HSV-1 exploits host heterochromatin for nuclear egress

Abstract

Herpes simplex virus (HSV-1) progeny form in the nucleus and exit to successfully infect other cells. Newly formed capsids navigate complex chromatin architecture to reach the inner nuclear membrane (INM) and egress. Here, we demonstrate by transmission electron microscopy (TEM) that HSV-1 capsids traverse heterochromatin associated with trimethylation on histone H3 lysine 27 (H3K27me3) and the histone variant macroH2A1. Through chromatin profiling during infection, we revealed global redistribution of these marks whereby massive host genomic regions bound by macroH2A1 and H3K27me3 correlate with decreased host transcription in active compartments. We found that the loss of these markers resulted in significantly lower viral titers but did not impact viral genome or protein accumulation. Strikingly, we discovered that loss of macroH2A1 or H3K27me3 resulted in nuclear trapping of capsids. Finally, by live-capsid tracking, we quantified this decreased capsid movement. Thus, our work demonstrates that HSV-1 takes advantage of the dynamic nature of host heterochromatin formation during infection for efficient nuclear egress.

Figure 1.

HSV-1 capsids navigate through regions of less dense chromatin to reach the inner nuclear membrane in HFF cells. (a) TEM images of representative uninfected nuclei in WT HFF-Ts. Regions outside of the nucleus are colorized yellow. Dark regions represent high-density chromatin (arrowhead). Scale bars as indicated. (b) TEM images of representative WT nuclei at 18 hpi with HSV-1. Inset shows an enlarged view of the respective boxed area. The arrowhead indicates high-density chromatin, arrows indicate HSV-1 capsids. (c) Representative Western blots with proteins as indicated showing macroH2A1 KO and H3K27me3 depletion (Dep.). (d) TEM images of representative uninfected nuclei in macroH2A1 knockout HFF-T cells. Regions outside of the nucleus are colorized green. (e) TEM images of representative uninfected nuclei in H3K27me3 depleted conditions. Regions outside of the nucleus are colorized blue. (f) Quantification of peripheral heterochromatin width in nuclei from a and d. Width was measured in nm from the nuclear periphery via binary thresholding from intensity profiles sampled every 10 pixels. Mean width was plotted for each nucleus. P = 0.0275 (n = 16 WT, n = 22 macroH2A1 KO) by unpaired t test. For the box plot, the box marks upper and lower quartiles, center line marks median, and error bars denote minimum and maximum values for the population. (g) Quantification as in f in nuclei from a and e. P = 0.0144 (n = 22 WT, n = 20 H3K27me3 depleted) by unpaired t test. For the box plot, the box marks upper and lower quartiles, center line marks median, and error bars denote minimum and maximum values for the population. Source data are available for this figure: SourceData F1.

Figure S1.

Quantification of macroH2A1 and H3K27me3 enrichment on host genomes during HSV-1 infection. (a) Relative peptide abundance of macroH2A1, H3K27me3, and H3K9me3 over the course of HSV-1 infection in primary HFFs from three biological replicates normalized to mock peptide counts (original data from Kulej et al. [2017]). (b) Representative Western blots of total viral and host proteins as indicated at mock (M) 4, 8, and 12 hpi with HSV-1 in WT and macroH2A1 KO HFF-T cells. H3 and actin are shown as loading controls. (c) Mean relative intensity of total macroH2A1, H3K27me3, ICP0, and VP16 as indicated. Error bars show ± SD of three replicates of Western blots as in b. (d) Spike-in normalized CUT&Tag read density of macroH2A1 on WT and macroH2A1 KO HFF-T host genomes during HSV-1 infection. M indicates mock-infected cells; 4, 8, and 12 indicate 4, 8, and 12 hpi. For box plots, the lower and upper hinges correspond to the first and third quartiles (the 25th and 75th percentiles). The upper whisker extends from the hinge to the largest value no further than 1.5 * IQR from the hinge (where IQR is the interquartile range or distance between the first and third quartiles). The lower whisker extends from the hinge to the smallest value at most 1.5 * IQR of the hinge. Data beyond the end of the whiskers are called “outlying” points and are plotted individually. (e) Spike-in normalized CUT&Tag read density of H3K27me3 on WT and macroH2A1 KO host genomes during HSV-1 infection presented as in d. (f) Total number of macroH2A1 and H3K27me3 domains over 1 kb as measured by CUT&Tag of WT and macroH2A1 KO host genomes during HSV-1 infection. (g) Correlation matrix of host genome enrichment from all datasets at domains as defined in Fig. 2. Pearson correlation coefficient is plotted both as a heatmap and the values are printed inside each cell. If the P value of a correlation coefficient is not significant, there is an asterisk over the number in that cell. Source data are available for this figure: SourceData FS1.

