Genus Beta Human Papillomavirus E6 Proteins Vary in Their Effects on the Transactivation of p53 Target Genes

ABSTRACT

The genus beta human papillomaviruses (beta HPVs) cause cutaneous lesions and are thought to be involved in the initiation of some nonmelanoma skin cancers (NMSCs), particularly in patients with the genetic disorder epidermodysplasia verruciformis (EV). We have previously reported that at least two of the genus beta HPV E6 proteins bind to and/or increase the steady-state levels of p53 in squamous epithelial cells. This is in contrast to a well-characterized ability of the E6 proteins of cancer-associated HPVs of genus alpha HPV, which inactivate p53 by targeting its ubiquitin-mediated proteolysis. In this study, we have investigated the ability of genus beta E6 proteins from eight different HPV types to block the transactivation of p53 target genes following DNA damage. We find that the E6 proteins from diverse beta HPV species and types vary in their capacity to block the induction of MDM2, p21, and proapoptotic genes after genotoxic stress. We conclude that some genus beta HPV E6 proteins inhibit at least some p53 target genes, although perhaps not by the same mechanism or to the same degree as the high-risk genus alpha HPV E6 proteins.

IMPORTANCE This study addresses the ability of various human papillomavirus E6 proteins to block the activation of p53-responsive cellular genes following DNA damage in human keratinocytes, the normal host cell for HPVs. The E6 proteins encoded by the high-risk, cancer-associated HPV types of genus alpha HPV have a well-established activity to target p53 degradation and thereby inhibit the response to DNA damage. In this study, we have investigated the ability of genus beta HPV E6 proteins from eight different HPV types to block the ability of p53 to transactivate downstream genes following DNA damage. We find that some, but not all, genus beta HPV E6 proteins can block the transactivation of some p53 target genes. This differential response to DNA damage furthers the understanding of cutaneous HPV biology and may help to explain the potential connection between some beta HPVs and cancer.

INTRODUCTION

Human papillomaviruses (HPVs) are DNA viruses with small double-stranded DNA genomes and a tropism for squamous epithelial cells. Different HPVs have been associated with various neoplastic diseases: they are the established cause of cervical cancer and have been associated with other anogenital cancers, they are responsible for an increasing proportion of head and neck oropharyngeal cancers, and they cause hyperproliferative lesions in patients with the genetic disorder epidermodysplasia verruciformis (EV) (1). Most HPV infections are benign, and the vast majority of HPV infections resolve without progression to cancer. This diversity of pathology is related to and reflected by the phylogenetic relationships between HPV types.

The more than 178 HPVs now identified are grouped on the basis of the sequence of their L1 gene into five genera (2, 3). Genus alpha human papillomaviruses (alpha HPVs) infect the mucosal epithelium, and a subset of these are associated with cancer. The “high-risk” virus types that cause cervical, anogenital, and head and neck cancers are found in species 7 and species 9 of genus alpha HPV. “Low-risk” genus alpha types cause genital warts and include the viruses in species 10. Genus beta virus types infect cutaneous epithelial cells. Human papillomavirus 5 (HPV5) and HPV8 (both genus beta HPV species 1) are the two types most frequently found in lesions of EV patients, but many other beta HPV types have been identified in such lesions (4).

Genus alpha HPVs encode three viral oncoproteins: E5, E6, and E7. The E6 and E7 proteins encoded by high-risk HPVs are sufficient to immortalize and/or transform human keratinocytes (5, 6) and consequently are thought to be the primary drivers of HPV-related cancers. The best understood function of E7 is its ability to bind and inactivate pRB1 (7–9), which allows cells in the differentiating epithelium to progress into an unscheduled S phase conducive to the replication of the viral DNA. pRB1 inactivation also triggers cell cycle control checkpoints, and one mechanism by which high-risk HPV oncoproteins inactivate these checkpoints is the targeting of the tumor suppressor p53 (10) for proteasome-mediated degradation (11). High-risk HPV E6 recruits the cellular ubiquitin ligase E6AP to ubiquitinate p53 in this reaction (10, 11), and critically, it is only the high-risk genus alpha HPV E6 that can bind both p53 and E6AP (12, 13). In cervical cancer cells, p53 levels are quite low due to E6-mediated proteolysis. Several studies have shown that the small amount of p53 remaining in these cells does retain some ability to increase p21 expression, the expression of other target genes, and translocate to the nucleus following DNA damage (14–16). Other studies have proposed mechanisms by which residual p53 might be inactivated via other functions of E6, largely through an interaction with CBP/p300 (17–19). The experiments in those reports use in vitro techniques or transient transfection of E6 expression constructs driven by the cytomegalovirus (CMV) promoter, possibly resulting in higher levels of E6 than in a normal infection. Overall, published reports do not agree on whether the p53 that remains in the presence of high-risk E6 is ever functional during a natural HPV infection or in its associated cancer.

p53 is a central factor in cellular sensing of genotoxic stress, including the stress resulting from DNA damage (reviewed in reference 20). Multiple pathways can be triggered by such stress and the subsequent activation of p53, including a paused cell cycle followed by DNA repair, senescence, or apoptosis. An important cellular negative regulator of p53 is MDM2, and the two proteins participate in a feedback loop in which p53 transactivates MDM2 expression via a p53 binding site in its promoter. Through its activity as an ubiquitin ligase MDM2 in turn downregulates p53 by targeting its ubiquitylation and proteolysis. The inactivation of p53 by high-risk E6 is one of the major mechanisms by which HPVs contribute directly to cancer development.

In addition to studies on the genus alpha HPVs, the idea that at least some of the genus beta HPVs might have some role in nonmelanoma skin cancers (NMSCs) has been a topic of some interest in recent years (21–24). The lesions that form in sun-exposed regions on the skin of EV patients make a clear link between beta HPVs and increased cellular proliferation; however, the contribution of the EVER1 or EVER2 mutations present in those patients' genomes may be mechanistically important as well (4, 25). The beta HPVs are ubiquitous. Their DNAs are frequently detected in NMSC but are very often on healthy skin (22), and recent high-throughput sequencing studies have not identified HPV transcription in squamous cell carcinomas (SCC) (26, 27). However, multiple reports link beta HPV E6 proteins (E6s) to an impaired DNA damage response through the ability of E6 to degrade Bak (28, 29), and several studies suggest that genus beta HPV E6 proteins block the ability of cells to repair DNA damage after UV irradiation (30, 31). These observations support the notion that some beta HPVs might contribute to SCC initiation but not be required for tumor maintenance.

