Mechanisms of Human Papillomavirus E2-Mediated Repression of Viral Oncogene Expression and Cervical Cancer Cell Growth Inhibition

Abstract

The papillomavirus E2 gene product plays a pivotal role in viral replication. E2 has multiple functions, including (i) transcriptional activation and repression of viral promoters and (ii) the enhancement of viral DNA replication. It was previously reported that E2 suppressed the growth of papillomavirus-positive cervical carcinoma cell lines. In the present study, we investigated the mechanisms of E2 growth inhibition. We found that the transcriptional activation function of E2 is required for inhibition of the growth of HeLa cells as well as for transcriptional repression of the viral E6/E7 promoter. It had been previously postulated that transcriptional repression of the E6/E7 promoter results from E2 binding its cognate sites proximal to the E6/E7 promoter and displacing other cellular transcriptional factors. In this study, we report a requirement for the transcription activation function for the binding of E2 to transcriptionally active templates.

The papillomavirus replication cycle is regulated by the viral E2 protein, a sequence-specific DNA binding protein (1, 35, 53). Depending on the promoter context, E2 can act either as a transcriptional activator or as a repressor of viral gene expression. The promoters for E6/E7 gene expression of human papillomavirus type 16 (HPV16) and HPV18 are negatively regulated by E2. This repression is thought to be mediated by the binding of E2 to its recognition sites within the promoter and the displacement of cellular transcriptional factors from the promoter (3, 12, 14, 15, 20, 23, 28, 41, 42, 54, 55, 58, 59, 61, 62). E2 is also involved in the regulation of viral DNA replication through its association with E1, the viral replication factor (36, 50, 63, 65, 66, 67, 68). The conserved N-terminal domain of E2 is required for transactivation (TA), E1 binding, and DNA replication functions. The conserved C-terminal domain forms a dimer and functions as a DNA binding domain. Both conserved domains are linked by a hinged region (reviewed in reference 24).

The loss of E2 expression has been also implicated in the development of HPV-induced carcinoma. Most human cervical carcinoma cells contain integrated HPV DNA and actively express E6/E7 genes (2, 52, 69). The E2 gene is frequently disrupted as a consequence of the integration of the viral genome, and it has been postulated that the loss of E2 somehow contributes to carcinogenic progression (9, 40, 47, 64). E6/E7 genes are invariably expressed in HPV-positive cancers and are considered to be involved in the development of HPV-associated cancers. E6 targets the ubiquitination and proteolysis of p53 through its association with the ubiquitin protein ligase, E6AP (25, 45, 46). E7 binds pRB and inactivates its tumor suppressor function (17, 38). Although E6 and E7 may have additional functions and cellular targets, it is believed that their inactivation of these important tumor suppressor proteins is critical for HPV-associated carcinogenesis. As mentioned above, E2 has the ability to suppress E6/E7 expression; thus, disruption of the E2 gene results in the deregulated expression of E6 and E7 and thus contributes to carcinogenic progression (24). This model is also supported by the finding that mutations in the E2 gene increased the immortalization activity of HPV16 DNA for human primary keratinocytes (43).

The significance of the disruption of the E2 gene in cervical carcinoma cells has been investigated by the reintroduction of E2 into HPV-positive carcinoma cell lines such as HeLa cells, which contain integrated HPV18 DNA and actively express E6/E7 (4, 49). The expression of E2 protein in HeLa cells markedly inhibits E6/E7 expression and inhibits cell proliferation (16, 26, 27, 62). The suppression of cell growth was also observed in other HPV-positive cancer cell lines, including SiHa and Caski, but not in HPV-negative cervical carcinoma cells such as C33A cells (16). These results suggested that the growth inhibition by E2 was specific for HPV-positive carcinoma cells and that the suppression of E6/E7 expression contributed to this process.

To further investigate the functions of the E2 protein in the growth inhibition of HPV-positive carcinoma cells, we tested the activities of various E2 mutants derived from HPV16 E2. We found a clear correlation between the TA and growth suppression activities of E2. The TA function of E2 was required for efficient repression of the E6/E7 promoter. Furthermore, results of in vivo footprinting and other functional assays strongly suggested that the TA function was required for the efficient binding of E2 to a transcriptionally active DNA target in vivo in order to affect the displacement of cellular factors from the E6/E7 promoter. The involvement of a TA function of some transcription factors in the binding to DNA forming chromatin structures has been reported previously (5, 34). It is possible that an intact E2 TA function may be required for interaction with specific cellular factors leading to a remodeling of the chromatin of transcriptionally active templates.

MATERIALS AND METHODS

Cell culture and transfection.

