|Home | About | Journals | Submit | Contact Us | Français|
Abortive infection of BALB/c mouse embryo fibroblasts differing in p53 gene status (p53+/+ versus p53−/−) with simian virus 40 (SV40) revealed a quantitatively and qualitatively decreased transformation efficiency in p53−/− cells compared to p53+/+ cells, suggesting a supportive effect of wild-type (wt) p53 in the SV40 transformation process. SV40 transformation efficiency also was low in immortalized p53−/− BALB/c 10-1 cells but could be restored to approximately the level in immortalized p53+/+ BALB/c 3T3 cells by reconstituting wt p53, but not mutant p53 (mutp53), expression. Stable expression of large T antigen (LT) in p53+/+ 3T3 cells resulted in full transformation, while LT expression in p53−/− 10-1 cells could not promote growth in suspension or in soft agar to a significant extent. The helper effect of wt p53 is mediated by its cooperation with LT and resides in the p53 N terminus, as an N-terminally truncated p53 (ΔNp53) could not rescue the p53-null phenotype. The p53 N terminus serves as a scaffold for recruiting transcriptional regulators like p300/CBP and Mdm2 into the LT-p53 complex. Consequently, LT affected global and specific gene expression in p53+/+ cells significantly more than in p53−/− cells. Our data suggest that recruitment of transcriptional regulators into the LT-p53 complex may help to modify cellular gene expression in response to the needs of cellular transformation.
The tumor suppressor p53 was discovered in 1979 as a cellular protein complexed to the simian virus 40 (SV40) large tumor antigen (LT) (49, 52). Based on the seeming analogy to the complex of the polyoma virus middle T-antigen with the cellular Src protein (12, 13), p53 initially was considered a cellular oncoprotein recruited by LT. This assumption was further supported by experiments demonstrating that p53 was able to immortalize certain primary cells and to cooperate with activated Ras in cellular transformation (27, 65, 72). However, these initial experiments had been performed with mutant p53 genes, and in 1989 the true nature of p53 as a tumor suppressor was established (31, 42). The discovery of p53 as a tumor suppressor not only spurred research on the role of p53 in tumorigenesis but also led to a renaissance of DNA tumor virus research, as the transforming proteins of polyoma-, papilloma-, and adenoviruses all target the tumor suppressor proteins p53 and pRb (16, 22, 39, 50, 53, 64, 92). Consistent with the functional inactivation of p53 in human and animal tumors, a wealth of evidence demonstrated that the interaction of transforming proteins with p53 also inactivated p53 function.
Cellular transformation by SV40 differs from cellular transformation by most other DNA tumor viruses insofar as SV40 LT, the major transforming protein of SV40, is able to perform both major steps of cellular transformation in vitro, immortalization and phenotypic transformation, by itself, while other DNA tumor viruses require the cooperation of at least two viral transforming proteins (11, 23, 44, 48). For example, immortalization of primary cells by the E1A protein of adenovirus or the E7 protein of human papilloma viruses (18, 69) requires the functional elimination of the proapoptotic, senescence-inducing functions of p53 by other viral proteins that inactivate p53 (e.g., its sequestration, degradation by the E1B 55K and E4orf6 proteins of adenovirus, or its degradation by the E6 protein of human papillomaviruses) (36, 68, 77, 85, 97). Thus, while it is undisputed that SV40 LT eliminates p53 tumor suppressor functions during cellular immortalization, the role of the LT-p53 complex in phenotypic transformation seems to be more complex (33, 61). In this respect, our laboratory already in 1987 provided evidence that metabolic stabilization of p53 is not simply a consequence of its physical interaction with LT (19, 20) but is an active cellular process that correlates with cellular transformation (21, 87-89, 94, 96). The finding raised the question of why the p53 protein, when destined to be functionally eliminated, should be stabilized during cellular transformation. Also, animal experiments revealed that SV40 is more efficient in promoting tumor growth when wt p53 is present (41). The suggestion that p53 in complex with SV40 LT might support SV40 phenotypic transformation has been received with due skepticism. However, recently it has been reported that human papillomavirus type 16 E6 protein mediated degradation of p53 in SV40-transformed human mesothelial cells induces a growth arrest by abrogating insulin-like growth factor 1 (IGF1) signaling (6). Those authors reported that LT and p53 associate with the IGF1 promoter within a multiprotein complex, in which p300 might be an important component required for IGF1 transcription.
In this study we analyzed the ability of SV40 virus and of LT alone to transform matched pairs of BALB/c fibroblasts differing in p53 gene status (p53+/+ cells versus p53−/− cells). Our results indicate that an N-terminal function of p53, most likely the recruitment of transcriptional modulators, like p300/CBP and Mdm2, provides a strong helper function to the LT-p53 complex during phenotypic transformation. We provide data suggesting that this recruitment is necessary for modulating global as well as local gene expression to the needs of SV40 transformation.
The BALB/c cell lines 3T3 (1) and 10-1 (37) as well as COS-1 cells used for SV40 virus stock production (34) were cultured in complete medium (Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum) maintained at 37°C or at the indicated temperature in a humidified atmosphere of 5% CO2.
