A viral histone-like protein exploits antagonism between linker histones and HMGB proteins to obstruct the cell cycle

Summary

Virus infection necessarily requires redirecting cellular resources towards viral progeny production. Adenovirus encodes the histone-like protein VII which causes catastrophic global reorganization of host chromatin to promote virus infection. Protein VII recruits the family of high mobility group box (HMGB) proteins to chromatin along with the histone chaperone SET. As a consequence of this recruitment, we find that protein VII causes chromatin-depletion of several linker histone H1 isoforms. The relationship between linker histone H1 and the functionally opposite HMGB proteins is critical for higher order chromatin structure. However, the physiological consequences of perturbing this relationship are largely unknown. Here, we employ complementary systems in Saccharomyces cerevisiae and human cells to demonstrate that adenovirus protein VII disrupts the H1-HMGB balance to obstruct the cell cycle. We find that protein VII causes an accumulation of G2/M cells both in yeast and human systems, underscoring the high conservation of this chromatin vulnerability. In contrast, adenovirus E1A and E1B proteins are well established to override cell cycle regulation and promote transformation of human cells. Strikingly, we find that protein VII obstructs the cell cycle even in the presence of E1A and E1B. We further show that in a protein VII-deleted infection, several cell cycle markers are regulated differently compared to wild type infection, supporting our model that protein VII plays an integral role in hijacking cell cycle regulation during infection. Together, our results demonstrate that protein VII targets H1-HMGB1 antagonism to obstruct cell cycle progression, revealing an unexpected chromatin vulnerability exploited for viral benefit.

eTOC Blurb:

Lynch et al. define a novel mechanism of chromatin disruption caused by adenovirus protein VII. Using complementary systems in budding yeast and human cells, they show that protein VII disrupts the balance between linker histones and high mobility group box proteins. This blocks cell cycle progression which ultimately promotes infection.

Graphical Abstract

Introduction

Viral takeover of cellular processes is essential for the success of these intracellular pathogens. Controlling host chromatin is an integral aspect of viral infection that includes re-directing host resources towards viral replication. Several examples have recently come to light that highlight viral alteration to host gene expression through modifications of core histone tails^1–5. Host chromatin may also be manipulated through displacement of linker histone H1. Linker histones are critical for genome compaction^6,7 and compete for binding sites with high mobility group (HMG) proteins⁸, which in contrast promote decompaction of chromatin and increase DNA accessibility^9,10. This interplay between linker histones and HMG proteins is a vulnerability that may be a target of viral manipulation.

The adenovirus genome is packaged with a small basic protein known as protein VII^11–13. This 20 kDa core protein recruits the histone chaperone protein SET^14,15 onto viral genomes as they enter the nucleus. SET is thought to populate the viral genomes with histones to promote viral gene expression^16,17. The first viral gene products, E1A and E1B, are expressed leading to viral gene activation, global changes in host histone modifications, and bypass of cell cycle checkpoints^18,19. E1A’s effects are a model for cellular transformation and epigenetic reprogramming by viruses^20,21, although it remains unclear why adenovirus infection does not transform humans^19,22. Protein VII is transcribed late during infection¹⁸, has a structural role inside virion cores, and localizes to host chromatin where it disrupts normal nuclear processes^23,24. Protein VII is post-translationally modified which impacts its localization to chromatin^23,24. Recombinant protein VII binds purified nucleosomes directly, protecting approximately 165 bp of DNA, suggesting that it has similar binding sites as linker histones²³.

Several host proteins are significantly enriched in chromatin in the presence of protein VII, including SET and high mobility group box (HMGB) proteins^23,25. Without protein VII, SET and HMGB proteins are transiently bound to chromatin^9,26. SET acts as a linker histone chaperone^27,28 while HMGB1 is an antagonist of histone H1¹⁰. Here, we use complementary systems in Saccharomyces cerevisiae and human cells to show that adenovirus protein VII requires HMGB1 and SET to disrupt chromatin and cause growth defects. These deficiencies in growth are due to deviations in cell cycle progression. We show that protein VII can inhibit cell cycle progression even in human cells that have been transformed in part by E1A-directed cell cycle bypass. Importantly, we show that during infection without protein VII, multiple cell cycle markers are altered compared to wild-type infection. Therefore, we propose that during infection protein VII obstructs mitosis to ensure cellular resources are directed towards viral progeny production rather than cell division. This study is the first to show viral exploitation of the H1-HMGB1 interaction and subsequent delay of the cell cycle for viral benefit.

Results

Human adenovirus protein VII causes growth defects in budding yeast

In addition to enrichment of SET and HMGBs in chromatin upon protein VII expression in human lung epithelial cells, we discovered that linker histone H1 isoforms were significantly depleted from chromatin (Figure S1A). Since H1 has recently been shown to be critical in maintaining chromatin architecture^29,30, we hypothesized that protein VII co-opts SET and HMGB1 to displace histone H1 from host chromatin (Figure S1B). We tested our model with genetic studies in Saccharomyces cerevisiae, which faithfully represents many aspects of mammalian chromatin with less genetic redundancy than mammalian cells³¹.

We expressed protein VII via an inducible promoter in the wild type (WT) lab strain W303 and assayed growth on solid media (Figure 1A and S1C). Strikingly, we found that protein VII expression resulted in a ~40% growth deficit compared to GFP (Figure 1B). Protein VII is post-translationally modified at five amino acid residues which are critical for its chromatin localization and subsequent phenotypes in human cells^23,24. Mutation of these sites (VII-ΔPTM) rescued the protein VII growth defect (Figure 1A–B). Liquid culture growth analysis also showed a significant growth reduction upon protein VII expression, but not GFP or VII-ΔPTM (Figure 1C–D). The protein VII growth defect was not unique to W303 as we observed similar results in a genetically diverged strain derived from S288C, FY602 (Figure 1E–H). Protein VII was expressed in yeast at approximately the same levels as ectopic expression human cell lines (Figure S1G), which are sufficient to bind and reorganize host chromatin²³. We found by immunofluorescence microscopy that protein VII is localized to the yeast nucleus (Figure S1I), likely due to its nuclear localization signal (NLS)³². Chromatin immunoprecipitation with sequencing (ChIP-seq) using a protein VII antibody³³ showed that it binds throughout the yeast genome (Figure S1J and S2). Comparing protein VII localization to either H3 or input control identified few peaks of protein VII enrichment, suggesting that protein VII coats yeast chromatin as uniformly as histone H3. Thus, we conclude that protein VII directly binds chromatin throughout the genome. These data demonstrate that protein VII effects on chromatin are direct and robust in yeast, indicating that eukaryotic chromatin is widely vulnerable to protein VII.

Figure 1. Adenovirus protein VII inhibits growth in budding yeast.

(A) Serial dilution growth assay of induced (+ expression) or repressed (No expression) WT W303 cells with indicated protein. (B) Growth in (A) as measured by spot #3 (red) growth intensity compared to spot #9 (blue) (see STAR Methods). ****p<0.0001, one-way ANOVA with multiple comparisons, n=7. (C) Exponential growth of WT W303 with the indicated plasmids, *p=0.0157, n=4. (D) Western blot analysis of WT W303 strains with and without HA-tagged VII or GFP expression during exponential growth. Phosphoglycerate kinase (Pgk1) loading control. (E-F) Same as (A-B) in WT FY602. ****p<0.0001, one-way ANOVA with multiple comparisons, n=18. (G) Same as (C) in WT FY602, **p=0.0098, n=3. (H) Same as (D) in WT FY602. All error bars represent SD. See Figure S1, S2, and Table S1.

