Acute Mountain Sickness (AMS), a kind of human pathological reaction to high altitude, is caused by acute exposure to low oxygen partial pressure at high altitude. It has a series of nonspecific symptoms, such as headache, dizziness, palpitation, disgust, fatigue, numbness in the extremities, convulsion, and so on. AMS may occur when people ascend to the height above 2500 meters. Severely, it can develop into high altitude pulmonary edema (HAPE) or high altitude cerebral edema (HACE), which is likely to be life-threatening. There is the largest plateau in China and many people settled on plateaus in the world. In addition, lots of tourists, mountaineers and frontier guards enter into plateaus above 3000 meters every year. AMS has become an urgent problem to be solved, especially in executing tasks on highland. During the earthquake occurring in Yushu of Qinghai province of China in 2008, the morbidity of AMS reached as high as 81.5% in the emergency rescue workers, seriously affecting relief worker.
Hypoxia-inducible factor 1 (HIF-1) is a transcription factor which is related to hypoxia. HIF-1, a heterodimer, consists of a α subunit and a β subunit. Under normoxia, HIF-1α is continuously synthesized and rapidly degraded via the ubiquitin-proteasome pathway. While under hypoxia, HIF-1α degradation is inhibited, then the accumulated HIF-1α binds with HIF-1β to form heterodimer HIF-1 with transcriptional activity, which could bind with hypoxia response elements (HREs) to activate transcription function of approximately 60 target genes, such as vascular endothelial growth factor (VEGF), inducible nitric oxide synthase (iNOS), erythropoietin (EPO) and so on.[7,8]
It has been reported that Rhodiola rosea, a traditional Tibetan medicine plant in China, can treat AMS through inhibiting HIF-1 degradation pathway. As we know, Traditional Chinese medicine (TCM) has been used in treatment of diseases as a main means in China for thousands of years. Different from western medicine, TCM has the lower side effects and exhibits synergistic effects of multi-compositions by multi-channels and multi-targets. However, separation and extraction are time-consuming due to numerous chemical compositions in TCM, the assay about biological activity is expensive for each composition, and the common drug research methods are not suitable to synergistic effects, so the studies on TCM mechanism is limited.
With the development of system biology and network biology, systematic docking and drug-target network analysis approach compensate for the limitation above. In recent 10 years, it has been widely applied to study molecular mechanism of TCMs.[11,12] Molecular docking method can explore the binding mode and binding strength between targets and compounds at molecular level. AutoDock Vina program, the most commonly used software for molecular docking, has been successfully applied in drug design and approved to be effective. Target-composition interaction network constructed by network biology method is a powerful tool for analyzing the relationship between targets and TCM compositions.[12,16] The combination of the two methods can contribute to more comprehensive understanding of TCM function.
To estimate the interaction between HIF-1 degradation pathway and a single composition of R rosea, we first collected 187 effective molecules from R rosea, and then docked all these molecules with the targets that could decrease HIF-1α degradation using AutoDock Vina software. Finally, we built a compound-target network using Cytoscape program commonly used for biological research. From the interaction network, we analyzed the potential target of compositions of R rosea and provided a reference for the therapeutic mechanism of R rosea.
2 Materials and methods
All study methods were approved by ethics committee of General Hospital of Chinese People's Armed Police Force.
2.1 Data preparation
A total of 187 compounds were collected from Panossian's and Zhou's articles[19,20] and were used to build compound database of R rosea. First, the structures of the 187 compounds were drawn using ChemDraw module of ChemBioOffice software and saved as 2D cdx format. Then, these compositions were converted into 3D conformation using chem3D module and were optimized with MM2 force field to get 3D mol2 format. Finally, we used AutoDockTools (ADT) to prepare ligands before docking simulation, including adding hydrogens and Gasteiger partial charges.
It is worth noting that the structure of pyridrde which is named compound 184 in Zhou's article, is different from that in other papers. Herein, we used 2-Benzhydrylpiperidine (right side in Table 1) as pyridrde with substitution of compound 184, and the other compounds, as well as numbers, are the same as those of Zhou's.
