Human cytomegalovirus infection coopts chromatin organization to diminish TEAD1 transcription factor activity

Center for Autoimmune Genomics and Etiology, Division of Human Genetics, Cincinnati Children’s Hospital Medical Center, Cincinnati, USA
Department of Molecular Genetics, Biochemistry & Microbiology, University of Cincinnati, Cincinnati, USA
Immunology Graduate Program, University of Cincinnati College of Medicine, Cincinnati, USA
Department of Medicine, Division of Infectious Diseases, Brigham and Women’s Hospital, Harvard Medical School, Boston, USA
Department of Microbiology, Harvard Program in Virology, Harvard Medical School, Boston, USA
Center for Integrated Solutions to Infectious Diseases, Broad Institute of Harvard and MIT, Cambridge, USA
Department of Pediatrics, University of Cincinnati College of Medicine, Cincinnati, USA
Research Service, Cincinnati VA Medical Center, Cincinnati, USA
Division of Allergy and Immunology, Cincinnati Children’s Hospital Medical Center, Cincinnati, USA
Division of Biomedical Informatics, Cincinnati Children’s Hospital Medical Center, Cincinnati, USA
Division of Developmental Biology, Cincinnati Children’s Hospital Medical Center, Cincinnati, USA

https://doi.org/10.7554/eLife.101578.1

Open access
Copyright information

Figures and data

Experimental design.
Schematic overview of the experimental design. A. Human foreskin fibroblasts or human retinal epithelial cells were infected with the TB40/E strain of Human Cytomegalovirus (HCMV) at a multiplicity of infection of 5 and 10, respectively. Uninfected and HCMV infected cells were harvested 48 hours post infection. B. Gene expression, chromatin accessibility, histone marks of active regulatory elements (H3K27ac), transcription factor occupancy (TEAD1 and CTCF), and chromatin looping were measured genome-wide using RNA-seq, ATAC-seq, ChIP-seq, and HiChIP, respectively. Differential analyses were employed to identify HCMV-dependent functional events on a genome-wide scale.

Extensive HCMV-mediated alterations to human chromatin accessibility.
A and B. Venn-diagram comparing ATAC-seq peaks in uninfected vs. HCMV infected fibroblasts (A) and retinal epithelial cells (B). C and D. ATAC-seq signal comparison in uninfected and HCMV infected fibroblasts (C) and retinal epithelial cells (D).

HCMV infection alters the accessibility of chromatin containing TEAD DNA binding motifs and avoids altering CTCF-containing sites.
Systematic prediction of human transcription factors with HCMV-altered binding. A. Left panel: transcription factor binding site motif enrichment comparison within ATAC-seq peaks that are unchanged with infection (X-axis) vs. peaks that are closed with infection (Y-axis). Right panel: same analysis comparing peaks that are unchanged with infection (X-axis) and peaks that are opened with infection (Y-axis). Each dot represents a human transcription factor binding site motif. Motifs are color-coded by TF family (see key). B. Same analysis in retinal epithelial cells. C. Percent of predicted binding sites for TEAD and CTCF in ATAC-seq peak regions unchanged with infection and regions closed by HCMV infection. D. Same analysis in retinal epithelial cells.

HCMV infection leads to widespread coincident loss of chromatin accessibility, TEAD1 binding, H3K27ac marks, and chromatin looping
A. HiChIP and ChIP-seq signal in the context of differentially accessible chromatin regions (ATAC-seq). Regions are split into those containing ATAC-seq signal that is unchanged (top), closed upon infection (middle), or opened upon infection (bottom). The corresponding normalized reads counts are depicted for (left to right): ATAC-seq, HiChIP, and ChIP-seq for TEAD1, CTCF, and H3K27ac. Each row in the heatmaps represents a single genomic locus. B. Genome browser images showing depletion of TEAD1, H3K27ac marks, and chromatin looping interactions proximal to Hippo pathway genes FRMD6 and RASSF2. Solid boxes highlight differential TEAD1 binding sites. The FRMD6 dashed box highlights a promoter and the RASSF2 dashed box highlights an enhancer. Chromatin looping interactions lost upon infection are highlighted in blue.