We determined whether other highly charged DNA-binding proteins elicit similar effects by testing adenovirus core protein V³⁴ and dinoflagellate viral nucleoprotein (DVNP), a viral-derived protein known to displace histones in yeast³⁵ (Figure S1C–F). Protein V did not significantly impact yeast growth but DVNP expression ablated growth, as previously reported³⁵. Given the range of impact that highly charged DNA-binding proteins had on growth, we next tested for protein VII’s specific genetic interactions in deletion mutants.

The genetic interactions of protein VII with the yeast homologs of H1, HMGB1, and SET support a linker histone displacement model

Unlike mammals, which have multiple linker H1 histones and HMGB proteins^7,9, yeast have primary homologs of H1, HMGB1 and SET, which are Hho1, Hmo1, and Nap1 respectively^36–38. Therefore, we hypothesized that Hmo1 and Nap1 facilitate protein VII dysregulation of chromatin in yeast (Figure 2A). Consistent with this prediction, we found that HMO1 deletion during protein VII expression rescued growth compared to WT W303 (Figure 2B–C). Likewise, loss of NAP1 led to slightly better relative growth than WT, though quantification did not reach significance. We also assayed growth of the alternate FY602 strain with the same set of gene deletions and found that deletion of either HMO1 or NAP1 significantly rescued growth (Figure 2E–F). Minor disparities in the degree of rescue between W303 and FY602 strains are likely due to genotypic variation³⁹. We also hypothesized that Hmo1 and Nap1 function cooperatively to facilitate protein VII’s disruption of chromatin and consequently assayed growth in a hmo1Δnap1Δ double deletion strain. Deletion of both genes rescued the protein VII growth phenotype to the same degree as a single gene deletion (Figure 2B and E). For both W303 and FY602, the hmo1Δnap1Δ deletion strain grew better than WT, but relative growth did not exceed that of the individual gene deletion strains (Figure 2C and F). Protein VII expression levels were largely consistent across mutants suggesting that protein stability alone does not explain the growth rescue (Figure 2D and G). Therefore, we conclude that Nap1 and Hmo1 work cooperatively with protein VII to cause robust growth defects in yeast.

Figure 2. The loss of binding partners Hmo1 or Nap1 rescues protein VII growth defects.

(A) Model for protein VII chromatin disruption in budding yeast. Protein VII recruits the homologs of HMGB1 and SET, Hmo1 and Nap1 respectively, to facilitate protein VII replacement of the linker histone Hho1. (B) Serial dilution assay for growth defects of W303 cells with or without protein VII expression in the indicated deletion strains. (C) Quantification of growth assays shown in (B) as described in Figure 1, ***p=0.0009 and **p=0.0090, n≥3. (D) Western blot of W303 gene deletion strains shown in (C) expressing protein VII-HA. Pgk1 loading control. (E) Same as (B) for FY602 deletion strains. (F) Same as (C) in FY602, **p=0.0035, ****p<0.0001, n≥4. (G) Same as (D) in FY602. Growth quantified on day 4(‡) for hmo1Δ and hmo1Δnap1Δ strains and day 3 for all others. All error bars show SD. See Figure S3.

Since protein VII expression reduces H1 occupancy in human cells (Figure S1A) and binds linker DNA in vitro²³, we predicted that protein VII and H1 compete for linker DNA binding sites. In both W303 and FY602, loss of HHO1 reduced growth compared to WT during protein VII induction although the reduction was not statistically significant (Figure 2B–C and E–F). Notably, GFP expression did not significantly alter deletion strains’ growth (Figure S3A–B). To test whether these mutations rescued protein VII-induced defects specifically and not those caused by any highly charged chromatin-binding protein that interrupts growth, we assayed DVNP growth defects in the same deletion strains (Figure S3C–D). There was no consistent rescue of the DVNP phenotype across the mutant strains which suggests that DVNP disruption does not follow the same mechanism as protein VII. Therefore, we conclude that the observed genetic interactions are specific to protein VII. Together, these results support our model in which protein VII interacts with Hmo1 and Nap1 to displace Hho1 and disrupt chromatin resulting in growth defects (Figure 2A).

Human HMGB1, SET, and H1.5 can replace their yeast homologs to facilitate protein VII’s disruption of chromatin in yeast

To determine whether human protein VII binding partners could replace their yeast homologs in our system, we co-expressed protein VII and the appropriate human homolog from a bidirectional promoter in the corresponding yeast mutant strain and examined growth in both W303 and FY602 (Figure 3). HMO1 deletion rescued growth while co-expression of HMGB1 with protein VII in the hmo1Δ strain restored reduced growth (Figure 3A–C and G–I). We performed the same set of assays using human SET to replace yeast Nap1 (Figure 3D–F and J–L). NAP1 deletion rescued the protein VII growth defect while human SET expression in the nap1Δ strain restored protein VII growth deficiency. These effects were not observed when GFP was expressed under the same conditions (Figure S3E–H). We next introduced human H1.5 expression and found that growth was significantly rescued compared to WT or hho1Δ during protein VII expression (Figure 3M–O). This is notable since exogenous human H1 expression slows yeast growth⁴⁰, which we recapitulated in the GFP control conditions (Figure S3I). The redundant paralogs Nhp6a and Nhp6b have been proposed as functional homologs of HMGB1 in yeast, despite containing a single HMG domain⁴¹ compared to double HMG domains in Hmo1 and human HMGB1⁴². We consequently expressed protein VII in a nhp6aΔnhp6bΔ strain alone and with co-expression of human HMGB1 (Figure 3P–R). NHP6A and NHP6B deletion did not rescue the protein VII growth defect but expressing HMGB1 significantly exacerbated the growth phenotype. This exacerbated growth deficit was not observed with GFP (Figure S3J). Thus, we conclude that yeast Hmo1 likely has the same role as HMGB1 in complex with protein VII rather than Nhp6a and Nhp6b. Together these data support a model in which protein VII interacts with HMGB1 and SET to displace H1 and disrupt chromatin.

Figure 3. Expressing homologous human factors HMGB1 and SET1 exacerbates the protein VII growth defects in yeast while H1.5 expression restores growth.

(A) Serial dilution assay for protein VII growth defects in WT W303 and W303 hmo1Δ strain with and without co-expression of human FLAG-HMGB1 from an inducible bidirectional promoter. Expression constructs and genotypes are indicated at left. (B) Quantification of (A) as described in Figure 1 and STAR Methods, ***p=0.0001, **p=0.0026. (C) Western blot showing tagged protein co-expression illustrated in (A). (D) Same as (A) in W303 nap1Δ with human FLAG-SET1. (E) Quantification as in (B) for series in (D), **p=0.0058. (F) Western blot of tagged protein co-expression shown in (D). (G-I) Same as (A-C) in FY602, *p=0.0180. (J-L) Same as (D-F) in FY602. ***p=0.0003, nap1Δ −/+ SET, *p=0.0240, WT vs nap1Δ + SET *p=0.0133. (M-O) Same as (A-C but with W303 hho1Δ and FLAG-H1.5 expression, **p=0.0078, ***p=0.0002. (P-R) Same as (A-C) but in W303 nhp6aΔnhpbΔ with FLAG-HMGB1 expression, ****p<0.0001. Cells without protein expression are shown after 3 days growth and those with protein expression after 6 days. P-values determined by one-way ANOVA with multiple comparisons. n≥3. All error bars represent SD. See Figure S3.