2.2 Selection of targets
As reported, Rhodiola genus plants could treat AMS by inhibiting HIF-1 degradation pathway. In this study, we focused on HIF-1α degradation pathway-related targets. HIF-1α degradation pathways mainly include von Hippel-–Lindau protein (pVHL) -dependent pathway and pVHL-independent pathway.
Under normoxia, in pVHL-dependent pathway, residues PRO402 and PRO564 of HIF-1α can be identified and hydroxylated by prolyl hydroxylase 2 (PHD2), then the hydroxylated HIF-1α binds with pVHL enriching elongin-C/elongin-B/cullin-2 E3 ubiguitin ligase, and is finally degraded by 26S proteasome. Under hypoxia, in the O2/pVHL/PHD pathway, arrest-defective-1 (ARD1), osteosarcoma-9 (OS-9), and spermidine/spermine-N1-acetyltransferase-2 (SSAT2) all can promote HIF-1α degradation by acetylating HIF-1α, binding with HIF-1α/PHD2 or HIF-1α/PHD3 complexes and binding with pVHL/E3 ligase, respectively.
In pVHL-independent pathway, receptor for activated C-kinase1 (RACK1) can enrich elongin-C/elongin-B/cullin-2 E3 ligase to induce HIF-1α degradation. Spermidine/spermine-N1-acetyltransferase-1 (SSAT1) can stabilize the reaction between HIF-1α and RACK1. In addition, glycogen synthase kinase 3β (GSK3β) can promote HIF-1α degradation by phosphorylation.
Besides above targets, E3 ligase, human double minute 2 (Hdm2), and forkhead transcription factor (FOXO4) have been proved to be also conducive to HIF-1α degradation.
In the end, we selected these targets including pVHL, PHD, ARD1, OS-9, SSAT2, RACK1, SSAT1, GSK3β, Hdm2, and FOXO4 as docking simulation.
2.3 Molecular docking
To prepare for docking simulation, water molecules, solvent molecules, and substrate ligands were removed using Pymol software; hydrogen and partial charges were added using ADT, and the ligand binding site for each protein was defined as either the location of the co-crystallized ligand or the site suggested by experimentalists.[31,32] The relevant parameters are shown in Table 2.
Prior to molecular docking, except the targets including RACK1 without ligand and DNA-binding FOXO4, crystal recovery test was performed in other 8 targets to ensure the reliability of method and parameter. AutoDock Vina software was employed to dock ligands with protein active site. Evaluation criterion was set to 2 angstrom or less by calculating root mean square deviation (RMSD) between docked ligand conformation and co-crystallized ligand conformation. RMSD value and AutoDock Vina score are shown in Table 2.
Except for OS-9 and SSAT1, RMSDs of other 6 targets were less than or about 2 angstrom, which was regarded as successful recovery. Sugar ring of OS-9 and purine ring of SSAT1 were thought to be the major responsibility for simulation failure. It can be seen in Figure 1 that sugar ring and purine ring were mostly outside the pocket and there were fewer interactions to maintain their orientations, so it was difficult to recover their locations. Except for the ring outside pocket, others in pocket could recover both location and major interaction (Fig. 1) in OS-9 and SSAT1. For above reason, the parameters of OS-9 and SSAT1 were still used in the following docking simulation.
For the docking simulation, AutoDock Vina software was employed to dock the 187 compounds with the structures of 10 proteins, respectively. The macromolecules were treated as rigid, while the ligands were flexible. The box parameters are shown in Table 2, and other parameters were set as default.
A total of 187 compounds were saved as pdbqt file. They were divided into 6 categories and are listed in Table 3.
All the scores of 187 compounds to the 10 proteins are listed in Table 4. To select compounds that could potentially inhibit the 10 proteins, we choose the top 5% of the compounds ranked by the lowest docking energy, respectively. If different compounds possessed the same score meeting with the top 5% criterion, these compounds all were regarded as positive results. Each target had 9 to 13 compounds with top 5% criterion (Table 4).
