The E8 repression domain can replace the E2 transactivation domain for growth inhibition of HeLa cells by papillomavirus E2 proteins

Abstract

Continuous expression of the human papillomavirus (HPV) oncoproteins E6 and E7 is required for the growth of cervical cancer cell lines. So far, only the overexpression of the wild type papillomavirus E2 protein has been shown to induce growth arrest in HPV18-positive HeLa cells by repressing E6/E7 transcription. Growth arrest by E2 requires the aminoterminal transcription activation domain in addition to the carboxyterminal DNA-binding domain. Several papillomaviruses such as the carcinogenic HPV31 express in addition to E2 an E8^∧E2C fusion protein in which the E8 domain, which is required for repression of replication and transcription, replaces the E2 activation domain. In this report, we demonstrate that the HPV31 E8^∧E2C protein is able to inhibit the growth of HeLa cells but not of HPV-negative C33A cervical cancer cells. Growth repression by E8^∧E2C correlates with repression of the endogenous HPV18 E6/E7 promoter and the reappearance of E6- and E7-regulated p53, pRb and p21 proteins, suggesting that E8^∧E2C inhibits growth by reactivating dormant tumor suppressor pathways. Growth inhibition requires an intact E8 repression domain in addition to the carboxyterminal E2C DNA binding domain. Chromatin immunoprecipitation experiments suggest that the E8 repression domain enhances binding to the HPV18 promoter sequence in vivo. In summary, our results demonstrate that the small E8 repression domain can functionally replace the large E2 transactivation domain for growth inhibition of HeLa cervical cancer cells. © 2007 Wiley-Liss, Inc.

Infections with certain human papillomaviruses (HPV), most notably types 16 and 18, are a necessary risk factor for the development of cervical cancer.1 While the HPV genome is present in an episomal state in low-grade lesions, the whole viral genome or fragments thereof are often integrated into the chromosomal DNA of the host cell in the majority of HPV-induced carcinomas.2 The HeLa cell line was derived from a cervical cancer and contains a partial HPV18 genome integrated into the host chromosomes.3 HeLa cells express the 2 major viral oncoproteins E6 and E7 but not the viral proteins involved in the control of viral replication and transcription such as E1, E2 and E8^∧E2C.3 Continuous expression of E6 and E7 is necessary for growth of HeLa cells and other HPV-positive cell lines derived from cervical cancers as shown by RNA interference experiments.4, 5, 6, 7 The E6 and E7 proteins derived from carcinogenic HPVs interfere with cell cycle control proteins. E6 forms a ternary complex with the E3 ubiquitin ligase E6AP and p53, which results in the degradation of p53 and therefore in the reduction of mRNA levels of p53 target genes.8, 9 The E7 protein binds to Rb family members and disrupts Rb/E2F complexes, resulting in increased expression of E2F-regulated genes.10, 11 In addition, the E7 protein enhances degradation of Rb family members.12, 13

A peculiar feature of carcinogenic HPV types is that the E6 and E7 transcripts are generated by alternative splicing from a common precursor RNA.14 Transfection studies using reporter plasmids of the corresponding promoters of HPV16, 18 or 31 demonstrated that coexpression of viral E2 proteins resulted in repression of promoter activity, which is due to binding of E2 to proximal recognition sequences.15, 16, 17, 18, 19 Similarly, repression of the endogenous HPV18 E6/E7 promoter in HeLa cells could be observed when E2 genes from bovine papillomavirus Type 1 (BPV1), HPV16 or 18 were introduced into HeLa cells.20, 21, 22 Overexpression of E2 resulted in growth arrest or apoptosis of HPV-positive cervical cancer cells.20, 21, 22 This was mainly due to the specific repression of the HPV promoter as the growth of HPV negative cells was not influenced by overexpression of E2.20, 21, 22, 23, 24 In line with this, the reduction of E6/E7 mRNA levels resulted in an increase in E6 and E7 target protein levels such as p53, p21 and pRb.20, 21, 22, 23, 24

E2 proteins bind as dimers to specific DNA sequences (E2 binding site, E2BS) to modulate transcription and viral DNA replication. The carboxyterminal domain which comprises ∼100 amino acids (aa) is sufficient for sequence-specific DNA recognition and dimerization25 (Fig. 1). The E2 aminoterminal domain of ∼200 aa is responsible for the stimulation of transcription, viral DNA replication and segregation of viral genomes upon binding to viral and cellular interaction partners26, 27, 28 (Fig. 1). Mutational analyses of E2 revealed that sequence-specific DNA binding activity is necessary for growth suppression.21, 29 BPV1 encodes in addition an aminoterminally truncated form of E2, E2TR, which is DNA-binding competent but cannot activate transcription or replication.30 Surprisingly, E2TR did not induce growth arrest of HeLa cells and increase p21 or p53 protein levels despite E2TR being able to inhibit the E2-dependent transcription of reporter plasmids in HeLa cells.20, 21, 22, 23 Consistent with the interpretation that the E2 transactivation domain is necessary for growth inhibition, mutational analyses of the E2 aminoterminal domain demonstrated a correlation between E2's ability to transactivate an E2-responsive reporter plasmid and the transcriptional repression of the endogenous HPV18E6/E7 promoter repression and therefore growth inhibition.24, 31 Interestingly, replacement of the E2 activation domain with the viral VP16 or cellular spi oncogene transactivation domains resulted in chimeras that were potent activators of E2-dependentreporter plasmids but unable to induce growth arrest of HeLa cells.21, 24, 31 Recent publications suggested that the aminoterminus of E2 interacts with the cellular Brd4 protein and that this interaction is involved in promoter repression and growth inhibition of HeLa cells.32

