Histone Modifications in Papillomavirus Virion Minichromosomes

ABSTRACT An unusual feature of papillomaviruses is that their genomes are packaged into virions along with host histones. Viral minichromosomes were visualized as “beads on a string” by electron microscopy in the 1970s but, to date, little is known about the posttranslational modifications of these histones. To investigate this, we analyzed the histone modifications in HPV16/18 quasivirions, wart-derived bovine papillomavirus (BPV1), and wart-derived human papillomavirus type 1 (HPV1) using quantitative mass spectrometry. The chromatin from all three virion samples had abundant posttranslational modifications (acetylation, methylation, and phosphorylation). These histone modifications were verified by acid urea polyacrylamide electrophoresis and immunoblot analysis. Compared to matched host cell controls, the virion minichromosome was enriched in histone modifications associated with active chromatin and depleted for those commonly found in repressed chromatin. We propose that the viral minichromosome acquires specific histone modifications late in infection that are coupled to the mechanisms of viral replication, late gene expression, and encapsidation. We predict that, in turn, these same modifications benefit early stages of infection by helping to evade detection, promoting localization of the viral chromosome to beneficial regions of the nucleus, and promoting early transcription and replication.

ABSTRACT An unusual feature of papillomaviruses is that their genomes are packaged into virions along with host histones. Viral minichromosomes were visualized as "beads on a string" by electron microscopy in the 1970s but, to date, little is known about the posttranslational modifications of these histones. To investigate this, we analyzed the histone modifications in HPV16/18 quasivirions, wart-derived bovine papillomavirus (BPV1), and wart-derived human papillomavirus type 1 (HPV1) using quantitative mass spectrometry. The chromatin from all three virion samples had abundant posttranslational modifications (acetylation, methylation, and phosphorylation). These histone modifications were verified by acid urea polyacrylamide electrophoresis and immunoblot analysis. Compared to matched host cell controls, the virion minichromosome was enriched in histone modifications associated with active chromatin and depleted for those commonly found in repressed chromatin. We propose that the viral minichromosome acquires specific histone modifications late in infection that are coupled to the mechanisms of viral replication, late gene expression, and encapsidation. We predict that, in turn, these same modifications benefit early stages of infection by helping to evade detection, promoting localization of the viral chromosome to beneficial regions of the nucleus, and promoting early transcription and replication. IMPORTANCE A relatively unique feature of papillomaviruses is that the viral genome is associated with host histones inside the virion. However, little is known about the nature of the epigenome within papillomavirions or its biological relevance to the infectious viral cycle. Here, we define the epigenetic signature of the H3 and H4 histones from HPV16 virions generated in cell culture and native human papillomavirus type 1 (HPV1) and bovine papillomavirus 1 (BPV1) virions isolated from bovine and human wart tissue. We show that native virions are enriched in posttranslational modifications associated with active chromatin and depleted with those associated with repressed chromatin compared to cellular chromatin. Native virions were also enriched in the histone variant H3.3 compared to the canonical histone H3. 1. We propose that the composition of virion-packaged chromatin reflects the late stages of the viral life cycle and promotes the early stages of infection by being primed for viral transcription. KEYWORDS HPV, chromatin, epigenetics, histones, papillomavirus P osttranslational modification of histones regulates many chromatin-associated processes, such as transcription, DNA replication, repair and recombination, and chromatin condensation. The genomes of several DNA viruses are also chromatinized at various stages of the viral life cycle, and this adds an additional layer of regulation to viral transcription and replication (1,2). In turn, host cell epigenetic processes can modify viral chromatin as part of the intrinsic immune response (3). The small, doublestranded, circular DNA genomes of papillomaviruses are bound by cellular histones to form minichromosomes with chromatin modifications that regulate many aspects of the persistent and productive phases of the viral life cycle (reviewed in reference 1). A relatively novel feature of the papillomaviruses is that the viral genomes are packaged with cellular histones inside the virion; the viral DNA is assembled into about 30 to 32 nucleosomes inside the capsid (4). Although polyomaviruses share this trait (5), other DNA virus families package their genomes either as naked DNA or in complex with other nonhistone DNA-binding proteins or molecules (6).
Papillomaviruses are species-specific and are thought to slowly coevolve with their unique host lineage over millions of years (7). These viruses infect and form reservoirs in dividing stem-cell-like keratinocytes in specific types of stratified epithelia. The productive phase of infection occurs only when these infected cells enter the process of terminal differentiation (7). Viral genomes amplify in cells in a G2-like phase of the cell cycle, and viral particles are produced in the most superficial, differentiated cells destined to be sloughed from the epithelial surface (8,9). The ;8 kb papillomavirus genome typically encodes just a few proteins that manipulate cellular processes at each stage of the life cycle, and it is almost certain that they have evolved to exploit chromatin modifications for regulation of their own genomes. These modifications could also promote genome packaging and virion assembly, as well as encapsidation of viral minichromosomes with specific histone modifications that could regulate immediate early viral transcription and replication upon infection of a new host cell.