HeLa and CV1 cells were maintained in Dulbecco's modified Eagle medium supplemented with 10% heat-inactivated fetal bovine serum. DNA transfection was done by a standard calcium phosphate precipitation method (31). The cells (1.5 × 10⁵) were seeded in a 6-cm-diameter dish 1 day before transfection. Plasmid and carrier DNA (total of 10 μg) were incubated with 500 μl of HEPES-buffered saline transfection buffer (140 mM NaCl, 0.75 mM Na₂HPO₄, 25 mM HEPES, 110 mM CaCl₂ [pH 6.90]) for 30 min at room temperature and then added to a culture dish; 20 h after transfection, cells were washed once with phosphate-buffered saline (PBS) and fresh growth medium was added. All analyses except the growth suppression assay were performed with the cells 2 days after transfection.

Plasmid preparations.

We have described construction of the original HPV16 E2 expression plasmid in a previous report (44). FLAG-tagged E2 expression plasmids were constructed from those original E2 expression plasmids. The FLAG tag sequence was added at the 5′ end of the E2 gene and recloned into pCMV₄ expression vector (10). The reporter plasmid p6xE2BStkCAT was described elsewhere (60). HPV16 and HPV18 replication origin (ori)-containing plasmids were constructed by inserting the DNA fragments amplified by a PCR into pCAT-basic (Promega Corp., Madison, Wis.). LCR16F contains the HPV16 long central region (LCR) fragment spanning nucleotides (nt) 7003 to 100 (GenBank accession no. K02718). LCR18F-CAT and LCR18S-CAT contain the HPV18 LCR fragments from nt 7000 to 110 and 7810 to 110, respectively (GenBank accession no. X05015). The sequences of the PCR-amplified portions of the final plasmids were confirmed by direct sequencing not to have any mutation. pSVβ was purchased from the manufacturer (Clontech Laboratories, Inc., Palo Alto, Calif.).

Gel-shift analysis of E2 DNA binding capacity.

Nuclear extracts were prepared from transfected HeLa cells (48). The transfected cells (ca. 5 × 10⁵ cells/6-cm-diameter plate) were washed once with phosphate-buffered saline (PBS), harvested into 1.5-ml microtubes, resuspended in 400 μl of buffer A (10 mM HEPES [pH 7.9], 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM dithiothreitol [DTT], benzamidine HCl [16 μg/ml], phenanthroline [10 μg/ml], aprotinin [10 μg/ml], leupeptin [10 μg/ml], pepstatin A [10 μg/ml], 1 mM phenylmethylsulfonyl fluoride [PMSF]) for 20 min on ice, mixed with 25 μl of 10% NP-40, and centrifuged at 2,000 rpm for 5 min at 4°C. The nuclear pellets were resuspended in 50 μl of buffer C (20 mM HEPES [pH 7.9], 0.4 mM KCl, 1 mM EDTA, 1 mM EGTA, 1 mM DTT, benzamidine HCl [16 μg/ml], phenanthroline [10 μg/ml], aprotinin [10 μg/ml], leupeptin [10 μg/ml], pepstatin A [10 μg/ml], 1 mM PMSF) for 20 min on ice and centrifuged at 12,000 rpm for 5 min at 4°C. The protein concentration of the supernatant was quantitated and used as the nuclear extract.

Ten micrograms of nuclear extract was incubated in 10 μl of binding reaction mixture [20 mM HEPES (pH 7.9), 5 mM MgCl₂, 100 mM KCl, 0.2 mM EDTA, 10% glycerol, 5 mM DTT, d(I-C) [1 μg/μl], ³²P-labeled probe (5 fmol/μl), benzamidine HCl (16 μg/ml), phenanthroline (10 μg/ml), aprotinin (10 μg/ml), leupeptin (10 μg/ml), pepstatin A (10 μg/ml), 1 mM PMSF] for 30 min at room temperature. The ³²P-labeled DNA probe was obtained by annealing two oligonucleotides (5′-GGT AAC CGA AAC CGG TTA-3′ and 5′-GGT AAC CGG TTT CGG TTA-3′), and [α-³²P]dCTP was incorporated by fill-in reaction with Klenow fragment DNA polymerase. The reaction mixture was then loaded on 8% acrylamide gel in 4⁻¹× Tris-borate-EDTA buffer. Electrophoresis was carried out at 200 V for 2 h. The image was obtained by exposing the gel to X-ray film.

Northern blot analysis and immunoblotting.

The cytoplasmic fraction of the transfected cells was obtained as a supernatant of buffer A-treated samples as described above. Cytoplasmic RNA was extracted by the acid guanidinium-phenol-chloroform method (8), and poly(A)⁺ RNA was enriched with oligo(dT)-cellulose beads (New England Biolabs Inc., Beverly, Mass.). One microgram of poly(A)⁺ RNA was separated through a formalin-containing 1.2% agarose gel and then transferred to a Hybond-N+ nylon membrane (Amersham Pharmacia Biotech UK Ltd., Little Chalfont, United Kingdom). Northern hybridization was performed according to a standard protocol (51). The probes used in this analysis were the HPV16 E2 DNA fragment, p53 cDNA, and the E6/E7 coding region of HPV18 labeled by a random prime labeling kit (TAKARA, Kusatsu, Japan).

