|Home | About | Journals | Submit | Contact Us | Français|
The papillomavirus life cycle parallels keratinocyte differentiation in stratifying epithelia. We have previously shown that the human papillomavirus type 8 (HPV8) E2 protein downregulates β4-integrin expression in normal human keratinocytes, which may trigger subsequent differentiation steps. Here, we demonstrate that the DNA binding domain of HPV8 E2 is sufficient to displace a cellular factor from the β4-integrin promoter. We identified the E2-displaceable factor as activator protein 1 (AP-1), a heteromeric transcription factor with differentiation-specific expression in the epithelium. β4-Integrin-positive epithelial cells displayed strong AP-1 binding activity. Both AP-1 binding activity and β4-integrin expression were coregulated during keratinocyte differentiation suggesting the involvement of AP-1 in β4-integrin expression. In normal human keratinocytes the AP-1 complex was composed of JunB and Fra-1 subunits. Chromatin immunoprecipitation assays confirmed that JunB/Fra-1 proteins interact in vivo with the β4-integrin promoter and that JunB/Fra-1 promoter occupancy is reduced during keratinocyte differentiation as well as in HPV8 E2 positive keratinocytes. Ectopic expression of the tethered JunB/Fra-1 heterodimer in normal human keratinocytes activated the β4-integrin promoter, while coexpression of HPV8 E2 reverted the JunB/Fra-1 effect. In summary, we identified a novel mechanism of human β4-integrin regulation that is specifically targeted by the HPV8 E2 protein mimicking transcriptional conditions of differentiation. This may explain the early steps of how HPV8 commits its host cells to the differentiation process required for the viral life cycle.
Human papillomaviruses (HPVs) are DNA viruses, which infect stratified epithelia of skin and mucosa. Depending on the virus type, they may give rise to a large spectrum of lesions ranging from benign skin warts to invasive carcinoma. Anogenital malignancies including cervical cancer and a group of head-and-neck squamous cell carcinomas are most commonly associated with high-risk mucosal HPV16 and HPV18 infection (38). For skin, the carcinogenic potential of HPV5 and HPV8 is fully accepted in epidermodysplasia verruciformis patients (24). Accumulating evidence supports a role of these viruses in the development of nonmelanoma skin cancer in the general population (36).
Actively dividing basal cells of stratified epithelia are the target for viral infection but the productive viral life cycle takes place in suprabasal, differentiating cells (27). HPV replication critically depends on the coexistence of the cellular DNA replication machinery and differentiation events in keratinocytes (30). However, in normal keratinocytes these processes are separated and confined to either basal cells or the suprabasal differentiating cell compartment. While the viral E6 and E7 oncoproteins force infected cells to retain cell cycle capabilities and repress differentiation, it is unclear which signals commit infected keratinocytes to enter the differentiation process.
The E2 protein is indispensable for both viral replication and gene expression. It acts predominantly as a DNA binding protein recognizing the consensus palindromic sequence ACCN6GGT (27, 30). Various studies have demonstrated that the papillomavirus E2 protein can display antiproliferative and prodifferentiation properties (4, 15, 19, 21, 33). Introduction of E2 from genital high-risk HPV into HPV-positive cancer cell lines results in repression of E6 and E7 oncogene transcription, leading to replicative senescence or apoptosis (13, 19). Repression is the consequence of E2 binding to low-affinity E2 binding sites in the viral genome and displacement of cellular transcription factors from their binding sites, which are adjacent or overlap the E2 cognate sequences. This mode of transcriptional regulation seems to be a conserved property of E2, since it was shown for E2 proteins from high-risk mucosal (HPV16 and 18) and cutaneous (HPV8) virus types (5, 12, 14, 43, 46, 48).
Apart from regulatory effects on viral gene expression we also observed that HPV8 E2 repressed a cellular gene, β4-integrin, and this may be an early event to commit basal keratinocytes to enter suprabasal differentiation steps. Our data showed that E2 directly interacts with the gene regulatory region at three novel binding sites of the β4-integrin promoter. Detailed analysis defined that HPV8 E2 was able to displace a yet-unrecognized cellular factor from the second binding site (33).
β4-Integrin is one of the major cell adhesion receptors in basal keratinocytes with fundamental functions in epithelial homeostasis. Expressed at the basolateral surface of keratinocytes, β4-integrin forms heterodimers with the α6 integrin subunit, and the α6β4 integrin binds to laminin-5. These specialized structures are crucial components of hemidesmosomes, which tightly attach basal cells to the underlying basement membrane and maintain proliferative potential (7). Since β4-integrin expression is undetectable in suprabasal keratinocytes, loss of β4-integrin expression may be among the first steps in keratinocyte differentiation, when cells leave the basal and enter the suprabasal layers (18).
The molecular mechanism of how HPV8 E2 exactly represses β4-integrin expression remains poorly understood. We therefore aimed to identify the cellular factor displaced by E2 (E2-displaceable factor [E2-DF]) from the β4-integrin promoter. We show here that E2-DF is a member of the activator protein 1 (AP-1) transcription factor family composed of JunB/Fra-1 heterodimers, which are expressed in proliferating basal keratinocytes (1). We demonstrate for the first time that JunB/Fra-1 activates the β4-integrin promoter and that its activity is targeted by the HPV8 E2 protein.
