pathways highly enriched in MF were protein kinase binding (GO:0019901)and oxidoreductase activity (GO:0016491).A total of 9 pathways were enriched in KEGG pathway,including the pathway carbon metabolism(hsa01200),arginine and proline metabolism(hsa00330),pancreatic secretion (hsa04972),and pyrimidine metabolism (hsa00240).

Fig.8 GO,KEGG analysis of common core target of Paenicoumacin C and liver cancer

According to the above-mentioned PPI analysis and GO and KEGG signal pathway analysis,we found that Paenicoumacin A,B and C played a potential role in regulating cell oxidative stress,nucleic acid metabolism,cancer process,viral carcinogenic process,secretion of granular cavity,nutrient metabolism,adhesion,etc.The functional enrichment analysis of their intersection targets with liver cancer diseases showed that they were closely related to various oxidation-related enzymes,such as hydrolases,peroxisomes,phosphotransferases,and protein kinases.Paenicoumacin A,B,and C share a common potential target with liver cancer diseases,namely CAT.It was found in GO and KEGG analysis that this target mainly existed in the pathways such as antioxidant activity,hydrolase activity,and secretory granule lumen,which indicated that this target might play an important role in the treatment of liver cancer with the secondary metabolite of amicoumacintype isocoumarin of

Paebibacillus

sp.Xy-2.In addition,Paenicoumacin A had unique potential effects on mRNA metabolism,positive regulation of immune processes,angiogenesis,circadian rhythm,etc.Paenicoumacin B had the advantages in regulating arachidonic acid metabolism,response to topologically incorrect proteins,and sulfide metabolism.Paenicoumacin C could regulate liver development,carbon metabolism,pancreatic secretions,etc.Paenicoumacin B had a negative regulatory effect on NF-kappaB transcription factor activity,while Paenicoumacin C exerted a positive regulation function.

4 Conclusion

Through prediction and analysis of amicoumacin isocoumarin-type compounds in the secondary metabolites of endophytic bacterium

Paenibacillus

sp.Xy-2,

three amicoumacin compounds,Paenicoumacin A,B and C,were found to play a potential value in the fight against liver cancer in terms of regulating cell oxidative stress process,cancer process,viral carcinogenic process,nucleic acid metabolism,and nutritional process.The above analysis provides useful information for the experimental verification of Paenicoumacin A,B and C for the treatment of liver cancer diseases.

5 Acknowledgments

This work was financially supported by Natural Science Foundation of Liaoning (2019-ZD-0457) and Foundation of Career Development Plan for young and middle-aged Teachers of Shenyang Pharmaceutical University (ZQN2018004).