The whole cell extract was prepared from transfected cells with extraction buffer (250 mM NaCl, 20 mM sodium phosphate buffer [pH 7.0], 30 mM sodium pyrophosphate, 5 mM EDTA, 10 mM NaF, 5 mM DTT, 0.1% NP-40, benzamidine HCl [16 μg/ml], phenanthroline [10 μg/ml], aprotinin [10 μg/ml], leupeptin [10 μg/ml], pepstatin A [10 μg/ml], 1 mM PMSF); 400 μl of extraction buffer was directly added to the culture dish and incubated for 20 min at 4°C with rocking. Then the cell lysate was transferred to a microcentritube and centrifuged at 15,000 rpm for 10 min at 4°C. The supernatant was used as the whole cell extract. Twenty micrograms of cell extract was applied to a polyacrylamide gel containing 10% sodium dodecyl sulfate, transferred to a Hybond-P polyvinylidene difluoride membrane, blocked in 5% dry fat-free milk–0.1% Tween 20–PBS (PBS-T), and incubated with a 1/1,000 dilution of anti-FLAG monoclonal antibody M5 (Sigma, St. Louis, Mo.) in 2% dry fat-free milk–PBS-T for 1 h. The filter was washed three times with PBS-T, incubated for 1 h with a 1/3,000 dilution of horseradish peroxidase-labeled goat anti-mouse immunoglobulin G monoclonal antibody (Amersham Pharmacia Biotech UK), subjected to four more washes with PBS-T, and visualized with the enhanced chemiluminescence detection reagent (Amersham Pharmacia Biotech UK).

TA and transrepression assays.

HeLa cells in 6-cm-diameter dishes were transfected with 1 μg of pCMV-E2, 1 μg of p6xE2BStkCAT or LCR18F-CAT, 1 μg of pSVβ, and 7 μg of herring sperm (HS) DNA. Two days after transfection, the bacterial chloramphenicol acetyltransferase (CAT) activity in the cell extract was measured by a standard protocol. The β-galactosidase activity from pSVβ was monitored by o-nitrophenylglycoside assay and used for normalizing the transfection efficiency of each sample. To obtain the transfection efficiency, the transfected cells were fixed with 0.25% glutaraldehyde–PBS and stained with 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal) by a standard protocol (6).

Growth suppression assays.

A total of 2 × 10⁵ HeLa cells transfected with 2 μg of pCMV-E2, 0.5 μg of pSV2neo, and 7.5 μg of HS DNA were cultured in growth medium supplemented with G418 (0.5 mg/ml). The surviving colonies were enumerated 14 days after transfection.

In vivo footprinting of the HeLa cell genome.

In vivo footprinting in HeLa cells transfected with 1 μg of pCMV-E2, 1 μg of pSVβ, and 8 μg of HS DNA was performed by following a standard protocol (37). The HeLa cells were treated with 0.1% dimethyl sulfate (DMS)-growth medium for 2 min, and then the genomic DNA was harvested by the proteinase K-phenol extraction method. In a parallel experiment, genomic DNA was obtained from mock-transfected HeLa cells without DMS treatment and then treated with 0.1% DMS in vitro (the sample indicated as naked). The DMS-treated DNA was cleaved with 1 M piperidine for 30 min at 90°C, purified, and used for ligation-mediated PCR (LM-PCR). The primers used in the LM-PCR were P1 (5′-GCA GTG AAG TGT TCA GTT CCG TGC ACA-3′), P2 (5′-GGT AGC TTG TAG GGT CGC CGT GTT GG-3′), and P3 (5′-GGT AGC TTG TGA GGT CGC CGT GTT GGA TCC TC-3′), LM PCR-1 (5′-GCG GTG ACC CGG GAG ATC TGA ATT C-3′), and LM PCR-2 (5′-GAA TTC AGA TC-3′). The annealing temperature was 60°C for P1, 63°C for P2, and 65°C for P3. P3 primer was end labeled with ³²P, and the products were visualized with a image analyzer (BAS-2000; Fuji Film, Tokyo, Japan).

Transient DNA replication assays of HPV16 and HPV18 ori-containing plasmids.

A transient DNA replication assay was performed as described previously (44, 67, 68). HPV16 ori- and HPV18 ori-containing plasmids (1 μg of each) were used for transfection into CV1 and HeLa cells, respectively, with 1 μg of pCMV-E2, 3 μg of pCMV-E1, and 5 μg of HS DNA. The extrachromosomal DNA was extracted from the transfected cells by Hirt's method, digested with BamHI and DpnI, and then analyzed by Southern blotting with a ³²P-labeled CAT gene probe.

RESULTS

Relationship between TA and growth suppression activities of E2.