No. of compounds that have interaction with corresponding target proteins are listed in Table 4. The interaction relationships between all the 10 proteins and related compounds were merged in compound-target network using Cytoscape software (Fig. 2).
For each target, Table 4 lists the compound which was regarded as possible inhibitors calculated by AutoDock Vina. Of the 42 interaction compounds, 18 have connection with more than 2 targets, while 27 compounds (64%) had more than 1 target. The relationships of targets and compounds are shown in Figure 2, which suggests TCM's multi-target/multi-component strategy for disease treatment.
Compounds which either act on more than 4 proteins or are proved as an active ingredient[20,33,34] are listed in Table 5. In these compounds, 5 compositions (Rhodiosin, Rhodiolatuntoside, Rhodionin,[30,36] Crenulatin, Pyridrde) have been proved as active ingredients, and 4 of them (80%) are inhibitors for more than 2 targets.
In this way, TCM offers an efficient treatment for diseases, which can be explained by molecular docking and network method to some extent. As some active compounds have been successfully selected, the method is regarded as an effective approach for exploring TCM's. Moreover, the identified 18 multi-target compounds with more than 2 targets are worth being noticed. However, these need validation for future experimental studies.
However, the common component, salidroside, was not identified, while 5-reported active ingredients were screened out from the 187 compounds. This may be that salidroside is likely to play a role in another pathway, such as HIF-1α synthetic pathway. Due to fewer study about this pathway, we didn’t include it in this study. In the following research, we will focus attention on HIF-1α synthetic pathway to explore the mechanism of salidroside.
Corresponding to the 10 targets, the compounds with approved active ingredients are listed in Table 6. From Table 6, we could see that targets ARD1, FOXO4, OS-9, PHD2, and Hdm2 existed in 2 different active compounds. Of the 5 targets, 3 targets (ARD1, OS-9, and PHD2) are located in pVHL-dependent pathway, while the remaining 2 targets (FOXO4 and Hdm2) belong to neither pVHL-dependent nor pVHL-independent pathway. We speculate that pVHL-dependent pathway is more important for HIF-1α degradation and also is a main mechanism for treatment of AMS. In addition, different from other targets, Hdm2 is a special target, because the 2 different compositions (Flavonoids and Other chemical constituent) (Table 3) all contain Hdm2 target.
Among all predicted interaction between target and compound, 3 target-compound interactions (PHD2-Rhodionin, RACK1-Rhodiosin, SSAT1-Rhodiolatuntoside) are worth being noticed due to their highest score in the 187 target-compound pairs. Based on this, we speculate that targets of the 3 active compounds [(Rhodionin (No.15), Rhodiosin (No.16), and Rhodiolatuntoside (No.4)] are PHD2, RACK1, and SSAT1, respectively. The predicted binding modes of them are shown in Figure 3 which indicates moderate hydrogen bond interaction, stacking interaction and van der Waals interaction in each binding graph. Further experiment will be needed to confirm their bindings.
This study had several limitations. First, this study did not cover entire proteome which might have been affected by R rosea. Second, the result is purely computational. Third, target proteins were selected based on the previous research about HIF-1 which requires biochemical or cell biological experiment to confirm the result conducted in this study.
We used molecular docking and network method to study the therapeutic mechanism of R rosea on AMS specifically investigated the relationship between the active compositions of R rosea and HIF-1α degradation pathway. Our results show that R rosea can inhibit HIF-1α degradation pathway through at least 5 constituents (Rhodiosin, Rhodiolatuntoside, Rhodionin,[30,36] Crenulatin, and Pyridrde). In addition, 5 targets (ARD1, FOXO4, OS-9, PHD2, Hdm2) of HIF-1α degradation pathway are worth being noticed. Our data also suggest that PHD2, RACK1 and SSAT1 are likely to be the targets of the 3 active compounds [(Rhodionin (No.15), Rhodiosin (No.16), and Rhodiolatuntoside (No.4)], respectively. Our results remain to be confirmed by further experiments.
