https://orcid.org/0000-0001-6394-4170
ABSTRACT
MicroRNAs (miRNAs) are small non-coding RNAs that regulate the post-transcriptional expression of target genes. Virus-encoded miRNAs play an important role in the replication of viruses, modulate gene expression in both the virus and host, and affect their persistence and immune evasion in hosts. This renders viral miRNAs as potential targets for therapeutic applications, especially against pathogenic viruses that infect humans and animals. Rift Valley fever virus (RVFV) is a mosquito-borne zoonotic RNA virus that causes severe disease in both humans and livestock. High mortality among newborn lambs and abortion storms are key characteristics of an RVF outbreak. To date, limited information is available on RVFV-derived miRNAs. In this study, computational methods were used to analyse the RVFV genome for putative pre-miRNA genes, which were then analysed for the presence of mature miRNAs. We detected 19 RVFV-encoded miRNAs and identified their potential mRNAs targets in sheep (Ovis aries), the most susceptible host. The identification of significantly enriched O. aries genes in association with RVFV miRNAs will help elucidate the molecular mechanisms underlying RVFV pathogenesis and potentially uncover novel drug targets for RVFV.
Introduction
miRNAs are endogenous small (~22 nucleotides) non-coding RNAs that are found in plants, animals, fungi, and viruses. These small molecules regulate gene expression by binding to complementary target mRNAs or promoter sequences in DNA (RNA). miRNAs that target promoters activate gene expression or silence genes. Recent studies have shown that miRNAs increase gene expression by binding to complementary promoter sequences. Furthermore, they regulate post-transcriptional gene expression through mRNA cleavage (mRNA decay) or translational repression (mRNA deposition). miRNAs regulate several key cellular processes, including growth and development, cell signalling pathways, external stresses, cancer-related gene expression, and immunological responses [Citation1,Citation2].
The discovery of new miRNA genes and determination of their function are important challenges in transcriptome profiling, which is based on advanced sequencing and computational approaches. Computational strategies provide an efficient method to predict miRNAs and their targets by surveying genomic sequences and databases, including expressed sequence tags (ESTs), which are based on secondary structure characteristics, phylogenetic conservation of both sequence and structure, and thermodynamic stability of hairpins [Citation2].
Most of our current knowledge of viral miRNAs is gained from DNA viruses, which may regulate the viral life cycle and affect host cellular gene expression. However, only limited data are available on RNA viruses, including a few reports on influenza A, Dengue West Nile or Ebola virus [Citation3]. Based on the last version of the miRBase database (22.1), there is still no information on the role of miRNA during the lifecycle of the Rift Valley fever virus (published miRNA sequences and annotation) [Citation4].
Rift Valley fever virus (RVFV) is a mosquito-borne zoonotic disease that was originally discovered in the Kenyan Rift Basin in the 1930s and was initially described as a massive abortion and necrotic hepatitis in sheep. Since then, the virus has spread throughout sub-Saharan Africa and, more recently, has expanded outside the African continent to the Arabian Peninsula. RVFV causes severe disease in both livestock and humans. Epidemics in livestock result in adult mortality rates of up to 20% and newborn mortality rates of up to 100% (especially high in sheep), and cause so-called abortion storms in ewes. In humans, the virus causes a variety of clinical manifestations, ranging from mild flu-like illnesses to severe complications, including blindness, meningoencephalitis, haemorrhagic fever, and death. Large outbreaks and epidemics are usually preceded by heavy rainfall and flooding, which provide ideal conditions for the mass reproduction of primary mosquito vector species, such as Aedes species, and an overall abundance of secondary mosquito vectors. RVFV has already shown to spread outside of endemic regions with devastating implications for public health and the economy [Citation5,Citation6].
