Systematic Genome-wide Screening and Prediction of microRNAs in EBOV During the 2014 Ebolavirus Outbreak

Recently, several thousand people have been killed by the Ebolavirus disease (EVD) in West Africa, yet no current antiviral medications and treatments are available. Systematic investigation of ebolavirus whole genomes during the 2014 outbreak may shed light on the underlying mechanisms of EVD development. Here, using the genome-wide screening in ebolavirus genome sequences, we predicted four putative viral microRNA precursors (pre-miRNAs) and seven putative mature microRNAs (miRNAs). Combing bioinformatics analysis and prediction of the potential ebolavirus miRNA target genes, we suggest that two ebolavirus coding possible miRNAs may be silence and down-regulate the target genes NFKBIE and RIPK1, which are the central mediator of the pathways related with host cell defense mechanism. Additionally, the ebolavirus exploits the miRNAs to inhibit the NF-kB and TNF factors to evade the host defense mechanisms that limit replication by killing infected cells, or to conversely trigger apoptosis as a mechanism to increase virus spreading. This is the first study to use the genome-wide scanning to predict microRNAs in the 2014 outbreak EVD and then to apply systematic bioinformatics to analyze their target genes. We revealed a potential mechanism of miRNAs in ebolavirus infection and possible therapeutic targets for Ebola viral infection treatment.

Structures of potential EBOV pre-miRNAs and mature miRNAs. In order to achieve more accurate prediction, we predicted the primary, secondary, and tertiary structure of the four potential EBOV pre-miRNAs 31,32 . RNAfold was used to predict consensus secondary structures of the pre-miRNAs. Attempts to fold these pre-miRNAs, along with their flanking sequences, into the expected hairpin structures were successful in candidates EBOV-pre-miRNAs-T1, EBOV-pre-miRNAs-T2, EBOV-pre-miRNAs-T3 and EBOV-pre-miRNAs-T4, while other pre-miRNAs failed to fold into any form of stable RNA structures because that they could only give rise to stem-loops that were too short to be authentic pre-miRNA stem-loops. These findings led us to believe that EBOV-pre-miRNAs-T1, EBOV-pre-miRNAs-T2, EBOV-pre-miRNAs-T3 and EBOV-pre-miRNAs-T4 are indeed authentic EBOV pre-miRNAs. The optimal secondary structures in dot-bracket notation with a minimum free energy of − 39.4 kcal/mol of EBOV-pre-miRNA-T1, − 18.8 kcal/mol of EBOV-pre-miRNA-T2, − 33.51 kcal/mol of EBOV-pre-miRNA-T3 and − 31.91 kcal/mol of EBOV-pre-miRNA-T4 are given in Fig. 3A,B shows the predicted hairpin structures of confirmed EBOV pre-miRNAs. Hairpins longer than 100 nt were truncated; these hairpins are indicated by a double slash preceding the stem. Furthermore, the 3D structures of the potential EBOV pre-miRNAs were computationally predicted. Even though EBOV-pre-miRNA-T3 and EBOV-pre-miRNA-T4 have similar 2D structures 31 , they own different 3D structures (Fig. 3C). The positions of possible mature miRNAs were colored in second and tertiary structures (Figs. 3B,C) 32 . For EBOV-pre-miRNAs-T1, EBOV-pre-miRNAs-T2, EBOV-pre-miRNAs-T3 and EBOV-pre-miRNAs-T4 alternative predictions, which are likely to represent the authentic stem-loop structures recognized and cleaved by Drosha are shown underneath the minimal energy structures.