Protein VII disrupts the cell cycle in yeast and human cells

To define the cause of slowed growth induced by protein VII, we analyzed yeast cell cycle progression with budding analysis. Since the yeast budding cycle is correlated to cell cycle progression and can be halted by cell cycle checkpoint activation, a change in the distribution of budding morphologies during log phase growth can indicate cells’ failure to progress from one phase of the cell cycle to the next⁴³. We found significant over-representation of cells with large buds when protein VII was expressed compared to GFP (Figure 4A). Over time, the proportion of large-budded cells was significantly greater for protein VII expressing cells than the GFP control (Figure 4B and S4A–C). Cells with large buds are likely unable to transition from S phase to G2 or from G2 to M at the same rate as normal. Thus, we conclude that protein VII expression disrupts the budding cycle consistent with cell cycle defects.

Figure 4. Protein VII dysregulates cell cycle progression in budding yeast and human cells.

(A) Budding analysis of WT W303 with GFP or protein VII during mid-log phase growth after 8 hours galactose induction. Bud size roughly corresponds to cell cycle phase as shown in schematic. Numbers indicate the percentage of the population with each bud type ± SD. The difference in large-budded cells upon expression of protein VII compared to the GFP control reached significance, p=0.0012, n=3 (multiple unpaired t-tests with two-stage step-up method of Benjamini, Krieger and Yekutieli to control FDR). (B) Percentage of large-budded cells during protein VII (pink) or GFP (green) expression in WT W303 population over time. X-axis indicates time from induction of protein expression. Unpaired t-test p-values for each comparison from left to right: *0.010250, **0.003903, **0.001382, and **0.001167, n=3. (C) Diploid RPE-1 cell proliferation for 48 hours after treatment with indicated recombinant adenovirus or CDK1 inhibitor. At 48 hours, there were significantly fewer VII-GFP expressing cells than no treatment (p<0.0001) or GFP control cells (p<0.0001), n=4 (one-way ANOVA with Tukey’s multiple comparisons test). (D) Cell cycle analysis of RPE-1 cells shown in (C) by flow cytometry analysis of DNA content. (E) Western blot analysis of cells shown in (C). NT - no treatment control; PT - pre-treatment; hpt - hours post-treatment. Histone H3 and vinculin loading controls. (F) HEK 293T proliferation with and without protein VII-HA expression via an inducible promoter. At 96 hours, there were significantly fewer VII-HA expressing cells than GFP control cells (p=0.0016) or untreated cells (p=0.0009), n=3 (one-way ANOVA with multiple comparisons). (G) Same as (D) for HEK 293T cells in (F). (H) Western blot analysis of HEK 293T cells shown in (F). E1A isoforms resolve as multiple bands. All error bars represent SD. See Figure S4.

To assess the impact of protein VII on human cells, we assayed proliferation and cell cycle progression in human diploid retinal pigment epithelial (RPE-1) cells expressing protein VII. Expression of GFP-tagged protein VII at levels comparable to WT infection (Figure S1G–H) blocked proliferation but did not induce cell death like cells arrested by treatment with CDK-1 inhibitor, RO-3306⁴⁴ (Figure 4C and S4D–F). Protein VII-GFP expression increased the proportion of G2/M DNA-content cells (Figure 4D) and increased levels of histone H3 serine 10 phosphorylation (H3S10ph), a mitotic marker⁴⁵ (Figure 4E). Together these results indicate that protein VII slows the cell cycle in diploid human cells, perhaps by allowing cells to enter mitosis but not complete it, leading to an enrichment of mitotic cells. This outcome echoes our observations in yeast, and we therefore conclude that expression of protein VII impairs cell cycle progression through mitosis.

Protein VII blocks proliferation of HEK 293T cells

During adenovirus infection, protein VII is expressed from the L2 late transcriptional unit¹⁸. The potent inhibition of proliferation and cell cycle progression in human cells by protein VII is in stark contrast with the effects of early-expressed adenovirus genes E1A and E1B which direct bypass of cell cycle checkpoints¹⁹. In fact, E1A and E1B are the primary drivers of transformation of the HEK 293 human cell line⁴⁶. Consequently, we asked whether protein VII could override the pro-proliferation effects of E1A and E1B. We expressed HA-tagged protein VII in HEK 293T cells under a doxycycline-inducible promoter and found that protein VII expression slowed growth significantly (Figure 4F and S4G–I). The loss of proliferation in HEK 293T cells occurred more gradually than in RPE-1 cells likely due to lower levels of protein VII (Figure S1G–H). Like the RPE-1 cells, protein VII expression decreased the number of G1 cells and increased the proportion of S and G2/M cells (Figure 4G) resulting in accumulation of H3S10ph (Figure 4H). The growth impairment was not caused by widespread cell death (Figure S4I) nor by loss of E1A nor E1B expression (Figure 4H and S4J). To determine that these effects are specific to protein VII, we expressed human sperm protamine 1 (PRM1)⁴⁷ and adenovirus protein V in the same cell type (Figure S4J–O). PRM1, which was unstable (Figure S4M), slowed cell growth in a non-specific manner as its expression had no effect on the cell cycle nor H3S10ph levels (Figure S4L–O). As in yeast, adenovirus protein V did not impact cell growth, cell cycle progression, or H3S10ph levels. Together these data suggest that during infection, the late-expressed protein VII could counteract the effects of E1A and E1B on cell cycle regulation and prevent cell division. Consequently, we next asked how protein VII impacts host cell cycle markers during adenovirus infection.

Protein VII contributes to host cell cycle dysregulation during late stages of adenovirus infection

We examined the levels of several cell cycle-dependent markers during infection with conditional deletion of protein VII. We infected HEK 293 cells constitutively expressing Cre recombinase (Figure S5A) with a floxxed protein VII gene adenovirus type 5 (loxP-VII-loxP Ad5)⁴⁸ and compared cell cycle markers during infection with the same virus in WT HEK 293 cells (Figure 5A). As previously reported, adenovirus DNA binding protein (DBP) abundance was unaffected by Cre suggesting that viral genome replication is robust despite protein VII deletion (VIIΔ)⁴⁸. CDT1 and cyclin E1, host cell cycle dependent proteins that are highly expressed during G1 or the G1/S transition respectively^49,50, were mainly unchanged by deletion of protein VII (Figure 5A–B and S5B). However, cyclin A2, phosphorylated CDK1 (ph-CDK1), thymidine kinase (TK1), geminin, cyclin B1, and H3S10ph were detected at higher levels in the VIIΔ infection compared to WT (compare quantification of lanes 2–4 to 6–8 in Figure 5A and S5B). These cell cycle regulated factors are all upregulated during G2 or mitosis (Figure 5B)^{45,49,51–55}. Their change in levels during infection without protein VII is consistent with our observation that ectopic protein VII impacts the transition from S phase to G2 to mitosis. Taken together, these results suggest that loss of protein VII during infection can lead to aberrant cell cycle progression.

Figure 5. Protein VII contributes to cell cycle manipulation during late adenovirus infection.