Beta HPVs are numerous and diverse, and no study has been able to examine all of the beta HPVs. A few studies have compared various beta HPV types, particularly with respect to the ability of their E6 and E7 proteins to transform rodent cells or to immortalize human foreskin keratinocytes (HFKs) (32, 33). Selected beta HPVs have been characterized in more detail with respect to their effects on p53, including HPV38 and HPV23. HFKs that express both HPV38 E6 and E7 proteins have an extended life span and stabilize p53 (34–36), but the cells were analyzed only in the absence of genotoxic stress, and under those conditions, there is no increase in expression of p53-regulated genes even though the level of p53 is increased. HPV23 E6 may alter p53 posttranslational modifications in a way that would block the expression of proapoptotic genes (37). Our own studies recently showed that at least two of the beta HPV E6 proteins, encoded by HPV38 and HPV92, can interact with p53, and that these two as well as HPV17a E6 can stabilize p53 (13). The steady-state level of p53 in cells expressing any one of these three HPV E6s is increased compared with E6-negative cells or with cells expressing a high-risk genus alpha HPV E6. This is not unique; p53 was first identified as a protein that binds to and is stabilized by its interaction with simian virus 40 (SV40) large T antigen (38, 39), and subsequently, adenovirus 5 (Ad5) E1B was also shown to bind p53 (40). Similar to the case for HPV E6 proteins from different types, Ad5 E1B binds and stabilizes p53, while Ad12 E1B does not bind but does stabilize p53 (41). p53 is not a target restricted to the tumor viruses, and a number of other viruses variously increase or decrease its expression and restrict or alter its targets for a variety of purposes (reviewed in reference 42).

Studies from a number of laboratories have suggested that E6 might target p53 through a variety of mechanisms. None, though, have systematically evaluated a panel of beta HPV types to assess whether p53 in the E6-positive cells remains functional to transactivate the expression of its target genes, including MDM2, p21, and proapoptotic genes. We therefore tested the functional status of p53 in genus beta HPV E6-expressing cells by examining whether E6 blocked the transactivation of p53 target genes. We examined eight different genus beta HPV E6 proteins and compared them to nine E6s from genus alpha HPV: five high-risk E6s known to functionally inactivate p53 and four low-risk or non-cancer-associated E6s that do not bind or degrade p53. After introducing the genes encoding these E6 proteins into human keratinocytes, we characterized the levels of p53 protein and RNA in cells, both before and after the induction of DNA damage. We measured protein and RNA levels for the p53 targets p21^WAF1 and MDM2 as well as proapoptotic genes BAX, TP53I3, and FAS. As anticipated, each high-risk genus alpha HPV E6 blocked the expression of these cellular genes after DNA damage. The E6s tested from genus beta HPV species 2, 3, and 4 affected p53 targets to different degrees. HPV76 E6 was uniquely able to block the induction of p53 and MDM2 after DNA damage at the protein level, and HPV17a, HPV38, and HPV92 E6 proteins blocked the induction of p21 RNA after DNA damage. In contrast, p53 target expression was not affected by the expression of other genus alpha HPV E6s or the genus beta HPV species 1 E6s in our study. On the basis of these results, we propose that genus beta HPV E6 proteins vary in their abilities to block the transcriptional activation functions of p53. In general, the high-risk alpha HPV E6 proteins affected p53 targets more broadly than did any single beta HPV E6 that we tested. Nevertheless, the effect of beta HPV E6s on p53 targets may have important consequences for the interaction of these viruses with host cells.

MATERIALS AND METHODS

Plasmids and cloning.

HPV11 E6 was PCR amplified from pGEM1-11E6 (10) and recombined into pDONR223. pDONR-11E6 (Howley laboratory plasmid number p7268) and pDONR-GFP (GFP stands for green fluorescent protein) (p6812) (gift of J. Wade Harper, Harvard Medical School) were both then recombined into pMSCV-N-HA-IRES-PURO GAW (IRES stands for internal ribosome entry site, and PURO stands for puromycin) (p6804) as previously described (43) to produce pMSCV-IP-N-HA 11E6 (p7267) and pMSCV-IP-N-HA GFP (p6575). The pMSCV-N-HA-IRES PURO EMPTY vector (p7270) was generated by digesting pMSCV-N-HA-IRES-PURO GAW (13) with BamHI and EcoRI to remove the chloramphenicol resistance and ccdB markers. The resulting 6.5-kbp fragment was end filled with Klenow fragment and ligated with T4 DNA ligase to produce a vector that contains the retroviral packaging sequence-HA tag-IRES-puromycin resistance cassette between retroviral long terminal repeats (LTRs). To produce E6 proteins without the hemagglutinin (HA) tag, pDONR-E6 vectors were generated as previously described (43), this time using primers that incorporated a 5′ Kozak sequence and methionine and a 3′ stop codon into the E6 open reading frame (ORF). Each E6 ORF was then recombined into pMSCV-P-C-HA GAW (p7353). The epitope tag encoded by the resulting expression vectors is located downstream of the stop codon at the 3′ end of the E6 ORF and is therefore not translated. Plasmids used in the study are listed in Data Set S1 in the supplemental material. All plasmids from the Howley laboratory described in this study are available through Addgene.

Tissue culture and drug treatments.

Three new N/Tert-HA-11E6, N/Tert-HA-GFP, and N/Tert-empty vector cell lines were generated as previously described and were propagated along with the 16 existing N/Tert-HA-E6 cell lines as before (13, 43). Both N/Tert-1 cells and G5-Ep primary human foreskin keratinocytes (generously provided by James Rheinwald [44–46]) were transduced with retroviruses generated from the pMSCV-E6-untagged vectors as well as previously described control vectors to generate seven new N/Tert-untagged E6 cell lines, seven new G5-Ep-untagged E6 cell lines, a G5-Ep-HA-GFP cell line, and a G5-Ep-empty vector cell line.

The cells were grown to approximately 30% confluence in keratinocyte serum-free medium (K-SFM) (Invitrogen) prior to harvest by trypsinization and freezing at −80°C for subsequent Western blotting and RNA isolation experiments. For drug treatments, mitomycin C (MMC) (Sigma-Aldrich) was resuspended in sterile water and diluted in K-SFM to the final concentrations indicated. The cells were refed with K-SFM containing mitomycin C and harvested 18 h after treatment. Neocarzinostatin (NCS) stock solution (Sigma-Aldrich) was used in the same way during a 4-h treatment. For UV irradiation, the cells were washed with phosphate-buffered saline (PBS) and exposed to 15-mJ/cm² UV light from a 306-nm UV-B bulb. The cells were then refed with fresh K-SFM and harvested 6 h after treatment.

Western blotting.