Figure 1. Colony reduction assay. HeLa cells were transfected with a puromycin-resistance plasmid and the empty pSG5 expression vector (vector) or derivates expressing E2 or HA-tagged E2 (E2-HA), E8^∧E2C or HA-tagged E8^∧E2C (E8^∧E2C-HA) or the E8^∧E2C-KWK mutant. The graph represents the average puromycin resistant colonies derived from 3 independent experiments. Standard deviations are indicated by error bars. Below the graph, a schematic representation of the structure of E2 and E8^∧E2C is shown. The aminoterminal domains unique for E2 and E8^∧E2C are shown as white and black boxes, respectively. The common carboxyterminus responsible for sequence specific DNA binding and dimerization is shown in gray.

Several papillomaviruses express a fusion protein in which the E8 gene substitutes for the E2 activation domain giving rise to E8^∧E2C.18, 33, 34, 35, 36, 37, 38 In addition, we found that an E8^∧E2C cDNA is also expressed in keratinocytes containing replicating HPV18 DNA (Stubenrauch, unpublished observation). The E8 domain encodes only 12 residues in HPV31 and bears no sequence homology to the aminoterminal domain of E2 (Fig. 1). HPV31 E8^∧E2C is a potent transcriptional repressor protein that inhibits the activity of the HPV31 E6/E7 promoter in transient transfection studies more efficiently than E2.18 In contrast to E2, E8^∧E2C is also able to repress transcription from promoter-distal E2BS and this is dependent upon the E8 part of the protein.39 The E8 domain represents a transferable transcription modulation domain and can confer repression activity to the Gal4 DNA binding domain.40

In this report, we demonstrate that E8^∧E2C protein derived from HPV31 is able to inhibit the growth of HeLa cells by repressing the endogenous HPV18 E6/E7 promoter and reactivating dormant tumor suppressor pathways. This activity depends upon the E8 repression domain that is required for efficient binding of the E8^∧E2C protein to the HPV18 promoter sequences in vivo as shown by chromatin immunoprecipitation assays.

Material and methods

Plasmids

The expression plasmids for HPV31 E2 (pSXE2), E8^∧E2C (pSGE8^∧E2C) and pSGE8^∧E2C KWK (pSGE8^∧E2C KWK), in which the residues KWK were replaced with AEA, are based upon pSG5 (Stratagene, Amsterdam, The Netherlands) and have been described.39, 41 An HA-tagged version of E2 was constructed by inserting a double stranded oligonucleotide (5′-AATTCTACC CATACGATGTTCCAGATTACGCTGAG-3′) into the EcoRI site of the HPV31 E2 sequence in the context of pSXE2. Sequences encoding the HA epitoptag were inserted in the E8^∧E2C and mutant genes into the EcoRI site by PCR. The DNA binding deficient mutant E8^∧E2C 118/119 was generated by site-directed mutagenesis and exchanges residues K118 to M and C119 to R, which correspond to residues 339 and 340 in the full length BPV1 E2 protein.29, 42 The luciferase reporter plasmid pC18-SP1-luc has been described.39 To generate the luciferase reporter plasmid p18URR-luc, a fragment (HPV18 nt. 7127-105) was amplified by PCR using the cloned HPV18 genome as template and primers 18URR MluI for (5′-TGTGCGTGTACGCGTCAGGAAGTAAT-3′) and 18URR NcoI rev (5′-TCAAAGCGCGCCATGGTATT GTGG-3′). The amplicon was digested with MluI and NcoI and cloned into MluI/NcoI-digested pGL3-basic (Promega, Mannheim, Germany). Plasmid pPur was used for selection of puromycin resistance (Clontech, Saint-Germain-en-Laye, France).

Cell culture

HeLa and C33A cells were maintained in DMEM (No. 41665-062; Invitrogen, Karlsruhe, Germany) supplemented with gentamycin and 10% fetal bovine serum (Seromed Biochrom, Berlin, Germany).

Colony reduction assay

Cells were transfected with lipofectamin (Invitrogen) according to the manufacturer's instructions. For colony reduction assays, 3.5 × 10⁵ HeLa or 2.5 × 10⁵ C33A cells were transfected in 60-mm dishes with (2 and 0.75 μg, respectively) the respective expression vectors and pPur (1 and 0.25 μg, respectively). One day after transfection, cells were split into 100-mm dishes and 24 hr later selected with 0.4 or 0.6 μg/ml puromycin, respectively. After 12–14 days of selection, colonies were fixed with acetone–methanol (1:1; v/v) solution, stained with eosin solution and then counted.

Transient luciferase expression assay

Approximately 5 × 10⁴ HeLa cells were seeded into 24 well dishes the day before transfection. Cells were cotransfected with 100 ng of luciferase reporters, 20 ng of pSG5 or the respective expression vector DNA as indicated in the figure legends. Transfections were carried out with 2 μl of lipofectamine (Invitrogen) and OptiMEM (Invitrogen) according to the manufacturer's recommendations. Luciferase assays were carried out 48 hr after transfection as previously described.39 The data presented in the figures are the average of at least 3 independent transfection experiments carried out in duplicate.