(A) Western blot analysis of cell cycle markers during adenovirus infection with conditional deletion of protein VII. VIIΔ - HEK 293 cells with constitutive Cre recombinase expression. VII – WT HEK 293 cells. Both cell types infected with adenovirus type 5 with loxP sites flanking the VII gene⁴⁸. Proteins probed as indicated. The relative amounts of protein detected in each lane were normalized to mock (M) and values are shown in italic numerals below each blot. H3 loading control. (B) Cell cycle protein expression schematic for uninfected cells (top) compared to WT adenovirus infection with (WT) and without protein VII (VIIΔ). Expression of protein VII late during infection contributes to manipulation of host cell cycle proteins that regulate the transition from G2 to mitosis. (C) Immunofluorescence microscopy of geminin (green) in A549 cells during adenovirus infection with conditional protein VII (red) deletion. The asterisk indicates a nucleus with high geminin and low protein VII. The arrow indicates a nucleus with low geminin and high protein VII. Scale bar is 10 μm. (D) Scoring of nuclei stained for geminin and protein VII as described in (C) and STAR methods, n≥30. +TC – Tat-Cre treated; NT – no treatment. Detection of protein VII was significantly reduced in the cells conditionally deleted for protein VII (p<0.0001). Geminin staining was significantly decreased in cells that also expressed high levels of protein VII (p<0.0001). P-values determined with Fisher’s exact test on absolute number of cells. (E) Immunofluorescence microscopy of HMGB1 (green) and protein VII (red) localization in cells as described in (C). Scale bar is 10 μm. (F) Model for protein VII on host chromatin during infection. An uninfected cell has normal cell cycle regulation. Upon infection, the early-expressed adenovirus genes E1A and E1B bypass cell cycle checkpoints and push the cell into an S phase-like gene expression program. E1A acts primarily through redistributing acetylation of histone H3K18 globally. Late during infection, protein VII recruits HMGB1 and SET to displace the linker histone from host chromatin. Protein VII chromatin disruption contributes to manipulation of cell cycle progression, culminating in cell lysis and maximizing the release of viral progeny. See Figure S5.

Next, we carried out the VIIΔ infection in a biologically relevant cell type, A549 lung epithelial cells, since adenovirus 5 causes respiratory infections⁵⁶. We treated the cells with purified Cre recombinase fused to HIV Tat⁵⁷ allowing for uptake and nuclear translocation of the nuclease prior to loxP-VII-loxP Ad5 infection (Figure S5A). This strategy resulted in less efficient gene deletion compared to HEK 293 cells and partial reduction in protein VII expression (Figure S5C–D). The amount of protein VII remaining was roughly comparable to the levels expressed ectopically by the inducible HEK 293T cell line, or roughly ~5–10% of WT infection abundance (Figure S1G–H). This residual protein VII may be sufficient to disrupt chromatin, and therefore the cell cycle, which would account for the subtle differences in cell cycle marker levels in the A549 cells (Figure S5D). Nevertheless, immunofluorescence microscopy allowed us to identify cells in the population without detectable protein VII (Figure S5E). We found that these low-protein VII cells also had high levels of geminin and, conversely, cells with high protein VII had low geminin (Figure 5C–D). Geminin abundance increases during G2 and decreases during mitosis as it is degraded during the metaphase-anaphase transition⁵⁴. In the WT infection, the number of cells with high protein VII signal corresponded with fewer cells with high geminin levels. The significant anticorrelation between VII and geminin abundance was lost by conditional depletion of protein VII. Taken together, these results suggest that protein VII functions late during infection to hijack host cell cycle regulation.

We next examined HMGB1 localization during infection without protein VII (Figure 5E). During WT infection, HMGB1 localization changes dramatically from diffuse nuclear staining to chromatin bound²³. Without protein VII, HMGB1 no longer localized to chromatin and remained diffuse in the infected nuclei (Figure 5E). Since the changes in cell cycle marker levels correspond to loss of protein VII which prevents HMGB1 localization to chromatin, it is likely that the presence of HMGB1 at chromatin along with protein VII directly contributes to cell cycle regulation effects during infection. Since protein VII is not fully deleted from A549 cells and the redundant HMGB2 and HMGB3 factors are present, we suspect other viral factors impact cell cycle manipulation late during infection. Collectively these results indicate that protein VII, together with chromatin-bound HMGB1 and SET, is intimately involved in disruption of cell cycle regulation during infection.

Discussion

For nuclear-replicating viruses like adenovirus, host chromatin is an impediment to and a bountiful resource for ensuring successful replication. Previous work showed that protein VII directly interacts with SET^14,15 and HMGB1 and recruits them to host chromatin during infection²³. Proteomic analysis of chromatin bound by protein VII revealed that linker histones are depleted while SET and HMGBs are enriched, suggesting that protein VII co-opts these host factors to disrupt linker histone occupancy. Thus, we hypothesized that protein VII exploits the antagonism between H1 and HMGB1 to undermine chromatin structure and maximize viral success. Here, we coupled budding yeast genetics and human cell experiments to show that protein VII disrupts chromatin via SET and HMGB proteins to displace linker histones and impede the cell cycle.

Protein VII expression elicited significant growth defects in two distinct lab yeast strains, demonstrating the robustness of protein VII’s chromatin perturbation. Growth defects were rescued by mutation of protein VII PTM sites which are responsible for chromatin localization in human cells²³. It is possible that protein VII is modified in yeast and that these modifications regulate protein VII’s interaction with and impact on chromatin. Alternatively, structural changes introduced by these point mutations could prevent protein VII from impacting chromatin despite its stable expression. At this time, it is not clear how PTM sites affect HMGB1 or SET binding.

Deleting HMO1 or NAP1, the yeast homologs for HMGB1 and SET, rescued the protein VII growth defects while deletion of the linker histone, HHO1, exacerbated this phenotype (Figure 2). We found slight variations in the degree of rescue between the two lab strains tested. Given that W303 has nearly ~9500 single nucleotide polymorphisms affecting nearly 12% of genes compared to FY602³⁹, it is likely that this variation gives rise to the different apparent impact of HMO1 or NAP1 deletion. Nonetheless, for both strains, expressing the homologous human factors when protein VII was present restored the growth defects in the corresponding deletion strains. The robust effect of protein VII and its specific genetic interactions shared by both strains emphasizes that the chromatin mechanisms exploited by protein VII are deeply conserved.

Consistent with our model that protein VII and H1 compete for DNA binding sites, deletion of HHO1 exacerbated the effects of protein VII on yeast growth (Figure 2) while human H1.5 alleviated the protein VII growth deficit (Figure 3). As such, H1 may protect chromatin from protein VII invasion, although the stoichiometry of the competition between protein VII and H1 remains unknown. Loci with higher H1 occupancy may be resistant to protein VII binding, and thus protein VII requires the action of SET and HMGB proteins to facilitate its chromatin invasion. This would explain why the effects of HHO1 deletion in yeast are relatively slight, while expression of human H1.5 resulted in significant rescue of protein VII growth defects. Compared to human cells, yeast have very little heterochromatin which is silenced by means other than linker H1⁵⁸. Therefore, a large portion of the yeast genome is vulnerable to protein VII invasion (Figure S1J and S2), even without HHO1 deletion, compared to the human genome which is thought to be comprised of ~25% heterochromatin⁵⁹.

As in yeast, protein VII led to dramatic loss of proliferation and cell cycle obstruction in diploid RPE-1 and transformed HEK 293 cells (Figure 4). The most profound effects on the cell cycle occurred at the G2/M transition, but also increased the S phase population for both cell types, which suggests that protein VII’s interruption of cell cycle progression occurs at multiple points. In HEK 293 cells, E1A promotes proliferation and transforms the cells by re-wiring their basal epigenetic state¹⁹. In contrast, protein VII expression does not cause significant changes to H3K18ac or most other histone marks²³. This suggests that the mechanism of protein VII’s block of proliferation in HEK 293 cells is not through changes to histone modifications. Likewise, although protein VII and E1A directly interact⁶⁰, the effects of protein VII on proliferation and cell cycle regulation do not require this interaction given that protein VII’s effects are observed in human cells without E1A (Figure 4C–E) and in budding yeast (Figure 1).