Figure 2.

MacroH2A1 and H3K27me3 bind broad chromatin regions on the host genome that are redistributed over the course of HSV-1 infection. (a) Representative genome browser snapshots of spike-in normalized CUT&Tag enrichment of macroH2A1 and H3K27me3 showing increases at 8 hpi of HSV-1 infection as measured by CUT&Tag in WT or macroH2A1 KO HFF-T cells as indicated. The region shown is found on chromosome 1. (b) Changes in log₂ enrichment of spike-in normalized CUT&Tag of macroH2A1 and H3K27me3 over IgG compared with mock treatment are shown as a heatmap. Each line in the heatmap represents a domain of macroH2A1. (c) Quantification of heat maps from b showing macroH2A1 enrichment in WT HFF-T cells across each cluster. (d) Quantification of heat maps from b showing H3K27me3 enrichment in clusters defined by macroH2A1 in HFF-T cells. (e) Changes in log₂ enrichment as in b of H3K27me3 over IgG in macroH2A1 KO HFF-T cells. (f) Quantification as in c in macroH2A1 KO HFF-T cells.

Figure S2.

Quantification of macroH2A1 and H3K27me3 enrichment on viral genomes during HSV-1 infection. (a) Spike-in normalized CUT&Tag read density of macroH2A1 on HSV-1 genomes during infection of WT and macroH2A1 KO HFF-T cells. The lower and upper hinges correspond to the first and third quartiles (the 25th and 75th percentiles). The upper whisker extends from the hinge to the largest value no further than 1.5 * IQR from the hinge (where IQR is the interquartile range or distance between the first and third quartiles). The lower whisker extends from the hinge to the smallest value at most 1.5 * IQR of the hinge. Data beyond the end of the whiskers are called “outlying” points and are plotted individually. (b) Spike-in normalized CUT&Tag read density of H3K27me3 on HSV-1 genomes during infection of WT and macroH2A1 KO HFF-T cells. Data graphed as in a. (c) Genome browser snapshots of spike-in normalized CUT&Tag enrichment of macroH2A1 on viral genomes in WT or macroH2A1 KO HFF-T cells as indicated. The full viral genome is shown. (d) Genome browser snapshots of spike-in normalized CUT&Tag enrichment of H3K27me3 on viral genomes in WT or macroH2A1 KO HFF-T cells as indicated. The full viral genome is shown. The same IgG control was used to generate IgG WT and KO control graphs for S2c-d. (e) Correlation matrix of CUT&Tag signals aligned to the viral genome from all datasets. Pearson correlation coefficient is plotted both as a heatmap and printed inside each cell. If the P value of a correlation coefficient is not significant, there is an asterisk over the number in that cell. (f) Top: Representative immunofluorescence of total macroH2A1 (yellow), viral DNA binding protein ICP8 (red), and DAPI (blue) at 10 hpi in WT and macroH2A1 KO HFF-T cells infected with HSV-1. Scale bar is 10 µm. Bottom: Histogram of macroH2A1 (green) and ICP8 (red) intensity at dotted transect line. (g) Top: Representative Immunofluorescence of H3K27me3 (yellow), ICP8 (red), and DAPI (blue) at 10 hpi in HSV-1 infected WT and H3K27me3 depleted HFF-T cells. Bottom: Histogram of H3K27me3 (green) and ICP8 (red) intensity at dotted transect line.

