The Human Papillomavirus Oncoprotein E7 Attenuates NF-κB Activation by Targeting the IκB Kinase Complex

Infection with high-risk human papillomaviruses (HPV) can lead to the development of cervical carcinomas. This process critically depends on the virus-encoded E6 and E7 oncoproteins, which stimulate proliferation by manipulating the function of a variety of host key regulatory proteins. Here we show that both viral proteins dose-dependently interfere with the transcriptional activity of NF-κB. A variety of experimental approaches revealed that a fraction of the E7 proteins is found in association with the IκB kinase complex and attenuates induced kinase activity of IκB kinase α (IKKα) and IKKβ, thus resulting in impaired IκBα phosphorylation and degradation. Indirect immunofluorescence shows that E7 impairs TNFα-induced nuclear translocation of NF-κB, thus preventing NF-κB from binding to its cognate DNA. While E7 obviates IKK activation in the cytoplasm, the E6 protein reduces NF-κB p65-dependent transcriptional activity within the nucleus. We suggest that HPV oncogene-mediated suppression of NF-κB activity contributes to HPV escape from the immune system.

HPVs 1 are small DNA viruses, and specific high-risk types such as the HPV type 16 (HPV16) or HPV18 are causative agents of some forms of anogenital and oral cancers (1). HPV16 encodes six early proteins including the major oncoproteins E6 and E7. Both proteins play a central role in the induction of benign proliferation and malignant transformation (2), and at least the persistence of E7 is necessary to maintain the transformed phenotype (3). These two oncoproteins are selectively and continuously expressed in HPV-induced tumors and manipulate cell proliferation upon physical and functional interaction with several master cell cycle regulators (4). E6 binds to p53 (5) and causes its ubiquitin-dependent degradation (6), thereby interfering with p53 functions in cell cycle control and apoptosis. In addition, the E6 protein binds to the protein kinase PKN (7) and other regulators including interferon regulatory factor 3 (8) and the proapoptotic Bak protein (9). The E7 protein interacts with so-called "pocket proteins" such as the retinoblastoma protein pRb, p107, and p130 (10), resulting in their enhanced phosphorylation and degradation (11). pRb destruction results in the release of E2F family transcription factors and subsequent activation of genes promoting cell proliferation (12). But the stimulatory effects of E7 on cell proliferation depends not only on its association with pRb (13,14), because E7 targets the function of a plethora of regulators including cyclin E (15), acid alpha-glucosidase (16), and M2 pyruvate kinase (17). E7 also interferes with the activity of a variety of transcription factors such as AP-1 (18), interferon regulatory factor-1 (19), fork head domain transcription factor MPP2 (20), and TATA-box-binding protein (21). This multiplicity of interaction partners and additional levels of functional E7 regulation by phosphorylations (22), protein stability (23), and the oligomerization state (24) allow a highly complex and sophisticated manipulation of the expression program by E7 oncoproteins (1,25). Most of the E6/E7-regulated genes allow the virus to interfere either with cell proliferation and apoptosis or enable viral escape from the immune system. Immunological tolerance is induced by various mechanisms including transcriptional down-regulation of the major histocompatibility complex (MHC) class I gene (26) and selected proinflammatory cytokines (4,27).
Some E6/E7-regulated gene products are target genes of NF-B, a dimeric transcription factor involved in the expression of proteins necessary for innate immunity (28), apoptosis, and cell proliferation (29). NF-B is typically a heterodimer between the p50 and p65 (RelA) subunits and is mainly regulated by intracellular compartmentalization. The inactive form of NF-B is kept in the cytoplasm upon association with an inhibitory IB protein (30). Triggering cells with a variety of stimuli including TNF␣, IL-1, or phorbol ester induces phosphorylation of IB, which allows subsequent ubiquitinylation and degradation of the inhibitor, thus leading to nuclear entry and DNA binding of NF-B (31). The inducible phosphorylation of IB␣ at serines 32 and 36 is mediated by the IB kinase complex (IKC), which contains the IB kinases IKK␣ and IKK␤ and the regulatory subunit IKK␥/NEMO (30). The IKKs are activated by direct phosphorylation mediated by upstream kinases. Alternatively, IKKs can be recruited to intracellular domains of cell surface receptors that lead to an increased local concentration of the IKKs and allow their auto-and crossphosphorylation (28).