The plasmid pEYFP-HPV8-E2fl, encoding enhanced yellow fluorescent protein (EYFP) fused to full-length HPV8 E2, pEYFP-HPV8-E2C, encoding EYFP fused to the DNA binding and dimerization domain of HPV8 E2 and tethered JunB/Fra-1 heterodimers expressed from a cytomegalovirus-driven pCG-based vector have all been described previously (2, 21). The L5.5K luciferase reporter construct containing a fragment (−5197 to +333) of the human β4-integrin promoter region fused upstream of the firefly luciferase gene in the pGL3-basic vector (Promega, Madison, WI) was a generous gift from S. Hirohashi (47). The pET14b-HPV8-E2C expression vector was kindly provided by G. Steger (Cologne, Germany) and used to express the C terminus of HPV8-E2 as an N-terminal His-tagged fusion protein in bacteria. pLXSN-HPV8E2 was created by insertion of the HPV8 E2 open reading frame into the pLXSN vector (BD Biosciences Clontech, Palo Alto, CA).
Primary normal human foreskin keratinocytes (NHK) (BioWhittaker, Vervier, Belgium) were cultured under low-calcium (0.15 mM) conditions in KGM-2 medium (BioWhittaker), if not otherwise indicated. The medium further contained insulin, epidermal growth factor, hydrocortisone, epinephrine, and bovine pituitary extract according to the supplier's instructions. To induce differentiation, NHK at 50% confluence were treated with 1.2 mM or 2.4 mM CaCl2 (POCH S.A., Poland) for 48 h (37). The HPV-negative keratinocyte cell line RTS3b derived from a human cutaneous squamous cell carcinoma (39) was maintained in Dulbecco modified Eagle medium containing (DMEM) 25% Ham F-12 medium, 9% fetal calf serum, 50 μg of gentamicin/ml, and 10 ng of epidermal growth factor/ml (all from Invitrogen, Karlsruhe, Germany) and 0.4 μg of hydrocortisone/ml, 10 ng of cholera toxin/ml, 5 μg of transferrin/ml, 2 × 10−11 M triiodothyronine, 1.8 × 10−4 M adenine, and 5 μg of insulin/ml (all from Sigma-Aldrich, Taufkirchen, Germany).
The HPV-negative cervical carcinoma cell line C33A (American Type Culture Collection [ATCC] HTB-31, Manassas, VA), the skin keratinocyte cell line HaCaT (8), the hepatocarcinoma cell line HepG2 (ATCC HB-8065), and 293T cells (34), as well as the HPV16- and HPV18-positive cervical carcinoma cell lines CaSki (ATCC CRL-1550, HPV16), SiHa (ATCC HTB-35, HPV16), HeLa (ATCC CCL-2, HPV18), SW756 (HPV18; kindly provided by M. von Knebel-Doeberitz, Heidelberg, Germany), and C4-1 (ATCC CRL-1594, HPV18) and the nonmalignant HPV16-transformed foreskin keratinocyte cell line HPKIA (kindly provided by M. Dürst, Jena, Germany), were all grown in DMEM with Glutamax I supplemented with 9% fetal calf serum, 100 U of penicillin/ml, 0.1 mg of streptomycin/ml, and 1 mM sodium pyruvate (all from Invitrogen).
NHK were seeded onto six-well plates at a density of 2.5 × 105cells per well. After 24 h, the culture medium was aspirated, and the cells were transfected in a serum-free fresh medium with a total amount of 2 μg of DNA per well with the TransFast transfection reagent (Promega) used according to the manufacturer's instructions. At 48 h after transfection the cells were washed twice in phosphate-buffered saline (PBS) and lysed by adding 150 μl of luciferase extraction buffer containing 0.1 M potassium phosphate buffer (pH 7.5) and 0.5% Nonidet P-40. Luciferase activity was determined with a Victor II (Perkin-Elmer LAS, Rodgau, Germany) luminometer and normalized to protein concentrations in the respective extracts. Each condition was tested in triplicates.
Cells (4 × 105) were incubated with a monoclonal anti-β4-integrin antibody, clone 3E1 (Chemicon, Hofheim, Germany), or with an isotype-matched antibody at a 1:200 dilution in 2% bovine serum albumin (BSA) in PBS for 45 min at 4°C. Cells were then washed, stained with fluorescein isothiocyanate-conjugated goat anti-mouse IgG (Dianova, Hamburg, Germany) for 45 min at 4°C, and fixed in 1% paraformaldehyde. Flow cytometry analysis was performed with a FACSCalibur instrument (Becton Dickinson, San Jose, CA), and the percentage of β4-integrin-positive cells was calculated by using CellQuest software (BD Biosciences, San Jose, CA).