Western blot analysis

Transfected cells were lysed 48 hr post transfection in 20 μl of 4 × SDS gel loading buffer (Carl Roth, Karlsruhe, Germany) heated to 95°C and separated in a 12% SDS-PAGE. Proteins were transferred in 10 mM CAPS (pH = 10.3) on a nitrocellulose membrane (Protran, Whatman, Dassel, Germany). Membranes were blocked by incubation in TBS-0.1% Tween-20–5% nonfat dry milk for 30 min and then incubated with diluted primary antibodies (HA-probe, Santa Cruz sc-805, 1:1500; α-tubulin, Oncogene CP06, 1:1500; CDKN1A/p21, BD Pharmingen 556430, 1:1500). Bound antibodies were detected with anti-rabbit (polyclonal swine anti-rabbit immunoglobulin-HRP; Dako, Hamburg, Germany) or mouse antibodies conjugated to horseradish peroxidase (polyclonal rabbit anti-mouse immunoglobulin-HRP; Dako) and SuperSignal West Pico reagent (Perbio Science, Bonn, Germany). Chemoluminescent signals were recorded with a FluorSMax Imaging system (Biorad, Munich, Germany).

Immunofluorescence analysis

HeLa (1 × 10⁵) or C33A (1.5 × 10⁵) cells were seeded into collagen-coated MatTek, collagen coated glass bottom culture dishes (MatTek Corp., Ashland, MA), and transfected the next day with 1 μg of the respective expression vectors. Cells were fixed and permeabilized with acetone–methanol (1:1; v/v) solution, at room temperature (RT) for 2 min. Diluted primary antibodies (HA-probe, Santa Cruz sc-805, 1:50; HA.11, BabCO MMS-101P, 1:50; pRb, Santa Cruz sc-50, 1:50; p53, Santa Cruz sc-126, 1:50) were incubated for 1 hr at RT. After extensive washing with phosphate buffered saline (PBS)-0.05% Tween 20, Alexa-488-conjugated anti-rabbit or Cy3-conjugated anti-mouse antibodies were added and incubated for 1 hr at RT. Unbound antibodies were removed by extensive washing with PBS; 4′,6-diamidino-2-phenylindole (DAPI) in PBS was gently added to stain DNA and then unbound DAPI was removed by washing with PBS. Fluorescence signals were visualized with a Zeiss Axiovert 200 M microscope with a 63× objective and the respective fluorescence filter sets for DAPI, Alexa-488 and Cy3. Pictures were taken with an AxioCam MRm camera and processed with AxioVision software version 4.3 (Carl Zeiss AG, Oberkochen, Germany).

RNA isolation and real time PCR analysis

RNA was isolated from HeLa cells 72 hr post transfection with the RNeasy kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. RNA (1 μg) was DNAse-treated and reverse transcribed with the Quantitect reverse transcription kit (Qiagen). Duplicate samples were analyzed by real time PCR with an SDS5700 system (Applied Biosystems, Darmstadt, Germany) using qPCR™ Mastermix for SYBR Green I reagents (Eurogentec, Cologne, Germany). To determine relative transcript amounts, aliquots of cDNA corresponding to 50 ng of total RNA were analyzed with the following primer sets: HPV18E7 548F and HPV18E7 655R to detect HPV18 E6/E7 transcripts,43 phosphoglycerate kinase 1 (PGK1; Quantitect primer assay; Qiagen) or CDKN1A/p21 (Quantitect primer assay). To determine the amplification efficiency of the primer sets, standard curves from 50, 5 and 0.5 ng of cDNA were generated. Relative transcript amounts were calculated using PGK1 as a reference gene and the equations as described by Pfaffl.44

Chromatin immunoprecipitation analysis

For chromatin immunoprecipitation experiments, 2.8 × 10⁷ HeLa cells were transfected with each construct using calcium phosphate coprecipitation. After 48 hr cells were washed with PBS and treated with 1% formaldehyde in prewarmed PBS for 10 min at 37°C. Crosslinking was stopped by the addition of glycine to the final concentration of 0.125 M. Cells were collected by centrifugation and washed with ice-cold PBS. To prepare nuclei, cells were resuspended in hypotonic RSB buffer [10 mM Tris-HCl (pH 8.0), 3 mM MgCl₂], incubated for 10 min on ice and disrupted by Dounce homogenization. Nuclei were washed with RSB and resuspended in NSB buffer [10 mM NaCl, 10 mM Tris-HCl (pH 8.0), 0.1% NP-40, 1 mM EDTA]. After centrifugation through a sucrose cushion, the supernatant was removed and DNA–protein complexes were resuspended in 3 ml TE. To achieve an average length of the DNA of 500 bp, DNA–protein-complexes were sonicated (Bandelin Sonoplus, 40% amplitude, 90 sec in 1 sec intervals) and then digested with 10 U micrococcal nuclease (Fermentas, St. Leon-Rot, Germany) in the presence of 2 mM CaCl₂ for 15 min at 37°C. The reaction was stopped by adding 5 mM EDTA. For immunoprecipitation, the extract was adjusted with 1/10 volume of 11 × NET (550 mM Tris-HCl pH 7.4, 1.65 M NaCl, 5.5 mM EDTA, 5.5% NP40). Five micrograms of antibody (anti-HA, Abcam ab1424) was added to comparable amounts of chromatin as determined by adsorption at 260 nm. After over night incubation, 20 μl of protein A/G beads-slurry (Santa Cruz, Heidelberg, Germany) equilibrated in 1 × NET-buffer were added to the chromatin solution. After 3 hr, beads were washed thrice with 1 ml RIPA buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 5 mM EDTA, 0.5 % NP-40), thrice with 1 ml LiCl buffer (10 mM Tris-HCl pH 7.5, 250 mM LiCl, 1 mM EDTA, 0.5 % NP-40, 0.5% DOC) and thrice with 1 ml TE-buffer. To elute DNA the beads were resuspended in 200-μl TE/1% SDS and incubated for 15 min at 37°C. After centrifugation, the supernatant was transferred to a new tube and 200 μl TE and ProteinaseK (180 μg; Fermentas) were added and incubated over night at 55°C. DNA was purified by phenol–chloroform extraction, concentrated by ethanol precipitation and then analyzed by quantitative real-time PCR using SYBR green. To detect the HPV18 promoter sequence primers HPV18 7826 F (5′-AAATAGGTTGGGCAGCA CAT-3′) and HPV18 163 R (5′-TCCGTGCACAGATCAGGTAG-3′) were used.