Cell pellets were lysed in 1× reducing sample buffer (50 mM Tris [pH 6.8], 0.2% sodium dodecyl sulfate, 10% glycerol, 5% 2-mercaptoethanol, 25 mM sodium fluoride, 1 mM sodium orthovanadate, 5 mM β-glycerophosphate, 1 mM phenylmethylsulfonyl fluoride, 50 μM leupeptin, and 100 μM pepstatin A) and quantified by Bradford assay. Equal amounts of protein were separated on NuPAGE (Invitrogen) gels and transferred onto polyvinylidene difluoride (PVDF) membranes. After blocking in 5% nonfat dried milk in TBS-T (Tris-buffered saline [pH 7.4] with 0.05% Tween 20), the blots were incubated with primary antibodies as follows: actin (Millipore), p53 (Santa Cruz Biotechnology), MDM2 (Millipore), p21 (LabVision/ThermoFisher), phospho-p53 (Ser15) [P-p53(Ser15)] (Cell Signaling Technology), P-p53(Ser392) (Cell Signaling Technology), and acetyl-p53(Lys382) (Cell Signaling Technology). The membranes were washed in TBS-T and incubated with horseradish peroxidase (HRP)-coupled anti-mouse or anti-rabbit antibodies or an Alexa Fluor 680-coupled anti-mouse antibody and detected using Western Lightning chemiluminescent substrate or a Li-COR infrared imaging system. HA-tagged proteins were detected using an HA antibody conjugated to HRP (Roche) and visualized on film.

RNA isolation and qRT-PCR.

Total RNA was isolated from cell pellets using the RNeasy minikit (Qiagen) or the Nucleospin RNA kit (Clontech) according to the manufacturer's instructions. Equal amounts of RNA were reverse transcribed using the high-capacity cDNA reverse transcription kit (Applied Biosystems). Quantitative real-time PCR (qRT-PCR) was performed in an Applied Biosystems ABI 7500 fast sequence detection system using TaqMan fast universal PCR master mix or TaqMan fast advanced master mix (Invitrogen) and TaqMan gene expression assays (Invitrogen) for the following genes: MDM2, p53, BAX, TP53I3/PIG3, FAS, and p21 (see Data Set S2 in the supplemental material). Sequences for the custom glucose-6-phosphate dehydrogenase (G6PD) primers and TaqMan probe (Integrated DNA Technologies, Coralville, IA) have been previously published (47). The relative amount of cDNA in each sample was calculated based on a standard curve prepared using serial dilutions of one reference cDNA. These were normalized to the values obtained by analyzing the same samples with primers and TaqMan probe specific to the cellular housekeeping gene G6PD. The fold change of transcription of the gene in a sample is displayed relative to the empty-vector-containing cell control sample.

Affymetrix arrays and analysis.

Total RNA was isolated from N/Tert-control and N/Tert-HA-E6 cells that were treated with 2 μg/ml mitomycin C for 18 h (or untreated controls) using the RNeasy minikit (Qiagen) as described above. The RNA concentration was measured by a NanoDrop spectrophotometer, and the RNA integrity was checked by a 2100 bioanalyzer (Agilent). The GeneChip 3′ IVT Express kit (Affymetrix) and a GeneAmp PCR system 9700 (Applied Biosystems) were used to synthesize the first-strand cDNA by reverse transcription, followed by synthesis of second-strand cDNA and subsequently biotin-labeled amplified RNA (aRNA). Hybridization was performed by injecting the hybridization cocktail containing fragmented aRNA into the U133 plus 2.0 human genome microarrays (Affymetrix) and incubating at 45°C for about 16 h with 60 rpm rotation in a GeneChip hybridization oven 640 (Affymetrix). Staining and washing were performed in a GeneChip fluidic station 450 (Affymetrix). Array images were acquired by scanning the arrays using a GeneChip 7G Plus scanner (Affymetrix). Raw data (.cel files) from individual gene chips were compiled and normalized using Affymetrix power tools using the standard RMA work flow (background adjustment, quantile normalization, and median polish probe set summarization) and then filtered to exclude expression values of <80. The GENE-E module of GenePattern was used to collapse probe set data to genes, to select the 1,000 most differentially expressed genes from the combined data set, and to perform hierarchical clustering of the remaining data based on both the genes and the samples. The Functional Annotation Tool in DAVID (Database for Annotation, Visualization and Integrated Discovery) (http://david.abcc.ncifcrf.gov) was used to assign functional identities to subsets of the genes identified using GENE-E.

Microarray data accession number.

The data discussed in this publication have been deposited in NCBI's Gene Expression Omnibus (48) and are accessible through GEO Series accession number GSE50568 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE50568).

RESULTS

High-risk alpha HPV genus E6 proteins and HPV76 E6 inhibit the increase of p53 and MDM2 protein levels following DNA damage.

Our initial studies to determine the effects of genus beta HPV E6 proteins on p53 activity utilized the N/Tert-1 human keratinocyte cell lines that were previously used in our E6 proteomic studies (13). These cell lines stably express one of 17 different E6 proteins from various HPV types: five high-risk HPVs from genus alpha HPV (HPV16, -33, -52, -18, and -45), two low-risk HPVs from genus alpha (HPV6b and -11), two non-cancer-associated genus alpha HPVs (HPV2a and -57), and eight genus beta HPV types (HPV8, -20, -25, -98, -17a, -38, -76, and -92). The beta HPVs included in the study come from four of the five species that comprise that genus. The HPV18 and HPV45 E6 ORFs contain a mutation that was introduced into the splice donor site to promote translation of the full-length E6 protein as previously described (13).

N/Tert-1 cells are human foreskin keratinocytes immortalized by human telomerase reverse transcriptase (hTert). They do not produce p16^INK4a protein (44) but are competent for the induction of p53 and its direct target genes following DNA damage (Fig. 1). The N/Tert-HA-E6 cell lines are the same ones that were used in our proteomic studies and that we used to validate the type-specific interactions of HPV38 E6 and HPV92 E6 with p53, as well as the extension of p53 half-life that occurs in cells expressing HPV17a, -38, and -92 E6 proteins. The use of the promoter present in the retroviral LTR to drive E6 transcription in these cell lines ensures that the stable cells maintain a low level of E6 protein expression.