The following plasmids have been described before: pCMVNc9 (mouse wild-type p53), pCMVc5 (p53 cDNA derived from MethA fibrosarcoma cells with point mutations at positions corresponding to residues 168 and 234 of the protein), and pCMVxk (control; first 15 amino acids of mouse p53) (26). pSV is based on pAT153 (29) and carries the SV40 genome with a deletion of 270 bp in the coding region of small T antigen.
Primary mouse embryo fibroblasts (MEF) were isolated according to standard methods from 17-day-old embryos of BALB/c mice (88). From heterozygous intercross matings (p53+/−) MEF were generated with p53 wild-type (+/+), heterozygous (+/−), and homozygous (−/−) genotypes (termed E1, E2, and E3, respectively). pMEF were generated from homozygous matings (p53+/+). MEF were cultured in complete medium and used for experiments at passage 4.
DNA was extracted from isolated primary MEF, and PCR was performed with p53-specific (G036/G037; 400-bp fragment) and deletion-specific (G037/G038; 600-bp fragment) primer pairs: G036, 5′-AGAGCGTGGTGGTACCTTAT-3′; G037, 5′-TATACTCAGAGCCGGCCT-3′; G038, 5′-CTATCAGGACATAGCGTTGG-3′.
pEF1αΔNtsp53 was generated by excision of ΔNtsp53 from pCIdNtsp53 (24) by NheI/NotI digestion and insertion into pEF1alphaneo (kindly provided by C. Bauer, CCS, Hamburg, Germany), which was also digested with NheI/NotI.
BALB/c 10-1 cells (1 × 105) were seeded in 10-cm-diameter plates and 24 h later transfected with 2 μg plasmid DNA using Effectene transfection reagent. After 2 weeks of selection with 500 μg/ml Geneticin (G418; Gibco) Geneticin-resistant colonies were picked and expanded into cell lines for further analysis. One clone, 10-1ΔNtsp53, was selected for further studies.
For growth analysis cells (3 × 104) were seeded in six-well plates, harvested, and counted at 24-h to 48-h intervals by using a hemocytometer. Each sample was assayed in duplicate. Population doubling time (PDT) was calculated by using an exponential regression curve between days 4 and 5. Saturation density was determined at the last day of counting.
Cells were seeded as indicated in 10-cm-diameter plates 1 day before infection. Depending on the SV40 virus stock (strain 776), cells were incubated for 2 h at 37°C with 1.5 to 2 ml SV40 per plate (multiplicity of infection [MOI] of 1, as previously described ). The virus stock was replaced by fresh complete medium and cells were subjected to Western blot analysis or colony formation at indicated time points.
Adherently growing cells were harvested, counted, and centrifuged. The pellet was resuspended in 1.5 ml SV40 virus stock and incubated for 1 h at 37°C in a water bath. After addition of 8.5 ml complete medium, cells were mixed and used in Western blot analysis or colony formation assays at the indicated time points.
For the soft agar colony formation assay, cells were harvested and counted. Single cells (numbers as indicated) were mixed thoroughly in complete medium containing 0.33% agar (end concentration; A9915; Sigma) and seeded in 10-cm-diamater plates over a bottom 0.5% agar layer. The plates were preincubated for 15 min at 4°C and then kept at the indicated temperatures. After 21 to 28 days of incubation, colonies were counted for each plate. Each soft agar assay was performed at least in triplicate. Crystal violet staining of colonies was performed by incubation with staining solution (0.25% hexamethyl-p-rosaniline chloride, 3.7% formaldehyde, 80% methanol) for 30 min at room temperature. After extensive washes with water the plates were scanned.
The vector pMYcLT was generated by excision of full-length LT cDNA from pBABE-zeo large T cDNA (Addgene) by BamHI digestion and insertion into pMYsIG (47) at the unique BamHI site. pMYsIG contains an internal ribosome entry site between the multiple-cloning site and the green fluorescent protein (GFP) coding region. To produce retroviral particles, 5 × 105 EcoPack2-293 (BD-Clontech) packaging cells were seeded in 10-cm diameter plates and 24 h later cotransfected with retroviral vector plasmid DNA (3.2 μg), either pMYsIG or pMYcLT and a plasmid expressing the ecotropic Env protein (3 μg) (63), using Lipofectamine 2000. The medium was changed 4 h later to complete medium. The cells were incubated at 37°C, and the media were harvested after 24 h and 48 h. The supernatant containing the retrovirus was filtered through a 0.22-μm filter and stored at −80°C.
3T3, 10-1, and 10-1ΔNtsp53 cells (1 × 105) were seeded in six-well plates and cultured for 6 h in 2 ml complete medium. For infection, the medium was replaced with 1-ml viral aliquots diluted in 2 ml complete medium supplemented with 8 μg/ml Polybrene (Sigma) and added to cell monolayers. After centrifugation of the six-well plate for 1 h at 700 × g and 20°C and incubation at 37°C for 24 h, the medium was replaced with fresh medium. The transduced cells were sorted 24 h later for GFP-expressing cells with a fluorescence-activated cell sorter (Aria Flow Cytometer; Becton Dickinson) and termed 3T3 LT, 10-1 LT, and ΔNp53 LT cells (pMYcLT) or 3T3 IG, 10-1 IG, and ΔNp53 IG cells (pMYsIG). The cells were maintained in complete medium.