Results

Overexpression of E8^∧E2C inhibits growth of HeLa cells

Based on the observation that E8^∧E2C inhibits the activity of the HPV31 E6/E7 promoter in transient transfection studies more efficiently than E2,8 we tested the influence of E8^∧E2C overexpression on the growth of HPV18-positive HeLa cervical cancer cells in a colony reduction assay as described for E2.19, 21, 45 HeLa cells were cotransfected with the pPur plasmid conferring puromycin resistance (1 μg) and the eukaryotic expression vector pSG5 or derivatives (2 μg) expressing HPV31 E2 or E8^∧E2C. Cells were selected and puromycin-resistant colonies were counted after 12–14 days. Figure 1 shows the average colony numbers derived from 3 independent experiments. Overexpression of HPV31 E2 reduced colony formation by 6-fold indicating that the E2 protein derived from HPV31 is able to inhibit growth of HeLa cells as has been described for the E2 proteins derived from BPV1, HPV16 and 18.20, 21 Interestingly, a similar reduction in colony numbers was also seen when E8^∧E2C was overexpressed (Fig. 1) indicating that the E8 repression domain can functionally replace the E2 transactivation domain for the inhibition of HeLa cell growth.20, 21, 22, 23 To analyze whether a functional aminoterminus of E8^∧E2C is required for growth inhibition of HeLa cells, a mutant E8 E2C protein (E8^∧E2C KWK), that is devoid of transcriptional repression activity from promoter-distal E2 binding sites, was also tested in this assay.39 In contrast to E8^∧E2C and E2, E8^∧E2C-KWK was greatly impaired in inhibiting colony formation suggesting an important role for the transcriptional repression domain (Fig. 1). To test whether this could be due to differential levels of expression between E8^∧E2C and the E8^∧E2C-KWK mutant proteins in HeLa cells, HA-tagged versions of the different proteins were generated to facilitate detection. First, the HA-tagged versions of E2 and E8^∧E2C were tested in colony reduction assays along with the untagged proteins. As can be seen in Figure 1, the HA tagged proteins inhibited HeLa growth as efficiently as the corresponding wt proteins. In addition to the repression-negative mutant E8^∧E2C-KWK-HA, a DNA-binding deficient, HA-tagged E8^∧E2C protein was generated (E8^∧E2C-K118M/C119R-HA). The ability of the different HA-tagged proteins to regulate a luciferase reporter plasmid, their expression levels and intracellular localization was tested after transfection into HeLa cells (Fig. 2). As can be seen in Figure 2a, only E8^∧E2C and E8^∧E2C-HA but not E8^∧E2C-KWK, E8^∧E2C-KWK-HA or E8^∧E2C-K118M/C119R-HA were able to inhibit activity from the E2BS-dependent reporter plasmid pC18-SP1-luc. In contrast, E2-HA strongly activated the reporter plasmid as has been described for untagged HPV31 E2.39 Western blot analysis revealed that E8^∧E2C-HA was present at similar levels as E2-HA or the E8^∧E2C-K118M/C119R-HA mutant, whereas E8^∧E2C-KWK-HA protein was present at much higher levels (Fig. 2b). Immunofluorescence studies using an HA-specific antibody also indicated that all mutants are expressed and localized to the nucleus similar to the E2 and E8^∧E2C proteins (Fig. 2c). Taken together these results rule out the possibility that the inability of the E8^∧E2C-KWK mutant to inhibit cell growth is due to reduced expression levels or subcellular mislocalization in HeLa cells.