Recent transcriptome profiling studies have indicated that small non-coding RNAs may play important roles in both virus-host association and viral pathogenesis. These studies suggest that small non-coding RNAs could affect virus replication and the changes in the host transcriptome during virus infection and play an important role in the antiviral response. Some small non-coding RNAs encoded by viruses are expressed in the host genome and may contribute to the regulation of both host and viral gene expression during infection. Therefore, small non-coding RNAs may be part of a complex mechanism of transcriptional regulation of viral propagation in susceptible hosts [Citation7–9].
Although some studies have been published on transcriptome alterations and differentially expressed genes during RVFV pathogenesis in infected hosts, RVFV-encoded miRNAs and their underlying influence on the cellular machinery remain largely unknown. The main goal of this study was to identify candidate pre-miRNAs in RVFV using in silico analysis based on multiple computational approaches and software tools. Finally, potential sheep mRNA targets of the newly discovered miRNAs were identified, and their potential functions were annotated.
Materials and methods
Genome analysis, pre-miRNA classification and fold energies
The complete genome sequence of the virulent Rift Valley fever virus strain ZH-548 was obtained from the NCBI database (NC_014397.1, NC_014395.1, and NC_014396.1). To identify candidate pre-miRNAs in particular, customized in-house scripts for extracting sequential and structural features were written in MATLAB and C# (these scripts are available upon request). The in-house script utilizes pseudo-code that leverages a support vector machine (SVM) model as its classifier. Various classifiers were assessed, and the optimal one was selected based on its Mathew’s correlation coefficient (MCC). To provide a brief overview, sequences were uploaded in FASTA format as input, followed by sequence and thermodynamic feature extraction. Subsequently, normalization and feature selection were applied, with SVM serving as the chosen classifier. The VMir software package (v2.3) was used to identify pre-miRNAs in the RVFV genome. The parameters of the window and step sizes were set to 500 and 1bp, respectively [Citation10]. Secondary structures with minimum free energy (MFE) and positional entropy were generated using the RNAfold algorithm [Citation11]. ViennaRNA package 2.0. and Mipred were used for the classification to distinguish real pre-miRNAs from false-positive pre-miRNAs (pseudo pre-miRNAs) [Citation12,Citation13] ().
Figure 1. Flowchart of the data analysis to enrich functional miRnas gene in RFVF genome (the schematic figure was drawn with Biorender.com).
Identification of mature miRNAs
The MatureBayes program was used with default parameters to predict mature miRNAs [Citation14]. The Naïve Bayes classifier was used by MatureBayes to identify probable mature miRNA molecules based on the sequence and structure of the miRNA precursors. RNAfold from the Vienna package was used to predict the secondary structure of the input pre-miRNA hairpins.
Prediction of miRNA targets and functional annotation
All conserved 3′-and 5′-UTR of O. aries genes were extracted using UTR sequences downloaded from the Targetscan database (http://www.targetscan.org/) [Citation15,Citation16] and searched against 60,641 O. aries mRNA RefSeq sequences downloaded from the NCBI database. RNAhybrid [Citation17], Miranda and miRDB (http://mirdb.org/) were applied to predict the potential target sites of the identified O. aries miRNAs. The Gene Ontology analysis was performed using the AmiGO (http://amigo.geneontology.org/amigo), ShinyGO (http://bioinformatics.sdstate.edu/go/), and Enrichr (https://maayanlab.cloud/Enrichr/) websites.
Results
In total, 727 pre-miRNA-like molecules in the RFVF genome were detected using home scripts and VMir ( and ). We further enriched 19 pre-miRNA sequences in MiPred and validated their MFE values using the Mfold web server tool ( and ). shows the characteristics of pre-miRNAs that were proven to be authentic in MiPred and Mfold. The average length of the sequences was 99 nt with an average ΔG of −33.20 kcal/mol. We detected the sequences and positions of mature miRNAs within the validated pre-miRNA sequences using MaturePred (). From the RVFV genome, we identified 727 new pre-miRNA candidates, none of which have been reported previously.
