Vaccinia and influenza A viruses select rather than adjust tRNAs to optimize translation

Abstract

Transfer RNAs (tRNAs) are central to protein synthesis and impact translational speed and fidelity by their abundance. Here we examine the extent to which viruses manipulate tRNA populations to favor translation of their own genes. We study two very different viruses: influenza A virus (IAV), a medium-sized (13 kB genome) RNA virus; and vaccinia virus (VV), a large (200 kB genome) DNA virus. We show that the total cellular tRNA population remains unchanged following viral infection, whereas the polysome-associated tRNA population changes dramatically in a virus-specific manner. The changes in polysome-associated tRNA levels reflect the codon usage of viral genes, suggesting the existence of local tRNA pools optimized for viral translation.

Figure 1.

Codon usage of virus compared with human. Viral codon usage (IAV in gray, VV in black) is plotted against human codon usage. Codon usage is expressed as frequency per 1000 codons (top) or RSCU (bottom). Regardless of the index used, viral codon usage correlates poorly with human codon usage. When codon usage is expressed as frequency per 1000, the R² correlation coefficient between viral and human codon usage is 0.26 for IAV and 0.01 for VV. When codon usage is expressed as RSCU, the R² correlation coefficient is 0.10 for IAV and 0.08 for VV.

Figure 2.

Changes in tRNA abundance after viral infection. HeLa cells were infected with IAV or VV; total cellular RNA or polysome RNA was isolated 6 h post-infection. tRNA abundance was measured by microarray relative to an uninfected control. Data are averages of three replicate experiments; error bars indicate standard deviation. (A) Median tRNA abundance after viral infection. Median values for nuclear-encoded tRNAs (black) and mitochondrial-encoded tRNAs (gray) are shown for IAV- and VV-infected cells relative to an uninfected control (set to 1). No mitochondrial-encoded tRNAs were detected in the polysome RNA samples. No significant changes in median tRNA abundance are detected in total cellular RNA (left) or polysome RNA (right). (B) Individual tRNA abundance after viral infection. Individual tRNA abundance values are shown for IAV (gray) and VV (black) infected cells relative to an uninfected control (set to 1, black line). A value of 1 indicates no change, a value <1 indicates a decrease and a value >1 indicates an increase after viral infection. No significant changes in individual tRNA abundances are detected in total cellular RNA (top), but distinct and virus-specific changes are observed in polysome RNA (bottom). One sample t-tests were performed to determine the statistical significance of the changes: * indicates P-value <0.0014 applying the Bonferroni correction for measuring multiple events (P-value/number of events, or 0.05/37).

Figure 3.

Polysome tRNAs reflect viral translation. (A) Translation patterns after viral infection. HeLa cells were infected with IAV, VV or mock-treated for the uninfected control. At 6 h post-infection, cells were pulsed with S³⁵-methionine and lysed. Cell lysate was loaded on an SDS-PAGE gel for visualization of S³⁵-labeled protein products (left). Signal intensity was quantitated to better compare the infected (IAV or VV) and the control samples (middle and right panels). Translation is significantly altered upon viral infection. However, VV is far more efficient than IAV in shutting down host translation (R² IAV/Ctrl > R² VV/Ctrl). (B) Quantitation of translation after viral infection. Cell lysate was obtained as in (A) and S³⁵-labeled protein products were quantitated on a scintillation counter after TCA precipitation. Values are averages of six technical replicates, error bars indicate standard deviation. A modest increase in translation is observed in IAV-infected cells but not VV-infected cells relative to the uninfected control. (C) tRNA—codon usage analysis. Relative tRNA abundance values after viral infection (IAV left, VV right) measured by microarray are plotted against normalized viral codon usage. Polysome tRNA values (gray) correlate well with viral codon usage, whereas total tRNA values (black) do not correlate with viral codon usage.

Figure 4.

Changes in tRNA abundance after IFN treatment. HeLa cells were treated with IFN-β or IFN-γ for 16 h and total cellular RNA was isolated. tRNA abundance was measured by microarray relative to an untreated control. (A) Median tRNA abundance after IFN treatment. Median values for nuclear-encoded tRNAs (black) and mitochondrial-encoded tRNAs (gray) are shown for IFN-β or IFN-γ−treated cells relative to an untreated control (set to 1). IFN-β slightly decreases global nuclear-encoded tRNA levels, whereas IFN-γ treatment increases global nuclear-encoded tRNA levels. (B) Individual tRNA abundance after IFN treatment. Individual tRNA abundance values are shown for IFN-β (black) or IFN-γ (gray) treated cells relative to an untreated control (set to 1, black line). Data are averages from one dye-swapped experiment; error bars indicate standard deviation. A value of 1 indicates no change, a value <1 indicates a decrease and a value >1 indicates an increase after IFN treatment relative to untreated. tRNA^Ile(UAU) is markedly increased upon IFN-β and IFN-γ treatment (black arrow). One sample t-tests were performed to determine the statistical significance of the changes: * indicates P-value <0.0009, applying the Bonferroni correction for measuring multiple events (P-value/number of events, or 0.05/58). (C) tRNA—codon usage analysis. Relative tRNA abundance values after IFN treatment (IFN-β in black and IFN-γ in gray) are plotted against normalized IAV codon usage. The Ile-AUA codon is over-represented in the IAV genome, correlating with an increase in tRNA^Ile(UAU) abundance after IFN treatment (black arrow). Similar results are obtained when plotting against VV codon usage (not shown). These data were referred in (25) but not shown.