Asian Zika Virus Isolate Significantly Changes the Transcriptional Profile and Alternative RNA Splicing Events in a Neuroblastoma Cell Line

Abstract

The alternative splicing of pre-mRNAs expands a single genetic blueprint to encode multiple, functionally diverse protein isoforms. Viruses have previously been shown to interact with, depend on, and alter host splicing machinery. The consequences, however, incited by viral infection on the global alternative slicing (AS) landscape are under-appreciated. Here, we investigated the transcriptional and alternative splicing profile of neuronal cells infected with a contemporary Puerto Rican Zika virus (ZIKV^PR) isolate, an isolate of the prototypical Ugandan ZIKV (ZIKV^MR), and dengue virus 2 (DENV2). Our analyses revealed that ZIKV^PR induced significantly more differential changes in expressed genes compared to ZIKV^MR or DENV2, despite all three viruses showing equivalent infectivity and viral RNA levels. Consistent with the transcriptional profile, ZIKV^PR induced a higher number of alternative splicing events compared to ZIKV^MR or DENV2, and gene ontology analyses highlighted alternative splicing changes in genes associated with mRNA splicing. In summary, we show that ZIKV affects cellular RNA homeostasis not only at the transcriptional levels but also through the alternative splicing of cellular transcripts. These findings could provide new molecular insights into the neuropathologies associated with this virus.

Keywords: Flavivirus; SH-SY5Y; Zika virus; alternative splicing; transcriptome.

Figure 1

SH-SY5Y cells are infected by ZIKV and DENV2. (A) Immunofluorescence images of virus-infected SH-SY5Y cells. SH-SY5Y cells seeded in 48-well plates were infected at a moi of 5 and fixed 24 h post infection. Virus-infected cells were visualized using an antibody that detects the replication intermediate dsRNA, and all cells in the field of view were visualized by staining cell nuclei with Hoechst and at 10× magnification. (B) Quantification of the percentage of infected cells. Uninfected and virus-infected SH-SY5Y cells seeded in 48-well plates were fixed and stained for dsRNA and Hoechst. A 2F0× objective was used to image three sections per well per virus where more than 400 cells for each independent infection were counted. The percentages of infected cells were determined. At least three biological replicates were performed. (C) Titers of virus released into the medium from ZIKV and DENV2-infected SH-SY5Y cells were determined by plaque assays. (D) RT-qPCR analysis of SH-SY5Y cells infected at a moi of 5. Primers targeting the coding regions of each virus were used along with primers for β-actin mRNA. Relative viral RNA levels were calculated by standardizing relative fluorescent units at Ct for each virus against β-actin mRNA. Error bars represent standard deviations established from three independent infections. No significant difference was determined for the number of cells counted in panel 1B, or the relative abundance of viral RNA in panel 1D. For statistical analysis, two-tailed student T-tests were performed (* p < 0.05; ** p < 0.01).

Figure 2

Transcriptome analysis of mock- and virus-infected SH-SY5Y cells. (A) Schematic showing the pipeline from cells to differential gene expression and alternative splicing analysis. (B) Venn diagram of differentially expressed (DE) transcripts between cells infected with different viruses versus mock-infected SH-SY5Y cells. DE transcripts from ZIKV^PR versus mock, ZIKV^MR versus mock, and DENV2 versus mock are highlighted in the red, yellow, and blue circles, respectively. (C) All differentially expressed genes for each condition were input into ShinyGO(2.0), and the top 25 biological GO terms were categorized into five types, with the distribution of each type presented in a pie chart. (D,F,H) Top five functional categories derived from statistically significant upregulated genes for the indicated condition. (E,G,I) Top five functional categories derived from statistically significant downregulated genes for the indicated condition. GO terms are annotated on the y-axes, and false discovery rates (FDR) are represented on the x-axes.

Figure 3

Upregulation and downregulation of genes in SH-SY5Y cells infected with ZIKV^PR, ZIKV^MR, and DENV2. (A) Venn diagram of genes differentially expressed between infected SH-SY5Y cells that were differentially upregulated (top) or downregulated (bottom). DE transcripts from ZIKV^PR, ZIKV^MR, and DENV2 are within the respective red, blue, and yellow circles. (B) Heatmap of the top 50 statistically significant differentially expressed immune response genes in ZIKV^PR-infected cells compared to mock. The expression of these genes in ZIKV^MR and DENV2-infected SH-SY5Y cells is shown for comparison. Three replicates from each condition were collapsed and normalized to mock. The color scale shows the Z-score of the immune response genes.

Figure 4

Analysis of alternative splicing events in virus versus mock-infected SH-SY5Y cells. (A) Schematic illustrating the five alternative splice events characterized in our analyses. The unchanged exons are black, while the differentially spliced exons are color coded. (B) Venn diagram of shared and unique genes that were alternatively spliced between virus and mock-infected cells. Alternative slicing (AS) events from ZIKV^PR versus mock, ZIKV^MR versus mock, and DENV2 versus mock are highlighted in the red, blue, and yellow circles respectively. (C) Pie charts depicting the percentages of all unique AS events for the three viruses versus mock-infected SH-SY5Y cells. (D) Pie charts representing the percentages of each type of AS event unique to each virus infection in SH-SY5Y cells. The segment colors match the AS events illustrated in (A). Values in (B) account for unique genes while (C) represents all unique splicing events.

Figure 5

Validation of select alternative splicing events in mock-, ZIKV^PR-, ZIKV^MR-, and DENV2-infected SH-SY5Y cells. (A–H) RNA from each of the three biological replicates was used for RT-PCR for the indicated gene. The percent spliced in (PSI) was calculated as described in the Materials and Methods. PSI indicates the change in the inclusion of a regulated exon. The significance of the data was determined from three independent experiments using the GraphPad software and performing an unpaired Student’s t-test; n.s. denotes not significant, * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001. Figure S6 shows the corresponding schematics of the gene exons examined as well as the RT-PCR products and Sashimi plots for the genes analyzed in this figure. Figure S7 shows the correlation between the PSI values obtained from RNA-seq and RT-PCR.