In 2000, Strahl and Allis predicted that posttranslational modifications of histone tails would modulate chromatin-associated processes (10). The N-terminal tails of all four core histones (H2A, H2B, H3, and H4) can be modified by phosphorylation, methylation, acetylation, ubiquitylation, and sumoylation (11). These modifications can affect the overall structure of chromatin, but also recruit reader proteins with specific domains that recognize these modifications (e.g., bromo-, chromo-, PHD, and Tudor domains) (12). The role of acetylation and methylation of the lysine residues on the N-terminal tails of H3 and H4 are well studied. These modifications generate novel binding sites for chromatin reader proteins and, in addition, acetylation neutralizes the positive charge on lysine residues, resulting in destabilization of the histone-DNA interaction and increased accessibility of the chromatin to other binding factors (13). In general, acetylation is a signature of active chromatin, and the key acetylated residues in H3 are K4, K9, K14, K18, K23, and K27 (14). Methylation can be a mark of active (H3K4) or repressed (H3K9 or K27) chromatin (15). In H4, acetylation of K5, K8, K12, and K16 are signatures of active chromatin (though K16ac can signify either active or repressed chromatin) (16). H4K20 is methylated with mono-, di-, or tri-methylation being associated with different functions (17).
Although it has been known for decades that papillomavirus DNA is packaged in nucleosomes within the virion (4), little is known about the modifications of histones on the packaged genome. Here, we profile the posttranslational modifications on H3 and H4 histones packaged by human papillomavirus type 1 (HPV1), HPV18, and bovine papillomavirus type 1 (BPV1). We show that papillomaviruses contain histones enriched in modifications usually associated with "active" chromatin. These findings provide insight into a relatively unexplored area of papillomavirus biology and may suggest therapeutic epigenetic approaches that could interfere with viral transmission.

RESULTS
The viral minichromosome in the HPV capsid. To visualize the HPV capsid and encapsidated minichromosome, we developed a structural model of 32 host nucleosomes packaged within an HPV16 capsid. Figure 1A shows the HPV16 capsid derived from cryo-electron microscopy (cryoEM) packed with a mix of H3.1-and H3.3containing nucleosomes and ;8,000-bp circular DNA. HPV capsids also contain between 12 and 72 copies of the L2 protein (not shown), most likely positioned under the pentameric L1 capsomeres (18).
Histone modifications in HPV18 quasivirions. Initial experiments compared the modifications on histones contained in purified HPV16/18 quasivirions (virions produced using recombinant expression plasmids [19]) to those on histones from 293TT packaging cells to optimize the mass spectrometry workflow (shown in Fig. S1) and to determine whether there was preferential packaging of specific histone modifications. HPV16 quasiviruses were produced with HPV16 capsids containing an HPV18 minicircle genome (supercoiled, recircularized viral DNA generated in bacteria) that was transfected into the 293TT packaging cells. HPV16 capsids were used since they are more stable and infectious than their HPV18 counterparts. Virions were isolated using Benzonase to increase the virion yields necessary for mass spectrometry, but this increases the proportion of cellular chromatin packed in the particles, and we previously determined that 0.1% of particles contained a viral genome (20). Therefore, we also cotransfected plasmids that express the HPV18 E1 and E2 replication proteins to promote replication of the HPV18 genome. This increases packaging of viral genomes by 20-fold (20). Previous studies have also shown that the fine structure of chromatin is more authentic on replicated compared to unreplicated, transfected SV40 DNA (21 equivalents [VGE]) were denatured, separated by SDS-PAGE, and stained with methylene blue. A region of the gel corresponding to the molecular weight of the core histones (H2A, H2B, H3, and H4) was extracted and processed for mass spectrometry. Histones from the 293TT packaging cells were acid extracted and used as "input" controls. The trypsin digestion used for standard mass spectrometry cleaves the N-terminal histone tails into multiple short (1 to 3 residue) peptides that are difficult to retain chromatographically or identify by mass spectrometry. Therefore, histone samples were propionylated both before and after trypsin digestion. This modification restricts digestion at Lys residues and generates cleavage only at Arg residues, resulting in longer peptides. In addition, propionylation reduces the charge state of primary amines and increases the hydrophobicity of the peptide, which allows better retention and separation on-column. Therefore, the histone samples were propionylated prior to trypsin digestion to block cleavage at lysine residues and improve hydrophobicity (22). The peptides were separated by reverse-phase chromatography and injected into the mass spectrometer, where data-independent acquisition (DIA) analysis was performed. The data were analyzed using EpiProfile 2.3 software (23), which calculates the relative abundance of modifications on each histone peptide. The mass spectrometry (MS) data collected in this study included all histone and histone-like proteins present in the samples. However, we chose to focus on H3 and H4 modifications because these peptides constituted the highest-confidence identifications and because antibodies specific for the modifications we report are widely available. The most common modifications detected were acetylation and methylation of the N-terminal tails of histones H3 (including the variants H3.1 and H3.3) and H4. These peptides, and their sites of acetylation and methylation, are shown in Fig. 1B. The proportion of each modified peptide relative to the total amount of peptide is shown for H3 ( Fig. 2A) and H4 (Fig. 2B) and listed in Table S1. The HPV16/18 quasiviruses contained histones with a wide variety and combination of modifications, but for the most part, they were very similar to those of the host 293TT cells. To determine the modification status of each lysine residue in the H3 and H4 tails, the proportion of each modification was summed and is displayed in the pie charts in Fig. 2C. Only the modifications detected on quasivirions are shown, since they were very similar to those of the 293TT host cells. For histone H3, a small amount of K4 was monomethylated (me1), and the majority of K9, K27, and K36 were modified by mono-, di-and tri-methylation. K14, K18, and K23 were all acetylated to various degrees. For histone H4, K5, K8, K12, and K16 showed various degrees of acetylation, while H4 K20 was primarily dimethylated. These modifications are very similar to the epigenetic profile of an average cell. In summary, the HPV16/18 quasiviruses incorporate DNA complexed with histones with the most common modifications, similar in abundance to those of the 293TT packaging cells. Thus, there appears to be no preference for packaging chromatin with specific histone modifications in this system.