It is remarkable that, although budding yeast and humans are separated by one billion years of divergent evolution, protein VII disrupts yeast chromatin as it does human chromatin. In light of this new role for protein VII in undermining cell cycle progression, we propose that protein VII expression late in the infection cycle maximizes progeny production by counterbalancing checkpoint bypass and transcriptional reprogramming initiated by the early-expressed factors E1A and E1B (Figure 5B and 5F). These earliest expressed proteins induce a “viral S phase”⁶¹ while a complement of other early adenovirus proteins inactivate the host DNA damage response⁶². Since both the intra-S phase and G2 host cell cycle checkpoints rely on sensing DNA damage to stall cell cycle progression, it is conceivable that, because of this unchecked “viral S phase”, an infected cell could initiate mitosis and cytokinesis before viral progeny assembly and release. In this scenario, breakdown of the nucleus would disrupt progeny assembly and cell division would redirect nuclear resources away from viral replication thereby reducing progeny. Therefore, we propose that protein VII functions as a backstop against cell cycle checkpoint bypass by early adenovirus proteins to prevent an infected cell from initiating cytokinesis. Other adenovirus proteins may also prevent aberrant mitosis to safeguard against any potential transformation in vivo. In sum, our findings in yeast and human cells reveal a highly conserved chromatin vulnerability that adenovirus exploits, underpinning the evolutionary conservation of H1-HMGB antagonism that may be exploited by a wide range of intracellular pathogens.

STAR Methods

RESOURCE AVAILABILITY

Lead contact

Information and requests for reagents and resources should be directed to and will be fulfilled by the lead contact, Daphne Avgousti (avgousti@fredhutch.org).

Materials availability

Plasmids, yeast strains, and human cell lines generated in this study are available through the lead contact, Daphne Avgousti.

Data and code availability

ChIP-seq data have been deposited at GEO and are publicly available as of the date of publication. Accession numbers are listed in the key resources table. Other data reported in this paper will be shared by the lead contact upon request. This paper does not report original code. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

KEY RESOURCES TABLE.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit Anti-GFP	Abcam	Cat: ab290
Mouse Anti-HA	BioLegend	Cat: MMS-101R
Rabbit Anti-Histone H3	Abcam	Cat: ab1791
Mouse Anti-PGK1	Abcam	Cat: ab113687
Mouse Anti-FLAG (M2)	Sigma	Cat: F3165
Mouse Anti-Vinculin	Sigma	Cat: V9131
Rabbit Anti-H3S10ph	EMD Millipore	Cat: 06–570
Mouse Anti-E1A	BD Biosciences	Cat: 554155
Mouse Anti-E1B-55 kDa	Laboratory of A. Levine	N/A
Rabbit Anti-Cre	EMD Millipore	Cat: 69050
Rabbit Anti-VII	Laboratory of L. Gerace	N/A
Mouse Anti-VII	Laboratory of H. Wodrich33	N/A
Rabbit Anti-Geminin	Cell Signaling Technology	Cat: 52508
Rabbit Anti-CDT1	Cell Signaling Technology	Cat: 8064S
Rabbit Anti-Thymidine Kinase 1	Cell Signaling Technology	Cat: 28755S
Rabbit Anti-H3S10ph	Cell Signaling Technology	Cat: 53348
Rabbit Anti-Cyclin A2	Cell Signaling Technology	Cat: 91500S
Rabbit Anti-Cyclin B1	Cell Signaling Technology	Cat: 12231
Rabbit Anti-Cyclin E1	Cell Signaling Technology	Cat: 20808S
Rabbit Anti-phCDC2	Cell Signaling Technology	Cat: 4539
Mouse Anti-CDC2 p34	Santa Cruz Biotechnology	Cat: sc-54
Peroxidase AffiniPure Goat Anti-Rabbit IgG (H+L)	Jackson ImmunoResearch	Cat: 111-035-045
Peroxidase AffiniPure Goat Anti-Mouse IgG (H+L)	Jackson ImmunoResearch	Cat: 115-035-003
Goat anti-Rabbit IgG (H+L) Cross-Adsorbed Secondary Antibody, Alexa Fluor 488	Invitrogen	Cat: A11008
Goat anti-Mouse IgG (H+L), Superclonal™ Recombinant Secondary Antibody, Alexa Fluor 555	Invitrogen	Cat: A28180
Bacterial and virus strains
rAdV-GFP	Laboratory of D. Curiel23	N/A
rAdV-VII-GFP	Laboratory of D. Curiel23	N/A
Tuner™(DE3) pLacI Competent E. coli Cells	Millipore Sigma	Cat: 70625-4
WT Adenovirus Type 5	Laboratory of M. Weitzman	N/A
loxP-VII-loxP Adenovirus Type 5	Laboratory of P. Hearing48	N/A
Biological samples
VII-3xHA gBlock Gene Fragment	Integrated DNA Technologies	N/A
ΔPTM-3xHA gBlock Gene Fragment	Integrated DNA Technologies	N/A
V-3xHA PriorityGENE	GeneWiz	N/A
3xFLAG-HMGB1 PriorityGENE	GeneWiz	N/A
3xFLAG-SET PriorityGENE	GeneWiz	N/A
3xFLAG-H1.5 PriorityGENE	GeneWiz	N/A
Chemicals, peptides, and recombinant proteins
SC-Ura Powder	Sunrise Science Products	Cat: 1306-030
SC-His Powder	Sunrise Science Products	Cat: 1303-030
Yeast Nitrogen Base	Sunrise Science Products	Cat: 1501-250
Raffinose	Fisher	Cat: AC195675000
Galactose	Fisher	Cat: BP656-500
Dextrose	Fisher	Cat: D16-500
ProLong Gold Antifade Mounting Solution	Invitrogen	Cat: P36934
Invitrogen Protein G Dynabeads for Immunoprecipitation	ThermoFisher	Cat: 10004D
AMPure XP Beads	Beckman Coulter	Cat: A63880
Tetracycline-free Fetal Bovine Serum	Gemini Bio-Products	Cat: 100-800; Lot #A69G00J
Doxycycline	Fisher Bioreagents	Cat: BP26535
MilliporeSigma™ Calbiochem™ DAPI Stain	MilliporeSigma	Cat: 50-874-10001
Propidium iodide	Fisher	Cat: AC440300250
RO-3306	Sigma-Aldrich	Cat: SML0569-5MG
Tat-Cre	57	N/A
Opti-MEM Reduced Serum Media	ThermoFisher	Cat: 31985070
Critical commercial assays
Unique Dual-Indexed Adapter Kit	KAPA Biosystems	Cat: KK8727
Hyper Prep Kit	KAPA Biosystems	Cat: KK8504
Qubit dsDNA HS Assay Kit	Invitrogen	Cat: Q32851
High Sensitivity D5000 ScreenTape System	Agilent	Cat: 5067-5592
Clarity Western ECL Substrate	BioRad	Cat: 1705061
RNeasy Plus Mini Kit	Qiagen	Cat: 74104
iScript Reverse Transcription Supermix	BioRad	Cat: 1708841BUN
QIAamp DNA Mini Kit	Qiagen	Cat: 51304
Deposited data
Analyzed ChIP-seq data	This paper	GEO: #GSE164684
Experimental models: Cell lines
Human: RPE-1 cell line (hTERT immortalized)	Laboratory of E. Hatch	N/A
Human: HEK 293T HILO acceptor cells	Laboratory of E. Makeyev23,63	N/A
Human: HEK 293 Cre	Laboratory of P. Hearing48	N/A
Human: HEK 293 cell line	Laboratory of M. Weitzman	N/A
Human: A549 cell line	Laboratory of A. Berger	N/A
Experimental models: Organisms/strains
S. cerevisiae: Strain background: WT W303	Laboratory of B. Brewer and M.K. Raghuraman	N/A
S. cerevisiae: Strain background: W303 hmo1::KanMx	Laboratory of T. Tsukiyama	N/A
S. cerevisiae: Strain background: W303 hho1::KanMx	Laboratory of T. Tsukiyama	N/A
S. cerevisiae: Strain background: W303 nap1::HYG	Laboratory of T. Tsukiyama	N/A
S. cerevisiae: Strain background: W303 nap1::HYG hmo1::HIS3	This paper	N/A
S. cerevisiae: Strain background: W303 nhp6a::HYG nhp6b::NAT	Laboratory of T. Tsukiyama	N/A
S. cerevisiae: Strain background: WT FY602	Laboratory of L. Howe35	N/A
S. cerevisiae: Strain background: FY602 hmo1::HIS3	This paper	N/A
S. cerevisiae: Strain background: FY602 nap1::KanMx	This paper	N/A
S. cerevisiae: Strain background: FY602 hho1::KanMx	This paper	N/A
S. cerevisiae: Strain background: FY602 hmo1::HIS3 nap1::KanMx	This paper	N/A
Oligonucleotides
See Table S2 for oligonucleotide information
Recombinant DNA
pRS416-GAL1-10-DVNP-3xHA-NLS	Laboratory of L. Howe35	N/A
pRS416-GAL1-10-GFP	This paper	N/A
pRS416-GAL1-10-VII-3xHA	This paper	N/A
pRS416-GAL1-10-VII-ΔPTM-3xHA	This paper	N/A
pRS416-GAL1-10-V-3xHA	This paper	N/A
pRS416-3xFLAG-HMGB1-biGAL1-10-GFP	This paper	N/A
pRS416-3xFLAG-SET-biGAL1-10-GFP	This paper	N/A
pRS416-3xFLAG-H1.5-biGAL1-10-GFP	This paper	N/A
pRS416-3xFLAG-HMGB1-biGAL1-10-VII-3xHA	This paper	N/A
pRS416-3xFLAG-SET-biGAL1-10-VII-3xHA	This paper	N/A
pRS416-3xFLAG-H1.5-biGAL1-10-VII-3xHA	This paper	N/A
pTriEx-HTNC	57	Addgene Plasmid #13763
DA0045-VII-HA	23	N/A
DA0173-GFP	23	N/A
DA0048-PRM1-HA	23	N/A
DA0049-V-HA	23	N/A
DA0021-Cre	23	N/A
Software and algorithms
GraphPad Prism v9.0	GraphPad Software	https://www.graphpad.com/
Image Lab v6.1	BioRad	https://www.bio-rad.com/en-us/product/image-lab-software?ID=KRE6P5E8Z
Bowtie2	Publicly available	http://bowtie-bio.sourceforge.net/bowtie2/index.shtml
deepTools2	Publicly available	https://deeptools.readthedocs.io/en/develop/
MACS2	Publicly available	https://github.com/macs3-project/MACS
Adobe Creative Cloud Photoshop	Adobe Inc.	https://www.adobe.com/creativecloud.html
Adobe Creative Cloud Illustrator	Adobe Inc.	https://www.adobe.com/creativecloud.html
FlowJo v10	FlowJo, Inc.	https://www.flowjo.com/solutions/flowjo
CFX Maestro v2.2	BioRad	https://www.bio-rad.com/en-us/sku/12013758-cfx-maestro-software-2-0-for-windows-pc?ID=12013758
Leica Application Suite V4.12	Leica Microsystems	https://www.leica-microsystems.com/products/microscope-software/p/leica-application-suite/downloads/
Fiji for Mac OS X	Publicly available	https://fiji.sc/