We previously reported a structure-function analysis of the HPV16 E2 N-terminal domain TA and DNA replication activities (44). Since growth inhibition of HPV-positive carcinoma cell lines also requires a function mapping to the N-terminal domain, we used a subset of the E2 mutants described in our previous report to examine whether the TA and/or DNA replication functions were required for suppression of cell growth. Four E2 mutants that had clear phenotypic characteristics were used in this analysis: R37A and I73A, defective in TA function; E39A, which maintains wild-type (wt) TA activity but fails to support DNA replication because of insufficient association with E1; and L79A, with wt TA and DNA replication activities. The wt and mutant E2 proteins were tagged at the N terminus with the FLAG epitope for convenience of detection. We confirmed that the addition of the epitope tag did not affect the phenotypes of the original E2 mutants. The tagged E2 mutants were expressed at comparable levels in the transfected HeLa cells (Fig. 1A), and their nuclear localization was confirmed by immunostaining (data not shown). To verify the DNA binding capacity of the FLAG-tagged E2 proteins, a gel shift assay with a probe containing a single E2 binding site was performed with both in vitro-translated E2 (data not shown) and the nuclear extracts of the HeLa cells transfected with the E2 expression plasmids (Fig. 1B). The results demonstrated that neither the N-terminal tag nor any of the specific alanine substitution mutations significantly affected the expression, stability, subcellular localization, or DNA binding capacity of E2.

FIG. 1.

Expression of epitope-tagged HPV16 E2 proteins. (A) Total cell extracts and nuclear extracts were prepared from HeLa cells 2 days following transfection with the epitope-tagged E2 expression plasmids. The E2 proteins were detected with anti-FLAG antibody M5. The positions of E2 proteins and molecular weight markers are indicated with open and closed arrowheads, respectively. (B) A gel shift assay with the single E2 binding site probe was performed with the nuclear extracts. The three left-hand lanes contain probe only (probe), probe plus nuclear extract from mock-transfected cells (control), and the extract from the cells transfected with 3 μg of wt E2 expression plasmid plus probe with a 50-fold excess of cold probe as a competitor (wt+competitor). Each of the other pairs of lanes contain the probe plus the nuclear extract from HeLa cells transfected with either 1 or 3 μg of the indicated E2 expression plasmid. Black and white arrowheads indicate the positions of E2-DNA complex and free probe, respectively.

The phenotype of the TA activity of each tagged E2 mutant was identical to that previously described by us for the untagged E2 proteins (44) (Fig. 2A). The levels of TA activities were comparable between the tagged and untagged E2 (data not shown). In addition, there was a strong correlation between the TA and growth suppression activities of the E2 mutants, suggesting that the TA function may be required for efficient growth inhibition of HeLa cells (Fig. 2B). The number of colonies observed with the I73A mutant was nearly the same as that in the control plate, but the colonies transfected with I73A were smaller than those on the control plates, suggesting that the I73A mutant may have retained some residual growth inhibition activity. We also examined the relationship between TA and growth inhibition activities with other mutant E2s described in our previous report (44) and found a clear correlation (data not shown). As shown previously (16), a requirement for an activity mapping to the N-terminal domain of E2 for growth inhibition was demonstrated with bovine papillomavirus type 1 (BPV1) E2TA and E2TR proteins and with a chimeric protein having the VP16 TA domain. Using N-terminal domain mutants of BPV E2, Goodwin et al. reported a similar correlation (21), suggesting this is a conserved feature of papillomavirus E2 proteins.

FIG. 2.

Comparison of TA and growth suppression activities of E2 proteins. For the TA assay (A), HeLa cells were transfected with 2 μg of effector expression plasmid and 1 μg of the p6xE2BS-CAT reporter plasmid. CAT assays were performed 2 days after transfection. HeLa cell growth suppression (B) was determined by the colony formation assay. Percent suppression is expressed as (number of colonies on control plate − number of colonies on sample plate)/(number of colonies on control plate) × 100. In these experiments, the number of colonies on the control plates ranged from 1.2 × 10³ to 3.0 × 10³. BPV-TA, BPV-TR, and VP16-E2 plasmids have been described previously (see text) and were not FLAG tagged. Data represent the average of three independent experiments, and standard deviations are indicated as error bars.

Repression of E6/E7 transcription by E2 in HeLa cells.

E2 can suppress E6/E7 expression from the HPV18 P₁₀₅ promoter and from the analogous promoter in HPV16, P₉₇ (3, 15, 20, 42, 61, 62). Since the HPV E6 and E7 genes encode the major transforming activities for the high-risk-type HPVs and are invariably expressed in HPV-positive cancers, we tested whether there was a correlation between the transcriptional repression of the E6/E7 promoter and the growth inhibition by E2. HeLa cells contain approximately 10 copies of HPV18 DNA integrated into host chromosomal DNA and actively express the E6 and E7 genes. We investigated the transcriptional repression of the P₁₀₅ promoter of the integrated HPV18 genome in HeLa cells by E2. HeLa cells were therefore transfected with an E2 expression plasmid and analyzed for the expression of E6/E7 mRNA. The transfection efficiency was over 80%, as estimated by X-Gal staining (see Materials and Methods). The levels of the E6/E7 mRNA were markedly diminished by the TA-competent E2 mutants (E39A and L79A) but not the TA-defective E2 mutants (R37A and I73A) (Fig. 3). This result indicated a strong correlation between the TA function of E2 and its ability to efficiently repress E6/E7 gene expression from the integrated HPV18 genomes in HeLa cells.