Histone modifications in BPV1 virions extracted from bovine fibropapillomas. The quasivirion analysis demonstrated that we can successfully quantify histone modifications in both virion and cell samples, and therefore, these methods can be used to compare histone modifications on more biologically relevant wart-derived PV virions with those of their corresponding host cells. Bovine papillomavirus virions were purified from three batches of bovine fibropapilloma tissue previously isolated from cows experimentally infected with bovine papillomavirus 1 (BPV1) (24). Virus was purified on OptiPrep gradients, and each fraction was tested for the presence of BPV L1 protein and histone H3 by immunoblot, BPV1 DNA by quantitative PCR (qPCR), and total protein by Sypro Ruby (Fig. 3). Fractions positive for L1, H3, and viral DNA were pooled. The virion samples were imaged by transmission electron microscopy, showing that they were mostly intact capsids with the expected size and morphology (Fig. 3).
Papillomaviruses assemble virions in the differentiated layers of the host epithelium and cause extensive changes in the pathology of the infected tissue; this makes it difficult to match perfectly the relevant host cells for histone extraction. Therefore, we generated three types of bovine keratinocyte cell and tissue samples to serve as controls and also to determine the effect of differentiation on host histone modifications. Primary bovine keratinocytes were cultured as a monolayer under normal proliferative conditions, induced to partially differentiate by the addition of high levels of calcium to the culture medium, or cultured as organotypic skin equivalents (rafts). The keratinocytes in the organotypic raft differentiate into all layers that are present in skin ( Fig. 3E and Fig. S2A).
Histone modifications in wart-derived BPV1 virions and each of the three keratinocyte culture conditions were identified by quantitative mass spectrometry, and the relative abundance of each modification was calculated (Fig. 4). Similar to the HPV quasiviruses, BPV1 capsids contained histones with many of the modifications detected in the cellular controls. However, the proportion of multiple modifications were significantly different between the virus and the three cellular controls (Fig. 4). The data for individual peptides with the greatest change are shown in detail in Fig. S3 for histone H3, and the raw data are shown in Table S2. To determine the individual modification status of each lysine residue in the H3 tail, the proportion of each modification was summed and displayed in the pie charts in Fig. 4B.
Histone H3 residue K4 can be acetylated or methylated, but .90% was unmodified in all three cell samples (Fig. 4, Fig. S3, Table S2). In contrast, ;50% of K4 residues were methylated (mostly K4me1) in the virion particle. H3K4ac was low in all samples but increased by nearly 6-fold in the virions, at ;0.6% compared to ;0.1% of the cellular controls. Both acetylation and methylation of H3K4 are usually associated with active  Table S1; raw data are shown in Table S1 for quasivirus versus host.

Viral Epigenome in Papillomavirus Virions
® chromatin (15). H3 residue K9 can also be acetylated or methylated; acetylation is found in active chromatin, and methylation in repressed chromatin (25). Similar to the activating H3K4 modification, H3K9ac was present, either alone or combinatorially with other modifications, at low levels in all cell samples (;2%) but was enriched in the virus (;9%). In contrast, ;44% H3K9 was methylated (K9me1, K9me2, or K9me3), alone or combinatorially with other modifications, in the virion, but about ;75% of cellular histones contained this repressive modification. Acetylation of H3K14, also associated with active chromatin, was found in ;65% of virion H3 peptides compared to ;47% on average in the cellular samples. Histone H3 peptide 18-26 contains two lysine residues (K18 and K23) that can be acetylated or methylated. Methylation of these residues was extremely low in all samples, but acetylation of both residues was much more abundant in the virion particles. The virions contained H3 peptides with ;22% K18ac and 54% K23ac, compared to the cell samples that contained, on average, ;6% and 28%, respectively.
H3 peptide 27-40 differs by a single amino acid, residue 31, between the histone variants H3.1 (Ala) and H3.3 (Ser; Fig. 4), and the key modified residues on these peptides are K27 and K36 (modified as K27ac, K27me1, K27me2, K27me3, and K36me1, K36me2, K36me3). The K27 acetylation mark is associated with active chromatin and di-and tri-methylation with repressed chromatin, while H3K36 methylation is most often found in transcriptionally active chromatin (26,27). Similar to H3K4, H3K27ac was very low in all samples but more abundant in the virus. Residues K27 and K36 were highly methylated in both H3.1 and H3.3 across all virion and cell samples. K27 was methylated in 82% of H3.1 and 77% of H3.3 peptides in virions compared to 86% of H3.1 and 67% of H3.3 peptides in the cell samples; in all samples, the most abundant peptide form was the single modification H3K27me2. Similarly, K36 was highly methylated in both virus and cells samples: 62% (H3.1) and 78% (H3.3) of residues were methylated in the virion compared to a wider range of 36 to 43% (H3.1) and 43 to 68% (H3.3) across the host cell samples. Overall, HK36 was mostly mono-and di-methylated, and overall, the virion was enriched in these two modified residues (Table S2; BPV versus host).