Figure S3.

MacroH2A1 and H3K27me3 presence on host genomes correlates with decreased transcription. (a) Box plots of total host RNA from RNA-seq in HFF-T WT for genes overlapping clusters from Fig. 2. Kolmogrov–Smirnov test (in R) was performed comparing logCPM values of genes at each time point with logCPM values of genes from the mock dataset for each cluster. The P values were corrected for multiple testing, and the time points and clusters with corrected P value <0.05 are shaded. (b) Same as a for macroH2A1 KO HFF-T cells. (c) Host differential RNA levels between WT and macroH2A1 KO HFF-T cells from RNA-seq at 4, 8, and 12 hpi by cluster as indicated. N = 3 biological replicates. Volcano plot for host genes in each cluster is plotted as in Fig. 3 a with log(Fold Change) on x-axis and −1 × log(False Discovery Rate) plotted on the y-axis. Points without significant change in expression are plotted in black, significant reduction in expression are plotted in blue, and significant increase in expression are plotted in red. (d) Enrichment gene sets from gene ontology (GO) analysis of genes belonging to each cluster as indicated. P value <0.001 and FDR was <0.05 for all shown GO clusters.

Figure S4.

Clusters with more macroH2A1 and decreasing transcription during HSV-1 infection also show less transcription after salt stress or heat shock and correlate with active compartments. (a) 4sU-RNA counts for genes from Hennig et al. (2018) overlapping with clusters from Fig. 2 shown as box plots. Kolmogrov–Smirnov test (in R) was performed comparing logCPM values of genes at each time point with logCPM values of genes from the mock dataset for each cluster. The P values were corrected for multiple testing, and the time points and clusters with corrected P value <0.05 are shaded. For all box plots in this figure, lower and upper hinges correspond to the first and third quartiles (the 25th and 75th percentiles). The upper whisker extends from the hinge to the largest value no further than 1.5 * IQR from the hinge (where IQR is the interquartile range, or distance between the first and third quartiles). The lower whisker extends from the hinge to the smallest value at most 1.5 * IQR of the hinge. Data beyond the end of the whiskers are called “outlying” points and are plotted individually. (b) Replicate RPKM values from published 4sU-RNA-seq data (Hennig et al., 2018; GSE100469) from HFF treated with 80 mM KCl (increase in final concentration) for 1 and 2 h were converted to TPM, averaged, and intersected with gene lists for each of the six clusters. The distribution of the log₂(TPM+1) for genes in each cluster is shown as boxplots. (c) Same as b for heat stress (44°C) also from GSE100469. (d) Box plots of Hi-C eigenvector scores of regions overlapping with clusters from Fig. 2. The Hi-C compartment eigenvector scores were obtained from 4DN project website. Wilcoxon signed rank test with alternate hypothesis that the true location is not equal to 0 was performed on the distribution of eigenvector scores for each cluster. The P values were corrected for multiple testing, and the clusters with corrected P value <0.05 are shaded. Bonferroni correction was used for P value adjustments. (e) Cell counts for experiments as described in Fig. 3, b–h. N = 3 biological replicates. No significance by paired t-test. Error bars represent the SEM of three biological replicates. (f) Cell viability of HFF-T in DMSO or 10 µM Tazemetostat for 4 d posttreatment. N = 3 biological replicates. No significance by paired t test. Error bars represent the SEM of three biological replicates. (g) Representative Western blots of HSV-1 protein UL34 in WT and macroH2A1 KO HFF-T cells during HSV-1 infection at mock-infected (M) or 4, 8, and 12 hpi. Actin is shown as the loading control. (h) Representative Western blots of HSV-1 protein UL34 in WT and H3K27me3 depleted cells during HSV-1 infection at mock-infected (M) or 4, 8, and 12 hpi. Ponceau stain is shown as the loading control. Source data are available for this figure: SourceData FS4.