We thank Doctor Jun-qi Yang from Beijing University of Technology for providing the ChemBioOffice software.
Conceptualization: Fan Wang.
Data curation: Hai-feng Liu.
Investigation: Kai Zhang.
Methodology: Xu-yi Zhang.
Writing – original draft: Zi-liang Liang.
Writing – review & editing: Hui-liang Liu.
. Imray C, Wright A, Subudhi A, et al. Acute mountain sickness: pathophysiology, prevention, and treatment. Prog Cardiovasc Dis 2010;52:467–84.
. Paralikar SJ, Paralikar JH. High-altitude medicine. Indian J Occup Environ Med 2010;24:6–12.
. Grocott M, Montgomery H, Vercueil A. High-altitude physiology and pathophysiology: implications and relevance for intensive care medicine. Crit Care 2007;11:203.
. Wu T. Mountain rescue. The highest earthquake in Yushu. High Alt Med Bio 2011;12:93–5.
. Smith TG, Robbins PA, Ratcliffe PJ. The human side of hypoxia-inducible factor. Br J Haematol 2008;141:325–34.
. Lee JW, Bae SH, Jeong JW, et al. Hypoxia-inducible factor (HIF-1)alpha: its protein stability and biological functions. Exp Mol Med 2004;36:1–2.
. Ke Q, Costa M. Hypoxia-inducible factor-1 (HIF-1). Mol Pharmacol 2006;70:1469–80.
. Kaur B, Khwaja FW, Severson EA, et al. Hypoxia and the hypoxia-inducible-factor pathway in glioma growth and angiogenesis. Neuro Oncol 2005;7:134–53.
. Wang YJ, He LQ, Sun W, et al. Optimized project of traditional Chinese medicine in treating chronic kidney disease stage 3: a multicenter double-blinded randomized controlled trial. J Ethnopharmacol 2012;139:757–64.
. Stone R. Biochemistry. Lifting the veil on traditional Chinese medicine. Science 2008;319:709–10.
. Shi SH, Cai YP, Cai XJ, et al. A network pharmacology approach to understanding the mechanisms of action of traditional medicine: bushenhuoxue formula for treatment of chronic kidney disease. PLoS One 2014;9:e89123.
. Gu S, Yin N, Pei J, et al. Understanding molecular mechanisms of traditional Chinese medicine for the treatment of influenza viruses infection by computational approaches. Mol Biosyst 2013;9:2696–700.
. Kitchen DB, Decornez H, Furr JR, et al. Docking and scoring in virtual screening for drug discovery: methods and applications. Nat Rev Drug Discov 2004;3:935–49.
. Trott O, Olson AJ. AutoDock Vina. improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. J Comput Chem 2010;31:455–61.
. Lim S, Othman R, Yusof R, et al. Rational drug discovery of HCV helicase inhibitor: improved docking accuracy with multiple seedings of Autodock Vina and in situ minimization. Curr Comput Aided Drug Des 2017;31:160–9.
. Liang H, Ruan H, Ouyang Q, et al. Herb-target interaction network analysis helps to disclose molecular mechanism of traditional Chinese medicine. Sci Rep 2016;6:36767.
. Shannon P, Markiel A, Ozier O, et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res 2003;13:2498–504.
. Psomopoulos FE, Vitsios DM, Baichoo S, et al. BioPAXViz: a cytoscape application for the visual exploration of metabolic pathway evolution. Bioinformatics 2017;33:1418–20.
. Panossian A, Wikman G, Sarris J. Rosenroot (Rhodiola rosea
): traditional use, chemical composition, pharmacology and clinical efficacy. Phytomedicine 2010;17:481–93.
. Jiang-tao Zhou, Lei Xu, Yan-yan Chen, et al. Chemical constituents of the genus phodiola. J Chin Pharm Sci 2014;433–45.