Figure 2. (a) Modulation of the activity of an E2BS-dependent reporter plasmid by HA-tagged E2 and E8^∧E2C proteins. HeLa cells were transfected with 100 ng of pC18-SP1-luc and 20 ng of expression plasmids as indicated and luciferase activities were determined 48 hr after transfection. The data represent the average of 3 independent transfections performed in duplicate. Luciferase activities are relative to the activity of pC18-SP1-luc in the presence of pSG5, which was set to 1. Error bars indicate the standard deviations. (b) Expression of HA-tagged E8^∧E2C proteins in HeLa cells. HeLa cells were transfected with pSG5 (Lane 1), pSXE2-HA (Lane 2), pSG E8^∧E2C-HA (Lane 3), pSG E8^∧E2C-KWK-HA (Lane 4) and pSG E8^∧E2C-K118M/C119R-HA (Lane 5). Whole cell lysates were prepared and analyzed by immunoblotting with an anti-HA antibody (lower panel) or alpha-tubulin (upper panel). Arrows indicate the positions of E2 and E8^∧E2C proteins. (c) Immunofluorescence analysis of transfected HeLa cells. HeLa cells were transfected with pSX-E2-HA, pSG E8^∧E2C-HA, pSG E8^∧E2C-KWK-HA or pSG E8^∧E2C-K118M/C119R-HA, fixed and stained with DAPI (DNA) to indicate the nucleus and an anti-HA antibody (HA). Pictures were taken with a Zeiss Axiovert 200 M microscope using a 63× objective and the respective fluorescence filter sets for DAPI and Cy3.

To investigate whether the growth inhibition of HeLa cells observed by E8^∧E2C occurs also with HPV-negative C33A cervical carcinoma cells colony reduction assays were performed. C33A cells were cotransfected with pPur and pSG5 or expression plasmids for E2 or the E8^∧E2C proteins and selected for puromycin resistance. In contrast to HeLa cells, all expression plasmids gave rise to similar numbers of puromycin resistant colonies (Fig. 3a). To rule out that the proteins are not expressed or mislocalized in C33A cells and therefore do not inhibit cell growth, immunofluorescence studies of transfected cells using an anti-HA antibody were carried out. Identical to HeLa cells, all proteins can be detected in the nuclei of transfected cells (Fig. 3b).

Figure 3. (a) Colony reduction assay. C33A cells were transfected with a puromycin-resistance plasmid and the empty pSG5 expression vector (vector) or derivates expressing E2-HA, E8^∧E2C-HA, E8^∧E2C-KWK-HA or E8^∧E2C-K118M/C119R-HA. The graph represents the average puromycin resistant colonies derived from 3 independent experiments. Standard deviations are indicated by error bars. Immunofluorescence analysis of transfected C33A cells. C33A cells were transfected with pSX-E2-HA, pSG E8^∧E2C-HA, pSG E8^∧E2C-KWK-HA or pSG E8^∧E2C-K118M/C119R-HA, fixed and stained with DAPI (DNA) and an anti-HA antibody (HA). Pictures were taken with a Zeiss Axiovert 200 M microscope using a 63× objective and the respective fluorescence filter sets for DAPI and Alexa-488. Merged pictures were generated by the AxioVison software 4.3.

E8^∧E2C decreases E6/E7 mRNA levels

Inhibition of HeLa cell growth by E2 is mainly due to the transcriptional repression of the endogenous HPV18E6/E7 promoter, since transduction of HeLa cells with an E6/E7 expression vector that is not regulated by E2, renders HeLa cells resistant to E2.45, 46 To test whether E8^∧E2C inhibits E6/E7 RNA expression, HeLa cells were transfected with the respective expression plasmids and total RNA was isolated 72 hr post transfection. DNAse-treated and reverse-transcribed RNA was analyzed by real time PCR with primer pairs for HPV18E7 and PGK1 as a cellular reference gene. The HPV18E7 primer pair is localised in the E7 gene and will detect both unspliced E6/E7 RNA coding for the E6 protein and spliced E6*/E7 RNA encoding the E7 protein.14, 47 Figure 4 shows that transfection of E8^∧E2C reduces E6/E7 RNA levels 2.5-fold compared to the empty vector pSG5. A similar reduction was observed by transfection of E2 (3-fold). In contrast, neither the DNA-binding deficient mutant E8^∧E2C-K118M/C119R-HA nor the repression domain mutant E8^∧E2C-KWK-HA reduced E6/E7 RNA levels compared to the control (Fig. 4). These data suggested that the growth arrest of HeLa cells by E8^∧E2C is mediated by repressing the activity of the endogenous HPV18E6/E7 promoter and this requires both the sequence-specific DNA binding and the E8 repression domain of E8^∧E2C.

Figure 4. HeLa cells were transfected with pSG5, pSX E2-HA, pSG E8^∧E2C-HA, pSG E8^∧E2C-KWK-HA or pSG E8^∧E2C-K118M/C119R-HA. Total RNA was isolated 72 hr post transfection and analyzed by reverse transcription followed by quantitative real time PCR using primer pairs within the HPV18 E7 gene and for PGK1 as a reference gene. RNA levels are given relative to pSG5 transfected cells and the expression levels of PGK1. The graph represents the average E7 RNA levels derived from 3 independent experiments. Standard deviations are indicated by error bars.