The proportion of the acetylation and methylation modifications on each lysine residue in the H3 tail is displayed in the pie charts in Fig. 4B for both the BPV1 virions and the bovine cellular controls (the mean of all three cellular controls). These epigenetic signatures highlight the major differences between the BPV1 virions and the cell controls. Overwhelmingly, the virion H3 peptides are enriched in active chromatin modifications (K4me1, K4me2, K4me3, K14ac, K18ac, K23ac, K36me1, K36me2) compared to all three host cell samples (Table S2).
Validation of BPV1 histone H3 modifications. To validate the mass spectrometry results, we performed immunoblot analysis using antibodies specific to each of the H3 modifications with the greatest enrichment in the BPV1 virions (human and bovine N-terminal histone sequences are identical). As shown in Fig. 5, the levels of H3K4me1,  Table S2. Viral Epigenome in Papillomavirus Virions ® H3K4me3, H3K9ac, H3K14ac, H3K18ac, and H3K23ac were confirmed to be higher in the virions than in cellular controls when the samples were normalized to the level of total histone H3. In agreement with the mass spectrometry data, the BPV virions were strongly depleted for H3K9me3, but not consistently for H3K27me3.
We also performed acid-urea (AU) PAGE to further confirm increased acetylation of histone H3 in the BPV virions. In AU PAGE, proteins are separated on the basis of size and charge. Acetylation of lysines reduces their positive charge, so histones with a higher degree of acetylation will migrate more slowly through the gel. This technique allows us to examine overall protein acetylation levels without the need for specific  Fig. 5, most of the signal from the histones extracted from BPV virions was in the more slowly migrating, upper bands, compared to the cellular lanes. This supports the mass spectrometry data and immunoblots and shows that H3 histones packaged in the virions are acetylated to a higher degree than those of the cellular controls.
BPV1 virions are enriched in acetylated histone H4. Mass spectrometry showed that the majority of H4 lysine residues K5, K8, K12, and K16 were unmodified in the virion and cellular samples. The BPV1 virions were enriched in acetylated K5, K8, and K12, as well as combinatorial acetylation of these lysines compared to each of the cellular controls (Fig. 6 and Fig. S4). In contrast, H4K16ac was depleted in the virions. This was notable since acetylation of K5, K8, and K12 is associated with active chromatin, while the K16ac modification is more complex. No consistent trend was found for the H4K20 methylation, with variation across the cell samples, although a small amount of H4K20ac was present in the virion but not in the cell samples.
To confirm the enrichment in lysine acetylation of H4 packaged in BPV virions, we performed immunoblots using antibodies specific to H4 acetylated on several different lysines. As shown in Fig. 7, the native BPV virions showed considerably higher levels of acetylation of K5ac, K8ac, and K12ac than those of the control keratinocytes. Additionally, the virion samples had higher signals of tetra-acetylated K5acK8acK12acK16ac. Together, these data suggest that bovine papillomavirus virions contain H4 histones that are more highly acetylated than host keratinocytes.
Histone modifications in wart-derived HPV1 virions. To extend these findings to human-specific virus, we were able to isolate sufficient quantities of HPV1 from a palmar wart to perform quantitative mass spectrometry. Primary human foreskin  Table S2.

Viral Epigenome in Papillomavirus Virions
® keratinocytes (HFKs) served as control host cells and were cultured in the three conditions described above: proliferating cells, cells partially differentiated with high calcium, and three-dimensional (3D) organotypic rafts (Fig. S2B). Histones were extracted from the HPV1 virions, and cellular controls and were analyzed for histone modifications and variants by quantitative mass spectrometry. As found in the BPV1 virions, the HPV1 virions were enriched in acetylated histone H3, primarily on lysines 14, 18, and 23 (Fig. 8, Fig. S5, Table S3). In all samples, H3K4 was primarily unmodified or monomethylated, and although H3K4me1 was increased in the virion compared to the raft sample, it was similar to the levels in growing and partially differentiated cells. As found in BPV1, the human virions were depleted for H3K9me1 and -me3 and contained a small amount of H3K9ac.
Histone H4 in the virions was enriched in acetylation on K5, K8, and K12 in comparison to the proliferating and calcium differentiating HFKs (Fig. 9A, Fig. S5, Table S3), but the amounts observed were similar to those of raft-derived histones. There was also less combinatorial acetylation of K5, K8, K12, and K16 in the virions compared to all cell samples, and overall acetylation of H4 K5, K8, K12, and K16 was less in the virion samples than in the average cell sample (Fig. 9B). As was observed for BPV1, virions were enriched for H4K20ac compared to all cellular samples, but there was no consistent pattern of enrichment in methylation patterns on K20.