FIG. 3.

Effects of HPV16 E2 mutants on endogenous HPV18 E6/E7 mRNA expression and on p53 protein levels. HPV18 E6/E7, p53, and E2 mRNAs in the E2-transfected HeLa cells were measured by Northern blot analysis 48 h after transfection. The p53 protein was detected by Western blotting using the DO-1 antibody. An extract from C33A cells was used as a control. The TA capacity of each E2 mutant is indicated at the bottom.

We also examined the consequence of E6/E7 mRNA repression by E2 in HeLa cells by examining p53 mRNA and protein levels. Similar to what has been described for BPV E2 in HeLa cells (16, 26), the amount of p53 protein increased in the HeLa cells expressing TA-competent HPV16 E2 proteins (Fig. 3). The amount of p53 mRNA was not affected by E2 expression, indicating that the modification of the expression level of p53 protein was a posttranscriptional effect. In addition, as previously described with BPV E2, we observed an increase of p21 protein and the downregulation of Cdk activity as downstream effects of p53 activation with the TA-competent E2 expression (data not shown). These results suggested that the TA function of E2 was connected to the growth inhibition of HeLa cells via transcriptional suppression of E6/E7 expression.

An intact TA domain is required for E2 repression of the P₁₀₅ promoter in a DNA template-dependent manner.

We analyzed the transrepression activities of E2 mutants with a CAT reporter plasmid in order to investigate the role of TA function in transcriptional repression. Two different reporter plasmids, LCR18F-CAT and LCR18S-CAT, were used in this experiment. LCR18F-CAT contains the entire LCR directing CAT expression from the P₁₀₅ promoter, but LCR18S-CAT has a deletion of all LCR sequences upstream of nt 7800, just upstream of the HPV18 ori region (Fig. 4A). The latter plasmid contains the core P₁₀₅ promoter elements but is devoid of the upstream enhancer element, resulting in a 95 to 98% loss of transcriptional activity compared to the LCR18F-CAT plasmid in HeLa cells (data not shown).

FIG. 4.

Transcriptional repression of LCR18F-CAT by E2. (A) Schematic presentation of the structure of HPV18 LCR. Data for transcription factors bound on the LCR are based on other reports (11, 39). The LCR portions inserted in LCR18F-CAT and LCR18S-CAT are indicated. (B) HeLa cells were transfected with 2 μg of the indicated effector expression plasmid and 1 μg of LCR18F-CAT. Mock-transfected cells were used as control. Two days after transfection, CAT activity in the cells was determined as described in Materials and Methods. The data represent the average of four independent experiments, and standard deviations are indicated as error bars.

Wild-type E2 strongly suppressed CAT expression from the P₁₀₅ promoter of LCR18F-CAT as did the TA-competent mutants, E39A and L79A. On the other hand, the TA-defective mutants, R37A and I73A, failed to suppress the P₁₀₅ promoter, indicating a correlation between an intact TA function with repression of the P₁₀₅ promoter as observed with the integrated HPV18 genome in HeLa cells (Fig. 4B). A requirement for an intact E2 TA function was also noted for repression of a CAT reporter plasmid containing the entire HPV16 LCR in which CAT is expressed from the P₉₇ promoter that directs expression of E6 and E7 of HPV16 (data not shown). No repression was obtained in the case of VP16-E2 chimeric protein, suggesting that a function specific to the E2 TA domain is involved in the E6/E7 promoter suppression.

We could demonstrate about twofold repression of the LCR18S-CAT reporter plasmid by wt HPV18 E2 (Fig. 5). Interestingly, each of the E2 mutants assayed could also repress expression from the LCR18S-CAT reporter approximately twofold. This was true for the TA-competent E2 mutants (E39A and L79A) as well as for the E2 mutants (R37A and I73A) which are impaired in TA function. Thus, the modest E2 repression of the HPV18 P₁₀₅ promoter devoid of the LCR upstream elements does not require an intact TA domain.

FIG. 5.

Transcriptional repression of LCR18S-CAT by E2. HeLa cells were transfected with LCR18S-CAT and E2 expression plasmids. Two days after transfection, cell extracts were prepared and CAT activity was assayed by a standard method. The CAT activity of mock-transfected cells was taken as 100% (control). The data represent the average of three independent experiments, and standard deviations are indicated as error bars.

The E2 TA function is involved in E2 promoter binding in HeLa cells.

It had been postulated that the negative regulation of P₁₀₅ activity by E2 was due to the binding of E2 to its cognate sites and the displacement of cellular transcription factors such as SP1 from the initiation complex. This model predicts that the DNA binding activity of E2 would then be sufficient for the P₁₀₅ suppression. Each of the E2 proteins tested in our study maintained the DNA binding capacity by in vitro assay (Fig. 1B), yet R37A and I73A failed to suppress P₁₀₅ activity of the integrated HPV18 genome and LCR18F-CAT plasmid (Fig. 3 and 4B). One possible explanation for our findings is that an intact TA function may be required for E2 binding in vivo, and the in vitro assay used may not reflect the true situation in HeLa cells. Since the LCR18F-CAT plasmid is a transcriptionally active template in HeLa cells, it may more closely resemble the chromatin structure of the integrated HPV18 genomes.