Given the overlapping set of genes regulated by E6/E7 and NF-B, we analyzed the effects of HPV E6 and E7 proteins on NF-B activity. These experiments revealed a dose-dependent interference of these HPV oncoproteins with NF-B functions.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfections-Human U2OS and H1299 cells were grown in DulbeccoЈs modified Eagle medium supplemented with 10% (v/v) fetal calf serum and 1% (v/v) penicillin/streptomycin (all from Invitrogen). These cell lines were transfected using Superfect ® reagent (Qiagen) according to the instructions of the manufacturer. Primary kidney epithelial cells stably transfected with the E6 or E7 gene under the control of the mouse mammary tumor virus (MMTV) promoter (32) were grown in the presence of 1 M dexamethasone.
Luciferase Assays-Luciferase activity in cell extracts was measured in a luminometer (Duo Lumat LB 9507, Berthold) by automatically injecting 50 l of assay buffer and measuring light emission for 10 s after injection according to the instructions of the manufacturer (Promega). To ensure comparable transfection efficiencies, results were normalized to ␤-galactosidase produced by a cotransfected RSV-␤-galactosidase expression vector.
Electrophoretic Mobility Shift Assays (EMSAs)-Cells stably transfected with MMTV-E6/E7 were washed twice with phosphate-buffered saline. Nuclear extracts were prepared by resuspending the cell pellet in 200 l of cold buffer A (10 mM Hepes/KOH (pH 7.9), 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM dithiothreitol, and 0.5 mM phenylmethylsulfonyl fluoride). After incubation for 10 min on ice, 5 l of 10% (v/v) Nonidet P-40 was added, and cells were lysed by vortexing. Cell nuclei were isolated by short centrifugation and dissolved in 30 l of buffer C (20 mM Hepes/KOH (pH 7.9), 0.4 M NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, and (10 g/ ml) aprotinin). After incubation on ice, the extract was centrifuged for 5 min in a microcentrifuge at 4°C, and the supernatant was used for EMSAs essentially as described (33). The supershift experiments were performed by preincubating the nuclear extracts with 2 g of ␣p65 antibodies for 15 min at 4°C.
Immunofluorescence-H1299 cells were grown on cover slips and analyzed 1 day post-transfection by immunofluorescence. 20 min after TNF␣ stimulation, cells were fixed with 3.5% (w/v) paraformaldehyde for 15 min at room temperature. After permeabilization with 0.02% (v/v) Nonidet P-40 in phosphate-buffered saline for 1 min, cells were incubated for 2 h with 10% (v/v) goat serum in phosphate-buffered saline containing 0.2% (v/v) Triton X-100. The primary antibodies were diluted to 1 g/ml and added for 1 h at 22°C. After further washing steps, the following secondary antibodies were added: Alexa-488-coupled goat ␣-rabbit (Molecular Probes) and Cy3-coupled goat ␣-mouse (Dianova). Chromosomal DNA was visualized by 4Ј,6-diamidino-2-phenylindole (DAPI), and stained cells were mounted on glass slides and examined using a Zeiss Axiophot microscope. The stained cells were further analyzed using Axiovision software.
Co-precipitation Experiments and Immunoblotting-Cells were washed with phosphate-buffered saline, and pellets were resuspended on ice for 15 min in 250 l of Nonidet P-40 lysis buffer (20 mM Tris/HCl (pH 7.5), 150 mM NaCl, 1 mM phenylmethylsulfonyl fluoride, 10 mM NaF, 0.5 mM sodium vanadate, leupeptin (10 g/ml), aprotinin (10 g/ml), 1% (v/v) Nonidet P-40, and 10% (v/v) glycerol). Cell debris was pelleted upon centrifugation, and extracts were precleared with protein A/G-Sepharose. Equal amounts of protein contained in the supernatants were mixed with 1-2 g of antibodies and 25 l of protein A/G-Sepharose and rotated for 4 h on a spinning wheel at 4°C. The immunoprecipitates were washed five times in Nonidet P-40 lysis buffer and then boiled in 1ϫ SDS sample buffer prior to SDS-PAGE and further analysis by semidry Western blotting.
IB Kinase Assays-Cells were transfected and lysed in Nonidet P-40 lysis buffer. An aliquot of the cell extract was directly analyzed by immunoblotting. The tagged IKK proteins contained in the remaining cell lysate were immunoprecipitated using ␣HA antibodies. The precipitate was washed three times in lysis buffer and two times in kinase buffer ( SDS sample buffer. Proteins were separated by SDS-PAGE, and the fixed gel was dried and quantified using a phosphorimager.