Figure 2. Potential stem-loop structures of pre-miRnas of RVFV genome. Hairpins are plotted according to genomic position (X-axis) and score (Y-axis). (hairpins in direct orientation represented as blue triangles; and hairpins in position (X-axis) and reverse orientation represented as green diamonds). The hairpin score indicates which hairpins (based on the secondary structures, minimum free energy, and positional entropy) are the best candidates as pre-miRnas.
Figure 3. The MFE, secondary structure and positional entropy of the predicted pre-miRnas were generated using RNAFold. L: L segment, M: M segment and S: S segment.
Figure 3. (Continued).
Table 1. Total number of pre-miRnas like molecules and real pre-miRnas in the RVFV genome.
Table 2. Minimum free energy (MFE) for real pre-miRnas in the RVFV genome.
Table 3. Sequence of the mature miRnas.
Furthermore, we predicted sheep 3′-and 5′-UTR targeted using RNAhybrid and Miranda. miRNA targets were functionally annotated in AmiGO, ShinyGO, and Enrichr using gene ontology (GO) terms ().
Table 4. Functionally annotated sheep gene targets of RVFV miRnas.
The main target genes for 19 of the real pre-miRNAs were Ash1-like histone lysine methyltransferase and astrotactin 2. Ash1-like histone lysine methyltransferases belong to the trithorax group, a group of transcriptional activators involved in body segment identity.
Idh1 and Ifnar1 have been identified as the main targets of miRNAs. Idh1 refers to Isocitrate Dehydrogenase 1, an enzyme involved in cellular metabolism and the regulation of cellular redox balance. Ifnar1 , on the other hand, represents Interferon Alpha and Beta Receptor Subunit 1, a receptor that plays a crucial role in the cellular response to interferons, including interferon-alpha and interferon-beta. These genes are significant targets of miRNAs, suggesting their potential involvement in the regulatory networks and biological processes modulated by these miRNAs.
Furthermore, we found that astrotactin 2, another miRNA target, a perforin-like integral membrane protein, is a vertebrate-specific gene that plays an important role in neurodevelopment. Other notable miRNA targets included Oxt, Cav1, Tgf-beta 3, Mip, Kitlg, Cyp1a1, Actb, Krt25, Csn1s2, Oxtr, and Ash1l. Each of these genes represents a potential regulatory target of miRNAs, suggesting their involvement in various biological processes and signalling pathways ().
Discussion
Using a novel computational approach with a coordinated use of home scripts, the VMir analyser program, the ViennaRNA package, and Maturebayes, combined with the miPred method and Mfold-related features, we identified a total of 19 miRNAs that originated from RVFV genome sequences. Subsequently, a pairwise comparison of whole ovine 3′/5’-UTRs against these 19 mature miRNAs was conducted to detect the target genes in the ovine host genome. For this purpose, an miRanda algorithm was applied, which includes thermodynamic stability calculations of miRNA:mRNA duplexes in combination with the miRDB database and Enrichr, a web-based tool for gene over-representation analysis.
Two main targets were identified: First, Ash-1 like histone lysine methyltransferase, which is involved in chromatin epigenetic modification and regulation of homeobox (Hox) genes, a class of transcription factors that are key regulators during developmental processes such as regional specification, patterning, and differentiation. Histone modifications across Hox gene clusters play a key role in transcriptionally activating Hox genes [Citation18]. Dysregulation of the Hox gene either by mutations or repression by Ash1-like histone lysine methyltransferase may be responsible for foetal malformations, including limb malformations seen in humans [Citation19] or mice [Citation20]. The same or similar deregulation of the interplay between Ash-1 and Hox genes may be caused by RVF virus-derived mechanisms after intrauterine infection, leading to disruption of the normal body shape of the O. aries foetus or even miscarriage. In this context, RVF-induced malformations have been frequently observed both in the field [Citation21] and after vaccination with vaccine strains MP-12 [Citation22] and clone 13 [Citation23].