To address the role of the intact TA function in the binding of E2 to its cognate sites in the P₁₀₅ promoter, we performed in vivo footprinting of the integrated HPV18 LCR, using HeLa cells transfected with E2 expression plasmids. As depicted in Fig. 6, E2 binding sites proximal to P₁₀₅ were protected by wt E2 and the TA-competent E2 mutants E39A and L79A. In contrast, TA-defective mutants R37A and I73A were impaired in the ability to protect. This experiment has been repeated five times and is highly reproducible. These results indicate an intact E2 TA function is required for the efficient binding to DNA in the context of the chromosome in vivo. It is noteworthy that even in the in vivo situation, the E2 TA function is not required for transcriptional suppression of the LCR18S-CAT reporter (Fig. 5), suggesting that the TA requirement for DNA binding is dependent on the transcriptional activity of the DNA template. LCR18S-CAT may reflect the DNA status of in vitro binding assay because it does not form a complex structure for active transcription.

FIG. 6.

In vivo footprinting analysis with the integrated copies of HPV18 genomes. In vivo footprinting was done with HeLa cells transfected with E2 expression plasmids. Two days after transfection, DNA was prepared and transfection efficiency was monitored in a parallel experiment. Over 80% of cells appeared to be positive for β-galactosidase expression. The corresponding sequence of the promoter region of HPV18 is shown at the top, and two promoter-proximal E2 binding sites (E2BS-1 and E2BS-2) are indicated. Positions of G residues protected by TA-competent E2 binding are indicated with dots. Lanes: naked, purified genomic DNA of HeLa cells used for the substrate; control, mock-transfected cells.

Although the in vivo footprinting experiment was useful for revealing the protein-DNA association in a more physiological state, the changes in the observed protection pattern could still either be a direct or an indirect consequence of E2 expression. We therefore extended our analysis to examine whether the TA activity was required for the E2 binding to DNA in vivo, using a DNA replication assay that employed an origin in the context of the full-length LCR.

The requirement for E2 TA function for DNA replication is dependent on the template.

DNA replication of an HPV ori-containing plasmid requires both E1 and E2 (7, 10, 63). Although E1 is the major DNA replication factor for the papillomaviruses, it has little affinity by itself for binding to the origin and must be recruited through its protein-protein interaction with E2 (29, 65, 66, 68). Therefore, the critical role of E2 as an auxiliary factor in transient ori-dependent DNA replication depends on its ability to bind DNA as well as to form a complex with E1. We therefore used a DNA replication assay utilizing the LCR18F-CAT reporter plasmid to assess the role of the TA function of E2 in augmenting E1-dependent DNA replication. In this experiment, both wt and L79A mutant E2s supported the DNA replication of the HPV18 (LCR18F) plasmid efficiently, whereas the TA-defective R37A and I73A mutants had reduced activity for DNA replication (Fig. 7, LCR18F). This experiment revealed a role for the E2 TA function in E1-mediated ori-dependent DNA replication, presumably by permitting more efficient E2-DNA binding in vivo. The E2 mutants tested in this experiment are known to have comparable levels of E1-E2 binding activity (44). The E39A mutant of E2 that is defective in forming a complex with E1 served as a negative control. The requirement for an intact TA domain to support E2-dependent DNA replication efficiently was also observed with a corresponding HPV16 ori-containing construct, LCR16F-CAT (Fig. 7).

FIG. 7.

Involvement of the TA activity of E2 in DNA replication. Transient DNA replication was performed with an HPV18 ori-containing plasmid, LCR18F-CAT or LCR18S-CAT, in HeLa cells. LCR16F-CAT was used for the transient DNA replication assay in CV1 cells. Note that LCR16 is active for transcription in CV1 cells. Signals of replicated DNA are indicated with black arrowheads, and those of input DNA are marked with white arrowheads. Positions of DNA size markers are indicated at the left.

In the transcriptional repression experiment with LCR18S-CAT, we found TA function-independent suppression of P₁₀₅ promoter activity, suggesting that TA function is not required for DNA binding with transcriptionally inactive templates (Fig. 5). We tested the DNA replication activity with the LCR18S-CAT reporter plasmid. In contrast to LCR18F-CAT, LCR18S-CAT could be replicated equally with each of the E2 mutants tested except E39A, which served as the negative control (Fig. 7). The result of the DNA replication assay suggested the requirement of the TA function of E2 for DNA binding was dependent on the transcriptional activity of the DNA template.