E8^∧E2C induces p53, pRb and CDKN1A/p21 protein levels

We next investigated whether p53 and pRb proteins that are targeted by the HPV18E6 and E7 proteins for degradation are induced by E8^∧E2C expression.8 HeLa cells were transfected with expression vectors for E2-HA and E8^∧E2C-HA proteins and analyzed by double immunofluorescence analyses for the presence of the HA-fusion proteins and p53 or pRb. As can be seen in Figure 5a cells that stain positive for E2-HA or E8^∧E2C-HA are also strongly positive for nuclear p53 or pRb, whereas untransfected cells have no p53 or pRb signal (Fig. 5a). Immunoblot analyses of transfected cells confirmed these findings and revealed that only the expression of E8^∧E2C-HA or E2-HA but not of the DNA-binding deficient mutant E8^∧E2C-K118M/C119R-HA or the repression domain mutant E8^∧E2C-KWK-HA increased p53 and pRB proteins levels (Fig. 5b). This suggested that the reduction of E6/E7 RNA levels by E8^∧E2C is sufficient to induce the reappearance of p53 and pRb in HeLa cells. To further prove that the induction of nuclear p53 protein by E8^∧E2C evokes a physiological response, we analyzed the expression of the p53 target gene CDKN1A/p21 whose expression has been linked to the growth arrest of HeLa cells upon E2 overexpression.23 HeLa cells were transfected with expression vectors for E2-HA, E8^∧E2C-HA, E8^∧E2C-KWK-HA, E8^∧E2C-K118M/C119R-HA or the empty expression vector pSG5. Total RNA was isolated 72 hr post transfection, reverse transcribed and analyzed by quantitative real time for the expression of CDKN1A/p21 using PGK1 as a reference gene. Figure 6a shows that the transfection of E8^∧E2C-HA or E2-HA induced CDKN1A/p21 RNA levels 25- and 13-fold, respectively, relative to empty vector transfected cells. In contrast, both the repression-negative E8^∧E2C-KWK-HA and DNA-binding deficient E8^∧E2C-K118M/C119R-HA mutants displayed no induction of CDKN1A/p21 RNA in line with both mutants being unable to repress E6/E7 RNA levels or induce pRb and p53 proteins (Figs. 4, 5 and 6a). The analysis of whole cell extracts by immunoblotting with anti-p21 and α-tubulin antibodies revealed that the induction of CDKN1A/p21 RNA correlated well with an increase in CDKN1A/p21 protein levels in transfected HeLa cells (Fig. 6b). Only E8^∧E2C-HA and E2-HA but not the mutants were able to induce CDKN1A/p21 protein (Fig. 6b). Taken together, overexpression of E8^∧E2C in HeLa cells results in a down-regulation of E6/E7 RNA which correlates with the reappearance of pRb and p53 and also in the expression of the p53 target gene CDKN1A/p21. Thus, E8^∧E2C induces growth arrest by similar mechanisms as has been described for E2.20, 22, 23, 24, 45, 46, 48

Figure 5. E8^∧E2C and E2 induce p53 and pRb. (a) HeLa cells were transfected with pSX E2-HA or pSG E8^∧E2C-HA and fixed 48 hr post transfection. Cells were stained in the upper panel with a mouse monoclonal antibody specific for p53 and a rabbit polyclonal antibody specific for the HA tag. Bound antibodies were detected with secondary Cy3-conjugated anti-mouse and Alexa-488 conjugated anti-rabbit antibodies. Cells were stained in the lower panel with a rabbit polyclonal antibody specific for pRb and a mouse monoclonal antibody specific for the HA tag. Bound antibodies were detected with secondary Cy3-conjugated anti-mouse and Alexa-488 conjugated anti-rabbit antibodies. Nuclei were visualized by DAPI (DNA) staining. Pictures were taken with a Zeiss Axiovert 200 M microscope using a 63× objective and the respective fluorescence filter sets for DAPI, Cy3 and Alexa-488. (b) HeLa cells were transfected with pSG5 (Lane 1), pSXE2-HA (Lane 2), pSGE8^∧E2C-HA (Lane 3), pSGE8^∧E2C-KWK-HA (Lane 4) or pSG E8^∧E2C-K118M/C119R-HA (Lane 5) and lysed 48 hr post transfection. Cell extracts were analyzed by immunoblotting for p53, pRb or alpha-tubulin with specific antibodies.

Figure 6. (a) Immunoblot analysis of transfected HeLa cells. HeLa cells were transfected with pSG5 (Lane 1), pSX E2-HA (Lane 2), pSG E8^∧E2C-HA (Lane 3), pSG E8^∧E2C-KWK-HA (Lane 4) or pSG E8^∧E2C-K118M/C119R-HA (Lane 5). Whole cell lysates were prepared and analyzed by immunoblotting with an anti-CDKN1A/p21 antibody and an α-tubulin antibody as a loading control. Positions of bands specific for CDKN1A/p21 and α-tubulin are indicated by arrows. (b) HeLa cells were transfected with pSG5, pSX E2-HA, pSG E8^∧E2C-HA, pSG E8^∧E2C-KWK-HA or pSG E8^∧E2C-K118M/C119R-HA. Total RNA was isolated 72 hr post transfection and analyzed by reverse transcription followed by quantitative real time PCR using primer pairs specific for the CDKN1A/p21 gene and for PGK1 as a reference gene. RNA levels are given relative to pSG5 transfected cells and the expression levels of PGK1. The graph represents the average CDKN1A/p21 RNA levels derived from 3 independent experiments. Standard deviations are indicated by error bars.