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Yeast strains and growth

We introduced gene deletions for HMO1, HHO1, and NAP1 using one-step homologous recombination gene replacement to haploid FY602, a derivative of S288C, or haploid W303. HIS3 (deletion of HMO1) was amplified from pRS413 and the KanMx6 gene (deletion of HHO1 or NAP1) was amplified from pUG6. The linear deletion fragments or plasmids were introduced by standard lithium acetate transformation and growth on selective media. Transformed colonies were streaked out for single isolates on selective media and candidate gene deletions were confirmed by diagnostic PCR and Sanger sequencing of the insertion junctions. For strains obtained from other labs, putative gene deletions were confirmed by diagnostic PCR to confirm integration of the selection gene. See Table S2 for primer sequences. Expression plasmids were transformed into relevant strains every four weeks. Colonies were streaked onto selective media, grown at 30°C until single colonies formed, and then stored at 4°C. Strains containing expression plasmids were maintained in selective media for all pre-growth steps.

Standard rich media (YEP+D) was prepared with 2% dextrose and 2.4% agar with 400 mg/mL kanamycin for selection. Complete media without uracil (SC-ura) or histidine (SC-his) was purchased from Sunrise Science Products and prepared per the manufacturer’s instructions with yeast nitrogen base and 2.4% agar for solid media growth assay plates. To derepress the galactose-promoter, cells were grown in SC-ura + 2% raffinose media. To induce protein expression, cells were grown in SC-ura + 2% galactose media. To repress protein expression, cells were grown in SC-ura + 2% dextrose media.

Human cell lines and growth

hTERT-immortalized RPE-1 cells (sex: female) were grown in 1:1 DMEM/F12 media with 10% FBS, 1% penicillin-streptomycin, and 0.01% hygromycin to maintain selection for hTERT expression. HEK293T HILO cells (sex: female) with either a GFP negative control, HA-tagged protein VII, PRM1, or protein V under a doxycycline-inducible promoter were generated as previously reported^23,63. After plasmid transfection and puromycin selection, HEK293T HILO cell lines were grown in DMEM with tetracycline-free 10% FBS, 1% penicillin-streptomycin and 1 ug/mL puromycin to maintain selection for the inducible construct. The HEK 293 cell line with constitutive Cre expression was grown in DMEM with 10% FBS, 1% penicillin-streptomycin, and 0.25 mg/mL geneticin. Standard HEK 293 cells were grown in DMEM with 10% FBS, and 1% penicillin-streptomycin. A549 cells (sex: male) were grown in Ham’s F 12K (Kaighn’s) Medium with 10% FBS, and 1% penicillin-streptomycin. All cell lines were grown at 37°C with 5% CO₂. Cell authentication was not performed. Cell lines were monitored for mycoplasma contamination approximately once per month.

METHOD DETAILS

Yeast Expression Plasmids

We used a standard unidirectional GAL1–10 promoter plasmid based on pRS416 and inserted a protein VII sequence which was codon optimized using the Integrated DNA Technologies (IDT) algorithm for expression in S. cerevisiae. To do so, we synthesized gBlock Gene Fragments (IDT) corresponding to mature protein VII from human adenovirus type 5, a version of the same protein coded for alanine substitutions at five residues previously found to be post-translationally modified at sites K2K3, K25, T31, T50 and S159²³, and adenovirus protein V from human adenovirus type 5. All were encoded with three consecutive HA epitope tags at the C terminus. These DNA fragments, as well as yeast codon-optimized GFP sequence from the plasmid PFA6A-GFP-KanMx, were PCR amplified to have 3’ and 5’ SpeI and HindIII restriction sites respectively. The amplified sequences were then cloned and confirmed by Sanger sequencing across both junctions. The DVNP expression plasmid was a gift from L. Howe. To generate the co-expression plasmids with either WT protein VII or GFP, the truncated, unidirectional GAL promoter was first replaced with the full-length bidirectional GAL1–10 that was PCR amplified from the FY602 strain and engineered with XbaI and SacI restriction sites. We confirmed successful replacement with the bidirectional GAL1–10 promoter by Sanger sequencing. Next, we PCR amplified human SET, human HMGB1, or H1.5 DNA fragments (all codon optimized for yeast) containing SacI restriction sites at both ends of the fragments from plasmids synthesized by GeneWiz. These fragments were then cloned opposite either GFP or protein VII allowing for co-expression of either gene with human SET, HMGB1 or H1.5. Plasmid sequences are available from lead contact.