As we have previously reported, there is no requirement of TA function for the E2-enhanced replication of a plasmid containing only the HPV16 ori region without the upstream LCR enhancer sequences, indicating that E1 binding is the major function of the E2 TA domain in DNA replication and that the TA activity of E2 is dispensable for ori-dependent DNA replication (44). Similar observations have been reported by other groups who studied DNA replication analysis with papillomavirus E2 mutants (18, 22). These previous reports are consistent with the result shown here with LCR18S-CAT. In contrast, the TA function of E2 was required for the E2-enhanced DNA replication of the LCR18F-CAT template that contains the full-length LCR and that is active for transcription from the P₁₀₅ promoter in HeLa cells. The results obtained from the transient DNA replication assay showed that the TA function of E2 was vital for the binding of E2 to the transcriptionally active promoter region.

DISCUSSION

Mechanisms of E2-mediated growth inhibition.

In this study, we investigated the molecular mechanisms by which HPV16 E2 protein inhibits the growth of HeLa cells. We have previously reported that growth inhibition by E2 is specific for HPV-positive carcinoma cell lines (16). The continuous expression of E6 and E7 genes from integrated HPV DNA is a hallmark in the majority of HPV-positive carcinomas and in cell lines derived from them. E6 and E7 are considered to play central roles in the progression to and maintenance of cancers. There is considerable information on the mechanisms of the carcinogenesis induced by the expression of high-risk-type E6 and E7 proteins (24). Both E6 and E7 disturb the G₁ and G₂ checkpoints, resulting in the accumulation of genetic mutations, which can culminate eventually in the progression to full malignancy.

HPV16 E2 suppressed the expression of E6 and E7 in HeLa cells, and we demonstrate here a clear correlation between the downregulation of E6/E7 expression and cell growth arrest. Our results are in agreement with previous studies that have also suggested that the expression of E6 and/or E7 is indispensable for the continued proliferation of carcinoma cells. The repression of E6/E7 expression by E2 reactivates the cell cycle checkpoints controlled by p53 and pRB. We have not yet determined the relative contributions of the p53 and pRB pathways to growth arrest in HeLa cells. In addition to their well-characterized roles in regulating the pRB and p53 pathways, E6 and E7 each likely have other important functions in cellular transformation and immortalization. It is anticipated that E2 repression of E6 and E7 expression in HPV-positive cancer cells could also lead to the activation of the pathways governed by these additional targets. For instance, E6 has been shown to upregulate cellular telomerase activity (32). The downregulation of E6 by E2 would presumably suppress the telomerase activity, and the subsequent detection of critical short telomeres might then signal cell growth arrest and/or senescence. There have been several reports describing the associations of E6 and E7 with cellular factors, and it will be interesting to explore the functional correlation between these associations and growth inhibition. It is also possible that functions of E2 independent of its ability to repress expression of E6/E7 could be involved in the growth inhibition of the cervical carcinoma cells. For instance, there may be cellular genes whose expression is either activated or repressed by E2, or there may be cellular factors whose activities are somehow modulated by E2. Additional experiments will be necessary to determine whether repression of E6/E7 by E2 is indeed sufficient to induce growth inhibition in HeLa cells.

Correlation between the TA and growth inhibition functions of E2 protein.

It has previously been reported that both the TA and DNA binding domains of BPV1 E2 protein are required for the growth inhibition effect (16). A detailed analysis using a panel of E2 proteins with mutations in their transactivation domains was performed with BPV1 E2 by Goodwin et al., who found that the TA function was required for growth inhibition (21). Our results confirm their data and extend them to the HPV16 E2 protein.

Transcriptional repression of the E6/E7 promoter is believed to occur via the binding of E2 to its recognition sites proximal to the E6/E7 promoter and displacement of cellular transcriptional factors necessary for assembly or activation of the initiation complex. It has also been thought that the DNA binding activity of E2 was sufficient for this transcriptional repression (13). We found here that the TA function is also required for the transcriptional repression of P₁₀₅ by E2, in accordance with the published results of Goodwin et al. (21). Desaintes et al. reported that E2TR of BPV1, which lacks the N-terminal transactivation domain, could suppress the HPV18 P₁₀₅ promoter as efficiently as E2TA (13). In our experiments using LCR18F-CAT, we observed a requirement for the N-terminal transactivation domain of BPV1 E2 for repression of the P₁₀₅ promoter (Fig. 4). The results obtained using BPV1 E2 by Goodwin et al. (21) also support an involvement of TA function in P₁₀₅ repression. We must point out that our data are in disagreement with those of Desaintes et al., but a recent study by Francis et al. also could not reproduce the observations by Desaintes et al. that the BPV E2TR can repress the endogenous HPV18 E6/E7 promoter in HeLa cells (19).

Involvement of the TA activity in the repression of the E6/E7 promoter.