The E8 domain is required for the binding of E8^∧E2C to the HPV18E6/E7 promoter sequences

Surprisingly, the E8^∧E2C-KWK mutant was unable to suppress cell growth, reduce E6/E7 RNA levels, induce CDKN1A/p21 RNA and protein or pRb and p53 proteins, despite being expressed in HeLa cells at levels even higher than the wt protein (Figs. 1, 2b, 2c and 4–6). To test the influence of the E8^∧E2C-KWK mutation on the HPV18 E6/E7 promoter activity, we constructed an HPV18URR construct (p18URR-luc) in which sequences upstream of the ATG start codon of E6 (7127-105) were cloned into pGL3-basic and cotransfected with the empty expression vector or expression vectors for E8^∧E2C-HA, E8^∧E2C-KWK-HA or E8^∧E2C-K118M/C119R-HA into HeLa cells (Fig. 7a). E8^∧E2C was able to reduce basal activity of the HPV18E6/E7 promoter ∼9-fold, in contrast the E8^∧E2C-KWK protein reduced activity only 2-fold (Fig. 7a). Repression of the reporter construct was dependent on an intact DNA binding domain as the E8^∧E2C-K118M/C119R mutant had no effect suggesting that the E8 domain in the absence of specific DNA binding activity has no effect on HPV18 promoter (Fig. 7a). The impaired inhibition of the transfected and the lack of repression of the endogenous HPV18 promoter in HeLa cells by the E8^∧E2C-KWK mutant protein could either be due to a lack of binding to the promoter region and/or due to a lack of repression activity. To test whether the E8^∧E2C-KWK mutant interacts with the HPV18 promoter in HeLa cells, in vivo DNA binding experiments using chromatin immunoprecipitation assays were carried out. HeLa cells (∼2.8 × 10⁷) were transfected with the empty expression vector, E2-HA, E8^∧E2C-HA or E8^∧E2C-KWK-HA. Consistent with the data shown in Figure 2b, the analysis of cell extracts prepared from a separate plate transfected identically in parallel by immunoblot with the anti-HA antibody used for the chromatin immunoprecipitation revealed similar expression levels for E2-HA and E8^∧E2C-HA and increased levels of E8^∧E2C-KWK-HA (data not shown). Crosslinked DNA–protein complexes from transfected cells were isolated, sheared by sonication and then precipitated with the anti-HA-antibody. Immunoprecipitates were washed extensively, the crosslinks reversed and then analyzed by quantitative real time PCR for the presence of the HPV18 nucleotides 7826-7875/1-163 which encompasses E2BS 2, 3 and 4 and the TATA-box of the early promoter. Figure 7b shows the average relative enrichment of the HPV18 promoter fragment compared to the amount of DNA precipitated by the anti-HA antibody from pSG5 transfected cells. It can be seen that the HPV18 promoter fragment is enriched 5- and 7-fold in E8^∧E2C-HA and E2-HA precipitates, respectively (Fig. 7b). In contrast, no enrichment over background was seen with precipitates obtained from cells transfected with E8^∧E2C-KWK-HA (Fig. 7b). These results indicate that the E8^∧E2C-KWK mutant protein despite being expressed at higher levels than the E8^∧E2C wt protein differs in its ability to interact with the endogenous HPV18 early promoter. Therefore, the most likely explanation for the inability of the E8^∧E2C-KWK mutant protein to inhibit E6/E7 RNA expression and HeLa cell growth is its diminished interaction with the integrated HPV18 promoter region in HeLa cells.

Figure 7. (a) HeLa cells were transfected with 100 ng of p18URR-luc and 20 ng of expression plasmids as indicated and luciferase activities were determined 48 hr after transfection. The data represent the average of 3 independent transfections performed in duplicate. Luciferase activities are relative to the activity of p18URR-luc in the presence of pSG5, which was set to 1. Error bars indicate the standard deviations. (b) Chromatin immunoprecipitation analyses of pSG5, pSX E2-HA, pSG E8^∧E2C-HA and pSG E8^∧E2C-KWK-HA transfected HeLa cells. Crosslinked DNA–protein complexes were precipitated with an anti-HA-tag antibody and analyzed for the presence of the HPV18 fragment encompassing 3 E2BS and the early promoter by quantitative real time PCR. The bars represent the average enrichment of the HPV18 sequences relative to pSG5 transfected cells. The graph shows the average of 3 independent experiments. Standard deviations are indicated by error bars.

Discussion

E8^∧E2C inhibits HeLa cell growth by binding to and inhibiting the HPV18E6/E7 promoter

In this report, we demonstrate for the first time that the naturally occurring E8^∧E2C repressor protein derived from the carcinogenic HPV31 specifically inhibits the growth of HPV18 positive HeLa cells but not of HPV-negative C33a cervical cancer cells. Our side by side analysis of HPV31 E8^∧E2C and E2 proteins demonstrated that both proteins inhibit expression of the HPV18 E6/E7 RNA to a similar extent (Fig. 4). Chromatin immunoprecipitation analyses provide direct evidence that both E2 and E8^∧E2C bind to the P105 sequence in vivo (Fig. 7b). Mutations within the E2C part of E8^∧E2C (E8^∧E2C-K118M/C119R) that prevent sequence specific DNA-binding of E2 proteins,42 abolish repression of the transfected and endogenous HPV18 E6/E7 promoter and also CDKN1A/p21, p53 and pRb induction (Figs. 2, 4–6 and 7a). These data strongly correlate growth inhibition by E8^∧E2C with binding to the HPV18 P105 promoter and repression of the E6/E7 mRNAs. The reduction of E6/E7 mRNA correlates with the reappearance of nuclear p53 protein, the transcriptional induction of the p53-target gene CDKN1A/p21 and also p21 protein and is consistent with a significant reduction of E6 protein levels by E8^∧E2C (Figs. 5 and 6). This appears also to be true for E7 protein levels as expression of E8^∧E2C in HeLa increases Rb protein levels (Fig. 5). Taken together these data strongly suggest that E8^∧E2C inhibits HeLa cell growth by transcriptionally repressing E6 and E7 oncoprotein expression which results in the reactivation of dormant tumor suppressor pathways similar to what has been described for E2.6, 20, 22, 23, 24, 45, 46, 48, 49

