Multi-platform omics analysis of Nipah virus infection reveals viral glycoprotein modulation of mitochondria

Abstract

The recent global pandemic illustrates the importance of understanding the host cellular infection processes of emerging zoonotic viruses. Nipah virus (NiV) is a deadly zoonotic biosafety level 4 encephalitic and respiratory paramyxovirus. Our knowledge of the molecular cell biology of NiV infection is extremely limited. This study identified changes in cellular components during NiV infection of human cells using a multi-platform, high-throughput transcriptomics, proteomics, lipidomics, and metabolomics approach. Remarkably, validation via multi-disciplinary approaches implicated viral glycoproteins in enriching mitochondria-associated proteins despite an overall decrease in protein translation. Our approach also allowed the mapping of significant fluctuations in the metabolism of glucose, lipids, and several amino acids, suggesting periodic changes in glycolysis and a transition to fatty acid oxidation and glutamine anaplerosis to support mitochondrial ATP synthesis. Notably, these analyses provide an atlas of cellular changes during NiV infections, which is helpful in designing therapeutics against the rapidly growing Henipavirus genus and related viral infections.

Keywords: CP: Microbiology; Nipah virus; host; infection; lipidomics; metabolism; metabolomics; omics; paramyxovirus; proteomics; transcriptomics.

Figure 1.. Overview of Nipah virus infection omics study

(A) Five pairs of mock and NiV infections were made at a multiplicity of infection (MOI) of 1 in HEK293T cells for each of four time points (4, 8, 12, and 16 h post-infection). All the biomolecules are listed depending on their abundance. (B) Formation of cell-cell fusion (syncytial formation) (black arrows) starting at 8 hpi and increases during infection to almost complete fusion at 16 hpi. Scale bars, 100 μm.

Figure 2.. Proteomic analyses identified highly significant increases to proteins associated with RNA processing and mitochondria

(A) Significantly increased (p < 0.05; log₂FC ≥ 0.58) proteins were presented using InteractiVenn. (B) Increased cellular processes are mapped using GOTERMFINDER. (C) Increased subcellular components and locations are mapped using REVIGO. (D) Significantly decreased (p < 0.05; log₂FC ≤ −0.58) proteins were presented using InteractiVenn. (E) Decreased cellular processes are mapped using GOTERMFINDER. (F) Decreased subcellular components and locations are mapped using REVIGO.

Figure 3.. Proteins associated with the cytosol, exosomes, and translation were most likely to be reduced during infection

Statistically significant proteins are grouped in main cellular pathways (nodes). Branches display where individual proteins are integrated. The heatmap displays the relative abundance.

Figure 4.. Transcriptional changes during late infection

(A) The microarray data are summarized in a heatmap demonstrating visual representation of transcriptional modulation during infection. (B) REVIGO is used for process enrichment, up-regulation. Log₁₀ Bonferroni-corrected p values were calculated for each GO term enrichment. (C) REVIGO is used for process enrichment, down-regulation. Log₁₀ Bonferroni-corrected p values were calculated for each GO term enrichment. (D) Limited overlap between transcriptomics and proteomics is highlighted in Venn diagrams.

Figure 5.. Summary diagram shows the metabolomic and lipidomic analyses of NiV infection

(A) All metabolites exhibiting significantly altered levels (p ≤ 0.05) at 8, 12, or 16 hpi were included. (B) Average log₂FCs of lipid sub-classes, up-regulation. (C) Average log₂FCs of lipid sub-classes, down-regulation. (B and C) The numbers in parentheses indicate how many lipid species fall into each category.

Figure 6.. Overall model for metabolic changes in an infected cell using a multi-omics approach

Data outputs from proteomics, lipidomics, and metabolomics were combined and used to construct a simplified model of cellular metabolic changes during NiV infection. Log₂-transformed FCs are shown for each time point where possible. The absence of detection for an item is shown in dark gray; non-significant change at all time points is indicated in light gray.

Figure 7.. F and G co-transfection displays an infection-like mitochondrial phenotype in proteomics analysis

(A) Expression of individual NiV proteins in HEK293T cells was detected by western blot analysis. (B) Mitochondrial content of infected and control HEK293T cells was measured using MitoTracker staining. (C) Transmission electron micrograph of F- and G-transfected HEK293T cells. Scale bars, 2 μm (left image) and 1 μm (right image). White triangles indicate mitochondria. (D) The number of mitochondria and mitochondria-like structures was counted and analyzed using ImageJ software. t-test p < 0.01. (E) Fluorescence imaging of fusion formation at 24 hpi. HEK293T cells were transfected with F, G, FG, pcDNA, and RFP-HA. HR2 peptide is included in the FG+ wells to block syncytia formation. Red is RFP-HA (cytosol) and blue is Hoechst (nuclei) staining. (F) Respirometry analysis on F- and G-transfected cells. Seahorse assay with reagents specific for oxygen consumption rate (left) was performed, with a secondary readout of extracellular acidification rate (right), indicating no increase in the metabolic rate. (G) Mitochondrial content of FG transfected into A549 cells. (H) The number of mitochondria in cells was imaged via TEM, counted, and then analyzed using GraphPad Prism. A t test yielded a p value of 0.0757. (I) qPCR analyses showing FC (2^−ΔΔCt) compared to the empty vector control (pcDNA). Statistical analyses for qPCR were performed using one-way ANOVA (****p < 0.0001).