Bioinformatics analysis of mature EBOV miRNAs and prediction of their target genes. Based on the sequences of the seven EBOV miRNAs, EBV-miR-T1-5p, EBV-miR-T1-3p, EBV-miR-T2-5p, EBV-miR-T2-3p, EBV-miR-T3-5p, EBV-miR-T3-3p, EBV-miR-T4-5p and EBV-miR-T4-3p, their target genes were searched by TargetScan 33 and the other miRNA regulation database (the integration database of miRecords, TarBase, starbase, and miR2Disease) [34][35][36][37] . All these potential mature miRNAs have Scientific RepoRts | 5:09912 | DOi: 10.1038/srep09912 total 138 possible target genes in human genome (Table 4). Interestingly, the genes P11, EFNA3, FKRP, AMBRA1, HEPACAM, HCCA2, LPHN1 and PHF21B were all targeted by at least two potential mature EBOV miRNAs. Using the GO and DAVID databases 38,39 , identified proteins were clustered into groups based on their biological processes, molecular function, and cellular compartment. Such clustering allowed us to determine how the differentially expressed proteins were distributed according to biological process (Fig. 4A), molecular function (Fig. 4B) and cellular compartment (Fig. 4C). Clusters with the biological process included proteins that are involved in immune system process (GO:0002376), multicellular organismal process (GO:0032501) and response to stimulus (GO:0050896), which are important for human antiviral response (Fig. 4A). This analysis grouped proteins into nine molecular functional classes (Fig. 4B), which were associated with transcription regulation, such as transporter activity (GO:0005215), translation regulator activity (GO:0045182), protein binding transcription factor activity (GO:0000988) and enzyme regulator activity (GO:0030234). Protein clusters were then further categorized into six subcellular distributions according to the cellular compartment (Fig. 4C). Fig. 4D shows the pathway enrichment analysis of the target genes. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis was performed using the DAVID bioinformatics tool 40,41 . The top enriched pathways are listed here (p-value < 0.01). Fig. 4D clearly shows that these proteins are functionally closely related, and are involved in multiple pathways. Inflammation are mediated by chemokine and cytokine signaling pathway   Table 2. The predicted mature miRNAs in EBOV. The EBV-miR-T3-5p and EBV-miR-T4-6p have the same sequences, which are highlighted in grey color.
(P00031), PDGF signaling pathway (P00047), Purine metabolism (P02769), PI3 kinase pathway (P00048) and Ras Pathway (P04393), which are important in human immune response to virus infection. The gene regulation network (GRN) of the target genes 42,43 is shown in the Fig. 4E. We find that the possible target genes, HDAC5 and JARID2 44 , are important epigenetic factors in transcription regulation, which have strong interaction with the transcription regulator in human cells, such as Histone h3, STAT1, CTCF, CCND2 and CTCF. It is illustrated that the EBOV down-regulated the epigenetic factor to large-scale epigenetic alterations in the host gene expressions 45 , which will damage the normal molecular regulation in immune signaling and thus disorder human immune system to block the anti-viral response. The target proteins, NFKBIE and RIPK, are the key co-regulator with IL22 and IFNG or the part of complex of NFκ B in human immune system 46,47 . It also suggests that the EBOV affects the signaling pathway of immune system by the non-coding RNA to inhibit the infection response. Further network analysis was conducted to see the potential biological pathways and processes in which these Signaling pathway analysis of the target genes. According to our gene regulation network analysis, NFKBIE and RIPK1 are two target genes of possible mature miRNAs, EBOV-miR-T1-5p and EBOV-miR-T2-3p, which plays important roles in immune system. Fig. 5A,B show the cytotoxin-associated target genes. NFKBIE 48 and RIPK 49 genes are both involved in the NF-κ B and TNF signaling pathways (Fig. 5C,D). NFKBIE encodes Iκ B epsilon (Iκ Bε ), a member of the Iκ B family. Its binding to NF-κ B inhibits the nuclear translocation of NF-κ B 50 . RIPK1, Receptor-interacting protein 1, is a key effector molecule in the TNFα -induced activation of the transcription factor NF-κ B 51 . The signaling pathway analysis shows that the TNF-a-mediated NF-kB signaling pathway plays a key role in inflammatory response (Fig. 5C,D). Through