HPV E6 and E7
Interfere with Transcriptional Activity of NF-B-To investigate the impact of HPV E6 and E7 proteins on NF-B activity, human U2OS cells were transfected with a NF-B-dependent luciferase reporter gene and various combinations of expression vectors encoding the two HPV oncoproteins and the adenovirus-encoded control protein E1A 12S. TNF␣-induced NF-B activity was only mildly impaired by E6, whereas the inhibitory effect of E7 on NF-B was more pronounced (Fig. 1A). Maximal NF-B inhibition was achieved with intermediate amounts of expression vector, whereas NF-B inhibition was diminished upon expression of higher levels of viral proteins. Also, E6 and E7 encoded by the HPV high-risk strain 18 were negatively interfering with NF-B activity (data not shown), revealing that the inhibitory activity is not restricted to HPV16. To test whether E7 also inhibits NF-B that is activated by further stimuli, we determined the effects of E6/E7 expression on NF-B-dependent reporter gene activity that is induced by IL-1␤ or phorbol ester. The HPV oncoproteins inhibited NF-B activity in response to both stimuli (Fig. 1B), indicating that the viral proteins interfere with one or more common steps during NF-B activation rather than with a single event that is specific for an individual stimulus. Accordingly, E7 also inhibited NF-B activity that was induced by expression of MEKK1 (data not shown). The inhibitory effect occurred also in other human and murine cell lines (data not shown), excluding cell type-specific effects.
E6 Interferes with NF-B-dependent Transactivation, whereas E7 Impairs Induced DNA Binding-The step within the NF-B activation cascade affected by HPV E6/E7 expression was further analyzed by testing the impact of both proteins on TNF␣-induced DNA binding of NF-B. Primary cells stably transfected with the E6 or E7 gene under the control of the glucocorticoid-inducible MMTV promoter were cultured in the absence or presence of the synthetic steroid dexamethasone. Because of the lack of E6/E7-recognizing antibodies, the hormone-inducible production of both oncoproteins was confirmed by cell proliferation assays (data not shown). TNF␣ stimulation induced DNA binding of nuclear NF-B, as deter-mined by EMSAs. Supershift assays showed that the NF-B/ DNA complex contained the transactivating p65 subunit. Whereas dexamethasone-triggered expression of E6 failed to interfere with DNA binding of NF-B ( Fig. 2A), the E7 protein strongly reduced DNA binding of NF-B (Fig. 2B). The E6 protein interferes with NF-B-dependent transcription without changing induced DNA binding, raising the possibility that E6 affects NF-B-dependent transactivation. To test this hypothesis, a nuclear fusion protein between the transactivating NF-B p65 subunit and the Gal4 DNA-binding domain was tested for its activity in the presence of HPV E6 and E7 proteins. Gal4-p65-induced transcription of a Gal4-dependent luciferase reporter gene was slightly but significantly impaired by E6, whereas E7 had no impact (Fig. 2C). In a control experiment, E6 expression did not interfere with transcription in- FIG. 3. HPV E7 interferes with TNF␣-induced nuclear translocation of NF-B. H1299 cells were transiently transfected with an expression vector encoding FLAG-tagged E7 and were either left untreated (A) or stimulated for 20 min with TNF␣ (B). Intracellular localization of p65 was investigated by indirect immunofluorescence using ␣p65 antibodies. The E7 protein was detected with ␣FLAG antibodies. NF-B p65 (green) and E7 (red) localization is shown. An overlay of both stains reveals areas of co-localization in yellow. Nuclear DNA was visualized with DAPI (4Ј,6-diamidino-2-phenylindole).

FIG. 4. E7 is constitutively associated with the IKC.
A, U2OS cells were transfected with either empty expression vector or an expression vector for FLAG-tagged E7. One day later cells were lysed, and the E7 protein was immunoprecipitated (IP) from an aliquot of the lysates with ␣FLAG antibodies. The co-precipitating IKK␣ protein was detected by immunoblotting using ␣IKK␣ antibodies (upper). Aliquots of the whole cell extract (WCE) were tested by immunoblotting for the expression of IKK␣ and E7 (lower). The position of molecular mass markers is indicated at the left. B, expression vectors for Myc-tagged IKK␣ and IKK␤ and FLAG-tagged E7 were transfected into U2OS cells at the indicated combinations. One day later cells were lysed, and an aliquot of the lysate was used to immunoprecipitate the E7 protein with ␣FLAG antibodies. The co-precipitating IKK proteins were detected by immunoblotting using ␣Myc antibodies. In a complementary experiment, another aliquot of the cell lysate was used to immunoprecipitate the IKKs with ␣Myc antibodies followed by detection of co-precipitating E7 protein with ␣FLAG antibodies (middle). Five percent of the WCE was analyzed by immunoblotting for the correct expression of the ectopically expressed proteins (lower). Representative results are shown. duced either by the DNA-binding domain of Gal4 alone or by the Gal4-VP16 fusion protein (Fig. 2C).