Supernatants containing recombinant retrovirus were generated by transfecting the pLXSN-HPV8E2 vector or the control pLXSN vector into the packaging PT67 cell line (BD Biosciences Clontech) and subsequent selection with G418. NHK in the first passage were seeded onto six-well plates at a density of 1.5 × 105 cells per well. After 24 h, the culture medium was replaced by DMEM containing 5 μg of Sequabrene (Sigma-Aldrich)/ml, followed by incubation for 10 min at 37°C. Next, the medium was changed to the retroviral supernatants 50% (vol/vol) of DMEM containing 5 μg of Sequabrene/ml. Cells were then centrifuged at 300 × g for 1 h at room temperature and washed with PBS, and finally KGM-2 medium was added. Selection with G418 sulfate (100 μg/ml; Invitrogen) began 2 days postinfection and continued for 8 days. Expression of the HPV8 E2 protein was verified by quantitative PCR.
Total RNA was extracted from retrovirally infected NFK by using an RNeasy minikit (Qiagen, Hilden, Germany) and cDNA was generated from 1 μg of RNA with the First-Strand cDNA Synthesis kit (Fermentas, St. Leon-Rot, Germany). Real-time PCR was performed with the LightCycler 480 instrument (Roche, Mannheim, Germany). PCR primers (Operon, Cologne, Germany) and probes (Roche Universal Probe Library) were designed by using ProbeFinder software, version 2.35 (Roche). For the β4-integrin quantitative PCR (qPCR), the primer pair 5′-CACACTGCCCAGGGACTAC-3′ and 5′-CAGCAGTCAGGCGAGAGTC-3′ and probe 52; for the GAPDH (glyceraldehyde-3-phosphate dehydrogenase) qPCR, we used the primer pair 5′-TGCTGTAGCCAAATTCGTTGT-3′ and 5′-CTGACTTCAACAGCGACACC-3′ and probe 25; and for the HPV8 E2 qPCR, we used the primer pair 5′-GACGGCGATCAACCTCAA-3′ and 5′-CTCCCCTTTGTGACCGTTT-3′ and probe 22. Changes in β4-integrin expression were quantified by using Roche LightCycler 480 software version 1.5 and normalized to GAPDH.
A total of 3 × 106 NHK, C33A, SiHa, CaSki, and HepG2 cells were grown in 100-mm dishes, washed twice with PBS, and detached with a cell scraper. Cells were pelleted and resuspended in 100 μl of buffer containing 5 mM Tris (pH 7.4), 75 mM NaCl, 0.5% Triton X-100, 0.5% sodium deoxycholate, 2.5 mM EDTA, and a cocktail of protease inhibitors. Then, 25 μg of the sample was separated by SDS-10% PAGE. After transfer onto Immobilon-P polyvinylidene difluoride transfer membranes (Millipore, Schwalbach, Germany), the membranes were blocked with 5% skim milk in PBS-Tween and incubated overnight at 4°C with anti-JunB (sc-73X) and anti-Fra-1 (sc-605X) (both from Santa Cruz Biotechnology, Santa Cruz, CA) and anti-c-Fos (catalog no. 06-341; Millipore, Schwalbach, Germany) antibodies at a dilution of 1:2,000. The blots were washed in PBS with 0.1% Tween, incubated with peroxidase-labeled secondary antibodies (Santa Cruz Biotechnology), and developed by using a chemiluminescence detection system (Pierce, Rockford, IL) according to the manufacturer's instructions.
Nuclear extracts were prepared from NHK, cell lines and transiently transfected 293T cells according to the protocol of Schreiber et al. (44). In brief, untransfected cells were seeded onto 60-mm dishes at the density of 1.25 × 105 cells per dish and then harvested after 24 h. 293T cells (1.5 × 106 cells per dish) were plated onto 100-mm dishes and grown for 24 h prior to transfection with 28 μg of pJunB/Fra-1 or 28 μg of pEYFP-HPV8-E2C DNA and their respective control vectors. Nuclear extracts were isolated after 48 h. For electromobility shift assay, oligonucleotides (BS-II) containing the second E2 binding site (underlined) of the β4-integrin promoter (33) (TCTCAGATGACCAGGAAAGGTGGCGACTCACACATTTG; nucleotides 3590 to 3601 according to GenBank accession no. AB012286) were radiolabeled with [γ-32P]dATP by using T4 polynucleotide kinase (New England Biolabs, Ipswich, MA). For these experiments also, BS-II oligonucleotides sequentially truncated by 5 bp from the 5′ and/or 3′ end (Table (Table1)1) and AP-1 consensus oligonucleotides 5′-CGCTTGATGACTCAGCCGGAA-3′ (AP-1 binding site, underlined) (Santa Cruz Biotechnology) were used.
The indicated amounts of nuclear extracts or purified His-tagged HPV8-E2C fusion proteins were incubated with 32P-labeled DNA probe in 30 μl of binding buffer [20 mM HEPES-K (pH 7.9), 60 mM KCl, 4 mM MgCl2, 1 mM spermidine, 0.1 mM EDTA, 1 mM dithiothreitol, 1 mg of BSA/ml, 50 μg of poly(dI-dC) (Amersham Pharmacia, Piscataway, NJ)/ml, 10% glycerol, 1.5 mM phenylmethylsulfonyl fluoride, and 0.3 mg of aprotinin (Sigma)/ml in final concentrations] on ice for 30 min.