Monoclonal antibodies PAb240, PAb246, and PAb248 (hybridoma supernatants), a polyclonal antiserum (sheep), and the polyclonal antibody FL393 (sc-6243-G) obtained from Santa Cruz Biotechnology were used to detect p53. For LT detection the following mouse monoclonal antibodies were used: KT3 (ascites) and also PAb108, PAb419, and PAb416 (hybridoma supernatants). Anti-Ac-Lys (9441) was purchased from Cell Signaling, anti-α-tubulin (Ab-1; CP06) was from Oncogene, and anti-GFP (11 814 460 0019) was obtained from Roche.
Cells were washed with ice-cold phosphate-buffered saline (PBS), harvested, and either lysed immediately or stored at −80°C. Pellets were lysed for 30 min on ice with mixing in E1A lysis buffer (50 mM HEPES/KOH, pH 7.4, 150 mM NaCl, 0.1% NP-40) supplemented with proteinase inhibitors, 5 μg/ml leupeptin, 10 μg/ml pepstatin A, 250 μg/ml Pefablock SC, and 1% Trasylol, and clarified by centrifugation for 30 min at 4°C. The protein concentration was determined by the Bradford assay (Bio-Rad). Equal amounts of protein were separated by sodium dodecyl sulfate-polyacrylamide electrophoresis (SDS-PAGE) gels (see figure legends) and transferred onto polyvinylidene membranes (Immobilon-P; Millipore). Following blocking with 5% skim milk-Tris-buffered saline with 0.1% Tween 20 (TBS-T) for 1 h, the membranes were incubated with primary antibody in TBS-T for 1 h at room temperature or overnight at 4°C. After washing the membranes were incubated with the respective horseradish peroxidase-conjugated secondary antibodies (Rockland) for 1 h. Protein bands were visualized with the enhanced chemiluminescence reagent (Pierce Biotechnology) and by exposure to an X-ray film.
Cell lysis was as described above in E1A buffer supplemented with proteinase inhibitors and addition of 4 mM EGTA, 1 mM EDTA, and 5 mM NaF. Equal amounts of protein were immunoprecipitated with the specific primary antibody for 1 to 2 h at 4°C. Immunoprecipitates were collected with protein A-Sepharose for 1 h and washed three times with E1A buffer. The pellet was resuspended in 2× Laemmli buffer (with 200 mM dithiothreitol), heated for 5 min at 95°C, and analyzed by electrophoresis.
Cells were grown on coverslips and fixed in −20°C acetone. Following rehydration in PBS and blocking with 1% normal donkey serum (10 to 30 min), cells were incubated with specific primary antibodies for 30 min to 1 h in a humidified chamber in PBS. After washing for 15 min in PBS cells were incubated with Alexa 488- and Alexa 555-labeled secondary antibodies (Molecular Probes) for 30 min and washed again in PBS, and the nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI) for 10 min. Slides were mounted in Mowiol and images were captured and acquired on Leica DMRA microscope.
Total RNA was isolated from cells using TRIzol (Invitrogen) and digested with RNase-free DNase I (Qiagen) according to the manufacturer's instructions. The quality and integrity of the total RNA were evaluated with an Agilent 2100 Bioanalyzer, and the concentration was measured using a NanoDrop spectrophotometer. Microarray hybridization was performed on One-Color Agilent whole-mouse genome oligo microarray chips. The Agilent microarray chip contains 41,534 60-mer oligonucleotide probes representing ~41,000 known genes. Synthesis of cRNA, hybridization, and image data processing were completed by imaGenes GmbH (Berlin, Germany). Microarray data were normalized and analyzed using dChip software (http://www.dChip.org/). A gene was considered to be differentially expressed if a Student's t test statistic had an associated P value of ≤0.05 and a fold change of ≥2.
For quantitative reverse transcription-PCR (qRT-PCR), reverse transcription of 1 μg RNA was performed with the High Capacity cDNA reverse transcription kit (Applied Biosystems) in the presence of random primers in a total volume of 20 μl. Five microliters of cDNA (1:5 diluted) was subjected to real-time qPCR analysis performed with 10 μl Power SYBR green PCR master mix (Applied Biosystems) and forward and reverse primers (200 nM) in a final volume of 20 μl, using the 7500 Fast real-time PCR system (Applied Biosystems). All reactions were carried out in triplicate. The cycling profile was 95°C for 10 min for preincubation, followed by 40 cycles of 95°C for 15 s and 60°C for 1 min. Appropriate nontemplate controls were included in each run, and dissociation analysis was performed at the end of each run to confirm the specificity of the PCR. Primer sets were used as follows: Igf1, forward, 5′-AAAATCAGCAGCCTTCCAACT-3′, and reverse, 5′-GTCTCTGGTCCAGCTGTGGT-3′ (murine Igf1; accession number NM_010512); Gapdh, forward, 5′-GGTGAAGGTCGGTGTGAAC-3′, and reverse, 5′-GGGGTCTCGCTCCTGGAA-3′ (murine Gapdh; accession number NM_008084). Standard curves were generated for each primer set with triplicate fivefold dilutions of one sample to determine the PCR efficiency, which was close to 100%. Analysis was carried out with the 7500 system software v1.4. Ct (threshold cycle) values were determined using the automatic baseline setting. Ct values for Igf1 were between 26 and 30. Igf1 expression was normalized to the endogenous control, Gapdh. Relative gene expression (RQ) was obtained by using the comparative 2−ΔΔCt method (78). The standard error of the mean of the expression levels is represented by the calculated minimum (RQ Min) and maximum (RQ Max) expression levels (error bars), based on a 95% confidence level.