FIG 1.

p53 targets are induced after mitomycin C treatment in some genus beta HPV E6 cells, but not in the presence of HPV76 E6. (A) N/Tert-1 parental cells (−) or N/Tert-1 cells stably expressing HA-HPV16 E6 (16E6) (+) were treated with 0, 0.5, 2, 5, or 10 μg/ml mitomycin C (MMC) and harvested 18 h after treatment. Lysates were analyzed by Western blotting using antibodies to p53, MDM2, HA, and actin. The positions of molecular mass markers (in kilodaltons) are indicated to the left of the blots. (B) N/Tert-1 parental cells or N/Tert-1 cells stably expressing one of several HA-HPV E6 proteins were treated with 2 μg/ml MMC and harvested 18 h after treatment. The E6 proteins were from HPV16, HPV8, HPV20, HPV17a, HPV76, and HPV92. The lysates were analyzed by Western blotting using antibodies to p53, p53 phosphorylated on serine 15, p53 acetylated on lysine 382, p53 phosphorylated on serine 392, MDM2, p21, HA, and actin. (C) N/Tert-1 cells containing an empty vector or N/Tert-1 cells stably expressing one of several untagged HPV E6 proteins were treated with 2 μg/ml MMC and harvested 18 h after treatment. The lysates were analyzed by Western blotting using antibodies to p53, MDM2, p21, HA, and actin. N/Tert HA-HPV16 E6 cells were included as a control.

First, we established conditions for the induction of DNA damage in N/Tert-HA-E6 cells to assess p53 responses in the presence of various E6 proteins. We first used mitomycin C (MMC), which alkylates DNA and results in cross-links between complementary strands (49). On the basis of the observation of Butz and colleagues (50) that MMC treatment can decrease expression of E6 and E7 in cervical cancer cell lines, we tested a range of doses of MMC to identify the treatment conditions that would lead to a maximal p53 response in E6-negative cells with minimal decrease in E6 expression. Parental N/Tert-1 cells or cells stably expressing HA-tagged HPV16 E6 (HA-HPV16 E6) were treated with a range of doses of MMC from 0 to 10 μg/ml for 18 h (Fig. 1A). In parental (E6-negative) cells, the increase in p53 expression after MMC treatment was apparent at the 2-μg/ml dose and stronger at higher doses. MDM2 is a direct transcriptional target of p53 and increases maximally at 2 and 5 μg/ml MMC in control cells. In cells expressing HA-HPV16 E6, we detected a low level of p53 in untreated cells and a modest induction of p53 and MDM2 after MMC treatment. This is consistent with the published reports showing that the p53 remaining in high-risk E6-positive cells remains able to transactivate target promoters (14, 15). The expression of HA-tagged E6 was decreased at the 5- and 10-μg/ml MMC doses. We therefore determined that the 2-μg/ml dose of MMC would maximize the p53 response while minimizing the effects on E6 protein levels, and this dose was used in all subsequent experiments in the study.

We next tested the induction of p53 in the presence of six different HPV E6 proteins from genus beta HPV: two of the four genus beta HPV species 1 types used in our proteomic studies (HPV8 and -20), both beta-2 types (HPV17a and -38), the beta-3 HPV76 E6, and the beta-4 HPV92 E6. HA-HPV16 E6 cells and parental N/Tert-1 cells were included as controls. As expected, the steady-state level of p53 in the absence of MMC treatment varied in the different cell lines and was higher in the HA-HPV17a, -38, and -92 E6 cells than in parental cells (Fig. 1B). MMC caused an increase in p53 protein levels in N/Tert-1 parental cells and a modest increase in N/Tert-HA-16E6 cells that was consistent with the low initial amount of p53 in those cells. In four of the six genus beta HPV E6-expressing cell lines (expressing HPV8, -20, -17a, and -38 E6s), MMC treatment caused an increase in p53 levels greater than or equal to that in parental cells. In untreated N/Tert-HA-HPV92 E6 cells, p53 levels were elevated than the levels of the controls and did not increase significantly after MMC treatment. Finally, in N/Tert-HA-HPV76 E6 cells, p53 levels increased modestly, similar to the increase observed in the presence of HPV16 E6, after MMC treatment.

To assess the functionality of the p53 in the various cell lines, we examined modifications on p53 after DNA damage (Fig. 1B). We chose three well-characterized modifications on p53: phosphorylation of p53 serine 15 (P Ser15), acetylation of p53 lysine 382 (Ac Lys382), and phosphorylation of p53 serine 392 (P Ser392). P Ser 15 is induced after DNA damage and inhibits the interaction of p53 with its negative regulator MDM2 (51). Ac Lys382 is added to p53 after DNA damage by the acetyltransferase p300 and promotes interaction of p53 with DNA (52). P Ser392 is elevated in human cancers, and a serine-to-alanine mutation at this site results in reduced p53 transactivation capacity (53, 54). P Ser 15 and P Ser 392 have previously been reported to be present at elevated levels on p53 in HPV38 E6/E7 cells (34). Each of these modifications was induced in N/Tert-1 parental cells following MMC treatment (Fig. 1B). Each of these modifications was also induced in all of the N/Tert-HA-E6 cell lines tested, and the degree of induction and the level of modified p53 protein were consistent with the degree of induction of total p53 protein in most of the cell lines. There were a few exceptions, however. For example, consistent with the earlier result (34), in untreated cells, P Ser 392 levels are high in the presence of HPV38 E6 relative to the other E6s. HA-HPV92 E6 cells do not show much induction of total p53 after DNA damage, but each modification is added after MMC treatment even in the presence of HPV92 E6. Taken together, these data indicate that the E6 proteins do not block the addition of these modifications, which are known to regulate the transcriptional activities of p53.

Next we examined the induction of the p53 transcriptional target MDM2 in the various cell lines expressing E6. MMC caused an increase in MDM2 protein levels in N/Tert-1 parental cells. In most of the beta HPV E6 cell lines, MDM2 was induced after MMC treatment. The degree of induction varied somewhat, but only HPV76 E6 was able to limit MDM2 induction to a degree comparable to that of HPV16 E6. Additional replicates of this experiment included additional HPV types, always with the result that induction of p53 and MDM2 was minimal after DNA damage in the presence of the high-risk alpha HPV genus E6 proteins and was affected similarly only by the genus beta HPV76 E6 protein. We do note that in some experiments in HA-HPV92 E6 cells, p53 levels were already high and did not increase further after DNA damage. However, MDM2 protein levels did increase severalfold after MMC treatment in the HPV92 E6-expressing cells (Fig. 1B).

To examine whether the HA tag present on the N terminus of each of these E6 proteins affected the ability of E6 to interfere with p53 transactivation functions, we generated N/Tert-1 cell lines that expressed an untagged version of each E6 protein. Each of these cell lines was tested for the induction of p53 and MDM2 proteins after mitomycin C treatment (Fig. 1C). Here untagged HPV16 E6 blocked p53 and MDM2 induction even more efficiently than HA-tagged HPV16 E6 did. Other E6s exhibited effects on p53 and MDM2 protein levels similar to those seen with the tagged versions. Briefly, HPV8, -20, -17a, -38, and -92 E6s did not block p53 or MDM2 induction after DNA damage, although it should be noted that p53 levels were increased in untreated cells in the presence of the HPV17a, -38, and -92 E6 proteins (as for the tagged versions). Untagged HPV76 E6 exhibited the same ability to block p53 and MDM2 induction after DNA damage as did HA-HPV76 E6.