analyzing the molecular mechanism in the NF-kB and TNF signaling pathways, those two pathways were discovered to participate in the host immune response, thus they are also attractive targets of viral pathogens 52 . The EBOV coding possible miRNAs, EBOV-miR-T1-5p and EBOV-miR-T2-3p, may be silence and down-regulate the NFKBIE and RIPK1, which are the central mediator of those pathways. In addition, the EBOV exploits the non-coding RNA to inhibit the NF-kB and TNF factor to evade the host defense mechanisms that limit replication by killing infected cells, or to conversely trigger apoptosis as a mechanism to increase virus spreading. NF-kB is activated by multiple families of viruses, including HIV-1, HTLV-1, hepatitis B virus (HBV), hepatitis C virus (HCV), EBV, and influenza virus 53 . This activation may serve several functions of promoting viral replication, preventing virus-induced apoptosis, and suprressing the immune response to the invading pathogen 54 . The genes NF-kB and TNF are affected by synthetic dsRNAs. It suggests that viruses that generate dsRNA replicative intermediates employ a common mechanism to enhance viral replication. The EBOV glycoprotein (GP) and the other viral proteins also induce the NF-kB and TNF signaling pathways 55 . The miRNA and target gene details of EBV-miR-T1-5p and EBV-miR-T2-3p are listed in Fig. 5E. The target gene of EBV-miR-T1-5p is named NFKBIE with ID No. 4794, and 1053-bp long. The target gene of EBV-miR-T2-3p is named RIPK1 with ID No. 8737, and 1848-bp long. Normally when a virus enters a human body, the target genes mentioned above should increase their expression levels. Based on our analysis, once EBOV enters the human body, the miRNAs EBV-miR-T1-5p and EBV-miR-T2-3p down-regulate the target genes of NFKBIE and RIPK1 to evade host innate immune responses and finally leads to multiple organ dysfunction syndrome (MODS).
EBOV miRNA regulated the expression of target genes NFKIBIE, RIPIK1, HDAC5 and JARID2. The miRNA mimics was synthesized and transfected into Hela cells for 24 h/37 °C. The results showed that EBOV-miRNA mimics directly regulate the expression of target genes, NFKIBIE, RIPIK1, JARID2 and HDAC5, in Hela cells. Real time (RT)-PCR was used to quantify and compare the mRNA expression of target genes in the transfected Hela cells with the EBOV-miRNA mimics and the scramble mimics as negative control. The transfection efficiency of miRNA mimics, including miR-T1-3p, miR-T2-5p, miR-T3-3p, miR-T4-3p and their mixed miRNA (MIX miRNA), and the scrambled miRNA mimics were detected by fluorescence activated cell sorter (FACS), which were more than 80% (Fig. 6A). The RT-PCR analysis indicated that expression of NFKIBIE was significantly down-regulated by 1.526 times in miR-T1-5p mimics-transfected Hela cells and the expression of RIPIK1 and JARID2 were down-regulated by miR-T2-3p mimics with 1.947 times and 2.25 times separately (Fig. 6B). Interestingly, in the transfected Hela cells with miR-T3-5p mimics, the expression of HDAC5 is increasing compared with the transfected Hela cells with scrambled miRNA mimics (Fig. 6B). We also found that the mRNA expression of NFKIBIE were inhibited by miR-T1-3p, miR-T3-5p, miR-T4-3p and MIX miRNA mimics,    while all EBOV-miRNA mimics significantly can block the expression of RIPK1 (Fig. 6C). Such results illustrated that these miRNA may have important functions in human immune and antiviral response systems.

Conclusion
From March 2014 to October 2014, the largest EBOV outbreak in the history has killed several thousand people in West Africa, yet no current specific treatment is validated for EBOV. miRNAs has been demonstrated to circulate in a highly stable, cell-free form in body fluids thus circulating miRNAs can be used as non-invasive biomarkers for molecular diagnostics and prognostics 24 . Since the majority of known viral miRNAs are encoded by DNA viruses, RNA-virus-encoded miRNAs remains controversial.
Recently, despite that several novel miRNAs have been experimentally identified in the West Nile virus (WNV) 28 , Dengue virus (DENV) 56 , hepatitis A virus (HAV) 23 , the viral miRNAs in numerous viruses have been identified using computational prediction followed by experimental validation 23,[57][58][59] . This year H. Liang et al. 24 have successfully identified two EBOV miRNAs, which provide new evidence that EBOV can encode functional miRNAs.