The microarray data sets discussed in the manuscript have been deposited in NCBI's Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/) and are accessible through GEO series accession number GSE17178.
MEF were prepared from day 17 mouse embryos of p53+/+, p53+/−, and p53−/− genotypes (see Materials and Methods). p53+/+ and p53+/− MEF grew well up to passage 5 but went into crisis starting at passage 6. As previously reported, p53−/− MEF did not undergo crisis but immortalized spontaneously (38). Primary cultures of MEF of all genotypes showed a low efficiency of infection by SV40 (<20% infected cells) compared to established 3T3 cells (>90%), which increased to ~40 to 50% after four passages. Furthermore, primary cultures of p53−/−, p53+/−, and p53+/+ MEF showed a low rate of spontaneous colony formation in soft agar (~60 colonies/106 cells for p53−/− MEF and ~20 colonies/106 cells for p53+/− and p53+/+ MEF), which decreased to less than 10 colonies/106 cells in passage 4. Therefore, we used MEF of all p53 genotypes in passage 4 in transformation assays. Cells were abortively infected with SV40, and colony formation in soft agar was determined as described in Materials and Methods. Table Table11 shows that MEF of a p53+/+ (pMEF and E1) and a p53+/− genotype (E2) were transformed by SV40 with an efficiency of ~5 × 10−4, while SV40 was unable to transform p53−/− MEF (E3) to a significant extent.
As the “spontaneous” soft agar colonies formed by MEF might obscure the evaluation of SV40 transformation efficiencies described above, we analyzed the growth potential of “spontaneous” colonies as well as of colonies arising from SV40 transformation of p53+/+ and p53−/− MEF. None of the “spontaneous” colonies arising from p53+/+ MEF could be expanded and established as cultures, suggesting that these cells were in a senescent state. In contrast, a significant fraction (6/11) of the colonies arising from p53−/− MEF could be expanded in culture, indicating that these cells had become immortalized, but not transformed, as they were not able to grow efficiently in soft agar.
Cells from colonies expanded from SV40-infected p53+/+ MEF all expressed SV40 LT as analyzed by immunofluorescent staining and exhibited a fully transformed phenotype (E1SV cells), with a high efficiency to form colonies in soft agar after reseeding (~50%). In contrast, expansion of colonies from SV40-infected p53−/− MEF proved to be rather inefficient, and only one SV40 LT-expressing culture could be established (E3SV cells). When reseeded in soft agar, E3SV cells (Fig. (Fig.1A)1A) showed ~3-fold-lower efficiency to form large colonies in soft agar than p53+/+ MEF-derived E1SV cells (Fig. (Fig.1B),1B), although the total number of colonies was similar. All E1SV colonies showed a compact morphology (Fig. (Fig.1C),1C), whereas E3SV cells mostly formed fluffy colonies (Fig. (Fig.1D).1D). Beyond that, E1SV colonies reseeded onto plastic dishes rapidly formed transformed foci (Fig. (Fig.1E),1E), whereas E3SV cells grew slowly and never formed foci (Fig. (Fig.1F).1F). The data indicate that SV40 transformation of cells expressing p53 results in full transformation, whereas in cells deficient in p53 expression a fully transformed phenotype is not achieved during either spontaneous immortalization or after SV40 infection.
The experiments described above support the assumption that wt p53 exerts a helper effect in SV40 transformation. However, primary MEF constitute a complex system to analyze transformation due to their heterogeneity and their inherent genetic and epigenetic instabilities, especially if they originate from p53−/− mice (15, 32, 91). As the necessity for eliminating the growth control functions of p53 in cellular immortalization is undisputed (37, 71), we focused on the role of p53 in phenotypic transformation of immortalized MEF by SV40. As a cell system we employed spontaneously immortalized p53−/− 10-1 cells (37). 10-1 cells exhibit a normal phenotype with regard to transformation-relevant parameters, like serum dependence of growth, saturation density, contact inhibition, their inability to form colonies in soft agar, and their lack of tumor formation in nude mice (data not shown). In this respect 10-1 cells are similar to p53+/+ BALB/c 3T3 cells, a cell line often used in SV40 transformation experiments and employed in our experiments as controls.
10-1 and 3T3 cells were abortively infected with SV40 (efficiency of infection of 60%, the highest one achievable in 10-1 cells). Although the infected 10-1 and 3T3 cells expressed LT to comparable levels (Fig. (Fig.2A),2A), SV40 did not significantly support growth of infected 10-1 cells in soft agar (Table (Table2).2). In contrast, SV40 infection of 3T3 cells, as expected, strongly enhanced the outgrowth of colonies of transformed cells in soft agar, resulting in a transformation efficiency of ~1 × 10−4 (8.5 × 10−5 to 1.5 × 10−4 in different experiments).