Furthermore, we found Astrotactin 2 (Astn-2) as another main miRNA target, a perforin-like transmembrane membrane protein. It is a vertebrate-specific gene that plays a central role in cerebellar development and is expressed in migrating cerebellar granule neurons during glia-driven migration [Citation24]. Defects in neuronal migration are associated with several developmental malformations [Citation25]. As a precisely coordinated expression pattern of Astn-2 is required for normal cortical development, the interference of RVF-induced miRNAs with Astn-2 could cause similar neuronal migration defects in sheep brain development up to stillbirths.
In the context of RVF-induced developmental anomalies, a recent publication, which presented transcriptomic data of the human host response following RVFV infection, demonstrated downregulation of the host gene for miR-17-92 [Citation26]. The miR-17–92 family collectively encodes a total of 15 miRNAs that are essential for the development of vertebrates, and deregulation is involved in the pathogenesis of a variety of human diseases, especially congenital developmental disorders [Citation27].
RVF strain MP-12 caused downregulation of host-encoded miRNA-201 (MIR201) after infection of human HEK293 cells. MIR201 is involved in numerous biological processes, including mitochondrial metabolism, angiogenesis, cell proliferation and apoptosis [Citation28]. For example, miR-201 acts as an anti-apoptotic factor in endothelial (HUVEC) cells under oxidative stress by reducing reactive oxygen species (ROS) generation and downregulating the CASP8AP2 pathway [Citation29]. This counteracts the mode of RVF virus action, in which infection of liver cells, triggered by RVF NSs protein, leads to an increase in ROS.
To date, the presence of RNA virus miRNAs has been confirmed in various RNA virus families, including Retroviridae, Orthomyxoviridae, Flaviviridae, Filoviridae, and Coronaviridae [Citation3]. Although recent evidence has demonstrated the existence of miRNAs encoded by RNA viruses, their physiological significance in viral life cycles and pathogenesis remains largely unclear, necessitating further investigation. The reliability of the reported data is limited because of the low expression levels of certain miRNAs and the available detection methods. Although recent advancements in comprehensive computational prediction have facilitated the identification of miRNAs with low expression levels, current algorithms still generate a substantial number of false positives [Citation30]. Therefore, a comprehensive approach integrating computational and experimental methodologies is essential to validate existing data and gain further insights into the role of miRNAs in RNA virus biology.
In our study, several target functions were attributed to foetal malformations, abortions, and neurodevelopment. The identification of significantly enriched O. aries gene clusters in RVFV infection will aid the elucidation of the molecular mechanisms underlying RVFV pathogenesis and potentially novel drug targets for RVFV. Further experimental validation is required to confirm the expression of these pre-miRNAs.
Conclusion
This is the first report of the identification of pre-miRNAs in RVFV using an in-silico approach. From the RVFV genome, we identified 727 new pre-miRNA candidates, none of which have been reported previously. In this study, the bias function kernel was used as a support vector machine (SVM) model to predict pre-miRNA genes. To feed the SVM, structural and thermodynamic characteristics were extracted from pre-miRNA genes. To validate our method, we used VMir software. Briefly, our classifier is suitable for predicting RVFV pre-miRNAs. The large number of novel pre-miRNA candidates indicates that there are many unidentified pre-miRNAs in RVFV. In our study, we considered pre-miRNA candidates as real RVFV pre-miRNAs using bioinformatics. Further experiments are required to validate the expression of these pre-miRNAs. The identification of novel pre-microRNA genes in RVFV through in-silico methods offers a promising avenue for investigating pathomechanisms linked to embryo-foetal dysgenesis. The findings from this study enhance our comprehension of the intricate molecular interactions within the virus, illuminating potential factors that contribute to developmental abnormalities. Furthermore, our research not only suggests the need for future experimental verification of RVFV pre-miRNA but also provides a framework for exploring gene function and the genetic mechanisms underlying complex traits at the genomic level.