HCMV infection alters Hippo signaling gene and protein expression levels
A. Differentially expressed genes in fibroblasts with HCMV infection. B. KEGG pathway enrichment analysis of differentially expressed genes. Pathways relevant to infection are highlighted with blue boxes. Key developmental pathways are highlighted with red boxes. C. Differentially expressed genes within the Hippo pathway. D. Gene expression profiles (transcripts per million [TPM] values) of all four TEAD family transcription factors with and without HCMV infection. E. Western blots of established TEAD1 targets THBS1 and CCN1 using whole cell lysates of uninfected and HCMV infected cells. GAPDH is used as control.

HCMV impairs TEAD1 activity through multiple distinct mechanisms
A. Representative Western blots for HCMV proteins IE1/2, YAP1, pYAP1, TEAD1 and the H3K27ac mark from whole cell lysates of uninfected and HCMV infected cells. GAPDH was used as a loading control. Additional Western blots (biological triplicates) are provided in Supplemental Figure S3, along with quantifications and p-values. B. Western blots using cytosolic and nuclear fractions obtained from uninfected and HCMV infected cells indicating the localization of YAP1 and pYAP1. GAPDH and Histone H3 were used as controls for cytoplasmic and nuclear fractions, respectively. C. Agarose gel image of RT-PCR products of TEAD1 exon-6 splicing events. The full length TEAD1 targeted region is 91 bp long; it is 79 bp without exon-6. Jurkat cells, which have approximately equal expression of TEAD1 with and without exon 6 ³⁹, were used as a control. D. Model depicting four distinct mechanisms by which HCMV reduces the activity of the TEAD1 transcription factor. E. Enrichment of phenotype-associated genetic variants at HCMV-altered TEAD1 binding events. Enrichment values calculated by the RELI algorithm are presented for TEAD1 binding events for the lobe attachment (A) and ocular cup (B) phenotypes. Enrichment for TEAD1 binding loss with HCMV infection is statistically significant for each assessment except for the group for lobe attachment (shown in purple). For each bar, dots represent RELI enrichment results from different ancestral groups as defined in the original GWAS studies.

A. Principal component analysis of ATAC-seq replicates of uninfected and HCMV infected fibroblasts and retinal epithelial cells. B. Heat map of gene expression profiles between replicates of uninfected and HCMV infected fibroblasts.

Heatmap of ChIP-seq peaks for TEAD1, CTCF, and H3K27ac between replicates in uninfected and HCMV infected fibroblasts.

Western blots of uninfected and HCMV infected fibroblasts in biological triplicates for TEAD1, H3K27ac, YAP1, pYAP1, and HCMV Immediate Early proteins IE1/2.
For the HCMV IE1/2 western blot, β-Actin (red color) was used as a loading control. For TEAD1, H3K27ac, YAP1, and pYAP1, GAPDH was used as a control (red color). Quantification of Western blot signal intensities for each protein are shown next to the blot images. Signal quantification was performed using Emperia Studio software and p-values were calculated using two-tailed t-test on GraphPad Prism software. NS, not significant.

A. Visualization of the two significant TEAD1 splicing changes in HCMV infected and uninfected cells. The splicing model of these events shows the percent spliced in (PSI) values, reported as percentages, for uninfected and HMCV infected cells. The red splice junction indicates the splicing outcome that is upregulated in HCMV infection. B. Sashimi plots depicting TEAD1 splice junctions and read coverage from exon 4 to exon 7, with the significant splicing changes of exon 6 (E6) skipping and partial inclusion of intron 5 (I5) highlighted in grey C. DNA sequence flanking TEAD1 exon 6 (green). RT-PCR primers are depicted in red.

Comparison of HiChIP signal between replicates of uninfected and HCMV infected cells. Axes represent genomic bin location.

Sign up for email alerts