To analyze AP-1 binding activity, 10 μg of each nuclear extract was incubated in a total volume of 20 μl with 10 μl of buffer A (20 mM HEPES [pH 7.9], 20% glycerol, 100 mM KCl, 0.2 mM EDTA, 0.2 mM phenylmethylsulfonyl fluoride, 1 mM DTT), 2 μl of buffer B (150 mM HEPES [pH 7.6], 100 mM KCl, 2.5 mM DTT, 2 mg of BSA/ml, 20 mM MgCl2), 1 μl of a 1-mg/ml concentration of poly(dI-dC), and 32P-labeled DNA probe.
For competition assay, unlabeled competitor oligonucleotides (40- and 400-fold molar excess) were preincubated with nuclear extracts for 15 min at 4°C before the addition of labeled probe. In some competition assays HPV8-E2C protein or EYPF-HPV8-E2C fusion protein containing extracts were used as indicated in the figure legends.
For supershift experiments, polyclonal rabbit anti-c-Jun (sc-45X), anti-JunB (sc-73X), anti-JunD (sc-74X), anti-Fra-1 (sc-605X), anti-Fra-2 (sc-604X), and goat anti-c-Fos (sc-52X) antibodies (all from Santa Cruz Biotechnology) were used at dilutions of 1:15 each.
Reaction products were separated in 5% polyacrylamide gels in 0.5× Tris-borate-EDTA buffer. The gels were dried and exposed to autoradiographic films.
RTS3b cells transfected for 24 h with 13.2 μg of pJunB/Fra-1 or control vector using Fugene 6 transfection reagent (Roche, Basel, Switzerland) according to the manufacturer's protocol and retrovirally infected NHK and NHK treated with 1.2 mM CaCl2 for 3 or 8 days were used for ChIP analysis. Cells (106) were seeded onto 100-mm dishes, treated as described above, treated with trypsin, and pelleted at 240 × g for 5 min, and ChIP was performed according to the protocol of the ChIP assay kit from Upstate Biotechnology (Lake Placid, NY) as described previously (26). Briefly, cells were resuspended in culture medium and treated with formaldehyde (Roth, Karlsruhe, Germany) at a final concentration of 1% and sonicated twice for 7 min with a Bioruptor apparatus (Diagenode, Liege, Belgium). For preclearing, the cell lysates were incubated on ice for 30 min with protein A/G-Sepharose (Santa Cruz) containing 200 μg of salmon sperm DNA/ml. The cleared supernatant was then incubated overnight at 4°C with 2 μg of either anti-Fra-1 (sc-605X; Santa Cruz), anti-JunB (sc-73X; Santa Cruz), or rabbit IgG (Chromopure, Dianova, Germany) antibodies. The immune complexes were precipitated with 25 μl of protein A/G-Sepharose in the presence of salmon sperm DNA for 1 h at 4°C. After elution of the protein-DNA complexes and reversing the cross-links, DNA was isolated, and the β4-integrin promoter sequence was detected and quantified by using real-time PCR and the Human Universal Probe Library System (Light Cycler; Roche). Probe 86 was used in combination with promoter-specific forward (5′-AAAGGTGGCGACTCACACATTTGC-3′) and reverse (5′-TGAGCCTCCATGCTGTCATCTGTA-3′) primers. The 98-bp PCR product was detected by electrophoresis on a 2.5% agarose gel.
As previously shown, the viral E2 protein strongly represses the β4-integrin promoter in primary human keratinocytes. In these experiments, ca. 40% of repression was still achieved with a fragment containing the carboxy-terminal DNA binding and dimerization domain of HPV8 E2. Further analysis revealed at least three papillomavirus E2 binding sites within the human β4-integrin enhancer. The second binding site was shown to be involved in E2-mediated repression and BS-II oligonucleotides comprising this E2 binding site displayed the highest affinity for E2. In addition to E2, cellular factors also present in nuclear extracts from normal human keratinocytes bound to BS-II oligonucleotides. The DNA binding activity, which could be displaced by full-length HPV8 E2 protein, was therefore termed E2-displaceable factor or E2-DF. The identity of this factor remained elusive at that time (33).
To further investigate whether the carboxy terminus of HPV8 E2 can compete directly with E2-DF for DNA binding, we performed gel retardation assays. Nuclear extracts from NHK were incubated with BS-II oligonucleotides. The HPV8 E2 carboxy terminus was added to this reaction mixture either as bacterially expressed, purified His-tagged protein (His-8E2C) or in nuclear extracts containing the fusion protein EYFP-8E2C consisting of EYFP and the C terminus of HPV8 E2. Nuclear extracts from pEYFP-transfected cells served as controls. In both cases, HPV8-E2C partially displaced the as-yet-unknown cellular factor from its binding to BS-II (Fig. (Fig.11).