The requirement of TA function in transcriptional repression of E6/E7 expression in HeLa cells raised the possibility that the TA function of E2 is involved in DNA binding in vivo. To test this hypothesis, in vivo footprinting was used with the integrated HPV18 genome in HeLa cells. Although HeLa cells harbor approximately 10 copies of HPV18 DNA in the genome, protection of the E2 binding sites was observed with wt and TA-competent E2 mutants. No such protection was seen with TA-defective mutants, indicating that the TA function of E2 was required for its efficient binding to the integrated P₁₀₅ promoter sequence in HeLa cells. All of the E2 mutants tested here exhibited similar DNA binding activities in vitro. Thus, the requirement of the TA activity for DNA binding was specific for the in vivo situation of the integrated HPV18 genomes in HeLa cells. These results support our model, although they do not rule out the possible importance of cellular factors interacting with a TA-competent E2 protein to affect the repression of the P₁₀₅ promoter.

E2 is known to participate in papillomavirus DNA replication in association with viral replication factor E1. The function of E2 in DNA replication is thought to be the recruitment of E1 to the ori through the E2 binding sites located adjacent to the ori. We and others have reported that the TA activity of HPV16 E2 is dispensable for transient DNA replication; however, these previous studies used a minimal ori in the replication assay (18, 44). Utilizing a replication template that includes the entire HPV18 LCR (or the entire HPV16 LCR), we now see a requirement for an intact TA function for efficient DNA replication. Thus, there are elements within the LCRs of the HPV16 and HPV18 genomes that require the TA function of E2 in addition to its ability to complex E1 to support ori-dependent DNA replication. It seems likely that these elements are the enhancer sequences within the LCRs that confer strong transcriptional activity to the templates themselves. In the transient DNA replication assay, the amplification of transcriptionally active templates was dependent on the TA activity of E2, but the amplification of templates with low transcriptional activities (LCR18S and LCR16S) was not. Therefore, the TA function of E2 was not required for DNA binding in the case of templates of low transcriptional activity even in vivo (Fig. 7). There was a report that the TA function of E2 was dispensable for the episomal DNA replication of HPV31 in human foreskin keratinocytes (56). Although we do not know why our result conflicts with this finding, some assay conditions might be critical for TA dependency of HPV DNA replication. The correlation between TA function and DNA binding activity of E2 can be made with repression of the HPV18 P₁₀₅ promoter. Using a P₁₀₅ reporter plasmid with a low level of transcriptional activity and devoid of the LCR enhancer elements, all of the HPV16 E2 mutants (regardless of their TA function status) could repress transcription approximately twofold, indicating that the TA function of E2 was not required for the binding to E2 binding sites on the transcriptionally silent template. Both the DNA replication assay and the P₁₀₅ repression assay indicated that the requirement for an intact E2 TA function was dependent on the transcriptional status of the DNA templates (Fig. 8).

FIG. 8.

TA activity of E2 is involved in DNA binding to a transcriptionally active template (model). The requirement of E2 TA activity for DNA binding is dependent on the state of DNA template. With transcriptionally active DNA template (left), the TA activity is required for DNA binding of E2. With transcriptionally silent template (right), the DNA binding domain may be sufficient for DNA binding of E2.

The TA domains of several transcription factors have been reported to enhance sequence-specific binding to chromatin (5, 34). Physical interaction between the TA domain of transcription factors and chromatin remodeling factor(s) is supposed to play an important role in enhancing this DNA binding (30, 33). It is possible that elements within the LCR that affect the transcriptional activity of the HPV16 and HPV18 templates used in our study may also affect chromatin structure. The chromatin organization of the HPV LCR is supposed to have an important regulatory function in the viral life cycle (57). Although transiently transfected DNA is not thought to form well-organized chromatin structures, it is possible that the TA-competent E2 mutants could associate with some cellular factors that are involved in the chromatin remodeling activity. The short LCR constructs containing the ori and flanking E2 binding sites and core promoter sequences may more closely resemble the naked DNA template used in in vitro assays, but the more transcriptionally active templates containing the full LCR might form a more complex structure with cellular DNA binding proteins including histones and transcriptional factors. The cellular factor(s) associated with E2 TA activity might then be required for the remodeling of the complex DNA template to facilitate the DNA binding of E2. This model would account for our results demonstrating a requirement for a functional E2 TA domain for binding to the E2 binding sites of the integrated HPV18 promoter sequences in HeLa cells (Fig. 6) and for supporting E2-dependent DNA replication or E2 repression of the P₁₀₅ promoter in assays that use the full LCR-containing templates and reporters (Fig. 3, 4, and 7).

As Dowhanick et al. (16) and Goodwin et al. (21) reported, heterologous TA domains such as that of VP16, p300, or Spi can not functionally replace the E2 TA domain in growth inhibition of cervical carcinoma cell lines or transcriptional suppression of E6/E7 genes. These results indicate that a function specific for E2 is therefore necessary for these activities. We do not yet know the basis of this specificity, but possibly it is mediated through enhanced DNA binding exerted through a specific interaction between the E2 TA domain and some cellular factor(s). It is also possible that cellular factors bound to the HPV LCR participate in the recruitment of E2 to its binding site. More detailed analyses will be required to elucidate the molecular mechanism of this E2-specific inhibition of cell growth and transcription.