Not all reported growth inhibiting properties of E2 proteins require binding to integrated HPV sequences and repression of E6/E7 RNA levels. Overexpression of E2 proteins derived from HPV16 or 31 has been shown to induce cell cycle disregulation or apoptosis markers in some HPV-negative cells.50, 51 Also adenoviral overexpression of HPV16 and 18 E2 proteins fused to GFP induces apoptosis in HeLa cells, which is independent of the DNA binding domain and correlates with the appearance of E2 in the cytoplasm that may point to E2 functions outside from the nucleus.52, 53 However, with the expression system used in this study E8^∧E2C and E2 proteins are exclusively localized to the nucleus (Figs. 2, 3 and 5). Nevertheless, at this point it cannot be excluded that E8^∧E2C has growth modulating properties that are independent of binding to the HPV18 promoter and only becomes evident in other cell types or by using different expression systems. It will also be interesting to determine whether overexpression of E8^∧E2C, in contrast to E2, is also able to inhibit the early promoter on autonomously replicating HPV genomes.54

The E8 repression domain is required for binding to and inhibition of the integrated HPV18 promoter sequences

Surprisingly, the E8^∧E2C-KWK protein in which the E8 repression activity is inactivated by mutation did not inhibit growth of HeLa cells. This is similar to what has been described for the BPV1 E2TR protein lacking the E2 transactivation domain and several E2 mutants with mutations within the aminoterminal domain.20, 21, 22, 23, 29, 31 The E8^∧E2C-KWK mutant did not repress E6/E7 mRNA expression, or induce p53, pRb and CDKN1A/p21 protein levels (Figs. 4–6). In general, it behaved functionally similar to the DNA binding defective mutant E8^∧E2C-K118M/C119R. Chromatin immunoprecipitation experiments revealed that indeed E8^∧E2C-KWK is less active than E8^∧E2C in binding to the endogenous HPV18 promoter DNA in vivo (Fig. 7b). In line with this, E8^∧E2C-KWK was also impaired in inhibiting a transfected HPV18 promoter construct (Fig. 7a). Thus, the most likely explanation for the lack of growth inhibiting activity by E8^∧E2C-KWK is that the E8 repression domain enhances recognition of the HPV18 promoter sequences within cells and this is especially relevant for the integrated HPV18 promoter sequence in HeLa cells.

We have previously shown that the E8^∧E2C-KWK protein and other mutant proteins with impaired repression activity are able to bind to E2BS in gel retardation assays comparable to the wt protein suggesting that mutations in the aminoterminus do not markedly influence DNA binding activity in vitro.39 This is also supported by quantitative studies demonstrating that the purified BPV1 E2 and E2TR proteins have similar DNA binding affinities in vitro which suggests that the aminotermini of E2 proteins are not involved in sequence-specific DNA binding to naked DNA.55 The E8^∧E2C-KWK mutant protein retains the ability to inhibit E2 activity in transcription and replication assays and is able to reduce the activities of transfected HPV18 and 31 E6/E7 promoter constructs, but it was always less efficient than the wt E8^∧E2C protein, which could be explained by a reduced binding to E2BS in vivo.39, 40 It has been described that E2 mutant proteins unable to inhibit the growth of HeLa cells show a reduced binding to the endogenous HPV18 promoter by in vivo footprint analyses.31 A very recent publication describes that the interaction with the cellular Brd4 protein and the HPV11 E2 protein enhances binding of E2 to the endogenous HPV18 promoter DNA and chromatinized templates in vitro and that this interaction is required for promoter repression.32 In line with this, the ability of the HPV16 E2 protein to bind to Brd4 correlates well with growth inhibition of HeLa cells.31, 56 Brd4 has been described as a protein that binds to acetylated histones.57 These data suggest that in addition to an interaction of the E2C DNA binding domain with DNA, an interaction with modified histones via Brd4 mediated by the transactivation domain of E2 might be required for efficient binding to chromatinized templates and a similar mechanism may account for the differences between E8^∧E2C wt and E8^∧E2C-KWK mutant. However, several observations argue against an involvement of Brd4 in the E8^∧E2C mediated repression: First, the Brd4 interaction domain in E2 proteins is completely missing from E8^∧E2C.27, 32, 56, 58 Second, in contrast to E2, E8^∧E2C represses transcription not only from promoter-proximal binding sites but also from promoter-distal binding sites.39 Third, the E8^∧E2C protein acts as a repressor of transcription and replication18, 39, 40 whereas Brd4 has been reported to contribute also to the activation of transcription and replication by E2.32, 56, 59 Therefore, it is likely that the E8 domain interacts with host cell proteins different from Brd4 that may enhance binding of E8^∧E2C to chromatin and also contribute to transcriptional repression. Taken together it can be hypothesized that a common function of the different aminoterminal domains in E2 or E8^∧E2C is to allow for an efficient recognition of chromatinized HPV promoter sequences to achieve repression of the promoter.

Acknowledgements

The authors thank Dr. A. Schepers for technical advice on the chromatin immunoprecipitation experiments.