During keratinocyte differentiation the expression of β4-integrin is downregulated. To determine the binding activity of E2-DF in relation to keratinocyte differentiation, 1.2 or 2.4 mM CaCl2 was added to NHK cultures, and nuclear extracts were analyzed by electrophoretic mobility shift assay (EMSA). Raising calcium levels by 1.2 or 2.4 mM resulted in strong reduction of β4-integrin levels, as measured by flow cytometry. This was paralleled by a strong decrease of E2-DF DNA binding activity (Fig. (Fig.2).2). This suggested that the analyzed transcription factor may act as a positive regulator of β4-integrin expression.
Based on two different databases (TRANSFAC public version [www.gene-regulation.com] and MatInspector [www.genomatix.de]), in silico analysis of the BS-II oligonucleotide sequence was performed in order to identify binding sites for cellular transcription factors. Two binding sites (for NeuroD1 and PAX3) were predicted; however, no binding of these factors could be confirmed in NHK nuclear extracts (data not shown). Therefore, we chose an experimental approach to identify the nature of E2-DF.
To narrow down the nucleotide sequence bound by E2-DF, we sequentially truncated BS-II oligonucleotides (Table (Table1)1) and subsequently performed EMSA with extracts from NHK. Neither the loss of 5 bp from the 3′ end nor up to 20-bp deletions from the 5′ end abolished the DNA binding activity of E2-DF. Complete loss of DNA binding activity was only observed when 25 bp at the 5′ end were lacking (Table (Table11 and Fig. Fig.33).
Three additional DNA probes were designed lacking 5 bp at the 3′ end and 20, 22, or 24 bp at the 5′ end, respectively (Table (Table1).1). For these experiments, 10 times more nuclear extracts were used in EMSA to visualize DNA-protein complexes. The DNA binding activity for the probe lacking 22 bp at 5′ was significantly decreased and became visible only after overexposing the gel. A complete loss of DNA binding activity was detected with the shortest DNA probe lacking 24 bp at the 5′ end and 5 bp at the 3′ end (data not shown). The sequence of the shortest oligonucleotide still binding the cellular transcription factor was thoroughly analyzed (Table (Table1).1). It comprised a sequence 5′-CGACTCA-3′ similar to the phorbol ester response element (TRE) 5′-TGA(C/G)TCA-3′, representing an AP-1 transcription factor binding site (6). From the original sequence it differed by a C (versus T) at the first nucleotide position.
To further characterize binding activities of the E2-DF, competition experiments were performed. Nuclear extracts from NHK were preincubated with unlabeled AP-1 consensus sequence, BS-II oligonucleotides, or nonspecific target sequence prior to the addition of 32P-labeled BS-II probe. The competitors were in 40- and 400-fold excess to the labeled BS-II probe. A 40-fold excess of AP-1 oligonucleotides completely abolished the binding of the 32P-labeled BS-II oligonucleotides. In the case of unlabeled BS-II a 400-fold excess was necessary for complete competition, whereas a 40-fold excess of these oligonucleotides inhibited binding only partially. Addition of nonspecific control oligonucleotides did not affect binding of E2-DF to the 32P-labeled BS-II (Fig. (Fig.4A4A).
Vice versa, BS-II oligonucleotides were tested for their ability to compete with the AP-1 target sequence for AP-1 binding. Here, AP-1 oligonucleotides served as a labeled DNA probe. Incubation of nuclear extracts with unlabeled AP-1 target sequence completely inhibited nuclear extracts binding to the labeled DNA probe, whereas the addition of unlabeled BS-II oligonucleotides led to a weaker dose-dependent binding inhibition. Nonspecific oligonucleotides did not compete for AP-1 binding (Fig. (Fig.4B).4B). Thus, BS-II oligonucleotides competed with AP-1 oligonucleotides for the same nuclear factors, although BS-II appeared to bind with lower affinity than the AP-1 consensus sequence. Taken together, these findings strongly suggested that BS-II was bound by AP-1 transcription factors.
Next, we examined whether BS-II and AP-1 binding activity correlates with β4-integrin expression in different epithelial cell types. As shown by flow cytometry and EMSA, all cell lines expressing β4-integrin also displayed DNA binding activity to BS-II and AP-1 oligonucleotides (Fig. (Fig.5).5). In nuclear extracts from β4-integrin-positive cell lines, more than one protein complex binding to BS-II was frequently detected. The pattern of the most retarded complex, corresponding to E2-DF, resembled that of AP-1. However, in all of the tested cell lines the binding activity of the E2-DF to BS-II was considerably lower than the binding activity of AP-1 to its consensus binding site (Fig. (Fig.5B5B).
Strong DNA binding activity was observed in nuclear extracts from NHK, HaCaT, and HPV18-positive HeLa and SW756 cells. Three HPV-negative cell lines—C33A, 293T, and HepG2 (Fig. (Fig.5A)—did5A)—did not express β4-integrin. Correspondingly, neither C33A nor 293T cell lines displayed E2-DF or AP-1 binding activities. A weak band was detected by prolonged exposure of an autoradiogram only in hepatocarcinoma cells (Fig. (Fig.5B).5B). This was observed regardless of different EMSA binding buffers, which were optimized for either the viral E2 or the cellular AP-1 proteins. In two HPV16-positive cell lines (SiHa and CaSki), despite the presence of E2-DF, the expression of β4-integrin was absent or low (Fig. (Fig.55).
