Real-time transcriptional profiling of cellular and viral gene expression during lytic cytomegalovirus infection

Abstract

During viral infections cellular gene expression is subject to rapid alterations induced by both viral and antiviral mechanisms. In this study, we applied metabolic labeling of newly transcribed RNA with 4-thiouridine (4sU-tagging) to dissect the real-time kinetics of cellular and viral transcriptional activity during lytic murine cytomegalovirus (MCMV) infection. Microarray profiling on newly transcribed RNA obtained at different times during the first six hours of MCMV infection revealed discrete functional clusters of cellular genes regulated with distinct kinetics at surprising temporal resolution. Immediately upon virus entry, a cluster of NF-κB- and interferon-regulated genes was induced. Rapid viral counter-regulation of this coincided with a very transient DNA-damage response, followed by a delayed ER-stress response. Rapid counter-regulation of all three clusters indicated the involvement of novel viral regulators targeting these pathways. In addition, down-regulation of two clusters involved in cell-differentiation (rapid repression) and cell-cycle (delayed repression) was observed. Promoter analysis revealed all five clusters to be associated with distinct transcription factors, of which NF-κB and c-Myc were validated to precisely match the respective transcriptional changes observed in newly transcribed RNA. 4sU-tagging also allowed us to study the real-time kinetics of viral gene expression in the absence of any interfering virion-associated-RNA. Both qRT-PCR and next-generation sequencing demonstrated a sharp peak of viral gene expression during the first two hours of infection including transcription of immediate-early, early and even well characterized late genes. Interestingly, this was subject to rapid gene silencing by 5-6 hours post infection. Despite the rapid increase in viral DNA load during viral DNA replication, transcriptional activity of some viral genes remained remarkably constant until late-stage infection, or was subject to further continuous decline. In summary, this study pioneers real-time transcriptional analysis during a lytic herpesvirus infection and highlights numerous novel regulatory aspects of virus-host-cell interaction.

Figure 1. Establishment of 4sU-tagging for lytic MCMV infection.

(A) Incorporation of 4sU throughout MCMV infection. Cells were infected with MCMV at an MOI of 10 and exposed to 200 µM 4sU for 1 h at different times of infection before total RNA was isolated. Thiol-specifically biotinylated RNA was subjected to dot blot analysis in 10-fold dilutions (1 µg down to 1 ng). A biotinylated oligonucleotide of 81 nt (PC, 100 ng down to 0.1 ng) was used to quantify 4sU-incorporation; M = mock control. (B)–(D) Comparison of genes identified to be regulated in newly transcribed RNA to genes regulated in total RNA. (B) Numbers of genes up- and down-regulated (>2-fold, p≤0.05) at different times of infection are shown for newly transcribed RNA and total RNA. (C) Venn diagrams of all genes regulated more then 2-fold in newly transcribed RNA and total RNA. (D) Venn diagrams showing genes regulated >2-fold in total RNA at 2, 4 and 6 hpi and in newly transcribed RNA at and prior to the indicated time point of infection; red = newly transcribed RNA, blue = total RNA.

Figure 2. Gene expression kinetics define distinct functional clusters.

(A) Heat-maps indicating the fold-changes are shown as matrices with rows representing genes and columns representing the time points post infection. Red represents up-regulation, blue down-regulation (>2-fold, p≤0.01) in newly transcribed RNA relative to uninfected cells. Ordering of genes in the heat-maps was determined using non-supervised hierarchical clustering. Shown are the 5 clusters of genes we identified. (B) All clusters are associated with distinct functional annotations. Enrichment analysis results of Gene Ontology ‘Biological Process’ terms and KEGG pathways are shown for each of the five clusters with the most significant (p≤0.01) categories displayed in the graphs as bars, sorted from bottom (most significant) to top. To reduce redundancy, similar terms are represented by the most significant and specific term. For complete list of functional annotations see Table S2a. (C) Specific transcription factor binding sites correlate with functional clusters. Shown are exemplary transcription factors with over-represented binding sites unique for the different clusters. For a complete list of over-represented transcription factor weight matrices see Table S2b. Illustrated are the transcription factor weight matrices, the percentage of promoters with sites and p-value.

Figure 3. Validation of exemplary transcription factors.

NIH-3T3 fibroblasts were infected with MCMV at an MOI of 10 for the indicated time points and lysates were prepared for western blot analysis (A), for immune staining (B) or luciferase assay (C). Western blot analysis was performed on samples prepared from uninfected and infected NIH-3T3 cells probed for RelA and IkBα (A). GAPDH was probed as loading control. For the immunofluorescence staining (B) cells were fixed and stained with the indicated antibodies; white circle indicating nucleus, nuclear dimensions were acquired by DAPI staining and its outline was overlaid into the shown channels; green, RelA; red, viral IE1. For the luciferase assays cells were transfected with a c-Myc-reporter construct (C) and infected 48 hours post transfection with MCMV at an MOI of 10. At the indicated times post infection, Firefly-Luciferase measurements were performed in triplicates. Shown is the mean +/− SD of a representative of three experiment; mpi = minutes post infection, hpi = hours post infection.

Figure 4. Identification of virus-specific regulation and counter-regulation using UV-inactivated virus.

Quantitative RT-PCR was employed to measure transcription rates of exemplary genes of the five gene clusters (A–F). NIH-3T3 were infected with wild-type (wt) or UV-inactivated (1500 J, 15 min) MCMV (MOI of 10) for the indicated time points. Displayed are the fold-changes relative to uninfected cells normalized to Lbr. Fold-changes in between 0.5- and 2-fold were considered as non-regulated. Shown are the combined data (means +/− SD) of three independent experiments.

Figure 5. Real-time kinetics of viral gene expression.

(A) 4sU-tagging allows efficient removal of viral DNA and virion-associated RNA. Newly transcribed RNA was labeled with 200 µM 4sU for 1 h in MCMV-infected NIH-3T3 cells at −1 to 0 (mock), 1–2, 3–4 and 7–8 hpi. As a negative control, Actinomycin-D was added to cells prior to infection to block transcription and thus 4sU-incorporation. Total RNA was isolated, treated with DNaseI and newly transcribed RNA was purified. qRT-PCR analysis was performed on newly transcribed RNA for viral ie1 and cellular Lbr. Shown are the combined data (means +/− SD) of three independent experiments. (B–F) Gene expression kinetics of exemplary viral genes. Shown are qRT-PCR measurements of newly transcribed RNA for ie1 (B), the early genes m169 (C) and m152 (D) as well as for the late genes m129/131 (E) and M94 (F). Synthesis rates were normalized to Lbr expression. Shown are the combined data (means +/− SD) of three independent experiments. (G) Contribution of viral transcripts to all coding sequence reads (CDS). RNA-seq was performed on newly transcribed, total and unlabeled pre-existing RNA samples (n = 1). Reads were mapped to both the cellular and viral transcriptome/genome. The contribution of viral reads to all CDS reads at different times of infection is shown.

Figure 6. Analysis of MCMV gene expression by RNA-seq.

Shown are the coverages of all viral genes across the MCMV genome obtained by RNA-seq on newly transcribed RNA from uninfected cells, 1–2 hpi, 5–6 hpi, 11–12 hpi, 18–19 hpi, 24–25 hpi and 47–48 hpi, normalized to the total number of reads mapped to mouse exons and annotated MCMV coding sequences. Genomic positions of exemplary genes as well as concordantly regulated gene clusters are marked; SCP = small capsid protein (m48.2), POL = polymerase (M54). Blue represents transcripts matching the direct DNA strand and red the complementary DNA strand.