It is clear that stable levels of the various E6 proteins vary for the different HPV types. For example, in this experiment (Fig. 1B) and in subsequent experiments, HPV8 E6 was always detected at a much lower level than HPV17a E6, perhaps limiting our ability to compare HPV8 E6 to other E6s with regard to its ability to block p53 transactivation functions. It should be noted, however, that the level of an E6 from a given HPV type relative to other E6s was consistent across different cell types and in different infections and that all of the E6 genes were expressed from the same retroviral promoter-E6-IRES-puro construct. We believe that the different levels of HPV E6 proteins observed in cells may be due to posttranscriptional regulation. Therefore, to best represent these differences, we have chosen to display intermediate exposures of all the HA-E6 Western blots. Each of the E6 proteins (e.g., HPV8 E6) was confirmed to be present in puromycin-resistant stable cell lines in each experiment using longer Western blot exposures.

To examine whether the E6 effects on p53 were restricted to the cell lines in the immortalized N/Tert-1 background, we generated primary human keratinocyte (G5-Ep) cell lines expressing untagged E6 proteins from HPV16 and six different genus beta HPVs (expressing HPV8, -20, -17a, -38, -76, and -92 E6s). G5-Ep cells are normal human newborn foreskin primary epidermal keratinocytes that have been previously described (45, 46). Following treatment with mitomycin C, p53 and MDM2 levels were assessed by Western blotting (Fig. 2). In these cells, E6s had the same effects on p53 and MDM2 protein levels as observed in the N/Tert-1 cells. Mitomycin C treatment induced the expression of both proteins in cells containing empty vector and GFP control cells as well as in HPV8, -20, -17a, and -38 E6 cells. In HPV92 E6 cells, the levels of p53 protein were elevated relative to controls both before and after DNA damage, but MDM2 expression was clearly induced after MMC treatment in the presence of HPV92 E6. As in the N/Tert cell lines both with and without the HA tag, only HPV16 E6 and HPV76 E6 were able to block the induction of p53 and MDM2 proteins after DNA damage.

FIG 2.

p53 targets are induced after mitomycin C treatment in primary keratinocytes, and induction is blocked by a subset of HPV E6 proteins. G5-Ep cells containing an empty vector or stably expressing HA-GFP or one of several untagged HPV E6 proteins were treated with 2 μg/ml MMC and harvested 18 h after treatment. The lysates were analyzed by Western blotting using antibodies to p53, MDM2, p21, HA, and actin.

To confirm that these effects were not restricted to DNA damage induced by a single agent, we also tested p53 activity in the presence of various E6s after treatment with neocarzinostatin (NCS), a radiomimetic and inducer of double-stranded DNA breaks. Again, we treated N/Tert-1 cells that harbor the empty vector and HA-HPV16 E6 cells with a range of doses of NCS, from 0 to 300 ng/ml, for 4 h. NCS treatment efficiently induced p53 protein and had minimal to no effect on E6 levels (Fig. 3A), and we chose to use 100 ng/ml NCS in the subsequent experiment. Because it is a radiomimetic, NCS reproduces the mechanism and effects of DNA damage resulting from exposure to UV radiation. An evolving model of beta HPV disease proposes that UV radiation and beta HPV infection may be cofactors in the development of NMSCs.

FIG 3.

p53 targets are induced after neocarzinostatin treatment in some genus beta HPV E6 cells. (A) N/Tert-1 cells containing an empty vector (−) or N/Tert-1 cells stably expressing HA-HPV16 E6 (+) were treated with 0, 50, 100, 200, or 300 ng/ml neocarzinostatin (NCS) and harvested 4 h after treatment. The lysates were analyzed by Western blotting using antibodies to p53, HA, and actin. (B) N/Tert-1 cells containing an empty vector or N/Tert-1 cells stably expressing one of several HA-HPV E6 proteins were treated with 100 ng/ml NCS and harvested 4 h after treatment. The lysates were analyzed by Western blotting using antibodies to p53, MDM2, p21, HA, and actin. (C) N/Tert-1 cells containing an empty vector or N/Tert-1 cells stably expressing one of several HA-HPV E6 proteins were exposed to 15-mJ/cm² UV light and harvested 6 h after treatment. The lysates were analyzed by Western blotting using antibodies to p53, MDM2, HA, and actin.

As in the MMC experiment, we treated a panel of N/Tert-HA-E6 cells with NCS and assayed p53 and target gene protein levels by Western blotting (Fig. 3B). Cells containing the empty vector were included as a control, as were high-risk genus alpha HPV HA-HPV16 E6 cells and low-risk genus alpha HPV HA-HPV6b E6 cells. We also tested one genus beta HPV E6 protein from each of the four species included in our study. The control cells behaved as expected: p53 and MDM2 were induced following NCS treatment in cells containing empty vector and cells expressing low-risk HA-HPV6b E6, and minimal to no induction was detected in the presence of HPV16 E6. In the beta HPV E6 cell lines, p53 increased after NCS treatment in the presence of HPV8, -17a, and -76 E6s, although the initial level of p53 and the magnitude of the increase were smallest in the HA-HPV76 E6 cells. Consistent with previous results, p53 levels were elevated in HA-HPV92 E6 cells in the absence of DNA damage and did not increase further after NCS treatment. The effect of NCS treatment on MDM2 protein was generally consistent with our earlier results—induction was blocked by HPV76 E6 but not by HPV17a E6 or HPV 92 E6—but here, HPV8 E6 also had some ability to block MDM2 protein induction. This result was not observed after treatment of HPV8 E6 cells with other genotoxic stressors.

Since UV radiation is a likely source of DNA damage in cells naturally infected with beta HPV, we next tested UV treatment in a similar experiment. N/Tert-HA-E6 cells were treated with 15-mJ/cm² UV and then incubated for 6 h. Again, p53 and MDM2 protein levels rose after UV treatment in control cells containing empty vector but not in cells expressing HA-HPV16 E6. The proteins were also induced in the presence of beta E6s, and in particular, we note that HPV76 E6 in the experiment shown here (Fig. 3C) did not block the induction of MDM2 protein after UV treatment, even though the increase in p53 protein in these cells was minimal.

Only high-risk genus alpha HPV E6 proteins are able to block the induction of MDM2 at the RNA level after DNA damage.