Author contributions
Conceptualization, B.S., M.H.G., and M. E.; methodology, B.S. and M.E; data curation, B.S. and M.E; software, B.S.; validation, B.S., M.H.G., and M. E.; formal analysis, B.S. and M.E; writing original draft preparation, B.S., M.H.G., and M.E; writing review and editing, B.S., M.H.G., and M.E; visualization, B.S. and M.E; funding acquisition, M.H.G. and M.E. All authors have read and agreed to the published version of the manuscript.
Acknowledgements
We are grateful to Felicitas Bergmann (FLI) for producing .
Disclosure statement
No potential conflict of interest was reported by the author(s).
Data Availability statement
All data generated or analysed during this study are included in this published article and its Supporting Information files.
Additional information
Funding
References
- Adhami M, MotieGhader H, Haghdoost AA, et al. Gene co-expression network approach for predicting prognostic microRNA biomarkers in different subtypes of breast cancer. Genomics. 2020;112(1):135–10. doi: 10.1016/j.ygeno.2019.01.010
- Sadeghi B, Ahmadi H, Azimzadeh-Jamalkandi S, et al., BosFinder: a novel pre-microRNA gene prediction algorithm in Bos taurus. Anim Genet, 2014. 45(4): 479–484. 10.1111/age.12170
- Nanbo A, Furuyama W, Lin Z. RNA virus-encoded miRnas: current insights and future challenges. Front Microbiol. 2021;12:12. doi: 10.3389/fmicb.2021.679210
- Kozomara A, Birgaoanu M, Griffiths-Jones S. miRBase: from microRNA sequences to function. Nucleic Acids Res. 2019;47(D1):D155–D162. doi: 10.1093/nar/gky1141
- Pepin M, Bouloy M, Bird BH, et al., Rift Valley fever virus (Bunyaviridae: Phlebovirus): an update on pathogenesis, molecular epidemiology, vectors, diagnostics and prevention. Vet Res, 2010. 41(6): 61. 10.1051/vetres/2010033
- Rissmann M, Stoek F, Pickin MJ, et al. Mechanisms of inter-epidemic maintenance of Rift Valley fever phlebovirus. Antiviral Res. 2020;174:104692. doi: 10.1016/j.antiviral.2019.104692
- Diosa-Toro M, Echavarría-Consuegra L, Flipse J, et al. MicroRNA profiling of human primary macrophages exposed to dengue virus identifies miRNA-3614-5p as antiviral and regulator of ADAR1 expression. PLoS Negl Trop Dis. 2017;11(10):e0005981. doi: 10.1371/journal.pntd.0005981
- Geddes VEV, de Oliveira AS, Tanuri A, et al., MicroRNA and cellular targets profiling reveal miR-217 and miR-576-3p as proviral factors during oropouche infection, 2018. 12(5): e0006508. PLoS negl trop dis 10.1371/journal.pntd.0006508
- Havranek KE, White LA, Lanchy J-M, et al. Transcriptome profiling in Rift Valley fever virus infected cells reveals modified transcriptional and alternative splicing programs. PloS One. 2019;14(5):e0217497. doi: 10.1371/journal.pone.0217497
- Grundhoff A. Computational prediction of viral miRnas. In: Antiviral RNAi. Springer; 2011. pp. 143–152.