In this paper, based on the whole genome alignment of the 2014 outbreak EBOVs 29 , which is different from the analysis method of Liang et al. 24 , we predicted and experimentally verified seven mature EBOV miRNAs and their regulated target genes. By analyzing the signaling pathways of the target genes, we hypothesized that EBOV miRNAs, especially EBV-miR-T1-5p and EBV-miR-T2-3p, down-regulate the target genes of NF-KB and TNF expression which are involved in virus-cell interaction, immune escape, and cell apoptosis 48,49 . As viruses evolve under the highly selective pressures of the immune system, they acquire the capacity to target critical steps in the host cell life, hijacking vital cellular functions to promote viral pathogenesis. Many viruses have evolved mechanisms to target the NF-kB pathway to facilitate their replication, cell survival, and evasion of immune responses. In addition, some viruses use the NF-kB pathway either for its anti-apoptotic properties to evade the host defense mechanisms or to trigger apoptosis as a mechanism of virus spread. Not surprisingly, recent studies focusing on the interrelations between NF-kB and virus pathogenesis reveal that many viral products bypass signal-induced stimulation and/or receptor-proximal steps to directly interface with the IKK complex. Continuous activation of NF-kB as a consequence of viral persistence in some instances leads to oncogenic transformation, which is an area of increasing scientific study and fascination.
It is noteworthy that this is the first paper to describe a strategy that the EBOV non-coding RNA may be evolved to modulate the NF-kB and TNF signaling pathways, to enhance viral replication, host cell survival, and evasion of immune responses. According to our prediction results, the EBOV-miRNA mimics of miR-T1-5p can significantly inhibit the expression of NFKIBIE, while the miR-T2-3p mimics can block the expression of RIPK1 significantly.
Such results illustrated that these miRNA may have important functions in human immune and antiviral response systems. Ongoing research will undoubtedly continue to reveal other novel interactions between viruses and the NF-kB pathway that will permit a more precise molecular dissection of this parasitic fatal attraction. Considering that the target genes regulated by our predicted miRNAs play key roles in human immune response, we highlight our prediction work raises the hope of quickly finding an effective EBOV disease treatment.

Materials and Methods
EBOV whole genome sequencing. To show the genome-wide nucleotide and amino acid signatures, we retrieved full genome sequences (as of October 22, 2014) from the genome browser at NCBI Database. One-hundread-and-two human EBOV genomes were derived from the 2014 outbreak 29,30 . Multiple genome-wide alignments. The multiple sequence alignment tools ClustalW and MUSCLE were applied for the alignment of the EBOV genomes. The alignments were then analyzed to identify the characteristic sites as potential signatures to distinguish different virus genomes and proteins 30 . Bioinformatics prediction of the miRNAs. A flowchart describing the computational prediction of the putative miRNAs is shown in Fig. 2. Briefly, the viral genome was scanned for stem-loop structures of miRNA precursor (pre-miRNA) using VMir 60 (http://www.hpi-hamburg.de/research/ departments-and-research-groups/antiviral-defense-mechanism/software-down-load.html), a computational analyzer program for the prediction of putative pre-miRNAs. The complete genome sequences of 138 different human EBOV strains were acquired from the PubMed, where 102 human EBOV strains belong to 2014 EBOV outbreak. VMir predictions were carried out using the default parameters. The putative pre-miRNAs that satisfied the filter parameters of a VMir score ≥ 150 and a window count ≥ 35 were selected for further assessment. Then, the source code of MiPred 61 was used to distinguish real and pseudo miRNA precursors from the obtained sequences with a prediction confidence equal to or greater than 70%. Subsequently, mature miRNA sequences were predicted from the pre-miRNA stem-loops. Then we used the MatureBayes tool 62 to extend the prediction coverage of the mature miRNAs. The default conditions were used for the MatureBayes tool.
Prediction of Novel Viral miRNA Targets. Human target genes of novel EBOV miRNAs were predicted using TargetScan 33 custom miRNA prediction methods. Putative targets within the viral genome were predicted using TargetScan Perl script 34,36,37,60 . Gene ontology (GO) analysis. GO analysis of the significant probe list was performed using PANTHER (http://www.pantherdb.org/) 38,39 , using text files containing the Gene ID list and accession numbers of the Illumina probe ID. All data analysis and visualization of differentially expressed genes were conducted using R 2.4.1 (www.r-project.org). In addition, the DAVID Functional Annotation Bioinformatics Microarray Analysis tools (http://david.abcc.ncifcrf.gov/) were used to study the biological function of the regulated genes.

Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis. Kyoto Encyclopedia of
Genes and Genomes (KEGG) 40,41 is a collection of online databases dealing with genomes, enzymatic pathways, and biological chemicals. The PATHWAY database records networks of molecular interactions in the cell that includes organism-specific network maps (http://www.genome.jp/kegg/). A total of 9 pathways, involving 49 genes, were collected from KEGG.
Constructing Gene Regulation Network. The integrated regulatory network was constructed based on systematic integration of various high-throughput datasets 42,43 . The target genes associated with candidate miRNAs were selected from miRNA regulation database (the integration database of miRecords, TarBase, starbase, and miR2Disease) and TFs associated with those target genes were selected from Transcriptional Regulatory Element Database (TRED). The integrated target genes-TF regulatory network was constructed by using Cytoscape software (http://cytoscape.org/), which is an open source software for visualizing complex networks and integrating these networks with any type of attribute data.
Cell culture and transfection. Hela cells at 37 °C and 5% CO 2 in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (Hyclone, USA) . Synthetic duplex EBV-miRNA mimics, scramble oligonucleotides used as negative control (NC) (GenePharma, Shanghai, China) at a final concentration of 50 nM were introduced into Hela cells by siPORT™ NeoFX™ transfection agent (AM4511, Applied Biosystems Inc., USA). according to the manufacturer's instructions. Cells were harvested at 24 hours after transfection.

RNA extraction and quantitative real-time PCR analysis. The harvested cells were placed in
TRIzolH Reagent (Invitrogen, Carlsbad,CA), and total RNAs were extracted according to the TRIzol manufacturer'sinstructions. Total RNA (3 μ g) was primed by oilgodT and converted intocDNA using SuperScript III (Invitrogen). The SYBR green-based real-time PCR was performed in Light Cycle 2.0 System(Roche) and Relative quantification of mRNA was measured using 2-Δ Δ Ct method which normalized to GAPDH. Each PCR assay was performed in a final volume of 20 ul, containing 2 ul of DNA template, 25 pmol of each type-specific primer set, 2X Maxima SYBR Green (Thermo Scientific), Reactions were incubated at 50 °C for2 minutes, followed by PCR amplification under the following conditions: 95 °C for10 minutes, 95 °C for 30 seconds, 57 °C for 30 seconds, 72 °C for 30 minute for a total of 40 cycles, Following the PCR step, a melt curve analysis was performed by increasing the temperature from 65 °C to 95 °C at increments of 0.1 °C/s for each fluorescence reading. The primer sequences for qPCR analysis were as follows: GAPDH, forward 5′ -AGAAGGCTGGGGCTCATTTG-3′ and reverse 5′ -AGGGGCCATCCACAGTCTTC-5′ ; NFKBIE, forward 5′ -TGCCAACAGATGGCCCATAC-3′ and reverse 5′ -TGTTCTTTTCACTAGAGGCACCA-3′ ; RIPK1, forward 5′ -TGGGCGTCATCATGAGGAAG-3′ and reverse 5′ -CGCCTTTTCCATGTAAGTAGCA-3′ ; MAPKAP3, forward 5′ -ATGAGAACATGCACCATGGCAAGC-3′ and reverse 5′ -GGGCAATGTTATGGCTGTGCAGAA -3′ ; HDAC5, forward 5′ -TTCTTTGGACCAGAGTTCCC-3′ and reverse 5′ -GTTGGGTTCAGAGGCTGTTT-3′ ; JARID2, forward 5′ -GAGCATGTGTTTCAGCAAGG-3′ and reverse 5′ -CTTCTCTTCCACTAGCCTCCAG-3′ Flow cytometry. Green fluorescence after fluorescent dye labeled miRNA delivery into Hela cells were assessed in an Accuri C6 BD Biosciences flow cytometer, excited by an argon 480 nm laser and detected by use of a 520 nm optical filter. The mean fluorescence intensity (MFI) data were collected from these cell populations. All data was analyzed with a minimum of 15000 events setting the respective negative control to less than 1% in BD CFlow software. Post-transfection in Hela cells as percentage of fluorescent dye labeled miRNA positive cells.