In order to expand these tests to a larger number of E6s and to enable an accurate quantification of the effects of DNA damage, we proceeded to analyze the levels of selected transcripts in the N/Tert-HA-E6 and control cell lines. The complete panel of N/Tert-HA-E6 cell lines as well as N/Tert-empty vector and N/Tert-HA-GFP cells as controls were treated with 2 μg/ml MMC for 18 h. RNA was isolated from these cells as well as from an untreated control for each cell line. The level of p53 RNA in each sample was determined by quantitative real-time reverse transcription (RT-PCR) (Fig. 4A). Values displayed in the graphs are the average of three independent experiments. p53 is predominantly regulated at a posttranscriptional level following DNA damage, and so not surprisingly there was little variation in the level of p53 RNA across the various cell lines and in the presence and absence of MMC treatment.

FIG 4.

High-risk HPV E6 proteins block the induction of MDM2 transcription after DNA damage. N/Tert-empty vector, N/Tert-HA-GFP, or N/Tert-HA-11E6 cells carrying high-risk (hr) or low-risk (lr) or non-cancer-associated (nc) alpha or beta HPVs were treated with 2 μg/ml MMC or vehicle control for 18 h and harvested for RNA isolation. Total cellular RNA was reverse transcribed, and cDNA was analyzed by quantitative real-time PCR to measure p53 (A) and MDM2 (B) transcript levels. Values are averages of three independent experiments, each normalized to G6PD transcript levels. Error bars indicate standard deviations. P values for all of the qRT-PCR experiments are listed in Data Set S3 in the supplemental material.

In contrast, MDM2 is a direct transcriptional target of p53 (55), and MDM2 RNA levels increased 4- to 6-fold in the control cell lines after MMC treatment (Fig. 4B). There was little to no induction of MDM2 RNA after DNA damage in cells expressing any one of the five high-risk E6s, with at most a 2-fold increase in the presence of HPV18 E6. None of the other E6s tested, whether low-risk, non-cancer-associated genus alpha HPV, or genus beta HPV, were able to block MDM2 induction, and each of the cell lines expressing these proteins exhibited the 4- to 6-fold increase in MDM2 RNA after MMC treatment. Although the HA-HPV76 E6 cells in several previous experiments did not show an elevated level of p53 or MDM2 proteins after DNA damage, here we saw that MDM2 RNA increased more than 4-fold after mitomycin C treatment in the same HA-HPV76 E6 cells.

Genome-wide transcriptional profiling indicates that HPV16 E6 cells respond differently to DNA damage than do selected genus beta HPV E6 cells.

Next we wished to generate a more comprehensive profile of gene expression changes following DNA damage in the N/Tert-E6 cell lines. We selected a subset of the cell lines for transcriptional analysis on microarrays. The five cell lines included in this experiment were the control cell line containing empty vector and the N/Tert-HPV16, -6b, -8, and -92 E6 cell lines. HPV16 E6 was included for its ability to bind and degrade p53, HPV6b E6 was included as a low-risk genus alpha HPV control unable to bind or degrade 53, and we tested two genus beta HPV E6s. HPV8 E6 does not bind or stabilize p53, while HPV92 E2 both binds and stabilizes p53. Cells were treated with 2 μg/ml MMC for 18 h or left untreated, and RNA was isolated for analysis on Affymetrix U133 plus 2.0 arrays. Three to five replicates of each condition were analyzed, with each replicate for a given cell line and treatment processed independently of the others at all steps from cell culture to microarray analysis.

Microarray data from 34 samples were collected, normalized with GenePattern, and filtered to remove low-intensity probe sets. The 1,000 most differentially expressed genes were selected for further analysis (see Data Set S4 in the supplemental material). Both the sample and gene lists were grouped by hierarchical clustering. After this analysis, the samples fell into two distinct categories: one group containing all five untreated N/Tert cell lines plus the HPV16 E6 cell line after MMC treatment and a second group containing the other four cell lines (empty vector plus HPV6b, -8, -92 E6 cells) that were treated with MMC (Fig. 5). Replicate data for each treatment condition clustered together. This sample grouping suggested that of the cell lines tested, only the cells expressing HPV16 E6 behaved significantly differently than the empty-vector-containing control cells after DNA damage.

FIG 5.

HPV16 E6 cells exhibit a unique pattern of global gene expression following DNA damage. N/Tert-empty vector or -E6 cells were treated with 2 μg/ml MMC or vehicle control for 18 h, and RNA was analyzed on Affymetrix U133 plus 2.0 arrays. The heat map shows the 1,000 most differentially expressed genes, after hierarchical clustering for both samples and genes. Enrichment signatures in gene clusters were determined by DAVID analysis. EGF, epidermal growth factor; min, minimum; max, maximum.

The differentially expressed genes also fell into several clusters, and five prominent clusters were chosen for further analysis. DAVID Functional Annotation Analysis was used to generate lists of genes that were overrepresented in each cluster (see Data Set S5 in the supplemental material). In particular, we noted that one cluster was enriched in genes involved in the positive regulation of apoptosis and cell cycle inhibition. A second cluster contained many genes required for cell cycle progression. We hypothesized that genes from these clusters, particularly from the proapoptotic/cell cycle inhibition group, would be differentially regulated in cases where E6 could inactivate p53 transactivation functions. We chose four genes that represented various Gene Ontology categories included in cluster 1 for validation studies: CDKN1A (p21), BAX, FAS, and TP53I3.

A subset of genus beta HPV E6 proteins block the induction of p21 after DNA damage.

The microarray experiment suggested that the regulation of proapoptotic and cell cycle inhibition genes might be altered by E6s from some HPV types. To investigate this idea and to extend the finding to more HPV types, we began by investigating the regulation of p21, a cell cycle inhibitor that is a target of p53 but has different functions and is regulated differently than MDM2. Again we used quantitative real-time RT-PCR to assess the level of p21 transcript in the same DNA damage experiments described above. In contrast to the MDM2 result, here we observed that p21 levels did not increase after MMC treatment in several of the genus beta HPV N/Tert-HA-E6 cell lines (Fig. 6). p21 RNA levels rose 2- to 4-fold in control cells treated with MMC. This increase was largely blocked by the high-risk genus alpha HPV E6 proteins, but not by the low-risk or non-cancer-associated alpha HPV E6s. Examining the beta HPV E6s, we noted that HPV8, -20, and -25 E6 proteins did not affect the induction of p21 after DNA damage and exhibited a 3- to 4-fold increase in p21 RNA following MMC treatment. In contrast, there was a negligible increase (comparable to that in high-risk HA-HPV E6 cell lines) in the HA-HPV98, -17a, -38, and -92 E6 cells. There was an intermediate effect in HA-HPV76 E6 cells.

FIG 6.