Together, these data indicate that wart-derived human virions are enriched in histone modifications associated with active chromatin compared to human keratinocyte control cells. Correspondingly, the data indicate that these virions are depleted for transcriptionally repressive histone modifications. This is in agreement with the data from the wart-derived BPV virions.
Histone variant H3.3 is enriched in BPV1 virions. The histone variants H3.1 and H3.3 are deposited in chromatin in replication-dependent or replication-independent processes, respectively (28). Mass spectrometry can distinguish between these variants since there is a four-amino acid difference between these variants and so can give us insight into the assembly processes of viral chromatin. In the peptide spanning residues 27 to 40 of histone H3, this results in an alanine to serine difference at residue 31. As measured by mass spectrometry, the ratio of H3.3 to total H3 was 5-to 8-fold higher in BPV1 and 3-to 5-fold higher in the HPV1 virions than the cellular controls (  and B). In comparison, the HPV quasivirions showed no enrichment of H3.3 compared to the control cells (Fig. 10C). There was insufficient material to perform Western blot analysis of H3.3 abundance in HPV1 virions, but to confirm this enrichment in the BPV virions, we performed immunoblot analysis with antibodies specific to H3.3 ( Fig. 10D and E). The histones from the BPV virions had much greater levels of H3.3 compared to all three cellular samples (Fig. 10D). To confirm this finding, H3.1 and H3.3 proteins were separated based on hydrophobicity using triton acid urea (TAU) PAGE and immunoblotted with an antibody against all variants of H3 ( Fig. 10F and G). This confirmed that about 40% of H3 in the BPV particles was H3.3, very similar to the quantitation by mass spectrometry and much higher than the cellular controls. Therefore, the chromatin in papillomavirus virions is enriched in the histone variant H3.3.

DISCUSSION
In this study, we used mass spectrometry to profile the histones of HPV16/18 quasivirions and native BPV1 and HPV1 virions extracted from warts. We demonstrate that the histone profiles of wart-derived virions are distinct from that of the host cell, suggesting that viral chromatin is assembled and/or modified independently from the cellular chromatin. In contrast, the HPV16/18 quasiviruses contained histone modifications similar to those found in the packaging cells, implying that the capsid did not exclude modified chromatin on the basis of space or chemical compatibility. This was  Table S3.

Viral Epigenome in Papillomavirus Virions
® not surprising as, due to the promiscuous nature of L1 and L2 capsid assembly, the majority of chromatin packaged by quasiviruses prepared using the Benzonase method is derived from cellular genomic DNA (29). Therefore, most likely, many of the histone modifications we detect in quasiviruses are derived from chromatin containing cellular DNA rather than viral minichromosomes. Nevertheless, these findings provide a good control for our methodology and show that the enhanced histone modifications found in the BPV1 and HPV1 virions are not artifactual. Quasiviruses are a useful model system for early HPV infection, and since quasivirion chromatin mirrors that of the host cells, treatment of packaging cells with small molecule modulators of chromatin-modifying enzymes is a powerful tool to examine the roles of specific modifications in infection studies (20).
The wart-derived HPV1 and BPV1 virions contained histones that were strongly enriched in acetylation of H3 and H4 histone tails. This acetylation could cause the viral minichromosome to be less compact but also provide binding sites for various transcriptional activators. Notably, we have shown that the double bromodomain protein, Brd4, activates early HPV transcription (20), so these modifications could ensure that the viral genome is primed for transcriptional activation by Brd4 immediately after infection. An enrichment of acetylated histones has also been reported in the capsids of the polyomaviruses, SV40, BKV, and mouse polyomavirus (30,31). We also observed enrichment of H3K4me1, particularly in the bovine virions. H3K4me1 is a mark of cellular enhancer regions, and we have previously shown that H3K4me1 is present throughout the chromatin in HPV replication foci that form in differentiated cells (32). BPV1  Table S3. Porter et al. ® particles also showed an enrichment for H3K4me3, a hallmark of actively transcribing genomic regions, and H3K4me2, which is associated with permissive euchromatin. Conversely, H3K9me1, me2, and me3, hallmarks of repressive chromatin, were depleted in the BPV1 and HPV1 virions. The strategy of packaging the viral genome in a chromatin state primed for transcription is likely advantageous for a small virus.
Papillomaviruses also promote various aspects of their life cycle by modulating host epigenetic mechanisms (33), which induces numerous epigenetic changes in host chromatin. Viral proteins bind to and manipulate several host epigenetic modulating enzymes. For example, the E7 protein binds to histone deacetylases (HDACs) and causes a global increase in acetylation (34)(35)(36). E7 also reprograms host cells by activating the lysine demethylases KDM6A/B and thus decreasing overall levels of H3K27 trimethylation (35). Both of these trends were observed in the chromatin from wartderived virions, and it seems likely that viral manipulation of host epigenetic processes also shapes the chromatin in the virion particles.