- Hofacker IL, Fontana W, Stadler PF, et al., Fast folding and comparison of RNA secondary structures. Monatshefte für Chemie/Chem Monthly, 1994. 125(2): 167–188. 10.1007/BF00818163
- Lorenz R, Bernhart SH, Höner Zu Siederdissen C, et al. ViennaRNA package 2.0. algorithms for molecular biology. Algorithms Mol Biol. 2011;6(1):1–14. doi: 10.1186/1748-7188-6-26
- Jiang P, Wu H, Wang W, et al., MiPred: classification of real and pseudo microRNA precursors using random forest prediction model with combined features. Nucleic Acids Res, 2007. 35(suppl_2): W339–W344. 10.1093/nar/gkm368
- Gkirtzou K, Tsamardinos I, Tsakalides P, et al. MatureBayes: a probabilistic algorithm for identifying the mature miRNA within novel precursors. PLoS One. 2010;5(8):e11843. doi: 10.1371/journal.pone.0011843
- Lewis BP, Burge CB, Bartel DP. Conserved seed pairing, often flanked by adenosines, indicates that thousands of human genes are microRNA targets. Cell. 2005;120(1):15–20. doi: 10.1016/j.cell.2004.12.035
- Lewis BP, et al., Prediction of mammalian microRNA targets. Cell, 2003. 115(7): 787–798.10.1016/S0092-8674(03)01018-3
- Rehmsmeier M, steffen P, Höchsmann M, et al. Fast and effective prediction of microRna/target duplexes. RNA. 2004;10(10):1507–1517. doi: 10.1261/rna.5248604
- Afzal Z, Krumlauf R, Transcriptional regulation and implications for controlling hox gene expression. J Dev Biol, 2022. 10(1): 4. 10.3390/jdb10010004
- Quinonez SC, Innis JW. Human HOX gene disorders. molecular genetics and metabolism. Mol Gene Metabol. 2014;111(1):4–15. doi: 10.1016/j.ymgme.2013.10.012
- Brinkmeier ML, Geister KA, Jones M, et al. The histone methyltransferase gene absent, small, or homeotic discs-1 like is required for normal hox gene expression and fertility in mice. Biol Reprod. 2015;93(5):121, 1–12. doi: 10.1095/biolreprod.115.131516
- Coetzer JA, Barnard B. Hydrops amnii in sheep associated with hydranencephaly and arthrogryposis with wesselsbron disease and rift valley fever viruses as aetiological agents. Onderstepoort J Vet Res. 1977;44(2):119–26.
- Hunter P, Erasmus B, Vorster JH. Teratogenicity of a mutagenised rift valley fever virus (MVP 12) in sheep. Onderstepoort J Vet Res. 2002;69(1):95–8.
- Makoschey B, van Kilsdonk E, Hubers WR, et al. Rift Valley fever vaccine virus clone 13 is able to cross the ovine placental barrier associated with foetal infections, malformations, and stillbirths. PLoS Negl Trop Dis. 2016;10(3):e0004550. doi: 10.1371/journal.pntd.0004550
- Wilson PM, Fryer RH, Fang Y, et al. Astn2, a novel member of the astrotactin gene family, regulates the trafficking of ASTN1 during glial-guided neuronal migration. J Neurosci. 2010;30(25):8529–8540. doi: 10.1523/JNEUROSCI.0032-10.2010
- Subramanian L, Calcagnotto ME, Paredes MF. Cortical malformations: lessons in human brain development. Front Cell Neurosci. 2020;13:576. doi: 10.3389/fncel.2019.00576
- Pinkham C, Ahmed A, Bracci N, et al. Host-based processes as therapeutic targets for rift valley fever virus. Antiviral Res. 2018;160:64–78. doi: 10.1016/j.antiviral.2018.10.004
- Concepcion CP, Han Y-C, Mu P, et al. Intact p53-dependent responses in miR-34–deficient mice. PLoS Genet. 2012;8(7):e1002797. doi: 10.1371/journal.pgen.1002797
- Yuan Y, Fei M, Yang Y-X, et al., MiR-201-5p alleviates lipopolysaccharide-induced renal cell dysfunction by targeting NOTCH3. Eur Rev Med Pharmacol Sci, 2020. 24(10): p. 5592–5603. 10.26355/eurrev_202005_21345
- Li T, Song X, Zhang J, et al. Protection of human umbilical vein endothelial cells against oxidative stress by MicroRNA. Oxid Med Cell Longevity. 2017;2017:1–11. doi: 10.1155/2017/3565613
- Pinzón N, Li B, Martinez L, et al. microRNA target prediction programs predict many false positives. Genome Res. 2017;27(2):234–245. doi: 10.1101/gr.205146.116