To support the assumption that the deficiency of 10-1 cells for SV40 transformation is due to lack of p53 expression, we reconstituted p53 expression in these cells. As overexpression of wt p53 in 10-1 cells leads to growth arrest and apoptosis (data not shown), cells were first abortively infected in suspension with SV40, which then could “neutralize” the growth-suppressive effects of a subsequently transfected wt p53. Infected cells were reseeded onto culture plates and then transfected with vector pCMVNc9, encoding wt p53, and vector pCMVxk as a control. To discriminate between specific effects of the coexpressed wt p53 associating with SV40 LT, and possible unrelated effects of a mutant p53 with gain-of-function properties, 10-1 cells were also transfected with vector pCMVc5, encoding the oncogenic mouse mutp53D168G/M234I (MethA) protein (17, 25). MethA mutp53 does not significantly interact with SV40 LT (95). Immunofluorescence analysis for LT and p53 expression demonstrated that ~60% of the cells were infected with SV40 and that ~30% of the cells coexpressed the transfected p53 genes. Due to the instability of wt p53, wt p53 and mutp53 levels were significantly different in the transfected cells (Fig. (Fig.2B).2B). However, despite being present at much higher levels, MethA mutp53 in SV40-infected 10-1 cells was unable to enhance the transformation efficiency of SV40 above that of the control vector. In contrast, transformation efficiency was strongly enhanced in SV40-infected 10-1 cells that received the wt p53-expressing vector (Table (Table3).3). Importantly, SV10-1 (wt p53) cells isolated from soft agar colonies exhibited a similar transformed phenotype as SV3T3 cells isolated in a control experiment (Table (Table4).4). wt p53 thus is able to cooperate with LT in cellular transformation of 10-1 cells. Considering that only 30% of the cells coexpressed SV40 LT and p53, the transformation efficiency of SV40 in 10-1 cells cotransfected with a wt p53-encoding vector closely approached that of SV40 in 3T3 cells. The data thus demonstrate that ectopic expression of wt p53 in 10-1 cells is able to rescue the SV40 transformation deficiency of these cells. An additional important finding was that mutp53 gain of function seems to be unrelated to the wt p53 helper function in SV40 transformation. However, we cannot exclude that the inability of MethA mutp53 to physically interact with SV40 LT precludes such a helper function.
As a reverse control, we intended to knock down p53 expression in 3T3 cells by using an RNA interference approach. However, even a p53 knockdown at the mRNA level of >80% did not completely prevent accumulation of p53 protein after UV irradiation of the cells (data not shown). Considering the tight physical interaction of p53 with LT and its ensuing metabolic stabilization during transformation, we considered this approach impracticable.
The SV40 early gene region in addition to LT encodes the small t and 17kT proteins, which cooperate with LT in SV40 transformation (8, 73). It thus is conceivable that the p53 helper function results from cooperation of p53 with all early SV40 proteins. In order to specifically analyze the cooperation of p53 with LT, we created 10-1-derived (p53−/−) and 3T3-derived (p53+/+) cell lines stably expressing LT only. As shown above, 3T3 cells can be used in such experiments as positive controls, as with SV40 transformation 3T3 cells correspond to 10-1 cells with reconstituted wt p53 expression.
For generation of the respective cells we employed retroviral transduction using the vector pMYcLT, based on the vector pMYsIG (47) (see Materials and Methods for details). pMYcLT in addition to LT also encodes GFP, with an internal ribosome entry site separating the two genes. GFP expression thus could be used to enrich successfully transduced cells by fluorescence-activated cell sorting to a purity of over 90%.
Figure Figure3A3A shows the growth curves of the different LT-expressing cells and their corresponding controls. It is evident that both adherent 3T3 LT and 10-1 LT cells grew to higher saturation densities than their respective control cells. An important difference between 3T3 LT and 10-1 LT cells was that 3T3 LT cells at high density formed three-dimensional foci, while 10-1 LT cells failed to do so (Fig. (Fig.3B,3B, panels a and b). The insert in Fig. Fig.3A3A shows the PDT for the different cells, revealing that LT expression shortened the PDT in all LT-expressing cells from ~33 h to ~25 h.
Next we analyzed the ability of the cells to form transformed foci when seeded at low cell density (103 cells per 10-cm-diameter plate). LT-expressing 3T3 cells formed more foci than 3T3 IG- and LT-expressing 10-1 cells which, however, did not form more foci than control 10-1 IG cells (Fig. (Fig.3C).3C). A more decisive criterion differentiating focus formation in 3T3 LT and 10-1 LT cells than quantity was the quality of the foci. While 3T3 LT cells grew three-dimensional foci, foci of 10-1 LT cells were “flat” and seldom consisted of more than one layer of cells.
The most stringent parameter for phenotypic transformation in vitro is anchorage-independent growth. 10-1 LT and 3T3 LT cells were subjected to growth in suspension by seeding the cells in NUNC low cell binding (LCB) plates, whose surface prevents cell attachment. Already, visual inspection of the culture plates after 14 days showed that 10-1 LT cells had formed many fewer swimming colonies than 3T3 LT cells. In addition, 10-1 LT colonies were much smaller in size and had a “pancake-like” structure, while colonies of 3T3 LT cells were bigger and grew as “spheres” (Fig. (Fig.4A).4A). The insert in Fig. Fig.4A4A provides a semiquantitative evaluation of the size of colonies of 3T3 LT and of 10-1 LT cells in suspension culture.