These data revealed that in all of the cell lines analyzed, the E2-DF binding activity matched the AP-1 binding activity and suggested that E2-DF might be an important though not the sole determinant for β4-integrin expression.
In order to identify the composition of the putative AP-1 complex bound to BS-II in NHK, EMSA was carried out in the presence of supershifting antibodies directed against specific Jun and Fos family proteins. Among all tested antibodies, the addition of either anti-JunB or anti-Fra-1 antibodies strongly reduced binding activity and led to supershifted bands. The simultaneous addition of antibodies directed against all tested AP-1 family members, i.e., c-Jun, JunB, JunD, c-Fos, Fra-1, and Fra-2 resulted in a complete loss of the AP-1 specific band. Similarly, the addition of only anti-JunB and anti-Fra-1 antibodies was sufficient for supershifting the DNA-protein complex (Fig. (Fig.6A).6A). These experiments identified the nature of BS-II binding activity as AP-1, composed of the JunB and Fra-1 subunits in NHK.
Consistent with the EMSA results, NHK displayed high constitutive JunB and Fra-1 expression, as revealed by Western blots with specific antibodies. CaSki cells with low β4-integrin expression showed low expression of both JunB and Fra-1 subunits. Interestingly, JunB or Fra-1 subunits could not be detected in either C33A or HepG2 cells. In SiHa cells, small amounts of JunB were present, whereas Fra-1 was completely absent (Fig. (Fig.6B).6B). This indicated that SiHa cells cannot form the JunB/Fra-1 heterodimer either. These data revealed that cells lacking the JunB/Fra-1 heterodimer do not express β4-integrin and further suggested the involvement of JunB/Fra-1 in the regulation of β4-integrin expression. Another AP-1 subunit, c-Fos, was detected in all of the tested cells (Fig. (Fig.6B),6B), supporting a critical role of Fra-1 and not c-Fos in β4-integrin expression in human keratinocytes.
By ChIP analysis we confirmed that JunB/Fra-1 proteins interacted in vivo with the BS-II comprising genomic DNA from the RTS3b keratinocyte cell line. Both anti-JunB and anti-Fra-1 antibodies led to seven- to eightfold enrichment of BS-II comprising DNA compared to control antibodies (Fig. (Fig.6C6C).
ChIP assays were performed in NHK cells. The results confirmed the occupancy of the β4-integrin promoter by JunB and Fra-1 in nondifferentiated keratinocytes. A gradual decrease in Fra-1 binding was observed during differentiation with 1.2 mM CaCl2. The binding of Fra-1 was reduced by 38% after a 3-day stimulation with calcium and by 72% after an 8-day stimulation (Fig. (Fig.6D),6D), and the binding of JunB to BS-II was reduced by 65 and 84%, respectively (data not shown).
To determine whether the HPV8 E2 protein indeed interferes with the binding of the cellular JunB/Fra-1 dimer to BS-II, nuclear extracts containing recombinant tethered JunB/Fra-1 heterodimers (2) generated from transiently transfected 293T cells were incubated with EYFP-8E2C containing nuclear extracts. Gel retardation assays showed that the complex corresponding to JunB/Fra-1 bound to BS-II (arrow) was displaced from DNA by increasing amounts of EYFP-8E2C but not by the highest dose of EYFP alone (Fig. (Fig.7A).7A). The binding of EYFP-8E2C to the probe is indicated by an asterisk in Fig. Fig.7A.7A. Thus, HPV8 E2 was able to compete with recombinant JunB/Fra-1 for binding to BS-II of the β4-integrin promoter.
To evaluate the impact of the HPV8 E2 protein on JunB/Fra-1 binding to the β4-integrin promoter in NHK, cells stably expressing untagged HPV8 E2 were generated by retroviral gene transfer. In these pLXSN-HPV8 E2 cells, expression of β4-integrin was reduced to 54% of the corresponding control cells (Fig. (Fig.7B,7B, left panel). These data confirmed and extended previous results obtained with NHK transiently transfected with HPV8 E2 (33). In ChIP assays performed with keratinocytes stably expressing HPV8 E2, binding of the critical AP-1 subunit Fra-1 to BS-II within the β4-integrin promoter was almost completely lost (Fig. (Fig.7B7B).
The functional consequences of JunB/Fra-1 and HPV8 E2 expression on β4-integrin promoter activity in NHK were analyzed. Although the basal levels were already high, overexpression of JunB/Fra-1 increased the β4-integrin promoter activity by twofold. To investigate whether HPV8 E2 has any effect on JunB/Fra-1-mediated transcriptional activity, cotransfection experiments were performed. In the presence of EYFP-HPV8-E2, activation of the β4-integrin promoter by JunB/Fra-1 was reduced to basal values (Fig. (Fig.7C7C).