Yeast Growth Assays and Quantification

For serial dilution growth assays, plasmid-bearing strains were pre-grown for 24 hours in SC-ura + 2% dextrose liquid media at 30°C with agitation. Cultures were sonicated at 35% amplitude for 7 seconds using a Fisher Scientific FB50 Sonic Dismembrator to disaggregate the cells before dilution. Cells were diluted in ultrapure water and counted by hemocytometer, then six 1:3 serial dilutions were prepared in a 96-well plate such that the most dilute cell mixture contained 3 cells/μL. After vigorous trituration, 5 μL of each solution was placed onto SC-ura + 2% dextrose agar media to repress protein expression or SC-ura + 2% galactose agar media to induce protein expression. The plates were dried at room temperature and then incubated at 30°C for six days. Each day, the plates were imaged on a BioRad ChemiDoc MP Imaging System. The images were processed using Adobe Creative Cloud Photoshop and Illustrator. The serial dilution assays were quantified using BioRad Image Lab v6.1 software. Using the Volume Tools analysis package, a box with the same pixel dimensions was drawn around the third position in the dilution series (see Figure 1) for both the dextrose- and galactose-grown cells and the signal intensity values were measured for each position independently. The local subtraction method was used to remove background signal. To determine the relative growth value, the signal intensity of grown on galactose media was divided by the signal intensity from dextrose media growth. For WT, nap1Δ and hho1Δ strains with the single protein expression plasmids, which all had the same growth rate, the relative growth values were calculated from the images taken after three days growth at 30°C. To compensate for slowed growth caused by HMO1 deletion, the relative growth for hmo1Δ and hmo1Δnap1Δ strains was measured after four days growth. For strains with the co-expression plasmids, which had slightly different growth rates, relative growth was calculated after five days growth during GFP expression or six days growth during protein VII expression.

For the exponential growth analysis, strains were pre-grown for 24 hours in SC-ura + 2% dextrose liquid media at 30°C with agitation, then scaled up to a larger culture volume in SC-ura + 2% raffinose liquid media and grown for another 24 hours. To start the exponential growth experiments, the raffinose-grown cultures were diluted to OD₆₆₀ of 0.15 in SC-ura + 2% dextrose or SC-ura + 2% galactose and incubated at 30°C with agitation. The OD₆₆₀ was measured every hour and graphs were generated in real time. To determine exponential growth rate, the OD₆₆₀ was graphed on a log-10 scale and an exponential growth equation with least squares fit was fit to at least five datapoints using GraphPad Prism v9.0 software. The bar charts represent the average exponential growth value (k) from at least three biological replicate experiments.

Yeast Western Blotting

To confirm protein expression, western blot samples were recovered after six hours induction or repression during exponential growth. The cells were washed once with ultrapure water and dry pellets were stored at −80°C. Whole cell lysate was prepared by cell lysis in 0.1M NaOH and extraction by heating at 65°C in SUMEB loading buffer (1% SDS, 8 M Urea, 10 mM MOPS, pH 6.8, 10 mM EDTA, 0.02% bromophenol blue) with MilliPore Sigma protease inhibitor cocktail and 5% 2-betamercaptoethanol for 10 minutes. Samples were loaded onto 15% SDS-PAGE gels and run via standard methods in NuPAGE MOPS SDS running buffer. Gels were transferred (Transfer Buffer: 25 mM Tris Base, 100 mM glycine, 20% methanol) onto nitrocellulose membrane and Ponceau stained, blocked with 5% milk in TBS-T for 30 minutes, and probed with primary antibody overnight. Primary antibodies and dilutions used were: α-GFP (1:2500), α-HA (1:500), α-PGK1 (1:5000), α-H3 (1:2500), and α-FLAG (1:1000). Blots were probed with α-mouse (1:5000) or α-rabbit (1:5000) HRP-conjugated secondary antibody and developed with ECL per the manufacturer’s instructions. See Key Resource Table for more antibody information. Images were acquired using the BioRad ChemiDoc MP Imaging System.

Yeast Immunofluorescence

Cultures were pre-grown in liquid SC-ura + 2% raffinose media as previously described, then scaled to 50 mL culture volume at an OD₆₆₀ of 0.3 in SC-ura + 2% dextrose or SC-ura + 2% galactose liquid media. Cultures were incubated for 4 hours at 30°C with agitation. The number of cells collected was normalized by OD₆₆₀. Cells were sonicated for 10 seconds at 50% duty cycle, output 2 using a Branson Sonifier 250 Model SSE-1 probe sonicator. The cells were pelleted by centrifugation at 3000 rpm for 5 minutes, washed twice in ultrapure water, resuspended, and fixed in 1 mL 4% PFA in PBS by incubation for 1 hour at room temperature on a nutating platform mixer. Cells were pelleted by centrifugation, washed and then resuspended in 1 mL Wash Buffer (0.1 M KH₂PO₄ in 1.2M sorbitol solution). Cells were counted by hemocytometer and 40 million cells were moved to a fresh microtube and brought up to 200 μL total volume in Wash Buffer. Zymolyase and 2-mercaptoethanol were added to the cells which were incubated for 50 minutes at 30°C to spheroplast the cells. Cells were washed twice in Wash Buffer then suspended in 200 μL 3% BSA in PBS and incubated at room temperature with agitation for 30 min to block. Cells were incubated in primary antibody solution containing α-HA (1:400) and α-PGK1 (1:400) in 3% BSA solution, then incubated for 1 hour with rocking at room temperature. Cells were washed thrice in 3% BSA solution, then suspended in 200 μL secondary antibody solution containing α-rabbit Alexa Fluor 488 (1:300), α-mouse Alexa Fluor 555 (1:300), and 2 ug/mL DAPI. Cells were washed once in 3% BSA solution, three times in PBS, and resuspended in 50 μL PBS. Cells were counted by hemocytometer and diluted to a final concentration of 500,000 cells per μL in PBS. 2 μL of cell solution was mixed with mounting solution and mounted on poly-lysine slides. Slides were visualized with a Zeiss 780 LSM confocal microscope, maximum intensity projection images were processed in FIJI and assembled using Adobe Photoshop and Illustrator.

Yeast Native ChIP-seq

After 6 hours galactose induction during log-phase growth in liquid culture, 500 mL culture volume of cells was collected, nuclei purified, chromatin isolated, and native ChIP performed as previously described⁶⁴. The ChIP protocol was performed using α-H3 and α-protein VII (raised in mouse) antibodies and Invitrogen Protein G Dynabeads for Immunoprecipitation. Libraries were prepared from the ChIP samples as well as a 10% input, following manufacturer recommendations, with KAPA Biosystems Unique Dual-Indexed Adapter Kit, Hyper Prep Kit, and Beckman Coulter AMPure XP Beads for library cleanup and size selection. Library concentrations were measured with Invitrogen Qubit dsDNA HS Assay Kit then analyzed with Agilent High Sensitivity D5000 ScreenTape System, and pooled. Libraries were sequenced with 50-bp paired-end reads on an Illumina NovaSeq 6000 SP sequencer at the Fred Hutch Genomics Core Facility. Reads were aligned to the Saccer3 yeast genome assembly (GenBank accession GCA_000146045.2) using Bowtie2. Read coverage was assessed using bamCoverage package from DeepTools2.0. Reads aligning uniquely to the yeast genome were used to identify protein VII peaks using the MACS2 peak caller. Peaks designated as “narrow” and bedGraph tracks showing read coverage for the input, H3, and protein VII ChIP samples for one replicate are shown in Figures S1J and S2. ChIP-seq was performed in duplicate. See Key Resource Table for GEO accession number.