Mass spectrometry can distinguish between the histone variants H3.1 and H3.3 (37). The H3.1 variant is known to be deposited on DNA in a replication-dependent manner, while H3.3 deposition is associated with transcription and DNA repair (38). We found that H3.3 is enriched several fold in the wart-derived virions compared to the cellular samples. This enrichment could reflect the transcriptionally active nature of the viral DNA before assembly or could reflect the mechanism of DNA replication at late stages of the viral infectious cycle. Papillomaviruses amplify viral DNA in differentiated cells in a G2-like phase of the cell cycle; this is dependent on host DNA damage machinery (39), and likely uses a recombination-directed replication mechanism (40) that does not utilize H3.1. The H3.3 variant is also required for transcription after DNA damage which could promote late viral transcription from amplified DNA templates (41). Several additional modifications enriched in the virions (H3K14ac, H3K18ac, H3K23ac, H4K8ac, H4K12ac, H4K16ac) are linked to DNA damage repair (42,43). Furthermore, acetyltransferases such as Tip60 aid the loading of DNA repair proteins by relaxing the surrounding chromatin; Tip60 modulates loading of repair proteins and repair of DNA double-strand breaks (44). Thus, an alternative explanation for enrichment of acetylated H3 and H4 could be that these modifications are related to the unique replication mechanism of papillomaviruses at the late stages of infection and may also facilitate the initial replication of the viral genome upon infection of a new host. This is consistent with our understanding of how papillomaviruses hijack specific epigenetic processes for viral replication and transcription throughout the infectious cycle and how they also modulate the host epigenome (1). Encapsidated viral chromatin confers other advantages to the virus. Long stretches of naked DNA signal the cell that the nucleic acid is foreign and activate the cGAS-STING pathway (45). Trafficking of HPV particles to the nucleus in vesicles helps the virus to evade cGAS/STING sensing (46), but assembly of DNA in chromatin also helps evade detection (45). Our study did not detect histone H1 in virions because we size selected for the smaller canonical histones, but chromatin without the linker histone H1 activates cGAS more efficiently (47). Recent studies have determined the structural mechanism of histone octamer-dependent inhibition of cGAS (48)(49)(50)(51)(52) and shown that cGAS interacts with the acidic patch on H2A-H2B heterodimers within adjacent nucleosomes (53); the N-terminal tail of H4 also interacts with this acidic patch to regulate nucleosome stacking and chromatin compaction (47,54). Therefore, hypothetically, viral minichromosomes with certain histone modifications could more efficiently bypass and evade innate immune surveillance.
The minor capsid protein L2 facilitates genome packaging, but no specific DNA packaging sequence has been identified (55). Could histone modifications on viral chromatin form a packaging signal for viral minichromosomes? On one hand, this seems unlikely because of the permissive nature of chromatin within the quasivirus particles (although this is mostly cellular chromatin), but on the other hand, assembly of authentic papillomavirus particles could take place in concert with specific replication processes and histone modifications in differentiated keratinocytes. The L2 protein localizes to viral replication foci at late stages of the papillomavirus life cycle (56,57), and in early infection, L2 traffics the viral DNA first to mitotic chromosomes and then to ND10 bodies to establish infection (58,59). Interaction of L2 with specifically modified viral chromatin could be key to these processes.
In conclusion, we propose that the viral chromatin acquires specific histone modifications late in infection that are coupled to the mechanisms of viral replication, late gene expression, and packaging. We predict that, in turn, these same modifications promote early infection by evading detection, promoting localization of the viral chromosome to beneficial regions of the nucleus, and facilitating early transcription and replication.
Differentiated (Ca 11 ) keratinocytes were as follows: cells were grown to confluence, the feeders removed with Versene, and the medium was changed to low-calcium basal medium (Lonza) supplemented with SingleQuots for keratinocytes (bovine pituitary extract, hydrocortisone, and epidermal growth factor) for 24 h. The medium was changed to basal medium supplemented with 1.5 mM CaCl 2 , and cells cultured for 5 additional days. Cells were collected by trypsinization, pelleted, flash-frozen, and stored at -80°C until protein extraction.
Organotypic raft cultures were as follows: a collagen dermal equivalent was formed by mixing 0.8 ml each of reconstitution buffer (2.2% NaHCO 3 , 0.05 N NaOH, 200 mM HEPES free acid) and 10Â Ham's F12 medium (Life Technologies) with 7.6 ml rat tail collagen type I (Sigma). Two million NIH 3T3 mouse fibroblasts (ATCC 1658) were suspended in 0.4 ml FBS and mixed with the collagen solution, and 0.75 ml collagen/cell suspension was aliquoted per well of a 12-well tissue culture plate. After 1 h of incubation at 37°C and 5% CO 2, the solidified gel was overlaid with 1 ml raft culture medium (3:1 DMEM/F12, 10% fetal calf serum [FCS], 0.4 mg/ml hydrocortisone, 0.01 nM cholera toxin, 5 mg/ml transferrin) and incubated at 37°C for 2 days. Then, 1 Â 10 5 keratinocytes were plated on each collagen disk and cultured in raft culture medium supplemented with 5 ng/ml epidermal growth factor until confluent. The collagen gels were transferred to the liquid-air interface on a metal grid and cultured for 10 days in raft culture medium with 5 ng/ml EGF. To harvest, each raft was bisected and either fixed in 3.7% formaldehyde before paraffin embedding, sectioning, and hematoxylin and eosin staining, or else the epithelial layer was removed and frozen at -80°C for subsequent protein extraction.