E7 Partially Co-localizes with NF-B p65 and Prevents Its TNF␣-induced Nuclear Translocation-To test whether impaired DNA binding of NF-B in the presence of E7 may be caused by the effects on nuclear import of NF-B, H1299 cells were transfected with an expression vector encoding FLAGtagged E7 or the empty expression vector as a control. The next day, cells were either left untreated or stimulated with TNF␣. Double staining with ␣p65 and ␣FLAG antibodies was used to identify transfected cells by immunofluorescent staining (Fig.  3A). In unstimulated cells, the p65 subunit was found in the cytoplasm. In agreement with previous reports (38,39), the E7 protein occurred in the nucleus and the cytoplasm. Double staining for NF-B p65 and E7 revealed areas of overlapping localization in the cytoplasm, which are shown in yellow. In the absence of E7, TNF␣ treatment induced complete nuclear localization of p65, whereas expression of this viral oncogene strongly interfered with nuclear uptake of p65 (Fig. 3B).
E7 Is Constitutively Associated with the IB Kinase Complex-The overlapping subcellular localization of E7 and NF-B p65 raises the possibility of a physical association between this viral protein and the IKC, which was experimentally tested by co-immunoprecipitation experiments. In the absence of antibodies immunoprecipitating the endogenous E7 protein, a FLAG-tagged E7 protein was expressed in U2OS cells. Cells were lysed, and the viral protein was immunoprecipitated from the cell extracts with ␣FLAG antibodies. Subsequent immunoblotting revealed the occurrence of endogenous IKK␣ in E7 immunoprecipitates (Fig. 4A), showing an association between the endogenous IKC and the E7 protein. The association between the IKC and E7 was not modulated after TNF␣ stimulation (data not shown). This interaction was further characterized upon transfection of cells with Myc-tagged versions of IKK␣ and IKK␤ either alone or together with a FLAG-E7 encoding vector. The E7 protein was immunoprecipitated from an aliquot of the cell lysates, and Western blotting showed its association with IKK␣ and IKK␤ (Fig. 4B). In a complementary experimental approach, the IKKs were immunoprecipitated from another aliquot of the cell lysate followed by the detection of associated E7 proteins by Western blotting. These experiments confirmed the E7-IKK interaction by an independent experimental approach. Of note, these biochemical experiments and the co-localization studies revealed only a fraction of the total cellular E7 proteins in association with the IKC.
E7 Interferes with Induced IB␣ Phosphorylation and IKK Kinase Activity-To test whether the IKK association of E7 has any consequences for induced IB␣ phosphorylation and degradation, U2OS cells were transfected at high efficiency with an E7 expression vector or empty control vector. Cells were either left untreated or stimulated with TNF␣, and extracts were tested by immunoblotting for the occurrence of IB␣. Expression of E7 impaired TNF␣-induced degradation of IB␣ (Fig. 5A). The incomplete protection of TNF␣-induced IB proteolysis by E7 can be attributed to the limited transfection efficiency of cells. In parallel, determination of IB␣ phosphorylation by Western blotting using phosphospecific antibodies FIG. 5. Expression of E7 impairs IKK activity and IB␣ phosphorylation. A, U2OS cells transfected to express E7 were either left untreated or stimulated for 5 min with TNF␣. Cell extracts were analyzed by immunoblotting for the abundance of IB␣ (upper). In parallel, the extract was tested by Western blotting for the phosphorylation of IB␣ as detected by phosphospecific antibodies (lower). The position of a nonspecific (ns) band is shown. B, HA-tagged IKK␣ was expressed either alone or in combination with NIK and increasing amounts of E7 in U2OS cells as shown. 24 h post-transfection, cell lysates were prepared, and IKK␣ was immunoprecipitated. Kinase activity was determined by immune complex kinase assays (KA) using purified GST-IB␣ (1-54) as substrate. An autoradiogram from a reducing SDS gel shows IKK␣ phosphorylation (upper) and phosphorylation of the recombinant substrate protein and a quantitative evaluation obtained by phosphorimaging (middle). A fraction of the cell lysate was analyzed by Western blotting (WB) for expression of IKK␣, NIK, and E7 (lower). C, the experiment was done and analyzed as in (B) with the exception that an expression vector encoding IKK␤ instead of IKK␣ was used. Representative experiments are shown.