Figure 3.

HSV-1 requires heterochromatin marks macroH2A1 and H3K27me3 for progeny production but not replication or protein production. (a) Volcano plots comparing differential viral RNA levels between WT and macroH2A1 KO HFF-T cells from RNA-seq at 4, 8, and 12 hpi. N = 3 biological replicates. Genes are plotted with log(Fold Change) on x-axis and −1 × log(False Discovery Rate) plotted on the y-axis. Points without significant change in expression are plotted in black, significant reduction in expression are plotted in blue, and significant increase in expression are plotted in red. There are no significantly changing genes. (b) Representative Western blots of proteins were shown as indicated upon H3K27me3 depletion or DMSO control–treated HFF-T cells during HSV-1 infection at mock-infected (M) or 4, 8, and 12 hpi. Actin and H3 are shown as loading controls. (c) Representative Western blots as in b for macroH2A1 KO HFF-T cells. (d) Mean relative intensity of H3K27me3, ICP0, VP16, and gH normalized to H3 quantified from Western blots as in b and c as indicated from WT or macroH2A1 KO. Error bars represent ± SD of three biological replicates. (e) Droplet digital (ddPCR) quantification of HSV-1 genomes extracted from infected cells as indicated for each time point. Error bars represent the SEM of three biological replicates. No significance by Dunnett's multiple comparisons test. (f) ddPCR quantification of HSV-1 genomes released from cells treated as indicated and isolated from supernatants (sups). Error bars represent the SEM of three biological replicates, **P < 0.01, ***P < 0.001 by Dunnett’s multiple comparisons test. (g) Infectious progeny produced from HSV-1 infected cells treated as indicated and quantified by plaque assay. Viral yield is indicated as the percent yield compared to WT at each indicated time point. Error bars represent the SEM of three biological replicates, **P < 0.01, ***P < 0.001 by Dunnett's multiple comparisons test. (h) Genome to PFU ratio 8 and 12 hpi in indicated conditions as measured by paired plaque assays and ddPCR of genomes isolated from supernatants as in f. Error bars represent the SEM of three biological replicates, no significance by Dunnett’s multiple comparison test. Source data are available for this figure: SourceData F3.

Figure S5.

MacroH2A1 is also required for efficient viral egress, but not protein production, in RPE cells. (a) Western blot of proteins as indicated in mock(M) or 4, 8, and 12 hpi in WT and macroH2A1 KO RPE cells. H3 is shown as a loading control. (b) Mean relative intensity of ICP0 over H3 at indicated time points during HSV-1 infection in WT or macroH2A1 KO RPEs as in a. Error bars represent ± SD from three biological replicates. (c) Same as b for VP16. (d) Same as b for gH. (e) Same as b for H3K27me3. (f) Same as b for macroH2A1. (g) Infectious progeny produced by WT or macroH2A1 knockout RPE cells quantified by plaque assay on cell-free supernatant. Viral yield calculated as in Fig. 3 g for conditions as indicated, **P < 0.01 by unpaired t test. Error bars represent the SEM of three biological replicates. (h) Average MSD plots ± SEM of nuclear capsid tracks in HSV-1 mCherry-VP26 infected WT or macroH2A1 KO RPE cells at 8 hpi. Non-linear fits of mean MSD plots and exponential plateau as dotted lines (YM) are indicated. (i) Representative live-cell images of HSV-1 mCherry-VP26 infected WT or macroH2A1 KO RPE nuclei at 6 hpi and corresponding tracks from single-particle tracking used for MSD analysis in h. Scale bar is 5 µm. Source data are available for this figure: SourceData FS5.

Figure 4.