Production of HPV16 quasiviruses. HPV16/18 quasiviruses were produced in 293TT cells as described in reference 61. Briefly, a 75-cm 2 flask of 293TT cells was transfected with 19 mg minicircle HPV18 genome, 19 mg pShell16 L1/L2 expression vector, and 6 mg each of pMEP9-HPV18 E1 and pMEP4-HPV18 E2 using Lipofectamine 2000 (Invitrogen). Then, 48 h posttransfection, quasivirus particles were isolated in the presence of Benzonase and purified on an OptiPrep density gradient. Fractions were analyzed for total protein content using Sypro Ruby and for HPV18 DNA using real-time PCR on a QuantStudio 7 PCR system and SYBR green PCR master mix (Roche). A 10-fold dilution series was included to generate a standard curve of cycle threshold versus log 10 quantity. PCR was performed in triplicate using forward 59-CACAATACTATGGCGCGCTTT-39 and reverse 59-CCGTGCACAGATCAGGTAGCT-39 primers, as described previously (62). Fractions that contained L1 and L2 proteins, histones, and HPV18 DNA were selected and pooled. Viral genome equivalents (VGE) and number of virion particles were calculated as previously described (61).
Isolation of bovine papillomavirus virions. Bovine wart tissue (3 to 5 g) (harvested from BPV1infected cows by Carl Olson, University of Wisconsin School of Veterinary Medicine; a gift from Carl C. Baker) (24) was minced and vortexed in 10 mM Tris-HCl, pH 7.5, 1 mM MgCl 2 , and 1% (wt/vol) Brij58 in PBS. Samples were digested with 75 U Benzonase (Millipore) and 50 U Plasmid-Safe (Lucigen) at 37°C for 1 h. Then, 2 mg collagenase H (Sigma) was added, the pH was adjusted to 7 to 7.5, and samples were incubated at 37°C for 15 min followed by 4°C incubation overnight on a rotator. The next day, 0.17 volumes of 5 M NaCl were added, and the tissue was pelleted by centrifugation at 1,000 Â g for 5 min. The supernatant was reserved, and the pellet was suspended in salt extraction buffer (10 mM Tris-HCl, pH 7.5, 800 mM NaCl, 1% Brij58, 1% PBS) and sonicated until the solution no longer changed in consistency. Debris was pelleted by centrifugation at 1,000 Â g for 5 min, and the supernatant was combined with the previous supernatant. After another clarification step, the supernatant was underlaid with a 1.5-ml cushion of 39% OptiPrep and centrifuged in a Sw32Ti rotor at 30,000 rpm for 2 h at 16°C. The bottom 3 ml was vortexed and centrifuged at 1,000 Â g for 10 min at 4°C. The supernatant was separated on an OptiPrep density gradient, and fractions were screened for the presence of BPV1 DNA by qPCR using forward 59-TTGGTGAGGACAAGCTACAAGTTG-39 and reverse 59-TTGGTGAGGACAAGCTACAAGTTG-39 primers. Fractions that contained L1 and L2 proteins, histone H3 (all identified by immunoblotting), and BPV1 DNA were combined, aliquoted, and stored at -80°C.
Isolation of human papillomavirus virions from wart tissue. A 0.65-cm-diameter biopsy specimen of a palmar wart was minced in 0.5 ml Dulbecco's PBS (DPBS), 10 mM MgCl 2 , and 1% Brij58 and digested with 50 U Benzonase and 20 U Plasmid-Safe for 20 min at 37°C. Then, 5 mg collagenase H was added and incubated for 30 min at 37°C and then overnight at 4°C. Next, 180 ml 5 M NaCl was added, and the sample was rocked at 4°C for 2 h. The sample was sonicated, centrifuged at 5,000 Â g for 5 min, and the supernatant was transferred to a new tube. The pellet was washed with 0.5 ml DPBS 1 0.8 M NaCl before centrifuging was done again. The supernatants were combined, and then virions were purified by OptiPrep ultracentrifugation (as described for quasiviruses). Viral DNA was extracted, analyzed with restriction digest, and cloned into a plasmid vector. Sequencing revealed it to be HPV1.
Mass spectrometry. Virus histone extraction. BPV1, HPV1, or HPV16/18 quasiviruses (;1.4 Â 10 9 VGE) were precipitated in 33% trichloroacetic acid (TCA) overnight at 4°C. The precipitate was centrifuged at 20,000 Â g for 10 min at 4°C, and the supernatant was removed. The pellet and tube walls were washed with 1 ml cold acetone 1 0.1% HCl and centrifuged at 20,000 Â g at 4°C for 10 min, and the supernatant was removed. The pellet and tube walls were washed with 1 ml cold acetone and centrifuged at 20,000 Â g at 4°C for 10 min, and the supernatant was removed. The pellet was dried overnight in a SpeedVac on low vacuum with no heat. The pellet was dissolved in 60 ml 54 mM Tris-HCl, pH 8, 1Â LDS sample buffer (Invitrogen), and 50 mM dithiothreitol (DTT) and heated at 70°C for 10 min. Samples were centrifuged at 16,000 Â g for 1 min, and the supernatant was separated on three adjacent wells of a NuPAGE 4 to 12% SDS-PAGE gel (Thermo Fisher) by electrophoresis at 100 V. Proteins were detected with Coomassie G-250 staining (1% Coomassie G-250, 2.55% ortho-phosphoric acid, 10% [wt/vol] ammonium sulfate, 20% ethanol), and bands corresponding to the molecular weight of the core histones (molecular weight, 10 to 20 kDa) were excised from the gel and stored at -80°C.