. Morris GM, Huey R, Lindstrom W, et al. AutoDock4 and AutoDockTools4: automated docking with selective receptor flexibility. J Comput Chem 2009;30:2785–91.
. Majmundar AJ, Wong WJ, Simon MC. Hypoxia-inducible factors and the response to hypoxic stress. Mol Cell 2010;40:294–309.
. Tribioli C, Mancini M, Plassart E, et al. Isolation of new genes in distal Xq28: transcriptional map and identification of a human homologue of the ARD1 N-acetyl transferase of Saccharomyces cerevisiae
. Hum Mol Genet 1994;3:1061–7.
. Baek JH, Mahon PC, Oh J, et al. OS-9 interacts with hypoxia-inducible factor 1alpha and prolyl hydroxylases to promote oxygen-dependent degradation of HIF-1alpha. Mol Cell 2005;17:503–12.
. Zahedi K, Bissler JJ, Wang Z, et al. Spermidine/spermine N1-acetyltransferase overexpression in kidney epithelial cells disrupts polyamine homeostasis, leads to DNA damage, and causes G2 arrest. Am J Physiol Cell Physiol 2007;292:C1204–15.
. Liu YV, Baek JH, Zhang H, et al. RACK1 competes with HSP90 for binding to HIF-1alpha and is required for O(2)-independent and HSP90 inhibitor-induced degradation of HIF-1alpha. Mol Cell 2007;25:207–17.
. Baek JH, Liu YV, McDonald KR, et al. Spermidine/spermine-N1-acetyltransferase 2 is an essential component of the ubiquitin ligase complex that regulates hypoxia-inducible factor 1alpha. J Biol Chem 282D 2007:23572–82580.
. Jope RS, Yuskaitis CJ, Beurel E. Glycogen synthase kinase-3 (GSK3): inflammation, diseases, and therapeutics. Neurochem Res 2007;32:577–95.
. Ravi R, Mookerjee B, Bhujwalla ZM, et al. Regulation of tumor angiogenesis by p53-induced degradation of hypoxia-inducible factor 1alpha. Genes Dev 2000;14:34–44.
. Tang TT, Lasky LA. The forkhead transcription factor FOXO4 induces the down-regulation of hypoxia-inducible factor 1 alpha by a von Hippel–Lindau protein-independent mechanism. J Biol Chem 2003;278:30125–35.
. Adams DR, Ron D, Kiely PA. RACK1, A multifaceted scaffolding protein: structure and function. Cell Commun Signal 2011;9:22.
. Liu YV, Hubbi ME, Pan F, et al. Calcineurin promotes hypoxia-inducible factor 1alpha expression by dephosphorylating RACK1 and blocking RACK1 dimerization. J Biol Chem 2007;282:37064–73.
. Cheng-fang Tang, Si-ping Jiang, Bin Chen, et al. Analysis of the main chemical composition of Rhodiola Magnoliae from different regions of Tibet by LC-MS /MS method. Acta Sci Nat Univ Sunyatseni 2013;6:99–103.
. Ping Wang, Yu-xia Xiong, Xian-li Meng, et al. Screening of effective parts of rhubarb and baicalin against endotoxin. J Chengdu Univ Tradit Chin Med 2006;3:56–8.
. Yan-mei Cu, An-ru Lou, Chang-yi Zhao. Research Progress on the chemical constituents and pharmacological effects of Rhodiola. J Beijing Norm Univ 2008;03:328–33.
. Tolonen A, György Z, Jalonen J, et al. LC/MS/MS identification of glycosides produced by biotransformation of cinnamyl alcohol in Rhodiola rosea
compact callus aggregates. Biomed Chromatogr 2004;18:550–8.
. Xiu-fang Gao, Hai-ming Shi, Ying Shan, et al. The effect of Rhodiola on the expression of HIF-1 alpha, HIF-1 beta and VEGF in endothelial cells under hypoxia. Chin J Tradit Chin Med Pharm 2010;4:582–5.