Next, 10-1 LT and 3T3 LT cells were subjected to transformation assays in soft agar. While neither 3T3 IG cells nor 10-1 IG cells formed transformed colonies in soft agar, LT allowed colony formation in soft agar in both 3T3 LT and 10-1 LT cells. However, the efficiency for colony formation of 10-1 LT cells was strongly reduced compared to that of 3T3 LT cells. In addition, colonies formed by 3T3 LT cells were significantly larger than colonies formed by 10-1 LT cells. Table Table55 and Fig. Fig.4B4B provide a semiquantitative evaluation of frequency of colony formation and colony size for 3T3 LT and 10-1 LT cells.
In summary, the data demonstrate that in the absence of p53 SV40 LT can change the phenotype of immortalized mouse fibroblasts only to a minimal transformed phenotype, while full transformation is obtained in the presence of p53.
We next investigated which domain of p53 is required for the p53 helper function. As LT targets the p53 core domain, leading to inactivation of its DNA binding properties (4, 45, 51, 81), the N- and C-terminal domains of p53 are largely unaffected by the interaction with LT (7). Both domains serve as platforms for the assembly of p53-interacting proteins (46) and thus could be potentially involved in the p53 helper function. We reckoned that the p53 N-terminal transactivation domain 1 (TAD1) might be of special importance, as this domain is required for recruiting p53-interacting proteins involved in transcriptional regulation, e.g., Mdm2 and p300/CBP, into the LT-p53 complex (7, 30, 40). Therefore, we constructed vector pEF1αΔNtsp53 driving expression of tsp53 lacking the first 40 amino acids of the full-length protein on the basis of the tsp53 gene contained in vector pLTRcG9 (see Materials and Methods for details). At 32°C, the encoded ΔNtsp53 corresponds to the naturally occurring 47-kDa ΔNp53 variant (14, 93). The tsp53val135 mutation in ΔNtsp53 would allow us to establish cells stably expressing ΔNtsp53 at the nonpermissive temperature, in case (overexpressed) ΔNtsp53 should exert growth inhibitory effects, as it still contains the p53 TAD2. 10-1 cells were transfected with vector pEF1αΔNtsp53. After clonal selection, only 1 out of 15 tested G418-resistant colonies expressed ΔNtsp53 (Fig. (Fig.5A).5A). Expression of LT by transfecting 10-1ΔNtsp53 cells at 32°C with the LT-encoding vector pSV showed that LT complexes with ΔNtsp53 (Fig. (Fig.5B),5B), thereby allowing the analysis of the contribution of the p53 N terminus to the p53 helper function in the LT-p53 complex. By transducing 10-1ΔNtsp53 cells with vector pMYcLT and selecting for GFP-expressing cells we established ΔNp53 LT cells. Figure Figure6A6A shows the comparison of ΔNp53 LT, 10-1 LT, and 3T3 LT cells at 32°C with regard to LT, GFP, p53, and ΔNtsp53 expression, respectively. Comparison of 10-1 LT, 3T3 LT, and ΔNp53 LT cells and ΔNp53 IG control cells for growth parameters (saturation density and PDT) (Fig. (Fig.6B)6B) and growth in suspension in LCB plates (Fig. (Fig.6C),6C), as well as their ability to form colonies in soft agar (Fig. (Fig.6D;6D; Table Table6)6) revealed that ΔNp53 LT cells were indistinguishable from 10-1 LT cells, thereby demonstrating that ΔNtsp53 in ΔNp53 LT cells was unable to rescue the p53-null phenotype of 10-1 LT cells.
Considering that the SV40 transformation efficiency of 10-1(wt p53) was similar to that of 3T3 cells (Tables (Tables22 and and3),3), we conclude that the p53 helper function for LT-mediated phenotypic transformation is provided by the p53 N terminus.
Acetylation of LT at lysine residue K697 is dependent on p53 and mediated by p300/CBP (66). As p300/CBP associate with p53 via the p53 N terminus, LT should be acetylated in 3T3 LT cells, but not in 10-1 LT and ΔNp53 LT cells. Figure 7A and B show that a fraction of LT is indeed acetylated in 3T3 LT but not in 10-1 LT cells and that, in contrast to p53 in 3T3 LT cells, ΔNp53 in ΔNp53 LT cells is not acetylated. The data provide further evidence that the N terminus of p53 recruits p300/CBP into the LT-p53 complex and is responsible for LT and p53 acetylation in 3T3 LT cells.
LT possesses an N-terminal transactivation domain (98) and is able to associate with cellular chromatin and the nuclear matrix (43, 70, 84). LT thus fulfills the criteria for a transcriptional regulator. Indeed, modulation of cellular transcription by SV40 LT has long been observed and correlates with cellular transformation (58, 59, 80, 90). Recruitment of transcriptional regulators like Mdm2 and p300/CBP into the LT-p53 complexes via their association with the p53 N terminus should expand the transcriptional potential of LT and therefore provide an additional means for tuning cellular gene expression to the needs of cellular transformation.