Gel band processing. Excised gel bands were diced into ;1-mm cubes and destained with successive washes of liquid chromatography-mass spectrometry (LC-MS)-grade water and 50% acetonitrile until no dye was visible (3 to 4 wash cycles). Gel pieces were fully dehydrated in acetonitrile, rehydrated with a 1:2 (vol/vol) solution of 100 mM ammonium bicarbonate and propionic anhydride (Sigma; catalog no. 240311-60G), and incubated at room temperature for 20 min. The supernatant was removed, and gel pieces were washed with alternating rounds of 50% acetonitrile and 100 mM ammonium bicarbonate until the supernatant pH was ;8.0. After a second round of propionylation and pH adjustment, gel pieces were dehydrated in 100% acetonitrile and rehydrated in 50 mM ammonium bicarbonate containing 12.5 ng/ml trypsin (Promega; catalog no. V5113). Digestion proceeded overnight at room temperature. Peptides were extracted from the gel with two rounds of 100% acetonitrile dehydration and water rehydration. Extracted peptides were propionylated for two additional rounds with 1:3 (vol/vol) propionic anhydride:2-propanol and adjusted to pH ;8 with ammonium hydroxide. Fully derivatized samples were desalted over Oligo R3 reverse-phase resin (Thermo Scientific) packed over a C18 Empore base (3M), dried to completion, and resuspended in buffer A (water 1 0.1% [vol/vol] formic acid).
Organotypic raft controls were processed similarly, but with an initial homogenization step in a ground glass pestle to enable effective extraction.
Histone derivatization. After nuclear extraction, histone samples were resuspended in 50 mM ammonium bicarbonate and propionylated for two rounds by addition of 16% (vol/vol) propionic anhydride and 14% (vol/vol) ammonium hydroxide for 10 min at room temperature. Samples were digested with 0.2 mg trypsin at room temperature overnight and then propionylated for another two rounds. Fully derivatized samples were desalted over Oligo R3 reverse-phase resin (Thermo Scientific) packed over a C18 Empore base (3M), dried to completion, and resuspended in buffer A.
Mass spectrometry acquisition. All solvents used in the analysis of MS samples were LC-MS grade. All peptides were separated over a 75 mm inner diameter silica capillary column packed in-house with Repro-Sil Pur C18-AQ 3 mm resin.
HPV16 quasivirion samples and associated 293TT controls were analyzed with an Easy-nLC system (Thermo Fisher) running 0.1% (vol/vol) formic acid (buffer A) and 80% (vol/vol) acetonitrile with 0.1% (vol/vol) formic acid (buffer B), coupled to an Orbitrap Fusion Tribrid mass spectrometer. Peptides were eluted with a gradient of 4 to 38% buffer B over 45 min. Full MS scans from 300 to 1,100 m/z were analyzed in the Orbitrap device at 60,000 full-width half maximum (FWHM) resolution and 2 Â 10 5 automatic gain control (AGC) target value for 100 ms maximum injection time (MIT). MS2 analyses were performed in data-independent acquisition (DIA) mode with 50 m/z isolation windows. Spectra were acquired in the ion trap operating in rapid mode. High-energy collisional dissociation (HCD) fragmentation was applied at 28%, with a stepped collision energy of 65%. The AGC target was set to 1 Â 10 4 , and MIT, to 50 ms.
BPV virion samples and BEK6 controls were analyzed with an Easy-nLC system (Thermo Fisher) running 0.1% (vol/vol) formic acid (buffer A) and 80% (vol/vol) acetonitrile with 0.1% (vol/vol) formic acid (buffer B), coupled to an Orbitrap Fusion Tribrid mass spectrometer. Peptides were eluted with a gradient of 5 to 34% buffer B over 45 min. Full MS scans from 300 to 1,100 m/z were analyzed in the Orbitrap Immunodetection. For immunoblots probing for histone modifications, protein samples could not be loaded based on equal protein amounts. Therefore, a series of dilutions of each viral and cellular protein preparation was first compared until the signal for total H3 protein was in a comparable range. The optimized amounts of viral preparation and cellular extracts were analyzed by immunoblotting for histone modifications. Membranes were stained with Ponceau S (Sigma) to ensure uniform loading and transfer and blocked in 5% skim milk in TBST (Tris-buffered saline/0.1% Tween) (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.1% Tween) for 1 h. Primary antibodies were diluted in 5% milk/TBST and incubated overnight at 4°C . Horseradish peroxidase conjugated secondary antibodies (Invitrogen) were used at a dilution of 1:10,000. Proteins were detected with SuperSignal West Dura extended duration substrate (Thermo Fisher), and chemiluminescent signals were captured using a G:Box (Syngene) and quantitated using Syngene GeneTools software. The blots were stripped after imaging with One Minute Plus stripping buffer (GM Biosciences) and reprobed with the pan-histone H3 antibody to check for even loading. The histone modification signals were normalized to the total histone H3 signals, and the levels were calculated relative to the normalized signal for BPV1. P values were calculated by two-tailed, equal variance t tests.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.