We have previously shown that the HPV8 E2 protein downregulates β4-integrin expression in primary human keratinocytes at the transcriptional and protein level. Direct interaction between E2 and an E2-specific binding site within the regulatory region of the β4-integrin gene leads to displacement of a cellular DNA binding factor and substantially contributes to the E2-mediated suppressive effect (33). This was the first study demonstrating specific E2 binding to a cellular promoter and direct E2-mediated regulation of cellular gene transcription. However, the precise mechanism of this activity remained as yet unknown.
Here, we clarify the molecular mechanism utilized by HPV8 E2 to suppress β4-integrin promoter activity. We show for the first time that the AP-1 heterodimer JunB/Fra-1 binds to and activates the β4-integrin promoter in human keratinocytes. Our data further demonstrate that HPV8 E2 displaces JunB/Fra-1 from its binding site in the gene regulatory region, thus suppressing its transcriptional activity. We unraveled a novel viral mechanism, how HPV interferes with cellular gene expression, and a novel cellular mechanism involved in the regulation of β4-integrin expression in keratinocytes.
Several lines of evidence pointed to the direct competition of the HPV8 E2 protein with a positive regulatory cellular transcription factor as the molecular mechanism of E2-mediated β4-integrin suppression. A truncated form of E2 comprising only its C-terminal DNA binding domain, which was able to exhibit transcriptional repression (33), was also sufficient to displace this cellular factor from its cognate sequence within the β4-integrin enhancer. We demonstrate that binding of the E2-displaceable factor to the β4-integrin gene paralleled β4-integrin expression during in vitro-induced keratinocyte differentiation. A series of subsequent experiments led to the identification of the cellular E2-displaceable factor. Our experiments unraveled E2-DF as a member of the AP-1 transcription factor family.
AP-1 proteins are transcriptional regulators composed either of Jun homodimers or heterodimers of one Jun and one Fos protein. In epidermis they are expressed in a differentiation-specific manner. Fra-1 and JunB are expressed in the basal cell layer in vivo (1) and have been detected in basal keratinocytes in the present study. AP-1 factors are known to regulate keratinocyte gene expression and have important roles in many biological processes, including cell proliferation, transformation, motility, and differentiation (1, 50). The expression levels of various differentiation-specific genes, such as loricrin, filaggrin, involucrin, and transglutaminase, are controlled by AP-1 in the epidermis (16, 25, 42). Also, the expression of β4-integrin is regulated in a differentiation-specific manner. It is detected in the basal cell layer of stratified epithelia, where it attaches to laminins of the basement membrane and maintains firm association of the epidermis with the dermis. Loss of attachment to the basement membrane is a prerequisite for keratinocytes to move into suprabasal, more differentiated layers. In fact, β4-integrin is not expressed in postmitotic suprabasal keratinocytes in the epidermis (49).
The role of AP-1 in the regulation of β4-integrin expression was supported by the observation that all investigated epithelial cell lines expressing β4-integrin also displayed AP-1 DNA binding activity. Evidence that E2-DF is a member of the AP-1 family was first obtained in gel shift experiments. Sequentially truncated oligonucleotides revealed that the minimal E2-DF recognition sequence strongly resembled a TRE, which represents an AP-1 binding site (6). This novel putative AP-1 binding site differed from the TRE by the first nucleotide, which possibly explains its lower affinity. Risse et al. (41) evaluated several mutated AP-1 sites, one of which was identical to ours. It was shown to bind AP-1, albeit with lower affinity than the AP-1 consensus sequence (41), thus confirming our results. Additional studies have shown that nonconsensus AP-1 sites may be functionally active. Their sequence composition may influence the cooperation of AP-1 with other transcription factors and thus the transcriptional activity (40). Furthermore, it can be assumed that lower-affinity binding is a crucial prerequisite for AP-1 displacement by adjacent binding factors, as observed here for the HPV8 E2 protein.
In addition to the AP-1 binding site detected in the present study, other putative AP-1 binding sites have previously been proposed within the 5′ regulatory region of the β4-integrin gene as well as in the first intron (47). One of the binding sites located in the proximal promoter region accounted for ca. 35% of promoter activity in urinary bladder carcinoma cells. It was bound by c-Jun, JunB, JunD, and Fra-2 but not c-Fos subunits present in the respective nuclear extracts; Fra-1 was not examined (47). For maturing mammary epithelial cells forming acinar structures, a positive role of c-Jun/c-Fos heterodimers for β4-integrin expression has been demonstrated (32). Other integrins, such as α2 and α5, are suppressed by AP-1 during keratinocyte differentiation (9), whereas members of the β2-integrin family are upregulated during myeloid cell differentiation (28). Thus, AP-1-mediated gene regulation is a complex process, which obviously depends on the nature of the AP-1 binding site and the composition of AP-1 factors, as well as tissue-specific characteristics.
