RNA-Seq Revealed a Circular RNA-microRNA-mRNA Regulatory Network in Hantaan Virus Infection

Abstract

Hantaan virus (HTNV), a Hantavirus serotype that is prevalent in Asia, causes hemorrhagic fever with renal syndrome (HFRS) with high mortality in human race. However, the pathogenesis of HTNV infection remains elusive. Circular RNAs (circRNAs), a new type of non-coding RNAs, play a crucial role in various pathogenic processes. Nevertheless, circRNA expression profiles and their effects on pathogenesis of HTNV infection are still completely unknown. In the present study, RNA sequencing was performed to analyze the circRNA, microRNA (miRNA), and mRNA expression profiles in HTNV-infected and mock-infected human umbilical vein endothelial cells (HUVECs). A total of 70 circRNAs, 66 miRNAs, and 788 mRNAs were differently expressed. Several differentially expressed RNAs were validated by RT-qPCR. Moreover, we verified that some differentially expressed RNAs, such as circ_0000479, miR-149-5p, miR-330-5p, miR-411-3p, RIG-I, CMPK2, PARP10, and GBP1, promoted or inhibited HTNV replication. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis demonstrated that the host genes of differentially expressed circRNAs were principally involved in the innate immune response, the type I interferon (IFN) signaling pathway, and the cytokine-mediated signaling pathway. Additionally, the circRNA-miRNA-mRNA regulatory network was integrally analyzed. The data showed that there were many circRNA-miRNA-mRNA interactions in HTNV infection. By dual-luciferase reporter assay, we confirmed that circ_0000479 indirectly regulated RIG-I expression by sponging miR-149-5p, hampering viral replication. This study for the first time presents a comprehensive overview of circRNAs induced by HTNV and reveals that a network of enriched circRNAs and circRNA-associated competitive endogenous RNAs (ceRNAs) is involved in the regulation of HTNV infection, thus offering new insight into the mechanisms underlying HTNV-host interaction.

Keywords: HTNV; cellular response changes; circular RNA; competing endogenous RNA network; microRNA; viral replication.

Graphical Abstract

In this work, HUVECs were mock infected or infected with HTNV for 3 days. Total RNA of the cells were analyzed by RNA-seq and obtained circRNA, miRNA, mRNA library. Differentially expressed (DE) RNAs were identified and subjected to GO analysis, KEGG analysis and ceRNA network construction. Then 8 DE circRNAs, 8 DE mRNAs, 6 DE miRNAs were verified by RT-qPCR. Besides, mRNAs (CMPK2, PARP10, GBP1, and RIG-I), circRNAs (circ_0000479), miRNAs (miR-149-5p, miR-411-3p, and miR-330-5p) in the ceRNA network were found effective to inhibit or promote virus replication. And the circ_0000479-miR-149-5p-RIG-I ceRNA axis was verified in HTNV infection.

Figure 1

circRNA expression overview. (A) circRNAs category chart. (B) circRNAs length distribution. (C) Volcano plot of DE circRNAs upon HTNV infection in HUVECs. Red dots represent up-regulated circRNAs and green dots represent down-regulated circRNAs. (D) Chromosome distribution of DE circRNAs. (E) Heatmap and clustering analysis of DE circRNAs. Each row represents one circRNA and each column represents one sample; −2, −1, 0, 1, and 2 represent fold change. Red indicates high expression and blue represents low expression. CON-1, CON-2 and CON-3 represent three mock-infected samples; HTNV-1, HTNV-2, and HTNV-3 represent three HTNV-infected samples. (F) Verification of dysregulated circRNAs. HUVECs were infected with HTNV 76-118 (MOI = 1) for 3 days. Then the total RNA was extracted and the expression levels of circRNAs were measured by RT-qPCR. Student's t-test, mean ± standard deviation (SD), *P < 0.05; **P < 0.01; ***P < 0.001. The experiment was performed at least three times independently.

Figure 2

Verification of DE circRNAs after HTNV infection and exploration of their effect on HTNV infection. (A) GO functional enrichment analysis of parental genes of DE circRNAs. The x-axis shows the P-value and gene numbers, and the y-axis shows the GO term. (B) The 30 most enriched KEGG pathways based on hosting genes of dysregulated circRNA during HTNV infection. The x-axis shows the enrichment factor, and the y-axis shows the pathway names. The point size represents the number of genes enriched in a particular pathway. (C) The silencing efficiencies of siRNAs targeting the connected sites of circRNAs. Student's t-test, mean ± standard deviation (SD), *P < 0.05; **P < 0.01; ***P < 0.001. (D) Western blot assay of HTNV NP expression in HUVECs with knockdown of the circRNAs circ_0000479, circ_0006132, circ_0007793, and circ_0046034. One-way ANOVA with Dunnett's multiple comparison test. Mean ± SD, *P < 0.05; **P < 0.01; ***P < 0.001; ns, no significance. The experiments were performed at least three times independently.