Clinical HSV-1 isolates also require heterochromatin mark macroH2A1 for progeny production but not replication or protein production. (a) Western blots of proteins as indicated for mock (M) 4, 8, and 12 hpi in WT and macroH2A1 KO cells infected with a low-shedding HSV-1 clinical isolate. (b) Western blot as in a for a high-shedding HSV-1 clinical isolate. (c) Mean relative intensity of ICP0, VP16, and gH from low-shed (top) or high-shed (bottom) clinical isolate Western blots. Error bars represent ± SD of three biological replicates, no significance by unpaired t test. (d) ddPCR quantification of HSV-1 genomes extracted from infected cells as indicated for each time point in WT and macroH2A1 KO cells infected with HSV-1 low- and high-shedding clinical isolates. Error bars represent the SEM of three biological replicates. No significance by Tukey’s multiple comparisons test. (e) ddPCR quantification of HSV-1 genomes released from cells infected with clinical isolates of HSV-1 as indicated and isolated from supernatants (sups). Error bars represent the SEM of three biological replicates. No significance by Tukey’s multiple comparisons test. (f) Infectious progeny of cells as indicated infected with clinical HSV-1 isolates quantified by plaque assay. Viral yield is indicated as the percent yield compared to WT, error bars represent the SEM of three biological replicates, *P < 0.05, **P < 0.01 by Tukey’s multiple comparisons test. Source data are available for this figure: SourceData F4.

Figure 5.

HSV-1 requires macroH2A1- and H3K27me3-dependent heterochromatin for movement through host chromatin to access the inner nuclear membrane (INM). (a) TEM images of representative macroH2A1 KO HFF-T nuclei infected with HSV-1 at 18 hpi. Insets show enlarged views of the respective box. Arrows indicate HSV-1 capsids. Scale bars as indicated. (b) Quantification of capsids within nuclei compared to Fig. 1 b. Number of capsids was normalized according to nucleus area in µm², P = 0.0036 (n = 40 WT, n = 55 mH2A1 KO) by unpaired t test. For the box plot, the box marks upper and lower quartiles, center line marks median, and error bars denote minimum and maximum values for the population. (c) TEM images of representative H3K27me3 depleted nuclei infected with HSV-1 presented as in a. (d) Quantification of capsids accumulating at the INM. The number of capsids within 200 nm of the membrane scored per chain as capsids within 300 nm of another capsid, P = 0.0008 by Mann-Whitney test (n = 61 for DMSO, n = 51 for H3K27me3 depleted). Error bars represent ± SEM of the population. (e) TEM images of representative A (empty), B (intermediate; scaffolding proteins, but no genome), and C (full) HSV-1 capsids. Scale bars as noted. (f) Quantification of capsid type within nuclei from a in WT and macroH2A1 KO cells. Values for each capsid type are shown as a proportion of total capsids. No significance by Chi-square test, n = 120 capsids per condition. (g) Quantification of capsid type as in f in nuclei from c in WT and H3K27me3 depleted cells. No significance by Chi-square test, n = 120 capsids per condition. (h) Average mean squared displacement (MSD) plots ± SEM of nuclear capsid tracks in HSV-1 mCherry-VP26 infected WT or macroH2A1 KO HFF-T cells at 6 hpi. Non-linear fits of MSD plots and exponential plateau as dotted lines (YM) are indicated. MSD plots represent 1,148 tracks in six nuclei for WT and 993 tracks in six nuclei for macroH2A1. Plateaus are at YM_WT = 0.66 µm² and YM_KO = 0.44 µm². (i) Representative live-cell images of HSV-1 mCherry-VP26 infected WT or macroH2A1 KO HFF-T nuclei at 6 hpi and corresponding tracks from single-particle tracking used for MSD analysis in h. Scale bar is 5 µm.

Figure 6.

Model for heterochromatin support of HSV-1 nuclear egress. Left: Model of previously described chromatin movement during HSV-1 infection. Right: Proposed model of effects of heterochromatin disruption in HSV-1 infection.