In supershift experiments the subunit composition of the E2-DF/AP-1 complex in NHK was identified as JunB and Fra-1. ChIP assays confirmed that the JunB/Fra-1 heterodimer interacts with the investigated β4-integrin promoter region in vivo. In undifferentiated normal human keratinocytes with high β4-integrin expression the β4-integrin promoter was strongly bound by the JunB/Fra-1 heterodimer, whereas both β4-integrin expression and its promoter occupancy by JunB/Fra-1 gradually decreased during differentiation. In primary human keratinocytes, overexpression of the JunB/Fra-1 heterodimer activated the β4-integrin promoter. We thus identified the AP-1 subunits JunB and Fra-1 as positive regulators of the β4-integrin gene. This correlated very well with the expression pattern of β4-integrin in basal keratinocytes of normal human epidermis and the availability of the AP-1 components JunB and Fra-1 in these cells (1). AP-1 binding within the β4-integrin promoter was almost completely lost in keratinocytes stably expressing HPV8 E2, as detected by ChIP assays. Thus, our data demonstrate a quantitative decrease in JunB/Fra-1 binding in human keratinocytes under both conditions, i.e., the differentiation and expression of HPV8 E2.
In most of the investigated cell lines, β4-integrin expression correlated with AP-1 binding activity. An exception were the SiHa cells positive for AP-1 DNA binding activity but lacking β4-integrin surface expression. Interestingly, SiHa cells also lack detectable quantities of Fra-1 (45; the present study). Low JunB and Fra-1 expression was observed in CaSki cells, which weakly express β4-integrin. These data further indicated that the expression of β4-integrin corresponds to the availability of both AP-1 subunits and supported the significance of the JunB/Fra-1 heterodimer for β4-integrin expression in human keratinocytes.
As demonstrated by gel shift experiments, the analyzed novel AP-1 binding site in the β4-integrin promoter was neighboring and not overlapping the E2 binding site. This finding is consistent with our previous observation that mutation of the E2 binding site disrupted the binding of the E2 protein to DNA but did not affect the binding of E2-DF (33). Thus, competition for adjacent binding sites and steric hindrance most presumably accounted for the inhibition of JunB/Fra-1 binding. It is tempting to speculate that similar mechanisms may occur at additional E2 binding sites which have been recently identified in cellular DNA (23).
As shown previously, HPV8 E2-mediated suppression of the β4-integrin promoter was achieved by the E2 C terminus comprising its DNA binding domain but lacking its transactivation domain (33). It was sufficient to suppress the β4-integrin promoter, albeit to a lesser extent than the full-length protein. This indicated that full-length HPV8 E2 may use additional mechanisms of transcriptional regulation apart from AP-1 displacement from the β4-integrin regulatory region. In fact, E2-mediated gene regulation can also be a consequence of direct protein-protein interactions. The E2 protein not only interacts with other HPV-encoded proteins (11, 20, 22) but also functionally cooperates with cellular transcription factors, such as p53 and C/EBP and other regulatory host cell proteins affecting cellular gene expression (21, 29, 31, 35). We have previously demonstrated that interactions with C/EBP transcription factors can lead to transactivation of C/EBP-binding site containing promoters, even in the absence of E2 binding sites. One of these targets is the promoter of involucrin, a structural protein selectively expressed in differentiating stratified epithelia (21). Moreover, E2 derived from cottontail rabbit papillomavirus (CRPV) and also from some HPVs (HPV31, -8, and -6) induces MMP-9 expression involving the extracellular signal-regulated kinase pathway and AP-1 as its downstream effector (3). Thus, the viral E2 protein may engage various mechanisms to regulate cellular gene expression.
Infection by HPVs is believed to occur through microtrauma in the stratified epithelium, exposing the basal cells to entry by viruses. α6β4 integrin is used by papillomaviruses as a membrane receptor, which may activate the antiapoptotic and progrowth signals (17). In a recently proposed model of HPV infection, laminin-5, a component of the extracellular matrix and natural ligand for α6β4 integrin, is involved in virion adsorption and transfer to α6-integrin, which facilitates further carryover to adjacent cells (10). Considering the above, one can speculate that in papillomaviral disease downregulation of β4-integrin may fulfill at least two functions. It could protect the already-infected cells from subsequent HPV infections, and it may also ensure the commitment of virus-infected cells with higher HPV8 E2 expression to leave the basal cell layer and begin the differentiation program required for completion of the viral life cycle.
In conclusion, we have identified novel cellular and viral mechanisms regulating β4-integrin expression, which play an important role during HPV infection and keratinocyte differentiation required for the HPV life cycle. First, our data show that the heterodimeric AP-1 transcription factor JunB/Fra-1 is a positive regulator of β4-integrin transcription in undifferentiated primary human keratinocytes. During keratinocyte differentiation, both JunB/Fra-1 DNA binding activity and β4-integrin expression are decreased. Second, the viral transcription factor HPV8 E2 can specifically displace this cellular factor from DNA binding as part of its inhibitory effect on β4-integrin expression. Thus, the HPV8 E2-mediated displacement of JunB/Fra-1 from the β4-integrin promoter in proliferating basal cells of stratified epithelia apparently mimics the transcriptional regulation of differentiation.
This study was supported by the DFG through the SFB670 and the Cologne Fortune program to S.S., as well as the Medical University of Warsaw grants 1M15/WB3/2006 and 1M15/WB1/07-09 to M.O. M.O. has been awarded a Young Scientist Stipend from the Foundation for Polish Science.
We thank B. Best for excellent technical assistance.
Published ahead of print on 18 November 2009.