revealed an impaired TNF␣-triggered phosphorylation of IB␣ in the presence of E7. To directly test the consequences of E7 expression on IKK activity, cells were transfected to express moderate amounts of HA-tagged IKK␣, which allows their incorporation into functional cytokine-responsive high molecular weight IKCs (40,41) together with vectors encoding the IKK activator NIK and increasing amounts of E7. The tagged IKK␣ protein was immunoprecipitated, and its activity was examined by measuring the phosphorylation of the exogenously added substrate protein (GST-IB-␣ (1-54)) in immune complex kinase assays (Fig. 5B). NIK-induced kinase activity of IKK␣ was dose-dependently impaired upon coexpression of E7. The impact of E7 on IKK␤ activity was assayed by an analogous experimental approach by ectopically expressing IKK␤ instead of IKK␣. As already seen for IKK␣, the expression of E7 also inhibited NIK-activated IKK␤ activity (Fig. 5C), showing that the viral protein can directly interfere with induced kinase activity.

DISCUSSION
Given the important contribution of NF-B to central biological processes, this transcription factor and its activation pathways are frequently targeted by viruses. Early evidence for viral appropriation of the NF-B pathway came from the finding that the turkey retrovirus REV-T encodes the v-rel protein, an oncogenic homologue to the NF-B DNA-binding subunits (42). It is frequently observed that products of a variety of viruses induce NF-B activity in order to ensure NF-B-dependent expression of viral genes (43). On the other hand, the relevance of NF-B for innate immunity and the induction of apoptosis forced some viruses to evolve strategies to counteract NF-B activation. For example, the African swine fever virus encodes a functional and stable IB protein, which is able to inhibit NF-B activity by replacement of the proteasome-degraded endogenous IB␣ protein (44).
The repressive effect of NF-B on HPV transcription via binding to the long control region (45) raises the need to interfere with the inhibitory activity of this transcription factor. Since HPV transcription is prevented by some proinflammatory NF-B target proteins including TNF␣ and IL-1 (46,47), it is reasonable to assume that it is beneficial for the virus to disturb NF-B activity. Consistent with our data showing an inhibitory effect of HPV oncoproteins on NF-B-dependent transcription, expression of the bona fide NF-B target gene IL-6 occurs only in stromal cells surrounding cervical carcinomas but not in the tumor itself (48). Therefore, the inhibitory effect of HPV oncoproteins on NF-B may contribute to virus evasion from the host immune system. Interestingly, papillomavirus-encoded proteins might also inhibit NF-B by an additional mechanism. Nuclear extracts from human papillomavirus type 6-and 11-infected laryngeal papilloma tissues contain elevated levels of NF-B p50 homodimers (49), which act as repressors of NF-B-dependent transcription (50). The E6 protein encoded by HPV type 16 negatively interferes with NF-B activity in the human ovarian cancer cell line A2780, thus sensitizing these cells to TNF␣-induced cytotoxicity (51).
Biological evidence shows that HPV infection down-regulates only a subset of NF-B target genes (4). This might be explained by the fact that many cytokines are not solely regulated by NF-B and critically depend on the concerted activity of NF-B together with various other transcription factors, which themselves are potentially affected by E6/E7 proteins. The dose dependence of E7 and E6 activity raises the possibility that the amount of viral proteins or the quantity of IKCassociated E7 proteins determines the impact on NF-B function. The mechanistic basis for reduced NF-B inhibition in the presence of high amounts of E6/E7 is not clear. Possibly, ele-vated concentrations of viral protein favor the formation of E6 and E7 multimers (52), thus leading to decreased interactions between E7 and the IKC. This might also explain why retrovirus mediated very strong overexpression of E6 from HPV16 induces transcription of some NF-B target genes (25).
Since the IKC serves as an intracellular point of convergence for distinct NF-B activation signals (53), and IKK␤ is required for activation of the immune response in response to viral infections (54), this kinase complex is frequently usurped by a variety of viral proteins (43). Interference of E7 with IKK activity may be mechanistically explained by mutual binding and steric effects on the spatial conformation of the IKC.
Our data indicate that E6 does not prevent induced DNA binding activity of NF-B, but it does impair NF-B-dependent transactivation. Since the zinc finger domain of the nuclear E6 protein interacts with the coactivator CBP/p300, the negative effect of E6 on p65 function may also involve competition for this commonly used coactivator. This inhibitory activity on p65 activity may add to the negative effect of E6 on p53 within the nucleus, since p53-induced apoptosis is prevented upon inhibition of NF-B (55). Interestingly, NF-B activity is never completely switched off by E6/E7 proteins. This may be biologically meaningful, because residual NF-B activity could be necessary for proliferation of virus-infected cells.