To measure on a global scale the LT-mediated changes in gene transcription, samples of total RNA from 3T3 LT, 10-1 LT, and control cells (3T3 IG and 10-1 IG) were subjected to gene expression analysis on Agilent whole mouse genome microarray chips. We compared the transcription profiles of LT-expressing cells with the respective control cells and found that the number of regulated genes differed significantly between 3T3 LT and 10-1 LT cells. Whereas in 3T3 cells 861 genes were regulated with changes higher than twofold, only 181 genes in 10-1 LT cells satisfied this criterion (Fig. (Fig.8A8A and data not shown). In both cell lines ~80% of all differentially expressed genes were downregulated, indicating that LT exerts mainly repressive effects. The minimal overlap between the lists of the differentially expressed genes strongly supports our hypothesis that the transformation-relevant transcriptional activity of LT is dependent on p53. As expected for SV40-transformed rodent embryo fibroblasts (74), the transformed phenotype of 3T3 LT cells is associated with a significant downregulation of a number of genes encoding extracellular matrix components, including collagens, fibronectin, cadherins, and proteoglycans (Table (Table77).
Recently, Bocchetta et al. demonstrated that the Igf1 gene is a transcriptional target of the LT-p53-p300 complex (6), allowing us to use the activity of the Igf1 gene to track the presumed transcriptional cooperation between p53 and LT in SV40-transformed BALB/c embryo fibroblasts. We found that in control 3T3 cells Igf1 transcription was strongly affected by cell density (data not shown) and serum content in the culture medium. We determined by quantitative real-time PCR a 10-fold stronger Igf1 transcription in 3T3 cells cultured for 2 days at ~70% confluence in 1% fetal calf serum (FCS) compared to cells grown in the presence of 10% FCS (Fig. (Fig.8B).8B). However, this dependence between Igf1 transcription and reduced serum content was significantly attenuated in SV40-transformed BALB/c 3T3 (SV3T3) and p53+/+ MEF (E1SV) cells, as well as in SV40-transformed 10-1 cells with reconstituted wt p53 expression [SV10-1(wt p53) cells] (Fig. (Fig.8B).8B). In contrast, in the absence of p53 the effect of LT alone on Igf1 transcription was marginal (E3SV cells) (Fig. (Fig.8B),8B), indicating a cooperation between p53 and LT in attenuating Igf1 gene activity in SV40-transformed mouse fibroblasts.
The transforming activity of SV40 LT mainly results from its interaction with cellular proteins (2). Some of these proteins are tumor suppressors, like pRB and p53, which are functionally inactivated by their interaction with LT. In addition, LT interacts with a variety of other proteins, like Mdm2, p300/CBP, Cul7, Bub1, TEF-1, Nbs1, and Fbw7, as well as hsc70. These interactions could further modulate the transforming activities of LT. LT-mediated cell transformation thus is an extremely complex process which further is strongly influenced by the cellular context chosen for transformation (2).
In this report we further pursued the idea that elimination of p53 function by SV40 LT is only one aspect of the tight interaction of these proteins. The idea was originally conceived from the stimulatory effect of cotransfected p53 on SV40-induced transformation of rat embryo fibroblasts (61) and our observation that metabolic stabilization of p53 in complex with LT not simply is the result of its physical interaction with LT but rather a cellular process tightly connected to the acquisition of a transformed phenotype (20). We have shown here that SV40 exerts a strongly enhanced transforming activity in p53+/+ MEFs compared to their p53-null counterparts, leading to a fully transformed phenotype which is never reached by SV40-infected p53-deficient cells. Immortalized p53-deficient mouse 10-1 fibroblasts were almost deficient for transformation by SV40, but transformation competence could be fully restored by reconstituting wt p53 expression in these cells. Thus, in two different, close-to-isogenic cell systems, full transformation strongly depended on the p53 gene status and thus on a helper function provided by p53. This helper function was only exerted by wt p53, but not by mutp53, because it is either dependent on a physical interaction of p53 with LT or constitutes a function exerted only by wt p53.
The helper function of p53 in LT-mediated cellular transformation resides in its N-terminal TAD1, as truncation of this domain generated a ΔNp53 that still complexes LT but failed to exert this helper function when ectopically expressed in 10-1 cells. The TAD1 domain serves as a scaffold for the interaction of p53 with a variety of cellular proteins involved in transcriptional regulation of p53 (46). Of these, the histone acetyltransferases p300 and CBP are of special importance, as the interaction of p300/CBP with p53 and LT leads to acetylation of both these proteins.