Yeast Budding Index

WT W303 strains were pre-grown in raffinose and transferred to SC-ura + 2% dextrose or SC-ura + 2% galactose liquid media as during exponential growth analysis described above. Samples were collected every 1–2 hours during growth/induction. 500 μL of each culture was pelleted, washed once in ultrapure water, resuspended in 70% ethanol, and fixed overnight at 4°C. Samples were pelleted, washed once with 50 mM sodium citrate, resuspended and diluted in the same solution. Cells were sonicated as described above and then loaded onto a hemocytometer counting chamber. 120–170 cells per sample were classified as “unbudded”, “small-budded” (bud < 50% the size of the mother cell), “large-budded” (bud > 50% the size of the mother cell), or “irregular” (multiple buds or other irregularities). Data were graphed and statistical analyses performed as described in the figure using GraphPad Prism v9.0 software.

Exogenous Protein Expression in Human Cells

RPE-1 cells were plated at 20% confluence in 6-well plates and the following day cells were transduced with recombinant adenovirus with the E1 region replaced with either GFP-tagged protein VII (MOI = 50) or a GFP control (MOI = 50) under a CMV promoter using standard methods²³. Control cells were treated with 10 μM RO-3306. For all treatment groups, samples were collected every 12 hours post-treatment for 48 hours total as described below. The HEK 293T doxycycline-inducible cell lines were plated with 0.2 μg/mL doxycycline to induce protein expression and doxycycline was refreshed every 24 hours. All cell lines and treatment groups were imaged, and samples were collected every 24 hours post-treatment for 96 hours. Cells were imaged using Leica Microsystems DM IL LED Fluorescent microscope with Leica Application Suite V4.12 software. Images were processed using FIJI, Adobe Photoshop, and compiled with Adobe Illustrator.

Human Cell Proliferation Assay

RPE-1 cells were harvested every 12 hours after transduction for 48 hours as was a pre-transduction sample. HEK 293T-based cells were harvested every 24 hours after plating and induction up to 96 hours. For each collection, both cell types were lifted by trypsinization. Cells were pelleted by centrifugation (1200 rpm for 2 minutes) and then suspended in 1 mL PBS + 1 mM EDTA to prevent clumping. The cell suspension was diluted 1:1 with 0.4% Trypan Blue solution and cells were counted by hemocytometer or using a BioRad TC20 Automated Cell Counter. The total number of cells in each well and the number of trypan negative cells was reported.

Human Cell Flow Cytometry

Cells were lifted with trypsin, pelleted by centrifugation (1000 rpm for 5 minutes at 4°C), and washed in 5 mL FACS Buffer (PBS + 0.1% BSA). Cells were pelleted and resuspended in 1 mL 0.5% PFA in PBS and incubated for 5 minutes at room temperature to fix the cells. Cells were washed twice in FACS Buffer and kept on ice or at 4°C for all remaining processing. Cells were suspended in 70% ethanol and stored at −20°C overnight. The cells were washed three times in PBS + 0.1% Triton-X, then resuspended in 500 μL PBS + 0.2 mg/mL RNAse A and 20 ug/mL propidium iodide, then incubated in the dark at 4°C for at least 3 hours. Cells were run on a BD Biosciences FACSCanto II Cell Analyzer in the Fred Hutchinson Cancer Research Center Flow Cytometry core facility. Since GFP was present in >90% of the cells expected to express GFP, cells were not gated based on GFP expression. Instead, single cells were gated using propidium iodide area and height measurements. Histograms indicating DNA content were generated using FlowJo v10 software.

Human Cell Western Blotting

Western blot cell samples were counted and then pelleted and resuspended in 1x NuPAGE LDS sample buffer + 5% 2-betamercaptoethanol. The mixture was heated for 20 minutes at 95°C and stored at −20°C. Samples were run on a 15% SDS PAGE gel using standard methods as described above with the following primary antibodies: α-vinculin (1:20,000), α-GFP (1:2500), α-H3 (1:20,000), α-H3S10ph (1:1000), α-HA (1:500) or α-E1A (1:250) α-VII (1:250 rabbit), α-E1B-55 kDa (1:1000), α-Cre (1:5000), α-Geminin (1:500), α-CDT1 (1:1000), α-Thymidine Kinase 1 (1:1000), α-Cyclin A2 (1:000), α-Cyclin B1 (1:1000) α-Cyclin E1 (1:1000), or α-phCDC2 (1:1000); and HRP-conjugated secondary antibodies (α-mouse or α-rabbit; 1:5000). Blots were developed with BioRad Clarity Western ECL and images were processed as described above with antibody catalog numbers listed in the Key Resources Table.

qPCR of Transcription and Viral Replication

For analysis of protein VII transcription, cells were harvested at the indicated times post-infection or post-induction of ectopic expression by trypsinization. Cells were centrifuged at 5000g for 2 minutes and resulting cell pellets were flash-frozen in liquid nitrogen. RNA was isolated from pellets using the Qiagen RNeasy Plus Mini Kit. RNA was converted to cDNA using BioRad iScript Reverse Transcription Supermix according to manufacturer’s instructions. Quantitative PCR was performed using primers for the adenovirus protein VII gene and cellular 18s ribosomal RNA. Ct values for VII were normalized internally to 18s and then to VII transcript levels at 24 hpi from the WT adenovirus 5 infection of A549 cells. For viral DNA quantification during protein VII conditional deletion infection in A549 cells, cells were recovered as described above at 4 and 24 hpi. Genomic DNA was isolated using the Qiagen QIAamp DNA Mini Kit. Quantitative PCR was performed on 50 μg DNA using primers specific for VII, DBP, and cellular tubulin. Ct values for VII and DBP were normalized internally to tubulin and to then to the 4 hpi viral input sample. For both transcriptional and replication analyses, quantitative PCR was analyzed using the BioRad CFX384 Real-Time System. Experiments were completed in biological triplicate and statistical analysis was performed using Prism v10 (GraphPad Software). See Table S2 for oligonucleotide sequences.

Conditional Protein VII Deletion Infection

HEK 293 Cre and HEK 293 cells were infected using standard methods at MOI=10 with the loxP-VII-loxP adenovirus type 5. Protein samples were harvested at 18, 24, and 30 hours post infection, with the mock sample harvested at 24 hours. Prior to infection, A549 cells were incubated with 100 ug purified Tat-Cre⁵⁷ and suspended in a minimal volume of Opti-MEM reduced serum media for two hours with agitation every 20 minutes. The untreated control cells were incubated with Opti-MEM containing Tat-Cre storage buffer (50 mM Tris-HCl, pH 8.0, 500 mM NaCl, 150 mM Imidazole, 10% Glycerol) for the equivalent incubation period. After 2 hours, the Opti-MEM mixture was aspirated from the cells and replaced with complete Ham’s F12K (Kaighn’s) medium and placed back into the incubator at 37°C for 2 hours. The cells were then infected at MOI=10 with loxP-VII-loxP adenovirus type 5 and harvested as described above.

QUANTIFICATION AND STATISTICAL ANALYSIS

All statistical analyses were performed using Prism v9 or v10 (GraphPad Software). Statistical tests and exact value of n are described in the figure legends since these parameters varied by experiment. Bar graphs represent mean values and corresponding error bars show standard deviation from n≥3 experimental replicates. Statistical significance was defined as p≤0.05. All p-values less than the cutoff are reported in the figure legends.

Highlights:

• Protein VII causes cell cycle delays by interacting directly with chromatin

• HMGB1 and SET promote protein VII’s impact on chromatin

• Linker histones protect chromatin from protein VII’s effects

• Protein VII contributes to cell cycle hijacking by adenovirus during infection