High-risk HPV E6 proteins and a subset of genus beta HPV E6 proteins block transcription of p21 after DNA damage. N/Tert-empty vector, -GFP, or -E6 cells were treated with 2 μg/ml MMC or vehicle control for 18 h and harvested for RNA isolation. Total cellular RNA was reverse transcribed, and cDNA was analyzed by quantitative real-time PCR to measure p21 transcript levels. Values are averages of three independent experiments, each normalized to G6PD transcript levels. Error bars indicate standard deviations. P values for all of the qRT-PCR experiments are listed in Data Set S3 in the supplemental material.

Given these results, we proceeded to assess p21 protein levels in the Western blot experiments shown earlier (Fig. 1B and C, 2, and 3B). Taken together, these results indicated that p21 protein levels increase most after DNA damage in cells expressing HPV8 E6 and HPV20 E6. In the same experiments, p21 induction at the protein level was lower or absent in the other E6 cell lines. p21 protein is not induced after UV treatment (56), so we did not include it in the UV experiment.

Proapoptotic genes are induced after DNA damage in the presence of genus beta HPV E6 proteins.

The microarray and DAVID results also indicated that proapoptotic genes were differentially regulated after DNA damage. We chose three of these genes, BAX, TP53I3 (PIG3), and FAS, and measured these RNAs in untreated or MMC-treated N/Tert HA-E6 and control cells by quantitative real-time RT-PCR (Fig. 7). These genes were upregulated by 1.5- to 2-fold following MMC treatment in control cells, but the amount of RNA increased very little or not at all in the presence of any of the five high-risk E6 proteins. The 1.5- to 2-fold induction seen in the control cells was also present in some of the other cell lines expressing the low-risk, genus alpha HPV non-cancer-associated, or genus beta HPV E6 proteins. The small magnitude of the increase in RNA levels together with the high degree of variability in the amount of induction make it difficult to draw any conclusions on the effects of beta E6s on the expression of these genes. BAX transcript levels in beta HPV E6 cells look comparable to those in empty-vector-containing and control cell lines, but for FAS transcripts, the degree of induction in several beta E6 cell lines (HPV17a and -38 E6) is comparable to that in some of the high-risk alpha HPV E6 cell lines (HPV52 and -18 E6). TP53I3 transcripts exhibit an intermediate phenotype. Possible reasons for this observation, especially in light of the larger effect of some beta E6s on p21 RNA levels, are discussed further below.

FIG 7.

High-risk HPV E6 proteins block transcription of proapoptotic genes after DNA damage. N/Tert-empty vector, -GFP, or -E6 cells were treated with 2 μg/ml MMC or vehicle control for 18 h and harvested for RNA isolation. Total cellular RNA was reverse transcribed, and cDNA was analyzed by quantitative real-time PCR to measure BAX, TP53I3 (PIG3), and FAS transcript levels. Values are averages of three independent experiments, each normalized to G6PD transcript levels. Error bars indicate standard deviations. P values for all of the qRT-PCR experiments are listed in Data Set 3 in the supplemental material.

DISCUSSION

The tumor suppressor p53 was first identified though its binding to simian virus 40 (SV40) large T antigen, and years of subsequent study have established that p53 is a central and critical player in many cellular processes, largely in its role as a sensor of damage to the genome. With regard to the papillomaviruses, p53 is a key target of the high-risk HPV E6 proteins and is targeted for degradation via a high-risk E6- and E6AP-dependent mechanism. This is thought to permit HPV-infected cells to avoid apoptosis as a result of E7-mediated triggers that drive unscheduled DNA replication in otherwise terminally differentiated cells. The abilities of E6 to bind and degrade p53 remain a major correlate of the ability of HPVs to cause cancer.

Our recent systematic proteomic study showed that p53 is bound not only by high-risk HPV E6s but that p53 also associates with at least two of the genus beta HPV E6 proteins, from HPV38 and HPV92 (13). These two beta HPV types are from different virus species, and the other binding partners engaged by their E6 proteins and the downstream effects of the interaction with p53 could be rather different. Cells expressing these two beta HPV E6 proteins as well as cells that express HPV17a E6 can also stabilize p53 levels in cells, but it is possible that the stabilization of p53 in some or all of these beta HPV E6-expressing cell lines may not be a direct consequence of its interaction with E6. We note that further studies will be required to determine the nature of this interaction and which other proteins might be present in a complex that includes these beta HPV E6 proteins and p53.

In this study, we have performed experiments to assess the functional status of p53 in human keratinocytes in the presence of a variety of HPV E6 proteins. Regarding the high-risk genus alpha HPV E6 proteins, we note that the p53 remaining in HPV16 E6-expressing cells was able to mediate a modest increase in MDM2 and p21 proteins after MMC treatment (Fig. 1B). In this light, it is important to recognize that the resulting level of either p53 target in the cells is still lower than in control or genus beta HPV E6 cell lines. This result is consistent with reports that p53 protein retains its transactivating functions in high-risk E6 cells or cervical cancer cells (14–16) but that its functions are largely eliminated via the E6AP and proteasome-mediated degradation of the majority of p53.

With this in mind, we proceeded to test genus beta HPV E6 proteins for their abilities to inhibit the expression of p53 targets following DNA damage, and these findings are summarized in Table 1. We used DNA damage induced by the radiomimetic neocarzinostatin, the alkylating agent MMC, or UV irradiation as triggers of p53 stabilization and to activate p53 transactivation functions on its downstream targets. Beginning with a subset of the genus beta HPV E6 proteins from our previous studies, we tested the induction of p53 targets MDM2 and p21 at the protein level after DNA damage induced by either one of the drugs. Most of the beta HPV E6s were unable to block the induction of these targets after DNA damage (Fig. 1B and C, 2, and 3B and C). One notable exception was that in several experiments, the genus beta HPV species 3 HPV76 E6 had an effect comparable to that of HPV16 E6 (Fig. 1B and C, 2, and 3B and C) in that both cell lines contained a low level of p53 protein that was induced very little or not at all after DNA damage. HPV76 E6 has a less clear effect on MDM2. MDM2 protein induction in HPV76 E6 cells may depend on the nature of the DNA damage, as it was minimally induced after MMC or NCS treatment (Fig. 1B and C, 2, and 3B) but was induced more strongly following UV irradiation (Fig. 3C). p21 transcript levels also increased about 2-fold following MMC treatment of HPV76 E6 cells (Fig. 6). One possible way that HPV76 E6 might mediate these effects relates to our previous finding that p53 is less stable in HPV76 E6 cells than in control cells (but is slightly more stable in HPV76 E6 cells than in HPV16 E6 cells) (13). Perhaps this intermediate degree of p53 destabilization results in the variable effects on MDM2 and p21 observed in this study. Our proteomic studies identified few to no HPV76 E6-specific binding partners, so the proteins or other factors responsible for the destabilization of p53 in the presence of HPV76 E6 remain to be identified.