Association of p300/CBP with p53 and p53 acetylation are closely linked to p53 transcriptional activity. It is assumed that recruitment of these histone acetyltransferases is required for acetylation of p53-bound histones in the promoter of a p53 target gene (28) and for further recruitment of coactivators/histone acetyltransferases to the p53 transcription site (5), thereby enhancing the transcriptional activity of p53. The only consequence of the association of p300/CBP with the LT-p53 complex known so far is the acetylation of lysine at position K697 on LT. It has been suggested that acetylation destabilizes LT and thereby negatively regulates transformation (82). However, we observed similar levels of LT in 3T3 LT and in 10-1 LT cells (Fig. (Fig.2),2), suggesting that the stability of acetylated LT in 3T3 LT and nonacetylated LT 10-1 LT cells does not grossly differ. More importantly, we demonstrated that acetylated LT in 3T3 LT cells has a much higher transformation competence than LT in 10-1 LT cells, where LT is not acetylated. The finding at first glance would suggest that acetylation of LT positively correlates with its transformation competence. However, K697 is located in the LT host range region, which is neither required for cellular transformation in vitro (55, 86) nor for tumor formation in mice (79). Our data therefore suggest that acetylation of LT is not directly involved in the transforming activity of LT.
In a search for a role of the recruitment of p300/CBP and other transcriptional regulators by LT via complex formation with p53, we compared the effects of LT on global gene expression in 3T3 and in 10-1 cells. The data indicated that recruitment of p53 by LT strongly enhances global transcriptional regulation by LT. An important observation is that about 80% of the differentially regulated genes in both cell lines were downregulated, indicating that LT mainly acts as a repressor. In light of the major role of p300/CBP as a transcriptional coactivator, the pronounced repressive effects of the LT-p53 complex may seem somewhat surprising. However, accumulating evidence demonstrates that p300/CBP also can act as a repressor (3, 35, 75, 76) either by its endogenous chromatin-specific repressor function (76) or in cooperation with other proteins, such as YY1 and HDAC3 (75). That p300/CBP by itself can provide a protein scaffold upon which a multicomponent transcriptional regulatory complex can be built (9) highlights the complexity of p300/CBP recruitment to the LT-p53 complex. The fact that p53 bridges and stabilizes the complex between LT and p300/CBP (7) might explain the diminished repressive effects of LT in 10-1 LT cells.
The LT-p53 complex not only recruits p300/CBP but also other transcriptional regulators, like Mdm2 (40). Mdm2 not only directly inhibits p53 transcriptional activity by binding to its transactivation domain but also can exert transcriptional repression by histone monoubiquitination via its RING domain (62). Taken together, the recruitment of p53 by LT strongly expands the transcriptional potential of LT, thereby allowing efficient transformation.
In addition to global modulation of gene expression, modulation of the expression of specific genes is important for understanding gene regulation by the LT-p53 complex. However, analysis of the expression of genes specifically regulated by LT or the LT-p53 complex so far had been hampered by the virtual lack of bona fide cellular LT and LT-p53 target genes. Although LT can bind SV40 DNA in a sequence-specific manner (54, 60, 83), it is still unclear whether sequence-specific binding to cellular DNA by LT plays a role in transformation, as LT mutants defective in binding to the SV40 origin of replication transform cells with the same efficiency as wild-type LT (10, 55, 56, 67). Thus, in contrast to p53 target genes, possible LT target genes in cellular transformation cannot be detected by sequence homology of possible LT response elements. Recently, however, Bocchetta et al. (6) demonstrated that the IGF1 gene is a direct transcriptional target of the LT-p53 complex, which positively regulates IGF1 transcription in SV40-transformed human mesothelial cells. Our analyses of Igf1 gene activity in SV40-transformed mouse fibroblasts contradict this result. Igf1 transcription was significantly upregulated in 3T3 cells grown in 1% FCS but attenuated in SV3T3, E1SV (p53+/+), and SV10-1 (wt p53) cells compared to 3T3 cells, whereas Igf1 gene expression in E3SV (p53−/−) cells was only marginally affected by expression of LT, indicating a specific repressive function of the LT-p53 complex. It remains to be elucidated how the cellular context (human mesothelial cells compared to the mouse fibroblasts analyzed here) results in opposite effects on Igf1 (IGF1) transcription and to clarify the role of the LT-p53 complex in this regulation. It is conceivable that elimination of p53 by ectopic expression of the HPV16 E6 protein may account for the observed discrepancy. Although the E6 protein induces rapid degradation of p53, this is only one of many of its functions (reviewed in reference 57). Despite these differences, however, both sets of data show that the Igf1 (IGF1) gene is a target of the LT-p53 complex.
In summary, our data show a novel facet of the interaction of SV40 LT with p53: while elimination of p53's growth-suppressing functions through complex formation with LT has long been demonstrated, we have shown here that p53 complexed to LT in addition performs a helper function in SV40 transformation. Our data support the hypothesis that the p53 helper function is exerted by transcriptional regulators recruited by the p53 N terminus and that the additional transcriptional potential provided to LT endows the LT-p53 complex with the transcription/repression activity required for efficient transformation.
We thank Timo Quante for crucial advice about in expression analysis and Daniel Speidel for helpful discussions. We are grateful to Carol Stocking for providing the pMYsIG construct.
The work described herein is part of the Ph.D. thesis by A.H.
This work was funded by the Deutsche Forschungsgemeinschaft (DE 212/23-3), the Fonds der Chemischen Industrie, and by EC FP6 funding. The Heinrich-Pette-Institute is financially supported by the Freie und Hansestadt Hamburg and the Bundesministerium für Gesundheit.
The publication reflects only the authors' views, and the Community is not liable for any use that may be made of the information contained therein.
Published ahead of print on 22 July 2009.