Figure 3

Identification of dysregulated mRNAs. (A) Heatmap and clustering analysis of DE mRNAs. Each row represents one mRNA and each column represents one sample; −2, −1, 0, 1, and 2 represent fold change. Red indicates high expression and blue represents low expression. CON-1, CON-2, and CON-3 represent three mock-infected samples; HTNV-1, HTNV-2, and HTNV-3 represent three HTNV-infected samples. (B) DE mRNAs count. (C) Verification of dysregulated mRNAs. HUVECs were infected with HTNV 76-118 for 3 days (MOI = 1). Then the total RNA was extracted and expression levels of the selected mRNAs were measured by RT-qPCR. Student's t-test, mean ± SD, ***P < 0.001. The experiment was performed at least three times independently. (D) GO functional enrichment analysis of DE genes. (E) Top 10 KEGG pathways of DE genes.

Figure 4

Identification of dysregulated miRNAs. (A) Heatmap and clustering analysis of DE miRNAs. Each row represents one miRNA and each column represents one sample; −2, −1, 0, 1, and 2 represent fold change. Red indicates high expression and blue represents low expression. CON-1, CON-2, and CON-3 represent three mock-infected samples; HTNV-1, HTNV-2, and HTNV-3 represent three HTNV-infected samples. (B) DE miRNAs count. (C) Verification of DE miRNAs upon HTNV infection. HUVECs were infected with HTNV 76-118 (MOI = 1) for 3 days. Then the total RNA was extracted and the expression levels of miRNAs were measured by RT-qPCR. Student's t-test, mean ± SD, ***P < 0.001. The experiment was performed at least three times independently. (D) GO functional enrichment analysis of target genes of DE miRNAs. The x-axis shows the P-value and gene numbers, and the y-axis shows the GO term. (E) KEGG pathway analysis revealed the top 10 enriched pathways of target genes of DE miRNAs.

Figure 5

circRNA-miRNA-mRNA interaction network. (A) circRNA-miRNA regulatory network. Circles represent circRNA and triangles represent miRNA. (B) miRNA-mRNA regulatory network. Triangles represent miRNA and dots represent mRNA. (C) ceRNA co-expression network. Circles represent circRNA, triangles represent miRNA, and rhombi represent mRNA. Red and green represent up-regulated and down-regulated RNAs, respectively.

Figure 6

Exploration of the function of DE mRNAs and miRNAs upon HTNV infection in ceRNA network. (A) The interfering efficiencies of siRNAs targeting the different sites of mRNAs. (B) HUVECs were transfected with siCON or efficient siRNAs. 24 h after transfection, the cells were infected with HTNV at an MOI of 1 for 3 days. The expression of HTNV NP in HUVECs was measured by Western-blotting assay. (C) The expression of RIG-I and HTNV NP in HUVECs after interfering RIG-I and infected with HTNV. (D) The overexpression effect of miRNA mimics in HUVECs. (E) Western-blotting assay of HTNV NP expression in HUVECs transfected with miRNA mimics. Student's t-test, One-way ANOVA with Dunnett's multiple comparison test. Means ± SD, *P < 0.05; **P <0.01; ***P < 0.001; ns, no significance. The experiments were performed at least three times independently.

Figure 7

Verification of circRNA-miRNA-mRNA axis. (A) Putative binding sites of miR-149-5p to RIG-I and circ_0000479. (B) 293T cells were co-transfected with NC mimics or miR-149-5p and GLO plasmids harboring the RIG-I binding region (RIG-I-BR) and circ_0000479 for 48 h. Cells were lysed and then subjected to dual-luciferase reporter analysis. (C) 293T cells were co-transfected with miR-149-5p or negative control, GLO-RIG-I-BR, and circ_0000479 overexpressing plasmids. Cells were lysed and then subjected to dual-luciferase reporter analysis. Student's t-test, one-way ANOVA with Bonferroni's correction comparison test. Means ± SD, *P < 0.05; **P < 0.01; ***P < 0.001; ns, no significance. The experiments were performed at least three times independently.