TABLE 1.

Summary of effects of genus beta HPV E6 proteins on p53 target genes in human N/Tert-1 cell lines

HPV source of E6	Ability to bind p53	p53 steady-state level and stability	Change in protein or RNA level after DNA damagea:
			p53 protein	MDM2 protein	MDM2 RNA	p21 RNA
None	N/Ab	Normal	+	+	+	+
HPV16	Through E6AP	Low	−	−	−	−
HPV8	No	Normal	+	+	+	+
HPV20	No	Normal	+	+	+	+
HPV17a	Weakc	High	+	+	+	−
HPV38	Yes	High	+	+	+	−
HPV76	No	Low	−	+/−	+/−	+/−
HPV92	Yes	High	+	+	+	−

^aChanges in protein or RNA level are shown as follows: +, increased; −, no increase; +/−, slight increase.

^bN/A, not applicable.

^cHPV17a E6 binding to p53 was not detected in mass spectrometry experiments but is sometimes weakly detected in immunoprecipitation-Western blot experiments.

We expanded our studies to include 17 different E6 proteins, eight of them from genus beta HPV. Although none of the genus beta HPV E6 proteins was able to block the induction of MDM2 RNA after DNA damage (Fig. 4B), several of the genus beta HPV species 2, 3, and 4 E6s efficiently blocked p21 induction after DNA damage (Fig. 6). DNA damage induced the transcription of several proapoptotic genes less strongly, so it was more difficult to determine the effect of E6 on these transcripts (Fig. 7). The differential response of the cell cycle inhibitor p21 versus the proapoptotic genes to DNA damage observed is consistent with studies in the literature showing that these two classes of genes are regulated differently depending on many factors, including the dose and nature of the DNA damage incurred in a cell (reviewed in reference 57). Although more genotoxic stress relative to what we used here might clarify whether the species 2, 3, and 4 beta HPV E6s can also inhibit BAX, FAS, and TP53I3 after DNA damage, our ability to conduct these experiments was limited by the fact that the larger doses of mitomycin C decreased the expression of E6s (Fig. 1A).

Together, these data indicate that some beta HPV E6 proteins can at least partially inhibit some of the downstream targets of p53. We note that HPV16 E6 was able to inhibit p53 target expression in each of the experiments included here, while the beta HPV E6s from species 2 (HPV17a and -38), 3 (HPV76), and 4 (HPV92) were each able to reduce the levels of p53 targets in only a subset of our experiments. The genus beta HPV species 1 (HPV8 and -20) E6 proteins in our studies were never able to inhibit the expression of p53 targets, although we note that studies from other groups suggest that these E6s have other, perhaps complementary, effects on DNA repair after UV irradiation (28, 31, 58). We also note the recent publication of Wallace et al. reporting that some beta HPV E6s block the stabilization of p53 after genotoxic stress resulting from multinucleation and abnormal centrosome phenotypes (59). Overall, it is clear that heterogeneity is a characteristic of the p53-dependent response to HPV E6 expression. This heterogeneity is reflected not only in the diversity of the effects of the beta HPV E6s but also in the effects observed in response to genotoxic stress of different types and perhaps under different experimental conditions. Differences in experimental conditions or interpretation could also explain the discrepancy between our results and those of an earlier publication from the same research group (31). In particular, their work demonstrated a reduced level of acetylation on p53 lysine 382 in the presence of HPV5 and HPV8 E6 proteins after UV irradiation, which is in contrast to our finding that HPV8 E6 did not block p53 K382 acetylation after mitomycin C treatment. It is possible that the method used to induce a DNA damage response is critical to the presence of a modification at this site.

The potential for beta HPVs to cause cancer and the mechanisms by which they might do so have been of some interest in the field. Our data suggest that functional inactivation of p53 may be one of the strategies used by the genus beta HPV E6 proteins as part of their life cycle but that the mechanism of targeting p53 varies among the beta HPVs and may not be used by all beta HPV E6s to the same degree. Beta HPV E6 proteins may also target other pathways not dependent on p53 in order to effect some of the same downstream consequences. For example, other studies have identified functions of genus beta HPV E6s potentially related to cancer initiation such as the impaired response to UV-induced DNA damage mentioned earlier. Our own lab and others recently identified the interaction of MAML1 with beta HPV E6 proteins and showed that this functions to inhibit Notch signaling in keratinocytes. Since Notch has been demonstrated to be a tumor suppressor in epithelial cells, this represents an additional means by which E6 might promote dysregulated cell growth without direct involvement of p53 (60–63). Other studies have suggested mechanisms by which E6, particularly HPV38 E6, might act on p53-related effectors with some of the same downstream results. For example, it was proposed that HPV38 E6 and E7 increase expression of ΔNp73, which is an inhibitor of p53 (34). Overall, it is clear that the genus beta HPV E7 proteins have the ability to bind pRB1 (43, 62), presumably to inactivate it and allow progression into a cellular state conducive to DNA replication, and that proapoptotic signals induced by this event need to be overcome by the beta HPVs in order to allow their productive replication.

Our results characterize a broad selection of E6 proteins, including eight genus beta HPV E6s from four different HPV species, with regard to their ability to block p53 transactivation functions. While high-risk genus alpha HPV E6 proteins, by efficiently targeting p53 for degradation, were able to block efficient activation of all of the target genes we tested, genus beta HPVs varied in their ability to do so. HPV17a, -38, and -92 E6s bound and/or stabilized p53 and inhibited the induction of p21 after DNA damage but did not block the accumulation of MDM2 protein or RNA after UV irradiation, mitomycin C treatment, or neocarzinostatin treatment. Future experiments to understand the nature of the HPV38 E6 and HPV92 E6 interactions with p53 might investigate whether the p53-E6 complex is present on sites in the viral genome or whether E6 alters the binding sites of p53 in the human genome. We also hope that future experiments will address the role of this interaction in the HPV life cycle. In contrast, HPV76 E6 appeared to reduce the level of p53 in cells, which is consistent with our earlier results and also led to minimal induction of p53 and MDM2 protein and reduced induction of p21 RNA after DNA damage. HPV76 E6 does not appear to bind to the ubiquitin ligase E6AP, which is required for p53 degradation in the presence of high-risk HPV E6 proteins. Future experiments will therefore aim to determine how p53 is destabilized in the presence of HPV76 E6.

Taken together, it is clear that as for the genus alpha HPV E6s, genus beta HPV E6s have a range of effects on p53 target genes and presumably mediate these effects via multiple different mechanisms. This survey should serve as a resource for the beta HPV field and may help to inform continued research examining a potential role of beta HPVs in cancer.